Adding upstream version 6.1.76.upstream/6.1.76

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

714 files changed, 356197 insertions, 0 deletions

diff --git a/drivers/gpu/drm/i915/.gitignore b/drivers/gpu/drm/i915/.gitignore
new file mode 100644
index 000000000..81972dce1
--- /dev/null
+++ b/drivers/gpu/drm/i915/.gitignore
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0-only
+*.hdrtest
diff --git a/drivers/gpu/drm/i915/Kconfig b/drivers/gpu/drm/i915/Kconfig
new file mode 100644
index 000000000..6b10868ec
--- /dev/null
+++ b/drivers/gpu/drm/i915/Kconfig
@@ -0,0 +1,173 @@
+# SPDX-License-Identifier: GPL-2.0-only
+config DRM_I915
+	tristate "Intel 8xx/9xx/G3x/G4x/HD Graphics"
+	depends on DRM
+	depends on X86 && PCI
+	depends on !PREEMPT_RT
+	select INTEL_GTT if X86
+	select INTERVAL_TREE
+	# we need shmfs for the swappable backing store, and in particular
+	# the shmem_readpage() which depends upon tmpfs
+	select SHMEM
+	select TMPFS
+	select DRM_DISPLAY_DP_HELPER
+	select DRM_DISPLAY_HDCP_HELPER
+	select DRM_DISPLAY_HDMI_HELPER
+	select DRM_DISPLAY_HELPER
+	select DRM_KMS_HELPER
+	select DRM_PANEL
+	select DRM_MIPI_DSI
+	select RELAY
+	select IRQ_WORK
+	# i915 depends on ACPI_VIDEO when ACPI is enabled
+	# but for select to work, need to select ACPI_VIDEO's dependencies, ick
+	select BACKLIGHT_CLASS_DEVICE if ACPI
+	select INPUT if ACPI
+	select X86_PLATFORM_DEVICES if ACPI
+	select ACPI_WMI if ACPI
+	select ACPI_VIDEO if ACPI
+	select ACPI_BUTTON if ACPI
+	select SYNC_FILE
+	select IOSF_MBI if X86
+	select CRC32
+	select SND_HDA_I915 if SND_HDA_CORE
+	select CEC_CORE if CEC_NOTIFIER
+	select VMAP_PFN
+	select DRM_TTM
+	select DRM_BUDDY
+	select AUXILIARY_BUS
+	help
+	  Choose this option if you have a system that has "Intel Graphics
+	  Media Accelerator" or "HD Graphics" integrated graphics,
+	  including 830M, 845G, 852GM, 855GM, 865G, 915G, 945G, 965G,
+	  G35, G41, G43, G45 chipsets and Celeron, Pentium, Core i3,
+	  Core i5, Core i7 as well as Atom CPUs with integrated graphics.
+
+	  This driver is used by the Intel driver in X.org 6.8 and
+	  XFree86 4.4 and above. It replaces the older i830 module that
+	  supported a subset of the hardware in older X.org releases.
+
+	  Note that the older i810/i815 chipsets require the use of the
+	  i810 driver instead, and the Atom z5xx series has an entirely
+	  different implementation.
+
+	  If "M" is selected, the module will be called i915.
+
+config DRM_I915_FORCE_PROBE
+	string "Force probe i915 for selected Intel hardware IDs"
+	depends on DRM_I915
+	help
+	  This is the default value for the i915.force_probe module
+	  parameter. Using the module parameter overrides this option.
+
+	  Force probe the i915 driver for Intel graphics devices that are
+	  recognized but not properly supported by this kernel version. Force
+	  probing an unsupported device taints the kernel. It is recommended to
+	  upgrade to a kernel version with proper support as soon as it is
+	  available.
+
+	  It can also be used to block the probe of recognized and fully
+	  supported devices.
+
+	  Use "" to disable force probe. If in doubt, use this.
+
+	  Use "<pci-id>[,<pci-id>,...]" to force probe the i915 for listed
+	  devices. For example, "4500" or "4500,4571".
+
+	  Use "*" to force probe the driver for all known devices. Not
+	  recommended.
+
+	  Use "!" right before the ID to block the probe of the device. For
+	  example, "4500,!4571" forces the probe of 4500 and blocks the probe of
+	  4571.
+
+	  Use "!*" to block the probe of the driver for all known devices.
+
+config DRM_I915_CAPTURE_ERROR
+	bool "Enable capturing GPU state following a hang"
+	depends on DRM_I915
+	default y
+	help
+	  This option enables capturing the GPU state when a hang is detected.
+	  This information is vital for triaging hangs and assists in debugging.
+	  Please report any hang for triaging according to:
+	    https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs
+
+	  If in doubt, say "Y".
+
+config DRM_I915_COMPRESS_ERROR
+	bool "Compress GPU error state"
+	depends on DRM_I915_CAPTURE_ERROR
+	select ZLIB_DEFLATE
+	default y
+	help
+	  This option selects ZLIB_DEFLATE if it isn't already
+	  selected and causes any error state captured upon a GPU hang
+	  to be compressed using zlib.
+
+	  If in doubt, say "Y".
+
+config DRM_I915_USERPTR
+	bool "Always enable userptr support"
+	depends on DRM_I915
+	select MMU_NOTIFIER
+	default y
+	help
+	  This option selects CONFIG_MMU_NOTIFIER if it isn't already
+	  selected to enabled full userptr support.
+
+	  If in doubt, say "Y".
+
+config DRM_I915_GVT_KVMGT
+	tristate "Enable KVM host support Intel GVT-g graphics virtualization"
+	depends on DRM_I915
+	depends on X86
+	depends on 64BIT
+	depends on KVM
+	depends on VFIO_MDEV
+	select DRM_I915_GVT
+	select KVM_EXTERNAL_WRITE_TRACKING
+
+	help
+	  Choose this option if you want to enable Intel GVT-g graphics
+	  virtualization technology host support with integrated graphics.
+	  With GVT-g, it's possible to have one integrated graphics
+	  device shared by multiple VMs under KVM.
+
+	  Note that this driver only supports newer device from Broadwell on.
+	  For further information and setup guide, you can visit:
+	  http://01.org/igvt-g.
+
+	  If in doubt, say "N".
+
+config DRM_I915_PXP
+	bool "Enable Intel PXP support"
+	depends on DRM_I915
+	depends on INTEL_MEI && INTEL_MEI_PXP
+	default n
+	help
+	  PXP (Protected Xe Path) is an i915 component, available on graphics
+	  version 12 and newer GPUs, that helps to establish the hardware
+	  protected session and manage the status of the alive software session,
+	  as well as its life cycle.
+
+menu "drm/i915 Debugging"
+depends on DRM_I915
+depends on EXPERT
+source "drivers/gpu/drm/i915/Kconfig.debug"
+endmenu
+
+menu "drm/i915 Profile Guided Optimisation"
+	visible if EXPERT
+	depends on DRM_I915
+	source "drivers/gpu/drm/i915/Kconfig.profile"
+endmenu
+
+menu "drm/i915 Unstable Evolution"
+	visible if EXPERT && STAGING && BROKEN
+	depends on DRM_I915
+	source "drivers/gpu/drm/i915/Kconfig.unstable"
+endmenu
+
+config DRM_I915_GVT
+	bool
diff --git a/drivers/gpu/drm/i915/Kconfig.debug b/drivers/gpu/drm/i915/Kconfig.debug
new file mode 100644
index 000000000..47e845353
--- /dev/null
+++ b/drivers/gpu/drm/i915/Kconfig.debug
@@ -0,0 +1,241 @@
+# SPDX-License-Identifier: GPL-2.0-only
+config DRM_I915_WERROR
+	bool "Force GCC to throw an error instead of a warning when compiling"
+	# As this may inadvertently break the build, only allow the user
+	# to shoot oneself in the foot iff they aim really hard
+	depends on EXPERT
+	# We use the dependency on !COMPILE_TEST to not be enabled in
+	# allmodconfig or allyesconfig configurations
+	depends on !COMPILE_TEST
+	default n
+	help
+	  Add -Werror to the build flags for (and only for) i915.ko.
+	  Do not enable this unless you are writing code for the i915.ko module.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_DEBUG
+	bool "Enable additional driver debugging"
+	depends on DRM_I915
+	depends on EXPERT # only for developers
+	depends on !COMPILE_TEST # never built by robots
+	select DEBUG_FS
+	select PREEMPT_COUNT
+	select I2C_CHARDEV
+	select STACKDEPOT
+	select DRM_DP_AUX_CHARDEV
+	select X86_MSR # used by igt/pm_rpm
+	select DRM_VGEM # used by igt/prime_vgem (dmabuf interop checks)
+	select DRM_DEBUG_MM if DRM=y
+	select DRM_EXPORT_FOR_TESTS if m
+	select DRM_DEBUG_SELFTEST
+	select DMABUF_SELFTESTS
+	select SW_SYNC # signaling validation framework (igt/syncobj*)
+	select DRM_I915_WERROR
+	select DRM_I915_DEBUG_GEM
+	select DRM_I915_DEBUG_GEM_ONCE
+	select DRM_I915_DEBUG_MMIO
+	select DRM_I915_DEBUG_RUNTIME_PM
+	select DRM_I915_SW_FENCE_DEBUG_OBJECTS
+	select DRM_I915_SELFTEST
+	default n
+	help
+	  Choose this option to turn on extra driver debugging that may affect
+	  performance but will catch some internal issues.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_DEBUG_MMIO
+	bool "Always insert extra checks around mmio access by default"
+	default n
+	help
+	  By default, always enables the extra sanity checks (extra register
+	  reads) around every mmio (register) access that will slow the system
+	  down. This sets the default value of i915.mmio_debug to -1 and can
+	  be overridden at module load.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_DEBUG_GEM
+	bool "Insert extra checks into the GEM internals"
+	default n
+	depends on DRM_I915_WERROR
+	help
+	  Enable extra sanity checks (including BUGs) along the GEM driver
+	  paths that may slow the system down and if hit hang the machine.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_DEBUG_GEM_ONCE
+	bool "Make a GEM debug failure fatal"
+	default n
+	depends on DRM_I915_DEBUG_GEM
+	help
+	  During development, we often only want the very first failure
+	  as that would otherwise be lost in the deluge of subsequent
+	  failures. However, more casual testers may not want to trigger
+	  a hard BUG_ON and hope that the system remains sufficiently usable
+	  to capture a bug report in situ.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_ERRLOG_GEM
+	bool "Insert extra logging (very verbose) for common GEM errors"
+	default n
+	depends on DRM_I915_DEBUG_GEM
+	help
+	  Enable additional logging that may help track down the cause of
+	  principally userspace errors.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_TRACE_GEM
+	bool "Insert extra ftrace output from the GEM internals"
+	depends on DRM_I915_DEBUG_GEM
+	select TRACING
+	default n
+	help
+	  Enable additional and verbose debugging output that will spam
+	  ordinary tests, but may be vital for post-mortem debugging when
+	  used with /proc/sys/kernel/ftrace_dump_on_oops
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_TRACE_GTT
+	bool "Insert extra ftrace output from the GTT internals"
+	depends on DRM_I915_DEBUG_GEM
+	select TRACING
+	default n
+	help
+	  Enable additional and verbose debugging output that will spam
+	  ordinary tests, but may be vital for post-mortem debugging when
+	  used with /proc/sys/kernel/ftrace_dump_on_oops
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_SW_FENCE_DEBUG_OBJECTS
+	bool "Enable additional driver debugging for fence objects"
+	depends on DRM_I915
+	select DEBUG_OBJECTS
+	default n
+	help
+	  Choose this option to turn on extra driver debugging that may affect
+	  performance but will catch some internal issues.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_SW_FENCE_CHECK_DAG
+	bool "Enable additional driver debugging for detecting dependency cycles"
+	depends on DRM_I915
+	default n
+	help
+	  Choose this option to turn on extra driver debugging that may affect
+	  performance but will catch some internal issues.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_DEBUG_GUC
+	bool "Enable additional driver debugging for GuC"
+	depends on DRM_I915
+	default n
+	help
+	  Choose this option to turn on extra driver debugging that may affect
+	  performance but will help resolve GuC related issues.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_SELFTEST
+	bool "Enable selftests upon driver load"
+	depends on DRM_I915
+	default n
+	select DRM_EXPORT_FOR_TESTS if m
+	select FAULT_INJECTION
+	select PRIME_NUMBERS
+	select CRC32
+	help
+	  Choose this option to allow the driver to perform selftests upon
+	  loading; also requires the i915.selftest=1 module parameter. To
+	  exit the module after running the selftests (i.e. to prevent normal
+	  module initialisation afterwards) use i915.selftest=-1.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_SELFTEST_BROKEN
+	bool "Enable broken and dangerous selftests"
+	depends on DRM_I915_SELFTEST
+	depends on BROKEN
+	default n
+	help
+	  This option enables the execution of selftests that are "dangerous"
+	  and may trigger unintended HW side-effects as they break strict
+	  rules given in the HW specification. For science.
+
+	  Recommended for masochistic driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_LOW_LEVEL_TRACEPOINTS
+	bool "Enable low level request tracing events"
+	depends on DRM_I915
+	default n
+	help
+	  Choose this option to turn on low level request tracing events.
+	  This provides the ability to precisely monitor engine utilisation
+	  and also analyze the request dependency resolving timeline.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_DEBUG_VBLANK_EVADE
+	bool "Enable extra debug warnings for vblank evasion"
+	depends on DRM_I915
+	default n
+	help
+	  Choose this option to turn on extra debug warnings for the
+	  vblank evade mechanism. This gives a warning every time the
+	  the deadline allotted for the vblank evade critical section
+	  is exceeded, even if there isn't an actual risk of missing
+	  the vblank.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_DEBUG_RUNTIME_PM
+	bool "Enable extra state checking for runtime PM"
+	depends on DRM_I915
+	default n
+	select STACKDEPOT
+	help
+	  Choose this option to turn on extra state checking for the
+	  runtime PM functionality. This may introduce overhead during
+	  driver loading, suspend and resume operations.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N"
diff --git a/drivers/gpu/drm/i915/Kconfig.profile b/drivers/gpu/drm/i915/Kconfig.profile
new file mode 100644
index 000000000..39328567c
--- /dev/null
+++ b/drivers/gpu/drm/i915/Kconfig.profile
@@ -0,0 +1,121 @@
+config DRM_I915_REQUEST_TIMEOUT
+	int "Default timeout for requests (ms)"
+	default 20000 # milliseconds
+	help
+	  Configures the default timeout after which any user submissions will
+	  be forcefully terminated.
+
+	  Beware setting this value lower, or close to heartbeat interval
+	  rounded to whole seconds times three, in order to avoid allowing
+	  misbehaving applications causing total rendering failure in unrelated
+	  clients.
+
+	  May be 0 to disable the timeout.
+
+config DRM_I915_FENCE_TIMEOUT
+	int "Timeout for unsignaled foreign fences (ms, jiffy granularity)"
+	default 10000 # milliseconds
+	help
+	  When listening to a foreign fence, we install a supplementary timer
+	  to ensure that we are always signaled and our userspace is able to
+	  make forward progress. This value specifies the timeout used for an
+	  unsignaled foreign fence.
+
+	  May be 0 to disable the timeout, and rely on the foreign fence being
+	  eventually signaled.
+
+config DRM_I915_USERFAULT_AUTOSUSPEND
+	int "Runtime autosuspend delay for userspace GGTT mmaps (ms)"
+	default 250 # milliseconds
+	help
+	  On runtime suspend, as we suspend the device, we have to revoke
+	  userspace GGTT mmaps and force userspace to take a pagefault on
+	  their next access. The revocation and subsequent recreation of
+	  the GGTT mmap can be very slow and so we impose a small hysteris
+	  that complements the runtime-pm autosuspend and provides a lower
+	  floor on the autosuspend delay.
+
+	  May be 0 to disable the extra delay and solely use the device level
+	  runtime pm autosuspend delay tunable.
+
+config DRM_I915_HEARTBEAT_INTERVAL
+	int "Interval between heartbeat pulses (ms)"
+	default 2500 # milliseconds
+	help
+	  The driver sends a periodic heartbeat down all active engines to
+	  check the health of the GPU and undertake regular house-keeping of
+	  internal driver state.
+
+	  This is adjustable via
+	  /sys/class/drm/card?/engine/*/heartbeat_interval_ms
+
+	  May be 0 to disable heartbeats and therefore disable automatic GPU
+	  hang detection.
+
+config DRM_I915_PREEMPT_TIMEOUT
+	int "Preempt timeout (ms, jiffy granularity)"
+	default 640 # milliseconds
+	help
+	  How long to wait (in milliseconds) for a preemption event to occur
+	  when submitting a new context via execlists. If the current context
+	  does not hit an arbitration point and yield to HW before the timer
+	  expires, the HW will be reset to allow the more important context
+	  to execute.
+
+	  This is adjustable via
+	  /sys/class/drm/card?/engine/*/preempt_timeout_ms
+
+	  May be 0 to disable the timeout.
+
+	  The compiled in default may get overridden at driver probe time on
+	  certain platforms and certain engines which will be reflected in the
+	  sysfs control.
+
+config DRM_I915_MAX_REQUEST_BUSYWAIT
+	int "Busywait for request completion limit (ns)"
+	default 8000 # nanoseconds
+	help
+	  Before sleeping waiting for a request (GPU operation) to complete,
+	  we may spend some time polling for its completion. As the IRQ may
+	  take a non-negligible time to setup, we do a short spin first to
+	  check if the request will complete in the time it would have taken
+	  us to enable the interrupt.
+
+	  This is adjustable via
+	  /sys/class/drm/card?/engine/*/max_busywait_duration_ns
+
+	  May be 0 to disable the initial spin. In practice, we estimate
+	  the cost of enabling the interrupt (if currently disabled) to be
+	  a few microseconds.
+
+config DRM_I915_STOP_TIMEOUT
+	int "How long to wait for an engine to quiesce gracefully before reset (ms)"
+	default 100 # milliseconds
+	help
+	  By stopping submission and sleeping for a short time before resetting
+	  the GPU, we allow the innocent contexts also on the system to quiesce.
+	  It is then less likely for a hanging context to cause collateral
+	  damage as the system is reset in order to recover. The corollary is
+	  that the reset itself may take longer and so be more disruptive to
+	  interactive or low latency workloads.
+
+	  This is adjustable via
+	  /sys/class/drm/card?/engine/*/stop_timeout_ms
+
+config DRM_I915_TIMESLICE_DURATION
+	int "Scheduling quantum for userspace batches (ms, jiffy granularity)"
+	default 1 # milliseconds
+	help
+	  When two user batches of equal priority are executing, we will
+	  alternate execution of each batch to ensure forward progress of
+	  all users. This is necessary in some cases where there may be
+	  an implicit dependency between those batches that requires
+	  concurrent execution in order for them to proceed, e.g. they
+	  interact with each other via userspace semaphores. Each context
+	  is scheduled for execution for the timeslice duration, before
+	  switching to the next context.
+
+	  This is adjustable via
+	  /sys/class/drm/card?/engine/*/timeslice_duration_ms
+
+	  May be 0 to disable timeslicing.
diff --git a/drivers/gpu/drm/i915/Kconfig.unstable b/drivers/gpu/drm/i915/Kconfig.unstable
new file mode 100644
index 000000000..cf151a297
--- /dev/null
+++ b/drivers/gpu/drm/i915/Kconfig.unstable
@@ -0,0 +1,21 @@
+# SPDX-License-Identifier: GPL-2.0-only
+config DRM_I915_UNSTABLE
+	bool "Enable unstable API for early prototype development"
+	depends on EXPERT
+	depends on STAGING
+	depends on BROKEN # should never be enabled by distros!
+	# We use the dependency on !COMPILE_TEST to not be enabled in
+	# allmodconfig or allyesconfig configurations
+	depends on !COMPILE_TEST
+	default n
+	help
+	  Enable prototype uAPI under general discussion before they are
+	  finalized. Such prototypes may be withdrawn or substantially
+	  changed before release. They are only enabled here so that a wide
+	  number of interested parties (userspace driver developers) can
+	  verify that the uAPI meet their expectations. These uAPI should
+	  never be used in production.
+
+	  Recommended for driver developers _only_.
+
+	  If in the slightest bit of doubt, say "N".
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
new file mode 100644
index 000000000..cea00aaca
--- /dev/null
+++ b/drivers/gpu/drm/i915/Makefile
@@ -0,0 +1,364 @@
+# SPDX-License-Identifier: GPL-2.0
+#
+# Makefile for the drm device driver.  This driver provides support for the
+# Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher.
+
+# Add a set of useful warning flags and enable -Werror for CI to prevent
+# trivial mistakes from creeping in. We have to do this piecemeal as we reject
+# any patch that isn't warning clean, so turning on -Wall -Wextra (or W=1) we
+# need to filter out dubious warnings.  Still it is our interest
+# to keep running locally with W=1 C=1 until we are completely clean.
+#
+# Note the danger in using -Wall -Wextra is that when CI updates gcc we
+# will most likely get a sudden build breakage... Hopefully we will fix
+# new warnings before CI updates!
+subdir-ccflags-y := -Wall -Wextra
+subdir-ccflags-y += -Wno-format-security
+subdir-ccflags-y += -Wno-unused-parameter
+subdir-ccflags-y += -Wno-type-limits
+subdir-ccflags-y += -Wno-missing-field-initializers
+subdir-ccflags-y += -Wno-sign-compare
+subdir-ccflags-y += -Wno-shift-negative-value
+subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable)
+subdir-ccflags-y += $(call cc-disable-warning, frame-address)
+subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werror
+
+# Fine grained warnings disable
+CFLAGS_i915_pci.o = $(call cc-disable-warning, override-init)
+CFLAGS_display/intel_fbdev.o = $(call cc-disable-warning, override-init)
+
+subdir-ccflags-y += -I$(srctree)/$(src)
+
+# Please keep these build lists sorted!
+
+# core driver code
+i915-y += i915_driver.o \
+	  i915_drm_client.o \
+	  i915_config.o \
+	  i915_getparam.o \
+	  i915_ioctl.o \
+	  i915_irq.o \
+	  i915_mitigations.o \
+	  i915_module.o \
+	  i915_params.o \
+	  i915_pci.o \
+	  i915_scatterlist.o \
+	  i915_suspend.o \
+	  i915_switcheroo.o \
+	  i915_sysfs.o \
+	  i915_utils.o \
+	  intel_device_info.o \
+	  intel_dram.o \
+	  intel_memory_region.o \
+	  intel_pch.o \
+	  intel_pcode.o \
+	  intel_pm.o \
+	  intel_region_ttm.o \
+	  intel_runtime_pm.o \
+	  intel_sbi.o \
+	  intel_step.o \
+	  intel_uncore.o \
+	  intel_wakeref.o \
+	  vlv_sideband.o \
+	  vlv_suspend.o
+
+# core library code
+i915-y += \
+	i915_memcpy.o \
+	i915_mm.o \
+	i915_sw_fence.o \
+	i915_sw_fence_work.o \
+	i915_syncmap.o \
+	i915_user_extensions.o
+
+i915-$(CONFIG_COMPAT)   += i915_ioc32.o
+i915-$(CONFIG_DEBUG_FS) += \
+	i915_debugfs.o \
+	i915_debugfs_params.o \
+	display/intel_display_debugfs.o \
+	display/intel_pipe_crc.o
+i915-$(CONFIG_PERF_EVENTS) += i915_pmu.o
+
+# "Graphics Technology" (aka we talk to the gpu)
+gt-y += \
+	gt/gen2_engine_cs.o \
+	gt/gen6_engine_cs.o \
+	gt/gen6_ppgtt.o \
+	gt/gen7_renderclear.o \
+	gt/gen8_engine_cs.o \
+	gt/gen8_ppgtt.o \
+	gt/intel_breadcrumbs.o \
+	gt/intel_context.o \
+	gt/intel_context_sseu.o \
+	gt/intel_engine_cs.o \
+	gt/intel_engine_heartbeat.o \
+	gt/intel_engine_pm.o \
+	gt/intel_engine_user.o \
+	gt/intel_execlists_submission.o \
+	gt/intel_ggtt.o \
+	gt/intel_ggtt_fencing.o \
+	gt/intel_gt.o \
+	gt/intel_gt_buffer_pool.o \
+	gt/intel_gt_clock_utils.o \
+	gt/intel_gt_debugfs.o \
+	gt/intel_gt_engines_debugfs.o \
+	gt/intel_gt_irq.o \
+	gt/intel_gt_mcr.o \
+	gt/intel_gt_pm.o \
+	gt/intel_gt_pm_debugfs.o \
+	gt/intel_gt_pm_irq.o \
+	gt/intel_gt_requests.o \
+	gt/intel_gt_sysfs.o \
+	gt/intel_gt_sysfs_pm.o \
+	gt/intel_gtt.o \
+	gt/intel_llc.o \
+	gt/intel_lrc.o \
+	gt/intel_migrate.o \
+	gt/intel_mocs.o \
+	gt/intel_ppgtt.o \
+	gt/intel_rc6.o \
+	gt/intel_region_lmem.o \
+	gt/intel_renderstate.o \
+	gt/intel_reset.o \
+	gt/intel_ring.o \
+	gt/intel_ring_submission.o \
+	gt/intel_rps.o \
+	gt/intel_sa_media.o \
+	gt/intel_sseu.o \
+	gt/intel_sseu_debugfs.o \
+	gt/intel_timeline.o \
+	gt/intel_workarounds.o \
+	gt/shmem_utils.o \
+	gt/sysfs_engines.o
+# x86 intel-gtt module support
+gt-$(CONFIG_X86) += gt/intel_ggtt_gmch.o
+# autogenerated null render state
+gt-y += \
+	gt/gen6_renderstate.o \
+	gt/gen7_renderstate.o \
+	gt/gen8_renderstate.o \
+	gt/gen9_renderstate.o
+i915-y += $(gt-y)
+
+# GEM (Graphics Execution Management) code
+gem-y += \
+	gem/i915_gem_busy.o \
+	gem/i915_gem_clflush.o \
+	gem/i915_gem_context.o \
+	gem/i915_gem_create.o \
+	gem/i915_gem_dmabuf.o \
+	gem/i915_gem_domain.o \
+	gem/i915_gem_execbuffer.o \
+	gem/i915_gem_internal.o \
+	gem/i915_gem_object.o \
+	gem/i915_gem_lmem.o \
+	gem/i915_gem_mman.o \
+	gem/i915_gem_pages.o \
+	gem/i915_gem_phys.o \
+	gem/i915_gem_pm.o \
+	gem/i915_gem_region.o \
+	gem/i915_gem_shmem.o \
+	gem/i915_gem_shrinker.o \
+	gem/i915_gem_stolen.o \
+	gem/i915_gem_throttle.o \
+	gem/i915_gem_tiling.o \
+	gem/i915_gem_ttm.o \
+	gem/i915_gem_ttm_move.o \
+	gem/i915_gem_ttm_pm.o \
+	gem/i915_gem_userptr.o \
+	gem/i915_gem_wait.o \
+	gem/i915_gemfs.o
+i915-y += \
+	  $(gem-y) \
+	  i915_active.o \
+	  i915_cmd_parser.o \
+	  i915_deps.o \
+	  i915_gem_evict.o \
+	  i915_gem_gtt.o \
+	  i915_gem_ww.o \
+	  i915_gem.o \
+	  i915_query.o \
+	  i915_request.o \
+	  i915_scheduler.o \
+	  i915_trace_points.o \
+	  i915_ttm_buddy_manager.o \
+	  i915_vma.o \
+	  i915_vma_resource.o \
+	  intel_wopcm.o
+
+# general-purpose microcontroller (GuC) support
+i915-y += gt/uc/intel_uc.o \
+	  gt/uc/intel_uc_debugfs.o \
+	  gt/uc/intel_uc_fw.o \
+	  gt/uc/intel_guc.o \
+	  gt/uc/intel_guc_ads.o \
+	  gt/uc/intel_guc_capture.o \
+	  gt/uc/intel_guc_ct.o \
+	  gt/uc/intel_guc_debugfs.o \
+	  gt/uc/intel_guc_fw.o \
+	  gt/uc/intel_guc_hwconfig.o \
+	  gt/uc/intel_guc_log.o \
+	  gt/uc/intel_guc_log_debugfs.o \
+	  gt/uc/intel_guc_rc.o \
+	  gt/uc/intel_guc_slpc.o \
+	  gt/uc/intel_guc_submission.o \
+	  gt/uc/intel_huc.o \
+	  gt/uc/intel_huc_debugfs.o \
+	  gt/uc/intel_huc_fw.o
+
+# graphics system controller (GSC) support
+i915-y += gt/intel_gsc.o
+
+# modesetting core code
+i915-y += \
+	display/hsw_ips.o \
+	display/intel_atomic.o \
+	display/intel_atomic_plane.o \
+	display/intel_audio.o \
+	display/intel_bios.o \
+	display/intel_bw.o \
+	display/intel_cdclk.o \
+	display/intel_color.o \
+	display/intel_combo_phy.o \
+	display/intel_connector.o \
+	display/intel_crtc.o \
+	display/intel_crtc_state_dump.o \
+	display/intel_cursor.o \
+	display/intel_display.o \
+	display/intel_display_power.o \
+	display/intel_display_power_map.o \
+	display/intel_display_power_well.o \
+	display/intel_dmc.o \
+	display/intel_dpio_phy.o \
+	display/intel_dpll.o \
+	display/intel_dpll_mgr.o \
+	display/intel_dpt.o \
+	display/intel_drrs.o \
+	display/intel_dsb.o \
+	display/intel_fb.o \
+	display/intel_fb_pin.o \
+	display/intel_fbc.o \
+	display/intel_fdi.o \
+	display/intel_fifo_underrun.o \
+	display/intel_frontbuffer.o \
+	display/intel_global_state.o \
+	display/intel_hdcp.o \
+	display/intel_hotplug.o \
+	display/intel_lpe_audio.o \
+	display/intel_modeset_verify.o \
+	display/intel_modeset_setup.o \
+	display/intel_overlay.o \
+	display/intel_pch_display.o \
+	display/intel_pch_refclk.o \
+	display/intel_plane_initial.o \
+	display/intel_psr.o \
+	display/intel_quirks.o \
+	display/intel_sprite.o \
+	display/intel_tc.o \
+	display/intel_vga.o \
+	display/i9xx_plane.o \
+	display/skl_scaler.o \
+	display/skl_universal_plane.o \
+	display/skl_watermark.o
+i915-$(CONFIG_ACPI) += \
+	display/intel_acpi.o \
+	display/intel_opregion.o
+i915-$(CONFIG_DRM_FBDEV_EMULATION) += \
+	display/intel_fbdev.o
+
+# modesetting output/encoder code
+i915-y += \
+	display/dvo_ch7017.o \
+	display/dvo_ch7xxx.o \
+	display/dvo_ivch.o \
+	display/dvo_ns2501.o \
+	display/dvo_sil164.o \
+	display/dvo_tfp410.o \
+	display/g4x_dp.o \
+	display/g4x_hdmi.o \
+	display/icl_dsi.o \
+	display/intel_backlight.o \
+	display/intel_crt.o \
+	display/intel_ddi.o \
+	display/intel_ddi_buf_trans.o \
+	display/intel_display_trace.o \
+	display/intel_dkl_phy.o \
+	display/intel_dp.o \
+	display/intel_dp_aux.o \
+	display/intel_dp_aux_backlight.o \
+	display/intel_dp_hdcp.o \
+	display/intel_dp_link_training.o \
+	display/intel_dp_mst.o \
+	display/intel_dsi.o \
+	display/intel_dsi_dcs_backlight.o \
+	display/intel_dsi_vbt.o \
+	display/intel_dvo.o \
+	display/intel_gmbus.o \
+	display/intel_hdmi.o \
+	display/intel_lspcon.o \
+	display/intel_lvds.o \
+	display/intel_panel.o \
+	display/intel_pps.o \
+	display/intel_qp_tables.o \
+	display/intel_sdvo.o \
+	display/intel_snps_phy.o \
+	display/intel_tv.o \
+	display/intel_vdsc.o \
+	display/intel_vrr.o \
+	display/vlv_dsi.o \
+	display/vlv_dsi_pll.o
+
+i915-y += i915_perf.o
+
+# Protected execution platform (PXP) support
+i915-$(CONFIG_DRM_I915_PXP) += \
+	pxp/intel_pxp.o \
+	pxp/intel_pxp_cmd.o \
+	pxp/intel_pxp_debugfs.o \
+	pxp/intel_pxp_irq.o \
+	pxp/intel_pxp_pm.o \
+	pxp/intel_pxp_session.o \
+	pxp/intel_pxp_tee.o
+
+# Post-mortem debug and GPU hang state capture
+i915-$(CONFIG_DRM_I915_CAPTURE_ERROR) += i915_gpu_error.o
+i915-$(CONFIG_DRM_I915_SELFTEST) += \
+	gem/selftests/i915_gem_client_blt.o \
+	gem/selftests/igt_gem_utils.o \
+	selftests/intel_scheduler_helpers.o \
+	selftests/i915_random.o \
+	selftests/i915_selftest.o \
+	selftests/igt_atomic.o \
+	selftests/igt_flush_test.o \
+	selftests/igt_live_test.o \
+	selftests/igt_mmap.o \
+	selftests/igt_reset.o \
+	selftests/igt_spinner.o \
+	selftests/librapl.o
+
+# virtual gpu code
+i915-y += i915_vgpu.o
+
+i915-$(CONFIG_DRM_I915_GVT) += \
+	intel_gvt.o \
+	intel_gvt_mmio_table.o
+include $(src)/gvt/Makefile
+
+obj-$(CONFIG_DRM_I915) += i915.o
+obj-$(CONFIG_DRM_I915_GVT_KVMGT) += kvmgt.o
+
+# header test
+
+# exclude some broken headers from the test coverage
+no-header-test := \
+	display/intel_vbt_defs.h
+
+always-$(CONFIG_DRM_I915_WERROR) += \
+	$(patsubst %.h,%.hdrtest, $(filter-out $(no-header-test), \
+		$(shell cd $(srctree)/$(src) && find * -name '*.h')))
+
+quiet_cmd_hdrtest = HDRTEST $(patsubst %.hdrtest,%.h,$@)
+      cmd_hdrtest = $(CC) $(filter-out $(CFLAGS_GCOV), $(c_flags)) -S -o /dev/null -x c /dev/null -include $<; touch $@
+
+$(obj)/%.hdrtest: $(src)/%.h FORCE
+	$(call if_changed_dep,hdrtest)
diff --git a/drivers/gpu/drm/i915/TODO.txt b/drivers/gpu/drm/i915/TODO.txt
new file mode 100644
index 000000000..879b08ca3
--- /dev/null
+++ b/drivers/gpu/drm/i915/TODO.txt
@@ -0,0 +1,41 @@
+gem/gt TODO items
+-----------------
+
+- For discrete memory manager, merge enough dg1 to be able to refactor it to
+  TTM. Then land pci ids (just in case that turns up an uapi problem). TTM has
+  improved a lot the past 2 years, there's no reason anymore not to use it.
+
+- Come up with a plan what to do with drm/scheduler and how to get there.
+
+- Roll out dma_fence critical section annotations.
+
+- There's a lot of complexity added past few years to make relocations faster.
+  That doesn't make sense given hw and gpu apis moved away from this model years
+  ago:
+  1. Land a modern pre-bound uapi like VM_BIND
+  2. Any complexity added in this area past few years which can't be justified
+  with VM_BIND using userspace should be removed. Looking at amdgpu dma_resv on
+  the bo and vm, plus some lru locks is all that needed. No complex rcu,
+  refcounts, caching, ... on everything.
+  This is the matching task on the vm side compared to ttm/dma_resv on the
+  backing storage side.
+
+- i915_sw_fence seems to be the main structure for the i915-gem dma_fence model.
+  How-to-dma_fence is core and drivers really shouldn't build their own world
+  here, treating everything else as a fixed platform. i915_sw_fence concepts
+  should be moved to dma_fence, drm/scheduler or atomic commit helpers. Or
+  removed if dri-devel consensus is that it's not a good idea. Once that's done
+  maybe even remove it if there's nothing left.
+
+Smaller things:
+- i915_utils.h needs to be moved to the right places.
+
+- dma_fence_work should be in drivers/dma-buf
+
+- i915_mm.c should be moved to the right places. Some of the helpers also look a
+  bit fishy:
+
+  https://lore.kernel.org/linux-mm/20210301083320.943079-1-hch@lst.de/
+
+- tasklet helpers in i915_tasklet.h also look a bit misplaced and should
+  probably be moved to tasklet headers.
diff --git a/drivers/gpu/drm/i915/display/dvo_ch7017.c b/drivers/gpu/drm/i915/display/dvo_ch7017.c
new file mode 100644
index 000000000..0589994dd
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/dvo_ch7017.c
@@ -0,0 +1,415 @@
+/*
+ * Copyright © 2006 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *
+ */
+
+#include "intel_display_types.h"
+#include "intel_dvo_dev.h"
+
+#define CH7017_TV_DISPLAY_MODE		0x00
+#define CH7017_FLICKER_FILTER		0x01
+#define CH7017_VIDEO_BANDWIDTH		0x02
+#define CH7017_TEXT_ENHANCEMENT		0x03
+#define CH7017_START_ACTIVE_VIDEO	0x04
+#define CH7017_HORIZONTAL_POSITION	0x05
+#define CH7017_VERTICAL_POSITION	0x06
+#define CH7017_BLACK_LEVEL		0x07
+#define CH7017_CONTRAST_ENHANCEMENT	0x08
+#define CH7017_TV_PLL			0x09
+#define CH7017_TV_PLL_M			0x0a
+#define CH7017_TV_PLL_N			0x0b
+#define CH7017_SUB_CARRIER_0		0x0c
+#define CH7017_CIV_CONTROL		0x10
+#define CH7017_CIV_0			0x11
+#define CH7017_CHROMA_BOOST		0x14
+#define CH7017_CLOCK_MODE		0x1c
+#define CH7017_INPUT_CLOCK		0x1d
+#define CH7017_GPIO_CONTROL		0x1e
+#define CH7017_INPUT_DATA_FORMAT	0x1f
+#define CH7017_CONNECTION_DETECT	0x20
+#define CH7017_DAC_CONTROL		0x21
+#define CH7017_BUFFERED_CLOCK_OUTPUT	0x22
+#define CH7017_DEFEAT_VSYNC		0x47
+#define CH7017_TEST_PATTERN		0x48
+
+#define CH7017_POWER_MANAGEMENT		0x49
+/** Enables the TV output path. */
+#define CH7017_TV_EN			(1 << 0)
+#define CH7017_DAC0_POWER_DOWN		(1 << 1)
+#define CH7017_DAC1_POWER_DOWN		(1 << 2)
+#define CH7017_DAC2_POWER_DOWN		(1 << 3)
+#define CH7017_DAC3_POWER_DOWN		(1 << 4)
+/** Powers down the TV out block, and DAC0-3 */
+#define CH7017_TV_POWER_DOWN_EN		(1 << 5)
+
+#define CH7017_VERSION_ID		0x4a
+
+#define CH7017_DEVICE_ID		0x4b
+#define CH7017_DEVICE_ID_VALUE		0x1b
+#define CH7018_DEVICE_ID_VALUE		0x1a
+#define CH7019_DEVICE_ID_VALUE		0x19
+
+#define CH7017_XCLK_D2_ADJUST		0x53
+#define CH7017_UP_SCALER_COEFF_0	0x55
+#define CH7017_UP_SCALER_COEFF_1	0x56
+#define CH7017_UP_SCALER_COEFF_2	0x57
+#define CH7017_UP_SCALER_COEFF_3	0x58
+#define CH7017_UP_SCALER_COEFF_4	0x59
+#define CH7017_UP_SCALER_VERTICAL_INC_0	0x5a
+#define CH7017_UP_SCALER_VERTICAL_INC_1	0x5b
+#define CH7017_GPIO_INVERT		0x5c
+#define CH7017_UP_SCALER_HORIZONTAL_INC_0	0x5d
+#define CH7017_UP_SCALER_HORIZONTAL_INC_1	0x5e
+
+#define CH7017_HORIZONTAL_ACTIVE_PIXEL_INPUT	0x5f
+/**< Low bits of horizontal active pixel input */
+
+#define CH7017_ACTIVE_INPUT_LINE_OUTPUT	0x60
+/** High bits of horizontal active pixel input */
+#define CH7017_LVDS_HAP_INPUT_MASK	(0x7 << 0)
+/** High bits of vertical active line output */
+#define CH7017_LVDS_VAL_HIGH_MASK	(0x7 << 3)
+
+#define CH7017_VERTICAL_ACTIVE_LINE_OUTPUT	0x61
+/**< Low bits of vertical active line output */
+
+#define CH7017_HORIZONTAL_ACTIVE_PIXEL_OUTPUT	0x62
+/**< Low bits of horizontal active pixel output */
+
+#define CH7017_LVDS_POWER_DOWN		0x63
+/** High bits of horizontal active pixel output */
+#define CH7017_LVDS_HAP_HIGH_MASK	(0x7 << 0)
+/** Enables the LVDS power down state transition */
+#define CH7017_LVDS_POWER_DOWN_EN	(1 << 6)
+/** Enables the LVDS upscaler */
+#define CH7017_LVDS_UPSCALER_EN		(1 << 7)
+#define CH7017_LVDS_POWER_DOWN_DEFAULT_RESERVED 0x08
+
+#define CH7017_LVDS_ENCODING		0x64
+#define CH7017_LVDS_DITHER_2D		(1 << 2)
+#define CH7017_LVDS_DITHER_DIS		(1 << 3)
+#define CH7017_LVDS_DUAL_CHANNEL_EN	(1 << 4)
+#define CH7017_LVDS_24_BIT		(1 << 5)
+
+#define CH7017_LVDS_ENCODING_2		0x65
+
+#define CH7017_LVDS_PLL_CONTROL		0x66
+/** Enables the LVDS panel output path */
+#define CH7017_LVDS_PANEN		(1 << 0)
+/** Enables the LVDS panel backlight */
+#define CH7017_LVDS_BKLEN		(1 << 3)
+
+#define CH7017_POWER_SEQUENCING_T1	0x67
+#define CH7017_POWER_SEQUENCING_T2	0x68
+#define CH7017_POWER_SEQUENCING_T3	0x69
+#define CH7017_POWER_SEQUENCING_T4	0x6a
+#define CH7017_POWER_SEQUENCING_T5	0x6b
+#define CH7017_GPIO_DRIVER_TYPE		0x6c
+#define CH7017_GPIO_DATA		0x6d
+#define CH7017_GPIO_DIRECTION_CONTROL	0x6e
+
+#define CH7017_LVDS_PLL_FEEDBACK_DIV	0x71
+# define CH7017_LVDS_PLL_FEED_BACK_DIVIDER_SHIFT 4
+# define CH7017_LVDS_PLL_FEED_FORWARD_DIVIDER_SHIFT 0
+# define CH7017_LVDS_PLL_FEEDBACK_DEFAULT_RESERVED 0x80
+
+#define CH7017_LVDS_PLL_VCO_CONTROL	0x72
+# define CH7017_LVDS_PLL_VCO_DEFAULT_RESERVED 0x80
+# define CH7017_LVDS_PLL_VCO_SHIFT	4
+# define CH7017_LVDS_PLL_POST_SCALE_DIV_SHIFT 0
+
+#define CH7017_OUTPUTS_ENABLE		0x73
+# define CH7017_CHARGE_PUMP_LOW		0x0
+# define CH7017_CHARGE_PUMP_HIGH	0x3
+# define CH7017_LVDS_CHANNEL_A		(1 << 3)
+# define CH7017_LVDS_CHANNEL_B		(1 << 4)
+# define CH7017_TV_DAC_A		(1 << 5)
+# define CH7017_TV_DAC_B		(1 << 6)
+# define CH7017_DDC_SELECT_DC2		(1 << 7)
+
+#define CH7017_LVDS_OUTPUT_AMPLITUDE	0x74
+#define CH7017_LVDS_PLL_EMI_REDUCTION	0x75
+#define CH7017_LVDS_POWER_DOWN_FLICKER	0x76
+
+#define CH7017_LVDS_CONTROL_2		0x78
+# define CH7017_LOOP_FILTER_SHIFT	5
+# define CH7017_PHASE_DETECTOR_SHIFT	0
+
+#define CH7017_BANG_LIMIT_CONTROL	0x7f
+
+struct ch7017_priv {
+	u8 dummy;
+};
+
+static void ch7017_dump_regs(struct intel_dvo_device *dvo);
+static void ch7017_dpms(struct intel_dvo_device *dvo, bool enable);
+
+static bool ch7017_read(struct intel_dvo_device *dvo, u8 addr, u8 *val)
+{
+	struct i2c_msg msgs[] = {
+		{
+			.addr = dvo->slave_addr,
+			.flags = 0,
+			.len = 1,
+			.buf = &addr,
+		},
+		{
+			.addr = dvo->slave_addr,
+			.flags = I2C_M_RD,
+			.len = 1,
+			.buf = val,
+		}
+	};
+	return i2c_transfer(dvo->i2c_bus, msgs, 2) == 2;
+}
+
+static bool ch7017_write(struct intel_dvo_device *dvo, u8 addr, u8 val)
+{
+	u8 buf[2] = { addr, val };
+	struct i2c_msg msg = {
+		.addr = dvo->slave_addr,
+		.flags = 0,
+		.len = 2,
+		.buf = buf,
+	};
+	return i2c_transfer(dvo->i2c_bus, &msg, 1) == 1;
+}
+
+/** Probes for a CH7017 on the given bus and slave address. */
+static bool ch7017_init(struct intel_dvo_device *dvo,
+			struct i2c_adapter *adapter)
+{
+	struct ch7017_priv *priv;
+	const char *str;
+	u8 val;
+
+	priv = kzalloc(sizeof(struct ch7017_priv), GFP_KERNEL);
+	if (priv == NULL)
+		return false;
+
+	dvo->i2c_bus = adapter;
+	dvo->dev_priv = priv;
+
+	if (!ch7017_read(dvo, CH7017_DEVICE_ID, &val))
+		goto fail;
+
+	switch (val) {
+	case CH7017_DEVICE_ID_VALUE:
+		str = "ch7017";
+		break;
+	case CH7018_DEVICE_ID_VALUE:
+		str = "ch7018";
+		break;
+	case CH7019_DEVICE_ID_VALUE:
+		str = "ch7019";
+		break;
+	default:
+		DRM_DEBUG_KMS("ch701x not detected, got %d: from %s "
+			      "slave %d.\n",
+			      val, adapter->name, dvo->slave_addr);
+		goto fail;
+	}
+
+	DRM_DEBUG_KMS("%s detected on %s, addr %d\n",
+		      str, adapter->name, dvo->slave_addr);
+	return true;
+
+fail:
+	kfree(priv);
+	return false;
+}
+
+static enum drm_connector_status ch7017_detect(struct intel_dvo_device *dvo)
+{
+	return connector_status_connected;
+}
+
+static enum drm_mode_status ch7017_mode_valid(struct intel_dvo_device *dvo,
+					      struct drm_display_mode *mode)
+{
+	if (mode->clock > 160000)
+		return MODE_CLOCK_HIGH;
+
+	return MODE_OK;
+}
+
+static void ch7017_mode_set(struct intel_dvo_device *dvo,
+			    const struct drm_display_mode *mode,
+			    const struct drm_display_mode *adjusted_mode)
+{
+	u8 lvds_pll_feedback_div, lvds_pll_vco_control;
+	u8 outputs_enable, lvds_control_2, lvds_power_down;
+	u8 horizontal_active_pixel_input;
+	u8 horizontal_active_pixel_output, vertical_active_line_output;
+	u8 active_input_line_output;
+
+	DRM_DEBUG_KMS("Registers before mode setting\n");
+	ch7017_dump_regs(dvo);
+
+	/* LVDS PLL settings from page 75 of 7017-7017ds.pdf*/
+	if (mode->clock < 100000) {
+		outputs_enable = CH7017_LVDS_CHANNEL_A | CH7017_CHARGE_PUMP_LOW;
+		lvds_pll_feedback_div = CH7017_LVDS_PLL_FEEDBACK_DEFAULT_RESERVED |
+			(2 << CH7017_LVDS_PLL_FEED_BACK_DIVIDER_SHIFT) |
+			(13 << CH7017_LVDS_PLL_FEED_FORWARD_DIVIDER_SHIFT);
+		lvds_pll_vco_control = CH7017_LVDS_PLL_VCO_DEFAULT_RESERVED |
+			(2 << CH7017_LVDS_PLL_VCO_SHIFT) |
+			(3 << CH7017_LVDS_PLL_POST_SCALE_DIV_SHIFT);
+		lvds_control_2 = (1 << CH7017_LOOP_FILTER_SHIFT) |
+			(0 << CH7017_PHASE_DETECTOR_SHIFT);
+	} else {
+		outputs_enable = CH7017_LVDS_CHANNEL_A | CH7017_CHARGE_PUMP_HIGH;
+		lvds_pll_feedback_div =
+			CH7017_LVDS_PLL_FEEDBACK_DEFAULT_RESERVED |
+			(2 << CH7017_LVDS_PLL_FEED_BACK_DIVIDER_SHIFT) |
+			(3 << CH7017_LVDS_PLL_FEED_FORWARD_DIVIDER_SHIFT);
+		lvds_control_2 = (3 << CH7017_LOOP_FILTER_SHIFT) |
+			(0 << CH7017_PHASE_DETECTOR_SHIFT);
+		if (1) { /* XXX: dual channel panel detection.  Assume yes for now. */
+			outputs_enable |= CH7017_LVDS_CHANNEL_B;
+			lvds_pll_vco_control = CH7017_LVDS_PLL_VCO_DEFAULT_RESERVED |
+				(2 << CH7017_LVDS_PLL_VCO_SHIFT) |
+				(13 << CH7017_LVDS_PLL_POST_SCALE_DIV_SHIFT);
+		} else {
+			lvds_pll_vco_control = CH7017_LVDS_PLL_VCO_DEFAULT_RESERVED |
+				(1 << CH7017_LVDS_PLL_VCO_SHIFT) |
+				(13 << CH7017_LVDS_PLL_POST_SCALE_DIV_SHIFT);
+		}
+	}
+
+	horizontal_active_pixel_input = mode->hdisplay & 0x00ff;
+
+	vertical_active_line_output = mode->vdisplay & 0x00ff;
+	horizontal_active_pixel_output = mode->hdisplay & 0x00ff;
+
+	active_input_line_output = ((mode->hdisplay & 0x0700) >> 8) |
+				   (((mode->vdisplay & 0x0700) >> 8) << 3);
+
+	lvds_power_down = CH7017_LVDS_POWER_DOWN_DEFAULT_RESERVED |
+			  (mode->hdisplay & 0x0700) >> 8;
+
+	ch7017_dpms(dvo, false);
+	ch7017_write(dvo, CH7017_HORIZONTAL_ACTIVE_PIXEL_INPUT,
+			horizontal_active_pixel_input);
+	ch7017_write(dvo, CH7017_HORIZONTAL_ACTIVE_PIXEL_OUTPUT,
+			horizontal_active_pixel_output);
+	ch7017_write(dvo, CH7017_VERTICAL_ACTIVE_LINE_OUTPUT,
+			vertical_active_line_output);
+	ch7017_write(dvo, CH7017_ACTIVE_INPUT_LINE_OUTPUT,
+			active_input_line_output);
+	ch7017_write(dvo, CH7017_LVDS_PLL_VCO_CONTROL, lvds_pll_vco_control);
+	ch7017_write(dvo, CH7017_LVDS_PLL_FEEDBACK_DIV, lvds_pll_feedback_div);
+	ch7017_write(dvo, CH7017_LVDS_CONTROL_2, lvds_control_2);
+	ch7017_write(dvo, CH7017_OUTPUTS_ENABLE, outputs_enable);
+
+	/* Turn the LVDS back on with new settings. */
+	ch7017_write(dvo, CH7017_LVDS_POWER_DOWN, lvds_power_down);
+
+	DRM_DEBUG_KMS("Registers after mode setting\n");
+	ch7017_dump_regs(dvo);
+}
+
+/* set the CH7017 power state */
+static void ch7017_dpms(struct intel_dvo_device *dvo, bool enable)
+{
+	u8 val;
+
+	ch7017_read(dvo, CH7017_LVDS_POWER_DOWN, &val);
+
+	/* Turn off TV/VGA, and never turn it on since we don't support it. */
+	ch7017_write(dvo, CH7017_POWER_MANAGEMENT,
+			CH7017_DAC0_POWER_DOWN |
+			CH7017_DAC1_POWER_DOWN |
+			CH7017_DAC2_POWER_DOWN |
+			CH7017_DAC3_POWER_DOWN |
+			CH7017_TV_POWER_DOWN_EN);
+
+	if (enable) {
+		/* Turn on the LVDS */
+		ch7017_write(dvo, CH7017_LVDS_POWER_DOWN,
+			     val & ~CH7017_LVDS_POWER_DOWN_EN);
+	} else {
+		/* Turn off the LVDS */
+		ch7017_write(dvo, CH7017_LVDS_POWER_DOWN,
+			     val | CH7017_LVDS_POWER_DOWN_EN);
+	}
+
+	/* XXX: Should actually wait for update power status somehow */
+	msleep(20);
+}
+
+static bool ch7017_get_hw_state(struct intel_dvo_device *dvo)
+{
+	u8 val;
+
+	ch7017_read(dvo, CH7017_LVDS_POWER_DOWN, &val);
+
+	if (val & CH7017_LVDS_POWER_DOWN_EN)
+		return false;
+	else
+		return true;
+}
+
+static void ch7017_dump_regs(struct intel_dvo_device *dvo)
+{
+	u8 val;
+
+#define DUMP(reg)					\
+do {							\
+	ch7017_read(dvo, reg, &val);			\
+	DRM_DEBUG_KMS(#reg ": %02x\n", val);		\
+} while (0)
+
+	DUMP(CH7017_HORIZONTAL_ACTIVE_PIXEL_INPUT);
+	DUMP(CH7017_HORIZONTAL_ACTIVE_PIXEL_OUTPUT);
+	DUMP(CH7017_VERTICAL_ACTIVE_LINE_OUTPUT);
+	DUMP(CH7017_ACTIVE_INPUT_LINE_OUTPUT);
+	DUMP(CH7017_LVDS_PLL_VCO_CONTROL);
+	DUMP(CH7017_LVDS_PLL_FEEDBACK_DIV);
+	DUMP(CH7017_LVDS_CONTROL_2);
+	DUMP(CH7017_OUTPUTS_ENABLE);
+	DUMP(CH7017_LVDS_POWER_DOWN);
+}
+
+static void ch7017_destroy(struct intel_dvo_device *dvo)
+{
+	struct ch7017_priv *priv = dvo->dev_priv;
+
+	if (priv) {
+		kfree(priv);
+		dvo->dev_priv = NULL;
+	}
+}
+
+const struct intel_dvo_dev_ops ch7017_ops = {
+	.init = ch7017_init,
+	.detect = ch7017_detect,
+	.mode_valid = ch7017_mode_valid,
+	.mode_set = ch7017_mode_set,
+	.dpms = ch7017_dpms,
+	.get_hw_state = ch7017_get_hw_state,
+	.dump_regs = ch7017_dump_regs,
+	.destroy = ch7017_destroy,
+};
diff --git a/drivers/gpu/drm/i915/display/dvo_ch7xxx.c b/drivers/gpu/drm/i915/display/dvo_ch7xxx.c
new file mode 100644
index 000000000..54f58ba44
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/dvo_ch7xxx.c
@@ -0,0 +1,367 @@
+/**************************************************************************
+
+Copyright © 2006 Dave Airlie
+
+All Rights Reserved.
+
+Permission is hereby granted, free of charge, to any person obtaining a
+copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sub license, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice (including the
+next paragraph) shall be included in all copies or substantial portions
+of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+**************************************************************************/
+
+#include "intel_display_types.h"
+#include "intel_dvo_dev.h"
+
+#define CH7xxx_REG_VID		0x4a
+#define CH7xxx_REG_DID		0x4b
+
+#define CH7011_VID		0x83 /* 7010 as well */
+#define CH7010B_VID		0x05
+#define CH7009A_VID		0x84
+#define CH7009B_VID		0x85
+#define CH7301_VID		0x95
+
+#define CH7xxx_VID		0x84
+#define CH7xxx_DID		0x17
+#define CH7010_DID		0x16
+
+#define CH7xxx_NUM_REGS		0x4c
+
+#define CH7xxx_CM		0x1c
+#define CH7xxx_CM_XCM		(1<<0)
+#define CH7xxx_CM_MCP		(1<<2)
+#define CH7xxx_INPUT_CLOCK	0x1d
+#define CH7xxx_GPIO		0x1e
+#define CH7xxx_GPIO_HPIR	(1<<3)
+#define CH7xxx_IDF		0x1f
+
+#define CH7xxx_IDF_HSP		(1<<3)
+#define CH7xxx_IDF_VSP		(1<<4)
+
+#define CH7xxx_CONNECTION_DETECT 0x20
+#define CH7xxx_CDET_DVI		(1<<5)
+
+#define CH7301_DAC_CNTL		0x21
+#define CH7301_HOTPLUG		0x23
+#define CH7xxx_TCTL		0x31
+#define CH7xxx_TVCO		0x32
+#define CH7xxx_TPCP		0x33
+#define CH7xxx_TPD		0x34
+#define CH7xxx_TPVT		0x35
+#define CH7xxx_TLPF		0x36
+#define CH7xxx_TCT		0x37
+#define CH7301_TEST_PATTERN	0x48
+
+#define CH7xxx_PM		0x49
+#define CH7xxx_PM_FPD		(1<<0)
+#define CH7301_PM_DACPD0	(1<<1)
+#define CH7301_PM_DACPD1	(1<<2)
+#define CH7301_PM_DACPD2	(1<<3)
+#define CH7xxx_PM_DVIL		(1<<6)
+#define CH7xxx_PM_DVIP		(1<<7)
+
+#define CH7301_SYNC_POLARITY	0x56
+#define CH7301_SYNC_RGB_YUV	(1<<0)
+#define CH7301_SYNC_POL_DVI	(1<<5)
+
+/** @file
+ * driver for the Chrontel 7xxx DVI chip over DVO.
+ */
+
+static struct ch7xxx_id_struct {
+	u8 vid;
+	char *name;
+} ch7xxx_ids[] = {
+	{ CH7011_VID, "CH7011" },
+	{ CH7010B_VID, "CH7010B" },
+	{ CH7009A_VID, "CH7009A" },
+	{ CH7009B_VID, "CH7009B" },
+	{ CH7301_VID, "CH7301" },
+};
+
+static struct ch7xxx_did_struct {
+	u8 did;
+	char *name;
+} ch7xxx_dids[] = {
+	{ CH7xxx_DID, "CH7XXX" },
+	{ CH7010_DID, "CH7010B" },
+};
+
+struct ch7xxx_priv {
+	bool quiet;
+};
+
+static char *ch7xxx_get_id(u8 vid)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ch7xxx_ids); i++) {
+		if (ch7xxx_ids[i].vid == vid)
+			return ch7xxx_ids[i].name;
+	}
+
+	return NULL;
+}
+
+static char *ch7xxx_get_did(u8 did)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ch7xxx_dids); i++) {
+		if (ch7xxx_dids[i].did == did)
+			return ch7xxx_dids[i].name;
+	}
+
+	return NULL;
+}
+
+/** Reads an 8 bit register */
+static bool ch7xxx_readb(struct intel_dvo_device *dvo, int addr, u8 *ch)
+{
+	struct ch7xxx_priv *ch7xxx = dvo->dev_priv;
+	struct i2c_adapter *adapter = dvo->i2c_bus;
+	u8 out_buf[2];
+	u8 in_buf[2];
+
+	struct i2c_msg msgs[] = {
+		{
+			.addr = dvo->slave_addr,
+			.flags = 0,
+			.len = 1,
+			.buf = out_buf,
+		},
+		{
+			.addr = dvo->slave_addr,
+			.flags = I2C_M_RD,
+			.len = 1,
+			.buf = in_buf,
+		}
+	};
+
+	out_buf[0] = addr;
+	out_buf[1] = 0;
+
+	if (i2c_transfer(adapter, msgs, 2) == 2) {
+		*ch = in_buf[0];
+		return true;
+	}
+
+	if (!ch7xxx->quiet) {
+		DRM_DEBUG_KMS("Unable to read register 0x%02x from %s:%02x.\n",
+			  addr, adapter->name, dvo->slave_addr);
+	}
+	return false;
+}
+
+/** Writes an 8 bit register */
+static bool ch7xxx_writeb(struct intel_dvo_device *dvo, int addr, u8 ch)
+{
+	struct ch7xxx_priv *ch7xxx = dvo->dev_priv;
+	struct i2c_adapter *adapter = dvo->i2c_bus;
+	u8 out_buf[2];
+	struct i2c_msg msg = {
+		.addr = dvo->slave_addr,
+		.flags = 0,
+		.len = 2,
+		.buf = out_buf,
+	};
+
+	out_buf[0] = addr;
+	out_buf[1] = ch;
+
+	if (i2c_transfer(adapter, &msg, 1) == 1)
+		return true;
+
+	if (!ch7xxx->quiet) {
+		DRM_DEBUG_KMS("Unable to write register 0x%02x to %s:%d.\n",
+			  addr, adapter->name, dvo->slave_addr);
+	}
+
+	return false;
+}
+
+static bool ch7xxx_init(struct intel_dvo_device *dvo,
+			struct i2c_adapter *adapter)
+{
+	/* this will detect the CH7xxx chip on the specified i2c bus */
+	struct ch7xxx_priv *ch7xxx;
+	u8 vendor, device;
+	char *name, *devid;
+
+	ch7xxx = kzalloc(sizeof(struct ch7xxx_priv), GFP_KERNEL);
+	if (ch7xxx == NULL)
+		return false;
+
+	dvo->i2c_bus = adapter;
+	dvo->dev_priv = ch7xxx;
+	ch7xxx->quiet = true;
+
+	if (!ch7xxx_readb(dvo, CH7xxx_REG_VID, &vendor))
+		goto out;
+
+	name = ch7xxx_get_id(vendor);
+	if (!name) {
+		DRM_DEBUG_KMS("ch7xxx not detected; got VID 0x%02x from %s slave %d.\n",
+			      vendor, adapter->name, dvo->slave_addr);
+		goto out;
+	}
+
+
+	if (!ch7xxx_readb(dvo, CH7xxx_REG_DID, &device))
+		goto out;
+
+	devid = ch7xxx_get_did(device);
+	if (!devid) {
+		DRM_DEBUG_KMS("ch7xxx not detected; got DID 0x%02x from %s slave %d.\n",
+			      device, adapter->name, dvo->slave_addr);
+		goto out;
+	}
+
+	ch7xxx->quiet = false;
+	DRM_DEBUG_KMS("Detected %s chipset, vendor/device ID 0x%02x/0x%02x\n",
+		  name, vendor, device);
+	return true;
+out:
+	kfree(ch7xxx);
+	return false;
+}
+
+static enum drm_connector_status ch7xxx_detect(struct intel_dvo_device *dvo)
+{
+	u8 cdet, orig_pm, pm;
+
+	ch7xxx_readb(dvo, CH7xxx_PM, &orig_pm);
+
+	pm = orig_pm;
+	pm &= ~CH7xxx_PM_FPD;
+	pm |= CH7xxx_PM_DVIL | CH7xxx_PM_DVIP;
+
+	ch7xxx_writeb(dvo, CH7xxx_PM, pm);
+
+	ch7xxx_readb(dvo, CH7xxx_CONNECTION_DETECT, &cdet);
+
+	ch7xxx_writeb(dvo, CH7xxx_PM, orig_pm);
+
+	if (cdet & CH7xxx_CDET_DVI)
+		return connector_status_connected;
+	return connector_status_disconnected;
+}
+
+static enum drm_mode_status ch7xxx_mode_valid(struct intel_dvo_device *dvo,
+					      struct drm_display_mode *mode)
+{
+	if (mode->clock > 165000)
+		return MODE_CLOCK_HIGH;
+
+	return MODE_OK;
+}
+
+static void ch7xxx_mode_set(struct intel_dvo_device *dvo,
+			    const struct drm_display_mode *mode,
+			    const struct drm_display_mode *adjusted_mode)
+{
+	u8 tvco, tpcp, tpd, tlpf, idf;
+
+	if (mode->clock <= 65000) {
+		tvco = 0x23;
+		tpcp = 0x08;
+		tpd = 0x16;
+		tlpf = 0x60;
+	} else {
+		tvco = 0x2d;
+		tpcp = 0x06;
+		tpd = 0x26;
+		tlpf = 0xa0;
+	}
+
+	ch7xxx_writeb(dvo, CH7xxx_TCTL, 0x00);
+	ch7xxx_writeb(dvo, CH7xxx_TVCO, tvco);
+	ch7xxx_writeb(dvo, CH7xxx_TPCP, tpcp);
+	ch7xxx_writeb(dvo, CH7xxx_TPD, tpd);
+	ch7xxx_writeb(dvo, CH7xxx_TPVT, 0x30);
+	ch7xxx_writeb(dvo, CH7xxx_TLPF, tlpf);
+	ch7xxx_writeb(dvo, CH7xxx_TCT, 0x00);
+
+	ch7xxx_readb(dvo, CH7xxx_IDF, &idf);
+
+	idf &= ~(CH7xxx_IDF_HSP | CH7xxx_IDF_VSP);
+	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+		idf |= CH7xxx_IDF_HSP;
+
+	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+		idf |= CH7xxx_IDF_VSP;
+
+	ch7xxx_writeb(dvo, CH7xxx_IDF, idf);
+}
+
+/* set the CH7xxx power state */
+static void ch7xxx_dpms(struct intel_dvo_device *dvo, bool enable)
+{
+	if (enable)
+		ch7xxx_writeb(dvo, CH7xxx_PM, CH7xxx_PM_DVIL | CH7xxx_PM_DVIP);
+	else
+		ch7xxx_writeb(dvo, CH7xxx_PM, CH7xxx_PM_FPD);
+}
+
+static bool ch7xxx_get_hw_state(struct intel_dvo_device *dvo)
+{
+	u8 val;
+
+	ch7xxx_readb(dvo, CH7xxx_PM, &val);
+
+	if (val & (CH7xxx_PM_DVIL | CH7xxx_PM_DVIP))
+		return true;
+	else
+		return false;
+}
+
+static void ch7xxx_dump_regs(struct intel_dvo_device *dvo)
+{
+	int i;
+
+	for (i = 0; i < CH7xxx_NUM_REGS; i++) {
+		u8 val;
+		if ((i % 8) == 0)
+			DRM_DEBUG_KMS("\n %02X: ", i);
+		ch7xxx_readb(dvo, i, &val);
+		DRM_DEBUG_KMS("%02X ", val);
+	}
+}
+
+static void ch7xxx_destroy(struct intel_dvo_device *dvo)
+{
+	struct ch7xxx_priv *ch7xxx = dvo->dev_priv;
+
+	if (ch7xxx) {
+		kfree(ch7xxx);
+		dvo->dev_priv = NULL;
+	}
+}
+
+const struct intel_dvo_dev_ops ch7xxx_ops = {
+	.init = ch7xxx_init,
+	.detect = ch7xxx_detect,
+	.mode_valid = ch7xxx_mode_valid,
+	.mode_set = ch7xxx_mode_set,
+	.dpms = ch7xxx_dpms,
+	.get_hw_state = ch7xxx_get_hw_state,
+	.dump_regs = ch7xxx_dump_regs,
+	.destroy = ch7xxx_destroy,
+};
diff --git a/drivers/gpu/drm/i915/display/dvo_ivch.c b/drivers/gpu/drm/i915/display/dvo_ivch.c
new file mode 100644
index 000000000..f43d8c610
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/dvo_ivch.c
@@ -0,0 +1,503 @@
+/*
+ * Copyright © 2006 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *    Thomas Richter <thor@math.tu-berlin.de>
+ *
+ * Minor modifications (Dithering enable):
+ *    Thomas Richter <thor@math.tu-berlin.de>
+ *
+ */
+
+#include "intel_display_types.h"
+#include "intel_dvo_dev.h"
+
+/*
+ * register definitions for the i82807aa.
+ *
+ * Documentation on this chipset can be found in datasheet #29069001 at
+ * intel.com.
+ */
+
+/*
+ * VCH Revision & GMBus Base Addr
+ */
+#define VR00		0x00
+# define VR00_BASE_ADDRESS_MASK		0x007f
+
+/*
+ * Functionality Enable
+ */
+#define VR01		0x01
+
+/*
+ * Enable the panel fitter
+ */
+# define VR01_PANEL_FIT_ENABLE		(1 << 3)
+/*
+ * Enables the LCD display.
+ *
+ * This must not be set while VR01_DVO_BYPASS_ENABLE is set.
+ */
+# define VR01_LCD_ENABLE		(1 << 2)
+/* Enables the DVO repeater. */
+# define VR01_DVO_BYPASS_ENABLE		(1 << 1)
+/* Enables the DVO clock */
+# define VR01_DVO_ENABLE		(1 << 0)
+/* Enable dithering for 18bpp panels. Not documented. */
+# define VR01_DITHER_ENABLE             (1 << 4)
+
+/*
+ * LCD Interface Format
+ */
+#define VR10		0x10
+/* Enables LVDS output instead of CMOS */
+# define VR10_LVDS_ENABLE		(1 << 4)
+/* Enables 18-bit LVDS output. */
+# define VR10_INTERFACE_1X18		(0 << 2)
+/* Enables 24-bit LVDS or CMOS output */
+# define VR10_INTERFACE_1X24		(1 << 2)
+/* Enables 2x18-bit LVDS or CMOS output. */
+# define VR10_INTERFACE_2X18		(2 << 2)
+/* Enables 2x24-bit LVDS output */
+# define VR10_INTERFACE_2X24		(3 << 2)
+/* Mask that defines the depth of the pipeline */
+# define VR10_INTERFACE_DEPTH_MASK      (3 << 2)
+
+/*
+ * VR20 LCD Horizontal Display Size
+ */
+#define VR20	0x20
+
+/*
+ * LCD Vertical Display Size
+ */
+#define VR21	0x21
+
+/*
+ * Panel power down status
+ */
+#define VR30		0x30
+/* Read only bit indicating that the panel is not in a safe poweroff state. */
+# define VR30_PANEL_ON			(1 << 15)
+
+#define VR40		0x40
+# define VR40_STALL_ENABLE		(1 << 13)
+# define VR40_VERTICAL_INTERP_ENABLE	(1 << 12)
+# define VR40_ENHANCED_PANEL_FITTING	(1 << 11)
+# define VR40_HORIZONTAL_INTERP_ENABLE	(1 << 10)
+# define VR40_AUTO_RATIO_ENABLE		(1 << 9)
+# define VR40_CLOCK_GATING_ENABLE	(1 << 8)
+
+/*
+ * Panel Fitting Vertical Ratio
+ * (((image_height - 1) << 16) / ((panel_height - 1))) >> 2
+ */
+#define VR41		0x41
+
+/*
+ * Panel Fitting Horizontal Ratio
+ * (((image_width - 1) << 16) / ((panel_width - 1))) >> 2
+ */
+#define VR42		0x42
+
+/*
+ * Horizontal Image Size
+ */
+#define VR43		0x43
+
+/* VR80 GPIO 0
+ */
+#define VR80	    0x80
+#define VR81	    0x81
+#define VR82	    0x82
+#define VR83	    0x83
+#define VR84	    0x84
+#define VR85	    0x85
+#define VR86	    0x86
+#define VR87	    0x87
+
+/* VR88 GPIO 8
+ */
+#define VR88	    0x88
+
+/* Graphics BIOS scratch 0
+ */
+#define VR8E	    0x8E
+# define VR8E_PANEL_TYPE_MASK		(0xf << 0)
+# define VR8E_PANEL_INTERFACE_CMOS	(0 << 4)
+# define VR8E_PANEL_INTERFACE_LVDS	(1 << 4)
+# define VR8E_FORCE_DEFAULT_PANEL	(1 << 5)
+
+/* Graphics BIOS scratch 1
+ */
+#define VR8F	    0x8F
+# define VR8F_VCH_PRESENT		(1 << 0)
+# define VR8F_DISPLAY_CONN		(1 << 1)
+# define VR8F_POWER_MASK		(0x3c)
+# define VR8F_POWER_POS			(2)
+
+/* Some Bios implementations do not restore the DVO state upon
+ * resume from standby. Thus, this driver has to handle it
+ * instead. The following list contains all registers that
+ * require saving.
+ */
+static const u16 backup_addresses[] = {
+	0x11, 0x12,
+	0x18, 0x19, 0x1a, 0x1f,
+	0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27,
+	0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37,
+	0x8e, 0x8f,
+	0x10		/* this must come last */
+};
+
+
+struct ivch_priv {
+	bool quiet;
+
+	u16 width, height;
+
+	/* Register backup */
+
+	u16 reg_backup[ARRAY_SIZE(backup_addresses)];
+};
+
+
+static void ivch_dump_regs(struct intel_dvo_device *dvo);
+/*
+ * Reads a register on the ivch.
+ *
+ * Each of the 256 registers are 16 bits long.
+ */
+static bool ivch_read(struct intel_dvo_device *dvo, int addr, u16 *data)
+{
+	struct ivch_priv *priv = dvo->dev_priv;
+	struct i2c_adapter *adapter = dvo->i2c_bus;
+	u8 out_buf[1];
+	u8 in_buf[2];
+
+	struct i2c_msg msgs[] = {
+		{
+			.addr = dvo->slave_addr,
+			.flags = I2C_M_RD,
+			.len = 0,
+		},
+		{
+			.addr = 0,
+			.flags = I2C_M_NOSTART,
+			.len = 1,
+			.buf = out_buf,
+		},
+		{
+			.addr = dvo->slave_addr,
+			.flags = I2C_M_RD | I2C_M_NOSTART,
+			.len = 2,
+			.buf = in_buf,
+		}
+	};
+
+	out_buf[0] = addr;
+
+	if (i2c_transfer(adapter, msgs, 3) == 3) {
+		*data = (in_buf[1] << 8) | in_buf[0];
+		return true;
+	}
+
+	if (!priv->quiet) {
+		DRM_DEBUG_KMS("Unable to read register 0x%02x from "
+				"%s:%02x.\n",
+			  addr, adapter->name, dvo->slave_addr);
+	}
+	return false;
+}
+
+/* Writes a 16-bit register on the ivch */
+static bool ivch_write(struct intel_dvo_device *dvo, int addr, u16 data)
+{
+	struct ivch_priv *priv = dvo->dev_priv;
+	struct i2c_adapter *adapter = dvo->i2c_bus;
+	u8 out_buf[3];
+	struct i2c_msg msg = {
+		.addr = dvo->slave_addr,
+		.flags = 0,
+		.len = 3,
+		.buf = out_buf,
+	};
+
+	out_buf[0] = addr;
+	out_buf[1] = data & 0xff;
+	out_buf[2] = data >> 8;
+
+	if (i2c_transfer(adapter, &msg, 1) == 1)
+		return true;
+
+	if (!priv->quiet) {
+		DRM_DEBUG_KMS("Unable to write register 0x%02x to %s:%d.\n",
+			  addr, adapter->name, dvo->slave_addr);
+	}
+
+	return false;
+}
+
+/* Probes the given bus and slave address for an ivch */
+static bool ivch_init(struct intel_dvo_device *dvo,
+		      struct i2c_adapter *adapter)
+{
+	struct ivch_priv *priv;
+	u16 temp;
+	int i;
+
+	priv = kzalloc(sizeof(struct ivch_priv), GFP_KERNEL);
+	if (priv == NULL)
+		return false;
+
+	dvo->i2c_bus = adapter;
+	dvo->dev_priv = priv;
+	priv->quiet = true;
+
+	if (!ivch_read(dvo, VR00, &temp))
+		goto out;
+	priv->quiet = false;
+
+	/* Since the identification bits are probably zeroes, which doesn't seem
+	 * very unique, check that the value in the base address field matches
+	 * the address it's responding on.
+	 */
+	if ((temp & VR00_BASE_ADDRESS_MASK) != dvo->slave_addr) {
+		DRM_DEBUG_KMS("ivch detect failed due to address mismatch "
+			  "(%d vs %d)\n",
+			  (temp & VR00_BASE_ADDRESS_MASK), dvo->slave_addr);
+		goto out;
+	}
+
+	ivch_read(dvo, VR20, &priv->width);
+	ivch_read(dvo, VR21, &priv->height);
+
+	/* Make a backup of the registers to be able to restore them
+	 * upon suspend.
+	 */
+	for (i = 0; i < ARRAY_SIZE(backup_addresses); i++)
+		ivch_read(dvo, backup_addresses[i], priv->reg_backup + i);
+
+	ivch_dump_regs(dvo);
+
+	return true;
+
+out:
+	kfree(priv);
+	return false;
+}
+
+static enum drm_connector_status ivch_detect(struct intel_dvo_device *dvo)
+{
+	return connector_status_connected;
+}
+
+static enum drm_mode_status ivch_mode_valid(struct intel_dvo_device *dvo,
+					    struct drm_display_mode *mode)
+{
+	if (mode->clock > 112000)
+		return MODE_CLOCK_HIGH;
+
+	return MODE_OK;
+}
+
+/* Restore the DVO registers after a resume
+ * from RAM. Registers have been saved during
+ * the initialization.
+ */
+static void ivch_reset(struct intel_dvo_device *dvo)
+{
+	struct ivch_priv *priv = dvo->dev_priv;
+	int i;
+
+	DRM_DEBUG_KMS("Resetting the IVCH registers\n");
+
+	ivch_write(dvo, VR10, 0x0000);
+
+	for (i = 0; i < ARRAY_SIZE(backup_addresses); i++)
+		ivch_write(dvo, backup_addresses[i], priv->reg_backup[i]);
+}
+
+/* Sets the power state of the panel connected to the ivch */
+static void ivch_dpms(struct intel_dvo_device *dvo, bool enable)
+{
+	int i;
+	u16 vr01, vr30, backlight;
+
+	ivch_reset(dvo);
+
+	/* Set the new power state of the panel. */
+	if (!ivch_read(dvo, VR01, &vr01))
+		return;
+
+	if (enable)
+		backlight = 1;
+	else
+		backlight = 0;
+
+	ivch_write(dvo, VR80, backlight);
+
+	if (enable)
+		vr01 |= VR01_LCD_ENABLE | VR01_DVO_ENABLE;
+	else
+		vr01 &= ~(VR01_LCD_ENABLE | VR01_DVO_ENABLE);
+
+	ivch_write(dvo, VR01, vr01);
+
+	/* Wait for the panel to make its state transition */
+	for (i = 0; i < 100; i++) {
+		if (!ivch_read(dvo, VR30, &vr30))
+			break;
+
+		if (((vr30 & VR30_PANEL_ON) != 0) == enable)
+			break;
+		udelay(1000);
+	}
+	/* wait some more; vch may fail to resync sometimes without this */
+	udelay(16 * 1000);
+}
+
+static bool ivch_get_hw_state(struct intel_dvo_device *dvo)
+{
+	u16 vr01;
+
+	ivch_reset(dvo);
+
+	/* Set the new power state of the panel. */
+	if (!ivch_read(dvo, VR01, &vr01))
+		return false;
+
+	if (vr01 & VR01_LCD_ENABLE)
+		return true;
+	else
+		return false;
+}
+
+static void ivch_mode_set(struct intel_dvo_device *dvo,
+			  const struct drm_display_mode *mode,
+			  const struct drm_display_mode *adjusted_mode)
+{
+	struct ivch_priv *priv = dvo->dev_priv;
+	u16 vr40 = 0;
+	u16 vr01 = 0;
+	u16 vr10;
+
+	ivch_reset(dvo);
+
+	vr10 = priv->reg_backup[ARRAY_SIZE(backup_addresses) - 1];
+
+	/* Enable dithering for 18 bpp pipelines */
+	vr10 &= VR10_INTERFACE_DEPTH_MASK;
+	if (vr10 == VR10_INTERFACE_2X18 || vr10 == VR10_INTERFACE_1X18)
+		vr01 = VR01_DITHER_ENABLE;
+
+	vr40 = (VR40_STALL_ENABLE | VR40_VERTICAL_INTERP_ENABLE |
+		VR40_HORIZONTAL_INTERP_ENABLE);
+
+	if (mode->hdisplay != adjusted_mode->crtc_hdisplay ||
+	    mode->vdisplay != adjusted_mode->crtc_vdisplay) {
+		u16 x_ratio, y_ratio;
+
+		vr01 |= VR01_PANEL_FIT_ENABLE;
+		vr40 |= VR40_CLOCK_GATING_ENABLE;
+		x_ratio = (((mode->hdisplay - 1) << 16) /
+			   (adjusted_mode->crtc_hdisplay - 1)) >> 2;
+		y_ratio = (((mode->vdisplay - 1) << 16) /
+			   (adjusted_mode->crtc_vdisplay - 1)) >> 2;
+		ivch_write(dvo, VR42, x_ratio);
+		ivch_write(dvo, VR41, y_ratio);
+	} else {
+		vr01 &= ~VR01_PANEL_FIT_ENABLE;
+		vr40 &= ~VR40_CLOCK_GATING_ENABLE;
+	}
+	vr40 &= ~VR40_AUTO_RATIO_ENABLE;
+
+	ivch_write(dvo, VR01, vr01);
+	ivch_write(dvo, VR40, vr40);
+}
+
+static void ivch_dump_regs(struct intel_dvo_device *dvo)
+{
+	u16 val;
+
+	ivch_read(dvo, VR00, &val);
+	DRM_DEBUG_KMS("VR00: 0x%04x\n", val);
+	ivch_read(dvo, VR01, &val);
+	DRM_DEBUG_KMS("VR01: 0x%04x\n", val);
+	ivch_read(dvo, VR10, &val);
+	DRM_DEBUG_KMS("VR10: 0x%04x\n", val);
+	ivch_read(dvo, VR30, &val);
+	DRM_DEBUG_KMS("VR30: 0x%04x\n", val);
+	ivch_read(dvo, VR40, &val);
+	DRM_DEBUG_KMS("VR40: 0x%04x\n", val);
+
+	/* GPIO registers */
+	ivch_read(dvo, VR80, &val);
+	DRM_DEBUG_KMS("VR80: 0x%04x\n", val);
+	ivch_read(dvo, VR81, &val);
+	DRM_DEBUG_KMS("VR81: 0x%04x\n", val);
+	ivch_read(dvo, VR82, &val);
+	DRM_DEBUG_KMS("VR82: 0x%04x\n", val);
+	ivch_read(dvo, VR83, &val);
+	DRM_DEBUG_KMS("VR83: 0x%04x\n", val);
+	ivch_read(dvo, VR84, &val);
+	DRM_DEBUG_KMS("VR84: 0x%04x\n", val);
+	ivch_read(dvo, VR85, &val);
+	DRM_DEBUG_KMS("VR85: 0x%04x\n", val);
+	ivch_read(dvo, VR86, &val);
+	DRM_DEBUG_KMS("VR86: 0x%04x\n", val);
+	ivch_read(dvo, VR87, &val);
+	DRM_DEBUG_KMS("VR87: 0x%04x\n", val);
+	ivch_read(dvo, VR88, &val);
+	DRM_DEBUG_KMS("VR88: 0x%04x\n", val);
+
+	/* Scratch register 0 - AIM Panel type */
+	ivch_read(dvo, VR8E, &val);
+	DRM_DEBUG_KMS("VR8E: 0x%04x\n", val);
+
+	/* Scratch register 1 - Status register */
+	ivch_read(dvo, VR8F, &val);
+	DRM_DEBUG_KMS("VR8F: 0x%04x\n", val);
+}
+
+static void ivch_destroy(struct intel_dvo_device *dvo)
+{
+	struct ivch_priv *priv = dvo->dev_priv;
+
+	if (priv) {
+		kfree(priv);
+		dvo->dev_priv = NULL;
+	}
+}
+
+const struct intel_dvo_dev_ops ivch_ops = {
+	.init = ivch_init,
+	.dpms = ivch_dpms,
+	.get_hw_state = ivch_get_hw_state,
+	.mode_valid = ivch_mode_valid,
+	.mode_set = ivch_mode_set,
+	.detect = ivch_detect,
+	.dump_regs = ivch_dump_regs,
+	.destroy = ivch_destroy,
+};
diff --git a/drivers/gpu/drm/i915/display/dvo_ns2501.c b/drivers/gpu/drm/i915/display/dvo_ns2501.c
new file mode 100644
index 000000000..a724a8755
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/dvo_ns2501.c
@@ -0,0 +1,710 @@
+/*
+ *
+ * Copyright (c) 2012 Gilles Dartiguelongue, Thomas Richter
+ *
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_display_types.h"
+#include "intel_dvo_dev.h"
+
+#define NS2501_VID 0x1305
+#define NS2501_DID 0x6726
+
+#define NS2501_VID_LO 0x00
+#define NS2501_VID_HI 0x01
+#define NS2501_DID_LO 0x02
+#define NS2501_DID_HI 0x03
+#define NS2501_REV 0x04
+#define NS2501_RSVD 0x05
+#define NS2501_FREQ_LO 0x06
+#define NS2501_FREQ_HI 0x07
+
+#define NS2501_REG8 0x08
+#define NS2501_8_VEN (1<<5)
+#define NS2501_8_HEN (1<<4)
+#define NS2501_8_DSEL (1<<3)
+#define NS2501_8_BPAS (1<<2)
+#define NS2501_8_RSVD (1<<1)
+#define NS2501_8_PD (1<<0)
+
+#define NS2501_REG9 0x09
+#define NS2501_9_VLOW (1<<7)
+#define NS2501_9_MSEL_MASK (0x7<<4)
+#define NS2501_9_TSEL (1<<3)
+#define NS2501_9_RSEN (1<<2)
+#define NS2501_9_RSVD (1<<1)
+#define NS2501_9_MDI (1<<0)
+
+#define NS2501_REGC 0x0c
+
+/*
+ * The following registers are not part of the official datasheet
+ * and are the result of reverse engineering.
+ */
+
+/*
+ * Register c0 controls how the DVO synchronizes with
+ * its input.
+ */
+#define NS2501_REGC0 0xc0
+#define NS2501_C0_ENABLE (1<<0)	/* enable the DVO sync in general */
+#define NS2501_C0_HSYNC (1<<1)	/* synchronize horizontal with input */
+#define NS2501_C0_VSYNC (1<<2)	/* synchronize vertical with input */
+#define NS2501_C0_RESET (1<<7)	/* reset the synchronization flip/flops */
+
+/*
+ * Register 41 is somehow related to the sync register and sync
+ * configuration. It should be 0x32 whenever regC0 is 0x05 (hsync off)
+ * and 0x00 otherwise.
+ */
+#define NS2501_REG41 0x41
+
+/*
+ * this register controls the dithering of the DVO
+ * One bit enables it, the other define the dithering depth.
+ * The higher the value, the lower the dithering depth.
+ */
+#define NS2501_F9_REG 0xf9
+#define NS2501_F9_ENABLE (1<<0)		/* if set, dithering is enabled */
+#define NS2501_F9_DITHER_MASK (0x7f<<1)	/* controls the dither depth */
+#define NS2501_F9_DITHER_SHIFT 1	/* shifts the dither mask */
+
+/*
+ * PLL configuration register. This is a pair of registers,
+ * one single byte register at 1B, and a pair at 1C,1D.
+ * These registers are counters/dividers.
+ */
+#define NS2501_REG1B 0x1b /* one byte PLL control register */
+#define NS2501_REG1C 0x1c /* low-part of the second register */
+#define NS2501_REG1D 0x1d /* high-part of the second register */
+
+/*
+ * Scaler control registers. Horizontal at b8,b9,
+ * vertical at 10,11. The scale factor is computed as
+ * 2^16/control-value. The low-byte comes first.
+ */
+#define NS2501_REG10 0x10 /* low-byte vertical scaler */
+#define NS2501_REG11 0x11 /* high-byte vertical scaler */
+#define NS2501_REGB8 0xb8 /* low-byte horizontal scaler */
+#define NS2501_REGB9 0xb9 /* high-byte horizontal scaler */
+
+/*
+ * Display window definition. This consists of four registers
+ * per dimension. One register pair defines the start of the
+ * display, one the end.
+ * As far as I understand, this defines the window within which
+ * the scaler samples the input.
+ */
+#define NS2501_REGC1 0xc1 /* low-byte horizontal display start */
+#define NS2501_REGC2 0xc2 /* high-byte horizontal display start */
+#define NS2501_REGC3 0xc3 /* low-byte horizontal display stop */
+#define NS2501_REGC4 0xc4 /* high-byte horizontal display stop */
+#define NS2501_REGC5 0xc5 /* low-byte vertical display start */
+#define NS2501_REGC6 0xc6 /* high-byte vertical display start */
+#define NS2501_REGC7 0xc7 /* low-byte vertical display stop */
+#define NS2501_REGC8 0xc8 /* high-byte vertical display stop */
+
+/*
+ * The following register pair seems to define the start of
+ * the vertical sync. If automatic syncing is enabled, and the
+ * register value defines a sync pulse that is later than the
+ * incoming sync, then the register value is ignored and the
+ * external hsync triggers the synchronization.
+ */
+#define NS2501_REG80 0x80 /* low-byte vsync-start */
+#define NS2501_REG81 0x81 /* high-byte vsync-start */
+
+/*
+ * The following register pair seems to define the total number
+ * of lines created at the output side of the scaler.
+ * This is again a low-high register pair.
+ */
+#define NS2501_REG82 0x82 /* output display height, low byte */
+#define NS2501_REG83 0x83 /* output display height, high byte */
+
+/*
+ * The following registers define the end of the front-porch
+ * in horizontal and vertical position and hence allow to shift
+ * the image left/right or up/down.
+ */
+#define NS2501_REG98 0x98 /* horizontal start of display + 256, low */
+#define NS2501_REG99 0x99 /* horizontal start of display + 256, high */
+#define NS2501_REG8E 0x8e /* vertical start of the display, low byte */
+#define NS2501_REG8F 0x8f /* vertical start of the display, high byte */
+
+/*
+ * The following register pair control the function of the
+ * backlight and the DVO output. To enable the corresponding
+ * function, the corresponding bit must be set in both registers.
+ */
+#define NS2501_REG34 0x34 /* DVO enable functions, first register */
+#define NS2501_REG35 0x35 /* DVO enable functions, second register */
+#define NS2501_34_ENABLE_OUTPUT (1<<0) /* enable DVO output */
+#define NS2501_34_ENABLE_BACKLIGHT (1<<1) /* enable backlight */
+
+/*
+ * Registers 9C and 9D define the vertical output offset
+ * of the visible region.
+ */
+#define NS2501_REG9C 0x9c
+#define NS2501_REG9D 0x9d
+
+/*
+ * The register 9F defines the dithering. This requires the
+ * scaler to be ON. Bit 0 enables dithering, the remaining
+ * bits control the depth of the dither. The higher the value,
+ * the LOWER the dithering amplitude. A good value seems to be
+ * 15 (total register value).
+ */
+#define NS2501_REGF9 0xf9
+#define NS2501_F9_ENABLE_DITHER (1<<0) /* enable dithering */
+#define NS2501_F9_DITHER_MASK (0x7f<<1) /* dither masking */
+#define NS2501_F9_DITHER_SHIFT 1	/* upshift of the dither mask */
+
+enum {
+	MODE_640x480,
+	MODE_800x600,
+	MODE_1024x768,
+};
+
+struct ns2501_reg {
+	u8 offset;
+	u8 value;
+};
+
+/*
+ * The following structure keeps the complete configuration of
+ * the DVO, given a specific output configuration.
+ * This is pretty much guess-work from reverse-engineering, so
+ * read all this with a grain of salt.
+ */
+struct ns2501_configuration {
+	u8 sync;		/* configuration of the C0 register */
+	u8 conf;		/* configuration register 8 */
+	u8 syncb;		/* configuration register 41 */
+	u8 dither;		/* configuration of the dithering */
+	u8 pll_a;		/* PLL configuration, register A, 1B */
+	u16 pll_b;		/* PLL configuration, register B, 1C/1D */
+	u16 hstart;		/* horizontal start, registers C1/C2 */
+	u16 hstop;		/* horizontal total, registers C3/C4 */
+	u16 vstart;		/* vertical start, registers C5/C6 */
+	u16 vstop;		/* vertical total, registers C7/C8 */
+	u16 vsync;		/* manual vertical sync start, 80/81 */
+	u16 vtotal;		/* number of lines generated, 82/83 */
+	u16 hpos;		/* horizontal position + 256, 98/99  */
+	u16 vpos;		/* vertical position, 8e/8f */
+	u16 voffs;		/* vertical output offset, 9c/9d */
+	u16 hscale;		/* horizontal scaling factor, b8/b9 */
+	u16 vscale;		/* vertical scaling factor, 10/11 */
+};
+
+/*
+ * DVO configuration values, partially based on what the BIOS
+ * of the Fujitsu Lifebook S6010 writes into registers,
+ * partially found by manual tweaking. These configurations assume
+ * a 1024x768 panel.
+ */
+static const struct ns2501_configuration ns2501_modes[] = {
+	[MODE_640x480] = {
+		.sync	= NS2501_C0_ENABLE | NS2501_C0_VSYNC,
+		.conf	= NS2501_8_VEN | NS2501_8_HEN | NS2501_8_PD,
+		.syncb	= 0x32,
+		.dither	= 0x0f,
+		.pll_a	= 17,
+		.pll_b	= 852,
+		.hstart	= 144,
+		.hstop	= 783,
+		.vstart	= 22,
+		.vstop	= 514,
+		.vsync	= 2047, /* actually, ignored with this config */
+		.vtotal	= 1341,
+		.hpos	= 0,
+		.vpos	= 16,
+		.voffs	= 36,
+		.hscale	= 40960,
+		.vscale	= 40960
+	},
+	[MODE_800x600] = {
+		.sync	= NS2501_C0_ENABLE |
+			  NS2501_C0_HSYNC | NS2501_C0_VSYNC,
+		.conf   = NS2501_8_VEN | NS2501_8_HEN | NS2501_8_PD,
+		.syncb	= 0x00,
+		.dither	= 0x0f,
+		.pll_a	= 25,
+		.pll_b	= 612,
+		.hstart	= 215,
+		.hstop	= 1016,
+		.vstart	= 26,
+		.vstop	= 627,
+		.vsync	= 807,
+		.vtotal	= 1341,
+		.hpos	= 0,
+		.vpos	= 4,
+		.voffs	= 35,
+		.hscale	= 51248,
+		.vscale	= 51232
+	},
+	[MODE_1024x768] = {
+		.sync	= NS2501_C0_ENABLE | NS2501_C0_VSYNC,
+		.conf   = NS2501_8_VEN | NS2501_8_HEN | NS2501_8_PD,
+		.syncb	= 0x32,
+		.dither	= 0x0f,
+		.pll_a	= 11,
+		.pll_b	= 1350,
+		.hstart	= 276,
+		.hstop	= 1299,
+		.vstart	= 15,
+		.vstop	= 1056,
+		.vsync	= 2047,
+		.vtotal	= 1341,
+		.hpos	= 0,
+		.vpos	= 7,
+		.voffs	= 27,
+		.hscale	= 65535,
+		.vscale	= 65535
+	}
+};
+
+/*
+ * Other configuration values left by the BIOS of the
+ * Fujitsu S6010 in the DVO control registers. Their
+ * value does not depend on the BIOS and their meaning
+ * is unknown.
+ */
+
+static const struct ns2501_reg mode_agnostic_values[] = {
+	/* 08 is mode specific */
+	[0] = { .offset = 0x0a, .value = 0x81, },
+	/* 10,11 are part of the mode specific configuration */
+	[1] = { .offset = 0x12, .value = 0x02, },
+	[2] = { .offset = 0x18, .value = 0x07, },
+	[3] = { .offset = 0x19, .value = 0x00, },
+	[4] = { .offset = 0x1a, .value = 0x00, }, /* PLL?, ignored */
+	/* 1b,1c,1d are part of the mode specific configuration */
+	[5] = { .offset = 0x1e, .value = 0x02, },
+	[6] = { .offset = 0x1f, .value = 0x40, },
+	[7] = { .offset = 0x20, .value = 0x00, },
+	[8] = { .offset = 0x21, .value = 0x00, },
+	[9] = { .offset = 0x22, .value = 0x00, },
+	[10] = { .offset = 0x23, .value = 0x00, },
+	[11] = { .offset = 0x24, .value = 0x00, },
+	[12] = { .offset = 0x25, .value = 0x00, },
+	[13] = { .offset = 0x26, .value = 0x00, },
+	[14] = { .offset = 0x27, .value = 0x00, },
+	[15] = { .offset = 0x7e, .value = 0x18, },
+	/* 80-84 are part of the mode-specific configuration */
+	[16] = { .offset = 0x84, .value = 0x00, },
+	[17] = { .offset = 0x85, .value = 0x00, },
+	[18] = { .offset = 0x86, .value = 0x00, },
+	[19] = { .offset = 0x87, .value = 0x00, },
+	[20] = { .offset = 0x88, .value = 0x00, },
+	[21] = { .offset = 0x89, .value = 0x00, },
+	[22] = { .offset = 0x8a, .value = 0x00, },
+	[23] = { .offset = 0x8b, .value = 0x00, },
+	[24] = { .offset = 0x8c, .value = 0x10, },
+	[25] = { .offset = 0x8d, .value = 0x02, },
+	/* 8e,8f are part of the mode-specific configuration */
+	[26] = { .offset = 0x90, .value = 0xff, },
+	[27] = { .offset = 0x91, .value = 0x07, },
+	[28] = { .offset = 0x92, .value = 0xa0, },
+	[29] = { .offset = 0x93, .value = 0x02, },
+	[30] = { .offset = 0x94, .value = 0x00, },
+	[31] = { .offset = 0x95, .value = 0x00, },
+	[32] = { .offset = 0x96, .value = 0x05, },
+	[33] = { .offset = 0x97, .value = 0x00, },
+	/* 98,99 are part of the mode-specific configuration */
+	[34] = { .offset = 0x9a, .value = 0x88, },
+	[35] = { .offset = 0x9b, .value = 0x00, },
+	/* 9c,9d are part of the mode-specific configuration */
+	[36] = { .offset = 0x9e, .value = 0x25, },
+	[37] = { .offset = 0x9f, .value = 0x03, },
+	[38] = { .offset = 0xa0, .value = 0x28, },
+	[39] = { .offset = 0xa1, .value = 0x01, },
+	[40] = { .offset = 0xa2, .value = 0x28, },
+	[41] = { .offset = 0xa3, .value = 0x05, },
+	/* register 0xa4 is mode specific, but 0x80..0x84 works always */
+	[42] = { .offset = 0xa4, .value = 0x84, },
+	[43] = { .offset = 0xa5, .value = 0x00, },
+	[44] = { .offset = 0xa6, .value = 0x00, },
+	[45] = { .offset = 0xa7, .value = 0x00, },
+	[46] = { .offset = 0xa8, .value = 0x00, },
+	/* 0xa9 to 0xab are mode specific, but have no visible effect */
+	[47] = { .offset = 0xa9, .value = 0x04, },
+	[48] = { .offset = 0xaa, .value = 0x70, },
+	[49] = { .offset = 0xab, .value = 0x4f, },
+	[50] = { .offset = 0xac, .value = 0x00, },
+	[51] = { .offset = 0xad, .value = 0x00, },
+	[52] = { .offset = 0xb6, .value = 0x09, },
+	[53] = { .offset = 0xb7, .value = 0x03, },
+	/* b8,b9 are part of the mode-specific configuration */
+	[54] = { .offset = 0xba, .value = 0x00, },
+	[55] = { .offset = 0xbb, .value = 0x20, },
+	[56] = { .offset = 0xf3, .value = 0x90, },
+	[57] = { .offset = 0xf4, .value = 0x00, },
+	[58] = { .offset = 0xf7, .value = 0x88, },
+	/* f8 is mode specific, but the value does not matter */
+	[59] = { .offset = 0xf8, .value = 0x0a, },
+	[60] = { .offset = 0xf9, .value = 0x00, }
+};
+
+static const struct ns2501_reg regs_init[] = {
+	[0] = { .offset = 0x35, .value = 0xff, },
+	[1] = { .offset = 0x34, .value = 0x00, },
+	[2] = { .offset = 0x08, .value = 0x30, },
+};
+
+struct ns2501_priv {
+	bool quiet;
+	const struct ns2501_configuration *conf;
+};
+
+#define NSPTR(d) ((NS2501Ptr)(d->DriverPrivate.ptr))
+
+/*
+** Read a register from the ns2501.
+** Returns true if successful, false otherwise.
+** If it returns false, it might be wise to enable the
+** DVO with the above function.
+*/
+static bool ns2501_readb(struct intel_dvo_device *dvo, int addr, u8 *ch)
+{
+	struct ns2501_priv *ns = dvo->dev_priv;
+	struct i2c_adapter *adapter = dvo->i2c_bus;
+	u8 out_buf[2];
+	u8 in_buf[2];
+
+	struct i2c_msg msgs[] = {
+		{
+		 .addr = dvo->slave_addr,
+		 .flags = 0,
+		 .len = 1,
+		 .buf = out_buf,
+		 },
+		{
+		 .addr = dvo->slave_addr,
+		 .flags = I2C_M_RD,
+		 .len = 1,
+		 .buf = in_buf,
+		 }
+	};
+
+	out_buf[0] = addr;
+	out_buf[1] = 0;
+
+	if (i2c_transfer(adapter, msgs, 2) == 2) {
+		*ch = in_buf[0];
+		return true;
+	}
+
+	if (!ns->quiet) {
+		DRM_DEBUG_KMS
+		    ("Unable to read register 0x%02x from %s:0x%02x.\n", addr,
+		     adapter->name, dvo->slave_addr);
+	}
+
+	return false;
+}
+
+/*
+** Write a register to the ns2501.
+** Returns true if successful, false otherwise.
+** If it returns false, it might be wise to enable the
+** DVO with the above function.
+*/
+static bool ns2501_writeb(struct intel_dvo_device *dvo, int addr, u8 ch)
+{
+	struct ns2501_priv *ns = dvo->dev_priv;
+	struct i2c_adapter *adapter = dvo->i2c_bus;
+	u8 out_buf[2];
+
+	struct i2c_msg msg = {
+		.addr = dvo->slave_addr,
+		.flags = 0,
+		.len = 2,
+		.buf = out_buf,
+	};
+
+	out_buf[0] = addr;
+	out_buf[1] = ch;
+
+	if (i2c_transfer(adapter, &msg, 1) == 1) {
+		return true;
+	}
+
+	if (!ns->quiet) {
+		DRM_DEBUG_KMS("Unable to write register 0x%02x to %s:%d\n",
+			      addr, adapter->name, dvo->slave_addr);
+	}
+
+	return false;
+}
+
+/* National Semiconductor 2501 driver for chip on i2c bus
+ * scan for the chip on the bus.
+ * Hope the VBIOS initialized the PLL correctly so we can
+ * talk to it. If not, it will not be seen and not detected.
+ * Bummer!
+ */
+static bool ns2501_init(struct intel_dvo_device *dvo,
+			struct i2c_adapter *adapter)
+{
+	/* this will detect the NS2501 chip on the specified i2c bus */
+	struct ns2501_priv *ns;
+	unsigned char ch;
+
+	ns = kzalloc(sizeof(struct ns2501_priv), GFP_KERNEL);
+	if (ns == NULL)
+		return false;
+
+	dvo->i2c_bus = adapter;
+	dvo->dev_priv = ns;
+	ns->quiet = true;
+
+	if (!ns2501_readb(dvo, NS2501_VID_LO, &ch))
+		goto out;
+
+	if (ch != (NS2501_VID & 0xff)) {
+		DRM_DEBUG_KMS("ns2501 not detected got %d: from %s Slave %d.\n",
+			      ch, adapter->name, dvo->slave_addr);
+		goto out;
+	}
+
+	if (!ns2501_readb(dvo, NS2501_DID_LO, &ch))
+		goto out;
+
+	if (ch != (NS2501_DID & 0xff)) {
+		DRM_DEBUG_KMS("ns2501 not detected got %d: from %s Slave %d.\n",
+			      ch, adapter->name, dvo->slave_addr);
+		goto out;
+	}
+	ns->quiet = false;
+
+	DRM_DEBUG_KMS("init ns2501 dvo controller successfully!\n");
+
+	return true;
+
+out:
+	kfree(ns);
+	return false;
+}
+
+static enum drm_connector_status ns2501_detect(struct intel_dvo_device *dvo)
+{
+	/*
+	 * This is a Laptop display, it doesn't have hotplugging.
+	 * Even if not, the detection bit of the 2501 is unreliable as
+	 * it only works for some display types.
+	 * It is even more unreliable as the PLL must be active for
+	 * allowing reading from the chiop.
+	 */
+	return connector_status_connected;
+}
+
+static enum drm_mode_status ns2501_mode_valid(struct intel_dvo_device *dvo,
+					      struct drm_display_mode *mode)
+{
+	DRM_DEBUG_KMS
+	    ("is mode valid (hdisplay=%d,htotal=%d,vdisplay=%d,vtotal=%d)\n",
+	     mode->hdisplay, mode->htotal, mode->vdisplay, mode->vtotal);
+
+	/*
+	 * Currently, these are all the modes I have data from.
+	 * More might exist. Unclear how to find the native resolution
+	 * of the panel in here so we could always accept it
+	 * by disabling the scaler.
+	 */
+	if ((mode->hdisplay == 640 && mode->vdisplay == 480 && mode->clock == 25175) ||
+	    (mode->hdisplay == 800 && mode->vdisplay == 600 && mode->clock == 40000) ||
+	    (mode->hdisplay == 1024 && mode->vdisplay == 768 && mode->clock == 65000)) {
+		return MODE_OK;
+	} else {
+		return MODE_ONE_SIZE;	/* Is this a reasonable error? */
+	}
+}
+
+static void ns2501_mode_set(struct intel_dvo_device *dvo,
+			    const struct drm_display_mode *mode,
+			    const struct drm_display_mode *adjusted_mode)
+{
+	const struct ns2501_configuration *conf;
+	struct ns2501_priv *ns = (struct ns2501_priv *)(dvo->dev_priv);
+	int mode_idx, i;
+
+	DRM_DEBUG_KMS
+	    ("set mode (hdisplay=%d,htotal=%d,vdisplay=%d,vtotal=%d).\n",
+	     mode->hdisplay, mode->htotal, mode->vdisplay, mode->vtotal);
+
+	DRM_DEBUG_KMS("Detailed requested mode settings are:\n"
+			"clock		: %d kHz\n"
+			"hdisplay	: %d\n"
+			"hblank start	: %d\n"
+			"hblank end	: %d\n"
+			"hsync start	: %d\n"
+			"hsync end	: %d\n"
+			"htotal		: %d\n"
+			"hskew		: %d\n"
+			"vdisplay	: %d\n"
+			"vblank start	: %d\n"
+			"hblank end	: %d\n"
+			"vsync start	: %d\n"
+			"vsync end	: %d\n"
+			"vtotal		: %d\n",
+			adjusted_mode->crtc_clock,
+			adjusted_mode->crtc_hdisplay,
+			adjusted_mode->crtc_hblank_start,
+			adjusted_mode->crtc_hblank_end,
+			adjusted_mode->crtc_hsync_start,
+			adjusted_mode->crtc_hsync_end,
+			adjusted_mode->crtc_htotal,
+			adjusted_mode->crtc_hskew,
+			adjusted_mode->crtc_vdisplay,
+			adjusted_mode->crtc_vblank_start,
+			adjusted_mode->crtc_vblank_end,
+			adjusted_mode->crtc_vsync_start,
+			adjusted_mode->crtc_vsync_end,
+			adjusted_mode->crtc_vtotal);
+
+	if (mode->hdisplay == 640 && mode->vdisplay == 480)
+		mode_idx = MODE_640x480;
+	else if (mode->hdisplay == 800 && mode->vdisplay == 600)
+		mode_idx = MODE_800x600;
+	else if (mode->hdisplay == 1024 && mode->vdisplay == 768)
+		mode_idx = MODE_1024x768;
+	else
+		return;
+
+	/* Hopefully doing it every time won't hurt... */
+	for (i = 0; i < ARRAY_SIZE(regs_init); i++)
+		ns2501_writeb(dvo, regs_init[i].offset, regs_init[i].value);
+
+	/* Write the mode-agnostic values */
+	for (i = 0; i < ARRAY_SIZE(mode_agnostic_values); i++)
+		ns2501_writeb(dvo, mode_agnostic_values[i].offset,
+				mode_agnostic_values[i].value);
+
+	/* Write now the mode-specific configuration */
+	conf = ns2501_modes + mode_idx;
+	ns->conf = conf;
+
+	ns2501_writeb(dvo, NS2501_REG8, conf->conf);
+	ns2501_writeb(dvo, NS2501_REG1B, conf->pll_a);
+	ns2501_writeb(dvo, NS2501_REG1C, conf->pll_b & 0xff);
+	ns2501_writeb(dvo, NS2501_REG1D, conf->pll_b >> 8);
+	ns2501_writeb(dvo, NS2501_REGC1, conf->hstart & 0xff);
+	ns2501_writeb(dvo, NS2501_REGC2, conf->hstart >> 8);
+	ns2501_writeb(dvo, NS2501_REGC3, conf->hstop & 0xff);
+	ns2501_writeb(dvo, NS2501_REGC4, conf->hstop >> 8);
+	ns2501_writeb(dvo, NS2501_REGC5, conf->vstart & 0xff);
+	ns2501_writeb(dvo, NS2501_REGC6, conf->vstart >> 8);
+	ns2501_writeb(dvo, NS2501_REGC7, conf->vstop & 0xff);
+	ns2501_writeb(dvo, NS2501_REGC8, conf->vstop >> 8);
+	ns2501_writeb(dvo, NS2501_REG80, conf->vsync & 0xff);
+	ns2501_writeb(dvo, NS2501_REG81, conf->vsync >> 8);
+	ns2501_writeb(dvo, NS2501_REG82, conf->vtotal & 0xff);
+	ns2501_writeb(dvo, NS2501_REG83, conf->vtotal >> 8);
+	ns2501_writeb(dvo, NS2501_REG98, conf->hpos & 0xff);
+	ns2501_writeb(dvo, NS2501_REG99, conf->hpos >> 8);
+	ns2501_writeb(dvo, NS2501_REG8E, conf->vpos & 0xff);
+	ns2501_writeb(dvo, NS2501_REG8F, conf->vpos >> 8);
+	ns2501_writeb(dvo, NS2501_REG9C, conf->voffs & 0xff);
+	ns2501_writeb(dvo, NS2501_REG9D, conf->voffs >> 8);
+	ns2501_writeb(dvo, NS2501_REGB8, conf->hscale & 0xff);
+	ns2501_writeb(dvo, NS2501_REGB9, conf->hscale >> 8);
+	ns2501_writeb(dvo, NS2501_REG10, conf->vscale & 0xff);
+	ns2501_writeb(dvo, NS2501_REG11, conf->vscale >> 8);
+	ns2501_writeb(dvo, NS2501_REGF9, conf->dither);
+	ns2501_writeb(dvo, NS2501_REG41, conf->syncb);
+	ns2501_writeb(dvo, NS2501_REGC0, conf->sync);
+}
+
+/* set the NS2501 power state */
+static bool ns2501_get_hw_state(struct intel_dvo_device *dvo)
+{
+	unsigned char ch;
+
+	if (!ns2501_readb(dvo, NS2501_REG8, &ch))
+		return false;
+
+	return ch & NS2501_8_PD;
+}
+
+/* set the NS2501 power state */
+static void ns2501_dpms(struct intel_dvo_device *dvo, bool enable)
+{
+	struct ns2501_priv *ns = (struct ns2501_priv *)(dvo->dev_priv);
+
+	DRM_DEBUG_KMS("Trying set the dpms of the DVO to %i\n", enable);
+
+	if (enable) {
+		ns2501_writeb(dvo, NS2501_REGC0, ns->conf->sync | 0x08);
+
+		ns2501_writeb(dvo, NS2501_REG41, ns->conf->syncb);
+
+		ns2501_writeb(dvo, NS2501_REG34, NS2501_34_ENABLE_OUTPUT);
+		msleep(15);
+
+		ns2501_writeb(dvo, NS2501_REG8,
+				ns->conf->conf | NS2501_8_BPAS);
+		if (!(ns->conf->conf & NS2501_8_BPAS))
+			ns2501_writeb(dvo, NS2501_REG8, ns->conf->conf);
+		msleep(200);
+
+		ns2501_writeb(dvo, NS2501_REG34,
+			NS2501_34_ENABLE_OUTPUT | NS2501_34_ENABLE_BACKLIGHT);
+
+		ns2501_writeb(dvo, NS2501_REGC0, ns->conf->sync);
+	} else {
+		ns2501_writeb(dvo, NS2501_REG34, NS2501_34_ENABLE_OUTPUT);
+		msleep(200);
+
+		ns2501_writeb(dvo, NS2501_REG8, NS2501_8_VEN | NS2501_8_HEN |
+				NS2501_8_BPAS);
+		msleep(15);
+
+		ns2501_writeb(dvo, NS2501_REG34, 0x00);
+	}
+}
+
+static void ns2501_destroy(struct intel_dvo_device *dvo)
+{
+	struct ns2501_priv *ns = dvo->dev_priv;
+
+	if (ns) {
+		kfree(ns);
+		dvo->dev_priv = NULL;
+	}
+}
+
+const struct intel_dvo_dev_ops ns2501_ops = {
+	.init = ns2501_init,
+	.detect = ns2501_detect,
+	.mode_valid = ns2501_mode_valid,
+	.mode_set = ns2501_mode_set,
+	.dpms = ns2501_dpms,
+	.get_hw_state = ns2501_get_hw_state,
+	.destroy = ns2501_destroy,
+};
diff --git a/drivers/gpu/drm/i915/display/dvo_sil164.c b/drivers/gpu/drm/i915/display/dvo_sil164.c
new file mode 100644
index 000000000..0dfa0a020
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/dvo_sil164.c
@@ -0,0 +1,280 @@
+/**************************************************************************
+
+Copyright © 2006 Dave Airlie
+
+All Rights Reserved.
+
+Permission is hereby granted, free of charge, to any person obtaining a
+copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sub license, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject to
+the following conditions:
+
+The above copyright notice and this permission notice (including the
+next paragraph) shall be included in all copies or substantial portions
+of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+**************************************************************************/
+
+#include "intel_display_types.h"
+#include "intel_dvo_dev.h"
+
+#define SIL164_VID 0x0001
+#define SIL164_DID 0x0006
+
+#define SIL164_VID_LO 0x00
+#define SIL164_VID_HI 0x01
+#define SIL164_DID_LO 0x02
+#define SIL164_DID_HI 0x03
+#define SIL164_REV    0x04
+#define SIL164_RSVD   0x05
+#define SIL164_FREQ_LO 0x06
+#define SIL164_FREQ_HI 0x07
+
+#define SIL164_REG8 0x08
+#define SIL164_8_VEN (1<<5)
+#define SIL164_8_HEN (1<<4)
+#define SIL164_8_DSEL (1<<3)
+#define SIL164_8_BSEL (1<<2)
+#define SIL164_8_EDGE (1<<1)
+#define SIL164_8_PD   (1<<0)
+
+#define SIL164_REG9 0x09
+#define SIL164_9_VLOW (1<<7)
+#define SIL164_9_MSEL_MASK (0x7<<4)
+#define SIL164_9_TSEL (1<<3)
+#define SIL164_9_RSEN (1<<2)
+#define SIL164_9_HTPLG (1<<1)
+#define SIL164_9_MDI (1<<0)
+
+#define SIL164_REGC 0x0c
+
+struct sil164_priv {
+	//I2CDevRec d;
+	bool quiet;
+};
+
+#define SILPTR(d) ((SIL164Ptr)(d->DriverPrivate.ptr))
+
+static bool sil164_readb(struct intel_dvo_device *dvo, int addr, u8 *ch)
+{
+	struct sil164_priv *sil = dvo->dev_priv;
+	struct i2c_adapter *adapter = dvo->i2c_bus;
+	u8 out_buf[2];
+	u8 in_buf[2];
+
+	struct i2c_msg msgs[] = {
+		{
+			.addr = dvo->slave_addr,
+			.flags = 0,
+			.len = 1,
+			.buf = out_buf,
+		},
+		{
+			.addr = dvo->slave_addr,
+			.flags = I2C_M_RD,
+			.len = 1,
+			.buf = in_buf,
+		}
+	};
+
+	out_buf[0] = addr;
+	out_buf[1] = 0;
+
+	if (i2c_transfer(adapter, msgs, 2) == 2) {
+		*ch = in_buf[0];
+		return true;
+	}
+
+	if (!sil->quiet) {
+		DRM_DEBUG_KMS("Unable to read register 0x%02x from %s:%02x.\n",
+			  addr, adapter->name, dvo->slave_addr);
+	}
+	return false;
+}
+
+static bool sil164_writeb(struct intel_dvo_device *dvo, int addr, u8 ch)
+{
+	struct sil164_priv *sil = dvo->dev_priv;
+	struct i2c_adapter *adapter = dvo->i2c_bus;
+	u8 out_buf[2];
+	struct i2c_msg msg = {
+		.addr = dvo->slave_addr,
+		.flags = 0,
+		.len = 2,
+		.buf = out_buf,
+	};
+
+	out_buf[0] = addr;
+	out_buf[1] = ch;
+
+	if (i2c_transfer(adapter, &msg, 1) == 1)
+		return true;
+
+	if (!sil->quiet) {
+		DRM_DEBUG_KMS("Unable to write register 0x%02x to %s:%d.\n",
+			  addr, adapter->name, dvo->slave_addr);
+	}
+
+	return false;
+}
+
+/* Silicon Image 164 driver for chip on i2c bus */
+static bool sil164_init(struct intel_dvo_device *dvo,
+			struct i2c_adapter *adapter)
+{
+	/* this will detect the SIL164 chip on the specified i2c bus */
+	struct sil164_priv *sil;
+	unsigned char ch;
+
+	sil = kzalloc(sizeof(struct sil164_priv), GFP_KERNEL);
+	if (sil == NULL)
+		return false;
+
+	dvo->i2c_bus = adapter;
+	dvo->dev_priv = sil;
+	sil->quiet = true;
+
+	if (!sil164_readb(dvo, SIL164_VID_LO, &ch))
+		goto out;
+
+	if (ch != (SIL164_VID & 0xff)) {
+		DRM_DEBUG_KMS("sil164 not detected got %d: from %s Slave %d.\n",
+			  ch, adapter->name, dvo->slave_addr);
+		goto out;
+	}
+
+	if (!sil164_readb(dvo, SIL164_DID_LO, &ch))
+		goto out;
+
+	if (ch != (SIL164_DID & 0xff)) {
+		DRM_DEBUG_KMS("sil164 not detected got %d: from %s Slave %d.\n",
+			  ch, adapter->name, dvo->slave_addr);
+		goto out;
+	}
+	sil->quiet = false;
+
+	DRM_DEBUG_KMS("init sil164 dvo controller successfully!\n");
+	return true;
+
+out:
+	kfree(sil);
+	return false;
+}
+
+static enum drm_connector_status sil164_detect(struct intel_dvo_device *dvo)
+{
+	u8 reg9;
+
+	sil164_readb(dvo, SIL164_REG9, &reg9);
+
+	if (reg9 & SIL164_9_HTPLG)
+		return connector_status_connected;
+	else
+		return connector_status_disconnected;
+}
+
+static enum drm_mode_status sil164_mode_valid(struct intel_dvo_device *dvo,
+					      struct drm_display_mode *mode)
+{
+	return MODE_OK;
+}
+
+static void sil164_mode_set(struct intel_dvo_device *dvo,
+			    const struct drm_display_mode *mode,
+			    const struct drm_display_mode *adjusted_mode)
+{
+	/* As long as the basics are set up, since we don't have clock
+	 * dependencies in the mode setup, we can just leave the
+	 * registers alone and everything will work fine.
+	 */
+	/* recommended programming sequence from doc */
+	/*sil164_writeb(sil, 0x08, 0x30);
+	  sil164_writeb(sil, 0x09, 0x00);
+	  sil164_writeb(sil, 0x0a, 0x90);
+	  sil164_writeb(sil, 0x0c, 0x89);
+	  sil164_writeb(sil, 0x08, 0x31);*/
+	/* don't do much */
+	return;
+}
+
+/* set the SIL164 power state */
+static void sil164_dpms(struct intel_dvo_device *dvo, bool enable)
+{
+	int ret;
+	unsigned char ch;
+
+	ret = sil164_readb(dvo, SIL164_REG8, &ch);
+	if (ret == false)
+		return;
+
+	if (enable)
+		ch |= SIL164_8_PD;
+	else
+		ch &= ~SIL164_8_PD;
+
+	sil164_writeb(dvo, SIL164_REG8, ch);
+	return;
+}
+
+static bool sil164_get_hw_state(struct intel_dvo_device *dvo)
+{
+	int ret;
+	unsigned char ch;
+
+	ret = sil164_readb(dvo, SIL164_REG8, &ch);
+	if (ret == false)
+		return false;
+
+	if (ch & SIL164_8_PD)
+		return true;
+	else
+		return false;
+}
+
+static void sil164_dump_regs(struct intel_dvo_device *dvo)
+{
+	u8 val;
+
+	sil164_readb(dvo, SIL164_FREQ_LO, &val);
+	DRM_DEBUG_KMS("SIL164_FREQ_LO: 0x%02x\n", val);
+	sil164_readb(dvo, SIL164_FREQ_HI, &val);
+	DRM_DEBUG_KMS("SIL164_FREQ_HI: 0x%02x\n", val);
+	sil164_readb(dvo, SIL164_REG8, &val);
+	DRM_DEBUG_KMS("SIL164_REG8: 0x%02x\n", val);
+	sil164_readb(dvo, SIL164_REG9, &val);
+	DRM_DEBUG_KMS("SIL164_REG9: 0x%02x\n", val);
+	sil164_readb(dvo, SIL164_REGC, &val);
+	DRM_DEBUG_KMS("SIL164_REGC: 0x%02x\n", val);
+}
+
+static void sil164_destroy(struct intel_dvo_device *dvo)
+{
+	struct sil164_priv *sil = dvo->dev_priv;
+
+	if (sil) {
+		kfree(sil);
+		dvo->dev_priv = NULL;
+	}
+}
+
+const struct intel_dvo_dev_ops sil164_ops = {
+	.init = sil164_init,
+	.detect = sil164_detect,
+	.mode_valid = sil164_mode_valid,
+	.mode_set = sil164_mode_set,
+	.dpms = sil164_dpms,
+	.get_hw_state = sil164_get_hw_state,
+	.dump_regs = sil164_dump_regs,
+	.destroy = sil164_destroy,
+};
diff --git a/drivers/gpu/drm/i915/display/dvo_tfp410.c b/drivers/gpu/drm/i915/display/dvo_tfp410.c
new file mode 100644
index 000000000..009d65b0f
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/dvo_tfp410.c
@@ -0,0 +1,319 @@
+/*
+ * Copyright © 2007 Dave Mueller
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Dave Mueller <dave.mueller@gmx.ch>
+ *
+ */
+
+#include "intel_display_types.h"
+#include "intel_dvo_dev.h"
+
+/* register definitions according to the TFP410 data sheet */
+#define TFP410_VID		0x014C
+#define TFP410_DID		0x0410
+
+#define TFP410_VID_LO		0x00
+#define TFP410_VID_HI		0x01
+#define TFP410_DID_LO		0x02
+#define TFP410_DID_HI		0x03
+#define TFP410_REV		0x04
+
+#define TFP410_CTL_1		0x08
+#define TFP410_CTL_1_TDIS	(1<<6)
+#define TFP410_CTL_1_VEN	(1<<5)
+#define TFP410_CTL_1_HEN	(1<<4)
+#define TFP410_CTL_1_DSEL	(1<<3)
+#define TFP410_CTL_1_BSEL	(1<<2)
+#define TFP410_CTL_1_EDGE	(1<<1)
+#define TFP410_CTL_1_PD		(1<<0)
+
+#define TFP410_CTL_2		0x09
+#define TFP410_CTL_2_VLOW	(1<<7)
+#define TFP410_CTL_2_MSEL_MASK	(0x7<<4)
+#define TFP410_CTL_2_MSEL	(1<<4)
+#define TFP410_CTL_2_TSEL	(1<<3)
+#define TFP410_CTL_2_RSEN	(1<<2)
+#define TFP410_CTL_2_HTPLG	(1<<1)
+#define TFP410_CTL_2_MDI	(1<<0)
+
+#define TFP410_CTL_3		0x0A
+#define TFP410_CTL_3_DK_MASK	(0x7<<5)
+#define TFP410_CTL_3_DK		(1<<5)
+#define TFP410_CTL_3_DKEN	(1<<4)
+#define TFP410_CTL_3_CTL_MASK	(0x7<<1)
+#define TFP410_CTL_3_CTL	(1<<1)
+
+#define TFP410_USERCFG		0x0B
+
+#define TFP410_DE_DLY		0x32
+
+#define TFP410_DE_CTL		0x33
+#define TFP410_DE_CTL_DEGEN	(1<<6)
+#define TFP410_DE_CTL_VSPOL	(1<<5)
+#define TFP410_DE_CTL_HSPOL	(1<<4)
+#define TFP410_DE_CTL_DEDLY8	(1<<0)
+
+#define TFP410_DE_TOP		0x34
+
+#define TFP410_DE_CNT_LO	0x36
+#define TFP410_DE_CNT_HI	0x37
+
+#define TFP410_DE_LIN_LO	0x38
+#define TFP410_DE_LIN_HI	0x39
+
+#define TFP410_H_RES_LO		0x3A
+#define TFP410_H_RES_HI		0x3B
+
+#define TFP410_V_RES_LO		0x3C
+#define TFP410_V_RES_HI		0x3D
+
+struct tfp410_priv {
+	bool quiet;
+};
+
+static bool tfp410_readb(struct intel_dvo_device *dvo, int addr, u8 *ch)
+{
+	struct tfp410_priv *tfp = dvo->dev_priv;
+	struct i2c_adapter *adapter = dvo->i2c_bus;
+	u8 out_buf[2];
+	u8 in_buf[2];
+
+	struct i2c_msg msgs[] = {
+		{
+			.addr = dvo->slave_addr,
+			.flags = 0,
+			.len = 1,
+			.buf = out_buf,
+		},
+		{
+			.addr = dvo->slave_addr,
+			.flags = I2C_M_RD,
+			.len = 1,
+			.buf = in_buf,
+		}
+	};
+
+	out_buf[0] = addr;
+	out_buf[1] = 0;
+
+	if (i2c_transfer(adapter, msgs, 2) == 2) {
+		*ch = in_buf[0];
+		return true;
+	}
+
+	if (!tfp->quiet) {
+		DRM_DEBUG_KMS("Unable to read register 0x%02x from %s:%02x.\n",
+			  addr, adapter->name, dvo->slave_addr);
+	}
+	return false;
+}
+
+static bool tfp410_writeb(struct intel_dvo_device *dvo, int addr, u8 ch)
+{
+	struct tfp410_priv *tfp = dvo->dev_priv;
+	struct i2c_adapter *adapter = dvo->i2c_bus;
+	u8 out_buf[2];
+	struct i2c_msg msg = {
+		.addr = dvo->slave_addr,
+		.flags = 0,
+		.len = 2,
+		.buf = out_buf,
+	};
+
+	out_buf[0] = addr;
+	out_buf[1] = ch;
+
+	if (i2c_transfer(adapter, &msg, 1) == 1)
+		return true;
+
+	if (!tfp->quiet) {
+		DRM_DEBUG_KMS("Unable to write register 0x%02x to %s:%d.\n",
+			  addr, adapter->name, dvo->slave_addr);
+	}
+
+	return false;
+}
+
+static int tfp410_getid(struct intel_dvo_device *dvo, int addr)
+{
+	u8 ch1, ch2;
+
+	if (tfp410_readb(dvo, addr+0, &ch1) &&
+	    tfp410_readb(dvo, addr+1, &ch2))
+		return ((ch2 << 8) & 0xFF00) | (ch1 & 0x00FF);
+
+	return -1;
+}
+
+/* Ti TFP410 driver for chip on i2c bus */
+static bool tfp410_init(struct intel_dvo_device *dvo,
+			struct i2c_adapter *adapter)
+{
+	/* this will detect the tfp410 chip on the specified i2c bus */
+	struct tfp410_priv *tfp;
+	int id;
+
+	tfp = kzalloc(sizeof(struct tfp410_priv), GFP_KERNEL);
+	if (tfp == NULL)
+		return false;
+
+	dvo->i2c_bus = adapter;
+	dvo->dev_priv = tfp;
+	tfp->quiet = true;
+
+	if ((id = tfp410_getid(dvo, TFP410_VID_LO)) != TFP410_VID) {
+		DRM_DEBUG_KMS("tfp410 not detected got VID %X: from %s "
+				"Slave %d.\n",
+			  id, adapter->name, dvo->slave_addr);
+		goto out;
+	}
+
+	if ((id = tfp410_getid(dvo, TFP410_DID_LO)) != TFP410_DID) {
+		DRM_DEBUG_KMS("tfp410 not detected got DID %X: from %s "
+				"Slave %d.\n",
+			  id, adapter->name, dvo->slave_addr);
+		goto out;
+	}
+	tfp->quiet = false;
+	return true;
+out:
+	kfree(tfp);
+	return false;
+}
+
+static enum drm_connector_status tfp410_detect(struct intel_dvo_device *dvo)
+{
+	enum drm_connector_status ret = connector_status_disconnected;
+	u8 ctl2;
+
+	if (tfp410_readb(dvo, TFP410_CTL_2, &ctl2)) {
+		if (ctl2 & TFP410_CTL_2_RSEN)
+			ret = connector_status_connected;
+		else
+			ret = connector_status_disconnected;
+	}
+
+	return ret;
+}
+
+static enum drm_mode_status tfp410_mode_valid(struct intel_dvo_device *dvo,
+					      struct drm_display_mode *mode)
+{
+	return MODE_OK;
+}
+
+static void tfp410_mode_set(struct intel_dvo_device *dvo,
+			    const struct drm_display_mode *mode,
+			    const struct drm_display_mode *adjusted_mode)
+{
+	/* As long as the basics are set up, since we don't have clock dependencies
+	* in the mode setup, we can just leave the registers alone and everything
+	* will work fine.
+	*/
+	/* don't do much */
+	return;
+}
+
+/* set the tfp410 power state */
+static void tfp410_dpms(struct intel_dvo_device *dvo, bool enable)
+{
+	u8 ctl1;
+
+	if (!tfp410_readb(dvo, TFP410_CTL_1, &ctl1))
+		return;
+
+	if (enable)
+		ctl1 |= TFP410_CTL_1_PD;
+	else
+		ctl1 &= ~TFP410_CTL_1_PD;
+
+	tfp410_writeb(dvo, TFP410_CTL_1, ctl1);
+}
+
+static bool tfp410_get_hw_state(struct intel_dvo_device *dvo)
+{
+	u8 ctl1;
+
+	if (!tfp410_readb(dvo, TFP410_CTL_1, &ctl1))
+		return false;
+
+	if (ctl1 & TFP410_CTL_1_PD)
+		return true;
+	else
+		return false;
+}
+
+static void tfp410_dump_regs(struct intel_dvo_device *dvo)
+{
+	u8 val, val2;
+
+	tfp410_readb(dvo, TFP410_REV, &val);
+	DRM_DEBUG_KMS("TFP410_REV: 0x%02X\n", val);
+	tfp410_readb(dvo, TFP410_CTL_1, &val);
+	DRM_DEBUG_KMS("TFP410_CTL1: 0x%02X\n", val);
+	tfp410_readb(dvo, TFP410_CTL_2, &val);
+	DRM_DEBUG_KMS("TFP410_CTL2: 0x%02X\n", val);
+	tfp410_readb(dvo, TFP410_CTL_3, &val);
+	DRM_DEBUG_KMS("TFP410_CTL3: 0x%02X\n", val);
+	tfp410_readb(dvo, TFP410_USERCFG, &val);
+	DRM_DEBUG_KMS("TFP410_USERCFG: 0x%02X\n", val);
+	tfp410_readb(dvo, TFP410_DE_DLY, &val);
+	DRM_DEBUG_KMS("TFP410_DE_DLY: 0x%02X\n", val);
+	tfp410_readb(dvo, TFP410_DE_CTL, &val);
+	DRM_DEBUG_KMS("TFP410_DE_CTL: 0x%02X\n", val);
+	tfp410_readb(dvo, TFP410_DE_TOP, &val);
+	DRM_DEBUG_KMS("TFP410_DE_TOP: 0x%02X\n", val);
+	tfp410_readb(dvo, TFP410_DE_CNT_LO, &val);
+	tfp410_readb(dvo, TFP410_DE_CNT_HI, &val2);
+	DRM_DEBUG_KMS("TFP410_DE_CNT: 0x%02X%02X\n", val2, val);
+	tfp410_readb(dvo, TFP410_DE_LIN_LO, &val);
+	tfp410_readb(dvo, TFP410_DE_LIN_HI, &val2);
+	DRM_DEBUG_KMS("TFP410_DE_LIN: 0x%02X%02X\n", val2, val);
+	tfp410_readb(dvo, TFP410_H_RES_LO, &val);
+	tfp410_readb(dvo, TFP410_H_RES_HI, &val2);
+	DRM_DEBUG_KMS("TFP410_H_RES: 0x%02X%02X\n", val2, val);
+	tfp410_readb(dvo, TFP410_V_RES_LO, &val);
+	tfp410_readb(dvo, TFP410_V_RES_HI, &val2);
+	DRM_DEBUG_KMS("TFP410_V_RES: 0x%02X%02X\n", val2, val);
+}
+
+static void tfp410_destroy(struct intel_dvo_device *dvo)
+{
+	struct tfp410_priv *tfp = dvo->dev_priv;
+
+	if (tfp) {
+		kfree(tfp);
+		dvo->dev_priv = NULL;
+	}
+}
+
+const struct intel_dvo_dev_ops tfp410_ops = {
+	.init = tfp410_init,
+	.detect = tfp410_detect,
+	.mode_valid = tfp410_mode_valid,
+	.mode_set = tfp410_mode_set,
+	.dpms = tfp410_dpms,
+	.get_hw_state = tfp410_get_hw_state,
+	.dump_regs = tfp410_dump_regs,
+	.destroy = tfp410_destroy,
+};
diff --git a/drivers/gpu/drm/i915/display/g4x_dp.c b/drivers/gpu/drm/i915/display/g4x_dp.c
new file mode 100644
index 000000000..e3e3d27ff
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/g4x_dp.c
@@ -0,0 +1,1404 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ *
+ * DisplayPort support for G4x,ILK,SNB,IVB,VLV,CHV (HSW+ handled by the DDI code).
+ */
+
+#include <linux/string_helpers.h>
+
+#include "g4x_dp.h"
+#include "intel_audio.h"
+#include "intel_backlight.h"
+#include "intel_connector.h"
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display_power.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_dp_link_training.h"
+#include "intel_dpio_phy.h"
+#include "intel_fifo_underrun.h"
+#include "intel_hdmi.h"
+#include "intel_hotplug.h"
+#include "intel_pch_display.h"
+#include "intel_pps.h"
+#include "vlv_sideband.h"
+
+static const struct dpll g4x_dpll[] = {
+	{ .dot = 162000, .p1 = 2, .p2 = 10, .n = 2, .m1 = 23, .m2 = 8, },
+	{ .dot = 270000, .p1 = 1, .p2 = 10, .n = 1, .m1 = 14, .m2 = 2, },
+};
+
+static const struct dpll pch_dpll[] = {
+	{ .dot = 162000, .p1 = 2, .p2 = 10, .n = 1, .m1 = 12, .m2 = 9, },
+	{ .dot = 270000, .p1 = 1, .p2 = 10, .n = 2, .m1 = 14, .m2 = 8, },
+};
+
+static const struct dpll vlv_dpll[] = {
+	{ .dot = 162000, .p1 = 3, .p2 = 2, .n = 5, .m1 = 3, .m2 = 81, },
+	{ .dot = 270000, .p1 = 2, .p2 = 2, .n = 1, .m1 = 2, .m2 = 27, },
+};
+
+static const struct dpll chv_dpll[] = {
+	/* m2 is .22 binary fixed point  */
+	{ .dot = 162000, .p1 = 4, .p2 = 2, .n = 1, .m1 = 2, .m2 = 0x819999a /* 32.4 */ },
+	{ .dot = 270000, .p1 = 4, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6c00000 /* 27.0 */ },
+};
+
+const struct dpll *vlv_get_dpll(struct drm_i915_private *i915)
+{
+	return IS_CHERRYVIEW(i915) ? &chv_dpll[0] : &vlv_dpll[0];
+}
+
+void g4x_dp_set_clock(struct intel_encoder *encoder,
+		      struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	const struct dpll *divisor = NULL;
+	int i, count = 0;
+
+	if (IS_G4X(dev_priv)) {
+		divisor = g4x_dpll;
+		count = ARRAY_SIZE(g4x_dpll);
+	} else if (HAS_PCH_SPLIT(dev_priv)) {
+		divisor = pch_dpll;
+		count = ARRAY_SIZE(pch_dpll);
+	} else if (IS_CHERRYVIEW(dev_priv)) {
+		divisor = chv_dpll;
+		count = ARRAY_SIZE(chv_dpll);
+	} else if (IS_VALLEYVIEW(dev_priv)) {
+		divisor = vlv_dpll;
+		count = ARRAY_SIZE(vlv_dpll);
+	}
+
+	if (divisor && count) {
+		for (i = 0; i < count; i++) {
+			if (pipe_config->port_clock == divisor[i].dot) {
+				pipe_config->dpll = divisor[i];
+				pipe_config->clock_set = true;
+				break;
+			}
+		}
+	}
+}
+
+static void intel_dp_prepare(struct intel_encoder *encoder,
+			     const struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	enum port port = encoder->port;
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	const struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+
+	intel_dp_set_link_params(intel_dp,
+				 pipe_config->port_clock,
+				 pipe_config->lane_count);
+
+	/*
+	 * There are four kinds of DP registers:
+	 * IBX PCH
+	 * SNB CPU
+	 * IVB CPU
+	 * CPT PCH
+	 *
+	 * IBX PCH and CPU are the same for almost everything,
+	 * except that the CPU DP PLL is configured in this
+	 * register
+	 *
+	 * CPT PCH is quite different, having many bits moved
+	 * to the TRANS_DP_CTL register instead. That
+	 * configuration happens (oddly) in ilk_pch_enable
+	 */
+
+	/* Preserve the BIOS-computed detected bit. This is
+	 * supposed to be read-only.
+	 */
+	intel_dp->DP = intel_de_read(dev_priv, intel_dp->output_reg) & DP_DETECTED;
+
+	/* Handle DP bits in common between all three register formats */
+	intel_dp->DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
+	intel_dp->DP |= DP_PORT_WIDTH(pipe_config->lane_count);
+
+	/* Split out the IBX/CPU vs CPT settings */
+
+	if (IS_IVYBRIDGE(dev_priv) && port == PORT_A) {
+		if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
+			intel_dp->DP |= DP_SYNC_HS_HIGH;
+		if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
+			intel_dp->DP |= DP_SYNC_VS_HIGH;
+		intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT;
+
+		if (drm_dp_enhanced_frame_cap(intel_dp->dpcd))
+			intel_dp->DP |= DP_ENHANCED_FRAMING;
+
+		intel_dp->DP |= DP_PIPE_SEL_IVB(crtc->pipe);
+	} else if (HAS_PCH_CPT(dev_priv) && port != PORT_A) {
+		u32 trans_dp;
+
+		intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT;
+
+		trans_dp = intel_de_read(dev_priv, TRANS_DP_CTL(crtc->pipe));
+		if (drm_dp_enhanced_frame_cap(intel_dp->dpcd))
+			trans_dp |= TRANS_DP_ENH_FRAMING;
+		else
+			trans_dp &= ~TRANS_DP_ENH_FRAMING;
+		intel_de_write(dev_priv, TRANS_DP_CTL(crtc->pipe), trans_dp);
+	} else {
+		if (IS_G4X(dev_priv) && pipe_config->limited_color_range)
+			intel_dp->DP |= DP_COLOR_RANGE_16_235;
+
+		if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
+			intel_dp->DP |= DP_SYNC_HS_HIGH;
+		if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
+			intel_dp->DP |= DP_SYNC_VS_HIGH;
+		intel_dp->DP |= DP_LINK_TRAIN_OFF;
+
+		if (drm_dp_enhanced_frame_cap(intel_dp->dpcd))
+			intel_dp->DP |= DP_ENHANCED_FRAMING;
+
+		if (IS_CHERRYVIEW(dev_priv))
+			intel_dp->DP |= DP_PIPE_SEL_CHV(crtc->pipe);
+		else
+			intel_dp->DP |= DP_PIPE_SEL(crtc->pipe);
+	}
+}
+
+static void assert_dp_port(struct intel_dp *intel_dp, bool state)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
+	bool cur_state = intel_de_read(dev_priv, intel_dp->output_reg) & DP_PORT_EN;
+
+	I915_STATE_WARN(cur_state != state,
+			"[ENCODER:%d:%s] state assertion failure (expected %s, current %s)\n",
+			dig_port->base.base.base.id, dig_port->base.base.name,
+			str_on_off(state), str_on_off(cur_state));
+}
+#define assert_dp_port_disabled(d) assert_dp_port((d), false)
+
+static void assert_edp_pll(struct drm_i915_private *dev_priv, bool state)
+{
+	bool cur_state = intel_de_read(dev_priv, DP_A) & DP_PLL_ENABLE;
+
+	I915_STATE_WARN(cur_state != state,
+			"eDP PLL state assertion failure (expected %s, current %s)\n",
+			str_on_off(state), str_on_off(cur_state));
+}
+#define assert_edp_pll_enabled(d) assert_edp_pll((d), true)
+#define assert_edp_pll_disabled(d) assert_edp_pll((d), false)
+
+static void ilk_edp_pll_on(struct intel_dp *intel_dp,
+			   const struct intel_crtc_state *pipe_config)
+{
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	assert_transcoder_disabled(dev_priv, pipe_config->cpu_transcoder);
+	assert_dp_port_disabled(intel_dp);
+	assert_edp_pll_disabled(dev_priv);
+
+	drm_dbg_kms(&dev_priv->drm, "enabling eDP PLL for clock %d\n",
+		    pipe_config->port_clock);
+
+	intel_dp->DP &= ~DP_PLL_FREQ_MASK;
+
+	if (pipe_config->port_clock == 162000)
+		intel_dp->DP |= DP_PLL_FREQ_162MHZ;
+	else
+		intel_dp->DP |= DP_PLL_FREQ_270MHZ;
+
+	intel_de_write(dev_priv, DP_A, intel_dp->DP);
+	intel_de_posting_read(dev_priv, DP_A);
+	udelay(500);
+
+	/*
+	 * [DevILK] Work around required when enabling DP PLL
+	 * while a pipe is enabled going to FDI:
+	 * 1. Wait for the start of vertical blank on the enabled pipe going to FDI
+	 * 2. Program DP PLL enable
+	 */
+	if (IS_IRONLAKE(dev_priv))
+		intel_wait_for_vblank_if_active(dev_priv, !crtc->pipe);
+
+	intel_dp->DP |= DP_PLL_ENABLE;
+
+	intel_de_write(dev_priv, DP_A, intel_dp->DP);
+	intel_de_posting_read(dev_priv, DP_A);
+	udelay(200);
+}
+
+static void ilk_edp_pll_off(struct intel_dp *intel_dp,
+			    const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	assert_transcoder_disabled(dev_priv, old_crtc_state->cpu_transcoder);
+	assert_dp_port_disabled(intel_dp);
+	assert_edp_pll_enabled(dev_priv);
+
+	drm_dbg_kms(&dev_priv->drm, "disabling eDP PLL\n");
+
+	intel_dp->DP &= ~DP_PLL_ENABLE;
+
+	intel_de_write(dev_priv, DP_A, intel_dp->DP);
+	intel_de_posting_read(dev_priv, DP_A);
+	udelay(200);
+}
+
+static bool cpt_dp_port_selected(struct drm_i915_private *dev_priv,
+				 enum port port, enum pipe *pipe)
+{
+	enum pipe p;
+
+	for_each_pipe(dev_priv, p) {
+		u32 val = intel_de_read(dev_priv, TRANS_DP_CTL(p));
+
+		if ((val & TRANS_DP_PORT_SEL_MASK) == TRANS_DP_PORT_SEL(port)) {
+			*pipe = p;
+			return true;
+		}
+	}
+
+	drm_dbg_kms(&dev_priv->drm, "No pipe for DP port %c found\n",
+		    port_name(port));
+
+	/* must initialize pipe to something for the asserts */
+	*pipe = PIPE_A;
+
+	return false;
+}
+
+bool g4x_dp_port_enabled(struct drm_i915_private *dev_priv,
+			 i915_reg_t dp_reg, enum port port,
+			 enum pipe *pipe)
+{
+	bool ret;
+	u32 val;
+
+	val = intel_de_read(dev_priv, dp_reg);
+
+	ret = val & DP_PORT_EN;
+
+	/* asserts want to know the pipe even if the port is disabled */
+	if (IS_IVYBRIDGE(dev_priv) && port == PORT_A)
+		*pipe = (val & DP_PIPE_SEL_MASK_IVB) >> DP_PIPE_SEL_SHIFT_IVB;
+	else if (HAS_PCH_CPT(dev_priv) && port != PORT_A)
+		ret &= cpt_dp_port_selected(dev_priv, port, pipe);
+	else if (IS_CHERRYVIEW(dev_priv))
+		*pipe = (val & DP_PIPE_SEL_MASK_CHV) >> DP_PIPE_SEL_SHIFT_CHV;
+	else
+		*pipe = (val & DP_PIPE_SEL_MASK) >> DP_PIPE_SEL_SHIFT;
+
+	return ret;
+}
+
+static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
+				  enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	intel_wakeref_t wakeref;
+	bool ret;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     encoder->power_domain);
+	if (!wakeref)
+		return false;
+
+	ret = g4x_dp_port_enabled(dev_priv, intel_dp->output_reg,
+				  encoder->port, pipe);
+
+	intel_display_power_put(dev_priv, encoder->power_domain, wakeref);
+
+	return ret;
+}
+
+static void g4x_dp_get_m_n(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	if (crtc_state->has_pch_encoder) {
+		intel_pch_transcoder_get_m1_n1(crtc, &crtc_state->dp_m_n);
+		intel_pch_transcoder_get_m2_n2(crtc, &crtc_state->dp_m2_n2);
+	} else {
+		intel_cpu_transcoder_get_m1_n1(crtc, crtc_state->cpu_transcoder,
+					       &crtc_state->dp_m_n);
+		intel_cpu_transcoder_get_m2_n2(crtc, crtc_state->cpu_transcoder,
+					       &crtc_state->dp_m2_n2);
+	}
+}
+
+static void intel_dp_get_config(struct intel_encoder *encoder,
+				struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	u32 tmp, flags = 0;
+	enum port port = encoder->port;
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+
+	if (encoder->type == INTEL_OUTPUT_EDP)
+		pipe_config->output_types |= BIT(INTEL_OUTPUT_EDP);
+	else
+		pipe_config->output_types |= BIT(INTEL_OUTPUT_DP);
+
+	tmp = intel_de_read(dev_priv, intel_dp->output_reg);
+
+	pipe_config->has_audio = tmp & DP_AUDIO_OUTPUT_ENABLE && port != PORT_A;
+
+	if (HAS_PCH_CPT(dev_priv) && port != PORT_A) {
+		u32 trans_dp = intel_de_read(dev_priv,
+					     TRANS_DP_CTL(crtc->pipe));
+
+		if (trans_dp & TRANS_DP_HSYNC_ACTIVE_HIGH)
+			flags |= DRM_MODE_FLAG_PHSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NHSYNC;
+
+		if (trans_dp & TRANS_DP_VSYNC_ACTIVE_HIGH)
+			flags |= DRM_MODE_FLAG_PVSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NVSYNC;
+	} else {
+		if (tmp & DP_SYNC_HS_HIGH)
+			flags |= DRM_MODE_FLAG_PHSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NHSYNC;
+
+		if (tmp & DP_SYNC_VS_HIGH)
+			flags |= DRM_MODE_FLAG_PVSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NVSYNC;
+	}
+
+	pipe_config->hw.adjusted_mode.flags |= flags;
+
+	if (IS_G4X(dev_priv) && tmp & DP_COLOR_RANGE_16_235)
+		pipe_config->limited_color_range = true;
+
+	pipe_config->lane_count =
+		((tmp & DP_PORT_WIDTH_MASK) >> DP_PORT_WIDTH_SHIFT) + 1;
+
+	g4x_dp_get_m_n(pipe_config);
+
+	if (port == PORT_A) {
+		if ((intel_de_read(dev_priv, DP_A) & DP_PLL_FREQ_MASK) == DP_PLL_FREQ_162MHZ)
+			pipe_config->port_clock = 162000;
+		else
+			pipe_config->port_clock = 270000;
+	}
+
+	pipe_config->hw.adjusted_mode.crtc_clock =
+		intel_dotclock_calculate(pipe_config->port_clock,
+					 &pipe_config->dp_m_n);
+
+	if (intel_dp_is_edp(intel_dp))
+		intel_edp_fixup_vbt_bpp(encoder, pipe_config->pipe_bpp);
+}
+
+static void
+intel_dp_link_down(struct intel_encoder *encoder,
+		   const struct intel_crtc_state *old_crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	enum port port = encoder->port;
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			(intel_de_read(dev_priv, intel_dp->output_reg) &
+			 DP_PORT_EN) == 0))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	if ((IS_IVYBRIDGE(dev_priv) && port == PORT_A) ||
+	    (HAS_PCH_CPT(dev_priv) && port != PORT_A)) {
+		intel_dp->DP &= ~DP_LINK_TRAIN_MASK_CPT;
+		intel_dp->DP |= DP_LINK_TRAIN_PAT_IDLE_CPT;
+	} else {
+		intel_dp->DP &= ~DP_LINK_TRAIN_MASK;
+		intel_dp->DP |= DP_LINK_TRAIN_PAT_IDLE;
+	}
+	intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+
+	intel_dp->DP &= ~(DP_PORT_EN | DP_AUDIO_OUTPUT_ENABLE);
+	intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+
+	/*
+	 * HW workaround for IBX, we need to move the port
+	 * to transcoder A after disabling it to allow the
+	 * matching HDMI port to be enabled on transcoder A.
+	 */
+	if (HAS_PCH_IBX(dev_priv) && crtc->pipe == PIPE_B && port != PORT_A) {
+		/*
+		 * We get CPU/PCH FIFO underruns on the other pipe when
+		 * doing the workaround. Sweep them under the rug.
+		 */
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, PIPE_A, false);
+		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, false);
+
+		/* always enable with pattern 1 (as per spec) */
+		intel_dp->DP &= ~(DP_PIPE_SEL_MASK | DP_LINK_TRAIN_MASK);
+		intel_dp->DP |= DP_PORT_EN | DP_PIPE_SEL(PIPE_A) |
+			DP_LINK_TRAIN_PAT_1;
+		intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP);
+		intel_de_posting_read(dev_priv, intel_dp->output_reg);
+
+		intel_dp->DP &= ~DP_PORT_EN;
+		intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP);
+		intel_de_posting_read(dev_priv, intel_dp->output_reg);
+
+		intel_wait_for_vblank_if_active(dev_priv, PIPE_A);
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, PIPE_A, true);
+		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, true);
+	}
+
+	msleep(intel_dp->pps.panel_power_down_delay);
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		intel_wakeref_t wakeref;
+
+		with_intel_pps_lock(intel_dp, wakeref)
+			intel_dp->pps.active_pipe = INVALID_PIPE;
+	}
+}
+
+static void intel_disable_dp(struct intel_atomic_state *state,
+			     struct intel_encoder *encoder,
+			     const struct intel_crtc_state *old_crtc_state,
+			     const struct drm_connector_state *old_conn_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+	intel_dp->link_trained = false;
+
+	intel_audio_codec_disable(encoder, old_crtc_state, old_conn_state);
+
+	/*
+	 * Make sure the panel is off before trying to change the mode.
+	 * But also ensure that we have vdd while we switch off the panel.
+	 */
+	intel_pps_vdd_on(intel_dp);
+	intel_edp_backlight_off(old_conn_state);
+	intel_dp_set_power(intel_dp, DP_SET_POWER_D3);
+	intel_pps_off(intel_dp);
+}
+
+static void g4x_disable_dp(struct intel_atomic_state *state,
+			   struct intel_encoder *encoder,
+			   const struct intel_crtc_state *old_crtc_state,
+			   const struct drm_connector_state *old_conn_state)
+{
+	intel_disable_dp(state, encoder, old_crtc_state, old_conn_state);
+}
+
+static void vlv_disable_dp(struct intel_atomic_state *state,
+			   struct intel_encoder *encoder,
+			   const struct intel_crtc_state *old_crtc_state,
+			   const struct drm_connector_state *old_conn_state)
+{
+	intel_disable_dp(state, encoder, old_crtc_state, old_conn_state);
+}
+
+static void g4x_post_disable_dp(struct intel_atomic_state *state,
+				struct intel_encoder *encoder,
+				const struct intel_crtc_state *old_crtc_state,
+				const struct drm_connector_state *old_conn_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	enum port port = encoder->port;
+
+	/*
+	 * Bspec does not list a specific disable sequence for g4x DP.
+	 * Follow the ilk+ sequence (disable pipe before the port) for
+	 * g4x DP as it does not suffer from underruns like the normal
+	 * g4x modeset sequence (disable pipe after the port).
+	 */
+	intel_dp_link_down(encoder, old_crtc_state);
+
+	/* Only ilk+ has port A */
+	if (port == PORT_A)
+		ilk_edp_pll_off(intel_dp, old_crtc_state);
+}
+
+static void vlv_post_disable_dp(struct intel_atomic_state *state,
+				struct intel_encoder *encoder,
+				const struct intel_crtc_state *old_crtc_state,
+				const struct drm_connector_state *old_conn_state)
+{
+	intel_dp_link_down(encoder, old_crtc_state);
+}
+
+static void chv_post_disable_dp(struct intel_atomic_state *state,
+				struct intel_encoder *encoder,
+				const struct intel_crtc_state *old_crtc_state,
+				const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	intel_dp_link_down(encoder, old_crtc_state);
+
+	vlv_dpio_get(dev_priv);
+
+	/* Assert data lane reset */
+	chv_data_lane_soft_reset(encoder, old_crtc_state, true);
+
+	vlv_dpio_put(dev_priv);
+}
+
+static void
+cpt_set_link_train(struct intel_dp *intel_dp,
+		   const struct intel_crtc_state *crtc_state,
+		   u8 dp_train_pat)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	intel_dp->DP &= ~DP_LINK_TRAIN_MASK_CPT;
+
+	switch (intel_dp_training_pattern_symbol(dp_train_pat)) {
+	case DP_TRAINING_PATTERN_DISABLE:
+		intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT;
+		break;
+	case DP_TRAINING_PATTERN_1:
+		intel_dp->DP |= DP_LINK_TRAIN_PAT_1_CPT;
+		break;
+	case DP_TRAINING_PATTERN_2:
+		intel_dp->DP |= DP_LINK_TRAIN_PAT_2_CPT;
+		break;
+	default:
+		MISSING_CASE(intel_dp_training_pattern_symbol(dp_train_pat));
+		return;
+	}
+
+	intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+}
+
+static void
+g4x_set_link_train(struct intel_dp *intel_dp,
+		   const struct intel_crtc_state *crtc_state,
+		   u8 dp_train_pat)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	intel_dp->DP &= ~DP_LINK_TRAIN_MASK;
+
+	switch (intel_dp_training_pattern_symbol(dp_train_pat)) {
+	case DP_TRAINING_PATTERN_DISABLE:
+		intel_dp->DP |= DP_LINK_TRAIN_OFF;
+		break;
+	case DP_TRAINING_PATTERN_1:
+		intel_dp->DP |= DP_LINK_TRAIN_PAT_1;
+		break;
+	case DP_TRAINING_PATTERN_2:
+		intel_dp->DP |= DP_LINK_TRAIN_PAT_2;
+		break;
+	default:
+		MISSING_CASE(intel_dp_training_pattern_symbol(dp_train_pat));
+		return;
+	}
+
+	intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+}
+
+static void intel_dp_enable_port(struct intel_dp *intel_dp,
+				 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	/* enable with pattern 1 (as per spec) */
+
+	intel_dp_program_link_training_pattern(intel_dp, crtc_state,
+					       DP_PHY_DPRX, DP_TRAINING_PATTERN_1);
+
+	/*
+	 * Magic for VLV/CHV. We _must_ first set up the register
+	 * without actually enabling the port, and then do another
+	 * write to enable the port. Otherwise link training will
+	 * fail when the power sequencer is freshly used for this port.
+	 */
+	intel_dp->DP |= DP_PORT_EN;
+	if (crtc_state->has_audio)
+		intel_dp->DP |= DP_AUDIO_OUTPUT_ENABLE;
+
+	intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+}
+
+static void intel_enable_dp(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *pipe_config,
+			    const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	u32 dp_reg = intel_de_read(dev_priv, intel_dp->output_reg);
+	intel_wakeref_t wakeref;
+
+	if (drm_WARN_ON(&dev_priv->drm, dp_reg & DP_PORT_EN))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+			vlv_pps_init(encoder, pipe_config);
+
+		intel_dp_enable_port(intel_dp, pipe_config);
+
+		intel_pps_vdd_on_unlocked(intel_dp);
+		intel_pps_on_unlocked(intel_dp);
+		intel_pps_vdd_off_unlocked(intel_dp, true);
+	}
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		unsigned int lane_mask = 0x0;
+
+		if (IS_CHERRYVIEW(dev_priv))
+			lane_mask = intel_dp_unused_lane_mask(pipe_config->lane_count);
+
+		vlv_wait_port_ready(dev_priv, dp_to_dig_port(intel_dp),
+				    lane_mask);
+	}
+
+	intel_dp_set_power(intel_dp, DP_SET_POWER_D0);
+	intel_dp_configure_protocol_converter(intel_dp, pipe_config);
+	intel_dp_check_frl_training(intel_dp);
+	intel_dp_pcon_dsc_configure(intel_dp, pipe_config);
+	intel_dp_start_link_train(intel_dp, pipe_config);
+	intel_dp_stop_link_train(intel_dp, pipe_config);
+
+	intel_audio_codec_enable(encoder, pipe_config, conn_state);
+}
+
+static void g4x_enable_dp(struct intel_atomic_state *state,
+			  struct intel_encoder *encoder,
+			  const struct intel_crtc_state *pipe_config,
+			  const struct drm_connector_state *conn_state)
+{
+	intel_enable_dp(state, encoder, pipe_config, conn_state);
+	intel_edp_backlight_on(pipe_config, conn_state);
+}
+
+static void vlv_enable_dp(struct intel_atomic_state *state,
+			  struct intel_encoder *encoder,
+			  const struct intel_crtc_state *pipe_config,
+			  const struct drm_connector_state *conn_state)
+{
+	intel_edp_backlight_on(pipe_config, conn_state);
+}
+
+static void g4x_pre_enable_dp(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *pipe_config,
+			      const struct drm_connector_state *conn_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	enum port port = encoder->port;
+
+	intel_dp_prepare(encoder, pipe_config);
+
+	/* Only ilk+ has port A */
+	if (port == PORT_A)
+		ilk_edp_pll_on(intel_dp, pipe_config);
+}
+
+static void vlv_pre_enable_dp(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *pipe_config,
+			      const struct drm_connector_state *conn_state)
+{
+	vlv_phy_pre_encoder_enable(encoder, pipe_config);
+
+	intel_enable_dp(state, encoder, pipe_config, conn_state);
+}
+
+static void vlv_dp_pre_pll_enable(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *pipe_config,
+				  const struct drm_connector_state *conn_state)
+{
+	intel_dp_prepare(encoder, pipe_config);
+
+	vlv_phy_pre_pll_enable(encoder, pipe_config);
+}
+
+static void chv_pre_enable_dp(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *pipe_config,
+			      const struct drm_connector_state *conn_state)
+{
+	chv_phy_pre_encoder_enable(encoder, pipe_config);
+
+	intel_enable_dp(state, encoder, pipe_config, conn_state);
+
+	/* Second common lane will stay alive on its own now */
+	chv_phy_release_cl2_override(encoder);
+}
+
+static void chv_dp_pre_pll_enable(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *pipe_config,
+				  const struct drm_connector_state *conn_state)
+{
+	intel_dp_prepare(encoder, pipe_config);
+
+	chv_phy_pre_pll_enable(encoder, pipe_config);
+}
+
+static void chv_dp_post_pll_disable(struct intel_atomic_state *state,
+				    struct intel_encoder *encoder,
+				    const struct intel_crtc_state *old_crtc_state,
+				    const struct drm_connector_state *old_conn_state)
+{
+	chv_phy_post_pll_disable(encoder, old_crtc_state);
+}
+
+static u8 intel_dp_voltage_max_2(struct intel_dp *intel_dp,
+				 const struct intel_crtc_state *crtc_state)
+{
+	return DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
+}
+
+static u8 intel_dp_voltage_max_3(struct intel_dp *intel_dp,
+				 const struct intel_crtc_state *crtc_state)
+{
+	return DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
+}
+
+static u8 intel_dp_preemph_max_2(struct intel_dp *intel_dp)
+{
+	return DP_TRAIN_PRE_EMPH_LEVEL_2;
+}
+
+static u8 intel_dp_preemph_max_3(struct intel_dp *intel_dp)
+{
+	return DP_TRAIN_PRE_EMPH_LEVEL_3;
+}
+
+static void vlv_set_signal_levels(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	unsigned long demph_reg_value, preemph_reg_value,
+		uniqtranscale_reg_value;
+	u8 train_set = intel_dp->train_set[0];
+
+	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
+	case DP_TRAIN_PRE_EMPH_LEVEL_0:
+		preemph_reg_value = 0x0004000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			demph_reg_value = 0x2B405555;
+			uniqtranscale_reg_value = 0x552AB83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			demph_reg_value = 0x2B404040;
+			uniqtranscale_reg_value = 0x5548B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			demph_reg_value = 0x2B245555;
+			uniqtranscale_reg_value = 0x5560B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
+			demph_reg_value = 0x2B405555;
+			uniqtranscale_reg_value = 0x5598DA3A;
+			break;
+		default:
+			return;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPH_LEVEL_1:
+		preemph_reg_value = 0x0002000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			demph_reg_value = 0x2B404040;
+			uniqtranscale_reg_value = 0x5552B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			demph_reg_value = 0x2B404848;
+			uniqtranscale_reg_value = 0x5580B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			demph_reg_value = 0x2B404040;
+			uniqtranscale_reg_value = 0x55ADDA3A;
+			break;
+		default:
+			return;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPH_LEVEL_2:
+		preemph_reg_value = 0x0000000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			demph_reg_value = 0x2B305555;
+			uniqtranscale_reg_value = 0x5570B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			demph_reg_value = 0x2B2B4040;
+			uniqtranscale_reg_value = 0x55ADDA3A;
+			break;
+		default:
+			return;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPH_LEVEL_3:
+		preemph_reg_value = 0x0006000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			demph_reg_value = 0x1B405555;
+			uniqtranscale_reg_value = 0x55ADDA3A;
+			break;
+		default:
+			return;
+		}
+		break;
+	default:
+		return;
+	}
+
+	vlv_set_phy_signal_level(encoder, crtc_state,
+				 demph_reg_value, preemph_reg_value,
+				 uniqtranscale_reg_value, 0);
+}
+
+static void chv_set_signal_levels(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	u32 deemph_reg_value, margin_reg_value;
+	bool uniq_trans_scale = false;
+	u8 train_set = intel_dp->train_set[0];
+
+	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
+	case DP_TRAIN_PRE_EMPH_LEVEL_0:
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			deemph_reg_value = 128;
+			margin_reg_value = 52;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			deemph_reg_value = 128;
+			margin_reg_value = 77;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			deemph_reg_value = 128;
+			margin_reg_value = 102;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
+			deemph_reg_value = 128;
+			margin_reg_value = 154;
+			uniq_trans_scale = true;
+			break;
+		default:
+			return;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPH_LEVEL_1:
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			deemph_reg_value = 85;
+			margin_reg_value = 78;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			deemph_reg_value = 85;
+			margin_reg_value = 116;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+			deemph_reg_value = 85;
+			margin_reg_value = 154;
+			break;
+		default:
+			return;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPH_LEVEL_2:
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			deemph_reg_value = 64;
+			margin_reg_value = 104;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+			deemph_reg_value = 64;
+			margin_reg_value = 154;
+			break;
+		default:
+			return;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPH_LEVEL_3:
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+			deemph_reg_value = 43;
+			margin_reg_value = 154;
+			break;
+		default:
+			return;
+		}
+		break;
+	default:
+		return;
+	}
+
+	chv_set_phy_signal_level(encoder, crtc_state,
+				 deemph_reg_value, margin_reg_value,
+				 uniq_trans_scale);
+}
+
+static u32 g4x_signal_levels(u8 train_set)
+{
+	u32 signal_levels = 0;
+
+	switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0:
+	default:
+		signal_levels |= DP_VOLTAGE_0_4;
+		break;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1:
+		signal_levels |= DP_VOLTAGE_0_6;
+		break;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2:
+		signal_levels |= DP_VOLTAGE_0_8;
+		break;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_3:
+		signal_levels |= DP_VOLTAGE_1_2;
+		break;
+	}
+	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
+	case DP_TRAIN_PRE_EMPH_LEVEL_0:
+	default:
+		signal_levels |= DP_PRE_EMPHASIS_0;
+		break;
+	case DP_TRAIN_PRE_EMPH_LEVEL_1:
+		signal_levels |= DP_PRE_EMPHASIS_3_5;
+		break;
+	case DP_TRAIN_PRE_EMPH_LEVEL_2:
+		signal_levels |= DP_PRE_EMPHASIS_6;
+		break;
+	case DP_TRAIN_PRE_EMPH_LEVEL_3:
+		signal_levels |= DP_PRE_EMPHASIS_9_5;
+		break;
+	}
+	return signal_levels;
+}
+
+static void
+g4x_set_signal_levels(struct intel_encoder *encoder,
+		      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	u8 train_set = intel_dp->train_set[0];
+	u32 signal_levels;
+
+	signal_levels = g4x_signal_levels(train_set);
+
+	drm_dbg_kms(&dev_priv->drm, "Using signal levels %08x\n",
+		    signal_levels);
+
+	intel_dp->DP &= ~(DP_VOLTAGE_MASK | DP_PRE_EMPHASIS_MASK);
+	intel_dp->DP |= signal_levels;
+
+	intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+}
+
+/* SNB CPU eDP voltage swing and pre-emphasis control */
+static u32 snb_cpu_edp_signal_levels(u8 train_set)
+{
+	u8 signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
+					DP_TRAIN_PRE_EMPHASIS_MASK);
+
+	switch (signal_levels) {
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return EDP_LINK_TRAIN_400_600MV_0DB_SNB_B;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return EDP_LINK_TRAIN_400MV_3_5DB_SNB_B;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_2:
+		return EDP_LINK_TRAIN_400_600MV_6DB_SNB_B;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return EDP_LINK_TRAIN_600_800MV_3_5DB_SNB_B;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_3 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return EDP_LINK_TRAIN_800_1200MV_0DB_SNB_B;
+	default:
+		MISSING_CASE(signal_levels);
+		return EDP_LINK_TRAIN_400_600MV_0DB_SNB_B;
+	}
+}
+
+static void
+snb_cpu_edp_set_signal_levels(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	u8 train_set = intel_dp->train_set[0];
+	u32 signal_levels;
+
+	signal_levels = snb_cpu_edp_signal_levels(train_set);
+
+	drm_dbg_kms(&dev_priv->drm, "Using signal levels %08x\n",
+		    signal_levels);
+
+	intel_dp->DP &= ~EDP_LINK_TRAIN_VOL_EMP_MASK_SNB;
+	intel_dp->DP |= signal_levels;
+
+	intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+}
+
+/* IVB CPU eDP voltage swing and pre-emphasis control */
+static u32 ivb_cpu_edp_signal_levels(u8 train_set)
+{
+	u8 signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
+					DP_TRAIN_PRE_EMPHASIS_MASK);
+
+	switch (signal_levels) {
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return EDP_LINK_TRAIN_400MV_0DB_IVB;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return EDP_LINK_TRAIN_400MV_3_5DB_IVB;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2:
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_2:
+		return EDP_LINK_TRAIN_400MV_6DB_IVB;
+
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return EDP_LINK_TRAIN_600MV_0DB_IVB;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return EDP_LINK_TRAIN_600MV_3_5DB_IVB;
+
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return EDP_LINK_TRAIN_800MV_0DB_IVB;
+	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return EDP_LINK_TRAIN_800MV_3_5DB_IVB;
+
+	default:
+		MISSING_CASE(signal_levels);
+		return EDP_LINK_TRAIN_500MV_0DB_IVB;
+	}
+}
+
+static void
+ivb_cpu_edp_set_signal_levels(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	u8 train_set = intel_dp->train_set[0];
+	u32 signal_levels;
+
+	signal_levels = ivb_cpu_edp_signal_levels(train_set);
+
+	drm_dbg_kms(&dev_priv->drm, "Using signal levels %08x\n",
+		    signal_levels);
+
+	intel_dp->DP &= ~EDP_LINK_TRAIN_VOL_EMP_MASK_IVB;
+	intel_dp->DP |= signal_levels;
+
+	intel_de_write(dev_priv, intel_dp->output_reg, intel_dp->DP);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+}
+
+/*
+ * If display is now connected check links status,
+ * there has been known issues of link loss triggering
+ * long pulse.
+ *
+ * Some sinks (eg. ASUS PB287Q) seem to perform some
+ * weird HPD ping pong during modesets. So we can apparently
+ * end up with HPD going low during a modeset, and then
+ * going back up soon after. And once that happens we must
+ * retrain the link to get a picture. That's in case no
+ * userspace component reacted to intermittent HPD dip.
+ */
+static enum intel_hotplug_state
+intel_dp_hotplug(struct intel_encoder *encoder,
+		 struct intel_connector *connector)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct drm_modeset_acquire_ctx ctx;
+	enum intel_hotplug_state state;
+	int ret;
+
+	if (intel_dp->compliance.test_active &&
+	    intel_dp->compliance.test_type == DP_TEST_LINK_PHY_TEST_PATTERN) {
+		intel_dp_phy_test(encoder);
+		/* just do the PHY test and nothing else */
+		return INTEL_HOTPLUG_UNCHANGED;
+	}
+
+	state = intel_encoder_hotplug(encoder, connector);
+
+	drm_modeset_acquire_init(&ctx, 0);
+
+	for (;;) {
+		ret = intel_dp_retrain_link(encoder, &ctx);
+
+		if (ret == -EDEADLK) {
+			drm_modeset_backoff(&ctx);
+			continue;
+		}
+
+		break;
+	}
+
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+	drm_WARN(encoder->base.dev, ret,
+		 "Acquiring modeset locks failed with %i\n", ret);
+
+	/*
+	 * Keeping it consistent with intel_ddi_hotplug() and
+	 * intel_hdmi_hotplug().
+	 */
+	if (state == INTEL_HOTPLUG_UNCHANGED && !connector->hotplug_retries)
+		state = INTEL_HOTPLUG_RETRY;
+
+	return state;
+}
+
+static bool ibx_digital_port_connected(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 bit = dev_priv->display.hotplug.pch_hpd[encoder->hpd_pin];
+
+	return intel_de_read(dev_priv, SDEISR) & bit;
+}
+
+static bool g4x_digital_port_connected(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 bit;
+
+	switch (encoder->hpd_pin) {
+	case HPD_PORT_B:
+		bit = PORTB_HOTPLUG_LIVE_STATUS_G4X;
+		break;
+	case HPD_PORT_C:
+		bit = PORTC_HOTPLUG_LIVE_STATUS_G4X;
+		break;
+	case HPD_PORT_D:
+		bit = PORTD_HOTPLUG_LIVE_STATUS_G4X;
+		break;
+	default:
+		MISSING_CASE(encoder->hpd_pin);
+		return false;
+	}
+
+	return intel_de_read(dev_priv, PORT_HOTPLUG_STAT) & bit;
+}
+
+static bool gm45_digital_port_connected(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 bit;
+
+	switch (encoder->hpd_pin) {
+	case HPD_PORT_B:
+		bit = PORTB_HOTPLUG_LIVE_STATUS_GM45;
+		break;
+	case HPD_PORT_C:
+		bit = PORTC_HOTPLUG_LIVE_STATUS_GM45;
+		break;
+	case HPD_PORT_D:
+		bit = PORTD_HOTPLUG_LIVE_STATUS_GM45;
+		break;
+	default:
+		MISSING_CASE(encoder->hpd_pin);
+		return false;
+	}
+
+	return intel_de_read(dev_priv, PORT_HOTPLUG_STAT) & bit;
+}
+
+static bool ilk_digital_port_connected(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 bit = dev_priv->display.hotplug.hpd[encoder->hpd_pin];
+
+	return intel_de_read(dev_priv, DEISR) & bit;
+}
+
+static void intel_dp_encoder_destroy(struct drm_encoder *encoder)
+{
+	intel_dp_encoder_flush_work(encoder);
+
+	drm_encoder_cleanup(encoder);
+	kfree(enc_to_dig_port(to_intel_encoder(encoder)));
+}
+
+enum pipe vlv_active_pipe(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	enum pipe pipe;
+
+	if (g4x_dp_port_enabled(dev_priv, intel_dp->output_reg,
+				encoder->port, &pipe))
+		return pipe;
+
+	return INVALID_PIPE;
+}
+
+static void intel_dp_encoder_reset(struct drm_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(to_intel_encoder(encoder));
+
+	intel_dp->DP = intel_de_read(dev_priv, intel_dp->output_reg);
+
+	intel_dp->reset_link_params = true;
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		intel_wakeref_t wakeref;
+
+		with_intel_pps_lock(intel_dp, wakeref)
+			intel_dp->pps.active_pipe = vlv_active_pipe(intel_dp);
+	}
+
+	intel_pps_encoder_reset(intel_dp);
+}
+
+static const struct drm_encoder_funcs intel_dp_enc_funcs = {
+	.reset = intel_dp_encoder_reset,
+	.destroy = intel_dp_encoder_destroy,
+};
+
+bool g4x_dp_init(struct drm_i915_private *dev_priv,
+		 i915_reg_t output_reg, enum port port)
+{
+	struct intel_digital_port *dig_port;
+	struct intel_encoder *intel_encoder;
+	struct drm_encoder *encoder;
+	struct intel_connector *intel_connector;
+
+	dig_port = kzalloc(sizeof(*dig_port), GFP_KERNEL);
+	if (!dig_port)
+		return false;
+
+	intel_connector = intel_connector_alloc();
+	if (!intel_connector)
+		goto err_connector_alloc;
+
+	intel_encoder = &dig_port->base;
+	encoder = &intel_encoder->base;
+
+	mutex_init(&dig_port->hdcp_mutex);
+
+	if (drm_encoder_init(&dev_priv->drm, &intel_encoder->base,
+			     &intel_dp_enc_funcs, DRM_MODE_ENCODER_TMDS,
+			     "DP %c", port_name(port)))
+		goto err_encoder_init;
+
+	intel_encoder->hotplug = intel_dp_hotplug;
+	intel_encoder->compute_config = intel_dp_compute_config;
+	intel_encoder->get_hw_state = intel_dp_get_hw_state;
+	intel_encoder->get_config = intel_dp_get_config;
+	intel_encoder->sync_state = intel_dp_sync_state;
+	intel_encoder->initial_fastset_check = intel_dp_initial_fastset_check;
+	intel_encoder->update_pipe = intel_backlight_update;
+	intel_encoder->suspend = intel_dp_encoder_suspend;
+	intel_encoder->shutdown = intel_dp_encoder_shutdown;
+	if (IS_CHERRYVIEW(dev_priv)) {
+		intel_encoder->pre_pll_enable = chv_dp_pre_pll_enable;
+		intel_encoder->pre_enable = chv_pre_enable_dp;
+		intel_encoder->enable = vlv_enable_dp;
+		intel_encoder->disable = vlv_disable_dp;
+		intel_encoder->post_disable = chv_post_disable_dp;
+		intel_encoder->post_pll_disable = chv_dp_post_pll_disable;
+	} else if (IS_VALLEYVIEW(dev_priv)) {
+		intel_encoder->pre_pll_enable = vlv_dp_pre_pll_enable;
+		intel_encoder->pre_enable = vlv_pre_enable_dp;
+		intel_encoder->enable = vlv_enable_dp;
+		intel_encoder->disable = vlv_disable_dp;
+		intel_encoder->post_disable = vlv_post_disable_dp;
+	} else {
+		intel_encoder->pre_enable = g4x_pre_enable_dp;
+		intel_encoder->enable = g4x_enable_dp;
+		intel_encoder->disable = g4x_disable_dp;
+		intel_encoder->post_disable = g4x_post_disable_dp;
+	}
+
+	if ((IS_IVYBRIDGE(dev_priv) && port == PORT_A) ||
+	    (HAS_PCH_CPT(dev_priv) && port != PORT_A))
+		dig_port->dp.set_link_train = cpt_set_link_train;
+	else
+		dig_port->dp.set_link_train = g4x_set_link_train;
+
+	if (IS_CHERRYVIEW(dev_priv))
+		intel_encoder->set_signal_levels = chv_set_signal_levels;
+	else if (IS_VALLEYVIEW(dev_priv))
+		intel_encoder->set_signal_levels = vlv_set_signal_levels;
+	else if (IS_IVYBRIDGE(dev_priv) && port == PORT_A)
+		intel_encoder->set_signal_levels = ivb_cpu_edp_set_signal_levels;
+	else if (IS_SANDYBRIDGE(dev_priv) && port == PORT_A)
+		intel_encoder->set_signal_levels = snb_cpu_edp_set_signal_levels;
+	else
+		intel_encoder->set_signal_levels = g4x_set_signal_levels;
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv) ||
+	    (HAS_PCH_SPLIT(dev_priv) && port != PORT_A)) {
+		dig_port->dp.preemph_max = intel_dp_preemph_max_3;
+		dig_port->dp.voltage_max = intel_dp_voltage_max_3;
+	} else {
+		dig_port->dp.preemph_max = intel_dp_preemph_max_2;
+		dig_port->dp.voltage_max = intel_dp_voltage_max_2;
+	}
+
+	dig_port->dp.output_reg = output_reg;
+	dig_port->max_lanes = 4;
+
+	intel_encoder->type = INTEL_OUTPUT_DP;
+	intel_encoder->power_domain = intel_display_power_ddi_lanes_domain(dev_priv, port);
+	if (IS_CHERRYVIEW(dev_priv)) {
+		if (port == PORT_D)
+			intel_encoder->pipe_mask = BIT(PIPE_C);
+		else
+			intel_encoder->pipe_mask = BIT(PIPE_A) | BIT(PIPE_B);
+	} else {
+		intel_encoder->pipe_mask = ~0;
+	}
+	intel_encoder->cloneable = 0;
+	intel_encoder->port = port;
+	intel_encoder->hpd_pin = intel_hpd_pin_default(dev_priv, port);
+
+	dig_port->hpd_pulse = intel_dp_hpd_pulse;
+
+	if (HAS_GMCH(dev_priv)) {
+		if (IS_GM45(dev_priv))
+			dig_port->connected = gm45_digital_port_connected;
+		else
+			dig_port->connected = g4x_digital_port_connected;
+	} else {
+		if (port == PORT_A)
+			dig_port->connected = ilk_digital_port_connected;
+		else
+			dig_port->connected = ibx_digital_port_connected;
+	}
+
+	if (port != PORT_A)
+		intel_infoframe_init(dig_port);
+
+	dig_port->aux_ch = intel_bios_port_aux_ch(dev_priv, port);
+	if (!intel_dp_init_connector(dig_port, intel_connector))
+		goto err_init_connector;
+
+	return true;
+
+err_init_connector:
+	drm_encoder_cleanup(encoder);
+err_encoder_init:
+	kfree(intel_connector);
+err_connector_alloc:
+	kfree(dig_port);
+	return false;
+}
diff --git a/drivers/gpu/drm/i915/display/g4x_dp.h b/drivers/gpu/drm/i915/display/g4x_dp.h
new file mode 100644
index 000000000..e1f50263a
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/g4x_dp.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _G4X_DP_H_
+#define _G4X_DP_H_
+
+#include <linux/types.h>
+
+#include "i915_reg.h"
+
+enum pipe;
+enum port;
+struct drm_i915_private;
+struct intel_crtc_state;
+struct intel_dp;
+struct intel_encoder;
+
+const struct dpll *vlv_get_dpll(struct drm_i915_private *i915);
+enum pipe vlv_active_pipe(struct intel_dp *intel_dp);
+void g4x_dp_set_clock(struct intel_encoder *encoder,
+		      struct intel_crtc_state *pipe_config);
+bool g4x_dp_port_enabled(struct drm_i915_private *dev_priv,
+			 i915_reg_t dp_reg, enum port port,
+			 enum pipe *pipe);
+bool g4x_dp_init(struct drm_i915_private *dev_priv,
+		 i915_reg_t output_reg, enum port port);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/g4x_hdmi.c b/drivers/gpu/drm/i915/display/g4x_hdmi.c
new file mode 100644
index 000000000..2b73f5ff0
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/g4x_hdmi.c
@@ -0,0 +1,606 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ *
+ * HDMI support for G4x,ILK,SNB,IVB,VLV,CHV (HSW+ handled by the DDI code).
+ */
+
+#include "g4x_hdmi.h"
+#include "intel_audio.h"
+#include "intel_connector.h"
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display_power.h"
+#include "intel_display_types.h"
+#include "intel_dpio_phy.h"
+#include "intel_fifo_underrun.h"
+#include "intel_hdmi.h"
+#include "intel_hotplug.h"
+#include "intel_sdvo.h"
+#include "vlv_sideband.h"
+
+static void intel_hdmi_prepare(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	u32 hdmi_val;
+
+	intel_dp_dual_mode_set_tmds_output(intel_hdmi, true);
+
+	hdmi_val = SDVO_ENCODING_HDMI;
+	if (!HAS_PCH_SPLIT(dev_priv) && crtc_state->limited_color_range)
+		hdmi_val |= HDMI_COLOR_RANGE_16_235;
+	if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
+		hdmi_val |= SDVO_VSYNC_ACTIVE_HIGH;
+	if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
+		hdmi_val |= SDVO_HSYNC_ACTIVE_HIGH;
+
+	if (crtc_state->pipe_bpp > 24)
+		hdmi_val |= HDMI_COLOR_FORMAT_12bpc;
+	else
+		hdmi_val |= SDVO_COLOR_FORMAT_8bpc;
+
+	if (crtc_state->has_hdmi_sink)
+		hdmi_val |= HDMI_MODE_SELECT_HDMI;
+
+	if (HAS_PCH_CPT(dev_priv))
+		hdmi_val |= SDVO_PIPE_SEL_CPT(crtc->pipe);
+	else if (IS_CHERRYVIEW(dev_priv))
+		hdmi_val |= SDVO_PIPE_SEL_CHV(crtc->pipe);
+	else
+		hdmi_val |= SDVO_PIPE_SEL(crtc->pipe);
+
+	intel_de_write(dev_priv, intel_hdmi->hdmi_reg, hdmi_val);
+	intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+}
+
+static bool intel_hdmi_get_hw_state(struct intel_encoder *encoder,
+				    enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	intel_wakeref_t wakeref;
+	bool ret;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     encoder->power_domain);
+	if (!wakeref)
+		return false;
+
+	ret = intel_sdvo_port_enabled(dev_priv, intel_hdmi->hdmi_reg, pipe);
+
+	intel_display_power_put(dev_priv, encoder->power_domain, wakeref);
+
+	return ret;
+}
+
+static void intel_hdmi_get_config(struct intel_encoder *encoder,
+				  struct intel_crtc_state *pipe_config)
+{
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 tmp, flags = 0;
+	int dotclock;
+
+	pipe_config->output_types |= BIT(INTEL_OUTPUT_HDMI);
+
+	tmp = intel_de_read(dev_priv, intel_hdmi->hdmi_reg);
+
+	if (tmp & SDVO_HSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+
+	if (tmp & SDVO_VSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	if (tmp & HDMI_MODE_SELECT_HDMI)
+		pipe_config->has_hdmi_sink = true;
+
+	pipe_config->infoframes.enable |=
+		intel_hdmi_infoframes_enabled(encoder, pipe_config);
+
+	if (pipe_config->infoframes.enable)
+		pipe_config->has_infoframe = true;
+
+	if (tmp & HDMI_AUDIO_ENABLE)
+		pipe_config->has_audio = true;
+
+	if (!HAS_PCH_SPLIT(dev_priv) &&
+	    tmp & HDMI_COLOR_RANGE_16_235)
+		pipe_config->limited_color_range = true;
+
+	pipe_config->hw.adjusted_mode.flags |= flags;
+
+	if ((tmp & SDVO_COLOR_FORMAT_MASK) == HDMI_COLOR_FORMAT_12bpc)
+		dotclock = DIV_ROUND_CLOSEST(pipe_config->port_clock * 2, 3);
+	else
+		dotclock = pipe_config->port_clock;
+
+	if (pipe_config->pixel_multiplier)
+		dotclock /= pipe_config->pixel_multiplier;
+
+	pipe_config->hw.adjusted_mode.crtc_clock = dotclock;
+
+	pipe_config->lane_count = 4;
+
+	intel_hdmi_read_gcp_infoframe(encoder, pipe_config);
+
+	intel_read_infoframe(encoder, pipe_config,
+			     HDMI_INFOFRAME_TYPE_AVI,
+			     &pipe_config->infoframes.avi);
+	intel_read_infoframe(encoder, pipe_config,
+			     HDMI_INFOFRAME_TYPE_SPD,
+			     &pipe_config->infoframes.spd);
+	intel_read_infoframe(encoder, pipe_config,
+			     HDMI_INFOFRAME_TYPE_VENDOR,
+			     &pipe_config->infoframes.hdmi);
+}
+
+static void g4x_enable_hdmi(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *pipe_config,
+			    const struct drm_connector_state *conn_state)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	u32 temp;
+
+	temp = intel_de_read(dev_priv, intel_hdmi->hdmi_reg);
+
+	temp |= SDVO_ENABLE;
+	if (pipe_config->has_audio)
+		temp |= HDMI_AUDIO_ENABLE;
+
+	intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+	intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+
+	drm_WARN_ON(&dev_priv->drm, pipe_config->has_audio &&
+		    !pipe_config->has_hdmi_sink);
+	intel_audio_codec_enable(encoder, pipe_config, conn_state);
+}
+
+static void ibx_enable_hdmi(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *pipe_config,
+			    const struct drm_connector_state *conn_state)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	u32 temp;
+
+	temp = intel_de_read(dev_priv, intel_hdmi->hdmi_reg);
+
+	temp |= SDVO_ENABLE;
+	if (pipe_config->has_audio)
+		temp |= HDMI_AUDIO_ENABLE;
+
+	/*
+	 * HW workaround, need to write this twice for issue
+	 * that may result in first write getting masked.
+	 */
+	intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+	intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+	intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+	intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+
+	/*
+	 * HW workaround, need to toggle enable bit off and on
+	 * for 12bpc with pixel repeat.
+	 *
+	 * FIXME: BSpec says this should be done at the end of
+	 * the modeset sequence, so not sure if this isn't too soon.
+	 */
+	if (pipe_config->pipe_bpp > 24 &&
+	    pipe_config->pixel_multiplier > 1) {
+		intel_de_write(dev_priv, intel_hdmi->hdmi_reg,
+			       temp & ~SDVO_ENABLE);
+		intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+
+		/*
+		 * HW workaround, need to write this twice for issue
+		 * that may result in first write getting masked.
+		 */
+		intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+		intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+		intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+		intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+	}
+
+	drm_WARN_ON(&dev_priv->drm, pipe_config->has_audio &&
+		    !pipe_config->has_hdmi_sink);
+	intel_audio_codec_enable(encoder, pipe_config, conn_state);
+}
+
+static void cpt_enable_hdmi(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *pipe_config,
+			    const struct drm_connector_state *conn_state)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	enum pipe pipe = crtc->pipe;
+	u32 temp;
+
+	temp = intel_de_read(dev_priv, intel_hdmi->hdmi_reg);
+
+	temp |= SDVO_ENABLE;
+	if (pipe_config->has_audio)
+		temp |= HDMI_AUDIO_ENABLE;
+
+	/*
+	 * WaEnableHDMI8bpcBefore12bpc:snb,ivb
+	 *
+	 * The procedure for 12bpc is as follows:
+	 * 1. disable HDMI clock gating
+	 * 2. enable HDMI with 8bpc
+	 * 3. enable HDMI with 12bpc
+	 * 4. enable HDMI clock gating
+	 */
+
+	if (pipe_config->pipe_bpp > 24) {
+		intel_de_write(dev_priv, TRANS_CHICKEN1(pipe),
+			       intel_de_read(dev_priv, TRANS_CHICKEN1(pipe)) | TRANS_CHICKEN1_HDMIUNIT_GC_DISABLE);
+
+		temp &= ~SDVO_COLOR_FORMAT_MASK;
+		temp |= SDVO_COLOR_FORMAT_8bpc;
+	}
+
+	intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+	intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+
+	if (pipe_config->pipe_bpp > 24) {
+		temp &= ~SDVO_COLOR_FORMAT_MASK;
+		temp |= HDMI_COLOR_FORMAT_12bpc;
+
+		intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+		intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+
+		intel_de_write(dev_priv, TRANS_CHICKEN1(pipe),
+			       intel_de_read(dev_priv, TRANS_CHICKEN1(pipe)) & ~TRANS_CHICKEN1_HDMIUNIT_GC_DISABLE);
+	}
+
+	drm_WARN_ON(&dev_priv->drm, pipe_config->has_audio &&
+		    !pipe_config->has_hdmi_sink);
+	intel_audio_codec_enable(encoder, pipe_config, conn_state);
+}
+
+static void vlv_enable_hdmi(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *pipe_config,
+			    const struct drm_connector_state *conn_state)
+{
+}
+
+static void intel_disable_hdmi(struct intel_atomic_state *state,
+			       struct intel_encoder *encoder,
+			       const struct intel_crtc_state *old_crtc_state,
+			       const struct drm_connector_state *old_conn_state)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	struct intel_digital_port *dig_port =
+		hdmi_to_dig_port(intel_hdmi);
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	u32 temp;
+
+	temp = intel_de_read(dev_priv, intel_hdmi->hdmi_reg);
+
+	temp &= ~(SDVO_ENABLE | HDMI_AUDIO_ENABLE);
+	intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+	intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+
+	/*
+	 * HW workaround for IBX, we need to move the port
+	 * to transcoder A after disabling it to allow the
+	 * matching DP port to be enabled on transcoder A.
+	 */
+	if (HAS_PCH_IBX(dev_priv) && crtc->pipe == PIPE_B) {
+		/*
+		 * We get CPU/PCH FIFO underruns on the other pipe when
+		 * doing the workaround. Sweep them under the rug.
+		 */
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, PIPE_A, false);
+		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, false);
+
+		temp &= ~SDVO_PIPE_SEL_MASK;
+		temp |= SDVO_ENABLE | SDVO_PIPE_SEL(PIPE_A);
+		/*
+		 * HW workaround, need to write this twice for issue
+		 * that may result in first write getting masked.
+		 */
+		intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+		intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+		intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+		intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+
+		temp &= ~SDVO_ENABLE;
+		intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp);
+		intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg);
+
+		intel_wait_for_vblank_if_active(dev_priv, PIPE_A);
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, PIPE_A, true);
+		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, true);
+	}
+
+	dig_port->set_infoframes(encoder,
+				       false,
+				       old_crtc_state, old_conn_state);
+
+	intel_dp_dual_mode_set_tmds_output(intel_hdmi, false);
+}
+
+static void g4x_disable_hdmi(struct intel_atomic_state *state,
+			     struct intel_encoder *encoder,
+			     const struct intel_crtc_state *old_crtc_state,
+			     const struct drm_connector_state *old_conn_state)
+{
+	intel_audio_codec_disable(encoder, old_crtc_state, old_conn_state);
+
+	intel_disable_hdmi(state, encoder, old_crtc_state, old_conn_state);
+}
+
+static void pch_disable_hdmi(struct intel_atomic_state *state,
+			     struct intel_encoder *encoder,
+			     const struct intel_crtc_state *old_crtc_state,
+			     const struct drm_connector_state *old_conn_state)
+{
+	intel_audio_codec_disable(encoder, old_crtc_state, old_conn_state);
+}
+
+static void pch_post_disable_hdmi(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *old_crtc_state,
+				  const struct drm_connector_state *old_conn_state)
+{
+	intel_disable_hdmi(state, encoder, old_crtc_state, old_conn_state);
+}
+
+static void intel_hdmi_pre_enable(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *pipe_config,
+				  const struct drm_connector_state *conn_state)
+{
+	struct intel_digital_port *dig_port =
+		enc_to_dig_port(encoder);
+
+	intel_hdmi_prepare(encoder, pipe_config);
+
+	dig_port->set_infoframes(encoder,
+				       pipe_config->has_infoframe,
+				       pipe_config, conn_state);
+}
+
+static void vlv_hdmi_pre_enable(struct intel_atomic_state *state,
+				struct intel_encoder *encoder,
+				const struct intel_crtc_state *pipe_config,
+				const struct drm_connector_state *conn_state)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	vlv_phy_pre_encoder_enable(encoder, pipe_config);
+
+	/* HDMI 1.0V-2dB */
+	vlv_set_phy_signal_level(encoder, pipe_config,
+				 0x2b245f5f, 0x00002000,
+				 0x5578b83a, 0x2b247878);
+
+	dig_port->set_infoframes(encoder,
+			      pipe_config->has_infoframe,
+			      pipe_config, conn_state);
+
+	g4x_enable_hdmi(state, encoder, pipe_config, conn_state);
+
+	vlv_wait_port_ready(dev_priv, dig_port, 0x0);
+}
+
+static void vlv_hdmi_pre_pll_enable(struct intel_atomic_state *state,
+				    struct intel_encoder *encoder,
+				    const struct intel_crtc_state *pipe_config,
+				    const struct drm_connector_state *conn_state)
+{
+	intel_hdmi_prepare(encoder, pipe_config);
+
+	vlv_phy_pre_pll_enable(encoder, pipe_config);
+}
+
+static void chv_hdmi_pre_pll_enable(struct intel_atomic_state *state,
+				    struct intel_encoder *encoder,
+				    const struct intel_crtc_state *pipe_config,
+				    const struct drm_connector_state *conn_state)
+{
+	intel_hdmi_prepare(encoder, pipe_config);
+
+	chv_phy_pre_pll_enable(encoder, pipe_config);
+}
+
+static void chv_hdmi_post_pll_disable(struct intel_atomic_state *state,
+				      struct intel_encoder *encoder,
+				      const struct intel_crtc_state *old_crtc_state,
+				      const struct drm_connector_state *old_conn_state)
+{
+	chv_phy_post_pll_disable(encoder, old_crtc_state);
+}
+
+static void vlv_hdmi_post_disable(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *old_crtc_state,
+				  const struct drm_connector_state *old_conn_state)
+{
+	/* Reset lanes to avoid HDMI flicker (VLV w/a) */
+	vlv_phy_reset_lanes(encoder, old_crtc_state);
+}
+
+static void chv_hdmi_post_disable(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *old_crtc_state,
+				  const struct drm_connector_state *old_conn_state)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	vlv_dpio_get(dev_priv);
+
+	/* Assert data lane reset */
+	chv_data_lane_soft_reset(encoder, old_crtc_state, true);
+
+	vlv_dpio_put(dev_priv);
+}
+
+static void chv_hdmi_pre_enable(struct intel_atomic_state *state,
+				struct intel_encoder *encoder,
+				const struct intel_crtc_state *pipe_config,
+				const struct drm_connector_state *conn_state)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	chv_phy_pre_encoder_enable(encoder, pipe_config);
+
+	/* FIXME: Program the support xxx V-dB */
+	/* Use 800mV-0dB */
+	chv_set_phy_signal_level(encoder, pipe_config, 128, 102, false);
+
+	dig_port->set_infoframes(encoder,
+			      pipe_config->has_infoframe,
+			      pipe_config, conn_state);
+
+	g4x_enable_hdmi(state, encoder, pipe_config, conn_state);
+
+	vlv_wait_port_ready(dev_priv, dig_port, 0x0);
+
+	/* Second common lane will stay alive on its own now */
+	chv_phy_release_cl2_override(encoder);
+}
+
+static const struct drm_encoder_funcs intel_hdmi_enc_funcs = {
+	.destroy = intel_encoder_destroy,
+};
+
+static enum intel_hotplug_state
+intel_hdmi_hotplug(struct intel_encoder *encoder,
+		   struct intel_connector *connector)
+{
+	enum intel_hotplug_state state;
+
+	state = intel_encoder_hotplug(encoder, connector);
+
+	/*
+	 * On many platforms the HDMI live state signal is known to be
+	 * unreliable, so we can't use it to detect if a sink is connected or
+	 * not. Instead we detect if it's connected based on whether we can
+	 * read the EDID or not. That in turn has a problem during disconnect,
+	 * since the HPD interrupt may be raised before the DDC lines get
+	 * disconnected (due to how the required length of DDC vs. HPD
+	 * connector pins are specified) and so we'll still be able to get a
+	 * valid EDID. To solve this schedule another detection cycle if this
+	 * time around we didn't detect any change in the sink's connection
+	 * status.
+	 */
+	if (state == INTEL_HOTPLUG_UNCHANGED && !connector->hotplug_retries)
+		state = INTEL_HOTPLUG_RETRY;
+
+	return state;
+}
+
+void g4x_hdmi_init(struct drm_i915_private *dev_priv,
+		   i915_reg_t hdmi_reg, enum port port)
+{
+	struct intel_digital_port *dig_port;
+	struct intel_encoder *intel_encoder;
+	struct intel_connector *intel_connector;
+
+	dig_port = kzalloc(sizeof(*dig_port), GFP_KERNEL);
+	if (!dig_port)
+		return;
+
+	intel_connector = intel_connector_alloc();
+	if (!intel_connector) {
+		kfree(dig_port);
+		return;
+	}
+
+	intel_encoder = &dig_port->base;
+
+	mutex_init(&dig_port->hdcp_mutex);
+
+	drm_encoder_init(&dev_priv->drm, &intel_encoder->base,
+			 &intel_hdmi_enc_funcs, DRM_MODE_ENCODER_TMDS,
+			 "HDMI %c", port_name(port));
+
+	intel_encoder->hotplug = intel_hdmi_hotplug;
+	intel_encoder->compute_config = intel_hdmi_compute_config;
+	if (HAS_PCH_SPLIT(dev_priv)) {
+		intel_encoder->disable = pch_disable_hdmi;
+		intel_encoder->post_disable = pch_post_disable_hdmi;
+	} else {
+		intel_encoder->disable = g4x_disable_hdmi;
+	}
+	intel_encoder->get_hw_state = intel_hdmi_get_hw_state;
+	intel_encoder->get_config = intel_hdmi_get_config;
+	if (IS_CHERRYVIEW(dev_priv)) {
+		intel_encoder->pre_pll_enable = chv_hdmi_pre_pll_enable;
+		intel_encoder->pre_enable = chv_hdmi_pre_enable;
+		intel_encoder->enable = vlv_enable_hdmi;
+		intel_encoder->post_disable = chv_hdmi_post_disable;
+		intel_encoder->post_pll_disable = chv_hdmi_post_pll_disable;
+	} else if (IS_VALLEYVIEW(dev_priv)) {
+		intel_encoder->pre_pll_enable = vlv_hdmi_pre_pll_enable;
+		intel_encoder->pre_enable = vlv_hdmi_pre_enable;
+		intel_encoder->enable = vlv_enable_hdmi;
+		intel_encoder->post_disable = vlv_hdmi_post_disable;
+	} else {
+		intel_encoder->pre_enable = intel_hdmi_pre_enable;
+		if (HAS_PCH_CPT(dev_priv))
+			intel_encoder->enable = cpt_enable_hdmi;
+		else if (HAS_PCH_IBX(dev_priv))
+			intel_encoder->enable = ibx_enable_hdmi;
+		else
+			intel_encoder->enable = g4x_enable_hdmi;
+	}
+	intel_encoder->shutdown = intel_hdmi_encoder_shutdown;
+
+	intel_encoder->type = INTEL_OUTPUT_HDMI;
+	intel_encoder->power_domain = intel_display_power_ddi_lanes_domain(dev_priv, port);
+	intel_encoder->port = port;
+	if (IS_CHERRYVIEW(dev_priv)) {
+		if (port == PORT_D)
+			intel_encoder->pipe_mask = BIT(PIPE_C);
+		else
+			intel_encoder->pipe_mask = BIT(PIPE_A) | BIT(PIPE_B);
+	} else {
+		intel_encoder->pipe_mask = ~0;
+	}
+	intel_encoder->cloneable = 1 << INTEL_OUTPUT_ANALOG;
+	intel_encoder->hpd_pin = intel_hpd_pin_default(dev_priv, port);
+	/*
+	 * BSpec is unclear about HDMI+HDMI cloning on g4x, but it seems
+	 * to work on real hardware. And since g4x can send infoframes to
+	 * only one port anyway, nothing is lost by allowing it.
+	 */
+	if (IS_G4X(dev_priv))
+		intel_encoder->cloneable |= 1 << INTEL_OUTPUT_HDMI;
+
+	dig_port->hdmi.hdmi_reg = hdmi_reg;
+	dig_port->dp.output_reg = INVALID_MMIO_REG;
+	dig_port->max_lanes = 4;
+
+	intel_infoframe_init(dig_port);
+
+	dig_port->aux_ch = intel_bios_port_aux_ch(dev_priv, port);
+	intel_hdmi_init_connector(dig_port, intel_connector);
+}
diff --git a/drivers/gpu/drm/i915/display/g4x_hdmi.h b/drivers/gpu/drm/i915/display/g4x_hdmi.h
new file mode 100644
index 000000000..db9a93bc9
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/g4x_hdmi.h
@@ -0,0 +1,19 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _G4X_HDMI_H_
+#define _G4X_HDMI_H_
+
+#include <linux/types.h>
+
+#include "i915_reg_defs.h"
+
+enum port;
+struct drm_i915_private;
+
+void g4x_hdmi_init(struct drm_i915_private *dev_priv,
+		   i915_reg_t hdmi_reg, enum port port);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/hsw_ips.c b/drivers/gpu/drm/i915/display/hsw_ips.c
new file mode 100644
index 000000000..a5be4af79
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/hsw_ips.c
@@ -0,0 +1,271 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include "hsw_ips.h"
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_pcode.h"
+
+static void hsw_ips_enable(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	if (!crtc_state->ips_enabled)
+		return;
+
+	/*
+	 * We can only enable IPS after we enable a plane and wait for a vblank
+	 * This function is called from post_plane_update, which is run after
+	 * a vblank wait.
+	 */
+	drm_WARN_ON(&i915->drm,
+		    !(crtc_state->active_planes & ~BIT(PLANE_CURSOR)));
+
+	if (IS_BROADWELL(i915)) {
+		drm_WARN_ON(&i915->drm,
+			    snb_pcode_write(&i915->uncore, DISPLAY_IPS_CONTROL,
+					    IPS_ENABLE | IPS_PCODE_CONTROL));
+		/*
+		 * Quoting Art Runyan: "its not safe to expect any particular
+		 * value in IPS_CTL bit 31 after enabling IPS through the
+		 * mailbox." Moreover, the mailbox may return a bogus state,
+		 * so we need to just enable it and continue on.
+		 */
+	} else {
+		intel_de_write(i915, IPS_CTL, IPS_ENABLE);
+		/*
+		 * The bit only becomes 1 in the next vblank, so this wait here
+		 * is essentially intel_wait_for_vblank. If we don't have this
+		 * and don't wait for vblanks until the end of crtc_enable, then
+		 * the HW state readout code will complain that the expected
+		 * IPS_CTL value is not the one we read.
+		 */
+		if (intel_de_wait_for_set(i915, IPS_CTL, IPS_ENABLE, 50))
+			drm_err(&i915->drm,
+				"Timed out waiting for IPS enable\n");
+	}
+}
+
+bool hsw_ips_disable(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	bool need_vblank_wait = false;
+
+	if (!crtc_state->ips_enabled)
+		return need_vblank_wait;
+
+	if (IS_BROADWELL(i915)) {
+		drm_WARN_ON(&i915->drm,
+			    snb_pcode_write(&i915->uncore, DISPLAY_IPS_CONTROL, 0));
+		/*
+		 * Wait for PCODE to finish disabling IPS. The BSpec specified
+		 * 42ms timeout value leads to occasional timeouts so use 100ms
+		 * instead.
+		 */
+		if (intel_de_wait_for_clear(i915, IPS_CTL, IPS_ENABLE, 100))
+			drm_err(&i915->drm,
+				"Timed out waiting for IPS disable\n");
+	} else {
+		intel_de_write(i915, IPS_CTL, 0);
+		intel_de_posting_read(i915, IPS_CTL);
+	}
+
+	/* We need to wait for a vblank before we can disable the plane. */
+	need_vblank_wait = true;
+
+	return need_vblank_wait;
+}
+
+static bool hsw_ips_need_disable(struct intel_atomic_state *state,
+				 struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (!old_crtc_state->ips_enabled)
+		return false;
+
+	if (intel_crtc_needs_modeset(new_crtc_state))
+		return true;
+
+	/*
+	 * Workaround : Do not read or write the pipe palette/gamma data while
+	 * GAMMA_MODE is configured for split gamma and IPS_CTL has IPS enabled.
+	 *
+	 * Disable IPS before we program the LUT.
+	 */
+	if (IS_HASWELL(i915) &&
+	    (new_crtc_state->uapi.color_mgmt_changed ||
+	     new_crtc_state->update_pipe) &&
+	    new_crtc_state->gamma_mode == GAMMA_MODE_MODE_SPLIT)
+		return true;
+
+	return !new_crtc_state->ips_enabled;
+}
+
+bool hsw_ips_pre_update(struct intel_atomic_state *state,
+			struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+
+	if (!hsw_ips_need_disable(state, crtc))
+		return false;
+
+	return hsw_ips_disable(old_crtc_state);
+}
+
+static bool hsw_ips_need_enable(struct intel_atomic_state *state,
+				struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (!new_crtc_state->ips_enabled)
+		return false;
+
+	if (intel_crtc_needs_modeset(new_crtc_state))
+		return true;
+
+	/*
+	 * Workaround : Do not read or write the pipe palette/gamma data while
+	 * GAMMA_MODE is configured for split gamma and IPS_CTL has IPS enabled.
+	 *
+	 * Re-enable IPS after the LUT has been programmed.
+	 */
+	if (IS_HASWELL(i915) &&
+	    (new_crtc_state->uapi.color_mgmt_changed ||
+	     new_crtc_state->update_pipe) &&
+	    new_crtc_state->gamma_mode == GAMMA_MODE_MODE_SPLIT)
+		return true;
+
+	/*
+	 * We can't read out IPS on broadwell, assume the worst and
+	 * forcibly enable IPS on the first fastset.
+	 */
+	if (new_crtc_state->update_pipe && old_crtc_state->inherited)
+		return true;
+
+	return !old_crtc_state->ips_enabled;
+}
+
+void hsw_ips_post_update(struct intel_atomic_state *state,
+			 struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (!hsw_ips_need_enable(state, crtc))
+		return;
+
+	hsw_ips_enable(new_crtc_state);
+}
+
+/* IPS only exists on ULT machines and is tied to pipe A. */
+bool hsw_crtc_supports_ips(struct intel_crtc *crtc)
+{
+	return HAS_IPS(to_i915(crtc->base.dev)) && crtc->pipe == PIPE_A;
+}
+
+bool hsw_crtc_state_ips_capable(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	/* IPS only exists on ULT machines and is tied to pipe A. */
+	if (!hsw_crtc_supports_ips(crtc))
+		return false;
+
+	if (!i915->params.enable_ips)
+		return false;
+
+	if (crtc_state->pipe_bpp > 24)
+		return false;
+
+	/*
+	 * We compare against max which means we must take
+	 * the increased cdclk requirement into account when
+	 * calculating the new cdclk.
+	 *
+	 * Should measure whether using a lower cdclk w/o IPS
+	 */
+	if (IS_BROADWELL(i915) &&
+	    crtc_state->pixel_rate > i915->display.cdclk.max_cdclk_freq * 95 / 100)
+		return false;
+
+	return true;
+}
+
+int hsw_ips_compute_config(struct intel_atomic_state *state,
+			   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	crtc_state->ips_enabled = false;
+
+	if (!hsw_crtc_state_ips_capable(crtc_state))
+		return 0;
+
+	/*
+	 * When IPS gets enabled, the pipe CRC changes. Since IPS gets
+	 * enabled and disabled dynamically based on package C states,
+	 * user space can't make reliable use of the CRCs, so let's just
+	 * completely disable it.
+	 */
+	if (crtc_state->crc_enabled)
+		return 0;
+
+	/* IPS should be fine as long as at least one plane is enabled. */
+	if (!(crtc_state->active_planes & ~BIT(PLANE_CURSOR)))
+		return 0;
+
+	if (IS_BROADWELL(i915)) {
+		const struct intel_cdclk_state *cdclk_state;
+
+		cdclk_state = intel_atomic_get_cdclk_state(state);
+		if (IS_ERR(cdclk_state))
+			return PTR_ERR(cdclk_state);
+
+		/* pixel rate mustn't exceed 95% of cdclk with IPS on BDW */
+		if (crtc_state->pixel_rate > cdclk_state->logical.cdclk * 95 / 100)
+			return 0;
+	}
+
+	crtc_state->ips_enabled = true;
+
+	return 0;
+}
+
+void hsw_ips_get_config(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	if (!hsw_crtc_supports_ips(crtc))
+		return;
+
+	if (IS_HASWELL(i915)) {
+		crtc_state->ips_enabled = intel_de_read(i915, IPS_CTL) & IPS_ENABLE;
+	} else {
+		/*
+		 * We cannot readout IPS state on broadwell, set to
+		 * true so we can set it to a defined state on first
+		 * commit.
+		 */
+		crtc_state->ips_enabled = true;
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/hsw_ips.h b/drivers/gpu/drm/i915/display/hsw_ips.h
new file mode 100644
index 000000000..4564dee49
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/hsw_ips.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __HSW_IPS_H__
+#define __HSW_IPS_H__
+
+#include <linux/types.h>
+
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+
+bool hsw_ips_disable(const struct intel_crtc_state *crtc_state);
+bool hsw_ips_pre_update(struct intel_atomic_state *state,
+			struct intel_crtc *crtc);
+void hsw_ips_post_update(struct intel_atomic_state *state,
+			 struct intel_crtc *crtc);
+bool hsw_crtc_supports_ips(struct intel_crtc *crtc);
+bool hsw_crtc_state_ips_capable(const struct intel_crtc_state *crtc_state);
+int hsw_ips_compute_config(struct intel_atomic_state *state,
+			   struct intel_crtc *crtc);
+void hsw_ips_get_config(struct intel_crtc_state *crtc_state);
+
+#endif /* __HSW_IPS_H__ */
diff --git a/drivers/gpu/drm/i915/display/i9xx_plane.c b/drivers/gpu/drm/i915/display/i9xx_plane.c
new file mode 100644
index 000000000..5afbe3e98
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/i9xx_plane.c
@@ -0,0 +1,1056 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+#include <linux/kernel.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_blend.h>
+#include <drm/drm_fourcc.h>
+
+#include "intel_atomic.h"
+#include "intel_atomic_plane.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_fb.h"
+#include "intel_fbc.h"
+#include "intel_sprite.h"
+#include "i9xx_plane.h"
+
+/* Primary plane formats for gen <= 3 */
+static const u32 i8xx_primary_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+};
+
+/* Primary plane formats for ivb (no fp16 due to hw issue) */
+static const u32 ivb_primary_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+};
+
+/* Primary plane formats for gen >= 4, except ivb */
+static const u32 i965_primary_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_XBGR16161616F,
+};
+
+/* Primary plane formats for vlv/chv */
+static const u32 vlv_primary_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_ARGB2101010,
+	DRM_FORMAT_ABGR2101010,
+	DRM_FORMAT_XBGR16161616F,
+};
+
+static bool i8xx_plane_format_mod_supported(struct drm_plane *_plane,
+					    u32 format, u64 modifier)
+{
+	if (!intel_fb_plane_supports_modifier(to_intel_plane(_plane), modifier))
+		return false;
+
+	switch (format) {
+	case DRM_FORMAT_C8:
+	case DRM_FORMAT_RGB565:
+	case DRM_FORMAT_XRGB1555:
+	case DRM_FORMAT_XRGB8888:
+		return modifier == DRM_FORMAT_MOD_LINEAR ||
+			modifier == I915_FORMAT_MOD_X_TILED;
+	default:
+		return false;
+	}
+}
+
+static bool i965_plane_format_mod_supported(struct drm_plane *_plane,
+					    u32 format, u64 modifier)
+{
+	if (!intel_fb_plane_supports_modifier(to_intel_plane(_plane), modifier))
+		return false;
+
+	switch (format) {
+	case DRM_FORMAT_C8:
+	case DRM_FORMAT_RGB565:
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_ARGB8888:
+	case DRM_FORMAT_ABGR8888:
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_XBGR2101010:
+	case DRM_FORMAT_ARGB2101010:
+	case DRM_FORMAT_ABGR2101010:
+	case DRM_FORMAT_XBGR16161616F:
+		return modifier == DRM_FORMAT_MOD_LINEAR ||
+			modifier == I915_FORMAT_MOD_X_TILED;
+	default:
+		return false;
+	}
+}
+
+static bool i9xx_plane_has_fbc(struct drm_i915_private *dev_priv,
+			       enum i9xx_plane_id i9xx_plane)
+{
+	if (!HAS_FBC(dev_priv))
+		return false;
+
+	if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		return i9xx_plane == PLANE_A; /* tied to pipe A */
+	else if (IS_IVYBRIDGE(dev_priv))
+		return i9xx_plane == PLANE_A || i9xx_plane == PLANE_B ||
+			i9xx_plane == PLANE_C;
+	else if (DISPLAY_VER(dev_priv) >= 4)
+		return i9xx_plane == PLANE_A || i9xx_plane == PLANE_B;
+	else
+		return i9xx_plane == PLANE_A;
+}
+
+static struct intel_fbc *i9xx_plane_fbc(struct drm_i915_private *dev_priv,
+					enum i9xx_plane_id i9xx_plane)
+{
+	if (i9xx_plane_has_fbc(dev_priv, i9xx_plane))
+		return dev_priv->display.fbc[INTEL_FBC_A];
+	else
+		return NULL;
+}
+
+static bool i9xx_plane_has_windowing(struct intel_plane *plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+
+	if (IS_CHERRYVIEW(dev_priv))
+		return i9xx_plane == PLANE_B;
+	else if (DISPLAY_VER(dev_priv) >= 5 || IS_G4X(dev_priv))
+		return false;
+	else if (DISPLAY_VER(dev_priv) == 4)
+		return i9xx_plane == PLANE_C;
+	else
+		return i9xx_plane == PLANE_B ||
+			i9xx_plane == PLANE_C;
+}
+
+static u32 i9xx_plane_ctl(const struct intel_crtc_state *crtc_state,
+			  const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+	u32 dspcntr;
+
+	dspcntr = DISP_ENABLE;
+
+	if (IS_G4X(dev_priv) || IS_IRONLAKE(dev_priv) ||
+	    IS_SANDYBRIDGE(dev_priv) || IS_IVYBRIDGE(dev_priv))
+		dspcntr |= DISP_TRICKLE_FEED_DISABLE;
+
+	switch (fb->format->format) {
+	case DRM_FORMAT_C8:
+		dspcntr |= DISP_FORMAT_8BPP;
+		break;
+	case DRM_FORMAT_XRGB1555:
+		dspcntr |= DISP_FORMAT_BGRX555;
+		break;
+	case DRM_FORMAT_ARGB1555:
+		dspcntr |= DISP_FORMAT_BGRA555;
+		break;
+	case DRM_FORMAT_RGB565:
+		dspcntr |= DISP_FORMAT_BGRX565;
+		break;
+	case DRM_FORMAT_XRGB8888:
+		dspcntr |= DISP_FORMAT_BGRX888;
+		break;
+	case DRM_FORMAT_XBGR8888:
+		dspcntr |= DISP_FORMAT_RGBX888;
+		break;
+	case DRM_FORMAT_ARGB8888:
+		dspcntr |= DISP_FORMAT_BGRA888;
+		break;
+	case DRM_FORMAT_ABGR8888:
+		dspcntr |= DISP_FORMAT_RGBA888;
+		break;
+	case DRM_FORMAT_XRGB2101010:
+		dspcntr |= DISP_FORMAT_BGRX101010;
+		break;
+	case DRM_FORMAT_XBGR2101010:
+		dspcntr |= DISP_FORMAT_RGBX101010;
+		break;
+	case DRM_FORMAT_ARGB2101010:
+		dspcntr |= DISP_FORMAT_BGRA101010;
+		break;
+	case DRM_FORMAT_ABGR2101010:
+		dspcntr |= DISP_FORMAT_RGBA101010;
+		break;
+	case DRM_FORMAT_XBGR16161616F:
+		dspcntr |= DISP_FORMAT_RGBX161616;
+		break;
+	default:
+		MISSING_CASE(fb->format->format);
+		return 0;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 4 &&
+	    fb->modifier == I915_FORMAT_MOD_X_TILED)
+		dspcntr |= DISP_TILED;
+
+	if (rotation & DRM_MODE_ROTATE_180)
+		dspcntr |= DISP_ROTATE_180;
+
+	if (rotation & DRM_MODE_REFLECT_X)
+		dspcntr |= DISP_MIRROR;
+
+	return dspcntr;
+}
+
+int i9xx_check_plane_surface(struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int src_x, src_y, src_w;
+	u32 offset;
+	int ret;
+
+	ret = intel_plane_compute_gtt(plane_state);
+	if (ret)
+		return ret;
+
+	if (!plane_state->uapi.visible)
+		return 0;
+
+	src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+	src_x = plane_state->uapi.src.x1 >> 16;
+	src_y = plane_state->uapi.src.y1 >> 16;
+
+	/* Undocumented hardware limit on i965/g4x/vlv/chv */
+	if (HAS_GMCH(dev_priv) && fb->format->cpp[0] == 8 && src_w > 2048)
+		return -EINVAL;
+
+	intel_add_fb_offsets(&src_x, &src_y, plane_state, 0);
+
+	if (DISPLAY_VER(dev_priv) >= 4)
+		offset = intel_plane_compute_aligned_offset(&src_x, &src_y,
+							    plane_state, 0);
+	else
+		offset = 0;
+
+	/*
+	 * When using an X-tiled surface the plane starts to
+	 * misbehave if the x offset + width exceeds the stride.
+	 * hsw/bdw: underrun galore
+	 * ilk/snb/ivb: wrap to the next tile row mid scanout
+	 * i965/g4x: so far appear immune to this
+	 * vlv/chv: TODO check
+	 *
+	 * Linear surfaces seem to work just fine, even on hsw/bdw
+	 * despite them not using the linear offset anymore.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 4 && fb->modifier == I915_FORMAT_MOD_X_TILED) {
+		u32 alignment = intel_surf_alignment(fb, 0);
+		int cpp = fb->format->cpp[0];
+
+		while ((src_x + src_w) * cpp > plane_state->view.color_plane[0].mapping_stride) {
+			if (offset == 0) {
+				drm_dbg_kms(&dev_priv->drm,
+					    "Unable to find suitable display surface offset due to X-tiling\n");
+				return -EINVAL;
+			}
+
+			offset = intel_plane_adjust_aligned_offset(&src_x, &src_y, plane_state, 0,
+								   offset, offset - alignment);
+		}
+	}
+
+	/*
+	 * Put the final coordinates back so that the src
+	 * coordinate checks will see the right values.
+	 */
+	drm_rect_translate_to(&plane_state->uapi.src,
+			      src_x << 16, src_y << 16);
+
+	/* HSW/BDW do this automagically in hardware */
+	if (!IS_HASWELL(dev_priv) && !IS_BROADWELL(dev_priv)) {
+		unsigned int rotation = plane_state->hw.rotation;
+		int src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+		int src_h = drm_rect_height(&plane_state->uapi.src) >> 16;
+
+		if (rotation & DRM_MODE_ROTATE_180) {
+			src_x += src_w - 1;
+			src_y += src_h - 1;
+		} else if (rotation & DRM_MODE_REFLECT_X) {
+			src_x += src_w - 1;
+		}
+	}
+
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
+		drm_WARN_ON(&dev_priv->drm, src_x > 8191 || src_y > 4095);
+	} else if (DISPLAY_VER(dev_priv) >= 4 &&
+		   fb->modifier == I915_FORMAT_MOD_X_TILED) {
+		drm_WARN_ON(&dev_priv->drm, src_x > 4095 || src_y > 4095);
+	}
+
+	plane_state->view.color_plane[0].offset = offset;
+	plane_state->view.color_plane[0].x = src_x;
+	plane_state->view.color_plane[0].y = src_y;
+
+	return 0;
+}
+
+static int
+i9xx_plane_check(struct intel_crtc_state *crtc_state,
+		 struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	int ret;
+
+	ret = chv_plane_check_rotation(plane_state);
+	if (ret)
+		return ret;
+
+	ret = intel_atomic_plane_check_clipping(plane_state, crtc_state,
+						DRM_PLANE_NO_SCALING,
+						DRM_PLANE_NO_SCALING,
+						i9xx_plane_has_windowing(plane));
+	if (ret)
+		return ret;
+
+	ret = i9xx_check_plane_surface(plane_state);
+	if (ret)
+		return ret;
+
+	if (!plane_state->uapi.visible)
+		return 0;
+
+	ret = intel_plane_check_src_coordinates(plane_state);
+	if (ret)
+		return ret;
+
+	plane_state->ctl = i9xx_plane_ctl(crtc_state, plane_state);
+
+	return 0;
+}
+
+static u32 i9xx_plane_ctl_crtc(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 dspcntr = 0;
+
+	if (crtc_state->gamma_enable)
+		dspcntr |= DISP_PIPE_GAMMA_ENABLE;
+
+	if (crtc_state->csc_enable)
+		dspcntr |= DISP_PIPE_CSC_ENABLE;
+
+	if (DISPLAY_VER(dev_priv) < 5)
+		dspcntr |= DISP_PIPE_SEL(crtc->pipe);
+
+	return dspcntr;
+}
+
+static void i9xx_plane_ratio(const struct intel_crtc_state *crtc_state,
+			     const struct intel_plane_state *plane_state,
+			     unsigned int *num, unsigned int *den)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int cpp = fb->format->cpp[0];
+
+	/*
+	 * g4x bspec says 64bpp pixel rate can't exceed 80%
+	 * of cdclk when the sprite plane is enabled on the
+	 * same pipe. ilk/snb bspec says 64bpp pixel rate is
+	 * never allowed to exceed 80% of cdclk. Let's just go
+	 * with the ilk/snb limit always.
+	 */
+	if (cpp == 8) {
+		*num = 10;
+		*den = 8;
+	} else {
+		*num = 1;
+		*den = 1;
+	}
+}
+
+static int i9xx_plane_min_cdclk(const struct intel_crtc_state *crtc_state,
+				const struct intel_plane_state *plane_state)
+{
+	unsigned int pixel_rate;
+	unsigned int num, den;
+
+	/*
+	 * Note that crtc_state->pixel_rate accounts for both
+	 * horizontal and vertical panel fitter downscaling factors.
+	 * Pre-HSW bspec tells us to only consider the horizontal
+	 * downscaling factor here. We ignore that and just consider
+	 * both for simplicity.
+	 */
+	pixel_rate = crtc_state->pixel_rate;
+
+	i9xx_plane_ratio(crtc_state, plane_state, &num, &den);
+
+	/* two pixels per clock with double wide pipe */
+	if (crtc_state->double_wide)
+		den *= 2;
+
+	return DIV_ROUND_UP(pixel_rate * num, den);
+}
+
+static void i9xx_plane_update_noarm(struct intel_plane *plane,
+				    const struct intel_crtc_state *crtc_state,
+				    const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+
+	intel_de_write_fw(dev_priv, DSPSTRIDE(i9xx_plane),
+			  plane_state->view.color_plane[0].mapping_stride);
+
+	if (DISPLAY_VER(dev_priv) < 4) {
+		int crtc_x = plane_state->uapi.dst.x1;
+		int crtc_y = plane_state->uapi.dst.y1;
+		int crtc_w = drm_rect_width(&plane_state->uapi.dst);
+		int crtc_h = drm_rect_height(&plane_state->uapi.dst);
+
+		/*
+		 * PLANE_A doesn't actually have a full window
+		 * generator but let's assume we still need to
+		 * program whatever is there.
+		 */
+		intel_de_write_fw(dev_priv, DSPPOS(i9xx_plane),
+				  DISP_POS_Y(crtc_y) | DISP_POS_X(crtc_x));
+		intel_de_write_fw(dev_priv, DSPSIZE(i9xx_plane),
+				  DISP_HEIGHT(crtc_h - 1) | DISP_WIDTH(crtc_w - 1));
+	}
+}
+
+static void i9xx_plane_update_arm(struct intel_plane *plane,
+				  const struct intel_crtc_state *crtc_state,
+				  const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+	int x = plane_state->view.color_plane[0].x;
+	int y = plane_state->view.color_plane[0].y;
+	u32 dspcntr, dspaddr_offset, linear_offset;
+
+	dspcntr = plane_state->ctl | i9xx_plane_ctl_crtc(crtc_state);
+
+	linear_offset = intel_fb_xy_to_linear(x, y, plane_state, 0);
+
+	if (DISPLAY_VER(dev_priv) >= 4)
+		dspaddr_offset = plane_state->view.color_plane[0].offset;
+	else
+		dspaddr_offset = linear_offset;
+
+	if (IS_CHERRYVIEW(dev_priv) && i9xx_plane == PLANE_B) {
+		int crtc_x = plane_state->uapi.dst.x1;
+		int crtc_y = plane_state->uapi.dst.y1;
+		int crtc_w = drm_rect_width(&plane_state->uapi.dst);
+		int crtc_h = drm_rect_height(&plane_state->uapi.dst);
+
+		intel_de_write_fw(dev_priv, PRIMPOS(i9xx_plane),
+				  PRIM_POS_Y(crtc_y) | PRIM_POS_X(crtc_x));
+		intel_de_write_fw(dev_priv, PRIMSIZE(i9xx_plane),
+				  PRIM_HEIGHT(crtc_h - 1) | PRIM_WIDTH(crtc_w - 1));
+		intel_de_write_fw(dev_priv, PRIMCNSTALPHA(i9xx_plane), 0);
+	}
+
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
+		intel_de_write_fw(dev_priv, DSPOFFSET(i9xx_plane),
+				  DISP_OFFSET_Y(y) | DISP_OFFSET_X(x));
+	} else if (DISPLAY_VER(dev_priv) >= 4) {
+		intel_de_write_fw(dev_priv, DSPLINOFF(i9xx_plane),
+				  linear_offset);
+		intel_de_write_fw(dev_priv, DSPTILEOFF(i9xx_plane),
+				  DISP_OFFSET_Y(y) | DISP_OFFSET_X(x));
+	}
+
+	/*
+	 * The control register self-arms if the plane was previously
+	 * disabled. Try to make the plane enable atomic by writing
+	 * the control register just before the surface register.
+	 */
+	intel_de_write_fw(dev_priv, DSPCNTR(i9xx_plane), dspcntr);
+
+	if (DISPLAY_VER(dev_priv) >= 4)
+		intel_de_write_fw(dev_priv, DSPSURF(i9xx_plane),
+				  intel_plane_ggtt_offset(plane_state) + dspaddr_offset);
+	else
+		intel_de_write_fw(dev_priv, DSPADDR(i9xx_plane),
+				  intel_plane_ggtt_offset(plane_state) + dspaddr_offset);
+}
+
+static void i830_plane_update_arm(struct intel_plane *plane,
+				  const struct intel_crtc_state *crtc_state,
+				  const struct intel_plane_state *plane_state)
+{
+	/*
+	 * On i830/i845 all registers are self-arming [ALM040].
+	 *
+	 * Additional breakage on i830 causes register reads to return
+	 * the last latched value instead of the last written value [ALM026].
+	 */
+	i9xx_plane_update_noarm(plane, crtc_state, plane_state);
+	i9xx_plane_update_arm(plane, crtc_state, plane_state);
+}
+
+static void i9xx_plane_disable_arm(struct intel_plane *plane,
+				   const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+	u32 dspcntr;
+
+	/*
+	 * DSPCNTR pipe gamma enable on g4x+ and pipe csc
+	 * enable on ilk+ affect the pipe bottom color as
+	 * well, so we must configure them even if the plane
+	 * is disabled.
+	 *
+	 * On pre-g4x there is no way to gamma correct the
+	 * pipe bottom color but we'll keep on doing this
+	 * anyway so that the crtc state readout works correctly.
+	 */
+	dspcntr = i9xx_plane_ctl_crtc(crtc_state);
+
+	intel_de_write_fw(dev_priv, DSPCNTR(i9xx_plane), dspcntr);
+
+	if (DISPLAY_VER(dev_priv) >= 4)
+		intel_de_write_fw(dev_priv, DSPSURF(i9xx_plane), 0);
+	else
+		intel_de_write_fw(dev_priv, DSPADDR(i9xx_plane), 0);
+}
+
+static void
+g4x_primary_async_flip(struct intel_plane *plane,
+		       const struct intel_crtc_state *crtc_state,
+		       const struct intel_plane_state *plane_state,
+		       bool async_flip)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	u32 dspcntr = plane_state->ctl | i9xx_plane_ctl_crtc(crtc_state);
+	u32 dspaddr_offset = plane_state->view.color_plane[0].offset;
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+
+	if (async_flip)
+		dspcntr |= DISP_ASYNC_FLIP;
+
+	intel_de_write_fw(dev_priv, DSPCNTR(i9xx_plane), dspcntr);
+
+	intel_de_write_fw(dev_priv, DSPSURF(i9xx_plane),
+			  intel_plane_ggtt_offset(plane_state) + dspaddr_offset);
+}
+
+static void
+vlv_primary_async_flip(struct intel_plane *plane,
+		       const struct intel_crtc_state *crtc_state,
+		       const struct intel_plane_state *plane_state,
+		       bool async_flip)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	u32 dspaddr_offset = plane_state->view.color_plane[0].offset;
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+
+	intel_de_write_fw(dev_priv, DSPADDR_VLV(i9xx_plane),
+			  intel_plane_ggtt_offset(plane_state) + dspaddr_offset);
+}
+
+static void
+bdw_primary_enable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	bdw_enable_pipe_irq(i915, pipe, GEN8_PIPE_PRIMARY_FLIP_DONE);
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+bdw_primary_disable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	bdw_disable_pipe_irq(i915, pipe, GEN8_PIPE_PRIMARY_FLIP_DONE);
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+ivb_primary_enable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	spin_lock_irq(&i915->irq_lock);
+	ilk_enable_display_irq(i915, DE_PLANE_FLIP_DONE_IVB(plane->i9xx_plane));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+ivb_primary_disable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	spin_lock_irq(&i915->irq_lock);
+	ilk_disable_display_irq(i915, DE_PLANE_FLIP_DONE_IVB(plane->i9xx_plane));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+ilk_primary_enable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	spin_lock_irq(&i915->irq_lock);
+	ilk_enable_display_irq(i915, DE_PLANE_FLIP_DONE(plane->i9xx_plane));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+ilk_primary_disable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	spin_lock_irq(&i915->irq_lock);
+	ilk_disable_display_irq(i915, DE_PLANE_FLIP_DONE(plane->i9xx_plane));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+vlv_primary_enable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	i915_enable_pipestat(i915, pipe, PLANE_FLIP_DONE_INT_STATUS_VLV);
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+vlv_primary_disable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	i915_disable_pipestat(i915, pipe, PLANE_FLIP_DONE_INT_STATUS_VLV);
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static bool i9xx_plane_get_hw_state(struct intel_plane *plane,
+				    enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum intel_display_power_domain power_domain;
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+	intel_wakeref_t wakeref;
+	bool ret;
+	u32 val;
+
+	/*
+	 * Not 100% correct for planes that can move between pipes,
+	 * but that's only the case for gen2-4 which don't have any
+	 * display power wells.
+	 */
+	power_domain = POWER_DOMAIN_PIPE(plane->pipe);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref)
+		return false;
+
+	val = intel_de_read(dev_priv, DSPCNTR(i9xx_plane));
+
+	ret = val & DISP_ENABLE;
+
+	if (DISPLAY_VER(dev_priv) >= 5)
+		*pipe = plane->pipe;
+	else
+		*pipe = REG_FIELD_GET(DISP_PIPE_SEL_MASK, val);
+
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+
+	return ret;
+}
+
+static unsigned int
+hsw_primary_max_stride(struct intel_plane *plane,
+		       u32 pixel_format, u64 modifier,
+		       unsigned int rotation)
+{
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+
+	/* Limit to 8k pixels to guarantee OFFSET.x doesn't get too big. */
+	return min(8192 * cpp, 32 * 1024);
+}
+
+static unsigned int
+ilk_primary_max_stride(struct intel_plane *plane,
+		       u32 pixel_format, u64 modifier,
+		       unsigned int rotation)
+{
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+
+	/* Limit to 4k pixels to guarantee TILEOFF.x doesn't get too big. */
+	if (modifier == I915_FORMAT_MOD_X_TILED)
+		return min(4096 * cpp, 32 * 1024);
+	else
+		return 32 * 1024;
+}
+
+unsigned int
+i965_plane_max_stride(struct intel_plane *plane,
+		      u32 pixel_format, u64 modifier,
+		      unsigned int rotation)
+{
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+
+	/* Limit to 4k pixels to guarantee TILEOFF.x doesn't get too big. */
+	if (modifier == I915_FORMAT_MOD_X_TILED)
+		return min(4096 * cpp, 16 * 1024);
+	else
+		return 32 * 1024;
+}
+
+static unsigned int
+i9xx_plane_max_stride(struct intel_plane *plane,
+		      u32 pixel_format, u64 modifier,
+		      unsigned int rotation)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+
+	if (DISPLAY_VER(dev_priv) >= 3) {
+		if (modifier == I915_FORMAT_MOD_X_TILED)
+			return 8*1024;
+		else
+			return 16*1024;
+	} else {
+		if (plane->i9xx_plane == PLANE_C)
+			return 4*1024;
+		else
+			return 8*1024;
+	}
+}
+
+static const struct drm_plane_funcs i965_plane_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = intel_plane_destroy,
+	.atomic_duplicate_state = intel_plane_duplicate_state,
+	.atomic_destroy_state = intel_plane_destroy_state,
+	.format_mod_supported = i965_plane_format_mod_supported,
+};
+
+static const struct drm_plane_funcs i8xx_plane_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = intel_plane_destroy,
+	.atomic_duplicate_state = intel_plane_duplicate_state,
+	.atomic_destroy_state = intel_plane_destroy_state,
+	.format_mod_supported = i8xx_plane_format_mod_supported,
+};
+
+struct intel_plane *
+intel_primary_plane_create(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	struct intel_plane *plane;
+	const struct drm_plane_funcs *plane_funcs;
+	unsigned int supported_rotations;
+	const u64 *modifiers;
+	const u32 *formats;
+	int num_formats;
+	int ret, zpos;
+
+	plane = intel_plane_alloc();
+	if (IS_ERR(plane))
+		return plane;
+
+	plane->pipe = pipe;
+	/*
+	 * On gen2/3 only plane A can do FBC, but the panel fitter and LVDS
+	 * port is hooked to pipe B. Hence we want plane A feeding pipe B.
+	 */
+	if (HAS_FBC(dev_priv) && DISPLAY_VER(dev_priv) < 4 &&
+	    INTEL_NUM_PIPES(dev_priv) == 2)
+		plane->i9xx_plane = (enum i9xx_plane_id) !pipe;
+	else
+		plane->i9xx_plane = (enum i9xx_plane_id) pipe;
+	plane->id = PLANE_PRIMARY;
+	plane->frontbuffer_bit = INTEL_FRONTBUFFER(pipe, plane->id);
+
+	intel_fbc_add_plane(i9xx_plane_fbc(dev_priv, plane->i9xx_plane), plane);
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		formats = vlv_primary_formats;
+		num_formats = ARRAY_SIZE(vlv_primary_formats);
+	} else if (DISPLAY_VER(dev_priv) >= 4) {
+		/*
+		 * WaFP16GammaEnabling:ivb
+		 * "Workaround : When using the 64-bit format, the plane
+		 *  output on each color channel has one quarter amplitude.
+		 *  It can be brought up to full amplitude by using pipe
+		 *  gamma correction or pipe color space conversion to
+		 *  multiply the plane output by four."
+		 *
+		 * There is no dedicated plane gamma for the primary plane,
+		 * and using the pipe gamma/csc could conflict with other
+		 * planes, so we choose not to expose fp16 on IVB primary
+		 * planes. HSW primary planes no longer have this problem.
+		 */
+		if (IS_IVYBRIDGE(dev_priv)) {
+			formats = ivb_primary_formats;
+			num_formats = ARRAY_SIZE(ivb_primary_formats);
+		} else {
+			formats = i965_primary_formats;
+			num_formats = ARRAY_SIZE(i965_primary_formats);
+		}
+	} else {
+		formats = i8xx_primary_formats;
+		num_formats = ARRAY_SIZE(i8xx_primary_formats);
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 4)
+		plane_funcs = &i965_plane_funcs;
+	else
+		plane_funcs = &i8xx_plane_funcs;
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		plane->min_cdclk = vlv_plane_min_cdclk;
+	else if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		plane->min_cdclk = hsw_plane_min_cdclk;
+	else if (IS_IVYBRIDGE(dev_priv))
+		plane->min_cdclk = ivb_plane_min_cdclk;
+	else
+		plane->min_cdclk = i9xx_plane_min_cdclk;
+
+	if (HAS_GMCH(dev_priv)) {
+		if (DISPLAY_VER(dev_priv) >= 4)
+			plane->max_stride = i965_plane_max_stride;
+		else
+			plane->max_stride = i9xx_plane_max_stride;
+	} else {
+		if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+			plane->max_stride = hsw_primary_max_stride;
+		else
+			plane->max_stride = ilk_primary_max_stride;
+	}
+
+	if (IS_I830(dev_priv) || IS_I845G(dev_priv)) {
+		plane->update_arm = i830_plane_update_arm;
+	} else {
+		plane->update_noarm = i9xx_plane_update_noarm;
+		plane->update_arm = i9xx_plane_update_arm;
+	}
+	plane->disable_arm = i9xx_plane_disable_arm;
+	plane->get_hw_state = i9xx_plane_get_hw_state;
+	plane->check_plane = i9xx_plane_check;
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		plane->async_flip = vlv_primary_async_flip;
+		plane->enable_flip_done = vlv_primary_enable_flip_done;
+		plane->disable_flip_done = vlv_primary_disable_flip_done;
+	} else if (IS_BROADWELL(dev_priv)) {
+		plane->need_async_flip_disable_wa = true;
+		plane->async_flip = g4x_primary_async_flip;
+		plane->enable_flip_done = bdw_primary_enable_flip_done;
+		plane->disable_flip_done = bdw_primary_disable_flip_done;
+	} else if (DISPLAY_VER(dev_priv) >= 7) {
+		plane->async_flip = g4x_primary_async_flip;
+		plane->enable_flip_done = ivb_primary_enable_flip_done;
+		plane->disable_flip_done = ivb_primary_disable_flip_done;
+	} else if (DISPLAY_VER(dev_priv) >= 5) {
+		plane->async_flip = g4x_primary_async_flip;
+		plane->enable_flip_done = ilk_primary_enable_flip_done;
+		plane->disable_flip_done = ilk_primary_disable_flip_done;
+	}
+
+	modifiers = intel_fb_plane_get_modifiers(dev_priv, INTEL_PLANE_CAP_TILING_X);
+
+	if (DISPLAY_VER(dev_priv) >= 5 || IS_G4X(dev_priv))
+		ret = drm_universal_plane_init(&dev_priv->drm, &plane->base,
+					       0, plane_funcs,
+					       formats, num_formats,
+					       modifiers,
+					       DRM_PLANE_TYPE_PRIMARY,
+					       "primary %c", pipe_name(pipe));
+	else
+		ret = drm_universal_plane_init(&dev_priv->drm, &plane->base,
+					       0, plane_funcs,
+					       formats, num_formats,
+					       modifiers,
+					       DRM_PLANE_TYPE_PRIMARY,
+					       "plane %c",
+					       plane_name(plane->i9xx_plane));
+
+	kfree(modifiers);
+
+	if (ret)
+		goto fail;
+
+	if (IS_CHERRYVIEW(dev_priv) && pipe == PIPE_B) {
+		supported_rotations =
+			DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180 |
+			DRM_MODE_REFLECT_X;
+	} else if (DISPLAY_VER(dev_priv) >= 4) {
+		supported_rotations =
+			DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180;
+	} else {
+		supported_rotations = DRM_MODE_ROTATE_0;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 4)
+		drm_plane_create_rotation_property(&plane->base,
+						   DRM_MODE_ROTATE_0,
+						   supported_rotations);
+
+	zpos = 0;
+	drm_plane_create_zpos_immutable_property(&plane->base, zpos);
+
+	intel_plane_helper_add(plane);
+
+	return plane;
+
+fail:
+	intel_plane_free(plane);
+
+	return ERR_PTR(ret);
+}
+
+static int i9xx_format_to_fourcc(int format)
+{
+	switch (format) {
+	case DISP_FORMAT_8BPP:
+		return DRM_FORMAT_C8;
+	case DISP_FORMAT_BGRA555:
+		return DRM_FORMAT_ARGB1555;
+	case DISP_FORMAT_BGRX555:
+		return DRM_FORMAT_XRGB1555;
+	case DISP_FORMAT_BGRX565:
+		return DRM_FORMAT_RGB565;
+	default:
+	case DISP_FORMAT_BGRX888:
+		return DRM_FORMAT_XRGB8888;
+	case DISP_FORMAT_RGBX888:
+		return DRM_FORMAT_XBGR8888;
+	case DISP_FORMAT_BGRA888:
+		return DRM_FORMAT_ARGB8888;
+	case DISP_FORMAT_RGBA888:
+		return DRM_FORMAT_ABGR8888;
+	case DISP_FORMAT_BGRX101010:
+		return DRM_FORMAT_XRGB2101010;
+	case DISP_FORMAT_RGBX101010:
+		return DRM_FORMAT_XBGR2101010;
+	case DISP_FORMAT_BGRA101010:
+		return DRM_FORMAT_ARGB2101010;
+	case DISP_FORMAT_RGBA101010:
+		return DRM_FORMAT_ABGR2101010;
+	case DISP_FORMAT_RGBX161616:
+		return DRM_FORMAT_XBGR16161616F;
+	}
+}
+
+void
+i9xx_get_initial_plane_config(struct intel_crtc *crtc,
+			      struct intel_initial_plane_config *plane_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_plane *plane = to_intel_plane(crtc->base.primary);
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+	enum pipe pipe;
+	u32 val, base, offset;
+	int fourcc, pixel_format;
+	unsigned int aligned_height;
+	struct drm_framebuffer *fb;
+	struct intel_framebuffer *intel_fb;
+
+	if (!plane->get_hw_state(plane, &pipe))
+		return;
+
+	drm_WARN_ON(dev, pipe != crtc->pipe);
+
+	intel_fb = kzalloc(sizeof(*intel_fb), GFP_KERNEL);
+	if (!intel_fb) {
+		drm_dbg_kms(&dev_priv->drm, "failed to alloc fb\n");
+		return;
+	}
+
+	fb = &intel_fb->base;
+
+	fb->dev = dev;
+
+	val = intel_de_read(dev_priv, DSPCNTR(i9xx_plane));
+
+	if (DISPLAY_VER(dev_priv) >= 4) {
+		if (val & DISP_TILED) {
+			plane_config->tiling = I915_TILING_X;
+			fb->modifier = I915_FORMAT_MOD_X_TILED;
+		}
+
+		if (val & DISP_ROTATE_180)
+			plane_config->rotation = DRM_MODE_ROTATE_180;
+	}
+
+	if (IS_CHERRYVIEW(dev_priv) && pipe == PIPE_B &&
+	    val & DISP_MIRROR)
+		plane_config->rotation |= DRM_MODE_REFLECT_X;
+
+	pixel_format = val & DISP_FORMAT_MASK;
+	fourcc = i9xx_format_to_fourcc(pixel_format);
+	fb->format = drm_format_info(fourcc);
+
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
+		offset = intel_de_read(dev_priv, DSPOFFSET(i9xx_plane));
+		base = intel_de_read(dev_priv, DSPSURF(i9xx_plane)) & DISP_ADDR_MASK;
+	} else if (DISPLAY_VER(dev_priv) >= 4) {
+		if (plane_config->tiling)
+			offset = intel_de_read(dev_priv,
+					       DSPTILEOFF(i9xx_plane));
+		else
+			offset = intel_de_read(dev_priv,
+					       DSPLINOFF(i9xx_plane));
+		base = intel_de_read(dev_priv, DSPSURF(i9xx_plane)) & DISP_ADDR_MASK;
+	} else {
+		base = intel_de_read(dev_priv, DSPADDR(i9xx_plane));
+	}
+	plane_config->base = base;
+
+	val = intel_de_read(dev_priv, PIPESRC(pipe));
+	fb->width = REG_FIELD_GET(PIPESRC_WIDTH_MASK, val) + 1;
+	fb->height = REG_FIELD_GET(PIPESRC_HEIGHT_MASK, val) + 1;
+
+	val = intel_de_read(dev_priv, DSPSTRIDE(i9xx_plane));
+	fb->pitches[0] = val & 0xffffffc0;
+
+	aligned_height = intel_fb_align_height(fb, 0, fb->height);
+
+	plane_config->size = fb->pitches[0] * aligned_height;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "%s/%s with fb: size=%dx%d@%d, offset=%x, pitch %d, size 0x%x\n",
+		    crtc->base.name, plane->base.name, fb->width, fb->height,
+		    fb->format->cpp[0] * 8, base, fb->pitches[0],
+		    plane_config->size);
+
+	plane_config->fb = intel_fb;
+}
diff --git a/drivers/gpu/drm/i915/display/i9xx_plane.h b/drivers/gpu/drm/i915/display/i9xx_plane.h
new file mode 100644
index 000000000..027b66053
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/i9xx_plane.h
@@ -0,0 +1,28 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _I9XX_PLANE_H_
+#define _I9XX_PLANE_H_
+
+#include <linux/types.h>
+
+enum pipe;
+struct drm_i915_private;
+struct intel_crtc;
+struct intel_initial_plane_config;
+struct intel_plane;
+struct intel_plane_state;
+
+unsigned int i965_plane_max_stride(struct intel_plane *plane,
+				   u32 pixel_format, u64 modifier,
+				   unsigned int rotation);
+int i9xx_check_plane_surface(struct intel_plane_state *plane_state);
+
+struct intel_plane *
+intel_primary_plane_create(struct drm_i915_private *dev_priv, enum pipe pipe);
+
+void i9xx_get_initial_plane_config(struct intel_crtc *crtc,
+				   struct intel_initial_plane_config *plane_config);
+#endif
diff --git a/drivers/gpu/drm/i915/display/icl_dsi.c b/drivers/gpu/drm/i915/display/icl_dsi.c
new file mode 100644
index 000000000..f7422f0cf
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/icl_dsi.c
@@ -0,0 +1,2127 @@
+/*
+ * Copyright © 2018 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *   Madhav Chauhan <madhav.chauhan@intel.com>
+ *   Jani Nikula <jani.nikula@intel.com>
+ */
+
+#include <drm/display/drm_dsc_helper.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_mipi_dsi.h>
+
+#include "icl_dsi.h"
+#include "icl_dsi_regs.h"
+#include "intel_atomic.h"
+#include "intel_backlight.h"
+#include "intel_backlight_regs.h"
+#include "intel_combo_phy.h"
+#include "intel_combo_phy_regs.h"
+#include "intel_connector.h"
+#include "intel_crtc.h"
+#include "intel_ddi.h"
+#include "intel_de.h"
+#include "intel_dsi.h"
+#include "intel_dsi_vbt.h"
+#include "intel_panel.h"
+#include "intel_vdsc.h"
+#include "skl_scaler.h"
+#include "skl_universal_plane.h"
+
+static int header_credits_available(struct drm_i915_private *dev_priv,
+				    enum transcoder dsi_trans)
+{
+	return (intel_de_read(dev_priv, DSI_CMD_TXCTL(dsi_trans)) & FREE_HEADER_CREDIT_MASK)
+		>> FREE_HEADER_CREDIT_SHIFT;
+}
+
+static int payload_credits_available(struct drm_i915_private *dev_priv,
+				     enum transcoder dsi_trans)
+{
+	return (intel_de_read(dev_priv, DSI_CMD_TXCTL(dsi_trans)) & FREE_PLOAD_CREDIT_MASK)
+		>> FREE_PLOAD_CREDIT_SHIFT;
+}
+
+static bool wait_for_header_credits(struct drm_i915_private *dev_priv,
+				    enum transcoder dsi_trans, int hdr_credit)
+{
+	if (wait_for_us(header_credits_available(dev_priv, dsi_trans) >=
+			hdr_credit, 100)) {
+		drm_err(&dev_priv->drm, "DSI header credits not released\n");
+		return false;
+	}
+
+	return true;
+}
+
+static bool wait_for_payload_credits(struct drm_i915_private *dev_priv,
+				     enum transcoder dsi_trans, int payld_credit)
+{
+	if (wait_for_us(payload_credits_available(dev_priv, dsi_trans) >=
+			payld_credit, 100)) {
+		drm_err(&dev_priv->drm, "DSI payload credits not released\n");
+		return false;
+	}
+
+	return true;
+}
+
+static enum transcoder dsi_port_to_transcoder(enum port port)
+{
+	if (port == PORT_A)
+		return TRANSCODER_DSI_0;
+	else
+		return TRANSCODER_DSI_1;
+}
+
+static void wait_for_cmds_dispatched_to_panel(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	struct mipi_dsi_device *dsi;
+	enum port port;
+	enum transcoder dsi_trans;
+	int ret;
+
+	/* wait for header/payload credits to be released */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		wait_for_header_credits(dev_priv, dsi_trans, MAX_HEADER_CREDIT);
+		wait_for_payload_credits(dev_priv, dsi_trans, MAX_PLOAD_CREDIT);
+	}
+
+	/* send nop DCS command */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi = intel_dsi->dsi_hosts[port]->device;
+		dsi->mode_flags |= MIPI_DSI_MODE_LPM;
+		dsi->channel = 0;
+		ret = mipi_dsi_dcs_nop(dsi);
+		if (ret < 0)
+			drm_err(&dev_priv->drm,
+				"error sending DCS NOP command\n");
+	}
+
+	/* wait for header credits to be released */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		wait_for_header_credits(dev_priv, dsi_trans, MAX_HEADER_CREDIT);
+	}
+
+	/* wait for LP TX in progress bit to be cleared */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		if (wait_for_us(!(intel_de_read(dev_priv, DSI_LP_MSG(dsi_trans)) &
+				  LPTX_IN_PROGRESS), 20))
+			drm_err(&dev_priv->drm, "LPTX bit not cleared\n");
+	}
+}
+
+static int dsi_send_pkt_payld(struct intel_dsi_host *host,
+			      const struct mipi_dsi_packet *packet)
+{
+	struct intel_dsi *intel_dsi = host->intel_dsi;
+	struct drm_i915_private *i915 = to_i915(intel_dsi->base.base.dev);
+	enum transcoder dsi_trans = dsi_port_to_transcoder(host->port);
+	const u8 *data = packet->payload;
+	u32 len = packet->payload_length;
+	int i, j;
+
+	/* payload queue can accept *256 bytes*, check limit */
+	if (len > MAX_PLOAD_CREDIT * 4) {
+		drm_err(&i915->drm, "payload size exceeds max queue limit\n");
+		return -EINVAL;
+	}
+
+	for (i = 0; i < len; i += 4) {
+		u32 tmp = 0;
+
+		if (!wait_for_payload_credits(i915, dsi_trans, 1))
+			return -EBUSY;
+
+		for (j = 0; j < min_t(u32, len - i, 4); j++)
+			tmp |= *data++ << 8 * j;
+
+		intel_de_write(i915, DSI_CMD_TXPYLD(dsi_trans), tmp);
+	}
+
+	return 0;
+}
+
+static int dsi_send_pkt_hdr(struct intel_dsi_host *host,
+			    const struct mipi_dsi_packet *packet,
+			    bool enable_lpdt)
+{
+	struct intel_dsi *intel_dsi = host->intel_dsi;
+	struct drm_i915_private *dev_priv = to_i915(intel_dsi->base.base.dev);
+	enum transcoder dsi_trans = dsi_port_to_transcoder(host->port);
+	u32 tmp;
+
+	if (!wait_for_header_credits(dev_priv, dsi_trans, 1))
+		return -EBUSY;
+
+	tmp = intel_de_read(dev_priv, DSI_CMD_TXHDR(dsi_trans));
+
+	if (packet->payload)
+		tmp |= PAYLOAD_PRESENT;
+	else
+		tmp &= ~PAYLOAD_PRESENT;
+
+	tmp &= ~VBLANK_FENCE;
+
+	if (enable_lpdt)
+		tmp |= LP_DATA_TRANSFER;
+	else
+		tmp &= ~LP_DATA_TRANSFER;
+
+	tmp &= ~(PARAM_WC_MASK | VC_MASK | DT_MASK);
+	tmp |= ((packet->header[0] & VC_MASK) << VC_SHIFT);
+	tmp |= ((packet->header[0] & DT_MASK) << DT_SHIFT);
+	tmp |= (packet->header[1] << PARAM_WC_LOWER_SHIFT);
+	tmp |= (packet->header[2] << PARAM_WC_UPPER_SHIFT);
+	intel_de_write(dev_priv, DSI_CMD_TXHDR(dsi_trans), tmp);
+
+	return 0;
+}
+
+void icl_dsi_frame_update(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 tmp, mode_flags;
+	enum port port;
+
+	mode_flags = crtc_state->mode_flags;
+
+	/*
+	 * case 1 also covers dual link
+	 * In case of dual link, frame update should be set on
+	 * DSI_0
+	 */
+	if (mode_flags & I915_MODE_FLAG_DSI_USE_TE0)
+		port = PORT_A;
+	else if (mode_flags & I915_MODE_FLAG_DSI_USE_TE1)
+		port = PORT_B;
+	else
+		return;
+
+	tmp = intel_de_read(dev_priv, DSI_CMD_FRMCTL(port));
+	tmp |= DSI_FRAME_UPDATE_REQUEST;
+	intel_de_write(dev_priv, DSI_CMD_FRMCTL(port), tmp);
+}
+
+static void dsi_program_swing_and_deemphasis(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum phy phy;
+	u32 tmp;
+	int lane;
+
+	for_each_dsi_phy(phy, intel_dsi->phys) {
+		/*
+		 * Program voltage swing and pre-emphasis level values as per
+		 * table in BSPEC under DDI buffer programing
+		 */
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW5_LN(0, phy));
+		tmp &= ~(SCALING_MODE_SEL_MASK | RTERM_SELECT_MASK);
+		tmp |= SCALING_MODE_SEL(0x2);
+		tmp |= TAP2_DISABLE | TAP3_DISABLE;
+		tmp |= RTERM_SELECT(0x6);
+		intel_de_write(dev_priv, ICL_PORT_TX_DW5_GRP(phy), tmp);
+
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW5_AUX(phy));
+		tmp &= ~(SCALING_MODE_SEL_MASK | RTERM_SELECT_MASK);
+		tmp |= SCALING_MODE_SEL(0x2);
+		tmp |= TAP2_DISABLE | TAP3_DISABLE;
+		tmp |= RTERM_SELECT(0x6);
+		intel_de_write(dev_priv, ICL_PORT_TX_DW5_AUX(phy), tmp);
+
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW2_LN(0, phy));
+		tmp &= ~(SWING_SEL_LOWER_MASK | SWING_SEL_UPPER_MASK |
+			 RCOMP_SCALAR_MASK);
+		tmp |= SWING_SEL_UPPER(0x2);
+		tmp |= SWING_SEL_LOWER(0x2);
+		tmp |= RCOMP_SCALAR(0x98);
+		intel_de_write(dev_priv, ICL_PORT_TX_DW2_GRP(phy), tmp);
+
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW2_AUX(phy));
+		tmp &= ~(SWING_SEL_LOWER_MASK | SWING_SEL_UPPER_MASK |
+			 RCOMP_SCALAR_MASK);
+		tmp |= SWING_SEL_UPPER(0x2);
+		tmp |= SWING_SEL_LOWER(0x2);
+		tmp |= RCOMP_SCALAR(0x98);
+		intel_de_write(dev_priv, ICL_PORT_TX_DW2_AUX(phy), tmp);
+
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW4_AUX(phy));
+		tmp &= ~(POST_CURSOR_1_MASK | POST_CURSOR_2_MASK |
+			 CURSOR_COEFF_MASK);
+		tmp |= POST_CURSOR_1(0x0);
+		tmp |= POST_CURSOR_2(0x0);
+		tmp |= CURSOR_COEFF(0x3f);
+		intel_de_write(dev_priv, ICL_PORT_TX_DW4_AUX(phy), tmp);
+
+		for (lane = 0; lane <= 3; lane++) {
+			/* Bspec: must not use GRP register for write */
+			tmp = intel_de_read(dev_priv,
+					    ICL_PORT_TX_DW4_LN(lane, phy));
+			tmp &= ~(POST_CURSOR_1_MASK | POST_CURSOR_2_MASK |
+				 CURSOR_COEFF_MASK);
+			tmp |= POST_CURSOR_1(0x0);
+			tmp |= POST_CURSOR_2(0x0);
+			tmp |= CURSOR_COEFF(0x3f);
+			intel_de_write(dev_priv,
+				       ICL_PORT_TX_DW4_LN(lane, phy), tmp);
+		}
+	}
+}
+
+static void configure_dual_link_mode(struct intel_encoder *encoder,
+				     const struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	i915_reg_t dss_ctl1_reg, dss_ctl2_reg;
+	u32 dss_ctl1;
+
+	/* FIXME: Move all DSS handling to intel_vdsc.c */
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+
+		dss_ctl1_reg = ICL_PIPE_DSS_CTL1(crtc->pipe);
+		dss_ctl2_reg = ICL_PIPE_DSS_CTL2(crtc->pipe);
+	} else {
+		dss_ctl1_reg = DSS_CTL1;
+		dss_ctl2_reg = DSS_CTL2;
+	}
+
+	dss_ctl1 = intel_de_read(dev_priv, dss_ctl1_reg);
+	dss_ctl1 |= SPLITTER_ENABLE;
+	dss_ctl1 &= ~OVERLAP_PIXELS_MASK;
+	dss_ctl1 |= OVERLAP_PIXELS(intel_dsi->pixel_overlap);
+
+	if (intel_dsi->dual_link == DSI_DUAL_LINK_FRONT_BACK) {
+		const struct drm_display_mode *adjusted_mode =
+					&pipe_config->hw.adjusted_mode;
+		u32 dss_ctl2;
+		u16 hactive = adjusted_mode->crtc_hdisplay;
+		u16 dl_buffer_depth;
+
+		dss_ctl1 &= ~DUAL_LINK_MODE_INTERLEAVE;
+		dl_buffer_depth = hactive / 2 + intel_dsi->pixel_overlap;
+
+		if (dl_buffer_depth > MAX_DL_BUFFER_TARGET_DEPTH)
+			drm_err(&dev_priv->drm,
+				"DL buffer depth exceed max value\n");
+
+		dss_ctl1 &= ~LEFT_DL_BUF_TARGET_DEPTH_MASK;
+		dss_ctl1 |= LEFT_DL_BUF_TARGET_DEPTH(dl_buffer_depth);
+		dss_ctl2 = intel_de_read(dev_priv, dss_ctl2_reg);
+		dss_ctl2 &= ~RIGHT_DL_BUF_TARGET_DEPTH_MASK;
+		dss_ctl2 |= RIGHT_DL_BUF_TARGET_DEPTH(dl_buffer_depth);
+		intel_de_write(dev_priv, dss_ctl2_reg, dss_ctl2);
+	} else {
+		/* Interleave */
+		dss_ctl1 |= DUAL_LINK_MODE_INTERLEAVE;
+	}
+
+	intel_de_write(dev_priv, dss_ctl1_reg, dss_ctl1);
+}
+
+/* aka DSI 8X clock */
+static int afe_clk(struct intel_encoder *encoder,
+		   const struct intel_crtc_state *crtc_state)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	int bpp;
+
+	if (crtc_state->dsc.compression_enable)
+		bpp = crtc_state->dsc.compressed_bpp;
+	else
+		bpp = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format);
+
+	return DIV_ROUND_CLOSEST(intel_dsi->pclk * bpp, intel_dsi->lane_count);
+}
+
+static void gen11_dsi_program_esc_clk_div(struct intel_encoder *encoder,
+					  const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	int afe_clk_khz;
+	int theo_word_clk, act_word_clk;
+	u32 esc_clk_div_m, esc_clk_div_m_phy;
+
+	afe_clk_khz = afe_clk(encoder, crtc_state);
+
+	if (IS_ALDERLAKE_S(dev_priv) || IS_ALDERLAKE_P(dev_priv)) {
+		theo_word_clk = DIV_ROUND_UP(afe_clk_khz, 8 * DSI_MAX_ESC_CLK);
+		act_word_clk = max(3, theo_word_clk + (theo_word_clk + 1) % 2);
+		esc_clk_div_m = act_word_clk * 8;
+		esc_clk_div_m_phy = (act_word_clk - 1) / 2;
+	} else {
+		esc_clk_div_m = DIV_ROUND_UP(afe_clk_khz, DSI_MAX_ESC_CLK);
+	}
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		intel_de_write(dev_priv, ICL_DSI_ESC_CLK_DIV(port),
+			       esc_clk_div_m & ICL_ESC_CLK_DIV_MASK);
+		intel_de_posting_read(dev_priv, ICL_DSI_ESC_CLK_DIV(port));
+	}
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		intel_de_write(dev_priv, ICL_DPHY_ESC_CLK_DIV(port),
+			       esc_clk_div_m & ICL_ESC_CLK_DIV_MASK);
+		intel_de_posting_read(dev_priv, ICL_DPHY_ESC_CLK_DIV(port));
+	}
+
+	if (IS_ALDERLAKE_S(dev_priv) || IS_ALDERLAKE_P(dev_priv)) {
+		for_each_dsi_port(port, intel_dsi->ports) {
+			intel_de_write(dev_priv, ADL_MIPIO_DW(port, 8),
+				       esc_clk_div_m_phy & TX_ESC_CLK_DIV_PHY);
+			intel_de_posting_read(dev_priv, ADL_MIPIO_DW(port, 8));
+		}
+	}
+}
+
+static void get_dsi_io_power_domains(struct drm_i915_private *dev_priv,
+				     struct intel_dsi *intel_dsi)
+{
+	enum port port;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		drm_WARN_ON(&dev_priv->drm, intel_dsi->io_wakeref[port]);
+		intel_dsi->io_wakeref[port] =
+			intel_display_power_get(dev_priv,
+						port == PORT_A ?
+						POWER_DOMAIN_PORT_DDI_IO_A :
+						POWER_DOMAIN_PORT_DDI_IO_B);
+	}
+}
+
+static void gen11_dsi_enable_io_power(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 tmp;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = intel_de_read(dev_priv, ICL_DSI_IO_MODECTL(port));
+		tmp |= COMBO_PHY_MODE_DSI;
+		intel_de_write(dev_priv, ICL_DSI_IO_MODECTL(port), tmp);
+	}
+
+	get_dsi_io_power_domains(dev_priv, intel_dsi);
+}
+
+static void gen11_dsi_power_up_lanes(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum phy phy;
+
+	for_each_dsi_phy(phy, intel_dsi->phys)
+		intel_combo_phy_power_up_lanes(dev_priv, phy, true,
+					       intel_dsi->lane_count, false);
+}
+
+static void gen11_dsi_config_phy_lanes_sequence(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum phy phy;
+	u32 tmp;
+	int lane;
+
+	/* Step 4b(i) set loadgen select for transmit and aux lanes */
+	for_each_dsi_phy(phy, intel_dsi->phys) {
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW4_AUX(phy));
+		tmp &= ~LOADGEN_SELECT;
+		intel_de_write(dev_priv, ICL_PORT_TX_DW4_AUX(phy), tmp);
+		for (lane = 0; lane <= 3; lane++) {
+			tmp = intel_de_read(dev_priv,
+					    ICL_PORT_TX_DW4_LN(lane, phy));
+			tmp &= ~LOADGEN_SELECT;
+			if (lane != 2)
+				tmp |= LOADGEN_SELECT;
+			intel_de_write(dev_priv,
+				       ICL_PORT_TX_DW4_LN(lane, phy), tmp);
+		}
+	}
+
+	/* Step 4b(ii) set latency optimization for transmit and aux lanes */
+	for_each_dsi_phy(phy, intel_dsi->phys) {
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW2_AUX(phy));
+		tmp &= ~FRC_LATENCY_OPTIM_MASK;
+		tmp |= FRC_LATENCY_OPTIM_VAL(0x5);
+		intel_de_write(dev_priv, ICL_PORT_TX_DW2_AUX(phy), tmp);
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW2_LN(0, phy));
+		tmp &= ~FRC_LATENCY_OPTIM_MASK;
+		tmp |= FRC_LATENCY_OPTIM_VAL(0x5);
+		intel_de_write(dev_priv, ICL_PORT_TX_DW2_GRP(phy), tmp);
+
+		/* For EHL, TGL, set latency optimization for PCS_DW1 lanes */
+		if (IS_JSL_EHL(dev_priv) || (DISPLAY_VER(dev_priv) >= 12)) {
+			tmp = intel_de_read(dev_priv,
+					    ICL_PORT_PCS_DW1_AUX(phy));
+			tmp &= ~LATENCY_OPTIM_MASK;
+			tmp |= LATENCY_OPTIM_VAL(0);
+			intel_de_write(dev_priv, ICL_PORT_PCS_DW1_AUX(phy),
+				       tmp);
+
+			tmp = intel_de_read(dev_priv,
+					    ICL_PORT_PCS_DW1_LN(0, phy));
+			tmp &= ~LATENCY_OPTIM_MASK;
+			tmp |= LATENCY_OPTIM_VAL(0x1);
+			intel_de_write(dev_priv, ICL_PORT_PCS_DW1_GRP(phy),
+				       tmp);
+		}
+	}
+
+}
+
+static void gen11_dsi_voltage_swing_program_seq(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	u32 tmp;
+	enum phy phy;
+
+	/* clear common keeper enable bit */
+	for_each_dsi_phy(phy, intel_dsi->phys) {
+		tmp = intel_de_read(dev_priv, ICL_PORT_PCS_DW1_LN(0, phy));
+		tmp &= ~COMMON_KEEPER_EN;
+		intel_de_write(dev_priv, ICL_PORT_PCS_DW1_GRP(phy), tmp);
+		tmp = intel_de_read(dev_priv, ICL_PORT_PCS_DW1_AUX(phy));
+		tmp &= ~COMMON_KEEPER_EN;
+		intel_de_write(dev_priv, ICL_PORT_PCS_DW1_AUX(phy), tmp);
+	}
+
+	/*
+	 * Set SUS Clock Config bitfield to 11b
+	 * Note: loadgen select program is done
+	 * as part of lane phy sequence configuration
+	 */
+	for_each_dsi_phy(phy, intel_dsi->phys) {
+		tmp = intel_de_read(dev_priv, ICL_PORT_CL_DW5(phy));
+		tmp |= SUS_CLOCK_CONFIG;
+		intel_de_write(dev_priv, ICL_PORT_CL_DW5(phy), tmp);
+	}
+
+	/* Clear training enable to change swing values */
+	for_each_dsi_phy(phy, intel_dsi->phys) {
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW5_LN(0, phy));
+		tmp &= ~TX_TRAINING_EN;
+		intel_de_write(dev_priv, ICL_PORT_TX_DW5_GRP(phy), tmp);
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW5_AUX(phy));
+		tmp &= ~TX_TRAINING_EN;
+		intel_de_write(dev_priv, ICL_PORT_TX_DW5_AUX(phy), tmp);
+	}
+
+	/* Program swing and de-emphasis */
+	dsi_program_swing_and_deemphasis(encoder);
+
+	/* Set training enable to trigger update */
+	for_each_dsi_phy(phy, intel_dsi->phys) {
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW5_LN(0, phy));
+		tmp |= TX_TRAINING_EN;
+		intel_de_write(dev_priv, ICL_PORT_TX_DW5_GRP(phy), tmp);
+		tmp = intel_de_read(dev_priv, ICL_PORT_TX_DW5_AUX(phy));
+		tmp |= TX_TRAINING_EN;
+		intel_de_write(dev_priv, ICL_PORT_TX_DW5_AUX(phy), tmp);
+	}
+}
+
+static void gen11_dsi_enable_ddi_buffer(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	u32 tmp;
+	enum port port;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = intel_de_read(dev_priv, DDI_BUF_CTL(port));
+		tmp |= DDI_BUF_CTL_ENABLE;
+		intel_de_write(dev_priv, DDI_BUF_CTL(port), tmp);
+
+		if (wait_for_us(!(intel_de_read(dev_priv, DDI_BUF_CTL(port)) &
+				  DDI_BUF_IS_IDLE),
+				  500))
+			drm_err(&dev_priv->drm, "DDI port:%c buffer idle\n",
+				port_name(port));
+	}
+}
+
+static void
+gen11_dsi_setup_dphy_timings(struct intel_encoder *encoder,
+			     const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	u32 tmp;
+	enum port port;
+	enum phy phy;
+
+	/* Program T-INIT master registers */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = intel_de_read(dev_priv, ICL_DSI_T_INIT_MASTER(port));
+		tmp &= ~DSI_T_INIT_MASTER_MASK;
+		tmp |= intel_dsi->init_count;
+		intel_de_write(dev_priv, ICL_DSI_T_INIT_MASTER(port), tmp);
+	}
+
+	/* Program DPHY clock lanes timings */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		intel_de_write(dev_priv, DPHY_CLK_TIMING_PARAM(port),
+			       intel_dsi->dphy_reg);
+
+		/* shadow register inside display core */
+		intel_de_write(dev_priv, DSI_CLK_TIMING_PARAM(port),
+			       intel_dsi->dphy_reg);
+	}
+
+	/* Program DPHY data lanes timings */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		intel_de_write(dev_priv, DPHY_DATA_TIMING_PARAM(port),
+			       intel_dsi->dphy_data_lane_reg);
+
+		/* shadow register inside display core */
+		intel_de_write(dev_priv, DSI_DATA_TIMING_PARAM(port),
+			       intel_dsi->dphy_data_lane_reg);
+	}
+
+	/*
+	 * If DSI link operating at or below an 800 MHz,
+	 * TA_SURE should be override and programmed to
+	 * a value '0' inside TA_PARAM_REGISTERS otherwise
+	 * leave all fields at HW default values.
+	 */
+	if (DISPLAY_VER(dev_priv) == 11) {
+		if (afe_clk(encoder, crtc_state) <= 800000) {
+			for_each_dsi_port(port, intel_dsi->ports) {
+				tmp = intel_de_read(dev_priv,
+						    DPHY_TA_TIMING_PARAM(port));
+				tmp &= ~TA_SURE_MASK;
+				tmp |= TA_SURE_OVERRIDE | TA_SURE(0);
+				intel_de_write(dev_priv,
+					       DPHY_TA_TIMING_PARAM(port),
+					       tmp);
+
+				/* shadow register inside display core */
+				tmp = intel_de_read(dev_priv,
+						    DSI_TA_TIMING_PARAM(port));
+				tmp &= ~TA_SURE_MASK;
+				tmp |= TA_SURE_OVERRIDE | TA_SURE(0);
+				intel_de_write(dev_priv,
+					       DSI_TA_TIMING_PARAM(port), tmp);
+			}
+		}
+	}
+
+	if (IS_JSL_EHL(dev_priv)) {
+		for_each_dsi_phy(phy, intel_dsi->phys) {
+			tmp = intel_de_read(dev_priv, ICL_DPHY_CHKN(phy));
+			tmp |= ICL_DPHY_CHKN_AFE_OVER_PPI_STRAP;
+			intel_de_write(dev_priv, ICL_DPHY_CHKN(phy), tmp);
+		}
+	}
+}
+
+static void gen11_dsi_gate_clocks(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	u32 tmp;
+	enum phy phy;
+
+	mutex_lock(&dev_priv->display.dpll.lock);
+	tmp = intel_de_read(dev_priv, ICL_DPCLKA_CFGCR0);
+	for_each_dsi_phy(phy, intel_dsi->phys)
+		tmp |= ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy);
+
+	intel_de_write(dev_priv, ICL_DPCLKA_CFGCR0, tmp);
+	mutex_unlock(&dev_priv->display.dpll.lock);
+}
+
+static void gen11_dsi_ungate_clocks(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	u32 tmp;
+	enum phy phy;
+
+	mutex_lock(&dev_priv->display.dpll.lock);
+	tmp = intel_de_read(dev_priv, ICL_DPCLKA_CFGCR0);
+	for_each_dsi_phy(phy, intel_dsi->phys)
+		tmp &= ~ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy);
+
+	intel_de_write(dev_priv, ICL_DPCLKA_CFGCR0, tmp);
+	mutex_unlock(&dev_priv->display.dpll.lock);
+}
+
+static bool gen11_dsi_is_clock_enabled(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	bool clock_enabled = false;
+	enum phy phy;
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, ICL_DPCLKA_CFGCR0);
+
+	for_each_dsi_phy(phy, intel_dsi->phys) {
+		if (!(tmp & ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy)))
+			clock_enabled = true;
+	}
+
+	return clock_enabled;
+}
+
+static void gen11_dsi_map_pll(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	enum phy phy;
+	u32 val;
+
+	mutex_lock(&dev_priv->display.dpll.lock);
+
+	val = intel_de_read(dev_priv, ICL_DPCLKA_CFGCR0);
+	for_each_dsi_phy(phy, intel_dsi->phys) {
+		val &= ~ICL_DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(phy);
+		val |= ICL_DPCLKA_CFGCR0_DDI_CLK_SEL(pll->info->id, phy);
+	}
+	intel_de_write(dev_priv, ICL_DPCLKA_CFGCR0, val);
+
+	for_each_dsi_phy(phy, intel_dsi->phys) {
+		val &= ~ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy);
+	}
+	intel_de_write(dev_priv, ICL_DPCLKA_CFGCR0, val);
+
+	intel_de_posting_read(dev_priv, ICL_DPCLKA_CFGCR0);
+
+	mutex_unlock(&dev_priv->display.dpll.lock);
+}
+
+static void
+gen11_dsi_configure_transcoder(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	enum pipe pipe = crtc->pipe;
+	u32 tmp;
+	enum port port;
+	enum transcoder dsi_trans;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		tmp = intel_de_read(dev_priv, DSI_TRANS_FUNC_CONF(dsi_trans));
+
+		if (intel_dsi->eotp_pkt)
+			tmp &= ~EOTP_DISABLED;
+		else
+			tmp |= EOTP_DISABLED;
+
+		/* enable link calibration if freq > 1.5Gbps */
+		if (afe_clk(encoder, pipe_config) >= 1500 * 1000) {
+			tmp &= ~LINK_CALIBRATION_MASK;
+			tmp |= CALIBRATION_ENABLED_INITIAL_ONLY;
+		}
+
+		/* configure continuous clock */
+		tmp &= ~CONTINUOUS_CLK_MASK;
+		if (intel_dsi->clock_stop)
+			tmp |= CLK_ENTER_LP_AFTER_DATA;
+		else
+			tmp |= CLK_HS_CONTINUOUS;
+
+		/* configure buffer threshold limit to minimum */
+		tmp &= ~PIX_BUF_THRESHOLD_MASK;
+		tmp |= PIX_BUF_THRESHOLD_1_4;
+
+		/* set virtual channel to '0' */
+		tmp &= ~PIX_VIRT_CHAN_MASK;
+		tmp |= PIX_VIRT_CHAN(0);
+
+		/* program BGR transmission */
+		if (intel_dsi->bgr_enabled)
+			tmp |= BGR_TRANSMISSION;
+
+		/* select pixel format */
+		tmp &= ~PIX_FMT_MASK;
+		if (pipe_config->dsc.compression_enable) {
+			tmp |= PIX_FMT_COMPRESSED;
+		} else {
+			switch (intel_dsi->pixel_format) {
+			default:
+				MISSING_CASE(intel_dsi->pixel_format);
+				fallthrough;
+			case MIPI_DSI_FMT_RGB565:
+				tmp |= PIX_FMT_RGB565;
+				break;
+			case MIPI_DSI_FMT_RGB666_PACKED:
+				tmp |= PIX_FMT_RGB666_PACKED;
+				break;
+			case MIPI_DSI_FMT_RGB666:
+				tmp |= PIX_FMT_RGB666_LOOSE;
+				break;
+			case MIPI_DSI_FMT_RGB888:
+				tmp |= PIX_FMT_RGB888;
+				break;
+			}
+		}
+
+		if (DISPLAY_VER(dev_priv) >= 12) {
+			if (is_vid_mode(intel_dsi))
+				tmp |= BLANKING_PACKET_ENABLE;
+		}
+
+		/* program DSI operation mode */
+		if (is_vid_mode(intel_dsi)) {
+			tmp &= ~OP_MODE_MASK;
+			switch (intel_dsi->video_mode) {
+			default:
+				MISSING_CASE(intel_dsi->video_mode);
+				fallthrough;
+			case NON_BURST_SYNC_EVENTS:
+				tmp |= VIDEO_MODE_SYNC_EVENT;
+				break;
+			case NON_BURST_SYNC_PULSE:
+				tmp |= VIDEO_MODE_SYNC_PULSE;
+				break;
+			}
+		} else {
+			/*
+			 * FIXME: Retrieve this info from VBT.
+			 * As per the spec when dsi transcoder is operating
+			 * in TE GATE mode, TE comes from GPIO
+			 * which is UTIL PIN for DSI 0.
+			 * Also this GPIO would not be used for other
+			 * purposes is an assumption.
+			 */
+			tmp &= ~OP_MODE_MASK;
+			tmp |= CMD_MODE_TE_GATE;
+			tmp |= TE_SOURCE_GPIO;
+		}
+
+		intel_de_write(dev_priv, DSI_TRANS_FUNC_CONF(dsi_trans), tmp);
+	}
+
+	/* enable port sync mode if dual link */
+	if (intel_dsi->dual_link) {
+		for_each_dsi_port(port, intel_dsi->ports) {
+			dsi_trans = dsi_port_to_transcoder(port);
+			tmp = intel_de_read(dev_priv,
+					    TRANS_DDI_FUNC_CTL2(dsi_trans));
+			tmp |= PORT_SYNC_MODE_ENABLE;
+			intel_de_write(dev_priv,
+				       TRANS_DDI_FUNC_CTL2(dsi_trans), tmp);
+		}
+
+		/* configure stream splitting */
+		configure_dual_link_mode(encoder, pipe_config);
+	}
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+
+		/* select data lane width */
+		tmp = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(dsi_trans));
+		tmp &= ~DDI_PORT_WIDTH_MASK;
+		tmp |= DDI_PORT_WIDTH(intel_dsi->lane_count);
+
+		/* select input pipe */
+		tmp &= ~TRANS_DDI_EDP_INPUT_MASK;
+		switch (pipe) {
+		default:
+			MISSING_CASE(pipe);
+			fallthrough;
+		case PIPE_A:
+			tmp |= TRANS_DDI_EDP_INPUT_A_ON;
+			break;
+		case PIPE_B:
+			tmp |= TRANS_DDI_EDP_INPUT_B_ONOFF;
+			break;
+		case PIPE_C:
+			tmp |= TRANS_DDI_EDP_INPUT_C_ONOFF;
+			break;
+		case PIPE_D:
+			tmp |= TRANS_DDI_EDP_INPUT_D_ONOFF;
+			break;
+		}
+
+		/* enable DDI buffer */
+		tmp |= TRANS_DDI_FUNC_ENABLE;
+		intel_de_write(dev_priv, TRANS_DDI_FUNC_CTL(dsi_trans), tmp);
+	}
+
+	/* wait for link ready */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		if (wait_for_us((intel_de_read(dev_priv, DSI_TRANS_FUNC_CONF(dsi_trans)) &
+				 LINK_READY), 2500))
+			drm_err(&dev_priv->drm, "DSI link not ready\n");
+	}
+}
+
+static void
+gen11_dsi_set_transcoder_timings(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	enum port port;
+	enum transcoder dsi_trans;
+	/* horizontal timings */
+	u16 htotal, hactive, hsync_start, hsync_end, hsync_size;
+	u16 hback_porch;
+	/* vertical timings */
+	u16 vtotal, vactive, vsync_start, vsync_end, vsync_shift;
+	int mul = 1, div = 1;
+
+	/*
+	 * Adjust horizontal timings (htotal, hsync_start, hsync_end) to account
+	 * for slower link speed if DSC is enabled.
+	 *
+	 * The compression frequency ratio is the ratio between compressed and
+	 * non-compressed link speeds, and simplifies down to the ratio between
+	 * compressed and non-compressed bpp.
+	 */
+	if (crtc_state->dsc.compression_enable) {
+		mul = crtc_state->dsc.compressed_bpp;
+		div = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format);
+	}
+
+	hactive = adjusted_mode->crtc_hdisplay;
+
+	if (is_vid_mode(intel_dsi))
+		htotal = DIV_ROUND_UP(adjusted_mode->crtc_htotal * mul, div);
+	else
+		htotal = DIV_ROUND_UP((hactive + 160) * mul, div);
+
+	hsync_start = DIV_ROUND_UP(adjusted_mode->crtc_hsync_start * mul, div);
+	hsync_end = DIV_ROUND_UP(adjusted_mode->crtc_hsync_end * mul, div);
+	hsync_size  = hsync_end - hsync_start;
+	hback_porch = (adjusted_mode->crtc_htotal -
+		       adjusted_mode->crtc_hsync_end);
+	vactive = adjusted_mode->crtc_vdisplay;
+
+	if (is_vid_mode(intel_dsi)) {
+		vtotal = adjusted_mode->crtc_vtotal;
+	} else {
+		int bpp, line_time_us, byte_clk_period_ns;
+
+		if (crtc_state->dsc.compression_enable)
+			bpp = crtc_state->dsc.compressed_bpp;
+		else
+			bpp = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format);
+
+		byte_clk_period_ns = 1000000 / afe_clk(encoder, crtc_state);
+		line_time_us = (htotal * (bpp / 8) * byte_clk_period_ns) / (1000 * intel_dsi->lane_count);
+		vtotal = vactive + DIV_ROUND_UP(400, line_time_us);
+	}
+	vsync_start = adjusted_mode->crtc_vsync_start;
+	vsync_end = adjusted_mode->crtc_vsync_end;
+	vsync_shift = hsync_start - htotal / 2;
+
+	if (intel_dsi->dual_link) {
+		hactive /= 2;
+		if (intel_dsi->dual_link == DSI_DUAL_LINK_FRONT_BACK)
+			hactive += intel_dsi->pixel_overlap;
+		htotal /= 2;
+	}
+
+	/* minimum hactive as per bspec: 256 pixels */
+	if (adjusted_mode->crtc_hdisplay < 256)
+		drm_err(&dev_priv->drm, "hactive is less then 256 pixels\n");
+
+	/* if RGB666 format, then hactive must be multiple of 4 pixels */
+	if (intel_dsi->pixel_format == MIPI_DSI_FMT_RGB666 && hactive % 4 != 0)
+		drm_err(&dev_priv->drm,
+			"hactive pixels are not multiple of 4\n");
+
+	/* program TRANS_HTOTAL register */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		intel_de_write(dev_priv, HTOTAL(dsi_trans),
+			       (hactive - 1) | ((htotal - 1) << 16));
+	}
+
+	/* TRANS_HSYNC register to be programmed only for video mode */
+	if (is_vid_mode(intel_dsi)) {
+		if (intel_dsi->video_mode == NON_BURST_SYNC_PULSE) {
+			/* BSPEC: hsync size should be atleast 16 pixels */
+			if (hsync_size < 16)
+				drm_err(&dev_priv->drm,
+					"hsync size < 16 pixels\n");
+		}
+
+		if (hback_porch < 16)
+			drm_err(&dev_priv->drm, "hback porch < 16 pixels\n");
+
+		if (intel_dsi->dual_link) {
+			hsync_start /= 2;
+			hsync_end /= 2;
+		}
+
+		for_each_dsi_port(port, intel_dsi->ports) {
+			dsi_trans = dsi_port_to_transcoder(port);
+			intel_de_write(dev_priv, HSYNC(dsi_trans),
+				       (hsync_start - 1) | ((hsync_end - 1) << 16));
+		}
+	}
+
+	/* program TRANS_VTOTAL register */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		/*
+		 * FIXME: Programing this by assuming progressive mode, since
+		 * non-interlaced info from VBT is not saved inside
+		 * struct drm_display_mode.
+		 * For interlace mode: program required pixel minus 2
+		 */
+		intel_de_write(dev_priv, VTOTAL(dsi_trans),
+			       (vactive - 1) | ((vtotal - 1) << 16));
+	}
+
+	if (vsync_end < vsync_start || vsync_end > vtotal)
+		drm_err(&dev_priv->drm, "Invalid vsync_end value\n");
+
+	if (vsync_start < vactive)
+		drm_err(&dev_priv->drm, "vsync_start less than vactive\n");
+
+	/* program TRANS_VSYNC register for video mode only */
+	if (is_vid_mode(intel_dsi)) {
+		for_each_dsi_port(port, intel_dsi->ports) {
+			dsi_trans = dsi_port_to_transcoder(port);
+			intel_de_write(dev_priv, VSYNC(dsi_trans),
+				       (vsync_start - 1) | ((vsync_end - 1) << 16));
+		}
+	}
+
+	/*
+	 * FIXME: It has to be programmed only for video modes and interlaced
+	 * modes. Put the check condition here once interlaced
+	 * info available as described above.
+	 * program TRANS_VSYNCSHIFT register
+	 */
+	if (is_vid_mode(intel_dsi)) {
+		for_each_dsi_port(port, intel_dsi->ports) {
+			dsi_trans = dsi_port_to_transcoder(port);
+			intel_de_write(dev_priv, VSYNCSHIFT(dsi_trans),
+				       vsync_shift);
+		}
+	}
+
+	/* program TRANS_VBLANK register, should be same as vtotal programmed */
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		for_each_dsi_port(port, intel_dsi->ports) {
+			dsi_trans = dsi_port_to_transcoder(port);
+			intel_de_write(dev_priv, VBLANK(dsi_trans),
+				       (vactive - 1) | ((vtotal - 1) << 16));
+		}
+	}
+}
+
+static void gen11_dsi_enable_transcoder(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	enum transcoder dsi_trans;
+	u32 tmp;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		tmp = intel_de_read(dev_priv, PIPECONF(dsi_trans));
+		tmp |= PIPECONF_ENABLE;
+		intel_de_write(dev_priv, PIPECONF(dsi_trans), tmp);
+
+		/* wait for transcoder to be enabled */
+		if (intel_de_wait_for_set(dev_priv, PIPECONF(dsi_trans),
+					  PIPECONF_STATE_ENABLE, 10))
+			drm_err(&dev_priv->drm,
+				"DSI transcoder not enabled\n");
+	}
+}
+
+static void gen11_dsi_setup_timeouts(struct intel_encoder *encoder,
+				     const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	enum transcoder dsi_trans;
+	u32 tmp, hs_tx_timeout, lp_rx_timeout, ta_timeout, divisor, mul;
+
+	/*
+	 * escape clock count calculation:
+	 * BYTE_CLK_COUNT = TIME_NS/(8 * UI)
+	 * UI (nsec) = (10^6)/Bitrate
+	 * TIME_NS = (BYTE_CLK_COUNT * 8 * 10^6)/ Bitrate
+	 * ESCAPE_CLK_COUNT  = TIME_NS/ESC_CLK_NS
+	 */
+	divisor = intel_dsi_tlpx_ns(intel_dsi) * afe_clk(encoder, crtc_state) * 1000;
+	mul = 8 * 1000000;
+	hs_tx_timeout = DIV_ROUND_UP(intel_dsi->hs_tx_timeout * mul,
+				     divisor);
+	lp_rx_timeout = DIV_ROUND_UP(intel_dsi->lp_rx_timeout * mul, divisor);
+	ta_timeout = DIV_ROUND_UP(intel_dsi->turn_arnd_val * mul, divisor);
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+
+		/* program hst_tx_timeout */
+		tmp = intel_de_read(dev_priv, DSI_HSTX_TO(dsi_trans));
+		tmp &= ~HSTX_TIMEOUT_VALUE_MASK;
+		tmp |= HSTX_TIMEOUT_VALUE(hs_tx_timeout);
+		intel_de_write(dev_priv, DSI_HSTX_TO(dsi_trans), tmp);
+
+		/* FIXME: DSI_CALIB_TO */
+
+		/* program lp_rx_host timeout */
+		tmp = intel_de_read(dev_priv, DSI_LPRX_HOST_TO(dsi_trans));
+		tmp &= ~LPRX_TIMEOUT_VALUE_MASK;
+		tmp |= LPRX_TIMEOUT_VALUE(lp_rx_timeout);
+		intel_de_write(dev_priv, DSI_LPRX_HOST_TO(dsi_trans), tmp);
+
+		/* FIXME: DSI_PWAIT_TO */
+
+		/* program turn around timeout */
+		tmp = intel_de_read(dev_priv, DSI_TA_TO(dsi_trans));
+		tmp &= ~TA_TIMEOUT_VALUE_MASK;
+		tmp |= TA_TIMEOUT_VALUE(ta_timeout);
+		intel_de_write(dev_priv, DSI_TA_TO(dsi_trans), tmp);
+	}
+}
+
+static void gen11_dsi_config_util_pin(struct intel_encoder *encoder,
+				      bool enable)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	u32 tmp;
+
+	/*
+	 * used as TE i/p for DSI0,
+	 * for dual link/DSI1 TE is from slave DSI1
+	 * through GPIO.
+	 */
+	if (is_vid_mode(intel_dsi) || (intel_dsi->ports & BIT(PORT_B)))
+		return;
+
+	tmp = intel_de_read(dev_priv, UTIL_PIN_CTL);
+
+	if (enable) {
+		tmp |= UTIL_PIN_DIRECTION_INPUT;
+		tmp |= UTIL_PIN_ENABLE;
+	} else {
+		tmp &= ~UTIL_PIN_ENABLE;
+	}
+	intel_de_write(dev_priv, UTIL_PIN_CTL, tmp);
+}
+
+static void
+gen11_dsi_enable_port_and_phy(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state)
+{
+	/* step 4a: power up all lanes of the DDI used by DSI */
+	gen11_dsi_power_up_lanes(encoder);
+
+	/* step 4b: configure lane sequencing of the Combo-PHY transmitters */
+	gen11_dsi_config_phy_lanes_sequence(encoder);
+
+	/* step 4c: configure voltage swing and skew */
+	gen11_dsi_voltage_swing_program_seq(encoder);
+
+	/* enable DDI buffer */
+	gen11_dsi_enable_ddi_buffer(encoder);
+
+	/* setup D-PHY timings */
+	gen11_dsi_setup_dphy_timings(encoder, crtc_state);
+
+	/* Since transcoder is configured to take events from GPIO */
+	gen11_dsi_config_util_pin(encoder, true);
+
+	/* step 4h: setup DSI protocol timeouts */
+	gen11_dsi_setup_timeouts(encoder, crtc_state);
+
+	/* Step (4h, 4i, 4j, 4k): Configure transcoder */
+	gen11_dsi_configure_transcoder(encoder, crtc_state);
+
+	/* Step 4l: Gate DDI clocks */
+	gen11_dsi_gate_clocks(encoder);
+}
+
+static void gen11_dsi_powerup_panel(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	struct mipi_dsi_device *dsi;
+	enum port port;
+	enum transcoder dsi_trans;
+	u32 tmp;
+	int ret;
+
+	/* set maximum return packet size */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+
+		/*
+		 * FIXME: This uses the number of DW's currently in the payload
+		 * receive queue. This is probably not what we want here.
+		 */
+		tmp = intel_de_read(dev_priv, DSI_CMD_RXCTL(dsi_trans));
+		tmp &= NUMBER_RX_PLOAD_DW_MASK;
+		/* multiply "Number Rx Payload DW" by 4 to get max value */
+		tmp = tmp * 4;
+		dsi = intel_dsi->dsi_hosts[port]->device;
+		ret = mipi_dsi_set_maximum_return_packet_size(dsi, tmp);
+		if (ret < 0)
+			drm_err(&dev_priv->drm,
+				"error setting max return pkt size%d\n", tmp);
+	}
+
+	/* panel power on related mipi dsi vbt sequences */
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_POWER_ON);
+	msleep(intel_dsi->panel_on_delay);
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DEASSERT_RESET);
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_INIT_OTP);
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DISPLAY_ON);
+
+	/* ensure all panel commands dispatched before enabling transcoder */
+	wait_for_cmds_dispatched_to_panel(encoder);
+}
+
+static void gen11_dsi_pre_pll_enable(struct intel_atomic_state *state,
+				     struct intel_encoder *encoder,
+				     const struct intel_crtc_state *crtc_state,
+				     const struct drm_connector_state *conn_state)
+{
+	/* step2: enable IO power */
+	gen11_dsi_enable_io_power(encoder);
+
+	/* step3: enable DSI PLL */
+	gen11_dsi_program_esc_clk_div(encoder, crtc_state);
+}
+
+static void gen11_dsi_pre_enable(struct intel_atomic_state *state,
+				 struct intel_encoder *encoder,
+				 const struct intel_crtc_state *pipe_config,
+				 const struct drm_connector_state *conn_state)
+{
+	/* step3b */
+	gen11_dsi_map_pll(encoder, pipe_config);
+
+	/* step4: enable DSI port and DPHY */
+	gen11_dsi_enable_port_and_phy(encoder, pipe_config);
+
+	/* step5: program and powerup panel */
+	gen11_dsi_powerup_panel(encoder);
+
+	intel_dsc_dsi_pps_write(encoder, pipe_config);
+
+	/* step6c: configure transcoder timings */
+	gen11_dsi_set_transcoder_timings(encoder, pipe_config);
+}
+
+/*
+ * Wa_1409054076:icl,jsl,ehl
+ * When pipe A is disabled and MIPI DSI is enabled on pipe B,
+ * the AMT KVMR feature will incorrectly see pipe A as enabled.
+ * Set 0x42080 bit 23=1 before enabling DSI on pipe B and leave
+ * it set while DSI is enabled on pipe B
+ */
+static void icl_apply_kvmr_pipe_a_wa(struct intel_encoder *encoder,
+				     enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (DISPLAY_VER(dev_priv) == 11 && pipe == PIPE_B)
+		intel_de_rmw(dev_priv, CHICKEN_PAR1_1,
+			     IGNORE_KVMR_PIPE_A,
+			     enable ? IGNORE_KVMR_PIPE_A : 0);
+}
+
+/*
+ * Wa_16012360555:adl-p
+ * SW will have to program the "LP to HS Wakeup Guardband"
+ * to account for the repeaters on the HS Request/Ready
+ * PPI signaling between the Display engine and the DPHY.
+ */
+static void adlp_set_lp_hs_wakeup_gb(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+
+	if (DISPLAY_VER(i915) == 13) {
+		for_each_dsi_port(port, intel_dsi->ports)
+			intel_de_rmw(i915, TGL_DSI_CHKN_REG(port),
+				     TGL_DSI_CHKN_LSHS_GB_MASK,
+				     TGL_DSI_CHKN_LSHS_GB(4));
+	}
+}
+
+static void gen11_dsi_enable(struct intel_atomic_state *state,
+			     struct intel_encoder *encoder,
+			     const struct intel_crtc_state *crtc_state,
+			     const struct drm_connector_state *conn_state)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(conn_state->crtc);
+
+	drm_WARN_ON(state->base.dev, crtc_state->has_pch_encoder);
+
+	/* Wa_1409054076:icl,jsl,ehl */
+	icl_apply_kvmr_pipe_a_wa(encoder, crtc->pipe, true);
+
+	/* Wa_16012360555:adl-p */
+	adlp_set_lp_hs_wakeup_gb(encoder);
+
+	/* step6d: enable dsi transcoder */
+	gen11_dsi_enable_transcoder(encoder);
+
+	/* step7: enable backlight */
+	intel_backlight_enable(crtc_state, conn_state);
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_BACKLIGHT_ON);
+
+	intel_crtc_vblank_on(crtc_state);
+}
+
+static void gen11_dsi_disable_transcoder(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	enum transcoder dsi_trans;
+	u32 tmp;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+
+		/* disable transcoder */
+		tmp = intel_de_read(dev_priv, PIPECONF(dsi_trans));
+		tmp &= ~PIPECONF_ENABLE;
+		intel_de_write(dev_priv, PIPECONF(dsi_trans), tmp);
+
+		/* wait for transcoder to be disabled */
+		if (intel_de_wait_for_clear(dev_priv, PIPECONF(dsi_trans),
+					    PIPECONF_STATE_ENABLE, 50))
+			drm_err(&dev_priv->drm,
+				"DSI trancoder not disabled\n");
+	}
+}
+
+static void gen11_dsi_powerdown_panel(struct intel_encoder *encoder)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DISPLAY_OFF);
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_ASSERT_RESET);
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_POWER_OFF);
+
+	/* ensure cmds dispatched to panel */
+	wait_for_cmds_dispatched_to_panel(encoder);
+}
+
+static void gen11_dsi_deconfigure_trancoder(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	enum transcoder dsi_trans;
+	u32 tmp;
+
+	/* disable periodic update mode */
+	if (is_cmd_mode(intel_dsi)) {
+		for_each_dsi_port(port, intel_dsi->ports) {
+			tmp = intel_de_read(dev_priv, DSI_CMD_FRMCTL(port));
+			tmp &= ~DSI_PERIODIC_FRAME_UPDATE_ENABLE;
+			intel_de_write(dev_priv, DSI_CMD_FRMCTL(port), tmp);
+		}
+	}
+
+	/* put dsi link in ULPS */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		tmp = intel_de_read(dev_priv, DSI_LP_MSG(dsi_trans));
+		tmp |= LINK_ENTER_ULPS;
+		tmp &= ~LINK_ULPS_TYPE_LP11;
+		intel_de_write(dev_priv, DSI_LP_MSG(dsi_trans), tmp);
+
+		if (wait_for_us((intel_de_read(dev_priv, DSI_LP_MSG(dsi_trans)) &
+				 LINK_IN_ULPS),
+				10))
+			drm_err(&dev_priv->drm, "DSI link not in ULPS\n");
+	}
+
+	/* disable ddi function */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		tmp = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(dsi_trans));
+		tmp &= ~TRANS_DDI_FUNC_ENABLE;
+		intel_de_write(dev_priv, TRANS_DDI_FUNC_CTL(dsi_trans), tmp);
+	}
+
+	/* disable port sync mode if dual link */
+	if (intel_dsi->dual_link) {
+		for_each_dsi_port(port, intel_dsi->ports) {
+			dsi_trans = dsi_port_to_transcoder(port);
+			tmp = intel_de_read(dev_priv,
+					    TRANS_DDI_FUNC_CTL2(dsi_trans));
+			tmp &= ~PORT_SYNC_MODE_ENABLE;
+			intel_de_write(dev_priv,
+				       TRANS_DDI_FUNC_CTL2(dsi_trans), tmp);
+		}
+	}
+}
+
+static void gen11_dsi_disable_port(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	u32 tmp;
+	enum port port;
+
+	gen11_dsi_ungate_clocks(encoder);
+	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = intel_de_read(dev_priv, DDI_BUF_CTL(port));
+		tmp &= ~DDI_BUF_CTL_ENABLE;
+		intel_de_write(dev_priv, DDI_BUF_CTL(port), tmp);
+
+		if (wait_for_us((intel_de_read(dev_priv, DDI_BUF_CTL(port)) &
+				 DDI_BUF_IS_IDLE),
+				 8))
+			drm_err(&dev_priv->drm,
+				"DDI port:%c buffer not idle\n",
+				port_name(port));
+	}
+	gen11_dsi_gate_clocks(encoder);
+}
+
+static void gen11_dsi_disable_io_power(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 tmp;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		intel_wakeref_t wakeref;
+
+		wakeref = fetch_and_zero(&intel_dsi->io_wakeref[port]);
+		intel_display_power_put(dev_priv,
+					port == PORT_A ?
+					POWER_DOMAIN_PORT_DDI_IO_A :
+					POWER_DOMAIN_PORT_DDI_IO_B,
+					wakeref);
+	}
+
+	/* set mode to DDI */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = intel_de_read(dev_priv, ICL_DSI_IO_MODECTL(port));
+		tmp &= ~COMBO_PHY_MODE_DSI;
+		intel_de_write(dev_priv, ICL_DSI_IO_MODECTL(port), tmp);
+	}
+}
+
+static void gen11_dsi_disable(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *old_crtc_state,
+			      const struct drm_connector_state *old_conn_state)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(old_conn_state->crtc);
+
+	/* step1: turn off backlight */
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_BACKLIGHT_OFF);
+	intel_backlight_disable(old_conn_state);
+
+	/* step2d,e: disable transcoder and wait */
+	gen11_dsi_disable_transcoder(encoder);
+
+	/* Wa_1409054076:icl,jsl,ehl */
+	icl_apply_kvmr_pipe_a_wa(encoder, crtc->pipe, false);
+
+	/* step2f,g: powerdown panel */
+	gen11_dsi_powerdown_panel(encoder);
+
+	/* step2h,i,j: deconfig trancoder */
+	gen11_dsi_deconfigure_trancoder(encoder);
+
+	/* step3: disable port */
+	gen11_dsi_disable_port(encoder);
+
+	gen11_dsi_config_util_pin(encoder, false);
+
+	/* step4: disable IO power */
+	gen11_dsi_disable_io_power(encoder);
+}
+
+static void gen11_dsi_post_disable(struct intel_atomic_state *state,
+				   struct intel_encoder *encoder,
+				   const struct intel_crtc_state *old_crtc_state,
+				   const struct drm_connector_state *old_conn_state)
+{
+	intel_crtc_vblank_off(old_crtc_state);
+
+	intel_dsc_disable(old_crtc_state);
+
+	skl_scaler_disable(old_crtc_state);
+}
+
+static enum drm_mode_status gen11_dsi_mode_valid(struct drm_connector *connector,
+						 struct drm_display_mode *mode)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	enum drm_mode_status status;
+
+	status = intel_cpu_transcoder_mode_valid(i915, mode);
+	if (status != MODE_OK)
+		return status;
+
+	/* FIXME: DSC? */
+	return intel_dsi_mode_valid(connector, mode);
+}
+
+static void gen11_dsi_get_timings(struct intel_encoder *encoder,
+				  struct intel_crtc_state *pipe_config)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	struct drm_display_mode *adjusted_mode =
+					&pipe_config->hw.adjusted_mode;
+
+	if (pipe_config->dsc.compressed_bpp) {
+		int div = pipe_config->dsc.compressed_bpp;
+		int mul = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format);
+
+		adjusted_mode->crtc_htotal =
+			DIV_ROUND_UP(adjusted_mode->crtc_htotal * mul, div);
+		adjusted_mode->crtc_hsync_start =
+			DIV_ROUND_UP(adjusted_mode->crtc_hsync_start * mul, div);
+		adjusted_mode->crtc_hsync_end =
+			DIV_ROUND_UP(adjusted_mode->crtc_hsync_end * mul, div);
+	}
+
+	if (intel_dsi->dual_link) {
+		adjusted_mode->crtc_hdisplay *= 2;
+		if (intel_dsi->dual_link == DSI_DUAL_LINK_FRONT_BACK)
+			adjusted_mode->crtc_hdisplay -=
+						intel_dsi->pixel_overlap;
+		adjusted_mode->crtc_htotal *= 2;
+	}
+	adjusted_mode->crtc_hblank_start = adjusted_mode->crtc_hdisplay;
+	adjusted_mode->crtc_hblank_end = adjusted_mode->crtc_htotal;
+
+	if (intel_dsi->operation_mode == INTEL_DSI_VIDEO_MODE) {
+		if (intel_dsi->dual_link) {
+			adjusted_mode->crtc_hsync_start *= 2;
+			adjusted_mode->crtc_hsync_end *= 2;
+		}
+	}
+	adjusted_mode->crtc_vblank_start = adjusted_mode->crtc_vdisplay;
+	adjusted_mode->crtc_vblank_end = adjusted_mode->crtc_vtotal;
+}
+
+static bool gen11_dsi_is_periodic_cmd_mode(struct intel_dsi *intel_dsi)
+{
+	struct drm_device *dev = intel_dsi->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum transcoder dsi_trans;
+	u32 val;
+
+	if (intel_dsi->ports == BIT(PORT_B))
+		dsi_trans = TRANSCODER_DSI_1;
+	else
+		dsi_trans = TRANSCODER_DSI_0;
+
+	val = intel_de_read(dev_priv, DSI_TRANS_FUNC_CONF(dsi_trans));
+	return (val & DSI_PERIODIC_FRAME_UPDATE_ENABLE);
+}
+
+static void gen11_dsi_get_cmd_mode_config(struct intel_dsi *intel_dsi,
+					  struct intel_crtc_state *pipe_config)
+{
+	if (intel_dsi->ports == (BIT(PORT_B) | BIT(PORT_A)))
+		pipe_config->mode_flags |= I915_MODE_FLAG_DSI_USE_TE1 |
+					    I915_MODE_FLAG_DSI_USE_TE0;
+	else if (intel_dsi->ports == BIT(PORT_B))
+		pipe_config->mode_flags |= I915_MODE_FLAG_DSI_USE_TE1;
+	else
+		pipe_config->mode_flags |= I915_MODE_FLAG_DSI_USE_TE0;
+}
+
+static void gen11_dsi_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_state *pipe_config)
+{
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+
+	intel_ddi_get_clock(encoder, pipe_config, icl_ddi_combo_get_pll(encoder));
+
+	pipe_config->hw.adjusted_mode.crtc_clock = intel_dsi->pclk;
+	if (intel_dsi->dual_link)
+		pipe_config->hw.adjusted_mode.crtc_clock *= 2;
+
+	gen11_dsi_get_timings(encoder, pipe_config);
+	pipe_config->output_types |= BIT(INTEL_OUTPUT_DSI);
+	pipe_config->pipe_bpp = bdw_get_pipemisc_bpp(crtc);
+
+	/* Get the details on which TE should be enabled */
+	if (is_cmd_mode(intel_dsi))
+		gen11_dsi_get_cmd_mode_config(intel_dsi, pipe_config);
+
+	if (gen11_dsi_is_periodic_cmd_mode(intel_dsi))
+		pipe_config->mode_flags |= I915_MODE_FLAG_DSI_PERIODIC_CMD_MODE;
+}
+
+static void gen11_dsi_sync_state(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *intel_crtc;
+	enum pipe pipe;
+
+	if (!crtc_state)
+		return;
+
+	intel_crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	pipe = intel_crtc->pipe;
+
+	/* wa verify 1409054076:icl,jsl,ehl */
+	if (DISPLAY_VER(dev_priv) == 11 && pipe == PIPE_B &&
+	    !(intel_de_read(dev_priv, CHICKEN_PAR1_1) & IGNORE_KVMR_PIPE_A))
+		drm_dbg_kms(&dev_priv->drm,
+			    "[ENCODER:%d:%s] BIOS left IGNORE_KVMR_PIPE_A cleared with pipe B enabled\n",
+			    encoder->base.base.id,
+			    encoder->base.name);
+}
+
+static int gen11_dsi_dsc_compute_config(struct intel_encoder *encoder,
+					struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct drm_dsc_config *vdsc_cfg = &crtc_state->dsc.config;
+	int dsc_max_bpc = DISPLAY_VER(dev_priv) >= 12 ? 12 : 10;
+	bool use_dsc;
+	int ret;
+
+	use_dsc = intel_bios_get_dsc_params(encoder, crtc_state, dsc_max_bpc);
+	if (!use_dsc)
+		return 0;
+
+	if (crtc_state->pipe_bpp < 8 * 3)
+		return -EINVAL;
+
+	/* FIXME: split only when necessary */
+	if (crtc_state->dsc.slice_count > 1)
+		crtc_state->dsc.dsc_split = true;
+
+	vdsc_cfg->convert_rgb = true;
+
+	/* FIXME: initialize from VBT */
+	vdsc_cfg->rc_model_size = DSC_RC_MODEL_SIZE_CONST;
+
+	vdsc_cfg->pic_height = crtc_state->hw.adjusted_mode.crtc_vdisplay;
+
+	ret = intel_dsc_compute_params(crtc_state);
+	if (ret)
+		return ret;
+
+	/* DSI specific sanity checks on the common code */
+	drm_WARN_ON(&dev_priv->drm, vdsc_cfg->vbr_enable);
+	drm_WARN_ON(&dev_priv->drm, vdsc_cfg->simple_422);
+	drm_WARN_ON(&dev_priv->drm,
+		    vdsc_cfg->pic_width % vdsc_cfg->slice_width);
+	drm_WARN_ON(&dev_priv->drm, vdsc_cfg->slice_height < 8);
+	drm_WARN_ON(&dev_priv->drm,
+		    vdsc_cfg->pic_height % vdsc_cfg->slice_height);
+
+	ret = drm_dsc_compute_rc_parameters(vdsc_cfg);
+	if (ret)
+		return ret;
+
+	crtc_state->dsc.compression_enable = true;
+
+	return 0;
+}
+
+static int gen11_dsi_compute_config(struct intel_encoder *encoder,
+				    struct intel_crtc_state *pipe_config,
+				    struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = container_of(encoder, struct intel_dsi,
+						   base);
+	struct intel_connector *intel_connector = intel_dsi->attached_connector;
+	struct drm_display_mode *adjusted_mode =
+		&pipe_config->hw.adjusted_mode;
+	int ret;
+
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+
+	ret = intel_panel_compute_config(intel_connector, adjusted_mode);
+	if (ret)
+		return ret;
+
+	ret = intel_panel_fitting(pipe_config, conn_state);
+	if (ret)
+		return ret;
+
+	adjusted_mode->flags = 0;
+
+	/* Dual link goes to trancoder DSI'0' */
+	if (intel_dsi->ports == BIT(PORT_B))
+		pipe_config->cpu_transcoder = TRANSCODER_DSI_1;
+	else
+		pipe_config->cpu_transcoder = TRANSCODER_DSI_0;
+
+	if (intel_dsi->pixel_format == MIPI_DSI_FMT_RGB888)
+		pipe_config->pipe_bpp = 24;
+	else
+		pipe_config->pipe_bpp = 18;
+
+	pipe_config->clock_set = true;
+
+	if (gen11_dsi_dsc_compute_config(encoder, pipe_config))
+		drm_dbg_kms(&i915->drm, "Attempting to use DSC failed\n");
+
+	pipe_config->port_clock = afe_clk(encoder, pipe_config) / 5;
+
+	/*
+	 * In case of TE GATE cmd mode, we
+	 * receive TE from the slave if
+	 * dual link is enabled
+	 */
+	if (is_cmd_mode(intel_dsi))
+		gen11_dsi_get_cmd_mode_config(intel_dsi, pipe_config);
+
+	return 0;
+}
+
+static void gen11_dsi_get_power_domains(struct intel_encoder *encoder,
+					struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+
+	get_dsi_io_power_domains(i915,
+				 enc_to_intel_dsi(encoder));
+}
+
+static bool gen11_dsi_get_hw_state(struct intel_encoder *encoder,
+				   enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum transcoder dsi_trans;
+	intel_wakeref_t wakeref;
+	enum port port;
+	bool ret = false;
+	u32 tmp;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     encoder->power_domain);
+	if (!wakeref)
+		return false;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi_trans = dsi_port_to_transcoder(port);
+		tmp = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(dsi_trans));
+		switch (tmp & TRANS_DDI_EDP_INPUT_MASK) {
+		case TRANS_DDI_EDP_INPUT_A_ON:
+			*pipe = PIPE_A;
+			break;
+		case TRANS_DDI_EDP_INPUT_B_ONOFF:
+			*pipe = PIPE_B;
+			break;
+		case TRANS_DDI_EDP_INPUT_C_ONOFF:
+			*pipe = PIPE_C;
+			break;
+		case TRANS_DDI_EDP_INPUT_D_ONOFF:
+			*pipe = PIPE_D;
+			break;
+		default:
+			drm_err(&dev_priv->drm, "Invalid PIPE input\n");
+			goto out;
+		}
+
+		tmp = intel_de_read(dev_priv, PIPECONF(dsi_trans));
+		ret = tmp & PIPECONF_ENABLE;
+	}
+out:
+	intel_display_power_put(dev_priv, encoder->power_domain, wakeref);
+	return ret;
+}
+
+static bool gen11_dsi_initial_fastset_check(struct intel_encoder *encoder,
+					    struct intel_crtc_state *crtc_state)
+{
+	if (crtc_state->dsc.compression_enable) {
+		drm_dbg_kms(encoder->base.dev, "Forcing full modeset due to DSC being enabled\n");
+		crtc_state->uapi.mode_changed = true;
+
+		return false;
+	}
+
+	return true;
+}
+
+static void gen11_dsi_encoder_destroy(struct drm_encoder *encoder)
+{
+	intel_encoder_destroy(encoder);
+}
+
+static const struct drm_encoder_funcs gen11_dsi_encoder_funcs = {
+	.destroy = gen11_dsi_encoder_destroy,
+};
+
+static const struct drm_connector_funcs gen11_dsi_connector_funcs = {
+	.detect = intel_panel_detect,
+	.late_register = intel_connector_register,
+	.early_unregister = intel_connector_unregister,
+	.destroy = intel_connector_destroy,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.atomic_get_property = intel_digital_connector_atomic_get_property,
+	.atomic_set_property = intel_digital_connector_atomic_set_property,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.atomic_duplicate_state = intel_digital_connector_duplicate_state,
+};
+
+static const struct drm_connector_helper_funcs gen11_dsi_connector_helper_funcs = {
+	.get_modes = intel_dsi_get_modes,
+	.mode_valid = gen11_dsi_mode_valid,
+	.atomic_check = intel_digital_connector_atomic_check,
+};
+
+static int gen11_dsi_host_attach(struct mipi_dsi_host *host,
+				 struct mipi_dsi_device *dsi)
+{
+	return 0;
+}
+
+static int gen11_dsi_host_detach(struct mipi_dsi_host *host,
+				 struct mipi_dsi_device *dsi)
+{
+	return 0;
+}
+
+static ssize_t gen11_dsi_host_transfer(struct mipi_dsi_host *host,
+				       const struct mipi_dsi_msg *msg)
+{
+	struct intel_dsi_host *intel_dsi_host = to_intel_dsi_host(host);
+	struct mipi_dsi_packet dsi_pkt;
+	ssize_t ret;
+	bool enable_lpdt = false;
+
+	ret = mipi_dsi_create_packet(&dsi_pkt, msg);
+	if (ret < 0)
+		return ret;
+
+	if (msg->flags & MIPI_DSI_MSG_USE_LPM)
+		enable_lpdt = true;
+
+	/* only long packet contains payload */
+	if (mipi_dsi_packet_format_is_long(msg->type)) {
+		ret = dsi_send_pkt_payld(intel_dsi_host, &dsi_pkt);
+		if (ret < 0)
+			return ret;
+	}
+
+	/* send packet header */
+	ret  = dsi_send_pkt_hdr(intel_dsi_host, &dsi_pkt, enable_lpdt);
+	if (ret < 0)
+		return ret;
+
+	//TODO: add payload receive code if needed
+
+	ret = sizeof(dsi_pkt.header) + dsi_pkt.payload_length;
+
+	return ret;
+}
+
+static const struct mipi_dsi_host_ops gen11_dsi_host_ops = {
+	.attach = gen11_dsi_host_attach,
+	.detach = gen11_dsi_host_detach,
+	.transfer = gen11_dsi_host_transfer,
+};
+
+#define ICL_PREPARE_CNT_MAX	0x7
+#define ICL_CLK_ZERO_CNT_MAX	0xf
+#define ICL_TRAIL_CNT_MAX	0x7
+#define ICL_TCLK_PRE_CNT_MAX	0x3
+#define ICL_TCLK_POST_CNT_MAX	0x7
+#define ICL_HS_ZERO_CNT_MAX	0xf
+#define ICL_EXIT_ZERO_CNT_MAX	0x7
+
+static void icl_dphy_param_init(struct intel_dsi *intel_dsi)
+{
+	struct drm_device *dev = intel_dsi->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_connector *connector = intel_dsi->attached_connector;
+	struct mipi_config *mipi_config = connector->panel.vbt.dsi.config;
+	u32 tlpx_ns;
+	u32 prepare_cnt, exit_zero_cnt, clk_zero_cnt, trail_cnt;
+	u32 ths_prepare_ns, tclk_trail_ns;
+	u32 hs_zero_cnt;
+	u32 tclk_pre_cnt, tclk_post_cnt;
+
+	tlpx_ns = intel_dsi_tlpx_ns(intel_dsi);
+
+	tclk_trail_ns = max(mipi_config->tclk_trail, mipi_config->ths_trail);
+	ths_prepare_ns = max(mipi_config->ths_prepare,
+			     mipi_config->tclk_prepare);
+
+	/*
+	 * prepare cnt in escape clocks
+	 * this field represents a hexadecimal value with a precision
+	 * of 1.2 – i.e. the most significant bit is the integer
+	 * and the least significant 2 bits are fraction bits.
+	 * so, the field can represent a range of 0.25 to 1.75
+	 */
+	prepare_cnt = DIV_ROUND_UP(ths_prepare_ns * 4, tlpx_ns);
+	if (prepare_cnt > ICL_PREPARE_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm, "prepare_cnt out of range (%d)\n",
+			    prepare_cnt);
+		prepare_cnt = ICL_PREPARE_CNT_MAX;
+	}
+
+	/* clk zero count in escape clocks */
+	clk_zero_cnt = DIV_ROUND_UP(mipi_config->tclk_prepare_clkzero -
+				    ths_prepare_ns, tlpx_ns);
+	if (clk_zero_cnt > ICL_CLK_ZERO_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "clk_zero_cnt out of range (%d)\n", clk_zero_cnt);
+		clk_zero_cnt = ICL_CLK_ZERO_CNT_MAX;
+	}
+
+	/* trail cnt in escape clocks*/
+	trail_cnt = DIV_ROUND_UP(tclk_trail_ns, tlpx_ns);
+	if (trail_cnt > ICL_TRAIL_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm, "trail_cnt out of range (%d)\n",
+			    trail_cnt);
+		trail_cnt = ICL_TRAIL_CNT_MAX;
+	}
+
+	/* tclk pre count in escape clocks */
+	tclk_pre_cnt = DIV_ROUND_UP(mipi_config->tclk_pre, tlpx_ns);
+	if (tclk_pre_cnt > ICL_TCLK_PRE_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "tclk_pre_cnt out of range (%d)\n", tclk_pre_cnt);
+		tclk_pre_cnt = ICL_TCLK_PRE_CNT_MAX;
+	}
+
+	/* tclk post count in escape clocks */
+	tclk_post_cnt = DIV_ROUND_UP(mipi_config->tclk_post, tlpx_ns);
+	if (tclk_post_cnt > ICL_TCLK_POST_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "tclk_post_cnt out of range (%d)\n",
+			    tclk_post_cnt);
+		tclk_post_cnt = ICL_TCLK_POST_CNT_MAX;
+	}
+
+	/* hs zero cnt in escape clocks */
+	hs_zero_cnt = DIV_ROUND_UP(mipi_config->ths_prepare_hszero -
+				   ths_prepare_ns, tlpx_ns);
+	if (hs_zero_cnt > ICL_HS_ZERO_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm, "hs_zero_cnt out of range (%d)\n",
+			    hs_zero_cnt);
+		hs_zero_cnt = ICL_HS_ZERO_CNT_MAX;
+	}
+
+	/* hs exit zero cnt in escape clocks */
+	exit_zero_cnt = DIV_ROUND_UP(mipi_config->ths_exit, tlpx_ns);
+	if (exit_zero_cnt > ICL_EXIT_ZERO_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "exit_zero_cnt out of range (%d)\n",
+			    exit_zero_cnt);
+		exit_zero_cnt = ICL_EXIT_ZERO_CNT_MAX;
+	}
+
+	/* clock lane dphy timings */
+	intel_dsi->dphy_reg = (CLK_PREPARE_OVERRIDE |
+			       CLK_PREPARE(prepare_cnt) |
+			       CLK_ZERO_OVERRIDE |
+			       CLK_ZERO(clk_zero_cnt) |
+			       CLK_PRE_OVERRIDE |
+			       CLK_PRE(tclk_pre_cnt) |
+			       CLK_POST_OVERRIDE |
+			       CLK_POST(tclk_post_cnt) |
+			       CLK_TRAIL_OVERRIDE |
+			       CLK_TRAIL(trail_cnt));
+
+	/* data lanes dphy timings */
+	intel_dsi->dphy_data_lane_reg = (HS_PREPARE_OVERRIDE |
+					 HS_PREPARE(prepare_cnt) |
+					 HS_ZERO_OVERRIDE |
+					 HS_ZERO(hs_zero_cnt) |
+					 HS_TRAIL_OVERRIDE |
+					 HS_TRAIL(trail_cnt) |
+					 HS_EXIT_OVERRIDE |
+					 HS_EXIT(exit_zero_cnt));
+
+	intel_dsi_log_params(intel_dsi);
+}
+
+static void icl_dsi_add_properties(struct intel_connector *connector)
+{
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_preferred_fixed_mode(connector);
+	u32 allowed_scalers;
+
+	allowed_scalers = BIT(DRM_MODE_SCALE_ASPECT) |
+			   BIT(DRM_MODE_SCALE_FULLSCREEN) |
+			   BIT(DRM_MODE_SCALE_CENTER);
+
+	drm_connector_attach_scaling_mode_property(&connector->base,
+						   allowed_scalers);
+
+	connector->base.state->scaling_mode = DRM_MODE_SCALE_ASPECT;
+
+	drm_connector_set_panel_orientation_with_quirk(&connector->base,
+						       intel_dsi_get_panel_orientation(connector),
+						       fixed_mode->hdisplay,
+						       fixed_mode->vdisplay);
+}
+
+void icl_dsi_init(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_dsi *intel_dsi;
+	struct intel_encoder *encoder;
+	struct intel_connector *intel_connector;
+	struct drm_connector *connector;
+	enum port port;
+
+	if (!intel_bios_is_dsi_present(dev_priv, &port))
+		return;
+
+	intel_dsi = kzalloc(sizeof(*intel_dsi), GFP_KERNEL);
+	if (!intel_dsi)
+		return;
+
+	intel_connector = intel_connector_alloc();
+	if (!intel_connector) {
+		kfree(intel_dsi);
+		return;
+	}
+
+	encoder = &intel_dsi->base;
+	intel_dsi->attached_connector = intel_connector;
+	connector = &intel_connector->base;
+
+	/* register DSI encoder with DRM subsystem */
+	drm_encoder_init(dev, &encoder->base, &gen11_dsi_encoder_funcs,
+			 DRM_MODE_ENCODER_DSI, "DSI %c", port_name(port));
+
+	encoder->pre_pll_enable = gen11_dsi_pre_pll_enable;
+	encoder->pre_enable = gen11_dsi_pre_enable;
+	encoder->enable = gen11_dsi_enable;
+	encoder->disable = gen11_dsi_disable;
+	encoder->post_disable = gen11_dsi_post_disable;
+	encoder->port = port;
+	encoder->get_config = gen11_dsi_get_config;
+	encoder->sync_state = gen11_dsi_sync_state;
+	encoder->update_pipe = intel_backlight_update;
+	encoder->compute_config = gen11_dsi_compute_config;
+	encoder->get_hw_state = gen11_dsi_get_hw_state;
+	encoder->initial_fastset_check = gen11_dsi_initial_fastset_check;
+	encoder->type = INTEL_OUTPUT_DSI;
+	encoder->cloneable = 0;
+	encoder->pipe_mask = ~0;
+	encoder->power_domain = POWER_DOMAIN_PORT_DSI;
+	encoder->get_power_domains = gen11_dsi_get_power_domains;
+	encoder->disable_clock = gen11_dsi_gate_clocks;
+	encoder->is_clock_enabled = gen11_dsi_is_clock_enabled;
+
+	/* register DSI connector with DRM subsystem */
+	drm_connector_init(dev, connector, &gen11_dsi_connector_funcs,
+			   DRM_MODE_CONNECTOR_DSI);
+	drm_connector_helper_add(connector, &gen11_dsi_connector_helper_funcs);
+	connector->display_info.subpixel_order = SubPixelHorizontalRGB;
+	connector->interlace_allowed = false;
+	connector->doublescan_allowed = false;
+	intel_connector->get_hw_state = intel_connector_get_hw_state;
+
+	/* attach connector to encoder */
+	intel_connector_attach_encoder(intel_connector, encoder);
+
+	encoder->devdata = intel_bios_encoder_data_lookup(dev_priv, port);
+	intel_bios_init_panel_late(dev_priv, &intel_connector->panel, encoder->devdata, NULL);
+
+	mutex_lock(&dev->mode_config.mutex);
+	intel_panel_add_vbt_lfp_fixed_mode(intel_connector);
+	mutex_unlock(&dev->mode_config.mutex);
+
+	if (!intel_panel_preferred_fixed_mode(intel_connector)) {
+		drm_err(&dev_priv->drm, "DSI fixed mode info missing\n");
+		goto err;
+	}
+
+	intel_panel_init(intel_connector);
+
+	intel_backlight_setup(intel_connector, INVALID_PIPE);
+
+	if (intel_connector->panel.vbt.dsi.config->dual_link)
+		intel_dsi->ports = BIT(PORT_A) | BIT(PORT_B);
+	else
+		intel_dsi->ports = BIT(port);
+
+	if (drm_WARN_ON(&dev_priv->drm, intel_connector->panel.vbt.dsi.bl_ports & ~intel_dsi->ports))
+		intel_connector->panel.vbt.dsi.bl_ports &= intel_dsi->ports;
+
+	if (drm_WARN_ON(&dev_priv->drm, intel_connector->panel.vbt.dsi.cabc_ports & ~intel_dsi->ports))
+		intel_connector->panel.vbt.dsi.cabc_ports &= intel_dsi->ports;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		struct intel_dsi_host *host;
+
+		host = intel_dsi_host_init(intel_dsi, &gen11_dsi_host_ops, port);
+		if (!host)
+			goto err;
+
+		intel_dsi->dsi_hosts[port] = host;
+	}
+
+	if (!intel_dsi_vbt_init(intel_dsi, MIPI_DSI_GENERIC_PANEL_ID)) {
+		drm_dbg_kms(&dev_priv->drm, "no device found\n");
+		goto err;
+	}
+
+	icl_dphy_param_init(intel_dsi);
+
+	icl_dsi_add_properties(intel_connector);
+	return;
+
+err:
+	drm_connector_cleanup(connector);
+	drm_encoder_cleanup(&encoder->base);
+	kfree(intel_dsi);
+	kfree(intel_connector);
+}
diff --git a/drivers/gpu/drm/i915/display/icl_dsi.h b/drivers/gpu/drm/i915/display/icl_dsi.h
new file mode 100644
index 000000000..b4861b56b
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/icl_dsi.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __ICL_DSI_H__
+#define __ICL_DSI_H__
+
+struct drm_i915_private;
+struct intel_crtc_state;
+
+void icl_dsi_init(struct drm_i915_private *i915);
+void icl_dsi_frame_update(struct intel_crtc_state *crtc_state);
+
+#endif /* __ICL_DSI_H__ */
diff --git a/drivers/gpu/drm/i915/display/icl_dsi_regs.h b/drivers/gpu/drm/i915/display/icl_dsi_regs.h
new file mode 100644
index 000000000..f78f28b8d
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/icl_dsi_regs.h
@@ -0,0 +1,342 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __ICL_DSI_REGS_H__
+#define __ICL_DSI_REGS_H__
+
+#include "i915_reg_defs.h"
+
+/* Gen11 DSI */
+#define _MMIO_DSI(tc, dsi0, dsi1)	_MMIO_TRANS((tc) - TRANSCODER_DSI_0, \
+						    dsi0, dsi1)
+#define _ICL_DSI_ESC_CLK_DIV0		0x6b090
+#define _ICL_DSI_ESC_CLK_DIV1		0x6b890
+#define ICL_DSI_ESC_CLK_DIV(port)	_MMIO_PORT((port),	\
+							_ICL_DSI_ESC_CLK_DIV0, \
+							_ICL_DSI_ESC_CLK_DIV1)
+#define _ICL_DPHY_ESC_CLK_DIV0		0x162190
+#define _ICL_DPHY_ESC_CLK_DIV1		0x6C190
+#define ICL_DPHY_ESC_CLK_DIV(port)	_MMIO_PORT((port),	\
+						_ICL_DPHY_ESC_CLK_DIV0, \
+						_ICL_DPHY_ESC_CLK_DIV1)
+#define  ICL_BYTE_CLK_PER_ESC_CLK_MASK		(0x1f << 16)
+#define  ICL_BYTE_CLK_PER_ESC_CLK_SHIFT	16
+#define  ICL_ESC_CLK_DIV_MASK			0x1ff
+#define  ICL_ESC_CLK_DIV_SHIFT			0
+#define DSI_MAX_ESC_CLK			20000		/* in KHz */
+
+#define _ADL_MIPIO_REG			0x180
+#define ADL_MIPIO_DW(port, dw)		_MMIO(_ICL_COMBOPHY(port) + _ADL_MIPIO_REG + 4 * (dw))
+#define   TX_ESC_CLK_DIV_PHY_SEL	REGBIT(16)
+#define   TX_ESC_CLK_DIV_PHY_MASK	REG_GENMASK(23, 16)
+#define   TX_ESC_CLK_DIV_PHY		REG_FIELD_PREP(TX_ESC_CLK_DIV_PHY_MASK, 0x7f)
+
+#define _DSI_CMD_FRMCTL_0		0x6b034
+#define _DSI_CMD_FRMCTL_1		0x6b834
+#define DSI_CMD_FRMCTL(port)		_MMIO_PORT(port,	\
+						   _DSI_CMD_FRMCTL_0,\
+						   _DSI_CMD_FRMCTL_1)
+#define   DSI_FRAME_UPDATE_REQUEST		(1 << 31)
+#define   DSI_PERIODIC_FRAME_UPDATE_ENABLE	(1 << 29)
+#define   DSI_NULL_PACKET_ENABLE		(1 << 28)
+#define   DSI_FRAME_IN_PROGRESS			(1 << 0)
+
+#define _DSI_INTR_MASK_REG_0		0x6b070
+#define _DSI_INTR_MASK_REG_1		0x6b870
+#define DSI_INTR_MASK_REG(port)		_MMIO_PORT(port,	\
+						   _DSI_INTR_MASK_REG_0,\
+						   _DSI_INTR_MASK_REG_1)
+
+#define _DSI_INTR_IDENT_REG_0		0x6b074
+#define _DSI_INTR_IDENT_REG_1		0x6b874
+#define DSI_INTR_IDENT_REG(port)	_MMIO_PORT(port,	\
+						   _DSI_INTR_IDENT_REG_0,\
+						   _DSI_INTR_IDENT_REG_1)
+#define   DSI_TE_EVENT				(1 << 31)
+#define   DSI_RX_DATA_OR_BTA_TERMINATED		(1 << 30)
+#define   DSI_TX_DATA				(1 << 29)
+#define   DSI_ULPS_ENTRY_DONE			(1 << 28)
+#define   DSI_NON_TE_TRIGGER_RECEIVED		(1 << 27)
+#define   DSI_HOST_CHKSUM_ERROR			(1 << 26)
+#define   DSI_HOST_MULTI_ECC_ERROR		(1 << 25)
+#define   DSI_HOST_SINGL_ECC_ERROR		(1 << 24)
+#define   DSI_HOST_CONTENTION_DETECTED		(1 << 23)
+#define   DSI_HOST_FALSE_CONTROL_ERROR		(1 << 22)
+#define   DSI_HOST_TIMEOUT_ERROR		(1 << 21)
+#define   DSI_HOST_LOW_POWER_TX_SYNC_ERROR	(1 << 20)
+#define   DSI_HOST_ESCAPE_MODE_ENTRY_ERROR	(1 << 19)
+#define   DSI_FRAME_UPDATE_DONE			(1 << 16)
+#define   DSI_PROTOCOL_VIOLATION_REPORTED	(1 << 15)
+#define   DSI_INVALID_TX_LENGTH			(1 << 13)
+#define   DSI_INVALID_VC			(1 << 12)
+#define   DSI_INVALID_DATA_TYPE			(1 << 11)
+#define   DSI_PERIPHERAL_CHKSUM_ERROR		(1 << 10)
+#define   DSI_PERIPHERAL_MULTI_ECC_ERROR	(1 << 9)
+#define   DSI_PERIPHERAL_SINGLE_ECC_ERROR	(1 << 8)
+#define   DSI_PERIPHERAL_CONTENTION_DETECTED	(1 << 7)
+#define   DSI_PERIPHERAL_FALSE_CTRL_ERROR	(1 << 6)
+#define   DSI_PERIPHERAL_TIMEOUT_ERROR		(1 << 5)
+#define   DSI_PERIPHERAL_LP_TX_SYNC_ERROR	(1 << 4)
+#define   DSI_PERIPHERAL_ESC_MODE_ENTRY_CMD_ERR	(1 << 3)
+#define   DSI_EOT_SYNC_ERROR			(1 << 2)
+#define   DSI_SOT_SYNC_ERROR			(1 << 1)
+#define   DSI_SOT_ERROR				(1 << 0)
+
+/* ICL DSI MODE control */
+#define _ICL_DSI_IO_MODECTL_0				0x6B094
+#define _ICL_DSI_IO_MODECTL_1				0x6B894
+#define ICL_DSI_IO_MODECTL(port)	_MMIO_PORT(port,	\
+						    _ICL_DSI_IO_MODECTL_0, \
+						    _ICL_DSI_IO_MODECTL_1)
+#define  COMBO_PHY_MODE_DSI				(1 << 0)
+
+/* TGL DSI Chicken register */
+#define _TGL_DSI_CHKN_REG_0			0x6B0C0
+#define _TGL_DSI_CHKN_REG_1			0x6B8C0
+#define TGL_DSI_CHKN_REG(port)		_MMIO_PORT(port,	\
+						    _TGL_DSI_CHKN_REG_0, \
+						    _TGL_DSI_CHKN_REG_1)
+#define TGL_DSI_CHKN_LSHS_GB_MASK		REG_GENMASK(15, 12)
+#define TGL_DSI_CHKN_LSHS_GB(byte_clocks)	REG_FIELD_PREP(TGL_DSI_CHKN_LSHS_GB_MASK, \
+							       (byte_clocks))
+#define _ICL_DSI_T_INIT_MASTER_0	0x6b088
+#define _ICL_DSI_T_INIT_MASTER_1	0x6b888
+#define ICL_DSI_T_INIT_MASTER(port)	_MMIO_PORT(port,	\
+						   _ICL_DSI_T_INIT_MASTER_0,\
+						   _ICL_DSI_T_INIT_MASTER_1)
+#define   DSI_T_INIT_MASTER_MASK	REG_GENMASK(15, 0)
+
+#define _DPHY_CLK_TIMING_PARAM_0	0x162180
+#define _DPHY_CLK_TIMING_PARAM_1	0x6c180
+#define DPHY_CLK_TIMING_PARAM(port)	_MMIO_PORT(port,	\
+						   _DPHY_CLK_TIMING_PARAM_0,\
+						   _DPHY_CLK_TIMING_PARAM_1)
+#define _DSI_CLK_TIMING_PARAM_0		0x6b080
+#define _DSI_CLK_TIMING_PARAM_1		0x6b880
+#define DSI_CLK_TIMING_PARAM(port)	_MMIO_PORT(port,	\
+						   _DSI_CLK_TIMING_PARAM_0,\
+						   _DSI_CLK_TIMING_PARAM_1)
+#define  CLK_PREPARE_OVERRIDE		(1 << 31)
+#define  CLK_PREPARE(x)		((x) << 28)
+#define  CLK_PREPARE_MASK		(0x7 << 28)
+#define  CLK_PREPARE_SHIFT		28
+#define  CLK_ZERO_OVERRIDE		(1 << 27)
+#define  CLK_ZERO(x)			((x) << 20)
+#define  CLK_ZERO_MASK			(0xf << 20)
+#define  CLK_ZERO_SHIFT		20
+#define  CLK_PRE_OVERRIDE		(1 << 19)
+#define  CLK_PRE(x)			((x) << 16)
+#define  CLK_PRE_MASK			(0x3 << 16)
+#define  CLK_PRE_SHIFT			16
+#define  CLK_POST_OVERRIDE		(1 << 15)
+#define  CLK_POST(x)			((x) << 8)
+#define  CLK_POST_MASK			(0x7 << 8)
+#define  CLK_POST_SHIFT		8
+#define  CLK_TRAIL_OVERRIDE		(1 << 7)
+#define  CLK_TRAIL(x)			((x) << 0)
+#define  CLK_TRAIL_MASK		(0xf << 0)
+#define  CLK_TRAIL_SHIFT		0
+
+#define _DPHY_DATA_TIMING_PARAM_0	0x162184
+#define _DPHY_DATA_TIMING_PARAM_1	0x6c184
+#define DPHY_DATA_TIMING_PARAM(port)	_MMIO_PORT(port,	\
+						   _DPHY_DATA_TIMING_PARAM_0,\
+						   _DPHY_DATA_TIMING_PARAM_1)
+#define _DSI_DATA_TIMING_PARAM_0	0x6B084
+#define _DSI_DATA_TIMING_PARAM_1	0x6B884
+#define DSI_DATA_TIMING_PARAM(port)	_MMIO_PORT(port,	\
+						   _DSI_DATA_TIMING_PARAM_0,\
+						   _DSI_DATA_TIMING_PARAM_1)
+#define  HS_PREPARE_OVERRIDE		(1 << 31)
+#define  HS_PREPARE(x)			((x) << 24)
+#define  HS_PREPARE_MASK		(0x7 << 24)
+#define  HS_PREPARE_SHIFT		24
+#define  HS_ZERO_OVERRIDE		(1 << 23)
+#define  HS_ZERO(x)			((x) << 16)
+#define  HS_ZERO_MASK			(0xf << 16)
+#define  HS_ZERO_SHIFT			16
+#define  HS_TRAIL_OVERRIDE		(1 << 15)
+#define  HS_TRAIL(x)			((x) << 8)
+#define  HS_TRAIL_MASK			(0x7 << 8)
+#define  HS_TRAIL_SHIFT		8
+#define  HS_EXIT_OVERRIDE		(1 << 7)
+#define  HS_EXIT(x)			((x) << 0)
+#define  HS_EXIT_MASK			(0x7 << 0)
+#define  HS_EXIT_SHIFT			0
+
+#define _DPHY_TA_TIMING_PARAM_0		0x162188
+#define _DPHY_TA_TIMING_PARAM_1		0x6c188
+#define DPHY_TA_TIMING_PARAM(port)	_MMIO_PORT(port,	\
+						   _DPHY_TA_TIMING_PARAM_0,\
+						   _DPHY_TA_TIMING_PARAM_1)
+#define _DSI_TA_TIMING_PARAM_0		0x6b098
+#define _DSI_TA_TIMING_PARAM_1		0x6b898
+#define DSI_TA_TIMING_PARAM(port)	_MMIO_PORT(port,	\
+						   _DSI_TA_TIMING_PARAM_0,\
+						   _DSI_TA_TIMING_PARAM_1)
+#define  TA_SURE_OVERRIDE		(1 << 31)
+#define  TA_SURE(x)			((x) << 16)
+#define  TA_SURE_MASK			(0x1f << 16)
+#define  TA_SURE_SHIFT			16
+#define  TA_GO_OVERRIDE		(1 << 15)
+#define  TA_GO(x)			((x) << 8)
+#define  TA_GO_MASK			(0xf << 8)
+#define  TA_GO_SHIFT			8
+#define  TA_GET_OVERRIDE		(1 << 7)
+#define  TA_GET(x)			((x) << 0)
+#define  TA_GET_MASK			(0xf << 0)
+#define  TA_GET_SHIFT			0
+
+/* DSI transcoder configuration */
+#define _DSI_TRANS_FUNC_CONF_0		0x6b030
+#define _DSI_TRANS_FUNC_CONF_1		0x6b830
+#define DSI_TRANS_FUNC_CONF(tc)		_MMIO_DSI(tc,	\
+						  _DSI_TRANS_FUNC_CONF_0,\
+						  _DSI_TRANS_FUNC_CONF_1)
+#define  OP_MODE_MASK			(0x3 << 28)
+#define  OP_MODE_SHIFT			28
+#define  CMD_MODE_NO_GATE		(0x0 << 28)
+#define  CMD_MODE_TE_GATE		(0x1 << 28)
+#define  VIDEO_MODE_SYNC_EVENT		(0x2 << 28)
+#define  VIDEO_MODE_SYNC_PULSE		(0x3 << 28)
+#define  TE_SOURCE_GPIO			(1 << 27)
+#define  LINK_READY			(1 << 20)
+#define  PIX_FMT_MASK			(0x3 << 16)
+#define  PIX_FMT_SHIFT			16
+#define  PIX_FMT_RGB565			(0x0 << 16)
+#define  PIX_FMT_RGB666_PACKED		(0x1 << 16)
+#define  PIX_FMT_RGB666_LOOSE		(0x2 << 16)
+#define  PIX_FMT_RGB888			(0x3 << 16)
+#define  PIX_FMT_RGB101010		(0x4 << 16)
+#define  PIX_FMT_RGB121212		(0x5 << 16)
+#define  PIX_FMT_COMPRESSED		(0x6 << 16)
+#define  BGR_TRANSMISSION		(1 << 15)
+#define  PIX_VIRT_CHAN(x)		((x) << 12)
+#define  PIX_VIRT_CHAN_MASK		(0x3 << 12)
+#define  PIX_VIRT_CHAN_SHIFT		12
+#define  PIX_BUF_THRESHOLD_MASK		(0x3 << 10)
+#define  PIX_BUF_THRESHOLD_SHIFT	10
+#define  PIX_BUF_THRESHOLD_1_4		(0x0 << 10)
+#define  PIX_BUF_THRESHOLD_1_2		(0x1 << 10)
+#define  PIX_BUF_THRESHOLD_3_4		(0x2 << 10)
+#define  PIX_BUF_THRESHOLD_FULL		(0x3 << 10)
+#define  CONTINUOUS_CLK_MASK		(0x3 << 8)
+#define  CONTINUOUS_CLK_SHIFT		8
+#define  CLK_ENTER_LP_AFTER_DATA	(0x0 << 8)
+#define  CLK_HS_OR_LP			(0x2 << 8)
+#define  CLK_HS_CONTINUOUS		(0x3 << 8)
+#define  LINK_CALIBRATION_MASK		(0x3 << 4)
+#define  LINK_CALIBRATION_SHIFT		4
+#define  CALIBRATION_DISABLED		(0x0 << 4)
+#define  CALIBRATION_ENABLED_INITIAL_ONLY	(0x2 << 4)
+#define  CALIBRATION_ENABLED_INITIAL_PERIODIC	(0x3 << 4)
+#define  BLANKING_PACKET_ENABLE		(1 << 2)
+#define  S3D_ORIENTATION_LANDSCAPE	(1 << 1)
+#define  EOTP_DISABLED			(1 << 0)
+
+#define _DSI_CMD_RXCTL_0		0x6b0d4
+#define _DSI_CMD_RXCTL_1		0x6b8d4
+#define DSI_CMD_RXCTL(tc)		_MMIO_DSI(tc,	\
+						  _DSI_CMD_RXCTL_0,\
+						  _DSI_CMD_RXCTL_1)
+#define  READ_UNLOADS_DW		(1 << 16)
+#define  RECEIVED_UNASSIGNED_TRIGGER	(1 << 15)
+#define  RECEIVED_ACKNOWLEDGE_TRIGGER	(1 << 14)
+#define  RECEIVED_TEAR_EFFECT_TRIGGER	(1 << 13)
+#define  RECEIVED_RESET_TRIGGER		(1 << 12)
+#define  RECEIVED_PAYLOAD_WAS_LOST	(1 << 11)
+#define  RECEIVED_CRC_WAS_LOST		(1 << 10)
+#define  NUMBER_RX_PLOAD_DW_MASK	(0xff << 0)
+#define  NUMBER_RX_PLOAD_DW_SHIFT	0
+
+#define _DSI_CMD_TXCTL_0		0x6b0d0
+#define _DSI_CMD_TXCTL_1		0x6b8d0
+#define DSI_CMD_TXCTL(tc)		_MMIO_DSI(tc,	\
+						  _DSI_CMD_TXCTL_0,\
+						  _DSI_CMD_TXCTL_1)
+#define  KEEP_LINK_IN_HS		(1 << 24)
+#define  FREE_HEADER_CREDIT_MASK	(0x1f << 8)
+#define  FREE_HEADER_CREDIT_SHIFT	0x8
+#define  FREE_PLOAD_CREDIT_MASK		(0xff << 0)
+#define  FREE_PLOAD_CREDIT_SHIFT	0
+#define  MAX_HEADER_CREDIT		0x10
+#define  MAX_PLOAD_CREDIT		0x40
+
+#define _DSI_CMD_TXHDR_0		0x6b100
+#define _DSI_CMD_TXHDR_1		0x6b900
+#define DSI_CMD_TXHDR(tc)		_MMIO_DSI(tc,	\
+						  _DSI_CMD_TXHDR_0,\
+						  _DSI_CMD_TXHDR_1)
+#define  PAYLOAD_PRESENT		(1 << 31)
+#define  LP_DATA_TRANSFER		(1 << 30)
+#define  VBLANK_FENCE			(1 << 29)
+#define  PARAM_WC_MASK			(0xffff << 8)
+#define  PARAM_WC_LOWER_SHIFT		8
+#define  PARAM_WC_UPPER_SHIFT		16
+#define  VC_MASK			(0x3 << 6)
+#define  VC_SHIFT			6
+#define  DT_MASK			(0x3f << 0)
+#define  DT_SHIFT			0
+
+#define _DSI_CMD_TXPYLD_0		0x6b104
+#define _DSI_CMD_TXPYLD_1		0x6b904
+#define DSI_CMD_TXPYLD(tc)		_MMIO_DSI(tc,	\
+						  _DSI_CMD_TXPYLD_0,\
+						  _DSI_CMD_TXPYLD_1)
+
+#define _DSI_LP_MSG_0			0x6b0d8
+#define _DSI_LP_MSG_1			0x6b8d8
+#define DSI_LP_MSG(tc)			_MMIO_DSI(tc,	\
+						  _DSI_LP_MSG_0,\
+						  _DSI_LP_MSG_1)
+#define  LPTX_IN_PROGRESS		(1 << 17)
+#define  LINK_IN_ULPS			(1 << 16)
+#define  LINK_ULPS_TYPE_LP11		(1 << 8)
+#define  LINK_ENTER_ULPS		(1 << 0)
+
+/* DSI timeout registers */
+#define _DSI_HSTX_TO_0			0x6b044
+#define _DSI_HSTX_TO_1			0x6b844
+#define DSI_HSTX_TO(tc)			_MMIO_DSI(tc,	\
+						  _DSI_HSTX_TO_0,\
+						  _DSI_HSTX_TO_1)
+#define  HSTX_TIMEOUT_VALUE_MASK	(0xffff << 16)
+#define  HSTX_TIMEOUT_VALUE_SHIFT	16
+#define  HSTX_TIMEOUT_VALUE(x)		((x) << 16)
+#define  HSTX_TIMED_OUT			(1 << 0)
+
+#define _DSI_LPRX_HOST_TO_0		0x6b048
+#define _DSI_LPRX_HOST_TO_1		0x6b848
+#define DSI_LPRX_HOST_TO(tc)		_MMIO_DSI(tc,	\
+						  _DSI_LPRX_HOST_TO_0,\
+						  _DSI_LPRX_HOST_TO_1)
+#define  LPRX_TIMED_OUT			(1 << 16)
+#define  LPRX_TIMEOUT_VALUE_MASK	(0xffff << 0)
+#define  LPRX_TIMEOUT_VALUE_SHIFT	0
+#define  LPRX_TIMEOUT_VALUE(x)		((x) << 0)
+
+#define _DSI_PWAIT_TO_0			0x6b040
+#define _DSI_PWAIT_TO_1			0x6b840
+#define DSI_PWAIT_TO(tc)		_MMIO_DSI(tc,	\
+						  _DSI_PWAIT_TO_0,\
+						  _DSI_PWAIT_TO_1)
+#define  PRESET_TIMEOUT_VALUE_MASK	(0xffff << 16)
+#define  PRESET_TIMEOUT_VALUE_SHIFT	16
+#define  PRESET_TIMEOUT_VALUE(x)	((x) << 16)
+#define  PRESPONSE_TIMEOUT_VALUE_MASK	(0xffff << 0)
+#define  PRESPONSE_TIMEOUT_VALUE_SHIFT	0
+#define  PRESPONSE_TIMEOUT_VALUE(x)	((x) << 0)
+
+#define _DSI_TA_TO_0			0x6b04c
+#define _DSI_TA_TO_1			0x6b84c
+#define DSI_TA_TO(tc)			_MMIO_DSI(tc,	\
+						  _DSI_TA_TO_0,\
+						  _DSI_TA_TO_1)
+#define  TA_TIMED_OUT			(1 << 16)
+#define  TA_TIMEOUT_VALUE_MASK		(0xffff << 0)
+#define  TA_TIMEOUT_VALUE_SHIFT		0
+#define  TA_TIMEOUT_VALUE(x)		((x) << 0)
+
+#endif /* __ICL_DSI_REGS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_acpi.c b/drivers/gpu/drm/i915/display/intel_acpi.c
new file mode 100644
index 000000000..9df78e7ca
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_acpi.c
@@ -0,0 +1,360 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Intel ACPI functions
+ *
+ * _DSM related code stolen from nouveau_acpi.c.
+ */
+
+#include <linux/pci.h>
+#include <linux/acpi.h>
+#include <acpi/video.h>
+
+#include "i915_drv.h"
+#include "intel_acpi.h"
+#include "intel_display_types.h"
+
+#define INTEL_DSM_REVISION_ID 1 /* For Calpella anyway... */
+#define INTEL_DSM_FN_PLATFORM_MUX_INFO 1 /* No args */
+
+static const guid_t intel_dsm_guid =
+	GUID_INIT(0x7ed873d3, 0xc2d0, 0x4e4f,
+		  0xa8, 0x54, 0x0f, 0x13, 0x17, 0xb0, 0x1c, 0x2c);
+
+#define INTEL_DSM_FN_GET_BIOS_DATA_FUNCS_SUPPORTED 0 /* No args */
+
+static const guid_t intel_dsm_guid2 =
+	GUID_INIT(0x3e5b41c6, 0xeb1d, 0x4260,
+		  0x9d, 0x15, 0xc7, 0x1f, 0xba, 0xda, 0xe4, 0x14);
+
+static char *intel_dsm_port_name(u8 id)
+{
+	switch (id) {
+	case 0:
+		return "Reserved";
+	case 1:
+		return "Analog VGA";
+	case 2:
+		return "LVDS";
+	case 3:
+		return "Reserved";
+	case 4:
+		return "HDMI/DVI_B";
+	case 5:
+		return "HDMI/DVI_C";
+	case 6:
+		return "HDMI/DVI_D";
+	case 7:
+		return "DisplayPort_A";
+	case 8:
+		return "DisplayPort_B";
+	case 9:
+		return "DisplayPort_C";
+	case 0xa:
+		return "DisplayPort_D";
+	case 0xb:
+	case 0xc:
+	case 0xd:
+		return "Reserved";
+	case 0xe:
+		return "WiDi";
+	default:
+		return "bad type";
+	}
+}
+
+static char *intel_dsm_mux_type(u8 type)
+{
+	switch (type) {
+	case 0:
+		return "unknown";
+	case 1:
+		return "No MUX, iGPU only";
+	case 2:
+		return "No MUX, dGPU only";
+	case 3:
+		return "MUXed between iGPU and dGPU";
+	default:
+		return "bad type";
+	}
+}
+
+static void intel_dsm_platform_mux_info(acpi_handle dhandle)
+{
+	int i;
+	union acpi_object *pkg, *connector_count;
+
+	pkg = acpi_evaluate_dsm_typed(dhandle, &intel_dsm_guid,
+			INTEL_DSM_REVISION_ID, INTEL_DSM_FN_PLATFORM_MUX_INFO,
+			NULL, ACPI_TYPE_PACKAGE);
+	if (!pkg) {
+		DRM_DEBUG_DRIVER("failed to evaluate _DSM\n");
+		return;
+	}
+
+	if (!pkg->package.count) {
+		DRM_DEBUG_DRIVER("no connection in _DSM\n");
+		return;
+	}
+
+	connector_count = &pkg->package.elements[0];
+	DRM_DEBUG_DRIVER("MUX info connectors: %lld\n",
+		  (unsigned long long)connector_count->integer.value);
+	for (i = 1; i < pkg->package.count; i++) {
+		union acpi_object *obj = &pkg->package.elements[i];
+		union acpi_object *connector_id;
+		union acpi_object *info;
+
+		if (obj->type != ACPI_TYPE_PACKAGE || obj->package.count < 2) {
+			DRM_DEBUG_DRIVER("Invalid object for MUX #%d\n", i);
+			continue;
+		}
+
+		connector_id = &obj->package.elements[0];
+		info = &obj->package.elements[1];
+		if (info->type != ACPI_TYPE_BUFFER || info->buffer.length < 4) {
+			DRM_DEBUG_DRIVER("Invalid info for MUX obj #%d\n", i);
+			continue;
+		}
+
+		DRM_DEBUG_DRIVER("Connector id: 0x%016llx\n",
+			  (unsigned long long)connector_id->integer.value);
+		DRM_DEBUG_DRIVER("  port id: %s\n",
+		       intel_dsm_port_name(info->buffer.pointer[0]));
+		DRM_DEBUG_DRIVER("  display mux info: %s\n",
+		       intel_dsm_mux_type(info->buffer.pointer[1]));
+		DRM_DEBUG_DRIVER("  aux/dc mux info: %s\n",
+		       intel_dsm_mux_type(info->buffer.pointer[2]));
+		DRM_DEBUG_DRIVER("  hpd mux info: %s\n",
+		       intel_dsm_mux_type(info->buffer.pointer[3]));
+	}
+
+	ACPI_FREE(pkg);
+}
+
+static acpi_handle intel_dsm_pci_probe(struct pci_dev *pdev)
+{
+	acpi_handle dhandle;
+
+	dhandle = ACPI_HANDLE(&pdev->dev);
+	if (!dhandle)
+		return NULL;
+
+	if (!acpi_check_dsm(dhandle, &intel_dsm_guid, INTEL_DSM_REVISION_ID,
+			    1 << INTEL_DSM_FN_PLATFORM_MUX_INFO)) {
+		DRM_DEBUG_KMS("no _DSM method for intel device\n");
+		return NULL;
+	}
+
+	intel_dsm_platform_mux_info(dhandle);
+
+	return dhandle;
+}
+
+static bool intel_dsm_detect(void)
+{
+	acpi_handle dhandle = NULL;
+	char acpi_method_name[255] = { 0 };
+	struct acpi_buffer buffer = {sizeof(acpi_method_name), acpi_method_name};
+	struct pci_dev *pdev = NULL;
+	int vga_count = 0;
+
+	while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, pdev)) != NULL) {
+		vga_count++;
+		dhandle = intel_dsm_pci_probe(pdev) ?: dhandle;
+	}
+
+	if (vga_count == 2 && dhandle) {
+		acpi_get_name(dhandle, ACPI_FULL_PATHNAME, &buffer);
+		DRM_DEBUG_DRIVER("vga_switcheroo: detected DSM switching method %s handle\n",
+				 acpi_method_name);
+		return true;
+	}
+
+	return false;
+}
+
+void intel_register_dsm_handler(void)
+{
+	if (!intel_dsm_detect())
+		return;
+}
+
+void intel_unregister_dsm_handler(void)
+{
+}
+
+void intel_dsm_get_bios_data_funcs_supported(struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	acpi_handle dhandle;
+	union acpi_object *obj;
+
+	dhandle = ACPI_HANDLE(&pdev->dev);
+	if (!dhandle)
+		return;
+
+	obj = acpi_evaluate_dsm(dhandle, &intel_dsm_guid2, INTEL_DSM_REVISION_ID,
+				INTEL_DSM_FN_GET_BIOS_DATA_FUNCS_SUPPORTED, NULL);
+	if (obj)
+		ACPI_FREE(obj);
+}
+
+/*
+ * ACPI Specification, Revision 5.0, Appendix B.3.2 _DOD (Enumerate All Devices
+ * Attached to the Display Adapter).
+ */
+#define ACPI_DISPLAY_INDEX_SHIFT		0
+#define ACPI_DISPLAY_INDEX_MASK			(0xf << 0)
+#define ACPI_DISPLAY_PORT_ATTACHMENT_SHIFT	4
+#define ACPI_DISPLAY_PORT_ATTACHMENT_MASK	(0xf << 4)
+#define ACPI_DISPLAY_TYPE_SHIFT			8
+#define ACPI_DISPLAY_TYPE_MASK			(0xf << 8)
+#define ACPI_DISPLAY_TYPE_OTHER			(0 << 8)
+#define ACPI_DISPLAY_TYPE_VGA			(1 << 8)
+#define ACPI_DISPLAY_TYPE_TV			(2 << 8)
+#define ACPI_DISPLAY_TYPE_EXTERNAL_DIGITAL	(3 << 8)
+#define ACPI_DISPLAY_TYPE_INTERNAL_DIGITAL	(4 << 8)
+#define ACPI_VENDOR_SPECIFIC_SHIFT		12
+#define ACPI_VENDOR_SPECIFIC_MASK		(0xf << 12)
+#define ACPI_BIOS_CAN_DETECT			(1 << 16)
+#define ACPI_DEPENDS_ON_VGA			(1 << 17)
+#define ACPI_PIPE_ID_SHIFT			18
+#define ACPI_PIPE_ID_MASK			(7 << 18)
+#define ACPI_DEVICE_ID_SCHEME			(1ULL << 31)
+
+static u32 acpi_display_type(struct intel_connector *connector)
+{
+	u32 display_type;
+
+	switch (connector->base.connector_type) {
+	case DRM_MODE_CONNECTOR_VGA:
+	case DRM_MODE_CONNECTOR_DVIA:
+		display_type = ACPI_DISPLAY_TYPE_VGA;
+		break;
+	case DRM_MODE_CONNECTOR_Composite:
+	case DRM_MODE_CONNECTOR_SVIDEO:
+	case DRM_MODE_CONNECTOR_Component:
+	case DRM_MODE_CONNECTOR_9PinDIN:
+	case DRM_MODE_CONNECTOR_TV:
+		display_type = ACPI_DISPLAY_TYPE_TV;
+		break;
+	case DRM_MODE_CONNECTOR_DVII:
+	case DRM_MODE_CONNECTOR_DVID:
+	case DRM_MODE_CONNECTOR_DisplayPort:
+	case DRM_MODE_CONNECTOR_HDMIA:
+	case DRM_MODE_CONNECTOR_HDMIB:
+		display_type = ACPI_DISPLAY_TYPE_EXTERNAL_DIGITAL;
+		break;
+	case DRM_MODE_CONNECTOR_LVDS:
+	case DRM_MODE_CONNECTOR_eDP:
+	case DRM_MODE_CONNECTOR_DSI:
+		display_type = ACPI_DISPLAY_TYPE_INTERNAL_DIGITAL;
+		break;
+	case DRM_MODE_CONNECTOR_Unknown:
+	case DRM_MODE_CONNECTOR_VIRTUAL:
+		display_type = ACPI_DISPLAY_TYPE_OTHER;
+		break;
+	default:
+		MISSING_CASE(connector->base.connector_type);
+		display_type = ACPI_DISPLAY_TYPE_OTHER;
+		break;
+	}
+
+	return display_type;
+}
+
+void intel_acpi_device_id_update(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *drm_dev = &dev_priv->drm;
+	struct intel_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	u8 display_index[16] = {};
+
+	/* Populate the ACPI IDs for all connectors for a given drm_device */
+	drm_connector_list_iter_begin(drm_dev, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		u32 device_id, type;
+
+		device_id = acpi_display_type(connector);
+
+		/* Use display type specific display index. */
+		type = (device_id & ACPI_DISPLAY_TYPE_MASK)
+			>> ACPI_DISPLAY_TYPE_SHIFT;
+		device_id |= display_index[type]++ << ACPI_DISPLAY_INDEX_SHIFT;
+
+		connector->acpi_device_id = device_id;
+	}
+	drm_connector_list_iter_end(&conn_iter);
+}
+
+/* NOTE: The connector order must be final before this is called. */
+void intel_acpi_assign_connector_fwnodes(struct drm_i915_private *i915)
+{
+	struct drm_connector_list_iter conn_iter;
+	struct drm_device *drm_dev = &i915->drm;
+	struct fwnode_handle *fwnode = NULL;
+	struct drm_connector *connector;
+	struct acpi_device *adev;
+
+	drm_connector_list_iter_begin(drm_dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		/* Always getting the next, even when the last was not used. */
+		fwnode = device_get_next_child_node(drm_dev->dev, fwnode);
+		if (!fwnode)
+			break;
+
+		switch (connector->connector_type) {
+		case DRM_MODE_CONNECTOR_LVDS:
+		case DRM_MODE_CONNECTOR_eDP:
+		case DRM_MODE_CONNECTOR_DSI:
+			/*
+			 * Integrated displays have a specific address 0x1f on
+			 * most Intel platforms, but not on all of them.
+			 */
+			adev = acpi_find_child_device(ACPI_COMPANION(drm_dev->dev),
+						      0x1f, 0);
+			if (adev) {
+				connector->fwnode =
+					fwnode_handle_get(acpi_fwnode_handle(adev));
+				break;
+			}
+			fallthrough;
+		default:
+			connector->fwnode = fwnode_handle_get(fwnode);
+			break;
+		}
+	}
+	drm_connector_list_iter_end(&conn_iter);
+	/*
+	 * device_get_next_child_node() takes a reference on the fwnode, if
+	 * we stopped iterating because we are out of connectors we need to
+	 * put this, otherwise fwnode is NULL and the put is a no-op.
+	 */
+	fwnode_handle_put(fwnode);
+}
+
+void intel_acpi_video_register(struct drm_i915_private *i915)
+{
+	struct drm_connector_list_iter conn_iter;
+	struct drm_connector *connector;
+
+	acpi_video_register();
+
+	/*
+	 * If i915 is driving an internal panel without registering its native
+	 * backlight handler try to register the acpi_video backlight.
+	 * For panels not driven by i915 another GPU driver may still register
+	 * a native backlight later and acpi_video_register_backlight() should
+	 * only be called after any native backlights have been registered.
+	 */
+	drm_connector_list_iter_begin(&i915->drm, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		struct intel_panel *panel = &to_intel_connector(connector)->panel;
+
+		if (panel->backlight.funcs && !panel->backlight.device) {
+			acpi_video_register_backlight();
+			break;
+		}
+	}
+	drm_connector_list_iter_end(&conn_iter);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_acpi.h b/drivers/gpu/drm/i915/display/intel_acpi.h
new file mode 100644
index 000000000..6a0007452
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_acpi.h
@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_ACPI_H__
+#define __INTEL_ACPI_H__
+
+struct drm_i915_private;
+
+#ifdef CONFIG_ACPI
+void intel_register_dsm_handler(void);
+void intel_unregister_dsm_handler(void);
+void intel_dsm_get_bios_data_funcs_supported(struct drm_i915_private *i915);
+void intel_acpi_device_id_update(struct drm_i915_private *i915);
+void intel_acpi_assign_connector_fwnodes(struct drm_i915_private *i915);
+void intel_acpi_video_register(struct drm_i915_private *i915);
+#else
+static inline void intel_register_dsm_handler(void) { return; }
+static inline void intel_unregister_dsm_handler(void) { return; }
+static inline
+void intel_dsm_get_bios_data_funcs_supported(struct drm_i915_private *i915) { return; }
+static inline
+void intel_acpi_device_id_update(struct drm_i915_private *i915) { return; }
+static inline
+void intel_acpi_assign_connector_fwnodes(struct drm_i915_private *i915) { return; }
+static inline
+void intel_acpi_video_register(struct drm_i915_private *i915) { return; }
+#endif /* CONFIG_ACPI */
+
+#endif /* __INTEL_ACPI_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_atomic.c b/drivers/gpu/drm/i915/display/intel_atomic.c
new file mode 100644
index 000000000..a502af0b6
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_atomic.c
@@ -0,0 +1,351 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+/**
+ * DOC: atomic modeset support
+ *
+ * The functions here implement the state management and hardware programming
+ * dispatch required by the atomic modeset infrastructure.
+ * See intel_atomic_plane.c for the plane-specific atomic functionality.
+ */
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_fourcc.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_atomic.h"
+#include "intel_cdclk.h"
+#include "intel_display_types.h"
+#include "intel_global_state.h"
+#include "intel_hdcp.h"
+#include "intel_psr.h"
+#include "intel_fb.h"
+#include "skl_universal_plane.h"
+
+/**
+ * intel_digital_connector_atomic_get_property - hook for connector->atomic_get_property.
+ * @connector: Connector to get the property for.
+ * @state: Connector state to retrieve the property from.
+ * @property: Property to retrieve.
+ * @val: Return value for the property.
+ *
+ * Returns the atomic property value for a digital connector.
+ */
+int intel_digital_connector_atomic_get_property(struct drm_connector *connector,
+						const struct drm_connector_state *state,
+						struct drm_property *property,
+						u64 *val)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_digital_connector_state *intel_conn_state =
+		to_intel_digital_connector_state(state);
+
+	if (property == dev_priv->display.properties.force_audio)
+		*val = intel_conn_state->force_audio;
+	else if (property == dev_priv->display.properties.broadcast_rgb)
+		*val = intel_conn_state->broadcast_rgb;
+	else {
+		drm_dbg_atomic(&dev_priv->drm,
+			       "Unknown property [PROP:%d:%s]\n",
+			       property->base.id, property->name);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/**
+ * intel_digital_connector_atomic_set_property - hook for connector->atomic_set_property.
+ * @connector: Connector to set the property for.
+ * @state: Connector state to set the property on.
+ * @property: Property to set.
+ * @val: New value for the property.
+ *
+ * Sets the atomic property value for a digital connector.
+ */
+int intel_digital_connector_atomic_set_property(struct drm_connector *connector,
+						struct drm_connector_state *state,
+						struct drm_property *property,
+						u64 val)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_digital_connector_state *intel_conn_state =
+		to_intel_digital_connector_state(state);
+
+	if (property == dev_priv->display.properties.force_audio) {
+		intel_conn_state->force_audio = val;
+		return 0;
+	}
+
+	if (property == dev_priv->display.properties.broadcast_rgb) {
+		intel_conn_state->broadcast_rgb = val;
+		return 0;
+	}
+
+	drm_dbg_atomic(&dev_priv->drm, "Unknown property [PROP:%d:%s]\n",
+		       property->base.id, property->name);
+	return -EINVAL;
+}
+
+int intel_digital_connector_atomic_check(struct drm_connector *conn,
+					 struct drm_atomic_state *state)
+{
+	struct drm_connector_state *new_state =
+		drm_atomic_get_new_connector_state(state, conn);
+	struct intel_digital_connector_state *new_conn_state =
+		to_intel_digital_connector_state(new_state);
+	struct drm_connector_state *old_state =
+		drm_atomic_get_old_connector_state(state, conn);
+	struct intel_digital_connector_state *old_conn_state =
+		to_intel_digital_connector_state(old_state);
+	struct drm_crtc_state *crtc_state;
+
+	intel_hdcp_atomic_check(conn, old_state, new_state);
+
+	if (!new_state->crtc)
+		return 0;
+
+	crtc_state = drm_atomic_get_new_crtc_state(state, new_state->crtc);
+
+	/*
+	 * These properties are handled by fastset, and might not end
+	 * up in a modeset.
+	 */
+	if (new_conn_state->force_audio != old_conn_state->force_audio ||
+	    new_conn_state->broadcast_rgb != old_conn_state->broadcast_rgb ||
+	    new_conn_state->base.colorspace != old_conn_state->base.colorspace ||
+	    new_conn_state->base.picture_aspect_ratio != old_conn_state->base.picture_aspect_ratio ||
+	    new_conn_state->base.content_type != old_conn_state->base.content_type ||
+	    new_conn_state->base.scaling_mode != old_conn_state->base.scaling_mode ||
+	    new_conn_state->base.privacy_screen_sw_state != old_conn_state->base.privacy_screen_sw_state ||
+	    !drm_connector_atomic_hdr_metadata_equal(old_state, new_state))
+		crtc_state->mode_changed = true;
+
+	return 0;
+}
+
+/**
+ * intel_digital_connector_duplicate_state - duplicate connector state
+ * @connector: digital connector
+ *
+ * Allocates and returns a copy of the connector state (both common and
+ * digital connector specific) for the specified connector.
+ *
+ * Returns: The newly allocated connector state, or NULL on failure.
+ */
+struct drm_connector_state *
+intel_digital_connector_duplicate_state(struct drm_connector *connector)
+{
+	struct intel_digital_connector_state *state;
+
+	state = kmemdup(connector->state, sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return NULL;
+
+	__drm_atomic_helper_connector_duplicate_state(connector, &state->base);
+	return &state->base;
+}
+
+/**
+ * intel_connector_needs_modeset - check if connector needs a modeset
+ * @state: the atomic state corresponding to this modeset
+ * @connector: the connector
+ */
+bool
+intel_connector_needs_modeset(struct intel_atomic_state *state,
+			      struct drm_connector *connector)
+{
+	const struct drm_connector_state *old_conn_state, *new_conn_state;
+
+	old_conn_state = drm_atomic_get_old_connector_state(&state->base, connector);
+	new_conn_state = drm_atomic_get_new_connector_state(&state->base, connector);
+
+	return old_conn_state->crtc != new_conn_state->crtc ||
+	       (new_conn_state->crtc &&
+		drm_atomic_crtc_needs_modeset(drm_atomic_get_new_crtc_state(&state->base,
+									    new_conn_state->crtc)));
+}
+
+/**
+ * intel_any_crtc_needs_modeset - check if any CRTC needs a modeset
+ * @state: the atomic state corresponding to this modeset
+ *
+ * Returns true if any CRTC in @state needs a modeset.
+ */
+bool intel_any_crtc_needs_modeset(struct intel_atomic_state *state)
+{
+	struct intel_crtc *crtc;
+	struct intel_crtc_state *crtc_state;
+	int i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		if (intel_crtc_needs_modeset(crtc_state))
+			return true;
+	}
+
+	return false;
+}
+
+struct intel_digital_connector_state *
+intel_atomic_get_digital_connector_state(struct intel_atomic_state *state,
+					 struct intel_connector *connector)
+{
+	struct drm_connector_state *conn_state;
+
+	conn_state = drm_atomic_get_connector_state(&state->base,
+						    &connector->base);
+	if (IS_ERR(conn_state))
+		return ERR_CAST(conn_state);
+
+	return to_intel_digital_connector_state(conn_state);
+}
+
+/**
+ * intel_crtc_duplicate_state - duplicate crtc state
+ * @crtc: drm crtc
+ *
+ * Allocates and returns a copy of the crtc state (both common and
+ * Intel-specific) for the specified crtc.
+ *
+ * Returns: The newly allocated crtc state, or NULL on failure.
+ */
+struct drm_crtc_state *
+intel_crtc_duplicate_state(struct drm_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state = to_intel_crtc_state(crtc->state);
+	struct intel_crtc_state *crtc_state;
+
+	crtc_state = kmemdup(old_crtc_state, sizeof(*crtc_state), GFP_KERNEL);
+	if (!crtc_state)
+		return NULL;
+
+	__drm_atomic_helper_crtc_duplicate_state(crtc, &crtc_state->uapi);
+
+	/* copy color blobs */
+	if (crtc_state->hw.degamma_lut)
+		drm_property_blob_get(crtc_state->hw.degamma_lut);
+	if (crtc_state->hw.ctm)
+		drm_property_blob_get(crtc_state->hw.ctm);
+	if (crtc_state->hw.gamma_lut)
+		drm_property_blob_get(crtc_state->hw.gamma_lut);
+
+	crtc_state->update_pipe = false;
+	crtc_state->disable_lp_wm = false;
+	crtc_state->disable_cxsr = false;
+	crtc_state->update_wm_pre = false;
+	crtc_state->update_wm_post = false;
+	crtc_state->fifo_changed = false;
+	crtc_state->preload_luts = false;
+	crtc_state->inherited = false;
+	crtc_state->wm.need_postvbl_update = false;
+	crtc_state->do_async_flip = false;
+	crtc_state->fb_bits = 0;
+	crtc_state->update_planes = 0;
+	crtc_state->dsb = NULL;
+
+	return &crtc_state->uapi;
+}
+
+static void intel_crtc_put_color_blobs(struct intel_crtc_state *crtc_state)
+{
+	drm_property_blob_put(crtc_state->hw.degamma_lut);
+	drm_property_blob_put(crtc_state->hw.gamma_lut);
+	drm_property_blob_put(crtc_state->hw.ctm);
+}
+
+void intel_crtc_free_hw_state(struct intel_crtc_state *crtc_state)
+{
+	intel_crtc_put_color_blobs(crtc_state);
+}
+
+/**
+ * intel_crtc_destroy_state - destroy crtc state
+ * @crtc: drm crtc
+ * @state: the state to destroy
+ *
+ * Destroys the crtc state (both common and Intel-specific) for the
+ * specified crtc.
+ */
+void
+intel_crtc_destroy_state(struct drm_crtc *crtc,
+			 struct drm_crtc_state *state)
+{
+	struct intel_crtc_state *crtc_state = to_intel_crtc_state(state);
+
+	drm_WARN_ON(crtc->dev, crtc_state->dsb);
+
+	__drm_atomic_helper_crtc_destroy_state(&crtc_state->uapi);
+	intel_crtc_free_hw_state(crtc_state);
+	kfree(crtc_state);
+}
+
+struct drm_atomic_state *
+intel_atomic_state_alloc(struct drm_device *dev)
+{
+	struct intel_atomic_state *state = kzalloc(sizeof(*state), GFP_KERNEL);
+
+	if (!state || drm_atomic_state_init(dev, &state->base) < 0) {
+		kfree(state);
+		return NULL;
+	}
+
+	return &state->base;
+}
+
+void intel_atomic_state_free(struct drm_atomic_state *_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(_state);
+
+	drm_atomic_state_default_release(&state->base);
+	kfree(state->global_objs);
+
+	i915_sw_fence_fini(&state->commit_ready);
+
+	kfree(state);
+}
+
+void intel_atomic_state_clear(struct drm_atomic_state *s)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(s);
+
+	drm_atomic_state_default_clear(&state->base);
+	intel_atomic_clear_global_state(state);
+
+	state->dpll_set = state->modeset = false;
+}
+
+struct intel_crtc_state *
+intel_atomic_get_crtc_state(struct drm_atomic_state *state,
+			    struct intel_crtc *crtc)
+{
+	struct drm_crtc_state *crtc_state;
+	crtc_state = drm_atomic_get_crtc_state(state, &crtc->base);
+	if (IS_ERR(crtc_state))
+		return ERR_CAST(crtc_state);
+
+	return to_intel_crtc_state(crtc_state);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_atomic.h b/drivers/gpu/drm/i915/display/intel_atomic.h
new file mode 100644
index 000000000..e506f6a87
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_atomic.h
@@ -0,0 +1,55 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_ATOMIC_H__
+#define __INTEL_ATOMIC_H__
+
+#include <linux/types.h>
+
+struct drm_atomic_state;
+struct drm_connector;
+struct drm_connector_state;
+struct drm_crtc;
+struct drm_crtc_state;
+struct drm_device;
+struct drm_i915_private;
+struct drm_property;
+struct intel_atomic_state;
+struct intel_connector;
+struct intel_crtc;
+struct intel_crtc_state;
+
+int intel_digital_connector_atomic_get_property(struct drm_connector *connector,
+						const struct drm_connector_state *state,
+						struct drm_property *property,
+						u64 *val);
+int intel_digital_connector_atomic_set_property(struct drm_connector *connector,
+						struct drm_connector_state *state,
+						struct drm_property *property,
+						u64 val);
+int intel_digital_connector_atomic_check(struct drm_connector *conn,
+					 struct drm_atomic_state *state);
+struct drm_connector_state *
+intel_digital_connector_duplicate_state(struct drm_connector *connector);
+bool intel_connector_needs_modeset(struct intel_atomic_state *state,
+				   struct drm_connector *connector);
+bool intel_any_crtc_needs_modeset(struct intel_atomic_state *state);
+struct intel_digital_connector_state *
+intel_atomic_get_digital_connector_state(struct intel_atomic_state *state,
+					 struct intel_connector *connector);
+
+struct drm_crtc_state *intel_crtc_duplicate_state(struct drm_crtc *crtc);
+void intel_crtc_destroy_state(struct drm_crtc *crtc,
+			       struct drm_crtc_state *state);
+void intel_crtc_free_hw_state(struct intel_crtc_state *crtc_state);
+struct drm_atomic_state *intel_atomic_state_alloc(struct drm_device *dev);
+void intel_atomic_state_free(struct drm_atomic_state *state);
+void intel_atomic_state_clear(struct drm_atomic_state *state);
+
+struct intel_crtc_state *
+intel_atomic_get_crtc_state(struct drm_atomic_state *state,
+			    struct intel_crtc *crtc);
+
+#endif /* __INTEL_ATOMIC_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_atomic_plane.c b/drivers/gpu/drm/i915/display/intel_atomic_plane.c
new file mode 100644
index 000000000..82826454b
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_atomic_plane.c
@@ -0,0 +1,1120 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+/**
+ * DOC: atomic plane helpers
+ *
+ * The functions here are used by the atomic plane helper functions to
+ * implement legacy plane updates (i.e., drm_plane->update_plane() and
+ * drm_plane->disable_plane()).  This allows plane updates to use the
+ * atomic state infrastructure and perform plane updates as separate
+ * prepare/check/commit/cleanup steps.
+ */
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_fourcc.h>
+
+#include "gt/intel_rps.h"
+
+#include "intel_atomic_plane.h"
+#include "intel_cdclk.h"
+#include "intel_display_trace.h"
+#include "intel_display_types.h"
+#include "intel_fb.h"
+#include "intel_fb_pin.h"
+#include "intel_sprite.h"
+#include "skl_scaler.h"
+#include "skl_watermark.h"
+
+static void intel_plane_state_reset(struct intel_plane_state *plane_state,
+				    struct intel_plane *plane)
+{
+	memset(plane_state, 0, sizeof(*plane_state));
+
+	__drm_atomic_helper_plane_state_reset(&plane_state->uapi, &plane->base);
+
+	plane_state->scaler_id = -1;
+}
+
+struct intel_plane *intel_plane_alloc(void)
+{
+	struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+
+	plane = kzalloc(sizeof(*plane), GFP_KERNEL);
+	if (!plane)
+		return ERR_PTR(-ENOMEM);
+
+	plane_state = kzalloc(sizeof(*plane_state), GFP_KERNEL);
+	if (!plane_state) {
+		kfree(plane);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	intel_plane_state_reset(plane_state, plane);
+
+	plane->base.state = &plane_state->uapi;
+
+	return plane;
+}
+
+void intel_plane_free(struct intel_plane *plane)
+{
+	intel_plane_destroy_state(&plane->base, plane->base.state);
+	kfree(plane);
+}
+
+/**
+ * intel_plane_duplicate_state - duplicate plane state
+ * @plane: drm plane
+ *
+ * Allocates and returns a copy of the plane state (both common and
+ * Intel-specific) for the specified plane.
+ *
+ * Returns: The newly allocated plane state, or NULL on failure.
+ */
+struct drm_plane_state *
+intel_plane_duplicate_state(struct drm_plane *plane)
+{
+	struct intel_plane_state *intel_state;
+
+	intel_state = to_intel_plane_state(plane->state);
+	intel_state = kmemdup(intel_state, sizeof(*intel_state), GFP_KERNEL);
+
+	if (!intel_state)
+		return NULL;
+
+	__drm_atomic_helper_plane_duplicate_state(plane, &intel_state->uapi);
+
+	intel_state->ggtt_vma = NULL;
+	intel_state->dpt_vma = NULL;
+	intel_state->flags = 0;
+
+	/* add reference to fb */
+	if (intel_state->hw.fb)
+		drm_framebuffer_get(intel_state->hw.fb);
+
+	return &intel_state->uapi;
+}
+
+/**
+ * intel_plane_destroy_state - destroy plane state
+ * @plane: drm plane
+ * @state: state object to destroy
+ *
+ * Destroys the plane state (both common and Intel-specific) for the
+ * specified plane.
+ */
+void
+intel_plane_destroy_state(struct drm_plane *plane,
+			  struct drm_plane_state *state)
+{
+	struct intel_plane_state *plane_state = to_intel_plane_state(state);
+
+	drm_WARN_ON(plane->dev, plane_state->ggtt_vma);
+	drm_WARN_ON(plane->dev, plane_state->dpt_vma);
+
+	__drm_atomic_helper_plane_destroy_state(&plane_state->uapi);
+	if (plane_state->hw.fb)
+		drm_framebuffer_put(plane_state->hw.fb);
+	kfree(plane_state);
+}
+
+unsigned int intel_adjusted_rate(const struct drm_rect *src,
+				 const struct drm_rect *dst,
+				 unsigned int rate)
+{
+	unsigned int src_w, src_h, dst_w, dst_h;
+
+	src_w = drm_rect_width(src) >> 16;
+	src_h = drm_rect_height(src) >> 16;
+	dst_w = drm_rect_width(dst);
+	dst_h = drm_rect_height(dst);
+
+	/* Downscaling limits the maximum pixel rate */
+	dst_w = min(src_w, dst_w);
+	dst_h = min(src_h, dst_h);
+
+	return DIV_ROUND_UP_ULL(mul_u32_u32(rate, src_w * src_h),
+				dst_w * dst_h);
+}
+
+unsigned int intel_plane_pixel_rate(const struct intel_crtc_state *crtc_state,
+				    const struct intel_plane_state *plane_state)
+{
+	/*
+	 * Note we don't check for plane visibility here as
+	 * we want to use this when calculating the cursor
+	 * watermarks even if the cursor is fully offscreen.
+	 * That depends on the src/dst rectangles being
+	 * correctly populated whenever the watermark code
+	 * considers the cursor to be visible, whether or not
+	 * it is actually visible.
+	 *
+	 * See: intel_wm_plane_visible() and intel_check_cursor()
+	 */
+
+	return intel_adjusted_rate(&plane_state->uapi.src,
+				   &plane_state->uapi.dst,
+				   crtc_state->pixel_rate);
+}
+
+unsigned int intel_plane_data_rate(const struct intel_crtc_state *crtc_state,
+				   const struct intel_plane_state *plane_state,
+				   int color_plane)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+
+	if (!plane_state->uapi.visible)
+		return 0;
+
+	return intel_plane_pixel_rate(crtc_state, plane_state) *
+		fb->format->cpp[color_plane];
+}
+
+static bool
+use_min_ddb(const struct intel_crtc_state *crtc_state,
+	    struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	return DISPLAY_VER(i915) >= 13 &&
+	       crtc_state->uapi.async_flip &&
+	       plane->async_flip;
+}
+
+static unsigned int
+intel_plane_relative_data_rate(const struct intel_crtc_state *crtc_state,
+			       const struct intel_plane_state *plane_state,
+			       int color_plane)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int width, height;
+
+	if (plane->id == PLANE_CURSOR)
+		return 0;
+
+	if (!plane_state->uapi.visible)
+		return 0;
+
+	/*
+	 * We calculate extra ddb based on ratio plane rate/total data rate
+	 * in case, in some cases we should not allocate extra ddb for the plane,
+	 * so do not count its data rate, if this is the case.
+	 */
+	if (use_min_ddb(crtc_state, plane))
+		return 0;
+
+	/*
+	 * Src coordinates are already rotated by 270 degrees for
+	 * the 90/270 degree plane rotation cases (to match the
+	 * GTT mapping), hence no need to account for rotation here.
+	 */
+	width = drm_rect_width(&plane_state->uapi.src) >> 16;
+	height = drm_rect_height(&plane_state->uapi.src) >> 16;
+
+	/* UV plane does 1/2 pixel sub-sampling */
+	if (color_plane == 1) {
+		width /= 2;
+		height /= 2;
+	}
+
+	return width * height * fb->format->cpp[color_plane];
+}
+
+int intel_plane_calc_min_cdclk(struct intel_atomic_state *state,
+			       struct intel_plane *plane,
+			       bool *need_cdclk_calc)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct intel_plane_state *plane_state =
+		intel_atomic_get_new_plane_state(state, plane);
+	struct intel_crtc *crtc = to_intel_crtc(plane_state->hw.crtc);
+	const struct intel_cdclk_state *cdclk_state;
+	const struct intel_crtc_state *old_crtc_state;
+	struct intel_crtc_state *new_crtc_state;
+
+	if (!plane_state->uapi.visible || !plane->min_cdclk)
+		return 0;
+
+	old_crtc_state = intel_atomic_get_old_crtc_state(state, crtc);
+	new_crtc_state = intel_atomic_get_new_crtc_state(state, crtc);
+
+	new_crtc_state->min_cdclk[plane->id] =
+		plane->min_cdclk(new_crtc_state, plane_state);
+
+	/*
+	 * No need to check against the cdclk state if
+	 * the min cdclk for the plane doesn't increase.
+	 *
+	 * Ie. we only ever increase the cdclk due to plane
+	 * requirements. This can reduce back and forth
+	 * display blinking due to constant cdclk changes.
+	 */
+	if (new_crtc_state->min_cdclk[plane->id] <=
+	    old_crtc_state->min_cdclk[plane->id])
+		return 0;
+
+	cdclk_state = intel_atomic_get_cdclk_state(state);
+	if (IS_ERR(cdclk_state))
+		return PTR_ERR(cdclk_state);
+
+	/*
+	 * No need to recalculate the cdclk state if
+	 * the min cdclk for the pipe doesn't increase.
+	 *
+	 * Ie. we only ever increase the cdclk due to plane
+	 * requirements. This can reduce back and forth
+	 * display blinking due to constant cdclk changes.
+	 */
+	if (new_crtc_state->min_cdclk[plane->id] <=
+	    cdclk_state->min_cdclk[crtc->pipe])
+		return 0;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "[PLANE:%d:%s] min cdclk (%d kHz) > [CRTC:%d:%s] min cdclk (%d kHz)\n",
+		    plane->base.base.id, plane->base.name,
+		    new_crtc_state->min_cdclk[plane->id],
+		    crtc->base.base.id, crtc->base.name,
+		    cdclk_state->min_cdclk[crtc->pipe]);
+	*need_cdclk_calc = true;
+
+	return 0;
+}
+
+static void intel_plane_clear_hw_state(struct intel_plane_state *plane_state)
+{
+	if (plane_state->hw.fb)
+		drm_framebuffer_put(plane_state->hw.fb);
+
+	memset(&plane_state->hw, 0, sizeof(plane_state->hw));
+}
+
+void intel_plane_copy_uapi_to_hw_state(struct intel_plane_state *plane_state,
+				       const struct intel_plane_state *from_plane_state,
+				       struct intel_crtc *crtc)
+{
+	intel_plane_clear_hw_state(plane_state);
+
+	/*
+	 * For the bigjoiner slave uapi.crtc will point at
+	 * the master crtc. So we explicitly assign the right
+	 * slave crtc to hw.crtc. uapi.crtc!=NULL simply indicates
+	 * the plane is logically enabled on the uapi level.
+	 */
+	plane_state->hw.crtc = from_plane_state->uapi.crtc ? &crtc->base : NULL;
+
+	plane_state->hw.fb = from_plane_state->uapi.fb;
+	if (plane_state->hw.fb)
+		drm_framebuffer_get(plane_state->hw.fb);
+
+	plane_state->hw.alpha = from_plane_state->uapi.alpha;
+	plane_state->hw.pixel_blend_mode =
+		from_plane_state->uapi.pixel_blend_mode;
+	plane_state->hw.rotation = from_plane_state->uapi.rotation;
+	plane_state->hw.color_encoding = from_plane_state->uapi.color_encoding;
+	plane_state->hw.color_range = from_plane_state->uapi.color_range;
+	plane_state->hw.scaling_filter = from_plane_state->uapi.scaling_filter;
+
+	plane_state->uapi.src = drm_plane_state_src(&from_plane_state->uapi);
+	plane_state->uapi.dst = drm_plane_state_dest(&from_plane_state->uapi);
+}
+
+void intel_plane_copy_hw_state(struct intel_plane_state *plane_state,
+			       const struct intel_plane_state *from_plane_state)
+{
+	intel_plane_clear_hw_state(plane_state);
+
+	memcpy(&plane_state->hw, &from_plane_state->hw,
+	       sizeof(plane_state->hw));
+
+	if (plane_state->hw.fb)
+		drm_framebuffer_get(plane_state->hw.fb);
+}
+
+void intel_plane_set_invisible(struct intel_crtc_state *crtc_state,
+			       struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+
+	crtc_state->active_planes &= ~BIT(plane->id);
+	crtc_state->scaled_planes &= ~BIT(plane->id);
+	crtc_state->nv12_planes &= ~BIT(plane->id);
+	crtc_state->c8_planes &= ~BIT(plane->id);
+	crtc_state->data_rate[plane->id] = 0;
+	crtc_state->data_rate_y[plane->id] = 0;
+	crtc_state->rel_data_rate[plane->id] = 0;
+	crtc_state->rel_data_rate_y[plane->id] = 0;
+	crtc_state->min_cdclk[plane->id] = 0;
+
+	plane_state->uapi.visible = false;
+}
+
+/* FIXME nuke when all wm code is atomic */
+static bool intel_wm_need_update(const struct intel_plane_state *cur,
+				 struct intel_plane_state *new)
+{
+	/* Update watermarks on tiling or size changes. */
+	if (new->uapi.visible != cur->uapi.visible)
+		return true;
+
+	if (!cur->hw.fb || !new->hw.fb)
+		return false;
+
+	if (cur->hw.fb->modifier != new->hw.fb->modifier ||
+	    cur->hw.rotation != new->hw.rotation ||
+	    drm_rect_width(&new->uapi.src) != drm_rect_width(&cur->uapi.src) ||
+	    drm_rect_height(&new->uapi.src) != drm_rect_height(&cur->uapi.src) ||
+	    drm_rect_width(&new->uapi.dst) != drm_rect_width(&cur->uapi.dst) ||
+	    drm_rect_height(&new->uapi.dst) != drm_rect_height(&cur->uapi.dst))
+		return true;
+
+	return false;
+}
+
+static bool intel_plane_is_scaled(const struct intel_plane_state *plane_state)
+{
+	int src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+	int src_h = drm_rect_height(&plane_state->uapi.src) >> 16;
+	int dst_w = drm_rect_width(&plane_state->uapi.dst);
+	int dst_h = drm_rect_height(&plane_state->uapi.dst);
+
+	return src_w != dst_w || src_h != dst_h;
+}
+
+static bool intel_plane_do_async_flip(struct intel_plane *plane,
+				      const struct intel_crtc_state *old_crtc_state,
+				      const struct intel_crtc_state *new_crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	if (!plane->async_flip)
+		return false;
+
+	if (!new_crtc_state->uapi.async_flip)
+		return false;
+
+	/*
+	 * In platforms after DISPLAY13, we might need to override
+	 * first async flip in order to change watermark levels
+	 * as part of optimization.
+	 * So for those, we are checking if this is a first async flip.
+	 * For platforms earlier than DISPLAY13 we always do async flip.
+	 */
+	return DISPLAY_VER(i915) < 13 || old_crtc_state->uapi.async_flip;
+}
+
+static int intel_plane_atomic_calc_changes(const struct intel_crtc_state *old_crtc_state,
+					   struct intel_crtc_state *new_crtc_state,
+					   const struct intel_plane_state *old_plane_state,
+					   struct intel_plane_state *new_plane_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct intel_plane *plane = to_intel_plane(new_plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	bool mode_changed = intel_crtc_needs_modeset(new_crtc_state);
+	bool was_crtc_enabled = old_crtc_state->hw.active;
+	bool is_crtc_enabled = new_crtc_state->hw.active;
+	bool turn_off, turn_on, visible, was_visible;
+	int ret;
+
+	if (DISPLAY_VER(dev_priv) >= 9 && plane->id != PLANE_CURSOR) {
+		ret = skl_update_scaler_plane(new_crtc_state, new_plane_state);
+		if (ret)
+			return ret;
+	}
+
+	was_visible = old_plane_state->uapi.visible;
+	visible = new_plane_state->uapi.visible;
+
+	if (!was_crtc_enabled && drm_WARN_ON(&dev_priv->drm, was_visible))
+		was_visible = false;
+
+	/*
+	 * Visibility is calculated as if the crtc was on, but
+	 * after scaler setup everything depends on it being off
+	 * when the crtc isn't active.
+	 *
+	 * FIXME this is wrong for watermarks. Watermarks should also
+	 * be computed as if the pipe would be active. Perhaps move
+	 * per-plane wm computation to the .check_plane() hook, and
+	 * only combine the results from all planes in the current place?
+	 */
+	if (!is_crtc_enabled) {
+		intel_plane_set_invisible(new_crtc_state, new_plane_state);
+		visible = false;
+	}
+
+	if (!was_visible && !visible)
+		return 0;
+
+	turn_off = was_visible && (!visible || mode_changed);
+	turn_on = visible && (!was_visible || mode_changed);
+
+	drm_dbg_atomic(&dev_priv->drm,
+		       "[CRTC:%d:%s] with [PLANE:%d:%s] visible %i -> %i, off %i, on %i, ms %i\n",
+		       crtc->base.base.id, crtc->base.name,
+		       plane->base.base.id, plane->base.name,
+		       was_visible, visible,
+		       turn_off, turn_on, mode_changed);
+
+	if (turn_on) {
+		if (DISPLAY_VER(dev_priv) < 5 && !IS_G4X(dev_priv))
+			new_crtc_state->update_wm_pre = true;
+
+		/* must disable cxsr around plane enable/disable */
+		if (plane->id != PLANE_CURSOR)
+			new_crtc_state->disable_cxsr = true;
+	} else if (turn_off) {
+		if (DISPLAY_VER(dev_priv) < 5 && !IS_G4X(dev_priv))
+			new_crtc_state->update_wm_post = true;
+
+		/* must disable cxsr around plane enable/disable */
+		if (plane->id != PLANE_CURSOR)
+			new_crtc_state->disable_cxsr = true;
+	} else if (intel_wm_need_update(old_plane_state, new_plane_state)) {
+		if (DISPLAY_VER(dev_priv) < 5 && !IS_G4X(dev_priv)) {
+			/* FIXME bollocks */
+			new_crtc_state->update_wm_pre = true;
+			new_crtc_state->update_wm_post = true;
+		}
+	}
+
+	if (visible || was_visible)
+		new_crtc_state->fb_bits |= plane->frontbuffer_bit;
+
+	/*
+	 * ILK/SNB DVSACNTR/Sprite Enable
+	 * IVB SPR_CTL/Sprite Enable
+	 * "When in Self Refresh Big FIFO mode, a write to enable the
+	 *  plane will be internally buffered and delayed while Big FIFO
+	 *  mode is exiting."
+	 *
+	 * Which means that enabling the sprite can take an extra frame
+	 * when we start in big FIFO mode (LP1+). Thus we need to drop
+	 * down to LP0 and wait for vblank in order to make sure the
+	 * sprite gets enabled on the next vblank after the register write.
+	 * Doing otherwise would risk enabling the sprite one frame after
+	 * we've already signalled flip completion. We can resume LP1+
+	 * once the sprite has been enabled.
+	 *
+	 *
+	 * WaCxSRDisabledForSpriteScaling:ivb
+	 * IVB SPR_SCALE/Scaling Enable
+	 * "Low Power watermarks must be disabled for at least one
+	 *  frame before enabling sprite scaling, and kept disabled
+	 *  until sprite scaling is disabled."
+	 *
+	 * ILK/SNB DVSASCALE/Scaling Enable
+	 * "When in Self Refresh Big FIFO mode, scaling enable will be
+	 *  masked off while Big FIFO mode is exiting."
+	 *
+	 * Despite the w/a only being listed for IVB we assume that
+	 * the ILK/SNB note has similar ramifications, hence we apply
+	 * the w/a on all three platforms.
+	 *
+	 * With experimental results seems this is needed also for primary
+	 * plane, not only sprite plane.
+	 */
+	if (plane->id != PLANE_CURSOR &&
+	    (IS_IRONLAKE(dev_priv) || IS_SANDYBRIDGE(dev_priv) ||
+	     IS_IVYBRIDGE(dev_priv)) &&
+	    (turn_on || (!intel_plane_is_scaled(old_plane_state) &&
+			 intel_plane_is_scaled(new_plane_state))))
+		new_crtc_state->disable_lp_wm = true;
+
+	if (intel_plane_do_async_flip(plane, old_crtc_state, new_crtc_state))
+		new_crtc_state->do_async_flip = true;
+
+	return 0;
+}
+
+int intel_plane_atomic_check_with_state(const struct intel_crtc_state *old_crtc_state,
+					struct intel_crtc_state *new_crtc_state,
+					const struct intel_plane_state *old_plane_state,
+					struct intel_plane_state *new_plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(new_plane_state->uapi.plane);
+	const struct drm_framebuffer *fb = new_plane_state->hw.fb;
+	int ret;
+
+	intel_plane_set_invisible(new_crtc_state, new_plane_state);
+	new_crtc_state->enabled_planes &= ~BIT(plane->id);
+
+	if (!new_plane_state->hw.crtc && !old_plane_state->hw.crtc)
+		return 0;
+
+	ret = plane->check_plane(new_crtc_state, new_plane_state);
+	if (ret)
+		return ret;
+
+	if (fb)
+		new_crtc_state->enabled_planes |= BIT(plane->id);
+
+	/* FIXME pre-g4x don't work like this */
+	if (new_plane_state->uapi.visible)
+		new_crtc_state->active_planes |= BIT(plane->id);
+
+	if (new_plane_state->uapi.visible &&
+	    intel_plane_is_scaled(new_plane_state))
+		new_crtc_state->scaled_planes |= BIT(plane->id);
+
+	if (new_plane_state->uapi.visible &&
+	    intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier))
+		new_crtc_state->nv12_planes |= BIT(plane->id);
+
+	if (new_plane_state->uapi.visible &&
+	    fb->format->format == DRM_FORMAT_C8)
+		new_crtc_state->c8_planes |= BIT(plane->id);
+
+	if (new_plane_state->uapi.visible || old_plane_state->uapi.visible)
+		new_crtc_state->update_planes |= BIT(plane->id);
+
+	if (new_plane_state->uapi.visible &&
+	    intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier)) {
+		new_crtc_state->data_rate_y[plane->id] =
+			intel_plane_data_rate(new_crtc_state, new_plane_state, 0);
+		new_crtc_state->data_rate[plane->id] =
+			intel_plane_data_rate(new_crtc_state, new_plane_state, 1);
+
+		new_crtc_state->rel_data_rate_y[plane->id] =
+			intel_plane_relative_data_rate(new_crtc_state,
+						       new_plane_state, 0);
+		new_crtc_state->rel_data_rate[plane->id] =
+			intel_plane_relative_data_rate(new_crtc_state,
+						       new_plane_state, 1);
+	} else if (new_plane_state->uapi.visible) {
+		new_crtc_state->data_rate[plane->id] =
+			intel_plane_data_rate(new_crtc_state, new_plane_state, 0);
+
+		new_crtc_state->rel_data_rate[plane->id] =
+			intel_plane_relative_data_rate(new_crtc_state,
+						       new_plane_state, 0);
+	}
+
+	return intel_plane_atomic_calc_changes(old_crtc_state, new_crtc_state,
+					       old_plane_state, new_plane_state);
+}
+
+static struct intel_plane *
+intel_crtc_get_plane(struct intel_crtc *crtc, enum plane_id plane_id)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_plane *plane;
+
+	for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) {
+		if (plane->id == plane_id)
+			return plane;
+	}
+
+	return NULL;
+}
+
+int intel_plane_atomic_check(struct intel_atomic_state *state,
+			     struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_plane_state *new_plane_state =
+		intel_atomic_get_new_plane_state(state, plane);
+	const struct intel_plane_state *old_plane_state =
+		intel_atomic_get_old_plane_state(state, plane);
+	const struct intel_plane_state *new_master_plane_state;
+	struct intel_crtc *crtc = intel_crtc_for_pipe(i915, plane->pipe);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (new_crtc_state && intel_crtc_is_bigjoiner_slave(new_crtc_state)) {
+		struct intel_crtc *master_crtc =
+			intel_master_crtc(new_crtc_state);
+		struct intel_plane *master_plane =
+			intel_crtc_get_plane(master_crtc, plane->id);
+
+		new_master_plane_state =
+			intel_atomic_get_new_plane_state(state, master_plane);
+	} else {
+		new_master_plane_state = new_plane_state;
+	}
+
+	intel_plane_copy_uapi_to_hw_state(new_plane_state,
+					  new_master_plane_state,
+					  crtc);
+
+	new_plane_state->uapi.visible = false;
+	if (!new_crtc_state)
+		return 0;
+
+	return intel_plane_atomic_check_with_state(old_crtc_state,
+						   new_crtc_state,
+						   old_plane_state,
+						   new_plane_state);
+}
+
+static struct intel_plane *
+skl_next_plane_to_commit(struct intel_atomic_state *state,
+			 struct intel_crtc *crtc,
+			 struct skl_ddb_entry ddb[I915_MAX_PLANES],
+			 struct skl_ddb_entry ddb_y[I915_MAX_PLANES],
+			 unsigned int *update_mask)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	if (*update_mask == 0)
+		return NULL;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		enum plane_id plane_id = plane->id;
+
+		if (crtc->pipe != plane->pipe ||
+		    !(*update_mask & BIT(plane_id)))
+			continue;
+
+		if (skl_ddb_allocation_overlaps(&crtc_state->wm.skl.plane_ddb[plane_id],
+						ddb, I915_MAX_PLANES, plane_id) ||
+		    skl_ddb_allocation_overlaps(&crtc_state->wm.skl.plane_ddb_y[plane_id],
+						ddb_y, I915_MAX_PLANES, plane_id))
+			continue;
+
+		*update_mask &= ~BIT(plane_id);
+		ddb[plane_id] = crtc_state->wm.skl.plane_ddb[plane_id];
+		ddb_y[plane_id] = crtc_state->wm.skl.plane_ddb_y[plane_id];
+
+		return plane;
+	}
+
+	/* should never happen */
+	drm_WARN_ON(state->base.dev, 1);
+
+	return NULL;
+}
+
+void intel_plane_update_noarm(struct intel_plane *plane,
+			      const struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	trace_intel_plane_update_noarm(&plane->base, crtc);
+
+	if (plane->update_noarm)
+		plane->update_noarm(plane, crtc_state, plane_state);
+}
+
+void intel_plane_update_arm(struct intel_plane *plane,
+			    const struct intel_crtc_state *crtc_state,
+			    const struct intel_plane_state *plane_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	trace_intel_plane_update_arm(&plane->base, crtc);
+
+	if (crtc_state->do_async_flip && plane->async_flip)
+		plane->async_flip(plane, crtc_state, plane_state, true);
+	else
+		plane->update_arm(plane, crtc_state, plane_state);
+}
+
+void intel_plane_disable_arm(struct intel_plane *plane,
+			     const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	trace_intel_plane_disable_arm(&plane->base, crtc);
+	plane->disable_arm(plane, crtc_state);
+}
+
+void intel_crtc_planes_update_noarm(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	u32 update_mask = new_crtc_state->update_planes;
+	struct intel_plane_state *new_plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	if (new_crtc_state->do_async_flip)
+		return;
+
+	/*
+	 * Since we only write non-arming registers here,
+	 * the order does not matter even for skl+.
+	 */
+	for_each_new_intel_plane_in_state(state, plane, new_plane_state, i) {
+		if (crtc->pipe != plane->pipe ||
+		    !(update_mask & BIT(plane->id)))
+			continue;
+
+		/* TODO: for mailbox updates this should be skipped */
+		if (new_plane_state->uapi.visible ||
+		    new_plane_state->planar_slave)
+			intel_plane_update_noarm(plane, new_crtc_state, new_plane_state);
+	}
+}
+
+static void skl_crtc_planes_update_arm(struct intel_atomic_state *state,
+				       struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct skl_ddb_entry ddb[I915_MAX_PLANES];
+	struct skl_ddb_entry ddb_y[I915_MAX_PLANES];
+	u32 update_mask = new_crtc_state->update_planes;
+	struct intel_plane *plane;
+
+	memcpy(ddb, old_crtc_state->wm.skl.plane_ddb,
+	       sizeof(old_crtc_state->wm.skl.plane_ddb));
+	memcpy(ddb_y, old_crtc_state->wm.skl.plane_ddb_y,
+	       sizeof(old_crtc_state->wm.skl.plane_ddb_y));
+
+	while ((plane = skl_next_plane_to_commit(state, crtc, ddb, ddb_y, &update_mask))) {
+		struct intel_plane_state *new_plane_state =
+			intel_atomic_get_new_plane_state(state, plane);
+
+		/*
+		 * TODO: for mailbox updates intel_plane_update_noarm()
+		 * would have to be called here as well.
+		 */
+		if (new_plane_state->uapi.visible ||
+		    new_plane_state->planar_slave)
+			intel_plane_update_arm(plane, new_crtc_state, new_plane_state);
+		else
+			intel_plane_disable_arm(plane, new_crtc_state);
+	}
+}
+
+static void i9xx_crtc_planes_update_arm(struct intel_atomic_state *state,
+					struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	u32 update_mask = new_crtc_state->update_planes;
+	struct intel_plane_state *new_plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, new_plane_state, i) {
+		if (crtc->pipe != plane->pipe ||
+		    !(update_mask & BIT(plane->id)))
+			continue;
+
+		/*
+		 * TODO: for mailbox updates intel_plane_update_noarm()
+		 * would have to be called here as well.
+		 */
+		if (new_plane_state->uapi.visible)
+			intel_plane_update_arm(plane, new_crtc_state, new_plane_state);
+		else
+			intel_plane_disable_arm(plane, new_crtc_state);
+	}
+}
+
+void intel_crtc_planes_update_arm(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+
+	if (DISPLAY_VER(i915) >= 9)
+		skl_crtc_planes_update_arm(state, crtc);
+	else
+		i9xx_crtc_planes_update_arm(state, crtc);
+}
+
+int intel_atomic_plane_check_clipping(struct intel_plane_state *plane_state,
+				      struct intel_crtc_state *crtc_state,
+				      int min_scale, int max_scale,
+				      bool can_position)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	struct drm_framebuffer *fb = plane_state->hw.fb;
+	struct drm_rect *src = &plane_state->uapi.src;
+	struct drm_rect *dst = &plane_state->uapi.dst;
+	const struct drm_rect *clip = &crtc_state->pipe_src;
+	unsigned int rotation = plane_state->hw.rotation;
+	int hscale, vscale;
+
+	if (!fb) {
+		plane_state->uapi.visible = false;
+		return 0;
+	}
+
+	drm_rect_rotate(src, fb->width << 16, fb->height << 16, rotation);
+
+	/* Check scaling */
+	hscale = drm_rect_calc_hscale(src, dst, min_scale, max_scale);
+	vscale = drm_rect_calc_vscale(src, dst, min_scale, max_scale);
+	if (hscale < 0 || vscale < 0) {
+		drm_dbg_kms(&i915->drm, "Invalid scaling of plane\n");
+		drm_rect_debug_print("src: ", src, true);
+		drm_rect_debug_print("dst: ", dst, false);
+		return -ERANGE;
+	}
+
+	/*
+	 * FIXME: This might need further adjustment for seamless scaling
+	 * with phase information, for the 2p2 and 2p1 scenarios.
+	 */
+	plane_state->uapi.visible = drm_rect_clip_scaled(src, dst, clip);
+
+	drm_rect_rotate_inv(src, fb->width << 16, fb->height << 16, rotation);
+
+	if (!can_position && plane_state->uapi.visible &&
+	    !drm_rect_equals(dst, clip)) {
+		drm_dbg_kms(&i915->drm, "Plane must cover entire CRTC\n");
+		drm_rect_debug_print("dst: ", dst, false);
+		drm_rect_debug_print("clip: ", clip, false);
+		return -EINVAL;
+	}
+
+	/* final plane coordinates will be relative to the plane's pipe */
+	drm_rect_translate(dst, -clip->x1, -clip->y1);
+
+	return 0;
+}
+
+struct wait_rps_boost {
+	struct wait_queue_entry wait;
+
+	struct drm_crtc *crtc;
+	struct i915_request *request;
+};
+
+static int do_rps_boost(struct wait_queue_entry *_wait,
+			unsigned mode, int sync, void *key)
+{
+	struct wait_rps_boost *wait = container_of(_wait, typeof(*wait), wait);
+	struct i915_request *rq = wait->request;
+
+	/*
+	 * If we missed the vblank, but the request is already running it
+	 * is reasonable to assume that it will complete before the next
+	 * vblank without our intervention, so leave RPS alone.
+	 */
+	if (!i915_request_started(rq))
+		intel_rps_boost(rq);
+	i915_request_put(rq);
+
+	drm_crtc_vblank_put(wait->crtc);
+
+	list_del(&wait->wait.entry);
+	kfree(wait);
+	return 1;
+}
+
+static void add_rps_boost_after_vblank(struct drm_crtc *crtc,
+				       struct dma_fence *fence)
+{
+	struct wait_rps_boost *wait;
+
+	if (!dma_fence_is_i915(fence))
+		return;
+
+	if (DISPLAY_VER(to_i915(crtc->dev)) < 6)
+		return;
+
+	if (drm_crtc_vblank_get(crtc))
+		return;
+
+	wait = kmalloc(sizeof(*wait), GFP_KERNEL);
+	if (!wait) {
+		drm_crtc_vblank_put(crtc);
+		return;
+	}
+
+	wait->request = to_request(dma_fence_get(fence));
+	wait->crtc = crtc;
+
+	wait->wait.func = do_rps_boost;
+	wait->wait.flags = 0;
+
+	add_wait_queue(drm_crtc_vblank_waitqueue(crtc), &wait->wait);
+}
+
+/**
+ * intel_prepare_plane_fb - Prepare fb for usage on plane
+ * @_plane: drm plane to prepare for
+ * @_new_plane_state: the plane state being prepared
+ *
+ * Prepares a framebuffer for usage on a display plane.  Generally this
+ * involves pinning the underlying object and updating the frontbuffer tracking
+ * bits.  Some older platforms need special physical address handling for
+ * cursor planes.
+ *
+ * Returns 0 on success, negative error code on failure.
+ */
+static int
+intel_prepare_plane_fb(struct drm_plane *_plane,
+		       struct drm_plane_state *_new_plane_state)
+{
+	struct i915_sched_attr attr = { .priority = I915_PRIORITY_DISPLAY };
+	struct intel_plane *plane = to_intel_plane(_plane);
+	struct intel_plane_state *new_plane_state =
+		to_intel_plane_state(_new_plane_state);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(new_plane_state->uapi.state);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct intel_plane_state *old_plane_state =
+		intel_atomic_get_old_plane_state(state, plane);
+	struct drm_i915_gem_object *obj = intel_fb_obj(new_plane_state->hw.fb);
+	struct drm_i915_gem_object *old_obj = intel_fb_obj(old_plane_state->hw.fb);
+	int ret;
+
+	if (old_obj) {
+		const struct intel_crtc_state *new_crtc_state =
+			intel_atomic_get_new_crtc_state(state,
+							to_intel_crtc(old_plane_state->hw.crtc));
+
+		/* Big Hammer, we also need to ensure that any pending
+		 * MI_WAIT_FOR_EVENT inside a user batch buffer on the
+		 * current scanout is retired before unpinning the old
+		 * framebuffer. Note that we rely on userspace rendering
+		 * into the buffer attached to the pipe they are waiting
+		 * on. If not, userspace generates a GPU hang with IPEHR
+		 * point to the MI_WAIT_FOR_EVENT.
+		 *
+		 * This should only fail upon a hung GPU, in which case we
+		 * can safely continue.
+		 */
+		if (new_crtc_state && intel_crtc_needs_modeset(new_crtc_state)) {
+			ret = i915_sw_fence_await_reservation(&state->commit_ready,
+							      old_obj->base.resv, NULL,
+							      false, 0,
+							      GFP_KERNEL);
+			if (ret < 0)
+				return ret;
+		}
+	}
+
+	if (new_plane_state->uapi.fence) { /* explicit fencing */
+		i915_gem_fence_wait_priority(new_plane_state->uapi.fence,
+					     &attr);
+		ret = i915_sw_fence_await_dma_fence(&state->commit_ready,
+						    new_plane_state->uapi.fence,
+						    i915_fence_timeout(dev_priv),
+						    GFP_KERNEL);
+		if (ret < 0)
+			return ret;
+	}
+
+	if (!obj)
+		return 0;
+
+
+	ret = intel_plane_pin_fb(new_plane_state);
+	if (ret)
+		return ret;
+
+	i915_gem_object_wait_priority(obj, 0, &attr);
+
+	if (!new_plane_state->uapi.fence) { /* implicit fencing */
+		struct dma_resv_iter cursor;
+		struct dma_fence *fence;
+
+		ret = i915_sw_fence_await_reservation(&state->commit_ready,
+						      obj->base.resv, NULL,
+						      false,
+						      i915_fence_timeout(dev_priv),
+						      GFP_KERNEL);
+		if (ret < 0)
+			goto unpin_fb;
+
+		dma_resv_iter_begin(&cursor, obj->base.resv,
+				    DMA_RESV_USAGE_WRITE);
+		dma_resv_for_each_fence_unlocked(&cursor, fence) {
+			add_rps_boost_after_vblank(new_plane_state->hw.crtc,
+						   fence);
+		}
+		dma_resv_iter_end(&cursor);
+	} else {
+		add_rps_boost_after_vblank(new_plane_state->hw.crtc,
+					   new_plane_state->uapi.fence);
+	}
+
+	/*
+	 * We declare pageflips to be interactive and so merit a small bias
+	 * towards upclocking to deliver the frame on time. By only changing
+	 * the RPS thresholds to sample more regularly and aim for higher
+	 * clocks we can hopefully deliver low power workloads (like kodi)
+	 * that are not quite steady state without resorting to forcing
+	 * maximum clocks following a vblank miss (see do_rps_boost()).
+	 */
+	if (!state->rps_interactive) {
+		intel_rps_mark_interactive(&to_gt(dev_priv)->rps, true);
+		state->rps_interactive = true;
+	}
+
+	return 0;
+
+unpin_fb:
+	intel_plane_unpin_fb(new_plane_state);
+
+	return ret;
+}
+
+/**
+ * intel_cleanup_plane_fb - Cleans up an fb after plane use
+ * @plane: drm plane to clean up for
+ * @_old_plane_state: the state from the previous modeset
+ *
+ * Cleans up a framebuffer that has just been removed from a plane.
+ */
+static void
+intel_cleanup_plane_fb(struct drm_plane *plane,
+		       struct drm_plane_state *_old_plane_state)
+{
+	struct intel_plane_state *old_plane_state =
+		to_intel_plane_state(_old_plane_state);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(old_plane_state->uapi.state);
+	struct drm_i915_private *dev_priv = to_i915(plane->dev);
+	struct drm_i915_gem_object *obj = intel_fb_obj(old_plane_state->hw.fb);
+
+	if (!obj)
+		return;
+
+	if (state->rps_interactive) {
+		intel_rps_mark_interactive(&to_gt(dev_priv)->rps, false);
+		state->rps_interactive = false;
+	}
+
+	/* Should only be called after a successful intel_prepare_plane_fb()! */
+	intel_plane_unpin_fb(old_plane_state);
+}
+
+static const struct drm_plane_helper_funcs intel_plane_helper_funcs = {
+	.prepare_fb = intel_prepare_plane_fb,
+	.cleanup_fb = intel_cleanup_plane_fb,
+};
+
+void intel_plane_helper_add(struct intel_plane *plane)
+{
+	drm_plane_helper_add(&plane->base, &intel_plane_helper_funcs);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_atomic_plane.h b/drivers/gpu/drm/i915/display/intel_atomic_plane.h
new file mode 100644
index 000000000..74b6d3b16
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_atomic_plane.h
@@ -0,0 +1,69 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_ATOMIC_PLANE_H__
+#define __INTEL_ATOMIC_PLANE_H__
+
+#include <linux/types.h>
+
+struct drm_plane;
+struct drm_property;
+struct drm_rect;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_plane;
+struct intel_plane_state;
+enum plane_id;
+
+unsigned int intel_adjusted_rate(const struct drm_rect *src,
+				 const struct drm_rect *dst,
+				 unsigned int rate);
+unsigned int intel_plane_pixel_rate(const struct intel_crtc_state *crtc_state,
+				    const struct intel_plane_state *plane_state);
+
+unsigned int intel_plane_data_rate(const struct intel_crtc_state *crtc_state,
+				   const struct intel_plane_state *plane_state,
+				   int color_plane);
+void intel_plane_copy_uapi_to_hw_state(struct intel_plane_state *plane_state,
+				       const struct intel_plane_state *from_plane_state,
+				       struct intel_crtc *crtc);
+void intel_plane_copy_hw_state(struct intel_plane_state *plane_state,
+			       const struct intel_plane_state *from_plane_state);
+void intel_plane_update_noarm(struct intel_plane *plane,
+			      const struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state);
+void intel_plane_update_arm(struct intel_plane *plane,
+			    const struct intel_crtc_state *crtc_state,
+			    const struct intel_plane_state *plane_state);
+void intel_plane_disable_arm(struct intel_plane *plane,
+			     const struct intel_crtc_state *crtc_state);
+struct intel_plane *intel_plane_alloc(void);
+void intel_plane_free(struct intel_plane *plane);
+struct drm_plane_state *intel_plane_duplicate_state(struct drm_plane *plane);
+void intel_plane_destroy_state(struct drm_plane *plane,
+			       struct drm_plane_state *state);
+void intel_crtc_planes_update_noarm(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc);
+void intel_crtc_planes_update_arm(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc);
+int intel_plane_atomic_check_with_state(const struct intel_crtc_state *old_crtc_state,
+					struct intel_crtc_state *crtc_state,
+					const struct intel_plane_state *old_plane_state,
+					struct intel_plane_state *intel_state);
+int intel_plane_atomic_check(struct intel_atomic_state *state,
+			     struct intel_plane *plane);
+int intel_plane_calc_min_cdclk(struct intel_atomic_state *state,
+			       struct intel_plane *plane,
+			       bool *need_cdclk_calc);
+int intel_atomic_plane_check_clipping(struct intel_plane_state *plane_state,
+				      struct intel_crtc_state *crtc_state,
+				      int min_scale, int max_scale,
+				      bool can_position);
+void intel_plane_set_invisible(struct intel_crtc_state *crtc_state,
+			       struct intel_plane_state *plane_state);
+void intel_plane_helper_add(struct intel_plane *plane);
+
+#endif /* __INTEL_ATOMIC_PLANE_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_audio.c b/drivers/gpu/drm/i915/display/intel_audio.c
new file mode 100644
index 000000000..aacbc6da8
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_audio.c
@@ -0,0 +1,1410 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/component.h>
+#include <linux/kernel.h>
+
+#include <drm/drm_edid.h>
+#include <drm/i915_component.h>
+
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_audio.h"
+#include "intel_audio_regs.h"
+#include "intel_cdclk.h"
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_lpe_audio.h"
+
+/**
+ * DOC: High Definition Audio over HDMI and Display Port
+ *
+ * The graphics and audio drivers together support High Definition Audio over
+ * HDMI and Display Port. The audio programming sequences are divided into audio
+ * codec and controller enable and disable sequences. The graphics driver
+ * handles the audio codec sequences, while the audio driver handles the audio
+ * controller sequences.
+ *
+ * The disable sequences must be performed before disabling the transcoder or
+ * port. The enable sequences may only be performed after enabling the
+ * transcoder and port, and after completed link training. Therefore the audio
+ * enable/disable sequences are part of the modeset sequence.
+ *
+ * The codec and controller sequences could be done either parallel or serial,
+ * but generally the ELDV/PD change in the codec sequence indicates to the audio
+ * driver that the controller sequence should start. Indeed, most of the
+ * co-operation between the graphics and audio drivers is handled via audio
+ * related registers. (The notable exception is the power management, not
+ * covered here.)
+ *
+ * The struct &i915_audio_component is used to interact between the graphics
+ * and audio drivers. The struct &i915_audio_component_ops @ops in it is
+ * defined in graphics driver and called in audio driver. The
+ * struct &i915_audio_component_audio_ops @audio_ops is called from i915 driver.
+ */
+
+struct intel_audio_funcs {
+	void (*audio_codec_enable)(struct intel_encoder *encoder,
+				   const struct intel_crtc_state *crtc_state,
+				   const struct drm_connector_state *conn_state);
+	void (*audio_codec_disable)(struct intel_encoder *encoder,
+				    const struct intel_crtc_state *old_crtc_state,
+				    const struct drm_connector_state *old_conn_state);
+};
+
+/* DP N/M table */
+#define LC_810M	810000
+#define LC_540M	540000
+#define LC_270M	270000
+#define LC_162M	162000
+
+struct dp_aud_n_m {
+	int sample_rate;
+	int clock;
+	u16 m;
+	u16 n;
+};
+
+struct hdmi_aud_ncts {
+	int sample_rate;
+	int clock;
+	int n;
+	int cts;
+};
+
+/* Values according to DP 1.4 Table 2-104 */
+static const struct dp_aud_n_m dp_aud_n_m[] = {
+	{ 32000, LC_162M, 1024, 10125 },
+	{ 44100, LC_162M, 784, 5625 },
+	{ 48000, LC_162M, 512, 3375 },
+	{ 64000, LC_162M, 2048, 10125 },
+	{ 88200, LC_162M, 1568, 5625 },
+	{ 96000, LC_162M, 1024, 3375 },
+	{ 128000, LC_162M, 4096, 10125 },
+	{ 176400, LC_162M, 3136, 5625 },
+	{ 192000, LC_162M, 2048, 3375 },
+	{ 32000, LC_270M, 1024, 16875 },
+	{ 44100, LC_270M, 784, 9375 },
+	{ 48000, LC_270M, 512, 5625 },
+	{ 64000, LC_270M, 2048, 16875 },
+	{ 88200, LC_270M, 1568, 9375 },
+	{ 96000, LC_270M, 1024, 5625 },
+	{ 128000, LC_270M, 4096, 16875 },
+	{ 176400, LC_270M, 3136, 9375 },
+	{ 192000, LC_270M, 2048, 5625 },
+	{ 32000, LC_540M, 1024, 33750 },
+	{ 44100, LC_540M, 784, 18750 },
+	{ 48000, LC_540M, 512, 11250 },
+	{ 64000, LC_540M, 2048, 33750 },
+	{ 88200, LC_540M, 1568, 18750 },
+	{ 96000, LC_540M, 1024, 11250 },
+	{ 128000, LC_540M, 4096, 33750 },
+	{ 176400, LC_540M, 3136, 18750 },
+	{ 192000, LC_540M, 2048, 11250 },
+	{ 32000, LC_810M, 1024, 50625 },
+	{ 44100, LC_810M, 784, 28125 },
+	{ 48000, LC_810M, 512, 16875 },
+	{ 64000, LC_810M, 2048, 50625 },
+	{ 88200, LC_810M, 1568, 28125 },
+	{ 96000, LC_810M, 1024, 16875 },
+	{ 128000, LC_810M, 4096, 50625 },
+	{ 176400, LC_810M, 3136, 28125 },
+	{ 192000, LC_810M, 2048, 16875 },
+};
+
+static const struct dp_aud_n_m *
+audio_config_dp_get_n_m(const struct intel_crtc_state *crtc_state, int rate)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(dp_aud_n_m); i++) {
+		if (rate == dp_aud_n_m[i].sample_rate &&
+		    crtc_state->port_clock == dp_aud_n_m[i].clock)
+			return &dp_aud_n_m[i];
+	}
+
+	return NULL;
+}
+
+static const struct {
+	int clock;
+	u32 config;
+} hdmi_audio_clock[] = {
+	{ 25175, AUD_CONFIG_PIXEL_CLOCK_HDMI_25175 },
+	{ 25200, AUD_CONFIG_PIXEL_CLOCK_HDMI_25200 }, /* default per bspec */
+	{ 27000, AUD_CONFIG_PIXEL_CLOCK_HDMI_27000 },
+	{ 27027, AUD_CONFIG_PIXEL_CLOCK_HDMI_27027 },
+	{ 54000, AUD_CONFIG_PIXEL_CLOCK_HDMI_54000 },
+	{ 54054, AUD_CONFIG_PIXEL_CLOCK_HDMI_54054 },
+	{ 74176, AUD_CONFIG_PIXEL_CLOCK_HDMI_74176 },
+	{ 74250, AUD_CONFIG_PIXEL_CLOCK_HDMI_74250 },
+	{ 148352, AUD_CONFIG_PIXEL_CLOCK_HDMI_148352 },
+	{ 148500, AUD_CONFIG_PIXEL_CLOCK_HDMI_148500 },
+	{ 296703, AUD_CONFIG_PIXEL_CLOCK_HDMI_296703 },
+	{ 297000, AUD_CONFIG_PIXEL_CLOCK_HDMI_297000 },
+	{ 593407, AUD_CONFIG_PIXEL_CLOCK_HDMI_593407 },
+	{ 594000, AUD_CONFIG_PIXEL_CLOCK_HDMI_594000 },
+};
+
+/* HDMI N/CTS table */
+#define TMDS_297M 297000
+#define TMDS_296M 296703
+#define TMDS_594M 594000
+#define TMDS_593M 593407
+
+static const struct hdmi_aud_ncts hdmi_aud_ncts_24bpp[] = {
+	{ 32000, TMDS_296M, 5824, 421875 },
+	{ 32000, TMDS_297M, 3072, 222750 },
+	{ 32000, TMDS_593M, 5824, 843750 },
+	{ 32000, TMDS_594M, 3072, 445500 },
+	{ 44100, TMDS_296M, 4459, 234375 },
+	{ 44100, TMDS_297M, 4704, 247500 },
+	{ 44100, TMDS_593M, 8918, 937500 },
+	{ 44100, TMDS_594M, 9408, 990000 },
+	{ 88200, TMDS_296M, 8918, 234375 },
+	{ 88200, TMDS_297M, 9408, 247500 },
+	{ 88200, TMDS_593M, 17836, 937500 },
+	{ 88200, TMDS_594M, 18816, 990000 },
+	{ 176400, TMDS_296M, 17836, 234375 },
+	{ 176400, TMDS_297M, 18816, 247500 },
+	{ 176400, TMDS_593M, 35672, 937500 },
+	{ 176400, TMDS_594M, 37632, 990000 },
+	{ 48000, TMDS_296M, 5824, 281250 },
+	{ 48000, TMDS_297M, 5120, 247500 },
+	{ 48000, TMDS_593M, 5824, 562500 },
+	{ 48000, TMDS_594M, 6144, 594000 },
+	{ 96000, TMDS_296M, 11648, 281250 },
+	{ 96000, TMDS_297M, 10240, 247500 },
+	{ 96000, TMDS_593M, 11648, 562500 },
+	{ 96000, TMDS_594M, 12288, 594000 },
+	{ 192000, TMDS_296M, 23296, 281250 },
+	{ 192000, TMDS_297M, 20480, 247500 },
+	{ 192000, TMDS_593M, 23296, 562500 },
+	{ 192000, TMDS_594M, 24576, 594000 },
+};
+
+/* Appendix C - N & CTS values for deep color from HDMI 2.0 spec*/
+/* HDMI N/CTS table for 10 bit deep color(30 bpp)*/
+#define TMDS_371M 371250
+#define TMDS_370M 370878
+
+static const struct hdmi_aud_ncts hdmi_aud_ncts_30bpp[] = {
+	{ 32000, TMDS_370M, 5824, 527344 },
+	{ 32000, TMDS_371M, 6144, 556875 },
+	{ 44100, TMDS_370M, 8918, 585938 },
+	{ 44100, TMDS_371M, 4704, 309375 },
+	{ 88200, TMDS_370M, 17836, 585938 },
+	{ 88200, TMDS_371M, 9408, 309375 },
+	{ 176400, TMDS_370M, 35672, 585938 },
+	{ 176400, TMDS_371M, 18816, 309375 },
+	{ 48000, TMDS_370M, 11648, 703125 },
+	{ 48000, TMDS_371M, 5120, 309375 },
+	{ 96000, TMDS_370M, 23296, 703125 },
+	{ 96000, TMDS_371M, 10240, 309375 },
+	{ 192000, TMDS_370M, 46592, 703125 },
+	{ 192000, TMDS_371M, 20480, 309375 },
+};
+
+/* HDMI N/CTS table for 12 bit deep color(36 bpp)*/
+#define TMDS_445_5M 445500
+#define TMDS_445M 445054
+
+static const struct hdmi_aud_ncts hdmi_aud_ncts_36bpp[] = {
+	{ 32000, TMDS_445M, 5824, 632813 },
+	{ 32000, TMDS_445_5M, 4096, 445500 },
+	{ 44100, TMDS_445M, 8918, 703125 },
+	{ 44100, TMDS_445_5M, 4704, 371250 },
+	{ 88200, TMDS_445M, 17836, 703125 },
+	{ 88200, TMDS_445_5M, 9408, 371250 },
+	{ 176400, TMDS_445M, 35672, 703125 },
+	{ 176400, TMDS_445_5M, 18816, 371250 },
+	{ 48000, TMDS_445M, 5824, 421875 },
+	{ 48000, TMDS_445_5M, 5120, 371250 },
+	{ 96000, TMDS_445M, 11648, 421875 },
+	{ 96000, TMDS_445_5M, 10240, 371250 },
+	{ 192000, TMDS_445M, 23296, 421875 },
+	{ 192000, TMDS_445_5M, 20480, 371250 },
+};
+
+/* get AUD_CONFIG_PIXEL_CLOCK_HDMI_* value for mode */
+static u32 audio_config_hdmi_pixel_clock(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(hdmi_audio_clock); i++) {
+		if (adjusted_mode->crtc_clock == hdmi_audio_clock[i].clock)
+			break;
+	}
+
+	if (DISPLAY_VER(dev_priv) < 12 && adjusted_mode->crtc_clock > 148500)
+		i = ARRAY_SIZE(hdmi_audio_clock);
+
+	if (i == ARRAY_SIZE(hdmi_audio_clock)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "HDMI audio pixel clock setting for %d not found, falling back to defaults\n",
+			    adjusted_mode->crtc_clock);
+		i = 1;
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Configuring HDMI audio for pixel clock %d (0x%08x)\n",
+		    hdmi_audio_clock[i].clock,
+		    hdmi_audio_clock[i].config);
+
+	return hdmi_audio_clock[i].config;
+}
+
+static int audio_config_hdmi_get_n(const struct intel_crtc_state *crtc_state,
+				   int rate)
+{
+	const struct hdmi_aud_ncts *hdmi_ncts_table;
+	int i, size;
+
+	if (crtc_state->pipe_bpp == 36) {
+		hdmi_ncts_table = hdmi_aud_ncts_36bpp;
+		size = ARRAY_SIZE(hdmi_aud_ncts_36bpp);
+	} else if (crtc_state->pipe_bpp == 30) {
+		hdmi_ncts_table = hdmi_aud_ncts_30bpp;
+		size = ARRAY_SIZE(hdmi_aud_ncts_30bpp);
+	} else {
+		hdmi_ncts_table = hdmi_aud_ncts_24bpp;
+		size = ARRAY_SIZE(hdmi_aud_ncts_24bpp);
+	}
+
+	for (i = 0; i < size; i++) {
+		if (rate == hdmi_ncts_table[i].sample_rate &&
+		    crtc_state->port_clock == hdmi_ncts_table[i].clock) {
+			return hdmi_ncts_table[i].n;
+		}
+	}
+	return 0;
+}
+
+static bool intel_eld_uptodate(struct drm_connector *connector,
+			       i915_reg_t reg_eldv, u32 bits_eldv,
+			       i915_reg_t reg_elda, u32 bits_elda,
+			       i915_reg_t reg_edid)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	const u8 *eld = connector->eld;
+	u32 tmp;
+	int i;
+
+	tmp = intel_de_read(dev_priv, reg_eldv);
+	tmp &= bits_eldv;
+
+	if (!tmp)
+		return false;
+
+	tmp = intel_de_read(dev_priv, reg_elda);
+	tmp &= ~bits_elda;
+	intel_de_write(dev_priv, reg_elda, tmp);
+
+	for (i = 0; i < drm_eld_size(eld) / 4; i++)
+		if (intel_de_read(dev_priv, reg_edid) != *((const u32 *)eld + i))
+			return false;
+
+	return true;
+}
+
+static void g4x_audio_codec_disable(struct intel_encoder *encoder,
+				    const struct intel_crtc_state *old_crtc_state,
+				    const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 eldv, tmp;
+
+	tmp = intel_de_read(dev_priv, G4X_AUD_VID_DID);
+	if (tmp == INTEL_AUDIO_DEVBLC || tmp == INTEL_AUDIO_DEVCL)
+		eldv = G4X_ELDV_DEVCL_DEVBLC;
+	else
+		eldv = G4X_ELDV_DEVCTG;
+
+	/* Invalidate ELD */
+	tmp = intel_de_read(dev_priv, G4X_AUD_CNTL_ST);
+	tmp &= ~eldv;
+	intel_de_write(dev_priv, G4X_AUD_CNTL_ST, tmp);
+}
+
+static void g4x_audio_codec_enable(struct intel_encoder *encoder,
+				   const struct intel_crtc_state *crtc_state,
+				   const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct drm_connector *connector = conn_state->connector;
+	const u8 *eld = connector->eld;
+	u32 eldv;
+	u32 tmp;
+	int len, i;
+
+	tmp = intel_de_read(dev_priv, G4X_AUD_VID_DID);
+	if (tmp == INTEL_AUDIO_DEVBLC || tmp == INTEL_AUDIO_DEVCL)
+		eldv = G4X_ELDV_DEVCL_DEVBLC;
+	else
+		eldv = G4X_ELDV_DEVCTG;
+
+	if (intel_eld_uptodate(connector,
+			       G4X_AUD_CNTL_ST, eldv,
+			       G4X_AUD_CNTL_ST, G4X_ELD_ADDR_MASK,
+			       G4X_HDMIW_HDMIEDID))
+		return;
+
+	tmp = intel_de_read(dev_priv, G4X_AUD_CNTL_ST);
+	tmp &= ~(eldv | G4X_ELD_ADDR_MASK);
+	len = (tmp >> 9) & 0x1f;		/* ELD buffer size */
+	intel_de_write(dev_priv, G4X_AUD_CNTL_ST, tmp);
+
+	len = min(drm_eld_size(eld) / 4, len);
+	for (i = 0; i < len; i++)
+		intel_de_write(dev_priv, G4X_HDMIW_HDMIEDID,
+			       *((const u32 *)eld + i));
+
+	tmp = intel_de_read(dev_priv, G4X_AUD_CNTL_ST);
+	tmp |= eldv;
+	intel_de_write(dev_priv, G4X_AUD_CNTL_ST, tmp);
+}
+
+static void
+hsw_dp_audio_config_update(struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct i915_audio_component *acomp = dev_priv->display.audio.component;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	enum port port = encoder->port;
+	const struct dp_aud_n_m *nm;
+	int rate;
+	u32 tmp;
+
+	rate = acomp ? acomp->aud_sample_rate[port] : 0;
+	nm = audio_config_dp_get_n_m(crtc_state, rate);
+	if (nm)
+		drm_dbg_kms(&dev_priv->drm, "using Maud %u, Naud %u\n", nm->m,
+			    nm->n);
+	else
+		drm_dbg_kms(&dev_priv->drm, "using automatic Maud, Naud\n");
+
+	tmp = intel_de_read(dev_priv, HSW_AUD_CFG(cpu_transcoder));
+	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
+	tmp &= ~AUD_CONFIG_PIXEL_CLOCK_HDMI_MASK;
+	tmp &= ~AUD_CONFIG_N_PROG_ENABLE;
+	tmp |= AUD_CONFIG_N_VALUE_INDEX;
+
+	if (nm) {
+		tmp &= ~AUD_CONFIG_N_MASK;
+		tmp |= AUD_CONFIG_N(nm->n);
+		tmp |= AUD_CONFIG_N_PROG_ENABLE;
+	}
+
+	intel_de_write(dev_priv, HSW_AUD_CFG(cpu_transcoder), tmp);
+
+	tmp = intel_de_read(dev_priv, HSW_AUD_M_CTS_ENABLE(cpu_transcoder));
+	tmp &= ~AUD_CONFIG_M_MASK;
+	tmp &= ~AUD_M_CTS_M_VALUE_INDEX;
+	tmp &= ~AUD_M_CTS_M_PROG_ENABLE;
+
+	if (nm) {
+		tmp |= nm->m;
+		tmp |= AUD_M_CTS_M_VALUE_INDEX;
+		tmp |= AUD_M_CTS_M_PROG_ENABLE;
+	}
+
+	intel_de_write(dev_priv, HSW_AUD_M_CTS_ENABLE(cpu_transcoder), tmp);
+}
+
+static void
+hsw_hdmi_audio_config_update(struct intel_encoder *encoder,
+			     const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct i915_audio_component *acomp = dev_priv->display.audio.component;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	enum port port = encoder->port;
+	int n, rate;
+	u32 tmp;
+
+	rate = acomp ? acomp->aud_sample_rate[port] : 0;
+
+	tmp = intel_de_read(dev_priv, HSW_AUD_CFG(cpu_transcoder));
+	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
+	tmp &= ~AUD_CONFIG_PIXEL_CLOCK_HDMI_MASK;
+	tmp &= ~AUD_CONFIG_N_PROG_ENABLE;
+	tmp |= audio_config_hdmi_pixel_clock(crtc_state);
+
+	n = audio_config_hdmi_get_n(crtc_state, rate);
+	if (n != 0) {
+		drm_dbg_kms(&dev_priv->drm, "using N %d\n", n);
+
+		tmp &= ~AUD_CONFIG_N_MASK;
+		tmp |= AUD_CONFIG_N(n);
+		tmp |= AUD_CONFIG_N_PROG_ENABLE;
+	} else {
+		drm_dbg_kms(&dev_priv->drm, "using automatic N\n");
+	}
+
+	intel_de_write(dev_priv, HSW_AUD_CFG(cpu_transcoder), tmp);
+
+	/*
+	 * Let's disable "Enable CTS or M Prog bit"
+	 * and let HW calculate the value
+	 */
+	tmp = intel_de_read(dev_priv, HSW_AUD_M_CTS_ENABLE(cpu_transcoder));
+	tmp &= ~AUD_M_CTS_M_PROG_ENABLE;
+	tmp &= ~AUD_M_CTS_M_VALUE_INDEX;
+	intel_de_write(dev_priv, HSW_AUD_M_CTS_ENABLE(cpu_transcoder), tmp);
+}
+
+static void
+hsw_audio_config_update(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state)
+{
+	if (intel_crtc_has_dp_encoder(crtc_state))
+		hsw_dp_audio_config_update(encoder, crtc_state);
+	else
+		hsw_hdmi_audio_config_update(encoder, crtc_state);
+}
+
+static void hsw_audio_codec_disable(struct intel_encoder *encoder,
+				    const struct intel_crtc_state *old_crtc_state,
+				    const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum transcoder cpu_transcoder = old_crtc_state->cpu_transcoder;
+	u32 tmp;
+
+	mutex_lock(&dev_priv->display.audio.mutex);
+
+	/* Disable timestamps */
+	tmp = intel_de_read(dev_priv, HSW_AUD_CFG(cpu_transcoder));
+	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
+	tmp |= AUD_CONFIG_N_PROG_ENABLE;
+	tmp &= ~AUD_CONFIG_UPPER_N_MASK;
+	tmp &= ~AUD_CONFIG_LOWER_N_MASK;
+	if (intel_crtc_has_dp_encoder(old_crtc_state))
+		tmp |= AUD_CONFIG_N_VALUE_INDEX;
+	intel_de_write(dev_priv, HSW_AUD_CFG(cpu_transcoder), tmp);
+
+	/* Invalidate ELD */
+	tmp = intel_de_read(dev_priv, HSW_AUD_PIN_ELD_CP_VLD);
+	tmp &= ~AUDIO_ELD_VALID(cpu_transcoder);
+	tmp &= ~AUDIO_OUTPUT_ENABLE(cpu_transcoder);
+	intel_de_write(dev_priv, HSW_AUD_PIN_ELD_CP_VLD, tmp);
+
+	mutex_unlock(&dev_priv->display.audio.mutex);
+}
+
+static unsigned int calc_hblank_early_prog(struct intel_encoder *encoder,
+					   const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	unsigned int link_clks_available, link_clks_required;
+	unsigned int tu_data, tu_line, link_clks_active;
+	unsigned int h_active, h_total, hblank_delta, pixel_clk;
+	unsigned int fec_coeff, cdclk, vdsc_bpp;
+	unsigned int link_clk, lanes;
+	unsigned int hblank_rise;
+
+	h_active = crtc_state->hw.adjusted_mode.crtc_hdisplay;
+	h_total = crtc_state->hw.adjusted_mode.crtc_htotal;
+	pixel_clk = crtc_state->hw.adjusted_mode.crtc_clock;
+	vdsc_bpp = crtc_state->dsc.compressed_bpp;
+	cdclk = i915->display.cdclk.hw.cdclk;
+	/* fec= 0.972261, using rounding multiplier of 1000000 */
+	fec_coeff = 972261;
+	link_clk = crtc_state->port_clock;
+	lanes = crtc_state->lane_count;
+
+	drm_dbg_kms(&i915->drm, "h_active = %u link_clk = %u :"
+		    "lanes = %u vdsc_bpp = %u cdclk = %u\n",
+		    h_active, link_clk, lanes, vdsc_bpp, cdclk);
+
+	if (WARN_ON(!link_clk || !pixel_clk || !lanes || !vdsc_bpp || !cdclk))
+		return 0;
+
+	link_clks_available = (h_total - h_active) * link_clk / pixel_clk - 28;
+	link_clks_required = DIV_ROUND_UP(192000 * h_total, 1000 * pixel_clk) * (48 / lanes + 2);
+
+	if (link_clks_available > link_clks_required)
+		hblank_delta = 32;
+	else
+		hblank_delta = DIV64_U64_ROUND_UP(mul_u32_u32(5 * (link_clk + cdclk), pixel_clk),
+						  mul_u32_u32(link_clk, cdclk));
+
+	tu_data = div64_u64(mul_u32_u32(pixel_clk * vdsc_bpp * 8, 1000000),
+			    mul_u32_u32(link_clk * lanes, fec_coeff));
+	tu_line = div64_u64(h_active * mul_u32_u32(link_clk, fec_coeff),
+			    mul_u32_u32(64 * pixel_clk, 1000000));
+	link_clks_active  = (tu_line - 1) * 64 + tu_data;
+
+	hblank_rise = (link_clks_active + 6 * DIV_ROUND_UP(link_clks_active, 250) + 4) * pixel_clk / link_clk;
+
+	return h_active - hblank_rise + hblank_delta;
+}
+
+static unsigned int calc_samples_room(const struct intel_crtc_state *crtc_state)
+{
+	unsigned int h_active, h_total, pixel_clk;
+	unsigned int link_clk, lanes;
+
+	h_active = crtc_state->hw.adjusted_mode.hdisplay;
+	h_total = crtc_state->hw.adjusted_mode.htotal;
+	pixel_clk = crtc_state->hw.adjusted_mode.clock;
+	link_clk = crtc_state->port_clock;
+	lanes = crtc_state->lane_count;
+
+	return ((h_total - h_active) * link_clk - 12 * pixel_clk) /
+		(pixel_clk * (48 / lanes + 2));
+}
+
+static void enable_audio_dsc_wa(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum pipe pipe = crtc->pipe;
+	unsigned int hblank_early_prog, samples_room;
+	unsigned int val;
+
+	if (DISPLAY_VER(i915) < 11)
+		return;
+
+	val = intel_de_read(i915, AUD_CONFIG_BE);
+
+	if (DISPLAY_VER(i915) == 11)
+		val |= HBLANK_EARLY_ENABLE_ICL(pipe);
+	else if (DISPLAY_VER(i915) >= 12)
+		val |= HBLANK_EARLY_ENABLE_TGL(pipe);
+
+	if (crtc_state->dsc.compression_enable &&
+	    crtc_state->hw.adjusted_mode.hdisplay >= 3840 &&
+	    crtc_state->hw.adjusted_mode.vdisplay >= 2160) {
+		/* Get hblank early enable value required */
+		val &= ~HBLANK_START_COUNT_MASK(pipe);
+		hblank_early_prog = calc_hblank_early_prog(encoder, crtc_state);
+		if (hblank_early_prog < 32)
+			val |= HBLANK_START_COUNT(pipe, HBLANK_START_COUNT_32);
+		else if (hblank_early_prog < 64)
+			val |= HBLANK_START_COUNT(pipe, HBLANK_START_COUNT_64);
+		else if (hblank_early_prog < 96)
+			val |= HBLANK_START_COUNT(pipe, HBLANK_START_COUNT_96);
+		else
+			val |= HBLANK_START_COUNT(pipe, HBLANK_START_COUNT_128);
+
+		/* Get samples room value required */
+		val &= ~NUMBER_SAMPLES_PER_LINE_MASK(pipe);
+		samples_room = calc_samples_room(crtc_state);
+		if (samples_room < 3)
+			val |= NUMBER_SAMPLES_PER_LINE(pipe, samples_room);
+		else /* Program 0 i.e "All Samples available in buffer" */
+			val |= NUMBER_SAMPLES_PER_LINE(pipe, 0x0);
+	}
+
+	intel_de_write(i915, AUD_CONFIG_BE, val);
+}
+
+#undef ROUNDING_FACTOR
+
+static void hsw_audio_codec_enable(struct intel_encoder *encoder,
+				   const struct intel_crtc_state *crtc_state,
+				   const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct drm_connector *connector = conn_state->connector;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	const u8 *eld = connector->eld;
+	u32 tmp;
+	int len, i;
+
+	mutex_lock(&dev_priv->display.audio.mutex);
+
+	/* Enable Audio WA for 4k DSC usecases */
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP))
+		enable_audio_dsc_wa(encoder, crtc_state);
+
+	/* Enable audio presence detect, invalidate ELD */
+	tmp = intel_de_read(dev_priv, HSW_AUD_PIN_ELD_CP_VLD);
+	tmp |= AUDIO_OUTPUT_ENABLE(cpu_transcoder);
+	tmp &= ~AUDIO_ELD_VALID(cpu_transcoder);
+	intel_de_write(dev_priv, HSW_AUD_PIN_ELD_CP_VLD, tmp);
+
+	/*
+	 * FIXME: We're supposed to wait for vblank here, but we have vblanks
+	 * disabled during the mode set. The proper fix would be to push the
+	 * rest of the setup into a vblank work item, queued here, but the
+	 * infrastructure is not there yet.
+	 */
+
+	/* Reset ELD write address */
+	tmp = intel_de_read(dev_priv, HSW_AUD_DIP_ELD_CTRL(cpu_transcoder));
+	tmp &= ~IBX_ELD_ADDRESS_MASK;
+	intel_de_write(dev_priv, HSW_AUD_DIP_ELD_CTRL(cpu_transcoder), tmp);
+
+	/* Up to 84 bytes of hw ELD buffer */
+	len = min(drm_eld_size(eld), 84);
+	for (i = 0; i < len / 4; i++)
+		intel_de_write(dev_priv, HSW_AUD_EDID_DATA(cpu_transcoder),
+			       *((const u32 *)eld + i));
+
+	/* ELD valid */
+	tmp = intel_de_read(dev_priv, HSW_AUD_PIN_ELD_CP_VLD);
+	tmp |= AUDIO_ELD_VALID(cpu_transcoder);
+	intel_de_write(dev_priv, HSW_AUD_PIN_ELD_CP_VLD, tmp);
+
+	/* Enable timestamps */
+	hsw_audio_config_update(encoder, crtc_state);
+
+	mutex_unlock(&dev_priv->display.audio.mutex);
+}
+
+static void ilk_audio_codec_disable(struct intel_encoder *encoder,
+				    const struct intel_crtc_state *old_crtc_state,
+				    const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	enum pipe pipe = crtc->pipe;
+	enum port port = encoder->port;
+	u32 tmp, eldv;
+	i915_reg_t aud_config, aud_cntrl_st2;
+
+	if (drm_WARN_ON(&dev_priv->drm, port == PORT_A))
+		return;
+
+	if (HAS_PCH_IBX(dev_priv)) {
+		aud_config = IBX_AUD_CFG(pipe);
+		aud_cntrl_st2 = IBX_AUD_CNTL_ST2;
+	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		aud_config = VLV_AUD_CFG(pipe);
+		aud_cntrl_st2 = VLV_AUD_CNTL_ST2;
+	} else {
+		aud_config = CPT_AUD_CFG(pipe);
+		aud_cntrl_st2 = CPT_AUD_CNTRL_ST2;
+	}
+
+	/* Disable timestamps */
+	tmp = intel_de_read(dev_priv, aud_config);
+	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
+	tmp |= AUD_CONFIG_N_PROG_ENABLE;
+	tmp &= ~AUD_CONFIG_UPPER_N_MASK;
+	tmp &= ~AUD_CONFIG_LOWER_N_MASK;
+	if (intel_crtc_has_dp_encoder(old_crtc_state))
+		tmp |= AUD_CONFIG_N_VALUE_INDEX;
+	intel_de_write(dev_priv, aud_config, tmp);
+
+	eldv = IBX_ELD_VALID(port);
+
+	/* Invalidate ELD */
+	tmp = intel_de_read(dev_priv, aud_cntrl_st2);
+	tmp &= ~eldv;
+	intel_de_write(dev_priv, aud_cntrl_st2, tmp);
+}
+
+static void ilk_audio_codec_enable(struct intel_encoder *encoder,
+				   const struct intel_crtc_state *crtc_state,
+				   const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_connector *connector = conn_state->connector;
+	enum pipe pipe = crtc->pipe;
+	enum port port = encoder->port;
+	const u8 *eld = connector->eld;
+	u32 tmp, eldv;
+	int len, i;
+	i915_reg_t hdmiw_hdmiedid, aud_config, aud_cntl_st, aud_cntrl_st2;
+
+	if (drm_WARN_ON(&dev_priv->drm, port == PORT_A))
+		return;
+
+	/*
+	 * FIXME: We're supposed to wait for vblank here, but we have vblanks
+	 * disabled during the mode set. The proper fix would be to push the
+	 * rest of the setup into a vblank work item, queued here, but the
+	 * infrastructure is not there yet.
+	 */
+
+	if (HAS_PCH_IBX(dev_priv)) {
+		hdmiw_hdmiedid = IBX_HDMIW_HDMIEDID(pipe);
+		aud_config = IBX_AUD_CFG(pipe);
+		aud_cntl_st = IBX_AUD_CNTL_ST(pipe);
+		aud_cntrl_st2 = IBX_AUD_CNTL_ST2;
+	} else if (IS_VALLEYVIEW(dev_priv) ||
+		   IS_CHERRYVIEW(dev_priv)) {
+		hdmiw_hdmiedid = VLV_HDMIW_HDMIEDID(pipe);
+		aud_config = VLV_AUD_CFG(pipe);
+		aud_cntl_st = VLV_AUD_CNTL_ST(pipe);
+		aud_cntrl_st2 = VLV_AUD_CNTL_ST2;
+	} else {
+		hdmiw_hdmiedid = CPT_HDMIW_HDMIEDID(pipe);
+		aud_config = CPT_AUD_CFG(pipe);
+		aud_cntl_st = CPT_AUD_CNTL_ST(pipe);
+		aud_cntrl_st2 = CPT_AUD_CNTRL_ST2;
+	}
+
+	eldv = IBX_ELD_VALID(port);
+
+	/* Invalidate ELD */
+	tmp = intel_de_read(dev_priv, aud_cntrl_st2);
+	tmp &= ~eldv;
+	intel_de_write(dev_priv, aud_cntrl_st2, tmp);
+
+	/* Reset ELD write address */
+	tmp = intel_de_read(dev_priv, aud_cntl_st);
+	tmp &= ~IBX_ELD_ADDRESS_MASK;
+	intel_de_write(dev_priv, aud_cntl_st, tmp);
+
+	/* Up to 84 bytes of hw ELD buffer */
+	len = min(drm_eld_size(eld), 84);
+	for (i = 0; i < len / 4; i++)
+		intel_de_write(dev_priv, hdmiw_hdmiedid,
+			       *((const u32 *)eld + i));
+
+	/* ELD valid */
+	tmp = intel_de_read(dev_priv, aud_cntrl_st2);
+	tmp |= eldv;
+	intel_de_write(dev_priv, aud_cntrl_st2, tmp);
+
+	/* Enable timestamps */
+	tmp = intel_de_read(dev_priv, aud_config);
+	tmp &= ~AUD_CONFIG_N_VALUE_INDEX;
+	tmp &= ~AUD_CONFIG_N_PROG_ENABLE;
+	tmp &= ~AUD_CONFIG_PIXEL_CLOCK_HDMI_MASK;
+	if (intel_crtc_has_dp_encoder(crtc_state))
+		tmp |= AUD_CONFIG_N_VALUE_INDEX;
+	else
+		tmp |= audio_config_hdmi_pixel_clock(crtc_state);
+	intel_de_write(dev_priv, aud_config, tmp);
+}
+
+/**
+ * intel_audio_codec_enable - Enable the audio codec for HD audio
+ * @encoder: encoder on which to enable audio
+ * @crtc_state: pointer to the current crtc state.
+ * @conn_state: pointer to the current connector state.
+ *
+ * The enable sequences may only be performed after enabling the transcoder and
+ * port, and after completed link training.
+ */
+void intel_audio_codec_enable(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state,
+			      const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct i915_audio_component *acomp = dev_priv->display.audio.component;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_connector *connector = conn_state->connector;
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	enum port port = encoder->port;
+	enum pipe pipe = crtc->pipe;
+
+	if (!crtc_state->has_audio)
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "[CONNECTOR:%d:%s][ENCODER:%d:%s] Enable audio codec on pipe %c, %u bytes ELD\n",
+		    connector->base.id, connector->name,
+		    encoder->base.base.id, encoder->base.name,
+		    pipe_name(pipe), drm_eld_size(connector->eld));
+
+	/* FIXME precompute the ELD in .compute_config() */
+	if (!connector->eld[0])
+		drm_dbg_kms(&dev_priv->drm,
+			    "Bogus ELD on [CONNECTOR:%d:%s]\n",
+			    connector->base.id, connector->name);
+
+	connector->eld[6] = drm_av_sync_delay(connector, adjusted_mode) / 2;
+
+	if (dev_priv->display.funcs.audio)
+		dev_priv->display.funcs.audio->audio_codec_enable(encoder,
+								  crtc_state,
+								  conn_state);
+
+	mutex_lock(&dev_priv->display.audio.mutex);
+	encoder->audio_connector = connector;
+
+	/* referred in audio callbacks */
+	dev_priv->display.audio.encoder_map[pipe] = encoder;
+	mutex_unlock(&dev_priv->display.audio.mutex);
+
+	if (acomp && acomp->base.audio_ops &&
+	    acomp->base.audio_ops->pin_eld_notify) {
+		/* audio drivers expect pipe = -1 to indicate Non-MST cases */
+		if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST))
+			pipe = -1;
+		acomp->base.audio_ops->pin_eld_notify(acomp->base.audio_ops->audio_ptr,
+						 (int) port, (int) pipe);
+	}
+
+	intel_lpe_audio_notify(dev_priv, pipe, port, connector->eld,
+			       crtc_state->port_clock,
+			       intel_crtc_has_dp_encoder(crtc_state));
+}
+
+/**
+ * intel_audio_codec_disable - Disable the audio codec for HD audio
+ * @encoder: encoder on which to disable audio
+ * @old_crtc_state: pointer to the old crtc state.
+ * @old_conn_state: pointer to the old connector state.
+ *
+ * The disable sequences must be performed before disabling the transcoder or
+ * port.
+ */
+void intel_audio_codec_disable(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *old_crtc_state,
+			       const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct i915_audio_component *acomp = dev_priv->display.audio.component;
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	struct drm_connector *connector = old_conn_state->connector;
+	enum port port = encoder->port;
+	enum pipe pipe = crtc->pipe;
+
+	if (!old_crtc_state->has_audio)
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "[CONNECTOR:%d:%s][ENCODER:%d:%s] Disable audio codec on pipe %c\n",
+		    connector->base.id, connector->name,
+		    encoder->base.base.id, encoder->base.name, pipe_name(pipe));
+
+	if (dev_priv->display.funcs.audio)
+		dev_priv->display.funcs.audio->audio_codec_disable(encoder,
+								   old_crtc_state,
+								   old_conn_state);
+
+	mutex_lock(&dev_priv->display.audio.mutex);
+	encoder->audio_connector = NULL;
+	dev_priv->display.audio.encoder_map[pipe] = NULL;
+	mutex_unlock(&dev_priv->display.audio.mutex);
+
+	if (acomp && acomp->base.audio_ops &&
+	    acomp->base.audio_ops->pin_eld_notify) {
+		/* audio drivers expect pipe = -1 to indicate Non-MST cases */
+		if (!intel_crtc_has_type(old_crtc_state, INTEL_OUTPUT_DP_MST))
+			pipe = -1;
+		acomp->base.audio_ops->pin_eld_notify(acomp->base.audio_ops->audio_ptr,
+						 (int) port, (int) pipe);
+	}
+
+	intel_lpe_audio_notify(dev_priv, pipe, port, NULL, 0, false);
+}
+
+static const struct intel_audio_funcs g4x_audio_funcs = {
+	.audio_codec_enable = g4x_audio_codec_enable,
+	.audio_codec_disable = g4x_audio_codec_disable,
+};
+
+static const struct intel_audio_funcs ilk_audio_funcs = {
+	.audio_codec_enable = ilk_audio_codec_enable,
+	.audio_codec_disable = ilk_audio_codec_disable,
+};
+
+static const struct intel_audio_funcs hsw_audio_funcs = {
+	.audio_codec_enable = hsw_audio_codec_enable,
+	.audio_codec_disable = hsw_audio_codec_disable,
+};
+
+/**
+ * intel_audio_hooks_init - Set up chip specific audio hooks
+ * @dev_priv: device private
+ */
+void intel_audio_hooks_init(struct drm_i915_private *dev_priv)
+{
+	if (IS_G4X(dev_priv)) {
+		dev_priv->display.funcs.audio = &g4x_audio_funcs;
+	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		dev_priv->display.funcs.audio = &ilk_audio_funcs;
+	} else if (IS_HASWELL(dev_priv) || DISPLAY_VER(dev_priv) >= 8) {
+		dev_priv->display.funcs.audio = &hsw_audio_funcs;
+	} else if (HAS_PCH_SPLIT(dev_priv)) {
+		dev_priv->display.funcs.audio = &ilk_audio_funcs;
+	}
+}
+
+struct aud_ts_cdclk_m_n {
+	u8 m;
+	u16 n;
+};
+
+void intel_audio_cdclk_change_pre(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 13)
+		intel_de_rmw(i915, AUD_TS_CDCLK_M, AUD_TS_CDCLK_M_EN, 0);
+}
+
+static void get_aud_ts_cdclk_m_n(int refclk, int cdclk, struct aud_ts_cdclk_m_n *aud_ts)
+{
+	if (refclk == 24000)
+		aud_ts->m = 12;
+	else
+		aud_ts->m = 15;
+
+	aud_ts->n = cdclk * aud_ts->m / 24000;
+}
+
+void intel_audio_cdclk_change_post(struct drm_i915_private *i915)
+{
+	struct aud_ts_cdclk_m_n aud_ts;
+
+	if (DISPLAY_VER(i915) >= 13) {
+		get_aud_ts_cdclk_m_n(i915->display.cdclk.hw.ref, i915->display.cdclk.hw.cdclk, &aud_ts);
+
+		intel_de_write(i915, AUD_TS_CDCLK_N, aud_ts.n);
+		intel_de_write(i915, AUD_TS_CDCLK_M, aud_ts.m | AUD_TS_CDCLK_M_EN);
+		drm_dbg_kms(&i915->drm, "aud_ts_cdclk set to M=%u, N=%u\n", aud_ts.m, aud_ts.n);
+	}
+}
+
+static int glk_force_audio_cdclk_commit(struct intel_atomic_state *state,
+					struct intel_crtc *crtc,
+					bool enable)
+{
+	struct intel_cdclk_state *cdclk_state;
+	int ret;
+
+	/* need to hold at least one crtc lock for the global state */
+	ret = drm_modeset_lock(&crtc->base.mutex, state->base.acquire_ctx);
+	if (ret)
+		return ret;
+
+	cdclk_state = intel_atomic_get_cdclk_state(state);
+	if (IS_ERR(cdclk_state))
+		return PTR_ERR(cdclk_state);
+
+	cdclk_state->force_min_cdclk = enable ? 2 * 96000 : 0;
+
+	return drm_atomic_commit(&state->base);
+}
+
+static void glk_force_audio_cdclk(struct drm_i915_private *dev_priv,
+				  bool enable)
+{
+	struct drm_modeset_acquire_ctx ctx;
+	struct drm_atomic_state *state;
+	struct intel_crtc *crtc;
+	int ret;
+
+	crtc = intel_first_crtc(dev_priv);
+	if (!crtc)
+		return;
+
+	drm_modeset_acquire_init(&ctx, 0);
+	state = drm_atomic_state_alloc(&dev_priv->drm);
+	if (drm_WARN_ON(&dev_priv->drm, !state))
+		return;
+
+	state->acquire_ctx = &ctx;
+
+retry:
+	ret = glk_force_audio_cdclk_commit(to_intel_atomic_state(state), crtc,
+					   enable);
+	if (ret == -EDEADLK) {
+		drm_atomic_state_clear(state);
+		drm_modeset_backoff(&ctx);
+		goto retry;
+	}
+
+	drm_WARN_ON(&dev_priv->drm, ret);
+
+	drm_atomic_state_put(state);
+
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+}
+
+static unsigned long i915_audio_component_get_power(struct device *kdev)
+{
+	struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
+	intel_wakeref_t ret;
+
+	/* Catch potential impedance mismatches before they occur! */
+	BUILD_BUG_ON(sizeof(intel_wakeref_t) > sizeof(unsigned long));
+
+	ret = intel_display_power_get(dev_priv, POWER_DOMAIN_AUDIO_PLAYBACK);
+
+	if (dev_priv->display.audio.power_refcount++ == 0) {
+		if (DISPLAY_VER(dev_priv) >= 9) {
+			intel_de_write(dev_priv, AUD_FREQ_CNTRL,
+				       dev_priv->display.audio.freq_cntrl);
+			drm_dbg_kms(&dev_priv->drm,
+				    "restored AUD_FREQ_CNTRL to 0x%x\n",
+				    dev_priv->display.audio.freq_cntrl);
+		}
+
+		/* Force CDCLK to 2*BCLK as long as we need audio powered. */
+		if (IS_GEMINILAKE(dev_priv))
+			glk_force_audio_cdclk(dev_priv, true);
+
+		if (DISPLAY_VER(dev_priv) >= 10)
+			intel_de_write(dev_priv, AUD_PIN_BUF_CTL,
+				       (intel_de_read(dev_priv, AUD_PIN_BUF_CTL) | AUD_PIN_BUF_ENABLE));
+	}
+
+	return ret;
+}
+
+static void i915_audio_component_put_power(struct device *kdev,
+					   unsigned long cookie)
+{
+	struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
+
+	/* Stop forcing CDCLK to 2*BCLK if no need for audio to be powered. */
+	if (--dev_priv->display.audio.power_refcount == 0)
+		if (IS_GEMINILAKE(dev_priv))
+			glk_force_audio_cdclk(dev_priv, false);
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_AUDIO_PLAYBACK, cookie);
+}
+
+static void i915_audio_component_codec_wake_override(struct device *kdev,
+						     bool enable)
+{
+	struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
+	unsigned long cookie;
+	u32 tmp;
+
+	if (DISPLAY_VER(dev_priv) < 9)
+		return;
+
+	cookie = i915_audio_component_get_power(kdev);
+
+	/*
+	 * Enable/disable generating the codec wake signal, overriding the
+	 * internal logic to generate the codec wake to controller.
+	 */
+	tmp = intel_de_read(dev_priv, HSW_AUD_CHICKENBIT);
+	tmp &= ~SKL_AUD_CODEC_WAKE_SIGNAL;
+	intel_de_write(dev_priv, HSW_AUD_CHICKENBIT, tmp);
+	usleep_range(1000, 1500);
+
+	if (enable) {
+		tmp = intel_de_read(dev_priv, HSW_AUD_CHICKENBIT);
+		tmp |= SKL_AUD_CODEC_WAKE_SIGNAL;
+		intel_de_write(dev_priv, HSW_AUD_CHICKENBIT, tmp);
+		usleep_range(1000, 1500);
+	}
+
+	i915_audio_component_put_power(kdev, cookie);
+}
+
+/* Get CDCLK in kHz  */
+static int i915_audio_component_get_cdclk_freq(struct device *kdev)
+{
+	struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
+
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !HAS_DDI(dev_priv)))
+		return -ENODEV;
+
+	return dev_priv->display.cdclk.hw.cdclk;
+}
+
+/*
+ * get the intel_encoder according to the parameter port and pipe
+ * intel_encoder is saved by the index of pipe
+ * MST & (pipe >= 0): return the audio.encoder_map[pipe],
+ *   when port is matched
+ * MST & (pipe < 0): this is invalid
+ * Non-MST & (pipe >= 0): only pipe = 0 (the first device entry)
+ *   will get the right intel_encoder with port matched
+ * Non-MST & (pipe < 0): get the right intel_encoder with port matched
+ */
+static struct intel_encoder *get_saved_enc(struct drm_i915_private *dev_priv,
+					       int port, int pipe)
+{
+	struct intel_encoder *encoder;
+
+	/* MST */
+	if (pipe >= 0) {
+		if (drm_WARN_ON(&dev_priv->drm,
+				pipe >= ARRAY_SIZE(dev_priv->display.audio.encoder_map)))
+			return NULL;
+
+		encoder = dev_priv->display.audio.encoder_map[pipe];
+		/*
+		 * when bootup, audio driver may not know it is
+		 * MST or not. So it will poll all the port & pipe
+		 * combinations
+		 */
+		if (encoder != NULL && encoder->port == port &&
+		    encoder->type == INTEL_OUTPUT_DP_MST)
+			return encoder;
+	}
+
+	/* Non-MST */
+	if (pipe > 0)
+		return NULL;
+
+	for_each_pipe(dev_priv, pipe) {
+		encoder = dev_priv->display.audio.encoder_map[pipe];
+		if (encoder == NULL)
+			continue;
+
+		if (encoder->type == INTEL_OUTPUT_DP_MST)
+			continue;
+
+		if (port == encoder->port)
+			return encoder;
+	}
+
+	return NULL;
+}
+
+static int i915_audio_component_sync_audio_rate(struct device *kdev, int port,
+						int pipe, int rate)
+{
+	struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
+	struct i915_audio_component *acomp = dev_priv->display.audio.component;
+	struct intel_encoder *encoder;
+	struct intel_crtc *crtc;
+	unsigned long cookie;
+	int err = 0;
+
+	if (!HAS_DDI(dev_priv))
+		return 0;
+
+	cookie = i915_audio_component_get_power(kdev);
+	mutex_lock(&dev_priv->display.audio.mutex);
+
+	/* 1. get the pipe */
+	encoder = get_saved_enc(dev_priv, port, pipe);
+	if (!encoder || !encoder->base.crtc) {
+		drm_dbg_kms(&dev_priv->drm, "Not valid for port %c\n",
+			    port_name(port));
+		err = -ENODEV;
+		goto unlock;
+	}
+
+	crtc = to_intel_crtc(encoder->base.crtc);
+
+	/* port must be valid now, otherwise the pipe will be invalid */
+	acomp->aud_sample_rate[port] = rate;
+
+	hsw_audio_config_update(encoder, crtc->config);
+
+ unlock:
+	mutex_unlock(&dev_priv->display.audio.mutex);
+	i915_audio_component_put_power(kdev, cookie);
+	return err;
+}
+
+static int i915_audio_component_get_eld(struct device *kdev, int port,
+					int pipe, bool *enabled,
+					unsigned char *buf, int max_bytes)
+{
+	struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
+	struct intel_encoder *intel_encoder;
+	const u8 *eld;
+	int ret = -EINVAL;
+
+	mutex_lock(&dev_priv->display.audio.mutex);
+
+	intel_encoder = get_saved_enc(dev_priv, port, pipe);
+	if (!intel_encoder) {
+		drm_dbg_kms(&dev_priv->drm, "Not valid for port %c\n",
+			    port_name(port));
+		mutex_unlock(&dev_priv->display.audio.mutex);
+		return ret;
+	}
+
+	ret = 0;
+	*enabled = intel_encoder->audio_connector != NULL;
+	if (*enabled) {
+		eld = intel_encoder->audio_connector->eld;
+		ret = drm_eld_size(eld);
+		memcpy(buf, eld, min(max_bytes, ret));
+	}
+
+	mutex_unlock(&dev_priv->display.audio.mutex);
+	return ret;
+}
+
+static const struct drm_audio_component_ops i915_audio_component_ops = {
+	.owner		= THIS_MODULE,
+	.get_power	= i915_audio_component_get_power,
+	.put_power	= i915_audio_component_put_power,
+	.codec_wake_override = i915_audio_component_codec_wake_override,
+	.get_cdclk_freq	= i915_audio_component_get_cdclk_freq,
+	.sync_audio_rate = i915_audio_component_sync_audio_rate,
+	.get_eld	= i915_audio_component_get_eld,
+};
+
+static int i915_audio_component_bind(struct device *i915_kdev,
+				     struct device *hda_kdev, void *data)
+{
+	struct i915_audio_component *acomp = data;
+	struct drm_i915_private *dev_priv = kdev_to_i915(i915_kdev);
+	int i;
+
+	if (drm_WARN_ON(&dev_priv->drm, acomp->base.ops || acomp->base.dev))
+		return -EEXIST;
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			!device_link_add(hda_kdev, i915_kdev,
+					 DL_FLAG_STATELESS)))
+		return -ENOMEM;
+
+	drm_modeset_lock_all(&dev_priv->drm);
+	acomp->base.ops = &i915_audio_component_ops;
+	acomp->base.dev = i915_kdev;
+	BUILD_BUG_ON(MAX_PORTS != I915_MAX_PORTS);
+	for (i = 0; i < ARRAY_SIZE(acomp->aud_sample_rate); i++)
+		acomp->aud_sample_rate[i] = 0;
+	dev_priv->display.audio.component = acomp;
+	drm_modeset_unlock_all(&dev_priv->drm);
+
+	return 0;
+}
+
+static void i915_audio_component_unbind(struct device *i915_kdev,
+					struct device *hda_kdev, void *data)
+{
+	struct i915_audio_component *acomp = data;
+	struct drm_i915_private *dev_priv = kdev_to_i915(i915_kdev);
+
+	drm_modeset_lock_all(&dev_priv->drm);
+	acomp->base.ops = NULL;
+	acomp->base.dev = NULL;
+	dev_priv->display.audio.component = NULL;
+	drm_modeset_unlock_all(&dev_priv->drm);
+
+	device_link_remove(hda_kdev, i915_kdev);
+
+	if (dev_priv->display.audio.power_refcount)
+		drm_err(&dev_priv->drm, "audio power refcount %d after unbind\n",
+			dev_priv->display.audio.power_refcount);
+}
+
+static const struct component_ops i915_audio_component_bind_ops = {
+	.bind	= i915_audio_component_bind,
+	.unbind	= i915_audio_component_unbind,
+};
+
+#define AUD_FREQ_TMODE_SHIFT	14
+#define AUD_FREQ_4T		0
+#define AUD_FREQ_8T		(2 << AUD_FREQ_TMODE_SHIFT)
+#define AUD_FREQ_PULLCLKS(x)	(((x) & 0x3) << 11)
+#define AUD_FREQ_BCLK_96M	BIT(4)
+
+#define AUD_FREQ_GEN12          (AUD_FREQ_8T | AUD_FREQ_PULLCLKS(0) | AUD_FREQ_BCLK_96M)
+#define AUD_FREQ_TGL_BROKEN     (AUD_FREQ_8T | AUD_FREQ_PULLCLKS(2) | AUD_FREQ_BCLK_96M)
+
+/**
+ * i915_audio_component_init - initialize and register the audio component
+ * @dev_priv: i915 device instance
+ *
+ * This will register with the component framework a child component which
+ * will bind dynamically to the snd_hda_intel driver's corresponding master
+ * component when the latter is registered. During binding the child
+ * initializes an instance of struct i915_audio_component which it receives
+ * from the master. The master can then start to use the interface defined by
+ * this struct. Each side can break the binding at any point by deregistering
+ * its own component after which each side's component unbind callback is
+ * called.
+ *
+ * We ignore any error during registration and continue with reduced
+ * functionality (i.e. without HDMI audio).
+ */
+static void i915_audio_component_init(struct drm_i915_private *dev_priv)
+{
+	u32 aud_freq, aud_freq_init;
+	int ret;
+
+	ret = component_add_typed(dev_priv->drm.dev,
+				  &i915_audio_component_bind_ops,
+				  I915_COMPONENT_AUDIO);
+	if (ret < 0) {
+		drm_err(&dev_priv->drm,
+			"failed to add audio component (%d)\n", ret);
+		/* continue with reduced functionality */
+		return;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 9) {
+		aud_freq_init = intel_de_read(dev_priv, AUD_FREQ_CNTRL);
+
+		if (DISPLAY_VER(dev_priv) >= 12)
+			aud_freq = AUD_FREQ_GEN12;
+		else
+			aud_freq = aud_freq_init;
+
+		/* use BIOS provided value for TGL and RKL unless it is a known bad value */
+		if ((IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv)) &&
+		    aud_freq_init != AUD_FREQ_TGL_BROKEN)
+			aud_freq = aud_freq_init;
+
+		drm_dbg_kms(&dev_priv->drm, "use AUD_FREQ_CNTRL of 0x%x (init value 0x%x)\n",
+			    aud_freq, aud_freq_init);
+
+		dev_priv->display.audio.freq_cntrl = aud_freq;
+	}
+
+	/* init with current cdclk */
+	intel_audio_cdclk_change_post(dev_priv);
+
+	dev_priv->display.audio.component_registered = true;
+}
+
+/**
+ * i915_audio_component_cleanup - deregister the audio component
+ * @dev_priv: i915 device instance
+ *
+ * Deregisters the audio component, breaking any existing binding to the
+ * corresponding snd_hda_intel driver's master component.
+ */
+static void i915_audio_component_cleanup(struct drm_i915_private *dev_priv)
+{
+	if (!dev_priv->display.audio.component_registered)
+		return;
+
+	component_del(dev_priv->drm.dev, &i915_audio_component_bind_ops);
+	dev_priv->display.audio.component_registered = false;
+}
+
+/**
+ * intel_audio_init() - Initialize the audio driver either using
+ * component framework or using lpe audio bridge
+ * @dev_priv: the i915 drm device private data
+ *
+ */
+void intel_audio_init(struct drm_i915_private *dev_priv)
+{
+	if (intel_lpe_audio_init(dev_priv) < 0)
+		i915_audio_component_init(dev_priv);
+}
+
+/**
+ * intel_audio_deinit() - deinitialize the audio driver
+ * @dev_priv: the i915 drm device private data
+ *
+ */
+void intel_audio_deinit(struct drm_i915_private *dev_priv)
+{
+	if (dev_priv->display.audio.lpe.platdev != NULL)
+		intel_lpe_audio_teardown(dev_priv);
+	else
+		i915_audio_component_cleanup(dev_priv);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_audio.h b/drivers/gpu/drm/i915/display/intel_audio.h
new file mode 100644
index 000000000..63b22131d
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_audio.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_AUDIO_H__
+#define __INTEL_AUDIO_H__
+
+struct drm_connector_state;
+struct drm_i915_private;
+struct intel_crtc_state;
+struct intel_encoder;
+
+void intel_audio_hooks_init(struct drm_i915_private *dev_priv);
+void intel_audio_codec_enable(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state,
+			      const struct drm_connector_state *conn_state);
+void intel_audio_codec_disable(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *old_crtc_state,
+			       const struct drm_connector_state *old_conn_state);
+void intel_audio_cdclk_change_pre(struct drm_i915_private *dev_priv);
+void intel_audio_cdclk_change_post(struct drm_i915_private *dev_priv);
+void intel_audio_init(struct drm_i915_private *dev_priv);
+void intel_audio_deinit(struct drm_i915_private *dev_priv);
+
+#endif /* __INTEL_AUDIO_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_audio_regs.h b/drivers/gpu/drm/i915/display/intel_audio_regs.h
new file mode 100644
index 000000000..d1e5844e3
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_audio_regs.h
@@ -0,0 +1,160 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_AUDIO_REGS_H__
+#define __INTEL_AUDIO_REGS_H__
+
+#include "i915_reg_defs.h"
+
+#define G4X_AUD_VID_DID			_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x62020)
+#define   INTEL_AUDIO_DEVCL		0x808629FB
+#define   INTEL_AUDIO_DEVBLC		0x80862801
+#define   INTEL_AUDIO_DEVCTG		0x80862802
+
+#define G4X_AUD_CNTL_ST			_MMIO(0x620B4)
+#define   G4X_ELDV_DEVCL_DEVBLC		(1 << 13)
+#define   G4X_ELDV_DEVCTG		(1 << 14)
+#define   G4X_ELD_ADDR_MASK		(0xf << 5)
+#define   G4X_ELD_ACK			(1 << 4)
+#define G4X_HDMIW_HDMIEDID		_MMIO(0x6210C)
+
+#define _IBX_HDMIW_HDMIEDID_A		0xE2050
+#define _IBX_HDMIW_HDMIEDID_B		0xE2150
+#define IBX_HDMIW_HDMIEDID(pipe)	_MMIO_PIPE(pipe, _IBX_HDMIW_HDMIEDID_A, \
+						  _IBX_HDMIW_HDMIEDID_B)
+#define _IBX_AUD_CNTL_ST_A		0xE20B4
+#define _IBX_AUD_CNTL_ST_B		0xE21B4
+#define IBX_AUD_CNTL_ST(pipe)		_MMIO_PIPE(pipe, _IBX_AUD_CNTL_ST_A, \
+						  _IBX_AUD_CNTL_ST_B)
+#define   IBX_ELD_BUFFER_SIZE_MASK	(0x1f << 10)
+#define   IBX_ELD_ADDRESS_MASK		(0x1f << 5)
+#define   IBX_ELD_ACK			(1 << 4)
+#define IBX_AUD_CNTL_ST2		_MMIO(0xE20C0)
+#define   IBX_CP_READY(port)		((1 << 1) << (((port) - 1) * 4))
+#define   IBX_ELD_VALID(port)		((1 << 0) << (((port) - 1) * 4))
+
+#define _CPT_HDMIW_HDMIEDID_A		0xE5050
+#define _CPT_HDMIW_HDMIEDID_B		0xE5150
+#define CPT_HDMIW_HDMIEDID(pipe)	_MMIO_PIPE(pipe, _CPT_HDMIW_HDMIEDID_A, _CPT_HDMIW_HDMIEDID_B)
+#define _CPT_AUD_CNTL_ST_A		0xE50B4
+#define _CPT_AUD_CNTL_ST_B		0xE51B4
+#define CPT_AUD_CNTL_ST(pipe)		_MMIO_PIPE(pipe, _CPT_AUD_CNTL_ST_A, _CPT_AUD_CNTL_ST_B)
+#define CPT_AUD_CNTRL_ST2		_MMIO(0xE50C0)
+
+#define _VLV_HDMIW_HDMIEDID_A		(VLV_DISPLAY_BASE + 0x62050)
+#define _VLV_HDMIW_HDMIEDID_B		(VLV_DISPLAY_BASE + 0x62150)
+#define VLV_HDMIW_HDMIEDID(pipe)	_MMIO_PIPE(pipe, _VLV_HDMIW_HDMIEDID_A, _VLV_HDMIW_HDMIEDID_B)
+#define _VLV_AUD_CNTL_ST_A		(VLV_DISPLAY_BASE + 0x620B4)
+#define _VLV_AUD_CNTL_ST_B		(VLV_DISPLAY_BASE + 0x621B4)
+#define VLV_AUD_CNTL_ST(pipe)		_MMIO_PIPE(pipe, _VLV_AUD_CNTL_ST_A, _VLV_AUD_CNTL_ST_B)
+#define VLV_AUD_CNTL_ST2		_MMIO(VLV_DISPLAY_BASE + 0x620C0)
+
+#define _IBX_AUD_CONFIG_A		0xe2000
+#define _IBX_AUD_CONFIG_B		0xe2100
+#define IBX_AUD_CFG(pipe)		_MMIO_PIPE(pipe, _IBX_AUD_CONFIG_A, _IBX_AUD_CONFIG_B)
+#define _CPT_AUD_CONFIG_A		0xe5000
+#define _CPT_AUD_CONFIG_B		0xe5100
+#define CPT_AUD_CFG(pipe)		_MMIO_PIPE(pipe, _CPT_AUD_CONFIG_A, _CPT_AUD_CONFIG_B)
+#define _VLV_AUD_CONFIG_A		(VLV_DISPLAY_BASE + 0x62000)
+#define _VLV_AUD_CONFIG_B		(VLV_DISPLAY_BASE + 0x62100)
+#define VLV_AUD_CFG(pipe)		_MMIO_PIPE(pipe, _VLV_AUD_CONFIG_A, _VLV_AUD_CONFIG_B)
+
+#define   AUD_CONFIG_N_VALUE_INDEX		(1 << 29)
+#define   AUD_CONFIG_N_PROG_ENABLE		(1 << 28)
+#define   AUD_CONFIG_UPPER_N_SHIFT		20
+#define   AUD_CONFIG_UPPER_N_MASK		(0xff << 20)
+#define   AUD_CONFIG_LOWER_N_SHIFT		4
+#define   AUD_CONFIG_LOWER_N_MASK		(0xfff << 4)
+#define   AUD_CONFIG_N_MASK			(AUD_CONFIG_UPPER_N_MASK | AUD_CONFIG_LOWER_N_MASK)
+#define   AUD_CONFIG_N(n) \
+	(((((n) >> 12) & 0xff) << AUD_CONFIG_UPPER_N_SHIFT) |	\
+	 (((n) & 0xfff) << AUD_CONFIG_LOWER_N_SHIFT))
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_SHIFT	16
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_MASK	(0xf << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_25175	(0 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_25200	(1 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_27000	(2 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_27027	(3 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_54000	(4 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_54054	(5 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_74176	(6 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_74250	(7 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_148352	(8 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_148500	(9 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_296703	(10 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_297000	(11 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_593407	(12 << 16)
+#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_594000	(13 << 16)
+#define   AUD_CONFIG_DISABLE_NCTS		(1 << 3)
+
+#define _HSW_AUD_CONFIG_A		0x65000
+#define _HSW_AUD_CONFIG_B		0x65100
+#define HSW_AUD_CFG(trans)		_MMIO_TRANS(trans, _HSW_AUD_CONFIG_A, _HSW_AUD_CONFIG_B)
+
+#define _HSW_AUD_MISC_CTRL_A		0x65010
+#define _HSW_AUD_MISC_CTRL_B		0x65110
+#define HSW_AUD_MISC_CTRL(trans)	_MMIO_TRANS(trans, _HSW_AUD_MISC_CTRL_A, _HSW_AUD_MISC_CTRL_B)
+
+#define _HSW_AUD_M_CTS_ENABLE_A		0x65028
+#define _HSW_AUD_M_CTS_ENABLE_B		0x65128
+#define HSW_AUD_M_CTS_ENABLE(trans)	_MMIO_TRANS(trans, _HSW_AUD_M_CTS_ENABLE_A, _HSW_AUD_M_CTS_ENABLE_B)
+#define   AUD_M_CTS_M_VALUE_INDEX	(1 << 21)
+#define   AUD_M_CTS_M_PROG_ENABLE	(1 << 20)
+#define   AUD_CONFIG_M_MASK		0xfffff
+
+#define _HSW_AUD_DIP_ELD_CTRL_ST_A	0x650b4
+#define _HSW_AUD_DIP_ELD_CTRL_ST_B	0x651b4
+#define HSW_AUD_DIP_ELD_CTRL(trans)	_MMIO_TRANS(trans, _HSW_AUD_DIP_ELD_CTRL_ST_A, _HSW_AUD_DIP_ELD_CTRL_ST_B)
+
+/* Audio Digital Converter */
+#define _HSW_AUD_DIG_CNVT_1		0x65080
+#define _HSW_AUD_DIG_CNVT_2		0x65180
+#define AUD_DIG_CNVT(trans)		_MMIO_TRANS(trans, _HSW_AUD_DIG_CNVT_1, _HSW_AUD_DIG_CNVT_2)
+#define DIP_PORT_SEL_MASK		0x3
+
+#define _HSW_AUD_EDID_DATA_A		0x65050
+#define _HSW_AUD_EDID_DATA_B		0x65150
+#define HSW_AUD_EDID_DATA(trans)	_MMIO_TRANS(trans, _HSW_AUD_EDID_DATA_A, _HSW_AUD_EDID_DATA_B)
+
+#define HSW_AUD_PIPE_CONV_CFG		_MMIO(0x6507c)
+#define HSW_AUD_PIN_ELD_CP_VLD		_MMIO(0x650c0)
+#define   AUDIO_INACTIVE(trans)		((1 << 3) << ((trans) * 4))
+#define   AUDIO_OUTPUT_ENABLE(trans)	((1 << 2) << ((trans) * 4))
+#define   AUDIO_CP_READY(trans)		((1 << 1) << ((trans) * 4))
+#define   AUDIO_ELD_VALID(trans)	((1 << 0) << ((trans) * 4))
+
+#define _AUD_TCA_DP_2DOT0_CTRL		0x650bc
+#define _AUD_TCB_DP_2DOT0_CTRL		0x651bc
+#define AUD_DP_2DOT0_CTRL(trans)	_MMIO_TRANS(trans, _AUD_TCA_DP_2DOT0_CTRL, _AUD_TCB_DP_2DOT0_CTRL)
+#define  AUD_ENABLE_SDP_SPLIT		REG_BIT(31)
+
+#define HSW_AUD_CHICKENBIT			_MMIO(0x65f10)
+#define   SKL_AUD_CODEC_WAKE_SIGNAL		(1 << 15)
+
+#define AUD_FREQ_CNTRL			_MMIO(0x65900)
+#define AUD_PIN_BUF_CTL		_MMIO(0x48414)
+#define   AUD_PIN_BUF_ENABLE		REG_BIT(31)
+
+#define AUD_TS_CDCLK_M			_MMIO(0x65ea0)
+#define   AUD_TS_CDCLK_M_EN		REG_BIT(31)
+#define AUD_TS_CDCLK_N			_MMIO(0x65ea4)
+
+/* Display Audio Config Reg */
+#define AUD_CONFIG_BE			_MMIO(0x65ef0)
+#define HBLANK_EARLY_ENABLE_ICL(pipe)		(0x1 << (20 - (pipe)))
+#define HBLANK_EARLY_ENABLE_TGL(pipe)		(0x1 << (24 + (pipe)))
+#define HBLANK_START_COUNT_MASK(pipe)		(0x7 << (3 + ((pipe) * 6)))
+#define HBLANK_START_COUNT(pipe, val)		(((val) & 0x7) << (3 + ((pipe)) * 6))
+#define NUMBER_SAMPLES_PER_LINE_MASK(pipe)	(0x3 << ((pipe) * 6))
+#define NUMBER_SAMPLES_PER_LINE(pipe, val)	(((val) & 0x3) << ((pipe) * 6))
+
+#define HBLANK_START_COUNT_8	0
+#define HBLANK_START_COUNT_16	1
+#define HBLANK_START_COUNT_32	2
+#define HBLANK_START_COUNT_64	3
+#define HBLANK_START_COUNT_96	4
+#define HBLANK_START_COUNT_128	5
+
+#endif /* __INTEL_AUDIO_REGS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_backlight.c b/drivers/gpu/drm/i915/display/intel_backlight.c
new file mode 100644
index 000000000..beba39a38
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_backlight.c
@@ -0,0 +1,1794 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/backlight.h>
+#include <linux/kernel.h>
+#include <linux/pwm.h>
+#include <linux/string_helpers.h>
+
+#include <acpi/video.h>
+
+#include "intel_backlight.h"
+#include "intel_backlight_regs.h"
+#include "intel_connector.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dp_aux_backlight.h"
+#include "intel_dsi_dcs_backlight.h"
+#include "intel_panel.h"
+#include "intel_pci_config.h"
+#include "intel_pps.h"
+#include "intel_quirks.h"
+
+/**
+ * scale - scale values from one range to another
+ * @source_val: value in range [@source_min..@source_max]
+ * @source_min: minimum legal value for @source_val
+ * @source_max: maximum legal value for @source_val
+ * @target_min: corresponding target value for @source_min
+ * @target_max: corresponding target value for @source_max
+ *
+ * Return @source_val in range [@source_min..@source_max] scaled to range
+ * [@target_min..@target_max].
+ */
+static u32 scale(u32 source_val,
+		 u32 source_min, u32 source_max,
+		 u32 target_min, u32 target_max)
+{
+	u64 target_val;
+
+	WARN_ON(source_min > source_max);
+	WARN_ON(target_min > target_max);
+
+	/* defensive */
+	source_val = clamp(source_val, source_min, source_max);
+
+	/* avoid overflows */
+	target_val = mul_u32_u32(source_val - source_min,
+				 target_max - target_min);
+	target_val = DIV_ROUND_CLOSEST_ULL(target_val, source_max - source_min);
+	target_val += target_min;
+
+	return target_val;
+}
+
+/*
+ * Scale user_level in range [0..user_max] to [0..hw_max], clamping the result
+ * to [hw_min..hw_max].
+ */
+static u32 clamp_user_to_hw(struct intel_connector *connector,
+			    u32 user_level, u32 user_max)
+{
+	struct intel_panel *panel = &connector->panel;
+	u32 hw_level;
+
+	hw_level = scale(user_level, 0, user_max, 0, panel->backlight.max);
+	hw_level = clamp(hw_level, panel->backlight.min, panel->backlight.max);
+
+	return hw_level;
+}
+
+/* Scale hw_level in range [hw_min..hw_max] to [0..user_max]. */
+static u32 scale_hw_to_user(struct intel_connector *connector,
+			    u32 hw_level, u32 user_max)
+{
+	struct intel_panel *panel = &connector->panel;
+
+	return scale(hw_level, panel->backlight.min, panel->backlight.max,
+		     0, user_max);
+}
+
+u32 intel_backlight_invert_pwm_level(struct intel_connector *connector, u32 val)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	drm_WARN_ON(&dev_priv->drm, panel->backlight.pwm_level_max == 0);
+
+	if (dev_priv->params.invert_brightness < 0)
+		return val;
+
+	if (dev_priv->params.invert_brightness > 0 ||
+	    intel_has_quirk(dev_priv, QUIRK_INVERT_BRIGHTNESS)) {
+		return panel->backlight.pwm_level_max - val + panel->backlight.pwm_level_min;
+	}
+
+	return val;
+}
+
+void intel_backlight_set_pwm_level(const struct drm_connector_state *conn_state, u32 val)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	drm_dbg_kms(&i915->drm, "set backlight PWM = %d\n", val);
+	panel->backlight.pwm_funcs->set(conn_state, val);
+}
+
+u32 intel_backlight_level_to_pwm(struct intel_connector *connector, u32 val)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	drm_WARN_ON_ONCE(&dev_priv->drm,
+			 panel->backlight.max == 0 || panel->backlight.pwm_level_max == 0);
+
+	val = scale(val, panel->backlight.min, panel->backlight.max,
+		    panel->backlight.pwm_level_min, panel->backlight.pwm_level_max);
+
+	return intel_backlight_invert_pwm_level(connector, val);
+}
+
+u32 intel_backlight_level_from_pwm(struct intel_connector *connector, u32 val)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	drm_WARN_ON_ONCE(&dev_priv->drm,
+			 panel->backlight.max == 0 || panel->backlight.pwm_level_max == 0);
+
+	if (dev_priv->params.invert_brightness > 0 ||
+	    (dev_priv->params.invert_brightness == 0 && intel_has_quirk(dev_priv, QUIRK_INVERT_BRIGHTNESS)))
+		val = panel->backlight.pwm_level_max - (val - panel->backlight.pwm_level_min);
+
+	return scale(val, panel->backlight.pwm_level_min, panel->backlight.pwm_level_max,
+		     panel->backlight.min, panel->backlight.max);
+}
+
+static u32 lpt_get_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+
+	return intel_de_read(dev_priv, BLC_PWM_PCH_CTL2) & BACKLIGHT_DUTY_CYCLE_MASK;
+}
+
+static u32 pch_get_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+
+	return intel_de_read(dev_priv, BLC_PWM_CPU_CTL) & BACKLIGHT_DUTY_CYCLE_MASK;
+}
+
+static u32 i9xx_get_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 val;
+
+	val = intel_de_read(dev_priv, BLC_PWM_CTL) & BACKLIGHT_DUTY_CYCLE_MASK;
+	if (DISPLAY_VER(dev_priv) < 4)
+		val >>= 1;
+
+	if (panel->backlight.combination_mode) {
+		u8 lbpc;
+
+		pci_read_config_byte(to_pci_dev(dev_priv->drm.dev), LBPC, &lbpc);
+		val *= lbpc;
+	}
+
+	return val;
+}
+
+static u32 vlv_get_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+
+	if (drm_WARN_ON(&dev_priv->drm, pipe != PIPE_A && pipe != PIPE_B))
+		return 0;
+
+	return intel_de_read(dev_priv, VLV_BLC_PWM_CTL(pipe)) & BACKLIGHT_DUTY_CYCLE_MASK;
+}
+
+static u32 bxt_get_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	return intel_de_read(dev_priv,
+			     BXT_BLC_PWM_DUTY(panel->backlight.controller));
+}
+
+static u32 ext_pwm_get_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct intel_panel *panel = &connector->panel;
+	struct pwm_state state;
+
+	pwm_get_state(panel->backlight.pwm, &state);
+	return pwm_get_relative_duty_cycle(&state, 100);
+}
+
+static void lpt_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+
+	u32 val = intel_de_read(dev_priv, BLC_PWM_PCH_CTL2) & ~BACKLIGHT_DUTY_CYCLE_MASK;
+	intel_de_write(dev_priv, BLC_PWM_PCH_CTL2, val | level);
+}
+
+static void pch_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, BLC_PWM_CPU_CTL) & ~BACKLIGHT_DUTY_CYCLE_MASK;
+	intel_de_write(dev_priv, BLC_PWM_CPU_CTL, tmp | level);
+}
+
+static void i9xx_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 tmp, mask;
+
+	drm_WARN_ON(&dev_priv->drm, panel->backlight.pwm_level_max == 0);
+
+	if (panel->backlight.combination_mode) {
+		u8 lbpc;
+
+		lbpc = level * 0xfe / panel->backlight.pwm_level_max + 1;
+		level /= lbpc;
+		pci_write_config_byte(to_pci_dev(dev_priv->drm.dev), LBPC, lbpc);
+	}
+
+	if (DISPLAY_VER(dev_priv) == 4) {
+		mask = BACKLIGHT_DUTY_CYCLE_MASK;
+	} else {
+		level <<= 1;
+		mask = BACKLIGHT_DUTY_CYCLE_MASK_PNV;
+	}
+
+	tmp = intel_de_read(dev_priv, BLC_PWM_CTL) & ~mask;
+	intel_de_write(dev_priv, BLC_PWM_CTL, tmp | level);
+}
+
+static void vlv_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	enum pipe pipe = to_intel_crtc(conn_state->crtc)->pipe;
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, VLV_BLC_PWM_CTL(pipe)) & ~BACKLIGHT_DUTY_CYCLE_MASK;
+	intel_de_write(dev_priv, VLV_BLC_PWM_CTL(pipe), tmp | level);
+}
+
+static void bxt_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	intel_de_write(dev_priv,
+		       BXT_BLC_PWM_DUTY(panel->backlight.controller), level);
+}
+
+static void ext_pwm_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_panel *panel = &to_intel_connector(conn_state->connector)->panel;
+
+	pwm_set_relative_duty_cycle(&panel->backlight.pwm_state, level, 100);
+	pwm_apply_state(panel->backlight.pwm, &panel->backlight.pwm_state);
+}
+
+static void
+intel_panel_actually_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	drm_dbg_kms(&i915->drm, "set backlight level = %d\n", level);
+
+	panel->backlight.funcs->set(conn_state, level);
+}
+
+/* set backlight brightness to level in range [0..max], assuming hw min is
+ * respected.
+ */
+void intel_backlight_set_acpi(const struct drm_connector_state *conn_state,
+			      u32 user_level, u32 user_max)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 hw_level;
+
+	/*
+	 * Lack of crtc may occur during driver init because
+	 * connection_mutex isn't held across the entire backlight
+	 * setup + modeset readout, and the BIOS can issue the
+	 * requests at any time.
+	 */
+	if (!panel->backlight.present || !conn_state->crtc)
+		return;
+
+	mutex_lock(&dev_priv->display.backlight.lock);
+
+	drm_WARN_ON(&dev_priv->drm, panel->backlight.max == 0);
+
+	hw_level = clamp_user_to_hw(connector, user_level, user_max);
+	panel->backlight.level = hw_level;
+
+	if (panel->backlight.device)
+		panel->backlight.device->props.brightness =
+			scale_hw_to_user(connector,
+					 panel->backlight.level,
+					 panel->backlight.device->props.max_brightness);
+
+	if (panel->backlight.enabled)
+		intel_panel_actually_set_backlight(conn_state, hw_level);
+
+	mutex_unlock(&dev_priv->display.backlight.lock);
+}
+
+static void lpt_disable_backlight(const struct drm_connector_state *old_conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(old_conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	u32 tmp;
+
+	intel_backlight_set_pwm_level(old_conn_state, level);
+
+	/*
+	 * Although we don't support or enable CPU PWM with LPT/SPT based
+	 * systems, it may have been enabled prior to loading the
+	 * driver. Disable to avoid warnings on LCPLL disable.
+	 *
+	 * This needs rework if we need to add support for CPU PWM on PCH split
+	 * platforms.
+	 */
+	tmp = intel_de_read(dev_priv, BLC_PWM_CPU_CTL2);
+	if (tmp & BLM_PWM_ENABLE) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "cpu backlight was enabled, disabling\n");
+		intel_de_write(dev_priv, BLC_PWM_CPU_CTL2,
+			       tmp & ~BLM_PWM_ENABLE);
+	}
+
+	tmp = intel_de_read(dev_priv, BLC_PWM_PCH_CTL1);
+	intel_de_write(dev_priv, BLC_PWM_PCH_CTL1, tmp & ~BLM_PCH_PWM_ENABLE);
+}
+
+static void pch_disable_backlight(const struct drm_connector_state *old_conn_state, u32 val)
+{
+	struct intel_connector *connector = to_intel_connector(old_conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	u32 tmp;
+
+	intel_backlight_set_pwm_level(old_conn_state, val);
+
+	tmp = intel_de_read(dev_priv, BLC_PWM_CPU_CTL2);
+	intel_de_write(dev_priv, BLC_PWM_CPU_CTL2, tmp & ~BLM_PWM_ENABLE);
+
+	tmp = intel_de_read(dev_priv, BLC_PWM_PCH_CTL1);
+	intel_de_write(dev_priv, BLC_PWM_PCH_CTL1, tmp & ~BLM_PCH_PWM_ENABLE);
+}
+
+static void i9xx_disable_backlight(const struct drm_connector_state *old_conn_state, u32 val)
+{
+	intel_backlight_set_pwm_level(old_conn_state, val);
+}
+
+static void i965_disable_backlight(const struct drm_connector_state *old_conn_state, u32 val)
+{
+	struct drm_i915_private *dev_priv = to_i915(old_conn_state->connector->dev);
+	u32 tmp;
+
+	intel_backlight_set_pwm_level(old_conn_state, val);
+
+	tmp = intel_de_read(dev_priv, BLC_PWM_CTL2);
+	intel_de_write(dev_priv, BLC_PWM_CTL2, tmp & ~BLM_PWM_ENABLE);
+}
+
+static void vlv_disable_backlight(const struct drm_connector_state *old_conn_state, u32 val)
+{
+	struct intel_connector *connector = to_intel_connector(old_conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	enum pipe pipe = to_intel_crtc(old_conn_state->crtc)->pipe;
+	u32 tmp;
+
+	intel_backlight_set_pwm_level(old_conn_state, val);
+
+	tmp = intel_de_read(dev_priv, VLV_BLC_PWM_CTL2(pipe));
+	intel_de_write(dev_priv, VLV_BLC_PWM_CTL2(pipe),
+		       tmp & ~BLM_PWM_ENABLE);
+}
+
+static void bxt_disable_backlight(const struct drm_connector_state *old_conn_state, u32 val)
+{
+	struct intel_connector *connector = to_intel_connector(old_conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 tmp;
+
+	intel_backlight_set_pwm_level(old_conn_state, val);
+
+	tmp = intel_de_read(dev_priv,
+			    BXT_BLC_PWM_CTL(panel->backlight.controller));
+	intel_de_write(dev_priv, BXT_BLC_PWM_CTL(panel->backlight.controller),
+		       tmp & ~BXT_BLC_PWM_ENABLE);
+
+	if (panel->backlight.controller == 1) {
+		val = intel_de_read(dev_priv, UTIL_PIN_CTL);
+		val &= ~UTIL_PIN_ENABLE;
+		intel_de_write(dev_priv, UTIL_PIN_CTL, val);
+	}
+}
+
+static void cnp_disable_backlight(const struct drm_connector_state *old_conn_state, u32 val)
+{
+	struct intel_connector *connector = to_intel_connector(old_conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 tmp;
+
+	intel_backlight_set_pwm_level(old_conn_state, val);
+
+	tmp = intel_de_read(dev_priv,
+			    BXT_BLC_PWM_CTL(panel->backlight.controller));
+	intel_de_write(dev_priv, BXT_BLC_PWM_CTL(panel->backlight.controller),
+		       tmp & ~BXT_BLC_PWM_ENABLE);
+}
+
+static void ext_pwm_disable_backlight(const struct drm_connector_state *old_conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(old_conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	intel_backlight_set_pwm_level(old_conn_state, level);
+
+	panel->backlight.pwm_state.enabled = false;
+	pwm_apply_state(panel->backlight.pwm, &panel->backlight.pwm_state);
+}
+
+void intel_backlight_disable(const struct drm_connector_state *old_conn_state)
+{
+	struct intel_connector *connector = to_intel_connector(old_conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	if (!panel->backlight.present)
+		return;
+
+	/*
+	 * Do not disable backlight on the vga_switcheroo path. When switching
+	 * away from i915, the other client may depend on i915 to handle the
+	 * backlight. This will leave the backlight on unnecessarily when
+	 * another client is not activated.
+	 */
+	if (dev_priv->drm.switch_power_state == DRM_SWITCH_POWER_CHANGING) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Skipping backlight disable on vga switch\n");
+		return;
+	}
+
+	mutex_lock(&dev_priv->display.backlight.lock);
+
+	if (panel->backlight.device)
+		panel->backlight.device->props.power = FB_BLANK_POWERDOWN;
+	panel->backlight.enabled = false;
+	panel->backlight.funcs->disable(old_conn_state, 0);
+
+	mutex_unlock(&dev_priv->display.backlight.lock);
+}
+
+static void lpt_enable_backlight(const struct intel_crtc_state *crtc_state,
+				 const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 pch_ctl1, pch_ctl2, schicken;
+
+	pch_ctl1 = intel_de_read(dev_priv, BLC_PWM_PCH_CTL1);
+	if (pch_ctl1 & BLM_PCH_PWM_ENABLE) {
+		drm_dbg_kms(&dev_priv->drm, "pch backlight already enabled\n");
+		pch_ctl1 &= ~BLM_PCH_PWM_ENABLE;
+		intel_de_write(dev_priv, BLC_PWM_PCH_CTL1, pch_ctl1);
+	}
+
+	if (HAS_PCH_LPT(dev_priv)) {
+		schicken = intel_de_read(dev_priv, SOUTH_CHICKEN2);
+		if (panel->backlight.alternate_pwm_increment)
+			schicken |= LPT_PWM_GRANULARITY;
+		else
+			schicken &= ~LPT_PWM_GRANULARITY;
+		intel_de_write(dev_priv, SOUTH_CHICKEN2, schicken);
+	} else {
+		schicken = intel_de_read(dev_priv, SOUTH_CHICKEN1);
+		if (panel->backlight.alternate_pwm_increment)
+			schicken |= SPT_PWM_GRANULARITY;
+		else
+			schicken &= ~SPT_PWM_GRANULARITY;
+		intel_de_write(dev_priv, SOUTH_CHICKEN1, schicken);
+	}
+
+	pch_ctl2 = panel->backlight.pwm_level_max << 16;
+	intel_de_write(dev_priv, BLC_PWM_PCH_CTL2, pch_ctl2);
+
+	pch_ctl1 = 0;
+	if (panel->backlight.active_low_pwm)
+		pch_ctl1 |= BLM_PCH_POLARITY;
+
+	/* After LPT, override is the default. */
+	if (HAS_PCH_LPT(dev_priv))
+		pch_ctl1 |= BLM_PCH_OVERRIDE_ENABLE;
+
+	intel_de_write(dev_priv, BLC_PWM_PCH_CTL1, pch_ctl1);
+	intel_de_posting_read(dev_priv, BLC_PWM_PCH_CTL1);
+	intel_de_write(dev_priv, BLC_PWM_PCH_CTL1,
+		       pch_ctl1 | BLM_PCH_PWM_ENABLE);
+
+	/* This won't stick until the above enable. */
+	intel_backlight_set_pwm_level(conn_state, level);
+}
+
+static void pch_enable_backlight(const struct intel_crtc_state *crtc_state,
+				 const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 cpu_ctl2, pch_ctl1, pch_ctl2;
+
+	cpu_ctl2 = intel_de_read(dev_priv, BLC_PWM_CPU_CTL2);
+	if (cpu_ctl2 & BLM_PWM_ENABLE) {
+		drm_dbg_kms(&dev_priv->drm, "cpu backlight already enabled\n");
+		cpu_ctl2 &= ~BLM_PWM_ENABLE;
+		intel_de_write(dev_priv, BLC_PWM_CPU_CTL2, cpu_ctl2);
+	}
+
+	pch_ctl1 = intel_de_read(dev_priv, BLC_PWM_PCH_CTL1);
+	if (pch_ctl1 & BLM_PCH_PWM_ENABLE) {
+		drm_dbg_kms(&dev_priv->drm, "pch backlight already enabled\n");
+		pch_ctl1 &= ~BLM_PCH_PWM_ENABLE;
+		intel_de_write(dev_priv, BLC_PWM_PCH_CTL1, pch_ctl1);
+	}
+
+	if (cpu_transcoder == TRANSCODER_EDP)
+		cpu_ctl2 = BLM_TRANSCODER_EDP;
+	else
+		cpu_ctl2 = BLM_PIPE(cpu_transcoder);
+	intel_de_write(dev_priv, BLC_PWM_CPU_CTL2, cpu_ctl2);
+	intel_de_posting_read(dev_priv, BLC_PWM_CPU_CTL2);
+	intel_de_write(dev_priv, BLC_PWM_CPU_CTL2, cpu_ctl2 | BLM_PWM_ENABLE);
+
+	/* This won't stick until the above enable. */
+	intel_backlight_set_pwm_level(conn_state, level);
+
+	pch_ctl2 = panel->backlight.pwm_level_max << 16;
+	intel_de_write(dev_priv, BLC_PWM_PCH_CTL2, pch_ctl2);
+
+	pch_ctl1 = 0;
+	if (panel->backlight.active_low_pwm)
+		pch_ctl1 |= BLM_PCH_POLARITY;
+
+	intel_de_write(dev_priv, BLC_PWM_PCH_CTL1, pch_ctl1);
+	intel_de_posting_read(dev_priv, BLC_PWM_PCH_CTL1);
+	intel_de_write(dev_priv, BLC_PWM_PCH_CTL1,
+		       pch_ctl1 | BLM_PCH_PWM_ENABLE);
+}
+
+static void i9xx_enable_backlight(const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 ctl, freq;
+
+	ctl = intel_de_read(dev_priv, BLC_PWM_CTL);
+	if (ctl & BACKLIGHT_DUTY_CYCLE_MASK_PNV) {
+		drm_dbg_kms(&dev_priv->drm, "backlight already enabled\n");
+		intel_de_write(dev_priv, BLC_PWM_CTL, 0);
+	}
+
+	freq = panel->backlight.pwm_level_max;
+	if (panel->backlight.combination_mode)
+		freq /= 0xff;
+
+	ctl = freq << 17;
+	if (panel->backlight.combination_mode)
+		ctl |= BLM_LEGACY_MODE;
+	if (IS_PINEVIEW(dev_priv) && panel->backlight.active_low_pwm)
+		ctl |= BLM_POLARITY_PNV;
+
+	intel_de_write(dev_priv, BLC_PWM_CTL, ctl);
+	intel_de_posting_read(dev_priv, BLC_PWM_CTL);
+
+	/* XXX: combine this into above write? */
+	intel_backlight_set_pwm_level(conn_state, level);
+
+	/*
+	 * Needed to enable backlight on some 855gm models. BLC_HIST_CTL is
+	 * 855gm only, but checking for gen2 is safe, as 855gm is the only gen2
+	 * that has backlight.
+	 */
+	if (DISPLAY_VER(dev_priv) == 2)
+		intel_de_write(dev_priv, BLC_HIST_CTL, BLM_HISTOGRAM_ENABLE);
+}
+
+static void i965_enable_backlight(const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	enum pipe pipe = to_intel_crtc(conn_state->crtc)->pipe;
+	u32 ctl, ctl2, freq;
+
+	ctl2 = intel_de_read(dev_priv, BLC_PWM_CTL2);
+	if (ctl2 & BLM_PWM_ENABLE) {
+		drm_dbg_kms(&dev_priv->drm, "backlight already enabled\n");
+		ctl2 &= ~BLM_PWM_ENABLE;
+		intel_de_write(dev_priv, BLC_PWM_CTL2, ctl2);
+	}
+
+	freq = panel->backlight.pwm_level_max;
+	if (panel->backlight.combination_mode)
+		freq /= 0xff;
+
+	ctl = freq << 16;
+	intel_de_write(dev_priv, BLC_PWM_CTL, ctl);
+
+	ctl2 = BLM_PIPE(pipe);
+	if (panel->backlight.combination_mode)
+		ctl2 |= BLM_COMBINATION_MODE;
+	if (panel->backlight.active_low_pwm)
+		ctl2 |= BLM_POLARITY_I965;
+	intel_de_write(dev_priv, BLC_PWM_CTL2, ctl2);
+	intel_de_posting_read(dev_priv, BLC_PWM_CTL2);
+	intel_de_write(dev_priv, BLC_PWM_CTL2, ctl2 | BLM_PWM_ENABLE);
+
+	intel_backlight_set_pwm_level(conn_state, level);
+}
+
+static void vlv_enable_backlight(const struct intel_crtc_state *crtc_state,
+				 const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	enum pipe pipe = to_intel_crtc(crtc_state->uapi.crtc)->pipe;
+	u32 ctl, ctl2;
+
+	ctl2 = intel_de_read(dev_priv, VLV_BLC_PWM_CTL2(pipe));
+	if (ctl2 & BLM_PWM_ENABLE) {
+		drm_dbg_kms(&dev_priv->drm, "backlight already enabled\n");
+		ctl2 &= ~BLM_PWM_ENABLE;
+		intel_de_write(dev_priv, VLV_BLC_PWM_CTL2(pipe), ctl2);
+	}
+
+	ctl = panel->backlight.pwm_level_max << 16;
+	intel_de_write(dev_priv, VLV_BLC_PWM_CTL(pipe), ctl);
+
+	/* XXX: combine this into above write? */
+	intel_backlight_set_pwm_level(conn_state, level);
+
+	ctl2 = 0;
+	if (panel->backlight.active_low_pwm)
+		ctl2 |= BLM_POLARITY_I965;
+	intel_de_write(dev_priv, VLV_BLC_PWM_CTL2(pipe), ctl2);
+	intel_de_posting_read(dev_priv, VLV_BLC_PWM_CTL2(pipe));
+	intel_de_write(dev_priv, VLV_BLC_PWM_CTL2(pipe),
+		       ctl2 | BLM_PWM_ENABLE);
+}
+
+static void bxt_enable_backlight(const struct intel_crtc_state *crtc_state,
+				 const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	enum pipe pipe = to_intel_crtc(crtc_state->uapi.crtc)->pipe;
+	u32 pwm_ctl, val;
+
+	/* Controller 1 uses the utility pin. */
+	if (panel->backlight.controller == 1) {
+		val = intel_de_read(dev_priv, UTIL_PIN_CTL);
+		if (val & UTIL_PIN_ENABLE) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "util pin already enabled\n");
+			val &= ~UTIL_PIN_ENABLE;
+			intel_de_write(dev_priv, UTIL_PIN_CTL, val);
+		}
+
+		val = 0;
+		if (panel->backlight.util_pin_active_low)
+			val |= UTIL_PIN_POLARITY;
+		intel_de_write(dev_priv, UTIL_PIN_CTL,
+			       val | UTIL_PIN_PIPE(pipe) | UTIL_PIN_MODE_PWM | UTIL_PIN_ENABLE);
+	}
+
+	pwm_ctl = intel_de_read(dev_priv,
+				BXT_BLC_PWM_CTL(panel->backlight.controller));
+	if (pwm_ctl & BXT_BLC_PWM_ENABLE) {
+		drm_dbg_kms(&dev_priv->drm, "backlight already enabled\n");
+		pwm_ctl &= ~BXT_BLC_PWM_ENABLE;
+		intel_de_write(dev_priv,
+			       BXT_BLC_PWM_CTL(panel->backlight.controller),
+			       pwm_ctl);
+	}
+
+	intel_de_write(dev_priv,
+		       BXT_BLC_PWM_FREQ(panel->backlight.controller),
+		       panel->backlight.pwm_level_max);
+
+	intel_backlight_set_pwm_level(conn_state, level);
+
+	pwm_ctl = 0;
+	if (panel->backlight.active_low_pwm)
+		pwm_ctl |= BXT_BLC_PWM_POLARITY;
+
+	intel_de_write(dev_priv, BXT_BLC_PWM_CTL(panel->backlight.controller),
+		       pwm_ctl);
+	intel_de_posting_read(dev_priv,
+			      BXT_BLC_PWM_CTL(panel->backlight.controller));
+	intel_de_write(dev_priv, BXT_BLC_PWM_CTL(panel->backlight.controller),
+		       pwm_ctl | BXT_BLC_PWM_ENABLE);
+}
+
+static void cnp_enable_backlight(const struct intel_crtc_state *crtc_state,
+				 const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 pwm_ctl;
+
+	pwm_ctl = intel_de_read(dev_priv,
+				BXT_BLC_PWM_CTL(panel->backlight.controller));
+	if (pwm_ctl & BXT_BLC_PWM_ENABLE) {
+		drm_dbg_kms(&dev_priv->drm, "backlight already enabled\n");
+		pwm_ctl &= ~BXT_BLC_PWM_ENABLE;
+		intel_de_write(dev_priv,
+			       BXT_BLC_PWM_CTL(panel->backlight.controller),
+			       pwm_ctl);
+	}
+
+	intel_de_write(dev_priv,
+		       BXT_BLC_PWM_FREQ(panel->backlight.controller),
+		       panel->backlight.pwm_level_max);
+
+	intel_backlight_set_pwm_level(conn_state, level);
+
+	pwm_ctl = 0;
+	if (panel->backlight.active_low_pwm)
+		pwm_ctl |= BXT_BLC_PWM_POLARITY;
+
+	intel_de_write(dev_priv, BXT_BLC_PWM_CTL(panel->backlight.controller),
+		       pwm_ctl);
+	intel_de_posting_read(dev_priv,
+			      BXT_BLC_PWM_CTL(panel->backlight.controller));
+	intel_de_write(dev_priv, BXT_BLC_PWM_CTL(panel->backlight.controller),
+		       pwm_ctl | BXT_BLC_PWM_ENABLE);
+}
+
+static void ext_pwm_enable_backlight(const struct intel_crtc_state *crtc_state,
+				     const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	pwm_set_relative_duty_cycle(&panel->backlight.pwm_state, level, 100);
+	panel->backlight.pwm_state.enabled = true;
+	pwm_apply_state(panel->backlight.pwm, &panel->backlight.pwm_state);
+}
+
+static void __intel_backlight_enable(const struct intel_crtc_state *crtc_state,
+				     const struct drm_connector_state *conn_state)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	WARN_ON(panel->backlight.max == 0);
+
+	if (panel->backlight.level <= panel->backlight.min) {
+		panel->backlight.level = panel->backlight.max;
+		if (panel->backlight.device)
+			panel->backlight.device->props.brightness =
+				scale_hw_to_user(connector,
+						 panel->backlight.level,
+						 panel->backlight.device->props.max_brightness);
+	}
+
+	panel->backlight.funcs->enable(crtc_state, conn_state, panel->backlight.level);
+	panel->backlight.enabled = true;
+	if (panel->backlight.device)
+		panel->backlight.device->props.power = FB_BLANK_UNBLANK;
+}
+
+void intel_backlight_enable(const struct intel_crtc_state *crtc_state,
+			    const struct drm_connector_state *conn_state)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	enum pipe pipe = to_intel_crtc(crtc_state->uapi.crtc)->pipe;
+
+	if (!panel->backlight.present)
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "pipe %c\n", pipe_name(pipe));
+
+	mutex_lock(&dev_priv->display.backlight.lock);
+
+	__intel_backlight_enable(crtc_state, conn_state);
+
+	mutex_unlock(&dev_priv->display.backlight.lock);
+}
+
+#if IS_ENABLED(CONFIG_BACKLIGHT_CLASS_DEVICE)
+static u32 intel_panel_get_backlight(struct intel_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 val = 0;
+
+	mutex_lock(&dev_priv->display.backlight.lock);
+
+	if (panel->backlight.enabled)
+		val = panel->backlight.funcs->get(connector, intel_connector_get_pipe(connector));
+
+	mutex_unlock(&dev_priv->display.backlight.lock);
+
+	drm_dbg_kms(&dev_priv->drm, "get backlight PWM = %d\n", val);
+	return val;
+}
+
+/* Scale user_level in range [0..user_max] to [hw_min..hw_max]. */
+static u32 scale_user_to_hw(struct intel_connector *connector,
+			    u32 user_level, u32 user_max)
+{
+	struct intel_panel *panel = &connector->panel;
+
+	return scale(user_level, 0, user_max,
+		     panel->backlight.min, panel->backlight.max);
+}
+
+/* set backlight brightness to level in range [0..max], scaling wrt hw min */
+static void intel_panel_set_backlight(const struct drm_connector_state *conn_state,
+				      u32 user_level, u32 user_max)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 hw_level;
+
+	if (!panel->backlight.present)
+		return;
+
+	mutex_lock(&dev_priv->display.backlight.lock);
+
+	drm_WARN_ON(&dev_priv->drm, panel->backlight.max == 0);
+
+	hw_level = scale_user_to_hw(connector, user_level, user_max);
+	panel->backlight.level = hw_level;
+
+	if (panel->backlight.enabled)
+		intel_panel_actually_set_backlight(conn_state, hw_level);
+
+	mutex_unlock(&dev_priv->display.backlight.lock);
+}
+
+static int intel_backlight_device_update_status(struct backlight_device *bd)
+{
+	struct intel_connector *connector = bl_get_data(bd);
+	struct intel_panel *panel = &connector->panel;
+	struct drm_device *dev = connector->base.dev;
+
+	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+	DRM_DEBUG_KMS("updating intel_backlight, brightness=%d/%d\n",
+		      bd->props.brightness, bd->props.max_brightness);
+	intel_panel_set_backlight(connector->base.state, bd->props.brightness,
+				  bd->props.max_brightness);
+
+	/*
+	 * Allow flipping bl_power as a sub-state of enabled. Sadly the
+	 * backlight class device does not make it easy to differentiate
+	 * between callbacks for brightness and bl_power, so our backlight_power
+	 * callback needs to take this into account.
+	 */
+	if (panel->backlight.enabled) {
+		if (panel->backlight.power) {
+			bool enable = bd->props.power == FB_BLANK_UNBLANK &&
+				bd->props.brightness != 0;
+			panel->backlight.power(connector, enable);
+		}
+	} else {
+		bd->props.power = FB_BLANK_POWERDOWN;
+	}
+
+	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+	return 0;
+}
+
+static int intel_backlight_device_get_brightness(struct backlight_device *bd)
+{
+	struct intel_connector *connector = bl_get_data(bd);
+	struct drm_device *dev = connector->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	intel_wakeref_t wakeref;
+	int ret = 0;
+
+	with_intel_runtime_pm(&dev_priv->runtime_pm, wakeref) {
+		u32 hw_level;
+
+		drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+
+		hw_level = intel_panel_get_backlight(connector);
+		ret = scale_hw_to_user(connector,
+				       hw_level, bd->props.max_brightness);
+
+		drm_modeset_unlock(&dev->mode_config.connection_mutex);
+	}
+
+	return ret;
+}
+
+static const struct backlight_ops intel_backlight_device_ops = {
+	.update_status = intel_backlight_device_update_status,
+	.get_brightness = intel_backlight_device_get_brightness,
+};
+
+int intel_backlight_device_register(struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	struct backlight_properties props;
+	struct backlight_device *bd;
+	const char *name;
+	int ret = 0;
+
+	if (WARN_ON(panel->backlight.device))
+		return -ENODEV;
+
+	if (!panel->backlight.present)
+		return 0;
+
+	WARN_ON(panel->backlight.max == 0);
+
+	if (!acpi_video_backlight_use_native()) {
+		drm_info(&i915->drm, "Skipping intel_backlight registration\n");
+		return 0;
+	}
+
+	memset(&props, 0, sizeof(props));
+	props.type = BACKLIGHT_RAW;
+
+	/*
+	 * Note: Everything should work even if the backlight device max
+	 * presented to the userspace is arbitrarily chosen.
+	 */
+	props.max_brightness = panel->backlight.max;
+	props.brightness = scale_hw_to_user(connector,
+					    panel->backlight.level,
+					    props.max_brightness);
+
+	if (panel->backlight.enabled)
+		props.power = FB_BLANK_UNBLANK;
+	else
+		props.power = FB_BLANK_POWERDOWN;
+
+	name = kstrdup("intel_backlight", GFP_KERNEL);
+	if (!name)
+		return -ENOMEM;
+
+	bd = backlight_device_get_by_name(name);
+	if (bd) {
+		put_device(&bd->dev);
+		/*
+		 * Using the same name independent of the drm device or connector
+		 * prevents registration of multiple backlight devices in the
+		 * driver. However, we need to use the default name for backward
+		 * compatibility. Use unique names for subsequent backlight devices as a
+		 * fallback when the default name already exists.
+		 */
+		kfree(name);
+		name = kasprintf(GFP_KERNEL, "card%d-%s-backlight",
+				 i915->drm.primary->index, connector->base.name);
+		if (!name)
+			return -ENOMEM;
+	}
+	bd = backlight_device_register(name, connector->base.kdev, connector,
+				       &intel_backlight_device_ops, &props);
+
+	if (IS_ERR(bd)) {
+		drm_err(&i915->drm,
+			"[CONNECTOR:%d:%s] backlight device %s register failed: %ld\n",
+			connector->base.base.id, connector->base.name, name, PTR_ERR(bd));
+		ret = PTR_ERR(bd);
+		goto out;
+	}
+
+	panel->backlight.device = bd;
+
+	drm_dbg_kms(&i915->drm,
+		    "[CONNECTOR:%d:%s] backlight device %s registered\n",
+		    connector->base.base.id, connector->base.name, name);
+
+out:
+	kfree(name);
+
+	return ret;
+}
+
+void intel_backlight_device_unregister(struct intel_connector *connector)
+{
+	struct intel_panel *panel = &connector->panel;
+
+	if (panel->backlight.device) {
+		backlight_device_unregister(panel->backlight.device);
+		panel->backlight.device = NULL;
+	}
+}
+#endif /* CONFIG_BACKLIGHT_CLASS_DEVICE */
+
+/*
+ * CNP: PWM clock frequency is 19.2 MHz or 24 MHz.
+ *      PWM increment = 1
+ */
+static u32 cnp_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+
+	return DIV_ROUND_CLOSEST(KHz(RUNTIME_INFO(dev_priv)->rawclk_freq),
+				 pwm_freq_hz);
+}
+
+/*
+ * BXT: PWM clock frequency = 19.2 MHz.
+ */
+static u32 bxt_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
+{
+	return DIV_ROUND_CLOSEST(KHz(19200), pwm_freq_hz);
+}
+
+/*
+ * SPT: This value represents the period of the PWM stream in clock periods
+ * multiplied by 16 (default increment) or 128 (alternate increment selected in
+ * SCHICKEN_1 bit 0). PWM clock is 24 MHz.
+ */
+static u32 spt_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
+{
+	struct intel_panel *panel = &connector->panel;
+	u32 mul;
+
+	if (panel->backlight.alternate_pwm_increment)
+		mul = 128;
+	else
+		mul = 16;
+
+	return DIV_ROUND_CLOSEST(MHz(24), pwm_freq_hz * mul);
+}
+
+/*
+ * LPT: This value represents the period of the PWM stream in clock periods
+ * multiplied by 128 (default increment) or 16 (alternate increment, selected in
+ * LPT SOUTH_CHICKEN2 register bit 5).
+ */
+static u32 lpt_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 mul, clock;
+
+	if (panel->backlight.alternate_pwm_increment)
+		mul = 16;
+	else
+		mul = 128;
+
+	if (HAS_PCH_LPT_H(dev_priv))
+		clock = MHz(135); /* LPT:H */
+	else
+		clock = MHz(24); /* LPT:LP */
+
+	return DIV_ROUND_CLOSEST(clock, pwm_freq_hz * mul);
+}
+
+/*
+ * ILK/SNB/IVB: This value represents the period of the PWM stream in PCH
+ * display raw clocks multiplied by 128.
+ */
+static u32 pch_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+
+	return DIV_ROUND_CLOSEST(KHz(RUNTIME_INFO(dev_priv)->rawclk_freq),
+				 pwm_freq_hz * 128);
+}
+
+/*
+ * Gen2: This field determines the number of time base events (display core
+ * clock frequency/32) in total for a complete cycle of modulated backlight
+ * control.
+ *
+ * Gen3: A time base event equals the display core clock ([DevPNV] HRAW clock)
+ * divided by 32.
+ */
+static u32 i9xx_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	int clock;
+
+	if (IS_PINEVIEW(dev_priv))
+		clock = KHz(RUNTIME_INFO(dev_priv)->rawclk_freq);
+	else
+		clock = KHz(dev_priv->display.cdclk.hw.cdclk);
+
+	return DIV_ROUND_CLOSEST(clock, pwm_freq_hz * 32);
+}
+
+/*
+ * Gen4: This value represents the period of the PWM stream in display core
+ * clocks ([DevCTG] HRAW clocks) multiplied by 128.
+ *
+ */
+static u32 i965_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	int clock;
+
+	if (IS_G4X(dev_priv))
+		clock = KHz(RUNTIME_INFO(dev_priv)->rawclk_freq);
+	else
+		clock = KHz(dev_priv->display.cdclk.hw.cdclk);
+
+	return DIV_ROUND_CLOSEST(clock, pwm_freq_hz * 128);
+}
+
+/*
+ * VLV: This value represents the period of the PWM stream in display core
+ * clocks ([DevCTG] 200MHz HRAW clocks) multiplied by 128 or 25MHz S0IX clocks
+ * multiplied by 16. CHV uses a 19.2MHz S0IX clock.
+ */
+static u32 vlv_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	int mul, clock;
+
+	if ((intel_de_read(dev_priv, CBR1_VLV) & CBR_PWM_CLOCK_MUX_SELECT) == 0) {
+		if (IS_CHERRYVIEW(dev_priv))
+			clock = KHz(19200);
+		else
+			clock = MHz(25);
+		mul = 16;
+	} else {
+		clock = KHz(RUNTIME_INFO(dev_priv)->rawclk_freq);
+		mul = 128;
+	}
+
+	return DIV_ROUND_CLOSEST(clock, pwm_freq_hz * mul);
+}
+
+static u16 get_vbt_pwm_freq(struct intel_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	u16 pwm_freq_hz = connector->panel.vbt.backlight.pwm_freq_hz;
+
+	if (pwm_freq_hz) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "VBT defined backlight frequency %u Hz\n",
+			    pwm_freq_hz);
+	} else {
+		pwm_freq_hz = 200;
+		drm_dbg_kms(&dev_priv->drm,
+			    "default backlight frequency %u Hz\n",
+			    pwm_freq_hz);
+	}
+
+	return pwm_freq_hz;
+}
+
+static u32 get_backlight_max_vbt(struct intel_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u16 pwm_freq_hz = get_vbt_pwm_freq(connector);
+	u32 pwm;
+
+	if (!panel->backlight.pwm_funcs->hz_to_pwm) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "backlight frequency conversion not supported\n");
+		return 0;
+	}
+
+	pwm = panel->backlight.pwm_funcs->hz_to_pwm(connector, pwm_freq_hz);
+	if (!pwm) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "backlight frequency conversion failed\n");
+		return 0;
+	}
+
+	return pwm;
+}
+
+/*
+ * Note: The setup hooks can't assume pipe is set!
+ */
+static u32 get_backlight_min_vbt(struct intel_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	int min;
+
+	drm_WARN_ON(&dev_priv->drm, panel->backlight.pwm_level_max == 0);
+
+	/*
+	 * XXX: If the vbt value is 255, it makes min equal to max, which leads
+	 * to problems. There are such machines out there. Either our
+	 * interpretation is wrong or the vbt has bogus data. Or both. Safeguard
+	 * against this by letting the minimum be at most (arbitrarily chosen)
+	 * 25% of the max.
+	 */
+	min = clamp_t(int, connector->panel.vbt.backlight.min_brightness, 0, 64);
+	if (min != connector->panel.vbt.backlight.min_brightness) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "clamping VBT min backlight %d/255 to %d/255\n",
+			    connector->panel.vbt.backlight.min_brightness, min);
+	}
+
+	/* vbt value is a coefficient in range [0..255] */
+	return scale(min, 0, 255, 0, panel->backlight.pwm_level_max);
+}
+
+static int lpt_setup_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 cpu_ctl2, pch_ctl1, pch_ctl2, val;
+	bool alt, cpu_mode;
+
+	if (HAS_PCH_LPT(dev_priv))
+		alt = intel_de_read(dev_priv, SOUTH_CHICKEN2) & LPT_PWM_GRANULARITY;
+	else
+		alt = intel_de_read(dev_priv, SOUTH_CHICKEN1) & SPT_PWM_GRANULARITY;
+	panel->backlight.alternate_pwm_increment = alt;
+
+	pch_ctl1 = intel_de_read(dev_priv, BLC_PWM_PCH_CTL1);
+	panel->backlight.active_low_pwm = pch_ctl1 & BLM_PCH_POLARITY;
+
+	pch_ctl2 = intel_de_read(dev_priv, BLC_PWM_PCH_CTL2);
+	panel->backlight.pwm_level_max = pch_ctl2 >> 16;
+
+	cpu_ctl2 = intel_de_read(dev_priv, BLC_PWM_CPU_CTL2);
+
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
+
+	if (!panel->backlight.pwm_level_max)
+		return -ENODEV;
+
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
+
+	panel->backlight.pwm_enabled = pch_ctl1 & BLM_PCH_PWM_ENABLE;
+
+	cpu_mode = panel->backlight.pwm_enabled && HAS_PCH_LPT(dev_priv) &&
+		   !(pch_ctl1 & BLM_PCH_OVERRIDE_ENABLE) &&
+		   (cpu_ctl2 & BLM_PWM_ENABLE);
+
+	if (cpu_mode) {
+		val = pch_get_backlight(connector, unused);
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "CPU backlight register was enabled, switching to PCH override\n");
+
+		/* Write converted CPU PWM value to PCH override register */
+		lpt_set_backlight(connector->base.state, val);
+		intel_de_write(dev_priv, BLC_PWM_PCH_CTL1,
+			       pch_ctl1 | BLM_PCH_OVERRIDE_ENABLE);
+
+		intel_de_write(dev_priv, BLC_PWM_CPU_CTL2,
+			       cpu_ctl2 & ~BLM_PWM_ENABLE);
+	}
+
+	return 0;
+}
+
+static int pch_setup_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 cpu_ctl2, pch_ctl1, pch_ctl2;
+
+	pch_ctl1 = intel_de_read(dev_priv, BLC_PWM_PCH_CTL1);
+	panel->backlight.active_low_pwm = pch_ctl1 & BLM_PCH_POLARITY;
+
+	pch_ctl2 = intel_de_read(dev_priv, BLC_PWM_PCH_CTL2);
+	panel->backlight.pwm_level_max = pch_ctl2 >> 16;
+
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
+
+	if (!panel->backlight.pwm_level_max)
+		return -ENODEV;
+
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
+
+	cpu_ctl2 = intel_de_read(dev_priv, BLC_PWM_CPU_CTL2);
+	panel->backlight.pwm_enabled = (cpu_ctl2 & BLM_PWM_ENABLE) &&
+		(pch_ctl1 & BLM_PCH_PWM_ENABLE);
+
+	return 0;
+}
+
+static int i9xx_setup_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 ctl, val;
+
+	ctl = intel_de_read(dev_priv, BLC_PWM_CTL);
+
+	if (DISPLAY_VER(dev_priv) == 2 || IS_I915GM(dev_priv) || IS_I945GM(dev_priv))
+		panel->backlight.combination_mode = ctl & BLM_LEGACY_MODE;
+
+	if (IS_PINEVIEW(dev_priv))
+		panel->backlight.active_low_pwm = ctl & BLM_POLARITY_PNV;
+
+	panel->backlight.pwm_level_max = ctl >> 17;
+
+	if (!panel->backlight.pwm_level_max) {
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
+		panel->backlight.pwm_level_max >>= 1;
+	}
+
+	if (!panel->backlight.pwm_level_max)
+		return -ENODEV;
+
+	if (panel->backlight.combination_mode)
+		panel->backlight.pwm_level_max *= 0xff;
+
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
+
+	val = i9xx_get_backlight(connector, unused);
+	val = intel_backlight_invert_pwm_level(connector, val);
+	val = clamp(val, panel->backlight.pwm_level_min, panel->backlight.pwm_level_max);
+
+	panel->backlight.pwm_enabled = val != 0;
+
+	return 0;
+}
+
+static int i965_setup_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 ctl, ctl2;
+
+	ctl2 = intel_de_read(dev_priv, BLC_PWM_CTL2);
+	panel->backlight.combination_mode = ctl2 & BLM_COMBINATION_MODE;
+	panel->backlight.active_low_pwm = ctl2 & BLM_POLARITY_I965;
+
+	ctl = intel_de_read(dev_priv, BLC_PWM_CTL);
+	panel->backlight.pwm_level_max = ctl >> 16;
+
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
+
+	if (!panel->backlight.pwm_level_max)
+		return -ENODEV;
+
+	if (panel->backlight.combination_mode)
+		panel->backlight.pwm_level_max *= 0xff;
+
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
+
+	panel->backlight.pwm_enabled = ctl2 & BLM_PWM_ENABLE;
+
+	return 0;
+}
+
+static int vlv_setup_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 ctl, ctl2;
+
+	if (drm_WARN_ON(&dev_priv->drm, pipe != PIPE_A && pipe != PIPE_B))
+		return -ENODEV;
+
+	ctl2 = intel_de_read(dev_priv, VLV_BLC_PWM_CTL2(pipe));
+	panel->backlight.active_low_pwm = ctl2 & BLM_POLARITY_I965;
+
+	ctl = intel_de_read(dev_priv, VLV_BLC_PWM_CTL(pipe));
+	panel->backlight.pwm_level_max = ctl >> 16;
+
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
+
+	if (!panel->backlight.pwm_level_max)
+		return -ENODEV;
+
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
+
+	panel->backlight.pwm_enabled = ctl2 & BLM_PWM_ENABLE;
+
+	return 0;
+}
+
+static int
+bxt_setup_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 pwm_ctl, val;
+
+	panel->backlight.controller = connector->panel.vbt.backlight.controller;
+
+	pwm_ctl = intel_de_read(dev_priv,
+				BXT_BLC_PWM_CTL(panel->backlight.controller));
+
+	/* Controller 1 uses the utility pin. */
+	if (panel->backlight.controller == 1) {
+		val = intel_de_read(dev_priv, UTIL_PIN_CTL);
+		panel->backlight.util_pin_active_low =
+					val & UTIL_PIN_POLARITY;
+	}
+
+	panel->backlight.active_low_pwm = pwm_ctl & BXT_BLC_PWM_POLARITY;
+	panel->backlight.pwm_level_max =
+		intel_de_read(dev_priv, BXT_BLC_PWM_FREQ(panel->backlight.controller));
+
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
+
+	if (!panel->backlight.pwm_level_max)
+		return -ENODEV;
+
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
+
+	panel->backlight.pwm_enabled = pwm_ctl & BXT_BLC_PWM_ENABLE;
+
+	return 0;
+}
+
+static int
+cnp_setup_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	u32 pwm_ctl;
+
+	/*
+	 * CNP has the BXT implementation of backlight, but with only one
+	 * controller. TODO: ICP has multiple controllers but we only use
+	 * controller 0 for now.
+	 */
+	panel->backlight.controller = 0;
+
+	pwm_ctl = intel_de_read(dev_priv,
+				BXT_BLC_PWM_CTL(panel->backlight.controller));
+
+	panel->backlight.active_low_pwm = pwm_ctl & BXT_BLC_PWM_POLARITY;
+	panel->backlight.pwm_level_max =
+		intel_de_read(dev_priv, BXT_BLC_PWM_FREQ(panel->backlight.controller));
+
+	if (!panel->backlight.pwm_level_max)
+		panel->backlight.pwm_level_max = get_backlight_max_vbt(connector);
+
+	if (!panel->backlight.pwm_level_max)
+		return -ENODEV;
+
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
+
+	panel->backlight.pwm_enabled = pwm_ctl & BXT_BLC_PWM_ENABLE;
+
+	return 0;
+}
+
+static int ext_pwm_setup_backlight(struct intel_connector *connector,
+				   enum pipe pipe)
+{
+	struct drm_device *dev = connector->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_panel *panel = &connector->panel;
+	const char *desc;
+	u32 level;
+
+	/* Get the right PWM chip for DSI backlight according to VBT */
+	if (connector->panel.vbt.dsi.config->pwm_blc == PPS_BLC_PMIC) {
+		panel->backlight.pwm = pwm_get(dev->dev, "pwm_pmic_backlight");
+		desc = "PMIC";
+	} else {
+		panel->backlight.pwm = pwm_get(dev->dev, "pwm_soc_backlight");
+		desc = "SoC";
+	}
+
+	if (IS_ERR(panel->backlight.pwm)) {
+		drm_err(&dev_priv->drm, "Failed to get the %s PWM chip\n",
+			desc);
+		panel->backlight.pwm = NULL;
+		return -ENODEV;
+	}
+
+	panel->backlight.pwm_level_max = 100; /* 100% */
+	panel->backlight.pwm_level_min = get_backlight_min_vbt(connector);
+
+	if (pwm_is_enabled(panel->backlight.pwm)) {
+		/* PWM is already enabled, use existing settings */
+		pwm_get_state(panel->backlight.pwm, &panel->backlight.pwm_state);
+
+		level = pwm_get_relative_duty_cycle(&panel->backlight.pwm_state,
+						    100);
+		level = intel_backlight_invert_pwm_level(connector, level);
+		panel->backlight.pwm_enabled = true;
+
+		drm_dbg_kms(&dev_priv->drm, "PWM already enabled at freq %ld, VBT freq %d, level %d\n",
+			    NSEC_PER_SEC / (unsigned long)panel->backlight.pwm_state.period,
+			    get_vbt_pwm_freq(connector), level);
+	} else {
+		/* Set period from VBT frequency, leave other settings at 0. */
+		panel->backlight.pwm_state.period =
+			NSEC_PER_SEC / get_vbt_pwm_freq(connector);
+	}
+
+	drm_info(&dev_priv->drm, "Using %s PWM for LCD backlight control\n",
+		 desc);
+	return 0;
+}
+
+static void intel_pwm_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	panel->backlight.pwm_funcs->set(conn_state,
+					intel_backlight_invert_pwm_level(connector, level));
+}
+
+static u32 intel_pwm_get_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct intel_panel *panel = &connector->panel;
+
+	return intel_backlight_invert_pwm_level(connector,
+					    panel->backlight.pwm_funcs->get(connector, pipe));
+}
+
+static void intel_pwm_enable_backlight(const struct intel_crtc_state *crtc_state,
+				       const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	panel->backlight.pwm_funcs->enable(crtc_state, conn_state,
+					   intel_backlight_invert_pwm_level(connector, level));
+}
+
+static void intel_pwm_disable_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	panel->backlight.pwm_funcs->disable(conn_state,
+					    intel_backlight_invert_pwm_level(connector, level));
+}
+
+static int intel_pwm_setup_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct intel_panel *panel = &connector->panel;
+	int ret = panel->backlight.pwm_funcs->setup(connector, pipe);
+
+	if (ret < 0)
+		return ret;
+
+	panel->backlight.min = panel->backlight.pwm_level_min;
+	panel->backlight.max = panel->backlight.pwm_level_max;
+	panel->backlight.level = intel_pwm_get_backlight(connector, pipe);
+	panel->backlight.enabled = panel->backlight.pwm_enabled;
+
+	return 0;
+}
+
+void intel_backlight_update(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    const struct drm_connector_state *conn_state)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+
+	if (!panel->backlight.present)
+		return;
+
+	mutex_lock(&dev_priv->display.backlight.lock);
+	if (!panel->backlight.enabled)
+		__intel_backlight_enable(crtc_state, conn_state);
+
+	mutex_unlock(&dev_priv->display.backlight.lock);
+}
+
+int intel_backlight_setup(struct intel_connector *connector, enum pipe pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	int ret;
+
+	if (!connector->panel.vbt.backlight.present) {
+		if (intel_has_quirk(dev_priv, QUIRK_BACKLIGHT_PRESENT)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "no backlight present per VBT, but present per quirk\n");
+		} else {
+			drm_dbg_kms(&dev_priv->drm,
+				    "no backlight present per VBT\n");
+			return 0;
+		}
+	}
+
+	/* ensure intel_panel has been initialized first */
+	if (drm_WARN_ON(&dev_priv->drm, !panel->backlight.funcs))
+		return -ENODEV;
+
+	/* set level and max in panel struct */
+	mutex_lock(&dev_priv->display.backlight.lock);
+	ret = panel->backlight.funcs->setup(connector, pipe);
+	mutex_unlock(&dev_priv->display.backlight.lock);
+
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "failed to setup backlight for connector %s\n",
+			    connector->base.name);
+		return ret;
+	}
+
+	panel->backlight.present = true;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Connector %s backlight initialized, %s, brightness %u/%u\n",
+		    connector->base.name,
+		    str_enabled_disabled(panel->backlight.enabled),
+		    panel->backlight.level, panel->backlight.max);
+
+	return 0;
+}
+
+void intel_backlight_destroy(struct intel_panel *panel)
+{
+	/* dispose of the pwm */
+	if (panel->backlight.pwm)
+		pwm_put(panel->backlight.pwm);
+
+	panel->backlight.present = false;
+}
+
+static const struct intel_panel_bl_funcs bxt_pwm_funcs = {
+	.setup = bxt_setup_backlight,
+	.enable = bxt_enable_backlight,
+	.disable = bxt_disable_backlight,
+	.set = bxt_set_backlight,
+	.get = bxt_get_backlight,
+	.hz_to_pwm = bxt_hz_to_pwm,
+};
+
+static const struct intel_panel_bl_funcs cnp_pwm_funcs = {
+	.setup = cnp_setup_backlight,
+	.enable = cnp_enable_backlight,
+	.disable = cnp_disable_backlight,
+	.set = bxt_set_backlight,
+	.get = bxt_get_backlight,
+	.hz_to_pwm = cnp_hz_to_pwm,
+};
+
+static const struct intel_panel_bl_funcs lpt_pwm_funcs = {
+	.setup = lpt_setup_backlight,
+	.enable = lpt_enable_backlight,
+	.disable = lpt_disable_backlight,
+	.set = lpt_set_backlight,
+	.get = lpt_get_backlight,
+	.hz_to_pwm = lpt_hz_to_pwm,
+};
+
+static const struct intel_panel_bl_funcs spt_pwm_funcs = {
+	.setup = lpt_setup_backlight,
+	.enable = lpt_enable_backlight,
+	.disable = lpt_disable_backlight,
+	.set = lpt_set_backlight,
+	.get = lpt_get_backlight,
+	.hz_to_pwm = spt_hz_to_pwm,
+};
+
+static const struct intel_panel_bl_funcs pch_pwm_funcs = {
+	.setup = pch_setup_backlight,
+	.enable = pch_enable_backlight,
+	.disable = pch_disable_backlight,
+	.set = pch_set_backlight,
+	.get = pch_get_backlight,
+	.hz_to_pwm = pch_hz_to_pwm,
+};
+
+static const struct intel_panel_bl_funcs ext_pwm_funcs = {
+	.setup = ext_pwm_setup_backlight,
+	.enable = ext_pwm_enable_backlight,
+	.disable = ext_pwm_disable_backlight,
+	.set = ext_pwm_set_backlight,
+	.get = ext_pwm_get_backlight,
+};
+
+static const struct intel_panel_bl_funcs vlv_pwm_funcs = {
+	.setup = vlv_setup_backlight,
+	.enable = vlv_enable_backlight,
+	.disable = vlv_disable_backlight,
+	.set = vlv_set_backlight,
+	.get = vlv_get_backlight,
+	.hz_to_pwm = vlv_hz_to_pwm,
+};
+
+static const struct intel_panel_bl_funcs i965_pwm_funcs = {
+	.setup = i965_setup_backlight,
+	.enable = i965_enable_backlight,
+	.disable = i965_disable_backlight,
+	.set = i9xx_set_backlight,
+	.get = i9xx_get_backlight,
+	.hz_to_pwm = i965_hz_to_pwm,
+};
+
+static const struct intel_panel_bl_funcs i9xx_pwm_funcs = {
+	.setup = i9xx_setup_backlight,
+	.enable = i9xx_enable_backlight,
+	.disable = i9xx_disable_backlight,
+	.set = i9xx_set_backlight,
+	.get = i9xx_get_backlight,
+	.hz_to_pwm = i9xx_hz_to_pwm,
+};
+
+static const struct intel_panel_bl_funcs pwm_bl_funcs = {
+	.setup = intel_pwm_setup_backlight,
+	.enable = intel_pwm_enable_backlight,
+	.disable = intel_pwm_disable_backlight,
+	.set = intel_pwm_set_backlight,
+	.get = intel_pwm_get_backlight,
+};
+
+/* Set up chip specific backlight functions */
+void intel_backlight_init_funcs(struct intel_panel *panel)
+{
+	struct intel_connector *connector =
+		container_of(panel, struct intel_connector, panel);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+
+	if (connector->base.connector_type == DRM_MODE_CONNECTOR_DSI &&
+	    intel_dsi_dcs_init_backlight_funcs(connector) == 0)
+		return;
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		panel->backlight.pwm_funcs = &bxt_pwm_funcs;
+	} else if (INTEL_PCH_TYPE(dev_priv) >= PCH_CNP) {
+		panel->backlight.pwm_funcs = &cnp_pwm_funcs;
+	} else if (INTEL_PCH_TYPE(dev_priv) >= PCH_LPT) {
+		if (HAS_PCH_LPT(dev_priv))
+			panel->backlight.pwm_funcs = &lpt_pwm_funcs;
+		else
+			panel->backlight.pwm_funcs = &spt_pwm_funcs;
+	} else if (HAS_PCH_SPLIT(dev_priv)) {
+		panel->backlight.pwm_funcs = &pch_pwm_funcs;
+	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		if (connector->base.connector_type == DRM_MODE_CONNECTOR_DSI) {
+			panel->backlight.pwm_funcs = &ext_pwm_funcs;
+		} else {
+			panel->backlight.pwm_funcs = &vlv_pwm_funcs;
+		}
+	} else if (DISPLAY_VER(dev_priv) == 4) {
+		panel->backlight.pwm_funcs = &i965_pwm_funcs;
+	} else {
+		panel->backlight.pwm_funcs = &i9xx_pwm_funcs;
+	}
+
+	if (connector->base.connector_type == DRM_MODE_CONNECTOR_eDP) {
+		if (intel_dp_aux_init_backlight_funcs(connector) == 0)
+			return;
+
+		if (!intel_has_quirk(dev_priv, QUIRK_NO_PPS_BACKLIGHT_POWER_HOOK))
+			connector->panel.backlight.power = intel_pps_backlight_power;
+	}
+
+	/* We're using a standard PWM backlight interface */
+	panel->backlight.funcs = &pwm_bl_funcs;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_backlight.h b/drivers/gpu/drm/i915/display/intel_backlight.h
new file mode 100644
index 000000000..339643f63
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_backlight.h
@@ -0,0 +1,52 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __INTEL_BACKLIGHT_H__
+#define __INTEL_BACKLIGHT_H__
+
+#include <linux/types.h>
+
+struct drm_connector_state;
+struct intel_atomic_state;
+struct intel_connector;
+struct intel_crtc_state;
+struct intel_encoder;
+struct intel_panel;
+enum pipe;
+
+void intel_backlight_init_funcs(struct intel_panel *panel);
+int intel_backlight_setup(struct intel_connector *connector, enum pipe pipe);
+void intel_backlight_destroy(struct intel_panel *panel);
+
+void intel_backlight_enable(const struct intel_crtc_state *crtc_state,
+			    const struct drm_connector_state *conn_state);
+void intel_backlight_update(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    const struct drm_connector_state *conn_state);
+void intel_backlight_disable(const struct drm_connector_state *old_conn_state);
+
+void intel_backlight_set_acpi(const struct drm_connector_state *conn_state,
+			      u32 level, u32 max);
+void intel_backlight_set_pwm_level(const struct drm_connector_state *conn_state,
+				   u32 level);
+u32 intel_backlight_invert_pwm_level(struct intel_connector *connector, u32 level);
+u32 intel_backlight_level_to_pwm(struct intel_connector *connector, u32 level);
+u32 intel_backlight_level_from_pwm(struct intel_connector *connector, u32 val);
+
+#if IS_ENABLED(CONFIG_BACKLIGHT_CLASS_DEVICE)
+int intel_backlight_device_register(struct intel_connector *connector);
+void intel_backlight_device_unregister(struct intel_connector *connector);
+#else /* CONFIG_BACKLIGHT_CLASS_DEVICE */
+static inline int intel_backlight_device_register(struct intel_connector *connector)
+{
+	return 0;
+}
+static inline void intel_backlight_device_unregister(struct intel_connector *connector)
+{
+}
+#endif /* CONFIG_BACKLIGHT_CLASS_DEVICE */
+
+#endif /* __INTEL_BACKLIGHT_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_backlight_regs.h b/drivers/gpu/drm/i915/display/intel_backlight_regs.h
new file mode 100644
index 000000000..50c1210f6
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_backlight_regs.h
@@ -0,0 +1,124 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_BACKLIGHT_REGS_H__
+#define __INTEL_BACKLIGHT_REGS_H__
+
+#include "i915_reg_defs.h"
+
+#define _VLV_BLC_PWM_CTL2_A (DISPLAY_MMIO_BASE(dev_priv) + 0x61250)
+#define _VLV_BLC_PWM_CTL2_B (DISPLAY_MMIO_BASE(dev_priv) + 0x61350)
+#define VLV_BLC_PWM_CTL2(pipe) _MMIO_PIPE(pipe, _VLV_BLC_PWM_CTL2_A, \
+					 _VLV_BLC_PWM_CTL2_B)
+
+#define _VLV_BLC_PWM_CTL_A (DISPLAY_MMIO_BASE(dev_priv) + 0x61254)
+#define _VLV_BLC_PWM_CTL_B (DISPLAY_MMIO_BASE(dev_priv) + 0x61354)
+#define VLV_BLC_PWM_CTL(pipe) _MMIO_PIPE(pipe, _VLV_BLC_PWM_CTL_A, \
+					_VLV_BLC_PWM_CTL_B)
+
+#define _VLV_BLC_HIST_CTL_A (DISPLAY_MMIO_BASE(dev_priv) + 0x61260)
+#define _VLV_BLC_HIST_CTL_B (DISPLAY_MMIO_BASE(dev_priv) + 0x61360)
+#define VLV_BLC_HIST_CTL(pipe) _MMIO_PIPE(pipe, _VLV_BLC_HIST_CTL_A, \
+					 _VLV_BLC_HIST_CTL_B)
+
+/* Backlight control */
+#define BLC_PWM_CTL2	_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x61250) /* 965+ only */
+#define   BLM_PWM_ENABLE		(1 << 31)
+#define   BLM_COMBINATION_MODE		(1 << 30) /* gen4 only */
+#define   BLM_PIPE_SELECT		(1 << 29)
+#define   BLM_PIPE_SELECT_IVB		(3 << 29)
+#define   BLM_PIPE_A			(0 << 29)
+#define   BLM_PIPE_B			(1 << 29)
+#define   BLM_PIPE_C			(2 << 29) /* ivb + */
+#define   BLM_TRANSCODER_A		BLM_PIPE_A /* hsw */
+#define   BLM_TRANSCODER_B		BLM_PIPE_B
+#define   BLM_TRANSCODER_C		BLM_PIPE_C
+#define   BLM_TRANSCODER_EDP		(3 << 29)
+#define   BLM_PIPE(pipe)		((pipe) << 29)
+#define   BLM_POLARITY_I965		(1 << 28) /* gen4 only */
+#define   BLM_PHASE_IN_INTERUPT_STATUS	(1 << 26)
+#define   BLM_PHASE_IN_ENABLE		(1 << 25)
+#define   BLM_PHASE_IN_INTERUPT_ENABL	(1 << 24)
+#define   BLM_PHASE_IN_TIME_BASE_SHIFT	(16)
+#define   BLM_PHASE_IN_TIME_BASE_MASK	(0xff << 16)
+#define   BLM_PHASE_IN_COUNT_SHIFT	(8)
+#define   BLM_PHASE_IN_COUNT_MASK	(0xff << 8)
+#define   BLM_PHASE_IN_INCR_SHIFT	(0)
+#define   BLM_PHASE_IN_INCR_MASK	(0xff << 0)
+#define BLC_PWM_CTL	_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x61254)
+/*
+ * This is the most significant 15 bits of the number of backlight cycles in a
+ * complete cycle of the modulated backlight control.
+ *
+ * The actual value is this field multiplied by two.
+ */
+#define   BACKLIGHT_MODULATION_FREQ_SHIFT	(17)
+#define   BACKLIGHT_MODULATION_FREQ_MASK	(0x7fff << 17)
+#define   BLM_LEGACY_MODE			(1 << 16) /* gen2 only */
+/*
+ * This is the number of cycles out of the backlight modulation cycle for which
+ * the backlight is on.
+ *
+ * This field must be no greater than the number of cycles in the complete
+ * backlight modulation cycle.
+ */
+#define   BACKLIGHT_DUTY_CYCLE_SHIFT		(0)
+#define   BACKLIGHT_DUTY_CYCLE_MASK		(0xffff)
+#define   BACKLIGHT_DUTY_CYCLE_MASK_PNV		(0xfffe)
+#define   BLM_POLARITY_PNV			(1 << 0) /* pnv only */
+
+#define BLC_HIST_CTL	_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x61260)
+#define  BLM_HISTOGRAM_ENABLE			(1 << 31)
+
+/* New registers for PCH-split platforms. Safe where new bits show up, the
+ * register layout machtes with gen4 BLC_PWM_CTL[12]. */
+#define BLC_PWM_CPU_CTL2	_MMIO(0x48250)
+#define BLC_PWM_CPU_CTL		_MMIO(0x48254)
+
+#define HSW_BLC_PWM2_CTL	_MMIO(0x48350)
+
+/* PCH CTL1 is totally different, all but the below bits are reserved. CTL2 is
+ * like the normal CTL from gen4 and earlier. Hooray for confusing naming. */
+#define BLC_PWM_PCH_CTL1	_MMIO(0xc8250)
+#define   BLM_PCH_PWM_ENABLE			(1 << 31)
+#define   BLM_PCH_OVERRIDE_ENABLE		(1 << 30)
+#define   BLM_PCH_POLARITY			(1 << 29)
+#define BLC_PWM_PCH_CTL2	_MMIO(0xc8254)
+
+/* BXT backlight register definition. */
+#define _BXT_BLC_PWM_CTL1			0xC8250
+#define   BXT_BLC_PWM_ENABLE			(1 << 31)
+#define   BXT_BLC_PWM_POLARITY			(1 << 29)
+#define _BXT_BLC_PWM_FREQ1			0xC8254
+#define _BXT_BLC_PWM_DUTY1			0xC8258
+
+#define _BXT_BLC_PWM_CTL2			0xC8350
+#define _BXT_BLC_PWM_FREQ2			0xC8354
+#define _BXT_BLC_PWM_DUTY2			0xC8358
+
+#define BXT_BLC_PWM_CTL(controller)    _MMIO_PIPE(controller,		\
+					_BXT_BLC_PWM_CTL1, _BXT_BLC_PWM_CTL2)
+#define BXT_BLC_PWM_FREQ(controller)   _MMIO_PIPE(controller, \
+					_BXT_BLC_PWM_FREQ1, _BXT_BLC_PWM_FREQ2)
+#define BXT_BLC_PWM_DUTY(controller)   _MMIO_PIPE(controller, \
+					_BXT_BLC_PWM_DUTY1, _BXT_BLC_PWM_DUTY2)
+
+/* Utility pin */
+#define UTIL_PIN_CTL			_MMIO(0x48400)
+#define   UTIL_PIN_ENABLE		(1 << 31)
+#define   UTIL_PIN_PIPE_MASK		(3 << 29)
+#define   UTIL_PIN_PIPE(x)		((x) << 29)
+#define   UTIL_PIN_MODE_MASK		(0xf << 24)
+#define   UTIL_PIN_MODE_DATA		(0 << 24)
+#define   UTIL_PIN_MODE_PWM		(1 << 24)
+#define   UTIL_PIN_MODE_VBLANK		(4 << 24)
+#define   UTIL_PIN_MODE_VSYNC		(5 << 24)
+#define   UTIL_PIN_MODE_EYE_LEVEL	(8 << 24)
+#define   UTIL_PIN_OUTPUT_DATA		(1 << 23)
+#define   UTIL_PIN_POLARITY		(1 << 22)
+#define   UTIL_PIN_DIRECTION_INPUT	(1 << 19)
+#define   UTIL_PIN_INPUT_DATA		(1 << 16)
+
+#endif /* __INTEL_BACKLIGHT_REGS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_bios.c b/drivers/gpu/drm/i915/display/intel_bios.c
new file mode 100644
index 000000000..a70b70617
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_bios.c
@@ -0,0 +1,3790 @@
+/*
+ * Copyright © 2006 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *
+ */
+
+#include <drm/drm_edid.h>
+#include <drm/display/drm_dp_helper.h>
+#include <drm/display/drm_dsc_helper.h>
+
+#include "display/intel_display.h"
+#include "display/intel_display_types.h"
+#include "display/intel_gmbus.h"
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+
+#define _INTEL_BIOS_PRIVATE
+#include "intel_vbt_defs.h"
+
+/**
+ * DOC: Video BIOS Table (VBT)
+ *
+ * The Video BIOS Table, or VBT, provides platform and board specific
+ * configuration information to the driver that is not discoverable or available
+ * through other means. The configuration is mostly related to display
+ * hardware. The VBT is available via the ACPI OpRegion or, on older systems, in
+ * the PCI ROM.
+ *
+ * The VBT consists of a VBT Header (defined as &struct vbt_header), a BDB
+ * Header (&struct bdb_header), and a number of BIOS Data Blocks (BDB) that
+ * contain the actual configuration information. The VBT Header, and thus the
+ * VBT, begins with "$VBT" signature. The VBT Header contains the offset of the
+ * BDB Header. The data blocks are concatenated after the BDB Header. The data
+ * blocks have a 1-byte Block ID, 2-byte Block Size, and Block Size bytes of
+ * data. (Block 53, the MIPI Sequence Block is an exception.)
+ *
+ * The driver parses the VBT during load. The relevant information is stored in
+ * driver private data for ease of use, and the actual VBT is not read after
+ * that.
+ */
+
+/* Wrapper for VBT child device config */
+struct intel_bios_encoder_data {
+	struct drm_i915_private *i915;
+
+	struct child_device_config child;
+	struct dsc_compression_parameters_entry *dsc;
+	struct list_head node;
+};
+
+#define	SLAVE_ADDR1	0x70
+#define	SLAVE_ADDR2	0x72
+
+/* Get BDB block size given a pointer to Block ID. */
+static u32 _get_blocksize(const u8 *block_base)
+{
+	/* The MIPI Sequence Block v3+ has a separate size field. */
+	if (*block_base == BDB_MIPI_SEQUENCE && *(block_base + 3) >= 3)
+		return *((const u32 *)(block_base + 4));
+	else
+		return *((const u16 *)(block_base + 1));
+}
+
+/* Get BDB block size give a pointer to data after Block ID and Block Size. */
+static u32 get_blocksize(const void *block_data)
+{
+	return _get_blocksize(block_data - 3);
+}
+
+static const void *
+find_raw_section(const void *_bdb, enum bdb_block_id section_id)
+{
+	const struct bdb_header *bdb = _bdb;
+	const u8 *base = _bdb;
+	int index = 0;
+	u32 total, current_size;
+	enum bdb_block_id current_id;
+
+	/* skip to first section */
+	index += bdb->header_size;
+	total = bdb->bdb_size;
+
+	/* walk the sections looking for section_id */
+	while (index + 3 < total) {
+		current_id = *(base + index);
+		current_size = _get_blocksize(base + index);
+		index += 3;
+
+		if (index + current_size > total)
+			return NULL;
+
+		if (current_id == section_id)
+			return base + index;
+
+		index += current_size;
+	}
+
+	return NULL;
+}
+
+/*
+ * Offset from the start of BDB to the start of the
+ * block data (just past the block header).
+ */
+static u32 raw_block_offset(const void *bdb, enum bdb_block_id section_id)
+{
+	const void *block;
+
+	block = find_raw_section(bdb, section_id);
+	if (!block)
+		return 0;
+
+	return block - bdb;
+}
+
+struct bdb_block_entry {
+	struct list_head node;
+	enum bdb_block_id section_id;
+	u8 data[];
+};
+
+static const void *
+find_section(struct drm_i915_private *i915,
+	     enum bdb_block_id section_id)
+{
+	struct bdb_block_entry *entry;
+
+	list_for_each_entry(entry, &i915->display.vbt.bdb_blocks, node) {
+		if (entry->section_id == section_id)
+			return entry->data + 3;
+	}
+
+	return NULL;
+}
+
+static const struct {
+	enum bdb_block_id section_id;
+	size_t min_size;
+} bdb_blocks[] = {
+	{ .section_id = BDB_GENERAL_FEATURES,
+	  .min_size = sizeof(struct bdb_general_features), },
+	{ .section_id = BDB_GENERAL_DEFINITIONS,
+	  .min_size = sizeof(struct bdb_general_definitions), },
+	{ .section_id = BDB_PSR,
+	  .min_size = sizeof(struct bdb_psr), },
+	{ .section_id = BDB_DRIVER_FEATURES,
+	  .min_size = sizeof(struct bdb_driver_features), },
+	{ .section_id = BDB_SDVO_LVDS_OPTIONS,
+	  .min_size = sizeof(struct bdb_sdvo_lvds_options), },
+	{ .section_id = BDB_SDVO_PANEL_DTDS,
+	  .min_size = sizeof(struct bdb_sdvo_panel_dtds), },
+	{ .section_id = BDB_EDP,
+	  .min_size = sizeof(struct bdb_edp), },
+	{ .section_id = BDB_LVDS_OPTIONS,
+	  .min_size = sizeof(struct bdb_lvds_options), },
+	/*
+	 * BDB_LVDS_LFP_DATA depends on BDB_LVDS_LFP_DATA_PTRS,
+	 * so keep the two ordered.
+	 */
+	{ .section_id = BDB_LVDS_LFP_DATA_PTRS,
+	  .min_size = sizeof(struct bdb_lvds_lfp_data_ptrs), },
+	{ .section_id = BDB_LVDS_LFP_DATA,
+	  .min_size = 0, /* special case */ },
+	{ .section_id = BDB_LVDS_BACKLIGHT,
+	  .min_size = sizeof(struct bdb_lfp_backlight_data), },
+	{ .section_id = BDB_LFP_POWER,
+	  .min_size = sizeof(struct bdb_lfp_power), },
+	{ .section_id = BDB_MIPI_CONFIG,
+	  .min_size = sizeof(struct bdb_mipi_config), },
+	{ .section_id = BDB_MIPI_SEQUENCE,
+	  .min_size = sizeof(struct bdb_mipi_sequence) },
+	{ .section_id = BDB_COMPRESSION_PARAMETERS,
+	  .min_size = sizeof(struct bdb_compression_parameters), },
+	{ .section_id = BDB_GENERIC_DTD,
+	  .min_size = sizeof(struct bdb_generic_dtd), },
+};
+
+static size_t lfp_data_min_size(struct drm_i915_private *i915)
+{
+	const struct bdb_lvds_lfp_data_ptrs *ptrs;
+	size_t size;
+
+	ptrs = find_section(i915, BDB_LVDS_LFP_DATA_PTRS);
+	if (!ptrs)
+		return 0;
+
+	size = sizeof(struct bdb_lvds_lfp_data);
+	if (ptrs->panel_name.table_size)
+		size = max(size, ptrs->panel_name.offset +
+			   sizeof(struct bdb_lvds_lfp_data_tail));
+
+	return size;
+}
+
+static bool validate_lfp_data_ptrs(const void *bdb,
+				   const struct bdb_lvds_lfp_data_ptrs *ptrs)
+{
+	int fp_timing_size, dvo_timing_size, panel_pnp_id_size, panel_name_size;
+	int data_block_size, lfp_data_size;
+	const void *data_block;
+	int i;
+
+	data_block = find_raw_section(bdb, BDB_LVDS_LFP_DATA);
+	if (!data_block)
+		return false;
+
+	data_block_size = get_blocksize(data_block);
+	if (data_block_size == 0)
+		return false;
+
+	/* always 3 indicating the presence of fp_timing+dvo_timing+panel_pnp_id */
+	if (ptrs->lvds_entries != 3)
+		return false;
+
+	fp_timing_size = ptrs->ptr[0].fp_timing.table_size;
+	dvo_timing_size = ptrs->ptr[0].dvo_timing.table_size;
+	panel_pnp_id_size = ptrs->ptr[0].panel_pnp_id.table_size;
+	panel_name_size = ptrs->panel_name.table_size;
+
+	/* fp_timing has variable size */
+	if (fp_timing_size < 32 ||
+	    dvo_timing_size != sizeof(struct lvds_dvo_timing) ||
+	    panel_pnp_id_size != sizeof(struct lvds_pnp_id))
+		return false;
+
+	/* panel_name is not present in old VBTs */
+	if (panel_name_size != 0 &&
+	    panel_name_size != sizeof(struct lvds_lfp_panel_name))
+		return false;
+
+	lfp_data_size = ptrs->ptr[1].fp_timing.offset - ptrs->ptr[0].fp_timing.offset;
+	if (16 * lfp_data_size > data_block_size)
+		return false;
+
+	/* make sure the table entries have uniform size */
+	for (i = 1; i < 16; i++) {
+		if (ptrs->ptr[i].fp_timing.table_size != fp_timing_size ||
+		    ptrs->ptr[i].dvo_timing.table_size != dvo_timing_size ||
+		    ptrs->ptr[i].panel_pnp_id.table_size != panel_pnp_id_size)
+			return false;
+
+		if (ptrs->ptr[i].fp_timing.offset - ptrs->ptr[i-1].fp_timing.offset != lfp_data_size ||
+		    ptrs->ptr[i].dvo_timing.offset - ptrs->ptr[i-1].dvo_timing.offset != lfp_data_size ||
+		    ptrs->ptr[i].panel_pnp_id.offset - ptrs->ptr[i-1].panel_pnp_id.offset != lfp_data_size)
+			return false;
+	}
+
+	/*
+	 * Except for vlv/chv machines all real VBTs seem to have 6
+	 * unaccounted bytes in the fp_timing table. And it doesn't
+	 * appear to be a really intentional hole as the fp_timing
+	 * 0xffff terminator is always within those 6 missing bytes.
+	 */
+	if (fp_timing_size + 6 + dvo_timing_size + panel_pnp_id_size == lfp_data_size)
+		fp_timing_size += 6;
+
+	if (fp_timing_size + dvo_timing_size + panel_pnp_id_size != lfp_data_size)
+		return false;
+
+	if (ptrs->ptr[0].fp_timing.offset + fp_timing_size != ptrs->ptr[0].dvo_timing.offset ||
+	    ptrs->ptr[0].dvo_timing.offset + dvo_timing_size != ptrs->ptr[0].panel_pnp_id.offset ||
+	    ptrs->ptr[0].panel_pnp_id.offset + panel_pnp_id_size != lfp_data_size)
+		return false;
+
+	/* make sure the tables fit inside the data block */
+	for (i = 0; i < 16; i++) {
+		if (ptrs->ptr[i].fp_timing.offset + fp_timing_size > data_block_size ||
+		    ptrs->ptr[i].dvo_timing.offset + dvo_timing_size > data_block_size ||
+		    ptrs->ptr[i].panel_pnp_id.offset + panel_pnp_id_size > data_block_size)
+			return false;
+	}
+
+	if (ptrs->panel_name.offset + 16 * panel_name_size > data_block_size)
+		return false;
+
+	/* make sure fp_timing terminators are present at expected locations */
+	for (i = 0; i < 16; i++) {
+		const u16 *t = data_block + ptrs->ptr[i].fp_timing.offset +
+			fp_timing_size - 2;
+
+		if (*t != 0xffff)
+			return false;
+	}
+
+	return true;
+}
+
+/* make the data table offsets relative to the data block */
+static bool fixup_lfp_data_ptrs(const void *bdb, void *ptrs_block)
+{
+	struct bdb_lvds_lfp_data_ptrs *ptrs = ptrs_block;
+	u32 offset;
+	int i;
+
+	offset = raw_block_offset(bdb, BDB_LVDS_LFP_DATA);
+
+	for (i = 0; i < 16; i++) {
+		if (ptrs->ptr[i].fp_timing.offset < offset ||
+		    ptrs->ptr[i].dvo_timing.offset < offset ||
+		    ptrs->ptr[i].panel_pnp_id.offset < offset)
+			return false;
+
+		ptrs->ptr[i].fp_timing.offset -= offset;
+		ptrs->ptr[i].dvo_timing.offset -= offset;
+		ptrs->ptr[i].panel_pnp_id.offset -= offset;
+	}
+
+	if (ptrs->panel_name.table_size) {
+		if (ptrs->panel_name.offset < offset)
+			return false;
+
+		ptrs->panel_name.offset -= offset;
+	}
+
+	return validate_lfp_data_ptrs(bdb, ptrs);
+}
+
+static int make_lfp_data_ptr(struct lvds_lfp_data_ptr_table *table,
+			     int table_size, int total_size)
+{
+	if (total_size < table_size)
+		return total_size;
+
+	table->table_size = table_size;
+	table->offset = total_size - table_size;
+
+	return total_size - table_size;
+}
+
+static void next_lfp_data_ptr(struct lvds_lfp_data_ptr_table *next,
+			      const struct lvds_lfp_data_ptr_table *prev,
+			      int size)
+{
+	next->table_size = prev->table_size;
+	next->offset = prev->offset + size;
+}
+
+static void *generate_lfp_data_ptrs(struct drm_i915_private *i915,
+				    const void *bdb)
+{
+	int i, size, table_size, block_size, offset, fp_timing_size;
+	struct bdb_lvds_lfp_data_ptrs *ptrs;
+	const void *block;
+	void *ptrs_block;
+
+	/*
+	 * The hardcoded fp_timing_size is only valid for
+	 * modernish VBTs. All older VBTs definitely should
+	 * include block 41 and thus we don't need to
+	 * generate one.
+	 */
+	if (i915->display.vbt.version < 155)
+		return NULL;
+
+	fp_timing_size = 38;
+
+	block = find_raw_section(bdb, BDB_LVDS_LFP_DATA);
+	if (!block)
+		return NULL;
+
+	drm_dbg_kms(&i915->drm, "Generating LFP data table pointers\n");
+
+	block_size = get_blocksize(block);
+
+	size = fp_timing_size + sizeof(struct lvds_dvo_timing) +
+		sizeof(struct lvds_pnp_id);
+	if (size * 16 > block_size)
+		return NULL;
+
+	ptrs_block = kzalloc(sizeof(*ptrs) + 3, GFP_KERNEL);
+	if (!ptrs_block)
+		return NULL;
+
+	*(u8 *)(ptrs_block + 0) = BDB_LVDS_LFP_DATA_PTRS;
+	*(u16 *)(ptrs_block + 1) = sizeof(*ptrs);
+	ptrs = ptrs_block + 3;
+
+	table_size = sizeof(struct lvds_pnp_id);
+	size = make_lfp_data_ptr(&ptrs->ptr[0].panel_pnp_id, table_size, size);
+
+	table_size = sizeof(struct lvds_dvo_timing);
+	size = make_lfp_data_ptr(&ptrs->ptr[0].dvo_timing, table_size, size);
+
+	table_size = fp_timing_size;
+	size = make_lfp_data_ptr(&ptrs->ptr[0].fp_timing, table_size, size);
+
+	if (ptrs->ptr[0].fp_timing.table_size)
+		ptrs->lvds_entries++;
+	if (ptrs->ptr[0].dvo_timing.table_size)
+		ptrs->lvds_entries++;
+	if (ptrs->ptr[0].panel_pnp_id.table_size)
+		ptrs->lvds_entries++;
+
+	if (size != 0 || ptrs->lvds_entries != 3) {
+		kfree(ptrs_block);
+		return NULL;
+	}
+
+	size = fp_timing_size + sizeof(struct lvds_dvo_timing) +
+		sizeof(struct lvds_pnp_id);
+	for (i = 1; i < 16; i++) {
+		next_lfp_data_ptr(&ptrs->ptr[i].fp_timing, &ptrs->ptr[i-1].fp_timing, size);
+		next_lfp_data_ptr(&ptrs->ptr[i].dvo_timing, &ptrs->ptr[i-1].dvo_timing, size);
+		next_lfp_data_ptr(&ptrs->ptr[i].panel_pnp_id, &ptrs->ptr[i-1].panel_pnp_id, size);
+	}
+
+	table_size = sizeof(struct lvds_lfp_panel_name);
+
+	if (16 * (size + table_size) <= block_size) {
+		ptrs->panel_name.table_size = table_size;
+		ptrs->panel_name.offset = size * 16;
+	}
+
+	offset = block - bdb;
+
+	for (i = 0; i < 16; i++) {
+		ptrs->ptr[i].fp_timing.offset += offset;
+		ptrs->ptr[i].dvo_timing.offset += offset;
+		ptrs->ptr[i].panel_pnp_id.offset += offset;
+	}
+
+	if (ptrs->panel_name.table_size)
+		ptrs->panel_name.offset += offset;
+
+	return ptrs_block;
+}
+
+static void
+init_bdb_block(struct drm_i915_private *i915,
+	       const void *bdb, enum bdb_block_id section_id,
+	       size_t min_size)
+{
+	struct bdb_block_entry *entry;
+	void *temp_block = NULL;
+	const void *block;
+	size_t block_size;
+
+	block = find_raw_section(bdb, section_id);
+
+	/* Modern VBTs lack the LFP data table pointers block, make one up */
+	if (!block && section_id == BDB_LVDS_LFP_DATA_PTRS) {
+		temp_block = generate_lfp_data_ptrs(i915, bdb);
+		if (temp_block)
+			block = temp_block + 3;
+	}
+	if (!block)
+		return;
+
+	drm_WARN(&i915->drm, min_size == 0,
+		 "Block %d min_size is zero\n", section_id);
+
+	block_size = get_blocksize(block);
+
+	/*
+	 * Version number and new block size are considered
+	 * part of the header for MIPI sequenece block v3+.
+	 */
+	if (section_id == BDB_MIPI_SEQUENCE && *(const u8 *)block >= 3)
+		block_size += 5;
+
+	entry = kzalloc(struct_size(entry, data, max(min_size, block_size) + 3),
+			GFP_KERNEL);
+	if (!entry) {
+		kfree(temp_block);
+		return;
+	}
+
+	entry->section_id = section_id;
+	memcpy(entry->data, block - 3, block_size + 3);
+
+	kfree(temp_block);
+
+	drm_dbg_kms(&i915->drm, "Found BDB block %d (size %zu, min size %zu)\n",
+		    section_id, block_size, min_size);
+
+	if (section_id == BDB_LVDS_LFP_DATA_PTRS &&
+	    !fixup_lfp_data_ptrs(bdb, entry->data + 3)) {
+		drm_err(&i915->drm, "VBT has malformed LFP data table pointers\n");
+		kfree(entry);
+		return;
+	}
+
+	list_add_tail(&entry->node, &i915->display.vbt.bdb_blocks);
+}
+
+static void init_bdb_blocks(struct drm_i915_private *i915,
+			    const void *bdb)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(bdb_blocks); i++) {
+		enum bdb_block_id section_id = bdb_blocks[i].section_id;
+		size_t min_size = bdb_blocks[i].min_size;
+
+		if (section_id == BDB_LVDS_LFP_DATA)
+			min_size = lfp_data_min_size(i915);
+
+		init_bdb_block(i915, bdb, section_id, min_size);
+	}
+}
+
+static void
+fill_detail_timing_data(struct drm_display_mode *panel_fixed_mode,
+			const struct lvds_dvo_timing *dvo_timing)
+{
+	panel_fixed_mode->hdisplay = (dvo_timing->hactive_hi << 8) |
+		dvo_timing->hactive_lo;
+	panel_fixed_mode->hsync_start = panel_fixed_mode->hdisplay +
+		((dvo_timing->hsync_off_hi << 8) | dvo_timing->hsync_off_lo);
+	panel_fixed_mode->hsync_end = panel_fixed_mode->hsync_start +
+		((dvo_timing->hsync_pulse_width_hi << 8) |
+			dvo_timing->hsync_pulse_width_lo);
+	panel_fixed_mode->htotal = panel_fixed_mode->hdisplay +
+		((dvo_timing->hblank_hi << 8) | dvo_timing->hblank_lo);
+
+	panel_fixed_mode->vdisplay = (dvo_timing->vactive_hi << 8) |
+		dvo_timing->vactive_lo;
+	panel_fixed_mode->vsync_start = panel_fixed_mode->vdisplay +
+		((dvo_timing->vsync_off_hi << 4) | dvo_timing->vsync_off_lo);
+	panel_fixed_mode->vsync_end = panel_fixed_mode->vsync_start +
+		((dvo_timing->vsync_pulse_width_hi << 4) |
+			dvo_timing->vsync_pulse_width_lo);
+	panel_fixed_mode->vtotal = panel_fixed_mode->vdisplay +
+		((dvo_timing->vblank_hi << 8) | dvo_timing->vblank_lo);
+	panel_fixed_mode->clock = dvo_timing->clock * 10;
+	panel_fixed_mode->type = DRM_MODE_TYPE_PREFERRED;
+
+	if (dvo_timing->hsync_positive)
+		panel_fixed_mode->flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		panel_fixed_mode->flags |= DRM_MODE_FLAG_NHSYNC;
+
+	if (dvo_timing->vsync_positive)
+		panel_fixed_mode->flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		panel_fixed_mode->flags |= DRM_MODE_FLAG_NVSYNC;
+
+	panel_fixed_mode->width_mm = (dvo_timing->himage_hi << 8) |
+		dvo_timing->himage_lo;
+	panel_fixed_mode->height_mm = (dvo_timing->vimage_hi << 8) |
+		dvo_timing->vimage_lo;
+
+	/* Some VBTs have bogus h/vtotal values */
+	if (panel_fixed_mode->hsync_end > panel_fixed_mode->htotal)
+		panel_fixed_mode->htotal = panel_fixed_mode->hsync_end + 1;
+	if (panel_fixed_mode->vsync_end > panel_fixed_mode->vtotal)
+		panel_fixed_mode->vtotal = panel_fixed_mode->vsync_end + 1;
+
+	drm_mode_set_name(panel_fixed_mode);
+}
+
+static const struct lvds_dvo_timing *
+get_lvds_dvo_timing(const struct bdb_lvds_lfp_data *data,
+		    const struct bdb_lvds_lfp_data_ptrs *ptrs,
+		    int index)
+{
+	return (const void *)data + ptrs->ptr[index].dvo_timing.offset;
+}
+
+static const struct lvds_fp_timing *
+get_lvds_fp_timing(const struct bdb_lvds_lfp_data *data,
+		   const struct bdb_lvds_lfp_data_ptrs *ptrs,
+		   int index)
+{
+	return (const void *)data + ptrs->ptr[index].fp_timing.offset;
+}
+
+static const struct lvds_pnp_id *
+get_lvds_pnp_id(const struct bdb_lvds_lfp_data *data,
+		const struct bdb_lvds_lfp_data_ptrs *ptrs,
+		int index)
+{
+	return (const void *)data + ptrs->ptr[index].panel_pnp_id.offset;
+}
+
+static const struct bdb_lvds_lfp_data_tail *
+get_lfp_data_tail(const struct bdb_lvds_lfp_data *data,
+		  const struct bdb_lvds_lfp_data_ptrs *ptrs)
+{
+	if (ptrs->panel_name.table_size)
+		return (const void *)data + ptrs->panel_name.offset;
+	else
+		return NULL;
+}
+
+static void dump_pnp_id(struct drm_i915_private *i915,
+			const struct lvds_pnp_id *pnp_id,
+			const char *name)
+{
+	u16 mfg_name = be16_to_cpu((__force __be16)pnp_id->mfg_name);
+	char vend[4];
+
+	drm_dbg_kms(&i915->drm, "%s PNPID mfg: %s (0x%x), prod: %u, serial: %u, week: %d, year: %d\n",
+		    name, drm_edid_decode_mfg_id(mfg_name, vend),
+		    pnp_id->mfg_name, pnp_id->product_code, pnp_id->serial,
+		    pnp_id->mfg_week, pnp_id->mfg_year + 1990);
+}
+
+static int opregion_get_panel_type(struct drm_i915_private *i915,
+				   const struct intel_bios_encoder_data *devdata,
+				   const struct edid *edid, bool use_fallback)
+{
+	return intel_opregion_get_panel_type(i915);
+}
+
+static int vbt_get_panel_type(struct drm_i915_private *i915,
+			      const struct intel_bios_encoder_data *devdata,
+			      const struct edid *edid, bool use_fallback)
+{
+	const struct bdb_lvds_options *lvds_options;
+
+	lvds_options = find_section(i915, BDB_LVDS_OPTIONS);
+	if (!lvds_options)
+		return -1;
+
+	if (lvds_options->panel_type > 0xf &&
+	    lvds_options->panel_type != 0xff) {
+		drm_dbg_kms(&i915->drm, "Invalid VBT panel type 0x%x\n",
+			    lvds_options->panel_type);
+		return -1;
+	}
+
+	if (devdata && devdata->child.handle == DEVICE_HANDLE_LFP2)
+		return lvds_options->panel_type2;
+
+	drm_WARN_ON(&i915->drm, devdata && devdata->child.handle != DEVICE_HANDLE_LFP1);
+
+	return lvds_options->panel_type;
+}
+
+static int pnpid_get_panel_type(struct drm_i915_private *i915,
+				const struct intel_bios_encoder_data *devdata,
+				const struct edid *edid, bool use_fallback)
+{
+	const struct bdb_lvds_lfp_data *data;
+	const struct bdb_lvds_lfp_data_ptrs *ptrs;
+	const struct lvds_pnp_id *edid_id;
+	struct lvds_pnp_id edid_id_nodate;
+	int i, best = -1;
+
+	if (!edid)
+		return -1;
+
+	edid_id = (const void *)&edid->mfg_id[0];
+
+	edid_id_nodate = *edid_id;
+	edid_id_nodate.mfg_week = 0;
+	edid_id_nodate.mfg_year = 0;
+
+	dump_pnp_id(i915, edid_id, "EDID");
+
+	ptrs = find_section(i915, BDB_LVDS_LFP_DATA_PTRS);
+	if (!ptrs)
+		return -1;
+
+	data = find_section(i915, BDB_LVDS_LFP_DATA);
+	if (!data)
+		return -1;
+
+	for (i = 0; i < 16; i++) {
+		const struct lvds_pnp_id *vbt_id =
+			get_lvds_pnp_id(data, ptrs, i);
+
+		/* full match? */
+		if (!memcmp(vbt_id, edid_id, sizeof(*vbt_id)))
+			return i;
+
+		/*
+		 * Accept a match w/o date if no full match is found,
+		 * and the VBT entry does not specify a date.
+		 */
+		if (best < 0 &&
+		    !memcmp(vbt_id, &edid_id_nodate, sizeof(*vbt_id)))
+			best = i;
+	}
+
+	return best;
+}
+
+static int fallback_get_panel_type(struct drm_i915_private *i915,
+				   const struct intel_bios_encoder_data *devdata,
+				   const struct edid *edid, bool use_fallback)
+{
+	return use_fallback ? 0 : -1;
+}
+
+enum panel_type {
+	PANEL_TYPE_OPREGION,
+	PANEL_TYPE_VBT,
+	PANEL_TYPE_PNPID,
+	PANEL_TYPE_FALLBACK,
+};
+
+static int get_panel_type(struct drm_i915_private *i915,
+			  const struct intel_bios_encoder_data *devdata,
+			  const struct edid *edid, bool use_fallback)
+{
+	struct {
+		const char *name;
+		int (*get_panel_type)(struct drm_i915_private *i915,
+				      const struct intel_bios_encoder_data *devdata,
+				      const struct edid *edid, bool use_fallback);
+		int panel_type;
+	} panel_types[] = {
+		[PANEL_TYPE_OPREGION] = {
+			.name = "OpRegion",
+			.get_panel_type = opregion_get_panel_type,
+		},
+		[PANEL_TYPE_VBT] = {
+			.name = "VBT",
+			.get_panel_type = vbt_get_panel_type,
+		},
+		[PANEL_TYPE_PNPID] = {
+			.name = "PNPID",
+			.get_panel_type = pnpid_get_panel_type,
+		},
+		[PANEL_TYPE_FALLBACK] = {
+			.name = "fallback",
+			.get_panel_type = fallback_get_panel_type,
+		},
+	};
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(panel_types); i++) {
+		panel_types[i].panel_type = panel_types[i].get_panel_type(i915, devdata,
+									  edid, use_fallback);
+
+		drm_WARN_ON(&i915->drm, panel_types[i].panel_type > 0xf &&
+			    panel_types[i].panel_type != 0xff);
+
+		if (panel_types[i].panel_type >= 0)
+			drm_dbg_kms(&i915->drm, "Panel type (%s): %d\n",
+				    panel_types[i].name, panel_types[i].panel_type);
+	}
+
+	if (panel_types[PANEL_TYPE_OPREGION].panel_type >= 0)
+		i = PANEL_TYPE_OPREGION;
+	else if (panel_types[PANEL_TYPE_VBT].panel_type == 0xff &&
+		 panel_types[PANEL_TYPE_PNPID].panel_type >= 0)
+		i = PANEL_TYPE_PNPID;
+	else if (panel_types[PANEL_TYPE_VBT].panel_type != 0xff &&
+		 panel_types[PANEL_TYPE_VBT].panel_type >= 0)
+		i = PANEL_TYPE_VBT;
+	else
+		i = PANEL_TYPE_FALLBACK;
+
+	drm_dbg_kms(&i915->drm, "Selected panel type (%s): %d\n",
+		    panel_types[i].name, panel_types[i].panel_type);
+
+	return panel_types[i].panel_type;
+}
+
+static unsigned int panel_bits(unsigned int value, int panel_type, int num_bits)
+{
+	return (value >> (panel_type * num_bits)) & (BIT(num_bits) - 1);
+}
+
+static bool panel_bool(unsigned int value, int panel_type)
+{
+	return panel_bits(value, panel_type, 1);
+}
+
+/* Parse general panel options */
+static void
+parse_panel_options(struct drm_i915_private *i915,
+		    struct intel_panel *panel)
+{
+	const struct bdb_lvds_options *lvds_options;
+	int panel_type = panel->vbt.panel_type;
+	int drrs_mode;
+
+	lvds_options = find_section(i915, BDB_LVDS_OPTIONS);
+	if (!lvds_options)
+		return;
+
+	panel->vbt.lvds_dither = lvds_options->pixel_dither;
+
+	/*
+	 * Empirical evidence indicates the block size can be
+	 * either 4,14,16,24+ bytes. For older VBTs no clear
+	 * relationship between the block size vs. BDB version.
+	 */
+	if (get_blocksize(lvds_options) < 16)
+		return;
+
+	drrs_mode = panel_bits(lvds_options->dps_panel_type_bits,
+			       panel_type, 2);
+	/*
+	 * VBT has static DRRS = 0 and seamless DRRS = 2.
+	 * The below piece of code is required to adjust vbt.drrs_type
+	 * to match the enum drrs_support_type.
+	 */
+	switch (drrs_mode) {
+	case 0:
+		panel->vbt.drrs_type = DRRS_TYPE_STATIC;
+		drm_dbg_kms(&i915->drm, "DRRS supported mode is static\n");
+		break;
+	case 2:
+		panel->vbt.drrs_type = DRRS_TYPE_SEAMLESS;
+		drm_dbg_kms(&i915->drm,
+			    "DRRS supported mode is seamless\n");
+		break;
+	default:
+		panel->vbt.drrs_type = DRRS_TYPE_NONE;
+		drm_dbg_kms(&i915->drm,
+			    "DRRS not supported (VBT input)\n");
+		break;
+	}
+}
+
+static void
+parse_lfp_panel_dtd(struct drm_i915_private *i915,
+		    struct intel_panel *panel,
+		    const struct bdb_lvds_lfp_data *lvds_lfp_data,
+		    const struct bdb_lvds_lfp_data_ptrs *lvds_lfp_data_ptrs)
+{
+	const struct lvds_dvo_timing *panel_dvo_timing;
+	const struct lvds_fp_timing *fp_timing;
+	struct drm_display_mode *panel_fixed_mode;
+	int panel_type = panel->vbt.panel_type;
+
+	panel_dvo_timing = get_lvds_dvo_timing(lvds_lfp_data,
+					       lvds_lfp_data_ptrs,
+					       panel_type);
+
+	panel_fixed_mode = kzalloc(sizeof(*panel_fixed_mode), GFP_KERNEL);
+	if (!panel_fixed_mode)
+		return;
+
+	fill_detail_timing_data(panel_fixed_mode, panel_dvo_timing);
+
+	panel->vbt.lfp_lvds_vbt_mode = panel_fixed_mode;
+
+	drm_dbg_kms(&i915->drm,
+		    "Found panel mode in BIOS VBT legacy lfp table: " DRM_MODE_FMT "\n",
+		    DRM_MODE_ARG(panel_fixed_mode));
+
+	fp_timing = get_lvds_fp_timing(lvds_lfp_data,
+				       lvds_lfp_data_ptrs,
+				       panel_type);
+
+	/* check the resolution, just to be sure */
+	if (fp_timing->x_res == panel_fixed_mode->hdisplay &&
+	    fp_timing->y_res == panel_fixed_mode->vdisplay) {
+		panel->vbt.bios_lvds_val = fp_timing->lvds_reg_val;
+		drm_dbg_kms(&i915->drm,
+			    "VBT initial LVDS value %x\n",
+			    panel->vbt.bios_lvds_val);
+	}
+}
+
+static void
+parse_lfp_data(struct drm_i915_private *i915,
+	       struct intel_panel *panel)
+{
+	const struct bdb_lvds_lfp_data *data;
+	const struct bdb_lvds_lfp_data_tail *tail;
+	const struct bdb_lvds_lfp_data_ptrs *ptrs;
+	const struct lvds_pnp_id *pnp_id;
+	int panel_type = panel->vbt.panel_type;
+
+	ptrs = find_section(i915, BDB_LVDS_LFP_DATA_PTRS);
+	if (!ptrs)
+		return;
+
+	data = find_section(i915, BDB_LVDS_LFP_DATA);
+	if (!data)
+		return;
+
+	if (!panel->vbt.lfp_lvds_vbt_mode)
+		parse_lfp_panel_dtd(i915, panel, data, ptrs);
+
+	pnp_id = get_lvds_pnp_id(data, ptrs, panel_type);
+	dump_pnp_id(i915, pnp_id, "Panel");
+
+	tail = get_lfp_data_tail(data, ptrs);
+	if (!tail)
+		return;
+
+	drm_dbg_kms(&i915->drm, "Panel name: %.*s\n",
+		    (int)sizeof(tail->panel_name[0].name),
+		    tail->panel_name[panel_type].name);
+
+	if (i915->display.vbt.version >= 188) {
+		panel->vbt.seamless_drrs_min_refresh_rate =
+			tail->seamless_drrs_min_refresh_rate[panel_type];
+		drm_dbg_kms(&i915->drm,
+			    "Seamless DRRS min refresh rate: %d Hz\n",
+			    panel->vbt.seamless_drrs_min_refresh_rate);
+	}
+}
+
+static void
+parse_generic_dtd(struct drm_i915_private *i915,
+		  struct intel_panel *panel)
+{
+	const struct bdb_generic_dtd *generic_dtd;
+	const struct generic_dtd_entry *dtd;
+	struct drm_display_mode *panel_fixed_mode;
+	int num_dtd;
+
+	/*
+	 * Older VBTs provided DTD information for internal displays through
+	 * the "LFP panel tables" block (42).  As of VBT revision 229 the
+	 * DTD information should be provided via a newer "generic DTD"
+	 * block (58).  Just to be safe, we'll try the new generic DTD block
+	 * first on VBT >= 229, but still fall back to trying the old LFP
+	 * block if that fails.
+	 */
+	if (i915->display.vbt.version < 229)
+		return;
+
+	generic_dtd = find_section(i915, BDB_GENERIC_DTD);
+	if (!generic_dtd)
+		return;
+
+	if (generic_dtd->gdtd_size < sizeof(struct generic_dtd_entry)) {
+		drm_err(&i915->drm, "GDTD size %u is too small.\n",
+			generic_dtd->gdtd_size);
+		return;
+	} else if (generic_dtd->gdtd_size !=
+		   sizeof(struct generic_dtd_entry)) {
+		drm_err(&i915->drm, "Unexpected GDTD size %u\n",
+			generic_dtd->gdtd_size);
+		/* DTD has unknown fields, but keep going */
+	}
+
+	num_dtd = (get_blocksize(generic_dtd) -
+		   sizeof(struct bdb_generic_dtd)) / generic_dtd->gdtd_size;
+	if (panel->vbt.panel_type >= num_dtd) {
+		drm_err(&i915->drm,
+			"Panel type %d not found in table of %d DTD's\n",
+			panel->vbt.panel_type, num_dtd);
+		return;
+	}
+
+	dtd = &generic_dtd->dtd[panel->vbt.panel_type];
+
+	panel_fixed_mode = kzalloc(sizeof(*panel_fixed_mode), GFP_KERNEL);
+	if (!panel_fixed_mode)
+		return;
+
+	panel_fixed_mode->hdisplay = dtd->hactive;
+	panel_fixed_mode->hsync_start =
+		panel_fixed_mode->hdisplay + dtd->hfront_porch;
+	panel_fixed_mode->hsync_end =
+		panel_fixed_mode->hsync_start + dtd->hsync;
+	panel_fixed_mode->htotal =
+		panel_fixed_mode->hdisplay + dtd->hblank;
+
+	panel_fixed_mode->vdisplay = dtd->vactive;
+	panel_fixed_mode->vsync_start =
+		panel_fixed_mode->vdisplay + dtd->vfront_porch;
+	panel_fixed_mode->vsync_end =
+		panel_fixed_mode->vsync_start + dtd->vsync;
+	panel_fixed_mode->vtotal =
+		panel_fixed_mode->vdisplay + dtd->vblank;
+
+	panel_fixed_mode->clock = dtd->pixel_clock;
+	panel_fixed_mode->width_mm = dtd->width_mm;
+	panel_fixed_mode->height_mm = dtd->height_mm;
+
+	panel_fixed_mode->type = DRM_MODE_TYPE_PREFERRED;
+	drm_mode_set_name(panel_fixed_mode);
+
+	if (dtd->hsync_positive_polarity)
+		panel_fixed_mode->flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		panel_fixed_mode->flags |= DRM_MODE_FLAG_NHSYNC;
+
+	if (dtd->vsync_positive_polarity)
+		panel_fixed_mode->flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		panel_fixed_mode->flags |= DRM_MODE_FLAG_NVSYNC;
+
+	drm_dbg_kms(&i915->drm,
+		    "Found panel mode in BIOS VBT generic dtd table: " DRM_MODE_FMT "\n",
+		    DRM_MODE_ARG(panel_fixed_mode));
+
+	panel->vbt.lfp_lvds_vbt_mode = panel_fixed_mode;
+}
+
+static void
+parse_lfp_backlight(struct drm_i915_private *i915,
+		    struct intel_panel *panel)
+{
+	const struct bdb_lfp_backlight_data *backlight_data;
+	const struct lfp_backlight_data_entry *entry;
+	int panel_type = panel->vbt.panel_type;
+	u16 level;
+
+	backlight_data = find_section(i915, BDB_LVDS_BACKLIGHT);
+	if (!backlight_data)
+		return;
+
+	if (backlight_data->entry_size != sizeof(backlight_data->data[0])) {
+		drm_dbg_kms(&i915->drm,
+			    "Unsupported backlight data entry size %u\n",
+			    backlight_data->entry_size);
+		return;
+	}
+
+	entry = &backlight_data->data[panel_type];
+
+	panel->vbt.backlight.present = entry->type == BDB_BACKLIGHT_TYPE_PWM;
+	if (!panel->vbt.backlight.present) {
+		drm_dbg_kms(&i915->drm,
+			    "PWM backlight not present in VBT (type %u)\n",
+			    entry->type);
+		return;
+	}
+
+	panel->vbt.backlight.type = INTEL_BACKLIGHT_DISPLAY_DDI;
+	if (i915->display.vbt.version >= 191) {
+		size_t exp_size;
+
+		if (i915->display.vbt.version >= 236)
+			exp_size = sizeof(struct bdb_lfp_backlight_data);
+		else if (i915->display.vbt.version >= 234)
+			exp_size = EXP_BDB_LFP_BL_DATA_SIZE_REV_234;
+		else
+			exp_size = EXP_BDB_LFP_BL_DATA_SIZE_REV_191;
+
+		if (get_blocksize(backlight_data) >= exp_size) {
+			const struct lfp_backlight_control_method *method;
+
+			method = &backlight_data->backlight_control[panel_type];
+			panel->vbt.backlight.type = method->type;
+			panel->vbt.backlight.controller = method->controller;
+		}
+	}
+
+	panel->vbt.backlight.pwm_freq_hz = entry->pwm_freq_hz;
+	panel->vbt.backlight.active_low_pwm = entry->active_low_pwm;
+
+	if (i915->display.vbt.version >= 234) {
+		u16 min_level;
+		bool scale;
+
+		level = backlight_data->brightness_level[panel_type].level;
+		min_level = backlight_data->brightness_min_level[panel_type].level;
+
+		if (i915->display.vbt.version >= 236)
+			scale = backlight_data->brightness_precision_bits[panel_type] == 16;
+		else
+			scale = level > 255;
+
+		if (scale)
+			min_level = min_level / 255;
+
+		if (min_level > 255) {
+			drm_warn(&i915->drm, "Brightness min level > 255\n");
+			level = 255;
+		}
+		panel->vbt.backlight.min_brightness = min_level;
+
+		panel->vbt.backlight.brightness_precision_bits =
+			backlight_data->brightness_precision_bits[panel_type];
+	} else {
+		level = backlight_data->level[panel_type];
+		panel->vbt.backlight.min_brightness = entry->min_brightness;
+	}
+
+	drm_dbg_kms(&i915->drm,
+		    "VBT backlight PWM modulation frequency %u Hz, "
+		    "active %s, min brightness %u, level %u, controller %u\n",
+		    panel->vbt.backlight.pwm_freq_hz,
+		    panel->vbt.backlight.active_low_pwm ? "low" : "high",
+		    panel->vbt.backlight.min_brightness,
+		    level,
+		    panel->vbt.backlight.controller);
+}
+
+/* Try to find sdvo panel data */
+static void
+parse_sdvo_panel_data(struct drm_i915_private *i915,
+		      struct intel_panel *panel)
+{
+	const struct bdb_sdvo_panel_dtds *dtds;
+	struct drm_display_mode *panel_fixed_mode;
+	int index;
+
+	index = i915->params.vbt_sdvo_panel_type;
+	if (index == -2) {
+		drm_dbg_kms(&i915->drm,
+			    "Ignore SDVO panel mode from BIOS VBT tables.\n");
+		return;
+	}
+
+	if (index == -1) {
+		const struct bdb_sdvo_lvds_options *sdvo_lvds_options;
+
+		sdvo_lvds_options = find_section(i915, BDB_SDVO_LVDS_OPTIONS);
+		if (!sdvo_lvds_options)
+			return;
+
+		index = sdvo_lvds_options->panel_type;
+	}
+
+	dtds = find_section(i915, BDB_SDVO_PANEL_DTDS);
+	if (!dtds)
+		return;
+
+	panel_fixed_mode = kzalloc(sizeof(*panel_fixed_mode), GFP_KERNEL);
+	if (!panel_fixed_mode)
+		return;
+
+	fill_detail_timing_data(panel_fixed_mode, &dtds->dtds[index]);
+
+	panel->vbt.sdvo_lvds_vbt_mode = panel_fixed_mode;
+
+	drm_dbg_kms(&i915->drm,
+		    "Found SDVO panel mode in BIOS VBT tables: " DRM_MODE_FMT "\n",
+		    DRM_MODE_ARG(panel_fixed_mode));
+}
+
+static int intel_bios_ssc_frequency(struct drm_i915_private *i915,
+				    bool alternate)
+{
+	switch (DISPLAY_VER(i915)) {
+	case 2:
+		return alternate ? 66667 : 48000;
+	case 3:
+	case 4:
+		return alternate ? 100000 : 96000;
+	default:
+		return alternate ? 100000 : 120000;
+	}
+}
+
+static void
+parse_general_features(struct drm_i915_private *i915)
+{
+	const struct bdb_general_features *general;
+
+	general = find_section(i915, BDB_GENERAL_FEATURES);
+	if (!general)
+		return;
+
+	i915->display.vbt.int_tv_support = general->int_tv_support;
+	/* int_crt_support can't be trusted on earlier platforms */
+	if (i915->display.vbt.version >= 155 &&
+	    (HAS_DDI(i915) || IS_VALLEYVIEW(i915)))
+		i915->display.vbt.int_crt_support = general->int_crt_support;
+	i915->display.vbt.lvds_use_ssc = general->enable_ssc;
+	i915->display.vbt.lvds_ssc_freq =
+		intel_bios_ssc_frequency(i915, general->ssc_freq);
+	i915->display.vbt.display_clock_mode = general->display_clock_mode;
+	i915->display.vbt.fdi_rx_polarity_inverted = general->fdi_rx_polarity_inverted;
+	if (i915->display.vbt.version >= 181) {
+		i915->display.vbt.orientation = general->rotate_180 ?
+			DRM_MODE_PANEL_ORIENTATION_BOTTOM_UP :
+			DRM_MODE_PANEL_ORIENTATION_NORMAL;
+	} else {
+		i915->display.vbt.orientation = DRM_MODE_PANEL_ORIENTATION_UNKNOWN;
+	}
+
+	if (i915->display.vbt.version >= 249 && general->afc_startup_config) {
+		i915->display.vbt.override_afc_startup = true;
+		i915->display.vbt.override_afc_startup_val = general->afc_startup_config == 0x1 ? 0x0 : 0x7;
+	}
+
+	drm_dbg_kms(&i915->drm,
+		    "BDB_GENERAL_FEATURES int_tv_support %d int_crt_support %d lvds_use_ssc %d lvds_ssc_freq %d display_clock_mode %d fdi_rx_polarity_inverted %d\n",
+		    i915->display.vbt.int_tv_support,
+		    i915->display.vbt.int_crt_support,
+		    i915->display.vbt.lvds_use_ssc,
+		    i915->display.vbt.lvds_ssc_freq,
+		    i915->display.vbt.display_clock_mode,
+		    i915->display.vbt.fdi_rx_polarity_inverted);
+}
+
+static const struct child_device_config *
+child_device_ptr(const struct bdb_general_definitions *defs, int i)
+{
+	return (const void *) &defs->devices[i * defs->child_dev_size];
+}
+
+static void
+parse_sdvo_device_mapping(struct drm_i915_private *i915)
+{
+	struct sdvo_device_mapping *mapping;
+	const struct intel_bios_encoder_data *devdata;
+	const struct child_device_config *child;
+	int count = 0;
+
+	/*
+	 * Only parse SDVO mappings on gens that could have SDVO. This isn't
+	 * accurate and doesn't have to be, as long as it's not too strict.
+	 */
+	if (!IS_DISPLAY_VER(i915, 3, 7)) {
+		drm_dbg_kms(&i915->drm, "Skipping SDVO device mapping\n");
+		return;
+	}
+
+	list_for_each_entry(devdata, &i915->display.vbt.display_devices, node) {
+		child = &devdata->child;
+
+		if (child->slave_addr != SLAVE_ADDR1 &&
+		    child->slave_addr != SLAVE_ADDR2) {
+			/*
+			 * If the slave address is neither 0x70 nor 0x72,
+			 * it is not a SDVO device. Skip it.
+			 */
+			continue;
+		}
+		if (child->dvo_port != DEVICE_PORT_DVOB &&
+		    child->dvo_port != DEVICE_PORT_DVOC) {
+			/* skip the incorrect SDVO port */
+			drm_dbg_kms(&i915->drm,
+				    "Incorrect SDVO port. Skip it\n");
+			continue;
+		}
+		drm_dbg_kms(&i915->drm,
+			    "the SDVO device with slave addr %2x is found on"
+			    " %s port\n",
+			    child->slave_addr,
+			    (child->dvo_port == DEVICE_PORT_DVOB) ?
+			    "SDVOB" : "SDVOC");
+		mapping = &i915->display.vbt.sdvo_mappings[child->dvo_port - 1];
+		if (!mapping->initialized) {
+			mapping->dvo_port = child->dvo_port;
+			mapping->slave_addr = child->slave_addr;
+			mapping->dvo_wiring = child->dvo_wiring;
+			mapping->ddc_pin = child->ddc_pin;
+			mapping->i2c_pin = child->i2c_pin;
+			mapping->initialized = 1;
+			drm_dbg_kms(&i915->drm,
+				    "SDVO device: dvo=%x, addr=%x, wiring=%d, ddc_pin=%d, i2c_pin=%d\n",
+				    mapping->dvo_port, mapping->slave_addr,
+				    mapping->dvo_wiring, mapping->ddc_pin,
+				    mapping->i2c_pin);
+		} else {
+			drm_dbg_kms(&i915->drm,
+				    "Maybe one SDVO port is shared by "
+				    "two SDVO device.\n");
+		}
+		if (child->slave2_addr) {
+			/* Maybe this is a SDVO device with multiple inputs */
+			/* And the mapping info is not added */
+			drm_dbg_kms(&i915->drm,
+				    "there exists the slave2_addr. Maybe this"
+				    " is a SDVO device with multiple inputs.\n");
+		}
+		count++;
+	}
+
+	if (!count) {
+		/* No SDVO device info is found */
+		drm_dbg_kms(&i915->drm,
+			    "No SDVO device info is found in VBT\n");
+	}
+}
+
+static void
+parse_driver_features(struct drm_i915_private *i915)
+{
+	const struct bdb_driver_features *driver;
+
+	driver = find_section(i915, BDB_DRIVER_FEATURES);
+	if (!driver)
+		return;
+
+	if (DISPLAY_VER(i915) >= 5) {
+		/*
+		 * Note that we consider BDB_DRIVER_FEATURE_INT_SDVO_LVDS
+		 * to mean "eDP". The VBT spec doesn't agree with that
+		 * interpretation, but real world VBTs seem to.
+		 */
+		if (driver->lvds_config != BDB_DRIVER_FEATURE_INT_LVDS)
+			i915->display.vbt.int_lvds_support = 0;
+	} else {
+		/*
+		 * FIXME it's not clear which BDB version has the LVDS config
+		 * bits defined. Revision history in the VBT spec says:
+		 * "0.92 | Add two definitions for VBT value of LVDS Active
+		 *  Config (00b and 11b values defined) | 06/13/2005"
+		 * but does not the specify the BDB version.
+		 *
+		 * So far version 134 (on i945gm) is the oldest VBT observed
+		 * in the wild with the bits correctly populated. Version
+		 * 108 (on i85x) does not have the bits correctly populated.
+		 */
+		if (i915->display.vbt.version >= 134 &&
+		    driver->lvds_config != BDB_DRIVER_FEATURE_INT_LVDS &&
+		    driver->lvds_config != BDB_DRIVER_FEATURE_INT_SDVO_LVDS)
+			i915->display.vbt.int_lvds_support = 0;
+	}
+}
+
+static void
+parse_panel_driver_features(struct drm_i915_private *i915,
+			    struct intel_panel *panel)
+{
+	const struct bdb_driver_features *driver;
+
+	driver = find_section(i915, BDB_DRIVER_FEATURES);
+	if (!driver)
+		return;
+
+	if (i915->display.vbt.version < 228) {
+		drm_dbg_kms(&i915->drm, "DRRS State Enabled:%d\n",
+			    driver->drrs_enabled);
+		/*
+		 * If DRRS is not supported, drrs_type has to be set to 0.
+		 * This is because, VBT is configured in such a way that
+		 * static DRRS is 0 and DRRS not supported is represented by
+		 * driver->drrs_enabled=false
+		 */
+		if (!driver->drrs_enabled && panel->vbt.drrs_type != DRRS_TYPE_NONE) {
+			/*
+			 * FIXME Should DMRRS perhaps be treated as seamless
+			 * but without the automatic downclocking?
+			 */
+			if (driver->dmrrs_enabled)
+				panel->vbt.drrs_type = DRRS_TYPE_STATIC;
+			else
+				panel->vbt.drrs_type = DRRS_TYPE_NONE;
+		}
+
+		panel->vbt.psr.enable = driver->psr_enabled;
+	}
+}
+
+static void
+parse_power_conservation_features(struct drm_i915_private *i915,
+				  struct intel_panel *panel)
+{
+	const struct bdb_lfp_power *power;
+	u8 panel_type = panel->vbt.panel_type;
+
+	panel->vbt.vrr = true; /* matches Windows behaviour */
+
+	if (i915->display.vbt.version < 228)
+		return;
+
+	power = find_section(i915, BDB_LFP_POWER);
+	if (!power)
+		return;
+
+	panel->vbt.psr.enable = panel_bool(power->psr, panel_type);
+
+	/*
+	 * If DRRS is not supported, drrs_type has to be set to 0.
+	 * This is because, VBT is configured in such a way that
+	 * static DRRS is 0 and DRRS not supported is represented by
+	 * power->drrs & BIT(panel_type)=false
+	 */
+	if (!panel_bool(power->drrs, panel_type) && panel->vbt.drrs_type != DRRS_TYPE_NONE) {
+		/*
+		 * FIXME Should DMRRS perhaps be treated as seamless
+		 * but without the automatic downclocking?
+		 */
+		if (panel_bool(power->dmrrs, panel_type))
+			panel->vbt.drrs_type = DRRS_TYPE_STATIC;
+		else
+			panel->vbt.drrs_type = DRRS_TYPE_NONE;
+	}
+
+	if (i915->display.vbt.version >= 232)
+		panel->vbt.edp.hobl = panel_bool(power->hobl, panel_type);
+
+	if (i915->display.vbt.version >= 233)
+		panel->vbt.vrr = panel_bool(power->vrr_feature_enabled,
+					    panel_type);
+}
+
+static void
+parse_edp(struct drm_i915_private *i915,
+	  struct intel_panel *panel)
+{
+	const struct bdb_edp *edp;
+	const struct edp_power_seq *edp_pps;
+	const struct edp_fast_link_params *edp_link_params;
+	int panel_type = panel->vbt.panel_type;
+
+	edp = find_section(i915, BDB_EDP);
+	if (!edp)
+		return;
+
+	switch (panel_bits(edp->color_depth, panel_type, 2)) {
+	case EDP_18BPP:
+		panel->vbt.edp.bpp = 18;
+		break;
+	case EDP_24BPP:
+		panel->vbt.edp.bpp = 24;
+		break;
+	case EDP_30BPP:
+		panel->vbt.edp.bpp = 30;
+		break;
+	}
+
+	/* Get the eDP sequencing and link info */
+	edp_pps = &edp->power_seqs[panel_type];
+	edp_link_params = &edp->fast_link_params[panel_type];
+
+	panel->vbt.edp.pps = *edp_pps;
+
+	if (i915->display.vbt.version >= 224) {
+		panel->vbt.edp.rate =
+			edp->edp_fast_link_training_rate[panel_type] * 20;
+	} else {
+		switch (edp_link_params->rate) {
+		case EDP_RATE_1_62:
+			panel->vbt.edp.rate = 162000;
+			break;
+		case EDP_RATE_2_7:
+			panel->vbt.edp.rate = 270000;
+			break;
+		case EDP_RATE_5_4:
+			panel->vbt.edp.rate = 540000;
+			break;
+		default:
+			drm_dbg_kms(&i915->drm,
+				    "VBT has unknown eDP link rate value %u\n",
+				    edp_link_params->rate);
+			break;
+		}
+	}
+
+	switch (edp_link_params->lanes) {
+	case EDP_LANE_1:
+		panel->vbt.edp.lanes = 1;
+		break;
+	case EDP_LANE_2:
+		panel->vbt.edp.lanes = 2;
+		break;
+	case EDP_LANE_4:
+		panel->vbt.edp.lanes = 4;
+		break;
+	default:
+		drm_dbg_kms(&i915->drm,
+			    "VBT has unknown eDP lane count value %u\n",
+			    edp_link_params->lanes);
+		break;
+	}
+
+	switch (edp_link_params->preemphasis) {
+	case EDP_PREEMPHASIS_NONE:
+		panel->vbt.edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_0;
+		break;
+	case EDP_PREEMPHASIS_3_5dB:
+		panel->vbt.edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_1;
+		break;
+	case EDP_PREEMPHASIS_6dB:
+		panel->vbt.edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_2;
+		break;
+	case EDP_PREEMPHASIS_9_5dB:
+		panel->vbt.edp.preemphasis = DP_TRAIN_PRE_EMPH_LEVEL_3;
+		break;
+	default:
+		drm_dbg_kms(&i915->drm,
+			    "VBT has unknown eDP pre-emphasis value %u\n",
+			    edp_link_params->preemphasis);
+		break;
+	}
+
+	switch (edp_link_params->vswing) {
+	case EDP_VSWING_0_4V:
+		panel->vbt.edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_0;
+		break;
+	case EDP_VSWING_0_6V:
+		panel->vbt.edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_1;
+		break;
+	case EDP_VSWING_0_8V:
+		panel->vbt.edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
+		break;
+	case EDP_VSWING_1_2V:
+		panel->vbt.edp.vswing = DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
+		break;
+	default:
+		drm_dbg_kms(&i915->drm,
+			    "VBT has unknown eDP voltage swing value %u\n",
+			    edp_link_params->vswing);
+		break;
+	}
+
+	if (i915->display.vbt.version >= 173) {
+		u8 vswing;
+
+		/* Don't read from VBT if module parameter has valid value*/
+		if (i915->params.edp_vswing) {
+			panel->vbt.edp.low_vswing =
+				i915->params.edp_vswing == 1;
+		} else {
+			vswing = (edp->edp_vswing_preemph >> (panel_type * 4)) & 0xF;
+			panel->vbt.edp.low_vswing = vswing == 0;
+		}
+	}
+
+	panel->vbt.edp.drrs_msa_timing_delay =
+		panel_bits(edp->sdrrs_msa_timing_delay, panel_type, 2);
+
+	if (i915->display.vbt.version >= 244)
+		panel->vbt.edp.max_link_rate =
+			edp->edp_max_port_link_rate[panel_type] * 20;
+}
+
+static void
+parse_psr(struct drm_i915_private *i915,
+	  struct intel_panel *panel)
+{
+	const struct bdb_psr *psr;
+	const struct psr_table *psr_table;
+	int panel_type = panel->vbt.panel_type;
+
+	psr = find_section(i915, BDB_PSR);
+	if (!psr) {
+		drm_dbg_kms(&i915->drm, "No PSR BDB found.\n");
+		return;
+	}
+
+	psr_table = &psr->psr_table[panel_type];
+
+	panel->vbt.psr.full_link = psr_table->full_link;
+	panel->vbt.psr.require_aux_wakeup = psr_table->require_aux_to_wakeup;
+
+	/* Allowed VBT values goes from 0 to 15 */
+	panel->vbt.psr.idle_frames = psr_table->idle_frames < 0 ? 0 :
+		psr_table->idle_frames > 15 ? 15 : psr_table->idle_frames;
+
+	/*
+	 * New psr options 0=500us, 1=100us, 2=2500us, 3=0us
+	 * Old decimal value is wake up time in multiples of 100 us.
+	 */
+	if (i915->display.vbt.version >= 205 &&
+	    (DISPLAY_VER(i915) >= 9 && !IS_BROXTON(i915))) {
+		switch (psr_table->tp1_wakeup_time) {
+		case 0:
+			panel->vbt.psr.tp1_wakeup_time_us = 500;
+			break;
+		case 1:
+			panel->vbt.psr.tp1_wakeup_time_us = 100;
+			break;
+		case 3:
+			panel->vbt.psr.tp1_wakeup_time_us = 0;
+			break;
+		default:
+			drm_dbg_kms(&i915->drm,
+				    "VBT tp1 wakeup time value %d is outside range[0-3], defaulting to max value 2500us\n",
+				    psr_table->tp1_wakeup_time);
+			fallthrough;
+		case 2:
+			panel->vbt.psr.tp1_wakeup_time_us = 2500;
+			break;
+		}
+
+		switch (psr_table->tp2_tp3_wakeup_time) {
+		case 0:
+			panel->vbt.psr.tp2_tp3_wakeup_time_us = 500;
+			break;
+		case 1:
+			panel->vbt.psr.tp2_tp3_wakeup_time_us = 100;
+			break;
+		case 3:
+			panel->vbt.psr.tp2_tp3_wakeup_time_us = 0;
+			break;
+		default:
+			drm_dbg_kms(&i915->drm,
+				    "VBT tp2_tp3 wakeup time value %d is outside range[0-3], defaulting to max value 2500us\n",
+				    psr_table->tp2_tp3_wakeup_time);
+			fallthrough;
+		case 2:
+			panel->vbt.psr.tp2_tp3_wakeup_time_us = 2500;
+		break;
+		}
+	} else {
+		panel->vbt.psr.tp1_wakeup_time_us = psr_table->tp1_wakeup_time * 100;
+		panel->vbt.psr.tp2_tp3_wakeup_time_us = psr_table->tp2_tp3_wakeup_time * 100;
+	}
+
+	if (i915->display.vbt.version >= 226) {
+		u32 wakeup_time = psr->psr2_tp2_tp3_wakeup_time;
+
+		wakeup_time = panel_bits(wakeup_time, panel_type, 2);
+		switch (wakeup_time) {
+		case 0:
+			wakeup_time = 500;
+			break;
+		case 1:
+			wakeup_time = 100;
+			break;
+		case 3:
+			wakeup_time = 50;
+			break;
+		default:
+		case 2:
+			wakeup_time = 2500;
+			break;
+		}
+		panel->vbt.psr.psr2_tp2_tp3_wakeup_time_us = wakeup_time;
+	} else {
+		/* Reusing PSR1 wakeup time for PSR2 in older VBTs */
+		panel->vbt.psr.psr2_tp2_tp3_wakeup_time_us = panel->vbt.psr.tp2_tp3_wakeup_time_us;
+	}
+}
+
+static void parse_dsi_backlight_ports(struct drm_i915_private *i915,
+				      struct intel_panel *panel,
+				      enum port port)
+{
+	enum port port_bc = DISPLAY_VER(i915) >= 11 ? PORT_B : PORT_C;
+
+	if (!panel->vbt.dsi.config->dual_link || i915->display.vbt.version < 197) {
+		panel->vbt.dsi.bl_ports = BIT(port);
+		if (panel->vbt.dsi.config->cabc_supported)
+			panel->vbt.dsi.cabc_ports = BIT(port);
+
+		return;
+	}
+
+	switch (panel->vbt.dsi.config->dl_dcs_backlight_ports) {
+	case DL_DCS_PORT_A:
+		panel->vbt.dsi.bl_ports = BIT(PORT_A);
+		break;
+	case DL_DCS_PORT_C:
+		panel->vbt.dsi.bl_ports = BIT(port_bc);
+		break;
+	default:
+	case DL_DCS_PORT_A_AND_C:
+		panel->vbt.dsi.bl_ports = BIT(PORT_A) | BIT(port_bc);
+		break;
+	}
+
+	if (!panel->vbt.dsi.config->cabc_supported)
+		return;
+
+	switch (panel->vbt.dsi.config->dl_dcs_cabc_ports) {
+	case DL_DCS_PORT_A:
+		panel->vbt.dsi.cabc_ports = BIT(PORT_A);
+		break;
+	case DL_DCS_PORT_C:
+		panel->vbt.dsi.cabc_ports = BIT(port_bc);
+		break;
+	default:
+	case DL_DCS_PORT_A_AND_C:
+		panel->vbt.dsi.cabc_ports =
+					BIT(PORT_A) | BIT(port_bc);
+		break;
+	}
+}
+
+static void
+parse_mipi_config(struct drm_i915_private *i915,
+		  struct intel_panel *panel)
+{
+	const struct bdb_mipi_config *start;
+	const struct mipi_config *config;
+	const struct mipi_pps_data *pps;
+	int panel_type = panel->vbt.panel_type;
+	enum port port;
+
+	/* parse MIPI blocks only if LFP type is MIPI */
+	if (!intel_bios_is_dsi_present(i915, &port))
+		return;
+
+	/* Initialize this to undefined indicating no generic MIPI support */
+	panel->vbt.dsi.panel_id = MIPI_DSI_UNDEFINED_PANEL_ID;
+
+	/* Block #40 is already parsed and panel_fixed_mode is
+	 * stored in i915->lfp_lvds_vbt_mode
+	 * resuse this when needed
+	 */
+
+	/* Parse #52 for panel index used from panel_type already
+	 * parsed
+	 */
+	start = find_section(i915, BDB_MIPI_CONFIG);
+	if (!start) {
+		drm_dbg_kms(&i915->drm, "No MIPI config BDB found");
+		return;
+	}
+
+	drm_dbg(&i915->drm, "Found MIPI Config block, panel index = %d\n",
+		panel_type);
+
+	/*
+	 * get hold of the correct configuration block and pps data as per
+	 * the panel_type as index
+	 */
+	config = &start->config[panel_type];
+	pps = &start->pps[panel_type];
+
+	/* store as of now full data. Trim when we realise all is not needed */
+	panel->vbt.dsi.config = kmemdup(config, sizeof(struct mipi_config), GFP_KERNEL);
+	if (!panel->vbt.dsi.config)
+		return;
+
+	panel->vbt.dsi.pps = kmemdup(pps, sizeof(struct mipi_pps_data), GFP_KERNEL);
+	if (!panel->vbt.dsi.pps) {
+		kfree(panel->vbt.dsi.config);
+		return;
+	}
+
+	parse_dsi_backlight_ports(i915, panel, port);
+
+	/* FIXME is the 90 vs. 270 correct? */
+	switch (config->rotation) {
+	case ENABLE_ROTATION_0:
+		/*
+		 * Most (all?) VBTs claim 0 degrees despite having
+		 * an upside down panel, thus we do not trust this.
+		 */
+		panel->vbt.dsi.orientation =
+			DRM_MODE_PANEL_ORIENTATION_UNKNOWN;
+		break;
+	case ENABLE_ROTATION_90:
+		panel->vbt.dsi.orientation =
+			DRM_MODE_PANEL_ORIENTATION_RIGHT_UP;
+		break;
+	case ENABLE_ROTATION_180:
+		panel->vbt.dsi.orientation =
+			DRM_MODE_PANEL_ORIENTATION_BOTTOM_UP;
+		break;
+	case ENABLE_ROTATION_270:
+		panel->vbt.dsi.orientation =
+			DRM_MODE_PANEL_ORIENTATION_LEFT_UP;
+		break;
+	}
+
+	/* We have mandatory mipi config blocks. Initialize as generic panel */
+	panel->vbt.dsi.panel_id = MIPI_DSI_GENERIC_PANEL_ID;
+}
+
+/* Find the sequence block and size for the given panel. */
+static const u8 *
+find_panel_sequence_block(const struct bdb_mipi_sequence *sequence,
+			  u16 panel_id, u32 *seq_size)
+{
+	u32 total = get_blocksize(sequence);
+	const u8 *data = &sequence->data[0];
+	u8 current_id;
+	u32 current_size;
+	int header_size = sequence->version >= 3 ? 5 : 3;
+	int index = 0;
+	int i;
+
+	/* skip new block size */
+	if (sequence->version >= 3)
+		data += 4;
+
+	for (i = 0; i < MAX_MIPI_CONFIGURATIONS && index < total; i++) {
+		if (index + header_size > total) {
+			DRM_ERROR("Invalid sequence block (header)\n");
+			return NULL;
+		}
+
+		current_id = *(data + index);
+		if (sequence->version >= 3)
+			current_size = *((const u32 *)(data + index + 1));
+		else
+			current_size = *((const u16 *)(data + index + 1));
+
+		index += header_size;
+
+		if (index + current_size > total) {
+			DRM_ERROR("Invalid sequence block\n");
+			return NULL;
+		}
+
+		if (current_id == panel_id) {
+			*seq_size = current_size;
+			return data + index;
+		}
+
+		index += current_size;
+	}
+
+	DRM_ERROR("Sequence block detected but no valid configuration\n");
+
+	return NULL;
+}
+
+static int goto_next_sequence(const u8 *data, int index, int total)
+{
+	u16 len;
+
+	/* Skip Sequence Byte. */
+	for (index = index + 1; index < total; index += len) {
+		u8 operation_byte = *(data + index);
+		index++;
+
+		switch (operation_byte) {
+		case MIPI_SEQ_ELEM_END:
+			return index;
+		case MIPI_SEQ_ELEM_SEND_PKT:
+			if (index + 4 > total)
+				return 0;
+
+			len = *((const u16 *)(data + index + 2)) + 4;
+			break;
+		case MIPI_SEQ_ELEM_DELAY:
+			len = 4;
+			break;
+		case MIPI_SEQ_ELEM_GPIO:
+			len = 2;
+			break;
+		case MIPI_SEQ_ELEM_I2C:
+			if (index + 7 > total)
+				return 0;
+			len = *(data + index + 6) + 7;
+			break;
+		default:
+			DRM_ERROR("Unknown operation byte\n");
+			return 0;
+		}
+	}
+
+	return 0;
+}
+
+static int goto_next_sequence_v3(const u8 *data, int index, int total)
+{
+	int seq_end;
+	u16 len;
+	u32 size_of_sequence;
+
+	/*
+	 * Could skip sequence based on Size of Sequence alone, but also do some
+	 * checking on the structure.
+	 */
+	if (total < 5) {
+		DRM_ERROR("Too small sequence size\n");
+		return 0;
+	}
+
+	/* Skip Sequence Byte. */
+	index++;
+
+	/*
+	 * Size of Sequence. Excludes the Sequence Byte and the size itself,
+	 * includes MIPI_SEQ_ELEM_END byte, excludes the final MIPI_SEQ_END
+	 * byte.
+	 */
+	size_of_sequence = *((const u32 *)(data + index));
+	index += 4;
+
+	seq_end = index + size_of_sequence;
+	if (seq_end > total) {
+		DRM_ERROR("Invalid sequence size\n");
+		return 0;
+	}
+
+	for (; index < total; index += len) {
+		u8 operation_byte = *(data + index);
+		index++;
+
+		if (operation_byte == MIPI_SEQ_ELEM_END) {
+			if (index != seq_end) {
+				DRM_ERROR("Invalid element structure\n");
+				return 0;
+			}
+			return index;
+		}
+
+		len = *(data + index);
+		index++;
+
+		/*
+		 * FIXME: Would be nice to check elements like for v1/v2 in
+		 * goto_next_sequence() above.
+		 */
+		switch (operation_byte) {
+		case MIPI_SEQ_ELEM_SEND_PKT:
+		case MIPI_SEQ_ELEM_DELAY:
+		case MIPI_SEQ_ELEM_GPIO:
+		case MIPI_SEQ_ELEM_I2C:
+		case MIPI_SEQ_ELEM_SPI:
+		case MIPI_SEQ_ELEM_PMIC:
+			break;
+		default:
+			DRM_ERROR("Unknown operation byte %u\n",
+				  operation_byte);
+			break;
+		}
+	}
+
+	return 0;
+}
+
+/*
+ * Get len of pre-fixed deassert fragment from a v1 init OTP sequence,
+ * skip all delay + gpio operands and stop at the first DSI packet op.
+ */
+static int get_init_otp_deassert_fragment_len(struct drm_i915_private *i915,
+					      struct intel_panel *panel)
+{
+	const u8 *data = panel->vbt.dsi.sequence[MIPI_SEQ_INIT_OTP];
+	int index, len;
+
+	if (drm_WARN_ON(&i915->drm,
+			!data || panel->vbt.dsi.seq_version != 1))
+		return 0;
+
+	/* index = 1 to skip sequence byte */
+	for (index = 1; data[index] != MIPI_SEQ_ELEM_END; index += len) {
+		switch (data[index]) {
+		case MIPI_SEQ_ELEM_SEND_PKT:
+			return index == 1 ? 0 : index;
+		case MIPI_SEQ_ELEM_DELAY:
+			len = 5; /* 1 byte for operand + uint32 */
+			break;
+		case MIPI_SEQ_ELEM_GPIO:
+			len = 3; /* 1 byte for op, 1 for gpio_nr, 1 for value */
+			break;
+		default:
+			return 0;
+		}
+	}
+
+	return 0;
+}
+
+/*
+ * Some v1 VBT MIPI sequences do the deassert in the init OTP sequence.
+ * The deassert must be done before calling intel_dsi_device_ready, so for
+ * these devices we split the init OTP sequence into a deassert sequence and
+ * the actual init OTP part.
+ */
+static void fixup_mipi_sequences(struct drm_i915_private *i915,
+				 struct intel_panel *panel)
+{
+	u8 *init_otp;
+	int len;
+
+	/* Limit this to VLV for now. */
+	if (!IS_VALLEYVIEW(i915))
+		return;
+
+	/* Limit this to v1 vid-mode sequences */
+	if (panel->vbt.dsi.config->is_cmd_mode ||
+	    panel->vbt.dsi.seq_version != 1)
+		return;
+
+	/* Only do this if there are otp and assert seqs and no deassert seq */
+	if (!panel->vbt.dsi.sequence[MIPI_SEQ_INIT_OTP] ||
+	    !panel->vbt.dsi.sequence[MIPI_SEQ_ASSERT_RESET] ||
+	    panel->vbt.dsi.sequence[MIPI_SEQ_DEASSERT_RESET])
+		return;
+
+	/* The deassert-sequence ends at the first DSI packet */
+	len = get_init_otp_deassert_fragment_len(i915, panel);
+	if (!len)
+		return;
+
+	drm_dbg_kms(&i915->drm,
+		    "Using init OTP fragment to deassert reset\n");
+
+	/* Copy the fragment, update seq byte and terminate it */
+	init_otp = (u8 *)panel->vbt.dsi.sequence[MIPI_SEQ_INIT_OTP];
+	panel->vbt.dsi.deassert_seq = kmemdup(init_otp, len + 1, GFP_KERNEL);
+	if (!panel->vbt.dsi.deassert_seq)
+		return;
+	panel->vbt.dsi.deassert_seq[0] = MIPI_SEQ_DEASSERT_RESET;
+	panel->vbt.dsi.deassert_seq[len] = MIPI_SEQ_ELEM_END;
+	/* Use the copy for deassert */
+	panel->vbt.dsi.sequence[MIPI_SEQ_DEASSERT_RESET] =
+		panel->vbt.dsi.deassert_seq;
+	/* Replace the last byte of the fragment with init OTP seq byte */
+	init_otp[len - 1] = MIPI_SEQ_INIT_OTP;
+	/* And make MIPI_MIPI_SEQ_INIT_OTP point to it */
+	panel->vbt.dsi.sequence[MIPI_SEQ_INIT_OTP] = init_otp + len - 1;
+}
+
+static void
+parse_mipi_sequence(struct drm_i915_private *i915,
+		    struct intel_panel *panel)
+{
+	int panel_type = panel->vbt.panel_type;
+	const struct bdb_mipi_sequence *sequence;
+	const u8 *seq_data;
+	u32 seq_size;
+	u8 *data;
+	int index = 0;
+
+	/* Only our generic panel driver uses the sequence block. */
+	if (panel->vbt.dsi.panel_id != MIPI_DSI_GENERIC_PANEL_ID)
+		return;
+
+	sequence = find_section(i915, BDB_MIPI_SEQUENCE);
+	if (!sequence) {
+		drm_dbg_kms(&i915->drm,
+			    "No MIPI Sequence found, parsing complete\n");
+		return;
+	}
+
+	/* Fail gracefully for forward incompatible sequence block. */
+	if (sequence->version >= 4) {
+		drm_err(&i915->drm,
+			"Unable to parse MIPI Sequence Block v%u\n",
+			sequence->version);
+		return;
+	}
+
+	drm_dbg(&i915->drm, "Found MIPI sequence block v%u\n",
+		sequence->version);
+
+	seq_data = find_panel_sequence_block(sequence, panel_type, &seq_size);
+	if (!seq_data)
+		return;
+
+	data = kmemdup(seq_data, seq_size, GFP_KERNEL);
+	if (!data)
+		return;
+
+	/* Parse the sequences, store pointers to each sequence. */
+	for (;;) {
+		u8 seq_id = *(data + index);
+		if (seq_id == MIPI_SEQ_END)
+			break;
+
+		if (seq_id >= MIPI_SEQ_MAX) {
+			drm_err(&i915->drm, "Unknown sequence %u\n",
+				seq_id);
+			goto err;
+		}
+
+		/* Log about presence of sequences we won't run. */
+		if (seq_id == MIPI_SEQ_TEAR_ON || seq_id == MIPI_SEQ_TEAR_OFF)
+			drm_dbg_kms(&i915->drm,
+				    "Unsupported sequence %u\n", seq_id);
+
+		panel->vbt.dsi.sequence[seq_id] = data + index;
+
+		if (sequence->version >= 3)
+			index = goto_next_sequence_v3(data, index, seq_size);
+		else
+			index = goto_next_sequence(data, index, seq_size);
+		if (!index) {
+			drm_err(&i915->drm, "Invalid sequence %u\n",
+				seq_id);
+			goto err;
+		}
+	}
+
+	panel->vbt.dsi.data = data;
+	panel->vbt.dsi.size = seq_size;
+	panel->vbt.dsi.seq_version = sequence->version;
+
+	fixup_mipi_sequences(i915, panel);
+
+	drm_dbg(&i915->drm, "MIPI related VBT parsing complete\n");
+	return;
+
+err:
+	kfree(data);
+	memset(panel->vbt.dsi.sequence, 0, sizeof(panel->vbt.dsi.sequence));
+}
+
+static void
+parse_compression_parameters(struct drm_i915_private *i915)
+{
+	const struct bdb_compression_parameters *params;
+	struct intel_bios_encoder_data *devdata;
+	const struct child_device_config *child;
+	u16 block_size;
+	int index;
+
+	if (i915->display.vbt.version < 198)
+		return;
+
+	params = find_section(i915, BDB_COMPRESSION_PARAMETERS);
+	if (params) {
+		/* Sanity checks */
+		if (params->entry_size != sizeof(params->data[0])) {
+			drm_dbg_kms(&i915->drm,
+				    "VBT: unsupported compression param entry size\n");
+			return;
+		}
+
+		block_size = get_blocksize(params);
+		if (block_size < sizeof(*params)) {
+			drm_dbg_kms(&i915->drm,
+				    "VBT: expected 16 compression param entries\n");
+			return;
+		}
+	}
+
+	list_for_each_entry(devdata, &i915->display.vbt.display_devices, node) {
+		child = &devdata->child;
+
+		if (!child->compression_enable)
+			continue;
+
+		if (!params) {
+			drm_dbg_kms(&i915->drm,
+				    "VBT: compression params not available\n");
+			continue;
+		}
+
+		if (child->compression_method_cps) {
+			drm_dbg_kms(&i915->drm,
+				    "VBT: CPS compression not supported\n");
+			continue;
+		}
+
+		index = child->compression_structure_index;
+
+		devdata->dsc = kmemdup(&params->data[index],
+				       sizeof(*devdata->dsc), GFP_KERNEL);
+	}
+}
+
+static u8 translate_iboost(u8 val)
+{
+	static const u8 mapping[] = { 1, 3, 7 }; /* See VBT spec */
+
+	if (val >= ARRAY_SIZE(mapping)) {
+		DRM_DEBUG_KMS("Unsupported I_boost value found in VBT (%d), display may not work properly\n", val);
+		return 0;
+	}
+	return mapping[val];
+}
+
+static const u8 cnp_ddc_pin_map[] = {
+	[0] = 0, /* N/A */
+	[DDC_BUS_DDI_B] = GMBUS_PIN_1_BXT,
+	[DDC_BUS_DDI_C] = GMBUS_PIN_2_BXT,
+	[DDC_BUS_DDI_D] = GMBUS_PIN_4_CNP, /* sic */
+	[DDC_BUS_DDI_F] = GMBUS_PIN_3_BXT, /* sic */
+};
+
+static const u8 icp_ddc_pin_map[] = {
+	[ICL_DDC_BUS_DDI_A] = GMBUS_PIN_1_BXT,
+	[ICL_DDC_BUS_DDI_B] = GMBUS_PIN_2_BXT,
+	[TGL_DDC_BUS_DDI_C] = GMBUS_PIN_3_BXT,
+	[ICL_DDC_BUS_PORT_1] = GMBUS_PIN_9_TC1_ICP,
+	[ICL_DDC_BUS_PORT_2] = GMBUS_PIN_10_TC2_ICP,
+	[ICL_DDC_BUS_PORT_3] = GMBUS_PIN_11_TC3_ICP,
+	[ICL_DDC_BUS_PORT_4] = GMBUS_PIN_12_TC4_ICP,
+	[TGL_DDC_BUS_PORT_5] = GMBUS_PIN_13_TC5_TGP,
+	[TGL_DDC_BUS_PORT_6] = GMBUS_PIN_14_TC6_TGP,
+};
+
+static const u8 rkl_pch_tgp_ddc_pin_map[] = {
+	[ICL_DDC_BUS_DDI_A] = GMBUS_PIN_1_BXT,
+	[ICL_DDC_BUS_DDI_B] = GMBUS_PIN_2_BXT,
+	[RKL_DDC_BUS_DDI_D] = GMBUS_PIN_9_TC1_ICP,
+	[RKL_DDC_BUS_DDI_E] = GMBUS_PIN_10_TC2_ICP,
+};
+
+static const u8 adls_ddc_pin_map[] = {
+	[ICL_DDC_BUS_DDI_A] = GMBUS_PIN_1_BXT,
+	[ADLS_DDC_BUS_PORT_TC1] = GMBUS_PIN_9_TC1_ICP,
+	[ADLS_DDC_BUS_PORT_TC2] = GMBUS_PIN_10_TC2_ICP,
+	[ADLS_DDC_BUS_PORT_TC3] = GMBUS_PIN_11_TC3_ICP,
+	[ADLS_DDC_BUS_PORT_TC4] = GMBUS_PIN_12_TC4_ICP,
+};
+
+static const u8 gen9bc_tgp_ddc_pin_map[] = {
+	[DDC_BUS_DDI_B] = GMBUS_PIN_2_BXT,
+	[DDC_BUS_DDI_C] = GMBUS_PIN_9_TC1_ICP,
+	[DDC_BUS_DDI_D] = GMBUS_PIN_10_TC2_ICP,
+};
+
+static const u8 adlp_ddc_pin_map[] = {
+	[ICL_DDC_BUS_DDI_A] = GMBUS_PIN_1_BXT,
+	[ICL_DDC_BUS_DDI_B] = GMBUS_PIN_2_BXT,
+	[ADLP_DDC_BUS_PORT_TC1] = GMBUS_PIN_9_TC1_ICP,
+	[ADLP_DDC_BUS_PORT_TC2] = GMBUS_PIN_10_TC2_ICP,
+	[ADLP_DDC_BUS_PORT_TC3] = GMBUS_PIN_11_TC3_ICP,
+	[ADLP_DDC_BUS_PORT_TC4] = GMBUS_PIN_12_TC4_ICP,
+};
+
+static u8 map_ddc_pin(struct drm_i915_private *i915, u8 vbt_pin)
+{
+	const u8 *ddc_pin_map;
+	int n_entries;
+
+	if (IS_ALDERLAKE_P(i915)) {
+		ddc_pin_map = adlp_ddc_pin_map;
+		n_entries = ARRAY_SIZE(adlp_ddc_pin_map);
+	} else if (IS_ALDERLAKE_S(i915)) {
+		ddc_pin_map = adls_ddc_pin_map;
+		n_entries = ARRAY_SIZE(adls_ddc_pin_map);
+	} else if (INTEL_PCH_TYPE(i915) >= PCH_DG1) {
+		return vbt_pin;
+	} else if (IS_ROCKETLAKE(i915) && INTEL_PCH_TYPE(i915) == PCH_TGP) {
+		ddc_pin_map = rkl_pch_tgp_ddc_pin_map;
+		n_entries = ARRAY_SIZE(rkl_pch_tgp_ddc_pin_map);
+	} else if (HAS_PCH_TGP(i915) && DISPLAY_VER(i915) == 9) {
+		ddc_pin_map = gen9bc_tgp_ddc_pin_map;
+		n_entries = ARRAY_SIZE(gen9bc_tgp_ddc_pin_map);
+	} else if (INTEL_PCH_TYPE(i915) >= PCH_ICP) {
+		ddc_pin_map = icp_ddc_pin_map;
+		n_entries = ARRAY_SIZE(icp_ddc_pin_map);
+	} else if (HAS_PCH_CNP(i915)) {
+		ddc_pin_map = cnp_ddc_pin_map;
+		n_entries = ARRAY_SIZE(cnp_ddc_pin_map);
+	} else {
+		/* Assuming direct map */
+		return vbt_pin;
+	}
+
+	if (vbt_pin < n_entries && ddc_pin_map[vbt_pin] != 0)
+		return ddc_pin_map[vbt_pin];
+
+	drm_dbg_kms(&i915->drm,
+		    "Ignoring alternate pin: VBT claims DDC pin %d, which is not valid for this platform\n",
+		    vbt_pin);
+	return 0;
+}
+
+static enum port get_port_by_ddc_pin(struct drm_i915_private *i915, u8 ddc_pin)
+{
+	const struct intel_bios_encoder_data *devdata;
+	enum port port;
+
+	if (!ddc_pin)
+		return PORT_NONE;
+
+	for_each_port(port) {
+		devdata = i915->display.vbt.ports[port];
+
+		if (devdata && ddc_pin == devdata->child.ddc_pin)
+			return port;
+	}
+
+	return PORT_NONE;
+}
+
+static void sanitize_ddc_pin(struct intel_bios_encoder_data *devdata,
+			     enum port port)
+{
+	struct drm_i915_private *i915 = devdata->i915;
+	struct child_device_config *child;
+	u8 mapped_ddc_pin;
+	enum port p;
+
+	if (!devdata->child.ddc_pin)
+		return;
+
+	mapped_ddc_pin = map_ddc_pin(i915, devdata->child.ddc_pin);
+	if (!intel_gmbus_is_valid_pin(i915, mapped_ddc_pin)) {
+		drm_dbg_kms(&i915->drm,
+			    "Port %c has invalid DDC pin %d, "
+			    "sticking to defaults\n",
+			    port_name(port), mapped_ddc_pin);
+		devdata->child.ddc_pin = 0;
+		return;
+	}
+
+	p = get_port_by_ddc_pin(i915, devdata->child.ddc_pin);
+	if (p == PORT_NONE)
+		return;
+
+	drm_dbg_kms(&i915->drm,
+		    "port %c trying to use the same DDC pin (0x%x) as port %c, "
+		    "disabling port %c DVI/HDMI support\n",
+		    port_name(port), mapped_ddc_pin,
+		    port_name(p), port_name(p));
+
+	/*
+	 * If we have multiple ports supposedly sharing the pin, then dvi/hdmi
+	 * couldn't exist on the shared port. Otherwise they share the same ddc
+	 * pin and system couldn't communicate with them separately.
+	 *
+	 * Give inverse child device order the priority, last one wins. Yes,
+	 * there are real machines (eg. Asrock B250M-HDV) where VBT has both
+	 * port A and port E with the same AUX ch and we must pick port E :(
+	 */
+	child = &i915->display.vbt.ports[p]->child;
+
+	child->device_type &= ~DEVICE_TYPE_TMDS_DVI_SIGNALING;
+	child->device_type |= DEVICE_TYPE_NOT_HDMI_OUTPUT;
+
+	child->ddc_pin = 0;
+}
+
+static enum port get_port_by_aux_ch(struct drm_i915_private *i915, u8 aux_ch)
+{
+	const struct intel_bios_encoder_data *devdata;
+	enum port port;
+
+	if (!aux_ch)
+		return PORT_NONE;
+
+	for_each_port(port) {
+		devdata = i915->display.vbt.ports[port];
+
+		if (devdata && aux_ch == devdata->child.aux_channel)
+			return port;
+	}
+
+	return PORT_NONE;
+}
+
+static void sanitize_aux_ch(struct intel_bios_encoder_data *devdata,
+			    enum port port)
+{
+	struct drm_i915_private *i915 = devdata->i915;
+	struct child_device_config *child;
+	enum port p;
+
+	p = get_port_by_aux_ch(i915, devdata->child.aux_channel);
+	if (p == PORT_NONE)
+		return;
+
+	drm_dbg_kms(&i915->drm,
+		    "port %c trying to use the same AUX CH (0x%x) as port %c, "
+		    "disabling port %c DP support\n",
+		    port_name(port), devdata->child.aux_channel,
+		    port_name(p), port_name(p));
+
+	/*
+	 * If we have multiple ports supposedly sharing the aux channel, then DP
+	 * couldn't exist on the shared port. Otherwise they share the same aux
+	 * channel and system couldn't communicate with them separately.
+	 *
+	 * Give inverse child device order the priority, last one wins. Yes,
+	 * there are real machines (eg. Asrock B250M-HDV) where VBT has both
+	 * port A and port E with the same AUX ch and we must pick port E :(
+	 */
+	child = &i915->display.vbt.ports[p]->child;
+
+	child->device_type &= ~DEVICE_TYPE_DISPLAYPORT_OUTPUT;
+	child->aux_channel = 0;
+}
+
+static u8 dvo_port_type(u8 dvo_port)
+{
+	switch (dvo_port) {
+	case DVO_PORT_HDMIA:
+	case DVO_PORT_HDMIB:
+	case DVO_PORT_HDMIC:
+	case DVO_PORT_HDMID:
+	case DVO_PORT_HDMIE:
+	case DVO_PORT_HDMIF:
+	case DVO_PORT_HDMIG:
+	case DVO_PORT_HDMIH:
+	case DVO_PORT_HDMII:
+		return DVO_PORT_HDMIA;
+	case DVO_PORT_DPA:
+	case DVO_PORT_DPB:
+	case DVO_PORT_DPC:
+	case DVO_PORT_DPD:
+	case DVO_PORT_DPE:
+	case DVO_PORT_DPF:
+	case DVO_PORT_DPG:
+	case DVO_PORT_DPH:
+	case DVO_PORT_DPI:
+		return DVO_PORT_DPA;
+	case DVO_PORT_MIPIA:
+	case DVO_PORT_MIPIB:
+	case DVO_PORT_MIPIC:
+	case DVO_PORT_MIPID:
+		return DVO_PORT_MIPIA;
+	default:
+		return dvo_port;
+	}
+}
+
+static enum port __dvo_port_to_port(int n_ports, int n_dvo,
+				    const int port_mapping[][3], u8 dvo_port)
+{
+	enum port port;
+	int i;
+
+	for (port = PORT_A; port < n_ports; port++) {
+		for (i = 0; i < n_dvo; i++) {
+			if (port_mapping[port][i] == -1)
+				break;
+
+			if (dvo_port == port_mapping[port][i])
+				return port;
+		}
+	}
+
+	return PORT_NONE;
+}
+
+static enum port dvo_port_to_port(struct drm_i915_private *i915,
+				  u8 dvo_port)
+{
+	/*
+	 * Each DDI port can have more than one value on the "DVO Port" field,
+	 * so look for all the possible values for each port.
+	 */
+	static const int port_mapping[][3] = {
+		[PORT_A] = { DVO_PORT_HDMIA, DVO_PORT_DPA, -1 },
+		[PORT_B] = { DVO_PORT_HDMIB, DVO_PORT_DPB, -1 },
+		[PORT_C] = { DVO_PORT_HDMIC, DVO_PORT_DPC, -1 },
+		[PORT_D] = { DVO_PORT_HDMID, DVO_PORT_DPD, -1 },
+		[PORT_E] = { DVO_PORT_HDMIE, DVO_PORT_DPE, DVO_PORT_CRT },
+		[PORT_F] = { DVO_PORT_HDMIF, DVO_PORT_DPF, -1 },
+		[PORT_G] = { DVO_PORT_HDMIG, DVO_PORT_DPG, -1 },
+		[PORT_H] = { DVO_PORT_HDMIH, DVO_PORT_DPH, -1 },
+		[PORT_I] = { DVO_PORT_HDMII, DVO_PORT_DPI, -1 },
+	};
+	/*
+	 * RKL VBT uses PHY based mapping. Combo PHYs A,B,C,D
+	 * map to DDI A,B,TC1,TC2 respectively.
+	 */
+	static const int rkl_port_mapping[][3] = {
+		[PORT_A] = { DVO_PORT_HDMIA, DVO_PORT_DPA, -1 },
+		[PORT_B] = { DVO_PORT_HDMIB, DVO_PORT_DPB, -1 },
+		[PORT_C] = { -1 },
+		[PORT_TC1] = { DVO_PORT_HDMIC, DVO_PORT_DPC, -1 },
+		[PORT_TC2] = { DVO_PORT_HDMID, DVO_PORT_DPD, -1 },
+	};
+	/*
+	 * Alderlake S ports used in the driver are PORT_A, PORT_D, PORT_E,
+	 * PORT_F and PORT_G, we need to map that to correct VBT sections.
+	 */
+	static const int adls_port_mapping[][3] = {
+		[PORT_A] = { DVO_PORT_HDMIA, DVO_PORT_DPA, -1 },
+		[PORT_B] = { -1 },
+		[PORT_C] = { -1 },
+		[PORT_TC1] = { DVO_PORT_HDMIB, DVO_PORT_DPB, -1 },
+		[PORT_TC2] = { DVO_PORT_HDMIC, DVO_PORT_DPC, -1 },
+		[PORT_TC3] = { DVO_PORT_HDMID, DVO_PORT_DPD, -1 },
+		[PORT_TC4] = { DVO_PORT_HDMIE, DVO_PORT_DPE, -1 },
+	};
+	static const int xelpd_port_mapping[][3] = {
+		[PORT_A] = { DVO_PORT_HDMIA, DVO_PORT_DPA, -1 },
+		[PORT_B] = { DVO_PORT_HDMIB, DVO_PORT_DPB, -1 },
+		[PORT_C] = { DVO_PORT_HDMIC, DVO_PORT_DPC, -1 },
+		[PORT_D_XELPD] = { DVO_PORT_HDMID, DVO_PORT_DPD, -1 },
+		[PORT_E_XELPD] = { DVO_PORT_HDMIE, DVO_PORT_DPE, -1 },
+		[PORT_TC1] = { DVO_PORT_HDMIF, DVO_PORT_DPF, -1 },
+		[PORT_TC2] = { DVO_PORT_HDMIG, DVO_PORT_DPG, -1 },
+		[PORT_TC3] = { DVO_PORT_HDMIH, DVO_PORT_DPH, -1 },
+		[PORT_TC4] = { DVO_PORT_HDMII, DVO_PORT_DPI, -1 },
+	};
+
+	if (DISPLAY_VER(i915) >= 13)
+		return __dvo_port_to_port(ARRAY_SIZE(xelpd_port_mapping),
+					  ARRAY_SIZE(xelpd_port_mapping[0]),
+					  xelpd_port_mapping,
+					  dvo_port);
+	else if (IS_ALDERLAKE_S(i915))
+		return __dvo_port_to_port(ARRAY_SIZE(adls_port_mapping),
+					  ARRAY_SIZE(adls_port_mapping[0]),
+					  adls_port_mapping,
+					  dvo_port);
+	else if (IS_DG1(i915) || IS_ROCKETLAKE(i915))
+		return __dvo_port_to_port(ARRAY_SIZE(rkl_port_mapping),
+					  ARRAY_SIZE(rkl_port_mapping[0]),
+					  rkl_port_mapping,
+					  dvo_port);
+	else
+		return __dvo_port_to_port(ARRAY_SIZE(port_mapping),
+					  ARRAY_SIZE(port_mapping[0]),
+					  port_mapping,
+					  dvo_port);
+}
+
+static enum port
+dsi_dvo_port_to_port(struct drm_i915_private *i915, u8 dvo_port)
+{
+	switch (dvo_port) {
+	case DVO_PORT_MIPIA:
+		return PORT_A;
+	case DVO_PORT_MIPIC:
+		if (DISPLAY_VER(i915) >= 11)
+			return PORT_B;
+		else
+			return PORT_C;
+	default:
+		return PORT_NONE;
+	}
+}
+
+static int parse_bdb_230_dp_max_link_rate(const int vbt_max_link_rate)
+{
+	switch (vbt_max_link_rate) {
+	default:
+	case BDB_230_VBT_DP_MAX_LINK_RATE_DEF:
+		return 0;
+	case BDB_230_VBT_DP_MAX_LINK_RATE_UHBR20:
+		return 2000000;
+	case BDB_230_VBT_DP_MAX_LINK_RATE_UHBR13P5:
+		return 1350000;
+	case BDB_230_VBT_DP_MAX_LINK_RATE_UHBR10:
+		return 1000000;
+	case BDB_230_VBT_DP_MAX_LINK_RATE_HBR3:
+		return 810000;
+	case BDB_230_VBT_DP_MAX_LINK_RATE_HBR2:
+		return 540000;
+	case BDB_230_VBT_DP_MAX_LINK_RATE_HBR:
+		return 270000;
+	case BDB_230_VBT_DP_MAX_LINK_RATE_LBR:
+		return 162000;
+	}
+}
+
+static int parse_bdb_216_dp_max_link_rate(const int vbt_max_link_rate)
+{
+	switch (vbt_max_link_rate) {
+	default:
+	case BDB_216_VBT_DP_MAX_LINK_RATE_HBR3:
+		return 810000;
+	case BDB_216_VBT_DP_MAX_LINK_RATE_HBR2:
+		return 540000;
+	case BDB_216_VBT_DP_MAX_LINK_RATE_HBR:
+		return 270000;
+	case BDB_216_VBT_DP_MAX_LINK_RATE_LBR:
+		return 162000;
+	}
+}
+
+static int _intel_bios_dp_max_link_rate(const struct intel_bios_encoder_data *devdata)
+{
+	if (!devdata || devdata->i915->display.vbt.version < 216)
+		return 0;
+
+	if (devdata->i915->display.vbt.version >= 230)
+		return parse_bdb_230_dp_max_link_rate(devdata->child.dp_max_link_rate);
+	else
+		return parse_bdb_216_dp_max_link_rate(devdata->child.dp_max_link_rate);
+}
+
+static int _intel_bios_dp_max_lane_count(const struct intel_bios_encoder_data *devdata)
+{
+	if (!devdata || devdata->i915->display.vbt.version < 244)
+		return 0;
+
+	return devdata->child.dp_max_lane_count + 1;
+}
+
+static void sanitize_device_type(struct intel_bios_encoder_data *devdata,
+				 enum port port)
+{
+	struct drm_i915_private *i915 = devdata->i915;
+	bool is_hdmi;
+
+	if (port != PORT_A || DISPLAY_VER(i915) >= 12)
+		return;
+
+	if (!intel_bios_encoder_supports_dvi(devdata))
+		return;
+
+	is_hdmi = intel_bios_encoder_supports_hdmi(devdata);
+
+	drm_dbg_kms(&i915->drm, "VBT claims port A supports DVI%s, ignoring\n",
+		    is_hdmi ? "/HDMI" : "");
+
+	devdata->child.device_type &= ~DEVICE_TYPE_TMDS_DVI_SIGNALING;
+	devdata->child.device_type |= DEVICE_TYPE_NOT_HDMI_OUTPUT;
+}
+
+static bool
+intel_bios_encoder_supports_crt(const struct intel_bios_encoder_data *devdata)
+{
+	return devdata->child.device_type & DEVICE_TYPE_ANALOG_OUTPUT;
+}
+
+bool
+intel_bios_encoder_supports_dvi(const struct intel_bios_encoder_data *devdata)
+{
+	return devdata->child.device_type & DEVICE_TYPE_TMDS_DVI_SIGNALING;
+}
+
+bool
+intel_bios_encoder_supports_hdmi(const struct intel_bios_encoder_data *devdata)
+{
+	return intel_bios_encoder_supports_dvi(devdata) &&
+		(devdata->child.device_type & DEVICE_TYPE_NOT_HDMI_OUTPUT) == 0;
+}
+
+bool
+intel_bios_encoder_supports_dp(const struct intel_bios_encoder_data *devdata)
+{
+	return devdata->child.device_type & DEVICE_TYPE_DISPLAYPORT_OUTPUT;
+}
+
+static bool
+intel_bios_encoder_supports_edp(const struct intel_bios_encoder_data *devdata)
+{
+	return intel_bios_encoder_supports_dp(devdata) &&
+		devdata->child.device_type & DEVICE_TYPE_INTERNAL_CONNECTOR;
+}
+
+static bool
+intel_bios_encoder_supports_dsi(const struct intel_bios_encoder_data *devdata)
+{
+	return devdata->child.device_type & DEVICE_TYPE_MIPI_OUTPUT;
+}
+
+static int _intel_bios_hdmi_level_shift(const struct intel_bios_encoder_data *devdata)
+{
+	if (!devdata || devdata->i915->display.vbt.version < 158)
+		return -1;
+
+	return devdata->child.hdmi_level_shifter_value;
+}
+
+static int _intel_bios_max_tmds_clock(const struct intel_bios_encoder_data *devdata)
+{
+	if (!devdata || devdata->i915->display.vbt.version < 204)
+		return 0;
+
+	switch (devdata->child.hdmi_max_data_rate) {
+	default:
+		MISSING_CASE(devdata->child.hdmi_max_data_rate);
+		fallthrough;
+	case HDMI_MAX_DATA_RATE_PLATFORM:
+		return 0;
+	case HDMI_MAX_DATA_RATE_594:
+		return 594000;
+	case HDMI_MAX_DATA_RATE_340:
+		return 340000;
+	case HDMI_MAX_DATA_RATE_300:
+		return 300000;
+	case HDMI_MAX_DATA_RATE_297:
+		return 297000;
+	case HDMI_MAX_DATA_RATE_165:
+		return 165000;
+	}
+}
+
+static bool is_port_valid(struct drm_i915_private *i915, enum port port)
+{
+	/*
+	 * On some ICL SKUs port F is not present, but broken VBTs mark
+	 * the port as present. Only try to initialize port F for the
+	 * SKUs that may actually have it.
+	 */
+	if (port == PORT_F && IS_ICELAKE(i915))
+		return IS_ICL_WITH_PORT_F(i915);
+
+	return true;
+}
+
+static void print_ddi_port(const struct intel_bios_encoder_data *devdata,
+			   enum port port)
+{
+	struct drm_i915_private *i915 = devdata->i915;
+	const struct child_device_config *child = &devdata->child;
+	bool is_dvi, is_hdmi, is_dp, is_edp, is_dsi, is_crt, supports_typec_usb, supports_tbt;
+	int dp_boost_level, dp_max_link_rate, hdmi_boost_level, hdmi_level_shift, max_tmds_clock;
+
+	is_dvi = intel_bios_encoder_supports_dvi(devdata);
+	is_dp = intel_bios_encoder_supports_dp(devdata);
+	is_crt = intel_bios_encoder_supports_crt(devdata);
+	is_hdmi = intel_bios_encoder_supports_hdmi(devdata);
+	is_edp = intel_bios_encoder_supports_edp(devdata);
+	is_dsi = intel_bios_encoder_supports_dsi(devdata);
+
+	supports_typec_usb = intel_bios_encoder_supports_typec_usb(devdata);
+	supports_tbt = intel_bios_encoder_supports_tbt(devdata);
+
+	drm_dbg_kms(&i915->drm,
+		    "Port %c VBT info: CRT:%d DVI:%d HDMI:%d DP:%d eDP:%d DSI:%d LSPCON:%d USB-Type-C:%d TBT:%d DSC:%d\n",
+		    port_name(port), is_crt, is_dvi, is_hdmi, is_dp, is_edp, is_dsi,
+		    HAS_LSPCON(i915) && child->lspcon,
+		    supports_typec_usb, supports_tbt,
+		    devdata->dsc != NULL);
+
+	hdmi_level_shift = _intel_bios_hdmi_level_shift(devdata);
+	if (hdmi_level_shift >= 0) {
+		drm_dbg_kms(&i915->drm,
+			    "Port %c VBT HDMI level shift: %d\n",
+			    port_name(port), hdmi_level_shift);
+	}
+
+	max_tmds_clock = _intel_bios_max_tmds_clock(devdata);
+	if (max_tmds_clock)
+		drm_dbg_kms(&i915->drm,
+			    "Port %c VBT HDMI max TMDS clock: %d kHz\n",
+			    port_name(port), max_tmds_clock);
+
+	/* I_boost config for SKL and above */
+	dp_boost_level = intel_bios_encoder_dp_boost_level(devdata);
+	if (dp_boost_level)
+		drm_dbg_kms(&i915->drm,
+			    "Port %c VBT (e)DP boost level: %d\n",
+			    port_name(port), dp_boost_level);
+
+	hdmi_boost_level = intel_bios_encoder_hdmi_boost_level(devdata);
+	if (hdmi_boost_level)
+		drm_dbg_kms(&i915->drm,
+			    "Port %c VBT HDMI boost level: %d\n",
+			    port_name(port), hdmi_boost_level);
+
+	dp_max_link_rate = _intel_bios_dp_max_link_rate(devdata);
+	if (dp_max_link_rate)
+		drm_dbg_kms(&i915->drm,
+			    "Port %c VBT DP max link rate: %d\n",
+			    port_name(port), dp_max_link_rate);
+}
+
+static void parse_ddi_port(struct intel_bios_encoder_data *devdata)
+{
+	struct drm_i915_private *i915 = devdata->i915;
+	const struct child_device_config *child = &devdata->child;
+	enum port port;
+
+	port = dvo_port_to_port(i915, child->dvo_port);
+	if (port == PORT_NONE && DISPLAY_VER(i915) >= 11)
+		port = dsi_dvo_port_to_port(i915, child->dvo_port);
+	if (port == PORT_NONE)
+		return;
+
+	if (!is_port_valid(i915, port)) {
+		drm_dbg_kms(&i915->drm,
+			    "VBT reports port %c as supported, but that can't be true: skipping\n",
+			    port_name(port));
+		return;
+	}
+
+	if (i915->display.vbt.ports[port]) {
+		drm_dbg_kms(&i915->drm,
+			    "More than one child device for port %c in VBT, using the first.\n",
+			    port_name(port));
+		return;
+	}
+
+	sanitize_device_type(devdata, port);
+
+	if (intel_bios_encoder_supports_dvi(devdata))
+		sanitize_ddc_pin(devdata, port);
+
+	if (intel_bios_encoder_supports_dp(devdata))
+		sanitize_aux_ch(devdata, port);
+
+	i915->display.vbt.ports[port] = devdata;
+}
+
+static bool has_ddi_port_info(struct drm_i915_private *i915)
+{
+	return DISPLAY_VER(i915) >= 5 || IS_G4X(i915);
+}
+
+static void parse_ddi_ports(struct drm_i915_private *i915)
+{
+	struct intel_bios_encoder_data *devdata;
+	enum port port;
+
+	if (!has_ddi_port_info(i915))
+		return;
+
+	list_for_each_entry(devdata, &i915->display.vbt.display_devices, node)
+		parse_ddi_port(devdata);
+
+	for_each_port(port) {
+		if (i915->display.vbt.ports[port])
+			print_ddi_port(i915->display.vbt.ports[port], port);
+	}
+}
+
+static void
+parse_general_definitions(struct drm_i915_private *i915)
+{
+	const struct bdb_general_definitions *defs;
+	struct intel_bios_encoder_data *devdata;
+	const struct child_device_config *child;
+	int i, child_device_num;
+	u8 expected_size;
+	u16 block_size;
+	int bus_pin;
+
+	defs = find_section(i915, BDB_GENERAL_DEFINITIONS);
+	if (!defs) {
+		drm_dbg_kms(&i915->drm,
+			    "No general definition block is found, no devices defined.\n");
+		return;
+	}
+
+	block_size = get_blocksize(defs);
+	if (block_size < sizeof(*defs)) {
+		drm_dbg_kms(&i915->drm,
+			    "General definitions block too small (%u)\n",
+			    block_size);
+		return;
+	}
+
+	bus_pin = defs->crt_ddc_gmbus_pin;
+	drm_dbg_kms(&i915->drm, "crt_ddc_bus_pin: %d\n", bus_pin);
+	if (intel_gmbus_is_valid_pin(i915, bus_pin))
+		i915->display.vbt.crt_ddc_pin = bus_pin;
+
+	if (i915->display.vbt.version < 106) {
+		expected_size = 22;
+	} else if (i915->display.vbt.version < 111) {
+		expected_size = 27;
+	} else if (i915->display.vbt.version < 195) {
+		expected_size = LEGACY_CHILD_DEVICE_CONFIG_SIZE;
+	} else if (i915->display.vbt.version == 195) {
+		expected_size = 37;
+	} else if (i915->display.vbt.version <= 215) {
+		expected_size = 38;
+	} else if (i915->display.vbt.version <= 237) {
+		expected_size = 39;
+	} else {
+		expected_size = sizeof(*child);
+		BUILD_BUG_ON(sizeof(*child) < 39);
+		drm_dbg(&i915->drm,
+			"Expected child device config size for VBT version %u not known; assuming %u\n",
+			i915->display.vbt.version, expected_size);
+	}
+
+	/* Flag an error for unexpected size, but continue anyway. */
+	if (defs->child_dev_size != expected_size)
+		drm_err(&i915->drm,
+			"Unexpected child device config size %u (expected %u for VBT version %u)\n",
+			defs->child_dev_size, expected_size, i915->display.vbt.version);
+
+	/* The legacy sized child device config is the minimum we need. */
+	if (defs->child_dev_size < LEGACY_CHILD_DEVICE_CONFIG_SIZE) {
+		drm_dbg_kms(&i915->drm,
+			    "Child device config size %u is too small.\n",
+			    defs->child_dev_size);
+		return;
+	}
+
+	/* get the number of child device */
+	child_device_num = (block_size - sizeof(*defs)) / defs->child_dev_size;
+
+	for (i = 0; i < child_device_num; i++) {
+		child = child_device_ptr(defs, i);
+		if (!child->device_type)
+			continue;
+
+		drm_dbg_kms(&i915->drm,
+			    "Found VBT child device with type 0x%x\n",
+			    child->device_type);
+
+		devdata = kzalloc(sizeof(*devdata), GFP_KERNEL);
+		if (!devdata)
+			break;
+
+		devdata->i915 = i915;
+
+		/*
+		 * Copy as much as we know (sizeof) and is available
+		 * (child_dev_size) of the child device config. Accessing the
+		 * data must depend on VBT version.
+		 */
+		memcpy(&devdata->child, child,
+		       min_t(size_t, defs->child_dev_size, sizeof(*child)));
+
+		list_add_tail(&devdata->node, &i915->display.vbt.display_devices);
+	}
+
+	if (list_empty(&i915->display.vbt.display_devices))
+		drm_dbg_kms(&i915->drm,
+			    "no child dev is parsed from VBT\n");
+}
+
+/* Common defaults which may be overridden by VBT. */
+static void
+init_vbt_defaults(struct drm_i915_private *i915)
+{
+	i915->display.vbt.crt_ddc_pin = GMBUS_PIN_VGADDC;
+
+	/* general features */
+	i915->display.vbt.int_tv_support = 1;
+	i915->display.vbt.int_crt_support = 1;
+
+	/* driver features */
+	i915->display.vbt.int_lvds_support = 1;
+
+	/* Default to using SSC */
+	i915->display.vbt.lvds_use_ssc = 1;
+	/*
+	 * Core/SandyBridge/IvyBridge use alternative (120MHz) reference
+	 * clock for LVDS.
+	 */
+	i915->display.vbt.lvds_ssc_freq = intel_bios_ssc_frequency(i915,
+								   !HAS_PCH_SPLIT(i915));
+	drm_dbg_kms(&i915->drm, "Set default to SSC at %d kHz\n",
+		    i915->display.vbt.lvds_ssc_freq);
+}
+
+/* Common defaults which may be overridden by VBT. */
+static void
+init_vbt_panel_defaults(struct intel_panel *panel)
+{
+	/* Default to having backlight */
+	panel->vbt.backlight.present = true;
+
+	/* LFP panel data */
+	panel->vbt.lvds_dither = true;
+}
+
+/* Defaults to initialize only if there is no VBT. */
+static void
+init_vbt_missing_defaults(struct drm_i915_private *i915)
+{
+	enum port port;
+	int ports = BIT(PORT_A) | BIT(PORT_B) | BIT(PORT_C) |
+		    BIT(PORT_D) | BIT(PORT_E) | BIT(PORT_F);
+
+	if (!HAS_DDI(i915) && !IS_CHERRYVIEW(i915))
+		return;
+
+	for_each_port_masked(port, ports) {
+		struct intel_bios_encoder_data *devdata;
+		struct child_device_config *child;
+		enum phy phy = intel_port_to_phy(i915, port);
+
+		/*
+		 * VBT has the TypeC mode (native,TBT/USB) and we don't want
+		 * to detect it.
+		 */
+		if (intel_phy_is_tc(i915, phy))
+			continue;
+
+		/* Create fake child device config */
+		devdata = kzalloc(sizeof(*devdata), GFP_KERNEL);
+		if (!devdata)
+			break;
+
+		devdata->i915 = i915;
+		child = &devdata->child;
+
+		if (port == PORT_F)
+			child->dvo_port = DVO_PORT_HDMIF;
+		else if (port == PORT_E)
+			child->dvo_port = DVO_PORT_HDMIE;
+		else
+			child->dvo_port = DVO_PORT_HDMIA + port;
+
+		if (port != PORT_A && port != PORT_E)
+			child->device_type |= DEVICE_TYPE_TMDS_DVI_SIGNALING;
+
+		if (port != PORT_E)
+			child->device_type |= DEVICE_TYPE_DISPLAYPORT_OUTPUT;
+
+		if (port == PORT_A)
+			child->device_type |= DEVICE_TYPE_INTERNAL_CONNECTOR;
+
+		list_add_tail(&devdata->node, &i915->display.vbt.display_devices);
+
+		drm_dbg_kms(&i915->drm,
+			    "Generating default VBT child device with type 0x04%x on port %c\n",
+			    child->device_type, port_name(port));
+	}
+
+	/* Bypass some minimum baseline VBT version checks */
+	i915->display.vbt.version = 155;
+}
+
+static const struct bdb_header *get_bdb_header(const struct vbt_header *vbt)
+{
+	const void *_vbt = vbt;
+
+	return _vbt + vbt->bdb_offset;
+}
+
+/**
+ * intel_bios_is_valid_vbt - does the given buffer contain a valid VBT
+ * @buf:	pointer to a buffer to validate
+ * @size:	size of the buffer
+ *
+ * Returns true on valid VBT.
+ */
+bool intel_bios_is_valid_vbt(const void *buf, size_t size)
+{
+	const struct vbt_header *vbt = buf;
+	const struct bdb_header *bdb;
+
+	if (!vbt)
+		return false;
+
+	if (sizeof(struct vbt_header) > size) {
+		DRM_DEBUG_DRIVER("VBT header incomplete\n");
+		return false;
+	}
+
+	if (memcmp(vbt->signature, "$VBT", 4)) {
+		DRM_DEBUG_DRIVER("VBT invalid signature\n");
+		return false;
+	}
+
+	if (vbt->vbt_size > size) {
+		DRM_DEBUG_DRIVER("VBT incomplete (vbt_size overflows)\n");
+		return false;
+	}
+
+	size = vbt->vbt_size;
+
+	if (range_overflows_t(size_t,
+			      vbt->bdb_offset,
+			      sizeof(struct bdb_header),
+			      size)) {
+		DRM_DEBUG_DRIVER("BDB header incomplete\n");
+		return false;
+	}
+
+	bdb = get_bdb_header(vbt);
+	if (range_overflows_t(size_t, vbt->bdb_offset, bdb->bdb_size, size)) {
+		DRM_DEBUG_DRIVER("BDB incomplete\n");
+		return false;
+	}
+
+	return vbt;
+}
+
+static struct vbt_header *spi_oprom_get_vbt(struct drm_i915_private *i915)
+{
+	u32 count, data, found, store = 0;
+	u32 static_region, oprom_offset;
+	u32 oprom_size = 0x200000;
+	u16 vbt_size;
+	u32 *vbt;
+
+	static_region = intel_uncore_read(&i915->uncore, SPI_STATIC_REGIONS);
+	static_region &= OPTIONROM_SPI_REGIONID_MASK;
+	intel_uncore_write(&i915->uncore, PRIMARY_SPI_REGIONID, static_region);
+
+	oprom_offset = intel_uncore_read(&i915->uncore, OROM_OFFSET);
+	oprom_offset &= OROM_OFFSET_MASK;
+
+	for (count = 0; count < oprom_size; count += 4) {
+		intel_uncore_write(&i915->uncore, PRIMARY_SPI_ADDRESS, oprom_offset + count);
+		data = intel_uncore_read(&i915->uncore, PRIMARY_SPI_TRIGGER);
+
+		if (data == *((const u32 *)"$VBT")) {
+			found = oprom_offset + count;
+			break;
+		}
+	}
+
+	if (count >= oprom_size)
+		goto err_not_found;
+
+	/* Get VBT size and allocate space for the VBT */
+	intel_uncore_write(&i915->uncore, PRIMARY_SPI_ADDRESS, found +
+		   offsetof(struct vbt_header, vbt_size));
+	vbt_size = intel_uncore_read(&i915->uncore, PRIMARY_SPI_TRIGGER);
+	vbt_size &= 0xffff;
+
+	vbt = kzalloc(round_up(vbt_size, 4), GFP_KERNEL);
+	if (!vbt)
+		goto err_not_found;
+
+	for (count = 0; count < vbt_size; count += 4) {
+		intel_uncore_write(&i915->uncore, PRIMARY_SPI_ADDRESS, found + count);
+		data = intel_uncore_read(&i915->uncore, PRIMARY_SPI_TRIGGER);
+		*(vbt + store++) = data;
+	}
+
+	if (!intel_bios_is_valid_vbt(vbt, vbt_size))
+		goto err_free_vbt;
+
+	drm_dbg_kms(&i915->drm, "Found valid VBT in SPI flash\n");
+
+	return (struct vbt_header *)vbt;
+
+err_free_vbt:
+	kfree(vbt);
+err_not_found:
+	return NULL;
+}
+
+static struct vbt_header *oprom_get_vbt(struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	void __iomem *p = NULL, *oprom;
+	struct vbt_header *vbt;
+	u16 vbt_size;
+	size_t i, size;
+
+	oprom = pci_map_rom(pdev, &size);
+	if (!oprom)
+		return NULL;
+
+	/* Scour memory looking for the VBT signature. */
+	for (i = 0; i + 4 < size; i += 4) {
+		if (ioread32(oprom + i) != *((const u32 *)"$VBT"))
+			continue;
+
+		p = oprom + i;
+		size -= i;
+		break;
+	}
+
+	if (!p)
+		goto err_unmap_oprom;
+
+	if (sizeof(struct vbt_header) > size) {
+		drm_dbg(&i915->drm, "VBT header incomplete\n");
+		goto err_unmap_oprom;
+	}
+
+	vbt_size = ioread16(p + offsetof(struct vbt_header, vbt_size));
+	if (vbt_size > size) {
+		drm_dbg(&i915->drm,
+			"VBT incomplete (vbt_size overflows)\n");
+		goto err_unmap_oprom;
+	}
+
+	/* The rest will be validated by intel_bios_is_valid_vbt() */
+	vbt = kmalloc(vbt_size, GFP_KERNEL);
+	if (!vbt)
+		goto err_unmap_oprom;
+
+	memcpy_fromio(vbt, p, vbt_size);
+
+	if (!intel_bios_is_valid_vbt(vbt, vbt_size))
+		goto err_free_vbt;
+
+	pci_unmap_rom(pdev, oprom);
+
+	drm_dbg_kms(&i915->drm, "Found valid VBT in PCI ROM\n");
+
+	return vbt;
+
+err_free_vbt:
+	kfree(vbt);
+err_unmap_oprom:
+	pci_unmap_rom(pdev, oprom);
+
+	return NULL;
+}
+
+/**
+ * intel_bios_init - find VBT and initialize settings from the BIOS
+ * @i915: i915 device instance
+ *
+ * Parse and initialize settings from the Video BIOS Tables (VBT). If the VBT
+ * was not found in ACPI OpRegion, try to find it in PCI ROM first. Also
+ * initialize some defaults if the VBT is not present at all.
+ */
+void intel_bios_init(struct drm_i915_private *i915)
+{
+	const struct vbt_header *vbt = i915->display.opregion.vbt;
+	struct vbt_header *oprom_vbt = NULL;
+	const struct bdb_header *bdb;
+
+	INIT_LIST_HEAD(&i915->display.vbt.display_devices);
+	INIT_LIST_HEAD(&i915->display.vbt.bdb_blocks);
+
+	if (!HAS_DISPLAY(i915)) {
+		drm_dbg_kms(&i915->drm,
+			    "Skipping VBT init due to disabled display.\n");
+		return;
+	}
+
+	init_vbt_defaults(i915);
+
+	/*
+	 * If the OpRegion does not have VBT, look in SPI flash through MMIO or
+	 * PCI mapping
+	 */
+	if (!vbt && IS_DGFX(i915)) {
+		oprom_vbt = spi_oprom_get_vbt(i915);
+		vbt = oprom_vbt;
+	}
+
+	if (!vbt) {
+		oprom_vbt = oprom_get_vbt(i915);
+		vbt = oprom_vbt;
+	}
+
+	if (!vbt)
+		goto out;
+
+	bdb = get_bdb_header(vbt);
+	i915->display.vbt.version = bdb->version;
+
+	drm_dbg_kms(&i915->drm,
+		    "VBT signature \"%.*s\", BDB version %d\n",
+		    (int)sizeof(vbt->signature), vbt->signature, i915->display.vbt.version);
+
+	init_bdb_blocks(i915, bdb);
+
+	/* Grab useful general definitions */
+	parse_general_features(i915);
+	parse_general_definitions(i915);
+	parse_driver_features(i915);
+
+	/* Depends on child device list */
+	parse_compression_parameters(i915);
+
+out:
+	if (!vbt) {
+		drm_info(&i915->drm,
+			 "Failed to find VBIOS tables (VBT)\n");
+		init_vbt_missing_defaults(i915);
+	}
+
+	/* Further processing on pre-parsed or generated child device data */
+	parse_sdvo_device_mapping(i915);
+	parse_ddi_ports(i915);
+
+	kfree(oprom_vbt);
+}
+
+static void intel_bios_init_panel(struct drm_i915_private *i915,
+				  struct intel_panel *panel,
+				  const struct intel_bios_encoder_data *devdata,
+				  const struct edid *edid,
+				  bool use_fallback)
+{
+	/* already have it? */
+	if (panel->vbt.panel_type >= 0) {
+		drm_WARN_ON(&i915->drm, !use_fallback);
+		return;
+	}
+
+	panel->vbt.panel_type = get_panel_type(i915, devdata,
+					       edid, use_fallback);
+	if (panel->vbt.panel_type < 0) {
+		drm_WARN_ON(&i915->drm, use_fallback);
+		return;
+	}
+
+	init_vbt_panel_defaults(panel);
+
+	parse_panel_options(i915, panel);
+	parse_generic_dtd(i915, panel);
+	parse_lfp_data(i915, panel);
+	parse_lfp_backlight(i915, panel);
+	parse_sdvo_panel_data(i915, panel);
+	parse_panel_driver_features(i915, panel);
+	parse_power_conservation_features(i915, panel);
+	parse_edp(i915, panel);
+	parse_psr(i915, panel);
+	parse_mipi_config(i915, panel);
+	parse_mipi_sequence(i915, panel);
+}
+
+void intel_bios_init_panel_early(struct drm_i915_private *i915,
+				 struct intel_panel *panel,
+				 const struct intel_bios_encoder_data *devdata)
+{
+	intel_bios_init_panel(i915, panel, devdata, NULL, false);
+}
+
+void intel_bios_init_panel_late(struct drm_i915_private *i915,
+				struct intel_panel *panel,
+				const struct intel_bios_encoder_data *devdata,
+				const struct edid *edid)
+{
+	intel_bios_init_panel(i915, panel, devdata, edid, true);
+}
+
+/**
+ * intel_bios_driver_remove - Free any resources allocated by intel_bios_init()
+ * @i915: i915 device instance
+ */
+void intel_bios_driver_remove(struct drm_i915_private *i915)
+{
+	struct intel_bios_encoder_data *devdata, *nd;
+	struct bdb_block_entry *entry, *ne;
+
+	list_for_each_entry_safe(devdata, nd, &i915->display.vbt.display_devices, node) {
+		list_del(&devdata->node);
+		kfree(devdata->dsc);
+		kfree(devdata);
+	}
+
+	list_for_each_entry_safe(entry, ne, &i915->display.vbt.bdb_blocks, node) {
+		list_del(&entry->node);
+		kfree(entry);
+	}
+}
+
+void intel_bios_fini_panel(struct intel_panel *panel)
+{
+	kfree(panel->vbt.sdvo_lvds_vbt_mode);
+	panel->vbt.sdvo_lvds_vbt_mode = NULL;
+	kfree(panel->vbt.lfp_lvds_vbt_mode);
+	panel->vbt.lfp_lvds_vbt_mode = NULL;
+	kfree(panel->vbt.dsi.data);
+	panel->vbt.dsi.data = NULL;
+	kfree(panel->vbt.dsi.pps);
+	panel->vbt.dsi.pps = NULL;
+	kfree(panel->vbt.dsi.config);
+	panel->vbt.dsi.config = NULL;
+	kfree(panel->vbt.dsi.deassert_seq);
+	panel->vbt.dsi.deassert_seq = NULL;
+}
+
+/**
+ * intel_bios_is_tv_present - is integrated TV present in VBT
+ * @i915: i915 device instance
+ *
+ * Return true if TV is present. If no child devices were parsed from VBT,
+ * assume TV is present.
+ */
+bool intel_bios_is_tv_present(struct drm_i915_private *i915)
+{
+	const struct intel_bios_encoder_data *devdata;
+	const struct child_device_config *child;
+
+	if (!i915->display.vbt.int_tv_support)
+		return false;
+
+	if (list_empty(&i915->display.vbt.display_devices))
+		return true;
+
+	list_for_each_entry(devdata, &i915->display.vbt.display_devices, node) {
+		child = &devdata->child;
+
+		/*
+		 * If the device type is not TV, continue.
+		 */
+		switch (child->device_type) {
+		case DEVICE_TYPE_INT_TV:
+		case DEVICE_TYPE_TV:
+		case DEVICE_TYPE_TV_SVIDEO_COMPOSITE:
+			break;
+		default:
+			continue;
+		}
+		/* Only when the addin_offset is non-zero, it is regarded
+		 * as present.
+		 */
+		if (child->addin_offset)
+			return true;
+	}
+
+	return false;
+}
+
+/**
+ * intel_bios_is_lvds_present - is LVDS present in VBT
+ * @i915:	i915 device instance
+ * @i2c_pin:	i2c pin for LVDS if present
+ *
+ * Return true if LVDS is present. If no child devices were parsed from VBT,
+ * assume LVDS is present.
+ */
+bool intel_bios_is_lvds_present(struct drm_i915_private *i915, u8 *i2c_pin)
+{
+	const struct intel_bios_encoder_data *devdata;
+	const struct child_device_config *child;
+
+	if (list_empty(&i915->display.vbt.display_devices))
+		return true;
+
+	list_for_each_entry(devdata, &i915->display.vbt.display_devices, node) {
+		child = &devdata->child;
+
+		/* If the device type is not LFP, continue.
+		 * We have to check both the new identifiers as well as the
+		 * old for compatibility with some BIOSes.
+		 */
+		if (child->device_type != DEVICE_TYPE_INT_LFP &&
+		    child->device_type != DEVICE_TYPE_LFP)
+			continue;
+
+		if (intel_gmbus_is_valid_pin(i915, child->i2c_pin))
+			*i2c_pin = child->i2c_pin;
+
+		/* However, we cannot trust the BIOS writers to populate
+		 * the VBT correctly.  Since LVDS requires additional
+		 * information from AIM blocks, a non-zero addin offset is
+		 * a good indicator that the LVDS is actually present.
+		 */
+		if (child->addin_offset)
+			return true;
+
+		/* But even then some BIOS writers perform some black magic
+		 * and instantiate the device without reference to any
+		 * additional data.  Trust that if the VBT was written into
+		 * the OpRegion then they have validated the LVDS's existence.
+		 */
+		if (i915->display.opregion.vbt)
+			return true;
+	}
+
+	return false;
+}
+
+/**
+ * intel_bios_is_port_present - is the specified digital port present
+ * @i915:	i915 device instance
+ * @port:	port to check
+ *
+ * Return true if the device in %port is present.
+ */
+bool intel_bios_is_port_present(struct drm_i915_private *i915, enum port port)
+{
+	if (WARN_ON(!has_ddi_port_info(i915)))
+		return true;
+
+	return i915->display.vbt.ports[port];
+}
+
+/**
+ * intel_bios_is_port_edp - is the device in given port eDP
+ * @i915:	i915 device instance
+ * @port:	port to check
+ *
+ * Return true if the device in %port is eDP.
+ */
+bool intel_bios_is_port_edp(struct drm_i915_private *i915, enum port port)
+{
+	const struct intel_bios_encoder_data *devdata =
+		intel_bios_encoder_data_lookup(i915, port);
+
+	return devdata && intel_bios_encoder_supports_edp(devdata);
+}
+
+static bool intel_bios_encoder_supports_dp_dual_mode(const struct intel_bios_encoder_data *devdata)
+{
+	const struct child_device_config *child = &devdata->child;
+
+	if (!intel_bios_encoder_supports_dp(devdata) ||
+	    !intel_bios_encoder_supports_hdmi(devdata))
+		return false;
+
+	if (dvo_port_type(child->dvo_port) == DVO_PORT_DPA)
+		return true;
+
+	/* Only accept a HDMI dvo_port as DP++ if it has an AUX channel */
+	if (dvo_port_type(child->dvo_port) == DVO_PORT_HDMIA &&
+	    child->aux_channel != 0)
+		return true;
+
+	return false;
+}
+
+bool intel_bios_is_port_dp_dual_mode(struct drm_i915_private *i915,
+				     enum port port)
+{
+	const struct intel_bios_encoder_data *devdata =
+		intel_bios_encoder_data_lookup(i915, port);
+
+	return devdata && intel_bios_encoder_supports_dp_dual_mode(devdata);
+}
+
+/**
+ * intel_bios_is_dsi_present - is DSI present in VBT
+ * @i915:	i915 device instance
+ * @port:	port for DSI if present
+ *
+ * Return true if DSI is present, and return the port in %port.
+ */
+bool intel_bios_is_dsi_present(struct drm_i915_private *i915,
+			       enum port *port)
+{
+	const struct intel_bios_encoder_data *devdata;
+	const struct child_device_config *child;
+	u8 dvo_port;
+
+	list_for_each_entry(devdata, &i915->display.vbt.display_devices, node) {
+		child = &devdata->child;
+
+		if (!(child->device_type & DEVICE_TYPE_MIPI_OUTPUT))
+			continue;
+
+		dvo_port = child->dvo_port;
+
+		if (dsi_dvo_port_to_port(i915, dvo_port) == PORT_NONE) {
+			drm_dbg_kms(&i915->drm,
+				    "VBT has unsupported DSI port %c\n",
+				    port_name(dvo_port - DVO_PORT_MIPIA));
+			continue;
+		}
+
+		if (port)
+			*port = dsi_dvo_port_to_port(i915, dvo_port);
+		return true;
+	}
+
+	return false;
+}
+
+static void fill_dsc(struct intel_crtc_state *crtc_state,
+		     struct dsc_compression_parameters_entry *dsc,
+		     int dsc_max_bpc)
+{
+	struct drm_dsc_config *vdsc_cfg = &crtc_state->dsc.config;
+	int bpc = 8;
+
+	vdsc_cfg->dsc_version_major = dsc->version_major;
+	vdsc_cfg->dsc_version_minor = dsc->version_minor;
+
+	if (dsc->support_12bpc && dsc_max_bpc >= 12)
+		bpc = 12;
+	else if (dsc->support_10bpc && dsc_max_bpc >= 10)
+		bpc = 10;
+	else if (dsc->support_8bpc && dsc_max_bpc >= 8)
+		bpc = 8;
+	else
+		DRM_DEBUG_KMS("VBT: Unsupported BPC %d for DCS\n",
+			      dsc_max_bpc);
+
+	crtc_state->pipe_bpp = bpc * 3;
+
+	crtc_state->dsc.compressed_bpp = min(crtc_state->pipe_bpp,
+					     VBT_DSC_MAX_BPP(dsc->max_bpp));
+
+	/*
+	 * FIXME: This is ugly, and slice count should take DSC engine
+	 * throughput etc. into account.
+	 *
+	 * Also, per spec DSI supports 1, 2, 3 or 4 horizontal slices.
+	 */
+	if (dsc->slices_per_line & BIT(2)) {
+		crtc_state->dsc.slice_count = 4;
+	} else if (dsc->slices_per_line & BIT(1)) {
+		crtc_state->dsc.slice_count = 2;
+	} else {
+		/* FIXME */
+		if (!(dsc->slices_per_line & BIT(0)))
+			DRM_DEBUG_KMS("VBT: Unsupported DSC slice count for DSI\n");
+
+		crtc_state->dsc.slice_count = 1;
+	}
+
+	if (crtc_state->hw.adjusted_mode.crtc_hdisplay %
+	    crtc_state->dsc.slice_count != 0)
+		DRM_DEBUG_KMS("VBT: DSC hdisplay %d not divisible by slice count %d\n",
+			      crtc_state->hw.adjusted_mode.crtc_hdisplay,
+			      crtc_state->dsc.slice_count);
+
+	/*
+	 * The VBT rc_buffer_block_size and rc_buffer_size definitions
+	 * correspond to DP 1.4 DPCD offsets 0x62 and 0x63.
+	 */
+	vdsc_cfg->rc_model_size = drm_dsc_dp_rc_buffer_size(dsc->rc_buffer_block_size,
+							    dsc->rc_buffer_size);
+
+	/* FIXME: DSI spec says bpc + 1 for this one */
+	vdsc_cfg->line_buf_depth = VBT_DSC_LINE_BUFFER_DEPTH(dsc->line_buffer_depth);
+
+	vdsc_cfg->block_pred_enable = dsc->block_prediction_enable;
+
+	vdsc_cfg->slice_height = dsc->slice_height;
+}
+
+/* FIXME: initially DSI specific */
+bool intel_bios_get_dsc_params(struct intel_encoder *encoder,
+			       struct intel_crtc_state *crtc_state,
+			       int dsc_max_bpc)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_bios_encoder_data *devdata;
+	const struct child_device_config *child;
+
+	list_for_each_entry(devdata, &i915->display.vbt.display_devices, node) {
+		child = &devdata->child;
+
+		if (!(child->device_type & DEVICE_TYPE_MIPI_OUTPUT))
+			continue;
+
+		if (dsi_dvo_port_to_port(i915, child->dvo_port) == encoder->port) {
+			if (!devdata->dsc)
+				return false;
+
+			if (crtc_state)
+				fill_dsc(crtc_state, devdata->dsc, dsc_max_bpc);
+
+			return true;
+		}
+	}
+
+	return false;
+}
+
+/**
+ * intel_bios_is_port_hpd_inverted - is HPD inverted for %port
+ * @i915:	i915 device instance
+ * @port:	port to check
+ *
+ * Return true if HPD should be inverted for %port.
+ */
+bool
+intel_bios_is_port_hpd_inverted(const struct drm_i915_private *i915,
+				enum port port)
+{
+	const struct intel_bios_encoder_data *devdata = i915->display.vbt.ports[port];
+
+	if (drm_WARN_ON_ONCE(&i915->drm,
+			     !IS_GEMINILAKE(i915) && !IS_BROXTON(i915)))
+		return false;
+
+	return devdata && devdata->child.hpd_invert;
+}
+
+/**
+ * intel_bios_is_lspcon_present - if LSPCON is attached on %port
+ * @i915:	i915 device instance
+ * @port:	port to check
+ *
+ * Return true if LSPCON is present on this port
+ */
+bool
+intel_bios_is_lspcon_present(const struct drm_i915_private *i915,
+			     enum port port)
+{
+	const struct intel_bios_encoder_data *devdata = i915->display.vbt.ports[port];
+
+	return HAS_LSPCON(i915) && devdata && devdata->child.lspcon;
+}
+
+/**
+ * intel_bios_is_lane_reversal_needed - if lane reversal needed on port
+ * @i915:       i915 device instance
+ * @port:       port to check
+ *
+ * Return true if port requires lane reversal
+ */
+bool
+intel_bios_is_lane_reversal_needed(const struct drm_i915_private *i915,
+				   enum port port)
+{
+	const struct intel_bios_encoder_data *devdata = i915->display.vbt.ports[port];
+
+	return devdata && devdata->child.lane_reversal;
+}
+
+enum aux_ch intel_bios_port_aux_ch(struct drm_i915_private *i915,
+				   enum port port)
+{
+	const struct intel_bios_encoder_data *devdata = i915->display.vbt.ports[port];
+	enum aux_ch aux_ch;
+
+	if (!devdata || !devdata->child.aux_channel) {
+		aux_ch = (enum aux_ch)port;
+
+		drm_dbg_kms(&i915->drm,
+			    "using AUX %c for port %c (platform default)\n",
+			    aux_ch_name(aux_ch), port_name(port));
+		return aux_ch;
+	}
+
+	/*
+	 * RKL/DG1 VBT uses PHY based mapping. Combo PHYs A,B,C,D
+	 * map to DDI A,B,TC1,TC2 respectively.
+	 *
+	 * ADL-S VBT uses PHY based mapping. Combo PHYs A,B,C,D,E
+	 * map to DDI A,TC1,TC2,TC3,TC4 respectively.
+	 */
+	switch (devdata->child.aux_channel) {
+	case DP_AUX_A:
+		aux_ch = AUX_CH_A;
+		break;
+	case DP_AUX_B:
+		if (IS_ALDERLAKE_S(i915))
+			aux_ch = AUX_CH_USBC1;
+		else
+			aux_ch = AUX_CH_B;
+		break;
+	case DP_AUX_C:
+		if (IS_ALDERLAKE_S(i915))
+			aux_ch = AUX_CH_USBC2;
+		else if (IS_DG1(i915) || IS_ROCKETLAKE(i915))
+			aux_ch = AUX_CH_USBC1;
+		else
+			aux_ch = AUX_CH_C;
+		break;
+	case DP_AUX_D:
+		if (DISPLAY_VER(i915) >= 13)
+			aux_ch = AUX_CH_D_XELPD;
+		else if (IS_ALDERLAKE_S(i915))
+			aux_ch = AUX_CH_USBC3;
+		else if (IS_DG1(i915) || IS_ROCKETLAKE(i915))
+			aux_ch = AUX_CH_USBC2;
+		else
+			aux_ch = AUX_CH_D;
+		break;
+	case DP_AUX_E:
+		if (DISPLAY_VER(i915) >= 13)
+			aux_ch = AUX_CH_E_XELPD;
+		else if (IS_ALDERLAKE_S(i915))
+			aux_ch = AUX_CH_USBC4;
+		else
+			aux_ch = AUX_CH_E;
+		break;
+	case DP_AUX_F:
+		if (DISPLAY_VER(i915) >= 13)
+			aux_ch = AUX_CH_USBC1;
+		else
+			aux_ch = AUX_CH_F;
+		break;
+	case DP_AUX_G:
+		if (DISPLAY_VER(i915) >= 13)
+			aux_ch = AUX_CH_USBC2;
+		else
+			aux_ch = AUX_CH_G;
+		break;
+	case DP_AUX_H:
+		if (DISPLAY_VER(i915) >= 13)
+			aux_ch = AUX_CH_USBC3;
+		else
+			aux_ch = AUX_CH_H;
+		break;
+	case DP_AUX_I:
+		if (DISPLAY_VER(i915) >= 13)
+			aux_ch = AUX_CH_USBC4;
+		else
+			aux_ch = AUX_CH_I;
+		break;
+	default:
+		MISSING_CASE(devdata->child.aux_channel);
+		aux_ch = AUX_CH_A;
+		break;
+	}
+
+	drm_dbg_kms(&i915->drm, "using AUX %c for port %c (VBT)\n",
+		    aux_ch_name(aux_ch), port_name(port));
+
+	return aux_ch;
+}
+
+int intel_bios_max_tmds_clock(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_bios_encoder_data *devdata = i915->display.vbt.ports[encoder->port];
+
+	return _intel_bios_max_tmds_clock(devdata);
+}
+
+/* This is an index in the HDMI/DVI DDI buffer translation table, or -1 */
+int intel_bios_hdmi_level_shift(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_bios_encoder_data *devdata = i915->display.vbt.ports[encoder->port];
+
+	return _intel_bios_hdmi_level_shift(devdata);
+}
+
+int intel_bios_encoder_dp_boost_level(const struct intel_bios_encoder_data *devdata)
+{
+	if (!devdata || devdata->i915->display.vbt.version < 196 || !devdata->child.iboost)
+		return 0;
+
+	return translate_iboost(devdata->child.dp_iboost_level);
+}
+
+int intel_bios_encoder_hdmi_boost_level(const struct intel_bios_encoder_data *devdata)
+{
+	if (!devdata || devdata->i915->display.vbt.version < 196 || !devdata->child.iboost)
+		return 0;
+
+	return translate_iboost(devdata->child.hdmi_iboost_level);
+}
+
+int intel_bios_dp_max_link_rate(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_bios_encoder_data *devdata = i915->display.vbt.ports[encoder->port];
+
+	return _intel_bios_dp_max_link_rate(devdata);
+}
+
+int intel_bios_dp_max_lane_count(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_bios_encoder_data *devdata = i915->display.vbt.ports[encoder->port];
+
+	return _intel_bios_dp_max_lane_count(devdata);
+}
+
+int intel_bios_alternate_ddc_pin(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_bios_encoder_data *devdata = i915->display.vbt.ports[encoder->port];
+
+	if (!devdata || !devdata->child.ddc_pin)
+		return 0;
+
+	return map_ddc_pin(i915, devdata->child.ddc_pin);
+}
+
+bool intel_bios_encoder_supports_typec_usb(const struct intel_bios_encoder_data *devdata)
+{
+	return devdata->i915->display.vbt.version >= 195 && devdata->child.dp_usb_type_c;
+}
+
+bool intel_bios_encoder_supports_tbt(const struct intel_bios_encoder_data *devdata)
+{
+	return devdata->i915->display.vbt.version >= 209 && devdata->child.tbt;
+}
+
+const struct intel_bios_encoder_data *
+intel_bios_encoder_data_lookup(struct drm_i915_private *i915, enum port port)
+{
+	return i915->display.vbt.ports[port];
+}
diff --git a/drivers/gpu/drm/i915/display/intel_bios.h b/drivers/gpu/drm/i915/display/intel_bios.h
new file mode 100644
index 000000000..ff1fdd2e0
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_bios.h
@@ -0,0 +1,280 @@
+/*
+ * Copyright © 2016-2019 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+/*
+ * Please use intel_vbt_defs.h for VBT private data, to hide and abstract away
+ * the VBT from the rest of the driver. Add the parsed, clean data to struct
+ * intel_vbt_data within struct drm_i915_private.
+ */
+
+#ifndef _INTEL_BIOS_H_
+#define _INTEL_BIOS_H_
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct edid;
+struct intel_bios_encoder_data;
+struct intel_crtc_state;
+struct intel_encoder;
+struct intel_panel;
+enum port;
+
+enum intel_backlight_type {
+	INTEL_BACKLIGHT_PMIC,
+	INTEL_BACKLIGHT_LPSS,
+	INTEL_BACKLIGHT_DISPLAY_DDI,
+	INTEL_BACKLIGHT_DSI_DCS,
+	INTEL_BACKLIGHT_PANEL_DRIVER_INTERFACE,
+	INTEL_BACKLIGHT_VESA_EDP_AUX_INTERFACE,
+};
+
+struct edp_power_seq {
+	u16 t1_t3;
+	u16 t8;
+	u16 t9;
+	u16 t10;
+	u16 t11_t12;
+} __packed;
+
+/*
+ * MIPI Sequence Block definitions
+ *
+ * Note the VBT spec has AssertReset / DeassertReset swapped from their
+ * usual naming, we use the proper names here to avoid confusion when
+ * reading the code.
+ */
+enum mipi_seq {
+	MIPI_SEQ_END = 0,
+	MIPI_SEQ_DEASSERT_RESET,	/* Spec says MipiAssertResetPin */
+	MIPI_SEQ_INIT_OTP,
+	MIPI_SEQ_DISPLAY_ON,
+	MIPI_SEQ_DISPLAY_OFF,
+	MIPI_SEQ_ASSERT_RESET,		/* Spec says MipiDeassertResetPin */
+	MIPI_SEQ_BACKLIGHT_ON,		/* sequence block v2+ */
+	MIPI_SEQ_BACKLIGHT_OFF,		/* sequence block v2+ */
+	MIPI_SEQ_TEAR_ON,		/* sequence block v2+ */
+	MIPI_SEQ_TEAR_OFF,		/* sequence block v3+ */
+	MIPI_SEQ_POWER_ON,		/* sequence block v3+ */
+	MIPI_SEQ_POWER_OFF,		/* sequence block v3+ */
+	MIPI_SEQ_MAX
+};
+
+enum mipi_seq_element {
+	MIPI_SEQ_ELEM_END = 0,
+	MIPI_SEQ_ELEM_SEND_PKT,
+	MIPI_SEQ_ELEM_DELAY,
+	MIPI_SEQ_ELEM_GPIO,
+	MIPI_SEQ_ELEM_I2C,		/* sequence block v2+ */
+	MIPI_SEQ_ELEM_SPI,		/* sequence block v3+ */
+	MIPI_SEQ_ELEM_PMIC,		/* sequence block v3+ */
+	MIPI_SEQ_ELEM_MAX
+};
+
+#define MIPI_DSI_UNDEFINED_PANEL_ID	0
+#define MIPI_DSI_GENERIC_PANEL_ID	1
+
+struct mipi_config {
+	u16 panel_id;
+
+	/* General Params */
+	u32 enable_dithering:1;
+	u32 rsvd1:1;
+	u32 is_bridge:1;
+
+	u32 panel_arch_type:2;
+	u32 is_cmd_mode:1;
+
+#define NON_BURST_SYNC_PULSE	0x1
+#define NON_BURST_SYNC_EVENTS	0x2
+#define BURST_MODE		0x3
+	u32 video_transfer_mode:2;
+
+	u32 cabc_supported:1;
+#define PPS_BLC_PMIC   0
+#define PPS_BLC_SOC    1
+	u32 pwm_blc:1;
+
+	/* Bit 13:10 */
+#define PIXEL_FORMAT_RGB565			0x1
+#define PIXEL_FORMAT_RGB666			0x2
+#define PIXEL_FORMAT_RGB666_LOOSELY_PACKED	0x3
+#define PIXEL_FORMAT_RGB888			0x4
+	u32 videomode_color_format:4;
+
+	/* Bit 15:14 */
+#define ENABLE_ROTATION_0	0x0
+#define ENABLE_ROTATION_90	0x1
+#define ENABLE_ROTATION_180	0x2
+#define ENABLE_ROTATION_270	0x3
+	u32 rotation:2;
+	u32 bta_enabled:1;
+	u32 rsvd2:15;
+
+	/* 2 byte Port Description */
+#define DUAL_LINK_NOT_SUPPORTED	0
+#define DUAL_LINK_FRONT_BACK	1
+#define DUAL_LINK_PIXEL_ALT	2
+	u16 dual_link:2;
+	u16 lane_cnt:2;
+	u16 pixel_overlap:3;
+	u16 rgb_flip:1;
+#define DL_DCS_PORT_A			0x00
+#define DL_DCS_PORT_C			0x01
+#define DL_DCS_PORT_A_AND_C		0x02
+	u16 dl_dcs_cabc_ports:2;
+	u16 dl_dcs_backlight_ports:2;
+	u16 rsvd3:4;
+
+	u16 rsvd4;
+
+	u8 rsvd5;
+	u32 target_burst_mode_freq;
+	u32 dsi_ddr_clk;
+	u32 bridge_ref_clk;
+
+#define  BYTE_CLK_SEL_20MHZ		0
+#define  BYTE_CLK_SEL_10MHZ		1
+#define  BYTE_CLK_SEL_5MHZ		2
+	u8 byte_clk_sel:2;
+
+	u8 rsvd6:6;
+
+	/* DPHY Flags */
+	u16 dphy_param_valid:1;
+	u16 eot_pkt_disabled:1;
+	u16 enable_clk_stop:1;
+	u16 rsvd7:13;
+
+	u32 hs_tx_timeout;
+	u32 lp_rx_timeout;
+	u32 turn_around_timeout;
+	u32 device_reset_timer;
+	u32 master_init_timer;
+	u32 dbi_bw_timer;
+	u32 lp_byte_clk_val;
+
+	/*  4 byte Dphy Params */
+	u32 prepare_cnt:6;
+	u32 rsvd8:2;
+	u32 clk_zero_cnt:8;
+	u32 trail_cnt:5;
+	u32 rsvd9:3;
+	u32 exit_zero_cnt:6;
+	u32 rsvd10:2;
+
+	u32 clk_lane_switch_cnt;
+	u32 hl_switch_cnt;
+
+	u32 rsvd11[6];
+
+	/* timings based on dphy spec */
+	u8 tclk_miss;
+	u8 tclk_post;
+	u8 rsvd12;
+	u8 tclk_pre;
+	u8 tclk_prepare;
+	u8 tclk_settle;
+	u8 tclk_term_enable;
+	u8 tclk_trail;
+	u16 tclk_prepare_clkzero;
+	u8 rsvd13;
+	u8 td_term_enable;
+	u8 teot;
+	u8 ths_exit;
+	u8 ths_prepare;
+	u16 ths_prepare_hszero;
+	u8 rsvd14;
+	u8 ths_settle;
+	u8 ths_skip;
+	u8 ths_trail;
+	u8 tinit;
+	u8 tlpx;
+	u8 rsvd15[3];
+
+	/* GPIOs */
+	u8 panel_enable;
+	u8 bl_enable;
+	u8 pwm_enable;
+	u8 reset_r_n;
+	u8 pwr_down_r;
+	u8 stdby_r_n;
+
+} __packed;
+
+/* all delays have a unit of 100us */
+struct mipi_pps_data {
+	u16 panel_on_delay;
+	u16 bl_enable_delay;
+	u16 bl_disable_delay;
+	u16 panel_off_delay;
+	u16 panel_power_cycle_delay;
+} __packed;
+
+void intel_bios_init(struct drm_i915_private *dev_priv);
+void intel_bios_init_panel_early(struct drm_i915_private *dev_priv,
+				 struct intel_panel *panel,
+				 const struct intel_bios_encoder_data *devdata);
+void intel_bios_init_panel_late(struct drm_i915_private *dev_priv,
+				struct intel_panel *panel,
+				const struct intel_bios_encoder_data *devdata,
+				const struct edid *edid);
+void intel_bios_fini_panel(struct intel_panel *panel);
+void intel_bios_driver_remove(struct drm_i915_private *dev_priv);
+bool intel_bios_is_valid_vbt(const void *buf, size_t size);
+bool intel_bios_is_tv_present(struct drm_i915_private *dev_priv);
+bool intel_bios_is_lvds_present(struct drm_i915_private *dev_priv, u8 *i2c_pin);
+bool intel_bios_is_port_present(struct drm_i915_private *dev_priv, enum port port);
+bool intel_bios_is_port_edp(struct drm_i915_private *dev_priv, enum port port);
+bool intel_bios_is_port_dp_dual_mode(struct drm_i915_private *dev_priv, enum port port);
+bool intel_bios_is_dsi_present(struct drm_i915_private *dev_priv, enum port *port);
+bool intel_bios_is_port_hpd_inverted(const struct drm_i915_private *i915,
+				     enum port port);
+bool intel_bios_is_lspcon_present(const struct drm_i915_private *i915,
+				  enum port port);
+bool intel_bios_is_lane_reversal_needed(const struct drm_i915_private *i915,
+					enum port port);
+enum aux_ch intel_bios_port_aux_ch(struct drm_i915_private *dev_priv, enum port port);
+bool intel_bios_get_dsc_params(struct intel_encoder *encoder,
+			       struct intel_crtc_state *crtc_state,
+			       int dsc_max_bpc);
+int intel_bios_max_tmds_clock(struct intel_encoder *encoder);
+int intel_bios_hdmi_level_shift(struct intel_encoder *encoder);
+int intel_bios_dp_max_link_rate(struct intel_encoder *encoder);
+int intel_bios_dp_max_lane_count(struct intel_encoder *encoder);
+int intel_bios_alternate_ddc_pin(struct intel_encoder *encoder);
+bool intel_bios_port_supports_typec_usb(struct drm_i915_private *i915, enum port port);
+bool intel_bios_port_supports_tbt(struct drm_i915_private *i915, enum port port);
+
+const struct intel_bios_encoder_data *
+intel_bios_encoder_data_lookup(struct drm_i915_private *i915, enum port port);
+
+bool intel_bios_encoder_supports_dvi(const struct intel_bios_encoder_data *devdata);
+bool intel_bios_encoder_supports_hdmi(const struct intel_bios_encoder_data *devdata);
+bool intel_bios_encoder_supports_dp(const struct intel_bios_encoder_data *devdata);
+bool intel_bios_encoder_supports_typec_usb(const struct intel_bios_encoder_data *devdata);
+bool intel_bios_encoder_supports_tbt(const struct intel_bios_encoder_data *devdata);
+int intel_bios_encoder_dp_boost_level(const struct intel_bios_encoder_data *devdata);
+int intel_bios_encoder_hdmi_boost_level(const struct intel_bios_encoder_data *devdata);
+
+#endif /* _INTEL_BIOS_H_ */
diff --git a/drivers/gpu/drm/i915/display/intel_bw.c b/drivers/gpu/drm/i915/display/intel_bw.c
new file mode 100644
index 000000000..4ace026b2
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_bw.c
@@ -0,0 +1,1203 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <drm/drm_atomic_state_helper.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "i915_utils.h"
+#include "intel_atomic.h"
+#include "intel_bw.h"
+#include "intel_cdclk.h"
+#include "intel_display_core.h"
+#include "intel_display_types.h"
+#include "skl_watermark.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pcode.h"
+
+/* Parameters for Qclk Geyserville (QGV) */
+struct intel_qgv_point {
+	u16 dclk, t_rp, t_rdpre, t_rc, t_ras, t_rcd;
+};
+
+struct intel_psf_gv_point {
+	u8 clk; /* clock in multiples of 16.6666 MHz */
+};
+
+struct intel_qgv_info {
+	struct intel_qgv_point points[I915_NUM_QGV_POINTS];
+	struct intel_psf_gv_point psf_points[I915_NUM_PSF_GV_POINTS];
+	u8 num_points;
+	u8 num_psf_points;
+	u8 t_bl;
+	u8 max_numchannels;
+	u8 channel_width;
+	u8 deinterleave;
+};
+
+static int dg1_mchbar_read_qgv_point_info(struct drm_i915_private *dev_priv,
+					  struct intel_qgv_point *sp,
+					  int point)
+{
+	u32 dclk_ratio, dclk_reference;
+	u32 val;
+
+	val = intel_uncore_read(&dev_priv->uncore, SA_PERF_STATUS_0_0_0_MCHBAR_PC);
+	dclk_ratio = REG_FIELD_GET(DG1_QCLK_RATIO_MASK, val);
+	if (val & DG1_QCLK_REFERENCE)
+		dclk_reference = 6; /* 6 * 16.666 MHz = 100 MHz */
+	else
+		dclk_reference = 8; /* 8 * 16.666 MHz = 133 MHz */
+	sp->dclk = DIV_ROUND_UP((16667 * dclk_ratio * dclk_reference) + 500, 1000);
+
+	val = intel_uncore_read(&dev_priv->uncore, SKL_MC_BIOS_DATA_0_0_0_MCHBAR_PCU);
+	if (val & DG1_GEAR_TYPE)
+		sp->dclk *= 2;
+
+	if (sp->dclk == 0)
+		return -EINVAL;
+
+	val = intel_uncore_read(&dev_priv->uncore, MCHBAR_CH0_CR_TC_PRE_0_0_0_MCHBAR);
+	sp->t_rp = REG_FIELD_GET(DG1_DRAM_T_RP_MASK, val);
+	sp->t_rdpre = REG_FIELD_GET(DG1_DRAM_T_RDPRE_MASK, val);
+
+	val = intel_uncore_read(&dev_priv->uncore, MCHBAR_CH0_CR_TC_PRE_0_0_0_MCHBAR_HIGH);
+	sp->t_rcd = REG_FIELD_GET(DG1_DRAM_T_RCD_MASK, val);
+	sp->t_ras = REG_FIELD_GET(DG1_DRAM_T_RAS_MASK, val);
+
+	sp->t_rc = sp->t_rp + sp->t_ras;
+
+	return 0;
+}
+
+static int icl_pcode_read_qgv_point_info(struct drm_i915_private *dev_priv,
+					 struct intel_qgv_point *sp,
+					 int point)
+{
+	u32 val = 0, val2 = 0;
+	u16 dclk;
+	int ret;
+
+	ret = snb_pcode_read(&dev_priv->uncore, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
+			     ICL_PCODE_MEM_SS_READ_QGV_POINT_INFO(point),
+			     &val, &val2);
+	if (ret)
+		return ret;
+
+	dclk = val & 0xffff;
+	sp->dclk = DIV_ROUND_UP((16667 * dclk) + (DISPLAY_VER(dev_priv) > 11 ? 500 : 0), 1000);
+	sp->t_rp = (val & 0xff0000) >> 16;
+	sp->t_rcd = (val & 0xff000000) >> 24;
+
+	sp->t_rdpre = val2 & 0xff;
+	sp->t_ras = (val2 & 0xff00) >> 8;
+
+	sp->t_rc = sp->t_rp + sp->t_ras;
+
+	return 0;
+}
+
+static int adls_pcode_read_psf_gv_point_info(struct drm_i915_private *dev_priv,
+					    struct intel_psf_gv_point *points)
+{
+	u32 val = 0;
+	int ret;
+	int i;
+
+	ret = snb_pcode_read(&dev_priv->uncore, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
+			     ADL_PCODE_MEM_SS_READ_PSF_GV_INFO, &val, NULL);
+	if (ret)
+		return ret;
+
+	for (i = 0; i < I915_NUM_PSF_GV_POINTS; i++) {
+		points[i].clk = val & 0xff;
+		val >>= 8;
+	}
+
+	return 0;
+}
+
+int icl_pcode_restrict_qgv_points(struct drm_i915_private *dev_priv,
+				  u32 points_mask)
+{
+	int ret;
+
+	/* bspec says to keep retrying for at least 1 ms */
+	ret = skl_pcode_request(&dev_priv->uncore, ICL_PCODE_SAGV_DE_MEM_SS_CONFIG,
+				points_mask,
+				ICL_PCODE_REP_QGV_MASK | ADLS_PCODE_REP_PSF_MASK,
+				ICL_PCODE_REP_QGV_SAFE | ADLS_PCODE_REP_PSF_SAFE,
+				1);
+
+	if (ret < 0) {
+		drm_err(&dev_priv->drm, "Failed to disable qgv points (%d) points: 0x%x\n", ret, points_mask);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int mtl_read_qgv_point_info(struct drm_i915_private *dev_priv,
+				   struct intel_qgv_point *sp, int point)
+{
+	u32 val, val2;
+	u16 dclk;
+
+	val = intel_uncore_read(&dev_priv->uncore,
+				MTL_MEM_SS_INFO_QGV_POINT_LOW(point));
+	val2 = intel_uncore_read(&dev_priv->uncore,
+				 MTL_MEM_SS_INFO_QGV_POINT_HIGH(point));
+	dclk = REG_FIELD_GET(MTL_DCLK_MASK, val);
+	sp->dclk = DIV_ROUND_UP((16667 * dclk), 1000);
+	sp->t_rp = REG_FIELD_GET(MTL_TRP_MASK, val);
+	sp->t_rcd = REG_FIELD_GET(MTL_TRCD_MASK, val);
+
+	sp->t_rdpre = REG_FIELD_GET(MTL_TRDPRE_MASK, val2);
+	sp->t_ras = REG_FIELD_GET(MTL_TRAS_MASK, val2);
+
+	sp->t_rc = sp->t_rp + sp->t_ras;
+
+	return 0;
+}
+
+static int
+intel_read_qgv_point_info(struct drm_i915_private *dev_priv,
+			  struct intel_qgv_point *sp,
+			  int point)
+{
+	if (DISPLAY_VER(dev_priv) >= 14)
+		return mtl_read_qgv_point_info(dev_priv, sp, point);
+	else if (IS_DG1(dev_priv))
+		return dg1_mchbar_read_qgv_point_info(dev_priv, sp, point);
+	else
+		return icl_pcode_read_qgv_point_info(dev_priv, sp, point);
+}
+
+static int icl_get_qgv_points(struct drm_i915_private *dev_priv,
+			      struct intel_qgv_info *qi,
+			      bool is_y_tile)
+{
+	const struct dram_info *dram_info = &dev_priv->dram_info;
+	int i, ret;
+
+	qi->num_points = dram_info->num_qgv_points;
+	qi->num_psf_points = dram_info->num_psf_gv_points;
+
+	if (DISPLAY_VER(dev_priv) >= 14) {
+		switch (dram_info->type) {
+		case INTEL_DRAM_DDR4:
+			qi->t_bl = 4;
+			qi->max_numchannels = 2;
+			qi->channel_width = 64;
+			qi->deinterleave = 2;
+			break;
+		case INTEL_DRAM_DDR5:
+			qi->t_bl = 8;
+			qi->max_numchannels = 4;
+			qi->channel_width = 32;
+			qi->deinterleave = 2;
+			break;
+		case INTEL_DRAM_LPDDR4:
+		case INTEL_DRAM_LPDDR5:
+			qi->t_bl = 16;
+			qi->max_numchannels = 8;
+			qi->channel_width = 16;
+			qi->deinterleave = 4;
+			break;
+		default:
+			MISSING_CASE(dram_info->type);
+			return -EINVAL;
+		}
+	} else if (DISPLAY_VER(dev_priv) >= 12) {
+		switch (dram_info->type) {
+		case INTEL_DRAM_DDR4:
+			qi->t_bl = is_y_tile ? 8 : 4;
+			qi->max_numchannels = 2;
+			qi->channel_width = 64;
+			qi->deinterleave = is_y_tile ? 1 : 2;
+			break;
+		case INTEL_DRAM_DDR5:
+			qi->t_bl = is_y_tile ? 16 : 8;
+			qi->max_numchannels = 4;
+			qi->channel_width = 32;
+			qi->deinterleave = is_y_tile ? 1 : 2;
+			break;
+		case INTEL_DRAM_LPDDR4:
+			if (IS_ROCKETLAKE(dev_priv)) {
+				qi->t_bl = 8;
+				qi->max_numchannels = 4;
+				qi->channel_width = 32;
+				qi->deinterleave = 2;
+				break;
+			}
+			fallthrough;
+		case INTEL_DRAM_LPDDR5:
+			qi->t_bl = 16;
+			qi->max_numchannels = 8;
+			qi->channel_width = 16;
+			qi->deinterleave = is_y_tile ? 2 : 4;
+			break;
+		default:
+			qi->t_bl = 16;
+			qi->max_numchannels = 1;
+			break;
+		}
+	} else if (DISPLAY_VER(dev_priv) == 11) {
+		qi->t_bl = dev_priv->dram_info.type == INTEL_DRAM_DDR4 ? 4 : 8;
+		qi->max_numchannels = 1;
+	}
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			qi->num_points > ARRAY_SIZE(qi->points)))
+		qi->num_points = ARRAY_SIZE(qi->points);
+
+	for (i = 0; i < qi->num_points; i++) {
+		struct intel_qgv_point *sp = &qi->points[i];
+
+		ret = intel_read_qgv_point_info(dev_priv, sp, i);
+		if (ret)
+			return ret;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "QGV %d: DCLK=%d tRP=%d tRDPRE=%d tRAS=%d tRCD=%d tRC=%d\n",
+			    i, sp->dclk, sp->t_rp, sp->t_rdpre, sp->t_ras,
+			    sp->t_rcd, sp->t_rc);
+	}
+
+	if (qi->num_psf_points > 0) {
+		ret = adls_pcode_read_psf_gv_point_info(dev_priv, qi->psf_points);
+		if (ret) {
+			drm_err(&dev_priv->drm, "Failed to read PSF point data; PSF points will not be considered in bandwidth calculations.\n");
+			qi->num_psf_points = 0;
+		}
+
+		for (i = 0; i < qi->num_psf_points; i++)
+			drm_dbg_kms(&dev_priv->drm,
+				    "PSF GV %d: CLK=%d \n",
+				    i, qi->psf_points[i].clk);
+	}
+
+	return 0;
+}
+
+static int adl_calc_psf_bw(int clk)
+{
+	/*
+	 * clk is multiples of 16.666MHz (100/6)
+	 * According to BSpec PSF GV bandwidth is
+	 * calculated as BW = 64 * clk * 16.666Mhz
+	 */
+	return DIV_ROUND_CLOSEST(64 * clk * 100, 6);
+}
+
+static int icl_sagv_max_dclk(const struct intel_qgv_info *qi)
+{
+	u16 dclk = 0;
+	int i;
+
+	for (i = 0; i < qi->num_points; i++)
+		dclk = max(dclk, qi->points[i].dclk);
+
+	return dclk;
+}
+
+struct intel_sa_info {
+	u16 displayrtids;
+	u8 deburst, deprogbwlimit, derating;
+};
+
+static const struct intel_sa_info icl_sa_info = {
+	.deburst = 8,
+	.deprogbwlimit = 25, /* GB/s */
+	.displayrtids = 128,
+	.derating = 10,
+};
+
+static const struct intel_sa_info tgl_sa_info = {
+	.deburst = 16,
+	.deprogbwlimit = 34, /* GB/s */
+	.displayrtids = 256,
+	.derating = 10,
+};
+
+static const struct intel_sa_info rkl_sa_info = {
+	.deburst = 8,
+	.deprogbwlimit = 20, /* GB/s */
+	.displayrtids = 128,
+	.derating = 10,
+};
+
+static const struct intel_sa_info adls_sa_info = {
+	.deburst = 16,
+	.deprogbwlimit = 38, /* GB/s */
+	.displayrtids = 256,
+	.derating = 10,
+};
+
+static const struct intel_sa_info adlp_sa_info = {
+	.deburst = 16,
+	.deprogbwlimit = 38, /* GB/s */
+	.displayrtids = 256,
+	.derating = 20,
+};
+
+static const struct intel_sa_info mtl_sa_info = {
+	.deburst = 32,
+	.deprogbwlimit = 38, /* GB/s */
+	.displayrtids = 256,
+	.derating = 20,
+};
+
+static int icl_get_bw_info(struct drm_i915_private *dev_priv, const struct intel_sa_info *sa)
+{
+	struct intel_qgv_info qi = {};
+	bool is_y_tile = true; /* assume y tile may be used */
+	int num_channels = max_t(u8, 1, dev_priv->dram_info.num_channels);
+	int ipqdepth, ipqdepthpch = 16;
+	int dclk_max;
+	int maxdebw;
+	int num_groups = ARRAY_SIZE(dev_priv->display.bw.max);
+	int i, ret;
+
+	ret = icl_get_qgv_points(dev_priv, &qi, is_y_tile);
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Failed to get memory subsystem information, ignoring bandwidth limits");
+		return ret;
+	}
+
+	dclk_max = icl_sagv_max_dclk(&qi);
+	maxdebw = min(sa->deprogbwlimit * 1000, dclk_max * 16 * 6 / 10);
+	ipqdepth = min(ipqdepthpch, sa->displayrtids / num_channels);
+	qi.deinterleave = DIV_ROUND_UP(num_channels, is_y_tile ? 4 : 2);
+
+	for (i = 0; i < num_groups; i++) {
+		struct intel_bw_info *bi = &dev_priv->display.bw.max[i];
+		int clpchgroup;
+		int j;
+
+		clpchgroup = (sa->deburst * qi.deinterleave / num_channels) << i;
+		bi->num_planes = (ipqdepth - clpchgroup) / clpchgroup + 1;
+
+		bi->num_qgv_points = qi.num_points;
+		bi->num_psf_gv_points = qi.num_psf_points;
+
+		for (j = 0; j < qi.num_points; j++) {
+			const struct intel_qgv_point *sp = &qi.points[j];
+			int ct, bw;
+
+			/*
+			 * Max row cycle time
+			 *
+			 * FIXME what is the logic behind the
+			 * assumed burst length?
+			 */
+			ct = max_t(int, sp->t_rc, sp->t_rp + sp->t_rcd +
+				   (clpchgroup - 1) * qi.t_bl + sp->t_rdpre);
+			bw = DIV_ROUND_UP(sp->dclk * clpchgroup * 32 * num_channels, ct);
+
+			bi->deratedbw[j] = min(maxdebw,
+					       bw * (100 - sa->derating) / 100);
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "BW%d / QGV %d: num_planes=%d deratedbw=%u\n",
+				    i, j, bi->num_planes, bi->deratedbw[j]);
+		}
+	}
+	/*
+	 * In case if SAGV is disabled in BIOS, we always get 1
+	 * SAGV point, but we can't send PCode commands to restrict it
+	 * as it will fail and pointless anyway.
+	 */
+	if (qi.num_points == 1)
+		dev_priv->display.sagv.status = I915_SAGV_NOT_CONTROLLED;
+	else
+		dev_priv->display.sagv.status = I915_SAGV_ENABLED;
+
+	return 0;
+}
+
+static int tgl_get_bw_info(struct drm_i915_private *dev_priv, const struct intel_sa_info *sa)
+{
+	struct intel_qgv_info qi = {};
+	const struct dram_info *dram_info = &dev_priv->dram_info;
+	bool is_y_tile = true; /* assume y tile may be used */
+	int num_channels = max_t(u8, 1, dev_priv->dram_info.num_channels);
+	int ipqdepth, ipqdepthpch = 16;
+	int dclk_max;
+	int maxdebw, peakbw;
+	int clperchgroup;
+	int num_groups = ARRAY_SIZE(dev_priv->display.bw.max);
+	int i, ret;
+
+	ret = icl_get_qgv_points(dev_priv, &qi, is_y_tile);
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Failed to get memory subsystem information, ignoring bandwidth limits");
+		return ret;
+	}
+
+	if (dram_info->type == INTEL_DRAM_LPDDR4 || dram_info->type == INTEL_DRAM_LPDDR5)
+		num_channels *= 2;
+
+	qi.deinterleave = qi.deinterleave ? : DIV_ROUND_UP(num_channels, is_y_tile ? 4 : 2);
+
+	if (num_channels < qi.max_numchannels && DISPLAY_VER(dev_priv) >= 12)
+		qi.deinterleave = max(DIV_ROUND_UP(qi.deinterleave, 2), 1);
+
+	if (DISPLAY_VER(dev_priv) > 11 && num_channels > qi.max_numchannels)
+		drm_warn(&dev_priv->drm, "Number of channels exceeds max number of channels.");
+	if (qi.max_numchannels != 0)
+		num_channels = min_t(u8, num_channels, qi.max_numchannels);
+
+	dclk_max = icl_sagv_max_dclk(&qi);
+
+	peakbw = num_channels * DIV_ROUND_UP(qi.channel_width, 8) * dclk_max;
+	maxdebw = min(sa->deprogbwlimit * 1000, peakbw * 6 / 10); /* 60% */
+
+	ipqdepth = min(ipqdepthpch, sa->displayrtids / num_channels);
+	/*
+	 * clperchgroup = 4kpagespermempage * clperchperblock,
+	 * clperchperblock = 8 / num_channels * interleave
+	 */
+	clperchgroup = 4 * DIV_ROUND_UP(8, num_channels) * qi.deinterleave;
+
+	for (i = 0; i < num_groups; i++) {
+		struct intel_bw_info *bi = &dev_priv->display.bw.max[i];
+		struct intel_bw_info *bi_next;
+		int clpchgroup;
+		int j;
+
+		clpchgroup = (sa->deburst * qi.deinterleave / num_channels) << i;
+
+		if (i < num_groups - 1) {
+			bi_next = &dev_priv->display.bw.max[i + 1];
+
+			if (clpchgroup < clperchgroup)
+				bi_next->num_planes = (ipqdepth - clpchgroup) /
+						       clpchgroup + 1;
+			else
+				bi_next->num_planes = 0;
+		}
+
+		bi->num_qgv_points = qi.num_points;
+		bi->num_psf_gv_points = qi.num_psf_points;
+
+		for (j = 0; j < qi.num_points; j++) {
+			const struct intel_qgv_point *sp = &qi.points[j];
+			int ct, bw;
+
+			/*
+			 * Max row cycle time
+			 *
+			 * FIXME what is the logic behind the
+			 * assumed burst length?
+			 */
+			ct = max_t(int, sp->t_rc, sp->t_rp + sp->t_rcd +
+				   (clpchgroup - 1) * qi.t_bl + sp->t_rdpre);
+			bw = DIV_ROUND_UP(sp->dclk * clpchgroup * 32 * num_channels, ct);
+
+			bi->deratedbw[j] = min(maxdebw,
+					       bw * (100 - sa->derating) / 100);
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "BW%d / QGV %d: num_planes=%d deratedbw=%u\n",
+				    i, j, bi->num_planes, bi->deratedbw[j]);
+		}
+
+		for (j = 0; j < qi.num_psf_points; j++) {
+			const struct intel_psf_gv_point *sp = &qi.psf_points[j];
+
+			bi->psf_bw[j] = adl_calc_psf_bw(sp->clk);
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "BW%d / PSF GV %d: num_planes=%d bw=%u\n",
+				    i, j, bi->num_planes, bi->psf_bw[j]);
+		}
+	}
+
+	/*
+	 * In case if SAGV is disabled in BIOS, we always get 1
+	 * SAGV point, but we can't send PCode commands to restrict it
+	 * as it will fail and pointless anyway.
+	 */
+	if (qi.num_points == 1)
+		dev_priv->display.sagv.status = I915_SAGV_NOT_CONTROLLED;
+	else
+		dev_priv->display.sagv.status = I915_SAGV_ENABLED;
+
+	return 0;
+}
+
+static void dg2_get_bw_info(struct drm_i915_private *i915)
+{
+	unsigned int deratedbw = IS_DG2_G11(i915) ? 38000 : 50000;
+	int num_groups = ARRAY_SIZE(i915->display.bw.max);
+	int i;
+
+	/*
+	 * DG2 doesn't have SAGV or QGV points, just a constant max bandwidth
+	 * that doesn't depend on the number of planes enabled. So fill all the
+	 * plane group with constant bw information for uniformity with other
+	 * platforms. DG2-G10 platforms have a constant 50 GB/s bandwidth,
+	 * whereas DG2-G11 platforms have 38 GB/s.
+	 */
+	for (i = 0; i < num_groups; i++) {
+		struct intel_bw_info *bi = &i915->display.bw.max[i];
+
+		bi->num_planes = 1;
+		/* Need only one dummy QGV point per group */
+		bi->num_qgv_points = 1;
+		bi->deratedbw[0] = deratedbw;
+	}
+
+	i915->display.sagv.status = I915_SAGV_NOT_CONTROLLED;
+}
+
+static unsigned int icl_max_bw(struct drm_i915_private *dev_priv,
+			       int num_planes, int qgv_point)
+{
+	int i;
+
+	/*
+	 * Let's return max bw for 0 planes
+	 */
+	num_planes = max(1, num_planes);
+
+	for (i = 0; i < ARRAY_SIZE(dev_priv->display.bw.max); i++) {
+		const struct intel_bw_info *bi =
+			&dev_priv->display.bw.max[i];
+
+		/*
+		 * Pcode will not expose all QGV points when
+		 * SAGV is forced to off/min/med/max.
+		 */
+		if (qgv_point >= bi->num_qgv_points)
+			return UINT_MAX;
+
+		if (num_planes >= bi->num_planes)
+			return bi->deratedbw[qgv_point];
+	}
+
+	return 0;
+}
+
+static unsigned int tgl_max_bw(struct drm_i915_private *dev_priv,
+			       int num_planes, int qgv_point)
+{
+	int i;
+
+	/*
+	 * Let's return max bw for 0 planes
+	 */
+	num_planes = max(1, num_planes);
+
+	for (i = ARRAY_SIZE(dev_priv->display.bw.max) - 1; i >= 0; i--) {
+		const struct intel_bw_info *bi =
+			&dev_priv->display.bw.max[i];
+
+		/*
+		 * Pcode will not expose all QGV points when
+		 * SAGV is forced to off/min/med/max.
+		 */
+		if (qgv_point >= bi->num_qgv_points)
+			return UINT_MAX;
+
+		if (num_planes <= bi->num_planes)
+			return bi->deratedbw[qgv_point];
+	}
+
+	return dev_priv->display.bw.max[0].deratedbw[qgv_point];
+}
+
+static unsigned int adl_psf_bw(struct drm_i915_private *dev_priv,
+			       int psf_gv_point)
+{
+	const struct intel_bw_info *bi =
+			&dev_priv->display.bw.max[0];
+
+	return bi->psf_bw[psf_gv_point];
+}
+
+void intel_bw_init_hw(struct drm_i915_private *dev_priv)
+{
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	if (DISPLAY_VER(dev_priv) >= 14)
+		tgl_get_bw_info(dev_priv, &mtl_sa_info);
+	else if (IS_DG2(dev_priv))
+		dg2_get_bw_info(dev_priv);
+	else if (IS_ALDERLAKE_P(dev_priv))
+		tgl_get_bw_info(dev_priv, &adlp_sa_info);
+	else if (IS_ALDERLAKE_S(dev_priv))
+		tgl_get_bw_info(dev_priv, &adls_sa_info);
+	else if (IS_ROCKETLAKE(dev_priv))
+		tgl_get_bw_info(dev_priv, &rkl_sa_info);
+	else if (DISPLAY_VER(dev_priv) == 12)
+		tgl_get_bw_info(dev_priv, &tgl_sa_info);
+	else if (DISPLAY_VER(dev_priv) == 11)
+		icl_get_bw_info(dev_priv, &icl_sa_info);
+}
+
+static unsigned int intel_bw_crtc_num_active_planes(const struct intel_crtc_state *crtc_state)
+{
+	/*
+	 * We assume cursors are small enough
+	 * to not not cause bandwidth problems.
+	 */
+	return hweight8(crtc_state->active_planes & ~BIT(PLANE_CURSOR));
+}
+
+static unsigned int intel_bw_crtc_data_rate(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	unsigned int data_rate = 0;
+	enum plane_id plane_id;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		/*
+		 * We assume cursors are small enough
+		 * to not not cause bandwidth problems.
+		 */
+		if (plane_id == PLANE_CURSOR)
+			continue;
+
+		data_rate += crtc_state->data_rate[plane_id];
+
+		if (DISPLAY_VER(i915) < 11)
+			data_rate += crtc_state->data_rate_y[plane_id];
+	}
+
+	return data_rate;
+}
+
+/* "Maximum Pipe Read Bandwidth" */
+static int intel_bw_crtc_min_cdclk(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	if (DISPLAY_VER(i915) < 12)
+		return 0;
+
+	return DIV_ROUND_UP_ULL(mul_u32_u32(intel_bw_crtc_data_rate(crtc_state), 10), 512);
+}
+
+void intel_bw_crtc_update(struct intel_bw_state *bw_state,
+			  const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	bw_state->data_rate[crtc->pipe] =
+		intel_bw_crtc_data_rate(crtc_state);
+	bw_state->num_active_planes[crtc->pipe] =
+		intel_bw_crtc_num_active_planes(crtc_state);
+
+	drm_dbg_kms(&i915->drm, "pipe %c data rate %u num active planes %u\n",
+		    pipe_name(crtc->pipe),
+		    bw_state->data_rate[crtc->pipe],
+		    bw_state->num_active_planes[crtc->pipe]);
+}
+
+static unsigned int intel_bw_num_active_planes(struct drm_i915_private *dev_priv,
+					       const struct intel_bw_state *bw_state)
+{
+	unsigned int num_active_planes = 0;
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe)
+		num_active_planes += bw_state->num_active_planes[pipe];
+
+	return num_active_planes;
+}
+
+static unsigned int intel_bw_data_rate(struct drm_i915_private *dev_priv,
+				       const struct intel_bw_state *bw_state)
+{
+	unsigned int data_rate = 0;
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe)
+		data_rate += bw_state->data_rate[pipe];
+
+	if (DISPLAY_VER(dev_priv) >= 13 && i915_vtd_active(dev_priv))
+		data_rate = DIV_ROUND_UP(data_rate * 105, 100);
+
+	return data_rate;
+}
+
+struct intel_bw_state *
+intel_atomic_get_old_bw_state(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_global_state *bw_state;
+
+	bw_state = intel_atomic_get_old_global_obj_state(state, &dev_priv->display.bw.obj);
+
+	return to_intel_bw_state(bw_state);
+}
+
+struct intel_bw_state *
+intel_atomic_get_new_bw_state(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_global_state *bw_state;
+
+	bw_state = intel_atomic_get_new_global_obj_state(state, &dev_priv->display.bw.obj);
+
+	return to_intel_bw_state(bw_state);
+}
+
+struct intel_bw_state *
+intel_atomic_get_bw_state(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_global_state *bw_state;
+
+	bw_state = intel_atomic_get_global_obj_state(state, &dev_priv->display.bw.obj);
+	if (IS_ERR(bw_state))
+		return ERR_CAST(bw_state);
+
+	return to_intel_bw_state(bw_state);
+}
+
+static bool intel_bw_state_changed(struct drm_i915_private *i915,
+				   const struct intel_bw_state *old_bw_state,
+				   const struct intel_bw_state *new_bw_state)
+{
+	enum pipe pipe;
+
+	for_each_pipe(i915, pipe) {
+		const struct intel_dbuf_bw *old_crtc_bw =
+			&old_bw_state->dbuf_bw[pipe];
+		const struct intel_dbuf_bw *new_crtc_bw =
+			&new_bw_state->dbuf_bw[pipe];
+		enum dbuf_slice slice;
+
+		for_each_dbuf_slice(i915, slice) {
+			if (old_crtc_bw->max_bw[slice] != new_crtc_bw->max_bw[slice] ||
+			    old_crtc_bw->active_planes[slice] != new_crtc_bw->active_planes[slice])
+				return true;
+		}
+
+		if (old_bw_state->min_cdclk[pipe] != new_bw_state->min_cdclk[pipe])
+			return true;
+	}
+
+	return false;
+}
+
+static void skl_plane_calc_dbuf_bw(struct intel_bw_state *bw_state,
+				   struct intel_crtc *crtc,
+				   enum plane_id plane_id,
+				   const struct skl_ddb_entry *ddb,
+				   unsigned int data_rate)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_dbuf_bw *crtc_bw = &bw_state->dbuf_bw[crtc->pipe];
+	unsigned int dbuf_mask = skl_ddb_dbuf_slice_mask(i915, ddb);
+	enum dbuf_slice slice;
+
+	/*
+	 * The arbiter can only really guarantee an
+	 * equal share of the total bw to each plane.
+	 */
+	for_each_dbuf_slice_in_mask(i915, slice, dbuf_mask) {
+		crtc_bw->max_bw[slice] = max(crtc_bw->max_bw[slice], data_rate);
+		crtc_bw->active_planes[slice] |= BIT(plane_id);
+	}
+}
+
+static void skl_crtc_calc_dbuf_bw(struct intel_bw_state *bw_state,
+				  const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_dbuf_bw *crtc_bw = &bw_state->dbuf_bw[crtc->pipe];
+	enum plane_id plane_id;
+
+	memset(crtc_bw, 0, sizeof(*crtc_bw));
+
+	if (!crtc_state->hw.active)
+		return;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		/*
+		 * We assume cursors are small enough
+		 * to not cause bandwidth problems.
+		 */
+		if (plane_id == PLANE_CURSOR)
+			continue;
+
+		skl_plane_calc_dbuf_bw(bw_state, crtc, plane_id,
+				       &crtc_state->wm.skl.plane_ddb[plane_id],
+				       crtc_state->data_rate[plane_id]);
+
+		if (DISPLAY_VER(i915) < 11)
+			skl_plane_calc_dbuf_bw(bw_state, crtc, plane_id,
+					       &crtc_state->wm.skl.plane_ddb_y[plane_id],
+					       crtc_state->data_rate[plane_id]);
+	}
+}
+
+/* "Maximum Data Buffer Bandwidth" */
+static int
+intel_bw_dbuf_min_cdclk(struct drm_i915_private *i915,
+			const struct intel_bw_state *bw_state)
+{
+	unsigned int total_max_bw = 0;
+	enum dbuf_slice slice;
+
+	for_each_dbuf_slice(i915, slice) {
+		int num_active_planes = 0;
+		unsigned int max_bw = 0;
+		enum pipe pipe;
+
+		/*
+		 * The arbiter can only really guarantee an
+		 * equal share of the total bw to each plane.
+		 */
+		for_each_pipe(i915, pipe) {
+			const struct intel_dbuf_bw *crtc_bw = &bw_state->dbuf_bw[pipe];
+
+			max_bw = max(crtc_bw->max_bw[slice], max_bw);
+			num_active_planes += hweight8(crtc_bw->active_planes[slice]);
+		}
+		max_bw *= num_active_planes;
+
+		total_max_bw = max(total_max_bw, max_bw);
+	}
+
+	return DIV_ROUND_UP(total_max_bw, 64);
+}
+
+int intel_bw_min_cdclk(struct drm_i915_private *i915,
+		       const struct intel_bw_state *bw_state)
+{
+	enum pipe pipe;
+	int min_cdclk;
+
+	min_cdclk = intel_bw_dbuf_min_cdclk(i915, bw_state);
+
+	for_each_pipe(i915, pipe)
+		min_cdclk = max(bw_state->min_cdclk[pipe], min_cdclk);
+
+	return min_cdclk;
+}
+
+int intel_bw_calc_min_cdclk(struct intel_atomic_state *state,
+			    bool *need_cdclk_calc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_bw_state *new_bw_state = NULL;
+	const struct intel_bw_state *old_bw_state = NULL;
+	const struct intel_cdclk_state *cdclk_state;
+	const struct intel_crtc_state *crtc_state;
+	int old_min_cdclk, new_min_cdclk;
+	struct intel_crtc *crtc;
+	int i;
+
+	if (DISPLAY_VER(dev_priv) < 9)
+		return 0;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		new_bw_state = intel_atomic_get_bw_state(state);
+		if (IS_ERR(new_bw_state))
+			return PTR_ERR(new_bw_state);
+
+		old_bw_state = intel_atomic_get_old_bw_state(state);
+
+		skl_crtc_calc_dbuf_bw(new_bw_state, crtc_state);
+
+		new_bw_state->min_cdclk[crtc->pipe] =
+			intel_bw_crtc_min_cdclk(crtc_state);
+	}
+
+	if (!old_bw_state)
+		return 0;
+
+	if (intel_bw_state_changed(dev_priv, old_bw_state, new_bw_state)) {
+		int ret = intel_atomic_lock_global_state(&new_bw_state->base);
+		if (ret)
+			return ret;
+	}
+
+	old_min_cdclk = intel_bw_min_cdclk(dev_priv, old_bw_state);
+	new_min_cdclk = intel_bw_min_cdclk(dev_priv, new_bw_state);
+
+	/*
+	 * No need to check against the cdclk state if
+	 * the min cdclk doesn't increase.
+	 *
+	 * Ie. we only ever increase the cdclk due to bandwidth
+	 * requirements. This can reduce back and forth
+	 * display blinking due to constant cdclk changes.
+	 */
+	if (new_min_cdclk <= old_min_cdclk)
+		return 0;
+
+	cdclk_state = intel_atomic_get_cdclk_state(state);
+	if (IS_ERR(cdclk_state))
+		return PTR_ERR(cdclk_state);
+
+	/*
+	 * No need to recalculate the cdclk state if
+	 * the min cdclk doesn't increase.
+	 *
+	 * Ie. we only ever increase the cdclk due to bandwidth
+	 * requirements. This can reduce back and forth
+	 * display blinking due to constant cdclk changes.
+	 */
+	if (new_min_cdclk <= cdclk_state->bw_min_cdclk)
+		return 0;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "new bandwidth min cdclk (%d kHz) > old min cdclk (%d kHz)\n",
+		    new_min_cdclk, cdclk_state->bw_min_cdclk);
+	*need_cdclk_calc = true;
+
+	return 0;
+}
+
+static u16 icl_qgv_points_mask(struct drm_i915_private *i915)
+{
+	unsigned int num_psf_gv_points = i915->display.bw.max[0].num_psf_gv_points;
+	unsigned int num_qgv_points = i915->display.bw.max[0].num_qgv_points;
+	u16 qgv_points = 0, psf_points = 0;
+
+	/*
+	 * We can _not_ use the whole ADLS_QGV_PT_MASK here, as PCode rejects
+	 * it with failure if we try masking any unadvertised points.
+	 * So need to operate only with those returned from PCode.
+	 */
+	if (num_qgv_points > 0)
+		qgv_points = GENMASK(num_qgv_points - 1, 0);
+
+	if (num_psf_gv_points > 0)
+		psf_points = GENMASK(num_psf_gv_points - 1, 0);
+
+	return ICL_PCODE_REQ_QGV_PT(qgv_points) | ADLS_PCODE_REQ_PSF_PT(psf_points);
+}
+
+static int intel_bw_check_data_rate(struct intel_atomic_state *state, bool *changed)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *new_crtc_state, *old_crtc_state;
+	struct intel_crtc *crtc;
+	int i;
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		unsigned int old_data_rate =
+			intel_bw_crtc_data_rate(old_crtc_state);
+		unsigned int new_data_rate =
+			intel_bw_crtc_data_rate(new_crtc_state);
+		unsigned int old_active_planes =
+			intel_bw_crtc_num_active_planes(old_crtc_state);
+		unsigned int new_active_planes =
+			intel_bw_crtc_num_active_planes(new_crtc_state);
+		struct intel_bw_state *new_bw_state;
+
+		/*
+		 * Avoid locking the bw state when
+		 * nothing significant has changed.
+		 */
+		if (old_data_rate == new_data_rate &&
+		    old_active_planes == new_active_planes)
+			continue;
+
+		new_bw_state = intel_atomic_get_bw_state(state);
+		if (IS_ERR(new_bw_state))
+			return PTR_ERR(new_bw_state);
+
+		new_bw_state->data_rate[crtc->pipe] = new_data_rate;
+		new_bw_state->num_active_planes[crtc->pipe] = new_active_planes;
+
+		*changed = true;
+
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] data rate %u num active planes %u\n",
+			    crtc->base.base.id, crtc->base.name,
+			    new_bw_state->data_rate[crtc->pipe],
+			    new_bw_state->num_active_planes[crtc->pipe]);
+	}
+
+	return 0;
+}
+
+int intel_bw_atomic_check(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_bw_state *old_bw_state;
+	struct intel_bw_state *new_bw_state;
+	unsigned int data_rate;
+	unsigned int num_active_planes;
+	int i, ret;
+	u16 qgv_points = 0, psf_points = 0;
+	unsigned int max_bw_point = 0, max_bw = 0;
+	unsigned int num_qgv_points = dev_priv->display.bw.max[0].num_qgv_points;
+	unsigned int num_psf_gv_points = dev_priv->display.bw.max[0].num_psf_gv_points;
+	bool changed = false;
+
+	/* FIXME earlier gens need some checks too */
+	if (DISPLAY_VER(dev_priv) < 11)
+		return 0;
+
+	ret = intel_bw_check_data_rate(state, &changed);
+	if (ret)
+		return ret;
+
+	old_bw_state = intel_atomic_get_old_bw_state(state);
+	new_bw_state = intel_atomic_get_new_bw_state(state);
+
+	if (new_bw_state &&
+	    intel_can_enable_sagv(dev_priv, old_bw_state) !=
+	    intel_can_enable_sagv(dev_priv, new_bw_state))
+		changed = true;
+
+	/*
+	 * If none of our inputs (data rates, number of active
+	 * planes, SAGV yes/no) changed then nothing to do here.
+	 */
+	if (!changed)
+		return 0;
+
+	ret = intel_atomic_lock_global_state(&new_bw_state->base);
+	if (ret)
+		return ret;
+
+	data_rate = intel_bw_data_rate(dev_priv, new_bw_state);
+	data_rate = DIV_ROUND_UP(data_rate, 1000);
+
+	num_active_planes = intel_bw_num_active_planes(dev_priv, new_bw_state);
+
+	for (i = 0; i < num_qgv_points; i++) {
+		unsigned int max_data_rate;
+
+		if (DISPLAY_VER(dev_priv) > 11)
+			max_data_rate = tgl_max_bw(dev_priv, num_active_planes, i);
+		else
+			max_data_rate = icl_max_bw(dev_priv, num_active_planes, i);
+		/*
+		 * We need to know which qgv point gives us
+		 * maximum bandwidth in order to disable SAGV
+		 * if we find that we exceed SAGV block time
+		 * with watermarks. By that moment we already
+		 * have those, as it is calculated earlier in
+		 * intel_atomic_check,
+		 */
+		if (max_data_rate > max_bw) {
+			max_bw_point = i;
+			max_bw = max_data_rate;
+		}
+		if (max_data_rate >= data_rate)
+			qgv_points |= BIT(i);
+
+		drm_dbg_kms(&dev_priv->drm, "QGV point %d: max bw %d required %d\n",
+			    i, max_data_rate, data_rate);
+	}
+
+	for (i = 0; i < num_psf_gv_points; i++) {
+		unsigned int max_data_rate = adl_psf_bw(dev_priv, i);
+
+		if (max_data_rate >= data_rate)
+			psf_points |= BIT(i);
+
+		drm_dbg_kms(&dev_priv->drm, "PSF GV point %d: max bw %d"
+			    " required %d\n",
+			    i, max_data_rate, data_rate);
+	}
+
+	/*
+	 * BSpec states that we always should have at least one allowed point
+	 * left, so if we couldn't - simply reject the configuration for obvious
+	 * reasons.
+	 */
+	if (qgv_points == 0) {
+		drm_dbg_kms(&dev_priv->drm, "No QGV points provide sufficient memory"
+			    " bandwidth %d for display configuration(%d active planes).\n",
+			    data_rate, num_active_planes);
+		return -EINVAL;
+	}
+
+	if (num_psf_gv_points > 0 && psf_points == 0) {
+		drm_dbg_kms(&dev_priv->drm, "No PSF GV points provide sufficient memory"
+			    " bandwidth %d for display configuration(%d active planes).\n",
+			    data_rate, num_active_planes);
+		return -EINVAL;
+	}
+
+	/*
+	 * Leave only single point with highest bandwidth, if
+	 * we can't enable SAGV due to the increased memory latency it may
+	 * cause.
+	 */
+	if (!intel_can_enable_sagv(dev_priv, new_bw_state)) {
+		qgv_points = BIT(max_bw_point);
+		drm_dbg_kms(&dev_priv->drm, "No SAGV, using single QGV point %d\n",
+			    max_bw_point);
+	}
+
+	/*
+	 * We store the ones which need to be masked as that is what PCode
+	 * actually accepts as a parameter.
+	 */
+	new_bw_state->qgv_points_mask =
+		~(ICL_PCODE_REQ_QGV_PT(qgv_points) |
+		  ADLS_PCODE_REQ_PSF_PT(psf_points)) &
+		icl_qgv_points_mask(dev_priv);
+
+	/*
+	 * If the actual mask had changed we need to make sure that
+	 * the commits are serialized(in case this is a nomodeset, nonblocking)
+	 */
+	if (new_bw_state->qgv_points_mask != old_bw_state->qgv_points_mask) {
+		ret = intel_atomic_serialize_global_state(&new_bw_state->base);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static struct intel_global_state *
+intel_bw_duplicate_state(struct intel_global_obj *obj)
+{
+	struct intel_bw_state *state;
+
+	state = kmemdup(obj->state, sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return NULL;
+
+	return &state->base;
+}
+
+static void intel_bw_destroy_state(struct intel_global_obj *obj,
+				   struct intel_global_state *state)
+{
+	kfree(state);
+}
+
+static const struct intel_global_state_funcs intel_bw_funcs = {
+	.atomic_duplicate_state = intel_bw_duplicate_state,
+	.atomic_destroy_state = intel_bw_destroy_state,
+};
+
+int intel_bw_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_bw_state *state;
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return -ENOMEM;
+
+	intel_atomic_global_obj_init(dev_priv, &dev_priv->display.bw.obj,
+				     &state->base, &intel_bw_funcs);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_bw.h b/drivers/gpu/drm/i915/display/intel_bw.h
new file mode 100644
index 000000000..cb7ee3a24
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_bw.h
@@ -0,0 +1,72 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_BW_H__
+#define __INTEL_BW_H__
+
+#include <drm/drm_atomic.h>
+
+#include "intel_display.h"
+#include "intel_display_power.h"
+#include "intel_global_state.h"
+
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc_state;
+
+struct intel_dbuf_bw {
+	unsigned int max_bw[I915_MAX_DBUF_SLICES];
+	u8 active_planes[I915_MAX_DBUF_SLICES];
+};
+
+struct intel_bw_state {
+	struct intel_global_state base;
+	struct intel_dbuf_bw dbuf_bw[I915_MAX_PIPES];
+
+	/*
+	 * Contains a bit mask, used to determine, whether correspondent
+	 * pipe allows SAGV or not.
+	 */
+	u8 pipe_sagv_reject;
+
+	/* bitmask of active pipes */
+	u8 active_pipes;
+
+	/*
+	 * Current QGV points mask, which restricts
+	 * some particular SAGV states, not to confuse
+	 * with pipe_sagv_mask.
+	 */
+	u16 qgv_points_mask;
+
+	int min_cdclk[I915_MAX_PIPES];
+	unsigned int data_rate[I915_MAX_PIPES];
+	u8 num_active_planes[I915_MAX_PIPES];
+};
+
+#define to_intel_bw_state(x) container_of((x), struct intel_bw_state, base)
+
+struct intel_bw_state *
+intel_atomic_get_old_bw_state(struct intel_atomic_state *state);
+
+struct intel_bw_state *
+intel_atomic_get_new_bw_state(struct intel_atomic_state *state);
+
+struct intel_bw_state *
+intel_atomic_get_bw_state(struct intel_atomic_state *state);
+
+void intel_bw_init_hw(struct drm_i915_private *dev_priv);
+int intel_bw_init(struct drm_i915_private *dev_priv);
+int intel_bw_atomic_check(struct intel_atomic_state *state);
+void intel_bw_crtc_update(struct intel_bw_state *bw_state,
+			  const struct intel_crtc_state *crtc_state);
+int icl_pcode_restrict_qgv_points(struct drm_i915_private *dev_priv,
+				  u32 points_mask);
+int intel_bw_calc_min_cdclk(struct intel_atomic_state *state,
+			    bool *need_cdclk_calc);
+int intel_bw_min_cdclk(struct drm_i915_private *i915,
+		       const struct intel_bw_state *bw_state);
+
+#endif /* __INTEL_BW_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_cdclk.c b/drivers/gpu/drm/i915/display/intel_cdclk.c
new file mode 100644
index 000000000..25dcdde5f
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_cdclk.c
@@ -0,0 +1,3283 @@
+/*
+ * Copyright © 2006-2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/time.h>
+
+#include "hsw_ips.h"
+#include "intel_atomic.h"
+#include "intel_atomic_plane.h"
+#include "intel_audio.h"
+#include "intel_bw.h"
+#include "intel_cdclk.h"
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pci_config.h"
+#include "intel_pcode.h"
+#include "intel_psr.h"
+#include "vlv_sideband.h"
+
+/**
+ * DOC: CDCLK / RAWCLK
+ *
+ * The display engine uses several different clocks to do its work. There
+ * are two main clocks involved that aren't directly related to the actual
+ * pixel clock or any symbol/bit clock of the actual output port. These
+ * are the core display clock (CDCLK) and RAWCLK.
+ *
+ * CDCLK clocks most of the display pipe logic, and thus its frequency
+ * must be high enough to support the rate at which pixels are flowing
+ * through the pipes. Downscaling must also be accounted as that increases
+ * the effective pixel rate.
+ *
+ * On several platforms the CDCLK frequency can be changed dynamically
+ * to minimize power consumption for a given display configuration.
+ * Typically changes to the CDCLK frequency require all the display pipes
+ * to be shut down while the frequency is being changed.
+ *
+ * On SKL+ the DMC will toggle the CDCLK off/on during DC5/6 entry/exit.
+ * DMC will not change the active CDCLK frequency however, so that part
+ * will still be performed by the driver directly.
+ *
+ * RAWCLK is a fixed frequency clock, often used by various auxiliary
+ * blocks such as AUX CH or backlight PWM. Hence the only thing we
+ * really need to know about RAWCLK is its frequency so that various
+ * dividers can be programmed correctly.
+ */
+
+struct intel_cdclk_funcs {
+	void (*get_cdclk)(struct drm_i915_private *i915,
+			  struct intel_cdclk_config *cdclk_config);
+	void (*set_cdclk)(struct drm_i915_private *i915,
+			  const struct intel_cdclk_config *cdclk_config,
+			  enum pipe pipe);
+	int (*modeset_calc_cdclk)(struct intel_cdclk_state *state);
+	u8 (*calc_voltage_level)(int cdclk);
+};
+
+void intel_cdclk_get_cdclk(struct drm_i915_private *dev_priv,
+			   struct intel_cdclk_config *cdclk_config)
+{
+	dev_priv->display.funcs.cdclk->get_cdclk(dev_priv, cdclk_config);
+}
+
+static void intel_cdclk_set_cdclk(struct drm_i915_private *dev_priv,
+				  const struct intel_cdclk_config *cdclk_config,
+				  enum pipe pipe)
+{
+	dev_priv->display.funcs.cdclk->set_cdclk(dev_priv, cdclk_config, pipe);
+}
+
+static int intel_cdclk_modeset_calc_cdclk(struct drm_i915_private *dev_priv,
+					  struct intel_cdclk_state *cdclk_config)
+{
+	return dev_priv->display.funcs.cdclk->modeset_calc_cdclk(cdclk_config);
+}
+
+static u8 intel_cdclk_calc_voltage_level(struct drm_i915_private *dev_priv,
+					 int cdclk)
+{
+	return dev_priv->display.funcs.cdclk->calc_voltage_level(cdclk);
+}
+
+static void fixed_133mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_config *cdclk_config)
+{
+	cdclk_config->cdclk = 133333;
+}
+
+static void fixed_200mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_config *cdclk_config)
+{
+	cdclk_config->cdclk = 200000;
+}
+
+static void fixed_266mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_config *cdclk_config)
+{
+	cdclk_config->cdclk = 266667;
+}
+
+static void fixed_333mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_config *cdclk_config)
+{
+	cdclk_config->cdclk = 333333;
+}
+
+static void fixed_400mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_config *cdclk_config)
+{
+	cdclk_config->cdclk = 400000;
+}
+
+static void fixed_450mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_config *cdclk_config)
+{
+	cdclk_config->cdclk = 450000;
+}
+
+static void i85x_get_cdclk(struct drm_i915_private *dev_priv,
+			   struct intel_cdclk_config *cdclk_config)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	u16 hpllcc = 0;
+
+	/*
+	 * 852GM/852GMV only supports 133 MHz and the HPLLCC
+	 * encoding is different :(
+	 * FIXME is this the right way to detect 852GM/852GMV?
+	 */
+	if (pdev->revision == 0x1) {
+		cdclk_config->cdclk = 133333;
+		return;
+	}
+
+	pci_bus_read_config_word(pdev->bus,
+				 PCI_DEVFN(0, 3), HPLLCC, &hpllcc);
+
+	/* Assume that the hardware is in the high speed state.  This
+	 * should be the default.
+	 */
+	switch (hpllcc & GC_CLOCK_CONTROL_MASK) {
+	case GC_CLOCK_133_200:
+	case GC_CLOCK_133_200_2:
+	case GC_CLOCK_100_200:
+		cdclk_config->cdclk = 200000;
+		break;
+	case GC_CLOCK_166_250:
+		cdclk_config->cdclk = 250000;
+		break;
+	case GC_CLOCK_100_133:
+		cdclk_config->cdclk = 133333;
+		break;
+	case GC_CLOCK_133_266:
+	case GC_CLOCK_133_266_2:
+	case GC_CLOCK_166_266:
+		cdclk_config->cdclk = 266667;
+		break;
+	}
+}
+
+static void i915gm_get_cdclk(struct drm_i915_private *dev_priv,
+			     struct intel_cdclk_config *cdclk_config)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	u16 gcfgc = 0;
+
+	pci_read_config_word(pdev, GCFGC, &gcfgc);
+
+	if (gcfgc & GC_LOW_FREQUENCY_ENABLE) {
+		cdclk_config->cdclk = 133333;
+		return;
+	}
+
+	switch (gcfgc & GC_DISPLAY_CLOCK_MASK) {
+	case GC_DISPLAY_CLOCK_333_320_MHZ:
+		cdclk_config->cdclk = 333333;
+		break;
+	default:
+	case GC_DISPLAY_CLOCK_190_200_MHZ:
+		cdclk_config->cdclk = 190000;
+		break;
+	}
+}
+
+static void i945gm_get_cdclk(struct drm_i915_private *dev_priv,
+			     struct intel_cdclk_config *cdclk_config)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	u16 gcfgc = 0;
+
+	pci_read_config_word(pdev, GCFGC, &gcfgc);
+
+	if (gcfgc & GC_LOW_FREQUENCY_ENABLE) {
+		cdclk_config->cdclk = 133333;
+		return;
+	}
+
+	switch (gcfgc & GC_DISPLAY_CLOCK_MASK) {
+	case GC_DISPLAY_CLOCK_333_320_MHZ:
+		cdclk_config->cdclk = 320000;
+		break;
+	default:
+	case GC_DISPLAY_CLOCK_190_200_MHZ:
+		cdclk_config->cdclk = 200000;
+		break;
+	}
+}
+
+static unsigned int intel_hpll_vco(struct drm_i915_private *dev_priv)
+{
+	static const unsigned int blb_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 4800000,
+		[4] = 6400000,
+	};
+	static const unsigned int pnv_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 4800000,
+		[4] = 2666667,
+	};
+	static const unsigned int cl_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 6400000,
+		[4] = 3333333,
+		[5] = 3566667,
+		[6] = 4266667,
+	};
+	static const unsigned int elk_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 4800000,
+	};
+	static const unsigned int ctg_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 6400000,
+		[4] = 2666667,
+		[5] = 4266667,
+	};
+	const unsigned int *vco_table;
+	unsigned int vco;
+	u8 tmp = 0;
+
+	/* FIXME other chipsets? */
+	if (IS_GM45(dev_priv))
+		vco_table = ctg_vco;
+	else if (IS_G45(dev_priv))
+		vco_table = elk_vco;
+	else if (IS_I965GM(dev_priv))
+		vco_table = cl_vco;
+	else if (IS_PINEVIEW(dev_priv))
+		vco_table = pnv_vco;
+	else if (IS_G33(dev_priv))
+		vco_table = blb_vco;
+	else
+		return 0;
+
+	tmp = intel_de_read(dev_priv,
+			    IS_PINEVIEW(dev_priv) || IS_MOBILE(dev_priv) ? HPLLVCO_MOBILE : HPLLVCO);
+
+	vco = vco_table[tmp & 0x7];
+	if (vco == 0)
+		drm_err(&dev_priv->drm, "Bad HPLL VCO (HPLLVCO=0x%02x)\n",
+			tmp);
+	else
+		drm_dbg_kms(&dev_priv->drm, "HPLL VCO %u kHz\n", vco);
+
+	return vco;
+}
+
+static void g33_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_config *cdclk_config)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	static const u8 div_3200[] = { 12, 10,  8,  7, 5, 16 };
+	static const u8 div_4000[] = { 14, 12, 10,  8, 6, 20 };
+	static const u8 div_4800[] = { 20, 14, 12, 10, 8, 24 };
+	static const u8 div_5333[] = { 20, 16, 12, 12, 8, 28 };
+	const u8 *div_table;
+	unsigned int cdclk_sel;
+	u16 tmp = 0;
+
+	cdclk_config->vco = intel_hpll_vco(dev_priv);
+
+	pci_read_config_word(pdev, GCFGC, &tmp);
+
+	cdclk_sel = (tmp >> 4) & 0x7;
+
+	if (cdclk_sel >= ARRAY_SIZE(div_3200))
+		goto fail;
+
+	switch (cdclk_config->vco) {
+	case 3200000:
+		div_table = div_3200;
+		break;
+	case 4000000:
+		div_table = div_4000;
+		break;
+	case 4800000:
+		div_table = div_4800;
+		break;
+	case 5333333:
+		div_table = div_5333;
+		break;
+	default:
+		goto fail;
+	}
+
+	cdclk_config->cdclk = DIV_ROUND_CLOSEST(cdclk_config->vco,
+						div_table[cdclk_sel]);
+	return;
+
+fail:
+	drm_err(&dev_priv->drm,
+		"Unable to determine CDCLK. HPLL VCO=%u kHz, CFGC=0x%08x\n",
+		cdclk_config->vco, tmp);
+	cdclk_config->cdclk = 190476;
+}
+
+static void pnv_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_config *cdclk_config)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	u16 gcfgc = 0;
+
+	pci_read_config_word(pdev, GCFGC, &gcfgc);
+
+	switch (gcfgc & GC_DISPLAY_CLOCK_MASK) {
+	case GC_DISPLAY_CLOCK_267_MHZ_PNV:
+		cdclk_config->cdclk = 266667;
+		break;
+	case GC_DISPLAY_CLOCK_333_MHZ_PNV:
+		cdclk_config->cdclk = 333333;
+		break;
+	case GC_DISPLAY_CLOCK_444_MHZ_PNV:
+		cdclk_config->cdclk = 444444;
+		break;
+	case GC_DISPLAY_CLOCK_200_MHZ_PNV:
+		cdclk_config->cdclk = 200000;
+		break;
+	default:
+		drm_err(&dev_priv->drm,
+			"Unknown pnv display core clock 0x%04x\n", gcfgc);
+		fallthrough;
+	case GC_DISPLAY_CLOCK_133_MHZ_PNV:
+		cdclk_config->cdclk = 133333;
+		break;
+	case GC_DISPLAY_CLOCK_167_MHZ_PNV:
+		cdclk_config->cdclk = 166667;
+		break;
+	}
+}
+
+static void i965gm_get_cdclk(struct drm_i915_private *dev_priv,
+			     struct intel_cdclk_config *cdclk_config)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	static const u8 div_3200[] = { 16, 10,  8 };
+	static const u8 div_4000[] = { 20, 12, 10 };
+	static const u8 div_5333[] = { 24, 16, 14 };
+	const u8 *div_table;
+	unsigned int cdclk_sel;
+	u16 tmp = 0;
+
+	cdclk_config->vco = intel_hpll_vco(dev_priv);
+
+	pci_read_config_word(pdev, GCFGC, &tmp);
+
+	cdclk_sel = ((tmp >> 8) & 0x1f) - 1;
+
+	if (cdclk_sel >= ARRAY_SIZE(div_3200))
+		goto fail;
+
+	switch (cdclk_config->vco) {
+	case 3200000:
+		div_table = div_3200;
+		break;
+	case 4000000:
+		div_table = div_4000;
+		break;
+	case 5333333:
+		div_table = div_5333;
+		break;
+	default:
+		goto fail;
+	}
+
+	cdclk_config->cdclk = DIV_ROUND_CLOSEST(cdclk_config->vco,
+						div_table[cdclk_sel]);
+	return;
+
+fail:
+	drm_err(&dev_priv->drm,
+		"Unable to determine CDCLK. HPLL VCO=%u kHz, CFGC=0x%04x\n",
+		cdclk_config->vco, tmp);
+	cdclk_config->cdclk = 200000;
+}
+
+static void gm45_get_cdclk(struct drm_i915_private *dev_priv,
+			   struct intel_cdclk_config *cdclk_config)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	unsigned int cdclk_sel;
+	u16 tmp = 0;
+
+	cdclk_config->vco = intel_hpll_vco(dev_priv);
+
+	pci_read_config_word(pdev, GCFGC, &tmp);
+
+	cdclk_sel = (tmp >> 12) & 0x1;
+
+	switch (cdclk_config->vco) {
+	case 2666667:
+	case 4000000:
+	case 5333333:
+		cdclk_config->cdclk = cdclk_sel ? 333333 : 222222;
+		break;
+	case 3200000:
+		cdclk_config->cdclk = cdclk_sel ? 320000 : 228571;
+		break;
+	default:
+		drm_err(&dev_priv->drm,
+			"Unable to determine CDCLK. HPLL VCO=%u, CFGC=0x%04x\n",
+			cdclk_config->vco, tmp);
+		cdclk_config->cdclk = 222222;
+		break;
+	}
+}
+
+static void hsw_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_config *cdclk_config)
+{
+	u32 lcpll = intel_de_read(dev_priv, LCPLL_CTL);
+	u32 freq = lcpll & LCPLL_CLK_FREQ_MASK;
+
+	if (lcpll & LCPLL_CD_SOURCE_FCLK)
+		cdclk_config->cdclk = 800000;
+	else if (intel_de_read(dev_priv, FUSE_STRAP) & HSW_CDCLK_LIMIT)
+		cdclk_config->cdclk = 450000;
+	else if (freq == LCPLL_CLK_FREQ_450)
+		cdclk_config->cdclk = 450000;
+	else if (IS_HSW_ULT(dev_priv))
+		cdclk_config->cdclk = 337500;
+	else
+		cdclk_config->cdclk = 540000;
+}
+
+static int vlv_calc_cdclk(struct drm_i915_private *dev_priv, int min_cdclk)
+{
+	int freq_320 = (dev_priv->hpll_freq <<  1) % 320000 != 0 ?
+		333333 : 320000;
+
+	/*
+	 * We seem to get an unstable or solid color picture at 200MHz.
+	 * Not sure what's wrong. For now use 200MHz only when all pipes
+	 * are off.
+	 */
+	if (IS_VALLEYVIEW(dev_priv) && min_cdclk > freq_320)
+		return 400000;
+	else if (min_cdclk > 266667)
+		return freq_320;
+	else if (min_cdclk > 0)
+		return 266667;
+	else
+		return 200000;
+}
+
+static u8 vlv_calc_voltage_level(struct drm_i915_private *dev_priv, int cdclk)
+{
+	if (IS_VALLEYVIEW(dev_priv)) {
+		if (cdclk >= 320000) /* jump to highest voltage for 400MHz too */
+			return 2;
+		else if (cdclk >= 266667)
+			return 1;
+		else
+			return 0;
+	} else {
+		/*
+		 * Specs are full of misinformation, but testing on actual
+		 * hardware has shown that we just need to write the desired
+		 * CCK divider into the Punit register.
+		 */
+		return DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1, cdclk) - 1;
+	}
+}
+
+static void vlv_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_config *cdclk_config)
+{
+	u32 val;
+
+	vlv_iosf_sb_get(dev_priv,
+			BIT(VLV_IOSF_SB_CCK) | BIT(VLV_IOSF_SB_PUNIT));
+
+	cdclk_config->vco = vlv_get_hpll_vco(dev_priv);
+	cdclk_config->cdclk = vlv_get_cck_clock(dev_priv, "cdclk",
+						CCK_DISPLAY_CLOCK_CONTROL,
+						cdclk_config->vco);
+
+	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM);
+
+	vlv_iosf_sb_put(dev_priv,
+			BIT(VLV_IOSF_SB_CCK) | BIT(VLV_IOSF_SB_PUNIT));
+
+	if (IS_VALLEYVIEW(dev_priv))
+		cdclk_config->voltage_level = (val & DSPFREQGUAR_MASK) >>
+			DSPFREQGUAR_SHIFT;
+	else
+		cdclk_config->voltage_level = (val & DSPFREQGUAR_MASK_CHV) >>
+			DSPFREQGUAR_SHIFT_CHV;
+}
+
+static void vlv_program_pfi_credits(struct drm_i915_private *dev_priv)
+{
+	unsigned int credits, default_credits;
+
+	if (IS_CHERRYVIEW(dev_priv))
+		default_credits = PFI_CREDIT(12);
+	else
+		default_credits = PFI_CREDIT(8);
+
+	if (dev_priv->display.cdclk.hw.cdclk >= dev_priv->czclk_freq) {
+		/* CHV suggested value is 31 or 63 */
+		if (IS_CHERRYVIEW(dev_priv))
+			credits = PFI_CREDIT_63;
+		else
+			credits = PFI_CREDIT(15);
+	} else {
+		credits = default_credits;
+	}
+
+	/*
+	 * WA - write default credits before re-programming
+	 * FIXME: should we also set the resend bit here?
+	 */
+	intel_de_write(dev_priv, GCI_CONTROL,
+		       VGA_FAST_MODE_DISABLE | default_credits);
+
+	intel_de_write(dev_priv, GCI_CONTROL,
+		       VGA_FAST_MODE_DISABLE | credits | PFI_CREDIT_RESEND);
+
+	/*
+	 * FIXME is this guaranteed to clear
+	 * immediately or should we poll for it?
+	 */
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, GCI_CONTROL) & PFI_CREDIT_RESEND);
+}
+
+static void vlv_set_cdclk(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_config *cdclk_config,
+			  enum pipe pipe)
+{
+	int cdclk = cdclk_config->cdclk;
+	u32 val, cmd = cdclk_config->voltage_level;
+	intel_wakeref_t wakeref;
+
+	switch (cdclk) {
+	case 400000:
+	case 333333:
+	case 320000:
+	case 266667:
+	case 200000:
+		break;
+	default:
+		MISSING_CASE(cdclk);
+		return;
+	}
+
+	/* There are cases where we can end up here with power domains
+	 * off and a CDCLK frequency other than the minimum, like when
+	 * issuing a modeset without actually changing any display after
+	 * a system suspend.  So grab the display core domain, which covers
+	 * the HW blocks needed for the following programming.
+	 */
+	wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_DISPLAY_CORE);
+
+	vlv_iosf_sb_get(dev_priv,
+			BIT(VLV_IOSF_SB_CCK) |
+			BIT(VLV_IOSF_SB_BUNIT) |
+			BIT(VLV_IOSF_SB_PUNIT));
+
+	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM);
+	val &= ~DSPFREQGUAR_MASK;
+	val |= (cmd << DSPFREQGUAR_SHIFT);
+	vlv_punit_write(dev_priv, PUNIT_REG_DSPSSPM, val);
+	if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM) &
+		      DSPFREQSTAT_MASK) == (cmd << DSPFREQSTAT_SHIFT),
+		     50)) {
+		drm_err(&dev_priv->drm,
+			"timed out waiting for CDclk change\n");
+	}
+
+	if (cdclk == 400000) {
+		u32 divider;
+
+		divider = DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1,
+					    cdclk) - 1;
+
+		/* adjust cdclk divider */
+		val = vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL);
+		val &= ~CCK_FREQUENCY_VALUES;
+		val |= divider;
+		vlv_cck_write(dev_priv, CCK_DISPLAY_CLOCK_CONTROL, val);
+
+		if (wait_for((vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL) &
+			      CCK_FREQUENCY_STATUS) == (divider << CCK_FREQUENCY_STATUS_SHIFT),
+			     50))
+			drm_err(&dev_priv->drm,
+				"timed out waiting for CDclk change\n");
+	}
+
+	/* adjust self-refresh exit latency value */
+	val = vlv_bunit_read(dev_priv, BUNIT_REG_BISOC);
+	val &= ~0x7f;
+
+	/*
+	 * For high bandwidth configs, we set a higher latency in the bunit
+	 * so that the core display fetch happens in time to avoid underruns.
+	 */
+	if (cdclk == 400000)
+		val |= 4500 / 250; /* 4.5 usec */
+	else
+		val |= 3000 / 250; /* 3.0 usec */
+	vlv_bunit_write(dev_priv, BUNIT_REG_BISOC, val);
+
+	vlv_iosf_sb_put(dev_priv,
+			BIT(VLV_IOSF_SB_CCK) |
+			BIT(VLV_IOSF_SB_BUNIT) |
+			BIT(VLV_IOSF_SB_PUNIT));
+
+	intel_update_cdclk(dev_priv);
+
+	vlv_program_pfi_credits(dev_priv);
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+}
+
+static void chv_set_cdclk(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_config *cdclk_config,
+			  enum pipe pipe)
+{
+	int cdclk = cdclk_config->cdclk;
+	u32 val, cmd = cdclk_config->voltage_level;
+	intel_wakeref_t wakeref;
+
+	switch (cdclk) {
+	case 333333:
+	case 320000:
+	case 266667:
+	case 200000:
+		break;
+	default:
+		MISSING_CASE(cdclk);
+		return;
+	}
+
+	/* There are cases where we can end up here with power domains
+	 * off and a CDCLK frequency other than the minimum, like when
+	 * issuing a modeset without actually changing any display after
+	 * a system suspend.  So grab the display core domain, which covers
+	 * the HW blocks needed for the following programming.
+	 */
+	wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_DISPLAY_CORE);
+
+	vlv_punit_get(dev_priv);
+	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM);
+	val &= ~DSPFREQGUAR_MASK_CHV;
+	val |= (cmd << DSPFREQGUAR_SHIFT_CHV);
+	vlv_punit_write(dev_priv, PUNIT_REG_DSPSSPM, val);
+	if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM) &
+		      DSPFREQSTAT_MASK_CHV) == (cmd << DSPFREQSTAT_SHIFT_CHV),
+		     50)) {
+		drm_err(&dev_priv->drm,
+			"timed out waiting for CDclk change\n");
+	}
+
+	vlv_punit_put(dev_priv);
+
+	intel_update_cdclk(dev_priv);
+
+	vlv_program_pfi_credits(dev_priv);
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+}
+
+static int bdw_calc_cdclk(int min_cdclk)
+{
+	if (min_cdclk > 540000)
+		return 675000;
+	else if (min_cdclk > 450000)
+		return 540000;
+	else if (min_cdclk > 337500)
+		return 450000;
+	else
+		return 337500;
+}
+
+static u8 bdw_calc_voltage_level(int cdclk)
+{
+	switch (cdclk) {
+	default:
+	case 337500:
+		return 2;
+	case 450000:
+		return 0;
+	case 540000:
+		return 1;
+	case 675000:
+		return 3;
+	}
+}
+
+static void bdw_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_config *cdclk_config)
+{
+	u32 lcpll = intel_de_read(dev_priv, LCPLL_CTL);
+	u32 freq = lcpll & LCPLL_CLK_FREQ_MASK;
+
+	if (lcpll & LCPLL_CD_SOURCE_FCLK)
+		cdclk_config->cdclk = 800000;
+	else if (intel_de_read(dev_priv, FUSE_STRAP) & HSW_CDCLK_LIMIT)
+		cdclk_config->cdclk = 450000;
+	else if (freq == LCPLL_CLK_FREQ_450)
+		cdclk_config->cdclk = 450000;
+	else if (freq == LCPLL_CLK_FREQ_54O_BDW)
+		cdclk_config->cdclk = 540000;
+	else if (freq == LCPLL_CLK_FREQ_337_5_BDW)
+		cdclk_config->cdclk = 337500;
+	else
+		cdclk_config->cdclk = 675000;
+
+	/*
+	 * Can't read this out :( Let's assume it's
+	 * at least what the CDCLK frequency requires.
+	 */
+	cdclk_config->voltage_level =
+		bdw_calc_voltage_level(cdclk_config->cdclk);
+}
+
+static u32 bdw_cdclk_freq_sel(int cdclk)
+{
+	switch (cdclk) {
+	default:
+		MISSING_CASE(cdclk);
+		fallthrough;
+	case 337500:
+		return LCPLL_CLK_FREQ_337_5_BDW;
+	case 450000:
+		return LCPLL_CLK_FREQ_450;
+	case 540000:
+		return LCPLL_CLK_FREQ_54O_BDW;
+	case 675000:
+		return LCPLL_CLK_FREQ_675_BDW;
+	}
+}
+
+static void bdw_set_cdclk(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_config *cdclk_config,
+			  enum pipe pipe)
+{
+	int cdclk = cdclk_config->cdclk;
+	int ret;
+
+	if (drm_WARN(&dev_priv->drm,
+		     (intel_de_read(dev_priv, LCPLL_CTL) &
+		      (LCPLL_PLL_DISABLE | LCPLL_PLL_LOCK |
+		       LCPLL_CD_CLOCK_DISABLE | LCPLL_ROOT_CD_CLOCK_DISABLE |
+		       LCPLL_CD2X_CLOCK_DISABLE | LCPLL_POWER_DOWN_ALLOW |
+		       LCPLL_CD_SOURCE_FCLK)) != LCPLL_PLL_LOCK,
+		     "trying to change cdclk frequency with cdclk not enabled\n"))
+		return;
+
+	ret = snb_pcode_write(&dev_priv->uncore, BDW_PCODE_DISPLAY_FREQ_CHANGE_REQ, 0x0);
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"failed to inform pcode about cdclk change\n");
+		return;
+	}
+
+	intel_de_rmw(dev_priv, LCPLL_CTL,
+		     0, LCPLL_CD_SOURCE_FCLK);
+
+	/*
+	 * According to the spec, it should be enough to poll for this 1 us.
+	 * However, extensive testing shows that this can take longer.
+	 */
+	if (wait_for_us(intel_de_read(dev_priv, LCPLL_CTL) &
+			LCPLL_CD_SOURCE_FCLK_DONE, 100))
+		drm_err(&dev_priv->drm, "Switching to FCLK failed\n");
+
+	intel_de_rmw(dev_priv, LCPLL_CTL,
+		     LCPLL_CLK_FREQ_MASK, bdw_cdclk_freq_sel(cdclk));
+
+	intel_de_rmw(dev_priv, LCPLL_CTL,
+		     LCPLL_CD_SOURCE_FCLK, 0);
+
+	if (wait_for_us((intel_de_read(dev_priv, LCPLL_CTL) &
+			 LCPLL_CD_SOURCE_FCLK_DONE) == 0, 1))
+		drm_err(&dev_priv->drm, "Switching back to LCPLL failed\n");
+
+	snb_pcode_write(&dev_priv->uncore, HSW_PCODE_DE_WRITE_FREQ_REQ,
+			cdclk_config->voltage_level);
+
+	intel_de_write(dev_priv, CDCLK_FREQ,
+		       DIV_ROUND_CLOSEST(cdclk, 1000) - 1);
+
+	intel_update_cdclk(dev_priv);
+}
+
+static int skl_calc_cdclk(int min_cdclk, int vco)
+{
+	if (vco == 8640000) {
+		if (min_cdclk > 540000)
+			return 617143;
+		else if (min_cdclk > 432000)
+			return 540000;
+		else if (min_cdclk > 308571)
+			return 432000;
+		else
+			return 308571;
+	} else {
+		if (min_cdclk > 540000)
+			return 675000;
+		else if (min_cdclk > 450000)
+			return 540000;
+		else if (min_cdclk > 337500)
+			return 450000;
+		else
+			return 337500;
+	}
+}
+
+static u8 skl_calc_voltage_level(int cdclk)
+{
+	if (cdclk > 540000)
+		return 3;
+	else if (cdclk > 450000)
+		return 2;
+	else if (cdclk > 337500)
+		return 1;
+	else
+		return 0;
+}
+
+static void skl_dpll0_update(struct drm_i915_private *dev_priv,
+			     struct intel_cdclk_config *cdclk_config)
+{
+	u32 val;
+
+	cdclk_config->ref = 24000;
+	cdclk_config->vco = 0;
+
+	val = intel_de_read(dev_priv, LCPLL1_CTL);
+	if ((val & LCPLL_PLL_ENABLE) == 0)
+		return;
+
+	if (drm_WARN_ON(&dev_priv->drm, (val & LCPLL_PLL_LOCK) == 0))
+		return;
+
+	val = intel_de_read(dev_priv, DPLL_CTRL1);
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			(val & (DPLL_CTRL1_HDMI_MODE(SKL_DPLL0) |
+				DPLL_CTRL1_SSC(SKL_DPLL0) |
+				DPLL_CTRL1_OVERRIDE(SKL_DPLL0))) !=
+			DPLL_CTRL1_OVERRIDE(SKL_DPLL0)))
+		return;
+
+	switch (val & DPLL_CTRL1_LINK_RATE_MASK(SKL_DPLL0)) {
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810, SKL_DPLL0):
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1350, SKL_DPLL0):
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1620, SKL_DPLL0):
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_2700, SKL_DPLL0):
+		cdclk_config->vco = 8100000;
+		break;
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1080, SKL_DPLL0):
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_2160, SKL_DPLL0):
+		cdclk_config->vco = 8640000;
+		break;
+	default:
+		MISSING_CASE(val & DPLL_CTRL1_LINK_RATE_MASK(SKL_DPLL0));
+		break;
+	}
+}
+
+static void skl_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_config *cdclk_config)
+{
+	u32 cdctl;
+
+	skl_dpll0_update(dev_priv, cdclk_config);
+
+	cdclk_config->cdclk = cdclk_config->bypass = cdclk_config->ref;
+
+	if (cdclk_config->vco == 0)
+		goto out;
+
+	cdctl = intel_de_read(dev_priv, CDCLK_CTL);
+
+	if (cdclk_config->vco == 8640000) {
+		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
+		case CDCLK_FREQ_450_432:
+			cdclk_config->cdclk = 432000;
+			break;
+		case CDCLK_FREQ_337_308:
+			cdclk_config->cdclk = 308571;
+			break;
+		case CDCLK_FREQ_540:
+			cdclk_config->cdclk = 540000;
+			break;
+		case CDCLK_FREQ_675_617:
+			cdclk_config->cdclk = 617143;
+			break;
+		default:
+			MISSING_CASE(cdctl & CDCLK_FREQ_SEL_MASK);
+			break;
+		}
+	} else {
+		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
+		case CDCLK_FREQ_450_432:
+			cdclk_config->cdclk = 450000;
+			break;
+		case CDCLK_FREQ_337_308:
+			cdclk_config->cdclk = 337500;
+			break;
+		case CDCLK_FREQ_540:
+			cdclk_config->cdclk = 540000;
+			break;
+		case CDCLK_FREQ_675_617:
+			cdclk_config->cdclk = 675000;
+			break;
+		default:
+			MISSING_CASE(cdctl & CDCLK_FREQ_SEL_MASK);
+			break;
+		}
+	}
+
+ out:
+	/*
+	 * Can't read this out :( Let's assume it's
+	 * at least what the CDCLK frequency requires.
+	 */
+	cdclk_config->voltage_level =
+		skl_calc_voltage_level(cdclk_config->cdclk);
+}
+
+/* convert from kHz to .1 fixpoint MHz with -1MHz offset */
+static int skl_cdclk_decimal(int cdclk)
+{
+	return DIV_ROUND_CLOSEST(cdclk - 1000, 500);
+}
+
+static void skl_set_preferred_cdclk_vco(struct drm_i915_private *dev_priv,
+					int vco)
+{
+	bool changed = dev_priv->skl_preferred_vco_freq != vco;
+
+	dev_priv->skl_preferred_vco_freq = vco;
+
+	if (changed)
+		intel_update_max_cdclk(dev_priv);
+}
+
+static u32 skl_dpll0_link_rate(struct drm_i915_private *dev_priv, int vco)
+{
+	drm_WARN_ON(&dev_priv->drm, vco != 8100000 && vco != 8640000);
+
+	/*
+	 * We always enable DPLL0 with the lowest link rate possible, but still
+	 * taking into account the VCO required to operate the eDP panel at the
+	 * desired frequency. The usual DP link rates operate with a VCO of
+	 * 8100 while the eDP 1.4 alternate link rates need a VCO of 8640.
+	 * The modeset code is responsible for the selection of the exact link
+	 * rate later on, with the constraint of choosing a frequency that
+	 * works with vco.
+	 */
+	if (vco == 8640000)
+		return DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1080, SKL_DPLL0);
+	else
+		return DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810, SKL_DPLL0);
+}
+
+static void skl_dpll0_enable(struct drm_i915_private *dev_priv, int vco)
+{
+	intel_de_rmw(dev_priv, DPLL_CTRL1,
+		     DPLL_CTRL1_HDMI_MODE(SKL_DPLL0) |
+		     DPLL_CTRL1_SSC(SKL_DPLL0) |
+		     DPLL_CTRL1_LINK_RATE_MASK(SKL_DPLL0),
+		     DPLL_CTRL1_OVERRIDE(SKL_DPLL0) |
+		     skl_dpll0_link_rate(dev_priv, vco));
+	intel_de_posting_read(dev_priv, DPLL_CTRL1);
+
+	intel_de_rmw(dev_priv, LCPLL1_CTL,
+		     0, LCPLL_PLL_ENABLE);
+
+	if (intel_de_wait_for_set(dev_priv, LCPLL1_CTL, LCPLL_PLL_LOCK, 5))
+		drm_err(&dev_priv->drm, "DPLL0 not locked\n");
+
+	dev_priv->display.cdclk.hw.vco = vco;
+
+	/* We'll want to keep using the current vco from now on. */
+	skl_set_preferred_cdclk_vco(dev_priv, vco);
+}
+
+static void skl_dpll0_disable(struct drm_i915_private *dev_priv)
+{
+	intel_de_rmw(dev_priv, LCPLL1_CTL,
+		     LCPLL_PLL_ENABLE, 0);
+
+	if (intel_de_wait_for_clear(dev_priv, LCPLL1_CTL, LCPLL_PLL_LOCK, 1))
+		drm_err(&dev_priv->drm, "Couldn't disable DPLL0\n");
+
+	dev_priv->display.cdclk.hw.vco = 0;
+}
+
+static u32 skl_cdclk_freq_sel(struct drm_i915_private *dev_priv,
+			      int cdclk, int vco)
+{
+	switch (cdclk) {
+	default:
+		drm_WARN_ON(&dev_priv->drm,
+			    cdclk != dev_priv->display.cdclk.hw.bypass);
+		drm_WARN_ON(&dev_priv->drm, vco != 0);
+		fallthrough;
+	case 308571:
+	case 337500:
+		return CDCLK_FREQ_337_308;
+	case 450000:
+	case 432000:
+		return CDCLK_FREQ_450_432;
+	case 540000:
+		return CDCLK_FREQ_540;
+	case 617143:
+	case 675000:
+		return CDCLK_FREQ_675_617;
+	}
+}
+
+static void skl_set_cdclk(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_config *cdclk_config,
+			  enum pipe pipe)
+{
+	int cdclk = cdclk_config->cdclk;
+	int vco = cdclk_config->vco;
+	u32 freq_select, cdclk_ctl;
+	int ret;
+
+	/*
+	 * Based on WA#1183 CDCLK rates 308 and 617MHz CDCLK rates are
+	 * unsupported on SKL. In theory this should never happen since only
+	 * the eDP1.4 2.16 and 4.32Gbps rates require it, but eDP1.4 is not
+	 * supported on SKL either, see the above WA. WARN whenever trying to
+	 * use the corresponding VCO freq as that always leads to using the
+	 * minimum 308MHz CDCLK.
+	 */
+	drm_WARN_ON_ONCE(&dev_priv->drm,
+			 IS_SKYLAKE(dev_priv) && vco == 8640000);
+
+	ret = skl_pcode_request(&dev_priv->uncore, SKL_PCODE_CDCLK_CONTROL,
+				SKL_CDCLK_PREPARE_FOR_CHANGE,
+				SKL_CDCLK_READY_FOR_CHANGE,
+				SKL_CDCLK_READY_FOR_CHANGE, 3);
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"Failed to inform PCU about cdclk change (%d)\n", ret);
+		return;
+	}
+
+	freq_select = skl_cdclk_freq_sel(dev_priv, cdclk, vco);
+
+	if (dev_priv->display.cdclk.hw.vco != 0 &&
+	    dev_priv->display.cdclk.hw.vco != vco)
+		skl_dpll0_disable(dev_priv);
+
+	cdclk_ctl = intel_de_read(dev_priv, CDCLK_CTL);
+
+	if (dev_priv->display.cdclk.hw.vco != vco) {
+		/* Wa Display #1183: skl,kbl,cfl */
+		cdclk_ctl &= ~(CDCLK_FREQ_SEL_MASK | CDCLK_FREQ_DECIMAL_MASK);
+		cdclk_ctl |= freq_select | skl_cdclk_decimal(cdclk);
+		intel_de_write(dev_priv, CDCLK_CTL, cdclk_ctl);
+	}
+
+	/* Wa Display #1183: skl,kbl,cfl */
+	cdclk_ctl |= CDCLK_DIVMUX_CD_OVERRIDE;
+	intel_de_write(dev_priv, CDCLK_CTL, cdclk_ctl);
+	intel_de_posting_read(dev_priv, CDCLK_CTL);
+
+	if (dev_priv->display.cdclk.hw.vco != vco)
+		skl_dpll0_enable(dev_priv, vco);
+
+	/* Wa Display #1183: skl,kbl,cfl */
+	cdclk_ctl &= ~(CDCLK_FREQ_SEL_MASK | CDCLK_FREQ_DECIMAL_MASK);
+	intel_de_write(dev_priv, CDCLK_CTL, cdclk_ctl);
+
+	cdclk_ctl |= freq_select | skl_cdclk_decimal(cdclk);
+	intel_de_write(dev_priv, CDCLK_CTL, cdclk_ctl);
+
+	/* Wa Display #1183: skl,kbl,cfl */
+	cdclk_ctl &= ~CDCLK_DIVMUX_CD_OVERRIDE;
+	intel_de_write(dev_priv, CDCLK_CTL, cdclk_ctl);
+	intel_de_posting_read(dev_priv, CDCLK_CTL);
+
+	/* inform PCU of the change */
+	snb_pcode_write(&dev_priv->uncore, SKL_PCODE_CDCLK_CONTROL,
+			cdclk_config->voltage_level);
+
+	intel_update_cdclk(dev_priv);
+}
+
+static void skl_sanitize_cdclk(struct drm_i915_private *dev_priv)
+{
+	u32 cdctl, expected;
+
+	/*
+	 * check if the pre-os initialized the display
+	 * There is SWF18 scratchpad register defined which is set by the
+	 * pre-os which can be used by the OS drivers to check the status
+	 */
+	if ((intel_de_read(dev_priv, SWF_ILK(0x18)) & 0x00FFFFFF) == 0)
+		goto sanitize;
+
+	intel_update_cdclk(dev_priv);
+	intel_cdclk_dump_config(dev_priv, &dev_priv->display.cdclk.hw, "Current CDCLK");
+
+	/* Is PLL enabled and locked ? */
+	if (dev_priv->display.cdclk.hw.vco == 0 ||
+	    dev_priv->display.cdclk.hw.cdclk == dev_priv->display.cdclk.hw.bypass)
+		goto sanitize;
+
+	/* DPLL okay; verify the cdclock
+	 *
+	 * Noticed in some instances that the freq selection is correct but
+	 * decimal part is programmed wrong from BIOS where pre-os does not
+	 * enable display. Verify the same as well.
+	 */
+	cdctl = intel_de_read(dev_priv, CDCLK_CTL);
+	expected = (cdctl & CDCLK_FREQ_SEL_MASK) |
+		skl_cdclk_decimal(dev_priv->display.cdclk.hw.cdclk);
+	if (cdctl == expected)
+		/* All well; nothing to sanitize */
+		return;
+
+sanitize:
+	drm_dbg_kms(&dev_priv->drm, "Sanitizing cdclk programmed by pre-os\n");
+
+	/* force cdclk programming */
+	dev_priv->display.cdclk.hw.cdclk = 0;
+	/* force full PLL disable + enable */
+	dev_priv->display.cdclk.hw.vco = -1;
+}
+
+static void skl_cdclk_init_hw(struct drm_i915_private *dev_priv)
+{
+	struct intel_cdclk_config cdclk_config;
+
+	skl_sanitize_cdclk(dev_priv);
+
+	if (dev_priv->display.cdclk.hw.cdclk != 0 &&
+	    dev_priv->display.cdclk.hw.vco != 0) {
+		/*
+		 * Use the current vco as our initial
+		 * guess as to what the preferred vco is.
+		 */
+		if (dev_priv->skl_preferred_vco_freq == 0)
+			skl_set_preferred_cdclk_vco(dev_priv,
+						    dev_priv->display.cdclk.hw.vco);
+		return;
+	}
+
+	cdclk_config = dev_priv->display.cdclk.hw;
+
+	cdclk_config.vco = dev_priv->skl_preferred_vco_freq;
+	if (cdclk_config.vco == 0)
+		cdclk_config.vco = 8100000;
+	cdclk_config.cdclk = skl_calc_cdclk(0, cdclk_config.vco);
+	cdclk_config.voltage_level = skl_calc_voltage_level(cdclk_config.cdclk);
+
+	skl_set_cdclk(dev_priv, &cdclk_config, INVALID_PIPE);
+}
+
+static void skl_cdclk_uninit_hw(struct drm_i915_private *dev_priv)
+{
+	struct intel_cdclk_config cdclk_config = dev_priv->display.cdclk.hw;
+
+	cdclk_config.cdclk = cdclk_config.bypass;
+	cdclk_config.vco = 0;
+	cdclk_config.voltage_level = skl_calc_voltage_level(cdclk_config.cdclk);
+
+	skl_set_cdclk(dev_priv, &cdclk_config, INVALID_PIPE);
+}
+
+static bool has_cdclk_squasher(struct drm_i915_private *i915)
+{
+	return IS_DG2(i915);
+}
+
+struct intel_cdclk_vals {
+	u32 cdclk;
+	u16 refclk;
+	u16 waveform;
+	u8 divider;	/* CD2X divider * 2 */
+	u8 ratio;
+};
+
+static const struct intel_cdclk_vals bxt_cdclk_table[] = {
+	{ .refclk = 19200, .cdclk = 144000, .divider = 8, .ratio = 60 },
+	{ .refclk = 19200, .cdclk = 288000, .divider = 4, .ratio = 60 },
+	{ .refclk = 19200, .cdclk = 384000, .divider = 3, .ratio = 60 },
+	{ .refclk = 19200, .cdclk = 576000, .divider = 2, .ratio = 60 },
+	{ .refclk = 19200, .cdclk = 624000, .divider = 2, .ratio = 65 },
+	{}
+};
+
+static const struct intel_cdclk_vals glk_cdclk_table[] = {
+	{ .refclk = 19200, .cdclk =  79200, .divider = 8, .ratio = 33 },
+	{ .refclk = 19200, .cdclk = 158400, .divider = 4, .ratio = 33 },
+	{ .refclk = 19200, .cdclk = 316800, .divider = 2, .ratio = 33 },
+	{}
+};
+
+static const struct intel_cdclk_vals icl_cdclk_table[] = {
+	{ .refclk = 19200, .cdclk = 172800, .divider = 2, .ratio = 18 },
+	{ .refclk = 19200, .cdclk = 192000, .divider = 2, .ratio = 20 },
+	{ .refclk = 19200, .cdclk = 307200, .divider = 2, .ratio = 32 },
+	{ .refclk = 19200, .cdclk = 326400, .divider = 4, .ratio = 68 },
+	{ .refclk = 19200, .cdclk = 556800, .divider = 2, .ratio = 58 },
+	{ .refclk = 19200, .cdclk = 652800, .divider = 2, .ratio = 68 },
+
+	{ .refclk = 24000, .cdclk = 180000, .divider = 2, .ratio = 15 },
+	{ .refclk = 24000, .cdclk = 192000, .divider = 2, .ratio = 16 },
+	{ .refclk = 24000, .cdclk = 312000, .divider = 2, .ratio = 26 },
+	{ .refclk = 24000, .cdclk = 324000, .divider = 4, .ratio = 54 },
+	{ .refclk = 24000, .cdclk = 552000, .divider = 2, .ratio = 46 },
+	{ .refclk = 24000, .cdclk = 648000, .divider = 2, .ratio = 54 },
+
+	{ .refclk = 38400, .cdclk = 172800, .divider = 2, .ratio =  9 },
+	{ .refclk = 38400, .cdclk = 192000, .divider = 2, .ratio = 10 },
+	{ .refclk = 38400, .cdclk = 307200, .divider = 2, .ratio = 16 },
+	{ .refclk = 38400, .cdclk = 326400, .divider = 4, .ratio = 34 },
+	{ .refclk = 38400, .cdclk = 556800, .divider = 2, .ratio = 29 },
+	{ .refclk = 38400, .cdclk = 652800, .divider = 2, .ratio = 34 },
+	{}
+};
+
+static const struct intel_cdclk_vals rkl_cdclk_table[] = {
+	{ .refclk = 19200, .cdclk = 172800, .divider = 4, .ratio =  36 },
+	{ .refclk = 19200, .cdclk = 192000, .divider = 4, .ratio =  40 },
+	{ .refclk = 19200, .cdclk = 307200, .divider = 4, .ratio =  64 },
+	{ .refclk = 19200, .cdclk = 326400, .divider = 8, .ratio = 136 },
+	{ .refclk = 19200, .cdclk = 556800, .divider = 4, .ratio = 116 },
+	{ .refclk = 19200, .cdclk = 652800, .divider = 4, .ratio = 136 },
+
+	{ .refclk = 24000, .cdclk = 180000, .divider = 4, .ratio =  30 },
+	{ .refclk = 24000, .cdclk = 192000, .divider = 4, .ratio =  32 },
+	{ .refclk = 24000, .cdclk = 312000, .divider = 4, .ratio =  52 },
+	{ .refclk = 24000, .cdclk = 324000, .divider = 8, .ratio = 108 },
+	{ .refclk = 24000, .cdclk = 552000, .divider = 4, .ratio =  92 },
+	{ .refclk = 24000, .cdclk = 648000, .divider = 4, .ratio = 108 },
+
+	{ .refclk = 38400, .cdclk = 172800, .divider = 4, .ratio = 18 },
+	{ .refclk = 38400, .cdclk = 192000, .divider = 4, .ratio = 20 },
+	{ .refclk = 38400, .cdclk = 307200, .divider = 4, .ratio = 32 },
+	{ .refclk = 38400, .cdclk = 326400, .divider = 8, .ratio = 68 },
+	{ .refclk = 38400, .cdclk = 556800, .divider = 4, .ratio = 58 },
+	{ .refclk = 38400, .cdclk = 652800, .divider = 4, .ratio = 68 },
+	{}
+};
+
+static const struct intel_cdclk_vals adlp_a_step_cdclk_table[] = {
+	{ .refclk = 19200, .cdclk = 307200, .divider = 2, .ratio = 32 },
+	{ .refclk = 19200, .cdclk = 556800, .divider = 2, .ratio = 58 },
+	{ .refclk = 19200, .cdclk = 652800, .divider = 2, .ratio = 68 },
+
+	{ .refclk = 24000, .cdclk = 312000, .divider = 2, .ratio = 26 },
+	{ .refclk = 24000, .cdclk = 552000, .divider = 2, .ratio = 46 },
+	{ .refclk = 24400, .cdclk = 648000, .divider = 2, .ratio = 54 },
+
+	{ .refclk = 38400, .cdclk = 307200, .divider = 2, .ratio = 16 },
+	{ .refclk = 38400, .cdclk = 556800, .divider = 2, .ratio = 29 },
+	{ .refclk = 38400, .cdclk = 652800, .divider = 2, .ratio = 34 },
+	{}
+};
+
+static const struct intel_cdclk_vals adlp_cdclk_table[] = {
+	{ .refclk = 19200, .cdclk = 172800, .divider = 3, .ratio = 27 },
+	{ .refclk = 19200, .cdclk = 192000, .divider = 2, .ratio = 20 },
+	{ .refclk = 19200, .cdclk = 307200, .divider = 2, .ratio = 32 },
+	{ .refclk = 19200, .cdclk = 556800, .divider = 2, .ratio = 58 },
+	{ .refclk = 19200, .cdclk = 652800, .divider = 2, .ratio = 68 },
+
+	{ .refclk = 24000, .cdclk = 176000, .divider = 3, .ratio = 22 },
+	{ .refclk = 24000, .cdclk = 192000, .divider = 2, .ratio = 16 },
+	{ .refclk = 24000, .cdclk = 312000, .divider = 2, .ratio = 26 },
+	{ .refclk = 24000, .cdclk = 552000, .divider = 2, .ratio = 46 },
+	{ .refclk = 24000, .cdclk = 648000, .divider = 2, .ratio = 54 },
+
+	{ .refclk = 38400, .cdclk = 179200, .divider = 3, .ratio = 14 },
+	{ .refclk = 38400, .cdclk = 192000, .divider = 2, .ratio = 10 },
+	{ .refclk = 38400, .cdclk = 307200, .divider = 2, .ratio = 16 },
+	{ .refclk = 38400, .cdclk = 556800, .divider = 2, .ratio = 29 },
+	{ .refclk = 38400, .cdclk = 652800, .divider = 2, .ratio = 34 },
+	{}
+};
+
+static const struct intel_cdclk_vals dg2_cdclk_table[] = {
+	{ .refclk = 38400, .cdclk = 163200, .divider = 2, .ratio = 34, .waveform = 0x8888 },
+	{ .refclk = 38400, .cdclk = 204000, .divider = 2, .ratio = 34, .waveform = 0x9248 },
+	{ .refclk = 38400, .cdclk = 244800, .divider = 2, .ratio = 34, .waveform = 0xa4a4 },
+	{ .refclk = 38400, .cdclk = 285600, .divider = 2, .ratio = 34, .waveform = 0xa54a },
+	{ .refclk = 38400, .cdclk = 326400, .divider = 2, .ratio = 34, .waveform = 0xaaaa },
+	{ .refclk = 38400, .cdclk = 367200, .divider = 2, .ratio = 34, .waveform = 0xad5a },
+	{ .refclk = 38400, .cdclk = 408000, .divider = 2, .ratio = 34, .waveform = 0xb6b6 },
+	{ .refclk = 38400, .cdclk = 448800, .divider = 2, .ratio = 34, .waveform = 0xdbb6 },
+	{ .refclk = 38400, .cdclk = 489600, .divider = 2, .ratio = 34, .waveform = 0xeeee },
+	{ .refclk = 38400, .cdclk = 530400, .divider = 2, .ratio = 34, .waveform = 0xf7de },
+	{ .refclk = 38400, .cdclk = 571200, .divider = 2, .ratio = 34, .waveform = 0xfefe },
+	{ .refclk = 38400, .cdclk = 612000, .divider = 2, .ratio = 34, .waveform = 0xfffe },
+	{ .refclk = 38400, .cdclk = 652800, .divider = 2, .ratio = 34, .waveform = 0xffff },
+	{}
+};
+
+static int bxt_calc_cdclk(struct drm_i915_private *dev_priv, int min_cdclk)
+{
+	const struct intel_cdclk_vals *table = dev_priv->display.cdclk.table;
+	int i;
+
+	for (i = 0; table[i].refclk; i++)
+		if (table[i].refclk == dev_priv->display.cdclk.hw.ref &&
+		    table[i].cdclk >= min_cdclk)
+			return table[i].cdclk;
+
+	drm_WARN(&dev_priv->drm, 1,
+		 "Cannot satisfy minimum cdclk %d with refclk %u\n",
+		 min_cdclk, dev_priv->display.cdclk.hw.ref);
+	return 0;
+}
+
+static int bxt_calc_cdclk_pll_vco(struct drm_i915_private *dev_priv, int cdclk)
+{
+	const struct intel_cdclk_vals *table = dev_priv->display.cdclk.table;
+	int i;
+
+	if (cdclk == dev_priv->display.cdclk.hw.bypass)
+		return 0;
+
+	for (i = 0; table[i].refclk; i++)
+		if (table[i].refclk == dev_priv->display.cdclk.hw.ref &&
+		    table[i].cdclk == cdclk)
+			return dev_priv->display.cdclk.hw.ref * table[i].ratio;
+
+	drm_WARN(&dev_priv->drm, 1, "cdclk %d not valid for refclk %u\n",
+		 cdclk, dev_priv->display.cdclk.hw.ref);
+	return 0;
+}
+
+static u8 bxt_calc_voltage_level(int cdclk)
+{
+	return DIV_ROUND_UP(cdclk, 25000);
+}
+
+static u8 icl_calc_voltage_level(int cdclk)
+{
+	if (cdclk > 556800)
+		return 2;
+	else if (cdclk > 312000)
+		return 1;
+	else
+		return 0;
+}
+
+static u8 ehl_calc_voltage_level(int cdclk)
+{
+	if (cdclk > 326400)
+		return 3;
+	else if (cdclk > 312000)
+		return 2;
+	else if (cdclk > 180000)
+		return 1;
+	else
+		return 0;
+}
+
+static u8 tgl_calc_voltage_level(int cdclk)
+{
+	if (cdclk > 556800)
+		return 3;
+	else if (cdclk > 326400)
+		return 2;
+	else if (cdclk > 312000)
+		return 1;
+	else
+		return 0;
+}
+
+static void icl_readout_refclk(struct drm_i915_private *dev_priv,
+			       struct intel_cdclk_config *cdclk_config)
+{
+	u32 dssm = intel_de_read(dev_priv, SKL_DSSM) & ICL_DSSM_CDCLK_PLL_REFCLK_MASK;
+
+	switch (dssm) {
+	default:
+		MISSING_CASE(dssm);
+		fallthrough;
+	case ICL_DSSM_CDCLK_PLL_REFCLK_24MHz:
+		cdclk_config->ref = 24000;
+		break;
+	case ICL_DSSM_CDCLK_PLL_REFCLK_19_2MHz:
+		cdclk_config->ref = 19200;
+		break;
+	case ICL_DSSM_CDCLK_PLL_REFCLK_38_4MHz:
+		cdclk_config->ref = 38400;
+		break;
+	}
+}
+
+static void bxt_de_pll_readout(struct drm_i915_private *dev_priv,
+			       struct intel_cdclk_config *cdclk_config)
+{
+	u32 val, ratio;
+
+	if (IS_DG2(dev_priv))
+		cdclk_config->ref = 38400;
+	else if (DISPLAY_VER(dev_priv) >= 11)
+		icl_readout_refclk(dev_priv, cdclk_config);
+	else
+		cdclk_config->ref = 19200;
+
+	val = intel_de_read(dev_priv, BXT_DE_PLL_ENABLE);
+	if ((val & BXT_DE_PLL_PLL_ENABLE) == 0 ||
+	    (val & BXT_DE_PLL_LOCK) == 0) {
+		/*
+		 * CDCLK PLL is disabled, the VCO/ratio doesn't matter, but
+		 * setting it to zero is a way to signal that.
+		 */
+		cdclk_config->vco = 0;
+		return;
+	}
+
+	/*
+	 * DISPLAY_VER >= 11 have the ratio directly in the PLL enable register,
+	 * gen9lp had it in a separate PLL control register.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 11)
+		ratio = val & ICL_CDCLK_PLL_RATIO_MASK;
+	else
+		ratio = intel_de_read(dev_priv, BXT_DE_PLL_CTL) & BXT_DE_PLL_RATIO_MASK;
+
+	cdclk_config->vco = ratio * cdclk_config->ref;
+}
+
+static void bxt_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_config *cdclk_config)
+{
+	u32 squash_ctl = 0;
+	u32 divider;
+	int div;
+
+	bxt_de_pll_readout(dev_priv, cdclk_config);
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		cdclk_config->bypass = cdclk_config->ref / 2;
+	else if (DISPLAY_VER(dev_priv) >= 11)
+		cdclk_config->bypass = 50000;
+	else
+		cdclk_config->bypass = cdclk_config->ref;
+
+	if (cdclk_config->vco == 0) {
+		cdclk_config->cdclk = cdclk_config->bypass;
+		goto out;
+	}
+
+	divider = intel_de_read(dev_priv, CDCLK_CTL) & BXT_CDCLK_CD2X_DIV_SEL_MASK;
+
+	switch (divider) {
+	case BXT_CDCLK_CD2X_DIV_SEL_1:
+		div = 2;
+		break;
+	case BXT_CDCLK_CD2X_DIV_SEL_1_5:
+		div = 3;
+		break;
+	case BXT_CDCLK_CD2X_DIV_SEL_2:
+		div = 4;
+		break;
+	case BXT_CDCLK_CD2X_DIV_SEL_4:
+		div = 8;
+		break;
+	default:
+		MISSING_CASE(divider);
+		return;
+	}
+
+	if (has_cdclk_squasher(dev_priv))
+		squash_ctl = intel_de_read(dev_priv, CDCLK_SQUASH_CTL);
+
+	if (squash_ctl & CDCLK_SQUASH_ENABLE) {
+		u16 waveform;
+		int size;
+
+		size = REG_FIELD_GET(CDCLK_SQUASH_WINDOW_SIZE_MASK, squash_ctl) + 1;
+		waveform = REG_FIELD_GET(CDCLK_SQUASH_WAVEFORM_MASK, squash_ctl) >> (16 - size);
+
+		cdclk_config->cdclk = DIV_ROUND_CLOSEST(hweight16(waveform) *
+							cdclk_config->vco, size * div);
+	} else {
+		cdclk_config->cdclk = DIV_ROUND_CLOSEST(cdclk_config->vco, div);
+	}
+
+ out:
+	/*
+	 * Can't read this out :( Let's assume it's
+	 * at least what the CDCLK frequency requires.
+	 */
+	cdclk_config->voltage_level =
+		intel_cdclk_calc_voltage_level(dev_priv, cdclk_config->cdclk);
+}
+
+static void bxt_de_pll_disable(struct drm_i915_private *dev_priv)
+{
+	intel_de_write(dev_priv, BXT_DE_PLL_ENABLE, 0);
+
+	/* Timeout 200us */
+	if (intel_de_wait_for_clear(dev_priv,
+				    BXT_DE_PLL_ENABLE, BXT_DE_PLL_LOCK, 1))
+		drm_err(&dev_priv->drm, "timeout waiting for DE PLL unlock\n");
+
+	dev_priv->display.cdclk.hw.vco = 0;
+}
+
+static void bxt_de_pll_enable(struct drm_i915_private *dev_priv, int vco)
+{
+	int ratio = DIV_ROUND_CLOSEST(vco, dev_priv->display.cdclk.hw.ref);
+
+	intel_de_rmw(dev_priv, BXT_DE_PLL_CTL,
+		     BXT_DE_PLL_RATIO_MASK, BXT_DE_PLL_RATIO(ratio));
+
+	intel_de_write(dev_priv, BXT_DE_PLL_ENABLE, BXT_DE_PLL_PLL_ENABLE);
+
+	/* Timeout 200us */
+	if (intel_de_wait_for_set(dev_priv,
+				  BXT_DE_PLL_ENABLE, BXT_DE_PLL_LOCK, 1))
+		drm_err(&dev_priv->drm, "timeout waiting for DE PLL lock\n");
+
+	dev_priv->display.cdclk.hw.vco = vco;
+}
+
+static void icl_cdclk_pll_disable(struct drm_i915_private *dev_priv)
+{
+	intel_de_rmw(dev_priv, BXT_DE_PLL_ENABLE,
+		     BXT_DE_PLL_PLL_ENABLE, 0);
+
+	/* Timeout 200us */
+	if (intel_de_wait_for_clear(dev_priv, BXT_DE_PLL_ENABLE, BXT_DE_PLL_LOCK, 1))
+		drm_err(&dev_priv->drm, "timeout waiting for CDCLK PLL unlock\n");
+
+	dev_priv->display.cdclk.hw.vco = 0;
+}
+
+static void icl_cdclk_pll_enable(struct drm_i915_private *dev_priv, int vco)
+{
+	int ratio = DIV_ROUND_CLOSEST(vco, dev_priv->display.cdclk.hw.ref);
+	u32 val;
+
+	val = ICL_CDCLK_PLL_RATIO(ratio);
+	intel_de_write(dev_priv, BXT_DE_PLL_ENABLE, val);
+
+	val |= BXT_DE_PLL_PLL_ENABLE;
+	intel_de_write(dev_priv, BXT_DE_PLL_ENABLE, val);
+
+	/* Timeout 200us */
+	if (intel_de_wait_for_set(dev_priv, BXT_DE_PLL_ENABLE, BXT_DE_PLL_LOCK, 1))
+		drm_err(&dev_priv->drm, "timeout waiting for CDCLK PLL lock\n");
+
+	dev_priv->display.cdclk.hw.vco = vco;
+}
+
+static void adlp_cdclk_pll_crawl(struct drm_i915_private *dev_priv, int vco)
+{
+	int ratio = DIV_ROUND_CLOSEST(vco, dev_priv->display.cdclk.hw.ref);
+	u32 val;
+
+	/* Write PLL ratio without disabling */
+	val = ICL_CDCLK_PLL_RATIO(ratio) | BXT_DE_PLL_PLL_ENABLE;
+	intel_de_write(dev_priv, BXT_DE_PLL_ENABLE, val);
+
+	/* Submit freq change request */
+	val |= BXT_DE_PLL_FREQ_REQ;
+	intel_de_write(dev_priv, BXT_DE_PLL_ENABLE, val);
+
+	/* Timeout 200us */
+	if (intel_de_wait_for_set(dev_priv, BXT_DE_PLL_ENABLE,
+				  BXT_DE_PLL_LOCK | BXT_DE_PLL_FREQ_REQ_ACK, 1))
+		drm_err(&dev_priv->drm, "timeout waiting for FREQ change request ack\n");
+
+	val &= ~BXT_DE_PLL_FREQ_REQ;
+	intel_de_write(dev_priv, BXT_DE_PLL_ENABLE, val);
+
+	dev_priv->display.cdclk.hw.vco = vco;
+}
+
+static u32 bxt_cdclk_cd2x_pipe(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		if (pipe == INVALID_PIPE)
+			return TGL_CDCLK_CD2X_PIPE_NONE;
+		else
+			return TGL_CDCLK_CD2X_PIPE(pipe);
+	} else if (DISPLAY_VER(dev_priv) >= 11) {
+		if (pipe == INVALID_PIPE)
+			return ICL_CDCLK_CD2X_PIPE_NONE;
+		else
+			return ICL_CDCLK_CD2X_PIPE(pipe);
+	} else {
+		if (pipe == INVALID_PIPE)
+			return BXT_CDCLK_CD2X_PIPE_NONE;
+		else
+			return BXT_CDCLK_CD2X_PIPE(pipe);
+	}
+}
+
+static u32 bxt_cdclk_cd2x_div_sel(struct drm_i915_private *dev_priv,
+				  int cdclk, int vco)
+{
+	/* cdclk = vco / 2 / div{1,1.5,2,4} */
+	switch (DIV_ROUND_CLOSEST(vco, cdclk)) {
+	default:
+		drm_WARN_ON(&dev_priv->drm,
+			    cdclk != dev_priv->display.cdclk.hw.bypass);
+		drm_WARN_ON(&dev_priv->drm, vco != 0);
+		fallthrough;
+	case 2:
+		return BXT_CDCLK_CD2X_DIV_SEL_1;
+	case 3:
+		return BXT_CDCLK_CD2X_DIV_SEL_1_5;
+	case 4:
+		return BXT_CDCLK_CD2X_DIV_SEL_2;
+	case 8:
+		return BXT_CDCLK_CD2X_DIV_SEL_4;
+	}
+}
+
+static u32 cdclk_squash_waveform(struct drm_i915_private *dev_priv,
+				 int cdclk)
+{
+	const struct intel_cdclk_vals *table = dev_priv->display.cdclk.table;
+	int i;
+
+	if (cdclk == dev_priv->display.cdclk.hw.bypass)
+		return 0;
+
+	for (i = 0; table[i].refclk; i++)
+		if (table[i].refclk == dev_priv->display.cdclk.hw.ref &&
+		    table[i].cdclk == cdclk)
+			return table[i].waveform;
+
+	drm_WARN(&dev_priv->drm, 1, "cdclk %d not valid for refclk %u\n",
+		 cdclk, dev_priv->display.cdclk.hw.ref);
+
+	return 0xffff;
+}
+
+static void bxt_set_cdclk(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_config *cdclk_config,
+			  enum pipe pipe)
+{
+	int cdclk = cdclk_config->cdclk;
+	int vco = cdclk_config->vco;
+	u32 val;
+	u16 waveform;
+	int clock;
+	int ret;
+
+	/* Inform power controller of upcoming frequency change. */
+	if (DISPLAY_VER(dev_priv) >= 11)
+		ret = skl_pcode_request(&dev_priv->uncore, SKL_PCODE_CDCLK_CONTROL,
+					SKL_CDCLK_PREPARE_FOR_CHANGE,
+					SKL_CDCLK_READY_FOR_CHANGE,
+					SKL_CDCLK_READY_FOR_CHANGE, 3);
+	else
+		/*
+		 * BSpec requires us to wait up to 150usec, but that leads to
+		 * timeouts; the 2ms used here is based on experiment.
+		 */
+		ret = snb_pcode_write_timeout(&dev_priv->uncore,
+					      HSW_PCODE_DE_WRITE_FREQ_REQ,
+					      0x80000000, 150, 2);
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"Failed to inform PCU about cdclk change (err %d, freq %d)\n",
+			ret, cdclk);
+		return;
+	}
+
+	if (HAS_CDCLK_CRAWL(dev_priv) && dev_priv->display.cdclk.hw.vco > 0 && vco > 0) {
+		if (dev_priv->display.cdclk.hw.vco != vco)
+			adlp_cdclk_pll_crawl(dev_priv, vco);
+	} else if (DISPLAY_VER(dev_priv) >= 11) {
+		if (dev_priv->display.cdclk.hw.vco != 0 &&
+		    dev_priv->display.cdclk.hw.vco != vco)
+			icl_cdclk_pll_disable(dev_priv);
+
+		if (dev_priv->display.cdclk.hw.vco != vco)
+			icl_cdclk_pll_enable(dev_priv, vco);
+	} else {
+		if (dev_priv->display.cdclk.hw.vco != 0 &&
+		    dev_priv->display.cdclk.hw.vco != vco)
+			bxt_de_pll_disable(dev_priv);
+
+		if (dev_priv->display.cdclk.hw.vco != vco)
+			bxt_de_pll_enable(dev_priv, vco);
+	}
+
+	waveform = cdclk_squash_waveform(dev_priv, cdclk);
+
+	if (waveform)
+		clock = vco / 2;
+	else
+		clock = cdclk;
+
+	if (has_cdclk_squasher(dev_priv)) {
+		u32 squash_ctl = 0;
+
+		if (waveform)
+			squash_ctl = CDCLK_SQUASH_ENABLE |
+				CDCLK_SQUASH_WINDOW_SIZE(0xf) | waveform;
+
+		intel_de_write(dev_priv, CDCLK_SQUASH_CTL, squash_ctl);
+	}
+
+	val = bxt_cdclk_cd2x_div_sel(dev_priv, clock, vco) |
+		bxt_cdclk_cd2x_pipe(dev_priv, pipe) |
+		skl_cdclk_decimal(cdclk);
+
+	/*
+	 * Disable SSA Precharge when CD clock frequency < 500 MHz,
+	 * enable otherwise.
+	 */
+	if ((IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) &&
+	    cdclk >= 500000)
+		val |= BXT_CDCLK_SSA_PRECHARGE_ENABLE;
+	intel_de_write(dev_priv, CDCLK_CTL, val);
+
+	if (pipe != INVALID_PIPE)
+		intel_crtc_wait_for_next_vblank(intel_crtc_for_pipe(dev_priv, pipe));
+
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		ret = snb_pcode_write(&dev_priv->uncore, SKL_PCODE_CDCLK_CONTROL,
+				      cdclk_config->voltage_level);
+	} else {
+		/*
+		 * The timeout isn't specified, the 2ms used here is based on
+		 * experiment.
+		 * FIXME: Waiting for the request completion could be delayed
+		 * until the next PCODE request based on BSpec.
+		 */
+		ret = snb_pcode_write_timeout(&dev_priv->uncore,
+					      HSW_PCODE_DE_WRITE_FREQ_REQ,
+					      cdclk_config->voltage_level,
+					      150, 2);
+	}
+
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"PCode CDCLK freq set failed, (err %d, freq %d)\n",
+			ret, cdclk);
+		return;
+	}
+
+	intel_update_cdclk(dev_priv);
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		/*
+		 * Can't read out the voltage level :(
+		 * Let's just assume everything is as expected.
+		 */
+		dev_priv->display.cdclk.hw.voltage_level = cdclk_config->voltage_level;
+}
+
+static void bxt_sanitize_cdclk(struct drm_i915_private *dev_priv)
+{
+	u32 cdctl, expected;
+	int cdclk, clock, vco;
+
+	intel_update_cdclk(dev_priv);
+	intel_cdclk_dump_config(dev_priv, &dev_priv->display.cdclk.hw, "Current CDCLK");
+
+	if (dev_priv->display.cdclk.hw.vco == 0 ||
+	    dev_priv->display.cdclk.hw.cdclk == dev_priv->display.cdclk.hw.bypass)
+		goto sanitize;
+
+	/* DPLL okay; verify the cdclock
+	 *
+	 * Some BIOS versions leave an incorrect decimal frequency value and
+	 * set reserved MBZ bits in CDCLK_CTL at least during exiting from S4,
+	 * so sanitize this register.
+	 */
+	cdctl = intel_de_read(dev_priv, CDCLK_CTL);
+	/*
+	 * Let's ignore the pipe field, since BIOS could have configured the
+	 * dividers both synching to an active pipe, or asynchronously
+	 * (PIPE_NONE).
+	 */
+	cdctl &= ~bxt_cdclk_cd2x_pipe(dev_priv, INVALID_PIPE);
+
+	/* Make sure this is a legal cdclk value for the platform */
+	cdclk = bxt_calc_cdclk(dev_priv, dev_priv->display.cdclk.hw.cdclk);
+	if (cdclk != dev_priv->display.cdclk.hw.cdclk)
+		goto sanitize;
+
+	/* Make sure the VCO is correct for the cdclk */
+	vco = bxt_calc_cdclk_pll_vco(dev_priv, cdclk);
+	if (vco != dev_priv->display.cdclk.hw.vco)
+		goto sanitize;
+
+	expected = skl_cdclk_decimal(cdclk);
+
+	/* Figure out what CD2X divider we should be using for this cdclk */
+	if (has_cdclk_squasher(dev_priv))
+		clock = dev_priv->display.cdclk.hw.vco / 2;
+	else
+		clock = dev_priv->display.cdclk.hw.cdclk;
+
+	expected |= bxt_cdclk_cd2x_div_sel(dev_priv, clock,
+					   dev_priv->display.cdclk.hw.vco);
+
+	/*
+	 * Disable SSA Precharge when CD clock frequency < 500 MHz,
+	 * enable otherwise.
+	 */
+	if ((IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) &&
+	    dev_priv->display.cdclk.hw.cdclk >= 500000)
+		expected |= BXT_CDCLK_SSA_PRECHARGE_ENABLE;
+
+	if (cdctl == expected)
+		/* All well; nothing to sanitize */
+		return;
+
+sanitize:
+	drm_dbg_kms(&dev_priv->drm, "Sanitizing cdclk programmed by pre-os\n");
+
+	/* force cdclk programming */
+	dev_priv->display.cdclk.hw.cdclk = 0;
+
+	/* force full PLL disable + enable */
+	dev_priv->display.cdclk.hw.vco = -1;
+}
+
+static void bxt_cdclk_init_hw(struct drm_i915_private *dev_priv)
+{
+	struct intel_cdclk_config cdclk_config;
+
+	bxt_sanitize_cdclk(dev_priv);
+
+	if (dev_priv->display.cdclk.hw.cdclk != 0 &&
+	    dev_priv->display.cdclk.hw.vco != 0)
+		return;
+
+	cdclk_config = dev_priv->display.cdclk.hw;
+
+	/*
+	 * FIXME:
+	 * - The initial CDCLK needs to be read from VBT.
+	 *   Need to make this change after VBT has changes for BXT.
+	 */
+	cdclk_config.cdclk = bxt_calc_cdclk(dev_priv, 0);
+	cdclk_config.vco = bxt_calc_cdclk_pll_vco(dev_priv, cdclk_config.cdclk);
+	cdclk_config.voltage_level =
+		intel_cdclk_calc_voltage_level(dev_priv, cdclk_config.cdclk);
+
+	bxt_set_cdclk(dev_priv, &cdclk_config, INVALID_PIPE);
+}
+
+static void bxt_cdclk_uninit_hw(struct drm_i915_private *dev_priv)
+{
+	struct intel_cdclk_config cdclk_config = dev_priv->display.cdclk.hw;
+
+	cdclk_config.cdclk = cdclk_config.bypass;
+	cdclk_config.vco = 0;
+	cdclk_config.voltage_level =
+		intel_cdclk_calc_voltage_level(dev_priv, cdclk_config.cdclk);
+
+	bxt_set_cdclk(dev_priv, &cdclk_config, INVALID_PIPE);
+}
+
+/**
+ * intel_cdclk_init_hw - Initialize CDCLK hardware
+ * @i915: i915 device
+ *
+ * Initialize CDCLK. This consists mainly of initializing dev_priv->display.cdclk.hw and
+ * sanitizing the state of the hardware if needed. This is generally done only
+ * during the display core initialization sequence, after which the DMC will
+ * take care of turning CDCLK off/on as needed.
+ */
+void intel_cdclk_init_hw(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 10 || IS_BROXTON(i915))
+		bxt_cdclk_init_hw(i915);
+	else if (DISPLAY_VER(i915) == 9)
+		skl_cdclk_init_hw(i915);
+}
+
+/**
+ * intel_cdclk_uninit_hw - Uninitialize CDCLK hardware
+ * @i915: i915 device
+ *
+ * Uninitialize CDCLK. This is done only during the display core
+ * uninitialization sequence.
+ */
+void intel_cdclk_uninit_hw(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 10 || IS_BROXTON(i915))
+		bxt_cdclk_uninit_hw(i915);
+	else if (DISPLAY_VER(i915) == 9)
+		skl_cdclk_uninit_hw(i915);
+}
+
+static bool intel_cdclk_can_crawl(struct drm_i915_private *dev_priv,
+				  const struct intel_cdclk_config *a,
+				  const struct intel_cdclk_config *b)
+{
+	int a_div, b_div;
+
+	if (!HAS_CDCLK_CRAWL(dev_priv))
+		return false;
+
+	/*
+	 * The vco and cd2x divider will change independently
+	 * from each, so we disallow cd2x change when crawling.
+	 */
+	a_div = DIV_ROUND_CLOSEST(a->vco, a->cdclk);
+	b_div = DIV_ROUND_CLOSEST(b->vco, b->cdclk);
+
+	return a->vco != 0 && b->vco != 0 &&
+		a->vco != b->vco &&
+		a_div == b_div &&
+		a->ref == b->ref;
+}
+
+static bool intel_cdclk_can_squash(struct drm_i915_private *dev_priv,
+				   const struct intel_cdclk_config *a,
+				   const struct intel_cdclk_config *b)
+{
+	/*
+	 * FIXME should store a bit more state in intel_cdclk_config
+	 * to differentiate squasher vs. cd2x divider properly. For
+	 * the moment all platforms with squasher use a fixed cd2x
+	 * divider.
+	 */
+	if (!has_cdclk_squasher(dev_priv))
+		return false;
+
+	return a->cdclk != b->cdclk &&
+		a->vco != 0 &&
+		a->vco == b->vco &&
+		a->ref == b->ref;
+}
+
+/**
+ * intel_cdclk_needs_modeset - Determine if changong between the CDCLK
+ *                             configurations requires a modeset on all pipes
+ * @a: first CDCLK configuration
+ * @b: second CDCLK configuration
+ *
+ * Returns:
+ * True if changing between the two CDCLK configurations
+ * requires all pipes to be off, false if not.
+ */
+bool intel_cdclk_needs_modeset(const struct intel_cdclk_config *a,
+			       const struct intel_cdclk_config *b)
+{
+	return a->cdclk != b->cdclk ||
+		a->vco != b->vco ||
+		a->ref != b->ref;
+}
+
+/**
+ * intel_cdclk_can_cd2x_update - Determine if changing between the two CDCLK
+ *                               configurations requires only a cd2x divider update
+ * @dev_priv: i915 device
+ * @a: first CDCLK configuration
+ * @b: second CDCLK configuration
+ *
+ * Returns:
+ * True if changing between the two CDCLK configurations
+ * can be done with just a cd2x divider update, false if not.
+ */
+static bool intel_cdclk_can_cd2x_update(struct drm_i915_private *dev_priv,
+					const struct intel_cdclk_config *a,
+					const struct intel_cdclk_config *b)
+{
+	/* Older hw doesn't have the capability */
+	if (DISPLAY_VER(dev_priv) < 10 && !IS_BROXTON(dev_priv))
+		return false;
+
+	/*
+	 * FIXME should store a bit more state in intel_cdclk_config
+	 * to differentiate squasher vs. cd2x divider properly. For
+	 * the moment all platforms with squasher use a fixed cd2x
+	 * divider.
+	 */
+	if (has_cdclk_squasher(dev_priv))
+		return false;
+
+	return a->cdclk != b->cdclk &&
+		a->vco != 0 &&
+		a->vco == b->vco &&
+		a->ref == b->ref;
+}
+
+/**
+ * intel_cdclk_changed - Determine if two CDCLK configurations are different
+ * @a: first CDCLK configuration
+ * @b: second CDCLK configuration
+ *
+ * Returns:
+ * True if the CDCLK configurations don't match, false if they do.
+ */
+static bool intel_cdclk_changed(const struct intel_cdclk_config *a,
+				const struct intel_cdclk_config *b)
+{
+	return intel_cdclk_needs_modeset(a, b) ||
+		a->voltage_level != b->voltage_level;
+}
+
+void intel_cdclk_dump_config(struct drm_i915_private *i915,
+			     const struct intel_cdclk_config *cdclk_config,
+			     const char *context)
+{
+	drm_dbg_kms(&i915->drm, "%s %d kHz, VCO %d kHz, ref %d kHz, bypass %d kHz, voltage level %d\n",
+		    context, cdclk_config->cdclk, cdclk_config->vco,
+		    cdclk_config->ref, cdclk_config->bypass,
+		    cdclk_config->voltage_level);
+}
+
+/**
+ * intel_set_cdclk - Push the CDCLK configuration to the hardware
+ * @dev_priv: i915 device
+ * @cdclk_config: new CDCLK configuration
+ * @pipe: pipe with which to synchronize the update
+ *
+ * Program the hardware based on the passed in CDCLK state,
+ * if necessary.
+ */
+static void intel_set_cdclk(struct drm_i915_private *dev_priv,
+			    const struct intel_cdclk_config *cdclk_config,
+			    enum pipe pipe)
+{
+	struct intel_encoder *encoder;
+
+	if (!intel_cdclk_changed(&dev_priv->display.cdclk.hw, cdclk_config))
+		return;
+
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !dev_priv->display.funcs.cdclk->set_cdclk))
+		return;
+
+	intel_cdclk_dump_config(dev_priv, cdclk_config, "Changing CDCLK to");
+
+	for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		intel_psr_pause(intel_dp);
+	}
+
+	intel_audio_cdclk_change_pre(dev_priv);
+
+	/*
+	 * Lock aux/gmbus while we change cdclk in case those
+	 * functions use cdclk. Not all platforms/ports do,
+	 * but we'll lock them all for simplicity.
+	 */
+	mutex_lock(&dev_priv->display.gmbus.mutex);
+	for_each_intel_dp(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		mutex_lock_nest_lock(&intel_dp->aux.hw_mutex,
+				     &dev_priv->display.gmbus.mutex);
+	}
+
+	intel_cdclk_set_cdclk(dev_priv, cdclk_config, pipe);
+
+	for_each_intel_dp(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		mutex_unlock(&intel_dp->aux.hw_mutex);
+	}
+	mutex_unlock(&dev_priv->display.gmbus.mutex);
+
+	for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		intel_psr_resume(intel_dp);
+	}
+
+	intel_audio_cdclk_change_post(dev_priv);
+
+	if (drm_WARN(&dev_priv->drm,
+		     intel_cdclk_changed(&dev_priv->display.cdclk.hw, cdclk_config),
+		     "cdclk state doesn't match!\n")) {
+		intel_cdclk_dump_config(dev_priv, &dev_priv->display.cdclk.hw, "[hw state]");
+		intel_cdclk_dump_config(dev_priv, cdclk_config, "[sw state]");
+	}
+}
+
+/**
+ * intel_set_cdclk_pre_plane_update - Push the CDCLK state to the hardware
+ * @state: intel atomic state
+ *
+ * Program the hardware before updating the HW plane state based on the
+ * new CDCLK state, if necessary.
+ */
+void
+intel_set_cdclk_pre_plane_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_cdclk_state *old_cdclk_state =
+		intel_atomic_get_old_cdclk_state(state);
+	const struct intel_cdclk_state *new_cdclk_state =
+		intel_atomic_get_new_cdclk_state(state);
+	enum pipe pipe = new_cdclk_state->pipe;
+
+	if (!intel_cdclk_changed(&old_cdclk_state->actual,
+				 &new_cdclk_state->actual))
+		return;
+
+	if (pipe == INVALID_PIPE ||
+	    old_cdclk_state->actual.cdclk <= new_cdclk_state->actual.cdclk) {
+		drm_WARN_ON(&dev_priv->drm, !new_cdclk_state->base.changed);
+
+		intel_set_cdclk(dev_priv, &new_cdclk_state->actual, pipe);
+	}
+}
+
+/**
+ * intel_set_cdclk_post_plane_update - Push the CDCLK state to the hardware
+ * @state: intel atomic state
+ *
+ * Program the hardware after updating the HW plane state based on the
+ * new CDCLK state, if necessary.
+ */
+void
+intel_set_cdclk_post_plane_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_cdclk_state *old_cdclk_state =
+		intel_atomic_get_old_cdclk_state(state);
+	const struct intel_cdclk_state *new_cdclk_state =
+		intel_atomic_get_new_cdclk_state(state);
+	enum pipe pipe = new_cdclk_state->pipe;
+
+	if (!intel_cdclk_changed(&old_cdclk_state->actual,
+				 &new_cdclk_state->actual))
+		return;
+
+	if (pipe != INVALID_PIPE &&
+	    old_cdclk_state->actual.cdclk > new_cdclk_state->actual.cdclk) {
+		drm_WARN_ON(&dev_priv->drm, !new_cdclk_state->base.changed);
+
+		intel_set_cdclk(dev_priv, &new_cdclk_state->actual, pipe);
+	}
+}
+
+static int intel_pixel_rate_to_cdclk(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	int pixel_rate = crtc_state->pixel_rate;
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		return DIV_ROUND_UP(pixel_rate, 2);
+	else if (DISPLAY_VER(dev_priv) == 9 ||
+		 IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		return pixel_rate;
+	else if (IS_CHERRYVIEW(dev_priv))
+		return DIV_ROUND_UP(pixel_rate * 100, 95);
+	else if (crtc_state->double_wide)
+		return DIV_ROUND_UP(pixel_rate * 100, 90 * 2);
+	else
+		return DIV_ROUND_UP(pixel_rate * 100, 90);
+}
+
+static int intel_planes_min_cdclk(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_plane *plane;
+	int min_cdclk = 0;
+
+	for_each_intel_plane_on_crtc(&dev_priv->drm, crtc, plane)
+		min_cdclk = max(crtc_state->min_cdclk[plane->id], min_cdclk);
+
+	return min_cdclk;
+}
+
+int intel_crtc_compute_min_cdclk(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(crtc_state->uapi.crtc->dev);
+	int min_cdclk;
+
+	if (!crtc_state->hw.enable)
+		return 0;
+
+	min_cdclk = intel_pixel_rate_to_cdclk(crtc_state);
+
+	/* pixel rate mustn't exceed 95% of cdclk with IPS on BDW */
+	if (IS_BROADWELL(dev_priv) && hsw_crtc_state_ips_capable(crtc_state))
+		min_cdclk = DIV_ROUND_UP(min_cdclk * 100, 95);
+
+	/* BSpec says "Do not use DisplayPort with CDCLK less than 432 MHz,
+	 * audio enabled, port width x4, and link rate HBR2 (5.4 GHz), or else
+	 * there may be audio corruption or screen corruption." This cdclk
+	 * restriction for GLK is 316.8 MHz.
+	 */
+	if (intel_crtc_has_dp_encoder(crtc_state) &&
+	    crtc_state->has_audio &&
+	    crtc_state->port_clock >= 540000 &&
+	    crtc_state->lane_count == 4) {
+		if (DISPLAY_VER(dev_priv) == 10) {
+			/* Display WA #1145: glk */
+			min_cdclk = max(316800, min_cdclk);
+		} else if (DISPLAY_VER(dev_priv) == 9 || IS_BROADWELL(dev_priv)) {
+			/* Display WA #1144: skl,bxt */
+			min_cdclk = max(432000, min_cdclk);
+		}
+	}
+
+	/*
+	 * According to BSpec, "The CD clock frequency must be at least twice
+	 * the frequency of the Azalia BCLK." and BCLK is 96 MHz by default.
+	 */
+	if (crtc_state->has_audio && DISPLAY_VER(dev_priv) >= 9)
+		min_cdclk = max(2 * 96000, min_cdclk);
+
+	/*
+	 * "For DP audio configuration, cdclk frequency shall be set to
+	 *  meet the following requirements:
+	 *  DP Link Frequency(MHz) | Cdclk frequency(MHz)
+	 *  270                    | 320 or higher
+	 *  162                    | 200 or higher"
+	 */
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    intel_crtc_has_dp_encoder(crtc_state) && crtc_state->has_audio)
+		min_cdclk = max(crtc_state->port_clock, min_cdclk);
+
+	/*
+	 * On Valleyview some DSI panels lose (v|h)sync when the clock is lower
+	 * than 320000KHz.
+	 */
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI) &&
+	    IS_VALLEYVIEW(dev_priv))
+		min_cdclk = max(320000, min_cdclk);
+
+	/*
+	 * On Geminilake once the CDCLK gets as low as 79200
+	 * picture gets unstable, despite that values are
+	 * correct for DSI PLL and DE PLL.
+	 */
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI) &&
+	    IS_GEMINILAKE(dev_priv))
+		min_cdclk = max(158400, min_cdclk);
+
+	/* Account for additional needs from the planes */
+	min_cdclk = max(intel_planes_min_cdclk(crtc_state), min_cdclk);
+
+	/*
+	 * When we decide to use only one VDSC engine, since
+	 * each VDSC operates with 1 ppc throughput, pixel clock
+	 * cannot be higher than the VDSC clock (cdclk)
+	 */
+	if (crtc_state->dsc.compression_enable && !crtc_state->dsc.dsc_split)
+		min_cdclk = max(min_cdclk, (int)crtc_state->pixel_rate);
+
+	/*
+	 * HACK. Currently for TGL/DG2 platforms we calculate
+	 * min_cdclk initially based on pixel_rate divided
+	 * by 2, accounting for also plane requirements,
+	 * however in some cases the lowest possible CDCLK
+	 * doesn't work and causing the underruns.
+	 * Explicitly stating here that this seems to be currently
+	 * rather a Hack, than final solution.
+	 */
+	if (IS_TIGERLAKE(dev_priv) || IS_DG2(dev_priv)) {
+		/*
+		 * Clamp to max_cdclk_freq in case pixel rate is higher,
+		 * in order not to break an 8K, but still leave W/A at place.
+		 */
+		min_cdclk = max_t(int, min_cdclk,
+				  min_t(int, crtc_state->pixel_rate,
+					dev_priv->display.cdclk.max_cdclk_freq));
+	}
+
+	return min_cdclk;
+}
+
+static int intel_compute_min_cdclk(struct intel_cdclk_state *cdclk_state)
+{
+	struct intel_atomic_state *state = cdclk_state->base.state;
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_bw_state *bw_state;
+	struct intel_crtc *crtc;
+	struct intel_crtc_state *crtc_state;
+	int min_cdclk, i;
+	enum pipe pipe;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		int ret;
+
+		min_cdclk = intel_crtc_compute_min_cdclk(crtc_state);
+		if (min_cdclk < 0)
+			return min_cdclk;
+
+		if (cdclk_state->min_cdclk[crtc->pipe] == min_cdclk)
+			continue;
+
+		cdclk_state->min_cdclk[crtc->pipe] = min_cdclk;
+
+		ret = intel_atomic_lock_global_state(&cdclk_state->base);
+		if (ret)
+			return ret;
+	}
+
+	bw_state = intel_atomic_get_new_bw_state(state);
+	if (bw_state) {
+		min_cdclk = intel_bw_min_cdclk(dev_priv, bw_state);
+
+		if (cdclk_state->bw_min_cdclk != min_cdclk) {
+			int ret;
+
+			cdclk_state->bw_min_cdclk = min_cdclk;
+
+			ret = intel_atomic_lock_global_state(&cdclk_state->base);
+			if (ret)
+				return ret;
+		}
+	}
+
+	min_cdclk = max(cdclk_state->force_min_cdclk,
+			cdclk_state->bw_min_cdclk);
+	for_each_pipe(dev_priv, pipe)
+		min_cdclk = max(cdclk_state->min_cdclk[pipe], min_cdclk);
+
+	/*
+	 * Avoid glk_force_audio_cdclk() causing excessive screen
+	 * blinking when multiple pipes are active by making sure
+	 * CDCLK frequency is always high enough for audio. With a
+	 * single active pipe we can always change CDCLK frequency
+	 * by changing the cd2x divider (see glk_cdclk_table[]) and
+	 * thus a full modeset won't be needed then.
+	 */
+	if (IS_GEMINILAKE(dev_priv) && cdclk_state->active_pipes &&
+	    !is_power_of_2(cdclk_state->active_pipes))
+		min_cdclk = max(2 * 96000, min_cdclk);
+
+	if (min_cdclk > dev_priv->display.cdclk.max_cdclk_freq) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "required cdclk (%d kHz) exceeds max (%d kHz)\n",
+			    min_cdclk, dev_priv->display.cdclk.max_cdclk_freq);
+		return -EINVAL;
+	}
+
+	return min_cdclk;
+}
+
+/*
+ * Account for port clock min voltage level requirements.
+ * This only really does something on DISPLA_VER >= 11 but can be
+ * called on earlier platforms as well.
+ *
+ * Note that this functions assumes that 0 is
+ * the lowest voltage value, and higher values
+ * correspond to increasingly higher voltages.
+ *
+ * Should that relationship no longer hold on
+ * future platforms this code will need to be
+ * adjusted.
+ */
+static int bxt_compute_min_voltage_level(struct intel_cdclk_state *cdclk_state)
+{
+	struct intel_atomic_state *state = cdclk_state->base.state;
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc *crtc;
+	struct intel_crtc_state *crtc_state;
+	u8 min_voltage_level;
+	int i;
+	enum pipe pipe;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		int ret;
+
+		if (crtc_state->hw.enable)
+			min_voltage_level = crtc_state->min_voltage_level;
+		else
+			min_voltage_level = 0;
+
+		if (cdclk_state->min_voltage_level[crtc->pipe] == min_voltage_level)
+			continue;
+
+		cdclk_state->min_voltage_level[crtc->pipe] = min_voltage_level;
+
+		ret = intel_atomic_lock_global_state(&cdclk_state->base);
+		if (ret)
+			return ret;
+	}
+
+	min_voltage_level = 0;
+	for_each_pipe(dev_priv, pipe)
+		min_voltage_level = max(cdclk_state->min_voltage_level[pipe],
+					min_voltage_level);
+
+	return min_voltage_level;
+}
+
+static int vlv_modeset_calc_cdclk(struct intel_cdclk_state *cdclk_state)
+{
+	struct intel_atomic_state *state = cdclk_state->base.state;
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	int min_cdclk, cdclk;
+
+	min_cdclk = intel_compute_min_cdclk(cdclk_state);
+	if (min_cdclk < 0)
+		return min_cdclk;
+
+	cdclk = vlv_calc_cdclk(dev_priv, min_cdclk);
+
+	cdclk_state->logical.cdclk = cdclk;
+	cdclk_state->logical.voltage_level =
+		vlv_calc_voltage_level(dev_priv, cdclk);
+
+	if (!cdclk_state->active_pipes) {
+		cdclk = vlv_calc_cdclk(dev_priv, cdclk_state->force_min_cdclk);
+
+		cdclk_state->actual.cdclk = cdclk;
+		cdclk_state->actual.voltage_level =
+			vlv_calc_voltage_level(dev_priv, cdclk);
+	} else {
+		cdclk_state->actual = cdclk_state->logical;
+	}
+
+	return 0;
+}
+
+static int bdw_modeset_calc_cdclk(struct intel_cdclk_state *cdclk_state)
+{
+	int min_cdclk, cdclk;
+
+	min_cdclk = intel_compute_min_cdclk(cdclk_state);
+	if (min_cdclk < 0)
+		return min_cdclk;
+
+	/*
+	 * FIXME should also account for plane ratio
+	 * once 64bpp pixel formats are supported.
+	 */
+	cdclk = bdw_calc_cdclk(min_cdclk);
+
+	cdclk_state->logical.cdclk = cdclk;
+	cdclk_state->logical.voltage_level =
+		bdw_calc_voltage_level(cdclk);
+
+	if (!cdclk_state->active_pipes) {
+		cdclk = bdw_calc_cdclk(cdclk_state->force_min_cdclk);
+
+		cdclk_state->actual.cdclk = cdclk;
+		cdclk_state->actual.voltage_level =
+			bdw_calc_voltage_level(cdclk);
+	} else {
+		cdclk_state->actual = cdclk_state->logical;
+	}
+
+	return 0;
+}
+
+static int skl_dpll0_vco(struct intel_cdclk_state *cdclk_state)
+{
+	struct intel_atomic_state *state = cdclk_state->base.state;
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc *crtc;
+	struct intel_crtc_state *crtc_state;
+	int vco, i;
+
+	vco = cdclk_state->logical.vco;
+	if (!vco)
+		vco = dev_priv->skl_preferred_vco_freq;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		if (!crtc_state->hw.enable)
+			continue;
+
+		if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP))
+			continue;
+
+		/*
+		 * DPLL0 VCO may need to be adjusted to get the correct
+		 * clock for eDP. This will affect cdclk as well.
+		 */
+		switch (crtc_state->port_clock / 2) {
+		case 108000:
+		case 216000:
+			vco = 8640000;
+			break;
+		default:
+			vco = 8100000;
+			break;
+		}
+	}
+
+	return vco;
+}
+
+static int skl_modeset_calc_cdclk(struct intel_cdclk_state *cdclk_state)
+{
+	int min_cdclk, cdclk, vco;
+
+	min_cdclk = intel_compute_min_cdclk(cdclk_state);
+	if (min_cdclk < 0)
+		return min_cdclk;
+
+	vco = skl_dpll0_vco(cdclk_state);
+
+	/*
+	 * FIXME should also account for plane ratio
+	 * once 64bpp pixel formats are supported.
+	 */
+	cdclk = skl_calc_cdclk(min_cdclk, vco);
+
+	cdclk_state->logical.vco = vco;
+	cdclk_state->logical.cdclk = cdclk;
+	cdclk_state->logical.voltage_level =
+		skl_calc_voltage_level(cdclk);
+
+	if (!cdclk_state->active_pipes) {
+		cdclk = skl_calc_cdclk(cdclk_state->force_min_cdclk, vco);
+
+		cdclk_state->actual.vco = vco;
+		cdclk_state->actual.cdclk = cdclk;
+		cdclk_state->actual.voltage_level =
+			skl_calc_voltage_level(cdclk);
+	} else {
+		cdclk_state->actual = cdclk_state->logical;
+	}
+
+	return 0;
+}
+
+static int bxt_modeset_calc_cdclk(struct intel_cdclk_state *cdclk_state)
+{
+	struct intel_atomic_state *state = cdclk_state->base.state;
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	int min_cdclk, min_voltage_level, cdclk, vco;
+
+	min_cdclk = intel_compute_min_cdclk(cdclk_state);
+	if (min_cdclk < 0)
+		return min_cdclk;
+
+	min_voltage_level = bxt_compute_min_voltage_level(cdclk_state);
+	if (min_voltage_level < 0)
+		return min_voltage_level;
+
+	cdclk = bxt_calc_cdclk(dev_priv, min_cdclk);
+	vco = bxt_calc_cdclk_pll_vco(dev_priv, cdclk);
+
+	cdclk_state->logical.vco = vco;
+	cdclk_state->logical.cdclk = cdclk;
+	cdclk_state->logical.voltage_level =
+		max_t(int, min_voltage_level,
+		      intel_cdclk_calc_voltage_level(dev_priv, cdclk));
+
+	if (!cdclk_state->active_pipes) {
+		cdclk = bxt_calc_cdclk(dev_priv, cdclk_state->force_min_cdclk);
+		vco = bxt_calc_cdclk_pll_vco(dev_priv, cdclk);
+
+		cdclk_state->actual.vco = vco;
+		cdclk_state->actual.cdclk = cdclk;
+		cdclk_state->actual.voltage_level =
+			intel_cdclk_calc_voltage_level(dev_priv, cdclk);
+	} else {
+		cdclk_state->actual = cdclk_state->logical;
+	}
+
+	return 0;
+}
+
+static int fixed_modeset_calc_cdclk(struct intel_cdclk_state *cdclk_state)
+{
+	int min_cdclk;
+
+	/*
+	 * We can't change the cdclk frequency, but we still want to
+	 * check that the required minimum frequency doesn't exceed
+	 * the actual cdclk frequency.
+	 */
+	min_cdclk = intel_compute_min_cdclk(cdclk_state);
+	if (min_cdclk < 0)
+		return min_cdclk;
+
+	return 0;
+}
+
+static struct intel_global_state *intel_cdclk_duplicate_state(struct intel_global_obj *obj)
+{
+	struct intel_cdclk_state *cdclk_state;
+
+	cdclk_state = kmemdup(obj->state, sizeof(*cdclk_state), GFP_KERNEL);
+	if (!cdclk_state)
+		return NULL;
+
+	cdclk_state->pipe = INVALID_PIPE;
+
+	return &cdclk_state->base;
+}
+
+static void intel_cdclk_destroy_state(struct intel_global_obj *obj,
+				      struct intel_global_state *state)
+{
+	kfree(state);
+}
+
+static const struct intel_global_state_funcs intel_cdclk_funcs = {
+	.atomic_duplicate_state = intel_cdclk_duplicate_state,
+	.atomic_destroy_state = intel_cdclk_destroy_state,
+};
+
+struct intel_cdclk_state *
+intel_atomic_get_cdclk_state(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_global_state *cdclk_state;
+
+	cdclk_state = intel_atomic_get_global_obj_state(state, &dev_priv->display.cdclk.obj);
+	if (IS_ERR(cdclk_state))
+		return ERR_CAST(cdclk_state);
+
+	return to_intel_cdclk_state(cdclk_state);
+}
+
+int intel_cdclk_atomic_check(struct intel_atomic_state *state,
+			     bool *need_cdclk_calc)
+{
+	const struct intel_cdclk_state *old_cdclk_state;
+	const struct intel_cdclk_state *new_cdclk_state;
+	struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	int ret;
+	int i;
+
+	/*
+	 * active_planes bitmask has been updated, and potentially affected
+	 * planes are part of the state. We can now compute the minimum cdclk
+	 * for each plane.
+	 */
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		ret = intel_plane_calc_min_cdclk(state, plane, need_cdclk_calc);
+		if (ret)
+			return ret;
+	}
+
+	ret = intel_bw_calc_min_cdclk(state, need_cdclk_calc);
+	if (ret)
+		return ret;
+
+	old_cdclk_state = intel_atomic_get_old_cdclk_state(state);
+	new_cdclk_state = intel_atomic_get_new_cdclk_state(state);
+
+	if (new_cdclk_state &&
+	    old_cdclk_state->force_min_cdclk != new_cdclk_state->force_min_cdclk)
+		*need_cdclk_calc = true;
+
+	return 0;
+}
+
+int intel_cdclk_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_cdclk_state *cdclk_state;
+
+	cdclk_state = kzalloc(sizeof(*cdclk_state), GFP_KERNEL);
+	if (!cdclk_state)
+		return -ENOMEM;
+
+	intel_atomic_global_obj_init(dev_priv, &dev_priv->display.cdclk.obj,
+				     &cdclk_state->base, &intel_cdclk_funcs);
+
+	return 0;
+}
+
+int intel_modeset_calc_cdclk(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_cdclk_state *old_cdclk_state;
+	struct intel_cdclk_state *new_cdclk_state;
+	enum pipe pipe = INVALID_PIPE;
+	int ret;
+
+	new_cdclk_state = intel_atomic_get_cdclk_state(state);
+	if (IS_ERR(new_cdclk_state))
+		return PTR_ERR(new_cdclk_state);
+
+	old_cdclk_state = intel_atomic_get_old_cdclk_state(state);
+
+	new_cdclk_state->active_pipes =
+		intel_calc_active_pipes(state, old_cdclk_state->active_pipes);
+
+	ret = intel_cdclk_modeset_calc_cdclk(dev_priv, new_cdclk_state);
+	if (ret)
+		return ret;
+
+	if (intel_cdclk_changed(&old_cdclk_state->actual,
+				&new_cdclk_state->actual)) {
+		/*
+		 * Also serialize commits across all crtcs
+		 * if the actual hw needs to be poked.
+		 */
+		ret = intel_atomic_serialize_global_state(&new_cdclk_state->base);
+		if (ret)
+			return ret;
+	} else if (old_cdclk_state->active_pipes != new_cdclk_state->active_pipes ||
+		   old_cdclk_state->force_min_cdclk != new_cdclk_state->force_min_cdclk ||
+		   intel_cdclk_changed(&old_cdclk_state->logical,
+				       &new_cdclk_state->logical)) {
+		ret = intel_atomic_lock_global_state(&new_cdclk_state->base);
+		if (ret)
+			return ret;
+	} else {
+		return 0;
+	}
+
+	if (is_power_of_2(new_cdclk_state->active_pipes) &&
+	    intel_cdclk_can_cd2x_update(dev_priv,
+					&old_cdclk_state->actual,
+					&new_cdclk_state->actual)) {
+		struct intel_crtc *crtc;
+		struct intel_crtc_state *crtc_state;
+
+		pipe = ilog2(new_cdclk_state->active_pipes);
+		crtc = intel_crtc_for_pipe(dev_priv, pipe);
+
+		crtc_state = intel_atomic_get_crtc_state(&state->base, crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+
+		if (drm_atomic_crtc_needs_modeset(&crtc_state->uapi))
+			pipe = INVALID_PIPE;
+	}
+
+	if (intel_cdclk_can_squash(dev_priv,
+				   &old_cdclk_state->actual,
+				   &new_cdclk_state->actual)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Can change cdclk via squasher\n");
+	} else if (intel_cdclk_can_crawl(dev_priv,
+					 &old_cdclk_state->actual,
+					 &new_cdclk_state->actual)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Can change cdclk via crawl\n");
+	} else if (pipe != INVALID_PIPE) {
+		new_cdclk_state->pipe = pipe;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "Can change cdclk cd2x divider with pipe %c active\n",
+			    pipe_name(pipe));
+	} else if (intel_cdclk_needs_modeset(&old_cdclk_state->actual,
+					     &new_cdclk_state->actual)) {
+		/* All pipes must be switched off while we change the cdclk. */
+		ret = intel_modeset_all_pipes(state);
+		if (ret)
+			return ret;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "Modeset required for cdclk change\n");
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "New cdclk calculated to be logical %u kHz, actual %u kHz\n",
+		    new_cdclk_state->logical.cdclk,
+		    new_cdclk_state->actual.cdclk);
+	drm_dbg_kms(&dev_priv->drm,
+		    "New voltage level calculated to be logical %u, actual %u\n",
+		    new_cdclk_state->logical.voltage_level,
+		    new_cdclk_state->actual.voltage_level);
+
+	return 0;
+}
+
+static int intel_compute_max_dotclk(struct drm_i915_private *dev_priv)
+{
+	int max_cdclk_freq = dev_priv->display.cdclk.max_cdclk_freq;
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		return 2 * max_cdclk_freq;
+	else if (DISPLAY_VER(dev_priv) == 9 ||
+		 IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		return max_cdclk_freq;
+	else if (IS_CHERRYVIEW(dev_priv))
+		return max_cdclk_freq*95/100;
+	else if (DISPLAY_VER(dev_priv) < 4)
+		return 2*max_cdclk_freq*90/100;
+	else
+		return max_cdclk_freq*90/100;
+}
+
+/**
+ * intel_update_max_cdclk - Determine the maximum support CDCLK frequency
+ * @dev_priv: i915 device
+ *
+ * Determine the maximum CDCLK frequency the platform supports, and also
+ * derive the maximum dot clock frequency the maximum CDCLK frequency
+ * allows.
+ */
+void intel_update_max_cdclk(struct drm_i915_private *dev_priv)
+{
+	if (IS_JSL_EHL(dev_priv)) {
+		if (dev_priv->display.cdclk.hw.ref == 24000)
+			dev_priv->display.cdclk.max_cdclk_freq = 552000;
+		else
+			dev_priv->display.cdclk.max_cdclk_freq = 556800;
+	} else if (DISPLAY_VER(dev_priv) >= 11) {
+		if (dev_priv->display.cdclk.hw.ref == 24000)
+			dev_priv->display.cdclk.max_cdclk_freq = 648000;
+		else
+			dev_priv->display.cdclk.max_cdclk_freq = 652800;
+	} else if (IS_GEMINILAKE(dev_priv)) {
+		dev_priv->display.cdclk.max_cdclk_freq = 316800;
+	} else if (IS_BROXTON(dev_priv)) {
+		dev_priv->display.cdclk.max_cdclk_freq = 624000;
+	} else if (DISPLAY_VER(dev_priv) == 9) {
+		u32 limit = intel_de_read(dev_priv, SKL_DFSM) & SKL_DFSM_CDCLK_LIMIT_MASK;
+		int max_cdclk, vco;
+
+		vco = dev_priv->skl_preferred_vco_freq;
+		drm_WARN_ON(&dev_priv->drm, vco != 8100000 && vco != 8640000);
+
+		/*
+		 * Use the lower (vco 8640) cdclk values as a
+		 * first guess. skl_calc_cdclk() will correct it
+		 * if the preferred vco is 8100 instead.
+		 */
+		if (limit == SKL_DFSM_CDCLK_LIMIT_675)
+			max_cdclk = 617143;
+		else if (limit == SKL_DFSM_CDCLK_LIMIT_540)
+			max_cdclk = 540000;
+		else if (limit == SKL_DFSM_CDCLK_LIMIT_450)
+			max_cdclk = 432000;
+		else
+			max_cdclk = 308571;
+
+		dev_priv->display.cdclk.max_cdclk_freq = skl_calc_cdclk(max_cdclk, vco);
+	} else if (IS_BROADWELL(dev_priv))  {
+		/*
+		 * FIXME with extra cooling we can allow
+		 * 540 MHz for ULX and 675 Mhz for ULT.
+		 * How can we know if extra cooling is
+		 * available? PCI ID, VTB, something else?
+		 */
+		if (intel_de_read(dev_priv, FUSE_STRAP) & HSW_CDCLK_LIMIT)
+			dev_priv->display.cdclk.max_cdclk_freq = 450000;
+		else if (IS_BDW_ULX(dev_priv))
+			dev_priv->display.cdclk.max_cdclk_freq = 450000;
+		else if (IS_BDW_ULT(dev_priv))
+			dev_priv->display.cdclk.max_cdclk_freq = 540000;
+		else
+			dev_priv->display.cdclk.max_cdclk_freq = 675000;
+	} else if (IS_CHERRYVIEW(dev_priv)) {
+		dev_priv->display.cdclk.max_cdclk_freq = 320000;
+	} else if (IS_VALLEYVIEW(dev_priv)) {
+		dev_priv->display.cdclk.max_cdclk_freq = 400000;
+	} else {
+		/* otherwise assume cdclk is fixed */
+		dev_priv->display.cdclk.max_cdclk_freq = dev_priv->display.cdclk.hw.cdclk;
+	}
+
+	dev_priv->max_dotclk_freq = intel_compute_max_dotclk(dev_priv);
+
+	drm_dbg(&dev_priv->drm, "Max CD clock rate: %d kHz\n",
+		dev_priv->display.cdclk.max_cdclk_freq);
+
+	drm_dbg(&dev_priv->drm, "Max dotclock rate: %d kHz\n",
+		dev_priv->max_dotclk_freq);
+}
+
+/**
+ * intel_update_cdclk - Determine the current CDCLK frequency
+ * @dev_priv: i915 device
+ *
+ * Determine the current CDCLK frequency.
+ */
+void intel_update_cdclk(struct drm_i915_private *dev_priv)
+{
+	intel_cdclk_get_cdclk(dev_priv, &dev_priv->display.cdclk.hw);
+
+	/*
+	 * 9:0 CMBUS [sic] CDCLK frequency (cdfreq):
+	 * Programmng [sic] note: bit[9:2] should be programmed to the number
+	 * of cdclk that generates 4MHz reference clock freq which is used to
+	 * generate GMBus clock. This will vary with the cdclk freq.
+	 */
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		intel_de_write(dev_priv, GMBUSFREQ_VLV,
+			       DIV_ROUND_UP(dev_priv->display.cdclk.hw.cdclk, 1000));
+}
+
+static int dg1_rawclk(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * DG1 always uses a 38.4 MHz rawclk.  The bspec tells us
+	 * "Program Numerator=2, Denominator=4, Divider=37 decimal."
+	 */
+	intel_de_write(dev_priv, PCH_RAWCLK_FREQ,
+		       CNP_RAWCLK_DEN(4) | CNP_RAWCLK_DIV(37) | ICP_RAWCLK_NUM(2));
+
+	return 38400;
+}
+
+static int cnp_rawclk(struct drm_i915_private *dev_priv)
+{
+	u32 rawclk;
+	int divider, fraction;
+
+	if (intel_de_read(dev_priv, SFUSE_STRAP) & SFUSE_STRAP_RAW_FREQUENCY) {
+		/* 24 MHz */
+		divider = 24000;
+		fraction = 0;
+	} else {
+		/* 19.2 MHz */
+		divider = 19000;
+		fraction = 200;
+	}
+
+	rawclk = CNP_RAWCLK_DIV(divider / 1000);
+	if (fraction) {
+		int numerator = 1;
+
+		rawclk |= CNP_RAWCLK_DEN(DIV_ROUND_CLOSEST(numerator * 1000,
+							   fraction) - 1);
+		if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP)
+			rawclk |= ICP_RAWCLK_NUM(numerator);
+	}
+
+	intel_de_write(dev_priv, PCH_RAWCLK_FREQ, rawclk);
+	return divider + fraction;
+}
+
+static int pch_rawclk(struct drm_i915_private *dev_priv)
+{
+	return (intel_de_read(dev_priv, PCH_RAWCLK_FREQ) & RAWCLK_FREQ_MASK) * 1000;
+}
+
+static int vlv_hrawclk(struct drm_i915_private *dev_priv)
+{
+	/* RAWCLK_FREQ_VLV register updated from power well code */
+	return vlv_get_cck_clock_hpll(dev_priv, "hrawclk",
+				      CCK_DISPLAY_REF_CLOCK_CONTROL);
+}
+
+static int i9xx_hrawclk(struct drm_i915_private *dev_priv)
+{
+	u32 clkcfg;
+
+	/*
+	 * hrawclock is 1/4 the FSB frequency
+	 *
+	 * Note that this only reads the state of the FSB
+	 * straps, not the actual FSB frequency. Some BIOSen
+	 * let you configure each independently. Ideally we'd
+	 * read out the actual FSB frequency but sadly we
+	 * don't know which registers have that information,
+	 * and all the relevant docs have gone to bit heaven :(
+	 */
+	clkcfg = intel_de_read(dev_priv, CLKCFG) & CLKCFG_FSB_MASK;
+
+	if (IS_MOBILE(dev_priv)) {
+		switch (clkcfg) {
+		case CLKCFG_FSB_400:
+			return 100000;
+		case CLKCFG_FSB_533:
+			return 133333;
+		case CLKCFG_FSB_667:
+			return 166667;
+		case CLKCFG_FSB_800:
+			return 200000;
+		case CLKCFG_FSB_1067:
+			return 266667;
+		case CLKCFG_FSB_1333:
+			return 333333;
+		default:
+			MISSING_CASE(clkcfg);
+			return 133333;
+		}
+	} else {
+		switch (clkcfg) {
+		case CLKCFG_FSB_400_ALT:
+			return 100000;
+		case CLKCFG_FSB_533:
+			return 133333;
+		case CLKCFG_FSB_667:
+			return 166667;
+		case CLKCFG_FSB_800:
+			return 200000;
+		case CLKCFG_FSB_1067_ALT:
+			return 266667;
+		case CLKCFG_FSB_1333_ALT:
+			return 333333;
+		case CLKCFG_FSB_1600_ALT:
+			return 400000;
+		default:
+			return 133333;
+		}
+	}
+}
+
+/**
+ * intel_read_rawclk - Determine the current RAWCLK frequency
+ * @dev_priv: i915 device
+ *
+ * Determine the current RAWCLK frequency. RAWCLK is a fixed
+ * frequency clock so this needs to done only once.
+ */
+u32 intel_read_rawclk(struct drm_i915_private *dev_priv)
+{
+	u32 freq;
+
+	if (INTEL_PCH_TYPE(dev_priv) >= PCH_DG1)
+		freq = dg1_rawclk(dev_priv);
+	else if (INTEL_PCH_TYPE(dev_priv) >= PCH_MTP)
+		/*
+		 * MTL always uses a 38.4 MHz rawclk.  The bspec tells us
+		 * "RAWCLK_FREQ defaults to the values for 38.4 and does
+		 * not need to be programmed."
+		 */
+		freq = 38400;
+	else if (INTEL_PCH_TYPE(dev_priv) >= PCH_CNP)
+		freq = cnp_rawclk(dev_priv);
+	else if (HAS_PCH_SPLIT(dev_priv))
+		freq = pch_rawclk(dev_priv);
+	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		freq = vlv_hrawclk(dev_priv);
+	else if (DISPLAY_VER(dev_priv) >= 3)
+		freq = i9xx_hrawclk(dev_priv);
+	else
+		/* no rawclk on other platforms, or no need to know it */
+		return 0;
+
+	return freq;
+}
+
+static const struct intel_cdclk_funcs tgl_cdclk_funcs = {
+	.get_cdclk = bxt_get_cdclk,
+	.set_cdclk = bxt_set_cdclk,
+	.modeset_calc_cdclk = bxt_modeset_calc_cdclk,
+	.calc_voltage_level = tgl_calc_voltage_level,
+};
+
+static const struct intel_cdclk_funcs ehl_cdclk_funcs = {
+	.get_cdclk = bxt_get_cdclk,
+	.set_cdclk = bxt_set_cdclk,
+	.modeset_calc_cdclk = bxt_modeset_calc_cdclk,
+	.calc_voltage_level = ehl_calc_voltage_level,
+};
+
+static const struct intel_cdclk_funcs icl_cdclk_funcs = {
+	.get_cdclk = bxt_get_cdclk,
+	.set_cdclk = bxt_set_cdclk,
+	.modeset_calc_cdclk = bxt_modeset_calc_cdclk,
+	.calc_voltage_level = icl_calc_voltage_level,
+};
+
+static const struct intel_cdclk_funcs bxt_cdclk_funcs = {
+	.get_cdclk = bxt_get_cdclk,
+	.set_cdclk = bxt_set_cdclk,
+	.modeset_calc_cdclk = bxt_modeset_calc_cdclk,
+	.calc_voltage_level = bxt_calc_voltage_level,
+};
+
+static const struct intel_cdclk_funcs skl_cdclk_funcs = {
+	.get_cdclk = skl_get_cdclk,
+	.set_cdclk = skl_set_cdclk,
+	.modeset_calc_cdclk = skl_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs bdw_cdclk_funcs = {
+	.get_cdclk = bdw_get_cdclk,
+	.set_cdclk = bdw_set_cdclk,
+	.modeset_calc_cdclk = bdw_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs chv_cdclk_funcs = {
+	.get_cdclk = vlv_get_cdclk,
+	.set_cdclk = chv_set_cdclk,
+	.modeset_calc_cdclk = vlv_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs vlv_cdclk_funcs = {
+	.get_cdclk = vlv_get_cdclk,
+	.set_cdclk = vlv_set_cdclk,
+	.modeset_calc_cdclk = vlv_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs hsw_cdclk_funcs = {
+	.get_cdclk = hsw_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+/* SNB, IVB, 965G, 945G */
+static const struct intel_cdclk_funcs fixed_400mhz_cdclk_funcs = {
+	.get_cdclk = fixed_400mhz_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs ilk_cdclk_funcs = {
+	.get_cdclk = fixed_450mhz_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs gm45_cdclk_funcs = {
+	.get_cdclk = gm45_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+/* G45 uses G33 */
+
+static const struct intel_cdclk_funcs i965gm_cdclk_funcs = {
+	.get_cdclk = i965gm_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+/* i965G uses fixed 400 */
+
+static const struct intel_cdclk_funcs pnv_cdclk_funcs = {
+	.get_cdclk = pnv_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs g33_cdclk_funcs = {
+	.get_cdclk = g33_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs i945gm_cdclk_funcs = {
+	.get_cdclk = i945gm_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+/* i945G uses fixed 400 */
+
+static const struct intel_cdclk_funcs i915gm_cdclk_funcs = {
+	.get_cdclk = i915gm_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs i915g_cdclk_funcs = {
+	.get_cdclk = fixed_333mhz_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs i865g_cdclk_funcs = {
+	.get_cdclk = fixed_266mhz_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs i85x_cdclk_funcs = {
+	.get_cdclk = i85x_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs i845g_cdclk_funcs = {
+	.get_cdclk = fixed_200mhz_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+static const struct intel_cdclk_funcs i830_cdclk_funcs = {
+	.get_cdclk = fixed_133mhz_get_cdclk,
+	.modeset_calc_cdclk = fixed_modeset_calc_cdclk,
+};
+
+/**
+ * intel_init_cdclk_hooks - Initialize CDCLK related modesetting hooks
+ * @dev_priv: i915 device
+ */
+void intel_init_cdclk_hooks(struct drm_i915_private *dev_priv)
+{
+	if (IS_DG2(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &tgl_cdclk_funcs;
+		dev_priv->display.cdclk.table = dg2_cdclk_table;
+	} else if (IS_ALDERLAKE_P(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &tgl_cdclk_funcs;
+		/* Wa_22011320316:adl-p[a0] */
+		if (IS_ADLP_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0))
+			dev_priv->display.cdclk.table = adlp_a_step_cdclk_table;
+		else
+			dev_priv->display.cdclk.table = adlp_cdclk_table;
+	} else if (IS_ROCKETLAKE(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &tgl_cdclk_funcs;
+		dev_priv->display.cdclk.table = rkl_cdclk_table;
+	} else if (DISPLAY_VER(dev_priv) >= 12) {
+		dev_priv->display.funcs.cdclk = &tgl_cdclk_funcs;
+		dev_priv->display.cdclk.table = icl_cdclk_table;
+	} else if (IS_JSL_EHL(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &ehl_cdclk_funcs;
+		dev_priv->display.cdclk.table = icl_cdclk_table;
+	} else if (DISPLAY_VER(dev_priv) >= 11) {
+		dev_priv->display.funcs.cdclk = &icl_cdclk_funcs;
+		dev_priv->display.cdclk.table = icl_cdclk_table;
+	} else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &bxt_cdclk_funcs;
+		if (IS_GEMINILAKE(dev_priv))
+			dev_priv->display.cdclk.table = glk_cdclk_table;
+		else
+			dev_priv->display.cdclk.table = bxt_cdclk_table;
+	} else if (DISPLAY_VER(dev_priv) == 9) {
+		dev_priv->display.funcs.cdclk = &skl_cdclk_funcs;
+	} else if (IS_BROADWELL(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &bdw_cdclk_funcs;
+	} else if (IS_HASWELL(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &hsw_cdclk_funcs;
+	} else if (IS_CHERRYVIEW(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &chv_cdclk_funcs;
+	} else if (IS_VALLEYVIEW(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &vlv_cdclk_funcs;
+	} else if (IS_SANDYBRIDGE(dev_priv) || IS_IVYBRIDGE(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &fixed_400mhz_cdclk_funcs;
+	} else if (IS_IRONLAKE(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &ilk_cdclk_funcs;
+	} else if (IS_GM45(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &gm45_cdclk_funcs;
+	} else if (IS_G45(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &g33_cdclk_funcs;
+	} else if (IS_I965GM(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &i965gm_cdclk_funcs;
+	} else if (IS_I965G(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &fixed_400mhz_cdclk_funcs;
+	} else if (IS_PINEVIEW(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &pnv_cdclk_funcs;
+	} else if (IS_G33(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &g33_cdclk_funcs;
+	} else if (IS_I945GM(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &i945gm_cdclk_funcs;
+	} else if (IS_I945G(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &fixed_400mhz_cdclk_funcs;
+	} else if (IS_I915GM(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &i915gm_cdclk_funcs;
+	} else if (IS_I915G(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &i915g_cdclk_funcs;
+	} else if (IS_I865G(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &i865g_cdclk_funcs;
+	} else if (IS_I85X(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &i85x_cdclk_funcs;
+	} else if (IS_I845G(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &i845g_cdclk_funcs;
+	} else if (IS_I830(dev_priv)) {
+		dev_priv->display.funcs.cdclk = &i830_cdclk_funcs;
+	}
+
+	if (drm_WARN(&dev_priv->drm, !dev_priv->display.funcs.cdclk,
+		     "Unknown platform. Assuming i830\n"))
+		dev_priv->display.funcs.cdclk = &i830_cdclk_funcs;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_cdclk.h b/drivers/gpu/drm/i915/display/intel_cdclk.h
new file mode 100644
index 000000000..c674879a8
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_cdclk.h
@@ -0,0 +1,86 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_CDCLK_H__
+#define __INTEL_CDCLK_H__
+
+#include <linux/types.h>
+
+#include "intel_display.h"
+#include "intel_global_state.h"
+
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc_state;
+
+struct intel_cdclk_config {
+	unsigned int cdclk, vco, ref, bypass;
+	u8 voltage_level;
+};
+
+struct intel_cdclk_state {
+	struct intel_global_state base;
+
+	/*
+	 * Logical configuration of cdclk (used for all scaling,
+	 * watermark, etc. calculations and checks). This is
+	 * computed as if all enabled crtcs were active.
+	 */
+	struct intel_cdclk_config logical;
+
+	/*
+	 * Actual configuration of cdclk, can be different from the
+	 * logical configuration only when all crtc's are DPMS off.
+	 */
+	struct intel_cdclk_config actual;
+
+	/* minimum acceptable cdclk to satisfy bandwidth requirements */
+	int bw_min_cdclk;
+	/* minimum acceptable cdclk for each pipe */
+	int min_cdclk[I915_MAX_PIPES];
+	/* minimum acceptable voltage level for each pipe */
+	u8 min_voltage_level[I915_MAX_PIPES];
+
+	/* pipe to which cd2x update is synchronized */
+	enum pipe pipe;
+
+	/* forced minimum cdclk for glk+ audio w/a */
+	int force_min_cdclk;
+
+	/* bitmask of active pipes */
+	u8 active_pipes;
+};
+
+int intel_crtc_compute_min_cdclk(const struct intel_crtc_state *crtc_state);
+void intel_cdclk_init_hw(struct drm_i915_private *i915);
+void intel_cdclk_uninit_hw(struct drm_i915_private *i915);
+void intel_init_cdclk_hooks(struct drm_i915_private *dev_priv);
+void intel_update_max_cdclk(struct drm_i915_private *dev_priv);
+void intel_update_cdclk(struct drm_i915_private *dev_priv);
+u32 intel_read_rawclk(struct drm_i915_private *dev_priv);
+bool intel_cdclk_needs_modeset(const struct intel_cdclk_config *a,
+			       const struct intel_cdclk_config *b);
+void intel_set_cdclk_pre_plane_update(struct intel_atomic_state *state);
+void intel_set_cdclk_post_plane_update(struct intel_atomic_state *state);
+void intel_cdclk_dump_config(struct drm_i915_private *i915,
+			     const struct intel_cdclk_config *cdclk_config,
+			     const char *context);
+int intel_modeset_calc_cdclk(struct intel_atomic_state *state);
+void intel_cdclk_get_cdclk(struct drm_i915_private *dev_priv,
+			   struct intel_cdclk_config *cdclk_config);
+int intel_cdclk_atomic_check(struct intel_atomic_state *state,
+			     bool *need_cdclk_calc);
+struct intel_cdclk_state *
+intel_atomic_get_cdclk_state(struct intel_atomic_state *state);
+
+#define to_intel_cdclk_state(x) container_of((x), struct intel_cdclk_state, base)
+#define intel_atomic_get_old_cdclk_state(state) \
+	to_intel_cdclk_state(intel_atomic_get_old_global_obj_state(state, &to_i915(state->base.dev)->display.cdclk.obj))
+#define intel_atomic_get_new_cdclk_state(state) \
+	to_intel_cdclk_state(intel_atomic_get_new_global_obj_state(state, &to_i915(state->base.dev)->display.cdclk.obj))
+
+int intel_cdclk_init(struct drm_i915_private *dev_priv);
+
+#endif /* __INTEL_CDCLK_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_color.c b/drivers/gpu/drm/i915/display/intel_color.c
new file mode 100644
index 000000000..78211e583
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_color.c
@@ -0,0 +1,2295 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "intel_color.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dpll.h"
+#include "intel_dsb.h"
+#include "vlv_dsi_pll.h"
+
+struct intel_color_funcs {
+	int (*color_check)(struct intel_crtc_state *crtc_state);
+	/*
+	 * Program non-arming double buffered color management registers
+	 * before vblank evasion. The registers should then latch after
+	 * the arming register is written (by color_commit_arm()) during
+	 * the next vblank start, alongside any other double buffered
+	 * registers involved with the same commit. This hook is optional.
+	 */
+	void (*color_commit_noarm)(const struct intel_crtc_state *crtc_state);
+	/*
+	 * Program arming double buffered color management registers
+	 * during vblank evasion. The registers (and whatever other registers
+	 * they arm that were written by color_commit_noarm) should then latch
+	 * during the next vblank start, alongside any other double buffered
+	 * registers involved with the same commit.
+	 */
+	void (*color_commit_arm)(const struct intel_crtc_state *crtc_state);
+	/*
+	 * Load LUTs (and other single buffered color management
+	 * registers). Will (hopefully) be called during the vblank
+	 * following the latching of any double buffered registers
+	 * involved with the same commit.
+	 */
+	void (*load_luts)(const struct intel_crtc_state *crtc_state);
+	void (*read_luts)(struct intel_crtc_state *crtc_state);
+};
+
+#define CTM_COEFF_SIGN	(1ULL << 63)
+
+#define CTM_COEFF_1_0	(1ULL << 32)
+#define CTM_COEFF_2_0	(CTM_COEFF_1_0 << 1)
+#define CTM_COEFF_4_0	(CTM_COEFF_2_0 << 1)
+#define CTM_COEFF_8_0	(CTM_COEFF_4_0 << 1)
+#define CTM_COEFF_0_5	(CTM_COEFF_1_0 >> 1)
+#define CTM_COEFF_0_25	(CTM_COEFF_0_5 >> 1)
+#define CTM_COEFF_0_125	(CTM_COEFF_0_25 >> 1)
+
+#define CTM_COEFF_LIMITED_RANGE ((235ULL - 16ULL) * CTM_COEFF_1_0 / 255)
+
+#define CTM_COEFF_NEGATIVE(coeff)	(((coeff) & CTM_COEFF_SIGN) != 0)
+#define CTM_COEFF_ABS(coeff)		((coeff) & (CTM_COEFF_SIGN - 1))
+
+#define LEGACY_LUT_LENGTH		256
+
+/*
+ * ILK+ csc matrix:
+ *
+ * |R/Cr|   | c0 c1 c2 |   ( |R/Cr|   |preoff0| )   |postoff0|
+ * |G/Y | = | c3 c4 c5 | x ( |G/Y | + |preoff1| ) + |postoff1|
+ * |B/Cb|   | c6 c7 c8 |   ( |B/Cb|   |preoff2| )   |postoff2|
+ *
+ * ILK/SNB don't have explicit post offsets, and instead
+ * CSC_MODE_YUV_TO_RGB and CSC_BLACK_SCREEN_OFFSET are used:
+ *  CSC_MODE_YUV_TO_RGB=0 + CSC_BLACK_SCREEN_OFFSET=0 -> 1/2, 0, 1/2
+ *  CSC_MODE_YUV_TO_RGB=0 + CSC_BLACK_SCREEN_OFFSET=1 -> 1/2, 1/16, 1/2
+ *  CSC_MODE_YUV_TO_RGB=1 + CSC_BLACK_SCREEN_OFFSET=0 -> 0, 0, 0
+ *  CSC_MODE_YUV_TO_RGB=1 + CSC_BLACK_SCREEN_OFFSET=1 -> 1/16, 1/16, 1/16
+ */
+
+/*
+ * Extract the CSC coefficient from a CTM coefficient (in U32.32 fixed point
+ * format). This macro takes the coefficient we want transformed and the
+ * number of fractional bits.
+ *
+ * We only have a 9 bits precision window which slides depending on the value
+ * of the CTM coefficient and we write the value from bit 3. We also round the
+ * value.
+ */
+#define ILK_CSC_COEFF_FP(coeff, fbits)	\
+	(clamp_val(((coeff) >> (32 - (fbits) - 3)) + 4, 0, 0xfff) & 0xff8)
+
+#define ILK_CSC_COEFF_LIMITED_RANGE 0x0dc0
+#define ILK_CSC_COEFF_1_0 0x7800
+
+#define ILK_CSC_POSTOFF_LIMITED_RANGE (16 * (1 << 12) / 255)
+
+/* Nop pre/post offsets */
+static const u16 ilk_csc_off_zero[3] = {};
+
+/* Identity matrix */
+static const u16 ilk_csc_coeff_identity[9] = {
+	ILK_CSC_COEFF_1_0, 0, 0,
+	0, ILK_CSC_COEFF_1_0, 0,
+	0, 0, ILK_CSC_COEFF_1_0,
+};
+
+/* Limited range RGB post offsets */
+static const u16 ilk_csc_postoff_limited_range[3] = {
+	ILK_CSC_POSTOFF_LIMITED_RANGE,
+	ILK_CSC_POSTOFF_LIMITED_RANGE,
+	ILK_CSC_POSTOFF_LIMITED_RANGE,
+};
+
+/* Full range RGB -> limited range RGB matrix */
+static const u16 ilk_csc_coeff_limited_range[9] = {
+	ILK_CSC_COEFF_LIMITED_RANGE, 0, 0,
+	0, ILK_CSC_COEFF_LIMITED_RANGE, 0,
+	0, 0, ILK_CSC_COEFF_LIMITED_RANGE,
+};
+
+/* BT.709 full range RGB -> limited range YCbCr matrix */
+static const u16 ilk_csc_coeff_rgb_to_ycbcr[9] = {
+	0x1e08, 0x9cc0, 0xb528,
+	0x2ba8, 0x09d8, 0x37e8,
+	0xbce8, 0x9ad8, 0x1e08,
+};
+
+/* Limited range YCbCr post offsets */
+static const u16 ilk_csc_postoff_rgb_to_ycbcr[3] = {
+	0x0800, 0x0100, 0x0800,
+};
+
+static bool lut_is_legacy(const struct drm_property_blob *lut)
+{
+	return drm_color_lut_size(lut) == LEGACY_LUT_LENGTH;
+}
+
+static bool crtc_state_is_legacy_gamma(const struct intel_crtc_state *crtc_state)
+{
+	return !crtc_state->hw.degamma_lut &&
+		!crtc_state->hw.ctm &&
+		crtc_state->hw.gamma_lut &&
+		lut_is_legacy(crtc_state->hw.gamma_lut);
+}
+
+/*
+ * When using limited range, multiply the matrix given by userspace by
+ * the matrix that we would use for the limited range.
+ */
+static u64 *ctm_mult_by_limited(u64 *result, const u64 *input)
+{
+	int i;
+
+	for (i = 0; i < 9; i++) {
+		u64 user_coeff = input[i];
+		u32 limited_coeff = CTM_COEFF_LIMITED_RANGE;
+		u32 abs_coeff = clamp_val(CTM_COEFF_ABS(user_coeff), 0,
+					  CTM_COEFF_4_0 - 1) >> 2;
+
+		/*
+		 * By scaling every co-efficient with limited range (16-235)
+		 * vs full range (0-255) the final o/p will be scaled down to
+		 * fit in the limited range supported by the panel.
+		 */
+		result[i] = mul_u32_u32(limited_coeff, abs_coeff) >> 30;
+		result[i] |= user_coeff & CTM_COEFF_SIGN;
+	}
+
+	return result;
+}
+
+static void ilk_update_pipe_csc(struct intel_crtc *crtc,
+				const u16 preoff[3],
+				const u16 coeff[9],
+				const u16 postoff[3])
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_PREOFF_HI(pipe), preoff[0]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_PREOFF_ME(pipe), preoff[1]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_PREOFF_LO(pipe), preoff[2]);
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_COEFF_RY_GY(pipe),
+			  coeff[0] << 16 | coeff[1]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_COEFF_BY(pipe), coeff[2] << 16);
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_COEFF_RU_GU(pipe),
+			  coeff[3] << 16 | coeff[4]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_COEFF_BU(pipe), coeff[5] << 16);
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_COEFF_RV_GV(pipe),
+			  coeff[6] << 16 | coeff[7]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_COEFF_BV(pipe), coeff[8] << 16);
+
+	if (DISPLAY_VER(dev_priv) >= 7) {
+		intel_de_write_fw(dev_priv, PIPE_CSC_POSTOFF_HI(pipe),
+				  postoff[0]);
+		intel_de_write_fw(dev_priv, PIPE_CSC_POSTOFF_ME(pipe),
+				  postoff[1]);
+		intel_de_write_fw(dev_priv, PIPE_CSC_POSTOFF_LO(pipe),
+				  postoff[2]);
+	}
+}
+
+static void icl_update_output_csc(struct intel_crtc *crtc,
+				  const u16 preoff[3],
+				  const u16 coeff[9],
+				  const u16 postoff[3])
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_PREOFF_HI(pipe), preoff[0]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_PREOFF_ME(pipe), preoff[1]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_PREOFF_LO(pipe), preoff[2]);
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_COEFF_RY_GY(pipe),
+			  coeff[0] << 16 | coeff[1]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_COEFF_BY(pipe),
+			  coeff[2] << 16);
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_COEFF_RU_GU(pipe),
+			  coeff[3] << 16 | coeff[4]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_COEFF_BU(pipe),
+			  coeff[5] << 16);
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_COEFF_RV_GV(pipe),
+			  coeff[6] << 16 | coeff[7]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_COEFF_BV(pipe),
+			  coeff[8] << 16);
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_POSTOFF_HI(pipe), postoff[0]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_POSTOFF_ME(pipe), postoff[1]);
+	intel_de_write_fw(dev_priv, PIPE_CSC_OUTPUT_POSTOFF_LO(pipe), postoff[2]);
+}
+
+static bool ilk_csc_limited_range(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	/*
+	 * FIXME if there's a gamma LUT after the CSC, we should
+	 * do the range compression using the gamma LUT instead.
+	 */
+	return crtc_state->limited_color_range &&
+		(IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv) ||
+		 IS_DISPLAY_VER(dev_priv, 9, 10));
+}
+
+static void ilk_csc_convert_ctm(const struct intel_crtc_state *crtc_state,
+				u16 coeffs[9])
+{
+	const struct drm_color_ctm *ctm = crtc_state->hw.ctm->data;
+	const u64 *input;
+	u64 temp[9];
+	int i;
+
+	if (ilk_csc_limited_range(crtc_state))
+		input = ctm_mult_by_limited(temp, ctm->matrix);
+	else
+		input = ctm->matrix;
+
+	/*
+	 * Convert fixed point S31.32 input to format supported by the
+	 * hardware.
+	 */
+	for (i = 0; i < 9; i++) {
+		u64 abs_coeff = ((1ULL << 63) - 1) & input[i];
+
+		/*
+		 * Clamp input value to min/max supported by
+		 * hardware.
+		 */
+		abs_coeff = clamp_val(abs_coeff, 0, CTM_COEFF_4_0 - 1);
+
+		coeffs[i] = 0;
+
+		/* sign bit */
+		if (CTM_COEFF_NEGATIVE(input[i]))
+			coeffs[i] |= 1 << 15;
+
+		if (abs_coeff < CTM_COEFF_0_125)
+			coeffs[i] |= (3 << 12) |
+				ILK_CSC_COEFF_FP(abs_coeff, 12);
+		else if (abs_coeff < CTM_COEFF_0_25)
+			coeffs[i] |= (2 << 12) |
+				ILK_CSC_COEFF_FP(abs_coeff, 11);
+		else if (abs_coeff < CTM_COEFF_0_5)
+			coeffs[i] |= (1 << 12) |
+				ILK_CSC_COEFF_FP(abs_coeff, 10);
+		else if (abs_coeff < CTM_COEFF_1_0)
+			coeffs[i] |= ILK_CSC_COEFF_FP(abs_coeff, 9);
+		else if (abs_coeff < CTM_COEFF_2_0)
+			coeffs[i] |= (7 << 12) |
+				ILK_CSC_COEFF_FP(abs_coeff, 8);
+		else
+			coeffs[i] |= (6 << 12) |
+				ILK_CSC_COEFF_FP(abs_coeff, 7);
+	}
+}
+
+static void ilk_load_csc_matrix(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	bool limited_color_range = ilk_csc_limited_range(crtc_state);
+
+	if (crtc_state->hw.ctm) {
+		u16 coeff[9];
+
+		ilk_csc_convert_ctm(crtc_state, coeff);
+		ilk_update_pipe_csc(crtc, ilk_csc_off_zero, coeff,
+				    limited_color_range ?
+				    ilk_csc_postoff_limited_range :
+				    ilk_csc_off_zero);
+	} else if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB) {
+		ilk_update_pipe_csc(crtc, ilk_csc_off_zero,
+				    ilk_csc_coeff_rgb_to_ycbcr,
+				    ilk_csc_postoff_rgb_to_ycbcr);
+	} else if (limited_color_range) {
+		ilk_update_pipe_csc(crtc, ilk_csc_off_zero,
+				    ilk_csc_coeff_limited_range,
+				    ilk_csc_postoff_limited_range);
+	} else if (crtc_state->csc_enable) {
+		/*
+		 * On GLK both pipe CSC and degamma LUT are controlled
+		 * by csc_enable. Hence for the cases where the degama
+		 * LUT is needed but CSC is not we need to load an
+		 * identity matrix.
+		 */
+		drm_WARN_ON(&dev_priv->drm, !IS_GEMINILAKE(dev_priv));
+
+		ilk_update_pipe_csc(crtc, ilk_csc_off_zero,
+				    ilk_csc_coeff_identity,
+				    ilk_csc_off_zero);
+	}
+}
+
+static void icl_load_csc_matrix(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	if (crtc_state->hw.ctm) {
+		u16 coeff[9];
+
+		ilk_csc_convert_ctm(crtc_state, coeff);
+		ilk_update_pipe_csc(crtc, ilk_csc_off_zero,
+				    coeff, ilk_csc_off_zero);
+	}
+
+	if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB) {
+		icl_update_output_csc(crtc, ilk_csc_off_zero,
+				      ilk_csc_coeff_rgb_to_ycbcr,
+				      ilk_csc_postoff_rgb_to_ycbcr);
+	} else if (crtc_state->limited_color_range) {
+		icl_update_output_csc(crtc, ilk_csc_off_zero,
+				      ilk_csc_coeff_limited_range,
+				      ilk_csc_postoff_limited_range);
+	}
+}
+
+static void chv_load_cgm_csc(struct intel_crtc *crtc,
+			     const struct drm_property_blob *blob)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_color_ctm *ctm = blob->data;
+	enum pipe pipe = crtc->pipe;
+	u16 coeffs[9];
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(coeffs); i++) {
+		u64 abs_coeff = ((1ULL << 63) - 1) & ctm->matrix[i];
+
+		/* Round coefficient. */
+		abs_coeff += 1 << (32 - 13);
+		/* Clamp to hardware limits. */
+		abs_coeff = clamp_val(abs_coeff, 0, CTM_COEFF_8_0 - 1);
+
+		coeffs[i] = 0;
+
+		/* Write coefficients in S3.12 format. */
+		if (ctm->matrix[i] & (1ULL << 63))
+			coeffs[i] |= 1 << 15;
+
+		coeffs[i] |= ((abs_coeff >> 32) & 7) << 12;
+		coeffs[i] |= (abs_coeff >> 20) & 0xfff;
+	}
+
+	intel_de_write_fw(dev_priv, CGM_PIPE_CSC_COEFF01(pipe),
+			  coeffs[1] << 16 | coeffs[0]);
+	intel_de_write_fw(dev_priv, CGM_PIPE_CSC_COEFF23(pipe),
+			  coeffs[3] << 16 | coeffs[2]);
+	intel_de_write_fw(dev_priv, CGM_PIPE_CSC_COEFF45(pipe),
+			  coeffs[5] << 16 | coeffs[4]);
+	intel_de_write_fw(dev_priv, CGM_PIPE_CSC_COEFF67(pipe),
+			  coeffs[7] << 16 | coeffs[6]);
+	intel_de_write_fw(dev_priv, CGM_PIPE_CSC_COEFF8(pipe),
+			  coeffs[8]);
+}
+
+/* convert hw value with given bit_precision to lut property val */
+static u32 intel_color_lut_pack(u32 val, int bit_precision)
+{
+	u32 max = 0xffff >> (16 - bit_precision);
+
+	val = clamp_val(val, 0, max);
+
+	if (bit_precision < 16)
+		val <<= 16 - bit_precision;
+
+	return val;
+}
+
+static u32 i9xx_lut_8(const struct drm_color_lut *color)
+{
+	return drm_color_lut_extract(color->red, 8) << 16 |
+		drm_color_lut_extract(color->green, 8) << 8 |
+		drm_color_lut_extract(color->blue, 8);
+}
+
+static void i9xx_lut_8_pack(struct drm_color_lut *entry, u32 val)
+{
+	entry->red = intel_color_lut_pack(REG_FIELD_GET(LGC_PALETTE_RED_MASK, val), 8);
+	entry->green = intel_color_lut_pack(REG_FIELD_GET(LGC_PALETTE_GREEN_MASK, val), 8);
+	entry->blue = intel_color_lut_pack(REG_FIELD_GET(LGC_PALETTE_BLUE_MASK, val), 8);
+}
+
+/* i965+ "10.6" bit interpolated format "even DW" (low 8 bits) */
+static u32 i965_lut_10p6_ldw(const struct drm_color_lut *color)
+{
+	return (color->red & 0xff) << 16 |
+		(color->green & 0xff) << 8 |
+		(color->blue & 0xff);
+}
+
+/* i965+ "10.6" interpolated format "odd DW" (high 8 bits) */
+static u32 i965_lut_10p6_udw(const struct drm_color_lut *color)
+{
+	return (color->red >> 8) << 16 |
+		(color->green >> 8) << 8 |
+		(color->blue >> 8);
+}
+
+static void i965_lut_10p6_pack(struct drm_color_lut *entry, u32 ldw, u32 udw)
+{
+	entry->red = REG_FIELD_GET(PALETTE_RED_MASK, udw) << 8 |
+		REG_FIELD_GET(PALETTE_RED_MASK, ldw);
+	entry->green = REG_FIELD_GET(PALETTE_GREEN_MASK, udw) << 8 |
+		REG_FIELD_GET(PALETTE_GREEN_MASK, ldw);
+	entry->blue = REG_FIELD_GET(PALETTE_BLUE_MASK, udw) << 8 |
+		REG_FIELD_GET(PALETTE_BLUE_MASK, ldw);
+}
+
+static u16 i965_lut_11p6_max_pack(u32 val)
+{
+	/* PIPEGCMAX is 11.6, clamp to 10.6 */
+	return clamp_val(val, 0, 0xffff);
+}
+
+static u32 ilk_lut_10(const struct drm_color_lut *color)
+{
+	return drm_color_lut_extract(color->red, 10) << 20 |
+		drm_color_lut_extract(color->green, 10) << 10 |
+		drm_color_lut_extract(color->blue, 10);
+}
+
+static void ilk_lut_10_pack(struct drm_color_lut *entry, u32 val)
+{
+	entry->red = intel_color_lut_pack(REG_FIELD_GET(PREC_PALETTE_RED_MASK, val), 10);
+	entry->green = intel_color_lut_pack(REG_FIELD_GET(PREC_PALETTE_GREEN_MASK, val), 10);
+	entry->blue = intel_color_lut_pack(REG_FIELD_GET(PREC_PALETTE_BLUE_MASK, val), 10);
+}
+
+static void icl_lut_multi_seg_pack(struct drm_color_lut *entry, u32 ldw, u32 udw)
+{
+	entry->red = REG_FIELD_GET(PAL_PREC_MULTI_SEG_RED_UDW_MASK, udw) << 6 |
+				   REG_FIELD_GET(PAL_PREC_MULTI_SEG_RED_LDW_MASK, ldw);
+	entry->green = REG_FIELD_GET(PAL_PREC_MULTI_SEG_GREEN_UDW_MASK, udw) << 6 |
+				     REG_FIELD_GET(PAL_PREC_MULTI_SEG_GREEN_LDW_MASK, ldw);
+	entry->blue = REG_FIELD_GET(PAL_PREC_MULTI_SEG_BLUE_UDW_MASK, udw) << 6 |
+				    REG_FIELD_GET(PAL_PREC_MULTI_SEG_BLUE_LDW_MASK, ldw);
+}
+
+static void icl_color_commit_noarm(const struct intel_crtc_state *crtc_state)
+{
+	icl_load_csc_matrix(crtc_state);
+}
+
+static void skl_color_commit_noarm(const struct intel_crtc_state *crtc_state)
+{
+	/*
+	 * Possibly related to display WA #1184, SKL CSC loses the latched
+	 * CSC coeff/offset register values if the CSC registers are disarmed
+	 * between DC5 exit and PSR exit. This will cause the plane(s) to
+	 * output all black (until CSC_MODE is rearmed and properly latched).
+	 * Once PSR exit (and proper register latching) has occurred the
+	 * danger is over. Thus when PSR is enabled the CSC coeff/offset
+	 * register programming will be peformed from skl_color_commit_arm()
+	 * which is called after PSR exit.
+	 */
+	if (!crtc_state->has_psr)
+		ilk_load_csc_matrix(crtc_state);
+}
+
+static void ilk_color_commit_noarm(const struct intel_crtc_state *crtc_state)
+{
+	ilk_load_csc_matrix(crtc_state);
+}
+
+static void i9xx_color_commit_arm(const struct intel_crtc_state *crtc_state)
+{
+	/* update PIPECONF GAMMA_MODE */
+	i9xx_set_pipeconf(crtc_state);
+}
+
+static void ilk_color_commit_arm(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	/* update PIPECONF GAMMA_MODE */
+	ilk_set_pipeconf(crtc_state);
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_MODE(crtc->pipe),
+			  crtc_state->csc_mode);
+}
+
+static void hsw_color_commit_arm(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	intel_de_write(dev_priv, GAMMA_MODE(crtc->pipe),
+		       crtc_state->gamma_mode);
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_MODE(crtc->pipe),
+			  crtc_state->csc_mode);
+}
+
+static void skl_color_commit_arm(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	u32 val = 0;
+
+	if (crtc_state->has_psr)
+		ilk_load_csc_matrix(crtc_state);
+
+	/*
+	 * We don't (yet) allow userspace to control the pipe background color,
+	 * so force it to black, but apply pipe gamma and CSC appropriately
+	 * so that its handling will match how we program our planes.
+	 */
+	if (crtc_state->gamma_enable)
+		val |= SKL_BOTTOM_COLOR_GAMMA_ENABLE;
+	if (crtc_state->csc_enable)
+		val |= SKL_BOTTOM_COLOR_CSC_ENABLE;
+	intel_de_write(dev_priv, SKL_BOTTOM_COLOR(pipe), val);
+
+	intel_de_write(dev_priv, GAMMA_MODE(crtc->pipe),
+		       crtc_state->gamma_mode);
+
+	intel_de_write_fw(dev_priv, PIPE_CSC_MODE(crtc->pipe),
+			  crtc_state->csc_mode);
+}
+
+static void i9xx_load_lut_8(struct intel_crtc *crtc,
+			    const struct drm_property_blob *blob)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_color_lut *lut;
+	enum pipe pipe = crtc->pipe;
+	int i;
+
+	if (!blob)
+		return;
+
+	lut = blob->data;
+
+	for (i = 0; i < 256; i++)
+		intel_de_write_fw(dev_priv, PALETTE(pipe, i),
+				  i9xx_lut_8(&lut[i]));
+}
+
+static void i9xx_load_luts(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_property_blob *gamma_lut = crtc_state->hw.gamma_lut;
+
+	assert_pll_enabled(dev_priv, crtc->pipe);
+
+	i9xx_load_lut_8(crtc, gamma_lut);
+}
+
+static void i965_load_lut_10p6(struct intel_crtc *crtc,
+			       const struct drm_property_blob *blob)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_color_lut *lut = blob->data;
+	int i, lut_size = drm_color_lut_size(blob);
+	enum pipe pipe = crtc->pipe;
+
+	for (i = 0; i < lut_size - 1; i++) {
+		intel_de_write_fw(dev_priv, PALETTE(pipe, 2 * i + 0),
+				  i965_lut_10p6_ldw(&lut[i]));
+		intel_de_write_fw(dev_priv, PALETTE(pipe, 2 * i + 1),
+				  i965_lut_10p6_udw(&lut[i]));
+	}
+
+	intel_de_write_fw(dev_priv, PIPEGCMAX(pipe, 0), lut[i].red);
+	intel_de_write_fw(dev_priv, PIPEGCMAX(pipe, 1), lut[i].green);
+	intel_de_write_fw(dev_priv, PIPEGCMAX(pipe, 2), lut[i].blue);
+}
+
+static void i965_load_luts(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_property_blob *gamma_lut = crtc_state->hw.gamma_lut;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI))
+		assert_dsi_pll_enabled(dev_priv);
+	else
+		assert_pll_enabled(dev_priv, crtc->pipe);
+
+	if (crtc_state->gamma_mode == GAMMA_MODE_MODE_8BIT)
+		i9xx_load_lut_8(crtc, gamma_lut);
+	else
+		i965_load_lut_10p6(crtc, gamma_lut);
+}
+
+static void ilk_load_lut_8(struct intel_crtc *crtc,
+			   const struct drm_property_blob *blob)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_color_lut *lut;
+	enum pipe pipe = crtc->pipe;
+	int i;
+
+	if (!blob)
+		return;
+
+	lut = blob->data;
+
+	for (i = 0; i < 256; i++)
+		intel_de_write_fw(dev_priv, LGC_PALETTE(pipe, i),
+				  i9xx_lut_8(&lut[i]));
+}
+
+static void ilk_load_lut_10(struct intel_crtc *crtc,
+			    const struct drm_property_blob *blob)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_color_lut *lut = blob->data;
+	int i, lut_size = drm_color_lut_size(blob);
+	enum pipe pipe = crtc->pipe;
+
+	for (i = 0; i < lut_size; i++)
+		intel_de_write_fw(dev_priv, PREC_PALETTE(pipe, i),
+				  ilk_lut_10(&lut[i]));
+}
+
+static void ilk_load_luts(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	const struct drm_property_blob *gamma_lut = crtc_state->hw.gamma_lut;
+
+	switch (crtc_state->gamma_mode) {
+	case GAMMA_MODE_MODE_8BIT:
+		ilk_load_lut_8(crtc, gamma_lut);
+		break;
+	case GAMMA_MODE_MODE_10BIT:
+		ilk_load_lut_10(crtc, gamma_lut);
+		break;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		break;
+	}
+}
+
+static int ivb_lut_10_size(u32 prec_index)
+{
+	if (prec_index & PAL_PREC_SPLIT_MODE)
+		return 512;
+	else
+		return 1024;
+}
+
+/*
+ * IVB/HSW Bspec / PAL_PREC_INDEX:
+ * "Restriction : Index auto increment mode is not
+ *  supported and must not be enabled."
+ */
+static void ivb_load_lut_10(struct intel_crtc *crtc,
+			    const struct drm_property_blob *blob,
+			    u32 prec_index)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	int hw_lut_size = ivb_lut_10_size(prec_index);
+	const struct drm_color_lut *lut = blob->data;
+	int i, lut_size = drm_color_lut_size(blob);
+	enum pipe pipe = crtc->pipe;
+
+	for (i = 0; i < hw_lut_size; i++) {
+		/* We discard half the user entries in split gamma mode */
+		const struct drm_color_lut *entry =
+			&lut[i * (lut_size - 1) / (hw_lut_size - 1)];
+
+		intel_de_write_fw(dev_priv, PREC_PAL_INDEX(pipe), prec_index++);
+		intel_de_write_fw(dev_priv, PREC_PAL_DATA(pipe),
+				  ilk_lut_10(entry));
+	}
+
+	/*
+	 * Reset the index, otherwise it prevents the legacy palette to be
+	 * written properly.
+	 */
+	intel_de_write_fw(dev_priv, PREC_PAL_INDEX(pipe), 0);
+}
+
+/* On BDW+ the index auto increment mode actually works */
+static void bdw_load_lut_10(struct intel_crtc *crtc,
+			    const struct drm_property_blob *blob,
+			    u32 prec_index)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	int hw_lut_size = ivb_lut_10_size(prec_index);
+	const struct drm_color_lut *lut = blob->data;
+	int i, lut_size = drm_color_lut_size(blob);
+	enum pipe pipe = crtc->pipe;
+
+	intel_de_write_fw(dev_priv, PREC_PAL_INDEX(pipe),
+			  prec_index | PAL_PREC_AUTO_INCREMENT);
+
+	for (i = 0; i < hw_lut_size; i++) {
+		/* We discard half the user entries in split gamma mode */
+		const struct drm_color_lut *entry =
+			&lut[i * (lut_size - 1) / (hw_lut_size - 1)];
+
+		intel_de_write_fw(dev_priv, PREC_PAL_DATA(pipe),
+				  ilk_lut_10(entry));
+	}
+
+	/*
+	 * Reset the index, otherwise it prevents the legacy palette to be
+	 * written properly.
+	 */
+	intel_de_write_fw(dev_priv, PREC_PAL_INDEX(pipe), 0);
+}
+
+static void ivb_load_lut_ext_max(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	/* Program the max register to clamp values > 1.0. */
+	intel_dsb_reg_write(crtc_state, PREC_PAL_EXT_GC_MAX(pipe, 0), 1 << 16);
+	intel_dsb_reg_write(crtc_state, PREC_PAL_EXT_GC_MAX(pipe, 1), 1 << 16);
+	intel_dsb_reg_write(crtc_state, PREC_PAL_EXT_GC_MAX(pipe, 2), 1 << 16);
+
+	/*
+	 * Program the gc max 2 register to clamp values > 1.0.
+	 * ToDo: Extend the ABI to be able to program values
+	 * from 3.0 to 7.0
+	 */
+	if (DISPLAY_VER(dev_priv) >= 10) {
+		intel_dsb_reg_write(crtc_state, PREC_PAL_EXT2_GC_MAX(pipe, 0),
+				    1 << 16);
+		intel_dsb_reg_write(crtc_state, PREC_PAL_EXT2_GC_MAX(pipe, 1),
+				    1 << 16);
+		intel_dsb_reg_write(crtc_state, PREC_PAL_EXT2_GC_MAX(pipe, 2),
+				    1 << 16);
+	}
+}
+
+static void ivb_load_luts(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	const struct drm_property_blob *gamma_lut = crtc_state->hw.gamma_lut;
+	const struct drm_property_blob *degamma_lut = crtc_state->hw.degamma_lut;
+	const struct drm_property_blob *blob = gamma_lut ?: degamma_lut;
+
+	switch (crtc_state->gamma_mode) {
+	case GAMMA_MODE_MODE_8BIT:
+		ilk_load_lut_8(crtc, blob);
+		break;
+	case GAMMA_MODE_MODE_SPLIT:
+		ivb_load_lut_10(crtc, degamma_lut, PAL_PREC_SPLIT_MODE |
+				PAL_PREC_INDEX_VALUE(0));
+		ivb_load_lut_ext_max(crtc_state);
+		ivb_load_lut_10(crtc, gamma_lut, PAL_PREC_SPLIT_MODE |
+				PAL_PREC_INDEX_VALUE(512));
+		break;
+	case GAMMA_MODE_MODE_10BIT:
+		ivb_load_lut_10(crtc, blob,
+				PAL_PREC_INDEX_VALUE(0));
+		ivb_load_lut_ext_max(crtc_state);
+		break;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		break;
+	}
+}
+
+static void bdw_load_luts(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	const struct drm_property_blob *gamma_lut = crtc_state->hw.gamma_lut;
+	const struct drm_property_blob *degamma_lut = crtc_state->hw.degamma_lut;
+	const struct drm_property_blob *blob = gamma_lut ?: degamma_lut;
+
+	switch (crtc_state->gamma_mode) {
+	case GAMMA_MODE_MODE_8BIT:
+		ilk_load_lut_8(crtc, blob);
+		break;
+	case GAMMA_MODE_MODE_SPLIT:
+		bdw_load_lut_10(crtc, degamma_lut, PAL_PREC_SPLIT_MODE |
+				PAL_PREC_INDEX_VALUE(0));
+		ivb_load_lut_ext_max(crtc_state);
+		bdw_load_lut_10(crtc, gamma_lut, PAL_PREC_SPLIT_MODE |
+				PAL_PREC_INDEX_VALUE(512));
+		break;
+	case GAMMA_MODE_MODE_10BIT:
+
+		bdw_load_lut_10(crtc, blob,
+				PAL_PREC_INDEX_VALUE(0));
+		ivb_load_lut_ext_max(crtc_state);
+		break;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		break;
+	}
+}
+
+static int glk_degamma_lut_size(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 13)
+		return 131;
+	else
+		return 35;
+}
+
+static void glk_load_degamma_lut(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	int i, lut_size = INTEL_INFO(dev_priv)->display.color.degamma_lut_size;
+	const struct drm_color_lut *lut = crtc_state->hw.degamma_lut->data;
+
+	/*
+	 * When setting the auto-increment bit, the hardware seems to
+	 * ignore the index bits, so we need to reset it to index 0
+	 * separately.
+	 */
+	intel_de_write_fw(dev_priv, PRE_CSC_GAMC_INDEX(pipe), 0);
+	intel_de_write_fw(dev_priv, PRE_CSC_GAMC_INDEX(pipe),
+			  PRE_CSC_GAMC_AUTO_INCREMENT);
+
+	for (i = 0; i < lut_size; i++) {
+		/*
+		 * First lut_size entries represent range from 0 to 1.0
+		 * 3 additional lut entries will represent extended range
+		 * inputs 3.0 and 7.0 respectively, currently clamped
+		 * at 1.0. Since the precision is 16bit, the user
+		 * value can be directly filled to register.
+		 * The pipe degamma table in GLK+ onwards doesn't
+		 * support different values per channel, so this just
+		 * programs green value which will be equal to Red and
+		 * Blue into the lut registers.
+		 * ToDo: Extend to max 7.0. Enable 32 bit input value
+		 * as compared to just 16 to achieve this.
+		 */
+		intel_de_write_fw(dev_priv, PRE_CSC_GAMC_DATA(pipe),
+				  lut[i].green);
+	}
+
+	/* Clamp values > 1.0. */
+	while (i++ < glk_degamma_lut_size(dev_priv))
+		intel_de_write_fw(dev_priv, PRE_CSC_GAMC_DATA(pipe), 1 << 16);
+
+	intel_de_write_fw(dev_priv, PRE_CSC_GAMC_INDEX(pipe), 0);
+}
+
+static void glk_load_degamma_lut_linear(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	int i, lut_size = INTEL_INFO(dev_priv)->display.color.degamma_lut_size;
+
+	/*
+	 * When setting the auto-increment bit, the hardware seems to
+	 * ignore the index bits, so we need to reset it to index 0
+	 * separately.
+	 */
+	intel_de_write_fw(dev_priv, PRE_CSC_GAMC_INDEX(pipe), 0);
+	intel_de_write_fw(dev_priv, PRE_CSC_GAMC_INDEX(pipe),
+			  PRE_CSC_GAMC_AUTO_INCREMENT);
+
+	for (i = 0; i < lut_size; i++) {
+		u32 v = (i << 16) / (lut_size - 1);
+
+		intel_de_write_fw(dev_priv, PRE_CSC_GAMC_DATA(pipe), v);
+	}
+
+	/* Clamp values > 1.0. */
+	while (i++ < 35)
+		intel_de_write_fw(dev_priv, PRE_CSC_GAMC_DATA(pipe), 1 << 16);
+
+	intel_de_write_fw(dev_priv, PRE_CSC_GAMC_INDEX(pipe), 0);
+}
+
+static void glk_load_luts(const struct intel_crtc_state *crtc_state)
+{
+	const struct drm_property_blob *gamma_lut = crtc_state->hw.gamma_lut;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	/*
+	 * On GLK+ both pipe CSC and degamma LUT are controlled
+	 * by csc_enable. Hence for the cases where the CSC is
+	 * needed but degamma LUT is not we need to load a
+	 * linear degamma LUT. In fact we'll just always load
+	 * the degama LUT so that we don't have to reload
+	 * it every time the pipe CSC is being enabled.
+	 */
+	if (crtc_state->hw.degamma_lut)
+		glk_load_degamma_lut(crtc_state);
+	else
+		glk_load_degamma_lut_linear(crtc_state);
+
+	switch (crtc_state->gamma_mode) {
+	case GAMMA_MODE_MODE_8BIT:
+		ilk_load_lut_8(crtc, gamma_lut);
+		break;
+	case GAMMA_MODE_MODE_10BIT:
+		bdw_load_lut_10(crtc, gamma_lut, PAL_PREC_INDEX_VALUE(0));
+		ivb_load_lut_ext_max(crtc_state);
+		break;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		break;
+	}
+}
+
+/* ilk+ "12.4" interpolated format (high 10 bits) */
+static u32 ilk_lut_12p4_udw(const struct drm_color_lut *color)
+{
+	return (color->red >> 6) << 20 | (color->green >> 6) << 10 |
+		(color->blue >> 6);
+}
+
+/* ilk+ "12.4" interpolated format (low 6 bits) */
+static u32 ilk_lut_12p4_ldw(const struct drm_color_lut *color)
+{
+	return (color->red & 0x3f) << 24 | (color->green & 0x3f) << 14 |
+		(color->blue & 0x3f) << 4;
+}
+
+static void
+icl_load_gcmax(const struct intel_crtc_state *crtc_state,
+	       const struct drm_color_lut *color)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum pipe pipe = crtc->pipe;
+
+	/* FIXME LUT entries are 16 bit only, so we can prog 0xFFFF max */
+	intel_dsb_reg_write(crtc_state, PREC_PAL_GC_MAX(pipe, 0), color->red);
+	intel_dsb_reg_write(crtc_state, PREC_PAL_GC_MAX(pipe, 1), color->green);
+	intel_dsb_reg_write(crtc_state, PREC_PAL_GC_MAX(pipe, 2), color->blue);
+}
+
+static void
+icl_program_gamma_superfine_segment(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	const struct drm_property_blob *blob = crtc_state->hw.gamma_lut;
+	const struct drm_color_lut *lut = blob->data;
+	enum pipe pipe = crtc->pipe;
+	int i;
+
+	/*
+	 * Program Super Fine segment (let's call it seg1)...
+	 *
+	 * Super Fine segment's step is 1/(8 * 128 * 256) and it has
+	 * 9 entries, corresponding to values 0, 1/(8 * 128 * 256),
+	 * 2/(8 * 128 * 256) ... 8/(8 * 128 * 256).
+	 */
+	intel_dsb_reg_write(crtc_state, PREC_PAL_MULTI_SEG_INDEX(pipe),
+			    PAL_PREC_AUTO_INCREMENT);
+
+	for (i = 0; i < 9; i++) {
+		const struct drm_color_lut *entry = &lut[i];
+
+		intel_dsb_indexed_reg_write(crtc_state, PREC_PAL_MULTI_SEG_DATA(pipe),
+					    ilk_lut_12p4_ldw(entry));
+		intel_dsb_indexed_reg_write(crtc_state, PREC_PAL_MULTI_SEG_DATA(pipe),
+					    ilk_lut_12p4_udw(entry));
+	}
+}
+
+static void
+icl_program_gamma_multi_segment(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	const struct drm_property_blob *blob = crtc_state->hw.gamma_lut;
+	const struct drm_color_lut *lut = blob->data;
+	const struct drm_color_lut *entry;
+	enum pipe pipe = crtc->pipe;
+	int i;
+
+	/*
+	 * Program Fine segment (let's call it seg2)...
+	 *
+	 * Fine segment's step is 1/(128 * 256) i.e. 1/(128 * 256), 2/(128 * 256)
+	 * ... 256/(128 * 256). So in order to program fine segment of LUT we
+	 * need to pick every 8th entry in the LUT, and program 256 indexes.
+	 *
+	 * PAL_PREC_INDEX[0] and PAL_PREC_INDEX[1] map to seg2[1],
+	 * seg2[0] being unused by the hardware.
+	 */
+	intel_dsb_reg_write(crtc_state, PREC_PAL_INDEX(pipe),
+			    PAL_PREC_AUTO_INCREMENT);
+	for (i = 1; i < 257; i++) {
+		entry = &lut[i * 8];
+		intel_dsb_indexed_reg_write(crtc_state, PREC_PAL_DATA(pipe),
+					    ilk_lut_12p4_ldw(entry));
+		intel_dsb_indexed_reg_write(crtc_state, PREC_PAL_DATA(pipe),
+					    ilk_lut_12p4_udw(entry));
+	}
+
+	/*
+	 * Program Coarse segment (let's call it seg3)...
+	 *
+	 * Coarse segment starts from index 0 and it's step is 1/256 ie 0,
+	 * 1/256, 2/256 ... 256/256. As per the description of each entry in LUT
+	 * above, we need to pick every (8 * 128)th entry in LUT, and
+	 * program 256 of those.
+	 *
+	 * Spec is not very clear about if entries seg3[0] and seg3[1] are
+	 * being used or not, but we still need to program these to advance
+	 * the index.
+	 */
+	for (i = 0; i < 256; i++) {
+		entry = &lut[i * 8 * 128];
+		intel_dsb_indexed_reg_write(crtc_state, PREC_PAL_DATA(pipe),
+					    ilk_lut_12p4_ldw(entry));
+		intel_dsb_indexed_reg_write(crtc_state, PREC_PAL_DATA(pipe),
+					    ilk_lut_12p4_udw(entry));
+	}
+
+	/* The last entry in the LUT is to be programmed in GCMAX */
+	entry = &lut[256 * 8 * 128];
+	icl_load_gcmax(crtc_state, entry);
+	ivb_load_lut_ext_max(crtc_state);
+}
+
+static void icl_load_luts(const struct intel_crtc_state *crtc_state)
+{
+	const struct drm_property_blob *gamma_lut = crtc_state->hw.gamma_lut;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	if (crtc_state->hw.degamma_lut)
+		glk_load_degamma_lut(crtc_state);
+
+	switch (crtc_state->gamma_mode & GAMMA_MODE_MODE_MASK) {
+	case GAMMA_MODE_MODE_8BIT:
+		ilk_load_lut_8(crtc, gamma_lut);
+		break;
+	case GAMMA_MODE_MODE_12BIT_MULTI_SEGMENTED:
+		icl_program_gamma_superfine_segment(crtc_state);
+		icl_program_gamma_multi_segment(crtc_state);
+		break;
+	case GAMMA_MODE_MODE_10BIT:
+		bdw_load_lut_10(crtc, gamma_lut, PAL_PREC_INDEX_VALUE(0));
+		ivb_load_lut_ext_max(crtc_state);
+		break;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		break;
+	}
+
+	intel_dsb_commit(crtc_state);
+}
+
+static u32 chv_cgm_degamma_ldw(const struct drm_color_lut *color)
+{
+	return drm_color_lut_extract(color->green, 14) << 16 |
+		drm_color_lut_extract(color->blue, 14);
+}
+
+static u32 chv_cgm_degamma_udw(const struct drm_color_lut *color)
+{
+	return drm_color_lut_extract(color->red, 14);
+}
+
+static void chv_load_cgm_degamma(struct intel_crtc *crtc,
+				 const struct drm_property_blob *blob)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_color_lut *lut = blob->data;
+	int i, lut_size = drm_color_lut_size(blob);
+	enum pipe pipe = crtc->pipe;
+
+	for (i = 0; i < lut_size; i++) {
+		intel_de_write_fw(dev_priv, CGM_PIPE_DEGAMMA(pipe, i, 0),
+				  chv_cgm_degamma_ldw(&lut[i]));
+		intel_de_write_fw(dev_priv, CGM_PIPE_DEGAMMA(pipe, i, 1),
+				  chv_cgm_degamma_udw(&lut[i]));
+	}
+}
+
+static u32 chv_cgm_gamma_ldw(const struct drm_color_lut *color)
+{
+	return drm_color_lut_extract(color->green, 10) << 16 |
+		drm_color_lut_extract(color->blue, 10);
+}
+
+static u32 chv_cgm_gamma_udw(const struct drm_color_lut *color)
+{
+	return drm_color_lut_extract(color->red, 10);
+}
+
+static void chv_cgm_gamma_pack(struct drm_color_lut *entry, u32 ldw, u32 udw)
+{
+	entry->green = intel_color_lut_pack(REG_FIELD_GET(CGM_PIPE_GAMMA_GREEN_MASK, ldw), 10);
+	entry->blue = intel_color_lut_pack(REG_FIELD_GET(CGM_PIPE_GAMMA_BLUE_MASK, ldw), 10);
+	entry->red = intel_color_lut_pack(REG_FIELD_GET(CGM_PIPE_GAMMA_RED_MASK, udw), 10);
+}
+
+static void chv_load_cgm_gamma(struct intel_crtc *crtc,
+			       const struct drm_property_blob *blob)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_color_lut *lut = blob->data;
+	int i, lut_size = drm_color_lut_size(blob);
+	enum pipe pipe = crtc->pipe;
+
+	for (i = 0; i < lut_size; i++) {
+		intel_de_write_fw(dev_priv, CGM_PIPE_GAMMA(pipe, i, 0),
+				  chv_cgm_gamma_ldw(&lut[i]));
+		intel_de_write_fw(dev_priv, CGM_PIPE_GAMMA(pipe, i, 1),
+				  chv_cgm_gamma_udw(&lut[i]));
+	}
+}
+
+static void chv_load_luts(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_property_blob *degamma_lut = crtc_state->hw.degamma_lut;
+	const struct drm_property_blob *gamma_lut = crtc_state->hw.gamma_lut;
+	const struct drm_property_blob *ctm = crtc_state->hw.ctm;
+
+	if (crtc_state->cgm_mode & CGM_PIPE_MODE_CSC)
+		chv_load_cgm_csc(crtc, ctm);
+
+	if (crtc_state->cgm_mode & CGM_PIPE_MODE_DEGAMMA)
+		chv_load_cgm_degamma(crtc, degamma_lut);
+
+	if (crtc_state->cgm_mode & CGM_PIPE_MODE_GAMMA)
+		chv_load_cgm_gamma(crtc, gamma_lut);
+	else
+		i965_load_luts(crtc_state);
+
+	intel_de_write_fw(dev_priv, CGM_PIPE_MODE(crtc->pipe),
+			  crtc_state->cgm_mode);
+}
+
+void intel_color_load_luts(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	dev_priv->display.funcs.color->load_luts(crtc_state);
+}
+
+void intel_color_commit_noarm(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	if (dev_priv->display.funcs.color->color_commit_noarm)
+		dev_priv->display.funcs.color->color_commit_noarm(crtc_state);
+}
+
+void intel_color_commit_arm(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	dev_priv->display.funcs.color->color_commit_arm(crtc_state);
+}
+
+static bool intel_can_preload_luts(const struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(new_crtc_state->uapi.state);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+
+	return !old_crtc_state->hw.gamma_lut &&
+		!old_crtc_state->hw.degamma_lut;
+}
+
+static bool chv_can_preload_luts(const struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(new_crtc_state->uapi.state);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+
+	/*
+	 * CGM_PIPE_MODE is itself single buffered. We'd have to
+	 * somehow split it out from chv_load_luts() if we wanted
+	 * the ability to preload the CGM LUTs/CSC without tearing.
+	 */
+	if (old_crtc_state->cgm_mode || new_crtc_state->cgm_mode)
+		return false;
+
+	return !old_crtc_state->hw.gamma_lut;
+}
+
+static bool glk_can_preload_luts(const struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(new_crtc_state->uapi.state);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+
+	/*
+	 * The hardware degamma is active whenever the pipe
+	 * CSC is active. Thus even if the old state has no
+	 * software degamma we need to avoid clobbering the
+	 * linear hardware degamma mid scanout.
+	 */
+	return !old_crtc_state->csc_enable &&
+		!old_crtc_state->hw.gamma_lut;
+}
+
+int intel_color_check(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	return dev_priv->display.funcs.color->color_check(crtc_state);
+}
+
+void intel_color_get_config(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	if (dev_priv->display.funcs.color->read_luts)
+		dev_priv->display.funcs.color->read_luts(crtc_state);
+}
+
+static bool need_plane_update(struct intel_plane *plane,
+			      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+
+	/*
+	 * On pre-SKL the pipe gamma enable and pipe csc enable for
+	 * the pipe bottom color are configured via the primary plane.
+	 * We have to reconfigure that even if the plane is inactive.
+	 */
+	return crtc_state->active_planes & BIT(plane->id) ||
+		(DISPLAY_VER(dev_priv) < 9 &&
+		 plane->id == PLANE_PRIMARY);
+}
+
+static int
+intel_color_add_affected_planes(struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(new_crtc_state->uapi.state);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct intel_plane *plane;
+
+	if (!new_crtc_state->hw.active ||
+	    drm_atomic_crtc_needs_modeset(&new_crtc_state->uapi))
+		return 0;
+
+	if (new_crtc_state->gamma_enable == old_crtc_state->gamma_enable &&
+	    new_crtc_state->csc_enable == old_crtc_state->csc_enable)
+		return 0;
+
+	for_each_intel_plane_on_crtc(&dev_priv->drm, crtc, plane) {
+		struct intel_plane_state *plane_state;
+
+		if (!need_plane_update(plane, new_crtc_state))
+			continue;
+
+		plane_state = intel_atomic_get_plane_state(state, plane);
+		if (IS_ERR(plane_state))
+			return PTR_ERR(plane_state);
+
+		new_crtc_state->update_planes |= BIT(plane->id);
+	}
+
+	return 0;
+}
+
+static int check_lut_size(const struct drm_property_blob *lut, int expected)
+{
+	int len;
+
+	if (!lut)
+		return 0;
+
+	len = drm_color_lut_size(lut);
+	if (len != expected) {
+		DRM_DEBUG_KMS("Invalid LUT size; got %d, expected %d\n",
+			      len, expected);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int check_luts(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	const struct drm_property_blob *gamma_lut = crtc_state->hw.gamma_lut;
+	const struct drm_property_blob *degamma_lut = crtc_state->hw.degamma_lut;
+	int gamma_length, degamma_length;
+	u32 gamma_tests, degamma_tests;
+
+	/* Always allow legacy gamma LUT with no further checking. */
+	if (crtc_state_is_legacy_gamma(crtc_state))
+		return 0;
+
+	/* C8 relies on its palette being stored in the legacy LUT */
+	if (crtc_state->c8_planes) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "C8 pixelformat requires the legacy LUT\n");
+		return -EINVAL;
+	}
+
+	degamma_length = INTEL_INFO(dev_priv)->display.color.degamma_lut_size;
+	gamma_length = INTEL_INFO(dev_priv)->display.color.gamma_lut_size;
+	degamma_tests = INTEL_INFO(dev_priv)->display.color.degamma_lut_tests;
+	gamma_tests = INTEL_INFO(dev_priv)->display.color.gamma_lut_tests;
+
+	if (check_lut_size(degamma_lut, degamma_length) ||
+	    check_lut_size(gamma_lut, gamma_length))
+		return -EINVAL;
+
+	if (drm_color_lut_check(degamma_lut, degamma_tests) ||
+	    drm_color_lut_check(gamma_lut, gamma_tests))
+		return -EINVAL;
+
+	return 0;
+}
+
+static u32 i9xx_gamma_mode(struct intel_crtc_state *crtc_state)
+{
+	if (!crtc_state->gamma_enable ||
+	    crtc_state_is_legacy_gamma(crtc_state))
+		return GAMMA_MODE_MODE_8BIT;
+	else
+		return GAMMA_MODE_MODE_10BIT; /* i965+ only */
+}
+
+static int i9xx_color_check(struct intel_crtc_state *crtc_state)
+{
+	int ret;
+
+	ret = check_luts(crtc_state);
+	if (ret)
+		return ret;
+
+	crtc_state->gamma_enable =
+		crtc_state->hw.gamma_lut &&
+		!crtc_state->c8_planes;
+
+	crtc_state->gamma_mode = i9xx_gamma_mode(crtc_state);
+
+	ret = intel_color_add_affected_planes(crtc_state);
+	if (ret)
+		return ret;
+
+	crtc_state->preload_luts = intel_can_preload_luts(crtc_state);
+
+	return 0;
+}
+
+static u32 chv_cgm_mode(const struct intel_crtc_state *crtc_state)
+{
+	u32 cgm_mode = 0;
+
+	if (crtc_state_is_legacy_gamma(crtc_state))
+		return 0;
+
+	if (crtc_state->hw.degamma_lut)
+		cgm_mode |= CGM_PIPE_MODE_DEGAMMA;
+	if (crtc_state->hw.ctm)
+		cgm_mode |= CGM_PIPE_MODE_CSC;
+	if (crtc_state->hw.gamma_lut)
+		cgm_mode |= CGM_PIPE_MODE_GAMMA;
+
+	return cgm_mode;
+}
+
+/*
+ * CHV color pipeline:
+ * u0.10 -> CGM degamma -> u0.14 -> CGM csc -> u0.14 -> CGM gamma ->
+ * u0.10 -> WGC csc -> u0.10 -> pipe gamma -> u0.10
+ *
+ * We always bypass the WGC csc and use the CGM csc
+ * instead since it has degamma and better precision.
+ */
+static int chv_color_check(struct intel_crtc_state *crtc_state)
+{
+	int ret;
+
+	ret = check_luts(crtc_state);
+	if (ret)
+		return ret;
+
+	/*
+	 * Pipe gamma will be used only for the legacy LUT.
+	 * Otherwise we bypass it and use the CGM gamma instead.
+	 */
+	crtc_state->gamma_enable =
+		crtc_state_is_legacy_gamma(crtc_state) &&
+		!crtc_state->c8_planes;
+
+	crtc_state->gamma_mode = GAMMA_MODE_MODE_8BIT;
+
+	crtc_state->cgm_mode = chv_cgm_mode(crtc_state);
+
+	ret = intel_color_add_affected_planes(crtc_state);
+	if (ret)
+		return ret;
+
+	crtc_state->preload_luts = chv_can_preload_luts(crtc_state);
+
+	return 0;
+}
+
+static u32 ilk_gamma_mode(const struct intel_crtc_state *crtc_state)
+{
+	if (!crtc_state->gamma_enable ||
+	    crtc_state_is_legacy_gamma(crtc_state))
+		return GAMMA_MODE_MODE_8BIT;
+	else
+		return GAMMA_MODE_MODE_10BIT;
+}
+
+static u32 ilk_csc_mode(const struct intel_crtc_state *crtc_state)
+{
+	/*
+	 * CSC comes after the LUT in RGB->YCbCr mode.
+	 * RGB->YCbCr needs the limited range offsets added to
+	 * the output. RGB limited range output is handled by
+	 * the hw automagically elsewhere.
+	 */
+	if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB)
+		return CSC_BLACK_SCREEN_OFFSET;
+
+	return CSC_MODE_YUV_TO_RGB |
+		CSC_POSITION_BEFORE_GAMMA;
+}
+
+static int ilk_color_check(struct intel_crtc_state *crtc_state)
+{
+	int ret;
+
+	ret = check_luts(crtc_state);
+	if (ret)
+		return ret;
+
+	crtc_state->gamma_enable =
+		crtc_state->hw.gamma_lut &&
+		!crtc_state->c8_planes;
+
+	/*
+	 * We don't expose the ctm on ilk/snb currently, also RGB
+	 * limited range output is handled by the hw automagically.
+	 */
+	crtc_state->csc_enable =
+		crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB;
+
+	crtc_state->gamma_mode = ilk_gamma_mode(crtc_state);
+
+	crtc_state->csc_mode = ilk_csc_mode(crtc_state);
+
+	ret = intel_color_add_affected_planes(crtc_state);
+	if (ret)
+		return ret;
+
+	crtc_state->preload_luts = intel_can_preload_luts(crtc_state);
+
+	return 0;
+}
+
+static u32 ivb_gamma_mode(const struct intel_crtc_state *crtc_state)
+{
+	if (!crtc_state->gamma_enable ||
+	    crtc_state_is_legacy_gamma(crtc_state))
+		return GAMMA_MODE_MODE_8BIT;
+	else if (crtc_state->hw.gamma_lut &&
+		 crtc_state->hw.degamma_lut)
+		return GAMMA_MODE_MODE_SPLIT;
+	else
+		return GAMMA_MODE_MODE_10BIT;
+}
+
+static u32 ivb_csc_mode(const struct intel_crtc_state *crtc_state)
+{
+	bool limited_color_range = ilk_csc_limited_range(crtc_state);
+
+	/*
+	 * CSC comes after the LUT in degamma, RGB->YCbCr,
+	 * and RGB full->limited range mode.
+	 */
+	if (crtc_state->hw.degamma_lut ||
+	    crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB ||
+	    limited_color_range)
+		return 0;
+
+	return CSC_POSITION_BEFORE_GAMMA;
+}
+
+static int ivb_color_check(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	bool limited_color_range = ilk_csc_limited_range(crtc_state);
+	int ret;
+
+	ret = check_luts(crtc_state);
+	if (ret)
+		return ret;
+
+	if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB &&
+	    crtc_state->hw.ctm) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "YCBCR and CTM together are not possible\n");
+		return -EINVAL;
+	}
+
+	crtc_state->gamma_enable =
+		(crtc_state->hw.gamma_lut ||
+		 crtc_state->hw.degamma_lut) &&
+		!crtc_state->c8_planes;
+
+	crtc_state->csc_enable =
+		crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB ||
+		crtc_state->hw.ctm || limited_color_range;
+
+	crtc_state->gamma_mode = ivb_gamma_mode(crtc_state);
+
+	crtc_state->csc_mode = ivb_csc_mode(crtc_state);
+
+	ret = intel_color_add_affected_planes(crtc_state);
+	if (ret)
+		return ret;
+
+	crtc_state->preload_luts = intel_can_preload_luts(crtc_state);
+
+	return 0;
+}
+
+static u32 glk_gamma_mode(const struct intel_crtc_state *crtc_state)
+{
+	if (!crtc_state->gamma_enable ||
+	    crtc_state_is_legacy_gamma(crtc_state))
+		return GAMMA_MODE_MODE_8BIT;
+	else
+		return GAMMA_MODE_MODE_10BIT;
+}
+
+static int glk_color_check(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	int ret;
+
+	ret = check_luts(crtc_state);
+	if (ret)
+		return ret;
+
+	if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB &&
+	    crtc_state->hw.ctm) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "YCBCR and CTM together are not possible\n");
+		return -EINVAL;
+	}
+
+	crtc_state->gamma_enable =
+		crtc_state->hw.gamma_lut &&
+		!crtc_state->c8_planes;
+
+	/* On GLK+ degamma LUT is controlled by csc_enable */
+	crtc_state->csc_enable =
+		crtc_state->hw.degamma_lut ||
+		crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB ||
+		crtc_state->hw.ctm || crtc_state->limited_color_range;
+
+	crtc_state->gamma_mode = glk_gamma_mode(crtc_state);
+
+	crtc_state->csc_mode = 0;
+
+	ret = intel_color_add_affected_planes(crtc_state);
+	if (ret)
+		return ret;
+
+	crtc_state->preload_luts = glk_can_preload_luts(crtc_state);
+
+	return 0;
+}
+
+static u32 icl_gamma_mode(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	u32 gamma_mode = 0;
+
+	if (crtc_state->hw.degamma_lut)
+		gamma_mode |= PRE_CSC_GAMMA_ENABLE;
+
+	if (crtc_state->hw.gamma_lut &&
+	    !crtc_state->c8_planes)
+		gamma_mode |= POST_CSC_GAMMA_ENABLE;
+
+	if (!crtc_state->hw.gamma_lut ||
+	    crtc_state_is_legacy_gamma(crtc_state))
+		gamma_mode |= GAMMA_MODE_MODE_8BIT;
+	/*
+	 * Enable 10bit gamma for D13
+	 * ToDo: Extend to Logarithmic Gamma once the new UAPI
+	 * is accepted and implemented by a userspace consumer
+	 */
+	else if (DISPLAY_VER(i915) >= 13)
+		gamma_mode |= GAMMA_MODE_MODE_10BIT;
+	else
+		gamma_mode |= GAMMA_MODE_MODE_12BIT_MULTI_SEGMENTED;
+
+	return gamma_mode;
+}
+
+static u32 icl_csc_mode(const struct intel_crtc_state *crtc_state)
+{
+	u32 csc_mode = 0;
+
+	if (crtc_state->hw.ctm)
+		csc_mode |= ICL_CSC_ENABLE;
+
+	if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB ||
+	    crtc_state->limited_color_range)
+		csc_mode |= ICL_OUTPUT_CSC_ENABLE;
+
+	return csc_mode;
+}
+
+static int icl_color_check(struct intel_crtc_state *crtc_state)
+{
+	int ret;
+
+	ret = check_luts(crtc_state);
+	if (ret)
+		return ret;
+
+	crtc_state->gamma_mode = icl_gamma_mode(crtc_state);
+
+	crtc_state->csc_mode = icl_csc_mode(crtc_state);
+
+	crtc_state->preload_luts = intel_can_preload_luts(crtc_state);
+
+	return 0;
+}
+
+static int i9xx_gamma_precision(const struct intel_crtc_state *crtc_state)
+{
+	if (!crtc_state->gamma_enable)
+		return 0;
+
+	switch (crtc_state->gamma_mode) {
+	case GAMMA_MODE_MODE_8BIT:
+		return 8;
+	case GAMMA_MODE_MODE_10BIT:
+		return 16;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		return 0;
+	}
+}
+
+static int ilk_gamma_precision(const struct intel_crtc_state *crtc_state)
+{
+	if (!crtc_state->gamma_enable)
+		return 0;
+
+	if ((crtc_state->csc_mode & CSC_POSITION_BEFORE_GAMMA) == 0)
+		return 0;
+
+	switch (crtc_state->gamma_mode) {
+	case GAMMA_MODE_MODE_8BIT:
+		return 8;
+	case GAMMA_MODE_MODE_10BIT:
+		return 10;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		return 0;
+	}
+}
+
+static int chv_gamma_precision(const struct intel_crtc_state *crtc_state)
+{
+	if (crtc_state->cgm_mode & CGM_PIPE_MODE_GAMMA)
+		return 10;
+	else
+		return i9xx_gamma_precision(crtc_state);
+}
+
+static int glk_gamma_precision(const struct intel_crtc_state *crtc_state)
+{
+	if (!crtc_state->gamma_enable)
+		return 0;
+
+	switch (crtc_state->gamma_mode) {
+	case GAMMA_MODE_MODE_8BIT:
+		return 8;
+	case GAMMA_MODE_MODE_10BIT:
+		return 10;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		return 0;
+	}
+}
+
+static int icl_gamma_precision(const struct intel_crtc_state *crtc_state)
+{
+	if ((crtc_state->gamma_mode & POST_CSC_GAMMA_ENABLE) == 0)
+		return 0;
+
+	switch (crtc_state->gamma_mode & GAMMA_MODE_MODE_MASK) {
+	case GAMMA_MODE_MODE_8BIT:
+		return 8;
+	case GAMMA_MODE_MODE_10BIT:
+		return 10;
+	case GAMMA_MODE_MODE_12BIT_MULTI_SEGMENTED:
+		return 16;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		return 0;
+	}
+}
+
+int intel_color_get_gamma_bit_precision(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	if (HAS_GMCH(dev_priv)) {
+		if (IS_CHERRYVIEW(dev_priv))
+			return chv_gamma_precision(crtc_state);
+		else
+			return i9xx_gamma_precision(crtc_state);
+	} else {
+		if (DISPLAY_VER(dev_priv) >= 11)
+			return icl_gamma_precision(crtc_state);
+		else if (DISPLAY_VER(dev_priv) == 10)
+			return glk_gamma_precision(crtc_state);
+		else if (IS_IRONLAKE(dev_priv))
+			return ilk_gamma_precision(crtc_state);
+	}
+
+	return 0;
+}
+
+static bool err_check(struct drm_color_lut *lut1,
+		      struct drm_color_lut *lut2, u32 err)
+{
+	return ((abs((long)lut2->red - lut1->red)) <= err) &&
+		((abs((long)lut2->blue - lut1->blue)) <= err) &&
+		((abs((long)lut2->green - lut1->green)) <= err);
+}
+
+static bool intel_color_lut_entries_equal(struct drm_color_lut *lut1,
+					  struct drm_color_lut *lut2,
+					  int lut_size, u32 err)
+{
+	int i;
+
+	for (i = 0; i < lut_size; i++) {
+		if (!err_check(&lut1[i], &lut2[i], err))
+			return false;
+	}
+
+	return true;
+}
+
+bool intel_color_lut_equal(struct drm_property_blob *blob1,
+			   struct drm_property_blob *blob2,
+			   u32 gamma_mode, u32 bit_precision)
+{
+	struct drm_color_lut *lut1, *lut2;
+	int lut_size1, lut_size2;
+	u32 err;
+
+	if (!blob1 != !blob2)
+		return false;
+
+	if (!blob1)
+		return true;
+
+	lut_size1 = drm_color_lut_size(blob1);
+	lut_size2 = drm_color_lut_size(blob2);
+
+	/* check sw and hw lut size */
+	if (lut_size1 != lut_size2)
+		return false;
+
+	lut1 = blob1->data;
+	lut2 = blob2->data;
+
+	err = 0xffff >> bit_precision;
+
+	/* check sw and hw lut entry to be equal */
+	switch (gamma_mode & GAMMA_MODE_MODE_MASK) {
+	case GAMMA_MODE_MODE_8BIT:
+	case GAMMA_MODE_MODE_10BIT:
+		if (!intel_color_lut_entries_equal(lut1, lut2,
+						   lut_size2, err))
+			return false;
+		break;
+	case GAMMA_MODE_MODE_12BIT_MULTI_SEGMENTED:
+		if (!intel_color_lut_entries_equal(lut1, lut2,
+						   9, err))
+			return false;
+		break;
+	default:
+		MISSING_CASE(gamma_mode);
+		return false;
+	}
+
+	return true;
+}
+
+static struct drm_property_blob *i9xx_read_lut_8(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	struct drm_property_blob *blob;
+	struct drm_color_lut *lut;
+	int i;
+
+	blob = drm_property_create_blob(&dev_priv->drm,
+					sizeof(struct drm_color_lut) * LEGACY_LUT_LENGTH,
+					NULL);
+	if (IS_ERR(blob))
+		return NULL;
+
+	lut = blob->data;
+
+	for (i = 0; i < LEGACY_LUT_LENGTH; i++) {
+		u32 val = intel_de_read_fw(dev_priv, PALETTE(pipe, i));
+
+		i9xx_lut_8_pack(&lut[i], val);
+	}
+
+	return blob;
+}
+
+static void i9xx_read_luts(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	if (!crtc_state->gamma_enable)
+		return;
+
+	crtc_state->hw.gamma_lut = i9xx_read_lut_8(crtc);
+}
+
+static struct drm_property_blob *i965_read_lut_10p6(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	int i, lut_size = INTEL_INFO(dev_priv)->display.color.gamma_lut_size;
+	enum pipe pipe = crtc->pipe;
+	struct drm_property_blob *blob;
+	struct drm_color_lut *lut;
+
+	blob = drm_property_create_blob(&dev_priv->drm,
+					sizeof(struct drm_color_lut) * lut_size,
+					NULL);
+	if (IS_ERR(blob))
+		return NULL;
+
+	lut = blob->data;
+
+	for (i = 0; i < lut_size - 1; i++) {
+		u32 ldw = intel_de_read_fw(dev_priv, PALETTE(pipe, 2 * i + 0));
+		u32 udw = intel_de_read_fw(dev_priv, PALETTE(pipe, 2 * i + 1));
+
+		i965_lut_10p6_pack(&lut[i], ldw, udw);
+	}
+
+	lut[i].red = i965_lut_11p6_max_pack(intel_de_read_fw(dev_priv, PIPEGCMAX(pipe, 0)));
+	lut[i].green = i965_lut_11p6_max_pack(intel_de_read_fw(dev_priv, PIPEGCMAX(pipe, 1)));
+	lut[i].blue = i965_lut_11p6_max_pack(intel_de_read_fw(dev_priv, PIPEGCMAX(pipe, 2)));
+
+	return blob;
+}
+
+static void i965_read_luts(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	if (!crtc_state->gamma_enable)
+		return;
+
+	if (crtc_state->gamma_mode == GAMMA_MODE_MODE_8BIT)
+		crtc_state->hw.gamma_lut = i9xx_read_lut_8(crtc);
+	else
+		crtc_state->hw.gamma_lut = i965_read_lut_10p6(crtc);
+}
+
+static struct drm_property_blob *chv_read_cgm_gamma(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	int i, lut_size = INTEL_INFO(dev_priv)->display.color.gamma_lut_size;
+	enum pipe pipe = crtc->pipe;
+	struct drm_property_blob *blob;
+	struct drm_color_lut *lut;
+
+	blob = drm_property_create_blob(&dev_priv->drm,
+					sizeof(struct drm_color_lut) * lut_size,
+					NULL);
+	if (IS_ERR(blob))
+		return NULL;
+
+	lut = blob->data;
+
+	for (i = 0; i < lut_size; i++) {
+		u32 ldw = intel_de_read_fw(dev_priv, CGM_PIPE_GAMMA(pipe, i, 0));
+		u32 udw = intel_de_read_fw(dev_priv, CGM_PIPE_GAMMA(pipe, i, 1));
+
+		chv_cgm_gamma_pack(&lut[i], ldw, udw);
+	}
+
+	return blob;
+}
+
+static void chv_read_luts(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	if (crtc_state->cgm_mode & CGM_PIPE_MODE_GAMMA)
+		crtc_state->hw.gamma_lut = chv_read_cgm_gamma(crtc);
+	else
+		i965_read_luts(crtc_state);
+}
+
+static struct drm_property_blob *ilk_read_lut_8(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	struct drm_property_blob *blob;
+	struct drm_color_lut *lut;
+	int i;
+
+	blob = drm_property_create_blob(&dev_priv->drm,
+					sizeof(struct drm_color_lut) * LEGACY_LUT_LENGTH,
+					NULL);
+	if (IS_ERR(blob))
+		return NULL;
+
+	lut = blob->data;
+
+	for (i = 0; i < LEGACY_LUT_LENGTH; i++) {
+		u32 val = intel_de_read_fw(dev_priv, LGC_PALETTE(pipe, i));
+
+		i9xx_lut_8_pack(&lut[i], val);
+	}
+
+	return blob;
+}
+
+static struct drm_property_blob *ilk_read_lut_10(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	int i, lut_size = INTEL_INFO(dev_priv)->display.color.gamma_lut_size;
+	enum pipe pipe = crtc->pipe;
+	struct drm_property_blob *blob;
+	struct drm_color_lut *lut;
+
+	blob = drm_property_create_blob(&dev_priv->drm,
+					sizeof(struct drm_color_lut) * lut_size,
+					NULL);
+	if (IS_ERR(blob))
+		return NULL;
+
+	lut = blob->data;
+
+	for (i = 0; i < lut_size; i++) {
+		u32 val = intel_de_read_fw(dev_priv, PREC_PALETTE(pipe, i));
+
+		ilk_lut_10_pack(&lut[i], val);
+	}
+
+	return blob;
+}
+
+static void ilk_read_luts(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	if (!crtc_state->gamma_enable)
+		return;
+
+	if ((crtc_state->csc_mode & CSC_POSITION_BEFORE_GAMMA) == 0)
+		return;
+
+	switch (crtc_state->gamma_mode) {
+	case GAMMA_MODE_MODE_8BIT:
+		crtc_state->hw.gamma_lut = ilk_read_lut_8(crtc);
+		break;
+	case GAMMA_MODE_MODE_10BIT:
+		crtc_state->hw.gamma_lut = ilk_read_lut_10(crtc);
+		break;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		break;
+	}
+}
+
+/* On BDW+ the index auto increment mode actually works */
+static struct drm_property_blob *bdw_read_lut_10(struct intel_crtc *crtc,
+						 u32 prec_index)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	int i, hw_lut_size = ivb_lut_10_size(prec_index);
+	int lut_size = INTEL_INFO(dev_priv)->display.color.gamma_lut_size;
+	enum pipe pipe = crtc->pipe;
+	struct drm_property_blob *blob;
+	struct drm_color_lut *lut;
+
+	drm_WARN_ON(&dev_priv->drm, lut_size != hw_lut_size);
+
+	blob = drm_property_create_blob(&dev_priv->drm,
+					sizeof(struct drm_color_lut) * lut_size,
+					NULL);
+	if (IS_ERR(blob))
+		return NULL;
+
+	lut = blob->data;
+
+	intel_de_write_fw(dev_priv, PREC_PAL_INDEX(pipe),
+			  prec_index | PAL_PREC_AUTO_INCREMENT);
+
+	for (i = 0; i < lut_size; i++) {
+		u32 val = intel_de_read_fw(dev_priv, PREC_PAL_DATA(pipe));
+
+		ilk_lut_10_pack(&lut[i], val);
+	}
+
+	intel_de_write_fw(dev_priv, PREC_PAL_INDEX(pipe), 0);
+
+	return blob;
+}
+
+static void glk_read_luts(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	if (!crtc_state->gamma_enable)
+		return;
+
+	switch (crtc_state->gamma_mode) {
+	case GAMMA_MODE_MODE_8BIT:
+		crtc_state->hw.gamma_lut = ilk_read_lut_8(crtc);
+		break;
+	case GAMMA_MODE_MODE_10BIT:
+		crtc_state->hw.gamma_lut = bdw_read_lut_10(crtc, PAL_PREC_INDEX_VALUE(0));
+		break;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		break;
+	}
+}
+
+static void icl_color_commit_arm(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	/*
+	 * We don't (yet) allow userspace to control the pipe background color,
+	 * so force it to black.
+	 */
+	intel_de_write(i915, SKL_BOTTOM_COLOR(pipe), 0);
+
+	intel_de_write(i915, GAMMA_MODE(crtc->pipe),
+		       crtc_state->gamma_mode);
+
+	intel_de_write_fw(i915, PIPE_CSC_MODE(crtc->pipe),
+			  crtc_state->csc_mode);
+}
+
+static struct drm_property_blob *
+icl_read_lut_multi_segment(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	int i, lut_size = INTEL_INFO(dev_priv)->display.color.gamma_lut_size;
+	enum pipe pipe = crtc->pipe;
+	struct drm_property_blob *blob;
+	struct drm_color_lut *lut;
+
+	blob = drm_property_create_blob(&dev_priv->drm,
+					sizeof(struct drm_color_lut) * lut_size,
+					NULL);
+	if (IS_ERR(blob))
+		return NULL;
+
+	lut = blob->data;
+
+	intel_de_write_fw(dev_priv, PREC_PAL_MULTI_SEG_INDEX(pipe),
+			  PAL_PREC_AUTO_INCREMENT);
+
+	for (i = 0; i < 9; i++) {
+		u32 ldw = intel_de_read_fw(dev_priv, PREC_PAL_MULTI_SEG_DATA(pipe));
+		u32 udw = intel_de_read_fw(dev_priv, PREC_PAL_MULTI_SEG_DATA(pipe));
+
+		icl_lut_multi_seg_pack(&lut[i], ldw, udw);
+	}
+
+	intel_de_write_fw(dev_priv, PREC_PAL_MULTI_SEG_INDEX(pipe), 0);
+
+	/*
+	 * FIXME readouts from PAL_PREC_DATA register aren't giving
+	 * correct values in the case of fine and coarse segments.
+	 * Restricting readouts only for super fine segment as of now.
+	 */
+
+	return blob;
+}
+
+static void icl_read_luts(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	if ((crtc_state->gamma_mode & POST_CSC_GAMMA_ENABLE) == 0)
+		return;
+
+	switch (crtc_state->gamma_mode & GAMMA_MODE_MODE_MASK) {
+	case GAMMA_MODE_MODE_8BIT:
+		crtc_state->hw.gamma_lut = ilk_read_lut_8(crtc);
+		break;
+	case GAMMA_MODE_MODE_10BIT:
+		crtc_state->hw.gamma_lut = bdw_read_lut_10(crtc, PAL_PREC_INDEX_VALUE(0));
+		break;
+	case GAMMA_MODE_MODE_12BIT_MULTI_SEGMENTED:
+		crtc_state->hw.gamma_lut = icl_read_lut_multi_segment(crtc);
+		break;
+	default:
+		MISSING_CASE(crtc_state->gamma_mode);
+		break;
+	}
+}
+
+static const struct intel_color_funcs chv_color_funcs = {
+	.color_check = chv_color_check,
+	.color_commit_arm = i9xx_color_commit_arm,
+	.load_luts = chv_load_luts,
+	.read_luts = chv_read_luts,
+};
+
+static const struct intel_color_funcs i965_color_funcs = {
+	.color_check = i9xx_color_check,
+	.color_commit_arm = i9xx_color_commit_arm,
+	.load_luts = i965_load_luts,
+	.read_luts = i965_read_luts,
+};
+
+static const struct intel_color_funcs i9xx_color_funcs = {
+	.color_check = i9xx_color_check,
+	.color_commit_arm = i9xx_color_commit_arm,
+	.load_luts = i9xx_load_luts,
+	.read_luts = i9xx_read_luts,
+};
+
+static const struct intel_color_funcs icl_color_funcs = {
+	.color_check = icl_color_check,
+	.color_commit_noarm = icl_color_commit_noarm,
+	.color_commit_arm = icl_color_commit_arm,
+	.load_luts = icl_load_luts,
+	.read_luts = icl_read_luts,
+};
+
+static const struct intel_color_funcs glk_color_funcs = {
+	.color_check = glk_color_check,
+	.color_commit_noarm = skl_color_commit_noarm,
+	.color_commit_arm = skl_color_commit_arm,
+	.load_luts = glk_load_luts,
+	.read_luts = glk_read_luts,
+};
+
+static const struct intel_color_funcs skl_color_funcs = {
+	.color_check = ivb_color_check,
+	.color_commit_noarm = skl_color_commit_noarm,
+	.color_commit_arm = skl_color_commit_arm,
+	.load_luts = bdw_load_luts,
+	.read_luts = NULL,
+};
+
+static const struct intel_color_funcs bdw_color_funcs = {
+	.color_check = ivb_color_check,
+	.color_commit_noarm = ilk_color_commit_noarm,
+	.color_commit_arm = hsw_color_commit_arm,
+	.load_luts = bdw_load_luts,
+	.read_luts = NULL,
+};
+
+static const struct intel_color_funcs hsw_color_funcs = {
+	.color_check = ivb_color_check,
+	.color_commit_noarm = ilk_color_commit_noarm,
+	.color_commit_arm = hsw_color_commit_arm,
+	.load_luts = ivb_load_luts,
+	.read_luts = NULL,
+};
+
+static const struct intel_color_funcs ivb_color_funcs = {
+	.color_check = ivb_color_check,
+	.color_commit_noarm = ilk_color_commit_noarm,
+	.color_commit_arm = ilk_color_commit_arm,
+	.load_luts = ivb_load_luts,
+	.read_luts = NULL,
+};
+
+static const struct intel_color_funcs ilk_color_funcs = {
+	.color_check = ilk_color_check,
+	.color_commit_noarm = ilk_color_commit_noarm,
+	.color_commit_arm = ilk_color_commit_arm,
+	.load_luts = ilk_load_luts,
+	.read_luts = ilk_read_luts,
+};
+
+void intel_color_init(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	bool has_ctm = INTEL_INFO(dev_priv)->display.color.degamma_lut_size != 0;
+
+	drm_mode_crtc_set_gamma_size(&crtc->base, 256);
+
+	if (HAS_GMCH(dev_priv)) {
+		if (IS_CHERRYVIEW(dev_priv)) {
+			dev_priv->display.funcs.color = &chv_color_funcs;
+		} else if (DISPLAY_VER(dev_priv) >= 4) {
+			dev_priv->display.funcs.color = &i965_color_funcs;
+		} else {
+			dev_priv->display.funcs.color = &i9xx_color_funcs;
+		}
+	} else {
+		if (DISPLAY_VER(dev_priv) >= 11)
+			dev_priv->display.funcs.color = &icl_color_funcs;
+		else if (DISPLAY_VER(dev_priv) == 10)
+			dev_priv->display.funcs.color = &glk_color_funcs;
+		else if (DISPLAY_VER(dev_priv) == 9)
+			dev_priv->display.funcs.color = &skl_color_funcs;
+		else if (DISPLAY_VER(dev_priv) == 8)
+			dev_priv->display.funcs.color = &bdw_color_funcs;
+		else if (DISPLAY_VER(dev_priv) == 7) {
+			if (IS_HASWELL(dev_priv))
+				dev_priv->display.funcs.color = &hsw_color_funcs;
+			else
+				dev_priv->display.funcs.color = &ivb_color_funcs;
+		} else
+			dev_priv->display.funcs.color = &ilk_color_funcs;
+	}
+
+	drm_crtc_enable_color_mgmt(&crtc->base,
+				   INTEL_INFO(dev_priv)->display.color.degamma_lut_size,
+				   has_ctm,
+				   INTEL_INFO(dev_priv)->display.color.gamma_lut_size);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_color.h b/drivers/gpu/drm/i915/display/intel_color.h
new file mode 100644
index 000000000..fd873425e
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_color.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_COLOR_H__
+#define __INTEL_COLOR_H__
+
+#include <linux/types.h>
+
+struct intel_crtc_state;
+struct intel_crtc;
+struct drm_property_blob;
+
+void intel_color_init(struct intel_crtc *crtc);
+int intel_color_check(struct intel_crtc_state *crtc_state);
+void intel_color_commit_noarm(const struct intel_crtc_state *crtc_state);
+void intel_color_commit_arm(const struct intel_crtc_state *crtc_state);
+void intel_color_load_luts(const struct intel_crtc_state *crtc_state);
+void intel_color_get_config(struct intel_crtc_state *crtc_state);
+int intel_color_get_gamma_bit_precision(const struct intel_crtc_state *crtc_state);
+bool intel_color_lut_equal(struct drm_property_blob *blob1,
+			   struct drm_property_blob *blob2,
+			   u32 gamma_mode, u32 bit_precision);
+
+#endif /* __INTEL_COLOR_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_combo_phy.c b/drivers/gpu/drm/i915/display/intel_combo_phy.c
new file mode 100644
index 000000000..64890f39c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_combo_phy.c
@@ -0,0 +1,439 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include "intel_combo_phy.h"
+#include "intel_combo_phy_regs.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+
+#define for_each_combo_phy(__dev_priv, __phy) \
+	for ((__phy) = PHY_A; (__phy) < I915_MAX_PHYS; (__phy)++)	\
+		for_each_if(intel_phy_is_combo(__dev_priv, __phy))
+
+#define for_each_combo_phy_reverse(__dev_priv, __phy) \
+	for ((__phy) = I915_MAX_PHYS; (__phy)-- > PHY_A;) \
+		for_each_if(intel_phy_is_combo(__dev_priv, __phy))
+
+enum {
+	PROCMON_0_85V_DOT_0,
+	PROCMON_0_95V_DOT_0,
+	PROCMON_0_95V_DOT_1,
+	PROCMON_1_05V_DOT_0,
+	PROCMON_1_05V_DOT_1,
+};
+
+static const struct icl_procmon {
+	const char *name;
+	u32 dw1, dw9, dw10;
+} icl_procmon_values[] = {
+	[PROCMON_0_85V_DOT_0] = {
+		.name = "0.85V dot0 (low-voltage)",
+		.dw1 = 0x00000000, .dw9 = 0x62AB67BB, .dw10 = 0x51914F96,
+	},
+	[PROCMON_0_95V_DOT_0] = {
+		.name = "0.95V dot0",
+		.dw1 = 0x00000000, .dw9 = 0x86E172C7, .dw10 = 0x77CA5EAB,
+	},
+	[PROCMON_0_95V_DOT_1] = {
+		.name = "0.95V dot1",
+		.dw1 = 0x00000000, .dw9 = 0x93F87FE1, .dw10 = 0x8AE871C5,
+	},
+	[PROCMON_1_05V_DOT_0] = {
+		.name = "1.05V dot0",
+		.dw1 = 0x00000000, .dw9 = 0x98FA82DD, .dw10 = 0x89E46DC1,
+	},
+	[PROCMON_1_05V_DOT_1] = {
+		.name = "1.05V dot1",
+		.dw1 = 0x00440000, .dw9 = 0x9A00AB25, .dw10 = 0x8AE38FF1,
+	},
+};
+
+static const struct icl_procmon *
+icl_get_procmon_ref_values(struct drm_i915_private *dev_priv, enum phy phy)
+{
+	const struct icl_procmon *procmon;
+	u32 val;
+
+	val = intel_de_read(dev_priv, ICL_PORT_COMP_DW3(phy));
+	switch (val & (PROCESS_INFO_MASK | VOLTAGE_INFO_MASK)) {
+	default:
+		MISSING_CASE(val);
+		fallthrough;
+	case VOLTAGE_INFO_0_85V | PROCESS_INFO_DOT_0:
+		procmon = &icl_procmon_values[PROCMON_0_85V_DOT_0];
+		break;
+	case VOLTAGE_INFO_0_95V | PROCESS_INFO_DOT_0:
+		procmon = &icl_procmon_values[PROCMON_0_95V_DOT_0];
+		break;
+	case VOLTAGE_INFO_0_95V | PROCESS_INFO_DOT_1:
+		procmon = &icl_procmon_values[PROCMON_0_95V_DOT_1];
+		break;
+	case VOLTAGE_INFO_1_05V | PROCESS_INFO_DOT_0:
+		procmon = &icl_procmon_values[PROCMON_1_05V_DOT_0];
+		break;
+	case VOLTAGE_INFO_1_05V | PROCESS_INFO_DOT_1:
+		procmon = &icl_procmon_values[PROCMON_1_05V_DOT_1];
+		break;
+	}
+
+	return procmon;
+}
+
+static void icl_set_procmon_ref_values(struct drm_i915_private *dev_priv,
+				       enum phy phy)
+{
+	const struct icl_procmon *procmon;
+	u32 val;
+
+	procmon = icl_get_procmon_ref_values(dev_priv, phy);
+
+	val = intel_de_read(dev_priv, ICL_PORT_COMP_DW1(phy));
+	val &= ~((0xff << 16) | 0xff);
+	val |= procmon->dw1;
+	intel_de_write(dev_priv, ICL_PORT_COMP_DW1(phy), val);
+
+	intel_de_write(dev_priv, ICL_PORT_COMP_DW9(phy), procmon->dw9);
+	intel_de_write(dev_priv, ICL_PORT_COMP_DW10(phy), procmon->dw10);
+}
+
+static bool check_phy_reg(struct drm_i915_private *dev_priv,
+			  enum phy phy, i915_reg_t reg, u32 mask,
+			  u32 expected_val)
+{
+	u32 val = intel_de_read(dev_priv, reg);
+
+	if ((val & mask) != expected_val) {
+		drm_dbg(&dev_priv->drm,
+			"Combo PHY %c reg %08x state mismatch: "
+			"current %08x mask %08x expected %08x\n",
+			phy_name(phy),
+			reg.reg, val, mask, expected_val);
+		return false;
+	}
+
+	return true;
+}
+
+static bool icl_verify_procmon_ref_values(struct drm_i915_private *dev_priv,
+					  enum phy phy)
+{
+	const struct icl_procmon *procmon;
+	bool ret;
+
+	procmon = icl_get_procmon_ref_values(dev_priv, phy);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Combo PHY %c Voltage/Process Info : %s\n",
+		    phy_name(phy), procmon->name);
+
+	ret = check_phy_reg(dev_priv, phy, ICL_PORT_COMP_DW1(phy),
+			    (0xff << 16) | 0xff, procmon->dw1);
+	ret &= check_phy_reg(dev_priv, phy, ICL_PORT_COMP_DW9(phy),
+			     -1U, procmon->dw9);
+	ret &= check_phy_reg(dev_priv, phy, ICL_PORT_COMP_DW10(phy),
+			     -1U, procmon->dw10);
+
+	return ret;
+}
+
+static bool has_phy_misc(struct drm_i915_private *i915, enum phy phy)
+{
+	/*
+	 * Some platforms only expect PHY_MISC to be programmed for PHY-A and
+	 * PHY-B and may not even have instances of the register for the
+	 * other combo PHY's.
+	 *
+	 * ADL-S technically has three instances of PHY_MISC, but only requires
+	 * that we program it for PHY A.
+	 */
+
+	if (IS_ALDERLAKE_S(i915))
+		return phy == PHY_A;
+	else if (IS_JSL_EHL(i915) ||
+		 IS_ROCKETLAKE(i915) ||
+		 IS_DG1(i915))
+		return phy < PHY_C;
+
+	return true;
+}
+
+static bool icl_combo_phy_enabled(struct drm_i915_private *dev_priv,
+				  enum phy phy)
+{
+	/* The PHY C added by EHL has no PHY_MISC register */
+	if (!has_phy_misc(dev_priv, phy))
+		return intel_de_read(dev_priv, ICL_PORT_COMP_DW0(phy)) & COMP_INIT;
+	else
+		return !(intel_de_read(dev_priv, ICL_PHY_MISC(phy)) &
+			 ICL_PHY_MISC_DE_IO_COMP_PWR_DOWN) &&
+			(intel_de_read(dev_priv, ICL_PORT_COMP_DW0(phy)) & COMP_INIT);
+}
+
+static bool ehl_vbt_ddi_d_present(struct drm_i915_private *i915)
+{
+	bool ddi_a_present = intel_bios_is_port_present(i915, PORT_A);
+	bool ddi_d_present = intel_bios_is_port_present(i915, PORT_D);
+	bool dsi_present = intel_bios_is_dsi_present(i915, NULL);
+
+	/*
+	 * VBT's 'dvo port' field for child devices references the DDI, not
+	 * the PHY.  So if combo PHY A is wired up to drive an external
+	 * display, we should see a child device present on PORT_D and
+	 * nothing on PORT_A and no DSI.
+	 */
+	if (ddi_d_present && !ddi_a_present && !dsi_present)
+		return true;
+
+	/*
+	 * If we encounter a VBT that claims to have an external display on
+	 * DDI-D _and_ an internal display on DDI-A/DSI leave an error message
+	 * in the log and let the internal display win.
+	 */
+	if (ddi_d_present)
+		drm_err(&i915->drm,
+			"VBT claims to have both internal and external displays on PHY A.  Configuring for internal.\n");
+
+	return false;
+}
+
+static bool phy_is_master(struct drm_i915_private *dev_priv, enum phy phy)
+{
+	/*
+	 * Certain PHYs are connected to compensation resistors and act
+	 * as masters to other PHYs.
+	 *
+	 * ICL,TGL:
+	 *   A(master) -> B(slave), C(slave)
+	 * RKL,DG1:
+	 *   A(master) -> B(slave)
+	 *   C(master) -> D(slave)
+	 * ADL-S:
+	 *   A(master) -> B(slave), C(slave)
+	 *   D(master) -> E(slave)
+	 *
+	 * We must set the IREFGEN bit for any PHY acting as a master
+	 * to another PHY.
+	 */
+	if (phy == PHY_A)
+		return true;
+	else if (IS_ALDERLAKE_S(dev_priv))
+		return phy == PHY_D;
+	else if (IS_DG1(dev_priv) || IS_ROCKETLAKE(dev_priv))
+		return phy == PHY_C;
+
+	return false;
+}
+
+static bool icl_combo_phy_verify_state(struct drm_i915_private *dev_priv,
+				       enum phy phy)
+{
+	bool ret = true;
+	u32 expected_val = 0;
+
+	if (!icl_combo_phy_enabled(dev_priv, phy))
+		return false;
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		ret &= check_phy_reg(dev_priv, phy, ICL_PORT_TX_DW8_LN(0, phy),
+				     ICL_PORT_TX_DW8_ODCC_CLK_SEL |
+				     ICL_PORT_TX_DW8_ODCC_CLK_DIV_SEL_MASK,
+				     ICL_PORT_TX_DW8_ODCC_CLK_SEL |
+				     ICL_PORT_TX_DW8_ODCC_CLK_DIV_SEL_DIV2);
+
+		ret &= check_phy_reg(dev_priv, phy, ICL_PORT_PCS_DW1_LN(0, phy),
+				     DCC_MODE_SELECT_MASK,
+				     DCC_MODE_SELECT_CONTINUOSLY);
+	}
+
+	ret &= icl_verify_procmon_ref_values(dev_priv, phy);
+
+	if (phy_is_master(dev_priv, phy)) {
+		ret &= check_phy_reg(dev_priv, phy, ICL_PORT_COMP_DW8(phy),
+				     IREFGEN, IREFGEN);
+
+		if (IS_JSL_EHL(dev_priv)) {
+			if (ehl_vbt_ddi_d_present(dev_priv))
+				expected_val = ICL_PHY_MISC_MUX_DDID;
+
+			ret &= check_phy_reg(dev_priv, phy, ICL_PHY_MISC(phy),
+					     ICL_PHY_MISC_MUX_DDID,
+					     expected_val);
+		}
+	}
+
+	ret &= check_phy_reg(dev_priv, phy, ICL_PORT_CL_DW5(phy),
+			     CL_POWER_DOWN_ENABLE, CL_POWER_DOWN_ENABLE);
+
+	return ret;
+}
+
+void intel_combo_phy_power_up_lanes(struct drm_i915_private *dev_priv,
+				    enum phy phy, bool is_dsi,
+				    int lane_count, bool lane_reversal)
+{
+	u8 lane_mask;
+	u32 val;
+
+	if (is_dsi) {
+		drm_WARN_ON(&dev_priv->drm, lane_reversal);
+
+		switch (lane_count) {
+		case 1:
+			lane_mask = PWR_DOWN_LN_3_1_0;
+			break;
+		case 2:
+			lane_mask = PWR_DOWN_LN_3_1;
+			break;
+		case 3:
+			lane_mask = PWR_DOWN_LN_3;
+			break;
+		default:
+			MISSING_CASE(lane_count);
+			fallthrough;
+		case 4:
+			lane_mask = PWR_UP_ALL_LANES;
+			break;
+		}
+	} else {
+		switch (lane_count) {
+		case 1:
+			lane_mask = lane_reversal ? PWR_DOWN_LN_2_1_0 :
+						    PWR_DOWN_LN_3_2_1;
+			break;
+		case 2:
+			lane_mask = lane_reversal ? PWR_DOWN_LN_1_0 :
+						    PWR_DOWN_LN_3_2;
+			break;
+		default:
+			MISSING_CASE(lane_count);
+			fallthrough;
+		case 4:
+			lane_mask = PWR_UP_ALL_LANES;
+			break;
+		}
+	}
+
+	val = intel_de_read(dev_priv, ICL_PORT_CL_DW10(phy));
+	val &= ~PWR_DOWN_LN_MASK;
+	val |= lane_mask;
+	intel_de_write(dev_priv, ICL_PORT_CL_DW10(phy), val);
+}
+
+static void icl_combo_phys_init(struct drm_i915_private *dev_priv)
+{
+	enum phy phy;
+
+	for_each_combo_phy(dev_priv, phy) {
+		u32 val;
+
+		if (icl_combo_phy_verify_state(dev_priv, phy)) {
+			drm_dbg(&dev_priv->drm,
+				"Combo PHY %c already enabled, won't reprogram it.\n",
+				phy_name(phy));
+			continue;
+		}
+
+		if (!has_phy_misc(dev_priv, phy))
+			goto skip_phy_misc;
+
+		/*
+		 * EHL's combo PHY A can be hooked up to either an external
+		 * display (via DDI-D) or an internal display (via DDI-A or
+		 * the DSI DPHY).  This is a motherboard design decision that
+		 * can't be changed on the fly, so initialize the PHY's mux
+		 * based on whether our VBT indicates the presence of any
+		 * "internal" child devices.
+		 */
+		val = intel_de_read(dev_priv, ICL_PHY_MISC(phy));
+		if (IS_JSL_EHL(dev_priv) && phy == PHY_A) {
+			val &= ~ICL_PHY_MISC_MUX_DDID;
+
+			if (ehl_vbt_ddi_d_present(dev_priv))
+				val |= ICL_PHY_MISC_MUX_DDID;
+		}
+
+		val &= ~ICL_PHY_MISC_DE_IO_COMP_PWR_DOWN;
+		intel_de_write(dev_priv, ICL_PHY_MISC(phy), val);
+
+skip_phy_misc:
+		if (DISPLAY_VER(dev_priv) >= 12) {
+			val = intel_de_read(dev_priv, ICL_PORT_TX_DW8_LN(0, phy));
+			val &= ~ICL_PORT_TX_DW8_ODCC_CLK_DIV_SEL_MASK;
+			val |= ICL_PORT_TX_DW8_ODCC_CLK_SEL;
+			val |= ICL_PORT_TX_DW8_ODCC_CLK_DIV_SEL_DIV2;
+			intel_de_write(dev_priv, ICL_PORT_TX_DW8_GRP(phy), val);
+
+			val = intel_de_read(dev_priv, ICL_PORT_PCS_DW1_LN(0, phy));
+			val &= ~DCC_MODE_SELECT_MASK;
+			val |= DCC_MODE_SELECT_CONTINUOSLY;
+			intel_de_write(dev_priv, ICL_PORT_PCS_DW1_GRP(phy), val);
+		}
+
+		icl_set_procmon_ref_values(dev_priv, phy);
+
+		if (phy_is_master(dev_priv, phy)) {
+			val = intel_de_read(dev_priv, ICL_PORT_COMP_DW8(phy));
+			val |= IREFGEN;
+			intel_de_write(dev_priv, ICL_PORT_COMP_DW8(phy), val);
+		}
+
+		val = intel_de_read(dev_priv, ICL_PORT_COMP_DW0(phy));
+		val |= COMP_INIT;
+		intel_de_write(dev_priv, ICL_PORT_COMP_DW0(phy), val);
+
+		val = intel_de_read(dev_priv, ICL_PORT_CL_DW5(phy));
+		val |= CL_POWER_DOWN_ENABLE;
+		intel_de_write(dev_priv, ICL_PORT_CL_DW5(phy), val);
+	}
+}
+
+static void icl_combo_phys_uninit(struct drm_i915_private *dev_priv)
+{
+	enum phy phy;
+
+	for_each_combo_phy_reverse(dev_priv, phy) {
+		u32 val;
+
+		if (phy == PHY_A &&
+		    !icl_combo_phy_verify_state(dev_priv, phy)) {
+			if (IS_TIGERLAKE(dev_priv) || IS_DG1(dev_priv)) {
+				/*
+				 * A known problem with old ifwi:
+				 * https://gitlab.freedesktop.org/drm/intel/-/issues/2411
+				 * Suppress the warning for CI. Remove ASAP!
+				 */
+				drm_dbg_kms(&dev_priv->drm,
+					    "Combo PHY %c HW state changed unexpectedly\n",
+					    phy_name(phy));
+			} else {
+				drm_warn(&dev_priv->drm,
+					 "Combo PHY %c HW state changed unexpectedly\n",
+					 phy_name(phy));
+			}
+		}
+
+		if (!has_phy_misc(dev_priv, phy))
+			goto skip_phy_misc;
+
+		val = intel_de_read(dev_priv, ICL_PHY_MISC(phy));
+		val |= ICL_PHY_MISC_DE_IO_COMP_PWR_DOWN;
+		intel_de_write(dev_priv, ICL_PHY_MISC(phy), val);
+
+skip_phy_misc:
+		val = intel_de_read(dev_priv, ICL_PORT_COMP_DW0(phy));
+		val &= ~COMP_INIT;
+		intel_de_write(dev_priv, ICL_PORT_COMP_DW0(phy), val);
+	}
+}
+
+void intel_combo_phy_init(struct drm_i915_private *i915)
+{
+	icl_combo_phys_init(i915);
+}
+
+void intel_combo_phy_uninit(struct drm_i915_private *i915)
+{
+	icl_combo_phys_uninit(i915);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_combo_phy.h b/drivers/gpu/drm/i915/display/intel_combo_phy.h
new file mode 100644
index 000000000..660886f86
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_combo_phy.h
@@ -0,0 +1,20 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_COMBO_PHY_H__
+#define __INTEL_COMBO_PHY_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+enum phy;
+
+void intel_combo_phy_init(struct drm_i915_private *dev_priv);
+void intel_combo_phy_uninit(struct drm_i915_private *dev_priv);
+void intel_combo_phy_power_up_lanes(struct drm_i915_private *dev_priv,
+				    enum phy phy, bool is_dsi,
+				    int lane_count, bool lane_reversal);
+
+#endif /* __INTEL_COMBO_PHY_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_combo_phy_regs.h b/drivers/gpu/drm/i915/display/intel_combo_phy_regs.h
new file mode 100644
index 000000000..2ed65193c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_combo_phy_regs.h
@@ -0,0 +1,162 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_COMBO_PHY_REGS__
+#define __INTEL_COMBO_PHY_REGS__
+
+#include "i915_reg_defs.h"
+
+#define _ICL_COMBOPHY_A				0x162000
+#define _ICL_COMBOPHY_B				0x6C000
+#define _EHL_COMBOPHY_C				0x160000
+#define _RKL_COMBOPHY_D				0x161000
+#define _ADL_COMBOPHY_E				0x16B000
+
+#define _ICL_COMBOPHY(phy)			_PICK(phy, _ICL_COMBOPHY_A, \
+						      _ICL_COMBOPHY_B, \
+						      _EHL_COMBOPHY_C, \
+						      _RKL_COMBOPHY_D, \
+						      _ADL_COMBOPHY_E)
+
+/* ICL Port CL_DW registers */
+#define _ICL_PORT_CL_DW(dw, phy)		(_ICL_COMBOPHY(phy) + \
+						 4 * (dw))
+
+#define ICL_PORT_CL_DW5(phy)			_MMIO(_ICL_PORT_CL_DW(5, phy))
+#define   CL_POWER_DOWN_ENABLE			(1 << 4)
+#define   SUS_CLOCK_CONFIG			(3 << 0)
+
+#define ICL_PORT_CL_DW10(phy)			_MMIO(_ICL_PORT_CL_DW(10, phy))
+#define  PG_SEQ_DELAY_OVERRIDE_MASK		(3 << 25)
+#define  PG_SEQ_DELAY_OVERRIDE_SHIFT		25
+#define  PG_SEQ_DELAY_OVERRIDE_ENABLE		(1 << 24)
+#define  PWR_UP_ALL_LANES			(0x0 << 4)
+#define  PWR_DOWN_LN_3_2_1			(0xe << 4)
+#define  PWR_DOWN_LN_3_2			(0xc << 4)
+#define  PWR_DOWN_LN_3				(0x8 << 4)
+#define  PWR_DOWN_LN_2_1_0			(0x7 << 4)
+#define  PWR_DOWN_LN_1_0			(0x3 << 4)
+#define  PWR_DOWN_LN_3_1			(0xa << 4)
+#define  PWR_DOWN_LN_3_1_0			(0xb << 4)
+#define  PWR_DOWN_LN_MASK			(0xf << 4)
+#define  PWR_DOWN_LN_SHIFT			4
+#define  EDP4K2K_MODE_OVRD_EN			(1 << 3)
+#define  EDP4K2K_MODE_OVRD_OPTIMIZED		(1 << 2)
+
+#define ICL_PORT_CL_DW12(phy)			_MMIO(_ICL_PORT_CL_DW(12, phy))
+#define   ICL_LANE_ENABLE_AUX			(1 << 0)
+
+/* ICL Port COMP_DW registers */
+#define _ICL_PORT_COMP				0x100
+#define _ICL_PORT_COMP_DW(dw, phy)		(_ICL_COMBOPHY(phy) + \
+						 _ICL_PORT_COMP + 4 * (dw))
+
+#define ICL_PORT_COMP_DW0(phy)			_MMIO(_ICL_PORT_COMP_DW(0, phy))
+#define   COMP_INIT				(1 << 31)
+
+#define ICL_PORT_COMP_DW1(phy)			_MMIO(_ICL_PORT_COMP_DW(1, phy))
+
+#define ICL_PORT_COMP_DW3(phy)			_MMIO(_ICL_PORT_COMP_DW(3, phy))
+#define   PROCESS_INFO_DOT_0			(0 << 26)
+#define   PROCESS_INFO_DOT_1			(1 << 26)
+#define   PROCESS_INFO_DOT_4			(2 << 26)
+#define   PROCESS_INFO_MASK			(7 << 26)
+#define   PROCESS_INFO_SHIFT			26
+#define   VOLTAGE_INFO_0_85V			(0 << 24)
+#define   VOLTAGE_INFO_0_95V			(1 << 24)
+#define   VOLTAGE_INFO_1_05V			(2 << 24)
+#define   VOLTAGE_INFO_MASK			(3 << 24)
+#define   VOLTAGE_INFO_SHIFT			24
+
+#define ICL_PORT_COMP_DW8(phy)			_MMIO(_ICL_PORT_COMP_DW(8, phy))
+#define   IREFGEN				(1 << 24)
+
+#define ICL_PORT_COMP_DW9(phy)			_MMIO(_ICL_PORT_COMP_DW(9, phy))
+
+#define ICL_PORT_COMP_DW10(phy)			_MMIO(_ICL_PORT_COMP_DW(10, phy))
+
+/* ICL Port PCS registers */
+#define _ICL_PORT_PCS_AUX			0x300
+#define _ICL_PORT_PCS_GRP			0x600
+#define _ICL_PORT_PCS_LN(ln)			(0x800 + (ln) * 0x100)
+#define _ICL_PORT_PCS_DW_AUX(dw, phy)		(_ICL_COMBOPHY(phy) + \
+						 _ICL_PORT_PCS_AUX + 4 * (dw))
+#define _ICL_PORT_PCS_DW_GRP(dw, phy)		(_ICL_COMBOPHY(phy) + \
+						 _ICL_PORT_PCS_GRP + 4 * (dw))
+#define _ICL_PORT_PCS_DW_LN(dw, ln, phy)	 (_ICL_COMBOPHY(phy) + \
+						  _ICL_PORT_PCS_LN(ln) + 4 * (dw))
+#define ICL_PORT_PCS_DW1_AUX(phy)		_MMIO(_ICL_PORT_PCS_DW_AUX(1, phy))
+#define ICL_PORT_PCS_DW1_GRP(phy)		_MMIO(_ICL_PORT_PCS_DW_GRP(1, phy))
+#define ICL_PORT_PCS_DW1_LN(ln, phy)		_MMIO(_ICL_PORT_PCS_DW_LN(1, ln, phy))
+#define   DCC_MODE_SELECT_MASK			(0x3 << 20)
+#define   DCC_MODE_SELECT_CONTINUOSLY		(0x3 << 20)
+#define   COMMON_KEEPER_EN			(1 << 26)
+#define   LATENCY_OPTIM_MASK			(0x3 << 2)
+#define   LATENCY_OPTIM_VAL(x)			((x) << 2)
+
+/* ICL Port TX registers */
+#define _ICL_PORT_TX_AUX			0x380
+#define _ICL_PORT_TX_GRP			0x680
+#define _ICL_PORT_TX_LN(ln)			(0x880 + (ln) * 0x100)
+
+#define _ICL_PORT_TX_DW_AUX(dw, phy)		(_ICL_COMBOPHY(phy) + \
+						 _ICL_PORT_TX_AUX + 4 * (dw))
+#define _ICL_PORT_TX_DW_GRP(dw, phy)		(_ICL_COMBOPHY(phy) + \
+						 _ICL_PORT_TX_GRP + 4 * (dw))
+#define _ICL_PORT_TX_DW_LN(dw, ln, phy) 	(_ICL_COMBOPHY(phy) + \
+						  _ICL_PORT_TX_LN(ln) + 4 * (dw))
+
+#define ICL_PORT_TX_DW2_AUX(phy)		_MMIO(_ICL_PORT_TX_DW_AUX(2, phy))
+#define ICL_PORT_TX_DW2_GRP(phy)		_MMIO(_ICL_PORT_TX_DW_GRP(2, phy))
+#define ICL_PORT_TX_DW2_LN(ln, phy)		_MMIO(_ICL_PORT_TX_DW_LN(2, ln, phy))
+#define   SWING_SEL_UPPER(x)			(((x) >> 3) << 15)
+#define   SWING_SEL_UPPER_MASK			(1 << 15)
+#define   SWING_SEL_LOWER(x)			(((x) & 0x7) << 11)
+#define   SWING_SEL_LOWER_MASK			(0x7 << 11)
+#define   FRC_LATENCY_OPTIM_MASK		(0x7 << 8)
+#define   FRC_LATENCY_OPTIM_VAL(x)		((x) << 8)
+#define   RCOMP_SCALAR(x)			((x) << 0)
+#define   RCOMP_SCALAR_MASK			(0xFF << 0)
+
+#define ICL_PORT_TX_DW4_AUX(phy)		_MMIO(_ICL_PORT_TX_DW_AUX(4, phy))
+#define ICL_PORT_TX_DW4_GRP(phy)		_MMIO(_ICL_PORT_TX_DW_GRP(4, phy))
+#define ICL_PORT_TX_DW4_LN(ln, phy)		_MMIO(_ICL_PORT_TX_DW_LN(4, ln, phy))
+#define   LOADGEN_SELECT			(1 << 31)
+#define   POST_CURSOR_1(x)			((x) << 12)
+#define   POST_CURSOR_1_MASK			(0x3F << 12)
+#define   POST_CURSOR_2(x)			((x) << 6)
+#define   POST_CURSOR_2_MASK			(0x3F << 6)
+#define   CURSOR_COEFF(x)			((x) << 0)
+#define   CURSOR_COEFF_MASK			(0x3F << 0)
+
+#define ICL_PORT_TX_DW5_AUX(phy)		_MMIO(_ICL_PORT_TX_DW_AUX(5, phy))
+#define ICL_PORT_TX_DW5_GRP(phy)		_MMIO(_ICL_PORT_TX_DW_GRP(5, phy))
+#define ICL_PORT_TX_DW5_LN(ln, phy)		_MMIO(_ICL_PORT_TX_DW_LN(5, ln, phy))
+#define   TX_TRAINING_EN			(1 << 31)
+#define   TAP2_DISABLE				(1 << 30)
+#define   TAP3_DISABLE				(1 << 29)
+#define   SCALING_MODE_SEL(x)			((x) << 18)
+#define   SCALING_MODE_SEL_MASK			(0x7 << 18)
+#define   RTERM_SELECT(x)			((x) << 3)
+#define   RTERM_SELECT_MASK			(0x7 << 3)
+
+#define ICL_PORT_TX_DW7_AUX(phy)		_MMIO(_ICL_PORT_TX_DW_AUX(7, phy))
+#define ICL_PORT_TX_DW7_GRP(phy)		_MMIO(_ICL_PORT_TX_DW_GRP(7, phy))
+#define ICL_PORT_TX_DW7_LN(ln, phy)		_MMIO(_ICL_PORT_TX_DW_LN(7, ln, phy))
+#define   N_SCALAR(x)				((x) << 24)
+#define   N_SCALAR_MASK				(0x7F << 24)
+
+#define ICL_PORT_TX_DW8_AUX(phy)		_MMIO(_ICL_PORT_TX_DW_AUX(8, phy))
+#define ICL_PORT_TX_DW8_GRP(phy)		_MMIO(_ICL_PORT_TX_DW_GRP(8, phy))
+#define ICL_PORT_TX_DW8_LN(ln, phy)		_MMIO(_ICL_PORT_TX_DW_LN(8, ln, phy))
+#define   ICL_PORT_TX_DW8_ODCC_CLK_SEL		REG_BIT(31)
+#define   ICL_PORT_TX_DW8_ODCC_CLK_DIV_SEL_MASK	REG_GENMASK(30, 29)
+#define   ICL_PORT_TX_DW8_ODCC_CLK_DIV_SEL_DIV2	REG_FIELD_PREP(ICL_PORT_TX_DW8_ODCC_CLK_DIV_SEL_MASK, 0x1)
+
+#define _ICL_DPHY_CHKN_REG			0x194
+#define ICL_DPHY_CHKN(port)			_MMIO(_ICL_COMBOPHY(port) + _ICL_DPHY_CHKN_REG)
+#define   ICL_DPHY_CHKN_AFE_OVER_PPI_STRAP	REG_BIT(7)
+
+#endif /* __INTEL_COMBO_PHY_REGS__ */
diff --git a/drivers/gpu/drm/i915/display/intel_connector.c b/drivers/gpu/drm/i915/display/intel_connector.c
new file mode 100644
index 000000000..8bb296f3d
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_connector.c
@@ -0,0 +1,295 @@
+/*
+ * Copyright (c) 2007 Dave Airlie <airlied@linux.ie>
+ * Copyright (c) 2007, 2010 Intel Corporation
+ *   Jesse Barnes <jesse.barnes@intel.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/i2c.h>
+#include <linux/slab.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_edid.h>
+
+#include "i915_drv.h"
+#include "intel_backlight.h"
+#include "intel_connector.h"
+#include "intel_display_debugfs.h"
+#include "intel_display_types.h"
+#include "intel_hdcp.h"
+#include "intel_panel.h"
+
+int intel_connector_init(struct intel_connector *connector)
+{
+	struct intel_digital_connector_state *conn_state;
+
+	/*
+	 * Allocate enough memory to hold intel_digital_connector_state,
+	 * This might be a few bytes too many, but for connectors that don't
+	 * need it we'll free the state and allocate a smaller one on the first
+	 * successful commit anyway.
+	 */
+	conn_state = kzalloc(sizeof(*conn_state), GFP_KERNEL);
+	if (!conn_state)
+		return -ENOMEM;
+
+	__drm_atomic_helper_connector_reset(&connector->base,
+					    &conn_state->base);
+
+	intel_panel_init_alloc(connector);
+
+	return 0;
+}
+
+struct intel_connector *intel_connector_alloc(void)
+{
+	struct intel_connector *connector;
+
+	connector = kzalloc(sizeof(*connector), GFP_KERNEL);
+	if (!connector)
+		return NULL;
+
+	if (intel_connector_init(connector) < 0) {
+		kfree(connector);
+		return NULL;
+	}
+
+	return connector;
+}
+
+/*
+ * Free the bits allocated by intel_connector_alloc.
+ * This should only be used after intel_connector_alloc has returned
+ * successfully, and before drm_connector_init returns successfully.
+ * Otherwise the destroy callbacks for the connector and the state should
+ * take care of proper cleanup/free (see intel_connector_destroy).
+ */
+void intel_connector_free(struct intel_connector *connector)
+{
+	kfree(to_intel_digital_connector_state(connector->base.state));
+	kfree(connector);
+}
+
+/*
+ * Connector type independent destroy hook for drm_connector_funcs.
+ */
+void intel_connector_destroy(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+
+	kfree(intel_connector->detect_edid);
+
+	intel_hdcp_cleanup(intel_connector);
+
+	if (!IS_ERR_OR_NULL(intel_connector->edid))
+		kfree(intel_connector->edid);
+
+	intel_panel_fini(intel_connector);
+
+	drm_connector_cleanup(connector);
+
+	if (intel_connector->port)
+		drm_dp_mst_put_port_malloc(intel_connector->port);
+
+	kfree(connector);
+}
+
+int intel_connector_register(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	int ret;
+
+	ret = intel_backlight_device_register(intel_connector);
+	if (ret)
+		goto err;
+
+	if (i915_inject_probe_failure(to_i915(connector->dev))) {
+		ret = -EFAULT;
+		goto err_backlight;
+	}
+
+	intel_connector_debugfs_add(intel_connector);
+
+	return 0;
+
+err_backlight:
+	intel_backlight_device_unregister(intel_connector);
+err:
+	return ret;
+}
+
+void intel_connector_unregister(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+
+	intel_backlight_device_unregister(intel_connector);
+}
+
+void intel_connector_attach_encoder(struct intel_connector *connector,
+				    struct intel_encoder *encoder)
+{
+	connector->encoder = encoder;
+	drm_connector_attach_encoder(&connector->base, &encoder->base);
+}
+
+/*
+ * Simple connector->get_hw_state implementation for encoders that support only
+ * one connector and no cloning and hence the encoder state determines the state
+ * of the connector.
+ */
+bool intel_connector_get_hw_state(struct intel_connector *connector)
+{
+	enum pipe pipe = 0;
+	struct intel_encoder *encoder = intel_attached_encoder(connector);
+
+	return encoder->get_hw_state(encoder, &pipe);
+}
+
+enum pipe intel_connector_get_pipe(struct intel_connector *connector)
+{
+	struct drm_device *dev = connector->base.dev;
+
+	drm_WARN_ON(dev,
+		    !drm_modeset_is_locked(&dev->mode_config.connection_mutex));
+
+	if (!connector->base.state->crtc)
+		return INVALID_PIPE;
+
+	return to_intel_crtc(connector->base.state->crtc)->pipe;
+}
+
+/**
+ * intel_connector_update_modes - update connector from edid
+ * @connector: DRM connector device to use
+ * @edid: previously read EDID information
+ */
+int intel_connector_update_modes(struct drm_connector *connector,
+				struct edid *edid)
+{
+	int ret;
+
+	drm_connector_update_edid_property(connector, edid);
+	ret = drm_add_edid_modes(connector, edid);
+
+	return ret;
+}
+
+/**
+ * intel_ddc_get_modes - get modelist from monitor
+ * @connector: DRM connector device to use
+ * @adapter: i2c adapter
+ *
+ * Fetch the EDID information from @connector using the DDC bus.
+ */
+int intel_ddc_get_modes(struct drm_connector *connector,
+			struct i2c_adapter *adapter)
+{
+	struct edid *edid;
+	int ret;
+
+	edid = drm_get_edid(connector, adapter);
+	if (!edid)
+		return 0;
+
+	ret = intel_connector_update_modes(connector, edid);
+	kfree(edid);
+
+	return ret;
+}
+
+static const struct drm_prop_enum_list force_audio_names[] = {
+	{ HDMI_AUDIO_OFF_DVI, "force-dvi" },
+	{ HDMI_AUDIO_OFF, "off" },
+	{ HDMI_AUDIO_AUTO, "auto" },
+	{ HDMI_AUDIO_ON, "on" },
+};
+
+void
+intel_attach_force_audio_property(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_property *prop;
+
+	prop = dev_priv->display.properties.force_audio;
+	if (prop == NULL) {
+		prop = drm_property_create_enum(dev, 0,
+					   "audio",
+					   force_audio_names,
+					   ARRAY_SIZE(force_audio_names));
+		if (prop == NULL)
+			return;
+
+		dev_priv->display.properties.force_audio = prop;
+	}
+	drm_object_attach_property(&connector->base, prop, 0);
+}
+
+static const struct drm_prop_enum_list broadcast_rgb_names[] = {
+	{ INTEL_BROADCAST_RGB_AUTO, "Automatic" },
+	{ INTEL_BROADCAST_RGB_FULL, "Full" },
+	{ INTEL_BROADCAST_RGB_LIMITED, "Limited 16:235" },
+};
+
+void
+intel_attach_broadcast_rgb_property(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_property *prop;
+
+	prop = dev_priv->display.properties.broadcast_rgb;
+	if (prop == NULL) {
+		prop = drm_property_create_enum(dev, DRM_MODE_PROP_ENUM,
+					   "Broadcast RGB",
+					   broadcast_rgb_names,
+					   ARRAY_SIZE(broadcast_rgb_names));
+		if (prop == NULL)
+			return;
+
+		dev_priv->display.properties.broadcast_rgb = prop;
+	}
+
+	drm_object_attach_property(&connector->base, prop, 0);
+}
+
+void
+intel_attach_aspect_ratio_property(struct drm_connector *connector)
+{
+	if (!drm_mode_create_aspect_ratio_property(connector->dev))
+		drm_object_attach_property(&connector->base,
+			connector->dev->mode_config.aspect_ratio_property,
+			DRM_MODE_PICTURE_ASPECT_NONE);
+}
+
+void
+intel_attach_hdmi_colorspace_property(struct drm_connector *connector)
+{
+	if (!drm_mode_create_hdmi_colorspace_property(connector))
+		drm_connector_attach_colorspace_property(connector);
+}
+
+void
+intel_attach_dp_colorspace_property(struct drm_connector *connector)
+{
+	if (!drm_mode_create_dp_colorspace_property(connector))
+		drm_connector_attach_colorspace_property(connector);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_connector.h b/drivers/gpu/drm/i915/display/intel_connector.h
new file mode 100644
index 000000000..661a37a3c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_connector.h
@@ -0,0 +1,36 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_CONNECTOR_H__
+#define __INTEL_CONNECTOR_H__
+
+#include "intel_display.h"
+
+struct drm_connector;
+struct edid;
+struct i2c_adapter;
+struct intel_connector;
+struct intel_encoder;
+
+int intel_connector_init(struct intel_connector *connector);
+struct intel_connector *intel_connector_alloc(void);
+void intel_connector_free(struct intel_connector *connector);
+void intel_connector_destroy(struct drm_connector *connector);
+int intel_connector_register(struct drm_connector *connector);
+void intel_connector_unregister(struct drm_connector *connector);
+void intel_connector_attach_encoder(struct intel_connector *connector,
+				    struct intel_encoder *encoder);
+bool intel_connector_get_hw_state(struct intel_connector *connector);
+enum pipe intel_connector_get_pipe(struct intel_connector *connector);
+int intel_connector_update_modes(struct drm_connector *connector,
+				 struct edid *edid);
+int intel_ddc_get_modes(struct drm_connector *c, struct i2c_adapter *adapter);
+void intel_attach_force_audio_property(struct drm_connector *connector);
+void intel_attach_broadcast_rgb_property(struct drm_connector *connector);
+void intel_attach_aspect_ratio_property(struct drm_connector *connector);
+void intel_attach_hdmi_colorspace_property(struct drm_connector *connector);
+void intel_attach_dp_colorspace_property(struct drm_connector *connector);
+
+#endif /* __INTEL_CONNECTOR_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_crt.c b/drivers/gpu/drm/i915/display/intel_crt.c
new file mode 100644
index 000000000..e60b2cf84
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_crt.c
@@ -0,0 +1,1124 @@
+/*
+ * Copyright © 2006-2007 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Eric Anholt <eric@anholt.net>
+ */
+
+#include <linux/dmi.h>
+#include <linux/i2c.h>
+#include <linux/slab.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_probe_helper.h>
+
+#include "i915_drv.h"
+#include "intel_connector.h"
+#include "intel_crt.h"
+#include "intel_crtc.h"
+#include "intel_ddi.h"
+#include "intel_ddi_buf_trans.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_fdi.h"
+#include "intel_fifo_underrun.h"
+#include "intel_gmbus.h"
+#include "intel_hotplug.h"
+#include "intel_pch_display.h"
+#include "intel_pch_refclk.h"
+
+/* Here's the desired hotplug mode */
+#define ADPA_HOTPLUG_BITS (ADPA_CRT_HOTPLUG_PERIOD_128 |		\
+			   ADPA_CRT_HOTPLUG_WARMUP_10MS |		\
+			   ADPA_CRT_HOTPLUG_SAMPLE_4S |			\
+			   ADPA_CRT_HOTPLUG_VOLTAGE_50 |		\
+			   ADPA_CRT_HOTPLUG_VOLREF_325MV |		\
+			   ADPA_CRT_HOTPLUG_ENABLE)
+
+struct intel_crt {
+	struct intel_encoder base;
+	/* DPMS state is stored in the connector, which we need in the
+	 * encoder's enable/disable callbacks */
+	struct intel_connector *connector;
+	bool force_hotplug_required;
+	i915_reg_t adpa_reg;
+};
+
+static struct intel_crt *intel_encoder_to_crt(struct intel_encoder *encoder)
+{
+	return container_of(encoder, struct intel_crt, base);
+}
+
+static struct intel_crt *intel_attached_crt(struct intel_connector *connector)
+{
+	return intel_encoder_to_crt(intel_attached_encoder(connector));
+}
+
+bool intel_crt_port_enabled(struct drm_i915_private *dev_priv,
+			    i915_reg_t adpa_reg, enum pipe *pipe)
+{
+	u32 val;
+
+	val = intel_de_read(dev_priv, adpa_reg);
+
+	/* asserts want to know the pipe even if the port is disabled */
+	if (HAS_PCH_CPT(dev_priv))
+		*pipe = (val & ADPA_PIPE_SEL_MASK_CPT) >> ADPA_PIPE_SEL_SHIFT_CPT;
+	else
+		*pipe = (val & ADPA_PIPE_SEL_MASK) >> ADPA_PIPE_SEL_SHIFT;
+
+	return val & ADPA_DAC_ENABLE;
+}
+
+static bool intel_crt_get_hw_state(struct intel_encoder *encoder,
+				   enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crt *crt = intel_encoder_to_crt(encoder);
+	intel_wakeref_t wakeref;
+	bool ret;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     encoder->power_domain);
+	if (!wakeref)
+		return false;
+
+	ret = intel_crt_port_enabled(dev_priv, crt->adpa_reg, pipe);
+
+	intel_display_power_put(dev_priv, encoder->power_domain, wakeref);
+
+	return ret;
+}
+
+static unsigned int intel_crt_get_flags(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crt *crt = intel_encoder_to_crt(encoder);
+	u32 tmp, flags = 0;
+
+	tmp = intel_de_read(dev_priv, crt->adpa_reg);
+
+	if (tmp & ADPA_HSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+
+	if (tmp & ADPA_VSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	return flags;
+}
+
+static void intel_crt_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_state *pipe_config)
+{
+	pipe_config->output_types |= BIT(INTEL_OUTPUT_ANALOG);
+
+	pipe_config->hw.adjusted_mode.flags |= intel_crt_get_flags(encoder);
+
+	pipe_config->hw.adjusted_mode.crtc_clock = pipe_config->port_clock;
+}
+
+static void hsw_crt_get_config(struct intel_encoder *encoder,
+			       struct intel_crtc_state *pipe_config)
+{
+	lpt_pch_get_config(pipe_config);
+
+	hsw_ddi_get_config(encoder, pipe_config);
+
+	pipe_config->hw.adjusted_mode.flags &= ~(DRM_MODE_FLAG_PHSYNC |
+					      DRM_MODE_FLAG_NHSYNC |
+					      DRM_MODE_FLAG_PVSYNC |
+					      DRM_MODE_FLAG_NVSYNC);
+	pipe_config->hw.adjusted_mode.flags |= intel_crt_get_flags(encoder);
+}
+
+/* Note: The caller is required to filter out dpms modes not supported by the
+ * platform. */
+static void intel_crt_set_dpms(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state,
+			       int mode)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crt *crt = intel_encoder_to_crt(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	u32 adpa;
+
+	if (DISPLAY_VER(dev_priv) >= 5)
+		adpa = ADPA_HOTPLUG_BITS;
+	else
+		adpa = 0;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
+		adpa |= ADPA_HSYNC_ACTIVE_HIGH;
+	if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
+		adpa |= ADPA_VSYNC_ACTIVE_HIGH;
+
+	/* For CPT allow 3 pipe config, for others just use A or B */
+	if (HAS_PCH_LPT(dev_priv))
+		; /* Those bits don't exist here */
+	else if (HAS_PCH_CPT(dev_priv))
+		adpa |= ADPA_PIPE_SEL_CPT(crtc->pipe);
+	else
+		adpa |= ADPA_PIPE_SEL(crtc->pipe);
+
+	if (!HAS_PCH_SPLIT(dev_priv))
+		intel_de_write(dev_priv, BCLRPAT(crtc->pipe), 0);
+
+	switch (mode) {
+	case DRM_MODE_DPMS_ON:
+		adpa |= ADPA_DAC_ENABLE;
+		break;
+	case DRM_MODE_DPMS_STANDBY:
+		adpa |= ADPA_DAC_ENABLE | ADPA_HSYNC_CNTL_DISABLE;
+		break;
+	case DRM_MODE_DPMS_SUSPEND:
+		adpa |= ADPA_DAC_ENABLE | ADPA_VSYNC_CNTL_DISABLE;
+		break;
+	case DRM_MODE_DPMS_OFF:
+		adpa |= ADPA_HSYNC_CNTL_DISABLE | ADPA_VSYNC_CNTL_DISABLE;
+		break;
+	}
+
+	intel_de_write(dev_priv, crt->adpa_reg, adpa);
+}
+
+static void intel_disable_crt(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *old_crtc_state,
+			      const struct drm_connector_state *old_conn_state)
+{
+	intel_crt_set_dpms(encoder, old_crtc_state, DRM_MODE_DPMS_OFF);
+}
+
+static void pch_disable_crt(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *old_crtc_state,
+			    const struct drm_connector_state *old_conn_state)
+{
+}
+
+static void pch_post_disable_crt(struct intel_atomic_state *state,
+				 struct intel_encoder *encoder,
+				 const struct intel_crtc_state *old_crtc_state,
+				 const struct drm_connector_state *old_conn_state)
+{
+	intel_disable_crt(state, encoder, old_crtc_state, old_conn_state);
+}
+
+static void hsw_disable_crt(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *old_crtc_state,
+			    const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	drm_WARN_ON(&dev_priv->drm, !old_crtc_state->has_pch_encoder);
+
+	intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, false);
+}
+
+static void hsw_post_disable_crt(struct intel_atomic_state *state,
+				 struct intel_encoder *encoder,
+				 const struct intel_crtc_state *old_crtc_state,
+				 const struct drm_connector_state *old_conn_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	intel_crtc_vblank_off(old_crtc_state);
+
+	intel_disable_transcoder(old_crtc_state);
+
+	intel_ddi_disable_transcoder_func(old_crtc_state);
+
+	ilk_pfit_disable(old_crtc_state);
+
+	intel_ddi_disable_pipe_clock(old_crtc_state);
+
+	pch_post_disable_crt(state, encoder, old_crtc_state, old_conn_state);
+
+	lpt_pch_disable(state, crtc);
+
+	hsw_fdi_disable(encoder);
+
+	drm_WARN_ON(&dev_priv->drm, !old_crtc_state->has_pch_encoder);
+
+	intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, true);
+}
+
+static void hsw_pre_pll_enable_crt(struct intel_atomic_state *state,
+				   struct intel_encoder *encoder,
+				   const struct intel_crtc_state *crtc_state,
+				   const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	drm_WARN_ON(&dev_priv->drm, !crtc_state->has_pch_encoder);
+
+	intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, false);
+}
+
+static void hsw_pre_enable_crt(struct intel_atomic_state *state,
+			       struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state,
+			       const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum pipe pipe = crtc->pipe;
+
+	drm_WARN_ON(&dev_priv->drm, !crtc_state->has_pch_encoder);
+
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false);
+
+	hsw_fdi_link_train(encoder, crtc_state);
+
+	intel_ddi_enable_pipe_clock(encoder, crtc_state);
+}
+
+static void hsw_enable_crt(struct intel_atomic_state *state,
+			   struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state,
+			   const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum pipe pipe = crtc->pipe;
+
+	drm_WARN_ON(&dev_priv->drm, !crtc_state->has_pch_encoder);
+
+	intel_ddi_enable_transcoder_func(encoder, crtc_state);
+
+	intel_enable_transcoder(crtc_state);
+
+	lpt_pch_enable(state, crtc);
+
+	intel_crtc_vblank_on(crtc_state);
+
+	intel_crt_set_dpms(encoder, crtc_state, DRM_MODE_DPMS_ON);
+
+	intel_crtc_wait_for_next_vblank(crtc);
+	intel_crtc_wait_for_next_vblank(crtc);
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
+	intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, true);
+}
+
+static void intel_enable_crt(struct intel_atomic_state *state,
+			     struct intel_encoder *encoder,
+			     const struct intel_crtc_state *crtc_state,
+			     const struct drm_connector_state *conn_state)
+{
+	intel_crt_set_dpms(encoder, crtc_state, DRM_MODE_DPMS_ON);
+}
+
+static enum drm_mode_status
+intel_crt_mode_valid(struct drm_connector *connector,
+		     struct drm_display_mode *mode)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	int max_dotclk = dev_priv->max_dotclk_freq;
+	enum drm_mode_status status;
+	int max_clock;
+
+	status = intel_cpu_transcoder_mode_valid(dev_priv, mode);
+	if (status != MODE_OK)
+		return status;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return MODE_NO_DBLESCAN;
+
+	if (mode->clock < 25000)
+		return MODE_CLOCK_LOW;
+
+	if (HAS_PCH_LPT(dev_priv))
+		max_clock = 180000;
+	else if (IS_VALLEYVIEW(dev_priv))
+		/*
+		 * 270 MHz due to current DPLL limits,
+		 * DAC limit supposedly 355 MHz.
+		 */
+		max_clock = 270000;
+	else if (IS_DISPLAY_VER(dev_priv, 3, 4))
+		max_clock = 400000;
+	else
+		max_clock = 350000;
+	if (mode->clock > max_clock)
+		return MODE_CLOCK_HIGH;
+
+	if (mode->clock > max_dotclk)
+		return MODE_CLOCK_HIGH;
+
+	/* The FDI receiver on LPT only supports 8bpc and only has 2 lanes. */
+	if (HAS_PCH_LPT(dev_priv) &&
+	    ilk_get_lanes_required(mode->clock, 270000, 24) > 2)
+		return MODE_CLOCK_HIGH;
+
+	/* HSW/BDW FDI limited to 4k */
+	if (mode->hdisplay > 4096)
+		return MODE_H_ILLEGAL;
+
+	return MODE_OK;
+}
+
+static int intel_crt_compute_config(struct intel_encoder *encoder,
+				    struct intel_crtc_state *pipe_config,
+				    struct drm_connector_state *conn_state)
+{
+	struct drm_display_mode *adjusted_mode =
+		&pipe_config->hw.adjusted_mode;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+
+	return 0;
+}
+
+static int pch_crt_compute_config(struct intel_encoder *encoder,
+				  struct intel_crtc_state *pipe_config,
+				  struct drm_connector_state *conn_state)
+{
+	struct drm_display_mode *adjusted_mode =
+		&pipe_config->hw.adjusted_mode;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	pipe_config->has_pch_encoder = true;
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+
+	return 0;
+}
+
+static int hsw_crt_compute_config(struct intel_encoder *encoder,
+				  struct intel_crtc_state *pipe_config,
+				  struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct drm_display_mode *adjusted_mode =
+		&pipe_config->hw.adjusted_mode;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	/* HSW/BDW FDI limited to 4k */
+	if (adjusted_mode->crtc_hdisplay > 4096 ||
+	    adjusted_mode->crtc_hblank_start > 4096)
+		return -EINVAL;
+
+	pipe_config->has_pch_encoder = true;
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+
+	/* LPT FDI RX only supports 8bpc. */
+	if (HAS_PCH_LPT(dev_priv)) {
+		if (pipe_config->bw_constrained && pipe_config->pipe_bpp < 24) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "LPT only supports 24bpp\n");
+			return -EINVAL;
+		}
+
+		pipe_config->pipe_bpp = 24;
+	}
+
+	/* FDI must always be 2.7 GHz */
+	pipe_config->port_clock = 135000 * 2;
+
+	adjusted_mode->crtc_clock = lpt_iclkip(pipe_config);
+
+	return 0;
+}
+
+static bool ilk_crt_detect_hotplug(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct intel_crt *crt = intel_attached_crt(to_intel_connector(connector));
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 adpa;
+	bool ret;
+
+	/* The first time through, trigger an explicit detection cycle */
+	if (crt->force_hotplug_required) {
+		bool turn_off_dac = HAS_PCH_SPLIT(dev_priv);
+		u32 save_adpa;
+
+		crt->force_hotplug_required = false;
+
+		save_adpa = adpa = intel_de_read(dev_priv, crt->adpa_reg);
+		drm_dbg_kms(&dev_priv->drm,
+			    "trigger hotplug detect cycle: adpa=0x%x\n", adpa);
+
+		adpa |= ADPA_CRT_HOTPLUG_FORCE_TRIGGER;
+		if (turn_off_dac)
+			adpa &= ~ADPA_DAC_ENABLE;
+
+		intel_de_write(dev_priv, crt->adpa_reg, adpa);
+
+		if (intel_de_wait_for_clear(dev_priv,
+					    crt->adpa_reg,
+					    ADPA_CRT_HOTPLUG_FORCE_TRIGGER,
+					    1000))
+			drm_dbg_kms(&dev_priv->drm,
+				    "timed out waiting for FORCE_TRIGGER");
+
+		if (turn_off_dac) {
+			intel_de_write(dev_priv, crt->adpa_reg, save_adpa);
+			intel_de_posting_read(dev_priv, crt->adpa_reg);
+		}
+	}
+
+	/* Check the status to see if both blue and green are on now */
+	adpa = intel_de_read(dev_priv, crt->adpa_reg);
+	if ((adpa & ADPA_CRT_HOTPLUG_MONITOR_MASK) != 0)
+		ret = true;
+	else
+		ret = false;
+	drm_dbg_kms(&dev_priv->drm, "ironlake hotplug adpa=0x%x, result %d\n",
+		    adpa, ret);
+
+	return ret;
+}
+
+static bool valleyview_crt_detect_hotplug(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct intel_crt *crt = intel_attached_crt(to_intel_connector(connector));
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	bool reenable_hpd;
+	u32 adpa;
+	bool ret;
+	u32 save_adpa;
+
+	/*
+	 * Doing a force trigger causes a hpd interrupt to get sent, which can
+	 * get us stuck in a loop if we're polling:
+	 *  - We enable power wells and reset the ADPA
+	 *  - output_poll_exec does force probe on VGA, triggering a hpd
+	 *  - HPD handler waits for poll to unlock dev->mode_config.mutex
+	 *  - output_poll_exec shuts off the ADPA, unlocks
+	 *    dev->mode_config.mutex
+	 *  - HPD handler runs, resets ADPA and brings us back to the start
+	 *
+	 * Just disable HPD interrupts here to prevent this
+	 */
+	reenable_hpd = intel_hpd_disable(dev_priv, crt->base.hpd_pin);
+
+	save_adpa = adpa = intel_de_read(dev_priv, crt->adpa_reg);
+	drm_dbg_kms(&dev_priv->drm,
+		    "trigger hotplug detect cycle: adpa=0x%x\n", adpa);
+
+	adpa |= ADPA_CRT_HOTPLUG_FORCE_TRIGGER;
+
+	intel_de_write(dev_priv, crt->adpa_reg, adpa);
+
+	if (intel_de_wait_for_clear(dev_priv, crt->adpa_reg,
+				    ADPA_CRT_HOTPLUG_FORCE_TRIGGER, 1000)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "timed out waiting for FORCE_TRIGGER");
+		intel_de_write(dev_priv, crt->adpa_reg, save_adpa);
+	}
+
+	/* Check the status to see if both blue and green are on now */
+	adpa = intel_de_read(dev_priv, crt->adpa_reg);
+	if ((adpa & ADPA_CRT_HOTPLUG_MONITOR_MASK) != 0)
+		ret = true;
+	else
+		ret = false;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "valleyview hotplug adpa=0x%x, result %d\n", adpa, ret);
+
+	if (reenable_hpd)
+		intel_hpd_enable(dev_priv, crt->base.hpd_pin);
+
+	return ret;
+}
+
+static bool intel_crt_detect_hotplug(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 stat;
+	bool ret = false;
+	int i, tries = 0;
+
+	if (HAS_PCH_SPLIT(dev_priv))
+		return ilk_crt_detect_hotplug(connector);
+
+	if (IS_VALLEYVIEW(dev_priv))
+		return valleyview_crt_detect_hotplug(connector);
+
+	/*
+	 * On 4 series desktop, CRT detect sequence need to be done twice
+	 * to get a reliable result.
+	 */
+
+	if (IS_G45(dev_priv))
+		tries = 2;
+	else
+		tries = 1;
+
+	for (i = 0; i < tries ; i++) {
+		/* turn on the FORCE_DETECT */
+		i915_hotplug_interrupt_update(dev_priv,
+					      CRT_HOTPLUG_FORCE_DETECT,
+					      CRT_HOTPLUG_FORCE_DETECT);
+		/* wait for FORCE_DETECT to go off */
+		if (intel_de_wait_for_clear(dev_priv, PORT_HOTPLUG_EN,
+					    CRT_HOTPLUG_FORCE_DETECT, 1000))
+			drm_dbg_kms(&dev_priv->drm,
+				    "timed out waiting for FORCE_DETECT to go off");
+	}
+
+	stat = intel_de_read(dev_priv, PORT_HOTPLUG_STAT);
+	if ((stat & CRT_HOTPLUG_MONITOR_MASK) != CRT_HOTPLUG_MONITOR_NONE)
+		ret = true;
+
+	/* clear the interrupt we just generated, if any */
+	intel_de_write(dev_priv, PORT_HOTPLUG_STAT, CRT_HOTPLUG_INT_STATUS);
+
+	i915_hotplug_interrupt_update(dev_priv, CRT_HOTPLUG_FORCE_DETECT, 0);
+
+	return ret;
+}
+
+static struct edid *intel_crt_get_edid(struct drm_connector *connector,
+				struct i2c_adapter *i2c)
+{
+	struct edid *edid;
+
+	edid = drm_get_edid(connector, i2c);
+
+	if (!edid && !intel_gmbus_is_forced_bit(i2c)) {
+		drm_dbg_kms(connector->dev,
+			    "CRT GMBUS EDID read failed, retry using GPIO bit-banging\n");
+		intel_gmbus_force_bit(i2c, true);
+		edid = drm_get_edid(connector, i2c);
+		intel_gmbus_force_bit(i2c, false);
+	}
+
+	return edid;
+}
+
+/* local version of intel_ddc_get_modes() to use intel_crt_get_edid() */
+static int intel_crt_ddc_get_modes(struct drm_connector *connector,
+				struct i2c_adapter *adapter)
+{
+	struct edid *edid;
+	int ret;
+
+	edid = intel_crt_get_edid(connector, adapter);
+	if (!edid)
+		return 0;
+
+	ret = intel_connector_update_modes(connector, edid);
+	kfree(edid);
+
+	return ret;
+}
+
+static bool intel_crt_detect_ddc(struct drm_connector *connector)
+{
+	struct intel_crt *crt = intel_attached_crt(to_intel_connector(connector));
+	struct drm_i915_private *dev_priv = to_i915(crt->base.base.dev);
+	struct edid *edid;
+	struct i2c_adapter *i2c;
+	bool ret = false;
+
+	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->display.vbt.crt_ddc_pin);
+	edid = intel_crt_get_edid(connector, i2c);
+
+	if (edid) {
+		bool is_digital = edid->input & DRM_EDID_INPUT_DIGITAL;
+
+		/*
+		 * This may be a DVI-I connector with a shared DDC
+		 * link between analog and digital outputs, so we
+		 * have to check the EDID input spec of the attached device.
+		 */
+		if (!is_digital) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "CRT detected via DDC:0x50 [EDID]\n");
+			ret = true;
+		} else {
+			drm_dbg_kms(&dev_priv->drm,
+				    "CRT not detected via DDC:0x50 [EDID reports a digital panel]\n");
+		}
+	} else {
+		drm_dbg_kms(&dev_priv->drm,
+			    "CRT not detected via DDC:0x50 [no valid EDID found]\n");
+	}
+
+	kfree(edid);
+
+	return ret;
+}
+
+static enum drm_connector_status
+intel_crt_load_detect(struct intel_crt *crt, u32 pipe)
+{
+	struct drm_device *dev = crt->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	u32 save_bclrpat;
+	u32 save_vtotal;
+	u32 vtotal, vactive;
+	u32 vsample;
+	u32 vblank, vblank_start, vblank_end;
+	u32 dsl;
+	i915_reg_t bclrpat_reg, vtotal_reg,
+		vblank_reg, vsync_reg, pipeconf_reg, pipe_dsl_reg;
+	u8 st00;
+	enum drm_connector_status status;
+
+	drm_dbg_kms(&dev_priv->drm, "starting load-detect on CRT\n");
+
+	bclrpat_reg = BCLRPAT(pipe);
+	vtotal_reg = VTOTAL(pipe);
+	vblank_reg = VBLANK(pipe);
+	vsync_reg = VSYNC(pipe);
+	pipeconf_reg = PIPECONF(pipe);
+	pipe_dsl_reg = PIPEDSL(pipe);
+
+	save_bclrpat = intel_uncore_read(uncore, bclrpat_reg);
+	save_vtotal = intel_uncore_read(uncore, vtotal_reg);
+	vblank = intel_uncore_read(uncore, vblank_reg);
+
+	vtotal = ((save_vtotal >> 16) & 0xfff) + 1;
+	vactive = (save_vtotal & 0x7ff) + 1;
+
+	vblank_start = (vblank & 0xfff) + 1;
+	vblank_end = ((vblank >> 16) & 0xfff) + 1;
+
+	/* Set the border color to purple. */
+	intel_uncore_write(uncore, bclrpat_reg, 0x500050);
+
+	if (DISPLAY_VER(dev_priv) != 2) {
+		u32 pipeconf = intel_uncore_read(uncore, pipeconf_reg);
+		intel_uncore_write(uncore,
+				   pipeconf_reg,
+				   pipeconf | PIPECONF_FORCE_BORDER);
+		intel_uncore_posting_read(uncore, pipeconf_reg);
+		/* Wait for next Vblank to substitue
+		 * border color for Color info */
+		intel_crtc_wait_for_next_vblank(intel_crtc_for_pipe(dev_priv, pipe));
+		st00 = intel_uncore_read8(uncore, _VGA_MSR_WRITE);
+		status = ((st00 & (1 << 4)) != 0) ?
+			connector_status_connected :
+			connector_status_disconnected;
+
+		intel_uncore_write(uncore, pipeconf_reg, pipeconf);
+	} else {
+		bool restore_vblank = false;
+		int count, detect;
+
+		/*
+		* If there isn't any border, add some.
+		* Yes, this will flicker
+		*/
+		if (vblank_start <= vactive && vblank_end >= vtotal) {
+			u32 vsync = intel_de_read(dev_priv, vsync_reg);
+			u32 vsync_start = (vsync & 0xffff) + 1;
+
+			vblank_start = vsync_start;
+			intel_uncore_write(uncore,
+					   vblank_reg,
+					   (vblank_start - 1) |
+					   ((vblank_end - 1) << 16));
+			restore_vblank = true;
+		}
+		/* sample in the vertical border, selecting the larger one */
+		if (vblank_start - vactive >= vtotal - vblank_end)
+			vsample = (vblank_start + vactive) >> 1;
+		else
+			vsample = (vtotal + vblank_end) >> 1;
+
+		/*
+		 * Wait for the border to be displayed
+		 */
+		while (intel_uncore_read(uncore, pipe_dsl_reg) >= vactive)
+			;
+		while ((dsl = intel_uncore_read(uncore, pipe_dsl_reg)) <=
+		       vsample)
+			;
+		/*
+		 * Watch ST00 for an entire scanline
+		 */
+		detect = 0;
+		count = 0;
+		do {
+			count++;
+			/* Read the ST00 VGA status register */
+			st00 = intel_uncore_read8(uncore, _VGA_MSR_WRITE);
+			if (st00 & (1 << 4))
+				detect++;
+		} while ((intel_uncore_read(uncore, pipe_dsl_reg) == dsl));
+
+		/* restore vblank if necessary */
+		if (restore_vblank)
+			intel_uncore_write(uncore, vblank_reg, vblank);
+		/*
+		 * If more than 3/4 of the scanline detected a monitor,
+		 * then it is assumed to be present. This works even on i830,
+		 * where there isn't any way to force the border color across
+		 * the screen
+		 */
+		status = detect * 4 > count * 3 ?
+			 connector_status_connected :
+			 connector_status_disconnected;
+	}
+
+	/* Restore previous settings */
+	intel_uncore_write(uncore, bclrpat_reg, save_bclrpat);
+
+	return status;
+}
+
+static int intel_spurious_crt_detect_dmi_callback(const struct dmi_system_id *id)
+{
+	DRM_DEBUG_DRIVER("Skipping CRT detection for %s\n", id->ident);
+	return 1;
+}
+
+static const struct dmi_system_id intel_spurious_crt_detect[] = {
+	{
+		.callback = intel_spurious_crt_detect_dmi_callback,
+		.ident = "ACER ZGB",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "ACER"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "ZGB"),
+		},
+	},
+	{
+		.callback = intel_spurious_crt_detect_dmi_callback,
+		.ident = "Intel DZ77BH-55K",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Intel Corporation"),
+			DMI_MATCH(DMI_BOARD_NAME, "DZ77BH-55K"),
+		},
+	},
+	{ }
+};
+
+static int
+intel_crt_detect(struct drm_connector *connector,
+		 struct drm_modeset_acquire_ctx *ctx,
+		 bool force)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_crt *crt = intel_attached_crt(to_intel_connector(connector));
+	struct intel_encoder *intel_encoder = &crt->base;
+	intel_wakeref_t wakeref;
+	int status, ret;
+	struct intel_load_detect_pipe tmp;
+
+	drm_dbg_kms(&dev_priv->drm, "[CONNECTOR:%d:%s] force=%d\n",
+		    connector->base.id, connector->name,
+		    force);
+
+	if (!INTEL_DISPLAY_ENABLED(dev_priv))
+		return connector_status_disconnected;
+
+	if (dev_priv->params.load_detect_test) {
+		wakeref = intel_display_power_get(dev_priv,
+						  intel_encoder->power_domain);
+		goto load_detect;
+	}
+
+	/* Skip machines without VGA that falsely report hotplug events */
+	if (dmi_check_system(intel_spurious_crt_detect))
+		return connector_status_disconnected;
+
+	wakeref = intel_display_power_get(dev_priv,
+					  intel_encoder->power_domain);
+
+	if (I915_HAS_HOTPLUG(dev_priv)) {
+		/* We can not rely on the HPD pin always being correctly wired
+		 * up, for example many KVM do not pass it through, and so
+		 * only trust an assertion that the monitor is connected.
+		 */
+		if (intel_crt_detect_hotplug(connector)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "CRT detected via hotplug\n");
+			status = connector_status_connected;
+			goto out;
+		} else
+			drm_dbg_kms(&dev_priv->drm,
+				    "CRT not detected via hotplug\n");
+	}
+
+	if (intel_crt_detect_ddc(connector)) {
+		status = connector_status_connected;
+		goto out;
+	}
+
+	/* Load detection is broken on HPD capable machines. Whoever wants a
+	 * broken monitor (without edid) to work behind a broken kvm (that fails
+	 * to have the right resistors for HP detection) needs to fix this up.
+	 * For now just bail out. */
+	if (I915_HAS_HOTPLUG(dev_priv)) {
+		status = connector_status_disconnected;
+		goto out;
+	}
+
+load_detect:
+	if (!force) {
+		status = connector->status;
+		goto out;
+	}
+
+	/* for pre-945g platforms use load detect */
+	ret = intel_get_load_detect_pipe(connector, &tmp, ctx);
+	if (ret > 0) {
+		if (intel_crt_detect_ddc(connector))
+			status = connector_status_connected;
+		else if (DISPLAY_VER(dev_priv) < 4)
+			status = intel_crt_load_detect(crt,
+				to_intel_crtc(connector->state->crtc)->pipe);
+		else if (dev_priv->params.load_detect_test)
+			status = connector_status_disconnected;
+		else
+			status = connector_status_unknown;
+		intel_release_load_detect_pipe(connector, &tmp, ctx);
+	} else if (ret == 0) {
+		status = connector_status_unknown;
+	} else {
+		status = ret;
+	}
+
+out:
+	intel_display_power_put(dev_priv, intel_encoder->power_domain, wakeref);
+
+	/*
+	 * Make sure the refs for power wells enabled during detect are
+	 * dropped to avoid a new detect cycle triggered by HPD polling.
+	 */
+	intel_display_power_flush_work(dev_priv);
+
+	return status;
+}
+
+static int intel_crt_get_modes(struct drm_connector *connector)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_crt *crt = intel_attached_crt(to_intel_connector(connector));
+	struct intel_encoder *intel_encoder = &crt->base;
+	intel_wakeref_t wakeref;
+	struct i2c_adapter *i2c;
+	int ret;
+
+	wakeref = intel_display_power_get(dev_priv,
+					  intel_encoder->power_domain);
+
+	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->display.vbt.crt_ddc_pin);
+	ret = intel_crt_ddc_get_modes(connector, i2c);
+	if (ret || !IS_G4X(dev_priv))
+		goto out;
+
+	/* Try to probe digital port for output in DVI-I -> VGA mode. */
+	i2c = intel_gmbus_get_adapter(dev_priv, GMBUS_PIN_DPB);
+	ret = intel_crt_ddc_get_modes(connector, i2c);
+
+out:
+	intel_display_power_put(dev_priv, intel_encoder->power_domain, wakeref);
+
+	return ret;
+}
+
+void intel_crt_reset(struct drm_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
+	struct intel_crt *crt = intel_encoder_to_crt(to_intel_encoder(encoder));
+
+	if (DISPLAY_VER(dev_priv) >= 5) {
+		u32 adpa;
+
+		adpa = intel_de_read(dev_priv, crt->adpa_reg);
+		adpa &= ~ADPA_CRT_HOTPLUG_MASK;
+		adpa |= ADPA_HOTPLUG_BITS;
+		intel_de_write(dev_priv, crt->adpa_reg, adpa);
+		intel_de_posting_read(dev_priv, crt->adpa_reg);
+
+		drm_dbg_kms(&dev_priv->drm, "crt adpa set to 0x%x\n", adpa);
+		crt->force_hotplug_required = true;
+	}
+
+}
+
+/*
+ * Routines for controlling stuff on the analog port
+ */
+
+static const struct drm_connector_funcs intel_crt_connector_funcs = {
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.late_register = intel_connector_register,
+	.early_unregister = intel_connector_unregister,
+	.destroy = intel_connector_destroy,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+};
+
+static const struct drm_connector_helper_funcs intel_crt_connector_helper_funcs = {
+	.detect_ctx = intel_crt_detect,
+	.mode_valid = intel_crt_mode_valid,
+	.get_modes = intel_crt_get_modes,
+};
+
+static const struct drm_encoder_funcs intel_crt_enc_funcs = {
+	.reset = intel_crt_reset,
+	.destroy = intel_encoder_destroy,
+};
+
+void intel_crt_init(struct drm_i915_private *dev_priv)
+{
+	struct drm_connector *connector;
+	struct intel_crt *crt;
+	struct intel_connector *intel_connector;
+	i915_reg_t adpa_reg;
+	u32 adpa;
+
+	if (HAS_PCH_SPLIT(dev_priv))
+		adpa_reg = PCH_ADPA;
+	else if (IS_VALLEYVIEW(dev_priv))
+		adpa_reg = VLV_ADPA;
+	else
+		adpa_reg = ADPA;
+
+	adpa = intel_de_read(dev_priv, adpa_reg);
+	if ((adpa & ADPA_DAC_ENABLE) == 0) {
+		/*
+		 * On some machines (some IVB at least) CRT can be
+		 * fused off, but there's no known fuse bit to
+		 * indicate that. On these machine the ADPA register
+		 * works normally, except the DAC enable bit won't
+		 * take. So the only way to tell is attempt to enable
+		 * it and see what happens.
+		 */
+		intel_de_write(dev_priv, adpa_reg,
+			       adpa | ADPA_DAC_ENABLE | ADPA_HSYNC_CNTL_DISABLE | ADPA_VSYNC_CNTL_DISABLE);
+		if ((intel_de_read(dev_priv, adpa_reg) & ADPA_DAC_ENABLE) == 0)
+			return;
+		intel_de_write(dev_priv, adpa_reg, adpa);
+	}
+
+	crt = kzalloc(sizeof(struct intel_crt), GFP_KERNEL);
+	if (!crt)
+		return;
+
+	intel_connector = intel_connector_alloc();
+	if (!intel_connector) {
+		kfree(crt);
+		return;
+	}
+
+	connector = &intel_connector->base;
+	crt->connector = intel_connector;
+	drm_connector_init(&dev_priv->drm, &intel_connector->base,
+			   &intel_crt_connector_funcs, DRM_MODE_CONNECTOR_VGA);
+
+	drm_encoder_init(&dev_priv->drm, &crt->base.base, &intel_crt_enc_funcs,
+			 DRM_MODE_ENCODER_DAC, "CRT");
+
+	intel_connector_attach_encoder(intel_connector, &crt->base);
+
+	crt->base.type = INTEL_OUTPUT_ANALOG;
+	crt->base.cloneable = (1 << INTEL_OUTPUT_DVO) | (1 << INTEL_OUTPUT_HDMI);
+	if (IS_I830(dev_priv))
+		crt->base.pipe_mask = BIT(PIPE_A);
+	else
+		crt->base.pipe_mask = ~0;
+
+	if (DISPLAY_VER(dev_priv) == 2)
+		connector->interlace_allowed = 0;
+	else
+		connector->interlace_allowed = 1;
+	connector->doublescan_allowed = 0;
+
+	crt->adpa_reg = adpa_reg;
+
+	crt->base.power_domain = POWER_DOMAIN_PORT_CRT;
+
+	if (I915_HAS_HOTPLUG(dev_priv) &&
+	    !dmi_check_system(intel_spurious_crt_detect)) {
+		crt->base.hpd_pin = HPD_CRT;
+		crt->base.hotplug = intel_encoder_hotplug;
+		intel_connector->polled = DRM_CONNECTOR_POLL_HPD;
+	} else {
+		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT;
+	}
+
+	if (HAS_DDI(dev_priv)) {
+		crt->base.port = PORT_E;
+		crt->base.get_config = hsw_crt_get_config;
+		crt->base.get_hw_state = intel_ddi_get_hw_state;
+		crt->base.compute_config = hsw_crt_compute_config;
+		crt->base.pre_pll_enable = hsw_pre_pll_enable_crt;
+		crt->base.pre_enable = hsw_pre_enable_crt;
+		crt->base.enable = hsw_enable_crt;
+		crt->base.disable = hsw_disable_crt;
+		crt->base.post_disable = hsw_post_disable_crt;
+		crt->base.enable_clock = hsw_ddi_enable_clock;
+		crt->base.disable_clock = hsw_ddi_disable_clock;
+		crt->base.is_clock_enabled = hsw_ddi_is_clock_enabled;
+
+		intel_ddi_buf_trans_init(&crt->base);
+	} else {
+		if (HAS_PCH_SPLIT(dev_priv)) {
+			crt->base.compute_config = pch_crt_compute_config;
+			crt->base.disable = pch_disable_crt;
+			crt->base.post_disable = pch_post_disable_crt;
+		} else {
+			crt->base.compute_config = intel_crt_compute_config;
+			crt->base.disable = intel_disable_crt;
+		}
+		crt->base.port = PORT_NONE;
+		crt->base.get_config = intel_crt_get_config;
+		crt->base.get_hw_state = intel_crt_get_hw_state;
+		crt->base.enable = intel_enable_crt;
+	}
+	intel_connector->get_hw_state = intel_connector_get_hw_state;
+
+	drm_connector_helper_add(connector, &intel_crt_connector_helper_funcs);
+
+	/*
+	 * TODO: find a proper way to discover whether we need to set the the
+	 * polarity and link reversal bits or not, instead of relying on the
+	 * BIOS.
+	 */
+	if (HAS_PCH_LPT(dev_priv)) {
+		u32 fdi_config = FDI_RX_POLARITY_REVERSED_LPT |
+				 FDI_RX_LINK_REVERSAL_OVERRIDE;
+
+		dev_priv->display.fdi.rx_config = intel_de_read(dev_priv,
+								FDI_RX_CTL(PIPE_A)) & fdi_config;
+	}
+
+	intel_crt_reset(&crt->base.base);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_crt.h b/drivers/gpu/drm/i915/display/intel_crt.h
new file mode 100644
index 000000000..c6071efd9
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_crt.h
@@ -0,0 +1,20 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_CRT_H__
+#define __INTEL_CRT_H__
+
+#include "i915_reg_defs.h"
+
+enum pipe;
+struct drm_encoder;
+struct drm_i915_private;
+
+bool intel_crt_port_enabled(struct drm_i915_private *dev_priv,
+			    i915_reg_t adpa_reg, enum pipe *pipe);
+void intel_crt_init(struct drm_i915_private *dev_priv);
+void intel_crt_reset(struct drm_encoder *encoder);
+
+#endif /* __INTEL_CRT_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_crtc.c b/drivers/gpu/drm/i915/display/intel_crtc.c
new file mode 100644
index 000000000..6792a9056
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_crtc.c
@@ -0,0 +1,705 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+#include <linux/kernel.h>
+#include <linux/pm_qos.h>
+#include <linux/slab.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_plane.h>
+#include <drm/drm_vblank_work.h>
+
+#include "i915_irq.h"
+#include "i915_vgpu.h"
+#include "i9xx_plane.h"
+#include "icl_dsi.h"
+#include "intel_atomic.h"
+#include "intel_atomic_plane.h"
+#include "intel_color.h"
+#include "intel_crtc.h"
+#include "intel_cursor.h"
+#include "intel_display_debugfs.h"
+#include "intel_display_trace.h"
+#include "intel_display_types.h"
+#include "intel_drrs.h"
+#include "intel_dsi.h"
+#include "intel_pipe_crc.h"
+#include "intel_psr.h"
+#include "intel_sprite.h"
+#include "intel_vrr.h"
+#include "skl_universal_plane.h"
+
+static void assert_vblank_disabled(struct drm_crtc *crtc)
+{
+	if (I915_STATE_WARN_ON(drm_crtc_vblank_get(crtc) == 0))
+		drm_crtc_vblank_put(crtc);
+}
+
+struct intel_crtc *intel_first_crtc(struct drm_i915_private *i915)
+{
+	return to_intel_crtc(drm_crtc_from_index(&i915->drm, 0));
+}
+
+struct intel_crtc *intel_crtc_for_pipe(struct drm_i915_private *i915,
+				       enum pipe pipe)
+{
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		if (crtc->pipe == pipe)
+			return crtc;
+	}
+
+	return NULL;
+}
+
+void intel_crtc_wait_for_next_vblank(struct intel_crtc *crtc)
+{
+	drm_crtc_wait_one_vblank(&crtc->base);
+}
+
+void intel_wait_for_vblank_if_active(struct drm_i915_private *i915,
+				     enum pipe pipe)
+{
+	struct intel_crtc *crtc = intel_crtc_for_pipe(i915, pipe);
+
+	if (crtc->active)
+		intel_crtc_wait_for_next_vblank(crtc);
+}
+
+u32 intel_crtc_get_vblank_counter(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_vblank_crtc *vblank = &dev->vblank[drm_crtc_index(&crtc->base)];
+
+	if (!crtc->active)
+		return 0;
+
+	if (!vblank->max_vblank_count)
+		return (u32)drm_crtc_accurate_vblank_count(&crtc->base);
+
+	return crtc->base.funcs->get_vblank_counter(&crtc->base);
+}
+
+u32 intel_crtc_max_vblank_count(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	/*
+	 * From Gen 11, In case of dsi cmd mode, frame counter wouldnt
+	 * have updated at the beginning of TE, if we want to use
+	 * the hw counter, then we would find it updated in only
+	 * the next TE, hence switching to sw counter.
+	 */
+	if (crtc_state->mode_flags & (I915_MODE_FLAG_DSI_USE_TE0 |
+				      I915_MODE_FLAG_DSI_USE_TE1))
+		return 0;
+
+	/*
+	 * On i965gm the hardware frame counter reads
+	 * zero when the TV encoder is enabled :(
+	 */
+	if (IS_I965GM(dev_priv) &&
+	    (crtc_state->output_types & BIT(INTEL_OUTPUT_TVOUT)))
+		return 0;
+
+	if (DISPLAY_VER(dev_priv) >= 5 || IS_G4X(dev_priv))
+		return 0xffffffff; /* full 32 bit counter */
+	else if (DISPLAY_VER(dev_priv) >= 3)
+		return 0xffffff; /* only 24 bits of frame count */
+	else
+		return 0; /* Gen2 doesn't have a hardware frame counter */
+}
+
+void intel_crtc_vblank_on(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	assert_vblank_disabled(&crtc->base);
+	drm_crtc_set_max_vblank_count(&crtc->base,
+				      intel_crtc_max_vblank_count(crtc_state));
+	drm_crtc_vblank_on(&crtc->base);
+
+	/*
+	 * Should really happen exactly when we enable the pipe
+	 * but we want the frame counters in the trace, and that
+	 * requires vblank support on some platforms/outputs.
+	 */
+	trace_intel_pipe_enable(crtc);
+}
+
+void intel_crtc_vblank_off(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	/*
+	 * Should really happen exactly when we disable the pipe
+	 * but we want the frame counters in the trace, and that
+	 * requires vblank support on some platforms/outputs.
+	 */
+	trace_intel_pipe_disable(crtc);
+
+	drm_crtc_vblank_off(&crtc->base);
+	assert_vblank_disabled(&crtc->base);
+}
+
+struct intel_crtc_state *intel_crtc_state_alloc(struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state;
+
+	crtc_state = kmalloc(sizeof(*crtc_state), GFP_KERNEL);
+
+	if (crtc_state)
+		intel_crtc_state_reset(crtc_state, crtc);
+
+	return crtc_state;
+}
+
+void intel_crtc_state_reset(struct intel_crtc_state *crtc_state,
+			    struct intel_crtc *crtc)
+{
+	memset(crtc_state, 0, sizeof(*crtc_state));
+
+	__drm_atomic_helper_crtc_state_reset(&crtc_state->uapi, &crtc->base);
+
+	crtc_state->cpu_transcoder = INVALID_TRANSCODER;
+	crtc_state->master_transcoder = INVALID_TRANSCODER;
+	crtc_state->hsw_workaround_pipe = INVALID_PIPE;
+	crtc_state->scaler_state.scaler_id = -1;
+	crtc_state->mst_master_transcoder = INVALID_TRANSCODER;
+}
+
+static struct intel_crtc *intel_crtc_alloc(void)
+{
+	struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+
+	crtc = kzalloc(sizeof(*crtc), GFP_KERNEL);
+	if (!crtc)
+		return ERR_PTR(-ENOMEM);
+
+	crtc_state = intel_crtc_state_alloc(crtc);
+	if (!crtc_state) {
+		kfree(crtc);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	crtc->base.state = &crtc_state->uapi;
+	crtc->config = crtc_state;
+
+	return crtc;
+}
+
+static void intel_crtc_free(struct intel_crtc *crtc)
+{
+	intel_crtc_destroy_state(&crtc->base, crtc->base.state);
+	kfree(crtc);
+}
+
+static void intel_crtc_destroy(struct drm_crtc *_crtc)
+{
+	struct intel_crtc *crtc = to_intel_crtc(_crtc);
+
+	cpu_latency_qos_remove_request(&crtc->vblank_pm_qos);
+
+	drm_crtc_cleanup(&crtc->base);
+	kfree(crtc);
+}
+
+static int intel_crtc_late_register(struct drm_crtc *crtc)
+{
+	intel_crtc_debugfs_add(crtc);
+	return 0;
+}
+
+#define INTEL_CRTC_FUNCS \
+	.set_config = drm_atomic_helper_set_config, \
+	.destroy = intel_crtc_destroy, \
+	.page_flip = drm_atomic_helper_page_flip, \
+	.atomic_duplicate_state = intel_crtc_duplicate_state, \
+	.atomic_destroy_state = intel_crtc_destroy_state, \
+	.set_crc_source = intel_crtc_set_crc_source, \
+	.verify_crc_source = intel_crtc_verify_crc_source, \
+	.get_crc_sources = intel_crtc_get_crc_sources, \
+	.late_register = intel_crtc_late_register
+
+static const struct drm_crtc_funcs bdw_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = g4x_get_vblank_counter,
+	.enable_vblank = bdw_enable_vblank,
+	.disable_vblank = bdw_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs ilk_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = g4x_get_vblank_counter,
+	.enable_vblank = ilk_enable_vblank,
+	.disable_vblank = ilk_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs g4x_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = g4x_get_vblank_counter,
+	.enable_vblank = i965_enable_vblank,
+	.disable_vblank = i965_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs i965_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = i915_get_vblank_counter,
+	.enable_vblank = i965_enable_vblank,
+	.disable_vblank = i965_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs i915gm_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = i915_get_vblank_counter,
+	.enable_vblank = i915gm_enable_vblank,
+	.disable_vblank = i915gm_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs i915_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	.get_vblank_counter = i915_get_vblank_counter,
+	.enable_vblank = i8xx_enable_vblank,
+	.disable_vblank = i8xx_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+static const struct drm_crtc_funcs i8xx_crtc_funcs = {
+	INTEL_CRTC_FUNCS,
+
+	/* no hw vblank counter */
+	.enable_vblank = i8xx_enable_vblank,
+	.disable_vblank = i8xx_disable_vblank,
+	.get_vblank_timestamp = intel_crtc_get_vblank_timestamp,
+};
+
+int intel_crtc_init(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	struct intel_plane *primary, *cursor;
+	const struct drm_crtc_funcs *funcs;
+	struct intel_crtc *crtc;
+	int sprite, ret;
+
+	crtc = intel_crtc_alloc();
+	if (IS_ERR(crtc))
+		return PTR_ERR(crtc);
+
+	crtc->pipe = pipe;
+	crtc->num_scalers = RUNTIME_INFO(dev_priv)->num_scalers[pipe];
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		primary = skl_universal_plane_create(dev_priv, pipe,
+						     PLANE_PRIMARY);
+	else
+		primary = intel_primary_plane_create(dev_priv, pipe);
+	if (IS_ERR(primary)) {
+		ret = PTR_ERR(primary);
+		goto fail;
+	}
+	crtc->plane_ids_mask |= BIT(primary->id);
+
+	for_each_sprite(dev_priv, pipe, sprite) {
+		struct intel_plane *plane;
+
+		if (DISPLAY_VER(dev_priv) >= 9)
+			plane = skl_universal_plane_create(dev_priv, pipe,
+							   PLANE_SPRITE0 + sprite);
+		else
+			plane = intel_sprite_plane_create(dev_priv, pipe, sprite);
+		if (IS_ERR(plane)) {
+			ret = PTR_ERR(plane);
+			goto fail;
+		}
+		crtc->plane_ids_mask |= BIT(plane->id);
+	}
+
+	cursor = intel_cursor_plane_create(dev_priv, pipe);
+	if (IS_ERR(cursor)) {
+		ret = PTR_ERR(cursor);
+		goto fail;
+	}
+	crtc->plane_ids_mask |= BIT(cursor->id);
+
+	if (HAS_GMCH(dev_priv)) {
+		if (IS_CHERRYVIEW(dev_priv) ||
+		    IS_VALLEYVIEW(dev_priv) || IS_G4X(dev_priv))
+			funcs = &g4x_crtc_funcs;
+		else if (DISPLAY_VER(dev_priv) == 4)
+			funcs = &i965_crtc_funcs;
+		else if (IS_I945GM(dev_priv) || IS_I915GM(dev_priv))
+			funcs = &i915gm_crtc_funcs;
+		else if (DISPLAY_VER(dev_priv) == 3)
+			funcs = &i915_crtc_funcs;
+		else
+			funcs = &i8xx_crtc_funcs;
+	} else {
+		if (DISPLAY_VER(dev_priv) >= 8)
+			funcs = &bdw_crtc_funcs;
+		else
+			funcs = &ilk_crtc_funcs;
+	}
+
+	ret = drm_crtc_init_with_planes(&dev_priv->drm, &crtc->base,
+					&primary->base, &cursor->base,
+					funcs, "pipe %c", pipe_name(pipe));
+	if (ret)
+		goto fail;
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		drm_crtc_create_scaling_filter_property(&crtc->base,
+						BIT(DRM_SCALING_FILTER_DEFAULT) |
+						BIT(DRM_SCALING_FILTER_NEAREST_NEIGHBOR));
+
+	intel_color_init(crtc);
+
+	intel_crtc_drrs_init(crtc);
+	intel_crtc_crc_init(crtc);
+
+	cpu_latency_qos_add_request(&crtc->vblank_pm_qos, PM_QOS_DEFAULT_VALUE);
+
+	drm_WARN_ON(&dev_priv->drm, drm_crtc_index(&crtc->base) != crtc->pipe);
+
+	return 0;
+
+fail:
+	intel_crtc_free(crtc);
+
+	return ret;
+}
+
+static bool intel_crtc_needs_vblank_work(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->hw.active &&
+		!intel_crtc_needs_modeset(crtc_state) &&
+		!crtc_state->preload_luts &&
+		(crtc_state->uapi.color_mgmt_changed ||
+		 crtc_state->update_pipe);
+}
+
+static void intel_crtc_vblank_work(struct kthread_work *base)
+{
+	struct drm_vblank_work *work = to_drm_vblank_work(base);
+	struct intel_crtc_state *crtc_state =
+		container_of(work, typeof(*crtc_state), vblank_work);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	trace_intel_crtc_vblank_work_start(crtc);
+
+	intel_color_load_luts(crtc_state);
+
+	if (crtc_state->uapi.event) {
+		spin_lock_irq(&crtc->base.dev->event_lock);
+		drm_crtc_send_vblank_event(&crtc->base, crtc_state->uapi.event);
+		crtc_state->uapi.event = NULL;
+		spin_unlock_irq(&crtc->base.dev->event_lock);
+	}
+
+	trace_intel_crtc_vblank_work_end(crtc);
+}
+
+static void intel_crtc_vblank_work_init(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	drm_vblank_work_init(&crtc_state->vblank_work, &crtc->base,
+			     intel_crtc_vblank_work);
+	/*
+	 * Interrupt latency is critical for getting the vblank
+	 * work executed as early as possible during the vblank.
+	 */
+	cpu_latency_qos_update_request(&crtc->vblank_pm_qos, 0);
+}
+
+void intel_wait_for_vblank_workers(struct intel_atomic_state *state)
+{
+	struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+	int i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		if (!intel_crtc_needs_vblank_work(crtc_state))
+			continue;
+
+		drm_vblank_work_flush(&crtc_state->vblank_work);
+		cpu_latency_qos_update_request(&crtc->vblank_pm_qos,
+					       PM_QOS_DEFAULT_VALUE);
+	}
+}
+
+int intel_usecs_to_scanlines(const struct drm_display_mode *adjusted_mode,
+			     int usecs)
+{
+	/* paranoia */
+	if (!adjusted_mode->crtc_htotal)
+		return 1;
+
+	return DIV_ROUND_UP(usecs * adjusted_mode->crtc_clock,
+			    1000 * adjusted_mode->crtc_htotal);
+}
+
+static int intel_mode_vblank_start(const struct drm_display_mode *mode)
+{
+	int vblank_start = mode->crtc_vblank_start;
+
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		vblank_start = DIV_ROUND_UP(vblank_start, 2);
+
+	return vblank_start;
+}
+
+/**
+ * intel_pipe_update_start() - start update of a set of display registers
+ * @new_crtc_state: the new crtc state
+ *
+ * Mark the start of an update to pipe registers that should be updated
+ * atomically regarding vblank. If the next vblank will happens within
+ * the next 100 us, this function waits until the vblank passes.
+ *
+ * After a successful call to this function, interrupts will be disabled
+ * until a subsequent call to intel_pipe_update_end(). That is done to
+ * avoid random delays.
+ */
+void intel_pipe_update_start(struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_display_mode *adjusted_mode = &new_crtc_state->hw.adjusted_mode;
+	long timeout = msecs_to_jiffies_timeout(1);
+	int scanline, min, max, vblank_start;
+	wait_queue_head_t *wq = drm_crtc_vblank_waitqueue(&crtc->base);
+	bool need_vlv_dsi_wa = (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+		intel_crtc_has_type(new_crtc_state, INTEL_OUTPUT_DSI);
+	DEFINE_WAIT(wait);
+
+	intel_psr_lock(new_crtc_state);
+
+	if (new_crtc_state->do_async_flip)
+		return;
+
+	if (intel_crtc_needs_vblank_work(new_crtc_state))
+		intel_crtc_vblank_work_init(new_crtc_state);
+
+	if (new_crtc_state->vrr.enable) {
+		if (intel_vrr_is_push_sent(new_crtc_state))
+			vblank_start = intel_vrr_vmin_vblank_start(new_crtc_state);
+		else
+			vblank_start = intel_vrr_vmax_vblank_start(new_crtc_state);
+	} else {
+		vblank_start = intel_mode_vblank_start(adjusted_mode);
+	}
+
+	/* FIXME needs to be calibrated sensibly */
+	min = vblank_start - intel_usecs_to_scanlines(adjusted_mode,
+						      VBLANK_EVASION_TIME_US);
+	max = vblank_start - 1;
+
+	if (min <= 0 || max <= 0)
+		goto irq_disable;
+
+	if (drm_WARN_ON(&dev_priv->drm, drm_crtc_vblank_get(&crtc->base)))
+		goto irq_disable;
+
+	/*
+	 * Wait for psr to idle out after enabling the VBL interrupts
+	 * VBL interrupts will start the PSR exit and prevent a PSR
+	 * re-entry as well.
+	 */
+	intel_psr_wait_for_idle_locked(new_crtc_state);
+
+	local_irq_disable();
+
+	crtc->debug.min_vbl = min;
+	crtc->debug.max_vbl = max;
+	trace_intel_pipe_update_start(crtc);
+
+	for (;;) {
+		/*
+		 * prepare_to_wait() has a memory barrier, which guarantees
+		 * other CPUs can see the task state update by the time we
+		 * read the scanline.
+		 */
+		prepare_to_wait(wq, &wait, TASK_UNINTERRUPTIBLE);
+
+		scanline = intel_get_crtc_scanline(crtc);
+		if (scanline < min || scanline > max)
+			break;
+
+		if (!timeout) {
+			drm_err(&dev_priv->drm,
+				"Potential atomic update failure on pipe %c\n",
+				pipe_name(crtc->pipe));
+			break;
+		}
+
+		local_irq_enable();
+
+		timeout = schedule_timeout(timeout);
+
+		local_irq_disable();
+	}
+
+	finish_wait(wq, &wait);
+
+	drm_crtc_vblank_put(&crtc->base);
+
+	/*
+	 * On VLV/CHV DSI the scanline counter would appear to
+	 * increment approx. 1/3 of a scanline before start of vblank.
+	 * The registers still get latched at start of vblank however.
+	 * This means we must not write any registers on the first
+	 * line of vblank (since not the whole line is actually in
+	 * vblank). And unfortunately we can't use the interrupt to
+	 * wait here since it will fire too soon. We could use the
+	 * frame start interrupt instead since it will fire after the
+	 * critical scanline, but that would require more changes
+	 * in the interrupt code. So for now we'll just do the nasty
+	 * thing and poll for the bad scanline to pass us by.
+	 *
+	 * FIXME figure out if BXT+ DSI suffers from this as well
+	 */
+	while (need_vlv_dsi_wa && scanline == vblank_start)
+		scanline = intel_get_crtc_scanline(crtc);
+
+	crtc->debug.scanline_start = scanline;
+	crtc->debug.start_vbl_time = ktime_get();
+	crtc->debug.start_vbl_count = intel_crtc_get_vblank_counter(crtc);
+
+	trace_intel_pipe_update_vblank_evaded(crtc);
+	return;
+
+irq_disable:
+	local_irq_disable();
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_VBLANK_EVADE)
+static void dbg_vblank_evade(struct intel_crtc *crtc, ktime_t end)
+{
+	u64 delta = ktime_to_ns(ktime_sub(end, crtc->debug.start_vbl_time));
+	unsigned int h;
+
+	h = ilog2(delta >> 9);
+	if (h >= ARRAY_SIZE(crtc->debug.vbl.times))
+		h = ARRAY_SIZE(crtc->debug.vbl.times) - 1;
+	crtc->debug.vbl.times[h]++;
+
+	crtc->debug.vbl.sum += delta;
+	if (!crtc->debug.vbl.min || delta < crtc->debug.vbl.min)
+		crtc->debug.vbl.min = delta;
+	if (delta > crtc->debug.vbl.max)
+		crtc->debug.vbl.max = delta;
+
+	if (delta > 1000 * VBLANK_EVASION_TIME_US) {
+		drm_dbg_kms(crtc->base.dev,
+			    "Atomic update on pipe (%c) took %lld us, max time under evasion is %u us\n",
+			    pipe_name(crtc->pipe),
+			    div_u64(delta, 1000),
+			    VBLANK_EVASION_TIME_US);
+		crtc->debug.vbl.over++;
+	}
+}
+#else
+static void dbg_vblank_evade(struct intel_crtc *crtc, ktime_t end) {}
+#endif
+
+/**
+ * intel_pipe_update_end() - end update of a set of display registers
+ * @new_crtc_state: the new crtc state
+ *
+ * Mark the end of an update started with intel_pipe_update_start(). This
+ * re-enables interrupts and verifies the update was actually completed
+ * before a vblank.
+ */
+void intel_pipe_update_end(struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	enum pipe pipe = crtc->pipe;
+	int scanline_end = intel_get_crtc_scanline(crtc);
+	u32 end_vbl_count = intel_crtc_get_vblank_counter(crtc);
+	ktime_t end_vbl_time = ktime_get();
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	intel_psr_unlock(new_crtc_state);
+
+	if (new_crtc_state->do_async_flip)
+		return;
+
+	trace_intel_pipe_update_end(crtc, end_vbl_count, scanline_end);
+
+	/*
+	 * Incase of mipi dsi command mode, we need to set frame update
+	 * request for every commit.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 11 &&
+	    intel_crtc_has_type(new_crtc_state, INTEL_OUTPUT_DSI))
+		icl_dsi_frame_update(new_crtc_state);
+
+	/* We're still in the vblank-evade critical section, this can't race.
+	 * Would be slightly nice to just grab the vblank count and arm the
+	 * event outside of the critical section - the spinlock might spin for a
+	 * while ... */
+	if (intel_crtc_needs_vblank_work(new_crtc_state)) {
+		drm_vblank_work_schedule(&new_crtc_state->vblank_work,
+					 drm_crtc_accurate_vblank_count(&crtc->base) + 1,
+					 false);
+	} else if (new_crtc_state->uapi.event) {
+		drm_WARN_ON(&dev_priv->drm,
+			    drm_crtc_vblank_get(&crtc->base) != 0);
+
+		spin_lock(&crtc->base.dev->event_lock);
+		drm_crtc_arm_vblank_event(&crtc->base,
+					  new_crtc_state->uapi.event);
+		spin_unlock(&crtc->base.dev->event_lock);
+
+		new_crtc_state->uapi.event = NULL;
+	}
+
+	/*
+	 * Send VRR Push to terminate Vblank. If we are already in vblank
+	 * this has to be done _after_ sampling the frame counter, as
+	 * otherwise the push would immediately terminate the vblank and
+	 * the sampled frame counter would correspond to the next frame
+	 * instead of the current frame.
+	 *
+	 * There is a tiny race here (iff vblank evasion failed us) where
+	 * we might sample the frame counter just before vmax vblank start
+	 * but the push would be sent just after it. That would cause the
+	 * push to affect the next frame instead of the current frame,
+	 * which would cause the next frame to terminate already at vmin
+	 * vblank start instead of vmax vblank start.
+	 */
+	intel_vrr_send_push(new_crtc_state);
+
+	local_irq_enable();
+
+	if (intel_vgpu_active(dev_priv))
+		return;
+
+	if (crtc->debug.start_vbl_count &&
+	    crtc->debug.start_vbl_count != end_vbl_count) {
+		drm_err(&dev_priv->drm,
+			"Atomic update failure on pipe %c (start=%u end=%u) time %lld us, min %d, max %d, scanline start %d, end %d\n",
+			pipe_name(pipe), crtc->debug.start_vbl_count,
+			end_vbl_count,
+			ktime_us_delta(end_vbl_time,
+				       crtc->debug.start_vbl_time),
+			crtc->debug.min_vbl, crtc->debug.max_vbl,
+			crtc->debug.scanline_start, scanline_end);
+	}
+
+	dbg_vblank_evade(crtc, end_vbl_time);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_crtc.h b/drivers/gpu/drm/i915/display/intel_crtc.h
new file mode 100644
index 000000000..73077137f
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_crtc.h
@@ -0,0 +1,39 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _INTEL_CRTC_H_
+#define _INTEL_CRTC_H_
+
+#include <linux/types.h>
+
+enum i9xx_plane_id;
+enum pipe;
+struct drm_display_mode;
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+
+int intel_usecs_to_scanlines(const struct drm_display_mode *adjusted_mode,
+			     int usecs);
+u32 intel_crtc_max_vblank_count(const struct intel_crtc_state *crtc_state);
+int intel_crtc_init(struct drm_i915_private *dev_priv, enum pipe pipe);
+struct intel_crtc_state *intel_crtc_state_alloc(struct intel_crtc *crtc);
+void intel_crtc_state_reset(struct intel_crtc_state *crtc_state,
+			    struct intel_crtc *crtc);
+u32 intel_crtc_get_vblank_counter(struct intel_crtc *crtc);
+void intel_crtc_vblank_on(const struct intel_crtc_state *crtc_state);
+void intel_crtc_vblank_off(const struct intel_crtc_state *crtc_state);
+void intel_pipe_update_start(struct intel_crtc_state *new_crtc_state);
+void intel_pipe_update_end(struct intel_crtc_state *new_crtc_state);
+void intel_wait_for_vblank_workers(struct intel_atomic_state *state);
+struct intel_crtc *intel_first_crtc(struct drm_i915_private *i915);
+struct intel_crtc *intel_crtc_for_pipe(struct drm_i915_private *i915,
+				       enum pipe pipe);
+void intel_wait_for_vblank_if_active(struct drm_i915_private *i915,
+				     enum pipe pipe);
+void intel_crtc_wait_for_next_vblank(struct intel_crtc *crtc);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_crtc_state_dump.c b/drivers/gpu/drm/i915/display/intel_crtc_state_dump.c
new file mode 100644
index 000000000..e9212f69c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_crtc_state_dump.c
@@ -0,0 +1,315 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "intel_crtc_state_dump.h"
+#include "intel_display_types.h"
+#include "intel_hdmi.h"
+#include "intel_vrr.h"
+
+static void intel_dump_crtc_timings(struct drm_i915_private *i915,
+				    const struct drm_display_mode *mode)
+{
+	drm_dbg_kms(&i915->drm, "crtc timings: %d %d %d %d %d %d %d %d %d, "
+		    "type: 0x%x flags: 0x%x\n",
+		    mode->crtc_clock,
+		    mode->crtc_hdisplay, mode->crtc_hsync_start,
+		    mode->crtc_hsync_end, mode->crtc_htotal,
+		    mode->crtc_vdisplay, mode->crtc_vsync_start,
+		    mode->crtc_vsync_end, mode->crtc_vtotal,
+		    mode->type, mode->flags);
+}
+
+static void
+intel_dump_m_n_config(const struct intel_crtc_state *pipe_config,
+		      const char *id, unsigned int lane_count,
+		      const struct intel_link_m_n *m_n)
+{
+	struct drm_i915_private *i915 = to_i915(pipe_config->uapi.crtc->dev);
+
+	drm_dbg_kms(&i915->drm,
+		    "%s: lanes: %i; data_m: %u, data_n: %u, link_m: %u, link_n: %u, tu: %u\n",
+		    id, lane_count,
+		    m_n->data_m, m_n->data_n,
+		    m_n->link_m, m_n->link_n, m_n->tu);
+}
+
+static void
+intel_dump_infoframe(struct drm_i915_private *i915,
+		     const union hdmi_infoframe *frame)
+{
+	if (!drm_debug_enabled(DRM_UT_KMS))
+		return;
+
+	hdmi_infoframe_log(KERN_DEBUG, i915->drm.dev, frame);
+}
+
+static void
+intel_dump_dp_vsc_sdp(struct drm_i915_private *i915,
+		      const struct drm_dp_vsc_sdp *vsc)
+{
+	if (!drm_debug_enabled(DRM_UT_KMS))
+		return;
+
+	drm_dp_vsc_sdp_log(KERN_DEBUG, i915->drm.dev, vsc);
+}
+
+#define OUTPUT_TYPE(x) [INTEL_OUTPUT_ ## x] = #x
+
+static const char * const output_type_str[] = {
+	OUTPUT_TYPE(UNUSED),
+	OUTPUT_TYPE(ANALOG),
+	OUTPUT_TYPE(DVO),
+	OUTPUT_TYPE(SDVO),
+	OUTPUT_TYPE(LVDS),
+	OUTPUT_TYPE(TVOUT),
+	OUTPUT_TYPE(HDMI),
+	OUTPUT_TYPE(DP),
+	OUTPUT_TYPE(EDP),
+	OUTPUT_TYPE(DSI),
+	OUTPUT_TYPE(DDI),
+	OUTPUT_TYPE(DP_MST),
+};
+
+#undef OUTPUT_TYPE
+
+static void snprintf_output_types(char *buf, size_t len,
+				  unsigned int output_types)
+{
+	char *str = buf;
+	int i;
+
+	str[0] = '\0';
+
+	for (i = 0; i < ARRAY_SIZE(output_type_str); i++) {
+		int r;
+
+		if ((output_types & BIT(i)) == 0)
+			continue;
+
+		r = snprintf(str, len, "%s%s",
+			     str != buf ? "," : "", output_type_str[i]);
+		if (r >= len)
+			break;
+		str += r;
+		len -= r;
+
+		output_types &= ~BIT(i);
+	}
+
+	WARN_ON_ONCE(output_types != 0);
+}
+
+static const char * const output_format_str[] = {
+	[INTEL_OUTPUT_FORMAT_RGB] = "RGB",
+	[INTEL_OUTPUT_FORMAT_YCBCR420] = "YCBCR4:2:0",
+	[INTEL_OUTPUT_FORMAT_YCBCR444] = "YCBCR4:4:4",
+};
+
+static const char *output_formats(enum intel_output_format format)
+{
+	if (format >= ARRAY_SIZE(output_format_str))
+		return "invalid";
+	return output_format_str[format];
+}
+
+static void intel_dump_plane_state(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+
+	if (!fb) {
+		drm_dbg_kms(&i915->drm,
+			    "[PLANE:%d:%s] fb: [NOFB], visible: %s\n",
+			    plane->base.base.id, plane->base.name,
+			    str_yes_no(plane_state->uapi.visible));
+		return;
+	}
+
+	drm_dbg_kms(&i915->drm,
+		    "[PLANE:%d:%s] fb: [FB:%d] %ux%u format = %p4cc modifier = 0x%llx, visible: %s\n",
+		    plane->base.base.id, plane->base.name,
+		    fb->base.id, fb->width, fb->height, &fb->format->format,
+		    fb->modifier, str_yes_no(plane_state->uapi.visible));
+	drm_dbg_kms(&i915->drm, "\trotation: 0x%x, scaler: %d, scaling_filter: %d\n",
+		    plane_state->hw.rotation, plane_state->scaler_id, plane_state->hw.scaling_filter);
+	if (plane_state->uapi.visible)
+		drm_dbg_kms(&i915->drm,
+			    "\tsrc: " DRM_RECT_FP_FMT " dst: " DRM_RECT_FMT "\n",
+			    DRM_RECT_FP_ARG(&plane_state->uapi.src),
+			    DRM_RECT_ARG(&plane_state->uapi.dst));
+}
+
+void intel_crtc_state_dump(const struct intel_crtc_state *pipe_config,
+			   struct intel_atomic_state *state,
+			   const char *context)
+{
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	const struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	char buf[64];
+	int i;
+
+	drm_dbg_kms(&i915->drm, "[CRTC:%d:%s] enable: %s [%s]\n",
+		    crtc->base.base.id, crtc->base.name,
+		    str_yes_no(pipe_config->hw.enable), context);
+
+	if (!pipe_config->hw.enable)
+		goto dump_planes;
+
+	snprintf_output_types(buf, sizeof(buf), pipe_config->output_types);
+	drm_dbg_kms(&i915->drm,
+		    "active: %s, output_types: %s (0x%x), output format: %s\n",
+		    str_yes_no(pipe_config->hw.active),
+		    buf, pipe_config->output_types,
+		    output_formats(pipe_config->output_format));
+
+	drm_dbg_kms(&i915->drm,
+		    "cpu_transcoder: %s, pipe bpp: %i, dithering: %i\n",
+		    transcoder_name(pipe_config->cpu_transcoder),
+		    pipe_config->pipe_bpp, pipe_config->dither);
+
+	drm_dbg_kms(&i915->drm, "MST master transcoder: %s\n",
+		    transcoder_name(pipe_config->mst_master_transcoder));
+
+	drm_dbg_kms(&i915->drm,
+		    "port sync: master transcoder: %s, slave transcoder bitmask = 0x%x\n",
+		    transcoder_name(pipe_config->master_transcoder),
+		    pipe_config->sync_mode_slaves_mask);
+
+	drm_dbg_kms(&i915->drm, "bigjoiner: %s, pipes: 0x%x\n",
+		    intel_crtc_is_bigjoiner_slave(pipe_config) ? "slave" :
+		    intel_crtc_is_bigjoiner_master(pipe_config) ? "master" : "no",
+		    pipe_config->bigjoiner_pipes);
+
+	drm_dbg_kms(&i915->drm, "splitter: %s, link count %d, overlap %d\n",
+		    str_enabled_disabled(pipe_config->splitter.enable),
+		    pipe_config->splitter.link_count,
+		    pipe_config->splitter.pixel_overlap);
+
+	if (pipe_config->has_pch_encoder)
+		intel_dump_m_n_config(pipe_config, "fdi",
+				      pipe_config->fdi_lanes,
+				      &pipe_config->fdi_m_n);
+
+	if (intel_crtc_has_dp_encoder(pipe_config)) {
+		intel_dump_m_n_config(pipe_config, "dp m_n",
+				      pipe_config->lane_count,
+				      &pipe_config->dp_m_n);
+		intel_dump_m_n_config(pipe_config, "dp m2_n2",
+				      pipe_config->lane_count,
+				      &pipe_config->dp_m2_n2);
+	}
+
+	drm_dbg_kms(&i915->drm, "framestart delay: %d, MSA timing delay: %d\n",
+		    pipe_config->framestart_delay, pipe_config->msa_timing_delay);
+
+	drm_dbg_kms(&i915->drm,
+		    "audio: %i, infoframes: %i, infoframes enabled: 0x%x\n",
+		    pipe_config->has_audio, pipe_config->has_infoframe,
+		    pipe_config->infoframes.enable);
+
+	if (pipe_config->infoframes.enable &
+	    intel_hdmi_infoframe_enable(HDMI_PACKET_TYPE_GENERAL_CONTROL))
+		drm_dbg_kms(&i915->drm, "GCP: 0x%x\n",
+			    pipe_config->infoframes.gcp);
+	if (pipe_config->infoframes.enable &
+	    intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_AVI))
+		intel_dump_infoframe(i915, &pipe_config->infoframes.avi);
+	if (pipe_config->infoframes.enable &
+	    intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_SPD))
+		intel_dump_infoframe(i915, &pipe_config->infoframes.spd);
+	if (pipe_config->infoframes.enable &
+	    intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_VENDOR))
+		intel_dump_infoframe(i915, &pipe_config->infoframes.hdmi);
+	if (pipe_config->infoframes.enable &
+	    intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_DRM))
+		intel_dump_infoframe(i915, &pipe_config->infoframes.drm);
+	if (pipe_config->infoframes.enable &
+	    intel_hdmi_infoframe_enable(HDMI_PACKET_TYPE_GAMUT_METADATA))
+		intel_dump_infoframe(i915, &pipe_config->infoframes.drm);
+	if (pipe_config->infoframes.enable &
+	    intel_hdmi_infoframe_enable(DP_SDP_VSC))
+		intel_dump_dp_vsc_sdp(i915, &pipe_config->infoframes.vsc);
+
+	drm_dbg_kms(&i915->drm, "vrr: %s, vmin: %d, vmax: %d, pipeline full: %d, guardband: %d flipline: %d, vmin vblank: %d, vmax vblank: %d\n",
+		    str_yes_no(pipe_config->vrr.enable),
+		    pipe_config->vrr.vmin, pipe_config->vrr.vmax,
+		    pipe_config->vrr.pipeline_full, pipe_config->vrr.guardband,
+		    pipe_config->vrr.flipline,
+		    intel_vrr_vmin_vblank_start(pipe_config),
+		    intel_vrr_vmax_vblank_start(pipe_config));
+
+	drm_dbg_kms(&i915->drm, "requested mode: " DRM_MODE_FMT "\n",
+		    DRM_MODE_ARG(&pipe_config->hw.mode));
+	drm_dbg_kms(&i915->drm, "adjusted mode: " DRM_MODE_FMT "\n",
+		    DRM_MODE_ARG(&pipe_config->hw.adjusted_mode));
+	intel_dump_crtc_timings(i915, &pipe_config->hw.adjusted_mode);
+	drm_dbg_kms(&i915->drm, "pipe mode: " DRM_MODE_FMT "\n",
+		    DRM_MODE_ARG(&pipe_config->hw.pipe_mode));
+	intel_dump_crtc_timings(i915, &pipe_config->hw.pipe_mode);
+	drm_dbg_kms(&i915->drm,
+		    "port clock: %d, pipe src: " DRM_RECT_FMT ", pixel rate %d\n",
+		    pipe_config->port_clock, DRM_RECT_ARG(&pipe_config->pipe_src),
+		    pipe_config->pixel_rate);
+
+	drm_dbg_kms(&i915->drm, "linetime: %d, ips linetime: %d\n",
+		    pipe_config->linetime, pipe_config->ips_linetime);
+
+	if (DISPLAY_VER(i915) >= 9)
+		drm_dbg_kms(&i915->drm,
+			    "num_scalers: %d, scaler_users: 0x%x, scaler_id: %d, scaling_filter: %d\n",
+			    crtc->num_scalers,
+			    pipe_config->scaler_state.scaler_users,
+			    pipe_config->scaler_state.scaler_id,
+			    pipe_config->hw.scaling_filter);
+
+	if (HAS_GMCH(i915))
+		drm_dbg_kms(&i915->drm,
+			    "gmch pfit: control: 0x%08x, ratios: 0x%08x, lvds border: 0x%08x\n",
+			    pipe_config->gmch_pfit.control,
+			    pipe_config->gmch_pfit.pgm_ratios,
+			    pipe_config->gmch_pfit.lvds_border_bits);
+	else
+		drm_dbg_kms(&i915->drm,
+			    "pch pfit: " DRM_RECT_FMT ", %s, force thru: %s\n",
+			    DRM_RECT_ARG(&pipe_config->pch_pfit.dst),
+			    str_enabled_disabled(pipe_config->pch_pfit.enabled),
+			    str_yes_no(pipe_config->pch_pfit.force_thru));
+
+	drm_dbg_kms(&i915->drm, "ips: %i, double wide: %i, drrs: %i\n",
+		    pipe_config->ips_enabled, pipe_config->double_wide,
+		    pipe_config->has_drrs);
+
+	intel_dpll_dump_hw_state(i915, &pipe_config->dpll_hw_state);
+
+	if (IS_CHERRYVIEW(i915))
+		drm_dbg_kms(&i915->drm,
+			    "cgm_mode: 0x%x gamma_mode: 0x%x gamma_enable: %d csc_enable: %d\n",
+			    pipe_config->cgm_mode, pipe_config->gamma_mode,
+			    pipe_config->gamma_enable, pipe_config->csc_enable);
+	else
+		drm_dbg_kms(&i915->drm,
+			    "csc_mode: 0x%x gamma_mode: 0x%x gamma_enable: %d csc_enable: %d\n",
+			    pipe_config->csc_mode, pipe_config->gamma_mode,
+			    pipe_config->gamma_enable, pipe_config->csc_enable);
+
+	drm_dbg_kms(&i915->drm, "degamma lut: %d entries, gamma lut: %d entries\n",
+		    pipe_config->hw.degamma_lut ?
+		    drm_color_lut_size(pipe_config->hw.degamma_lut) : 0,
+		    pipe_config->hw.gamma_lut ?
+		    drm_color_lut_size(pipe_config->hw.gamma_lut) : 0);
+
+dump_planes:
+	if (!state)
+		return;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		if (plane->pipe == crtc->pipe)
+			intel_dump_plane_state(plane_state);
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_crtc_state_dump.h b/drivers/gpu/drm/i915/display/intel_crtc_state_dump.h
new file mode 100644
index 000000000..9399c35b7
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_crtc_state_dump.h
@@ -0,0 +1,16 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_CRTC_STATE_DUMP_H__
+#define __INTEL_CRTC_STATE_DUMP_H__
+
+struct intel_crtc_state;
+struct intel_atomic_state;
+
+void intel_crtc_state_dump(const struct intel_crtc_state *crtc_state,
+			   struct intel_atomic_state *state,
+			   const char *context);
+
+#endif /* __INTEL_CRTC_STATE_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_cursor.c b/drivers/gpu/drm/i915/display/intel_cursor.c
new file mode 100644
index 000000000..87899e89b
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_cursor.c
@@ -0,0 +1,828 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+#include <linux/kernel.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_atomic_uapi.h>
+#include <drm/drm_blend.h>
+#include <drm/drm_damage_helper.h>
+#include <drm/drm_fourcc.h>
+
+#include "intel_atomic.h"
+#include "intel_atomic_plane.h"
+#include "intel_cursor.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_display.h"
+#include "intel_fb.h"
+#include "intel_fb_pin.h"
+#include "intel_frontbuffer.h"
+#include "intel_psr.h"
+#include "intel_sprite.h"
+#include "skl_watermark.h"
+
+/* Cursor formats */
+static const u32 intel_cursor_formats[] = {
+	DRM_FORMAT_ARGB8888,
+};
+
+static u32 intel_cursor_base(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	const struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	u32 base;
+
+	if (INTEL_INFO(dev_priv)->display.cursor_needs_physical)
+		base = sg_dma_address(obj->mm.pages->sgl);
+	else
+		base = intel_plane_ggtt_offset(plane_state);
+
+	return base + plane_state->view.color_plane[0].offset;
+}
+
+static u32 intel_cursor_position(const struct intel_plane_state *plane_state)
+{
+	int x = plane_state->uapi.dst.x1;
+	int y = plane_state->uapi.dst.y1;
+	u32 pos = 0;
+
+	if (x < 0) {
+		pos |= CURSOR_POS_X_SIGN;
+		x = -x;
+	}
+	pos |= CURSOR_POS_X(x);
+
+	if (y < 0) {
+		pos |= CURSOR_POS_Y_SIGN;
+		y = -y;
+	}
+	pos |= CURSOR_POS_Y(y);
+
+	return pos;
+}
+
+static bool intel_cursor_size_ok(const struct intel_plane_state *plane_state)
+{
+	const struct drm_mode_config *config =
+		&plane_state->uapi.plane->dev->mode_config;
+	int width = drm_rect_width(&plane_state->uapi.dst);
+	int height = drm_rect_height(&plane_state->uapi.dst);
+
+	return width > 0 && width <= config->cursor_width &&
+		height > 0 && height <= config->cursor_height;
+}
+
+static int intel_cursor_check_surface(struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	unsigned int rotation = plane_state->hw.rotation;
+	int src_x, src_y;
+	u32 offset;
+	int ret;
+
+	ret = intel_plane_compute_gtt(plane_state);
+	if (ret)
+		return ret;
+
+	if (!plane_state->uapi.visible)
+		return 0;
+
+	src_x = plane_state->uapi.src.x1 >> 16;
+	src_y = plane_state->uapi.src.y1 >> 16;
+
+	intel_add_fb_offsets(&src_x, &src_y, plane_state, 0);
+	offset = intel_plane_compute_aligned_offset(&src_x, &src_y,
+						    plane_state, 0);
+
+	if (src_x != 0 || src_y != 0) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Arbitrary cursor panning not supported\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * Put the final coordinates back so that the src
+	 * coordinate checks will see the right values.
+	 */
+	drm_rect_translate_to(&plane_state->uapi.src,
+			      src_x << 16, src_y << 16);
+
+	/* ILK+ do this automagically in hardware */
+	if (HAS_GMCH(dev_priv) && rotation & DRM_MODE_ROTATE_180) {
+		const struct drm_framebuffer *fb = plane_state->hw.fb;
+		int src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+		int src_h = drm_rect_height(&plane_state->uapi.src) >> 16;
+
+		offset += (src_h * src_w - 1) * fb->format->cpp[0];
+	}
+
+	plane_state->view.color_plane[0].offset = offset;
+	plane_state->view.color_plane[0].x = src_x;
+	plane_state->view.color_plane[0].y = src_y;
+
+	return 0;
+}
+
+static int intel_check_cursor(struct intel_crtc_state *crtc_state,
+			      struct intel_plane_state *plane_state)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	const struct drm_rect src = plane_state->uapi.src;
+	const struct drm_rect dst = plane_state->uapi.dst;
+	int ret;
+
+	if (fb && fb->modifier != DRM_FORMAT_MOD_LINEAR) {
+		drm_dbg_kms(&i915->drm, "cursor cannot be tiled\n");
+		return -EINVAL;
+	}
+
+	ret = intel_atomic_plane_check_clipping(plane_state, crtc_state,
+						DRM_PLANE_NO_SCALING,
+						DRM_PLANE_NO_SCALING,
+						true);
+	if (ret)
+		return ret;
+
+	/* Use the unclipped src/dst rectangles, which we program to hw */
+	plane_state->uapi.src = src;
+	plane_state->uapi.dst = dst;
+
+	/* final plane coordinates will be relative to the plane's pipe */
+	drm_rect_translate(&plane_state->uapi.dst,
+			   -crtc_state->pipe_src.x1,
+			   -crtc_state->pipe_src.y1);
+
+	ret = intel_cursor_check_surface(plane_state);
+	if (ret)
+		return ret;
+
+	if (!plane_state->uapi.visible)
+		return 0;
+
+	ret = intel_plane_check_src_coordinates(plane_state);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static unsigned int
+i845_cursor_max_stride(struct intel_plane *plane,
+		       u32 pixel_format, u64 modifier,
+		       unsigned int rotation)
+{
+	return 2048;
+}
+
+static u32 i845_cursor_ctl_crtc(const struct intel_crtc_state *crtc_state)
+{
+	u32 cntl = 0;
+
+	if (crtc_state->gamma_enable)
+		cntl |= CURSOR_PIPE_GAMMA_ENABLE;
+
+	return cntl;
+}
+
+static u32 i845_cursor_ctl(const struct intel_crtc_state *crtc_state,
+			   const struct intel_plane_state *plane_state)
+{
+	return CURSOR_ENABLE |
+		CURSOR_FORMAT_ARGB |
+		CURSOR_STRIDE(plane_state->view.color_plane[0].mapping_stride);
+}
+
+static bool i845_cursor_size_ok(const struct intel_plane_state *plane_state)
+{
+	int width = drm_rect_width(&plane_state->uapi.dst);
+
+	/*
+	 * 845g/865g are only limited by the width of their cursors,
+	 * the height is arbitrary up to the precision of the register.
+	 */
+	return intel_cursor_size_ok(plane_state) && IS_ALIGNED(width, 64);
+}
+
+static int i845_check_cursor(struct intel_crtc_state *crtc_state,
+			     struct intel_plane_state *plane_state)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	int ret;
+
+	ret = intel_check_cursor(crtc_state, plane_state);
+	if (ret)
+		return ret;
+
+	/* if we want to turn off the cursor ignore width and height */
+	if (!fb)
+		return 0;
+
+	/* Check for which cursor types we support */
+	if (!i845_cursor_size_ok(plane_state)) {
+		drm_dbg_kms(&i915->drm,
+			    "Cursor dimension %dx%d not supported\n",
+			    drm_rect_width(&plane_state->uapi.dst),
+			    drm_rect_height(&plane_state->uapi.dst));
+		return -EINVAL;
+	}
+
+	drm_WARN_ON(&i915->drm, plane_state->uapi.visible &&
+		    plane_state->view.color_plane[0].mapping_stride != fb->pitches[0]);
+
+	switch (fb->pitches[0]) {
+	case 256:
+	case 512:
+	case 1024:
+	case 2048:
+		break;
+	default:
+		 drm_dbg_kms(&i915->drm, "Invalid cursor stride (%u)\n",
+			     fb->pitches[0]);
+		return -EINVAL;
+	}
+
+	plane_state->ctl = i845_cursor_ctl(crtc_state, plane_state);
+
+	return 0;
+}
+
+/* TODO: split into noarm+arm pair */
+static void i845_cursor_update_arm(struct intel_plane *plane,
+				   const struct intel_crtc_state *crtc_state,
+				   const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	u32 cntl = 0, base = 0, pos = 0, size = 0;
+
+	if (plane_state && plane_state->uapi.visible) {
+		unsigned int width = drm_rect_width(&plane_state->uapi.dst);
+		unsigned int height = drm_rect_height(&plane_state->uapi.dst);
+
+		cntl = plane_state->ctl |
+			i845_cursor_ctl_crtc(crtc_state);
+
+		size = CURSOR_HEIGHT(height) | CURSOR_WIDTH(width);
+
+		base = intel_cursor_base(plane_state);
+		pos = intel_cursor_position(plane_state);
+	}
+
+	/* On these chipsets we can only modify the base/size/stride
+	 * whilst the cursor is disabled.
+	 */
+	if (plane->cursor.base != base ||
+	    plane->cursor.size != size ||
+	    plane->cursor.cntl != cntl) {
+		intel_de_write_fw(dev_priv, CURCNTR(PIPE_A), 0);
+		intel_de_write_fw(dev_priv, CURBASE(PIPE_A), base);
+		intel_de_write_fw(dev_priv, CURSIZE(PIPE_A), size);
+		intel_de_write_fw(dev_priv, CURPOS(PIPE_A), pos);
+		intel_de_write_fw(dev_priv, CURCNTR(PIPE_A), cntl);
+
+		plane->cursor.base = base;
+		plane->cursor.size = size;
+		plane->cursor.cntl = cntl;
+	} else {
+		intel_de_write_fw(dev_priv, CURPOS(PIPE_A), pos);
+	}
+}
+
+static void i845_cursor_disable_arm(struct intel_plane *plane,
+				    const struct intel_crtc_state *crtc_state)
+{
+	i845_cursor_update_arm(plane, crtc_state, NULL);
+}
+
+static bool i845_cursor_get_hw_state(struct intel_plane *plane,
+				     enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum intel_display_power_domain power_domain;
+	intel_wakeref_t wakeref;
+	bool ret;
+
+	power_domain = POWER_DOMAIN_PIPE(PIPE_A);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref)
+		return false;
+
+	ret = intel_de_read(dev_priv, CURCNTR(PIPE_A)) & CURSOR_ENABLE;
+
+	*pipe = PIPE_A;
+
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+
+	return ret;
+}
+
+static unsigned int
+i9xx_cursor_max_stride(struct intel_plane *plane,
+		       u32 pixel_format, u64 modifier,
+		       unsigned int rotation)
+{
+	return plane->base.dev->mode_config.cursor_width * 4;
+}
+
+static u32 i9xx_cursor_ctl_crtc(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 cntl = 0;
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		return cntl;
+
+	if (crtc_state->gamma_enable)
+		cntl = MCURSOR_PIPE_GAMMA_ENABLE;
+
+	if (crtc_state->csc_enable)
+		cntl |= MCURSOR_PIPE_CSC_ENABLE;
+
+	if (DISPLAY_VER(dev_priv) < 5 && !IS_G4X(dev_priv))
+		cntl |= MCURSOR_PIPE_SEL(crtc->pipe);
+
+	return cntl;
+}
+
+static u32 i9xx_cursor_ctl(const struct intel_crtc_state *crtc_state,
+			   const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	u32 cntl = 0;
+
+	if (IS_SANDYBRIDGE(dev_priv) || IS_IVYBRIDGE(dev_priv))
+		cntl |= MCURSOR_TRICKLE_FEED_DISABLE;
+
+	switch (drm_rect_width(&plane_state->uapi.dst)) {
+	case 64:
+		cntl |= MCURSOR_MODE_64_ARGB_AX;
+		break;
+	case 128:
+		cntl |= MCURSOR_MODE_128_ARGB_AX;
+		break;
+	case 256:
+		cntl |= MCURSOR_MODE_256_ARGB_AX;
+		break;
+	default:
+		MISSING_CASE(drm_rect_width(&plane_state->uapi.dst));
+		return 0;
+	}
+
+	if (plane_state->hw.rotation & DRM_MODE_ROTATE_180)
+		cntl |= MCURSOR_ROTATE_180;
+
+	/* Wa_22012358565:adl-p */
+	if (DISPLAY_VER(dev_priv) == 13)
+		cntl |= MCURSOR_ARB_SLOTS(1);
+
+	return cntl;
+}
+
+static bool i9xx_cursor_size_ok(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	int width = drm_rect_width(&plane_state->uapi.dst);
+	int height = drm_rect_height(&plane_state->uapi.dst);
+
+	if (!intel_cursor_size_ok(plane_state))
+		return false;
+
+	/* Cursor width is limited to a few power-of-two sizes */
+	switch (width) {
+	case 256:
+	case 128:
+	case 64:
+		break;
+	default:
+		return false;
+	}
+
+	/*
+	 * IVB+ have CUR_FBC_CTL which allows an arbitrary cursor
+	 * height from 8 lines up to the cursor width, when the
+	 * cursor is not rotated. Everything else requires square
+	 * cursors.
+	 */
+	if (HAS_CUR_FBC(dev_priv) &&
+	    plane_state->hw.rotation & DRM_MODE_ROTATE_0) {
+		if (height < 8 || height > width)
+			return false;
+	} else {
+		if (height != width)
+			return false;
+	}
+
+	return true;
+}
+
+static int i9xx_check_cursor(struct intel_crtc_state *crtc_state,
+			     struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	enum pipe pipe = plane->pipe;
+	int ret;
+
+	ret = intel_check_cursor(crtc_state, plane_state);
+	if (ret)
+		return ret;
+
+	/* if we want to turn off the cursor ignore width and height */
+	if (!fb)
+		return 0;
+
+	/* Check for which cursor types we support */
+	if (!i9xx_cursor_size_ok(plane_state)) {
+		drm_dbg(&dev_priv->drm,
+			"Cursor dimension %dx%d not supported\n",
+			drm_rect_width(&plane_state->uapi.dst),
+			drm_rect_height(&plane_state->uapi.dst));
+		return -EINVAL;
+	}
+
+	drm_WARN_ON(&dev_priv->drm, plane_state->uapi.visible &&
+		    plane_state->view.color_plane[0].mapping_stride != fb->pitches[0]);
+
+	if (fb->pitches[0] !=
+	    drm_rect_width(&plane_state->uapi.dst) * fb->format->cpp[0]) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Invalid cursor stride (%u) (cursor width %d)\n",
+			    fb->pitches[0],
+			    drm_rect_width(&plane_state->uapi.dst));
+		return -EINVAL;
+	}
+
+	/*
+	 * There's something wrong with the cursor on CHV pipe C.
+	 * If it straddles the left edge of the screen then
+	 * moving it away from the edge or disabling it often
+	 * results in a pipe underrun, and often that can lead to
+	 * dead pipe (constant underrun reported, and it scans
+	 * out just a solid color). To recover from that, the
+	 * display power well must be turned off and on again.
+	 * Refuse the put the cursor into that compromised position.
+	 */
+	if (IS_CHERRYVIEW(dev_priv) && pipe == PIPE_C &&
+	    plane_state->uapi.visible && plane_state->uapi.dst.x1 < 0) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "CHV cursor C not allowed to straddle the left screen edge\n");
+		return -EINVAL;
+	}
+
+	plane_state->ctl = i9xx_cursor_ctl(crtc_state, plane_state);
+
+	return 0;
+}
+
+/* TODO: split into noarm+arm pair */
+static void i9xx_cursor_update_arm(struct intel_plane *plane,
+				   const struct intel_crtc_state *crtc_state,
+				   const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	u32 cntl = 0, base = 0, pos = 0, fbc_ctl = 0;
+
+	if (plane_state && plane_state->uapi.visible) {
+		int width = drm_rect_width(&plane_state->uapi.dst);
+		int height = drm_rect_height(&plane_state->uapi.dst);
+
+		cntl = plane_state->ctl |
+			i9xx_cursor_ctl_crtc(crtc_state);
+
+		if (width != height)
+			fbc_ctl = CUR_FBC_EN | CUR_FBC_HEIGHT(height - 1);
+
+		base = intel_cursor_base(plane_state);
+		pos = intel_cursor_position(plane_state);
+	}
+
+	/*
+	 * On some platforms writing CURCNTR first will also
+	 * cause CURPOS to be armed by the CURBASE write.
+	 * Without the CURCNTR write the CURPOS write would
+	 * arm itself. Thus we always update CURCNTR before
+	 * CURPOS.
+	 *
+	 * On other platforms CURPOS always requires the
+	 * CURBASE write to arm the update. Additonally
+	 * a write to any of the cursor register will cancel
+	 * an already armed cursor update. Thus leaving out
+	 * the CURBASE write after CURPOS could lead to a
+	 * cursor that doesn't appear to move, or even change
+	 * shape. Thus we always write CURBASE.
+	 *
+	 * The other registers are armed by the CURBASE write
+	 * except when the plane is getting enabled at which time
+	 * the CURCNTR write arms the update.
+	 */
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		skl_write_cursor_wm(plane, crtc_state);
+
+	if (plane_state)
+		intel_psr2_program_plane_sel_fetch(plane, crtc_state, plane_state, 0);
+	else
+		intel_psr2_disable_plane_sel_fetch(plane, crtc_state);
+
+	if (plane->cursor.base != base ||
+	    plane->cursor.size != fbc_ctl ||
+	    plane->cursor.cntl != cntl) {
+		if (HAS_CUR_FBC(dev_priv))
+			intel_de_write_fw(dev_priv, CUR_FBC_CTL(pipe),
+					  fbc_ctl);
+		intel_de_write_fw(dev_priv, CURCNTR(pipe), cntl);
+		intel_de_write_fw(dev_priv, CURPOS(pipe), pos);
+		intel_de_write_fw(dev_priv, CURBASE(pipe), base);
+
+		plane->cursor.base = base;
+		plane->cursor.size = fbc_ctl;
+		plane->cursor.cntl = cntl;
+	} else {
+		intel_de_write_fw(dev_priv, CURPOS(pipe), pos);
+		intel_de_write_fw(dev_priv, CURBASE(pipe), base);
+	}
+}
+
+static void i9xx_cursor_disable_arm(struct intel_plane *plane,
+				    const struct intel_crtc_state *crtc_state)
+{
+	i9xx_cursor_update_arm(plane, crtc_state, NULL);
+}
+
+static bool i9xx_cursor_get_hw_state(struct intel_plane *plane,
+				     enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum intel_display_power_domain power_domain;
+	intel_wakeref_t wakeref;
+	bool ret;
+	u32 val;
+
+	/*
+	 * Not 100% correct for planes that can move between pipes,
+	 * but that's only the case for gen2-3 which don't have any
+	 * display power wells.
+	 */
+	power_domain = POWER_DOMAIN_PIPE(plane->pipe);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref)
+		return false;
+
+	val = intel_de_read(dev_priv, CURCNTR(plane->pipe));
+
+	ret = val & MCURSOR_MODE_MASK;
+
+	if (DISPLAY_VER(dev_priv) >= 5 || IS_G4X(dev_priv))
+		*pipe = plane->pipe;
+	else
+		*pipe = REG_FIELD_GET(MCURSOR_PIPE_SEL_MASK, val);
+
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+
+	return ret;
+}
+
+static bool intel_cursor_format_mod_supported(struct drm_plane *_plane,
+					      u32 format, u64 modifier)
+{
+	if (!intel_fb_plane_supports_modifier(to_intel_plane(_plane), modifier))
+		return false;
+
+	return format == DRM_FORMAT_ARGB8888;
+}
+
+static int
+intel_legacy_cursor_update(struct drm_plane *_plane,
+			   struct drm_crtc *_crtc,
+			   struct drm_framebuffer *fb,
+			   int crtc_x, int crtc_y,
+			   unsigned int crtc_w, unsigned int crtc_h,
+			   u32 src_x, u32 src_y,
+			   u32 src_w, u32 src_h,
+			   struct drm_modeset_acquire_ctx *ctx)
+{
+	struct intel_plane *plane = to_intel_plane(_plane);
+	struct intel_crtc *crtc = to_intel_crtc(_crtc);
+	struct intel_plane_state *old_plane_state =
+		to_intel_plane_state(plane->base.state);
+	struct intel_plane_state *new_plane_state;
+	struct intel_crtc_state *crtc_state =
+		to_intel_crtc_state(crtc->base.state);
+	struct intel_crtc_state *new_crtc_state;
+	int ret;
+
+	/*
+	 * When crtc is inactive or there is a modeset pending,
+	 * wait for it to complete in the slowpath.
+	 * PSR2 selective fetch also requires the slow path as
+	 * PSR2 plane and transcoder registers can only be updated during
+	 * vblank.
+	 *
+	 * FIXME bigjoiner fastpath would be good
+	 */
+	if (!crtc_state->hw.active || intel_crtc_needs_modeset(crtc_state) ||
+	    crtc_state->update_pipe || crtc_state->bigjoiner_pipes)
+		goto slow;
+
+	/*
+	 * Don't do an async update if there is an outstanding commit modifying
+	 * the plane.  This prevents our async update's changes from getting
+	 * overridden by a previous synchronous update's state.
+	 */
+	if (old_plane_state->uapi.commit &&
+	    !try_wait_for_completion(&old_plane_state->uapi.commit->hw_done))
+		goto slow;
+
+	/*
+	 * If any parameters change that may affect watermarks,
+	 * take the slowpath. Only changing fb or position should be
+	 * in the fastpath.
+	 */
+	if (old_plane_state->uapi.crtc != &crtc->base ||
+	    old_plane_state->uapi.src_w != src_w ||
+	    old_plane_state->uapi.src_h != src_h ||
+	    old_plane_state->uapi.crtc_w != crtc_w ||
+	    old_plane_state->uapi.crtc_h != crtc_h ||
+	    !old_plane_state->uapi.fb != !fb)
+		goto slow;
+
+	new_plane_state = to_intel_plane_state(intel_plane_duplicate_state(&plane->base));
+	if (!new_plane_state)
+		return -ENOMEM;
+
+	new_crtc_state = to_intel_crtc_state(intel_crtc_duplicate_state(&crtc->base));
+	if (!new_crtc_state) {
+		ret = -ENOMEM;
+		goto out_free;
+	}
+
+	drm_atomic_set_fb_for_plane(&new_plane_state->uapi, fb);
+
+	new_plane_state->uapi.src_x = src_x;
+	new_plane_state->uapi.src_y = src_y;
+	new_plane_state->uapi.src_w = src_w;
+	new_plane_state->uapi.src_h = src_h;
+	new_plane_state->uapi.crtc_x = crtc_x;
+	new_plane_state->uapi.crtc_y = crtc_y;
+	new_plane_state->uapi.crtc_w = crtc_w;
+	new_plane_state->uapi.crtc_h = crtc_h;
+
+	intel_plane_copy_uapi_to_hw_state(new_plane_state, new_plane_state, crtc);
+
+	ret = intel_plane_atomic_check_with_state(crtc_state, new_crtc_state,
+						  old_plane_state, new_plane_state);
+	if (ret)
+		goto out_free;
+
+	ret = intel_plane_pin_fb(new_plane_state);
+	if (ret)
+		goto out_free;
+
+	intel_frontbuffer_flush(to_intel_frontbuffer(new_plane_state->hw.fb),
+				ORIGIN_CURSOR_UPDATE);
+	intel_frontbuffer_track(to_intel_frontbuffer(old_plane_state->hw.fb),
+				to_intel_frontbuffer(new_plane_state->hw.fb),
+				plane->frontbuffer_bit);
+
+	/* Swap plane state */
+	plane->base.state = &new_plane_state->uapi;
+
+	/*
+	 * We cannot swap crtc_state as it may be in use by an atomic commit or
+	 * page flip that's running simultaneously. If we swap crtc_state and
+	 * destroy the old state, we will cause a use-after-free there.
+	 *
+	 * Only update active_planes, which is needed for our internal
+	 * bookkeeping. Either value will do the right thing when updating
+	 * planes atomically. If the cursor was part of the atomic update then
+	 * we would have taken the slowpath.
+	 */
+	crtc_state->active_planes = new_crtc_state->active_planes;
+
+	/*
+	 * Technically we should do a vblank evasion here to make
+	 * sure all the cursor registers update on the same frame.
+	 * For now just make sure the register writes happen as
+	 * quickly as possible to minimize the race window.
+	 */
+	local_irq_disable();
+
+	if (new_plane_state->uapi.visible) {
+		intel_plane_update_noarm(plane, crtc_state, new_plane_state);
+		intel_plane_update_arm(plane, crtc_state, new_plane_state);
+	} else {
+		intel_plane_disable_arm(plane, crtc_state);
+	}
+
+	local_irq_enable();
+
+	intel_plane_unpin_fb(old_plane_state);
+
+out_free:
+	if (new_crtc_state)
+		intel_crtc_destroy_state(&crtc->base, &new_crtc_state->uapi);
+	if (ret)
+		intel_plane_destroy_state(&plane->base, &new_plane_state->uapi);
+	else
+		intel_plane_destroy_state(&plane->base, &old_plane_state->uapi);
+	return ret;
+
+slow:
+	return drm_atomic_helper_update_plane(&plane->base, &crtc->base, fb,
+					      crtc_x, crtc_y, crtc_w, crtc_h,
+					      src_x, src_y, src_w, src_h, ctx);
+}
+
+static const struct drm_plane_funcs intel_cursor_plane_funcs = {
+	.update_plane = intel_legacy_cursor_update,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = intel_plane_destroy,
+	.atomic_duplicate_state = intel_plane_duplicate_state,
+	.atomic_destroy_state = intel_plane_destroy_state,
+	.format_mod_supported = intel_cursor_format_mod_supported,
+};
+
+struct intel_plane *
+intel_cursor_plane_create(struct drm_i915_private *dev_priv,
+			  enum pipe pipe)
+{
+	struct intel_plane *cursor;
+	int ret, zpos;
+	u64 *modifiers;
+
+	cursor = intel_plane_alloc();
+	if (IS_ERR(cursor))
+		return cursor;
+
+	cursor->pipe = pipe;
+	cursor->i9xx_plane = (enum i9xx_plane_id) pipe;
+	cursor->id = PLANE_CURSOR;
+	cursor->frontbuffer_bit = INTEL_FRONTBUFFER(pipe, cursor->id);
+
+	if (IS_I845G(dev_priv) || IS_I865G(dev_priv)) {
+		cursor->max_stride = i845_cursor_max_stride;
+		cursor->update_arm = i845_cursor_update_arm;
+		cursor->disable_arm = i845_cursor_disable_arm;
+		cursor->get_hw_state = i845_cursor_get_hw_state;
+		cursor->check_plane = i845_check_cursor;
+	} else {
+		cursor->max_stride = i9xx_cursor_max_stride;
+		cursor->update_arm = i9xx_cursor_update_arm;
+		cursor->disable_arm = i9xx_cursor_disable_arm;
+		cursor->get_hw_state = i9xx_cursor_get_hw_state;
+		cursor->check_plane = i9xx_check_cursor;
+	}
+
+	cursor->cursor.base = ~0;
+	cursor->cursor.cntl = ~0;
+
+	if (IS_I845G(dev_priv) || IS_I865G(dev_priv) || HAS_CUR_FBC(dev_priv))
+		cursor->cursor.size = ~0;
+
+	modifiers = intel_fb_plane_get_modifiers(dev_priv, INTEL_PLANE_CAP_NONE);
+
+	ret = drm_universal_plane_init(&dev_priv->drm, &cursor->base,
+				       0, &intel_cursor_plane_funcs,
+				       intel_cursor_formats,
+				       ARRAY_SIZE(intel_cursor_formats),
+				       modifiers,
+				       DRM_PLANE_TYPE_CURSOR,
+				       "cursor %c", pipe_name(pipe));
+
+	kfree(modifiers);
+
+	if (ret)
+		goto fail;
+
+	if (DISPLAY_VER(dev_priv) >= 4)
+		drm_plane_create_rotation_property(&cursor->base,
+						   DRM_MODE_ROTATE_0,
+						   DRM_MODE_ROTATE_0 |
+						   DRM_MODE_ROTATE_180);
+
+	zpos = RUNTIME_INFO(dev_priv)->num_sprites[pipe] + 1;
+	drm_plane_create_zpos_immutable_property(&cursor->base, zpos);
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		drm_plane_enable_fb_damage_clips(&cursor->base);
+
+	intel_plane_helper_add(cursor);
+
+	return cursor;
+
+fail:
+	intel_plane_free(cursor);
+
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_cursor.h b/drivers/gpu/drm/i915/display/intel_cursor.h
new file mode 100644
index 000000000..ce333bf4c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_cursor.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _INTEL_CURSOR_H_
+#define _INTEL_CURSOR_H_
+
+enum pipe;
+struct drm_i915_private;
+struct intel_plane;
+
+struct intel_plane *
+intel_cursor_plane_create(struct drm_i915_private *dev_priv,
+			  enum pipe pipe);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c
new file mode 100644
index 000000000..706e2d956
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_ddi.c
@@ -0,0 +1,4525 @@
+/*
+ * Copyright © 2012 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eugeni Dodonov <eugeni.dodonov@intel.com>
+ *
+ */
+
+#include <linux/string_helpers.h>
+
+#include <drm/display/drm_scdc_helper.h>
+#include <drm/drm_privacy_screen_consumer.h>
+
+#include "i915_drv.h"
+#include "intel_audio.h"
+#include "intel_audio_regs.h"
+#include "intel_backlight.h"
+#include "intel_combo_phy.h"
+#include "intel_combo_phy_regs.h"
+#include "intel_connector.h"
+#include "intel_crtc.h"
+#include "intel_ddi.h"
+#include "intel_ddi_buf_trans.h"
+#include "intel_de.h"
+#include "intel_display_power.h"
+#include "intel_display_types.h"
+#include "intel_dkl_phy.h"
+#include "intel_dp.h"
+#include "intel_dp_link_training.h"
+#include "intel_dp_mst.h"
+#include "intel_dpio_phy.h"
+#include "intel_dsi.h"
+#include "intel_fdi.h"
+#include "intel_fifo_underrun.h"
+#include "intel_gmbus.h"
+#include "intel_hdcp.h"
+#include "intel_hdmi.h"
+#include "intel_hotplug.h"
+#include "intel_lspcon.h"
+#include "intel_pps.h"
+#include "intel_psr.h"
+#include "intel_quirks.h"
+#include "intel_snps_phy.h"
+#include "intel_sprite.h"
+#include "intel_tc.h"
+#include "intel_tc_phy_regs.h"
+#include "intel_vdsc.h"
+#include "intel_vrr.h"
+#include "skl_scaler.h"
+#include "skl_universal_plane.h"
+
+static const u8 index_to_dp_signal_levels[] = {
+	[0] = DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0,
+	[1] = DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1,
+	[2] = DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2,
+	[3] = DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_3,
+	[4] = DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0,
+	[5] = DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1,
+	[6] = DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_2,
+	[7] = DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0,
+	[8] = DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1,
+	[9] = DP_TRAIN_VOLTAGE_SWING_LEVEL_3 | DP_TRAIN_PRE_EMPH_LEVEL_0,
+};
+
+static int intel_ddi_hdmi_level(struct intel_encoder *encoder,
+				const struct intel_ddi_buf_trans *trans)
+{
+	int level;
+
+	level = intel_bios_hdmi_level_shift(encoder);
+	if (level < 0)
+		level = trans->hdmi_default_entry;
+
+	return level;
+}
+
+static bool has_buf_trans_select(struct drm_i915_private *i915)
+{
+	return DISPLAY_VER(i915) < 10 && !IS_BROXTON(i915);
+}
+
+static bool has_iboost(struct drm_i915_private *i915)
+{
+	return DISPLAY_VER(i915) == 9 && !IS_BROXTON(i915);
+}
+
+/*
+ * Starting with Haswell, DDI port buffers must be programmed with correct
+ * values in advance. This function programs the correct values for
+ * DP/eDP/FDI use cases.
+ */
+void hsw_prepare_dp_ddi_buffers(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 iboost_bit = 0;
+	int i, n_entries;
+	enum port port = encoder->port;
+	const struct intel_ddi_buf_trans *trans;
+
+	trans = encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !trans))
+		return;
+
+	/* If we're boosting the current, set bit 31 of trans1 */
+	if (has_iboost(dev_priv) &&
+	    intel_bios_encoder_dp_boost_level(encoder->devdata))
+		iboost_bit = DDI_BUF_BALANCE_LEG_ENABLE;
+
+	for (i = 0; i < n_entries; i++) {
+		intel_de_write(dev_priv, DDI_BUF_TRANS_LO(port, i),
+			       trans->entries[i].hsw.trans1 | iboost_bit);
+		intel_de_write(dev_priv, DDI_BUF_TRANS_HI(port, i),
+			       trans->entries[i].hsw.trans2);
+	}
+}
+
+/*
+ * Starting with Haswell, DDI port buffers must be programmed with correct
+ * values in advance. This function programs the correct values for
+ * HDMI/DVI use cases.
+ */
+static void hsw_prepare_hdmi_ddi_buffers(struct intel_encoder *encoder,
+					 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	int level = intel_ddi_level(encoder, crtc_state, 0);
+	u32 iboost_bit = 0;
+	int n_entries;
+	enum port port = encoder->port;
+	const struct intel_ddi_buf_trans *trans;
+
+	trans = encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !trans))
+		return;
+
+	/* If we're boosting the current, set bit 31 of trans1 */
+	if (has_iboost(dev_priv) &&
+	    intel_bios_encoder_hdmi_boost_level(encoder->devdata))
+		iboost_bit = DDI_BUF_BALANCE_LEG_ENABLE;
+
+	/* Entry 9 is for HDMI: */
+	intel_de_write(dev_priv, DDI_BUF_TRANS_LO(port, 9),
+		       trans->entries[level].hsw.trans1 | iboost_bit);
+	intel_de_write(dev_priv, DDI_BUF_TRANS_HI(port, 9),
+		       trans->entries[level].hsw.trans2);
+}
+
+void intel_wait_ddi_buf_idle(struct drm_i915_private *dev_priv,
+			     enum port port)
+{
+	if (IS_BROXTON(dev_priv)) {
+		udelay(16);
+		return;
+	}
+
+	if (wait_for_us((intel_de_read(dev_priv, DDI_BUF_CTL(port)) &
+			 DDI_BUF_IS_IDLE), 8))
+		drm_err(&dev_priv->drm, "Timeout waiting for DDI BUF %c to get idle\n",
+			port_name(port));
+}
+
+static void intel_wait_ddi_buf_active(struct drm_i915_private *dev_priv,
+				      enum port port)
+{
+	int ret;
+
+	/* Wait > 518 usecs for DDI_BUF_CTL to be non idle */
+	if (DISPLAY_VER(dev_priv) < 10) {
+		usleep_range(518, 1000);
+		return;
+	}
+
+	ret = _wait_for(!(intel_de_read(dev_priv, DDI_BUF_CTL(port)) &
+			  DDI_BUF_IS_IDLE), IS_DG2(dev_priv) ? 1200 : 500, 10, 10);
+
+	if (ret)
+		drm_err(&dev_priv->drm, "Timeout waiting for DDI BUF %c to get active\n",
+			port_name(port));
+}
+
+static u32 hsw_pll_to_ddi_pll_sel(const struct intel_shared_dpll *pll)
+{
+	switch (pll->info->id) {
+	case DPLL_ID_WRPLL1:
+		return PORT_CLK_SEL_WRPLL1;
+	case DPLL_ID_WRPLL2:
+		return PORT_CLK_SEL_WRPLL2;
+	case DPLL_ID_SPLL:
+		return PORT_CLK_SEL_SPLL;
+	case DPLL_ID_LCPLL_810:
+		return PORT_CLK_SEL_LCPLL_810;
+	case DPLL_ID_LCPLL_1350:
+		return PORT_CLK_SEL_LCPLL_1350;
+	case DPLL_ID_LCPLL_2700:
+		return PORT_CLK_SEL_LCPLL_2700;
+	default:
+		MISSING_CASE(pll->info->id);
+		return PORT_CLK_SEL_NONE;
+	}
+}
+
+static u32 icl_pll_to_ddi_clk_sel(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state)
+{
+	const struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	int clock = crtc_state->port_clock;
+	const enum intel_dpll_id id = pll->info->id;
+
+	switch (id) {
+	default:
+		/*
+		 * DPLL_ID_ICL_DPLL0 and DPLL_ID_ICL_DPLL1 should not be used
+		 * here, so do warn if this get passed in
+		 */
+		MISSING_CASE(id);
+		return DDI_CLK_SEL_NONE;
+	case DPLL_ID_ICL_TBTPLL:
+		switch (clock) {
+		case 162000:
+			return DDI_CLK_SEL_TBT_162;
+		case 270000:
+			return DDI_CLK_SEL_TBT_270;
+		case 540000:
+			return DDI_CLK_SEL_TBT_540;
+		case 810000:
+			return DDI_CLK_SEL_TBT_810;
+		default:
+			MISSING_CASE(clock);
+			return DDI_CLK_SEL_NONE;
+		}
+	case DPLL_ID_ICL_MGPLL1:
+	case DPLL_ID_ICL_MGPLL2:
+	case DPLL_ID_ICL_MGPLL3:
+	case DPLL_ID_ICL_MGPLL4:
+	case DPLL_ID_TGL_MGPLL5:
+	case DPLL_ID_TGL_MGPLL6:
+		return DDI_CLK_SEL_MG;
+	}
+}
+
+static u32 ddi_buf_phy_link_rate(int port_clock)
+{
+	switch (port_clock) {
+	case 162000:
+		return DDI_BUF_PHY_LINK_RATE(0);
+	case 216000:
+		return DDI_BUF_PHY_LINK_RATE(4);
+	case 243000:
+		return DDI_BUF_PHY_LINK_RATE(5);
+	case 270000:
+		return DDI_BUF_PHY_LINK_RATE(1);
+	case 324000:
+		return DDI_BUF_PHY_LINK_RATE(6);
+	case 432000:
+		return DDI_BUF_PHY_LINK_RATE(7);
+	case 540000:
+		return DDI_BUF_PHY_LINK_RATE(2);
+	case 810000:
+		return DDI_BUF_PHY_LINK_RATE(3);
+	default:
+		MISSING_CASE(port_clock);
+		return DDI_BUF_PHY_LINK_RATE(0);
+	}
+}
+
+static void intel_ddi_init_dp_buf_reg(struct intel_encoder *encoder,
+				      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	/* DDI_BUF_CTL_ENABLE will be set by intel_ddi_prepare_link_retrain() later */
+	intel_dp->DP = dig_port->saved_port_bits |
+		DDI_PORT_WIDTH(crtc_state->lane_count) |
+		DDI_BUF_TRANS_SELECT(0);
+
+	if (IS_ALDERLAKE_P(i915) && intel_phy_is_tc(i915, phy)) {
+		intel_dp->DP |= ddi_buf_phy_link_rate(crtc_state->port_clock);
+		if (!intel_tc_port_in_tbt_alt_mode(dig_port))
+			intel_dp->DP |= DDI_BUF_CTL_TC_PHY_OWNERSHIP;
+	}
+}
+
+static int icl_calc_tbt_pll_link(struct drm_i915_private *dev_priv,
+				 enum port port)
+{
+	u32 val = intel_de_read(dev_priv, DDI_CLK_SEL(port)) & DDI_CLK_SEL_MASK;
+
+	switch (val) {
+	case DDI_CLK_SEL_NONE:
+		return 0;
+	case DDI_CLK_SEL_TBT_162:
+		return 162000;
+	case DDI_CLK_SEL_TBT_270:
+		return 270000;
+	case DDI_CLK_SEL_TBT_540:
+		return 540000;
+	case DDI_CLK_SEL_TBT_810:
+		return 810000;
+	default:
+		MISSING_CASE(val);
+		return 0;
+	}
+}
+
+static void ddi_dotclock_get(struct intel_crtc_state *pipe_config)
+{
+	/* CRT dotclock is determined via other means */
+	if (pipe_config->has_pch_encoder)
+		return;
+
+	pipe_config->hw.adjusted_mode.crtc_clock =
+		intel_crtc_dotclock(pipe_config);
+}
+
+void intel_ddi_set_dp_msa(const struct intel_crtc_state *crtc_state,
+			  const struct drm_connector_state *conn_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 temp;
+
+	if (!intel_crtc_has_dp_encoder(crtc_state))
+		return;
+
+	drm_WARN_ON(&dev_priv->drm, transcoder_is_dsi(cpu_transcoder));
+
+	temp = DP_MSA_MISC_SYNC_CLOCK;
+
+	switch (crtc_state->pipe_bpp) {
+	case 18:
+		temp |= DP_MSA_MISC_6_BPC;
+		break;
+	case 24:
+		temp |= DP_MSA_MISC_8_BPC;
+		break;
+	case 30:
+		temp |= DP_MSA_MISC_10_BPC;
+		break;
+	case 36:
+		temp |= DP_MSA_MISC_12_BPC;
+		break;
+	default:
+		MISSING_CASE(crtc_state->pipe_bpp);
+		break;
+	}
+
+	/* nonsense combination */
+	drm_WARN_ON(&dev_priv->drm, crtc_state->limited_color_range &&
+		    crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB);
+
+	if (crtc_state->limited_color_range)
+		temp |= DP_MSA_MISC_COLOR_CEA_RGB;
+
+	/*
+	 * As per DP 1.2 spec section 2.3.4.3 while sending
+	 * YCBCR 444 signals we should program MSA MISC1/0 fields with
+	 * colorspace information.
+	 */
+	if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR444)
+		temp |= DP_MSA_MISC_COLOR_YCBCR_444_BT709;
+
+	/*
+	 * As per DP 1.4a spec section 2.2.4.3 [MSA Field for Indication
+	 * of Color Encoding Format and Content Color Gamut] while sending
+	 * YCBCR 420, HDR BT.2020 signals we should program MSA MISC1 fields
+	 * which indicate VSC SDP for the Pixel Encoding/Colorimetry Format.
+	 */
+	if (intel_dp_needs_vsc_sdp(crtc_state, conn_state))
+		temp |= DP_MSA_MISC_COLOR_VSC_SDP;
+
+	intel_de_write(dev_priv, TRANS_MSA_MISC(cpu_transcoder), temp);
+}
+
+static u32 bdw_trans_port_sync_master_select(enum transcoder master_transcoder)
+{
+	if (master_transcoder == TRANSCODER_EDP)
+		return 0;
+	else
+		return master_transcoder + 1;
+}
+
+static void
+intel_ddi_config_transcoder_dp2(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 val = 0;
+
+	if (intel_dp_is_uhbr(crtc_state))
+		val = TRANS_DP2_128B132B_CHANNEL_CODING;
+
+	intel_de_write(i915, TRANS_DP2_CTL(cpu_transcoder), val);
+}
+
+/*
+ * Returns the TRANS_DDI_FUNC_CTL value based on CRTC state.
+ *
+ * Only intended to be used by intel_ddi_enable_transcoder_func() and
+ * intel_ddi_config_transcoder_func().
+ */
+static u32
+intel_ddi_transcoder_func_reg_val_get(struct intel_encoder *encoder,
+				      const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	enum port port = encoder->port;
+	u32 temp;
+
+	/* Enable TRANS_DDI_FUNC_CTL for the pipe to work in HDMI mode */
+	temp = TRANS_DDI_FUNC_ENABLE;
+	if (DISPLAY_VER(dev_priv) >= 12)
+		temp |= TGL_TRANS_DDI_SELECT_PORT(port);
+	else
+		temp |= TRANS_DDI_SELECT_PORT(port);
+
+	switch (crtc_state->pipe_bpp) {
+	default:
+		MISSING_CASE(crtc_state->pipe_bpp);
+		fallthrough;
+	case 18:
+		temp |= TRANS_DDI_BPC_6;
+		break;
+	case 24:
+		temp |= TRANS_DDI_BPC_8;
+		break;
+	case 30:
+		temp |= TRANS_DDI_BPC_10;
+		break;
+	case 36:
+		temp |= TRANS_DDI_BPC_12;
+		break;
+	}
+
+	if (crtc_state->hw.adjusted_mode.flags & DRM_MODE_FLAG_PVSYNC)
+		temp |= TRANS_DDI_PVSYNC;
+	if (crtc_state->hw.adjusted_mode.flags & DRM_MODE_FLAG_PHSYNC)
+		temp |= TRANS_DDI_PHSYNC;
+
+	if (cpu_transcoder == TRANSCODER_EDP) {
+		switch (pipe) {
+		default:
+			MISSING_CASE(pipe);
+			fallthrough;
+		case PIPE_A:
+			/* On Haswell, can only use the always-on power well for
+			 * eDP when not using the panel fitter, and when not
+			 * using motion blur mitigation (which we don't
+			 * support). */
+			if (crtc_state->pch_pfit.force_thru)
+				temp |= TRANS_DDI_EDP_INPUT_A_ONOFF;
+			else
+				temp |= TRANS_DDI_EDP_INPUT_A_ON;
+			break;
+		case PIPE_B:
+			temp |= TRANS_DDI_EDP_INPUT_B_ONOFF;
+			break;
+		case PIPE_C:
+			temp |= TRANS_DDI_EDP_INPUT_C_ONOFF;
+			break;
+		}
+	}
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI)) {
+		if (crtc_state->has_hdmi_sink)
+			temp |= TRANS_DDI_MODE_SELECT_HDMI;
+		else
+			temp |= TRANS_DDI_MODE_SELECT_DVI;
+
+		if (crtc_state->hdmi_scrambling)
+			temp |= TRANS_DDI_HDMI_SCRAMBLING;
+		if (crtc_state->hdmi_high_tmds_clock_ratio)
+			temp |= TRANS_DDI_HIGH_TMDS_CHAR_RATE;
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG)) {
+		temp |= TRANS_DDI_MODE_SELECT_FDI_OR_128B132B;
+		temp |= (crtc_state->fdi_lanes - 1) << 1;
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST)) {
+		if (intel_dp_is_uhbr(crtc_state))
+			temp |= TRANS_DDI_MODE_SELECT_FDI_OR_128B132B;
+		else
+			temp |= TRANS_DDI_MODE_SELECT_DP_MST;
+		temp |= DDI_PORT_WIDTH(crtc_state->lane_count);
+
+		if (DISPLAY_VER(dev_priv) >= 12) {
+			enum transcoder master;
+
+			master = crtc_state->mst_master_transcoder;
+			drm_WARN_ON(&dev_priv->drm,
+				    master == INVALID_TRANSCODER);
+			temp |= TRANS_DDI_MST_TRANSPORT_SELECT(master);
+		}
+	} else {
+		temp |= TRANS_DDI_MODE_SELECT_DP_SST;
+		temp |= DDI_PORT_WIDTH(crtc_state->lane_count);
+	}
+
+	if (IS_DISPLAY_VER(dev_priv, 8, 10) &&
+	    crtc_state->master_transcoder != INVALID_TRANSCODER) {
+		u8 master_select =
+			bdw_trans_port_sync_master_select(crtc_state->master_transcoder);
+
+		temp |= TRANS_DDI_PORT_SYNC_ENABLE |
+			TRANS_DDI_PORT_SYNC_MASTER_SELECT(master_select);
+	}
+
+	return temp;
+}
+
+void intel_ddi_enable_transcoder_func(struct intel_encoder *encoder,
+				      const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		enum transcoder master_transcoder = crtc_state->master_transcoder;
+		u32 ctl2 = 0;
+
+		if (master_transcoder != INVALID_TRANSCODER) {
+			u8 master_select =
+				bdw_trans_port_sync_master_select(master_transcoder);
+
+			ctl2 |= PORT_SYNC_MODE_ENABLE |
+				PORT_SYNC_MODE_MASTER_SELECT(master_select);
+		}
+
+		intel_de_write(dev_priv,
+			       TRANS_DDI_FUNC_CTL2(cpu_transcoder), ctl2);
+	}
+
+	intel_de_write(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder),
+		       intel_ddi_transcoder_func_reg_val_get(encoder,
+							     crtc_state));
+}
+
+/*
+ * Same as intel_ddi_enable_transcoder_func(), but it does not set the enable
+ * bit.
+ */
+static void
+intel_ddi_config_transcoder_func(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 ctl;
+
+	ctl = intel_ddi_transcoder_func_reg_val_get(encoder, crtc_state);
+	ctl &= ~TRANS_DDI_FUNC_ENABLE;
+	intel_de_write(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder), ctl);
+}
+
+void intel_ddi_disable_transcoder_func(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 ctl;
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		intel_de_write(dev_priv,
+			       TRANS_DDI_FUNC_CTL2(cpu_transcoder), 0);
+
+	ctl = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder));
+
+	drm_WARN_ON(crtc->base.dev, ctl & TRANS_DDI_HDCP_SIGNALLING);
+
+	ctl &= ~TRANS_DDI_FUNC_ENABLE;
+
+	if (IS_DISPLAY_VER(dev_priv, 8, 10))
+		ctl &= ~(TRANS_DDI_PORT_SYNC_ENABLE |
+			 TRANS_DDI_PORT_SYNC_MASTER_SELECT_MASK);
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		if (!intel_dp_mst_is_master_trans(crtc_state)) {
+			ctl &= ~(TGL_TRANS_DDI_PORT_MASK |
+				 TRANS_DDI_MODE_SELECT_MASK);
+		}
+	} else {
+		ctl &= ~(TRANS_DDI_PORT_MASK | TRANS_DDI_MODE_SELECT_MASK);
+	}
+
+	intel_de_write(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder), ctl);
+
+	if (intel_has_quirk(dev_priv, QUIRK_INCREASE_DDI_DISABLED_TIME) &&
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Quirk Increase DDI disabled time\n");
+		/* Quirk time at 100ms for reliable operation */
+		msleep(100);
+	}
+}
+
+int intel_ddi_toggle_hdcp_bits(struct intel_encoder *intel_encoder,
+			       enum transcoder cpu_transcoder,
+			       bool enable, u32 hdcp_mask)
+{
+	struct drm_device *dev = intel_encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	intel_wakeref_t wakeref;
+	int ret = 0;
+	u32 tmp;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     intel_encoder->power_domain);
+	if (drm_WARN_ON(dev, !wakeref))
+		return -ENXIO;
+
+	tmp = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder));
+	if (enable)
+		tmp |= hdcp_mask;
+	else
+		tmp &= ~hdcp_mask;
+	intel_de_write(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder), tmp);
+	intel_display_power_put(dev_priv, intel_encoder->power_domain, wakeref);
+	return ret;
+}
+
+bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector)
+{
+	struct drm_device *dev = intel_connector->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_encoder *encoder = intel_attached_encoder(intel_connector);
+	int type = intel_connector->base.connector_type;
+	enum port port = encoder->port;
+	enum transcoder cpu_transcoder;
+	intel_wakeref_t wakeref;
+	enum pipe pipe = 0;
+	u32 tmp;
+	bool ret;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     encoder->power_domain);
+	if (!wakeref)
+		return false;
+
+	if (!encoder->get_hw_state(encoder, &pipe)) {
+		ret = false;
+		goto out;
+	}
+
+	if (HAS_TRANSCODER(dev_priv, TRANSCODER_EDP) && port == PORT_A)
+		cpu_transcoder = TRANSCODER_EDP;
+	else
+		cpu_transcoder = (enum transcoder) pipe;
+
+	tmp = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder));
+
+	switch (tmp & TRANS_DDI_MODE_SELECT_MASK) {
+	case TRANS_DDI_MODE_SELECT_HDMI:
+	case TRANS_DDI_MODE_SELECT_DVI:
+		ret = type == DRM_MODE_CONNECTOR_HDMIA;
+		break;
+
+	case TRANS_DDI_MODE_SELECT_DP_SST:
+		ret = type == DRM_MODE_CONNECTOR_eDP ||
+		      type == DRM_MODE_CONNECTOR_DisplayPort;
+		break;
+
+	case TRANS_DDI_MODE_SELECT_DP_MST:
+		/* if the transcoder is in MST state then
+		 * connector isn't connected */
+		ret = false;
+		break;
+
+	case TRANS_DDI_MODE_SELECT_FDI_OR_128B132B:
+		if (HAS_DP20(dev_priv))
+			/* 128b/132b */
+			ret = false;
+		else
+			/* FDI */
+			ret = type == DRM_MODE_CONNECTOR_VGA;
+		break;
+
+	default:
+		ret = false;
+		break;
+	}
+
+out:
+	intel_display_power_put(dev_priv, encoder->power_domain, wakeref);
+
+	return ret;
+}
+
+static void intel_ddi_get_encoder_pipes(struct intel_encoder *encoder,
+					u8 *pipe_mask, bool *is_dp_mst)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum port port = encoder->port;
+	intel_wakeref_t wakeref;
+	enum pipe p;
+	u32 tmp;
+	u8 mst_pipe_mask;
+
+	*pipe_mask = 0;
+	*is_dp_mst = false;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     encoder->power_domain);
+	if (!wakeref)
+		return;
+
+	tmp = intel_de_read(dev_priv, DDI_BUF_CTL(port));
+	if (!(tmp & DDI_BUF_CTL_ENABLE))
+		goto out;
+
+	if (HAS_TRANSCODER(dev_priv, TRANSCODER_EDP) && port == PORT_A) {
+		tmp = intel_de_read(dev_priv,
+				    TRANS_DDI_FUNC_CTL(TRANSCODER_EDP));
+
+		switch (tmp & TRANS_DDI_EDP_INPUT_MASK) {
+		default:
+			MISSING_CASE(tmp & TRANS_DDI_EDP_INPUT_MASK);
+			fallthrough;
+		case TRANS_DDI_EDP_INPUT_A_ON:
+		case TRANS_DDI_EDP_INPUT_A_ONOFF:
+			*pipe_mask = BIT(PIPE_A);
+			break;
+		case TRANS_DDI_EDP_INPUT_B_ONOFF:
+			*pipe_mask = BIT(PIPE_B);
+			break;
+		case TRANS_DDI_EDP_INPUT_C_ONOFF:
+			*pipe_mask = BIT(PIPE_C);
+			break;
+		}
+
+		goto out;
+	}
+
+	mst_pipe_mask = 0;
+	for_each_pipe(dev_priv, p) {
+		enum transcoder cpu_transcoder = (enum transcoder)p;
+		unsigned int port_mask, ddi_select;
+		intel_wakeref_t trans_wakeref;
+
+		trans_wakeref = intel_display_power_get_if_enabled(dev_priv,
+								   POWER_DOMAIN_TRANSCODER(cpu_transcoder));
+		if (!trans_wakeref)
+			continue;
+
+		if (DISPLAY_VER(dev_priv) >= 12) {
+			port_mask = TGL_TRANS_DDI_PORT_MASK;
+			ddi_select = TGL_TRANS_DDI_SELECT_PORT(port);
+		} else {
+			port_mask = TRANS_DDI_PORT_MASK;
+			ddi_select = TRANS_DDI_SELECT_PORT(port);
+		}
+
+		tmp = intel_de_read(dev_priv,
+				    TRANS_DDI_FUNC_CTL(cpu_transcoder));
+		intel_display_power_put(dev_priv, POWER_DOMAIN_TRANSCODER(cpu_transcoder),
+					trans_wakeref);
+
+		if ((tmp & port_mask) != ddi_select)
+			continue;
+
+		if ((tmp & TRANS_DDI_MODE_SELECT_MASK) == TRANS_DDI_MODE_SELECT_DP_MST ||
+		    (HAS_DP20(dev_priv) &&
+		     (tmp & TRANS_DDI_MODE_SELECT_MASK) == TRANS_DDI_MODE_SELECT_FDI_OR_128B132B))
+			mst_pipe_mask |= BIT(p);
+
+		*pipe_mask |= BIT(p);
+	}
+
+	if (!*pipe_mask)
+		drm_dbg_kms(&dev_priv->drm,
+			    "No pipe for [ENCODER:%d:%s] found\n",
+			    encoder->base.base.id, encoder->base.name);
+
+	if (!mst_pipe_mask && hweight8(*pipe_mask) > 1) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Multiple pipes for [ENCODER:%d:%s] (pipe_mask %02x)\n",
+			    encoder->base.base.id, encoder->base.name,
+			    *pipe_mask);
+		*pipe_mask = BIT(ffs(*pipe_mask) - 1);
+	}
+
+	if (mst_pipe_mask && mst_pipe_mask != *pipe_mask)
+		drm_dbg_kms(&dev_priv->drm,
+			    "Conflicting MST and non-MST state for [ENCODER:%d:%s] (pipe_mask %02x mst_pipe_mask %02x)\n",
+			    encoder->base.base.id, encoder->base.name,
+			    *pipe_mask, mst_pipe_mask);
+	else
+		*is_dp_mst = mst_pipe_mask;
+
+out:
+	if (*pipe_mask && (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))) {
+		tmp = intel_de_read(dev_priv, BXT_PHY_CTL(port));
+		if ((tmp & (BXT_PHY_CMNLANE_POWERDOWN_ACK |
+			    BXT_PHY_LANE_POWERDOWN_ACK |
+			    BXT_PHY_LANE_ENABLED)) != BXT_PHY_LANE_ENABLED)
+			drm_err(&dev_priv->drm,
+				"[ENCODER:%d:%s] enabled but PHY powered down? (PHY_CTL %08x)\n",
+				encoder->base.base.id, encoder->base.name, tmp);
+	}
+
+	intel_display_power_put(dev_priv, encoder->power_domain, wakeref);
+}
+
+bool intel_ddi_get_hw_state(struct intel_encoder *encoder,
+			    enum pipe *pipe)
+{
+	u8 pipe_mask;
+	bool is_mst;
+
+	intel_ddi_get_encoder_pipes(encoder, &pipe_mask, &is_mst);
+
+	if (is_mst || !pipe_mask)
+		return false;
+
+	*pipe = ffs(pipe_mask) - 1;
+
+	return true;
+}
+
+static enum intel_display_power_domain
+intel_ddi_main_link_aux_domain(struct intel_digital_port *dig_port)
+{
+	/* ICL+ HW requires corresponding AUX IOs to be powered up for PSR with
+	 * DC states enabled at the same time, while for driver initiated AUX
+	 * transfers we need the same AUX IOs to be powered but with DC states
+	 * disabled. Accordingly use the AUX power domain here which leaves DC
+	 * states enabled.
+	 * However, for non-A AUX ports the corresponding non-EDP transcoders
+	 * would have already enabled power well 2 and DC_OFF. This means we can
+	 * acquire a wider POWER_DOMAIN_AUX_{B,C,D,F} reference instead of a
+	 * specific AUX_IO reference without powering up any extra wells.
+	 * Note that PSR is enabled only on Port A even though this function
+	 * returns the correct domain for other ports too.
+	 */
+	return dig_port->aux_ch == AUX_CH_A ? POWER_DOMAIN_AUX_IO_A :
+					      intel_aux_power_domain(dig_port);
+}
+
+static void intel_ddi_get_power_domains(struct intel_encoder *encoder,
+					struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port;
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+
+	/*
+	 * TODO: Add support for MST encoders. Atm, the following should never
+	 * happen since fake-MST encoders don't set their get_power_domains()
+	 * hook.
+	 */
+	if (drm_WARN_ON(&dev_priv->drm,
+			intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST)))
+		return;
+
+	dig_port = enc_to_dig_port(encoder);
+
+	if (!intel_tc_port_in_tbt_alt_mode(dig_port)) {
+		drm_WARN_ON(&dev_priv->drm, dig_port->ddi_io_wakeref);
+		dig_port->ddi_io_wakeref = intel_display_power_get(dev_priv,
+								   dig_port->ddi_io_power_domain);
+	}
+
+	/*
+	 * AUX power is only needed for (e)DP mode, and for HDMI mode on TC
+	 * ports.
+	 */
+	if (intel_crtc_has_dp_encoder(crtc_state) ||
+	    intel_phy_is_tc(dev_priv, phy)) {
+		drm_WARN_ON(&dev_priv->drm, dig_port->aux_wakeref);
+		dig_port->aux_wakeref =
+			intel_display_power_get(dev_priv,
+						intel_ddi_main_link_aux_domain(dig_port));
+	}
+}
+
+void intel_ddi_enable_pipe_clock(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+	u32 val;
+
+	if (cpu_transcoder != TRANSCODER_EDP) {
+		if (DISPLAY_VER(dev_priv) >= 13)
+			val = TGL_TRANS_CLK_SEL_PORT(phy);
+		else if (DISPLAY_VER(dev_priv) >= 12)
+			val = TGL_TRANS_CLK_SEL_PORT(encoder->port);
+		else
+			val = TRANS_CLK_SEL_PORT(encoder->port);
+
+		intel_de_write(dev_priv, TRANS_CLK_SEL(cpu_transcoder), val);
+	}
+}
+
+void intel_ddi_disable_pipe_clock(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	if (cpu_transcoder != TRANSCODER_EDP) {
+		if (DISPLAY_VER(dev_priv) >= 12)
+			intel_de_write(dev_priv,
+				       TRANS_CLK_SEL(cpu_transcoder),
+				       TGL_TRANS_CLK_SEL_DISABLED);
+		else
+			intel_de_write(dev_priv,
+				       TRANS_CLK_SEL(cpu_transcoder),
+				       TRANS_CLK_SEL_DISABLED);
+	}
+}
+
+static void _skl_ddi_set_iboost(struct drm_i915_private *dev_priv,
+				enum port port, u8 iboost)
+{
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, DISPIO_CR_TX_BMU_CR0);
+	tmp &= ~(BALANCE_LEG_MASK(port) | BALANCE_LEG_DISABLE(port));
+	if (iboost)
+		tmp |= iboost << BALANCE_LEG_SHIFT(port);
+	else
+		tmp |= BALANCE_LEG_DISABLE(port);
+	intel_de_write(dev_priv, DISPIO_CR_TX_BMU_CR0, tmp);
+}
+
+static void skl_ddi_set_iboost(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state,
+			       int level)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u8 iboost;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		iboost = intel_bios_encoder_hdmi_boost_level(encoder->devdata);
+	else
+		iboost = intel_bios_encoder_dp_boost_level(encoder->devdata);
+
+	if (iboost == 0) {
+		const struct intel_ddi_buf_trans *trans;
+		int n_entries;
+
+		trans = encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+		if (drm_WARN_ON_ONCE(&dev_priv->drm, !trans))
+			return;
+
+		iboost = trans->entries[level].hsw.i_boost;
+	}
+
+	/* Make sure that the requested I_boost is valid */
+	if (iboost && iboost != 0x1 && iboost != 0x3 && iboost != 0x7) {
+		drm_err(&dev_priv->drm, "Invalid I_boost value %u\n", iboost);
+		return;
+	}
+
+	_skl_ddi_set_iboost(dev_priv, encoder->port, iboost);
+
+	if (encoder->port == PORT_A && dig_port->max_lanes == 4)
+		_skl_ddi_set_iboost(dev_priv, PORT_E, iboost);
+}
+
+static u8 intel_ddi_dp_voltage_max(struct intel_dp *intel_dp,
+				   const struct intel_crtc_state *crtc_state)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	int n_entries;
+
+	encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+
+	if (drm_WARN_ON(&dev_priv->drm, n_entries < 1))
+		n_entries = 1;
+	if (drm_WARN_ON(&dev_priv->drm,
+			n_entries > ARRAY_SIZE(index_to_dp_signal_levels)))
+		n_entries = ARRAY_SIZE(index_to_dp_signal_levels);
+
+	return index_to_dp_signal_levels[n_entries - 1] &
+		DP_TRAIN_VOLTAGE_SWING_MASK;
+}
+
+/*
+ * We assume that the full set of pre-emphasis values can be
+ * used on all DDI platforms. Should that change we need to
+ * rethink this code.
+ */
+static u8 intel_ddi_dp_preemph_max(struct intel_dp *intel_dp)
+{
+	return DP_TRAIN_PRE_EMPH_LEVEL_3;
+}
+
+static u32 icl_combo_phy_loadgen_select(const struct intel_crtc_state *crtc_state,
+					int lane)
+{
+	if (crtc_state->port_clock > 600000)
+		return 0;
+
+	if (crtc_state->lane_count == 4)
+		return lane >= 1 ? LOADGEN_SELECT : 0;
+	else
+		return lane == 1 || lane == 2 ? LOADGEN_SELECT : 0;
+}
+
+static void icl_ddi_combo_vswing_program(struct intel_encoder *encoder,
+					 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	const struct intel_ddi_buf_trans *trans;
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+	int n_entries, ln;
+	u32 val;
+
+	trans = encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !trans))
+		return;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP)) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		val = EDP4K2K_MODE_OVRD_EN | EDP4K2K_MODE_OVRD_OPTIMIZED;
+		intel_dp->hobl_active = is_hobl_buf_trans(trans);
+		intel_de_rmw(dev_priv, ICL_PORT_CL_DW10(phy), val,
+			     intel_dp->hobl_active ? val : 0);
+	}
+
+	/* Set PORT_TX_DW5 */
+	val = intel_de_read(dev_priv, ICL_PORT_TX_DW5_LN(0, phy));
+	val &= ~(SCALING_MODE_SEL_MASK | RTERM_SELECT_MASK |
+		  TAP2_DISABLE | TAP3_DISABLE);
+	val |= SCALING_MODE_SEL(0x2);
+	val |= RTERM_SELECT(0x6);
+	val |= TAP3_DISABLE;
+	intel_de_write(dev_priv, ICL_PORT_TX_DW5_GRP(phy), val);
+
+	/* Program PORT_TX_DW2 */
+	for (ln = 0; ln < 4; ln++) {
+		int level = intel_ddi_level(encoder, crtc_state, ln);
+
+		intel_de_rmw(dev_priv, ICL_PORT_TX_DW2_LN(ln, phy),
+			     SWING_SEL_UPPER_MASK | SWING_SEL_LOWER_MASK | RCOMP_SCALAR_MASK,
+			     SWING_SEL_UPPER(trans->entries[level].icl.dw2_swing_sel) |
+			     SWING_SEL_LOWER(trans->entries[level].icl.dw2_swing_sel) |
+			     RCOMP_SCALAR(0x98));
+	}
+
+	/* Program PORT_TX_DW4 */
+	/* We cannot write to GRP. It would overwrite individual loadgen. */
+	for (ln = 0; ln < 4; ln++) {
+		int level = intel_ddi_level(encoder, crtc_state, ln);
+
+		intel_de_rmw(dev_priv, ICL_PORT_TX_DW4_LN(ln, phy),
+			     POST_CURSOR_1_MASK | POST_CURSOR_2_MASK | CURSOR_COEFF_MASK,
+			     POST_CURSOR_1(trans->entries[level].icl.dw4_post_cursor_1) |
+			     POST_CURSOR_2(trans->entries[level].icl.dw4_post_cursor_2) |
+			     CURSOR_COEFF(trans->entries[level].icl.dw4_cursor_coeff));
+	}
+
+	/* Program PORT_TX_DW7 */
+	for (ln = 0; ln < 4; ln++) {
+		int level = intel_ddi_level(encoder, crtc_state, ln);
+
+		intel_de_rmw(dev_priv, ICL_PORT_TX_DW7_LN(ln, phy),
+			     N_SCALAR_MASK,
+			     N_SCALAR(trans->entries[level].icl.dw7_n_scalar));
+	}
+}
+
+static void icl_combo_phy_set_signal_levels(struct intel_encoder *encoder,
+					    const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+	u32 val;
+	int ln;
+
+	/*
+	 * 1. If port type is eDP or DP,
+	 * set PORT_PCS_DW1 cmnkeeper_enable to 1b,
+	 * else clear to 0b.
+	 */
+	val = intel_de_read(dev_priv, ICL_PORT_PCS_DW1_LN(0, phy));
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		val &= ~COMMON_KEEPER_EN;
+	else
+		val |= COMMON_KEEPER_EN;
+	intel_de_write(dev_priv, ICL_PORT_PCS_DW1_GRP(phy), val);
+
+	/* 2. Program loadgen select */
+	/*
+	 * Program PORT_TX_DW4 depending on Bit rate and used lanes
+	 * <= 6 GHz and 4 lanes (LN0=0, LN1=1, LN2=1, LN3=1)
+	 * <= 6 GHz and 1,2 lanes (LN0=0, LN1=1, LN2=1, LN3=0)
+	 * > 6 GHz (LN0=0, LN1=0, LN2=0, LN3=0)
+	 */
+	for (ln = 0; ln < 4; ln++) {
+		intel_de_rmw(dev_priv, ICL_PORT_TX_DW4_LN(ln, phy),
+			     LOADGEN_SELECT,
+			     icl_combo_phy_loadgen_select(crtc_state, ln));
+	}
+
+	/* 3. Set PORT_CL_DW5 SUS Clock Config to 11b */
+	intel_de_rmw(dev_priv, ICL_PORT_CL_DW5(phy),
+		     0, SUS_CLOCK_CONFIG);
+
+	/* 4. Clear training enable to change swing values */
+	val = intel_de_read(dev_priv, ICL_PORT_TX_DW5_LN(0, phy));
+	val &= ~TX_TRAINING_EN;
+	intel_de_write(dev_priv, ICL_PORT_TX_DW5_GRP(phy), val);
+
+	/* 5. Program swing and de-emphasis */
+	icl_ddi_combo_vswing_program(encoder, crtc_state);
+
+	/* 6. Set training enable to trigger update */
+	val = intel_de_read(dev_priv, ICL_PORT_TX_DW5_LN(0, phy));
+	val |= TX_TRAINING_EN;
+	intel_de_write(dev_priv, ICL_PORT_TX_DW5_GRP(phy), val);
+}
+
+static void icl_mg_phy_set_signal_levels(struct intel_encoder *encoder,
+					 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum tc_port tc_port = intel_port_to_tc(dev_priv, encoder->port);
+	const struct intel_ddi_buf_trans *trans;
+	int n_entries, ln;
+
+	if (intel_tc_port_in_tbt_alt_mode(enc_to_dig_port(encoder)))
+		return;
+
+	trans = encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !trans))
+		return;
+
+	for (ln = 0; ln < 2; ln++) {
+		intel_de_rmw(dev_priv, MG_TX1_LINK_PARAMS(ln, tc_port),
+			     CRI_USE_FS32, 0);
+		intel_de_rmw(dev_priv, MG_TX2_LINK_PARAMS(ln, tc_port),
+			     CRI_USE_FS32, 0);
+	}
+
+	/* Program MG_TX_SWINGCTRL with values from vswing table */
+	for (ln = 0; ln < 2; ln++) {
+		int level;
+
+		level = intel_ddi_level(encoder, crtc_state, 2*ln+0);
+
+		intel_de_rmw(dev_priv, MG_TX1_SWINGCTRL(ln, tc_port),
+			     CRI_TXDEEMPH_OVERRIDE_17_12_MASK,
+			     CRI_TXDEEMPH_OVERRIDE_17_12(trans->entries[level].mg.cri_txdeemph_override_17_12));
+
+		level = intel_ddi_level(encoder, crtc_state, 2*ln+1);
+
+		intel_de_rmw(dev_priv, MG_TX2_SWINGCTRL(ln, tc_port),
+			     CRI_TXDEEMPH_OVERRIDE_17_12_MASK,
+			     CRI_TXDEEMPH_OVERRIDE_17_12(trans->entries[level].mg.cri_txdeemph_override_17_12));
+	}
+
+	/* Program MG_TX_DRVCTRL with values from vswing table */
+	for (ln = 0; ln < 2; ln++) {
+		int level;
+
+		level = intel_ddi_level(encoder, crtc_state, 2*ln+0);
+
+		intel_de_rmw(dev_priv, MG_TX1_DRVCTRL(ln, tc_port),
+			     CRI_TXDEEMPH_OVERRIDE_11_6_MASK |
+			     CRI_TXDEEMPH_OVERRIDE_5_0_MASK,
+			     CRI_TXDEEMPH_OVERRIDE_11_6(trans->entries[level].mg.cri_txdeemph_override_11_6) |
+			     CRI_TXDEEMPH_OVERRIDE_5_0(trans->entries[level].mg.cri_txdeemph_override_5_0) |
+			     CRI_TXDEEMPH_OVERRIDE_EN);
+
+		level = intel_ddi_level(encoder, crtc_state, 2*ln+1);
+
+		intel_de_rmw(dev_priv, MG_TX2_DRVCTRL(ln, tc_port),
+			     CRI_TXDEEMPH_OVERRIDE_11_6_MASK |
+			     CRI_TXDEEMPH_OVERRIDE_5_0_MASK,
+			     CRI_TXDEEMPH_OVERRIDE_11_6(trans->entries[level].mg.cri_txdeemph_override_11_6) |
+			     CRI_TXDEEMPH_OVERRIDE_5_0(trans->entries[level].mg.cri_txdeemph_override_5_0) |
+			     CRI_TXDEEMPH_OVERRIDE_EN);
+
+		/* FIXME: Program CRI_LOADGEN_SEL after the spec is updated */
+	}
+
+	/*
+	 * Program MG_CLKHUB<LN, port being used> with value from frequency table
+	 * In case of Legacy mode on MG PHY, both TX1 and TX2 enabled so use the
+	 * values from table for which TX1 and TX2 enabled.
+	 */
+	for (ln = 0; ln < 2; ln++) {
+		intel_de_rmw(dev_priv, MG_CLKHUB(ln, tc_port),
+			     CFG_LOW_RATE_LKREN_EN,
+			     crtc_state->port_clock < 300000 ? CFG_LOW_RATE_LKREN_EN : 0);
+	}
+
+	/* Program the MG_TX_DCC<LN, port being used> based on the link frequency */
+	for (ln = 0; ln < 2; ln++) {
+		intel_de_rmw(dev_priv, MG_TX1_DCC(ln, tc_port),
+			     CFG_AMI_CK_DIV_OVERRIDE_VAL_MASK |
+			     CFG_AMI_CK_DIV_OVERRIDE_EN,
+			     crtc_state->port_clock > 500000 ?
+			     CFG_AMI_CK_DIV_OVERRIDE_VAL(1) |
+			     CFG_AMI_CK_DIV_OVERRIDE_EN : 0);
+
+		intel_de_rmw(dev_priv, MG_TX2_DCC(ln, tc_port),
+			     CFG_AMI_CK_DIV_OVERRIDE_VAL_MASK |
+			     CFG_AMI_CK_DIV_OVERRIDE_EN,
+			     crtc_state->port_clock > 500000 ?
+			     CFG_AMI_CK_DIV_OVERRIDE_VAL(1) |
+			     CFG_AMI_CK_DIV_OVERRIDE_EN : 0);
+	}
+
+	/* Program MG_TX_PISO_READLOAD with values from vswing table */
+	for (ln = 0; ln < 2; ln++) {
+		intel_de_rmw(dev_priv, MG_TX1_PISO_READLOAD(ln, tc_port),
+			     0, CRI_CALCINIT);
+		intel_de_rmw(dev_priv, MG_TX2_PISO_READLOAD(ln, tc_port),
+			     0, CRI_CALCINIT);
+	}
+}
+
+static void tgl_dkl_phy_set_signal_levels(struct intel_encoder *encoder,
+					  const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum tc_port tc_port = intel_port_to_tc(dev_priv, encoder->port);
+	const struct intel_ddi_buf_trans *trans;
+	int n_entries, ln;
+
+	if (intel_tc_port_in_tbt_alt_mode(enc_to_dig_port(encoder)))
+		return;
+
+	trans = encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !trans))
+		return;
+
+	for (ln = 0; ln < 2; ln++) {
+		int level;
+
+		intel_dkl_phy_write(dev_priv, DKL_TX_PMD_LANE_SUS(tc_port), ln, 0);
+
+		level = intel_ddi_level(encoder, crtc_state, 2*ln+0);
+
+		intel_dkl_phy_rmw(dev_priv, DKL_TX_DPCNTL0(tc_port), ln,
+				  DKL_TX_PRESHOOT_COEFF_MASK |
+				  DKL_TX_DE_EMPAHSIS_COEFF_MASK |
+				  DKL_TX_VSWING_CONTROL_MASK,
+				  DKL_TX_PRESHOOT_COEFF(trans->entries[level].dkl.preshoot) |
+				  DKL_TX_DE_EMPHASIS_COEFF(trans->entries[level].dkl.de_emphasis) |
+				  DKL_TX_VSWING_CONTROL(trans->entries[level].dkl.vswing));
+
+		level = intel_ddi_level(encoder, crtc_state, 2*ln+1);
+
+		intel_dkl_phy_rmw(dev_priv, DKL_TX_DPCNTL1(tc_port), ln,
+				  DKL_TX_PRESHOOT_COEFF_MASK |
+				  DKL_TX_DE_EMPAHSIS_COEFF_MASK |
+				  DKL_TX_VSWING_CONTROL_MASK,
+				  DKL_TX_PRESHOOT_COEFF(trans->entries[level].dkl.preshoot) |
+				  DKL_TX_DE_EMPHASIS_COEFF(trans->entries[level].dkl.de_emphasis) |
+				  DKL_TX_VSWING_CONTROL(trans->entries[level].dkl.vswing));
+
+		intel_dkl_phy_rmw(dev_priv, DKL_TX_DPCNTL2(tc_port), ln,
+				  DKL_TX_DP20BITMODE, 0);
+
+		if (IS_ALDERLAKE_P(dev_priv)) {
+			u32 val;
+
+			if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI)) {
+				if (ln == 0) {
+					val = DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX1(0);
+					val |= DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX2(2);
+				} else {
+					val = DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX1(3);
+					val |= DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX2(3);
+				}
+			} else {
+				val = DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX1(0);
+				val |= DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX2(0);
+			}
+
+			intel_dkl_phy_rmw(dev_priv, DKL_TX_DPCNTL2(tc_port), ln,
+					  DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX1_MASK |
+					  DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX2_MASK,
+					  val);
+		}
+	}
+}
+
+static int translate_signal_level(struct intel_dp *intel_dp,
+				  u8 signal_levels)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(index_to_dp_signal_levels); i++) {
+		if (index_to_dp_signal_levels[i] == signal_levels)
+			return i;
+	}
+
+	drm_WARN(&i915->drm, 1,
+		 "Unsupported voltage swing/pre-emphasis level: 0x%x\n",
+		 signal_levels);
+
+	return 0;
+}
+
+static int intel_ddi_dp_level(struct intel_dp *intel_dp,
+			      const struct intel_crtc_state *crtc_state,
+			      int lane)
+{
+	u8 train_set = intel_dp->train_set[lane];
+
+	if (intel_dp_is_uhbr(crtc_state)) {
+		return train_set & DP_TX_FFE_PRESET_VALUE_MASK;
+	} else {
+		u8 signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
+						DP_TRAIN_PRE_EMPHASIS_MASK);
+
+		return translate_signal_level(intel_dp, signal_levels);
+	}
+}
+
+int intel_ddi_level(struct intel_encoder *encoder,
+		    const struct intel_crtc_state *crtc_state,
+		    int lane)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_ddi_buf_trans *trans;
+	int level, n_entries;
+
+	trans = encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+	if (drm_WARN_ON_ONCE(&i915->drm, !trans))
+		return 0;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		level = intel_ddi_hdmi_level(encoder, trans);
+	else
+		level = intel_ddi_dp_level(enc_to_intel_dp(encoder), crtc_state,
+					   lane);
+
+	if (drm_WARN_ON_ONCE(&i915->drm, level >= n_entries))
+		level = n_entries - 1;
+
+	return level;
+}
+
+static void
+hsw_set_signal_levels(struct intel_encoder *encoder,
+		      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	int level = intel_ddi_level(encoder, crtc_state, 0);
+	enum port port = encoder->port;
+	u32 signal_levels;
+
+	if (has_iboost(dev_priv))
+		skl_ddi_set_iboost(encoder, crtc_state, level);
+
+	/* HDMI ignores the rest */
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return;
+
+	signal_levels = DDI_BUF_TRANS_SELECT(level);
+
+	drm_dbg_kms(&dev_priv->drm, "Using signal levels %08x\n",
+		    signal_levels);
+
+	intel_dp->DP &= ~DDI_BUF_EMP_MASK;
+	intel_dp->DP |= signal_levels;
+
+	intel_de_write(dev_priv, DDI_BUF_CTL(port), intel_dp->DP);
+	intel_de_posting_read(dev_priv, DDI_BUF_CTL(port));
+}
+
+static void _icl_ddi_enable_clock(struct drm_i915_private *i915, i915_reg_t reg,
+				  u32 clk_sel_mask, u32 clk_sel, u32 clk_off)
+{
+	mutex_lock(&i915->display.dpll.lock);
+
+	intel_de_rmw(i915, reg, clk_sel_mask, clk_sel);
+
+	/*
+	 * "This step and the step before must be
+	 *  done with separate register writes."
+	 */
+	intel_de_rmw(i915, reg, clk_off, 0);
+
+	mutex_unlock(&i915->display.dpll.lock);
+}
+
+static void _icl_ddi_disable_clock(struct drm_i915_private *i915, i915_reg_t reg,
+				   u32 clk_off)
+{
+	mutex_lock(&i915->display.dpll.lock);
+
+	intel_de_rmw(i915, reg, 0, clk_off);
+
+	mutex_unlock(&i915->display.dpll.lock);
+}
+
+static bool _icl_ddi_is_clock_enabled(struct drm_i915_private *i915, i915_reg_t reg,
+				      u32 clk_off)
+{
+	return !(intel_de_read(i915, reg) & clk_off);
+}
+
+static struct intel_shared_dpll *
+_icl_ddi_get_pll(struct drm_i915_private *i915, i915_reg_t reg,
+		 u32 clk_sel_mask, u32 clk_sel_shift)
+{
+	enum intel_dpll_id id;
+
+	id = (intel_de_read(i915, reg) & clk_sel_mask) >> clk_sel_shift;
+
+	return intel_get_shared_dpll_by_id(i915, id);
+}
+
+static void adls_ddi_enable_clock(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	if (drm_WARN_ON(&i915->drm, !pll))
+		return;
+
+	_icl_ddi_enable_clock(i915, ADLS_DPCLKA_CFGCR(phy),
+			      ADLS_DPCLKA_CFGCR_DDI_CLK_SEL_MASK(phy),
+			      pll->info->id << ADLS_DPCLKA_CFGCR_DDI_SHIFT(phy),
+			      ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static void adls_ddi_disable_clock(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	_icl_ddi_disable_clock(i915, ADLS_DPCLKA_CFGCR(phy),
+			       ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static bool adls_ddi_is_clock_enabled(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	return _icl_ddi_is_clock_enabled(i915, ADLS_DPCLKA_CFGCR(phy),
+					 ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static struct intel_shared_dpll *adls_ddi_get_pll(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	return _icl_ddi_get_pll(i915, ADLS_DPCLKA_CFGCR(phy),
+				ADLS_DPCLKA_CFGCR_DDI_CLK_SEL_MASK(phy),
+				ADLS_DPCLKA_CFGCR_DDI_SHIFT(phy));
+}
+
+static void rkl_ddi_enable_clock(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	if (drm_WARN_ON(&i915->drm, !pll))
+		return;
+
+	_icl_ddi_enable_clock(i915, ICL_DPCLKA_CFGCR0,
+			      RKL_DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(phy),
+			      RKL_DPCLKA_CFGCR0_DDI_CLK_SEL(pll->info->id, phy),
+			      RKL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static void rkl_ddi_disable_clock(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	_icl_ddi_disable_clock(i915, ICL_DPCLKA_CFGCR0,
+			       RKL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static bool rkl_ddi_is_clock_enabled(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	return _icl_ddi_is_clock_enabled(i915, ICL_DPCLKA_CFGCR0,
+					 RKL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static struct intel_shared_dpll *rkl_ddi_get_pll(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	return _icl_ddi_get_pll(i915, ICL_DPCLKA_CFGCR0,
+				RKL_DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(phy),
+				RKL_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy));
+}
+
+static void dg1_ddi_enable_clock(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	if (drm_WARN_ON(&i915->drm, !pll))
+		return;
+
+	/*
+	 * If we fail this, something went very wrong: first 2 PLLs should be
+	 * used by first 2 phys and last 2 PLLs by last phys
+	 */
+	if (drm_WARN_ON(&i915->drm,
+			(pll->info->id < DPLL_ID_DG1_DPLL2 && phy >= PHY_C) ||
+			(pll->info->id >= DPLL_ID_DG1_DPLL2 && phy < PHY_C)))
+		return;
+
+	_icl_ddi_enable_clock(i915, DG1_DPCLKA_CFGCR0(phy),
+			      DG1_DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(phy),
+			      DG1_DPCLKA_CFGCR0_DDI_CLK_SEL(pll->info->id, phy),
+			      DG1_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static void dg1_ddi_disable_clock(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	_icl_ddi_disable_clock(i915, DG1_DPCLKA_CFGCR0(phy),
+			       DG1_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static bool dg1_ddi_is_clock_enabled(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	return _icl_ddi_is_clock_enabled(i915, DG1_DPCLKA_CFGCR0(phy),
+					 DG1_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static struct intel_shared_dpll *dg1_ddi_get_pll(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+	enum intel_dpll_id id;
+	u32 val;
+
+	val = intel_de_read(i915, DG1_DPCLKA_CFGCR0(phy));
+	val &= DG1_DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(phy);
+	val >>= DG1_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy);
+	id = val;
+
+	/*
+	 * _DG1_DPCLKA0_CFGCR0 maps between DPLL 0 and 1 with one bit for phy A
+	 * and B while _DG1_DPCLKA1_CFGCR0 maps between DPLL 2 and 3 with one
+	 * bit for phy C and D.
+	 */
+	if (phy >= PHY_C)
+		id += DPLL_ID_DG1_DPLL2;
+
+	return intel_get_shared_dpll_by_id(i915, id);
+}
+
+static void icl_ddi_combo_enable_clock(struct intel_encoder *encoder,
+				       const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	if (drm_WARN_ON(&i915->drm, !pll))
+		return;
+
+	_icl_ddi_enable_clock(i915, ICL_DPCLKA_CFGCR0,
+			      ICL_DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(phy),
+			      ICL_DPCLKA_CFGCR0_DDI_CLK_SEL(pll->info->id, phy),
+			      ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static void icl_ddi_combo_disable_clock(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	_icl_ddi_disable_clock(i915, ICL_DPCLKA_CFGCR0,
+			       ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+static bool icl_ddi_combo_is_clock_enabled(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	return _icl_ddi_is_clock_enabled(i915, ICL_DPCLKA_CFGCR0,
+					 ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy));
+}
+
+struct intel_shared_dpll *icl_ddi_combo_get_pll(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	return _icl_ddi_get_pll(i915, ICL_DPCLKA_CFGCR0,
+				ICL_DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(phy),
+				ICL_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy));
+}
+
+static void jsl_ddi_tc_enable_clock(struct intel_encoder *encoder,
+				    const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	enum port port = encoder->port;
+
+	if (drm_WARN_ON(&i915->drm, !pll))
+		return;
+
+	/*
+	 * "For DDIC and DDID, program DDI_CLK_SEL to map the MG clock to the port.
+	 *  MG does not exist, but the programming is required to ungate DDIC and DDID."
+	 */
+	intel_de_write(i915, DDI_CLK_SEL(port), DDI_CLK_SEL_MG);
+
+	icl_ddi_combo_enable_clock(encoder, crtc_state);
+}
+
+static void jsl_ddi_tc_disable_clock(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+
+	icl_ddi_combo_disable_clock(encoder);
+
+	intel_de_write(i915, DDI_CLK_SEL(port), DDI_CLK_SEL_NONE);
+}
+
+static bool jsl_ddi_tc_is_clock_enabled(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	u32 tmp;
+
+	tmp = intel_de_read(i915, DDI_CLK_SEL(port));
+
+	if ((tmp & DDI_CLK_SEL_MASK) == DDI_CLK_SEL_NONE)
+		return false;
+
+	return icl_ddi_combo_is_clock_enabled(encoder);
+}
+
+static void icl_ddi_tc_enable_clock(struct intel_encoder *encoder,
+				    const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	enum tc_port tc_port = intel_port_to_tc(i915, encoder->port);
+	enum port port = encoder->port;
+
+	if (drm_WARN_ON(&i915->drm, !pll))
+		return;
+
+	intel_de_write(i915, DDI_CLK_SEL(port),
+		       icl_pll_to_ddi_clk_sel(encoder, crtc_state));
+
+	mutex_lock(&i915->display.dpll.lock);
+
+	intel_de_rmw(i915, ICL_DPCLKA_CFGCR0,
+		     ICL_DPCLKA_CFGCR0_TC_CLK_OFF(tc_port), 0);
+
+	mutex_unlock(&i915->display.dpll.lock);
+}
+
+static void icl_ddi_tc_disable_clock(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum tc_port tc_port = intel_port_to_tc(i915, encoder->port);
+	enum port port = encoder->port;
+
+	mutex_lock(&i915->display.dpll.lock);
+
+	intel_de_rmw(i915, ICL_DPCLKA_CFGCR0,
+		     0, ICL_DPCLKA_CFGCR0_TC_CLK_OFF(tc_port));
+
+	mutex_unlock(&i915->display.dpll.lock);
+
+	intel_de_write(i915, DDI_CLK_SEL(port), DDI_CLK_SEL_NONE);
+}
+
+static bool icl_ddi_tc_is_clock_enabled(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum tc_port tc_port = intel_port_to_tc(i915, encoder->port);
+	enum port port = encoder->port;
+	u32 tmp;
+
+	tmp = intel_de_read(i915, DDI_CLK_SEL(port));
+
+	if ((tmp & DDI_CLK_SEL_MASK) == DDI_CLK_SEL_NONE)
+		return false;
+
+	tmp = intel_de_read(i915, ICL_DPCLKA_CFGCR0);
+
+	return !(tmp & ICL_DPCLKA_CFGCR0_TC_CLK_OFF(tc_port));
+}
+
+static struct intel_shared_dpll *icl_ddi_tc_get_pll(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum tc_port tc_port = intel_port_to_tc(i915, encoder->port);
+	enum port port = encoder->port;
+	enum intel_dpll_id id;
+	u32 tmp;
+
+	tmp = intel_de_read(i915, DDI_CLK_SEL(port));
+
+	switch (tmp & DDI_CLK_SEL_MASK) {
+	case DDI_CLK_SEL_TBT_162:
+	case DDI_CLK_SEL_TBT_270:
+	case DDI_CLK_SEL_TBT_540:
+	case DDI_CLK_SEL_TBT_810:
+		id = DPLL_ID_ICL_TBTPLL;
+		break;
+	case DDI_CLK_SEL_MG:
+		id = icl_tc_port_to_pll_id(tc_port);
+		break;
+	default:
+		MISSING_CASE(tmp);
+		fallthrough;
+	case DDI_CLK_SEL_NONE:
+		return NULL;
+	}
+
+	return intel_get_shared_dpll_by_id(i915, id);
+}
+
+static struct intel_shared_dpll *bxt_ddi_get_pll(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum intel_dpll_id id;
+
+	switch (encoder->port) {
+	case PORT_A:
+		id = DPLL_ID_SKL_DPLL0;
+		break;
+	case PORT_B:
+		id = DPLL_ID_SKL_DPLL1;
+		break;
+	case PORT_C:
+		id = DPLL_ID_SKL_DPLL2;
+		break;
+	default:
+		MISSING_CASE(encoder->port);
+		return NULL;
+	}
+
+	return intel_get_shared_dpll_by_id(i915, id);
+}
+
+static void skl_ddi_enable_clock(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	enum port port = encoder->port;
+
+	if (drm_WARN_ON(&i915->drm, !pll))
+		return;
+
+	mutex_lock(&i915->display.dpll.lock);
+
+	intel_de_rmw(i915, DPLL_CTRL2,
+		     DPLL_CTRL2_DDI_CLK_OFF(port) |
+		     DPLL_CTRL2_DDI_CLK_SEL_MASK(port),
+		     DPLL_CTRL2_DDI_CLK_SEL(pll->info->id, port) |
+		     DPLL_CTRL2_DDI_SEL_OVERRIDE(port));
+
+	mutex_unlock(&i915->display.dpll.lock);
+}
+
+static void skl_ddi_disable_clock(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+
+	mutex_lock(&i915->display.dpll.lock);
+
+	intel_de_rmw(i915, DPLL_CTRL2,
+		     0, DPLL_CTRL2_DDI_CLK_OFF(port));
+
+	mutex_unlock(&i915->display.dpll.lock);
+}
+
+static bool skl_ddi_is_clock_enabled(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+
+	/*
+	 * FIXME Not sure if the override affects both
+	 * the PLL selection and the CLK_OFF bit.
+	 */
+	return !(intel_de_read(i915, DPLL_CTRL2) & DPLL_CTRL2_DDI_CLK_OFF(port));
+}
+
+static struct intel_shared_dpll *skl_ddi_get_pll(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	enum intel_dpll_id id;
+	u32 tmp;
+
+	tmp = intel_de_read(i915, DPLL_CTRL2);
+
+	/*
+	 * FIXME Not sure if the override affects both
+	 * the PLL selection and the CLK_OFF bit.
+	 */
+	if ((tmp & DPLL_CTRL2_DDI_SEL_OVERRIDE(port)) == 0)
+		return NULL;
+
+	id = (tmp & DPLL_CTRL2_DDI_CLK_SEL_MASK(port)) >>
+		DPLL_CTRL2_DDI_CLK_SEL_SHIFT(port);
+
+	return intel_get_shared_dpll_by_id(i915, id);
+}
+
+void hsw_ddi_enable_clock(struct intel_encoder *encoder,
+			  const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	enum port port = encoder->port;
+
+	if (drm_WARN_ON(&i915->drm, !pll))
+		return;
+
+	intel_de_write(i915, PORT_CLK_SEL(port), hsw_pll_to_ddi_pll_sel(pll));
+}
+
+void hsw_ddi_disable_clock(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+
+	intel_de_write(i915, PORT_CLK_SEL(port), PORT_CLK_SEL_NONE);
+}
+
+bool hsw_ddi_is_clock_enabled(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+
+	return intel_de_read(i915, PORT_CLK_SEL(port)) != PORT_CLK_SEL_NONE;
+}
+
+static struct intel_shared_dpll *hsw_ddi_get_pll(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	enum intel_dpll_id id;
+	u32 tmp;
+
+	tmp = intel_de_read(i915, PORT_CLK_SEL(port));
+
+	switch (tmp & PORT_CLK_SEL_MASK) {
+	case PORT_CLK_SEL_WRPLL1:
+		id = DPLL_ID_WRPLL1;
+		break;
+	case PORT_CLK_SEL_WRPLL2:
+		id = DPLL_ID_WRPLL2;
+		break;
+	case PORT_CLK_SEL_SPLL:
+		id = DPLL_ID_SPLL;
+		break;
+	case PORT_CLK_SEL_LCPLL_810:
+		id = DPLL_ID_LCPLL_810;
+		break;
+	case PORT_CLK_SEL_LCPLL_1350:
+		id = DPLL_ID_LCPLL_1350;
+		break;
+	case PORT_CLK_SEL_LCPLL_2700:
+		id = DPLL_ID_LCPLL_2700;
+		break;
+	default:
+		MISSING_CASE(tmp);
+		fallthrough;
+	case PORT_CLK_SEL_NONE:
+		return NULL;
+	}
+
+	return intel_get_shared_dpll_by_id(i915, id);
+}
+
+void intel_ddi_enable_clock(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state)
+{
+	if (encoder->enable_clock)
+		encoder->enable_clock(encoder, crtc_state);
+}
+
+void intel_ddi_disable_clock(struct intel_encoder *encoder)
+{
+	if (encoder->disable_clock)
+		encoder->disable_clock(encoder);
+}
+
+void intel_ddi_sanitize_encoder_pll_mapping(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	u32 port_mask;
+	bool ddi_clk_needed;
+
+	/*
+	 * In case of DP MST, we sanitize the primary encoder only, not the
+	 * virtual ones.
+	 */
+	if (encoder->type == INTEL_OUTPUT_DP_MST)
+		return;
+
+	if (!encoder->base.crtc && intel_encoder_is_dp(encoder)) {
+		u8 pipe_mask;
+		bool is_mst;
+
+		intel_ddi_get_encoder_pipes(encoder, &pipe_mask, &is_mst);
+		/*
+		 * In the unlikely case that BIOS enables DP in MST mode, just
+		 * warn since our MST HW readout is incomplete.
+		 */
+		if (drm_WARN_ON(&i915->drm, is_mst))
+			return;
+	}
+
+	port_mask = BIT(encoder->port);
+	ddi_clk_needed = encoder->base.crtc;
+
+	if (encoder->type == INTEL_OUTPUT_DSI) {
+		struct intel_encoder *other_encoder;
+
+		port_mask = intel_dsi_encoder_ports(encoder);
+		/*
+		 * Sanity check that we haven't incorrectly registered another
+		 * encoder using any of the ports of this DSI encoder.
+		 */
+		for_each_intel_encoder(&i915->drm, other_encoder) {
+			if (other_encoder == encoder)
+				continue;
+
+			if (drm_WARN_ON(&i915->drm,
+					port_mask & BIT(other_encoder->port)))
+				return;
+		}
+		/*
+		 * For DSI we keep the ddi clocks gated
+		 * except during enable/disable sequence.
+		 */
+		ddi_clk_needed = false;
+	}
+
+	if (ddi_clk_needed || !encoder->is_clock_enabled ||
+	    !encoder->is_clock_enabled(encoder))
+		return;
+
+	drm_notice(&i915->drm,
+		   "[ENCODER:%d:%s] is disabled/in DSI mode with an ungated DDI clock, gate it\n",
+		   encoder->base.base.id, encoder->base.name);
+
+	encoder->disable_clock(encoder);
+}
+
+static void
+icl_program_mg_dp_mode(struct intel_digital_port *dig_port,
+		       const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
+	enum tc_port tc_port = intel_port_to_tc(dev_priv, dig_port->base.port);
+	enum phy phy = intel_port_to_phy(dev_priv, dig_port->base.port);
+	u32 ln0, ln1, pin_assignment;
+	u8 width;
+
+	if (!intel_phy_is_tc(dev_priv, phy) ||
+	    intel_tc_port_in_tbt_alt_mode(dig_port))
+		return;
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		ln0 = intel_dkl_phy_read(dev_priv, DKL_DP_MODE(tc_port), 0);
+		ln1 = intel_dkl_phy_read(dev_priv, DKL_DP_MODE(tc_port), 1);
+	} else {
+		ln0 = intel_de_read(dev_priv, MG_DP_MODE(0, tc_port));
+		ln1 = intel_de_read(dev_priv, MG_DP_MODE(1, tc_port));
+	}
+
+	ln0 &= ~(MG_DP_MODE_CFG_DP_X1_MODE | MG_DP_MODE_CFG_DP_X2_MODE);
+	ln1 &= ~(MG_DP_MODE_CFG_DP_X1_MODE | MG_DP_MODE_CFG_DP_X2_MODE);
+
+	/* DPPATC */
+	pin_assignment = intel_tc_port_get_pin_assignment_mask(dig_port);
+	width = crtc_state->lane_count;
+
+	switch (pin_assignment) {
+	case 0x0:
+		drm_WARN_ON(&dev_priv->drm,
+			    !intel_tc_port_in_legacy_mode(dig_port));
+		if (width == 1) {
+			ln1 |= MG_DP_MODE_CFG_DP_X1_MODE;
+		} else {
+			ln0 |= MG_DP_MODE_CFG_DP_X2_MODE;
+			ln1 |= MG_DP_MODE_CFG_DP_X2_MODE;
+		}
+		break;
+	case 0x1:
+		if (width == 4) {
+			ln0 |= MG_DP_MODE_CFG_DP_X2_MODE;
+			ln1 |= MG_DP_MODE_CFG_DP_X2_MODE;
+		}
+		break;
+	case 0x2:
+		if (width == 2) {
+			ln0 |= MG_DP_MODE_CFG_DP_X2_MODE;
+			ln1 |= MG_DP_MODE_CFG_DP_X2_MODE;
+		}
+		break;
+	case 0x3:
+	case 0x5:
+		if (width == 1) {
+			ln0 |= MG_DP_MODE_CFG_DP_X1_MODE;
+			ln1 |= MG_DP_MODE_CFG_DP_X1_MODE;
+		} else {
+			ln0 |= MG_DP_MODE_CFG_DP_X2_MODE;
+			ln1 |= MG_DP_MODE_CFG_DP_X2_MODE;
+		}
+		break;
+	case 0x4:
+	case 0x6:
+		if (width == 1) {
+			ln0 |= MG_DP_MODE_CFG_DP_X1_MODE;
+			ln1 |= MG_DP_MODE_CFG_DP_X1_MODE;
+		} else {
+			ln0 |= MG_DP_MODE_CFG_DP_X2_MODE;
+			ln1 |= MG_DP_MODE_CFG_DP_X2_MODE;
+		}
+		break;
+	default:
+		MISSING_CASE(pin_assignment);
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		intel_dkl_phy_write(dev_priv, DKL_DP_MODE(tc_port), 0, ln0);
+		intel_dkl_phy_write(dev_priv, DKL_DP_MODE(tc_port), 1, ln1);
+	} else {
+		intel_de_write(dev_priv, MG_DP_MODE(0, tc_port), ln0);
+		intel_de_write(dev_priv, MG_DP_MODE(1, tc_port), ln1);
+	}
+}
+
+static enum transcoder
+tgl_dp_tp_transcoder(const struct intel_crtc_state *crtc_state)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST))
+		return crtc_state->mst_master_transcoder;
+	else
+		return crtc_state->cpu_transcoder;
+}
+
+i915_reg_t dp_tp_ctl_reg(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		return TGL_DP_TP_CTL(tgl_dp_tp_transcoder(crtc_state));
+	else
+		return DP_TP_CTL(encoder->port);
+}
+
+i915_reg_t dp_tp_status_reg(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		return TGL_DP_TP_STATUS(tgl_dp_tp_transcoder(crtc_state));
+	else
+		return DP_TP_STATUS(encoder->port);
+}
+
+static void intel_dp_sink_set_msa_timing_par_ignore_state(struct intel_dp *intel_dp,
+							  const struct intel_crtc_state *crtc_state,
+							  bool enable)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	if (!crtc_state->vrr.enable)
+		return;
+
+	if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_DOWNSPREAD_CTRL,
+			       enable ? DP_MSA_TIMING_PAR_IGNORE_EN : 0) <= 0)
+		drm_dbg_kms(&i915->drm,
+			    "Failed to %s MSA_TIMING_PAR_IGNORE in the sink\n",
+			    str_enable_disable(enable));
+}
+
+static void intel_dp_sink_set_fec_ready(struct intel_dp *intel_dp,
+					const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	if (!crtc_state->fec_enable)
+		return;
+
+	if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_FEC_CONFIGURATION, DP_FEC_READY) <= 0)
+		drm_dbg_kms(&i915->drm,
+			    "Failed to set FEC_READY in the sink\n");
+}
+
+static void intel_ddi_enable_fec(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp;
+	u32 val;
+
+	if (!crtc_state->fec_enable)
+		return;
+
+	intel_dp = enc_to_intel_dp(encoder);
+	val = intel_de_read(dev_priv, dp_tp_ctl_reg(encoder, crtc_state));
+	val |= DP_TP_CTL_FEC_ENABLE;
+	intel_de_write(dev_priv, dp_tp_ctl_reg(encoder, crtc_state), val);
+}
+
+static void intel_ddi_disable_fec_state(struct intel_encoder *encoder,
+					const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp;
+	u32 val;
+
+	if (!crtc_state->fec_enable)
+		return;
+
+	intel_dp = enc_to_intel_dp(encoder);
+	val = intel_de_read(dev_priv, dp_tp_ctl_reg(encoder, crtc_state));
+	val &= ~DP_TP_CTL_FEC_ENABLE;
+	intel_de_write(dev_priv, dp_tp_ctl_reg(encoder, crtc_state), val);
+	intel_de_posting_read(dev_priv, dp_tp_ctl_reg(encoder, crtc_state));
+}
+
+static void intel_ddi_power_up_lanes(struct intel_encoder *encoder,
+				     const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	if (intel_phy_is_combo(i915, phy)) {
+		bool lane_reversal =
+			dig_port->saved_port_bits & DDI_BUF_PORT_REVERSAL;
+
+		intel_combo_phy_power_up_lanes(i915, phy, false,
+					       crtc_state->lane_count,
+					       lane_reversal);
+	}
+}
+
+/* Splitter enable for eDP MSO is limited to certain pipes. */
+static u8 intel_ddi_splitter_pipe_mask(struct drm_i915_private *i915)
+{
+	if (IS_ALDERLAKE_P(i915))
+		return BIT(PIPE_A) | BIT(PIPE_B);
+	else
+		return BIT(PIPE_A);
+}
+
+static void intel_ddi_mso_get_config(struct intel_encoder *encoder,
+				     struct intel_crtc_state *pipe_config)
+{
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	u32 dss1;
+
+	if (!HAS_MSO(i915))
+		return;
+
+	dss1 = intel_de_read(i915, ICL_PIPE_DSS_CTL1(pipe));
+
+	pipe_config->splitter.enable = dss1 & SPLITTER_ENABLE;
+	if (!pipe_config->splitter.enable)
+		return;
+
+	if (drm_WARN_ON(&i915->drm, !(intel_ddi_splitter_pipe_mask(i915) & BIT(pipe)))) {
+		pipe_config->splitter.enable = false;
+		return;
+	}
+
+	switch (dss1 & SPLITTER_CONFIGURATION_MASK) {
+	default:
+		drm_WARN(&i915->drm, true,
+			 "Invalid splitter configuration, dss1=0x%08x\n", dss1);
+		fallthrough;
+	case SPLITTER_CONFIGURATION_2_SEGMENT:
+		pipe_config->splitter.link_count = 2;
+		break;
+	case SPLITTER_CONFIGURATION_4_SEGMENT:
+		pipe_config->splitter.link_count = 4;
+		break;
+	}
+
+	pipe_config->splitter.pixel_overlap = REG_FIELD_GET(OVERLAP_PIXELS_MASK, dss1);
+}
+
+static void intel_ddi_mso_configure(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	u32 dss1 = 0;
+
+	if (!HAS_MSO(i915))
+		return;
+
+	if (crtc_state->splitter.enable) {
+		dss1 |= SPLITTER_ENABLE;
+		dss1 |= OVERLAP_PIXELS(crtc_state->splitter.pixel_overlap);
+		if (crtc_state->splitter.link_count == 2)
+			dss1 |= SPLITTER_CONFIGURATION_2_SEGMENT;
+		else
+			dss1 |= SPLITTER_CONFIGURATION_4_SEGMENT;
+	}
+
+	intel_de_rmw(i915, ICL_PIPE_DSS_CTL1(pipe),
+		     SPLITTER_ENABLE | SPLITTER_CONFIGURATION_MASK |
+		     OVERLAP_PIXELS_MASK, dss1);
+}
+
+static void tgl_ddi_pre_enable_dp(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	bool is_mst = intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST);
+
+	intel_dp_set_link_params(intel_dp,
+				 crtc_state->port_clock,
+				 crtc_state->lane_count);
+
+	/*
+	 * We only configure what the register value will be here.  Actual
+	 * enabling happens during link training farther down.
+	 */
+	intel_ddi_init_dp_buf_reg(encoder, crtc_state);
+
+	/*
+	 * 1. Enable Power Wells
+	 *
+	 * This was handled at the beginning of intel_atomic_commit_tail(),
+	 * before we called down into this function.
+	 */
+
+	/* 2. Enable Panel Power if PPS is required */
+	intel_pps_on(intel_dp);
+
+	/*
+	 * 3. For non-TBT Type-C ports, set FIA lane count
+	 * (DFLEXDPSP.DPX4TXLATC)
+	 *
+	 * This was done before tgl_ddi_pre_enable_dp by
+	 * hsw_crtc_enable()->intel_encoders_pre_pll_enable().
+	 */
+
+	/*
+	 * 4. Enable the port PLL.
+	 *
+	 * The PLL enabling itself was already done before this function by
+	 * hsw_crtc_enable()->intel_enable_shared_dpll().  We need only
+	 * configure the PLL to port mapping here.
+	 */
+	intel_ddi_enable_clock(encoder, crtc_state);
+
+	/* 5. If IO power is controlled through PWR_WELL_CTL, Enable IO Power */
+	if (!intel_tc_port_in_tbt_alt_mode(dig_port)) {
+		drm_WARN_ON(&dev_priv->drm, dig_port->ddi_io_wakeref);
+		dig_port->ddi_io_wakeref = intel_display_power_get(dev_priv,
+								   dig_port->ddi_io_power_domain);
+	}
+
+	/* 6. Program DP_MODE */
+	icl_program_mg_dp_mode(dig_port, crtc_state);
+
+	/*
+	 * 7. The rest of the below are substeps under the bspec's "Enable and
+	 * Train Display Port" step.  Note that steps that are specific to
+	 * MST will be handled by intel_mst_pre_enable_dp() before/after it
+	 * calls into this function.  Also intel_mst_pre_enable_dp() only calls
+	 * us when active_mst_links==0, so any steps designated for "single
+	 * stream or multi-stream master transcoder" can just be performed
+	 * unconditionally here.
+	 */
+
+	/*
+	 * 7.a Configure Transcoder Clock Select to direct the Port clock to the
+	 * Transcoder.
+	 */
+	intel_ddi_enable_pipe_clock(encoder, crtc_state);
+
+	if (HAS_DP20(dev_priv))
+		intel_ddi_config_transcoder_dp2(encoder, crtc_state);
+
+	/*
+	 * 7.b Configure TRANS_DDI_FUNC_CTL DDI Select, DDI Mode Select & MST
+	 * Transport Select
+	 */
+	intel_ddi_config_transcoder_func(encoder, crtc_state);
+
+	/*
+	 * 7.c Configure & enable DP_TP_CTL with link training pattern 1
+	 * selected
+	 *
+	 * This will be handled by the intel_dp_start_link_train() farther
+	 * down this function.
+	 */
+
+	/* 7.e Configure voltage swing and related IO settings */
+	encoder->set_signal_levels(encoder, crtc_state);
+
+	/*
+	 * 7.f Combo PHY: Configure PORT_CL_DW10 Static Power Down to power up
+	 * the used lanes of the DDI.
+	 */
+	intel_ddi_power_up_lanes(encoder, crtc_state);
+
+	/*
+	 * 7.g Program CoG/MSO configuration bits in DSS_CTL1 if selected.
+	 */
+	intel_ddi_mso_configure(crtc_state);
+
+	if (!is_mst)
+		intel_dp_set_power(intel_dp, DP_SET_POWER_D0);
+
+	intel_dp_configure_protocol_converter(intel_dp, crtc_state);
+	intel_dp_sink_set_decompression_state(intel_dp, crtc_state, true);
+	/*
+	 * DDI FEC: "anticipates enabling FEC encoding sets the FEC_READY bit
+	 * in the FEC_CONFIGURATION register to 1 before initiating link
+	 * training
+	 */
+	intel_dp_sink_set_fec_ready(intel_dp, crtc_state);
+
+	intel_dp_check_frl_training(intel_dp);
+	intel_dp_pcon_dsc_configure(intel_dp, crtc_state);
+
+	/*
+	 * 7.i Follow DisplayPort specification training sequence (see notes for
+	 *     failure handling)
+	 * 7.j If DisplayPort multi-stream - Set DP_TP_CTL link training to Idle
+	 *     Pattern, wait for 5 idle patterns (DP_TP_STATUS Min_Idles_Sent)
+	 *     (timeout after 800 us)
+	 */
+	intel_dp_start_link_train(intel_dp, crtc_state);
+
+	/* 7.k Set DP_TP_CTL link training to Normal */
+	if (!is_trans_port_sync_mode(crtc_state))
+		intel_dp_stop_link_train(intel_dp, crtc_state);
+
+	/* 7.l Configure and enable FEC if needed */
+	intel_ddi_enable_fec(encoder, crtc_state);
+
+	intel_dsc_dp_pps_write(encoder, crtc_state);
+}
+
+static void hsw_ddi_pre_enable_dp(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	bool is_mst = intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST);
+
+	if (DISPLAY_VER(dev_priv) < 11)
+		drm_WARN_ON(&dev_priv->drm,
+			    is_mst && (port == PORT_A || port == PORT_E));
+	else
+		drm_WARN_ON(&dev_priv->drm, is_mst && port == PORT_A);
+
+	intel_dp_set_link_params(intel_dp,
+				 crtc_state->port_clock,
+				 crtc_state->lane_count);
+
+	/*
+	 * We only configure what the register value will be here.  Actual
+	 * enabling happens during link training farther down.
+	 */
+	intel_ddi_init_dp_buf_reg(encoder, crtc_state);
+
+	intel_pps_on(intel_dp);
+
+	intel_ddi_enable_clock(encoder, crtc_state);
+
+	if (!intel_tc_port_in_tbt_alt_mode(dig_port)) {
+		drm_WARN_ON(&dev_priv->drm, dig_port->ddi_io_wakeref);
+		dig_port->ddi_io_wakeref = intel_display_power_get(dev_priv,
+								   dig_port->ddi_io_power_domain);
+	}
+
+	icl_program_mg_dp_mode(dig_port, crtc_state);
+
+	if (has_buf_trans_select(dev_priv))
+		hsw_prepare_dp_ddi_buffers(encoder, crtc_state);
+
+	encoder->set_signal_levels(encoder, crtc_state);
+
+	intel_ddi_power_up_lanes(encoder, crtc_state);
+
+	if (!is_mst)
+		intel_dp_set_power(intel_dp, DP_SET_POWER_D0);
+	intel_dp_configure_protocol_converter(intel_dp, crtc_state);
+	intel_dp_sink_set_decompression_state(intel_dp, crtc_state,
+					      true);
+	intel_dp_sink_set_fec_ready(intel_dp, crtc_state);
+	intel_dp_start_link_train(intel_dp, crtc_state);
+	if ((port != PORT_A || DISPLAY_VER(dev_priv) >= 9) &&
+	    !is_trans_port_sync_mode(crtc_state))
+		intel_dp_stop_link_train(intel_dp, crtc_state);
+
+	intel_ddi_enable_fec(encoder, crtc_state);
+
+	if (!is_mst)
+		intel_ddi_enable_pipe_clock(encoder, crtc_state);
+
+	intel_dsc_dp_pps_write(encoder, crtc_state);
+}
+
+static void intel_ddi_pre_enable_dp(struct intel_atomic_state *state,
+				    struct intel_encoder *encoder,
+				    const struct intel_crtc_state *crtc_state,
+				    const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		tgl_ddi_pre_enable_dp(state, encoder, crtc_state, conn_state);
+	else
+		hsw_ddi_pre_enable_dp(state, encoder, crtc_state, conn_state);
+
+	/* MST will call a setting of MSA after an allocating of Virtual Channel
+	 * from MST encoder pre_enable callback.
+	 */
+	if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST))
+		intel_ddi_set_dp_msa(crtc_state, conn_state);
+}
+
+static void intel_ddi_pre_enable_hdmi(struct intel_atomic_state *state,
+				      struct intel_encoder *encoder,
+				      const struct intel_crtc_state *crtc_state,
+				      const struct drm_connector_state *conn_state)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_hdmi *intel_hdmi = &dig_port->hdmi;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	intel_dp_dual_mode_set_tmds_output(intel_hdmi, true);
+	intel_ddi_enable_clock(encoder, crtc_state);
+
+	drm_WARN_ON(&dev_priv->drm, dig_port->ddi_io_wakeref);
+	dig_port->ddi_io_wakeref = intel_display_power_get(dev_priv,
+							   dig_port->ddi_io_power_domain);
+
+	icl_program_mg_dp_mode(dig_port, crtc_state);
+
+	intel_ddi_enable_pipe_clock(encoder, crtc_state);
+
+	dig_port->set_infoframes(encoder,
+				 crtc_state->has_infoframe,
+				 crtc_state, conn_state);
+}
+
+static void intel_ddi_pre_enable(struct intel_atomic_state *state,
+				 struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state,
+				 const struct drm_connector_state *conn_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	/*
+	 * When called from DP MST code:
+	 * - conn_state will be NULL
+	 * - encoder will be the main encoder (ie. mst->primary)
+	 * - the main connector associated with this port
+	 *   won't be active or linked to a crtc
+	 * - crtc_state will be the state of the first stream to
+	 *   be activated on this port, and it may not be the same
+	 *   stream that will be deactivated last, but each stream
+	 *   should have a state that is identical when it comes to
+	 *   the DP link parameteres
+	 */
+
+	drm_WARN_ON(&dev_priv->drm, crtc_state->has_pch_encoder);
+
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI)) {
+		intel_ddi_pre_enable_hdmi(state, encoder, crtc_state,
+					  conn_state);
+	} else {
+		struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+
+		intel_ddi_pre_enable_dp(state, encoder, crtc_state,
+					conn_state);
+
+		/* FIXME precompute everything properly */
+		/* FIXME how do we turn infoframes off again? */
+		if (dig_port->lspcon.active && dig_port->dp.has_hdmi_sink)
+			dig_port->set_infoframes(encoder,
+						 crtc_state->has_infoframe,
+						 crtc_state, conn_state);
+	}
+}
+
+static void intel_disable_ddi_buf(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	bool wait = false;
+	u32 val;
+
+	val = intel_de_read(dev_priv, DDI_BUF_CTL(port));
+	if (val & DDI_BUF_CTL_ENABLE) {
+		val &= ~DDI_BUF_CTL_ENABLE;
+		intel_de_write(dev_priv, DDI_BUF_CTL(port), val);
+		wait = true;
+	}
+
+	if (intel_crtc_has_dp_encoder(crtc_state)) {
+		val = intel_de_read(dev_priv, dp_tp_ctl_reg(encoder, crtc_state));
+		val &= ~(DP_TP_CTL_ENABLE | DP_TP_CTL_LINK_TRAIN_MASK);
+		val |= DP_TP_CTL_LINK_TRAIN_PAT1;
+		intel_de_write(dev_priv, dp_tp_ctl_reg(encoder, crtc_state), val);
+	}
+
+	/* Disable FEC in DP Sink */
+	intel_ddi_disable_fec_state(encoder, crtc_state);
+
+	if (wait)
+		intel_wait_ddi_buf_idle(dev_priv, port);
+}
+
+static void intel_ddi_post_disable_dp(struct intel_atomic_state *state,
+				      struct intel_encoder *encoder,
+				      const struct intel_crtc_state *old_crtc_state,
+				      const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_dp *intel_dp = &dig_port->dp;
+	bool is_mst = intel_crtc_has_type(old_crtc_state,
+					  INTEL_OUTPUT_DP_MST);
+
+	if (!is_mst)
+		intel_dp_set_infoframes(encoder, false,
+					old_crtc_state, old_conn_state);
+
+	/*
+	 * Power down sink before disabling the port, otherwise we end
+	 * up getting interrupts from the sink on detecting link loss.
+	 */
+	intel_dp_set_power(intel_dp, DP_SET_POWER_D3);
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		if (is_mst) {
+			enum transcoder cpu_transcoder = old_crtc_state->cpu_transcoder;
+			u32 val;
+
+			val = intel_de_read(dev_priv,
+					    TRANS_DDI_FUNC_CTL(cpu_transcoder));
+			val &= ~(TGL_TRANS_DDI_PORT_MASK |
+				 TRANS_DDI_MODE_SELECT_MASK);
+			intel_de_write(dev_priv,
+				       TRANS_DDI_FUNC_CTL(cpu_transcoder),
+				       val);
+		}
+	} else {
+		if (!is_mst)
+			intel_ddi_disable_pipe_clock(old_crtc_state);
+	}
+
+	intel_disable_ddi_buf(encoder, old_crtc_state);
+
+	/*
+	 * From TGL spec: "If single stream or multi-stream master transcoder:
+	 * Configure Transcoder Clock select to direct no clock to the
+	 * transcoder"
+	 */
+	if (DISPLAY_VER(dev_priv) >= 12)
+		intel_ddi_disable_pipe_clock(old_crtc_state);
+
+	intel_pps_vdd_on(intel_dp);
+	intel_pps_off(intel_dp);
+
+	if (!intel_tc_port_in_tbt_alt_mode(dig_port))
+		intel_display_power_put(dev_priv,
+					dig_port->ddi_io_power_domain,
+					fetch_and_zero(&dig_port->ddi_io_wakeref));
+
+	intel_ddi_disable_clock(encoder);
+}
+
+static void intel_ddi_post_disable_hdmi(struct intel_atomic_state *state,
+					struct intel_encoder *encoder,
+					const struct intel_crtc_state *old_crtc_state,
+					const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_hdmi *intel_hdmi = &dig_port->hdmi;
+
+	dig_port->set_infoframes(encoder, false,
+				 old_crtc_state, old_conn_state);
+
+	if (DISPLAY_VER(dev_priv) < 12)
+		intel_ddi_disable_pipe_clock(old_crtc_state);
+
+	intel_disable_ddi_buf(encoder, old_crtc_state);
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		intel_ddi_disable_pipe_clock(old_crtc_state);
+
+	intel_display_power_put(dev_priv,
+				dig_port->ddi_io_power_domain,
+				fetch_and_zero(&dig_port->ddi_io_wakeref));
+
+	intel_ddi_disable_clock(encoder);
+
+	intel_dp_dual_mode_set_tmds_output(intel_hdmi, false);
+}
+
+static void intel_ddi_post_disable(struct intel_atomic_state *state,
+				   struct intel_encoder *encoder,
+				   const struct intel_crtc_state *old_crtc_state,
+				   const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+	bool is_tc_port = intel_phy_is_tc(dev_priv, phy);
+	struct intel_crtc *slave_crtc;
+
+	if (!intel_crtc_has_type(old_crtc_state, INTEL_OUTPUT_DP_MST)) {
+		intel_crtc_vblank_off(old_crtc_state);
+
+		intel_disable_transcoder(old_crtc_state);
+
+		intel_vrr_disable(old_crtc_state);
+
+		intel_ddi_disable_transcoder_func(old_crtc_state);
+
+		intel_dsc_disable(old_crtc_state);
+
+		if (DISPLAY_VER(dev_priv) >= 9)
+			skl_scaler_disable(old_crtc_state);
+		else
+			ilk_pfit_disable(old_crtc_state);
+	}
+
+	for_each_intel_crtc_in_pipe_mask(&dev_priv->drm, slave_crtc,
+					 intel_crtc_bigjoiner_slave_pipes(old_crtc_state)) {
+		const struct intel_crtc_state *old_slave_crtc_state =
+			intel_atomic_get_old_crtc_state(state, slave_crtc);
+
+		intel_crtc_vblank_off(old_slave_crtc_state);
+
+		intel_dsc_disable(old_slave_crtc_state);
+		skl_scaler_disable(old_slave_crtc_state);
+	}
+
+	/*
+	 * When called from DP MST code:
+	 * - old_conn_state will be NULL
+	 * - encoder will be the main encoder (ie. mst->primary)
+	 * - the main connector associated with this port
+	 *   won't be active or linked to a crtc
+	 * - old_crtc_state will be the state of the last stream to
+	 *   be deactivated on this port, and it may not be the same
+	 *   stream that was activated last, but each stream
+	 *   should have a state that is identical when it comes to
+	 *   the DP link parameteres
+	 */
+
+	if (intel_crtc_has_type(old_crtc_state, INTEL_OUTPUT_HDMI))
+		intel_ddi_post_disable_hdmi(state, encoder, old_crtc_state,
+					    old_conn_state);
+	else
+		intel_ddi_post_disable_dp(state, encoder, old_crtc_state,
+					  old_conn_state);
+
+	if (intel_crtc_has_dp_encoder(old_crtc_state) || is_tc_port)
+		intel_display_power_put(dev_priv,
+					intel_ddi_main_link_aux_domain(dig_port),
+					fetch_and_zero(&dig_port->aux_wakeref));
+
+	if (is_tc_port)
+		intel_tc_port_put_link(dig_port);
+}
+
+static void trans_port_sync_stop_link_train(struct intel_atomic_state *state,
+					    struct intel_encoder *encoder,
+					    const struct intel_crtc_state *crtc_state)
+{
+	const struct drm_connector_state *conn_state;
+	struct drm_connector *conn;
+	int i;
+
+	if (!crtc_state->sync_mode_slaves_mask)
+		return;
+
+	for_each_new_connector_in_state(&state->base, conn, conn_state, i) {
+		struct intel_encoder *slave_encoder =
+			to_intel_encoder(conn_state->best_encoder);
+		struct intel_crtc *slave_crtc = to_intel_crtc(conn_state->crtc);
+		const struct intel_crtc_state *slave_crtc_state;
+
+		if (!slave_crtc)
+			continue;
+
+		slave_crtc_state =
+			intel_atomic_get_new_crtc_state(state, slave_crtc);
+
+		if (slave_crtc_state->master_transcoder !=
+		    crtc_state->cpu_transcoder)
+			continue;
+
+		intel_dp_stop_link_train(enc_to_intel_dp(slave_encoder),
+					 slave_crtc_state);
+	}
+
+	usleep_range(200, 400);
+
+	intel_dp_stop_link_train(enc_to_intel_dp(encoder),
+				 crtc_state);
+}
+
+static void intel_enable_ddi_dp(struct intel_atomic_state *state,
+				struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state,
+				const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	enum port port = encoder->port;
+
+	if (port == PORT_A && DISPLAY_VER(dev_priv) < 9)
+		intel_dp_stop_link_train(intel_dp, crtc_state);
+
+	drm_connector_update_privacy_screen(conn_state);
+	intel_edp_backlight_on(crtc_state, conn_state);
+
+	if (!dig_port->lspcon.active || dig_port->dp.has_hdmi_sink)
+		intel_dp_set_infoframes(encoder, true, crtc_state, conn_state);
+
+	intel_audio_codec_enable(encoder, crtc_state, conn_state);
+
+	trans_port_sync_stop_link_train(state, encoder, crtc_state);
+}
+
+static i915_reg_t
+gen9_chicken_trans_reg_by_port(struct drm_i915_private *dev_priv,
+			       enum port port)
+{
+	static const enum transcoder trans[] = {
+		[PORT_A] = TRANSCODER_EDP,
+		[PORT_B] = TRANSCODER_A,
+		[PORT_C] = TRANSCODER_B,
+		[PORT_D] = TRANSCODER_C,
+		[PORT_E] = TRANSCODER_A,
+	};
+
+	drm_WARN_ON(&dev_priv->drm, DISPLAY_VER(dev_priv) < 9);
+
+	if (drm_WARN_ON(&dev_priv->drm, port < PORT_A || port > PORT_E))
+		port = PORT_A;
+
+	return CHICKEN_TRANS(trans[port]);
+}
+
+static void intel_enable_ddi_hdmi(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_connector *connector = conn_state->connector;
+	enum port port = encoder->port;
+	enum phy phy = intel_port_to_phy(dev_priv, port);
+	u32 buf_ctl;
+
+	if (!intel_hdmi_handle_sink_scrambling(encoder, connector,
+					       crtc_state->hdmi_high_tmds_clock_ratio,
+					       crtc_state->hdmi_scrambling))
+		drm_dbg_kms(&dev_priv->drm,
+			    "[CONNECTOR:%d:%s] Failed to configure sink scrambling/TMDS bit clock ratio\n",
+			    connector->base.id, connector->name);
+
+	if (has_buf_trans_select(dev_priv))
+		hsw_prepare_hdmi_ddi_buffers(encoder, crtc_state);
+
+	encoder->set_signal_levels(encoder, crtc_state);
+
+	/* Display WA #1143: skl,kbl,cfl */
+	if (DISPLAY_VER(dev_priv) == 9 && !IS_BROXTON(dev_priv)) {
+		/*
+		 * For some reason these chicken bits have been
+		 * stuffed into a transcoder register, event though
+		 * the bits affect a specific DDI port rather than
+		 * a specific transcoder.
+		 */
+		i915_reg_t reg = gen9_chicken_trans_reg_by_port(dev_priv, port);
+		u32 val;
+
+		val = intel_de_read(dev_priv, reg);
+
+		if (port == PORT_E)
+			val |= DDIE_TRAINING_OVERRIDE_ENABLE |
+				DDIE_TRAINING_OVERRIDE_VALUE;
+		else
+			val |= DDI_TRAINING_OVERRIDE_ENABLE |
+				DDI_TRAINING_OVERRIDE_VALUE;
+
+		intel_de_write(dev_priv, reg, val);
+		intel_de_posting_read(dev_priv, reg);
+
+		udelay(1);
+
+		if (port == PORT_E)
+			val &= ~(DDIE_TRAINING_OVERRIDE_ENABLE |
+				 DDIE_TRAINING_OVERRIDE_VALUE);
+		else
+			val &= ~(DDI_TRAINING_OVERRIDE_ENABLE |
+				 DDI_TRAINING_OVERRIDE_VALUE);
+
+		intel_de_write(dev_priv, reg, val);
+	}
+
+	intel_ddi_power_up_lanes(encoder, crtc_state);
+
+	/* In HDMI/DVI mode, the port width, and swing/emphasis values
+	 * are ignored so nothing special needs to be done besides
+	 * enabling the port.
+	 *
+	 * On ADL_P the PHY link rate and lane count must be programmed but
+	 * these are both 0 for HDMI.
+	 */
+	buf_ctl = dig_port->saved_port_bits | DDI_BUF_CTL_ENABLE;
+	if (IS_ALDERLAKE_P(dev_priv) && intel_phy_is_tc(dev_priv, phy)) {
+		drm_WARN_ON(&dev_priv->drm, !intel_tc_port_in_legacy_mode(dig_port));
+		buf_ctl |= DDI_BUF_CTL_TC_PHY_OWNERSHIP;
+	}
+	intel_de_write(dev_priv, DDI_BUF_CTL(port), buf_ctl);
+
+	intel_audio_codec_enable(encoder, crtc_state, conn_state);
+}
+
+static void intel_enable_ddi(struct intel_atomic_state *state,
+			     struct intel_encoder *encoder,
+			     const struct intel_crtc_state *crtc_state,
+			     const struct drm_connector_state *conn_state)
+{
+	drm_WARN_ON(state->base.dev, crtc_state->has_pch_encoder);
+
+	if (!intel_crtc_is_bigjoiner_slave(crtc_state))
+		intel_ddi_enable_transcoder_func(encoder, crtc_state);
+
+	intel_vrr_enable(encoder, crtc_state);
+
+	intel_enable_transcoder(crtc_state);
+
+	intel_crtc_vblank_on(crtc_state);
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		intel_enable_ddi_hdmi(state, encoder, crtc_state, conn_state);
+	else
+		intel_enable_ddi_dp(state, encoder, crtc_state, conn_state);
+
+	/* Enable hdcp if it's desired */
+	if (conn_state->content_protection ==
+	    DRM_MODE_CONTENT_PROTECTION_DESIRED)
+		intel_hdcp_enable(to_intel_connector(conn_state->connector),
+				  crtc_state,
+				  (u8)conn_state->hdcp_content_type);
+}
+
+static void intel_disable_ddi_dp(struct intel_atomic_state *state,
+				 struct intel_encoder *encoder,
+				 const struct intel_crtc_state *old_crtc_state,
+				 const struct drm_connector_state *old_conn_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+	intel_dp->link_trained = false;
+
+	intel_audio_codec_disable(encoder, old_crtc_state, old_conn_state);
+
+	intel_psr_disable(intel_dp, old_crtc_state);
+	intel_edp_backlight_off(old_conn_state);
+	/* Disable the decompression in DP Sink */
+	intel_dp_sink_set_decompression_state(intel_dp, old_crtc_state,
+					      false);
+	/* Disable Ignore_MSA bit in DP Sink */
+	intel_dp_sink_set_msa_timing_par_ignore_state(intel_dp, old_crtc_state,
+						      false);
+}
+
+static void intel_disable_ddi_hdmi(struct intel_atomic_state *state,
+				   struct intel_encoder *encoder,
+				   const struct intel_crtc_state *old_crtc_state,
+				   const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct drm_connector *connector = old_conn_state->connector;
+
+	intel_audio_codec_disable(encoder, old_crtc_state, old_conn_state);
+
+	if (!intel_hdmi_handle_sink_scrambling(encoder, connector,
+					       false, false))
+		drm_dbg_kms(&i915->drm,
+			    "[CONNECTOR:%d:%s] Failed to reset sink scrambling/TMDS bit clock ratio\n",
+			    connector->base.id, connector->name);
+}
+
+static void intel_disable_ddi(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *old_crtc_state,
+			      const struct drm_connector_state *old_conn_state)
+{
+	intel_hdcp_disable(to_intel_connector(old_conn_state->connector));
+
+	if (intel_crtc_has_type(old_crtc_state, INTEL_OUTPUT_HDMI))
+		intel_disable_ddi_hdmi(state, encoder, old_crtc_state,
+				       old_conn_state);
+	else
+		intel_disable_ddi_dp(state, encoder, old_crtc_state,
+				     old_conn_state);
+}
+
+static void intel_ddi_update_pipe_dp(struct intel_atomic_state *state,
+				     struct intel_encoder *encoder,
+				     const struct intel_crtc_state *crtc_state,
+				     const struct drm_connector_state *conn_state)
+{
+	intel_ddi_set_dp_msa(crtc_state, conn_state);
+
+	intel_dp_set_infoframes(encoder, true, crtc_state, conn_state);
+
+	intel_backlight_update(state, encoder, crtc_state, conn_state);
+	drm_connector_update_privacy_screen(conn_state);
+}
+
+void intel_ddi_update_pipe(struct intel_atomic_state *state,
+			   struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state,
+			   const struct drm_connector_state *conn_state)
+{
+
+	if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI) &&
+	    !intel_encoder_is_mst(encoder))
+		intel_ddi_update_pipe_dp(state, encoder, crtc_state,
+					 conn_state);
+
+	intel_hdcp_update_pipe(state, encoder, crtc_state, conn_state);
+}
+
+static void
+intel_ddi_update_prepare(struct intel_atomic_state *state,
+			 struct intel_encoder *encoder,
+			 struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		crtc ? intel_atomic_get_new_crtc_state(state, crtc) : NULL;
+	int required_lanes = crtc_state ? crtc_state->lane_count : 1;
+
+	drm_WARN_ON(state->base.dev, crtc && crtc->active);
+
+	intel_tc_port_get_link(enc_to_dig_port(encoder),
+		               required_lanes);
+	if (crtc_state && crtc_state->hw.active) {
+		struct intel_crtc *slave_crtc;
+
+		intel_update_active_dpll(state, crtc, encoder);
+
+		for_each_intel_crtc_in_pipe_mask(&i915->drm, slave_crtc,
+						 intel_crtc_bigjoiner_slave_pipes(crtc_state))
+			intel_update_active_dpll(state, slave_crtc, encoder);
+	}
+}
+
+static void
+intel_ddi_update_complete(struct intel_atomic_state *state,
+			  struct intel_encoder *encoder,
+			  struct intel_crtc *crtc)
+{
+	intel_tc_port_put_link(enc_to_dig_port(encoder));
+}
+
+static void
+intel_ddi_pre_pll_enable(struct intel_atomic_state *state,
+			 struct intel_encoder *encoder,
+			 const struct intel_crtc_state *crtc_state,
+			 const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+	bool is_tc_port = intel_phy_is_tc(dev_priv, phy);
+
+	if (is_tc_port)
+		intel_tc_port_get_link(dig_port, crtc_state->lane_count);
+
+	if (intel_crtc_has_dp_encoder(crtc_state) || is_tc_port) {
+		drm_WARN_ON(&dev_priv->drm, dig_port->aux_wakeref);
+		dig_port->aux_wakeref =
+			intel_display_power_get(dev_priv,
+						intel_ddi_main_link_aux_domain(dig_port));
+	}
+
+	if (is_tc_port && !intel_tc_port_in_tbt_alt_mode(dig_port))
+		/*
+		 * Program the lane count for static/dynamic connections on
+		 * Type-C ports.  Skip this step for TBT.
+		 */
+		intel_tc_port_set_fia_lane_count(dig_port, crtc_state->lane_count);
+	else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		bxt_ddi_phy_set_lane_optim_mask(encoder,
+						crtc_state->lane_lat_optim_mask);
+}
+
+static void adlp_tbt_to_dp_alt_switch_wa(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum tc_port tc_port = intel_port_to_tc(i915, encoder->port);
+	int ln;
+
+	for (ln = 0; ln < 2; ln++)
+		intel_dkl_phy_rmw(i915, DKL_PCS_DW5(tc_port), ln, DKL_PCS_DW5_CORE_SOFTRESET, 0);
+}
+
+static void intel_ddi_prepare_link_retrain(struct intel_dp *intel_dp,
+					   const struct intel_crtc_state *crtc_state)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &dig_port->base;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	u32 dp_tp_ctl, ddi_buf_ctl;
+	bool wait = false;
+
+	dp_tp_ctl = intel_de_read(dev_priv, dp_tp_ctl_reg(encoder, crtc_state));
+
+	if (dp_tp_ctl & DP_TP_CTL_ENABLE) {
+		ddi_buf_ctl = intel_de_read(dev_priv, DDI_BUF_CTL(port));
+		if (ddi_buf_ctl & DDI_BUF_CTL_ENABLE) {
+			intel_de_write(dev_priv, DDI_BUF_CTL(port),
+				       ddi_buf_ctl & ~DDI_BUF_CTL_ENABLE);
+			wait = true;
+		}
+
+		dp_tp_ctl &= ~(DP_TP_CTL_ENABLE | DP_TP_CTL_LINK_TRAIN_MASK);
+		dp_tp_ctl |= DP_TP_CTL_LINK_TRAIN_PAT1;
+		intel_de_write(dev_priv, dp_tp_ctl_reg(encoder, crtc_state), dp_tp_ctl);
+		intel_de_posting_read(dev_priv, dp_tp_ctl_reg(encoder, crtc_state));
+
+		if (wait)
+			intel_wait_ddi_buf_idle(dev_priv, port);
+	}
+
+	dp_tp_ctl = DP_TP_CTL_ENABLE | DP_TP_CTL_LINK_TRAIN_PAT1;
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST)) {
+		dp_tp_ctl |= DP_TP_CTL_MODE_MST;
+	} else {
+		dp_tp_ctl |= DP_TP_CTL_MODE_SST;
+		if (drm_dp_enhanced_frame_cap(intel_dp->dpcd))
+			dp_tp_ctl |= DP_TP_CTL_ENHANCED_FRAME_ENABLE;
+	}
+	intel_de_write(dev_priv, dp_tp_ctl_reg(encoder, crtc_state), dp_tp_ctl);
+	intel_de_posting_read(dev_priv, dp_tp_ctl_reg(encoder, crtc_state));
+
+	if (IS_ALDERLAKE_P(dev_priv) &&
+	    (intel_tc_port_in_dp_alt_mode(dig_port) || intel_tc_port_in_legacy_mode(dig_port)))
+		adlp_tbt_to_dp_alt_switch_wa(encoder);
+
+	intel_dp->DP |= DDI_BUF_CTL_ENABLE;
+	intel_de_write(dev_priv, DDI_BUF_CTL(port), intel_dp->DP);
+	intel_de_posting_read(dev_priv, DDI_BUF_CTL(port));
+
+	intel_wait_ddi_buf_active(dev_priv, port);
+}
+
+static void intel_ddi_set_link_train(struct intel_dp *intel_dp,
+				     const struct intel_crtc_state *crtc_state,
+				     u8 dp_train_pat)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 temp;
+
+	temp = intel_de_read(dev_priv, dp_tp_ctl_reg(encoder, crtc_state));
+
+	temp &= ~DP_TP_CTL_LINK_TRAIN_MASK;
+	switch (intel_dp_training_pattern_symbol(dp_train_pat)) {
+	case DP_TRAINING_PATTERN_DISABLE:
+		temp |= DP_TP_CTL_LINK_TRAIN_NORMAL;
+		break;
+	case DP_TRAINING_PATTERN_1:
+		temp |= DP_TP_CTL_LINK_TRAIN_PAT1;
+		break;
+	case DP_TRAINING_PATTERN_2:
+		temp |= DP_TP_CTL_LINK_TRAIN_PAT2;
+		break;
+	case DP_TRAINING_PATTERN_3:
+		temp |= DP_TP_CTL_LINK_TRAIN_PAT3;
+		break;
+	case DP_TRAINING_PATTERN_4:
+		temp |= DP_TP_CTL_LINK_TRAIN_PAT4;
+		break;
+	}
+
+	intel_de_write(dev_priv, dp_tp_ctl_reg(encoder, crtc_state), temp);
+}
+
+static void intel_ddi_set_idle_link_train(struct intel_dp *intel_dp,
+					  const struct intel_crtc_state *crtc_state)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	u32 val;
+
+	val = intel_de_read(dev_priv, dp_tp_ctl_reg(encoder, crtc_state));
+	val &= ~DP_TP_CTL_LINK_TRAIN_MASK;
+	val |= DP_TP_CTL_LINK_TRAIN_IDLE;
+	intel_de_write(dev_priv, dp_tp_ctl_reg(encoder, crtc_state), val);
+
+	/*
+	 * Until TGL on PORT_A we can have only eDP in SST mode. There the only
+	 * reason we need to set idle transmission mode is to work around a HW
+	 * issue where we enable the pipe while not in idle link-training mode.
+	 * In this case there is requirement to wait for a minimum number of
+	 * idle patterns to be sent.
+	 */
+	if (port == PORT_A && DISPLAY_VER(dev_priv) < 12)
+		return;
+
+	if (intel_de_wait_for_set(dev_priv,
+				  dp_tp_status_reg(encoder, crtc_state),
+				  DP_TP_STATUS_IDLE_DONE, 1))
+		drm_err(&dev_priv->drm,
+			"Timed out waiting for DP idle patterns\n");
+}
+
+static bool intel_ddi_is_audio_enabled(struct drm_i915_private *dev_priv,
+				       enum transcoder cpu_transcoder)
+{
+	if (cpu_transcoder == TRANSCODER_EDP)
+		return false;
+
+	if (!intel_display_power_is_enabled(dev_priv, POWER_DOMAIN_AUDIO_MMIO))
+		return false;
+
+	return intel_de_read(dev_priv, HSW_AUD_PIN_ELD_CP_VLD) &
+		AUDIO_OUTPUT_ENABLE(cpu_transcoder);
+}
+
+void intel_ddi_compute_min_voltage_level(struct drm_i915_private *dev_priv,
+					 struct intel_crtc_state *crtc_state)
+{
+	if (DISPLAY_VER(dev_priv) >= 12 && crtc_state->port_clock > 594000)
+		crtc_state->min_voltage_level = 2;
+	else if (IS_JSL_EHL(dev_priv) && crtc_state->port_clock > 594000)
+		crtc_state->min_voltage_level = 3;
+	else if (DISPLAY_VER(dev_priv) >= 11 && crtc_state->port_clock > 594000)
+		crtc_state->min_voltage_level = 1;
+}
+
+static enum transcoder bdw_transcoder_master_readout(struct drm_i915_private *dev_priv,
+						     enum transcoder cpu_transcoder)
+{
+	u32 master_select;
+
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		u32 ctl2 = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL2(cpu_transcoder));
+
+		if ((ctl2 & PORT_SYNC_MODE_ENABLE) == 0)
+			return INVALID_TRANSCODER;
+
+		master_select = REG_FIELD_GET(PORT_SYNC_MODE_MASTER_SELECT_MASK, ctl2);
+	} else {
+		u32 ctl = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder));
+
+		if ((ctl & TRANS_DDI_PORT_SYNC_ENABLE) == 0)
+			return INVALID_TRANSCODER;
+
+		master_select = REG_FIELD_GET(TRANS_DDI_PORT_SYNC_MASTER_SELECT_MASK, ctl);
+	}
+
+	if (master_select == 0)
+		return TRANSCODER_EDP;
+	else
+		return master_select - 1;
+}
+
+static void bdw_get_trans_port_sync_config(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	u32 transcoders = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) |
+		BIT(TRANSCODER_C) | BIT(TRANSCODER_D);
+	enum transcoder cpu_transcoder;
+
+	crtc_state->master_transcoder =
+		bdw_transcoder_master_readout(dev_priv, crtc_state->cpu_transcoder);
+
+	for_each_cpu_transcoder_masked(dev_priv, cpu_transcoder, transcoders) {
+		enum intel_display_power_domain power_domain;
+		intel_wakeref_t trans_wakeref;
+
+		power_domain = POWER_DOMAIN_TRANSCODER(cpu_transcoder);
+		trans_wakeref = intel_display_power_get_if_enabled(dev_priv,
+								   power_domain);
+
+		if (!trans_wakeref)
+			continue;
+
+		if (bdw_transcoder_master_readout(dev_priv, cpu_transcoder) ==
+		    crtc_state->cpu_transcoder)
+			crtc_state->sync_mode_slaves_mask |= BIT(cpu_transcoder);
+
+		intel_display_power_put(dev_priv, power_domain, trans_wakeref);
+	}
+
+	drm_WARN_ON(&dev_priv->drm,
+		    crtc_state->master_transcoder != INVALID_TRANSCODER &&
+		    crtc_state->sync_mode_slaves_mask);
+}
+
+static void intel_ddi_read_func_ctl(struct intel_encoder *encoder,
+				    struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	enum transcoder cpu_transcoder = pipe_config->cpu_transcoder;
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	u32 temp, flags = 0;
+
+	temp = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder));
+	if (temp & TRANS_DDI_PHSYNC)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+	if (temp & TRANS_DDI_PVSYNC)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	pipe_config->hw.adjusted_mode.flags |= flags;
+
+	switch (temp & TRANS_DDI_BPC_MASK) {
+	case TRANS_DDI_BPC_6:
+		pipe_config->pipe_bpp = 18;
+		break;
+	case TRANS_DDI_BPC_8:
+		pipe_config->pipe_bpp = 24;
+		break;
+	case TRANS_DDI_BPC_10:
+		pipe_config->pipe_bpp = 30;
+		break;
+	case TRANS_DDI_BPC_12:
+		pipe_config->pipe_bpp = 36;
+		break;
+	default:
+		break;
+	}
+
+	switch (temp & TRANS_DDI_MODE_SELECT_MASK) {
+	case TRANS_DDI_MODE_SELECT_HDMI:
+		pipe_config->has_hdmi_sink = true;
+
+		pipe_config->infoframes.enable |=
+			intel_hdmi_infoframes_enabled(encoder, pipe_config);
+
+		if (pipe_config->infoframes.enable)
+			pipe_config->has_infoframe = true;
+
+		if (temp & TRANS_DDI_HDMI_SCRAMBLING)
+			pipe_config->hdmi_scrambling = true;
+		if (temp & TRANS_DDI_HIGH_TMDS_CHAR_RATE)
+			pipe_config->hdmi_high_tmds_clock_ratio = true;
+		fallthrough;
+	case TRANS_DDI_MODE_SELECT_DVI:
+		pipe_config->output_types |= BIT(INTEL_OUTPUT_HDMI);
+		pipe_config->lane_count = 4;
+		break;
+	case TRANS_DDI_MODE_SELECT_DP_SST:
+		if (encoder->type == INTEL_OUTPUT_EDP)
+			pipe_config->output_types |= BIT(INTEL_OUTPUT_EDP);
+		else
+			pipe_config->output_types |= BIT(INTEL_OUTPUT_DP);
+		pipe_config->lane_count =
+			((temp & DDI_PORT_WIDTH_MASK) >> DDI_PORT_WIDTH_SHIFT) + 1;
+
+		intel_cpu_transcoder_get_m1_n1(crtc, cpu_transcoder,
+					       &pipe_config->dp_m_n);
+		intel_cpu_transcoder_get_m2_n2(crtc, cpu_transcoder,
+					       &pipe_config->dp_m2_n2);
+
+		if (DISPLAY_VER(dev_priv) >= 11) {
+			i915_reg_t dp_tp_ctl = dp_tp_ctl_reg(encoder, pipe_config);
+
+			pipe_config->fec_enable =
+				intel_de_read(dev_priv, dp_tp_ctl) & DP_TP_CTL_FEC_ENABLE;
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "[ENCODER:%d:%s] Fec status: %u\n",
+				    encoder->base.base.id, encoder->base.name,
+				    pipe_config->fec_enable);
+		}
+
+		if (dig_port->lspcon.active && dig_port->dp.has_hdmi_sink)
+			pipe_config->infoframes.enable |=
+				intel_lspcon_infoframes_enabled(encoder, pipe_config);
+		else
+			pipe_config->infoframes.enable |=
+				intel_hdmi_infoframes_enabled(encoder, pipe_config);
+		break;
+	case TRANS_DDI_MODE_SELECT_FDI_OR_128B132B:
+		if (!HAS_DP20(dev_priv)) {
+			/* FDI */
+			pipe_config->output_types |= BIT(INTEL_OUTPUT_ANALOG);
+			break;
+		}
+		fallthrough; /* 128b/132b */
+	case TRANS_DDI_MODE_SELECT_DP_MST:
+		pipe_config->output_types |= BIT(INTEL_OUTPUT_DP_MST);
+		pipe_config->lane_count =
+			((temp & DDI_PORT_WIDTH_MASK) >> DDI_PORT_WIDTH_SHIFT) + 1;
+
+		if (DISPLAY_VER(dev_priv) >= 12)
+			pipe_config->mst_master_transcoder =
+					REG_FIELD_GET(TRANS_DDI_MST_TRANSPORT_SELECT_MASK, temp);
+
+		intel_cpu_transcoder_get_m1_n1(crtc, cpu_transcoder,
+					       &pipe_config->dp_m_n);
+
+		pipe_config->infoframes.enable |=
+			intel_hdmi_infoframes_enabled(encoder, pipe_config);
+		break;
+	default:
+		break;
+	}
+}
+
+static void intel_ddi_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum transcoder cpu_transcoder = pipe_config->cpu_transcoder;
+
+	/* XXX: DSI transcoder paranoia */
+	if (drm_WARN_ON(&dev_priv->drm, transcoder_is_dsi(cpu_transcoder)))
+		return;
+
+	intel_ddi_read_func_ctl(encoder, pipe_config);
+
+	intel_ddi_mso_get_config(encoder, pipe_config);
+
+	pipe_config->has_audio =
+		intel_ddi_is_audio_enabled(dev_priv, cpu_transcoder);
+
+	if (encoder->type == INTEL_OUTPUT_EDP)
+		intel_edp_fixup_vbt_bpp(encoder, pipe_config->pipe_bpp);
+
+	ddi_dotclock_get(pipe_config);
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		pipe_config->lane_lat_optim_mask =
+			bxt_ddi_phy_get_lane_lat_optim_mask(encoder);
+
+	intel_ddi_compute_min_voltage_level(dev_priv, pipe_config);
+
+	intel_hdmi_read_gcp_infoframe(encoder, pipe_config);
+
+	intel_read_infoframe(encoder, pipe_config,
+			     HDMI_INFOFRAME_TYPE_AVI,
+			     &pipe_config->infoframes.avi);
+	intel_read_infoframe(encoder, pipe_config,
+			     HDMI_INFOFRAME_TYPE_SPD,
+			     &pipe_config->infoframes.spd);
+	intel_read_infoframe(encoder, pipe_config,
+			     HDMI_INFOFRAME_TYPE_VENDOR,
+			     &pipe_config->infoframes.hdmi);
+	intel_read_infoframe(encoder, pipe_config,
+			     HDMI_INFOFRAME_TYPE_DRM,
+			     &pipe_config->infoframes.drm);
+
+	if (DISPLAY_VER(dev_priv) >= 8)
+		bdw_get_trans_port_sync_config(pipe_config);
+
+	intel_read_dp_sdp(encoder, pipe_config, HDMI_PACKET_TYPE_GAMUT_METADATA);
+	intel_read_dp_sdp(encoder, pipe_config, DP_SDP_VSC);
+
+	intel_psr_get_config(encoder, pipe_config);
+}
+
+void intel_ddi_get_clock(struct intel_encoder *encoder,
+			 struct intel_crtc_state *crtc_state,
+			 struct intel_shared_dpll *pll)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum icl_port_dpll_id port_dpll_id = ICL_PORT_DPLL_DEFAULT;
+	struct icl_port_dpll *port_dpll = &crtc_state->icl_port_dplls[port_dpll_id];
+	bool pll_active;
+
+	if (drm_WARN_ON(&i915->drm, !pll))
+		return;
+
+	port_dpll->pll = pll;
+	pll_active = intel_dpll_get_hw_state(i915, pll, &port_dpll->hw_state);
+	drm_WARN_ON(&i915->drm, !pll_active);
+
+	icl_set_active_port_dpll(crtc_state, port_dpll_id);
+
+	crtc_state->port_clock = intel_dpll_get_freq(i915, crtc_state->shared_dpll,
+						     &crtc_state->dpll_hw_state);
+}
+
+static void dg2_ddi_get_config(struct intel_encoder *encoder,
+				struct intel_crtc_state *crtc_state)
+{
+	intel_mpllb_readout_hw_state(encoder, &crtc_state->mpllb_state);
+	crtc_state->port_clock = intel_mpllb_calc_port_clock(encoder, &crtc_state->mpllb_state);
+
+	intel_ddi_get_config(encoder, crtc_state);
+}
+
+static void adls_ddi_get_config(struct intel_encoder *encoder,
+				struct intel_crtc_state *crtc_state)
+{
+	intel_ddi_get_clock(encoder, crtc_state, adls_ddi_get_pll(encoder));
+	intel_ddi_get_config(encoder, crtc_state);
+}
+
+static void rkl_ddi_get_config(struct intel_encoder *encoder,
+			       struct intel_crtc_state *crtc_state)
+{
+	intel_ddi_get_clock(encoder, crtc_state, rkl_ddi_get_pll(encoder));
+	intel_ddi_get_config(encoder, crtc_state);
+}
+
+static void dg1_ddi_get_config(struct intel_encoder *encoder,
+			       struct intel_crtc_state *crtc_state)
+{
+	intel_ddi_get_clock(encoder, crtc_state, dg1_ddi_get_pll(encoder));
+	intel_ddi_get_config(encoder, crtc_state);
+}
+
+static void icl_ddi_combo_get_config(struct intel_encoder *encoder,
+				     struct intel_crtc_state *crtc_state)
+{
+	intel_ddi_get_clock(encoder, crtc_state, icl_ddi_combo_get_pll(encoder));
+	intel_ddi_get_config(encoder, crtc_state);
+}
+
+static void icl_ddi_tc_get_clock(struct intel_encoder *encoder,
+				 struct intel_crtc_state *crtc_state,
+				 struct intel_shared_dpll *pll)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum icl_port_dpll_id port_dpll_id;
+	struct icl_port_dpll *port_dpll;
+	bool pll_active;
+
+	if (drm_WARN_ON(&i915->drm, !pll))
+		return;
+
+	if (intel_get_shared_dpll_id(i915, pll) == DPLL_ID_ICL_TBTPLL)
+		port_dpll_id = ICL_PORT_DPLL_DEFAULT;
+	else
+		port_dpll_id = ICL_PORT_DPLL_MG_PHY;
+
+	port_dpll = &crtc_state->icl_port_dplls[port_dpll_id];
+
+	port_dpll->pll = pll;
+	pll_active = intel_dpll_get_hw_state(i915, pll, &port_dpll->hw_state);
+	drm_WARN_ON(&i915->drm, !pll_active);
+
+	icl_set_active_port_dpll(crtc_state, port_dpll_id);
+
+	if (intel_get_shared_dpll_id(i915, crtc_state->shared_dpll) == DPLL_ID_ICL_TBTPLL)
+		crtc_state->port_clock = icl_calc_tbt_pll_link(i915, encoder->port);
+	else
+		crtc_state->port_clock = intel_dpll_get_freq(i915, crtc_state->shared_dpll,
+							     &crtc_state->dpll_hw_state);
+}
+
+static void icl_ddi_tc_get_config(struct intel_encoder *encoder,
+				  struct intel_crtc_state *crtc_state)
+{
+	icl_ddi_tc_get_clock(encoder, crtc_state, icl_ddi_tc_get_pll(encoder));
+	intel_ddi_get_config(encoder, crtc_state);
+}
+
+static void bxt_ddi_get_config(struct intel_encoder *encoder,
+			       struct intel_crtc_state *crtc_state)
+{
+	intel_ddi_get_clock(encoder, crtc_state, bxt_ddi_get_pll(encoder));
+	intel_ddi_get_config(encoder, crtc_state);
+}
+
+static void skl_ddi_get_config(struct intel_encoder *encoder,
+			       struct intel_crtc_state *crtc_state)
+{
+	intel_ddi_get_clock(encoder, crtc_state, skl_ddi_get_pll(encoder));
+	intel_ddi_get_config(encoder, crtc_state);
+}
+
+void hsw_ddi_get_config(struct intel_encoder *encoder,
+			struct intel_crtc_state *crtc_state)
+{
+	intel_ddi_get_clock(encoder, crtc_state, hsw_ddi_get_pll(encoder));
+	intel_ddi_get_config(encoder, crtc_state);
+}
+
+static void intel_ddi_sync_state(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	if (intel_phy_is_tc(i915, phy))
+		intel_tc_port_sanitize_mode(enc_to_dig_port(encoder));
+
+	if (crtc_state && intel_crtc_has_dp_encoder(crtc_state))
+		intel_dp_sync_state(encoder, crtc_state);
+}
+
+static bool intel_ddi_initial_fastset_check(struct intel_encoder *encoder,
+					    struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+	bool fastset = true;
+
+	if (intel_phy_is_tc(i915, phy)) {
+		drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s] Forcing full modeset to compute TC port DPLLs\n",
+			    encoder->base.base.id, encoder->base.name);
+		crtc_state->uapi.mode_changed = true;
+		fastset = false;
+	}
+
+	if (intel_crtc_has_dp_encoder(crtc_state) &&
+	    !intel_dp_initial_fastset_check(encoder, crtc_state))
+		fastset = false;
+
+	return fastset;
+}
+
+static enum intel_output_type
+intel_ddi_compute_output_type(struct intel_encoder *encoder,
+			      struct intel_crtc_state *crtc_state,
+			      struct drm_connector_state *conn_state)
+{
+	switch (conn_state->connector->connector_type) {
+	case DRM_MODE_CONNECTOR_HDMIA:
+		return INTEL_OUTPUT_HDMI;
+	case DRM_MODE_CONNECTOR_eDP:
+		return INTEL_OUTPUT_EDP;
+	case DRM_MODE_CONNECTOR_DisplayPort:
+		return INTEL_OUTPUT_DP;
+	default:
+		MISSING_CASE(conn_state->connector->connector_type);
+		return INTEL_OUTPUT_UNUSED;
+	}
+}
+
+static int intel_ddi_compute_config(struct intel_encoder *encoder,
+				    struct intel_crtc_state *pipe_config,
+				    struct drm_connector_state *conn_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	int ret;
+
+	if (HAS_TRANSCODER(dev_priv, TRANSCODER_EDP) && port == PORT_A)
+		pipe_config->cpu_transcoder = TRANSCODER_EDP;
+
+	if (intel_crtc_has_type(pipe_config, INTEL_OUTPUT_HDMI)) {
+		ret = intel_hdmi_compute_config(encoder, pipe_config, conn_state);
+	} else {
+		ret = intel_dp_compute_config(encoder, pipe_config, conn_state);
+	}
+
+	if (ret)
+		return ret;
+
+	if (IS_HASWELL(dev_priv) && crtc->pipe == PIPE_A &&
+	    pipe_config->cpu_transcoder == TRANSCODER_EDP)
+		pipe_config->pch_pfit.force_thru =
+			pipe_config->pch_pfit.enabled ||
+			pipe_config->crc_enabled;
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		pipe_config->lane_lat_optim_mask =
+			bxt_ddi_phy_calc_lane_lat_optim_mask(pipe_config->lane_count);
+
+	intel_ddi_compute_min_voltage_level(dev_priv, pipe_config);
+
+	return 0;
+}
+
+static bool mode_equal(const struct drm_display_mode *mode1,
+		       const struct drm_display_mode *mode2)
+{
+	return drm_mode_match(mode1, mode2,
+			      DRM_MODE_MATCH_TIMINGS |
+			      DRM_MODE_MATCH_FLAGS |
+			      DRM_MODE_MATCH_3D_FLAGS) &&
+		mode1->clock == mode2->clock; /* we want an exact match */
+}
+
+static bool m_n_equal(const struct intel_link_m_n *m_n_1,
+		      const struct intel_link_m_n *m_n_2)
+{
+	return m_n_1->tu == m_n_2->tu &&
+		m_n_1->data_m == m_n_2->data_m &&
+		m_n_1->data_n == m_n_2->data_n &&
+		m_n_1->link_m == m_n_2->link_m &&
+		m_n_1->link_n == m_n_2->link_n;
+}
+
+static bool crtcs_port_sync_compatible(const struct intel_crtc_state *crtc_state1,
+				       const struct intel_crtc_state *crtc_state2)
+{
+	return crtc_state1->hw.active && crtc_state2->hw.active &&
+		crtc_state1->output_types == crtc_state2->output_types &&
+		crtc_state1->output_format == crtc_state2->output_format &&
+		crtc_state1->lane_count == crtc_state2->lane_count &&
+		crtc_state1->port_clock == crtc_state2->port_clock &&
+		mode_equal(&crtc_state1->hw.adjusted_mode,
+			   &crtc_state2->hw.adjusted_mode) &&
+		m_n_equal(&crtc_state1->dp_m_n, &crtc_state2->dp_m_n);
+}
+
+static u8
+intel_ddi_port_sync_transcoders(const struct intel_crtc_state *ref_crtc_state,
+				int tile_group_id)
+{
+	struct drm_connector *connector;
+	const struct drm_connector_state *conn_state;
+	struct drm_i915_private *dev_priv = to_i915(ref_crtc_state->uapi.crtc->dev);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(ref_crtc_state->uapi.state);
+	u8 transcoders = 0;
+	int i;
+
+	/*
+	 * We don't enable port sync on BDW due to missing w/as and
+	 * due to not having adjusted the modeset sequence appropriately.
+	 */
+	if (DISPLAY_VER(dev_priv) < 9)
+		return 0;
+
+	if (!intel_crtc_has_type(ref_crtc_state, INTEL_OUTPUT_DP))
+		return 0;
+
+	for_each_new_connector_in_state(&state->base, connector, conn_state, i) {
+		struct intel_crtc *crtc = to_intel_crtc(conn_state->crtc);
+		const struct intel_crtc_state *crtc_state;
+
+		if (!crtc)
+			continue;
+
+		if (!connector->has_tile ||
+		    connector->tile_group->id !=
+		    tile_group_id)
+			continue;
+		crtc_state = intel_atomic_get_new_crtc_state(state,
+							     crtc);
+		if (!crtcs_port_sync_compatible(ref_crtc_state,
+						crtc_state))
+			continue;
+		transcoders |= BIT(crtc_state->cpu_transcoder);
+	}
+
+	return transcoders;
+}
+
+static int intel_ddi_compute_config_late(struct intel_encoder *encoder,
+					 struct intel_crtc_state *crtc_state,
+					 struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct drm_connector *connector = conn_state->connector;
+	u8 port_sync_transcoders = 0;
+
+	drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s] [CRTC:%d:%s]",
+		    encoder->base.base.id, encoder->base.name,
+		    crtc_state->uapi.crtc->base.id, crtc_state->uapi.crtc->name);
+
+	if (connector->has_tile)
+		port_sync_transcoders = intel_ddi_port_sync_transcoders(crtc_state,
+									connector->tile_group->id);
+
+	/*
+	 * EDP Transcoders cannot be ensalved
+	 * make them a master always when present
+	 */
+	if (port_sync_transcoders & BIT(TRANSCODER_EDP))
+		crtc_state->master_transcoder = TRANSCODER_EDP;
+	else
+		crtc_state->master_transcoder = ffs(port_sync_transcoders) - 1;
+
+	if (crtc_state->master_transcoder == crtc_state->cpu_transcoder) {
+		crtc_state->master_transcoder = INVALID_TRANSCODER;
+		crtc_state->sync_mode_slaves_mask =
+			port_sync_transcoders & ~BIT(crtc_state->cpu_transcoder);
+	}
+
+	return 0;
+}
+
+static void intel_ddi_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(to_intel_encoder(encoder));
+	enum phy phy = intel_port_to_phy(i915, dig_port->base.port);
+
+	intel_dp_encoder_flush_work(encoder);
+	if (intel_phy_is_tc(i915, phy))
+		intel_tc_port_flush_work(dig_port);
+	intel_display_power_flush_work(i915);
+
+	drm_encoder_cleanup(encoder);
+	kfree(dig_port->hdcp_port_data.streams);
+	kfree(dig_port);
+}
+
+static void intel_ddi_encoder_reset(struct drm_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(to_intel_encoder(encoder));
+	struct intel_digital_port *dig_port = enc_to_dig_port(to_intel_encoder(encoder));
+	enum phy phy = intel_port_to_phy(i915, dig_port->base.port);
+
+	intel_dp->reset_link_params = true;
+
+	intel_pps_encoder_reset(intel_dp);
+
+	if (intel_phy_is_tc(i915, phy))
+		intel_tc_port_init_mode(dig_port);
+}
+
+static const struct drm_encoder_funcs intel_ddi_funcs = {
+	.reset = intel_ddi_encoder_reset,
+	.destroy = intel_ddi_encoder_destroy,
+};
+
+static struct intel_connector *
+intel_ddi_init_dp_connector(struct intel_digital_port *dig_port)
+{
+	struct intel_connector *connector;
+	enum port port = dig_port->base.port;
+
+	connector = intel_connector_alloc();
+	if (!connector)
+		return NULL;
+
+	dig_port->dp.output_reg = DDI_BUF_CTL(port);
+	dig_port->dp.prepare_link_retrain = intel_ddi_prepare_link_retrain;
+	dig_port->dp.set_link_train = intel_ddi_set_link_train;
+	dig_port->dp.set_idle_link_train = intel_ddi_set_idle_link_train;
+
+	dig_port->dp.voltage_max = intel_ddi_dp_voltage_max;
+	dig_port->dp.preemph_max = intel_ddi_dp_preemph_max;
+
+	if (!intel_dp_init_connector(dig_port, connector)) {
+		kfree(connector);
+		return NULL;
+	}
+
+	if (dig_port->base.type == INTEL_OUTPUT_EDP) {
+		struct drm_device *dev = dig_port->base.base.dev;
+		struct drm_privacy_screen *privacy_screen;
+
+		privacy_screen = drm_privacy_screen_get(dev->dev, NULL);
+		if (!IS_ERR(privacy_screen)) {
+			drm_connector_attach_privacy_screen_provider(&connector->base,
+								     privacy_screen);
+		} else if (PTR_ERR(privacy_screen) != -ENODEV) {
+			drm_warn(dev, "Error getting privacy-screen\n");
+		}
+	}
+
+	return connector;
+}
+
+static int modeset_pipe(struct drm_crtc *crtc,
+			struct drm_modeset_acquire_ctx *ctx)
+{
+	struct drm_atomic_state *state;
+	struct drm_crtc_state *crtc_state;
+	int ret;
+
+	state = drm_atomic_state_alloc(crtc->dev);
+	if (!state)
+		return -ENOMEM;
+
+	state->acquire_ctx = ctx;
+
+	crtc_state = drm_atomic_get_crtc_state(state, crtc);
+	if (IS_ERR(crtc_state)) {
+		ret = PTR_ERR(crtc_state);
+		goto out;
+	}
+
+	crtc_state->connectors_changed = true;
+
+	ret = drm_atomic_commit(state);
+out:
+	drm_atomic_state_put(state);
+
+	return ret;
+}
+
+static int intel_hdmi_reset_link(struct intel_encoder *encoder,
+				 struct drm_modeset_acquire_ctx *ctx)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_hdmi *hdmi = enc_to_intel_hdmi(encoder);
+	struct intel_connector *connector = hdmi->attached_connector;
+	struct i2c_adapter *adapter =
+		intel_gmbus_get_adapter(dev_priv, hdmi->ddc_bus);
+	struct drm_connector_state *conn_state;
+	struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+	u8 config;
+	int ret;
+
+	if (!connector || connector->base.status != connector_status_connected)
+		return 0;
+
+	ret = drm_modeset_lock(&dev_priv->drm.mode_config.connection_mutex,
+			       ctx);
+	if (ret)
+		return ret;
+
+	conn_state = connector->base.state;
+
+	crtc = to_intel_crtc(conn_state->crtc);
+	if (!crtc)
+		return 0;
+
+	ret = drm_modeset_lock(&crtc->base.mutex, ctx);
+	if (ret)
+		return ret;
+
+	crtc_state = to_intel_crtc_state(crtc->base.state);
+
+	drm_WARN_ON(&dev_priv->drm,
+		    !intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI));
+
+	if (!crtc_state->hw.active)
+		return 0;
+
+	if (!crtc_state->hdmi_high_tmds_clock_ratio &&
+	    !crtc_state->hdmi_scrambling)
+		return 0;
+
+	if (conn_state->commit &&
+	    !try_wait_for_completion(&conn_state->commit->hw_done))
+		return 0;
+
+	ret = drm_scdc_readb(adapter, SCDC_TMDS_CONFIG, &config);
+	if (ret < 0) {
+		drm_err(&dev_priv->drm, "Failed to read TMDS config: %d\n",
+			ret);
+		return 0;
+	}
+
+	if (!!(config & SCDC_TMDS_BIT_CLOCK_RATIO_BY_40) ==
+	    crtc_state->hdmi_high_tmds_clock_ratio &&
+	    !!(config & SCDC_SCRAMBLING_ENABLE) ==
+	    crtc_state->hdmi_scrambling)
+		return 0;
+
+	/*
+	 * HDMI 2.0 says that one should not send scrambled data
+	 * prior to configuring the sink scrambling, and that
+	 * TMDS clock/data transmission should be suspended when
+	 * changing the TMDS clock rate in the sink. So let's
+	 * just do a full modeset here, even though some sinks
+	 * would be perfectly happy if were to just reconfigure
+	 * the SCDC settings on the fly.
+	 */
+	return modeset_pipe(&crtc->base, ctx);
+}
+
+static enum intel_hotplug_state
+intel_ddi_hotplug(struct intel_encoder *encoder,
+		  struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_dp *intel_dp = &dig_port->dp;
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+	bool is_tc = intel_phy_is_tc(i915, phy);
+	struct drm_modeset_acquire_ctx ctx;
+	enum intel_hotplug_state state;
+	int ret;
+
+	if (intel_dp->compliance.test_active &&
+	    intel_dp->compliance.test_type == DP_TEST_LINK_PHY_TEST_PATTERN) {
+		intel_dp_phy_test(encoder);
+		/* just do the PHY test and nothing else */
+		return INTEL_HOTPLUG_UNCHANGED;
+	}
+
+	state = intel_encoder_hotplug(encoder, connector);
+
+	drm_modeset_acquire_init(&ctx, 0);
+
+	for (;;) {
+		if (connector->base.connector_type == DRM_MODE_CONNECTOR_HDMIA)
+			ret = intel_hdmi_reset_link(encoder, &ctx);
+		else
+			ret = intel_dp_retrain_link(encoder, &ctx);
+
+		if (ret == -EDEADLK) {
+			drm_modeset_backoff(&ctx);
+			continue;
+		}
+
+		break;
+	}
+
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+	drm_WARN(encoder->base.dev, ret,
+		 "Acquiring modeset locks failed with %i\n", ret);
+
+	/*
+	 * Unpowered type-c dongles can take some time to boot and be
+	 * responsible, so here giving some time to those dongles to power up
+	 * and then retrying the probe.
+	 *
+	 * On many platforms the HDMI live state signal is known to be
+	 * unreliable, so we can't use it to detect if a sink is connected or
+	 * not. Instead we detect if it's connected based on whether we can
+	 * read the EDID or not. That in turn has a problem during disconnect,
+	 * since the HPD interrupt may be raised before the DDC lines get
+	 * disconnected (due to how the required length of DDC vs. HPD
+	 * connector pins are specified) and so we'll still be able to get a
+	 * valid EDID. To solve this schedule another detection cycle if this
+	 * time around we didn't detect any change in the sink's connection
+	 * status.
+	 *
+	 * Type-c connectors which get their HPD signal deasserted then
+	 * reasserted, without unplugging/replugging the sink from the
+	 * connector, introduce a delay until the AUX channel communication
+	 * becomes functional. Retry the detection for 5 seconds on type-c
+	 * connectors to account for this delay.
+	 */
+	if (state == INTEL_HOTPLUG_UNCHANGED &&
+	    connector->hotplug_retries < (is_tc ? 5 : 1) &&
+	    !dig_port->dp.is_mst)
+		state = INTEL_HOTPLUG_RETRY;
+
+	return state;
+}
+
+static bool lpt_digital_port_connected(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 bit = dev_priv->display.hotplug.pch_hpd[encoder->hpd_pin];
+
+	return intel_de_read(dev_priv, SDEISR) & bit;
+}
+
+static bool hsw_digital_port_connected(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 bit = dev_priv->display.hotplug.hpd[encoder->hpd_pin];
+
+	return intel_de_read(dev_priv, DEISR) & bit;
+}
+
+static bool bdw_digital_port_connected(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 bit = dev_priv->display.hotplug.hpd[encoder->hpd_pin];
+
+	return intel_de_read(dev_priv, GEN8_DE_PORT_ISR) & bit;
+}
+
+static struct intel_connector *
+intel_ddi_init_hdmi_connector(struct intel_digital_port *dig_port)
+{
+	struct intel_connector *connector;
+	enum port port = dig_port->base.port;
+
+	connector = intel_connector_alloc();
+	if (!connector)
+		return NULL;
+
+	dig_port->hdmi.hdmi_reg = DDI_BUF_CTL(port);
+	intel_hdmi_init_connector(dig_port, connector);
+
+	return connector;
+}
+
+static bool intel_ddi_a_force_4_lanes(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
+
+	if (dig_port->base.port != PORT_A)
+		return false;
+
+	if (dig_port->saved_port_bits & DDI_A_4_LANES)
+		return false;
+
+	/* Broxton/Geminilake: Bspec says that DDI_A_4_LANES is the only
+	 *                     supported configuration
+	 */
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		return true;
+
+	return false;
+}
+
+static int
+intel_ddi_max_lanes(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
+	enum port port = dig_port->base.port;
+	int max_lanes = 4;
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		return max_lanes;
+
+	if (port == PORT_A || port == PORT_E) {
+		if (intel_de_read(dev_priv, DDI_BUF_CTL(PORT_A)) & DDI_A_4_LANES)
+			max_lanes = port == PORT_A ? 4 : 0;
+		else
+			/* Both A and E share 2 lanes */
+			max_lanes = 2;
+	}
+
+	/*
+	 * Some BIOS might fail to set this bit on port A if eDP
+	 * wasn't lit up at boot.  Force this bit set when needed
+	 * so we use the proper lane count for our calculations.
+	 */
+	if (intel_ddi_a_force_4_lanes(dig_port)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Forcing DDI_A_4_LANES for port A\n");
+		dig_port->saved_port_bits |= DDI_A_4_LANES;
+		max_lanes = 4;
+	}
+
+	return max_lanes;
+}
+
+static bool hti_uses_phy(struct drm_i915_private *i915, enum phy phy)
+{
+	return i915->hti_state & HDPORT_ENABLED &&
+	       i915->hti_state & HDPORT_DDI_USED(phy);
+}
+
+static enum hpd_pin xelpd_hpd_pin(struct drm_i915_private *dev_priv,
+				  enum port port)
+{
+	if (port >= PORT_D_XELPD)
+		return HPD_PORT_D + port - PORT_D_XELPD;
+	else if (port >= PORT_TC1)
+		return HPD_PORT_TC1 + port - PORT_TC1;
+	else
+		return HPD_PORT_A + port - PORT_A;
+}
+
+static enum hpd_pin dg1_hpd_pin(struct drm_i915_private *dev_priv,
+				enum port port)
+{
+	if (port >= PORT_TC1)
+		return HPD_PORT_C + port - PORT_TC1;
+	else
+		return HPD_PORT_A + port - PORT_A;
+}
+
+static enum hpd_pin tgl_hpd_pin(struct drm_i915_private *dev_priv,
+				enum port port)
+{
+	if (port >= PORT_TC1)
+		return HPD_PORT_TC1 + port - PORT_TC1;
+	else
+		return HPD_PORT_A + port - PORT_A;
+}
+
+static enum hpd_pin rkl_hpd_pin(struct drm_i915_private *dev_priv,
+				enum port port)
+{
+	if (HAS_PCH_TGP(dev_priv))
+		return tgl_hpd_pin(dev_priv, port);
+
+	if (port >= PORT_TC1)
+		return HPD_PORT_C + port - PORT_TC1;
+	else
+		return HPD_PORT_A + port - PORT_A;
+}
+
+static enum hpd_pin icl_hpd_pin(struct drm_i915_private *dev_priv,
+				enum port port)
+{
+	if (port >= PORT_C)
+		return HPD_PORT_TC1 + port - PORT_C;
+	else
+		return HPD_PORT_A + port - PORT_A;
+}
+
+static enum hpd_pin ehl_hpd_pin(struct drm_i915_private *dev_priv,
+				enum port port)
+{
+	if (port == PORT_D)
+		return HPD_PORT_A;
+
+	if (HAS_PCH_TGP(dev_priv))
+		return icl_hpd_pin(dev_priv, port);
+
+	return HPD_PORT_A + port - PORT_A;
+}
+
+static enum hpd_pin skl_hpd_pin(struct drm_i915_private *dev_priv, enum port port)
+{
+	if (HAS_PCH_TGP(dev_priv))
+		return icl_hpd_pin(dev_priv, port);
+
+	return HPD_PORT_A + port - PORT_A;
+}
+
+static bool intel_ddi_is_tc(struct drm_i915_private *i915, enum port port)
+{
+	if (DISPLAY_VER(i915) >= 12)
+		return port >= PORT_TC1;
+	else if (DISPLAY_VER(i915) >= 11)
+		return port >= PORT_C;
+	else
+		return false;
+}
+
+static void intel_ddi_encoder_suspend(struct intel_encoder *encoder)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	intel_dp_encoder_suspend(encoder);
+
+	if (!intel_phy_is_tc(i915, phy))
+		return;
+
+	intel_tc_port_flush_work(dig_port);
+}
+
+static void intel_ddi_encoder_shutdown(struct intel_encoder *encoder)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	intel_dp_encoder_shutdown(encoder);
+	intel_hdmi_encoder_shutdown(encoder);
+
+	if (!intel_phy_is_tc(i915, phy))
+		return;
+
+	intel_tc_port_flush_work(dig_port);
+}
+
+#define port_tc_name(port) ((port) - PORT_TC1 + '1')
+#define tc_port_name(tc_port) ((tc_port) - TC_PORT_1 + '1')
+
+void intel_ddi_init(struct drm_i915_private *dev_priv, enum port port)
+{
+	struct intel_digital_port *dig_port;
+	struct intel_encoder *encoder;
+	const struct intel_bios_encoder_data *devdata;
+	bool init_hdmi, init_dp;
+	enum phy phy = intel_port_to_phy(dev_priv, port);
+
+	/*
+	 * On platforms with HTI (aka HDPORT), if it's enabled at boot it may
+	 * have taken over some of the PHYs and made them unavailable to the
+	 * driver.  In that case we should skip initializing the corresponding
+	 * outputs.
+	 */
+	if (hti_uses_phy(dev_priv, phy)) {
+		drm_dbg_kms(&dev_priv->drm, "PORT %c / PHY %c reserved by HTI\n",
+			    port_name(port), phy_name(phy));
+		return;
+	}
+
+	devdata = intel_bios_encoder_data_lookup(dev_priv, port);
+	if (!devdata) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "VBT says port %c is not present\n",
+			    port_name(port));
+		return;
+	}
+
+	init_hdmi = intel_bios_encoder_supports_dvi(devdata) ||
+		intel_bios_encoder_supports_hdmi(devdata);
+	init_dp = intel_bios_encoder_supports_dp(devdata);
+
+	if (intel_bios_is_lspcon_present(dev_priv, port)) {
+		/*
+		 * Lspcon device needs to be driven with DP connector
+		 * with special detection sequence. So make sure DP
+		 * is initialized before lspcon.
+		 */
+		init_dp = true;
+		init_hdmi = false;
+		drm_dbg_kms(&dev_priv->drm, "VBT says port %c has lspcon\n",
+			    port_name(port));
+	}
+
+	if (!init_dp && !init_hdmi) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "VBT says port %c is not DVI/HDMI/DP compatible, respect it\n",
+			    port_name(port));
+		return;
+	}
+
+	if (intel_phy_is_snps(dev_priv, phy) &&
+	    dev_priv->snps_phy_failed_calibration & BIT(phy)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "SNPS PHY %c failed to calibrate, proceeding anyway\n",
+			    phy_name(phy));
+	}
+
+	dig_port = kzalloc(sizeof(*dig_port), GFP_KERNEL);
+	if (!dig_port)
+		return;
+
+	encoder = &dig_port->base;
+	encoder->devdata = devdata;
+
+	if (DISPLAY_VER(dev_priv) >= 13 && port >= PORT_D_XELPD) {
+		drm_encoder_init(&dev_priv->drm, &encoder->base, &intel_ddi_funcs,
+				 DRM_MODE_ENCODER_TMDS,
+				 "DDI %c/PHY %c",
+				 port_name(port - PORT_D_XELPD + PORT_D),
+				 phy_name(phy));
+	} else if (DISPLAY_VER(dev_priv) >= 12) {
+		enum tc_port tc_port = intel_port_to_tc(dev_priv, port);
+
+		drm_encoder_init(&dev_priv->drm, &encoder->base, &intel_ddi_funcs,
+				 DRM_MODE_ENCODER_TMDS,
+				 "DDI %s%c/PHY %s%c",
+				 port >= PORT_TC1 ? "TC" : "",
+				 port >= PORT_TC1 ? port_tc_name(port) : port_name(port),
+				 tc_port != TC_PORT_NONE ? "TC" : "",
+				 tc_port != TC_PORT_NONE ? tc_port_name(tc_port) : phy_name(phy));
+	} else if (DISPLAY_VER(dev_priv) >= 11) {
+		enum tc_port tc_port = intel_port_to_tc(dev_priv, port);
+
+		drm_encoder_init(&dev_priv->drm, &encoder->base, &intel_ddi_funcs,
+				 DRM_MODE_ENCODER_TMDS,
+				 "DDI %c%s/PHY %s%c",
+				 port_name(port),
+				 port >= PORT_C ? " (TC)" : "",
+				 tc_port != TC_PORT_NONE ? "TC" : "",
+				 tc_port != TC_PORT_NONE ? tc_port_name(tc_port) : phy_name(phy));
+	} else {
+		drm_encoder_init(&dev_priv->drm, &encoder->base, &intel_ddi_funcs,
+				 DRM_MODE_ENCODER_TMDS,
+				 "DDI %c/PHY %c", port_name(port),  phy_name(phy));
+	}
+
+	mutex_init(&dig_port->hdcp_mutex);
+	dig_port->num_hdcp_streams = 0;
+
+	encoder->hotplug = intel_ddi_hotplug;
+	encoder->compute_output_type = intel_ddi_compute_output_type;
+	encoder->compute_config = intel_ddi_compute_config;
+	encoder->compute_config_late = intel_ddi_compute_config_late;
+	encoder->enable = intel_enable_ddi;
+	encoder->pre_pll_enable = intel_ddi_pre_pll_enable;
+	encoder->pre_enable = intel_ddi_pre_enable;
+	encoder->disable = intel_disable_ddi;
+	encoder->post_disable = intel_ddi_post_disable;
+	encoder->update_pipe = intel_ddi_update_pipe;
+	encoder->get_hw_state = intel_ddi_get_hw_state;
+	encoder->sync_state = intel_ddi_sync_state;
+	encoder->initial_fastset_check = intel_ddi_initial_fastset_check;
+	encoder->suspend = intel_ddi_encoder_suspend;
+	encoder->shutdown = intel_ddi_encoder_shutdown;
+	encoder->get_power_domains = intel_ddi_get_power_domains;
+
+	encoder->type = INTEL_OUTPUT_DDI;
+	encoder->power_domain = intel_display_power_ddi_lanes_domain(dev_priv, port);
+	encoder->port = port;
+	encoder->cloneable = 0;
+	encoder->pipe_mask = ~0;
+
+	if (IS_DG2(dev_priv)) {
+		encoder->enable_clock = intel_mpllb_enable;
+		encoder->disable_clock = intel_mpllb_disable;
+		encoder->get_config = dg2_ddi_get_config;
+	} else if (IS_ALDERLAKE_S(dev_priv)) {
+		encoder->enable_clock = adls_ddi_enable_clock;
+		encoder->disable_clock = adls_ddi_disable_clock;
+		encoder->is_clock_enabled = adls_ddi_is_clock_enabled;
+		encoder->get_config = adls_ddi_get_config;
+	} else if (IS_ROCKETLAKE(dev_priv)) {
+		encoder->enable_clock = rkl_ddi_enable_clock;
+		encoder->disable_clock = rkl_ddi_disable_clock;
+		encoder->is_clock_enabled = rkl_ddi_is_clock_enabled;
+		encoder->get_config = rkl_ddi_get_config;
+	} else if (IS_DG1(dev_priv)) {
+		encoder->enable_clock = dg1_ddi_enable_clock;
+		encoder->disable_clock = dg1_ddi_disable_clock;
+		encoder->is_clock_enabled = dg1_ddi_is_clock_enabled;
+		encoder->get_config = dg1_ddi_get_config;
+	} else if (IS_JSL_EHL(dev_priv)) {
+		if (intel_ddi_is_tc(dev_priv, port)) {
+			encoder->enable_clock = jsl_ddi_tc_enable_clock;
+			encoder->disable_clock = jsl_ddi_tc_disable_clock;
+			encoder->is_clock_enabled = jsl_ddi_tc_is_clock_enabled;
+			encoder->get_config = icl_ddi_combo_get_config;
+		} else {
+			encoder->enable_clock = icl_ddi_combo_enable_clock;
+			encoder->disable_clock = icl_ddi_combo_disable_clock;
+			encoder->is_clock_enabled = icl_ddi_combo_is_clock_enabled;
+			encoder->get_config = icl_ddi_combo_get_config;
+		}
+	} else if (DISPLAY_VER(dev_priv) >= 11) {
+		if (intel_ddi_is_tc(dev_priv, port)) {
+			encoder->enable_clock = icl_ddi_tc_enable_clock;
+			encoder->disable_clock = icl_ddi_tc_disable_clock;
+			encoder->is_clock_enabled = icl_ddi_tc_is_clock_enabled;
+			encoder->get_config = icl_ddi_tc_get_config;
+		} else {
+			encoder->enable_clock = icl_ddi_combo_enable_clock;
+			encoder->disable_clock = icl_ddi_combo_disable_clock;
+			encoder->is_clock_enabled = icl_ddi_combo_is_clock_enabled;
+			encoder->get_config = icl_ddi_combo_get_config;
+		}
+	} else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		/* BXT/GLK have fixed PLL->port mapping */
+		encoder->get_config = bxt_ddi_get_config;
+	} else if (DISPLAY_VER(dev_priv) == 9) {
+		encoder->enable_clock = skl_ddi_enable_clock;
+		encoder->disable_clock = skl_ddi_disable_clock;
+		encoder->is_clock_enabled = skl_ddi_is_clock_enabled;
+		encoder->get_config = skl_ddi_get_config;
+	} else if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv)) {
+		encoder->enable_clock = hsw_ddi_enable_clock;
+		encoder->disable_clock = hsw_ddi_disable_clock;
+		encoder->is_clock_enabled = hsw_ddi_is_clock_enabled;
+		encoder->get_config = hsw_ddi_get_config;
+	}
+
+	if (IS_DG2(dev_priv)) {
+		encoder->set_signal_levels = intel_snps_phy_set_signal_levels;
+	} else if (DISPLAY_VER(dev_priv) >= 12) {
+		if (intel_phy_is_combo(dev_priv, phy))
+			encoder->set_signal_levels = icl_combo_phy_set_signal_levels;
+		else
+			encoder->set_signal_levels = tgl_dkl_phy_set_signal_levels;
+	} else if (DISPLAY_VER(dev_priv) >= 11) {
+		if (intel_phy_is_combo(dev_priv, phy))
+			encoder->set_signal_levels = icl_combo_phy_set_signal_levels;
+		else
+			encoder->set_signal_levels = icl_mg_phy_set_signal_levels;
+	} else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		encoder->set_signal_levels = bxt_ddi_phy_set_signal_levels;
+	} else {
+		encoder->set_signal_levels = hsw_set_signal_levels;
+	}
+
+	intel_ddi_buf_trans_init(encoder);
+
+	if (DISPLAY_VER(dev_priv) >= 13)
+		encoder->hpd_pin = xelpd_hpd_pin(dev_priv, port);
+	else if (IS_DG1(dev_priv))
+		encoder->hpd_pin = dg1_hpd_pin(dev_priv, port);
+	else if (IS_ROCKETLAKE(dev_priv))
+		encoder->hpd_pin = rkl_hpd_pin(dev_priv, port);
+	else if (DISPLAY_VER(dev_priv) >= 12)
+		encoder->hpd_pin = tgl_hpd_pin(dev_priv, port);
+	else if (IS_JSL_EHL(dev_priv))
+		encoder->hpd_pin = ehl_hpd_pin(dev_priv, port);
+	else if (DISPLAY_VER(dev_priv) == 11)
+		encoder->hpd_pin = icl_hpd_pin(dev_priv, port);
+	else if (DISPLAY_VER(dev_priv) == 9 && !IS_BROXTON(dev_priv))
+		encoder->hpd_pin = skl_hpd_pin(dev_priv, port);
+	else
+		encoder->hpd_pin = intel_hpd_pin_default(dev_priv, port);
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		dig_port->saved_port_bits =
+			intel_de_read(dev_priv, DDI_BUF_CTL(port))
+			& DDI_BUF_PORT_REVERSAL;
+	else
+		dig_port->saved_port_bits =
+			intel_de_read(dev_priv, DDI_BUF_CTL(port))
+			& (DDI_BUF_PORT_REVERSAL | DDI_A_4_LANES);
+
+	if (intel_bios_is_lane_reversal_needed(dev_priv, port))
+		dig_port->saved_port_bits |= DDI_BUF_PORT_REVERSAL;
+
+	dig_port->dp.output_reg = INVALID_MMIO_REG;
+	dig_port->max_lanes = intel_ddi_max_lanes(dig_port);
+	dig_port->aux_ch = intel_bios_port_aux_ch(dev_priv, port);
+
+	if (intel_phy_is_tc(dev_priv, phy)) {
+		bool is_legacy =
+			!intel_bios_encoder_supports_typec_usb(devdata) &&
+			!intel_bios_encoder_supports_tbt(devdata);
+
+		intel_tc_port_init(dig_port, is_legacy);
+
+		encoder->update_prepare = intel_ddi_update_prepare;
+		encoder->update_complete = intel_ddi_update_complete;
+	}
+
+	drm_WARN_ON(&dev_priv->drm, port > PORT_I);
+	dig_port->ddi_io_power_domain = intel_display_power_ddi_io_domain(dev_priv, port);
+
+	if (init_dp) {
+		if (!intel_ddi_init_dp_connector(dig_port))
+			goto err;
+
+		dig_port->hpd_pulse = intel_dp_hpd_pulse;
+
+		if (dig_port->dp.mso_link_count)
+			encoder->pipe_mask = intel_ddi_splitter_pipe_mask(dev_priv);
+	}
+
+	/* In theory we don't need the encoder->type check, but leave it just in
+	 * case we have some really bad VBTs... */
+	if (encoder->type != INTEL_OUTPUT_EDP && init_hdmi) {
+		if (!intel_ddi_init_hdmi_connector(dig_port))
+			goto err;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		if (intel_phy_is_tc(dev_priv, phy))
+			dig_port->connected = intel_tc_port_connected;
+		else
+			dig_port->connected = lpt_digital_port_connected;
+	} else if (DISPLAY_VER(dev_priv) >= 8) {
+		if (port == PORT_A || IS_GEMINILAKE(dev_priv) ||
+		    IS_BROXTON(dev_priv))
+			dig_port->connected = bdw_digital_port_connected;
+		else
+			dig_port->connected = lpt_digital_port_connected;
+	} else {
+		if (port == PORT_A)
+			dig_port->connected = hsw_digital_port_connected;
+		else
+			dig_port->connected = lpt_digital_port_connected;
+	}
+
+	intel_infoframe_init(dig_port);
+
+	return;
+
+err:
+	drm_encoder_cleanup(&encoder->base);
+	kfree(dig_port);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_ddi.h b/drivers/gpu/drm/i915/display/intel_ddi.h
new file mode 100644
index 000000000..d39076fac
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_ddi.h
@@ -0,0 +1,73 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DDI_H__
+#define __INTEL_DDI_H__
+
+#include "i915_reg_defs.h"
+
+struct drm_connector_state;
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_connector;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_dp;
+struct intel_dpll_hw_state;
+struct intel_encoder;
+struct intel_shared_dpll;
+enum pipe;
+enum port;
+enum transcoder;
+
+i915_reg_t dp_tp_ctl_reg(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *crtc_state);
+i915_reg_t dp_tp_status_reg(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state);
+void intel_ddi_fdi_post_disable(struct intel_atomic_state *state,
+				struct intel_encoder *intel_encoder,
+				const struct intel_crtc_state *old_crtc_state,
+				const struct drm_connector_state *old_conn_state);
+void intel_ddi_enable_clock(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state);
+void intel_ddi_disable_clock(struct intel_encoder *encoder);
+void intel_ddi_get_clock(struct intel_encoder *encoder,
+			 struct intel_crtc_state *crtc_state,
+			 struct intel_shared_dpll *pll);
+void hsw_ddi_enable_clock(struct intel_encoder *encoder,
+			  const struct intel_crtc_state *crtc_state);
+void hsw_ddi_disable_clock(struct intel_encoder *encoder);
+bool hsw_ddi_is_clock_enabled(struct intel_encoder *encoder);
+void hsw_ddi_get_config(struct intel_encoder *encoder,
+			struct intel_crtc_state *crtc_state);
+struct intel_shared_dpll *icl_ddi_combo_get_pll(struct intel_encoder *encoder);
+void hsw_prepare_dp_ddi_buffers(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state);
+void intel_wait_ddi_buf_idle(struct drm_i915_private *dev_priv,
+			     enum port port);
+void intel_ddi_init(struct drm_i915_private *dev_priv, enum port port);
+bool intel_ddi_get_hw_state(struct intel_encoder *encoder, enum pipe *pipe);
+void intel_ddi_enable_transcoder_func(struct intel_encoder *encoder,
+				      const struct intel_crtc_state *crtc_state);
+void intel_ddi_disable_transcoder_func(const struct intel_crtc_state *crtc_state);
+void intel_ddi_enable_pipe_clock(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state);
+void intel_ddi_disable_pipe_clock(const  struct intel_crtc_state *crtc_state);
+void intel_ddi_set_dp_msa(const struct intel_crtc_state *crtc_state,
+			  const struct drm_connector_state *conn_state);
+bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector);
+void intel_ddi_set_vc_payload_alloc(const struct intel_crtc_state *crtc_state,
+				    bool state);
+void intel_ddi_compute_min_voltage_level(struct drm_i915_private *dev_priv,
+					 struct intel_crtc_state *crtc_state);
+int intel_ddi_toggle_hdcp_bits(struct intel_encoder *intel_encoder,
+			       enum transcoder cpu_transcoder,
+			       bool enable, u32 hdcp_mask);
+void intel_ddi_sanitize_encoder_pll_mapping(struct intel_encoder *encoder);
+int intel_ddi_level(struct intel_encoder *encoder,
+		    const struct intel_crtc_state *crtc_state,
+		    int lane);
+
+#endif /* __INTEL_DDI_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_ddi_buf_trans.c b/drivers/gpu/drm/i915/display/intel_ddi_buf_trans.c
new file mode 100644
index 000000000..006a2e979
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_ddi_buf_trans.c
@@ -0,0 +1,1662 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "intel_ddi.h"
+#include "intel_ddi_buf_trans.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+
+/* HDMI/DVI modes ignore everything but the last 2 items. So we share
+ * them for both DP and FDI transports, allowing those ports to
+ * automatically adapt to HDMI connections as well
+ */
+static const union intel_ddi_buf_trans_entry _hsw_trans_dp[] = {
+	{ .hsw = { 0x00FFFFFF, 0x0006000E, 0x0 } },
+	{ .hsw = { 0x00D75FFF, 0x0005000A, 0x0 } },
+	{ .hsw = { 0x00C30FFF, 0x00040006, 0x0 } },
+	{ .hsw = { 0x80AAAFFF, 0x000B0000, 0x0 } },
+	{ .hsw = { 0x00FFFFFF, 0x0005000A, 0x0 } },
+	{ .hsw = { 0x00D75FFF, 0x000C0004, 0x0 } },
+	{ .hsw = { 0x80C30FFF, 0x000B0000, 0x0 } },
+	{ .hsw = { 0x00FFFFFF, 0x00040006, 0x0 } },
+	{ .hsw = { 0x80D75FFF, 0x000B0000, 0x0 } },
+};
+
+static const struct intel_ddi_buf_trans hsw_trans_dp = {
+	.entries = _hsw_trans_dp,
+	.num_entries = ARRAY_SIZE(_hsw_trans_dp),
+};
+
+static const union intel_ddi_buf_trans_entry _hsw_trans_fdi[] = {
+	{ .hsw = { 0x00FFFFFF, 0x0007000E, 0x0 } },
+	{ .hsw = { 0x00D75FFF, 0x000F000A, 0x0 } },
+	{ .hsw = { 0x00C30FFF, 0x00060006, 0x0 } },
+	{ .hsw = { 0x00AAAFFF, 0x001E0000, 0x0 } },
+	{ .hsw = { 0x00FFFFFF, 0x000F000A, 0x0 } },
+	{ .hsw = { 0x00D75FFF, 0x00160004, 0x0 } },
+	{ .hsw = { 0x00C30FFF, 0x001E0000, 0x0 } },
+	{ .hsw = { 0x00FFFFFF, 0x00060006, 0x0 } },
+	{ .hsw = { 0x00D75FFF, 0x001E0000, 0x0 } },
+};
+
+static const struct intel_ddi_buf_trans hsw_trans_fdi = {
+	.entries = _hsw_trans_fdi,
+	.num_entries = ARRAY_SIZE(_hsw_trans_fdi),
+};
+
+static const union intel_ddi_buf_trans_entry _hsw_trans_hdmi[] = {
+							/* Idx	NT mV d	T mV d	db	*/
+	{ .hsw = { 0x00FFFFFF, 0x0006000E, 0x0 } },	/* 0:	400	400	0	*/
+	{ .hsw = { 0x00E79FFF, 0x000E000C, 0x0 } },	/* 1:	400	500	2	*/
+	{ .hsw = { 0x00D75FFF, 0x0005000A, 0x0 } },	/* 2:	400	600	3.5	*/
+	{ .hsw = { 0x00FFFFFF, 0x0005000A, 0x0 } },	/* 3:	600	600	0	*/
+	{ .hsw = { 0x00E79FFF, 0x001D0007, 0x0 } },	/* 4:	600	750	2	*/
+	{ .hsw = { 0x00D75FFF, 0x000C0004, 0x0 } },	/* 5:	600	900	3.5	*/
+	{ .hsw = { 0x00FFFFFF, 0x00040006, 0x0 } },	/* 6:	800	800	0	*/
+	{ .hsw = { 0x80E79FFF, 0x00030002, 0x0 } },	/* 7:	800	1000	2	*/
+	{ .hsw = { 0x00FFFFFF, 0x00140005, 0x0 } },	/* 8:	850	850	0	*/
+	{ .hsw = { 0x00FFFFFF, 0x000C0004, 0x0 } },	/* 9:	900	900	0	*/
+	{ .hsw = { 0x00FFFFFF, 0x001C0003, 0x0 } },	/* 10:	950	950	0	*/
+	{ .hsw = { 0x80FFFFFF, 0x00030002, 0x0 } },	/* 11:	1000	1000	0	*/
+};
+
+static const struct intel_ddi_buf_trans hsw_trans_hdmi = {
+	.entries = _hsw_trans_hdmi,
+	.num_entries = ARRAY_SIZE(_hsw_trans_hdmi),
+	.hdmi_default_entry = 6,
+};
+
+static const union intel_ddi_buf_trans_entry _bdw_trans_edp[] = {
+	{ .hsw = { 0x00FFFFFF, 0x00000012, 0x0 } },
+	{ .hsw = { 0x00EBAFFF, 0x00020011, 0x0 } },
+	{ .hsw = { 0x00C71FFF, 0x0006000F, 0x0 } },
+	{ .hsw = { 0x00AAAFFF, 0x000E000A, 0x0 } },
+	{ .hsw = { 0x00FFFFFF, 0x00020011, 0x0 } },
+	{ .hsw = { 0x00DB6FFF, 0x0005000F, 0x0 } },
+	{ .hsw = { 0x00BEEFFF, 0x000A000C, 0x0 } },
+	{ .hsw = { 0x00FFFFFF, 0x0005000F, 0x0 } },
+	{ .hsw = { 0x00DB6FFF, 0x000A000C, 0x0 } },
+};
+
+static const struct intel_ddi_buf_trans bdw_trans_edp = {
+	.entries = _bdw_trans_edp,
+	.num_entries = ARRAY_SIZE(_bdw_trans_edp),
+};
+
+static const union intel_ddi_buf_trans_entry _bdw_trans_dp[] = {
+	{ .hsw = { 0x00FFFFFF, 0x0007000E, 0x0 } },
+	{ .hsw = { 0x00D75FFF, 0x000E000A, 0x0 } },
+	{ .hsw = { 0x00BEFFFF, 0x00140006, 0x0 } },
+	{ .hsw = { 0x80B2CFFF, 0x001B0002, 0x0 } },
+	{ .hsw = { 0x00FFFFFF, 0x000E000A, 0x0 } },
+	{ .hsw = { 0x00DB6FFF, 0x00160005, 0x0 } },
+	{ .hsw = { 0x80C71FFF, 0x001A0002, 0x0 } },
+	{ .hsw = { 0x00F7DFFF, 0x00180004, 0x0 } },
+	{ .hsw = { 0x80D75FFF, 0x001B0002, 0x0 } },
+};
+
+static const struct intel_ddi_buf_trans bdw_trans_dp = {
+	.entries = _bdw_trans_dp,
+	.num_entries = ARRAY_SIZE(_bdw_trans_dp),
+};
+
+static const union intel_ddi_buf_trans_entry _bdw_trans_fdi[] = {
+	{ .hsw = { 0x00FFFFFF, 0x0001000E, 0x0 } },
+	{ .hsw = { 0x00D75FFF, 0x0004000A, 0x0 } },
+	{ .hsw = { 0x00C30FFF, 0x00070006, 0x0 } },
+	{ .hsw = { 0x00AAAFFF, 0x000C0000, 0x0 } },
+	{ .hsw = { 0x00FFFFFF, 0x0004000A, 0x0 } },
+	{ .hsw = { 0x00D75FFF, 0x00090004, 0x0 } },
+	{ .hsw = { 0x00C30FFF, 0x000C0000, 0x0 } },
+	{ .hsw = { 0x00FFFFFF, 0x00070006, 0x0 } },
+	{ .hsw = { 0x00D75FFF, 0x000C0000, 0x0 } },
+};
+
+static const struct intel_ddi_buf_trans bdw_trans_fdi = {
+	.entries = _bdw_trans_fdi,
+	.num_entries = ARRAY_SIZE(_bdw_trans_fdi),
+};
+
+static const union intel_ddi_buf_trans_entry _bdw_trans_hdmi[] = {
+							/* Idx	NT mV d	T mV df	db	*/
+	{ .hsw = { 0x00FFFFFF, 0x0007000E, 0x0 } },	/* 0:	400	400	0	*/
+	{ .hsw = { 0x00D75FFF, 0x000E000A, 0x0 } },	/* 1:	400	600	3.5	*/
+	{ .hsw = { 0x00BEFFFF, 0x00140006, 0x0 } },	/* 2:	400	800	6	*/
+	{ .hsw = { 0x00FFFFFF, 0x0009000D, 0x0 } },	/* 3:	450	450	0	*/
+	{ .hsw = { 0x00FFFFFF, 0x000E000A, 0x0 } },	/* 4:	600	600	0	*/
+	{ .hsw = { 0x00D7FFFF, 0x00140006, 0x0 } },	/* 5:	600	800	2.5	*/
+	{ .hsw = { 0x80CB2FFF, 0x001B0002, 0x0 } },	/* 6:	600	1000	4.5	*/
+	{ .hsw = { 0x00FFFFFF, 0x00140006, 0x0 } },	/* 7:	800	800	0	*/
+	{ .hsw = { 0x80E79FFF, 0x001B0002, 0x0 } },	/* 8:	800	1000	2	*/
+	{ .hsw = { 0x80FFFFFF, 0x001B0002, 0x0 } },	/* 9:	1000	1000	0	*/
+};
+
+static const struct intel_ddi_buf_trans bdw_trans_hdmi = {
+	.entries = _bdw_trans_hdmi,
+	.num_entries = ARRAY_SIZE(_bdw_trans_hdmi),
+	.hdmi_default_entry = 7,
+};
+
+/* Skylake H and S */
+static const union intel_ddi_buf_trans_entry _skl_trans_dp[] = {
+	{ .hsw = { 0x00002016, 0x000000A0, 0x0 } },
+	{ .hsw = { 0x00005012, 0x0000009B, 0x0 } },
+	{ .hsw = { 0x00007011, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80009010, 0x000000C0, 0x1 } },
+	{ .hsw = { 0x00002016, 0x0000009B, 0x0 } },
+	{ .hsw = { 0x00005012, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80007011, 0x000000C0, 0x1 } },
+	{ .hsw = { 0x00002016, 0x000000DF, 0x0 } },
+	{ .hsw = { 0x80005012, 0x000000C0, 0x1 } },
+};
+
+static const struct intel_ddi_buf_trans skl_trans_dp = {
+	.entries = _skl_trans_dp,
+	.num_entries = ARRAY_SIZE(_skl_trans_dp),
+};
+
+/* Skylake U */
+static const union intel_ddi_buf_trans_entry _skl_u_trans_dp[] = {
+	{ .hsw = { 0x0000201B, 0x000000A2, 0x0 } },
+	{ .hsw = { 0x00005012, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80007011, 0x000000CD, 0x1 } },
+	{ .hsw = { 0x80009010, 0x000000C0, 0x1 } },
+	{ .hsw = { 0x0000201B, 0x0000009D, 0x0 } },
+	{ .hsw = { 0x80005012, 0x000000C0, 0x1 } },
+	{ .hsw = { 0x80007011, 0x000000C0, 0x1 } },
+	{ .hsw = { 0x00002016, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80005012, 0x000000C0, 0x1 } },
+};
+
+static const struct intel_ddi_buf_trans skl_u_trans_dp = {
+	.entries = _skl_u_trans_dp,
+	.num_entries = ARRAY_SIZE(_skl_u_trans_dp),
+};
+
+/* Skylake Y */
+static const union intel_ddi_buf_trans_entry _skl_y_trans_dp[] = {
+	{ .hsw = { 0x00000018, 0x000000A2, 0x0 } },
+	{ .hsw = { 0x00005012, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80007011, 0x000000CD, 0x3 } },
+	{ .hsw = { 0x80009010, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x00000018, 0x0000009D, 0x0 } },
+	{ .hsw = { 0x80005012, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x80007011, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x00000018, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80005012, 0x000000C0, 0x3 } },
+};
+
+static const struct intel_ddi_buf_trans skl_y_trans_dp = {
+	.entries = _skl_y_trans_dp,
+	.num_entries = ARRAY_SIZE(_skl_y_trans_dp),
+};
+
+/* Kabylake H and S */
+static const union intel_ddi_buf_trans_entry _kbl_trans_dp[] = {
+	{ .hsw = { 0x00002016, 0x000000A0, 0x0 } },
+	{ .hsw = { 0x00005012, 0x0000009B, 0x0 } },
+	{ .hsw = { 0x00007011, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80009010, 0x000000C0, 0x1 } },
+	{ .hsw = { 0x00002016, 0x0000009B, 0x0 } },
+	{ .hsw = { 0x00005012, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80007011, 0x000000C0, 0x1 } },
+	{ .hsw = { 0x00002016, 0x00000097, 0x0 } },
+	{ .hsw = { 0x80005012, 0x000000C0, 0x1 } },
+};
+
+static const struct intel_ddi_buf_trans kbl_trans_dp = {
+	.entries = _kbl_trans_dp,
+	.num_entries = ARRAY_SIZE(_kbl_trans_dp),
+};
+
+/* Kabylake U */
+static const union intel_ddi_buf_trans_entry _kbl_u_trans_dp[] = {
+	{ .hsw = { 0x0000201B, 0x000000A1, 0x0 } },
+	{ .hsw = { 0x00005012, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80007011, 0x000000CD, 0x3 } },
+	{ .hsw = { 0x80009010, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x0000201B, 0x0000009D, 0x0 } },
+	{ .hsw = { 0x80005012, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x80007011, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x00002016, 0x0000004F, 0x0 } },
+	{ .hsw = { 0x80005012, 0x000000C0, 0x3 } },
+};
+
+static const struct intel_ddi_buf_trans kbl_u_trans_dp = {
+	.entries = _kbl_u_trans_dp,
+	.num_entries = ARRAY_SIZE(_kbl_u_trans_dp),
+};
+
+/* Kabylake Y */
+static const union intel_ddi_buf_trans_entry _kbl_y_trans_dp[] = {
+	{ .hsw = { 0x00001017, 0x000000A1, 0x0 } },
+	{ .hsw = { 0x00005012, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80007011, 0x000000CD, 0x3 } },
+	{ .hsw = { 0x8000800F, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x00001017, 0x0000009D, 0x0 } },
+	{ .hsw = { 0x80005012, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x80007011, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x00001017, 0x0000004C, 0x0 } },
+	{ .hsw = { 0x80005012, 0x000000C0, 0x3 } },
+};
+
+static const struct intel_ddi_buf_trans kbl_y_trans_dp = {
+	.entries = _kbl_y_trans_dp,
+	.num_entries = ARRAY_SIZE(_kbl_y_trans_dp),
+};
+
+/*
+ * Skylake/Kabylake H and S
+ * eDP 1.4 low vswing translation parameters
+ */
+static const union intel_ddi_buf_trans_entry _skl_trans_edp[] = {
+	{ .hsw = { 0x00000018, 0x000000A8, 0x0 } },
+	{ .hsw = { 0x00004013, 0x000000A9, 0x0 } },
+	{ .hsw = { 0x00007011, 0x000000A2, 0x0 } },
+	{ .hsw = { 0x00009010, 0x0000009C, 0x0 } },
+	{ .hsw = { 0x00000018, 0x000000A9, 0x0 } },
+	{ .hsw = { 0x00006013, 0x000000A2, 0x0 } },
+	{ .hsw = { 0x00007011, 0x000000A6, 0x0 } },
+	{ .hsw = { 0x00000018, 0x000000AB, 0x0 } },
+	{ .hsw = { 0x00007013, 0x0000009F, 0x0 } },
+	{ .hsw = { 0x00000018, 0x000000DF, 0x0 } },
+};
+
+static const struct intel_ddi_buf_trans skl_trans_edp = {
+	.entries = _skl_trans_edp,
+	.num_entries = ARRAY_SIZE(_skl_trans_edp),
+};
+
+/*
+ * Skylake/Kabylake U
+ * eDP 1.4 low vswing translation parameters
+ */
+static const union intel_ddi_buf_trans_entry _skl_u_trans_edp[] = {
+	{ .hsw = { 0x00000018, 0x000000A8, 0x0 } },
+	{ .hsw = { 0x00004013, 0x000000A9, 0x0 } },
+	{ .hsw = { 0x00007011, 0x000000A2, 0x0 } },
+	{ .hsw = { 0x00009010, 0x0000009C, 0x0 } },
+	{ .hsw = { 0x00000018, 0x000000A9, 0x0 } },
+	{ .hsw = { 0x00006013, 0x000000A2, 0x0 } },
+	{ .hsw = { 0x00007011, 0x000000A6, 0x0 } },
+	{ .hsw = { 0x00002016, 0x000000AB, 0x0 } },
+	{ .hsw = { 0x00005013, 0x0000009F, 0x0 } },
+	{ .hsw = { 0x00000018, 0x000000DF, 0x0 } },
+};
+
+static const struct intel_ddi_buf_trans skl_u_trans_edp = {
+	.entries = _skl_u_trans_edp,
+	.num_entries = ARRAY_SIZE(_skl_u_trans_edp),
+};
+
+/*
+ * Skylake/Kabylake Y
+ * eDP 1.4 low vswing translation parameters
+ */
+static const union intel_ddi_buf_trans_entry _skl_y_trans_edp[] = {
+	{ .hsw = { 0x00000018, 0x000000A8, 0x0 } },
+	{ .hsw = { 0x00004013, 0x000000AB, 0x0 } },
+	{ .hsw = { 0x00007011, 0x000000A4, 0x0 } },
+	{ .hsw = { 0x00009010, 0x000000DF, 0x0 } },
+	{ .hsw = { 0x00000018, 0x000000AA, 0x0 } },
+	{ .hsw = { 0x00006013, 0x000000A4, 0x0 } },
+	{ .hsw = { 0x00007011, 0x0000009D, 0x0 } },
+	{ .hsw = { 0x00000018, 0x000000A0, 0x0 } },
+	{ .hsw = { 0x00006012, 0x000000DF, 0x0 } },
+	{ .hsw = { 0x00000018, 0x0000008A, 0x0 } },
+};
+
+static const struct intel_ddi_buf_trans skl_y_trans_edp = {
+	.entries = _skl_y_trans_edp,
+	.num_entries = ARRAY_SIZE(_skl_y_trans_edp),
+};
+
+/* Skylake/Kabylake U, H and S */
+static const union intel_ddi_buf_trans_entry _skl_trans_hdmi[] = {
+	{ .hsw = { 0x00000018, 0x000000AC, 0x0 } },
+	{ .hsw = { 0x00005012, 0x0000009D, 0x0 } },
+	{ .hsw = { 0x00007011, 0x00000088, 0x0 } },
+	{ .hsw = { 0x00000018, 0x000000A1, 0x0 } },
+	{ .hsw = { 0x00000018, 0x00000098, 0x0 } },
+	{ .hsw = { 0x00004013, 0x00000088, 0x0 } },
+	{ .hsw = { 0x80006012, 0x000000CD, 0x1 } },
+	{ .hsw = { 0x00000018, 0x000000DF, 0x0 } },
+	{ .hsw = { 0x80003015, 0x000000CD, 0x1 } },	/* Default */
+	{ .hsw = { 0x80003015, 0x000000C0, 0x1 } },
+	{ .hsw = { 0x80000018, 0x000000C0, 0x1 } },
+};
+
+static const struct intel_ddi_buf_trans skl_trans_hdmi = {
+	.entries = _skl_trans_hdmi,
+	.num_entries = ARRAY_SIZE(_skl_trans_hdmi),
+	.hdmi_default_entry = 8,
+};
+
+/* Skylake/Kabylake Y */
+static const union intel_ddi_buf_trans_entry _skl_y_trans_hdmi[] = {
+	{ .hsw = { 0x00000018, 0x000000A1, 0x0 } },
+	{ .hsw = { 0x00005012, 0x000000DF, 0x0 } },
+	{ .hsw = { 0x80007011, 0x000000CB, 0x3 } },
+	{ .hsw = { 0x00000018, 0x000000A4, 0x0 } },
+	{ .hsw = { 0x00000018, 0x0000009D, 0x0 } },
+	{ .hsw = { 0x00004013, 0x00000080, 0x0 } },
+	{ .hsw = { 0x80006013, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x00000018, 0x0000008A, 0x0 } },
+	{ .hsw = { 0x80003015, 0x000000C0, 0x3 } },	/* Default */
+	{ .hsw = { 0x80003015, 0x000000C0, 0x3 } },
+	{ .hsw = { 0x80000018, 0x000000C0, 0x3 } },
+};
+
+static const struct intel_ddi_buf_trans skl_y_trans_hdmi = {
+	.entries = _skl_y_trans_hdmi,
+	.num_entries = ARRAY_SIZE(_skl_y_trans_hdmi),
+	.hdmi_default_entry = 8,
+};
+
+static const union intel_ddi_buf_trans_entry _bxt_trans_dp[] = {
+						/* Idx	NT mV diff	db  */
+	{ .bxt = { 52,  0x9A, 0, 128, } },	/* 0:	400		0   */
+	{ .bxt = { 78,  0x9A, 0, 85,  } },	/* 1:	400		3.5 */
+	{ .bxt = { 104, 0x9A, 0, 64,  } },	/* 2:	400		6   */
+	{ .bxt = { 154, 0x9A, 0, 43,  } },	/* 3:	400		9.5 */
+	{ .bxt = { 77,  0x9A, 0, 128, } },	/* 4:	600		0   */
+	{ .bxt = { 116, 0x9A, 0, 85,  } },	/* 5:	600		3.5 */
+	{ .bxt = { 154, 0x9A, 0, 64,  } },	/* 6:	600		6   */
+	{ .bxt = { 102, 0x9A, 0, 128, } },	/* 7:	800		0   */
+	{ .bxt = { 154, 0x9A, 0, 85,  } },	/* 8:	800		3.5 */
+	{ .bxt = { 154, 0x9A, 1, 128, } },	/* 9:	1200		0   */
+};
+
+static const struct intel_ddi_buf_trans bxt_trans_dp = {
+	.entries = _bxt_trans_dp,
+	.num_entries = ARRAY_SIZE(_bxt_trans_dp),
+};
+
+static const union intel_ddi_buf_trans_entry _bxt_trans_edp[] = {
+					/* Idx	NT mV diff	db  */
+	{ .bxt = { 26, 0, 0, 128, } },	/* 0:	200		0   */
+	{ .bxt = { 38, 0, 0, 112, } },	/* 1:	200		1.5 */
+	{ .bxt = { 48, 0, 0, 96,  } },	/* 2:	200		4   */
+	{ .bxt = { 54, 0, 0, 69,  } },	/* 3:	200		6   */
+	{ .bxt = { 32, 0, 0, 128, } },	/* 4:	250		0   */
+	{ .bxt = { 48, 0, 0, 104, } },	/* 5:	250		1.5 */
+	{ .bxt = { 54, 0, 0, 85,  } },	/* 6:	250		4   */
+	{ .bxt = { 43, 0, 0, 128, } },	/* 7:	300		0   */
+	{ .bxt = { 54, 0, 0, 101, } },	/* 8:	300		1.5 */
+	{ .bxt = { 48, 0, 0, 128, } },	/* 9:	300		0   */
+};
+
+static const struct intel_ddi_buf_trans bxt_trans_edp = {
+	.entries = _bxt_trans_edp,
+	.num_entries = ARRAY_SIZE(_bxt_trans_edp),
+};
+
+/* BSpec has 2 recommended values - entries 0 and 8.
+ * Using the entry with higher vswing.
+ */
+static const union intel_ddi_buf_trans_entry _bxt_trans_hdmi[] = {
+						/* Idx	NT mV diff	db  */
+	{ .bxt = { 52,  0x9A, 0, 128, } },	/* 0:	400		0   */
+	{ .bxt = { 52,  0x9A, 0, 85,  } },	/* 1:	400		3.5 */
+	{ .bxt = { 52,  0x9A, 0, 64,  } },	/* 2:	400		6   */
+	{ .bxt = { 42,  0x9A, 0, 43,  } },	/* 3:	400		9.5 */
+	{ .bxt = { 77,  0x9A, 0, 128, } },	/* 4:	600		0   */
+	{ .bxt = { 77,  0x9A, 0, 85,  } },	/* 5:	600		3.5 */
+	{ .bxt = { 77,  0x9A, 0, 64,  } },	/* 6:	600		6   */
+	{ .bxt = { 102, 0x9A, 0, 128, } },	/* 7:	800		0   */
+	{ .bxt = { 102, 0x9A, 0, 85,  } },	/* 8:	800		3.5 */
+	{ .bxt = { 154, 0x9A, 1, 128, } },	/* 9:	1200		0   */
+};
+
+static const struct intel_ddi_buf_trans bxt_trans_hdmi = {
+	.entries = _bxt_trans_hdmi,
+	.num_entries = ARRAY_SIZE(_bxt_trans_hdmi),
+	.hdmi_default_entry = ARRAY_SIZE(_bxt_trans_hdmi) - 1,
+};
+
+/* icl_combo_phy_trans */
+static const union intel_ddi_buf_trans_entry _icl_combo_phy_trans_dp_hbr2_edp_hbr3[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x4F, 0x37, 0x00, 0x08 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x71, 0x2F, 0x00, 0x10 } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2B, 0x00, 0x14 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x4C, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x73, 0x34, 0x00, 0x0B } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x2F, 0x00, 0x10 } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x6C, 0x3C, 0x00, 0x03 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x35, 0x00, 0x0A } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans icl_combo_phy_trans_dp_hbr2_edp_hbr3 = {
+	.entries = _icl_combo_phy_trans_dp_hbr2_edp_hbr3,
+	.num_entries = ARRAY_SIZE(_icl_combo_phy_trans_dp_hbr2_edp_hbr3),
+};
+
+static const union intel_ddi_buf_trans_entry _icl_combo_phy_trans_edp_hbr2[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0x0, 0x7F, 0x3F, 0x00, 0x00 } },	/* 200   200      0.0   */
+	{ .icl = { 0x8, 0x7F, 0x38, 0x00, 0x07 } },	/* 200   250      1.9   */
+	{ .icl = { 0x1, 0x7F, 0x33, 0x00, 0x0C } },	/* 200   300      3.5   */
+	{ .icl = { 0x9, 0x7F, 0x31, 0x00, 0x0E } },	/* 200   350      4.9   */
+	{ .icl = { 0x8, 0x7F, 0x3F, 0x00, 0x00 } },	/* 250   250      0.0   */
+	{ .icl = { 0x1, 0x7F, 0x38, 0x00, 0x07 } },	/* 250   300      1.6   */
+	{ .icl = { 0x9, 0x7F, 0x35, 0x00, 0x0A } },	/* 250   350      2.9   */
+	{ .icl = { 0x1, 0x7F, 0x3F, 0x00, 0x00 } },	/* 300   300      0.0   */
+	{ .icl = { 0x9, 0x7F, 0x38, 0x00, 0x07 } },	/* 300   350      1.3   */
+	{ .icl = { 0x9, 0x7F, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+};
+
+static const struct intel_ddi_buf_trans icl_combo_phy_trans_edp_hbr2 = {
+	.entries = _icl_combo_phy_trans_edp_hbr2,
+	.num_entries = ARRAY_SIZE(_icl_combo_phy_trans_edp_hbr2),
+};
+
+static const union intel_ddi_buf_trans_entry _icl_combo_phy_trans_hdmi[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x60, 0x3F, 0x00, 0x00 } },	/* 450   450      0.0   */
+	{ .icl = { 0xB, 0x73, 0x36, 0x00, 0x09 } },	/* 450   650      3.2   */
+	{ .icl = { 0x6, 0x7F, 0x31, 0x00, 0x0E } },	/* 450   850      5.5   */
+	{ .icl = { 0xB, 0x73, 0x3F, 0x00, 0x00 } },	/* 650   650      0.0   ALS */
+	{ .icl = { 0x6, 0x7F, 0x37, 0x00, 0x08 } },	/* 650   850      2.3   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 850   850      0.0   */
+	{ .icl = { 0x6, 0x7F, 0x35, 0x00, 0x0A } },	/* 600   850      3.0   */
+};
+
+static const struct intel_ddi_buf_trans icl_combo_phy_trans_hdmi = {
+	.entries = _icl_combo_phy_trans_hdmi,
+	.num_entries = ARRAY_SIZE(_icl_combo_phy_trans_hdmi),
+	.hdmi_default_entry = ARRAY_SIZE(_icl_combo_phy_trans_hdmi) - 1,
+};
+
+static const union intel_ddi_buf_trans_entry _ehl_combo_phy_trans_dp[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x33, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x47, 0x38, 0x00, 0x07 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x64, 0x33, 0x00, 0x0C } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2F, 0x00, 0x10 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x46, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x64, 0x37, 0x00, 0x08 } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x32, 0x00, 0x0D } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x61, 0x3F, 0x00, 0x00 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x37, 0x00, 0x08 } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans ehl_combo_phy_trans_dp = {
+	.entries = _ehl_combo_phy_trans_dp,
+	.num_entries = ARRAY_SIZE(_ehl_combo_phy_trans_dp),
+};
+
+static const union intel_ddi_buf_trans_entry _ehl_combo_phy_trans_edp_hbr2[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0x8, 0x7F, 0x3F, 0x00, 0x00 } },	/* 200   200      0.0   */
+	{ .icl = { 0x8, 0x7F, 0x3F, 0x00, 0x00 } },	/* 200   250      1.9   */
+	{ .icl = { 0x1, 0x7F, 0x3D, 0x00, 0x02 } },	/* 200   300      3.5   */
+	{ .icl = { 0xA, 0x35, 0x39, 0x00, 0x06 } },	/* 200   350      4.9   */
+	{ .icl = { 0x8, 0x7F, 0x3F, 0x00, 0x00 } },	/* 250   250      0.0   */
+	{ .icl = { 0x1, 0x7F, 0x3C, 0x00, 0x03 } },	/* 250   300      1.6   */
+	{ .icl = { 0xA, 0x35, 0x39, 0x00, 0x06 } },	/* 250   350      2.9   */
+	{ .icl = { 0x1, 0x7F, 0x3F, 0x00, 0x00 } },	/* 300   300      0.0   */
+	{ .icl = { 0xA, 0x35, 0x38, 0x00, 0x07 } },	/* 300   350      1.3   */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+};
+
+static const struct intel_ddi_buf_trans ehl_combo_phy_trans_edp_hbr2 = {
+	.entries = _ehl_combo_phy_trans_edp_hbr2,
+	.num_entries = ARRAY_SIZE(_ehl_combo_phy_trans_edp_hbr2),
+};
+
+static const union intel_ddi_buf_trans_entry _jsl_combo_phy_trans_edp_hbr[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0x8, 0x7F, 0x3F, 0x00, 0x00 } },	/* 200   200      0.0   */
+	{ .icl = { 0x8, 0x7F, 0x38, 0x00, 0x07 } },	/* 200   250      1.9   */
+	{ .icl = { 0x1, 0x7F, 0x33, 0x00, 0x0C } },	/* 200   300      3.5   */
+	{ .icl = { 0xA, 0x35, 0x36, 0x00, 0x09 } },	/* 200   350      4.9   */
+	{ .icl = { 0x8, 0x7F, 0x3F, 0x00, 0x00 } },	/* 250   250      0.0   */
+	{ .icl = { 0x1, 0x7F, 0x38, 0x00, 0x07 } },	/* 250   300      1.6   */
+	{ .icl = { 0xA, 0x35, 0x35, 0x00, 0x0A } },	/* 250   350      2.9   */
+	{ .icl = { 0x1, 0x7F, 0x3F, 0x00, 0x00 } },	/* 300   300      0.0   */
+	{ .icl = { 0xA, 0x35, 0x38, 0x00, 0x07 } },	/* 300   350      1.3   */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+};
+
+static const struct intel_ddi_buf_trans jsl_combo_phy_trans_edp_hbr = {
+	.entries = _jsl_combo_phy_trans_edp_hbr,
+	.num_entries = ARRAY_SIZE(_jsl_combo_phy_trans_edp_hbr),
+};
+
+static const union intel_ddi_buf_trans_entry _jsl_combo_phy_trans_edp_hbr2[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0x8, 0x7F, 0x3F, 0x00, 0x00 } },	/* 200   200      0.0   */
+	{ .icl = { 0x8, 0x7F, 0x3F, 0x00, 0x00 } },	/* 200   250      1.9   */
+	{ .icl = { 0x1, 0x7F, 0x3D, 0x00, 0x02 } },	/* 200   300      3.5   */
+	{ .icl = { 0xA, 0x35, 0x38, 0x00, 0x07 } },	/* 200   350      4.9   */
+	{ .icl = { 0x8, 0x7F, 0x3F, 0x00, 0x00 } },	/* 250   250      0.0   */
+	{ .icl = { 0x1, 0x7F, 0x3F, 0x00, 0x00 } },	/* 250   300      1.6   */
+	{ .icl = { 0xA, 0x35, 0x3A, 0x00, 0x05 } },	/* 250   350      2.9   */
+	{ .icl = { 0x1, 0x7F, 0x3F, 0x00, 0x00 } },	/* 300   300      0.0   */
+	{ .icl = { 0xA, 0x35, 0x38, 0x00, 0x07 } },	/* 300   350      1.3   */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+};
+
+static const struct intel_ddi_buf_trans jsl_combo_phy_trans_edp_hbr2 = {
+	.entries = _jsl_combo_phy_trans_edp_hbr2,
+	.num_entries = ARRAY_SIZE(_jsl_combo_phy_trans_edp_hbr2),
+};
+
+static const union intel_ddi_buf_trans_entry _dg1_combo_phy_trans_dp_rbr_hbr[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x32, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x48, 0x35, 0x00, 0x0A } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x63, 0x2F, 0x00, 0x10 } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2C, 0x00, 0x13 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x43, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x60, 0x36, 0x00, 0x09 } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x30, 0x00, 0x0F } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x60, 0x3F, 0x00, 0x00 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x37, 0x00, 0x08 } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans dg1_combo_phy_trans_dp_rbr_hbr = {
+	.entries = _dg1_combo_phy_trans_dp_rbr_hbr,
+	.num_entries = ARRAY_SIZE(_dg1_combo_phy_trans_dp_rbr_hbr),
+};
+
+static const union intel_ddi_buf_trans_entry _dg1_combo_phy_trans_dp_hbr2_hbr3[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x32, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x48, 0x35, 0x00, 0x0A } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x63, 0x2F, 0x00, 0x10 } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2C, 0x00, 0x13 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x43, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x60, 0x36, 0x00, 0x09 } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x30, 0x00, 0x0F } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x58, 0x3F, 0x00, 0x00 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x35, 0x00, 0x0A } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans dg1_combo_phy_trans_dp_hbr2_hbr3 = {
+	.entries = _dg1_combo_phy_trans_dp_hbr2_hbr3,
+	.num_entries = ARRAY_SIZE(_dg1_combo_phy_trans_dp_hbr2_hbr3),
+};
+
+static const union intel_ddi_buf_trans_entry _icl_mg_phy_trans_rbr_hbr[] = {
+					/* Voltage swing  pre-emphasis */
+	{ .mg = { 0x18, 0x00, 0x00 } },	/* 0              0   */
+	{ .mg = { 0x1D, 0x00, 0x05 } },	/* 0              1   */
+	{ .mg = { 0x24, 0x00, 0x0C } },	/* 0              2   */
+	{ .mg = { 0x2B, 0x00, 0x14 } },	/* 0              3   */
+	{ .mg = { 0x21, 0x00, 0x00 } },	/* 1              0   */
+	{ .mg = { 0x2B, 0x00, 0x08 } },	/* 1              1   */
+	{ .mg = { 0x30, 0x00, 0x0F } },	/* 1              2   */
+	{ .mg = { 0x31, 0x00, 0x03 } },	/* 2              0   */
+	{ .mg = { 0x34, 0x00, 0x0B } },	/* 2              1   */
+	{ .mg = { 0x3F, 0x00, 0x00 } },	/* 3              0   */
+};
+
+static const struct intel_ddi_buf_trans icl_mg_phy_trans_rbr_hbr = {
+	.entries = _icl_mg_phy_trans_rbr_hbr,
+	.num_entries = ARRAY_SIZE(_icl_mg_phy_trans_rbr_hbr),
+};
+
+static const union intel_ddi_buf_trans_entry _icl_mg_phy_trans_hbr2_hbr3[] = {
+					/* Voltage swing  pre-emphasis */
+	{ .mg = { 0x18, 0x00, 0x00 } },	/* 0              0   */
+	{ .mg = { 0x1D, 0x00, 0x05 } },	/* 0              1   */
+	{ .mg = { 0x24, 0x00, 0x0C } },	/* 0              2   */
+	{ .mg = { 0x2B, 0x00, 0x14 } },	/* 0              3   */
+	{ .mg = { 0x26, 0x00, 0x00 } },	/* 1              0   */
+	{ .mg = { 0x2C, 0x00, 0x07 } },	/* 1              1   */
+	{ .mg = { 0x33, 0x00, 0x0C } },	/* 1              2   */
+	{ .mg = { 0x2E, 0x00, 0x00 } },	/* 2              0   */
+	{ .mg = { 0x36, 0x00, 0x09 } },	/* 2              1   */
+	{ .mg = { 0x3F, 0x00, 0x00 } },	/* 3              0   */
+};
+
+static const struct intel_ddi_buf_trans icl_mg_phy_trans_hbr2_hbr3 = {
+	.entries = _icl_mg_phy_trans_hbr2_hbr3,
+	.num_entries = ARRAY_SIZE(_icl_mg_phy_trans_hbr2_hbr3),
+};
+
+static const union intel_ddi_buf_trans_entry _icl_mg_phy_trans_hdmi[] = {
+					/* HDMI Preset	VS	Pre-emph */
+	{ .mg = { 0x1A, 0x0, 0x0 } },	/* 1		400mV	0dB */
+	{ .mg = { 0x20, 0x0, 0x0 } },	/* 2		500mV	0dB */
+	{ .mg = { 0x29, 0x0, 0x0 } },	/* 3		650mV	0dB */
+	{ .mg = { 0x32, 0x0, 0x0 } },	/* 4		800mV	0dB */
+	{ .mg = { 0x3F, 0x0, 0x0 } },	/* 5		1000mV	0dB */
+	{ .mg = { 0x3A, 0x0, 0x5 } },	/* 6		Full	-1.5 dB */
+	{ .mg = { 0x39, 0x0, 0x6 } },	/* 7		Full	-1.8 dB */
+	{ .mg = { 0x38, 0x0, 0x7 } },	/* 8		Full	-2 dB */
+	{ .mg = { 0x37, 0x0, 0x8 } },	/* 9		Full	-2.5 dB */
+	{ .mg = { 0x36, 0x0, 0x9 } },	/* 10		Full	-3 dB */
+};
+
+static const struct intel_ddi_buf_trans icl_mg_phy_trans_hdmi = {
+	.entries = _icl_mg_phy_trans_hdmi,
+	.num_entries = ARRAY_SIZE(_icl_mg_phy_trans_hdmi),
+	.hdmi_default_entry = ARRAY_SIZE(_icl_mg_phy_trans_hdmi) - 1,
+};
+
+static const union intel_ddi_buf_trans_entry _tgl_dkl_phy_trans_dp_hbr[] = {
+					/* VS	pre-emp	Non-trans mV	Pre-emph dB */
+	{ .dkl = { 0x7, 0x0, 0x00 } },	/* 0	0	400mV		0 dB */
+	{ .dkl = { 0x5, 0x0, 0x05 } },	/* 0	1	400mV		3.5 dB */
+	{ .dkl = { 0x2, 0x0, 0x0B } },	/* 0	2	400mV		6 dB */
+	{ .dkl = { 0x0, 0x0, 0x18 } },	/* 0	3	400mV		9.5 dB */
+	{ .dkl = { 0x5, 0x0, 0x00 } },	/* 1	0	600mV		0 dB */
+	{ .dkl = { 0x2, 0x0, 0x08 } },	/* 1	1	600mV		3.5 dB */
+	{ .dkl = { 0x0, 0x0, 0x14 } },	/* 1	2	600mV		6 dB */
+	{ .dkl = { 0x2, 0x0, 0x00 } },	/* 2	0	800mV		0 dB */
+	{ .dkl = { 0x0, 0x0, 0x0B } },	/* 2	1	800mV		3.5 dB */
+	{ .dkl = { 0x0, 0x0, 0x00 } },	/* 3	0	1200mV		0 dB HDMI default */
+};
+
+static const struct intel_ddi_buf_trans tgl_dkl_phy_trans_dp_hbr = {
+	.entries = _tgl_dkl_phy_trans_dp_hbr,
+	.num_entries = ARRAY_SIZE(_tgl_dkl_phy_trans_dp_hbr),
+};
+
+static const union intel_ddi_buf_trans_entry _tgl_dkl_phy_trans_dp_hbr2[] = {
+					/* VS	pre-emp	Non-trans mV	Pre-emph dB */
+	{ .dkl = { 0x7, 0x0, 0x00 } },	/* 0	0	400mV		0 dB */
+	{ .dkl = { 0x5, 0x0, 0x05 } },	/* 0	1	400mV		3.5 dB */
+	{ .dkl = { 0x2, 0x0, 0x0B } },	/* 0	2	400mV		6 dB */
+	{ .dkl = { 0x0, 0x0, 0x19 } },	/* 0	3	400mV		9.5 dB */
+	{ .dkl = { 0x5, 0x0, 0x00 } },	/* 1	0	600mV		0 dB */
+	{ .dkl = { 0x2, 0x0, 0x08 } },	/* 1	1	600mV		3.5 dB */
+	{ .dkl = { 0x0, 0x0, 0x14 } },	/* 1	2	600mV		6 dB */
+	{ .dkl = { 0x2, 0x0, 0x00 } },	/* 2	0	800mV		0 dB */
+	{ .dkl = { 0x0, 0x0, 0x0B } },	/* 2	1	800mV		3.5 dB */
+	{ .dkl = { 0x0, 0x0, 0x00 } },	/* 3	0	1200mV		0 dB HDMI default */
+};
+
+static const struct intel_ddi_buf_trans tgl_dkl_phy_trans_dp_hbr2 = {
+	.entries = _tgl_dkl_phy_trans_dp_hbr2,
+	.num_entries = ARRAY_SIZE(_tgl_dkl_phy_trans_dp_hbr2),
+};
+
+static const union intel_ddi_buf_trans_entry _tgl_dkl_phy_trans_hdmi[] = {
+					/* HDMI Preset	VS	Pre-emph */
+	{ .dkl = { 0x7, 0x0, 0x0 } },	/* 1		400mV	0dB */
+	{ .dkl = { 0x6, 0x0, 0x0 } },	/* 2		500mV	0dB */
+	{ .dkl = { 0x4, 0x0, 0x0 } },	/* 3		650mV	0dB */
+	{ .dkl = { 0x2, 0x0, 0x0 } },	/* 4		800mV	0dB */
+	{ .dkl = { 0x0, 0x0, 0x0 } },	/* 5		1000mV	0dB */
+	{ .dkl = { 0x0, 0x0, 0x5 } },	/* 6		Full	-1.5 dB */
+	{ .dkl = { 0x0, 0x0, 0x6 } },	/* 7		Full	-1.8 dB */
+	{ .dkl = { 0x0, 0x0, 0x7 } },	/* 8		Full	-2 dB */
+	{ .dkl = { 0x0, 0x0, 0x8 } },	/* 9		Full	-2.5 dB */
+	{ .dkl = { 0x0, 0x0, 0xA } },	/* 10		Full	-3 dB */
+};
+
+static const struct intel_ddi_buf_trans tgl_dkl_phy_trans_hdmi = {
+	.entries = _tgl_dkl_phy_trans_hdmi,
+	.num_entries = ARRAY_SIZE(_tgl_dkl_phy_trans_hdmi),
+	.hdmi_default_entry = ARRAY_SIZE(_tgl_dkl_phy_trans_hdmi) - 1,
+};
+
+static const union intel_ddi_buf_trans_entry _tgl_combo_phy_trans_dp_hbr[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x32, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x4F, 0x37, 0x00, 0x08 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x71, 0x2F, 0x00, 0x10 } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7D, 0x2B, 0x00, 0x14 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x4C, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x73, 0x34, 0x00, 0x0B } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x2F, 0x00, 0x10 } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x6C, 0x3C, 0x00, 0x03 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x35, 0x00, 0x0A } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans tgl_combo_phy_trans_dp_hbr = {
+	.entries = _tgl_combo_phy_trans_dp_hbr,
+	.num_entries = ARRAY_SIZE(_tgl_combo_phy_trans_dp_hbr),
+};
+
+static const union intel_ddi_buf_trans_entry _tgl_combo_phy_trans_dp_hbr2[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x4F, 0x37, 0x00, 0x08 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x63, 0x2F, 0x00, 0x10 } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2B, 0x00, 0x14 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x47, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x63, 0x34, 0x00, 0x0B } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x2F, 0x00, 0x10 } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x61, 0x3C, 0x00, 0x03 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7B, 0x35, 0x00, 0x0A } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans tgl_combo_phy_trans_dp_hbr2 = {
+	.entries = _tgl_combo_phy_trans_dp_hbr2,
+	.num_entries = ARRAY_SIZE(_tgl_combo_phy_trans_dp_hbr2),
+};
+
+static const union intel_ddi_buf_trans_entry _tgl_uy_combo_phy_trans_dp_hbr2[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x4F, 0x36, 0x00, 0x09 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x60, 0x32, 0x00, 0x0D } },	/* 350   700      6.0   */
+	{ .icl = { 0xC, 0x7F, 0x2D, 0x00, 0x12 } },	/* 350   900      8.2   */
+	{ .icl = { 0xC, 0x47, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x6F, 0x36, 0x00, 0x09 } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7D, 0x32, 0x00, 0x0D } },	/* 500   900      5.1   */
+	{ .icl = { 0x6, 0x60, 0x3C, 0x00, 0x03 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x34, 0x00, 0x0B } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans tgl_uy_combo_phy_trans_dp_hbr2 = {
+	.entries = _tgl_uy_combo_phy_trans_dp_hbr2,
+	.num_entries = ARRAY_SIZE(_tgl_uy_combo_phy_trans_dp_hbr2),
+};
+
+/*
+ * Cloned the HOBL entry to comply with the voltage and pre-emphasis entries
+ * that DisplayPort specification requires
+ */
+static const union intel_ddi_buf_trans_entry _tgl_combo_phy_trans_edp_hbr2_hobl[] = {
+							/* VS	pre-emp	*/
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 0	0	*/
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 0	1	*/
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 0	2	*/
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 0	3	*/
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 1	0	*/
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 1	1	*/
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 1	2	*/
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 2	0	*/
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 2	1	*/
+};
+
+static const struct intel_ddi_buf_trans tgl_combo_phy_trans_edp_hbr2_hobl = {
+	.entries = _tgl_combo_phy_trans_edp_hbr2_hobl,
+	.num_entries = ARRAY_SIZE(_tgl_combo_phy_trans_edp_hbr2_hobl),
+};
+
+static const union intel_ddi_buf_trans_entry _rkl_combo_phy_trans_dp_hbr[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x2F, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x4F, 0x37, 0x00, 0x08 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x63, 0x2F, 0x00, 0x10 } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7D, 0x2A, 0x00, 0x15 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x4C, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x73, 0x34, 0x00, 0x0B } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x2F, 0x00, 0x10 } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x6E, 0x3E, 0x00, 0x01 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x35, 0x00, 0x0A } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans rkl_combo_phy_trans_dp_hbr = {
+	.entries = _rkl_combo_phy_trans_dp_hbr,
+	.num_entries = ARRAY_SIZE(_rkl_combo_phy_trans_dp_hbr),
+};
+
+static const union intel_ddi_buf_trans_entry _rkl_combo_phy_trans_dp_hbr2_hbr3[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x50, 0x38, 0x00, 0x07 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x61, 0x33, 0x00, 0x0C } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2E, 0x00, 0x11 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x47, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x5F, 0x38, 0x00, 0x07 } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x2F, 0x00, 0x10 } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x5F, 0x3F, 0x00, 0x00 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7E, 0x36, 0x00, 0x09 } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans rkl_combo_phy_trans_dp_hbr2_hbr3 = {
+	.entries = _rkl_combo_phy_trans_dp_hbr2_hbr3,
+	.num_entries = ARRAY_SIZE(_rkl_combo_phy_trans_dp_hbr2_hbr3),
+};
+
+static const union intel_ddi_buf_trans_entry _adls_combo_phy_trans_dp_hbr2_hbr3[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x4F, 0x37, 0x00, 0x08 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x63, 0x31, 0x00, 0x0E } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2C, 0x00, 0x13 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x47, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x63, 0x37, 0x00, 0x08 } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x73, 0x32, 0x00, 0x0D } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x58, 0x3F, 0x00, 0x00 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x35, 0x00, 0x0A } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans adls_combo_phy_trans_dp_hbr2_hbr3 = {
+	.entries = _adls_combo_phy_trans_dp_hbr2_hbr3,
+	.num_entries = ARRAY_SIZE(_adls_combo_phy_trans_dp_hbr2_hbr3),
+};
+
+static const union intel_ddi_buf_trans_entry _adls_combo_phy_trans_edp_hbr2[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0x9, 0x73, 0x3D, 0x00, 0x02 } },	/* 200   200      0.0   */
+	{ .icl = { 0x9, 0x7A, 0x3C, 0x00, 0x03 } },	/* 200   250      1.9   */
+	{ .icl = { 0x9, 0x7F, 0x3B, 0x00, 0x04 } },	/* 200   300      3.5   */
+	{ .icl = { 0x4, 0x6C, 0x33, 0x00, 0x0C } },	/* 200   350      4.9   */
+	{ .icl = { 0x2, 0x73, 0x3A, 0x00, 0x05 } },	/* 250   250      0.0   */
+	{ .icl = { 0x2, 0x7C, 0x38, 0x00, 0x07 } },	/* 250   300      1.6   */
+	{ .icl = { 0x4, 0x5A, 0x36, 0x00, 0x09 } },	/* 250   350      2.9   */
+	{ .icl = { 0x4, 0x57, 0x3D, 0x00, 0x02 } },	/* 300   300      0.0   */
+	{ .icl = { 0x4, 0x65, 0x38, 0x00, 0x07 } },	/* 300   350      1.3   */
+	{ .icl = { 0x4, 0x6C, 0x3A, 0x00, 0x05 } },	/* 350   350      0.0   */
+};
+
+static const struct intel_ddi_buf_trans adls_combo_phy_trans_edp_hbr2 = {
+	.entries = _adls_combo_phy_trans_edp_hbr2,
+	.num_entries = ARRAY_SIZE(_adls_combo_phy_trans_edp_hbr2),
+};
+
+static const union intel_ddi_buf_trans_entry _adls_combo_phy_trans_edp_hbr3[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x4F, 0x37, 0x00, 0x08 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x63, 0x31, 0x00, 0x0E } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2C, 0x00, 0x13 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x47, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x63, 0x37, 0x00, 0x08 } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x73, 0x32, 0x00, 0x0D } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x58, 0x3F, 0x00, 0x00 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x35, 0x00, 0x0A } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans adls_combo_phy_trans_edp_hbr3 = {
+	.entries = _adls_combo_phy_trans_edp_hbr3,
+	.num_entries = ARRAY_SIZE(_adls_combo_phy_trans_edp_hbr3),
+};
+
+static const union intel_ddi_buf_trans_entry _adlp_combo_phy_trans_dp_hbr[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x4F, 0x37, 0x00, 0x08 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x71, 0x31, 0x00, 0x0E } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2C, 0x00, 0x13 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x4C, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x73, 0x34, 0x00, 0x0B } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x2F, 0x00, 0x10 } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x7C, 0x3C, 0x00, 0x03 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x35, 0x00, 0x0A } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans adlp_combo_phy_trans_dp_hbr = {
+	.entries = _adlp_combo_phy_trans_dp_hbr,
+	.num_entries = ARRAY_SIZE(_adlp_combo_phy_trans_dp_hbr),
+};
+
+static const union intel_ddi_buf_trans_entry _adlp_combo_phy_trans_dp_hbr2_hbr3[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x4F, 0x37, 0x00, 0x08 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x71, 0x30, 0x00, 0x0F } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2B, 0x00, 0x14 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x4C, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x73, 0x34, 0x00, 0x0B } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x30, 0x00, 0x0F } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x63, 0x3F, 0x00, 0x00 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x38, 0x00, 0x07 } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const union intel_ddi_buf_trans_entry _adlp_combo_phy_trans_edp_hbr2[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0x4, 0x50, 0x38, 0x00, 0x07 } },	/* 200   200      0.0   */
+	{ .icl = { 0x4, 0x58, 0x35, 0x00, 0x0A } },	/* 200   250      1.9   */
+	{ .icl = { 0x4, 0x60, 0x34, 0x00, 0x0B } },	/* 200   300      3.5   */
+	{ .icl = { 0x4, 0x6A, 0x32, 0x00, 0x0D } },	/* 200   350      4.9   */
+	{ .icl = { 0x4, 0x5E, 0x38, 0x00, 0x07 } },	/* 250   250      0.0   */
+	{ .icl = { 0x4, 0x61, 0x36, 0x00, 0x09 } },	/* 250   300      1.6   */
+	{ .icl = { 0x4, 0x6B, 0x34, 0x00, 0x0B } },	/* 250   350      2.9   */
+	{ .icl = { 0x4, 0x69, 0x39, 0x00, 0x06 } },	/* 300   300      0.0   */
+	{ .icl = { 0x4, 0x73, 0x37, 0x00, 0x08 } },	/* 300   350      1.3   */
+	{ .icl = { 0x4, 0x7A, 0x38, 0x00, 0x07 } },	/* 350   350      0.0   */
+};
+
+static const union intel_ddi_buf_trans_entry _adlp_combo_phy_trans_dp_hbr2_edp_hbr3[] = {
+							/* NT mV Trans mV db    */
+	{ .icl = { 0xA, 0x35, 0x3F, 0x00, 0x00 } },	/* 350   350      0.0   */
+	{ .icl = { 0xA, 0x4F, 0x37, 0x00, 0x08 } },	/* 350   500      3.1   */
+	{ .icl = { 0xC, 0x71, 0x30, 0x00, 0x0f } },	/* 350   700      6.0   */
+	{ .icl = { 0x6, 0x7F, 0x2B, 0x00, 0x14 } },	/* 350   900      8.2   */
+	{ .icl = { 0xA, 0x4C, 0x3F, 0x00, 0x00 } },	/* 500   500      0.0   */
+	{ .icl = { 0xC, 0x73, 0x34, 0x00, 0x0B } },	/* 500   700      2.9   */
+	{ .icl = { 0x6, 0x7F, 0x30, 0x00, 0x0F } },	/* 500   900      5.1   */
+	{ .icl = { 0xC, 0x63, 0x3F, 0x00, 0x00 } },	/* 650   700      0.6   */
+	{ .icl = { 0x6, 0x7F, 0x38, 0x00, 0x07 } },	/* 600   900      3.5   */
+	{ .icl = { 0x6, 0x7F, 0x3F, 0x00, 0x00 } },	/* 900   900      0.0   */
+};
+
+static const struct intel_ddi_buf_trans adlp_combo_phy_trans_dp_hbr2_hbr3 = {
+	.entries = _adlp_combo_phy_trans_dp_hbr2_hbr3,
+	.num_entries = ARRAY_SIZE(_adlp_combo_phy_trans_dp_hbr2_hbr3),
+};
+
+static const struct intel_ddi_buf_trans adlp_combo_phy_trans_edp_hbr3 = {
+	.entries = _adlp_combo_phy_trans_dp_hbr2_edp_hbr3,
+	.num_entries = ARRAY_SIZE(_adlp_combo_phy_trans_dp_hbr2_edp_hbr3),
+};
+
+static const struct intel_ddi_buf_trans adlp_combo_phy_trans_edp_up_to_hbr2 = {
+	.entries = _adlp_combo_phy_trans_edp_hbr2,
+	.num_entries = ARRAY_SIZE(_adlp_combo_phy_trans_edp_hbr2),
+};
+
+static const union intel_ddi_buf_trans_entry _adlp_dkl_phy_trans_dp_hbr[] = {
+					/* VS	pre-emp	Non-trans mV	Pre-emph dB */
+	{ .dkl = { 0x7, 0x0, 0x01 } },	/* 0	0	400mV		0 dB */
+	{ .dkl = { 0x5, 0x0, 0x06 } },	/* 0	1	400mV		3.5 dB */
+	{ .dkl = { 0x2, 0x0, 0x0B } },	/* 0	2	400mV		6 dB */
+	{ .dkl = { 0x0, 0x0, 0x17 } },	/* 0	3	400mV		9.5 dB */
+	{ .dkl = { 0x5, 0x0, 0x00 } },	/* 1	0	600mV		0 dB */
+	{ .dkl = { 0x2, 0x0, 0x08 } },	/* 1	1	600mV		3.5 dB */
+	{ .dkl = { 0x0, 0x0, 0x14 } },	/* 1	2	600mV		6 dB */
+	{ .dkl = { 0x2, 0x0, 0x00 } },	/* 2	0	800mV		0 dB */
+	{ .dkl = { 0x0, 0x0, 0x0B } },	/* 2	1	800mV		3.5 dB */
+	{ .dkl = { 0x0, 0x0, 0x00 } },	/* 3	0	1200mV		0 dB */
+};
+
+static const struct intel_ddi_buf_trans adlp_dkl_phy_trans_dp_hbr = {
+	.entries = _adlp_dkl_phy_trans_dp_hbr,
+	.num_entries = ARRAY_SIZE(_adlp_dkl_phy_trans_dp_hbr),
+};
+
+static const union intel_ddi_buf_trans_entry _adlp_dkl_phy_trans_dp_hbr2_hbr3[] = {
+					/* VS	pre-emp	Non-trans mV	Pre-emph dB */
+	{ .dkl = { 0x7, 0x0, 0x00 } },	/* 0	0	400mV		0 dB */
+	{ .dkl = { 0x5, 0x0, 0x04 } },	/* 0	1	400mV		3.5 dB */
+	{ .dkl = { 0x2, 0x0, 0x0A } },	/* 0	2	400mV		6 dB */
+	{ .dkl = { 0x0, 0x0, 0x18 } },	/* 0	3	400mV		9.5 dB */
+	{ .dkl = { 0x5, 0x0, 0x00 } },	/* 1	0	600mV		0 dB */
+	{ .dkl = { 0x2, 0x0, 0x06 } },	/* 1	1	600mV		3.5 dB */
+	{ .dkl = { 0x0, 0x0, 0x14 } },	/* 1	2	600mV		6 dB */
+	{ .dkl = { 0x2, 0x0, 0x00 } },	/* 2	0	800mV		0 dB */
+	{ .dkl = { 0x0, 0x0, 0x09 } },	/* 2	1	800mV		3.5 dB */
+	{ .dkl = { 0x0, 0x0, 0x00 } },	/* 3	0	1200mV		0 dB */
+};
+
+static const struct intel_ddi_buf_trans adlp_dkl_phy_trans_dp_hbr2_hbr3 = {
+	.entries = _adlp_dkl_phy_trans_dp_hbr2_hbr3,
+	.num_entries = ARRAY_SIZE(_adlp_dkl_phy_trans_dp_hbr2_hbr3),
+};
+
+static const union intel_ddi_buf_trans_entry _dg2_snps_trans[] = {
+	{ .snps = { 25, 0, 0 } },	/* VS 0, pre-emph 0 */
+	{ .snps = { 32, 0, 6 } },	/* VS 0, pre-emph 1 */
+	{ .snps = { 35, 0, 10 } },	/* VS 0, pre-emph 2 */
+	{ .snps = { 43, 0, 17 } },	/* VS 0, pre-emph 3 */
+	{ .snps = { 35, 0, 0 } },	/* VS 1, pre-emph 0 */
+	{ .snps = { 45, 0, 8 } },	/* VS 1, pre-emph 1 */
+	{ .snps = { 48, 0, 14 } },	/* VS 1, pre-emph 2 */
+	{ .snps = { 47, 0, 0 } },	/* VS 2, pre-emph 0 */
+	{ .snps = { 55, 0, 7 } },	/* VS 2, pre-emph 1 */
+	{ .snps = { 62, 0, 0 } },	/* VS 3, pre-emph 0 */
+};
+
+static const struct intel_ddi_buf_trans dg2_snps_trans = {
+	.entries = _dg2_snps_trans,
+	.num_entries = ARRAY_SIZE(_dg2_snps_trans),
+	.hdmi_default_entry = ARRAY_SIZE(_dg2_snps_trans) - 1,
+};
+
+static const union intel_ddi_buf_trans_entry _dg2_snps_trans_uhbr[] = {
+	{ .snps = { 62, 0, 0 } },	/* preset 0 */
+	{ .snps = { 55, 0, 7 } },	/* preset 1 */
+	{ .snps = { 50, 0, 12 } },	/* preset 2 */
+	{ .snps = { 44, 0, 18 } },	/* preset 3 */
+	{ .snps = { 35, 0, 21 } },	/* preset 4 */
+	{ .snps = { 59, 3, 0 } },	/* preset 5 */
+	{ .snps = { 53, 3, 6 } },	/* preset 6 */
+	{ .snps = { 48, 3, 11 } },	/* preset 7 */
+	{ .snps = { 42, 5, 15 } },	/* preset 8 */
+	{ .snps = { 37, 5, 20 } },	/* preset 9 */
+	{ .snps = { 56, 6, 0 } },	/* preset 10 */
+	{ .snps = { 48, 7, 7 } },	/* preset 11 */
+	{ .snps = { 45, 7, 10 } },	/* preset 12 */
+	{ .snps = { 39, 8, 15 } },	/* preset 13 */
+	{ .snps = { 48, 14, 0 } },	/* preset 14 */
+	{ .snps = { 45, 4, 4 } },	/* preset 15 */
+};
+
+static const struct intel_ddi_buf_trans dg2_snps_trans_uhbr = {
+	.entries = _dg2_snps_trans_uhbr,
+	.num_entries = ARRAY_SIZE(_dg2_snps_trans_uhbr),
+};
+
+bool is_hobl_buf_trans(const struct intel_ddi_buf_trans *table)
+{
+	return table == &tgl_combo_phy_trans_edp_hbr2_hobl;
+}
+
+static bool use_edp_hobl(struct intel_encoder *encoder)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_connector *connector = intel_dp->attached_connector;
+
+	return connector->panel.vbt.edp.hobl && !intel_dp->hobl_failed;
+}
+
+static bool use_edp_low_vswing(struct intel_encoder *encoder)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_connector *connector = intel_dp->attached_connector;
+
+	return connector->panel.vbt.edp.low_vswing;
+}
+
+static const struct intel_ddi_buf_trans *
+intel_get_buf_trans(const struct intel_ddi_buf_trans *trans, int *num_entries)
+{
+	*num_entries = trans->num_entries;
+	return trans;
+}
+
+static const struct intel_ddi_buf_trans *
+hsw_get_buf_trans(struct intel_encoder *encoder,
+		  const struct intel_crtc_state *crtc_state,
+		  int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG))
+		return intel_get_buf_trans(&hsw_trans_fdi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&hsw_trans_hdmi, n_entries);
+	else
+		return intel_get_buf_trans(&hsw_trans_dp, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+bdw_get_buf_trans(struct intel_encoder *encoder,
+		  const struct intel_crtc_state *crtc_state,
+		  int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG))
+		return intel_get_buf_trans(&bdw_trans_fdi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&bdw_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP) &&
+		 use_edp_low_vswing(encoder))
+		return intel_get_buf_trans(&bdw_trans_edp, n_entries);
+	else
+		return intel_get_buf_trans(&bdw_trans_dp, n_entries);
+}
+
+static int skl_buf_trans_num_entries(enum port port, int n_entries)
+{
+	/* Only DDIA and DDIE can select the 10th register with DP */
+	if (port == PORT_A || port == PORT_E)
+		return min(n_entries, 10);
+	else
+		return min(n_entries, 9);
+}
+
+static const struct intel_ddi_buf_trans *
+_skl_get_buf_trans_dp(struct intel_encoder *encoder,
+		      const struct intel_ddi_buf_trans *trans,
+		      int *n_entries)
+{
+	trans = intel_get_buf_trans(trans, n_entries);
+	*n_entries = skl_buf_trans_num_entries(encoder->port, *n_entries);
+	return trans;
+}
+
+static const struct intel_ddi_buf_trans *
+skl_y_get_buf_trans(struct intel_encoder *encoder,
+		    const struct intel_crtc_state *crtc_state,
+		    int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&skl_y_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP) &&
+		 use_edp_low_vswing(encoder))
+		return _skl_get_buf_trans_dp(encoder, &skl_y_trans_edp, n_entries);
+	else
+		return _skl_get_buf_trans_dp(encoder, &skl_y_trans_dp, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+skl_u_get_buf_trans(struct intel_encoder *encoder,
+		    const struct intel_crtc_state *crtc_state,
+		    int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&skl_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP) &&
+		 use_edp_low_vswing(encoder))
+		return _skl_get_buf_trans_dp(encoder, &skl_u_trans_edp, n_entries);
+	else
+		return _skl_get_buf_trans_dp(encoder, &skl_u_trans_dp, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+skl_get_buf_trans(struct intel_encoder *encoder,
+		  const struct intel_crtc_state *crtc_state,
+		  int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&skl_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP) &&
+		 use_edp_low_vswing(encoder))
+		return _skl_get_buf_trans_dp(encoder, &skl_trans_edp, n_entries);
+	else
+		return _skl_get_buf_trans_dp(encoder, &skl_trans_dp, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+kbl_y_get_buf_trans(struct intel_encoder *encoder,
+		    const struct intel_crtc_state *crtc_state,
+		    int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&skl_y_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP) &&
+		 use_edp_low_vswing(encoder))
+		return _skl_get_buf_trans_dp(encoder, &skl_y_trans_edp, n_entries);
+	else
+		return _skl_get_buf_trans_dp(encoder, &kbl_y_trans_dp, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+kbl_u_get_buf_trans(struct intel_encoder *encoder,
+		    const struct intel_crtc_state *crtc_state,
+		    int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&skl_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP) &&
+		 use_edp_low_vswing(encoder))
+		return _skl_get_buf_trans_dp(encoder, &skl_u_trans_edp, n_entries);
+	else
+		return _skl_get_buf_trans_dp(encoder, &kbl_u_trans_dp, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+kbl_get_buf_trans(struct intel_encoder *encoder,
+		  const struct intel_crtc_state *crtc_state,
+		  int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&skl_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP) &&
+		 use_edp_low_vswing(encoder))
+		return _skl_get_buf_trans_dp(encoder, &skl_trans_edp, n_entries);
+	else
+		return _skl_get_buf_trans_dp(encoder, &kbl_trans_dp, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+bxt_get_buf_trans(struct intel_encoder *encoder,
+		  const struct intel_crtc_state *crtc_state,
+		  int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&bxt_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP) &&
+		 use_edp_low_vswing(encoder))
+		return intel_get_buf_trans(&bxt_trans_edp, n_entries);
+	else
+		return intel_get_buf_trans(&bxt_trans_dp, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+icl_get_combo_buf_trans_dp(struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state,
+			   int *n_entries)
+{
+	return intel_get_buf_trans(&icl_combo_phy_trans_dp_hbr2_edp_hbr3,
+				   n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+icl_get_combo_buf_trans_edp(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    int *n_entries)
+{
+	if (crtc_state->port_clock > 540000) {
+		return intel_get_buf_trans(&icl_combo_phy_trans_dp_hbr2_edp_hbr3,
+					   n_entries);
+	} else if (use_edp_low_vswing(encoder)) {
+		return intel_get_buf_trans(&icl_combo_phy_trans_edp_hbr2,
+					   n_entries);
+	}
+
+	return icl_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+icl_get_combo_buf_trans(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state,
+			int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&icl_combo_phy_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP))
+		return icl_get_combo_buf_trans_edp(encoder, crtc_state, n_entries);
+	else
+		return icl_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+icl_get_mg_buf_trans_dp(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state,
+			int *n_entries)
+{
+	if (crtc_state->port_clock > 270000) {
+		return intel_get_buf_trans(&icl_mg_phy_trans_hbr2_hbr3,
+					   n_entries);
+	} else {
+		return intel_get_buf_trans(&icl_mg_phy_trans_rbr_hbr,
+					   n_entries);
+	}
+}
+
+static const struct intel_ddi_buf_trans *
+icl_get_mg_buf_trans(struct intel_encoder *encoder,
+		     const struct intel_crtc_state *crtc_state,
+		     int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&icl_mg_phy_trans_hdmi, n_entries);
+	else
+		return icl_get_mg_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+ehl_get_combo_buf_trans_edp(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    int *n_entries)
+{
+	if (crtc_state->port_clock > 270000)
+		return intel_get_buf_trans(&ehl_combo_phy_trans_edp_hbr2, n_entries);
+	else
+		return intel_get_buf_trans(&icl_combo_phy_trans_edp_hbr2, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+ehl_get_combo_buf_trans(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state,
+			int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&icl_combo_phy_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP) &&
+		 use_edp_low_vswing(encoder))
+		return ehl_get_combo_buf_trans_edp(encoder, crtc_state, n_entries);
+	else
+		return intel_get_buf_trans(&ehl_combo_phy_trans_dp, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+jsl_get_combo_buf_trans_edp(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    int *n_entries)
+{
+	if (crtc_state->port_clock > 270000)
+		return intel_get_buf_trans(&jsl_combo_phy_trans_edp_hbr2, n_entries);
+	else
+		return intel_get_buf_trans(&jsl_combo_phy_trans_edp_hbr, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+jsl_get_combo_buf_trans(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state,
+			int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&icl_combo_phy_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP) &&
+		 use_edp_low_vswing(encoder))
+		return jsl_get_combo_buf_trans_edp(encoder, crtc_state, n_entries);
+	else
+		return intel_get_buf_trans(&icl_combo_phy_trans_dp_hbr2_edp_hbr3, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+tgl_get_combo_buf_trans_dp(struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state,
+			   int *n_entries)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (crtc_state->port_clock > 270000) {
+		if (IS_TGL_UY(dev_priv)) {
+			return intel_get_buf_trans(&tgl_uy_combo_phy_trans_dp_hbr2,
+						   n_entries);
+		} else {
+			return intel_get_buf_trans(&tgl_combo_phy_trans_dp_hbr2,
+						   n_entries);
+		}
+	} else {
+		return intel_get_buf_trans(&tgl_combo_phy_trans_dp_hbr,
+					   n_entries);
+	}
+}
+
+static const struct intel_ddi_buf_trans *
+tgl_get_combo_buf_trans_edp(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    int *n_entries)
+{
+	if (crtc_state->port_clock > 540000) {
+		return intel_get_buf_trans(&icl_combo_phy_trans_dp_hbr2_edp_hbr3,
+					   n_entries);
+	} else if (use_edp_hobl(encoder)) {
+		return intel_get_buf_trans(&tgl_combo_phy_trans_edp_hbr2_hobl,
+					   n_entries);
+	} else if (use_edp_low_vswing(encoder)) {
+		return intel_get_buf_trans(&icl_combo_phy_trans_edp_hbr2,
+					   n_entries);
+	}
+
+	return tgl_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+tgl_get_combo_buf_trans(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state,
+			int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&icl_combo_phy_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP))
+		return tgl_get_combo_buf_trans_edp(encoder, crtc_state, n_entries);
+	else
+		return tgl_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+dg1_get_combo_buf_trans_dp(struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state,
+			   int *n_entries)
+{
+	if (crtc_state->port_clock > 270000)
+		return intel_get_buf_trans(&dg1_combo_phy_trans_dp_hbr2_hbr3,
+					   n_entries);
+	else
+		return intel_get_buf_trans(&dg1_combo_phy_trans_dp_rbr_hbr,
+					   n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+dg1_get_combo_buf_trans_edp(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    int *n_entries)
+{
+	if (crtc_state->port_clock > 540000)
+		return intel_get_buf_trans(&icl_combo_phy_trans_dp_hbr2_edp_hbr3,
+					   n_entries);
+	else if (use_edp_hobl(encoder))
+		return intel_get_buf_trans(&tgl_combo_phy_trans_edp_hbr2_hobl,
+					   n_entries);
+	else if (use_edp_low_vswing(encoder))
+		return intel_get_buf_trans(&icl_combo_phy_trans_edp_hbr2,
+					   n_entries);
+	else
+		return dg1_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+dg1_get_combo_buf_trans(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state,
+			int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&icl_combo_phy_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP))
+		return dg1_get_combo_buf_trans_edp(encoder, crtc_state, n_entries);
+	else
+		return dg1_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+rkl_get_combo_buf_trans_dp(struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state,
+			   int *n_entries)
+{
+	if (crtc_state->port_clock > 270000)
+		return intel_get_buf_trans(&rkl_combo_phy_trans_dp_hbr2_hbr3, n_entries);
+	else
+		return intel_get_buf_trans(&rkl_combo_phy_trans_dp_hbr, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+rkl_get_combo_buf_trans_edp(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    int *n_entries)
+{
+	if (crtc_state->port_clock > 540000) {
+		return intel_get_buf_trans(&icl_combo_phy_trans_dp_hbr2_edp_hbr3,
+					   n_entries);
+	} else if (use_edp_hobl(encoder)) {
+		return intel_get_buf_trans(&tgl_combo_phy_trans_edp_hbr2_hobl,
+					   n_entries);
+	} else if (use_edp_low_vswing(encoder)) {
+		return intel_get_buf_trans(&icl_combo_phy_trans_edp_hbr2,
+					   n_entries);
+	}
+
+	return rkl_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+rkl_get_combo_buf_trans(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state,
+			int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&icl_combo_phy_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP))
+		return rkl_get_combo_buf_trans_edp(encoder, crtc_state, n_entries);
+	else
+		return rkl_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+adls_get_combo_buf_trans_dp(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    int *n_entries)
+{
+	if (crtc_state->port_clock > 270000)
+		return intel_get_buf_trans(&adls_combo_phy_trans_dp_hbr2_hbr3, n_entries);
+	else
+		return intel_get_buf_trans(&tgl_combo_phy_trans_dp_hbr, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+adls_get_combo_buf_trans_edp(struct intel_encoder *encoder,
+			     const struct intel_crtc_state *crtc_state,
+			     int *n_entries)
+{
+	if (crtc_state->port_clock > 540000)
+		return intel_get_buf_trans(&adls_combo_phy_trans_edp_hbr3, n_entries);
+	else if (use_edp_hobl(encoder))
+		return intel_get_buf_trans(&tgl_combo_phy_trans_edp_hbr2_hobl, n_entries);
+	else if (use_edp_low_vswing(encoder))
+		return intel_get_buf_trans(&adls_combo_phy_trans_edp_hbr2, n_entries);
+	else
+		return adls_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+adls_get_combo_buf_trans(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *crtc_state,
+			 int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&icl_combo_phy_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP))
+		return adls_get_combo_buf_trans_edp(encoder, crtc_state, n_entries);
+	else
+		return adls_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+adlp_get_combo_buf_trans_dp(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    int *n_entries)
+{
+	if (crtc_state->port_clock > 270000)
+		return intel_get_buf_trans(&adlp_combo_phy_trans_dp_hbr2_hbr3, n_entries);
+	else
+		return intel_get_buf_trans(&adlp_combo_phy_trans_dp_hbr, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+adlp_get_combo_buf_trans_edp(struct intel_encoder *encoder,
+			     const struct intel_crtc_state *crtc_state,
+			     int *n_entries)
+{
+	if (crtc_state->port_clock > 540000) {
+		return intel_get_buf_trans(&adlp_combo_phy_trans_edp_hbr3,
+					   n_entries);
+	} else if (use_edp_hobl(encoder)) {
+		return intel_get_buf_trans(&tgl_combo_phy_trans_edp_hbr2_hobl,
+					   n_entries);
+	} else if (use_edp_low_vswing(encoder)) {
+		return intel_get_buf_trans(&adlp_combo_phy_trans_edp_up_to_hbr2,
+					   n_entries);
+	}
+
+	return adlp_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+adlp_get_combo_buf_trans(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *crtc_state,
+			 int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&icl_combo_phy_trans_hdmi, n_entries);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP))
+		return adlp_get_combo_buf_trans_edp(encoder, crtc_state, n_entries);
+	else
+		return adlp_get_combo_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+tgl_get_dkl_buf_trans_dp(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *crtc_state,
+			 int *n_entries)
+{
+	if (crtc_state->port_clock > 270000) {
+		return intel_get_buf_trans(&tgl_dkl_phy_trans_dp_hbr2,
+					   n_entries);
+	} else {
+		return intel_get_buf_trans(&tgl_dkl_phy_trans_dp_hbr,
+					   n_entries);
+	}
+}
+
+static const struct intel_ddi_buf_trans *
+tgl_get_dkl_buf_trans(struct intel_encoder *encoder,
+		      const struct intel_crtc_state *crtc_state,
+		      int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&tgl_dkl_phy_trans_hdmi, n_entries);
+	else
+		return tgl_get_dkl_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+adlp_get_dkl_buf_trans_dp(struct intel_encoder *encoder,
+			  const struct intel_crtc_state *crtc_state,
+			  int *n_entries)
+{
+	if (crtc_state->port_clock > 270000) {
+		return intel_get_buf_trans(&adlp_dkl_phy_trans_dp_hbr2_hbr3,
+					   n_entries);
+	} else {
+		return intel_get_buf_trans(&adlp_dkl_phy_trans_dp_hbr,
+					   n_entries);
+	}
+}
+
+static const struct intel_ddi_buf_trans *
+adlp_get_dkl_buf_trans(struct intel_encoder *encoder,
+		       const struct intel_crtc_state *crtc_state,
+		       int *n_entries)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return intel_get_buf_trans(&tgl_dkl_phy_trans_hdmi, n_entries);
+	else
+		return adlp_get_dkl_buf_trans_dp(encoder, crtc_state, n_entries);
+}
+
+static const struct intel_ddi_buf_trans *
+dg2_get_snps_buf_trans(struct intel_encoder *encoder,
+		       const struct intel_crtc_state *crtc_state,
+		       int *n_entries)
+{
+	if (intel_crtc_has_dp_encoder(crtc_state) &&
+	    intel_dp_is_uhbr(crtc_state))
+		return intel_get_buf_trans(&dg2_snps_trans_uhbr, n_entries);
+	else
+		return intel_get_buf_trans(&dg2_snps_trans, n_entries);
+}
+
+void intel_ddi_buf_trans_init(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+
+	if (IS_DG2(i915)) {
+		encoder->get_buf_trans = dg2_get_snps_buf_trans;
+	} else if (IS_ALDERLAKE_P(i915)) {
+		if (intel_phy_is_combo(i915, phy))
+			encoder->get_buf_trans = adlp_get_combo_buf_trans;
+		else
+			encoder->get_buf_trans = adlp_get_dkl_buf_trans;
+	} else if (IS_ALDERLAKE_S(i915)) {
+		encoder->get_buf_trans = adls_get_combo_buf_trans;
+	} else if (IS_ROCKETLAKE(i915)) {
+		encoder->get_buf_trans = rkl_get_combo_buf_trans;
+	} else if (IS_DG1(i915)) {
+		encoder->get_buf_trans = dg1_get_combo_buf_trans;
+	} else if (DISPLAY_VER(i915) >= 12) {
+		if (intel_phy_is_combo(i915, phy))
+			encoder->get_buf_trans = tgl_get_combo_buf_trans;
+		else
+			encoder->get_buf_trans = tgl_get_dkl_buf_trans;
+	} else if (DISPLAY_VER(i915) == 11) {
+		if (IS_PLATFORM(i915, INTEL_JASPERLAKE))
+			encoder->get_buf_trans = jsl_get_combo_buf_trans;
+		else if (IS_PLATFORM(i915, INTEL_ELKHARTLAKE))
+			encoder->get_buf_trans = ehl_get_combo_buf_trans;
+		else if (intel_phy_is_combo(i915, phy))
+			encoder->get_buf_trans = icl_get_combo_buf_trans;
+		else
+			encoder->get_buf_trans = icl_get_mg_buf_trans;
+	} else if (IS_GEMINILAKE(i915) || IS_BROXTON(i915)) {
+		encoder->get_buf_trans = bxt_get_buf_trans;
+	} else if (IS_CML_ULX(i915) || IS_CFL_ULX(i915) || IS_KBL_ULX(i915)) {
+		encoder->get_buf_trans = kbl_y_get_buf_trans;
+	} else if (IS_CML_ULT(i915) || IS_CFL_ULT(i915) || IS_KBL_ULT(i915)) {
+		encoder->get_buf_trans = kbl_u_get_buf_trans;
+	} else if (IS_COMETLAKE(i915) || IS_COFFEELAKE(i915) || IS_KABYLAKE(i915)) {
+		encoder->get_buf_trans = kbl_get_buf_trans;
+	} else if (IS_SKL_ULX(i915)) {
+		encoder->get_buf_trans = skl_y_get_buf_trans;
+	} else if (IS_SKL_ULT(i915)) {
+		encoder->get_buf_trans = skl_u_get_buf_trans;
+	} else if (IS_SKYLAKE(i915)) {
+		encoder->get_buf_trans = skl_get_buf_trans;
+	} else if (IS_BROADWELL(i915)) {
+		encoder->get_buf_trans = bdw_get_buf_trans;
+	} else if (IS_HASWELL(i915)) {
+		encoder->get_buf_trans = hsw_get_buf_trans;
+	} else {
+		MISSING_CASE(INTEL_INFO(i915)->platform);
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_ddi_buf_trans.h b/drivers/gpu/drm/i915/display/intel_ddi_buf_trans.h
new file mode 100644
index 000000000..2133984a5
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_ddi_buf_trans.h
@@ -0,0 +1,73 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _INTEL_DDI_BUF_TRANS_H_
+#define _INTEL_DDI_BUF_TRANS_H_
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct intel_encoder;
+struct intel_crtc_state;
+
+struct hsw_ddi_buf_trans {
+	u32 trans1;	/* balance leg enable, de-emph level */
+	u32 trans2;	/* vref sel, vswing */
+	u8 i_boost;	/* SKL: I_boost; valid: 0x0, 0x1, 0x3, 0x7 */
+};
+
+struct bxt_ddi_buf_trans {
+	u8 margin;	/* swing value */
+	u8 scale;	/* scale value */
+	u8 enable;	/* scale enable */
+	u8 deemphasis;
+};
+
+struct icl_ddi_buf_trans {
+	u8 dw2_swing_sel;
+	u8 dw7_n_scalar;
+	u8 dw4_cursor_coeff;
+	u8 dw4_post_cursor_2;
+	u8 dw4_post_cursor_1;
+};
+
+struct icl_mg_phy_ddi_buf_trans {
+	u8 cri_txdeemph_override_11_6;
+	u8 cri_txdeemph_override_5_0;
+	u8 cri_txdeemph_override_17_12;
+};
+
+struct tgl_dkl_phy_ddi_buf_trans {
+	u8 vswing;
+	u8 preshoot;
+	u8 de_emphasis;
+};
+
+struct dg2_snps_phy_buf_trans {
+	u8 vswing;
+	u8 pre_cursor;
+	u8 post_cursor;
+};
+
+union intel_ddi_buf_trans_entry {
+	struct hsw_ddi_buf_trans hsw;
+	struct bxt_ddi_buf_trans bxt;
+	struct icl_ddi_buf_trans icl;
+	struct icl_mg_phy_ddi_buf_trans mg;
+	struct tgl_dkl_phy_ddi_buf_trans dkl;
+	struct dg2_snps_phy_buf_trans snps;
+};
+
+struct intel_ddi_buf_trans {
+	const union intel_ddi_buf_trans_entry *entries;
+	u8 num_entries;
+	u8 hdmi_default_entry;
+};
+
+bool is_hobl_buf_trans(const struct intel_ddi_buf_trans *table);
+
+void intel_ddi_buf_trans_init(struct intel_encoder *encoder);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_de.h b/drivers/gpu/drm/i915/display/intel_de.h
new file mode 100644
index 000000000..9c104f65e
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_de.h
@@ -0,0 +1,84 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DE_H__
+#define __INTEL_DE_H__
+
+#include "i915_drv.h"
+#include "i915_trace.h"
+#include "intel_uncore.h"
+
+static inline u32
+intel_de_read(struct drm_i915_private *i915, i915_reg_t reg)
+{
+	return intel_uncore_read(&i915->uncore, reg);
+}
+
+static inline void
+intel_de_posting_read(struct drm_i915_private *i915, i915_reg_t reg)
+{
+	intel_uncore_posting_read(&i915->uncore, reg);
+}
+
+static inline void
+intel_de_write(struct drm_i915_private *i915, i915_reg_t reg, u32 val)
+{
+	intel_uncore_write(&i915->uncore, reg, val);
+}
+
+static inline void
+intel_de_rmw(struct drm_i915_private *i915, i915_reg_t reg, u32 clear, u32 set)
+{
+	intel_uncore_rmw(&i915->uncore, reg, clear, set);
+}
+
+static inline int
+intel_de_wait_for_register(struct drm_i915_private *i915, i915_reg_t reg,
+			   u32 mask, u32 value, unsigned int timeout)
+{
+	return intel_wait_for_register(&i915->uncore, reg, mask, value, timeout);
+}
+
+static inline int
+intel_de_wait_for_set(struct drm_i915_private *i915, i915_reg_t reg,
+		      u32 mask, unsigned int timeout)
+{
+	return intel_de_wait_for_register(i915, reg, mask, mask, timeout);
+}
+
+static inline int
+intel_de_wait_for_clear(struct drm_i915_private *i915, i915_reg_t reg,
+			u32 mask, unsigned int timeout)
+{
+	return intel_de_wait_for_register(i915, reg, mask, 0, timeout);
+}
+
+/*
+ * Unlocked mmio-accessors, think carefully before using these.
+ *
+ * Certain architectures will die if the same cacheline is concurrently accessed
+ * by different clients (e.g. on Ivybridge). Access to registers should
+ * therefore generally be serialised, by either the dev_priv->uncore.lock or
+ * a more localised lock guarding all access to that bank of registers.
+ */
+static inline u32
+intel_de_read_fw(struct drm_i915_private *i915, i915_reg_t reg)
+{
+	u32 val;
+
+	val = intel_uncore_read_fw(&i915->uncore, reg);
+	trace_i915_reg_rw(false, reg, val, sizeof(val), true);
+
+	return val;
+}
+
+static inline void
+intel_de_write_fw(struct drm_i915_private *i915, i915_reg_t reg, u32 val)
+{
+	trace_i915_reg_rw(true, reg, val, sizeof(val), true);
+	intel_uncore_write_fw(&i915->uncore, reg, val);
+}
+
+#endif /* __INTEL_DE_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
new file mode 100644
index 000000000..fb8d1d634
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display.c
@@ -0,0 +1,9140 @@
+/*
+ * Copyright © 2006-2007 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Eric Anholt <eric@anholt.net>
+ */
+
+#include <acpi/video.h>
+#include <linux/i2c.h>
+#include <linux/input.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/dma-resv.h>
+#include <linux/slab.h>
+#include <linux/string_helpers.h>
+#include <linux/vga_switcheroo.h>
+
+#include <drm/display/drm_dp_helper.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_atomic_uapi.h>
+#include <drm/drm_damage_helper.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_privacy_screen_consumer.h>
+#include <drm/drm_probe_helper.h>
+#include <drm/drm_rect.h>
+
+#include "display/intel_audio.h"
+#include "display/intel_crt.h"
+#include "display/intel_ddi.h"
+#include "display/intel_display_debugfs.h"
+#include "display/intel_display_power.h"
+#include "display/intel_dp.h"
+#include "display/intel_dp_mst.h"
+#include "display/intel_dpll.h"
+#include "display/intel_dpll_mgr.h"
+#include "display/intel_drrs.h"
+#include "display/intel_dsi.h"
+#include "display/intel_dvo.h"
+#include "display/intel_fb.h"
+#include "display/intel_gmbus.h"
+#include "display/intel_hdmi.h"
+#include "display/intel_lvds.h"
+#include "display/intel_sdvo.h"
+#include "display/intel_snps_phy.h"
+#include "display/intel_tv.h"
+#include "display/intel_vdsc.h"
+#include "display/intel_vrr.h"
+
+#include "gem/i915_gem_lmem.h"
+#include "gem/i915_gem_object.h"
+
+#include "gt/gen8_ppgtt.h"
+
+#include "g4x_dp.h"
+#include "g4x_hdmi.h"
+#include "hsw_ips.h"
+#include "i915_drv.h"
+#include "i915_utils.h"
+#include "icl_dsi.h"
+#include "intel_acpi.h"
+#include "intel_atomic.h"
+#include "intel_atomic_plane.h"
+#include "intel_bw.h"
+#include "intel_cdclk.h"
+#include "intel_color.h"
+#include "intel_crtc.h"
+#include "intel_crtc_state_dump.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dmc.h"
+#include "intel_dp_link_training.h"
+#include "intel_dpt.h"
+#include "intel_dsb.h"
+#include "intel_fbc.h"
+#include "intel_fbdev.h"
+#include "intel_fdi.h"
+#include "intel_fifo_underrun.h"
+#include "intel_frontbuffer.h"
+#include "intel_hdcp.h"
+#include "intel_hotplug.h"
+#include "intel_modeset_verify.h"
+#include "intel_modeset_setup.h"
+#include "intel_overlay.h"
+#include "intel_panel.h"
+#include "intel_pch_display.h"
+#include "intel_pch_refclk.h"
+#include "intel_pcode.h"
+#include "intel_pipe_crc.h"
+#include "intel_plane_initial.h"
+#include "intel_pm.h"
+#include "intel_pps.h"
+#include "intel_psr.h"
+#include "intel_quirks.h"
+#include "intel_sprite.h"
+#include "intel_tc.h"
+#include "intel_vga.h"
+#include "i9xx_plane.h"
+#include "skl_scaler.h"
+#include "skl_universal_plane.h"
+#include "skl_watermark.h"
+#include "vlv_dsi.h"
+#include "vlv_dsi_pll.h"
+#include "vlv_dsi_regs.h"
+#include "vlv_sideband.h"
+
+static void intel_set_transcoder_timings(const struct intel_crtc_state *crtc_state);
+static void intel_set_pipe_src_size(const struct intel_crtc_state *crtc_state);
+static void hsw_set_transconf(const struct intel_crtc_state *crtc_state);
+static void bdw_set_pipemisc(const struct intel_crtc_state *crtc_state);
+static void ilk_pfit_enable(const struct intel_crtc_state *crtc_state);
+
+/**
+ * intel_update_watermarks - update FIFO watermark values based on current modes
+ * @dev_priv: i915 device
+ *
+ * Calculate watermark values for the various WM regs based on current mode
+ * and plane configuration.
+ *
+ * There are several cases to deal with here:
+ *   - normal (i.e. non-self-refresh)
+ *   - self-refresh (SR) mode
+ *   - lines are large relative to FIFO size (buffer can hold up to 2)
+ *   - lines are small relative to FIFO size (buffer can hold more than 2
+ *     lines), so need to account for TLB latency
+ *
+ *   The normal calculation is:
+ *     watermark = dotclock * bytes per pixel * latency
+ *   where latency is platform & configuration dependent (we assume pessimal
+ *   values here).
+ *
+ *   The SR calculation is:
+ *     watermark = (trunc(latency/line time)+1) * surface width *
+ *       bytes per pixel
+ *   where
+ *     line time = htotal / dotclock
+ *     surface width = hdisplay for normal plane and 64 for cursor
+ *   and latency is assumed to be high, as above.
+ *
+ * The final value programmed to the register should always be rounded up,
+ * and include an extra 2 entries to account for clock crossings.
+ *
+ * We don't use the sprite, so we can ignore that.  And on Crestline we have
+ * to set the non-SR watermarks to 8.
+ */
+void intel_update_watermarks(struct drm_i915_private *dev_priv)
+{
+	if (dev_priv->display.funcs.wm->update_wm)
+		dev_priv->display.funcs.wm->update_wm(dev_priv);
+}
+
+static int intel_compute_pipe_wm(struct intel_atomic_state *state,
+				 struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	if (dev_priv->display.funcs.wm->compute_pipe_wm)
+		return dev_priv->display.funcs.wm->compute_pipe_wm(state, crtc);
+	return 0;
+}
+
+static int intel_compute_intermediate_wm(struct intel_atomic_state *state,
+					 struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	if (!dev_priv->display.funcs.wm->compute_intermediate_wm)
+		return 0;
+	if (drm_WARN_ON(&dev_priv->drm,
+			!dev_priv->display.funcs.wm->compute_pipe_wm))
+		return 0;
+	return dev_priv->display.funcs.wm->compute_intermediate_wm(state, crtc);
+}
+
+static bool intel_initial_watermarks(struct intel_atomic_state *state,
+				     struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	if (dev_priv->display.funcs.wm->initial_watermarks) {
+		dev_priv->display.funcs.wm->initial_watermarks(state, crtc);
+		return true;
+	}
+	return false;
+}
+
+static void intel_atomic_update_watermarks(struct intel_atomic_state *state,
+					   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	if (dev_priv->display.funcs.wm->atomic_update_watermarks)
+		dev_priv->display.funcs.wm->atomic_update_watermarks(state, crtc);
+}
+
+static void intel_optimize_watermarks(struct intel_atomic_state *state,
+				      struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	if (dev_priv->display.funcs.wm->optimize_watermarks)
+		dev_priv->display.funcs.wm->optimize_watermarks(state, crtc);
+}
+
+static int intel_compute_global_watermarks(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	if (dev_priv->display.funcs.wm->compute_global_watermarks)
+		return dev_priv->display.funcs.wm->compute_global_watermarks(state);
+	return 0;
+}
+
+/* returns HPLL frequency in kHz */
+int vlv_get_hpll_vco(struct drm_i915_private *dev_priv)
+{
+	int hpll_freq, vco_freq[] = { 800, 1600, 2000, 2400 };
+
+	/* Obtain SKU information */
+	hpll_freq = vlv_cck_read(dev_priv, CCK_FUSE_REG) &
+		CCK_FUSE_HPLL_FREQ_MASK;
+
+	return vco_freq[hpll_freq] * 1000;
+}
+
+int vlv_get_cck_clock(struct drm_i915_private *dev_priv,
+		      const char *name, u32 reg, int ref_freq)
+{
+	u32 val;
+	int divider;
+
+	val = vlv_cck_read(dev_priv, reg);
+	divider = val & CCK_FREQUENCY_VALUES;
+
+	drm_WARN(&dev_priv->drm, (val & CCK_FREQUENCY_STATUS) !=
+		 (divider << CCK_FREQUENCY_STATUS_SHIFT),
+		 "%s change in progress\n", name);
+
+	return DIV_ROUND_CLOSEST(ref_freq << 1, divider + 1);
+}
+
+int vlv_get_cck_clock_hpll(struct drm_i915_private *dev_priv,
+			   const char *name, u32 reg)
+{
+	int hpll;
+
+	vlv_cck_get(dev_priv);
+
+	if (dev_priv->hpll_freq == 0)
+		dev_priv->hpll_freq = vlv_get_hpll_vco(dev_priv);
+
+	hpll = vlv_get_cck_clock(dev_priv, name, reg, dev_priv->hpll_freq);
+
+	vlv_cck_put(dev_priv);
+
+	return hpll;
+}
+
+static void intel_update_czclk(struct drm_i915_private *dev_priv)
+{
+	if (!(IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)))
+		return;
+
+	dev_priv->czclk_freq = vlv_get_cck_clock_hpll(dev_priv, "czclk",
+						      CCK_CZ_CLOCK_CONTROL);
+
+	drm_dbg(&dev_priv->drm, "CZ clock rate: %d kHz\n",
+		dev_priv->czclk_freq);
+}
+
+static bool is_hdr_mode(const struct intel_crtc_state *crtc_state)
+{
+	return (crtc_state->active_planes &
+		~(icl_hdr_plane_mask() | BIT(PLANE_CURSOR))) == 0;
+}
+
+/* WA Display #0827: Gen9:all */
+static void
+skl_wa_827(struct drm_i915_private *dev_priv, enum pipe pipe, bool enable)
+{
+	if (enable)
+		intel_de_write(dev_priv, CLKGATE_DIS_PSL(pipe),
+		               intel_de_read(dev_priv, CLKGATE_DIS_PSL(pipe)) | DUPS1_GATING_DIS | DUPS2_GATING_DIS);
+	else
+		intel_de_write(dev_priv, CLKGATE_DIS_PSL(pipe),
+		               intel_de_read(dev_priv, CLKGATE_DIS_PSL(pipe)) & ~(DUPS1_GATING_DIS | DUPS2_GATING_DIS));
+}
+
+/* Wa_2006604312:icl,ehl */
+static void
+icl_wa_scalerclkgating(struct drm_i915_private *dev_priv, enum pipe pipe,
+		       bool enable)
+{
+	if (enable)
+		intel_de_write(dev_priv, CLKGATE_DIS_PSL(pipe),
+		               intel_de_read(dev_priv, CLKGATE_DIS_PSL(pipe)) | DPFR_GATING_DIS);
+	else
+		intel_de_write(dev_priv, CLKGATE_DIS_PSL(pipe),
+		               intel_de_read(dev_priv, CLKGATE_DIS_PSL(pipe)) & ~DPFR_GATING_DIS);
+}
+
+/* Wa_1604331009:icl,jsl,ehl */
+static void
+icl_wa_cursorclkgating(struct drm_i915_private *dev_priv, enum pipe pipe,
+		       bool enable)
+{
+	intel_de_rmw(dev_priv, CLKGATE_DIS_PSL(pipe), CURSOR_GATING_DIS,
+		     enable ? CURSOR_GATING_DIS : 0);
+}
+
+static bool
+is_trans_port_sync_slave(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->master_transcoder != INVALID_TRANSCODER;
+}
+
+static bool
+is_trans_port_sync_master(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->sync_mode_slaves_mask != 0;
+}
+
+bool
+is_trans_port_sync_mode(const struct intel_crtc_state *crtc_state)
+{
+	return is_trans_port_sync_master(crtc_state) ||
+		is_trans_port_sync_slave(crtc_state);
+}
+
+static enum pipe bigjoiner_master_pipe(const struct intel_crtc_state *crtc_state)
+{
+	return ffs(crtc_state->bigjoiner_pipes) - 1;
+}
+
+u8 intel_crtc_bigjoiner_slave_pipes(const struct intel_crtc_state *crtc_state)
+{
+	if (crtc_state->bigjoiner_pipes)
+		return crtc_state->bigjoiner_pipes & ~BIT(bigjoiner_master_pipe(crtc_state));
+	else
+		return 0;
+}
+
+bool intel_crtc_is_bigjoiner_slave(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	return crtc_state->bigjoiner_pipes &&
+		crtc->pipe != bigjoiner_master_pipe(crtc_state);
+}
+
+bool intel_crtc_is_bigjoiner_master(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	return crtc_state->bigjoiner_pipes &&
+		crtc->pipe == bigjoiner_master_pipe(crtc_state);
+}
+
+static int intel_bigjoiner_num_pipes(const struct intel_crtc_state *crtc_state)
+{
+	return hweight8(crtc_state->bigjoiner_pipes);
+}
+
+struct intel_crtc *intel_master_crtc(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+
+	if (intel_crtc_is_bigjoiner_slave(crtc_state))
+		return intel_crtc_for_pipe(i915, bigjoiner_master_pipe(crtc_state));
+	else
+		return to_intel_crtc(crtc_state->uapi.crtc);
+}
+
+static bool pipe_scanline_is_moving(struct drm_i915_private *dev_priv,
+				    enum pipe pipe)
+{
+	i915_reg_t reg = PIPEDSL(pipe);
+	u32 line1, line2;
+
+	line1 = intel_de_read(dev_priv, reg) & PIPEDSL_LINE_MASK;
+	msleep(5);
+	line2 = intel_de_read(dev_priv, reg) & PIPEDSL_LINE_MASK;
+
+	return line1 != line2;
+}
+
+static void wait_for_pipe_scanline_moving(struct intel_crtc *crtc, bool state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	/* Wait for the display line to settle/start moving */
+	if (wait_for(pipe_scanline_is_moving(dev_priv, pipe) == state, 100))
+		drm_err(&dev_priv->drm,
+			"pipe %c scanline %s wait timed out\n",
+			pipe_name(pipe), str_on_off(state));
+}
+
+static void intel_wait_for_pipe_scanline_stopped(struct intel_crtc *crtc)
+{
+	wait_for_pipe_scanline_moving(crtc, false);
+}
+
+static void intel_wait_for_pipe_scanline_moving(struct intel_crtc *crtc)
+{
+	wait_for_pipe_scanline_moving(crtc, true);
+}
+
+static void
+intel_wait_for_pipe_off(const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	if (DISPLAY_VER(dev_priv) >= 4) {
+		enum transcoder cpu_transcoder = old_crtc_state->cpu_transcoder;
+
+		/* Wait for the Pipe State to go off */
+		if (intel_de_wait_for_clear(dev_priv, PIPECONF(cpu_transcoder),
+					    PIPECONF_STATE_ENABLE, 100))
+			drm_WARN(&dev_priv->drm, 1, "pipe_off wait timed out\n");
+	} else {
+		intel_wait_for_pipe_scanline_stopped(crtc);
+	}
+}
+
+void assert_transcoder(struct drm_i915_private *dev_priv,
+		       enum transcoder cpu_transcoder, bool state)
+{
+	bool cur_state;
+	enum intel_display_power_domain power_domain;
+	intel_wakeref_t wakeref;
+
+	/* we keep both pipes enabled on 830 */
+	if (IS_I830(dev_priv))
+		state = true;
+
+	power_domain = POWER_DOMAIN_TRANSCODER(cpu_transcoder);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (wakeref) {
+		u32 val = intel_de_read(dev_priv, PIPECONF(cpu_transcoder));
+		cur_state = !!(val & PIPECONF_ENABLE);
+
+		intel_display_power_put(dev_priv, power_domain, wakeref);
+	} else {
+		cur_state = false;
+	}
+
+	I915_STATE_WARN(cur_state != state,
+			"transcoder %s assertion failure (expected %s, current %s)\n",
+			transcoder_name(cpu_transcoder),
+			str_on_off(state), str_on_off(cur_state));
+}
+
+static void assert_plane(struct intel_plane *plane, bool state)
+{
+	enum pipe pipe;
+	bool cur_state;
+
+	cur_state = plane->get_hw_state(plane, &pipe);
+
+	I915_STATE_WARN(cur_state != state,
+			"%s assertion failure (expected %s, current %s)\n",
+			plane->base.name, str_on_off(state),
+			str_on_off(cur_state));
+}
+
+#define assert_plane_enabled(p) assert_plane(p, true)
+#define assert_plane_disabled(p) assert_plane(p, false)
+
+static void assert_planes_disabled(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_plane *plane;
+
+	for_each_intel_plane_on_crtc(&dev_priv->drm, crtc, plane)
+		assert_plane_disabled(plane);
+}
+
+void vlv_wait_port_ready(struct drm_i915_private *dev_priv,
+			 struct intel_digital_port *dig_port,
+			 unsigned int expected_mask)
+{
+	u32 port_mask;
+	i915_reg_t dpll_reg;
+
+	switch (dig_port->base.port) {
+	default:
+		MISSING_CASE(dig_port->base.port);
+		fallthrough;
+	case PORT_B:
+		port_mask = DPLL_PORTB_READY_MASK;
+		dpll_reg = DPLL(0);
+		break;
+	case PORT_C:
+		port_mask = DPLL_PORTC_READY_MASK;
+		dpll_reg = DPLL(0);
+		expected_mask <<= 4;
+		break;
+	case PORT_D:
+		port_mask = DPLL_PORTD_READY_MASK;
+		dpll_reg = DPIO_PHY_STATUS;
+		break;
+	}
+
+	if (intel_de_wait_for_register(dev_priv, dpll_reg,
+				       port_mask, expected_mask, 1000))
+		drm_WARN(&dev_priv->drm, 1,
+			 "timed out waiting for [ENCODER:%d:%s] port ready: got 0x%x, expected 0x%x\n",
+			 dig_port->base.base.base.id, dig_port->base.base.name,
+			 intel_de_read(dev_priv, dpll_reg) & port_mask,
+			 expected_mask);
+}
+
+void intel_enable_transcoder(const struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = new_crtc_state->cpu_transcoder;
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 val;
+
+	drm_dbg_kms(&dev_priv->drm, "enabling pipe %c\n", pipe_name(pipe));
+
+	assert_planes_disabled(crtc);
+
+	/*
+	 * A pipe without a PLL won't actually be able to drive bits from
+	 * a plane.  On ILK+ the pipe PLLs are integrated, so we don't
+	 * need the check.
+	 */
+	if (HAS_GMCH(dev_priv)) {
+		if (intel_crtc_has_type(new_crtc_state, INTEL_OUTPUT_DSI))
+			assert_dsi_pll_enabled(dev_priv);
+		else
+			assert_pll_enabled(dev_priv, pipe);
+	} else {
+		if (new_crtc_state->has_pch_encoder) {
+			/* if driving the PCH, we need FDI enabled */
+			assert_fdi_rx_pll_enabled(dev_priv,
+						  intel_crtc_pch_transcoder(crtc));
+			assert_fdi_tx_pll_enabled(dev_priv,
+						  (enum pipe) cpu_transcoder);
+		}
+		/* FIXME: assert CPU port conditions for SNB+ */
+	}
+
+	/* Wa_22012358565:adl-p */
+	if (DISPLAY_VER(dev_priv) == 13)
+		intel_de_rmw(dev_priv, PIPE_ARB_CTL(pipe),
+			     0, PIPE_ARB_USE_PROG_SLOTS);
+
+	reg = PIPECONF(cpu_transcoder);
+	val = intel_de_read(dev_priv, reg);
+	if (val & PIPECONF_ENABLE) {
+		/* we keep both pipes enabled on 830 */
+		drm_WARN_ON(&dev_priv->drm, !IS_I830(dev_priv));
+		return;
+	}
+
+	intel_de_write(dev_priv, reg, val | PIPECONF_ENABLE);
+	intel_de_posting_read(dev_priv, reg);
+
+	/*
+	 * Until the pipe starts PIPEDSL reads will return a stale value,
+	 * which causes an apparent vblank timestamp jump when PIPEDSL
+	 * resets to its proper value. That also messes up the frame count
+	 * when it's derived from the timestamps. So let's wait for the
+	 * pipe to start properly before we call drm_crtc_vblank_on()
+	 */
+	if (intel_crtc_max_vblank_count(new_crtc_state) == 0)
+		intel_wait_for_pipe_scanline_moving(crtc);
+}
+
+void intel_disable_transcoder(const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = old_crtc_state->cpu_transcoder;
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 val;
+
+	drm_dbg_kms(&dev_priv->drm, "disabling pipe %c\n", pipe_name(pipe));
+
+	/*
+	 * Make sure planes won't keep trying to pump pixels to us,
+	 * or we might hang the display.
+	 */
+	assert_planes_disabled(crtc);
+
+	reg = PIPECONF(cpu_transcoder);
+	val = intel_de_read(dev_priv, reg);
+	if ((val & PIPECONF_ENABLE) == 0)
+		return;
+
+	/*
+	 * Double wide has implications for planes
+	 * so best keep it disabled when not needed.
+	 */
+	if (old_crtc_state->double_wide)
+		val &= ~PIPECONF_DOUBLE_WIDE;
+
+	/* Don't disable pipe or pipe PLLs if needed */
+	if (!IS_I830(dev_priv))
+		val &= ~PIPECONF_ENABLE;
+
+	if (DISPLAY_VER(dev_priv) >= 14)
+		intel_de_rmw(dev_priv, MTL_CHICKEN_TRANS(cpu_transcoder),
+			     FECSTALL_DIS_DPTSTREAM_DPTTG, 0);
+	else if (DISPLAY_VER(dev_priv) >= 12)
+		intel_de_rmw(dev_priv, CHICKEN_TRANS(cpu_transcoder),
+			     FECSTALL_DIS_DPTSTREAM_DPTTG, 0);
+
+	intel_de_write(dev_priv, reg, val);
+	if ((val & PIPECONF_ENABLE) == 0)
+		intel_wait_for_pipe_off(old_crtc_state);
+}
+
+unsigned int intel_rotation_info_size(const struct intel_rotation_info *rot_info)
+{
+	unsigned int size = 0;
+	int i;
+
+	for (i = 0 ; i < ARRAY_SIZE(rot_info->plane); i++)
+		size += rot_info->plane[i].dst_stride * rot_info->plane[i].width;
+
+	return size;
+}
+
+unsigned int intel_remapped_info_size(const struct intel_remapped_info *rem_info)
+{
+	unsigned int size = 0;
+	int i;
+
+	for (i = 0 ; i < ARRAY_SIZE(rem_info->plane); i++) {
+		unsigned int plane_size;
+
+		if (rem_info->plane[i].linear)
+			plane_size = rem_info->plane[i].size;
+		else
+			plane_size = rem_info->plane[i].dst_stride * rem_info->plane[i].height;
+
+		if (plane_size == 0)
+			continue;
+
+		if (rem_info->plane_alignment)
+			size = ALIGN(size, rem_info->plane_alignment);
+
+		size += plane_size;
+	}
+
+	return size;
+}
+
+bool intel_plane_uses_fence(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+
+	return DISPLAY_VER(dev_priv) < 4 ||
+		(plane->fbc &&
+		 plane_state->view.gtt.type == I915_GTT_VIEW_NORMAL);
+}
+
+/*
+ * Convert the x/y offsets into a linear offset.
+ * Only valid with 0/180 degree rotation, which is fine since linear
+ * offset is only used with linear buffers on pre-hsw and tiled buffers
+ * with gen2/3, and 90/270 degree rotations isn't supported on any of them.
+ */
+u32 intel_fb_xy_to_linear(int x, int y,
+			  const struct intel_plane_state *state,
+			  int color_plane)
+{
+	const struct drm_framebuffer *fb = state->hw.fb;
+	unsigned int cpp = fb->format->cpp[color_plane];
+	unsigned int pitch = state->view.color_plane[color_plane].mapping_stride;
+
+	return y * pitch + x * cpp;
+}
+
+/*
+ * Add the x/y offsets derived from fb->offsets[] to the user
+ * specified plane src x/y offsets. The resulting x/y offsets
+ * specify the start of scanout from the beginning of the gtt mapping.
+ */
+void intel_add_fb_offsets(int *x, int *y,
+			  const struct intel_plane_state *state,
+			  int color_plane)
+
+{
+	*x += state->view.color_plane[color_plane].x;
+	*y += state->view.color_plane[color_plane].y;
+}
+
+u32 intel_plane_fb_max_stride(struct drm_i915_private *dev_priv,
+			      u32 pixel_format, u64 modifier)
+{
+	struct intel_crtc *crtc;
+	struct intel_plane *plane;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return 0;
+
+	/*
+	 * We assume the primary plane for pipe A has
+	 * the highest stride limits of them all,
+	 * if in case pipe A is disabled, use the first pipe from pipe_mask.
+	 */
+	crtc = intel_first_crtc(dev_priv);
+	if (!crtc)
+		return 0;
+
+	plane = to_intel_plane(crtc->base.primary);
+
+	return plane->max_stride(plane, pixel_format, modifier,
+				 DRM_MODE_ROTATE_0);
+}
+
+void intel_set_plane_visible(struct intel_crtc_state *crtc_state,
+			     struct intel_plane_state *plane_state,
+			     bool visible)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+
+	plane_state->uapi.visible = visible;
+
+	if (visible)
+		crtc_state->uapi.plane_mask |= drm_plane_mask(&plane->base);
+	else
+		crtc_state->uapi.plane_mask &= ~drm_plane_mask(&plane->base);
+}
+
+void intel_plane_fixup_bitmasks(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	struct drm_plane *plane;
+
+	/*
+	 * Active_planes aliases if multiple "primary" or cursor planes
+	 * have been used on the same (or wrong) pipe. plane_mask uses
+	 * unique ids, hence we can use that to reconstruct active_planes.
+	 */
+	crtc_state->enabled_planes = 0;
+	crtc_state->active_planes = 0;
+
+	drm_for_each_plane_mask(plane, &dev_priv->drm,
+				crtc_state->uapi.plane_mask) {
+		crtc_state->enabled_planes |= BIT(to_intel_plane(plane)->id);
+		crtc_state->active_planes |= BIT(to_intel_plane(plane)->id);
+	}
+}
+
+void intel_plane_disable_noatomic(struct intel_crtc *crtc,
+				  struct intel_plane *plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state =
+		to_intel_crtc_state(crtc->base.state);
+	struct intel_plane_state *plane_state =
+		to_intel_plane_state(plane->base.state);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Disabling [PLANE:%d:%s] on [CRTC:%d:%s]\n",
+		    plane->base.base.id, plane->base.name,
+		    crtc->base.base.id, crtc->base.name);
+
+	intel_set_plane_visible(crtc_state, plane_state, false);
+	intel_plane_fixup_bitmasks(crtc_state);
+	crtc_state->data_rate[plane->id] = 0;
+	crtc_state->data_rate_y[plane->id] = 0;
+	crtc_state->rel_data_rate[plane->id] = 0;
+	crtc_state->rel_data_rate_y[plane->id] = 0;
+	crtc_state->min_cdclk[plane->id] = 0;
+
+	if ((crtc_state->active_planes & ~BIT(PLANE_CURSOR)) == 0 &&
+	    hsw_ips_disable(crtc_state)) {
+		crtc_state->ips_enabled = false;
+		intel_crtc_wait_for_next_vblank(crtc);
+	}
+
+	/*
+	 * Vblank time updates from the shadow to live plane control register
+	 * are blocked if the memory self-refresh mode is active at that
+	 * moment. So to make sure the plane gets truly disabled, disable
+	 * first the self-refresh mode. The self-refresh enable bit in turn
+	 * will be checked/applied by the HW only at the next frame start
+	 * event which is after the vblank start event, so we need to have a
+	 * wait-for-vblank between disabling the plane and the pipe.
+	 */
+	if (HAS_GMCH(dev_priv) &&
+	    intel_set_memory_cxsr(dev_priv, false))
+		intel_crtc_wait_for_next_vblank(crtc);
+
+	/*
+	 * Gen2 reports pipe underruns whenever all planes are disabled.
+	 * So disable underrun reporting before all the planes get disabled.
+	 */
+	if (DISPLAY_VER(dev_priv) == 2 && !crtc_state->active_planes)
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, crtc->pipe, false);
+
+	intel_plane_disable_arm(plane, crtc_state);
+	intel_crtc_wait_for_next_vblank(crtc);
+}
+
+unsigned int
+intel_plane_fence_y_offset(const struct intel_plane_state *plane_state)
+{
+	int x = 0, y = 0;
+
+	intel_plane_adjust_aligned_offset(&x, &y, plane_state, 0,
+					  plane_state->view.color_plane[0].offset, 0);
+
+	return y;
+}
+
+static int
+__intel_display_resume(struct drm_i915_private *i915,
+		       struct drm_atomic_state *state,
+		       struct drm_modeset_acquire_ctx *ctx)
+{
+	struct drm_crtc_state *crtc_state;
+	struct drm_crtc *crtc;
+	int i, ret;
+
+	intel_modeset_setup_hw_state(i915, ctx);
+	intel_vga_redisable(i915);
+
+	if (!state)
+		return 0;
+
+	/*
+	 * We've duplicated the state, pointers to the old state are invalid.
+	 *
+	 * Don't attempt to use the old state until we commit the duplicated state.
+	 */
+	for_each_new_crtc_in_state(state, crtc, crtc_state, i) {
+		/*
+		 * Force recalculation even if we restore
+		 * current state. With fast modeset this may not result
+		 * in a modeset when the state is compatible.
+		 */
+		crtc_state->mode_changed = true;
+	}
+
+	/* ignore any reset values/BIOS leftovers in the WM registers */
+	if (!HAS_GMCH(i915))
+		to_intel_atomic_state(state)->skip_intermediate_wm = true;
+
+	ret = drm_atomic_helper_commit_duplicated_state(state, ctx);
+
+	drm_WARN_ON(&i915->drm, ret == -EDEADLK);
+
+	return ret;
+}
+
+static bool gpu_reset_clobbers_display(struct drm_i915_private *dev_priv)
+{
+	return (INTEL_INFO(dev_priv)->gpu_reset_clobbers_display &&
+		intel_has_gpu_reset(to_gt(dev_priv)));
+}
+
+void intel_display_prepare_reset(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_modeset_acquire_ctx *ctx = &dev_priv->reset_ctx;
+	struct drm_atomic_state *state;
+	int ret;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	/* reset doesn't touch the display */
+	if (!dev_priv->params.force_reset_modeset_test &&
+	    !gpu_reset_clobbers_display(dev_priv))
+		return;
+
+	/* We have a modeset vs reset deadlock, defensively unbreak it. */
+	set_bit(I915_RESET_MODESET, &to_gt(dev_priv)->reset.flags);
+	smp_mb__after_atomic();
+	wake_up_bit(&to_gt(dev_priv)->reset.flags, I915_RESET_MODESET);
+
+	if (atomic_read(&dev_priv->gpu_error.pending_fb_pin)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Modeset potentially stuck, unbreaking through wedging\n");
+		intel_gt_set_wedged(to_gt(dev_priv));
+	}
+
+	/*
+	 * Need mode_config.mutex so that we don't
+	 * trample ongoing ->detect() and whatnot.
+	 */
+	mutex_lock(&dev->mode_config.mutex);
+	drm_modeset_acquire_init(ctx, 0);
+	while (1) {
+		ret = drm_modeset_lock_all_ctx(dev, ctx);
+		if (ret != -EDEADLK)
+			break;
+
+		drm_modeset_backoff(ctx);
+	}
+	/*
+	 * Disabling the crtcs gracefully seems nicer. Also the
+	 * g33 docs say we should at least disable all the planes.
+	 */
+	state = drm_atomic_helper_duplicate_state(dev, ctx);
+	if (IS_ERR(state)) {
+		ret = PTR_ERR(state);
+		drm_err(&dev_priv->drm, "Duplicating state failed with %i\n",
+			ret);
+		return;
+	}
+
+	ret = drm_atomic_helper_disable_all(dev, ctx);
+	if (ret) {
+		drm_err(&dev_priv->drm, "Suspending crtc's failed with %i\n",
+			ret);
+		drm_atomic_state_put(state);
+		return;
+	}
+
+	dev_priv->modeset_restore_state = state;
+	state->acquire_ctx = ctx;
+}
+
+void intel_display_finish_reset(struct drm_i915_private *i915)
+{
+	struct drm_modeset_acquire_ctx *ctx = &i915->reset_ctx;
+	struct drm_atomic_state *state;
+	int ret;
+
+	if (!HAS_DISPLAY(i915))
+		return;
+
+	/* reset doesn't touch the display */
+	if (!test_bit(I915_RESET_MODESET, &to_gt(i915)->reset.flags))
+		return;
+
+	state = fetch_and_zero(&i915->modeset_restore_state);
+	if (!state)
+		goto unlock;
+
+	/* reset doesn't touch the display */
+	if (!gpu_reset_clobbers_display(i915)) {
+		/* for testing only restore the display */
+		ret = __intel_display_resume(i915, state, ctx);
+		if (ret)
+			drm_err(&i915->drm,
+				"Restoring old state failed with %i\n", ret);
+	} else {
+		/*
+		 * The display has been reset as well,
+		 * so need a full re-initialization.
+		 */
+		intel_pps_unlock_regs_wa(i915);
+		intel_modeset_init_hw(i915);
+		intel_init_clock_gating(i915);
+		intel_hpd_init(i915);
+
+		ret = __intel_display_resume(i915, state, ctx);
+		if (ret)
+			drm_err(&i915->drm,
+				"Restoring old state failed with %i\n", ret);
+
+		intel_hpd_poll_disable(i915);
+	}
+
+	drm_atomic_state_put(state);
+unlock:
+	drm_modeset_drop_locks(ctx);
+	drm_modeset_acquire_fini(ctx);
+	mutex_unlock(&i915->drm.mode_config.mutex);
+
+	clear_bit_unlock(I915_RESET_MODESET, &to_gt(i915)->reset.flags);
+}
+
+static void icl_set_pipe_chicken(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, PIPE_CHICKEN(pipe));
+
+	/*
+	 * Display WA #1153: icl
+	 * enable hardware to bypass the alpha math
+	 * and rounding for per-pixel values 00 and 0xff
+	 */
+	tmp |= PER_PIXEL_ALPHA_BYPASS_EN;
+	/*
+	 * Display WA # 1605353570: icl
+	 * Set the pixel rounding bit to 1 for allowing
+	 * passthrough of Frame buffer pixels unmodified
+	 * across pipe
+	 */
+	tmp |= PIXEL_ROUNDING_TRUNC_FB_PASSTHRU;
+
+	/*
+	 * Underrun recovery must always be disabled on display 13+.
+	 * DG2 chicken bit meaning is inverted compared to other platforms.
+	 */
+	if (IS_DG2(dev_priv))
+		tmp &= ~UNDERRUN_RECOVERY_ENABLE_DG2;
+	else if (DISPLAY_VER(dev_priv) >= 13)
+		tmp |= UNDERRUN_RECOVERY_DISABLE_ADLP;
+
+	/* Wa_14010547955:dg2 */
+	if (IS_DG2_DISPLAY_STEP(dev_priv, STEP_B0, STEP_FOREVER))
+		tmp |= DG2_RENDER_CCSTAG_4_3_EN;
+
+	intel_de_write(dev_priv, PIPE_CHICKEN(pipe), tmp);
+}
+
+bool intel_has_pending_fb_unpin(struct drm_i915_private *dev_priv)
+{
+	struct drm_crtc *crtc;
+	bool cleanup_done;
+
+	drm_for_each_crtc(crtc, &dev_priv->drm) {
+		struct drm_crtc_commit *commit;
+		spin_lock(&crtc->commit_lock);
+		commit = list_first_entry_or_null(&crtc->commit_list,
+						  struct drm_crtc_commit, commit_entry);
+		cleanup_done = commit ?
+			try_wait_for_completion(&commit->cleanup_done) : true;
+		spin_unlock(&crtc->commit_lock);
+
+		if (cleanup_done)
+			continue;
+
+		intel_crtc_wait_for_next_vblank(to_intel_crtc(crtc));
+
+		return true;
+	}
+
+	return false;
+}
+
+/*
+ * Finds the encoder associated with the given CRTC. This can only be
+ * used when we know that the CRTC isn't feeding multiple encoders!
+ */
+struct intel_encoder *
+intel_get_crtc_new_encoder(const struct intel_atomic_state *state,
+			   const struct intel_crtc_state *crtc_state)
+{
+	const struct drm_connector_state *connector_state;
+	const struct drm_connector *connector;
+	struct intel_encoder *encoder = NULL;
+	struct intel_crtc *master_crtc;
+	int num_encoders = 0;
+	int i;
+
+	master_crtc = intel_master_crtc(crtc_state);
+
+	for_each_new_connector_in_state(&state->base, connector, connector_state, i) {
+		if (connector_state->crtc != &master_crtc->base)
+			continue;
+
+		encoder = to_intel_encoder(connector_state->best_encoder);
+		num_encoders++;
+	}
+
+	drm_WARN(state->base.dev, num_encoders != 1,
+		 "%d encoders for pipe %c\n",
+		 num_encoders, pipe_name(master_crtc->pipe));
+
+	return encoder;
+}
+
+static void cpt_verify_modeset(struct drm_i915_private *dev_priv,
+			       enum pipe pipe)
+{
+	i915_reg_t dslreg = PIPEDSL(pipe);
+	u32 temp;
+
+	temp = intel_de_read(dev_priv, dslreg);
+	udelay(500);
+	if (wait_for(intel_de_read(dev_priv, dslreg) != temp, 5)) {
+		if (wait_for(intel_de_read(dev_priv, dslreg) != temp, 5))
+			drm_err(&dev_priv->drm,
+				"mode set failed: pipe %c stuck\n",
+				pipe_name(pipe));
+	}
+}
+
+static void ilk_pfit_enable(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_rect *dst = &crtc_state->pch_pfit.dst;
+	enum pipe pipe = crtc->pipe;
+	int width = drm_rect_width(dst);
+	int height = drm_rect_height(dst);
+	int x = dst->x1;
+	int y = dst->y1;
+
+	if (!crtc_state->pch_pfit.enabled)
+		return;
+
+	/* Force use of hard-coded filter coefficients
+	 * as some pre-programmed values are broken,
+	 * e.g. x201.
+	 */
+	if (IS_IVYBRIDGE(dev_priv) || IS_HASWELL(dev_priv))
+		intel_de_write_fw(dev_priv, PF_CTL(pipe), PF_ENABLE |
+				  PF_FILTER_MED_3x3 | PF_PIPE_SEL_IVB(pipe));
+	else
+		intel_de_write_fw(dev_priv, PF_CTL(pipe), PF_ENABLE |
+				  PF_FILTER_MED_3x3);
+	intel_de_write_fw(dev_priv, PF_WIN_POS(pipe), x << 16 | y);
+	intel_de_write_fw(dev_priv, PF_WIN_SZ(pipe), width << 16 | height);
+}
+
+static void intel_crtc_dpms_overlay_disable(struct intel_crtc *crtc)
+{
+	if (crtc->overlay)
+		(void) intel_overlay_switch_off(crtc->overlay);
+
+	/* Let userspace switch the overlay on again. In most cases userspace
+	 * has to recompute where to put it anyway.
+	 */
+}
+
+static bool needs_nv12_wa(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	if (!crtc_state->nv12_planes)
+		return false;
+
+	/* WA Display #0827: Gen9:all */
+	if (DISPLAY_VER(dev_priv) == 9)
+		return true;
+
+	return false;
+}
+
+static bool needs_scalerclk_wa(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	/* Wa_2006604312:icl,ehl */
+	if (crtc_state->scaler_state.scaler_users > 0 && DISPLAY_VER(dev_priv) == 11)
+		return true;
+
+	return false;
+}
+
+static bool needs_cursorclk_wa(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	/* Wa_1604331009:icl,jsl,ehl */
+	if (is_hdr_mode(crtc_state) &&
+	    crtc_state->active_planes & BIT(PLANE_CURSOR) &&
+	    DISPLAY_VER(dev_priv) == 11)
+		return true;
+
+	return false;
+}
+
+static void intel_async_flip_vtd_wa(struct drm_i915_private *i915,
+				    enum pipe pipe, bool enable)
+{
+	if (DISPLAY_VER(i915) == 9) {
+		/*
+		 * "Plane N strech max must be programmed to 11b (x1)
+		 *  when Async flips are enabled on that plane."
+		 */
+		intel_de_rmw(i915, CHICKEN_PIPESL_1(pipe),
+			     SKL_PLANE1_STRETCH_MAX_MASK,
+			     enable ? SKL_PLANE1_STRETCH_MAX_X1 : SKL_PLANE1_STRETCH_MAX_X8);
+	} else {
+		/* Also needed on HSW/BDW albeit undocumented */
+		intel_de_rmw(i915, CHICKEN_PIPESL_1(pipe),
+			     HSW_PRI_STRETCH_MAX_MASK,
+			     enable ? HSW_PRI_STRETCH_MAX_X1 : HSW_PRI_STRETCH_MAX_X8);
+	}
+}
+
+static bool needs_async_flip_vtd_wa(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+
+	return crtc_state->uapi.async_flip && i915_vtd_active(i915) &&
+		(DISPLAY_VER(i915) == 9 || IS_BROADWELL(i915) || IS_HASWELL(i915));
+}
+
+static bool planes_enabling(const struct intel_crtc_state *old_crtc_state,
+			    const struct intel_crtc_state *new_crtc_state)
+{
+	return (!old_crtc_state->active_planes || intel_crtc_needs_modeset(new_crtc_state)) &&
+		new_crtc_state->active_planes;
+}
+
+static bool planes_disabling(const struct intel_crtc_state *old_crtc_state,
+			     const struct intel_crtc_state *new_crtc_state)
+{
+	return old_crtc_state->active_planes &&
+		(!new_crtc_state->active_planes || intel_crtc_needs_modeset(new_crtc_state));
+}
+
+static void intel_post_plane_update(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	enum pipe pipe = crtc->pipe;
+
+	intel_frontbuffer_flip(dev_priv, new_crtc_state->fb_bits);
+
+	if (new_crtc_state->update_wm_post && new_crtc_state->hw.active)
+		intel_update_watermarks(dev_priv);
+
+	hsw_ips_post_update(state, crtc);
+	intel_fbc_post_update(state, crtc);
+
+	if (needs_async_flip_vtd_wa(old_crtc_state) &&
+	    !needs_async_flip_vtd_wa(new_crtc_state))
+		intel_async_flip_vtd_wa(dev_priv, pipe, false);
+
+	if (needs_nv12_wa(old_crtc_state) &&
+	    !needs_nv12_wa(new_crtc_state))
+		skl_wa_827(dev_priv, pipe, false);
+
+	if (needs_scalerclk_wa(old_crtc_state) &&
+	    !needs_scalerclk_wa(new_crtc_state))
+		icl_wa_scalerclkgating(dev_priv, pipe, false);
+
+	if (needs_cursorclk_wa(old_crtc_state) &&
+	    !needs_cursorclk_wa(new_crtc_state))
+		icl_wa_cursorclkgating(dev_priv, pipe, false);
+
+	intel_drrs_activate(new_crtc_state);
+}
+
+static void intel_crtc_enable_flip_done(struct intel_atomic_state *state,
+					struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	u8 update_planes = crtc_state->update_planes;
+	const struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		if (plane->pipe == crtc->pipe &&
+		    update_planes & BIT(plane->id))
+			plane->enable_flip_done(plane);
+	}
+}
+
+static void intel_crtc_disable_flip_done(struct intel_atomic_state *state,
+					 struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	u8 update_planes = crtc_state->update_planes;
+	const struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		if (plane->pipe == crtc->pipe &&
+		    update_planes & BIT(plane->id))
+			plane->disable_flip_done(plane);
+	}
+}
+
+static void intel_crtc_async_flip_disable_wa(struct intel_atomic_state *state,
+					     struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	u8 update_planes = new_crtc_state->update_planes;
+	const struct intel_plane_state *old_plane_state;
+	struct intel_plane *plane;
+	bool need_vbl_wait = false;
+	int i;
+
+	for_each_old_intel_plane_in_state(state, plane, old_plane_state, i) {
+		if (plane->need_async_flip_disable_wa &&
+		    plane->pipe == crtc->pipe &&
+		    update_planes & BIT(plane->id)) {
+			/*
+			 * Apart from the async flip bit we want to
+			 * preserve the old state for the plane.
+			 */
+			plane->async_flip(plane, old_crtc_state,
+					  old_plane_state, false);
+			need_vbl_wait = true;
+		}
+	}
+
+	if (need_vbl_wait)
+		intel_crtc_wait_for_next_vblank(crtc);
+}
+
+static void intel_pre_plane_update(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	enum pipe pipe = crtc->pipe;
+
+	intel_drrs_deactivate(old_crtc_state);
+
+	intel_psr_pre_plane_update(state, crtc);
+
+	if (hsw_ips_pre_update(state, crtc))
+		intel_crtc_wait_for_next_vblank(crtc);
+
+	if (intel_fbc_pre_update(state, crtc))
+		intel_crtc_wait_for_next_vblank(crtc);
+
+	if (!needs_async_flip_vtd_wa(old_crtc_state) &&
+	    needs_async_flip_vtd_wa(new_crtc_state))
+		intel_async_flip_vtd_wa(dev_priv, pipe, true);
+
+	/* Display WA 827 */
+	if (!needs_nv12_wa(old_crtc_state) &&
+	    needs_nv12_wa(new_crtc_state))
+		skl_wa_827(dev_priv, pipe, true);
+
+	/* Wa_2006604312:icl,ehl */
+	if (!needs_scalerclk_wa(old_crtc_state) &&
+	    needs_scalerclk_wa(new_crtc_state))
+		icl_wa_scalerclkgating(dev_priv, pipe, true);
+
+	/* Wa_1604331009:icl,jsl,ehl */
+	if (!needs_cursorclk_wa(old_crtc_state) &&
+	    needs_cursorclk_wa(new_crtc_state))
+		icl_wa_cursorclkgating(dev_priv, pipe, true);
+
+	/*
+	 * Vblank time updates from the shadow to live plane control register
+	 * are blocked if the memory self-refresh mode is active at that
+	 * moment. So to make sure the plane gets truly disabled, disable
+	 * first the self-refresh mode. The self-refresh enable bit in turn
+	 * will be checked/applied by the HW only at the next frame start
+	 * event which is after the vblank start event, so we need to have a
+	 * wait-for-vblank between disabling the plane and the pipe.
+	 */
+	if (HAS_GMCH(dev_priv) && old_crtc_state->hw.active &&
+	    new_crtc_state->disable_cxsr && intel_set_memory_cxsr(dev_priv, false))
+		intel_crtc_wait_for_next_vblank(crtc);
+
+	/*
+	 * IVB workaround: must disable low power watermarks for at least
+	 * one frame before enabling scaling.  LP watermarks can be re-enabled
+	 * when scaling is disabled.
+	 *
+	 * WaCxSRDisabledForSpriteScaling:ivb
+	 */
+	if (old_crtc_state->hw.active &&
+	    new_crtc_state->disable_lp_wm && ilk_disable_lp_wm(dev_priv))
+		intel_crtc_wait_for_next_vblank(crtc);
+
+	/*
+	 * If we're doing a modeset we don't need to do any
+	 * pre-vblank watermark programming here.
+	 */
+	if (!intel_crtc_needs_modeset(new_crtc_state)) {
+		/*
+		 * For platforms that support atomic watermarks, program the
+		 * 'intermediate' watermarks immediately.  On pre-gen9 platforms, these
+		 * will be the intermediate values that are safe for both pre- and
+		 * post- vblank; when vblank happens, the 'active' values will be set
+		 * to the final 'target' values and we'll do this again to get the
+		 * optimal watermarks.  For gen9+ platforms, the values we program here
+		 * will be the final target values which will get automatically latched
+		 * at vblank time; no further programming will be necessary.
+		 *
+		 * If a platform hasn't been transitioned to atomic watermarks yet,
+		 * we'll continue to update watermarks the old way, if flags tell
+		 * us to.
+		 */
+		if (!intel_initial_watermarks(state, crtc))
+			if (new_crtc_state->update_wm_pre)
+				intel_update_watermarks(dev_priv);
+	}
+
+	/*
+	 * Gen2 reports pipe underruns whenever all planes are disabled.
+	 * So disable underrun reporting before all the planes get disabled.
+	 *
+	 * We do this after .initial_watermarks() so that we have a
+	 * chance of catching underruns with the intermediate watermarks
+	 * vs. the old plane configuration.
+	 */
+	if (DISPLAY_VER(dev_priv) == 2 && planes_disabling(old_crtc_state, new_crtc_state))
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false);
+
+	/*
+	 * WA for platforms where async address update enable bit
+	 * is double buffered and only latched at start of vblank.
+	 */
+	if (old_crtc_state->uapi.async_flip && !new_crtc_state->uapi.async_flip)
+		intel_crtc_async_flip_disable_wa(state, crtc);
+}
+
+static void intel_crtc_disable_planes(struct intel_atomic_state *state,
+				      struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	unsigned int update_mask = new_crtc_state->update_planes;
+	const struct intel_plane_state *old_plane_state;
+	struct intel_plane *plane;
+	unsigned fb_bits = 0;
+	int i;
+
+	intel_crtc_dpms_overlay_disable(crtc);
+
+	for_each_old_intel_plane_in_state(state, plane, old_plane_state, i) {
+		if (crtc->pipe != plane->pipe ||
+		    !(update_mask & BIT(plane->id)))
+			continue;
+
+		intel_plane_disable_arm(plane, new_crtc_state);
+
+		if (old_plane_state->uapi.visible)
+			fb_bits |= plane->frontbuffer_bit;
+	}
+
+	intel_frontbuffer_flip(dev_priv, fb_bits);
+}
+
+/*
+ * intel_connector_primary_encoder - get the primary encoder for a connector
+ * @connector: connector for which to return the encoder
+ *
+ * Returns the primary encoder for a connector. There is a 1:1 mapping from
+ * all connectors to their encoder, except for DP-MST connectors which have
+ * both a virtual and a primary encoder. These DP-MST primary encoders can be
+ * pointed to by as many DP-MST connectors as there are pipes.
+ */
+static struct intel_encoder *
+intel_connector_primary_encoder(struct intel_connector *connector)
+{
+	struct intel_encoder *encoder;
+
+	if (connector->mst_port)
+		return &dp_to_dig_port(connector->mst_port)->base;
+
+	encoder = intel_attached_encoder(connector);
+	drm_WARN_ON(connector->base.dev, !encoder);
+
+	return encoder;
+}
+
+static void intel_encoders_update_prepare(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_crtc_state *new_crtc_state, *old_crtc_state;
+	struct intel_crtc *crtc;
+	struct drm_connector_state *new_conn_state;
+	struct drm_connector *connector;
+	int i;
+
+	/*
+	 * Make sure the DPLL state is up-to-date for fastset TypeC ports after non-blocking commits.
+	 * TODO: Update the DPLL state for all cases in the encoder->update_prepare() hook.
+	 */
+	if (i915->display.dpll.mgr) {
+		for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
+			if (intel_crtc_needs_modeset(new_crtc_state))
+				continue;
+
+			new_crtc_state->shared_dpll = old_crtc_state->shared_dpll;
+			new_crtc_state->dpll_hw_state = old_crtc_state->dpll_hw_state;
+		}
+	}
+
+	if (!state->modeset)
+		return;
+
+	for_each_new_connector_in_state(&state->base, connector, new_conn_state,
+					i) {
+		struct intel_connector *intel_connector;
+		struct intel_encoder *encoder;
+		struct intel_crtc *crtc;
+
+		if (!intel_connector_needs_modeset(state, connector))
+			continue;
+
+		intel_connector = to_intel_connector(connector);
+		encoder = intel_connector_primary_encoder(intel_connector);
+		if (!encoder->update_prepare)
+			continue;
+
+		crtc = new_conn_state->crtc ?
+			to_intel_crtc(new_conn_state->crtc) : NULL;
+		encoder->update_prepare(state, encoder, crtc);
+	}
+}
+
+static void intel_encoders_update_complete(struct intel_atomic_state *state)
+{
+	struct drm_connector_state *new_conn_state;
+	struct drm_connector *connector;
+	int i;
+
+	if (!state->modeset)
+		return;
+
+	for_each_new_connector_in_state(&state->base, connector, new_conn_state,
+					i) {
+		struct intel_connector *intel_connector;
+		struct intel_encoder *encoder;
+		struct intel_crtc *crtc;
+
+		if (!intel_connector_needs_modeset(state, connector))
+			continue;
+
+		intel_connector = to_intel_connector(connector);
+		encoder = intel_connector_primary_encoder(intel_connector);
+		if (!encoder->update_complete)
+			continue;
+
+		crtc = new_conn_state->crtc ?
+			to_intel_crtc(new_conn_state->crtc) : NULL;
+		encoder->update_complete(state, encoder, crtc);
+	}
+}
+
+static void intel_encoders_pre_pll_enable(struct intel_atomic_state *state,
+					  struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct drm_connector_state *conn_state;
+	struct drm_connector *conn;
+	int i;
+
+	for_each_new_connector_in_state(&state->base, conn, conn_state, i) {
+		struct intel_encoder *encoder =
+			to_intel_encoder(conn_state->best_encoder);
+
+		if (conn_state->crtc != &crtc->base)
+			continue;
+
+		if (encoder->pre_pll_enable)
+			encoder->pre_pll_enable(state, encoder,
+						crtc_state, conn_state);
+	}
+}
+
+static void intel_encoders_pre_enable(struct intel_atomic_state *state,
+				      struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct drm_connector_state *conn_state;
+	struct drm_connector *conn;
+	int i;
+
+	for_each_new_connector_in_state(&state->base, conn, conn_state, i) {
+		struct intel_encoder *encoder =
+			to_intel_encoder(conn_state->best_encoder);
+
+		if (conn_state->crtc != &crtc->base)
+			continue;
+
+		if (encoder->pre_enable)
+			encoder->pre_enable(state, encoder,
+					    crtc_state, conn_state);
+	}
+}
+
+static void intel_encoders_enable(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct drm_connector_state *conn_state;
+	struct drm_connector *conn;
+	int i;
+
+	for_each_new_connector_in_state(&state->base, conn, conn_state, i) {
+		struct intel_encoder *encoder =
+			to_intel_encoder(conn_state->best_encoder);
+
+		if (conn_state->crtc != &crtc->base)
+			continue;
+
+		if (encoder->enable)
+			encoder->enable(state, encoder,
+					crtc_state, conn_state);
+		intel_opregion_notify_encoder(encoder, true);
+	}
+}
+
+static void intel_encoders_disable(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct drm_connector_state *old_conn_state;
+	struct drm_connector *conn;
+	int i;
+
+	for_each_old_connector_in_state(&state->base, conn, old_conn_state, i) {
+		struct intel_encoder *encoder =
+			to_intel_encoder(old_conn_state->best_encoder);
+
+		if (old_conn_state->crtc != &crtc->base)
+			continue;
+
+		intel_opregion_notify_encoder(encoder, false);
+		if (encoder->disable)
+			encoder->disable(state, encoder,
+					 old_crtc_state, old_conn_state);
+	}
+}
+
+static void intel_encoders_post_disable(struct intel_atomic_state *state,
+					struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct drm_connector_state *old_conn_state;
+	struct drm_connector *conn;
+	int i;
+
+	for_each_old_connector_in_state(&state->base, conn, old_conn_state, i) {
+		struct intel_encoder *encoder =
+			to_intel_encoder(old_conn_state->best_encoder);
+
+		if (old_conn_state->crtc != &crtc->base)
+			continue;
+
+		if (encoder->post_disable)
+			encoder->post_disable(state, encoder,
+					      old_crtc_state, old_conn_state);
+	}
+}
+
+static void intel_encoders_post_pll_disable(struct intel_atomic_state *state,
+					    struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct drm_connector_state *old_conn_state;
+	struct drm_connector *conn;
+	int i;
+
+	for_each_old_connector_in_state(&state->base, conn, old_conn_state, i) {
+		struct intel_encoder *encoder =
+			to_intel_encoder(old_conn_state->best_encoder);
+
+		if (old_conn_state->crtc != &crtc->base)
+			continue;
+
+		if (encoder->post_pll_disable)
+			encoder->post_pll_disable(state, encoder,
+						  old_crtc_state, old_conn_state);
+	}
+}
+
+static void intel_encoders_update_pipe(struct intel_atomic_state *state,
+				       struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct drm_connector_state *conn_state;
+	struct drm_connector *conn;
+	int i;
+
+	for_each_new_connector_in_state(&state->base, conn, conn_state, i) {
+		struct intel_encoder *encoder =
+			to_intel_encoder(conn_state->best_encoder);
+
+		if (conn_state->crtc != &crtc->base)
+			continue;
+
+		if (encoder->update_pipe)
+			encoder->update_pipe(state, encoder,
+					     crtc_state, conn_state);
+	}
+}
+
+static void intel_disable_primary_plane(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct intel_plane *plane = to_intel_plane(crtc->base.primary);
+
+	plane->disable_arm(plane, crtc_state);
+}
+
+static void ilk_configure_cpu_transcoder(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	if (crtc_state->has_pch_encoder) {
+		intel_cpu_transcoder_set_m1_n1(crtc, cpu_transcoder,
+					       &crtc_state->fdi_m_n);
+	} else if (intel_crtc_has_dp_encoder(crtc_state)) {
+		intel_cpu_transcoder_set_m1_n1(crtc, cpu_transcoder,
+					       &crtc_state->dp_m_n);
+		intel_cpu_transcoder_set_m2_n2(crtc, cpu_transcoder,
+					       &crtc_state->dp_m2_n2);
+	}
+
+	intel_set_transcoder_timings(crtc_state);
+
+	ilk_set_pipeconf(crtc_state);
+}
+
+static void ilk_crtc_enable(struct intel_atomic_state *state,
+			    struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	if (drm_WARN_ON(&dev_priv->drm, crtc->active))
+		return;
+
+	/*
+	 * Sometimes spurious CPU pipe underruns happen during FDI
+	 * training, at least with VGA+HDMI cloning. Suppress them.
+	 *
+	 * On ILK we get an occasional spurious CPU pipe underruns
+	 * between eDP port A enable and vdd enable. Also PCH port
+	 * enable seems to result in the occasional CPU pipe underrun.
+	 *
+	 * Spurious PCH underruns also occur during PCH enabling.
+	 */
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false);
+	intel_set_pch_fifo_underrun_reporting(dev_priv, pipe, false);
+
+	ilk_configure_cpu_transcoder(new_crtc_state);
+
+	intel_set_pipe_src_size(new_crtc_state);
+
+	crtc->active = true;
+
+	intel_encoders_pre_enable(state, crtc);
+
+	if (new_crtc_state->has_pch_encoder) {
+		ilk_pch_pre_enable(state, crtc);
+	} else {
+		assert_fdi_tx_disabled(dev_priv, pipe);
+		assert_fdi_rx_disabled(dev_priv, pipe);
+	}
+
+	ilk_pfit_enable(new_crtc_state);
+
+	/*
+	 * On ILK+ LUT must be loaded before the pipe is running but with
+	 * clocks enabled
+	 */
+	intel_color_load_luts(new_crtc_state);
+	intel_color_commit_noarm(new_crtc_state);
+	intel_color_commit_arm(new_crtc_state);
+	/* update DSPCNTR to configure gamma for pipe bottom color */
+	intel_disable_primary_plane(new_crtc_state);
+
+	intel_initial_watermarks(state, crtc);
+	intel_enable_transcoder(new_crtc_state);
+
+	if (new_crtc_state->has_pch_encoder)
+		ilk_pch_enable(state, crtc);
+
+	intel_crtc_vblank_on(new_crtc_state);
+
+	intel_encoders_enable(state, crtc);
+
+	if (HAS_PCH_CPT(dev_priv))
+		cpt_verify_modeset(dev_priv, pipe);
+
+	/*
+	 * Must wait for vblank to avoid spurious PCH FIFO underruns.
+	 * And a second vblank wait is needed at least on ILK with
+	 * some interlaced HDMI modes. Let's do the double wait always
+	 * in case there are more corner cases we don't know about.
+	 */
+	if (new_crtc_state->has_pch_encoder) {
+		intel_crtc_wait_for_next_vblank(crtc);
+		intel_crtc_wait_for_next_vblank(crtc);
+	}
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
+	intel_set_pch_fifo_underrun_reporting(dev_priv, pipe, true);
+}
+
+static void glk_pipe_scaler_clock_gating_wa(struct drm_i915_private *dev_priv,
+					    enum pipe pipe, bool apply)
+{
+	u32 val = intel_de_read(dev_priv, CLKGATE_DIS_PSL(pipe));
+	u32 mask = DPF_GATING_DIS | DPF_RAM_GATING_DIS | DPFR_GATING_DIS;
+
+	if (apply)
+		val |= mask;
+	else
+		val &= ~mask;
+
+	intel_de_write(dev_priv, CLKGATE_DIS_PSL(pipe), val);
+}
+
+static void hsw_set_linetime_wm(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	intel_de_write(dev_priv, WM_LINETIME(crtc->pipe),
+		       HSW_LINETIME(crtc_state->linetime) |
+		       HSW_IPS_LINETIME(crtc_state->ips_linetime));
+}
+
+static void hsw_set_frame_start_delay(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder transcoder = crtc_state->cpu_transcoder;
+	i915_reg_t reg = DISPLAY_VER(dev_priv) >= 14 ? MTL_CHICKEN_TRANS(transcoder) :
+			 CHICKEN_TRANS(transcoder);
+	u32 val;
+
+	val = intel_de_read(dev_priv, reg);
+	val &= ~HSW_FRAME_START_DELAY_MASK;
+	val |= HSW_FRAME_START_DELAY(crtc_state->framestart_delay - 1);
+	intel_de_write(dev_priv, reg, val);
+}
+
+static void icl_ddi_bigjoiner_pre_enable(struct intel_atomic_state *state,
+					 const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *master_crtc = intel_master_crtc(crtc_state);
+
+	/*
+	 * Enable sequence steps 1-7 on bigjoiner master
+	 */
+	if (intel_crtc_is_bigjoiner_slave(crtc_state))
+		intel_encoders_pre_pll_enable(state, master_crtc);
+
+	if (crtc_state->shared_dpll)
+		intel_enable_shared_dpll(crtc_state);
+
+	if (intel_crtc_is_bigjoiner_slave(crtc_state))
+		intel_encoders_pre_enable(state, master_crtc);
+}
+
+static void hsw_configure_cpu_transcoder(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	if (crtc_state->has_pch_encoder) {
+		intel_cpu_transcoder_set_m1_n1(crtc, cpu_transcoder,
+					       &crtc_state->fdi_m_n);
+	} else if (intel_crtc_has_dp_encoder(crtc_state)) {
+		intel_cpu_transcoder_set_m1_n1(crtc, cpu_transcoder,
+					       &crtc_state->dp_m_n);
+		intel_cpu_transcoder_set_m2_n2(crtc, cpu_transcoder,
+					       &crtc_state->dp_m2_n2);
+	}
+
+	intel_set_transcoder_timings(crtc_state);
+
+	if (cpu_transcoder != TRANSCODER_EDP)
+		intel_de_write(dev_priv, PIPE_MULT(cpu_transcoder),
+			       crtc_state->pixel_multiplier - 1);
+
+	hsw_set_frame_start_delay(crtc_state);
+
+	hsw_set_transconf(crtc_state);
+}
+
+static void hsw_crtc_enable(struct intel_atomic_state *state,
+			    struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe, hsw_workaround_pipe;
+	enum transcoder cpu_transcoder = new_crtc_state->cpu_transcoder;
+	bool psl_clkgate_wa;
+
+	if (drm_WARN_ON(&dev_priv->drm, crtc->active))
+		return;
+
+	if (!new_crtc_state->bigjoiner_pipes) {
+		intel_encoders_pre_pll_enable(state, crtc);
+
+		if (new_crtc_state->shared_dpll)
+			intel_enable_shared_dpll(new_crtc_state);
+
+		intel_encoders_pre_enable(state, crtc);
+	} else {
+		icl_ddi_bigjoiner_pre_enable(state, new_crtc_state);
+	}
+
+	intel_dsc_enable(new_crtc_state);
+
+	if (DISPLAY_VER(dev_priv) >= 13)
+		intel_uncompressed_joiner_enable(new_crtc_state);
+
+	intel_set_pipe_src_size(new_crtc_state);
+	if (DISPLAY_VER(dev_priv) >= 9 || IS_BROADWELL(dev_priv))
+		bdw_set_pipemisc(new_crtc_state);
+
+	if (!intel_crtc_is_bigjoiner_slave(new_crtc_state) &&
+	    !transcoder_is_dsi(cpu_transcoder))
+		hsw_configure_cpu_transcoder(new_crtc_state);
+
+	crtc->active = true;
+
+	/* Display WA #1180: WaDisableScalarClockGating: glk */
+	psl_clkgate_wa = DISPLAY_VER(dev_priv) == 10 &&
+		new_crtc_state->pch_pfit.enabled;
+	if (psl_clkgate_wa)
+		glk_pipe_scaler_clock_gating_wa(dev_priv, pipe, true);
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		skl_pfit_enable(new_crtc_state);
+	else
+		ilk_pfit_enable(new_crtc_state);
+
+	/*
+	 * On ILK+ LUT must be loaded before the pipe is running but with
+	 * clocks enabled
+	 */
+	intel_color_load_luts(new_crtc_state);
+	intel_color_commit_noarm(new_crtc_state);
+	intel_color_commit_arm(new_crtc_state);
+	/* update DSPCNTR to configure gamma/csc for pipe bottom color */
+	if (DISPLAY_VER(dev_priv) < 9)
+		intel_disable_primary_plane(new_crtc_state);
+
+	hsw_set_linetime_wm(new_crtc_state);
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		icl_set_pipe_chicken(new_crtc_state);
+
+	intel_initial_watermarks(state, crtc);
+
+	if (intel_crtc_is_bigjoiner_slave(new_crtc_state))
+		intel_crtc_vblank_on(new_crtc_state);
+
+	intel_encoders_enable(state, crtc);
+
+	if (psl_clkgate_wa) {
+		intel_crtc_wait_for_next_vblank(crtc);
+		glk_pipe_scaler_clock_gating_wa(dev_priv, pipe, false);
+	}
+
+	/* If we change the relative order between pipe/planes enabling, we need
+	 * to change the workaround. */
+	hsw_workaround_pipe = new_crtc_state->hsw_workaround_pipe;
+	if (IS_HASWELL(dev_priv) && hsw_workaround_pipe != INVALID_PIPE) {
+		struct intel_crtc *wa_crtc;
+
+		wa_crtc = intel_crtc_for_pipe(dev_priv, hsw_workaround_pipe);
+
+		intel_crtc_wait_for_next_vblank(wa_crtc);
+		intel_crtc_wait_for_next_vblank(wa_crtc);
+	}
+}
+
+void ilk_pfit_disable(const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	/* To avoid upsetting the power well on haswell only disable the pfit if
+	 * it's in use. The hw state code will make sure we get this right. */
+	if (!old_crtc_state->pch_pfit.enabled)
+		return;
+
+	intel_de_write_fw(dev_priv, PF_CTL(pipe), 0);
+	intel_de_write_fw(dev_priv, PF_WIN_POS(pipe), 0);
+	intel_de_write_fw(dev_priv, PF_WIN_SZ(pipe), 0);
+}
+
+static void ilk_crtc_disable(struct intel_atomic_state *state,
+			     struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	/*
+	 * Sometimes spurious CPU pipe underruns happen when the
+	 * pipe is already disabled, but FDI RX/TX is still enabled.
+	 * Happens at least with VGA+HDMI cloning. Suppress them.
+	 */
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false);
+	intel_set_pch_fifo_underrun_reporting(dev_priv, pipe, false);
+
+	intel_encoders_disable(state, crtc);
+
+	intel_crtc_vblank_off(old_crtc_state);
+
+	intel_disable_transcoder(old_crtc_state);
+
+	ilk_pfit_disable(old_crtc_state);
+
+	if (old_crtc_state->has_pch_encoder)
+		ilk_pch_disable(state, crtc);
+
+	intel_encoders_post_disable(state, crtc);
+
+	if (old_crtc_state->has_pch_encoder)
+		ilk_pch_post_disable(state, crtc);
+
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
+	intel_set_pch_fifo_underrun_reporting(dev_priv, pipe, true);
+}
+
+static void hsw_crtc_disable(struct intel_atomic_state *state,
+			     struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+
+	/*
+	 * FIXME collapse everything to one hook.
+	 * Need care with mst->ddi interactions.
+	 */
+	if (!intel_crtc_is_bigjoiner_slave(old_crtc_state)) {
+		intel_encoders_disable(state, crtc);
+		intel_encoders_post_disable(state, crtc);
+	}
+}
+
+static void i9xx_pfit_enable(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	if (!crtc_state->gmch_pfit.control)
+		return;
+
+	/*
+	 * The panel fitter should only be adjusted whilst the pipe is disabled,
+	 * according to register description and PRM.
+	 */
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, PFIT_CONTROL) & PFIT_ENABLE);
+	assert_transcoder_disabled(dev_priv, crtc_state->cpu_transcoder);
+
+	intel_de_write(dev_priv, PFIT_PGM_RATIOS,
+		       crtc_state->gmch_pfit.pgm_ratios);
+	intel_de_write(dev_priv, PFIT_CONTROL, crtc_state->gmch_pfit.control);
+
+	/* Border color in case we don't scale up to the full screen. Black by
+	 * default, change to something else for debugging. */
+	intel_de_write(dev_priv, BCLRPAT(crtc->pipe), 0);
+}
+
+bool intel_phy_is_combo(struct drm_i915_private *dev_priv, enum phy phy)
+{
+	if (phy == PHY_NONE)
+		return false;
+	else if (IS_ALDERLAKE_S(dev_priv))
+		return phy <= PHY_E;
+	else if (IS_DG1(dev_priv) || IS_ROCKETLAKE(dev_priv))
+		return phy <= PHY_D;
+	else if (IS_JSL_EHL(dev_priv))
+		return phy <= PHY_C;
+	else if (IS_ALDERLAKE_P(dev_priv) || IS_DISPLAY_VER(dev_priv, 11, 12))
+		return phy <= PHY_B;
+	else
+		/*
+		 * DG2 outputs labelled as "combo PHY" in the bspec use
+		 * SNPS PHYs with completely different programming,
+		 * hence we always return false here.
+		 */
+		return false;
+}
+
+bool intel_phy_is_tc(struct drm_i915_private *dev_priv, enum phy phy)
+{
+	if (IS_DG2(dev_priv))
+		/* DG2's "TC1" output uses a SNPS PHY */
+		return false;
+	else if (IS_ALDERLAKE_P(dev_priv))
+		return phy >= PHY_F && phy <= PHY_I;
+	else if (IS_TIGERLAKE(dev_priv))
+		return phy >= PHY_D && phy <= PHY_I;
+	else if (IS_ICELAKE(dev_priv))
+		return phy >= PHY_C && phy <= PHY_F;
+	else
+		return false;
+}
+
+bool intel_phy_is_snps(struct drm_i915_private *dev_priv, enum phy phy)
+{
+	if (phy == PHY_NONE)
+		return false;
+	else if (IS_DG2(dev_priv))
+		/*
+		 * All four "combo" ports and the TC1 port (PHY E) use
+		 * Synopsis PHYs.
+		 */
+		return phy <= PHY_E;
+
+	return false;
+}
+
+enum phy intel_port_to_phy(struct drm_i915_private *i915, enum port port)
+{
+	if (DISPLAY_VER(i915) >= 13 && port >= PORT_D_XELPD)
+		return PHY_D + port - PORT_D_XELPD;
+	else if (DISPLAY_VER(i915) >= 13 && port >= PORT_TC1)
+		return PHY_F + port - PORT_TC1;
+	else if (IS_ALDERLAKE_S(i915) && port >= PORT_TC1)
+		return PHY_B + port - PORT_TC1;
+	else if ((IS_DG1(i915) || IS_ROCKETLAKE(i915)) && port >= PORT_TC1)
+		return PHY_C + port - PORT_TC1;
+	else if (IS_JSL_EHL(i915) && port == PORT_D)
+		return PHY_A;
+
+	return PHY_A + port - PORT_A;
+}
+
+enum tc_port intel_port_to_tc(struct drm_i915_private *dev_priv, enum port port)
+{
+	if (!intel_phy_is_tc(dev_priv, intel_port_to_phy(dev_priv, port)))
+		return TC_PORT_NONE;
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		return TC_PORT_1 + port - PORT_TC1;
+	else
+		return TC_PORT_1 + port - PORT_C;
+}
+
+enum intel_display_power_domain
+intel_aux_power_domain(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	if (intel_tc_port_in_tbt_alt_mode(dig_port))
+		return intel_display_power_tbt_aux_domain(i915, dig_port->aux_ch);
+
+	return intel_display_power_legacy_aux_domain(i915, dig_port->aux_ch);
+}
+
+static void get_crtc_power_domains(struct intel_crtc_state *crtc_state,
+				   struct intel_power_domain_mask *mask)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	struct drm_encoder *encoder;
+	enum pipe pipe = crtc->pipe;
+
+	bitmap_zero(mask->bits, POWER_DOMAIN_NUM);
+
+	if (!crtc_state->hw.active)
+		return;
+
+	set_bit(POWER_DOMAIN_PIPE(pipe), mask->bits);
+	set_bit(POWER_DOMAIN_TRANSCODER(cpu_transcoder), mask->bits);
+	if (crtc_state->pch_pfit.enabled ||
+	    crtc_state->pch_pfit.force_thru)
+		set_bit(POWER_DOMAIN_PIPE_PANEL_FITTER(pipe), mask->bits);
+
+	drm_for_each_encoder_mask(encoder, &dev_priv->drm,
+				  crtc_state->uapi.encoder_mask) {
+		struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
+
+		set_bit(intel_encoder->power_domain, mask->bits);
+	}
+
+	if (HAS_DDI(dev_priv) && crtc_state->has_audio)
+		set_bit(POWER_DOMAIN_AUDIO_MMIO, mask->bits);
+
+	if (crtc_state->shared_dpll)
+		set_bit(POWER_DOMAIN_DISPLAY_CORE, mask->bits);
+
+	if (crtc_state->dsc.compression_enable)
+		set_bit(intel_dsc_power_domain(crtc, cpu_transcoder), mask->bits);
+}
+
+void intel_modeset_get_crtc_power_domains(struct intel_crtc_state *crtc_state,
+					  struct intel_power_domain_mask *old_domains)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum intel_display_power_domain domain;
+	struct intel_power_domain_mask domains, new_domains;
+
+	get_crtc_power_domains(crtc_state, &domains);
+
+	bitmap_andnot(new_domains.bits,
+		      domains.bits,
+		      crtc->enabled_power_domains.mask.bits,
+		      POWER_DOMAIN_NUM);
+	bitmap_andnot(old_domains->bits,
+		      crtc->enabled_power_domains.mask.bits,
+		      domains.bits,
+		      POWER_DOMAIN_NUM);
+
+	for_each_power_domain(domain, &new_domains)
+		intel_display_power_get_in_set(dev_priv,
+					       &crtc->enabled_power_domains,
+					       domain);
+}
+
+void intel_modeset_put_crtc_power_domains(struct intel_crtc *crtc,
+					  struct intel_power_domain_mask *domains)
+{
+	intel_display_power_put_mask_in_set(to_i915(crtc->base.dev),
+					    &crtc->enabled_power_domains,
+					    domains);
+}
+
+static void i9xx_configure_cpu_transcoder(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	if (intel_crtc_has_dp_encoder(crtc_state)) {
+		intel_cpu_transcoder_set_m1_n1(crtc, cpu_transcoder,
+					       &crtc_state->dp_m_n);
+		intel_cpu_transcoder_set_m2_n2(crtc, cpu_transcoder,
+					       &crtc_state->dp_m2_n2);
+	}
+
+	intel_set_transcoder_timings(crtc_state);
+
+	i9xx_set_pipeconf(crtc_state);
+}
+
+static void valleyview_crtc_enable(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	if (drm_WARN_ON(&dev_priv->drm, crtc->active))
+		return;
+
+	i9xx_configure_cpu_transcoder(new_crtc_state);
+
+	intel_set_pipe_src_size(new_crtc_state);
+
+	if (IS_CHERRYVIEW(dev_priv) && pipe == PIPE_B) {
+		intel_de_write(dev_priv, CHV_BLEND(pipe), CHV_BLEND_LEGACY);
+		intel_de_write(dev_priv, CHV_CANVAS(pipe), 0);
+	}
+
+	crtc->active = true;
+
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
+
+	intel_encoders_pre_pll_enable(state, crtc);
+
+	if (IS_CHERRYVIEW(dev_priv))
+		chv_enable_pll(new_crtc_state);
+	else
+		vlv_enable_pll(new_crtc_state);
+
+	intel_encoders_pre_enable(state, crtc);
+
+	i9xx_pfit_enable(new_crtc_state);
+
+	intel_color_load_luts(new_crtc_state);
+	intel_color_commit_noarm(new_crtc_state);
+	intel_color_commit_arm(new_crtc_state);
+	/* update DSPCNTR to configure gamma for pipe bottom color */
+	intel_disable_primary_plane(new_crtc_state);
+
+	intel_initial_watermarks(state, crtc);
+	intel_enable_transcoder(new_crtc_state);
+
+	intel_crtc_vblank_on(new_crtc_state);
+
+	intel_encoders_enable(state, crtc);
+}
+
+static void i9xx_crtc_enable(struct intel_atomic_state *state,
+			     struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	if (drm_WARN_ON(&dev_priv->drm, crtc->active))
+		return;
+
+	i9xx_configure_cpu_transcoder(new_crtc_state);
+
+	intel_set_pipe_src_size(new_crtc_state);
+
+	crtc->active = true;
+
+	if (DISPLAY_VER(dev_priv) != 2)
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
+
+	intel_encoders_pre_enable(state, crtc);
+
+	i9xx_enable_pll(new_crtc_state);
+
+	i9xx_pfit_enable(new_crtc_state);
+
+	intel_color_load_luts(new_crtc_state);
+	intel_color_commit_noarm(new_crtc_state);
+	intel_color_commit_arm(new_crtc_state);
+	/* update DSPCNTR to configure gamma for pipe bottom color */
+	intel_disable_primary_plane(new_crtc_state);
+
+	if (!intel_initial_watermarks(state, crtc))
+		intel_update_watermarks(dev_priv);
+	intel_enable_transcoder(new_crtc_state);
+
+	intel_crtc_vblank_on(new_crtc_state);
+
+	intel_encoders_enable(state, crtc);
+
+	/* prevents spurious underruns */
+	if (DISPLAY_VER(dev_priv) == 2)
+		intel_crtc_wait_for_next_vblank(crtc);
+}
+
+static void i9xx_pfit_disable(const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	if (!old_crtc_state->gmch_pfit.control)
+		return;
+
+	assert_transcoder_disabled(dev_priv, old_crtc_state->cpu_transcoder);
+
+	drm_dbg_kms(&dev_priv->drm, "disabling pfit, current: 0x%08x\n",
+		    intel_de_read(dev_priv, PFIT_CONTROL));
+	intel_de_write(dev_priv, PFIT_CONTROL, 0);
+}
+
+static void i9xx_crtc_disable(struct intel_atomic_state *state,
+			      struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	/*
+	 * On gen2 planes are double buffered but the pipe isn't, so we must
+	 * wait for planes to fully turn off before disabling the pipe.
+	 */
+	if (DISPLAY_VER(dev_priv) == 2)
+		intel_crtc_wait_for_next_vblank(crtc);
+
+	intel_encoders_disable(state, crtc);
+
+	intel_crtc_vblank_off(old_crtc_state);
+
+	intel_disable_transcoder(old_crtc_state);
+
+	i9xx_pfit_disable(old_crtc_state);
+
+	intel_encoders_post_disable(state, crtc);
+
+	if (!intel_crtc_has_type(old_crtc_state, INTEL_OUTPUT_DSI)) {
+		if (IS_CHERRYVIEW(dev_priv))
+			chv_disable_pll(dev_priv, pipe);
+		else if (IS_VALLEYVIEW(dev_priv))
+			vlv_disable_pll(dev_priv, pipe);
+		else
+			i9xx_disable_pll(old_crtc_state);
+	}
+
+	intel_encoders_post_pll_disable(state, crtc);
+
+	if (DISPLAY_VER(dev_priv) != 2)
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false);
+
+	if (!dev_priv->display.funcs.wm->initial_watermarks)
+		intel_update_watermarks(dev_priv);
+
+	/* clock the pipe down to 640x480@60 to potentially save power */
+	if (IS_I830(dev_priv))
+		i830_enable_pipe(dev_priv, pipe);
+}
+
+
+/*
+ * turn all crtc's off, but do not adjust state
+ * This has to be paired with a call to intel_modeset_setup_hw_state.
+ */
+int intel_display_suspend(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_atomic_state *state;
+	int ret;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return 0;
+
+	state = drm_atomic_helper_suspend(dev);
+	ret = PTR_ERR_OR_ZERO(state);
+	if (ret)
+		drm_err(&dev_priv->drm, "Suspending crtc's failed with %i\n",
+			ret);
+	else
+		dev_priv->modeset_restore_state = state;
+	return ret;
+}
+
+void intel_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
+
+	drm_encoder_cleanup(encoder);
+	kfree(intel_encoder);
+}
+
+static bool intel_crtc_supports_double_wide(const struct intel_crtc *crtc)
+{
+	const struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	/* GDG double wide on either pipe, otherwise pipe A only */
+	return DISPLAY_VER(dev_priv) < 4 &&
+		(crtc->pipe == PIPE_A || IS_I915G(dev_priv));
+}
+
+static u32 ilk_pipe_pixel_rate(const struct intel_crtc_state *crtc_state)
+{
+	u32 pixel_rate = crtc_state->hw.pipe_mode.crtc_clock;
+	struct drm_rect src;
+
+	/*
+	 * We only use IF-ID interlacing. If we ever use
+	 * PF-ID we'll need to adjust the pixel_rate here.
+	 */
+
+	if (!crtc_state->pch_pfit.enabled)
+		return pixel_rate;
+
+	drm_rect_init(&src, 0, 0,
+		      drm_rect_width(&crtc_state->pipe_src) << 16,
+		      drm_rect_height(&crtc_state->pipe_src) << 16);
+
+	return intel_adjusted_rate(&src, &crtc_state->pch_pfit.dst,
+				   pixel_rate);
+}
+
+static void intel_mode_from_crtc_timings(struct drm_display_mode *mode,
+					 const struct drm_display_mode *timings)
+{
+	mode->hdisplay = timings->crtc_hdisplay;
+	mode->htotal = timings->crtc_htotal;
+	mode->hsync_start = timings->crtc_hsync_start;
+	mode->hsync_end = timings->crtc_hsync_end;
+
+	mode->vdisplay = timings->crtc_vdisplay;
+	mode->vtotal = timings->crtc_vtotal;
+	mode->vsync_start = timings->crtc_vsync_start;
+	mode->vsync_end = timings->crtc_vsync_end;
+
+	mode->flags = timings->flags;
+	mode->type = DRM_MODE_TYPE_DRIVER;
+
+	mode->clock = timings->crtc_clock;
+
+	drm_mode_set_name(mode);
+}
+
+static void intel_crtc_compute_pixel_rate(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	if (HAS_GMCH(dev_priv))
+		/* FIXME calculate proper pipe pixel rate for GMCH pfit */
+		crtc_state->pixel_rate =
+			crtc_state->hw.pipe_mode.crtc_clock;
+	else
+		crtc_state->pixel_rate =
+			ilk_pipe_pixel_rate(crtc_state);
+}
+
+static void intel_bigjoiner_adjust_timings(const struct intel_crtc_state *crtc_state,
+					   struct drm_display_mode *mode)
+{
+	int num_pipes = intel_bigjoiner_num_pipes(crtc_state);
+
+	if (num_pipes < 2)
+		return;
+
+	mode->crtc_clock /= num_pipes;
+	mode->crtc_hdisplay /= num_pipes;
+	mode->crtc_hblank_start /= num_pipes;
+	mode->crtc_hblank_end /= num_pipes;
+	mode->crtc_hsync_start /= num_pipes;
+	mode->crtc_hsync_end /= num_pipes;
+	mode->crtc_htotal /= num_pipes;
+}
+
+static void intel_splitter_adjust_timings(const struct intel_crtc_state *crtc_state,
+					  struct drm_display_mode *mode)
+{
+	int overlap = crtc_state->splitter.pixel_overlap;
+	int n = crtc_state->splitter.link_count;
+
+	if (!crtc_state->splitter.enable)
+		return;
+
+	/*
+	 * eDP MSO uses segment timings from EDID for transcoder
+	 * timings, but full mode for everything else.
+	 *
+	 * h_full = (h_segment - pixel_overlap) * link_count
+	 */
+	mode->crtc_hdisplay = (mode->crtc_hdisplay - overlap) * n;
+	mode->crtc_hblank_start = (mode->crtc_hblank_start - overlap) * n;
+	mode->crtc_hblank_end = (mode->crtc_hblank_end - overlap) * n;
+	mode->crtc_hsync_start = (mode->crtc_hsync_start - overlap) * n;
+	mode->crtc_hsync_end = (mode->crtc_hsync_end - overlap) * n;
+	mode->crtc_htotal = (mode->crtc_htotal - overlap) * n;
+	mode->crtc_clock *= n;
+}
+
+static void intel_crtc_readout_derived_state(struct intel_crtc_state *crtc_state)
+{
+	struct drm_display_mode *mode = &crtc_state->hw.mode;
+	struct drm_display_mode *pipe_mode = &crtc_state->hw.pipe_mode;
+	struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+
+	/*
+	 * Start with the adjusted_mode crtc timings, which
+	 * have been filled with the transcoder timings.
+	 */
+	drm_mode_copy(pipe_mode, adjusted_mode);
+
+	/* Expand MSO per-segment transcoder timings to full */
+	intel_splitter_adjust_timings(crtc_state, pipe_mode);
+
+	/*
+	 * We want the full numbers in adjusted_mode normal timings,
+	 * adjusted_mode crtc timings are left with the raw transcoder
+	 * timings.
+	 */
+	intel_mode_from_crtc_timings(adjusted_mode, pipe_mode);
+
+	/* Populate the "user" mode with full numbers */
+	drm_mode_copy(mode, pipe_mode);
+	intel_mode_from_crtc_timings(mode, mode);
+	mode->hdisplay = drm_rect_width(&crtc_state->pipe_src) *
+		(intel_bigjoiner_num_pipes(crtc_state) ?: 1);
+	mode->vdisplay = drm_rect_height(&crtc_state->pipe_src);
+
+	/* Derive per-pipe timings in case bigjoiner is used */
+	intel_bigjoiner_adjust_timings(crtc_state, pipe_mode);
+	intel_mode_from_crtc_timings(pipe_mode, pipe_mode);
+
+	intel_crtc_compute_pixel_rate(crtc_state);
+}
+
+void intel_encoder_get_config(struct intel_encoder *encoder,
+			      struct intel_crtc_state *crtc_state)
+{
+	encoder->get_config(encoder, crtc_state);
+
+	intel_crtc_readout_derived_state(crtc_state);
+}
+
+static void intel_bigjoiner_compute_pipe_src(struct intel_crtc_state *crtc_state)
+{
+	int num_pipes = intel_bigjoiner_num_pipes(crtc_state);
+	int width, height;
+
+	if (num_pipes < 2)
+		return;
+
+	width = drm_rect_width(&crtc_state->pipe_src);
+	height = drm_rect_height(&crtc_state->pipe_src);
+
+	drm_rect_init(&crtc_state->pipe_src, 0, 0,
+		      width / num_pipes, height);
+}
+
+static int intel_crtc_compute_pipe_src(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	intel_bigjoiner_compute_pipe_src(crtc_state);
+
+	/*
+	 * Pipe horizontal size must be even in:
+	 * - DVO ganged mode
+	 * - LVDS dual channel mode
+	 * - Double wide pipe
+	 */
+	if (drm_rect_width(&crtc_state->pipe_src) & 1) {
+		if (crtc_state->double_wide) {
+			drm_dbg_kms(&i915->drm,
+				    "[CRTC:%d:%s] Odd pipe source width not supported with double wide pipe\n",
+				    crtc->base.base.id, crtc->base.name);
+			return -EINVAL;
+		}
+
+		if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+		    intel_is_dual_link_lvds(i915)) {
+			drm_dbg_kms(&i915->drm,
+				    "[CRTC:%d:%s] Odd pipe source width not supported with dual link LVDS\n",
+				    crtc->base.base.id, crtc->base.name);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static int intel_crtc_compute_pipe_mode(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	struct drm_display_mode *pipe_mode = &crtc_state->hw.pipe_mode;
+	int clock_limit = i915->max_dotclk_freq;
+
+	/*
+	 * Start with the adjusted_mode crtc timings, which
+	 * have been filled with the transcoder timings.
+	 */
+	drm_mode_copy(pipe_mode, adjusted_mode);
+
+	/* Expand MSO per-segment transcoder timings to full */
+	intel_splitter_adjust_timings(crtc_state, pipe_mode);
+
+	/* Derive per-pipe timings in case bigjoiner is used */
+	intel_bigjoiner_adjust_timings(crtc_state, pipe_mode);
+	intel_mode_from_crtc_timings(pipe_mode, pipe_mode);
+
+	if (DISPLAY_VER(i915) < 4) {
+		clock_limit = i915->display.cdclk.max_cdclk_freq * 9 / 10;
+
+		/*
+		 * Enable double wide mode when the dot clock
+		 * is > 90% of the (display) core speed.
+		 */
+		if (intel_crtc_supports_double_wide(crtc) &&
+		    pipe_mode->crtc_clock > clock_limit) {
+			clock_limit = i915->max_dotclk_freq;
+			crtc_state->double_wide = true;
+		}
+	}
+
+	if (pipe_mode->crtc_clock > clock_limit) {
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] requested pixel clock (%d kHz) too high (max: %d kHz, double wide: %s)\n",
+			    crtc->base.base.id, crtc->base.name,
+			    pipe_mode->crtc_clock, clock_limit,
+			    str_yes_no(crtc_state->double_wide));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int intel_crtc_compute_config(struct intel_atomic_state *state,
+				     struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	int ret;
+
+	ret = intel_dpll_crtc_compute_clock(state, crtc);
+	if (ret)
+		return ret;
+
+	ret = intel_crtc_compute_pipe_src(crtc_state);
+	if (ret)
+		return ret;
+
+	ret = intel_crtc_compute_pipe_mode(crtc_state);
+	if (ret)
+		return ret;
+
+	intel_crtc_compute_pixel_rate(crtc_state);
+
+	if (crtc_state->has_pch_encoder)
+		return ilk_fdi_compute_config(crtc, crtc_state);
+
+	return 0;
+}
+
+static void
+intel_reduce_m_n_ratio(u32 *num, u32 *den)
+{
+	while (*num > DATA_LINK_M_N_MASK ||
+	       *den > DATA_LINK_M_N_MASK) {
+		*num >>= 1;
+		*den >>= 1;
+	}
+}
+
+static void compute_m_n(u32 *ret_m, u32 *ret_n,
+			u32 m, u32 n, u32 constant_n)
+{
+	if (constant_n)
+		*ret_n = constant_n;
+	else
+		*ret_n = min_t(unsigned int, roundup_pow_of_two(n), DATA_LINK_N_MAX);
+
+	*ret_m = div_u64(mul_u32_u32(m, *ret_n), n);
+	intel_reduce_m_n_ratio(ret_m, ret_n);
+}
+
+void
+intel_link_compute_m_n(u16 bits_per_pixel, int nlanes,
+		       int pixel_clock, int link_clock,
+		       struct intel_link_m_n *m_n,
+		       bool fec_enable)
+{
+	u32 data_clock = bits_per_pixel * pixel_clock;
+
+	if (fec_enable)
+		data_clock = intel_dp_mode_to_fec_clock(data_clock);
+
+	/*
+	 * Windows/BIOS uses fixed M/N values always. Follow suit.
+	 *
+	 * Also several DP dongles in particular seem to be fussy
+	 * about too large link M/N values. Presumably the 20bit
+	 * value used by Windows/BIOS is acceptable to everyone.
+	 */
+	m_n->tu = 64;
+	compute_m_n(&m_n->data_m, &m_n->data_n,
+		    data_clock, link_clock * nlanes * 8,
+		    0x8000000);
+
+	compute_m_n(&m_n->link_m, &m_n->link_n,
+		    pixel_clock, link_clock,
+		    0x80000);
+}
+
+static void intel_panel_sanitize_ssc(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * There may be no VBT; and if the BIOS enabled SSC we can
+	 * just keep using it to avoid unnecessary flicker.  Whereas if the
+	 * BIOS isn't using it, don't assume it will work even if the VBT
+	 * indicates as much.
+	 */
+	if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)) {
+		bool bios_lvds_use_ssc = intel_de_read(dev_priv,
+						       PCH_DREF_CONTROL) &
+			DREF_SSC1_ENABLE;
+
+		if (dev_priv->display.vbt.lvds_use_ssc != bios_lvds_use_ssc) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "SSC %s by BIOS, overriding VBT which says %s\n",
+				    str_enabled_disabled(bios_lvds_use_ssc),
+				    str_enabled_disabled(dev_priv->display.vbt.lvds_use_ssc));
+			dev_priv->display.vbt.lvds_use_ssc = bios_lvds_use_ssc;
+		}
+	}
+}
+
+void intel_zero_m_n(struct intel_link_m_n *m_n)
+{
+	/* corresponds to 0 register value */
+	memset(m_n, 0, sizeof(*m_n));
+	m_n->tu = 1;
+}
+
+void intel_set_m_n(struct drm_i915_private *i915,
+		   const struct intel_link_m_n *m_n,
+		   i915_reg_t data_m_reg, i915_reg_t data_n_reg,
+		   i915_reg_t link_m_reg, i915_reg_t link_n_reg)
+{
+	intel_de_write(i915, data_m_reg, TU_SIZE(m_n->tu) | m_n->data_m);
+	intel_de_write(i915, data_n_reg, m_n->data_n);
+	intel_de_write(i915, link_m_reg, m_n->link_m);
+	/*
+	 * On BDW+ writing LINK_N arms the double buffered update
+	 * of all the M/N registers, so it must be written last.
+	 */
+	intel_de_write(i915, link_n_reg, m_n->link_n);
+}
+
+bool intel_cpu_transcoder_has_m2_n2(struct drm_i915_private *dev_priv,
+				    enum transcoder transcoder)
+{
+	if (IS_HASWELL(dev_priv))
+		return transcoder == TRANSCODER_EDP;
+
+	return IS_DISPLAY_VER(dev_priv, 5, 7) || IS_CHERRYVIEW(dev_priv);
+}
+
+void intel_cpu_transcoder_set_m1_n1(struct intel_crtc *crtc,
+				    enum transcoder transcoder,
+				    const struct intel_link_m_n *m_n)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	if (DISPLAY_VER(dev_priv) >= 5)
+		intel_set_m_n(dev_priv, m_n,
+			      PIPE_DATA_M1(transcoder), PIPE_DATA_N1(transcoder),
+			      PIPE_LINK_M1(transcoder), PIPE_LINK_N1(transcoder));
+	else
+		intel_set_m_n(dev_priv, m_n,
+			      PIPE_DATA_M_G4X(pipe), PIPE_DATA_N_G4X(pipe),
+			      PIPE_LINK_M_G4X(pipe), PIPE_LINK_N_G4X(pipe));
+}
+
+void intel_cpu_transcoder_set_m2_n2(struct intel_crtc *crtc,
+				    enum transcoder transcoder,
+				    const struct intel_link_m_n *m_n)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	if (!intel_cpu_transcoder_has_m2_n2(dev_priv, transcoder))
+		return;
+
+	intel_set_m_n(dev_priv, m_n,
+		      PIPE_DATA_M2(transcoder), PIPE_DATA_N2(transcoder),
+		      PIPE_LINK_M2(transcoder), PIPE_LINK_N2(transcoder));
+}
+
+static void intel_set_transcoder_timings(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	u32 crtc_vtotal, crtc_vblank_end;
+	int vsyncshift = 0;
+
+	/* We need to be careful not to changed the adjusted mode, for otherwise
+	 * the hw state checker will get angry at the mismatch. */
+	crtc_vtotal = adjusted_mode->crtc_vtotal;
+	crtc_vblank_end = adjusted_mode->crtc_vblank_end;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
+		/* the chip adds 2 halflines automatically */
+		crtc_vtotal -= 1;
+		crtc_vblank_end -= 1;
+
+		if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO))
+			vsyncshift = (adjusted_mode->crtc_htotal - 1) / 2;
+		else
+			vsyncshift = adjusted_mode->crtc_hsync_start -
+				adjusted_mode->crtc_htotal / 2;
+		if (vsyncshift < 0)
+			vsyncshift += adjusted_mode->crtc_htotal;
+	}
+
+	if (DISPLAY_VER(dev_priv) > 3)
+		intel_de_write(dev_priv, VSYNCSHIFT(cpu_transcoder),
+		               vsyncshift);
+
+	intel_de_write(dev_priv, HTOTAL(cpu_transcoder),
+		       (adjusted_mode->crtc_hdisplay - 1) | ((adjusted_mode->crtc_htotal - 1) << 16));
+	intel_de_write(dev_priv, HBLANK(cpu_transcoder),
+		       (adjusted_mode->crtc_hblank_start - 1) | ((adjusted_mode->crtc_hblank_end - 1) << 16));
+	intel_de_write(dev_priv, HSYNC(cpu_transcoder),
+		       (adjusted_mode->crtc_hsync_start - 1) | ((adjusted_mode->crtc_hsync_end - 1) << 16));
+
+	intel_de_write(dev_priv, VTOTAL(cpu_transcoder),
+		       (adjusted_mode->crtc_vdisplay - 1) | ((crtc_vtotal - 1) << 16));
+	intel_de_write(dev_priv, VBLANK(cpu_transcoder),
+		       (adjusted_mode->crtc_vblank_start - 1) | ((crtc_vblank_end - 1) << 16));
+	intel_de_write(dev_priv, VSYNC(cpu_transcoder),
+		       (adjusted_mode->crtc_vsync_start - 1) | ((adjusted_mode->crtc_vsync_end - 1) << 16));
+
+	/* Workaround: when the EDP input selection is B, the VTOTAL_B must be
+	 * programmed with the VTOTAL_EDP value. Same for VTOTAL_C. This is
+	 * documented on the DDI_FUNC_CTL register description, EDP Input Select
+	 * bits. */
+	if (IS_HASWELL(dev_priv) && cpu_transcoder == TRANSCODER_EDP &&
+	    (pipe == PIPE_B || pipe == PIPE_C))
+		intel_de_write(dev_priv, VTOTAL(pipe),
+		               intel_de_read(dev_priv, VTOTAL(cpu_transcoder)));
+
+}
+
+static void intel_set_pipe_src_size(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	int width = drm_rect_width(&crtc_state->pipe_src);
+	int height = drm_rect_height(&crtc_state->pipe_src);
+	enum pipe pipe = crtc->pipe;
+
+	/* pipesrc controls the size that is scaled from, which should
+	 * always be the user's requested size.
+	 */
+	intel_de_write(dev_priv, PIPESRC(pipe),
+		       PIPESRC_WIDTH(width - 1) | PIPESRC_HEIGHT(height - 1));
+}
+
+static bool intel_pipe_is_interlaced(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	if (DISPLAY_VER(dev_priv) == 2)
+		return false;
+
+	if (DISPLAY_VER(dev_priv) >= 9 ||
+	    IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		return intel_de_read(dev_priv, PIPECONF(cpu_transcoder)) & PIPECONF_INTERLACE_MASK_HSW;
+	else
+		return intel_de_read(dev_priv, PIPECONF(cpu_transcoder)) & PIPECONF_INTERLACE_MASK;
+}
+
+static void intel_get_transcoder_timings(struct intel_crtc *crtc,
+					 struct intel_crtc_state *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum transcoder cpu_transcoder = pipe_config->cpu_transcoder;
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, HTOTAL(cpu_transcoder));
+	pipe_config->hw.adjusted_mode.crtc_hdisplay = (tmp & 0xffff) + 1;
+	pipe_config->hw.adjusted_mode.crtc_htotal = ((tmp >> 16) & 0xffff) + 1;
+
+	if (!transcoder_is_dsi(cpu_transcoder)) {
+		tmp = intel_de_read(dev_priv, HBLANK(cpu_transcoder));
+		pipe_config->hw.adjusted_mode.crtc_hblank_start =
+							(tmp & 0xffff) + 1;
+		pipe_config->hw.adjusted_mode.crtc_hblank_end =
+						((tmp >> 16) & 0xffff) + 1;
+	}
+	tmp = intel_de_read(dev_priv, HSYNC(cpu_transcoder));
+	pipe_config->hw.adjusted_mode.crtc_hsync_start = (tmp & 0xffff) + 1;
+	pipe_config->hw.adjusted_mode.crtc_hsync_end = ((tmp >> 16) & 0xffff) + 1;
+
+	tmp = intel_de_read(dev_priv, VTOTAL(cpu_transcoder));
+	pipe_config->hw.adjusted_mode.crtc_vdisplay = (tmp & 0xffff) + 1;
+	pipe_config->hw.adjusted_mode.crtc_vtotal = ((tmp >> 16) & 0xffff) + 1;
+
+	if (!transcoder_is_dsi(cpu_transcoder)) {
+		tmp = intel_de_read(dev_priv, VBLANK(cpu_transcoder));
+		pipe_config->hw.adjusted_mode.crtc_vblank_start =
+							(tmp & 0xffff) + 1;
+		pipe_config->hw.adjusted_mode.crtc_vblank_end =
+						((tmp >> 16) & 0xffff) + 1;
+	}
+	tmp = intel_de_read(dev_priv, VSYNC(cpu_transcoder));
+	pipe_config->hw.adjusted_mode.crtc_vsync_start = (tmp & 0xffff) + 1;
+	pipe_config->hw.adjusted_mode.crtc_vsync_end = ((tmp >> 16) & 0xffff) + 1;
+
+	if (intel_pipe_is_interlaced(pipe_config)) {
+		pipe_config->hw.adjusted_mode.flags |= DRM_MODE_FLAG_INTERLACE;
+		pipe_config->hw.adjusted_mode.crtc_vtotal += 1;
+		pipe_config->hw.adjusted_mode.crtc_vblank_end += 1;
+	}
+}
+
+static void intel_bigjoiner_adjust_pipe_src(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	int num_pipes = intel_bigjoiner_num_pipes(crtc_state);
+	enum pipe master_pipe, pipe = crtc->pipe;
+	int width;
+
+	if (num_pipes < 2)
+		return;
+
+	master_pipe = bigjoiner_master_pipe(crtc_state);
+	width = drm_rect_width(&crtc_state->pipe_src);
+
+	drm_rect_translate_to(&crtc_state->pipe_src,
+			      (pipe - master_pipe) * width, 0);
+}
+
+static void intel_get_pipe_src_size(struct intel_crtc *crtc,
+				    struct intel_crtc_state *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, PIPESRC(crtc->pipe));
+
+	drm_rect_init(&pipe_config->pipe_src, 0, 0,
+		      REG_FIELD_GET(PIPESRC_WIDTH_MASK, tmp) + 1,
+		      REG_FIELD_GET(PIPESRC_HEIGHT_MASK, tmp) + 1);
+
+	intel_bigjoiner_adjust_pipe_src(pipe_config);
+}
+
+void i9xx_set_pipeconf(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 pipeconf = 0;
+
+	/*
+	 * - We keep both pipes enabled on 830
+	 * - During modeset the pipe is still disabled and must remain so
+	 * - During fastset the pipe is already enabled and must remain so
+	 */
+	if (IS_I830(dev_priv) || !intel_crtc_needs_modeset(crtc_state))
+		pipeconf |= PIPECONF_ENABLE;
+
+	if (crtc_state->double_wide)
+		pipeconf |= PIPECONF_DOUBLE_WIDE;
+
+	/* only g4x and later have fancy bpc/dither controls */
+	if (IS_G4X(dev_priv) || IS_VALLEYVIEW(dev_priv) ||
+	    IS_CHERRYVIEW(dev_priv)) {
+		/* Bspec claims that we can't use dithering for 30bpp pipes. */
+		if (crtc_state->dither && crtc_state->pipe_bpp != 30)
+			pipeconf |= PIPECONF_DITHER_EN |
+				    PIPECONF_DITHER_TYPE_SP;
+
+		switch (crtc_state->pipe_bpp) {
+		default:
+			/* Case prevented by intel_choose_pipe_bpp_dither. */
+			MISSING_CASE(crtc_state->pipe_bpp);
+			fallthrough;
+		case 18:
+			pipeconf |= PIPECONF_BPC_6;
+			break;
+		case 24:
+			pipeconf |= PIPECONF_BPC_8;
+			break;
+		case 30:
+			pipeconf |= PIPECONF_BPC_10;
+			break;
+		}
+	}
+
+	if (crtc_state->hw.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE) {
+		if (DISPLAY_VER(dev_priv) < 4 ||
+		    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO))
+			pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
+		else
+			pipeconf |= PIPECONF_INTERLACE_W_SYNC_SHIFT;
+	} else {
+		pipeconf |= PIPECONF_INTERLACE_PROGRESSIVE;
+	}
+
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	     crtc_state->limited_color_range)
+		pipeconf |= PIPECONF_COLOR_RANGE_SELECT;
+
+	pipeconf |= PIPECONF_GAMMA_MODE(crtc_state->gamma_mode);
+
+	pipeconf |= PIPECONF_FRAME_START_DELAY(crtc_state->framestart_delay - 1);
+
+	intel_de_write(dev_priv, PIPECONF(crtc->pipe), pipeconf);
+	intel_de_posting_read(dev_priv, PIPECONF(crtc->pipe));
+}
+
+static bool i9xx_has_pfit(struct drm_i915_private *dev_priv)
+{
+	if (IS_I830(dev_priv))
+		return false;
+
+	return DISPLAY_VER(dev_priv) >= 4 ||
+		IS_PINEVIEW(dev_priv) || IS_MOBILE(dev_priv);
+}
+
+static void i9xx_get_pfit_config(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 tmp;
+
+	if (!i9xx_has_pfit(dev_priv))
+		return;
+
+	tmp = intel_de_read(dev_priv, PFIT_CONTROL);
+	if (!(tmp & PFIT_ENABLE))
+		return;
+
+	/* Check whether the pfit is attached to our pipe. */
+	if (DISPLAY_VER(dev_priv) < 4) {
+		if (crtc->pipe != PIPE_B)
+			return;
+	} else {
+		if ((tmp & PFIT_PIPE_MASK) != (crtc->pipe << PFIT_PIPE_SHIFT))
+			return;
+	}
+
+	crtc_state->gmch_pfit.control = tmp;
+	crtc_state->gmch_pfit.pgm_ratios =
+		intel_de_read(dev_priv, PFIT_PGM_RATIOS);
+}
+
+static void vlv_crtc_clock_get(struct intel_crtc *crtc,
+			       struct intel_crtc_state *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	struct dpll clock;
+	u32 mdiv;
+	int refclk = 100000;
+
+	/* In case of DSI, DPLL will not be used */
+	if ((pipe_config->dpll_hw_state.dpll & DPLL_VCO_ENABLE) == 0)
+		return;
+
+	vlv_dpio_get(dev_priv);
+	mdiv = vlv_dpio_read(dev_priv, pipe, VLV_PLL_DW3(pipe));
+	vlv_dpio_put(dev_priv);
+
+	clock.m1 = (mdiv >> DPIO_M1DIV_SHIFT) & 7;
+	clock.m2 = mdiv & DPIO_M2DIV_MASK;
+	clock.n = (mdiv >> DPIO_N_SHIFT) & 0xf;
+	clock.p1 = (mdiv >> DPIO_P1_SHIFT) & 7;
+	clock.p2 = (mdiv >> DPIO_P2_SHIFT) & 0x1f;
+
+	pipe_config->port_clock = vlv_calc_dpll_params(refclk, &clock);
+}
+
+static void chv_crtc_clock_get(struct intel_crtc *crtc,
+			       struct intel_crtc_state *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	enum dpio_channel port = vlv_pipe_to_channel(pipe);
+	struct dpll clock;
+	u32 cmn_dw13, pll_dw0, pll_dw1, pll_dw2, pll_dw3;
+	int refclk = 100000;
+
+	/* In case of DSI, DPLL will not be used */
+	if ((pipe_config->dpll_hw_state.dpll & DPLL_VCO_ENABLE) == 0)
+		return;
+
+	vlv_dpio_get(dev_priv);
+	cmn_dw13 = vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW13(port));
+	pll_dw0 = vlv_dpio_read(dev_priv, pipe, CHV_PLL_DW0(port));
+	pll_dw1 = vlv_dpio_read(dev_priv, pipe, CHV_PLL_DW1(port));
+	pll_dw2 = vlv_dpio_read(dev_priv, pipe, CHV_PLL_DW2(port));
+	pll_dw3 = vlv_dpio_read(dev_priv, pipe, CHV_PLL_DW3(port));
+	vlv_dpio_put(dev_priv);
+
+	clock.m1 = (pll_dw1 & 0x7) == DPIO_CHV_M1_DIV_BY_2 ? 2 : 0;
+	clock.m2 = (pll_dw0 & 0xff) << 22;
+	if (pll_dw3 & DPIO_CHV_FRAC_DIV_EN)
+		clock.m2 |= pll_dw2 & 0x3fffff;
+	clock.n = (pll_dw1 >> DPIO_CHV_N_DIV_SHIFT) & 0xf;
+	clock.p1 = (cmn_dw13 >> DPIO_CHV_P1_DIV_SHIFT) & 0x7;
+	clock.p2 = (cmn_dw13 >> DPIO_CHV_P2_DIV_SHIFT) & 0x1f;
+
+	pipe_config->port_clock = chv_calc_dpll_params(refclk, &clock);
+}
+
+static enum intel_output_format
+bdw_get_pipemisc_output_format(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, PIPEMISC(crtc->pipe));
+
+	if (tmp & PIPEMISC_YUV420_ENABLE) {
+		/* We support 4:2:0 in full blend mode only */
+		drm_WARN_ON(&dev_priv->drm,
+			    (tmp & PIPEMISC_YUV420_MODE_FULL_BLEND) == 0);
+
+		return INTEL_OUTPUT_FORMAT_YCBCR420;
+	} else if (tmp & PIPEMISC_OUTPUT_COLORSPACE_YUV) {
+		return INTEL_OUTPUT_FORMAT_YCBCR444;
+	} else {
+		return INTEL_OUTPUT_FORMAT_RGB;
+	}
+}
+
+static void i9xx_get_pipe_color_config(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct intel_plane *plane = to_intel_plane(crtc->base.primary);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum i9xx_plane_id i9xx_plane = plane->i9xx_plane;
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, DSPCNTR(i9xx_plane));
+
+	if (tmp & DISP_PIPE_GAMMA_ENABLE)
+		crtc_state->gamma_enable = true;
+
+	if (!HAS_GMCH(dev_priv) &&
+	    tmp & DISP_PIPE_CSC_ENABLE)
+		crtc_state->csc_enable = true;
+}
+
+static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
+				 struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum intel_display_power_domain power_domain;
+	intel_wakeref_t wakeref;
+	u32 tmp;
+	bool ret;
+
+	power_domain = POWER_DOMAIN_PIPE(crtc->pipe);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref)
+		return false;
+
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+	pipe_config->cpu_transcoder = (enum transcoder) crtc->pipe;
+	pipe_config->shared_dpll = NULL;
+
+	ret = false;
+
+	tmp = intel_de_read(dev_priv, PIPECONF(crtc->pipe));
+	if (!(tmp & PIPECONF_ENABLE))
+		goto out;
+
+	if (IS_G4X(dev_priv) || IS_VALLEYVIEW(dev_priv) ||
+	    IS_CHERRYVIEW(dev_priv)) {
+		switch (tmp & PIPECONF_BPC_MASK) {
+		case PIPECONF_BPC_6:
+			pipe_config->pipe_bpp = 18;
+			break;
+		case PIPECONF_BPC_8:
+			pipe_config->pipe_bpp = 24;
+			break;
+		case PIPECONF_BPC_10:
+			pipe_config->pipe_bpp = 30;
+			break;
+		default:
+			MISSING_CASE(tmp);
+			break;
+		}
+	}
+
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    (tmp & PIPECONF_COLOR_RANGE_SELECT))
+		pipe_config->limited_color_range = true;
+
+	pipe_config->gamma_mode = REG_FIELD_GET(PIPECONF_GAMMA_MODE_MASK_I9XX, tmp);
+
+	pipe_config->framestart_delay = REG_FIELD_GET(PIPECONF_FRAME_START_DELAY_MASK, tmp) + 1;
+
+	if (IS_CHERRYVIEW(dev_priv))
+		pipe_config->cgm_mode = intel_de_read(dev_priv,
+						      CGM_PIPE_MODE(crtc->pipe));
+
+	i9xx_get_pipe_color_config(pipe_config);
+	intel_color_get_config(pipe_config);
+
+	if (DISPLAY_VER(dev_priv) < 4)
+		pipe_config->double_wide = tmp & PIPECONF_DOUBLE_WIDE;
+
+	intel_get_transcoder_timings(crtc, pipe_config);
+	intel_get_pipe_src_size(crtc, pipe_config);
+
+	i9xx_get_pfit_config(pipe_config);
+
+	if (DISPLAY_VER(dev_priv) >= 4) {
+		/* No way to read it out on pipes B and C */
+		if (IS_CHERRYVIEW(dev_priv) && crtc->pipe != PIPE_A)
+			tmp = dev_priv->chv_dpll_md[crtc->pipe];
+		else
+			tmp = intel_de_read(dev_priv, DPLL_MD(crtc->pipe));
+		pipe_config->pixel_multiplier =
+			((tmp & DPLL_MD_UDI_MULTIPLIER_MASK)
+			 >> DPLL_MD_UDI_MULTIPLIER_SHIFT) + 1;
+		pipe_config->dpll_hw_state.dpll_md = tmp;
+	} else if (IS_I945G(dev_priv) || IS_I945GM(dev_priv) ||
+		   IS_G33(dev_priv) || IS_PINEVIEW(dev_priv)) {
+		tmp = intel_de_read(dev_priv, DPLL(crtc->pipe));
+		pipe_config->pixel_multiplier =
+			((tmp & SDVO_MULTIPLIER_MASK)
+			 >> SDVO_MULTIPLIER_SHIFT_HIRES) + 1;
+	} else {
+		/* Note that on i915G/GM the pixel multiplier is in the sdvo
+		 * port and will be fixed up in the encoder->get_config
+		 * function. */
+		pipe_config->pixel_multiplier = 1;
+	}
+	pipe_config->dpll_hw_state.dpll = intel_de_read(dev_priv,
+							DPLL(crtc->pipe));
+	if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv)) {
+		pipe_config->dpll_hw_state.fp0 = intel_de_read(dev_priv,
+							       FP0(crtc->pipe));
+		pipe_config->dpll_hw_state.fp1 = intel_de_read(dev_priv,
+							       FP1(crtc->pipe));
+	} else {
+		/* Mask out read-only status bits. */
+		pipe_config->dpll_hw_state.dpll &= ~(DPLL_LOCK_VLV |
+						     DPLL_PORTC_READY_MASK |
+						     DPLL_PORTB_READY_MASK);
+	}
+
+	if (IS_CHERRYVIEW(dev_priv))
+		chv_crtc_clock_get(crtc, pipe_config);
+	else if (IS_VALLEYVIEW(dev_priv))
+		vlv_crtc_clock_get(crtc, pipe_config);
+	else
+		i9xx_crtc_clock_get(crtc, pipe_config);
+
+	/*
+	 * Normally the dotclock is filled in by the encoder .get_config()
+	 * but in case the pipe is enabled w/o any ports we need a sane
+	 * default.
+	 */
+	pipe_config->hw.adjusted_mode.crtc_clock =
+		pipe_config->port_clock / pipe_config->pixel_multiplier;
+
+	ret = true;
+
+out:
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+
+	return ret;
+}
+
+void ilk_set_pipeconf(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	u32 val = 0;
+
+	/*
+	 * - During modeset the pipe is still disabled and must remain so
+	 * - During fastset the pipe is already enabled and must remain so
+	 */
+	if (!intel_crtc_needs_modeset(crtc_state))
+		val |= PIPECONF_ENABLE;
+
+	switch (crtc_state->pipe_bpp) {
+	default:
+		/* Case prevented by intel_choose_pipe_bpp_dither. */
+		MISSING_CASE(crtc_state->pipe_bpp);
+		fallthrough;
+	case 18:
+		val |= PIPECONF_BPC_6;
+		break;
+	case 24:
+		val |= PIPECONF_BPC_8;
+		break;
+	case 30:
+		val |= PIPECONF_BPC_10;
+		break;
+	case 36:
+		val |= PIPECONF_BPC_12;
+		break;
+	}
+
+	if (crtc_state->dither)
+		val |= PIPECONF_DITHER_EN | PIPECONF_DITHER_TYPE_SP;
+
+	if (crtc_state->hw.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
+		val |= PIPECONF_INTERLACE_IF_ID_ILK;
+	else
+		val |= PIPECONF_INTERLACE_PF_PD_ILK;
+
+	/*
+	 * This would end up with an odd purple hue over
+	 * the entire display. Make sure we don't do it.
+	 */
+	drm_WARN_ON(&dev_priv->drm, crtc_state->limited_color_range &&
+		    crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB);
+
+	if (crtc_state->limited_color_range &&
+	    !intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO))
+		val |= PIPECONF_COLOR_RANGE_SELECT;
+
+	if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB)
+		val |= PIPECONF_OUTPUT_COLORSPACE_YUV709;
+
+	val |= PIPECONF_GAMMA_MODE(crtc_state->gamma_mode);
+
+	val |= PIPECONF_FRAME_START_DELAY(crtc_state->framestart_delay - 1);
+	val |= PIPECONF_MSA_TIMING_DELAY(crtc_state->msa_timing_delay);
+
+	intel_de_write(dev_priv, PIPECONF(pipe), val);
+	intel_de_posting_read(dev_priv, PIPECONF(pipe));
+}
+
+static void hsw_set_transconf(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 val = 0;
+
+	/*
+	 * - During modeset the pipe is still disabled and must remain so
+	 * - During fastset the pipe is already enabled and must remain so
+	 */
+	if (!intel_crtc_needs_modeset(crtc_state))
+		val |= PIPECONF_ENABLE;
+
+	if (IS_HASWELL(dev_priv) && crtc_state->dither)
+		val |= PIPECONF_DITHER_EN | PIPECONF_DITHER_TYPE_SP;
+
+	if (crtc_state->hw.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
+		val |= PIPECONF_INTERLACE_IF_ID_ILK;
+	else
+		val |= PIPECONF_INTERLACE_PF_PD_ILK;
+
+	if (IS_HASWELL(dev_priv) &&
+	    crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB)
+		val |= PIPECONF_OUTPUT_COLORSPACE_YUV_HSW;
+
+	intel_de_write(dev_priv, PIPECONF(cpu_transcoder), val);
+	intel_de_posting_read(dev_priv, PIPECONF(cpu_transcoder));
+}
+
+static void bdw_set_pipemisc(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 val = 0;
+
+	switch (crtc_state->pipe_bpp) {
+	case 18:
+		val |= PIPEMISC_BPC_6;
+		break;
+	case 24:
+		val |= PIPEMISC_BPC_8;
+		break;
+	case 30:
+		val |= PIPEMISC_BPC_10;
+		break;
+	case 36:
+		/* Port output 12BPC defined for ADLP+ */
+		if (DISPLAY_VER(dev_priv) > 12)
+			val |= PIPEMISC_BPC_12_ADLP;
+		break;
+	default:
+		MISSING_CASE(crtc_state->pipe_bpp);
+		break;
+	}
+
+	if (crtc_state->dither)
+		val |= PIPEMISC_DITHER_ENABLE | PIPEMISC_DITHER_TYPE_SP;
+
+	if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR420 ||
+	    crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR444)
+		val |= PIPEMISC_OUTPUT_COLORSPACE_YUV;
+
+	if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR420)
+		val |= PIPEMISC_YUV420_ENABLE |
+			PIPEMISC_YUV420_MODE_FULL_BLEND;
+
+	if (DISPLAY_VER(dev_priv) >= 11 && is_hdr_mode(crtc_state))
+		val |= PIPEMISC_HDR_MODE_PRECISION;
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		val |= PIPEMISC_PIXEL_ROUNDING_TRUNC;
+
+	intel_de_write(dev_priv, PIPEMISC(crtc->pipe), val);
+}
+
+int bdw_get_pipemisc_bpp(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, PIPEMISC(crtc->pipe));
+
+	switch (tmp & PIPEMISC_BPC_MASK) {
+	case PIPEMISC_BPC_6:
+		return 18;
+	case PIPEMISC_BPC_8:
+		return 24;
+	case PIPEMISC_BPC_10:
+		return 30;
+	/*
+	 * PORT OUTPUT 12 BPC defined for ADLP+.
+	 *
+	 * TODO:
+	 * For previous platforms with DSI interface, bits 5:7
+	 * are used for storing pipe_bpp irrespective of dithering.
+	 * Since the value of 12 BPC is not defined for these bits
+	 * on older platforms, need to find a workaround for 12 BPC
+	 * MIPI DSI HW readout.
+	 */
+	case PIPEMISC_BPC_12_ADLP:
+		if (DISPLAY_VER(dev_priv) > 12)
+			return 36;
+		fallthrough;
+	default:
+		MISSING_CASE(tmp);
+		return 0;
+	}
+}
+
+int ilk_get_lanes_required(int target_clock, int link_bw, int bpp)
+{
+	/*
+	 * Account for spread spectrum to avoid
+	 * oversubscribing the link. Max center spread
+	 * is 2.5%; use 5% for safety's sake.
+	 */
+	u32 bps = target_clock * bpp * 21 / 20;
+	return DIV_ROUND_UP(bps, link_bw * 8);
+}
+
+void intel_get_m_n(struct drm_i915_private *i915,
+		   struct intel_link_m_n *m_n,
+		   i915_reg_t data_m_reg, i915_reg_t data_n_reg,
+		   i915_reg_t link_m_reg, i915_reg_t link_n_reg)
+{
+	m_n->link_m = intel_de_read(i915, link_m_reg) & DATA_LINK_M_N_MASK;
+	m_n->link_n = intel_de_read(i915, link_n_reg) & DATA_LINK_M_N_MASK;
+	m_n->data_m = intel_de_read(i915, data_m_reg) & DATA_LINK_M_N_MASK;
+	m_n->data_n = intel_de_read(i915, data_n_reg) & DATA_LINK_M_N_MASK;
+	m_n->tu = REG_FIELD_GET(TU_SIZE_MASK, intel_de_read(i915, data_m_reg)) + 1;
+}
+
+void intel_cpu_transcoder_get_m1_n1(struct intel_crtc *crtc,
+				    enum transcoder transcoder,
+				    struct intel_link_m_n *m_n)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	if (DISPLAY_VER(dev_priv) >= 5)
+		intel_get_m_n(dev_priv, m_n,
+			      PIPE_DATA_M1(transcoder), PIPE_DATA_N1(transcoder),
+			      PIPE_LINK_M1(transcoder), PIPE_LINK_N1(transcoder));
+	else
+		intel_get_m_n(dev_priv, m_n,
+			      PIPE_DATA_M_G4X(pipe), PIPE_DATA_N_G4X(pipe),
+			      PIPE_LINK_M_G4X(pipe), PIPE_LINK_N_G4X(pipe));
+}
+
+void intel_cpu_transcoder_get_m2_n2(struct intel_crtc *crtc,
+				    enum transcoder transcoder,
+				    struct intel_link_m_n *m_n)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	if (!intel_cpu_transcoder_has_m2_n2(dev_priv, transcoder))
+		return;
+
+	intel_get_m_n(dev_priv, m_n,
+		      PIPE_DATA_M2(transcoder), PIPE_DATA_N2(transcoder),
+		      PIPE_LINK_M2(transcoder), PIPE_LINK_N2(transcoder));
+}
+
+static void ilk_get_pfit_pos_size(struct intel_crtc_state *crtc_state,
+				  u32 pos, u32 size)
+{
+	drm_rect_init(&crtc_state->pch_pfit.dst,
+		      pos >> 16, pos & 0xffff,
+		      size >> 16, size & 0xffff);
+}
+
+static void skl_get_pfit_config(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_scaler_state *scaler_state = &crtc_state->scaler_state;
+	int id = -1;
+	int i;
+
+	/* find scaler attached to this pipe */
+	for (i = 0; i < crtc->num_scalers; i++) {
+		u32 ctl, pos, size;
+
+		ctl = intel_de_read(dev_priv, SKL_PS_CTRL(crtc->pipe, i));
+		if ((ctl & (PS_SCALER_EN | PS_PLANE_SEL_MASK)) != PS_SCALER_EN)
+			continue;
+
+		id = i;
+		crtc_state->pch_pfit.enabled = true;
+
+		pos = intel_de_read(dev_priv, SKL_PS_WIN_POS(crtc->pipe, i));
+		size = intel_de_read(dev_priv, SKL_PS_WIN_SZ(crtc->pipe, i));
+
+		ilk_get_pfit_pos_size(crtc_state, pos, size);
+
+		scaler_state->scalers[i].in_use = true;
+		break;
+	}
+
+	scaler_state->scaler_id = id;
+	if (id >= 0)
+		scaler_state->scaler_users |= (1 << SKL_CRTC_INDEX);
+	else
+		scaler_state->scaler_users &= ~(1 << SKL_CRTC_INDEX);
+}
+
+static void ilk_get_pfit_config(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 ctl, pos, size;
+
+	ctl = intel_de_read(dev_priv, PF_CTL(crtc->pipe));
+	if ((ctl & PF_ENABLE) == 0)
+		return;
+
+	crtc_state->pch_pfit.enabled = true;
+
+	pos = intel_de_read(dev_priv, PF_WIN_POS(crtc->pipe));
+	size = intel_de_read(dev_priv, PF_WIN_SZ(crtc->pipe));
+
+	ilk_get_pfit_pos_size(crtc_state, pos, size);
+
+	/*
+	 * We currently do not free assignements of panel fitters on
+	 * ivb/hsw (since we don't use the higher upscaling modes which
+	 * differentiates them) so just WARN about this case for now.
+	 */
+	drm_WARN_ON(&dev_priv->drm, DISPLAY_VER(dev_priv) == 7 &&
+		    (ctl & PF_PIPE_SEL_MASK_IVB) != PF_PIPE_SEL_IVB(crtc->pipe));
+}
+
+static bool ilk_get_pipe_config(struct intel_crtc *crtc,
+				struct intel_crtc_state *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum intel_display_power_domain power_domain;
+	intel_wakeref_t wakeref;
+	u32 tmp;
+	bool ret;
+
+	power_domain = POWER_DOMAIN_PIPE(crtc->pipe);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref)
+		return false;
+
+	pipe_config->cpu_transcoder = (enum transcoder) crtc->pipe;
+	pipe_config->shared_dpll = NULL;
+
+	ret = false;
+	tmp = intel_de_read(dev_priv, PIPECONF(crtc->pipe));
+	if (!(tmp & PIPECONF_ENABLE))
+		goto out;
+
+	switch (tmp & PIPECONF_BPC_MASK) {
+	case PIPECONF_BPC_6:
+		pipe_config->pipe_bpp = 18;
+		break;
+	case PIPECONF_BPC_8:
+		pipe_config->pipe_bpp = 24;
+		break;
+	case PIPECONF_BPC_10:
+		pipe_config->pipe_bpp = 30;
+		break;
+	case PIPECONF_BPC_12:
+		pipe_config->pipe_bpp = 36;
+		break;
+	default:
+		break;
+	}
+
+	if (tmp & PIPECONF_COLOR_RANGE_SELECT)
+		pipe_config->limited_color_range = true;
+
+	switch (tmp & PIPECONF_OUTPUT_COLORSPACE_MASK) {
+	case PIPECONF_OUTPUT_COLORSPACE_YUV601:
+	case PIPECONF_OUTPUT_COLORSPACE_YUV709:
+		pipe_config->output_format = INTEL_OUTPUT_FORMAT_YCBCR444;
+		break;
+	default:
+		pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+		break;
+	}
+
+	pipe_config->gamma_mode = REG_FIELD_GET(PIPECONF_GAMMA_MODE_MASK_ILK, tmp);
+
+	pipe_config->framestart_delay = REG_FIELD_GET(PIPECONF_FRAME_START_DELAY_MASK, tmp) + 1;
+
+	pipe_config->msa_timing_delay = REG_FIELD_GET(PIPECONF_MSA_TIMING_DELAY_MASK, tmp);
+
+	pipe_config->csc_mode = intel_de_read(dev_priv,
+					      PIPE_CSC_MODE(crtc->pipe));
+
+	i9xx_get_pipe_color_config(pipe_config);
+	intel_color_get_config(pipe_config);
+
+	pipe_config->pixel_multiplier = 1;
+
+	ilk_pch_get_config(pipe_config);
+
+	intel_get_transcoder_timings(crtc, pipe_config);
+	intel_get_pipe_src_size(crtc, pipe_config);
+
+	ilk_get_pfit_config(pipe_config);
+
+	ret = true;
+
+out:
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+
+	return ret;
+}
+
+static u8 bigjoiner_pipes(struct drm_i915_private *i915)
+{
+	u8 pipes;
+
+	if (DISPLAY_VER(i915) >= 12)
+		pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D);
+	else if (DISPLAY_VER(i915) >= 11)
+		pipes = BIT(PIPE_B) | BIT(PIPE_C);
+	else
+		pipes = 0;
+
+	return pipes & RUNTIME_INFO(i915)->pipe_mask;
+}
+
+static bool transcoder_ddi_func_is_enabled(struct drm_i915_private *dev_priv,
+					   enum transcoder cpu_transcoder)
+{
+	enum intel_display_power_domain power_domain;
+	intel_wakeref_t wakeref;
+	u32 tmp = 0;
+
+	power_domain = POWER_DOMAIN_TRANSCODER(cpu_transcoder);
+
+	with_intel_display_power_if_enabled(dev_priv, power_domain, wakeref)
+		tmp = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder));
+
+	return tmp & TRANS_DDI_FUNC_ENABLE;
+}
+
+static void enabled_bigjoiner_pipes(struct drm_i915_private *dev_priv,
+				    u8 *master_pipes, u8 *slave_pipes)
+{
+	struct intel_crtc *crtc;
+
+	*master_pipes = 0;
+	*slave_pipes = 0;
+
+	for_each_intel_crtc_in_pipe_mask(&dev_priv->drm, crtc,
+					 bigjoiner_pipes(dev_priv)) {
+		enum intel_display_power_domain power_domain;
+		enum pipe pipe = crtc->pipe;
+		intel_wakeref_t wakeref;
+
+		power_domain = intel_dsc_power_domain(crtc, (enum transcoder) pipe);
+		with_intel_display_power_if_enabled(dev_priv, power_domain, wakeref) {
+			u32 tmp = intel_de_read(dev_priv, ICL_PIPE_DSS_CTL1(pipe));
+
+			if (!(tmp & BIG_JOINER_ENABLE))
+				continue;
+
+			if (tmp & MASTER_BIG_JOINER_ENABLE)
+				*master_pipes |= BIT(pipe);
+			else
+				*slave_pipes |= BIT(pipe);
+		}
+
+		if (DISPLAY_VER(dev_priv) < 13)
+			continue;
+
+		power_domain = POWER_DOMAIN_PIPE(pipe);
+		with_intel_display_power_if_enabled(dev_priv, power_domain, wakeref) {
+			u32 tmp = intel_de_read(dev_priv, ICL_PIPE_DSS_CTL1(pipe));
+
+			if (tmp & UNCOMPRESSED_JOINER_MASTER)
+				*master_pipes |= BIT(pipe);
+			if (tmp & UNCOMPRESSED_JOINER_SLAVE)
+				*slave_pipes |= BIT(pipe);
+		}
+	}
+
+	/* Bigjoiner pipes should always be consecutive master and slave */
+	drm_WARN(&dev_priv->drm, *slave_pipes != *master_pipes << 1,
+		 "Bigjoiner misconfigured (master pipes 0x%x, slave pipes 0x%x)\n",
+		 *master_pipes, *slave_pipes);
+}
+
+static enum pipe get_bigjoiner_master_pipe(enum pipe pipe, u8 master_pipes, u8 slave_pipes)
+{
+	if ((slave_pipes & BIT(pipe)) == 0)
+		return pipe;
+
+	/* ignore everything above our pipe */
+	master_pipes &= ~GENMASK(7, pipe);
+
+	/* highest remaining bit should be our master pipe */
+	return fls(master_pipes) - 1;
+}
+
+static u8 get_bigjoiner_slave_pipes(enum pipe pipe, u8 master_pipes, u8 slave_pipes)
+{
+	enum pipe master_pipe, next_master_pipe;
+
+	master_pipe = get_bigjoiner_master_pipe(pipe, master_pipes, slave_pipes);
+
+	if ((master_pipes & BIT(master_pipe)) == 0)
+		return 0;
+
+	/* ignore our master pipe and everything below it */
+	master_pipes &= ~GENMASK(master_pipe, 0);
+	/* make sure a high bit is set for the ffs() */
+	master_pipes |= BIT(7);
+	/* lowest remaining bit should be the next master pipe */
+	next_master_pipe = ffs(master_pipes) - 1;
+
+	return slave_pipes & GENMASK(next_master_pipe - 1, master_pipe);
+}
+
+static u8 hsw_panel_transcoders(struct drm_i915_private *i915)
+{
+	u8 panel_transcoder_mask = BIT(TRANSCODER_EDP);
+
+	if (DISPLAY_VER(i915) >= 11)
+		panel_transcoder_mask |= BIT(TRANSCODER_DSI_0) | BIT(TRANSCODER_DSI_1);
+
+	return panel_transcoder_mask;
+}
+
+static u8 hsw_enabled_transcoders(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u8 panel_transcoder_mask = hsw_panel_transcoders(dev_priv);
+	enum transcoder cpu_transcoder;
+	u8 master_pipes, slave_pipes;
+	u8 enabled_transcoders = 0;
+
+	/*
+	 * XXX: Do intel_display_power_get_if_enabled before reading this (for
+	 * consistency and less surprising code; it's in always on power).
+	 */
+	for_each_cpu_transcoder_masked(dev_priv, cpu_transcoder,
+				       panel_transcoder_mask) {
+		enum intel_display_power_domain power_domain;
+		intel_wakeref_t wakeref;
+		enum pipe trans_pipe;
+		u32 tmp = 0;
+
+		power_domain = POWER_DOMAIN_TRANSCODER(cpu_transcoder);
+		with_intel_display_power_if_enabled(dev_priv, power_domain, wakeref)
+			tmp = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder));
+
+		if (!(tmp & TRANS_DDI_FUNC_ENABLE))
+			continue;
+
+		switch (tmp & TRANS_DDI_EDP_INPUT_MASK) {
+		default:
+			drm_WARN(dev, 1,
+				 "unknown pipe linked to transcoder %s\n",
+				 transcoder_name(cpu_transcoder));
+			fallthrough;
+		case TRANS_DDI_EDP_INPUT_A_ONOFF:
+		case TRANS_DDI_EDP_INPUT_A_ON:
+			trans_pipe = PIPE_A;
+			break;
+		case TRANS_DDI_EDP_INPUT_B_ONOFF:
+			trans_pipe = PIPE_B;
+			break;
+		case TRANS_DDI_EDP_INPUT_C_ONOFF:
+			trans_pipe = PIPE_C;
+			break;
+		case TRANS_DDI_EDP_INPUT_D_ONOFF:
+			trans_pipe = PIPE_D;
+			break;
+		}
+
+		if (trans_pipe == crtc->pipe)
+			enabled_transcoders |= BIT(cpu_transcoder);
+	}
+
+	/* single pipe or bigjoiner master */
+	cpu_transcoder = (enum transcoder) crtc->pipe;
+	if (transcoder_ddi_func_is_enabled(dev_priv, cpu_transcoder))
+		enabled_transcoders |= BIT(cpu_transcoder);
+
+	/* bigjoiner slave -> consider the master pipe's transcoder as well */
+	enabled_bigjoiner_pipes(dev_priv, &master_pipes, &slave_pipes);
+	if (slave_pipes & BIT(crtc->pipe)) {
+		cpu_transcoder = (enum transcoder)
+			get_bigjoiner_master_pipe(crtc->pipe, master_pipes, slave_pipes);
+		if (transcoder_ddi_func_is_enabled(dev_priv, cpu_transcoder))
+			enabled_transcoders |= BIT(cpu_transcoder);
+	}
+
+	return enabled_transcoders;
+}
+
+static bool has_edp_transcoders(u8 enabled_transcoders)
+{
+	return enabled_transcoders & BIT(TRANSCODER_EDP);
+}
+
+static bool has_dsi_transcoders(u8 enabled_transcoders)
+{
+	return enabled_transcoders & (BIT(TRANSCODER_DSI_0) |
+				      BIT(TRANSCODER_DSI_1));
+}
+
+static bool has_pipe_transcoders(u8 enabled_transcoders)
+{
+	return enabled_transcoders & ~(BIT(TRANSCODER_EDP) |
+				       BIT(TRANSCODER_DSI_0) |
+				       BIT(TRANSCODER_DSI_1));
+}
+
+static void assert_enabled_transcoders(struct drm_i915_private *i915,
+				       u8 enabled_transcoders)
+{
+	/* Only one type of transcoder please */
+	drm_WARN_ON(&i915->drm,
+		    has_edp_transcoders(enabled_transcoders) +
+		    has_dsi_transcoders(enabled_transcoders) +
+		    has_pipe_transcoders(enabled_transcoders) > 1);
+
+	/* Only DSI transcoders can be ganged */
+	drm_WARN_ON(&i915->drm,
+		    !has_dsi_transcoders(enabled_transcoders) &&
+		    !is_power_of_2(enabled_transcoders));
+}
+
+static bool hsw_get_transcoder_state(struct intel_crtc *crtc,
+				     struct intel_crtc_state *pipe_config,
+				     struct intel_display_power_domain_set *power_domain_set)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	unsigned long enabled_transcoders;
+	u32 tmp;
+
+	enabled_transcoders = hsw_enabled_transcoders(crtc);
+	if (!enabled_transcoders)
+		return false;
+
+	assert_enabled_transcoders(dev_priv, enabled_transcoders);
+
+	/*
+	 * With the exception of DSI we should only ever have
+	 * a single enabled transcoder. With DSI let's just
+	 * pick the first one.
+	 */
+	pipe_config->cpu_transcoder = ffs(enabled_transcoders) - 1;
+
+	if (!intel_display_power_get_in_set_if_enabled(dev_priv, power_domain_set,
+						       POWER_DOMAIN_TRANSCODER(pipe_config->cpu_transcoder)))
+		return false;
+
+	if (hsw_panel_transcoders(dev_priv) & BIT(pipe_config->cpu_transcoder)) {
+		tmp = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(pipe_config->cpu_transcoder));
+
+		if ((tmp & TRANS_DDI_EDP_INPUT_MASK) == TRANS_DDI_EDP_INPUT_A_ONOFF)
+			pipe_config->pch_pfit.force_thru = true;
+	}
+
+	tmp = intel_de_read(dev_priv, PIPECONF(pipe_config->cpu_transcoder));
+
+	return tmp & PIPECONF_ENABLE;
+}
+
+static bool bxt_get_dsi_transcoder_state(struct intel_crtc *crtc,
+					 struct intel_crtc_state *pipe_config,
+					 struct intel_display_power_domain_set *power_domain_set)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum transcoder cpu_transcoder;
+	enum port port;
+	u32 tmp;
+
+	for_each_port_masked(port, BIT(PORT_A) | BIT(PORT_C)) {
+		if (port == PORT_A)
+			cpu_transcoder = TRANSCODER_DSI_A;
+		else
+			cpu_transcoder = TRANSCODER_DSI_C;
+
+		if (!intel_display_power_get_in_set_if_enabled(dev_priv, power_domain_set,
+							       POWER_DOMAIN_TRANSCODER(cpu_transcoder)))
+			continue;
+
+		/*
+		 * The PLL needs to be enabled with a valid divider
+		 * configuration, otherwise accessing DSI registers will hang
+		 * the machine. See BSpec North Display Engine
+		 * registers/MIPI[BXT]. We can break out here early, since we
+		 * need the same DSI PLL to be enabled for both DSI ports.
+		 */
+		if (!bxt_dsi_pll_is_enabled(dev_priv))
+			break;
+
+		/* XXX: this works for video mode only */
+		tmp = intel_de_read(dev_priv, BXT_MIPI_PORT_CTRL(port));
+		if (!(tmp & DPI_ENABLE))
+			continue;
+
+		tmp = intel_de_read(dev_priv, MIPI_CTRL(port));
+		if ((tmp & BXT_PIPE_SELECT_MASK) != BXT_PIPE_SELECT(crtc->pipe))
+			continue;
+
+		pipe_config->cpu_transcoder = cpu_transcoder;
+		break;
+	}
+
+	return transcoder_is_dsi(pipe_config->cpu_transcoder);
+}
+
+static void intel_bigjoiner_get_config(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	u8 master_pipes, slave_pipes;
+	enum pipe pipe = crtc->pipe;
+
+	enabled_bigjoiner_pipes(i915, &master_pipes, &slave_pipes);
+
+	if (((master_pipes | slave_pipes) & BIT(pipe)) == 0)
+		return;
+
+	crtc_state->bigjoiner_pipes =
+		BIT(get_bigjoiner_master_pipe(pipe, master_pipes, slave_pipes)) |
+		get_bigjoiner_slave_pipes(pipe, master_pipes, slave_pipes);
+}
+
+static bool hsw_get_pipe_config(struct intel_crtc *crtc,
+				struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_display_power_domain_set power_domain_set = { };
+	bool active;
+	u32 tmp;
+
+	if (!intel_display_power_get_in_set_if_enabled(dev_priv, &power_domain_set,
+						       POWER_DOMAIN_PIPE(crtc->pipe)))
+		return false;
+
+	pipe_config->shared_dpll = NULL;
+
+	active = hsw_get_transcoder_state(crtc, pipe_config, &power_domain_set);
+
+	if ((IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) &&
+	    bxt_get_dsi_transcoder_state(crtc, pipe_config, &power_domain_set)) {
+		drm_WARN_ON(&dev_priv->drm, active);
+		active = true;
+	}
+
+	if (!active)
+		goto out;
+
+	intel_dsc_get_config(pipe_config);
+	intel_bigjoiner_get_config(pipe_config);
+
+	if (!transcoder_is_dsi(pipe_config->cpu_transcoder) ||
+	    DISPLAY_VER(dev_priv) >= 11)
+		intel_get_transcoder_timings(crtc, pipe_config);
+
+	if (HAS_VRR(dev_priv) && !transcoder_is_dsi(pipe_config->cpu_transcoder))
+		intel_vrr_get_config(crtc, pipe_config);
+
+	intel_get_pipe_src_size(crtc, pipe_config);
+
+	if (IS_HASWELL(dev_priv)) {
+		u32 tmp = intel_de_read(dev_priv,
+					PIPECONF(pipe_config->cpu_transcoder));
+
+		if (tmp & PIPECONF_OUTPUT_COLORSPACE_YUV_HSW)
+			pipe_config->output_format = INTEL_OUTPUT_FORMAT_YCBCR444;
+		else
+			pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+	} else {
+		pipe_config->output_format =
+			bdw_get_pipemisc_output_format(crtc);
+	}
+
+	pipe_config->gamma_mode = intel_de_read(dev_priv,
+						GAMMA_MODE(crtc->pipe));
+
+	pipe_config->csc_mode = intel_de_read(dev_priv,
+					      PIPE_CSC_MODE(crtc->pipe));
+
+	if (DISPLAY_VER(dev_priv) >= 9) {
+		tmp = intel_de_read(dev_priv, SKL_BOTTOM_COLOR(crtc->pipe));
+
+		if (tmp & SKL_BOTTOM_COLOR_GAMMA_ENABLE)
+			pipe_config->gamma_enable = true;
+
+		if (tmp & SKL_BOTTOM_COLOR_CSC_ENABLE)
+			pipe_config->csc_enable = true;
+	} else {
+		i9xx_get_pipe_color_config(pipe_config);
+	}
+
+	intel_color_get_config(pipe_config);
+
+	tmp = intel_de_read(dev_priv, WM_LINETIME(crtc->pipe));
+	pipe_config->linetime = REG_FIELD_GET(HSW_LINETIME_MASK, tmp);
+	if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		pipe_config->ips_linetime =
+			REG_FIELD_GET(HSW_IPS_LINETIME_MASK, tmp);
+
+	if (intel_display_power_get_in_set_if_enabled(dev_priv, &power_domain_set,
+						      POWER_DOMAIN_PIPE_PANEL_FITTER(crtc->pipe))) {
+		if (DISPLAY_VER(dev_priv) >= 9)
+			skl_get_pfit_config(pipe_config);
+		else
+			ilk_get_pfit_config(pipe_config);
+	}
+
+	hsw_ips_get_config(pipe_config);
+
+	if (pipe_config->cpu_transcoder != TRANSCODER_EDP &&
+	    !transcoder_is_dsi(pipe_config->cpu_transcoder)) {
+		pipe_config->pixel_multiplier =
+			intel_de_read(dev_priv,
+				      PIPE_MULT(pipe_config->cpu_transcoder)) + 1;
+	} else {
+		pipe_config->pixel_multiplier = 1;
+	}
+
+	if (!transcoder_is_dsi(pipe_config->cpu_transcoder)) {
+		tmp = intel_de_read(dev_priv, DISPLAY_VER(dev_priv) >= 14 ?
+				    MTL_CHICKEN_TRANS(pipe_config->cpu_transcoder) :
+				    CHICKEN_TRANS(pipe_config->cpu_transcoder));
+
+		pipe_config->framestart_delay = REG_FIELD_GET(HSW_FRAME_START_DELAY_MASK, tmp) + 1;
+	} else {
+		/* no idea if this is correct */
+		pipe_config->framestart_delay = 1;
+	}
+
+out:
+	intel_display_power_put_all_in_set(dev_priv, &power_domain_set);
+
+	return active;
+}
+
+bool intel_crtc_get_pipe_config(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	if (!i915->display.funcs.display->get_pipe_config(crtc, crtc_state))
+		return false;
+
+	crtc_state->hw.active = true;
+
+	intel_crtc_readout_derived_state(crtc_state);
+
+	return true;
+}
+
+/* VESA 640x480x72Hz mode to set on the pipe */
+static const struct drm_display_mode load_detect_mode = {
+	DRM_MODE("640x480", DRM_MODE_TYPE_DEFAULT, 31500, 640, 664,
+		 704, 832, 0, 480, 489, 491, 520, 0, DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC),
+};
+
+static int intel_modeset_disable_planes(struct drm_atomic_state *state,
+					struct drm_crtc *crtc)
+{
+	struct drm_plane *plane;
+	struct drm_plane_state *plane_state;
+	int ret, i;
+
+	ret = drm_atomic_add_affected_planes(state, crtc);
+	if (ret)
+		return ret;
+
+	for_each_new_plane_in_state(state, plane, plane_state, i) {
+		if (plane_state->crtc != crtc)
+			continue;
+
+		ret = drm_atomic_set_crtc_for_plane(plane_state, NULL);
+		if (ret)
+			return ret;
+
+		drm_atomic_set_fb_for_plane(plane_state, NULL);
+	}
+
+	return 0;
+}
+
+int intel_get_load_detect_pipe(struct drm_connector *connector,
+			       struct intel_load_detect_pipe *old,
+			       struct drm_modeset_acquire_ctx *ctx)
+{
+	struct intel_encoder *encoder =
+		intel_attached_encoder(to_intel_connector(connector));
+	struct intel_crtc *possible_crtc;
+	struct intel_crtc *crtc = NULL;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_mode_config *config = &dev->mode_config;
+	struct drm_atomic_state *state = NULL, *restore_state = NULL;
+	struct drm_connector_state *connector_state;
+	struct intel_crtc_state *crtc_state;
+	int ret;
+
+	drm_dbg_kms(&dev_priv->drm, "[CONNECTOR:%d:%s], [ENCODER:%d:%s]\n",
+		    connector->base.id, connector->name,
+		    encoder->base.base.id, encoder->base.name);
+
+	old->restore_state = NULL;
+
+	drm_WARN_ON(dev, !drm_modeset_is_locked(&config->connection_mutex));
+
+	/*
+	 * Algorithm gets a little messy:
+	 *
+	 *   - if the connector already has an assigned crtc, use it (but make
+	 *     sure it's on first)
+	 *
+	 *   - try to find the first unused crtc that can drive this connector,
+	 *     and use that if we find one
+	 */
+
+	/* See if we already have a CRTC for this connector */
+	if (connector->state->crtc) {
+		crtc = to_intel_crtc(connector->state->crtc);
+
+		ret = drm_modeset_lock(&crtc->base.mutex, ctx);
+		if (ret)
+			goto fail;
+
+		/* Make sure the crtc and connector are running */
+		goto found;
+	}
+
+	/* Find an unused one (if possible) */
+	for_each_intel_crtc(dev, possible_crtc) {
+		if (!(encoder->base.possible_crtcs &
+		      drm_crtc_mask(&possible_crtc->base)))
+			continue;
+
+		ret = drm_modeset_lock(&possible_crtc->base.mutex, ctx);
+		if (ret)
+			goto fail;
+
+		if (possible_crtc->base.state->enable) {
+			drm_modeset_unlock(&possible_crtc->base.mutex);
+			continue;
+		}
+
+		crtc = possible_crtc;
+		break;
+	}
+
+	/*
+	 * If we didn't find an unused CRTC, don't use any.
+	 */
+	if (!crtc) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "no pipe available for load-detect\n");
+		ret = -ENODEV;
+		goto fail;
+	}
+
+found:
+	state = drm_atomic_state_alloc(dev);
+	restore_state = drm_atomic_state_alloc(dev);
+	if (!state || !restore_state) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	state->acquire_ctx = ctx;
+	restore_state->acquire_ctx = ctx;
+
+	connector_state = drm_atomic_get_connector_state(state, connector);
+	if (IS_ERR(connector_state)) {
+		ret = PTR_ERR(connector_state);
+		goto fail;
+	}
+
+	ret = drm_atomic_set_crtc_for_connector(connector_state, &crtc->base);
+	if (ret)
+		goto fail;
+
+	crtc_state = intel_atomic_get_crtc_state(state, crtc);
+	if (IS_ERR(crtc_state)) {
+		ret = PTR_ERR(crtc_state);
+		goto fail;
+	}
+
+	crtc_state->uapi.active = true;
+
+	ret = drm_atomic_set_mode_for_crtc(&crtc_state->uapi,
+					   &load_detect_mode);
+	if (ret)
+		goto fail;
+
+	ret = intel_modeset_disable_planes(state, &crtc->base);
+	if (ret)
+		goto fail;
+
+	ret = PTR_ERR_OR_ZERO(drm_atomic_get_connector_state(restore_state, connector));
+	if (!ret)
+		ret = PTR_ERR_OR_ZERO(drm_atomic_get_crtc_state(restore_state, &crtc->base));
+	if (!ret)
+		ret = drm_atomic_add_affected_planes(restore_state, &crtc->base);
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Failed to create a copy of old state to restore: %i\n",
+			    ret);
+		goto fail;
+	}
+
+	ret = drm_atomic_commit(state);
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "failed to set mode on load-detect pipe\n");
+		goto fail;
+	}
+
+	old->restore_state = restore_state;
+	drm_atomic_state_put(state);
+
+	/* let the connector get through one full cycle before testing */
+	intel_crtc_wait_for_next_vblank(crtc);
+
+	return true;
+
+fail:
+	if (state) {
+		drm_atomic_state_put(state);
+		state = NULL;
+	}
+	if (restore_state) {
+		drm_atomic_state_put(restore_state);
+		restore_state = NULL;
+	}
+
+	if (ret == -EDEADLK)
+		return ret;
+
+	return false;
+}
+
+void intel_release_load_detect_pipe(struct drm_connector *connector,
+				    struct intel_load_detect_pipe *old,
+				    struct drm_modeset_acquire_ctx *ctx)
+{
+	struct intel_encoder *intel_encoder =
+		intel_attached_encoder(to_intel_connector(connector));
+	struct drm_i915_private *i915 = to_i915(intel_encoder->base.dev);
+	struct drm_encoder *encoder = &intel_encoder->base;
+	struct drm_atomic_state *state = old->restore_state;
+	int ret;
+
+	drm_dbg_kms(&i915->drm, "[CONNECTOR:%d:%s], [ENCODER:%d:%s]\n",
+		    connector->base.id, connector->name,
+		    encoder->base.id, encoder->name);
+
+	if (!state)
+		return;
+
+	ret = drm_atomic_helper_commit_duplicated_state(state, ctx);
+	if (ret)
+		drm_dbg_kms(&i915->drm,
+			    "Couldn't release load detect pipe: %i\n", ret);
+	drm_atomic_state_put(state);
+}
+
+static int i9xx_pll_refclk(struct drm_device *dev,
+			   const struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 dpll = pipe_config->dpll_hw_state.dpll;
+
+	if ((dpll & PLL_REF_INPUT_MASK) == PLLB_REF_INPUT_SPREADSPECTRUMIN)
+		return dev_priv->display.vbt.lvds_ssc_freq;
+	else if (HAS_PCH_SPLIT(dev_priv))
+		return 120000;
+	else if (DISPLAY_VER(dev_priv) != 2)
+		return 96000;
+	else
+		return 48000;
+}
+
+/* Returns the clock of the currently programmed mode of the given pipe. */
+void i9xx_crtc_clock_get(struct intel_crtc *crtc,
+			 struct intel_crtc_state *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 dpll = pipe_config->dpll_hw_state.dpll;
+	u32 fp;
+	struct dpll clock;
+	int port_clock;
+	int refclk = i9xx_pll_refclk(dev, pipe_config);
+
+	if ((dpll & DISPLAY_RATE_SELECT_FPA1) == 0)
+		fp = pipe_config->dpll_hw_state.fp0;
+	else
+		fp = pipe_config->dpll_hw_state.fp1;
+
+	clock.m1 = (fp & FP_M1_DIV_MASK) >> FP_M1_DIV_SHIFT;
+	if (IS_PINEVIEW(dev_priv)) {
+		clock.n = ffs((fp & FP_N_PINEVIEW_DIV_MASK) >> FP_N_DIV_SHIFT) - 1;
+		clock.m2 = (fp & FP_M2_PINEVIEW_DIV_MASK) >> FP_M2_DIV_SHIFT;
+	} else {
+		clock.n = (fp & FP_N_DIV_MASK) >> FP_N_DIV_SHIFT;
+		clock.m2 = (fp & FP_M2_DIV_MASK) >> FP_M2_DIV_SHIFT;
+	}
+
+	if (DISPLAY_VER(dev_priv) != 2) {
+		if (IS_PINEVIEW(dev_priv))
+			clock.p1 = ffs((dpll & DPLL_FPA01_P1_POST_DIV_MASK_PINEVIEW) >>
+				DPLL_FPA01_P1_POST_DIV_SHIFT_PINEVIEW);
+		else
+			clock.p1 = ffs((dpll & DPLL_FPA01_P1_POST_DIV_MASK) >>
+			       DPLL_FPA01_P1_POST_DIV_SHIFT);
+
+		switch (dpll & DPLL_MODE_MASK) {
+		case DPLLB_MODE_DAC_SERIAL:
+			clock.p2 = dpll & DPLL_DAC_SERIAL_P2_CLOCK_DIV_5 ?
+				5 : 10;
+			break;
+		case DPLLB_MODE_LVDS:
+			clock.p2 = dpll & DPLLB_LVDS_P2_CLOCK_DIV_7 ?
+				7 : 14;
+			break;
+		default:
+			drm_dbg_kms(&dev_priv->drm,
+				    "Unknown DPLL mode %08x in programmed "
+				    "mode\n", (int)(dpll & DPLL_MODE_MASK));
+			return;
+		}
+
+		if (IS_PINEVIEW(dev_priv))
+			port_clock = pnv_calc_dpll_params(refclk, &clock);
+		else
+			port_clock = i9xx_calc_dpll_params(refclk, &clock);
+	} else {
+		enum pipe lvds_pipe;
+
+		if (IS_I85X(dev_priv) &&
+		    intel_lvds_port_enabled(dev_priv, LVDS, &lvds_pipe) &&
+		    lvds_pipe == crtc->pipe) {
+			u32 lvds = intel_de_read(dev_priv, LVDS);
+
+			clock.p1 = ffs((dpll & DPLL_FPA01_P1_POST_DIV_MASK_I830_LVDS) >>
+				       DPLL_FPA01_P1_POST_DIV_SHIFT);
+
+			if (lvds & LVDS_CLKB_POWER_UP)
+				clock.p2 = 7;
+			else
+				clock.p2 = 14;
+		} else {
+			if (dpll & PLL_P1_DIVIDE_BY_TWO)
+				clock.p1 = 2;
+			else {
+				clock.p1 = ((dpll & DPLL_FPA01_P1_POST_DIV_MASK_I830) >>
+					    DPLL_FPA01_P1_POST_DIV_SHIFT) + 2;
+			}
+			if (dpll & PLL_P2_DIVIDE_BY_4)
+				clock.p2 = 4;
+			else
+				clock.p2 = 2;
+		}
+
+		port_clock = i9xx_calc_dpll_params(refclk, &clock);
+	}
+
+	/*
+	 * This value includes pixel_multiplier. We will use
+	 * port_clock to compute adjusted_mode.crtc_clock in the
+	 * encoder's get_config() function.
+	 */
+	pipe_config->port_clock = port_clock;
+}
+
+int intel_dotclock_calculate(int link_freq,
+			     const struct intel_link_m_n *m_n)
+{
+	/*
+	 * The calculation for the data clock is:
+	 * pixel_clock = ((m/n)*(link_clock * nr_lanes))/bpp
+	 * But we want to avoid losing precison if possible, so:
+	 * pixel_clock = ((m * link_clock * nr_lanes)/(n*bpp))
+	 *
+	 * and the link clock is simpler:
+	 * link_clock = (m * link_clock) / n
+	 */
+
+	if (!m_n->link_n)
+		return 0;
+
+	return DIV_ROUND_UP_ULL(mul_u32_u32(m_n->link_m, link_freq),
+				m_n->link_n);
+}
+
+int intel_crtc_dotclock(const struct intel_crtc_state *pipe_config)
+{
+	int dotclock;
+
+	if (intel_crtc_has_dp_encoder(pipe_config))
+		dotclock = intel_dotclock_calculate(pipe_config->port_clock,
+						    &pipe_config->dp_m_n);
+	else if (pipe_config->has_hdmi_sink && pipe_config->pipe_bpp > 24)
+		dotclock = DIV_ROUND_CLOSEST(pipe_config->port_clock * 24,
+					     pipe_config->pipe_bpp);
+	else
+		dotclock = pipe_config->port_clock;
+
+	if (pipe_config->output_format == INTEL_OUTPUT_FORMAT_YCBCR420 &&
+	    !intel_crtc_has_dp_encoder(pipe_config))
+		dotclock *= 2;
+
+	if (pipe_config->pixel_multiplier)
+		dotclock /= pipe_config->pixel_multiplier;
+
+	return dotclock;
+}
+
+/* Returns the currently programmed mode of the given encoder. */
+struct drm_display_mode *
+intel_encoder_current_mode(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc_state *crtc_state;
+	struct drm_display_mode *mode;
+	struct intel_crtc *crtc;
+	enum pipe pipe;
+
+	if (!encoder->get_hw_state(encoder, &pipe))
+		return NULL;
+
+	crtc = intel_crtc_for_pipe(dev_priv, pipe);
+
+	mode = kzalloc(sizeof(*mode), GFP_KERNEL);
+	if (!mode)
+		return NULL;
+
+	crtc_state = intel_crtc_state_alloc(crtc);
+	if (!crtc_state) {
+		kfree(mode);
+		return NULL;
+	}
+
+	if (!intel_crtc_get_pipe_config(crtc_state)) {
+		kfree(crtc_state);
+		kfree(mode);
+		return NULL;
+	}
+
+	intel_encoder_get_config(encoder, crtc_state);
+
+	intel_mode_from_crtc_timings(mode, &crtc_state->hw.adjusted_mode);
+
+	kfree(crtc_state);
+
+	return mode;
+}
+
+static bool encoders_cloneable(const struct intel_encoder *a,
+			       const struct intel_encoder *b)
+{
+	/* masks could be asymmetric, so check both ways */
+	return a == b || (a->cloneable & (1 << b->type) &&
+			  b->cloneable & (1 << a->type));
+}
+
+static bool check_single_encoder_cloning(struct intel_atomic_state *state,
+					 struct intel_crtc *crtc,
+					 struct intel_encoder *encoder)
+{
+	struct intel_encoder *source_encoder;
+	struct drm_connector *connector;
+	struct drm_connector_state *connector_state;
+	int i;
+
+	for_each_new_connector_in_state(&state->base, connector, connector_state, i) {
+		if (connector_state->crtc != &crtc->base)
+			continue;
+
+		source_encoder =
+			to_intel_encoder(connector_state->best_encoder);
+		if (!encoders_cloneable(encoder, source_encoder))
+			return false;
+	}
+
+	return true;
+}
+
+static int icl_add_linked_planes(struct intel_atomic_state *state)
+{
+	struct intel_plane *plane, *linked;
+	struct intel_plane_state *plane_state, *linked_plane_state;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		linked = plane_state->planar_linked_plane;
+
+		if (!linked)
+			continue;
+
+		linked_plane_state = intel_atomic_get_plane_state(state, linked);
+		if (IS_ERR(linked_plane_state))
+			return PTR_ERR(linked_plane_state);
+
+		drm_WARN_ON(state->base.dev,
+			    linked_plane_state->planar_linked_plane != plane);
+		drm_WARN_ON(state->base.dev,
+			    linked_plane_state->planar_slave == plane_state->planar_slave);
+	}
+
+	return 0;
+}
+
+static int icl_check_nv12_planes(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_atomic_state *state = to_intel_atomic_state(crtc_state->uapi.state);
+	struct intel_plane *plane, *linked;
+	struct intel_plane_state *plane_state;
+	int i;
+
+	if (DISPLAY_VER(dev_priv) < 11)
+		return 0;
+
+	/*
+	 * Destroy all old plane links and make the slave plane invisible
+	 * in the crtc_state->active_planes mask.
+	 */
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		if (plane->pipe != crtc->pipe || !plane_state->planar_linked_plane)
+			continue;
+
+		plane_state->planar_linked_plane = NULL;
+		if (plane_state->planar_slave && !plane_state->uapi.visible) {
+			crtc_state->enabled_planes &= ~BIT(plane->id);
+			crtc_state->active_planes &= ~BIT(plane->id);
+			crtc_state->update_planes |= BIT(plane->id);
+			crtc_state->data_rate[plane->id] = 0;
+			crtc_state->rel_data_rate[plane->id] = 0;
+		}
+
+		plane_state->planar_slave = false;
+	}
+
+	if (!crtc_state->nv12_planes)
+		return 0;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		struct intel_plane_state *linked_state = NULL;
+
+		if (plane->pipe != crtc->pipe ||
+		    !(crtc_state->nv12_planes & BIT(plane->id)))
+			continue;
+
+		for_each_intel_plane_on_crtc(&dev_priv->drm, crtc, linked) {
+			if (!icl_is_nv12_y_plane(dev_priv, linked->id))
+				continue;
+
+			if (crtc_state->active_planes & BIT(linked->id))
+				continue;
+
+			linked_state = intel_atomic_get_plane_state(state, linked);
+			if (IS_ERR(linked_state))
+				return PTR_ERR(linked_state);
+
+			break;
+		}
+
+		if (!linked_state) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Need %d free Y planes for planar YUV\n",
+				    hweight8(crtc_state->nv12_planes));
+
+			return -EINVAL;
+		}
+
+		plane_state->planar_linked_plane = linked;
+
+		linked_state->planar_slave = true;
+		linked_state->planar_linked_plane = plane;
+		crtc_state->enabled_planes |= BIT(linked->id);
+		crtc_state->active_planes |= BIT(linked->id);
+		crtc_state->update_planes |= BIT(linked->id);
+		crtc_state->data_rate[linked->id] =
+			crtc_state->data_rate_y[plane->id];
+		crtc_state->rel_data_rate[linked->id] =
+			crtc_state->rel_data_rate_y[plane->id];
+		drm_dbg_kms(&dev_priv->drm, "Using %s as Y plane for %s\n",
+			    linked->base.name, plane->base.name);
+
+		/* Copy parameters to slave plane */
+		linked_state->ctl = plane_state->ctl | PLANE_CTL_YUV420_Y_PLANE;
+		linked_state->color_ctl = plane_state->color_ctl;
+		linked_state->view = plane_state->view;
+		linked_state->decrypt = plane_state->decrypt;
+
+		intel_plane_copy_hw_state(linked_state, plane_state);
+		linked_state->uapi.src = plane_state->uapi.src;
+		linked_state->uapi.dst = plane_state->uapi.dst;
+
+		if (icl_is_hdr_plane(dev_priv, plane->id)) {
+			if (linked->id == PLANE_SPRITE5)
+				plane_state->cus_ctl |= PLANE_CUS_Y_PLANE_7_ICL;
+			else if (linked->id == PLANE_SPRITE4)
+				plane_state->cus_ctl |= PLANE_CUS_Y_PLANE_6_ICL;
+			else if (linked->id == PLANE_SPRITE3)
+				plane_state->cus_ctl |= PLANE_CUS_Y_PLANE_5_RKL;
+			else if (linked->id == PLANE_SPRITE2)
+				plane_state->cus_ctl |= PLANE_CUS_Y_PLANE_4_RKL;
+			else
+				MISSING_CASE(linked->id);
+		}
+	}
+
+	return 0;
+}
+
+static bool c8_planes_changed(const struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(new_crtc_state->uapi.state);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+
+	return !old_crtc_state->c8_planes != !new_crtc_state->c8_planes;
+}
+
+static u16 hsw_linetime_wm(const struct intel_crtc_state *crtc_state)
+{
+	const struct drm_display_mode *pipe_mode =
+		&crtc_state->hw.pipe_mode;
+	int linetime_wm;
+
+	if (!crtc_state->hw.enable)
+		return 0;
+
+	linetime_wm = DIV_ROUND_CLOSEST(pipe_mode->crtc_htotal * 1000 * 8,
+					pipe_mode->crtc_clock);
+
+	return min(linetime_wm, 0x1ff);
+}
+
+static u16 hsw_ips_linetime_wm(const struct intel_crtc_state *crtc_state,
+			       const struct intel_cdclk_state *cdclk_state)
+{
+	const struct drm_display_mode *pipe_mode =
+		&crtc_state->hw.pipe_mode;
+	int linetime_wm;
+
+	if (!crtc_state->hw.enable)
+		return 0;
+
+	linetime_wm = DIV_ROUND_CLOSEST(pipe_mode->crtc_htotal * 1000 * 8,
+					cdclk_state->logical.cdclk);
+
+	return min(linetime_wm, 0x1ff);
+}
+
+static u16 skl_linetime_wm(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_display_mode *pipe_mode =
+		&crtc_state->hw.pipe_mode;
+	int linetime_wm;
+
+	if (!crtc_state->hw.enable)
+		return 0;
+
+	linetime_wm = DIV_ROUND_UP(pipe_mode->crtc_htotal * 1000 * 8,
+				   crtc_state->pixel_rate);
+
+	/* Display WA #1135: BXT:ALL GLK:ALL */
+	if ((IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) &&
+	    skl_watermark_ipc_enabled(dev_priv))
+		linetime_wm /= 2;
+
+	return min(linetime_wm, 0x1ff);
+}
+
+static int hsw_compute_linetime_wm(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_cdclk_state *cdclk_state;
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		crtc_state->linetime = skl_linetime_wm(crtc_state);
+	else
+		crtc_state->linetime = hsw_linetime_wm(crtc_state);
+
+	if (!hsw_crtc_supports_ips(crtc))
+		return 0;
+
+	cdclk_state = intel_atomic_get_cdclk_state(state);
+	if (IS_ERR(cdclk_state))
+		return PTR_ERR(cdclk_state);
+
+	crtc_state->ips_linetime = hsw_ips_linetime_wm(crtc_state,
+						       cdclk_state);
+
+	return 0;
+}
+
+static int intel_crtc_atomic_check(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	bool mode_changed = intel_crtc_needs_modeset(crtc_state);
+	int ret;
+
+	if (DISPLAY_VER(dev_priv) < 5 && !IS_G4X(dev_priv) &&
+	    mode_changed && !crtc_state->hw.active)
+		crtc_state->update_wm_post = true;
+
+	if (mode_changed) {
+		ret = intel_dpll_crtc_get_shared_dpll(state, crtc);
+		if (ret)
+			return ret;
+	}
+
+	/*
+	 * May need to update pipe gamma enable bits
+	 * when C8 planes are getting enabled/disabled.
+	 */
+	if (c8_planes_changed(crtc_state))
+		crtc_state->uapi.color_mgmt_changed = true;
+
+	if (mode_changed || crtc_state->update_pipe ||
+	    crtc_state->uapi.color_mgmt_changed) {
+		ret = intel_color_check(crtc_state);
+		if (ret)
+			return ret;
+	}
+
+	ret = intel_compute_pipe_wm(state, crtc);
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Target pipe watermarks are invalid\n");
+		return ret;
+	}
+
+	/*
+	 * Calculate 'intermediate' watermarks that satisfy both the
+	 * old state and the new state.  We can program these
+	 * immediately.
+	 */
+	ret = intel_compute_intermediate_wm(state, crtc);
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "No valid intermediate pipe watermarks are possible\n");
+		return ret;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 9) {
+		if (mode_changed || crtc_state->update_pipe) {
+			ret = skl_update_scaler_crtc(crtc_state);
+			if (ret)
+				return ret;
+		}
+
+		ret = intel_atomic_setup_scalers(dev_priv, crtc, crtc_state);
+		if (ret)
+			return ret;
+	}
+
+	if (HAS_IPS(dev_priv)) {
+		ret = hsw_ips_compute_config(state, crtc);
+		if (ret)
+			return ret;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 9 ||
+	    IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv)) {
+		ret = hsw_compute_linetime_wm(state, crtc);
+		if (ret)
+			return ret;
+
+	}
+
+	ret = intel_psr2_sel_fetch_update(state, crtc);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int
+compute_sink_pipe_bpp(const struct drm_connector_state *conn_state,
+		      struct intel_crtc_state *crtc_state)
+{
+	struct drm_connector *connector = conn_state->connector;
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	const struct drm_display_info *info = &connector->display_info;
+	int bpp;
+
+	switch (conn_state->max_bpc) {
+	case 6 ... 7:
+		bpp = 6 * 3;
+		break;
+	case 8 ... 9:
+		bpp = 8 * 3;
+		break;
+	case 10 ... 11:
+		bpp = 10 * 3;
+		break;
+	case 12 ... 16:
+		bpp = 12 * 3;
+		break;
+	default:
+		MISSING_CASE(conn_state->max_bpc);
+		return -EINVAL;
+	}
+
+	if (bpp < crtc_state->pipe_bpp) {
+		drm_dbg_kms(&i915->drm,
+			    "[CONNECTOR:%d:%s] Limiting display bpp to %d "
+			    "(EDID bpp %d, max requested bpp %d, max platform bpp %d)\n",
+			    connector->base.id, connector->name,
+			    bpp, 3 * info->bpc,
+			    3 * conn_state->max_requested_bpc,
+			    crtc_state->pipe_bpp);
+
+		crtc_state->pipe_bpp = bpp;
+	}
+
+	return 0;
+}
+
+static int
+compute_baseline_pipe_bpp(struct intel_atomic_state *state,
+			  struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_connector *connector;
+	struct drm_connector_state *connector_state;
+	int bpp, i;
+
+	if ((IS_G4X(dev_priv) || IS_VALLEYVIEW(dev_priv) ||
+	    IS_CHERRYVIEW(dev_priv)))
+		bpp = 10*3;
+	else if (DISPLAY_VER(dev_priv) >= 5)
+		bpp = 12*3;
+	else
+		bpp = 8*3;
+
+	crtc_state->pipe_bpp = bpp;
+
+	/* Clamp display bpp to connector max bpp */
+	for_each_new_connector_in_state(&state->base, connector, connector_state, i) {
+		int ret;
+
+		if (connector_state->crtc != &crtc->base)
+			continue;
+
+		ret = compute_sink_pipe_bpp(connector_state, crtc_state);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static bool check_digital_port_conflicts(struct intel_atomic_state *state)
+{
+	struct drm_device *dev = state->base.dev;
+	struct drm_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	unsigned int used_ports = 0;
+	unsigned int used_mst_ports = 0;
+	bool ret = true;
+
+	/*
+	 * We're going to peek into connector->state,
+	 * hence connection_mutex must be held.
+	 */
+	drm_modeset_lock_assert_held(&dev->mode_config.connection_mutex);
+
+	/*
+	 * Walk the connector list instead of the encoder
+	 * list to detect the problem on ddi platforms
+	 * where there's just one encoder per digital port.
+	 */
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		struct drm_connector_state *connector_state;
+		struct intel_encoder *encoder;
+
+		connector_state =
+			drm_atomic_get_new_connector_state(&state->base,
+							   connector);
+		if (!connector_state)
+			connector_state = connector->state;
+
+		if (!connector_state->best_encoder)
+			continue;
+
+		encoder = to_intel_encoder(connector_state->best_encoder);
+
+		drm_WARN_ON(dev, !connector_state->crtc);
+
+		switch (encoder->type) {
+		case INTEL_OUTPUT_DDI:
+			if (drm_WARN_ON(dev, !HAS_DDI(to_i915(dev))))
+				break;
+			fallthrough;
+		case INTEL_OUTPUT_DP:
+		case INTEL_OUTPUT_HDMI:
+		case INTEL_OUTPUT_EDP:
+			/* the same port mustn't appear more than once */
+			if (used_ports & BIT(encoder->port))
+				ret = false;
+
+			used_ports |= BIT(encoder->port);
+			break;
+		case INTEL_OUTPUT_DP_MST:
+			used_mst_ports |=
+				1 << encoder->port;
+			break;
+		default:
+			break;
+		}
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	/* can't mix MST and SST/HDMI on the same port */
+	if (used_ports & used_mst_ports)
+		return false;
+
+	return ret;
+}
+
+static void
+intel_crtc_copy_uapi_to_hw_state_nomodeset(struct intel_atomic_state *state,
+					   struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	WARN_ON(intel_crtc_is_bigjoiner_slave(crtc_state));
+
+	drm_property_replace_blob(&crtc_state->hw.degamma_lut,
+				  crtc_state->uapi.degamma_lut);
+	drm_property_replace_blob(&crtc_state->hw.gamma_lut,
+				  crtc_state->uapi.gamma_lut);
+	drm_property_replace_blob(&crtc_state->hw.ctm,
+				  crtc_state->uapi.ctm);
+}
+
+static void
+intel_crtc_copy_uapi_to_hw_state_modeset(struct intel_atomic_state *state,
+					 struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	WARN_ON(intel_crtc_is_bigjoiner_slave(crtc_state));
+
+	crtc_state->hw.enable = crtc_state->uapi.enable;
+	crtc_state->hw.active = crtc_state->uapi.active;
+	drm_mode_copy(&crtc_state->hw.mode,
+		      &crtc_state->uapi.mode);
+	drm_mode_copy(&crtc_state->hw.adjusted_mode,
+		      &crtc_state->uapi.adjusted_mode);
+	crtc_state->hw.scaling_filter = crtc_state->uapi.scaling_filter;
+
+	intel_crtc_copy_uapi_to_hw_state_nomodeset(state, crtc);
+}
+
+static void
+copy_bigjoiner_crtc_state_nomodeset(struct intel_atomic_state *state,
+				    struct intel_crtc *slave_crtc)
+{
+	struct intel_crtc_state *slave_crtc_state =
+		intel_atomic_get_new_crtc_state(state, slave_crtc);
+	struct intel_crtc *master_crtc = intel_master_crtc(slave_crtc_state);
+	const struct intel_crtc_state *master_crtc_state =
+		intel_atomic_get_new_crtc_state(state, master_crtc);
+
+	drm_property_replace_blob(&slave_crtc_state->hw.degamma_lut,
+				  master_crtc_state->hw.degamma_lut);
+	drm_property_replace_blob(&slave_crtc_state->hw.gamma_lut,
+				  master_crtc_state->hw.gamma_lut);
+	drm_property_replace_blob(&slave_crtc_state->hw.ctm,
+				  master_crtc_state->hw.ctm);
+
+	slave_crtc_state->uapi.color_mgmt_changed = master_crtc_state->uapi.color_mgmt_changed;
+}
+
+static int
+copy_bigjoiner_crtc_state_modeset(struct intel_atomic_state *state,
+				  struct intel_crtc *slave_crtc)
+{
+	struct intel_crtc_state *slave_crtc_state =
+		intel_atomic_get_new_crtc_state(state, slave_crtc);
+	struct intel_crtc *master_crtc = intel_master_crtc(slave_crtc_state);
+	const struct intel_crtc_state *master_crtc_state =
+		intel_atomic_get_new_crtc_state(state, master_crtc);
+	struct intel_crtc_state *saved_state;
+
+	WARN_ON(master_crtc_state->bigjoiner_pipes !=
+		slave_crtc_state->bigjoiner_pipes);
+
+	saved_state = kmemdup(master_crtc_state, sizeof(*saved_state), GFP_KERNEL);
+	if (!saved_state)
+		return -ENOMEM;
+
+	/* preserve some things from the slave's original crtc state */
+	saved_state->uapi = slave_crtc_state->uapi;
+	saved_state->scaler_state = slave_crtc_state->scaler_state;
+	saved_state->shared_dpll = slave_crtc_state->shared_dpll;
+	saved_state->crc_enabled = slave_crtc_state->crc_enabled;
+
+	intel_crtc_free_hw_state(slave_crtc_state);
+	memcpy(slave_crtc_state, saved_state, sizeof(*slave_crtc_state));
+	kfree(saved_state);
+
+	/* Re-init hw state */
+	memset(&slave_crtc_state->hw, 0, sizeof(slave_crtc_state->hw));
+	slave_crtc_state->hw.enable = master_crtc_state->hw.enable;
+	slave_crtc_state->hw.active = master_crtc_state->hw.active;
+	drm_mode_copy(&slave_crtc_state->hw.mode,
+		      &master_crtc_state->hw.mode);
+	drm_mode_copy(&slave_crtc_state->hw.pipe_mode,
+		      &master_crtc_state->hw.pipe_mode);
+	drm_mode_copy(&slave_crtc_state->hw.adjusted_mode,
+		      &master_crtc_state->hw.adjusted_mode);
+	slave_crtc_state->hw.scaling_filter = master_crtc_state->hw.scaling_filter;
+
+	copy_bigjoiner_crtc_state_nomodeset(state, slave_crtc);
+
+	slave_crtc_state->uapi.mode_changed = master_crtc_state->uapi.mode_changed;
+	slave_crtc_state->uapi.connectors_changed = master_crtc_state->uapi.connectors_changed;
+	slave_crtc_state->uapi.active_changed = master_crtc_state->uapi.active_changed;
+
+	WARN_ON(master_crtc_state->bigjoiner_pipes !=
+		slave_crtc_state->bigjoiner_pipes);
+
+	return 0;
+}
+
+static int
+intel_crtc_prepare_cleared_state(struct intel_atomic_state *state,
+				 struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *saved_state;
+
+	saved_state = intel_crtc_state_alloc(crtc);
+	if (!saved_state)
+		return -ENOMEM;
+
+	/* free the old crtc_state->hw members */
+	intel_crtc_free_hw_state(crtc_state);
+
+	/* FIXME: before the switch to atomic started, a new pipe_config was
+	 * kzalloc'd. Code that depends on any field being zero should be
+	 * fixed, so that the crtc_state can be safely duplicated. For now,
+	 * only fields that are know to not cause problems are preserved. */
+
+	saved_state->uapi = crtc_state->uapi;
+	saved_state->inherited = crtc_state->inherited;
+	saved_state->scaler_state = crtc_state->scaler_state;
+	saved_state->shared_dpll = crtc_state->shared_dpll;
+	saved_state->dpll_hw_state = crtc_state->dpll_hw_state;
+	memcpy(saved_state->icl_port_dplls, crtc_state->icl_port_dplls,
+	       sizeof(saved_state->icl_port_dplls));
+	saved_state->crc_enabled = crtc_state->crc_enabled;
+	if (IS_G4X(dev_priv) ||
+	    IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		saved_state->wm = crtc_state->wm;
+
+	memcpy(crtc_state, saved_state, sizeof(*crtc_state));
+	kfree(saved_state);
+
+	intel_crtc_copy_uapi_to_hw_state_modeset(state, crtc);
+
+	return 0;
+}
+
+static int
+intel_modeset_pipe_config(struct intel_atomic_state *state,
+			  struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_connector *connector;
+	struct drm_connector_state *connector_state;
+	int pipe_src_w, pipe_src_h;
+	int base_bpp, ret, i;
+	bool retry = true;
+
+	crtc_state->cpu_transcoder = (enum transcoder) crtc->pipe;
+
+	crtc_state->framestart_delay = 1;
+
+	/*
+	 * Sanitize sync polarity flags based on requested ones. If neither
+	 * positive or negative polarity is requested, treat this as meaning
+	 * negative polarity.
+	 */
+	if (!(crtc_state->hw.adjusted_mode.flags &
+	      (DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NHSYNC)))
+		crtc_state->hw.adjusted_mode.flags |= DRM_MODE_FLAG_NHSYNC;
+
+	if (!(crtc_state->hw.adjusted_mode.flags &
+	      (DRM_MODE_FLAG_PVSYNC | DRM_MODE_FLAG_NVSYNC)))
+		crtc_state->hw.adjusted_mode.flags |= DRM_MODE_FLAG_NVSYNC;
+
+	ret = compute_baseline_pipe_bpp(state, crtc);
+	if (ret)
+		return ret;
+
+	base_bpp = crtc_state->pipe_bpp;
+
+	/*
+	 * Determine the real pipe dimensions. Note that stereo modes can
+	 * increase the actual pipe size due to the frame doubling and
+	 * insertion of additional space for blanks between the frame. This
+	 * is stored in the crtc timings. We use the requested mode to do this
+	 * computation to clearly distinguish it from the adjusted mode, which
+	 * can be changed by the connectors in the below retry loop.
+	 */
+	drm_mode_get_hv_timing(&crtc_state->hw.mode,
+			       &pipe_src_w, &pipe_src_h);
+	drm_rect_init(&crtc_state->pipe_src, 0, 0,
+		      pipe_src_w, pipe_src_h);
+
+	for_each_new_connector_in_state(&state->base, connector, connector_state, i) {
+		struct intel_encoder *encoder =
+			to_intel_encoder(connector_state->best_encoder);
+
+		if (connector_state->crtc != &crtc->base)
+			continue;
+
+		if (!check_single_encoder_cloning(state, crtc, encoder)) {
+			drm_dbg_kms(&i915->drm,
+				    "[ENCODER:%d:%s] rejecting invalid cloning configuration\n",
+				    encoder->base.base.id, encoder->base.name);
+			return -EINVAL;
+		}
+
+		/*
+		 * Determine output_types before calling the .compute_config()
+		 * hooks so that the hooks can use this information safely.
+		 */
+		if (encoder->compute_output_type)
+			crtc_state->output_types |=
+				BIT(encoder->compute_output_type(encoder, crtc_state,
+								 connector_state));
+		else
+			crtc_state->output_types |= BIT(encoder->type);
+	}
+
+encoder_retry:
+	/* Ensure the port clock defaults are reset when retrying. */
+	crtc_state->port_clock = 0;
+	crtc_state->pixel_multiplier = 1;
+
+	/* Fill in default crtc timings, allow encoders to overwrite them. */
+	drm_mode_set_crtcinfo(&crtc_state->hw.adjusted_mode,
+			      CRTC_STEREO_DOUBLE);
+
+	/* Pass our mode to the connectors and the CRTC to give them a chance to
+	 * adjust it according to limitations or connector properties, and also
+	 * a chance to reject the mode entirely.
+	 */
+	for_each_new_connector_in_state(&state->base, connector, connector_state, i) {
+		struct intel_encoder *encoder =
+			to_intel_encoder(connector_state->best_encoder);
+
+		if (connector_state->crtc != &crtc->base)
+			continue;
+
+		ret = encoder->compute_config(encoder, crtc_state,
+					      connector_state);
+		if (ret == -EDEADLK)
+			return ret;
+		if (ret < 0) {
+			drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s] config failure: %d\n",
+				    encoder->base.base.id, encoder->base.name, ret);
+			return ret;
+		}
+	}
+
+	/* Set default port clock if not overwritten by the encoder. Needs to be
+	 * done afterwards in case the encoder adjusts the mode. */
+	if (!crtc_state->port_clock)
+		crtc_state->port_clock = crtc_state->hw.adjusted_mode.crtc_clock
+			* crtc_state->pixel_multiplier;
+
+	ret = intel_crtc_compute_config(state, crtc);
+	if (ret == -EDEADLK)
+		return ret;
+	if (ret == -EAGAIN) {
+		if (drm_WARN(&i915->drm, !retry,
+			     "[CRTC:%d:%s] loop in pipe configuration computation\n",
+			     crtc->base.base.id, crtc->base.name))
+			return -EINVAL;
+
+		drm_dbg_kms(&i915->drm, "[CRTC:%d:%s] bw constrained, retrying\n",
+			    crtc->base.base.id, crtc->base.name);
+		retry = false;
+		goto encoder_retry;
+	}
+	if (ret < 0) {
+		drm_dbg_kms(&i915->drm, "[CRTC:%d:%s] config failure: %d\n",
+			    crtc->base.base.id, crtc->base.name, ret);
+		return ret;
+	}
+
+	/* Dithering seems to not pass-through bits correctly when it should, so
+	 * only enable it on 6bpc panels and when its not a compliance
+	 * test requesting 6bpc video pattern.
+	 */
+	crtc_state->dither = (crtc_state->pipe_bpp == 6*3) &&
+		!crtc_state->dither_force_disable;
+	drm_dbg_kms(&i915->drm,
+		    "[CRTC:%d:%s] hw max bpp: %i, pipe bpp: %i, dithering: %i\n",
+		    crtc->base.base.id, crtc->base.name,
+		    base_bpp, crtc_state->pipe_bpp, crtc_state->dither);
+
+	return 0;
+}
+
+static int
+intel_modeset_pipe_config_late(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_connector_state *conn_state;
+	struct drm_connector *connector;
+	int i;
+
+	intel_bigjoiner_adjust_pipe_src(crtc_state);
+
+	for_each_new_connector_in_state(&state->base, connector,
+					conn_state, i) {
+		struct intel_encoder *encoder =
+			to_intel_encoder(conn_state->best_encoder);
+		int ret;
+
+		if (conn_state->crtc != &crtc->base ||
+		    !encoder->compute_config_late)
+			continue;
+
+		ret = encoder->compute_config_late(encoder, crtc_state,
+						   conn_state);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+bool intel_fuzzy_clock_check(int clock1, int clock2)
+{
+	int diff;
+
+	if (clock1 == clock2)
+		return true;
+
+	if (!clock1 || !clock2)
+		return false;
+
+	diff = abs(clock1 - clock2);
+
+	if (((((diff + clock1 + clock2) * 100)) / (clock1 + clock2)) < 105)
+		return true;
+
+	return false;
+}
+
+static bool
+intel_compare_link_m_n(const struct intel_link_m_n *m_n,
+		       const struct intel_link_m_n *m2_n2)
+{
+	return m_n->tu == m2_n2->tu &&
+		m_n->data_m == m2_n2->data_m &&
+		m_n->data_n == m2_n2->data_n &&
+		m_n->link_m == m2_n2->link_m &&
+		m_n->link_n == m2_n2->link_n;
+}
+
+static bool
+intel_compare_infoframe(const union hdmi_infoframe *a,
+			const union hdmi_infoframe *b)
+{
+	return memcmp(a, b, sizeof(*a)) == 0;
+}
+
+static bool
+intel_compare_dp_vsc_sdp(const struct drm_dp_vsc_sdp *a,
+			 const struct drm_dp_vsc_sdp *b)
+{
+	return memcmp(a, b, sizeof(*a)) == 0;
+}
+
+static void
+pipe_config_infoframe_mismatch(struct drm_i915_private *dev_priv,
+			       bool fastset, const char *name,
+			       const union hdmi_infoframe *a,
+			       const union hdmi_infoframe *b)
+{
+	if (fastset) {
+		if (!drm_debug_enabled(DRM_UT_KMS))
+			return;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "fastset mismatch in %s infoframe\n", name);
+		drm_dbg_kms(&dev_priv->drm, "expected:\n");
+		hdmi_infoframe_log(KERN_DEBUG, dev_priv->drm.dev, a);
+		drm_dbg_kms(&dev_priv->drm, "found:\n");
+		hdmi_infoframe_log(KERN_DEBUG, dev_priv->drm.dev, b);
+	} else {
+		drm_err(&dev_priv->drm, "mismatch in %s infoframe\n", name);
+		drm_err(&dev_priv->drm, "expected:\n");
+		hdmi_infoframe_log(KERN_ERR, dev_priv->drm.dev, a);
+		drm_err(&dev_priv->drm, "found:\n");
+		hdmi_infoframe_log(KERN_ERR, dev_priv->drm.dev, b);
+	}
+}
+
+static void
+pipe_config_dp_vsc_sdp_mismatch(struct drm_i915_private *dev_priv,
+				bool fastset, const char *name,
+				const struct drm_dp_vsc_sdp *a,
+				const struct drm_dp_vsc_sdp *b)
+{
+	if (fastset) {
+		if (!drm_debug_enabled(DRM_UT_KMS))
+			return;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "fastset mismatch in %s dp sdp\n", name);
+		drm_dbg_kms(&dev_priv->drm, "expected:\n");
+		drm_dp_vsc_sdp_log(KERN_DEBUG, dev_priv->drm.dev, a);
+		drm_dbg_kms(&dev_priv->drm, "found:\n");
+		drm_dp_vsc_sdp_log(KERN_DEBUG, dev_priv->drm.dev, b);
+	} else {
+		drm_err(&dev_priv->drm, "mismatch in %s dp sdp\n", name);
+		drm_err(&dev_priv->drm, "expected:\n");
+		drm_dp_vsc_sdp_log(KERN_ERR, dev_priv->drm.dev, a);
+		drm_err(&dev_priv->drm, "found:\n");
+		drm_dp_vsc_sdp_log(KERN_ERR, dev_priv->drm.dev, b);
+	}
+}
+
+static void __printf(4, 5)
+pipe_config_mismatch(bool fastset, const struct intel_crtc *crtc,
+		     const char *name, const char *format, ...)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct va_format vaf;
+	va_list args;
+
+	va_start(args, format);
+	vaf.fmt = format;
+	vaf.va = &args;
+
+	if (fastset)
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] fastset mismatch in %s %pV\n",
+			    crtc->base.base.id, crtc->base.name, name, &vaf);
+	else
+		drm_err(&i915->drm, "[CRTC:%d:%s] mismatch in %s %pV\n",
+			crtc->base.base.id, crtc->base.name, name, &vaf);
+
+	va_end(args);
+}
+
+static bool fastboot_enabled(struct drm_i915_private *dev_priv)
+{
+	if (dev_priv->params.fastboot != -1)
+		return dev_priv->params.fastboot;
+
+	/* Enable fastboot by default on Skylake and newer */
+	if (DISPLAY_VER(dev_priv) >= 9)
+		return true;
+
+	/* Enable fastboot by default on VLV and CHV */
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		return true;
+
+	/* Disabled by default on all others */
+	return false;
+}
+
+bool
+intel_pipe_config_compare(const struct intel_crtc_state *current_config,
+			  const struct intel_crtc_state *pipe_config,
+			  bool fastset)
+{
+	struct drm_i915_private *dev_priv = to_i915(current_config->uapi.crtc->dev);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	bool ret = true;
+	u32 bp_gamma = 0;
+	bool fixup_inherited = fastset &&
+		current_config->inherited && !pipe_config->inherited;
+
+	if (fixup_inherited && !fastboot_enabled(dev_priv)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "initial modeset and fastboot not set\n");
+		ret = false;
+	}
+
+#define PIPE_CONF_CHECK_X(name) do { \
+	if (current_config->name != pipe_config->name) { \
+		pipe_config_mismatch(fastset, crtc, __stringify(name), \
+				     "(expected 0x%08x, found 0x%08x)", \
+				     current_config->name, \
+				     pipe_config->name); \
+		ret = false; \
+	} \
+} while (0)
+
+#define PIPE_CONF_CHECK_X_WITH_MASK(name, mask) do { \
+	if ((current_config->name & (mask)) != (pipe_config->name & (mask))) { \
+		pipe_config_mismatch(fastset, crtc, __stringify(name), \
+				     "(expected 0x%08x, found 0x%08x)", \
+				     current_config->name & (mask), \
+				     pipe_config->name & (mask)); \
+		ret = false; \
+	} \
+} while (0)
+
+#define PIPE_CONF_CHECK_I(name) do { \
+	if (current_config->name != pipe_config->name) { \
+		pipe_config_mismatch(fastset, crtc, __stringify(name), \
+				     "(expected %i, found %i)", \
+				     current_config->name, \
+				     pipe_config->name); \
+		ret = false; \
+	} \
+} while (0)
+
+#define PIPE_CONF_CHECK_BOOL(name) do { \
+	if (current_config->name != pipe_config->name) { \
+		pipe_config_mismatch(fastset, crtc,  __stringify(name), \
+				     "(expected %s, found %s)", \
+				     str_yes_no(current_config->name), \
+				     str_yes_no(pipe_config->name)); \
+		ret = false; \
+	} \
+} while (0)
+
+/*
+ * Checks state where we only read out the enabling, but not the entire
+ * state itself (like full infoframes or ELD for audio). These states
+ * require a full modeset on bootup to fix up.
+ */
+#define PIPE_CONF_CHECK_BOOL_INCOMPLETE(name) do { \
+	if (!fixup_inherited || (!current_config->name && !pipe_config->name)) { \
+		PIPE_CONF_CHECK_BOOL(name); \
+	} else { \
+		pipe_config_mismatch(fastset, crtc, __stringify(name), \
+				     "unable to verify whether state matches exactly, forcing modeset (expected %s, found %s)", \
+				     str_yes_no(current_config->name), \
+				     str_yes_no(pipe_config->name)); \
+		ret = false; \
+	} \
+} while (0)
+
+#define PIPE_CONF_CHECK_P(name) do { \
+	if (current_config->name != pipe_config->name) { \
+		pipe_config_mismatch(fastset, crtc, __stringify(name), \
+				     "(expected %p, found %p)", \
+				     current_config->name, \
+				     pipe_config->name); \
+		ret = false; \
+	} \
+} while (0)
+
+#define PIPE_CONF_CHECK_M_N(name) do { \
+	if (!intel_compare_link_m_n(&current_config->name, \
+				    &pipe_config->name)) { \
+		pipe_config_mismatch(fastset, crtc, __stringify(name), \
+				     "(expected tu %i data %i/%i link %i/%i, " \
+				     "found tu %i, data %i/%i link %i/%i)", \
+				     current_config->name.tu, \
+				     current_config->name.data_m, \
+				     current_config->name.data_n, \
+				     current_config->name.link_m, \
+				     current_config->name.link_n, \
+				     pipe_config->name.tu, \
+				     pipe_config->name.data_m, \
+				     pipe_config->name.data_n, \
+				     pipe_config->name.link_m, \
+				     pipe_config->name.link_n); \
+		ret = false; \
+	} \
+} while (0)
+
+#define PIPE_CONF_CHECK_TIMINGS(name) do { \
+	PIPE_CONF_CHECK_I(name.crtc_hdisplay); \
+	PIPE_CONF_CHECK_I(name.crtc_htotal); \
+	PIPE_CONF_CHECK_I(name.crtc_hblank_start); \
+	PIPE_CONF_CHECK_I(name.crtc_hblank_end); \
+	PIPE_CONF_CHECK_I(name.crtc_hsync_start); \
+	PIPE_CONF_CHECK_I(name.crtc_hsync_end); \
+	PIPE_CONF_CHECK_I(name.crtc_vdisplay); \
+	PIPE_CONF_CHECK_I(name.crtc_vtotal); \
+	PIPE_CONF_CHECK_I(name.crtc_vblank_start); \
+	PIPE_CONF_CHECK_I(name.crtc_vblank_end); \
+	PIPE_CONF_CHECK_I(name.crtc_vsync_start); \
+	PIPE_CONF_CHECK_I(name.crtc_vsync_end); \
+} while (0)
+
+#define PIPE_CONF_CHECK_RECT(name) do { \
+	PIPE_CONF_CHECK_I(name.x1); \
+	PIPE_CONF_CHECK_I(name.x2); \
+	PIPE_CONF_CHECK_I(name.y1); \
+	PIPE_CONF_CHECK_I(name.y2); \
+} while (0)
+
+/* This is required for BDW+ where there is only one set of registers for
+ * switching between high and low RR.
+ * This macro can be used whenever a comparison has to be made between one
+ * hw state and multiple sw state variables.
+ */
+#define PIPE_CONF_CHECK_M_N_ALT(name, alt_name) do { \
+	if (!intel_compare_link_m_n(&current_config->name, \
+				    &pipe_config->name) && \
+	    !intel_compare_link_m_n(&current_config->alt_name, \
+				    &pipe_config->name)) { \
+		pipe_config_mismatch(fastset, crtc, __stringify(name), \
+				     "(expected tu %i data %i/%i link %i/%i, " \
+				     "or tu %i data %i/%i link %i/%i, " \
+				     "found tu %i, data %i/%i link %i/%i)", \
+				     current_config->name.tu, \
+				     current_config->name.data_m, \
+				     current_config->name.data_n, \
+				     current_config->name.link_m, \
+				     current_config->name.link_n, \
+				     current_config->alt_name.tu, \
+				     current_config->alt_name.data_m, \
+				     current_config->alt_name.data_n, \
+				     current_config->alt_name.link_m, \
+				     current_config->alt_name.link_n, \
+				     pipe_config->name.tu, \
+				     pipe_config->name.data_m, \
+				     pipe_config->name.data_n, \
+				     pipe_config->name.link_m, \
+				     pipe_config->name.link_n); \
+		ret = false; \
+	} \
+} while (0)
+
+#define PIPE_CONF_CHECK_FLAGS(name, mask) do { \
+	if ((current_config->name ^ pipe_config->name) & (mask)) { \
+		pipe_config_mismatch(fastset, crtc, __stringify(name), \
+				     "(%x) (expected %i, found %i)", \
+				     (mask), \
+				     current_config->name & (mask), \
+				     pipe_config->name & (mask)); \
+		ret = false; \
+	} \
+} while (0)
+
+#define PIPE_CONF_CHECK_INFOFRAME(name) do { \
+	if (!intel_compare_infoframe(&current_config->infoframes.name, \
+				     &pipe_config->infoframes.name)) { \
+		pipe_config_infoframe_mismatch(dev_priv, fastset, __stringify(name), \
+					       &current_config->infoframes.name, \
+					       &pipe_config->infoframes.name); \
+		ret = false; \
+	} \
+} while (0)
+
+#define PIPE_CONF_CHECK_DP_VSC_SDP(name) do { \
+	if (!current_config->has_psr && !pipe_config->has_psr && \
+	    !intel_compare_dp_vsc_sdp(&current_config->infoframes.name, \
+				      &pipe_config->infoframes.name)) { \
+		pipe_config_dp_vsc_sdp_mismatch(dev_priv, fastset, __stringify(name), \
+						&current_config->infoframes.name, \
+						&pipe_config->infoframes.name); \
+		ret = false; \
+	} \
+} while (0)
+
+#define PIPE_CONF_CHECK_COLOR_LUT(name1, name2, bit_precision) do { \
+	if (current_config->name1 != pipe_config->name1) { \
+		pipe_config_mismatch(fastset, crtc, __stringify(name1), \
+				"(expected %i, found %i, won't compare lut values)", \
+				current_config->name1, \
+				pipe_config->name1); \
+		ret = false;\
+	} else { \
+		if (!intel_color_lut_equal(current_config->name2, \
+					pipe_config->name2, pipe_config->name1, \
+					bit_precision)) { \
+			pipe_config_mismatch(fastset, crtc, __stringify(name2), \
+					"hw_state doesn't match sw_state"); \
+			ret = false; \
+		} \
+	} \
+} while (0)
+
+#define PIPE_CONF_QUIRK(quirk) \
+	((current_config->quirks | pipe_config->quirks) & (quirk))
+
+	PIPE_CONF_CHECK_I(hw.enable);
+	PIPE_CONF_CHECK_I(hw.active);
+
+	PIPE_CONF_CHECK_I(cpu_transcoder);
+	PIPE_CONF_CHECK_I(mst_master_transcoder);
+
+	PIPE_CONF_CHECK_BOOL(has_pch_encoder);
+	PIPE_CONF_CHECK_I(fdi_lanes);
+	PIPE_CONF_CHECK_M_N(fdi_m_n);
+
+	PIPE_CONF_CHECK_I(lane_count);
+	PIPE_CONF_CHECK_X(lane_lat_optim_mask);
+
+	if (HAS_DOUBLE_BUFFERED_M_N(dev_priv)) {
+		if (!fastset || !pipe_config->seamless_m_n)
+			PIPE_CONF_CHECK_M_N_ALT(dp_m_n, dp_m2_n2);
+	} else {
+		PIPE_CONF_CHECK_M_N(dp_m_n);
+		PIPE_CONF_CHECK_M_N(dp_m2_n2);
+	}
+
+	PIPE_CONF_CHECK_X(output_types);
+
+	PIPE_CONF_CHECK_I(framestart_delay);
+	PIPE_CONF_CHECK_I(msa_timing_delay);
+
+	PIPE_CONF_CHECK_TIMINGS(hw.pipe_mode);
+	PIPE_CONF_CHECK_TIMINGS(hw.adjusted_mode);
+
+	PIPE_CONF_CHECK_I(pixel_multiplier);
+
+	PIPE_CONF_CHECK_FLAGS(hw.adjusted_mode.flags,
+			      DRM_MODE_FLAG_INTERLACE);
+
+	if (!PIPE_CONF_QUIRK(PIPE_CONFIG_QUIRK_MODE_SYNC_FLAGS)) {
+		PIPE_CONF_CHECK_FLAGS(hw.adjusted_mode.flags,
+				      DRM_MODE_FLAG_PHSYNC);
+		PIPE_CONF_CHECK_FLAGS(hw.adjusted_mode.flags,
+				      DRM_MODE_FLAG_NHSYNC);
+		PIPE_CONF_CHECK_FLAGS(hw.adjusted_mode.flags,
+				      DRM_MODE_FLAG_PVSYNC);
+		PIPE_CONF_CHECK_FLAGS(hw.adjusted_mode.flags,
+				      DRM_MODE_FLAG_NVSYNC);
+	}
+
+	PIPE_CONF_CHECK_I(output_format);
+	PIPE_CONF_CHECK_BOOL(has_hdmi_sink);
+	if ((DISPLAY_VER(dev_priv) < 8 && !IS_HASWELL(dev_priv)) ||
+	    IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		PIPE_CONF_CHECK_BOOL(limited_color_range);
+
+	PIPE_CONF_CHECK_BOOL(hdmi_scrambling);
+	PIPE_CONF_CHECK_BOOL(hdmi_high_tmds_clock_ratio);
+	PIPE_CONF_CHECK_BOOL(has_infoframe);
+	PIPE_CONF_CHECK_BOOL(fec_enable);
+
+	PIPE_CONF_CHECK_BOOL_INCOMPLETE(has_audio);
+
+	PIPE_CONF_CHECK_X(gmch_pfit.control);
+	/* pfit ratios are autocomputed by the hw on gen4+ */
+	if (DISPLAY_VER(dev_priv) < 4)
+		PIPE_CONF_CHECK_X(gmch_pfit.pgm_ratios);
+	PIPE_CONF_CHECK_X(gmch_pfit.lvds_border_bits);
+
+	/*
+	 * Changing the EDP transcoder input mux
+	 * (A_ONOFF vs. A_ON) requires a full modeset.
+	 */
+	PIPE_CONF_CHECK_BOOL(pch_pfit.force_thru);
+
+	if (!fastset) {
+		PIPE_CONF_CHECK_RECT(pipe_src);
+
+		PIPE_CONF_CHECK_BOOL(pch_pfit.enabled);
+		PIPE_CONF_CHECK_RECT(pch_pfit.dst);
+
+		PIPE_CONF_CHECK_I(scaler_state.scaler_id);
+		PIPE_CONF_CHECK_I(pixel_rate);
+
+		PIPE_CONF_CHECK_X(gamma_mode);
+		if (IS_CHERRYVIEW(dev_priv))
+			PIPE_CONF_CHECK_X(cgm_mode);
+		else
+			PIPE_CONF_CHECK_X(csc_mode);
+		PIPE_CONF_CHECK_BOOL(gamma_enable);
+		PIPE_CONF_CHECK_BOOL(csc_enable);
+
+		PIPE_CONF_CHECK_I(linetime);
+		PIPE_CONF_CHECK_I(ips_linetime);
+
+		bp_gamma = intel_color_get_gamma_bit_precision(pipe_config);
+		if (bp_gamma)
+			PIPE_CONF_CHECK_COLOR_LUT(gamma_mode, hw.gamma_lut, bp_gamma);
+
+		if (current_config->active_planes) {
+			PIPE_CONF_CHECK_BOOL(has_psr);
+			PIPE_CONF_CHECK_BOOL(has_psr2);
+			PIPE_CONF_CHECK_BOOL(enable_psr2_sel_fetch);
+			PIPE_CONF_CHECK_I(dc3co_exitline);
+		}
+	}
+
+	PIPE_CONF_CHECK_BOOL(double_wide);
+
+	if (dev_priv->display.dpll.mgr) {
+		PIPE_CONF_CHECK_P(shared_dpll);
+
+		PIPE_CONF_CHECK_X(dpll_hw_state.dpll);
+		PIPE_CONF_CHECK_X(dpll_hw_state.dpll_md);
+		PIPE_CONF_CHECK_X(dpll_hw_state.fp0);
+		PIPE_CONF_CHECK_X(dpll_hw_state.fp1);
+		PIPE_CONF_CHECK_X(dpll_hw_state.wrpll);
+		PIPE_CONF_CHECK_X(dpll_hw_state.spll);
+		PIPE_CONF_CHECK_X(dpll_hw_state.ctrl1);
+		PIPE_CONF_CHECK_X(dpll_hw_state.cfgcr1);
+		PIPE_CONF_CHECK_X(dpll_hw_state.cfgcr2);
+		PIPE_CONF_CHECK_X(dpll_hw_state.cfgcr0);
+		PIPE_CONF_CHECK_X(dpll_hw_state.div0);
+		PIPE_CONF_CHECK_X(dpll_hw_state.ebb0);
+		PIPE_CONF_CHECK_X(dpll_hw_state.ebb4);
+		PIPE_CONF_CHECK_X(dpll_hw_state.pll0);
+		PIPE_CONF_CHECK_X(dpll_hw_state.pll1);
+		PIPE_CONF_CHECK_X(dpll_hw_state.pll2);
+		PIPE_CONF_CHECK_X(dpll_hw_state.pll3);
+		PIPE_CONF_CHECK_X(dpll_hw_state.pll6);
+		PIPE_CONF_CHECK_X(dpll_hw_state.pll8);
+		PIPE_CONF_CHECK_X(dpll_hw_state.pll9);
+		PIPE_CONF_CHECK_X(dpll_hw_state.pll10);
+		PIPE_CONF_CHECK_X(dpll_hw_state.pcsdw12);
+		PIPE_CONF_CHECK_X(dpll_hw_state.mg_refclkin_ctl);
+		PIPE_CONF_CHECK_X(dpll_hw_state.mg_clktop2_coreclkctl1);
+		PIPE_CONF_CHECK_X(dpll_hw_state.mg_clktop2_hsclkctl);
+		PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_div0);
+		PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_div1);
+		PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_lf);
+		PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_frac_lock);
+		PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_ssc);
+		PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_bias);
+		PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_tdc_coldst_bias);
+	}
+
+	PIPE_CONF_CHECK_X(dsi_pll.ctrl);
+	PIPE_CONF_CHECK_X(dsi_pll.div);
+
+	if (IS_G4X(dev_priv) || DISPLAY_VER(dev_priv) >= 5)
+		PIPE_CONF_CHECK_I(pipe_bpp);
+
+	if (!fastset || !pipe_config->seamless_m_n) {
+		PIPE_CONF_CHECK_I(hw.pipe_mode.crtc_clock);
+		PIPE_CONF_CHECK_I(hw.adjusted_mode.crtc_clock);
+	}
+	PIPE_CONF_CHECK_I(port_clock);
+
+	PIPE_CONF_CHECK_I(min_voltage_level);
+
+	if (current_config->has_psr || pipe_config->has_psr)
+		PIPE_CONF_CHECK_X_WITH_MASK(infoframes.enable,
+					    ~intel_hdmi_infoframe_enable(DP_SDP_VSC));
+	else
+		PIPE_CONF_CHECK_X(infoframes.enable);
+
+	PIPE_CONF_CHECK_X(infoframes.gcp);
+	PIPE_CONF_CHECK_INFOFRAME(avi);
+	PIPE_CONF_CHECK_INFOFRAME(spd);
+	PIPE_CONF_CHECK_INFOFRAME(hdmi);
+	PIPE_CONF_CHECK_INFOFRAME(drm);
+	PIPE_CONF_CHECK_DP_VSC_SDP(vsc);
+
+	PIPE_CONF_CHECK_X(sync_mode_slaves_mask);
+	PIPE_CONF_CHECK_I(master_transcoder);
+	PIPE_CONF_CHECK_X(bigjoiner_pipes);
+
+	PIPE_CONF_CHECK_I(dsc.compression_enable);
+	PIPE_CONF_CHECK_I(dsc.dsc_split);
+	PIPE_CONF_CHECK_I(dsc.compressed_bpp);
+
+	PIPE_CONF_CHECK_BOOL(splitter.enable);
+	PIPE_CONF_CHECK_I(splitter.link_count);
+	PIPE_CONF_CHECK_I(splitter.pixel_overlap);
+
+	PIPE_CONF_CHECK_BOOL(vrr.enable);
+	PIPE_CONF_CHECK_I(vrr.vmin);
+	PIPE_CONF_CHECK_I(vrr.vmax);
+	PIPE_CONF_CHECK_I(vrr.flipline);
+	PIPE_CONF_CHECK_I(vrr.pipeline_full);
+	PIPE_CONF_CHECK_I(vrr.guardband);
+
+#undef PIPE_CONF_CHECK_X
+#undef PIPE_CONF_CHECK_I
+#undef PIPE_CONF_CHECK_BOOL
+#undef PIPE_CONF_CHECK_BOOL_INCOMPLETE
+#undef PIPE_CONF_CHECK_P
+#undef PIPE_CONF_CHECK_FLAGS
+#undef PIPE_CONF_CHECK_COLOR_LUT
+#undef PIPE_CONF_CHECK_TIMINGS
+#undef PIPE_CONF_CHECK_RECT
+#undef PIPE_CONF_QUIRK
+
+	return ret;
+}
+
+static void
+intel_verify_planes(struct intel_atomic_state *state)
+{
+	struct intel_plane *plane;
+	const struct intel_plane_state *plane_state;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane,
+					  plane_state, i)
+		assert_plane(plane, plane_state->planar_slave ||
+			     plane_state->uapi.visible);
+}
+
+int intel_modeset_all_pipes(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc *crtc;
+
+	/*
+	 * Add all pipes to the state, and force
+	 * a modeset on all the active ones.
+	 */
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		struct intel_crtc_state *crtc_state;
+		int ret;
+
+		crtc_state = intel_atomic_get_crtc_state(&state->base, crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+
+		if (!crtc_state->hw.active ||
+		    drm_atomic_crtc_needs_modeset(&crtc_state->uapi))
+			continue;
+
+		crtc_state->uapi.mode_changed = true;
+
+		ret = drm_atomic_add_affected_connectors(&state->base,
+							 &crtc->base);
+		if (ret)
+			return ret;
+
+		ret = intel_dp_mst_add_topology_state_for_crtc(state, crtc);
+		if (ret)
+			return ret;
+
+		ret = intel_atomic_add_affected_planes(state, crtc);
+		if (ret)
+			return ret;
+
+		crtc_state->update_planes |= crtc_state->active_planes;
+	}
+
+	return 0;
+}
+
+void intel_crtc_update_active_timings(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct drm_display_mode adjusted_mode;
+
+	drm_mode_init(&adjusted_mode, &crtc_state->hw.adjusted_mode);
+
+	if (crtc_state->vrr.enable) {
+		adjusted_mode.crtc_vtotal = crtc_state->vrr.vmax;
+		adjusted_mode.crtc_vblank_end = crtc_state->vrr.vmax;
+		adjusted_mode.crtc_vblank_start = intel_vrr_vmin_vblank_start(crtc_state);
+		crtc->vmax_vblank_start = intel_vrr_vmax_vblank_start(crtc_state);
+	}
+
+	drm_calc_timestamping_constants(&crtc->base, &adjusted_mode);
+
+	crtc->mode_flags = crtc_state->mode_flags;
+
+	/*
+	 * The scanline counter increments at the leading edge of hsync.
+	 *
+	 * On most platforms it starts counting from vtotal-1 on the
+	 * first active line. That means the scanline counter value is
+	 * always one less than what we would expect. Ie. just after
+	 * start of vblank, which also occurs at start of hsync (on the
+	 * last active line), the scanline counter will read vblank_start-1.
+	 *
+	 * On gen2 the scanline counter starts counting from 1 instead
+	 * of vtotal-1, so we have to subtract one (or rather add vtotal-1
+	 * to keep the value positive), instead of adding one.
+	 *
+	 * On HSW+ the behaviour of the scanline counter depends on the output
+	 * type. For DP ports it behaves like most other platforms, but on HDMI
+	 * there's an extra 1 line difference. So we need to add two instead of
+	 * one to the value.
+	 *
+	 * On VLV/CHV DSI the scanline counter would appear to increment
+	 * approx. 1/3 of a scanline before start of vblank. Unfortunately
+	 * that means we can't tell whether we're in vblank or not while
+	 * we're on that particular line. We must still set scanline_offset
+	 * to 1 so that the vblank timestamps come out correct when we query
+	 * the scanline counter from within the vblank interrupt handler.
+	 * However if queried just before the start of vblank we'll get an
+	 * answer that's slightly in the future.
+	 */
+	if (DISPLAY_VER(dev_priv) == 2) {
+		int vtotal;
+
+		vtotal = adjusted_mode.crtc_vtotal;
+		if (adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
+			vtotal /= 2;
+
+		crtc->scanline_offset = vtotal - 1;
+	} else if (HAS_DDI(dev_priv) &&
+		   intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI)) {
+		crtc->scanline_offset = 2;
+	} else {
+		crtc->scanline_offset = 1;
+	}
+}
+
+/*
+ * This implements the workaround described in the "notes" section of the mode
+ * set sequence documentation. When going from no pipes or single pipe to
+ * multiple pipes, and planes are enabled after the pipe, we need to wait at
+ * least 2 vblanks on the first pipe before enabling planes on the second pipe.
+ */
+static int hsw_mode_set_planes_workaround(struct intel_atomic_state *state)
+{
+	struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+	struct intel_crtc_state *first_crtc_state = NULL;
+	struct intel_crtc_state *other_crtc_state = NULL;
+	enum pipe first_pipe = INVALID_PIPE, enabled_pipe = INVALID_PIPE;
+	int i;
+
+	/* look at all crtc's that are going to be enabled in during modeset */
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		if (!crtc_state->hw.active ||
+		    !intel_crtc_needs_modeset(crtc_state))
+			continue;
+
+		if (first_crtc_state) {
+			other_crtc_state = crtc_state;
+			break;
+		} else {
+			first_crtc_state = crtc_state;
+			first_pipe = crtc->pipe;
+		}
+	}
+
+	/* No workaround needed? */
+	if (!first_crtc_state)
+		return 0;
+
+	/* w/a possibly needed, check how many crtc's are already enabled. */
+	for_each_intel_crtc(state->base.dev, crtc) {
+		crtc_state = intel_atomic_get_crtc_state(&state->base, crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+
+		crtc_state->hsw_workaround_pipe = INVALID_PIPE;
+
+		if (!crtc_state->hw.active ||
+		    intel_crtc_needs_modeset(crtc_state))
+			continue;
+
+		/* 2 or more enabled crtcs means no need for w/a */
+		if (enabled_pipe != INVALID_PIPE)
+			return 0;
+
+		enabled_pipe = crtc->pipe;
+	}
+
+	if (enabled_pipe != INVALID_PIPE)
+		first_crtc_state->hsw_workaround_pipe = enabled_pipe;
+	else if (other_crtc_state)
+		other_crtc_state->hsw_workaround_pipe = first_pipe;
+
+	return 0;
+}
+
+u8 intel_calc_active_pipes(struct intel_atomic_state *state,
+			   u8 active_pipes)
+{
+	const struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+	int i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		if (crtc_state->hw.active)
+			active_pipes |= BIT(crtc->pipe);
+		else
+			active_pipes &= ~BIT(crtc->pipe);
+	}
+
+	return active_pipes;
+}
+
+static int intel_modeset_checks(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+
+	state->modeset = true;
+
+	if (IS_HASWELL(dev_priv))
+		return hsw_mode_set_planes_workaround(state);
+
+	return 0;
+}
+
+static void intel_crtc_check_fastset(const struct intel_crtc_state *old_crtc_state,
+				     struct intel_crtc_state *new_crtc_state)
+{
+	if (!intel_pipe_config_compare(old_crtc_state, new_crtc_state, true))
+		return;
+
+	new_crtc_state->uapi.mode_changed = false;
+	new_crtc_state->update_pipe = true;
+}
+
+static int intel_crtc_add_planes_to_state(struct intel_atomic_state *state,
+					  struct intel_crtc *crtc,
+					  u8 plane_ids_mask)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_plane *plane;
+
+	for_each_intel_plane_on_crtc(&dev_priv->drm, crtc, plane) {
+		struct intel_plane_state *plane_state;
+
+		if ((plane_ids_mask & BIT(plane->id)) == 0)
+			continue;
+
+		plane_state = intel_atomic_get_plane_state(state, plane);
+		if (IS_ERR(plane_state))
+			return PTR_ERR(plane_state);
+	}
+
+	return 0;
+}
+
+int intel_atomic_add_affected_planes(struct intel_atomic_state *state,
+				     struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	return intel_crtc_add_planes_to_state(state, crtc,
+					      old_crtc_state->enabled_planes |
+					      new_crtc_state->enabled_planes);
+}
+
+static bool active_planes_affects_min_cdclk(struct drm_i915_private *dev_priv)
+{
+	/* See {hsw,vlv,ivb}_plane_ratio() */
+	return IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv) ||
+		IS_CHERRYVIEW(dev_priv) || IS_VALLEYVIEW(dev_priv) ||
+		IS_IVYBRIDGE(dev_priv);
+}
+
+static int intel_crtc_add_bigjoiner_planes(struct intel_atomic_state *state,
+					   struct intel_crtc *crtc,
+					   struct intel_crtc *other)
+{
+	const struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	u8 plane_ids = 0;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		if (plane->pipe == crtc->pipe)
+			plane_ids |= BIT(plane->id);
+	}
+
+	return intel_crtc_add_planes_to_state(state, other, plane_ids);
+}
+
+static int intel_bigjoiner_add_affected_planes(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+	int i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		struct intel_crtc *other;
+
+		for_each_intel_crtc_in_pipe_mask(&i915->drm, other,
+						 crtc_state->bigjoiner_pipes) {
+			int ret;
+
+			if (crtc == other)
+				continue;
+
+			ret = intel_crtc_add_bigjoiner_planes(state, crtc, other);
+			if (ret)
+				return ret;
+		}
+	}
+
+	return 0;
+}
+
+static int intel_atomic_check_planes(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *old_crtc_state, *new_crtc_state;
+	struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	struct intel_crtc *crtc;
+	int i, ret;
+
+	ret = icl_add_linked_planes(state);
+	if (ret)
+		return ret;
+
+	ret = intel_bigjoiner_add_affected_planes(state);
+	if (ret)
+		return ret;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		ret = intel_plane_atomic_check(state, plane);
+		if (ret) {
+			drm_dbg_atomic(&dev_priv->drm,
+				       "[PLANE:%d:%s] atomic driver check failed\n",
+				       plane->base.base.id, plane->base.name);
+			return ret;
+		}
+	}
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		u8 old_active_planes, new_active_planes;
+
+		ret = icl_check_nv12_planes(new_crtc_state);
+		if (ret)
+			return ret;
+
+		/*
+		 * On some platforms the number of active planes affects
+		 * the planes' minimum cdclk calculation. Add such planes
+		 * to the state before we compute the minimum cdclk.
+		 */
+		if (!active_planes_affects_min_cdclk(dev_priv))
+			continue;
+
+		old_active_planes = old_crtc_state->active_planes & ~BIT(PLANE_CURSOR);
+		new_active_planes = new_crtc_state->active_planes & ~BIT(PLANE_CURSOR);
+
+		if (hweight8(old_active_planes) == hweight8(new_active_planes))
+			continue;
+
+		ret = intel_crtc_add_planes_to_state(state, crtc, new_active_planes);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int intel_atomic_check_crtcs(struct intel_atomic_state *state)
+{
+	struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+	int i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+		int ret;
+
+		ret = intel_crtc_atomic_check(state, crtc);
+		if (ret) {
+			drm_dbg_atomic(&i915->drm,
+				       "[CRTC:%d:%s] atomic driver check failed\n",
+				       crtc->base.base.id, crtc->base.name);
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+static bool intel_cpu_transcoders_need_modeset(struct intel_atomic_state *state,
+					       u8 transcoders)
+{
+	const struct intel_crtc_state *new_crtc_state;
+	struct intel_crtc *crtc;
+	int i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		if (new_crtc_state->hw.enable &&
+		    transcoders & BIT(new_crtc_state->cpu_transcoder) &&
+		    intel_crtc_needs_modeset(new_crtc_state))
+			return true;
+	}
+
+	return false;
+}
+
+static bool intel_pipes_need_modeset(struct intel_atomic_state *state,
+				     u8 pipes)
+{
+	const struct intel_crtc_state *new_crtc_state;
+	struct intel_crtc *crtc;
+	int i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		if (new_crtc_state->hw.enable &&
+		    pipes & BIT(crtc->pipe) &&
+		    intel_crtc_needs_modeset(new_crtc_state))
+			return true;
+	}
+
+	return false;
+}
+
+static int intel_atomic_check_bigjoiner(struct intel_atomic_state *state,
+					struct intel_crtc *master_crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_crtc_state *master_crtc_state =
+		intel_atomic_get_new_crtc_state(state, master_crtc);
+	struct intel_crtc *slave_crtc;
+
+	if (!master_crtc_state->bigjoiner_pipes)
+		return 0;
+
+	/* sanity check */
+	if (drm_WARN_ON(&i915->drm,
+			master_crtc->pipe != bigjoiner_master_pipe(master_crtc_state)))
+		return -EINVAL;
+
+	if (master_crtc_state->bigjoiner_pipes & ~bigjoiner_pipes(i915)) {
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] Cannot act as big joiner master "
+			    "(need 0x%x as pipes, only 0x%x possible)\n",
+			    master_crtc->base.base.id, master_crtc->base.name,
+			    master_crtc_state->bigjoiner_pipes, bigjoiner_pipes(i915));
+		return -EINVAL;
+	}
+
+	for_each_intel_crtc_in_pipe_mask(&i915->drm, slave_crtc,
+					 intel_crtc_bigjoiner_slave_pipes(master_crtc_state)) {
+		struct intel_crtc_state *slave_crtc_state;
+		int ret;
+
+		slave_crtc_state = intel_atomic_get_crtc_state(&state->base, slave_crtc);
+		if (IS_ERR(slave_crtc_state))
+			return PTR_ERR(slave_crtc_state);
+
+		/* master being enabled, slave was already configured? */
+		if (slave_crtc_state->uapi.enable) {
+			drm_dbg_kms(&i915->drm,
+				    "[CRTC:%d:%s] Slave is enabled as normal CRTC, but "
+				    "[CRTC:%d:%s] claiming this CRTC for bigjoiner.\n",
+				    slave_crtc->base.base.id, slave_crtc->base.name,
+				    master_crtc->base.base.id, master_crtc->base.name);
+			return -EINVAL;
+		}
+
+		/*
+		 * The state copy logic assumes the master crtc gets processed
+		 * before the slave crtc during the main compute_config loop.
+		 * This works because the crtcs are created in pipe order,
+		 * and the hardware requires master pipe < slave pipe as well.
+		 * Should that change we need to rethink the logic.
+		 */
+		if (WARN_ON(drm_crtc_index(&master_crtc->base) >
+			    drm_crtc_index(&slave_crtc->base)))
+			return -EINVAL;
+
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] Used as slave for big joiner master [CRTC:%d:%s]\n",
+			    slave_crtc->base.base.id, slave_crtc->base.name,
+			    master_crtc->base.base.id, master_crtc->base.name);
+
+		slave_crtc_state->bigjoiner_pipes =
+			master_crtc_state->bigjoiner_pipes;
+
+		ret = copy_bigjoiner_crtc_state_modeset(state, slave_crtc);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static void kill_bigjoiner_slave(struct intel_atomic_state *state,
+				 struct intel_crtc *master_crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_crtc_state *master_crtc_state =
+		intel_atomic_get_new_crtc_state(state, master_crtc);
+	struct intel_crtc *slave_crtc;
+
+	for_each_intel_crtc_in_pipe_mask(&i915->drm, slave_crtc,
+					 intel_crtc_bigjoiner_slave_pipes(master_crtc_state)) {
+		struct intel_crtc_state *slave_crtc_state =
+			intel_atomic_get_new_crtc_state(state, slave_crtc);
+
+		slave_crtc_state->bigjoiner_pipes = 0;
+
+		intel_crtc_copy_uapi_to_hw_state_modeset(state, slave_crtc);
+	}
+
+	master_crtc_state->bigjoiner_pipes = 0;
+}
+
+/**
+ * DOC: asynchronous flip implementation
+ *
+ * Asynchronous page flip is the implementation for the DRM_MODE_PAGE_FLIP_ASYNC
+ * flag. Currently async flip is only supported via the drmModePageFlip IOCTL.
+ * Correspondingly, support is currently added for primary plane only.
+ *
+ * Async flip can only change the plane surface address, so anything else
+ * changing is rejected from the intel_async_flip_check_hw() function.
+ * Once this check is cleared, flip done interrupt is enabled using
+ * the intel_crtc_enable_flip_done() function.
+ *
+ * As soon as the surface address register is written, flip done interrupt is
+ * generated and the requested events are sent to the usersapce in the interrupt
+ * handler itself. The timestamp and sequence sent during the flip done event
+ * correspond to the last vblank and have no relation to the actual time when
+ * the flip done event was sent.
+ */
+static int intel_async_flip_check_uapi(struct intel_atomic_state *state,
+				       struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_plane_state *old_plane_state;
+	struct intel_plane_state *new_plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	if (!new_crtc_state->uapi.async_flip)
+		return 0;
+
+	if (!new_crtc_state->uapi.active) {
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] not active\n",
+			    crtc->base.base.id, crtc->base.name);
+		return -EINVAL;
+	}
+
+	if (intel_crtc_needs_modeset(new_crtc_state)) {
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] modeset required\n",
+			    crtc->base.base.id, crtc->base.name);
+		return -EINVAL;
+	}
+
+	/*
+	 * FIXME: Bigjoiner+async flip is busted currently.
+	 * Remove this check once the issues are fixed.
+	 */
+	if (new_crtc_state->bigjoiner_pipes) {
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] async flip disallowed with bigjoiner\n",
+			    crtc->base.base.id, crtc->base.name);
+		return -EINVAL;
+	}
+
+	for_each_oldnew_intel_plane_in_state(state, plane, old_plane_state,
+					     new_plane_state, i) {
+		if (plane->pipe != crtc->pipe)
+			continue;
+
+		/*
+		 * TODO: Async flip is only supported through the page flip IOCTL
+		 * as of now. So support currently added for primary plane only.
+		 * Support for other planes on platforms on which supports
+		 * this(vlv/chv and icl+) should be added when async flip is
+		 * enabled in the atomic IOCTL path.
+		 */
+		if (!plane->async_flip) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] async flip not supported\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (!old_plane_state->uapi.fb || !new_plane_state->uapi.fb) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] no old or new framebuffer\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static int intel_async_flip_check_hw(struct intel_atomic_state *state, struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *old_crtc_state, *new_crtc_state;
+	const struct intel_plane_state *new_plane_state, *old_plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	old_crtc_state = intel_atomic_get_old_crtc_state(state, crtc);
+	new_crtc_state = intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (!new_crtc_state->uapi.async_flip)
+		return 0;
+
+	if (!new_crtc_state->hw.active) {
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] not active\n",
+			    crtc->base.base.id, crtc->base.name);
+		return -EINVAL;
+	}
+
+	if (intel_crtc_needs_modeset(new_crtc_state)) {
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] modeset required\n",
+			    crtc->base.base.id, crtc->base.name);
+		return -EINVAL;
+	}
+
+	if (old_crtc_state->active_planes != new_crtc_state->active_planes) {
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] Active planes cannot be in async flip\n",
+			    crtc->base.base.id, crtc->base.name);
+		return -EINVAL;
+	}
+
+	for_each_oldnew_intel_plane_in_state(state, plane, old_plane_state,
+					     new_plane_state, i) {
+		if (plane->pipe != crtc->pipe)
+			continue;
+
+		/*
+		 * Only async flip capable planes should be in the state
+		 * if we're really about to ask the hardware to perform
+		 * an async flip. We should never get this far otherwise.
+		 */
+		if (drm_WARN_ON(&i915->drm,
+				new_crtc_state->do_async_flip && !plane->async_flip))
+			return -EINVAL;
+
+		/*
+		 * Only check async flip capable planes other planes
+		 * may be involved in the initial commit due to
+		 * the wm0/ddb optimization.
+		 *
+		 * TODO maybe should track which planes actually
+		 * were requested to do the async flip...
+		 */
+		if (!plane->async_flip)
+			continue;
+
+		/*
+		 * FIXME: This check is kept generic for all platforms.
+		 * Need to verify this for all gen9 platforms to enable
+		 * this selectively if required.
+		 */
+		switch (new_plane_state->hw.fb->modifier) {
+		case I915_FORMAT_MOD_X_TILED:
+		case I915_FORMAT_MOD_Y_TILED:
+		case I915_FORMAT_MOD_Yf_TILED:
+		case I915_FORMAT_MOD_4_TILED:
+			break;
+		default:
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Modifier does not support async flips\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (new_plane_state->hw.fb->format->num_planes > 1) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Planar formats do not support async flips\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (old_plane_state->view.color_plane[0].mapping_stride !=
+		    new_plane_state->view.color_plane[0].mapping_stride) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Stride cannot be changed in async flip\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (old_plane_state->hw.fb->modifier !=
+		    new_plane_state->hw.fb->modifier) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Modifier cannot be changed in async flip\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (old_plane_state->hw.fb->format !=
+		    new_plane_state->hw.fb->format) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Pixel format cannot be changed in async flip\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (old_plane_state->hw.rotation !=
+		    new_plane_state->hw.rotation) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Rotation cannot be changed in async flip\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (!drm_rect_equals(&old_plane_state->uapi.src, &new_plane_state->uapi.src) ||
+		    !drm_rect_equals(&old_plane_state->uapi.dst, &new_plane_state->uapi.dst)) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Size/co-ordinates cannot be changed in async flip\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (old_plane_state->hw.alpha != new_plane_state->hw.alpha) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANES:%d:%s] Alpha value cannot be changed in async flip\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (old_plane_state->hw.pixel_blend_mode !=
+		    new_plane_state->hw.pixel_blend_mode) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Pixel blend mode cannot be changed in async flip\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (old_plane_state->hw.color_encoding != new_plane_state->hw.color_encoding) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Color encoding cannot be changed in async flip\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		if (old_plane_state->hw.color_range != new_plane_state->hw.color_range) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Color range cannot be changed in async flip\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+
+		/* plane decryption is allow to change only in synchronous flips */
+		if (old_plane_state->decrypt != new_plane_state->decrypt) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] Decryption cannot be changed in async flip\n",
+				    plane->base.base.id, plane->base.name);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static int intel_bigjoiner_add_affected_crtcs(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+	u8 affected_pipes = 0;
+	u8 modeset_pipes = 0;
+	int i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		affected_pipes |= crtc_state->bigjoiner_pipes;
+		if (intel_crtc_needs_modeset(crtc_state))
+			modeset_pipes |= crtc_state->bigjoiner_pipes;
+	}
+
+	for_each_intel_crtc_in_pipe_mask(&i915->drm, crtc, affected_pipes) {
+		crtc_state = intel_atomic_get_crtc_state(&state->base, crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+	}
+
+	for_each_intel_crtc_in_pipe_mask(&i915->drm, crtc, modeset_pipes) {
+		int ret;
+
+		crtc_state = intel_atomic_get_new_crtc_state(state, crtc);
+
+		crtc_state->uapi.mode_changed = true;
+
+		ret = drm_atomic_add_affected_connectors(&state->base, &crtc->base);
+		if (ret)
+			return ret;
+
+		ret = intel_atomic_add_affected_planes(state, crtc);
+		if (ret)
+			return ret;
+	}
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		/* Kill old bigjoiner link, we may re-establish afterwards */
+		if (intel_crtc_needs_modeset(crtc_state) &&
+		    intel_crtc_is_bigjoiner_master(crtc_state))
+			kill_bigjoiner_slave(state, crtc);
+	}
+
+	return 0;
+}
+
+/**
+ * intel_atomic_check - validate state object
+ * @dev: drm device
+ * @_state: state to validate
+ */
+static int intel_atomic_check(struct drm_device *dev,
+			      struct drm_atomic_state *_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_atomic_state *state = to_intel_atomic_state(_state);
+	struct intel_crtc_state *old_crtc_state, *new_crtc_state;
+	struct intel_crtc *crtc;
+	int ret, i;
+	bool any_ms = false;
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		if (new_crtc_state->inherited != old_crtc_state->inherited)
+			new_crtc_state->uapi.mode_changed = true;
+
+		if (new_crtc_state->uapi.scaling_filter !=
+		    old_crtc_state->uapi.scaling_filter)
+			new_crtc_state->uapi.mode_changed = true;
+	}
+
+	intel_vrr_check_modeset(state);
+
+	ret = drm_atomic_helper_check_modeset(dev, &state->base);
+	if (ret)
+		goto fail;
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		ret = intel_async_flip_check_uapi(state, crtc);
+		if (ret)
+			return ret;
+	}
+
+	ret = intel_bigjoiner_add_affected_crtcs(state);
+	if (ret)
+		goto fail;
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		if (!intel_crtc_needs_modeset(new_crtc_state)) {
+			if (intel_crtc_is_bigjoiner_slave(new_crtc_state))
+				copy_bigjoiner_crtc_state_nomodeset(state, crtc);
+			else
+				intel_crtc_copy_uapi_to_hw_state_nomodeset(state, crtc);
+			continue;
+		}
+
+		if (intel_crtc_is_bigjoiner_slave(new_crtc_state)) {
+			drm_WARN_ON(&dev_priv->drm, new_crtc_state->uapi.enable);
+			continue;
+		}
+
+		ret = intel_crtc_prepare_cleared_state(state, crtc);
+		if (ret)
+			goto fail;
+
+		if (!new_crtc_state->hw.enable)
+			continue;
+
+		ret = intel_modeset_pipe_config(state, crtc);
+		if (ret)
+			goto fail;
+
+		ret = intel_atomic_check_bigjoiner(state, crtc);
+		if (ret)
+			goto fail;
+	}
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		if (!intel_crtc_needs_modeset(new_crtc_state))
+			continue;
+
+		if (new_crtc_state->hw.enable) {
+			ret = intel_modeset_pipe_config_late(state, crtc);
+			if (ret)
+				goto fail;
+		}
+
+		intel_crtc_check_fastset(old_crtc_state, new_crtc_state);
+	}
+
+	/**
+	 * Check if fastset is allowed by external dependencies like other
+	 * pipes and transcoders.
+	 *
+	 * Right now it only forces a fullmodeset when the MST master
+	 * transcoder did not changed but the pipe of the master transcoder
+	 * needs a fullmodeset so all slaves also needs to do a fullmodeset or
+	 * in case of port synced crtcs, if one of the synced crtcs
+	 * needs a full modeset, all other synced crtcs should be
+	 * forced a full modeset.
+	 */
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		if (!new_crtc_state->hw.enable || intel_crtc_needs_modeset(new_crtc_state))
+			continue;
+
+		if (intel_dp_mst_is_slave_trans(new_crtc_state)) {
+			enum transcoder master = new_crtc_state->mst_master_transcoder;
+
+			if (intel_cpu_transcoders_need_modeset(state, BIT(master))) {
+				new_crtc_state->uapi.mode_changed = true;
+				new_crtc_state->update_pipe = false;
+			}
+		}
+
+		if (is_trans_port_sync_mode(new_crtc_state)) {
+			u8 trans = new_crtc_state->sync_mode_slaves_mask;
+
+			if (new_crtc_state->master_transcoder != INVALID_TRANSCODER)
+				trans |= BIT(new_crtc_state->master_transcoder);
+
+			if (intel_cpu_transcoders_need_modeset(state, trans)) {
+				new_crtc_state->uapi.mode_changed = true;
+				new_crtc_state->update_pipe = false;
+			}
+		}
+
+		if (new_crtc_state->bigjoiner_pipes) {
+			if (intel_pipes_need_modeset(state, new_crtc_state->bigjoiner_pipes)) {
+				new_crtc_state->uapi.mode_changed = true;
+				new_crtc_state->update_pipe = false;
+			}
+		}
+	}
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		if (!intel_crtc_needs_modeset(new_crtc_state))
+			continue;
+
+		any_ms = true;
+
+		intel_release_shared_dplls(state, crtc);
+	}
+
+	if (any_ms && !check_digital_port_conflicts(state)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "rejecting conflicting digital port configuration\n");
+		ret = -EINVAL;
+		goto fail;
+	}
+
+	ret = drm_dp_mst_atomic_check(&state->base);
+	if (ret)
+		goto fail;
+
+	ret = intel_atomic_check_planes(state);
+	if (ret)
+		goto fail;
+
+	ret = intel_compute_global_watermarks(state);
+	if (ret)
+		goto fail;
+
+	ret = intel_bw_atomic_check(state);
+	if (ret)
+		goto fail;
+
+	ret = intel_cdclk_atomic_check(state, &any_ms);
+	if (ret)
+		goto fail;
+
+	if (intel_any_crtc_needs_modeset(state))
+		any_ms = true;
+
+	if (any_ms) {
+		ret = intel_modeset_checks(state);
+		if (ret)
+			goto fail;
+
+		ret = intel_modeset_calc_cdclk(state);
+		if (ret)
+			return ret;
+	}
+
+	ret = intel_atomic_check_crtcs(state);
+	if (ret)
+		goto fail;
+
+	ret = intel_fbc_atomic_check(state);
+	if (ret)
+		goto fail;
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		ret = intel_async_flip_check_hw(state, crtc);
+		if (ret)
+			goto fail;
+
+		if (!intel_crtc_needs_modeset(new_crtc_state) &&
+		    !new_crtc_state->update_pipe)
+			continue;
+
+		intel_crtc_state_dump(new_crtc_state, state,
+				      intel_crtc_needs_modeset(new_crtc_state) ?
+				      "modeset" : "fastset");
+	}
+
+	return 0;
+
+ fail:
+	if (ret == -EDEADLK)
+		return ret;
+
+	/*
+	 * FIXME would probably be nice to know which crtc specifically
+	 * caused the failure, in cases where we can pinpoint it.
+	 */
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i)
+		intel_crtc_state_dump(new_crtc_state, state, "failed");
+
+	return ret;
+}
+
+static int intel_atomic_prepare_commit(struct intel_atomic_state *state)
+{
+	struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+	int i, ret;
+
+	ret = drm_atomic_helper_prepare_planes(state->base.dev, &state->base);
+	if (ret < 0)
+		return ret;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) {
+		bool mode_changed = intel_crtc_needs_modeset(crtc_state);
+
+		if (mode_changed || crtc_state->update_pipe ||
+		    crtc_state->uapi.color_mgmt_changed) {
+			intel_dsb_prepare(crtc_state);
+		}
+	}
+
+	return 0;
+}
+
+void intel_crtc_arm_fifo_underrun(struct intel_crtc *crtc,
+				  struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	if (DISPLAY_VER(dev_priv) != 2 || crtc_state->active_planes)
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, crtc->pipe, true);
+
+	if (crtc_state->has_pch_encoder) {
+		enum pipe pch_transcoder =
+			intel_crtc_pch_transcoder(crtc);
+
+		intel_set_pch_fifo_underrun_reporting(dev_priv, pch_transcoder, true);
+	}
+}
+
+static void intel_pipe_fastset(const struct intel_crtc_state *old_crtc_state,
+			       const struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	/*
+	 * Update pipe size and adjust fitter if needed: the reason for this is
+	 * that in compute_mode_changes we check the native mode (not the pfit
+	 * mode) to see if we can flip rather than do a full mode set. In the
+	 * fastboot case, we'll flip, but if we don't update the pipesrc and
+	 * pfit state, we'll end up with a big fb scanned out into the wrong
+	 * sized surface.
+	 */
+	intel_set_pipe_src_size(new_crtc_state);
+
+	/* on skylake this is done by detaching scalers */
+	if (DISPLAY_VER(dev_priv) >= 9) {
+		if (new_crtc_state->pch_pfit.enabled)
+			skl_pfit_enable(new_crtc_state);
+	} else if (HAS_PCH_SPLIT(dev_priv)) {
+		if (new_crtc_state->pch_pfit.enabled)
+			ilk_pfit_enable(new_crtc_state);
+		else if (old_crtc_state->pch_pfit.enabled)
+			ilk_pfit_disable(old_crtc_state);
+	}
+
+	/*
+	 * The register is supposedly single buffered so perhaps
+	 * not 100% correct to do this here. But SKL+ calculate
+	 * this based on the adjust pixel rate so pfit changes do
+	 * affect it and so it must be updated for fastsets.
+	 * HSW/BDW only really need this here for fastboot, after
+	 * that the value should not change without a full modeset.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 9 ||
+	    IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		hsw_set_linetime_wm(new_crtc_state);
+
+	if (new_crtc_state->seamless_m_n)
+		intel_cpu_transcoder_set_m1_n1(crtc, new_crtc_state->cpu_transcoder,
+					       &new_crtc_state->dp_m_n);
+}
+
+static void commit_pipe_pre_planes(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	bool modeset = intel_crtc_needs_modeset(new_crtc_state);
+
+	/*
+	 * During modesets pipe configuration was programmed as the
+	 * CRTC was enabled.
+	 */
+	if (!modeset) {
+		if (new_crtc_state->uapi.color_mgmt_changed ||
+		    new_crtc_state->update_pipe)
+			intel_color_commit_arm(new_crtc_state);
+
+		if (DISPLAY_VER(dev_priv) >= 9 || IS_BROADWELL(dev_priv))
+			bdw_set_pipemisc(new_crtc_state);
+
+		if (new_crtc_state->update_pipe)
+			intel_pipe_fastset(old_crtc_state, new_crtc_state);
+	}
+
+	intel_psr2_program_trans_man_trk_ctl(new_crtc_state);
+
+	intel_atomic_update_watermarks(state, crtc);
+}
+
+static void commit_pipe_post_planes(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	/*
+	 * Disable the scaler(s) after the plane(s) so that we don't
+	 * get a catastrophic underrun even if the two operations
+	 * end up happening in two different frames.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 9 &&
+	    !intel_crtc_needs_modeset(new_crtc_state))
+		skl_detach_scalers(new_crtc_state);
+}
+
+static void intel_enable_crtc(struct intel_atomic_state *state,
+			      struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (!intel_crtc_needs_modeset(new_crtc_state))
+		return;
+
+	intel_crtc_update_active_timings(new_crtc_state);
+
+	dev_priv->display.funcs.display->crtc_enable(state, crtc);
+
+	if (intel_crtc_is_bigjoiner_slave(new_crtc_state))
+		return;
+
+	/* vblanks work again, re-enable pipe CRC. */
+	intel_crtc_enable_pipe_crc(crtc);
+}
+
+static void intel_update_crtc(struct intel_atomic_state *state,
+			      struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	bool modeset = intel_crtc_needs_modeset(new_crtc_state);
+
+	if (!modeset) {
+		if (new_crtc_state->preload_luts &&
+		    (new_crtc_state->uapi.color_mgmt_changed ||
+		     new_crtc_state->update_pipe))
+			intel_color_load_luts(new_crtc_state);
+
+		intel_pre_plane_update(state, crtc);
+
+		if (new_crtc_state->update_pipe)
+			intel_encoders_update_pipe(state, crtc);
+
+		if (DISPLAY_VER(i915) >= 11 &&
+		    new_crtc_state->update_pipe)
+			icl_set_pipe_chicken(new_crtc_state);
+	}
+
+	intel_fbc_update(state, crtc);
+
+	if (!modeset &&
+	    (new_crtc_state->uapi.color_mgmt_changed ||
+	     new_crtc_state->update_pipe))
+		intel_color_commit_noarm(new_crtc_state);
+
+	intel_crtc_planes_update_noarm(state, crtc);
+
+	/* Perform vblank evasion around commit operation */
+	intel_pipe_update_start(new_crtc_state);
+
+	commit_pipe_pre_planes(state, crtc);
+
+	intel_crtc_planes_update_arm(state, crtc);
+
+	commit_pipe_post_planes(state, crtc);
+
+	intel_pipe_update_end(new_crtc_state);
+
+	/*
+	 * We usually enable FIFO underrun interrupts as part of the
+	 * CRTC enable sequence during modesets.  But when we inherit a
+	 * valid pipe configuration from the BIOS we need to take care
+	 * of enabling them on the CRTC's first fastset.
+	 */
+	if (new_crtc_state->update_pipe && !modeset &&
+	    old_crtc_state->inherited)
+		intel_crtc_arm_fifo_underrun(crtc, new_crtc_state);
+}
+
+static void intel_old_crtc_state_disables(struct intel_atomic_state *state,
+					  struct intel_crtc_state *old_crtc_state,
+					  struct intel_crtc_state *new_crtc_state,
+					  struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+
+	/*
+	 * We need to disable pipe CRC before disabling the pipe,
+	 * or we race against vblank off.
+	 */
+	intel_crtc_disable_pipe_crc(crtc);
+
+	dev_priv->display.funcs.display->crtc_disable(state, crtc);
+	crtc->active = false;
+	intel_fbc_disable(crtc);
+	intel_disable_shared_dpll(old_crtc_state);
+
+	/* FIXME unify this for all platforms */
+	if (!new_crtc_state->hw.active &&
+	    !HAS_GMCH(dev_priv))
+		intel_initial_watermarks(state, crtc);
+}
+
+static void intel_commit_modeset_disables(struct intel_atomic_state *state)
+{
+	struct intel_crtc_state *new_crtc_state, *old_crtc_state;
+	struct intel_crtc *crtc;
+	u32 handled = 0;
+	int i;
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		if (!intel_crtc_needs_modeset(new_crtc_state))
+			continue;
+
+		if (!old_crtc_state->hw.active)
+			continue;
+
+		intel_pre_plane_update(state, crtc);
+		intel_crtc_disable_planes(state, crtc);
+	}
+
+	/* Only disable port sync and MST slaves */
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		if (!intel_crtc_needs_modeset(new_crtc_state))
+			continue;
+
+		if (!old_crtc_state->hw.active)
+			continue;
+
+		/* In case of Transcoder port Sync master slave CRTCs can be
+		 * assigned in any order and we need to make sure that
+		 * slave CRTCs are disabled first and then master CRTC since
+		 * Slave vblanks are masked till Master Vblanks.
+		 */
+		if (!is_trans_port_sync_slave(old_crtc_state) &&
+		    !intel_dp_mst_is_slave_trans(old_crtc_state) &&
+		    !intel_crtc_is_bigjoiner_slave(old_crtc_state))
+			continue;
+
+		intel_old_crtc_state_disables(state, old_crtc_state,
+					      new_crtc_state, crtc);
+		handled |= BIT(crtc->pipe);
+	}
+
+	/* Disable everything else left on */
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		if (!intel_crtc_needs_modeset(new_crtc_state) ||
+		    (handled & BIT(crtc->pipe)))
+			continue;
+
+		if (!old_crtc_state->hw.active)
+			continue;
+
+		intel_old_crtc_state_disables(state, old_crtc_state,
+					      new_crtc_state, crtc);
+	}
+}
+
+static void intel_commit_modeset_enables(struct intel_atomic_state *state)
+{
+	struct intel_crtc_state *new_crtc_state;
+	struct intel_crtc *crtc;
+	int i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		if (!new_crtc_state->hw.active)
+			continue;
+
+		intel_enable_crtc(state, crtc);
+		intel_update_crtc(state, crtc);
+	}
+}
+
+static void skl_commit_modeset_enables(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc *crtc;
+	struct intel_crtc_state *old_crtc_state, *new_crtc_state;
+	struct skl_ddb_entry entries[I915_MAX_PIPES] = {};
+	u8 update_pipes = 0, modeset_pipes = 0;
+	int i;
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
+		enum pipe pipe = crtc->pipe;
+
+		if (!new_crtc_state->hw.active)
+			continue;
+
+		/* ignore allocations for crtc's that have been turned off. */
+		if (!intel_crtc_needs_modeset(new_crtc_state)) {
+			entries[pipe] = old_crtc_state->wm.skl.ddb;
+			update_pipes |= BIT(pipe);
+		} else {
+			modeset_pipes |= BIT(pipe);
+		}
+	}
+
+	/*
+	 * Whenever the number of active pipes changes, we need to make sure we
+	 * update the pipes in the right order so that their ddb allocations
+	 * never overlap with each other between CRTC updates. Otherwise we'll
+	 * cause pipe underruns and other bad stuff.
+	 *
+	 * So first lets enable all pipes that do not need a fullmodeset as
+	 * those don't have any external dependency.
+	 */
+	while (update_pipes) {
+		for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+						    new_crtc_state, i) {
+			enum pipe pipe = crtc->pipe;
+
+			if ((update_pipes & BIT(pipe)) == 0)
+				continue;
+
+			if (skl_ddb_allocation_overlaps(&new_crtc_state->wm.skl.ddb,
+							entries, I915_MAX_PIPES, pipe))
+				continue;
+
+			entries[pipe] = new_crtc_state->wm.skl.ddb;
+			update_pipes &= ~BIT(pipe);
+
+			intel_update_crtc(state, crtc);
+
+			/*
+			 * If this is an already active pipe, it's DDB changed,
+			 * and this isn't the last pipe that needs updating
+			 * then we need to wait for a vblank to pass for the
+			 * new ddb allocation to take effect.
+			 */
+			if (!skl_ddb_entry_equal(&new_crtc_state->wm.skl.ddb,
+						 &old_crtc_state->wm.skl.ddb) &&
+			    (update_pipes | modeset_pipes))
+				intel_crtc_wait_for_next_vblank(crtc);
+		}
+	}
+
+	update_pipes = modeset_pipes;
+
+	/*
+	 * Enable all pipes that needs a modeset and do not depends on other
+	 * pipes
+	 */
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		enum pipe pipe = crtc->pipe;
+
+		if ((modeset_pipes & BIT(pipe)) == 0)
+			continue;
+
+		if (intel_dp_mst_is_slave_trans(new_crtc_state) ||
+		    is_trans_port_sync_master(new_crtc_state) ||
+		    intel_crtc_is_bigjoiner_master(new_crtc_state))
+			continue;
+
+		modeset_pipes &= ~BIT(pipe);
+
+		intel_enable_crtc(state, crtc);
+	}
+
+	/*
+	 * Then we enable all remaining pipes that depend on other
+	 * pipes: MST slaves and port sync masters, big joiner master
+	 */
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		enum pipe pipe = crtc->pipe;
+
+		if ((modeset_pipes & BIT(pipe)) == 0)
+			continue;
+
+		modeset_pipes &= ~BIT(pipe);
+
+		intel_enable_crtc(state, crtc);
+	}
+
+	/*
+	 * Finally we do the plane updates/etc. for all pipes that got enabled.
+	 */
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		enum pipe pipe = crtc->pipe;
+
+		if ((update_pipes & BIT(pipe)) == 0)
+			continue;
+
+		drm_WARN_ON(&dev_priv->drm, skl_ddb_allocation_overlaps(&new_crtc_state->wm.skl.ddb,
+									entries, I915_MAX_PIPES, pipe));
+
+		entries[pipe] = new_crtc_state->wm.skl.ddb;
+		update_pipes &= ~BIT(pipe);
+
+		intel_update_crtc(state, crtc);
+	}
+
+	drm_WARN_ON(&dev_priv->drm, modeset_pipes);
+	drm_WARN_ON(&dev_priv->drm, update_pipes);
+}
+
+static void intel_atomic_helper_free_state(struct drm_i915_private *dev_priv)
+{
+	struct intel_atomic_state *state, *next;
+	struct llist_node *freed;
+
+	freed = llist_del_all(&dev_priv->display.atomic_helper.free_list);
+	llist_for_each_entry_safe(state, next, freed, freed)
+		drm_atomic_state_put(&state->base);
+}
+
+static void intel_atomic_helper_free_state_worker(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, typeof(*dev_priv), display.atomic_helper.free_work);
+
+	intel_atomic_helper_free_state(dev_priv);
+}
+
+static void intel_atomic_commit_fence_wait(struct intel_atomic_state *intel_state)
+{
+	struct wait_queue_entry wait_fence, wait_reset;
+	struct drm_i915_private *dev_priv = to_i915(intel_state->base.dev);
+
+	init_wait_entry(&wait_fence, 0);
+	init_wait_entry(&wait_reset, 0);
+	for (;;) {
+		prepare_to_wait(&intel_state->commit_ready.wait,
+				&wait_fence, TASK_UNINTERRUPTIBLE);
+		prepare_to_wait(bit_waitqueue(&to_gt(dev_priv)->reset.flags,
+					      I915_RESET_MODESET),
+				&wait_reset, TASK_UNINTERRUPTIBLE);
+
+
+		if (i915_sw_fence_done(&intel_state->commit_ready) ||
+		    test_bit(I915_RESET_MODESET, &to_gt(dev_priv)->reset.flags))
+			break;
+
+		schedule();
+	}
+	finish_wait(&intel_state->commit_ready.wait, &wait_fence);
+	finish_wait(bit_waitqueue(&to_gt(dev_priv)->reset.flags,
+				  I915_RESET_MODESET),
+		    &wait_reset);
+}
+
+static void intel_cleanup_dsbs(struct intel_atomic_state *state)
+{
+	struct intel_crtc_state *old_crtc_state, *new_crtc_state;
+	struct intel_crtc *crtc;
+	int i;
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i)
+		intel_dsb_cleanup(old_crtc_state);
+}
+
+static void intel_atomic_cleanup_work(struct work_struct *work)
+{
+	struct intel_atomic_state *state =
+		container_of(work, struct intel_atomic_state, base.commit_work);
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+
+	intel_cleanup_dsbs(state);
+	drm_atomic_helper_cleanup_planes(&i915->drm, &state->base);
+	drm_atomic_helper_commit_cleanup_done(&state->base);
+	drm_atomic_state_put(&state->base);
+
+	intel_atomic_helper_free_state(i915);
+}
+
+static void intel_atomic_prepare_plane_clear_colors(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_plane *plane;
+	struct intel_plane_state *plane_state;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		struct drm_framebuffer *fb = plane_state->hw.fb;
+		int cc_plane;
+		int ret;
+
+		if (!fb)
+			continue;
+
+		cc_plane = intel_fb_rc_ccs_cc_plane(fb);
+		if (cc_plane < 0)
+			continue;
+
+		/*
+		 * The layout of the fast clear color value expected by HW
+		 * (the DRM ABI requiring this value to be located in fb at
+		 * offset 0 of cc plane, plane #2 previous generations or
+		 * plane #1 for flat ccs):
+		 * - 4 x 4 bytes per-channel value
+		 *   (in surface type specific float/int format provided by the fb user)
+		 * - 8 bytes native color value used by the display
+		 *   (converted/written by GPU during a fast clear operation using the
+		 *    above per-channel values)
+		 *
+		 * The commit's FB prepare hook already ensured that FB obj is pinned and the
+		 * caller made sure that the object is synced wrt. the related color clear value
+		 * GPU write on it.
+		 */
+		ret = i915_gem_object_read_from_page(intel_fb_obj(fb),
+						     fb->offsets[cc_plane] + 16,
+						     &plane_state->ccval,
+						     sizeof(plane_state->ccval));
+		/* The above could only fail if the FB obj has an unexpected backing store type. */
+		drm_WARN_ON(&i915->drm, ret);
+	}
+}
+
+static void intel_atomic_commit_tail(struct intel_atomic_state *state)
+{
+	struct drm_device *dev = state->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_crtc_state *new_crtc_state, *old_crtc_state;
+	struct intel_crtc *crtc;
+	struct intel_power_domain_mask put_domains[I915_MAX_PIPES] = {};
+	intel_wakeref_t wakeref = 0;
+	int i;
+
+	intel_atomic_commit_fence_wait(state);
+
+	drm_atomic_helper_wait_for_dependencies(&state->base);
+	drm_dp_mst_atomic_wait_for_dependencies(&state->base);
+
+	if (state->modeset)
+		wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_MODESET);
+
+	intel_atomic_prepare_plane_clear_colors(state);
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		if (intel_crtc_needs_modeset(new_crtc_state) ||
+		    new_crtc_state->update_pipe) {
+			intel_modeset_get_crtc_power_domains(new_crtc_state, &put_domains[crtc->pipe]);
+		}
+	}
+
+	intel_commit_modeset_disables(state);
+
+	/* FIXME: Eventually get rid of our crtc->config pointer */
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i)
+		crtc->config = new_crtc_state;
+
+	if (state->modeset) {
+		drm_atomic_helper_update_legacy_modeset_state(dev, &state->base);
+
+		intel_set_cdclk_pre_plane_update(state);
+
+		intel_modeset_verify_disabled(dev_priv, state);
+	}
+
+	intel_sagv_pre_plane_update(state);
+
+	/* Complete the events for pipes that have now been disabled */
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		bool modeset = intel_crtc_needs_modeset(new_crtc_state);
+
+		/* Complete events for now disable pipes here. */
+		if (modeset && !new_crtc_state->hw.active && new_crtc_state->uapi.event) {
+			spin_lock_irq(&dev->event_lock);
+			drm_crtc_send_vblank_event(&crtc->base,
+						   new_crtc_state->uapi.event);
+			spin_unlock_irq(&dev->event_lock);
+
+			new_crtc_state->uapi.event = NULL;
+		}
+	}
+
+	intel_encoders_update_prepare(state);
+
+	intel_dbuf_pre_plane_update(state);
+	intel_mbus_dbox_update(state);
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		if (new_crtc_state->do_async_flip)
+			intel_crtc_enable_flip_done(state, crtc);
+	}
+
+	/* Now enable the clocks, plane, pipe, and connectors that we set up. */
+	dev_priv->display.funcs.display->commit_modeset_enables(state);
+
+	intel_encoders_update_complete(state);
+
+	if (state->modeset)
+		intel_set_cdclk_post_plane_update(state);
+
+	intel_wait_for_vblank_workers(state);
+
+	/* FIXME: We should call drm_atomic_helper_commit_hw_done() here
+	 * already, but still need the state for the delayed optimization. To
+	 * fix this:
+	 * - wrap the optimization/post_plane_update stuff into a per-crtc work.
+	 * - schedule that vblank worker _before_ calling hw_done
+	 * - at the start of commit_tail, cancel it _synchrously
+	 * - switch over to the vblank wait helper in the core after that since
+	 *   we don't need out special handling any more.
+	 */
+	drm_atomic_helper_wait_for_flip_done(dev, &state->base);
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		if (new_crtc_state->do_async_flip)
+			intel_crtc_disable_flip_done(state, crtc);
+	}
+
+	/*
+	 * Now that the vblank has passed, we can go ahead and program the
+	 * optimal watermarks on platforms that need two-step watermark
+	 * programming.
+	 *
+	 * TODO: Move this (and other cleanup) to an async worker eventually.
+	 */
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		/*
+		 * Gen2 reports pipe underruns whenever all planes are disabled.
+		 * So re-enable underrun reporting after some planes get enabled.
+		 *
+		 * We do this before .optimize_watermarks() so that we have a
+		 * chance of catching underruns with the intermediate watermarks
+		 * vs. the new plane configuration.
+		 */
+		if (DISPLAY_VER(dev_priv) == 2 && planes_enabling(old_crtc_state, new_crtc_state))
+			intel_set_cpu_fifo_underrun_reporting(dev_priv, crtc->pipe, true);
+
+		intel_optimize_watermarks(state, crtc);
+	}
+
+	intel_dbuf_post_plane_update(state);
+	intel_psr_post_plane_update(state);
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
+		intel_post_plane_update(state, crtc);
+
+		intel_modeset_put_crtc_power_domains(crtc, &put_domains[crtc->pipe]);
+
+		intel_modeset_verify_crtc(crtc, state, old_crtc_state, new_crtc_state);
+
+		/*
+		 * DSB cleanup is done in cleanup_work aligning with framebuffer
+		 * cleanup. So copy and reset the dsb structure to sync with
+		 * commit_done and later do dsb cleanup in cleanup_work.
+		 */
+		old_crtc_state->dsb = fetch_and_zero(&new_crtc_state->dsb);
+	}
+
+	/* Underruns don't always raise interrupts, so check manually */
+	intel_check_cpu_fifo_underruns(dev_priv);
+	intel_check_pch_fifo_underruns(dev_priv);
+
+	if (state->modeset)
+		intel_verify_planes(state);
+
+	intel_sagv_post_plane_update(state);
+
+	drm_atomic_helper_commit_hw_done(&state->base);
+
+	if (state->modeset) {
+		/* As one of the primary mmio accessors, KMS has a high
+		 * likelihood of triggering bugs in unclaimed access. After we
+		 * finish modesetting, see if an error has been flagged, and if
+		 * so enable debugging for the next modeset - and hope we catch
+		 * the culprit.
+		 */
+		intel_uncore_arm_unclaimed_mmio_detection(&dev_priv->uncore);
+		intel_display_power_put(dev_priv, POWER_DOMAIN_MODESET, wakeref);
+	}
+	intel_runtime_pm_put(&dev_priv->runtime_pm, state->wakeref);
+
+	/*
+	 * Defer the cleanup of the old state to a separate worker to not
+	 * impede the current task (userspace for blocking modesets) that
+	 * are executed inline. For out-of-line asynchronous modesets/flips,
+	 * deferring to a new worker seems overkill, but we would place a
+	 * schedule point (cond_resched()) here anyway to keep latencies
+	 * down.
+	 */
+	INIT_WORK(&state->base.commit_work, intel_atomic_cleanup_work);
+	queue_work(system_highpri_wq, &state->base.commit_work);
+}
+
+static void intel_atomic_commit_work(struct work_struct *work)
+{
+	struct intel_atomic_state *state =
+		container_of(work, struct intel_atomic_state, base.commit_work);
+
+	intel_atomic_commit_tail(state);
+}
+
+static int
+intel_atomic_commit_ready(struct i915_sw_fence *fence,
+			  enum i915_sw_fence_notify notify)
+{
+	struct intel_atomic_state *state =
+		container_of(fence, struct intel_atomic_state, commit_ready);
+
+	switch (notify) {
+	case FENCE_COMPLETE:
+		/* we do blocking waits in the worker, nothing to do here */
+		break;
+	case FENCE_FREE:
+		{
+			struct intel_atomic_helper *helper =
+				&to_i915(state->base.dev)->display.atomic_helper;
+
+			if (llist_add(&state->freed, &helper->free_list))
+				schedule_work(&helper->free_work);
+			break;
+		}
+	}
+
+	return NOTIFY_DONE;
+}
+
+static void intel_atomic_track_fbs(struct intel_atomic_state *state)
+{
+	struct intel_plane_state *old_plane_state, *new_plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	for_each_oldnew_intel_plane_in_state(state, plane, old_plane_state,
+					     new_plane_state, i)
+		intel_frontbuffer_track(to_intel_frontbuffer(old_plane_state->hw.fb),
+					to_intel_frontbuffer(new_plane_state->hw.fb),
+					plane->frontbuffer_bit);
+}
+
+static int intel_atomic_commit(struct drm_device *dev,
+			       struct drm_atomic_state *_state,
+			       bool nonblock)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(_state);
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	int ret = 0;
+
+	state->wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+
+	drm_atomic_state_get(&state->base);
+	i915_sw_fence_init(&state->commit_ready,
+			   intel_atomic_commit_ready);
+
+	/*
+	 * The intel_legacy_cursor_update() fast path takes care
+	 * of avoiding the vblank waits for simple cursor
+	 * movement and flips. For cursor on/off and size changes,
+	 * we want to perform the vblank waits so that watermark
+	 * updates happen during the correct frames. Gen9+ have
+	 * double buffered watermarks and so shouldn't need this.
+	 *
+	 * Unset state->legacy_cursor_update before the call to
+	 * drm_atomic_helper_setup_commit() because otherwise
+	 * drm_atomic_helper_wait_for_flip_done() is a noop and
+	 * we get FIFO underruns because we didn't wait
+	 * for vblank.
+	 *
+	 * FIXME doing watermarks and fb cleanup from a vblank worker
+	 * (assuming we had any) would solve these problems.
+	 */
+	if (DISPLAY_VER(dev_priv) < 9 && state->base.legacy_cursor_update) {
+		struct intel_crtc_state *new_crtc_state;
+		struct intel_crtc *crtc;
+		int i;
+
+		for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i)
+			if (new_crtc_state->wm.need_postvbl_update ||
+			    new_crtc_state->update_wm_post)
+				state->base.legacy_cursor_update = false;
+	}
+
+	ret = intel_atomic_prepare_commit(state);
+	if (ret) {
+		drm_dbg_atomic(&dev_priv->drm,
+			       "Preparing state failed with %i\n", ret);
+		i915_sw_fence_commit(&state->commit_ready);
+		intel_runtime_pm_put(&dev_priv->runtime_pm, state->wakeref);
+		return ret;
+	}
+
+	ret = drm_atomic_helper_setup_commit(&state->base, nonblock);
+	if (!ret)
+		ret = drm_atomic_helper_swap_state(&state->base, true);
+	if (!ret)
+		intel_atomic_swap_global_state(state);
+
+	if (ret) {
+		struct intel_crtc_state *new_crtc_state;
+		struct intel_crtc *crtc;
+		int i;
+
+		i915_sw_fence_commit(&state->commit_ready);
+
+		for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i)
+			intel_dsb_cleanup(new_crtc_state);
+
+		drm_atomic_helper_cleanup_planes(dev, &state->base);
+		intel_runtime_pm_put(&dev_priv->runtime_pm, state->wakeref);
+		return ret;
+	}
+	intel_shared_dpll_swap_state(state);
+	intel_atomic_track_fbs(state);
+
+	drm_atomic_state_get(&state->base);
+	INIT_WORK(&state->base.commit_work, intel_atomic_commit_work);
+
+	i915_sw_fence_commit(&state->commit_ready);
+	if (nonblock && state->modeset) {
+		queue_work(dev_priv->display.wq.modeset, &state->base.commit_work);
+	} else if (nonblock) {
+		queue_work(dev_priv->display.wq.flip, &state->base.commit_work);
+	} else {
+		if (state->modeset)
+			flush_workqueue(dev_priv->display.wq.modeset);
+		intel_atomic_commit_tail(state);
+	}
+
+	return 0;
+}
+
+/**
+ * intel_plane_destroy - destroy a plane
+ * @plane: plane to destroy
+ *
+ * Common destruction function for all types of planes (primary, cursor,
+ * sprite).
+ */
+void intel_plane_destroy(struct drm_plane *plane)
+{
+	drm_plane_cleanup(plane);
+	kfree(to_intel_plane(plane));
+}
+
+static void intel_plane_possible_crtcs_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_plane *plane;
+
+	for_each_intel_plane(&dev_priv->drm, plane) {
+		struct intel_crtc *crtc = intel_crtc_for_pipe(dev_priv,
+							      plane->pipe);
+
+		plane->base.possible_crtcs = drm_crtc_mask(&crtc->base);
+	}
+}
+
+
+int intel_get_pipe_from_crtc_id_ioctl(struct drm_device *dev, void *data,
+				      struct drm_file *file)
+{
+	struct drm_i915_get_pipe_from_crtc_id *pipe_from_crtc_id = data;
+	struct drm_crtc *drmmode_crtc;
+	struct intel_crtc *crtc;
+
+	drmmode_crtc = drm_crtc_find(dev, file, pipe_from_crtc_id->crtc_id);
+	if (!drmmode_crtc)
+		return -ENOENT;
+
+	crtc = to_intel_crtc(drmmode_crtc);
+	pipe_from_crtc_id->pipe = crtc->pipe;
+
+	return 0;
+}
+
+static u32 intel_encoder_possible_clones(struct intel_encoder *encoder)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct intel_encoder *source_encoder;
+	u32 possible_clones = 0;
+
+	for_each_intel_encoder(dev, source_encoder) {
+		if (encoders_cloneable(encoder, source_encoder))
+			possible_clones |= drm_encoder_mask(&source_encoder->base);
+	}
+
+	return possible_clones;
+}
+
+static u32 intel_encoder_possible_crtcs(struct intel_encoder *encoder)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct intel_crtc *crtc;
+	u32 possible_crtcs = 0;
+
+	for_each_intel_crtc_in_pipe_mask(dev, crtc, encoder->pipe_mask)
+		possible_crtcs |= drm_crtc_mask(&crtc->base);
+
+	return possible_crtcs;
+}
+
+static bool ilk_has_edp_a(struct drm_i915_private *dev_priv)
+{
+	if (!IS_MOBILE(dev_priv))
+		return false;
+
+	if ((intel_de_read(dev_priv, DP_A) & DP_DETECTED) == 0)
+		return false;
+
+	if (IS_IRONLAKE(dev_priv) && (intel_de_read(dev_priv, FUSE_STRAP) & ILK_eDP_A_DISABLE))
+		return false;
+
+	return true;
+}
+
+static bool intel_ddi_crt_present(struct drm_i915_private *dev_priv)
+{
+	if (DISPLAY_VER(dev_priv) >= 9)
+		return false;
+
+	if (IS_HSW_ULT(dev_priv) || IS_BDW_ULT(dev_priv))
+		return false;
+
+	if (HAS_PCH_LPT_H(dev_priv) &&
+	    intel_de_read(dev_priv, SFUSE_STRAP) & SFUSE_STRAP_CRT_DISABLED)
+		return false;
+
+	/* DDI E can't be used if DDI A requires 4 lanes */
+	if (intel_de_read(dev_priv, DDI_BUF_CTL(PORT_A)) & DDI_A_4_LANES)
+		return false;
+
+	if (!dev_priv->display.vbt.int_crt_support)
+		return false;
+
+	return true;
+}
+
+static void intel_setup_outputs(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+	bool dpd_is_edp = false;
+
+	intel_pps_unlock_regs_wa(dev_priv);
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	if (IS_DG2(dev_priv)) {
+		intel_ddi_init(dev_priv, PORT_A);
+		intel_ddi_init(dev_priv, PORT_B);
+		intel_ddi_init(dev_priv, PORT_C);
+		intel_ddi_init(dev_priv, PORT_D_XELPD);
+		intel_ddi_init(dev_priv, PORT_TC1);
+	} else if (IS_ALDERLAKE_P(dev_priv)) {
+		intel_ddi_init(dev_priv, PORT_A);
+		intel_ddi_init(dev_priv, PORT_B);
+		intel_ddi_init(dev_priv, PORT_TC1);
+		intel_ddi_init(dev_priv, PORT_TC2);
+		intel_ddi_init(dev_priv, PORT_TC3);
+		intel_ddi_init(dev_priv, PORT_TC4);
+		icl_dsi_init(dev_priv);
+	} else if (IS_ALDERLAKE_S(dev_priv)) {
+		intel_ddi_init(dev_priv, PORT_A);
+		intel_ddi_init(dev_priv, PORT_TC1);
+		intel_ddi_init(dev_priv, PORT_TC2);
+		intel_ddi_init(dev_priv, PORT_TC3);
+		intel_ddi_init(dev_priv, PORT_TC4);
+	} else if (IS_DG1(dev_priv) || IS_ROCKETLAKE(dev_priv)) {
+		intel_ddi_init(dev_priv, PORT_A);
+		intel_ddi_init(dev_priv, PORT_B);
+		intel_ddi_init(dev_priv, PORT_TC1);
+		intel_ddi_init(dev_priv, PORT_TC2);
+	} else if (DISPLAY_VER(dev_priv) >= 12) {
+		intel_ddi_init(dev_priv, PORT_A);
+		intel_ddi_init(dev_priv, PORT_B);
+		intel_ddi_init(dev_priv, PORT_TC1);
+		intel_ddi_init(dev_priv, PORT_TC2);
+		intel_ddi_init(dev_priv, PORT_TC3);
+		intel_ddi_init(dev_priv, PORT_TC4);
+		intel_ddi_init(dev_priv, PORT_TC5);
+		intel_ddi_init(dev_priv, PORT_TC6);
+		icl_dsi_init(dev_priv);
+	} else if (IS_JSL_EHL(dev_priv)) {
+		intel_ddi_init(dev_priv, PORT_A);
+		intel_ddi_init(dev_priv, PORT_B);
+		intel_ddi_init(dev_priv, PORT_C);
+		intel_ddi_init(dev_priv, PORT_D);
+		icl_dsi_init(dev_priv);
+	} else if (DISPLAY_VER(dev_priv) == 11) {
+		intel_ddi_init(dev_priv, PORT_A);
+		intel_ddi_init(dev_priv, PORT_B);
+		intel_ddi_init(dev_priv, PORT_C);
+		intel_ddi_init(dev_priv, PORT_D);
+		intel_ddi_init(dev_priv, PORT_E);
+		intel_ddi_init(dev_priv, PORT_F);
+		icl_dsi_init(dev_priv);
+	} else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		intel_ddi_init(dev_priv, PORT_A);
+		intel_ddi_init(dev_priv, PORT_B);
+		intel_ddi_init(dev_priv, PORT_C);
+		vlv_dsi_init(dev_priv);
+	} else if (DISPLAY_VER(dev_priv) >= 9) {
+		intel_ddi_init(dev_priv, PORT_A);
+		intel_ddi_init(dev_priv, PORT_B);
+		intel_ddi_init(dev_priv, PORT_C);
+		intel_ddi_init(dev_priv, PORT_D);
+		intel_ddi_init(dev_priv, PORT_E);
+	} else if (HAS_DDI(dev_priv)) {
+		u32 found;
+
+		if (intel_ddi_crt_present(dev_priv))
+			intel_crt_init(dev_priv);
+
+		/* Haswell uses DDI functions to detect digital outputs. */
+		found = intel_de_read(dev_priv, DDI_BUF_CTL(PORT_A)) & DDI_INIT_DISPLAY_DETECTED;
+		if (found)
+			intel_ddi_init(dev_priv, PORT_A);
+
+		found = intel_de_read(dev_priv, SFUSE_STRAP);
+		if (found & SFUSE_STRAP_DDIB_DETECTED)
+			intel_ddi_init(dev_priv, PORT_B);
+		if (found & SFUSE_STRAP_DDIC_DETECTED)
+			intel_ddi_init(dev_priv, PORT_C);
+		if (found & SFUSE_STRAP_DDID_DETECTED)
+			intel_ddi_init(dev_priv, PORT_D);
+		if (found & SFUSE_STRAP_DDIF_DETECTED)
+			intel_ddi_init(dev_priv, PORT_F);
+	} else if (HAS_PCH_SPLIT(dev_priv)) {
+		int found;
+
+		/*
+		 * intel_edp_init_connector() depends on this completing first,
+		 * to prevent the registration of both eDP and LVDS and the
+		 * incorrect sharing of the PPS.
+		 */
+		intel_lvds_init(dev_priv);
+		intel_crt_init(dev_priv);
+
+		dpd_is_edp = intel_dp_is_port_edp(dev_priv, PORT_D);
+
+		if (ilk_has_edp_a(dev_priv))
+			g4x_dp_init(dev_priv, DP_A, PORT_A);
+
+		if (intel_de_read(dev_priv, PCH_HDMIB) & SDVO_DETECTED) {
+			/* PCH SDVOB multiplex with HDMIB */
+			found = intel_sdvo_init(dev_priv, PCH_SDVOB, PORT_B);
+			if (!found)
+				g4x_hdmi_init(dev_priv, PCH_HDMIB, PORT_B);
+			if (!found && (intel_de_read(dev_priv, PCH_DP_B) & DP_DETECTED))
+				g4x_dp_init(dev_priv, PCH_DP_B, PORT_B);
+		}
+
+		if (intel_de_read(dev_priv, PCH_HDMIC) & SDVO_DETECTED)
+			g4x_hdmi_init(dev_priv, PCH_HDMIC, PORT_C);
+
+		if (!dpd_is_edp && intel_de_read(dev_priv, PCH_HDMID) & SDVO_DETECTED)
+			g4x_hdmi_init(dev_priv, PCH_HDMID, PORT_D);
+
+		if (intel_de_read(dev_priv, PCH_DP_C) & DP_DETECTED)
+			g4x_dp_init(dev_priv, PCH_DP_C, PORT_C);
+
+		if (intel_de_read(dev_priv, PCH_DP_D) & DP_DETECTED)
+			g4x_dp_init(dev_priv, PCH_DP_D, PORT_D);
+	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		bool has_edp, has_port;
+
+		if (IS_VALLEYVIEW(dev_priv) && dev_priv->display.vbt.int_crt_support)
+			intel_crt_init(dev_priv);
+
+		/*
+		 * The DP_DETECTED bit is the latched state of the DDC
+		 * SDA pin at boot. However since eDP doesn't require DDC
+		 * (no way to plug in a DP->HDMI dongle) the DDC pins for
+		 * eDP ports may have been muxed to an alternate function.
+		 * Thus we can't rely on the DP_DETECTED bit alone to detect
+		 * eDP ports. Consult the VBT as well as DP_DETECTED to
+		 * detect eDP ports.
+		 *
+		 * Sadly the straps seem to be missing sometimes even for HDMI
+		 * ports (eg. on Voyo V3 - CHT x7-Z8700), so check both strap
+		 * and VBT for the presence of the port. Additionally we can't
+		 * trust the port type the VBT declares as we've seen at least
+		 * HDMI ports that the VBT claim are DP or eDP.
+		 */
+		has_edp = intel_dp_is_port_edp(dev_priv, PORT_B);
+		has_port = intel_bios_is_port_present(dev_priv, PORT_B);
+		if (intel_de_read(dev_priv, VLV_DP_B) & DP_DETECTED || has_port)
+			has_edp &= g4x_dp_init(dev_priv, VLV_DP_B, PORT_B);
+		if ((intel_de_read(dev_priv, VLV_HDMIB) & SDVO_DETECTED || has_port) && !has_edp)
+			g4x_hdmi_init(dev_priv, VLV_HDMIB, PORT_B);
+
+		has_edp = intel_dp_is_port_edp(dev_priv, PORT_C);
+		has_port = intel_bios_is_port_present(dev_priv, PORT_C);
+		if (intel_de_read(dev_priv, VLV_DP_C) & DP_DETECTED || has_port)
+			has_edp &= g4x_dp_init(dev_priv, VLV_DP_C, PORT_C);
+		if ((intel_de_read(dev_priv, VLV_HDMIC) & SDVO_DETECTED || has_port) && !has_edp)
+			g4x_hdmi_init(dev_priv, VLV_HDMIC, PORT_C);
+
+		if (IS_CHERRYVIEW(dev_priv)) {
+			/*
+			 * eDP not supported on port D,
+			 * so no need to worry about it
+			 */
+			has_port = intel_bios_is_port_present(dev_priv, PORT_D);
+			if (intel_de_read(dev_priv, CHV_DP_D) & DP_DETECTED || has_port)
+				g4x_dp_init(dev_priv, CHV_DP_D, PORT_D);
+			if (intel_de_read(dev_priv, CHV_HDMID) & SDVO_DETECTED || has_port)
+				g4x_hdmi_init(dev_priv, CHV_HDMID, PORT_D);
+		}
+
+		vlv_dsi_init(dev_priv);
+	} else if (IS_PINEVIEW(dev_priv)) {
+		intel_lvds_init(dev_priv);
+		intel_crt_init(dev_priv);
+	} else if (IS_DISPLAY_VER(dev_priv, 3, 4)) {
+		bool found = false;
+
+		if (IS_MOBILE(dev_priv))
+			intel_lvds_init(dev_priv);
+
+		intel_crt_init(dev_priv);
+
+		if (intel_de_read(dev_priv, GEN3_SDVOB) & SDVO_DETECTED) {
+			drm_dbg_kms(&dev_priv->drm, "probing SDVOB\n");
+			found = intel_sdvo_init(dev_priv, GEN3_SDVOB, PORT_B);
+			if (!found && IS_G4X(dev_priv)) {
+				drm_dbg_kms(&dev_priv->drm,
+					    "probing HDMI on SDVOB\n");
+				g4x_hdmi_init(dev_priv, GEN4_HDMIB, PORT_B);
+			}
+
+			if (!found && IS_G4X(dev_priv))
+				g4x_dp_init(dev_priv, DP_B, PORT_B);
+		}
+
+		/* Before G4X SDVOC doesn't have its own detect register */
+
+		if (intel_de_read(dev_priv, GEN3_SDVOB) & SDVO_DETECTED) {
+			drm_dbg_kms(&dev_priv->drm, "probing SDVOC\n");
+			found = intel_sdvo_init(dev_priv, GEN3_SDVOC, PORT_C);
+		}
+
+		if (!found && (intel_de_read(dev_priv, GEN3_SDVOC) & SDVO_DETECTED)) {
+
+			if (IS_G4X(dev_priv)) {
+				drm_dbg_kms(&dev_priv->drm,
+					    "probing HDMI on SDVOC\n");
+				g4x_hdmi_init(dev_priv, GEN4_HDMIC, PORT_C);
+			}
+			if (IS_G4X(dev_priv))
+				g4x_dp_init(dev_priv, DP_C, PORT_C);
+		}
+
+		if (IS_G4X(dev_priv) && (intel_de_read(dev_priv, DP_D) & DP_DETECTED))
+			g4x_dp_init(dev_priv, DP_D, PORT_D);
+
+		if (SUPPORTS_TV(dev_priv))
+			intel_tv_init(dev_priv);
+	} else if (DISPLAY_VER(dev_priv) == 2) {
+		if (IS_I85X(dev_priv))
+			intel_lvds_init(dev_priv);
+
+		intel_crt_init(dev_priv);
+		intel_dvo_init(dev_priv);
+	}
+
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		encoder->base.possible_crtcs =
+			intel_encoder_possible_crtcs(encoder);
+		encoder->base.possible_clones =
+			intel_encoder_possible_clones(encoder);
+	}
+
+	intel_init_pch_refclk(dev_priv);
+
+	drm_helper_move_panel_connectors_to_head(&dev_priv->drm);
+}
+
+static int max_dotclock(struct drm_i915_private *i915)
+{
+	int max_dotclock = i915->max_dotclk_freq;
+
+	/* icl+ might use bigjoiner */
+	if (DISPLAY_VER(i915) >= 11)
+		max_dotclock *= 2;
+
+	return max_dotclock;
+}
+
+static enum drm_mode_status
+intel_mode_valid(struct drm_device *dev,
+		 const struct drm_display_mode *mode)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	int hdisplay_max, htotal_max;
+	int vdisplay_max, vtotal_max;
+
+	/*
+	 * Can't reject DBLSCAN here because Xorg ddxen can add piles
+	 * of DBLSCAN modes to the output's mode list when they detect
+	 * the scaling mode property on the connector. And they don't
+	 * ask the kernel to validate those modes in any way until
+	 * modeset time at which point the client gets a protocol error.
+	 * So in order to not upset those clients we silently ignore the
+	 * DBLSCAN flag on such connectors. For other connectors we will
+	 * reject modes with the DBLSCAN flag in encoder->compute_config().
+	 * And we always reject DBLSCAN modes in connector->mode_valid()
+	 * as we never want such modes on the connector's mode list.
+	 */
+
+	if (mode->vscan > 1)
+		return MODE_NO_VSCAN;
+
+	if (mode->flags & DRM_MODE_FLAG_HSKEW)
+		return MODE_H_ILLEGAL;
+
+	if (mode->flags & (DRM_MODE_FLAG_CSYNC |
+			   DRM_MODE_FLAG_NCSYNC |
+			   DRM_MODE_FLAG_PCSYNC))
+		return MODE_HSYNC;
+
+	if (mode->flags & (DRM_MODE_FLAG_BCAST |
+			   DRM_MODE_FLAG_PIXMUX |
+			   DRM_MODE_FLAG_CLKDIV2))
+		return MODE_BAD;
+
+	/*
+	 * Reject clearly excessive dotclocks early to
+	 * avoid having to worry about huge integers later.
+	 */
+	if (mode->clock > max_dotclock(dev_priv))
+		return MODE_CLOCK_HIGH;
+
+	/* Transcoder timing limits */
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		hdisplay_max = 16384;
+		vdisplay_max = 8192;
+		htotal_max = 16384;
+		vtotal_max = 8192;
+	} else if (DISPLAY_VER(dev_priv) >= 9 ||
+		   IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv)) {
+		hdisplay_max = 8192; /* FDI max 4096 handled elsewhere */
+		vdisplay_max = 4096;
+		htotal_max = 8192;
+		vtotal_max = 8192;
+	} else if (DISPLAY_VER(dev_priv) >= 3) {
+		hdisplay_max = 4096;
+		vdisplay_max = 4096;
+		htotal_max = 8192;
+		vtotal_max = 8192;
+	} else {
+		hdisplay_max = 2048;
+		vdisplay_max = 2048;
+		htotal_max = 4096;
+		vtotal_max = 4096;
+	}
+
+	if (mode->hdisplay > hdisplay_max ||
+	    mode->hsync_start > htotal_max ||
+	    mode->hsync_end > htotal_max ||
+	    mode->htotal > htotal_max)
+		return MODE_H_ILLEGAL;
+
+	if (mode->vdisplay > vdisplay_max ||
+	    mode->vsync_start > vtotal_max ||
+	    mode->vsync_end > vtotal_max ||
+	    mode->vtotal > vtotal_max)
+		return MODE_V_ILLEGAL;
+
+	return MODE_OK;
+}
+
+enum drm_mode_status intel_cpu_transcoder_mode_valid(struct drm_i915_private *dev_priv,
+						     const struct drm_display_mode *mode)
+{
+	/*
+	 * Additional transcoder timing limits,
+	 * excluding BXT/GLK DSI transcoders.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 5) {
+		if (mode->hdisplay < 64 ||
+		    mode->htotal - mode->hdisplay < 32)
+			return MODE_H_ILLEGAL;
+
+		if (mode->vtotal - mode->vdisplay < 5)
+			return MODE_V_ILLEGAL;
+	} else {
+		if (mode->htotal - mode->hdisplay < 32)
+			return MODE_H_ILLEGAL;
+
+		if (mode->vtotal - mode->vdisplay < 3)
+			return MODE_V_ILLEGAL;
+	}
+
+	/*
+	 * Cantiga+ cannot handle modes with a hsync front porch of 0.
+	 * WaPruneModeWithIncorrectHsyncOffset:ctg,elk,ilk,snb,ivb,vlv,hsw.
+	 */
+	if ((DISPLAY_VER(dev_priv) > 4 || IS_G4X(dev_priv)) &&
+	    mode->hsync_start == mode->hdisplay)
+		return MODE_H_ILLEGAL;
+
+	return MODE_OK;
+}
+
+enum drm_mode_status
+intel_mode_valid_max_plane_size(struct drm_i915_private *dev_priv,
+				const struct drm_display_mode *mode,
+				bool bigjoiner)
+{
+	int plane_width_max, plane_height_max;
+
+	/*
+	 * intel_mode_valid() should be
+	 * sufficient on older platforms.
+	 */
+	if (DISPLAY_VER(dev_priv) < 9)
+		return MODE_OK;
+
+	/*
+	 * Most people will probably want a fullscreen
+	 * plane so let's not advertize modes that are
+	 * too big for that.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		plane_width_max = 5120 << bigjoiner;
+		plane_height_max = 4320;
+	} else {
+		plane_width_max = 5120;
+		plane_height_max = 4096;
+	}
+
+	if (mode->hdisplay > plane_width_max)
+		return MODE_H_ILLEGAL;
+
+	if (mode->vdisplay > plane_height_max)
+		return MODE_V_ILLEGAL;
+
+	return MODE_OK;
+}
+
+static const struct drm_mode_config_funcs intel_mode_funcs = {
+	.fb_create = intel_user_framebuffer_create,
+	.get_format_info = intel_fb_get_format_info,
+	.output_poll_changed = intel_fbdev_output_poll_changed,
+	.mode_valid = intel_mode_valid,
+	.atomic_check = intel_atomic_check,
+	.atomic_commit = intel_atomic_commit,
+	.atomic_state_alloc = intel_atomic_state_alloc,
+	.atomic_state_clear = intel_atomic_state_clear,
+	.atomic_state_free = intel_atomic_state_free,
+};
+
+static const struct intel_display_funcs skl_display_funcs = {
+	.get_pipe_config = hsw_get_pipe_config,
+	.crtc_enable = hsw_crtc_enable,
+	.crtc_disable = hsw_crtc_disable,
+	.commit_modeset_enables = skl_commit_modeset_enables,
+	.get_initial_plane_config = skl_get_initial_plane_config,
+};
+
+static const struct intel_display_funcs ddi_display_funcs = {
+	.get_pipe_config = hsw_get_pipe_config,
+	.crtc_enable = hsw_crtc_enable,
+	.crtc_disable = hsw_crtc_disable,
+	.commit_modeset_enables = intel_commit_modeset_enables,
+	.get_initial_plane_config = i9xx_get_initial_plane_config,
+};
+
+static const struct intel_display_funcs pch_split_display_funcs = {
+	.get_pipe_config = ilk_get_pipe_config,
+	.crtc_enable = ilk_crtc_enable,
+	.crtc_disable = ilk_crtc_disable,
+	.commit_modeset_enables = intel_commit_modeset_enables,
+	.get_initial_plane_config = i9xx_get_initial_plane_config,
+};
+
+static const struct intel_display_funcs vlv_display_funcs = {
+	.get_pipe_config = i9xx_get_pipe_config,
+	.crtc_enable = valleyview_crtc_enable,
+	.crtc_disable = i9xx_crtc_disable,
+	.commit_modeset_enables = intel_commit_modeset_enables,
+	.get_initial_plane_config = i9xx_get_initial_plane_config,
+};
+
+static const struct intel_display_funcs i9xx_display_funcs = {
+	.get_pipe_config = i9xx_get_pipe_config,
+	.crtc_enable = i9xx_crtc_enable,
+	.crtc_disable = i9xx_crtc_disable,
+	.commit_modeset_enables = intel_commit_modeset_enables,
+	.get_initial_plane_config = i9xx_get_initial_plane_config,
+};
+
+/**
+ * intel_init_display_hooks - initialize the display modesetting hooks
+ * @dev_priv: device private
+ */
+void intel_init_display_hooks(struct drm_i915_private *dev_priv)
+{
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	intel_init_cdclk_hooks(dev_priv);
+	intel_audio_hooks_init(dev_priv);
+
+	intel_dpll_init_clock_hook(dev_priv);
+
+	if (DISPLAY_VER(dev_priv) >= 9) {
+		dev_priv->display.funcs.display = &skl_display_funcs;
+	} else if (HAS_DDI(dev_priv)) {
+		dev_priv->display.funcs.display = &ddi_display_funcs;
+	} else if (HAS_PCH_SPLIT(dev_priv)) {
+		dev_priv->display.funcs.display = &pch_split_display_funcs;
+	} else if (IS_CHERRYVIEW(dev_priv) ||
+		   IS_VALLEYVIEW(dev_priv)) {
+		dev_priv->display.funcs.display = &vlv_display_funcs;
+	} else {
+		dev_priv->display.funcs.display = &i9xx_display_funcs;
+	}
+
+	intel_fdi_init_hook(dev_priv);
+}
+
+void intel_modeset_init_hw(struct drm_i915_private *i915)
+{
+	struct intel_cdclk_state *cdclk_state;
+
+	if (!HAS_DISPLAY(i915))
+		return;
+
+	cdclk_state = to_intel_cdclk_state(i915->display.cdclk.obj.state);
+
+	intel_update_cdclk(i915);
+	intel_cdclk_dump_config(i915, &i915->display.cdclk.hw, "Current CDCLK");
+	cdclk_state->logical = cdclk_state->actual = i915->display.cdclk.hw;
+}
+
+static int sanitize_watermarks_add_affected(struct drm_atomic_state *state)
+{
+	struct drm_plane *plane;
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(state->dev, crtc) {
+		struct intel_crtc_state *crtc_state;
+
+		crtc_state = intel_atomic_get_crtc_state(state, crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+
+		if (crtc_state->hw.active) {
+			/*
+			 * Preserve the inherited flag to avoid
+			 * taking the full modeset path.
+			 */
+			crtc_state->inherited = true;
+		}
+	}
+
+	drm_for_each_plane(plane, state->dev) {
+		struct drm_plane_state *plane_state;
+
+		plane_state = drm_atomic_get_plane_state(state, plane);
+		if (IS_ERR(plane_state))
+			return PTR_ERR(plane_state);
+	}
+
+	return 0;
+}
+
+/*
+ * Calculate what we think the watermarks should be for the state we've read
+ * out of the hardware and then immediately program those watermarks so that
+ * we ensure the hardware settings match our internal state.
+ *
+ * We can calculate what we think WM's should be by creating a duplicate of the
+ * current state (which was constructed during hardware readout) and running it
+ * through the atomic check code to calculate new watermark values in the
+ * state object.
+ */
+static void sanitize_watermarks(struct drm_i915_private *dev_priv)
+{
+	struct drm_atomic_state *state;
+	struct intel_atomic_state *intel_state;
+	struct intel_crtc *crtc;
+	struct intel_crtc_state *crtc_state;
+	struct drm_modeset_acquire_ctx ctx;
+	int ret;
+	int i;
+
+	/* Only supported on platforms that use atomic watermark design */
+	if (!dev_priv->display.funcs.wm->optimize_watermarks)
+		return;
+
+	state = drm_atomic_state_alloc(&dev_priv->drm);
+	if (drm_WARN_ON(&dev_priv->drm, !state))
+		return;
+
+	intel_state = to_intel_atomic_state(state);
+
+	drm_modeset_acquire_init(&ctx, 0);
+
+retry:
+	state->acquire_ctx = &ctx;
+
+	/*
+	 * Hardware readout is the only time we don't want to calculate
+	 * intermediate watermarks (since we don't trust the current
+	 * watermarks).
+	 */
+	if (!HAS_GMCH(dev_priv))
+		intel_state->skip_intermediate_wm = true;
+
+	ret = sanitize_watermarks_add_affected(state);
+	if (ret)
+		goto fail;
+
+	ret = intel_atomic_check(&dev_priv->drm, state);
+	if (ret)
+		goto fail;
+
+	/* Write calculated watermark values back */
+	for_each_new_intel_crtc_in_state(intel_state, crtc, crtc_state, i) {
+		crtc_state->wm.need_postvbl_update = true;
+		intel_optimize_watermarks(intel_state, crtc);
+
+		to_intel_crtc_state(crtc->base.state)->wm = crtc_state->wm;
+	}
+
+fail:
+	if (ret == -EDEADLK) {
+		drm_atomic_state_clear(state);
+		drm_modeset_backoff(&ctx);
+		goto retry;
+	}
+
+	/*
+	 * If we fail here, it means that the hardware appears to be
+	 * programmed in a way that shouldn't be possible, given our
+	 * understanding of watermark requirements.  This might mean a
+	 * mistake in the hardware readout code or a mistake in the
+	 * watermark calculations for a given platform.  Raise a WARN
+	 * so that this is noticeable.
+	 *
+	 * If this actually happens, we'll have to just leave the
+	 * BIOS-programmed watermarks untouched and hope for the best.
+	 */
+	drm_WARN(&dev_priv->drm, ret,
+		 "Could not determine valid watermarks for inherited state\n");
+
+	drm_atomic_state_put(state);
+
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+}
+
+static int intel_initial_commit(struct drm_device *dev)
+{
+	struct drm_atomic_state *state = NULL;
+	struct drm_modeset_acquire_ctx ctx;
+	struct intel_crtc *crtc;
+	int ret = 0;
+
+	state = drm_atomic_state_alloc(dev);
+	if (!state)
+		return -ENOMEM;
+
+	drm_modeset_acquire_init(&ctx, 0);
+
+retry:
+	state->acquire_ctx = &ctx;
+
+	for_each_intel_crtc(dev, crtc) {
+		struct intel_crtc_state *crtc_state =
+			intel_atomic_get_crtc_state(state, crtc);
+
+		if (IS_ERR(crtc_state)) {
+			ret = PTR_ERR(crtc_state);
+			goto out;
+		}
+
+		if (crtc_state->hw.active) {
+			struct intel_encoder *encoder;
+
+			/*
+			 * We've not yet detected sink capabilities
+			 * (audio,infoframes,etc.) and thus we don't want to
+			 * force a full state recomputation yet. We want that to
+			 * happen only for the first real commit from userspace.
+			 * So preserve the inherited flag for the time being.
+			 */
+			crtc_state->inherited = true;
+
+			ret = drm_atomic_add_affected_planes(state, &crtc->base);
+			if (ret)
+				goto out;
+
+			/*
+			 * FIXME hack to force a LUT update to avoid the
+			 * plane update forcing the pipe gamma on without
+			 * having a proper LUT loaded. Remove once we
+			 * have readout for pipe gamma enable.
+			 */
+			crtc_state->uapi.color_mgmt_changed = true;
+
+			for_each_intel_encoder_mask(dev, encoder,
+						    crtc_state->uapi.encoder_mask) {
+				if (encoder->initial_fastset_check &&
+				    !encoder->initial_fastset_check(encoder, crtc_state)) {
+					ret = drm_atomic_add_affected_connectors(state,
+										 &crtc->base);
+					if (ret)
+						goto out;
+				}
+			}
+		}
+	}
+
+	ret = drm_atomic_commit(state);
+
+out:
+	if (ret == -EDEADLK) {
+		drm_atomic_state_clear(state);
+		drm_modeset_backoff(&ctx);
+		goto retry;
+	}
+
+	drm_atomic_state_put(state);
+
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+
+	return ret;
+}
+
+static const struct drm_mode_config_helper_funcs intel_mode_config_funcs = {
+	.atomic_commit_setup = drm_dp_mst_atomic_setup_commit,
+};
+
+static void intel_mode_config_init(struct drm_i915_private *i915)
+{
+	struct drm_mode_config *mode_config = &i915->drm.mode_config;
+
+	drm_mode_config_init(&i915->drm);
+	INIT_LIST_HEAD(&i915->global_obj_list);
+
+	mode_config->min_width = 0;
+	mode_config->min_height = 0;
+
+	mode_config->preferred_depth = 24;
+	mode_config->prefer_shadow = 1;
+
+	mode_config->funcs = &intel_mode_funcs;
+	mode_config->helper_private = &intel_mode_config_funcs;
+
+	mode_config->async_page_flip = HAS_ASYNC_FLIPS(i915);
+
+	/*
+	 * Maximum framebuffer dimensions, chosen to match
+	 * the maximum render engine surface size on gen4+.
+	 */
+	if (DISPLAY_VER(i915) >= 7) {
+		mode_config->max_width = 16384;
+		mode_config->max_height = 16384;
+	} else if (DISPLAY_VER(i915) >= 4) {
+		mode_config->max_width = 8192;
+		mode_config->max_height = 8192;
+	} else if (DISPLAY_VER(i915) == 3) {
+		mode_config->max_width = 4096;
+		mode_config->max_height = 4096;
+	} else {
+		mode_config->max_width = 2048;
+		mode_config->max_height = 2048;
+	}
+
+	if (IS_I845G(i915) || IS_I865G(i915)) {
+		mode_config->cursor_width = IS_I845G(i915) ? 64 : 512;
+		mode_config->cursor_height = 1023;
+	} else if (IS_I830(i915) || IS_I85X(i915) ||
+		   IS_I915G(i915) || IS_I915GM(i915)) {
+		mode_config->cursor_width = 64;
+		mode_config->cursor_height = 64;
+	} else {
+		mode_config->cursor_width = 256;
+		mode_config->cursor_height = 256;
+	}
+}
+
+static void intel_mode_config_cleanup(struct drm_i915_private *i915)
+{
+	intel_atomic_global_obj_cleanup(i915);
+	drm_mode_config_cleanup(&i915->drm);
+}
+
+/* part #1: call before irq install */
+int intel_modeset_init_noirq(struct drm_i915_private *i915)
+{
+	int ret;
+
+	if (i915_inject_probe_failure(i915))
+		return -ENODEV;
+
+	if (HAS_DISPLAY(i915)) {
+		ret = drm_vblank_init(&i915->drm,
+				      INTEL_NUM_PIPES(i915));
+		if (ret)
+			return ret;
+	}
+
+	intel_bios_init(i915);
+
+	ret = intel_vga_register(i915);
+	if (ret)
+		goto cleanup_bios;
+
+	/* FIXME: completely on the wrong abstraction layer */
+	intel_power_domains_init_hw(i915, false);
+
+	if (!HAS_DISPLAY(i915))
+		return 0;
+
+	intel_dmc_ucode_init(i915);
+
+	i915->display.wq.modeset = alloc_ordered_workqueue("i915_modeset", 0);
+	i915->display.wq.flip = alloc_workqueue("i915_flip", WQ_HIGHPRI |
+						WQ_UNBOUND, WQ_UNBOUND_MAX_ACTIVE);
+
+	intel_mode_config_init(i915);
+
+	ret = intel_cdclk_init(i915);
+	if (ret)
+		goto cleanup_vga_client_pw_domain_dmc;
+
+	ret = intel_dbuf_init(i915);
+	if (ret)
+		goto cleanup_vga_client_pw_domain_dmc;
+
+	ret = intel_bw_init(i915);
+	if (ret)
+		goto cleanup_vga_client_pw_domain_dmc;
+
+	init_llist_head(&i915->display.atomic_helper.free_list);
+	INIT_WORK(&i915->display.atomic_helper.free_work,
+		  intel_atomic_helper_free_state_worker);
+
+	intel_init_quirks(i915);
+
+	intel_fbc_init(i915);
+
+	return 0;
+
+cleanup_vga_client_pw_domain_dmc:
+	intel_dmc_ucode_fini(i915);
+	intel_power_domains_driver_remove(i915);
+	intel_vga_unregister(i915);
+cleanup_bios:
+	intel_bios_driver_remove(i915);
+
+	return ret;
+}
+
+/* part #2: call after irq install, but before gem init */
+int intel_modeset_init_nogem(struct drm_i915_private *i915)
+{
+	struct drm_device *dev = &i915->drm;
+	enum pipe pipe;
+	struct intel_crtc *crtc;
+	int ret;
+
+	if (!HAS_DISPLAY(i915))
+		return 0;
+
+	intel_init_pm(i915);
+
+	intel_panel_sanitize_ssc(i915);
+
+	intel_pps_setup(i915);
+
+	intel_gmbus_setup(i915);
+
+	drm_dbg_kms(&i915->drm, "%d display pipe%s available.\n",
+		    INTEL_NUM_PIPES(i915),
+		    INTEL_NUM_PIPES(i915) > 1 ? "s" : "");
+
+	for_each_pipe(i915, pipe) {
+		ret = intel_crtc_init(i915, pipe);
+		if (ret) {
+			intel_mode_config_cleanup(i915);
+			return ret;
+		}
+	}
+
+	intel_plane_possible_crtcs_init(i915);
+	intel_shared_dpll_init(i915);
+	intel_fdi_pll_freq_update(i915);
+
+	intel_update_czclk(i915);
+	intel_modeset_init_hw(i915);
+	intel_dpll_update_ref_clks(i915);
+
+	intel_hdcp_component_init(i915);
+
+	if (i915->display.cdclk.max_cdclk_freq == 0)
+		intel_update_max_cdclk(i915);
+
+	/*
+	 * If the platform has HTI, we need to find out whether it has reserved
+	 * any display resources before we create our display outputs.
+	 */
+	if (INTEL_INFO(i915)->display.has_hti)
+		i915->hti_state = intel_de_read(i915, HDPORT_STATE);
+
+	/* Just disable it once at startup */
+	intel_vga_disable(i915);
+	intel_setup_outputs(i915);
+
+	drm_modeset_lock_all(dev);
+	intel_modeset_setup_hw_state(i915, dev->mode_config.acquire_ctx);
+	intel_acpi_assign_connector_fwnodes(i915);
+	drm_modeset_unlock_all(dev);
+
+	for_each_intel_crtc(dev, crtc) {
+		if (!to_intel_crtc_state(crtc->base.state)->uapi.active)
+			continue;
+		intel_crtc_initial_plane_config(crtc);
+	}
+
+	/*
+	 * Make sure hardware watermarks really match the state we read out.
+	 * Note that we need to do this after reconstructing the BIOS fb's
+	 * since the watermark calculation done here will use pstate->fb.
+	 */
+	if (!HAS_GMCH(i915))
+		sanitize_watermarks(i915);
+
+	return 0;
+}
+
+/* part #3: call after gem init */
+int intel_modeset_init(struct drm_i915_private *i915)
+{
+	int ret;
+
+	if (!HAS_DISPLAY(i915))
+		return 0;
+
+	/*
+	 * Force all active planes to recompute their states. So that on
+	 * mode_setcrtc after probe, all the intel_plane_state variables
+	 * are already calculated and there is no assert_plane warnings
+	 * during bootup.
+	 */
+	ret = intel_initial_commit(&i915->drm);
+	if (ret)
+		drm_dbg_kms(&i915->drm, "Initial modeset failed, %d\n", ret);
+
+	intel_overlay_setup(i915);
+
+	ret = intel_fbdev_init(&i915->drm);
+	if (ret)
+		return ret;
+
+	/* Only enable hotplug handling once the fbdev is fully set up. */
+	intel_hpd_init(i915);
+	intel_hpd_poll_disable(i915);
+
+	skl_watermark_ipc_init(i915);
+
+	return 0;
+}
+
+void i830_enable_pipe(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	struct intel_crtc *crtc = intel_crtc_for_pipe(dev_priv, pipe);
+	/* 640x480@60Hz, ~25175 kHz */
+	struct dpll clock = {
+		.m1 = 18,
+		.m2 = 7,
+		.p1 = 13,
+		.p2 = 4,
+		.n = 2,
+	};
+	u32 dpll, fp;
+	int i;
+
+	drm_WARN_ON(&dev_priv->drm,
+		    i9xx_calc_dpll_params(48000, &clock) != 25154);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "enabling pipe %c due to force quirk (vco=%d dot=%d)\n",
+		    pipe_name(pipe), clock.vco, clock.dot);
+
+	fp = i9xx_dpll_compute_fp(&clock);
+	dpll = DPLL_DVO_2X_MODE |
+		DPLL_VGA_MODE_DIS |
+		((clock.p1 - 2) << DPLL_FPA01_P1_POST_DIV_SHIFT) |
+		PLL_P2_DIVIDE_BY_4 |
+		PLL_REF_INPUT_DREFCLK |
+		DPLL_VCO_ENABLE;
+
+	intel_de_write(dev_priv, HTOTAL(pipe), (640 - 1) | ((800 - 1) << 16));
+	intel_de_write(dev_priv, HBLANK(pipe), (640 - 1) | ((800 - 1) << 16));
+	intel_de_write(dev_priv, HSYNC(pipe), (656 - 1) | ((752 - 1) << 16));
+	intel_de_write(dev_priv, VTOTAL(pipe), (480 - 1) | ((525 - 1) << 16));
+	intel_de_write(dev_priv, VBLANK(pipe), (480 - 1) | ((525 - 1) << 16));
+	intel_de_write(dev_priv, VSYNC(pipe), (490 - 1) | ((492 - 1) << 16));
+	intel_de_write(dev_priv, PIPESRC(pipe), ((640 - 1) << 16) | (480 - 1));
+
+	intel_de_write(dev_priv, FP0(pipe), fp);
+	intel_de_write(dev_priv, FP1(pipe), fp);
+
+	/*
+	 * Apparently we need to have VGA mode enabled prior to changing
+	 * the P1/P2 dividers. Otherwise the DPLL will keep using the old
+	 * dividers, even though the register value does change.
+	 */
+	intel_de_write(dev_priv, DPLL(pipe), dpll & ~DPLL_VGA_MODE_DIS);
+	intel_de_write(dev_priv, DPLL(pipe), dpll);
+
+	/* Wait for the clocks to stabilize. */
+	intel_de_posting_read(dev_priv, DPLL(pipe));
+	udelay(150);
+
+	/* The pixel multiplier can only be updated once the
+	 * DPLL is enabled and the clocks are stable.
+	 *
+	 * So write it again.
+	 */
+	intel_de_write(dev_priv, DPLL(pipe), dpll);
+
+	/* We do this three times for luck */
+	for (i = 0; i < 3 ; i++) {
+		intel_de_write(dev_priv, DPLL(pipe), dpll);
+		intel_de_posting_read(dev_priv, DPLL(pipe));
+		udelay(150); /* wait for warmup */
+	}
+
+	intel_de_write(dev_priv, PIPECONF(pipe), PIPECONF_ENABLE);
+	intel_de_posting_read(dev_priv, PIPECONF(pipe));
+
+	intel_wait_for_pipe_scanline_moving(crtc);
+}
+
+void i830_disable_pipe(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	struct intel_crtc *crtc = intel_crtc_for_pipe(dev_priv, pipe);
+
+	drm_dbg_kms(&dev_priv->drm, "disabling pipe %c due to force quirk\n",
+		    pipe_name(pipe));
+
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, DSPCNTR(PLANE_A)) & DISP_ENABLE);
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, DSPCNTR(PLANE_B)) & DISP_ENABLE);
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, DSPCNTR(PLANE_C)) & DISP_ENABLE);
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, CURCNTR(PIPE_A)) & MCURSOR_MODE_MASK);
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, CURCNTR(PIPE_B)) & MCURSOR_MODE_MASK);
+
+	intel_de_write(dev_priv, PIPECONF(pipe), 0);
+	intel_de_posting_read(dev_priv, PIPECONF(pipe));
+
+	intel_wait_for_pipe_scanline_stopped(crtc);
+
+	intel_de_write(dev_priv, DPLL(pipe), DPLL_VGA_MODE_DIS);
+	intel_de_posting_read(dev_priv, DPLL(pipe));
+}
+
+void intel_display_resume(struct drm_device *dev)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_atomic_state *state = i915->modeset_restore_state;
+	struct drm_modeset_acquire_ctx ctx;
+	int ret;
+
+	if (!HAS_DISPLAY(i915))
+		return;
+
+	i915->modeset_restore_state = NULL;
+	if (state)
+		state->acquire_ctx = &ctx;
+
+	drm_modeset_acquire_init(&ctx, 0);
+
+	while (1) {
+		ret = drm_modeset_lock_all_ctx(dev, &ctx);
+		if (ret != -EDEADLK)
+			break;
+
+		drm_modeset_backoff(&ctx);
+	}
+
+	if (!ret)
+		ret = __intel_display_resume(i915, state, &ctx);
+
+	skl_watermark_ipc_update(i915);
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+
+	if (ret)
+		drm_err(&i915->drm,
+			"Restoring old state failed with %i\n", ret);
+	if (state)
+		drm_atomic_state_put(state);
+}
+
+static void intel_hpd_poll_fini(struct drm_i915_private *i915)
+{
+	struct intel_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+
+	/* Kill all the work that may have been queued by hpd. */
+	drm_connector_list_iter_begin(&i915->drm, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		if (connector->modeset_retry_work.func)
+			cancel_work_sync(&connector->modeset_retry_work);
+		if (connector->hdcp.shim) {
+			cancel_delayed_work_sync(&connector->hdcp.check_work);
+			cancel_work_sync(&connector->hdcp.prop_work);
+		}
+	}
+	drm_connector_list_iter_end(&conn_iter);
+}
+
+/* part #1: call before irq uninstall */
+void intel_modeset_driver_remove(struct drm_i915_private *i915)
+{
+	if (!HAS_DISPLAY(i915))
+		return;
+
+	flush_workqueue(i915->display.wq.flip);
+	flush_workqueue(i915->display.wq.modeset);
+
+	flush_work(&i915->display.atomic_helper.free_work);
+	drm_WARN_ON(&i915->drm, !llist_empty(&i915->display.atomic_helper.free_list));
+
+	/*
+	 * MST topology needs to be suspended so we don't have any calls to
+	 * fbdev after it's finalized. MST will be destroyed later as part of
+	 * drm_mode_config_cleanup()
+	 */
+	intel_dp_mst_suspend(i915);
+}
+
+/* part #2: call after irq uninstall */
+void intel_modeset_driver_remove_noirq(struct drm_i915_private *i915)
+{
+	if (!HAS_DISPLAY(i915))
+		return;
+
+	/*
+	 * Due to the hpd irq storm handling the hotplug work can re-arm the
+	 * poll handlers. Hence disable polling after hpd handling is shut down.
+	 */
+	intel_hpd_poll_fini(i915);
+
+	/* poll work can call into fbdev, hence clean that up afterwards */
+	intel_fbdev_fini(i915);
+
+	intel_unregister_dsm_handler();
+
+	/* flush any delayed tasks or pending work */
+	flush_scheduled_work();
+
+	intel_hdcp_component_fini(i915);
+
+	intel_mode_config_cleanup(i915);
+
+	intel_overlay_cleanup(i915);
+
+	intel_gmbus_teardown(i915);
+
+	destroy_workqueue(i915->display.wq.flip);
+	destroy_workqueue(i915->display.wq.modeset);
+
+	intel_fbc_cleanup(i915);
+}
+
+/* part #3: call after gem init */
+void intel_modeset_driver_remove_nogem(struct drm_i915_private *i915)
+{
+	intel_dmc_ucode_fini(i915);
+
+	intel_power_domains_driver_remove(i915);
+
+	intel_vga_unregister(i915);
+
+	intel_bios_driver_remove(i915);
+}
+
+bool intel_modeset_probe_defer(struct pci_dev *pdev)
+{
+	struct drm_privacy_screen *privacy_screen;
+
+	/*
+	 * apple-gmux is needed on dual GPU MacBook Pro
+	 * to probe the panel if we're the inactive GPU.
+	 */
+	if (vga_switcheroo_client_probe_defer(pdev))
+		return true;
+
+	/* If the LCD panel has a privacy-screen, wait for it */
+	privacy_screen = drm_privacy_screen_get(&pdev->dev, NULL);
+	if (IS_ERR(privacy_screen) && PTR_ERR(privacy_screen) == -EPROBE_DEFER)
+		return true;
+
+	drm_privacy_screen_put(privacy_screen);
+
+	return false;
+}
+
+void intel_display_driver_register(struct drm_i915_private *i915)
+{
+	if (!HAS_DISPLAY(i915))
+		return;
+
+	intel_display_debugfs_register(i915);
+
+	/* Must be done after probing outputs */
+	intel_opregion_register(i915);
+	intel_acpi_video_register(i915);
+
+	intel_audio_init(i915);
+
+	/*
+	 * Some ports require correctly set-up hpd registers for
+	 * detection to work properly (leading to ghost connected
+	 * connector status), e.g. VGA on gm45.  Hence we can only set
+	 * up the initial fbdev config after hpd irqs are fully
+	 * enabled. We do it last so that the async config cannot run
+	 * before the connectors are registered.
+	 */
+	intel_fbdev_initial_config_async(&i915->drm);
+
+	/*
+	 * We need to coordinate the hotplugs with the asynchronous
+	 * fbdev configuration, for which we use the
+	 * fbdev->async_cookie.
+	 */
+	drm_kms_helper_poll_init(&i915->drm);
+}
+
+void intel_display_driver_unregister(struct drm_i915_private *i915)
+{
+	if (!HAS_DISPLAY(i915))
+		return;
+
+	intel_fbdev_unregister(i915);
+	intel_audio_deinit(i915);
+
+	/*
+	 * After flushing the fbdev (incl. a late async config which
+	 * will have delayed queuing of a hotplug event), then flush
+	 * the hotplug events.
+	 */
+	drm_kms_helper_poll_fini(&i915->drm);
+	drm_atomic_helper_shutdown(&i915->drm);
+
+	acpi_video_unregister();
+	intel_opregion_unregister(i915);
+}
+
+bool intel_scanout_needs_vtd_wa(struct drm_i915_private *i915)
+{
+	return DISPLAY_VER(i915) >= 6 && i915_vtd_active(i915);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_display.h b/drivers/gpu/drm/i915/display/intel_display.h
new file mode 100644
index 000000000..b4f941674
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display.h
@@ -0,0 +1,721 @@
+/*
+ * Copyright © 2006-2019 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _INTEL_DISPLAY_H_
+#define _INTEL_DISPLAY_H_
+
+#include <drm/drm_util.h>
+
+#include "i915_reg_defs.h"
+
+enum drm_scaling_filter;
+struct dpll;
+struct drm_connector;
+struct drm_device;
+struct drm_display_mode;
+struct drm_encoder;
+struct drm_file;
+struct drm_format_info;
+struct drm_framebuffer;
+struct drm_i915_gem_object;
+struct drm_i915_private;
+struct drm_mode_fb_cmd2;
+struct drm_modeset_acquire_ctx;
+struct drm_plane;
+struct drm_plane_state;
+struct i915_address_space;
+struct i915_gtt_view;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_digital_port;
+struct intel_dp;
+struct intel_encoder;
+struct intel_initial_plane_config;
+struct intel_load_detect_pipe;
+struct intel_plane;
+struct intel_plane_state;
+struct intel_power_domain_mask;
+struct intel_remapped_info;
+struct intel_rotation_info;
+struct pci_dev;
+
+enum i915_gpio {
+	GPIOA,
+	GPIOB,
+	GPIOC,
+	GPIOD,
+	GPIOE,
+	GPIOF,
+	GPIOG,
+	GPIOH,
+	__GPIOI_UNUSED,
+	GPIOJ,
+	GPIOK,
+	GPIOL,
+	GPIOM,
+	GPION,
+	GPIOO,
+};
+
+/*
+ * Keep the pipe enum values fixed: the code assumes that PIPE_A=0, the
+ * rest have consecutive values and match the enum values of transcoders
+ * with a 1:1 transcoder -> pipe mapping.
+ */
+enum pipe {
+	INVALID_PIPE = -1,
+
+	PIPE_A = 0,
+	PIPE_B,
+	PIPE_C,
+	PIPE_D,
+	_PIPE_EDP,
+
+	I915_MAX_PIPES = _PIPE_EDP
+};
+
+#define pipe_name(p) ((p) + 'A')
+
+enum transcoder {
+	INVALID_TRANSCODER = -1,
+	/*
+	 * The following transcoders have a 1:1 transcoder -> pipe mapping,
+	 * keep their values fixed: the code assumes that TRANSCODER_A=0, the
+	 * rest have consecutive values and match the enum values of the pipes
+	 * they map to.
+	 */
+	TRANSCODER_A = PIPE_A,
+	TRANSCODER_B = PIPE_B,
+	TRANSCODER_C = PIPE_C,
+	TRANSCODER_D = PIPE_D,
+
+	/*
+	 * The following transcoders can map to any pipe, their enum value
+	 * doesn't need to stay fixed.
+	 */
+	TRANSCODER_EDP,
+	TRANSCODER_DSI_0,
+	TRANSCODER_DSI_1,
+	TRANSCODER_DSI_A = TRANSCODER_DSI_0,	/* legacy DSI */
+	TRANSCODER_DSI_C = TRANSCODER_DSI_1,	/* legacy DSI */
+
+	I915_MAX_TRANSCODERS
+};
+
+static inline const char *transcoder_name(enum transcoder transcoder)
+{
+	switch (transcoder) {
+	case TRANSCODER_A:
+		return "A";
+	case TRANSCODER_B:
+		return "B";
+	case TRANSCODER_C:
+		return "C";
+	case TRANSCODER_D:
+		return "D";
+	case TRANSCODER_EDP:
+		return "EDP";
+	case TRANSCODER_DSI_A:
+		return "DSI A";
+	case TRANSCODER_DSI_C:
+		return "DSI C";
+	default:
+		return "<invalid>";
+	}
+}
+
+static inline bool transcoder_is_dsi(enum transcoder transcoder)
+{
+	return transcoder == TRANSCODER_DSI_A || transcoder == TRANSCODER_DSI_C;
+}
+
+/*
+ * Global legacy plane identifier. Valid only for primary/sprite
+ * planes on pre-g4x, and only for primary planes on g4x-bdw.
+ */
+enum i9xx_plane_id {
+	PLANE_A,
+	PLANE_B,
+	PLANE_C,
+};
+
+#define plane_name(p) ((p) + 'A')
+#define sprite_name(p, s) ((p) * RUNTIME_INFO(dev_priv)->num_sprites[(p)] + (s) + 'A')
+
+/*
+ * Per-pipe plane identifier.
+ * I915_MAX_PLANES in the enum below is the maximum (across all platforms)
+ * number of planes per CRTC.  Not all platforms really have this many planes,
+ * which means some arrays of size I915_MAX_PLANES may have unused entries
+ * between the topmost sprite plane and the cursor plane.
+ *
+ * This is expected to be passed to various register macros
+ * (eg. PLANE_CTL(), PS_PLANE_SEL(), etc.) so adjust with care.
+ */
+enum plane_id {
+	PLANE_PRIMARY,
+	PLANE_SPRITE0,
+	PLANE_SPRITE1,
+	PLANE_SPRITE2,
+	PLANE_SPRITE3,
+	PLANE_SPRITE4,
+	PLANE_SPRITE5,
+	PLANE_CURSOR,
+
+	I915_MAX_PLANES,
+};
+
+#define for_each_plane_id_on_crtc(__crtc, __p) \
+	for ((__p) = PLANE_PRIMARY; (__p) < I915_MAX_PLANES; (__p)++) \
+		for_each_if((__crtc)->plane_ids_mask & BIT(__p))
+
+#define for_each_dbuf_slice(__dev_priv, __slice) \
+	for ((__slice) = DBUF_S1; (__slice) < I915_MAX_DBUF_SLICES; (__slice)++) \
+		for_each_if(INTEL_INFO(__dev_priv)->display.dbuf.slice_mask & BIT(__slice))
+
+#define for_each_dbuf_slice_in_mask(__dev_priv, __slice, __mask) \
+	for_each_dbuf_slice((__dev_priv), (__slice)) \
+		for_each_if((__mask) & BIT(__slice))
+
+enum port {
+	PORT_NONE = -1,
+
+	PORT_A = 0,
+	PORT_B,
+	PORT_C,
+	PORT_D,
+	PORT_E,
+	PORT_F,
+	PORT_G,
+	PORT_H,
+	PORT_I,
+
+	/* tgl+ */
+	PORT_TC1 = PORT_D,
+	PORT_TC2,
+	PORT_TC3,
+	PORT_TC4,
+	PORT_TC5,
+	PORT_TC6,
+
+	/* XE_LPD repositions D/E offsets and bitfields */
+	PORT_D_XELPD = PORT_TC5,
+	PORT_E_XELPD,
+
+	I915_MAX_PORTS
+};
+
+#define port_name(p) ((p) + 'A')
+
+/*
+ * Ports identifier referenced from other drivers.
+ * Expected to remain stable over time
+ */
+static inline const char *port_identifier(enum port port)
+{
+	switch (port) {
+	case PORT_A:
+		return "Port A";
+	case PORT_B:
+		return "Port B";
+	case PORT_C:
+		return "Port C";
+	case PORT_D:
+		return "Port D";
+	case PORT_E:
+		return "Port E";
+	case PORT_F:
+		return "Port F";
+	case PORT_G:
+		return "Port G";
+	case PORT_H:
+		return "Port H";
+	case PORT_I:
+		return "Port I";
+	default:
+		return "<invalid>";
+	}
+}
+
+enum tc_port {
+	TC_PORT_NONE = -1,
+
+	TC_PORT_1 = 0,
+	TC_PORT_2,
+	TC_PORT_3,
+	TC_PORT_4,
+	TC_PORT_5,
+	TC_PORT_6,
+
+	I915_MAX_TC_PORTS
+};
+
+enum tc_port_mode {
+	TC_PORT_DISCONNECTED,
+	TC_PORT_TBT_ALT,
+	TC_PORT_DP_ALT,
+	TC_PORT_LEGACY,
+};
+
+enum dpio_channel {
+	DPIO_CH0,
+	DPIO_CH1
+};
+
+enum dpio_phy {
+	DPIO_PHY0,
+	DPIO_PHY1,
+	DPIO_PHY2,
+};
+
+enum aux_ch {
+	AUX_CH_A,
+	AUX_CH_B,
+	AUX_CH_C,
+	AUX_CH_D,
+	AUX_CH_E, /* ICL+ */
+	AUX_CH_F,
+	AUX_CH_G,
+	AUX_CH_H,
+	AUX_CH_I,
+
+	/* tgl+ */
+	AUX_CH_USBC1 = AUX_CH_D,
+	AUX_CH_USBC2,
+	AUX_CH_USBC3,
+	AUX_CH_USBC4,
+	AUX_CH_USBC5,
+	AUX_CH_USBC6,
+
+	/* XE_LPD repositions D/E offsets and bitfields */
+	AUX_CH_D_XELPD = AUX_CH_USBC5,
+	AUX_CH_E_XELPD,
+};
+
+#define aux_ch_name(a) ((a) + 'A')
+
+/* Used by dp and fdi links */
+struct intel_link_m_n {
+	u32 tu;
+	u32 data_m;
+	u32 data_n;
+	u32 link_m;
+	u32 link_n;
+};
+
+enum phy {
+	PHY_NONE = -1,
+
+	PHY_A = 0,
+	PHY_B,
+	PHY_C,
+	PHY_D,
+	PHY_E,
+	PHY_F,
+	PHY_G,
+	PHY_H,
+	PHY_I,
+
+	I915_MAX_PHYS
+};
+
+#define phy_name(a) ((a) + 'A')
+
+enum phy_fia {
+	FIA1,
+	FIA2,
+	FIA3,
+};
+
+enum hpd_pin {
+	HPD_NONE = 0,
+	HPD_TV = HPD_NONE,     /* TV is known to be unreliable */
+	HPD_CRT,
+	HPD_SDVO_B,
+	HPD_SDVO_C,
+	HPD_PORT_A,
+	HPD_PORT_B,
+	HPD_PORT_C,
+	HPD_PORT_D,
+	HPD_PORT_E,
+	HPD_PORT_TC1,
+	HPD_PORT_TC2,
+	HPD_PORT_TC3,
+	HPD_PORT_TC4,
+	HPD_PORT_TC5,
+	HPD_PORT_TC6,
+
+	HPD_NUM_PINS
+};
+
+#define for_each_hpd_pin(__pin) \
+	for ((__pin) = (HPD_NONE + 1); (__pin) < HPD_NUM_PINS; (__pin)++)
+
+#define for_each_pipe(__dev_priv, __p) \
+	for ((__p) = 0; (__p) < I915_MAX_PIPES; (__p)++) \
+		for_each_if(RUNTIME_INFO(__dev_priv)->pipe_mask & BIT(__p))
+
+#define for_each_pipe_masked(__dev_priv, __p, __mask) \
+	for_each_pipe(__dev_priv, __p) \
+		for_each_if((__mask) & BIT(__p))
+
+#define for_each_cpu_transcoder(__dev_priv, __t) \
+	for ((__t) = 0; (__t) < I915_MAX_TRANSCODERS; (__t)++)	\
+		for_each_if (RUNTIME_INFO(__dev_priv)->cpu_transcoder_mask & BIT(__t))
+
+#define for_each_cpu_transcoder_masked(__dev_priv, __t, __mask) \
+	for_each_cpu_transcoder(__dev_priv, __t) \
+		for_each_if ((__mask) & BIT(__t))
+
+#define for_each_sprite(__dev_priv, __p, __s)				\
+	for ((__s) = 0;							\
+	     (__s) < RUNTIME_INFO(__dev_priv)->num_sprites[(__p)];	\
+	     (__s)++)
+
+#define for_each_port(__port) \
+	for ((__port) = PORT_A; (__port) < I915_MAX_PORTS; (__port)++)
+
+#define for_each_port_masked(__port, __ports_mask)			\
+	for_each_port(__port)						\
+		for_each_if((__ports_mask) & BIT(__port))
+
+#define for_each_phy_masked(__phy, __phys_mask) \
+	for ((__phy) = PHY_A; (__phy) < I915_MAX_PHYS; (__phy)++)	\
+		for_each_if((__phys_mask) & BIT(__phy))
+
+#define for_each_crtc(dev, crtc) \
+	list_for_each_entry(crtc, &(dev)->mode_config.crtc_list, head)
+
+#define for_each_intel_plane(dev, intel_plane) \
+	list_for_each_entry(intel_plane,			\
+			    &(dev)->mode_config.plane_list,	\
+			    base.head)
+
+#define for_each_intel_plane_mask(dev, intel_plane, plane_mask)		\
+	list_for_each_entry(intel_plane,				\
+			    &(dev)->mode_config.plane_list,		\
+			    base.head)					\
+		for_each_if((plane_mask) &				\
+			    drm_plane_mask(&intel_plane->base))
+
+#define for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane)	\
+	list_for_each_entry(intel_plane,				\
+			    &(dev)->mode_config.plane_list,		\
+			    base.head)					\
+		for_each_if((intel_plane)->pipe == (intel_crtc)->pipe)
+
+#define for_each_intel_crtc(dev, intel_crtc)				\
+	list_for_each_entry(intel_crtc,					\
+			    &(dev)->mode_config.crtc_list,		\
+			    base.head)
+
+#define for_each_intel_crtc_in_pipe_mask(dev, intel_crtc, pipe_mask)	\
+	list_for_each_entry(intel_crtc,					\
+			    &(dev)->mode_config.crtc_list,		\
+			    base.head)					\
+		for_each_if((pipe_mask) & BIT(intel_crtc->pipe))
+
+#define for_each_intel_encoder(dev, intel_encoder)		\
+	list_for_each_entry(intel_encoder,			\
+			    &(dev)->mode_config.encoder_list,	\
+			    base.head)
+
+#define for_each_intel_encoder_mask(dev, intel_encoder, encoder_mask)	\
+	list_for_each_entry(intel_encoder,				\
+			    &(dev)->mode_config.encoder_list,		\
+			    base.head)					\
+		for_each_if((encoder_mask) &				\
+			    drm_encoder_mask(&intel_encoder->base))
+
+#define for_each_intel_encoder_mask_with_psr(dev, intel_encoder, encoder_mask) \
+	list_for_each_entry((intel_encoder), &(dev)->mode_config.encoder_list, base.head) \
+		for_each_if(((encoder_mask) & drm_encoder_mask(&(intel_encoder)->base)) && \
+			    intel_encoder_can_psr(intel_encoder))
+
+#define for_each_intel_dp(dev, intel_encoder)			\
+	for_each_intel_encoder(dev, intel_encoder)		\
+		for_each_if(intel_encoder_is_dp(intel_encoder))
+
+#define for_each_intel_encoder_with_psr(dev, intel_encoder) \
+	for_each_intel_encoder((dev), (intel_encoder)) \
+		for_each_if(intel_encoder_can_psr(intel_encoder))
+
+#define for_each_intel_connector_iter(intel_connector, iter) \
+	while ((intel_connector = to_intel_connector(drm_connector_list_iter_next(iter))))
+
+#define for_each_encoder_on_crtc(dev, __crtc, intel_encoder) \
+	list_for_each_entry((intel_encoder), &(dev)->mode_config.encoder_list, base.head) \
+		for_each_if((intel_encoder)->base.crtc == (__crtc))
+
+#define for_each_connector_on_encoder(dev, __encoder, intel_connector) \
+	list_for_each_entry((intel_connector), &(dev)->mode_config.connector_list, base.head) \
+		for_each_if((intel_connector)->base.encoder == (__encoder))
+
+#define for_each_old_intel_plane_in_state(__state, plane, old_plane_state, __i) \
+	for ((__i) = 0; \
+	     (__i) < (__state)->base.dev->mode_config.num_total_plane && \
+		     ((plane) = to_intel_plane((__state)->base.planes[__i].ptr), \
+		      (old_plane_state) = to_intel_plane_state((__state)->base.planes[__i].old_state), 1); \
+	     (__i)++) \
+		for_each_if(plane)
+
+#define for_each_new_intel_plane_in_state(__state, plane, new_plane_state, __i) \
+	for ((__i) = 0; \
+	     (__i) < (__state)->base.dev->mode_config.num_total_plane && \
+		     ((plane) = to_intel_plane((__state)->base.planes[__i].ptr), \
+		      (new_plane_state) = to_intel_plane_state((__state)->base.planes[__i].new_state), 1); \
+	     (__i)++) \
+		for_each_if(plane)
+
+#define for_each_new_intel_crtc_in_state(__state, crtc, new_crtc_state, __i) \
+	for ((__i) = 0; \
+	     (__i) < (__state)->base.dev->mode_config.num_crtc && \
+		     ((crtc) = to_intel_crtc((__state)->base.crtcs[__i].ptr), \
+		      (new_crtc_state) = to_intel_crtc_state((__state)->base.crtcs[__i].new_state), 1); \
+	     (__i)++) \
+		for_each_if(crtc)
+
+#define for_each_oldnew_intel_plane_in_state(__state, plane, old_plane_state, new_plane_state, __i) \
+	for ((__i) = 0; \
+	     (__i) < (__state)->base.dev->mode_config.num_total_plane && \
+		     ((plane) = to_intel_plane((__state)->base.planes[__i].ptr), \
+		      (old_plane_state) = to_intel_plane_state((__state)->base.planes[__i].old_state), \
+		      (new_plane_state) = to_intel_plane_state((__state)->base.planes[__i].new_state), 1); \
+	     (__i)++) \
+		for_each_if(plane)
+
+#define for_each_oldnew_intel_crtc_in_state(__state, crtc, old_crtc_state, new_crtc_state, __i) \
+	for ((__i) = 0; \
+	     (__i) < (__state)->base.dev->mode_config.num_crtc && \
+		     ((crtc) = to_intel_crtc((__state)->base.crtcs[__i].ptr), \
+		      (old_crtc_state) = to_intel_crtc_state((__state)->base.crtcs[__i].old_state), \
+		      (new_crtc_state) = to_intel_crtc_state((__state)->base.crtcs[__i].new_state), 1); \
+	     (__i)++) \
+		for_each_if(crtc)
+
+#define for_each_oldnew_intel_crtc_in_state_reverse(__state, crtc, old_crtc_state, new_crtc_state, __i) \
+	for ((__i) = (__state)->base.dev->mode_config.num_crtc - 1; \
+	     (__i) >= 0  && \
+	     ((crtc) = to_intel_crtc((__state)->base.crtcs[__i].ptr), \
+	      (old_crtc_state) = to_intel_crtc_state((__state)->base.crtcs[__i].old_state), \
+	      (new_crtc_state) = to_intel_crtc_state((__state)->base.crtcs[__i].new_state), 1); \
+	     (__i)--) \
+		for_each_if(crtc)
+
+#define intel_atomic_crtc_state_for_each_plane_state( \
+		  plane, plane_state, \
+		  crtc_state) \
+	for_each_intel_plane_mask(((crtc_state)->uapi.state->dev), (plane), \
+				((crtc_state)->uapi.plane_mask)) \
+		for_each_if ((plane_state = \
+			      to_intel_plane_state(__drm_atomic_get_current_plane_state((crtc_state)->uapi.state, &plane->base))))
+
+#define for_each_new_intel_connector_in_state(__state, connector, new_connector_state, __i) \
+	for ((__i) = 0; \
+	     (__i) < (__state)->base.num_connector; \
+	     (__i)++) \
+		for_each_if ((__state)->base.connectors[__i].ptr && \
+			     ((connector) = to_intel_connector((__state)->base.connectors[__i].ptr), \
+			     (new_connector_state) = to_intel_digital_connector_state((__state)->base.connectors[__i].new_state), 1))
+
+int intel_atomic_add_affected_planes(struct intel_atomic_state *state,
+				     struct intel_crtc *crtc);
+u8 intel_calc_active_pipes(struct intel_atomic_state *state,
+			   u8 active_pipes);
+void intel_link_compute_m_n(u16 bpp, int nlanes,
+			    int pixel_clock, int link_clock,
+			    struct intel_link_m_n *m_n,
+			    bool fec_enable);
+u32 intel_plane_fb_max_stride(struct drm_i915_private *dev_priv,
+			      u32 pixel_format, u64 modifier);
+enum drm_mode_status
+intel_mode_valid_max_plane_size(struct drm_i915_private *dev_priv,
+				const struct drm_display_mode *mode,
+				bool bigjoiner);
+enum drm_mode_status
+intel_cpu_transcoder_mode_valid(struct drm_i915_private *i915,
+				const struct drm_display_mode *mode);
+enum phy intel_port_to_phy(struct drm_i915_private *i915, enum port port);
+bool is_trans_port_sync_mode(const struct intel_crtc_state *state);
+bool intel_crtc_is_bigjoiner_slave(const struct intel_crtc_state *crtc_state);
+bool intel_crtc_is_bigjoiner_master(const struct intel_crtc_state *crtc_state);
+u8 intel_crtc_bigjoiner_slave_pipes(const struct intel_crtc_state *crtc_state);
+struct intel_crtc *intel_master_crtc(const struct intel_crtc_state *crtc_state);
+bool intel_crtc_get_pipe_config(struct intel_crtc_state *crtc_state);
+bool intel_pipe_config_compare(const struct intel_crtc_state *current_config,
+			       const struct intel_crtc_state *pipe_config,
+			       bool fastset);
+void intel_crtc_update_active_timings(const struct intel_crtc_state *crtc_state);
+
+void intel_plane_destroy(struct drm_plane *plane);
+void i9xx_set_pipeconf(const struct intel_crtc_state *crtc_state);
+void ilk_set_pipeconf(const struct intel_crtc_state *crtc_state);
+void intel_enable_transcoder(const struct intel_crtc_state *new_crtc_state);
+void intel_disable_transcoder(const struct intel_crtc_state *old_crtc_state);
+void i830_enable_pipe(struct drm_i915_private *dev_priv, enum pipe pipe);
+void i830_disable_pipe(struct drm_i915_private *dev_priv, enum pipe pipe);
+int vlv_get_hpll_vco(struct drm_i915_private *dev_priv);
+int vlv_get_cck_clock(struct drm_i915_private *dev_priv,
+		      const char *name, u32 reg, int ref_freq);
+int vlv_get_cck_clock_hpll(struct drm_i915_private *dev_priv,
+			   const char *name, u32 reg);
+void intel_init_display_hooks(struct drm_i915_private *dev_priv);
+unsigned int intel_fb_xy_to_linear(int x, int y,
+				   const struct intel_plane_state *state,
+				   int plane);
+void intel_add_fb_offsets(int *x, int *y,
+			  const struct intel_plane_state *state, int plane);
+unsigned int intel_rotation_info_size(const struct intel_rotation_info *rot_info);
+unsigned int intel_remapped_info_size(const struct intel_remapped_info *rem_info);
+bool intel_has_pending_fb_unpin(struct drm_i915_private *dev_priv);
+int intel_display_suspend(struct drm_device *dev);
+void intel_encoder_destroy(struct drm_encoder *encoder);
+struct drm_display_mode *
+intel_encoder_current_mode(struct intel_encoder *encoder);
+void intel_encoder_get_config(struct intel_encoder *encoder,
+			      struct intel_crtc_state *crtc_state);
+bool intel_phy_is_combo(struct drm_i915_private *dev_priv, enum phy phy);
+bool intel_phy_is_tc(struct drm_i915_private *dev_priv, enum phy phy);
+bool intel_phy_is_snps(struct drm_i915_private *dev_priv, enum phy phy);
+enum tc_port intel_port_to_tc(struct drm_i915_private *dev_priv,
+			      enum port port);
+int intel_get_pipe_from_crtc_id_ioctl(struct drm_device *dev, void *data,
+				      struct drm_file *file_priv);
+
+int ilk_get_lanes_required(int target_clock, int link_bw, int bpp);
+void vlv_wait_port_ready(struct drm_i915_private *dev_priv,
+			 struct intel_digital_port *dig_port,
+			 unsigned int expected_mask);
+int intel_get_load_detect_pipe(struct drm_connector *connector,
+			       struct intel_load_detect_pipe *old,
+			       struct drm_modeset_acquire_ctx *ctx);
+void intel_release_load_detect_pipe(struct drm_connector *connector,
+				    struct intel_load_detect_pipe *old,
+				    struct drm_modeset_acquire_ctx *ctx);
+struct drm_framebuffer *
+intel_framebuffer_create(struct drm_i915_gem_object *obj,
+			 struct drm_mode_fb_cmd2 *mode_cmd);
+
+bool intel_fuzzy_clock_check(int clock1, int clock2);
+
+void intel_display_prepare_reset(struct drm_i915_private *dev_priv);
+void intel_display_finish_reset(struct drm_i915_private *dev_priv);
+void intel_zero_m_n(struct intel_link_m_n *m_n);
+void intel_set_m_n(struct drm_i915_private *i915,
+		   const struct intel_link_m_n *m_n,
+		   i915_reg_t data_m_reg, i915_reg_t data_n_reg,
+		   i915_reg_t link_m_reg, i915_reg_t link_n_reg);
+void intel_get_m_n(struct drm_i915_private *i915,
+		   struct intel_link_m_n *m_n,
+		   i915_reg_t data_m_reg, i915_reg_t data_n_reg,
+		   i915_reg_t link_m_reg, i915_reg_t link_n_reg);
+bool intel_cpu_transcoder_has_m2_n2(struct drm_i915_private *dev_priv,
+				    enum transcoder transcoder);
+void intel_cpu_transcoder_set_m1_n1(struct intel_crtc *crtc,
+				    enum transcoder cpu_transcoder,
+				    const struct intel_link_m_n *m_n);
+void intel_cpu_transcoder_set_m2_n2(struct intel_crtc *crtc,
+				    enum transcoder cpu_transcoder,
+				    const struct intel_link_m_n *m_n);
+void intel_cpu_transcoder_get_m1_n1(struct intel_crtc *crtc,
+				    enum transcoder cpu_transcoder,
+				    struct intel_link_m_n *m_n);
+void intel_cpu_transcoder_get_m2_n2(struct intel_crtc *crtc,
+				    enum transcoder cpu_transcoder,
+				    struct intel_link_m_n *m_n);
+void i9xx_crtc_clock_get(struct intel_crtc *crtc,
+			 struct intel_crtc_state *pipe_config);
+int intel_dotclock_calculate(int link_freq, const struct intel_link_m_n *m_n);
+int intel_crtc_dotclock(const struct intel_crtc_state *pipe_config);
+enum intel_display_power_domain intel_port_to_power_domain(struct intel_digital_port *dig_port);
+enum intel_display_power_domain
+intel_aux_power_domain(struct intel_digital_port *dig_port);
+void intel_crtc_arm_fifo_underrun(struct intel_crtc *crtc,
+				  struct intel_crtc_state *crtc_state);
+void ilk_pfit_disable(const struct intel_crtc_state *old_crtc_state);
+
+int bdw_get_pipemisc_bpp(struct intel_crtc *crtc);
+unsigned int intel_plane_fence_y_offset(const struct intel_plane_state *plane_state);
+
+bool intel_plane_uses_fence(const struct intel_plane_state *plane_state);
+
+struct intel_encoder *
+intel_get_crtc_new_encoder(const struct intel_atomic_state *state,
+			   const struct intel_crtc_state *crtc_state);
+void intel_plane_disable_noatomic(struct intel_crtc *crtc,
+				  struct intel_plane *plane);
+void intel_set_plane_visible(struct intel_crtc_state *crtc_state,
+			     struct intel_plane_state *plane_state,
+			     bool visible);
+void intel_plane_fixup_bitmasks(struct intel_crtc_state *crtc_state);
+
+void intel_display_driver_register(struct drm_i915_private *i915);
+void intel_display_driver_unregister(struct drm_i915_private *i915);
+
+void intel_update_watermarks(struct drm_i915_private *i915);
+
+/* modesetting */
+bool intel_modeset_probe_defer(struct pci_dev *pdev);
+void intel_modeset_init_hw(struct drm_i915_private *i915);
+int intel_modeset_init_noirq(struct drm_i915_private *i915);
+int intel_modeset_init_nogem(struct drm_i915_private *i915);
+int intel_modeset_init(struct drm_i915_private *i915);
+void intel_modeset_driver_remove(struct drm_i915_private *i915);
+void intel_modeset_driver_remove_noirq(struct drm_i915_private *i915);
+void intel_modeset_driver_remove_nogem(struct drm_i915_private *i915);
+void intel_display_resume(struct drm_device *dev);
+int intel_modeset_all_pipes(struct intel_atomic_state *state);
+void intel_modeset_get_crtc_power_domains(struct intel_crtc_state *crtc_state,
+					  struct intel_power_domain_mask *old_domains);
+void intel_modeset_put_crtc_power_domains(struct intel_crtc *crtc,
+					  struct intel_power_domain_mask *domains);
+
+/* modesetting asserts */
+void assert_transcoder(struct drm_i915_private *dev_priv,
+		       enum transcoder cpu_transcoder, bool state);
+#define assert_transcoder_enabled(d, t) assert_transcoder(d, t, true)
+#define assert_transcoder_disabled(d, t) assert_transcoder(d, t, false)
+
+/* Use I915_STATE_WARN(x) and I915_STATE_WARN_ON() (rather than WARN() and
+ * WARN_ON()) for hw state sanity checks to check for unexpected conditions
+ * which may not necessarily be a user visible problem.  This will either
+ * WARN() or DRM_ERROR() depending on the verbose_checks moduleparam, to
+ * enable distros and users to tailor their preferred amount of i915 abrt
+ * spam.
+ */
+#define I915_STATE_WARN(condition, format...) ({			\
+	int __ret_warn_on = !!(condition);				\
+	if (unlikely(__ret_warn_on))					\
+		if (!WARN(i915_modparams.verbose_state_checks, format))	\
+			DRM_ERROR(format);				\
+	unlikely(__ret_warn_on);					\
+})
+
+#define I915_STATE_WARN_ON(x)						\
+	I915_STATE_WARN((x), "%s", "WARN_ON(" __stringify(x) ")")
+
+bool intel_scanout_needs_vtd_wa(struct drm_i915_private *i915);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_display_core.h b/drivers/gpu/drm/i915/display/intel_display_core.h
new file mode 100644
index 000000000..9b51148e8
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_core.h
@@ -0,0 +1,426 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_DISPLAY_CORE_H__
+#define __INTEL_DISPLAY_CORE_H__
+
+#include <linux/list.h>
+#include <linux/llist.h>
+#include <linux/mutex.h>
+#include <linux/types.h>
+#include <linux/wait.h>
+#include <linux/workqueue.h>
+
+#include <drm/drm_connector.h>
+
+#include "intel_cdclk.h"
+#include "intel_display.h"
+#include "intel_display_power.h"
+#include "intel_dmc.h"
+#include "intel_dpll_mgr.h"
+#include "intel_fbc.h"
+#include "intel_global_state.h"
+#include "intel_gmbus.h"
+#include "intel_opregion.h"
+#include "intel_pm_types.h"
+
+struct drm_i915_private;
+struct drm_property;
+struct i915_audio_component;
+struct i915_hdcp_comp_master;
+struct intel_atomic_state;
+struct intel_audio_funcs;
+struct intel_bios_encoder_data;
+struct intel_cdclk_funcs;
+struct intel_cdclk_vals;
+struct intel_color_funcs;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_dpll_funcs;
+struct intel_dpll_mgr;
+struct intel_fbdev;
+struct intel_fdi_funcs;
+struct intel_hotplug_funcs;
+struct intel_initial_plane_config;
+struct intel_overlay;
+
+/* Amount of SAGV/QGV points, BSpec precisely defines this */
+#define I915_NUM_QGV_POINTS 8
+
+/* Amount of PSF GV points, BSpec precisely defines this */
+#define I915_NUM_PSF_GV_POINTS 3
+
+struct intel_display_funcs {
+	/*
+	 * Returns the active state of the crtc, and if the crtc is active,
+	 * fills out the pipe-config with the hw state.
+	 */
+	bool (*get_pipe_config)(struct intel_crtc *,
+				struct intel_crtc_state *);
+	void (*get_initial_plane_config)(struct intel_crtc *,
+					 struct intel_initial_plane_config *);
+	void (*crtc_enable)(struct intel_atomic_state *state,
+			    struct intel_crtc *crtc);
+	void (*crtc_disable)(struct intel_atomic_state *state,
+			     struct intel_crtc *crtc);
+	void (*commit_modeset_enables)(struct intel_atomic_state *state);
+};
+
+/* functions used for watermark calcs for display. */
+struct intel_wm_funcs {
+	/* update_wm is for legacy wm management */
+	void (*update_wm)(struct drm_i915_private *dev_priv);
+	int (*compute_pipe_wm)(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc);
+	int (*compute_intermediate_wm)(struct intel_atomic_state *state,
+				       struct intel_crtc *crtc);
+	void (*initial_watermarks)(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc);
+	void (*atomic_update_watermarks)(struct intel_atomic_state *state,
+					 struct intel_crtc *crtc);
+	void (*optimize_watermarks)(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc);
+	int (*compute_global_watermarks)(struct intel_atomic_state *state);
+};
+
+struct intel_audio {
+	/* hda/i915 audio component */
+	struct i915_audio_component *component;
+	bool component_registered;
+	/* mutex for audio/video sync */
+	struct mutex mutex;
+	int power_refcount;
+	u32 freq_cntrl;
+
+	/* Used to save the pipe-to-encoder mapping for audio */
+	struct intel_encoder *encoder_map[I915_MAX_PIPES];
+
+	/* necessary resource sharing with HDMI LPE audio driver. */
+	struct {
+		struct platform_device *platdev;
+		int irq;
+	} lpe;
+};
+
+/*
+ * dpll and cdclk state is protected by connection_mutex dpll.lock serializes
+ * intel_{prepare,enable,disable}_shared_dpll.  Must be global rather than per
+ * dpll, because on some platforms plls share registers.
+ */
+struct intel_dpll {
+	struct mutex lock;
+
+	int num_shared_dpll;
+	struct intel_shared_dpll shared_dplls[I915_NUM_PLLS];
+	const struct intel_dpll_mgr *mgr;
+
+	struct {
+		int nssc;
+		int ssc;
+	} ref_clks;
+};
+
+struct intel_frontbuffer_tracking {
+	spinlock_t lock;
+
+	/*
+	 * Tracking bits for delayed frontbuffer flushing du to gpu activity or
+	 * scheduled flips.
+	 */
+	unsigned busy_bits;
+	unsigned flip_bits;
+};
+
+struct intel_hotplug {
+	struct delayed_work hotplug_work;
+
+	const u32 *hpd, *pch_hpd;
+
+	struct {
+		unsigned long last_jiffies;
+		int count;
+		enum {
+			HPD_ENABLED = 0,
+			HPD_DISABLED = 1,
+			HPD_MARK_DISABLED = 2
+		} state;
+	} stats[HPD_NUM_PINS];
+	u32 event_bits;
+	u32 retry_bits;
+	struct delayed_work reenable_work;
+
+	u32 long_port_mask;
+	u32 short_port_mask;
+	struct work_struct dig_port_work;
+
+	struct work_struct poll_init_work;
+	bool poll_enabled;
+
+	unsigned int hpd_storm_threshold;
+	/* Whether or not to count short HPD IRQs in HPD storms */
+	u8 hpd_short_storm_enabled;
+
+	/*
+	 * if we get a HPD irq from DP and a HPD irq from non-DP
+	 * the non-DP HPD could block the workqueue on a mode config
+	 * mutex getting, that userspace may have taken. However
+	 * userspace is waiting on the DP workqueue to run which is
+	 * blocked behind the non-DP one.
+	 */
+	struct workqueue_struct *dp_wq;
+};
+
+struct intel_vbt_data {
+	/* bdb version */
+	u16 version;
+
+	/* Feature bits */
+	unsigned int int_tv_support:1;
+	unsigned int int_crt_support:1;
+	unsigned int lvds_use_ssc:1;
+	unsigned int int_lvds_support:1;
+	unsigned int display_clock_mode:1;
+	unsigned int fdi_rx_polarity_inverted:1;
+	int lvds_ssc_freq;
+	enum drm_panel_orientation orientation;
+
+	bool override_afc_startup;
+	u8 override_afc_startup_val;
+
+	int crt_ddc_pin;
+
+	struct list_head display_devices;
+	struct list_head bdb_blocks;
+
+	struct intel_bios_encoder_data *ports[I915_MAX_PORTS]; /* Non-NULL if port present. */
+	struct sdvo_device_mapping {
+		u8 initialized;
+		u8 dvo_port;
+		u8 slave_addr;
+		u8 dvo_wiring;
+		u8 i2c_pin;
+		u8 ddc_pin;
+	} sdvo_mappings[2];
+};
+
+struct intel_wm {
+	/*
+	 * Raw watermark latency values:
+	 * in 0.1us units for WM0,
+	 * in 0.5us units for WM1+.
+	 */
+	/* primary */
+	u16 pri_latency[5];
+	/* sprite */
+	u16 spr_latency[5];
+	/* cursor */
+	u16 cur_latency[5];
+	/*
+	 * Raw watermark memory latency values
+	 * for SKL for all 8 levels
+	 * in 1us units.
+	 */
+	u16 skl_latency[8];
+
+	/* current hardware state */
+	union {
+		struct ilk_wm_values hw;
+		struct vlv_wm_values vlv;
+		struct g4x_wm_values g4x;
+	};
+
+	u8 max_level;
+
+	/*
+	 * Should be held around atomic WM register writing; also
+	 * protects * intel_crtc->wm.active and
+	 * crtc_state->wm.need_postvbl_update.
+	 */
+	struct mutex wm_mutex;
+
+	bool ipc_enabled;
+};
+
+struct intel_display {
+	/* Display functions */
+	struct {
+		/* Top level crtc-ish functions */
+		const struct intel_display_funcs *display;
+
+		/* Display CDCLK functions */
+		const struct intel_cdclk_funcs *cdclk;
+
+		/* Display pll funcs */
+		const struct intel_dpll_funcs *dpll;
+
+		/* irq display functions */
+		const struct intel_hotplug_funcs *hotplug;
+
+		/* pm display functions */
+		const struct intel_wm_funcs *wm;
+
+		/* fdi display functions */
+		const struct intel_fdi_funcs *fdi;
+
+		/* Display internal color functions */
+		const struct intel_color_funcs *color;
+
+		/* Display internal audio functions */
+		const struct intel_audio_funcs *audio;
+	} funcs;
+
+	/* Grouping using anonymous structs. Keep sorted. */
+	struct intel_atomic_helper {
+		struct llist_head free_list;
+		struct work_struct free_work;
+	} atomic_helper;
+
+	struct {
+		/* backlight registers and fields in struct intel_panel */
+		struct mutex lock;
+	} backlight;
+
+	struct {
+		struct intel_global_obj obj;
+
+		struct intel_bw_info {
+			/* for each QGV point */
+			unsigned int deratedbw[I915_NUM_QGV_POINTS];
+			/* for each PSF GV point */
+			unsigned int psf_bw[I915_NUM_PSF_GV_POINTS];
+			u8 num_qgv_points;
+			u8 num_psf_gv_points;
+			u8 num_planes;
+		} max[6];
+	} bw;
+
+	struct {
+		/* The current hardware cdclk configuration */
+		struct intel_cdclk_config hw;
+
+		/* cdclk, divider, and ratio table from bspec */
+		const struct intel_cdclk_vals *table;
+
+		struct intel_global_obj obj;
+
+		unsigned int max_cdclk_freq;
+	} cdclk;
+
+	struct {
+		/* The current hardware dbuf configuration */
+		u8 enabled_slices;
+
+		struct intel_global_obj obj;
+	} dbuf;
+
+	struct {
+		/*
+		 * dkl.phy_lock protects against concurrent access of the
+		 * Dekel TypeC PHYs.
+		 */
+		spinlock_t phy_lock;
+	} dkl;
+
+	struct {
+		/* VLV/CHV/BXT/GLK DSI MMIO register base address */
+		u32 mmio_base;
+	} dsi;
+
+	struct {
+		/* list of fbdev register on this device */
+		struct intel_fbdev *fbdev;
+		struct work_struct suspend_work;
+	} fbdev;
+
+	struct {
+		unsigned int pll_freq;
+		u32 rx_config;
+	} fdi;
+
+	struct {
+		/*
+		 * Base address of where the gmbus and gpio blocks are located
+		 * (either on PCH or on SoC for platforms without PCH).
+		 */
+		u32 mmio_base;
+
+		/*
+		 * gmbus.mutex protects against concurrent usage of the single
+		 * hw gmbus controller on different i2c buses.
+		 */
+		struct mutex mutex;
+
+		struct intel_gmbus *bus[GMBUS_NUM_PINS];
+
+		wait_queue_head_t wait_queue;
+	} gmbus;
+
+	struct {
+		struct i915_hdcp_comp_master *master;
+		bool comp_added;
+
+		/* Mutex to protect the above hdcp component related values. */
+		struct mutex comp_mutex;
+	} hdcp;
+
+	struct {
+		struct i915_power_domains domains;
+
+		/* Shadow for DISPLAY_PHY_CONTROL which can't be safely read */
+		u32 chv_phy_control;
+
+		/* perform PHY state sanity checks? */
+		bool chv_phy_assert[2];
+	} power;
+
+	struct {
+		u32 mmio_base;
+
+		/* protects panel power sequencer state */
+		struct mutex mutex;
+	} pps;
+
+	struct {
+		struct drm_property *broadcast_rgb;
+		struct drm_property *force_audio;
+	} properties;
+
+	struct {
+		unsigned long mask;
+	} quirks;
+
+	struct {
+		enum {
+			I915_SAGV_UNKNOWN = 0,
+			I915_SAGV_DISABLED,
+			I915_SAGV_ENABLED,
+			I915_SAGV_NOT_CONTROLLED
+		} status;
+
+		u32 block_time_us;
+	} sagv;
+
+	struct {
+		/* ordered wq for modesets */
+		struct workqueue_struct *modeset;
+
+		/* unbound hipri wq for page flips/plane updates */
+		struct workqueue_struct *flip;
+	} wq;
+
+	/* Grouping using named structs. Keep sorted. */
+	struct intel_audio audio;
+	struct intel_dmc dmc;
+	struct intel_dpll dpll;
+	struct intel_fbc *fbc[I915_MAX_FBCS];
+	struct intel_frontbuffer_tracking fb_tracking;
+	struct intel_hotplug hotplug;
+	struct intel_opregion opregion;
+	struct intel_overlay *overlay;
+	struct intel_vbt_data vbt;
+	struct intel_wm wm;
+};
+
+#endif /* __INTEL_DISPLAY_CORE_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_display_debugfs.c b/drivers/gpu/drm/i915/display/intel_display_debugfs.c
new file mode 100644
index 000000000..7c7253a25
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_debugfs.c
@@ -0,0 +1,2254 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+
+#include <drm/drm_debugfs.h>
+#include <drm/drm_fourcc.h>
+
+#include "i915_debugfs.h"
+#include "intel_de.h"
+#include "intel_display_debugfs.h"
+#include "intel_display_power.h"
+#include "intel_display_power_well.h"
+#include "intel_display_types.h"
+#include "intel_dmc.h"
+#include "intel_dp.h"
+#include "intel_dp_mst.h"
+#include "intel_drrs.h"
+#include "intel_fbc.h"
+#include "intel_fbdev.h"
+#include "intel_hdcp.h"
+#include "intel_hdmi.h"
+#include "intel_panel.h"
+#include "intel_pm.h"
+#include "intel_psr.h"
+#include "intel_sprite.h"
+#include "skl_watermark.h"
+
+static inline struct drm_i915_private *node_to_i915(struct drm_info_node *node)
+{
+	return to_i915(node->minor->dev);
+}
+
+static int i915_frontbuffer_tracking(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+
+	seq_printf(m, "FB tracking busy bits: 0x%08x\n",
+		   dev_priv->display.fb_tracking.busy_bits);
+
+	seq_printf(m, "FB tracking flip bits: 0x%08x\n",
+		   dev_priv->display.fb_tracking.flip_bits);
+
+	return 0;
+}
+
+static int i915_ips_status(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	intel_wakeref_t wakeref;
+
+	if (!HAS_IPS(dev_priv))
+		return -ENODEV;
+
+	wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+
+	seq_printf(m, "Enabled by kernel parameter: %s\n",
+		   str_yes_no(dev_priv->params.enable_ips));
+
+	if (DISPLAY_VER(dev_priv) >= 8) {
+		seq_puts(m, "Currently: unknown\n");
+	} else {
+		if (intel_de_read(dev_priv, IPS_CTL) & IPS_ENABLE)
+			seq_puts(m, "Currently: enabled\n");
+		else
+			seq_puts(m, "Currently: disabled\n");
+	}
+
+	intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+
+	return 0;
+}
+
+static int i915_sr_status(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	intel_wakeref_t wakeref;
+	bool sr_enabled = false;
+
+	wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_INIT);
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		/* no global SR status; inspect per-plane WM */;
+	else if (HAS_PCH_SPLIT(dev_priv))
+		sr_enabled = intel_de_read(dev_priv, WM1_LP_ILK) & WM_LP_ENABLE;
+	else if (IS_I965GM(dev_priv) || IS_G4X(dev_priv) ||
+		 IS_I945G(dev_priv) || IS_I945GM(dev_priv))
+		sr_enabled = intel_de_read(dev_priv, FW_BLC_SELF) & FW_BLC_SELF_EN;
+	else if (IS_I915GM(dev_priv))
+		sr_enabled = intel_de_read(dev_priv, INSTPM) & INSTPM_SELF_EN;
+	else if (IS_PINEVIEW(dev_priv))
+		sr_enabled = intel_de_read(dev_priv, DSPFW3) & PINEVIEW_SELF_REFRESH_EN;
+	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		sr_enabled = intel_de_read(dev_priv, FW_BLC_SELF_VLV) & FW_CSPWRDWNEN;
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_INIT, wakeref);
+
+	seq_printf(m, "self-refresh: %s\n", str_enabled_disabled(sr_enabled));
+
+	return 0;
+}
+
+static int i915_opregion(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *i915 = node_to_i915(m->private);
+	struct intel_opregion *opregion = &i915->display.opregion;
+
+	if (opregion->header)
+		seq_write(m, opregion->header, OPREGION_SIZE);
+
+	return 0;
+}
+
+static int i915_vbt(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *i915 = node_to_i915(m->private);
+	struct intel_opregion *opregion = &i915->display.opregion;
+
+	if (opregion->vbt)
+		seq_write(m, opregion->vbt, opregion->vbt_size);
+
+	return 0;
+}
+
+static int i915_gem_framebuffer_info(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_framebuffer *fbdev_fb = NULL;
+	struct drm_framebuffer *drm_fb;
+
+#ifdef CONFIG_DRM_FBDEV_EMULATION
+	fbdev_fb = intel_fbdev_framebuffer(dev_priv->display.fbdev.fbdev);
+	if (fbdev_fb) {
+		seq_printf(m, "fbcon size: %d x %d, depth %d, %d bpp, modifier 0x%llx, refcount %d, obj ",
+			   fbdev_fb->base.width,
+			   fbdev_fb->base.height,
+			   fbdev_fb->base.format->depth,
+			   fbdev_fb->base.format->cpp[0] * 8,
+			   fbdev_fb->base.modifier,
+			   drm_framebuffer_read_refcount(&fbdev_fb->base));
+		i915_debugfs_describe_obj(m, intel_fb_obj(&fbdev_fb->base));
+		seq_putc(m, '\n');
+	}
+#endif
+
+	mutex_lock(&dev->mode_config.fb_lock);
+	drm_for_each_fb(drm_fb, dev) {
+		struct intel_framebuffer *fb = to_intel_framebuffer(drm_fb);
+		if (fb == fbdev_fb)
+			continue;
+
+		seq_printf(m, "user size: %d x %d, depth %d, %d bpp, modifier 0x%llx, refcount %d, obj ",
+			   fb->base.width,
+			   fb->base.height,
+			   fb->base.format->depth,
+			   fb->base.format->cpp[0] * 8,
+			   fb->base.modifier,
+			   drm_framebuffer_read_refcount(&fb->base));
+		i915_debugfs_describe_obj(m, intel_fb_obj(&fb->base));
+		seq_putc(m, '\n');
+	}
+	mutex_unlock(&dev->mode_config.fb_lock);
+
+	return 0;
+}
+
+static int i915_psr_sink_status_show(struct seq_file *m, void *data)
+{
+	u8 val;
+	static const char * const sink_status[] = {
+		"inactive",
+		"transition to active, capture and display",
+		"active, display from RFB",
+		"active, capture and display on sink device timings",
+		"transition to inactive, capture and display, timing re-sync",
+		"reserved",
+		"reserved",
+		"sink internal error",
+	};
+	struct drm_connector *connector = m->private;
+	struct intel_dp *intel_dp =
+		intel_attached_dp(to_intel_connector(connector));
+	int ret;
+
+	if (!CAN_PSR(intel_dp)) {
+		seq_puts(m, "PSR Unsupported\n");
+		return -ENODEV;
+	}
+
+	if (connector->status != connector_status_connected)
+		return -ENODEV;
+
+	ret = drm_dp_dpcd_readb(&intel_dp->aux, DP_PSR_STATUS, &val);
+
+	if (ret == 1) {
+		const char *str = "unknown";
+
+		val &= DP_PSR_SINK_STATE_MASK;
+		if (val < ARRAY_SIZE(sink_status))
+			str = sink_status[val];
+		seq_printf(m, "Sink PSR status: 0x%x [%s]\n", val, str);
+	} else {
+		return ret;
+	}
+
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(i915_psr_sink_status);
+
+static void
+psr_source_status(struct intel_dp *intel_dp, struct seq_file *m)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	const char *status = "unknown";
+	u32 val, status_val;
+
+	if (intel_dp->psr.psr2_enabled) {
+		static const char * const live_status[] = {
+			"IDLE",
+			"CAPTURE",
+			"CAPTURE_FS",
+			"SLEEP",
+			"BUFON_FW",
+			"ML_UP",
+			"SU_STANDBY",
+			"FAST_SLEEP",
+			"DEEP_SLEEP",
+			"BUF_ON",
+			"TG_ON"
+		};
+		val = intel_de_read(dev_priv,
+				    EDP_PSR2_STATUS(intel_dp->psr.transcoder));
+		status_val = REG_FIELD_GET(EDP_PSR2_STATUS_STATE_MASK, val);
+		if (status_val < ARRAY_SIZE(live_status))
+			status = live_status[status_val];
+	} else {
+		static const char * const live_status[] = {
+			"IDLE",
+			"SRDONACK",
+			"SRDENT",
+			"BUFOFF",
+			"BUFON",
+			"AUXACK",
+			"SRDOFFACK",
+			"SRDENT_ON",
+		};
+		val = intel_de_read(dev_priv,
+				    EDP_PSR_STATUS(intel_dp->psr.transcoder));
+		status_val = (val & EDP_PSR_STATUS_STATE_MASK) >>
+			      EDP_PSR_STATUS_STATE_SHIFT;
+		if (status_val < ARRAY_SIZE(live_status))
+			status = live_status[status_val];
+	}
+
+	seq_printf(m, "Source PSR status: %s [0x%08x]\n", status, val);
+}
+
+static int intel_psr_status(struct seq_file *m, struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_psr *psr = &intel_dp->psr;
+	intel_wakeref_t wakeref;
+	const char *status;
+	bool enabled;
+	u32 val;
+
+	seq_printf(m, "Sink support: %s", str_yes_no(psr->sink_support));
+	if (psr->sink_support)
+		seq_printf(m, " [0x%02x]", intel_dp->psr_dpcd[0]);
+	seq_puts(m, "\n");
+
+	if (!psr->sink_support)
+		return 0;
+
+	wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+	mutex_lock(&psr->lock);
+
+	if (psr->enabled)
+		status = psr->psr2_enabled ? "PSR2 enabled" : "PSR1 enabled";
+	else
+		status = "disabled";
+	seq_printf(m, "PSR mode: %s\n", status);
+
+	if (!psr->enabled) {
+		seq_printf(m, "PSR sink not reliable: %s\n",
+			   str_yes_no(psr->sink_not_reliable));
+
+		goto unlock;
+	}
+
+	if (psr->psr2_enabled) {
+		val = intel_de_read(dev_priv,
+				    EDP_PSR2_CTL(intel_dp->psr.transcoder));
+		enabled = val & EDP_PSR2_ENABLE;
+	} else {
+		val = intel_de_read(dev_priv,
+				    EDP_PSR_CTL(intel_dp->psr.transcoder));
+		enabled = val & EDP_PSR_ENABLE;
+	}
+	seq_printf(m, "Source PSR ctl: %s [0x%08x]\n",
+		   str_enabled_disabled(enabled), val);
+	psr_source_status(intel_dp, m);
+	seq_printf(m, "Busy frontbuffer bits: 0x%08x\n",
+		   psr->busy_frontbuffer_bits);
+
+	/*
+	 * SKL+ Perf counter is reset to 0 everytime DC state is entered
+	 */
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
+		val = intel_de_read(dev_priv,
+				    EDP_PSR_PERF_CNT(intel_dp->psr.transcoder));
+		val &= EDP_PSR_PERF_CNT_MASK;
+		seq_printf(m, "Performance counter: %u\n", val);
+	}
+
+	if (psr->debug & I915_PSR_DEBUG_IRQ) {
+		seq_printf(m, "Last attempted entry at: %lld\n",
+			   psr->last_entry_attempt);
+		seq_printf(m, "Last exit at: %lld\n", psr->last_exit);
+	}
+
+	if (psr->psr2_enabled) {
+		u32 su_frames_val[3];
+		int frame;
+
+		/*
+		 * Reading all 3 registers before hand to minimize crossing a
+		 * frame boundary between register reads
+		 */
+		for (frame = 0; frame < PSR2_SU_STATUS_FRAMES; frame += 3) {
+			val = intel_de_read(dev_priv,
+					    PSR2_SU_STATUS(intel_dp->psr.transcoder, frame));
+			su_frames_val[frame / 3] = val;
+		}
+
+		seq_puts(m, "Frame:\tPSR2 SU blocks:\n");
+
+		for (frame = 0; frame < PSR2_SU_STATUS_FRAMES; frame++) {
+			u32 su_blocks;
+
+			su_blocks = su_frames_val[frame / 3] &
+				    PSR2_SU_STATUS_MASK(frame);
+			su_blocks = su_blocks >> PSR2_SU_STATUS_SHIFT(frame);
+			seq_printf(m, "%d\t%d\n", frame, su_blocks);
+		}
+
+		seq_printf(m, "PSR2 selective fetch: %s\n",
+			   str_enabled_disabled(psr->psr2_sel_fetch_enabled));
+	}
+
+unlock:
+	mutex_unlock(&psr->lock);
+	intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+
+	return 0;
+}
+
+static int i915_edp_psr_status(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct intel_dp *intel_dp = NULL;
+	struct intel_encoder *encoder;
+
+	if (!HAS_PSR(dev_priv))
+		return -ENODEV;
+
+	/* Find the first EDP which supports PSR */
+	for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) {
+		intel_dp = enc_to_intel_dp(encoder);
+		break;
+	}
+
+	if (!intel_dp)
+		return -ENODEV;
+
+	return intel_psr_status(m, intel_dp);
+}
+
+static int
+i915_edp_psr_debug_set(void *data, u64 val)
+{
+	struct drm_i915_private *dev_priv = data;
+	struct intel_encoder *encoder;
+	intel_wakeref_t wakeref;
+	int ret = -ENODEV;
+
+	if (!HAS_PSR(dev_priv))
+		return ret;
+
+	for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		drm_dbg_kms(&dev_priv->drm, "Setting PSR debug to %llx\n", val);
+
+		wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+
+		// TODO: split to each transcoder's PSR debug state
+		ret = intel_psr_debug_set(intel_dp, val);
+
+		intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+	}
+
+	return ret;
+}
+
+static int
+i915_edp_psr_debug_get(void *data, u64 *val)
+{
+	struct drm_i915_private *dev_priv = data;
+	struct intel_encoder *encoder;
+
+	if (!HAS_PSR(dev_priv))
+		return -ENODEV;
+
+	for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		// TODO: split to each transcoder's PSR debug state
+		*val = READ_ONCE(intel_dp->psr.debug);
+		return 0;
+	}
+
+	return -ENODEV;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(i915_edp_psr_debug_fops,
+			i915_edp_psr_debug_get, i915_edp_psr_debug_set,
+			"%llu\n");
+
+static int i915_power_domain_info(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *i915 = node_to_i915(m->private);
+
+	intel_display_power_debug(i915, m);
+
+	return 0;
+}
+
+static void intel_seq_print_mode(struct seq_file *m, int tabs,
+				 const struct drm_display_mode *mode)
+{
+	int i;
+
+	for (i = 0; i < tabs; i++)
+		seq_putc(m, '\t');
+
+	seq_printf(m, DRM_MODE_FMT "\n", DRM_MODE_ARG(mode));
+}
+
+static void intel_encoder_info(struct seq_file *m,
+			       struct intel_crtc *crtc,
+			       struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct drm_connector_list_iter conn_iter;
+	struct drm_connector *connector;
+
+	seq_printf(m, "\t[ENCODER:%d:%s]: connectors:\n",
+		   encoder->base.base.id, encoder->base.name);
+
+	drm_connector_list_iter_begin(&dev_priv->drm, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		const struct drm_connector_state *conn_state =
+			connector->state;
+
+		if (conn_state->best_encoder != &encoder->base)
+			continue;
+
+		seq_printf(m, "\t\t[CONNECTOR:%d:%s]\n",
+			   connector->base.id, connector->name);
+	}
+	drm_connector_list_iter_end(&conn_iter);
+}
+
+static void intel_panel_info(struct seq_file *m,
+			     struct intel_connector *connector)
+{
+	const struct drm_display_mode *fixed_mode;
+
+	if (list_empty(&connector->panel.fixed_modes))
+		return;
+
+	seq_puts(m, "\tfixed modes:\n");
+
+	list_for_each_entry(fixed_mode, &connector->panel.fixed_modes, head)
+		intel_seq_print_mode(m, 2, fixed_mode);
+}
+
+static void intel_hdcp_info(struct seq_file *m,
+			    struct intel_connector *intel_connector)
+{
+	bool hdcp_cap, hdcp2_cap;
+
+	if (!intel_connector->hdcp.shim) {
+		seq_puts(m, "No Connector Support");
+		goto out;
+	}
+
+	hdcp_cap = intel_hdcp_capable(intel_connector);
+	hdcp2_cap = intel_hdcp2_capable(intel_connector);
+
+	if (hdcp_cap)
+		seq_puts(m, "HDCP1.4 ");
+	if (hdcp2_cap)
+		seq_puts(m, "HDCP2.2 ");
+
+	if (!hdcp_cap && !hdcp2_cap)
+		seq_puts(m, "None");
+
+out:
+	seq_puts(m, "\n");
+}
+
+static void intel_dp_info(struct seq_file *m,
+			  struct intel_connector *intel_connector)
+{
+	struct intel_encoder *intel_encoder = intel_attached_encoder(intel_connector);
+	struct intel_dp *intel_dp = enc_to_intel_dp(intel_encoder);
+	const struct drm_property_blob *edid = intel_connector->base.edid_blob_ptr;
+
+	seq_printf(m, "\tDPCD rev: %x\n", intel_dp->dpcd[DP_DPCD_REV]);
+	seq_printf(m, "\taudio support: %s\n",
+		   str_yes_no(intel_dp->has_audio));
+
+	drm_dp_downstream_debug(m, intel_dp->dpcd, intel_dp->downstream_ports,
+				edid ? edid->data : NULL, &intel_dp->aux);
+}
+
+static void intel_dp_mst_info(struct seq_file *m,
+			      struct intel_connector *intel_connector)
+{
+	bool has_audio = intel_connector->port->has_audio;
+
+	seq_printf(m, "\taudio support: %s\n", str_yes_no(has_audio));
+}
+
+static void intel_hdmi_info(struct seq_file *m,
+			    struct intel_connector *intel_connector)
+{
+	struct intel_encoder *intel_encoder = intel_attached_encoder(intel_connector);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(intel_encoder);
+
+	seq_printf(m, "\taudio support: %s\n",
+		   str_yes_no(intel_hdmi->has_audio));
+}
+
+static void intel_connector_info(struct seq_file *m,
+				 struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	const struct drm_connector_state *conn_state = connector->state;
+	struct intel_encoder *encoder =
+		to_intel_encoder(conn_state->best_encoder);
+	const struct drm_display_mode *mode;
+
+	seq_printf(m, "[CONNECTOR:%d:%s]: status: %s\n",
+		   connector->base.id, connector->name,
+		   drm_get_connector_status_name(connector->status));
+
+	if (connector->status == connector_status_disconnected)
+		return;
+
+	seq_printf(m, "\tphysical dimensions: %dx%dmm\n",
+		   connector->display_info.width_mm,
+		   connector->display_info.height_mm);
+	seq_printf(m, "\tsubpixel order: %s\n",
+		   drm_get_subpixel_order_name(connector->display_info.subpixel_order));
+	seq_printf(m, "\tCEA rev: %d\n", connector->display_info.cea_rev);
+
+	if (!encoder)
+		return;
+
+	switch (connector->connector_type) {
+	case DRM_MODE_CONNECTOR_DisplayPort:
+	case DRM_MODE_CONNECTOR_eDP:
+		if (encoder->type == INTEL_OUTPUT_DP_MST)
+			intel_dp_mst_info(m, intel_connector);
+		else
+			intel_dp_info(m, intel_connector);
+		break;
+	case DRM_MODE_CONNECTOR_HDMIA:
+		if (encoder->type == INTEL_OUTPUT_HDMI ||
+		    encoder->type == INTEL_OUTPUT_DDI)
+			intel_hdmi_info(m, intel_connector);
+		break;
+	default:
+		break;
+	}
+
+	seq_puts(m, "\tHDCP version: ");
+	intel_hdcp_info(m, intel_connector);
+
+	seq_printf(m, "\tmax bpc: %u\n", connector->display_info.bpc);
+
+	intel_panel_info(m, intel_connector);
+
+	seq_printf(m, "\tmodes:\n");
+	list_for_each_entry(mode, &connector->modes, head)
+		intel_seq_print_mode(m, 2, mode);
+}
+
+static const char *plane_type(enum drm_plane_type type)
+{
+	switch (type) {
+	case DRM_PLANE_TYPE_OVERLAY:
+		return "OVL";
+	case DRM_PLANE_TYPE_PRIMARY:
+		return "PRI";
+	case DRM_PLANE_TYPE_CURSOR:
+		return "CUR";
+	/*
+	 * Deliberately omitting default: to generate compiler warnings
+	 * when a new drm_plane_type gets added.
+	 */
+	}
+
+	return "unknown";
+}
+
+static void plane_rotation(char *buf, size_t bufsize, unsigned int rotation)
+{
+	/*
+	 * According to doc only one DRM_MODE_ROTATE_ is allowed but this
+	 * will print them all to visualize if the values are misused
+	 */
+	snprintf(buf, bufsize,
+		 "%s%s%s%s%s%s(0x%08x)",
+		 (rotation & DRM_MODE_ROTATE_0) ? "0 " : "",
+		 (rotation & DRM_MODE_ROTATE_90) ? "90 " : "",
+		 (rotation & DRM_MODE_ROTATE_180) ? "180 " : "",
+		 (rotation & DRM_MODE_ROTATE_270) ? "270 " : "",
+		 (rotation & DRM_MODE_REFLECT_X) ? "FLIPX " : "",
+		 (rotation & DRM_MODE_REFLECT_Y) ? "FLIPY " : "",
+		 rotation);
+}
+
+static const char *plane_visibility(const struct intel_plane_state *plane_state)
+{
+	if (plane_state->uapi.visible)
+		return "visible";
+
+	if (plane_state->planar_slave)
+		return "planar-slave";
+
+	return "hidden";
+}
+
+static void intel_plane_uapi_info(struct seq_file *m, struct intel_plane *plane)
+{
+	const struct intel_plane_state *plane_state =
+		to_intel_plane_state(plane->base.state);
+	const struct drm_framebuffer *fb = plane_state->uapi.fb;
+	struct drm_rect src, dst;
+	char rot_str[48];
+
+	src = drm_plane_state_src(&plane_state->uapi);
+	dst = drm_plane_state_dest(&plane_state->uapi);
+
+	plane_rotation(rot_str, sizeof(rot_str),
+		       plane_state->uapi.rotation);
+
+	seq_puts(m, "\t\tuapi: [FB:");
+	if (fb)
+		seq_printf(m, "%d] %p4cc,0x%llx,%dx%d", fb->base.id,
+			   &fb->format->format, fb->modifier, fb->width,
+			   fb->height);
+	else
+		seq_puts(m, "0] n/a,0x0,0x0,");
+	seq_printf(m, ", visible=%s, src=" DRM_RECT_FP_FMT ", dst=" DRM_RECT_FMT
+		   ", rotation=%s\n", plane_visibility(plane_state),
+		   DRM_RECT_FP_ARG(&src), DRM_RECT_ARG(&dst), rot_str);
+
+	if (plane_state->planar_linked_plane)
+		seq_printf(m, "\t\tplanar: Linked to [PLANE:%d:%s] as a %s\n",
+			   plane_state->planar_linked_plane->base.base.id, plane_state->planar_linked_plane->base.name,
+			   plane_state->planar_slave ? "slave" : "master");
+}
+
+static void intel_plane_hw_info(struct seq_file *m, struct intel_plane *plane)
+{
+	const struct intel_plane_state *plane_state =
+		to_intel_plane_state(plane->base.state);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	char rot_str[48];
+
+	if (!fb)
+		return;
+
+	plane_rotation(rot_str, sizeof(rot_str),
+		       plane_state->hw.rotation);
+
+	seq_printf(m, "\t\thw: [FB:%d] %p4cc,0x%llx,%dx%d, visible=%s, src="
+		   DRM_RECT_FP_FMT ", dst=" DRM_RECT_FMT ", rotation=%s\n",
+		   fb->base.id, &fb->format->format,
+		   fb->modifier, fb->width, fb->height,
+		   str_yes_no(plane_state->uapi.visible),
+		   DRM_RECT_FP_ARG(&plane_state->uapi.src),
+		   DRM_RECT_ARG(&plane_state->uapi.dst),
+		   rot_str);
+}
+
+static void intel_plane_info(struct seq_file *m, struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct intel_plane *plane;
+
+	for_each_intel_plane_on_crtc(&dev_priv->drm, crtc, plane) {
+		seq_printf(m, "\t[PLANE:%d:%s]: type=%s\n",
+			   plane->base.base.id, plane->base.name,
+			   plane_type(plane->base.type));
+		intel_plane_uapi_info(m, plane);
+		intel_plane_hw_info(m, plane);
+	}
+}
+
+static void intel_scaler_info(struct seq_file *m, struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *crtc_state =
+		to_intel_crtc_state(crtc->base.state);
+	int num_scalers = crtc->num_scalers;
+	int i;
+
+	/* Not all platformas have a scaler */
+	if (num_scalers) {
+		seq_printf(m, "\tnum_scalers=%d, scaler_users=%x scaler_id=%d scaling_filter=%d",
+			   num_scalers,
+			   crtc_state->scaler_state.scaler_users,
+			   crtc_state->scaler_state.scaler_id,
+			   crtc_state->hw.scaling_filter);
+
+		for (i = 0; i < num_scalers; i++) {
+			const struct intel_scaler *sc =
+				&crtc_state->scaler_state.scalers[i];
+
+			seq_printf(m, ", scalers[%d]: use=%s, mode=%x",
+				   i, str_yes_no(sc->in_use), sc->mode);
+		}
+		seq_puts(m, "\n");
+	} else {
+		seq_puts(m, "\tNo scalers available on this platform\n");
+	}
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_VBLANK_EVADE)
+static void crtc_updates_info(struct seq_file *m,
+			      struct intel_crtc *crtc,
+			      const char *hdr)
+{
+	u64 count;
+	int row;
+
+	count = 0;
+	for (row = 0; row < ARRAY_SIZE(crtc->debug.vbl.times); row++)
+		count += crtc->debug.vbl.times[row];
+	seq_printf(m, "%sUpdates: %llu\n", hdr, count);
+	if (!count)
+		return;
+
+	for (row = 0; row < ARRAY_SIZE(crtc->debug.vbl.times); row++) {
+		char columns[80] = "       |";
+		unsigned int x;
+
+		if (row & 1) {
+			const char *units;
+
+			if (row > 10) {
+				x = 1000000;
+				units = "ms";
+			} else {
+				x = 1000;
+				units = "us";
+			}
+
+			snprintf(columns, sizeof(columns), "%4ld%s |",
+				 DIV_ROUND_CLOSEST(BIT(row + 9), x), units);
+		}
+
+		if (crtc->debug.vbl.times[row]) {
+			x = ilog2(crtc->debug.vbl.times[row]);
+			memset(columns + 8, '*', x);
+			columns[8 + x] = '\0';
+		}
+
+		seq_printf(m, "%s%s\n", hdr, columns);
+	}
+
+	seq_printf(m, "%sMin update: %lluns\n",
+		   hdr, crtc->debug.vbl.min);
+	seq_printf(m, "%sMax update: %lluns\n",
+		   hdr, crtc->debug.vbl.max);
+	seq_printf(m, "%sAverage update: %lluns\n",
+		   hdr, div64_u64(crtc->debug.vbl.sum,  count));
+	seq_printf(m, "%sOverruns > %uus: %u\n",
+		   hdr, VBLANK_EVASION_TIME_US, crtc->debug.vbl.over);
+}
+
+static int crtc_updates_show(struct seq_file *m, void *data)
+{
+	crtc_updates_info(m, m->private, "");
+	return 0;
+}
+
+static int crtc_updates_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, crtc_updates_show, inode->i_private);
+}
+
+static ssize_t crtc_updates_write(struct file *file,
+				  const char __user *ubuf,
+				  size_t len, loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct intel_crtc *crtc = m->private;
+
+	/* May race with an update. Meh. */
+	memset(&crtc->debug.vbl, 0, sizeof(crtc->debug.vbl));
+
+	return len;
+}
+
+static const struct file_operations crtc_updates_fops = {
+	.owner = THIS_MODULE,
+	.open = crtc_updates_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = crtc_updates_write
+};
+
+static void crtc_updates_add(struct drm_crtc *crtc)
+{
+	debugfs_create_file("i915_update_info", 0644, crtc->debugfs_entry,
+			    to_intel_crtc(crtc), &crtc_updates_fops);
+}
+
+#else
+static void crtc_updates_info(struct seq_file *m,
+			      struct intel_crtc *crtc,
+			      const char *hdr)
+{
+}
+
+static void crtc_updates_add(struct drm_crtc *crtc)
+{
+}
+#endif
+
+static void intel_crtc_info(struct seq_file *m, struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	const struct intel_crtc_state *crtc_state =
+		to_intel_crtc_state(crtc->base.state);
+	struct intel_encoder *encoder;
+
+	seq_printf(m, "[CRTC:%d:%s]:\n",
+		   crtc->base.base.id, crtc->base.name);
+
+	seq_printf(m, "\tuapi: enable=%s, active=%s, mode=" DRM_MODE_FMT "\n",
+		   str_yes_no(crtc_state->uapi.enable),
+		   str_yes_no(crtc_state->uapi.active),
+		   DRM_MODE_ARG(&crtc_state->uapi.mode));
+
+	seq_printf(m, "\thw: enable=%s, active=%s\n",
+		   str_yes_no(crtc_state->hw.enable), str_yes_no(crtc_state->hw.active));
+	seq_printf(m, "\tadjusted_mode=" DRM_MODE_FMT "\n",
+		   DRM_MODE_ARG(&crtc_state->hw.adjusted_mode));
+	seq_printf(m, "\tpipe__mode=" DRM_MODE_FMT "\n",
+		   DRM_MODE_ARG(&crtc_state->hw.pipe_mode));
+
+	seq_printf(m, "\tpipe src=" DRM_RECT_FMT ", dither=%s, bpp=%d\n",
+		   DRM_RECT_ARG(&crtc_state->pipe_src),
+		   str_yes_no(crtc_state->dither), crtc_state->pipe_bpp);
+
+	intel_scaler_info(m, crtc);
+
+	if (crtc_state->bigjoiner_pipes)
+		seq_printf(m, "\tLinked to 0x%x pipes as a %s\n",
+			   crtc_state->bigjoiner_pipes,
+			   intel_crtc_is_bigjoiner_slave(crtc_state) ? "slave" : "master");
+
+	for_each_intel_encoder_mask(&dev_priv->drm, encoder,
+				    crtc_state->uapi.encoder_mask)
+		intel_encoder_info(m, crtc, encoder);
+
+	intel_plane_info(m, crtc);
+
+	seq_printf(m, "\tunderrun reporting: cpu=%s pch=%s\n",
+		   str_yes_no(!crtc->cpu_fifo_underrun_disabled),
+		   str_yes_no(!crtc->pch_fifo_underrun_disabled));
+
+	crtc_updates_info(m, crtc, "\t");
+}
+
+static int i915_display_info(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_crtc *crtc;
+	struct drm_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	intel_wakeref_t wakeref;
+
+	wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+
+	drm_modeset_lock_all(dev);
+
+	seq_printf(m, "CRTC info\n");
+	seq_printf(m, "---------\n");
+	for_each_intel_crtc(dev, crtc)
+		intel_crtc_info(m, crtc);
+
+	seq_printf(m, "\n");
+	seq_printf(m, "Connector info\n");
+	seq_printf(m, "--------------\n");
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter)
+		intel_connector_info(m, connector);
+	drm_connector_list_iter_end(&conn_iter);
+
+	drm_modeset_unlock_all(dev);
+
+	intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+
+	return 0;
+}
+
+static int i915_shared_dplls_info(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct drm_device *dev = &dev_priv->drm;
+	int i;
+
+	drm_modeset_lock_all(dev);
+
+	seq_printf(m, "PLL refclks: non-SSC: %d kHz, SSC: %d kHz\n",
+		   dev_priv->display.dpll.ref_clks.nssc,
+		   dev_priv->display.dpll.ref_clks.ssc);
+
+	for (i = 0; i < dev_priv->display.dpll.num_shared_dpll; i++) {
+		struct intel_shared_dpll *pll = &dev_priv->display.dpll.shared_dplls[i];
+
+		seq_printf(m, "DPLL%i: %s, id: %i\n", i, pll->info->name,
+			   pll->info->id);
+		seq_printf(m, " pipe_mask: 0x%x, active: 0x%x, on: %s\n",
+			   pll->state.pipe_mask, pll->active_mask,
+			   str_yes_no(pll->on));
+		seq_printf(m, " tracked hardware state:\n");
+		seq_printf(m, " dpll:    0x%08x\n", pll->state.hw_state.dpll);
+		seq_printf(m, " dpll_md: 0x%08x\n",
+			   pll->state.hw_state.dpll_md);
+		seq_printf(m, " fp0:     0x%08x\n", pll->state.hw_state.fp0);
+		seq_printf(m, " fp1:     0x%08x\n", pll->state.hw_state.fp1);
+		seq_printf(m, " wrpll:   0x%08x\n", pll->state.hw_state.wrpll);
+		seq_printf(m, " cfgcr0:  0x%08x\n", pll->state.hw_state.cfgcr0);
+		seq_printf(m, " cfgcr1:  0x%08x\n", pll->state.hw_state.cfgcr1);
+		seq_printf(m, " div0:    0x%08x\n", pll->state.hw_state.div0);
+		seq_printf(m, " mg_refclkin_ctl:        0x%08x\n",
+			   pll->state.hw_state.mg_refclkin_ctl);
+		seq_printf(m, " mg_clktop2_coreclkctl1: 0x%08x\n",
+			   pll->state.hw_state.mg_clktop2_coreclkctl1);
+		seq_printf(m, " mg_clktop2_hsclkctl:    0x%08x\n",
+			   pll->state.hw_state.mg_clktop2_hsclkctl);
+		seq_printf(m, " mg_pll_div0:  0x%08x\n",
+			   pll->state.hw_state.mg_pll_div0);
+		seq_printf(m, " mg_pll_div1:  0x%08x\n",
+			   pll->state.hw_state.mg_pll_div1);
+		seq_printf(m, " mg_pll_lf:    0x%08x\n",
+			   pll->state.hw_state.mg_pll_lf);
+		seq_printf(m, " mg_pll_frac_lock: 0x%08x\n",
+			   pll->state.hw_state.mg_pll_frac_lock);
+		seq_printf(m, " mg_pll_ssc:   0x%08x\n",
+			   pll->state.hw_state.mg_pll_ssc);
+		seq_printf(m, " mg_pll_bias:  0x%08x\n",
+			   pll->state.hw_state.mg_pll_bias);
+		seq_printf(m, " mg_pll_tdc_coldst_bias: 0x%08x\n",
+			   pll->state.hw_state.mg_pll_tdc_coldst_bias);
+	}
+	drm_modeset_unlock_all(dev);
+
+	return 0;
+}
+
+static int i915_ddb_info(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct drm_device *dev = &dev_priv->drm;
+	struct skl_ddb_entry *entry;
+	struct intel_crtc *crtc;
+
+	if (DISPLAY_VER(dev_priv) < 9)
+		return -ENODEV;
+
+	drm_modeset_lock_all(dev);
+
+	seq_printf(m, "%-15s%8s%8s%8s\n", "", "Start", "End", "Size");
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		enum pipe pipe = crtc->pipe;
+		enum plane_id plane_id;
+
+		seq_printf(m, "Pipe %c\n", pipe_name(pipe));
+
+		for_each_plane_id_on_crtc(crtc, plane_id) {
+			entry = &crtc_state->wm.skl.plane_ddb[plane_id];
+			seq_printf(m, "  Plane%-8d%8u%8u%8u\n", plane_id + 1,
+				   entry->start, entry->end,
+				   skl_ddb_entry_size(entry));
+		}
+
+		entry = &crtc_state->wm.skl.plane_ddb[PLANE_CURSOR];
+		seq_printf(m, "  %-13s%8u%8u%8u\n", "Cursor", entry->start,
+			   entry->end, skl_ddb_entry_size(entry));
+	}
+
+	drm_modeset_unlock_all(dev);
+
+	return 0;
+}
+
+static int i915_drrs_status(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct drm_connector_list_iter conn_iter;
+	struct intel_connector *connector;
+	struct intel_crtc *crtc;
+
+	drm_connector_list_iter_begin(&dev_priv->drm, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		seq_printf(m, "[CONNECTOR:%d:%s] DRRS type: %s\n",
+			   connector->base.base.id, connector->base.name,
+			   intel_drrs_type_str(intel_panel_drrs_type(connector)));
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	seq_puts(m, "\n");
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		const struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		seq_printf(m, "[CRTC:%d:%s]:\n",
+			   crtc->base.base.id, crtc->base.name);
+
+		mutex_lock(&crtc->drrs.mutex);
+
+		/* DRRS Supported */
+		seq_printf(m, "\tDRRS Enabled: %s\n",
+			   str_yes_no(crtc_state->has_drrs));
+
+		seq_printf(m, "\tDRRS Active: %s\n",
+			   str_yes_no(intel_drrs_is_active(crtc)));
+
+		seq_printf(m, "\tBusy_frontbuffer_bits: 0x%X\n",
+			   crtc->drrs.busy_frontbuffer_bits);
+
+		seq_printf(m, "\tDRRS refresh rate: %s\n",
+			   crtc->drrs.refresh_rate == DRRS_REFRESH_RATE_LOW ?
+			   "low" : "high");
+
+		mutex_unlock(&crtc->drrs.mutex);
+	}
+
+	return 0;
+}
+
+static bool
+intel_lpsp_power_well_enabled(struct drm_i915_private *i915,
+			      enum i915_power_well_id power_well_id)
+{
+	intel_wakeref_t wakeref;
+	bool is_enabled;
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+	is_enabled = intel_display_power_well_is_enabled(i915,
+							 power_well_id);
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+
+	return is_enabled;
+}
+
+static int i915_lpsp_status(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *i915 = node_to_i915(m->private);
+	bool lpsp_enabled = false;
+
+	if (DISPLAY_VER(i915) >= 13 || IS_DISPLAY_VER(i915, 9, 10)) {
+		lpsp_enabled = !intel_lpsp_power_well_enabled(i915, SKL_DISP_PW_2);
+	} else if (IS_DISPLAY_VER(i915, 11, 12)) {
+		lpsp_enabled = !intel_lpsp_power_well_enabled(i915, ICL_DISP_PW_3);
+	} else if (IS_HASWELL(i915) || IS_BROADWELL(i915)) {
+		lpsp_enabled = !intel_lpsp_power_well_enabled(i915, HSW_DISP_PW_GLOBAL);
+	} else {
+		seq_puts(m, "LPSP: not supported\n");
+		return 0;
+	}
+
+	seq_printf(m, "LPSP: %s\n", str_enabled_disabled(lpsp_enabled));
+
+	return 0;
+}
+
+static int i915_dp_mst_info(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_encoder *intel_encoder;
+	struct intel_digital_port *dig_port;
+	struct drm_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		if (connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort)
+			continue;
+
+		intel_encoder = intel_attached_encoder(to_intel_connector(connector));
+		if (!intel_encoder || intel_encoder->type == INTEL_OUTPUT_DP_MST)
+			continue;
+
+		dig_port = enc_to_dig_port(intel_encoder);
+		if (!intel_dp_mst_source_support(&dig_port->dp))
+			continue;
+
+		seq_printf(m, "MST Source Port [ENCODER:%d:%s]\n",
+			   dig_port->base.base.base.id,
+			   dig_port->base.base.name);
+		drm_dp_mst_dump_topology(m, &dig_port->dp.mst_mgr);
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	return 0;
+}
+
+static ssize_t i915_displayport_test_active_write(struct file *file,
+						  const char __user *ubuf,
+						  size_t len, loff_t *offp)
+{
+	char *input_buffer;
+	int status = 0;
+	struct drm_device *dev;
+	struct drm_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	struct intel_dp *intel_dp;
+	int val = 0;
+
+	dev = ((struct seq_file *)file->private_data)->private;
+
+	if (len == 0)
+		return 0;
+
+	input_buffer = memdup_user_nul(ubuf, len);
+	if (IS_ERR(input_buffer))
+		return PTR_ERR(input_buffer);
+
+	drm_dbg(&to_i915(dev)->drm,
+		"Copied %d bytes from user\n", (unsigned int)len);
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		struct intel_encoder *encoder;
+
+		if (connector->connector_type !=
+		    DRM_MODE_CONNECTOR_DisplayPort)
+			continue;
+
+		encoder = to_intel_encoder(connector->encoder);
+		if (encoder && encoder->type == INTEL_OUTPUT_DP_MST)
+			continue;
+
+		if (encoder && connector->status == connector_status_connected) {
+			intel_dp = enc_to_intel_dp(encoder);
+			status = kstrtoint(input_buffer, 10, &val);
+			if (status < 0)
+				break;
+			drm_dbg(&to_i915(dev)->drm,
+				"Got %d for test active\n", val);
+			/* To prevent erroneous activation of the compliance
+			 * testing code, only accept an actual value of 1 here
+			 */
+			if (val == 1)
+				intel_dp->compliance.test_active = true;
+			else
+				intel_dp->compliance.test_active = false;
+		}
+	}
+	drm_connector_list_iter_end(&conn_iter);
+	kfree(input_buffer);
+	if (status < 0)
+		return status;
+
+	*offp += len;
+	return len;
+}
+
+static int i915_displayport_test_active_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = m->private;
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	struct intel_dp *intel_dp;
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		struct intel_encoder *encoder;
+
+		if (connector->connector_type !=
+		    DRM_MODE_CONNECTOR_DisplayPort)
+			continue;
+
+		encoder = to_intel_encoder(connector->encoder);
+		if (encoder && encoder->type == INTEL_OUTPUT_DP_MST)
+			continue;
+
+		if (encoder && connector->status == connector_status_connected) {
+			intel_dp = enc_to_intel_dp(encoder);
+			if (intel_dp->compliance.test_active)
+				seq_puts(m, "1");
+			else
+				seq_puts(m, "0");
+		} else
+			seq_puts(m, "0");
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	return 0;
+}
+
+static int i915_displayport_test_active_open(struct inode *inode,
+					     struct file *file)
+{
+	return single_open(file, i915_displayport_test_active_show,
+			   inode->i_private);
+}
+
+static const struct file_operations i915_displayport_test_active_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_displayport_test_active_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = i915_displayport_test_active_write
+};
+
+static int i915_displayport_test_data_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = m->private;
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	struct intel_dp *intel_dp;
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		struct intel_encoder *encoder;
+
+		if (connector->connector_type !=
+		    DRM_MODE_CONNECTOR_DisplayPort)
+			continue;
+
+		encoder = to_intel_encoder(connector->encoder);
+		if (encoder && encoder->type == INTEL_OUTPUT_DP_MST)
+			continue;
+
+		if (encoder && connector->status == connector_status_connected) {
+			intel_dp = enc_to_intel_dp(encoder);
+			if (intel_dp->compliance.test_type ==
+			    DP_TEST_LINK_EDID_READ)
+				seq_printf(m, "%lx",
+					   intel_dp->compliance.test_data.edid);
+			else if (intel_dp->compliance.test_type ==
+				 DP_TEST_LINK_VIDEO_PATTERN) {
+				seq_printf(m, "hdisplay: %d\n",
+					   intel_dp->compliance.test_data.hdisplay);
+				seq_printf(m, "vdisplay: %d\n",
+					   intel_dp->compliance.test_data.vdisplay);
+				seq_printf(m, "bpc: %u\n",
+					   intel_dp->compliance.test_data.bpc);
+			} else if (intel_dp->compliance.test_type ==
+				   DP_TEST_LINK_PHY_TEST_PATTERN) {
+				seq_printf(m, "pattern: %d\n",
+					   intel_dp->compliance.test_data.phytest.phy_pattern);
+				seq_printf(m, "Number of lanes: %d\n",
+					   intel_dp->compliance.test_data.phytest.num_lanes);
+				seq_printf(m, "Link Rate: %d\n",
+					   intel_dp->compliance.test_data.phytest.link_rate);
+				seq_printf(m, "level: %02x\n",
+					   intel_dp->train_set[0]);
+			}
+		} else
+			seq_puts(m, "0");
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(i915_displayport_test_data);
+
+static int i915_displayport_test_type_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = m->private;
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	struct intel_dp *intel_dp;
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		struct intel_encoder *encoder;
+
+		if (connector->connector_type !=
+		    DRM_MODE_CONNECTOR_DisplayPort)
+			continue;
+
+		encoder = to_intel_encoder(connector->encoder);
+		if (encoder && encoder->type == INTEL_OUTPUT_DP_MST)
+			continue;
+
+		if (encoder && connector->status == connector_status_connected) {
+			intel_dp = enc_to_intel_dp(encoder);
+			seq_printf(m, "%02lx\n", intel_dp->compliance.test_type);
+		} else
+			seq_puts(m, "0");
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(i915_displayport_test_type);
+
+static void wm_latency_show(struct seq_file *m, const u16 wm[8])
+{
+	struct drm_i915_private *dev_priv = m->private;
+	struct drm_device *dev = &dev_priv->drm;
+	int level;
+	int num_levels;
+
+	if (IS_CHERRYVIEW(dev_priv))
+		num_levels = 3;
+	else if (IS_VALLEYVIEW(dev_priv))
+		num_levels = 1;
+	else if (IS_G4X(dev_priv))
+		num_levels = 3;
+	else
+		num_levels = ilk_wm_max_level(dev_priv) + 1;
+
+	drm_modeset_lock_all(dev);
+
+	for (level = 0; level < num_levels; level++) {
+		unsigned int latency = wm[level];
+
+		/*
+		 * - WM1+ latency values in 0.5us units
+		 * - latencies are in us on gen9/vlv/chv
+		 */
+		if (DISPLAY_VER(dev_priv) >= 9 ||
+		    IS_VALLEYVIEW(dev_priv) ||
+		    IS_CHERRYVIEW(dev_priv) ||
+		    IS_G4X(dev_priv))
+			latency *= 10;
+		else if (level > 0)
+			latency *= 5;
+
+		seq_printf(m, "WM%d %u (%u.%u usec)\n",
+			   level, wm[level], latency / 10, latency % 10);
+	}
+
+	drm_modeset_unlock_all(dev);
+}
+
+static int pri_wm_latency_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = m->private;
+	const u16 *latencies;
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		latencies = dev_priv->display.wm.skl_latency;
+	else
+		latencies = dev_priv->display.wm.pri_latency;
+
+	wm_latency_show(m, latencies);
+
+	return 0;
+}
+
+static int spr_wm_latency_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = m->private;
+	const u16 *latencies;
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		latencies = dev_priv->display.wm.skl_latency;
+	else
+		latencies = dev_priv->display.wm.spr_latency;
+
+	wm_latency_show(m, latencies);
+
+	return 0;
+}
+
+static int cur_wm_latency_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = m->private;
+	const u16 *latencies;
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		latencies = dev_priv->display.wm.skl_latency;
+	else
+		latencies = dev_priv->display.wm.cur_latency;
+
+	wm_latency_show(m, latencies);
+
+	return 0;
+}
+
+static int pri_wm_latency_open(struct inode *inode, struct file *file)
+{
+	struct drm_i915_private *dev_priv = inode->i_private;
+
+	if (DISPLAY_VER(dev_priv) < 5 && !IS_G4X(dev_priv))
+		return -ENODEV;
+
+	return single_open(file, pri_wm_latency_show, dev_priv);
+}
+
+static int spr_wm_latency_open(struct inode *inode, struct file *file)
+{
+	struct drm_i915_private *dev_priv = inode->i_private;
+
+	if (HAS_GMCH(dev_priv))
+		return -ENODEV;
+
+	return single_open(file, spr_wm_latency_show, dev_priv);
+}
+
+static int cur_wm_latency_open(struct inode *inode, struct file *file)
+{
+	struct drm_i915_private *dev_priv = inode->i_private;
+
+	if (HAS_GMCH(dev_priv))
+		return -ENODEV;
+
+	return single_open(file, cur_wm_latency_show, dev_priv);
+}
+
+static ssize_t wm_latency_write(struct file *file, const char __user *ubuf,
+				size_t len, loff_t *offp, u16 wm[8])
+{
+	struct seq_file *m = file->private_data;
+	struct drm_i915_private *dev_priv = m->private;
+	struct drm_device *dev = &dev_priv->drm;
+	u16 new[8] = { 0 };
+	int num_levels;
+	int level;
+	int ret;
+	char tmp[32];
+
+	if (IS_CHERRYVIEW(dev_priv))
+		num_levels = 3;
+	else if (IS_VALLEYVIEW(dev_priv))
+		num_levels = 1;
+	else if (IS_G4X(dev_priv))
+		num_levels = 3;
+	else
+		num_levels = ilk_wm_max_level(dev_priv) + 1;
+
+	if (len >= sizeof(tmp))
+		return -EINVAL;
+
+	if (copy_from_user(tmp, ubuf, len))
+		return -EFAULT;
+
+	tmp[len] = '\0';
+
+	ret = sscanf(tmp, "%hu %hu %hu %hu %hu %hu %hu %hu",
+		     &new[0], &new[1], &new[2], &new[3],
+		     &new[4], &new[5], &new[6], &new[7]);
+	if (ret != num_levels)
+		return -EINVAL;
+
+	drm_modeset_lock_all(dev);
+
+	for (level = 0; level < num_levels; level++)
+		wm[level] = new[level];
+
+	drm_modeset_unlock_all(dev);
+
+	return len;
+}
+
+
+static ssize_t pri_wm_latency_write(struct file *file, const char __user *ubuf,
+				    size_t len, loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct drm_i915_private *dev_priv = m->private;
+	u16 *latencies;
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		latencies = dev_priv->display.wm.skl_latency;
+	else
+		latencies = dev_priv->display.wm.pri_latency;
+
+	return wm_latency_write(file, ubuf, len, offp, latencies);
+}
+
+static ssize_t spr_wm_latency_write(struct file *file, const char __user *ubuf,
+				    size_t len, loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct drm_i915_private *dev_priv = m->private;
+	u16 *latencies;
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		latencies = dev_priv->display.wm.skl_latency;
+	else
+		latencies = dev_priv->display.wm.spr_latency;
+
+	return wm_latency_write(file, ubuf, len, offp, latencies);
+}
+
+static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf,
+				    size_t len, loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct drm_i915_private *dev_priv = m->private;
+	u16 *latencies;
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		latencies = dev_priv->display.wm.skl_latency;
+	else
+		latencies = dev_priv->display.wm.cur_latency;
+
+	return wm_latency_write(file, ubuf, len, offp, latencies);
+}
+
+static const struct file_operations i915_pri_wm_latency_fops = {
+	.owner = THIS_MODULE,
+	.open = pri_wm_latency_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = pri_wm_latency_write
+};
+
+static const struct file_operations i915_spr_wm_latency_fops = {
+	.owner = THIS_MODULE,
+	.open = spr_wm_latency_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = spr_wm_latency_write
+};
+
+static const struct file_operations i915_cur_wm_latency_fops = {
+	.owner = THIS_MODULE,
+	.open = cur_wm_latency_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = cur_wm_latency_write
+};
+
+static int i915_hpd_storm_ctl_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = m->private;
+	struct intel_hotplug *hotplug = &dev_priv->display.hotplug;
+
+	/* Synchronize with everything first in case there's been an HPD
+	 * storm, but we haven't finished handling it in the kernel yet
+	 */
+	intel_synchronize_irq(dev_priv);
+	flush_work(&dev_priv->display.hotplug.dig_port_work);
+	flush_delayed_work(&dev_priv->display.hotplug.hotplug_work);
+
+	seq_printf(m, "Threshold: %d\n", hotplug->hpd_storm_threshold);
+	seq_printf(m, "Detected: %s\n",
+		   str_yes_no(delayed_work_pending(&hotplug->reenable_work)));
+
+	return 0;
+}
+
+static ssize_t i915_hpd_storm_ctl_write(struct file *file,
+					const char __user *ubuf, size_t len,
+					loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct drm_i915_private *dev_priv = m->private;
+	struct intel_hotplug *hotplug = &dev_priv->display.hotplug;
+	unsigned int new_threshold;
+	int i;
+	char *newline;
+	char tmp[16];
+
+	if (len >= sizeof(tmp))
+		return -EINVAL;
+
+	if (copy_from_user(tmp, ubuf, len))
+		return -EFAULT;
+
+	tmp[len] = '\0';
+
+	/* Strip newline, if any */
+	newline = strchr(tmp, '\n');
+	if (newline)
+		*newline = '\0';
+
+	if (strcmp(tmp, "reset") == 0)
+		new_threshold = HPD_STORM_DEFAULT_THRESHOLD;
+	else if (kstrtouint(tmp, 10, &new_threshold) != 0)
+		return -EINVAL;
+
+	if (new_threshold > 0)
+		drm_dbg_kms(&dev_priv->drm,
+			    "Setting HPD storm detection threshold to %d\n",
+			    new_threshold);
+	else
+		drm_dbg_kms(&dev_priv->drm, "Disabling HPD storm detection\n");
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	hotplug->hpd_storm_threshold = new_threshold;
+	/* Reset the HPD storm stats so we don't accidentally trigger a storm */
+	for_each_hpd_pin(i)
+		hotplug->stats[i].count = 0;
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/* Re-enable hpd immediately if we were in an irq storm */
+	flush_delayed_work(&dev_priv->display.hotplug.reenable_work);
+
+	return len;
+}
+
+static int i915_hpd_storm_ctl_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, i915_hpd_storm_ctl_show, inode->i_private);
+}
+
+static const struct file_operations i915_hpd_storm_ctl_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_hpd_storm_ctl_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = i915_hpd_storm_ctl_write
+};
+
+static int i915_hpd_short_storm_ctl_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = m->private;
+
+	seq_printf(m, "Enabled: %s\n",
+		   str_yes_no(dev_priv->display.hotplug.hpd_short_storm_enabled));
+
+	return 0;
+}
+
+static int
+i915_hpd_short_storm_ctl_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, i915_hpd_short_storm_ctl_show,
+			   inode->i_private);
+}
+
+static ssize_t i915_hpd_short_storm_ctl_write(struct file *file,
+					      const char __user *ubuf,
+					      size_t len, loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct drm_i915_private *dev_priv = m->private;
+	struct intel_hotplug *hotplug = &dev_priv->display.hotplug;
+	char *newline;
+	char tmp[16];
+	int i;
+	bool new_state;
+
+	if (len >= sizeof(tmp))
+		return -EINVAL;
+
+	if (copy_from_user(tmp, ubuf, len))
+		return -EFAULT;
+
+	tmp[len] = '\0';
+
+	/* Strip newline, if any */
+	newline = strchr(tmp, '\n');
+	if (newline)
+		*newline = '\0';
+
+	/* Reset to the "default" state for this system */
+	if (strcmp(tmp, "reset") == 0)
+		new_state = !HAS_DP_MST(dev_priv);
+	else if (kstrtobool(tmp, &new_state) != 0)
+		return -EINVAL;
+
+	drm_dbg_kms(&dev_priv->drm, "%sabling HPD short storm detection\n",
+		    new_state ? "En" : "Dis");
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	hotplug->hpd_short_storm_enabled = new_state;
+	/* Reset the HPD storm stats so we don't accidentally trigger a storm */
+	for_each_hpd_pin(i)
+		hotplug->stats[i].count = 0;
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/* Re-enable hpd immediately if we were in an irq storm */
+	flush_delayed_work(&dev_priv->display.hotplug.reenable_work);
+
+	return len;
+}
+
+static const struct file_operations i915_hpd_short_storm_ctl_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_hpd_short_storm_ctl_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = i915_hpd_short_storm_ctl_write,
+};
+
+static int i915_drrs_ctl_set(void *data, u64 val)
+{
+	struct drm_i915_private *dev_priv = data;
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(dev, crtc) {
+		struct intel_crtc_state *crtc_state;
+		struct drm_crtc_commit *commit;
+		int ret;
+
+		ret = drm_modeset_lock_single_interruptible(&crtc->base.mutex);
+		if (ret)
+			return ret;
+
+		crtc_state = to_intel_crtc_state(crtc->base.state);
+
+		if (!crtc_state->hw.active ||
+		    !crtc_state->has_drrs)
+			goto out;
+
+		commit = crtc_state->uapi.commit;
+		if (commit) {
+			ret = wait_for_completion_interruptible(&commit->hw_done);
+			if (ret)
+				goto out;
+		}
+
+		drm_dbg(&dev_priv->drm,
+			"Manually %sactivating DRRS\n", val ? "" : "de");
+
+		if (val)
+			intel_drrs_activate(crtc_state);
+		else
+			intel_drrs_deactivate(crtc_state);
+
+out:
+		drm_modeset_unlock(&crtc->base.mutex);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(i915_drrs_ctl_fops, NULL, i915_drrs_ctl_set, "%llu\n");
+
+static ssize_t
+i915_fifo_underrun_reset_write(struct file *filp,
+			       const char __user *ubuf,
+			       size_t cnt, loff_t *ppos)
+{
+	struct drm_i915_private *dev_priv = filp->private_data;
+	struct intel_crtc *crtc;
+	struct drm_device *dev = &dev_priv->drm;
+	int ret;
+	bool reset;
+
+	ret = kstrtobool_from_user(ubuf, cnt, &reset);
+	if (ret)
+		return ret;
+
+	if (!reset)
+		return cnt;
+
+	for_each_intel_crtc(dev, crtc) {
+		struct drm_crtc_commit *commit;
+		struct intel_crtc_state *crtc_state;
+
+		ret = drm_modeset_lock_single_interruptible(&crtc->base.mutex);
+		if (ret)
+			return ret;
+
+		crtc_state = to_intel_crtc_state(crtc->base.state);
+		commit = crtc_state->uapi.commit;
+		if (commit) {
+			ret = wait_for_completion_interruptible(&commit->hw_done);
+			if (!ret)
+				ret = wait_for_completion_interruptible(&commit->flip_done);
+		}
+
+		if (!ret && crtc_state->hw.active) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Re-arming FIFO underruns on pipe %c\n",
+				    pipe_name(crtc->pipe));
+
+			intel_crtc_arm_fifo_underrun(crtc, crtc_state);
+		}
+
+		drm_modeset_unlock(&crtc->base.mutex);
+
+		if (ret)
+			return ret;
+	}
+
+	intel_fbc_reset_underrun(dev_priv);
+
+	return cnt;
+}
+
+static const struct file_operations i915_fifo_underrun_reset_ops = {
+	.owner = THIS_MODULE,
+	.open = simple_open,
+	.write = i915_fifo_underrun_reset_write,
+	.llseek = default_llseek,
+};
+
+static const struct drm_info_list intel_display_debugfs_list[] = {
+	{"i915_frontbuffer_tracking", i915_frontbuffer_tracking, 0},
+	{"i915_ips_status", i915_ips_status, 0},
+	{"i915_sr_status", i915_sr_status, 0},
+	{"i915_opregion", i915_opregion, 0},
+	{"i915_vbt", i915_vbt, 0},
+	{"i915_gem_framebuffer", i915_gem_framebuffer_info, 0},
+	{"i915_edp_psr_status", i915_edp_psr_status, 0},
+	{"i915_power_domain_info", i915_power_domain_info, 0},
+	{"i915_display_info", i915_display_info, 0},
+	{"i915_shared_dplls_info", i915_shared_dplls_info, 0},
+	{"i915_dp_mst_info", i915_dp_mst_info, 0},
+	{"i915_ddb_info", i915_ddb_info, 0},
+	{"i915_drrs_status", i915_drrs_status, 0},
+	{"i915_lpsp_status", i915_lpsp_status, 0},
+};
+
+static const struct {
+	const char *name;
+	const struct file_operations *fops;
+} intel_display_debugfs_files[] = {
+	{"i915_fifo_underrun_reset", &i915_fifo_underrun_reset_ops},
+	{"i915_pri_wm_latency", &i915_pri_wm_latency_fops},
+	{"i915_spr_wm_latency", &i915_spr_wm_latency_fops},
+	{"i915_cur_wm_latency", &i915_cur_wm_latency_fops},
+	{"i915_dp_test_data", &i915_displayport_test_data_fops},
+	{"i915_dp_test_type", &i915_displayport_test_type_fops},
+	{"i915_dp_test_active", &i915_displayport_test_active_fops},
+	{"i915_hpd_storm_ctl", &i915_hpd_storm_ctl_fops},
+	{"i915_hpd_short_storm_ctl", &i915_hpd_short_storm_ctl_fops},
+	{"i915_drrs_ctl", &i915_drrs_ctl_fops},
+	{"i915_edp_psr_debug", &i915_edp_psr_debug_fops},
+};
+
+void intel_display_debugfs_register(struct drm_i915_private *i915)
+{
+	struct drm_minor *minor = i915->drm.primary;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(intel_display_debugfs_files); i++) {
+		debugfs_create_file(intel_display_debugfs_files[i].name,
+				    S_IRUGO | S_IWUSR,
+				    minor->debugfs_root,
+				    to_i915(minor->dev),
+				    intel_display_debugfs_files[i].fops);
+	}
+
+	drm_debugfs_create_files(intel_display_debugfs_list,
+				 ARRAY_SIZE(intel_display_debugfs_list),
+				 minor->debugfs_root, minor);
+
+	intel_dmc_debugfs_register(i915);
+	intel_fbc_debugfs_register(i915);
+	skl_watermark_ipc_debugfs_register(i915);
+}
+
+static int i915_panel_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	struct intel_dp *intel_dp =
+		intel_attached_dp(to_intel_connector(connector));
+
+	if (connector->status != connector_status_connected)
+		return -ENODEV;
+
+	seq_printf(m, "Panel power up delay: %d\n",
+		   intel_dp->pps.panel_power_up_delay);
+	seq_printf(m, "Panel power down delay: %d\n",
+		   intel_dp->pps.panel_power_down_delay);
+	seq_printf(m, "Backlight on delay: %d\n",
+		   intel_dp->pps.backlight_on_delay);
+	seq_printf(m, "Backlight off delay: %d\n",
+		   intel_dp->pps.backlight_off_delay);
+
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(i915_panel);
+
+static int i915_hdcp_sink_capability_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	int ret;
+
+	ret = drm_modeset_lock_single_interruptible(&i915->drm.mode_config.connection_mutex);
+	if (ret)
+		return ret;
+
+	if (!connector->encoder || connector->status != connector_status_connected) {
+		ret = -ENODEV;
+		goto out;
+	}
+
+	seq_printf(m, "%s:%d HDCP version: ", connector->name,
+		   connector->base.id);
+	intel_hdcp_info(m, intel_connector);
+
+out:
+	drm_modeset_unlock(&i915->drm.mode_config.connection_mutex);
+
+	return ret;
+}
+DEFINE_SHOW_ATTRIBUTE(i915_hdcp_sink_capability);
+
+static int i915_psr_status_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	struct intel_dp *intel_dp =
+		intel_attached_dp(to_intel_connector(connector));
+
+	return intel_psr_status(m, intel_dp);
+}
+DEFINE_SHOW_ATTRIBUTE(i915_psr_status);
+
+static int i915_lpsp_capability_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct intel_encoder *encoder;
+	bool lpsp_capable = false;
+
+	encoder = intel_attached_encoder(to_intel_connector(connector));
+	if (!encoder)
+		return -ENODEV;
+
+	if (connector->status != connector_status_connected)
+		return -ENODEV;
+
+	if (DISPLAY_VER(i915) >= 13)
+		lpsp_capable = encoder->port <= PORT_B;
+	else if (DISPLAY_VER(i915) >= 12)
+		/*
+		 * Actually TGL can drive LPSP on port till DDI_C
+		 * but there is no physical connected DDI_C on TGL sku's,
+		 * even driver is not initilizing DDI_C port for gen12.
+		 */
+		lpsp_capable = encoder->port <= PORT_B;
+	else if (DISPLAY_VER(i915) == 11)
+		lpsp_capable = (connector->connector_type == DRM_MODE_CONNECTOR_DSI ||
+				connector->connector_type == DRM_MODE_CONNECTOR_eDP);
+	else if (IS_DISPLAY_VER(i915, 9, 10))
+		lpsp_capable = (encoder->port == PORT_A &&
+				(connector->connector_type == DRM_MODE_CONNECTOR_DSI ||
+				 connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
+				 connector->connector_type == DRM_MODE_CONNECTOR_DisplayPort));
+	else if (IS_HASWELL(i915) || IS_BROADWELL(i915))
+		lpsp_capable = connector->connector_type == DRM_MODE_CONNECTOR_eDP;
+
+	seq_printf(m, "LPSP: %s\n", lpsp_capable ? "capable" : "incapable");
+
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(i915_lpsp_capability);
+
+static int i915_dsc_fec_support_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	struct drm_device *dev = connector->dev;
+	struct drm_crtc *crtc;
+	struct intel_dp *intel_dp;
+	struct drm_modeset_acquire_ctx ctx;
+	struct intel_crtc_state *crtc_state = NULL;
+	int ret = 0;
+	bool try_again = false;
+
+	drm_modeset_acquire_init(&ctx, DRM_MODESET_ACQUIRE_INTERRUPTIBLE);
+
+	do {
+		try_again = false;
+		ret = drm_modeset_lock(&dev->mode_config.connection_mutex,
+				       &ctx);
+		if (ret) {
+			if (ret == -EDEADLK && !drm_modeset_backoff(&ctx)) {
+				try_again = true;
+				continue;
+			}
+			break;
+		}
+		crtc = connector->state->crtc;
+		if (connector->status != connector_status_connected || !crtc) {
+			ret = -ENODEV;
+			break;
+		}
+		ret = drm_modeset_lock(&crtc->mutex, &ctx);
+		if (ret == -EDEADLK) {
+			ret = drm_modeset_backoff(&ctx);
+			if (!ret) {
+				try_again = true;
+				continue;
+			}
+			break;
+		} else if (ret) {
+			break;
+		}
+		intel_dp = intel_attached_dp(to_intel_connector(connector));
+		crtc_state = to_intel_crtc_state(crtc->state);
+		seq_printf(m, "DSC_Enabled: %s\n",
+			   str_yes_no(crtc_state->dsc.compression_enable));
+		seq_printf(m, "DSC_Sink_Support: %s\n",
+			   str_yes_no(drm_dp_sink_supports_dsc(intel_dp->dsc_dpcd)));
+		seq_printf(m, "Force_DSC_Enable: %s\n",
+			   str_yes_no(intel_dp->force_dsc_en));
+		if (!intel_dp_is_edp(intel_dp))
+			seq_printf(m, "FEC_Sink_Support: %s\n",
+				   str_yes_no(drm_dp_sink_supports_fec(intel_dp->fec_capable)));
+	} while (try_again);
+
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+
+	return ret;
+}
+
+static ssize_t i915_dsc_fec_support_write(struct file *file,
+					  const char __user *ubuf,
+					  size_t len, loff_t *offp)
+{
+	bool dsc_enable = false;
+	int ret;
+	struct drm_connector *connector =
+		((struct seq_file *)file->private_data)->private;
+	struct intel_encoder *encoder = intel_attached_encoder(to_intel_connector(connector));
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+	if (len == 0)
+		return 0;
+
+	drm_dbg(&i915->drm,
+		"Copied %zu bytes from user to force DSC\n", len);
+
+	ret = kstrtobool_from_user(ubuf, len, &dsc_enable);
+	if (ret < 0)
+		return ret;
+
+	drm_dbg(&i915->drm, "Got %s for DSC Enable\n",
+		(dsc_enable) ? "true" : "false");
+	intel_dp->force_dsc_en = dsc_enable;
+
+	*offp += len;
+	return len;
+}
+
+static int i915_dsc_fec_support_open(struct inode *inode,
+				     struct file *file)
+{
+	return single_open(file, i915_dsc_fec_support_show,
+			   inode->i_private);
+}
+
+static const struct file_operations i915_dsc_fec_support_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_dsc_fec_support_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = i915_dsc_fec_support_write
+};
+
+static int i915_dsc_bpc_show(struct seq_file *m, void *data)
+{
+	struct drm_connector *connector = m->private;
+	struct drm_device *dev = connector->dev;
+	struct drm_crtc *crtc;
+	struct intel_crtc_state *crtc_state;
+	struct intel_encoder *encoder = intel_attached_encoder(to_intel_connector(connector));
+	int ret;
+
+	if (!encoder)
+		return -ENODEV;
+
+	ret = drm_modeset_lock_single_interruptible(&dev->mode_config.connection_mutex);
+	if (ret)
+		return ret;
+
+	crtc = connector->state->crtc;
+	if (connector->status != connector_status_connected || !crtc) {
+		ret = -ENODEV;
+		goto out;
+	}
+
+	crtc_state = to_intel_crtc_state(crtc->state);
+	seq_printf(m, "Input_BPC: %d\n", crtc_state->dsc.config.bits_per_component);
+
+out:	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+
+	return ret;
+}
+
+static ssize_t i915_dsc_bpc_write(struct file *file,
+				  const char __user *ubuf,
+				  size_t len, loff_t *offp)
+{
+	struct drm_connector *connector =
+		((struct seq_file *)file->private_data)->private;
+	struct intel_encoder *encoder = intel_attached_encoder(to_intel_connector(connector));
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	int dsc_bpc = 0;
+	int ret;
+
+	ret = kstrtoint_from_user(ubuf, len, 0, &dsc_bpc);
+	if (ret < 0)
+		return ret;
+
+	intel_dp->force_dsc_bpc = dsc_bpc;
+	*offp += len;
+
+	return len;
+}
+
+static int i915_dsc_bpc_open(struct inode *inode,
+			     struct file *file)
+{
+	return single_open(file, i915_dsc_bpc_show, inode->i_private);
+}
+
+static const struct file_operations i915_dsc_bpc_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_dsc_bpc_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = i915_dsc_bpc_write
+};
+
+/*
+ * Returns the Current CRTC's bpc.
+ * Example usage: cat /sys/kernel/debug/dri/0/crtc-0/i915_current_bpc
+ */
+static int i915_current_bpc_show(struct seq_file *m, void *data)
+{
+	struct intel_crtc *crtc = to_intel_crtc(m->private);
+	struct intel_crtc_state *crtc_state;
+	int ret;
+
+	ret = drm_modeset_lock_single_interruptible(&crtc->base.mutex);
+	if (ret)
+		return ret;
+
+	crtc_state = to_intel_crtc_state(crtc->base.state);
+	seq_printf(m, "Current: %u\n", crtc_state->pipe_bpp / 3);
+
+	drm_modeset_unlock(&crtc->base.mutex);
+
+	return ret;
+}
+DEFINE_SHOW_ATTRIBUTE(i915_current_bpc);
+
+/**
+ * intel_connector_debugfs_add - add i915 specific connector debugfs files
+ * @connector: pointer to a registered drm_connector
+ *
+ * Cleanup will be done by drm_connector_unregister() through a call to
+ * drm_debugfs_connector_remove().
+ */
+void intel_connector_debugfs_add(struct intel_connector *intel_connector)
+{
+	struct drm_connector *connector = &intel_connector->base;
+	struct dentry *root = connector->debugfs_entry;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+
+	/* The connector must have been registered beforehands. */
+	if (!root)
+		return;
+
+	if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) {
+		debugfs_create_file("i915_panel_timings", S_IRUGO, root,
+				    connector, &i915_panel_fops);
+		debugfs_create_file("i915_psr_sink_status", S_IRUGO, root,
+				    connector, &i915_psr_sink_status_fops);
+	}
+
+	if (HAS_PSR(dev_priv) &&
+	    connector->connector_type == DRM_MODE_CONNECTOR_eDP) {
+		debugfs_create_file("i915_psr_status", 0444, root,
+				    connector, &i915_psr_status_fops);
+	}
+
+	if (connector->connector_type == DRM_MODE_CONNECTOR_DisplayPort ||
+	    connector->connector_type == DRM_MODE_CONNECTOR_HDMIA ||
+	    connector->connector_type == DRM_MODE_CONNECTOR_HDMIB) {
+		debugfs_create_file("i915_hdcp_sink_capability", S_IRUGO, root,
+				    connector, &i915_hdcp_sink_capability_fops);
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 11 &&
+	    ((connector->connector_type == DRM_MODE_CONNECTOR_DisplayPort &&
+	    !to_intel_connector(connector)->mst_port) ||
+	    connector->connector_type == DRM_MODE_CONNECTOR_eDP)) {
+		debugfs_create_file("i915_dsc_fec_support", 0644, root,
+				    connector, &i915_dsc_fec_support_fops);
+
+		debugfs_create_file("i915_dsc_bpc", 0644, root,
+				    connector, &i915_dsc_bpc_fops);
+	}
+
+	if (connector->connector_type == DRM_MODE_CONNECTOR_DSI ||
+	    connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
+	    connector->connector_type == DRM_MODE_CONNECTOR_DisplayPort ||
+	    connector->connector_type == DRM_MODE_CONNECTOR_HDMIA ||
+	    connector->connector_type == DRM_MODE_CONNECTOR_HDMIB)
+		debugfs_create_file("i915_lpsp_capability", 0444, root,
+				    connector, &i915_lpsp_capability_fops);
+}
+
+/**
+ * intel_crtc_debugfs_add - add i915 specific crtc debugfs files
+ * @crtc: pointer to a drm_crtc
+ *
+ * Failure to add debugfs entries should generally be ignored.
+ */
+void intel_crtc_debugfs_add(struct drm_crtc *crtc)
+{
+	if (!crtc->debugfs_entry)
+		return;
+
+	crtc_updates_add(crtc);
+	intel_fbc_crtc_debugfs_add(to_intel_crtc(crtc));
+
+	debugfs_create_file("i915_current_bpc", 0444, crtc->debugfs_entry, crtc,
+			    &i915_current_bpc_fops);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_display_debugfs.h b/drivers/gpu/drm/i915/display/intel_display_debugfs.h
new file mode 100644
index 000000000..d3a79c07c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_debugfs.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __INTEL_DISPLAY_DEBUGFS_H__
+#define __INTEL_DISPLAY_DEBUGFS_H__
+
+struct drm_crtc;
+struct drm_i915_private;
+struct intel_connector;
+
+#ifdef CONFIG_DEBUG_FS
+void intel_display_debugfs_register(struct drm_i915_private *i915);
+void intel_connector_debugfs_add(struct intel_connector *connector);
+void intel_crtc_debugfs_add(struct drm_crtc *crtc);
+#else
+static inline void intel_display_debugfs_register(struct drm_i915_private *i915) {}
+static inline void intel_connector_debugfs_add(struct intel_connector *connector) {}
+static inline void intel_crtc_debugfs_add(struct drm_crtc *crtc) {}
+#endif
+
+#endif /* __INTEL_DISPLAY_DEBUGFS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c
new file mode 100644
index 000000000..1a63da28f
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_power.c
@@ -0,0 +1,2489 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+
+#include "i915_drv.h"
+#include "i915_irq.h"
+#include "intel_backlight_regs.h"
+#include "intel_cdclk.h"
+#include "intel_combo_phy.h"
+#include "intel_de.h"
+#include "intel_display_power.h"
+#include "intel_display_power_map.h"
+#include "intel_display_power_well.h"
+#include "intel_display_types.h"
+#include "intel_dmc.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pch_refclk.h"
+#include "intel_pcode.h"
+#include "intel_snps_phy.h"
+#include "skl_watermark.h"
+#include "vlv_sideband.h"
+
+#define for_each_power_domain_well(__dev_priv, __power_well, __domain)	\
+	for_each_power_well(__dev_priv, __power_well)				\
+		for_each_if(test_bit((__domain), (__power_well)->domains.bits))
+
+#define for_each_power_domain_well_reverse(__dev_priv, __power_well, __domain) \
+	for_each_power_well_reverse(__dev_priv, __power_well)		        \
+		for_each_if(test_bit((__domain), (__power_well)->domains.bits))
+
+const char *
+intel_display_power_domain_str(enum intel_display_power_domain domain)
+{
+	switch (domain) {
+	case POWER_DOMAIN_DISPLAY_CORE:
+		return "DISPLAY_CORE";
+	case POWER_DOMAIN_PIPE_A:
+		return "PIPE_A";
+	case POWER_DOMAIN_PIPE_B:
+		return "PIPE_B";
+	case POWER_DOMAIN_PIPE_C:
+		return "PIPE_C";
+	case POWER_DOMAIN_PIPE_D:
+		return "PIPE_D";
+	case POWER_DOMAIN_PIPE_PANEL_FITTER_A:
+		return "PIPE_PANEL_FITTER_A";
+	case POWER_DOMAIN_PIPE_PANEL_FITTER_B:
+		return "PIPE_PANEL_FITTER_B";
+	case POWER_DOMAIN_PIPE_PANEL_FITTER_C:
+		return "PIPE_PANEL_FITTER_C";
+	case POWER_DOMAIN_PIPE_PANEL_FITTER_D:
+		return "PIPE_PANEL_FITTER_D";
+	case POWER_DOMAIN_TRANSCODER_A:
+		return "TRANSCODER_A";
+	case POWER_DOMAIN_TRANSCODER_B:
+		return "TRANSCODER_B";
+	case POWER_DOMAIN_TRANSCODER_C:
+		return "TRANSCODER_C";
+	case POWER_DOMAIN_TRANSCODER_D:
+		return "TRANSCODER_D";
+	case POWER_DOMAIN_TRANSCODER_EDP:
+		return "TRANSCODER_EDP";
+	case POWER_DOMAIN_TRANSCODER_DSI_A:
+		return "TRANSCODER_DSI_A";
+	case POWER_DOMAIN_TRANSCODER_DSI_C:
+		return "TRANSCODER_DSI_C";
+	case POWER_DOMAIN_TRANSCODER_VDSC_PW2:
+		return "TRANSCODER_VDSC_PW2";
+	case POWER_DOMAIN_PORT_DDI_LANES_A:
+		return "PORT_DDI_LANES_A";
+	case POWER_DOMAIN_PORT_DDI_LANES_B:
+		return "PORT_DDI_LANES_B";
+	case POWER_DOMAIN_PORT_DDI_LANES_C:
+		return "PORT_DDI_LANES_C";
+	case POWER_DOMAIN_PORT_DDI_LANES_D:
+		return "PORT_DDI_LANES_D";
+	case POWER_DOMAIN_PORT_DDI_LANES_E:
+		return "PORT_DDI_LANES_E";
+	case POWER_DOMAIN_PORT_DDI_LANES_F:
+		return "PORT_DDI_LANES_F";
+	case POWER_DOMAIN_PORT_DDI_LANES_TC1:
+		return "PORT_DDI_LANES_TC1";
+	case POWER_DOMAIN_PORT_DDI_LANES_TC2:
+		return "PORT_DDI_LANES_TC2";
+	case POWER_DOMAIN_PORT_DDI_LANES_TC3:
+		return "PORT_DDI_LANES_TC3";
+	case POWER_DOMAIN_PORT_DDI_LANES_TC4:
+		return "PORT_DDI_LANES_TC4";
+	case POWER_DOMAIN_PORT_DDI_LANES_TC5:
+		return "PORT_DDI_LANES_TC5";
+	case POWER_DOMAIN_PORT_DDI_LANES_TC6:
+		return "PORT_DDI_LANES_TC6";
+	case POWER_DOMAIN_PORT_DDI_IO_A:
+		return "PORT_DDI_IO_A";
+	case POWER_DOMAIN_PORT_DDI_IO_B:
+		return "PORT_DDI_IO_B";
+	case POWER_DOMAIN_PORT_DDI_IO_C:
+		return "PORT_DDI_IO_C";
+	case POWER_DOMAIN_PORT_DDI_IO_D:
+		return "PORT_DDI_IO_D";
+	case POWER_DOMAIN_PORT_DDI_IO_E:
+		return "PORT_DDI_IO_E";
+	case POWER_DOMAIN_PORT_DDI_IO_F:
+		return "PORT_DDI_IO_F";
+	case POWER_DOMAIN_PORT_DDI_IO_TC1:
+		return "PORT_DDI_IO_TC1";
+	case POWER_DOMAIN_PORT_DDI_IO_TC2:
+		return "PORT_DDI_IO_TC2";
+	case POWER_DOMAIN_PORT_DDI_IO_TC3:
+		return "PORT_DDI_IO_TC3";
+	case POWER_DOMAIN_PORT_DDI_IO_TC4:
+		return "PORT_DDI_IO_TC4";
+	case POWER_DOMAIN_PORT_DDI_IO_TC5:
+		return "PORT_DDI_IO_TC5";
+	case POWER_DOMAIN_PORT_DDI_IO_TC6:
+		return "PORT_DDI_IO_TC6";
+	case POWER_DOMAIN_PORT_DSI:
+		return "PORT_DSI";
+	case POWER_DOMAIN_PORT_CRT:
+		return "PORT_CRT";
+	case POWER_DOMAIN_PORT_OTHER:
+		return "PORT_OTHER";
+	case POWER_DOMAIN_VGA:
+		return "VGA";
+	case POWER_DOMAIN_AUDIO_MMIO:
+		return "AUDIO_MMIO";
+	case POWER_DOMAIN_AUDIO_PLAYBACK:
+		return "AUDIO_PLAYBACK";
+	case POWER_DOMAIN_AUX_A:
+		return "AUX_A";
+	case POWER_DOMAIN_AUX_B:
+		return "AUX_B";
+	case POWER_DOMAIN_AUX_C:
+		return "AUX_C";
+	case POWER_DOMAIN_AUX_D:
+		return "AUX_D";
+	case POWER_DOMAIN_AUX_E:
+		return "AUX_E";
+	case POWER_DOMAIN_AUX_F:
+		return "AUX_F";
+	case POWER_DOMAIN_AUX_USBC1:
+		return "AUX_USBC1";
+	case POWER_DOMAIN_AUX_USBC2:
+		return "AUX_USBC2";
+	case POWER_DOMAIN_AUX_USBC3:
+		return "AUX_USBC3";
+	case POWER_DOMAIN_AUX_USBC4:
+		return "AUX_USBC4";
+	case POWER_DOMAIN_AUX_USBC5:
+		return "AUX_USBC5";
+	case POWER_DOMAIN_AUX_USBC6:
+		return "AUX_USBC6";
+	case POWER_DOMAIN_AUX_IO_A:
+		return "AUX_IO_A";
+	case POWER_DOMAIN_AUX_TBT1:
+		return "AUX_TBT1";
+	case POWER_DOMAIN_AUX_TBT2:
+		return "AUX_TBT2";
+	case POWER_DOMAIN_AUX_TBT3:
+		return "AUX_TBT3";
+	case POWER_DOMAIN_AUX_TBT4:
+		return "AUX_TBT4";
+	case POWER_DOMAIN_AUX_TBT5:
+		return "AUX_TBT5";
+	case POWER_DOMAIN_AUX_TBT6:
+		return "AUX_TBT6";
+	case POWER_DOMAIN_GMBUS:
+		return "GMBUS";
+	case POWER_DOMAIN_INIT:
+		return "INIT";
+	case POWER_DOMAIN_MODESET:
+		return "MODESET";
+	case POWER_DOMAIN_GT_IRQ:
+		return "GT_IRQ";
+	case POWER_DOMAIN_DC_OFF:
+		return "DC_OFF";
+	case POWER_DOMAIN_TC_COLD_OFF:
+		return "TC_COLD_OFF";
+	default:
+		MISSING_CASE(domain);
+		return "?";
+	}
+}
+
+/**
+ * __intel_display_power_is_enabled - unlocked check for a power domain
+ * @dev_priv: i915 device instance
+ * @domain: power domain to check
+ *
+ * This is the unlocked version of intel_display_power_is_enabled() and should
+ * only be used from error capture and recovery code where deadlocks are
+ * possible.
+ *
+ * Returns:
+ * True when the power domain is enabled, false otherwise.
+ */
+bool __intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
+				      enum intel_display_power_domain domain)
+{
+	struct i915_power_well *power_well;
+	bool is_enabled;
+
+	if (dev_priv->runtime_pm.suspended)
+		return false;
+
+	is_enabled = true;
+
+	for_each_power_domain_well_reverse(dev_priv, power_well, domain) {
+		if (intel_power_well_is_always_on(power_well))
+			continue;
+
+		if (!intel_power_well_is_enabled_cached(power_well)) {
+			is_enabled = false;
+			break;
+		}
+	}
+
+	return is_enabled;
+}
+
+/**
+ * intel_display_power_is_enabled - check for a power domain
+ * @dev_priv: i915 device instance
+ * @domain: power domain to check
+ *
+ * This function can be used to check the hw power domain state. It is mostly
+ * used in hardware state readout functions. Everywhere else code should rely
+ * upon explicit power domain reference counting to ensure that the hardware
+ * block is powered up before accessing it.
+ *
+ * Callers must hold the relevant modesetting locks to ensure that concurrent
+ * threads can't disable the power well while the caller tries to read a few
+ * registers.
+ *
+ * Returns:
+ * True when the power domain is enabled, false otherwise.
+ */
+bool intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
+				    enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains;
+	bool ret;
+
+	power_domains = &dev_priv->display.power.domains;
+
+	mutex_lock(&power_domains->lock);
+	ret = __intel_display_power_is_enabled(dev_priv, domain);
+	mutex_unlock(&power_domains->lock);
+
+	return ret;
+}
+
+static u32
+sanitize_target_dc_state(struct drm_i915_private *dev_priv,
+			 u32 target_dc_state)
+{
+	static const u32 states[] = {
+		DC_STATE_EN_UPTO_DC6,
+		DC_STATE_EN_UPTO_DC5,
+		DC_STATE_EN_DC3CO,
+		DC_STATE_DISABLE,
+	};
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(states) - 1; i++) {
+		if (target_dc_state != states[i])
+			continue;
+
+		if (dev_priv->display.dmc.allowed_dc_mask & target_dc_state)
+			break;
+
+		target_dc_state = states[i + 1];
+	}
+
+	return target_dc_state;
+}
+
+/**
+ * intel_display_power_set_target_dc_state - Set target dc state.
+ * @dev_priv: i915 device
+ * @state: state which needs to be set as target_dc_state.
+ *
+ * This function set the "DC off" power well target_dc_state,
+ * based upon this target_dc_stste, "DC off" power well will
+ * enable desired DC state.
+ */
+void intel_display_power_set_target_dc_state(struct drm_i915_private *dev_priv,
+					     u32 state)
+{
+	struct i915_power_well *power_well;
+	bool dc_off_enabled;
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+
+	mutex_lock(&power_domains->lock);
+	power_well = lookup_power_well(dev_priv, SKL_DISP_DC_OFF);
+
+	if (drm_WARN_ON(&dev_priv->drm, !power_well))
+		goto unlock;
+
+	state = sanitize_target_dc_state(dev_priv, state);
+
+	if (state == dev_priv->display.dmc.target_dc_state)
+		goto unlock;
+
+	dc_off_enabled = intel_power_well_is_enabled(dev_priv, power_well);
+	/*
+	 * If DC off power well is disabled, need to enable and disable the
+	 * DC off power well to effect target DC state.
+	 */
+	if (!dc_off_enabled)
+		intel_power_well_enable(dev_priv, power_well);
+
+	dev_priv->display.dmc.target_dc_state = state;
+
+	if (!dc_off_enabled)
+		intel_power_well_disable(dev_priv, power_well);
+
+unlock:
+	mutex_unlock(&power_domains->lock);
+}
+
+#define POWER_DOMAIN_MASK (GENMASK_ULL(POWER_DOMAIN_NUM - 1, 0))
+
+static void __async_put_domains_mask(struct i915_power_domains *power_domains,
+				     struct intel_power_domain_mask *mask)
+{
+	bitmap_or(mask->bits,
+		  power_domains->async_put_domains[0].bits,
+		  power_domains->async_put_domains[1].bits,
+		  POWER_DOMAIN_NUM);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+
+static bool
+assert_async_put_domain_masks_disjoint(struct i915_power_domains *power_domains)
+{
+	struct drm_i915_private *i915 = container_of(power_domains,
+						     struct drm_i915_private,
+						     display.power.domains);
+
+	return !drm_WARN_ON(&i915->drm,
+			    bitmap_intersects(power_domains->async_put_domains[0].bits,
+					      power_domains->async_put_domains[1].bits,
+					      POWER_DOMAIN_NUM));
+}
+
+static bool
+__async_put_domains_state_ok(struct i915_power_domains *power_domains)
+{
+	struct drm_i915_private *i915 = container_of(power_domains,
+						     struct drm_i915_private,
+						     display.power.domains);
+	struct intel_power_domain_mask async_put_mask;
+	enum intel_display_power_domain domain;
+	bool err = false;
+
+	err |= !assert_async_put_domain_masks_disjoint(power_domains);
+	__async_put_domains_mask(power_domains, &async_put_mask);
+	err |= drm_WARN_ON(&i915->drm,
+			   !!power_domains->async_put_wakeref !=
+			   !bitmap_empty(async_put_mask.bits, POWER_DOMAIN_NUM));
+
+	for_each_power_domain(domain, &async_put_mask)
+		err |= drm_WARN_ON(&i915->drm,
+				   power_domains->domain_use_count[domain] != 1);
+
+	return !err;
+}
+
+static void print_power_domains(struct i915_power_domains *power_domains,
+				const char *prefix, struct intel_power_domain_mask *mask)
+{
+	struct drm_i915_private *i915 = container_of(power_domains,
+						     struct drm_i915_private,
+						     display.power.domains);
+	enum intel_display_power_domain domain;
+
+	drm_dbg(&i915->drm, "%s (%d):\n", prefix, bitmap_weight(mask->bits, POWER_DOMAIN_NUM));
+	for_each_power_domain(domain, mask)
+		drm_dbg(&i915->drm, "%s use_count %d\n",
+			intel_display_power_domain_str(domain),
+			power_domains->domain_use_count[domain]);
+}
+
+static void
+print_async_put_domains_state(struct i915_power_domains *power_domains)
+{
+	struct drm_i915_private *i915 = container_of(power_domains,
+						     struct drm_i915_private,
+						     display.power.domains);
+
+	drm_dbg(&i915->drm, "async_put_wakeref %u\n",
+		power_domains->async_put_wakeref);
+
+	print_power_domains(power_domains, "async_put_domains[0]",
+			    &power_domains->async_put_domains[0]);
+	print_power_domains(power_domains, "async_put_domains[1]",
+			    &power_domains->async_put_domains[1]);
+}
+
+static void
+verify_async_put_domains_state(struct i915_power_domains *power_domains)
+{
+	if (!__async_put_domains_state_ok(power_domains))
+		print_async_put_domains_state(power_domains);
+}
+
+#else
+
+static void
+assert_async_put_domain_masks_disjoint(struct i915_power_domains *power_domains)
+{
+}
+
+static void
+verify_async_put_domains_state(struct i915_power_domains *power_domains)
+{
+}
+
+#endif /* CONFIG_DRM_I915_DEBUG_RUNTIME_PM */
+
+static void async_put_domains_mask(struct i915_power_domains *power_domains,
+				   struct intel_power_domain_mask *mask)
+
+{
+	assert_async_put_domain_masks_disjoint(power_domains);
+
+	__async_put_domains_mask(power_domains, mask);
+}
+
+static void
+async_put_domains_clear_domain(struct i915_power_domains *power_domains,
+			       enum intel_display_power_domain domain)
+{
+	assert_async_put_domain_masks_disjoint(power_domains);
+
+	clear_bit(domain, power_domains->async_put_domains[0].bits);
+	clear_bit(domain, power_domains->async_put_domains[1].bits);
+}
+
+static bool
+intel_display_power_grab_async_put_ref(struct drm_i915_private *dev_priv,
+				       enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	struct intel_power_domain_mask async_put_mask;
+	bool ret = false;
+
+	async_put_domains_mask(power_domains, &async_put_mask);
+	if (!test_bit(domain, async_put_mask.bits))
+		goto out_verify;
+
+	async_put_domains_clear_domain(power_domains, domain);
+
+	ret = true;
+
+	async_put_domains_mask(power_domains, &async_put_mask);
+	if (!bitmap_empty(async_put_mask.bits, POWER_DOMAIN_NUM))
+		goto out_verify;
+
+	cancel_delayed_work(&power_domains->async_put_work);
+	intel_runtime_pm_put_raw(&dev_priv->runtime_pm,
+				 fetch_and_zero(&power_domains->async_put_wakeref));
+out_verify:
+	verify_async_put_domains_state(power_domains);
+
+	return ret;
+}
+
+static void
+__intel_display_power_get_domain(struct drm_i915_private *dev_priv,
+				 enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	struct i915_power_well *power_well;
+
+	if (intel_display_power_grab_async_put_ref(dev_priv, domain))
+		return;
+
+	for_each_power_domain_well(dev_priv, power_well, domain)
+		intel_power_well_get(dev_priv, power_well);
+
+	power_domains->domain_use_count[domain]++;
+}
+
+/**
+ * intel_display_power_get - grab a power domain reference
+ * @dev_priv: i915 device instance
+ * @domain: power domain to reference
+ *
+ * This function grabs a power domain reference for @domain and ensures that the
+ * power domain and all its parents are powered up. Therefore users should only
+ * grab a reference to the innermost power domain they need.
+ *
+ * Any power domain reference obtained by this function must have a symmetric
+ * call to intel_display_power_put() to release the reference again.
+ */
+intel_wakeref_t intel_display_power_get(struct drm_i915_private *dev_priv,
+					enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	intel_wakeref_t wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+
+	mutex_lock(&power_domains->lock);
+	__intel_display_power_get_domain(dev_priv, domain);
+	mutex_unlock(&power_domains->lock);
+
+	return wakeref;
+}
+
+/**
+ * intel_display_power_get_if_enabled - grab a reference for an enabled display power domain
+ * @dev_priv: i915 device instance
+ * @domain: power domain to reference
+ *
+ * This function grabs a power domain reference for @domain and ensures that the
+ * power domain and all its parents are powered up. Therefore users should only
+ * grab a reference to the innermost power domain they need.
+ *
+ * Any power domain reference obtained by this function must have a symmetric
+ * call to intel_display_power_put() to release the reference again.
+ */
+intel_wakeref_t
+intel_display_power_get_if_enabled(struct drm_i915_private *dev_priv,
+				   enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	intel_wakeref_t wakeref;
+	bool is_enabled;
+
+	wakeref = intel_runtime_pm_get_if_in_use(&dev_priv->runtime_pm);
+	if (!wakeref)
+		return false;
+
+	mutex_lock(&power_domains->lock);
+
+	if (__intel_display_power_is_enabled(dev_priv, domain)) {
+		__intel_display_power_get_domain(dev_priv, domain);
+		is_enabled = true;
+	} else {
+		is_enabled = false;
+	}
+
+	mutex_unlock(&power_domains->lock);
+
+	if (!is_enabled) {
+		intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+		wakeref = 0;
+	}
+
+	return wakeref;
+}
+
+static void
+__intel_display_power_put_domain(struct drm_i915_private *dev_priv,
+				 enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains;
+	struct i915_power_well *power_well;
+	const char *name = intel_display_power_domain_str(domain);
+	struct intel_power_domain_mask async_put_mask;
+
+	power_domains = &dev_priv->display.power.domains;
+
+	drm_WARN(&dev_priv->drm, !power_domains->domain_use_count[domain],
+		 "Use count on domain %s is already zero\n",
+		 name);
+	async_put_domains_mask(power_domains, &async_put_mask);
+	drm_WARN(&dev_priv->drm,
+		 test_bit(domain, async_put_mask.bits),
+		 "Async disabling of domain %s is pending\n",
+		 name);
+
+	power_domains->domain_use_count[domain]--;
+
+	for_each_power_domain_well_reverse(dev_priv, power_well, domain)
+		intel_power_well_put(dev_priv, power_well);
+}
+
+static void __intel_display_power_put(struct drm_i915_private *dev_priv,
+				      enum intel_display_power_domain domain)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+
+	mutex_lock(&power_domains->lock);
+	__intel_display_power_put_domain(dev_priv, domain);
+	mutex_unlock(&power_domains->lock);
+}
+
+static void
+queue_async_put_domains_work(struct i915_power_domains *power_domains,
+			     intel_wakeref_t wakeref)
+{
+	struct drm_i915_private *i915 = container_of(power_domains,
+						     struct drm_i915_private,
+						     display.power.domains);
+	drm_WARN_ON(&i915->drm, power_domains->async_put_wakeref);
+	power_domains->async_put_wakeref = wakeref;
+	drm_WARN_ON(&i915->drm, !queue_delayed_work(system_unbound_wq,
+						    &power_domains->async_put_work,
+						    msecs_to_jiffies(100)));
+}
+
+static void
+release_async_put_domains(struct i915_power_domains *power_domains,
+			  struct intel_power_domain_mask *mask)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(power_domains, struct drm_i915_private,
+			     display.power.domains);
+	struct intel_runtime_pm *rpm = &dev_priv->runtime_pm;
+	enum intel_display_power_domain domain;
+	intel_wakeref_t wakeref;
+
+	/*
+	 * The caller must hold already raw wakeref, upgrade that to a proper
+	 * wakeref to make the state checker happy about the HW access during
+	 * power well disabling.
+	 */
+	assert_rpm_raw_wakeref_held(rpm);
+	wakeref = intel_runtime_pm_get(rpm);
+
+	for_each_power_domain(domain, mask) {
+		/* Clear before put, so put's sanity check is happy. */
+		async_put_domains_clear_domain(power_domains, domain);
+		__intel_display_power_put_domain(dev_priv, domain);
+	}
+
+	intel_runtime_pm_put(rpm, wakeref);
+}
+
+static void
+intel_display_power_put_async_work(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, struct drm_i915_private,
+			     display.power.domains.async_put_work.work);
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	struct intel_runtime_pm *rpm = &dev_priv->runtime_pm;
+	intel_wakeref_t new_work_wakeref = intel_runtime_pm_get_raw(rpm);
+	intel_wakeref_t old_work_wakeref = 0;
+
+	mutex_lock(&power_domains->lock);
+
+	/*
+	 * Bail out if all the domain refs pending to be released were grabbed
+	 * by subsequent gets or a flush_work.
+	 */
+	old_work_wakeref = fetch_and_zero(&power_domains->async_put_wakeref);
+	if (!old_work_wakeref)
+		goto out_verify;
+
+	release_async_put_domains(power_domains,
+				  &power_domains->async_put_domains[0]);
+
+	/* Requeue the work if more domains were async put meanwhile. */
+	if (!bitmap_empty(power_domains->async_put_domains[1].bits, POWER_DOMAIN_NUM)) {
+		bitmap_copy(power_domains->async_put_domains[0].bits,
+			    power_domains->async_put_domains[1].bits,
+			    POWER_DOMAIN_NUM);
+		bitmap_zero(power_domains->async_put_domains[1].bits,
+			    POWER_DOMAIN_NUM);
+		queue_async_put_domains_work(power_domains,
+					     fetch_and_zero(&new_work_wakeref));
+	} else {
+		/*
+		 * Cancel the work that got queued after this one got dequeued,
+		 * since here we released the corresponding async-put reference.
+		 */
+		cancel_delayed_work(&power_domains->async_put_work);
+	}
+
+out_verify:
+	verify_async_put_domains_state(power_domains);
+
+	mutex_unlock(&power_domains->lock);
+
+	if (old_work_wakeref)
+		intel_runtime_pm_put_raw(rpm, old_work_wakeref);
+	if (new_work_wakeref)
+		intel_runtime_pm_put_raw(rpm, new_work_wakeref);
+}
+
+/**
+ * intel_display_power_put_async - release a power domain reference asynchronously
+ * @i915: i915 device instance
+ * @domain: power domain to reference
+ * @wakeref: wakeref acquired for the reference that is being released
+ *
+ * This function drops the power domain reference obtained by
+ * intel_display_power_get*() and schedules a work to power down the
+ * corresponding hardware block if this is the last reference.
+ */
+void __intel_display_power_put_async(struct drm_i915_private *i915,
+				     enum intel_display_power_domain domain,
+				     intel_wakeref_t wakeref)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+	struct intel_runtime_pm *rpm = &i915->runtime_pm;
+	intel_wakeref_t work_wakeref = intel_runtime_pm_get_raw(rpm);
+
+	mutex_lock(&power_domains->lock);
+
+	if (power_domains->domain_use_count[domain] > 1) {
+		__intel_display_power_put_domain(i915, domain);
+
+		goto out_verify;
+	}
+
+	drm_WARN_ON(&i915->drm, power_domains->domain_use_count[domain] != 1);
+
+	/* Let a pending work requeue itself or queue a new one. */
+	if (power_domains->async_put_wakeref) {
+		set_bit(domain, power_domains->async_put_domains[1].bits);
+	} else {
+		set_bit(domain, power_domains->async_put_domains[0].bits);
+		queue_async_put_domains_work(power_domains,
+					     fetch_and_zero(&work_wakeref));
+	}
+
+out_verify:
+	verify_async_put_domains_state(power_domains);
+
+	mutex_unlock(&power_domains->lock);
+
+	if (work_wakeref)
+		intel_runtime_pm_put_raw(rpm, work_wakeref);
+
+	intel_runtime_pm_put(rpm, wakeref);
+}
+
+/**
+ * intel_display_power_flush_work - flushes the async display power disabling work
+ * @i915: i915 device instance
+ *
+ * Flushes any pending work that was scheduled by a preceding
+ * intel_display_power_put_async() call, completing the disabling of the
+ * corresponding power domains.
+ *
+ * Note that the work handler function may still be running after this
+ * function returns; to ensure that the work handler isn't running use
+ * intel_display_power_flush_work_sync() instead.
+ */
+void intel_display_power_flush_work(struct drm_i915_private *i915)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+	struct intel_power_domain_mask async_put_mask;
+	intel_wakeref_t work_wakeref;
+
+	mutex_lock(&power_domains->lock);
+
+	work_wakeref = fetch_and_zero(&power_domains->async_put_wakeref);
+	if (!work_wakeref)
+		goto out_verify;
+
+	async_put_domains_mask(power_domains, &async_put_mask);
+	release_async_put_domains(power_domains, &async_put_mask);
+	cancel_delayed_work(&power_domains->async_put_work);
+
+out_verify:
+	verify_async_put_domains_state(power_domains);
+
+	mutex_unlock(&power_domains->lock);
+
+	if (work_wakeref)
+		intel_runtime_pm_put_raw(&i915->runtime_pm, work_wakeref);
+}
+
+/**
+ * intel_display_power_flush_work_sync - flushes and syncs the async display power disabling work
+ * @i915: i915 device instance
+ *
+ * Like intel_display_power_flush_work(), but also ensure that the work
+ * handler function is not running any more when this function returns.
+ */
+static void
+intel_display_power_flush_work_sync(struct drm_i915_private *i915)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+
+	intel_display_power_flush_work(i915);
+	cancel_delayed_work_sync(&power_domains->async_put_work);
+
+	verify_async_put_domains_state(power_domains);
+
+	drm_WARN_ON(&i915->drm, power_domains->async_put_wakeref);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+/**
+ * intel_display_power_put - release a power domain reference
+ * @dev_priv: i915 device instance
+ * @domain: power domain to reference
+ * @wakeref: wakeref acquired for the reference that is being released
+ *
+ * This function drops the power domain reference obtained by
+ * intel_display_power_get() and might power down the corresponding hardware
+ * block right away if this is the last reference.
+ */
+void intel_display_power_put(struct drm_i915_private *dev_priv,
+			     enum intel_display_power_domain domain,
+			     intel_wakeref_t wakeref)
+{
+	__intel_display_power_put(dev_priv, domain);
+	intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+}
+#else
+/**
+ * intel_display_power_put_unchecked - release an unchecked power domain reference
+ * @dev_priv: i915 device instance
+ * @domain: power domain to reference
+ *
+ * This function drops the power domain reference obtained by
+ * intel_display_power_get() and might power down the corresponding hardware
+ * block right away if this is the last reference.
+ *
+ * This function is only for the power domain code's internal use to suppress wakeref
+ * tracking when the correspondig debug kconfig option is disabled, should not
+ * be used otherwise.
+ */
+void intel_display_power_put_unchecked(struct drm_i915_private *dev_priv,
+				       enum intel_display_power_domain domain)
+{
+	__intel_display_power_put(dev_priv, domain);
+	intel_runtime_pm_put_unchecked(&dev_priv->runtime_pm);
+}
+#endif
+
+void
+intel_display_power_get_in_set(struct drm_i915_private *i915,
+			       struct intel_display_power_domain_set *power_domain_set,
+			       enum intel_display_power_domain domain)
+{
+	intel_wakeref_t __maybe_unused wf;
+
+	drm_WARN_ON(&i915->drm, test_bit(domain, power_domain_set->mask.bits));
+
+	wf = intel_display_power_get(i915, domain);
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+	power_domain_set->wakerefs[domain] = wf;
+#endif
+	set_bit(domain, power_domain_set->mask.bits);
+}
+
+bool
+intel_display_power_get_in_set_if_enabled(struct drm_i915_private *i915,
+					  struct intel_display_power_domain_set *power_domain_set,
+					  enum intel_display_power_domain domain)
+{
+	intel_wakeref_t wf;
+
+	drm_WARN_ON(&i915->drm, test_bit(domain, power_domain_set->mask.bits));
+
+	wf = intel_display_power_get_if_enabled(i915, domain);
+	if (!wf)
+		return false;
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+	power_domain_set->wakerefs[domain] = wf;
+#endif
+	set_bit(domain, power_domain_set->mask.bits);
+
+	return true;
+}
+
+void
+intel_display_power_put_mask_in_set(struct drm_i915_private *i915,
+				    struct intel_display_power_domain_set *power_domain_set,
+				    struct intel_power_domain_mask *mask)
+{
+	enum intel_display_power_domain domain;
+
+	drm_WARN_ON(&i915->drm,
+		    !bitmap_subset(mask->bits, power_domain_set->mask.bits, POWER_DOMAIN_NUM));
+
+	for_each_power_domain(domain, mask) {
+		intel_wakeref_t __maybe_unused wf = -1;
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+		wf = fetch_and_zero(&power_domain_set->wakerefs[domain]);
+#endif
+		intel_display_power_put(i915, domain, wf);
+		clear_bit(domain, power_domain_set->mask.bits);
+	}
+}
+
+static int
+sanitize_disable_power_well_option(const struct drm_i915_private *dev_priv,
+				   int disable_power_well)
+{
+	if (disable_power_well >= 0)
+		return !!disable_power_well;
+
+	return 1;
+}
+
+static u32 get_allowed_dc_mask(const struct drm_i915_private *dev_priv,
+			       int enable_dc)
+{
+	u32 mask;
+	int requested_dc;
+	int max_dc;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return 0;
+
+	if (IS_DG2(dev_priv))
+		max_dc = 1;
+	else if (IS_DG1(dev_priv))
+		max_dc = 3;
+	else if (DISPLAY_VER(dev_priv) >= 12)
+		max_dc = 4;
+	else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		max_dc = 1;
+	else if (DISPLAY_VER(dev_priv) >= 9)
+		max_dc = 2;
+	else
+		max_dc = 0;
+
+	/*
+	 * DC9 has a separate HW flow from the rest of the DC states,
+	 * not depending on the DMC firmware. It's needed by system
+	 * suspend/resume, so allow it unconditionally.
+	 */
+	mask = IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv) ||
+		DISPLAY_VER(dev_priv) >= 11 ?
+	       DC_STATE_EN_DC9 : 0;
+
+	if (!dev_priv->params.disable_power_well)
+		max_dc = 0;
+
+	if (enable_dc >= 0 && enable_dc <= max_dc) {
+		requested_dc = enable_dc;
+	} else if (enable_dc == -1) {
+		requested_dc = max_dc;
+	} else if (enable_dc > max_dc && enable_dc <= 4) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Adjusting requested max DC state (%d->%d)\n",
+			    enable_dc, max_dc);
+		requested_dc = max_dc;
+	} else {
+		drm_err(&dev_priv->drm,
+			"Unexpected value for enable_dc (%d)\n", enable_dc);
+		requested_dc = max_dc;
+	}
+
+	switch (requested_dc) {
+	case 4:
+		mask |= DC_STATE_EN_DC3CO | DC_STATE_EN_UPTO_DC6;
+		break;
+	case 3:
+		mask |= DC_STATE_EN_DC3CO | DC_STATE_EN_UPTO_DC5;
+		break;
+	case 2:
+		mask |= DC_STATE_EN_UPTO_DC6;
+		break;
+	case 1:
+		mask |= DC_STATE_EN_UPTO_DC5;
+		break;
+	}
+
+	drm_dbg_kms(&dev_priv->drm, "Allowed DC state mask %02x\n", mask);
+
+	return mask;
+}
+
+/**
+ * intel_power_domains_init - initializes the power domain structures
+ * @dev_priv: i915 device instance
+ *
+ * Initializes the power domain structures for @dev_priv depending upon the
+ * supported platform.
+ */
+int intel_power_domains_init(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+
+	dev_priv->params.disable_power_well =
+		sanitize_disable_power_well_option(dev_priv,
+						   dev_priv->params.disable_power_well);
+	dev_priv->display.dmc.allowed_dc_mask =
+		get_allowed_dc_mask(dev_priv, dev_priv->params.enable_dc);
+
+	dev_priv->display.dmc.target_dc_state =
+		sanitize_target_dc_state(dev_priv, DC_STATE_EN_UPTO_DC6);
+
+	mutex_init(&power_domains->lock);
+
+	INIT_DELAYED_WORK(&power_domains->async_put_work,
+			  intel_display_power_put_async_work);
+
+	return intel_display_power_map_init(power_domains);
+}
+
+/**
+ * intel_power_domains_cleanup - clean up power domains resources
+ * @dev_priv: i915 device instance
+ *
+ * Release any resources acquired by intel_power_domains_init()
+ */
+void intel_power_domains_cleanup(struct drm_i915_private *dev_priv)
+{
+	intel_display_power_map_cleanup(&dev_priv->display.power.domains);
+}
+
+static void intel_power_domains_sync_hw(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	struct i915_power_well *power_well;
+
+	mutex_lock(&power_domains->lock);
+	for_each_power_well(dev_priv, power_well)
+		intel_power_well_sync_hw(dev_priv, power_well);
+	mutex_unlock(&power_domains->lock);
+}
+
+static void gen9_dbuf_slice_set(struct drm_i915_private *dev_priv,
+				enum dbuf_slice slice, bool enable)
+{
+	i915_reg_t reg = DBUF_CTL_S(slice);
+	bool state;
+
+	intel_de_rmw(dev_priv, reg, DBUF_POWER_REQUEST,
+		     enable ? DBUF_POWER_REQUEST : 0);
+	intel_de_posting_read(dev_priv, reg);
+	udelay(10);
+
+	state = intel_de_read(dev_priv, reg) & DBUF_POWER_STATE;
+	drm_WARN(&dev_priv->drm, enable != state,
+		 "DBuf slice %d power %s timeout!\n",
+		 slice, str_enable_disable(enable));
+}
+
+void gen9_dbuf_slices_update(struct drm_i915_private *dev_priv,
+			     u8 req_slices)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	u8 slice_mask = INTEL_INFO(dev_priv)->display.dbuf.slice_mask;
+	enum dbuf_slice slice;
+
+	drm_WARN(&dev_priv->drm, req_slices & ~slice_mask,
+		 "Invalid set of dbuf slices (0x%x) requested (total dbuf slices 0x%x)\n",
+		 req_slices, slice_mask);
+
+	drm_dbg_kms(&dev_priv->drm, "Updating dbuf slices to 0x%x\n",
+		    req_slices);
+
+	/*
+	 * Might be running this in parallel to gen9_dc_off_power_well_enable
+	 * being called from intel_dp_detect for instance,
+	 * which causes assertion triggered by race condition,
+	 * as gen9_assert_dbuf_enabled might preempt this when registers
+	 * were already updated, while dev_priv was not.
+	 */
+	mutex_lock(&power_domains->lock);
+
+	for_each_dbuf_slice(dev_priv, slice)
+		gen9_dbuf_slice_set(dev_priv, slice, req_slices & BIT(slice));
+
+	dev_priv->display.dbuf.enabled_slices = req_slices;
+
+	mutex_unlock(&power_domains->lock);
+}
+
+static void gen9_dbuf_enable(struct drm_i915_private *dev_priv)
+{
+	dev_priv->display.dbuf.enabled_slices =
+		intel_enabled_dbuf_slices_mask(dev_priv);
+
+	/*
+	 * Just power up at least 1 slice, we will
+	 * figure out later which slices we have and what we need.
+	 */
+	gen9_dbuf_slices_update(dev_priv, BIT(DBUF_S1) |
+				dev_priv->display.dbuf.enabled_slices);
+}
+
+static void gen9_dbuf_disable(struct drm_i915_private *dev_priv)
+{
+	gen9_dbuf_slices_update(dev_priv, 0);
+}
+
+static void gen12_dbuf_slices_config(struct drm_i915_private *dev_priv)
+{
+	enum dbuf_slice slice;
+
+	if (IS_ALDERLAKE_P(dev_priv))
+		return;
+
+	for_each_dbuf_slice(dev_priv, slice)
+		intel_de_rmw(dev_priv, DBUF_CTL_S(slice),
+			     DBUF_TRACKER_STATE_SERVICE_MASK,
+			     DBUF_TRACKER_STATE_SERVICE(8));
+}
+
+static void icl_mbus_init(struct drm_i915_private *dev_priv)
+{
+	unsigned long abox_regs = INTEL_INFO(dev_priv)->display.abox_mask;
+	u32 mask, val, i;
+
+	if (IS_ALDERLAKE_P(dev_priv) || DISPLAY_VER(dev_priv) >= 14)
+		return;
+
+	mask = MBUS_ABOX_BT_CREDIT_POOL1_MASK |
+		MBUS_ABOX_BT_CREDIT_POOL2_MASK |
+		MBUS_ABOX_B_CREDIT_MASK |
+		MBUS_ABOX_BW_CREDIT_MASK;
+	val = MBUS_ABOX_BT_CREDIT_POOL1(16) |
+		MBUS_ABOX_BT_CREDIT_POOL2(16) |
+		MBUS_ABOX_B_CREDIT(1) |
+		MBUS_ABOX_BW_CREDIT(1);
+
+	/*
+	 * gen12 platforms that use abox1 and abox2 for pixel data reads still
+	 * expect us to program the abox_ctl0 register as well, even though
+	 * we don't have to program other instance-0 registers like BW_BUDDY.
+	 */
+	if (DISPLAY_VER(dev_priv) == 12)
+		abox_regs |= BIT(0);
+
+	for_each_set_bit(i, &abox_regs, sizeof(abox_regs))
+		intel_de_rmw(dev_priv, MBUS_ABOX_CTL(i), mask, val);
+}
+
+static void hsw_assert_cdclk(struct drm_i915_private *dev_priv)
+{
+	u32 val = intel_de_read(dev_priv, LCPLL_CTL);
+
+	/*
+	 * The LCPLL register should be turned on by the BIOS. For now
+	 * let's just check its state and print errors in case
+	 * something is wrong.  Don't even try to turn it on.
+	 */
+
+	if (val & LCPLL_CD_SOURCE_FCLK)
+		drm_err(&dev_priv->drm, "CDCLK source is not LCPLL\n");
+
+	if (val & LCPLL_PLL_DISABLE)
+		drm_err(&dev_priv->drm, "LCPLL is disabled\n");
+
+	if ((val & LCPLL_REF_MASK) != LCPLL_REF_NON_SSC)
+		drm_err(&dev_priv->drm, "LCPLL not using non-SSC reference\n");
+}
+
+static void assert_can_disable_lcpll(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(dev, crtc)
+		I915_STATE_WARN(crtc->active, "CRTC for pipe %c enabled\n",
+				pipe_name(crtc->pipe));
+
+	I915_STATE_WARN(intel_de_read(dev_priv, HSW_PWR_WELL_CTL2),
+			"Display power well on\n");
+	I915_STATE_WARN(intel_de_read(dev_priv, SPLL_CTL) & SPLL_PLL_ENABLE,
+			"SPLL enabled\n");
+	I915_STATE_WARN(intel_de_read(dev_priv, WRPLL_CTL(0)) & WRPLL_PLL_ENABLE,
+			"WRPLL1 enabled\n");
+	I915_STATE_WARN(intel_de_read(dev_priv, WRPLL_CTL(1)) & WRPLL_PLL_ENABLE,
+			"WRPLL2 enabled\n");
+	I915_STATE_WARN(intel_de_read(dev_priv, PP_STATUS(0)) & PP_ON,
+			"Panel power on\n");
+	I915_STATE_WARN(intel_de_read(dev_priv, BLC_PWM_CPU_CTL2) & BLM_PWM_ENABLE,
+			"CPU PWM1 enabled\n");
+	if (IS_HASWELL(dev_priv))
+		I915_STATE_WARN(intel_de_read(dev_priv, HSW_BLC_PWM2_CTL) & BLM_PWM_ENABLE,
+				"CPU PWM2 enabled\n");
+	I915_STATE_WARN(intel_de_read(dev_priv, BLC_PWM_PCH_CTL1) & BLM_PCH_PWM_ENABLE,
+			"PCH PWM1 enabled\n");
+	I915_STATE_WARN(intel_de_read(dev_priv, UTIL_PIN_CTL) & UTIL_PIN_ENABLE,
+			"Utility pin enabled\n");
+	I915_STATE_WARN(intel_de_read(dev_priv, PCH_GTC_CTL) & PCH_GTC_ENABLE,
+			"PCH GTC enabled\n");
+
+	/*
+	 * In theory we can still leave IRQs enabled, as long as only the HPD
+	 * interrupts remain enabled. We used to check for that, but since it's
+	 * gen-specific and since we only disable LCPLL after we fully disable
+	 * the interrupts, the check below should be enough.
+	 */
+	I915_STATE_WARN(intel_irqs_enabled(dev_priv), "IRQs enabled\n");
+}
+
+static u32 hsw_read_dcomp(struct drm_i915_private *dev_priv)
+{
+	if (IS_HASWELL(dev_priv))
+		return intel_de_read(dev_priv, D_COMP_HSW);
+	else
+		return intel_de_read(dev_priv, D_COMP_BDW);
+}
+
+static void hsw_write_dcomp(struct drm_i915_private *dev_priv, u32 val)
+{
+	if (IS_HASWELL(dev_priv)) {
+		if (snb_pcode_write(&dev_priv->uncore, GEN6_PCODE_WRITE_D_COMP, val))
+			drm_dbg_kms(&dev_priv->drm,
+				    "Failed to write to D_COMP\n");
+	} else {
+		intel_de_write(dev_priv, D_COMP_BDW, val);
+		intel_de_posting_read(dev_priv, D_COMP_BDW);
+	}
+}
+
+/*
+ * This function implements pieces of two sequences from BSpec:
+ * - Sequence for display software to disable LCPLL
+ * - Sequence for display software to allow package C8+
+ * The steps implemented here are just the steps that actually touch the LCPLL
+ * register. Callers should take care of disabling all the display engine
+ * functions, doing the mode unset, fixing interrupts, etc.
+ */
+static void hsw_disable_lcpll(struct drm_i915_private *dev_priv,
+			      bool switch_to_fclk, bool allow_power_down)
+{
+	u32 val;
+
+	assert_can_disable_lcpll(dev_priv);
+
+	val = intel_de_read(dev_priv, LCPLL_CTL);
+
+	if (switch_to_fclk) {
+		val |= LCPLL_CD_SOURCE_FCLK;
+		intel_de_write(dev_priv, LCPLL_CTL, val);
+
+		if (wait_for_us(intel_de_read(dev_priv, LCPLL_CTL) &
+				LCPLL_CD_SOURCE_FCLK_DONE, 1))
+			drm_err(&dev_priv->drm, "Switching to FCLK failed\n");
+
+		val = intel_de_read(dev_priv, LCPLL_CTL);
+	}
+
+	val |= LCPLL_PLL_DISABLE;
+	intel_de_write(dev_priv, LCPLL_CTL, val);
+	intel_de_posting_read(dev_priv, LCPLL_CTL);
+
+	if (intel_de_wait_for_clear(dev_priv, LCPLL_CTL, LCPLL_PLL_LOCK, 1))
+		drm_err(&dev_priv->drm, "LCPLL still locked\n");
+
+	val = hsw_read_dcomp(dev_priv);
+	val |= D_COMP_COMP_DISABLE;
+	hsw_write_dcomp(dev_priv, val);
+	ndelay(100);
+
+	if (wait_for((hsw_read_dcomp(dev_priv) &
+		      D_COMP_RCOMP_IN_PROGRESS) == 0, 1))
+		drm_err(&dev_priv->drm, "D_COMP RCOMP still in progress\n");
+
+	if (allow_power_down) {
+		val = intel_de_read(dev_priv, LCPLL_CTL);
+		val |= LCPLL_POWER_DOWN_ALLOW;
+		intel_de_write(dev_priv, LCPLL_CTL, val);
+		intel_de_posting_read(dev_priv, LCPLL_CTL);
+	}
+}
+
+/*
+ * Fully restores LCPLL, disallowing power down and switching back to LCPLL
+ * source.
+ */
+static void hsw_restore_lcpll(struct drm_i915_private *dev_priv)
+{
+	u32 val;
+
+	val = intel_de_read(dev_priv, LCPLL_CTL);
+
+	if ((val & (LCPLL_PLL_LOCK | LCPLL_PLL_DISABLE | LCPLL_CD_SOURCE_FCLK |
+		    LCPLL_POWER_DOWN_ALLOW)) == LCPLL_PLL_LOCK)
+		return;
+
+	/*
+	 * Make sure we're not on PC8 state before disabling PC8, otherwise
+	 * we'll hang the machine. To prevent PC8 state, just enable force_wake.
+	 */
+	intel_uncore_forcewake_get(&dev_priv->uncore, FORCEWAKE_ALL);
+
+	if (val & LCPLL_POWER_DOWN_ALLOW) {
+		val &= ~LCPLL_POWER_DOWN_ALLOW;
+		intel_de_write(dev_priv, LCPLL_CTL, val);
+		intel_de_posting_read(dev_priv, LCPLL_CTL);
+	}
+
+	val = hsw_read_dcomp(dev_priv);
+	val |= D_COMP_COMP_FORCE;
+	val &= ~D_COMP_COMP_DISABLE;
+	hsw_write_dcomp(dev_priv, val);
+
+	val = intel_de_read(dev_priv, LCPLL_CTL);
+	val &= ~LCPLL_PLL_DISABLE;
+	intel_de_write(dev_priv, LCPLL_CTL, val);
+
+	if (intel_de_wait_for_set(dev_priv, LCPLL_CTL, LCPLL_PLL_LOCK, 5))
+		drm_err(&dev_priv->drm, "LCPLL not locked yet\n");
+
+	if (val & LCPLL_CD_SOURCE_FCLK) {
+		val = intel_de_read(dev_priv, LCPLL_CTL);
+		val &= ~LCPLL_CD_SOURCE_FCLK;
+		intel_de_write(dev_priv, LCPLL_CTL, val);
+
+		if (wait_for_us((intel_de_read(dev_priv, LCPLL_CTL) &
+				 LCPLL_CD_SOURCE_FCLK_DONE) == 0, 1))
+			drm_err(&dev_priv->drm,
+				"Switching back to LCPLL failed\n");
+	}
+
+	intel_uncore_forcewake_put(&dev_priv->uncore, FORCEWAKE_ALL);
+
+	intel_update_cdclk(dev_priv);
+	intel_cdclk_dump_config(dev_priv, &dev_priv->display.cdclk.hw, "Current CDCLK");
+}
+
+/*
+ * Package states C8 and deeper are really deep PC states that can only be
+ * reached when all the devices on the system allow it, so even if the graphics
+ * device allows PC8+, it doesn't mean the system will actually get to these
+ * states. Our driver only allows PC8+ when going into runtime PM.
+ *
+ * The requirements for PC8+ are that all the outputs are disabled, the power
+ * well is disabled and most interrupts are disabled, and these are also
+ * requirements for runtime PM. When these conditions are met, we manually do
+ * the other conditions: disable the interrupts, clocks and switch LCPLL refclk
+ * to Fclk. If we're in PC8+ and we get an non-hotplug interrupt, we can hard
+ * hang the machine.
+ *
+ * When we really reach PC8 or deeper states (not just when we allow it) we lose
+ * the state of some registers, so when we come back from PC8+ we need to
+ * restore this state. We don't get into PC8+ if we're not in RC6, so we don't
+ * need to take care of the registers kept by RC6. Notice that this happens even
+ * if we don't put the device in PCI D3 state (which is what currently happens
+ * because of the runtime PM support).
+ *
+ * For more, read "Display Sequences for Package C8" on the hardware
+ * documentation.
+ */
+static void hsw_enable_pc8(struct drm_i915_private *dev_priv)
+{
+	u32 val;
+
+	drm_dbg_kms(&dev_priv->drm, "Enabling package C8+\n");
+
+	if (HAS_PCH_LPT_LP(dev_priv)) {
+		val = intel_de_read(dev_priv, SOUTH_DSPCLK_GATE_D);
+		val &= ~PCH_LP_PARTITION_LEVEL_DISABLE;
+		intel_de_write(dev_priv, SOUTH_DSPCLK_GATE_D, val);
+	}
+
+	lpt_disable_clkout_dp(dev_priv);
+	hsw_disable_lcpll(dev_priv, true, true);
+}
+
+static void hsw_disable_pc8(struct drm_i915_private *dev_priv)
+{
+	u32 val;
+
+	drm_dbg_kms(&dev_priv->drm, "Disabling package C8+\n");
+
+	hsw_restore_lcpll(dev_priv);
+	intel_init_pch_refclk(dev_priv);
+
+	if (HAS_PCH_LPT_LP(dev_priv)) {
+		val = intel_de_read(dev_priv, SOUTH_DSPCLK_GATE_D);
+		val |= PCH_LP_PARTITION_LEVEL_DISABLE;
+		intel_de_write(dev_priv, SOUTH_DSPCLK_GATE_D, val);
+	}
+}
+
+static void intel_pch_reset_handshake(struct drm_i915_private *dev_priv,
+				      bool enable)
+{
+	i915_reg_t reg;
+	u32 reset_bits, val;
+
+	if (IS_IVYBRIDGE(dev_priv)) {
+		reg = GEN7_MSG_CTL;
+		reset_bits = WAIT_FOR_PCH_FLR_ACK | WAIT_FOR_PCH_RESET_ACK;
+	} else {
+		reg = HSW_NDE_RSTWRN_OPT;
+		reset_bits = RESET_PCH_HANDSHAKE_ENABLE;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 14)
+		reset_bits |= MTL_RESET_PICA_HANDSHAKE_EN;
+
+	val = intel_de_read(dev_priv, reg);
+
+	if (enable)
+		val |= reset_bits;
+	else
+		val &= ~reset_bits;
+
+	intel_de_write(dev_priv, reg, val);
+}
+
+static void skl_display_core_init(struct drm_i915_private *dev_priv,
+				  bool resume)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	struct i915_power_well *well;
+
+	gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
+
+	/* enable PCH reset handshake */
+	intel_pch_reset_handshake(dev_priv, !HAS_PCH_NOP(dev_priv));
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	/* enable PG1 and Misc I/O */
+	mutex_lock(&power_domains->lock);
+
+	well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
+	intel_power_well_enable(dev_priv, well);
+
+	well = lookup_power_well(dev_priv, SKL_DISP_PW_MISC_IO);
+	intel_power_well_enable(dev_priv, well);
+
+	mutex_unlock(&power_domains->lock);
+
+	intel_cdclk_init_hw(dev_priv);
+
+	gen9_dbuf_enable(dev_priv);
+
+	if (resume)
+		intel_dmc_load_program(dev_priv);
+}
+
+static void skl_display_core_uninit(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	struct i915_power_well *well;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	gen9_disable_dc_states(dev_priv);
+	/* TODO: disable DMC program */
+
+	gen9_dbuf_disable(dev_priv);
+
+	intel_cdclk_uninit_hw(dev_priv);
+
+	/* The spec doesn't call for removing the reset handshake flag */
+	/* disable PG1 and Misc I/O */
+
+	mutex_lock(&power_domains->lock);
+
+	/*
+	 * BSpec says to keep the MISC IO power well enabled here, only
+	 * remove our request for power well 1.
+	 * Note that even though the driver's request is removed power well 1
+	 * may stay enabled after this due to DMC's own request on it.
+	 */
+	well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
+	intel_power_well_disable(dev_priv, well);
+
+	mutex_unlock(&power_domains->lock);
+
+	usleep_range(10, 30);		/* 10 us delay per Bspec */
+}
+
+static void bxt_display_core_init(struct drm_i915_private *dev_priv, bool resume)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	struct i915_power_well *well;
+
+	gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
+
+	/*
+	 * NDE_RSTWRN_OPT RST PCH Handshake En must always be 0b on BXT
+	 * or else the reset will hang because there is no PCH to respond.
+	 * Move the handshake programming to initialization sequence.
+	 * Previously was left up to BIOS.
+	 */
+	intel_pch_reset_handshake(dev_priv, false);
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	/* Enable PG1 */
+	mutex_lock(&power_domains->lock);
+
+	well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
+	intel_power_well_enable(dev_priv, well);
+
+	mutex_unlock(&power_domains->lock);
+
+	intel_cdclk_init_hw(dev_priv);
+
+	gen9_dbuf_enable(dev_priv);
+
+	if (resume)
+		intel_dmc_load_program(dev_priv);
+}
+
+static void bxt_display_core_uninit(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	struct i915_power_well *well;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	gen9_disable_dc_states(dev_priv);
+	/* TODO: disable DMC program */
+
+	gen9_dbuf_disable(dev_priv);
+
+	intel_cdclk_uninit_hw(dev_priv);
+
+	/* The spec doesn't call for removing the reset handshake flag */
+
+	/*
+	 * Disable PW1 (PG1).
+	 * Note that even though the driver's request is removed power well 1
+	 * may stay enabled after this due to DMC's own request on it.
+	 */
+	mutex_lock(&power_domains->lock);
+
+	well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
+	intel_power_well_disable(dev_priv, well);
+
+	mutex_unlock(&power_domains->lock);
+
+	usleep_range(10, 30);		/* 10 us delay per Bspec */
+}
+
+struct buddy_page_mask {
+	u32 page_mask;
+	u8 type;
+	u8 num_channels;
+};
+
+static const struct buddy_page_mask tgl_buddy_page_masks[] = {
+	{ .num_channels = 1, .type = INTEL_DRAM_DDR4,   .page_mask = 0xF },
+	{ .num_channels = 1, .type = INTEL_DRAM_DDR5,	.page_mask = 0xF },
+	{ .num_channels = 2, .type = INTEL_DRAM_LPDDR4, .page_mask = 0x1C },
+	{ .num_channels = 2, .type = INTEL_DRAM_LPDDR5, .page_mask = 0x1C },
+	{ .num_channels = 2, .type = INTEL_DRAM_DDR4,   .page_mask = 0x1F },
+	{ .num_channels = 2, .type = INTEL_DRAM_DDR5,   .page_mask = 0x1E },
+	{ .num_channels = 4, .type = INTEL_DRAM_LPDDR4, .page_mask = 0x38 },
+	{ .num_channels = 4, .type = INTEL_DRAM_LPDDR5, .page_mask = 0x38 },
+	{}
+};
+
+static const struct buddy_page_mask wa_1409767108_buddy_page_masks[] = {
+	{ .num_channels = 1, .type = INTEL_DRAM_LPDDR4, .page_mask = 0x1 },
+	{ .num_channels = 1, .type = INTEL_DRAM_DDR4,   .page_mask = 0x1 },
+	{ .num_channels = 1, .type = INTEL_DRAM_DDR5,   .page_mask = 0x1 },
+	{ .num_channels = 1, .type = INTEL_DRAM_LPDDR5, .page_mask = 0x1 },
+	{ .num_channels = 2, .type = INTEL_DRAM_LPDDR4, .page_mask = 0x3 },
+	{ .num_channels = 2, .type = INTEL_DRAM_DDR4,   .page_mask = 0x3 },
+	{ .num_channels = 2, .type = INTEL_DRAM_DDR5,   .page_mask = 0x3 },
+	{ .num_channels = 2, .type = INTEL_DRAM_LPDDR5, .page_mask = 0x3 },
+	{}
+};
+
+static void tgl_bw_buddy_init(struct drm_i915_private *dev_priv)
+{
+	enum intel_dram_type type = dev_priv->dram_info.type;
+	u8 num_channels = dev_priv->dram_info.num_channels;
+	const struct buddy_page_mask *table;
+	unsigned long abox_mask = INTEL_INFO(dev_priv)->display.abox_mask;
+	int config, i;
+
+	/* BW_BUDDY registers are not used on dgpu's beyond DG1 */
+	if (IS_DGFX(dev_priv) && !IS_DG1(dev_priv))
+		return;
+
+	if (IS_ALDERLAKE_S(dev_priv) ||
+	    IS_DG1_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0) ||
+	    IS_RKL_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0) ||
+	    IS_TGL_DISPLAY_STEP(dev_priv, STEP_A0, STEP_C0))
+		/* Wa_1409767108:tgl,dg1,adl-s */
+		table = wa_1409767108_buddy_page_masks;
+	else
+		table = tgl_buddy_page_masks;
+
+	for (config = 0; table[config].page_mask != 0; config++)
+		if (table[config].num_channels == num_channels &&
+		    table[config].type == type)
+			break;
+
+	if (table[config].page_mask == 0) {
+		drm_dbg(&dev_priv->drm,
+			"Unknown memory configuration; disabling address buddy logic.\n");
+		for_each_set_bit(i, &abox_mask, sizeof(abox_mask))
+			intel_de_write(dev_priv, BW_BUDDY_CTL(i),
+				       BW_BUDDY_DISABLE);
+	} else {
+		for_each_set_bit(i, &abox_mask, sizeof(abox_mask)) {
+			intel_de_write(dev_priv, BW_BUDDY_PAGE_MASK(i),
+				       table[config].page_mask);
+
+			/* Wa_22010178259:tgl,dg1,rkl,adl-s */
+			if (DISPLAY_VER(dev_priv) == 12)
+				intel_de_rmw(dev_priv, BW_BUDDY_CTL(i),
+					     BW_BUDDY_TLB_REQ_TIMER_MASK,
+					     BW_BUDDY_TLB_REQ_TIMER(0x8));
+		}
+	}
+}
+
+static void icl_display_core_init(struct drm_i915_private *dev_priv,
+				  bool resume)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	struct i915_power_well *well;
+	u32 val;
+
+	gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
+
+	/* Wa_14011294188:ehl,jsl,tgl,rkl,adl-s */
+	if (INTEL_PCH_TYPE(dev_priv) >= PCH_TGP &&
+	    INTEL_PCH_TYPE(dev_priv) < PCH_DG1)
+		intel_de_rmw(dev_priv, SOUTH_DSPCLK_GATE_D, 0,
+			     PCH_DPMGUNIT_CLOCK_GATE_DISABLE);
+
+	/* 1. Enable PCH reset handshake. */
+	intel_pch_reset_handshake(dev_priv, !HAS_PCH_NOP(dev_priv));
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	/* 2. Initialize all combo phys */
+	intel_combo_phy_init(dev_priv);
+
+	/*
+	 * 3. Enable Power Well 1 (PG1).
+	 *    The AUX IO power wells will be enabled on demand.
+	 */
+	mutex_lock(&power_domains->lock);
+	well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
+	intel_power_well_enable(dev_priv, well);
+	mutex_unlock(&power_domains->lock);
+
+	/* 4. Enable CDCLK. */
+	intel_cdclk_init_hw(dev_priv);
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		gen12_dbuf_slices_config(dev_priv);
+
+	/* 5. Enable DBUF. */
+	gen9_dbuf_enable(dev_priv);
+
+	/* 6. Setup MBUS. */
+	icl_mbus_init(dev_priv);
+
+	/* 7. Program arbiter BW_BUDDY registers */
+	if (DISPLAY_VER(dev_priv) >= 12)
+		tgl_bw_buddy_init(dev_priv);
+
+	/* 8. Ensure PHYs have completed calibration and adaptation */
+	if (IS_DG2(dev_priv))
+		intel_snps_phy_wait_for_calibration(dev_priv);
+
+	if (resume)
+		intel_dmc_load_program(dev_priv);
+
+	/* Wa_14011508470:tgl,dg1,rkl,adl-s,adl-p */
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		val = DCPR_CLEAR_MEMSTAT_DIS | DCPR_SEND_RESP_IMM |
+		      DCPR_MASK_LPMODE | DCPR_MASK_MAXLATENCY_MEMUP_CLR;
+		intel_uncore_rmw(&dev_priv->uncore, GEN11_CHICKEN_DCPR_2, 0, val);
+	}
+
+	/* Wa_14011503030:xelpd */
+	if (DISPLAY_VER(dev_priv) >= 13)
+		intel_de_write(dev_priv, XELPD_DISPLAY_ERR_FATAL_MASK, ~0);
+}
+
+static void icl_display_core_uninit(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	struct i915_power_well *well;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	gen9_disable_dc_states(dev_priv);
+	intel_dmc_disable_program(dev_priv);
+
+	/* 1. Disable all display engine functions -> aready done */
+
+	/* 2. Disable DBUF */
+	gen9_dbuf_disable(dev_priv);
+
+	/* 3. Disable CD clock */
+	intel_cdclk_uninit_hw(dev_priv);
+
+	/*
+	 * 4. Disable Power Well 1 (PG1).
+	 *    The AUX IO power wells are toggled on demand, so they are already
+	 *    disabled at this point.
+	 */
+	mutex_lock(&power_domains->lock);
+	well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
+	intel_power_well_disable(dev_priv, well);
+	mutex_unlock(&power_domains->lock);
+
+	/* 5. */
+	intel_combo_phy_uninit(dev_priv);
+}
+
+static void chv_phy_control_init(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_well *cmn_bc =
+		lookup_power_well(dev_priv, VLV_DISP_PW_DPIO_CMN_BC);
+	struct i915_power_well *cmn_d =
+		lookup_power_well(dev_priv, CHV_DISP_PW_DPIO_CMN_D);
+
+	/*
+	 * DISPLAY_PHY_CONTROL can get corrupted if read. As a
+	 * workaround never ever read DISPLAY_PHY_CONTROL, and
+	 * instead maintain a shadow copy ourselves. Use the actual
+	 * power well state and lane status to reconstruct the
+	 * expected initial value.
+	 */
+	dev_priv->display.power.chv_phy_control =
+		PHY_LDO_SEQ_DELAY(PHY_LDO_DELAY_600NS, DPIO_PHY0) |
+		PHY_LDO_SEQ_DELAY(PHY_LDO_DELAY_600NS, DPIO_PHY1) |
+		PHY_CH_POWER_MODE(PHY_CH_DEEP_PSR, DPIO_PHY0, DPIO_CH0) |
+		PHY_CH_POWER_MODE(PHY_CH_DEEP_PSR, DPIO_PHY0, DPIO_CH1) |
+		PHY_CH_POWER_MODE(PHY_CH_DEEP_PSR, DPIO_PHY1, DPIO_CH0);
+
+	/*
+	 * If all lanes are disabled we leave the override disabled
+	 * with all power down bits cleared to match the state we
+	 * would use after disabling the port. Otherwise enable the
+	 * override and set the lane powerdown bits accding to the
+	 * current lane status.
+	 */
+	if (intel_power_well_is_enabled(dev_priv, cmn_bc)) {
+		u32 status = intel_de_read(dev_priv, DPLL(PIPE_A));
+		unsigned int mask;
+
+		mask = status & DPLL_PORTB_READY_MASK;
+		if (mask == 0xf)
+			mask = 0x0;
+		else
+			dev_priv->display.power.chv_phy_control |=
+				PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY0, DPIO_CH0);
+
+		dev_priv->display.power.chv_phy_control |=
+			PHY_CH_POWER_DOWN_OVRD(mask, DPIO_PHY0, DPIO_CH0);
+
+		mask = (status & DPLL_PORTC_READY_MASK) >> 4;
+		if (mask == 0xf)
+			mask = 0x0;
+		else
+			dev_priv->display.power.chv_phy_control |=
+				PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY0, DPIO_CH1);
+
+		dev_priv->display.power.chv_phy_control |=
+			PHY_CH_POWER_DOWN_OVRD(mask, DPIO_PHY0, DPIO_CH1);
+
+		dev_priv->display.power.chv_phy_control |= PHY_COM_LANE_RESET_DEASSERT(DPIO_PHY0);
+
+		dev_priv->display.power.chv_phy_assert[DPIO_PHY0] = false;
+	} else {
+		dev_priv->display.power.chv_phy_assert[DPIO_PHY0] = true;
+	}
+
+	if (intel_power_well_is_enabled(dev_priv, cmn_d)) {
+		u32 status = intel_de_read(dev_priv, DPIO_PHY_STATUS);
+		unsigned int mask;
+
+		mask = status & DPLL_PORTD_READY_MASK;
+
+		if (mask == 0xf)
+			mask = 0x0;
+		else
+			dev_priv->display.power.chv_phy_control |=
+				PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY1, DPIO_CH0);
+
+		dev_priv->display.power.chv_phy_control |=
+			PHY_CH_POWER_DOWN_OVRD(mask, DPIO_PHY1, DPIO_CH0);
+
+		dev_priv->display.power.chv_phy_control |= PHY_COM_LANE_RESET_DEASSERT(DPIO_PHY1);
+
+		dev_priv->display.power.chv_phy_assert[DPIO_PHY1] = false;
+	} else {
+		dev_priv->display.power.chv_phy_assert[DPIO_PHY1] = true;
+	}
+
+	drm_dbg_kms(&dev_priv->drm, "Initial PHY_CONTROL=0x%08x\n",
+		    dev_priv->display.power.chv_phy_control);
+
+	/* Defer application of initial phy_control to enabling the powerwell */
+}
+
+static void vlv_cmnlane_wa(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_well *cmn =
+		lookup_power_well(dev_priv, VLV_DISP_PW_DPIO_CMN_BC);
+	struct i915_power_well *disp2d =
+		lookup_power_well(dev_priv, VLV_DISP_PW_DISP2D);
+
+	/* If the display might be already active skip this */
+	if (intel_power_well_is_enabled(dev_priv, cmn) &&
+	    intel_power_well_is_enabled(dev_priv, disp2d) &&
+	    intel_de_read(dev_priv, DPIO_CTL) & DPIO_CMNRST)
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "toggling display PHY side reset\n");
+
+	/* cmnlane needs DPLL registers */
+	intel_power_well_enable(dev_priv, disp2d);
+
+	/*
+	 * From VLV2A0_DP_eDP_HDMI_DPIO_driver_vbios_notes_11.docx:
+	 * Need to assert and de-assert PHY SB reset by gating the
+	 * common lane power, then un-gating it.
+	 * Simply ungating isn't enough to reset the PHY enough to get
+	 * ports and lanes running.
+	 */
+	intel_power_well_disable(dev_priv, cmn);
+}
+
+static bool vlv_punit_is_power_gated(struct drm_i915_private *dev_priv, u32 reg0)
+{
+	bool ret;
+
+	vlv_punit_get(dev_priv);
+	ret = (vlv_punit_read(dev_priv, reg0) & SSPM0_SSC_MASK) == SSPM0_SSC_PWR_GATE;
+	vlv_punit_put(dev_priv);
+
+	return ret;
+}
+
+static void assert_ved_power_gated(struct drm_i915_private *dev_priv)
+{
+	drm_WARN(&dev_priv->drm,
+		 !vlv_punit_is_power_gated(dev_priv, PUNIT_REG_VEDSSPM0),
+		 "VED not power gated\n");
+}
+
+static void assert_isp_power_gated(struct drm_i915_private *dev_priv)
+{
+	static const struct pci_device_id isp_ids[] = {
+		{PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x0f38)},
+		{PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x22b8)},
+		{}
+	};
+
+	drm_WARN(&dev_priv->drm, !pci_dev_present(isp_ids) &&
+		 !vlv_punit_is_power_gated(dev_priv, PUNIT_REG_ISPSSPM0),
+		 "ISP not power gated\n");
+}
+
+static void intel_power_domains_verify_state(struct drm_i915_private *dev_priv);
+
+/**
+ * intel_power_domains_init_hw - initialize hardware power domain state
+ * @i915: i915 device instance
+ * @resume: Called from resume code paths or not
+ *
+ * This function initializes the hardware power domain state and enables all
+ * power wells belonging to the INIT power domain. Power wells in other
+ * domains (and not in the INIT domain) are referenced or disabled by
+ * intel_modeset_readout_hw_state(). After that the reference count of each
+ * power well must match its HW enabled state, see
+ * intel_power_domains_verify_state().
+ *
+ * It will return with power domains disabled (to be enabled later by
+ * intel_power_domains_enable()) and must be paired with
+ * intel_power_domains_driver_remove().
+ */
+void intel_power_domains_init_hw(struct drm_i915_private *i915, bool resume)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+
+	power_domains->initializing = true;
+
+	if (DISPLAY_VER(i915) >= 11) {
+		icl_display_core_init(i915, resume);
+	} else if (IS_GEMINILAKE(i915) || IS_BROXTON(i915)) {
+		bxt_display_core_init(i915, resume);
+	} else if (DISPLAY_VER(i915) == 9) {
+		skl_display_core_init(i915, resume);
+	} else if (IS_CHERRYVIEW(i915)) {
+		mutex_lock(&power_domains->lock);
+		chv_phy_control_init(i915);
+		mutex_unlock(&power_domains->lock);
+		assert_isp_power_gated(i915);
+	} else if (IS_VALLEYVIEW(i915)) {
+		mutex_lock(&power_domains->lock);
+		vlv_cmnlane_wa(i915);
+		mutex_unlock(&power_domains->lock);
+		assert_ved_power_gated(i915);
+		assert_isp_power_gated(i915);
+	} else if (IS_BROADWELL(i915) || IS_HASWELL(i915)) {
+		hsw_assert_cdclk(i915);
+		intel_pch_reset_handshake(i915, !HAS_PCH_NOP(i915));
+	} else if (IS_IVYBRIDGE(i915)) {
+		intel_pch_reset_handshake(i915, !HAS_PCH_NOP(i915));
+	}
+
+	/*
+	 * Keep all power wells enabled for any dependent HW access during
+	 * initialization and to make sure we keep BIOS enabled display HW
+	 * resources powered until display HW readout is complete. We drop
+	 * this reference in intel_power_domains_enable().
+	 */
+	drm_WARN_ON(&i915->drm, power_domains->init_wakeref);
+	power_domains->init_wakeref =
+		intel_display_power_get(i915, POWER_DOMAIN_INIT);
+
+	/* Disable power support if the user asked so. */
+	if (!i915->params.disable_power_well) {
+		drm_WARN_ON(&i915->drm, power_domains->disable_wakeref);
+		i915->display.power.domains.disable_wakeref = intel_display_power_get(i915,
+										      POWER_DOMAIN_INIT);
+	}
+	intel_power_domains_sync_hw(i915);
+
+	power_domains->initializing = false;
+}
+
+/**
+ * intel_power_domains_driver_remove - deinitialize hw power domain state
+ * @i915: i915 device instance
+ *
+ * De-initializes the display power domain HW state. It also ensures that the
+ * device stays powered up so that the driver can be reloaded.
+ *
+ * It must be called with power domains already disabled (after a call to
+ * intel_power_domains_disable()) and must be paired with
+ * intel_power_domains_init_hw().
+ */
+void intel_power_domains_driver_remove(struct drm_i915_private *i915)
+{
+	intel_wakeref_t wakeref __maybe_unused =
+		fetch_and_zero(&i915->display.power.domains.init_wakeref);
+
+	/* Remove the refcount we took to keep power well support disabled. */
+	if (!i915->params.disable_power_well)
+		intel_display_power_put(i915, POWER_DOMAIN_INIT,
+					fetch_and_zero(&i915->display.power.domains.disable_wakeref));
+
+	intel_display_power_flush_work_sync(i915);
+
+	intel_power_domains_verify_state(i915);
+
+	/* Keep the power well enabled, but cancel its rpm wakeref. */
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+}
+
+/**
+ * intel_power_domains_sanitize_state - sanitize power domains state
+ * @i915: i915 device instance
+ *
+ * Sanitize the power domains state during driver loading and system resume.
+ * The function will disable all display power wells that BIOS has enabled
+ * without a user for it (any user for a power well has taken a reference
+ * on it by the time this function is called, after the state of all the
+ * pipe, encoder, etc. HW resources have been sanitized).
+ */
+void intel_power_domains_sanitize_state(struct drm_i915_private *i915)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+	struct i915_power_well *power_well;
+
+	mutex_lock(&power_domains->lock);
+
+	for_each_power_well_reverse(i915, power_well) {
+		if (power_well->desc->always_on || power_well->count ||
+		    !intel_power_well_is_enabled(i915, power_well))
+			continue;
+
+		drm_dbg_kms(&i915->drm,
+			    "BIOS left unused %s power well enabled, disabling it\n",
+			    intel_power_well_name(power_well));
+		intel_power_well_disable(i915, power_well);
+	}
+
+	mutex_unlock(&power_domains->lock);
+}
+
+/**
+ * intel_power_domains_enable - enable toggling of display power wells
+ * @i915: i915 device instance
+ *
+ * Enable the ondemand enabling/disabling of the display power wells. Note that
+ * power wells not belonging to POWER_DOMAIN_INIT are allowed to be toggled
+ * only at specific points of the display modeset sequence, thus they are not
+ * affected by the intel_power_domains_enable()/disable() calls. The purpose
+ * of these function is to keep the rest of power wells enabled until the end
+ * of display HW readout (which will acquire the power references reflecting
+ * the current HW state).
+ */
+void intel_power_domains_enable(struct drm_i915_private *i915)
+{
+	intel_wakeref_t wakeref __maybe_unused =
+		fetch_and_zero(&i915->display.power.domains.init_wakeref);
+
+	intel_display_power_put(i915, POWER_DOMAIN_INIT, wakeref);
+	intel_power_domains_verify_state(i915);
+}
+
+/**
+ * intel_power_domains_disable - disable toggling of display power wells
+ * @i915: i915 device instance
+ *
+ * Disable the ondemand enabling/disabling of the display power wells. See
+ * intel_power_domains_enable() for which power wells this call controls.
+ */
+void intel_power_domains_disable(struct drm_i915_private *i915)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+
+	drm_WARN_ON(&i915->drm, power_domains->init_wakeref);
+	power_domains->init_wakeref =
+		intel_display_power_get(i915, POWER_DOMAIN_INIT);
+
+	intel_power_domains_verify_state(i915);
+}
+
+/**
+ * intel_power_domains_suspend - suspend power domain state
+ * @i915: i915 device instance
+ * @suspend_mode: specifies the target suspend state (idle, mem, hibernation)
+ *
+ * This function prepares the hardware power domain state before entering
+ * system suspend.
+ *
+ * It must be called with power domains already disabled (after a call to
+ * intel_power_domains_disable()) and paired with intel_power_domains_resume().
+ */
+void intel_power_domains_suspend(struct drm_i915_private *i915,
+				 enum i915_drm_suspend_mode suspend_mode)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+	intel_wakeref_t wakeref __maybe_unused =
+		fetch_and_zero(&power_domains->init_wakeref);
+
+	intel_display_power_put(i915, POWER_DOMAIN_INIT, wakeref);
+
+	/*
+	 * In case of suspend-to-idle (aka S0ix) on a DMC platform without DC9
+	 * support don't manually deinit the power domains. This also means the
+	 * DMC firmware will stay active, it will power down any HW
+	 * resources as required and also enable deeper system power states
+	 * that would be blocked if the firmware was inactive.
+	 */
+	if (!(i915->display.dmc.allowed_dc_mask & DC_STATE_EN_DC9) &&
+	    suspend_mode == I915_DRM_SUSPEND_IDLE &&
+	    intel_dmc_has_payload(i915)) {
+		intel_display_power_flush_work(i915);
+		intel_power_domains_verify_state(i915);
+		return;
+	}
+
+	/*
+	 * Even if power well support was disabled we still want to disable
+	 * power wells if power domains must be deinitialized for suspend.
+	 */
+	if (!i915->params.disable_power_well)
+		intel_display_power_put(i915, POWER_DOMAIN_INIT,
+					fetch_and_zero(&i915->display.power.domains.disable_wakeref));
+
+	intel_display_power_flush_work(i915);
+	intel_power_domains_verify_state(i915);
+
+	if (DISPLAY_VER(i915) >= 11)
+		icl_display_core_uninit(i915);
+	else if (IS_GEMINILAKE(i915) || IS_BROXTON(i915))
+		bxt_display_core_uninit(i915);
+	else if (DISPLAY_VER(i915) == 9)
+		skl_display_core_uninit(i915);
+
+	power_domains->display_core_suspended = true;
+}
+
+/**
+ * intel_power_domains_resume - resume power domain state
+ * @i915: i915 device instance
+ *
+ * This function resume the hardware power domain state during system resume.
+ *
+ * It will return with power domain support disabled (to be enabled later by
+ * intel_power_domains_enable()) and must be paired with
+ * intel_power_domains_suspend().
+ */
+void intel_power_domains_resume(struct drm_i915_private *i915)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+
+	if (power_domains->display_core_suspended) {
+		intel_power_domains_init_hw(i915, true);
+		power_domains->display_core_suspended = false;
+	} else {
+		drm_WARN_ON(&i915->drm, power_domains->init_wakeref);
+		power_domains->init_wakeref =
+			intel_display_power_get(i915, POWER_DOMAIN_INIT);
+	}
+
+	intel_power_domains_verify_state(i915);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+
+static void intel_power_domains_dump_info(struct drm_i915_private *i915)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+	struct i915_power_well *power_well;
+
+	for_each_power_well(i915, power_well) {
+		enum intel_display_power_domain domain;
+
+		drm_dbg(&i915->drm, "%-25s %d\n",
+			intel_power_well_name(power_well), intel_power_well_refcount(power_well));
+
+		for_each_power_domain(domain, intel_power_well_domains(power_well))
+			drm_dbg(&i915->drm, "  %-23s %d\n",
+				intel_display_power_domain_str(domain),
+				power_domains->domain_use_count[domain]);
+	}
+}
+
+/**
+ * intel_power_domains_verify_state - verify the HW/SW state for all power wells
+ * @i915: i915 device instance
+ *
+ * Verify if the reference count of each power well matches its HW enabled
+ * state and the total refcount of the domains it belongs to. This must be
+ * called after modeset HW state sanitization, which is responsible for
+ * acquiring reference counts for any power wells in use and disabling the
+ * ones left on by BIOS but not required by any active output.
+ */
+static void intel_power_domains_verify_state(struct drm_i915_private *i915)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+	struct i915_power_well *power_well;
+	bool dump_domain_info;
+
+	mutex_lock(&power_domains->lock);
+
+	verify_async_put_domains_state(power_domains);
+
+	dump_domain_info = false;
+	for_each_power_well(i915, power_well) {
+		enum intel_display_power_domain domain;
+		int domains_count;
+		bool enabled;
+
+		enabled = intel_power_well_is_enabled(i915, power_well);
+		if ((intel_power_well_refcount(power_well) ||
+		     intel_power_well_is_always_on(power_well)) !=
+		    enabled)
+			drm_err(&i915->drm,
+				"power well %s state mismatch (refcount %d/enabled %d)",
+				intel_power_well_name(power_well),
+				intel_power_well_refcount(power_well), enabled);
+
+		domains_count = 0;
+		for_each_power_domain(domain, intel_power_well_domains(power_well))
+			domains_count += power_domains->domain_use_count[domain];
+
+		if (intel_power_well_refcount(power_well) != domains_count) {
+			drm_err(&i915->drm,
+				"power well %s refcount/domain refcount mismatch "
+				"(refcount %d/domains refcount %d)\n",
+				intel_power_well_name(power_well),
+				intel_power_well_refcount(power_well),
+				domains_count);
+			dump_domain_info = true;
+		}
+	}
+
+	if (dump_domain_info) {
+		static bool dumped;
+
+		if (!dumped) {
+			intel_power_domains_dump_info(i915);
+			dumped = true;
+		}
+	}
+
+	mutex_unlock(&power_domains->lock);
+}
+
+#else
+
+static void intel_power_domains_verify_state(struct drm_i915_private *i915)
+{
+}
+
+#endif
+
+void intel_display_power_suspend_late(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 11 || IS_GEMINILAKE(i915) ||
+	    IS_BROXTON(i915)) {
+		bxt_enable_dc9(i915);
+	} else if (IS_HASWELL(i915) || IS_BROADWELL(i915)) {
+		hsw_enable_pc8(i915);
+	}
+
+	/* Tweaked Wa_14010685332:cnp,icp,jsp,mcc,tgp,adp */
+	if (INTEL_PCH_TYPE(i915) >= PCH_CNP && INTEL_PCH_TYPE(i915) < PCH_DG1)
+		intel_de_rmw(i915, SOUTH_CHICKEN1, SBCLK_RUN_REFCLK_DIS, SBCLK_RUN_REFCLK_DIS);
+}
+
+void intel_display_power_resume_early(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 11 || IS_GEMINILAKE(i915) ||
+	    IS_BROXTON(i915)) {
+		gen9_sanitize_dc_state(i915);
+		bxt_disable_dc9(i915);
+	} else if (IS_HASWELL(i915) || IS_BROADWELL(i915)) {
+		hsw_disable_pc8(i915);
+	}
+
+	/* Tweaked Wa_14010685332:cnp,icp,jsp,mcc,tgp,adp */
+	if (INTEL_PCH_TYPE(i915) >= PCH_CNP && INTEL_PCH_TYPE(i915) < PCH_DG1)
+		intel_de_rmw(i915, SOUTH_CHICKEN1, SBCLK_RUN_REFCLK_DIS, 0);
+}
+
+void intel_display_power_suspend(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 11) {
+		icl_display_core_uninit(i915);
+		bxt_enable_dc9(i915);
+	} else if (IS_GEMINILAKE(i915) || IS_BROXTON(i915)) {
+		bxt_display_core_uninit(i915);
+		bxt_enable_dc9(i915);
+	} else if (IS_HASWELL(i915) || IS_BROADWELL(i915)) {
+		hsw_enable_pc8(i915);
+	}
+}
+
+void intel_display_power_resume(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 11) {
+		bxt_disable_dc9(i915);
+		icl_display_core_init(i915, true);
+		if (intel_dmc_has_payload(i915)) {
+			if (i915->display.dmc.allowed_dc_mask &
+			    DC_STATE_EN_UPTO_DC6)
+				skl_enable_dc6(i915);
+			else if (i915->display.dmc.allowed_dc_mask &
+				 DC_STATE_EN_UPTO_DC5)
+				gen9_enable_dc5(i915);
+		}
+	} else if (IS_GEMINILAKE(i915) || IS_BROXTON(i915)) {
+		bxt_disable_dc9(i915);
+		bxt_display_core_init(i915, true);
+		if (intel_dmc_has_payload(i915) &&
+		    (i915->display.dmc.allowed_dc_mask & DC_STATE_EN_UPTO_DC5))
+			gen9_enable_dc5(i915);
+	} else if (IS_HASWELL(i915) || IS_BROADWELL(i915)) {
+		hsw_disable_pc8(i915);
+	}
+}
+
+void intel_display_power_debug(struct drm_i915_private *i915, struct seq_file *m)
+{
+	struct i915_power_domains *power_domains = &i915->display.power.domains;
+	int i;
+
+	mutex_lock(&power_domains->lock);
+
+	seq_printf(m, "%-25s %s\n", "Power well/domain", "Use count");
+	for (i = 0; i < power_domains->power_well_count; i++) {
+		struct i915_power_well *power_well;
+		enum intel_display_power_domain power_domain;
+
+		power_well = &power_domains->power_wells[i];
+		seq_printf(m, "%-25s %d\n", intel_power_well_name(power_well),
+			   intel_power_well_refcount(power_well));
+
+		for_each_power_domain(power_domain, intel_power_well_domains(power_well))
+			seq_printf(m, "  %-23s %d\n",
+				   intel_display_power_domain_str(power_domain),
+				   power_domains->domain_use_count[power_domain]);
+	}
+
+	mutex_unlock(&power_domains->lock);
+}
+
+struct intel_ddi_port_domains {
+	enum port port_start;
+	enum port port_end;
+	enum aux_ch aux_ch_start;
+	enum aux_ch aux_ch_end;
+
+	enum intel_display_power_domain ddi_lanes;
+	enum intel_display_power_domain ddi_io;
+	enum intel_display_power_domain aux_legacy_usbc;
+	enum intel_display_power_domain aux_tbt;
+};
+
+static const struct intel_ddi_port_domains
+i9xx_port_domains[] = {
+	{
+		.port_start = PORT_A,
+		.port_end = PORT_F,
+		.aux_ch_start = AUX_CH_A,
+		.aux_ch_end = AUX_CH_F,
+
+		.ddi_lanes = POWER_DOMAIN_PORT_DDI_LANES_A,
+		.ddi_io = POWER_DOMAIN_PORT_DDI_IO_A,
+		.aux_legacy_usbc = POWER_DOMAIN_AUX_A,
+		.aux_tbt = POWER_DOMAIN_INVALID,
+	},
+};
+
+static const struct intel_ddi_port_domains
+d11_port_domains[] = {
+	{
+		.port_start = PORT_A,
+		.port_end = PORT_B,
+		.aux_ch_start = AUX_CH_A,
+		.aux_ch_end = AUX_CH_B,
+
+		.ddi_lanes = POWER_DOMAIN_PORT_DDI_LANES_A,
+		.ddi_io = POWER_DOMAIN_PORT_DDI_IO_A,
+		.aux_legacy_usbc = POWER_DOMAIN_AUX_A,
+		.aux_tbt = POWER_DOMAIN_INVALID,
+	}, {
+		.port_start = PORT_C,
+		.port_end = PORT_F,
+		.aux_ch_start = AUX_CH_C,
+		.aux_ch_end = AUX_CH_F,
+
+		.ddi_lanes = POWER_DOMAIN_PORT_DDI_LANES_C,
+		.ddi_io = POWER_DOMAIN_PORT_DDI_IO_C,
+		.aux_legacy_usbc = POWER_DOMAIN_AUX_C,
+		.aux_tbt = POWER_DOMAIN_AUX_TBT1,
+	},
+};
+
+static const struct intel_ddi_port_domains
+d12_port_domains[] = {
+	{
+		.port_start = PORT_A,
+		.port_end = PORT_C,
+		.aux_ch_start = AUX_CH_A,
+		.aux_ch_end = AUX_CH_C,
+
+		.ddi_lanes = POWER_DOMAIN_PORT_DDI_LANES_A,
+		.ddi_io = POWER_DOMAIN_PORT_DDI_IO_A,
+		.aux_legacy_usbc = POWER_DOMAIN_AUX_A,
+		.aux_tbt = POWER_DOMAIN_INVALID,
+	}, {
+		.port_start = PORT_TC1,
+		.port_end = PORT_TC6,
+		.aux_ch_start = AUX_CH_USBC1,
+		.aux_ch_end = AUX_CH_USBC6,
+
+		.ddi_lanes = POWER_DOMAIN_PORT_DDI_LANES_TC1,
+		.ddi_io = POWER_DOMAIN_PORT_DDI_IO_TC1,
+		.aux_legacy_usbc = POWER_DOMAIN_AUX_USBC1,
+		.aux_tbt = POWER_DOMAIN_AUX_TBT1,
+	},
+};
+
+static const struct intel_ddi_port_domains
+d13_port_domains[] = {
+	{
+		.port_start = PORT_A,
+		.port_end = PORT_C,
+		.aux_ch_start = AUX_CH_A,
+		.aux_ch_end = AUX_CH_C,
+
+		.ddi_lanes = POWER_DOMAIN_PORT_DDI_LANES_A,
+		.ddi_io = POWER_DOMAIN_PORT_DDI_IO_A,
+		.aux_legacy_usbc = POWER_DOMAIN_AUX_A,
+		.aux_tbt = POWER_DOMAIN_INVALID,
+	}, {
+		.port_start = PORT_TC1,
+		.port_end = PORT_TC4,
+		.aux_ch_start = AUX_CH_USBC1,
+		.aux_ch_end = AUX_CH_USBC4,
+
+		.ddi_lanes = POWER_DOMAIN_PORT_DDI_LANES_TC1,
+		.ddi_io = POWER_DOMAIN_PORT_DDI_IO_TC1,
+		.aux_legacy_usbc = POWER_DOMAIN_AUX_USBC1,
+		.aux_tbt = POWER_DOMAIN_AUX_TBT1,
+	}, {
+		.port_start = PORT_D_XELPD,
+		.port_end = PORT_E_XELPD,
+		.aux_ch_start = AUX_CH_D_XELPD,
+		.aux_ch_end = AUX_CH_E_XELPD,
+
+		.ddi_lanes = POWER_DOMAIN_PORT_DDI_LANES_D,
+		.ddi_io = POWER_DOMAIN_PORT_DDI_IO_D,
+		.aux_legacy_usbc = POWER_DOMAIN_AUX_D,
+		.aux_tbt = POWER_DOMAIN_INVALID,
+	},
+};
+
+static void
+intel_port_domains_for_platform(struct drm_i915_private *i915,
+				const struct intel_ddi_port_domains **domains,
+				int *domains_size)
+{
+	if (DISPLAY_VER(i915) >= 13) {
+		*domains = d13_port_domains;
+		*domains_size = ARRAY_SIZE(d13_port_domains);
+	} else if (DISPLAY_VER(i915) >= 12) {
+		*domains = d12_port_domains;
+		*domains_size = ARRAY_SIZE(d12_port_domains);
+	} else if (DISPLAY_VER(i915) >= 11) {
+		*domains = d11_port_domains;
+		*domains_size = ARRAY_SIZE(d11_port_domains);
+	} else {
+		*domains = i9xx_port_domains;
+		*domains_size = ARRAY_SIZE(i9xx_port_domains);
+	}
+}
+
+static const struct intel_ddi_port_domains *
+intel_port_domains_for_port(struct drm_i915_private *i915, enum port port)
+{
+	const struct intel_ddi_port_domains *domains;
+	int domains_size;
+	int i;
+
+	intel_port_domains_for_platform(i915, &domains, &domains_size);
+	for (i = 0; i < domains_size; i++)
+		if (port >= domains[i].port_start && port <= domains[i].port_end)
+			return &domains[i];
+
+	return NULL;
+}
+
+enum intel_display_power_domain
+intel_display_power_ddi_io_domain(struct drm_i915_private *i915, enum port port)
+{
+	const struct intel_ddi_port_domains *domains = intel_port_domains_for_port(i915, port);
+
+	if (drm_WARN_ON(&i915->drm, !domains || domains->ddi_io == POWER_DOMAIN_INVALID))
+		return POWER_DOMAIN_PORT_DDI_IO_A;
+
+	return domains->ddi_io + (int)(port - domains->port_start);
+}
+
+enum intel_display_power_domain
+intel_display_power_ddi_lanes_domain(struct drm_i915_private *i915, enum port port)
+{
+	const struct intel_ddi_port_domains *domains = intel_port_domains_for_port(i915, port);
+
+	if (drm_WARN_ON(&i915->drm, !domains || domains->ddi_lanes == POWER_DOMAIN_INVALID))
+		return POWER_DOMAIN_PORT_DDI_LANES_A;
+
+	return domains->ddi_lanes + (int)(port - domains->port_start);
+}
+
+static const struct intel_ddi_port_domains *
+intel_port_domains_for_aux_ch(struct drm_i915_private *i915, enum aux_ch aux_ch)
+{
+	const struct intel_ddi_port_domains *domains;
+	int domains_size;
+	int i;
+
+	intel_port_domains_for_platform(i915, &domains, &domains_size);
+	for (i = 0; i < domains_size; i++)
+		if (aux_ch >= domains[i].aux_ch_start && aux_ch <= domains[i].aux_ch_end)
+			return &domains[i];
+
+	return NULL;
+}
+
+enum intel_display_power_domain
+intel_display_power_legacy_aux_domain(struct drm_i915_private *i915, enum aux_ch aux_ch)
+{
+	const struct intel_ddi_port_domains *domains = intel_port_domains_for_aux_ch(i915, aux_ch);
+
+	if (drm_WARN_ON(&i915->drm, !domains || domains->aux_legacy_usbc == POWER_DOMAIN_INVALID))
+		return POWER_DOMAIN_AUX_A;
+
+	return domains->aux_legacy_usbc + (int)(aux_ch - domains->aux_ch_start);
+}
+
+enum intel_display_power_domain
+intel_display_power_tbt_aux_domain(struct drm_i915_private *i915, enum aux_ch aux_ch)
+{
+	const struct intel_ddi_port_domains *domains = intel_port_domains_for_aux_ch(i915, aux_ch);
+
+	if (drm_WARN_ON(&i915->drm, !domains || domains->aux_tbt == POWER_DOMAIN_INVALID))
+		return POWER_DOMAIN_AUX_TBT1;
+
+	return domains->aux_tbt + (int)(aux_ch - domains->aux_ch_start);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_display_power.h b/drivers/gpu/drm/i915/display/intel_display_power.h
new file mode 100644
index 000000000..7136ea3f2
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_power.h
@@ -0,0 +1,279 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DISPLAY_POWER_H__
+#define __INTEL_DISPLAY_POWER_H__
+
+#include "intel_runtime_pm.h"
+
+enum aux_ch;
+enum dpio_channel;
+enum dpio_phy;
+enum port;
+struct drm_i915_private;
+struct i915_power_well;
+struct intel_encoder;
+
+/*
+ * Keep the pipe, transcoder, port (DDI_LANES,DDI_IO,AUX) domain instances
+ * consecutive, so that the pipe,transcoder,port -> power domain macros
+ * work correctly.
+ */
+enum intel_display_power_domain {
+	POWER_DOMAIN_DISPLAY_CORE,
+	POWER_DOMAIN_PIPE_A,
+	POWER_DOMAIN_PIPE_B,
+	POWER_DOMAIN_PIPE_C,
+	POWER_DOMAIN_PIPE_D,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_A,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_D,
+	POWER_DOMAIN_TRANSCODER_A,
+	POWER_DOMAIN_TRANSCODER_B,
+	POWER_DOMAIN_TRANSCODER_C,
+	POWER_DOMAIN_TRANSCODER_D,
+	POWER_DOMAIN_TRANSCODER_EDP,
+	POWER_DOMAIN_TRANSCODER_DSI_A,
+	POWER_DOMAIN_TRANSCODER_DSI_C,
+
+	/* VDSC/joining for eDP/DSI transcoder (ICL) or pipe A (TGL) */
+	POWER_DOMAIN_TRANSCODER_VDSC_PW2,
+
+	POWER_DOMAIN_PORT_DDI_LANES_A,
+	POWER_DOMAIN_PORT_DDI_LANES_B,
+	POWER_DOMAIN_PORT_DDI_LANES_C,
+	POWER_DOMAIN_PORT_DDI_LANES_D,
+	POWER_DOMAIN_PORT_DDI_LANES_E,
+	POWER_DOMAIN_PORT_DDI_LANES_F,
+
+	POWER_DOMAIN_PORT_DDI_LANES_TC1,
+	POWER_DOMAIN_PORT_DDI_LANES_TC2,
+	POWER_DOMAIN_PORT_DDI_LANES_TC3,
+	POWER_DOMAIN_PORT_DDI_LANES_TC4,
+	POWER_DOMAIN_PORT_DDI_LANES_TC5,
+	POWER_DOMAIN_PORT_DDI_LANES_TC6,
+
+	POWER_DOMAIN_PORT_DDI_IO_A,
+	POWER_DOMAIN_PORT_DDI_IO_B,
+	POWER_DOMAIN_PORT_DDI_IO_C,
+	POWER_DOMAIN_PORT_DDI_IO_D,
+	POWER_DOMAIN_PORT_DDI_IO_E,
+	POWER_DOMAIN_PORT_DDI_IO_F,
+
+	POWER_DOMAIN_PORT_DDI_IO_TC1,
+	POWER_DOMAIN_PORT_DDI_IO_TC2,
+	POWER_DOMAIN_PORT_DDI_IO_TC3,
+	POWER_DOMAIN_PORT_DDI_IO_TC4,
+	POWER_DOMAIN_PORT_DDI_IO_TC5,
+	POWER_DOMAIN_PORT_DDI_IO_TC6,
+
+	POWER_DOMAIN_PORT_DSI,
+	POWER_DOMAIN_PORT_CRT,
+	POWER_DOMAIN_PORT_OTHER,
+	POWER_DOMAIN_VGA,
+	POWER_DOMAIN_AUDIO_MMIO,
+	POWER_DOMAIN_AUDIO_PLAYBACK,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_AUX_C,
+	POWER_DOMAIN_AUX_D,
+	POWER_DOMAIN_AUX_E,
+	POWER_DOMAIN_AUX_F,
+
+	POWER_DOMAIN_AUX_USBC1,
+	POWER_DOMAIN_AUX_USBC2,
+	POWER_DOMAIN_AUX_USBC3,
+	POWER_DOMAIN_AUX_USBC4,
+	POWER_DOMAIN_AUX_USBC5,
+	POWER_DOMAIN_AUX_USBC6,
+
+	POWER_DOMAIN_AUX_IO_A,
+
+	POWER_DOMAIN_AUX_TBT1,
+	POWER_DOMAIN_AUX_TBT2,
+	POWER_DOMAIN_AUX_TBT3,
+	POWER_DOMAIN_AUX_TBT4,
+	POWER_DOMAIN_AUX_TBT5,
+	POWER_DOMAIN_AUX_TBT6,
+
+	POWER_DOMAIN_GMBUS,
+	POWER_DOMAIN_MODESET,
+	POWER_DOMAIN_GT_IRQ,
+	POWER_DOMAIN_DC_OFF,
+	POWER_DOMAIN_TC_COLD_OFF,
+	POWER_DOMAIN_INIT,
+
+	POWER_DOMAIN_NUM,
+	POWER_DOMAIN_INVALID = POWER_DOMAIN_NUM,
+};
+
+#define POWER_DOMAIN_PIPE(pipe) ((pipe) + POWER_DOMAIN_PIPE_A)
+#define POWER_DOMAIN_PIPE_PANEL_FITTER(pipe) \
+		((pipe) + POWER_DOMAIN_PIPE_PANEL_FITTER_A)
+#define POWER_DOMAIN_TRANSCODER(tran) \
+	((tran) == TRANSCODER_EDP ? POWER_DOMAIN_TRANSCODER_EDP : \
+	 (tran) + POWER_DOMAIN_TRANSCODER_A)
+
+struct intel_power_domain_mask {
+	DECLARE_BITMAP(bits, POWER_DOMAIN_NUM);
+};
+
+struct i915_power_domains {
+	/*
+	 * Power wells needed for initialization at driver init and suspend
+	 * time are on. They are kept on until after the first modeset.
+	 */
+	bool initializing;
+	bool display_core_suspended;
+	int power_well_count;
+
+	intel_wakeref_t init_wakeref;
+	intel_wakeref_t disable_wakeref;
+
+	struct mutex lock;
+	int domain_use_count[POWER_DOMAIN_NUM];
+
+	struct delayed_work async_put_work;
+	intel_wakeref_t async_put_wakeref;
+	struct intel_power_domain_mask async_put_domains[2];
+
+	struct i915_power_well *power_wells;
+};
+
+struct intel_display_power_domain_set {
+	struct intel_power_domain_mask mask;
+#ifdef CONFIG_DRM_I915_DEBUG_RUNTIME_PM
+	intel_wakeref_t wakerefs[POWER_DOMAIN_NUM];
+#endif
+};
+
+#define for_each_power_domain(__domain, __mask)				\
+	for ((__domain) = 0; (__domain) < POWER_DOMAIN_NUM; (__domain)++)	\
+		for_each_if(test_bit((__domain), (__mask)->bits))
+
+int intel_power_domains_init(struct drm_i915_private *dev_priv);
+void intel_power_domains_cleanup(struct drm_i915_private *dev_priv);
+void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume);
+void intel_power_domains_driver_remove(struct drm_i915_private *dev_priv);
+void intel_power_domains_enable(struct drm_i915_private *dev_priv);
+void intel_power_domains_disable(struct drm_i915_private *dev_priv);
+void intel_power_domains_suspend(struct drm_i915_private *dev_priv,
+				 enum i915_drm_suspend_mode);
+void intel_power_domains_resume(struct drm_i915_private *dev_priv);
+void intel_power_domains_sanitize_state(struct drm_i915_private *dev_priv);
+
+void intel_display_power_suspend_late(struct drm_i915_private *i915);
+void intel_display_power_resume_early(struct drm_i915_private *i915);
+void intel_display_power_suspend(struct drm_i915_private *i915);
+void intel_display_power_resume(struct drm_i915_private *i915);
+void intel_display_power_set_target_dc_state(struct drm_i915_private *dev_priv,
+					     u32 state);
+
+const char *
+intel_display_power_domain_str(enum intel_display_power_domain domain);
+
+bool intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
+				    enum intel_display_power_domain domain);
+bool __intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
+				      enum intel_display_power_domain domain);
+intel_wakeref_t intel_display_power_get(struct drm_i915_private *dev_priv,
+					enum intel_display_power_domain domain);
+intel_wakeref_t
+intel_display_power_get_if_enabled(struct drm_i915_private *dev_priv,
+				   enum intel_display_power_domain domain);
+void __intel_display_power_put_async(struct drm_i915_private *i915,
+				     enum intel_display_power_domain domain,
+				     intel_wakeref_t wakeref);
+void intel_display_power_flush_work(struct drm_i915_private *i915);
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+void intel_display_power_put(struct drm_i915_private *dev_priv,
+			     enum intel_display_power_domain domain,
+			     intel_wakeref_t wakeref);
+static inline void
+intel_display_power_put_async(struct drm_i915_private *i915,
+			      enum intel_display_power_domain domain,
+			      intel_wakeref_t wakeref)
+{
+	__intel_display_power_put_async(i915, domain, wakeref);
+}
+#else
+void intel_display_power_put_unchecked(struct drm_i915_private *dev_priv,
+				       enum intel_display_power_domain domain);
+
+static inline void
+intel_display_power_put(struct drm_i915_private *i915,
+			enum intel_display_power_domain domain,
+			intel_wakeref_t wakeref)
+{
+	intel_display_power_put_unchecked(i915, domain);
+}
+
+static inline void
+intel_display_power_put_async(struct drm_i915_private *i915,
+			      enum intel_display_power_domain domain,
+			      intel_wakeref_t wakeref)
+{
+	__intel_display_power_put_async(i915, domain, -1);
+}
+#endif
+
+void
+intel_display_power_get_in_set(struct drm_i915_private *i915,
+			       struct intel_display_power_domain_set *power_domain_set,
+			       enum intel_display_power_domain domain);
+
+bool
+intel_display_power_get_in_set_if_enabled(struct drm_i915_private *i915,
+					  struct intel_display_power_domain_set *power_domain_set,
+					  enum intel_display_power_domain domain);
+
+void
+intel_display_power_put_mask_in_set(struct drm_i915_private *i915,
+				    struct intel_display_power_domain_set *power_domain_set,
+				    struct intel_power_domain_mask *mask);
+
+static inline void
+intel_display_power_put_all_in_set(struct drm_i915_private *i915,
+				   struct intel_display_power_domain_set *power_domain_set)
+{
+	intel_display_power_put_mask_in_set(i915, power_domain_set, &power_domain_set->mask);
+}
+
+void intel_display_power_debug(struct drm_i915_private *i915, struct seq_file *m);
+
+enum intel_display_power_domain
+intel_display_power_ddi_lanes_domain(struct drm_i915_private *i915, enum port port);
+enum intel_display_power_domain
+intel_display_power_ddi_io_domain(struct drm_i915_private *i915, enum port port);
+enum intel_display_power_domain
+intel_display_power_legacy_aux_domain(struct drm_i915_private *i915, enum aux_ch aux_ch);
+enum intel_display_power_domain
+intel_display_power_tbt_aux_domain(struct drm_i915_private *i915, enum aux_ch aux_ch);
+
+/*
+ * FIXME: We should probably switch this to a 0-based scheme to be consistent
+ * with how we now name/number DBUF_CTL instances.
+ */
+enum dbuf_slice {
+	DBUF_S1,
+	DBUF_S2,
+	DBUF_S3,
+	DBUF_S4,
+	I915_MAX_DBUF_SLICES
+};
+
+void gen9_dbuf_slices_update(struct drm_i915_private *dev_priv,
+			     u8 req_slices);
+
+#define with_intel_display_power(i915, domain, wf) \
+	for ((wf) = intel_display_power_get((i915), (domain)); (wf); \
+	     intel_display_power_put_async((i915), (domain), (wf)), (wf) = 0)
+
+#define with_intel_display_power_if_enabled(i915, domain, wf) \
+	for ((wf) = intel_display_power_get_if_enabled((i915), (domain)); (wf); \
+	     intel_display_power_put_async((i915), (domain), (wf)), (wf) = 0)
+
+#endif /* __INTEL_DISPLAY_POWER_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_display_power_map.c b/drivers/gpu/drm/i915/display/intel_display_power_map.c
new file mode 100644
index 000000000..dc04afc6c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_power_map.c
@@ -0,0 +1,1614 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+
+#include "vlv_sideband_reg.h"
+
+#include "intel_display_power_map.h"
+#include "intel_display_power_well.h"
+
+#define __LIST_INLINE_ELEMS(__elem_type, ...) \
+	((__elem_type[]) { __VA_ARGS__ })
+
+#define __LIST(__elems) { \
+	.list = __elems, \
+	.count = ARRAY_SIZE(__elems), \
+}
+
+#define I915_PW_DOMAINS(...) \
+	(const struct i915_power_domain_list) \
+		__LIST(__LIST_INLINE_ELEMS(const enum intel_display_power_domain, __VA_ARGS__))
+
+#define I915_DECL_PW_DOMAINS(__name, ...) \
+	static const struct i915_power_domain_list __name = I915_PW_DOMAINS(__VA_ARGS__)
+
+/* Zero-length list assigns all power domains, a NULL list assigns none. */
+#define I915_PW_DOMAINS_NONE	NULL
+#define I915_PW_DOMAINS_ALL	/* zero-length list */
+
+#define I915_PW_INSTANCES(...) \
+	(const struct i915_power_well_instance_list) \
+		__LIST(__LIST_INLINE_ELEMS(const struct i915_power_well_instance, __VA_ARGS__))
+
+#define I915_PW(_name, _domain_list, ...) \
+	{ .name = _name, .domain_list = _domain_list, ## __VA_ARGS__ }
+
+
+struct i915_power_well_desc_list {
+	const struct i915_power_well_desc *list;
+	u8 count;
+};
+
+#define I915_PW_DESCRIPTORS(x) __LIST(x)
+
+
+I915_DECL_PW_DOMAINS(i9xx_pwdoms_always_on, I915_PW_DOMAINS_ALL);
+
+static const struct i915_power_well_desc i9xx_power_wells_always_on[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("always-on", &i9xx_pwdoms_always_on),
+		),
+		.ops = &i9xx_always_on_power_well_ops,
+		.always_on = true,
+	},
+};
+
+static const struct i915_power_well_desc_list i9xx_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+};
+
+I915_DECL_PW_DOMAINS(i830_pwdoms_pipes,
+	POWER_DOMAIN_PIPE_A,
+	POWER_DOMAIN_PIPE_B,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_A,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B,
+	POWER_DOMAIN_TRANSCODER_A,
+	POWER_DOMAIN_TRANSCODER_B,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc i830_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("pipes", &i830_pwdoms_pipes),
+		),
+		.ops = &i830_pipes_power_well_ops,
+	},
+};
+
+static const struct i915_power_well_desc_list i830_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(i830_power_wells_main),
+};
+
+I915_DECL_PW_DOMAINS(hsw_pwdoms_display,
+	POWER_DOMAIN_PIPE_B,
+	POWER_DOMAIN_PIPE_C,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_A,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C,
+	POWER_DOMAIN_TRANSCODER_A,
+	POWER_DOMAIN_TRANSCODER_B,
+	POWER_DOMAIN_TRANSCODER_C,
+	POWER_DOMAIN_PORT_DDI_LANES_B,
+	POWER_DOMAIN_PORT_DDI_LANES_C,
+	POWER_DOMAIN_PORT_DDI_LANES_D,
+	POWER_DOMAIN_PORT_CRT, /* DDI E */
+	POWER_DOMAIN_VGA,
+	POWER_DOMAIN_AUDIO_MMIO,
+	POWER_DOMAIN_AUDIO_PLAYBACK,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc hsw_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("display", &hsw_pwdoms_display,
+				.hsw.idx = HSW_PW_CTL_IDX_GLOBAL,
+				.id = HSW_DISP_PW_GLOBAL),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_vga = true,
+	},
+};
+
+static const struct i915_power_well_desc_list hsw_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(hsw_power_wells_main),
+};
+
+I915_DECL_PW_DOMAINS(bdw_pwdoms_display,
+	POWER_DOMAIN_PIPE_B,
+	POWER_DOMAIN_PIPE_C,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C,
+	POWER_DOMAIN_TRANSCODER_A,
+	POWER_DOMAIN_TRANSCODER_B,
+	POWER_DOMAIN_TRANSCODER_C,
+	POWER_DOMAIN_PORT_DDI_LANES_B,
+	POWER_DOMAIN_PORT_DDI_LANES_C,
+	POWER_DOMAIN_PORT_DDI_LANES_D,
+	POWER_DOMAIN_PORT_CRT, /* DDI E */
+	POWER_DOMAIN_VGA,
+	POWER_DOMAIN_AUDIO_MMIO,
+	POWER_DOMAIN_AUDIO_PLAYBACK,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc bdw_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("display", &bdw_pwdoms_display,
+				.hsw.idx = HSW_PW_CTL_IDX_GLOBAL,
+				.id = HSW_DISP_PW_GLOBAL),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_vga = true,
+		.irq_pipe_mask = BIT(PIPE_B) | BIT(PIPE_C),
+	},
+};
+
+static const struct i915_power_well_desc_list bdw_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(bdw_power_wells_main),
+};
+
+I915_DECL_PW_DOMAINS(vlv_pwdoms_display,
+	POWER_DOMAIN_DISPLAY_CORE,
+	POWER_DOMAIN_PIPE_A,
+	POWER_DOMAIN_PIPE_B,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_A,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B,
+	POWER_DOMAIN_TRANSCODER_A,
+	POWER_DOMAIN_TRANSCODER_B,
+	POWER_DOMAIN_PORT_DDI_LANES_B,
+	POWER_DOMAIN_PORT_DDI_LANES_C,
+	POWER_DOMAIN_PORT_DSI,
+	POWER_DOMAIN_PORT_CRT,
+	POWER_DOMAIN_VGA,
+	POWER_DOMAIN_AUDIO_MMIO,
+	POWER_DOMAIN_AUDIO_PLAYBACK,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_AUX_C,
+	POWER_DOMAIN_GMBUS,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(vlv_pwdoms_dpio_cmn_bc,
+	POWER_DOMAIN_PORT_DDI_LANES_B,
+	POWER_DOMAIN_PORT_DDI_LANES_C,
+	POWER_DOMAIN_PORT_CRT,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_AUX_C,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(vlv_pwdoms_dpio_tx_bc_lanes,
+	POWER_DOMAIN_PORT_DDI_LANES_B,
+	POWER_DOMAIN_PORT_DDI_LANES_C,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_AUX_C,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc vlv_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("display", &vlv_pwdoms_display,
+				.vlv.idx = PUNIT_PWGT_IDX_DISP2D,
+				.id = VLV_DISP_PW_DISP2D),
+		),
+		.ops = &vlv_display_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("dpio-tx-b-01", &vlv_pwdoms_dpio_tx_bc_lanes,
+				.vlv.idx = PUNIT_PWGT_IDX_DPIO_TX_B_LANES_01),
+			I915_PW("dpio-tx-b-23", &vlv_pwdoms_dpio_tx_bc_lanes,
+				.vlv.idx = PUNIT_PWGT_IDX_DPIO_TX_B_LANES_23),
+			I915_PW("dpio-tx-c-01", &vlv_pwdoms_dpio_tx_bc_lanes,
+				.vlv.idx = PUNIT_PWGT_IDX_DPIO_TX_C_LANES_01),
+			I915_PW("dpio-tx-c-23", &vlv_pwdoms_dpio_tx_bc_lanes,
+				.vlv.idx = PUNIT_PWGT_IDX_DPIO_TX_C_LANES_23),
+		),
+		.ops = &vlv_dpio_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("dpio-common", &vlv_pwdoms_dpio_cmn_bc,
+				.vlv.idx = PUNIT_PWGT_IDX_DPIO_CMN_BC,
+				.id = VLV_DISP_PW_DPIO_CMN_BC),
+		),
+		.ops = &vlv_dpio_cmn_power_well_ops,
+	},
+};
+
+static const struct i915_power_well_desc_list vlv_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(vlv_power_wells_main),
+};
+
+I915_DECL_PW_DOMAINS(chv_pwdoms_display,
+	POWER_DOMAIN_DISPLAY_CORE,
+	POWER_DOMAIN_PIPE_A,
+	POWER_DOMAIN_PIPE_B,
+	POWER_DOMAIN_PIPE_C,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_A,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C,
+	POWER_DOMAIN_TRANSCODER_A,
+	POWER_DOMAIN_TRANSCODER_B,
+	POWER_DOMAIN_TRANSCODER_C,
+	POWER_DOMAIN_PORT_DDI_LANES_B,
+	POWER_DOMAIN_PORT_DDI_LANES_C,
+	POWER_DOMAIN_PORT_DDI_LANES_D,
+	POWER_DOMAIN_PORT_DSI,
+	POWER_DOMAIN_VGA,
+	POWER_DOMAIN_AUDIO_MMIO,
+	POWER_DOMAIN_AUDIO_PLAYBACK,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_AUX_C,
+	POWER_DOMAIN_AUX_D,
+	POWER_DOMAIN_GMBUS,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(chv_pwdoms_dpio_cmn_bc,
+	POWER_DOMAIN_PORT_DDI_LANES_B,
+	POWER_DOMAIN_PORT_DDI_LANES_C,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_AUX_C,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(chv_pwdoms_dpio_cmn_d,
+	POWER_DOMAIN_PORT_DDI_LANES_D,
+	POWER_DOMAIN_AUX_D,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc chv_power_wells_main[] = {
+	{
+		/*
+		 * Pipe A power well is the new disp2d well. Pipe B and C
+		 * power wells don't actually exist. Pipe A power well is
+		 * required for any pipe to work.
+		 */
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("display", &chv_pwdoms_display),
+		),
+		.ops = &chv_pipe_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("dpio-common-bc", &chv_pwdoms_dpio_cmn_bc,
+				.vlv.idx = PUNIT_PWGT_IDX_DPIO_CMN_BC,
+				.id = VLV_DISP_PW_DPIO_CMN_BC),
+			I915_PW("dpio-common-d", &chv_pwdoms_dpio_cmn_d,
+				.vlv.idx = PUNIT_PWGT_IDX_DPIO_CMN_D,
+				.id = CHV_DISP_PW_DPIO_CMN_D),
+		),
+		.ops = &chv_dpio_cmn_power_well_ops,
+	},
+};
+
+static const struct i915_power_well_desc_list chv_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(chv_power_wells_main),
+};
+
+#define SKL_PW_2_POWER_DOMAINS \
+	POWER_DOMAIN_PIPE_B, \
+	POWER_DOMAIN_PIPE_C, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C, \
+	POWER_DOMAIN_TRANSCODER_A, \
+	POWER_DOMAIN_TRANSCODER_B, \
+	POWER_DOMAIN_TRANSCODER_C, \
+	POWER_DOMAIN_PORT_DDI_LANES_B, \
+	POWER_DOMAIN_PORT_DDI_LANES_C, \
+	POWER_DOMAIN_PORT_DDI_LANES_D, \
+	POWER_DOMAIN_PORT_DDI_LANES_E, \
+	POWER_DOMAIN_VGA, \
+	POWER_DOMAIN_AUDIO_MMIO, \
+	POWER_DOMAIN_AUDIO_PLAYBACK, \
+	POWER_DOMAIN_AUX_B, \
+	POWER_DOMAIN_AUX_C, \
+	POWER_DOMAIN_AUX_D
+
+I915_DECL_PW_DOMAINS(skl_pwdoms_pw_2,
+	SKL_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(skl_pwdoms_dc_off,
+	SKL_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_MODESET,
+	POWER_DOMAIN_GT_IRQ,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(skl_pwdoms_ddi_io_a_e,
+	POWER_DOMAIN_PORT_DDI_IO_A,
+	POWER_DOMAIN_PORT_DDI_IO_E,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(skl_pwdoms_ddi_io_b,
+	POWER_DOMAIN_PORT_DDI_IO_B,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(skl_pwdoms_ddi_io_c,
+	POWER_DOMAIN_PORT_DDI_IO_C,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(skl_pwdoms_ddi_io_d,
+	POWER_DOMAIN_PORT_DDI_IO_D,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc skl_power_wells_pw_1[] = {
+	{
+		/* Handled by the DMC firmware */
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_1", I915_PW_DOMAINS_NONE,
+				.hsw.idx = SKL_PW_CTL_IDX_PW_1,
+				.id = SKL_DISP_PW_1),
+		),
+		.ops = &hsw_power_well_ops,
+		.always_on = true,
+		.has_fuses = true,
+	},
+};
+
+static const struct i915_power_well_desc skl_power_wells_main[] = {
+	{
+		/* Handled by the DMC firmware */
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("MISC_IO", I915_PW_DOMAINS_NONE,
+				.hsw.idx = SKL_PW_CTL_IDX_MISC_IO,
+				.id = SKL_DISP_PW_MISC_IO),
+		),
+		.ops = &hsw_power_well_ops,
+		.always_on = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DC_off", &skl_pwdoms_dc_off,
+				.id = SKL_DISP_DC_OFF),
+		),
+		.ops = &gen9_dc_off_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_2", &skl_pwdoms_pw_2,
+				.hsw.idx = SKL_PW_CTL_IDX_PW_2,
+				.id = SKL_DISP_PW_2),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_vga = true,
+		.irq_pipe_mask = BIT(PIPE_B) | BIT(PIPE_C),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DDI_IO_A_E", &skl_pwdoms_ddi_io_a_e, .hsw.idx = SKL_PW_CTL_IDX_DDI_A_E),
+			I915_PW("DDI_IO_B", &skl_pwdoms_ddi_io_b, .hsw.idx = SKL_PW_CTL_IDX_DDI_B),
+			I915_PW("DDI_IO_C", &skl_pwdoms_ddi_io_c, .hsw.idx = SKL_PW_CTL_IDX_DDI_C),
+			I915_PW("DDI_IO_D", &skl_pwdoms_ddi_io_d, .hsw.idx = SKL_PW_CTL_IDX_DDI_D),
+		),
+		.ops = &hsw_power_well_ops,
+	},
+};
+
+static const struct i915_power_well_desc_list skl_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(skl_power_wells_pw_1),
+	I915_PW_DESCRIPTORS(skl_power_wells_main),
+};
+
+#define BXT_PW_2_POWER_DOMAINS \
+	POWER_DOMAIN_PIPE_B, \
+	POWER_DOMAIN_PIPE_C, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C, \
+	POWER_DOMAIN_TRANSCODER_A, \
+	POWER_DOMAIN_TRANSCODER_B, \
+	POWER_DOMAIN_TRANSCODER_C, \
+	POWER_DOMAIN_PORT_DDI_LANES_B, \
+	POWER_DOMAIN_PORT_DDI_LANES_C, \
+	POWER_DOMAIN_VGA, \
+	POWER_DOMAIN_AUDIO_MMIO, \
+	POWER_DOMAIN_AUDIO_PLAYBACK, \
+	POWER_DOMAIN_AUX_B, \
+	POWER_DOMAIN_AUX_C
+
+I915_DECL_PW_DOMAINS(bxt_pwdoms_pw_2,
+	BXT_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(bxt_pwdoms_dc_off,
+	BXT_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_GMBUS,
+	POWER_DOMAIN_MODESET,
+	POWER_DOMAIN_GT_IRQ,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(bxt_pwdoms_dpio_cmn_a,
+	POWER_DOMAIN_PORT_DDI_LANES_A,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(bxt_pwdoms_dpio_cmn_bc,
+	POWER_DOMAIN_PORT_DDI_LANES_B,
+	POWER_DOMAIN_PORT_DDI_LANES_C,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_AUX_C,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc bxt_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DC_off", &bxt_pwdoms_dc_off,
+				.id = SKL_DISP_DC_OFF),
+		),
+		.ops = &gen9_dc_off_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_2", &bxt_pwdoms_pw_2,
+				.hsw.idx = SKL_PW_CTL_IDX_PW_2,
+				.id = SKL_DISP_PW_2),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_vga = true,
+		.irq_pipe_mask = BIT(PIPE_B) | BIT(PIPE_C),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("dpio-common-a", &bxt_pwdoms_dpio_cmn_a,
+				.bxt.phy = DPIO_PHY1,
+				.id = BXT_DISP_PW_DPIO_CMN_A),
+			I915_PW("dpio-common-bc", &bxt_pwdoms_dpio_cmn_bc,
+				.bxt.phy = DPIO_PHY0,
+				.id = VLV_DISP_PW_DPIO_CMN_BC),
+		),
+		.ops = &bxt_dpio_cmn_power_well_ops,
+	},
+};
+
+static const struct i915_power_well_desc_list bxt_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(skl_power_wells_pw_1),
+	I915_PW_DESCRIPTORS(bxt_power_wells_main),
+};
+
+#define GLK_PW_2_POWER_DOMAINS \
+	POWER_DOMAIN_PIPE_B, \
+	POWER_DOMAIN_PIPE_C, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C, \
+	POWER_DOMAIN_TRANSCODER_A, \
+	POWER_DOMAIN_TRANSCODER_B, \
+	POWER_DOMAIN_TRANSCODER_C, \
+	POWER_DOMAIN_PORT_DDI_LANES_B, \
+	POWER_DOMAIN_PORT_DDI_LANES_C, \
+	POWER_DOMAIN_VGA, \
+	POWER_DOMAIN_AUDIO_MMIO, \
+	POWER_DOMAIN_AUDIO_PLAYBACK, \
+	POWER_DOMAIN_AUX_B, \
+	POWER_DOMAIN_AUX_C
+
+I915_DECL_PW_DOMAINS(glk_pwdoms_pw_2,
+	GLK_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(glk_pwdoms_dc_off,
+	GLK_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_GMBUS,
+	POWER_DOMAIN_MODESET,
+	POWER_DOMAIN_GT_IRQ,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(glk_pwdoms_ddi_io_a,	POWER_DOMAIN_PORT_DDI_IO_A);
+I915_DECL_PW_DOMAINS(glk_pwdoms_ddi_io_b,	POWER_DOMAIN_PORT_DDI_IO_B);
+I915_DECL_PW_DOMAINS(glk_pwdoms_ddi_io_c,	POWER_DOMAIN_PORT_DDI_IO_C);
+
+I915_DECL_PW_DOMAINS(glk_pwdoms_dpio_cmn_a,
+	POWER_DOMAIN_PORT_DDI_LANES_A,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(glk_pwdoms_dpio_cmn_b,
+	POWER_DOMAIN_PORT_DDI_LANES_B,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(glk_pwdoms_dpio_cmn_c,
+	POWER_DOMAIN_PORT_DDI_LANES_C,
+	POWER_DOMAIN_AUX_C,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(glk_pwdoms_aux_a,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_AUX_IO_A,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(glk_pwdoms_aux_b,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(glk_pwdoms_aux_c,
+	POWER_DOMAIN_AUX_C,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc glk_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DC_off", &glk_pwdoms_dc_off,
+				.id = SKL_DISP_DC_OFF),
+		),
+		.ops = &gen9_dc_off_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_2", &glk_pwdoms_pw_2,
+				.hsw.idx = SKL_PW_CTL_IDX_PW_2,
+				.id = SKL_DISP_PW_2),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_vga = true,
+		.irq_pipe_mask = BIT(PIPE_B) | BIT(PIPE_C),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("dpio-common-a", &glk_pwdoms_dpio_cmn_a,
+				.bxt.phy = DPIO_PHY1,
+				.id = BXT_DISP_PW_DPIO_CMN_A),
+			I915_PW("dpio-common-b", &glk_pwdoms_dpio_cmn_b,
+				.bxt.phy = DPIO_PHY0,
+				.id = VLV_DISP_PW_DPIO_CMN_BC),
+			I915_PW("dpio-common-c", &glk_pwdoms_dpio_cmn_c,
+				.bxt.phy = DPIO_PHY2,
+				.id = GLK_DISP_PW_DPIO_CMN_C),
+		),
+		.ops = &bxt_dpio_cmn_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("AUX_A", &glk_pwdoms_aux_a, .hsw.idx = GLK_PW_CTL_IDX_AUX_A),
+			I915_PW("AUX_B", &glk_pwdoms_aux_b, .hsw.idx = GLK_PW_CTL_IDX_AUX_B),
+			I915_PW("AUX_C", &glk_pwdoms_aux_c, .hsw.idx = GLK_PW_CTL_IDX_AUX_C),
+			I915_PW("DDI_IO_A", &glk_pwdoms_ddi_io_a, .hsw.idx = GLK_PW_CTL_IDX_DDI_A),
+			I915_PW("DDI_IO_B", &glk_pwdoms_ddi_io_b, .hsw.idx = SKL_PW_CTL_IDX_DDI_B),
+			I915_PW("DDI_IO_C", &glk_pwdoms_ddi_io_c, .hsw.idx = SKL_PW_CTL_IDX_DDI_C),
+		),
+		.ops = &hsw_power_well_ops,
+	},
+};
+
+static const struct i915_power_well_desc_list glk_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(skl_power_wells_pw_1),
+	I915_PW_DESCRIPTORS(glk_power_wells_main),
+};
+
+/*
+ * ICL PW_0/PG_0 domains (HW/DMC control):
+ * - PCI
+ * - clocks except port PLL
+ * - central power except FBC
+ * - shared functions except pipe interrupts, pipe MBUS, DBUF registers
+ * ICL PW_1/PG_1 domains (HW/DMC control):
+ * - DBUF function
+ * - PIPE_A and its planes, except VGA
+ * - transcoder EDP + PSR
+ * - transcoder DSI
+ * - DDI_A
+ * - FBC
+ */
+#define ICL_PW_4_POWER_DOMAINS \
+	POWER_DOMAIN_PIPE_C, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C
+
+I915_DECL_PW_DOMAINS(icl_pwdoms_pw_4,
+	ICL_PW_4_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+	/* VDSC/joining */
+
+#define ICL_PW_3_POWER_DOMAINS \
+	ICL_PW_4_POWER_DOMAINS, \
+	POWER_DOMAIN_PIPE_B, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B, \
+	POWER_DOMAIN_TRANSCODER_A, \
+	POWER_DOMAIN_TRANSCODER_B, \
+	POWER_DOMAIN_TRANSCODER_C, \
+	POWER_DOMAIN_PORT_DDI_LANES_B, \
+	POWER_DOMAIN_PORT_DDI_LANES_C, \
+	POWER_DOMAIN_PORT_DDI_LANES_D, \
+	POWER_DOMAIN_PORT_DDI_LANES_E, \
+	POWER_DOMAIN_PORT_DDI_LANES_F, \
+	POWER_DOMAIN_VGA, \
+	POWER_DOMAIN_AUDIO_MMIO, \
+	POWER_DOMAIN_AUDIO_PLAYBACK, \
+	POWER_DOMAIN_AUX_B, \
+	POWER_DOMAIN_AUX_C, \
+	POWER_DOMAIN_AUX_D, \
+	POWER_DOMAIN_AUX_E, \
+	POWER_DOMAIN_AUX_F, \
+	POWER_DOMAIN_AUX_TBT1, \
+	POWER_DOMAIN_AUX_TBT2, \
+	POWER_DOMAIN_AUX_TBT3, \
+	POWER_DOMAIN_AUX_TBT4
+
+I915_DECL_PW_DOMAINS(icl_pwdoms_pw_3,
+	ICL_PW_3_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+	/*
+	 * - transcoder WD
+	 * - KVMR (HW control)
+	 */
+
+#define ICL_PW_2_POWER_DOMAINS \
+	ICL_PW_3_POWER_DOMAINS, \
+	POWER_DOMAIN_TRANSCODER_VDSC_PW2
+
+I915_DECL_PW_DOMAINS(icl_pwdoms_pw_2,
+	ICL_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+	/*
+	 * - KVMR (HW control)
+	 */
+
+I915_DECL_PW_DOMAINS(icl_pwdoms_dc_off,
+	ICL_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_MODESET,
+	POWER_DOMAIN_DC_OFF,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(icl_pwdoms_ddi_io_d,	POWER_DOMAIN_PORT_DDI_IO_D);
+I915_DECL_PW_DOMAINS(icl_pwdoms_ddi_io_e,	POWER_DOMAIN_PORT_DDI_IO_E);
+I915_DECL_PW_DOMAINS(icl_pwdoms_ddi_io_f,	POWER_DOMAIN_PORT_DDI_IO_F);
+
+I915_DECL_PW_DOMAINS(icl_pwdoms_aux_a,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_AUX_IO_A);
+I915_DECL_PW_DOMAINS(icl_pwdoms_aux_b,		POWER_DOMAIN_AUX_B);
+I915_DECL_PW_DOMAINS(icl_pwdoms_aux_c,		POWER_DOMAIN_AUX_C);
+I915_DECL_PW_DOMAINS(icl_pwdoms_aux_d,		POWER_DOMAIN_AUX_D);
+I915_DECL_PW_DOMAINS(icl_pwdoms_aux_e,		POWER_DOMAIN_AUX_E);
+I915_DECL_PW_DOMAINS(icl_pwdoms_aux_f,		POWER_DOMAIN_AUX_F);
+I915_DECL_PW_DOMAINS(icl_pwdoms_aux_tbt1,	POWER_DOMAIN_AUX_TBT1);
+I915_DECL_PW_DOMAINS(icl_pwdoms_aux_tbt2,	POWER_DOMAIN_AUX_TBT2);
+I915_DECL_PW_DOMAINS(icl_pwdoms_aux_tbt3,	POWER_DOMAIN_AUX_TBT3);
+I915_DECL_PW_DOMAINS(icl_pwdoms_aux_tbt4,	POWER_DOMAIN_AUX_TBT4);
+
+static const struct i915_power_well_desc icl_power_wells_pw_1[] = {
+	{
+		/* Handled by the DMC firmware */
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_1", I915_PW_DOMAINS_NONE,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_1,
+				.id = SKL_DISP_PW_1),
+		),
+		.ops = &hsw_power_well_ops,
+		.always_on = true,
+		.has_fuses = true,
+	},
+};
+
+static const struct i915_power_well_desc icl_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DC_off", &icl_pwdoms_dc_off,
+				.id = SKL_DISP_DC_OFF),
+		),
+		.ops = &gen9_dc_off_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_2", &icl_pwdoms_pw_2,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_2,
+				.id = SKL_DISP_PW_2),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_3", &icl_pwdoms_pw_3,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_3,
+				.id = ICL_DISP_PW_3),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_vga = true,
+		.irq_pipe_mask = BIT(PIPE_B),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DDI_IO_A", &glk_pwdoms_ddi_io_a, .hsw.idx = ICL_PW_CTL_IDX_DDI_A),
+			I915_PW("DDI_IO_B", &glk_pwdoms_ddi_io_b, .hsw.idx = ICL_PW_CTL_IDX_DDI_B),
+			I915_PW("DDI_IO_C", &glk_pwdoms_ddi_io_c, .hsw.idx = ICL_PW_CTL_IDX_DDI_C),
+			I915_PW("DDI_IO_D", &icl_pwdoms_ddi_io_d, .hsw.idx = ICL_PW_CTL_IDX_DDI_D),
+			I915_PW("DDI_IO_E", &icl_pwdoms_ddi_io_e, .hsw.idx = ICL_PW_CTL_IDX_DDI_E),
+			I915_PW("DDI_IO_F", &icl_pwdoms_ddi_io_f, .hsw.idx = ICL_PW_CTL_IDX_DDI_F),
+		),
+		.ops = &icl_ddi_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("AUX_A", &icl_pwdoms_aux_a, .hsw.idx = ICL_PW_CTL_IDX_AUX_A),
+			I915_PW("AUX_B", &icl_pwdoms_aux_b, .hsw.idx = ICL_PW_CTL_IDX_AUX_B),
+			I915_PW("AUX_C", &icl_pwdoms_aux_c, .hsw.idx = ICL_PW_CTL_IDX_AUX_C),
+			I915_PW("AUX_D", &icl_pwdoms_aux_d, .hsw.idx = ICL_PW_CTL_IDX_AUX_D),
+			I915_PW("AUX_E", &icl_pwdoms_aux_e, .hsw.idx = ICL_PW_CTL_IDX_AUX_E),
+			I915_PW("AUX_F", &icl_pwdoms_aux_f, .hsw.idx = ICL_PW_CTL_IDX_AUX_F),
+		),
+		.ops = &icl_aux_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("AUX_TBT1", &icl_pwdoms_aux_tbt1, .hsw.idx = ICL_PW_CTL_IDX_AUX_TBT1),
+			I915_PW("AUX_TBT2", &icl_pwdoms_aux_tbt2, .hsw.idx = ICL_PW_CTL_IDX_AUX_TBT2),
+			I915_PW("AUX_TBT3", &icl_pwdoms_aux_tbt3, .hsw.idx = ICL_PW_CTL_IDX_AUX_TBT3),
+			I915_PW("AUX_TBT4", &icl_pwdoms_aux_tbt4, .hsw.idx = ICL_PW_CTL_IDX_AUX_TBT4),
+		),
+		.ops = &icl_aux_power_well_ops,
+		.is_tc_tbt = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_4", &icl_pwdoms_pw_4,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_4),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_C),
+		.has_fuses = true,
+	},
+};
+
+static const struct i915_power_well_desc_list icl_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(icl_power_wells_pw_1),
+	I915_PW_DESCRIPTORS(icl_power_wells_main),
+};
+
+#define TGL_PW_5_POWER_DOMAINS \
+	POWER_DOMAIN_PIPE_D, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_D, \
+	POWER_DOMAIN_TRANSCODER_D
+
+I915_DECL_PW_DOMAINS(tgl_pwdoms_pw_5,
+	TGL_PW_5_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+#define TGL_PW_4_POWER_DOMAINS \
+	TGL_PW_5_POWER_DOMAINS, \
+	POWER_DOMAIN_PIPE_C, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C, \
+	POWER_DOMAIN_TRANSCODER_C
+
+I915_DECL_PW_DOMAINS(tgl_pwdoms_pw_4,
+	TGL_PW_4_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+#define TGL_PW_3_POWER_DOMAINS \
+	TGL_PW_4_POWER_DOMAINS, \
+	POWER_DOMAIN_PIPE_B, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B, \
+	POWER_DOMAIN_TRANSCODER_B, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC1, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC2, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC3, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC4, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC5, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC6, \
+	POWER_DOMAIN_VGA, \
+	POWER_DOMAIN_AUDIO_MMIO, \
+	POWER_DOMAIN_AUDIO_PLAYBACK, \
+	POWER_DOMAIN_AUX_USBC1, \
+	POWER_DOMAIN_AUX_USBC2, \
+	POWER_DOMAIN_AUX_USBC3, \
+	POWER_DOMAIN_AUX_USBC4, \
+	POWER_DOMAIN_AUX_USBC5, \
+	POWER_DOMAIN_AUX_USBC6, \
+	POWER_DOMAIN_AUX_TBT1, \
+	POWER_DOMAIN_AUX_TBT2, \
+	POWER_DOMAIN_AUX_TBT3, \
+	POWER_DOMAIN_AUX_TBT4, \
+	POWER_DOMAIN_AUX_TBT5, \
+	POWER_DOMAIN_AUX_TBT6
+
+I915_DECL_PW_DOMAINS(tgl_pwdoms_pw_3,
+	TGL_PW_3_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(tgl_pwdoms_pw_2,
+	TGL_PW_3_POWER_DOMAINS,
+	POWER_DOMAIN_TRANSCODER_VDSC_PW2,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(tgl_pwdoms_dc_off,
+	TGL_PW_3_POWER_DOMAINS,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_AUX_C,
+	POWER_DOMAIN_MODESET,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(tgl_pwdoms_ddi_io_tc1,	POWER_DOMAIN_PORT_DDI_IO_TC1);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_ddi_io_tc2,	POWER_DOMAIN_PORT_DDI_IO_TC2);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_ddi_io_tc3,	POWER_DOMAIN_PORT_DDI_IO_TC3);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_ddi_io_tc4,	POWER_DOMAIN_PORT_DDI_IO_TC4);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_ddi_io_tc5,	POWER_DOMAIN_PORT_DDI_IO_TC5);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_ddi_io_tc6,	POWER_DOMAIN_PORT_DDI_IO_TC6);
+
+I915_DECL_PW_DOMAINS(tgl_pwdoms_aux_usbc1,	POWER_DOMAIN_AUX_USBC1);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_aux_usbc2,	POWER_DOMAIN_AUX_USBC2);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_aux_usbc3,	POWER_DOMAIN_AUX_USBC3);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_aux_usbc4,	POWER_DOMAIN_AUX_USBC4);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_aux_usbc5,	POWER_DOMAIN_AUX_USBC5);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_aux_usbc6,	POWER_DOMAIN_AUX_USBC6);
+
+I915_DECL_PW_DOMAINS(tgl_pwdoms_aux_tbt5,	POWER_DOMAIN_AUX_TBT5);
+I915_DECL_PW_DOMAINS(tgl_pwdoms_aux_tbt6,	POWER_DOMAIN_AUX_TBT6);
+
+I915_DECL_PW_DOMAINS(tgl_pwdoms_tc_cold_off,
+	POWER_DOMAIN_AUX_USBC1,
+	POWER_DOMAIN_AUX_USBC2,
+	POWER_DOMAIN_AUX_USBC3,
+	POWER_DOMAIN_AUX_USBC4,
+	POWER_DOMAIN_AUX_USBC5,
+	POWER_DOMAIN_AUX_USBC6,
+	POWER_DOMAIN_AUX_TBT1,
+	POWER_DOMAIN_AUX_TBT2,
+	POWER_DOMAIN_AUX_TBT3,
+	POWER_DOMAIN_AUX_TBT4,
+	POWER_DOMAIN_AUX_TBT5,
+	POWER_DOMAIN_AUX_TBT6,
+	POWER_DOMAIN_TC_COLD_OFF);
+
+static const struct i915_power_well_desc tgl_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DC_off", &tgl_pwdoms_dc_off,
+				.id = SKL_DISP_DC_OFF),
+		),
+		.ops = &gen9_dc_off_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_2", &tgl_pwdoms_pw_2,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_2,
+				.id = SKL_DISP_PW_2),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_3", &tgl_pwdoms_pw_3,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_3,
+				.id = ICL_DISP_PW_3),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_vga = true,
+		.irq_pipe_mask = BIT(PIPE_B),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DDI_IO_A", &glk_pwdoms_ddi_io_a, .hsw.idx = ICL_PW_CTL_IDX_DDI_A),
+			I915_PW("DDI_IO_B", &glk_pwdoms_ddi_io_b, .hsw.idx = ICL_PW_CTL_IDX_DDI_B),
+			I915_PW("DDI_IO_C", &glk_pwdoms_ddi_io_c, .hsw.idx = ICL_PW_CTL_IDX_DDI_C),
+			I915_PW("DDI_IO_TC1", &tgl_pwdoms_ddi_io_tc1, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC1),
+			I915_PW("DDI_IO_TC2", &tgl_pwdoms_ddi_io_tc2, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC2),
+			I915_PW("DDI_IO_TC3", &tgl_pwdoms_ddi_io_tc3, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC3),
+			I915_PW("DDI_IO_TC4", &tgl_pwdoms_ddi_io_tc4, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC4),
+			I915_PW("DDI_IO_TC5", &tgl_pwdoms_ddi_io_tc5, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC5),
+			I915_PW("DDI_IO_TC6", &tgl_pwdoms_ddi_io_tc6, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC6),
+		),
+		.ops = &icl_ddi_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_4", &tgl_pwdoms_pw_4,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_4),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_fuses = true,
+		.irq_pipe_mask = BIT(PIPE_C),
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_5", &tgl_pwdoms_pw_5,
+				.hsw.idx = TGL_PW_CTL_IDX_PW_5),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_fuses = true,
+		.irq_pipe_mask = BIT(PIPE_D),
+	},
+};
+
+static const struct i915_power_well_desc tgl_power_wells_tc_cold_off[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("TC_cold_off", &tgl_pwdoms_tc_cold_off,
+				.id = TGL_DISP_PW_TC_COLD_OFF),
+		),
+		.ops = &tgl_tc_cold_off_ops,
+	},
+};
+
+static const struct i915_power_well_desc tgl_power_wells_aux[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("AUX_A", &icl_pwdoms_aux_a, .hsw.idx = ICL_PW_CTL_IDX_AUX_A),
+			I915_PW("AUX_B", &icl_pwdoms_aux_b, .hsw.idx = ICL_PW_CTL_IDX_AUX_B),
+			I915_PW("AUX_C", &icl_pwdoms_aux_c, .hsw.idx = ICL_PW_CTL_IDX_AUX_C),
+			I915_PW("AUX_USBC1", &tgl_pwdoms_aux_usbc1, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC1),
+			I915_PW("AUX_USBC2", &tgl_pwdoms_aux_usbc2, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC2),
+			I915_PW("AUX_USBC3", &tgl_pwdoms_aux_usbc3, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC3),
+			I915_PW("AUX_USBC4", &tgl_pwdoms_aux_usbc4, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC4),
+			I915_PW("AUX_USBC5", &tgl_pwdoms_aux_usbc5, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC5),
+			I915_PW("AUX_USBC6", &tgl_pwdoms_aux_usbc6, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC6),
+		),
+		.ops = &icl_aux_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("AUX_TBT1", &icl_pwdoms_aux_tbt1, .hsw.idx = TGL_PW_CTL_IDX_AUX_TBT1),
+			I915_PW("AUX_TBT2", &icl_pwdoms_aux_tbt2, .hsw.idx = TGL_PW_CTL_IDX_AUX_TBT2),
+			I915_PW("AUX_TBT3", &icl_pwdoms_aux_tbt3, .hsw.idx = TGL_PW_CTL_IDX_AUX_TBT3),
+			I915_PW("AUX_TBT4", &icl_pwdoms_aux_tbt4, .hsw.idx = TGL_PW_CTL_IDX_AUX_TBT4),
+			I915_PW("AUX_TBT5", &tgl_pwdoms_aux_tbt5, .hsw.idx = TGL_PW_CTL_IDX_AUX_TBT5),
+			I915_PW("AUX_TBT6", &tgl_pwdoms_aux_tbt6, .hsw.idx = TGL_PW_CTL_IDX_AUX_TBT6),
+		),
+		.ops = &icl_aux_power_well_ops,
+		.is_tc_tbt = true,
+	},
+};
+
+static const struct i915_power_well_desc_list tgl_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(icl_power_wells_pw_1),
+	I915_PW_DESCRIPTORS(tgl_power_wells_main),
+	I915_PW_DESCRIPTORS(tgl_power_wells_tc_cold_off),
+	I915_PW_DESCRIPTORS(tgl_power_wells_aux),
+};
+
+static const struct i915_power_well_desc_list adls_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(icl_power_wells_pw_1),
+	I915_PW_DESCRIPTORS(tgl_power_wells_main),
+	I915_PW_DESCRIPTORS(tgl_power_wells_aux),
+};
+
+#define RKL_PW_4_POWER_DOMAINS \
+	POWER_DOMAIN_PIPE_C, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C, \
+	POWER_DOMAIN_TRANSCODER_C
+
+I915_DECL_PW_DOMAINS(rkl_pwdoms_pw_4,
+	RKL_PW_4_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+#define RKL_PW_3_POWER_DOMAINS \
+	RKL_PW_4_POWER_DOMAINS, \
+	POWER_DOMAIN_PIPE_B, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B, \
+	POWER_DOMAIN_TRANSCODER_B, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC1, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC2, \
+	POWER_DOMAIN_VGA, \
+	POWER_DOMAIN_AUDIO_MMIO, \
+	POWER_DOMAIN_AUDIO_PLAYBACK, \
+	POWER_DOMAIN_AUX_USBC1, \
+	POWER_DOMAIN_AUX_USBC2
+
+I915_DECL_PW_DOMAINS(rkl_pwdoms_pw_3,
+	RKL_PW_3_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+/*
+ * There is no PW_2/PG_2 on RKL.
+ *
+ * RKL PW_1/PG_1 domains (under HW/DMC control):
+ * - DBUF function (note: registers are in PW0)
+ * - PIPE_A and its planes and VDSC/joining, except VGA
+ * - transcoder A
+ * - DDI_A and DDI_B
+ * - FBC
+ *
+ * RKL PW_0/PG_0 domains (under HW/DMC control):
+ * - PCI
+ * - clocks except port PLL
+ * - shared functions:
+ *     * interrupts except pipe interrupts
+ *     * MBus except PIPE_MBUS_DBOX_CTL
+ *     * DBUF registers
+ * - central power except FBC
+ * - top-level GTC (DDI-level GTC is in the well associated with the DDI)
+ */
+
+I915_DECL_PW_DOMAINS(rkl_pwdoms_dc_off,
+	RKL_PW_3_POWER_DOMAINS,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_MODESET,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc rkl_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DC_off", &rkl_pwdoms_dc_off,
+				.id = SKL_DISP_DC_OFF),
+		),
+		.ops = &gen9_dc_off_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_3", &rkl_pwdoms_pw_3,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_3,
+				.id = ICL_DISP_PW_3),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_B),
+		.has_vga = true,
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_4", &rkl_pwdoms_pw_4,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_4),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_fuses = true,
+		.irq_pipe_mask = BIT(PIPE_C),
+	},
+};
+
+static const struct i915_power_well_desc rkl_power_wells_ddi_aux[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DDI_IO_A", &glk_pwdoms_ddi_io_a, .hsw.idx = ICL_PW_CTL_IDX_DDI_A),
+			I915_PW("DDI_IO_B", &glk_pwdoms_ddi_io_b, .hsw.idx = ICL_PW_CTL_IDX_DDI_B),
+			I915_PW("DDI_IO_TC1", &tgl_pwdoms_ddi_io_tc1, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC1),
+			I915_PW("DDI_IO_TC2", &tgl_pwdoms_ddi_io_tc2, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC2),
+		),
+		.ops = &icl_ddi_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("AUX_A", &icl_pwdoms_aux_a, .hsw.idx = ICL_PW_CTL_IDX_AUX_A),
+			I915_PW("AUX_B", &icl_pwdoms_aux_b, .hsw.idx = ICL_PW_CTL_IDX_AUX_B),
+			I915_PW("AUX_USBC1", &tgl_pwdoms_aux_usbc1, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC1),
+			I915_PW("AUX_USBC2", &tgl_pwdoms_aux_usbc2, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC2),
+		),
+		.ops = &icl_aux_power_well_ops,
+	},
+};
+
+static const struct i915_power_well_desc_list rkl_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(icl_power_wells_pw_1),
+	I915_PW_DESCRIPTORS(rkl_power_wells_main),
+	I915_PW_DESCRIPTORS(rkl_power_wells_ddi_aux),
+};
+
+/*
+ * DG1 onwards Audio MMIO/VERBS lies in PG0 power well.
+ */
+#define DG1_PW_3_POWER_DOMAINS \
+	TGL_PW_4_POWER_DOMAINS, \
+	POWER_DOMAIN_PIPE_B, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B, \
+	POWER_DOMAIN_TRANSCODER_B, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC1, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC2, \
+	POWER_DOMAIN_VGA, \
+	POWER_DOMAIN_AUDIO_PLAYBACK, \
+	POWER_DOMAIN_AUX_USBC1, \
+	POWER_DOMAIN_AUX_USBC2
+
+I915_DECL_PW_DOMAINS(dg1_pwdoms_pw_3,
+	DG1_PW_3_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(dg1_pwdoms_dc_off,
+	DG1_PW_3_POWER_DOMAINS,
+	POWER_DOMAIN_AUDIO_MMIO,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_MODESET,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(dg1_pwdoms_pw_2,
+	DG1_PW_3_POWER_DOMAINS,
+	POWER_DOMAIN_TRANSCODER_VDSC_PW2,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc dg1_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DC_off", &dg1_pwdoms_dc_off,
+				.id = SKL_DISP_DC_OFF),
+		),
+		.ops = &gen9_dc_off_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_2", &dg1_pwdoms_pw_2,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_2,
+				.id = SKL_DISP_PW_2),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_3", &dg1_pwdoms_pw_3,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_3,
+				.id = ICL_DISP_PW_3),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_B),
+		.has_vga = true,
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_4", &tgl_pwdoms_pw_4,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_4),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_fuses = true,
+		.irq_pipe_mask = BIT(PIPE_C),
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_5", &tgl_pwdoms_pw_5,
+				.hsw.idx = TGL_PW_CTL_IDX_PW_5),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_fuses = true,
+		.irq_pipe_mask = BIT(PIPE_D),
+	},
+};
+
+static const struct i915_power_well_desc_list dg1_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(icl_power_wells_pw_1),
+	I915_PW_DESCRIPTORS(dg1_power_wells_main),
+	I915_PW_DESCRIPTORS(rkl_power_wells_ddi_aux),
+};
+
+/*
+ * XE_LPD Power Domains
+ *
+ * Previous platforms required that PG(n-1) be enabled before PG(n).  That
+ * dependency chain turns into a dependency tree on XE_LPD:
+ *
+ *       PG0
+ *        |
+ *     --PG1--
+ *    /       \
+ *  PGA     --PG2--
+ *         /   |   \
+ *       PGB  PGC  PGD
+ *
+ * Power wells must be enabled from top to bottom and disabled from bottom
+ * to top.  This allows pipes to be power gated independently.
+ */
+
+#define XELPD_PW_D_POWER_DOMAINS \
+	POWER_DOMAIN_PIPE_D, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_D, \
+	POWER_DOMAIN_TRANSCODER_D
+
+I915_DECL_PW_DOMAINS(xelpd_pwdoms_pw_d,
+	XELPD_PW_D_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+#define XELPD_PW_C_POWER_DOMAINS \
+	POWER_DOMAIN_PIPE_C, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_C, \
+	POWER_DOMAIN_TRANSCODER_C
+
+I915_DECL_PW_DOMAINS(xelpd_pwdoms_pw_c,
+	XELPD_PW_C_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+#define XELPD_PW_B_POWER_DOMAINS \
+	POWER_DOMAIN_PIPE_B, \
+	POWER_DOMAIN_PIPE_PANEL_FITTER_B, \
+	POWER_DOMAIN_TRANSCODER_B
+
+I915_DECL_PW_DOMAINS(xelpd_pwdoms_pw_b,
+	XELPD_PW_B_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(xelpd_pwdoms_pw_a,
+	POWER_DOMAIN_PIPE_A,
+	POWER_DOMAIN_PIPE_PANEL_FITTER_A,
+	POWER_DOMAIN_INIT);
+
+#define XELPD_PW_2_POWER_DOMAINS \
+	XELPD_PW_B_POWER_DOMAINS, \
+	XELPD_PW_C_POWER_DOMAINS, \
+	XELPD_PW_D_POWER_DOMAINS, \
+	POWER_DOMAIN_PORT_DDI_LANES_C, \
+	POWER_DOMAIN_PORT_DDI_LANES_D, \
+	POWER_DOMAIN_PORT_DDI_LANES_E, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC1, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC2, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC3, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC4, \
+	POWER_DOMAIN_VGA, \
+	POWER_DOMAIN_AUDIO_PLAYBACK, \
+	POWER_DOMAIN_AUX_C, \
+	POWER_DOMAIN_AUX_D, \
+	POWER_DOMAIN_AUX_E, \
+	POWER_DOMAIN_AUX_USBC1, \
+	POWER_DOMAIN_AUX_USBC2, \
+	POWER_DOMAIN_AUX_USBC3, \
+	POWER_DOMAIN_AUX_USBC4, \
+	POWER_DOMAIN_AUX_TBT1, \
+	POWER_DOMAIN_AUX_TBT2, \
+	POWER_DOMAIN_AUX_TBT3, \
+	POWER_DOMAIN_AUX_TBT4
+
+I915_DECL_PW_DOMAINS(xelpd_pwdoms_pw_2,
+	XELPD_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+/*
+ * XELPD PW_1/PG_1 domains (under HW/DMC control):
+ *  - DBUF function (registers are in PW0)
+ *  - Transcoder A
+ *  - DDI_A and DDI_B
+ *
+ * XELPD PW_0/PW_1 domains (under HW/DMC control):
+ *  - PCI
+ *  - Clocks except port PLL
+ *  - Shared functions:
+ *     * interrupts except pipe interrupts
+ *     * MBus except PIPE_MBUS_DBOX_CTL
+ *     * DBUF registers
+ *  - Central power except FBC
+ *  - Top-level GTC (DDI-level GTC is in the well associated with the DDI)
+ */
+
+I915_DECL_PW_DOMAINS(xelpd_pwdoms_dc_off,
+	XELPD_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_PORT_DSI,
+	POWER_DOMAIN_AUDIO_MMIO,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_MODESET,
+	POWER_DOMAIN_INIT);
+
+static const struct i915_power_well_desc xelpd_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DC_off", &xelpd_pwdoms_dc_off,
+				.id = SKL_DISP_DC_OFF),
+		),
+		.ops = &gen9_dc_off_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_2", &xelpd_pwdoms_pw_2,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_2,
+				.id = SKL_DISP_PW_2),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_vga = true,
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_A", &xelpd_pwdoms_pw_a,
+				.hsw.idx = XELPD_PW_CTL_IDX_PW_A),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_A),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_B", &xelpd_pwdoms_pw_b,
+				.hsw.idx = XELPD_PW_CTL_IDX_PW_B),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_B),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_C", &xelpd_pwdoms_pw_c,
+				.hsw.idx = XELPD_PW_CTL_IDX_PW_C),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_C),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_D", &xelpd_pwdoms_pw_d,
+				.hsw.idx = XELPD_PW_CTL_IDX_PW_D),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_D),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DDI_IO_A", &glk_pwdoms_ddi_io_a, .hsw.idx = ICL_PW_CTL_IDX_DDI_A),
+			I915_PW("DDI_IO_B", &glk_pwdoms_ddi_io_b, .hsw.idx = ICL_PW_CTL_IDX_DDI_B),
+			I915_PW("DDI_IO_C", &glk_pwdoms_ddi_io_c, .hsw.idx = ICL_PW_CTL_IDX_DDI_C),
+			I915_PW("DDI_IO_D", &icl_pwdoms_ddi_io_d, .hsw.idx = XELPD_PW_CTL_IDX_DDI_D),
+			I915_PW("DDI_IO_E", &icl_pwdoms_ddi_io_e, .hsw.idx = XELPD_PW_CTL_IDX_DDI_E),
+			I915_PW("DDI_IO_TC1", &tgl_pwdoms_ddi_io_tc1, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC1),
+			I915_PW("DDI_IO_TC2", &tgl_pwdoms_ddi_io_tc2, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC2),
+			I915_PW("DDI_IO_TC3", &tgl_pwdoms_ddi_io_tc3, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC3),
+			I915_PW("DDI_IO_TC4", &tgl_pwdoms_ddi_io_tc4, .hsw.idx = TGL_PW_CTL_IDX_DDI_TC4),
+		),
+		.ops = &icl_ddi_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("AUX_A", &icl_pwdoms_aux_a, .hsw.idx = ICL_PW_CTL_IDX_AUX_A),
+			I915_PW("AUX_B", &icl_pwdoms_aux_b, .hsw.idx = ICL_PW_CTL_IDX_AUX_B),
+			I915_PW("AUX_C", &icl_pwdoms_aux_c, .hsw.idx = ICL_PW_CTL_IDX_AUX_C),
+			I915_PW("AUX_D", &icl_pwdoms_aux_d, .hsw.idx = XELPD_PW_CTL_IDX_AUX_D),
+			I915_PW("AUX_E", &icl_pwdoms_aux_e, .hsw.idx = XELPD_PW_CTL_IDX_AUX_E),
+			I915_PW("AUX_USBC1", &tgl_pwdoms_aux_usbc1, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC1),
+			I915_PW("AUX_USBC2", &tgl_pwdoms_aux_usbc2, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC2),
+			I915_PW("AUX_USBC3", &tgl_pwdoms_aux_usbc3, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC3),
+			I915_PW("AUX_USBC4", &tgl_pwdoms_aux_usbc4, .hsw.idx = TGL_PW_CTL_IDX_AUX_TC4),
+		),
+		.ops = &icl_aux_power_well_ops,
+		.fixed_enable_delay = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("AUX_TBT1", &icl_pwdoms_aux_tbt1, .hsw.idx = TGL_PW_CTL_IDX_AUX_TBT1),
+			I915_PW("AUX_TBT2", &icl_pwdoms_aux_tbt2, .hsw.idx = TGL_PW_CTL_IDX_AUX_TBT2),
+			I915_PW("AUX_TBT3", &icl_pwdoms_aux_tbt3, .hsw.idx = TGL_PW_CTL_IDX_AUX_TBT3),
+			I915_PW("AUX_TBT4", &icl_pwdoms_aux_tbt4, .hsw.idx = TGL_PW_CTL_IDX_AUX_TBT4),
+		),
+		.ops = &icl_aux_power_well_ops,
+		.is_tc_tbt = true,
+	},
+};
+
+static const struct i915_power_well_desc_list xelpd_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(icl_power_wells_pw_1),
+	I915_PW_DESCRIPTORS(xelpd_power_wells_main),
+};
+
+/*
+ * MTL is based on XELPD power domains with the exception of power gating for:
+ * - DDI_IO (moved to PLL logic)
+ * - AUX and AUX_IO functionality and register access for USBC1-4 (PICA always-on)
+ */
+#define XELPDP_PW_2_POWER_DOMAINS \
+	XELPD_PW_B_POWER_DOMAINS, \
+	XELPD_PW_C_POWER_DOMAINS, \
+	XELPD_PW_D_POWER_DOMAINS, \
+	POWER_DOMAIN_AUDIO_PLAYBACK, \
+	POWER_DOMAIN_VGA, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC1, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC2, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC3, \
+	POWER_DOMAIN_PORT_DDI_LANES_TC4
+
+I915_DECL_PW_DOMAINS(xelpdp_pwdoms_pw_2,
+	XELPDP_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(xelpdp_pwdoms_dc_off,
+	XELPDP_PW_2_POWER_DOMAINS,
+	POWER_DOMAIN_AUDIO_MMIO,
+	POWER_DOMAIN_MODESET,
+	POWER_DOMAIN_AUX_A,
+	POWER_DOMAIN_AUX_B,
+	POWER_DOMAIN_INIT);
+
+I915_DECL_PW_DOMAINS(xelpdp_pwdoms_aux_tc1,
+	POWER_DOMAIN_AUX_USBC1,
+	POWER_DOMAIN_AUX_TBT1);
+
+I915_DECL_PW_DOMAINS(xelpdp_pwdoms_aux_tc2,
+	POWER_DOMAIN_AUX_USBC2,
+	POWER_DOMAIN_AUX_TBT2);
+
+I915_DECL_PW_DOMAINS(xelpdp_pwdoms_aux_tc3,
+	POWER_DOMAIN_AUX_USBC3,
+	POWER_DOMAIN_AUX_TBT3);
+
+I915_DECL_PW_DOMAINS(xelpdp_pwdoms_aux_tc4,
+	POWER_DOMAIN_AUX_USBC4,
+	POWER_DOMAIN_AUX_TBT4);
+
+static const struct i915_power_well_desc xelpdp_power_wells_main[] = {
+	{
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("DC_off", &xelpdp_pwdoms_dc_off,
+				.id = SKL_DISP_DC_OFF),
+		),
+		.ops = &gen9_dc_off_power_well_ops,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_2", &xelpdp_pwdoms_pw_2,
+				.hsw.idx = ICL_PW_CTL_IDX_PW_2,
+				.id = SKL_DISP_PW_2),
+		),
+		.ops = &hsw_power_well_ops,
+		.has_vga = true,
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_A", &xelpd_pwdoms_pw_a,
+				.hsw.idx = XELPD_PW_CTL_IDX_PW_A),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_A),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_B", &xelpd_pwdoms_pw_b,
+				.hsw.idx = XELPD_PW_CTL_IDX_PW_B),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_B),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_C", &xelpd_pwdoms_pw_c,
+				.hsw.idx = XELPD_PW_CTL_IDX_PW_C),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_C),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("PW_D", &xelpd_pwdoms_pw_d,
+				.hsw.idx = XELPD_PW_CTL_IDX_PW_D),
+		),
+		.ops = &hsw_power_well_ops,
+		.irq_pipe_mask = BIT(PIPE_D),
+		.has_fuses = true,
+	}, {
+		.instances = &I915_PW_INSTANCES(
+			I915_PW("AUX_A", &icl_pwdoms_aux_a, .xelpdp.aux_ch = AUX_CH_A),
+			I915_PW("AUX_B", &icl_pwdoms_aux_b, .xelpdp.aux_ch = AUX_CH_B),
+			I915_PW("AUX_TC1", &xelpdp_pwdoms_aux_tc1, .xelpdp.aux_ch = AUX_CH_USBC1),
+			I915_PW("AUX_TC2", &xelpdp_pwdoms_aux_tc2, .xelpdp.aux_ch = AUX_CH_USBC2),
+			I915_PW("AUX_TC3", &xelpdp_pwdoms_aux_tc3, .xelpdp.aux_ch = AUX_CH_USBC3),
+			I915_PW("AUX_TC4", &xelpdp_pwdoms_aux_tc4, .xelpdp.aux_ch = AUX_CH_USBC4),
+		),
+		.ops = &xelpdp_aux_power_well_ops,
+	},
+};
+
+static const struct i915_power_well_desc_list xelpdp_power_wells[] = {
+	I915_PW_DESCRIPTORS(i9xx_power_wells_always_on),
+	I915_PW_DESCRIPTORS(icl_power_wells_pw_1),
+	I915_PW_DESCRIPTORS(xelpdp_power_wells_main),
+};
+
+static void init_power_well_domains(const struct i915_power_well_instance *inst,
+				    struct i915_power_well *power_well)
+{
+	int j;
+
+	if (!inst->domain_list)
+		return;
+
+	if (inst->domain_list->count == 0) {
+		bitmap_fill(power_well->domains.bits, POWER_DOMAIN_NUM);
+
+		return;
+	}
+
+	for (j = 0; j < inst->domain_list->count; j++)
+		set_bit(inst->domain_list->list[j], power_well->domains.bits);
+}
+
+#define for_each_power_well_instance_in_desc_list(_desc_list, _desc_count, _desc, _inst) \
+	for ((_desc) = (_desc_list); (_desc) - (_desc_list) < (_desc_count); (_desc)++) \
+		for ((_inst) = (_desc)->instances->list; \
+		     (_inst) - (_desc)->instances->list < (_desc)->instances->count; \
+		     (_inst)++)
+
+#define for_each_power_well_instance(_desc_list, _desc_count, _descs, _desc, _inst) \
+	for ((_descs) = (_desc_list); \
+	     (_descs) - (_desc_list) < (_desc_count); \
+	     (_descs)++) \
+		for_each_power_well_instance_in_desc_list((_descs)->list, (_descs)->count, \
+							  (_desc), (_inst))
+
+static int
+__set_power_wells(struct i915_power_domains *power_domains,
+		  const struct i915_power_well_desc_list *power_well_descs,
+		  int power_well_descs_sz)
+{
+	struct drm_i915_private *i915 = container_of(power_domains,
+						     struct drm_i915_private,
+						     display.power.domains);
+	u64 power_well_ids = 0;
+	const struct i915_power_well_desc_list *desc_list;
+	const struct i915_power_well_desc *desc;
+	const struct i915_power_well_instance *inst;
+	int power_well_count = 0;
+	int plt_idx = 0;
+
+	for_each_power_well_instance(power_well_descs, power_well_descs_sz, desc_list, desc, inst)
+		power_well_count++;
+
+	power_domains->power_well_count = power_well_count;
+	power_domains->power_wells =
+				kcalloc(power_well_count,
+					sizeof(*power_domains->power_wells),
+					GFP_KERNEL);
+	if (!power_domains->power_wells)
+		return -ENOMEM;
+
+	for_each_power_well_instance(power_well_descs, power_well_descs_sz, desc_list, desc, inst) {
+		struct i915_power_well *pw = &power_domains->power_wells[plt_idx];
+		enum i915_power_well_id id = inst->id;
+
+		pw->desc = desc;
+		drm_WARN_ON(&i915->drm,
+			    overflows_type(inst - desc->instances->list, pw->instance_idx));
+		pw->instance_idx = inst - desc->instances->list;
+
+		init_power_well_domains(inst, pw);
+
+		plt_idx++;
+
+		if (id == DISP_PW_ID_NONE)
+			continue;
+
+		drm_WARN_ON(&i915->drm, id >= sizeof(power_well_ids) * 8);
+		drm_WARN_ON(&i915->drm, power_well_ids & BIT_ULL(id));
+		power_well_ids |= BIT_ULL(id);
+	}
+
+	return 0;
+}
+
+#define set_power_wells(power_domains, __power_well_descs) \
+	__set_power_wells(power_domains, __power_well_descs, \
+			  ARRAY_SIZE(__power_well_descs))
+
+/**
+ * intel_display_power_map_init - initialize power domain -> power well mappings
+ * @power_domains: power domain state
+ *
+ * Creates all the power wells for the current platform, initializes the
+ * dynamic state for them and initializes the mapping of each power well to
+ * all the power domains the power well belongs to.
+ */
+int intel_display_power_map_init(struct i915_power_domains *power_domains)
+{
+	struct drm_i915_private *i915 = container_of(power_domains,
+						     struct drm_i915_private,
+						     display.power.domains);
+	/*
+	 * The enabling order will be from lower to higher indexed wells,
+	 * the disabling order is reversed.
+	 */
+	if (!HAS_DISPLAY(i915)) {
+		power_domains->power_well_count = 0;
+		return 0;
+	}
+
+	if (DISPLAY_VER(i915) >= 14)
+		return set_power_wells(power_domains, xelpdp_power_wells);
+	else if (DISPLAY_VER(i915) >= 13)
+		return set_power_wells(power_domains, xelpd_power_wells);
+	else if (IS_DG1(i915))
+		return set_power_wells(power_domains, dg1_power_wells);
+	else if (IS_ALDERLAKE_S(i915))
+		return set_power_wells(power_domains, adls_power_wells);
+	else if (IS_ROCKETLAKE(i915))
+		return set_power_wells(power_domains, rkl_power_wells);
+	else if (DISPLAY_VER(i915) == 12)
+		return set_power_wells(power_domains, tgl_power_wells);
+	else if (DISPLAY_VER(i915) == 11)
+		return set_power_wells(power_domains, icl_power_wells);
+	else if (IS_GEMINILAKE(i915))
+		return set_power_wells(power_domains, glk_power_wells);
+	else if (IS_BROXTON(i915))
+		return set_power_wells(power_domains, bxt_power_wells);
+	else if (DISPLAY_VER(i915) == 9)
+		return set_power_wells(power_domains, skl_power_wells);
+	else if (IS_CHERRYVIEW(i915))
+		return set_power_wells(power_domains, chv_power_wells);
+	else if (IS_BROADWELL(i915))
+		return set_power_wells(power_domains, bdw_power_wells);
+	else if (IS_HASWELL(i915))
+		return set_power_wells(power_domains, hsw_power_wells);
+	else if (IS_VALLEYVIEW(i915))
+		return set_power_wells(power_domains, vlv_power_wells);
+	else if (IS_I830(i915))
+		return set_power_wells(power_domains, i830_power_wells);
+	else
+		return set_power_wells(power_domains, i9xx_power_wells);
+}
+
+/**
+ * intel_display_power_map_cleanup - clean up power domain -> power well mappings
+ * @power_domains: power domain state
+ *
+ * Cleans up all the state that was initialized by intel_display_power_map_init().
+ */
+void intel_display_power_map_cleanup(struct i915_power_domains *power_domains)
+{
+	kfree(power_domains->power_wells);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_display_power_map.h b/drivers/gpu/drm/i915/display/intel_display_power_map.h
new file mode 100644
index 000000000..da8f7055a
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_power_map.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_DISPLAY_POWER_MAP_H__
+#define __INTEL_DISPLAY_POWER_MAP_H__
+
+struct i915_power_domains;
+
+int intel_display_power_map_init(struct i915_power_domains *power_domains);
+void intel_display_power_map_cleanup(struct i915_power_domains *power_domains);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_display_power_well.c b/drivers/gpu/drm/i915/display/intel_display_power_well.c
new file mode 100644
index 000000000..1d18eee56
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_power_well.c
@@ -0,0 +1,1956 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_irq.h"
+#include "intel_backlight_regs.h"
+#include "intel_combo_phy.h"
+#include "intel_combo_phy_regs.h"
+#include "intel_crt.h"
+#include "intel_de.h"
+#include "intel_display_power_well.h"
+#include "intel_display_types.h"
+#include "intel_dkl_phy.h"
+#include "intel_dmc.h"
+#include "intel_dpio_phy.h"
+#include "intel_dpll.h"
+#include "intel_hotplug.h"
+#include "intel_pcode.h"
+#include "intel_pps.h"
+#include "intel_tc.h"
+#include "intel_vga.h"
+#include "skl_watermark.h"
+#include "vlv_sideband.h"
+#include "vlv_sideband_reg.h"
+
+struct i915_power_well_regs {
+	i915_reg_t bios;
+	i915_reg_t driver;
+	i915_reg_t kvmr;
+	i915_reg_t debug;
+};
+
+struct i915_power_well_ops {
+	const struct i915_power_well_regs *regs;
+	/*
+	 * Synchronize the well's hw state to match the current sw state, for
+	 * example enable/disable it based on the current refcount. Called
+	 * during driver init and resume time, possibly after first calling
+	 * the enable/disable handlers.
+	 */
+	void (*sync_hw)(struct drm_i915_private *i915,
+			struct i915_power_well *power_well);
+	/*
+	 * Enable the well and resources that depend on it (for example
+	 * interrupts located on the well). Called after the 0->1 refcount
+	 * transition.
+	 */
+	void (*enable)(struct drm_i915_private *i915,
+		       struct i915_power_well *power_well);
+	/*
+	 * Disable the well and resources that depend on it. Called after
+	 * the 1->0 refcount transition.
+	 */
+	void (*disable)(struct drm_i915_private *i915,
+			struct i915_power_well *power_well);
+	/* Returns the hw enabled state. */
+	bool (*is_enabled)(struct drm_i915_private *i915,
+			   struct i915_power_well *power_well);
+};
+
+static const struct i915_power_well_instance *
+i915_power_well_instance(const struct i915_power_well *power_well)
+{
+	return &power_well->desc->instances->list[power_well->instance_idx];
+}
+
+struct i915_power_well *
+lookup_power_well(struct drm_i915_private *i915,
+		  enum i915_power_well_id power_well_id)
+{
+	struct i915_power_well *power_well;
+
+	for_each_power_well(i915, power_well)
+		if (i915_power_well_instance(power_well)->id == power_well_id)
+			return power_well;
+
+	/*
+	 * It's not feasible to add error checking code to the callers since
+	 * this condition really shouldn't happen and it doesn't even make sense
+	 * to abort things like display initialization sequences. Just return
+	 * the first power well and hope the WARN gets reported so we can fix
+	 * our driver.
+	 */
+	drm_WARN(&i915->drm, 1,
+		 "Power well %d not defined for this platform\n",
+		 power_well_id);
+	return &i915->display.power.domains.power_wells[0];
+}
+
+void intel_power_well_enable(struct drm_i915_private *i915,
+			     struct i915_power_well *power_well)
+{
+	drm_dbg_kms(&i915->drm, "enabling %s\n", intel_power_well_name(power_well));
+	power_well->desc->ops->enable(i915, power_well);
+	power_well->hw_enabled = true;
+}
+
+void intel_power_well_disable(struct drm_i915_private *i915,
+			      struct i915_power_well *power_well)
+{
+	drm_dbg_kms(&i915->drm, "disabling %s\n", intel_power_well_name(power_well));
+	power_well->hw_enabled = false;
+	power_well->desc->ops->disable(i915, power_well);
+}
+
+void intel_power_well_sync_hw(struct drm_i915_private *i915,
+			      struct i915_power_well *power_well)
+{
+	power_well->desc->ops->sync_hw(i915, power_well);
+	power_well->hw_enabled =
+		power_well->desc->ops->is_enabled(i915, power_well);
+}
+
+void intel_power_well_get(struct drm_i915_private *i915,
+			  struct i915_power_well *power_well)
+{
+	if (!power_well->count++)
+		intel_power_well_enable(i915, power_well);
+}
+
+void intel_power_well_put(struct drm_i915_private *i915,
+			  struct i915_power_well *power_well)
+{
+	drm_WARN(&i915->drm, !power_well->count,
+		 "Use count on power well %s is already zero",
+		 i915_power_well_instance(power_well)->name);
+
+	if (!--power_well->count)
+		intel_power_well_disable(i915, power_well);
+}
+
+bool intel_power_well_is_enabled(struct drm_i915_private *i915,
+				 struct i915_power_well *power_well)
+{
+	return power_well->desc->ops->is_enabled(i915, power_well);
+}
+
+bool intel_power_well_is_enabled_cached(struct i915_power_well *power_well)
+{
+	return power_well->hw_enabled;
+}
+
+bool intel_display_power_well_is_enabled(struct drm_i915_private *dev_priv,
+					 enum i915_power_well_id power_well_id)
+{
+	struct i915_power_well *power_well;
+
+	power_well = lookup_power_well(dev_priv, power_well_id);
+
+	return intel_power_well_is_enabled(dev_priv, power_well);
+}
+
+bool intel_power_well_is_always_on(struct i915_power_well *power_well)
+{
+	return power_well->desc->always_on;
+}
+
+const char *intel_power_well_name(struct i915_power_well *power_well)
+{
+	return i915_power_well_instance(power_well)->name;
+}
+
+struct intel_power_domain_mask *intel_power_well_domains(struct i915_power_well *power_well)
+{
+	return &power_well->domains;
+}
+
+int intel_power_well_refcount(struct i915_power_well *power_well)
+{
+	return power_well->count;
+}
+
+/*
+ * Starting with Haswell, we have a "Power Down Well" that can be turned off
+ * when not needed anymore. We have 4 registers that can request the power well
+ * to be enabled, and it will only be disabled if none of the registers is
+ * requesting it to be enabled.
+ */
+static void hsw_power_well_post_enable(struct drm_i915_private *dev_priv,
+				       u8 irq_pipe_mask, bool has_vga)
+{
+	if (has_vga)
+		intel_vga_reset_io_mem(dev_priv);
+
+	if (irq_pipe_mask)
+		gen8_irq_power_well_post_enable(dev_priv, irq_pipe_mask);
+}
+
+static void hsw_power_well_pre_disable(struct drm_i915_private *dev_priv,
+				       u8 irq_pipe_mask)
+{
+	if (irq_pipe_mask)
+		gen8_irq_power_well_pre_disable(dev_priv, irq_pipe_mask);
+}
+
+#define ICL_AUX_PW_TO_CH(pw_idx)	\
+	((pw_idx) - ICL_PW_CTL_IDX_AUX_A + AUX_CH_A)
+
+#define ICL_TBT_AUX_PW_TO_CH(pw_idx)	\
+	((pw_idx) - ICL_PW_CTL_IDX_AUX_TBT1 + AUX_CH_C)
+
+static enum aux_ch icl_aux_pw_to_ch(const struct i915_power_well *power_well)
+{
+	int pw_idx = i915_power_well_instance(power_well)->hsw.idx;
+
+	return power_well->desc->is_tc_tbt ? ICL_TBT_AUX_PW_TO_CH(pw_idx) :
+					     ICL_AUX_PW_TO_CH(pw_idx);
+}
+
+static struct intel_digital_port *
+aux_ch_to_digital_port(struct drm_i915_private *dev_priv,
+		       enum aux_ch aux_ch)
+{
+	struct intel_digital_port *dig_port = NULL;
+	struct intel_encoder *encoder;
+
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		/* We'll check the MST primary port */
+		if (encoder->type == INTEL_OUTPUT_DP_MST)
+			continue;
+
+		dig_port = enc_to_dig_port(encoder);
+		if (!dig_port)
+			continue;
+
+		if (dig_port->aux_ch != aux_ch) {
+			dig_port = NULL;
+			continue;
+		}
+
+		break;
+	}
+
+	return dig_port;
+}
+
+static enum phy icl_aux_pw_to_phy(struct drm_i915_private *i915,
+				  const struct i915_power_well *power_well)
+{
+	enum aux_ch aux_ch = icl_aux_pw_to_ch(power_well);
+	struct intel_digital_port *dig_port = aux_ch_to_digital_port(i915, aux_ch);
+
+	return intel_port_to_phy(i915, dig_port->base.port);
+}
+
+static void hsw_wait_for_power_well_enable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well,
+					   bool timeout_expected)
+{
+	const struct i915_power_well_regs *regs = power_well->desc->ops->regs;
+	int pw_idx = i915_power_well_instance(power_well)->hsw.idx;
+
+	/*
+	 * For some power wells we're not supposed to watch the status bit for
+	 * an ack, but rather just wait a fixed amount of time and then
+	 * proceed.  This is only used on DG2.
+	 */
+	if (IS_DG2(dev_priv) && power_well->desc->fixed_enable_delay) {
+		usleep_range(600, 1200);
+		return;
+	}
+
+	/* Timeout for PW1:10 us, AUX:not specified, other PWs:20 us. */
+	if (intel_de_wait_for_set(dev_priv, regs->driver,
+				  HSW_PWR_WELL_CTL_STATE(pw_idx), 1)) {
+		drm_dbg_kms(&dev_priv->drm, "%s power well enable timeout\n",
+			    intel_power_well_name(power_well));
+
+		drm_WARN_ON(&dev_priv->drm, !timeout_expected);
+
+	}
+}
+
+static u32 hsw_power_well_requesters(struct drm_i915_private *dev_priv,
+				     const struct i915_power_well_regs *regs,
+				     int pw_idx)
+{
+	u32 req_mask = HSW_PWR_WELL_CTL_REQ(pw_idx);
+	u32 ret;
+
+	ret = intel_de_read(dev_priv, regs->bios) & req_mask ? 1 : 0;
+	ret |= intel_de_read(dev_priv, regs->driver) & req_mask ? 2 : 0;
+	if (regs->kvmr.reg)
+		ret |= intel_de_read(dev_priv, regs->kvmr) & req_mask ? 4 : 0;
+	ret |= intel_de_read(dev_priv, regs->debug) & req_mask ? 8 : 0;
+
+	return ret;
+}
+
+static void hsw_wait_for_power_well_disable(struct drm_i915_private *dev_priv,
+					    struct i915_power_well *power_well)
+{
+	const struct i915_power_well_regs *regs = power_well->desc->ops->regs;
+	int pw_idx = i915_power_well_instance(power_well)->hsw.idx;
+	bool disabled;
+	u32 reqs;
+
+	/*
+	 * Bspec doesn't require waiting for PWs to get disabled, but still do
+	 * this for paranoia. The known cases where a PW will be forced on:
+	 * - a KVMR request on any power well via the KVMR request register
+	 * - a DMC request on PW1 and MISC_IO power wells via the BIOS and
+	 *   DEBUG request registers
+	 * Skip the wait in case any of the request bits are set and print a
+	 * diagnostic message.
+	 */
+	wait_for((disabled = !(intel_de_read(dev_priv, regs->driver) &
+			       HSW_PWR_WELL_CTL_STATE(pw_idx))) ||
+		 (reqs = hsw_power_well_requesters(dev_priv, regs, pw_idx)), 1);
+	if (disabled)
+		return;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "%s forced on (bios:%d driver:%d kvmr:%d debug:%d)\n",
+		    intel_power_well_name(power_well),
+		    !!(reqs & 1), !!(reqs & 2), !!(reqs & 4), !!(reqs & 8));
+}
+
+static void gen9_wait_for_power_well_fuses(struct drm_i915_private *dev_priv,
+					   enum skl_power_gate pg)
+{
+	/* Timeout 5us for PG#0, for other PGs 1us */
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_wait_for_set(dev_priv, SKL_FUSE_STATUS,
+					  SKL_FUSE_PG_DIST_STATUS(pg), 1));
+}
+
+static void hsw_power_well_enable(struct drm_i915_private *dev_priv,
+				  struct i915_power_well *power_well)
+{
+	const struct i915_power_well_regs *regs = power_well->desc->ops->regs;
+	int pw_idx = i915_power_well_instance(power_well)->hsw.idx;
+	u32 val;
+
+	if (power_well->desc->has_fuses) {
+		enum skl_power_gate pg;
+
+		pg = DISPLAY_VER(dev_priv) >= 11 ? ICL_PW_CTL_IDX_TO_PG(pw_idx) :
+						 SKL_PW_CTL_IDX_TO_PG(pw_idx);
+
+		/* Wa_16013190616:adlp */
+		if (IS_ALDERLAKE_P(dev_priv) && pg == SKL_PG1)
+			intel_de_rmw(dev_priv, GEN8_CHICKEN_DCPR_1, 0, DISABLE_FLR_SRC);
+
+		/*
+		 * For PW1 we have to wait both for the PW0/PG0 fuse state
+		 * before enabling the power well and PW1/PG1's own fuse
+		 * state after the enabling. For all other power wells with
+		 * fuses we only have to wait for that PW/PG's fuse state
+		 * after the enabling.
+		 */
+		if (pg == SKL_PG1)
+			gen9_wait_for_power_well_fuses(dev_priv, SKL_PG0);
+	}
+
+	val = intel_de_read(dev_priv, regs->driver);
+	intel_de_write(dev_priv, regs->driver,
+		       val | HSW_PWR_WELL_CTL_REQ(pw_idx));
+
+	hsw_wait_for_power_well_enable(dev_priv, power_well, false);
+
+	if (power_well->desc->has_fuses) {
+		enum skl_power_gate pg;
+
+		pg = DISPLAY_VER(dev_priv) >= 11 ? ICL_PW_CTL_IDX_TO_PG(pw_idx) :
+						 SKL_PW_CTL_IDX_TO_PG(pw_idx);
+		gen9_wait_for_power_well_fuses(dev_priv, pg);
+	}
+
+	hsw_power_well_post_enable(dev_priv,
+				   power_well->desc->irq_pipe_mask,
+				   power_well->desc->has_vga);
+}
+
+static void hsw_power_well_disable(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	const struct i915_power_well_regs *regs = power_well->desc->ops->regs;
+	int pw_idx = i915_power_well_instance(power_well)->hsw.idx;
+	u32 val;
+
+	hsw_power_well_pre_disable(dev_priv,
+				   power_well->desc->irq_pipe_mask);
+
+	val = intel_de_read(dev_priv, regs->driver);
+	intel_de_write(dev_priv, regs->driver,
+		       val & ~HSW_PWR_WELL_CTL_REQ(pw_idx));
+	hsw_wait_for_power_well_disable(dev_priv, power_well);
+}
+
+static void
+icl_combo_phy_aux_power_well_enable(struct drm_i915_private *dev_priv,
+				    struct i915_power_well *power_well)
+{
+	const struct i915_power_well_regs *regs = power_well->desc->ops->regs;
+	int pw_idx = i915_power_well_instance(power_well)->hsw.idx;
+	enum phy phy = icl_aux_pw_to_phy(dev_priv, power_well);
+	u32 val;
+
+	drm_WARN_ON(&dev_priv->drm, !IS_ICELAKE(dev_priv));
+
+	val = intel_de_read(dev_priv, regs->driver);
+	intel_de_write(dev_priv, regs->driver,
+		       val | HSW_PWR_WELL_CTL_REQ(pw_idx));
+
+	if (DISPLAY_VER(dev_priv) < 12) {
+		val = intel_de_read(dev_priv, ICL_PORT_CL_DW12(phy));
+		intel_de_write(dev_priv, ICL_PORT_CL_DW12(phy),
+			       val | ICL_LANE_ENABLE_AUX);
+	}
+
+	hsw_wait_for_power_well_enable(dev_priv, power_well, false);
+
+	/* Display WA #1178: icl */
+	if (pw_idx >= ICL_PW_CTL_IDX_AUX_A && pw_idx <= ICL_PW_CTL_IDX_AUX_B &&
+	    !intel_bios_is_port_edp(dev_priv, (enum port)phy)) {
+		val = intel_de_read(dev_priv, ICL_AUX_ANAOVRD1(pw_idx));
+		val |= ICL_AUX_ANAOVRD1_ENABLE | ICL_AUX_ANAOVRD1_LDO_BYPASS;
+		intel_de_write(dev_priv, ICL_AUX_ANAOVRD1(pw_idx), val);
+	}
+}
+
+static void
+icl_combo_phy_aux_power_well_disable(struct drm_i915_private *dev_priv,
+				     struct i915_power_well *power_well)
+{
+	const struct i915_power_well_regs *regs = power_well->desc->ops->regs;
+	int pw_idx = i915_power_well_instance(power_well)->hsw.idx;
+	enum phy phy = icl_aux_pw_to_phy(dev_priv, power_well);
+	u32 val;
+
+	drm_WARN_ON(&dev_priv->drm, !IS_ICELAKE(dev_priv));
+
+	val = intel_de_read(dev_priv, ICL_PORT_CL_DW12(phy));
+	intel_de_write(dev_priv, ICL_PORT_CL_DW12(phy),
+		       val & ~ICL_LANE_ENABLE_AUX);
+
+	val = intel_de_read(dev_priv, regs->driver);
+	intel_de_write(dev_priv, regs->driver,
+		       val & ~HSW_PWR_WELL_CTL_REQ(pw_idx));
+
+	hsw_wait_for_power_well_disable(dev_priv, power_well);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+
+static void icl_tc_port_assert_ref_held(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well,
+					struct intel_digital_port *dig_port)
+{
+	if (drm_WARN_ON(&dev_priv->drm, !dig_port))
+		return;
+
+	if (DISPLAY_VER(dev_priv) == 11 && intel_tc_cold_requires_aux_pw(dig_port))
+		return;
+
+	drm_WARN_ON(&dev_priv->drm, !intel_tc_port_ref_held(dig_port));
+}
+
+#else
+
+static void icl_tc_port_assert_ref_held(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well,
+					struct intel_digital_port *dig_port)
+{
+}
+
+#endif
+
+#define TGL_AUX_PW_TO_TC_PORT(pw_idx)	((pw_idx) - TGL_PW_CTL_IDX_AUX_TC1)
+
+static void icl_tc_cold_exit(struct drm_i915_private *i915)
+{
+	int ret, tries = 0;
+
+	while (1) {
+		ret = snb_pcode_write_timeout(&i915->uncore, ICL_PCODE_EXIT_TCCOLD, 0,
+					      250, 1);
+		if (ret != -EAGAIN || ++tries == 3)
+			break;
+		msleep(1);
+	}
+
+	/* Spec states that TC cold exit can take up to 1ms to complete */
+	if (!ret)
+		msleep(1);
+
+	/* TODO: turn failure into a error as soon i915 CI updates ICL IFWI */
+	drm_dbg_kms(&i915->drm, "TC cold block %s\n", ret ? "failed" :
+		    "succeeded");
+}
+
+static void
+icl_tc_phy_aux_power_well_enable(struct drm_i915_private *dev_priv,
+				 struct i915_power_well *power_well)
+{
+	enum aux_ch aux_ch = icl_aux_pw_to_ch(power_well);
+	struct intel_digital_port *dig_port = aux_ch_to_digital_port(dev_priv, aux_ch);
+	const struct i915_power_well_regs *regs = power_well->desc->ops->regs;
+	bool is_tbt = power_well->desc->is_tc_tbt;
+	bool timeout_expected;
+	u32 val;
+
+	icl_tc_port_assert_ref_held(dev_priv, power_well, dig_port);
+
+	val = intel_de_read(dev_priv, DP_AUX_CH_CTL(aux_ch));
+	val &= ~DP_AUX_CH_CTL_TBT_IO;
+	if (is_tbt)
+		val |= DP_AUX_CH_CTL_TBT_IO;
+	intel_de_write(dev_priv, DP_AUX_CH_CTL(aux_ch), val);
+
+	val = intel_de_read(dev_priv, regs->driver);
+	intel_de_write(dev_priv, regs->driver,
+		       val | HSW_PWR_WELL_CTL_REQ(i915_power_well_instance(power_well)->hsw.idx));
+
+	/*
+	 * An AUX timeout is expected if the TBT DP tunnel is down,
+	 * or need to enable AUX on a legacy TypeC port as part of the TC-cold
+	 * exit sequence.
+	 */
+	timeout_expected = is_tbt || intel_tc_cold_requires_aux_pw(dig_port);
+	if (DISPLAY_VER(dev_priv) == 11 && intel_tc_cold_requires_aux_pw(dig_port))
+		icl_tc_cold_exit(dev_priv);
+
+	hsw_wait_for_power_well_enable(dev_priv, power_well, timeout_expected);
+
+	if (DISPLAY_VER(dev_priv) >= 12 && !is_tbt) {
+		enum tc_port tc_port;
+
+		tc_port = TGL_AUX_PW_TO_TC_PORT(i915_power_well_instance(power_well)->hsw.idx);
+
+		if (wait_for(intel_dkl_phy_read(dev_priv, DKL_CMN_UC_DW_27(tc_port), 2) &
+			     DKL_CMN_UC_DW27_UC_HEALTH, 1))
+			drm_warn(&dev_priv->drm,
+				 "Timeout waiting TC uC health\n");
+	}
+}
+
+static void
+icl_aux_power_well_enable(struct drm_i915_private *dev_priv,
+			  struct i915_power_well *power_well)
+{
+	enum phy phy = icl_aux_pw_to_phy(dev_priv, power_well);
+
+	if (intel_phy_is_tc(dev_priv, phy))
+		return icl_tc_phy_aux_power_well_enable(dev_priv, power_well);
+	else if (IS_ICELAKE(dev_priv))
+		return icl_combo_phy_aux_power_well_enable(dev_priv,
+							   power_well);
+	else
+		return hsw_power_well_enable(dev_priv, power_well);
+}
+
+static void
+icl_aux_power_well_disable(struct drm_i915_private *dev_priv,
+			   struct i915_power_well *power_well)
+{
+	enum phy phy = icl_aux_pw_to_phy(dev_priv, power_well);
+
+	if (intel_phy_is_tc(dev_priv, phy))
+		return hsw_power_well_disable(dev_priv, power_well);
+	else if (IS_ICELAKE(dev_priv))
+		return icl_combo_phy_aux_power_well_disable(dev_priv,
+							    power_well);
+	else
+		return hsw_power_well_disable(dev_priv, power_well);
+}
+
+/*
+ * We should only use the power well if we explicitly asked the hardware to
+ * enable it, so check if it's enabled and also check if we've requested it to
+ * be enabled.
+ */
+static bool hsw_power_well_enabled(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	const struct i915_power_well_regs *regs = power_well->desc->ops->regs;
+	enum i915_power_well_id id = i915_power_well_instance(power_well)->id;
+	int pw_idx = i915_power_well_instance(power_well)->hsw.idx;
+	u32 mask = HSW_PWR_WELL_CTL_REQ(pw_idx) |
+		   HSW_PWR_WELL_CTL_STATE(pw_idx);
+	u32 val;
+
+	val = intel_de_read(dev_priv, regs->driver);
+
+	/*
+	 * On GEN9 big core due to a DMC bug the driver's request bits for PW1
+	 * and the MISC_IO PW will be not restored, so check instead for the
+	 * BIOS's own request bits, which are forced-on for these power wells
+	 * when exiting DC5/6.
+	 */
+	if (DISPLAY_VER(dev_priv) == 9 && !IS_BROXTON(dev_priv) &&
+	    (id == SKL_DISP_PW_1 || id == SKL_DISP_PW_MISC_IO))
+		val |= intel_de_read(dev_priv, regs->bios);
+
+	return (val & mask) == mask;
+}
+
+static void assert_can_enable_dc9(struct drm_i915_private *dev_priv)
+{
+	drm_WARN_ONCE(&dev_priv->drm,
+		      (intel_de_read(dev_priv, DC_STATE_EN) & DC_STATE_EN_DC9),
+		      "DC9 already programmed to be enabled.\n");
+	drm_WARN_ONCE(&dev_priv->drm,
+		      intel_de_read(dev_priv, DC_STATE_EN) &
+		      DC_STATE_EN_UPTO_DC5,
+		      "DC5 still not disabled to enable DC9.\n");
+	drm_WARN_ONCE(&dev_priv->drm,
+		      intel_de_read(dev_priv, HSW_PWR_WELL_CTL2) &
+		      HSW_PWR_WELL_CTL_REQ(SKL_PW_CTL_IDX_PW_2),
+		      "Power well 2 on.\n");
+	drm_WARN_ONCE(&dev_priv->drm, intel_irqs_enabled(dev_priv),
+		      "Interrupts not disabled yet.\n");
+
+	 /*
+	  * TODO: check for the following to verify the conditions to enter DC9
+	  * state are satisfied:
+	  * 1] Check relevant display engine registers to verify if mode set
+	  * disable sequence was followed.
+	  * 2] Check if display uninitialize sequence is initialized.
+	  */
+}
+
+static void assert_can_disable_dc9(struct drm_i915_private *dev_priv)
+{
+	drm_WARN_ONCE(&dev_priv->drm, intel_irqs_enabled(dev_priv),
+		      "Interrupts not disabled yet.\n");
+	drm_WARN_ONCE(&dev_priv->drm,
+		      intel_de_read(dev_priv, DC_STATE_EN) &
+		      DC_STATE_EN_UPTO_DC5,
+		      "DC5 still not disabled.\n");
+
+	 /*
+	  * TODO: check for the following to verify DC9 state was indeed
+	  * entered before programming to disable it:
+	  * 1] Check relevant display engine registers to verify if mode
+	  *  set disable sequence was followed.
+	  * 2] Check if display uninitialize sequence is initialized.
+	  */
+}
+
+static void gen9_write_dc_state(struct drm_i915_private *dev_priv,
+				u32 state)
+{
+	int rewrites = 0;
+	int rereads = 0;
+	u32 v;
+
+	intel_de_write(dev_priv, DC_STATE_EN, state);
+
+	/* It has been observed that disabling the dc6 state sometimes
+	 * doesn't stick and dmc keeps returning old value. Make sure
+	 * the write really sticks enough times and also force rewrite until
+	 * we are confident that state is exactly what we want.
+	 */
+	do  {
+		v = intel_de_read(dev_priv, DC_STATE_EN);
+
+		if (v != state) {
+			intel_de_write(dev_priv, DC_STATE_EN, state);
+			rewrites++;
+			rereads = 0;
+		} else if (rereads++ > 5) {
+			break;
+		}
+
+	} while (rewrites < 100);
+
+	if (v != state)
+		drm_err(&dev_priv->drm,
+			"Writing dc state to 0x%x failed, now 0x%x\n",
+			state, v);
+
+	/* Most of the times we need one retry, avoid spam */
+	if (rewrites > 1)
+		drm_dbg_kms(&dev_priv->drm,
+			    "Rewrote dc state to 0x%x %d times\n",
+			    state, rewrites);
+}
+
+static u32 gen9_dc_mask(struct drm_i915_private *dev_priv)
+{
+	u32 mask;
+
+	mask = DC_STATE_EN_UPTO_DC5;
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		mask |= DC_STATE_EN_DC3CO | DC_STATE_EN_UPTO_DC6
+					  | DC_STATE_EN_DC9;
+	else if (DISPLAY_VER(dev_priv) == 11)
+		mask |= DC_STATE_EN_UPTO_DC6 | DC_STATE_EN_DC9;
+	else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		mask |= DC_STATE_EN_DC9;
+	else
+		mask |= DC_STATE_EN_UPTO_DC6;
+
+	return mask;
+}
+
+void gen9_sanitize_dc_state(struct drm_i915_private *dev_priv)
+{
+	u32 val;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	val = intel_de_read(dev_priv, DC_STATE_EN) & gen9_dc_mask(dev_priv);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Resetting DC state tracking from %02x to %02x\n",
+		    dev_priv->display.dmc.dc_state, val);
+	dev_priv->display.dmc.dc_state = val;
+}
+
+/**
+ * gen9_set_dc_state - set target display C power state
+ * @dev_priv: i915 device instance
+ * @state: target DC power state
+ * - DC_STATE_DISABLE
+ * - DC_STATE_EN_UPTO_DC5
+ * - DC_STATE_EN_UPTO_DC6
+ * - DC_STATE_EN_DC9
+ *
+ * Signal to DMC firmware/HW the target DC power state passed in @state.
+ * DMC/HW can turn off individual display clocks and power rails when entering
+ * a deeper DC power state (higher in number) and turns these back when exiting
+ * that state to a shallower power state (lower in number). The HW will decide
+ * when to actually enter a given state on an on-demand basis, for instance
+ * depending on the active state of display pipes. The state of display
+ * registers backed by affected power rails are saved/restored as needed.
+ *
+ * Based on the above enabling a deeper DC power state is asynchronous wrt.
+ * enabling it. Disabling a deeper power state is synchronous: for instance
+ * setting %DC_STATE_DISABLE won't complete until all HW resources are turned
+ * back on and register state is restored. This is guaranteed by the MMIO write
+ * to DC_STATE_EN blocking until the state is restored.
+ */
+void gen9_set_dc_state(struct drm_i915_private *dev_priv, u32 state)
+{
+	u32 val;
+	u32 mask;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	if (drm_WARN_ON_ONCE(&dev_priv->drm,
+			     state & ~dev_priv->display.dmc.allowed_dc_mask))
+		state &= dev_priv->display.dmc.allowed_dc_mask;
+
+	val = intel_de_read(dev_priv, DC_STATE_EN);
+	mask = gen9_dc_mask(dev_priv);
+	drm_dbg_kms(&dev_priv->drm, "Setting DC state from %02x to %02x\n",
+		    val & mask, state);
+
+	/* Check if DMC is ignoring our DC state requests */
+	if ((val & mask) != dev_priv->display.dmc.dc_state)
+		drm_err(&dev_priv->drm, "DC state mismatch (0x%x -> 0x%x)\n",
+			dev_priv->display.dmc.dc_state, val & mask);
+
+	val &= ~mask;
+	val |= state;
+
+	gen9_write_dc_state(dev_priv, val);
+
+	dev_priv->display.dmc.dc_state = val & mask;
+}
+
+static void tgl_enable_dc3co(struct drm_i915_private *dev_priv)
+{
+	drm_dbg_kms(&dev_priv->drm, "Enabling DC3CO\n");
+	gen9_set_dc_state(dev_priv, DC_STATE_EN_DC3CO);
+}
+
+static void tgl_disable_dc3co(struct drm_i915_private *dev_priv)
+{
+	u32 val;
+
+	drm_dbg_kms(&dev_priv->drm, "Disabling DC3CO\n");
+	val = intel_de_read(dev_priv, DC_STATE_EN);
+	val &= ~DC_STATE_DC3CO_STATUS;
+	intel_de_write(dev_priv, DC_STATE_EN, val);
+	gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
+	/*
+	 * Delay of 200us DC3CO Exit time B.Spec 49196
+	 */
+	usleep_range(200, 210);
+}
+
+static void assert_can_enable_dc5(struct drm_i915_private *dev_priv)
+{
+	enum i915_power_well_id high_pg;
+
+	/* Power wells at this level and above must be disabled for DC5 entry */
+	if (DISPLAY_VER(dev_priv) == 12)
+		high_pg = ICL_DISP_PW_3;
+	else
+		high_pg = SKL_DISP_PW_2;
+
+	drm_WARN_ONCE(&dev_priv->drm,
+		      intel_display_power_well_is_enabled(dev_priv, high_pg),
+		      "Power wells above platform's DC5 limit still enabled.\n");
+
+	drm_WARN_ONCE(&dev_priv->drm,
+		      (intel_de_read(dev_priv, DC_STATE_EN) &
+		       DC_STATE_EN_UPTO_DC5),
+		      "DC5 already programmed to be enabled.\n");
+	assert_rpm_wakelock_held(&dev_priv->runtime_pm);
+
+	assert_dmc_loaded(dev_priv);
+}
+
+void gen9_enable_dc5(struct drm_i915_private *dev_priv)
+{
+	assert_can_enable_dc5(dev_priv);
+
+	drm_dbg_kms(&dev_priv->drm, "Enabling DC5\n");
+
+	/* Wa Display #1183: skl,kbl,cfl */
+	if (DISPLAY_VER(dev_priv) == 9 && !IS_BROXTON(dev_priv))
+		intel_de_write(dev_priv, GEN8_CHICKEN_DCPR_1,
+			       intel_de_read(dev_priv, GEN8_CHICKEN_DCPR_1) | SKL_SELECT_ALTERNATE_DC_EXIT);
+
+	gen9_set_dc_state(dev_priv, DC_STATE_EN_UPTO_DC5);
+}
+
+static void assert_can_enable_dc6(struct drm_i915_private *dev_priv)
+{
+	drm_WARN_ONCE(&dev_priv->drm,
+		      intel_de_read(dev_priv, UTIL_PIN_CTL) & UTIL_PIN_ENABLE,
+		      "Backlight is not disabled.\n");
+	drm_WARN_ONCE(&dev_priv->drm,
+		      (intel_de_read(dev_priv, DC_STATE_EN) &
+		       DC_STATE_EN_UPTO_DC6),
+		      "DC6 already programmed to be enabled.\n");
+
+	assert_dmc_loaded(dev_priv);
+}
+
+void skl_enable_dc6(struct drm_i915_private *dev_priv)
+{
+	assert_can_enable_dc6(dev_priv);
+
+	drm_dbg_kms(&dev_priv->drm, "Enabling DC6\n");
+
+	/* Wa Display #1183: skl,kbl,cfl */
+	if (DISPLAY_VER(dev_priv) == 9 && !IS_BROXTON(dev_priv))
+		intel_de_write(dev_priv, GEN8_CHICKEN_DCPR_1,
+			       intel_de_read(dev_priv, GEN8_CHICKEN_DCPR_1) | SKL_SELECT_ALTERNATE_DC_EXIT);
+
+	gen9_set_dc_state(dev_priv, DC_STATE_EN_UPTO_DC6);
+}
+
+void bxt_enable_dc9(struct drm_i915_private *dev_priv)
+{
+	assert_can_enable_dc9(dev_priv);
+
+	drm_dbg_kms(&dev_priv->drm, "Enabling DC9\n");
+	/*
+	 * Power sequencer reset is not needed on
+	 * platforms with South Display Engine on PCH,
+	 * because PPS registers are always on.
+	 */
+	if (!HAS_PCH_SPLIT(dev_priv))
+		intel_pps_reset_all(dev_priv);
+	gen9_set_dc_state(dev_priv, DC_STATE_EN_DC9);
+}
+
+void bxt_disable_dc9(struct drm_i915_private *dev_priv)
+{
+	assert_can_disable_dc9(dev_priv);
+
+	drm_dbg_kms(&dev_priv->drm, "Disabling DC9\n");
+
+	gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
+
+	intel_pps_unlock_regs_wa(dev_priv);
+}
+
+static void hsw_power_well_sync_hw(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	const struct i915_power_well_regs *regs = power_well->desc->ops->regs;
+	int pw_idx = i915_power_well_instance(power_well)->hsw.idx;
+	u32 mask = HSW_PWR_WELL_CTL_REQ(pw_idx);
+	u32 bios_req = intel_de_read(dev_priv, regs->bios);
+
+	/* Take over the request bit if set by BIOS. */
+	if (bios_req & mask) {
+		u32 drv_req = intel_de_read(dev_priv, regs->driver);
+
+		if (!(drv_req & mask))
+			intel_de_write(dev_priv, regs->driver, drv_req | mask);
+		intel_de_write(dev_priv, regs->bios, bios_req & ~mask);
+	}
+}
+
+static void bxt_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	bxt_ddi_phy_init(dev_priv, i915_power_well_instance(power_well)->bxt.phy);
+}
+
+static void bxt_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
+					    struct i915_power_well *power_well)
+{
+	bxt_ddi_phy_uninit(dev_priv, i915_power_well_instance(power_well)->bxt.phy);
+}
+
+static bool bxt_dpio_cmn_power_well_enabled(struct drm_i915_private *dev_priv,
+					    struct i915_power_well *power_well)
+{
+	return bxt_ddi_phy_is_enabled(dev_priv, i915_power_well_instance(power_well)->bxt.phy);
+}
+
+static void bxt_verify_ddi_phy_power_wells(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_well *power_well;
+
+	power_well = lookup_power_well(dev_priv, BXT_DISP_PW_DPIO_CMN_A);
+	if (intel_power_well_refcount(power_well) > 0)
+		bxt_ddi_phy_verify_state(dev_priv, i915_power_well_instance(power_well)->bxt.phy);
+
+	power_well = lookup_power_well(dev_priv, VLV_DISP_PW_DPIO_CMN_BC);
+	if (intel_power_well_refcount(power_well) > 0)
+		bxt_ddi_phy_verify_state(dev_priv, i915_power_well_instance(power_well)->bxt.phy);
+
+	if (IS_GEMINILAKE(dev_priv)) {
+		power_well = lookup_power_well(dev_priv,
+					       GLK_DISP_PW_DPIO_CMN_C);
+		if (intel_power_well_refcount(power_well) > 0)
+			bxt_ddi_phy_verify_state(dev_priv,
+						 i915_power_well_instance(power_well)->bxt.phy);
+	}
+}
+
+static bool gen9_dc_off_power_well_enabled(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	return ((intel_de_read(dev_priv, DC_STATE_EN) & DC_STATE_EN_DC3CO) == 0 &&
+		(intel_de_read(dev_priv, DC_STATE_EN) & DC_STATE_EN_UPTO_DC5_DC6_MASK) == 0);
+}
+
+static void gen9_assert_dbuf_enabled(struct drm_i915_private *dev_priv)
+{
+	u8 hw_enabled_dbuf_slices = intel_enabled_dbuf_slices_mask(dev_priv);
+	u8 enabled_dbuf_slices = dev_priv->display.dbuf.enabled_slices;
+
+	drm_WARN(&dev_priv->drm,
+		 hw_enabled_dbuf_slices != enabled_dbuf_slices,
+		 "Unexpected DBuf power power state (0x%08x, expected 0x%08x)\n",
+		 hw_enabled_dbuf_slices,
+		 enabled_dbuf_slices);
+}
+
+void gen9_disable_dc_states(struct drm_i915_private *dev_priv)
+{
+	struct intel_cdclk_config cdclk_config = {};
+
+	if (dev_priv->display.dmc.target_dc_state == DC_STATE_EN_DC3CO) {
+		tgl_disable_dc3co(dev_priv);
+		return;
+	}
+
+	gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	intel_cdclk_get_cdclk(dev_priv, &cdclk_config);
+	/* Can't read out voltage_level so can't use intel_cdclk_changed() */
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_cdclk_needs_modeset(&dev_priv->display.cdclk.hw,
+					      &cdclk_config));
+
+	gen9_assert_dbuf_enabled(dev_priv);
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		bxt_verify_ddi_phy_power_wells(dev_priv);
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		/*
+		 * DMC retains HW context only for port A, the other combo
+		 * PHY's HW context for port B is lost after DC transitions,
+		 * so we need to restore it manually.
+		 */
+		intel_combo_phy_init(dev_priv);
+}
+
+static void gen9_dc_off_power_well_enable(struct drm_i915_private *dev_priv,
+					  struct i915_power_well *power_well)
+{
+	gen9_disable_dc_states(dev_priv);
+}
+
+static void gen9_dc_off_power_well_disable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	if (!intel_dmc_has_payload(dev_priv))
+		return;
+
+	switch (dev_priv->display.dmc.target_dc_state) {
+	case DC_STATE_EN_DC3CO:
+		tgl_enable_dc3co(dev_priv);
+		break;
+	case DC_STATE_EN_UPTO_DC6:
+		skl_enable_dc6(dev_priv);
+		break;
+	case DC_STATE_EN_UPTO_DC5:
+		gen9_enable_dc5(dev_priv);
+		break;
+	}
+}
+
+static void i9xx_power_well_sync_hw_noop(struct drm_i915_private *dev_priv,
+					 struct i915_power_well *power_well)
+{
+}
+
+static void i9xx_always_on_power_well_noop(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+}
+
+static bool i9xx_always_on_power_well_enabled(struct drm_i915_private *dev_priv,
+					     struct i915_power_well *power_well)
+{
+	return true;
+}
+
+static void i830_pipes_power_well_enable(struct drm_i915_private *dev_priv,
+					 struct i915_power_well *power_well)
+{
+	if ((intel_de_read(dev_priv, PIPECONF(PIPE_A)) & PIPECONF_ENABLE) == 0)
+		i830_enable_pipe(dev_priv, PIPE_A);
+	if ((intel_de_read(dev_priv, PIPECONF(PIPE_B)) & PIPECONF_ENABLE) == 0)
+		i830_enable_pipe(dev_priv, PIPE_B);
+}
+
+static void i830_pipes_power_well_disable(struct drm_i915_private *dev_priv,
+					  struct i915_power_well *power_well)
+{
+	i830_disable_pipe(dev_priv, PIPE_B);
+	i830_disable_pipe(dev_priv, PIPE_A);
+}
+
+static bool i830_pipes_power_well_enabled(struct drm_i915_private *dev_priv,
+					  struct i915_power_well *power_well)
+{
+	return intel_de_read(dev_priv, PIPECONF(PIPE_A)) & PIPECONF_ENABLE &&
+		intel_de_read(dev_priv, PIPECONF(PIPE_B)) & PIPECONF_ENABLE;
+}
+
+static void i830_pipes_power_well_sync_hw(struct drm_i915_private *dev_priv,
+					  struct i915_power_well *power_well)
+{
+	if (intel_power_well_refcount(power_well) > 0)
+		i830_pipes_power_well_enable(dev_priv, power_well);
+	else
+		i830_pipes_power_well_disable(dev_priv, power_well);
+}
+
+static void vlv_set_power_well(struct drm_i915_private *dev_priv,
+			       struct i915_power_well *power_well, bool enable)
+{
+	int pw_idx = i915_power_well_instance(power_well)->vlv.idx;
+	u32 mask;
+	u32 state;
+	u32 ctrl;
+
+	mask = PUNIT_PWRGT_MASK(pw_idx);
+	state = enable ? PUNIT_PWRGT_PWR_ON(pw_idx) :
+			 PUNIT_PWRGT_PWR_GATE(pw_idx);
+
+	vlv_punit_get(dev_priv);
+
+#define COND \
+	((vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_STATUS) & mask) == state)
+
+	if (COND)
+		goto out;
+
+	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL);
+	ctrl &= ~mask;
+	ctrl |= state;
+	vlv_punit_write(dev_priv, PUNIT_REG_PWRGT_CTRL, ctrl);
+
+	if (wait_for(COND, 100))
+		drm_err(&dev_priv->drm,
+			"timeout setting power well state %08x (%08x)\n",
+			state,
+			vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL));
+
+#undef COND
+
+out:
+	vlv_punit_put(dev_priv);
+}
+
+static void vlv_power_well_enable(struct drm_i915_private *dev_priv,
+				  struct i915_power_well *power_well)
+{
+	vlv_set_power_well(dev_priv, power_well, true);
+}
+
+static void vlv_power_well_disable(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	vlv_set_power_well(dev_priv, power_well, false);
+}
+
+static bool vlv_power_well_enabled(struct drm_i915_private *dev_priv,
+				   struct i915_power_well *power_well)
+{
+	int pw_idx = i915_power_well_instance(power_well)->vlv.idx;
+	bool enabled = false;
+	u32 mask;
+	u32 state;
+	u32 ctrl;
+
+	mask = PUNIT_PWRGT_MASK(pw_idx);
+	ctrl = PUNIT_PWRGT_PWR_ON(pw_idx);
+
+	vlv_punit_get(dev_priv);
+
+	state = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_STATUS) & mask;
+	/*
+	 * We only ever set the power-on and power-gate states, anything
+	 * else is unexpected.
+	 */
+	drm_WARN_ON(&dev_priv->drm, state != PUNIT_PWRGT_PWR_ON(pw_idx) &&
+		    state != PUNIT_PWRGT_PWR_GATE(pw_idx));
+	if (state == ctrl)
+		enabled = true;
+
+	/*
+	 * A transient state at this point would mean some unexpected party
+	 * is poking at the power controls too.
+	 */
+	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL) & mask;
+	drm_WARN_ON(&dev_priv->drm, ctrl != state);
+
+	vlv_punit_put(dev_priv);
+
+	return enabled;
+}
+
+static void vlv_init_display_clock_gating(struct drm_i915_private *dev_priv)
+{
+	u32 val;
+
+	/*
+	 * On driver load, a pipe may be active and driving a DSI display.
+	 * Preserve DPOUNIT_CLOCK_GATE_DISABLE to avoid the pipe getting stuck
+	 * (and never recovering) in this case. intel_dsi_post_disable() will
+	 * clear it when we turn off the display.
+	 */
+	val = intel_de_read(dev_priv, DSPCLK_GATE_D(dev_priv));
+	val &= DPOUNIT_CLOCK_GATE_DISABLE;
+	val |= VRHUNIT_CLOCK_GATE_DISABLE;
+	intel_de_write(dev_priv, DSPCLK_GATE_D(dev_priv), val);
+
+	/*
+	 * Disable trickle feed and enable pnd deadline calculation
+	 */
+	intel_de_write(dev_priv, MI_ARB_VLV,
+		       MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE);
+	intel_de_write(dev_priv, CBR1_VLV, 0);
+
+	drm_WARN_ON(&dev_priv->drm, RUNTIME_INFO(dev_priv)->rawclk_freq == 0);
+	intel_de_write(dev_priv, RAWCLK_FREQ_VLV,
+		       DIV_ROUND_CLOSEST(RUNTIME_INFO(dev_priv)->rawclk_freq,
+					 1000));
+}
+
+static void vlv_display_power_well_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+	enum pipe pipe;
+
+	/*
+	 * Enable the CRI clock source so we can get at the
+	 * display and the reference clock for VGA
+	 * hotplug / manual detection. Supposedly DSI also
+	 * needs the ref clock up and running.
+	 *
+	 * CHV DPLL B/C have some issues if VGA mode is enabled.
+	 */
+	for_each_pipe(dev_priv, pipe) {
+		u32 val = intel_de_read(dev_priv, DPLL(pipe));
+
+		val |= DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
+		if (pipe != PIPE_A)
+			val |= DPLL_INTEGRATED_CRI_CLK_VLV;
+
+		intel_de_write(dev_priv, DPLL(pipe), val);
+	}
+
+	vlv_init_display_clock_gating(dev_priv);
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	valleyview_enable_display_irqs(dev_priv);
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/*
+	 * During driver initialization/resume we can avoid restoring the
+	 * part of the HW/SW state that will be inited anyway explicitly.
+	 */
+	if (dev_priv->display.power.domains.initializing)
+		return;
+
+	intel_hpd_init(dev_priv);
+	intel_hpd_poll_disable(dev_priv);
+
+	/* Re-enable the ADPA, if we have one */
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		if (encoder->type == INTEL_OUTPUT_ANALOG)
+			intel_crt_reset(&encoder->base);
+	}
+
+	intel_vga_redisable_power_on(dev_priv);
+
+	intel_pps_unlock_regs_wa(dev_priv);
+}
+
+static void vlv_display_power_well_deinit(struct drm_i915_private *dev_priv)
+{
+	spin_lock_irq(&dev_priv->irq_lock);
+	valleyview_disable_display_irqs(dev_priv);
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/* make sure we're done processing display irqs */
+	intel_synchronize_irq(dev_priv);
+
+	intel_pps_reset_all(dev_priv);
+
+	/* Prevent us from re-enabling polling on accident in late suspend */
+	if (!dev_priv->drm.dev->power.is_suspended)
+		intel_hpd_poll_enable(dev_priv);
+}
+
+static void vlv_display_power_well_enable(struct drm_i915_private *dev_priv,
+					  struct i915_power_well *power_well)
+{
+	vlv_set_power_well(dev_priv, power_well, true);
+
+	vlv_display_power_well_init(dev_priv);
+}
+
+static void vlv_display_power_well_disable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	vlv_display_power_well_deinit(dev_priv);
+
+	vlv_set_power_well(dev_priv, power_well, false);
+}
+
+static void vlv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	/* since ref/cri clock was enabled */
+	udelay(1); /* >10ns for cmnreset, >0ns for sidereset */
+
+	vlv_set_power_well(dev_priv, power_well, true);
+
+	/*
+	 * From VLV2A0_DP_eDP_DPIO_driver_vbios_notes_10.docx -
+	 *  6.	De-assert cmn_reset/side_reset. Same as VLV X0.
+	 *   a.	GUnit 0x2110 bit[0] set to 1 (def 0)
+	 *   b.	The other bits such as sfr settings / modesel may all
+	 *	be set to 0.
+	 *
+	 * This should only be done on init and resume from S3 with
+	 * both PLLs disabled, or we risk losing DPIO and PLL
+	 * synchronization.
+	 */
+	intel_de_write(dev_priv, DPIO_CTL,
+		       intel_de_read(dev_priv, DPIO_CTL) | DPIO_CMNRST);
+}
+
+static void vlv_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
+					    struct i915_power_well *power_well)
+{
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe)
+		assert_pll_disabled(dev_priv, pipe);
+
+	/* Assert common reset */
+	intel_de_write(dev_priv, DPIO_CTL,
+		       intel_de_read(dev_priv, DPIO_CTL) & ~DPIO_CMNRST);
+
+	vlv_set_power_well(dev_priv, power_well, false);
+}
+
+#define BITS_SET(val, bits) (((val) & (bits)) == (bits))
+
+static void assert_chv_phy_status(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_well *cmn_bc =
+		lookup_power_well(dev_priv, VLV_DISP_PW_DPIO_CMN_BC);
+	struct i915_power_well *cmn_d =
+		lookup_power_well(dev_priv, CHV_DISP_PW_DPIO_CMN_D);
+	u32 phy_control = dev_priv->display.power.chv_phy_control;
+	u32 phy_status = 0;
+	u32 phy_status_mask = 0xffffffff;
+
+	/*
+	 * The BIOS can leave the PHY is some weird state
+	 * where it doesn't fully power down some parts.
+	 * Disable the asserts until the PHY has been fully
+	 * reset (ie. the power well has been disabled at
+	 * least once).
+	 */
+	if (!dev_priv->display.power.chv_phy_assert[DPIO_PHY0])
+		phy_status_mask &= ~(PHY_STATUS_CMN_LDO(DPIO_PHY0, DPIO_CH0) |
+				     PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH0, 0) |
+				     PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH0, 1) |
+				     PHY_STATUS_CMN_LDO(DPIO_PHY0, DPIO_CH1) |
+				     PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH1, 0) |
+				     PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH1, 1));
+
+	if (!dev_priv->display.power.chv_phy_assert[DPIO_PHY1])
+		phy_status_mask &= ~(PHY_STATUS_CMN_LDO(DPIO_PHY1, DPIO_CH0) |
+				     PHY_STATUS_SPLINE_LDO(DPIO_PHY1, DPIO_CH0, 0) |
+				     PHY_STATUS_SPLINE_LDO(DPIO_PHY1, DPIO_CH0, 1));
+
+	if (intel_power_well_is_enabled(dev_priv, cmn_bc)) {
+		phy_status |= PHY_POWERGOOD(DPIO_PHY0);
+
+		/* this assumes override is only used to enable lanes */
+		if ((phy_control & PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY0, DPIO_CH0)) == 0)
+			phy_control |= PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY0, DPIO_CH0);
+
+		if ((phy_control & PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY0, DPIO_CH1)) == 0)
+			phy_control |= PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY0, DPIO_CH1);
+
+		/* CL1 is on whenever anything is on in either channel */
+		if (BITS_SET(phy_control,
+			     PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY0, DPIO_CH0) |
+			     PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY0, DPIO_CH1)))
+			phy_status |= PHY_STATUS_CMN_LDO(DPIO_PHY0, DPIO_CH0);
+
+		/*
+		 * The DPLLB check accounts for the pipe B + port A usage
+		 * with CL2 powered up but all the lanes in the second channel
+		 * powered down.
+		 */
+		if (BITS_SET(phy_control,
+			     PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY0, DPIO_CH1)) &&
+		    (intel_de_read(dev_priv, DPLL(PIPE_B)) & DPLL_VCO_ENABLE) == 0)
+			phy_status |= PHY_STATUS_CMN_LDO(DPIO_PHY0, DPIO_CH1);
+
+		if (BITS_SET(phy_control,
+			     PHY_CH_POWER_DOWN_OVRD(0x3, DPIO_PHY0, DPIO_CH0)))
+			phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH0, 0);
+		if (BITS_SET(phy_control,
+			     PHY_CH_POWER_DOWN_OVRD(0xc, DPIO_PHY0, DPIO_CH0)))
+			phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH0, 1);
+
+		if (BITS_SET(phy_control,
+			     PHY_CH_POWER_DOWN_OVRD(0x3, DPIO_PHY0, DPIO_CH1)))
+			phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH1, 0);
+		if (BITS_SET(phy_control,
+			     PHY_CH_POWER_DOWN_OVRD(0xc, DPIO_PHY0, DPIO_CH1)))
+			phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH1, 1);
+	}
+
+	if (intel_power_well_is_enabled(dev_priv, cmn_d)) {
+		phy_status |= PHY_POWERGOOD(DPIO_PHY1);
+
+		/* this assumes override is only used to enable lanes */
+		if ((phy_control & PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY1, DPIO_CH0)) == 0)
+			phy_control |= PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY1, DPIO_CH0);
+
+		if (BITS_SET(phy_control,
+			     PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY1, DPIO_CH0)))
+			phy_status |= PHY_STATUS_CMN_LDO(DPIO_PHY1, DPIO_CH0);
+
+		if (BITS_SET(phy_control,
+			     PHY_CH_POWER_DOWN_OVRD(0x3, DPIO_PHY1, DPIO_CH0)))
+			phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY1, DPIO_CH0, 0);
+		if (BITS_SET(phy_control,
+			     PHY_CH_POWER_DOWN_OVRD(0xc, DPIO_PHY1, DPIO_CH0)))
+			phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY1, DPIO_CH0, 1);
+	}
+
+	phy_status &= phy_status_mask;
+
+	/*
+	 * The PHY may be busy with some initial calibration and whatnot,
+	 * so the power state can take a while to actually change.
+	 */
+	if (intel_de_wait_for_register(dev_priv, DISPLAY_PHY_STATUS,
+				       phy_status_mask, phy_status, 10))
+		drm_err(&dev_priv->drm,
+			"Unexpected PHY_STATUS 0x%08x, expected 0x%08x (PHY_CONTROL=0x%08x)\n",
+			intel_de_read(dev_priv, DISPLAY_PHY_STATUS) & phy_status_mask,
+			phy_status, dev_priv->display.power.chv_phy_control);
+}
+
+#undef BITS_SET
+
+static void chv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
+					   struct i915_power_well *power_well)
+{
+	enum i915_power_well_id id = i915_power_well_instance(power_well)->id;
+	enum dpio_phy phy;
+	enum pipe pipe;
+	u32 tmp;
+
+	drm_WARN_ON_ONCE(&dev_priv->drm,
+			 id != VLV_DISP_PW_DPIO_CMN_BC &&
+			 id != CHV_DISP_PW_DPIO_CMN_D);
+
+	if (id == VLV_DISP_PW_DPIO_CMN_BC) {
+		pipe = PIPE_A;
+		phy = DPIO_PHY0;
+	} else {
+		pipe = PIPE_C;
+		phy = DPIO_PHY1;
+	}
+
+	/* since ref/cri clock was enabled */
+	udelay(1); /* >10ns for cmnreset, >0ns for sidereset */
+	vlv_set_power_well(dev_priv, power_well, true);
+
+	/* Poll for phypwrgood signal */
+	if (intel_de_wait_for_set(dev_priv, DISPLAY_PHY_STATUS,
+				  PHY_POWERGOOD(phy), 1))
+		drm_err(&dev_priv->drm, "Display PHY %d is not power up\n",
+			phy);
+
+	vlv_dpio_get(dev_priv);
+
+	/* Enable dynamic power down */
+	tmp = vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW28);
+	tmp |= DPIO_DYNPWRDOWNEN_CH0 | DPIO_CL1POWERDOWNEN |
+		DPIO_SUS_CLK_CONFIG_GATE_CLKREQ;
+	vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW28, tmp);
+
+	if (id == VLV_DISP_PW_DPIO_CMN_BC) {
+		tmp = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW6_CH1);
+		tmp |= DPIO_DYNPWRDOWNEN_CH1;
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW6_CH1, tmp);
+	} else {
+		/*
+		 * Force the non-existing CL2 off. BXT does this
+		 * too, so maybe it saves some power even though
+		 * CL2 doesn't exist?
+		 */
+		tmp = vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW30);
+		tmp |= DPIO_CL2_LDOFUSE_PWRENB;
+		vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW30, tmp);
+	}
+
+	vlv_dpio_put(dev_priv);
+
+	dev_priv->display.power.chv_phy_control |= PHY_COM_LANE_RESET_DEASSERT(phy);
+	intel_de_write(dev_priv, DISPLAY_PHY_CONTROL,
+		       dev_priv->display.power.chv_phy_control);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Enabled DPIO PHY%d (PHY_CONTROL=0x%08x)\n",
+		    phy, dev_priv->display.power.chv_phy_control);
+
+	assert_chv_phy_status(dev_priv);
+}
+
+static void chv_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
+					    struct i915_power_well *power_well)
+{
+	enum i915_power_well_id id = i915_power_well_instance(power_well)->id;
+	enum dpio_phy phy;
+
+	drm_WARN_ON_ONCE(&dev_priv->drm,
+			 id != VLV_DISP_PW_DPIO_CMN_BC &&
+			 id != CHV_DISP_PW_DPIO_CMN_D);
+
+	if (id == VLV_DISP_PW_DPIO_CMN_BC) {
+		phy = DPIO_PHY0;
+		assert_pll_disabled(dev_priv, PIPE_A);
+		assert_pll_disabled(dev_priv, PIPE_B);
+	} else {
+		phy = DPIO_PHY1;
+		assert_pll_disabled(dev_priv, PIPE_C);
+	}
+
+	dev_priv->display.power.chv_phy_control &= ~PHY_COM_LANE_RESET_DEASSERT(phy);
+	intel_de_write(dev_priv, DISPLAY_PHY_CONTROL,
+		       dev_priv->display.power.chv_phy_control);
+
+	vlv_set_power_well(dev_priv, power_well, false);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Disabled DPIO PHY%d (PHY_CONTROL=0x%08x)\n",
+		    phy, dev_priv->display.power.chv_phy_control);
+
+	/* PHY is fully reset now, so we can enable the PHY state asserts */
+	dev_priv->display.power.chv_phy_assert[phy] = true;
+
+	assert_chv_phy_status(dev_priv);
+}
+
+static void assert_chv_phy_powergate(struct drm_i915_private *dev_priv, enum dpio_phy phy,
+				     enum dpio_channel ch, bool override, unsigned int mask)
+{
+	enum pipe pipe = phy == DPIO_PHY0 ? PIPE_A : PIPE_C;
+	u32 reg, val, expected, actual;
+
+	/*
+	 * The BIOS can leave the PHY is some weird state
+	 * where it doesn't fully power down some parts.
+	 * Disable the asserts until the PHY has been fully
+	 * reset (ie. the power well has been disabled at
+	 * least once).
+	 */
+	if (!dev_priv->display.power.chv_phy_assert[phy])
+		return;
+
+	if (ch == DPIO_CH0)
+		reg = _CHV_CMN_DW0_CH0;
+	else
+		reg = _CHV_CMN_DW6_CH1;
+
+	vlv_dpio_get(dev_priv);
+	val = vlv_dpio_read(dev_priv, pipe, reg);
+	vlv_dpio_put(dev_priv);
+
+	/*
+	 * This assumes !override is only used when the port is disabled.
+	 * All lanes should power down even without the override when
+	 * the port is disabled.
+	 */
+	if (!override || mask == 0xf) {
+		expected = DPIO_ALLDL_POWERDOWN | DPIO_ANYDL_POWERDOWN;
+		/*
+		 * If CH1 common lane is not active anymore
+		 * (eg. for pipe B DPLL) the entire channel will
+		 * shut down, which causes the common lane registers
+		 * to read as 0. That means we can't actually check
+		 * the lane power down status bits, but as the entire
+		 * register reads as 0 it's a good indication that the
+		 * channel is indeed entirely powered down.
+		 */
+		if (ch == DPIO_CH1 && val == 0)
+			expected = 0;
+	} else if (mask != 0x0) {
+		expected = DPIO_ANYDL_POWERDOWN;
+	} else {
+		expected = 0;
+	}
+
+	if (ch == DPIO_CH0)
+		actual = val >> DPIO_ANYDL_POWERDOWN_SHIFT_CH0;
+	else
+		actual = val >> DPIO_ANYDL_POWERDOWN_SHIFT_CH1;
+	actual &= DPIO_ALLDL_POWERDOWN | DPIO_ANYDL_POWERDOWN;
+
+	drm_WARN(&dev_priv->drm, actual != expected,
+		 "Unexpected DPIO lane power down: all %d, any %d. Expected: all %d, any %d. (0x%x = 0x%08x)\n",
+		 !!(actual & DPIO_ALLDL_POWERDOWN),
+		 !!(actual & DPIO_ANYDL_POWERDOWN),
+		 !!(expected & DPIO_ALLDL_POWERDOWN),
+		 !!(expected & DPIO_ANYDL_POWERDOWN),
+		 reg, val);
+}
+
+bool chv_phy_powergate_ch(struct drm_i915_private *dev_priv, enum dpio_phy phy,
+			  enum dpio_channel ch, bool override)
+{
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	bool was_override;
+
+	mutex_lock(&power_domains->lock);
+
+	was_override = dev_priv->display.power.chv_phy_control & PHY_CH_POWER_DOWN_OVRD_EN(phy, ch);
+
+	if (override == was_override)
+		goto out;
+
+	if (override)
+		dev_priv->display.power.chv_phy_control |= PHY_CH_POWER_DOWN_OVRD_EN(phy, ch);
+	else
+		dev_priv->display.power.chv_phy_control &= ~PHY_CH_POWER_DOWN_OVRD_EN(phy, ch);
+
+	intel_de_write(dev_priv, DISPLAY_PHY_CONTROL,
+		       dev_priv->display.power.chv_phy_control);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Power gating DPIO PHY%d CH%d (DPIO_PHY_CONTROL=0x%08x)\n",
+		    phy, ch, dev_priv->display.power.chv_phy_control);
+
+	assert_chv_phy_status(dev_priv);
+
+out:
+	mutex_unlock(&power_domains->lock);
+
+	return was_override;
+}
+
+void chv_phy_powergate_lanes(struct intel_encoder *encoder,
+			     bool override, unsigned int mask)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct i915_power_domains *power_domains = &dev_priv->display.power.domains;
+	enum dpio_phy phy = vlv_dig_port_to_phy(enc_to_dig_port(encoder));
+	enum dpio_channel ch = vlv_dig_port_to_channel(enc_to_dig_port(encoder));
+
+	mutex_lock(&power_domains->lock);
+
+	dev_priv->display.power.chv_phy_control &= ~PHY_CH_POWER_DOWN_OVRD(0xf, phy, ch);
+	dev_priv->display.power.chv_phy_control |= PHY_CH_POWER_DOWN_OVRD(mask, phy, ch);
+
+	if (override)
+		dev_priv->display.power.chv_phy_control |= PHY_CH_POWER_DOWN_OVRD_EN(phy, ch);
+	else
+		dev_priv->display.power.chv_phy_control &= ~PHY_CH_POWER_DOWN_OVRD_EN(phy, ch);
+
+	intel_de_write(dev_priv, DISPLAY_PHY_CONTROL,
+		       dev_priv->display.power.chv_phy_control);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Power gating DPIO PHY%d CH%d lanes 0x%x (PHY_CONTROL=0x%08x)\n",
+		    phy, ch, mask, dev_priv->display.power.chv_phy_control);
+
+	assert_chv_phy_status(dev_priv);
+
+	assert_chv_phy_powergate(dev_priv, phy, ch, override, mask);
+
+	mutex_unlock(&power_domains->lock);
+}
+
+static bool chv_pipe_power_well_enabled(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well)
+{
+	enum pipe pipe = PIPE_A;
+	bool enabled;
+	u32 state, ctrl;
+
+	vlv_punit_get(dev_priv);
+
+	state = vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM) & DP_SSS_MASK(pipe);
+	/*
+	 * We only ever set the power-on and power-gate states, anything
+	 * else is unexpected.
+	 */
+	drm_WARN_ON(&dev_priv->drm, state != DP_SSS_PWR_ON(pipe) &&
+		    state != DP_SSS_PWR_GATE(pipe));
+	enabled = state == DP_SSS_PWR_ON(pipe);
+
+	/*
+	 * A transient state at this point would mean some unexpected party
+	 * is poking at the power controls too.
+	 */
+	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM) & DP_SSC_MASK(pipe);
+	drm_WARN_ON(&dev_priv->drm, ctrl << 16 != state);
+
+	vlv_punit_put(dev_priv);
+
+	return enabled;
+}
+
+static void chv_set_pipe_power_well(struct drm_i915_private *dev_priv,
+				    struct i915_power_well *power_well,
+				    bool enable)
+{
+	enum pipe pipe = PIPE_A;
+	u32 state;
+	u32 ctrl;
+
+	state = enable ? DP_SSS_PWR_ON(pipe) : DP_SSS_PWR_GATE(pipe);
+
+	vlv_punit_get(dev_priv);
+
+#define COND \
+	((vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM) & DP_SSS_MASK(pipe)) == state)
+
+	if (COND)
+		goto out;
+
+	ctrl = vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM);
+	ctrl &= ~DP_SSC_MASK(pipe);
+	ctrl |= enable ? DP_SSC_PWR_ON(pipe) : DP_SSC_PWR_GATE(pipe);
+	vlv_punit_write(dev_priv, PUNIT_REG_DSPSSPM, ctrl);
+
+	if (wait_for(COND, 100))
+		drm_err(&dev_priv->drm,
+			"timeout setting power well state %08x (%08x)\n",
+			state,
+			vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM));
+
+#undef COND
+
+out:
+	vlv_punit_put(dev_priv);
+}
+
+static void chv_pipe_power_well_sync_hw(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well)
+{
+	intel_de_write(dev_priv, DISPLAY_PHY_CONTROL,
+		       dev_priv->display.power.chv_phy_control);
+}
+
+static void chv_pipe_power_well_enable(struct drm_i915_private *dev_priv,
+				       struct i915_power_well *power_well)
+{
+	chv_set_pipe_power_well(dev_priv, power_well, true);
+
+	vlv_display_power_well_init(dev_priv);
+}
+
+static void chv_pipe_power_well_disable(struct drm_i915_private *dev_priv,
+					struct i915_power_well *power_well)
+{
+	vlv_display_power_well_deinit(dev_priv);
+
+	chv_set_pipe_power_well(dev_priv, power_well, false);
+}
+
+static void
+tgl_tc_cold_request(struct drm_i915_private *i915, bool block)
+{
+	u8 tries = 0;
+	int ret;
+
+	while (1) {
+		u32 low_val;
+		u32 high_val = 0;
+
+		if (block)
+			low_val = TGL_PCODE_EXIT_TCCOLD_DATA_L_BLOCK_REQ;
+		else
+			low_val = TGL_PCODE_EXIT_TCCOLD_DATA_L_UNBLOCK_REQ;
+
+		/*
+		 * Spec states that we should timeout the request after 200us
+		 * but the function below will timeout after 500us
+		 */
+		ret = snb_pcode_read(&i915->uncore, TGL_PCODE_TCCOLD, &low_val, &high_val);
+		if (ret == 0) {
+			if (block &&
+			    (low_val & TGL_PCODE_EXIT_TCCOLD_DATA_L_EXIT_FAILED))
+				ret = -EIO;
+			else
+				break;
+		}
+
+		if (++tries == 3)
+			break;
+
+		msleep(1);
+	}
+
+	if (ret)
+		drm_err(&i915->drm, "TC cold %sblock failed\n",
+			block ? "" : "un");
+	else
+		drm_dbg_kms(&i915->drm, "TC cold %sblock succeeded\n",
+			    block ? "" : "un");
+}
+
+static void
+tgl_tc_cold_off_power_well_enable(struct drm_i915_private *i915,
+				  struct i915_power_well *power_well)
+{
+	tgl_tc_cold_request(i915, true);
+}
+
+static void
+tgl_tc_cold_off_power_well_disable(struct drm_i915_private *i915,
+				   struct i915_power_well *power_well)
+{
+	tgl_tc_cold_request(i915, false);
+}
+
+static void
+tgl_tc_cold_off_power_well_sync_hw(struct drm_i915_private *i915,
+				   struct i915_power_well *power_well)
+{
+	if (intel_power_well_refcount(power_well) > 0)
+		tgl_tc_cold_off_power_well_enable(i915, power_well);
+	else
+		tgl_tc_cold_off_power_well_disable(i915, power_well);
+}
+
+static bool
+tgl_tc_cold_off_power_well_is_enabled(struct drm_i915_private *dev_priv,
+				      struct i915_power_well *power_well)
+{
+	/*
+	 * Not the correctly implementation but there is no way to just read it
+	 * from PCODE, so returning count to avoid state mismatch errors
+	 */
+	return intel_power_well_refcount(power_well);
+}
+
+static void xelpdp_aux_power_well_enable(struct drm_i915_private *dev_priv,
+					 struct i915_power_well *power_well)
+{
+	enum aux_ch aux_ch = i915_power_well_instance(power_well)->xelpdp.aux_ch;
+
+	intel_de_rmw(dev_priv, XELPDP_DP_AUX_CH_CTL(aux_ch),
+		     XELPDP_DP_AUX_CH_CTL_POWER_REQUEST,
+		     XELPDP_DP_AUX_CH_CTL_POWER_REQUEST);
+
+	/*
+	 * The power status flag cannot be used to determine whether aux
+	 * power wells have finished powering up.  Instead we're
+	 * expected to just wait a fixed 600us after raising the request
+	 * bit.
+	 */
+	usleep_range(600, 1200);
+}
+
+static void xelpdp_aux_power_well_disable(struct drm_i915_private *dev_priv,
+					  struct i915_power_well *power_well)
+{
+	enum aux_ch aux_ch = i915_power_well_instance(power_well)->xelpdp.aux_ch;
+
+	intel_de_rmw(dev_priv, XELPDP_DP_AUX_CH_CTL(aux_ch),
+		     XELPDP_DP_AUX_CH_CTL_POWER_REQUEST,
+		     0);
+	usleep_range(10, 30);
+}
+
+static bool xelpdp_aux_power_well_enabled(struct drm_i915_private *dev_priv,
+					  struct i915_power_well *power_well)
+{
+	enum aux_ch aux_ch = i915_power_well_instance(power_well)->xelpdp.aux_ch;
+
+	return intel_de_read(dev_priv, XELPDP_DP_AUX_CH_CTL(aux_ch)) &
+		XELPDP_DP_AUX_CH_CTL_POWER_STATUS;
+}
+
+const struct i915_power_well_ops i9xx_always_on_power_well_ops = {
+	.sync_hw = i9xx_power_well_sync_hw_noop,
+	.enable = i9xx_always_on_power_well_noop,
+	.disable = i9xx_always_on_power_well_noop,
+	.is_enabled = i9xx_always_on_power_well_enabled,
+};
+
+const struct i915_power_well_ops chv_pipe_power_well_ops = {
+	.sync_hw = chv_pipe_power_well_sync_hw,
+	.enable = chv_pipe_power_well_enable,
+	.disable = chv_pipe_power_well_disable,
+	.is_enabled = chv_pipe_power_well_enabled,
+};
+
+const struct i915_power_well_ops chv_dpio_cmn_power_well_ops = {
+	.sync_hw = i9xx_power_well_sync_hw_noop,
+	.enable = chv_dpio_cmn_power_well_enable,
+	.disable = chv_dpio_cmn_power_well_disable,
+	.is_enabled = vlv_power_well_enabled,
+};
+
+const struct i915_power_well_ops i830_pipes_power_well_ops = {
+	.sync_hw = i830_pipes_power_well_sync_hw,
+	.enable = i830_pipes_power_well_enable,
+	.disable = i830_pipes_power_well_disable,
+	.is_enabled = i830_pipes_power_well_enabled,
+};
+
+static const struct i915_power_well_regs hsw_power_well_regs = {
+	.bios	= HSW_PWR_WELL_CTL1,
+	.driver	= HSW_PWR_WELL_CTL2,
+	.kvmr	= HSW_PWR_WELL_CTL3,
+	.debug	= HSW_PWR_WELL_CTL4,
+};
+
+const struct i915_power_well_ops hsw_power_well_ops = {
+	.regs = &hsw_power_well_regs,
+	.sync_hw = hsw_power_well_sync_hw,
+	.enable = hsw_power_well_enable,
+	.disable = hsw_power_well_disable,
+	.is_enabled = hsw_power_well_enabled,
+};
+
+const struct i915_power_well_ops gen9_dc_off_power_well_ops = {
+	.sync_hw = i9xx_power_well_sync_hw_noop,
+	.enable = gen9_dc_off_power_well_enable,
+	.disable = gen9_dc_off_power_well_disable,
+	.is_enabled = gen9_dc_off_power_well_enabled,
+};
+
+const struct i915_power_well_ops bxt_dpio_cmn_power_well_ops = {
+	.sync_hw = i9xx_power_well_sync_hw_noop,
+	.enable = bxt_dpio_cmn_power_well_enable,
+	.disable = bxt_dpio_cmn_power_well_disable,
+	.is_enabled = bxt_dpio_cmn_power_well_enabled,
+};
+
+const struct i915_power_well_ops vlv_display_power_well_ops = {
+	.sync_hw = i9xx_power_well_sync_hw_noop,
+	.enable = vlv_display_power_well_enable,
+	.disable = vlv_display_power_well_disable,
+	.is_enabled = vlv_power_well_enabled,
+};
+
+const struct i915_power_well_ops vlv_dpio_cmn_power_well_ops = {
+	.sync_hw = i9xx_power_well_sync_hw_noop,
+	.enable = vlv_dpio_cmn_power_well_enable,
+	.disable = vlv_dpio_cmn_power_well_disable,
+	.is_enabled = vlv_power_well_enabled,
+};
+
+const struct i915_power_well_ops vlv_dpio_power_well_ops = {
+	.sync_hw = i9xx_power_well_sync_hw_noop,
+	.enable = vlv_power_well_enable,
+	.disable = vlv_power_well_disable,
+	.is_enabled = vlv_power_well_enabled,
+};
+
+static const struct i915_power_well_regs icl_aux_power_well_regs = {
+	.bios	= ICL_PWR_WELL_CTL_AUX1,
+	.driver	= ICL_PWR_WELL_CTL_AUX2,
+	.debug	= ICL_PWR_WELL_CTL_AUX4,
+};
+
+const struct i915_power_well_ops icl_aux_power_well_ops = {
+	.regs = &icl_aux_power_well_regs,
+	.sync_hw = hsw_power_well_sync_hw,
+	.enable = icl_aux_power_well_enable,
+	.disable = icl_aux_power_well_disable,
+	.is_enabled = hsw_power_well_enabled,
+};
+
+static const struct i915_power_well_regs icl_ddi_power_well_regs = {
+	.bios	= ICL_PWR_WELL_CTL_DDI1,
+	.driver	= ICL_PWR_WELL_CTL_DDI2,
+	.debug	= ICL_PWR_WELL_CTL_DDI4,
+};
+
+const struct i915_power_well_ops icl_ddi_power_well_ops = {
+	.regs = &icl_ddi_power_well_regs,
+	.sync_hw = hsw_power_well_sync_hw,
+	.enable = hsw_power_well_enable,
+	.disable = hsw_power_well_disable,
+	.is_enabled = hsw_power_well_enabled,
+};
+
+const struct i915_power_well_ops tgl_tc_cold_off_ops = {
+	.sync_hw = tgl_tc_cold_off_power_well_sync_hw,
+	.enable = tgl_tc_cold_off_power_well_enable,
+	.disable = tgl_tc_cold_off_power_well_disable,
+	.is_enabled = tgl_tc_cold_off_power_well_is_enabled,
+};
+
+const struct i915_power_well_ops xelpdp_aux_power_well_ops = {
+	.sync_hw = i9xx_power_well_sync_hw_noop,
+	.enable = xelpdp_aux_power_well_enable,
+	.disable = xelpdp_aux_power_well_disable,
+	.is_enabled = xelpdp_aux_power_well_enabled,
+};
diff --git a/drivers/gpu/drm/i915/display/intel_display_power_well.h b/drivers/gpu/drm/i915/display/intel_display_power_well.h
new file mode 100644
index 000000000..e13b521e3
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_power_well.h
@@ -0,0 +1,177 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+#ifndef __INTEL_DISPLAY_POWER_WELL_H__
+#define __INTEL_DISPLAY_POWER_WELL_H__
+
+#include <linux/types.h>
+
+#include "intel_display.h"
+#include "intel_display_power.h"
+
+struct drm_i915_private;
+struct i915_power_well;
+
+#define for_each_power_well(__dev_priv, __power_well)				\
+	for ((__power_well) = (__dev_priv)->display.power.domains.power_wells;	\
+	     (__power_well) - (__dev_priv)->display.power.domains.power_wells <	\
+		(__dev_priv)->display.power.domains.power_well_count;		\
+	     (__power_well)++)
+
+#define for_each_power_well_reverse(__dev_priv, __power_well)			\
+	for ((__power_well) = (__dev_priv)->display.power.domains.power_wells +		\
+			      (__dev_priv)->display.power.domains.power_well_count - 1;	\
+	     (__power_well) - (__dev_priv)->display.power.domains.power_wells >= 0;	\
+	     (__power_well)--)
+
+/*
+ * i915_power_well_id:
+ *
+ * IDs used to look up power wells. Power wells accessed directly bypassing
+ * the power domains framework must be assigned a unique ID. The rest of power
+ * wells must be assigned DISP_PW_ID_NONE.
+ */
+enum i915_power_well_id {
+	DISP_PW_ID_NONE = 0,		/* must be kept zero */
+
+	VLV_DISP_PW_DISP2D,
+	BXT_DISP_PW_DPIO_CMN_A,
+	VLV_DISP_PW_DPIO_CMN_BC,
+	GLK_DISP_PW_DPIO_CMN_C,
+	CHV_DISP_PW_DPIO_CMN_D,
+	HSW_DISP_PW_GLOBAL,
+	SKL_DISP_PW_MISC_IO,
+	SKL_DISP_PW_1,
+	SKL_DISP_PW_2,
+	ICL_DISP_PW_3,
+	SKL_DISP_DC_OFF,
+	TGL_DISP_PW_TC_COLD_OFF,
+};
+
+struct i915_power_well_instance {
+	const char *name;
+	const struct i915_power_domain_list {
+		const enum intel_display_power_domain *list;
+		u8 count;
+	} *domain_list;
+
+	/* unique identifier for this power well */
+	enum i915_power_well_id id;
+	/*
+	 * Arbitraty data associated with this power well. Platform and power
+	 * well specific.
+	 */
+	union {
+		struct {
+			/*
+			 * request/status flag index in the PUNIT power well
+			 * control/status registers.
+			 */
+			u8 idx;
+		} vlv;
+		struct {
+			enum dpio_phy phy;
+		} bxt;
+		struct {
+			/*
+			 * request/status flag index in the power well
+			 * constrol/status registers.
+			 */
+			u8 idx;
+		} hsw;
+		struct {
+			u8 aux_ch;
+		} xelpdp;
+	};
+};
+
+struct i915_power_well_desc {
+	const struct i915_power_well_ops *ops;
+	const struct i915_power_well_instance_list {
+		const struct i915_power_well_instance *list;
+		u8 count;
+	} *instances;
+
+	/* Mask of pipes whose IRQ logic is backed by the pw */
+	u16 irq_pipe_mask:4;
+	u16 always_on:1;
+	/*
+	 * Instead of waiting for the status bit to ack enables,
+	 * just wait a specific amount of time and then consider
+	 * the well enabled.
+	 */
+	u16 fixed_enable_delay:1;
+	/* The pw is backing the VGA functionality */
+	u16 has_vga:1;
+	u16 has_fuses:1;
+	/*
+	 * The pw is for an ICL+ TypeC PHY port in
+	 * Thunderbolt mode.
+	 */
+	u16 is_tc_tbt:1;
+};
+
+struct i915_power_well {
+	const struct i915_power_well_desc *desc;
+	struct intel_power_domain_mask domains;
+	/* power well enable/disable usage count */
+	int count;
+	/* cached hw enabled state */
+	bool hw_enabled;
+	/* index into desc->instances->list */
+	u8 instance_idx;
+};
+
+struct i915_power_well *lookup_power_well(struct drm_i915_private *i915,
+					  enum i915_power_well_id id);
+
+void intel_power_well_enable(struct drm_i915_private *i915,
+			     struct i915_power_well *power_well);
+void intel_power_well_disable(struct drm_i915_private *i915,
+			      struct i915_power_well *power_well);
+void intel_power_well_sync_hw(struct drm_i915_private *i915,
+			      struct i915_power_well *power_well);
+void intel_power_well_get(struct drm_i915_private *i915,
+			  struct i915_power_well *power_well);
+void intel_power_well_put(struct drm_i915_private *i915,
+			  struct i915_power_well *power_well);
+bool intel_power_well_is_enabled(struct drm_i915_private *i915,
+				 struct i915_power_well *power_well);
+bool intel_power_well_is_enabled_cached(struct i915_power_well *power_well);
+bool intel_display_power_well_is_enabled(struct drm_i915_private *dev_priv,
+					 enum i915_power_well_id power_well_id);
+bool intel_power_well_is_always_on(struct i915_power_well *power_well);
+const char *intel_power_well_name(struct i915_power_well *power_well);
+struct intel_power_domain_mask *intel_power_well_domains(struct i915_power_well *power_well);
+int intel_power_well_refcount(struct i915_power_well *power_well);
+
+void chv_phy_powergate_lanes(struct intel_encoder *encoder,
+			     bool override, unsigned int mask);
+bool chv_phy_powergate_ch(struct drm_i915_private *dev_priv, enum dpio_phy phy,
+			  enum dpio_channel ch, bool override);
+
+void gen9_enable_dc5(struct drm_i915_private *dev_priv);
+void skl_enable_dc6(struct drm_i915_private *dev_priv);
+void gen9_sanitize_dc_state(struct drm_i915_private *dev_priv);
+void gen9_set_dc_state(struct drm_i915_private *dev_priv, u32 state);
+void gen9_disable_dc_states(struct drm_i915_private *dev_priv);
+void bxt_enable_dc9(struct drm_i915_private *dev_priv);
+void bxt_disable_dc9(struct drm_i915_private *dev_priv);
+
+extern const struct i915_power_well_ops i9xx_always_on_power_well_ops;
+extern const struct i915_power_well_ops chv_pipe_power_well_ops;
+extern const struct i915_power_well_ops chv_dpio_cmn_power_well_ops;
+extern const struct i915_power_well_ops i830_pipes_power_well_ops;
+extern const struct i915_power_well_ops hsw_power_well_ops;
+extern const struct i915_power_well_ops gen9_dc_off_power_well_ops;
+extern const struct i915_power_well_ops bxt_dpio_cmn_power_well_ops;
+extern const struct i915_power_well_ops vlv_display_power_well_ops;
+extern const struct i915_power_well_ops vlv_dpio_cmn_power_well_ops;
+extern const struct i915_power_well_ops vlv_dpio_power_well_ops;
+extern const struct i915_power_well_ops icl_aux_power_well_ops;
+extern const struct i915_power_well_ops icl_ddi_power_well_ops;
+extern const struct i915_power_well_ops tgl_tc_cold_off_ops;
+extern const struct i915_power_well_ops xelpdp_aux_power_well_ops;
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_display_trace.c b/drivers/gpu/drm/i915/display/intel_display_trace.c
new file mode 100644
index 000000000..737979ada
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_trace.c
@@ -0,0 +1,9 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __CHECKER__
+#define CREATE_TRACE_POINTS
+#include "intel_display_trace.h"
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_display_trace.h b/drivers/gpu/drm/i915/display/intel_display_trace.h
new file mode 100644
index 000000000..2dd5a4b7f
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_trace.h
@@ -0,0 +1,589 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM i915
+
+#if !defined(__INTEL_DISPLAY_TRACE_H__) || defined(TRACE_HEADER_MULTI_READ)
+#define __INTEL_DISPLAY_TRACE_H__
+
+#include <linux/string_helpers.h>
+#include <linux/types.h>
+#include <linux/tracepoint.h>
+
+#include "i915_drv.h"
+#include "i915_irq.h"
+#include "intel_crtc.h"
+#include "intel_display_types.h"
+
+TRACE_EVENT(intel_pipe_enable,
+	    TP_PROTO(struct intel_crtc *crtc),
+	    TP_ARGS(crtc),
+
+	    TP_STRUCT__entry(
+			     __array(u32, frame, 3)
+			     __array(u32, scanline, 3)
+			     __field(enum pipe, pipe)
+			     ),
+	    TP_fast_assign(
+			   struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+			   struct intel_crtc *it__;
+			   for_each_intel_crtc(&dev_priv->drm, it__) {
+				   __entry->frame[it__->pipe] = intel_crtc_get_vblank_counter(it__);
+				   __entry->scanline[it__->pipe] = intel_get_crtc_scanline(it__);
+			   }
+			   __entry->pipe = crtc->pipe;
+			   ),
+
+	    TP_printk("pipe %c enable, pipe A: frame=%u, scanline=%u, pipe B: frame=%u, scanline=%u, pipe C: frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe),
+		      __entry->frame[PIPE_A], __entry->scanline[PIPE_A],
+		      __entry->frame[PIPE_B], __entry->scanline[PIPE_B],
+		      __entry->frame[PIPE_C], __entry->scanline[PIPE_C])
+);
+
+TRACE_EVENT(intel_pipe_disable,
+	    TP_PROTO(struct intel_crtc *crtc),
+	    TP_ARGS(crtc),
+
+	    TP_STRUCT__entry(
+			     __array(u32, frame, 3)
+			     __array(u32, scanline, 3)
+			     __field(enum pipe, pipe)
+			     ),
+
+	    TP_fast_assign(
+			   struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+			   struct intel_crtc *it__;
+			   for_each_intel_crtc(&dev_priv->drm, it__) {
+				   __entry->frame[it__->pipe] = intel_crtc_get_vblank_counter(it__);
+				   __entry->scanline[it__->pipe] = intel_get_crtc_scanline(it__);
+			   }
+			   __entry->pipe = crtc->pipe;
+			   ),
+
+	    TP_printk("pipe %c disable, pipe A: frame=%u, scanline=%u, pipe B: frame=%u, scanline=%u, pipe C: frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe),
+		      __entry->frame[PIPE_A], __entry->scanline[PIPE_A],
+		      __entry->frame[PIPE_B], __entry->scanline[PIPE_B],
+		      __entry->frame[PIPE_C], __entry->scanline[PIPE_C])
+);
+
+TRACE_EVENT(intel_pipe_crc,
+	    TP_PROTO(struct intel_crtc *crtc, const u32 *crcs),
+	    TP_ARGS(crtc, crcs),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __array(u32, crcs, 5)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   memcpy(__entry->crcs, crcs, sizeof(__entry->crcs));
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u crc=%08x %08x %08x %08x %08x",
+		      pipe_name(__entry->pipe), __entry->frame, __entry->scanline,
+		      __entry->crcs[0], __entry->crcs[1], __entry->crcs[2],
+		      __entry->crcs[3], __entry->crcs[4])
+);
+
+TRACE_EVENT(intel_cpu_fifo_underrun,
+	    TP_PROTO(struct drm_i915_private *dev_priv, enum pipe pipe),
+	    TP_ARGS(dev_priv, pipe),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			    struct intel_crtc *crtc = intel_crtc_for_pipe(dev_priv, pipe);
+			   __entry->pipe = pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe),
+		      __entry->frame, __entry->scanline)
+);
+
+TRACE_EVENT(intel_pch_fifo_underrun,
+	    TP_PROTO(struct drm_i915_private *dev_priv, enum pipe pch_transcoder),
+	    TP_ARGS(dev_priv, pch_transcoder),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			   enum pipe pipe = pch_transcoder;
+			   struct intel_crtc *crtc = intel_crtc_for_pipe(dev_priv, pipe);
+			   __entry->pipe = pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   ),
+
+	    TP_printk("pch transcoder %c, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe),
+		      __entry->frame, __entry->scanline)
+);
+
+TRACE_EVENT(intel_memory_cxsr,
+	    TP_PROTO(struct drm_i915_private *dev_priv, bool old, bool new),
+	    TP_ARGS(dev_priv, old, new),
+
+	    TP_STRUCT__entry(
+			     __array(u32, frame, 3)
+			     __array(u32, scanline, 3)
+			     __field(bool, old)
+			     __field(bool, new)
+			     ),
+
+	    TP_fast_assign(
+			   struct intel_crtc *crtc;
+			   for_each_intel_crtc(&dev_priv->drm, crtc) {
+				   __entry->frame[crtc->pipe] = intel_crtc_get_vblank_counter(crtc);
+				   __entry->scanline[crtc->pipe] = intel_get_crtc_scanline(crtc);
+			   }
+			   __entry->old = old;
+			   __entry->new = new;
+			   ),
+
+	    TP_printk("%s->%s, pipe A: frame=%u, scanline=%u, pipe B: frame=%u, scanline=%u, pipe C: frame=%u, scanline=%u",
+		      str_on_off(__entry->old), str_on_off(__entry->new),
+		      __entry->frame[PIPE_A], __entry->scanline[PIPE_A],
+		      __entry->frame[PIPE_B], __entry->scanline[PIPE_B],
+		      __entry->frame[PIPE_C], __entry->scanline[PIPE_C])
+);
+
+TRACE_EVENT(g4x_wm,
+	    TP_PROTO(struct intel_crtc *crtc, const struct g4x_wm_values *wm),
+	    TP_ARGS(crtc, wm),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __field(u16, primary)
+			     __field(u16, sprite)
+			     __field(u16, cursor)
+			     __field(u16, sr_plane)
+			     __field(u16, sr_cursor)
+			     __field(u16, sr_fbc)
+			     __field(u16, hpll_plane)
+			     __field(u16, hpll_cursor)
+			     __field(u16, hpll_fbc)
+			     __field(bool, cxsr)
+			     __field(bool, hpll)
+			     __field(bool, fbc)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   __entry->primary = wm->pipe[crtc->pipe].plane[PLANE_PRIMARY];
+			   __entry->sprite = wm->pipe[crtc->pipe].plane[PLANE_SPRITE0];
+			   __entry->cursor = wm->pipe[crtc->pipe].plane[PLANE_CURSOR];
+			   __entry->sr_plane = wm->sr.plane;
+			   __entry->sr_cursor = wm->sr.cursor;
+			   __entry->sr_fbc = wm->sr.fbc;
+			   __entry->hpll_plane = wm->hpll.plane;
+			   __entry->hpll_cursor = wm->hpll.cursor;
+			   __entry->hpll_fbc = wm->hpll.fbc;
+			   __entry->cxsr = wm->cxsr;
+			   __entry->hpll = wm->hpll_en;
+			   __entry->fbc = wm->fbc_en;
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u, wm %d/%d/%d, sr %s/%d/%d/%d, hpll %s/%d/%d/%d, fbc %s",
+		      pipe_name(__entry->pipe), __entry->frame, __entry->scanline,
+		      __entry->primary, __entry->sprite, __entry->cursor,
+		      str_yes_no(__entry->cxsr), __entry->sr_plane, __entry->sr_cursor, __entry->sr_fbc,
+		      str_yes_no(__entry->hpll), __entry->hpll_plane, __entry->hpll_cursor, __entry->hpll_fbc,
+		      str_yes_no(__entry->fbc))
+);
+
+TRACE_EVENT(vlv_wm,
+	    TP_PROTO(struct intel_crtc *crtc, const struct vlv_wm_values *wm),
+	    TP_ARGS(crtc, wm),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __field(u32, level)
+			     __field(u32, cxsr)
+			     __field(u32, primary)
+			     __field(u32, sprite0)
+			     __field(u32, sprite1)
+			     __field(u32, cursor)
+			     __field(u32, sr_plane)
+			     __field(u32, sr_cursor)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   __entry->level = wm->level;
+			   __entry->cxsr = wm->cxsr;
+			   __entry->primary = wm->pipe[crtc->pipe].plane[PLANE_PRIMARY];
+			   __entry->sprite0 = wm->pipe[crtc->pipe].plane[PLANE_SPRITE0];
+			   __entry->sprite1 = wm->pipe[crtc->pipe].plane[PLANE_SPRITE1];
+			   __entry->cursor = wm->pipe[crtc->pipe].plane[PLANE_CURSOR];
+			   __entry->sr_plane = wm->sr.plane;
+			   __entry->sr_cursor = wm->sr.cursor;
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u, level=%d, cxsr=%d, wm %d/%d/%d/%d, sr %d/%d",
+		      pipe_name(__entry->pipe), __entry->frame,
+		      __entry->scanline, __entry->level, __entry->cxsr,
+		      __entry->primary, __entry->sprite0, __entry->sprite1, __entry->cursor,
+		      __entry->sr_plane, __entry->sr_cursor)
+);
+
+TRACE_EVENT(vlv_fifo_size,
+	    TP_PROTO(struct intel_crtc *crtc, u32 sprite0_start, u32 sprite1_start, u32 fifo_size),
+	    TP_ARGS(crtc, sprite0_start, sprite1_start, fifo_size),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __field(u32, sprite0_start)
+			     __field(u32, sprite1_start)
+			     __field(u32, fifo_size)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   __entry->sprite0_start = sprite0_start;
+			   __entry->sprite1_start = sprite1_start;
+			   __entry->fifo_size = fifo_size;
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u, %d/%d/%d",
+		      pipe_name(__entry->pipe), __entry->frame,
+		      __entry->scanline, __entry->sprite0_start,
+		      __entry->sprite1_start, __entry->fifo_size)
+);
+
+TRACE_EVENT(intel_plane_update_noarm,
+	    TP_PROTO(struct drm_plane *plane, struct intel_crtc *crtc),
+	    TP_ARGS(plane, crtc),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __array(int, src, 4)
+			     __array(int, dst, 4)
+			     __string(name, plane->name)
+			     ),
+
+	    TP_fast_assign(
+			   __assign_str(name, plane->name);
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   memcpy(__entry->src, &plane->state->src, sizeof(__entry->src));
+			   memcpy(__entry->dst, &plane->state->dst, sizeof(__entry->dst));
+			   ),
+
+	    TP_printk("pipe %c, plane %s, frame=%u, scanline=%u, " DRM_RECT_FP_FMT " -> " DRM_RECT_FMT,
+		      pipe_name(__entry->pipe), __get_str(name),
+		      __entry->frame, __entry->scanline,
+		      DRM_RECT_FP_ARG((const struct drm_rect *)__entry->src),
+		      DRM_RECT_ARG((const struct drm_rect *)__entry->dst))
+);
+
+TRACE_EVENT(intel_plane_update_arm,
+	    TP_PROTO(struct drm_plane *plane, struct intel_crtc *crtc),
+	    TP_ARGS(plane, crtc),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __array(int, src, 4)
+			     __array(int, dst, 4)
+			     __string(name, plane->name)
+			     ),
+
+	    TP_fast_assign(
+			   __assign_str(name, plane->name);
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   memcpy(__entry->src, &plane->state->src, sizeof(__entry->src));
+			   memcpy(__entry->dst, &plane->state->dst, sizeof(__entry->dst));
+			   ),
+
+	    TP_printk("pipe %c, plane %s, frame=%u, scanline=%u, " DRM_RECT_FP_FMT " -> " DRM_RECT_FMT,
+		      pipe_name(__entry->pipe), __get_str(name),
+		      __entry->frame, __entry->scanline,
+		      DRM_RECT_FP_ARG((const struct drm_rect *)__entry->src),
+		      DRM_RECT_ARG((const struct drm_rect *)__entry->dst))
+);
+
+TRACE_EVENT(intel_plane_disable_arm,
+	    TP_PROTO(struct drm_plane *plane, struct intel_crtc *crtc),
+	    TP_ARGS(plane, crtc),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __string(name, plane->name)
+			     ),
+
+	    TP_fast_assign(
+			   __assign_str(name, plane->name);
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   ),
+
+	    TP_printk("pipe %c, plane %s, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe), __get_str(name),
+		      __entry->frame, __entry->scanline)
+);
+
+TRACE_EVENT(intel_fbc_activate,
+	    TP_PROTO(struct intel_plane *plane),
+	    TP_ARGS(plane),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			   struct intel_crtc *crtc = intel_crtc_for_pipe(to_i915(plane->base.dev),
+									 plane->pipe);
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe), __entry->frame, __entry->scanline)
+);
+
+TRACE_EVENT(intel_fbc_deactivate,
+	    TP_PROTO(struct intel_plane *plane),
+	    TP_ARGS(plane),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			   struct intel_crtc *crtc = intel_crtc_for_pipe(to_i915(plane->base.dev),
+									 plane->pipe);
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe), __entry->frame, __entry->scanline)
+);
+
+TRACE_EVENT(intel_fbc_nuke,
+	    TP_PROTO(struct intel_plane *plane),
+	    TP_ARGS(plane),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			   struct intel_crtc *crtc = intel_crtc_for_pipe(to_i915(plane->base.dev),
+									 plane->pipe);
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe), __entry->frame, __entry->scanline)
+);
+
+TRACE_EVENT(intel_crtc_vblank_work_start,
+	    TP_PROTO(struct intel_crtc *crtc),
+	    TP_ARGS(crtc),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe), __entry->frame,
+		       __entry->scanline)
+);
+
+TRACE_EVENT(intel_crtc_vblank_work_end,
+	    TP_PROTO(struct intel_crtc *crtc),
+	    TP_ARGS(crtc),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe), __entry->frame,
+		       __entry->scanline)
+);
+
+TRACE_EVENT(intel_pipe_update_start,
+	    TP_PROTO(struct intel_crtc *crtc),
+	    TP_ARGS(crtc),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __field(u32, min)
+			     __field(u32, max)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = intel_crtc_get_vblank_counter(crtc);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   __entry->min = crtc->debug.min_vbl;
+			   __entry->max = crtc->debug.max_vbl;
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u, min=%u, max=%u",
+		      pipe_name(__entry->pipe), __entry->frame,
+		       __entry->scanline, __entry->min, __entry->max)
+);
+
+TRACE_EVENT(intel_pipe_update_vblank_evaded,
+	    TP_PROTO(struct intel_crtc *crtc),
+	    TP_ARGS(crtc),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __field(u32, min)
+			     __field(u32, max)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = crtc->debug.start_vbl_count;
+			   __entry->scanline = crtc->debug.scanline_start;
+			   __entry->min = crtc->debug.min_vbl;
+			   __entry->max = crtc->debug.max_vbl;
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u, min=%u, max=%u",
+		      pipe_name(__entry->pipe), __entry->frame,
+		       __entry->scanline, __entry->min, __entry->max)
+);
+
+TRACE_EVENT(intel_pipe_update_end,
+	    TP_PROTO(struct intel_crtc *crtc, u32 frame, int scanline_end),
+	    TP_ARGS(crtc, frame, scanline_end),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = frame;
+			   __entry->scanline = scanline_end;
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe), __entry->frame,
+		      __entry->scanline)
+);
+
+TRACE_EVENT(intel_frontbuffer_invalidate,
+	    TP_PROTO(unsigned int frontbuffer_bits, unsigned int origin),
+	    TP_ARGS(frontbuffer_bits, origin),
+
+	    TP_STRUCT__entry(
+			     __field(unsigned int, frontbuffer_bits)
+			     __field(unsigned int, origin)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->frontbuffer_bits = frontbuffer_bits;
+			   __entry->origin = origin;
+			   ),
+
+	    TP_printk("frontbuffer_bits=0x%08x, origin=%u",
+		      __entry->frontbuffer_bits, __entry->origin)
+);
+
+TRACE_EVENT(intel_frontbuffer_flush,
+	    TP_PROTO(unsigned int frontbuffer_bits, unsigned int origin),
+	    TP_ARGS(frontbuffer_bits, origin),
+
+	    TP_STRUCT__entry(
+			     __field(unsigned int, frontbuffer_bits)
+			     __field(unsigned int, origin)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->frontbuffer_bits = frontbuffer_bits;
+			   __entry->origin = origin;
+			   ),
+
+	    TP_printk("frontbuffer_bits=0x%08x, origin=%u",
+		      __entry->frontbuffer_bits, __entry->origin)
+);
+
+#endif /* __INTEL_DISPLAY_TRACE_H__ */
+
+/* This part must be outside protection */
+#undef TRACE_INCLUDE_PATH
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_PATH ../../drivers/gpu/drm/i915/display
+#define TRACE_INCLUDE_FILE intel_display_trace
+#include <trace/define_trace.h>
diff --git a/drivers/gpu/drm/i915/display/intel_display_types.h b/drivers/gpu/drm/i915/display/intel_display_types.h
new file mode 100644
index 000000000..a8bf91a21
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_display_types.h
@@ -0,0 +1,2074 @@
+/*
+ * Copyright (c) 2006 Dave Airlie <airlied@linux.ie>
+ * Copyright (c) 2007-2008 Intel Corporation
+ *   Jesse Barnes <jesse.barnes@intel.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __INTEL_DISPLAY_TYPES_H__
+#define __INTEL_DISPLAY_TYPES_H__
+
+#include <linux/i2c.h>
+#include <linux/pm_qos.h>
+#include <linux/pwm.h>
+#include <linux/sched/clock.h>
+
+#include <drm/display/drm_dp_dual_mode_helper.h>
+#include <drm/display/drm_dp_mst_helper.h>
+#include <drm/display/drm_dsc.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_encoder.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_framebuffer.h>
+#include <drm/drm_probe_helper.h>
+#include <drm/drm_rect.h>
+#include <drm/drm_vblank.h>
+#include <drm/drm_vblank_work.h>
+#include <drm/i915_mei_hdcp_interface.h>
+#include <media/cec-notifier.h>
+
+#include "i915_vma.h"
+#include "i915_vma_types.h"
+#include "intel_bios.h"
+#include "intel_display.h"
+#include "intel_display_power.h"
+#include "intel_dpll_mgr.h"
+#include "intel_pm_types.h"
+
+struct drm_printer;
+struct __intel_global_objs_state;
+struct intel_ddi_buf_trans;
+struct intel_fbc;
+struct intel_connector;
+
+/*
+ * Display related stuff
+ */
+
+/* these are outputs from the chip - integrated only
+   external chips are via DVO or SDVO output */
+enum intel_output_type {
+	INTEL_OUTPUT_UNUSED = 0,
+	INTEL_OUTPUT_ANALOG = 1,
+	INTEL_OUTPUT_DVO = 2,
+	INTEL_OUTPUT_SDVO = 3,
+	INTEL_OUTPUT_LVDS = 4,
+	INTEL_OUTPUT_TVOUT = 5,
+	INTEL_OUTPUT_HDMI = 6,
+	INTEL_OUTPUT_DP = 7,
+	INTEL_OUTPUT_EDP = 8,
+	INTEL_OUTPUT_DSI = 9,
+	INTEL_OUTPUT_DDI = 10,
+	INTEL_OUTPUT_DP_MST = 11,
+};
+
+enum hdmi_force_audio {
+	HDMI_AUDIO_OFF_DVI = -2,	/* no aux data for HDMI-DVI converter */
+	HDMI_AUDIO_OFF,			/* force turn off HDMI audio */
+	HDMI_AUDIO_AUTO,		/* trust EDID */
+	HDMI_AUDIO_ON,			/* force turn on HDMI audio */
+};
+
+/* "Broadcast RGB" property */
+enum intel_broadcast_rgb {
+	INTEL_BROADCAST_RGB_AUTO,
+	INTEL_BROADCAST_RGB_FULL,
+	INTEL_BROADCAST_RGB_LIMITED,
+};
+
+struct intel_fb_view {
+	/*
+	 * The remap information used in the remapped and rotated views to
+	 * create the DMA scatter-gather list for each FB color plane. This sg
+	 * list is created along with the view type (gtt.type) specific
+	 * i915_vma object and contains the list of FB object pages (reordered
+	 * in the rotated view) that are visible in the view.
+	 * In the normal view the FB object's backing store sg list is used
+	 * directly and hence the remap information here is not used.
+	 */
+	struct i915_gtt_view gtt;
+
+	/*
+	 * The GTT view (gtt.type) specific information for each FB color
+	 * plane. In the normal GTT view all formats (up to 4 color planes),
+	 * in the rotated and remapped GTT view all no-CCS formats (up to 2
+	 * color planes) are supported.
+	 *
+	 * The view information shared by all FB color planes in the FB,
+	 * like dst x/y and src/dst width, is stored separately in
+	 * intel_plane_state.
+	 */
+	struct i915_color_plane_view {
+		u32 offset;
+		unsigned int x, y;
+		/*
+		 * Plane stride in:
+		 *   bytes for 0/180 degree rotation
+		 *   pixels for 90/270 degree rotation
+		 */
+		unsigned int mapping_stride;
+		unsigned int scanout_stride;
+	} color_plane[4];
+};
+
+struct intel_framebuffer {
+	struct drm_framebuffer base;
+	struct intel_frontbuffer *frontbuffer;
+
+	/* Params to remap the FB pages and program the plane registers in each view. */
+	struct intel_fb_view normal_view;
+	union {
+		struct intel_fb_view rotated_view;
+		struct intel_fb_view remapped_view;
+	};
+
+	struct i915_address_space *dpt_vm;
+};
+
+enum intel_hotplug_state {
+	INTEL_HOTPLUG_UNCHANGED,
+	INTEL_HOTPLUG_CHANGED,
+	INTEL_HOTPLUG_RETRY,
+};
+
+struct intel_encoder {
+	struct drm_encoder base;
+
+	enum intel_output_type type;
+	enum port port;
+	u16 cloneable;
+	u8 pipe_mask;
+	enum intel_hotplug_state (*hotplug)(struct intel_encoder *encoder,
+					    struct intel_connector *connector);
+	enum intel_output_type (*compute_output_type)(struct intel_encoder *,
+						      struct intel_crtc_state *,
+						      struct drm_connector_state *);
+	int (*compute_config)(struct intel_encoder *,
+			      struct intel_crtc_state *,
+			      struct drm_connector_state *);
+	int (*compute_config_late)(struct intel_encoder *,
+				   struct intel_crtc_state *,
+				   struct drm_connector_state *);
+	void (*update_prepare)(struct intel_atomic_state *,
+			       struct intel_encoder *,
+			       struct intel_crtc *);
+	void (*pre_pll_enable)(struct intel_atomic_state *,
+			       struct intel_encoder *,
+			       const struct intel_crtc_state *,
+			       const struct drm_connector_state *);
+	void (*pre_enable)(struct intel_atomic_state *,
+			   struct intel_encoder *,
+			   const struct intel_crtc_state *,
+			   const struct drm_connector_state *);
+	void (*enable)(struct intel_atomic_state *,
+		       struct intel_encoder *,
+		       const struct intel_crtc_state *,
+		       const struct drm_connector_state *);
+	void (*update_complete)(struct intel_atomic_state *,
+				struct intel_encoder *,
+				struct intel_crtc *);
+	void (*disable)(struct intel_atomic_state *,
+			struct intel_encoder *,
+			const struct intel_crtc_state *,
+			const struct drm_connector_state *);
+	void (*post_disable)(struct intel_atomic_state *,
+			     struct intel_encoder *,
+			     const struct intel_crtc_state *,
+			     const struct drm_connector_state *);
+	void (*post_pll_disable)(struct intel_atomic_state *,
+				 struct intel_encoder *,
+				 const struct intel_crtc_state *,
+				 const struct drm_connector_state *);
+	void (*update_pipe)(struct intel_atomic_state *,
+			    struct intel_encoder *,
+			    const struct intel_crtc_state *,
+			    const struct drm_connector_state *);
+	/* Read out the current hw state of this connector, returning true if
+	 * the encoder is active. If the encoder is enabled it also set the pipe
+	 * it is connected to in the pipe parameter. */
+	bool (*get_hw_state)(struct intel_encoder *, enum pipe *pipe);
+	/* Reconstructs the equivalent mode flags for the current hardware
+	 * state. This must be called _after_ display->get_pipe_config has
+	 * pre-filled the pipe config. Note that intel_encoder->base.crtc must
+	 * be set correctly before calling this function. */
+	void (*get_config)(struct intel_encoder *,
+			   struct intel_crtc_state *pipe_config);
+
+	/*
+	 * Optional hook called during init/resume to sync any state
+	 * stored in the encoder (eg. DP link parameters) wrt. the HW state.
+	 */
+	void (*sync_state)(struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state);
+
+	/*
+	 * Optional hook, returning true if this encoder allows a fastset
+	 * during the initial commit, false otherwise.
+	 */
+	bool (*initial_fastset_check)(struct intel_encoder *encoder,
+				      struct intel_crtc_state *crtc_state);
+
+	/*
+	 * Acquires the power domains needed for an active encoder during
+	 * hardware state readout.
+	 */
+	void (*get_power_domains)(struct intel_encoder *encoder,
+				  struct intel_crtc_state *crtc_state);
+	/*
+	 * Called during system suspend after all pending requests for the
+	 * encoder are flushed (for example for DP AUX transactions) and
+	 * device interrupts are disabled.
+	 */
+	void (*suspend)(struct intel_encoder *);
+	/*
+	 * Called during system reboot/shutdown after all the
+	 * encoders have been disabled and suspended.
+	 */
+	void (*shutdown)(struct intel_encoder *encoder);
+	/*
+	 * Enable/disable the clock to the port.
+	 */
+	void (*enable_clock)(struct intel_encoder *encoder,
+			     const struct intel_crtc_state *crtc_state);
+	void (*disable_clock)(struct intel_encoder *encoder);
+	/*
+	 * Returns whether the port clock is enabled or not.
+	 */
+	bool (*is_clock_enabled)(struct intel_encoder *encoder);
+	const struct intel_ddi_buf_trans *(*get_buf_trans)(struct intel_encoder *encoder,
+							   const struct intel_crtc_state *crtc_state,
+							   int *n_entries);
+	void (*set_signal_levels)(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state);
+
+	enum hpd_pin hpd_pin;
+	enum intel_display_power_domain power_domain;
+	/* for communication with audio component; protected by av_mutex */
+	const struct drm_connector *audio_connector;
+
+	/* VBT information for this encoder (may be NULL for older platforms) */
+	const struct intel_bios_encoder_data *devdata;
+};
+
+struct intel_panel_bl_funcs {
+	/* Connector and platform specific backlight functions */
+	int (*setup)(struct intel_connector *connector, enum pipe pipe);
+	u32 (*get)(struct intel_connector *connector, enum pipe pipe);
+	void (*set)(const struct drm_connector_state *conn_state, u32 level);
+	void (*disable)(const struct drm_connector_state *conn_state, u32 level);
+	void (*enable)(const struct intel_crtc_state *crtc_state,
+		       const struct drm_connector_state *conn_state, u32 level);
+	u32 (*hz_to_pwm)(struct intel_connector *connector, u32 hz);
+};
+
+enum drrs_type {
+	DRRS_TYPE_NONE,
+	DRRS_TYPE_STATIC,
+	DRRS_TYPE_SEAMLESS,
+};
+
+struct intel_vbt_panel_data {
+	struct drm_display_mode *lfp_lvds_vbt_mode; /* if any */
+	struct drm_display_mode *sdvo_lvds_vbt_mode; /* if any */
+
+	/* Feature bits */
+	int panel_type;
+	unsigned int lvds_dither:1;
+	unsigned int bios_lvds_val; /* initial [PCH_]LVDS reg val in VBIOS */
+
+	bool vrr;
+
+	u8 seamless_drrs_min_refresh_rate;
+	enum drrs_type drrs_type;
+
+	struct {
+		int max_link_rate;
+		int rate;
+		int lanes;
+		int preemphasis;
+		int vswing;
+		int bpp;
+		struct edp_power_seq pps;
+		u8 drrs_msa_timing_delay;
+		bool low_vswing;
+		bool initialized;
+		bool hobl;
+	} edp;
+
+	struct {
+		bool enable;
+		bool full_link;
+		bool require_aux_wakeup;
+		int idle_frames;
+		int tp1_wakeup_time_us;
+		int tp2_tp3_wakeup_time_us;
+		int psr2_tp2_tp3_wakeup_time_us;
+	} psr;
+
+	struct {
+		u16 pwm_freq_hz;
+		u16 brightness_precision_bits;
+		bool present;
+		bool active_low_pwm;
+		u8 min_brightness;	/* min_brightness/255 of max */
+		u8 controller;		/* brightness controller number */
+		enum intel_backlight_type type;
+	} backlight;
+
+	/* MIPI DSI */
+	struct {
+		u16 panel_id;
+		struct mipi_config *config;
+		struct mipi_pps_data *pps;
+		u16 bl_ports;
+		u16 cabc_ports;
+		u8 seq_version;
+		u32 size;
+		u8 *data;
+		const u8 *sequence[MIPI_SEQ_MAX];
+		u8 *deassert_seq; /* Used by fixup_mipi_sequences() */
+		enum drm_panel_orientation orientation;
+	} dsi;
+};
+
+struct intel_panel {
+	struct list_head fixed_modes;
+
+	/* backlight */
+	struct {
+		bool present;
+		u32 level;
+		u32 min;
+		u32 max;
+		bool enabled;
+		bool combination_mode;	/* gen 2/4 only */
+		bool active_low_pwm;
+		bool alternate_pwm_increment;	/* lpt+ */
+
+		/* PWM chip */
+		u32 pwm_level_min;
+		u32 pwm_level_max;
+		bool pwm_enabled;
+		bool util_pin_active_low;	/* bxt+ */
+		u8 controller;		/* bxt+ only */
+		struct pwm_device *pwm;
+		struct pwm_state pwm_state;
+
+		/* DPCD backlight */
+		union {
+			struct {
+				struct drm_edp_backlight_info info;
+			} vesa;
+			struct {
+				bool sdr_uses_aux;
+			} intel;
+		} edp;
+
+		struct backlight_device *device;
+
+		const struct intel_panel_bl_funcs *funcs;
+		const struct intel_panel_bl_funcs *pwm_funcs;
+		void (*power)(struct intel_connector *, bool enable);
+	} backlight;
+
+	struct intel_vbt_panel_data vbt;
+};
+
+struct intel_digital_port;
+
+enum check_link_response {
+	HDCP_LINK_PROTECTED	= 0,
+	HDCP_TOPOLOGY_CHANGE,
+	HDCP_LINK_INTEGRITY_FAILURE,
+	HDCP_REAUTH_REQUEST
+};
+
+/*
+ * This structure serves as a translation layer between the generic HDCP code
+ * and the bus-specific code. What that means is that HDCP over HDMI differs
+ * from HDCP over DP, so to account for these differences, we need to
+ * communicate with the receiver through this shim.
+ *
+ * For completeness, the 2 buses differ in the following ways:
+ *	- DP AUX vs. DDC
+ *		HDCP registers on the receiver are set via DP AUX for DP, and
+ *		they are set via DDC for HDMI.
+ *	- Receiver register offsets
+ *		The offsets of the registers are different for DP vs. HDMI
+ *	- Receiver register masks/offsets
+ *		For instance, the ready bit for the KSV fifo is in a different
+ *		place on DP vs HDMI
+ *	- Receiver register names
+ *		Seriously. In the DP spec, the 16-bit register containing
+ *		downstream information is called BINFO, on HDMI it's called
+ *		BSTATUS. To confuse matters further, DP has a BSTATUS register
+ *		with a completely different definition.
+ *	- KSV FIFO
+ *		On HDMI, the ksv fifo is read all at once, whereas on DP it must
+ *		be read 3 keys at a time
+ *	- Aksv output
+ *		Since Aksv is hidden in hardware, there's different procedures
+ *		to send it over DP AUX vs DDC
+ */
+struct intel_hdcp_shim {
+	/* Outputs the transmitter's An and Aksv values to the receiver. */
+	int (*write_an_aksv)(struct intel_digital_port *dig_port, u8 *an);
+
+	/* Reads the receiver's key selection vector */
+	int (*read_bksv)(struct intel_digital_port *dig_port, u8 *bksv);
+
+	/*
+	 * Reads BINFO from DP receivers and BSTATUS from HDMI receivers. The
+	 * definitions are the same in the respective specs, but the names are
+	 * different. Call it BSTATUS since that's the name the HDMI spec
+	 * uses and it was there first.
+	 */
+	int (*read_bstatus)(struct intel_digital_port *dig_port,
+			    u8 *bstatus);
+
+	/* Determines whether a repeater is present downstream */
+	int (*repeater_present)(struct intel_digital_port *dig_port,
+				bool *repeater_present);
+
+	/* Reads the receiver's Ri' value */
+	int (*read_ri_prime)(struct intel_digital_port *dig_port, u8 *ri);
+
+	/* Determines if the receiver's KSV FIFO is ready for consumption */
+	int (*read_ksv_ready)(struct intel_digital_port *dig_port,
+			      bool *ksv_ready);
+
+	/* Reads the ksv fifo for num_downstream devices */
+	int (*read_ksv_fifo)(struct intel_digital_port *dig_port,
+			     int num_downstream, u8 *ksv_fifo);
+
+	/* Reads a 32-bit part of V' from the receiver */
+	int (*read_v_prime_part)(struct intel_digital_port *dig_port,
+				 int i, u32 *part);
+
+	/* Enables HDCP signalling on the port */
+	int (*toggle_signalling)(struct intel_digital_port *dig_port,
+				 enum transcoder cpu_transcoder,
+				 bool enable);
+
+	/* Enable/Disable stream encryption on DP MST Transport Link */
+	int (*stream_encryption)(struct intel_connector *connector,
+				 bool enable);
+
+	/* Ensures the link is still protected */
+	bool (*check_link)(struct intel_digital_port *dig_port,
+			   struct intel_connector *connector);
+
+	/* Detects panel's hdcp capability. This is optional for HDMI. */
+	int (*hdcp_capable)(struct intel_digital_port *dig_port,
+			    bool *hdcp_capable);
+
+	/* HDCP adaptation(DP/HDMI) required on the port */
+	enum hdcp_wired_protocol protocol;
+
+	/* Detects whether sink is HDCP2.2 capable */
+	int (*hdcp_2_2_capable)(struct intel_digital_port *dig_port,
+				bool *capable);
+
+	/* Write HDCP2.2 messages */
+	int (*write_2_2_msg)(struct intel_digital_port *dig_port,
+			     void *buf, size_t size);
+
+	/* Read HDCP2.2 messages */
+	int (*read_2_2_msg)(struct intel_digital_port *dig_port,
+			    u8 msg_id, void *buf, size_t size);
+
+	/*
+	 * Implementation of DP HDCP2.2 Errata for the communication of stream
+	 * type to Receivers. In DP HDCP2.2 Stream type is one of the input to
+	 * the HDCP2.2 Cipher for En/De-Cryption. Not applicable for HDMI.
+	 */
+	int (*config_stream_type)(struct intel_digital_port *dig_port,
+				  bool is_repeater, u8 type);
+
+	/* Enable/Disable HDCP 2.2 stream encryption on DP MST Transport Link */
+	int (*stream_2_2_encryption)(struct intel_connector *connector,
+				     bool enable);
+
+	/* HDCP2.2 Link Integrity Check */
+	int (*check_2_2_link)(struct intel_digital_port *dig_port,
+			      struct intel_connector *connector);
+};
+
+struct intel_hdcp {
+	const struct intel_hdcp_shim *shim;
+	/* Mutex for hdcp state of the connector */
+	struct mutex mutex;
+	u64 value;
+	struct delayed_work check_work;
+	struct work_struct prop_work;
+
+	/* HDCP1.4 Encryption status */
+	bool hdcp_encrypted;
+
+	/* HDCP2.2 related definitions */
+	/* Flag indicates whether this connector supports HDCP2.2 or not. */
+	bool hdcp2_supported;
+
+	/* HDCP2.2 Encryption status */
+	bool hdcp2_encrypted;
+
+	/*
+	 * Content Stream Type defined by content owner. TYPE0(0x0) content can
+	 * flow in the link protected by HDCP2.2 or HDCP1.4, where as TYPE1(0x1)
+	 * content can flow only through a link protected by HDCP2.2.
+	 */
+	u8 content_type;
+
+	bool is_paired;
+	bool is_repeater;
+
+	/*
+	 * Count of ReceiverID_List received. Initialized to 0 at AKE_INIT.
+	 * Incremented after processing the RepeaterAuth_Send_ReceiverID_List.
+	 * When it rolls over re-auth has to be triggered.
+	 */
+	u32 seq_num_v;
+
+	/*
+	 * Count of RepeaterAuth_Stream_Manage msg propagated.
+	 * Initialized to 0 on AKE_INIT. Incremented after every successful
+	 * transmission of RepeaterAuth_Stream_Manage message. When it rolls
+	 * over re-Auth has to be triggered.
+	 */
+	u32 seq_num_m;
+
+	/*
+	 * Work queue to signal the CP_IRQ. Used for the waiters to read the
+	 * available information from HDCP DP sink.
+	 */
+	wait_queue_head_t cp_irq_queue;
+	atomic_t cp_irq_count;
+	int cp_irq_count_cached;
+
+	/*
+	 * HDCP register access for gen12+ need the transcoder associated.
+	 * Transcoder attached to the connector could be changed at modeset.
+	 * Hence caching the transcoder here.
+	 */
+	enum transcoder cpu_transcoder;
+	/* Only used for DP MST stream encryption */
+	enum transcoder stream_transcoder;
+};
+
+struct intel_connector {
+	struct drm_connector base;
+	/*
+	 * The fixed encoder this connector is connected to.
+	 */
+	struct intel_encoder *encoder;
+
+	/* ACPI device id for ACPI and driver cooperation */
+	u32 acpi_device_id;
+
+	/* Reads out the current hw, returning true if the connector is enabled
+	 * and active (i.e. dpms ON state). */
+	bool (*get_hw_state)(struct intel_connector *);
+
+	/* Panel info for eDP and LVDS */
+	struct intel_panel panel;
+
+	/* Cached EDID for eDP and LVDS. May hold ERR_PTR for invalid EDID. */
+	struct edid *edid;
+	struct edid *detect_edid;
+
+	/* Number of times hotplug detection was tried after an HPD interrupt */
+	int hotplug_retries;
+
+	/* since POLL and HPD connectors may use the same HPD line keep the native
+	   state of connector->polled in case hotplug storm detection changes it */
+	u8 polled;
+
+	struct drm_dp_mst_port *port;
+
+	struct intel_dp *mst_port;
+
+	/* Work struct to schedule a uevent on link train failure */
+	struct work_struct modeset_retry_work;
+
+	struct intel_hdcp hdcp;
+};
+
+struct intel_digital_connector_state {
+	struct drm_connector_state base;
+
+	enum hdmi_force_audio force_audio;
+	int broadcast_rgb;
+};
+
+#define to_intel_digital_connector_state(x) container_of(x, struct intel_digital_connector_state, base)
+
+struct dpll {
+	/* given values */
+	int n;
+	int m1, m2;
+	int p1, p2;
+	/* derived values */
+	int	dot;
+	int	vco;
+	int	m;
+	int	p;
+};
+
+struct intel_atomic_state {
+	struct drm_atomic_state base;
+
+	intel_wakeref_t wakeref;
+
+	struct __intel_global_objs_state *global_objs;
+	int num_global_objs;
+
+	bool dpll_set, modeset;
+
+	struct intel_shared_dpll_state shared_dpll[I915_NUM_PLLS];
+
+	/*
+	 * Current watermarks can't be trusted during hardware readout, so
+	 * don't bother calculating intermediate watermarks.
+	 */
+	bool skip_intermediate_wm;
+
+	bool rps_interactive;
+
+	struct i915_sw_fence commit_ready;
+
+	struct llist_node freed;
+};
+
+struct intel_plane_state {
+	struct drm_plane_state uapi;
+
+	/*
+	 * actual hardware state, the state we program to the hardware.
+	 * The following members are used to verify the hardware state:
+	 * During initial hw readout, they need to be copied from uapi.
+	 */
+	struct {
+		struct drm_crtc *crtc;
+		struct drm_framebuffer *fb;
+
+		u16 alpha;
+		u16 pixel_blend_mode;
+		unsigned int rotation;
+		enum drm_color_encoding color_encoding;
+		enum drm_color_range color_range;
+		enum drm_scaling_filter scaling_filter;
+	} hw;
+
+	struct i915_vma *ggtt_vma;
+	struct i915_vma *dpt_vma;
+	unsigned long flags;
+#define PLANE_HAS_FENCE BIT(0)
+
+	struct intel_fb_view view;
+
+	/* Plane pxp decryption state */
+	bool decrypt;
+
+	/* Plane state to display black pixels when pxp is borked */
+	bool force_black;
+
+	/* plane control register */
+	u32 ctl;
+
+	/* plane color control register */
+	u32 color_ctl;
+
+	/* chroma upsampler control register */
+	u32 cus_ctl;
+
+	/*
+	 * scaler_id
+	 *    = -1 : not using a scaler
+	 *    >=  0 : using a scalers
+	 *
+	 * plane requiring a scaler:
+	 *   - During check_plane, its bit is set in
+	 *     crtc_state->scaler_state.scaler_users by calling helper function
+	 *     update_scaler_plane.
+	 *   - scaler_id indicates the scaler it got assigned.
+	 *
+	 * plane doesn't require a scaler:
+	 *   - this can happen when scaling is no more required or plane simply
+	 *     got disabled.
+	 *   - During check_plane, corresponding bit is reset in
+	 *     crtc_state->scaler_state.scaler_users by calling helper function
+	 *     update_scaler_plane.
+	 */
+	int scaler_id;
+
+	/*
+	 * planar_linked_plane:
+	 *
+	 * ICL planar formats require 2 planes that are updated as pairs.
+	 * This member is used to make sure the other plane is also updated
+	 * when required, and for update_slave() to find the correct
+	 * plane_state to pass as argument.
+	 */
+	struct intel_plane *planar_linked_plane;
+
+	/*
+	 * planar_slave:
+	 * If set don't update use the linked plane's state for updating
+	 * this plane during atomic commit with the update_slave() callback.
+	 *
+	 * It's also used by the watermark code to ignore wm calculations on
+	 * this plane. They're calculated by the linked plane's wm code.
+	 */
+	u32 planar_slave;
+
+	struct drm_intel_sprite_colorkey ckey;
+
+	struct drm_rect psr2_sel_fetch_area;
+
+	/* Clear Color Value */
+	u64 ccval;
+
+	const char *no_fbc_reason;
+};
+
+struct intel_initial_plane_config {
+	struct intel_framebuffer *fb;
+	struct i915_vma *vma;
+	unsigned int tiling;
+	int size;
+	u32 base;
+	u8 rotation;
+};
+
+struct intel_scaler {
+	int in_use;
+	u32 mode;
+};
+
+struct intel_crtc_scaler_state {
+#define SKL_NUM_SCALERS 2
+	struct intel_scaler scalers[SKL_NUM_SCALERS];
+
+	/*
+	 * scaler_users: keeps track of users requesting scalers on this crtc.
+	 *
+	 *     If a bit is set, a user is using a scaler.
+	 *     Here user can be a plane or crtc as defined below:
+	 *       bits 0-30 - plane (bit position is index from drm_plane_index)
+	 *       bit 31    - crtc
+	 *
+	 * Instead of creating a new index to cover planes and crtc, using
+	 * existing drm_plane_index for planes which is well less than 31
+	 * planes and bit 31 for crtc. This should be fine to cover all
+	 * our platforms.
+	 *
+	 * intel_atomic_setup_scalers will setup available scalers to users
+	 * requesting scalers. It will gracefully fail if request exceeds
+	 * avilability.
+	 */
+#define SKL_CRTC_INDEX 31
+	unsigned scaler_users;
+
+	/* scaler used by crtc for panel fitting purpose */
+	int scaler_id;
+};
+
+/* {crtc,crtc_state}->mode_flags */
+/* Flag to get scanline using frame time stamps */
+#define I915_MODE_FLAG_GET_SCANLINE_FROM_TIMESTAMP (1<<1)
+/* Flag to use the scanline counter instead of the pixel counter */
+#define I915_MODE_FLAG_USE_SCANLINE_COUNTER (1<<2)
+/*
+ * TE0 or TE1 flag is set if the crtc has a DSI encoder which
+ * is operating in command mode.
+ * Flag to use TE from DSI0 instead of VBI in command mode
+ */
+#define I915_MODE_FLAG_DSI_USE_TE0 (1<<3)
+/* Flag to use TE from DSI1 instead of VBI in command mode */
+#define I915_MODE_FLAG_DSI_USE_TE1 (1<<4)
+/* Flag to indicate mipi dsi periodic command mode where we do not get TE */
+#define I915_MODE_FLAG_DSI_PERIODIC_CMD_MODE (1<<5)
+/* Do tricks to make vblank timestamps sane with VRR? */
+#define I915_MODE_FLAG_VRR (1<<6)
+
+struct intel_wm_level {
+	bool enable;
+	u32 pri_val;
+	u32 spr_val;
+	u32 cur_val;
+	u32 fbc_val;
+};
+
+struct intel_pipe_wm {
+	struct intel_wm_level wm[5];
+	bool fbc_wm_enabled;
+	bool pipe_enabled;
+	bool sprites_enabled;
+	bool sprites_scaled;
+};
+
+struct skl_wm_level {
+	u16 min_ddb_alloc;
+	u16 blocks;
+	u8 lines;
+	bool enable;
+	bool ignore_lines;
+	bool can_sagv;
+};
+
+struct skl_plane_wm {
+	struct skl_wm_level wm[8];
+	struct skl_wm_level uv_wm[8];
+	struct skl_wm_level trans_wm;
+	struct {
+		struct skl_wm_level wm0;
+		struct skl_wm_level trans_wm;
+	} sagv;
+	bool is_planar;
+};
+
+struct skl_pipe_wm {
+	struct skl_plane_wm planes[I915_MAX_PLANES];
+	bool use_sagv_wm;
+};
+
+enum vlv_wm_level {
+	VLV_WM_LEVEL_PM2,
+	VLV_WM_LEVEL_PM5,
+	VLV_WM_LEVEL_DDR_DVFS,
+	NUM_VLV_WM_LEVELS,
+};
+
+struct vlv_wm_state {
+	struct g4x_pipe_wm wm[NUM_VLV_WM_LEVELS];
+	struct g4x_sr_wm sr[NUM_VLV_WM_LEVELS];
+	u8 num_levels;
+	bool cxsr;
+};
+
+struct vlv_fifo_state {
+	u16 plane[I915_MAX_PLANES];
+};
+
+enum g4x_wm_level {
+	G4X_WM_LEVEL_NORMAL,
+	G4X_WM_LEVEL_SR,
+	G4X_WM_LEVEL_HPLL,
+	NUM_G4X_WM_LEVELS,
+};
+
+struct g4x_wm_state {
+	struct g4x_pipe_wm wm;
+	struct g4x_sr_wm sr;
+	struct g4x_sr_wm hpll;
+	bool cxsr;
+	bool hpll_en;
+	bool fbc_en;
+};
+
+struct intel_crtc_wm_state {
+	union {
+		/*
+		 * raw:
+		 * The "raw" watermark values produced by the formula
+		 * given the plane's current state. They do not consider
+		 * how much FIFO is actually allocated for each plane.
+		 *
+		 * optimal:
+		 * The "optimal" watermark values given the current
+		 * state of the planes and the amount of FIFO
+		 * allocated to each, ignoring any previous state
+		 * of the planes.
+		 *
+		 * intermediate:
+		 * The "intermediate" watermark values when transitioning
+		 * between the old and new "optimal" values. Used when
+		 * the watermark registers are single buffered and hence
+		 * their state changes asynchronously with regards to the
+		 * actual plane registers. These are essentially the
+		 * worst case combination of the old and new "optimal"
+		 * watermarks, which are therefore safe to use when the
+		 * plane is in either its old or new state.
+		 */
+		struct {
+			struct intel_pipe_wm intermediate;
+			struct intel_pipe_wm optimal;
+		} ilk;
+
+		struct {
+			struct skl_pipe_wm raw;
+			/* gen9+ only needs 1-step wm programming */
+			struct skl_pipe_wm optimal;
+			struct skl_ddb_entry ddb;
+			/*
+			 * pre-icl: for packed/planar CbCr
+			 * icl+: for everything
+			 */
+			struct skl_ddb_entry plane_ddb[I915_MAX_PLANES];
+			/* pre-icl: for planar Y */
+			struct skl_ddb_entry plane_ddb_y[I915_MAX_PLANES];
+		} skl;
+
+		struct {
+			struct g4x_pipe_wm raw[NUM_VLV_WM_LEVELS]; /* not inverted */
+			struct vlv_wm_state intermediate; /* inverted */
+			struct vlv_wm_state optimal; /* inverted */
+			struct vlv_fifo_state fifo_state;
+		} vlv;
+
+		struct {
+			struct g4x_pipe_wm raw[NUM_G4X_WM_LEVELS];
+			struct g4x_wm_state intermediate;
+			struct g4x_wm_state optimal;
+		} g4x;
+	};
+
+	/*
+	 * Platforms with two-step watermark programming will need to
+	 * update watermark programming post-vblank to switch from the
+	 * safe intermediate watermarks to the optimal final
+	 * watermarks.
+	 */
+	bool need_postvbl_update;
+};
+
+enum intel_output_format {
+	INTEL_OUTPUT_FORMAT_RGB,
+	INTEL_OUTPUT_FORMAT_YCBCR420,
+	INTEL_OUTPUT_FORMAT_YCBCR444,
+};
+
+struct intel_mpllb_state {
+	u32 clock; /* in KHz */
+	u32 ref_control;
+	u32 mpllb_cp;
+	u32 mpllb_div;
+	u32 mpllb_div2;
+	u32 mpllb_fracn1;
+	u32 mpllb_fracn2;
+	u32 mpllb_sscen;
+	u32 mpllb_sscstep;
+};
+
+struct intel_crtc_state {
+	/*
+	 * uapi (drm) state. This is the software state shown to userspace.
+	 * In particular, the following members are used for bookkeeping:
+	 * - crtc
+	 * - state
+	 * - *_changed
+	 * - event
+	 * - commit
+	 * - mode_blob
+	 */
+	struct drm_crtc_state uapi;
+
+	/*
+	 * actual hardware state, the state we program to the hardware.
+	 * The following members are used to verify the hardware state:
+	 * - enable
+	 * - active
+	 * - mode / pipe_mode / adjusted_mode
+	 * - color property blobs.
+	 *
+	 * During initial hw readout, they need to be copied to uapi.
+	 *
+	 * Bigjoiner will allow a transcoder mode that spans 2 pipes;
+	 * Use the pipe_mode for calculations like watermarks, pipe
+	 * scaler, and bandwidth.
+	 *
+	 * Use adjusted_mode for things that need to know the full
+	 * mode on the transcoder, which spans all pipes.
+	 */
+	struct {
+		bool active, enable;
+		struct drm_property_blob *degamma_lut, *gamma_lut, *ctm;
+		struct drm_display_mode mode, pipe_mode, adjusted_mode;
+		enum drm_scaling_filter scaling_filter;
+	} hw;
+
+	/**
+	 * quirks - bitfield with hw state readout quirks
+	 *
+	 * For various reasons the hw state readout code might not be able to
+	 * completely faithfully read out the current state. These cases are
+	 * tracked with quirk flags so that fastboot and state checker can act
+	 * accordingly.
+	 */
+#define PIPE_CONFIG_QUIRK_MODE_SYNC_FLAGS	(1<<0) /* unreliable sync mode.flags */
+	unsigned long quirks;
+
+	unsigned fb_bits; /* framebuffers to flip */
+	bool update_pipe; /* can a fast modeset be performed? */
+	bool disable_cxsr;
+	bool update_wm_pre, update_wm_post; /* watermarks are updated */
+	bool fifo_changed; /* FIFO split is changed */
+	bool preload_luts;
+	bool inherited; /* state inherited from BIOS? */
+
+	/* Ask the hardware to actually async flip? */
+	bool do_async_flip;
+
+	/* Pipe source size (ie. panel fitter input size)
+	 * All planes will be positioned inside this space,
+	 * and get clipped at the edges. */
+	struct drm_rect pipe_src;
+
+	/*
+	 * Pipe pixel rate, adjusted for
+	 * panel fitter/pipe scaler downscaling.
+	 */
+	unsigned int pixel_rate;
+
+	/* Whether to set up the PCH/FDI. Note that we never allow sharing
+	 * between pch encoders and cpu encoders. */
+	bool has_pch_encoder;
+
+	/* Are we sending infoframes on the attached port */
+	bool has_infoframe;
+
+	/* CPU Transcoder for the pipe. Currently this can only differ from the
+	 * pipe on Haswell and later (where we have a special eDP transcoder)
+	 * and Broxton (where we have special DSI transcoders). */
+	enum transcoder cpu_transcoder;
+
+	/*
+	 * Use reduced/limited/broadcast rbg range, compressing from the full
+	 * range fed into the crtcs.
+	 */
+	bool limited_color_range;
+
+	/* Bitmask of encoder types (enum intel_output_type)
+	 * driven by the pipe.
+	 */
+	unsigned int output_types;
+
+	/* Whether we should send NULL infoframes. Required for audio. */
+	bool has_hdmi_sink;
+
+	/* Audio enabled on this pipe. Only valid if either has_hdmi_sink or
+	 * has_dp_encoder is set. */
+	bool has_audio;
+
+	/*
+	 * Enable dithering, used when the selected pipe bpp doesn't match the
+	 * plane bpp.
+	 */
+	bool dither;
+
+	/*
+	 * Dither gets enabled for 18bpp which causes CRC mismatch errors for
+	 * compliance video pattern tests.
+	 * Disable dither only if it is a compliance test request for
+	 * 18bpp.
+	 */
+	bool dither_force_disable;
+
+	/* Controls for the clock computation, to override various stages. */
+	bool clock_set;
+
+	/* SDVO TV has a bunch of special case. To make multifunction encoders
+	 * work correctly, we need to track this at runtime.*/
+	bool sdvo_tv_clock;
+
+	/*
+	 * crtc bandwidth limit, don't increase pipe bpp or clock if not really
+	 * required. This is set in the 2nd loop of calling encoder's
+	 * ->compute_config if the first pick doesn't work out.
+	 */
+	bool bw_constrained;
+
+	/* Settings for the intel dpll used on pretty much everything but
+	 * haswell. */
+	struct dpll dpll;
+
+	/* Selected dpll when shared or NULL. */
+	struct intel_shared_dpll *shared_dpll;
+
+	/* Actual register state of the dpll, for shared dpll cross-checking. */
+	union {
+		struct intel_dpll_hw_state dpll_hw_state;
+		struct intel_mpllb_state mpllb_state;
+	};
+
+	/*
+	 * ICL reserved DPLLs for the CRTC/port. The active PLL is selected by
+	 * setting shared_dpll and dpll_hw_state to one of these reserved ones.
+	 */
+	struct icl_port_dpll {
+		struct intel_shared_dpll *pll;
+		struct intel_dpll_hw_state hw_state;
+	} icl_port_dplls[ICL_PORT_DPLL_COUNT];
+
+	/* DSI PLL registers */
+	struct {
+		u32 ctrl, div;
+	} dsi_pll;
+
+	int pipe_bpp;
+	struct intel_link_m_n dp_m_n;
+
+	/* m2_n2 for eDP downclock */
+	struct intel_link_m_n dp_m2_n2;
+	bool has_drrs;
+	bool seamless_m_n;
+
+	/* PSR is supported but might not be enabled due the lack of enabled planes */
+	bool has_psr;
+	bool has_psr2;
+	bool enable_psr2_sel_fetch;
+	bool req_psr2_sdp_prior_scanline;
+	u32 dc3co_exitline;
+	u16 su_y_granularity;
+	struct drm_dp_vsc_sdp psr_vsc;
+
+	/*
+	 * Frequence the dpll for the port should run at. Differs from the
+	 * adjusted dotclock e.g. for DP or 10/12bpc hdmi mode. This is also
+	 * already multiplied by pixel_multiplier.
+	 */
+	int port_clock;
+
+	/* Used by SDVO (and if we ever fix it, HDMI). */
+	unsigned pixel_multiplier;
+
+	/* I915_MODE_FLAG_* */
+	u8 mode_flags;
+
+	u8 lane_count;
+
+	/*
+	 * Used by platforms having DP/HDMI PHY with programmable lane
+	 * latency optimization.
+	 */
+	u8 lane_lat_optim_mask;
+
+	/* minimum acceptable voltage level */
+	u8 min_voltage_level;
+
+	/* Panel fitter controls for gen2-gen4 + VLV */
+	struct {
+		u32 control;
+		u32 pgm_ratios;
+		u32 lvds_border_bits;
+	} gmch_pfit;
+
+	/* Panel fitter placement and size for Ironlake+ */
+	struct {
+		struct drm_rect dst;
+		bool enabled;
+		bool force_thru;
+	} pch_pfit;
+
+	/* FDI configuration, only valid if has_pch_encoder is set. */
+	int fdi_lanes;
+	struct intel_link_m_n fdi_m_n;
+
+	bool ips_enabled;
+
+	bool crc_enabled;
+
+	bool double_wide;
+
+	int pbn;
+
+	struct intel_crtc_scaler_state scaler_state;
+
+	/* w/a for waiting 2 vblanks during crtc enable */
+	enum pipe hsw_workaround_pipe;
+
+	/* IVB sprite scaling w/a (WaCxSRDisabledForSpriteScaling:ivb) */
+	bool disable_lp_wm;
+
+	struct intel_crtc_wm_state wm;
+
+	int min_cdclk[I915_MAX_PLANES];
+
+	/* for packed/planar CbCr */
+	u32 data_rate[I915_MAX_PLANES];
+	/* for planar Y */
+	u32 data_rate_y[I915_MAX_PLANES];
+
+	/* FIXME unify with data_rate[]? */
+	u64 rel_data_rate[I915_MAX_PLANES];
+	u64 rel_data_rate_y[I915_MAX_PLANES];
+
+	/* Gamma mode programmed on the pipe */
+	u32 gamma_mode;
+
+	union {
+		/* CSC mode programmed on the pipe */
+		u32 csc_mode;
+
+		/* CHV CGM mode */
+		u32 cgm_mode;
+	};
+
+	/* bitmask of logically enabled planes (enum plane_id) */
+	u8 enabled_planes;
+
+	/* bitmask of actually visible planes (enum plane_id) */
+	u8 active_planes;
+	u8 scaled_planes;
+	u8 nv12_planes;
+	u8 c8_planes;
+
+	/* bitmask of planes that will be updated during the commit */
+	u8 update_planes;
+
+	u8 framestart_delay; /* 1-4 */
+	u8 msa_timing_delay; /* 0-3 */
+
+	struct {
+		u32 enable;
+		u32 gcp;
+		union hdmi_infoframe avi;
+		union hdmi_infoframe spd;
+		union hdmi_infoframe hdmi;
+		union hdmi_infoframe drm;
+		struct drm_dp_vsc_sdp vsc;
+	} infoframes;
+
+	/* HDMI scrambling status */
+	bool hdmi_scrambling;
+
+	/* HDMI High TMDS char rate ratio */
+	bool hdmi_high_tmds_clock_ratio;
+
+	/* Output format RGB/YCBCR etc */
+	enum intel_output_format output_format;
+
+	/* enable pipe gamma? */
+	bool gamma_enable;
+
+	/* enable pipe csc? */
+	bool csc_enable;
+
+	/* big joiner pipe bitmask */
+	u8 bigjoiner_pipes;
+
+	/* Display Stream compression state */
+	struct {
+		bool compression_enable;
+		bool dsc_split;
+		u16 compressed_bpp;
+		u8 slice_count;
+		struct drm_dsc_config config;
+	} dsc;
+
+	/* HSW+ linetime watermarks */
+	u16 linetime;
+	u16 ips_linetime;
+
+	/* Forward Error correction State */
+	bool fec_enable;
+
+	/* Pointer to master transcoder in case of tiled displays */
+	enum transcoder master_transcoder;
+
+	/* Bitmask to indicate slaves attached */
+	u8 sync_mode_slaves_mask;
+
+	/* Only valid on TGL+ */
+	enum transcoder mst_master_transcoder;
+
+	/* For DSB related info */
+	struct intel_dsb *dsb;
+
+	u32 psr2_man_track_ctl;
+
+	/* Variable Refresh Rate state */
+	struct {
+		bool enable;
+		u8 pipeline_full;
+		u16 flipline, vmin, vmax, guardband;
+	} vrr;
+
+	/* Stream Splitter for eDP MSO */
+	struct {
+		bool enable;
+		u8 link_count;
+		u8 pixel_overlap;
+	} splitter;
+
+	/* for loading single buffered registers during vblank */
+	struct drm_vblank_work vblank_work;
+};
+
+enum intel_pipe_crc_source {
+	INTEL_PIPE_CRC_SOURCE_NONE,
+	INTEL_PIPE_CRC_SOURCE_PLANE1,
+	INTEL_PIPE_CRC_SOURCE_PLANE2,
+	INTEL_PIPE_CRC_SOURCE_PLANE3,
+	INTEL_PIPE_CRC_SOURCE_PLANE4,
+	INTEL_PIPE_CRC_SOURCE_PLANE5,
+	INTEL_PIPE_CRC_SOURCE_PLANE6,
+	INTEL_PIPE_CRC_SOURCE_PLANE7,
+	INTEL_PIPE_CRC_SOURCE_PIPE,
+	/* TV/DP on pre-gen5/vlv can't use the pipe source. */
+	INTEL_PIPE_CRC_SOURCE_TV,
+	INTEL_PIPE_CRC_SOURCE_DP_B,
+	INTEL_PIPE_CRC_SOURCE_DP_C,
+	INTEL_PIPE_CRC_SOURCE_DP_D,
+	INTEL_PIPE_CRC_SOURCE_AUTO,
+	INTEL_PIPE_CRC_SOURCE_MAX,
+};
+
+enum drrs_refresh_rate {
+	DRRS_REFRESH_RATE_HIGH,
+	DRRS_REFRESH_RATE_LOW,
+};
+
+#define INTEL_PIPE_CRC_ENTRIES_NR	128
+struct intel_pipe_crc {
+	spinlock_t lock;
+	int skipped;
+	enum intel_pipe_crc_source source;
+};
+
+struct intel_crtc {
+	struct drm_crtc base;
+	enum pipe pipe;
+	/*
+	 * Whether the crtc and the connected output pipeline is active. Implies
+	 * that crtc->enabled is set, i.e. the current mode configuration has
+	 * some outputs connected to this crtc.
+	 */
+	bool active;
+	u8 plane_ids_mask;
+
+	/* I915_MODE_FLAG_* */
+	u8 mode_flags;
+
+	u16 vmax_vblank_start;
+
+	struct intel_display_power_domain_set enabled_power_domains;
+	struct intel_overlay *overlay;
+
+	struct intel_crtc_state *config;
+
+	/* Access to these should be protected by dev_priv->irq_lock. */
+	bool cpu_fifo_underrun_disabled;
+	bool pch_fifo_underrun_disabled;
+
+	/* per-pipe watermark state */
+	struct {
+		/* watermarks currently being used  */
+		union {
+			struct intel_pipe_wm ilk;
+			struct vlv_wm_state vlv;
+			struct g4x_wm_state g4x;
+		} active;
+	} wm;
+
+	struct {
+		struct mutex mutex;
+		struct delayed_work work;
+		enum drrs_refresh_rate refresh_rate;
+		unsigned int frontbuffer_bits;
+		unsigned int busy_frontbuffer_bits;
+		enum transcoder cpu_transcoder;
+		struct intel_link_m_n m_n, m2_n2;
+	} drrs;
+
+	int scanline_offset;
+
+	struct {
+		unsigned start_vbl_count;
+		ktime_t start_vbl_time;
+		int min_vbl, max_vbl;
+		int scanline_start;
+#ifdef CONFIG_DRM_I915_DEBUG_VBLANK_EVADE
+		struct {
+			u64 min;
+			u64 max;
+			u64 sum;
+			unsigned int over;
+			unsigned int times[17]; /* [1us, 16ms] */
+		} vbl;
+#endif
+	} debug;
+
+	/* scalers available on this crtc */
+	int num_scalers;
+
+	/* for loading single buffered registers during vblank */
+	struct pm_qos_request vblank_pm_qos;
+
+#ifdef CONFIG_DEBUG_FS
+	struct intel_pipe_crc pipe_crc;
+#endif
+};
+
+struct intel_plane {
+	struct drm_plane base;
+	enum i9xx_plane_id i9xx_plane;
+	enum plane_id id;
+	enum pipe pipe;
+	bool need_async_flip_disable_wa;
+	u32 frontbuffer_bit;
+
+	struct {
+		u32 base, cntl, size;
+	} cursor;
+
+	struct intel_fbc *fbc;
+
+	/*
+	 * NOTE: Do not place new plane state fields here (e.g., when adding
+	 * new plane properties).  New runtime state should now be placed in
+	 * the intel_plane_state structure and accessed via plane_state.
+	 */
+
+	int (*min_width)(const struct drm_framebuffer *fb,
+			 int color_plane,
+			 unsigned int rotation);
+	int (*max_width)(const struct drm_framebuffer *fb,
+			 int color_plane,
+			 unsigned int rotation);
+	int (*max_height)(const struct drm_framebuffer *fb,
+			  int color_plane,
+			  unsigned int rotation);
+	unsigned int (*max_stride)(struct intel_plane *plane,
+				   u32 pixel_format, u64 modifier,
+				   unsigned int rotation);
+	/* Write all non-self arming plane registers */
+	void (*update_noarm)(struct intel_plane *plane,
+			     const struct intel_crtc_state *crtc_state,
+			     const struct intel_plane_state *plane_state);
+	/* Write all self-arming plane registers */
+	void (*update_arm)(struct intel_plane *plane,
+			   const struct intel_crtc_state *crtc_state,
+			   const struct intel_plane_state *plane_state);
+	/* Disable the plane, must arm */
+	void (*disable_arm)(struct intel_plane *plane,
+			    const struct intel_crtc_state *crtc_state);
+	bool (*get_hw_state)(struct intel_plane *plane, enum pipe *pipe);
+	int (*check_plane)(struct intel_crtc_state *crtc_state,
+			   struct intel_plane_state *plane_state);
+	int (*min_cdclk)(const struct intel_crtc_state *crtc_state,
+			 const struct intel_plane_state *plane_state);
+	void (*async_flip)(struct intel_plane *plane,
+			   const struct intel_crtc_state *crtc_state,
+			   const struct intel_plane_state *plane_state,
+			   bool async_flip);
+	void (*enable_flip_done)(struct intel_plane *plane);
+	void (*disable_flip_done)(struct intel_plane *plane);
+};
+
+struct intel_watermark_params {
+	u16 fifo_size;
+	u16 max_wm;
+	u8 default_wm;
+	u8 guard_size;
+	u8 cacheline_size;
+};
+
+struct cxsr_latency {
+	bool is_desktop : 1;
+	bool is_ddr3 : 1;
+	u16 fsb_freq;
+	u16 mem_freq;
+	u16 display_sr;
+	u16 display_hpll_disable;
+	u16 cursor_sr;
+	u16 cursor_hpll_disable;
+};
+
+#define to_intel_atomic_state(x) container_of(x, struct intel_atomic_state, base)
+#define to_intel_crtc(x) container_of(x, struct intel_crtc, base)
+#define to_intel_crtc_state(x) container_of(x, struct intel_crtc_state, uapi)
+#define to_intel_connector(x) container_of(x, struct intel_connector, base)
+#define to_intel_encoder(x) container_of(x, struct intel_encoder, base)
+#define to_intel_framebuffer(x) container_of(x, struct intel_framebuffer, base)
+#define to_intel_plane(x) container_of(x, struct intel_plane, base)
+#define to_intel_plane_state(x) container_of(x, struct intel_plane_state, uapi)
+#define intel_fb_obj(x) ((x) ? to_intel_bo((x)->obj[0]) : NULL)
+
+struct intel_hdmi {
+	i915_reg_t hdmi_reg;
+	int ddc_bus;
+	struct {
+		enum drm_dp_dual_mode_type type;
+		int max_tmds_clock;
+	} dp_dual_mode;
+	bool has_hdmi_sink;
+	bool has_audio;
+	struct intel_connector *attached_connector;
+	struct cec_notifier *cec_notifier;
+};
+
+struct intel_dp_mst_encoder;
+
+struct intel_dp_compliance_data {
+	unsigned long edid;
+	u8 video_pattern;
+	u16 hdisplay, vdisplay;
+	u8 bpc;
+	struct drm_dp_phy_test_params phytest;
+};
+
+struct intel_dp_compliance {
+	unsigned long test_type;
+	struct intel_dp_compliance_data test_data;
+	bool test_active;
+	int test_link_rate;
+	u8 test_lane_count;
+};
+
+struct intel_dp_pcon_frl {
+	bool is_trained;
+	int trained_rate_gbps;
+};
+
+struct intel_pps {
+	int panel_power_up_delay;
+	int panel_power_down_delay;
+	int panel_power_cycle_delay;
+	int backlight_on_delay;
+	int backlight_off_delay;
+	struct delayed_work panel_vdd_work;
+	bool want_panel_vdd;
+	bool initializing;
+	unsigned long last_power_on;
+	unsigned long last_backlight_off;
+	ktime_t panel_power_off_time;
+	intel_wakeref_t vdd_wakeref;
+
+	/*
+	 * Pipe whose power sequencer is currently locked into
+	 * this port. Only relevant on VLV/CHV.
+	 */
+	enum pipe pps_pipe;
+	/*
+	 * Pipe currently driving the port. Used for preventing
+	 * the use of the PPS for any pipe currentrly driving
+	 * external DP as that will mess things up on VLV.
+	 */
+	enum pipe active_pipe;
+	/*
+	 * Set if the sequencer may be reset due to a power transition,
+	 * requiring a reinitialization. Only relevant on BXT.
+	 */
+	bool pps_reset;
+	struct edp_power_seq pps_delays;
+	struct edp_power_seq bios_pps_delays;
+};
+
+struct intel_psr {
+	/* Mutex for PSR state of the transcoder */
+	struct mutex lock;
+
+#define I915_PSR_DEBUG_MODE_MASK	0x0f
+#define I915_PSR_DEBUG_DEFAULT		0x00
+#define I915_PSR_DEBUG_DISABLE		0x01
+#define I915_PSR_DEBUG_ENABLE		0x02
+#define I915_PSR_DEBUG_FORCE_PSR1	0x03
+#define I915_PSR_DEBUG_ENABLE_SEL_FETCH	0x4
+#define I915_PSR_DEBUG_IRQ		0x10
+
+	u32 debug;
+	bool sink_support;
+	bool source_support;
+	bool enabled;
+	bool paused;
+	enum pipe pipe;
+	enum transcoder transcoder;
+	bool active;
+	struct work_struct work;
+	unsigned int busy_frontbuffer_bits;
+	bool sink_psr2_support;
+	bool link_standby;
+	bool colorimetry_support;
+	bool psr2_enabled;
+	bool psr2_sel_fetch_enabled;
+	bool psr2_sel_fetch_cff_enabled;
+	bool req_psr2_sdp_prior_scanline;
+	u8 sink_sync_latency;
+	u8 io_wake_lines;
+	u8 fast_wake_lines;
+	ktime_t last_entry_attempt;
+	ktime_t last_exit;
+	bool sink_not_reliable;
+	bool irq_aux_error;
+	u16 su_w_granularity;
+	u16 su_y_granularity;
+	u32 dc3co_exitline;
+	u32 dc3co_exit_delay;
+	struct delayed_work dc3co_work;
+};
+
+struct intel_dp {
+	i915_reg_t output_reg;
+	u32 DP;
+	int link_rate;
+	u8 lane_count;
+	u8 sink_count;
+	bool link_trained;
+	bool has_hdmi_sink;
+	bool has_audio;
+	bool reset_link_params;
+	bool use_max_params;
+	u8 dpcd[DP_RECEIVER_CAP_SIZE];
+	u8 psr_dpcd[EDP_PSR_RECEIVER_CAP_SIZE];
+	u8 downstream_ports[DP_MAX_DOWNSTREAM_PORTS];
+	u8 edp_dpcd[EDP_DISPLAY_CTL_CAP_SIZE];
+	u8 dsc_dpcd[DP_DSC_RECEIVER_CAP_SIZE];
+	u8 lttpr_common_caps[DP_LTTPR_COMMON_CAP_SIZE];
+	u8 lttpr_phy_caps[DP_MAX_LTTPR_COUNT][DP_LTTPR_PHY_CAP_SIZE];
+	u8 fec_capable;
+	u8 pcon_dsc_dpcd[DP_PCON_DSC_ENCODER_CAP_SIZE];
+	/* source rates */
+	int num_source_rates;
+	const int *source_rates;
+	/* sink rates as reported by DP_MAX_LINK_RATE/DP_SUPPORTED_LINK_RATES */
+	int num_sink_rates;
+	int sink_rates[DP_MAX_SUPPORTED_RATES];
+	bool use_rate_select;
+	/* Max sink lane count as reported by DP_MAX_LANE_COUNT */
+	int max_sink_lane_count;
+	/* intersection of source and sink rates */
+	int num_common_rates;
+	int common_rates[DP_MAX_SUPPORTED_RATES];
+	/* Max lane count for the current link */
+	int max_link_lane_count;
+	/* Max rate for the current link */
+	int max_link_rate;
+	int mso_link_count;
+	int mso_pixel_overlap;
+	/* sink or branch descriptor */
+	struct drm_dp_desc desc;
+	struct drm_dp_aux aux;
+	u32 aux_busy_last_status;
+	u8 train_set[4];
+
+	struct intel_pps pps;
+
+	bool is_mst;
+	int active_mst_links;
+
+	/* connector directly attached - won't be use for modeset in mst world */
+	struct intel_connector *attached_connector;
+
+	/* mst connector list */
+	struct intel_dp_mst_encoder *mst_encoders[I915_MAX_PIPES];
+	struct drm_dp_mst_topology_mgr mst_mgr;
+
+	u32 (*get_aux_clock_divider)(struct intel_dp *dp, int index);
+	/*
+	 * This function returns the value we have to program the AUX_CTL
+	 * register with to kick off an AUX transaction.
+	 */
+	u32 (*get_aux_send_ctl)(struct intel_dp *dp, int send_bytes,
+				u32 aux_clock_divider);
+
+	i915_reg_t (*aux_ch_ctl_reg)(struct intel_dp *dp);
+	i915_reg_t (*aux_ch_data_reg)(struct intel_dp *dp, int index);
+
+	/* This is called before a link training is starterd */
+	void (*prepare_link_retrain)(struct intel_dp *intel_dp,
+				     const struct intel_crtc_state *crtc_state);
+	void (*set_link_train)(struct intel_dp *intel_dp,
+			       const struct intel_crtc_state *crtc_state,
+			       u8 dp_train_pat);
+	void (*set_idle_link_train)(struct intel_dp *intel_dp,
+				    const struct intel_crtc_state *crtc_state);
+
+	u8 (*preemph_max)(struct intel_dp *intel_dp);
+	u8 (*voltage_max)(struct intel_dp *intel_dp,
+			  const struct intel_crtc_state *crtc_state);
+
+	/* Displayport compliance testing */
+	struct intel_dp_compliance compliance;
+
+	/* Downstream facing port caps */
+	struct {
+		int min_tmds_clock, max_tmds_clock;
+		int max_dotclock;
+		int pcon_max_frl_bw;
+		u8 max_bpc;
+		bool ycbcr_444_to_420;
+		bool rgb_to_ycbcr;
+	} dfp;
+
+	/* To control wakeup latency, e.g. for irq-driven dp aux transfers. */
+	struct pm_qos_request pm_qos;
+
+	/* Display stream compression testing */
+	bool force_dsc_en;
+	int force_dsc_bpc;
+
+	bool hobl_failed;
+	bool hobl_active;
+
+	struct intel_dp_pcon_frl frl;
+
+	struct intel_psr psr;
+
+	/* When we last wrote the OUI for eDP */
+	unsigned long last_oui_write;
+};
+
+enum lspcon_vendor {
+	LSPCON_VENDOR_MCA,
+	LSPCON_VENDOR_PARADE
+};
+
+struct intel_lspcon {
+	bool active;
+	bool hdr_supported;
+	enum drm_lspcon_mode mode;
+	enum lspcon_vendor vendor;
+};
+
+struct intel_digital_port {
+	struct intel_encoder base;
+	u32 saved_port_bits;
+	struct intel_dp dp;
+	struct intel_hdmi hdmi;
+	struct intel_lspcon lspcon;
+	enum irqreturn (*hpd_pulse)(struct intel_digital_port *, bool);
+	bool release_cl2_override;
+	u8 max_lanes;
+	/* Used for DP and ICL+ TypeC/DP and TypeC/HDMI ports. */
+	enum aux_ch aux_ch;
+	enum intel_display_power_domain ddi_io_power_domain;
+	intel_wakeref_t ddi_io_wakeref;
+	intel_wakeref_t aux_wakeref;
+
+	struct mutex tc_lock;	/* protects the TypeC port mode */
+	intel_wakeref_t tc_lock_wakeref;
+	enum intel_display_power_domain tc_lock_power_domain;
+	struct delayed_work tc_disconnect_phy_work;
+	int tc_link_refcount;
+	bool tc_legacy_port:1;
+	char tc_port_name[8];
+	enum tc_port_mode tc_mode;
+	enum tc_port_mode tc_init_mode;
+	enum phy_fia tc_phy_fia;
+	u8 tc_phy_fia_idx;
+
+	/* protects num_hdcp_streams reference count, hdcp_port_data and hdcp_auth_status */
+	struct mutex hdcp_mutex;
+	/* the number of pipes using HDCP signalling out of this port */
+	unsigned int num_hdcp_streams;
+	/* port HDCP auth status */
+	bool hdcp_auth_status;
+	/* HDCP port data need to pass to security f/w */
+	struct hdcp_port_data hdcp_port_data;
+	/* Whether the MST topology supports HDCP Type 1 Content */
+	bool hdcp_mst_type1_capable;
+
+	void (*write_infoframe)(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state,
+				unsigned int type,
+				const void *frame, ssize_t len);
+	void (*read_infoframe)(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state,
+			       unsigned int type,
+			       void *frame, ssize_t len);
+	void (*set_infoframes)(struct intel_encoder *encoder,
+			       bool enable,
+			       const struct intel_crtc_state *crtc_state,
+			       const struct drm_connector_state *conn_state);
+	u32 (*infoframes_enabled)(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *pipe_config);
+	bool (*connected)(struct intel_encoder *encoder);
+};
+
+struct intel_dp_mst_encoder {
+	struct intel_encoder base;
+	enum pipe pipe;
+	struct intel_digital_port *primary;
+	struct intel_connector *connector;
+};
+
+static inline enum dpio_channel
+vlv_dig_port_to_channel(struct intel_digital_port *dig_port)
+{
+	switch (dig_port->base.port) {
+	default:
+		MISSING_CASE(dig_port->base.port);
+		fallthrough;
+	case PORT_B:
+	case PORT_D:
+		return DPIO_CH0;
+	case PORT_C:
+		return DPIO_CH1;
+	}
+}
+
+static inline enum dpio_phy
+vlv_dig_port_to_phy(struct intel_digital_port *dig_port)
+{
+	switch (dig_port->base.port) {
+	default:
+		MISSING_CASE(dig_port->base.port);
+		fallthrough;
+	case PORT_B:
+	case PORT_C:
+		return DPIO_PHY0;
+	case PORT_D:
+		return DPIO_PHY1;
+	}
+}
+
+static inline enum dpio_channel
+vlv_pipe_to_channel(enum pipe pipe)
+{
+	switch (pipe) {
+	default:
+		MISSING_CASE(pipe);
+		fallthrough;
+	case PIPE_A:
+	case PIPE_C:
+		return DPIO_CH0;
+	case PIPE_B:
+		return DPIO_CH1;
+	}
+}
+
+struct intel_load_detect_pipe {
+	struct drm_atomic_state *restore_state;
+};
+
+static inline struct intel_encoder *
+intel_attached_encoder(struct intel_connector *connector)
+{
+	return connector->encoder;
+}
+
+static inline bool intel_encoder_is_dig_port(struct intel_encoder *encoder)
+{
+	switch (encoder->type) {
+	case INTEL_OUTPUT_DDI:
+	case INTEL_OUTPUT_DP:
+	case INTEL_OUTPUT_EDP:
+	case INTEL_OUTPUT_HDMI:
+		return true;
+	default:
+		return false;
+	}
+}
+
+static inline bool intel_encoder_is_mst(struct intel_encoder *encoder)
+{
+	return encoder->type == INTEL_OUTPUT_DP_MST;
+}
+
+static inline struct intel_dp_mst_encoder *
+enc_to_mst(struct intel_encoder *encoder)
+{
+	return container_of(&encoder->base, struct intel_dp_mst_encoder,
+			    base.base);
+}
+
+static inline struct intel_digital_port *
+enc_to_dig_port(struct intel_encoder *encoder)
+{
+	struct intel_encoder *intel_encoder = encoder;
+
+	if (intel_encoder_is_dig_port(intel_encoder))
+		return container_of(&encoder->base, struct intel_digital_port,
+				    base.base);
+	else if (intel_encoder_is_mst(intel_encoder))
+		return enc_to_mst(encoder)->primary;
+	else
+		return NULL;
+}
+
+static inline struct intel_digital_port *
+intel_attached_dig_port(struct intel_connector *connector)
+{
+	return enc_to_dig_port(intel_attached_encoder(connector));
+}
+
+static inline struct intel_hdmi *
+enc_to_intel_hdmi(struct intel_encoder *encoder)
+{
+	return &enc_to_dig_port(encoder)->hdmi;
+}
+
+static inline struct intel_hdmi *
+intel_attached_hdmi(struct intel_connector *connector)
+{
+	return enc_to_intel_hdmi(intel_attached_encoder(connector));
+}
+
+static inline struct intel_dp *enc_to_intel_dp(struct intel_encoder *encoder)
+{
+	return &enc_to_dig_port(encoder)->dp;
+}
+
+static inline struct intel_dp *intel_attached_dp(struct intel_connector *connector)
+{
+	return enc_to_intel_dp(intel_attached_encoder(connector));
+}
+
+static inline bool intel_encoder_is_dp(struct intel_encoder *encoder)
+{
+	switch (encoder->type) {
+	case INTEL_OUTPUT_DP:
+	case INTEL_OUTPUT_EDP:
+		return true;
+	case INTEL_OUTPUT_DDI:
+		/* Skip pure HDMI/DVI DDI encoders */
+		return i915_mmio_reg_valid(enc_to_intel_dp(encoder)->output_reg);
+	default:
+		return false;
+	}
+}
+
+static inline struct intel_lspcon *
+enc_to_intel_lspcon(struct intel_encoder *encoder)
+{
+	return &enc_to_dig_port(encoder)->lspcon;
+}
+
+static inline struct intel_digital_port *
+dp_to_dig_port(struct intel_dp *intel_dp)
+{
+	return container_of(intel_dp, struct intel_digital_port, dp);
+}
+
+static inline struct intel_lspcon *
+dp_to_lspcon(struct intel_dp *intel_dp)
+{
+	return &dp_to_dig_port(intel_dp)->lspcon;
+}
+
+#define dp_to_i915(__intel_dp) to_i915(dp_to_dig_port(__intel_dp)->base.base.dev)
+
+#define CAN_PSR(intel_dp) ((intel_dp)->psr.sink_support && \
+			   (intel_dp)->psr.source_support)
+
+static inline bool intel_encoder_can_psr(struct intel_encoder *encoder)
+{
+	if (!intel_encoder_is_dp(encoder))
+		return false;
+
+	return CAN_PSR(enc_to_intel_dp(encoder));
+}
+
+static inline struct intel_digital_port *
+hdmi_to_dig_port(struct intel_hdmi *intel_hdmi)
+{
+	return container_of(intel_hdmi, struct intel_digital_port, hdmi);
+}
+
+static inline struct intel_plane_state *
+intel_atomic_get_plane_state(struct intel_atomic_state *state,
+				 struct intel_plane *plane)
+{
+	struct drm_plane_state *ret =
+		drm_atomic_get_plane_state(&state->base, &plane->base);
+
+	if (IS_ERR(ret))
+		return ERR_CAST(ret);
+
+	return to_intel_plane_state(ret);
+}
+
+static inline struct intel_plane_state *
+intel_atomic_get_old_plane_state(struct intel_atomic_state *state,
+				 struct intel_plane *plane)
+{
+	return to_intel_plane_state(drm_atomic_get_old_plane_state(&state->base,
+								   &plane->base));
+}
+
+static inline struct intel_plane_state *
+intel_atomic_get_new_plane_state(struct intel_atomic_state *state,
+				 struct intel_plane *plane)
+{
+	return to_intel_plane_state(drm_atomic_get_new_plane_state(&state->base,
+								   &plane->base));
+}
+
+static inline struct intel_crtc_state *
+intel_atomic_get_old_crtc_state(struct intel_atomic_state *state,
+				struct intel_crtc *crtc)
+{
+	return to_intel_crtc_state(drm_atomic_get_old_crtc_state(&state->base,
+								 &crtc->base));
+}
+
+static inline struct intel_crtc_state *
+intel_atomic_get_new_crtc_state(struct intel_atomic_state *state,
+				struct intel_crtc *crtc)
+{
+	return to_intel_crtc_state(drm_atomic_get_new_crtc_state(&state->base,
+								 &crtc->base));
+}
+
+static inline struct intel_digital_connector_state *
+intel_atomic_get_new_connector_state(struct intel_atomic_state *state,
+				     struct intel_connector *connector)
+{
+	return to_intel_digital_connector_state(
+			drm_atomic_get_new_connector_state(&state->base,
+			&connector->base));
+}
+
+static inline struct intel_digital_connector_state *
+intel_atomic_get_old_connector_state(struct intel_atomic_state *state,
+				     struct intel_connector *connector)
+{
+	return to_intel_digital_connector_state(
+			drm_atomic_get_old_connector_state(&state->base,
+			&connector->base));
+}
+
+/* intel_display.c */
+static inline bool
+intel_crtc_has_type(const struct intel_crtc_state *crtc_state,
+		    enum intel_output_type type)
+{
+	return crtc_state->output_types & (1 << type);
+}
+static inline bool
+intel_crtc_has_dp_encoder(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->output_types &
+		((1 << INTEL_OUTPUT_DP) |
+		 (1 << INTEL_OUTPUT_DP_MST) |
+		 (1 << INTEL_OUTPUT_EDP));
+}
+
+static inline bool
+intel_crtc_needs_modeset(const struct intel_crtc_state *crtc_state)
+{
+	return drm_atomic_crtc_needs_modeset(&crtc_state->uapi);
+}
+
+static inline u32 intel_plane_ggtt_offset(const struct intel_plane_state *plane_state)
+{
+	return i915_ggtt_offset(plane_state->ggtt_vma);
+}
+
+static inline struct intel_frontbuffer *
+to_intel_frontbuffer(struct drm_framebuffer *fb)
+{
+	return fb ? to_intel_framebuffer(fb)->frontbuffer : NULL;
+}
+
+#endif /*  __INTEL_DISPLAY_TYPES_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dkl_phy.c b/drivers/gpu/drm/i915/display/intel_dkl_phy.c
new file mode 100644
index 000000000..710b030c7
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dkl_phy.c
@@ -0,0 +1,109 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+
+#include "intel_de.h"
+#include "intel_display.h"
+#include "intel_dkl_phy.h"
+
+static void
+dkl_phy_set_hip_idx(struct drm_i915_private *i915, i915_reg_t reg, int idx)
+{
+	enum tc_port tc_port = DKL_REG_TC_PORT(reg);
+
+	drm_WARN_ON(&i915->drm, tc_port < TC_PORT_1 || tc_port >= I915_MAX_TC_PORTS);
+
+	intel_de_write(i915,
+		       HIP_INDEX_REG(tc_port),
+		       HIP_INDEX_VAL(tc_port, idx));
+}
+
+/**
+ * intel_dkl_phy_read - read a Dekel PHY register
+ * @i915: i915 device instance
+ * @reg: Dekel PHY register
+ * @ln: lane instance of @reg
+ *
+ * Read the @reg Dekel PHY register.
+ *
+ * Returns the read value.
+ */
+u32
+intel_dkl_phy_read(struct drm_i915_private *i915, i915_reg_t reg, int ln)
+{
+	u32 val;
+
+	spin_lock(&i915->display.dkl.phy_lock);
+
+	dkl_phy_set_hip_idx(i915, reg, ln);
+	val = intel_de_read(i915, reg);
+
+	spin_unlock(&i915->display.dkl.phy_lock);
+
+	return val;
+}
+
+/**
+ * intel_dkl_phy_write - write a Dekel PHY register
+ * @i915: i915 device instance
+ * @reg: Dekel PHY register
+ * @ln: lane instance of @reg
+ * @val: value to write
+ *
+ * Write @val to the @reg Dekel PHY register.
+ */
+void
+intel_dkl_phy_write(struct drm_i915_private *i915, i915_reg_t reg, int ln, u32 val)
+{
+	spin_lock(&i915->display.dkl.phy_lock);
+
+	dkl_phy_set_hip_idx(i915, reg, ln);
+	intel_de_write(i915, reg, val);
+
+	spin_unlock(&i915->display.dkl.phy_lock);
+}
+
+/**
+ * intel_dkl_phy_rmw - read-modify-write a Dekel PHY register
+ * @i915: i915 device instance
+ * @reg: Dekel PHY register
+ * @ln: lane instance of @reg
+ * @clear: mask to clear
+ * @set: mask to set
+ *
+ * Read the @reg Dekel PHY register, clearing then setting the @clear/@set bits in it, and writing
+ * this value back to the register if the value differs from the read one.
+ */
+void
+intel_dkl_phy_rmw(struct drm_i915_private *i915, i915_reg_t reg, int ln, u32 clear, u32 set)
+{
+	spin_lock(&i915->display.dkl.phy_lock);
+
+	dkl_phy_set_hip_idx(i915, reg, ln);
+	intel_de_rmw(i915, reg, clear, set);
+
+	spin_unlock(&i915->display.dkl.phy_lock);
+}
+
+/**
+ * intel_dkl_phy_posting_read - do a posting read from a Dekel PHY register
+ * @i915: i915 device instance
+ * @reg: Dekel PHY register
+ * @ln: lane instance of @reg
+ *
+ * Read the @reg Dekel PHY register without returning the read value.
+ */
+void
+intel_dkl_phy_posting_read(struct drm_i915_private *i915, i915_reg_t reg, int ln)
+{
+	spin_lock(&i915->display.dkl.phy_lock);
+
+	dkl_phy_set_hip_idx(i915, reg, ln);
+	intel_de_posting_read(i915, reg);
+
+	spin_unlock(&i915->display.dkl.phy_lock);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dkl_phy.h b/drivers/gpu/drm/i915/display/intel_dkl_phy.h
new file mode 100644
index 000000000..260ad121a
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dkl_phy.h
@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_DKL_PHY_H__
+#define __INTEL_DKL_PHY_H__
+
+#include <linux/types.h>
+
+#include "i915_reg_defs.h"
+
+struct drm_i915_private;
+
+u32
+intel_dkl_phy_read(struct drm_i915_private *i915, i915_reg_t reg, int ln);
+void
+intel_dkl_phy_write(struct drm_i915_private *i915, i915_reg_t reg, int ln, u32 val);
+void
+intel_dkl_phy_rmw(struct drm_i915_private *i915, i915_reg_t reg, int ln, u32 clear, u32 set);
+void
+intel_dkl_phy_posting_read(struct drm_i915_private *i915, i915_reg_t reg, int ln);
+
+#endif /* __INTEL_DKL_PHY_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dmc.c b/drivers/gpu/drm/i915/display/intel_dmc.c
new file mode 100644
index 000000000..e52ecc073
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dmc.c
@@ -0,0 +1,1132 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_de.h"
+#include "intel_dmc.h"
+#include "intel_dmc_regs.h"
+
+/**
+ * DOC: DMC Firmware Support
+ *
+ * From gen9 onwards we have newly added DMC (Display microcontroller) in display
+ * engine to save and restore the state of display engine when it enter into
+ * low-power state and comes back to normal.
+ */
+
+#define DMC_VERSION(major, minor)	((major) << 16 | (minor))
+#define DMC_VERSION_MAJOR(version)	((version) >> 16)
+#define DMC_VERSION_MINOR(version)	((version) & 0xffff)
+
+#define DMC_PATH(platform, major, minor) \
+	"i915/"				 \
+	__stringify(platform) "_dmc_ver" \
+	__stringify(major) "_"		 \
+	__stringify(minor) ".bin"
+
+#define DISPLAY_VER13_DMC_MAX_FW_SIZE	0x20000
+
+#define DISPLAY_VER12_DMC_MAX_FW_SIZE	ICL_DMC_MAX_FW_SIZE
+
+#define DG2_DMC_PATH			DMC_PATH(dg2, 2, 07)
+#define DG2_DMC_VERSION_REQUIRED	DMC_VERSION(2, 07)
+MODULE_FIRMWARE(DG2_DMC_PATH);
+
+#define ADLP_DMC_PATH			DMC_PATH(adlp, 2, 16)
+#define ADLP_DMC_VERSION_REQUIRED	DMC_VERSION(2, 16)
+MODULE_FIRMWARE(ADLP_DMC_PATH);
+
+#define ADLS_DMC_PATH			DMC_PATH(adls, 2, 01)
+#define ADLS_DMC_VERSION_REQUIRED	DMC_VERSION(2, 1)
+MODULE_FIRMWARE(ADLS_DMC_PATH);
+
+#define DG1_DMC_PATH			DMC_PATH(dg1, 2, 02)
+#define DG1_DMC_VERSION_REQUIRED	DMC_VERSION(2, 2)
+MODULE_FIRMWARE(DG1_DMC_PATH);
+
+#define RKL_DMC_PATH			DMC_PATH(rkl, 2, 03)
+#define RKL_DMC_VERSION_REQUIRED	DMC_VERSION(2, 3)
+MODULE_FIRMWARE(RKL_DMC_PATH);
+
+#define TGL_DMC_PATH			DMC_PATH(tgl, 2, 12)
+#define TGL_DMC_VERSION_REQUIRED	DMC_VERSION(2, 12)
+MODULE_FIRMWARE(TGL_DMC_PATH);
+
+#define ICL_DMC_PATH			DMC_PATH(icl, 1, 09)
+#define ICL_DMC_VERSION_REQUIRED	DMC_VERSION(1, 9)
+#define ICL_DMC_MAX_FW_SIZE		0x6000
+MODULE_FIRMWARE(ICL_DMC_PATH);
+
+#define GLK_DMC_PATH			DMC_PATH(glk, 1, 04)
+#define GLK_DMC_VERSION_REQUIRED	DMC_VERSION(1, 4)
+#define GLK_DMC_MAX_FW_SIZE		0x4000
+MODULE_FIRMWARE(GLK_DMC_PATH);
+
+#define KBL_DMC_PATH			DMC_PATH(kbl, 1, 04)
+#define KBL_DMC_VERSION_REQUIRED	DMC_VERSION(1, 4)
+#define KBL_DMC_MAX_FW_SIZE		BXT_DMC_MAX_FW_SIZE
+MODULE_FIRMWARE(KBL_DMC_PATH);
+
+#define SKL_DMC_PATH			DMC_PATH(skl, 1, 27)
+#define SKL_DMC_VERSION_REQUIRED	DMC_VERSION(1, 27)
+#define SKL_DMC_MAX_FW_SIZE		BXT_DMC_MAX_FW_SIZE
+MODULE_FIRMWARE(SKL_DMC_PATH);
+
+#define BXT_DMC_PATH			DMC_PATH(bxt, 1, 07)
+#define BXT_DMC_VERSION_REQUIRED	DMC_VERSION(1, 7)
+#define BXT_DMC_MAX_FW_SIZE		0x3000
+MODULE_FIRMWARE(BXT_DMC_PATH);
+
+#define DMC_DEFAULT_FW_OFFSET		0xFFFFFFFF
+#define PACKAGE_MAX_FW_INFO_ENTRIES	20
+#define PACKAGE_V2_MAX_FW_INFO_ENTRIES	32
+#define DMC_V1_MAX_MMIO_COUNT		8
+#define DMC_V3_MAX_MMIO_COUNT		20
+#define DMC_V1_MMIO_START_RANGE		0x80000
+
+struct intel_css_header {
+	/* 0x09 for DMC */
+	u32 module_type;
+
+	/* Includes the DMC specific header in dwords */
+	u32 header_len;
+
+	/* always value would be 0x10000 */
+	u32 header_ver;
+
+	/* Not used */
+	u32 module_id;
+
+	/* Not used */
+	u32 module_vendor;
+
+	/* in YYYYMMDD format */
+	u32 date;
+
+	/* Size in dwords (CSS_Headerlen + PackageHeaderLen + dmc FWsLen)/4 */
+	u32 size;
+
+	/* Not used */
+	u32 key_size;
+
+	/* Not used */
+	u32 modulus_size;
+
+	/* Not used */
+	u32 exponent_size;
+
+	/* Not used */
+	u32 reserved1[12];
+
+	/* Major Minor */
+	u32 version;
+
+	/* Not used */
+	u32 reserved2[8];
+
+	/* Not used */
+	u32 kernel_header_info;
+} __packed;
+
+struct intel_fw_info {
+	u8 reserved1;
+
+	/* reserved on package_header version 1, must be 0 on version 2 */
+	u8 dmc_id;
+
+	/* Stepping (A, B, C, ..., *). * is a wildcard */
+	char stepping;
+
+	/* Sub-stepping (0, 1, ..., *). * is a wildcard */
+	char substepping;
+
+	u32 offset;
+	u32 reserved2;
+} __packed;
+
+struct intel_package_header {
+	/* DMC container header length in dwords */
+	u8 header_len;
+
+	/* 0x01, 0x02 */
+	u8 header_ver;
+
+	u8 reserved[10];
+
+	/* Number of valid entries in the FWInfo array below */
+	u32 num_entries;
+} __packed;
+
+struct intel_dmc_header_base {
+	/* always value would be 0x40403E3E */
+	u32 signature;
+
+	/* DMC binary header length */
+	u8 header_len;
+
+	/* 0x01 */
+	u8 header_ver;
+
+	/* Reserved */
+	u16 dmcc_ver;
+
+	/* Major, Minor */
+	u32 project;
+
+	/* Firmware program size (excluding header) in dwords */
+	u32 fw_size;
+
+	/* Major Minor version */
+	u32 fw_version;
+} __packed;
+
+struct intel_dmc_header_v1 {
+	struct intel_dmc_header_base base;
+
+	/* Number of valid MMIO cycles present. */
+	u32 mmio_count;
+
+	/* MMIO address */
+	u32 mmioaddr[DMC_V1_MAX_MMIO_COUNT];
+
+	/* MMIO data */
+	u32 mmiodata[DMC_V1_MAX_MMIO_COUNT];
+
+	/* FW filename  */
+	char dfile[32];
+
+	u32 reserved1[2];
+} __packed;
+
+struct intel_dmc_header_v3 {
+	struct intel_dmc_header_base base;
+
+	/* DMC RAM start MMIO address */
+	u32 start_mmioaddr;
+
+	u32 reserved[9];
+
+	/* FW filename */
+	char dfile[32];
+
+	/* Number of valid MMIO cycles present. */
+	u32 mmio_count;
+
+	/* MMIO address */
+	u32 mmioaddr[DMC_V3_MAX_MMIO_COUNT];
+
+	/* MMIO data */
+	u32 mmiodata[DMC_V3_MAX_MMIO_COUNT];
+} __packed;
+
+struct stepping_info {
+	char stepping;
+	char substepping;
+};
+
+static bool has_dmc_id_fw(struct drm_i915_private *i915, int dmc_id)
+{
+	return i915->display.dmc.dmc_info[dmc_id].payload;
+}
+
+bool intel_dmc_has_payload(struct drm_i915_private *i915)
+{
+	return has_dmc_id_fw(i915, DMC_FW_MAIN);
+}
+
+static const struct stepping_info *
+intel_get_stepping_info(struct drm_i915_private *i915,
+			struct stepping_info *si)
+{
+	const char *step_name = intel_step_name(RUNTIME_INFO(i915)->step.display_step);
+
+	si->stepping = step_name[0];
+	si->substepping = step_name[1];
+	return si;
+}
+
+static void gen9_set_dc_state_debugmask(struct drm_i915_private *dev_priv)
+{
+	/* The below bit doesn't need to be cleared ever afterwards */
+	intel_de_rmw(dev_priv, DC_STATE_DEBUG, 0,
+		     DC_STATE_DEBUG_MASK_CORES | DC_STATE_DEBUG_MASK_MEMORY_UP);
+	intel_de_posting_read(dev_priv, DC_STATE_DEBUG);
+}
+
+static void disable_event_handler(struct drm_i915_private *i915,
+				  i915_reg_t ctl_reg, i915_reg_t htp_reg)
+{
+	intel_de_write(i915, ctl_reg,
+		       REG_FIELD_PREP(DMC_EVT_CTL_TYPE_MASK,
+				      DMC_EVT_CTL_TYPE_EDGE_0_1) |
+		       REG_FIELD_PREP(DMC_EVT_CTL_EVENT_ID_MASK,
+				      DMC_EVT_CTL_EVENT_ID_FALSE));
+	intel_de_write(i915, htp_reg, 0);
+}
+
+static void
+disable_flip_queue_event(struct drm_i915_private *i915,
+			 i915_reg_t ctl_reg, i915_reg_t htp_reg)
+{
+	u32 event_ctl;
+	u32 event_htp;
+
+	event_ctl = intel_de_read(i915, ctl_reg);
+	event_htp = intel_de_read(i915, htp_reg);
+	if (event_ctl != (DMC_EVT_CTL_ENABLE |
+			  DMC_EVT_CTL_RECURRING |
+			  REG_FIELD_PREP(DMC_EVT_CTL_TYPE_MASK,
+					 DMC_EVT_CTL_TYPE_EDGE_0_1) |
+			  REG_FIELD_PREP(DMC_EVT_CTL_EVENT_ID_MASK,
+					 DMC_EVT_CTL_EVENT_ID_CLK_MSEC)) ||
+	    !event_htp) {
+		drm_dbg_kms(&i915->drm,
+			    "Unexpected DMC event configuration (control %08x htp %08x)\n",
+			    event_ctl, event_htp);
+		return;
+	}
+
+	disable_event_handler(i915, ctl_reg, htp_reg);
+}
+
+static bool
+get_flip_queue_event_regs(struct drm_i915_private *i915, int dmc_id,
+			  i915_reg_t *ctl_reg, i915_reg_t *htp_reg)
+{
+	switch (dmc_id) {
+	case DMC_FW_MAIN:
+		if (DISPLAY_VER(i915) == 12) {
+			*ctl_reg = DMC_EVT_CTL(i915, dmc_id, 3);
+			*htp_reg = DMC_EVT_HTP(i915, dmc_id, 3);
+
+			return true;
+		}
+		break;
+	case DMC_FW_PIPEA ... DMC_FW_PIPED:
+		if (IS_DG2(i915)) {
+			*ctl_reg = DMC_EVT_CTL(i915, dmc_id, 2);
+			*htp_reg = DMC_EVT_HTP(i915, dmc_id, 2);
+
+			return true;
+		}
+		break;
+	}
+
+	return false;
+}
+
+static void
+disable_all_flip_queue_events(struct drm_i915_private *i915)
+{
+	int dmc_id;
+
+	/* TODO: check if the following applies to all D13+ platforms. */
+	if (!IS_DG2(i915) && !IS_TIGERLAKE(i915))
+		return;
+
+	for (dmc_id = 0; dmc_id < DMC_FW_MAX; dmc_id++) {
+		i915_reg_t ctl_reg;
+		i915_reg_t htp_reg;
+
+		if (!has_dmc_id_fw(i915, dmc_id))
+			continue;
+
+		if (!get_flip_queue_event_regs(i915, dmc_id, &ctl_reg, &htp_reg))
+			continue;
+
+		disable_flip_queue_event(i915, ctl_reg, htp_reg);
+	}
+}
+
+static void disable_all_event_handlers(struct drm_i915_private *i915)
+{
+	int id;
+
+	/* TODO: disable the event handlers on pre-GEN12 platforms as well */
+	if (DISPLAY_VER(i915) < 12)
+		return;
+
+	for (id = DMC_FW_MAIN; id < DMC_FW_MAX; id++) {
+		int handler;
+
+		if (!has_dmc_id_fw(i915, id))
+			continue;
+
+		for (handler = 0; handler < DMC_EVENT_HANDLER_COUNT_GEN12; handler++)
+			disable_event_handler(i915,
+					      DMC_EVT_CTL(i915, id, handler),
+					      DMC_EVT_HTP(i915, id, handler));
+	}
+}
+
+static void pipedmc_clock_gating_wa(struct drm_i915_private *i915, bool enable)
+{
+	enum pipe pipe;
+
+	if (DISPLAY_VER(i915) != 13)
+		return;
+
+	/*
+	 * Wa_16015201720:adl-p,dg2
+	 * The WA requires clock gating to be disabled all the time
+	 * for pipe A and B.
+	 * For pipe C and D clock gating needs to be disabled only
+	 * during initializing the firmware.
+	 */
+	if (enable)
+		for (pipe = PIPE_A; pipe <= PIPE_D; pipe++)
+			intel_de_rmw(i915, CLKGATE_DIS_PSL_EXT(pipe),
+				     0, PIPEDMC_GATING_DIS);
+	else
+		for (pipe = PIPE_C; pipe <= PIPE_D; pipe++)
+			intel_de_rmw(i915, CLKGATE_DIS_PSL_EXT(pipe),
+				     PIPEDMC_GATING_DIS, 0);
+}
+
+/**
+ * intel_dmc_load_program() - write the firmware from memory to register.
+ * @dev_priv: i915 drm device.
+ *
+ * DMC firmware is read from a .bin file and kept in internal memory one time.
+ * Everytime display comes back from low power state this function is called to
+ * copy the firmware from internal memory to registers.
+ */
+void intel_dmc_load_program(struct drm_i915_private *dev_priv)
+{
+	struct intel_dmc *dmc = &dev_priv->display.dmc;
+	u32 id, i;
+
+	if (!intel_dmc_has_payload(dev_priv))
+		return;
+
+	pipedmc_clock_gating_wa(dev_priv, true);
+
+	disable_all_event_handlers(dev_priv);
+
+	assert_rpm_wakelock_held(&dev_priv->runtime_pm);
+
+	preempt_disable();
+
+	for (id = 0; id < DMC_FW_MAX; id++) {
+		for (i = 0; i < dmc->dmc_info[id].dmc_fw_size; i++) {
+			intel_uncore_write_fw(&dev_priv->uncore,
+					      DMC_PROGRAM(dmc->dmc_info[id].start_mmioaddr, i),
+					      dmc->dmc_info[id].payload[i]);
+		}
+	}
+
+	preempt_enable();
+
+	for (id = 0; id < DMC_FW_MAX; id++) {
+		for (i = 0; i < dmc->dmc_info[id].mmio_count; i++) {
+			intel_de_write(dev_priv, dmc->dmc_info[id].mmioaddr[i],
+				       dmc->dmc_info[id].mmiodata[i]);
+		}
+	}
+
+	dev_priv->display.dmc.dc_state = 0;
+
+	gen9_set_dc_state_debugmask(dev_priv);
+
+	/*
+	 * Flip queue events need to be disabled before enabling DC5/6.
+	 * i915 doesn't use the flip queue feature, so disable it already
+	 * here.
+	 */
+	disable_all_flip_queue_events(dev_priv);
+
+	pipedmc_clock_gating_wa(dev_priv, false);
+}
+
+/**
+ * intel_dmc_disable_program() - disable the firmware
+ * @i915: i915 drm device
+ *
+ * Disable all event handlers in the firmware, making sure the firmware is
+ * inactive after the display is uninitialized.
+ */
+void intel_dmc_disable_program(struct drm_i915_private *i915)
+{
+	if (!intel_dmc_has_payload(i915))
+		return;
+
+	pipedmc_clock_gating_wa(i915, true);
+	disable_all_event_handlers(i915);
+	pipedmc_clock_gating_wa(i915, false);
+}
+
+void assert_dmc_loaded(struct drm_i915_private *i915)
+{
+	drm_WARN_ONCE(&i915->drm,
+		      !intel_de_read(i915, DMC_PROGRAM(i915->display.dmc.dmc_info[DMC_FW_MAIN].start_mmioaddr, 0)),
+		      "DMC program storage start is NULL\n");
+	drm_WARN_ONCE(&i915->drm, !intel_de_read(i915, DMC_SSP_BASE),
+		      "DMC SSP Base Not fine\n");
+	drm_WARN_ONCE(&i915->drm, !intel_de_read(i915, DMC_HTP_SKL),
+		      "DMC HTP Not fine\n");
+}
+
+static bool fw_info_matches_stepping(const struct intel_fw_info *fw_info,
+				     const struct stepping_info *si)
+{
+	if ((fw_info->substepping == '*' && si->stepping == fw_info->stepping) ||
+	    (si->stepping == fw_info->stepping && si->substepping == fw_info->substepping) ||
+	    /*
+	     * If we don't find a more specific one from above two checks, we
+	     * then check for the generic one to be sure to work even with
+	     * "broken firmware"
+	     */
+	    (si->stepping == '*' && si->substepping == fw_info->substepping) ||
+	    (fw_info->stepping == '*' && fw_info->substepping == '*'))
+		return true;
+
+	return false;
+}
+
+/*
+ * Search fw_info table for dmc_offset to find firmware binary: num_entries is
+ * already sanitized.
+ */
+static void dmc_set_fw_offset(struct intel_dmc *dmc,
+			      const struct intel_fw_info *fw_info,
+			      unsigned int num_entries,
+			      const struct stepping_info *si,
+			      u8 package_ver)
+{
+	unsigned int i, id;
+
+	struct drm_i915_private *i915 = container_of(dmc, typeof(*i915), display.dmc);
+
+	for (i = 0; i < num_entries; i++) {
+		id = package_ver <= 1 ? DMC_FW_MAIN : fw_info[i].dmc_id;
+
+		if (id >= DMC_FW_MAX) {
+			drm_dbg(&i915->drm, "Unsupported firmware id: %u\n", id);
+			continue;
+		}
+
+		/* More specific versions come first, so we don't even have to
+		 * check for the stepping since we already found a previous FW
+		 * for this id.
+		 */
+		if (dmc->dmc_info[id].present)
+			continue;
+
+		if (fw_info_matches_stepping(&fw_info[i], si)) {
+			dmc->dmc_info[id].present = true;
+			dmc->dmc_info[id].dmc_offset = fw_info[i].offset;
+		}
+	}
+}
+
+static bool dmc_mmio_addr_sanity_check(struct intel_dmc *dmc,
+				       const u32 *mmioaddr, u32 mmio_count,
+				       int header_ver, u8 dmc_id)
+{
+	struct drm_i915_private *i915 = container_of(dmc, typeof(*i915), display.dmc);
+	u32 start_range, end_range;
+	int i;
+
+	if (dmc_id >= DMC_FW_MAX) {
+		drm_warn(&i915->drm, "Unsupported firmware id %u\n", dmc_id);
+		return false;
+	}
+
+	if (header_ver == 1) {
+		start_range = DMC_MMIO_START_RANGE;
+		end_range = DMC_MMIO_END_RANGE;
+	} else if (dmc_id == DMC_FW_MAIN) {
+		start_range = TGL_MAIN_MMIO_START;
+		end_range = TGL_MAIN_MMIO_END;
+	} else if (DISPLAY_VER(i915) >= 13) {
+		start_range = ADLP_PIPE_MMIO_START;
+		end_range = ADLP_PIPE_MMIO_END;
+	} else if (DISPLAY_VER(i915) >= 12) {
+		start_range = TGL_PIPE_MMIO_START(dmc_id);
+		end_range = TGL_PIPE_MMIO_END(dmc_id);
+	} else {
+		drm_warn(&i915->drm, "Unknown mmio range for sanity check");
+		return false;
+	}
+
+	for (i = 0; i < mmio_count; i++) {
+		if (mmioaddr[i] < start_range || mmioaddr[i] > end_range)
+			return false;
+	}
+
+	return true;
+}
+
+static u32 parse_dmc_fw_header(struct intel_dmc *dmc,
+			       const struct intel_dmc_header_base *dmc_header,
+			       size_t rem_size, u8 dmc_id)
+{
+	struct drm_i915_private *i915 = container_of(dmc, typeof(*i915), display.dmc);
+	struct dmc_fw_info *dmc_info = &dmc->dmc_info[dmc_id];
+	unsigned int header_len_bytes, dmc_header_size, payload_size, i;
+	const u32 *mmioaddr, *mmiodata;
+	u32 mmio_count, mmio_count_max, start_mmioaddr;
+	u8 *payload;
+
+	BUILD_BUG_ON(ARRAY_SIZE(dmc_info->mmioaddr) < DMC_V3_MAX_MMIO_COUNT ||
+		     ARRAY_SIZE(dmc_info->mmioaddr) < DMC_V1_MAX_MMIO_COUNT);
+
+	/*
+	 * Check if we can access common fields, we will checkc again below
+	 * after we have read the version
+	 */
+	if (rem_size < sizeof(struct intel_dmc_header_base))
+		goto error_truncated;
+
+	/* Cope with small differences between v1 and v3 */
+	if (dmc_header->header_ver == 3) {
+		const struct intel_dmc_header_v3 *v3 =
+			(const struct intel_dmc_header_v3 *)dmc_header;
+
+		if (rem_size < sizeof(struct intel_dmc_header_v3))
+			goto error_truncated;
+
+		mmioaddr = v3->mmioaddr;
+		mmiodata = v3->mmiodata;
+		mmio_count = v3->mmio_count;
+		mmio_count_max = DMC_V3_MAX_MMIO_COUNT;
+		/* header_len is in dwords */
+		header_len_bytes = dmc_header->header_len * 4;
+		start_mmioaddr = v3->start_mmioaddr;
+		dmc_header_size = sizeof(*v3);
+	} else if (dmc_header->header_ver == 1) {
+		const struct intel_dmc_header_v1 *v1 =
+			(const struct intel_dmc_header_v1 *)dmc_header;
+
+		if (rem_size < sizeof(struct intel_dmc_header_v1))
+			goto error_truncated;
+
+		mmioaddr = v1->mmioaddr;
+		mmiodata = v1->mmiodata;
+		mmio_count = v1->mmio_count;
+		mmio_count_max = DMC_V1_MAX_MMIO_COUNT;
+		header_len_bytes = dmc_header->header_len;
+		start_mmioaddr = DMC_V1_MMIO_START_RANGE;
+		dmc_header_size = sizeof(*v1);
+	} else {
+		drm_err(&i915->drm, "Unknown DMC fw header version: %u\n",
+			dmc_header->header_ver);
+		return 0;
+	}
+
+	if (header_len_bytes != dmc_header_size) {
+		drm_err(&i915->drm, "DMC firmware has wrong dmc header length "
+			"(%u bytes)\n", header_len_bytes);
+		return 0;
+	}
+
+	/* Cache the dmc header info. */
+	if (mmio_count > mmio_count_max) {
+		drm_err(&i915->drm, "DMC firmware has wrong mmio count %u\n", mmio_count);
+		return 0;
+	}
+
+	if (!dmc_mmio_addr_sanity_check(dmc, mmioaddr, mmio_count,
+					dmc_header->header_ver, dmc_id)) {
+		drm_err(&i915->drm, "DMC firmware has Wrong MMIO Addresses\n");
+		return 0;
+	}
+
+	for (i = 0; i < mmio_count; i++) {
+		dmc_info->mmioaddr[i] = _MMIO(mmioaddr[i]);
+		dmc_info->mmiodata[i] = mmiodata[i];
+	}
+	dmc_info->mmio_count = mmio_count;
+	dmc_info->start_mmioaddr = start_mmioaddr;
+
+	rem_size -= header_len_bytes;
+
+	/* fw_size is in dwords, so multiplied by 4 to convert into bytes. */
+	payload_size = dmc_header->fw_size * 4;
+	if (rem_size < payload_size)
+		goto error_truncated;
+
+	if (payload_size > dmc->max_fw_size) {
+		drm_err(&i915->drm, "DMC FW too big (%u bytes)\n", payload_size);
+		return 0;
+	}
+	dmc_info->dmc_fw_size = dmc_header->fw_size;
+
+	dmc_info->payload = kmalloc(payload_size, GFP_KERNEL);
+	if (!dmc_info->payload)
+		return 0;
+
+	payload = (u8 *)(dmc_header) + header_len_bytes;
+	memcpy(dmc_info->payload, payload, payload_size);
+
+	return header_len_bytes + payload_size;
+
+error_truncated:
+	drm_err(&i915->drm, "Truncated DMC firmware, refusing.\n");
+	return 0;
+}
+
+static u32
+parse_dmc_fw_package(struct intel_dmc *dmc,
+		     const struct intel_package_header *package_header,
+		     const struct stepping_info *si,
+		     size_t rem_size)
+{
+	struct drm_i915_private *i915 = container_of(dmc, typeof(*i915), display.dmc);
+	u32 package_size = sizeof(struct intel_package_header);
+	u32 num_entries, max_entries;
+	const struct intel_fw_info *fw_info;
+
+	if (rem_size < package_size)
+		goto error_truncated;
+
+	if (package_header->header_ver == 1) {
+		max_entries = PACKAGE_MAX_FW_INFO_ENTRIES;
+	} else if (package_header->header_ver == 2) {
+		max_entries = PACKAGE_V2_MAX_FW_INFO_ENTRIES;
+	} else {
+		drm_err(&i915->drm, "DMC firmware has unknown header version %u\n",
+			package_header->header_ver);
+		return 0;
+	}
+
+	/*
+	 * We should always have space for max_entries,
+	 * even if not all are used
+	 */
+	package_size += max_entries * sizeof(struct intel_fw_info);
+	if (rem_size < package_size)
+		goto error_truncated;
+
+	if (package_header->header_len * 4 != package_size) {
+		drm_err(&i915->drm, "DMC firmware has wrong package header length "
+			"(%u bytes)\n", package_size);
+		return 0;
+	}
+
+	num_entries = package_header->num_entries;
+	if (WARN_ON(package_header->num_entries > max_entries))
+		num_entries = max_entries;
+
+	fw_info = (const struct intel_fw_info *)
+		((u8 *)package_header + sizeof(*package_header));
+	dmc_set_fw_offset(dmc, fw_info, num_entries, si,
+			  package_header->header_ver);
+
+	/* dmc_offset is in dwords */
+	return package_size;
+
+error_truncated:
+	drm_err(&i915->drm, "Truncated DMC firmware, refusing.\n");
+	return 0;
+}
+
+/* Return number of bytes parsed or 0 on error */
+static u32 parse_dmc_fw_css(struct intel_dmc *dmc,
+			    struct intel_css_header *css_header,
+			    size_t rem_size)
+{
+	struct drm_i915_private *i915 = container_of(dmc, typeof(*i915), display.dmc);
+
+	if (rem_size < sizeof(struct intel_css_header)) {
+		drm_err(&i915->drm, "Truncated DMC firmware, refusing.\n");
+		return 0;
+	}
+
+	if (sizeof(struct intel_css_header) !=
+	    (css_header->header_len * 4)) {
+		drm_err(&i915->drm, "DMC firmware has wrong CSS header length "
+			"(%u bytes)\n",
+			(css_header->header_len * 4));
+		return 0;
+	}
+
+	if (dmc->required_version &&
+	    css_header->version != dmc->required_version) {
+		drm_info(&i915->drm, "Refusing to load DMC firmware v%u.%u,"
+			 " please use v%u.%u\n",
+			 DMC_VERSION_MAJOR(css_header->version),
+			 DMC_VERSION_MINOR(css_header->version),
+			 DMC_VERSION_MAJOR(dmc->required_version),
+			 DMC_VERSION_MINOR(dmc->required_version));
+		return 0;
+	}
+
+	dmc->version = css_header->version;
+
+	return sizeof(struct intel_css_header);
+}
+
+static void parse_dmc_fw(struct drm_i915_private *dev_priv,
+			 const struct firmware *fw)
+{
+	struct intel_css_header *css_header;
+	struct intel_package_header *package_header;
+	struct intel_dmc_header_base *dmc_header;
+	struct intel_dmc *dmc = &dev_priv->display.dmc;
+	struct stepping_info display_info = { '*', '*'};
+	const struct stepping_info *si = intel_get_stepping_info(dev_priv, &display_info);
+	u32 readcount = 0;
+	u32 r, offset;
+	int id;
+
+	if (!fw)
+		return;
+
+	/* Extract CSS Header information */
+	css_header = (struct intel_css_header *)fw->data;
+	r = parse_dmc_fw_css(dmc, css_header, fw->size);
+	if (!r)
+		return;
+
+	readcount += r;
+
+	/* Extract Package Header information */
+	package_header = (struct intel_package_header *)&fw->data[readcount];
+	r = parse_dmc_fw_package(dmc, package_header, si, fw->size - readcount);
+	if (!r)
+		return;
+
+	readcount += r;
+
+	for (id = 0; id < DMC_FW_MAX; id++) {
+		if (!dev_priv->display.dmc.dmc_info[id].present)
+			continue;
+
+		offset = readcount + dmc->dmc_info[id].dmc_offset * 4;
+		if (offset > fw->size) {
+			drm_err(&dev_priv->drm, "Reading beyond the fw_size\n");
+			continue;
+		}
+
+		dmc_header = (struct intel_dmc_header_base *)&fw->data[offset];
+		parse_dmc_fw_header(dmc, dmc_header, fw->size - offset, id);
+	}
+}
+
+static void intel_dmc_runtime_pm_get(struct drm_i915_private *dev_priv)
+{
+	drm_WARN_ON(&dev_priv->drm, dev_priv->display.dmc.wakeref);
+	dev_priv->display.dmc.wakeref =
+		intel_display_power_get(dev_priv, POWER_DOMAIN_INIT);
+}
+
+static void intel_dmc_runtime_pm_put(struct drm_i915_private *dev_priv)
+{
+	intel_wakeref_t wakeref __maybe_unused =
+		fetch_and_zero(&dev_priv->display.dmc.wakeref);
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_INIT, wakeref);
+}
+
+static void dmc_load_work_fn(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv;
+	struct intel_dmc *dmc;
+	const struct firmware *fw = NULL;
+
+	dev_priv = container_of(work, typeof(*dev_priv), display.dmc.work);
+	dmc = &dev_priv->display.dmc;
+
+	request_firmware(&fw, dev_priv->display.dmc.fw_path, dev_priv->drm.dev);
+	parse_dmc_fw(dev_priv, fw);
+
+	if (intel_dmc_has_payload(dev_priv)) {
+		intel_dmc_load_program(dev_priv);
+		intel_dmc_runtime_pm_put(dev_priv);
+
+		drm_info(&dev_priv->drm,
+			 "Finished loading DMC firmware %s (v%u.%u)\n",
+			 dev_priv->display.dmc.fw_path, DMC_VERSION_MAJOR(dmc->version),
+			 DMC_VERSION_MINOR(dmc->version));
+	} else {
+		drm_notice(&dev_priv->drm,
+			   "Failed to load DMC firmware %s."
+			   " Disabling runtime power management.\n",
+			   dmc->fw_path);
+		drm_notice(&dev_priv->drm, "DMC firmware homepage: %s",
+			   INTEL_UC_FIRMWARE_URL);
+	}
+
+	release_firmware(fw);
+}
+
+/**
+ * intel_dmc_ucode_init() - initialize the firmware loading.
+ * @dev_priv: i915 drm device.
+ *
+ * This function is called at the time of loading the display driver to read
+ * firmware from a .bin file and copied into a internal memory.
+ */
+void intel_dmc_ucode_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_dmc *dmc = &dev_priv->display.dmc;
+
+	INIT_WORK(&dev_priv->display.dmc.work, dmc_load_work_fn);
+
+	if (!HAS_DMC(dev_priv))
+		return;
+
+	/*
+	 * Obtain a runtime pm reference, until DMC is loaded, to avoid entering
+	 * runtime-suspend.
+	 *
+	 * On error, we return with the rpm wakeref held to prevent runtime
+	 * suspend as runtime suspend *requires* a working DMC for whatever
+	 * reason.
+	 */
+	intel_dmc_runtime_pm_get(dev_priv);
+
+	if (IS_DG2(dev_priv)) {
+		dmc->fw_path = DG2_DMC_PATH;
+		dmc->required_version = DG2_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = DISPLAY_VER13_DMC_MAX_FW_SIZE;
+	} else if (IS_ALDERLAKE_P(dev_priv)) {
+		dmc->fw_path = ADLP_DMC_PATH;
+		dmc->required_version = ADLP_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = DISPLAY_VER13_DMC_MAX_FW_SIZE;
+	} else if (IS_ALDERLAKE_S(dev_priv)) {
+		dmc->fw_path = ADLS_DMC_PATH;
+		dmc->required_version = ADLS_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = DISPLAY_VER12_DMC_MAX_FW_SIZE;
+	} else if (IS_DG1(dev_priv)) {
+		dmc->fw_path = DG1_DMC_PATH;
+		dmc->required_version = DG1_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = DISPLAY_VER12_DMC_MAX_FW_SIZE;
+	} else if (IS_ROCKETLAKE(dev_priv)) {
+		dmc->fw_path = RKL_DMC_PATH;
+		dmc->required_version = RKL_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = DISPLAY_VER12_DMC_MAX_FW_SIZE;
+	} else if (IS_TIGERLAKE(dev_priv)) {
+		dmc->fw_path = TGL_DMC_PATH;
+		dmc->required_version = TGL_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = DISPLAY_VER12_DMC_MAX_FW_SIZE;
+	} else if (DISPLAY_VER(dev_priv) == 11) {
+		dmc->fw_path = ICL_DMC_PATH;
+		dmc->required_version = ICL_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = ICL_DMC_MAX_FW_SIZE;
+	} else if (IS_GEMINILAKE(dev_priv)) {
+		dmc->fw_path = GLK_DMC_PATH;
+		dmc->required_version = GLK_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = GLK_DMC_MAX_FW_SIZE;
+	} else if (IS_KABYLAKE(dev_priv) ||
+		   IS_COFFEELAKE(dev_priv) ||
+		   IS_COMETLAKE(dev_priv)) {
+		dmc->fw_path = KBL_DMC_PATH;
+		dmc->required_version = KBL_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = KBL_DMC_MAX_FW_SIZE;
+	} else if (IS_SKYLAKE(dev_priv)) {
+		dmc->fw_path = SKL_DMC_PATH;
+		dmc->required_version = SKL_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = SKL_DMC_MAX_FW_SIZE;
+	} else if (IS_BROXTON(dev_priv)) {
+		dmc->fw_path = BXT_DMC_PATH;
+		dmc->required_version = BXT_DMC_VERSION_REQUIRED;
+		dmc->max_fw_size = BXT_DMC_MAX_FW_SIZE;
+	}
+
+	if (dev_priv->params.dmc_firmware_path) {
+		if (strlen(dev_priv->params.dmc_firmware_path) == 0) {
+			dmc->fw_path = NULL;
+			drm_info(&dev_priv->drm,
+				 "Disabling DMC firmware and runtime PM\n");
+			return;
+		}
+
+		dmc->fw_path = dev_priv->params.dmc_firmware_path;
+		/* Bypass version check for firmware override. */
+		dmc->required_version = 0;
+	}
+
+	if (!dmc->fw_path) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "No known DMC firmware for platform, disabling runtime PM\n");
+		return;
+	}
+
+	drm_dbg_kms(&dev_priv->drm, "Loading %s\n", dmc->fw_path);
+	schedule_work(&dev_priv->display.dmc.work);
+}
+
+/**
+ * intel_dmc_ucode_suspend() - prepare DMC firmware before system suspend
+ * @dev_priv: i915 drm device
+ *
+ * Prepare the DMC firmware before entering system suspend. This includes
+ * flushing pending work items and releasing any resources acquired during
+ * init.
+ */
+void intel_dmc_ucode_suspend(struct drm_i915_private *dev_priv)
+{
+	if (!HAS_DMC(dev_priv))
+		return;
+
+	flush_work(&dev_priv->display.dmc.work);
+
+	/* Drop the reference held in case DMC isn't loaded. */
+	if (!intel_dmc_has_payload(dev_priv))
+		intel_dmc_runtime_pm_put(dev_priv);
+}
+
+/**
+ * intel_dmc_ucode_resume() - init DMC firmware during system resume
+ * @dev_priv: i915 drm device
+ *
+ * Reinitialize the DMC firmware during system resume, reacquiring any
+ * resources released in intel_dmc_ucode_suspend().
+ */
+void intel_dmc_ucode_resume(struct drm_i915_private *dev_priv)
+{
+	if (!HAS_DMC(dev_priv))
+		return;
+
+	/*
+	 * Reacquire the reference to keep RPM disabled in case DMC isn't
+	 * loaded.
+	 */
+	if (!intel_dmc_has_payload(dev_priv))
+		intel_dmc_runtime_pm_get(dev_priv);
+}
+
+/**
+ * intel_dmc_ucode_fini() - unload the DMC firmware.
+ * @dev_priv: i915 drm device.
+ *
+ * Firmmware unloading includes freeing the internal memory and reset the
+ * firmware loading status.
+ */
+void intel_dmc_ucode_fini(struct drm_i915_private *dev_priv)
+{
+	int id;
+
+	if (!HAS_DMC(dev_priv))
+		return;
+
+	intel_dmc_ucode_suspend(dev_priv);
+	drm_WARN_ON(&dev_priv->drm, dev_priv->display.dmc.wakeref);
+
+	for (id = 0; id < DMC_FW_MAX; id++)
+		kfree(dev_priv->display.dmc.dmc_info[id].payload);
+}
+
+void intel_dmc_print_error_state(struct drm_i915_error_state_buf *m,
+				 struct drm_i915_private *i915)
+{
+	struct intel_dmc *dmc = &i915->display.dmc;
+
+	if (!HAS_DMC(i915))
+		return;
+
+	i915_error_printf(m, "DMC loaded: %s\n",
+			  str_yes_no(intel_dmc_has_payload(i915)));
+	i915_error_printf(m, "DMC fw version: %d.%d\n",
+			  DMC_VERSION_MAJOR(dmc->version),
+			  DMC_VERSION_MINOR(dmc->version));
+}
+
+static int intel_dmc_debugfs_status_show(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *i915 = m->private;
+	intel_wakeref_t wakeref;
+	struct intel_dmc *dmc;
+	i915_reg_t dc5_reg, dc6_reg = INVALID_MMIO_REG;
+
+	if (!HAS_DMC(i915))
+		return -ENODEV;
+
+	dmc = &i915->display.dmc;
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	seq_printf(m, "fw loaded: %s\n",
+		   str_yes_no(intel_dmc_has_payload(i915)));
+	seq_printf(m, "path: %s\n", dmc->fw_path);
+	seq_printf(m, "Pipe A fw support: %s\n",
+		   str_yes_no(GRAPHICS_VER(i915) >= 12));
+	seq_printf(m, "Pipe A fw loaded: %s\n",
+		   str_yes_no(dmc->dmc_info[DMC_FW_PIPEA].payload));
+	seq_printf(m, "Pipe B fw support: %s\n",
+		   str_yes_no(IS_ALDERLAKE_P(i915)));
+	seq_printf(m, "Pipe B fw loaded: %s\n",
+		   str_yes_no(dmc->dmc_info[DMC_FW_PIPEB].payload));
+
+	if (!intel_dmc_has_payload(i915))
+		goto out;
+
+	seq_printf(m, "version: %d.%d\n", DMC_VERSION_MAJOR(dmc->version),
+		   DMC_VERSION_MINOR(dmc->version));
+
+	if (DISPLAY_VER(i915) >= 12) {
+		if (IS_DGFX(i915)) {
+			dc5_reg = DG1_DMC_DEBUG_DC5_COUNT;
+		} else {
+			dc5_reg = TGL_DMC_DEBUG_DC5_COUNT;
+			dc6_reg = TGL_DMC_DEBUG_DC6_COUNT;
+		}
+
+		/*
+		 * NOTE: DMC_DEBUG3 is a general purpose reg.
+		 * According to B.Specs:49196 DMC f/w reuses DC5/6 counter
+		 * reg for DC3CO debugging and validation,
+		 * but TGL DMC f/w is using DMC_DEBUG3 reg for DC3CO counter.
+		 */
+		seq_printf(m, "DC3CO count: %d\n",
+			   intel_de_read(i915, IS_DGFX(i915) ?
+					 DG1_DMC_DEBUG3 : TGL_DMC_DEBUG3));
+	} else {
+		dc5_reg = IS_BROXTON(i915) ? BXT_DMC_DC3_DC5_COUNT :
+			SKL_DMC_DC3_DC5_COUNT;
+		if (!IS_GEMINILAKE(i915) && !IS_BROXTON(i915))
+			dc6_reg = SKL_DMC_DC5_DC6_COUNT;
+	}
+
+	seq_printf(m, "DC3 -> DC5 count: %d\n", intel_de_read(i915, dc5_reg));
+	if (i915_mmio_reg_valid(dc6_reg))
+		seq_printf(m, "DC5 -> DC6 count: %d\n",
+			   intel_de_read(i915, dc6_reg));
+
+out:
+	seq_printf(m, "program base: 0x%08x\n",
+		   intel_de_read(i915, DMC_PROGRAM(dmc->dmc_info[DMC_FW_MAIN].start_mmioaddr, 0)));
+	seq_printf(m, "ssp base: 0x%08x\n",
+		   intel_de_read(i915, DMC_SSP_BASE));
+	seq_printf(m, "htp: 0x%08x\n", intel_de_read(i915, DMC_HTP_SKL));
+
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+
+	return 0;
+}
+
+DEFINE_SHOW_ATTRIBUTE(intel_dmc_debugfs_status);
+
+void intel_dmc_debugfs_register(struct drm_i915_private *i915)
+{
+	struct drm_minor *minor = i915->drm.primary;
+
+	debugfs_create_file("i915_dmc_info", 0444, minor->debugfs_root,
+			    i915, &intel_dmc_debugfs_status_fops);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dmc.h b/drivers/gpu/drm/i915/display/intel_dmc.h
new file mode 100644
index 000000000..67e03315e
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dmc.h
@@ -0,0 +1,61 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DMC_H__
+#define __INTEL_DMC_H__
+
+#include "i915_reg_defs.h"
+#include "intel_wakeref.h"
+#include <linux/workqueue.h>
+
+struct drm_i915_error_state_buf;
+struct drm_i915_private;
+
+enum {
+	DMC_FW_MAIN = 0,
+	DMC_FW_PIPEA,
+	DMC_FW_PIPEB,
+	DMC_FW_PIPEC,
+	DMC_FW_PIPED,
+	DMC_FW_MAX
+};
+
+struct intel_dmc {
+	struct work_struct work;
+	const char *fw_path;
+	u32 required_version;
+	u32 max_fw_size; /* bytes */
+	u32 version;
+	struct dmc_fw_info {
+		u32 mmio_count;
+		i915_reg_t mmioaddr[20];
+		u32 mmiodata[20];
+		u32 dmc_offset;
+		u32 start_mmioaddr;
+		u32 dmc_fw_size; /*dwords */
+		u32 *payload;
+		bool present;
+	} dmc_info[DMC_FW_MAX];
+
+	u32 dc_state;
+	u32 target_dc_state;
+	u32 allowed_dc_mask;
+	intel_wakeref_t wakeref;
+};
+
+void intel_dmc_ucode_init(struct drm_i915_private *i915);
+void intel_dmc_load_program(struct drm_i915_private *i915);
+void intel_dmc_disable_program(struct drm_i915_private *i915);
+void intel_dmc_ucode_fini(struct drm_i915_private *i915);
+void intel_dmc_ucode_suspend(struct drm_i915_private *i915);
+void intel_dmc_ucode_resume(struct drm_i915_private *i915);
+bool intel_dmc_has_payload(struct drm_i915_private *i915);
+void intel_dmc_debugfs_register(struct drm_i915_private *i915);
+void intel_dmc_print_error_state(struct drm_i915_error_state_buf *m,
+				 struct drm_i915_private *i915);
+
+void assert_dmc_loaded(struct drm_i915_private *i915);
+
+#endif /* __INTEL_DMC_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dmc_regs.h b/drivers/gpu/drm/i915/display/intel_dmc_regs.h
new file mode 100644
index 000000000..5e5e41644
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dmc_regs.h
@@ -0,0 +1,89 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_DMC_REGS_H__
+#define __INTEL_DMC_REGS_H__
+
+#include "i915_reg_defs.h"
+
+#define DMC_PROGRAM(addr, i)	_MMIO((addr) + (i) * 4)
+#define DMC_SSP_BASE_ADDR_GEN9	0x00002FC0
+
+#define _ADLP_PIPEDMC_REG_MMIO_BASE_A	0x5f000
+#define _TGL_PIPEDMC_REG_MMIO_BASE_A	0x92000
+
+#define __PIPEDMC_REG_MMIO_BASE(i915, dmc_id) \
+	((DISPLAY_VER(i915) >= 13 ? _ADLP_PIPEDMC_REG_MMIO_BASE_A : \
+				    _TGL_PIPEDMC_REG_MMIO_BASE_A) + \
+	 0x400 * ((dmc_id) - 1))
+
+#define __DMC_REG_MMIO_BASE		0x8f000
+
+#define _DMC_REG_MMIO_BASE(i915, dmc_id) \
+	((dmc_id) == DMC_FW_MAIN ? __DMC_REG_MMIO_BASE : \
+				   __PIPEDMC_REG_MMIO_BASE(i915, dmc_id))
+
+#define _DMC_REG(i915, dmc_id, reg) \
+	((reg) - __DMC_REG_MMIO_BASE + _DMC_REG_MMIO_BASE(i915, dmc_id))
+
+#define DMC_EVENT_HANDLER_COUNT_GEN12	8
+
+#define _DMC_EVT_HTP_0			0x8f004
+
+#define DMC_EVT_HTP(i915, dmc_id, handler) \
+	_MMIO(_DMC_REG(i915, dmc_id, _DMC_EVT_HTP_0) + 4 * (handler))
+
+#define _DMC_EVT_CTL_0			0x8f034
+
+#define DMC_EVT_CTL(i915, dmc_id, handler) \
+	_MMIO(_DMC_REG(i915, dmc_id, _DMC_EVT_CTL_0) + 4 * (handler))
+
+#define DMC_EVT_CTL_ENABLE		REG_BIT(31)
+#define DMC_EVT_CTL_RECURRING		REG_BIT(30)
+#define DMC_EVT_CTL_TYPE_MASK		REG_GENMASK(17, 16)
+#define DMC_EVT_CTL_TYPE_LEVEL_0	0
+#define DMC_EVT_CTL_TYPE_LEVEL_1	1
+#define DMC_EVT_CTL_TYPE_EDGE_1_0	2
+#define DMC_EVT_CTL_TYPE_EDGE_0_1	3
+
+#define DMC_EVT_CTL_EVENT_ID_MASK	REG_GENMASK(15, 8)
+#define DMC_EVT_CTL_EVENT_ID_FALSE	0x01
+/* An event handler scheduled to run at a 1 kHz frequency. */
+#define DMC_EVT_CTL_EVENT_ID_CLK_MSEC	0xbf
+
+#define DMC_HTP_ADDR_SKL	0x00500034
+#define DMC_SSP_BASE		_MMIO(0x8F074)
+#define DMC_HTP_SKL		_MMIO(0x8F004)
+#define DMC_LAST_WRITE		_MMIO(0x8F034)
+#define DMC_LAST_WRITE_VALUE	0xc003b400
+#define DMC_MMIO_START_RANGE	0x80000
+#define DMC_MMIO_END_RANGE     0x8FFFF
+#define DMC_V1_MMIO_START_RANGE		0x80000
+#define TGL_MAIN_MMIO_START		0x8F000
+#define TGL_MAIN_MMIO_END		0x8FFFF
+#define _TGL_PIPEA_MMIO_START		0x92000
+#define _TGL_PIPEA_MMIO_END		0x93FFF
+#define _TGL_PIPEB_MMIO_START		0x96000
+#define _TGL_PIPEB_MMIO_END		0x97FFF
+#define ADLP_PIPE_MMIO_START		0x5F000
+#define ADLP_PIPE_MMIO_END		0x5FFFF
+
+#define TGL_PIPE_MMIO_START(dmc_id)	_PICK_EVEN(((dmc_id) - 1), _TGL_PIPEA_MMIO_START,\
+					      _TGL_PIPEB_MMIO_START)
+
+#define TGL_PIPE_MMIO_END(dmc_id)	_PICK_EVEN(((dmc_id) - 1), _TGL_PIPEA_MMIO_END,\
+					      _TGL_PIPEB_MMIO_END)
+
+#define SKL_DMC_DC3_DC5_COUNT	_MMIO(0x80030)
+#define SKL_DMC_DC5_DC6_COUNT	_MMIO(0x8002C)
+#define BXT_DMC_DC3_DC5_COUNT	_MMIO(0x80038)
+#define TGL_DMC_DEBUG_DC5_COUNT	_MMIO(0x101084)
+#define TGL_DMC_DEBUG_DC6_COUNT	_MMIO(0x101088)
+#define DG1_DMC_DEBUG_DC5_COUNT	_MMIO(0x134154)
+
+#define TGL_DMC_DEBUG3		_MMIO(0x101090)
+#define DG1_DMC_DEBUG3		_MMIO(0x13415c)
+
+#endif /* __INTEL_DMC_REGS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
new file mode 100644
index 000000000..4699c2110
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp.c
@@ -0,0 +1,5475 @@
+/*
+ * Copyright © 2008 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Keith Packard <keithp@keithp.com>
+ *
+ */
+
+#include <linux/export.h>
+#include <linux/i2c.h>
+#include <linux/notifier.h>
+#include <linux/slab.h>
+#include <linux/string_helpers.h>
+#include <linux/timekeeping.h>
+#include <linux/types.h>
+
+#include <asm/byteorder.h>
+
+#include <drm/display/drm_dp_helper.h>
+#include <drm/display/drm_dsc_helper.h>
+#include <drm/display/drm_hdmi_helper.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_probe_helper.h>
+
+#include "g4x_dp.h"
+#include "i915_debugfs.h"
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_audio.h"
+#include "intel_backlight.h"
+#include "intel_combo_phy_regs.h"
+#include "intel_connector.h"
+#include "intel_crtc.h"
+#include "intel_ddi.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_dp_aux.h"
+#include "intel_dp_hdcp.h"
+#include "intel_dp_link_training.h"
+#include "intel_dp_mst.h"
+#include "intel_dpio_phy.h"
+#include "intel_dpll.h"
+#include "intel_fifo_underrun.h"
+#include "intel_hdcp.h"
+#include "intel_hdmi.h"
+#include "intel_hotplug.h"
+#include "intel_lspcon.h"
+#include "intel_lvds.h"
+#include "intel_panel.h"
+#include "intel_pch_display.h"
+#include "intel_pps.h"
+#include "intel_psr.h"
+#include "intel_tc.h"
+#include "intel_vdsc.h"
+#include "intel_vrr.h"
+
+/* DP DSC throughput values used for slice count calculations KPixels/s */
+#define DP_DSC_PEAK_PIXEL_RATE			2720000
+#define DP_DSC_MAX_ENC_THROUGHPUT_0		340000
+#define DP_DSC_MAX_ENC_THROUGHPUT_1		400000
+
+/* DP DSC FEC Overhead factor = 1/(0.972261) */
+#define DP_DSC_FEC_OVERHEAD_FACTOR		972261
+
+/* Compliance test status bits  */
+#define INTEL_DP_RESOLUTION_SHIFT_MASK	0
+#define INTEL_DP_RESOLUTION_PREFERRED	(1 << INTEL_DP_RESOLUTION_SHIFT_MASK)
+#define INTEL_DP_RESOLUTION_STANDARD	(2 << INTEL_DP_RESOLUTION_SHIFT_MASK)
+#define INTEL_DP_RESOLUTION_FAILSAFE	(3 << INTEL_DP_RESOLUTION_SHIFT_MASK)
+
+
+/* Constants for DP DSC configurations */
+static const u8 valid_dsc_bpp[] = {6, 8, 10, 12, 15};
+
+/* With Single pipe configuration, HW is capable of supporting maximum
+ * of 4 slices per line.
+ */
+static const u8 valid_dsc_slicecount[] = {1, 2, 4};
+
+/**
+ * intel_dp_is_edp - is the given port attached to an eDP panel (either CPU or PCH)
+ * @intel_dp: DP struct
+ *
+ * If a CPU or PCH DP output is attached to an eDP panel, this function
+ * will return true, and false otherwise.
+ *
+ * This function is not safe to use prior to encoder type being set.
+ */
+bool intel_dp_is_edp(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+
+	return dig_port->base.type == INTEL_OUTPUT_EDP;
+}
+
+static void intel_dp_unset_edid(struct intel_dp *intel_dp);
+static int intel_dp_dsc_compute_bpp(struct intel_dp *intel_dp, u8 dsc_max_bpc);
+
+/* Is link rate UHBR and thus 128b/132b? */
+bool intel_dp_is_uhbr(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->port_clock >= 1000000;
+}
+
+static void intel_dp_set_default_sink_rates(struct intel_dp *intel_dp)
+{
+	intel_dp->sink_rates[0] = 162000;
+	intel_dp->num_sink_rates = 1;
+}
+
+/* update sink rates from dpcd */
+static void intel_dp_set_dpcd_sink_rates(struct intel_dp *intel_dp)
+{
+	static const int dp_rates[] = {
+		162000, 270000, 540000, 810000
+	};
+	int i, max_rate;
+	int max_lttpr_rate;
+
+	if (drm_dp_has_quirk(&intel_dp->desc, DP_DPCD_QUIRK_CAN_DO_MAX_LINK_RATE_3_24_GBPS)) {
+		/* Needed, e.g., for Apple MBP 2017, 15 inch eDP Retina panel */
+		static const int quirk_rates[] = { 162000, 270000, 324000 };
+
+		memcpy(intel_dp->sink_rates, quirk_rates, sizeof(quirk_rates));
+		intel_dp->num_sink_rates = ARRAY_SIZE(quirk_rates);
+
+		return;
+	}
+
+	/*
+	 * Sink rates for 8b/10b.
+	 */
+	max_rate = drm_dp_bw_code_to_link_rate(intel_dp->dpcd[DP_MAX_LINK_RATE]);
+	max_lttpr_rate = drm_dp_lttpr_max_link_rate(intel_dp->lttpr_common_caps);
+	if (max_lttpr_rate)
+		max_rate = min(max_rate, max_lttpr_rate);
+
+	for (i = 0; i < ARRAY_SIZE(dp_rates); i++) {
+		if (dp_rates[i] > max_rate)
+			break;
+		intel_dp->sink_rates[i] = dp_rates[i];
+	}
+
+	/*
+	 * Sink rates for 128b/132b. If set, sink should support all 8b/10b
+	 * rates and 10 Gbps.
+	 */
+	if (intel_dp->dpcd[DP_MAIN_LINK_CHANNEL_CODING] & DP_CAP_ANSI_128B132B) {
+		u8 uhbr_rates = 0;
+
+		BUILD_BUG_ON(ARRAY_SIZE(intel_dp->sink_rates) < ARRAY_SIZE(dp_rates) + 3);
+
+		drm_dp_dpcd_readb(&intel_dp->aux,
+				  DP_128B132B_SUPPORTED_LINK_RATES, &uhbr_rates);
+
+		if (drm_dp_lttpr_count(intel_dp->lttpr_common_caps)) {
+			/* We have a repeater */
+			if (intel_dp->lttpr_common_caps[0] >= 0x20 &&
+			    intel_dp->lttpr_common_caps[DP_MAIN_LINK_CHANNEL_CODING_PHY_REPEATER -
+							DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV] &
+			    DP_PHY_REPEATER_128B132B_SUPPORTED) {
+				/* Repeater supports 128b/132b, valid UHBR rates */
+				uhbr_rates &= intel_dp->lttpr_common_caps[DP_PHY_REPEATER_128B132B_RATES -
+									  DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
+			} else {
+				/* Does not support 128b/132b */
+				uhbr_rates = 0;
+			}
+		}
+
+		if (uhbr_rates & DP_UHBR10)
+			intel_dp->sink_rates[i++] = 1000000;
+		if (uhbr_rates & DP_UHBR13_5)
+			intel_dp->sink_rates[i++] = 1350000;
+		if (uhbr_rates & DP_UHBR20)
+			intel_dp->sink_rates[i++] = 2000000;
+	}
+
+	intel_dp->num_sink_rates = i;
+}
+
+static void intel_dp_set_sink_rates(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &intel_dig_port->base;
+
+	intel_dp_set_dpcd_sink_rates(intel_dp);
+
+	if (intel_dp->num_sink_rates)
+		return;
+
+	drm_err(&dp_to_i915(intel_dp)->drm,
+		"[CONNECTOR:%d:%s][ENCODER:%d:%s] Invalid DPCD with no link rates, using defaults\n",
+		connector->base.base.id, connector->base.name,
+		encoder->base.base.id, encoder->base.name);
+
+	intel_dp_set_default_sink_rates(intel_dp);
+}
+
+static void intel_dp_set_default_max_sink_lane_count(struct intel_dp *intel_dp)
+{
+	intel_dp->max_sink_lane_count = 1;
+}
+
+static void intel_dp_set_max_sink_lane_count(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &intel_dig_port->base;
+
+	intel_dp->max_sink_lane_count = drm_dp_max_lane_count(intel_dp->dpcd);
+
+	switch (intel_dp->max_sink_lane_count) {
+	case 1:
+	case 2:
+	case 4:
+		return;
+	}
+
+	drm_err(&dp_to_i915(intel_dp)->drm,
+		"[CONNECTOR:%d:%s][ENCODER:%d:%s] Invalid DPCD max lane count (%d), using default\n",
+		connector->base.base.id, connector->base.name,
+		encoder->base.base.id, encoder->base.name,
+		intel_dp->max_sink_lane_count);
+
+	intel_dp_set_default_max_sink_lane_count(intel_dp);
+}
+
+/* Get length of rates array potentially limited by max_rate. */
+static int intel_dp_rate_limit_len(const int *rates, int len, int max_rate)
+{
+	int i;
+
+	/* Limit results by potentially reduced max rate */
+	for (i = 0; i < len; i++) {
+		if (rates[len - i - 1] <= max_rate)
+			return len - i;
+	}
+
+	return 0;
+}
+
+/* Get length of common rates array potentially limited by max_rate. */
+static int intel_dp_common_len_rate_limit(const struct intel_dp *intel_dp,
+					  int max_rate)
+{
+	return intel_dp_rate_limit_len(intel_dp->common_rates,
+				       intel_dp->num_common_rates, max_rate);
+}
+
+static int intel_dp_common_rate(struct intel_dp *intel_dp, int index)
+{
+	if (drm_WARN_ON(&dp_to_i915(intel_dp)->drm,
+			index < 0 || index >= intel_dp->num_common_rates))
+		return 162000;
+
+	return intel_dp->common_rates[index];
+}
+
+/* Theoretical max between source and sink */
+static int intel_dp_max_common_rate(struct intel_dp *intel_dp)
+{
+	return intel_dp_common_rate(intel_dp, intel_dp->num_common_rates - 1);
+}
+
+static int intel_dp_max_source_lane_count(struct intel_digital_port *dig_port)
+{
+	int vbt_max_lanes = intel_bios_dp_max_lane_count(&dig_port->base);
+	int max_lanes = dig_port->max_lanes;
+
+	if (vbt_max_lanes)
+		max_lanes = min(max_lanes, vbt_max_lanes);
+
+	return max_lanes;
+}
+
+/* Theoretical max between source and sink */
+static int intel_dp_max_common_lane_count(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	int source_max = intel_dp_max_source_lane_count(dig_port);
+	int sink_max = intel_dp->max_sink_lane_count;
+	int fia_max = intel_tc_port_fia_max_lane_count(dig_port);
+	int lttpr_max = drm_dp_lttpr_max_lane_count(intel_dp->lttpr_common_caps);
+
+	if (lttpr_max)
+		sink_max = min(sink_max, lttpr_max);
+
+	return min3(source_max, sink_max, fia_max);
+}
+
+int intel_dp_max_lane_count(struct intel_dp *intel_dp)
+{
+	switch (intel_dp->max_link_lane_count) {
+	case 1:
+	case 2:
+	case 4:
+		return intel_dp->max_link_lane_count;
+	default:
+		MISSING_CASE(intel_dp->max_link_lane_count);
+		return 1;
+	}
+}
+
+/*
+ * The required data bandwidth for a mode with given pixel clock and bpp. This
+ * is the required net bandwidth independent of the data bandwidth efficiency.
+ */
+int
+intel_dp_link_required(int pixel_clock, int bpp)
+{
+	/* pixel_clock is in kHz, divide bpp by 8 for bit to Byte conversion */
+	return DIV_ROUND_UP(pixel_clock * bpp, 8);
+}
+
+/*
+ * Given a link rate and lanes, get the data bandwidth.
+ *
+ * Data bandwidth is the actual payload rate, which depends on the data
+ * bandwidth efficiency and the link rate.
+ *
+ * For 8b/10b channel encoding, SST and non-FEC, the data bandwidth efficiency
+ * is 80%. For example, for a 1.62 Gbps link, 1.62*10^9 bps * 0.80 * (1/8) =
+ * 162000 kBps. With 8-bit symbols, we have 162000 kHz symbol clock. Just by
+ * coincidence, the port clock in kHz matches the data bandwidth in kBps, and
+ * they equal the link bit rate in Gbps multiplied by 100000. (Note that this no
+ * longer holds for data bandwidth as soon as FEC or MST is taken into account!)
+ *
+ * For 128b/132b channel encoding, the data bandwidth efficiency is 96.71%. For
+ * example, for a 10 Gbps link, 10*10^9 bps * 0.9671 * (1/8) = 1208875
+ * kBps. With 32-bit symbols, we have 312500 kHz symbol clock. The value 1000000
+ * does not match the symbol clock, the port clock (not even if you think in
+ * terms of a byte clock), nor the data bandwidth. It only matches the link bit
+ * rate in units of 10000 bps.
+ */
+int
+intel_dp_max_data_rate(int max_link_rate, int max_lanes)
+{
+	if (max_link_rate >= 1000000) {
+		/*
+		 * UHBR rates always use 128b/132b channel encoding, and have
+		 * 97.71% data bandwidth efficiency. Consider max_link_rate the
+		 * link bit rate in units of 10000 bps.
+		 */
+		int max_link_rate_kbps = max_link_rate * 10;
+
+		max_link_rate_kbps = DIV_ROUND_CLOSEST_ULL(mul_u32_u32(max_link_rate_kbps, 9671), 10000);
+		max_link_rate = max_link_rate_kbps / 8;
+	}
+
+	/*
+	 * Lower than UHBR rates always use 8b/10b channel encoding, and have
+	 * 80% data bandwidth efficiency for SST non-FEC. However, this turns
+	 * out to be a nop by coincidence, and can be skipped:
+	 *
+	 *	int max_link_rate_kbps = max_link_rate * 10;
+	 *	max_link_rate_kbps = DIV_ROUND_CLOSEST_ULL(max_link_rate_kbps * 8, 10);
+	 *	max_link_rate = max_link_rate_kbps / 8;
+	 */
+
+	return max_link_rate * max_lanes;
+}
+
+bool intel_dp_can_bigjoiner(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &intel_dig_port->base;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	return DISPLAY_VER(dev_priv) >= 12 ||
+		(DISPLAY_VER(dev_priv) == 11 &&
+		 encoder->port != PORT_A);
+}
+
+static int dg2_max_source_rate(struct intel_dp *intel_dp)
+{
+	return intel_dp_is_edp(intel_dp) ? 810000 : 1350000;
+}
+
+static int icl_max_source_rate(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
+	enum phy phy = intel_port_to_phy(dev_priv, dig_port->base.port);
+
+	if (intel_phy_is_combo(dev_priv, phy) && !intel_dp_is_edp(intel_dp))
+		return 540000;
+
+	return 810000;
+}
+
+static int ehl_max_source_rate(struct intel_dp *intel_dp)
+{
+	if (intel_dp_is_edp(intel_dp))
+		return 540000;
+
+	return 810000;
+}
+
+static int vbt_max_link_rate(struct intel_dp *intel_dp)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	int max_rate;
+
+	max_rate = intel_bios_dp_max_link_rate(encoder);
+
+	if (intel_dp_is_edp(intel_dp)) {
+		struct intel_connector *connector = intel_dp->attached_connector;
+		int edp_max_rate = connector->panel.vbt.edp.max_link_rate;
+
+		if (max_rate && edp_max_rate)
+			max_rate = min(max_rate, edp_max_rate);
+		else if (edp_max_rate)
+			max_rate = edp_max_rate;
+	}
+
+	return max_rate;
+}
+
+static void
+intel_dp_set_source_rates(struct intel_dp *intel_dp)
+{
+	/* The values must be in increasing order */
+	static const int icl_rates[] = {
+		162000, 216000, 270000, 324000, 432000, 540000, 648000, 810000,
+		1000000, 1350000,
+	};
+	static const int bxt_rates[] = {
+		162000, 216000, 243000, 270000, 324000, 432000, 540000
+	};
+	static const int skl_rates[] = {
+		162000, 216000, 270000, 324000, 432000, 540000
+	};
+	static const int hsw_rates[] = {
+		162000, 270000, 540000
+	};
+	static const int g4x_rates[] = {
+		162000, 270000
+	};
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
+	const int *source_rates;
+	int size, max_rate = 0, vbt_max_rate;
+
+	/* This should only be done once */
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_dp->source_rates || intel_dp->num_source_rates);
+
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		source_rates = icl_rates;
+		size = ARRAY_SIZE(icl_rates);
+		if (IS_DG2(dev_priv))
+			max_rate = dg2_max_source_rate(intel_dp);
+		else if (IS_ALDERLAKE_P(dev_priv) || IS_ALDERLAKE_S(dev_priv) ||
+			 IS_DG1(dev_priv) || IS_ROCKETLAKE(dev_priv))
+			max_rate = 810000;
+		else if (IS_JSL_EHL(dev_priv))
+			max_rate = ehl_max_source_rate(intel_dp);
+		else
+			max_rate = icl_max_source_rate(intel_dp);
+	} else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		source_rates = bxt_rates;
+		size = ARRAY_SIZE(bxt_rates);
+	} else if (DISPLAY_VER(dev_priv) == 9) {
+		source_rates = skl_rates;
+		size = ARRAY_SIZE(skl_rates);
+	} else if ((IS_HASWELL(dev_priv) && !IS_HSW_ULX(dev_priv)) ||
+		   IS_BROADWELL(dev_priv)) {
+		source_rates = hsw_rates;
+		size = ARRAY_SIZE(hsw_rates);
+	} else {
+		source_rates = g4x_rates;
+		size = ARRAY_SIZE(g4x_rates);
+	}
+
+	vbt_max_rate = vbt_max_link_rate(intel_dp);
+	if (max_rate && vbt_max_rate)
+		max_rate = min(max_rate, vbt_max_rate);
+	else if (vbt_max_rate)
+		max_rate = vbt_max_rate;
+
+	if (max_rate)
+		size = intel_dp_rate_limit_len(source_rates, size, max_rate);
+
+	intel_dp->source_rates = source_rates;
+	intel_dp->num_source_rates = size;
+}
+
+static int intersect_rates(const int *source_rates, int source_len,
+			   const int *sink_rates, int sink_len,
+			   int *common_rates)
+{
+	int i = 0, j = 0, k = 0;
+
+	while (i < source_len && j < sink_len) {
+		if (source_rates[i] == sink_rates[j]) {
+			if (WARN_ON(k >= DP_MAX_SUPPORTED_RATES))
+				return k;
+			common_rates[k] = source_rates[i];
+			++k;
+			++i;
+			++j;
+		} else if (source_rates[i] < sink_rates[j]) {
+			++i;
+		} else {
+			++j;
+		}
+	}
+	return k;
+}
+
+/* return index of rate in rates array, or -1 if not found */
+static int intel_dp_rate_index(const int *rates, int len, int rate)
+{
+	int i;
+
+	for (i = 0; i < len; i++)
+		if (rate == rates[i])
+			return i;
+
+	return -1;
+}
+
+static void intel_dp_set_common_rates(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	drm_WARN_ON(&i915->drm,
+		    !intel_dp->num_source_rates || !intel_dp->num_sink_rates);
+
+	intel_dp->num_common_rates = intersect_rates(intel_dp->source_rates,
+						     intel_dp->num_source_rates,
+						     intel_dp->sink_rates,
+						     intel_dp->num_sink_rates,
+						     intel_dp->common_rates);
+
+	/* Paranoia, there should always be something in common. */
+	if (drm_WARN_ON(&i915->drm, intel_dp->num_common_rates == 0)) {
+		intel_dp->common_rates[0] = 162000;
+		intel_dp->num_common_rates = 1;
+	}
+}
+
+static bool intel_dp_link_params_valid(struct intel_dp *intel_dp, int link_rate,
+				       u8 lane_count)
+{
+	/*
+	 * FIXME: we need to synchronize the current link parameters with
+	 * hardware readout. Currently fast link training doesn't work on
+	 * boot-up.
+	 */
+	if (link_rate == 0 ||
+	    link_rate > intel_dp->max_link_rate)
+		return false;
+
+	if (lane_count == 0 ||
+	    lane_count > intel_dp_max_lane_count(intel_dp))
+		return false;
+
+	return true;
+}
+
+static bool intel_dp_can_link_train_fallback_for_edp(struct intel_dp *intel_dp,
+						     int link_rate,
+						     u8 lane_count)
+{
+	/* FIXME figure out what we actually want here */
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_preferred_fixed_mode(intel_dp->attached_connector);
+	int mode_rate, max_rate;
+
+	mode_rate = intel_dp_link_required(fixed_mode->clock, 18);
+	max_rate = intel_dp_max_data_rate(link_rate, lane_count);
+	if (mode_rate > max_rate)
+		return false;
+
+	return true;
+}
+
+int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,
+					    int link_rate, u8 lane_count)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int index;
+
+	/*
+	 * TODO: Enable fallback on MST links once MST link compute can handle
+	 * the fallback params.
+	 */
+	if (intel_dp->is_mst) {
+		drm_err(&i915->drm, "Link Training Unsuccessful\n");
+		return -1;
+	}
+
+	if (intel_dp_is_edp(intel_dp) && !intel_dp->use_max_params) {
+		drm_dbg_kms(&i915->drm,
+			    "Retrying Link training for eDP with max parameters\n");
+		intel_dp->use_max_params = true;
+		return 0;
+	}
+
+	index = intel_dp_rate_index(intel_dp->common_rates,
+				    intel_dp->num_common_rates,
+				    link_rate);
+	if (index > 0) {
+		if (intel_dp_is_edp(intel_dp) &&
+		    !intel_dp_can_link_train_fallback_for_edp(intel_dp,
+							      intel_dp_common_rate(intel_dp, index - 1),
+							      lane_count)) {
+			drm_dbg_kms(&i915->drm,
+				    "Retrying Link training for eDP with same parameters\n");
+			return 0;
+		}
+		intel_dp->max_link_rate = intel_dp_common_rate(intel_dp, index - 1);
+		intel_dp->max_link_lane_count = lane_count;
+	} else if (lane_count > 1) {
+		if (intel_dp_is_edp(intel_dp) &&
+		    !intel_dp_can_link_train_fallback_for_edp(intel_dp,
+							      intel_dp_max_common_rate(intel_dp),
+							      lane_count >> 1)) {
+			drm_dbg_kms(&i915->drm,
+				    "Retrying Link training for eDP with same parameters\n");
+			return 0;
+		}
+		intel_dp->max_link_rate = intel_dp_max_common_rate(intel_dp);
+		intel_dp->max_link_lane_count = lane_count >> 1;
+	} else {
+		drm_err(&i915->drm, "Link Training Unsuccessful\n");
+		return -1;
+	}
+
+	return 0;
+}
+
+u32 intel_dp_mode_to_fec_clock(u32 mode_clock)
+{
+	return div_u64(mul_u32_u32(mode_clock, 1000000U),
+		       DP_DSC_FEC_OVERHEAD_FACTOR);
+}
+
+static int
+small_joiner_ram_size_bits(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 13)
+		return 17280 * 8;
+	else if (DISPLAY_VER(i915) >= 11)
+		return 7680 * 8;
+	else
+		return 6144 * 8;
+}
+
+static u16 intel_dp_dsc_get_output_bpp(struct drm_i915_private *i915,
+				       u32 link_clock, u32 lane_count,
+				       u32 mode_clock, u32 mode_hdisplay,
+				       bool bigjoiner,
+				       u32 pipe_bpp)
+{
+	u32 bits_per_pixel, max_bpp_small_joiner_ram;
+	int i;
+
+	/*
+	 * Available Link Bandwidth(Kbits/sec) = (NumberOfLanes)*
+	 * (LinkSymbolClock)* 8 * (TimeSlotsPerMTP)
+	 * for SST -> TimeSlotsPerMTP is 1,
+	 * for MST -> TimeSlotsPerMTP has to be calculated
+	 */
+	bits_per_pixel = (link_clock * lane_count * 8) /
+			 intel_dp_mode_to_fec_clock(mode_clock);
+
+	/* Small Joiner Check: output bpp <= joiner RAM (bits) / Horiz. width */
+	max_bpp_small_joiner_ram = small_joiner_ram_size_bits(i915) /
+		mode_hdisplay;
+
+	if (bigjoiner)
+		max_bpp_small_joiner_ram *= 2;
+
+	/*
+	 * Greatest allowed DSC BPP = MIN (output BPP from available Link BW
+	 * check, output bpp from small joiner RAM check)
+	 */
+	bits_per_pixel = min(bits_per_pixel, max_bpp_small_joiner_ram);
+
+	if (bigjoiner) {
+		u32 max_bpp_bigjoiner =
+			i915->display.cdclk.max_cdclk_freq * 48 /
+			intel_dp_mode_to_fec_clock(mode_clock);
+
+		bits_per_pixel = min(bits_per_pixel, max_bpp_bigjoiner);
+	}
+
+	/* Error out if the max bpp is less than smallest allowed valid bpp */
+	if (bits_per_pixel < valid_dsc_bpp[0]) {
+		drm_dbg_kms(&i915->drm, "Unsupported BPP %u, min %u\n",
+			    bits_per_pixel, valid_dsc_bpp[0]);
+		return 0;
+	}
+
+	/* From XE_LPD onwards we support from bpc upto uncompressed bpp-1 BPPs */
+	if (DISPLAY_VER(i915) >= 13) {
+		bits_per_pixel = min(bits_per_pixel, pipe_bpp - 1);
+	} else {
+		/* Find the nearest match in the array of known BPPs from VESA */
+		for (i = 0; i < ARRAY_SIZE(valid_dsc_bpp) - 1; i++) {
+			if (bits_per_pixel < valid_dsc_bpp[i + 1])
+				break;
+		}
+		bits_per_pixel = valid_dsc_bpp[i];
+	}
+
+	/*
+	 * Compressed BPP in U6.4 format so multiply by 16, for Gen 11,
+	 * fractional part is 0
+	 */
+	return bits_per_pixel << 4;
+}
+
+static u8 intel_dp_dsc_get_slice_count(struct intel_dp *intel_dp,
+				       int mode_clock, int mode_hdisplay,
+				       bool bigjoiner)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 min_slice_count, i;
+	int max_slice_width;
+
+	if (mode_clock <= DP_DSC_PEAK_PIXEL_RATE)
+		min_slice_count = DIV_ROUND_UP(mode_clock,
+					       DP_DSC_MAX_ENC_THROUGHPUT_0);
+	else
+		min_slice_count = DIV_ROUND_UP(mode_clock,
+					       DP_DSC_MAX_ENC_THROUGHPUT_1);
+
+	max_slice_width = drm_dp_dsc_sink_max_slice_width(intel_dp->dsc_dpcd);
+	if (max_slice_width < DP_DSC_MIN_SLICE_WIDTH_VALUE) {
+		drm_dbg_kms(&i915->drm,
+			    "Unsupported slice width %d by DP DSC Sink device\n",
+			    max_slice_width);
+		return 0;
+	}
+	/* Also take into account max slice width */
+	min_slice_count = max_t(u8, min_slice_count,
+				DIV_ROUND_UP(mode_hdisplay,
+					     max_slice_width));
+
+	/* Find the closest match to the valid slice count values */
+	for (i = 0; i < ARRAY_SIZE(valid_dsc_slicecount); i++) {
+		u8 test_slice_count = valid_dsc_slicecount[i] << bigjoiner;
+
+		if (test_slice_count >
+		    drm_dp_dsc_sink_max_slice_count(intel_dp->dsc_dpcd, false))
+			break;
+
+		/* big joiner needs small joiner to be enabled */
+		if (bigjoiner && test_slice_count < 4)
+			continue;
+
+		if (min_slice_count <= test_slice_count)
+			return test_slice_count;
+	}
+
+	drm_dbg_kms(&i915->drm, "Unsupported Slice Count %d\n",
+		    min_slice_count);
+	return 0;
+}
+
+static enum intel_output_format
+intel_dp_output_format(struct intel_connector *connector,
+		       bool ycbcr_420_output)
+{
+	struct intel_dp *intel_dp = intel_attached_dp(connector);
+
+	if (!connector->base.ycbcr_420_allowed || !ycbcr_420_output)
+		return INTEL_OUTPUT_FORMAT_RGB;
+
+	if (intel_dp->dfp.rgb_to_ycbcr &&
+	    intel_dp->dfp.ycbcr_444_to_420)
+		return INTEL_OUTPUT_FORMAT_RGB;
+
+	if (intel_dp->dfp.ycbcr_444_to_420)
+		return INTEL_OUTPUT_FORMAT_YCBCR444;
+	else
+		return INTEL_OUTPUT_FORMAT_YCBCR420;
+}
+
+int intel_dp_min_bpp(enum intel_output_format output_format)
+{
+	if (output_format == INTEL_OUTPUT_FORMAT_RGB)
+		return 6 * 3;
+	else
+		return 8 * 3;
+}
+
+static int intel_dp_output_bpp(enum intel_output_format output_format, int bpp)
+{
+	/*
+	 * bpp value was assumed to RGB format. And YCbCr 4:2:0 output
+	 * format of the number of bytes per pixel will be half the number
+	 * of bytes of RGB pixel.
+	 */
+	if (output_format == INTEL_OUTPUT_FORMAT_YCBCR420)
+		bpp /= 2;
+
+	return bpp;
+}
+
+static int
+intel_dp_mode_min_output_bpp(struct intel_connector *connector,
+			     const struct drm_display_mode *mode)
+{
+	const struct drm_display_info *info = &connector->base.display_info;
+	enum intel_output_format output_format =
+		intel_dp_output_format(connector, drm_mode_is_420_only(info, mode));
+
+	return intel_dp_output_bpp(output_format, intel_dp_min_bpp(output_format));
+}
+
+static bool intel_dp_hdisplay_bad(struct drm_i915_private *dev_priv,
+				  int hdisplay)
+{
+	/*
+	 * Older platforms don't like hdisplay==4096 with DP.
+	 *
+	 * On ILK/SNB/IVB the pipe seems to be somewhat running (scanline
+	 * and frame counter increment), but we don't get vblank interrupts,
+	 * and the pipe underruns immediately. The link also doesn't seem
+	 * to get trained properly.
+	 *
+	 * On CHV the vblank interrupts don't seem to disappear but
+	 * otherwise the symptoms are similar.
+	 *
+	 * TODO: confirm the behaviour on HSW+
+	 */
+	return hdisplay == 4096 && !HAS_DDI(dev_priv);
+}
+
+static int intel_dp_max_tmds_clock(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+	const struct drm_display_info *info = &connector->base.display_info;
+	int max_tmds_clock = intel_dp->dfp.max_tmds_clock;
+
+	/* Only consider the sink's max TMDS clock if we know this is a HDMI DFP */
+	if (max_tmds_clock && info->max_tmds_clock)
+		max_tmds_clock = min(max_tmds_clock, info->max_tmds_clock);
+
+	return max_tmds_clock;
+}
+
+static enum drm_mode_status
+intel_dp_tmds_clock_valid(struct intel_dp *intel_dp,
+			  int clock, int bpc, bool ycbcr420_output,
+			  bool respect_downstream_limits)
+{
+	int tmds_clock, min_tmds_clock, max_tmds_clock;
+
+	if (!respect_downstream_limits)
+		return MODE_OK;
+
+	tmds_clock = intel_hdmi_tmds_clock(clock, bpc, ycbcr420_output);
+
+	min_tmds_clock = intel_dp->dfp.min_tmds_clock;
+	max_tmds_clock = intel_dp_max_tmds_clock(intel_dp);
+
+	if (min_tmds_clock && tmds_clock < min_tmds_clock)
+		return MODE_CLOCK_LOW;
+
+	if (max_tmds_clock && tmds_clock > max_tmds_clock)
+		return MODE_CLOCK_HIGH;
+
+	return MODE_OK;
+}
+
+static enum drm_mode_status
+intel_dp_mode_valid_downstream(struct intel_connector *connector,
+			       const struct drm_display_mode *mode,
+			       int target_clock)
+{
+	struct intel_dp *intel_dp = intel_attached_dp(connector);
+	const struct drm_display_info *info = &connector->base.display_info;
+	enum drm_mode_status status;
+	bool ycbcr_420_only;
+
+	/* If PCON supports FRL MODE, check FRL bandwidth constraints */
+	if (intel_dp->dfp.pcon_max_frl_bw) {
+		int target_bw;
+		int max_frl_bw;
+		int bpp = intel_dp_mode_min_output_bpp(connector, mode);
+
+		target_bw = bpp * target_clock;
+
+		max_frl_bw = intel_dp->dfp.pcon_max_frl_bw;
+
+		/* converting bw from Gbps to Kbps*/
+		max_frl_bw = max_frl_bw * 1000000;
+
+		if (target_bw > max_frl_bw)
+			return MODE_CLOCK_HIGH;
+
+		return MODE_OK;
+	}
+
+	if (intel_dp->dfp.max_dotclock &&
+	    target_clock > intel_dp->dfp.max_dotclock)
+		return MODE_CLOCK_HIGH;
+
+	ycbcr_420_only = drm_mode_is_420_only(info, mode);
+
+	/* Assume 8bpc for the DP++/HDMI/DVI TMDS clock check */
+	status = intel_dp_tmds_clock_valid(intel_dp, target_clock,
+					   8, ycbcr_420_only, true);
+
+	if (status != MODE_OK) {
+		if (ycbcr_420_only ||
+		    !connector->base.ycbcr_420_allowed ||
+		    !drm_mode_is_420_also(info, mode))
+			return status;
+
+		status = intel_dp_tmds_clock_valid(intel_dp, target_clock,
+						   8, true, true);
+		if (status != MODE_OK)
+			return status;
+	}
+
+	return MODE_OK;
+}
+
+static bool intel_dp_need_bigjoiner(struct intel_dp *intel_dp,
+				    int hdisplay, int clock)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	if (!intel_dp_can_bigjoiner(intel_dp))
+		return false;
+
+	return clock > i915->max_dotclk_freq || hdisplay > 5120;
+}
+
+static enum drm_mode_status
+intel_dp_mode_valid(struct drm_connector *_connector,
+		    struct drm_display_mode *mode)
+{
+	struct intel_connector *connector = to_intel_connector(_connector);
+	struct intel_dp *intel_dp = intel_attached_dp(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	const struct drm_display_mode *fixed_mode;
+	int target_clock = mode->clock;
+	int max_rate, mode_rate, max_lanes, max_link_clock;
+	int max_dotclk = dev_priv->max_dotclk_freq;
+	u16 dsc_max_output_bpp = 0;
+	u8 dsc_slice_count = 0;
+	enum drm_mode_status status;
+	bool dsc = false, bigjoiner = false;
+
+	status = intel_cpu_transcoder_mode_valid(dev_priv, mode);
+	if (status != MODE_OK)
+		return status;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLCLK)
+		return MODE_H_ILLEGAL;
+
+	fixed_mode = intel_panel_fixed_mode(connector, mode);
+	if (intel_dp_is_edp(intel_dp) && fixed_mode) {
+		status = intel_panel_mode_valid(connector, mode);
+		if (status != MODE_OK)
+			return status;
+
+		target_clock = fixed_mode->clock;
+	}
+
+	if (mode->clock < 10000)
+		return MODE_CLOCK_LOW;
+
+	if (intel_dp_need_bigjoiner(intel_dp, mode->hdisplay, target_clock)) {
+		bigjoiner = true;
+		max_dotclk *= 2;
+	}
+	if (target_clock > max_dotclk)
+		return MODE_CLOCK_HIGH;
+
+	max_link_clock = intel_dp_max_link_rate(intel_dp);
+	max_lanes = intel_dp_max_lane_count(intel_dp);
+
+	max_rate = intel_dp_max_data_rate(max_link_clock, max_lanes);
+	mode_rate = intel_dp_link_required(target_clock,
+					   intel_dp_mode_min_output_bpp(connector, mode));
+
+	if (intel_dp_hdisplay_bad(dev_priv, mode->hdisplay))
+		return MODE_H_ILLEGAL;
+
+	/*
+	 * Output bpp is stored in 6.4 format so right shift by 4 to get the
+	 * integer value since we support only integer values of bpp.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 10 &&
+	    drm_dp_sink_supports_dsc(intel_dp->dsc_dpcd)) {
+		/*
+		 * TBD pass the connector BPC,
+		 * for now U8_MAX so that max BPC on that platform would be picked
+		 */
+		int pipe_bpp = intel_dp_dsc_compute_bpp(intel_dp, U8_MAX);
+
+		if (intel_dp_is_edp(intel_dp)) {
+			dsc_max_output_bpp =
+				drm_edp_dsc_sink_output_bpp(intel_dp->dsc_dpcd) >> 4;
+			dsc_slice_count =
+				drm_dp_dsc_sink_max_slice_count(intel_dp->dsc_dpcd,
+								true);
+		} else if (drm_dp_sink_supports_fec(intel_dp->fec_capable)) {
+			dsc_max_output_bpp =
+				intel_dp_dsc_get_output_bpp(dev_priv,
+							    max_link_clock,
+							    max_lanes,
+							    target_clock,
+							    mode->hdisplay,
+							    bigjoiner,
+							    pipe_bpp) >> 4;
+			dsc_slice_count =
+				intel_dp_dsc_get_slice_count(intel_dp,
+							     target_clock,
+							     mode->hdisplay,
+							     bigjoiner);
+		}
+
+		dsc = dsc_max_output_bpp && dsc_slice_count;
+	}
+
+	/*
+	 * Big joiner configuration needs DSC for TGL which is not true for
+	 * XE_LPD where uncompressed joiner is supported.
+	 */
+	if (DISPLAY_VER(dev_priv) < 13 && bigjoiner && !dsc)
+		return MODE_CLOCK_HIGH;
+
+	if (mode_rate > max_rate && !dsc)
+		return MODE_CLOCK_HIGH;
+
+	status = intel_dp_mode_valid_downstream(connector, mode, target_clock);
+	if (status != MODE_OK)
+		return status;
+
+	return intel_mode_valid_max_plane_size(dev_priv, mode, bigjoiner);
+}
+
+bool intel_dp_source_supports_tps3(struct drm_i915_private *i915)
+{
+	return DISPLAY_VER(i915) >= 9 || IS_BROADWELL(i915) || IS_HASWELL(i915);
+}
+
+bool intel_dp_source_supports_tps4(struct drm_i915_private *i915)
+{
+	return DISPLAY_VER(i915) >= 10;
+}
+
+static void snprintf_int_array(char *str, size_t len,
+			       const int *array, int nelem)
+{
+	int i;
+
+	str[0] = '\0';
+
+	for (i = 0; i < nelem; i++) {
+		int r = snprintf(str, len, "%s%d", i ? ", " : "", array[i]);
+		if (r >= len)
+			return;
+		str += r;
+		len -= r;
+	}
+}
+
+static void intel_dp_print_rates(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	char str[128]; /* FIXME: too big for stack? */
+
+	if (!drm_debug_enabled(DRM_UT_KMS))
+		return;
+
+	snprintf_int_array(str, sizeof(str),
+			   intel_dp->source_rates, intel_dp->num_source_rates);
+	drm_dbg_kms(&i915->drm, "source rates: %s\n", str);
+
+	snprintf_int_array(str, sizeof(str),
+			   intel_dp->sink_rates, intel_dp->num_sink_rates);
+	drm_dbg_kms(&i915->drm, "sink rates: %s\n", str);
+
+	snprintf_int_array(str, sizeof(str),
+			   intel_dp->common_rates, intel_dp->num_common_rates);
+	drm_dbg_kms(&i915->drm, "common rates: %s\n", str);
+}
+
+int
+intel_dp_max_link_rate(struct intel_dp *intel_dp)
+{
+	int len;
+
+	len = intel_dp_common_len_rate_limit(intel_dp, intel_dp->max_link_rate);
+
+	return intel_dp_common_rate(intel_dp, len - 1);
+}
+
+int intel_dp_rate_select(struct intel_dp *intel_dp, int rate)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int i = intel_dp_rate_index(intel_dp->sink_rates,
+				    intel_dp->num_sink_rates, rate);
+
+	if (drm_WARN_ON(&i915->drm, i < 0))
+		i = 0;
+
+	return i;
+}
+
+void intel_dp_compute_rate(struct intel_dp *intel_dp, int port_clock,
+			   u8 *link_bw, u8 *rate_select)
+{
+	/* eDP 1.4 rate select method. */
+	if (intel_dp->use_rate_select) {
+		*link_bw = 0;
+		*rate_select =
+			intel_dp_rate_select(intel_dp, port_clock);
+	} else {
+		*link_bw = drm_dp_link_rate_to_bw_code(port_clock);
+		*rate_select = 0;
+	}
+}
+
+static bool intel_dp_source_supports_fec(struct intel_dp *intel_dp,
+					 const struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	/* On TGL, FEC is supported on all Pipes */
+	if (DISPLAY_VER(dev_priv) >= 12)
+		return true;
+
+	if (DISPLAY_VER(dev_priv) == 11 && pipe_config->cpu_transcoder != TRANSCODER_A)
+		return true;
+
+	return false;
+}
+
+static bool intel_dp_supports_fec(struct intel_dp *intel_dp,
+				  const struct intel_crtc_state *pipe_config)
+{
+	return intel_dp_source_supports_fec(intel_dp, pipe_config) &&
+		drm_dp_sink_supports_fec(intel_dp->fec_capable);
+}
+
+static bool intel_dp_supports_dsc(struct intel_dp *intel_dp,
+				  const struct intel_crtc_state *crtc_state)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP) && !crtc_state->fec_enable)
+		return false;
+
+	return intel_dsc_source_support(crtc_state) &&
+		drm_dp_sink_supports_dsc(intel_dp->dsc_dpcd);
+}
+
+static bool intel_dp_is_ycbcr420(struct intel_dp *intel_dp,
+				 const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR420 ||
+		(crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR444 &&
+		 intel_dp->dfp.ycbcr_444_to_420);
+}
+
+static int intel_dp_hdmi_compute_bpc(struct intel_dp *intel_dp,
+				     const struct intel_crtc_state *crtc_state,
+				     int bpc, bool respect_downstream_limits)
+{
+	bool ycbcr420_output = intel_dp_is_ycbcr420(intel_dp, crtc_state);
+	int clock = crtc_state->hw.adjusted_mode.crtc_clock;
+
+	/*
+	 * Current bpc could already be below 8bpc due to
+	 * FDI bandwidth constraints or other limits.
+	 * HDMI minimum is 8bpc however.
+	 */
+	bpc = max(bpc, 8);
+
+	/*
+	 * We will never exceed downstream TMDS clock limits while
+	 * attempting deep color. If the user insists on forcing an
+	 * out of spec mode they will have to be satisfied with 8bpc.
+	 */
+	if (!respect_downstream_limits)
+		bpc = 8;
+
+	for (; bpc >= 8; bpc -= 2) {
+		if (intel_hdmi_bpc_possible(crtc_state, bpc,
+					    intel_dp->has_hdmi_sink, ycbcr420_output) &&
+		    intel_dp_tmds_clock_valid(intel_dp, clock, bpc, ycbcr420_output,
+					      respect_downstream_limits) == MODE_OK)
+			return bpc;
+	}
+
+	return -EINVAL;
+}
+
+static int intel_dp_max_bpp(struct intel_dp *intel_dp,
+			    const struct intel_crtc_state *crtc_state,
+			    bool respect_downstream_limits)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
+	int bpp, bpc;
+
+	bpc = crtc_state->pipe_bpp / 3;
+
+	if (intel_dp->dfp.max_bpc)
+		bpc = min_t(int, bpc, intel_dp->dfp.max_bpc);
+
+	if (intel_dp->dfp.min_tmds_clock) {
+		int max_hdmi_bpc;
+
+		max_hdmi_bpc = intel_dp_hdmi_compute_bpc(intel_dp, crtc_state, bpc,
+							 respect_downstream_limits);
+		if (max_hdmi_bpc < 0)
+			return 0;
+
+		bpc = min(bpc, max_hdmi_bpc);
+	}
+
+	bpp = bpc * 3;
+	if (intel_dp_is_edp(intel_dp)) {
+		/* Get bpp from vbt only for panels that dont have bpp in edid */
+		if (intel_connector->base.display_info.bpc == 0 &&
+		    intel_connector->panel.vbt.edp.bpp &&
+		    intel_connector->panel.vbt.edp.bpp < bpp) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "clamping bpp for eDP panel to BIOS-provided %i\n",
+				    intel_connector->panel.vbt.edp.bpp);
+			bpp = intel_connector->panel.vbt.edp.bpp;
+		}
+	}
+
+	return bpp;
+}
+
+/* Adjust link config limits based on compliance test requests. */
+void
+intel_dp_adjust_compliance_config(struct intel_dp *intel_dp,
+				  struct intel_crtc_state *pipe_config,
+				  struct link_config_limits *limits)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	/* For DP Compliance we override the computed bpp for the pipe */
+	if (intel_dp->compliance.test_data.bpc != 0) {
+		int bpp = 3 * intel_dp->compliance.test_data.bpc;
+
+		limits->min_bpp = limits->max_bpp = bpp;
+		pipe_config->dither_force_disable = bpp == 6 * 3;
+
+		drm_dbg_kms(&i915->drm, "Setting pipe_bpp to %d\n", bpp);
+	}
+
+	/* Use values requested by Compliance Test Request */
+	if (intel_dp->compliance.test_type == DP_TEST_LINK_TRAINING) {
+		int index;
+
+		/* Validate the compliance test data since max values
+		 * might have changed due to link train fallback.
+		 */
+		if (intel_dp_link_params_valid(intel_dp, intel_dp->compliance.test_link_rate,
+					       intel_dp->compliance.test_lane_count)) {
+			index = intel_dp_rate_index(intel_dp->common_rates,
+						    intel_dp->num_common_rates,
+						    intel_dp->compliance.test_link_rate);
+			if (index >= 0)
+				limits->min_rate = limits->max_rate =
+					intel_dp->compliance.test_link_rate;
+			limits->min_lane_count = limits->max_lane_count =
+				intel_dp->compliance.test_lane_count;
+		}
+	}
+}
+
+static bool has_seamless_m_n(struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+
+	/*
+	 * Seamless M/N reprogramming only implemented
+	 * for BDW+ double buffered M/N registers so far.
+	 */
+	return HAS_DOUBLE_BUFFERED_M_N(i915) &&
+		intel_panel_drrs_type(connector) == DRRS_TYPE_SEAMLESS;
+}
+
+static int intel_dp_mode_clock(const struct intel_crtc_state *crtc_state,
+			       const struct drm_connector_state *conn_state)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+
+	/* FIXME a bit of a mess wrt clock vs. crtc_clock */
+	if (has_seamless_m_n(connector))
+		return intel_panel_highest_mode(connector, adjusted_mode)->clock;
+	else
+		return adjusted_mode->crtc_clock;
+}
+
+/* Optimize link config in order: max bpp, min clock, min lanes */
+static int
+intel_dp_compute_link_config_wide(struct intel_dp *intel_dp,
+				  struct intel_crtc_state *pipe_config,
+				  const struct drm_connector_state *conn_state,
+				  const struct link_config_limits *limits)
+{
+	int bpp, i, lane_count, clock = intel_dp_mode_clock(pipe_config, conn_state);
+	int mode_rate, link_rate, link_avail;
+
+	for (bpp = limits->max_bpp; bpp >= limits->min_bpp; bpp -= 2 * 3) {
+		int output_bpp = intel_dp_output_bpp(pipe_config->output_format, bpp);
+
+		mode_rate = intel_dp_link_required(clock, output_bpp);
+
+		for (i = 0; i < intel_dp->num_common_rates; i++) {
+			link_rate = intel_dp_common_rate(intel_dp, i);
+			if (link_rate < limits->min_rate ||
+			    link_rate > limits->max_rate)
+				continue;
+
+			for (lane_count = limits->min_lane_count;
+			     lane_count <= limits->max_lane_count;
+			     lane_count <<= 1) {
+				link_avail = intel_dp_max_data_rate(link_rate,
+								    lane_count);
+
+				if (mode_rate <= link_avail) {
+					pipe_config->lane_count = lane_count;
+					pipe_config->pipe_bpp = bpp;
+					pipe_config->port_clock = link_rate;
+
+					return 0;
+				}
+			}
+		}
+	}
+
+	return -EINVAL;
+}
+
+static int intel_dp_dsc_compute_bpp(struct intel_dp *intel_dp, u8 max_req_bpc)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int i, num_bpc;
+	u8 dsc_bpc[3] = {0};
+	u8 dsc_max_bpc;
+
+	/* Max DSC Input BPC for ICL is 10 and for TGL+ is 12 */
+	if (DISPLAY_VER(i915) >= 12)
+		dsc_max_bpc = min_t(u8, 12, max_req_bpc);
+	else
+		dsc_max_bpc = min_t(u8, 10, max_req_bpc);
+
+	num_bpc = drm_dp_dsc_sink_supported_input_bpcs(intel_dp->dsc_dpcd,
+						       dsc_bpc);
+	for (i = 0; i < num_bpc; i++) {
+		if (dsc_max_bpc >= dsc_bpc[i])
+			return dsc_bpc[i] * 3;
+	}
+
+	return 0;
+}
+
+static int intel_dp_source_dsc_version_minor(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	return DISPLAY_VER(i915) >= 14 ? 2 : 1;
+}
+
+static int intel_dp_sink_dsc_version_minor(struct intel_dp *intel_dp)
+{
+	return (intel_dp->dsc_dpcd[DP_DSC_REV - DP_DSC_SUPPORT] & DP_DSC_MINOR_MASK) >>
+		DP_DSC_MINOR_SHIFT;
+}
+
+static int intel_dp_dsc_compute_params(struct intel_encoder *encoder,
+				       struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct drm_dsc_config *vdsc_cfg = &crtc_state->dsc.config;
+	u8 line_buf_depth;
+	int ret;
+
+	/*
+	 * RC_MODEL_SIZE is currently a constant across all configurations.
+	 *
+	 * FIXME: Look into using sink defined DPCD DP_DSC_RC_BUF_BLK_SIZE and
+	 * DP_DSC_RC_BUF_SIZE for this.
+	 */
+	vdsc_cfg->rc_model_size = DSC_RC_MODEL_SIZE_CONST;
+	vdsc_cfg->pic_height = crtc_state->hw.adjusted_mode.crtc_vdisplay;
+
+	/*
+	 * Slice Height of 8 works for all currently available panels. So start
+	 * with that if pic_height is an integral multiple of 8. Eventually add
+	 * logic to try multiple slice heights.
+	 */
+	if (vdsc_cfg->pic_height % 8 == 0)
+		vdsc_cfg->slice_height = 8;
+	else if (vdsc_cfg->pic_height % 4 == 0)
+		vdsc_cfg->slice_height = 4;
+	else
+		vdsc_cfg->slice_height = 2;
+
+	ret = intel_dsc_compute_params(crtc_state);
+	if (ret)
+		return ret;
+
+	vdsc_cfg->dsc_version_major =
+		(intel_dp->dsc_dpcd[DP_DSC_REV - DP_DSC_SUPPORT] &
+		 DP_DSC_MAJOR_MASK) >> DP_DSC_MAJOR_SHIFT;
+	vdsc_cfg->dsc_version_minor =
+		min(intel_dp_source_dsc_version_minor(intel_dp),
+		    intel_dp_sink_dsc_version_minor(intel_dp));
+
+	vdsc_cfg->convert_rgb = intel_dp->dsc_dpcd[DP_DSC_DEC_COLOR_FORMAT_CAP - DP_DSC_SUPPORT] &
+		DP_DSC_RGB;
+
+	line_buf_depth = drm_dp_dsc_sink_line_buf_depth(intel_dp->dsc_dpcd);
+	if (!line_buf_depth) {
+		drm_dbg_kms(&i915->drm,
+			    "DSC Sink Line Buffer Depth invalid\n");
+		return -EINVAL;
+	}
+
+	if (vdsc_cfg->dsc_version_minor == 2)
+		vdsc_cfg->line_buf_depth = (line_buf_depth == DSC_1_2_MAX_LINEBUF_DEPTH_BITS) ?
+			DSC_1_2_MAX_LINEBUF_DEPTH_VAL : line_buf_depth;
+	else
+		vdsc_cfg->line_buf_depth = (line_buf_depth > DSC_1_1_MAX_LINEBUF_DEPTH_BITS) ?
+			DSC_1_1_MAX_LINEBUF_DEPTH_BITS : line_buf_depth;
+
+	vdsc_cfg->block_pred_enable =
+		intel_dp->dsc_dpcd[DP_DSC_BLK_PREDICTION_SUPPORT - DP_DSC_SUPPORT] &
+		DP_DSC_BLK_PREDICTION_IS_SUPPORTED;
+
+	return drm_dsc_compute_rc_parameters(vdsc_cfg);
+}
+
+static int intel_dp_dsc_compute_config(struct intel_dp *intel_dp,
+				       struct intel_crtc_state *pipe_config,
+				       struct drm_connector_state *conn_state,
+				       struct link_config_limits *limits)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
+	const struct drm_display_mode *adjusted_mode =
+		&pipe_config->hw.adjusted_mode;
+	int pipe_bpp;
+	int ret;
+
+	pipe_config->fec_enable = !intel_dp_is_edp(intel_dp) &&
+		intel_dp_supports_fec(intel_dp, pipe_config);
+
+	if (!intel_dp_supports_dsc(intel_dp, pipe_config))
+		return -EINVAL;
+
+	pipe_bpp = intel_dp_dsc_compute_bpp(intel_dp, conn_state->max_requested_bpc);
+
+	if (intel_dp->force_dsc_bpc) {
+		pipe_bpp = intel_dp->force_dsc_bpc * 3;
+		drm_dbg_kms(&dev_priv->drm, "Input DSC BPP forced to %d", pipe_bpp);
+	}
+
+	/* Min Input BPC for ICL+ is 8 */
+	if (pipe_bpp < 8 * 3) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "No DSC support for less than 8bpc\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * For now enable DSC for max bpp, max link rate, max lane count.
+	 * Optimize this later for the minimum possible link rate/lane count
+	 * with DSC enabled for the requested mode.
+	 */
+	pipe_config->pipe_bpp = pipe_bpp;
+	pipe_config->port_clock = limits->max_rate;
+	pipe_config->lane_count = limits->max_lane_count;
+
+	if (intel_dp_is_edp(intel_dp)) {
+		pipe_config->dsc.compressed_bpp =
+			min_t(u16, drm_edp_dsc_sink_output_bpp(intel_dp->dsc_dpcd) >> 4,
+			      pipe_config->pipe_bpp);
+		pipe_config->dsc.slice_count =
+			drm_dp_dsc_sink_max_slice_count(intel_dp->dsc_dpcd,
+							true);
+		if (!pipe_config->dsc.slice_count) {
+			drm_dbg_kms(&dev_priv->drm, "Unsupported Slice Count %d\n",
+				    pipe_config->dsc.slice_count);
+			return -EINVAL;
+		}
+	} else {
+		u16 dsc_max_output_bpp;
+		u8 dsc_dp_slice_count;
+
+		dsc_max_output_bpp =
+			intel_dp_dsc_get_output_bpp(dev_priv,
+						    pipe_config->port_clock,
+						    pipe_config->lane_count,
+						    adjusted_mode->crtc_clock,
+						    adjusted_mode->crtc_hdisplay,
+						    pipe_config->bigjoiner_pipes,
+						    pipe_bpp);
+		dsc_dp_slice_count =
+			intel_dp_dsc_get_slice_count(intel_dp,
+						     adjusted_mode->crtc_clock,
+						     adjusted_mode->crtc_hdisplay,
+						     pipe_config->bigjoiner_pipes);
+		if (!dsc_max_output_bpp || !dsc_dp_slice_count) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Compressed BPP/Slice Count not supported\n");
+			return -EINVAL;
+		}
+		pipe_config->dsc.compressed_bpp = min_t(u16,
+							       dsc_max_output_bpp >> 4,
+							       pipe_config->pipe_bpp);
+		pipe_config->dsc.slice_count = dsc_dp_slice_count;
+	}
+
+	/*
+	 * VDSC engine operates at 1 Pixel per clock, so if peak pixel rate
+	 * is greater than the maximum Cdclock and if slice count is even
+	 * then we need to use 2 VDSC instances.
+	 */
+	if (adjusted_mode->crtc_clock > dev_priv->display.cdclk.max_cdclk_freq ||
+	    pipe_config->bigjoiner_pipes) {
+		if (pipe_config->dsc.slice_count < 2) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Cannot split stream to use 2 VDSC instances\n");
+			return -EINVAL;
+		}
+
+		pipe_config->dsc.dsc_split = true;
+	}
+
+	ret = intel_dp_dsc_compute_params(&dig_port->base, pipe_config);
+	if (ret < 0) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Cannot compute valid DSC parameters for Input Bpp = %d "
+			    "Compressed BPP = %d\n",
+			    pipe_config->pipe_bpp,
+			    pipe_config->dsc.compressed_bpp);
+		return ret;
+	}
+
+	pipe_config->dsc.compression_enable = true;
+	drm_dbg_kms(&dev_priv->drm, "DP DSC computed with Input Bpp = %d "
+		    "Compressed Bpp = %d Slice Count = %d\n",
+		    pipe_config->pipe_bpp,
+		    pipe_config->dsc.compressed_bpp,
+		    pipe_config->dsc.slice_count);
+
+	return 0;
+}
+
+static int
+intel_dp_compute_link_config(struct intel_encoder *encoder,
+			     struct intel_crtc_state *pipe_config,
+			     struct drm_connector_state *conn_state,
+			     bool respect_downstream_limits)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	const struct drm_display_mode *adjusted_mode =
+		&pipe_config->hw.adjusted_mode;
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct link_config_limits limits;
+	bool joiner_needs_dsc = false;
+	int ret;
+
+	limits.min_rate = intel_dp_common_rate(intel_dp, 0);
+	limits.max_rate = intel_dp_max_link_rate(intel_dp);
+
+	limits.min_lane_count = 1;
+	limits.max_lane_count = intel_dp_max_lane_count(intel_dp);
+
+	limits.min_bpp = intel_dp_min_bpp(pipe_config->output_format);
+	limits.max_bpp = intel_dp_max_bpp(intel_dp, pipe_config, respect_downstream_limits);
+
+	if (intel_dp->use_max_params) {
+		/*
+		 * Use the maximum clock and number of lanes the eDP panel
+		 * advertizes being capable of in case the initial fast
+		 * optimal params failed us. The panels are generally
+		 * designed to support only a single clock and lane
+		 * configuration, and typically on older panels these
+		 * values correspond to the native resolution of the panel.
+		 */
+		limits.min_lane_count = limits.max_lane_count;
+		limits.min_rate = limits.max_rate;
+	}
+
+	intel_dp_adjust_compliance_config(intel_dp, pipe_config, &limits);
+
+	drm_dbg_kms(&i915->drm, "DP link computation with max lane count %i "
+		    "max rate %d max bpp %d pixel clock %iKHz\n",
+		    limits.max_lane_count, limits.max_rate,
+		    limits.max_bpp, adjusted_mode->crtc_clock);
+
+	if (intel_dp_need_bigjoiner(intel_dp, adjusted_mode->crtc_hdisplay,
+				    adjusted_mode->crtc_clock))
+		pipe_config->bigjoiner_pipes = GENMASK(crtc->pipe + 1, crtc->pipe);
+
+	/*
+	 * Pipe joiner needs compression up to display 12 due to bandwidth
+	 * limitation. DG2 onwards pipe joiner can be enabled without
+	 * compression.
+	 */
+	joiner_needs_dsc = DISPLAY_VER(i915) < 13 && pipe_config->bigjoiner_pipes;
+
+	/*
+	 * Optimize for slow and wide for everything, because there are some
+	 * eDP 1.3 and 1.4 panels don't work well with fast and narrow.
+	 */
+	ret = intel_dp_compute_link_config_wide(intel_dp, pipe_config, conn_state, &limits);
+
+	if (ret || joiner_needs_dsc || intel_dp->force_dsc_en) {
+		drm_dbg_kms(&i915->drm, "Try DSC (fallback=%s, joiner=%s, force=%s)\n",
+			    str_yes_no(ret), str_yes_no(joiner_needs_dsc),
+			    str_yes_no(intel_dp->force_dsc_en));
+		ret = intel_dp_dsc_compute_config(intel_dp, pipe_config,
+						  conn_state, &limits);
+		if (ret < 0)
+			return ret;
+	}
+
+	if (pipe_config->dsc.compression_enable) {
+		drm_dbg_kms(&i915->drm,
+			    "DP lane count %d clock %d Input bpp %d Compressed bpp %d\n",
+			    pipe_config->lane_count, pipe_config->port_clock,
+			    pipe_config->pipe_bpp,
+			    pipe_config->dsc.compressed_bpp);
+
+		drm_dbg_kms(&i915->drm,
+			    "DP link rate required %i available %i\n",
+			    intel_dp_link_required(adjusted_mode->crtc_clock,
+						   pipe_config->dsc.compressed_bpp),
+			    intel_dp_max_data_rate(pipe_config->port_clock,
+						   pipe_config->lane_count));
+	} else {
+		drm_dbg_kms(&i915->drm, "DP lane count %d clock %d bpp %d\n",
+			    pipe_config->lane_count, pipe_config->port_clock,
+			    pipe_config->pipe_bpp);
+
+		drm_dbg_kms(&i915->drm,
+			    "DP link rate required %i available %i\n",
+			    intel_dp_link_required(adjusted_mode->crtc_clock,
+						   pipe_config->pipe_bpp),
+			    intel_dp_max_data_rate(pipe_config->port_clock,
+						   pipe_config->lane_count));
+	}
+	return 0;
+}
+
+bool intel_dp_limited_color_range(const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state)
+{
+	const struct intel_digital_connector_state *intel_conn_state =
+		to_intel_digital_connector_state(conn_state);
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+
+	/*
+	 * Our YCbCr output is always limited range.
+	 * crtc_state->limited_color_range only applies to RGB,
+	 * and it must never be set for YCbCr or we risk setting
+	 * some conflicting bits in PIPECONF which will mess up
+	 * the colors on the monitor.
+	 */
+	if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB)
+		return false;
+
+	if (intel_conn_state->broadcast_rgb == INTEL_BROADCAST_RGB_AUTO) {
+		/*
+		 * See:
+		 * CEA-861-E - 5.1 Default Encoding Parameters
+		 * VESA DisplayPort Ver.1.2a - 5.1.1.1 Video Colorimetry
+		 */
+		return crtc_state->pipe_bpp != 18 &&
+			drm_default_rgb_quant_range(adjusted_mode) ==
+			HDMI_QUANTIZATION_RANGE_LIMITED;
+	} else {
+		return intel_conn_state->broadcast_rgb ==
+			INTEL_BROADCAST_RGB_LIMITED;
+	}
+}
+
+static bool intel_dp_port_has_audio(struct drm_i915_private *dev_priv,
+				    enum port port)
+{
+	if (IS_G4X(dev_priv))
+		return false;
+	if (DISPLAY_VER(dev_priv) < 12 && port == PORT_A)
+		return false;
+
+	return true;
+}
+
+static void intel_dp_compute_vsc_colorimetry(const struct intel_crtc_state *crtc_state,
+					     const struct drm_connector_state *conn_state,
+					     struct drm_dp_vsc_sdp *vsc)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	/*
+	 * Prepare VSC Header for SU as per DP 1.4 spec, Table 2-118
+	 * VSC SDP supporting 3D stereo, PSR2, and Pixel Encoding/
+	 * Colorimetry Format indication.
+	 */
+	vsc->revision = 0x5;
+	vsc->length = 0x13;
+
+	/* DP 1.4a spec, Table 2-120 */
+	switch (crtc_state->output_format) {
+	case INTEL_OUTPUT_FORMAT_YCBCR444:
+		vsc->pixelformat = DP_PIXELFORMAT_YUV444;
+		break;
+	case INTEL_OUTPUT_FORMAT_YCBCR420:
+		vsc->pixelformat = DP_PIXELFORMAT_YUV420;
+		break;
+	case INTEL_OUTPUT_FORMAT_RGB:
+	default:
+		vsc->pixelformat = DP_PIXELFORMAT_RGB;
+	}
+
+	switch (conn_state->colorspace) {
+	case DRM_MODE_COLORIMETRY_BT709_YCC:
+		vsc->colorimetry = DP_COLORIMETRY_BT709_YCC;
+		break;
+	case DRM_MODE_COLORIMETRY_XVYCC_601:
+		vsc->colorimetry = DP_COLORIMETRY_XVYCC_601;
+		break;
+	case DRM_MODE_COLORIMETRY_XVYCC_709:
+		vsc->colorimetry = DP_COLORIMETRY_XVYCC_709;
+		break;
+	case DRM_MODE_COLORIMETRY_SYCC_601:
+		vsc->colorimetry = DP_COLORIMETRY_SYCC_601;
+		break;
+	case DRM_MODE_COLORIMETRY_OPYCC_601:
+		vsc->colorimetry = DP_COLORIMETRY_OPYCC_601;
+		break;
+	case DRM_MODE_COLORIMETRY_BT2020_CYCC:
+		vsc->colorimetry = DP_COLORIMETRY_BT2020_CYCC;
+		break;
+	case DRM_MODE_COLORIMETRY_BT2020_RGB:
+		vsc->colorimetry = DP_COLORIMETRY_BT2020_RGB;
+		break;
+	case DRM_MODE_COLORIMETRY_BT2020_YCC:
+		vsc->colorimetry = DP_COLORIMETRY_BT2020_YCC;
+		break;
+	case DRM_MODE_COLORIMETRY_DCI_P3_RGB_D65:
+	case DRM_MODE_COLORIMETRY_DCI_P3_RGB_THEATER:
+		vsc->colorimetry = DP_COLORIMETRY_DCI_P3_RGB;
+		break;
+	default:
+		/*
+		 * RGB->YCBCR color conversion uses the BT.709
+		 * color space.
+		 */
+		if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR420)
+			vsc->colorimetry = DP_COLORIMETRY_BT709_YCC;
+		else
+			vsc->colorimetry = DP_COLORIMETRY_DEFAULT;
+		break;
+	}
+
+	vsc->bpc = crtc_state->pipe_bpp / 3;
+
+	/* only RGB pixelformat supports 6 bpc */
+	drm_WARN_ON(&dev_priv->drm,
+		    vsc->bpc == 6 && vsc->pixelformat != DP_PIXELFORMAT_RGB);
+
+	/* all YCbCr are always limited range */
+	vsc->dynamic_range = DP_DYNAMIC_RANGE_CTA;
+	vsc->content_type = DP_CONTENT_TYPE_NOT_DEFINED;
+}
+
+static void intel_dp_compute_vsc_sdp(struct intel_dp *intel_dp,
+				     struct intel_crtc_state *crtc_state,
+				     const struct drm_connector_state *conn_state)
+{
+	struct drm_dp_vsc_sdp *vsc = &crtc_state->infoframes.vsc;
+
+	/* When a crtc state has PSR, VSC SDP will be handled by PSR routine */
+	if (crtc_state->has_psr)
+		return;
+
+	if (!intel_dp_needs_vsc_sdp(crtc_state, conn_state))
+		return;
+
+	crtc_state->infoframes.enable |= intel_hdmi_infoframe_enable(DP_SDP_VSC);
+	vsc->sdp_type = DP_SDP_VSC;
+	intel_dp_compute_vsc_colorimetry(crtc_state, conn_state,
+					 &crtc_state->infoframes.vsc);
+}
+
+void intel_dp_compute_psr_vsc_sdp(struct intel_dp *intel_dp,
+				  const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state,
+				  struct drm_dp_vsc_sdp *vsc)
+{
+	vsc->sdp_type = DP_SDP_VSC;
+
+	if (crtc_state->has_psr2) {
+		if (intel_dp->psr.colorimetry_support &&
+		    intel_dp_needs_vsc_sdp(crtc_state, conn_state)) {
+			/* [PSR2, +Colorimetry] */
+			intel_dp_compute_vsc_colorimetry(crtc_state, conn_state,
+							 vsc);
+		} else {
+			/*
+			 * [PSR2, -Colorimetry]
+			 * Prepare VSC Header for SU as per eDP 1.4 spec, Table 6-11
+			 * 3D stereo + PSR/PSR2 + Y-coordinate.
+			 */
+			vsc->revision = 0x4;
+			vsc->length = 0xe;
+		}
+	} else {
+		/*
+		 * [PSR1]
+		 * Prepare VSC Header for SU as per DP 1.4 spec, Table 2-118
+		 * VSC SDP supporting 3D stereo + PSR (applies to eDP v1.3 or
+		 * higher).
+		 */
+		vsc->revision = 0x2;
+		vsc->length = 0x8;
+	}
+}
+
+static void
+intel_dp_compute_hdr_metadata_infoframe_sdp(struct intel_dp *intel_dp,
+					    struct intel_crtc_state *crtc_state,
+					    const struct drm_connector_state *conn_state)
+{
+	int ret;
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct hdmi_drm_infoframe *drm_infoframe = &crtc_state->infoframes.drm.drm;
+
+	if (!conn_state->hdr_output_metadata)
+		return;
+
+	ret = drm_hdmi_infoframe_set_hdr_metadata(drm_infoframe, conn_state);
+
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm, "couldn't set HDR metadata in infoframe\n");
+		return;
+	}
+
+	crtc_state->infoframes.enable |=
+		intel_hdmi_infoframe_enable(HDMI_PACKET_TYPE_GAMUT_METADATA);
+}
+
+static bool cpu_transcoder_has_drrs(struct drm_i915_private *i915,
+				    enum transcoder cpu_transcoder)
+{
+	if (HAS_DOUBLE_BUFFERED_M_N(i915))
+		return true;
+
+	return intel_cpu_transcoder_has_m2_n2(i915, cpu_transcoder);
+}
+
+static bool can_enable_drrs(struct intel_connector *connector,
+			    const struct intel_crtc_state *pipe_config,
+			    const struct drm_display_mode *downclock_mode)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+
+	if (pipe_config->vrr.enable)
+		return false;
+
+	/*
+	 * DRRS and PSR can't be enable together, so giving preference to PSR
+	 * as it allows more power-savings by complete shutting down display,
+	 * so to guarantee this, intel_drrs_compute_config() must be called
+	 * after intel_psr_compute_config().
+	 */
+	if (pipe_config->has_psr)
+		return false;
+
+	/* FIXME missing FDI M2/N2 etc. */
+	if (pipe_config->has_pch_encoder)
+		return false;
+
+	if (!cpu_transcoder_has_drrs(i915, pipe_config->cpu_transcoder))
+		return false;
+
+	return downclock_mode &&
+		intel_panel_drrs_type(connector) == DRRS_TYPE_SEAMLESS;
+}
+
+static void
+intel_dp_drrs_compute_config(struct intel_connector *connector,
+			     struct intel_crtc_state *pipe_config,
+			     int output_bpp)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	const struct drm_display_mode *downclock_mode =
+		intel_panel_downclock_mode(connector, &pipe_config->hw.adjusted_mode);
+	int pixel_clock;
+
+	if (has_seamless_m_n(connector))
+		pipe_config->seamless_m_n = true;
+
+	if (!can_enable_drrs(connector, pipe_config, downclock_mode)) {
+		if (intel_cpu_transcoder_has_m2_n2(i915, pipe_config->cpu_transcoder))
+			intel_zero_m_n(&pipe_config->dp_m2_n2);
+		return;
+	}
+
+	if (IS_IRONLAKE(i915) || IS_SANDYBRIDGE(i915) || IS_IVYBRIDGE(i915))
+		pipe_config->msa_timing_delay = connector->panel.vbt.edp.drrs_msa_timing_delay;
+
+	pipe_config->has_drrs = true;
+
+	pixel_clock = downclock_mode->clock;
+	if (pipe_config->splitter.enable)
+		pixel_clock /= pipe_config->splitter.link_count;
+
+	intel_link_compute_m_n(output_bpp, pipe_config->lane_count, pixel_clock,
+			       pipe_config->port_clock, &pipe_config->dp_m2_n2,
+			       pipe_config->fec_enable);
+
+	/* FIXME: abstract this better */
+	if (pipe_config->splitter.enable)
+		pipe_config->dp_m2_n2.data_m *= pipe_config->splitter.link_count;
+}
+
+static bool intel_dp_has_audio(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state,
+			       const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	const struct intel_digital_connector_state *intel_conn_state =
+		to_intel_digital_connector_state(conn_state);
+
+	if (!intel_dp_port_has_audio(i915, encoder->port))
+		return false;
+
+	if (intel_conn_state->force_audio == HDMI_AUDIO_AUTO)
+		return intel_dp->has_audio;
+	else
+		return intel_conn_state->force_audio == HDMI_AUDIO_ON;
+}
+
+static int
+intel_dp_compute_output_format(struct intel_encoder *encoder,
+			       struct intel_crtc_state *crtc_state,
+			       struct drm_connector_state *conn_state,
+			       bool respect_downstream_limits)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_connector *connector = intel_dp->attached_connector;
+	const struct drm_display_info *info = &connector->base.display_info;
+	const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	bool ycbcr_420_only;
+	int ret;
+
+	ycbcr_420_only = drm_mode_is_420_only(info, adjusted_mode);
+
+	crtc_state->output_format = intel_dp_output_format(connector, ycbcr_420_only);
+
+	if (ycbcr_420_only && !intel_dp_is_ycbcr420(intel_dp, crtc_state)) {
+		drm_dbg_kms(&i915->drm,
+			    "YCbCr 4:2:0 mode but YCbCr 4:2:0 output not possible. Falling back to RGB.\n");
+		crtc_state->output_format = INTEL_OUTPUT_FORMAT_RGB;
+	}
+
+	ret = intel_dp_compute_link_config(encoder, crtc_state, conn_state,
+					   respect_downstream_limits);
+	if (ret) {
+		if (intel_dp_is_ycbcr420(intel_dp, crtc_state) ||
+		    !connector->base.ycbcr_420_allowed ||
+		    !drm_mode_is_420_also(info, adjusted_mode))
+			return ret;
+
+		crtc_state->output_format = intel_dp_output_format(connector, true);
+		ret = intel_dp_compute_link_config(encoder, crtc_state, conn_state,
+						   respect_downstream_limits);
+	}
+
+	return ret;
+}
+
+int
+intel_dp_compute_config(struct intel_encoder *encoder,
+			struct intel_crtc_state *pipe_config,
+			struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	const struct drm_display_mode *fixed_mode;
+	struct intel_connector *connector = intel_dp->attached_connector;
+	int ret = 0, output_bpp;
+
+	if (HAS_PCH_SPLIT(dev_priv) && !HAS_DDI(dev_priv) && encoder->port != PORT_A)
+		pipe_config->has_pch_encoder = true;
+
+	pipe_config->has_audio = intel_dp_has_audio(encoder, pipe_config, conn_state);
+
+	fixed_mode = intel_panel_fixed_mode(connector, adjusted_mode);
+	if (intel_dp_is_edp(intel_dp) && fixed_mode) {
+		ret = intel_panel_compute_config(connector, adjusted_mode);
+		if (ret)
+			return ret;
+	}
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	if (HAS_GMCH(dev_priv) &&
+	    adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
+		return -EINVAL;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK)
+		return -EINVAL;
+
+	if (intel_dp_hdisplay_bad(dev_priv, adjusted_mode->crtc_hdisplay))
+		return -EINVAL;
+
+	/*
+	 * Try to respect downstream TMDS clock limits first, if
+	 * that fails assume the user might know something we don't.
+	 */
+	ret = intel_dp_compute_output_format(encoder, pipe_config, conn_state, true);
+	if (ret)
+		ret = intel_dp_compute_output_format(encoder, pipe_config, conn_state, false);
+	if (ret)
+		return ret;
+
+	if ((intel_dp_is_edp(intel_dp) && fixed_mode) ||
+	    pipe_config->output_format == INTEL_OUTPUT_FORMAT_YCBCR420) {
+		ret = intel_panel_fitting(pipe_config, conn_state);
+		if (ret)
+			return ret;
+	}
+
+	pipe_config->limited_color_range =
+		intel_dp_limited_color_range(pipe_config, conn_state);
+
+	if (pipe_config->dsc.compression_enable)
+		output_bpp = pipe_config->dsc.compressed_bpp;
+	else
+		output_bpp = intel_dp_output_bpp(pipe_config->output_format,
+						 pipe_config->pipe_bpp);
+
+	if (intel_dp->mso_link_count) {
+		int n = intel_dp->mso_link_count;
+		int overlap = intel_dp->mso_pixel_overlap;
+
+		pipe_config->splitter.enable = true;
+		pipe_config->splitter.link_count = n;
+		pipe_config->splitter.pixel_overlap = overlap;
+
+		drm_dbg_kms(&dev_priv->drm, "MSO link count %d, pixel overlap %d\n",
+			    n, overlap);
+
+		adjusted_mode->crtc_hdisplay = adjusted_mode->crtc_hdisplay / n + overlap;
+		adjusted_mode->crtc_hblank_start = adjusted_mode->crtc_hblank_start / n + overlap;
+		adjusted_mode->crtc_hblank_end = adjusted_mode->crtc_hblank_end / n + overlap;
+		adjusted_mode->crtc_hsync_start = adjusted_mode->crtc_hsync_start / n + overlap;
+		adjusted_mode->crtc_hsync_end = adjusted_mode->crtc_hsync_end / n + overlap;
+		adjusted_mode->crtc_htotal = adjusted_mode->crtc_htotal / n + overlap;
+		adjusted_mode->crtc_clock /= n;
+	}
+
+	intel_link_compute_m_n(output_bpp,
+			       pipe_config->lane_count,
+			       adjusted_mode->crtc_clock,
+			       pipe_config->port_clock,
+			       &pipe_config->dp_m_n,
+			       pipe_config->fec_enable);
+
+	/* FIXME: abstract this better */
+	if (pipe_config->splitter.enable)
+		pipe_config->dp_m_n.data_m *= pipe_config->splitter.link_count;
+
+	if (!HAS_DDI(dev_priv))
+		g4x_dp_set_clock(encoder, pipe_config);
+
+	intel_vrr_compute_config(pipe_config, conn_state);
+	intel_psr_compute_config(intel_dp, pipe_config, conn_state);
+	intel_dp_drrs_compute_config(connector, pipe_config, output_bpp);
+	intel_dp_compute_vsc_sdp(intel_dp, pipe_config, conn_state);
+	intel_dp_compute_hdr_metadata_infoframe_sdp(intel_dp, pipe_config, conn_state);
+
+	return 0;
+}
+
+void intel_dp_set_link_params(struct intel_dp *intel_dp,
+			      int link_rate, int lane_count)
+{
+	memset(intel_dp->train_set, 0, sizeof(intel_dp->train_set));
+	intel_dp->link_trained = false;
+	intel_dp->link_rate = link_rate;
+	intel_dp->lane_count = lane_count;
+}
+
+static void intel_dp_reset_max_link_params(struct intel_dp *intel_dp)
+{
+	intel_dp->max_link_lane_count = intel_dp_max_common_lane_count(intel_dp);
+	intel_dp->max_link_rate = intel_dp_max_common_rate(intel_dp);
+}
+
+/* Enable backlight PWM and backlight PP control. */
+void intel_edp_backlight_on(const struct intel_crtc_state *crtc_state,
+			    const struct drm_connector_state *conn_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(to_intel_encoder(conn_state->best_encoder));
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	drm_dbg_kms(&i915->drm, "\n");
+
+	intel_backlight_enable(crtc_state, conn_state);
+	intel_pps_backlight_on(intel_dp);
+}
+
+/* Disable backlight PP control and backlight PWM. */
+void intel_edp_backlight_off(const struct drm_connector_state *old_conn_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(to_intel_encoder(old_conn_state->best_encoder));
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	drm_dbg_kms(&i915->drm, "\n");
+
+	intel_pps_backlight_off(intel_dp);
+	intel_backlight_disable(old_conn_state);
+}
+
+static bool downstream_hpd_needs_d0(struct intel_dp *intel_dp)
+{
+	/*
+	 * DPCD 1.2+ should support BRANCH_DEVICE_CTRL, and thus
+	 * be capable of signalling downstream hpd with a long pulse.
+	 * Whether or not that means D3 is safe to use is not clear,
+	 * but let's assume so until proven otherwise.
+	 *
+	 * FIXME should really check all downstream ports...
+	 */
+	return intel_dp->dpcd[DP_DPCD_REV] == 0x11 &&
+		drm_dp_is_branch(intel_dp->dpcd) &&
+		intel_dp->downstream_ports[0] & DP_DS_PORT_HPD;
+}
+
+void intel_dp_sink_set_decompression_state(struct intel_dp *intel_dp,
+					   const struct intel_crtc_state *crtc_state,
+					   bool enable)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int ret;
+
+	if (!crtc_state->dsc.compression_enable)
+		return;
+
+	ret = drm_dp_dpcd_writeb(&intel_dp->aux, DP_DSC_ENABLE,
+				 enable ? DP_DECOMPRESSION_EN : 0);
+	if (ret < 0)
+		drm_dbg_kms(&i915->drm,
+			    "Failed to %s sink decompression state\n",
+			    str_enable_disable(enable));
+}
+
+static void
+intel_edp_init_source_oui(struct intel_dp *intel_dp, bool careful)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 oui[] = { 0x00, 0xaa, 0x01 };
+	u8 buf[3] = { 0 };
+
+	/*
+	 * During driver init, we want to be careful and avoid changing the source OUI if it's
+	 * already set to what we want, so as to avoid clearing any state by accident
+	 */
+	if (careful) {
+		if (drm_dp_dpcd_read(&intel_dp->aux, DP_SOURCE_OUI, buf, sizeof(buf)) < 0)
+			drm_err(&i915->drm, "Failed to read source OUI\n");
+
+		if (memcmp(oui, buf, sizeof(oui)) == 0)
+			return;
+	}
+
+	if (drm_dp_dpcd_write(&intel_dp->aux, DP_SOURCE_OUI, oui, sizeof(oui)) < 0)
+		drm_err(&i915->drm, "Failed to write source OUI\n");
+
+	intel_dp->last_oui_write = jiffies;
+}
+
+void intel_dp_wait_source_oui(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	drm_dbg_kms(&i915->drm, "Performing OUI wait\n");
+	wait_remaining_ms_from_jiffies(intel_dp->last_oui_write, 30);
+}
+
+/* If the device supports it, try to set the power state appropriately */
+void intel_dp_set_power(struct intel_dp *intel_dp, u8 mode)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	int ret, i;
+
+	/* Should have a valid DPCD by this point */
+	if (intel_dp->dpcd[DP_DPCD_REV] < 0x11)
+		return;
+
+	if (mode != DP_SET_POWER_D0) {
+		if (downstream_hpd_needs_d0(intel_dp))
+			return;
+
+		ret = drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER, mode);
+	} else {
+		struct intel_lspcon *lspcon = dp_to_lspcon(intel_dp);
+
+		lspcon_resume(dp_to_dig_port(intel_dp));
+
+		/* Write the source OUI as early as possible */
+		if (intel_dp_is_edp(intel_dp))
+			intel_edp_init_source_oui(intel_dp, false);
+
+		/*
+		 * When turning on, we need to retry for 1ms to give the sink
+		 * time to wake up.
+		 */
+		for (i = 0; i < 3; i++) {
+			ret = drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER, mode);
+			if (ret == 1)
+				break;
+			msleep(1);
+		}
+
+		if (ret == 1 && lspcon->active)
+			lspcon_wait_pcon_mode(lspcon);
+	}
+
+	if (ret != 1)
+		drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s] Set power to %s failed\n",
+			    encoder->base.base.id, encoder->base.name,
+			    mode == DP_SET_POWER_D0 ? "D0" : "D3");
+}
+
+static bool
+intel_dp_get_dpcd(struct intel_dp *intel_dp);
+
+/**
+ * intel_dp_sync_state - sync the encoder state during init/resume
+ * @encoder: intel encoder to sync
+ * @crtc_state: state for the CRTC connected to the encoder
+ *
+ * Sync any state stored in the encoder wrt. HW state during driver init
+ * and system resume.
+ */
+void intel_dp_sync_state(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *crtc_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+	if (!crtc_state)
+		return;
+
+	/*
+	 * Don't clobber DPCD if it's been already read out during output
+	 * setup (eDP) or detect.
+	 */
+	if (intel_dp->dpcd[DP_DPCD_REV] == 0)
+		intel_dp_get_dpcd(intel_dp);
+
+	intel_dp_reset_max_link_params(intel_dp);
+}
+
+bool intel_dp_initial_fastset_check(struct intel_encoder *encoder,
+				    struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+	/*
+	 * If BIOS has set an unsupported or non-standard link rate for some
+	 * reason force an encoder recompute and full modeset.
+	 */
+	if (intel_dp_rate_index(intel_dp->source_rates, intel_dp->num_source_rates,
+				crtc_state->port_clock) < 0) {
+		drm_dbg_kms(&i915->drm, "Forcing full modeset due to unsupported link rate\n");
+		crtc_state->uapi.connectors_changed = true;
+		return false;
+	}
+
+	/*
+	 * FIXME hack to force full modeset when DSC is being used.
+	 *
+	 * As long as we do not have full state readout and config comparison
+	 * of crtc_state->dsc, we have no way to ensure reliable fastset.
+	 * Remove once we have readout for DSC.
+	 */
+	if (crtc_state->dsc.compression_enable) {
+		drm_dbg_kms(&i915->drm, "Forcing full modeset due to DSC being enabled\n");
+		crtc_state->uapi.mode_changed = true;
+		return false;
+	}
+
+	if (CAN_PSR(intel_dp)) {
+		drm_dbg_kms(&i915->drm, "Forcing full modeset to compute PSR state\n");
+		crtc_state->uapi.mode_changed = true;
+		return false;
+	}
+
+	return true;
+}
+
+static void intel_dp_get_pcon_dsc_cap(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	/* Clear the cached register set to avoid using stale values */
+
+	memset(intel_dp->pcon_dsc_dpcd, 0, sizeof(intel_dp->pcon_dsc_dpcd));
+
+	if (drm_dp_dpcd_read(&intel_dp->aux, DP_PCON_DSC_ENCODER,
+			     intel_dp->pcon_dsc_dpcd,
+			     sizeof(intel_dp->pcon_dsc_dpcd)) < 0)
+		drm_err(&i915->drm, "Failed to read DPCD register 0x%x\n",
+			DP_PCON_DSC_ENCODER);
+
+	drm_dbg_kms(&i915->drm, "PCON ENCODER DSC DPCD: %*ph\n",
+		    (int)sizeof(intel_dp->pcon_dsc_dpcd), intel_dp->pcon_dsc_dpcd);
+}
+
+static int intel_dp_pcon_get_frl_mask(u8 frl_bw_mask)
+{
+	int bw_gbps[] = {9, 18, 24, 32, 40, 48};
+	int i;
+
+	for (i = ARRAY_SIZE(bw_gbps) - 1; i >= 0; i--) {
+		if (frl_bw_mask & (1 << i))
+			return bw_gbps[i];
+	}
+	return 0;
+}
+
+static int intel_dp_pcon_set_frl_mask(int max_frl)
+{
+	switch (max_frl) {
+	case 48:
+		return DP_PCON_FRL_BW_MASK_48GBPS;
+	case 40:
+		return DP_PCON_FRL_BW_MASK_40GBPS;
+	case 32:
+		return DP_PCON_FRL_BW_MASK_32GBPS;
+	case 24:
+		return DP_PCON_FRL_BW_MASK_24GBPS;
+	case 18:
+		return DP_PCON_FRL_BW_MASK_18GBPS;
+	case 9:
+		return DP_PCON_FRL_BW_MASK_9GBPS;
+	}
+
+	return 0;
+}
+
+static int intel_dp_hdmi_sink_max_frl(struct intel_dp *intel_dp)
+{
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
+	struct drm_connector *connector = &intel_connector->base;
+	int max_frl_rate;
+	int max_lanes, rate_per_lane;
+	int max_dsc_lanes, dsc_rate_per_lane;
+
+	max_lanes = connector->display_info.hdmi.max_lanes;
+	rate_per_lane = connector->display_info.hdmi.max_frl_rate_per_lane;
+	max_frl_rate = max_lanes * rate_per_lane;
+
+	if (connector->display_info.hdmi.dsc_cap.v_1p2) {
+		max_dsc_lanes = connector->display_info.hdmi.dsc_cap.max_lanes;
+		dsc_rate_per_lane = connector->display_info.hdmi.dsc_cap.max_frl_rate_per_lane;
+		if (max_dsc_lanes && dsc_rate_per_lane)
+			max_frl_rate = min(max_frl_rate, max_dsc_lanes * dsc_rate_per_lane);
+	}
+
+	return max_frl_rate;
+}
+
+static bool
+intel_dp_pcon_is_frl_trained(struct intel_dp *intel_dp,
+			     u8 max_frl_bw_mask, u8 *frl_trained_mask)
+{
+	if (drm_dp_pcon_hdmi_link_active(&intel_dp->aux) &&
+	    drm_dp_pcon_hdmi_link_mode(&intel_dp->aux, frl_trained_mask) == DP_PCON_HDMI_MODE_FRL &&
+	    *frl_trained_mask >= max_frl_bw_mask)
+		return true;
+
+	return false;
+}
+
+static int intel_dp_pcon_start_frl_training(struct intel_dp *intel_dp)
+{
+#define TIMEOUT_FRL_READY_MS 500
+#define TIMEOUT_HDMI_LINK_ACTIVE_MS 1000
+
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int max_frl_bw, max_pcon_frl_bw, max_edid_frl_bw, ret;
+	u8 max_frl_bw_mask = 0, frl_trained_mask;
+	bool is_active;
+
+	max_pcon_frl_bw = intel_dp->dfp.pcon_max_frl_bw;
+	drm_dbg(&i915->drm, "PCON max rate = %d Gbps\n", max_pcon_frl_bw);
+
+	max_edid_frl_bw = intel_dp_hdmi_sink_max_frl(intel_dp);
+	drm_dbg(&i915->drm, "Sink max rate from EDID = %d Gbps\n", max_edid_frl_bw);
+
+	max_frl_bw = min(max_edid_frl_bw, max_pcon_frl_bw);
+
+	if (max_frl_bw <= 0)
+		return -EINVAL;
+
+	max_frl_bw_mask = intel_dp_pcon_set_frl_mask(max_frl_bw);
+	drm_dbg(&i915->drm, "MAX_FRL_BW_MASK = %u\n", max_frl_bw_mask);
+
+	if (intel_dp_pcon_is_frl_trained(intel_dp, max_frl_bw_mask, &frl_trained_mask))
+		goto frl_trained;
+
+	ret = drm_dp_pcon_frl_prepare(&intel_dp->aux, false);
+	if (ret < 0)
+		return ret;
+	/* Wait for PCON to be FRL Ready */
+	wait_for(is_active = drm_dp_pcon_is_frl_ready(&intel_dp->aux) == true, TIMEOUT_FRL_READY_MS);
+
+	if (!is_active)
+		return -ETIMEDOUT;
+
+	ret = drm_dp_pcon_frl_configure_1(&intel_dp->aux, max_frl_bw,
+					  DP_PCON_ENABLE_SEQUENTIAL_LINK);
+	if (ret < 0)
+		return ret;
+	ret = drm_dp_pcon_frl_configure_2(&intel_dp->aux, max_frl_bw_mask,
+					  DP_PCON_FRL_LINK_TRAIN_NORMAL);
+	if (ret < 0)
+		return ret;
+	ret = drm_dp_pcon_frl_enable(&intel_dp->aux);
+	if (ret < 0)
+		return ret;
+	/*
+	 * Wait for FRL to be completed
+	 * Check if the HDMI Link is up and active.
+	 */
+	wait_for(is_active =
+		 intel_dp_pcon_is_frl_trained(intel_dp, max_frl_bw_mask, &frl_trained_mask),
+		 TIMEOUT_HDMI_LINK_ACTIVE_MS);
+
+	if (!is_active)
+		return -ETIMEDOUT;
+
+frl_trained:
+	drm_dbg(&i915->drm, "FRL_TRAINED_MASK = %u\n", frl_trained_mask);
+	intel_dp->frl.trained_rate_gbps = intel_dp_pcon_get_frl_mask(frl_trained_mask);
+	intel_dp->frl.is_trained = true;
+	drm_dbg(&i915->drm, "FRL trained with : %d Gbps\n", intel_dp->frl.trained_rate_gbps);
+
+	return 0;
+}
+
+static bool intel_dp_is_hdmi_2_1_sink(struct intel_dp *intel_dp)
+{
+	if (drm_dp_is_branch(intel_dp->dpcd) &&
+	    intel_dp->has_hdmi_sink &&
+	    intel_dp_hdmi_sink_max_frl(intel_dp) > 0)
+		return true;
+
+	return false;
+}
+
+static
+int intel_dp_pcon_set_tmds_mode(struct intel_dp *intel_dp)
+{
+	int ret;
+	u8 buf = 0;
+
+	/* Set PCON source control mode */
+	buf |= DP_PCON_ENABLE_SOURCE_CTL_MODE;
+
+	ret = drm_dp_dpcd_writeb(&intel_dp->aux, DP_PCON_HDMI_LINK_CONFIG_1, buf);
+	if (ret < 0)
+		return ret;
+
+	/* Set HDMI LINK ENABLE */
+	buf |= DP_PCON_ENABLE_HDMI_LINK;
+	ret = drm_dp_dpcd_writeb(&intel_dp->aux, DP_PCON_HDMI_LINK_CONFIG_1, buf);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+void intel_dp_check_frl_training(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	/*
+	 * Always go for FRL training if:
+	 * -PCON supports SRC_CTL_MODE (VESA DP2.0-HDMI2.1 PCON Spec Draft-1 Sec-7)
+	 * -sink is HDMI2.1
+	 */
+	if (!(intel_dp->downstream_ports[2] & DP_PCON_SOURCE_CTL_MODE) ||
+	    !intel_dp_is_hdmi_2_1_sink(intel_dp) ||
+	    intel_dp->frl.is_trained)
+		return;
+
+	if (intel_dp_pcon_start_frl_training(intel_dp) < 0) {
+		int ret, mode;
+
+		drm_dbg(&dev_priv->drm, "Couldn't set FRL mode, continuing with TMDS mode\n");
+		ret = intel_dp_pcon_set_tmds_mode(intel_dp);
+		mode = drm_dp_pcon_hdmi_link_mode(&intel_dp->aux, NULL);
+
+		if (ret < 0 || mode != DP_PCON_HDMI_MODE_TMDS)
+			drm_dbg(&dev_priv->drm, "Issue with PCON, cannot set TMDS mode\n");
+	} else {
+		drm_dbg(&dev_priv->drm, "FRL training Completed\n");
+	}
+}
+
+static int
+intel_dp_pcon_dsc_enc_slice_height(const struct intel_crtc_state *crtc_state)
+{
+	int vactive = crtc_state->hw.adjusted_mode.vdisplay;
+
+	return intel_hdmi_dsc_get_slice_height(vactive);
+}
+
+static int
+intel_dp_pcon_dsc_enc_slices(struct intel_dp *intel_dp,
+			     const struct intel_crtc_state *crtc_state)
+{
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
+	struct drm_connector *connector = &intel_connector->base;
+	int hdmi_throughput = connector->display_info.hdmi.dsc_cap.clk_per_slice;
+	int hdmi_max_slices = connector->display_info.hdmi.dsc_cap.max_slices;
+	int pcon_max_slices = drm_dp_pcon_dsc_max_slices(intel_dp->pcon_dsc_dpcd);
+	int pcon_max_slice_width = drm_dp_pcon_dsc_max_slice_width(intel_dp->pcon_dsc_dpcd);
+
+	return intel_hdmi_dsc_get_num_slices(crtc_state, pcon_max_slices,
+					     pcon_max_slice_width,
+					     hdmi_max_slices, hdmi_throughput);
+}
+
+static int
+intel_dp_pcon_dsc_enc_bpp(struct intel_dp *intel_dp,
+			  const struct intel_crtc_state *crtc_state,
+			  int num_slices, int slice_width)
+{
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
+	struct drm_connector *connector = &intel_connector->base;
+	int output_format = crtc_state->output_format;
+	bool hdmi_all_bpp = connector->display_info.hdmi.dsc_cap.all_bpp;
+	int pcon_fractional_bpp = drm_dp_pcon_dsc_bpp_incr(intel_dp->pcon_dsc_dpcd);
+	int hdmi_max_chunk_bytes =
+		connector->display_info.hdmi.dsc_cap.total_chunk_kbytes * 1024;
+
+	return intel_hdmi_dsc_get_bpp(pcon_fractional_bpp, slice_width,
+				      num_slices, output_format, hdmi_all_bpp,
+				      hdmi_max_chunk_bytes);
+}
+
+void
+intel_dp_pcon_dsc_configure(struct intel_dp *intel_dp,
+			    const struct intel_crtc_state *crtc_state)
+{
+	u8 pps_param[6];
+	int slice_height;
+	int slice_width;
+	int num_slices;
+	int bits_per_pixel;
+	int ret;
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct drm_connector *connector;
+	bool hdmi_is_dsc_1_2;
+
+	if (!intel_dp_is_hdmi_2_1_sink(intel_dp))
+		return;
+
+	if (!intel_connector)
+		return;
+	connector = &intel_connector->base;
+	hdmi_is_dsc_1_2 = connector->display_info.hdmi.dsc_cap.v_1p2;
+
+	if (!drm_dp_pcon_enc_is_dsc_1_2(intel_dp->pcon_dsc_dpcd) ||
+	    !hdmi_is_dsc_1_2)
+		return;
+
+	slice_height = intel_dp_pcon_dsc_enc_slice_height(crtc_state);
+	if (!slice_height)
+		return;
+
+	num_slices = intel_dp_pcon_dsc_enc_slices(intel_dp, crtc_state);
+	if (!num_slices)
+		return;
+
+	slice_width = DIV_ROUND_UP(crtc_state->hw.adjusted_mode.hdisplay,
+				   num_slices);
+
+	bits_per_pixel = intel_dp_pcon_dsc_enc_bpp(intel_dp, crtc_state,
+						   num_slices, slice_width);
+	if (!bits_per_pixel)
+		return;
+
+	pps_param[0] = slice_height & 0xFF;
+	pps_param[1] = slice_height >> 8;
+	pps_param[2] = slice_width & 0xFF;
+	pps_param[3] = slice_width >> 8;
+	pps_param[4] = bits_per_pixel & 0xFF;
+	pps_param[5] = (bits_per_pixel >> 8) & 0x3;
+
+	ret = drm_dp_pcon_pps_override_param(&intel_dp->aux, pps_param);
+	if (ret < 0)
+		drm_dbg_kms(&i915->drm, "Failed to set pcon DSC\n");
+}
+
+void intel_dp_configure_protocol_converter(struct intel_dp *intel_dp,
+					   const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 tmp;
+
+	if (intel_dp->dpcd[DP_DPCD_REV] < 0x13)
+		return;
+
+	if (!drm_dp_is_branch(intel_dp->dpcd))
+		return;
+
+	tmp = intel_dp->has_hdmi_sink ?
+		DP_HDMI_DVI_OUTPUT_CONFIG : 0;
+
+	if (drm_dp_dpcd_writeb(&intel_dp->aux,
+			       DP_PROTOCOL_CONVERTER_CONTROL_0, tmp) != 1)
+		drm_dbg_kms(&i915->drm, "Failed to %s protocol converter HDMI mode\n",
+			    str_enable_disable(intel_dp->has_hdmi_sink));
+
+	tmp = crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR444 &&
+		intel_dp->dfp.ycbcr_444_to_420 ? DP_CONVERSION_TO_YCBCR420_ENABLE : 0;
+
+	if (drm_dp_dpcd_writeb(&intel_dp->aux,
+			       DP_PROTOCOL_CONVERTER_CONTROL_1, tmp) != 1)
+		drm_dbg_kms(&i915->drm,
+			    "Failed to %s protocol converter YCbCr 4:2:0 conversion mode\n",
+			    str_enable_disable(intel_dp->dfp.ycbcr_444_to_420));
+
+	tmp = intel_dp->dfp.rgb_to_ycbcr ?
+		DP_CONVERSION_BT709_RGB_YCBCR_ENABLE : 0;
+
+	if (drm_dp_pcon_convert_rgb_to_ycbcr(&intel_dp->aux, tmp) < 0)
+		drm_dbg_kms(&i915->drm,
+			   "Failed to %s protocol converter RGB->YCbCr conversion mode\n",
+			   str_enable_disable(tmp));
+}
+
+
+bool intel_dp_get_colorimetry_status(struct intel_dp *intel_dp)
+{
+	u8 dprx = 0;
+
+	if (drm_dp_dpcd_readb(&intel_dp->aux, DP_DPRX_FEATURE_ENUMERATION_LIST,
+			      &dprx) != 1)
+		return false;
+	return dprx & DP_VSC_SDP_EXT_FOR_COLORIMETRY_SUPPORTED;
+}
+
+static void intel_dp_get_dsc_sink_cap(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	/*
+	 * Clear the cached register set to avoid using stale values
+	 * for the sinks that do not support DSC.
+	 */
+	memset(intel_dp->dsc_dpcd, 0, sizeof(intel_dp->dsc_dpcd));
+
+	/* Clear fec_capable to avoid using stale values */
+	intel_dp->fec_capable = 0;
+
+	/* Cache the DSC DPCD if eDP or DP rev >= 1.4 */
+	if (intel_dp->dpcd[DP_DPCD_REV] >= 0x14 ||
+	    intel_dp->edp_dpcd[0] >= DP_EDP_14) {
+		if (drm_dp_dpcd_read(&intel_dp->aux, DP_DSC_SUPPORT,
+				     intel_dp->dsc_dpcd,
+				     sizeof(intel_dp->dsc_dpcd)) < 0)
+			drm_err(&i915->drm,
+				"Failed to read DPCD register 0x%x\n",
+				DP_DSC_SUPPORT);
+
+		drm_dbg_kms(&i915->drm, "DSC DPCD: %*ph\n",
+			    (int)sizeof(intel_dp->dsc_dpcd),
+			    intel_dp->dsc_dpcd);
+
+		/* FEC is supported only on DP 1.4 */
+		if (!intel_dp_is_edp(intel_dp) &&
+		    drm_dp_dpcd_readb(&intel_dp->aux, DP_FEC_CAPABILITY,
+				      &intel_dp->fec_capable) < 0)
+			drm_err(&i915->drm,
+				"Failed to read FEC DPCD register\n");
+
+		drm_dbg_kms(&i915->drm, "FEC CAPABILITY: %x\n",
+			    intel_dp->fec_capable);
+	}
+}
+
+static void intel_edp_mso_mode_fixup(struct intel_connector *connector,
+				     struct drm_display_mode *mode)
+{
+	struct intel_dp *intel_dp = intel_attached_dp(connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	int n = intel_dp->mso_link_count;
+	int overlap = intel_dp->mso_pixel_overlap;
+
+	if (!mode || !n)
+		return;
+
+	mode->hdisplay = (mode->hdisplay - overlap) * n;
+	mode->hsync_start = (mode->hsync_start - overlap) * n;
+	mode->hsync_end = (mode->hsync_end - overlap) * n;
+	mode->htotal = (mode->htotal - overlap) * n;
+	mode->clock *= n;
+
+	drm_mode_set_name(mode);
+
+	drm_dbg_kms(&i915->drm,
+		    "[CONNECTOR:%d:%s] using generated MSO mode: " DRM_MODE_FMT "\n",
+		    connector->base.base.id, connector->base.name,
+		    DRM_MODE_ARG(mode));
+}
+
+void intel_edp_fixup_vbt_bpp(struct intel_encoder *encoder, int pipe_bpp)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_connector *connector = intel_dp->attached_connector;
+
+	if (connector->panel.vbt.edp.bpp && pipe_bpp > connector->panel.vbt.edp.bpp) {
+		/*
+		 * This is a big fat ugly hack.
+		 *
+		 * Some machines in UEFI boot mode provide us a VBT that has 18
+		 * bpp and 1.62 GHz link bandwidth for eDP, which for reasons
+		 * unknown we fail to light up. Yet the same BIOS boots up with
+		 * 24 bpp and 2.7 GHz link. Use the same bpp as the BIOS uses as
+		 * max, not what it tells us to use.
+		 *
+		 * Note: This will still be broken if the eDP panel is not lit
+		 * up by the BIOS, and thus we can't get the mode at module
+		 * load.
+		 */
+		drm_dbg_kms(&dev_priv->drm,
+			    "pipe has %d bpp for eDP panel, overriding BIOS-provided max %d bpp\n",
+			    pipe_bpp, connector->panel.vbt.edp.bpp);
+		connector->panel.vbt.edp.bpp = pipe_bpp;
+	}
+}
+
+static void intel_edp_mso_init(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct drm_display_info *info = &connector->base.display_info;
+	u8 mso;
+
+	if (intel_dp->edp_dpcd[0] < DP_EDP_14)
+		return;
+
+	if (drm_dp_dpcd_readb(&intel_dp->aux, DP_EDP_MSO_LINK_CAPABILITIES, &mso) != 1) {
+		drm_err(&i915->drm, "Failed to read MSO cap\n");
+		return;
+	}
+
+	/* Valid configurations are SST or MSO 2x1, 2x2, 4x1 */
+	mso &= DP_EDP_MSO_NUMBER_OF_LINKS_MASK;
+	if (mso % 2 || mso > drm_dp_max_lane_count(intel_dp->dpcd)) {
+		drm_err(&i915->drm, "Invalid MSO link count cap %u\n", mso);
+		mso = 0;
+	}
+
+	if (mso) {
+		drm_dbg_kms(&i915->drm, "Sink MSO %ux%u configuration, pixel overlap %u\n",
+			    mso, drm_dp_max_lane_count(intel_dp->dpcd) / mso,
+			    info->mso_pixel_overlap);
+		if (!HAS_MSO(i915)) {
+			drm_err(&i915->drm, "No source MSO support, disabling\n");
+			mso = 0;
+		}
+	}
+
+	intel_dp->mso_link_count = mso;
+	intel_dp->mso_pixel_overlap = mso ? info->mso_pixel_overlap : 0;
+}
+
+static bool
+intel_edp_init_dpcd(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(dp_to_dig_port(intel_dp)->base.base.dev);
+
+	/* this function is meant to be called only once */
+	drm_WARN_ON(&dev_priv->drm, intel_dp->dpcd[DP_DPCD_REV] != 0);
+
+	if (drm_dp_read_dpcd_caps(&intel_dp->aux, intel_dp->dpcd) != 0)
+		return false;
+
+	drm_dp_read_desc(&intel_dp->aux, &intel_dp->desc,
+			 drm_dp_is_branch(intel_dp->dpcd));
+
+	/*
+	 * Read the eDP display control registers.
+	 *
+	 * Do this independent of DP_DPCD_DISPLAY_CONTROL_CAPABLE bit in
+	 * DP_EDP_CONFIGURATION_CAP, because some buggy displays do not have it
+	 * set, but require eDP 1.4+ detection (e.g. for supported link rates
+	 * method). The display control registers should read zero if they're
+	 * not supported anyway.
+	 */
+	if (drm_dp_dpcd_read(&intel_dp->aux, DP_EDP_DPCD_REV,
+			     intel_dp->edp_dpcd, sizeof(intel_dp->edp_dpcd)) ==
+			     sizeof(intel_dp->edp_dpcd)) {
+		drm_dbg_kms(&dev_priv->drm, "eDP DPCD: %*ph\n",
+			    (int)sizeof(intel_dp->edp_dpcd),
+			    intel_dp->edp_dpcd);
+
+		intel_dp->use_max_params = intel_dp->edp_dpcd[0] < DP_EDP_14;
+	}
+
+	/*
+	 * This has to be called after intel_dp->edp_dpcd is filled, PSR checks
+	 * for SET_POWER_CAPABLE bit in intel_dp->edp_dpcd[1]
+	 */
+	intel_psr_init_dpcd(intel_dp);
+
+	/* Clear the default sink rates */
+	intel_dp->num_sink_rates = 0;
+
+	/* Read the eDP 1.4+ supported link rates. */
+	if (intel_dp->edp_dpcd[0] >= DP_EDP_14) {
+		__le16 sink_rates[DP_MAX_SUPPORTED_RATES];
+		int i;
+
+		drm_dp_dpcd_read(&intel_dp->aux, DP_SUPPORTED_LINK_RATES,
+				sink_rates, sizeof(sink_rates));
+
+		for (i = 0; i < ARRAY_SIZE(sink_rates); i++) {
+			int val = le16_to_cpu(sink_rates[i]);
+
+			if (val == 0)
+				break;
+
+			/* Value read multiplied by 200kHz gives the per-lane
+			 * link rate in kHz. The source rates are, however,
+			 * stored in terms of LS_Clk kHz. The full conversion
+			 * back to symbols is
+			 * (val * 200kHz)*(8/10 ch. encoding)*(1/8 bit to Byte)
+			 */
+			intel_dp->sink_rates[i] = (val * 200) / 10;
+		}
+		intel_dp->num_sink_rates = i;
+	}
+
+	/*
+	 * Use DP_LINK_RATE_SET if DP_SUPPORTED_LINK_RATES are available,
+	 * default to DP_MAX_LINK_RATE and DP_LINK_BW_SET otherwise.
+	 */
+	if (intel_dp->num_sink_rates)
+		intel_dp->use_rate_select = true;
+	else
+		intel_dp_set_sink_rates(intel_dp);
+	intel_dp_set_max_sink_lane_count(intel_dp);
+
+	/* Read the eDP DSC DPCD registers */
+	if (DISPLAY_VER(dev_priv) >= 10)
+		intel_dp_get_dsc_sink_cap(intel_dp);
+
+	/*
+	 * If needed, program our source OUI so we can make various Intel-specific AUX services
+	 * available (such as HDR backlight controls)
+	 */
+	intel_edp_init_source_oui(intel_dp, true);
+
+	return true;
+}
+
+static bool
+intel_dp_has_sink_count(struct intel_dp *intel_dp)
+{
+	if (!intel_dp->attached_connector)
+		return false;
+
+	return drm_dp_read_sink_count_cap(&intel_dp->attached_connector->base,
+					  intel_dp->dpcd,
+					  &intel_dp->desc);
+}
+
+static bool
+intel_dp_get_dpcd(struct intel_dp *intel_dp)
+{
+	int ret;
+
+	if (intel_dp_init_lttpr_and_dprx_caps(intel_dp) < 0)
+		return false;
+
+	/*
+	 * Don't clobber cached eDP rates. Also skip re-reading
+	 * the OUI/ID since we know it won't change.
+	 */
+	if (!intel_dp_is_edp(intel_dp)) {
+		drm_dp_read_desc(&intel_dp->aux, &intel_dp->desc,
+				 drm_dp_is_branch(intel_dp->dpcd));
+
+		intel_dp_set_sink_rates(intel_dp);
+		intel_dp_set_max_sink_lane_count(intel_dp);
+		intel_dp_set_common_rates(intel_dp);
+	}
+
+	if (intel_dp_has_sink_count(intel_dp)) {
+		ret = drm_dp_read_sink_count(&intel_dp->aux);
+		if (ret < 0)
+			return false;
+
+		/*
+		 * Sink count can change between short pulse hpd hence
+		 * a member variable in intel_dp will track any changes
+		 * between short pulse interrupts.
+		 */
+		intel_dp->sink_count = ret;
+
+		/*
+		 * SINK_COUNT == 0 and DOWNSTREAM_PORT_PRESENT == 1 implies that
+		 * a dongle is present but no display. Unless we require to know
+		 * if a dongle is present or not, we don't need to update
+		 * downstream port information. So, an early return here saves
+		 * time from performing other operations which are not required.
+		 */
+		if (!intel_dp->sink_count)
+			return false;
+	}
+
+	return drm_dp_read_downstream_info(&intel_dp->aux, intel_dp->dpcd,
+					   intel_dp->downstream_ports) == 0;
+}
+
+static bool
+intel_dp_can_mst(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	return i915->params.enable_dp_mst &&
+		intel_dp_mst_source_support(intel_dp) &&
+		drm_dp_read_mst_cap(&intel_dp->aux, intel_dp->dpcd);
+}
+
+static void
+intel_dp_configure_mst(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_encoder *encoder =
+		&dp_to_dig_port(intel_dp)->base;
+	bool sink_can_mst = drm_dp_read_mst_cap(&intel_dp->aux, intel_dp->dpcd);
+
+	drm_dbg_kms(&i915->drm,
+		    "[ENCODER:%d:%s] MST support: port: %s, sink: %s, modparam: %s\n",
+		    encoder->base.base.id, encoder->base.name,
+		    str_yes_no(intel_dp_mst_source_support(intel_dp)),
+		    str_yes_no(sink_can_mst),
+		    str_yes_no(i915->params.enable_dp_mst));
+
+	if (!intel_dp_mst_source_support(intel_dp))
+		return;
+
+	intel_dp->is_mst = sink_can_mst &&
+		i915->params.enable_dp_mst;
+
+	drm_dp_mst_topology_mgr_set_mst(&intel_dp->mst_mgr,
+					intel_dp->is_mst);
+}
+
+static bool
+intel_dp_get_sink_irq_esi(struct intel_dp *intel_dp, u8 *esi)
+{
+	return drm_dp_dpcd_read(&intel_dp->aux, DP_SINK_COUNT_ESI, esi, 4) == 4;
+}
+
+static bool intel_dp_ack_sink_irq_esi(struct intel_dp *intel_dp, u8 esi[4])
+{
+	int retry;
+
+	for (retry = 0; retry < 3; retry++) {
+		if (drm_dp_dpcd_write(&intel_dp->aux, DP_SINK_COUNT_ESI + 1,
+				      &esi[1], 3) == 3)
+			return true;
+	}
+
+	return false;
+}
+
+bool
+intel_dp_needs_vsc_sdp(const struct intel_crtc_state *crtc_state,
+		       const struct drm_connector_state *conn_state)
+{
+	/*
+	 * As per DP 1.4a spec section 2.2.4.3 [MSA Field for Indication
+	 * of Color Encoding Format and Content Color Gamut], in order to
+	 * sending YCBCR 420 or HDR BT.2020 signals we should use DP VSC SDP.
+	 */
+	if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR420)
+		return true;
+
+	switch (conn_state->colorspace) {
+	case DRM_MODE_COLORIMETRY_SYCC_601:
+	case DRM_MODE_COLORIMETRY_OPYCC_601:
+	case DRM_MODE_COLORIMETRY_BT2020_YCC:
+	case DRM_MODE_COLORIMETRY_BT2020_RGB:
+	case DRM_MODE_COLORIMETRY_BT2020_CYCC:
+		return true;
+	default:
+		break;
+	}
+
+	return false;
+}
+
+static ssize_t intel_dp_vsc_sdp_pack(const struct drm_dp_vsc_sdp *vsc,
+				     struct dp_sdp *sdp, size_t size)
+{
+	size_t length = sizeof(struct dp_sdp);
+
+	if (size < length)
+		return -ENOSPC;
+
+	memset(sdp, 0, size);
+
+	/*
+	 * Prepare VSC Header for SU as per DP 1.4a spec, Table 2-119
+	 * VSC SDP Header Bytes
+	 */
+	sdp->sdp_header.HB0 = 0; /* Secondary-Data Packet ID = 0 */
+	sdp->sdp_header.HB1 = vsc->sdp_type; /* Secondary-data Packet Type */
+	sdp->sdp_header.HB2 = vsc->revision; /* Revision Number */
+	sdp->sdp_header.HB3 = vsc->length; /* Number of Valid Data Bytes */
+
+	/*
+	 * Only revision 0x5 supports Pixel Encoding/Colorimetry Format as
+	 * per DP 1.4a spec.
+	 */
+	if (vsc->revision != 0x5)
+		goto out;
+
+	/* VSC SDP Payload for DB16 through DB18 */
+	/* Pixel Encoding and Colorimetry Formats  */
+	sdp->db[16] = (vsc->pixelformat & 0xf) << 4; /* DB16[7:4] */
+	sdp->db[16] |= vsc->colorimetry & 0xf; /* DB16[3:0] */
+
+	switch (vsc->bpc) {
+	case 6:
+		/* 6bpc: 0x0 */
+		break;
+	case 8:
+		sdp->db[17] = 0x1; /* DB17[3:0] */
+		break;
+	case 10:
+		sdp->db[17] = 0x2;
+		break;
+	case 12:
+		sdp->db[17] = 0x3;
+		break;
+	case 16:
+		sdp->db[17] = 0x4;
+		break;
+	default:
+		MISSING_CASE(vsc->bpc);
+		break;
+	}
+	/* Dynamic Range and Component Bit Depth */
+	if (vsc->dynamic_range == DP_DYNAMIC_RANGE_CTA)
+		sdp->db[17] |= 0x80;  /* DB17[7] */
+
+	/* Content Type */
+	sdp->db[18] = vsc->content_type & 0x7;
+
+out:
+	return length;
+}
+
+static ssize_t
+intel_dp_hdr_metadata_infoframe_sdp_pack(struct drm_i915_private *i915,
+					 const struct hdmi_drm_infoframe *drm_infoframe,
+					 struct dp_sdp *sdp,
+					 size_t size)
+{
+	size_t length = sizeof(struct dp_sdp);
+	const int infoframe_size = HDMI_INFOFRAME_HEADER_SIZE + HDMI_DRM_INFOFRAME_SIZE;
+	unsigned char buf[HDMI_INFOFRAME_HEADER_SIZE + HDMI_DRM_INFOFRAME_SIZE];
+	ssize_t len;
+
+	if (size < length)
+		return -ENOSPC;
+
+	memset(sdp, 0, size);
+
+	len = hdmi_drm_infoframe_pack_only(drm_infoframe, buf, sizeof(buf));
+	if (len < 0) {
+		drm_dbg_kms(&i915->drm, "buffer size is smaller than hdr metadata infoframe\n");
+		return -ENOSPC;
+	}
+
+	if (len != infoframe_size) {
+		drm_dbg_kms(&i915->drm, "wrong static hdr metadata size\n");
+		return -ENOSPC;
+	}
+
+	/*
+	 * Set up the infoframe sdp packet for HDR static metadata.
+	 * Prepare VSC Header for SU as per DP 1.4a spec,
+	 * Table 2-100 and Table 2-101
+	 */
+
+	/* Secondary-Data Packet ID, 00h for non-Audio INFOFRAME */
+	sdp->sdp_header.HB0 = 0;
+	/*
+	 * Packet Type 80h + Non-audio INFOFRAME Type value
+	 * HDMI_INFOFRAME_TYPE_DRM: 0x87
+	 * - 80h + Non-audio INFOFRAME Type value
+	 * - InfoFrame Type: 0x07
+	 *    [CTA-861-G Table-42 Dynamic Range and Mastering InfoFrame]
+	 */
+	sdp->sdp_header.HB1 = drm_infoframe->type;
+	/*
+	 * Least Significant Eight Bits of (Data Byte Count – 1)
+	 * infoframe_size - 1
+	 */
+	sdp->sdp_header.HB2 = 0x1D;
+	/* INFOFRAME SDP Version Number */
+	sdp->sdp_header.HB3 = (0x13 << 2);
+	/* CTA Header Byte 2 (INFOFRAME Version Number) */
+	sdp->db[0] = drm_infoframe->version;
+	/* CTA Header Byte 3 (Length of INFOFRAME): HDMI_DRM_INFOFRAME_SIZE */
+	sdp->db[1] = drm_infoframe->length;
+	/*
+	 * Copy HDMI_DRM_INFOFRAME_SIZE size from a buffer after
+	 * HDMI_INFOFRAME_HEADER_SIZE
+	 */
+	BUILD_BUG_ON(sizeof(sdp->db) < HDMI_DRM_INFOFRAME_SIZE + 2);
+	memcpy(&sdp->db[2], &buf[HDMI_INFOFRAME_HEADER_SIZE],
+	       HDMI_DRM_INFOFRAME_SIZE);
+
+	/*
+	 * Size of DP infoframe sdp packet for HDR static metadata consists of
+	 * - DP SDP Header(struct dp_sdp_header): 4 bytes
+	 * - Two Data Blocks: 2 bytes
+	 *    CTA Header Byte2 (INFOFRAME Version Number)
+	 *    CTA Header Byte3 (Length of INFOFRAME)
+	 * - HDMI_DRM_INFOFRAME_SIZE: 26 bytes
+	 *
+	 * Prior to GEN11's GMP register size is identical to DP HDR static metadata
+	 * infoframe size. But GEN11+ has larger than that size, write_infoframe
+	 * will pad rest of the size.
+	 */
+	return sizeof(struct dp_sdp_header) + 2 + HDMI_DRM_INFOFRAME_SIZE;
+}
+
+static void intel_write_dp_sdp(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state,
+			       unsigned int type)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct dp_sdp sdp = {};
+	ssize_t len;
+
+	if ((crtc_state->infoframes.enable &
+	     intel_hdmi_infoframe_enable(type)) == 0)
+		return;
+
+	switch (type) {
+	case DP_SDP_VSC:
+		len = intel_dp_vsc_sdp_pack(&crtc_state->infoframes.vsc, &sdp,
+					    sizeof(sdp));
+		break;
+	case HDMI_PACKET_TYPE_GAMUT_METADATA:
+		len = intel_dp_hdr_metadata_infoframe_sdp_pack(dev_priv,
+							       &crtc_state->infoframes.drm.drm,
+							       &sdp, sizeof(sdp));
+		break;
+	default:
+		MISSING_CASE(type);
+		return;
+	}
+
+	if (drm_WARN_ON(&dev_priv->drm, len < 0))
+		return;
+
+	dig_port->write_infoframe(encoder, crtc_state, type, &sdp, len);
+}
+
+void intel_write_dp_vsc_sdp(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    const struct drm_dp_vsc_sdp *vsc)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct dp_sdp sdp = {};
+	ssize_t len;
+
+	len = intel_dp_vsc_sdp_pack(vsc, &sdp, sizeof(sdp));
+
+	if (drm_WARN_ON(&dev_priv->drm, len < 0))
+		return;
+
+	dig_port->write_infoframe(encoder, crtc_state, DP_SDP_VSC,
+					&sdp, len);
+}
+
+void intel_dp_set_infoframes(struct intel_encoder *encoder,
+			     bool enable,
+			     const struct intel_crtc_state *crtc_state,
+			     const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	i915_reg_t reg = HSW_TVIDEO_DIP_CTL(crtc_state->cpu_transcoder);
+	u32 dip_enable = VIDEO_DIP_ENABLE_AVI_HSW | VIDEO_DIP_ENABLE_GCP_HSW |
+			 VIDEO_DIP_ENABLE_VS_HSW | VIDEO_DIP_ENABLE_GMP_HSW |
+			 VIDEO_DIP_ENABLE_SPD_HSW | VIDEO_DIP_ENABLE_DRM_GLK;
+	u32 val = intel_de_read(dev_priv, reg) & ~dip_enable;
+
+	/* TODO: Add DSC case (DIP_ENABLE_PPS) */
+	/* When PSR is enabled, this routine doesn't disable VSC DIP */
+	if (!crtc_state->has_psr)
+		val &= ~VIDEO_DIP_ENABLE_VSC_HSW;
+
+	intel_de_write(dev_priv, reg, val);
+	intel_de_posting_read(dev_priv, reg);
+
+	if (!enable)
+		return;
+
+	/* When PSR is enabled, VSC SDP is handled by PSR routine */
+	if (!crtc_state->has_psr)
+		intel_write_dp_sdp(encoder, crtc_state, DP_SDP_VSC);
+
+	intel_write_dp_sdp(encoder, crtc_state, HDMI_PACKET_TYPE_GAMUT_METADATA);
+}
+
+static int intel_dp_vsc_sdp_unpack(struct drm_dp_vsc_sdp *vsc,
+				   const void *buffer, size_t size)
+{
+	const struct dp_sdp *sdp = buffer;
+
+	if (size < sizeof(struct dp_sdp))
+		return -EINVAL;
+
+	memset(vsc, 0, sizeof(*vsc));
+
+	if (sdp->sdp_header.HB0 != 0)
+		return -EINVAL;
+
+	if (sdp->sdp_header.HB1 != DP_SDP_VSC)
+		return -EINVAL;
+
+	vsc->sdp_type = sdp->sdp_header.HB1;
+	vsc->revision = sdp->sdp_header.HB2;
+	vsc->length = sdp->sdp_header.HB3;
+
+	if ((sdp->sdp_header.HB2 == 0x2 && sdp->sdp_header.HB3 == 0x8) ||
+	    (sdp->sdp_header.HB2 == 0x4 && sdp->sdp_header.HB3 == 0xe)) {
+		/*
+		 * - HB2 = 0x2, HB3 = 0x8
+		 *   VSC SDP supporting 3D stereo + PSR
+		 * - HB2 = 0x4, HB3 = 0xe
+		 *   VSC SDP supporting 3D stereo + PSR2 with Y-coordinate of
+		 *   first scan line of the SU region (applies to eDP v1.4b
+		 *   and higher).
+		 */
+		return 0;
+	} else if (sdp->sdp_header.HB2 == 0x5 && sdp->sdp_header.HB3 == 0x13) {
+		/*
+		 * - HB2 = 0x5, HB3 = 0x13
+		 *   VSC SDP supporting 3D stereo + PSR2 + Pixel Encoding/Colorimetry
+		 *   Format.
+		 */
+		vsc->pixelformat = (sdp->db[16] >> 4) & 0xf;
+		vsc->colorimetry = sdp->db[16] & 0xf;
+		vsc->dynamic_range = (sdp->db[17] >> 7) & 0x1;
+
+		switch (sdp->db[17] & 0x7) {
+		case 0x0:
+			vsc->bpc = 6;
+			break;
+		case 0x1:
+			vsc->bpc = 8;
+			break;
+		case 0x2:
+			vsc->bpc = 10;
+			break;
+		case 0x3:
+			vsc->bpc = 12;
+			break;
+		case 0x4:
+			vsc->bpc = 16;
+			break;
+		default:
+			MISSING_CASE(sdp->db[17] & 0x7);
+			return -EINVAL;
+		}
+
+		vsc->content_type = sdp->db[18] & 0x7;
+	} else {
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+intel_dp_hdr_metadata_infoframe_sdp_unpack(struct hdmi_drm_infoframe *drm_infoframe,
+					   const void *buffer, size_t size)
+{
+	int ret;
+
+	const struct dp_sdp *sdp = buffer;
+
+	if (size < sizeof(struct dp_sdp))
+		return -EINVAL;
+
+	if (sdp->sdp_header.HB0 != 0)
+		return -EINVAL;
+
+	if (sdp->sdp_header.HB1 != HDMI_INFOFRAME_TYPE_DRM)
+		return -EINVAL;
+
+	/*
+	 * Least Significant Eight Bits of (Data Byte Count – 1)
+	 * 1Dh (i.e., Data Byte Count = 30 bytes).
+	 */
+	if (sdp->sdp_header.HB2 != 0x1D)
+		return -EINVAL;
+
+	/* Most Significant Two Bits of (Data Byte Count – 1), Clear to 00b. */
+	if ((sdp->sdp_header.HB3 & 0x3) != 0)
+		return -EINVAL;
+
+	/* INFOFRAME SDP Version Number */
+	if (((sdp->sdp_header.HB3 >> 2) & 0x3f) != 0x13)
+		return -EINVAL;
+
+	/* CTA Header Byte 2 (INFOFRAME Version Number) */
+	if (sdp->db[0] != 1)
+		return -EINVAL;
+
+	/* CTA Header Byte 3 (Length of INFOFRAME): HDMI_DRM_INFOFRAME_SIZE */
+	if (sdp->db[1] != HDMI_DRM_INFOFRAME_SIZE)
+		return -EINVAL;
+
+	ret = hdmi_drm_infoframe_unpack_only(drm_infoframe, &sdp->db[2],
+					     HDMI_DRM_INFOFRAME_SIZE);
+
+	return ret;
+}
+
+static void intel_read_dp_vsc_sdp(struct intel_encoder *encoder,
+				  struct intel_crtc_state *crtc_state,
+				  struct drm_dp_vsc_sdp *vsc)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	unsigned int type = DP_SDP_VSC;
+	struct dp_sdp sdp = {};
+	int ret;
+
+	/* When PSR is enabled, VSC SDP is handled by PSR routine */
+	if (crtc_state->has_psr)
+		return;
+
+	if ((crtc_state->infoframes.enable &
+	     intel_hdmi_infoframe_enable(type)) == 0)
+		return;
+
+	dig_port->read_infoframe(encoder, crtc_state, type, &sdp, sizeof(sdp));
+
+	ret = intel_dp_vsc_sdp_unpack(vsc, &sdp, sizeof(sdp));
+
+	if (ret)
+		drm_dbg_kms(&dev_priv->drm, "Failed to unpack DP VSC SDP\n");
+}
+
+static void intel_read_dp_hdr_metadata_infoframe_sdp(struct intel_encoder *encoder,
+						     struct intel_crtc_state *crtc_state,
+						     struct hdmi_drm_infoframe *drm_infoframe)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	unsigned int type = HDMI_PACKET_TYPE_GAMUT_METADATA;
+	struct dp_sdp sdp = {};
+	int ret;
+
+	if ((crtc_state->infoframes.enable &
+	    intel_hdmi_infoframe_enable(type)) == 0)
+		return;
+
+	dig_port->read_infoframe(encoder, crtc_state, type, &sdp,
+				 sizeof(sdp));
+
+	ret = intel_dp_hdr_metadata_infoframe_sdp_unpack(drm_infoframe, &sdp,
+							 sizeof(sdp));
+
+	if (ret)
+		drm_dbg_kms(&dev_priv->drm,
+			    "Failed to unpack DP HDR Metadata Infoframe SDP\n");
+}
+
+void intel_read_dp_sdp(struct intel_encoder *encoder,
+		       struct intel_crtc_state *crtc_state,
+		       unsigned int type)
+{
+	switch (type) {
+	case DP_SDP_VSC:
+		intel_read_dp_vsc_sdp(encoder, crtc_state,
+				      &crtc_state->infoframes.vsc);
+		break;
+	case HDMI_PACKET_TYPE_GAMUT_METADATA:
+		intel_read_dp_hdr_metadata_infoframe_sdp(encoder, crtc_state,
+							 &crtc_state->infoframes.drm.drm);
+		break;
+	default:
+		MISSING_CASE(type);
+		break;
+	}
+}
+
+static u8 intel_dp_autotest_link_training(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int status = 0;
+	int test_link_rate;
+	u8 test_lane_count, test_link_bw;
+	/* (DP CTS 1.2)
+	 * 4.3.1.11
+	 */
+	/* Read the TEST_LANE_COUNT and TEST_LINK_RTAE fields (DP CTS 3.1.4) */
+	status = drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_LANE_COUNT,
+				   &test_lane_count);
+
+	if (status <= 0) {
+		drm_dbg_kms(&i915->drm, "Lane count read failed\n");
+		return DP_TEST_NAK;
+	}
+	test_lane_count &= DP_MAX_LANE_COUNT_MASK;
+
+	status = drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_LINK_RATE,
+				   &test_link_bw);
+	if (status <= 0) {
+		drm_dbg_kms(&i915->drm, "Link Rate read failed\n");
+		return DP_TEST_NAK;
+	}
+	test_link_rate = drm_dp_bw_code_to_link_rate(test_link_bw);
+
+	/* Validate the requested link rate and lane count */
+	if (!intel_dp_link_params_valid(intel_dp, test_link_rate,
+					test_lane_count))
+		return DP_TEST_NAK;
+
+	intel_dp->compliance.test_lane_count = test_lane_count;
+	intel_dp->compliance.test_link_rate = test_link_rate;
+
+	return DP_TEST_ACK;
+}
+
+static u8 intel_dp_autotest_video_pattern(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 test_pattern;
+	u8 test_misc;
+	__be16 h_width, v_height;
+	int status = 0;
+
+	/* Read the TEST_PATTERN (DP CTS 3.1.5) */
+	status = drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_PATTERN,
+				   &test_pattern);
+	if (status <= 0) {
+		drm_dbg_kms(&i915->drm, "Test pattern read failed\n");
+		return DP_TEST_NAK;
+	}
+	if (test_pattern != DP_COLOR_RAMP)
+		return DP_TEST_NAK;
+
+	status = drm_dp_dpcd_read(&intel_dp->aux, DP_TEST_H_WIDTH_HI,
+				  &h_width, 2);
+	if (status <= 0) {
+		drm_dbg_kms(&i915->drm, "H Width read failed\n");
+		return DP_TEST_NAK;
+	}
+
+	status = drm_dp_dpcd_read(&intel_dp->aux, DP_TEST_V_HEIGHT_HI,
+				  &v_height, 2);
+	if (status <= 0) {
+		drm_dbg_kms(&i915->drm, "V Height read failed\n");
+		return DP_TEST_NAK;
+	}
+
+	status = drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_MISC0,
+				   &test_misc);
+	if (status <= 0) {
+		drm_dbg_kms(&i915->drm, "TEST MISC read failed\n");
+		return DP_TEST_NAK;
+	}
+	if ((test_misc & DP_TEST_COLOR_FORMAT_MASK) != DP_COLOR_FORMAT_RGB)
+		return DP_TEST_NAK;
+	if (test_misc & DP_TEST_DYNAMIC_RANGE_CEA)
+		return DP_TEST_NAK;
+	switch (test_misc & DP_TEST_BIT_DEPTH_MASK) {
+	case DP_TEST_BIT_DEPTH_6:
+		intel_dp->compliance.test_data.bpc = 6;
+		break;
+	case DP_TEST_BIT_DEPTH_8:
+		intel_dp->compliance.test_data.bpc = 8;
+		break;
+	default:
+		return DP_TEST_NAK;
+	}
+
+	intel_dp->compliance.test_data.video_pattern = test_pattern;
+	intel_dp->compliance.test_data.hdisplay = be16_to_cpu(h_width);
+	intel_dp->compliance.test_data.vdisplay = be16_to_cpu(v_height);
+	/* Set test active flag here so userspace doesn't interrupt things */
+	intel_dp->compliance.test_active = true;
+
+	return DP_TEST_ACK;
+}
+
+static u8 intel_dp_autotest_edid(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 test_result = DP_TEST_ACK;
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
+	struct drm_connector *connector = &intel_connector->base;
+
+	if (intel_connector->detect_edid == NULL ||
+	    connector->edid_corrupt ||
+	    intel_dp->aux.i2c_defer_count > 6) {
+		/* Check EDID read for NACKs, DEFERs and corruption
+		 * (DP CTS 1.2 Core r1.1)
+		 *    4.2.2.4 : Failed EDID read, I2C_NAK
+		 *    4.2.2.5 : Failed EDID read, I2C_DEFER
+		 *    4.2.2.6 : EDID corruption detected
+		 * Use failsafe mode for all cases
+		 */
+		if (intel_dp->aux.i2c_nack_count > 0 ||
+			intel_dp->aux.i2c_defer_count > 0)
+			drm_dbg_kms(&i915->drm,
+				    "EDID read had %d NACKs, %d DEFERs\n",
+				    intel_dp->aux.i2c_nack_count,
+				    intel_dp->aux.i2c_defer_count);
+		intel_dp->compliance.test_data.edid = INTEL_DP_RESOLUTION_FAILSAFE;
+	} else {
+		struct edid *block = intel_connector->detect_edid;
+
+		/* We have to write the checksum
+		 * of the last block read
+		 */
+		block += intel_connector->detect_edid->extensions;
+
+		if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_EDID_CHECKSUM,
+				       block->checksum) <= 0)
+			drm_dbg_kms(&i915->drm,
+				    "Failed to write EDID checksum\n");
+
+		test_result = DP_TEST_ACK | DP_TEST_EDID_CHECKSUM_WRITE;
+		intel_dp->compliance.test_data.edid = INTEL_DP_RESOLUTION_PREFERRED;
+	}
+
+	/* Set test active flag here so userspace doesn't interrupt things */
+	intel_dp->compliance.test_active = true;
+
+	return test_result;
+}
+
+static void intel_dp_phy_pattern_update(struct intel_dp *intel_dp,
+					const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv =
+			to_i915(dp_to_dig_port(intel_dp)->base.base.dev);
+	struct drm_dp_phy_test_params *data =
+			&intel_dp->compliance.test_data.phytest;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum pipe pipe = crtc->pipe;
+	u32 pattern_val;
+
+	switch (data->phy_pattern) {
+	case DP_PHY_TEST_PATTERN_NONE:
+		drm_dbg_kms(&dev_priv->drm, "Disable Phy Test Pattern\n");
+		intel_de_write(dev_priv, DDI_DP_COMP_CTL(pipe), 0x0);
+		break;
+	case DP_PHY_TEST_PATTERN_D10_2:
+		drm_dbg_kms(&dev_priv->drm, "Set D10.2 Phy Test Pattern\n");
+		intel_de_write(dev_priv, DDI_DP_COMP_CTL(pipe),
+			       DDI_DP_COMP_CTL_ENABLE | DDI_DP_COMP_CTL_D10_2);
+		break;
+	case DP_PHY_TEST_PATTERN_ERROR_COUNT:
+		drm_dbg_kms(&dev_priv->drm, "Set Error Count Phy Test Pattern\n");
+		intel_de_write(dev_priv, DDI_DP_COMP_CTL(pipe),
+			       DDI_DP_COMP_CTL_ENABLE |
+			       DDI_DP_COMP_CTL_SCRAMBLED_0);
+		break;
+	case DP_PHY_TEST_PATTERN_PRBS7:
+		drm_dbg_kms(&dev_priv->drm, "Set PRBS7 Phy Test Pattern\n");
+		intel_de_write(dev_priv, DDI_DP_COMP_CTL(pipe),
+			       DDI_DP_COMP_CTL_ENABLE | DDI_DP_COMP_CTL_PRBS7);
+		break;
+	case DP_PHY_TEST_PATTERN_80BIT_CUSTOM:
+		/*
+		 * FIXME: Ideally pattern should come from DPCD 0x250. As
+		 * current firmware of DPR-100 could not set it, so hardcoding
+		 * now for complaince test.
+		 */
+		drm_dbg_kms(&dev_priv->drm,
+			    "Set 80Bit Custom Phy Test Pattern 0x3e0f83e0 0x0f83e0f8 0x0000f83e\n");
+		pattern_val = 0x3e0f83e0;
+		intel_de_write(dev_priv, DDI_DP_COMP_PAT(pipe, 0), pattern_val);
+		pattern_val = 0x0f83e0f8;
+		intel_de_write(dev_priv, DDI_DP_COMP_PAT(pipe, 1), pattern_val);
+		pattern_val = 0x0000f83e;
+		intel_de_write(dev_priv, DDI_DP_COMP_PAT(pipe, 2), pattern_val);
+		intel_de_write(dev_priv, DDI_DP_COMP_CTL(pipe),
+			       DDI_DP_COMP_CTL_ENABLE |
+			       DDI_DP_COMP_CTL_CUSTOM80);
+		break;
+	case DP_PHY_TEST_PATTERN_CP2520:
+		/*
+		 * FIXME: Ideally pattern should come from DPCD 0x24A. As
+		 * current firmware of DPR-100 could not set it, so hardcoding
+		 * now for complaince test.
+		 */
+		drm_dbg_kms(&dev_priv->drm, "Set HBR2 compliance Phy Test Pattern\n");
+		pattern_val = 0xFB;
+		intel_de_write(dev_priv, DDI_DP_COMP_CTL(pipe),
+			       DDI_DP_COMP_CTL_ENABLE | DDI_DP_COMP_CTL_HBR2 |
+			       pattern_val);
+		break;
+	default:
+		WARN(1, "Invalid Phy Test Pattern\n");
+	}
+}
+
+static void intel_dp_process_phy_request(struct intel_dp *intel_dp,
+					 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct drm_dp_phy_test_params *data =
+		&intel_dp->compliance.test_data.phytest;
+	u8 link_status[DP_LINK_STATUS_SIZE];
+
+	if (drm_dp_dpcd_read_phy_link_status(&intel_dp->aux, DP_PHY_DPRX,
+					     link_status) < 0) {
+		drm_dbg_kms(&i915->drm, "failed to get link status\n");
+		return;
+	}
+
+	/* retrieve vswing & pre-emphasis setting */
+	intel_dp_get_adjust_train(intel_dp, crtc_state, DP_PHY_DPRX,
+				  link_status);
+
+	intel_dp_set_signal_levels(intel_dp, crtc_state, DP_PHY_DPRX);
+
+	intel_dp_phy_pattern_update(intel_dp, crtc_state);
+
+	drm_dp_dpcd_write(&intel_dp->aux, DP_TRAINING_LANE0_SET,
+			  intel_dp->train_set, crtc_state->lane_count);
+
+	drm_dp_set_phy_test_pattern(&intel_dp->aux, data,
+				    intel_dp->dpcd[DP_DPCD_REV]);
+}
+
+static u8 intel_dp_autotest_phy_pattern(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct drm_dp_phy_test_params *data =
+		&intel_dp->compliance.test_data.phytest;
+
+	if (drm_dp_get_phy_test_pattern(&intel_dp->aux, data)) {
+		drm_dbg_kms(&i915->drm, "DP Phy Test pattern AUX read failure\n");
+		return DP_TEST_NAK;
+	}
+
+	/* Set test active flag here so userspace doesn't interrupt things */
+	intel_dp->compliance.test_active = true;
+
+	return DP_TEST_ACK;
+}
+
+static void intel_dp_handle_test_request(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 response = DP_TEST_NAK;
+	u8 request = 0;
+	int status;
+
+	status = drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_REQUEST, &request);
+	if (status <= 0) {
+		drm_dbg_kms(&i915->drm,
+			    "Could not read test request from sink\n");
+		goto update_status;
+	}
+
+	switch (request) {
+	case DP_TEST_LINK_TRAINING:
+		drm_dbg_kms(&i915->drm, "LINK_TRAINING test requested\n");
+		response = intel_dp_autotest_link_training(intel_dp);
+		break;
+	case DP_TEST_LINK_VIDEO_PATTERN:
+		drm_dbg_kms(&i915->drm, "TEST_PATTERN test requested\n");
+		response = intel_dp_autotest_video_pattern(intel_dp);
+		break;
+	case DP_TEST_LINK_EDID_READ:
+		drm_dbg_kms(&i915->drm, "EDID test requested\n");
+		response = intel_dp_autotest_edid(intel_dp);
+		break;
+	case DP_TEST_LINK_PHY_TEST_PATTERN:
+		drm_dbg_kms(&i915->drm, "PHY_PATTERN test requested\n");
+		response = intel_dp_autotest_phy_pattern(intel_dp);
+		break;
+	default:
+		drm_dbg_kms(&i915->drm, "Invalid test request '%02x'\n",
+			    request);
+		break;
+	}
+
+	if (response & DP_TEST_ACK)
+		intel_dp->compliance.test_type = request;
+
+update_status:
+	status = drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_RESPONSE, response);
+	if (status <= 0)
+		drm_dbg_kms(&i915->drm,
+			    "Could not write test response to sink\n");
+}
+
+static bool intel_dp_link_ok(struct intel_dp *intel_dp,
+			     u8 link_status[DP_LINK_STATUS_SIZE])
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	bool uhbr = intel_dp->link_rate >= 1000000;
+	bool ok;
+
+	if (uhbr)
+		ok = drm_dp_128b132b_lane_channel_eq_done(link_status,
+							  intel_dp->lane_count);
+	else
+		ok = drm_dp_channel_eq_ok(link_status, intel_dp->lane_count);
+
+	if (ok)
+		return true;
+
+	intel_dp_dump_link_status(intel_dp, DP_PHY_DPRX, link_status);
+	drm_dbg_kms(&i915->drm,
+		    "[ENCODER:%d:%s] %s link not ok, retraining\n",
+		    encoder->base.base.id, encoder->base.name,
+		    uhbr ? "128b/132b" : "8b/10b");
+
+	return false;
+}
+
+static void
+intel_dp_mst_hpd_irq(struct intel_dp *intel_dp, u8 *esi, u8 *ack)
+{
+	bool handled = false;
+
+	drm_dp_mst_hpd_irq_handle_event(&intel_dp->mst_mgr, esi, ack, &handled);
+
+	if (esi[1] & DP_CP_IRQ) {
+		intel_hdcp_handle_cp_irq(intel_dp->attached_connector);
+		ack[1] |= DP_CP_IRQ;
+	}
+}
+
+static bool intel_dp_mst_link_status(struct intel_dp *intel_dp)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	u8 link_status[DP_LINK_STATUS_SIZE] = {};
+	const size_t esi_link_status_size = DP_LINK_STATUS_SIZE - 2;
+
+	if (drm_dp_dpcd_read(&intel_dp->aux, DP_LANE0_1_STATUS_ESI, link_status,
+			     esi_link_status_size) != esi_link_status_size) {
+		drm_err(&i915->drm,
+			"[ENCODER:%d:%s] Failed to read link status\n",
+			encoder->base.base.id, encoder->base.name);
+		return false;
+	}
+
+	return intel_dp_link_ok(intel_dp, link_status);
+}
+
+/**
+ * intel_dp_check_mst_status - service any pending MST interrupts, check link status
+ * @intel_dp: Intel DP struct
+ *
+ * Read any pending MST interrupts, call MST core to handle these and ack the
+ * interrupts. Check if the main and AUX link state is ok.
+ *
+ * Returns:
+ * - %true if pending interrupts were serviced (or no interrupts were
+ *   pending) w/o detecting an error condition.
+ * - %false if an error condition - like AUX failure or a loss of link - is
+ *   detected, which needs servicing from the hotplug work.
+ */
+static bool
+intel_dp_check_mst_status(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	bool link_ok = true;
+
+	drm_WARN_ON_ONCE(&i915->drm, intel_dp->active_mst_links < 0);
+
+	for (;;) {
+		u8 esi[4] = {};
+		u8 ack[4] = {};
+
+		if (!intel_dp_get_sink_irq_esi(intel_dp, esi)) {
+			drm_dbg_kms(&i915->drm,
+				    "failed to get ESI - device may have failed\n");
+			link_ok = false;
+
+			break;
+		}
+
+		drm_dbg_kms(&i915->drm, "DPRX ESI: %4ph\n", esi);
+
+		if (intel_dp->active_mst_links > 0 && link_ok &&
+		    esi[3] & LINK_STATUS_CHANGED) {
+			if (!intel_dp_mst_link_status(intel_dp))
+				link_ok = false;
+			ack[3] |= LINK_STATUS_CHANGED;
+		}
+
+		intel_dp_mst_hpd_irq(intel_dp, esi, ack);
+
+		if (!memchr_inv(ack, 0, sizeof(ack)))
+			break;
+
+		if (!intel_dp_ack_sink_irq_esi(intel_dp, ack))
+			drm_dbg_kms(&i915->drm, "Failed to ack ESI\n");
+
+		if (ack[1] & (DP_DOWN_REP_MSG_RDY | DP_UP_REQ_MSG_RDY))
+			drm_dp_mst_hpd_irq_send_new_request(&intel_dp->mst_mgr);
+	}
+
+	return link_ok;
+}
+
+static void
+intel_dp_handle_hdmi_link_status_change(struct intel_dp *intel_dp)
+{
+	bool is_active;
+	u8 buf = 0;
+
+	is_active = drm_dp_pcon_hdmi_link_active(&intel_dp->aux);
+	if (intel_dp->frl.is_trained && !is_active) {
+		if (drm_dp_dpcd_readb(&intel_dp->aux, DP_PCON_HDMI_LINK_CONFIG_1, &buf) < 0)
+			return;
+
+		buf &=  ~DP_PCON_ENABLE_HDMI_LINK;
+		if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_PCON_HDMI_LINK_CONFIG_1, buf) < 0)
+			return;
+
+		drm_dp_pcon_hdmi_frl_link_error_count(&intel_dp->aux, &intel_dp->attached_connector->base);
+
+		intel_dp->frl.is_trained = false;
+
+		/* Restart FRL training or fall back to TMDS mode */
+		intel_dp_check_frl_training(intel_dp);
+	}
+}
+
+static bool
+intel_dp_needs_link_retrain(struct intel_dp *intel_dp)
+{
+	u8 link_status[DP_LINK_STATUS_SIZE];
+
+	if (!intel_dp->link_trained)
+		return false;
+
+	/*
+	 * While PSR source HW is enabled, it will control main-link sending
+	 * frames, enabling and disabling it so trying to do a retrain will fail
+	 * as the link would or not be on or it could mix training patterns
+	 * and frame data at the same time causing retrain to fail.
+	 * Also when exiting PSR, HW will retrain the link anyways fixing
+	 * any link status error.
+	 */
+	if (intel_psr_enabled(intel_dp))
+		return false;
+
+	if (drm_dp_dpcd_read_phy_link_status(&intel_dp->aux, DP_PHY_DPRX,
+					     link_status) < 0)
+		return false;
+
+	/*
+	 * Validate the cached values of intel_dp->link_rate and
+	 * intel_dp->lane_count before attempting to retrain.
+	 *
+	 * FIXME would be nice to user the crtc state here, but since
+	 * we need to call this from the short HPD handler that seems
+	 * a bit hard.
+	 */
+	if (!intel_dp_link_params_valid(intel_dp, intel_dp->link_rate,
+					intel_dp->lane_count))
+		return false;
+
+	/* Retrain if link not ok */
+	return !intel_dp_link_ok(intel_dp, link_status);
+}
+
+static bool intel_dp_has_connector(struct intel_dp *intel_dp,
+				   const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_encoder *encoder;
+	enum pipe pipe;
+
+	if (!conn_state->best_encoder)
+		return false;
+
+	/* SST */
+	encoder = &dp_to_dig_port(intel_dp)->base;
+	if (conn_state->best_encoder == &encoder->base)
+		return true;
+
+	/* MST */
+	for_each_pipe(i915, pipe) {
+		encoder = &intel_dp->mst_encoders[pipe]->base;
+		if (conn_state->best_encoder == &encoder->base)
+			return true;
+	}
+
+	return false;
+}
+
+static int intel_dp_prep_link_retrain(struct intel_dp *intel_dp,
+				      struct drm_modeset_acquire_ctx *ctx,
+				      u8 *pipe_mask)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct drm_connector_list_iter conn_iter;
+	struct intel_connector *connector;
+	int ret = 0;
+
+	*pipe_mask = 0;
+
+	if (!intel_dp_needs_link_retrain(intel_dp))
+		return 0;
+
+	drm_connector_list_iter_begin(&i915->drm, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		struct drm_connector_state *conn_state =
+			connector->base.state;
+		struct intel_crtc_state *crtc_state;
+		struct intel_crtc *crtc;
+
+		if (!intel_dp_has_connector(intel_dp, conn_state))
+			continue;
+
+		crtc = to_intel_crtc(conn_state->crtc);
+		if (!crtc)
+			continue;
+
+		ret = drm_modeset_lock(&crtc->base.mutex, ctx);
+		if (ret)
+			break;
+
+		crtc_state = to_intel_crtc_state(crtc->base.state);
+
+		drm_WARN_ON(&i915->drm, !intel_crtc_has_dp_encoder(crtc_state));
+
+		if (!crtc_state->hw.active)
+			continue;
+
+		if (conn_state->commit &&
+		    !try_wait_for_completion(&conn_state->commit->hw_done))
+			continue;
+
+		*pipe_mask |= BIT(crtc->pipe);
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	if (!intel_dp_needs_link_retrain(intel_dp))
+		*pipe_mask = 0;
+
+	return ret;
+}
+
+static bool intel_dp_is_connected(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+
+	return connector->base.status == connector_status_connected ||
+		intel_dp->is_mst;
+}
+
+int intel_dp_retrain_link(struct intel_encoder *encoder,
+			  struct drm_modeset_acquire_ctx *ctx)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_crtc *crtc;
+	u8 pipe_mask;
+	int ret;
+
+	if (!intel_dp_is_connected(intel_dp))
+		return 0;
+
+	ret = drm_modeset_lock(&dev_priv->drm.mode_config.connection_mutex,
+			       ctx);
+	if (ret)
+		return ret;
+
+	ret = intel_dp_prep_link_retrain(intel_dp, ctx, &pipe_mask);
+	if (ret)
+		return ret;
+
+	if (pipe_mask == 0)
+		return 0;
+
+	drm_dbg_kms(&dev_priv->drm, "[ENCODER:%d:%s] retraining link\n",
+		    encoder->base.base.id, encoder->base.name);
+
+	for_each_intel_crtc_in_pipe_mask(&dev_priv->drm, crtc, pipe_mask) {
+		const struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		/* Suppress underruns caused by re-training */
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, crtc->pipe, false);
+		if (crtc_state->has_pch_encoder)
+			intel_set_pch_fifo_underrun_reporting(dev_priv,
+							      intel_crtc_pch_transcoder(crtc), false);
+	}
+
+	for_each_intel_crtc_in_pipe_mask(&dev_priv->drm, crtc, pipe_mask) {
+		const struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		/* retrain on the MST master transcoder */
+		if (DISPLAY_VER(dev_priv) >= 12 &&
+		    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST) &&
+		    !intel_dp_mst_is_master_trans(crtc_state))
+			continue;
+
+		intel_dp_check_frl_training(intel_dp);
+		intel_dp_pcon_dsc_configure(intel_dp, crtc_state);
+		intel_dp_start_link_train(intel_dp, crtc_state);
+		intel_dp_stop_link_train(intel_dp, crtc_state);
+		break;
+	}
+
+	for_each_intel_crtc_in_pipe_mask(&dev_priv->drm, crtc, pipe_mask) {
+		const struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		/* Keep underrun reporting disabled until things are stable */
+		intel_crtc_wait_for_next_vblank(crtc);
+
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, crtc->pipe, true);
+		if (crtc_state->has_pch_encoder)
+			intel_set_pch_fifo_underrun_reporting(dev_priv,
+							      intel_crtc_pch_transcoder(crtc), true);
+	}
+
+	return 0;
+}
+
+static int intel_dp_prep_phy_test(struct intel_dp *intel_dp,
+				  struct drm_modeset_acquire_ctx *ctx,
+				  u8 *pipe_mask)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct drm_connector_list_iter conn_iter;
+	struct intel_connector *connector;
+	int ret = 0;
+
+	*pipe_mask = 0;
+
+	drm_connector_list_iter_begin(&i915->drm, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		struct drm_connector_state *conn_state =
+			connector->base.state;
+		struct intel_crtc_state *crtc_state;
+		struct intel_crtc *crtc;
+
+		if (!intel_dp_has_connector(intel_dp, conn_state))
+			continue;
+
+		crtc = to_intel_crtc(conn_state->crtc);
+		if (!crtc)
+			continue;
+
+		ret = drm_modeset_lock(&crtc->base.mutex, ctx);
+		if (ret)
+			break;
+
+		crtc_state = to_intel_crtc_state(crtc->base.state);
+
+		drm_WARN_ON(&i915->drm, !intel_crtc_has_dp_encoder(crtc_state));
+
+		if (!crtc_state->hw.active)
+			continue;
+
+		if (conn_state->commit &&
+		    !try_wait_for_completion(&conn_state->commit->hw_done))
+			continue;
+
+		*pipe_mask |= BIT(crtc->pipe);
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	return ret;
+}
+
+static int intel_dp_do_phy_test(struct intel_encoder *encoder,
+				struct drm_modeset_acquire_ctx *ctx)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_crtc *crtc;
+	u8 pipe_mask;
+	int ret;
+
+	ret = drm_modeset_lock(&dev_priv->drm.mode_config.connection_mutex,
+			       ctx);
+	if (ret)
+		return ret;
+
+	ret = intel_dp_prep_phy_test(intel_dp, ctx, &pipe_mask);
+	if (ret)
+		return ret;
+
+	if (pipe_mask == 0)
+		return 0;
+
+	drm_dbg_kms(&dev_priv->drm, "[ENCODER:%d:%s] PHY test\n",
+		    encoder->base.base.id, encoder->base.name);
+
+	for_each_intel_crtc_in_pipe_mask(&dev_priv->drm, crtc, pipe_mask) {
+		const struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		/* test on the MST master transcoder */
+		if (DISPLAY_VER(dev_priv) >= 12 &&
+		    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST) &&
+		    !intel_dp_mst_is_master_trans(crtc_state))
+			continue;
+
+		intel_dp_process_phy_request(intel_dp, crtc_state);
+		break;
+	}
+
+	return 0;
+}
+
+void intel_dp_phy_test(struct intel_encoder *encoder)
+{
+	struct drm_modeset_acquire_ctx ctx;
+	int ret;
+
+	drm_modeset_acquire_init(&ctx, 0);
+
+	for (;;) {
+		ret = intel_dp_do_phy_test(encoder, &ctx);
+
+		if (ret == -EDEADLK) {
+			drm_modeset_backoff(&ctx);
+			continue;
+		}
+
+		break;
+	}
+
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+	drm_WARN(encoder->base.dev, ret,
+		 "Acquiring modeset locks failed with %i\n", ret);
+}
+
+static void intel_dp_check_device_service_irq(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 val;
+
+	if (intel_dp->dpcd[DP_DPCD_REV] < 0x11)
+		return;
+
+	if (drm_dp_dpcd_readb(&intel_dp->aux,
+			      DP_DEVICE_SERVICE_IRQ_VECTOR, &val) != 1 || !val)
+		return;
+
+	drm_dp_dpcd_writeb(&intel_dp->aux, DP_DEVICE_SERVICE_IRQ_VECTOR, val);
+
+	if (val & DP_AUTOMATED_TEST_REQUEST)
+		intel_dp_handle_test_request(intel_dp);
+
+	if (val & DP_CP_IRQ)
+		intel_hdcp_handle_cp_irq(intel_dp->attached_connector);
+
+	if (val & DP_SINK_SPECIFIC_IRQ)
+		drm_dbg_kms(&i915->drm, "Sink specific irq unhandled\n");
+}
+
+static void intel_dp_check_link_service_irq(struct intel_dp *intel_dp)
+{
+	u8 val;
+
+	if (intel_dp->dpcd[DP_DPCD_REV] < 0x11)
+		return;
+
+	if (drm_dp_dpcd_readb(&intel_dp->aux,
+			      DP_LINK_SERVICE_IRQ_VECTOR_ESI0, &val) != 1 || !val)
+		return;
+
+	if (drm_dp_dpcd_writeb(&intel_dp->aux,
+			       DP_LINK_SERVICE_IRQ_VECTOR_ESI0, val) != 1)
+		return;
+
+	if (val & HDMI_LINK_STATUS_CHANGED)
+		intel_dp_handle_hdmi_link_status_change(intel_dp);
+}
+
+/*
+ * According to DP spec
+ * 5.1.2:
+ *  1. Read DPCD
+ *  2. Configure link according to Receiver Capabilities
+ *  3. Use Link Training from 2.5.3.3 and 3.5.1.3
+ *  4. Check link status on receipt of hot-plug interrupt
+ *
+ * intel_dp_short_pulse -  handles short pulse interrupts
+ * when full detection is not required.
+ * Returns %true if short pulse is handled and full detection
+ * is NOT required and %false otherwise.
+ */
+static bool
+intel_dp_short_pulse(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u8 old_sink_count = intel_dp->sink_count;
+	bool ret;
+
+	/*
+	 * Clearing compliance test variables to allow capturing
+	 * of values for next automated test request.
+	 */
+	memset(&intel_dp->compliance, 0, sizeof(intel_dp->compliance));
+
+	/*
+	 * Now read the DPCD to see if it's actually running
+	 * If the current value of sink count doesn't match with
+	 * the value that was stored earlier or dpcd read failed
+	 * we need to do full detection
+	 */
+	ret = intel_dp_get_dpcd(intel_dp);
+
+	if ((old_sink_count != intel_dp->sink_count) || !ret) {
+		/* No need to proceed if we are going to do full detect */
+		return false;
+	}
+
+	intel_dp_check_device_service_irq(intel_dp);
+	intel_dp_check_link_service_irq(intel_dp);
+
+	/* Handle CEC interrupts, if any */
+	drm_dp_cec_irq(&intel_dp->aux);
+
+	/* defer to the hotplug work for link retraining if needed */
+	if (intel_dp_needs_link_retrain(intel_dp))
+		return false;
+
+	intel_psr_short_pulse(intel_dp);
+
+	switch (intel_dp->compliance.test_type) {
+	case DP_TEST_LINK_TRAINING:
+		drm_dbg_kms(&dev_priv->drm,
+			    "Link Training Compliance Test requested\n");
+		/* Send a Hotplug Uevent to userspace to start modeset */
+		drm_kms_helper_hotplug_event(&dev_priv->drm);
+		break;
+	case DP_TEST_LINK_PHY_TEST_PATTERN:
+		drm_dbg_kms(&dev_priv->drm,
+			    "PHY test pattern Compliance Test requested\n");
+		/*
+		 * Schedule long hpd to do the test
+		 *
+		 * FIXME get rid of the ad-hoc phy test modeset code
+		 * and properly incorporate it into the normal modeset.
+		 */
+		return false;
+	}
+
+	return true;
+}
+
+/* XXX this is probably wrong for multiple downstream ports */
+static enum drm_connector_status
+intel_dp_detect_dpcd(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	u8 *dpcd = intel_dp->dpcd;
+	u8 type;
+
+	if (drm_WARN_ON(&i915->drm, intel_dp_is_edp(intel_dp)))
+		return connector_status_connected;
+
+	lspcon_resume(dig_port);
+
+	if (!intel_dp_get_dpcd(intel_dp))
+		return connector_status_disconnected;
+
+	/* if there's no downstream port, we're done */
+	if (!drm_dp_is_branch(dpcd))
+		return connector_status_connected;
+
+	/* If we're HPD-aware, SINK_COUNT changes dynamically */
+	if (intel_dp_has_sink_count(intel_dp) &&
+	    intel_dp->downstream_ports[0] & DP_DS_PORT_HPD) {
+		return intel_dp->sink_count ?
+		connector_status_connected : connector_status_disconnected;
+	}
+
+	if (intel_dp_can_mst(intel_dp))
+		return connector_status_connected;
+
+	/* If no HPD, poke DDC gently */
+	if (drm_probe_ddc(&intel_dp->aux.ddc))
+		return connector_status_connected;
+
+	/* Well we tried, say unknown for unreliable port types */
+	if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11) {
+		type = intel_dp->downstream_ports[0] & DP_DS_PORT_TYPE_MASK;
+		if (type == DP_DS_PORT_TYPE_VGA ||
+		    type == DP_DS_PORT_TYPE_NON_EDID)
+			return connector_status_unknown;
+	} else {
+		type = intel_dp->dpcd[DP_DOWNSTREAMPORT_PRESENT] &
+			DP_DWN_STRM_PORT_TYPE_MASK;
+		if (type == DP_DWN_STRM_PORT_TYPE_ANALOG ||
+		    type == DP_DWN_STRM_PORT_TYPE_OTHER)
+			return connector_status_unknown;
+	}
+
+	/* Anything else is out of spec, warn and ignore */
+	drm_dbg_kms(&i915->drm, "Broken DP branch device, ignoring\n");
+	return connector_status_disconnected;
+}
+
+static enum drm_connector_status
+edp_detect(struct intel_dp *intel_dp)
+{
+	return connector_status_connected;
+}
+
+/*
+ * intel_digital_port_connected - is the specified port connected?
+ * @encoder: intel_encoder
+ *
+ * In cases where there's a connector physically connected but it can't be used
+ * by our hardware we also return false, since the rest of the driver should
+ * pretty much treat the port as disconnected. This is relevant for type-C
+ * (starting on ICL) where there's ownership involved.
+ *
+ * Return %true if port is connected, %false otherwise.
+ */
+bool intel_digital_port_connected(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	bool is_connected = false;
+	intel_wakeref_t wakeref;
+
+	with_intel_display_power(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref)
+		is_connected = dig_port->connected(encoder);
+
+	return is_connected;
+}
+
+static struct edid *
+intel_dp_get_edid(struct intel_dp *intel_dp)
+{
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
+
+	/* use cached edid if we have one */
+	if (intel_connector->edid) {
+		/* invalid edid */
+		if (IS_ERR(intel_connector->edid))
+			return NULL;
+
+		return drm_edid_duplicate(intel_connector->edid);
+	} else
+		return drm_get_edid(&intel_connector->base,
+				    &intel_dp->aux.ddc);
+}
+
+static void
+intel_dp_update_dfp(struct intel_dp *intel_dp,
+		    const struct edid *edid)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_connector *connector = intel_dp->attached_connector;
+
+	intel_dp->dfp.max_bpc =
+		drm_dp_downstream_max_bpc(intel_dp->dpcd,
+					  intel_dp->downstream_ports, edid);
+
+	intel_dp->dfp.max_dotclock =
+		drm_dp_downstream_max_dotclock(intel_dp->dpcd,
+					       intel_dp->downstream_ports);
+
+	intel_dp->dfp.min_tmds_clock =
+		drm_dp_downstream_min_tmds_clock(intel_dp->dpcd,
+						 intel_dp->downstream_ports,
+						 edid);
+	intel_dp->dfp.max_tmds_clock =
+		drm_dp_downstream_max_tmds_clock(intel_dp->dpcd,
+						 intel_dp->downstream_ports,
+						 edid);
+
+	intel_dp->dfp.pcon_max_frl_bw =
+		drm_dp_get_pcon_max_frl_bw(intel_dp->dpcd,
+					   intel_dp->downstream_ports);
+
+	drm_dbg_kms(&i915->drm,
+		    "[CONNECTOR:%d:%s] DFP max bpc %d, max dotclock %d, TMDS clock %d-%d, PCON Max FRL BW %dGbps\n",
+		    connector->base.base.id, connector->base.name,
+		    intel_dp->dfp.max_bpc,
+		    intel_dp->dfp.max_dotclock,
+		    intel_dp->dfp.min_tmds_clock,
+		    intel_dp->dfp.max_tmds_clock,
+		    intel_dp->dfp.pcon_max_frl_bw);
+
+	intel_dp_get_pcon_dsc_cap(intel_dp);
+}
+
+static void
+intel_dp_update_420(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_connector *connector = intel_dp->attached_connector;
+	bool is_branch, ycbcr_420_passthrough, ycbcr_444_to_420, rgb_to_ycbcr;
+
+	/* No YCbCr output support on gmch platforms */
+	if (HAS_GMCH(i915))
+		return;
+
+	/*
+	 * ILK doesn't seem capable of DP YCbCr output. The
+	 * displayed image is severly corrupted. SNB+ is fine.
+	 */
+	if (IS_IRONLAKE(i915))
+		return;
+
+	is_branch = drm_dp_is_branch(intel_dp->dpcd);
+	ycbcr_420_passthrough =
+		drm_dp_downstream_420_passthrough(intel_dp->dpcd,
+						  intel_dp->downstream_ports);
+	/* on-board LSPCON always assumed to support 4:4:4->4:2:0 conversion */
+	ycbcr_444_to_420 =
+		dp_to_dig_port(intel_dp)->lspcon.active ||
+		drm_dp_downstream_444_to_420_conversion(intel_dp->dpcd,
+							intel_dp->downstream_ports);
+	rgb_to_ycbcr = drm_dp_downstream_rgb_to_ycbcr_conversion(intel_dp->dpcd,
+								 intel_dp->downstream_ports,
+								 DP_DS_HDMI_BT709_RGB_YCBCR_CONV);
+
+	if (DISPLAY_VER(i915) >= 11) {
+		/* Let PCON convert from RGB->YCbCr if possible */
+		if (is_branch && rgb_to_ycbcr && ycbcr_444_to_420) {
+			intel_dp->dfp.rgb_to_ycbcr = true;
+			intel_dp->dfp.ycbcr_444_to_420 = true;
+			connector->base.ycbcr_420_allowed = true;
+		} else {
+		/* Prefer 4:2:0 passthrough over 4:4:4->4:2:0 conversion */
+			intel_dp->dfp.ycbcr_444_to_420 =
+				ycbcr_444_to_420 && !ycbcr_420_passthrough;
+
+			connector->base.ycbcr_420_allowed =
+				!is_branch || ycbcr_444_to_420 || ycbcr_420_passthrough;
+		}
+	} else {
+		/* 4:4:4->4:2:0 conversion is the only way */
+		intel_dp->dfp.ycbcr_444_to_420 = ycbcr_444_to_420;
+
+		connector->base.ycbcr_420_allowed = ycbcr_444_to_420;
+	}
+
+	drm_dbg_kms(&i915->drm,
+		    "[CONNECTOR:%d:%s] RGB->YcbCr conversion? %s, YCbCr 4:2:0 allowed? %s, YCbCr 4:4:4->4:2:0 conversion? %s\n",
+		    connector->base.base.id, connector->base.name,
+		    str_yes_no(intel_dp->dfp.rgb_to_ycbcr),
+		    str_yes_no(connector->base.ycbcr_420_allowed),
+		    str_yes_no(intel_dp->dfp.ycbcr_444_to_420));
+}
+
+static void
+intel_dp_set_edid(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct edid *edid;
+	bool vrr_capable;
+
+	intel_dp_unset_edid(intel_dp);
+	edid = intel_dp_get_edid(intel_dp);
+	connector->detect_edid = edid;
+
+	vrr_capable = intel_vrr_is_capable(connector);
+	drm_dbg_kms(&i915->drm, "[CONNECTOR:%d:%s] VRR capable: %s\n",
+		    connector->base.base.id, connector->base.name, str_yes_no(vrr_capable));
+	drm_connector_set_vrr_capable_property(&connector->base, vrr_capable);
+
+	intel_dp_update_dfp(intel_dp, edid);
+	intel_dp_update_420(intel_dp);
+
+	if (edid && edid->input & DRM_EDID_INPUT_DIGITAL) {
+		intel_dp->has_hdmi_sink = drm_detect_hdmi_monitor(edid);
+		intel_dp->has_audio = drm_detect_monitor_audio(edid);
+	}
+
+	drm_dp_cec_set_edid(&intel_dp->aux, edid);
+}
+
+static void
+intel_dp_unset_edid(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+
+	drm_dp_cec_unset_edid(&intel_dp->aux);
+	kfree(connector->detect_edid);
+	connector->detect_edid = NULL;
+
+	intel_dp->has_hdmi_sink = false;
+	intel_dp->has_audio = false;
+
+	intel_dp->dfp.max_bpc = 0;
+	intel_dp->dfp.max_dotclock = 0;
+	intel_dp->dfp.min_tmds_clock = 0;
+	intel_dp->dfp.max_tmds_clock = 0;
+
+	intel_dp->dfp.pcon_max_frl_bw = 0;
+
+	intel_dp->dfp.ycbcr_444_to_420 = false;
+	connector->base.ycbcr_420_allowed = false;
+
+	drm_connector_set_vrr_capable_property(&connector->base,
+					       false);
+}
+
+static int
+intel_dp_detect(struct drm_connector *connector,
+		struct drm_modeset_acquire_ctx *ctx,
+		bool force)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_dp *intel_dp = intel_attached_dp(to_intel_connector(connector));
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &dig_port->base;
+	enum drm_connector_status status;
+
+	drm_dbg_kms(&dev_priv->drm, "[CONNECTOR:%d:%s]\n",
+		    connector->base.id, connector->name);
+	drm_WARN_ON(&dev_priv->drm,
+		    !drm_modeset_is_locked(&dev_priv->drm.mode_config.connection_mutex));
+
+	if (!INTEL_DISPLAY_ENABLED(dev_priv))
+		return connector_status_disconnected;
+
+	/* Can't disconnect eDP */
+	if (intel_dp_is_edp(intel_dp))
+		status = edp_detect(intel_dp);
+	else if (intel_digital_port_connected(encoder))
+		status = intel_dp_detect_dpcd(intel_dp);
+	else
+		status = connector_status_disconnected;
+
+	if (status == connector_status_disconnected) {
+		memset(&intel_dp->compliance, 0, sizeof(intel_dp->compliance));
+		memset(intel_dp->dsc_dpcd, 0, sizeof(intel_dp->dsc_dpcd));
+
+		if (intel_dp->is_mst) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "MST device may have disappeared %d vs %d\n",
+				    intel_dp->is_mst,
+				    intel_dp->mst_mgr.mst_state);
+			intel_dp->is_mst = false;
+			drm_dp_mst_topology_mgr_set_mst(&intel_dp->mst_mgr,
+							intel_dp->is_mst);
+		}
+
+		goto out;
+	}
+
+	/* Read DP Sink DSC Cap DPCD regs for DP v1.4 */
+	if (DISPLAY_VER(dev_priv) >= 11)
+		intel_dp_get_dsc_sink_cap(intel_dp);
+
+	intel_dp_configure_mst(intel_dp);
+
+	/*
+	 * TODO: Reset link params when switching to MST mode, until MST
+	 * supports link training fallback params.
+	 */
+	if (intel_dp->reset_link_params || intel_dp->is_mst) {
+		intel_dp_reset_max_link_params(intel_dp);
+		intel_dp->reset_link_params = false;
+	}
+
+	intel_dp_print_rates(intel_dp);
+
+	if (intel_dp->is_mst) {
+		/*
+		 * If we are in MST mode then this connector
+		 * won't appear connected or have anything
+		 * with EDID on it
+		 */
+		status = connector_status_disconnected;
+		goto out;
+	}
+
+	/*
+	 * Some external monitors do not signal loss of link synchronization
+	 * with an IRQ_HPD, so force a link status check.
+	 */
+	if (!intel_dp_is_edp(intel_dp)) {
+		int ret;
+
+		ret = intel_dp_retrain_link(encoder, ctx);
+		if (ret)
+			return ret;
+	}
+
+	/*
+	 * Clearing NACK and defer counts to get their exact values
+	 * while reading EDID which are required by Compliance tests
+	 * 4.2.2.4 and 4.2.2.5
+	 */
+	intel_dp->aux.i2c_nack_count = 0;
+	intel_dp->aux.i2c_defer_count = 0;
+
+	intel_dp_set_edid(intel_dp);
+	if (intel_dp_is_edp(intel_dp) ||
+	    to_intel_connector(connector)->detect_edid)
+		status = connector_status_connected;
+
+	intel_dp_check_device_service_irq(intel_dp);
+
+out:
+	if (status != connector_status_connected && !intel_dp->is_mst)
+		intel_dp_unset_edid(intel_dp);
+
+	/*
+	 * Make sure the refs for power wells enabled during detect are
+	 * dropped to avoid a new detect cycle triggered by HPD polling.
+	 */
+	intel_display_power_flush_work(dev_priv);
+
+	if (!intel_dp_is_edp(intel_dp))
+		drm_dp_set_subconnector_property(connector,
+						 status,
+						 intel_dp->dpcd,
+						 intel_dp->downstream_ports);
+	return status;
+}
+
+static void
+intel_dp_force(struct drm_connector *connector)
+{
+	struct intel_dp *intel_dp = intel_attached_dp(to_intel_connector(connector));
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *intel_encoder = &dig_port->base;
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
+	enum intel_display_power_domain aux_domain =
+		intel_aux_power_domain(dig_port);
+	intel_wakeref_t wakeref;
+
+	drm_dbg_kms(&dev_priv->drm, "[CONNECTOR:%d:%s]\n",
+		    connector->base.id, connector->name);
+	intel_dp_unset_edid(intel_dp);
+
+	if (connector->status != connector_status_connected)
+		return;
+
+	wakeref = intel_display_power_get(dev_priv, aux_domain);
+
+	intel_dp_set_edid(intel_dp);
+
+	intel_display_power_put(dev_priv, aux_domain, wakeref);
+}
+
+static int intel_dp_get_modes(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct edid *edid;
+	int num_modes = 0;
+
+	edid = intel_connector->detect_edid;
+	if (edid)
+		num_modes = intel_connector_update_modes(connector, edid);
+
+	/* Also add fixed mode, which may or may not be present in EDID */
+	if (intel_dp_is_edp(intel_attached_dp(intel_connector)))
+		num_modes += intel_panel_get_modes(intel_connector);
+
+	if (num_modes)
+		return num_modes;
+
+	if (!edid) {
+		struct intel_dp *intel_dp = intel_attached_dp(intel_connector);
+		struct drm_display_mode *mode;
+
+		mode = drm_dp_downstream_mode(connector->dev,
+					      intel_dp->dpcd,
+					      intel_dp->downstream_ports);
+		if (mode) {
+			drm_mode_probed_add(connector, mode);
+			num_modes++;
+		}
+	}
+
+	return num_modes;
+}
+
+static int
+intel_dp_connector_register(struct drm_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct intel_dp *intel_dp = intel_attached_dp(to_intel_connector(connector));
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct intel_lspcon *lspcon = &dig_port->lspcon;
+	int ret;
+
+	ret = intel_connector_register(connector);
+	if (ret)
+		return ret;
+
+	drm_dbg_kms(&i915->drm, "registering %s bus for %s\n",
+		    intel_dp->aux.name, connector->kdev->kobj.name);
+
+	intel_dp->aux.dev = connector->kdev;
+	ret = drm_dp_aux_register(&intel_dp->aux);
+	if (!ret)
+		drm_dp_cec_register_connector(&intel_dp->aux, connector);
+
+	if (!intel_bios_is_lspcon_present(i915, dig_port->base.port))
+		return ret;
+
+	/*
+	 * ToDo: Clean this up to handle lspcon init and resume more
+	 * efficiently and streamlined.
+	 */
+	if (lspcon_init(dig_port)) {
+		lspcon_detect_hdr_capability(lspcon);
+		if (lspcon->hdr_supported)
+			drm_connector_attach_hdr_output_metadata_property(connector);
+	}
+
+	return ret;
+}
+
+static void
+intel_dp_connector_unregister(struct drm_connector *connector)
+{
+	struct intel_dp *intel_dp = intel_attached_dp(to_intel_connector(connector));
+
+	drm_dp_cec_unregister_connector(&intel_dp->aux);
+	drm_dp_aux_unregister(&intel_dp->aux);
+	intel_connector_unregister(connector);
+}
+
+void intel_dp_encoder_flush_work(struct drm_encoder *encoder)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(to_intel_encoder(encoder));
+	struct intel_dp *intel_dp = &dig_port->dp;
+
+	intel_dp_mst_encoder_cleanup(dig_port);
+
+	intel_pps_vdd_off_sync(intel_dp);
+
+	intel_dp_aux_fini(intel_dp);
+}
+
+void intel_dp_encoder_suspend(struct intel_encoder *intel_encoder)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(intel_encoder);
+
+	intel_pps_vdd_off_sync(intel_dp);
+}
+
+void intel_dp_encoder_shutdown(struct intel_encoder *intel_encoder)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(intel_encoder);
+
+	intel_pps_wait_power_cycle(intel_dp);
+}
+
+static int intel_modeset_tile_group(struct intel_atomic_state *state,
+				    int tile_group_id)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct drm_connector_list_iter conn_iter;
+	struct drm_connector *connector;
+	int ret = 0;
+
+	drm_connector_list_iter_begin(&dev_priv->drm, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		struct drm_connector_state *conn_state;
+		struct intel_crtc_state *crtc_state;
+		struct intel_crtc *crtc;
+
+		if (!connector->has_tile ||
+		    connector->tile_group->id != tile_group_id)
+			continue;
+
+		conn_state = drm_atomic_get_connector_state(&state->base,
+							    connector);
+		if (IS_ERR(conn_state)) {
+			ret = PTR_ERR(conn_state);
+			break;
+		}
+
+		crtc = to_intel_crtc(conn_state->crtc);
+
+		if (!crtc)
+			continue;
+
+		crtc_state = intel_atomic_get_new_crtc_state(state, crtc);
+		crtc_state->uapi.mode_changed = true;
+
+		ret = drm_atomic_add_affected_planes(&state->base, &crtc->base);
+		if (ret)
+			break;
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	return ret;
+}
+
+static int intel_modeset_affected_transcoders(struct intel_atomic_state *state, u8 transcoders)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc *crtc;
+
+	if (transcoders == 0)
+		return 0;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		struct intel_crtc_state *crtc_state;
+		int ret;
+
+		crtc_state = intel_atomic_get_crtc_state(&state->base, crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+
+		if (!crtc_state->hw.enable)
+			continue;
+
+		if (!(transcoders & BIT(crtc_state->cpu_transcoder)))
+			continue;
+
+		crtc_state->uapi.mode_changed = true;
+
+		ret = drm_atomic_add_affected_connectors(&state->base, &crtc->base);
+		if (ret)
+			return ret;
+
+		ret = drm_atomic_add_affected_planes(&state->base, &crtc->base);
+		if (ret)
+			return ret;
+
+		transcoders &= ~BIT(crtc_state->cpu_transcoder);
+	}
+
+	drm_WARN_ON(&dev_priv->drm, transcoders != 0);
+
+	return 0;
+}
+
+static int intel_modeset_synced_crtcs(struct intel_atomic_state *state,
+				      struct drm_connector *connector)
+{
+	const struct drm_connector_state *old_conn_state =
+		drm_atomic_get_old_connector_state(&state->base, connector);
+	const struct intel_crtc_state *old_crtc_state;
+	struct intel_crtc *crtc;
+	u8 transcoders;
+
+	crtc = to_intel_crtc(old_conn_state->crtc);
+	if (!crtc)
+		return 0;
+
+	old_crtc_state = intel_atomic_get_old_crtc_state(state, crtc);
+
+	if (!old_crtc_state->hw.active)
+		return 0;
+
+	transcoders = old_crtc_state->sync_mode_slaves_mask;
+	if (old_crtc_state->master_transcoder != INVALID_TRANSCODER)
+		transcoders |= BIT(old_crtc_state->master_transcoder);
+
+	return intel_modeset_affected_transcoders(state,
+						  transcoders);
+}
+
+static int intel_dp_connector_atomic_check(struct drm_connector *conn,
+					   struct drm_atomic_state *_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(conn->dev);
+	struct intel_atomic_state *state = to_intel_atomic_state(_state);
+	struct drm_connector_state *conn_state = drm_atomic_get_new_connector_state(_state, conn);
+	struct intel_connector *intel_conn = to_intel_connector(conn);
+	struct intel_dp *intel_dp = enc_to_intel_dp(intel_conn->encoder);
+	int ret;
+
+	ret = intel_digital_connector_atomic_check(conn, &state->base);
+	if (ret)
+		return ret;
+
+	if (intel_dp_mst_source_support(intel_dp)) {
+		ret = drm_dp_mst_root_conn_atomic_check(conn_state, &intel_dp->mst_mgr);
+		if (ret)
+			return ret;
+	}
+
+	/*
+	 * We don't enable port sync on BDW due to missing w/as and
+	 * due to not having adjusted the modeset sequence appropriately.
+	 */
+	if (DISPLAY_VER(dev_priv) < 9)
+		return 0;
+
+	if (!intel_connector_needs_modeset(state, conn))
+		return 0;
+
+	if (conn->has_tile) {
+		ret = intel_modeset_tile_group(state, conn->tile_group->id);
+		if (ret)
+			return ret;
+	}
+
+	return intel_modeset_synced_crtcs(state, conn);
+}
+
+static void intel_dp_oob_hotplug_event(struct drm_connector *connector)
+{
+	struct intel_encoder *encoder = intel_attached_encoder(to_intel_connector(connector));
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+
+	spin_lock_irq(&i915->irq_lock);
+	i915->display.hotplug.event_bits |= BIT(encoder->hpd_pin);
+	spin_unlock_irq(&i915->irq_lock);
+	queue_delayed_work(system_wq, &i915->display.hotplug.hotplug_work, 0);
+}
+
+static const struct drm_connector_funcs intel_dp_connector_funcs = {
+	.force = intel_dp_force,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.atomic_get_property = intel_digital_connector_atomic_get_property,
+	.atomic_set_property = intel_digital_connector_atomic_set_property,
+	.late_register = intel_dp_connector_register,
+	.early_unregister = intel_dp_connector_unregister,
+	.destroy = intel_connector_destroy,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.atomic_duplicate_state = intel_digital_connector_duplicate_state,
+	.oob_hotplug_event = intel_dp_oob_hotplug_event,
+};
+
+static const struct drm_connector_helper_funcs intel_dp_connector_helper_funcs = {
+	.detect_ctx = intel_dp_detect,
+	.get_modes = intel_dp_get_modes,
+	.mode_valid = intel_dp_mode_valid,
+	.atomic_check = intel_dp_connector_atomic_check,
+};
+
+enum irqreturn
+intel_dp_hpd_pulse(struct intel_digital_port *dig_port, bool long_hpd)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_dp *intel_dp = &dig_port->dp;
+
+	if (dig_port->base.type == INTEL_OUTPUT_EDP &&
+	    (long_hpd || !intel_pps_have_panel_power_or_vdd(intel_dp))) {
+		/*
+		 * vdd off can generate a long/short pulse on eDP which
+		 * would require vdd on to handle it, and thus we
+		 * would end up in an endless cycle of
+		 * "vdd off -> long/short hpd -> vdd on -> detect -> vdd off -> ..."
+		 */
+		drm_dbg_kms(&i915->drm,
+			    "ignoring %s hpd on eDP [ENCODER:%d:%s]\n",
+			    long_hpd ? "long" : "short",
+			    dig_port->base.base.base.id,
+			    dig_port->base.base.name);
+		return IRQ_HANDLED;
+	}
+
+	drm_dbg_kms(&i915->drm, "got hpd irq on [ENCODER:%d:%s] - %s\n",
+		    dig_port->base.base.base.id,
+		    dig_port->base.base.name,
+		    long_hpd ? "long" : "short");
+
+	if (long_hpd) {
+		intel_dp->reset_link_params = true;
+		return IRQ_NONE;
+	}
+
+	if (intel_dp->is_mst) {
+		if (!intel_dp_check_mst_status(intel_dp))
+			return IRQ_NONE;
+	} else if (!intel_dp_short_pulse(intel_dp)) {
+		return IRQ_NONE;
+	}
+
+	return IRQ_HANDLED;
+}
+
+/* check the VBT to see whether the eDP is on another port */
+bool intel_dp_is_port_edp(struct drm_i915_private *dev_priv, enum port port)
+{
+	/*
+	 * eDP not supported on g4x. so bail out early just
+	 * for a bit extra safety in case the VBT is bonkers.
+	 */
+	if (DISPLAY_VER(dev_priv) < 5)
+		return false;
+
+	if (DISPLAY_VER(dev_priv) < 9 && port == PORT_A)
+		return true;
+
+	return intel_bios_is_port_edp(dev_priv, port);
+}
+
+static bool
+has_gamut_metadata_dip(struct drm_i915_private *i915, enum port port)
+{
+	if (intel_bios_is_lspcon_present(i915, port))
+		return false;
+
+	if (DISPLAY_VER(i915) >= 11)
+		return true;
+
+	if (port == PORT_A)
+		return false;
+
+	if (IS_HASWELL(i915) || IS_BROADWELL(i915) ||
+	    DISPLAY_VER(i915) >= 9)
+		return true;
+
+	return false;
+}
+
+static void
+intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	enum port port = dp_to_dig_port(intel_dp)->base.port;
+
+	if (!intel_dp_is_edp(intel_dp))
+		drm_connector_attach_dp_subconnector_property(connector);
+
+	if (!IS_G4X(dev_priv) && port != PORT_A)
+		intel_attach_force_audio_property(connector);
+
+	intel_attach_broadcast_rgb_property(connector);
+	if (HAS_GMCH(dev_priv))
+		drm_connector_attach_max_bpc_property(connector, 6, 10);
+	else if (DISPLAY_VER(dev_priv) >= 5)
+		drm_connector_attach_max_bpc_property(connector, 6, 12);
+
+	/* Register HDMI colorspace for case of lspcon */
+	if (intel_bios_is_lspcon_present(dev_priv, port)) {
+		drm_connector_attach_content_type_property(connector);
+		intel_attach_hdmi_colorspace_property(connector);
+	} else {
+		intel_attach_dp_colorspace_property(connector);
+	}
+
+	if (has_gamut_metadata_dip(dev_priv, port))
+		drm_connector_attach_hdr_output_metadata_property(connector);
+
+	if (intel_dp_is_edp(intel_dp)) {
+		u32 allowed_scalers;
+
+		allowed_scalers = BIT(DRM_MODE_SCALE_ASPECT) | BIT(DRM_MODE_SCALE_FULLSCREEN);
+		if (!HAS_GMCH(dev_priv))
+			allowed_scalers |= BIT(DRM_MODE_SCALE_CENTER);
+
+		drm_connector_attach_scaling_mode_property(connector, allowed_scalers);
+
+		connector->state->scaling_mode = DRM_MODE_SCALE_ASPECT;
+
+	}
+
+	if (HAS_VRR(dev_priv))
+		drm_connector_attach_vrr_capable_property(connector);
+}
+
+static void
+intel_edp_add_properties(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_preferred_fixed_mode(connector);
+
+	if (!fixed_mode)
+		return;
+
+	drm_connector_set_panel_orientation_with_quirk(&connector->base,
+						       i915->display.vbt.orientation,
+						       fixed_mode->hdisplay,
+						       fixed_mode->vdisplay);
+}
+
+static bool intel_edp_init_connector(struct intel_dp *intel_dp,
+				     struct intel_connector *intel_connector)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_connector *connector = &intel_connector->base;
+	struct drm_display_mode *fixed_mode;
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	bool has_dpcd;
+	enum pipe pipe = INVALID_PIPE;
+	struct edid *edid;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return true;
+
+	/*
+	 * On IBX/CPT we may get here with LVDS already registered. Since the
+	 * driver uses the only internal power sequencer available for both
+	 * eDP and LVDS bail out early in this case to prevent interfering
+	 * with an already powered-on LVDS power sequencer.
+	 */
+	if (intel_get_lvds_encoder(dev_priv)) {
+		drm_WARN_ON(dev,
+			    !(HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)));
+		drm_info(&dev_priv->drm,
+			 "LVDS was detected, not registering eDP\n");
+
+		return false;
+	}
+
+	intel_bios_init_panel_early(dev_priv, &intel_connector->panel,
+				    encoder->devdata);
+
+	intel_pps_init(intel_dp);
+
+	/* Cache DPCD and EDID for edp. */
+	has_dpcd = intel_edp_init_dpcd(intel_dp);
+
+	if (!has_dpcd) {
+		/* if this fails, presume the device is a ghost */
+		drm_info(&dev_priv->drm,
+			 "failed to retrieve link info, disabling eDP\n");
+		goto out_vdd_off;
+	}
+
+	mutex_lock(&dev->mode_config.mutex);
+	edid = drm_get_edid(connector, &intel_dp->aux.ddc);
+	if (!edid) {
+		/* Fallback to EDID from ACPI OpRegion, if any */
+		edid = intel_opregion_get_edid(intel_connector);
+		if (edid)
+			drm_dbg_kms(&dev_priv->drm,
+				    "[CONNECTOR:%d:%s] Using OpRegion EDID\n",
+				    connector->base.id, connector->name);
+	}
+	if (edid) {
+		if (drm_add_edid_modes(connector, edid)) {
+			drm_connector_update_edid_property(connector, edid);
+		} else {
+			kfree(edid);
+			edid = ERR_PTR(-EINVAL);
+		}
+	} else {
+		edid = ERR_PTR(-ENOENT);
+	}
+	intel_connector->edid = edid;
+
+	intel_bios_init_panel_late(dev_priv, &intel_connector->panel,
+				   encoder->devdata, IS_ERR(edid) ? NULL : edid);
+
+	intel_panel_add_edid_fixed_modes(intel_connector, true);
+
+	/* MSO requires information from the EDID */
+	intel_edp_mso_init(intel_dp);
+
+	/* multiply the mode clock and horizontal timings for MSO */
+	list_for_each_entry(fixed_mode, &intel_connector->panel.fixed_modes, head)
+		intel_edp_mso_mode_fixup(intel_connector, fixed_mode);
+
+	/* fallback to VBT if available for eDP */
+	if (!intel_panel_preferred_fixed_mode(intel_connector))
+		intel_panel_add_vbt_lfp_fixed_mode(intel_connector);
+
+	mutex_unlock(&dev->mode_config.mutex);
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		/*
+		 * Figure out the current pipe for the initial backlight setup.
+		 * If the current pipe isn't valid, try the PPS pipe, and if that
+		 * fails just assume pipe A.
+		 */
+		pipe = vlv_active_pipe(intel_dp);
+
+		if (pipe != PIPE_A && pipe != PIPE_B)
+			pipe = intel_dp->pps.pps_pipe;
+
+		if (pipe != PIPE_A && pipe != PIPE_B)
+			pipe = PIPE_A;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "using pipe %c for initial backlight setup\n",
+			    pipe_name(pipe));
+	}
+
+	intel_panel_init(intel_connector);
+
+	intel_backlight_setup(intel_connector, pipe);
+
+	intel_edp_add_properties(intel_dp);
+
+	intel_pps_init_late(intel_dp);
+
+	return true;
+
+out_vdd_off:
+	intel_pps_vdd_off_sync(intel_dp);
+
+	return false;
+}
+
+static void intel_dp_modeset_retry_work_fn(struct work_struct *work)
+{
+	struct intel_connector *intel_connector;
+	struct drm_connector *connector;
+
+	intel_connector = container_of(work, typeof(*intel_connector),
+				       modeset_retry_work);
+	connector = &intel_connector->base;
+	drm_dbg_kms(connector->dev, "[CONNECTOR:%d:%s]\n", connector->base.id,
+		    connector->name);
+
+	/* Grab the locks before changing connector property*/
+	mutex_lock(&connector->dev->mode_config.mutex);
+	/* Set connector link status to BAD and send a Uevent to notify
+	 * userspace to do a modeset.
+	 */
+	drm_connector_set_link_status_property(connector,
+					       DRM_MODE_LINK_STATUS_BAD);
+	mutex_unlock(&connector->dev->mode_config.mutex);
+	/* Send Hotplug uevent so userspace can reprobe */
+	drm_kms_helper_connector_hotplug_event(connector);
+}
+
+bool
+intel_dp_init_connector(struct intel_digital_port *dig_port,
+			struct intel_connector *intel_connector)
+{
+	struct drm_connector *connector = &intel_connector->base;
+	struct intel_dp *intel_dp = &dig_port->dp;
+	struct intel_encoder *intel_encoder = &dig_port->base;
+	struct drm_device *dev = intel_encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum port port = intel_encoder->port;
+	enum phy phy = intel_port_to_phy(dev_priv, port);
+	int type;
+
+	/* Initialize the work for modeset in case of link train failure */
+	INIT_WORK(&intel_connector->modeset_retry_work,
+		  intel_dp_modeset_retry_work_fn);
+
+	if (drm_WARN(dev, dig_port->max_lanes < 1,
+		     "Not enough lanes (%d) for DP on [ENCODER:%d:%s]\n",
+		     dig_port->max_lanes, intel_encoder->base.base.id,
+		     intel_encoder->base.name))
+		return false;
+
+	intel_dp->reset_link_params = true;
+	intel_dp->pps.pps_pipe = INVALID_PIPE;
+	intel_dp->pps.active_pipe = INVALID_PIPE;
+
+	/* Preserve the current hw state. */
+	intel_dp->DP = intel_de_read(dev_priv, intel_dp->output_reg);
+	intel_dp->attached_connector = intel_connector;
+
+	if (intel_dp_is_port_edp(dev_priv, port)) {
+		/*
+		 * Currently we don't support eDP on TypeC ports, although in
+		 * theory it could work on TypeC legacy ports.
+		 */
+		drm_WARN_ON(dev, intel_phy_is_tc(dev_priv, phy));
+		type = DRM_MODE_CONNECTOR_eDP;
+		intel_encoder->type = INTEL_OUTPUT_EDP;
+
+		/* eDP only on port B and/or C on vlv/chv */
+		if (drm_WARN_ON(dev, (IS_VALLEYVIEW(dev_priv) ||
+				      IS_CHERRYVIEW(dev_priv)) &&
+				port != PORT_B && port != PORT_C))
+			return false;
+	} else {
+		type = DRM_MODE_CONNECTOR_DisplayPort;
+	}
+
+	intel_dp_set_default_sink_rates(intel_dp);
+	intel_dp_set_default_max_sink_lane_count(intel_dp);
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		intel_dp->pps.active_pipe = vlv_active_pipe(intel_dp);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Adding %s connector on [ENCODER:%d:%s]\n",
+		    type == DRM_MODE_CONNECTOR_eDP ? "eDP" : "DP",
+		    intel_encoder->base.base.id, intel_encoder->base.name);
+
+	drm_connector_init(dev, connector, &intel_dp_connector_funcs, type);
+	drm_connector_helper_add(connector, &intel_dp_connector_helper_funcs);
+
+	if (!HAS_GMCH(dev_priv))
+		connector->interlace_allowed = true;
+	connector->doublescan_allowed = 0;
+
+	intel_connector->polled = DRM_CONNECTOR_POLL_HPD;
+
+	intel_dp_aux_init(intel_dp);
+
+	intel_connector_attach_encoder(intel_connector, intel_encoder);
+
+	if (HAS_DDI(dev_priv))
+		intel_connector->get_hw_state = intel_ddi_connector_get_hw_state;
+	else
+		intel_connector->get_hw_state = intel_connector_get_hw_state;
+
+	if (!intel_edp_init_connector(intel_dp, intel_connector)) {
+		intel_dp_aux_fini(intel_dp);
+		goto fail;
+	}
+
+	intel_dp_set_source_rates(intel_dp);
+	intel_dp_set_common_rates(intel_dp);
+	intel_dp_reset_max_link_params(intel_dp);
+
+	/* init MST on ports that can support it */
+	intel_dp_mst_encoder_init(dig_port,
+				  intel_connector->base.base.id);
+
+	intel_dp_add_properties(intel_dp, connector);
+
+	if (is_hdcp_supported(dev_priv, port) && !intel_dp_is_edp(intel_dp)) {
+		int ret = intel_dp_hdcp_init(dig_port, intel_connector);
+		if (ret)
+			drm_dbg_kms(&dev_priv->drm,
+				    "HDCP init failed, skipping.\n");
+	}
+
+	/* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first be written
+	 * 0xd.  Failure to do so will result in spurious interrupts being
+	 * generated on the port when a cable is not attached.
+	 */
+	if (IS_G45(dev_priv)) {
+		u32 temp = intel_de_read(dev_priv, PEG_BAND_GAP_DATA);
+		intel_de_write(dev_priv, PEG_BAND_GAP_DATA,
+			       (temp & ~0xf) | 0xd);
+	}
+
+	intel_dp->frl.is_trained = false;
+	intel_dp->frl.trained_rate_gbps = 0;
+
+	intel_psr_init(intel_dp);
+
+	return true;
+
+fail:
+	drm_connector_cleanup(connector);
+
+	return false;
+}
+
+void intel_dp_mst_suspend(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp;
+
+		if (encoder->type != INTEL_OUTPUT_DDI)
+			continue;
+
+		intel_dp = enc_to_intel_dp(encoder);
+
+		if (!intel_dp_mst_source_support(intel_dp))
+			continue;
+
+		if (intel_dp->is_mst)
+			drm_dp_mst_topology_mgr_suspend(&intel_dp->mst_mgr);
+	}
+}
+
+void intel_dp_mst_resume(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp;
+		int ret;
+
+		if (encoder->type != INTEL_OUTPUT_DDI)
+			continue;
+
+		intel_dp = enc_to_intel_dp(encoder);
+
+		if (!intel_dp_mst_source_support(intel_dp))
+			continue;
+
+		ret = drm_dp_mst_topology_mgr_resume(&intel_dp->mst_mgr,
+						     true);
+		if (ret) {
+			intel_dp->is_mst = false;
+			drm_dp_mst_topology_mgr_set_mst(&intel_dp->mst_mgr,
+							false);
+		}
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dp.h b/drivers/gpu/drm/i915/display/intel_dp.h
new file mode 100644
index 000000000..a54902c71
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp.h
@@ -0,0 +1,124 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DP_H__
+#define __INTEL_DP_H__
+
+#include <linux/types.h>
+
+enum intel_output_format;
+enum pipe;
+enum port;
+struct drm_connector_state;
+struct drm_encoder;
+struct drm_i915_private;
+struct drm_modeset_acquire_ctx;
+struct drm_dp_vsc_sdp;
+struct intel_atomic_state;
+struct intel_connector;
+struct intel_crtc_state;
+struct intel_digital_port;
+struct intel_dp;
+struct intel_encoder;
+
+struct link_config_limits {
+	int min_rate, max_rate;
+	int min_lane_count, max_lane_count;
+	int min_bpp, max_bpp;
+};
+
+void intel_edp_fixup_vbt_bpp(struct intel_encoder *encoder, int pipe_bpp);
+void intel_dp_adjust_compliance_config(struct intel_dp *intel_dp,
+				       struct intel_crtc_state *pipe_config,
+				       struct link_config_limits *limits);
+bool intel_dp_limited_color_range(const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state);
+int intel_dp_min_bpp(enum intel_output_format output_format);
+bool intel_dp_init_connector(struct intel_digital_port *dig_port,
+			     struct intel_connector *intel_connector);
+void intel_dp_set_link_params(struct intel_dp *intel_dp,
+			      int link_rate, int lane_count);
+int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp,
+					    int link_rate, u8 lane_count);
+int intel_dp_retrain_link(struct intel_encoder *encoder,
+			  struct drm_modeset_acquire_ctx *ctx);
+void intel_dp_set_power(struct intel_dp *intel_dp, u8 mode);
+void intel_dp_configure_protocol_converter(struct intel_dp *intel_dp,
+					   const struct intel_crtc_state *crtc_state);
+void intel_dp_sink_set_decompression_state(struct intel_dp *intel_dp,
+					   const struct intel_crtc_state *crtc_state,
+					   bool enable);
+void intel_dp_encoder_suspend(struct intel_encoder *intel_encoder);
+void intel_dp_encoder_shutdown(struct intel_encoder *intel_encoder);
+void intel_dp_encoder_flush_work(struct drm_encoder *encoder);
+int intel_dp_compute_config(struct intel_encoder *encoder,
+			    struct intel_crtc_state *pipe_config,
+			    struct drm_connector_state *conn_state);
+bool intel_dp_is_edp(struct intel_dp *intel_dp);
+bool intel_dp_is_uhbr(const struct intel_crtc_state *crtc_state);
+bool intel_dp_is_port_edp(struct drm_i915_private *dev_priv, enum port port);
+enum irqreturn intel_dp_hpd_pulse(struct intel_digital_port *dig_port,
+				  bool long_hpd);
+void intel_edp_backlight_on(const struct intel_crtc_state *crtc_state,
+			    const struct drm_connector_state *conn_state);
+void intel_edp_backlight_off(const struct drm_connector_state *conn_state);
+void intel_edp_fixup_vbt_bpp(struct intel_encoder *encoder, int pipe_bpp);
+void intel_dp_mst_suspend(struct drm_i915_private *dev_priv);
+void intel_dp_mst_resume(struct drm_i915_private *dev_priv);
+int intel_dp_max_link_rate(struct intel_dp *intel_dp);
+int intel_dp_max_lane_count(struct intel_dp *intel_dp);
+int intel_dp_rate_select(struct intel_dp *intel_dp, int rate);
+
+void intel_dp_compute_rate(struct intel_dp *intel_dp, int port_clock,
+			   u8 *link_bw, u8 *rate_select);
+bool intel_dp_source_supports_tps3(struct drm_i915_private *i915);
+bool intel_dp_source_supports_tps4(struct drm_i915_private *i915);
+
+bool intel_dp_get_colorimetry_status(struct intel_dp *intel_dp);
+int intel_dp_link_required(int pixel_clock, int bpp);
+int intel_dp_max_data_rate(int max_link_rate, int max_lanes);
+bool intel_dp_can_bigjoiner(struct intel_dp *intel_dp);
+bool intel_dp_needs_vsc_sdp(const struct intel_crtc_state *crtc_state,
+			    const struct drm_connector_state *conn_state);
+void intel_dp_compute_psr_vsc_sdp(struct intel_dp *intel_dp,
+				  const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state,
+				  struct drm_dp_vsc_sdp *vsc);
+void intel_write_dp_vsc_sdp(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    const struct drm_dp_vsc_sdp *vsc);
+void intel_dp_set_infoframes(struct intel_encoder *encoder, bool enable,
+			     const struct intel_crtc_state *crtc_state,
+			     const struct drm_connector_state *conn_state);
+void intel_read_dp_sdp(struct intel_encoder *encoder,
+		       struct intel_crtc_state *crtc_state,
+		       unsigned int type);
+bool intel_digital_port_connected(struct intel_encoder *encoder);
+
+static inline unsigned int intel_dp_unused_lane_mask(int lane_count)
+{
+	return ~((1 << lane_count) - 1) & 0xf;
+}
+
+u32 intel_dp_mode_to_fec_clock(u32 mode_clock);
+
+void intel_ddi_update_pipe(struct intel_atomic_state *state,
+			   struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state,
+			   const struct drm_connector_state *conn_state);
+
+bool intel_dp_initial_fastset_check(struct intel_encoder *encoder,
+				    struct intel_crtc_state *crtc_state);
+void intel_dp_sync_state(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *crtc_state);
+
+void intel_dp_check_frl_training(struct intel_dp *intel_dp);
+void intel_dp_pcon_dsc_configure(struct intel_dp *intel_dp,
+				 const struct intel_crtc_state *crtc_state);
+void intel_dp_phy_test(struct intel_encoder *encoder);
+
+void intel_dp_wait_source_oui(struct intel_dp *intel_dp);
+
+#endif /* __INTEL_DP_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.c b/drivers/gpu/drm/i915/display/intel_dp_aux.c
new file mode 100644
index 000000000..ab357161c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_aux.c
@@ -0,0 +1,771 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020-2021 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_trace.h"
+#include "intel_display_types.h"
+#include "intel_dp_aux.h"
+#include "intel_pps.h"
+#include "intel_tc.h"
+
+static u32 intel_dp_aux_pack(const u8 *src, int src_bytes)
+{
+	int i;
+	u32 v = 0;
+
+	if (src_bytes > 4)
+		src_bytes = 4;
+	for (i = 0; i < src_bytes; i++)
+		v |= ((u32)src[i]) << ((3 - i) * 8);
+	return v;
+}
+
+static void intel_dp_aux_unpack(u32 src, u8 *dst, int dst_bytes)
+{
+	int i;
+
+	if (dst_bytes > 4)
+		dst_bytes = 4;
+	for (i = 0; i < dst_bytes; i++)
+		dst[i] = src >> ((3 - i) * 8);
+}
+
+static u32
+intel_dp_aux_wait_done(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	i915_reg_t ch_ctl = intel_dp->aux_ch_ctl_reg(intel_dp);
+	const unsigned int timeout_ms = 10;
+	u32 status;
+	bool done;
+
+#define C (((status = intel_uncore_read_notrace(&i915->uncore, ch_ctl)) & DP_AUX_CH_CTL_SEND_BUSY) == 0)
+	done = wait_event_timeout(i915->display.gmbus.wait_queue, C,
+				  msecs_to_jiffies_timeout(timeout_ms));
+
+	/* just trace the final value */
+	trace_i915_reg_rw(false, ch_ctl, status, sizeof(status), true);
+
+	if (!done)
+		drm_err(&i915->drm,
+			"%s: did not complete or timeout within %ums (status 0x%08x)\n",
+			intel_dp->aux.name, timeout_ms, status);
+#undef C
+
+	return status;
+}
+
+static u32 g4x_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	if (index)
+		return 0;
+
+	/*
+	 * The clock divider is based off the hrawclk, and would like to run at
+	 * 2MHz.  So, take the hrawclk value and divide by 2000 and use that
+	 */
+	return DIV_ROUND_CLOSEST(RUNTIME_INFO(dev_priv)->rawclk_freq, 2000);
+}
+
+static u32 ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	u32 freq;
+
+	if (index)
+		return 0;
+
+	/*
+	 * The clock divider is based off the cdclk or PCH rawclk, and would
+	 * like to run at 2MHz.  So, take the cdclk or PCH rawclk value and
+	 * divide by 2000 and use that
+	 */
+	if (dig_port->aux_ch == AUX_CH_A)
+		freq = dev_priv->display.cdclk.hw.cdclk;
+	else
+		freq = RUNTIME_INFO(dev_priv)->rawclk_freq;
+	return DIV_ROUND_CLOSEST(freq, 2000);
+}
+
+static u32 hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+
+	if (dig_port->aux_ch != AUX_CH_A && HAS_PCH_LPT_H(dev_priv)) {
+		/* Workaround for non-ULT HSW */
+		switch (index) {
+		case 0: return 63;
+		case 1: return 72;
+		default: return 0;
+		}
+	}
+
+	return ilk_get_aux_clock_divider(intel_dp, index);
+}
+
+static u32 skl_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
+{
+	/*
+	 * SKL doesn't need us to program the AUX clock divider (Hardware will
+	 * derive the clock from CDCLK automatically). We still implement the
+	 * get_aux_clock_divider vfunc to plug-in into the existing code.
+	 */
+	return index ? 0 : 1;
+}
+
+static int intel_dp_aux_sync_len(void)
+{
+	int precharge = 16; /* 10-16 */
+	int preamble = 16;
+
+	return precharge + preamble;
+}
+
+static int intel_dp_aux_fw_sync_len(void)
+{
+	int precharge = 10; /* 10-16 */
+	int preamble = 8;
+
+	return precharge + preamble;
+}
+
+static int g4x_dp_aux_precharge_len(void)
+{
+	int precharge_min = 10;
+	int preamble = 16;
+
+	/* HW wants the length of the extra precharge in 2us units */
+	return (intel_dp_aux_sync_len() -
+		precharge_min - preamble) / 2;
+}
+
+static u32 g4x_get_aux_send_ctl(struct intel_dp *intel_dp,
+				int send_bytes,
+				u32 aux_clock_divider)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv =
+			to_i915(dig_port->base.base.dev);
+	u32 timeout;
+
+	/* Max timeout value on G4x-BDW: 1.6ms */
+	if (IS_BROADWELL(dev_priv))
+		timeout = DP_AUX_CH_CTL_TIME_OUT_600us;
+	else
+		timeout = DP_AUX_CH_CTL_TIME_OUT_400us;
+
+	return DP_AUX_CH_CTL_SEND_BUSY |
+	       DP_AUX_CH_CTL_DONE |
+	       DP_AUX_CH_CTL_INTERRUPT |
+	       DP_AUX_CH_CTL_TIME_OUT_ERROR |
+	       timeout |
+	       DP_AUX_CH_CTL_RECEIVE_ERROR |
+	       (send_bytes << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
+	       (g4x_dp_aux_precharge_len() << DP_AUX_CH_CTL_PRECHARGE_2US_SHIFT) |
+	       (aux_clock_divider << DP_AUX_CH_CTL_BIT_CLOCK_2X_SHIFT);
+}
+
+static u32 skl_get_aux_send_ctl(struct intel_dp *intel_dp,
+				int send_bytes,
+				u32 unused)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *i915 =	to_i915(dig_port->base.base.dev);
+	u32 ret;
+
+	/*
+	 * Max timeout values:
+	 * SKL-GLK: 1.6ms
+	 * ICL+: 4ms
+	 */
+	ret = DP_AUX_CH_CTL_SEND_BUSY |
+	      DP_AUX_CH_CTL_DONE |
+	      DP_AUX_CH_CTL_INTERRUPT |
+	      DP_AUX_CH_CTL_TIME_OUT_ERROR |
+	      DP_AUX_CH_CTL_TIME_OUT_MAX |
+	      DP_AUX_CH_CTL_RECEIVE_ERROR |
+	      (send_bytes << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
+	      DP_AUX_CH_CTL_FW_SYNC_PULSE_SKL(intel_dp_aux_fw_sync_len()) |
+	      DP_AUX_CH_CTL_SYNC_PULSE_SKL(intel_dp_aux_sync_len());
+
+	if (intel_tc_port_in_tbt_alt_mode(dig_port))
+		ret |= DP_AUX_CH_CTL_TBT_IO;
+
+	/*
+	 * Power request bit is already set during aux power well enable.
+	 * Preserve the bit across aux transactions.
+	 */
+	if (DISPLAY_VER(i915) >= 14)
+		ret |= XELPDP_DP_AUX_CH_CTL_POWER_REQUEST;
+
+	return ret;
+}
+
+static int
+intel_dp_aux_xfer(struct intel_dp *intel_dp,
+		  const u8 *send, int send_bytes,
+		  u8 *recv, int recv_size,
+		  u32 aux_send_ctl_flags)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *i915 =
+			to_i915(dig_port->base.base.dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	enum phy phy = intel_port_to_phy(i915, dig_port->base.port);
+	bool is_tc_port = intel_phy_is_tc(i915, phy);
+	i915_reg_t ch_ctl, ch_data[5];
+	u32 aux_clock_divider;
+	enum intel_display_power_domain aux_domain;
+	intel_wakeref_t aux_wakeref;
+	intel_wakeref_t pps_wakeref;
+	int i, ret, recv_bytes;
+	int try, clock = 0;
+	u32 status;
+	bool vdd;
+
+	ch_ctl = intel_dp->aux_ch_ctl_reg(intel_dp);
+	for (i = 0; i < ARRAY_SIZE(ch_data); i++)
+		ch_data[i] = intel_dp->aux_ch_data_reg(intel_dp, i);
+
+	if (is_tc_port)
+		intel_tc_port_lock(dig_port);
+
+	aux_domain = intel_aux_power_domain(dig_port);
+
+	aux_wakeref = intel_display_power_get(i915, aux_domain);
+	pps_wakeref = intel_pps_lock(intel_dp);
+
+	/*
+	 * We will be called with VDD already enabled for dpcd/edid/oui reads.
+	 * In such cases we want to leave VDD enabled and it's up to upper layers
+	 * to turn it off. But for eg. i2c-dev access we need to turn it on/off
+	 * ourselves.
+	 */
+	vdd = intel_pps_vdd_on_unlocked(intel_dp);
+
+	/*
+	 * dp aux is extremely sensitive to irq latency, hence request the
+	 * lowest possible wakeup latency and so prevent the cpu from going into
+	 * deep sleep states.
+	 */
+	cpu_latency_qos_update_request(&intel_dp->pm_qos, 0);
+
+	intel_pps_check_power_unlocked(intel_dp);
+
+	/* Try to wait for any previous AUX channel activity */
+	for (try = 0; try < 3; try++) {
+		status = intel_uncore_read_notrace(uncore, ch_ctl);
+		if ((status & DP_AUX_CH_CTL_SEND_BUSY) == 0)
+			break;
+		msleep(1);
+	}
+	/* just trace the final value */
+	trace_i915_reg_rw(false, ch_ctl, status, sizeof(status), true);
+
+	if (try == 3) {
+		const u32 status = intel_uncore_read(uncore, ch_ctl);
+
+		if (status != intel_dp->aux_busy_last_status) {
+			drm_WARN(&i915->drm, 1,
+				 "%s: not started (status 0x%08x)\n",
+				 intel_dp->aux.name, status);
+			intel_dp->aux_busy_last_status = status;
+		}
+
+		ret = -EBUSY;
+		goto out;
+	}
+
+	/* Only 5 data registers! */
+	if (drm_WARN_ON(&i915->drm, send_bytes > 20 || recv_size > 20)) {
+		ret = -E2BIG;
+		goto out;
+	}
+
+	while ((aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, clock++))) {
+		u32 send_ctl = intel_dp->get_aux_send_ctl(intel_dp,
+							  send_bytes,
+							  aux_clock_divider);
+
+		send_ctl |= aux_send_ctl_flags;
+
+		/* Must try at least 3 times according to DP spec */
+		for (try = 0; try < 5; try++) {
+			/* Load the send data into the aux channel data registers */
+			for (i = 0; i < send_bytes; i += 4)
+				intel_uncore_write(uncore,
+						   ch_data[i >> 2],
+						   intel_dp_aux_pack(send + i,
+								     send_bytes - i));
+
+			/* Send the command and wait for it to complete */
+			intel_uncore_write(uncore, ch_ctl, send_ctl);
+
+			status = intel_dp_aux_wait_done(intel_dp);
+
+			/* Clear done status and any errors */
+			intel_uncore_write(uncore,
+					   ch_ctl,
+					   status |
+					   DP_AUX_CH_CTL_DONE |
+					   DP_AUX_CH_CTL_TIME_OUT_ERROR |
+					   DP_AUX_CH_CTL_RECEIVE_ERROR);
+
+			/*
+			 * DP CTS 1.2 Core Rev 1.1, 4.2.1.1 & 4.2.1.2
+			 *   400us delay required for errors and timeouts
+			 *   Timeout errors from the HW already meet this
+			 *   requirement so skip to next iteration
+			 */
+			if (status & DP_AUX_CH_CTL_TIME_OUT_ERROR)
+				continue;
+
+			if (status & DP_AUX_CH_CTL_RECEIVE_ERROR) {
+				usleep_range(400, 500);
+				continue;
+			}
+			if (status & DP_AUX_CH_CTL_DONE)
+				goto done;
+		}
+	}
+
+	if ((status & DP_AUX_CH_CTL_DONE) == 0) {
+		drm_err(&i915->drm, "%s: not done (status 0x%08x)\n",
+			intel_dp->aux.name, status);
+		ret = -EBUSY;
+		goto out;
+	}
+
+done:
+	/*
+	 * Check for timeout or receive error. Timeouts occur when the sink is
+	 * not connected.
+	 */
+	if (status & DP_AUX_CH_CTL_RECEIVE_ERROR) {
+		drm_err(&i915->drm, "%s: receive error (status 0x%08x)\n",
+			intel_dp->aux.name, status);
+		ret = -EIO;
+		goto out;
+	}
+
+	/*
+	 * Timeouts occur when the device isn't connected, so they're "normal"
+	 * -- don't fill the kernel log with these
+	 */
+	if (status & DP_AUX_CH_CTL_TIME_OUT_ERROR) {
+		drm_dbg_kms(&i915->drm, "%s: timeout (status 0x%08x)\n",
+			    intel_dp->aux.name, status);
+		ret = -ETIMEDOUT;
+		goto out;
+	}
+
+	/* Unload any bytes sent back from the other side */
+	recv_bytes = ((status & DP_AUX_CH_CTL_MESSAGE_SIZE_MASK) >>
+		      DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT);
+
+	/*
+	 * By BSpec: "Message sizes of 0 or >20 are not allowed."
+	 * We have no idea of what happened so we return -EBUSY so
+	 * drm layer takes care for the necessary retries.
+	 */
+	if (recv_bytes == 0 || recv_bytes > 20) {
+		drm_dbg_kms(&i915->drm,
+			    "%s: Forbidden recv_bytes = %d on aux transaction\n",
+			    intel_dp->aux.name, recv_bytes);
+		ret = -EBUSY;
+		goto out;
+	}
+
+	if (recv_bytes > recv_size)
+		recv_bytes = recv_size;
+
+	for (i = 0; i < recv_bytes; i += 4)
+		intel_dp_aux_unpack(intel_uncore_read(uncore, ch_data[i >> 2]),
+				    recv + i, recv_bytes - i);
+
+	ret = recv_bytes;
+out:
+	cpu_latency_qos_update_request(&intel_dp->pm_qos, PM_QOS_DEFAULT_VALUE);
+
+	if (vdd)
+		intel_pps_vdd_off_unlocked(intel_dp, false);
+
+	intel_pps_unlock(intel_dp, pps_wakeref);
+	intel_display_power_put_async(i915, aux_domain, aux_wakeref);
+
+	if (is_tc_port)
+		intel_tc_port_unlock(dig_port);
+
+	return ret;
+}
+
+#define BARE_ADDRESS_SIZE	3
+#define HEADER_SIZE		(BARE_ADDRESS_SIZE + 1)
+
+static void
+intel_dp_aux_header(u8 txbuf[HEADER_SIZE],
+		    const struct drm_dp_aux_msg *msg)
+{
+	txbuf[0] = (msg->request << 4) | ((msg->address >> 16) & 0xf);
+	txbuf[1] = (msg->address >> 8) & 0xff;
+	txbuf[2] = msg->address & 0xff;
+	txbuf[3] = msg->size - 1;
+}
+
+static u32 intel_dp_aux_xfer_flags(const struct drm_dp_aux_msg *msg)
+{
+	/*
+	 * If we're trying to send the HDCP Aksv, we need to set a the Aksv
+	 * select bit to inform the hardware to send the Aksv after our header
+	 * since we can't access that data from software.
+	 */
+	if ((msg->request & ~DP_AUX_I2C_MOT) == DP_AUX_NATIVE_WRITE &&
+	    msg->address == DP_AUX_HDCP_AKSV)
+		return DP_AUX_CH_CTL_AUX_AKSV_SELECT;
+
+	return 0;
+}
+
+static ssize_t
+intel_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg)
+{
+	struct intel_dp *intel_dp = container_of(aux, struct intel_dp, aux);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 txbuf[20], rxbuf[20];
+	size_t txsize, rxsize;
+	u32 flags = intel_dp_aux_xfer_flags(msg);
+	int ret;
+
+	intel_dp_aux_header(txbuf, msg);
+
+	switch (msg->request & ~DP_AUX_I2C_MOT) {
+	case DP_AUX_NATIVE_WRITE:
+	case DP_AUX_I2C_WRITE:
+	case DP_AUX_I2C_WRITE_STATUS_UPDATE:
+		txsize = msg->size ? HEADER_SIZE + msg->size : BARE_ADDRESS_SIZE;
+		rxsize = 2; /* 0 or 1 data bytes */
+
+		if (drm_WARN_ON(&i915->drm, txsize > 20))
+			return -E2BIG;
+
+		drm_WARN_ON(&i915->drm, !msg->buffer != !msg->size);
+
+		if (msg->buffer)
+			memcpy(txbuf + HEADER_SIZE, msg->buffer, msg->size);
+
+		ret = intel_dp_aux_xfer(intel_dp, txbuf, txsize,
+					rxbuf, rxsize, flags);
+		if (ret > 0) {
+			msg->reply = rxbuf[0] >> 4;
+
+			if (ret > 1) {
+				/* Number of bytes written in a short write. */
+				ret = clamp_t(int, rxbuf[1], 0, msg->size);
+			} else {
+				/* Return payload size. */
+				ret = msg->size;
+			}
+		}
+		break;
+
+	case DP_AUX_NATIVE_READ:
+	case DP_AUX_I2C_READ:
+		txsize = msg->size ? HEADER_SIZE : BARE_ADDRESS_SIZE;
+		rxsize = msg->size + 1;
+
+		if (drm_WARN_ON(&i915->drm, rxsize > 20))
+			return -E2BIG;
+
+		ret = intel_dp_aux_xfer(intel_dp, txbuf, txsize,
+					rxbuf, rxsize, flags);
+		if (ret > 0) {
+			msg->reply = rxbuf[0] >> 4;
+			/*
+			 * Assume happy day, and copy the data. The caller is
+			 * expected to check msg->reply before touching it.
+			 *
+			 * Return payload size.
+			 */
+			ret--;
+			memcpy(msg->buffer, rxbuf + 1, ret);
+		}
+		break;
+
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+static i915_reg_t g4x_aux_ctl_reg(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+		return DP_AUX_CH_CTL(aux_ch);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_CTL(AUX_CH_B);
+	}
+}
+
+static i915_reg_t g4x_aux_data_reg(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+		return DP_AUX_CH_DATA(aux_ch, index);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_DATA(AUX_CH_B, index);
+	}
+}
+
+static i915_reg_t ilk_aux_ctl_reg(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+		return DP_AUX_CH_CTL(aux_ch);
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+		return PCH_DP_AUX_CH_CTL(aux_ch);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_CTL(AUX_CH_A);
+	}
+}
+
+static i915_reg_t ilk_aux_data_reg(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+		return DP_AUX_CH_DATA(aux_ch, index);
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+		return PCH_DP_AUX_CH_DATA(aux_ch, index);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_DATA(AUX_CH_A, index);
+	}
+}
+
+static i915_reg_t skl_aux_ctl_reg(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+	case AUX_CH_E:
+	case AUX_CH_F:
+		return DP_AUX_CH_CTL(aux_ch);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_CTL(AUX_CH_A);
+	}
+}
+
+static i915_reg_t skl_aux_data_reg(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_D:
+	case AUX_CH_E:
+	case AUX_CH_F:
+		return DP_AUX_CH_DATA(aux_ch, index);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_DATA(AUX_CH_A, index);
+	}
+}
+
+static i915_reg_t tgl_aux_ctl_reg(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_USBC1:
+	case AUX_CH_USBC2:
+	case AUX_CH_USBC3:
+	case AUX_CH_USBC4:
+	case AUX_CH_USBC5:  /* aka AUX_CH_D_XELPD */
+	case AUX_CH_USBC6:  /* aka AUX_CH_E_XELPD */
+		return DP_AUX_CH_CTL(aux_ch);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_CTL(AUX_CH_A);
+	}
+}
+
+static i915_reg_t tgl_aux_data_reg(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+	case AUX_CH_B:
+	case AUX_CH_C:
+	case AUX_CH_USBC1:
+	case AUX_CH_USBC2:
+	case AUX_CH_USBC3:
+	case AUX_CH_USBC4:
+	case AUX_CH_USBC5:  /* aka AUX_CH_D_XELPD */
+	case AUX_CH_USBC6:  /* aka AUX_CH_E_XELPD */
+		return DP_AUX_CH_DATA(aux_ch, index);
+	default:
+		MISSING_CASE(aux_ch);
+		return DP_AUX_CH_DATA(AUX_CH_A, index);
+	}
+}
+
+static i915_reg_t xelpdp_aux_ctl_reg(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+	case AUX_CH_B:
+	case AUX_CH_USBC1:
+	case AUX_CH_USBC2:
+	case AUX_CH_USBC3:
+	case AUX_CH_USBC4:
+		return XELPDP_DP_AUX_CH_CTL(aux_ch);
+	default:
+		MISSING_CASE(aux_ch);
+		return XELPDP_DP_AUX_CH_CTL(AUX_CH_A);
+	}
+}
+
+static i915_reg_t xelpdp_aux_data_reg(struct intel_dp *intel_dp, int index)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	switch (aux_ch) {
+	case AUX_CH_A:
+	case AUX_CH_B:
+	case AUX_CH_USBC1:
+	case AUX_CH_USBC2:
+	case AUX_CH_USBC3:
+	case AUX_CH_USBC4:
+		return XELPDP_DP_AUX_CH_DATA(aux_ch, index);
+	default:
+		MISSING_CASE(aux_ch);
+		return XELPDP_DP_AUX_CH_DATA(AUX_CH_A, index);
+	}
+}
+
+void intel_dp_aux_fini(struct intel_dp *intel_dp)
+{
+	if (cpu_latency_qos_request_active(&intel_dp->pm_qos))
+		cpu_latency_qos_remove_request(&intel_dp->pm_qos);
+
+	kfree(intel_dp->aux.name);
+}
+
+void intel_dp_aux_init(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct intel_encoder *encoder = &dig_port->base;
+	enum aux_ch aux_ch = dig_port->aux_ch;
+
+	if (DISPLAY_VER(dev_priv) >= 14) {
+		intel_dp->aux_ch_ctl_reg = xelpdp_aux_ctl_reg;
+		intel_dp->aux_ch_data_reg = xelpdp_aux_data_reg;
+	} else if (DISPLAY_VER(dev_priv) >= 12) {
+		intel_dp->aux_ch_ctl_reg = tgl_aux_ctl_reg;
+		intel_dp->aux_ch_data_reg = tgl_aux_data_reg;
+	} else if (DISPLAY_VER(dev_priv) >= 9) {
+		intel_dp->aux_ch_ctl_reg = skl_aux_ctl_reg;
+		intel_dp->aux_ch_data_reg = skl_aux_data_reg;
+	} else if (HAS_PCH_SPLIT(dev_priv)) {
+		intel_dp->aux_ch_ctl_reg = ilk_aux_ctl_reg;
+		intel_dp->aux_ch_data_reg = ilk_aux_data_reg;
+	} else {
+		intel_dp->aux_ch_ctl_reg = g4x_aux_ctl_reg;
+		intel_dp->aux_ch_data_reg = g4x_aux_data_reg;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		intel_dp->get_aux_clock_divider = skl_get_aux_clock_divider;
+	else if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		intel_dp->get_aux_clock_divider = hsw_get_aux_clock_divider;
+	else if (HAS_PCH_SPLIT(dev_priv))
+		intel_dp->get_aux_clock_divider = ilk_get_aux_clock_divider;
+	else
+		intel_dp->get_aux_clock_divider = g4x_get_aux_clock_divider;
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		intel_dp->get_aux_send_ctl = skl_get_aux_send_ctl;
+	else
+		intel_dp->get_aux_send_ctl = g4x_get_aux_send_ctl;
+
+	intel_dp->aux.drm_dev = &dev_priv->drm;
+	drm_dp_aux_init(&intel_dp->aux);
+
+	/* Failure to allocate our preferred name is not critical */
+	if (DISPLAY_VER(dev_priv) >= 13 && aux_ch >= AUX_CH_D_XELPD)
+		intel_dp->aux.name = kasprintf(GFP_KERNEL, "AUX %c/%s",
+					       aux_ch_name(aux_ch - AUX_CH_D_XELPD + AUX_CH_D),
+					       encoder->base.name);
+	else if (DISPLAY_VER(dev_priv) >= 12 && aux_ch >= AUX_CH_USBC1)
+		intel_dp->aux.name = kasprintf(GFP_KERNEL, "AUX USBC%c/%s",
+					       aux_ch - AUX_CH_USBC1 + '1',
+					       encoder->base.name);
+	else
+		intel_dp->aux.name = kasprintf(GFP_KERNEL, "AUX %c/%s",
+					       aux_ch_name(aux_ch),
+					       encoder->base.name);
+
+	intel_dp->aux.transfer = intel_dp_aux_transfer;
+	cpu_latency_qos_add_request(&intel_dp->pm_qos, PM_QOS_DEFAULT_VALUE);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux.h b/drivers/gpu/drm/i915/display/intel_dp_aux.h
new file mode 100644
index 000000000..738577537
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_aux.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020-2021 Intel Corporation
+ */
+
+#ifndef __INTEL_DP_AUX_H__
+#define __INTEL_DP_AUX_H__
+
+struct intel_dp;
+
+void intel_dp_aux_fini(struct intel_dp *intel_dp);
+void intel_dp_aux_init(struct intel_dp *intel_dp);
+
+#endif /* __INTEL_DP_AUX_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c b/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c
new file mode 100644
index 000000000..83af95bce
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.c
@@ -0,0 +1,519 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+/*
+ * Laptops with Intel GPUs which have panels that support controlling the
+ * backlight through DP AUX can actually use two different interfaces: Intel's
+ * proprietary DP AUX backlight interface, and the standard VESA backlight
+ * interface. Unfortunately, at the time of writing this a lot of laptops will
+ * advertise support for the standard VESA backlight interface when they
+ * don't properly support it. However, on these systems the Intel backlight
+ * interface generally does work properly. Additionally, these systems will
+ * usually just indicate that they use PWM backlight controls in their VBIOS
+ * for some reason.
+ */
+
+#include "i915_drv.h"
+#include "intel_backlight.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_dp_aux_backlight.h"
+
+/* TODO:
+ * Implement HDR, right now we just implement the bare minimum to bring us back into SDR mode so we
+ * can make people's backlights work in the mean time
+ */
+
+/*
+ * DP AUX registers for Intel's proprietary HDR backlight interface. We define
+ * them here since we'll likely be the only driver to ever use these.
+ */
+#define INTEL_EDP_HDR_TCON_CAP0                                        0x340
+
+#define INTEL_EDP_HDR_TCON_CAP1                                        0x341
+# define INTEL_EDP_HDR_TCON_2084_DECODE_CAP                           BIT(0)
+# define INTEL_EDP_HDR_TCON_2020_GAMUT_CAP                            BIT(1)
+# define INTEL_EDP_HDR_TCON_TONE_MAPPING_CAP                          BIT(2)
+# define INTEL_EDP_HDR_TCON_SEGMENTED_BACKLIGHT_CAP                   BIT(3)
+# define INTEL_EDP_HDR_TCON_BRIGHTNESS_NITS_CAP                       BIT(4)
+# define INTEL_EDP_HDR_TCON_OPTIMIZATION_CAP                          BIT(5)
+# define INTEL_EDP_HDR_TCON_SDP_COLORIMETRY_CAP                       BIT(6)
+# define INTEL_EDP_HDR_TCON_SRGB_TO_PANEL_GAMUT_CONVERSION_CAP        BIT(7)
+
+#define INTEL_EDP_HDR_TCON_CAP2                                        0x342
+# define INTEL_EDP_SDR_TCON_BRIGHTNESS_AUX_CAP                        BIT(0)
+
+#define INTEL_EDP_HDR_TCON_CAP3                                        0x343
+
+#define INTEL_EDP_HDR_GETSET_CTRL_PARAMS                               0x344
+# define INTEL_EDP_HDR_TCON_2084_DECODE_ENABLE                        BIT(0)
+# define INTEL_EDP_HDR_TCON_2020_GAMUT_ENABLE                         BIT(1)
+# define INTEL_EDP_HDR_TCON_TONE_MAPPING_ENABLE                       BIT(2) /* Pre-TGL+ */
+# define INTEL_EDP_HDR_TCON_SEGMENTED_BACKLIGHT_ENABLE                BIT(3)
+# define INTEL_EDP_HDR_TCON_BRIGHTNESS_AUX_ENABLE                     BIT(4)
+# define INTEL_EDP_HDR_TCON_SRGB_TO_PANEL_GAMUT_ENABLE                BIT(5)
+/* Bit 6 is reserved */
+# define INTEL_EDP_HDR_TCON_SDP_COLORIMETRY_ENABLE                    BIT(7)
+
+#define INTEL_EDP_HDR_CONTENT_LUMINANCE                                0x346 /* Pre-TGL+ */
+#define INTEL_EDP_HDR_PANEL_LUMINANCE_OVERRIDE                         0x34A
+#define INTEL_EDP_SDR_LUMINANCE_LEVEL                                  0x352
+#define INTEL_EDP_BRIGHTNESS_NITS_LSB                                  0x354
+#define INTEL_EDP_BRIGHTNESS_NITS_MSB                                  0x355
+#define INTEL_EDP_BRIGHTNESS_DELAY_FRAMES                              0x356
+#define INTEL_EDP_BRIGHTNESS_PER_FRAME_STEPS                           0x357
+
+#define INTEL_EDP_BRIGHTNESS_OPTIMIZATION_0                            0x358
+# define INTEL_EDP_TCON_USAGE_MASK                             GENMASK(0, 3)
+# define INTEL_EDP_TCON_USAGE_UNKNOWN                                    0x0
+# define INTEL_EDP_TCON_USAGE_DESKTOP                                    0x1
+# define INTEL_EDP_TCON_USAGE_FULL_SCREEN_MEDIA                          0x2
+# define INTEL_EDP_TCON_USAGE_FULL_SCREEN_GAMING                         0x3
+# define INTEL_EDP_TCON_POWER_MASK                                    BIT(4)
+# define INTEL_EDP_TCON_POWER_DC                                    (0 << 4)
+# define INTEL_EDP_TCON_POWER_AC                                    (1 << 4)
+# define INTEL_EDP_TCON_OPTIMIZATION_STRENGTH_MASK             GENMASK(5, 7)
+
+#define INTEL_EDP_BRIGHTNESS_OPTIMIZATION_1                            0x359
+
+enum intel_dp_aux_backlight_modparam {
+	INTEL_DP_AUX_BACKLIGHT_AUTO = -1,
+	INTEL_DP_AUX_BACKLIGHT_OFF = 0,
+	INTEL_DP_AUX_BACKLIGHT_ON = 1,
+	INTEL_DP_AUX_BACKLIGHT_FORCE_VESA = 2,
+	INTEL_DP_AUX_BACKLIGHT_FORCE_INTEL = 3,
+};
+
+/* Intel EDP backlight callbacks */
+static bool
+intel_dp_aux_supports_hdr_backlight(struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+	struct drm_dp_aux *aux = &intel_dp->aux;
+	struct intel_panel *panel = &connector->panel;
+	int ret;
+	u8 tcon_cap[4];
+
+	intel_dp_wait_source_oui(intel_dp);
+
+	ret = drm_dp_dpcd_read(aux, INTEL_EDP_HDR_TCON_CAP0, tcon_cap, sizeof(tcon_cap));
+	if (ret != sizeof(tcon_cap))
+		return false;
+
+	if (!(tcon_cap[1] & INTEL_EDP_HDR_TCON_BRIGHTNESS_NITS_CAP))
+		return false;
+
+	if (tcon_cap[0] >= 1) {
+		drm_dbg_kms(&i915->drm, "Detected Intel HDR backlight interface version %d\n",
+			    tcon_cap[0]);
+	} else {
+		drm_dbg_kms(&i915->drm, "Detected unsupported HDR backlight interface version %d\n",
+			    tcon_cap[0]);
+		return false;
+	}
+
+	/*
+	 * If we don't have HDR static metadata there is no way to
+	 * runtime detect used range for nits based control. For now
+	 * do not use Intel proprietary eDP backlight control if we
+	 * don't have this data in panel EDID. In case we find panel
+	 * which supports only nits based control, but doesn't provide
+	 * HDR static metadata we need to start maintaining table of
+	 * ranges for such panels.
+	 */
+	if (i915->params.enable_dpcd_backlight != INTEL_DP_AUX_BACKLIGHT_FORCE_INTEL &&
+	    !(connector->base.hdr_sink_metadata.hdmi_type1.metadata_type &
+	      BIT(HDMI_STATIC_METADATA_TYPE1))) {
+		drm_info(&i915->drm,
+			 "Panel is missing HDR static metadata. Possible support for Intel HDR backlight interface is not used. If your backlight controls don't work try booting with i915.enable_dpcd_backlight=%d. needs this, please file a _new_ bug report on drm/i915, see " FDO_BUG_URL " for details.\n",
+			 INTEL_DP_AUX_BACKLIGHT_FORCE_INTEL);
+		return false;
+	}
+
+	panel->backlight.edp.intel.sdr_uses_aux =
+		tcon_cap[2] & INTEL_EDP_SDR_TCON_BRIGHTNESS_AUX_CAP;
+
+	return true;
+}
+
+static u32
+intel_dp_aux_hdr_get_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+	u8 tmp;
+	u8 buf[2] = { 0 };
+
+	if (drm_dp_dpcd_readb(&intel_dp->aux, INTEL_EDP_HDR_GETSET_CTRL_PARAMS, &tmp) != 1) {
+		drm_err(&i915->drm, "Failed to read current backlight mode from DPCD\n");
+		return 0;
+	}
+
+	if (!(tmp & INTEL_EDP_HDR_TCON_BRIGHTNESS_AUX_ENABLE)) {
+		if (!panel->backlight.edp.intel.sdr_uses_aux) {
+			u32 pwm_level = panel->backlight.pwm_funcs->get(connector, pipe);
+
+			return intel_backlight_level_from_pwm(connector, pwm_level);
+		}
+
+		/* Assume 100% brightness if backlight controls aren't enabled yet */
+		return panel->backlight.max;
+	}
+
+	if (drm_dp_dpcd_read(&intel_dp->aux, INTEL_EDP_BRIGHTNESS_NITS_LSB, buf,
+			     sizeof(buf)) != sizeof(buf)) {
+		drm_err(&i915->drm, "Failed to read brightness from DPCD\n");
+		return 0;
+	}
+
+	return (buf[1] << 8 | buf[0]);
+}
+
+static void
+intel_dp_aux_hdr_set_aux_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_device *dev = connector->base.dev;
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+	u8 buf[4] = { 0 };
+
+	buf[0] = level & 0xFF;
+	buf[1] = (level & 0xFF00) >> 8;
+
+	if (drm_dp_dpcd_write(&intel_dp->aux, INTEL_EDP_BRIGHTNESS_NITS_LSB, buf,
+			      sizeof(buf)) != sizeof(buf))
+		drm_err(dev, "Failed to write brightness level to DPCD\n");
+}
+
+static void
+intel_dp_aux_hdr_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	if (panel->backlight.edp.intel.sdr_uses_aux) {
+		intel_dp_aux_hdr_set_aux_backlight(conn_state, level);
+	} else {
+		const u32 pwm_level = intel_backlight_level_to_pwm(connector, level);
+
+		intel_backlight_set_pwm_level(conn_state, pwm_level);
+	}
+}
+
+static void
+intel_dp_aux_hdr_enable_backlight(const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+	int ret;
+	u8 old_ctrl, ctrl;
+
+	intel_dp_wait_source_oui(intel_dp);
+
+	ret = drm_dp_dpcd_readb(&intel_dp->aux, INTEL_EDP_HDR_GETSET_CTRL_PARAMS, &old_ctrl);
+	if (ret != 1) {
+		drm_err(&i915->drm, "Failed to read current backlight control mode: %d\n", ret);
+		return;
+	}
+
+	ctrl = old_ctrl;
+	if (panel->backlight.edp.intel.sdr_uses_aux) {
+		ctrl |= INTEL_EDP_HDR_TCON_BRIGHTNESS_AUX_ENABLE;
+		intel_dp_aux_hdr_set_aux_backlight(conn_state, level);
+	} else {
+		u32 pwm_level = intel_backlight_level_to_pwm(connector, level);
+
+		panel->backlight.pwm_funcs->enable(crtc_state, conn_state, pwm_level);
+
+		ctrl &= ~INTEL_EDP_HDR_TCON_BRIGHTNESS_AUX_ENABLE;
+	}
+
+	if (ctrl != old_ctrl)
+		if (drm_dp_dpcd_writeb(&intel_dp->aux, INTEL_EDP_HDR_GETSET_CTRL_PARAMS, ctrl) != 1)
+			drm_err(&i915->drm, "Failed to configure DPCD brightness controls\n");
+}
+
+static void
+intel_dp_aux_hdr_disable_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+
+	/* Nothing to do for AUX based backlight controls */
+	if (panel->backlight.edp.intel.sdr_uses_aux)
+		return;
+
+	/* Note we want the actual pwm_level to be 0, regardless of pwm_min */
+	panel->backlight.pwm_funcs->disable(conn_state, intel_backlight_invert_pwm_level(connector, 0));
+}
+
+static int
+intel_dp_aux_hdr_setup_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_panel *panel = &connector->panel;
+	struct drm_luminance_range_info *luminance_range =
+		&connector->base.display_info.luminance_range;
+	int ret;
+
+	if (panel->backlight.edp.intel.sdr_uses_aux) {
+		drm_dbg_kms(&i915->drm, "SDR backlight is controlled through DPCD\n");
+	} else {
+		drm_dbg_kms(&i915->drm, "SDR backlight is controlled through PWM\n");
+
+		ret = panel->backlight.pwm_funcs->setup(connector, pipe);
+		if (ret < 0) {
+			drm_err(&i915->drm,
+				"Failed to setup SDR backlight controls through PWM: %d\n", ret);
+			return ret;
+		}
+	}
+
+	if (luminance_range->max_luminance) {
+		panel->backlight.max = luminance_range->max_luminance;
+		panel->backlight.min = luminance_range->min_luminance;
+	} else {
+		panel->backlight.max = 512;
+		panel->backlight.min = 0;
+	}
+
+	drm_dbg_kms(&i915->drm, "Using backlight range %d..%d\n", panel->backlight.min,
+		    panel->backlight.max);
+
+	panel->backlight.level = intel_dp_aux_hdr_get_backlight(connector, pipe);
+	panel->backlight.enabled = panel->backlight.level != 0;
+
+	return 0;
+}
+
+/* VESA backlight callbacks */
+static u32 intel_dp_aux_vesa_get_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	return connector->panel.backlight.level;
+}
+
+static void
+intel_dp_aux_vesa_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+
+	if (!panel->backlight.edp.vesa.info.aux_set) {
+		const u32 pwm_level = intel_backlight_level_to_pwm(connector, level);
+
+		intel_backlight_set_pwm_level(conn_state, pwm_level);
+	}
+
+	drm_edp_backlight_set_level(&intel_dp->aux, &panel->backlight.edp.vesa.info, level);
+}
+
+static void
+intel_dp_aux_vesa_enable_backlight(const struct intel_crtc_state *crtc_state,
+				   const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+
+	if (!panel->backlight.edp.vesa.info.aux_enable) {
+		u32 pwm_level;
+
+		if (!panel->backlight.edp.vesa.info.aux_set)
+			pwm_level = intel_backlight_level_to_pwm(connector, level);
+		else
+			pwm_level = intel_backlight_invert_pwm_level(connector,
+								     panel->backlight.pwm_level_max);
+
+		panel->backlight.pwm_funcs->enable(crtc_state, conn_state, pwm_level);
+	}
+
+	drm_edp_backlight_enable(&intel_dp->aux, &panel->backlight.edp.vesa.info, level);
+}
+
+static void intel_dp_aux_vesa_disable_backlight(const struct drm_connector_state *old_conn_state,
+						u32 level)
+{
+	struct intel_connector *connector = to_intel_connector(old_conn_state->connector);
+	struct intel_panel *panel = &connector->panel;
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+
+	drm_edp_backlight_disable(&intel_dp->aux, &panel->backlight.edp.vesa.info);
+
+	if (!panel->backlight.edp.vesa.info.aux_enable)
+		panel->backlight.pwm_funcs->disable(old_conn_state,
+						    intel_backlight_invert_pwm_level(connector, 0));
+}
+
+static int intel_dp_aux_vesa_setup_backlight(struct intel_connector *connector, enum pipe pipe)
+{
+	struct intel_dp *intel_dp = intel_attached_dp(connector);
+	struct intel_panel *panel = &connector->panel;
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u16 current_level;
+	u8 current_mode;
+	int ret;
+
+	ret = drm_edp_backlight_init(&intel_dp->aux, &panel->backlight.edp.vesa.info,
+				     panel->vbt.backlight.pwm_freq_hz, intel_dp->edp_dpcd,
+				     &current_level, &current_mode);
+	if (ret < 0)
+		return ret;
+
+	if (!panel->backlight.edp.vesa.info.aux_set || !panel->backlight.edp.vesa.info.aux_enable) {
+		ret = panel->backlight.pwm_funcs->setup(connector, pipe);
+		if (ret < 0) {
+			drm_err(&i915->drm,
+				"Failed to setup PWM backlight controls for eDP backlight: %d\n",
+				ret);
+			return ret;
+		}
+	}
+
+	if (panel->backlight.edp.vesa.info.aux_set) {
+		panel->backlight.max = panel->backlight.edp.vesa.info.max;
+		panel->backlight.min = 0;
+		if (current_mode == DP_EDP_BACKLIGHT_CONTROL_MODE_DPCD) {
+			panel->backlight.level = current_level;
+			panel->backlight.enabled = panel->backlight.level != 0;
+		} else {
+			panel->backlight.level = panel->backlight.max;
+			panel->backlight.enabled = false;
+		}
+	} else {
+		panel->backlight.max = panel->backlight.pwm_level_max;
+		panel->backlight.min = panel->backlight.pwm_level_min;
+		if (current_mode == DP_EDP_BACKLIGHT_CONTROL_MODE_PWM) {
+			panel->backlight.level = panel->backlight.pwm_funcs->get(connector, pipe);
+			panel->backlight.enabled = panel->backlight.pwm_enabled;
+		} else {
+			panel->backlight.level = panel->backlight.max;
+			panel->backlight.enabled = false;
+		}
+	}
+
+	return 0;
+}
+
+static bool
+intel_dp_aux_supports_vesa_backlight(struct intel_connector *connector)
+{
+	struct intel_dp *intel_dp = intel_attached_dp(connector);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	if (drm_edp_backlight_supported(intel_dp->edp_dpcd)) {
+		drm_dbg_kms(&i915->drm, "AUX Backlight Control Supported!\n");
+		return true;
+	}
+	return false;
+}
+
+static const struct intel_panel_bl_funcs intel_dp_hdr_bl_funcs = {
+	.setup = intel_dp_aux_hdr_setup_backlight,
+	.enable = intel_dp_aux_hdr_enable_backlight,
+	.disable = intel_dp_aux_hdr_disable_backlight,
+	.set = intel_dp_aux_hdr_set_backlight,
+	.get = intel_dp_aux_hdr_get_backlight,
+};
+
+static const struct intel_panel_bl_funcs intel_dp_vesa_bl_funcs = {
+	.setup = intel_dp_aux_vesa_setup_backlight,
+	.enable = intel_dp_aux_vesa_enable_backlight,
+	.disable = intel_dp_aux_vesa_disable_backlight,
+	.set = intel_dp_aux_vesa_set_backlight,
+	.get = intel_dp_aux_vesa_get_backlight,
+};
+
+int intel_dp_aux_init_backlight_funcs(struct intel_connector *connector)
+{
+	struct drm_device *dev = connector->base.dev;
+	struct intel_panel *panel = &connector->panel;
+	struct intel_dp *intel_dp = enc_to_intel_dp(connector->encoder);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	bool try_intel_interface = false, try_vesa_interface = false;
+
+	/* Check the VBT and user's module parameters to figure out which
+	 * interfaces to probe
+	 */
+	switch (i915->params.enable_dpcd_backlight) {
+	case INTEL_DP_AUX_BACKLIGHT_OFF:
+		return -ENODEV;
+	case INTEL_DP_AUX_BACKLIGHT_AUTO:
+		switch (panel->vbt.backlight.type) {
+		case INTEL_BACKLIGHT_VESA_EDP_AUX_INTERFACE:
+			try_vesa_interface = true;
+			break;
+		case INTEL_BACKLIGHT_DISPLAY_DDI:
+			try_intel_interface = true;
+			break;
+		default:
+			return -ENODEV;
+		}
+		break;
+	case INTEL_DP_AUX_BACKLIGHT_ON:
+		if (panel->vbt.backlight.type != INTEL_BACKLIGHT_VESA_EDP_AUX_INTERFACE)
+			try_intel_interface = true;
+
+		try_vesa_interface = true;
+		break;
+	case INTEL_DP_AUX_BACKLIGHT_FORCE_VESA:
+		try_vesa_interface = true;
+		break;
+	case INTEL_DP_AUX_BACKLIGHT_FORCE_INTEL:
+		try_intel_interface = true;
+		break;
+	}
+
+	/*
+	 * Since Intel has their own backlight control interface, the majority of machines out there
+	 * using DPCD backlight controls with Intel GPUs will be using this interface as opposed to
+	 * the VESA interface. However, other GPUs (such as Nvidia's) will always use the VESA
+	 * interface. This means that there's quite a number of panels out there that will advertise
+	 * support for both interfaces, primarily systems with Intel/Nvidia hybrid GPU setups.
+	 *
+	 * There's a catch to this though: on many panels that advertise support for both
+	 * interfaces, the VESA backlight interface will stop working once we've programmed the
+	 * panel with Intel's OUI - which is also required for us to be able to detect Intel's
+	 * backlight interface at all. This means that the only sensible way for us to detect both
+	 * interfaces is to probe for Intel's first, and VESA's second.
+	 */
+	if (try_intel_interface && intel_dp_aux_supports_hdr_backlight(connector)) {
+		drm_dbg_kms(dev, "Using Intel proprietary eDP backlight controls\n");
+		panel->backlight.funcs = &intel_dp_hdr_bl_funcs;
+		return 0;
+	}
+
+	if (try_vesa_interface && intel_dp_aux_supports_vesa_backlight(connector)) {
+		drm_dbg_kms(dev, "Using VESA eDP backlight controls\n");
+		panel->backlight.funcs = &intel_dp_vesa_bl_funcs;
+		return 0;
+	}
+
+	return -ENODEV;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.h b/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.h
new file mode 100644
index 000000000..ed60c2858
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_aux_backlight.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DP_AUX_BACKLIGHT_H__
+#define __INTEL_DP_AUX_BACKLIGHT_H__
+
+struct intel_connector;
+
+int intel_dp_aux_init_backlight_funcs(struct intel_connector *intel_connector);
+
+#endif /* __INTEL_DP_AUX_BACKLIGHT_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dp_hdcp.c b/drivers/gpu/drm/i915/display/intel_dp_hdcp.c
new file mode 100644
index 000000000..88689124c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_hdcp.c
@@ -0,0 +1,823 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright (C) 2020 Google, Inc.
+ *
+ * Authors:
+ * Sean Paul <seanpaul@chromium.org>
+ */
+
+#include <drm/display/drm_dp_helper.h>
+#include <drm/display/drm_dp_mst_helper.h>
+#include <drm/display/drm_hdcp_helper.h>
+#include <drm/drm_print.h>
+
+#include "intel_ddi.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_dp_hdcp.h"
+#include "intel_hdcp.h"
+#include "intel_hdcp_regs.h"
+
+static unsigned int transcoder_to_stream_enc_status(enum transcoder cpu_transcoder)
+{
+	u32 stream_enc_mask;
+
+	switch (cpu_transcoder) {
+	case TRANSCODER_A:
+		stream_enc_mask = HDCP_STATUS_STREAM_A_ENC;
+		break;
+	case TRANSCODER_B:
+		stream_enc_mask = HDCP_STATUS_STREAM_B_ENC;
+		break;
+	case TRANSCODER_C:
+		stream_enc_mask = HDCP_STATUS_STREAM_C_ENC;
+		break;
+	case TRANSCODER_D:
+		stream_enc_mask = HDCP_STATUS_STREAM_D_ENC;
+		break;
+	default:
+		stream_enc_mask = 0;
+	}
+
+	return stream_enc_mask;
+}
+
+static void intel_dp_hdcp_wait_for_cp_irq(struct intel_hdcp *hdcp, int timeout)
+{
+	long ret;
+
+#define C (hdcp->cp_irq_count_cached != atomic_read(&hdcp->cp_irq_count))
+	ret = wait_event_interruptible_timeout(hdcp->cp_irq_queue, C,
+					       msecs_to_jiffies(timeout));
+
+	if (!ret)
+		DRM_DEBUG_KMS("Timedout at waiting for CP_IRQ\n");
+}
+
+static
+int intel_dp_hdcp_write_an_aksv(struct intel_digital_port *dig_port,
+				u8 *an)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	u8 aksv[DRM_HDCP_KSV_LEN] = {};
+	ssize_t dpcd_ret;
+
+	/* Output An first, that's easy */
+	dpcd_ret = drm_dp_dpcd_write(&dig_port->dp.aux, DP_AUX_HDCP_AN,
+				     an, DRM_HDCP_AN_LEN);
+	if (dpcd_ret != DRM_HDCP_AN_LEN) {
+		drm_dbg_kms(&i915->drm,
+			    "Failed to write An over DP/AUX (%zd)\n",
+			    dpcd_ret);
+		return dpcd_ret >= 0 ? -EIO : dpcd_ret;
+	}
+
+	/*
+	 * Since Aksv is Oh-So-Secret, we can't access it in software. So we
+	 * send an empty buffer of the correct length through the DP helpers. On
+	 * the other side, in the transfer hook, we'll generate a flag based on
+	 * the destination address which will tickle the hardware to output the
+	 * Aksv on our behalf after the header is sent.
+	 */
+	dpcd_ret = drm_dp_dpcd_write(&dig_port->dp.aux, DP_AUX_HDCP_AKSV,
+				     aksv, DRM_HDCP_KSV_LEN);
+	if (dpcd_ret != DRM_HDCP_KSV_LEN) {
+		drm_dbg_kms(&i915->drm,
+			    "Failed to write Aksv over DP/AUX (%zd)\n",
+			    dpcd_ret);
+		return dpcd_ret >= 0 ? -EIO : dpcd_ret;
+	}
+	return 0;
+}
+
+static int intel_dp_hdcp_read_bksv(struct intel_digital_port *dig_port,
+				   u8 *bksv)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	ssize_t ret;
+
+	ret = drm_dp_dpcd_read(&dig_port->dp.aux, DP_AUX_HDCP_BKSV, bksv,
+			       DRM_HDCP_KSV_LEN);
+	if (ret != DRM_HDCP_KSV_LEN) {
+		drm_dbg_kms(&i915->drm,
+			    "Read Bksv from DP/AUX failed (%zd)\n", ret);
+		return ret >= 0 ? -EIO : ret;
+	}
+	return 0;
+}
+
+static int intel_dp_hdcp_read_bstatus(struct intel_digital_port *dig_port,
+				      u8 *bstatus)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	ssize_t ret;
+
+	/*
+	 * For some reason the HDMI and DP HDCP specs call this register
+	 * definition by different names. In the HDMI spec, it's called BSTATUS,
+	 * but in DP it's called BINFO.
+	 */
+	ret = drm_dp_dpcd_read(&dig_port->dp.aux, DP_AUX_HDCP_BINFO,
+			       bstatus, DRM_HDCP_BSTATUS_LEN);
+	if (ret != DRM_HDCP_BSTATUS_LEN) {
+		drm_dbg_kms(&i915->drm,
+			    "Read bstatus from DP/AUX failed (%zd)\n", ret);
+		return ret >= 0 ? -EIO : ret;
+	}
+	return 0;
+}
+
+static
+int intel_dp_hdcp_read_bcaps(struct intel_digital_port *dig_port,
+			     u8 *bcaps)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	ssize_t ret;
+
+	ret = drm_dp_dpcd_read(&dig_port->dp.aux, DP_AUX_HDCP_BCAPS,
+			       bcaps, 1);
+	if (ret != 1) {
+		drm_dbg_kms(&i915->drm,
+			    "Read bcaps from DP/AUX failed (%zd)\n", ret);
+		return ret >= 0 ? -EIO : ret;
+	}
+
+	return 0;
+}
+
+static
+int intel_dp_hdcp_repeater_present(struct intel_digital_port *dig_port,
+				   bool *repeater_present)
+{
+	ssize_t ret;
+	u8 bcaps;
+
+	ret = intel_dp_hdcp_read_bcaps(dig_port, &bcaps);
+	if (ret)
+		return ret;
+
+	*repeater_present = bcaps & DP_BCAPS_REPEATER_PRESENT;
+	return 0;
+}
+
+static
+int intel_dp_hdcp_read_ri_prime(struct intel_digital_port *dig_port,
+				u8 *ri_prime)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	ssize_t ret;
+
+	ret = drm_dp_dpcd_read(&dig_port->dp.aux, DP_AUX_HDCP_RI_PRIME,
+			       ri_prime, DRM_HDCP_RI_LEN);
+	if (ret != DRM_HDCP_RI_LEN) {
+		drm_dbg_kms(&i915->drm, "Read Ri' from DP/AUX failed (%zd)\n",
+			    ret);
+		return ret >= 0 ? -EIO : ret;
+	}
+	return 0;
+}
+
+static
+int intel_dp_hdcp_read_ksv_ready(struct intel_digital_port *dig_port,
+				 bool *ksv_ready)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	ssize_t ret;
+	u8 bstatus;
+
+	ret = drm_dp_dpcd_read(&dig_port->dp.aux, DP_AUX_HDCP_BSTATUS,
+			       &bstatus, 1);
+	if (ret != 1) {
+		drm_dbg_kms(&i915->drm,
+			    "Read bstatus from DP/AUX failed (%zd)\n", ret);
+		return ret >= 0 ? -EIO : ret;
+	}
+	*ksv_ready = bstatus & DP_BSTATUS_READY;
+	return 0;
+}
+
+static
+int intel_dp_hdcp_read_ksv_fifo(struct intel_digital_port *dig_port,
+				int num_downstream, u8 *ksv_fifo)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	ssize_t ret;
+	int i;
+
+	/* KSV list is read via 15 byte window (3 entries @ 5 bytes each) */
+	for (i = 0; i < num_downstream; i += 3) {
+		size_t len = min(num_downstream - i, 3) * DRM_HDCP_KSV_LEN;
+		ret = drm_dp_dpcd_read(&dig_port->dp.aux,
+				       DP_AUX_HDCP_KSV_FIFO,
+				       ksv_fifo + i * DRM_HDCP_KSV_LEN,
+				       len);
+		if (ret != len) {
+			drm_dbg_kms(&i915->drm,
+				    "Read ksv[%d] from DP/AUX failed (%zd)\n",
+				    i, ret);
+			return ret >= 0 ? -EIO : ret;
+		}
+	}
+	return 0;
+}
+
+static
+int intel_dp_hdcp_read_v_prime_part(struct intel_digital_port *dig_port,
+				    int i, u32 *part)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	ssize_t ret;
+
+	if (i >= DRM_HDCP_V_PRIME_NUM_PARTS)
+		return -EINVAL;
+
+	ret = drm_dp_dpcd_read(&dig_port->dp.aux,
+			       DP_AUX_HDCP_V_PRIME(i), part,
+			       DRM_HDCP_V_PRIME_PART_LEN);
+	if (ret != DRM_HDCP_V_PRIME_PART_LEN) {
+		drm_dbg_kms(&i915->drm,
+			    "Read v'[%d] from DP/AUX failed (%zd)\n", i, ret);
+		return ret >= 0 ? -EIO : ret;
+	}
+	return 0;
+}
+
+static
+int intel_dp_hdcp_toggle_signalling(struct intel_digital_port *dig_port,
+				    enum transcoder cpu_transcoder,
+				    bool enable)
+{
+	/* Not used for single stream DisplayPort setups */
+	return 0;
+}
+
+static
+bool intel_dp_hdcp_check_link(struct intel_digital_port *dig_port,
+			      struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	ssize_t ret;
+	u8 bstatus;
+
+	ret = drm_dp_dpcd_read(&dig_port->dp.aux, DP_AUX_HDCP_BSTATUS,
+			       &bstatus, 1);
+	if (ret != 1) {
+		drm_dbg_kms(&i915->drm,
+			    "Read bstatus from DP/AUX failed (%zd)\n", ret);
+		return false;
+	}
+
+	return !(bstatus & (DP_BSTATUS_LINK_FAILURE | DP_BSTATUS_REAUTH_REQ));
+}
+
+static
+int intel_dp_hdcp_capable(struct intel_digital_port *dig_port,
+			  bool *hdcp_capable)
+{
+	ssize_t ret;
+	u8 bcaps;
+
+	ret = intel_dp_hdcp_read_bcaps(dig_port, &bcaps);
+	if (ret)
+		return ret;
+
+	*hdcp_capable = bcaps & DP_BCAPS_HDCP_CAPABLE;
+	return 0;
+}
+
+struct hdcp2_dp_errata_stream_type {
+	u8	msg_id;
+	u8	stream_type;
+} __packed;
+
+struct hdcp2_dp_msg_data {
+	u8 msg_id;
+	u32 offset;
+	bool msg_detectable;
+	u32 timeout;
+	u32 timeout2; /* Added for non_paired situation */
+	/* Timeout to read entire msg */
+	u32 msg_read_timeout;
+};
+
+static const struct hdcp2_dp_msg_data hdcp2_dp_msg_data[] = {
+	{ HDCP_2_2_AKE_INIT, DP_HDCP_2_2_AKE_INIT_OFFSET, false, 0, 0, 0},
+	{ HDCP_2_2_AKE_SEND_CERT, DP_HDCP_2_2_AKE_SEND_CERT_OFFSET,
+	  false, HDCP_2_2_CERT_TIMEOUT_MS, 0, HDCP_2_2_DP_CERT_READ_TIMEOUT_MS},
+	{ HDCP_2_2_AKE_NO_STORED_KM, DP_HDCP_2_2_AKE_NO_STORED_KM_OFFSET,
+	  false, 0, 0, 0 },
+	{ HDCP_2_2_AKE_STORED_KM, DP_HDCP_2_2_AKE_STORED_KM_OFFSET,
+	  false, 0, 0, 0 },
+	{ HDCP_2_2_AKE_SEND_HPRIME, DP_HDCP_2_2_AKE_SEND_HPRIME_OFFSET,
+	  true, HDCP_2_2_HPRIME_PAIRED_TIMEOUT_MS,
+	  HDCP_2_2_HPRIME_NO_PAIRED_TIMEOUT_MS, HDCP_2_2_DP_HPRIME_READ_TIMEOUT_MS},
+	{ HDCP_2_2_AKE_SEND_PAIRING_INFO,
+	  DP_HDCP_2_2_AKE_SEND_PAIRING_INFO_OFFSET, true,
+	  HDCP_2_2_PAIRING_TIMEOUT_MS, 0, HDCP_2_2_DP_PAIRING_READ_TIMEOUT_MS },
+	{ HDCP_2_2_LC_INIT, DP_HDCP_2_2_LC_INIT_OFFSET, false, 0, 0, 0 },
+	{ HDCP_2_2_LC_SEND_LPRIME, DP_HDCP_2_2_LC_SEND_LPRIME_OFFSET,
+	  false, HDCP_2_2_DP_LPRIME_TIMEOUT_MS, 0, 0 },
+	{ HDCP_2_2_SKE_SEND_EKS, DP_HDCP_2_2_SKE_SEND_EKS_OFFSET, false,
+	  0, 0, 0 },
+	{ HDCP_2_2_REP_SEND_RECVID_LIST,
+	  DP_HDCP_2_2_REP_SEND_RECVID_LIST_OFFSET, true,
+	  HDCP_2_2_RECVID_LIST_TIMEOUT_MS, 0, 0 },
+	{ HDCP_2_2_REP_SEND_ACK, DP_HDCP_2_2_REP_SEND_ACK_OFFSET, false,
+	  0, 0, 0 },
+	{ HDCP_2_2_REP_STREAM_MANAGE,
+	  DP_HDCP_2_2_REP_STREAM_MANAGE_OFFSET, false,
+	  0, 0, 0},
+	{ HDCP_2_2_REP_STREAM_READY, DP_HDCP_2_2_REP_STREAM_READY_OFFSET,
+	  false, HDCP_2_2_STREAM_READY_TIMEOUT_MS, 0, 0 },
+/* local define to shovel this through the write_2_2 interface */
+#define HDCP_2_2_ERRATA_DP_STREAM_TYPE	50
+	{ HDCP_2_2_ERRATA_DP_STREAM_TYPE,
+	  DP_HDCP_2_2_REG_STREAM_TYPE_OFFSET, false,
+	  0, 0 },
+};
+
+static int
+intel_dp_hdcp2_read_rx_status(struct intel_digital_port *dig_port,
+			      u8 *rx_status)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	ssize_t ret;
+
+	ret = drm_dp_dpcd_read(&dig_port->dp.aux,
+			       DP_HDCP_2_2_REG_RXSTATUS_OFFSET, rx_status,
+			       HDCP_2_2_DP_RXSTATUS_LEN);
+	if (ret != HDCP_2_2_DP_RXSTATUS_LEN) {
+		drm_dbg_kms(&i915->drm,
+			    "Read bstatus from DP/AUX failed (%zd)\n", ret);
+		return ret >= 0 ? -EIO : ret;
+	}
+
+	return 0;
+}
+
+static
+int hdcp2_detect_msg_availability(struct intel_digital_port *dig_port,
+				  u8 msg_id, bool *msg_ready)
+{
+	u8 rx_status;
+	int ret;
+
+	*msg_ready = false;
+	ret = intel_dp_hdcp2_read_rx_status(dig_port, &rx_status);
+	if (ret < 0)
+		return ret;
+
+	switch (msg_id) {
+	case HDCP_2_2_AKE_SEND_HPRIME:
+		if (HDCP_2_2_DP_RXSTATUS_H_PRIME(rx_status))
+			*msg_ready = true;
+		break;
+	case HDCP_2_2_AKE_SEND_PAIRING_INFO:
+		if (HDCP_2_2_DP_RXSTATUS_PAIRING(rx_status))
+			*msg_ready = true;
+		break;
+	case HDCP_2_2_REP_SEND_RECVID_LIST:
+		if (HDCP_2_2_DP_RXSTATUS_READY(rx_status))
+			*msg_ready = true;
+		break;
+	default:
+		DRM_ERROR("Unidentified msg_id: %d\n", msg_id);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static ssize_t
+intel_dp_hdcp2_wait_for_msg(struct intel_digital_port *dig_port,
+			    const struct hdcp2_dp_msg_data *hdcp2_msg_data)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_dp *dp = &dig_port->dp;
+	struct intel_hdcp *hdcp = &dp->attached_connector->hdcp;
+	u8 msg_id = hdcp2_msg_data->msg_id;
+	int ret, timeout;
+	bool msg_ready = false;
+
+	if (msg_id == HDCP_2_2_AKE_SEND_HPRIME && !hdcp->is_paired)
+		timeout = hdcp2_msg_data->timeout2;
+	else
+		timeout = hdcp2_msg_data->timeout;
+
+	/*
+	 * There is no way to detect the CERT, LPRIME and STREAM_READY
+	 * availability. So Wait for timeout and read the msg.
+	 */
+	if (!hdcp2_msg_data->msg_detectable) {
+		mdelay(timeout);
+		ret = 0;
+	} else {
+		/*
+		 * As we want to check the msg availability at timeout, Ignoring
+		 * the timeout at wait for CP_IRQ.
+		 */
+		intel_dp_hdcp_wait_for_cp_irq(hdcp, timeout);
+		ret = hdcp2_detect_msg_availability(dig_port,
+						    msg_id, &msg_ready);
+		if (!msg_ready)
+			ret = -ETIMEDOUT;
+	}
+
+	if (ret)
+		drm_dbg_kms(&i915->drm,
+			    "msg_id %d, ret %d, timeout(mSec): %d\n",
+			    hdcp2_msg_data->msg_id, ret, timeout);
+
+	return ret;
+}
+
+static const struct hdcp2_dp_msg_data *get_hdcp2_dp_msg_data(u8 msg_id)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(hdcp2_dp_msg_data); i++)
+		if (hdcp2_dp_msg_data[i].msg_id == msg_id)
+			return &hdcp2_dp_msg_data[i];
+
+	return NULL;
+}
+
+static
+int intel_dp_hdcp2_write_msg(struct intel_digital_port *dig_port,
+			     void *buf, size_t size)
+{
+	unsigned int offset;
+	u8 *byte = buf;
+	ssize_t ret, bytes_to_write, len;
+	const struct hdcp2_dp_msg_data *hdcp2_msg_data;
+
+	hdcp2_msg_data = get_hdcp2_dp_msg_data(*byte);
+	if (!hdcp2_msg_data)
+		return -EINVAL;
+
+	offset = hdcp2_msg_data->offset;
+
+	/* No msg_id in DP HDCP2.2 msgs */
+	bytes_to_write = size - 1;
+	byte++;
+
+	while (bytes_to_write) {
+		len = bytes_to_write > DP_AUX_MAX_PAYLOAD_BYTES ?
+				DP_AUX_MAX_PAYLOAD_BYTES : bytes_to_write;
+
+		ret = drm_dp_dpcd_write(&dig_port->dp.aux,
+					offset, (void *)byte, len);
+		if (ret < 0)
+			return ret;
+
+		bytes_to_write -= ret;
+		byte += ret;
+		offset += ret;
+	}
+
+	return size;
+}
+
+static
+ssize_t get_receiver_id_list_rx_info(struct intel_digital_port *dig_port, u32 *dev_cnt, u8 *byte)
+{
+	ssize_t ret;
+	u8 *rx_info = byte;
+
+	ret = drm_dp_dpcd_read(&dig_port->dp.aux,
+			       DP_HDCP_2_2_REG_RXINFO_OFFSET,
+			       (void *)rx_info, HDCP_2_2_RXINFO_LEN);
+	if (ret != HDCP_2_2_RXINFO_LEN)
+		return ret >= 0 ? -EIO : ret;
+
+	*dev_cnt = (HDCP_2_2_DEV_COUNT_HI(rx_info[0]) << 4 |
+		   HDCP_2_2_DEV_COUNT_LO(rx_info[1]));
+
+	if (*dev_cnt > HDCP_2_2_MAX_DEVICE_COUNT)
+		*dev_cnt = HDCP_2_2_MAX_DEVICE_COUNT;
+
+	return ret;
+}
+
+static
+int intel_dp_hdcp2_read_msg(struct intel_digital_port *dig_port,
+			    u8 msg_id, void *buf, size_t size)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_dp *dp = &dig_port->dp;
+	struct intel_hdcp *hdcp = &dp->attached_connector->hdcp;
+	unsigned int offset;
+	u8 *byte = buf;
+	ssize_t ret, bytes_to_recv, len;
+	const struct hdcp2_dp_msg_data *hdcp2_msg_data;
+	ktime_t msg_end = ktime_set(0, 0);
+	bool msg_expired;
+	u32 dev_cnt;
+
+	hdcp2_msg_data = get_hdcp2_dp_msg_data(msg_id);
+	if (!hdcp2_msg_data)
+		return -EINVAL;
+	offset = hdcp2_msg_data->offset;
+
+	ret = intel_dp_hdcp2_wait_for_msg(dig_port, hdcp2_msg_data);
+	if (ret < 0)
+		return ret;
+
+	hdcp->cp_irq_count_cached = atomic_read(&hdcp->cp_irq_count);
+
+	/* DP adaptation msgs has no msg_id */
+	byte++;
+
+	if (msg_id == HDCP_2_2_REP_SEND_RECVID_LIST) {
+		ret = get_receiver_id_list_rx_info(dig_port, &dev_cnt, byte);
+		if (ret < 0)
+			return ret;
+
+		byte += ret;
+		size = sizeof(struct hdcp2_rep_send_receiverid_list) -
+		HDCP_2_2_RXINFO_LEN - HDCP_2_2_RECEIVER_IDS_MAX_LEN +
+		(dev_cnt * HDCP_2_2_RECEIVER_ID_LEN);
+		offset += HDCP_2_2_RXINFO_LEN;
+	}
+
+	bytes_to_recv = size - 1;
+
+	while (bytes_to_recv) {
+		len = bytes_to_recv > DP_AUX_MAX_PAYLOAD_BYTES ?
+		      DP_AUX_MAX_PAYLOAD_BYTES : bytes_to_recv;
+
+		/* Entire msg read timeout since initiate of msg read */
+		if (bytes_to_recv == size - 1 && hdcp2_msg_data->msg_read_timeout > 0)
+			msg_end = ktime_add_ms(ktime_get_raw(),
+					       hdcp2_msg_data->msg_read_timeout);
+
+		ret = drm_dp_dpcd_read(&dig_port->dp.aux, offset,
+				       (void *)byte, len);
+		if (ret < 0) {
+			drm_dbg_kms(&i915->drm, "msg_id %d, ret %zd\n",
+				    msg_id, ret);
+			return ret;
+		}
+
+		bytes_to_recv -= ret;
+		byte += ret;
+		offset += ret;
+	}
+
+	if (hdcp2_msg_data->msg_read_timeout > 0) {
+		msg_expired = ktime_after(ktime_get_raw(), msg_end);
+		if (msg_expired) {
+			drm_dbg_kms(&i915->drm, "msg_id %d, entire msg read timeout(mSec): %d\n",
+				    msg_id, hdcp2_msg_data->msg_read_timeout);
+			return -ETIMEDOUT;
+		}
+	}
+
+	byte = buf;
+	*byte = msg_id;
+
+	return size;
+}
+
+static
+int intel_dp_hdcp2_config_stream_type(struct intel_digital_port *dig_port,
+				      bool is_repeater, u8 content_type)
+{
+	int ret;
+	struct hdcp2_dp_errata_stream_type stream_type_msg;
+
+	if (is_repeater)
+		return 0;
+
+	/*
+	 * Errata for DP: As Stream type is used for encryption, Receiver
+	 * should be communicated with stream type for the decryption of the
+	 * content.
+	 * Repeater will be communicated with stream type as a part of it's
+	 * auth later in time.
+	 */
+	stream_type_msg.msg_id = HDCP_2_2_ERRATA_DP_STREAM_TYPE;
+	stream_type_msg.stream_type = content_type;
+
+	ret =  intel_dp_hdcp2_write_msg(dig_port, &stream_type_msg,
+					sizeof(stream_type_msg));
+
+	return ret < 0 ? ret : 0;
+
+}
+
+static
+int intel_dp_hdcp2_check_link(struct intel_digital_port *dig_port,
+			      struct intel_connector *connector)
+{
+	u8 rx_status;
+	int ret;
+
+	ret = intel_dp_hdcp2_read_rx_status(dig_port, &rx_status);
+	if (ret)
+		return ret;
+
+	if (HDCP_2_2_DP_RXSTATUS_REAUTH_REQ(rx_status))
+		ret = HDCP_REAUTH_REQUEST;
+	else if (HDCP_2_2_DP_RXSTATUS_LINK_FAILED(rx_status))
+		ret = HDCP_LINK_INTEGRITY_FAILURE;
+	else if (HDCP_2_2_DP_RXSTATUS_READY(rx_status))
+		ret = HDCP_TOPOLOGY_CHANGE;
+
+	return ret;
+}
+
+static
+int intel_dp_hdcp2_capable(struct intel_digital_port *dig_port,
+			   bool *capable)
+{
+	u8 rx_caps[3];
+	int ret;
+
+	*capable = false;
+	ret = drm_dp_dpcd_read(&dig_port->dp.aux,
+			       DP_HDCP_2_2_REG_RX_CAPS_OFFSET,
+			       rx_caps, HDCP_2_2_RXCAPS_LEN);
+	if (ret != HDCP_2_2_RXCAPS_LEN)
+		return ret >= 0 ? -EIO : ret;
+
+	if (rx_caps[0] == HDCP_2_2_RX_CAPS_VERSION_VAL &&
+	    HDCP_2_2_DP_HDCP_CAPABLE(rx_caps[2]))
+		*capable = true;
+
+	return 0;
+}
+
+static const struct intel_hdcp_shim intel_dp_hdcp_shim = {
+	.write_an_aksv = intel_dp_hdcp_write_an_aksv,
+	.read_bksv = intel_dp_hdcp_read_bksv,
+	.read_bstatus = intel_dp_hdcp_read_bstatus,
+	.repeater_present = intel_dp_hdcp_repeater_present,
+	.read_ri_prime = intel_dp_hdcp_read_ri_prime,
+	.read_ksv_ready = intel_dp_hdcp_read_ksv_ready,
+	.read_ksv_fifo = intel_dp_hdcp_read_ksv_fifo,
+	.read_v_prime_part = intel_dp_hdcp_read_v_prime_part,
+	.toggle_signalling = intel_dp_hdcp_toggle_signalling,
+	.check_link = intel_dp_hdcp_check_link,
+	.hdcp_capable = intel_dp_hdcp_capable,
+	.write_2_2_msg = intel_dp_hdcp2_write_msg,
+	.read_2_2_msg = intel_dp_hdcp2_read_msg,
+	.config_stream_type = intel_dp_hdcp2_config_stream_type,
+	.check_2_2_link = intel_dp_hdcp2_check_link,
+	.hdcp_2_2_capable = intel_dp_hdcp2_capable,
+	.protocol = HDCP_PROTOCOL_DP,
+};
+
+static int
+intel_dp_mst_toggle_hdcp_stream_select(struct intel_connector *connector,
+				       bool enable)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	int ret;
+
+	ret = intel_ddi_toggle_hdcp_bits(&dig_port->base,
+					 hdcp->stream_transcoder, enable,
+					 TRANS_DDI_HDCP_SELECT);
+	if (ret)
+		drm_err(&i915->drm, "%s HDCP stream select failed (%d)\n",
+			enable ? "Enable" : "Disable", ret);
+	return ret;
+}
+
+static int
+intel_dp_mst_hdcp_stream_encryption(struct intel_connector *connector,
+				    bool enable)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum port port = dig_port->base.port;
+	enum transcoder cpu_transcoder = hdcp->stream_transcoder;
+	u32 stream_enc_status;
+	int ret;
+
+	ret = intel_dp_mst_toggle_hdcp_stream_select(connector, enable);
+	if (ret)
+		return ret;
+
+	stream_enc_status =  transcoder_to_stream_enc_status(cpu_transcoder);
+	if (!stream_enc_status)
+		return -EINVAL;
+
+	/* Wait for encryption confirmation */
+	if (intel_de_wait_for_register(i915,
+				       HDCP_STATUS(i915, cpu_transcoder, port),
+				       stream_enc_status,
+				       enable ? stream_enc_status : 0,
+				       HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS)) {
+		drm_err(&i915->drm, "Timed out waiting for transcoder: %s stream encryption %s\n",
+			transcoder_name(cpu_transcoder), enable ? "enabled" : "disabled");
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+static int
+intel_dp_mst_hdcp2_stream_encryption(struct intel_connector *connector,
+				     bool enable)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum transcoder cpu_transcoder = hdcp->stream_transcoder;
+	enum pipe pipe = (enum pipe)cpu_transcoder;
+	enum port port = dig_port->base.port;
+	int ret;
+
+	drm_WARN_ON(&i915->drm, enable &&
+		    !!(intel_de_read(i915, HDCP2_AUTH_STREAM(i915, cpu_transcoder, port))
+		    & AUTH_STREAM_TYPE) != data->streams[0].stream_type);
+
+	ret = intel_dp_mst_toggle_hdcp_stream_select(connector, enable);
+	if (ret)
+		return ret;
+
+	/* Wait for encryption confirmation */
+	if (intel_de_wait_for_register(i915,
+				       HDCP2_STREAM_STATUS(i915, cpu_transcoder, pipe),
+				       STREAM_ENCRYPTION_STATUS,
+				       enable ? STREAM_ENCRYPTION_STATUS : 0,
+				       HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS)) {
+		drm_err(&i915->drm, "Timed out waiting for transcoder: %s stream encryption %s\n",
+			transcoder_name(cpu_transcoder), enable ? "enabled" : "disabled");
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+static
+int intel_dp_mst_hdcp2_check_link(struct intel_digital_port *dig_port,
+				  struct intel_connector *connector)
+{
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	int ret;
+
+	/*
+	 * We do need to do the Link Check only for the connector involved with
+	 * HDCP port authentication and encryption.
+	 * We can re-use the hdcp->is_repeater flag to know that the connector
+	 * involved with HDCP port authentication and encryption.
+	 */
+	if (hdcp->is_repeater) {
+		ret = intel_dp_hdcp2_check_link(dig_port, connector);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static const struct intel_hdcp_shim intel_dp_mst_hdcp_shim = {
+	.write_an_aksv = intel_dp_hdcp_write_an_aksv,
+	.read_bksv = intel_dp_hdcp_read_bksv,
+	.read_bstatus = intel_dp_hdcp_read_bstatus,
+	.repeater_present = intel_dp_hdcp_repeater_present,
+	.read_ri_prime = intel_dp_hdcp_read_ri_prime,
+	.read_ksv_ready = intel_dp_hdcp_read_ksv_ready,
+	.read_ksv_fifo = intel_dp_hdcp_read_ksv_fifo,
+	.read_v_prime_part = intel_dp_hdcp_read_v_prime_part,
+	.toggle_signalling = intel_dp_hdcp_toggle_signalling,
+	.stream_encryption = intel_dp_mst_hdcp_stream_encryption,
+	.check_link = intel_dp_hdcp_check_link,
+	.hdcp_capable = intel_dp_hdcp_capable,
+	.write_2_2_msg = intel_dp_hdcp2_write_msg,
+	.read_2_2_msg = intel_dp_hdcp2_read_msg,
+	.config_stream_type = intel_dp_hdcp2_config_stream_type,
+	.stream_2_2_encryption = intel_dp_mst_hdcp2_stream_encryption,
+	.check_2_2_link = intel_dp_mst_hdcp2_check_link,
+	.hdcp_2_2_capable = intel_dp_hdcp2_capable,
+	.protocol = HDCP_PROTOCOL_DP,
+};
+
+int intel_dp_hdcp_init(struct intel_digital_port *dig_port,
+		       struct intel_connector *intel_connector)
+{
+	struct drm_device *dev = intel_connector->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_encoder *intel_encoder = &dig_port->base;
+	enum port port = intel_encoder->port;
+	struct intel_dp *intel_dp = &dig_port->dp;
+
+	if (!is_hdcp_supported(dev_priv, port))
+		return 0;
+
+	if (intel_connector->mst_port)
+		return intel_hdcp_init(intel_connector, dig_port,
+				       &intel_dp_mst_hdcp_shim);
+	else if (!intel_dp_is_edp(intel_dp))
+		return intel_hdcp_init(intel_connector, dig_port,
+				       &intel_dp_hdcp_shim);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dp_hdcp.h b/drivers/gpu/drm/i915/display/intel_dp_hdcp.h
new file mode 100644
index 000000000..eff5ec5c5
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_hdcp.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __INTEL_DP_HDCP___
+#define __INTEL_DP_HDCP___
+
+struct intel_connector;
+struct intel_digital_port;
+
+int intel_dp_hdcp_init(struct intel_digital_port *dig_port,
+		       struct intel_connector *intel_connector);
+
+#endif /* __INTEL_DP_HDCP___ */
diff --git a/drivers/gpu/drm/i915/display/intel_dp_link_training.c b/drivers/gpu/drm/i915/display/intel_dp_link_training.c
new file mode 100644
index 000000000..3d3efcf02
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_link_training.c
@@ -0,0 +1,1456 @@
+/*
+ * Copyright © 2008-2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "i915_drv.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_dp_link_training.h"
+
+static void intel_dp_reset_lttpr_common_caps(struct intel_dp *intel_dp)
+{
+	memset(intel_dp->lttpr_common_caps, 0, sizeof(intel_dp->lttpr_common_caps));
+}
+
+static void intel_dp_reset_lttpr_count(struct intel_dp *intel_dp)
+{
+	intel_dp->lttpr_common_caps[DP_PHY_REPEATER_CNT -
+				    DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV] = 0;
+}
+
+static u8 *intel_dp_lttpr_phy_caps(struct intel_dp *intel_dp,
+				   enum drm_dp_phy dp_phy)
+{
+	return intel_dp->lttpr_phy_caps[dp_phy - DP_PHY_LTTPR1];
+}
+
+static void intel_dp_read_lttpr_phy_caps(struct intel_dp *intel_dp,
+					 const u8 dpcd[DP_RECEIVER_CAP_SIZE],
+					 enum drm_dp_phy dp_phy)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	u8 *phy_caps = intel_dp_lttpr_phy_caps(intel_dp, dp_phy);
+
+	if (drm_dp_read_lttpr_phy_caps(&intel_dp->aux, dpcd, dp_phy, phy_caps) < 0) {
+		drm_dbg_kms(&dp_to_i915(intel_dp)->drm,
+			    "[ENCODER:%d:%s][%s] failed to read the PHY caps\n",
+			    encoder->base.base.id, encoder->base.name,
+			    drm_dp_phy_name(dp_phy));
+		return;
+	}
+
+	drm_dbg_kms(&dp_to_i915(intel_dp)->drm,
+		    "[ENCODER:%d:%s][%s] PHY capabilities: %*ph\n",
+		    encoder->base.base.id, encoder->base.name,
+		    drm_dp_phy_name(dp_phy),
+		    (int)sizeof(intel_dp->lttpr_phy_caps[0]),
+		    phy_caps);
+}
+
+static bool intel_dp_read_lttpr_common_caps(struct intel_dp *intel_dp,
+					    const u8 dpcd[DP_RECEIVER_CAP_SIZE])
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	int ret;
+
+	ret = drm_dp_read_lttpr_common_caps(&intel_dp->aux, dpcd,
+					    intel_dp->lttpr_common_caps);
+	if (ret < 0)
+		goto reset_caps;
+
+	drm_dbg_kms(&dp_to_i915(intel_dp)->drm,
+		    "[ENCODER:%d:%s] LTTPR common capabilities: %*ph\n",
+		    encoder->base.base.id, encoder->base.name,
+		    (int)sizeof(intel_dp->lttpr_common_caps),
+		    intel_dp->lttpr_common_caps);
+
+	/* The minimum value of LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV is 1.4 */
+	if (intel_dp->lttpr_common_caps[0] < 0x14)
+		goto reset_caps;
+
+	return true;
+
+reset_caps:
+	intel_dp_reset_lttpr_common_caps(intel_dp);
+	return false;
+}
+
+static bool
+intel_dp_set_lttpr_transparent_mode(struct intel_dp *intel_dp, bool enable)
+{
+	u8 val = enable ? DP_PHY_REPEATER_MODE_TRANSPARENT :
+			  DP_PHY_REPEATER_MODE_NON_TRANSPARENT;
+
+	return drm_dp_dpcd_write(&intel_dp->aux, DP_PHY_REPEATER_MODE, &val, 1) == 1;
+}
+
+static int intel_dp_init_lttpr(struct intel_dp *intel_dp, const u8 dpcd[DP_RECEIVER_CAP_SIZE])
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	int lttpr_count;
+	int i;
+
+	if (!intel_dp_read_lttpr_common_caps(intel_dp, dpcd))
+		return 0;
+
+	lttpr_count = drm_dp_lttpr_count(intel_dp->lttpr_common_caps);
+	/*
+	 * Prevent setting LTTPR transparent mode explicitly if no LTTPRs are
+	 * detected as this breaks link training at least on the Dell WD19TB
+	 * dock.
+	 */
+	if (lttpr_count == 0)
+		return 0;
+
+	/*
+	 * See DP Standard v2.0 3.6.6.1. about the explicit disabling of
+	 * non-transparent mode and the disable->enable non-transparent mode
+	 * sequence.
+	 */
+	intel_dp_set_lttpr_transparent_mode(intel_dp, true);
+
+	/*
+	 * In case of unsupported number of LTTPRs or failing to switch to
+	 * non-transparent mode fall-back to transparent link training mode,
+	 * still taking into account any LTTPR common lane- rate/count limits.
+	 */
+	if (lttpr_count < 0)
+		return 0;
+
+	if (!intel_dp_set_lttpr_transparent_mode(intel_dp, false)) {
+		drm_dbg_kms(&i915->drm,
+			    "[ENCODER:%d:%s] Switching to LTTPR non-transparent LT mode failed, fall-back to transparent mode\n",
+			    encoder->base.base.id, encoder->base.name);
+
+		intel_dp_set_lttpr_transparent_mode(intel_dp, true);
+		intel_dp_reset_lttpr_count(intel_dp);
+
+		return 0;
+	}
+
+	for (i = 0; i < lttpr_count; i++)
+		intel_dp_read_lttpr_phy_caps(intel_dp, dpcd, DP_PHY_LTTPR(i));
+
+	return lttpr_count;
+}
+
+/**
+ * intel_dp_init_lttpr_and_dprx_caps - detect LTTPR and DPRX caps, init the LTTPR link training mode
+ * @intel_dp: Intel DP struct
+ *
+ * Read the LTTPR common and DPRX capabilities and switch to non-transparent
+ * link training mode if any is detected and read the PHY capabilities for all
+ * detected LTTPRs. In case of an LTTPR detection error or if the number of
+ * LTTPRs is more than is supported (8), fall back to the no-LTTPR,
+ * transparent mode link training mode.
+ *
+ * Returns:
+ *   >0  if LTTPRs were detected and the non-transparent LT mode was set. The
+ *       DPRX capabilities are read out.
+ *    0  if no LTTPRs or more than 8 LTTPRs were detected or in case of a
+ *       detection failure and the transparent LT mode was set. The DPRX
+ *       capabilities are read out.
+ *   <0  Reading out the DPRX capabilities failed.
+ */
+int intel_dp_init_lttpr_and_dprx_caps(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int lttpr_count = 0;
+
+	/*
+	 * Detecting LTTPRs must be avoided on platforms with an AUX timeout
+	 * period < 3.2ms. (see DP Standard v2.0, 2.11.2, 3.6.6.1).
+	 */
+	if (!intel_dp_is_edp(intel_dp) &&
+	    (DISPLAY_VER(i915) >= 10 && !IS_GEMINILAKE(i915))) {
+		u8 dpcd[DP_RECEIVER_CAP_SIZE];
+
+		if (drm_dp_dpcd_probe(&intel_dp->aux, DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV))
+			return -EIO;
+
+		if (drm_dp_read_dpcd_caps(&intel_dp->aux, dpcd))
+			return -EIO;
+
+		lttpr_count = intel_dp_init_lttpr(intel_dp, dpcd);
+	}
+
+	/*
+	 * The DPTX shall read the DPRX caps after LTTPR detection, so re-read
+	 * it here.
+	 */
+	if (drm_dp_read_dpcd_caps(&intel_dp->aux, intel_dp->dpcd)) {
+		intel_dp_reset_lttpr_common_caps(intel_dp);
+		return -EIO;
+	}
+
+	return lttpr_count;
+}
+
+static u8 dp_voltage_max(u8 preemph)
+{
+	switch (preemph & DP_TRAIN_PRE_EMPHASIS_MASK) {
+	case DP_TRAIN_PRE_EMPH_LEVEL_0:
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
+	case DP_TRAIN_PRE_EMPH_LEVEL_1:
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
+	case DP_TRAIN_PRE_EMPH_LEVEL_2:
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_1;
+	case DP_TRAIN_PRE_EMPH_LEVEL_3:
+	default:
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_0;
+	}
+}
+
+static u8 intel_dp_lttpr_voltage_max(struct intel_dp *intel_dp,
+				     enum drm_dp_phy dp_phy)
+{
+	const u8 *phy_caps = intel_dp_lttpr_phy_caps(intel_dp, dp_phy);
+
+	if (drm_dp_lttpr_voltage_swing_level_3_supported(phy_caps))
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
+	else
+		return DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
+}
+
+static u8 intel_dp_lttpr_preemph_max(struct intel_dp *intel_dp,
+				     enum drm_dp_phy dp_phy)
+{
+	const u8 *phy_caps = intel_dp_lttpr_phy_caps(intel_dp, dp_phy);
+
+	if (drm_dp_lttpr_pre_emphasis_level_3_supported(phy_caps))
+		return DP_TRAIN_PRE_EMPH_LEVEL_3;
+	else
+		return DP_TRAIN_PRE_EMPH_LEVEL_2;
+}
+
+static bool
+intel_dp_phy_is_downstream_of_source(struct intel_dp *intel_dp,
+				     enum drm_dp_phy dp_phy)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int lttpr_count = drm_dp_lttpr_count(intel_dp->lttpr_common_caps);
+
+	drm_WARN_ON_ONCE(&i915->drm, lttpr_count <= 0 && dp_phy != DP_PHY_DPRX);
+
+	return lttpr_count <= 0 || dp_phy == DP_PHY_LTTPR(lttpr_count - 1);
+}
+
+static u8 intel_dp_phy_voltage_max(struct intel_dp *intel_dp,
+				   const struct intel_crtc_state *crtc_state,
+				   enum drm_dp_phy dp_phy)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 voltage_max;
+
+	/*
+	 * Get voltage_max from the DPTX_PHY (source or LTTPR) upstream from
+	 * the DPRX_PHY we train.
+	 */
+	if (intel_dp_phy_is_downstream_of_source(intel_dp, dp_phy))
+		voltage_max = intel_dp->voltage_max(intel_dp, crtc_state);
+	else
+		voltage_max = intel_dp_lttpr_voltage_max(intel_dp, dp_phy + 1);
+
+	drm_WARN_ON_ONCE(&i915->drm,
+			 voltage_max != DP_TRAIN_VOLTAGE_SWING_LEVEL_2 &&
+			 voltage_max != DP_TRAIN_VOLTAGE_SWING_LEVEL_3);
+
+	return voltage_max;
+}
+
+static u8 intel_dp_phy_preemph_max(struct intel_dp *intel_dp,
+				   enum drm_dp_phy dp_phy)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 preemph_max;
+
+	/*
+	 * Get preemph_max from the DPTX_PHY (source or LTTPR) upstream from
+	 * the DPRX_PHY we train.
+	 */
+	if (intel_dp_phy_is_downstream_of_source(intel_dp, dp_phy))
+		preemph_max = intel_dp->preemph_max(intel_dp);
+	else
+		preemph_max = intel_dp_lttpr_preemph_max(intel_dp, dp_phy + 1);
+
+	drm_WARN_ON_ONCE(&i915->drm,
+			 preemph_max != DP_TRAIN_PRE_EMPH_LEVEL_2 &&
+			 preemph_max != DP_TRAIN_PRE_EMPH_LEVEL_3);
+
+	return preemph_max;
+}
+
+static bool has_per_lane_signal_levels(struct intel_dp *intel_dp,
+				       enum drm_dp_phy dp_phy)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	return !intel_dp_phy_is_downstream_of_source(intel_dp, dp_phy) ||
+		DISPLAY_VER(i915) >= 11;
+}
+
+/* 128b/132b */
+static u8 intel_dp_get_lane_adjust_tx_ffe_preset(struct intel_dp *intel_dp,
+						 const struct intel_crtc_state *crtc_state,
+						 enum drm_dp_phy dp_phy,
+						 const u8 link_status[DP_LINK_STATUS_SIZE],
+						 int lane)
+{
+	u8 tx_ffe = 0;
+
+	if (has_per_lane_signal_levels(intel_dp, dp_phy)) {
+		lane = min(lane, crtc_state->lane_count - 1);
+		tx_ffe = drm_dp_get_adjust_tx_ffe_preset(link_status, lane);
+	} else {
+		for (lane = 0; lane < crtc_state->lane_count; lane++)
+			tx_ffe = max(tx_ffe, drm_dp_get_adjust_tx_ffe_preset(link_status, lane));
+	}
+
+	return tx_ffe;
+}
+
+/* 8b/10b */
+static u8 intel_dp_get_lane_adjust_vswing_preemph(struct intel_dp *intel_dp,
+						  const struct intel_crtc_state *crtc_state,
+						  enum drm_dp_phy dp_phy,
+						  const u8 link_status[DP_LINK_STATUS_SIZE],
+						  int lane)
+{
+	u8 v = 0;
+	u8 p = 0;
+	u8 voltage_max;
+	u8 preemph_max;
+
+	if (has_per_lane_signal_levels(intel_dp, dp_phy)) {
+		lane = min(lane, crtc_state->lane_count - 1);
+
+		v = drm_dp_get_adjust_request_voltage(link_status, lane);
+		p = drm_dp_get_adjust_request_pre_emphasis(link_status, lane);
+	} else {
+		for (lane = 0; lane < crtc_state->lane_count; lane++) {
+			v = max(v, drm_dp_get_adjust_request_voltage(link_status, lane));
+			p = max(p, drm_dp_get_adjust_request_pre_emphasis(link_status, lane));
+		}
+	}
+
+	preemph_max = intel_dp_phy_preemph_max(intel_dp, dp_phy);
+	if (p >= preemph_max)
+		p = preemph_max | DP_TRAIN_MAX_PRE_EMPHASIS_REACHED;
+
+	v = min(v, dp_voltage_max(p));
+
+	voltage_max = intel_dp_phy_voltage_max(intel_dp, crtc_state, dp_phy);
+	if (v >= voltage_max)
+		v = voltage_max | DP_TRAIN_MAX_SWING_REACHED;
+
+	return v | p;
+}
+
+static u8 intel_dp_get_lane_adjust_train(struct intel_dp *intel_dp,
+					 const struct intel_crtc_state *crtc_state,
+					 enum drm_dp_phy dp_phy,
+					 const u8 link_status[DP_LINK_STATUS_SIZE],
+					 int lane)
+{
+	if (intel_dp_is_uhbr(crtc_state))
+		return intel_dp_get_lane_adjust_tx_ffe_preset(intel_dp, crtc_state,
+							      dp_phy, link_status, lane);
+	else
+		return intel_dp_get_lane_adjust_vswing_preemph(intel_dp, crtc_state,
+							       dp_phy, link_status, lane);
+}
+
+#define TRAIN_REQ_FMT "%d/%d/%d/%d"
+#define _TRAIN_REQ_VSWING_ARGS(link_status, lane) \
+	(drm_dp_get_adjust_request_voltage((link_status), (lane)) >> DP_TRAIN_VOLTAGE_SWING_SHIFT)
+#define TRAIN_REQ_VSWING_ARGS(link_status) \
+	_TRAIN_REQ_VSWING_ARGS(link_status, 0), \
+	_TRAIN_REQ_VSWING_ARGS(link_status, 1), \
+	_TRAIN_REQ_VSWING_ARGS(link_status, 2), \
+	_TRAIN_REQ_VSWING_ARGS(link_status, 3)
+#define _TRAIN_REQ_PREEMPH_ARGS(link_status, lane) \
+	(drm_dp_get_adjust_request_pre_emphasis((link_status), (lane)) >> DP_TRAIN_PRE_EMPHASIS_SHIFT)
+#define TRAIN_REQ_PREEMPH_ARGS(link_status) \
+	_TRAIN_REQ_PREEMPH_ARGS(link_status, 0), \
+	_TRAIN_REQ_PREEMPH_ARGS(link_status, 1), \
+	_TRAIN_REQ_PREEMPH_ARGS(link_status, 2), \
+	_TRAIN_REQ_PREEMPH_ARGS(link_status, 3)
+#define _TRAIN_REQ_TX_FFE_ARGS(link_status, lane) \
+	drm_dp_get_adjust_tx_ffe_preset((link_status), (lane))
+#define TRAIN_REQ_TX_FFE_ARGS(link_status) \
+	_TRAIN_REQ_TX_FFE_ARGS(link_status, 0), \
+	_TRAIN_REQ_TX_FFE_ARGS(link_status, 1), \
+	_TRAIN_REQ_TX_FFE_ARGS(link_status, 2), \
+	_TRAIN_REQ_TX_FFE_ARGS(link_status, 3)
+
+void
+intel_dp_get_adjust_train(struct intel_dp *intel_dp,
+			  const struct intel_crtc_state *crtc_state,
+			  enum drm_dp_phy dp_phy,
+			  const u8 link_status[DP_LINK_STATUS_SIZE])
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	int lane;
+
+	if (intel_dp_is_uhbr(crtc_state)) {
+		drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s][%s] 128b/132b, lanes: %d, "
+			    "TX FFE request: " TRAIN_REQ_FMT "\n",
+			    encoder->base.base.id, encoder->base.name,
+			    drm_dp_phy_name(dp_phy),
+			    crtc_state->lane_count,
+			    TRAIN_REQ_TX_FFE_ARGS(link_status));
+	} else {
+		drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s][%s] 8b/10b, lanes: %d, "
+			    "vswing request: " TRAIN_REQ_FMT ", "
+			    "pre-emphasis request: " TRAIN_REQ_FMT "\n",
+			    encoder->base.base.id, encoder->base.name,
+			    drm_dp_phy_name(dp_phy),
+			    crtc_state->lane_count,
+			    TRAIN_REQ_VSWING_ARGS(link_status),
+			    TRAIN_REQ_PREEMPH_ARGS(link_status));
+	}
+
+	for (lane = 0; lane < 4; lane++)
+		intel_dp->train_set[lane] =
+			intel_dp_get_lane_adjust_train(intel_dp, crtc_state,
+						       dp_phy, link_status, lane);
+}
+
+static int intel_dp_training_pattern_set_reg(struct intel_dp *intel_dp,
+					     enum drm_dp_phy dp_phy)
+{
+	return dp_phy == DP_PHY_DPRX ?
+		DP_TRAINING_PATTERN_SET :
+		DP_TRAINING_PATTERN_SET_PHY_REPEATER(dp_phy);
+}
+
+static bool
+intel_dp_set_link_train(struct intel_dp *intel_dp,
+			const struct intel_crtc_state *crtc_state,
+			enum drm_dp_phy dp_phy,
+			u8 dp_train_pat)
+{
+	int reg = intel_dp_training_pattern_set_reg(intel_dp, dp_phy);
+	u8 buf[sizeof(intel_dp->train_set) + 1];
+	int len;
+
+	intel_dp_program_link_training_pattern(intel_dp, crtc_state,
+					       dp_phy, dp_train_pat);
+
+	buf[0] = dp_train_pat;
+	/* DP_TRAINING_LANEx_SET follow DP_TRAINING_PATTERN_SET */
+	memcpy(buf + 1, intel_dp->train_set, crtc_state->lane_count);
+	len = crtc_state->lane_count + 1;
+
+	return drm_dp_dpcd_write(&intel_dp->aux, reg, buf, len) == len;
+}
+
+static char dp_training_pattern_name(u8 train_pat)
+{
+	switch (train_pat) {
+	case DP_TRAINING_PATTERN_1:
+	case DP_TRAINING_PATTERN_2:
+	case DP_TRAINING_PATTERN_3:
+		return '0' + train_pat;
+	case DP_TRAINING_PATTERN_4:
+		return '4';
+	default:
+		MISSING_CASE(train_pat);
+		return '?';
+	}
+}
+
+void
+intel_dp_program_link_training_pattern(struct intel_dp *intel_dp,
+				       const struct intel_crtc_state *crtc_state,
+				       enum drm_dp_phy dp_phy,
+				       u8 dp_train_pat)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	u8 train_pat = intel_dp_training_pattern_symbol(dp_train_pat);
+
+	if (train_pat != DP_TRAINING_PATTERN_DISABLE)
+		drm_dbg_kms(&i915->drm,
+			    "[ENCODER:%d:%s][%s] Using DP training pattern TPS%c\n",
+			    encoder->base.base.id, encoder->base.name,
+			    drm_dp_phy_name(dp_phy),
+			    dp_training_pattern_name(train_pat));
+
+	intel_dp->set_link_train(intel_dp, crtc_state, dp_train_pat);
+}
+
+#define TRAIN_SET_FMT "%d%s/%d%s/%d%s/%d%s"
+#define _TRAIN_SET_VSWING_ARGS(train_set) \
+	((train_set) & DP_TRAIN_VOLTAGE_SWING_MASK) >> DP_TRAIN_VOLTAGE_SWING_SHIFT, \
+	(train_set) & DP_TRAIN_MAX_SWING_REACHED ? "(max)" : ""
+#define TRAIN_SET_VSWING_ARGS(train_set) \
+	_TRAIN_SET_VSWING_ARGS((train_set)[0]), \
+	_TRAIN_SET_VSWING_ARGS((train_set)[1]), \
+	_TRAIN_SET_VSWING_ARGS((train_set)[2]), \
+	_TRAIN_SET_VSWING_ARGS((train_set)[3])
+#define _TRAIN_SET_PREEMPH_ARGS(train_set) \
+	((train_set) & DP_TRAIN_PRE_EMPHASIS_MASK) >> DP_TRAIN_PRE_EMPHASIS_SHIFT, \
+	(train_set) & DP_TRAIN_MAX_PRE_EMPHASIS_REACHED ? "(max)" : ""
+#define TRAIN_SET_PREEMPH_ARGS(train_set) \
+	_TRAIN_SET_PREEMPH_ARGS((train_set)[0]), \
+	_TRAIN_SET_PREEMPH_ARGS((train_set)[1]), \
+	_TRAIN_SET_PREEMPH_ARGS((train_set)[2]), \
+	_TRAIN_SET_PREEMPH_ARGS((train_set)[3])
+#define _TRAIN_SET_TX_FFE_ARGS(train_set) \
+	((train_set) & DP_TX_FFE_PRESET_VALUE_MASK), ""
+#define TRAIN_SET_TX_FFE_ARGS(train_set) \
+	_TRAIN_SET_TX_FFE_ARGS((train_set)[0]), \
+	_TRAIN_SET_TX_FFE_ARGS((train_set)[1]), \
+	_TRAIN_SET_TX_FFE_ARGS((train_set)[2]), \
+	_TRAIN_SET_TX_FFE_ARGS((train_set)[3])
+
+void intel_dp_set_signal_levels(struct intel_dp *intel_dp,
+				const struct intel_crtc_state *crtc_state,
+				enum drm_dp_phy dp_phy)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+
+	if (intel_dp_is_uhbr(crtc_state)) {
+		drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s][%s] 128b/132b, lanes: %d, "
+			    "TX FFE presets: " TRAIN_SET_FMT "\n",
+			    encoder->base.base.id, encoder->base.name,
+			    drm_dp_phy_name(dp_phy),
+			    crtc_state->lane_count,
+			    TRAIN_SET_TX_FFE_ARGS(intel_dp->train_set));
+	} else {
+		drm_dbg_kms(&i915->drm, "[ENCODER:%d:%s][%s] 8b/10b, lanes: %d, "
+			    "vswing levels: " TRAIN_SET_FMT ", "
+			    "pre-emphasis levels: " TRAIN_SET_FMT "\n",
+			    encoder->base.base.id, encoder->base.name,
+			    drm_dp_phy_name(dp_phy),
+			    crtc_state->lane_count,
+			    TRAIN_SET_VSWING_ARGS(intel_dp->train_set),
+			    TRAIN_SET_PREEMPH_ARGS(intel_dp->train_set));
+	}
+
+	if (intel_dp_phy_is_downstream_of_source(intel_dp, dp_phy))
+		encoder->set_signal_levels(encoder, crtc_state);
+}
+
+static bool
+intel_dp_reset_link_train(struct intel_dp *intel_dp,
+			  const struct intel_crtc_state *crtc_state,
+			  enum drm_dp_phy dp_phy,
+			  u8 dp_train_pat)
+{
+	memset(intel_dp->train_set, 0, sizeof(intel_dp->train_set));
+	intel_dp_set_signal_levels(intel_dp, crtc_state, dp_phy);
+	return intel_dp_set_link_train(intel_dp, crtc_state, dp_phy, dp_train_pat);
+}
+
+static bool
+intel_dp_update_link_train(struct intel_dp *intel_dp,
+			   const struct intel_crtc_state *crtc_state,
+			   enum drm_dp_phy dp_phy)
+{
+	int reg = dp_phy == DP_PHY_DPRX ?
+			    DP_TRAINING_LANE0_SET :
+			    DP_TRAINING_LANE0_SET_PHY_REPEATER(dp_phy);
+	int ret;
+
+	intel_dp_set_signal_levels(intel_dp, crtc_state, dp_phy);
+
+	ret = drm_dp_dpcd_write(&intel_dp->aux, reg,
+				intel_dp->train_set, crtc_state->lane_count);
+
+	return ret == crtc_state->lane_count;
+}
+
+/* 128b/132b */
+static bool intel_dp_lane_max_tx_ffe_reached(u8 train_set_lane)
+{
+	return (train_set_lane & DP_TX_FFE_PRESET_VALUE_MASK) ==
+		DP_TX_FFE_PRESET_VALUE_MASK;
+}
+
+/*
+ * 8b/10b
+ *
+ * FIXME: The DP spec is very confusing here, also the Link CTS spec seems to
+ * have self contradicting tests around this area.
+ *
+ * In lieu of better ideas let's just stop when we've reached the max supported
+ * vswing with its max pre-emphasis, which is either 2+1 or 3+0 depending on
+ * whether vswing level 3 is supported or not.
+ */
+static bool intel_dp_lane_max_vswing_reached(u8 train_set_lane)
+{
+	u8 v = (train_set_lane & DP_TRAIN_VOLTAGE_SWING_MASK) >>
+		DP_TRAIN_VOLTAGE_SWING_SHIFT;
+	u8 p = (train_set_lane & DP_TRAIN_PRE_EMPHASIS_MASK) >>
+		DP_TRAIN_PRE_EMPHASIS_SHIFT;
+
+	if ((train_set_lane & DP_TRAIN_MAX_SWING_REACHED) == 0)
+		return false;
+
+	if (v + p != 3)
+		return false;
+
+	return true;
+}
+
+static bool intel_dp_link_max_vswing_reached(struct intel_dp *intel_dp,
+					     const struct intel_crtc_state *crtc_state)
+{
+	int lane;
+
+	for (lane = 0; lane < crtc_state->lane_count; lane++) {
+		u8 train_set_lane = intel_dp->train_set[lane];
+
+		if (intel_dp_is_uhbr(crtc_state)) {
+			if (!intel_dp_lane_max_tx_ffe_reached(train_set_lane))
+				return false;
+		} else {
+			if (!intel_dp_lane_max_vswing_reached(train_set_lane))
+				return false;
+		}
+	}
+
+	return true;
+}
+
+/*
+ * Prepare link training by configuring the link parameters. On DDI platforms
+ * also enable the port here.
+ */
+static bool
+intel_dp_prepare_link_train(struct intel_dp *intel_dp,
+			    const struct intel_crtc_state *crtc_state)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	u8 link_config[2];
+	u8 link_bw, rate_select;
+
+	if (intel_dp->prepare_link_retrain)
+		intel_dp->prepare_link_retrain(intel_dp, crtc_state);
+
+	intel_dp_compute_rate(intel_dp, crtc_state->port_clock,
+			      &link_bw, &rate_select);
+
+	/*
+	 * WaEdpLinkRateDataReload
+	 *
+	 * Parade PS8461E MUX (used on varius TGL+ laptops) needs
+	 * to snoop the link rates reported by the sink when we
+	 * use LINK_RATE_SET in order to operate in jitter cleaning
+	 * mode (as opposed to redriver mode). Unfortunately it
+	 * loses track of the snooped link rates when powered down,
+	 * so we need to make it re-snoop often. Without this high
+	 * link rates are not stable.
+	 */
+	if (!link_bw) {
+		struct intel_connector *connector = intel_dp->attached_connector;
+		__le16 sink_rates[DP_MAX_SUPPORTED_RATES];
+
+		drm_dbg_kms(&i915->drm, "[CONNECTOR:%d:%s] Reloading eDP link rates\n",
+			    connector->base.base.id, connector->base.name);
+
+		drm_dp_dpcd_read(&intel_dp->aux, DP_SUPPORTED_LINK_RATES,
+				 sink_rates, sizeof(sink_rates));
+	}
+
+	if (link_bw)
+		drm_dbg_kms(&i915->drm,
+			    "[ENCODER:%d:%s] Using LINK_BW_SET value %02x\n",
+			    encoder->base.base.id, encoder->base.name, link_bw);
+	else
+		drm_dbg_kms(&i915->drm,
+			    "[ENCODER:%d:%s] Using LINK_RATE_SET value %02x\n",
+			    encoder->base.base.id, encoder->base.name, rate_select);
+
+	/* Write the link configuration data */
+	link_config[0] = link_bw;
+	link_config[1] = crtc_state->lane_count;
+	if (drm_dp_enhanced_frame_cap(intel_dp->dpcd))
+		link_config[1] |= DP_LANE_COUNT_ENHANCED_FRAME_EN;
+	drm_dp_dpcd_write(&intel_dp->aux, DP_LINK_BW_SET, link_config, 2);
+
+	/* eDP 1.4 rate select method. */
+	if (!link_bw)
+		drm_dp_dpcd_write(&intel_dp->aux, DP_LINK_RATE_SET,
+				  &rate_select, 1);
+
+	link_config[0] = crtc_state->vrr.enable ? DP_MSA_TIMING_PAR_IGNORE_EN : 0;
+	link_config[1] = intel_dp_is_uhbr(crtc_state) ?
+		DP_SET_ANSI_128B132B : DP_SET_ANSI_8B10B;
+	drm_dp_dpcd_write(&intel_dp->aux, DP_DOWNSPREAD_CTRL, link_config, 2);
+
+	return true;
+}
+
+static bool intel_dp_adjust_request_changed(const struct intel_crtc_state *crtc_state,
+					    const u8 old_link_status[DP_LINK_STATUS_SIZE],
+					    const u8 new_link_status[DP_LINK_STATUS_SIZE])
+{
+	int lane;
+
+	for (lane = 0; lane < crtc_state->lane_count; lane++) {
+		u8 old, new;
+
+		if (intel_dp_is_uhbr(crtc_state)) {
+			old = drm_dp_get_adjust_tx_ffe_preset(old_link_status, lane);
+			new = drm_dp_get_adjust_tx_ffe_preset(new_link_status, lane);
+		} else {
+			old = drm_dp_get_adjust_request_voltage(old_link_status, lane) |
+				drm_dp_get_adjust_request_pre_emphasis(old_link_status, lane);
+			new = drm_dp_get_adjust_request_voltage(new_link_status, lane) |
+				drm_dp_get_adjust_request_pre_emphasis(new_link_status, lane);
+		}
+
+		if (old != new)
+			return true;
+	}
+
+	return false;
+}
+
+void
+intel_dp_dump_link_status(struct intel_dp *intel_dp, enum drm_dp_phy dp_phy,
+			  const u8 link_status[DP_LINK_STATUS_SIZE])
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+
+	drm_dbg_kms(&i915->drm,
+		    "[ENCODER:%d:%s][%s] ln0_1:0x%x ln2_3:0x%x align:0x%x sink:0x%x adj_req0_1:0x%x adj_req2_3:0x%x\n",
+		    encoder->base.base.id, encoder->base.name,
+		    drm_dp_phy_name(dp_phy),
+		    link_status[0], link_status[1], link_status[2],
+		    link_status[3], link_status[4], link_status[5]);
+}
+
+/*
+ * Perform the link training clock recovery phase on the given DP PHY using
+ * training pattern 1.
+ */
+static bool
+intel_dp_link_training_clock_recovery(struct intel_dp *intel_dp,
+				      const struct intel_crtc_state *crtc_state,
+				      enum drm_dp_phy dp_phy)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	u8 old_link_status[DP_LINK_STATUS_SIZE] = {};
+	int voltage_tries, cr_tries, max_cr_tries;
+	u8 link_status[DP_LINK_STATUS_SIZE];
+	bool max_vswing_reached = false;
+	int delay_us;
+
+	delay_us = drm_dp_read_clock_recovery_delay(&intel_dp->aux,
+						    intel_dp->dpcd, dp_phy,
+						    intel_dp_is_uhbr(crtc_state));
+
+	/* clock recovery */
+	if (!intel_dp_reset_link_train(intel_dp, crtc_state, dp_phy,
+				       DP_TRAINING_PATTERN_1 |
+				       DP_LINK_SCRAMBLING_DISABLE)) {
+		drm_err(&i915->drm, "[ENCODER:%d:%s][%s] Failed to enable link training\n",
+			encoder->base.base.id, encoder->base.name,
+			drm_dp_phy_name(dp_phy));
+		return false;
+	}
+
+	/*
+	 * The DP 1.4 spec defines the max clock recovery retries value
+	 * as 10 but for pre-DP 1.4 devices we set a very tolerant
+	 * retry limit of 80 (4 voltage levels x 4 preemphasis levels x
+	 * x 5 identical voltage retries). Since the previous specs didn't
+	 * define a limit and created the possibility of an infinite loop
+	 * we want to prevent any sync from triggering that corner case.
+	 */
+	if (intel_dp->dpcd[DP_DPCD_REV] >= DP_DPCD_REV_14)
+		max_cr_tries = 10;
+	else
+		max_cr_tries = 80;
+
+	voltage_tries = 1;
+	for (cr_tries = 0; cr_tries < max_cr_tries; ++cr_tries) {
+		usleep_range(delay_us, 2 * delay_us);
+
+		if (drm_dp_dpcd_read_phy_link_status(&intel_dp->aux, dp_phy,
+						     link_status) < 0) {
+			drm_err(&i915->drm, "[ENCODER:%d:%s][%s] Failed to get link status\n",
+				encoder->base.base.id, encoder->base.name,
+				drm_dp_phy_name(dp_phy));
+			return false;
+		}
+
+		if (drm_dp_clock_recovery_ok(link_status, crtc_state->lane_count)) {
+			drm_dbg_kms(&i915->drm,
+				    "[ENCODER:%d:%s][%s] Clock recovery OK\n",
+				    encoder->base.base.id, encoder->base.name,
+				    drm_dp_phy_name(dp_phy));
+			return true;
+		}
+
+		if (voltage_tries == 5) {
+			intel_dp_dump_link_status(intel_dp, dp_phy, link_status);
+			drm_dbg_kms(&i915->drm,
+				    "[ENCODER:%d:%s][%s] Same voltage tried 5 times\n",
+				    encoder->base.base.id, encoder->base.name,
+				    drm_dp_phy_name(dp_phy));
+			return false;
+		}
+
+		if (max_vswing_reached) {
+			intel_dp_dump_link_status(intel_dp, dp_phy, link_status);
+			drm_dbg_kms(&i915->drm,
+				    "[ENCODER:%d:%s][%s] Max Voltage Swing reached\n",
+				    encoder->base.base.id, encoder->base.name,
+				    drm_dp_phy_name(dp_phy));
+			return false;
+		}
+
+		/* Update training set as requested by target */
+		intel_dp_get_adjust_train(intel_dp, crtc_state, dp_phy,
+					  link_status);
+		if (!intel_dp_update_link_train(intel_dp, crtc_state, dp_phy)) {
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s][%s] Failed to update link training\n",
+				encoder->base.base.id, encoder->base.name,
+				drm_dp_phy_name(dp_phy));
+			return false;
+		}
+
+		if (!intel_dp_adjust_request_changed(crtc_state, old_link_status, link_status))
+			++voltage_tries;
+		else
+			voltage_tries = 1;
+
+		memcpy(old_link_status, link_status, sizeof(link_status));
+
+		if (intel_dp_link_max_vswing_reached(intel_dp, crtc_state))
+			max_vswing_reached = true;
+	}
+
+	intel_dp_dump_link_status(intel_dp, dp_phy, link_status);
+	drm_err(&i915->drm,
+		"[ENCODER:%d:%s][%s] Failed clock recovery %d times, giving up!\n",
+		encoder->base.base.id, encoder->base.name,
+		drm_dp_phy_name(dp_phy), max_cr_tries);
+
+	return false;
+}
+
+/*
+ * Pick Training Pattern Sequence (TPS) for channel equalization. 128b/132b TPS2
+ * for UHBR+, TPS4 for HBR3 or for 1.4 devices that support it, TPS3 for HBR2 or
+ * 1.2 devices that support it, TPS2 otherwise.
+ */
+static u32 intel_dp_training_pattern(struct intel_dp *intel_dp,
+				     const struct intel_crtc_state *crtc_state,
+				     enum drm_dp_phy dp_phy)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	bool source_tps3, sink_tps3, source_tps4, sink_tps4;
+
+	/* UHBR+ use separate 128b/132b TPS2 */
+	if (intel_dp_is_uhbr(crtc_state))
+		return DP_TRAINING_PATTERN_2;
+
+	/*
+	 * TPS4 support is mandatory for all downstream devices that
+	 * support HBR3. There are no known eDP panels that support
+	 * TPS4 as of Feb 2018 as per VESA eDP_v1.4b_E1 specification.
+	 * LTTPRs must support TPS4.
+	 */
+	source_tps4 = intel_dp_source_supports_tps4(i915);
+	sink_tps4 = dp_phy != DP_PHY_DPRX ||
+		    drm_dp_tps4_supported(intel_dp->dpcd);
+	if (source_tps4 && sink_tps4) {
+		return DP_TRAINING_PATTERN_4;
+	} else if (crtc_state->port_clock == 810000) {
+		if (!source_tps4)
+			drm_dbg_kms(&i915->drm,
+				    "8.1 Gbps link rate without source TPS4 support\n");
+		if (!sink_tps4)
+			drm_dbg_kms(&i915->drm,
+				    "8.1 Gbps link rate without sink TPS4 support\n");
+	}
+
+	/*
+	 * TPS3 support is mandatory for downstream devices that
+	 * support HBR2. However, not all sinks follow the spec.
+	 */
+	source_tps3 = intel_dp_source_supports_tps3(i915);
+	sink_tps3 = dp_phy != DP_PHY_DPRX ||
+		    drm_dp_tps3_supported(intel_dp->dpcd);
+	if (source_tps3 && sink_tps3) {
+		return  DP_TRAINING_PATTERN_3;
+	} else if (crtc_state->port_clock >= 540000) {
+		if (!source_tps3)
+			drm_dbg_kms(&i915->drm,
+				    ">=5.4/6.48 Gbps link rate without source TPS3 support\n");
+		if (!sink_tps3)
+			drm_dbg_kms(&i915->drm,
+				    ">=5.4/6.48 Gbps link rate without sink TPS3 support\n");
+	}
+
+	return DP_TRAINING_PATTERN_2;
+}
+
+/*
+ * Perform the link training channel equalization phase on the given DP PHY
+ * using one of training pattern 2, 3 or 4 depending on the source and
+ * sink capabilities.
+ */
+static bool
+intel_dp_link_training_channel_equalization(struct intel_dp *intel_dp,
+					    const struct intel_crtc_state *crtc_state,
+					    enum drm_dp_phy dp_phy)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	int tries;
+	u32 training_pattern;
+	u8 link_status[DP_LINK_STATUS_SIZE];
+	bool channel_eq = false;
+	int delay_us;
+
+	delay_us = drm_dp_read_channel_eq_delay(&intel_dp->aux,
+						intel_dp->dpcd, dp_phy,
+						intel_dp_is_uhbr(crtc_state));
+
+	training_pattern = intel_dp_training_pattern(intel_dp, crtc_state, dp_phy);
+	/* Scrambling is disabled for TPS2/3 and enabled for TPS4 */
+	if (training_pattern != DP_TRAINING_PATTERN_4)
+		training_pattern |= DP_LINK_SCRAMBLING_DISABLE;
+
+	/* channel equalization */
+	if (!intel_dp_set_link_train(intel_dp, crtc_state, dp_phy,
+				     training_pattern)) {
+		drm_err(&i915->drm,
+			"[ENCODER:%d:%s][%s] Failed to start channel equalization\n",
+			encoder->base.base.id, encoder->base.name,
+			drm_dp_phy_name(dp_phy));
+		return false;
+	}
+
+	for (tries = 0; tries < 5; tries++) {
+		usleep_range(delay_us, 2 * delay_us);
+
+		if (drm_dp_dpcd_read_phy_link_status(&intel_dp->aux, dp_phy,
+						     link_status) < 0) {
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s][%s] Failed to get link status\n",
+				encoder->base.base.id, encoder->base.name,
+				drm_dp_phy_name(dp_phy));
+			break;
+		}
+
+		/* Make sure clock is still ok */
+		if (!drm_dp_clock_recovery_ok(link_status,
+					      crtc_state->lane_count)) {
+			intel_dp_dump_link_status(intel_dp, dp_phy, link_status);
+			drm_dbg_kms(&i915->drm,
+				    "[ENCODER:%d:%s][%s] Clock recovery check failed, cannot "
+				    "continue channel equalization\n",
+				    encoder->base.base.id, encoder->base.name,
+				    drm_dp_phy_name(dp_phy));
+			break;
+		}
+
+		if (drm_dp_channel_eq_ok(link_status,
+					 crtc_state->lane_count)) {
+			channel_eq = true;
+			drm_dbg_kms(&i915->drm,
+				    "[ENCODER:%d:%s][%s] Channel EQ done. DP Training successful\n",
+				    encoder->base.base.id, encoder->base.name,
+				    drm_dp_phy_name(dp_phy));
+			break;
+		}
+
+		/* Update training set as requested by target */
+		intel_dp_get_adjust_train(intel_dp, crtc_state, dp_phy,
+					  link_status);
+		if (!intel_dp_update_link_train(intel_dp, crtc_state, dp_phy)) {
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s][%s] Failed to update link training\n",
+				encoder->base.base.id, encoder->base.name,
+				drm_dp_phy_name(dp_phy));
+			break;
+		}
+	}
+
+	/* Try 5 times, else fail and try at lower BW */
+	if (tries == 5) {
+		intel_dp_dump_link_status(intel_dp, dp_phy, link_status);
+		drm_dbg_kms(&i915->drm,
+			    "[ENCODER:%d:%s][%s] Channel equalization failed 5 times\n",
+			    encoder->base.base.id, encoder->base.name,
+			    drm_dp_phy_name(dp_phy));
+	}
+
+	return channel_eq;
+}
+
+static bool intel_dp_disable_dpcd_training_pattern(struct intel_dp *intel_dp,
+						   enum drm_dp_phy dp_phy)
+{
+	int reg = intel_dp_training_pattern_set_reg(intel_dp, dp_phy);
+	u8 val = DP_TRAINING_PATTERN_DISABLE;
+
+	return drm_dp_dpcd_write(&intel_dp->aux, reg, &val, 1) == 1;
+}
+
+static int
+intel_dp_128b132b_intra_hop(struct intel_dp *intel_dp,
+			    const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 sink_status;
+	int ret;
+
+	ret = drm_dp_dpcd_readb(&intel_dp->aux, DP_SINK_STATUS, &sink_status);
+	if (ret != 1) {
+		drm_dbg_kms(&i915->drm, "Failed to read sink status\n");
+		return ret < 0 ? ret : -EIO;
+	}
+
+	return sink_status & DP_INTRA_HOP_AUX_REPLY_INDICATION ? 1 : 0;
+}
+
+/**
+ * intel_dp_stop_link_train - stop link training
+ * @intel_dp: DP struct
+ * @crtc_state: state for CRTC attached to the encoder
+ *
+ * Stop the link training of the @intel_dp port, disabling the training
+ * pattern in the sink's DPCD, and disabling the test pattern symbol
+ * generation on the port.
+ *
+ * What symbols are output on the port after this point is
+ * platform specific: On DDI/VLV/CHV platforms it will be the idle pattern
+ * with the pipe being disabled, on older platforms it's HW specific if/how an
+ * idle pattern is generated, as the pipe is already enabled here for those.
+ *
+ * This function must be called after intel_dp_start_link_train().
+ */
+void intel_dp_stop_link_train(struct intel_dp *intel_dp,
+			      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+
+	intel_dp->link_trained = true;
+
+	intel_dp_disable_dpcd_training_pattern(intel_dp, DP_PHY_DPRX);
+	intel_dp_program_link_training_pattern(intel_dp, crtc_state, DP_PHY_DPRX,
+					       DP_TRAINING_PATTERN_DISABLE);
+
+	if (intel_dp_is_uhbr(crtc_state) &&
+	    wait_for(intel_dp_128b132b_intra_hop(intel_dp, crtc_state) == 0, 500)) {
+		drm_dbg_kms(&i915->drm,
+			    "[ENCODER:%d:%s] 128b/132b intra-hop not clearing\n",
+			    encoder->base.base.id, encoder->base.name);
+	}
+}
+
+static bool
+intel_dp_link_train_phy(struct intel_dp *intel_dp,
+			const struct intel_crtc_state *crtc_state,
+			enum drm_dp_phy dp_phy)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	bool ret = false;
+
+	if (!intel_dp_link_training_clock_recovery(intel_dp, crtc_state, dp_phy))
+		goto out;
+
+	if (!intel_dp_link_training_channel_equalization(intel_dp, crtc_state, dp_phy))
+		goto out;
+
+	ret = true;
+
+out:
+	drm_dbg_kms(&dp_to_i915(intel_dp)->drm,
+		    "[CONNECTOR:%d:%s][ENCODER:%d:%s][%s] Link Training %s at link rate = %d, lane count = %d\n",
+		    connector->base.base.id, connector->base.name,
+		    encoder->base.base.id, encoder->base.name,
+		    drm_dp_phy_name(dp_phy),
+		    ret ? "passed" : "failed",
+		    crtc_state->port_clock, crtc_state->lane_count);
+
+	return ret;
+}
+
+static void intel_dp_schedule_fallback_link_training(struct intel_dp *intel_dp,
+						     const struct intel_crtc_state *crtc_state)
+{
+	struct intel_connector *intel_connector = intel_dp->attached_connector;
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+
+	if (intel_dp->hobl_active) {
+		drm_dbg_kms(&dp_to_i915(intel_dp)->drm,
+			    "[ENCODER:%d:%s] Link Training failed with HOBL active, "
+			    "not enabling it from now on",
+			    encoder->base.base.id, encoder->base.name);
+		intel_dp->hobl_failed = true;
+	} else if (intel_dp_get_link_train_fallback_values(intel_dp,
+							   crtc_state->port_clock,
+							   crtc_state->lane_count)) {
+		return;
+	}
+
+	/* Schedule a Hotplug Uevent to userspace to start modeset */
+	schedule_work(&intel_connector->modeset_retry_work);
+}
+
+/* Perform the link training on all LTTPRs and the DPRX on a link. */
+static bool
+intel_dp_link_train_all_phys(struct intel_dp *intel_dp,
+			     const struct intel_crtc_state *crtc_state,
+			     int lttpr_count)
+{
+	bool ret = true;
+	int i;
+
+	for (i = lttpr_count - 1; i >= 0; i--) {
+		enum drm_dp_phy dp_phy = DP_PHY_LTTPR(i);
+
+		ret = intel_dp_link_train_phy(intel_dp, crtc_state, dp_phy);
+		intel_dp_disable_dpcd_training_pattern(intel_dp, dp_phy);
+
+		if (!ret)
+			break;
+	}
+
+	if (ret)
+		ret = intel_dp_link_train_phy(intel_dp, crtc_state, DP_PHY_DPRX);
+
+	if (intel_dp->set_idle_link_train)
+		intel_dp->set_idle_link_train(intel_dp, crtc_state);
+
+	return ret;
+}
+
+/*
+ * 128b/132b DP LANEx_EQ_DONE Sequence (DP 2.0 E11 3.5.2.16.1)
+ */
+static bool
+intel_dp_128b132b_lane_eq(struct intel_dp *intel_dp,
+			  const struct intel_crtc_state *crtc_state)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	u8 link_status[DP_LINK_STATUS_SIZE];
+	int delay_us;
+	int try, max_tries = 20;
+	unsigned long deadline;
+	bool timeout = false;
+
+	/*
+	 * Reset signal levels. Start transmitting 128b/132b TPS1.
+	 *
+	 * Put DPRX and LTTPRs (if any) into intra-hop AUX mode by writing TPS1
+	 * in DP_TRAINING_PATTERN_SET.
+	 */
+	if (!intel_dp_reset_link_train(intel_dp, crtc_state, DP_PHY_DPRX,
+				       DP_TRAINING_PATTERN_1)) {
+		drm_err(&i915->drm,
+			"[ENCODER:%d:%s] Failed to start 128b/132b TPS1\n",
+			encoder->base.base.id, encoder->base.name);
+		return false;
+	}
+
+	delay_us = drm_dp_128b132b_read_aux_rd_interval(&intel_dp->aux);
+
+	/* Read the initial TX FFE settings. */
+	if (drm_dp_dpcd_read_link_status(&intel_dp->aux, link_status) < 0) {
+		drm_err(&i915->drm,
+			"[ENCODER:%d:%s] Failed to read TX FFE presets\n",
+			encoder->base.base.id, encoder->base.name);
+		return false;
+	}
+
+	/* Update signal levels and training set as requested. */
+	intel_dp_get_adjust_train(intel_dp, crtc_state, DP_PHY_DPRX, link_status);
+	if (!intel_dp_update_link_train(intel_dp, crtc_state, DP_PHY_DPRX)) {
+		drm_err(&i915->drm,
+			"[ENCODER:%d:%s] Failed to set initial TX FFE settings\n",
+			encoder->base.base.id, encoder->base.name);
+		return false;
+	}
+
+	/* Start transmitting 128b/132b TPS2. */
+	if (!intel_dp_set_link_train(intel_dp, crtc_state, DP_PHY_DPRX,
+				     DP_TRAINING_PATTERN_2)) {
+		drm_err(&i915->drm,
+			"[ENCODER:%d:%s] Failed to start 128b/132b TPS2\n",
+			encoder->base.base.id, encoder->base.name);
+		return false;
+	}
+
+	/* Time budget for the LANEx_EQ_DONE Sequence */
+	deadline = jiffies + msecs_to_jiffies_timeout(400);
+
+	for (try = 0; try < max_tries; try++) {
+		usleep_range(delay_us, 2 * delay_us);
+
+		/*
+		 * The delay may get updated. The transmitter shall read the
+		 * delay before link status during link training.
+		 */
+		delay_us = drm_dp_128b132b_read_aux_rd_interval(&intel_dp->aux);
+
+		if (drm_dp_dpcd_read_link_status(&intel_dp->aux, link_status) < 0) {
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s] Failed to read link status\n",
+				encoder->base.base.id, encoder->base.name);
+			return false;
+		}
+
+		if (drm_dp_128b132b_link_training_failed(link_status)) {
+			intel_dp_dump_link_status(intel_dp, DP_PHY_DPRX, link_status);
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s] Downstream link training failure\n",
+				encoder->base.base.id, encoder->base.name);
+			return false;
+		}
+
+		if (drm_dp_128b132b_lane_channel_eq_done(link_status, crtc_state->lane_count)) {
+			drm_dbg_kms(&i915->drm,
+				    "[ENCODER:%d:%s] Lane channel eq done\n",
+				    encoder->base.base.id, encoder->base.name);
+			break;
+		}
+
+		if (timeout) {
+			intel_dp_dump_link_status(intel_dp, DP_PHY_DPRX, link_status);
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s] Lane channel eq timeout\n",
+				encoder->base.base.id, encoder->base.name);
+			return false;
+		}
+
+		if (time_after(jiffies, deadline))
+			timeout = true; /* try one last time after deadline */
+
+		/* Update signal levels and training set as requested. */
+		intel_dp_get_adjust_train(intel_dp, crtc_state, DP_PHY_DPRX, link_status);
+		if (!intel_dp_update_link_train(intel_dp, crtc_state, DP_PHY_DPRX)) {
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s] Failed to update TX FFE settings\n",
+				encoder->base.base.id, encoder->base.name);
+			return false;
+		}
+	}
+
+	if (try == max_tries) {
+		intel_dp_dump_link_status(intel_dp, DP_PHY_DPRX, link_status);
+		drm_err(&i915->drm,
+			"[ENCODER:%d:%s] Max loop count reached\n",
+			encoder->base.base.id, encoder->base.name);
+		return false;
+	}
+
+	for (;;) {
+		if (time_after(jiffies, deadline))
+			timeout = true; /* try one last time after deadline */
+
+		if (drm_dp_dpcd_read_link_status(&intel_dp->aux, link_status) < 0) {
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s] Failed to read link status\n",
+				encoder->base.base.id, encoder->base.name);
+			return false;
+		}
+
+		if (drm_dp_128b132b_link_training_failed(link_status)) {
+			intel_dp_dump_link_status(intel_dp, DP_PHY_DPRX, link_status);
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s] Downstream link training failure\n",
+				encoder->base.base.id, encoder->base.name);
+			return false;
+		}
+
+		if (drm_dp_128b132b_eq_interlane_align_done(link_status)) {
+			drm_dbg_kms(&i915->drm,
+				    "[ENCODER:%d:%s] Interlane align done\n",
+				    encoder->base.base.id, encoder->base.name);
+			break;
+		}
+
+		if (timeout) {
+			intel_dp_dump_link_status(intel_dp, DP_PHY_DPRX, link_status);
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s] Interlane align timeout\n",
+				encoder->base.base.id, encoder->base.name);
+			return false;
+		}
+
+		usleep_range(2000, 3000);
+	}
+
+	return true;
+}
+
+/*
+ * 128b/132b DP LANEx_CDS_DONE Sequence (DP 2.0 E11 3.5.2.16.2)
+ */
+static bool
+intel_dp_128b132b_lane_cds(struct intel_dp *intel_dp,
+			   const struct intel_crtc_state *crtc_state,
+			   int lttpr_count)
+{
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	u8 link_status[DP_LINK_STATUS_SIZE];
+	unsigned long deadline;
+
+	if (drm_dp_dpcd_writeb(&intel_dp->aux, DP_TRAINING_PATTERN_SET,
+			       DP_TRAINING_PATTERN_2_CDS) != 1) {
+		drm_err(&i915->drm,
+			"[ENCODER:%d:%s] Failed to start 128b/132b TPS2 CDS\n",
+			encoder->base.base.id, encoder->base.name);
+		return false;
+	}
+
+	/* Time budget for the LANEx_CDS_DONE Sequence */
+	deadline = jiffies + msecs_to_jiffies_timeout((lttpr_count + 1) * 20);
+
+	for (;;) {
+		bool timeout = false;
+
+		if (time_after(jiffies, deadline))
+			timeout = true; /* try one last time after deadline */
+
+		usleep_range(2000, 3000);
+
+		if (drm_dp_dpcd_read_link_status(&intel_dp->aux, link_status) < 0) {
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s] Failed to read link status\n",
+				encoder->base.base.id, encoder->base.name);
+			return false;
+		}
+
+		if (drm_dp_128b132b_eq_interlane_align_done(link_status) &&
+		    drm_dp_128b132b_cds_interlane_align_done(link_status) &&
+		    drm_dp_128b132b_lane_symbol_locked(link_status, crtc_state->lane_count)) {
+			drm_dbg_kms(&i915->drm,
+				    "[ENCODER:%d:%s] CDS interlane align done\n",
+				    encoder->base.base.id, encoder->base.name);
+			break;
+		}
+
+		if (drm_dp_128b132b_link_training_failed(link_status)) {
+			intel_dp_dump_link_status(intel_dp, DP_PHY_DPRX, link_status);
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s] Downstream link training failure\n",
+				encoder->base.base.id, encoder->base.name);
+			return false;
+		}
+
+		if (timeout) {
+			intel_dp_dump_link_status(intel_dp, DP_PHY_DPRX, link_status);
+			drm_err(&i915->drm,
+				"[ENCODER:%d:%s] CDS timeout\n",
+				encoder->base.base.id, encoder->base.name);
+			return false;
+		}
+	}
+
+	/* FIXME: Should DP_TRAINING_PATTERN_DISABLE be written first? */
+	if (intel_dp->set_idle_link_train)
+		intel_dp->set_idle_link_train(intel_dp, crtc_state);
+
+	return true;
+}
+
+/*
+ * 128b/132b link training sequence. (DP 2.0 E11 SCR on link training.)
+ */
+static bool
+intel_dp_128b132b_link_train(struct intel_dp *intel_dp,
+			     const struct intel_crtc_state *crtc_state,
+			     int lttpr_count)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+	bool passed = false;
+
+	if (wait_for(intel_dp_128b132b_intra_hop(intel_dp, crtc_state) == 0, 500)) {
+		drm_err(&i915->drm,
+			"[ENCODER:%d:%s] 128b/132b intra-hop not clear\n",
+			encoder->base.base.id, encoder->base.name);
+		return false;
+	}
+
+	if (intel_dp_128b132b_lane_eq(intel_dp, crtc_state) &&
+	    intel_dp_128b132b_lane_cds(intel_dp, crtc_state, lttpr_count))
+		passed = true;
+
+	drm_dbg_kms(&i915->drm,
+		    "[CONNECTOR:%d:%s][ENCODER:%d:%s] 128b/132b Link Training %s at link rate = %d, lane count = %d\n",
+		    connector->base.base.id, connector->base.name,
+		    encoder->base.base.id, encoder->base.name,
+		    passed ? "passed" : "failed",
+		    crtc_state->port_clock, crtc_state->lane_count);
+
+	return passed;
+}
+
+/**
+ * intel_dp_start_link_train - start link training
+ * @intel_dp: DP struct
+ * @crtc_state: state for CRTC attached to the encoder
+ *
+ * Start the link training of the @intel_dp port, scheduling a fallback
+ * retraining with reduced link rate/lane parameters if the link training
+ * fails.
+ * After calling this function intel_dp_stop_link_train() must be called.
+ */
+void intel_dp_start_link_train(struct intel_dp *intel_dp,
+			       const struct intel_crtc_state *crtc_state)
+{
+	bool passed;
+	/*
+	 * TODO: Reiniting LTTPRs here won't be needed once proper connector
+	 * HW state readout is added.
+	 */
+	int lttpr_count = intel_dp_init_lttpr_and_dprx_caps(intel_dp);
+
+	if (lttpr_count < 0)
+		/* Still continue with enabling the port and link training. */
+		lttpr_count = 0;
+
+	intel_dp_prepare_link_train(intel_dp, crtc_state);
+
+	if (intel_dp_is_uhbr(crtc_state))
+		passed = intel_dp_128b132b_link_train(intel_dp, crtc_state, lttpr_count);
+	else
+		passed = intel_dp_link_train_all_phys(intel_dp, crtc_state, lttpr_count);
+
+	if (!passed)
+		intel_dp_schedule_fallback_link_training(intel_dp, crtc_state);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dp_link_training.h b/drivers/gpu/drm/i915/display/intel_dp_link_training.h
new file mode 100644
index 000000000..7fa1c0833
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_link_training.h
@@ -0,0 +1,42 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DP_LINK_TRAINING_H__
+#define __INTEL_DP_LINK_TRAINING_H__
+
+#include <drm/display/drm_dp_helper.h>
+
+struct intel_crtc_state;
+struct intel_dp;
+
+int intel_dp_init_lttpr_and_dprx_caps(struct intel_dp *intel_dp);
+
+void intel_dp_get_adjust_train(struct intel_dp *intel_dp,
+			       const struct intel_crtc_state *crtc_state,
+			       enum drm_dp_phy dp_phy,
+			       const u8 link_status[DP_LINK_STATUS_SIZE]);
+void intel_dp_program_link_training_pattern(struct intel_dp *intel_dp,
+					    const struct intel_crtc_state *crtc_state,
+					    enum drm_dp_phy dp_phy,
+					    u8 dp_train_pat);
+void intel_dp_set_signal_levels(struct intel_dp *intel_dp,
+				const struct intel_crtc_state *crtc_state,
+				enum drm_dp_phy dp_phy);
+void intel_dp_start_link_train(struct intel_dp *intel_dp,
+			       const struct intel_crtc_state *crtc_state);
+void intel_dp_stop_link_train(struct intel_dp *intel_dp,
+			      const struct intel_crtc_state *crtc_state);
+
+void
+intel_dp_dump_link_status(struct intel_dp *intel_dp, enum drm_dp_phy dp_phy,
+			  const u8 link_status[DP_LINK_STATUS_SIZE]);
+
+/* Get the TPSx symbol type of the value programmed to DP_TRAINING_PATTERN_SET */
+static inline u8 intel_dp_training_pattern_symbol(u8 pattern)
+{
+	return pattern & ~DP_LINK_SCRAMBLING_DISABLE;
+}
+
+#endif /* __INTEL_DP_LINK_TRAINING_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dp_mst.c b/drivers/gpu/drm/i915/display/intel_dp_mst.c
new file mode 100644
index 000000000..eec32f682
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_mst.c
@@ -0,0 +1,1076 @@
+/*
+ * Copyright © 2008 Intel Corporation
+ *             2014 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_probe_helper.h>
+
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_audio.h"
+#include "intel_connector.h"
+#include "intel_crtc.h"
+#include "intel_ddi.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_dp_hdcp.h"
+#include "intel_dp_mst.h"
+#include "intel_dpio_phy.h"
+#include "intel_hdcp.h"
+#include "intel_hotplug.h"
+#include "skl_scaler.h"
+
+static int intel_dp_mst_compute_link_config(struct intel_encoder *encoder,
+					    struct intel_crtc_state *crtc_state,
+					    struct drm_connector_state *conn_state,
+					    struct link_config_limits *limits)
+{
+	struct drm_atomic_state *state = crtc_state->uapi.state;
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_dp *intel_dp = &intel_mst->primary->dp;
+	struct drm_dp_mst_topology_state *mst_state;
+	struct intel_connector *connector =
+		to_intel_connector(conn_state->connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	int bpp, slots = -EINVAL;
+
+	mst_state = drm_atomic_get_mst_topology_state(state, &intel_dp->mst_mgr);
+	if (IS_ERR(mst_state))
+		return PTR_ERR(mst_state);
+
+	crtc_state->lane_count = limits->max_lane_count;
+	crtc_state->port_clock = limits->max_rate;
+
+	// TODO: Handle pbn_div changes by adding a new MST helper
+	if (!mst_state->pbn_div) {
+		mst_state->pbn_div = drm_dp_get_vc_payload_bw(&intel_dp->mst_mgr,
+							      limits->max_rate,
+							      limits->max_lane_count);
+	}
+
+	for (bpp = limits->max_bpp; bpp >= limits->min_bpp; bpp -= 2 * 3) {
+		crtc_state->pipe_bpp = bpp;
+
+		crtc_state->pbn = drm_dp_calc_pbn_mode(adjusted_mode->crtc_clock,
+						       crtc_state->pipe_bpp,
+						       false);
+		slots = drm_dp_atomic_find_time_slots(state, &intel_dp->mst_mgr,
+						      connector->port, crtc_state->pbn);
+		if (slots == -EDEADLK)
+			return slots;
+		if (slots >= 0)
+			break;
+	}
+
+	if (slots < 0) {
+		drm_dbg_kms(&i915->drm, "failed finding vcpi slots:%d\n",
+			    slots);
+		return slots;
+	}
+
+	intel_link_compute_m_n(crtc_state->pipe_bpp,
+			       crtc_state->lane_count,
+			       adjusted_mode->crtc_clock,
+			       crtc_state->port_clock,
+			       &crtc_state->dp_m_n,
+			       crtc_state->fec_enable);
+	crtc_state->dp_m_n.tu = slots;
+
+	return 0;
+}
+
+static int intel_dp_mst_update_slots(struct intel_encoder *encoder,
+				     struct intel_crtc_state *crtc_state,
+				     struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_dp *intel_dp = &intel_mst->primary->dp;
+	struct drm_dp_mst_topology_mgr *mgr = &intel_dp->mst_mgr;
+	struct drm_dp_mst_topology_state *topology_state;
+	u8 link_coding_cap = intel_dp_is_uhbr(crtc_state) ?
+		DP_CAP_ANSI_128B132B : DP_CAP_ANSI_8B10B;
+
+	topology_state = drm_atomic_get_mst_topology_state(conn_state->state, mgr);
+	if (IS_ERR(topology_state)) {
+		drm_dbg_kms(&i915->drm, "slot update failed\n");
+		return PTR_ERR(topology_state);
+	}
+
+	drm_dp_mst_update_slots(topology_state, link_coding_cap);
+
+	return 0;
+}
+
+static int intel_dp_mst_compute_config(struct intel_encoder *encoder,
+				       struct intel_crtc_state *pipe_config,
+				       struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_dp *intel_dp = &intel_mst->primary->dp;
+	struct intel_connector *connector =
+		to_intel_connector(conn_state->connector);
+	struct intel_digital_connector_state *intel_conn_state =
+		to_intel_digital_connector_state(conn_state);
+	const struct drm_display_mode *adjusted_mode =
+		&pipe_config->hw.adjusted_mode;
+	struct link_config_limits limits;
+	int ret;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+	pipe_config->has_pch_encoder = false;
+
+	if (intel_conn_state->force_audio == HDMI_AUDIO_AUTO)
+		pipe_config->has_audio = connector->port->has_audio;
+	else
+		pipe_config->has_audio =
+			intel_conn_state->force_audio == HDMI_AUDIO_ON;
+
+	/*
+	 * for MST we always configure max link bw - the spec doesn't
+	 * seem to suggest we should do otherwise.
+	 */
+	limits.min_rate =
+	limits.max_rate = intel_dp_max_link_rate(intel_dp);
+
+	limits.min_lane_count =
+	limits.max_lane_count = intel_dp_max_lane_count(intel_dp);
+
+	limits.min_bpp = intel_dp_min_bpp(pipe_config->output_format);
+	/*
+	 * FIXME: If all the streams can't fit into the link with
+	 * their current pipe_bpp we should reduce pipe_bpp across
+	 * the board until things start to fit. Until then we
+	 * limit to <= 8bpc since that's what was hardcoded for all
+	 * MST streams previously. This hack should be removed once
+	 * we have the proper retry logic in place.
+	 */
+	limits.max_bpp = min(pipe_config->pipe_bpp, 24);
+
+	intel_dp_adjust_compliance_config(intel_dp, pipe_config, &limits);
+
+	ret = intel_dp_mst_compute_link_config(encoder, pipe_config,
+					       conn_state, &limits);
+	if (ret)
+		return ret;
+
+	ret = intel_dp_mst_update_slots(encoder, pipe_config, conn_state);
+	if (ret)
+		return ret;
+
+	pipe_config->limited_color_range =
+		intel_dp_limited_color_range(pipe_config, conn_state);
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		pipe_config->lane_lat_optim_mask =
+			bxt_ddi_phy_calc_lane_lat_optim_mask(pipe_config->lane_count);
+
+	intel_ddi_compute_min_voltage_level(dev_priv, pipe_config);
+
+	return 0;
+}
+
+/*
+ * Iterate over all connectors and return a mask of
+ * all CPU transcoders streaming over the same DP link.
+ */
+static unsigned int
+intel_dp_mst_transcoder_mask(struct intel_atomic_state *state,
+			     struct intel_dp *mst_port)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_digital_connector_state *conn_state;
+	struct intel_connector *connector;
+	u8 transcoders = 0;
+	int i;
+
+	if (DISPLAY_VER(dev_priv) < 12)
+		return 0;
+
+	for_each_new_intel_connector_in_state(state, connector, conn_state, i) {
+		const struct intel_crtc_state *crtc_state;
+		struct intel_crtc *crtc;
+
+		if (connector->mst_port != mst_port || !conn_state->base.crtc)
+			continue;
+
+		crtc = to_intel_crtc(conn_state->base.crtc);
+		crtc_state = intel_atomic_get_new_crtc_state(state, crtc);
+
+		if (!crtc_state->hw.active)
+			continue;
+
+		transcoders |= BIT(crtc_state->cpu_transcoder);
+	}
+
+	return transcoders;
+}
+
+static int intel_dp_mst_compute_config_late(struct intel_encoder *encoder,
+					    struct intel_crtc_state *crtc_state,
+					    struct drm_connector_state *conn_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(conn_state->state);
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_dp *intel_dp = &intel_mst->primary->dp;
+
+	/* lowest numbered transcoder will be designated master */
+	crtc_state->mst_master_transcoder =
+		ffs(intel_dp_mst_transcoder_mask(state, intel_dp)) - 1;
+
+	return 0;
+}
+
+/*
+ * If one of the connectors in a MST stream needs a modeset, mark all CRTCs
+ * that shares the same MST stream as mode changed,
+ * intel_modeset_pipe_config()+intel_crtc_check_fastset() will take care to do
+ * a fastset when possible.
+ */
+static int
+intel_dp_mst_atomic_master_trans_check(struct intel_connector *connector,
+				       struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct drm_connector_list_iter connector_list_iter;
+	struct intel_connector *connector_iter;
+	int ret = 0;
+
+	if (DISPLAY_VER(dev_priv) < 12)
+		return  0;
+
+	if (!intel_connector_needs_modeset(state, &connector->base))
+		return 0;
+
+	drm_connector_list_iter_begin(&dev_priv->drm, &connector_list_iter);
+	for_each_intel_connector_iter(connector_iter, &connector_list_iter) {
+		struct intel_digital_connector_state *conn_iter_state;
+		struct intel_crtc_state *crtc_state;
+		struct intel_crtc *crtc;
+
+		if (connector_iter->mst_port != connector->mst_port ||
+		    connector_iter == connector)
+			continue;
+
+		conn_iter_state = intel_atomic_get_digital_connector_state(state,
+									   connector_iter);
+		if (IS_ERR(conn_iter_state)) {
+			ret = PTR_ERR(conn_iter_state);
+			break;
+		}
+
+		if (!conn_iter_state->base.crtc)
+			continue;
+
+		crtc = to_intel_crtc(conn_iter_state->base.crtc);
+		crtc_state = intel_atomic_get_crtc_state(&state->base, crtc);
+		if (IS_ERR(crtc_state)) {
+			ret = PTR_ERR(crtc_state);
+			break;
+		}
+
+		ret = drm_atomic_add_affected_planes(&state->base, &crtc->base);
+		if (ret)
+			break;
+		crtc_state->uapi.mode_changed = true;
+	}
+	drm_connector_list_iter_end(&connector_list_iter);
+
+	return ret;
+}
+
+static int
+intel_dp_mst_atomic_check(struct drm_connector *connector,
+			  struct drm_atomic_state *_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(_state);
+	struct intel_connector *intel_connector =
+		to_intel_connector(connector);
+	int ret;
+
+	ret = intel_digital_connector_atomic_check(connector, &state->base);
+	if (ret)
+		return ret;
+
+	ret = intel_dp_mst_atomic_master_trans_check(intel_connector, state);
+	if (ret)
+		return ret;
+
+	return drm_dp_atomic_release_time_slots(&state->base,
+						&intel_connector->mst_port->mst_mgr,
+						intel_connector->port);
+}
+
+static void clear_act_sent(struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+
+	intel_de_write(i915, dp_tp_status_reg(encoder, crtc_state),
+		       DP_TP_STATUS_ACT_SENT);
+}
+
+static void wait_for_act_sent(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_dp *intel_dp = &intel_mst->primary->dp;
+
+	if (intel_de_wait_for_set(i915, dp_tp_status_reg(encoder, crtc_state),
+				  DP_TP_STATUS_ACT_SENT, 1))
+		drm_err(&i915->drm, "Timed out waiting for ACT sent\n");
+
+	drm_dp_check_act_status(&intel_dp->mst_mgr);
+}
+
+static void intel_mst_disable_dp(struct intel_atomic_state *state,
+				 struct intel_encoder *encoder,
+				 const struct intel_crtc_state *old_crtc_state,
+				 const struct drm_connector_state *old_conn_state)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_digital_port *dig_port = intel_mst->primary;
+	struct intel_dp *intel_dp = &dig_port->dp;
+	struct intel_connector *connector =
+		to_intel_connector(old_conn_state->connector);
+	struct drm_dp_mst_topology_state *old_mst_state =
+		drm_atomic_get_old_mst_topology_state(&state->base, &intel_dp->mst_mgr);
+	struct drm_dp_mst_topology_state *new_mst_state =
+		drm_atomic_get_new_mst_topology_state(&state->base, &intel_dp->mst_mgr);
+	const struct drm_dp_mst_atomic_payload *old_payload =
+		drm_atomic_get_mst_payload_state(old_mst_state, connector->port);
+	struct drm_dp_mst_atomic_payload *new_payload =
+		drm_atomic_get_mst_payload_state(new_mst_state, connector->port);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+
+	drm_dbg_kms(&i915->drm, "active links %d\n",
+		    intel_dp->active_mst_links);
+
+	intel_hdcp_disable(intel_mst->connector);
+
+	drm_dp_remove_payload(&intel_dp->mst_mgr, new_mst_state,
+			      old_payload, new_payload);
+
+	intel_audio_codec_disable(encoder, old_crtc_state, old_conn_state);
+}
+
+static void intel_mst_post_disable_dp(struct intel_atomic_state *state,
+				      struct intel_encoder *encoder,
+				      const struct intel_crtc_state *old_crtc_state,
+				      const struct drm_connector_state *old_conn_state)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_digital_port *dig_port = intel_mst->primary;
+	struct intel_dp *intel_dp = &dig_port->dp;
+	struct intel_connector *connector =
+		to_intel_connector(old_conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	bool last_mst_stream;
+
+	intel_dp->active_mst_links--;
+	last_mst_stream = intel_dp->active_mst_links == 0;
+	drm_WARN_ON(&dev_priv->drm,
+		    DISPLAY_VER(dev_priv) >= 12 && last_mst_stream &&
+		    !intel_dp_mst_is_master_trans(old_crtc_state));
+
+	intel_crtc_vblank_off(old_crtc_state);
+
+	intel_disable_transcoder(old_crtc_state);
+
+	clear_act_sent(encoder, old_crtc_state);
+
+	intel_de_rmw(dev_priv, TRANS_DDI_FUNC_CTL(old_crtc_state->cpu_transcoder),
+		     TRANS_DDI_DP_VC_PAYLOAD_ALLOC, 0);
+
+	wait_for_act_sent(encoder, old_crtc_state);
+
+	intel_ddi_disable_transcoder_func(old_crtc_state);
+
+	if (DISPLAY_VER(dev_priv) >= 9)
+		skl_scaler_disable(old_crtc_state);
+	else
+		ilk_pfit_disable(old_crtc_state);
+
+	/*
+	 * Power down mst path before disabling the port, otherwise we end
+	 * up getting interrupts from the sink upon detecting link loss.
+	 */
+	drm_dp_send_power_updown_phy(&intel_dp->mst_mgr, connector->port,
+				     false);
+
+	/*
+	 * BSpec 4287: disable DIP after the transcoder is disabled and before
+	 * the transcoder clock select is set to none.
+	 */
+	if (last_mst_stream)
+		intel_dp_set_infoframes(&dig_port->base, false,
+					old_crtc_state, NULL);
+	/*
+	 * From TGL spec: "If multi-stream slave transcoder: Configure
+	 * Transcoder Clock Select to direct no clock to the transcoder"
+	 *
+	 * From older GENs spec: "Configure Transcoder Clock Select to direct
+	 * no clock to the transcoder"
+	 */
+	if (DISPLAY_VER(dev_priv) < 12 || !last_mst_stream)
+		intel_ddi_disable_pipe_clock(old_crtc_state);
+
+
+	intel_mst->connector = NULL;
+	if (last_mst_stream)
+		dig_port->base.post_disable(state, &dig_port->base,
+						  old_crtc_state, NULL);
+
+	drm_dbg_kms(&dev_priv->drm, "active links %d\n",
+		    intel_dp->active_mst_links);
+}
+
+static void intel_mst_pre_pll_enable_dp(struct intel_atomic_state *state,
+					struct intel_encoder *encoder,
+					const struct intel_crtc_state *pipe_config,
+					const struct drm_connector_state *conn_state)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_digital_port *dig_port = intel_mst->primary;
+	struct intel_dp *intel_dp = &dig_port->dp;
+
+	if (intel_dp->active_mst_links == 0)
+		dig_port->base.pre_pll_enable(state, &dig_port->base,
+						    pipe_config, NULL);
+}
+
+static void intel_mst_pre_enable_dp(struct intel_atomic_state *state,
+				    struct intel_encoder *encoder,
+				    const struct intel_crtc_state *pipe_config,
+				    const struct drm_connector_state *conn_state)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_digital_port *dig_port = intel_mst->primary;
+	struct intel_dp *intel_dp = &dig_port->dp;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_connector *connector =
+		to_intel_connector(conn_state->connector);
+	struct drm_dp_mst_topology_state *mst_state =
+		drm_atomic_get_new_mst_topology_state(&state->base, &intel_dp->mst_mgr);
+	int ret;
+	bool first_mst_stream;
+
+	/* MST encoders are bound to a crtc, not to a connector,
+	 * force the mapping here for get_hw_state.
+	 */
+	connector->encoder = encoder;
+	intel_mst->connector = connector;
+	first_mst_stream = intel_dp->active_mst_links == 0;
+	drm_WARN_ON(&dev_priv->drm,
+		    DISPLAY_VER(dev_priv) >= 12 && first_mst_stream &&
+		    !intel_dp_mst_is_master_trans(pipe_config));
+
+	drm_dbg_kms(&dev_priv->drm, "active links %d\n",
+		    intel_dp->active_mst_links);
+
+	if (first_mst_stream)
+		intel_dp_set_power(intel_dp, DP_SET_POWER_D0);
+
+	drm_dp_send_power_updown_phy(&intel_dp->mst_mgr, connector->port, true);
+
+	if (first_mst_stream)
+		dig_port->base.pre_enable(state, &dig_port->base,
+						pipe_config, NULL);
+
+	intel_dp->active_mst_links++;
+
+	ret = drm_dp_add_payload_part1(&intel_dp->mst_mgr, mst_state,
+				       drm_atomic_get_mst_payload_state(mst_state, connector->port));
+	if (ret < 0)
+		drm_err(&dev_priv->drm, "Failed to create MST payload for %s: %d\n",
+			connector->base.name, ret);
+
+	/*
+	 * Before Gen 12 this is not done as part of
+	 * dig_port->base.pre_enable() and should be done here. For
+	 * Gen 12+ the step in which this should be done is different for the
+	 * first MST stream, so it's done on the DDI for the first stream and
+	 * here for the following ones.
+	 */
+	if (DISPLAY_VER(dev_priv) < 12 || !first_mst_stream)
+		intel_ddi_enable_pipe_clock(encoder, pipe_config);
+
+	intel_ddi_set_dp_msa(pipe_config, conn_state);
+}
+
+static void intel_mst_enable_dp(struct intel_atomic_state *state,
+				struct intel_encoder *encoder,
+				const struct intel_crtc_state *pipe_config,
+				const struct drm_connector_state *conn_state)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_digital_port *dig_port = intel_mst->primary;
+	struct intel_dp *intel_dp = &dig_port->dp;
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct drm_dp_mst_topology_state *mst_state =
+		drm_atomic_get_new_mst_topology_state(&state->base, &intel_dp->mst_mgr);
+	enum transcoder trans = pipe_config->cpu_transcoder;
+
+	drm_WARN_ON(&dev_priv->drm, pipe_config->has_pch_encoder);
+
+	clear_act_sent(encoder, pipe_config);
+
+	if (intel_dp_is_uhbr(pipe_config)) {
+		const struct drm_display_mode *adjusted_mode =
+			&pipe_config->hw.adjusted_mode;
+		u64 crtc_clock_hz = KHz(adjusted_mode->crtc_clock);
+
+		intel_de_write(dev_priv, TRANS_DP2_VFREQHIGH(pipe_config->cpu_transcoder),
+			       TRANS_DP2_VFREQ_PIXEL_CLOCK(crtc_clock_hz >> 24));
+		intel_de_write(dev_priv, TRANS_DP2_VFREQLOW(pipe_config->cpu_transcoder),
+			       TRANS_DP2_VFREQ_PIXEL_CLOCK(crtc_clock_hz & 0xffffff));
+	}
+
+	intel_ddi_enable_transcoder_func(encoder, pipe_config);
+
+	intel_de_rmw(dev_priv, TRANS_DDI_FUNC_CTL(trans), 0,
+		     TRANS_DDI_DP_VC_PAYLOAD_ALLOC);
+
+	drm_dbg_kms(&dev_priv->drm, "active links %d\n",
+		    intel_dp->active_mst_links);
+
+	wait_for_act_sent(encoder, pipe_config);
+
+	drm_dp_add_payload_part2(&intel_dp->mst_mgr, &state->base,
+				 drm_atomic_get_mst_payload_state(mst_state, connector->port));
+
+	if (DISPLAY_VER(dev_priv) >= 14 && pipe_config->fec_enable)
+		intel_de_rmw(dev_priv, MTL_CHICKEN_TRANS(trans), 0,
+			     FECSTALL_DIS_DPTSTREAM_DPTTG);
+	else if (DISPLAY_VER(dev_priv) >= 12 && pipe_config->fec_enable)
+		intel_de_rmw(dev_priv, CHICKEN_TRANS(trans), 0,
+			     FECSTALL_DIS_DPTSTREAM_DPTTG);
+
+	intel_enable_transcoder(pipe_config);
+
+	intel_crtc_vblank_on(pipe_config);
+
+	intel_audio_codec_enable(encoder, pipe_config, conn_state);
+
+	/* Enable hdcp if it's desired */
+	if (conn_state->content_protection ==
+	    DRM_MODE_CONTENT_PROTECTION_DESIRED)
+		intel_hdcp_enable(to_intel_connector(conn_state->connector),
+				  pipe_config,
+				  (u8)conn_state->hdcp_content_type);
+}
+
+static bool intel_dp_mst_enc_get_hw_state(struct intel_encoder *encoder,
+				      enum pipe *pipe)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	*pipe = intel_mst->pipe;
+	if (intel_mst->connector)
+		return true;
+	return false;
+}
+
+static void intel_dp_mst_enc_get_config(struct intel_encoder *encoder,
+					struct intel_crtc_state *pipe_config)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_digital_port *dig_port = intel_mst->primary;
+
+	dig_port->base.get_config(&dig_port->base, pipe_config);
+}
+
+static bool intel_dp_mst_initial_fastset_check(struct intel_encoder *encoder,
+					       struct intel_crtc_state *crtc_state)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(encoder);
+	struct intel_digital_port *dig_port = intel_mst->primary;
+
+	return intel_dp_initial_fastset_check(&dig_port->base, crtc_state);
+}
+
+static int intel_dp_mst_get_ddc_modes(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_dp *intel_dp = intel_connector->mst_port;
+	struct edid *edid;
+	int ret;
+
+	if (drm_connector_is_unregistered(connector))
+		return intel_connector_update_modes(connector, NULL);
+
+	edid = drm_dp_mst_get_edid(connector, &intel_dp->mst_mgr, intel_connector->port);
+	ret = intel_connector_update_modes(connector, edid);
+	kfree(edid);
+
+	return ret;
+}
+
+static int
+intel_dp_mst_connector_late_register(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	int ret;
+
+	ret = drm_dp_mst_connector_late_register(connector,
+						 intel_connector->port);
+	if (ret < 0)
+		return ret;
+
+	ret = intel_connector_register(connector);
+	if (ret < 0)
+		drm_dp_mst_connector_early_unregister(connector,
+						      intel_connector->port);
+
+	return ret;
+}
+
+static void
+intel_dp_mst_connector_early_unregister(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+
+	intel_connector_unregister(connector);
+	drm_dp_mst_connector_early_unregister(connector,
+					      intel_connector->port);
+}
+
+static const struct drm_connector_funcs intel_dp_mst_connector_funcs = {
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.atomic_get_property = intel_digital_connector_atomic_get_property,
+	.atomic_set_property = intel_digital_connector_atomic_set_property,
+	.late_register = intel_dp_mst_connector_late_register,
+	.early_unregister = intel_dp_mst_connector_early_unregister,
+	.destroy = intel_connector_destroy,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.atomic_duplicate_state = intel_digital_connector_duplicate_state,
+};
+
+static int intel_dp_mst_get_modes(struct drm_connector *connector)
+{
+	return intel_dp_mst_get_ddc_modes(connector);
+}
+
+static int
+intel_dp_mst_mode_valid_ctx(struct drm_connector *connector,
+			    struct drm_display_mode *mode,
+			    struct drm_modeset_acquire_ctx *ctx,
+			    enum drm_mode_status *status)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_dp *intel_dp = intel_connector->mst_port;
+	struct drm_dp_mst_topology_mgr *mgr = &intel_dp->mst_mgr;
+	struct drm_dp_mst_port *port = intel_connector->port;
+	const int min_bpp = 18;
+	int max_dotclk = to_i915(connector->dev)->max_dotclk_freq;
+	int max_rate, mode_rate, max_lanes, max_link_clock;
+	int ret;
+
+	if (drm_connector_is_unregistered(connector)) {
+		*status = MODE_ERROR;
+		return 0;
+	}
+
+	*status = intel_cpu_transcoder_mode_valid(dev_priv, mode);
+	if (*status != MODE_OK)
+		return 0;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLSCAN) {
+		*status = MODE_NO_DBLESCAN;
+		return 0;
+	}
+
+	max_link_clock = intel_dp_max_link_rate(intel_dp);
+	max_lanes = intel_dp_max_lane_count(intel_dp);
+
+	max_rate = intel_dp_max_data_rate(max_link_clock, max_lanes);
+	mode_rate = intel_dp_link_required(mode->clock, min_bpp);
+
+	ret = drm_modeset_lock(&mgr->base.lock, ctx);
+	if (ret)
+		return ret;
+
+	if (mode_rate > max_rate || mode->clock > max_dotclk ||
+	    drm_dp_calc_pbn_mode(mode->clock, min_bpp, false) > port->full_pbn) {
+		*status = MODE_CLOCK_HIGH;
+		return 0;
+	}
+
+	if (mode->clock < 10000) {
+		*status = MODE_CLOCK_LOW;
+		return 0;
+	}
+
+	if (mode->flags & DRM_MODE_FLAG_DBLCLK) {
+		*status = MODE_H_ILLEGAL;
+		return 0;
+	}
+
+	*status = intel_mode_valid_max_plane_size(dev_priv, mode, false);
+	return 0;
+}
+
+static struct drm_encoder *intel_mst_atomic_best_encoder(struct drm_connector *connector,
+							 struct drm_atomic_state *state)
+{
+	struct drm_connector_state *connector_state = drm_atomic_get_new_connector_state(state,
+											 connector);
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_dp *intel_dp = intel_connector->mst_port;
+	struct intel_crtc *crtc = to_intel_crtc(connector_state->crtc);
+
+	return &intel_dp->mst_encoders[crtc->pipe]->base.base;
+}
+
+static int
+intel_dp_mst_detect(struct drm_connector *connector,
+		    struct drm_modeset_acquire_ctx *ctx, bool force)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_dp *intel_dp = intel_connector->mst_port;
+
+	if (!INTEL_DISPLAY_ENABLED(i915))
+		return connector_status_disconnected;
+
+	if (drm_connector_is_unregistered(connector))
+		return connector_status_disconnected;
+
+	return drm_dp_mst_detect_port(connector, ctx, &intel_dp->mst_mgr,
+				      intel_connector->port);
+}
+
+static const struct drm_connector_helper_funcs intel_dp_mst_connector_helper_funcs = {
+	.get_modes = intel_dp_mst_get_modes,
+	.mode_valid_ctx = intel_dp_mst_mode_valid_ctx,
+	.atomic_best_encoder = intel_mst_atomic_best_encoder,
+	.atomic_check = intel_dp_mst_atomic_check,
+	.detect_ctx = intel_dp_mst_detect,
+};
+
+static void intel_dp_mst_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct intel_dp_mst_encoder *intel_mst = enc_to_mst(to_intel_encoder(encoder));
+
+	drm_encoder_cleanup(encoder);
+	kfree(intel_mst);
+}
+
+static const struct drm_encoder_funcs intel_dp_mst_enc_funcs = {
+	.destroy = intel_dp_mst_encoder_destroy,
+};
+
+static bool intel_dp_mst_get_hw_state(struct intel_connector *connector)
+{
+	if (intel_attached_encoder(connector) && connector->base.state->crtc) {
+		enum pipe pipe;
+		if (!intel_attached_encoder(connector)->get_hw_state(intel_attached_encoder(connector), &pipe))
+			return false;
+		return true;
+	}
+	return false;
+}
+
+static struct drm_connector *intel_dp_add_mst_connector(struct drm_dp_mst_topology_mgr *mgr, struct drm_dp_mst_port *port, const char *pathprop)
+{
+	struct intel_dp *intel_dp = container_of(mgr, struct intel_dp, mst_mgr);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_connector *intel_connector;
+	struct drm_connector *connector;
+	enum pipe pipe;
+	int ret;
+
+	intel_connector = intel_connector_alloc();
+	if (!intel_connector)
+		return NULL;
+
+	intel_connector->get_hw_state = intel_dp_mst_get_hw_state;
+	intel_connector->mst_port = intel_dp;
+	intel_connector->port = port;
+	drm_dp_mst_get_port_malloc(port);
+
+	connector = &intel_connector->base;
+	ret = drm_connector_init(dev, connector, &intel_dp_mst_connector_funcs,
+				 DRM_MODE_CONNECTOR_DisplayPort);
+	if (ret) {
+		drm_dp_mst_put_port_malloc(port);
+		intel_connector_free(intel_connector);
+		return NULL;
+	}
+
+	drm_connector_helper_add(connector, &intel_dp_mst_connector_helper_funcs);
+
+	for_each_pipe(dev_priv, pipe) {
+		struct drm_encoder *enc =
+			&intel_dp->mst_encoders[pipe]->base.base;
+
+		ret = drm_connector_attach_encoder(&intel_connector->base, enc);
+		if (ret)
+			goto err;
+	}
+
+	drm_object_attach_property(&connector->base, dev->mode_config.path_property, 0);
+	drm_object_attach_property(&connector->base, dev->mode_config.tile_property, 0);
+
+	ret = drm_connector_set_path_property(connector, pathprop);
+	if (ret)
+		goto err;
+
+	intel_attach_force_audio_property(connector);
+	intel_attach_broadcast_rgb_property(connector);
+
+	ret = intel_dp_hdcp_init(dig_port, intel_connector);
+	if (ret)
+		drm_dbg_kms(&dev_priv->drm, "[%s:%d] HDCP MST init failed, skipping.\n",
+			    connector->name, connector->base.id);
+	/*
+	 * Reuse the prop from the SST connector because we're
+	 * not allowed to create new props after device registration.
+	 */
+	connector->max_bpc_property =
+		intel_dp->attached_connector->base.max_bpc_property;
+	if (connector->max_bpc_property)
+		drm_connector_attach_max_bpc_property(connector, 6, 12);
+
+	return connector;
+
+err:
+	drm_connector_cleanup(connector);
+	return NULL;
+}
+
+static void
+intel_dp_mst_poll_hpd_irq(struct drm_dp_mst_topology_mgr *mgr)
+{
+	struct intel_dp *intel_dp = container_of(mgr, struct intel_dp, mst_mgr);
+
+	intel_hpd_trigger_irq(dp_to_dig_port(intel_dp));
+}
+
+static const struct drm_dp_mst_topology_cbs mst_cbs = {
+	.add_connector = intel_dp_add_mst_connector,
+	.poll_hpd_irq = intel_dp_mst_poll_hpd_irq,
+};
+
+static struct intel_dp_mst_encoder *
+intel_dp_create_fake_mst_encoder(struct intel_digital_port *dig_port, enum pipe pipe)
+{
+	struct intel_dp_mst_encoder *intel_mst;
+	struct intel_encoder *intel_encoder;
+	struct drm_device *dev = dig_port->base.base.dev;
+
+	intel_mst = kzalloc(sizeof(*intel_mst), GFP_KERNEL);
+
+	if (!intel_mst)
+		return NULL;
+
+	intel_mst->pipe = pipe;
+	intel_encoder = &intel_mst->base;
+	intel_mst->primary = dig_port;
+
+	drm_encoder_init(dev, &intel_encoder->base, &intel_dp_mst_enc_funcs,
+			 DRM_MODE_ENCODER_DPMST, "DP-MST %c", pipe_name(pipe));
+
+	intel_encoder->type = INTEL_OUTPUT_DP_MST;
+	intel_encoder->power_domain = dig_port->base.power_domain;
+	intel_encoder->port = dig_port->base.port;
+	intel_encoder->cloneable = 0;
+	/*
+	 * This is wrong, but broken userspace uses the intersection
+	 * of possible_crtcs of all the encoders of a given connector
+	 * to figure out which crtcs can drive said connector. What
+	 * should be used instead is the union of possible_crtcs.
+	 * To keep such userspace functioning we must misconfigure
+	 * this to make sure the intersection is not empty :(
+	 */
+	intel_encoder->pipe_mask = ~0;
+
+	intel_encoder->compute_config = intel_dp_mst_compute_config;
+	intel_encoder->compute_config_late = intel_dp_mst_compute_config_late;
+	intel_encoder->disable = intel_mst_disable_dp;
+	intel_encoder->post_disable = intel_mst_post_disable_dp;
+	intel_encoder->update_pipe = intel_ddi_update_pipe;
+	intel_encoder->pre_pll_enable = intel_mst_pre_pll_enable_dp;
+	intel_encoder->pre_enable = intel_mst_pre_enable_dp;
+	intel_encoder->enable = intel_mst_enable_dp;
+	intel_encoder->get_hw_state = intel_dp_mst_enc_get_hw_state;
+	intel_encoder->get_config = intel_dp_mst_enc_get_config;
+	intel_encoder->initial_fastset_check = intel_dp_mst_initial_fastset_check;
+
+	return intel_mst;
+
+}
+
+static bool
+intel_dp_create_fake_mst_encoders(struct intel_digital_port *dig_port)
+{
+	struct intel_dp *intel_dp = &dig_port->dp;
+	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe)
+		intel_dp->mst_encoders[pipe] = intel_dp_create_fake_mst_encoder(dig_port, pipe);
+	return true;
+}
+
+int
+intel_dp_mst_encoder_active_links(struct intel_digital_port *dig_port)
+{
+	return dig_port->dp.active_mst_links;
+}
+
+int
+intel_dp_mst_encoder_init(struct intel_digital_port *dig_port, int conn_base_id)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_dp *intel_dp = &dig_port->dp;
+	enum port port = dig_port->base.port;
+	int ret;
+
+	if (!HAS_DP_MST(i915) || intel_dp_is_edp(intel_dp))
+		return 0;
+
+	if (DISPLAY_VER(i915) < 12 && port == PORT_A)
+		return 0;
+
+	if (DISPLAY_VER(i915) < 11 && port == PORT_E)
+		return 0;
+
+	intel_dp->mst_mgr.cbs = &mst_cbs;
+
+	/* create encoders */
+	intel_dp_create_fake_mst_encoders(dig_port);
+	ret = drm_dp_mst_topology_mgr_init(&intel_dp->mst_mgr, &i915->drm,
+					   &intel_dp->aux, 16, 3, conn_base_id);
+	if (ret) {
+		intel_dp->mst_mgr.cbs = NULL;
+		return ret;
+	}
+
+	return 0;
+}
+
+bool intel_dp_mst_source_support(struct intel_dp *intel_dp)
+{
+	return intel_dp->mst_mgr.cbs;
+}
+
+void
+intel_dp_mst_encoder_cleanup(struct intel_digital_port *dig_port)
+{
+	struct intel_dp *intel_dp = &dig_port->dp;
+
+	if (!intel_dp_mst_source_support(intel_dp))
+		return;
+
+	drm_dp_mst_topology_mgr_destroy(&intel_dp->mst_mgr);
+	/* encoders will get killed by normal cleanup */
+
+	intel_dp->mst_mgr.cbs = NULL;
+}
+
+bool intel_dp_mst_is_master_trans(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->mst_master_transcoder == crtc_state->cpu_transcoder;
+}
+
+bool intel_dp_mst_is_slave_trans(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->mst_master_transcoder != INVALID_TRANSCODER &&
+	       crtc_state->mst_master_transcoder != crtc_state->cpu_transcoder;
+}
+
+/**
+ * intel_dp_mst_add_topology_state_for_connector - add MST topology state for a connector
+ * @state: atomic state
+ * @connector: connector to add the state for
+ * @crtc: the CRTC @connector is attached to
+ *
+ * Add the MST topology state for @connector to @state.
+ *
+ * Returns 0 on success, negative error code on failure.
+ */
+static int
+intel_dp_mst_add_topology_state_for_connector(struct intel_atomic_state *state,
+					      struct intel_connector *connector,
+					      struct intel_crtc *crtc)
+{
+	struct drm_dp_mst_topology_state *mst_state;
+
+	if (!connector->mst_port)
+		return 0;
+
+	mst_state = drm_atomic_get_mst_topology_state(&state->base,
+						      &connector->mst_port->mst_mgr);
+	if (IS_ERR(mst_state))
+		return PTR_ERR(mst_state);
+
+	mst_state->pending_crtc_mask |= drm_crtc_mask(&crtc->base);
+
+	return 0;
+}
+
+/**
+ * intel_dp_mst_add_topology_state_for_crtc - add MST topology state for a CRTC
+ * @state: atomic state
+ * @crtc: CRTC to add the state for
+ *
+ * Add the MST topology state for @crtc to @state.
+ *
+ * Returns 0 on success, negative error code on failure.
+ */
+int intel_dp_mst_add_topology_state_for_crtc(struct intel_atomic_state *state,
+					     struct intel_crtc *crtc)
+{
+	struct drm_connector *_connector;
+	struct drm_connector_state *conn_state;
+	int i;
+
+	for_each_new_connector_in_state(&state->base, _connector, conn_state, i) {
+		struct intel_connector *connector = to_intel_connector(_connector);
+		int ret;
+
+		if (conn_state->crtc != &crtc->base)
+			continue;
+
+		ret = intel_dp_mst_add_topology_state_for_connector(state, connector, crtc);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dp_mst.h b/drivers/gpu/drm/i915/display/intel_dp_mst.h
new file mode 100644
index 000000000..f1815bb72
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dp_mst.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DP_MST_H__
+#define __INTEL_DP_MST_H__
+
+#include <linux/types.h>
+
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_digital_port;
+struct intel_dp;
+
+int intel_dp_mst_encoder_init(struct intel_digital_port *dig_port, int conn_id);
+void intel_dp_mst_encoder_cleanup(struct intel_digital_port *dig_port);
+int intel_dp_mst_encoder_active_links(struct intel_digital_port *dig_port);
+bool intel_dp_mst_is_master_trans(const struct intel_crtc_state *crtc_state);
+bool intel_dp_mst_is_slave_trans(const struct intel_crtc_state *crtc_state);
+bool intel_dp_mst_source_support(struct intel_dp *intel_dp);
+int intel_dp_mst_add_topology_state_for_crtc(struct intel_atomic_state *state,
+					     struct intel_crtc *crtc);
+
+#endif /* __INTEL_DP_MST_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dpio_phy.c b/drivers/gpu/drm/i915/display/intel_dpio_phy.c
new file mode 100644
index 000000000..8732b8722
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dpio_phy.c
@@ -0,0 +1,1106 @@
+/*
+ * Copyright © 2014-2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include "intel_ddi.h"
+#include "intel_ddi_buf_trans.h"
+#include "intel_de.h"
+#include "intel_display_power_well.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_dpio_phy.h"
+#include "vlv_sideband.h"
+
+/**
+ * DOC: DPIO
+ *
+ * VLV, CHV and BXT have slightly peculiar display PHYs for driving DP/HDMI
+ * ports. DPIO is the name given to such a display PHY. These PHYs
+ * don't follow the standard programming model using direct MMIO
+ * registers, and instead their registers must be accessed trough IOSF
+ * sideband. VLV has one such PHY for driving ports B and C, and CHV
+ * adds another PHY for driving port D. Each PHY responds to specific
+ * IOSF-SB port.
+ *
+ * Each display PHY is made up of one or two channels. Each channel
+ * houses a common lane part which contains the PLL and other common
+ * logic. CH0 common lane also contains the IOSF-SB logic for the
+ * Common Register Interface (CRI) ie. the DPIO registers. CRI clock
+ * must be running when any DPIO registers are accessed.
+ *
+ * In addition to having their own registers, the PHYs are also
+ * controlled through some dedicated signals from the display
+ * controller. These include PLL reference clock enable, PLL enable,
+ * and CRI clock selection, for example.
+ *
+ * Eeach channel also has two splines (also called data lanes), and
+ * each spline is made up of one Physical Access Coding Sub-Layer
+ * (PCS) block and two TX lanes. So each channel has two PCS blocks
+ * and four TX lanes. The TX lanes are used as DP lanes or TMDS
+ * data/clock pairs depending on the output type.
+ *
+ * Additionally the PHY also contains an AUX lane with AUX blocks
+ * for each channel. This is used for DP AUX communication, but
+ * this fact isn't really relevant for the driver since AUX is
+ * controlled from the display controller side. No DPIO registers
+ * need to be accessed during AUX communication,
+ *
+ * Generally on VLV/CHV the common lane corresponds to the pipe and
+ * the spline (PCS/TX) corresponds to the port.
+ *
+ * For dual channel PHY (VLV/CHV):
+ *
+ *  pipe A == CMN/PLL/REF CH0
+ *
+ *  pipe B == CMN/PLL/REF CH1
+ *
+ *  port B == PCS/TX CH0
+ *
+ *  port C == PCS/TX CH1
+ *
+ * This is especially important when we cross the streams
+ * ie. drive port B with pipe B, or port C with pipe A.
+ *
+ * For single channel PHY (CHV):
+ *
+ *  pipe C == CMN/PLL/REF CH0
+ *
+ *  port D == PCS/TX CH0
+ *
+ * On BXT the entire PHY channel corresponds to the port. That means
+ * the PLL is also now associated with the port rather than the pipe,
+ * and so the clock needs to be routed to the appropriate transcoder.
+ * Port A PLL is directly connected to transcoder EDP and port B/C
+ * PLLs can be routed to any transcoder A/B/C.
+ *
+ * Note: DDI0 is digital port B, DD1 is digital port C, and DDI2 is
+ * digital port D (CHV) or port A (BXT). ::
+ *
+ *
+ *     Dual channel PHY (VLV/CHV/BXT)
+ *     ---------------------------------
+ *     |      CH0      |      CH1      |
+ *     |  CMN/PLL/REF  |  CMN/PLL/REF  |
+ *     |---------------|---------------| Display PHY
+ *     | PCS01 | PCS23 | PCS01 | PCS23 |
+ *     |-------|-------|-------|-------|
+ *     |TX0|TX1|TX2|TX3|TX0|TX1|TX2|TX3|
+ *     ---------------------------------
+ *     |     DDI0      |     DDI1      | DP/HDMI ports
+ *     ---------------------------------
+ *
+ *     Single channel PHY (CHV/BXT)
+ *     -----------------
+ *     |      CH0      |
+ *     |  CMN/PLL/REF  |
+ *     |---------------| Display PHY
+ *     | PCS01 | PCS23 |
+ *     |-------|-------|
+ *     |TX0|TX1|TX2|TX3|
+ *     -----------------
+ *     |     DDI2      | DP/HDMI port
+ *     -----------------
+ */
+
+/**
+ * struct bxt_ddi_phy_info - Hold info for a broxton DDI phy
+ */
+struct bxt_ddi_phy_info {
+	/**
+	 * @dual_channel: true if this phy has a second channel.
+	 */
+	bool dual_channel;
+
+	/**
+	 * @rcomp_phy: If -1, indicates this phy has its own rcomp resistor.
+	 * Otherwise the GRC value will be copied from the phy indicated by
+	 * this field.
+	 */
+	enum dpio_phy rcomp_phy;
+
+	/**
+	 * @reset_delay: delay in us to wait before setting the common reset
+	 * bit in BXT_PHY_CTL_FAMILY, which effectively enables the phy.
+	 */
+	int reset_delay;
+
+	/**
+	 * @pwron_mask: Mask with the appropriate bit set that would cause the
+	 * punit to power this phy if written to BXT_P_CR_GT_DISP_PWRON.
+	 */
+	u32 pwron_mask;
+
+	/**
+	 * @channel: struct containing per channel information.
+	 */
+	struct {
+		/**
+		 * @channel.port: which port maps to this channel.
+		 */
+		enum port port;
+	} channel[2];
+};
+
+static const struct bxt_ddi_phy_info bxt_ddi_phy_info[] = {
+	[DPIO_PHY0] = {
+		.dual_channel = true,
+		.rcomp_phy = DPIO_PHY1,
+		.pwron_mask = BIT(0),
+
+		.channel = {
+			[DPIO_CH0] = { .port = PORT_B },
+			[DPIO_CH1] = { .port = PORT_C },
+		}
+	},
+	[DPIO_PHY1] = {
+		.dual_channel = false,
+		.rcomp_phy = -1,
+		.pwron_mask = BIT(1),
+
+		.channel = {
+			[DPIO_CH0] = { .port = PORT_A },
+		}
+	},
+};
+
+static const struct bxt_ddi_phy_info glk_ddi_phy_info[] = {
+	[DPIO_PHY0] = {
+		.dual_channel = false,
+		.rcomp_phy = DPIO_PHY1,
+		.pwron_mask = BIT(0),
+		.reset_delay = 20,
+
+		.channel = {
+			[DPIO_CH0] = { .port = PORT_B },
+		}
+	},
+	[DPIO_PHY1] = {
+		.dual_channel = false,
+		.rcomp_phy = -1,
+		.pwron_mask = BIT(3),
+		.reset_delay = 20,
+
+		.channel = {
+			[DPIO_CH0] = { .port = PORT_A },
+		}
+	},
+	[DPIO_PHY2] = {
+		.dual_channel = false,
+		.rcomp_phy = DPIO_PHY1,
+		.pwron_mask = BIT(1),
+		.reset_delay = 20,
+
+		.channel = {
+			[DPIO_CH0] = { .port = PORT_C },
+		}
+	},
+};
+
+static const struct bxt_ddi_phy_info *
+bxt_get_phy_list(struct drm_i915_private *dev_priv, int *count)
+{
+	if (IS_GEMINILAKE(dev_priv)) {
+		*count =  ARRAY_SIZE(glk_ddi_phy_info);
+		return glk_ddi_phy_info;
+	} else {
+		*count =  ARRAY_SIZE(bxt_ddi_phy_info);
+		return bxt_ddi_phy_info;
+	}
+}
+
+static const struct bxt_ddi_phy_info *
+bxt_get_phy_info(struct drm_i915_private *dev_priv, enum dpio_phy phy)
+{
+	int count;
+	const struct bxt_ddi_phy_info *phy_list =
+		bxt_get_phy_list(dev_priv, &count);
+
+	return &phy_list[phy];
+}
+
+void bxt_port_to_phy_channel(struct drm_i915_private *dev_priv, enum port port,
+			     enum dpio_phy *phy, enum dpio_channel *ch)
+{
+	const struct bxt_ddi_phy_info *phy_info, *phys;
+	int i, count;
+
+	phys = bxt_get_phy_list(dev_priv, &count);
+
+	for (i = 0; i < count; i++) {
+		phy_info = &phys[i];
+
+		if (port == phy_info->channel[DPIO_CH0].port) {
+			*phy = i;
+			*ch = DPIO_CH0;
+			return;
+		}
+
+		if (phy_info->dual_channel &&
+		    port == phy_info->channel[DPIO_CH1].port) {
+			*phy = i;
+			*ch = DPIO_CH1;
+			return;
+		}
+	}
+
+	drm_WARN(&dev_priv->drm, 1, "PHY not found for PORT %c",
+		 port_name(port));
+	*phy = DPIO_PHY0;
+	*ch = DPIO_CH0;
+}
+
+void bxt_ddi_phy_set_signal_levels(struct intel_encoder *encoder,
+				   const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	int level = intel_ddi_level(encoder, crtc_state, 0);
+	const struct intel_ddi_buf_trans *trans;
+	enum dpio_channel ch;
+	enum dpio_phy phy;
+	int n_entries;
+	u32 val;
+
+	trans = encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !trans))
+		return;
+
+	bxt_port_to_phy_channel(dev_priv, encoder->port, &phy, &ch);
+
+	/*
+	 * While we write to the group register to program all lanes at once we
+	 * can read only lane registers and we pick lanes 0/1 for that.
+	 */
+	val = intel_de_read(dev_priv, BXT_PORT_PCS_DW10_LN01(phy, ch));
+	val &= ~(TX2_SWING_CALC_INIT | TX1_SWING_CALC_INIT);
+	intel_de_write(dev_priv, BXT_PORT_PCS_DW10_GRP(phy, ch), val);
+
+	val = intel_de_read(dev_priv, BXT_PORT_TX_DW2_LN0(phy, ch));
+	val &= ~(MARGIN_000 | UNIQ_TRANS_SCALE);
+	val |= trans->entries[level].bxt.margin << MARGIN_000_SHIFT |
+		trans->entries[level].bxt.scale << UNIQ_TRANS_SCALE_SHIFT;
+	intel_de_write(dev_priv, BXT_PORT_TX_DW2_GRP(phy, ch), val);
+
+	val = intel_de_read(dev_priv, BXT_PORT_TX_DW3_LN0(phy, ch));
+	val &= ~SCALE_DCOMP_METHOD;
+	if (trans->entries[level].bxt.enable)
+		val |= SCALE_DCOMP_METHOD;
+
+	if ((val & UNIQUE_TRANGE_EN_METHOD) && !(val & SCALE_DCOMP_METHOD))
+		drm_err(&dev_priv->drm,
+			"Disabled scaling while ouniqetrangenmethod was set");
+
+	intel_de_write(dev_priv, BXT_PORT_TX_DW3_GRP(phy, ch), val);
+
+	val = intel_de_read(dev_priv, BXT_PORT_TX_DW4_LN0(phy, ch));
+	val &= ~DE_EMPHASIS;
+	val |= trans->entries[level].bxt.deemphasis << DEEMPH_SHIFT;
+	intel_de_write(dev_priv, BXT_PORT_TX_DW4_GRP(phy, ch), val);
+
+	val = intel_de_read(dev_priv, BXT_PORT_PCS_DW10_LN01(phy, ch));
+	val |= TX2_SWING_CALC_INIT | TX1_SWING_CALC_INIT;
+	intel_de_write(dev_priv, BXT_PORT_PCS_DW10_GRP(phy, ch), val);
+}
+
+bool bxt_ddi_phy_is_enabled(struct drm_i915_private *dev_priv,
+			    enum dpio_phy phy)
+{
+	const struct bxt_ddi_phy_info *phy_info;
+
+	phy_info = bxt_get_phy_info(dev_priv, phy);
+
+	if (!(intel_de_read(dev_priv, BXT_P_CR_GT_DISP_PWRON) & phy_info->pwron_mask))
+		return false;
+
+	if ((intel_de_read(dev_priv, BXT_PORT_CL1CM_DW0(phy)) &
+	     (PHY_POWER_GOOD | PHY_RESERVED)) != PHY_POWER_GOOD) {
+		drm_dbg(&dev_priv->drm,
+			"DDI PHY %d powered, but power hasn't settled\n", phy);
+
+		return false;
+	}
+
+	if (!(intel_de_read(dev_priv, BXT_PHY_CTL_FAMILY(phy)) & COMMON_RESET_DIS)) {
+		drm_dbg(&dev_priv->drm,
+			"DDI PHY %d powered, but still in reset\n", phy);
+
+		return false;
+	}
+
+	return true;
+}
+
+static u32 bxt_get_grc(struct drm_i915_private *dev_priv, enum dpio_phy phy)
+{
+	u32 val = intel_de_read(dev_priv, BXT_PORT_REF_DW6(phy));
+
+	return (val & GRC_CODE_MASK) >> GRC_CODE_SHIFT;
+}
+
+static void bxt_phy_wait_grc_done(struct drm_i915_private *dev_priv,
+				  enum dpio_phy phy)
+{
+	if (intel_de_wait_for_set(dev_priv, BXT_PORT_REF_DW3(phy),
+				  GRC_DONE, 10))
+		drm_err(&dev_priv->drm, "timeout waiting for PHY%d GRC\n",
+			phy);
+}
+
+static void _bxt_ddi_phy_init(struct drm_i915_private *dev_priv,
+			      enum dpio_phy phy)
+{
+	const struct bxt_ddi_phy_info *phy_info;
+	u32 val;
+
+	phy_info = bxt_get_phy_info(dev_priv, phy);
+
+	if (bxt_ddi_phy_is_enabled(dev_priv, phy)) {
+		/* Still read out the GRC value for state verification */
+		if (phy_info->rcomp_phy != -1)
+			dev_priv->bxt_phy_grc = bxt_get_grc(dev_priv, phy);
+
+		if (bxt_ddi_phy_verify_state(dev_priv, phy)) {
+			drm_dbg(&dev_priv->drm, "DDI PHY %d already enabled, "
+				"won't reprogram it\n", phy);
+			return;
+		}
+
+		drm_dbg(&dev_priv->drm,
+			"DDI PHY %d enabled with invalid state, "
+			"force reprogramming it\n", phy);
+	}
+
+	val = intel_de_read(dev_priv, BXT_P_CR_GT_DISP_PWRON);
+	val |= phy_info->pwron_mask;
+	intel_de_write(dev_priv, BXT_P_CR_GT_DISP_PWRON, val);
+
+	/*
+	 * The PHY registers start out inaccessible and respond to reads with
+	 * all 1s.  Eventually they become accessible as they power up, then
+	 * the reserved bit will give the default 0.  Poll on the reserved bit
+	 * becoming 0 to find when the PHY is accessible.
+	 * The flag should get set in 100us according to the HW team, but
+	 * use 1ms due to occasional timeouts observed with that.
+	 */
+	if (intel_wait_for_register_fw(&dev_priv->uncore,
+				       BXT_PORT_CL1CM_DW0(phy),
+				       PHY_RESERVED | PHY_POWER_GOOD,
+				       PHY_POWER_GOOD,
+				       1))
+		drm_err(&dev_priv->drm, "timeout during PHY%d power on\n",
+			phy);
+
+	/* Program PLL Rcomp code offset */
+	val = intel_de_read(dev_priv, BXT_PORT_CL1CM_DW9(phy));
+	val &= ~IREF0RC_OFFSET_MASK;
+	val |= 0xE4 << IREF0RC_OFFSET_SHIFT;
+	intel_de_write(dev_priv, BXT_PORT_CL1CM_DW9(phy), val);
+
+	val = intel_de_read(dev_priv, BXT_PORT_CL1CM_DW10(phy));
+	val &= ~IREF1RC_OFFSET_MASK;
+	val |= 0xE4 << IREF1RC_OFFSET_SHIFT;
+	intel_de_write(dev_priv, BXT_PORT_CL1CM_DW10(phy), val);
+
+	/* Program power gating */
+	val = intel_de_read(dev_priv, BXT_PORT_CL1CM_DW28(phy));
+	val |= OCL1_POWER_DOWN_EN | DW28_OLDO_DYN_PWR_DOWN_EN |
+		SUS_CLK_CONFIG;
+	intel_de_write(dev_priv, BXT_PORT_CL1CM_DW28(phy), val);
+
+	if (phy_info->dual_channel) {
+		val = intel_de_read(dev_priv, BXT_PORT_CL2CM_DW6(phy));
+		val |= DW6_OLDO_DYN_PWR_DOWN_EN;
+		intel_de_write(dev_priv, BXT_PORT_CL2CM_DW6(phy), val);
+	}
+
+	if (phy_info->rcomp_phy != -1) {
+		u32 grc_code;
+
+		bxt_phy_wait_grc_done(dev_priv, phy_info->rcomp_phy);
+
+		/*
+		 * PHY0 isn't connected to an RCOMP resistor so copy over
+		 * the corresponding calibrated value from PHY1, and disable
+		 * the automatic calibration on PHY0.
+		 */
+		val = dev_priv->bxt_phy_grc = bxt_get_grc(dev_priv,
+							  phy_info->rcomp_phy);
+		grc_code = val << GRC_CODE_FAST_SHIFT |
+			   val << GRC_CODE_SLOW_SHIFT |
+			   val;
+		intel_de_write(dev_priv, BXT_PORT_REF_DW6(phy), grc_code);
+
+		val = intel_de_read(dev_priv, BXT_PORT_REF_DW8(phy));
+		val |= GRC_DIS | GRC_RDY_OVRD;
+		intel_de_write(dev_priv, BXT_PORT_REF_DW8(phy), val);
+	}
+
+	if (phy_info->reset_delay)
+		udelay(phy_info->reset_delay);
+
+	val = intel_de_read(dev_priv, BXT_PHY_CTL_FAMILY(phy));
+	val |= COMMON_RESET_DIS;
+	intel_de_write(dev_priv, BXT_PHY_CTL_FAMILY(phy), val);
+}
+
+void bxt_ddi_phy_uninit(struct drm_i915_private *dev_priv, enum dpio_phy phy)
+{
+	const struct bxt_ddi_phy_info *phy_info;
+	u32 val;
+
+	phy_info = bxt_get_phy_info(dev_priv, phy);
+
+	val = intel_de_read(dev_priv, BXT_PHY_CTL_FAMILY(phy));
+	val &= ~COMMON_RESET_DIS;
+	intel_de_write(dev_priv, BXT_PHY_CTL_FAMILY(phy), val);
+
+	val = intel_de_read(dev_priv, BXT_P_CR_GT_DISP_PWRON);
+	val &= ~phy_info->pwron_mask;
+	intel_de_write(dev_priv, BXT_P_CR_GT_DISP_PWRON, val);
+}
+
+void bxt_ddi_phy_init(struct drm_i915_private *dev_priv, enum dpio_phy phy)
+{
+	const struct bxt_ddi_phy_info *phy_info =
+		bxt_get_phy_info(dev_priv, phy);
+	enum dpio_phy rcomp_phy = phy_info->rcomp_phy;
+	bool was_enabled;
+
+	lockdep_assert_held(&dev_priv->display.power.domains.lock);
+
+	was_enabled = true;
+	if (rcomp_phy != -1)
+		was_enabled = bxt_ddi_phy_is_enabled(dev_priv, rcomp_phy);
+
+	/*
+	 * We need to copy the GRC calibration value from rcomp_phy,
+	 * so make sure it's powered up.
+	 */
+	if (!was_enabled)
+		_bxt_ddi_phy_init(dev_priv, rcomp_phy);
+
+	_bxt_ddi_phy_init(dev_priv, phy);
+
+	if (!was_enabled)
+		bxt_ddi_phy_uninit(dev_priv, rcomp_phy);
+}
+
+static bool __printf(6, 7)
+__phy_reg_verify_state(struct drm_i915_private *dev_priv, enum dpio_phy phy,
+		       i915_reg_t reg, u32 mask, u32 expected,
+		       const char *reg_fmt, ...)
+{
+	struct va_format vaf;
+	va_list args;
+	u32 val;
+
+	val = intel_de_read(dev_priv, reg);
+	if ((val & mask) == expected)
+		return true;
+
+	va_start(args, reg_fmt);
+	vaf.fmt = reg_fmt;
+	vaf.va = &args;
+
+	drm_dbg(&dev_priv->drm, "DDI PHY %d reg %pV [%08x] state mismatch: "
+			 "current %08x, expected %08x (mask %08x)\n",
+			 phy, &vaf, reg.reg, val, (val & ~mask) | expected,
+			 mask);
+
+	va_end(args);
+
+	return false;
+}
+
+bool bxt_ddi_phy_verify_state(struct drm_i915_private *dev_priv,
+			      enum dpio_phy phy)
+{
+	const struct bxt_ddi_phy_info *phy_info;
+	u32 mask;
+	bool ok;
+
+	phy_info = bxt_get_phy_info(dev_priv, phy);
+
+#define _CHK(reg, mask, exp, fmt, ...)					\
+	__phy_reg_verify_state(dev_priv, phy, reg, mask, exp, fmt,	\
+			       ## __VA_ARGS__)
+
+	if (!bxt_ddi_phy_is_enabled(dev_priv, phy))
+		return false;
+
+	ok = true;
+
+	/* PLL Rcomp code offset */
+	ok &= _CHK(BXT_PORT_CL1CM_DW9(phy),
+		    IREF0RC_OFFSET_MASK, 0xe4 << IREF0RC_OFFSET_SHIFT,
+		    "BXT_PORT_CL1CM_DW9(%d)", phy);
+	ok &= _CHK(BXT_PORT_CL1CM_DW10(phy),
+		    IREF1RC_OFFSET_MASK, 0xe4 << IREF1RC_OFFSET_SHIFT,
+		    "BXT_PORT_CL1CM_DW10(%d)", phy);
+
+	/* Power gating */
+	mask = OCL1_POWER_DOWN_EN | DW28_OLDO_DYN_PWR_DOWN_EN | SUS_CLK_CONFIG;
+	ok &= _CHK(BXT_PORT_CL1CM_DW28(phy), mask, mask,
+		    "BXT_PORT_CL1CM_DW28(%d)", phy);
+
+	if (phy_info->dual_channel)
+		ok &= _CHK(BXT_PORT_CL2CM_DW6(phy),
+			   DW6_OLDO_DYN_PWR_DOWN_EN, DW6_OLDO_DYN_PWR_DOWN_EN,
+			   "BXT_PORT_CL2CM_DW6(%d)", phy);
+
+	if (phy_info->rcomp_phy != -1) {
+		u32 grc_code = dev_priv->bxt_phy_grc;
+
+		grc_code = grc_code << GRC_CODE_FAST_SHIFT |
+			   grc_code << GRC_CODE_SLOW_SHIFT |
+			   grc_code;
+		mask = GRC_CODE_FAST_MASK | GRC_CODE_SLOW_MASK |
+		       GRC_CODE_NOM_MASK;
+		ok &= _CHK(BXT_PORT_REF_DW6(phy), mask, grc_code,
+			   "BXT_PORT_REF_DW6(%d)", phy);
+
+		mask = GRC_DIS | GRC_RDY_OVRD;
+		ok &= _CHK(BXT_PORT_REF_DW8(phy), mask, mask,
+			    "BXT_PORT_REF_DW8(%d)", phy);
+	}
+
+	return ok;
+#undef _CHK
+}
+
+u8
+bxt_ddi_phy_calc_lane_lat_optim_mask(u8 lane_count)
+{
+	switch (lane_count) {
+	case 1:
+		return 0;
+	case 2:
+		return BIT(2) | BIT(0);
+	case 4:
+		return BIT(3) | BIT(2) | BIT(0);
+	default:
+		MISSING_CASE(lane_count);
+
+		return 0;
+	}
+}
+
+void bxt_ddi_phy_set_lane_optim_mask(struct intel_encoder *encoder,
+				     u8 lane_lat_optim_mask)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	enum dpio_phy phy;
+	enum dpio_channel ch;
+	int lane;
+
+	bxt_port_to_phy_channel(dev_priv, port, &phy, &ch);
+
+	for (lane = 0; lane < 4; lane++) {
+		u32 val = intel_de_read(dev_priv,
+					BXT_PORT_TX_DW14_LN(phy, ch, lane));
+
+		/*
+		 * Note that on CHV this flag is called UPAR, but has
+		 * the same function.
+		 */
+		val &= ~LATENCY_OPTIM;
+		if (lane_lat_optim_mask & BIT(lane))
+			val |= LATENCY_OPTIM;
+
+		intel_de_write(dev_priv, BXT_PORT_TX_DW14_LN(phy, ch, lane),
+			       val);
+	}
+}
+
+u8
+bxt_ddi_phy_get_lane_lat_optim_mask(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	enum dpio_phy phy;
+	enum dpio_channel ch;
+	int lane;
+	u8 mask;
+
+	bxt_port_to_phy_channel(dev_priv, port, &phy, &ch);
+
+	mask = 0;
+	for (lane = 0; lane < 4; lane++) {
+		u32 val = intel_de_read(dev_priv,
+					BXT_PORT_TX_DW14_LN(phy, ch, lane));
+
+		if (val & LATENCY_OPTIM)
+			mask |= BIT(lane);
+	}
+
+	return mask;
+}
+
+void chv_set_phy_signal_level(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state,
+			      u32 deemph_reg_value, u32 margin_reg_value,
+			      bool uniq_trans_scale)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum dpio_channel ch = vlv_dig_port_to_channel(dig_port);
+	enum pipe pipe = crtc->pipe;
+	u32 val;
+	int i;
+
+	vlv_dpio_get(dev_priv);
+
+	/* Clear calc init */
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW10(ch));
+	val &= ~(DPIO_PCS_SWING_CALC_TX0_TX2 | DPIO_PCS_SWING_CALC_TX1_TX3);
+	val &= ~(DPIO_PCS_TX1DEEMP_MASK | DPIO_PCS_TX2DEEMP_MASK);
+	val |= DPIO_PCS_TX1DEEMP_9P5 | DPIO_PCS_TX2DEEMP_9P5;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW10(ch), val);
+
+	if (crtc_state->lane_count > 2) {
+		val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW10(ch));
+		val &= ~(DPIO_PCS_SWING_CALC_TX0_TX2 | DPIO_PCS_SWING_CALC_TX1_TX3);
+		val &= ~(DPIO_PCS_TX1DEEMP_MASK | DPIO_PCS_TX2DEEMP_MASK);
+		val |= DPIO_PCS_TX1DEEMP_9P5 | DPIO_PCS_TX2DEEMP_9P5;
+		vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW10(ch), val);
+	}
+
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW9(ch));
+	val &= ~(DPIO_PCS_TX1MARGIN_MASK | DPIO_PCS_TX2MARGIN_MASK);
+	val |= DPIO_PCS_TX1MARGIN_000 | DPIO_PCS_TX2MARGIN_000;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW9(ch), val);
+
+	if (crtc_state->lane_count > 2) {
+		val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW9(ch));
+		val &= ~(DPIO_PCS_TX1MARGIN_MASK | DPIO_PCS_TX2MARGIN_MASK);
+		val |= DPIO_PCS_TX1MARGIN_000 | DPIO_PCS_TX2MARGIN_000;
+		vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW9(ch), val);
+	}
+
+	/* Program swing deemph */
+	for (i = 0; i < crtc_state->lane_count; i++) {
+		val = vlv_dpio_read(dev_priv, pipe, CHV_TX_DW4(ch, i));
+		val &= ~DPIO_SWING_DEEMPH9P5_MASK;
+		val |= deemph_reg_value << DPIO_SWING_DEEMPH9P5_SHIFT;
+		vlv_dpio_write(dev_priv, pipe, CHV_TX_DW4(ch, i), val);
+	}
+
+	/* Program swing margin */
+	for (i = 0; i < crtc_state->lane_count; i++) {
+		val = vlv_dpio_read(dev_priv, pipe, CHV_TX_DW2(ch, i));
+
+		val &= ~DPIO_SWING_MARGIN000_MASK;
+		val |= margin_reg_value << DPIO_SWING_MARGIN000_SHIFT;
+
+		/*
+		 * Supposedly this value shouldn't matter when unique transition
+		 * scale is disabled, but in fact it does matter. Let's just
+		 * always program the same value and hope it's OK.
+		 */
+		val &= ~(0xff << DPIO_UNIQ_TRANS_SCALE_SHIFT);
+		val |= 0x9a << DPIO_UNIQ_TRANS_SCALE_SHIFT;
+
+		vlv_dpio_write(dev_priv, pipe, CHV_TX_DW2(ch, i), val);
+	}
+
+	/*
+	 * The document said it needs to set bit 27 for ch0 and bit 26
+	 * for ch1. Might be a typo in the doc.
+	 * For now, for this unique transition scale selection, set bit
+	 * 27 for ch0 and ch1.
+	 */
+	for (i = 0; i < crtc_state->lane_count; i++) {
+		val = vlv_dpio_read(dev_priv, pipe, CHV_TX_DW3(ch, i));
+		if (uniq_trans_scale)
+			val |= DPIO_TX_UNIQ_TRANS_SCALE_EN;
+		else
+			val &= ~DPIO_TX_UNIQ_TRANS_SCALE_EN;
+		vlv_dpio_write(dev_priv, pipe, CHV_TX_DW3(ch, i), val);
+	}
+
+	/* Start swing calculation */
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW10(ch));
+	val |= DPIO_PCS_SWING_CALC_TX0_TX2 | DPIO_PCS_SWING_CALC_TX1_TX3;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW10(ch), val);
+
+	if (crtc_state->lane_count > 2) {
+		val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW10(ch));
+		val |= DPIO_PCS_SWING_CALC_TX0_TX2 | DPIO_PCS_SWING_CALC_TX1_TX3;
+		vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW10(ch), val);
+	}
+
+	vlv_dpio_put(dev_priv);
+}
+
+void chv_data_lane_soft_reset(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state,
+			      bool reset)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum dpio_channel ch = vlv_dig_port_to_channel(enc_to_dig_port(encoder));
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum pipe pipe = crtc->pipe;
+	u32 val;
+
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW0(ch));
+	if (reset)
+		val &= ~(DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET);
+	else
+		val |= DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW0(ch), val);
+
+	if (crtc_state->lane_count > 2) {
+		val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW0(ch));
+		if (reset)
+			val &= ~(DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET);
+		else
+			val |= DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET;
+		vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW0(ch), val);
+	}
+
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW1(ch));
+	val |= CHV_PCS_REQ_SOFTRESET_EN;
+	if (reset)
+		val &= ~DPIO_PCS_CLK_SOFT_RESET;
+	else
+		val |= DPIO_PCS_CLK_SOFT_RESET;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW1(ch), val);
+
+	if (crtc_state->lane_count > 2) {
+		val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW1(ch));
+		val |= CHV_PCS_REQ_SOFTRESET_EN;
+		if (reset)
+			val &= ~DPIO_PCS_CLK_SOFT_RESET;
+		else
+			val |= DPIO_PCS_CLK_SOFT_RESET;
+		vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW1(ch), val);
+	}
+}
+
+void chv_phy_pre_pll_enable(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum dpio_channel ch = vlv_dig_port_to_channel(dig_port);
+	enum pipe pipe = crtc->pipe;
+	unsigned int lane_mask =
+		intel_dp_unused_lane_mask(crtc_state->lane_count);
+	u32 val;
+
+	/*
+	 * Must trick the second common lane into life.
+	 * Otherwise we can't even access the PLL.
+	 */
+	if (ch == DPIO_CH0 && pipe == PIPE_B)
+		dig_port->release_cl2_override =
+			!chv_phy_powergate_ch(dev_priv, DPIO_PHY0, DPIO_CH1, true);
+
+	chv_phy_powergate_lanes(encoder, true, lane_mask);
+
+	vlv_dpio_get(dev_priv);
+
+	/* Assert data lane reset */
+	chv_data_lane_soft_reset(encoder, crtc_state, true);
+
+	/* program left/right clock distribution */
+	if (pipe != PIPE_B) {
+		val = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW5_CH0);
+		val &= ~(CHV_BUFLEFTENA1_MASK | CHV_BUFRIGHTENA1_MASK);
+		if (ch == DPIO_CH0)
+			val |= CHV_BUFLEFTENA1_FORCE;
+		if (ch == DPIO_CH1)
+			val |= CHV_BUFRIGHTENA1_FORCE;
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW5_CH0, val);
+	} else {
+		val = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW1_CH1);
+		val &= ~(CHV_BUFLEFTENA2_MASK | CHV_BUFRIGHTENA2_MASK);
+		if (ch == DPIO_CH0)
+			val |= CHV_BUFLEFTENA2_FORCE;
+		if (ch == DPIO_CH1)
+			val |= CHV_BUFRIGHTENA2_FORCE;
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW1_CH1, val);
+	}
+
+	/* program clock channel usage */
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW8(ch));
+	val |= CHV_PCS_USEDCLKCHANNEL_OVRRIDE;
+	if (pipe != PIPE_B)
+		val &= ~CHV_PCS_USEDCLKCHANNEL;
+	else
+		val |= CHV_PCS_USEDCLKCHANNEL;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW8(ch), val);
+
+	if (crtc_state->lane_count > 2) {
+		val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW8(ch));
+		val |= CHV_PCS_USEDCLKCHANNEL_OVRRIDE;
+		if (pipe != PIPE_B)
+			val &= ~CHV_PCS_USEDCLKCHANNEL;
+		else
+			val |= CHV_PCS_USEDCLKCHANNEL;
+		vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW8(ch), val);
+	}
+
+	/*
+	 * This a a bit weird since generally CL
+	 * matches the pipe, but here we need to
+	 * pick the CL based on the port.
+	 */
+	val = vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW19(ch));
+	if (pipe != PIPE_B)
+		val &= ~CHV_CMN_USEDCLKCHANNEL;
+	else
+		val |= CHV_CMN_USEDCLKCHANNEL;
+	vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW19(ch), val);
+
+	vlv_dpio_put(dev_priv);
+}
+
+void chv_phy_pre_encoder_enable(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum dpio_channel ch = vlv_dig_port_to_channel(dig_port);
+	enum pipe pipe = crtc->pipe;
+	int data, i, stagger;
+	u32 val;
+
+	vlv_dpio_get(dev_priv);
+
+	/* allow hardware to manage TX FIFO reset source */
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW11(ch));
+	val &= ~DPIO_LANEDESKEW_STRAP_OVRD;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW11(ch), val);
+
+	if (crtc_state->lane_count > 2) {
+		val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW11(ch));
+		val &= ~DPIO_LANEDESKEW_STRAP_OVRD;
+		vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW11(ch), val);
+	}
+
+	/* Program Tx lane latency optimal setting*/
+	for (i = 0; i < crtc_state->lane_count; i++) {
+		/* Set the upar bit */
+		if (crtc_state->lane_count == 1)
+			data = 0x0;
+		else
+			data = (i == 1) ? 0x0 : 0x1;
+		vlv_dpio_write(dev_priv, pipe, CHV_TX_DW14(ch, i),
+				data << DPIO_UPAR_SHIFT);
+	}
+
+	/* Data lane stagger programming */
+	if (crtc_state->port_clock > 270000)
+		stagger = 0x18;
+	else if (crtc_state->port_clock > 135000)
+		stagger = 0xd;
+	else if (crtc_state->port_clock > 67500)
+		stagger = 0x7;
+	else if (crtc_state->port_clock > 33750)
+		stagger = 0x4;
+	else
+		stagger = 0x2;
+
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW11(ch));
+	val |= DPIO_TX2_STAGGER_MASK(0x1f);
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW11(ch), val);
+
+	if (crtc_state->lane_count > 2) {
+		val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW11(ch));
+		val |= DPIO_TX2_STAGGER_MASK(0x1f);
+		vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW11(ch), val);
+	}
+
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW12(ch),
+		       DPIO_LANESTAGGER_STRAP(stagger) |
+		       DPIO_LANESTAGGER_STRAP_OVRD |
+		       DPIO_TX1_STAGGER_MASK(0x1f) |
+		       DPIO_TX1_STAGGER_MULT(6) |
+		       DPIO_TX2_STAGGER_MULT(0));
+
+	if (crtc_state->lane_count > 2) {
+		vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW12(ch),
+			       DPIO_LANESTAGGER_STRAP(stagger) |
+			       DPIO_LANESTAGGER_STRAP_OVRD |
+			       DPIO_TX1_STAGGER_MASK(0x1f) |
+			       DPIO_TX1_STAGGER_MULT(7) |
+			       DPIO_TX2_STAGGER_MULT(5));
+	}
+
+	/* Deassert data lane reset */
+	chv_data_lane_soft_reset(encoder, crtc_state, false);
+
+	vlv_dpio_put(dev_priv);
+}
+
+void chv_phy_release_cl2_override(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (dig_port->release_cl2_override) {
+		chv_phy_powergate_ch(dev_priv, DPIO_PHY0, DPIO_CH1, false);
+		dig_port->release_cl2_override = false;
+	}
+}
+
+void chv_phy_post_pll_disable(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *old_crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum pipe pipe = to_intel_crtc(old_crtc_state->uapi.crtc)->pipe;
+	u32 val;
+
+	vlv_dpio_get(dev_priv);
+
+	/* disable left/right clock distribution */
+	if (pipe != PIPE_B) {
+		val = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW5_CH0);
+		val &= ~(CHV_BUFLEFTENA1_MASK | CHV_BUFRIGHTENA1_MASK);
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW5_CH0, val);
+	} else {
+		val = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW1_CH1);
+		val &= ~(CHV_BUFLEFTENA2_MASK | CHV_BUFRIGHTENA2_MASK);
+		vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW1_CH1, val);
+	}
+
+	vlv_dpio_put(dev_priv);
+
+	/*
+	 * Leave the power down bit cleared for at least one
+	 * lane so that chv_powergate_phy_ch() will power
+	 * on something when the channel is otherwise unused.
+	 * When the port is off and the override is removed
+	 * the lanes power down anyway, so otherwise it doesn't
+	 * really matter what the state of power down bits is
+	 * after this.
+	 */
+	chv_phy_powergate_lanes(encoder, false, 0x0);
+}
+
+void vlv_set_phy_signal_level(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state,
+			      u32 demph_reg_value, u32 preemph_reg_value,
+			      u32 uniqtranscale_reg_value, u32 tx3_demph)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum dpio_channel port = vlv_dig_port_to_channel(dig_port);
+	enum pipe pipe = crtc->pipe;
+
+	vlv_dpio_get(dev_priv);
+
+	vlv_dpio_write(dev_priv, pipe, VLV_TX_DW5(port), 0x00000000);
+	vlv_dpio_write(dev_priv, pipe, VLV_TX_DW4(port), demph_reg_value);
+	vlv_dpio_write(dev_priv, pipe, VLV_TX_DW2(port),
+			 uniqtranscale_reg_value);
+	vlv_dpio_write(dev_priv, pipe, VLV_TX_DW3(port), 0x0C782040);
+
+	if (tx3_demph)
+		vlv_dpio_write(dev_priv, pipe, VLV_TX3_DW4(port), tx3_demph);
+
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS_DW11(port), 0x00030000);
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS_DW9(port), preemph_reg_value);
+	vlv_dpio_write(dev_priv, pipe, VLV_TX_DW5(port), DPIO_TX_OCALINIT_EN);
+
+	vlv_dpio_put(dev_priv);
+}
+
+void vlv_phy_pre_pll_enable(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum dpio_channel port = vlv_dig_port_to_channel(dig_port);
+	enum pipe pipe = crtc->pipe;
+
+	/* Program Tx lane resets to default */
+	vlv_dpio_get(dev_priv);
+
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS_DW0(port),
+			 DPIO_PCS_TX_LANE2_RESET |
+			 DPIO_PCS_TX_LANE1_RESET);
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS_DW1(port),
+			 DPIO_PCS_CLK_CRI_RXEB_EIOS_EN |
+			 DPIO_PCS_CLK_CRI_RXDIGFILTSG_EN |
+			 (1<<DPIO_PCS_CLK_DATAWIDTH_SHIFT) |
+				 DPIO_PCS_CLK_SOFT_RESET);
+
+	/* Fix up inter-pair skew failure */
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS_DW12(port), 0x00750f00);
+	vlv_dpio_write(dev_priv, pipe, VLV_TX_DW11(port), 0x00001500);
+	vlv_dpio_write(dev_priv, pipe, VLV_TX_DW14(port), 0x40400000);
+
+	vlv_dpio_put(dev_priv);
+}
+
+void vlv_phy_pre_encoder_enable(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum dpio_channel port = vlv_dig_port_to_channel(dig_port);
+	enum pipe pipe = crtc->pipe;
+	u32 val;
+
+	vlv_dpio_get(dev_priv);
+
+	/* Enable clock channels for this port */
+	val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW8(port));
+	val = 0;
+	if (pipe)
+		val |= (1<<21);
+	else
+		val &= ~(1<<21);
+	val |= 0x001000c4;
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS_DW8(port), val);
+
+	/* Program lane clock */
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS_DW14(port), 0x00760018);
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS_DW23(port), 0x00400888);
+
+	vlv_dpio_put(dev_priv);
+}
+
+void vlv_phy_reset_lanes(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	enum dpio_channel port = vlv_dig_port_to_channel(dig_port);
+	enum pipe pipe = crtc->pipe;
+
+	vlv_dpio_get(dev_priv);
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS_DW0(port), 0x00000000);
+	vlv_dpio_write(dev_priv, pipe, VLV_PCS_DW1(port), 0x00e00060);
+	vlv_dpio_put(dev_priv);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dpio_phy.h b/drivers/gpu/drm/i915/display/intel_dpio_phy.h
new file mode 100644
index 000000000..9c3d008e8
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dpio_phy.h
@@ -0,0 +1,59 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DPIO_PHY_H__
+#define __INTEL_DPIO_PHY_H__
+
+#include <linux/types.h>
+
+enum dpio_channel;
+enum dpio_phy;
+enum port;
+struct drm_i915_private;
+struct intel_crtc_state;
+struct intel_encoder;
+
+void bxt_port_to_phy_channel(struct drm_i915_private *dev_priv, enum port port,
+			     enum dpio_phy *phy, enum dpio_channel *ch);
+void bxt_ddi_phy_set_signal_levels(struct intel_encoder *encoder,
+				   const struct intel_crtc_state *crtc_state);
+void bxt_ddi_phy_init(struct drm_i915_private *dev_priv, enum dpio_phy phy);
+void bxt_ddi_phy_uninit(struct drm_i915_private *dev_priv, enum dpio_phy phy);
+bool bxt_ddi_phy_is_enabled(struct drm_i915_private *dev_priv,
+			    enum dpio_phy phy);
+bool bxt_ddi_phy_verify_state(struct drm_i915_private *dev_priv,
+			      enum dpio_phy phy);
+u8 bxt_ddi_phy_calc_lane_lat_optim_mask(u8 lane_count);
+void bxt_ddi_phy_set_lane_optim_mask(struct intel_encoder *encoder,
+				     u8 lane_lat_optim_mask);
+u8 bxt_ddi_phy_get_lane_lat_optim_mask(struct intel_encoder *encoder);
+
+void chv_set_phy_signal_level(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state,
+			      u32 deemph_reg_value, u32 margin_reg_value,
+			      bool uniq_trans_scale);
+void chv_data_lane_soft_reset(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state,
+			      bool reset);
+void chv_phy_pre_pll_enable(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state);
+void chv_phy_pre_encoder_enable(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state);
+void chv_phy_release_cl2_override(struct intel_encoder *encoder);
+void chv_phy_post_pll_disable(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *old_crtc_state);
+
+void vlv_set_phy_signal_level(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *crtc_state,
+			      u32 demph_reg_value, u32 preemph_reg_value,
+			      u32 uniqtranscale_reg_value, u32 tx3_demph);
+void vlv_phy_pre_pll_enable(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state);
+void vlv_phy_pre_encoder_enable(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state);
+void vlv_phy_reset_lanes(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *old_crtc_state);
+
+#endif /* __INTEL_DPIO_PHY_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dpll.c b/drivers/gpu/drm/i915/display/intel_dpll.c
new file mode 100644
index 000000000..b15ba78d6
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dpll.c
@@ -0,0 +1,2061 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/kernel.h>
+#include <linux/string_helpers.h>
+
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display.h"
+#include "intel_display_types.h"
+#include "intel_dpll.h"
+#include "intel_lvds.h"
+#include "intel_panel.h"
+#include "intel_pps.h"
+#include "intel_snps_phy.h"
+#include "vlv_sideband.h"
+
+struct intel_dpll_funcs {
+	int (*crtc_compute_clock)(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc);
+	int (*crtc_get_shared_dpll)(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc);
+};
+
+struct intel_limit {
+	struct {
+		int min, max;
+	} dot, vco, n, m, m1, m2, p, p1;
+
+	struct {
+		int dot_limit;
+		int p2_slow, p2_fast;
+	} p2;
+};
+static const struct intel_limit intel_limits_i8xx_dac = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 908000, .max = 1512000 },
+	.n = { .min = 2, .max = 16 },
+	.m = { .min = 96, .max = 140 },
+	.m1 = { .min = 18, .max = 26 },
+	.m2 = { .min = 6, .max = 16 },
+	.p = { .min = 4, .max = 128 },
+	.p1 = { .min = 2, .max = 33 },
+	.p2 = { .dot_limit = 165000,
+		.p2_slow = 4, .p2_fast = 2 },
+};
+
+static const struct intel_limit intel_limits_i8xx_dvo = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 908000, .max = 1512000 },
+	.n = { .min = 2, .max = 16 },
+	.m = { .min = 96, .max = 140 },
+	.m1 = { .min = 18, .max = 26 },
+	.m2 = { .min = 6, .max = 16 },
+	.p = { .min = 4, .max = 128 },
+	.p1 = { .min = 2, .max = 33 },
+	.p2 = { .dot_limit = 165000,
+		.p2_slow = 4, .p2_fast = 4 },
+};
+
+static const struct intel_limit intel_limits_i8xx_lvds = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 908000, .max = 1512000 },
+	.n = { .min = 2, .max = 16 },
+	.m = { .min = 96, .max = 140 },
+	.m1 = { .min = 18, .max = 26 },
+	.m2 = { .min = 6, .max = 16 },
+	.p = { .min = 4, .max = 128 },
+	.p1 = { .min = 1, .max = 6 },
+	.p2 = { .dot_limit = 165000,
+		.p2_slow = 14, .p2_fast = 7 },
+};
+
+static const struct intel_limit intel_limits_i9xx_sdvo = {
+	.dot = { .min = 20000, .max = 400000 },
+	.vco = { .min = 1400000, .max = 2800000 },
+	.n = { .min = 1, .max = 6 },
+	.m = { .min = 70, .max = 120 },
+	.m1 = { .min = 8, .max = 18 },
+	.m2 = { .min = 3, .max = 7 },
+	.p = { .min = 5, .max = 80 },
+	.p1 = { .min = 1, .max = 8 },
+	.p2 = { .dot_limit = 200000,
+		.p2_slow = 10, .p2_fast = 5 },
+};
+
+static const struct intel_limit intel_limits_i9xx_lvds = {
+	.dot = { .min = 20000, .max = 400000 },
+	.vco = { .min = 1400000, .max = 2800000 },
+	.n = { .min = 1, .max = 6 },
+	.m = { .min = 70, .max = 120 },
+	.m1 = { .min = 8, .max = 18 },
+	.m2 = { .min = 3, .max = 7 },
+	.p = { .min = 7, .max = 98 },
+	.p1 = { .min = 1, .max = 8 },
+	.p2 = { .dot_limit = 112000,
+		.p2_slow = 14, .p2_fast = 7 },
+};
+
+
+static const struct intel_limit intel_limits_g4x_sdvo = {
+	.dot = { .min = 25000, .max = 270000 },
+	.vco = { .min = 1750000, .max = 3500000},
+	.n = { .min = 1, .max = 4 },
+	.m = { .min = 104, .max = 138 },
+	.m1 = { .min = 17, .max = 23 },
+	.m2 = { .min = 5, .max = 11 },
+	.p = { .min = 10, .max = 30 },
+	.p1 = { .min = 1, .max = 3},
+	.p2 = { .dot_limit = 270000,
+		.p2_slow = 10,
+		.p2_fast = 10
+	},
+};
+
+static const struct intel_limit intel_limits_g4x_hdmi = {
+	.dot = { .min = 22000, .max = 400000 },
+	.vco = { .min = 1750000, .max = 3500000},
+	.n = { .min = 1, .max = 4 },
+	.m = { .min = 104, .max = 138 },
+	.m1 = { .min = 16, .max = 23 },
+	.m2 = { .min = 5, .max = 11 },
+	.p = { .min = 5, .max = 80 },
+	.p1 = { .min = 1, .max = 8},
+	.p2 = { .dot_limit = 165000,
+		.p2_slow = 10, .p2_fast = 5 },
+};
+
+static const struct intel_limit intel_limits_g4x_single_channel_lvds = {
+	.dot = { .min = 20000, .max = 115000 },
+	.vco = { .min = 1750000, .max = 3500000 },
+	.n = { .min = 1, .max = 3 },
+	.m = { .min = 104, .max = 138 },
+	.m1 = { .min = 17, .max = 23 },
+	.m2 = { .min = 5, .max = 11 },
+	.p = { .min = 28, .max = 112 },
+	.p1 = { .min = 2, .max = 8 },
+	.p2 = { .dot_limit = 0,
+		.p2_slow = 14, .p2_fast = 14
+	},
+};
+
+static const struct intel_limit intel_limits_g4x_dual_channel_lvds = {
+	.dot = { .min = 80000, .max = 224000 },
+	.vco = { .min = 1750000, .max = 3500000 },
+	.n = { .min = 1, .max = 3 },
+	.m = { .min = 104, .max = 138 },
+	.m1 = { .min = 17, .max = 23 },
+	.m2 = { .min = 5, .max = 11 },
+	.p = { .min = 14, .max = 42 },
+	.p1 = { .min = 2, .max = 6 },
+	.p2 = { .dot_limit = 0,
+		.p2_slow = 7, .p2_fast = 7
+	},
+};
+
+static const struct intel_limit pnv_limits_sdvo = {
+	.dot = { .min = 20000, .max = 400000},
+	.vco = { .min = 1700000, .max = 3500000 },
+	/* Pineview's Ncounter is a ring counter */
+	.n = { .min = 3, .max = 6 },
+	.m = { .min = 2, .max = 256 },
+	/* Pineview only has one combined m divider, which we treat as m2. */
+	.m1 = { .min = 0, .max = 0 },
+	.m2 = { .min = 0, .max = 254 },
+	.p = { .min = 5, .max = 80 },
+	.p1 = { .min = 1, .max = 8 },
+	.p2 = { .dot_limit = 200000,
+		.p2_slow = 10, .p2_fast = 5 },
+};
+
+static const struct intel_limit pnv_limits_lvds = {
+	.dot = { .min = 20000, .max = 400000 },
+	.vco = { .min = 1700000, .max = 3500000 },
+	.n = { .min = 3, .max = 6 },
+	.m = { .min = 2, .max = 256 },
+	.m1 = { .min = 0, .max = 0 },
+	.m2 = { .min = 0, .max = 254 },
+	.p = { .min = 7, .max = 112 },
+	.p1 = { .min = 1, .max = 8 },
+	.p2 = { .dot_limit = 112000,
+		.p2_slow = 14, .p2_fast = 14 },
+};
+
+/* Ironlake / Sandybridge
+ *
+ * We calculate clock using (register_value + 2) for N/M1/M2, so here
+ * the range value for them is (actual_value - 2).
+ */
+static const struct intel_limit ilk_limits_dac = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 1760000, .max = 3510000 },
+	.n = { .min = 1, .max = 5 },
+	.m = { .min = 79, .max = 127 },
+	.m1 = { .min = 12, .max = 22 },
+	.m2 = { .min = 5, .max = 9 },
+	.p = { .min = 5, .max = 80 },
+	.p1 = { .min = 1, .max = 8 },
+	.p2 = { .dot_limit = 225000,
+		.p2_slow = 10, .p2_fast = 5 },
+};
+
+static const struct intel_limit ilk_limits_single_lvds = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 1760000, .max = 3510000 },
+	.n = { .min = 1, .max = 3 },
+	.m = { .min = 79, .max = 118 },
+	.m1 = { .min = 12, .max = 22 },
+	.m2 = { .min = 5, .max = 9 },
+	.p = { .min = 28, .max = 112 },
+	.p1 = { .min = 2, .max = 8 },
+	.p2 = { .dot_limit = 225000,
+		.p2_slow = 14, .p2_fast = 14 },
+};
+
+static const struct intel_limit ilk_limits_dual_lvds = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 1760000, .max = 3510000 },
+	.n = { .min = 1, .max = 3 },
+	.m = { .min = 79, .max = 127 },
+	.m1 = { .min = 12, .max = 22 },
+	.m2 = { .min = 5, .max = 9 },
+	.p = { .min = 14, .max = 56 },
+	.p1 = { .min = 2, .max = 8 },
+	.p2 = { .dot_limit = 225000,
+		.p2_slow = 7, .p2_fast = 7 },
+};
+
+/* LVDS 100mhz refclk limits. */
+static const struct intel_limit ilk_limits_single_lvds_100m = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 1760000, .max = 3510000 },
+	.n = { .min = 1, .max = 2 },
+	.m = { .min = 79, .max = 126 },
+	.m1 = { .min = 12, .max = 22 },
+	.m2 = { .min = 5, .max = 9 },
+	.p = { .min = 28, .max = 112 },
+	.p1 = { .min = 2, .max = 8 },
+	.p2 = { .dot_limit = 225000,
+		.p2_slow = 14, .p2_fast = 14 },
+};
+
+static const struct intel_limit ilk_limits_dual_lvds_100m = {
+	.dot = { .min = 25000, .max = 350000 },
+	.vco = { .min = 1760000, .max = 3510000 },
+	.n = { .min = 1, .max = 3 },
+	.m = { .min = 79, .max = 126 },
+	.m1 = { .min = 12, .max = 22 },
+	.m2 = { .min = 5, .max = 9 },
+	.p = { .min = 14, .max = 42 },
+	.p1 = { .min = 2, .max = 6 },
+	.p2 = { .dot_limit = 225000,
+		.p2_slow = 7, .p2_fast = 7 },
+};
+
+static const struct intel_limit intel_limits_vlv = {
+	 /*
+	  * These are based on the data rate limits (measured in fast clocks)
+	  * since those are the strictest limits we have. The fast
+	  * clock and actual rate limits are more relaxed, so checking
+	  * them would make no difference.
+	  */
+	.dot = { .min = 25000, .max = 270000 },
+	.vco = { .min = 4000000, .max = 6000000 },
+	.n = { .min = 1, .max = 7 },
+	.m1 = { .min = 2, .max = 3 },
+	.m2 = { .min = 11, .max = 156 },
+	.p1 = { .min = 2, .max = 3 },
+	.p2 = { .p2_slow = 2, .p2_fast = 20 }, /* slow=min, fast=max */
+};
+
+static const struct intel_limit intel_limits_chv = {
+	/*
+	 * These are based on the data rate limits (measured in fast clocks)
+	 * since those are the strictest limits we have.  The fast
+	 * clock and actual rate limits are more relaxed, so checking
+	 * them would make no difference.
+	 */
+	.dot = { .min = 25000, .max = 540000 },
+	.vco = { .min = 4800000, .max = 6480000 },
+	.n = { .min = 1, .max = 1 },
+	.m1 = { .min = 2, .max = 2 },
+	.m2 = { .min = 24 << 22, .max = 175 << 22 },
+	.p1 = { .min = 2, .max = 4 },
+	.p2 = {	.p2_slow = 1, .p2_fast = 14 },
+};
+
+static const struct intel_limit intel_limits_bxt = {
+	.dot = { .min = 25000, .max = 594000 },
+	.vco = { .min = 4800000, .max = 6700000 },
+	.n = { .min = 1, .max = 1 },
+	.m1 = { .min = 2, .max = 2 },
+	/* FIXME: find real m2 limits */
+	.m2 = { .min = 2 << 22, .max = 255 << 22 },
+	.p1 = { .min = 2, .max = 4 },
+	.p2 = { .p2_slow = 1, .p2_fast = 20 },
+};
+
+/*
+ * Platform specific helpers to calculate the port PLL loopback- (clock.m),
+ * and post-divider (clock.p) values, pre- (clock.vco) and post-divided fast
+ * (clock.dot) clock rates. This fast dot clock is fed to the port's IO logic.
+ * The helpers' return value is the rate of the clock that is fed to the
+ * display engine's pipe which can be the above fast dot clock rate or a
+ * divided-down version of it.
+ */
+/* m1 is reserved as 0 in Pineview, n is a ring counter */
+int pnv_calc_dpll_params(int refclk, struct dpll *clock)
+{
+	clock->m = clock->m2 + 2;
+	clock->p = clock->p1 * clock->p2;
+	if (WARN_ON(clock->n == 0 || clock->p == 0))
+		return 0;
+	clock->vco = DIV_ROUND_CLOSEST(refclk * clock->m, clock->n);
+	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
+
+	return clock->dot;
+}
+
+static u32 i9xx_dpll_compute_m(const struct dpll *dpll)
+{
+	return 5 * (dpll->m1 + 2) + (dpll->m2 + 2);
+}
+
+int i9xx_calc_dpll_params(int refclk, struct dpll *clock)
+{
+	clock->m = i9xx_dpll_compute_m(clock);
+	clock->p = clock->p1 * clock->p2;
+	if (WARN_ON(clock->n + 2 == 0 || clock->p == 0))
+		return 0;
+	clock->vco = DIV_ROUND_CLOSEST(refclk * clock->m, clock->n + 2);
+	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
+
+	return clock->dot;
+}
+
+int vlv_calc_dpll_params(int refclk, struct dpll *clock)
+{
+	clock->m = clock->m1 * clock->m2;
+	clock->p = clock->p1 * clock->p2 * 5;
+	if (WARN_ON(clock->n == 0 || clock->p == 0))
+		return 0;
+	clock->vco = DIV_ROUND_CLOSEST(refclk * clock->m, clock->n);
+	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
+
+	return clock->dot;
+}
+
+int chv_calc_dpll_params(int refclk, struct dpll *clock)
+{
+	clock->m = clock->m1 * clock->m2;
+	clock->p = clock->p1 * clock->p2 * 5;
+	if (WARN_ON(clock->n == 0 || clock->p == 0))
+		return 0;
+	clock->vco = DIV_ROUND_CLOSEST_ULL(mul_u32_u32(refclk, clock->m),
+					   clock->n << 22);
+	clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p);
+
+	return clock->dot;
+}
+
+/*
+ * Returns whether the given set of divisors are valid for a given refclk with
+ * the given connectors.
+ */
+static bool intel_pll_is_valid(struct drm_i915_private *dev_priv,
+			       const struct intel_limit *limit,
+			       const struct dpll *clock)
+{
+	if (clock->n < limit->n.min || limit->n.max < clock->n)
+		return false;
+	if (clock->p1 < limit->p1.min || limit->p1.max < clock->p1)
+		return false;
+	if (clock->m2 < limit->m2.min || limit->m2.max < clock->m2)
+		return false;
+	if (clock->m1 < limit->m1.min || limit->m1.max < clock->m1)
+		return false;
+
+	if (!IS_PINEVIEW(dev_priv) && !IS_LP(dev_priv))
+		if (clock->m1 <= clock->m2)
+			return false;
+
+	if (!IS_LP(dev_priv)) {
+		if (clock->p < limit->p.min || limit->p.max < clock->p)
+			return false;
+		if (clock->m < limit->m.min || limit->m.max < clock->m)
+			return false;
+	}
+
+	if (clock->vco < limit->vco.min || limit->vco.max < clock->vco)
+		return false;
+	/* XXX: We may need to be checking "Dot clock" depending on the multiplier,
+	 * connector, etc., rather than just a single range.
+	 */
+	if (clock->dot < limit->dot.min || limit->dot.max < clock->dot)
+		return false;
+
+	return true;
+}
+
+static int
+i9xx_select_p2_div(const struct intel_limit *limit,
+		   const struct intel_crtc_state *crtc_state,
+		   int target)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		/*
+		 * For LVDS just rely on its current settings for dual-channel.
+		 * We haven't figured out how to reliably set up different
+		 * single/dual channel state, if we even can.
+		 */
+		if (intel_is_dual_link_lvds(dev_priv))
+			return limit->p2.p2_fast;
+		else
+			return limit->p2.p2_slow;
+	} else {
+		if (target < limit->p2.dot_limit)
+			return limit->p2.p2_slow;
+		else
+			return limit->p2.p2_fast;
+	}
+}
+
+/*
+ * Returns a set of divisors for the desired target clock with the given
+ * refclk, or FALSE.
+ *
+ * Target and reference clocks are specified in kHz.
+ *
+ * If match_clock is provided, then best_clock P divider must match the P
+ * divider from @match_clock used for LVDS downclocking.
+ */
+static bool
+i9xx_find_best_dpll(const struct intel_limit *limit,
+		    struct intel_crtc_state *crtc_state,
+		    int target, int refclk,
+		    const struct dpll *match_clock,
+		    struct dpll *best_clock)
+{
+	struct drm_device *dev = crtc_state->uapi.crtc->dev;
+	struct dpll clock;
+	int err = target;
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	clock.p2 = i9xx_select_p2_div(limit, crtc_state, target);
+
+	for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max;
+	     clock.m1++) {
+		for (clock.m2 = limit->m2.min;
+		     clock.m2 <= limit->m2.max; clock.m2++) {
+			if (clock.m2 >= clock.m1)
+				break;
+			for (clock.n = limit->n.min;
+			     clock.n <= limit->n.max; clock.n++) {
+				for (clock.p1 = limit->p1.min;
+					clock.p1 <= limit->p1.max; clock.p1++) {
+					int this_err;
+
+					i9xx_calc_dpll_params(refclk, &clock);
+					if (!intel_pll_is_valid(to_i915(dev),
+								limit,
+								&clock))
+						continue;
+					if (match_clock &&
+					    clock.p != match_clock->p)
+						continue;
+
+					this_err = abs(clock.dot - target);
+					if (this_err < err) {
+						*best_clock = clock;
+						err = this_err;
+					}
+				}
+			}
+		}
+	}
+
+	return (err != target);
+}
+
+/*
+ * Returns a set of divisors for the desired target clock with the given
+ * refclk, or FALSE.
+ *
+ * Target and reference clocks are specified in kHz.
+ *
+ * If match_clock is provided, then best_clock P divider must match the P
+ * divider from @match_clock used for LVDS downclocking.
+ */
+static bool
+pnv_find_best_dpll(const struct intel_limit *limit,
+		   struct intel_crtc_state *crtc_state,
+		   int target, int refclk,
+		   const struct dpll *match_clock,
+		   struct dpll *best_clock)
+{
+	struct drm_device *dev = crtc_state->uapi.crtc->dev;
+	struct dpll clock;
+	int err = target;
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	clock.p2 = i9xx_select_p2_div(limit, crtc_state, target);
+
+	for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max;
+	     clock.m1++) {
+		for (clock.m2 = limit->m2.min;
+		     clock.m2 <= limit->m2.max; clock.m2++) {
+			for (clock.n = limit->n.min;
+			     clock.n <= limit->n.max; clock.n++) {
+				for (clock.p1 = limit->p1.min;
+					clock.p1 <= limit->p1.max; clock.p1++) {
+					int this_err;
+
+					pnv_calc_dpll_params(refclk, &clock);
+					if (!intel_pll_is_valid(to_i915(dev),
+								limit,
+								&clock))
+						continue;
+					if (match_clock &&
+					    clock.p != match_clock->p)
+						continue;
+
+					this_err = abs(clock.dot - target);
+					if (this_err < err) {
+						*best_clock = clock;
+						err = this_err;
+					}
+				}
+			}
+		}
+	}
+
+	return (err != target);
+}
+
+/*
+ * Returns a set of divisors for the desired target clock with the given
+ * refclk, or FALSE.
+ *
+ * Target and reference clocks are specified in kHz.
+ *
+ * If match_clock is provided, then best_clock P divider must match the P
+ * divider from @match_clock used for LVDS downclocking.
+ */
+static bool
+g4x_find_best_dpll(const struct intel_limit *limit,
+		   struct intel_crtc_state *crtc_state,
+		   int target, int refclk,
+		   const struct dpll *match_clock,
+		   struct dpll *best_clock)
+{
+	struct drm_device *dev = crtc_state->uapi.crtc->dev;
+	struct dpll clock;
+	int max_n;
+	bool found = false;
+	/* approximately equals target * 0.00585 */
+	int err_most = (target >> 8) + (target >> 9);
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	clock.p2 = i9xx_select_p2_div(limit, crtc_state, target);
+
+	max_n = limit->n.max;
+	/* based on hardware requirement, prefer smaller n to precision */
+	for (clock.n = limit->n.min; clock.n <= max_n; clock.n++) {
+		/* based on hardware requirement, prefere larger m1,m2 */
+		for (clock.m1 = limit->m1.max;
+		     clock.m1 >= limit->m1.min; clock.m1--) {
+			for (clock.m2 = limit->m2.max;
+			     clock.m2 >= limit->m2.min; clock.m2--) {
+				for (clock.p1 = limit->p1.max;
+				     clock.p1 >= limit->p1.min; clock.p1--) {
+					int this_err;
+
+					i9xx_calc_dpll_params(refclk, &clock);
+					if (!intel_pll_is_valid(to_i915(dev),
+								limit,
+								&clock))
+						continue;
+
+					this_err = abs(clock.dot - target);
+					if (this_err < err_most) {
+						*best_clock = clock;
+						err_most = this_err;
+						max_n = clock.n;
+						found = true;
+					}
+				}
+			}
+		}
+	}
+	return found;
+}
+
+/*
+ * Check if the calculated PLL configuration is more optimal compared to the
+ * best configuration and error found so far. Return the calculated error.
+ */
+static bool vlv_PLL_is_optimal(struct drm_device *dev, int target_freq,
+			       const struct dpll *calculated_clock,
+			       const struct dpll *best_clock,
+			       unsigned int best_error_ppm,
+			       unsigned int *error_ppm)
+{
+	/*
+	 * For CHV ignore the error and consider only the P value.
+	 * Prefer a bigger P value based on HW requirements.
+	 */
+	if (IS_CHERRYVIEW(to_i915(dev))) {
+		*error_ppm = 0;
+
+		return calculated_clock->p > best_clock->p;
+	}
+
+	if (drm_WARN_ON_ONCE(dev, !target_freq))
+		return false;
+
+	*error_ppm = div_u64(1000000ULL *
+				abs(target_freq - calculated_clock->dot),
+			     target_freq);
+	/*
+	 * Prefer a better P value over a better (smaller) error if the error
+	 * is small. Ensure this preference for future configurations too by
+	 * setting the error to 0.
+	 */
+	if (*error_ppm < 100 && calculated_clock->p > best_clock->p) {
+		*error_ppm = 0;
+
+		return true;
+	}
+
+	return *error_ppm + 10 < best_error_ppm;
+}
+
+/*
+ * Returns a set of divisors for the desired target clock with the given
+ * refclk, or FALSE.
+ */
+static bool
+vlv_find_best_dpll(const struct intel_limit *limit,
+		   struct intel_crtc_state *crtc_state,
+		   int target, int refclk,
+		   const struct dpll *match_clock,
+		   struct dpll *best_clock)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_device *dev = crtc->base.dev;
+	struct dpll clock;
+	unsigned int bestppm = 1000000;
+	/* min update 19.2 MHz */
+	int max_n = min(limit->n.max, refclk / 19200);
+	bool found = false;
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	/* based on hardware requirement, prefer smaller n to precision */
+	for (clock.n = limit->n.min; clock.n <= max_n; clock.n++) {
+		for (clock.p1 = limit->p1.max; clock.p1 >= limit->p1.min; clock.p1--) {
+			for (clock.p2 = limit->p2.p2_fast; clock.p2 >= limit->p2.p2_slow;
+			     clock.p2 -= clock.p2 > 10 ? 2 : 1) {
+				clock.p = clock.p1 * clock.p2 * 5;
+				/* based on hardware requirement, prefer bigger m1,m2 values */
+				for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max; clock.m1++) {
+					unsigned int ppm;
+
+					clock.m2 = DIV_ROUND_CLOSEST(target * clock.p * clock.n,
+								     refclk * clock.m1);
+
+					vlv_calc_dpll_params(refclk, &clock);
+
+					if (!intel_pll_is_valid(to_i915(dev),
+								limit,
+								&clock))
+						continue;
+
+					if (!vlv_PLL_is_optimal(dev, target,
+								&clock,
+								best_clock,
+								bestppm, &ppm))
+						continue;
+
+					*best_clock = clock;
+					bestppm = ppm;
+					found = true;
+				}
+			}
+		}
+	}
+
+	return found;
+}
+
+/*
+ * Returns a set of divisors for the desired target clock with the given
+ * refclk, or FALSE.
+ */
+static bool
+chv_find_best_dpll(const struct intel_limit *limit,
+		   struct intel_crtc_state *crtc_state,
+		   int target, int refclk,
+		   const struct dpll *match_clock,
+		   struct dpll *best_clock)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_device *dev = crtc->base.dev;
+	unsigned int best_error_ppm;
+	struct dpll clock;
+	u64 m2;
+	int found = false;
+
+	memset(best_clock, 0, sizeof(*best_clock));
+	best_error_ppm = 1000000;
+
+	/*
+	 * Based on hardware doc, the n always set to 1, and m1 always
+	 * set to 2.  If requires to support 200Mhz refclk, we need to
+	 * revisit this because n may not 1 anymore.
+	 */
+	clock.n = 1;
+	clock.m1 = 2;
+
+	for (clock.p1 = limit->p1.max; clock.p1 >= limit->p1.min; clock.p1--) {
+		for (clock.p2 = limit->p2.p2_fast;
+				clock.p2 >= limit->p2.p2_slow;
+				clock.p2 -= clock.p2 > 10 ? 2 : 1) {
+			unsigned int error_ppm;
+
+			clock.p = clock.p1 * clock.p2 * 5;
+
+			m2 = DIV_ROUND_CLOSEST_ULL(mul_u32_u32(target, clock.p * clock.n) << 22,
+						   refclk * clock.m1);
+
+			if (m2 > INT_MAX/clock.m1)
+				continue;
+
+			clock.m2 = m2;
+
+			chv_calc_dpll_params(refclk, &clock);
+
+			if (!intel_pll_is_valid(to_i915(dev), limit, &clock))
+				continue;
+
+			if (!vlv_PLL_is_optimal(dev, target, &clock, best_clock,
+						best_error_ppm, &error_ppm))
+				continue;
+
+			*best_clock = clock;
+			best_error_ppm = error_ppm;
+			found = true;
+		}
+	}
+
+	return found;
+}
+
+bool bxt_find_best_dpll(struct intel_crtc_state *crtc_state,
+			struct dpll *best_clock)
+{
+	const struct intel_limit *limit = &intel_limits_bxt;
+	int refclk = 100000;
+
+	return chv_find_best_dpll(limit, crtc_state,
+				  crtc_state->port_clock, refclk,
+				  NULL, best_clock);
+}
+
+u32 i9xx_dpll_compute_fp(const struct dpll *dpll)
+{
+	return dpll->n << 16 | dpll->m1 << 8 | dpll->m2;
+}
+
+static u32 pnv_dpll_compute_fp(const struct dpll *dpll)
+{
+	return (1 << dpll->n) << 16 | dpll->m2;
+}
+
+static void i9xx_update_pll_dividers(struct intel_crtc_state *crtc_state,
+				     const struct dpll *clock,
+				     const struct dpll *reduced_clock)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 fp, fp2;
+
+	if (IS_PINEVIEW(dev_priv)) {
+		fp = pnv_dpll_compute_fp(clock);
+		fp2 = pnv_dpll_compute_fp(reduced_clock);
+	} else {
+		fp = i9xx_dpll_compute_fp(clock);
+		fp2 = i9xx_dpll_compute_fp(reduced_clock);
+	}
+
+	crtc_state->dpll_hw_state.fp0 = fp;
+	crtc_state->dpll_hw_state.fp1 = fp2;
+}
+
+static void i9xx_compute_dpll(struct intel_crtc_state *crtc_state,
+			      const struct dpll *clock,
+			      const struct dpll *reduced_clock)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 dpll;
+
+	i9xx_update_pll_dividers(crtc_state, clock, reduced_clock);
+
+	dpll = DPLL_VGA_MODE_DIS;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS))
+		dpll |= DPLLB_MODE_LVDS;
+	else
+		dpll |= DPLLB_MODE_DAC_SERIAL;
+
+	if (IS_I945G(dev_priv) || IS_I945GM(dev_priv) ||
+	    IS_G33(dev_priv) || IS_PINEVIEW(dev_priv)) {
+		dpll |= (crtc_state->pixel_multiplier - 1)
+			<< SDVO_MULTIPLIER_SHIFT_HIRES;
+	}
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO) ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		dpll |= DPLL_SDVO_HIGH_SPEED;
+
+	if (intel_crtc_has_dp_encoder(crtc_state))
+		dpll |= DPLL_SDVO_HIGH_SPEED;
+
+	/* compute bitmask from p1 value */
+	if (IS_G4X(dev_priv)) {
+		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+		dpll |= (1 << (reduced_clock->p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
+	} else if (IS_PINEVIEW(dev_priv)) {
+		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT_PINEVIEW;
+		WARN_ON(reduced_clock->p1 != clock->p1);
+	} else {
+		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+		WARN_ON(reduced_clock->p1 != clock->p1);
+	}
+
+	switch (clock->p2) {
+	case 5:
+		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_5;
+		break;
+	case 7:
+		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_7;
+		break;
+	case 10:
+		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_10;
+		break;
+	case 14:
+		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_14;
+		break;
+	}
+	WARN_ON(reduced_clock->p2 != clock->p2);
+
+	if (DISPLAY_VER(dev_priv) >= 4)
+		dpll |= (6 << PLL_LOAD_PULSE_PHASE_SHIFT);
+
+	if (crtc_state->sdvo_tv_clock)
+		dpll |= PLL_REF_INPUT_TVCLKINBC;
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+		 intel_panel_use_ssc(dev_priv))
+		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
+	else
+		dpll |= PLL_REF_INPUT_DREFCLK;
+
+	dpll |= DPLL_VCO_ENABLE;
+	crtc_state->dpll_hw_state.dpll = dpll;
+
+	if (DISPLAY_VER(dev_priv) >= 4) {
+		u32 dpll_md = (crtc_state->pixel_multiplier - 1)
+			<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
+		crtc_state->dpll_hw_state.dpll_md = dpll_md;
+	}
+}
+
+static void i8xx_compute_dpll(struct intel_crtc_state *crtc_state,
+			      const struct dpll *clock,
+			      const struct dpll *reduced_clock)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 dpll;
+
+	i9xx_update_pll_dividers(crtc_state, clock, reduced_clock);
+
+	dpll = DPLL_VGA_MODE_DIS;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+	} else {
+		if (clock->p1 == 2)
+			dpll |= PLL_P1_DIVIDE_BY_TWO;
+		else
+			dpll |= (clock->p1 - 2) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+		if (clock->p2 == 4)
+			dpll |= PLL_P2_DIVIDE_BY_4;
+	}
+	WARN_ON(reduced_clock->p1 != clock->p1);
+	WARN_ON(reduced_clock->p2 != clock->p2);
+
+	/*
+	 * Bspec:
+	 * "[Almador Errata}: For the correct operation of the muxed DVO pins
+	 *  (GDEVSELB/I2Cdata, GIRDBY/I2CClk) and (GFRAMEB/DVI_Data,
+	 *  GTRDYB/DVI_Clk): Bit 31 (DPLL VCO Enable) and Bit 30 (2X Clock
+	 *  Enable) must be set to “1” in both the DPLL A Control Register
+	 *  (06014h-06017h) and DPLL B Control Register (06018h-0601Bh)."
+	 *
+	 * For simplicity We simply keep both bits always enabled in
+	 * both DPLLS. The spec says we should disable the DVO 2X clock
+	 * when not needed, but this seems to work fine in practice.
+	 */
+	if (IS_I830(dev_priv) ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DVO))
+		dpll |= DPLL_DVO_2X_MODE;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+	    intel_panel_use_ssc(dev_priv))
+		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
+	else
+		dpll |= PLL_REF_INPUT_DREFCLK;
+
+	dpll |= DPLL_VCO_ENABLE;
+	crtc_state->dpll_hw_state.dpll = dpll;
+}
+
+static int hsw_crtc_compute_clock(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_encoder *encoder =
+		intel_get_crtc_new_encoder(state, crtc_state);
+	int ret;
+
+	if (DISPLAY_VER(dev_priv) < 11 &&
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI))
+		return 0;
+
+	ret = intel_compute_shared_dplls(state, crtc, encoder);
+	if (ret)
+		return ret;
+
+	/* FIXME this is a mess */
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI))
+		return 0;
+
+	/* CRT dotclock is determined via other means */
+	if (!crtc_state->has_pch_encoder)
+		crtc_state->hw.adjusted_mode.crtc_clock = intel_crtc_dotclock(crtc_state);
+
+	return 0;
+}
+
+static int hsw_crtc_get_shared_dpll(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_encoder *encoder =
+		intel_get_crtc_new_encoder(state, crtc_state);
+
+	if (DISPLAY_VER(dev_priv) < 11 &&
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI))
+		return 0;
+
+	return intel_reserve_shared_dplls(state, crtc, encoder);
+}
+
+static int dg2_crtc_compute_clock(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_encoder *encoder =
+		intel_get_crtc_new_encoder(state, crtc_state);
+	int ret;
+
+	ret = intel_mpllb_calc_state(crtc_state, encoder);
+	if (ret)
+		return ret;
+
+	crtc_state->hw.adjusted_mode.crtc_clock = intel_crtc_dotclock(crtc_state);
+
+	return 0;
+}
+
+static bool ilk_needs_fb_cb_tune(const struct dpll *dpll, int factor)
+{
+	return dpll->m < factor * dpll->n;
+}
+
+static void ilk_update_pll_dividers(struct intel_crtc_state *crtc_state,
+				    const struct dpll *clock,
+				    const struct dpll *reduced_clock)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 fp, fp2;
+	int factor;
+
+	/* Enable autotuning of the PLL clock (if permissible) */
+	factor = 21;
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if ((intel_panel_use_ssc(dev_priv) &&
+		     dev_priv->display.vbt.lvds_ssc_freq == 100000) ||
+		    (HAS_PCH_IBX(dev_priv) &&
+		     intel_is_dual_link_lvds(dev_priv)))
+			factor = 25;
+	} else if (crtc_state->sdvo_tv_clock) {
+		factor = 20;
+	}
+
+	fp = i9xx_dpll_compute_fp(clock);
+	if (ilk_needs_fb_cb_tune(clock, factor))
+		fp |= FP_CB_TUNE;
+
+	fp2 = i9xx_dpll_compute_fp(reduced_clock);
+	if (ilk_needs_fb_cb_tune(reduced_clock, factor))
+		fp2 |= FP_CB_TUNE;
+
+	crtc_state->dpll_hw_state.fp0 = fp;
+	crtc_state->dpll_hw_state.fp1 = fp2;
+}
+
+static void ilk_compute_dpll(struct intel_crtc_state *crtc_state,
+			     const struct dpll *clock,
+			     const struct dpll *reduced_clock)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 dpll;
+
+	ilk_update_pll_dividers(crtc_state, clock, reduced_clock);
+
+	dpll = 0;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS))
+		dpll |= DPLLB_MODE_LVDS;
+	else
+		dpll |= DPLLB_MODE_DAC_SERIAL;
+
+	dpll |= (crtc_state->pixel_multiplier - 1)
+		<< PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO) ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		dpll |= DPLL_SDVO_HIGH_SPEED;
+
+	if (intel_crtc_has_dp_encoder(crtc_state))
+		dpll |= DPLL_SDVO_HIGH_SPEED;
+
+	/*
+	 * The high speed IO clock is only really required for
+	 * SDVO/HDMI/DP, but we also enable it for CRT to make it
+	 * possible to share the DPLL between CRT and HDMI. Enabling
+	 * the clock needlessly does no real harm, except use up a
+	 * bit of power potentially.
+	 *
+	 * We'll limit this to IVB with 3 pipes, since it has only two
+	 * DPLLs and so DPLL sharing is the only way to get three pipes
+	 * driving PCH ports at the same time. On SNB we could do this,
+	 * and potentially avoid enabling the second DPLL, but it's not
+	 * clear if it''s a win or loss power wise. No point in doing
+	 * this on ILK at all since it has a fixed DPLL<->pipe mapping.
+	 */
+	if (INTEL_NUM_PIPES(dev_priv) == 3 &&
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG))
+		dpll |= DPLL_SDVO_HIGH_SPEED;
+
+	/* compute bitmask from p1 value */
+	dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+	/* also FPA1 */
+	dpll |= (1 << (reduced_clock->p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
+
+	switch (clock->p2) {
+	case 5:
+		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_5;
+		break;
+	case 7:
+		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_7;
+		break;
+	case 10:
+		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_10;
+		break;
+	case 14:
+		dpll |= DPLLB_LVDS_P2_CLOCK_DIV_14;
+		break;
+	}
+	WARN_ON(reduced_clock->p2 != clock->p2);
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS) &&
+	    intel_panel_use_ssc(dev_priv))
+		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
+	else
+		dpll |= PLL_REF_INPUT_DREFCLK;
+
+	dpll |= DPLL_VCO_ENABLE;
+
+	crtc_state->dpll_hw_state.dpll = dpll;
+}
+
+static int ilk_crtc_compute_clock(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_limit *limit;
+	int refclk = 120000;
+	int ret;
+
+	/* CPU eDP is the only output that doesn't need a PCH PLL of its own. */
+	if (!crtc_state->has_pch_encoder)
+		return 0;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "using SSC reference clock of %d kHz\n",
+				    dev_priv->display.vbt.lvds_ssc_freq);
+			refclk = dev_priv->display.vbt.lvds_ssc_freq;
+		}
+
+		if (intel_is_dual_link_lvds(dev_priv)) {
+			if (refclk == 100000)
+				limit = &ilk_limits_dual_lvds_100m;
+			else
+				limit = &ilk_limits_dual_lvds;
+		} else {
+			if (refclk == 100000)
+				limit = &ilk_limits_single_lvds_100m;
+			else
+				limit = &ilk_limits_single_lvds;
+		}
+	} else {
+		limit = &ilk_limits_dac;
+	}
+
+	if (!crtc_state->clock_set &&
+	    !g4x_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				refclk, NULL, &crtc_state->dpll))
+		return -EINVAL;
+
+	ilk_compute_dpll(crtc_state, &crtc_state->dpll,
+			 &crtc_state->dpll);
+
+	ret = intel_compute_shared_dplls(state, crtc, NULL);
+	if (ret)
+		return ret;
+
+	crtc_state->port_clock = crtc_state->dpll.dot;
+	crtc_state->hw.adjusted_mode.crtc_clock = intel_crtc_dotclock(crtc_state);
+
+	return ret;
+}
+
+static int ilk_crtc_get_shared_dpll(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	/* CPU eDP is the only output that doesn't need a PCH PLL of its own. */
+	if (!crtc_state->has_pch_encoder)
+		return 0;
+
+	return intel_reserve_shared_dplls(state, crtc, NULL);
+}
+
+void vlv_compute_dpll(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	crtc_state->dpll_hw_state.dpll = DPLL_INTEGRATED_REF_CLK_VLV |
+		DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
+	if (crtc->pipe != PIPE_A)
+		crtc_state->dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
+
+	/* DPLL not used with DSI, but still need the rest set up */
+	if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI))
+		crtc_state->dpll_hw_state.dpll |= DPLL_VCO_ENABLE |
+			DPLL_EXT_BUFFER_ENABLE_VLV;
+
+	crtc_state->dpll_hw_state.dpll_md =
+		(crtc_state->pixel_multiplier - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
+}
+
+void chv_compute_dpll(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	crtc_state->dpll_hw_state.dpll = DPLL_SSC_REF_CLK_CHV |
+		DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
+	if (crtc->pipe != PIPE_A)
+		crtc_state->dpll_hw_state.dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
+
+	/* DPLL not used with DSI, but still need the rest set up */
+	if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI))
+		crtc_state->dpll_hw_state.dpll |= DPLL_VCO_ENABLE;
+
+	crtc_state->dpll_hw_state.dpll_md =
+		(crtc_state->pixel_multiplier - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT;
+}
+
+static int chv_crtc_compute_clock(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_limit *limit = &intel_limits_chv;
+	int refclk = 100000;
+
+	if (!crtc_state->clock_set &&
+	    !chv_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				refclk, NULL, &crtc_state->dpll))
+		return -EINVAL;
+
+	chv_compute_dpll(crtc_state);
+
+	/* FIXME this is a mess */
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI))
+		return 0;
+
+	crtc_state->port_clock = crtc_state->dpll.dot;
+	crtc_state->hw.adjusted_mode.crtc_clock = intel_crtc_dotclock(crtc_state);
+
+	return 0;
+}
+
+static int vlv_crtc_compute_clock(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_limit *limit = &intel_limits_vlv;
+	int refclk = 100000;
+
+	if (!crtc_state->clock_set &&
+	    !vlv_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				refclk, NULL, &crtc_state->dpll)) {
+		return -EINVAL;
+	}
+
+	vlv_compute_dpll(crtc_state);
+
+	/* FIXME this is a mess */
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI))
+		return 0;
+
+	crtc_state->port_clock = crtc_state->dpll.dot;
+	crtc_state->hw.adjusted_mode.crtc_clock = intel_crtc_dotclock(crtc_state);
+
+	return 0;
+}
+
+static int g4x_crtc_compute_clock(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_limit *limit;
+	int refclk = 96000;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			refclk = dev_priv->display.vbt.lvds_ssc_freq;
+			drm_dbg_kms(&dev_priv->drm,
+				    "using SSC reference clock of %d kHz\n",
+				    refclk);
+		}
+
+		if (intel_is_dual_link_lvds(dev_priv))
+			limit = &intel_limits_g4x_dual_channel_lvds;
+		else
+			limit = &intel_limits_g4x_single_channel_lvds;
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI) ||
+		   intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG)) {
+		limit = &intel_limits_g4x_hdmi;
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO)) {
+		limit = &intel_limits_g4x_sdvo;
+	} else {
+		/* The option is for other outputs */
+		limit = &intel_limits_i9xx_sdvo;
+	}
+
+	if (!crtc_state->clock_set &&
+	    !g4x_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				refclk, NULL, &crtc_state->dpll))
+		return -EINVAL;
+
+	i9xx_compute_dpll(crtc_state, &crtc_state->dpll,
+			  &crtc_state->dpll);
+
+	crtc_state->port_clock = crtc_state->dpll.dot;
+	/* FIXME this is a mess */
+	if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_TVOUT))
+		crtc_state->hw.adjusted_mode.crtc_clock = intel_crtc_dotclock(crtc_state);
+
+	return 0;
+}
+
+static int pnv_crtc_compute_clock(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_limit *limit;
+	int refclk = 96000;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			refclk = dev_priv->display.vbt.lvds_ssc_freq;
+			drm_dbg_kms(&dev_priv->drm,
+				    "using SSC reference clock of %d kHz\n",
+				    refclk);
+		}
+
+		limit = &pnv_limits_lvds;
+	} else {
+		limit = &pnv_limits_sdvo;
+	}
+
+	if (!crtc_state->clock_set &&
+	    !pnv_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				refclk, NULL, &crtc_state->dpll))
+		return -EINVAL;
+
+	i9xx_compute_dpll(crtc_state, &crtc_state->dpll,
+			  &crtc_state->dpll);
+
+	crtc_state->port_clock = crtc_state->dpll.dot;
+	crtc_state->hw.adjusted_mode.crtc_clock = intel_crtc_dotclock(crtc_state);
+
+	return 0;
+}
+
+static int i9xx_crtc_compute_clock(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_limit *limit;
+	int refclk = 96000;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			refclk = dev_priv->display.vbt.lvds_ssc_freq;
+			drm_dbg_kms(&dev_priv->drm,
+				    "using SSC reference clock of %d kHz\n",
+				    refclk);
+		}
+
+		limit = &intel_limits_i9xx_lvds;
+	} else {
+		limit = &intel_limits_i9xx_sdvo;
+	}
+
+	if (!crtc_state->clock_set &&
+	    !i9xx_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				 refclk, NULL, &crtc_state->dpll))
+		return -EINVAL;
+
+	i9xx_compute_dpll(crtc_state, &crtc_state->dpll,
+			  &crtc_state->dpll);
+
+	crtc_state->port_clock = crtc_state->dpll.dot;
+	/* FIXME this is a mess */
+	if (!intel_crtc_has_type(crtc_state, INTEL_OUTPUT_TVOUT))
+		crtc_state->hw.adjusted_mode.crtc_clock = intel_crtc_dotclock(crtc_state);
+
+	return 0;
+}
+
+static int i8xx_crtc_compute_clock(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_limit *limit;
+	int refclk = 48000;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_LVDS)) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			refclk = dev_priv->display.vbt.lvds_ssc_freq;
+			drm_dbg_kms(&dev_priv->drm,
+				    "using SSC reference clock of %d kHz\n",
+				    refclk);
+		}
+
+		limit = &intel_limits_i8xx_lvds;
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DVO)) {
+		limit = &intel_limits_i8xx_dvo;
+	} else {
+		limit = &intel_limits_i8xx_dac;
+	}
+
+	if (!crtc_state->clock_set &&
+	    !i9xx_find_best_dpll(limit, crtc_state, crtc_state->port_clock,
+				 refclk, NULL, &crtc_state->dpll))
+		return -EINVAL;
+
+	i8xx_compute_dpll(crtc_state, &crtc_state->dpll,
+			  &crtc_state->dpll);
+
+	crtc_state->port_clock = crtc_state->dpll.dot;
+	crtc_state->hw.adjusted_mode.crtc_clock = intel_crtc_dotclock(crtc_state);
+
+	return 0;
+}
+
+static const struct intel_dpll_funcs dg2_dpll_funcs = {
+	.crtc_compute_clock = dg2_crtc_compute_clock,
+};
+
+static const struct intel_dpll_funcs hsw_dpll_funcs = {
+	.crtc_compute_clock = hsw_crtc_compute_clock,
+	.crtc_get_shared_dpll = hsw_crtc_get_shared_dpll,
+};
+
+static const struct intel_dpll_funcs ilk_dpll_funcs = {
+	.crtc_compute_clock = ilk_crtc_compute_clock,
+	.crtc_get_shared_dpll = ilk_crtc_get_shared_dpll,
+};
+
+static const struct intel_dpll_funcs chv_dpll_funcs = {
+	.crtc_compute_clock = chv_crtc_compute_clock,
+};
+
+static const struct intel_dpll_funcs vlv_dpll_funcs = {
+	.crtc_compute_clock = vlv_crtc_compute_clock,
+};
+
+static const struct intel_dpll_funcs g4x_dpll_funcs = {
+	.crtc_compute_clock = g4x_crtc_compute_clock,
+};
+
+static const struct intel_dpll_funcs pnv_dpll_funcs = {
+	.crtc_compute_clock = pnv_crtc_compute_clock,
+};
+
+static const struct intel_dpll_funcs i9xx_dpll_funcs = {
+	.crtc_compute_clock = i9xx_crtc_compute_clock,
+};
+
+static const struct intel_dpll_funcs i8xx_dpll_funcs = {
+	.crtc_compute_clock = i8xx_crtc_compute_clock,
+};
+
+int intel_dpll_crtc_compute_clock(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	int ret;
+
+	drm_WARN_ON(&i915->drm, !intel_crtc_needs_modeset(crtc_state));
+
+	memset(&crtc_state->dpll_hw_state, 0,
+	       sizeof(crtc_state->dpll_hw_state));
+
+	if (!crtc_state->hw.enable)
+		return 0;
+
+	ret = i915->display.funcs.dpll->crtc_compute_clock(state, crtc);
+	if (ret) {
+		drm_dbg_kms(&i915->drm, "[CRTC:%d:%s] Couldn't calculate DPLL settings\n",
+			    crtc->base.base.id, crtc->base.name);
+		return ret;
+	}
+
+	return 0;
+}
+
+int intel_dpll_crtc_get_shared_dpll(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	int ret;
+
+	drm_WARN_ON(&i915->drm, !intel_crtc_needs_modeset(crtc_state));
+	drm_WARN_ON(&i915->drm, !crtc_state->hw.enable && crtc_state->shared_dpll);
+
+	if (!crtc_state->hw.enable || crtc_state->shared_dpll)
+		return 0;
+
+	if (!i915->display.funcs.dpll->crtc_get_shared_dpll)
+		return 0;
+
+	ret = i915->display.funcs.dpll->crtc_get_shared_dpll(state, crtc);
+	if (ret) {
+		drm_dbg_kms(&i915->drm, "[CRTC:%d:%s] Couldn't get a shared DPLL\n",
+			    crtc->base.base.id, crtc->base.name);
+		return ret;
+	}
+
+	return 0;
+}
+
+void
+intel_dpll_init_clock_hook(struct drm_i915_private *dev_priv)
+{
+	if (IS_DG2(dev_priv))
+		dev_priv->display.funcs.dpll = &dg2_dpll_funcs;
+	else if (DISPLAY_VER(dev_priv) >= 9 || HAS_DDI(dev_priv))
+		dev_priv->display.funcs.dpll = &hsw_dpll_funcs;
+	else if (HAS_PCH_SPLIT(dev_priv))
+		dev_priv->display.funcs.dpll = &ilk_dpll_funcs;
+	else if (IS_CHERRYVIEW(dev_priv))
+		dev_priv->display.funcs.dpll = &chv_dpll_funcs;
+	else if (IS_VALLEYVIEW(dev_priv))
+		dev_priv->display.funcs.dpll = &vlv_dpll_funcs;
+	else if (IS_G4X(dev_priv))
+		dev_priv->display.funcs.dpll = &g4x_dpll_funcs;
+	else if (IS_PINEVIEW(dev_priv))
+		dev_priv->display.funcs.dpll = &pnv_dpll_funcs;
+	else if (DISPLAY_VER(dev_priv) != 2)
+		dev_priv->display.funcs.dpll = &i9xx_dpll_funcs;
+	else
+		dev_priv->display.funcs.dpll = &i8xx_dpll_funcs;
+}
+
+static bool i9xx_has_pps(struct drm_i915_private *dev_priv)
+{
+	if (IS_I830(dev_priv))
+		return false;
+
+	return IS_PINEVIEW(dev_priv) || IS_MOBILE(dev_priv);
+}
+
+void i9xx_enable_pll(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 dpll = crtc_state->dpll_hw_state.dpll;
+	enum pipe pipe = crtc->pipe;
+	int i;
+
+	assert_transcoder_disabled(dev_priv, crtc_state->cpu_transcoder);
+
+	/* PLL is protected by panel, make sure we can write it */
+	if (i9xx_has_pps(dev_priv))
+		assert_pps_unlocked(dev_priv, pipe);
+
+	intel_de_write(dev_priv, FP0(pipe), crtc_state->dpll_hw_state.fp0);
+	intel_de_write(dev_priv, FP1(pipe), crtc_state->dpll_hw_state.fp1);
+
+	/*
+	 * Apparently we need to have VGA mode enabled prior to changing
+	 * the P1/P2 dividers. Otherwise the DPLL will keep using the old
+	 * dividers, even though the register value does change.
+	 */
+	intel_de_write(dev_priv, DPLL(pipe), dpll & ~DPLL_VGA_MODE_DIS);
+	intel_de_write(dev_priv, DPLL(pipe), dpll);
+
+	/* Wait for the clocks to stabilize. */
+	intel_de_posting_read(dev_priv, DPLL(pipe));
+	udelay(150);
+
+	if (DISPLAY_VER(dev_priv) >= 4) {
+		intel_de_write(dev_priv, DPLL_MD(pipe),
+			       crtc_state->dpll_hw_state.dpll_md);
+	} else {
+		/* The pixel multiplier can only be updated once the
+		 * DPLL is enabled and the clocks are stable.
+		 *
+		 * So write it again.
+		 */
+		intel_de_write(dev_priv, DPLL(pipe), dpll);
+	}
+
+	/* We do this three times for luck */
+	for (i = 0; i < 3; i++) {
+		intel_de_write(dev_priv, DPLL(pipe), dpll);
+		intel_de_posting_read(dev_priv, DPLL(pipe));
+		udelay(150); /* wait for warmup */
+	}
+}
+
+static void vlv_pllb_recal_opamp(struct drm_i915_private *dev_priv,
+				 enum pipe pipe)
+{
+	u32 reg_val;
+
+	/*
+	 * PLLB opamp always calibrates to max value of 0x3f, force enable it
+	 * and set it to a reasonable value instead.
+	 */
+	reg_val = vlv_dpio_read(dev_priv, pipe, VLV_PLL_DW9(1));
+	reg_val &= 0xffffff00;
+	reg_val |= 0x00000030;
+	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW9(1), reg_val);
+
+	reg_val = vlv_dpio_read(dev_priv, pipe, VLV_REF_DW13);
+	reg_val &= 0x00ffffff;
+	reg_val |= 0x8c000000;
+	vlv_dpio_write(dev_priv, pipe, VLV_REF_DW13, reg_val);
+
+	reg_val = vlv_dpio_read(dev_priv, pipe, VLV_PLL_DW9(1));
+	reg_val &= 0xffffff00;
+	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW9(1), reg_val);
+
+	reg_val = vlv_dpio_read(dev_priv, pipe, VLV_REF_DW13);
+	reg_val &= 0x00ffffff;
+	reg_val |= 0xb0000000;
+	vlv_dpio_write(dev_priv, pipe, VLV_REF_DW13, reg_val);
+}
+
+static void vlv_prepare_pll(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	u32 mdiv;
+	u32 bestn, bestm1, bestm2, bestp1, bestp2;
+	u32 coreclk, reg_val;
+
+	vlv_dpio_get(dev_priv);
+
+	bestn = crtc_state->dpll.n;
+	bestm1 = crtc_state->dpll.m1;
+	bestm2 = crtc_state->dpll.m2;
+	bestp1 = crtc_state->dpll.p1;
+	bestp2 = crtc_state->dpll.p2;
+
+	/* See eDP HDMI DPIO driver vbios notes doc */
+
+	/* PLL B needs special handling */
+	if (pipe == PIPE_B)
+		vlv_pllb_recal_opamp(dev_priv, pipe);
+
+	/* Set up Tx target for periodic Rcomp update */
+	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW9_BCAST, 0x0100000f);
+
+	/* Disable target IRef on PLL */
+	reg_val = vlv_dpio_read(dev_priv, pipe, VLV_PLL_DW8(pipe));
+	reg_val &= 0x00ffffff;
+	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW8(pipe), reg_val);
+
+	/* Disable fast lock */
+	vlv_dpio_write(dev_priv, pipe, VLV_CMN_DW0, 0x610);
+
+	/* Set idtafcrecal before PLL is enabled */
+	mdiv = ((bestm1 << DPIO_M1DIV_SHIFT) | (bestm2 & DPIO_M2DIV_MASK));
+	mdiv |= ((bestp1 << DPIO_P1_SHIFT) | (bestp2 << DPIO_P2_SHIFT));
+	mdiv |= ((bestn << DPIO_N_SHIFT));
+	mdiv |= (1 << DPIO_K_SHIFT);
+
+	/*
+	 * Post divider depends on pixel clock rate, DAC vs digital (and LVDS,
+	 * but we don't support that).
+	 * Note: don't use the DAC post divider as it seems unstable.
+	 */
+	mdiv |= (DPIO_POST_DIV_HDMIDP << DPIO_POST_DIV_SHIFT);
+	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW3(pipe), mdiv);
+
+	mdiv |= DPIO_ENABLE_CALIBRATION;
+	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW3(pipe), mdiv);
+
+	/* Set HBR and RBR LPF coefficients */
+	if (crtc_state->port_clock == 162000 ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG) ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW10(pipe),
+				 0x009f0003);
+	else
+		vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW10(pipe),
+				 0x00d0000f);
+
+	if (intel_crtc_has_dp_encoder(crtc_state)) {
+		/* Use SSC source */
+		if (pipe == PIPE_A)
+			vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW5(pipe),
+					 0x0df40000);
+		else
+			vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW5(pipe),
+					 0x0df70000);
+	} else { /* HDMI or VGA */
+		/* Use bend source */
+		if (pipe == PIPE_A)
+			vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW5(pipe),
+					 0x0df70000);
+		else
+			vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW5(pipe),
+					 0x0df40000);
+	}
+
+	coreclk = vlv_dpio_read(dev_priv, pipe, VLV_PLL_DW7(pipe));
+	coreclk = (coreclk & 0x0000ff00) | 0x01c00000;
+	if (intel_crtc_has_dp_encoder(crtc_state))
+		coreclk |= 0x01000000;
+	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW7(pipe), coreclk);
+
+	vlv_dpio_write(dev_priv, pipe, VLV_PLL_DW11(pipe), 0x87871000);
+
+	vlv_dpio_put(dev_priv);
+}
+
+static void _vlv_enable_pll(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	intel_de_write(dev_priv, DPLL(pipe), crtc_state->dpll_hw_state.dpll);
+	intel_de_posting_read(dev_priv, DPLL(pipe));
+	udelay(150);
+
+	if (intel_de_wait_for_set(dev_priv, DPLL(pipe), DPLL_LOCK_VLV, 1))
+		drm_err(&dev_priv->drm, "DPLL %d failed to lock\n", pipe);
+}
+
+void vlv_enable_pll(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	assert_transcoder_disabled(dev_priv, crtc_state->cpu_transcoder);
+
+	/* PLL is protected by panel, make sure we can write it */
+	assert_pps_unlocked(dev_priv, pipe);
+
+	/* Enable Refclk */
+	intel_de_write(dev_priv, DPLL(pipe),
+		       crtc_state->dpll_hw_state.dpll &
+		       ~(DPLL_VCO_ENABLE | DPLL_EXT_BUFFER_ENABLE_VLV));
+
+	if (crtc_state->dpll_hw_state.dpll & DPLL_VCO_ENABLE) {
+		vlv_prepare_pll(crtc_state);
+		_vlv_enable_pll(crtc_state);
+	}
+
+	intel_de_write(dev_priv, DPLL_MD(pipe),
+		       crtc_state->dpll_hw_state.dpll_md);
+	intel_de_posting_read(dev_priv, DPLL_MD(pipe));
+}
+
+static void chv_prepare_pll(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	enum dpio_channel port = vlv_pipe_to_channel(pipe);
+	u32 loopfilter, tribuf_calcntr;
+	u32 bestn, bestm1, bestm2, bestp1, bestp2, bestm2_frac;
+	u32 dpio_val;
+	int vco;
+
+	bestn = crtc_state->dpll.n;
+	bestm2_frac = crtc_state->dpll.m2 & 0x3fffff;
+	bestm1 = crtc_state->dpll.m1;
+	bestm2 = crtc_state->dpll.m2 >> 22;
+	bestp1 = crtc_state->dpll.p1;
+	bestp2 = crtc_state->dpll.p2;
+	vco = crtc_state->dpll.vco;
+	dpio_val = 0;
+	loopfilter = 0;
+
+	vlv_dpio_get(dev_priv);
+
+	/* p1 and p2 divider */
+	vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW13(port),
+			5 << DPIO_CHV_S1_DIV_SHIFT |
+			bestp1 << DPIO_CHV_P1_DIV_SHIFT |
+			bestp2 << DPIO_CHV_P2_DIV_SHIFT |
+			1 << DPIO_CHV_K_DIV_SHIFT);
+
+	/* Feedback post-divider - m2 */
+	vlv_dpio_write(dev_priv, pipe, CHV_PLL_DW0(port), bestm2);
+
+	/* Feedback refclk divider - n and m1 */
+	vlv_dpio_write(dev_priv, pipe, CHV_PLL_DW1(port),
+			DPIO_CHV_M1_DIV_BY_2 |
+			1 << DPIO_CHV_N_DIV_SHIFT);
+
+	/* M2 fraction division */
+	vlv_dpio_write(dev_priv, pipe, CHV_PLL_DW2(port), bestm2_frac);
+
+	/* M2 fraction division enable */
+	dpio_val = vlv_dpio_read(dev_priv, pipe, CHV_PLL_DW3(port));
+	dpio_val &= ~(DPIO_CHV_FEEDFWD_GAIN_MASK | DPIO_CHV_FRAC_DIV_EN);
+	dpio_val |= (2 << DPIO_CHV_FEEDFWD_GAIN_SHIFT);
+	if (bestm2_frac)
+		dpio_val |= DPIO_CHV_FRAC_DIV_EN;
+	vlv_dpio_write(dev_priv, pipe, CHV_PLL_DW3(port), dpio_val);
+
+	/* Program digital lock detect threshold */
+	dpio_val = vlv_dpio_read(dev_priv, pipe, CHV_PLL_DW9(port));
+	dpio_val &= ~(DPIO_CHV_INT_LOCK_THRESHOLD_MASK |
+					DPIO_CHV_INT_LOCK_THRESHOLD_SEL_COARSE);
+	dpio_val |= (0x5 << DPIO_CHV_INT_LOCK_THRESHOLD_SHIFT);
+	if (!bestm2_frac)
+		dpio_val |= DPIO_CHV_INT_LOCK_THRESHOLD_SEL_COARSE;
+	vlv_dpio_write(dev_priv, pipe, CHV_PLL_DW9(port), dpio_val);
+
+	/* Loop filter */
+	if (vco == 5400000) {
+		loopfilter |= (0x3 << DPIO_CHV_PROP_COEFF_SHIFT);
+		loopfilter |= (0x8 << DPIO_CHV_INT_COEFF_SHIFT);
+		loopfilter |= (0x1 << DPIO_CHV_GAIN_CTRL_SHIFT);
+		tribuf_calcntr = 0x9;
+	} else if (vco <= 6200000) {
+		loopfilter |= (0x5 << DPIO_CHV_PROP_COEFF_SHIFT);
+		loopfilter |= (0xB << DPIO_CHV_INT_COEFF_SHIFT);
+		loopfilter |= (0x3 << DPIO_CHV_GAIN_CTRL_SHIFT);
+		tribuf_calcntr = 0x9;
+	} else if (vco <= 6480000) {
+		loopfilter |= (0x4 << DPIO_CHV_PROP_COEFF_SHIFT);
+		loopfilter |= (0x9 << DPIO_CHV_INT_COEFF_SHIFT);
+		loopfilter |= (0x3 << DPIO_CHV_GAIN_CTRL_SHIFT);
+		tribuf_calcntr = 0x8;
+	} else {
+		/* Not supported. Apply the same limits as in the max case */
+		loopfilter |= (0x4 << DPIO_CHV_PROP_COEFF_SHIFT);
+		loopfilter |= (0x9 << DPIO_CHV_INT_COEFF_SHIFT);
+		loopfilter |= (0x3 << DPIO_CHV_GAIN_CTRL_SHIFT);
+		tribuf_calcntr = 0;
+	}
+	vlv_dpio_write(dev_priv, pipe, CHV_PLL_DW6(port), loopfilter);
+
+	dpio_val = vlv_dpio_read(dev_priv, pipe, CHV_PLL_DW8(port));
+	dpio_val &= ~DPIO_CHV_TDC_TARGET_CNT_MASK;
+	dpio_val |= (tribuf_calcntr << DPIO_CHV_TDC_TARGET_CNT_SHIFT);
+	vlv_dpio_write(dev_priv, pipe, CHV_PLL_DW8(port), dpio_val);
+
+	/* AFC Recal */
+	vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW14(port),
+			vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW14(port)) |
+			DPIO_AFC_RECAL);
+
+	vlv_dpio_put(dev_priv);
+}
+
+static void _chv_enable_pll(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	enum dpio_channel port = vlv_pipe_to_channel(pipe);
+	u32 tmp;
+
+	vlv_dpio_get(dev_priv);
+
+	/* Enable back the 10bit clock to display controller */
+	tmp = vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW14(port));
+	tmp |= DPIO_DCLKP_EN;
+	vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW14(port), tmp);
+
+	vlv_dpio_put(dev_priv);
+
+	/*
+	 * Need to wait > 100ns between dclkp clock enable bit and PLL enable.
+	 */
+	udelay(1);
+
+	/* Enable PLL */
+	intel_de_write(dev_priv, DPLL(pipe), crtc_state->dpll_hw_state.dpll);
+
+	/* Check PLL is locked */
+	if (intel_de_wait_for_set(dev_priv, DPLL(pipe), DPLL_LOCK_VLV, 1))
+		drm_err(&dev_priv->drm, "PLL %d failed to lock\n", pipe);
+}
+
+void chv_enable_pll(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	assert_transcoder_disabled(dev_priv, crtc_state->cpu_transcoder);
+
+	/* PLL is protected by panel, make sure we can write it */
+	assert_pps_unlocked(dev_priv, pipe);
+
+	/* Enable Refclk and SSC */
+	intel_de_write(dev_priv, DPLL(pipe),
+		       crtc_state->dpll_hw_state.dpll & ~DPLL_VCO_ENABLE);
+
+	if (crtc_state->dpll_hw_state.dpll & DPLL_VCO_ENABLE) {
+		chv_prepare_pll(crtc_state);
+		_chv_enable_pll(crtc_state);
+	}
+
+	if (pipe != PIPE_A) {
+		/*
+		 * WaPixelRepeatModeFixForC0:chv
+		 *
+		 * DPLLCMD is AWOL. Use chicken bits to propagate
+		 * the value from DPLLBMD to either pipe B or C.
+		 */
+		intel_de_write(dev_priv, CBR4_VLV, CBR_DPLLBMD_PIPE(pipe));
+		intel_de_write(dev_priv, DPLL_MD(PIPE_B),
+			       crtc_state->dpll_hw_state.dpll_md);
+		intel_de_write(dev_priv, CBR4_VLV, 0);
+		dev_priv->chv_dpll_md[pipe] = crtc_state->dpll_hw_state.dpll_md;
+
+		/*
+		 * DPLLB VGA mode also seems to cause problems.
+		 * We should always have it disabled.
+		 */
+		drm_WARN_ON(&dev_priv->drm,
+			    (intel_de_read(dev_priv, DPLL(PIPE_B)) &
+			     DPLL_VGA_MODE_DIS) == 0);
+	} else {
+		intel_de_write(dev_priv, DPLL_MD(pipe),
+			       crtc_state->dpll_hw_state.dpll_md);
+		intel_de_posting_read(dev_priv, DPLL_MD(pipe));
+	}
+}
+
+/**
+ * vlv_force_pll_on - forcibly enable just the PLL
+ * @dev_priv: i915 private structure
+ * @pipe: pipe PLL to enable
+ * @dpll: PLL configuration
+ *
+ * Enable the PLL for @pipe using the supplied @dpll config. To be used
+ * in cases where we need the PLL enabled even when @pipe is not going to
+ * be enabled.
+ */
+int vlv_force_pll_on(struct drm_i915_private *dev_priv, enum pipe pipe,
+		     const struct dpll *dpll)
+{
+	struct intel_crtc *crtc = intel_crtc_for_pipe(dev_priv, pipe);
+	struct intel_crtc_state *crtc_state;
+
+	crtc_state = intel_crtc_state_alloc(crtc);
+	if (!crtc_state)
+		return -ENOMEM;
+
+	crtc_state->cpu_transcoder = (enum transcoder)pipe;
+	crtc_state->pixel_multiplier = 1;
+	crtc_state->dpll = *dpll;
+	crtc_state->output_types = BIT(INTEL_OUTPUT_EDP);
+
+	if (IS_CHERRYVIEW(dev_priv)) {
+		chv_compute_dpll(crtc_state);
+		chv_enable_pll(crtc_state);
+	} else {
+		vlv_compute_dpll(crtc_state);
+		vlv_enable_pll(crtc_state);
+	}
+
+	kfree(crtc_state);
+
+	return 0;
+}
+
+void vlv_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	u32 val;
+
+	/* Make sure the pipe isn't still relying on us */
+	assert_transcoder_disabled(dev_priv, (enum transcoder)pipe);
+
+	val = DPLL_INTEGRATED_REF_CLK_VLV |
+		DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
+	if (pipe != PIPE_A)
+		val |= DPLL_INTEGRATED_CRI_CLK_VLV;
+
+	intel_de_write(dev_priv, DPLL(pipe), val);
+	intel_de_posting_read(dev_priv, DPLL(pipe));
+}
+
+void chv_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	enum dpio_channel port = vlv_pipe_to_channel(pipe);
+	u32 val;
+
+	/* Make sure the pipe isn't still relying on us */
+	assert_transcoder_disabled(dev_priv, (enum transcoder)pipe);
+
+	val = DPLL_SSC_REF_CLK_CHV |
+		DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
+	if (pipe != PIPE_A)
+		val |= DPLL_INTEGRATED_CRI_CLK_VLV;
+
+	intel_de_write(dev_priv, DPLL(pipe), val);
+	intel_de_posting_read(dev_priv, DPLL(pipe));
+
+	vlv_dpio_get(dev_priv);
+
+	/* Disable 10bit clock to display controller */
+	val = vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW14(port));
+	val &= ~DPIO_DCLKP_EN;
+	vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW14(port), val);
+
+	vlv_dpio_put(dev_priv);
+}
+
+void i9xx_disable_pll(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	/* Don't disable pipe or pipe PLLs if needed */
+	if (IS_I830(dev_priv))
+		return;
+
+	/* Make sure the pipe isn't still relying on us */
+	assert_transcoder_disabled(dev_priv, crtc_state->cpu_transcoder);
+
+	intel_de_write(dev_priv, DPLL(pipe), DPLL_VGA_MODE_DIS);
+	intel_de_posting_read(dev_priv, DPLL(pipe));
+}
+
+
+/**
+ * vlv_force_pll_off - forcibly disable just the PLL
+ * @dev_priv: i915 private structure
+ * @pipe: pipe PLL to disable
+ *
+ * Disable the PLL for @pipe. To be used in cases where we need
+ * the PLL enabled even when @pipe is not going to be enabled.
+ */
+void vlv_force_pll_off(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	if (IS_CHERRYVIEW(dev_priv))
+		chv_disable_pll(dev_priv, pipe);
+	else
+		vlv_disable_pll(dev_priv, pipe);
+}
+
+/* Only for pre-ILK configs */
+static void assert_pll(struct drm_i915_private *dev_priv,
+		       enum pipe pipe, bool state)
+{
+	bool cur_state;
+
+	cur_state = intel_de_read(dev_priv, DPLL(pipe)) & DPLL_VCO_ENABLE;
+	I915_STATE_WARN(cur_state != state,
+			"PLL state assertion failure (expected %s, current %s)\n",
+			str_on_off(state), str_on_off(cur_state));
+}
+
+void assert_pll_enabled(struct drm_i915_private *i915, enum pipe pipe)
+{
+	assert_pll(i915, pipe, true);
+}
+
+void assert_pll_disabled(struct drm_i915_private *i915, enum pipe pipe)
+{
+	assert_pll(i915, pipe, false);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dpll.h b/drivers/gpu/drm/i915/display/intel_dpll.h
new file mode 100644
index 000000000..bbc30542f
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dpll.h
@@ -0,0 +1,47 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _INTEL_DPLL_H_
+#define _INTEL_DPLL_H_
+
+#include <linux/types.h>
+
+struct dpll;
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+enum pipe;
+
+void intel_dpll_init_clock_hook(struct drm_i915_private *dev_priv);
+int intel_dpll_crtc_compute_clock(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc);
+int intel_dpll_crtc_get_shared_dpll(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc);
+int vlv_calc_dpll_params(int refclk, struct dpll *clock);
+int pnv_calc_dpll_params(int refclk, struct dpll *clock);
+int i9xx_calc_dpll_params(int refclk, struct dpll *clock);
+u32 i9xx_dpll_compute_fp(const struct dpll *dpll);
+void vlv_compute_dpll(struct intel_crtc_state *crtc_state);
+void chv_compute_dpll(struct intel_crtc_state *crtc_state);
+
+int vlv_force_pll_on(struct drm_i915_private *dev_priv, enum pipe pipe,
+		     const struct dpll *dpll);
+void vlv_force_pll_off(struct drm_i915_private *dev_priv, enum pipe pipe);
+
+void chv_enable_pll(const struct intel_crtc_state *crtc_state);
+void chv_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe);
+void vlv_enable_pll(const struct intel_crtc_state *crtc_state);
+void vlv_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe);
+void i9xx_enable_pll(const struct intel_crtc_state *crtc_state);
+void i9xx_disable_pll(const struct intel_crtc_state *crtc_state);
+bool bxt_find_best_dpll(struct intel_crtc_state *crtc_state,
+			struct dpll *best_clock);
+int chv_calc_dpll_params(int refclk, struct dpll *pll_clock);
+
+void assert_pll_enabled(struct drm_i915_private *i915, enum pipe pipe);
+void assert_pll_disabled(struct drm_i915_private *i915, enum pipe pipe);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_dpll_mgr.c b/drivers/gpu/drm/i915/display/intel_dpll_mgr.c
new file mode 100644
index 000000000..64dd603dc
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dpll_mgr.c
@@ -0,0 +1,4564 @@
+/*
+ * Copyright © 2006-2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include <linux/string_helpers.h>
+
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dkl_phy.h"
+#include "intel_dpio_phy.h"
+#include "intel_dpll.h"
+#include "intel_dpll_mgr.h"
+#include "intel_pch_refclk.h"
+#include "intel_tc.h"
+#include "intel_tc_phy_regs.h"
+
+/**
+ * DOC: Display PLLs
+ *
+ * Display PLLs used for driving outputs vary by platform. While some have
+ * per-pipe or per-encoder dedicated PLLs, others allow the use of any PLL
+ * from a pool. In the latter scenario, it is possible that multiple pipes
+ * share a PLL if their configurations match.
+ *
+ * This file provides an abstraction over display PLLs. The function
+ * intel_shared_dpll_init() initializes the PLLs for the given platform.  The
+ * users of a PLL are tracked and that tracking is integrated with the atomic
+ * modset interface. During an atomic operation, required PLLs can be reserved
+ * for a given CRTC and encoder configuration by calling
+ * intel_reserve_shared_dplls() and previously reserved PLLs can be released
+ * with intel_release_shared_dplls().
+ * Changes to the users are first staged in the atomic state, and then made
+ * effective by calling intel_shared_dpll_swap_state() during the atomic
+ * commit phase.
+ */
+
+/* platform specific hooks for managing DPLLs */
+struct intel_shared_dpll_funcs {
+	/*
+	 * Hook for enabling the pll, called from intel_enable_shared_dpll() if
+	 * the pll is not already enabled.
+	 */
+	void (*enable)(struct drm_i915_private *i915,
+		       struct intel_shared_dpll *pll);
+
+	/*
+	 * Hook for disabling the pll, called from intel_disable_shared_dpll()
+	 * only when it is safe to disable the pll, i.e., there are no more
+	 * tracked users for it.
+	 */
+	void (*disable)(struct drm_i915_private *i915,
+			struct intel_shared_dpll *pll);
+
+	/*
+	 * Hook for reading the values currently programmed to the DPLL
+	 * registers. This is used for initial hw state readout and state
+	 * verification after a mode set.
+	 */
+	bool (*get_hw_state)(struct drm_i915_private *i915,
+			     struct intel_shared_dpll *pll,
+			     struct intel_dpll_hw_state *hw_state);
+
+	/*
+	 * Hook for calculating the pll's output frequency based on its passed
+	 * in state.
+	 */
+	int (*get_freq)(struct drm_i915_private *i915,
+			const struct intel_shared_dpll *pll,
+			const struct intel_dpll_hw_state *pll_state);
+};
+
+struct intel_dpll_mgr {
+	const struct dpll_info *dpll_info;
+
+	int (*compute_dplls)(struct intel_atomic_state *state,
+			     struct intel_crtc *crtc,
+			     struct intel_encoder *encoder);
+	int (*get_dplls)(struct intel_atomic_state *state,
+			 struct intel_crtc *crtc,
+			 struct intel_encoder *encoder);
+	void (*put_dplls)(struct intel_atomic_state *state,
+			  struct intel_crtc *crtc);
+	void (*update_active_dpll)(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc,
+				   struct intel_encoder *encoder);
+	void (*update_ref_clks)(struct drm_i915_private *i915);
+	void (*dump_hw_state)(struct drm_i915_private *dev_priv,
+			      const struct intel_dpll_hw_state *hw_state);
+};
+
+static void
+intel_atomic_duplicate_dpll_state(struct drm_i915_private *dev_priv,
+				  struct intel_shared_dpll_state *shared_dpll)
+{
+	enum intel_dpll_id i;
+
+	/* Copy shared dpll state */
+	for (i = 0; i < dev_priv->display.dpll.num_shared_dpll; i++) {
+		struct intel_shared_dpll *pll = &dev_priv->display.dpll.shared_dplls[i];
+
+		shared_dpll[i] = pll->state;
+	}
+}
+
+static struct intel_shared_dpll_state *
+intel_atomic_get_shared_dpll_state(struct drm_atomic_state *s)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(s);
+
+	drm_WARN_ON(s->dev, !drm_modeset_is_locked(&s->dev->mode_config.connection_mutex));
+
+	if (!state->dpll_set) {
+		state->dpll_set = true;
+
+		intel_atomic_duplicate_dpll_state(to_i915(s->dev),
+						  state->shared_dpll);
+	}
+
+	return state->shared_dpll;
+}
+
+/**
+ * intel_get_shared_dpll_by_id - get a DPLL given its id
+ * @dev_priv: i915 device instance
+ * @id: pll id
+ *
+ * Returns:
+ * A pointer to the DPLL with @id
+ */
+struct intel_shared_dpll *
+intel_get_shared_dpll_by_id(struct drm_i915_private *dev_priv,
+			    enum intel_dpll_id id)
+{
+	return &dev_priv->display.dpll.shared_dplls[id];
+}
+
+/**
+ * intel_get_shared_dpll_id - get the id of a DPLL
+ * @dev_priv: i915 device instance
+ * @pll: the DPLL
+ *
+ * Returns:
+ * The id of @pll
+ */
+enum intel_dpll_id
+intel_get_shared_dpll_id(struct drm_i915_private *dev_priv,
+			 struct intel_shared_dpll *pll)
+{
+	long pll_idx = pll - dev_priv->display.dpll.shared_dplls;
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			pll_idx < 0 ||
+			pll_idx >= dev_priv->display.dpll.num_shared_dpll))
+		return -1;
+
+	return pll_idx;
+}
+
+/* For ILK+ */
+void assert_shared_dpll(struct drm_i915_private *dev_priv,
+			struct intel_shared_dpll *pll,
+			bool state)
+{
+	bool cur_state;
+	struct intel_dpll_hw_state hw_state;
+
+	if (drm_WARN(&dev_priv->drm, !pll,
+		     "asserting DPLL %s with no DPLL\n", str_on_off(state)))
+		return;
+
+	cur_state = intel_dpll_get_hw_state(dev_priv, pll, &hw_state);
+	I915_STATE_WARN(cur_state != state,
+	     "%s assertion failure (expected %s, current %s)\n",
+			pll->info->name, str_on_off(state),
+			str_on_off(cur_state));
+}
+
+static enum tc_port icl_pll_id_to_tc_port(enum intel_dpll_id id)
+{
+	return TC_PORT_1 + id - DPLL_ID_ICL_MGPLL1;
+}
+
+enum intel_dpll_id icl_tc_port_to_pll_id(enum tc_port tc_port)
+{
+	return tc_port - TC_PORT_1 + DPLL_ID_ICL_MGPLL1;
+}
+
+static i915_reg_t
+intel_combo_pll_enable_reg(struct drm_i915_private *i915,
+			   struct intel_shared_dpll *pll)
+{
+	if (IS_DG1(i915))
+		return DG1_DPLL_ENABLE(pll->info->id);
+	else if (IS_JSL_EHL(i915) && (pll->info->id == DPLL_ID_EHL_DPLL4))
+		return MG_PLL_ENABLE(0);
+
+	return ICL_DPLL_ENABLE(pll->info->id);
+}
+
+static i915_reg_t
+intel_tc_pll_enable_reg(struct drm_i915_private *i915,
+			struct intel_shared_dpll *pll)
+{
+	const enum intel_dpll_id id = pll->info->id;
+	enum tc_port tc_port = icl_pll_id_to_tc_port(id);
+
+	if (IS_ALDERLAKE_P(i915))
+		return ADLP_PORTTC_PLL_ENABLE(tc_port);
+
+	return MG_PLL_ENABLE(tc_port);
+}
+
+/**
+ * intel_enable_shared_dpll - enable a CRTC's shared DPLL
+ * @crtc_state: CRTC, and its state, which has a shared DPLL
+ *
+ * Enable the shared DPLL used by @crtc.
+ */
+void intel_enable_shared_dpll(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	unsigned int pipe_mask = BIT(crtc->pipe);
+	unsigned int old_mask;
+
+	if (drm_WARN_ON(&dev_priv->drm, pll == NULL))
+		return;
+
+	mutex_lock(&dev_priv->display.dpll.lock);
+	old_mask = pll->active_mask;
+
+	if (drm_WARN_ON(&dev_priv->drm, !(pll->state.pipe_mask & pipe_mask)) ||
+	    drm_WARN_ON(&dev_priv->drm, pll->active_mask & pipe_mask))
+		goto out;
+
+	pll->active_mask |= pipe_mask;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "enable %s (active 0x%x, on? %d) for [CRTC:%d:%s]\n",
+		    pll->info->name, pll->active_mask, pll->on,
+		    crtc->base.base.id, crtc->base.name);
+
+	if (old_mask) {
+		drm_WARN_ON(&dev_priv->drm, !pll->on);
+		assert_shared_dpll_enabled(dev_priv, pll);
+		goto out;
+	}
+	drm_WARN_ON(&dev_priv->drm, pll->on);
+
+	drm_dbg_kms(&dev_priv->drm, "enabling %s\n", pll->info->name);
+	pll->info->funcs->enable(dev_priv, pll);
+	pll->on = true;
+
+out:
+	mutex_unlock(&dev_priv->display.dpll.lock);
+}
+
+/**
+ * intel_disable_shared_dpll - disable a CRTC's shared DPLL
+ * @crtc_state: CRTC, and its state, which has a shared DPLL
+ *
+ * Disable the shared DPLL used by @crtc.
+ */
+void intel_disable_shared_dpll(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_shared_dpll *pll = crtc_state->shared_dpll;
+	unsigned int pipe_mask = BIT(crtc->pipe);
+
+	/* PCH only available on ILK+ */
+	if (DISPLAY_VER(dev_priv) < 5)
+		return;
+
+	if (pll == NULL)
+		return;
+
+	mutex_lock(&dev_priv->display.dpll.lock);
+	if (drm_WARN(&dev_priv->drm, !(pll->active_mask & pipe_mask),
+		     "%s not used by [CRTC:%d:%s]\n", pll->info->name,
+		     crtc->base.base.id, crtc->base.name))
+		goto out;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "disable %s (active 0x%x, on? %d) for [CRTC:%d:%s]\n",
+		    pll->info->name, pll->active_mask, pll->on,
+		    crtc->base.base.id, crtc->base.name);
+
+	assert_shared_dpll_enabled(dev_priv, pll);
+	drm_WARN_ON(&dev_priv->drm, !pll->on);
+
+	pll->active_mask &= ~pipe_mask;
+	if (pll->active_mask)
+		goto out;
+
+	drm_dbg_kms(&dev_priv->drm, "disabling %s\n", pll->info->name);
+	pll->info->funcs->disable(dev_priv, pll);
+	pll->on = false;
+
+out:
+	mutex_unlock(&dev_priv->display.dpll.lock);
+}
+
+static struct intel_shared_dpll *
+intel_find_shared_dpll(struct intel_atomic_state *state,
+		       const struct intel_crtc *crtc,
+		       const struct intel_dpll_hw_state *pll_state,
+		       unsigned long dpll_mask)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_shared_dpll *pll, *unused_pll = NULL;
+	struct intel_shared_dpll_state *shared_dpll;
+	enum intel_dpll_id i;
+
+	shared_dpll = intel_atomic_get_shared_dpll_state(&state->base);
+
+	drm_WARN_ON(&dev_priv->drm, dpll_mask & ~(BIT(I915_NUM_PLLS) - 1));
+
+	for_each_set_bit(i, &dpll_mask, I915_NUM_PLLS) {
+		pll = &dev_priv->display.dpll.shared_dplls[i];
+
+		/* Only want to check enabled timings first */
+		if (shared_dpll[i].pipe_mask == 0) {
+			if (!unused_pll)
+				unused_pll = pll;
+			continue;
+		}
+
+		if (memcmp(pll_state,
+			   &shared_dpll[i].hw_state,
+			   sizeof(*pll_state)) == 0) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "[CRTC:%d:%s] sharing existing %s (pipe mask 0x%x, active 0x%x)\n",
+				    crtc->base.base.id, crtc->base.name,
+				    pll->info->name,
+				    shared_dpll[i].pipe_mask,
+				    pll->active_mask);
+			return pll;
+		}
+	}
+
+	/* Ok no matching timings, maybe there's a free one? */
+	if (unused_pll) {
+		drm_dbg_kms(&dev_priv->drm, "[CRTC:%d:%s] allocated %s\n",
+			    crtc->base.base.id, crtc->base.name,
+			    unused_pll->info->name);
+		return unused_pll;
+	}
+
+	return NULL;
+}
+
+static void
+intel_reference_shared_dpll(struct intel_atomic_state *state,
+			    const struct intel_crtc *crtc,
+			    const struct intel_shared_dpll *pll,
+			    const struct intel_dpll_hw_state *pll_state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_shared_dpll_state *shared_dpll;
+	const enum intel_dpll_id id = pll->info->id;
+
+	shared_dpll = intel_atomic_get_shared_dpll_state(&state->base);
+
+	if (shared_dpll[id].pipe_mask == 0)
+		shared_dpll[id].hw_state = *pll_state;
+
+	drm_dbg(&i915->drm, "using %s for pipe %c\n", pll->info->name,
+		pipe_name(crtc->pipe));
+
+	shared_dpll[id].pipe_mask |= BIT(crtc->pipe);
+}
+
+static void intel_unreference_shared_dpll(struct intel_atomic_state *state,
+					  const struct intel_crtc *crtc,
+					  const struct intel_shared_dpll *pll)
+{
+	struct intel_shared_dpll_state *shared_dpll;
+
+	shared_dpll = intel_atomic_get_shared_dpll_state(&state->base);
+	shared_dpll[pll->info->id].pipe_mask &= ~BIT(crtc->pipe);
+}
+
+static void intel_put_dpll(struct intel_atomic_state *state,
+			   struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	new_crtc_state->shared_dpll = NULL;
+
+	if (!old_crtc_state->shared_dpll)
+		return;
+
+	intel_unreference_shared_dpll(state, crtc, old_crtc_state->shared_dpll);
+}
+
+/**
+ * intel_shared_dpll_swap_state - make atomic DPLL configuration effective
+ * @state: atomic state
+ *
+ * This is the dpll version of drm_atomic_helper_swap_state() since the
+ * helper does not handle driver-specific global state.
+ *
+ * For consistency with atomic helpers this function does a complete swap,
+ * i.e. it also puts the current state into @state, even though there is no
+ * need for that at this moment.
+ */
+void intel_shared_dpll_swap_state(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_shared_dpll_state *shared_dpll = state->shared_dpll;
+	enum intel_dpll_id i;
+
+	if (!state->dpll_set)
+		return;
+
+	for (i = 0; i < dev_priv->display.dpll.num_shared_dpll; i++) {
+		struct intel_shared_dpll *pll =
+			&dev_priv->display.dpll.shared_dplls[i];
+
+		swap(pll->state, shared_dpll[i]);
+	}
+}
+
+static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv,
+				      struct intel_shared_dpll *pll,
+				      struct intel_dpll_hw_state *hw_state)
+{
+	const enum intel_dpll_id id = pll->info->id;
+	intel_wakeref_t wakeref;
+	u32 val;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     POWER_DOMAIN_DISPLAY_CORE);
+	if (!wakeref)
+		return false;
+
+	val = intel_de_read(dev_priv, PCH_DPLL(id));
+	hw_state->dpll = val;
+	hw_state->fp0 = intel_de_read(dev_priv, PCH_FP0(id));
+	hw_state->fp1 = intel_de_read(dev_priv, PCH_FP1(id));
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+
+	return val & DPLL_VCO_ENABLE;
+}
+
+static void ibx_assert_pch_refclk_enabled(struct drm_i915_private *dev_priv)
+{
+	u32 val;
+	bool enabled;
+
+	I915_STATE_WARN_ON(!(HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)));
+
+	val = intel_de_read(dev_priv, PCH_DREF_CONTROL);
+	enabled = !!(val & (DREF_SSC_SOURCE_MASK | DREF_NONSPREAD_SOURCE_MASK |
+			    DREF_SUPERSPREAD_SOURCE_MASK));
+	I915_STATE_WARN(!enabled, "PCH refclk assertion failure, should be active but is disabled\n");
+}
+
+static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv,
+				struct intel_shared_dpll *pll)
+{
+	const enum intel_dpll_id id = pll->info->id;
+
+	/* PCH refclock must be enabled first */
+	ibx_assert_pch_refclk_enabled(dev_priv);
+
+	intel_de_write(dev_priv, PCH_FP0(id), pll->state.hw_state.fp0);
+	intel_de_write(dev_priv, PCH_FP1(id), pll->state.hw_state.fp1);
+
+	intel_de_write(dev_priv, PCH_DPLL(id), pll->state.hw_state.dpll);
+
+	/* Wait for the clocks to stabilize. */
+	intel_de_posting_read(dev_priv, PCH_DPLL(id));
+	udelay(150);
+
+	/* The pixel multiplier can only be updated once the
+	 * DPLL is enabled and the clocks are stable.
+	 *
+	 * So write it again.
+	 */
+	intel_de_write(dev_priv, PCH_DPLL(id), pll->state.hw_state.dpll);
+	intel_de_posting_read(dev_priv, PCH_DPLL(id));
+	udelay(200);
+}
+
+static void ibx_pch_dpll_disable(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll)
+{
+	const enum intel_dpll_id id = pll->info->id;
+
+	intel_de_write(dev_priv, PCH_DPLL(id), 0);
+	intel_de_posting_read(dev_priv, PCH_DPLL(id));
+	udelay(200);
+}
+
+static int ibx_compute_dpll(struct intel_atomic_state *state,
+			    struct intel_crtc *crtc,
+			    struct intel_encoder *encoder)
+{
+	return 0;
+}
+
+static int ibx_get_dpll(struct intel_atomic_state *state,
+			struct intel_crtc *crtc,
+			struct intel_encoder *encoder)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_shared_dpll *pll;
+	enum intel_dpll_id i;
+
+	if (HAS_PCH_IBX(dev_priv)) {
+		/* Ironlake PCH has a fixed PLL->PCH pipe mapping. */
+		i = (enum intel_dpll_id) crtc->pipe;
+		pll = &dev_priv->display.dpll.shared_dplls[i];
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "[CRTC:%d:%s] using pre-allocated %s\n",
+			    crtc->base.base.id, crtc->base.name,
+			    pll->info->name);
+	} else {
+		pll = intel_find_shared_dpll(state, crtc,
+					     &crtc_state->dpll_hw_state,
+					     BIT(DPLL_ID_PCH_PLL_B) |
+					     BIT(DPLL_ID_PCH_PLL_A));
+	}
+
+	if (!pll)
+		return -EINVAL;
+
+	/* reference the pll */
+	intel_reference_shared_dpll(state, crtc,
+				    pll, &crtc_state->dpll_hw_state);
+
+	crtc_state->shared_dpll = pll;
+
+	return 0;
+}
+
+static void ibx_dump_hw_state(struct drm_i915_private *dev_priv,
+			      const struct intel_dpll_hw_state *hw_state)
+{
+	drm_dbg_kms(&dev_priv->drm,
+		    "dpll_hw_state: dpll: 0x%x, dpll_md: 0x%x, "
+		    "fp0: 0x%x, fp1: 0x%x\n",
+		    hw_state->dpll,
+		    hw_state->dpll_md,
+		    hw_state->fp0,
+		    hw_state->fp1);
+}
+
+static const struct intel_shared_dpll_funcs ibx_pch_dpll_funcs = {
+	.enable = ibx_pch_dpll_enable,
+	.disable = ibx_pch_dpll_disable,
+	.get_hw_state = ibx_pch_dpll_get_hw_state,
+};
+
+static const struct dpll_info pch_plls[] = {
+	{ "PCH DPLL A", &ibx_pch_dpll_funcs, DPLL_ID_PCH_PLL_A, 0 },
+	{ "PCH DPLL B", &ibx_pch_dpll_funcs, DPLL_ID_PCH_PLL_B, 0 },
+	{ },
+};
+
+static const struct intel_dpll_mgr pch_pll_mgr = {
+	.dpll_info = pch_plls,
+	.compute_dplls = ibx_compute_dpll,
+	.get_dplls = ibx_get_dpll,
+	.put_dplls = intel_put_dpll,
+	.dump_hw_state = ibx_dump_hw_state,
+};
+
+static void hsw_ddi_wrpll_enable(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll)
+{
+	const enum intel_dpll_id id = pll->info->id;
+
+	intel_de_write(dev_priv, WRPLL_CTL(id), pll->state.hw_state.wrpll);
+	intel_de_posting_read(dev_priv, WRPLL_CTL(id));
+	udelay(20);
+}
+
+static void hsw_ddi_spll_enable(struct drm_i915_private *dev_priv,
+				struct intel_shared_dpll *pll)
+{
+	intel_de_write(dev_priv, SPLL_CTL, pll->state.hw_state.spll);
+	intel_de_posting_read(dev_priv, SPLL_CTL);
+	udelay(20);
+}
+
+static void hsw_ddi_wrpll_disable(struct drm_i915_private *dev_priv,
+				  struct intel_shared_dpll *pll)
+{
+	const enum intel_dpll_id id = pll->info->id;
+	u32 val;
+
+	val = intel_de_read(dev_priv, WRPLL_CTL(id));
+	intel_de_write(dev_priv, WRPLL_CTL(id), val & ~WRPLL_PLL_ENABLE);
+	intel_de_posting_read(dev_priv, WRPLL_CTL(id));
+
+	/*
+	 * Try to set up the PCH reference clock once all DPLLs
+	 * that depend on it have been shut down.
+	 */
+	if (dev_priv->pch_ssc_use & BIT(id))
+		intel_init_pch_refclk(dev_priv);
+}
+
+static void hsw_ddi_spll_disable(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll)
+{
+	enum intel_dpll_id id = pll->info->id;
+	u32 val;
+
+	val = intel_de_read(dev_priv, SPLL_CTL);
+	intel_de_write(dev_priv, SPLL_CTL, val & ~SPLL_PLL_ENABLE);
+	intel_de_posting_read(dev_priv, SPLL_CTL);
+
+	/*
+	 * Try to set up the PCH reference clock once all DPLLs
+	 * that depend on it have been shut down.
+	 */
+	if (dev_priv->pch_ssc_use & BIT(id))
+		intel_init_pch_refclk(dev_priv);
+}
+
+static bool hsw_ddi_wrpll_get_hw_state(struct drm_i915_private *dev_priv,
+				       struct intel_shared_dpll *pll,
+				       struct intel_dpll_hw_state *hw_state)
+{
+	const enum intel_dpll_id id = pll->info->id;
+	intel_wakeref_t wakeref;
+	u32 val;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     POWER_DOMAIN_DISPLAY_CORE);
+	if (!wakeref)
+		return false;
+
+	val = intel_de_read(dev_priv, WRPLL_CTL(id));
+	hw_state->wrpll = val;
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+
+	return val & WRPLL_PLL_ENABLE;
+}
+
+static bool hsw_ddi_spll_get_hw_state(struct drm_i915_private *dev_priv,
+				      struct intel_shared_dpll *pll,
+				      struct intel_dpll_hw_state *hw_state)
+{
+	intel_wakeref_t wakeref;
+	u32 val;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     POWER_DOMAIN_DISPLAY_CORE);
+	if (!wakeref)
+		return false;
+
+	val = intel_de_read(dev_priv, SPLL_CTL);
+	hw_state->spll = val;
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+
+	return val & SPLL_PLL_ENABLE;
+}
+
+#define LC_FREQ 2700
+#define LC_FREQ_2K U64_C(LC_FREQ * 2000)
+
+#define P_MIN 2
+#define P_MAX 64
+#define P_INC 2
+
+/* Constraints for PLL good behavior */
+#define REF_MIN 48
+#define REF_MAX 400
+#define VCO_MIN 2400
+#define VCO_MAX 4800
+
+struct hsw_wrpll_rnp {
+	unsigned p, n2, r2;
+};
+
+static unsigned hsw_wrpll_get_budget_for_freq(int clock)
+{
+	unsigned budget;
+
+	switch (clock) {
+	case 25175000:
+	case 25200000:
+	case 27000000:
+	case 27027000:
+	case 37762500:
+	case 37800000:
+	case 40500000:
+	case 40541000:
+	case 54000000:
+	case 54054000:
+	case 59341000:
+	case 59400000:
+	case 72000000:
+	case 74176000:
+	case 74250000:
+	case 81000000:
+	case 81081000:
+	case 89012000:
+	case 89100000:
+	case 108000000:
+	case 108108000:
+	case 111264000:
+	case 111375000:
+	case 148352000:
+	case 148500000:
+	case 162000000:
+	case 162162000:
+	case 222525000:
+	case 222750000:
+	case 296703000:
+	case 297000000:
+		budget = 0;
+		break;
+	case 233500000:
+	case 245250000:
+	case 247750000:
+	case 253250000:
+	case 298000000:
+		budget = 1500;
+		break;
+	case 169128000:
+	case 169500000:
+	case 179500000:
+	case 202000000:
+		budget = 2000;
+		break;
+	case 256250000:
+	case 262500000:
+	case 270000000:
+	case 272500000:
+	case 273750000:
+	case 280750000:
+	case 281250000:
+	case 286000000:
+	case 291750000:
+		budget = 4000;
+		break;
+	case 267250000:
+	case 268500000:
+		budget = 5000;
+		break;
+	default:
+		budget = 1000;
+		break;
+	}
+
+	return budget;
+}
+
+static void hsw_wrpll_update_rnp(u64 freq2k, unsigned int budget,
+				 unsigned int r2, unsigned int n2,
+				 unsigned int p,
+				 struct hsw_wrpll_rnp *best)
+{
+	u64 a, b, c, d, diff, diff_best;
+
+	/* No best (r,n,p) yet */
+	if (best->p == 0) {
+		best->p = p;
+		best->n2 = n2;
+		best->r2 = r2;
+		return;
+	}
+
+	/*
+	 * Output clock is (LC_FREQ_2K / 2000) * N / (P * R), which compares to
+	 * freq2k.
+	 *
+	 * delta = 1e6 *
+	 *	   abs(freq2k - (LC_FREQ_2K * n2/(p * r2))) /
+	 *	   freq2k;
+	 *
+	 * and we would like delta <= budget.
+	 *
+	 * If the discrepancy is above the PPM-based budget, always prefer to
+	 * improve upon the previous solution.  However, if you're within the
+	 * budget, try to maximize Ref * VCO, that is N / (P * R^2).
+	 */
+	a = freq2k * budget * p * r2;
+	b = freq2k * budget * best->p * best->r2;
+	diff = abs_diff(freq2k * p * r2, LC_FREQ_2K * n2);
+	diff_best = abs_diff(freq2k * best->p * best->r2,
+			     LC_FREQ_2K * best->n2);
+	c = 1000000 * diff;
+	d = 1000000 * diff_best;
+
+	if (a < c && b < d) {
+		/* If both are above the budget, pick the closer */
+		if (best->p * best->r2 * diff < p * r2 * diff_best) {
+			best->p = p;
+			best->n2 = n2;
+			best->r2 = r2;
+		}
+	} else if (a >= c && b < d) {
+		/* If A is below the threshold but B is above it?  Update. */
+		best->p = p;
+		best->n2 = n2;
+		best->r2 = r2;
+	} else if (a >= c && b >= d) {
+		/* Both are below the limit, so pick the higher n2/(r2*r2) */
+		if (n2 * best->r2 * best->r2 > best->n2 * r2 * r2) {
+			best->p = p;
+			best->n2 = n2;
+			best->r2 = r2;
+		}
+	}
+	/* Otherwise a < c && b >= d, do nothing */
+}
+
+static void
+hsw_ddi_calculate_wrpll(int clock /* in Hz */,
+			unsigned *r2_out, unsigned *n2_out, unsigned *p_out)
+{
+	u64 freq2k;
+	unsigned p, n2, r2;
+	struct hsw_wrpll_rnp best = {};
+	unsigned budget;
+
+	freq2k = clock / 100;
+
+	budget = hsw_wrpll_get_budget_for_freq(clock);
+
+	/* Special case handling for 540 pixel clock: bypass WR PLL entirely
+	 * and directly pass the LC PLL to it. */
+	if (freq2k == 5400000) {
+		*n2_out = 2;
+		*p_out = 1;
+		*r2_out = 2;
+		return;
+	}
+
+	/*
+	 * Ref = LC_FREQ / R, where Ref is the actual reference input seen by
+	 * the WR PLL.
+	 *
+	 * We want R so that REF_MIN <= Ref <= REF_MAX.
+	 * Injecting R2 = 2 * R gives:
+	 *   REF_MAX * r2 > LC_FREQ * 2 and
+	 *   REF_MIN * r2 < LC_FREQ * 2
+	 *
+	 * Which means the desired boundaries for r2 are:
+	 *  LC_FREQ * 2 / REF_MAX < r2 < LC_FREQ * 2 / REF_MIN
+	 *
+	 */
+	for (r2 = LC_FREQ * 2 / REF_MAX + 1;
+	     r2 <= LC_FREQ * 2 / REF_MIN;
+	     r2++) {
+
+		/*
+		 * VCO = N * Ref, that is: VCO = N * LC_FREQ / R
+		 *
+		 * Once again we want VCO_MIN <= VCO <= VCO_MAX.
+		 * Injecting R2 = 2 * R and N2 = 2 * N, we get:
+		 *   VCO_MAX * r2 > n2 * LC_FREQ and
+		 *   VCO_MIN * r2 < n2 * LC_FREQ)
+		 *
+		 * Which means the desired boundaries for n2 are:
+		 * VCO_MIN * r2 / LC_FREQ < n2 < VCO_MAX * r2 / LC_FREQ
+		 */
+		for (n2 = VCO_MIN * r2 / LC_FREQ + 1;
+		     n2 <= VCO_MAX * r2 / LC_FREQ;
+		     n2++) {
+
+			for (p = P_MIN; p <= P_MAX; p += P_INC)
+				hsw_wrpll_update_rnp(freq2k, budget,
+						     r2, n2, p, &best);
+		}
+	}
+
+	*n2_out = best.n2;
+	*p_out = best.p;
+	*r2_out = best.r2;
+}
+
+static int hsw_ddi_wrpll_get_freq(struct drm_i915_private *dev_priv,
+				  const struct intel_shared_dpll *pll,
+				  const struct intel_dpll_hw_state *pll_state)
+{
+	int refclk;
+	int n, p, r;
+	u32 wrpll = pll_state->wrpll;
+
+	switch (wrpll & WRPLL_REF_MASK) {
+	case WRPLL_REF_SPECIAL_HSW:
+		/* Muxed-SSC for BDW, non-SSC for non-ULT HSW. */
+		if (IS_HASWELL(dev_priv) && !IS_HSW_ULT(dev_priv)) {
+			refclk = dev_priv->display.dpll.ref_clks.nssc;
+			break;
+		}
+		fallthrough;
+	case WRPLL_REF_PCH_SSC:
+		/*
+		 * We could calculate spread here, but our checking
+		 * code only cares about 5% accuracy, and spread is a max of
+		 * 0.5% downspread.
+		 */
+		refclk = dev_priv->display.dpll.ref_clks.ssc;
+		break;
+	case WRPLL_REF_LCPLL:
+		refclk = 2700000;
+		break;
+	default:
+		MISSING_CASE(wrpll);
+		return 0;
+	}
+
+	r = wrpll & WRPLL_DIVIDER_REF_MASK;
+	p = (wrpll & WRPLL_DIVIDER_POST_MASK) >> WRPLL_DIVIDER_POST_SHIFT;
+	n = (wrpll & WRPLL_DIVIDER_FB_MASK) >> WRPLL_DIVIDER_FB_SHIFT;
+
+	/* Convert to KHz, p & r have a fixed point portion */
+	return (refclk * n / 10) / (p * r) * 2;
+}
+
+static int
+hsw_ddi_wrpll_compute_dpll(struct intel_atomic_state *state,
+			   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	unsigned int p, n2, r2;
+
+	hsw_ddi_calculate_wrpll(crtc_state->port_clock * 1000, &r2, &n2, &p);
+
+	crtc_state->dpll_hw_state.wrpll =
+		WRPLL_PLL_ENABLE | WRPLL_REF_LCPLL |
+		WRPLL_DIVIDER_REFERENCE(r2) | WRPLL_DIVIDER_FEEDBACK(n2) |
+		WRPLL_DIVIDER_POST(p);
+
+	crtc_state->port_clock = hsw_ddi_wrpll_get_freq(i915, NULL,
+							&crtc_state->dpll_hw_state);
+
+	return 0;
+}
+
+static struct intel_shared_dpll *
+hsw_ddi_wrpll_get_dpll(struct intel_atomic_state *state,
+		       struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	return intel_find_shared_dpll(state, crtc,
+				      &crtc_state->dpll_hw_state,
+				      BIT(DPLL_ID_WRPLL2) |
+				      BIT(DPLL_ID_WRPLL1));
+}
+
+static int
+hsw_ddi_lcpll_compute_dpll(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	int clock = crtc_state->port_clock;
+
+	switch (clock / 2) {
+	case 81000:
+	case 135000:
+	case 270000:
+		return 0;
+	default:
+		drm_dbg_kms(&dev_priv->drm, "Invalid clock for DP: %d\n",
+			    clock);
+		return -EINVAL;
+	}
+}
+
+static struct intel_shared_dpll *
+hsw_ddi_lcpll_get_dpll(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	struct intel_shared_dpll *pll;
+	enum intel_dpll_id pll_id;
+	int clock = crtc_state->port_clock;
+
+	switch (clock / 2) {
+	case 81000:
+		pll_id = DPLL_ID_LCPLL_810;
+		break;
+	case 135000:
+		pll_id = DPLL_ID_LCPLL_1350;
+		break;
+	case 270000:
+		pll_id = DPLL_ID_LCPLL_2700;
+		break;
+	default:
+		MISSING_CASE(clock / 2);
+		return NULL;
+	}
+
+	pll = intel_get_shared_dpll_by_id(dev_priv, pll_id);
+
+	if (!pll)
+		return NULL;
+
+	return pll;
+}
+
+static int hsw_ddi_lcpll_get_freq(struct drm_i915_private *i915,
+				  const struct intel_shared_dpll *pll,
+				  const struct intel_dpll_hw_state *pll_state)
+{
+	int link_clock = 0;
+
+	switch (pll->info->id) {
+	case DPLL_ID_LCPLL_810:
+		link_clock = 81000;
+		break;
+	case DPLL_ID_LCPLL_1350:
+		link_clock = 135000;
+		break;
+	case DPLL_ID_LCPLL_2700:
+		link_clock = 270000;
+		break;
+	default:
+		drm_WARN(&i915->drm, 1, "bad port clock sel\n");
+		break;
+	}
+
+	return link_clock * 2;
+}
+
+static int
+hsw_ddi_spll_compute_dpll(struct intel_atomic_state *state,
+			  struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (drm_WARN_ON(crtc->base.dev, crtc_state->port_clock / 2 != 135000))
+		return -EINVAL;
+
+	crtc_state->dpll_hw_state.spll =
+		SPLL_PLL_ENABLE | SPLL_FREQ_1350MHz | SPLL_REF_MUXED_SSC;
+
+	return 0;
+}
+
+static struct intel_shared_dpll *
+hsw_ddi_spll_get_dpll(struct intel_atomic_state *state,
+		      struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	return intel_find_shared_dpll(state, crtc, &crtc_state->dpll_hw_state,
+				      BIT(DPLL_ID_SPLL));
+}
+
+static int hsw_ddi_spll_get_freq(struct drm_i915_private *i915,
+				 const struct intel_shared_dpll *pll,
+				 const struct intel_dpll_hw_state *pll_state)
+{
+	int link_clock = 0;
+
+	switch (pll_state->spll & SPLL_FREQ_MASK) {
+	case SPLL_FREQ_810MHz:
+		link_clock = 81000;
+		break;
+	case SPLL_FREQ_1350MHz:
+		link_clock = 135000;
+		break;
+	case SPLL_FREQ_2700MHz:
+		link_clock = 270000;
+		break;
+	default:
+		drm_WARN(&i915->drm, 1, "bad spll freq\n");
+		break;
+	}
+
+	return link_clock * 2;
+}
+
+static int hsw_compute_dpll(struct intel_atomic_state *state,
+			    struct intel_crtc *crtc,
+			    struct intel_encoder *encoder)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return hsw_ddi_wrpll_compute_dpll(state, crtc);
+	else if (intel_crtc_has_dp_encoder(crtc_state))
+		return hsw_ddi_lcpll_compute_dpll(crtc_state);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG))
+		return hsw_ddi_spll_compute_dpll(state, crtc);
+	else
+		return -EINVAL;
+}
+
+static int hsw_get_dpll(struct intel_atomic_state *state,
+			struct intel_crtc *crtc,
+			struct intel_encoder *encoder)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_shared_dpll *pll = NULL;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		pll = hsw_ddi_wrpll_get_dpll(state, crtc);
+	else if (intel_crtc_has_dp_encoder(crtc_state))
+		pll = hsw_ddi_lcpll_get_dpll(crtc_state);
+	else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_ANALOG))
+		pll = hsw_ddi_spll_get_dpll(state, crtc);
+
+	if (!pll)
+		return -EINVAL;
+
+	intel_reference_shared_dpll(state, crtc,
+				    pll, &crtc_state->dpll_hw_state);
+
+	crtc_state->shared_dpll = pll;
+
+	return 0;
+}
+
+static void hsw_update_dpll_ref_clks(struct drm_i915_private *i915)
+{
+	i915->display.dpll.ref_clks.ssc = 135000;
+	/* Non-SSC is only used on non-ULT HSW. */
+	if (intel_de_read(i915, FUSE_STRAP3) & HSW_REF_CLK_SELECT)
+		i915->display.dpll.ref_clks.nssc = 24000;
+	else
+		i915->display.dpll.ref_clks.nssc = 135000;
+}
+
+static void hsw_dump_hw_state(struct drm_i915_private *dev_priv,
+			      const struct intel_dpll_hw_state *hw_state)
+{
+	drm_dbg_kms(&dev_priv->drm, "dpll_hw_state: wrpll: 0x%x spll: 0x%x\n",
+		    hw_state->wrpll, hw_state->spll);
+}
+
+static const struct intel_shared_dpll_funcs hsw_ddi_wrpll_funcs = {
+	.enable = hsw_ddi_wrpll_enable,
+	.disable = hsw_ddi_wrpll_disable,
+	.get_hw_state = hsw_ddi_wrpll_get_hw_state,
+	.get_freq = hsw_ddi_wrpll_get_freq,
+};
+
+static const struct intel_shared_dpll_funcs hsw_ddi_spll_funcs = {
+	.enable = hsw_ddi_spll_enable,
+	.disable = hsw_ddi_spll_disable,
+	.get_hw_state = hsw_ddi_spll_get_hw_state,
+	.get_freq = hsw_ddi_spll_get_freq,
+};
+
+static void hsw_ddi_lcpll_enable(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll)
+{
+}
+
+static void hsw_ddi_lcpll_disable(struct drm_i915_private *dev_priv,
+				  struct intel_shared_dpll *pll)
+{
+}
+
+static bool hsw_ddi_lcpll_get_hw_state(struct drm_i915_private *dev_priv,
+				       struct intel_shared_dpll *pll,
+				       struct intel_dpll_hw_state *hw_state)
+{
+	return true;
+}
+
+static const struct intel_shared_dpll_funcs hsw_ddi_lcpll_funcs = {
+	.enable = hsw_ddi_lcpll_enable,
+	.disable = hsw_ddi_lcpll_disable,
+	.get_hw_state = hsw_ddi_lcpll_get_hw_state,
+	.get_freq = hsw_ddi_lcpll_get_freq,
+};
+
+static const struct dpll_info hsw_plls[] = {
+	{ "WRPLL 1",    &hsw_ddi_wrpll_funcs, DPLL_ID_WRPLL1,     0 },
+	{ "WRPLL 2",    &hsw_ddi_wrpll_funcs, DPLL_ID_WRPLL2,     0 },
+	{ "SPLL",       &hsw_ddi_spll_funcs,  DPLL_ID_SPLL,       0 },
+	{ "LCPLL 810",  &hsw_ddi_lcpll_funcs, DPLL_ID_LCPLL_810,  INTEL_DPLL_ALWAYS_ON },
+	{ "LCPLL 1350", &hsw_ddi_lcpll_funcs, DPLL_ID_LCPLL_1350, INTEL_DPLL_ALWAYS_ON },
+	{ "LCPLL 2700", &hsw_ddi_lcpll_funcs, DPLL_ID_LCPLL_2700, INTEL_DPLL_ALWAYS_ON },
+	{ },
+};
+
+static const struct intel_dpll_mgr hsw_pll_mgr = {
+	.dpll_info = hsw_plls,
+	.compute_dplls = hsw_compute_dpll,
+	.get_dplls = hsw_get_dpll,
+	.put_dplls = intel_put_dpll,
+	.update_ref_clks = hsw_update_dpll_ref_clks,
+	.dump_hw_state = hsw_dump_hw_state,
+};
+
+struct skl_dpll_regs {
+	i915_reg_t ctl, cfgcr1, cfgcr2;
+};
+
+/* this array is indexed by the *shared* pll id */
+static const struct skl_dpll_regs skl_dpll_regs[4] = {
+	{
+		/* DPLL 0 */
+		.ctl = LCPLL1_CTL,
+		/* DPLL 0 doesn't support HDMI mode */
+	},
+	{
+		/* DPLL 1 */
+		.ctl = LCPLL2_CTL,
+		.cfgcr1 = DPLL_CFGCR1(SKL_DPLL1),
+		.cfgcr2 = DPLL_CFGCR2(SKL_DPLL1),
+	},
+	{
+		/* DPLL 2 */
+		.ctl = WRPLL_CTL(0),
+		.cfgcr1 = DPLL_CFGCR1(SKL_DPLL2),
+		.cfgcr2 = DPLL_CFGCR2(SKL_DPLL2),
+	},
+	{
+		/* DPLL 3 */
+		.ctl = WRPLL_CTL(1),
+		.cfgcr1 = DPLL_CFGCR1(SKL_DPLL3),
+		.cfgcr2 = DPLL_CFGCR2(SKL_DPLL3),
+	},
+};
+
+static void skl_ddi_pll_write_ctrl1(struct drm_i915_private *dev_priv,
+				    struct intel_shared_dpll *pll)
+{
+	const enum intel_dpll_id id = pll->info->id;
+	u32 val;
+
+	val = intel_de_read(dev_priv, DPLL_CTRL1);
+
+	val &= ~(DPLL_CTRL1_HDMI_MODE(id) |
+		 DPLL_CTRL1_SSC(id) |
+		 DPLL_CTRL1_LINK_RATE_MASK(id));
+	val |= pll->state.hw_state.ctrl1 << (id * 6);
+
+	intel_de_write(dev_priv, DPLL_CTRL1, val);
+	intel_de_posting_read(dev_priv, DPLL_CTRL1);
+}
+
+static void skl_ddi_pll_enable(struct drm_i915_private *dev_priv,
+			       struct intel_shared_dpll *pll)
+{
+	const struct skl_dpll_regs *regs = skl_dpll_regs;
+	const enum intel_dpll_id id = pll->info->id;
+
+	skl_ddi_pll_write_ctrl1(dev_priv, pll);
+
+	intel_de_write(dev_priv, regs[id].cfgcr1, pll->state.hw_state.cfgcr1);
+	intel_de_write(dev_priv, regs[id].cfgcr2, pll->state.hw_state.cfgcr2);
+	intel_de_posting_read(dev_priv, regs[id].cfgcr1);
+	intel_de_posting_read(dev_priv, regs[id].cfgcr2);
+
+	/* the enable bit is always bit 31 */
+	intel_de_write(dev_priv, regs[id].ctl,
+		       intel_de_read(dev_priv, regs[id].ctl) | LCPLL_PLL_ENABLE);
+
+	if (intel_de_wait_for_set(dev_priv, DPLL_STATUS, DPLL_LOCK(id), 5))
+		drm_err(&dev_priv->drm, "DPLL %d not locked\n", id);
+}
+
+static void skl_ddi_dpll0_enable(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll)
+{
+	skl_ddi_pll_write_ctrl1(dev_priv, pll);
+}
+
+static void skl_ddi_pll_disable(struct drm_i915_private *dev_priv,
+				struct intel_shared_dpll *pll)
+{
+	const struct skl_dpll_regs *regs = skl_dpll_regs;
+	const enum intel_dpll_id id = pll->info->id;
+
+	/* the enable bit is always bit 31 */
+	intel_de_write(dev_priv, regs[id].ctl,
+		       intel_de_read(dev_priv, regs[id].ctl) & ~LCPLL_PLL_ENABLE);
+	intel_de_posting_read(dev_priv, regs[id].ctl);
+}
+
+static void skl_ddi_dpll0_disable(struct drm_i915_private *dev_priv,
+				  struct intel_shared_dpll *pll)
+{
+}
+
+static bool skl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv,
+				     struct intel_shared_dpll *pll,
+				     struct intel_dpll_hw_state *hw_state)
+{
+	u32 val;
+	const struct skl_dpll_regs *regs = skl_dpll_regs;
+	const enum intel_dpll_id id = pll->info->id;
+	intel_wakeref_t wakeref;
+	bool ret;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     POWER_DOMAIN_DISPLAY_CORE);
+	if (!wakeref)
+		return false;
+
+	ret = false;
+
+	val = intel_de_read(dev_priv, regs[id].ctl);
+	if (!(val & LCPLL_PLL_ENABLE))
+		goto out;
+
+	val = intel_de_read(dev_priv, DPLL_CTRL1);
+	hw_state->ctrl1 = (val >> (id * 6)) & 0x3f;
+
+	/* avoid reading back stale values if HDMI mode is not enabled */
+	if (val & DPLL_CTRL1_HDMI_MODE(id)) {
+		hw_state->cfgcr1 = intel_de_read(dev_priv, regs[id].cfgcr1);
+		hw_state->cfgcr2 = intel_de_read(dev_priv, regs[id].cfgcr2);
+	}
+	ret = true;
+
+out:
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+
+	return ret;
+}
+
+static bool skl_ddi_dpll0_get_hw_state(struct drm_i915_private *dev_priv,
+				       struct intel_shared_dpll *pll,
+				       struct intel_dpll_hw_state *hw_state)
+{
+	const struct skl_dpll_regs *regs = skl_dpll_regs;
+	const enum intel_dpll_id id = pll->info->id;
+	intel_wakeref_t wakeref;
+	u32 val;
+	bool ret;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     POWER_DOMAIN_DISPLAY_CORE);
+	if (!wakeref)
+		return false;
+
+	ret = false;
+
+	/* DPLL0 is always enabled since it drives CDCLK */
+	val = intel_de_read(dev_priv, regs[id].ctl);
+	if (drm_WARN_ON(&dev_priv->drm, !(val & LCPLL_PLL_ENABLE)))
+		goto out;
+
+	val = intel_de_read(dev_priv, DPLL_CTRL1);
+	hw_state->ctrl1 = (val >> (id * 6)) & 0x3f;
+
+	ret = true;
+
+out:
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+
+	return ret;
+}
+
+struct skl_wrpll_context {
+	u64 min_deviation;		/* current minimal deviation */
+	u64 central_freq;		/* chosen central freq */
+	u64 dco_freq;			/* chosen dco freq */
+	unsigned int p;			/* chosen divider */
+};
+
+/* DCO freq must be within +1%/-6%  of the DCO central freq */
+#define SKL_DCO_MAX_PDEVIATION	100
+#define SKL_DCO_MAX_NDEVIATION	600
+
+static void skl_wrpll_try_divider(struct skl_wrpll_context *ctx,
+				  u64 central_freq,
+				  u64 dco_freq,
+				  unsigned int divider)
+{
+	u64 deviation;
+
+	deviation = div64_u64(10000 * abs_diff(dco_freq, central_freq),
+			      central_freq);
+
+	/* positive deviation */
+	if (dco_freq >= central_freq) {
+		if (deviation < SKL_DCO_MAX_PDEVIATION &&
+		    deviation < ctx->min_deviation) {
+			ctx->min_deviation = deviation;
+			ctx->central_freq = central_freq;
+			ctx->dco_freq = dco_freq;
+			ctx->p = divider;
+		}
+	/* negative deviation */
+	} else if (deviation < SKL_DCO_MAX_NDEVIATION &&
+		   deviation < ctx->min_deviation) {
+		ctx->min_deviation = deviation;
+		ctx->central_freq = central_freq;
+		ctx->dco_freq = dco_freq;
+		ctx->p = divider;
+	}
+}
+
+static void skl_wrpll_get_multipliers(unsigned int p,
+				      unsigned int *p0 /* out */,
+				      unsigned int *p1 /* out */,
+				      unsigned int *p2 /* out */)
+{
+	/* even dividers */
+	if (p % 2 == 0) {
+		unsigned int half = p / 2;
+
+		if (half == 1 || half == 2 || half == 3 || half == 5) {
+			*p0 = 2;
+			*p1 = 1;
+			*p2 = half;
+		} else if (half % 2 == 0) {
+			*p0 = 2;
+			*p1 = half / 2;
+			*p2 = 2;
+		} else if (half % 3 == 0) {
+			*p0 = 3;
+			*p1 = half / 3;
+			*p2 = 2;
+		} else if (half % 7 == 0) {
+			*p0 = 7;
+			*p1 = half / 7;
+			*p2 = 2;
+		}
+	} else if (p == 3 || p == 9) {  /* 3, 5, 7, 9, 15, 21, 35 */
+		*p0 = 3;
+		*p1 = 1;
+		*p2 = p / 3;
+	} else if (p == 5 || p == 7) {
+		*p0 = p;
+		*p1 = 1;
+		*p2 = 1;
+	} else if (p == 15) {
+		*p0 = 3;
+		*p1 = 1;
+		*p2 = 5;
+	} else if (p == 21) {
+		*p0 = 7;
+		*p1 = 1;
+		*p2 = 3;
+	} else if (p == 35) {
+		*p0 = 7;
+		*p1 = 1;
+		*p2 = 5;
+	}
+}
+
+struct skl_wrpll_params {
+	u32 dco_fraction;
+	u32 dco_integer;
+	u32 qdiv_ratio;
+	u32 qdiv_mode;
+	u32 kdiv;
+	u32 pdiv;
+	u32 central_freq;
+};
+
+static void skl_wrpll_params_populate(struct skl_wrpll_params *params,
+				      u64 afe_clock,
+				      int ref_clock,
+				      u64 central_freq,
+				      u32 p0, u32 p1, u32 p2)
+{
+	u64 dco_freq;
+
+	switch (central_freq) {
+	case 9600000000ULL:
+		params->central_freq = 0;
+		break;
+	case 9000000000ULL:
+		params->central_freq = 1;
+		break;
+	case 8400000000ULL:
+		params->central_freq = 3;
+	}
+
+	switch (p0) {
+	case 1:
+		params->pdiv = 0;
+		break;
+	case 2:
+		params->pdiv = 1;
+		break;
+	case 3:
+		params->pdiv = 2;
+		break;
+	case 7:
+		params->pdiv = 4;
+		break;
+	default:
+		WARN(1, "Incorrect PDiv\n");
+	}
+
+	switch (p2) {
+	case 5:
+		params->kdiv = 0;
+		break;
+	case 2:
+		params->kdiv = 1;
+		break;
+	case 3:
+		params->kdiv = 2;
+		break;
+	case 1:
+		params->kdiv = 3;
+		break;
+	default:
+		WARN(1, "Incorrect KDiv\n");
+	}
+
+	params->qdiv_ratio = p1;
+	params->qdiv_mode = (params->qdiv_ratio == 1) ? 0 : 1;
+
+	dco_freq = p0 * p1 * p2 * afe_clock;
+
+	/*
+	 * Intermediate values are in Hz.
+	 * Divide by MHz to match bsepc
+	 */
+	params->dco_integer = div_u64(dco_freq, ref_clock * KHz(1));
+	params->dco_fraction =
+		div_u64((div_u64(dco_freq, ref_clock / KHz(1)) -
+			 params->dco_integer * MHz(1)) * 0x8000, MHz(1));
+}
+
+static int
+skl_ddi_calculate_wrpll(int clock /* in Hz */,
+			int ref_clock,
+			struct skl_wrpll_params *wrpll_params)
+{
+	static const u64 dco_central_freq[3] = { 8400000000ULL,
+						 9000000000ULL,
+						 9600000000ULL };
+	static const u8 even_dividers[] = {  4,  6,  8, 10, 12, 14, 16, 18, 20,
+					    24, 28, 30, 32, 36, 40, 42, 44,
+					    48, 52, 54, 56, 60, 64, 66, 68,
+					    70, 72, 76, 78, 80, 84, 88, 90,
+					    92, 96, 98 };
+	static const u8 odd_dividers[] = { 3, 5, 7, 9, 15, 21, 35 };
+	static const struct {
+		const u8 *list;
+		int n_dividers;
+	} dividers[] = {
+		{ even_dividers, ARRAY_SIZE(even_dividers) },
+		{ odd_dividers, ARRAY_SIZE(odd_dividers) },
+	};
+	struct skl_wrpll_context ctx = {
+		.min_deviation = U64_MAX,
+	};
+	unsigned int dco, d, i;
+	unsigned int p0, p1, p2;
+	u64 afe_clock = clock * 5; /* AFE Clock is 5x Pixel clock */
+
+	for (d = 0; d < ARRAY_SIZE(dividers); d++) {
+		for (dco = 0; dco < ARRAY_SIZE(dco_central_freq); dco++) {
+			for (i = 0; i < dividers[d].n_dividers; i++) {
+				unsigned int p = dividers[d].list[i];
+				u64 dco_freq = p * afe_clock;
+
+				skl_wrpll_try_divider(&ctx,
+						      dco_central_freq[dco],
+						      dco_freq,
+						      p);
+				/*
+				 * Skip the remaining dividers if we're sure to
+				 * have found the definitive divider, we can't
+				 * improve a 0 deviation.
+				 */
+				if (ctx.min_deviation == 0)
+					goto skip_remaining_dividers;
+			}
+		}
+
+skip_remaining_dividers:
+		/*
+		 * If a solution is found with an even divider, prefer
+		 * this one.
+		 */
+		if (d == 0 && ctx.p)
+			break;
+	}
+
+	if (!ctx.p)
+		return -EINVAL;
+
+	/*
+	 * gcc incorrectly analyses that these can be used without being
+	 * initialized. To be fair, it's hard to guess.
+	 */
+	p0 = p1 = p2 = 0;
+	skl_wrpll_get_multipliers(ctx.p, &p0, &p1, &p2);
+	skl_wrpll_params_populate(wrpll_params, afe_clock, ref_clock,
+				  ctx.central_freq, p0, p1, p2);
+
+	return 0;
+}
+
+static int skl_ddi_wrpll_get_freq(struct drm_i915_private *i915,
+				  const struct intel_shared_dpll *pll,
+				  const struct intel_dpll_hw_state *pll_state)
+{
+	int ref_clock = i915->display.dpll.ref_clks.nssc;
+	u32 p0, p1, p2, dco_freq;
+
+	p0 = pll_state->cfgcr2 & DPLL_CFGCR2_PDIV_MASK;
+	p2 = pll_state->cfgcr2 & DPLL_CFGCR2_KDIV_MASK;
+
+	if (pll_state->cfgcr2 &  DPLL_CFGCR2_QDIV_MODE(1))
+		p1 = (pll_state->cfgcr2 & DPLL_CFGCR2_QDIV_RATIO_MASK) >> 8;
+	else
+		p1 = 1;
+
+
+	switch (p0) {
+	case DPLL_CFGCR2_PDIV_1:
+		p0 = 1;
+		break;
+	case DPLL_CFGCR2_PDIV_2:
+		p0 = 2;
+		break;
+	case DPLL_CFGCR2_PDIV_3:
+		p0 = 3;
+		break;
+	case DPLL_CFGCR2_PDIV_7_INVALID:
+		/*
+		 * Incorrect ASUS-Z170M BIOS setting, the HW seems to ignore bit#0,
+		 * handling it the same way as PDIV_7.
+		 */
+		drm_dbg_kms(&i915->drm, "Invalid WRPLL PDIV divider value, fixing it.\n");
+		fallthrough;
+	case DPLL_CFGCR2_PDIV_7:
+		p0 = 7;
+		break;
+	default:
+		MISSING_CASE(p0);
+		return 0;
+	}
+
+	switch (p2) {
+	case DPLL_CFGCR2_KDIV_5:
+		p2 = 5;
+		break;
+	case DPLL_CFGCR2_KDIV_2:
+		p2 = 2;
+		break;
+	case DPLL_CFGCR2_KDIV_3:
+		p2 = 3;
+		break;
+	case DPLL_CFGCR2_KDIV_1:
+		p2 = 1;
+		break;
+	default:
+		MISSING_CASE(p2);
+		return 0;
+	}
+
+	dco_freq = (pll_state->cfgcr1 & DPLL_CFGCR1_DCO_INTEGER_MASK) *
+		   ref_clock;
+
+	dco_freq += ((pll_state->cfgcr1 & DPLL_CFGCR1_DCO_FRACTION_MASK) >> 9) *
+		    ref_clock / 0x8000;
+
+	if (drm_WARN_ON(&i915->drm, p0 == 0 || p1 == 0 || p2 == 0))
+		return 0;
+
+	return dco_freq / (p0 * p1 * p2 * 5);
+}
+
+static int skl_ddi_hdmi_pll_dividers(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	struct skl_wrpll_params wrpll_params = {};
+	u32 ctrl1, cfgcr1, cfgcr2;
+	int ret;
+
+	/*
+	 * See comment in intel_dpll_hw_state to understand why we always use 0
+	 * as the DPLL id in this function.
+	 */
+	ctrl1 = DPLL_CTRL1_OVERRIDE(0);
+
+	ctrl1 |= DPLL_CTRL1_HDMI_MODE(0);
+
+	ret = skl_ddi_calculate_wrpll(crtc_state->port_clock * 1000,
+				      i915->display.dpll.ref_clks.nssc, &wrpll_params);
+	if (ret)
+		return ret;
+
+	cfgcr1 = DPLL_CFGCR1_FREQ_ENABLE |
+		DPLL_CFGCR1_DCO_FRACTION(wrpll_params.dco_fraction) |
+		wrpll_params.dco_integer;
+
+	cfgcr2 = DPLL_CFGCR2_QDIV_RATIO(wrpll_params.qdiv_ratio) |
+		DPLL_CFGCR2_QDIV_MODE(wrpll_params.qdiv_mode) |
+		DPLL_CFGCR2_KDIV(wrpll_params.kdiv) |
+		DPLL_CFGCR2_PDIV(wrpll_params.pdiv) |
+		wrpll_params.central_freq;
+
+	crtc_state->dpll_hw_state.ctrl1 = ctrl1;
+	crtc_state->dpll_hw_state.cfgcr1 = cfgcr1;
+	crtc_state->dpll_hw_state.cfgcr2 = cfgcr2;
+
+	crtc_state->port_clock = skl_ddi_wrpll_get_freq(i915, NULL,
+							&crtc_state->dpll_hw_state);
+
+	return 0;
+}
+
+static int
+skl_ddi_dp_set_dpll_hw_state(struct intel_crtc_state *crtc_state)
+{
+	u32 ctrl1;
+
+	/*
+	 * See comment in intel_dpll_hw_state to understand why we always use 0
+	 * as the DPLL id in this function.
+	 */
+	ctrl1 = DPLL_CTRL1_OVERRIDE(0);
+	switch (crtc_state->port_clock / 2) {
+	case 81000:
+		ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810, 0);
+		break;
+	case 135000:
+		ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1350, 0);
+		break;
+	case 270000:
+		ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_2700, 0);
+		break;
+		/* eDP 1.4 rates */
+	case 162000:
+		ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1620, 0);
+		break;
+	case 108000:
+		ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1080, 0);
+		break;
+	case 216000:
+		ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_2160, 0);
+		break;
+	}
+
+	crtc_state->dpll_hw_state.ctrl1 = ctrl1;
+
+	return 0;
+}
+
+static int skl_ddi_lcpll_get_freq(struct drm_i915_private *i915,
+				  const struct intel_shared_dpll *pll,
+				  const struct intel_dpll_hw_state *pll_state)
+{
+	int link_clock = 0;
+
+	switch ((pll_state->ctrl1 & DPLL_CTRL1_LINK_RATE_MASK(0)) >>
+		DPLL_CTRL1_LINK_RATE_SHIFT(0)) {
+	case DPLL_CTRL1_LINK_RATE_810:
+		link_clock = 81000;
+		break;
+	case DPLL_CTRL1_LINK_RATE_1080:
+		link_clock = 108000;
+		break;
+	case DPLL_CTRL1_LINK_RATE_1350:
+		link_clock = 135000;
+		break;
+	case DPLL_CTRL1_LINK_RATE_1620:
+		link_clock = 162000;
+		break;
+	case DPLL_CTRL1_LINK_RATE_2160:
+		link_clock = 216000;
+		break;
+	case DPLL_CTRL1_LINK_RATE_2700:
+		link_clock = 270000;
+		break;
+	default:
+		drm_WARN(&i915->drm, 1, "Unsupported link rate\n");
+		break;
+	}
+
+	return link_clock * 2;
+}
+
+static int skl_compute_dpll(struct intel_atomic_state *state,
+			    struct intel_crtc *crtc,
+			    struct intel_encoder *encoder)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return skl_ddi_hdmi_pll_dividers(crtc_state);
+	else if (intel_crtc_has_dp_encoder(crtc_state))
+		return skl_ddi_dp_set_dpll_hw_state(crtc_state);
+	else
+		return -EINVAL;
+}
+
+static int skl_get_dpll(struct intel_atomic_state *state,
+			struct intel_crtc *crtc,
+			struct intel_encoder *encoder)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_shared_dpll *pll;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP))
+		pll = intel_find_shared_dpll(state, crtc,
+					     &crtc_state->dpll_hw_state,
+					     BIT(DPLL_ID_SKL_DPLL0));
+	else
+		pll = intel_find_shared_dpll(state, crtc,
+					     &crtc_state->dpll_hw_state,
+					     BIT(DPLL_ID_SKL_DPLL3) |
+					     BIT(DPLL_ID_SKL_DPLL2) |
+					     BIT(DPLL_ID_SKL_DPLL1));
+	if (!pll)
+		return -EINVAL;
+
+	intel_reference_shared_dpll(state, crtc,
+				    pll, &crtc_state->dpll_hw_state);
+
+	crtc_state->shared_dpll = pll;
+
+	return 0;
+}
+
+static int skl_ddi_pll_get_freq(struct drm_i915_private *i915,
+				const struct intel_shared_dpll *pll,
+				const struct intel_dpll_hw_state *pll_state)
+{
+	/*
+	 * ctrl1 register is already shifted for each pll, just use 0 to get
+	 * the internal shift for each field
+	 */
+	if (pll_state->ctrl1 & DPLL_CTRL1_HDMI_MODE(0))
+		return skl_ddi_wrpll_get_freq(i915, pll, pll_state);
+	else
+		return skl_ddi_lcpll_get_freq(i915, pll, pll_state);
+}
+
+static void skl_update_dpll_ref_clks(struct drm_i915_private *i915)
+{
+	/* No SSC ref */
+	i915->display.dpll.ref_clks.nssc = i915->display.cdclk.hw.ref;
+}
+
+static void skl_dump_hw_state(struct drm_i915_private *dev_priv,
+			      const struct intel_dpll_hw_state *hw_state)
+{
+	drm_dbg_kms(&dev_priv->drm, "dpll_hw_state: "
+		      "ctrl1: 0x%x, cfgcr1: 0x%x, cfgcr2: 0x%x\n",
+		      hw_state->ctrl1,
+		      hw_state->cfgcr1,
+		      hw_state->cfgcr2);
+}
+
+static const struct intel_shared_dpll_funcs skl_ddi_pll_funcs = {
+	.enable = skl_ddi_pll_enable,
+	.disable = skl_ddi_pll_disable,
+	.get_hw_state = skl_ddi_pll_get_hw_state,
+	.get_freq = skl_ddi_pll_get_freq,
+};
+
+static const struct intel_shared_dpll_funcs skl_ddi_dpll0_funcs = {
+	.enable = skl_ddi_dpll0_enable,
+	.disable = skl_ddi_dpll0_disable,
+	.get_hw_state = skl_ddi_dpll0_get_hw_state,
+	.get_freq = skl_ddi_pll_get_freq,
+};
+
+static const struct dpll_info skl_plls[] = {
+	{ "DPLL 0", &skl_ddi_dpll0_funcs, DPLL_ID_SKL_DPLL0, INTEL_DPLL_ALWAYS_ON },
+	{ "DPLL 1", &skl_ddi_pll_funcs,   DPLL_ID_SKL_DPLL1, 0 },
+	{ "DPLL 2", &skl_ddi_pll_funcs,   DPLL_ID_SKL_DPLL2, 0 },
+	{ "DPLL 3", &skl_ddi_pll_funcs,   DPLL_ID_SKL_DPLL3, 0 },
+	{ },
+};
+
+static const struct intel_dpll_mgr skl_pll_mgr = {
+	.dpll_info = skl_plls,
+	.compute_dplls = skl_compute_dpll,
+	.get_dplls = skl_get_dpll,
+	.put_dplls = intel_put_dpll,
+	.update_ref_clks = skl_update_dpll_ref_clks,
+	.dump_hw_state = skl_dump_hw_state,
+};
+
+static void bxt_ddi_pll_enable(struct drm_i915_private *dev_priv,
+				struct intel_shared_dpll *pll)
+{
+	u32 temp;
+	enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */
+	enum dpio_phy phy;
+	enum dpio_channel ch;
+
+	bxt_port_to_phy_channel(dev_priv, port, &phy, &ch);
+
+	/* Non-SSC reference */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL_ENABLE(port));
+	temp |= PORT_PLL_REF_SEL;
+	intel_de_write(dev_priv, BXT_PORT_PLL_ENABLE(port), temp);
+
+	if (IS_GEMINILAKE(dev_priv)) {
+		temp = intel_de_read(dev_priv, BXT_PORT_PLL_ENABLE(port));
+		temp |= PORT_PLL_POWER_ENABLE;
+		intel_de_write(dev_priv, BXT_PORT_PLL_ENABLE(port), temp);
+
+		if (wait_for_us((intel_de_read(dev_priv, BXT_PORT_PLL_ENABLE(port)) &
+				 PORT_PLL_POWER_STATE), 200))
+			drm_err(&dev_priv->drm,
+				"Power state not set for PLL:%d\n", port);
+	}
+
+	/* Disable 10 bit clock */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL_EBB_4(phy, ch));
+	temp &= ~PORT_PLL_10BIT_CLK_ENABLE;
+	intel_de_write(dev_priv, BXT_PORT_PLL_EBB_4(phy, ch), temp);
+
+	/* Write P1 & P2 */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL_EBB_0(phy, ch));
+	temp &= ~(PORT_PLL_P1_MASK | PORT_PLL_P2_MASK);
+	temp |= pll->state.hw_state.ebb0;
+	intel_de_write(dev_priv, BXT_PORT_PLL_EBB_0(phy, ch), temp);
+
+	/* Write M2 integer */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 0));
+	temp &= ~PORT_PLL_M2_INT_MASK;
+	temp |= pll->state.hw_state.pll0;
+	intel_de_write(dev_priv, BXT_PORT_PLL(phy, ch, 0), temp);
+
+	/* Write N */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 1));
+	temp &= ~PORT_PLL_N_MASK;
+	temp |= pll->state.hw_state.pll1;
+	intel_de_write(dev_priv, BXT_PORT_PLL(phy, ch, 1), temp);
+
+	/* Write M2 fraction */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 2));
+	temp &= ~PORT_PLL_M2_FRAC_MASK;
+	temp |= pll->state.hw_state.pll2;
+	intel_de_write(dev_priv, BXT_PORT_PLL(phy, ch, 2), temp);
+
+	/* Write M2 fraction enable */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 3));
+	temp &= ~PORT_PLL_M2_FRAC_ENABLE;
+	temp |= pll->state.hw_state.pll3;
+	intel_de_write(dev_priv, BXT_PORT_PLL(phy, ch, 3), temp);
+
+	/* Write coeff */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 6));
+	temp &= ~PORT_PLL_PROP_COEFF_MASK;
+	temp &= ~PORT_PLL_INT_COEFF_MASK;
+	temp &= ~PORT_PLL_GAIN_CTL_MASK;
+	temp |= pll->state.hw_state.pll6;
+	intel_de_write(dev_priv, BXT_PORT_PLL(phy, ch, 6), temp);
+
+	/* Write calibration val */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 8));
+	temp &= ~PORT_PLL_TARGET_CNT_MASK;
+	temp |= pll->state.hw_state.pll8;
+	intel_de_write(dev_priv, BXT_PORT_PLL(phy, ch, 8), temp);
+
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 9));
+	temp &= ~PORT_PLL_LOCK_THRESHOLD_MASK;
+	temp |= pll->state.hw_state.pll9;
+	intel_de_write(dev_priv, BXT_PORT_PLL(phy, ch, 9), temp);
+
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 10));
+	temp &= ~PORT_PLL_DCO_AMP_OVR_EN_H;
+	temp &= ~PORT_PLL_DCO_AMP_MASK;
+	temp |= pll->state.hw_state.pll10;
+	intel_de_write(dev_priv, BXT_PORT_PLL(phy, ch, 10), temp);
+
+	/* Recalibrate with new settings */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL_EBB_4(phy, ch));
+	temp |= PORT_PLL_RECALIBRATE;
+	intel_de_write(dev_priv, BXT_PORT_PLL_EBB_4(phy, ch), temp);
+	temp &= ~PORT_PLL_10BIT_CLK_ENABLE;
+	temp |= pll->state.hw_state.ebb4;
+	intel_de_write(dev_priv, BXT_PORT_PLL_EBB_4(phy, ch), temp);
+
+	/* Enable PLL */
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL_ENABLE(port));
+	temp |= PORT_PLL_ENABLE;
+	intel_de_write(dev_priv, BXT_PORT_PLL_ENABLE(port), temp);
+	intel_de_posting_read(dev_priv, BXT_PORT_PLL_ENABLE(port));
+
+	if (wait_for_us((intel_de_read(dev_priv, BXT_PORT_PLL_ENABLE(port)) & PORT_PLL_LOCK),
+			200))
+		drm_err(&dev_priv->drm, "PLL %d not locked\n", port);
+
+	if (IS_GEMINILAKE(dev_priv)) {
+		temp = intel_de_read(dev_priv, BXT_PORT_TX_DW5_LN0(phy, ch));
+		temp |= DCC_DELAY_RANGE_2;
+		intel_de_write(dev_priv, BXT_PORT_TX_DW5_GRP(phy, ch), temp);
+	}
+
+	/*
+	 * While we write to the group register to program all lanes at once we
+	 * can read only lane registers and we pick lanes 0/1 for that.
+	 */
+	temp = intel_de_read(dev_priv, BXT_PORT_PCS_DW12_LN01(phy, ch));
+	temp &= ~LANE_STAGGER_MASK;
+	temp &= ~LANESTAGGER_STRAP_OVRD;
+	temp |= pll->state.hw_state.pcsdw12;
+	intel_de_write(dev_priv, BXT_PORT_PCS_DW12_GRP(phy, ch), temp);
+}
+
+static void bxt_ddi_pll_disable(struct drm_i915_private *dev_priv,
+					struct intel_shared_dpll *pll)
+{
+	enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */
+	u32 temp;
+
+	temp = intel_de_read(dev_priv, BXT_PORT_PLL_ENABLE(port));
+	temp &= ~PORT_PLL_ENABLE;
+	intel_de_write(dev_priv, BXT_PORT_PLL_ENABLE(port), temp);
+	intel_de_posting_read(dev_priv, BXT_PORT_PLL_ENABLE(port));
+
+	if (IS_GEMINILAKE(dev_priv)) {
+		temp = intel_de_read(dev_priv, BXT_PORT_PLL_ENABLE(port));
+		temp &= ~PORT_PLL_POWER_ENABLE;
+		intel_de_write(dev_priv, BXT_PORT_PLL_ENABLE(port), temp);
+
+		if (wait_for_us(!(intel_de_read(dev_priv, BXT_PORT_PLL_ENABLE(port)) &
+				  PORT_PLL_POWER_STATE), 200))
+			drm_err(&dev_priv->drm,
+				"Power state not reset for PLL:%d\n", port);
+	}
+}
+
+static bool bxt_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv,
+					struct intel_shared_dpll *pll,
+					struct intel_dpll_hw_state *hw_state)
+{
+	enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */
+	intel_wakeref_t wakeref;
+	enum dpio_phy phy;
+	enum dpio_channel ch;
+	u32 val;
+	bool ret;
+
+	bxt_port_to_phy_channel(dev_priv, port, &phy, &ch);
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     POWER_DOMAIN_DISPLAY_CORE);
+	if (!wakeref)
+		return false;
+
+	ret = false;
+
+	val = intel_de_read(dev_priv, BXT_PORT_PLL_ENABLE(port));
+	if (!(val & PORT_PLL_ENABLE))
+		goto out;
+
+	hw_state->ebb0 = intel_de_read(dev_priv, BXT_PORT_PLL_EBB_0(phy, ch));
+	hw_state->ebb0 &= PORT_PLL_P1_MASK | PORT_PLL_P2_MASK;
+
+	hw_state->ebb4 = intel_de_read(dev_priv, BXT_PORT_PLL_EBB_4(phy, ch));
+	hw_state->ebb4 &= PORT_PLL_10BIT_CLK_ENABLE;
+
+	hw_state->pll0 = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 0));
+	hw_state->pll0 &= PORT_PLL_M2_INT_MASK;
+
+	hw_state->pll1 = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 1));
+	hw_state->pll1 &= PORT_PLL_N_MASK;
+
+	hw_state->pll2 = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 2));
+	hw_state->pll2 &= PORT_PLL_M2_FRAC_MASK;
+
+	hw_state->pll3 = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 3));
+	hw_state->pll3 &= PORT_PLL_M2_FRAC_ENABLE;
+
+	hw_state->pll6 = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 6));
+	hw_state->pll6 &= PORT_PLL_PROP_COEFF_MASK |
+			  PORT_PLL_INT_COEFF_MASK |
+			  PORT_PLL_GAIN_CTL_MASK;
+
+	hw_state->pll8 = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 8));
+	hw_state->pll8 &= PORT_PLL_TARGET_CNT_MASK;
+
+	hw_state->pll9 = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 9));
+	hw_state->pll9 &= PORT_PLL_LOCK_THRESHOLD_MASK;
+
+	hw_state->pll10 = intel_de_read(dev_priv, BXT_PORT_PLL(phy, ch, 10));
+	hw_state->pll10 &= PORT_PLL_DCO_AMP_OVR_EN_H |
+			   PORT_PLL_DCO_AMP_MASK;
+
+	/*
+	 * While we write to the group register to program all lanes at once we
+	 * can read only lane registers. We configure all lanes the same way, so
+	 * here just read out lanes 0/1 and output a note if lanes 2/3 differ.
+	 */
+	hw_state->pcsdw12 = intel_de_read(dev_priv,
+					  BXT_PORT_PCS_DW12_LN01(phy, ch));
+	if (intel_de_read(dev_priv, BXT_PORT_PCS_DW12_LN23(phy, ch)) != hw_state->pcsdw12)
+		drm_dbg(&dev_priv->drm,
+			"lane stagger config different for lane 01 (%08x) and 23 (%08x)\n",
+			hw_state->pcsdw12,
+			intel_de_read(dev_priv,
+				      BXT_PORT_PCS_DW12_LN23(phy, ch)));
+	hw_state->pcsdw12 &= LANE_STAGGER_MASK | LANESTAGGER_STRAP_OVRD;
+
+	ret = true;
+
+out:
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+
+	return ret;
+}
+
+/* pre-calculated values for DP linkrates */
+static const struct dpll bxt_dp_clk_val[] = {
+	/* m2 is .22 binary fixed point */
+	{ .dot = 162000, .p1 = 4, .p2 = 2, .n = 1, .m1 = 2, .m2 = 0x819999a /* 32.4 */ },
+	{ .dot = 270000, .p1 = 4, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6c00000 /* 27.0 */ },
+	{ .dot = 540000, .p1 = 2, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6c00000 /* 27.0 */ },
+	{ .dot = 216000, .p1 = 3, .p2 = 2, .n = 1, .m1 = 2, .m2 = 0x819999a /* 32.4 */ },
+	{ .dot = 243000, .p1 = 4, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6133333 /* 24.3 */ },
+	{ .dot = 324000, .p1 = 4, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x819999a /* 32.4 */ },
+	{ .dot = 432000, .p1 = 3, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x819999a /* 32.4 */ },
+};
+
+static int
+bxt_ddi_hdmi_pll_dividers(struct intel_crtc_state *crtc_state,
+			  struct dpll *clk_div)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+
+	/* Calculate HDMI div */
+	/*
+	 * FIXME: tie the following calculation into
+	 * i9xx_crtc_compute_clock
+	 */
+	if (!bxt_find_best_dpll(crtc_state, clk_div))
+		return -EINVAL;
+
+	drm_WARN_ON(&i915->drm, clk_div->m1 != 2);
+
+	return 0;
+}
+
+static void bxt_ddi_dp_pll_dividers(struct intel_crtc_state *crtc_state,
+				    struct dpll *clk_div)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	int i;
+
+	*clk_div = bxt_dp_clk_val[0];
+	for (i = 0; i < ARRAY_SIZE(bxt_dp_clk_val); ++i) {
+		if (crtc_state->port_clock == bxt_dp_clk_val[i].dot) {
+			*clk_div = bxt_dp_clk_val[i];
+			break;
+		}
+	}
+
+	chv_calc_dpll_params(i915->display.dpll.ref_clks.nssc, clk_div);
+
+	drm_WARN_ON(&i915->drm, clk_div->vco == 0 ||
+		    clk_div->dot != crtc_state->port_clock);
+}
+
+static int bxt_ddi_set_dpll_hw_state(struct intel_crtc_state *crtc_state,
+				     const struct dpll *clk_div)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	struct intel_dpll_hw_state *dpll_hw_state = &crtc_state->dpll_hw_state;
+	int clock = crtc_state->port_clock;
+	int vco = clk_div->vco;
+	u32 prop_coef, int_coef, gain_ctl, targ_cnt;
+	u32 lanestagger;
+
+	if (vco >= 6200000 && vco <= 6700000) {
+		prop_coef = 4;
+		int_coef = 9;
+		gain_ctl = 3;
+		targ_cnt = 8;
+	} else if ((vco > 5400000 && vco < 6200000) ||
+			(vco >= 4800000 && vco < 5400000)) {
+		prop_coef = 5;
+		int_coef = 11;
+		gain_ctl = 3;
+		targ_cnt = 9;
+	} else if (vco == 5400000) {
+		prop_coef = 3;
+		int_coef = 8;
+		gain_ctl = 1;
+		targ_cnt = 9;
+	} else {
+		drm_err(&i915->drm, "Invalid VCO\n");
+		return -EINVAL;
+	}
+
+	if (clock > 270000)
+		lanestagger = 0x18;
+	else if (clock > 135000)
+		lanestagger = 0x0d;
+	else if (clock > 67000)
+		lanestagger = 0x07;
+	else if (clock > 33000)
+		lanestagger = 0x04;
+	else
+		lanestagger = 0x02;
+
+	dpll_hw_state->ebb0 = PORT_PLL_P1(clk_div->p1) | PORT_PLL_P2(clk_div->p2);
+	dpll_hw_state->pll0 = PORT_PLL_M2_INT(clk_div->m2 >> 22);
+	dpll_hw_state->pll1 = PORT_PLL_N(clk_div->n);
+	dpll_hw_state->pll2 = PORT_PLL_M2_FRAC(clk_div->m2 & 0x3fffff);
+
+	if (clk_div->m2 & 0x3fffff)
+		dpll_hw_state->pll3 = PORT_PLL_M2_FRAC_ENABLE;
+
+	dpll_hw_state->pll6 = PORT_PLL_PROP_COEFF(prop_coef) |
+		PORT_PLL_INT_COEFF(int_coef) |
+		PORT_PLL_GAIN_CTL(gain_ctl);
+
+	dpll_hw_state->pll8 = PORT_PLL_TARGET_CNT(targ_cnt);
+
+	dpll_hw_state->pll9 = PORT_PLL_LOCK_THRESHOLD(5);
+
+	dpll_hw_state->pll10 = PORT_PLL_DCO_AMP(15) |
+		PORT_PLL_DCO_AMP_OVR_EN_H;
+
+	dpll_hw_state->ebb4 = PORT_PLL_10BIT_CLK_ENABLE;
+
+	dpll_hw_state->pcsdw12 = LANESTAGGER_STRAP_OVRD | lanestagger;
+
+	return 0;
+}
+
+static int bxt_ddi_pll_get_freq(struct drm_i915_private *i915,
+				const struct intel_shared_dpll *pll,
+				const struct intel_dpll_hw_state *pll_state)
+{
+	struct dpll clock;
+
+	clock.m1 = 2;
+	clock.m2 = REG_FIELD_GET(PORT_PLL_M2_INT_MASK, pll_state->pll0) << 22;
+	if (pll_state->pll3 & PORT_PLL_M2_FRAC_ENABLE)
+		clock.m2 |= REG_FIELD_GET(PORT_PLL_M2_FRAC_MASK, pll_state->pll2);
+	clock.n = REG_FIELD_GET(PORT_PLL_N_MASK, pll_state->pll1);
+	clock.p1 = REG_FIELD_GET(PORT_PLL_P1_MASK, pll_state->ebb0);
+	clock.p2 = REG_FIELD_GET(PORT_PLL_P2_MASK, pll_state->ebb0);
+
+	return chv_calc_dpll_params(i915->display.dpll.ref_clks.nssc, &clock);
+}
+
+static int
+bxt_ddi_dp_set_dpll_hw_state(struct intel_crtc_state *crtc_state)
+{
+	struct dpll clk_div = {};
+
+	bxt_ddi_dp_pll_dividers(crtc_state, &clk_div);
+
+	return bxt_ddi_set_dpll_hw_state(crtc_state, &clk_div);
+}
+
+static int
+bxt_ddi_hdmi_set_dpll_hw_state(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	struct dpll clk_div = {};
+	int ret;
+
+	bxt_ddi_hdmi_pll_dividers(crtc_state, &clk_div);
+
+	ret = bxt_ddi_set_dpll_hw_state(crtc_state, &clk_div);
+	if (ret)
+		return ret;
+
+	crtc_state->port_clock = bxt_ddi_pll_get_freq(i915, NULL,
+						      &crtc_state->dpll_hw_state);
+
+	return 0;
+}
+
+static int bxt_compute_dpll(struct intel_atomic_state *state,
+			    struct intel_crtc *crtc,
+			    struct intel_encoder *encoder)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+		return bxt_ddi_hdmi_set_dpll_hw_state(crtc_state);
+	else if (intel_crtc_has_dp_encoder(crtc_state))
+		return bxt_ddi_dp_set_dpll_hw_state(crtc_state);
+	else
+		return -EINVAL;
+}
+
+static int bxt_get_dpll(struct intel_atomic_state *state,
+			struct intel_crtc *crtc,
+			struct intel_encoder *encoder)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_shared_dpll *pll;
+	enum intel_dpll_id id;
+
+	/* 1:1 mapping between ports and PLLs */
+	id = (enum intel_dpll_id) encoder->port;
+	pll = intel_get_shared_dpll_by_id(dev_priv, id);
+
+	drm_dbg_kms(&dev_priv->drm, "[CRTC:%d:%s] using pre-allocated %s\n",
+		    crtc->base.base.id, crtc->base.name, pll->info->name);
+
+	intel_reference_shared_dpll(state, crtc,
+				    pll, &crtc_state->dpll_hw_state);
+
+	crtc_state->shared_dpll = pll;
+
+	return 0;
+}
+
+static void bxt_update_dpll_ref_clks(struct drm_i915_private *i915)
+{
+	i915->display.dpll.ref_clks.ssc = 100000;
+	i915->display.dpll.ref_clks.nssc = 100000;
+	/* DSI non-SSC ref 19.2MHz */
+}
+
+static void bxt_dump_hw_state(struct drm_i915_private *dev_priv,
+			      const struct intel_dpll_hw_state *hw_state)
+{
+	drm_dbg_kms(&dev_priv->drm, "dpll_hw_state: ebb0: 0x%x, ebb4: 0x%x,"
+		    "pll0: 0x%x, pll1: 0x%x, pll2: 0x%x, pll3: 0x%x, "
+		    "pll6: 0x%x, pll8: 0x%x, pll9: 0x%x, pll10: 0x%x, pcsdw12: 0x%x\n",
+		    hw_state->ebb0,
+		    hw_state->ebb4,
+		    hw_state->pll0,
+		    hw_state->pll1,
+		    hw_state->pll2,
+		    hw_state->pll3,
+		    hw_state->pll6,
+		    hw_state->pll8,
+		    hw_state->pll9,
+		    hw_state->pll10,
+		    hw_state->pcsdw12);
+}
+
+static const struct intel_shared_dpll_funcs bxt_ddi_pll_funcs = {
+	.enable = bxt_ddi_pll_enable,
+	.disable = bxt_ddi_pll_disable,
+	.get_hw_state = bxt_ddi_pll_get_hw_state,
+	.get_freq = bxt_ddi_pll_get_freq,
+};
+
+static const struct dpll_info bxt_plls[] = {
+	{ "PORT PLL A", &bxt_ddi_pll_funcs, DPLL_ID_SKL_DPLL0, 0 },
+	{ "PORT PLL B", &bxt_ddi_pll_funcs, DPLL_ID_SKL_DPLL1, 0 },
+	{ "PORT PLL C", &bxt_ddi_pll_funcs, DPLL_ID_SKL_DPLL2, 0 },
+	{ },
+};
+
+static const struct intel_dpll_mgr bxt_pll_mgr = {
+	.dpll_info = bxt_plls,
+	.compute_dplls = bxt_compute_dpll,
+	.get_dplls = bxt_get_dpll,
+	.put_dplls = intel_put_dpll,
+	.update_ref_clks = bxt_update_dpll_ref_clks,
+	.dump_hw_state = bxt_dump_hw_state,
+};
+
+static void icl_wrpll_get_multipliers(int bestdiv, int *pdiv,
+				      int *qdiv, int *kdiv)
+{
+	/* even dividers */
+	if (bestdiv % 2 == 0) {
+		if (bestdiv == 2) {
+			*pdiv = 2;
+			*qdiv = 1;
+			*kdiv = 1;
+		} else if (bestdiv % 4 == 0) {
+			*pdiv = 2;
+			*qdiv = bestdiv / 4;
+			*kdiv = 2;
+		} else if (bestdiv % 6 == 0) {
+			*pdiv = 3;
+			*qdiv = bestdiv / 6;
+			*kdiv = 2;
+		} else if (bestdiv % 5 == 0) {
+			*pdiv = 5;
+			*qdiv = bestdiv / 10;
+			*kdiv = 2;
+		} else if (bestdiv % 14 == 0) {
+			*pdiv = 7;
+			*qdiv = bestdiv / 14;
+			*kdiv = 2;
+		}
+	} else {
+		if (bestdiv == 3 || bestdiv == 5 || bestdiv == 7) {
+			*pdiv = bestdiv;
+			*qdiv = 1;
+			*kdiv = 1;
+		} else { /* 9, 15, 21 */
+			*pdiv = bestdiv / 3;
+			*qdiv = 1;
+			*kdiv = 3;
+		}
+	}
+}
+
+static void icl_wrpll_params_populate(struct skl_wrpll_params *params,
+				      u32 dco_freq, u32 ref_freq,
+				      int pdiv, int qdiv, int kdiv)
+{
+	u32 dco;
+
+	switch (kdiv) {
+	case 1:
+		params->kdiv = 1;
+		break;
+	case 2:
+		params->kdiv = 2;
+		break;
+	case 3:
+		params->kdiv = 4;
+		break;
+	default:
+		WARN(1, "Incorrect KDiv\n");
+	}
+
+	switch (pdiv) {
+	case 2:
+		params->pdiv = 1;
+		break;
+	case 3:
+		params->pdiv = 2;
+		break;
+	case 5:
+		params->pdiv = 4;
+		break;
+	case 7:
+		params->pdiv = 8;
+		break;
+	default:
+		WARN(1, "Incorrect PDiv\n");
+	}
+
+	WARN_ON(kdiv != 2 && qdiv != 1);
+
+	params->qdiv_ratio = qdiv;
+	params->qdiv_mode = (qdiv == 1) ? 0 : 1;
+
+	dco = div_u64((u64)dco_freq << 15, ref_freq);
+
+	params->dco_integer = dco >> 15;
+	params->dco_fraction = dco & 0x7fff;
+}
+
+/*
+ * Display WA #22010492432: ehl, tgl, adl-s, adl-p
+ * Program half of the nominal DCO divider fraction value.
+ */
+static bool
+ehl_combo_pll_div_frac_wa_needed(struct drm_i915_private *i915)
+{
+	return ((IS_PLATFORM(i915, INTEL_ELKHARTLAKE) &&
+		 IS_JSL_EHL_DISPLAY_STEP(i915, STEP_B0, STEP_FOREVER)) ||
+		 IS_TIGERLAKE(i915) || IS_ALDERLAKE_S(i915) || IS_ALDERLAKE_P(i915)) &&
+		 i915->display.dpll.ref_clks.nssc == 38400;
+}
+
+struct icl_combo_pll_params {
+	int clock;
+	struct skl_wrpll_params wrpll;
+};
+
+/*
+ * These values alrea already adjusted: they're the bits we write to the
+ * registers, not the logical values.
+ */
+static const struct icl_combo_pll_params icl_dp_combo_pll_24MHz_values[] = {
+	{ 540000,
+	  { .dco_integer = 0x151, .dco_fraction = 0x4000,		/* [0]: 5.4 */
+	    .pdiv = 0x2 /* 3 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 270000,
+	  { .dco_integer = 0x151, .dco_fraction = 0x4000,		/* [1]: 2.7 */
+	    .pdiv = 0x2 /* 3 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 162000,
+	  { .dco_integer = 0x151, .dco_fraction = 0x4000,		/* [2]: 1.62 */
+	    .pdiv = 0x4 /* 5 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 324000,
+	  { .dco_integer = 0x151, .dco_fraction = 0x4000,		/* [3]: 3.24 */
+	    .pdiv = 0x4 /* 5 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 216000,
+	  { .dco_integer = 0x168, .dco_fraction = 0x0000,		/* [4]: 2.16 */
+	    .pdiv = 0x1 /* 2 */, .kdiv = 2, .qdiv_mode = 1, .qdiv_ratio = 2, }, },
+	{ 432000,
+	  { .dco_integer = 0x168, .dco_fraction = 0x0000,		/* [5]: 4.32 */
+	    .pdiv = 0x1 /* 2 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 648000,
+	  { .dco_integer = 0x195, .dco_fraction = 0x0000,		/* [6]: 6.48 */
+	    .pdiv = 0x2 /* 3 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 810000,
+	  { .dco_integer = 0x151, .dco_fraction = 0x4000,		/* [7]: 8.1 */
+	    .pdiv = 0x1 /* 2 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+};
+
+
+/* Also used for 38.4 MHz values. */
+static const struct icl_combo_pll_params icl_dp_combo_pll_19_2MHz_values[] = {
+	{ 540000,
+	  { .dco_integer = 0x1A5, .dco_fraction = 0x7000,		/* [0]: 5.4 */
+	    .pdiv = 0x2 /* 3 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 270000,
+	  { .dco_integer = 0x1A5, .dco_fraction = 0x7000,		/* [1]: 2.7 */
+	    .pdiv = 0x2 /* 3 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 162000,
+	  { .dco_integer = 0x1A5, .dco_fraction = 0x7000,		/* [2]: 1.62 */
+	    .pdiv = 0x4 /* 5 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 324000,
+	  { .dco_integer = 0x1A5, .dco_fraction = 0x7000,		/* [3]: 3.24 */
+	    .pdiv = 0x4 /* 5 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 216000,
+	  { .dco_integer = 0x1C2, .dco_fraction = 0x0000,		/* [4]: 2.16 */
+	    .pdiv = 0x1 /* 2 */, .kdiv = 2, .qdiv_mode = 1, .qdiv_ratio = 2, }, },
+	{ 432000,
+	  { .dco_integer = 0x1C2, .dco_fraction = 0x0000,		/* [5]: 4.32 */
+	    .pdiv = 0x1 /* 2 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 648000,
+	  { .dco_integer = 0x1FA, .dco_fraction = 0x2000,		/* [6]: 6.48 */
+	    .pdiv = 0x2 /* 3 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+	{ 810000,
+	  { .dco_integer = 0x1A5, .dco_fraction = 0x7000,		/* [7]: 8.1 */
+	    .pdiv = 0x1 /* 2 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0, }, },
+};
+
+static const struct skl_wrpll_params icl_tbt_pll_24MHz_values = {
+	.dco_integer = 0x151, .dco_fraction = 0x4000,
+	.pdiv = 0x4 /* 5 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0,
+};
+
+static const struct skl_wrpll_params icl_tbt_pll_19_2MHz_values = {
+	.dco_integer = 0x1A5, .dco_fraction = 0x7000,
+	.pdiv = 0x4 /* 5 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0,
+};
+
+static const struct skl_wrpll_params tgl_tbt_pll_19_2MHz_values = {
+	.dco_integer = 0x54, .dco_fraction = 0x3000,
+	/* the following params are unused */
+	.pdiv = 0, .kdiv = 0, .qdiv_mode = 0, .qdiv_ratio = 0,
+};
+
+static const struct skl_wrpll_params tgl_tbt_pll_24MHz_values = {
+	.dco_integer = 0x43, .dco_fraction = 0x4000,
+	/* the following params are unused */
+};
+
+static int icl_calc_dp_combo_pll(struct intel_crtc_state *crtc_state,
+				 struct skl_wrpll_params *pll_params)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	const struct icl_combo_pll_params *params =
+		dev_priv->display.dpll.ref_clks.nssc == 24000 ?
+		icl_dp_combo_pll_24MHz_values :
+		icl_dp_combo_pll_19_2MHz_values;
+	int clock = crtc_state->port_clock;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(icl_dp_combo_pll_24MHz_values); i++) {
+		if (clock == params[i].clock) {
+			*pll_params = params[i].wrpll;
+			return 0;
+		}
+	}
+
+	MISSING_CASE(clock);
+	return -EINVAL;
+}
+
+static int icl_calc_tbt_pll(struct intel_crtc_state *crtc_state,
+			    struct skl_wrpll_params *pll_params)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		switch (dev_priv->display.dpll.ref_clks.nssc) {
+		default:
+			MISSING_CASE(dev_priv->display.dpll.ref_clks.nssc);
+			fallthrough;
+		case 19200:
+		case 38400:
+			*pll_params = tgl_tbt_pll_19_2MHz_values;
+			break;
+		case 24000:
+			*pll_params = tgl_tbt_pll_24MHz_values;
+			break;
+		}
+	} else {
+		switch (dev_priv->display.dpll.ref_clks.nssc) {
+		default:
+			MISSING_CASE(dev_priv->display.dpll.ref_clks.nssc);
+			fallthrough;
+		case 19200:
+		case 38400:
+			*pll_params = icl_tbt_pll_19_2MHz_values;
+			break;
+		case 24000:
+			*pll_params = icl_tbt_pll_24MHz_values;
+			break;
+		}
+	}
+
+	return 0;
+}
+
+static int icl_ddi_tbt_pll_get_freq(struct drm_i915_private *i915,
+				    const struct intel_shared_dpll *pll,
+				    const struct intel_dpll_hw_state *pll_state)
+{
+	/*
+	 * The PLL outputs multiple frequencies at the same time, selection is
+	 * made at DDI clock mux level.
+	 */
+	drm_WARN_ON(&i915->drm, 1);
+
+	return 0;
+}
+
+static int icl_wrpll_ref_clock(struct drm_i915_private *i915)
+{
+	int ref_clock = i915->display.dpll.ref_clks.nssc;
+
+	/*
+	 * For ICL+, the spec states: if reference frequency is 38.4,
+	 * use 19.2 because the DPLL automatically divides that by 2.
+	 */
+	if (ref_clock == 38400)
+		ref_clock = 19200;
+
+	return ref_clock;
+}
+
+static int
+icl_calc_wrpll(struct intel_crtc_state *crtc_state,
+	       struct skl_wrpll_params *wrpll_params)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	int ref_clock = icl_wrpll_ref_clock(i915);
+	u32 afe_clock = crtc_state->port_clock * 5;
+	u32 dco_min = 7998000;
+	u32 dco_max = 10000000;
+	u32 dco_mid = (dco_min + dco_max) / 2;
+	static const int dividers[] = {  2,  4,  6,  8, 10, 12,  14,  16,
+					 18, 20, 24, 28, 30, 32,  36,  40,
+					 42, 44, 48, 50, 52, 54,  56,  60,
+					 64, 66, 68, 70, 72, 76,  78,  80,
+					 84, 88, 90, 92, 96, 98, 100, 102,
+					  3,  5,  7,  9, 15, 21 };
+	u32 dco, best_dco = 0, dco_centrality = 0;
+	u32 best_dco_centrality = U32_MAX; /* Spec meaning of 999999 MHz */
+	int d, best_div = 0, pdiv = 0, qdiv = 0, kdiv = 0;
+
+	for (d = 0; d < ARRAY_SIZE(dividers); d++) {
+		dco = afe_clock * dividers[d];
+
+		if (dco <= dco_max && dco >= dco_min) {
+			dco_centrality = abs(dco - dco_mid);
+
+			if (dco_centrality < best_dco_centrality) {
+				best_dco_centrality = dco_centrality;
+				best_div = dividers[d];
+				best_dco = dco;
+			}
+		}
+	}
+
+	if (best_div == 0)
+		return -EINVAL;
+
+	icl_wrpll_get_multipliers(best_div, &pdiv, &qdiv, &kdiv);
+	icl_wrpll_params_populate(wrpll_params, best_dco, ref_clock,
+				  pdiv, qdiv, kdiv);
+
+	return 0;
+}
+
+static int icl_ddi_combo_pll_get_freq(struct drm_i915_private *i915,
+				      const struct intel_shared_dpll *pll,
+				      const struct intel_dpll_hw_state *pll_state)
+{
+	int ref_clock = icl_wrpll_ref_clock(i915);
+	u32 dco_fraction;
+	u32 p0, p1, p2, dco_freq;
+
+	p0 = pll_state->cfgcr1 & DPLL_CFGCR1_PDIV_MASK;
+	p2 = pll_state->cfgcr1 & DPLL_CFGCR1_KDIV_MASK;
+
+	if (pll_state->cfgcr1 & DPLL_CFGCR1_QDIV_MODE(1))
+		p1 = (pll_state->cfgcr1 & DPLL_CFGCR1_QDIV_RATIO_MASK) >>
+			DPLL_CFGCR1_QDIV_RATIO_SHIFT;
+	else
+		p1 = 1;
+
+	switch (p0) {
+	case DPLL_CFGCR1_PDIV_2:
+		p0 = 2;
+		break;
+	case DPLL_CFGCR1_PDIV_3:
+		p0 = 3;
+		break;
+	case DPLL_CFGCR1_PDIV_5:
+		p0 = 5;
+		break;
+	case DPLL_CFGCR1_PDIV_7:
+		p0 = 7;
+		break;
+	}
+
+	switch (p2) {
+	case DPLL_CFGCR1_KDIV_1:
+		p2 = 1;
+		break;
+	case DPLL_CFGCR1_KDIV_2:
+		p2 = 2;
+		break;
+	case DPLL_CFGCR1_KDIV_3:
+		p2 = 3;
+		break;
+	}
+
+	dco_freq = (pll_state->cfgcr0 & DPLL_CFGCR0_DCO_INTEGER_MASK) *
+		   ref_clock;
+
+	dco_fraction = (pll_state->cfgcr0 & DPLL_CFGCR0_DCO_FRACTION_MASK) >>
+		       DPLL_CFGCR0_DCO_FRACTION_SHIFT;
+
+	if (ehl_combo_pll_div_frac_wa_needed(i915))
+		dco_fraction *= 2;
+
+	dco_freq += (dco_fraction * ref_clock) / 0x8000;
+
+	if (drm_WARN_ON(&i915->drm, p0 == 0 || p1 == 0 || p2 == 0))
+		return 0;
+
+	return dco_freq / (p0 * p1 * p2 * 5);
+}
+
+static void icl_calc_dpll_state(struct drm_i915_private *i915,
+				const struct skl_wrpll_params *pll_params,
+				struct intel_dpll_hw_state *pll_state)
+{
+	u32 dco_fraction = pll_params->dco_fraction;
+
+	if (ehl_combo_pll_div_frac_wa_needed(i915))
+		dco_fraction = DIV_ROUND_CLOSEST(dco_fraction, 2);
+
+	pll_state->cfgcr0 = DPLL_CFGCR0_DCO_FRACTION(dco_fraction) |
+			    pll_params->dco_integer;
+
+	pll_state->cfgcr1 = DPLL_CFGCR1_QDIV_RATIO(pll_params->qdiv_ratio) |
+			    DPLL_CFGCR1_QDIV_MODE(pll_params->qdiv_mode) |
+			    DPLL_CFGCR1_KDIV(pll_params->kdiv) |
+			    DPLL_CFGCR1_PDIV(pll_params->pdiv);
+
+	if (DISPLAY_VER(i915) >= 12)
+		pll_state->cfgcr1 |= TGL_DPLL_CFGCR1_CFSELOVRD_NORMAL_XTAL;
+	else
+		pll_state->cfgcr1 |= DPLL_CFGCR1_CENTRAL_FREQ_8400;
+
+	if (i915->display.vbt.override_afc_startup)
+		pll_state->div0 = TGL_DPLL0_DIV0_AFC_STARTUP(i915->display.vbt.override_afc_startup_val);
+}
+
+static int icl_mg_pll_find_divisors(int clock_khz, bool is_dp, bool use_ssc,
+				    u32 *target_dco_khz,
+				    struct intel_dpll_hw_state *state,
+				    bool is_dkl)
+{
+	static const u8 div1_vals[] = { 7, 5, 3, 2 };
+	u32 dco_min_freq, dco_max_freq;
+	unsigned int i;
+	int div2;
+
+	dco_min_freq = is_dp ? 8100000 : use_ssc ? 8000000 : 7992000;
+	dco_max_freq = is_dp ? 8100000 : 10000000;
+
+	for (i = 0; i < ARRAY_SIZE(div1_vals); i++) {
+		int div1 = div1_vals[i];
+
+		for (div2 = 10; div2 > 0; div2--) {
+			int dco = div1 * div2 * clock_khz * 5;
+			int a_divratio, tlinedrv, inputsel;
+			u32 hsdiv;
+
+			if (dco < dco_min_freq || dco > dco_max_freq)
+				continue;
+
+			if (div2 >= 2) {
+				/*
+				 * Note: a_divratio not matching TGL BSpec
+				 * algorithm but matching hardcoded values and
+				 * working on HW for DP alt-mode at least
+				 */
+				a_divratio = is_dp ? 10 : 5;
+				tlinedrv = is_dkl ? 1 : 2;
+			} else {
+				a_divratio = 5;
+				tlinedrv = 0;
+			}
+			inputsel = is_dp ? 0 : 1;
+
+			switch (div1) {
+			default:
+				MISSING_CASE(div1);
+				fallthrough;
+			case 2:
+				hsdiv = MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_2;
+				break;
+			case 3:
+				hsdiv = MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_3;
+				break;
+			case 5:
+				hsdiv = MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_5;
+				break;
+			case 7:
+				hsdiv = MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_7;
+				break;
+			}
+
+			*target_dco_khz = dco;
+
+			state->mg_refclkin_ctl = MG_REFCLKIN_CTL_OD_2_MUX(1);
+
+			state->mg_clktop2_coreclkctl1 =
+				MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO(a_divratio);
+
+			state->mg_clktop2_hsclkctl =
+				MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL(tlinedrv) |
+				MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL(inputsel) |
+				hsdiv |
+				MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO(div2);
+
+			return 0;
+		}
+	}
+
+	return -EINVAL;
+}
+
+/*
+ * The specification for this function uses real numbers, so the math had to be
+ * adapted to integer-only calculation, that's why it looks so different.
+ */
+static int icl_calc_mg_pll_state(struct intel_crtc_state *crtc_state,
+				 struct intel_dpll_hw_state *pll_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	int refclk_khz = dev_priv->display.dpll.ref_clks.nssc;
+	int clock = crtc_state->port_clock;
+	u32 dco_khz, m1div, m2div_int, m2div_rem, m2div_frac;
+	u32 iref_ndiv, iref_trim, iref_pulse_w;
+	u32 prop_coeff, int_coeff;
+	u32 tdc_targetcnt, feedfwgain;
+	u64 ssc_stepsize, ssc_steplen, ssc_steplog;
+	u64 tmp;
+	bool use_ssc = false;
+	bool is_dp = !intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI);
+	bool is_dkl = DISPLAY_VER(dev_priv) >= 12;
+	int ret;
+
+	ret = icl_mg_pll_find_divisors(clock, is_dp, use_ssc, &dco_khz,
+				       pll_state, is_dkl);
+	if (ret)
+		return ret;
+
+	m1div = 2;
+	m2div_int = dco_khz / (refclk_khz * m1div);
+	if (m2div_int > 255) {
+		if (!is_dkl) {
+			m1div = 4;
+			m2div_int = dco_khz / (refclk_khz * m1div);
+		}
+
+		if (m2div_int > 255)
+			return -EINVAL;
+	}
+	m2div_rem = dco_khz % (refclk_khz * m1div);
+
+	tmp = (u64)m2div_rem * (1 << 22);
+	do_div(tmp, refclk_khz * m1div);
+	m2div_frac = tmp;
+
+	switch (refclk_khz) {
+	case 19200:
+		iref_ndiv = 1;
+		iref_trim = 28;
+		iref_pulse_w = 1;
+		break;
+	case 24000:
+		iref_ndiv = 1;
+		iref_trim = 25;
+		iref_pulse_w = 2;
+		break;
+	case 38400:
+		iref_ndiv = 2;
+		iref_trim = 28;
+		iref_pulse_w = 1;
+		break;
+	default:
+		MISSING_CASE(refclk_khz);
+		return -EINVAL;
+	}
+
+	/*
+	 * tdc_res = 0.000003
+	 * tdc_targetcnt = int(2 / (tdc_res * 8 * 50 * 1.1) / refclk_mhz + 0.5)
+	 *
+	 * The multiplication by 1000 is due to refclk MHz to KHz conversion. It
+	 * was supposed to be a division, but we rearranged the operations of
+	 * the formula to avoid early divisions so we don't multiply the
+	 * rounding errors.
+	 *
+	 * 0.000003 * 8 * 50 * 1.1 = 0.00132, also known as 132 / 100000, which
+	 * we also rearrange to work with integers.
+	 *
+	 * The 0.5 transformed to 5 results in a multiplication by 10 and the
+	 * last division by 10.
+	 */
+	tdc_targetcnt = (2 * 1000 * 100000 * 10 / (132 * refclk_khz) + 5) / 10;
+
+	/*
+	 * Here we divide dco_khz by 10 in order to allow the dividend to fit in
+	 * 32 bits. That's not a problem since we round the division down
+	 * anyway.
+	 */
+	feedfwgain = (use_ssc || m2div_rem > 0) ?
+		m1div * 1000000 * 100 / (dco_khz * 3 / 10) : 0;
+
+	if (dco_khz >= 9000000) {
+		prop_coeff = 5;
+		int_coeff = 10;
+	} else {
+		prop_coeff = 4;
+		int_coeff = 8;
+	}
+
+	if (use_ssc) {
+		tmp = mul_u32_u32(dco_khz, 47 * 32);
+		do_div(tmp, refclk_khz * m1div * 10000);
+		ssc_stepsize = tmp;
+
+		tmp = mul_u32_u32(dco_khz, 1000);
+		ssc_steplen = DIV_ROUND_UP_ULL(tmp, 32 * 2 * 32);
+	} else {
+		ssc_stepsize = 0;
+		ssc_steplen = 0;
+	}
+	ssc_steplog = 4;
+
+	/* write pll_state calculations */
+	if (is_dkl) {
+		pll_state->mg_pll_div0 = DKL_PLL_DIV0_INTEG_COEFF(int_coeff) |
+					 DKL_PLL_DIV0_PROP_COEFF(prop_coeff) |
+					 DKL_PLL_DIV0_FBPREDIV(m1div) |
+					 DKL_PLL_DIV0_FBDIV_INT(m2div_int);
+		if (dev_priv->display.vbt.override_afc_startup) {
+			u8 val = dev_priv->display.vbt.override_afc_startup_val;
+
+			pll_state->mg_pll_div0 |= DKL_PLL_DIV0_AFC_STARTUP(val);
+		}
+
+		pll_state->mg_pll_div1 = DKL_PLL_DIV1_IREF_TRIM(iref_trim) |
+					 DKL_PLL_DIV1_TDC_TARGET_CNT(tdc_targetcnt);
+
+		pll_state->mg_pll_ssc = DKL_PLL_SSC_IREF_NDIV_RATIO(iref_ndiv) |
+					DKL_PLL_SSC_STEP_LEN(ssc_steplen) |
+					DKL_PLL_SSC_STEP_NUM(ssc_steplog) |
+					(use_ssc ? DKL_PLL_SSC_EN : 0);
+
+		pll_state->mg_pll_bias = (m2div_frac ? DKL_PLL_BIAS_FRAC_EN_H : 0) |
+					  DKL_PLL_BIAS_FBDIV_FRAC(m2div_frac);
+
+		pll_state->mg_pll_tdc_coldst_bias =
+				DKL_PLL_TDC_SSC_STEP_SIZE(ssc_stepsize) |
+				DKL_PLL_TDC_FEED_FWD_GAIN(feedfwgain);
+
+	} else {
+		pll_state->mg_pll_div0 =
+			(m2div_rem > 0 ? MG_PLL_DIV0_FRACNEN_H : 0) |
+			MG_PLL_DIV0_FBDIV_FRAC(m2div_frac) |
+			MG_PLL_DIV0_FBDIV_INT(m2div_int);
+
+		pll_state->mg_pll_div1 =
+			MG_PLL_DIV1_IREF_NDIVRATIO(iref_ndiv) |
+			MG_PLL_DIV1_DITHER_DIV_2 |
+			MG_PLL_DIV1_NDIVRATIO(1) |
+			MG_PLL_DIV1_FBPREDIV(m1div);
+
+		pll_state->mg_pll_lf =
+			MG_PLL_LF_TDCTARGETCNT(tdc_targetcnt) |
+			MG_PLL_LF_AFCCNTSEL_512 |
+			MG_PLL_LF_GAINCTRL(1) |
+			MG_PLL_LF_INT_COEFF(int_coeff) |
+			MG_PLL_LF_PROP_COEFF(prop_coeff);
+
+		pll_state->mg_pll_frac_lock =
+			MG_PLL_FRAC_LOCK_TRUELOCK_CRIT_32 |
+			MG_PLL_FRAC_LOCK_EARLYLOCK_CRIT_32 |
+			MG_PLL_FRAC_LOCK_LOCKTHRESH(10) |
+			MG_PLL_FRAC_LOCK_DCODITHEREN |
+			MG_PLL_FRAC_LOCK_FEEDFWRDGAIN(feedfwgain);
+		if (use_ssc || m2div_rem > 0)
+			pll_state->mg_pll_frac_lock |=
+				MG_PLL_FRAC_LOCK_FEEDFWRDCAL_EN;
+
+		pll_state->mg_pll_ssc =
+			(use_ssc ? MG_PLL_SSC_EN : 0) |
+			MG_PLL_SSC_TYPE(2) |
+			MG_PLL_SSC_STEPLENGTH(ssc_steplen) |
+			MG_PLL_SSC_STEPNUM(ssc_steplog) |
+			MG_PLL_SSC_FLLEN |
+			MG_PLL_SSC_STEPSIZE(ssc_stepsize);
+
+		pll_state->mg_pll_tdc_coldst_bias =
+			MG_PLL_TDC_COLDST_COLDSTART |
+			MG_PLL_TDC_COLDST_IREFINT_EN |
+			MG_PLL_TDC_COLDST_REFBIAS_START_PULSE_W(iref_pulse_w) |
+			MG_PLL_TDC_TDCOVCCORR_EN |
+			MG_PLL_TDC_TDCSEL(3);
+
+		pll_state->mg_pll_bias =
+			MG_PLL_BIAS_BIAS_GB_SEL(3) |
+			MG_PLL_BIAS_INIT_DCOAMP(0x3F) |
+			MG_PLL_BIAS_BIAS_BONUS(10) |
+			MG_PLL_BIAS_BIASCAL_EN |
+			MG_PLL_BIAS_CTRIM(12) |
+			MG_PLL_BIAS_VREF_RDAC(4) |
+			MG_PLL_BIAS_IREFTRIM(iref_trim);
+
+		if (refclk_khz == 38400) {
+			pll_state->mg_pll_tdc_coldst_bias_mask =
+				MG_PLL_TDC_COLDST_COLDSTART;
+			pll_state->mg_pll_bias_mask = 0;
+		} else {
+			pll_state->mg_pll_tdc_coldst_bias_mask = -1U;
+			pll_state->mg_pll_bias_mask = -1U;
+		}
+
+		pll_state->mg_pll_tdc_coldst_bias &=
+			pll_state->mg_pll_tdc_coldst_bias_mask;
+		pll_state->mg_pll_bias &= pll_state->mg_pll_bias_mask;
+	}
+
+	return 0;
+}
+
+static int icl_ddi_mg_pll_get_freq(struct drm_i915_private *dev_priv,
+				   const struct intel_shared_dpll *pll,
+				   const struct intel_dpll_hw_state *pll_state)
+{
+	u32 m1, m2_int, m2_frac, div1, div2, ref_clock;
+	u64 tmp;
+
+	ref_clock = dev_priv->display.dpll.ref_clks.nssc;
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		m1 = pll_state->mg_pll_div0 & DKL_PLL_DIV0_FBPREDIV_MASK;
+		m1 = m1 >> DKL_PLL_DIV0_FBPREDIV_SHIFT;
+		m2_int = pll_state->mg_pll_div0 & DKL_PLL_DIV0_FBDIV_INT_MASK;
+
+		if (pll_state->mg_pll_bias & DKL_PLL_BIAS_FRAC_EN_H) {
+			m2_frac = pll_state->mg_pll_bias &
+				  DKL_PLL_BIAS_FBDIV_FRAC_MASK;
+			m2_frac = m2_frac >> DKL_PLL_BIAS_FBDIV_SHIFT;
+		} else {
+			m2_frac = 0;
+		}
+	} else {
+		m1 = pll_state->mg_pll_div1 & MG_PLL_DIV1_FBPREDIV_MASK;
+		m2_int = pll_state->mg_pll_div0 & MG_PLL_DIV0_FBDIV_INT_MASK;
+
+		if (pll_state->mg_pll_div0 & MG_PLL_DIV0_FRACNEN_H) {
+			m2_frac = pll_state->mg_pll_div0 &
+				  MG_PLL_DIV0_FBDIV_FRAC_MASK;
+			m2_frac = m2_frac >> MG_PLL_DIV0_FBDIV_FRAC_SHIFT;
+		} else {
+			m2_frac = 0;
+		}
+	}
+
+	switch (pll_state->mg_clktop2_hsclkctl &
+		MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_MASK) {
+	case MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_2:
+		div1 = 2;
+		break;
+	case MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_3:
+		div1 = 3;
+		break;
+	case MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_5:
+		div1 = 5;
+		break;
+	case MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_7:
+		div1 = 7;
+		break;
+	default:
+		MISSING_CASE(pll_state->mg_clktop2_hsclkctl);
+		return 0;
+	}
+
+	div2 = (pll_state->mg_clktop2_hsclkctl &
+		MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_MASK) >>
+		MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_SHIFT;
+
+	/* div2 value of 0 is same as 1 means no div */
+	if (div2 == 0)
+		div2 = 1;
+
+	/*
+	 * Adjust the original formula to delay the division by 2^22 in order to
+	 * minimize possible rounding errors.
+	 */
+	tmp = (u64)m1 * m2_int * ref_clock +
+	      (((u64)m1 * m2_frac * ref_clock) >> 22);
+	tmp = div_u64(tmp, 5 * div1 * div2);
+
+	return tmp;
+}
+
+/**
+ * icl_set_active_port_dpll - select the active port DPLL for a given CRTC
+ * @crtc_state: state for the CRTC to select the DPLL for
+ * @port_dpll_id: the active @port_dpll_id to select
+ *
+ * Select the given @port_dpll_id instance from the DPLLs reserved for the
+ * CRTC.
+ */
+void icl_set_active_port_dpll(struct intel_crtc_state *crtc_state,
+			      enum icl_port_dpll_id port_dpll_id)
+{
+	struct icl_port_dpll *port_dpll =
+		&crtc_state->icl_port_dplls[port_dpll_id];
+
+	crtc_state->shared_dpll = port_dpll->pll;
+	crtc_state->dpll_hw_state = port_dpll->hw_state;
+}
+
+static void icl_update_active_dpll(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc,
+				   struct intel_encoder *encoder)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_digital_port *primary_port;
+	enum icl_port_dpll_id port_dpll_id = ICL_PORT_DPLL_DEFAULT;
+
+	primary_port = encoder->type == INTEL_OUTPUT_DP_MST ?
+		enc_to_mst(encoder)->primary :
+		enc_to_dig_port(encoder);
+
+	if (primary_port &&
+	    (intel_tc_port_in_dp_alt_mode(primary_port) ||
+	     intel_tc_port_in_legacy_mode(primary_port)))
+		port_dpll_id = ICL_PORT_DPLL_MG_PHY;
+
+	icl_set_active_port_dpll(crtc_state, port_dpll_id);
+}
+
+static u32 intel_get_hti_plls(struct drm_i915_private *i915)
+{
+	if (!(i915->hti_state & HDPORT_ENABLED))
+		return 0;
+
+	return REG_FIELD_GET(HDPORT_DPLL_USED_MASK, i915->hti_state);
+}
+
+static int icl_compute_combo_phy_dpll(struct intel_atomic_state *state,
+				      struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct icl_port_dpll *port_dpll =
+		&crtc_state->icl_port_dplls[ICL_PORT_DPLL_DEFAULT];
+	struct skl_wrpll_params pll_params = {};
+	int ret;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI) ||
+	    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI))
+		ret = icl_calc_wrpll(crtc_state, &pll_params);
+	else
+		ret = icl_calc_dp_combo_pll(crtc_state, &pll_params);
+
+	if (ret)
+		return ret;
+
+	icl_calc_dpll_state(dev_priv, &pll_params, &port_dpll->hw_state);
+
+	/* this is mainly for the fastset check */
+	icl_set_active_port_dpll(crtc_state, ICL_PORT_DPLL_DEFAULT);
+
+	crtc_state->port_clock = icl_ddi_combo_pll_get_freq(dev_priv, NULL,
+							    &port_dpll->hw_state);
+
+	return 0;
+}
+
+static int icl_get_combo_phy_dpll(struct intel_atomic_state *state,
+				  struct intel_crtc *crtc,
+				  struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct icl_port_dpll *port_dpll =
+		&crtc_state->icl_port_dplls[ICL_PORT_DPLL_DEFAULT];
+	enum port port = encoder->port;
+	unsigned long dpll_mask;
+
+	if (IS_ALDERLAKE_S(dev_priv)) {
+		dpll_mask =
+			BIT(DPLL_ID_DG1_DPLL3) |
+			BIT(DPLL_ID_DG1_DPLL2) |
+			BIT(DPLL_ID_ICL_DPLL1) |
+			BIT(DPLL_ID_ICL_DPLL0);
+	} else if (IS_DG1(dev_priv)) {
+		if (port == PORT_D || port == PORT_E) {
+			dpll_mask =
+				BIT(DPLL_ID_DG1_DPLL2) |
+				BIT(DPLL_ID_DG1_DPLL3);
+		} else {
+			dpll_mask =
+				BIT(DPLL_ID_DG1_DPLL0) |
+				BIT(DPLL_ID_DG1_DPLL1);
+		}
+	} else if (IS_ROCKETLAKE(dev_priv)) {
+		dpll_mask =
+			BIT(DPLL_ID_EHL_DPLL4) |
+			BIT(DPLL_ID_ICL_DPLL1) |
+			BIT(DPLL_ID_ICL_DPLL0);
+	} else if (IS_JSL_EHL(dev_priv) && port != PORT_A) {
+		dpll_mask =
+			BIT(DPLL_ID_EHL_DPLL4) |
+			BIT(DPLL_ID_ICL_DPLL1) |
+			BIT(DPLL_ID_ICL_DPLL0);
+	} else {
+		dpll_mask = BIT(DPLL_ID_ICL_DPLL1) | BIT(DPLL_ID_ICL_DPLL0);
+	}
+
+	/* Eliminate DPLLs from consideration if reserved by HTI */
+	dpll_mask &= ~intel_get_hti_plls(dev_priv);
+
+	port_dpll->pll = intel_find_shared_dpll(state, crtc,
+						&port_dpll->hw_state,
+						dpll_mask);
+	if (!port_dpll->pll)
+		return -EINVAL;
+
+	intel_reference_shared_dpll(state, crtc,
+				    port_dpll->pll, &port_dpll->hw_state);
+
+	icl_update_active_dpll(state, crtc, encoder);
+
+	return 0;
+}
+
+static int icl_compute_tc_phy_dplls(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct icl_port_dpll *port_dpll =
+		&crtc_state->icl_port_dplls[ICL_PORT_DPLL_DEFAULT];
+	struct skl_wrpll_params pll_params = {};
+	int ret;
+
+	port_dpll = &crtc_state->icl_port_dplls[ICL_PORT_DPLL_DEFAULT];
+	ret = icl_calc_tbt_pll(crtc_state, &pll_params);
+	if (ret)
+		return ret;
+
+	icl_calc_dpll_state(dev_priv, &pll_params, &port_dpll->hw_state);
+
+	port_dpll = &crtc_state->icl_port_dplls[ICL_PORT_DPLL_MG_PHY];
+	ret = icl_calc_mg_pll_state(crtc_state, &port_dpll->hw_state);
+	if (ret)
+		return ret;
+
+	/* this is mainly for the fastset check */
+	icl_set_active_port_dpll(crtc_state, ICL_PORT_DPLL_MG_PHY);
+
+	crtc_state->port_clock = icl_ddi_mg_pll_get_freq(dev_priv, NULL,
+							 &port_dpll->hw_state);
+
+	return 0;
+}
+
+static int icl_get_tc_phy_dplls(struct intel_atomic_state *state,
+				struct intel_crtc *crtc,
+				struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct icl_port_dpll *port_dpll =
+		&crtc_state->icl_port_dplls[ICL_PORT_DPLL_DEFAULT];
+	enum intel_dpll_id dpll_id;
+	int ret;
+
+	port_dpll = &crtc_state->icl_port_dplls[ICL_PORT_DPLL_DEFAULT];
+	port_dpll->pll = intel_find_shared_dpll(state, crtc,
+						&port_dpll->hw_state,
+						BIT(DPLL_ID_ICL_TBTPLL));
+	if (!port_dpll->pll)
+		return -EINVAL;
+	intel_reference_shared_dpll(state, crtc,
+				    port_dpll->pll, &port_dpll->hw_state);
+
+
+	port_dpll = &crtc_state->icl_port_dplls[ICL_PORT_DPLL_MG_PHY];
+	dpll_id = icl_tc_port_to_pll_id(intel_port_to_tc(dev_priv,
+							 encoder->port));
+	port_dpll->pll = intel_find_shared_dpll(state, crtc,
+						&port_dpll->hw_state,
+						BIT(dpll_id));
+	if (!port_dpll->pll) {
+		ret = -EINVAL;
+		goto err_unreference_tbt_pll;
+	}
+	intel_reference_shared_dpll(state, crtc,
+				    port_dpll->pll, &port_dpll->hw_state);
+
+	icl_update_active_dpll(state, crtc, encoder);
+
+	return 0;
+
+err_unreference_tbt_pll:
+	port_dpll = &crtc_state->icl_port_dplls[ICL_PORT_DPLL_DEFAULT];
+	intel_unreference_shared_dpll(state, crtc, port_dpll->pll);
+
+	return ret;
+}
+
+static int icl_compute_dplls(struct intel_atomic_state *state,
+			     struct intel_crtc *crtc,
+			     struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+
+	if (intel_phy_is_combo(dev_priv, phy))
+		return icl_compute_combo_phy_dpll(state, crtc);
+	else if (intel_phy_is_tc(dev_priv, phy))
+		return icl_compute_tc_phy_dplls(state, crtc);
+
+	MISSING_CASE(phy);
+
+	return 0;
+}
+
+static int icl_get_dplls(struct intel_atomic_state *state,
+			 struct intel_crtc *crtc,
+			 struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+
+	if (intel_phy_is_combo(dev_priv, phy))
+		return icl_get_combo_phy_dpll(state, crtc, encoder);
+	else if (intel_phy_is_tc(dev_priv, phy))
+		return icl_get_tc_phy_dplls(state, crtc, encoder);
+
+	MISSING_CASE(phy);
+
+	return -EINVAL;
+}
+
+static void icl_put_dplls(struct intel_atomic_state *state,
+			  struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	enum icl_port_dpll_id id;
+
+	new_crtc_state->shared_dpll = NULL;
+
+	for (id = ICL_PORT_DPLL_DEFAULT; id < ICL_PORT_DPLL_COUNT; id++) {
+		const struct icl_port_dpll *old_port_dpll =
+			&old_crtc_state->icl_port_dplls[id];
+		struct icl_port_dpll *new_port_dpll =
+			&new_crtc_state->icl_port_dplls[id];
+
+		new_port_dpll->pll = NULL;
+
+		if (!old_port_dpll->pll)
+			continue;
+
+		intel_unreference_shared_dpll(state, crtc, old_port_dpll->pll);
+	}
+}
+
+static bool mg_pll_get_hw_state(struct drm_i915_private *dev_priv,
+				struct intel_shared_dpll *pll,
+				struct intel_dpll_hw_state *hw_state)
+{
+	const enum intel_dpll_id id = pll->info->id;
+	enum tc_port tc_port = icl_pll_id_to_tc_port(id);
+	intel_wakeref_t wakeref;
+	bool ret = false;
+	u32 val;
+
+	i915_reg_t enable_reg = intel_tc_pll_enable_reg(dev_priv, pll);
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     POWER_DOMAIN_DISPLAY_CORE);
+	if (!wakeref)
+		return false;
+
+	val = intel_de_read(dev_priv, enable_reg);
+	if (!(val & PLL_ENABLE))
+		goto out;
+
+	hw_state->mg_refclkin_ctl = intel_de_read(dev_priv,
+						  MG_REFCLKIN_CTL(tc_port));
+	hw_state->mg_refclkin_ctl &= MG_REFCLKIN_CTL_OD_2_MUX_MASK;
+
+	hw_state->mg_clktop2_coreclkctl1 =
+		intel_de_read(dev_priv, MG_CLKTOP2_CORECLKCTL1(tc_port));
+	hw_state->mg_clktop2_coreclkctl1 &=
+		MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO_MASK;
+
+	hw_state->mg_clktop2_hsclkctl =
+		intel_de_read(dev_priv, MG_CLKTOP2_HSCLKCTL(tc_port));
+	hw_state->mg_clktop2_hsclkctl &=
+		MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL_MASK |
+		MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL_MASK |
+		MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_MASK |
+		MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_MASK;
+
+	hw_state->mg_pll_div0 = intel_de_read(dev_priv, MG_PLL_DIV0(tc_port));
+	hw_state->mg_pll_div1 = intel_de_read(dev_priv, MG_PLL_DIV1(tc_port));
+	hw_state->mg_pll_lf = intel_de_read(dev_priv, MG_PLL_LF(tc_port));
+	hw_state->mg_pll_frac_lock = intel_de_read(dev_priv,
+						   MG_PLL_FRAC_LOCK(tc_port));
+	hw_state->mg_pll_ssc = intel_de_read(dev_priv, MG_PLL_SSC(tc_port));
+
+	hw_state->mg_pll_bias = intel_de_read(dev_priv, MG_PLL_BIAS(tc_port));
+	hw_state->mg_pll_tdc_coldst_bias =
+		intel_de_read(dev_priv, MG_PLL_TDC_COLDST_BIAS(tc_port));
+
+	if (dev_priv->display.dpll.ref_clks.nssc == 38400) {
+		hw_state->mg_pll_tdc_coldst_bias_mask = MG_PLL_TDC_COLDST_COLDSTART;
+		hw_state->mg_pll_bias_mask = 0;
+	} else {
+		hw_state->mg_pll_tdc_coldst_bias_mask = -1U;
+		hw_state->mg_pll_bias_mask = -1U;
+	}
+
+	hw_state->mg_pll_tdc_coldst_bias &= hw_state->mg_pll_tdc_coldst_bias_mask;
+	hw_state->mg_pll_bias &= hw_state->mg_pll_bias_mask;
+
+	ret = true;
+out:
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+	return ret;
+}
+
+static bool dkl_pll_get_hw_state(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll,
+				 struct intel_dpll_hw_state *hw_state)
+{
+	const enum intel_dpll_id id = pll->info->id;
+	enum tc_port tc_port = icl_pll_id_to_tc_port(id);
+	intel_wakeref_t wakeref;
+	bool ret = false;
+	u32 val;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     POWER_DOMAIN_DISPLAY_CORE);
+	if (!wakeref)
+		return false;
+
+	val = intel_de_read(dev_priv, intel_tc_pll_enable_reg(dev_priv, pll));
+	if (!(val & PLL_ENABLE))
+		goto out;
+
+	/*
+	 * All registers read here have the same HIP_INDEX_REG even though
+	 * they are on different building blocks
+	 */
+	hw_state->mg_refclkin_ctl = intel_dkl_phy_read(dev_priv,
+						       DKL_REFCLKIN_CTL(tc_port), 2);
+	hw_state->mg_refclkin_ctl &= MG_REFCLKIN_CTL_OD_2_MUX_MASK;
+
+	hw_state->mg_clktop2_hsclkctl =
+		intel_dkl_phy_read(dev_priv, DKL_CLKTOP2_HSCLKCTL(tc_port), 2);
+	hw_state->mg_clktop2_hsclkctl &=
+		MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL_MASK |
+		MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL_MASK |
+		MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_MASK |
+		MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_MASK;
+
+	hw_state->mg_clktop2_coreclkctl1 =
+		intel_dkl_phy_read(dev_priv, DKL_CLKTOP2_CORECLKCTL1(tc_port), 2);
+	hw_state->mg_clktop2_coreclkctl1 &=
+		MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO_MASK;
+
+	hw_state->mg_pll_div0 = intel_dkl_phy_read(dev_priv, DKL_PLL_DIV0(tc_port), 2);
+	val = DKL_PLL_DIV0_MASK;
+	if (dev_priv->display.vbt.override_afc_startup)
+		val |= DKL_PLL_DIV0_AFC_STARTUP_MASK;
+	hw_state->mg_pll_div0 &= val;
+
+	hw_state->mg_pll_div1 = intel_dkl_phy_read(dev_priv, DKL_PLL_DIV1(tc_port), 2);
+	hw_state->mg_pll_div1 &= (DKL_PLL_DIV1_IREF_TRIM_MASK |
+				  DKL_PLL_DIV1_TDC_TARGET_CNT_MASK);
+
+	hw_state->mg_pll_ssc = intel_dkl_phy_read(dev_priv, DKL_PLL_SSC(tc_port), 2);
+	hw_state->mg_pll_ssc &= (DKL_PLL_SSC_IREF_NDIV_RATIO_MASK |
+				 DKL_PLL_SSC_STEP_LEN_MASK |
+				 DKL_PLL_SSC_STEP_NUM_MASK |
+				 DKL_PLL_SSC_EN);
+
+	hw_state->mg_pll_bias = intel_dkl_phy_read(dev_priv, DKL_PLL_BIAS(tc_port), 2);
+	hw_state->mg_pll_bias &= (DKL_PLL_BIAS_FRAC_EN_H |
+				  DKL_PLL_BIAS_FBDIV_FRAC_MASK);
+
+	hw_state->mg_pll_tdc_coldst_bias =
+		intel_dkl_phy_read(dev_priv, DKL_PLL_TDC_COLDST_BIAS(tc_port), 2);
+	hw_state->mg_pll_tdc_coldst_bias &= (DKL_PLL_TDC_SSC_STEP_SIZE_MASK |
+					     DKL_PLL_TDC_FEED_FWD_GAIN_MASK);
+
+	ret = true;
+out:
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+	return ret;
+}
+
+static bool icl_pll_get_hw_state(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll,
+				 struct intel_dpll_hw_state *hw_state,
+				 i915_reg_t enable_reg)
+{
+	const enum intel_dpll_id id = pll->info->id;
+	intel_wakeref_t wakeref;
+	bool ret = false;
+	u32 val;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     POWER_DOMAIN_DISPLAY_CORE);
+	if (!wakeref)
+		return false;
+
+	val = intel_de_read(dev_priv, enable_reg);
+	if (!(val & PLL_ENABLE))
+		goto out;
+
+	if (IS_ALDERLAKE_S(dev_priv)) {
+		hw_state->cfgcr0 = intel_de_read(dev_priv, ADLS_DPLL_CFGCR0(id));
+		hw_state->cfgcr1 = intel_de_read(dev_priv, ADLS_DPLL_CFGCR1(id));
+	} else if (IS_DG1(dev_priv)) {
+		hw_state->cfgcr0 = intel_de_read(dev_priv, DG1_DPLL_CFGCR0(id));
+		hw_state->cfgcr1 = intel_de_read(dev_priv, DG1_DPLL_CFGCR1(id));
+	} else if (IS_ROCKETLAKE(dev_priv)) {
+		hw_state->cfgcr0 = intel_de_read(dev_priv,
+						 RKL_DPLL_CFGCR0(id));
+		hw_state->cfgcr1 = intel_de_read(dev_priv,
+						 RKL_DPLL_CFGCR1(id));
+	} else if (DISPLAY_VER(dev_priv) >= 12) {
+		hw_state->cfgcr0 = intel_de_read(dev_priv,
+						 TGL_DPLL_CFGCR0(id));
+		hw_state->cfgcr1 = intel_de_read(dev_priv,
+						 TGL_DPLL_CFGCR1(id));
+		if (dev_priv->display.vbt.override_afc_startup) {
+			hw_state->div0 = intel_de_read(dev_priv, TGL_DPLL0_DIV0(id));
+			hw_state->div0 &= TGL_DPLL0_DIV0_AFC_STARTUP_MASK;
+		}
+	} else {
+		if (IS_JSL_EHL(dev_priv) && id == DPLL_ID_EHL_DPLL4) {
+			hw_state->cfgcr0 = intel_de_read(dev_priv,
+							 ICL_DPLL_CFGCR0(4));
+			hw_state->cfgcr1 = intel_de_read(dev_priv,
+							 ICL_DPLL_CFGCR1(4));
+		} else {
+			hw_state->cfgcr0 = intel_de_read(dev_priv,
+							 ICL_DPLL_CFGCR0(id));
+			hw_state->cfgcr1 = intel_de_read(dev_priv,
+							 ICL_DPLL_CFGCR1(id));
+		}
+	}
+
+	ret = true;
+out:
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+	return ret;
+}
+
+static bool combo_pll_get_hw_state(struct drm_i915_private *dev_priv,
+				   struct intel_shared_dpll *pll,
+				   struct intel_dpll_hw_state *hw_state)
+{
+	i915_reg_t enable_reg = intel_combo_pll_enable_reg(dev_priv, pll);
+
+	return icl_pll_get_hw_state(dev_priv, pll, hw_state, enable_reg);
+}
+
+static bool tbt_pll_get_hw_state(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll,
+				 struct intel_dpll_hw_state *hw_state)
+{
+	return icl_pll_get_hw_state(dev_priv, pll, hw_state, TBT_PLL_ENABLE);
+}
+
+static void icl_dpll_write(struct drm_i915_private *dev_priv,
+			   struct intel_shared_dpll *pll)
+{
+	struct intel_dpll_hw_state *hw_state = &pll->state.hw_state;
+	const enum intel_dpll_id id = pll->info->id;
+	i915_reg_t cfgcr0_reg, cfgcr1_reg, div0_reg = INVALID_MMIO_REG;
+
+	if (IS_ALDERLAKE_S(dev_priv)) {
+		cfgcr0_reg = ADLS_DPLL_CFGCR0(id);
+		cfgcr1_reg = ADLS_DPLL_CFGCR1(id);
+	} else if (IS_DG1(dev_priv)) {
+		cfgcr0_reg = DG1_DPLL_CFGCR0(id);
+		cfgcr1_reg = DG1_DPLL_CFGCR1(id);
+	} else if (IS_ROCKETLAKE(dev_priv)) {
+		cfgcr0_reg = RKL_DPLL_CFGCR0(id);
+		cfgcr1_reg = RKL_DPLL_CFGCR1(id);
+	} else if (DISPLAY_VER(dev_priv) >= 12) {
+		cfgcr0_reg = TGL_DPLL_CFGCR0(id);
+		cfgcr1_reg = TGL_DPLL_CFGCR1(id);
+		div0_reg = TGL_DPLL0_DIV0(id);
+	} else {
+		if (IS_JSL_EHL(dev_priv) && id == DPLL_ID_EHL_DPLL4) {
+			cfgcr0_reg = ICL_DPLL_CFGCR0(4);
+			cfgcr1_reg = ICL_DPLL_CFGCR1(4);
+		} else {
+			cfgcr0_reg = ICL_DPLL_CFGCR0(id);
+			cfgcr1_reg = ICL_DPLL_CFGCR1(id);
+		}
+	}
+
+	intel_de_write(dev_priv, cfgcr0_reg, hw_state->cfgcr0);
+	intel_de_write(dev_priv, cfgcr1_reg, hw_state->cfgcr1);
+	drm_WARN_ON_ONCE(&dev_priv->drm, dev_priv->display.vbt.override_afc_startup &&
+			 !i915_mmio_reg_valid(div0_reg));
+	if (dev_priv->display.vbt.override_afc_startup &&
+	    i915_mmio_reg_valid(div0_reg))
+		intel_de_rmw(dev_priv, div0_reg, TGL_DPLL0_DIV0_AFC_STARTUP_MASK,
+			     hw_state->div0);
+	intel_de_posting_read(dev_priv, cfgcr1_reg);
+}
+
+static void icl_mg_pll_write(struct drm_i915_private *dev_priv,
+			     struct intel_shared_dpll *pll)
+{
+	struct intel_dpll_hw_state *hw_state = &pll->state.hw_state;
+	enum tc_port tc_port = icl_pll_id_to_tc_port(pll->info->id);
+	u32 val;
+
+	/*
+	 * Some of the following registers have reserved fields, so program
+	 * these with RMW based on a mask. The mask can be fixed or generated
+	 * during the calc/readout phase if the mask depends on some other HW
+	 * state like refclk, see icl_calc_mg_pll_state().
+	 */
+	val = intel_de_read(dev_priv, MG_REFCLKIN_CTL(tc_port));
+	val &= ~MG_REFCLKIN_CTL_OD_2_MUX_MASK;
+	val |= hw_state->mg_refclkin_ctl;
+	intel_de_write(dev_priv, MG_REFCLKIN_CTL(tc_port), val);
+
+	val = intel_de_read(dev_priv, MG_CLKTOP2_CORECLKCTL1(tc_port));
+	val &= ~MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO_MASK;
+	val |= hw_state->mg_clktop2_coreclkctl1;
+	intel_de_write(dev_priv, MG_CLKTOP2_CORECLKCTL1(tc_port), val);
+
+	val = intel_de_read(dev_priv, MG_CLKTOP2_HSCLKCTL(tc_port));
+	val &= ~(MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL_MASK |
+		 MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL_MASK |
+		 MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_MASK |
+		 MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_MASK);
+	val |= hw_state->mg_clktop2_hsclkctl;
+	intel_de_write(dev_priv, MG_CLKTOP2_HSCLKCTL(tc_port), val);
+
+	intel_de_write(dev_priv, MG_PLL_DIV0(tc_port), hw_state->mg_pll_div0);
+	intel_de_write(dev_priv, MG_PLL_DIV1(tc_port), hw_state->mg_pll_div1);
+	intel_de_write(dev_priv, MG_PLL_LF(tc_port), hw_state->mg_pll_lf);
+	intel_de_write(dev_priv, MG_PLL_FRAC_LOCK(tc_port),
+		       hw_state->mg_pll_frac_lock);
+	intel_de_write(dev_priv, MG_PLL_SSC(tc_port), hw_state->mg_pll_ssc);
+
+	val = intel_de_read(dev_priv, MG_PLL_BIAS(tc_port));
+	val &= ~hw_state->mg_pll_bias_mask;
+	val |= hw_state->mg_pll_bias;
+	intel_de_write(dev_priv, MG_PLL_BIAS(tc_port), val);
+
+	val = intel_de_read(dev_priv, MG_PLL_TDC_COLDST_BIAS(tc_port));
+	val &= ~hw_state->mg_pll_tdc_coldst_bias_mask;
+	val |= hw_state->mg_pll_tdc_coldst_bias;
+	intel_de_write(dev_priv, MG_PLL_TDC_COLDST_BIAS(tc_port), val);
+
+	intel_de_posting_read(dev_priv, MG_PLL_TDC_COLDST_BIAS(tc_port));
+}
+
+static void dkl_pll_write(struct drm_i915_private *dev_priv,
+			  struct intel_shared_dpll *pll)
+{
+	struct intel_dpll_hw_state *hw_state = &pll->state.hw_state;
+	enum tc_port tc_port = icl_pll_id_to_tc_port(pll->info->id);
+	u32 val;
+
+	/*
+	 * All registers programmed here have the same HIP_INDEX_REG even
+	 * though on different building block
+	 */
+	/* All the registers are RMW */
+	val = intel_dkl_phy_read(dev_priv, DKL_REFCLKIN_CTL(tc_port), 2);
+	val &= ~MG_REFCLKIN_CTL_OD_2_MUX_MASK;
+	val |= hw_state->mg_refclkin_ctl;
+	intel_dkl_phy_write(dev_priv, DKL_REFCLKIN_CTL(tc_port), 2, val);
+
+	val = intel_dkl_phy_read(dev_priv, DKL_CLKTOP2_CORECLKCTL1(tc_port), 2);
+	val &= ~MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO_MASK;
+	val |= hw_state->mg_clktop2_coreclkctl1;
+	intel_dkl_phy_write(dev_priv, DKL_CLKTOP2_CORECLKCTL1(tc_port), 2, val);
+
+	val = intel_dkl_phy_read(dev_priv, DKL_CLKTOP2_HSCLKCTL(tc_port), 2);
+	val &= ~(MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL_MASK |
+		 MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL_MASK |
+		 MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_MASK |
+		 MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_MASK);
+	val |= hw_state->mg_clktop2_hsclkctl;
+	intel_dkl_phy_write(dev_priv, DKL_CLKTOP2_HSCLKCTL(tc_port), 2, val);
+
+	val = DKL_PLL_DIV0_MASK;
+	if (dev_priv->display.vbt.override_afc_startup)
+		val |= DKL_PLL_DIV0_AFC_STARTUP_MASK;
+	intel_dkl_phy_rmw(dev_priv, DKL_PLL_DIV0(tc_port), 2, val,
+			  hw_state->mg_pll_div0);
+
+	val = intel_dkl_phy_read(dev_priv, DKL_PLL_DIV1(tc_port), 2);
+	val &= ~(DKL_PLL_DIV1_IREF_TRIM_MASK |
+		 DKL_PLL_DIV1_TDC_TARGET_CNT_MASK);
+	val |= hw_state->mg_pll_div1;
+	intel_dkl_phy_write(dev_priv, DKL_PLL_DIV1(tc_port), 2, val);
+
+	val = intel_dkl_phy_read(dev_priv, DKL_PLL_SSC(tc_port), 2);
+	val &= ~(DKL_PLL_SSC_IREF_NDIV_RATIO_MASK |
+		 DKL_PLL_SSC_STEP_LEN_MASK |
+		 DKL_PLL_SSC_STEP_NUM_MASK |
+		 DKL_PLL_SSC_EN);
+	val |= hw_state->mg_pll_ssc;
+	intel_dkl_phy_write(dev_priv, DKL_PLL_SSC(tc_port), 2, val);
+
+	val = intel_dkl_phy_read(dev_priv, DKL_PLL_BIAS(tc_port), 2);
+	val &= ~(DKL_PLL_BIAS_FRAC_EN_H |
+		 DKL_PLL_BIAS_FBDIV_FRAC_MASK);
+	val |= hw_state->mg_pll_bias;
+	intel_dkl_phy_write(dev_priv, DKL_PLL_BIAS(tc_port), 2, val);
+
+	val = intel_dkl_phy_read(dev_priv, DKL_PLL_TDC_COLDST_BIAS(tc_port), 2);
+	val &= ~(DKL_PLL_TDC_SSC_STEP_SIZE_MASK |
+		 DKL_PLL_TDC_FEED_FWD_GAIN_MASK);
+	val |= hw_state->mg_pll_tdc_coldst_bias;
+	intel_dkl_phy_write(dev_priv, DKL_PLL_TDC_COLDST_BIAS(tc_port), 2, val);
+
+	intel_dkl_phy_posting_read(dev_priv, DKL_PLL_TDC_COLDST_BIAS(tc_port), 2);
+}
+
+static void icl_pll_power_enable(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll,
+				 i915_reg_t enable_reg)
+{
+	u32 val;
+
+	val = intel_de_read(dev_priv, enable_reg);
+	val |= PLL_POWER_ENABLE;
+	intel_de_write(dev_priv, enable_reg, val);
+
+	/*
+	 * The spec says we need to "wait" but it also says it should be
+	 * immediate.
+	 */
+	if (intel_de_wait_for_set(dev_priv, enable_reg, PLL_POWER_STATE, 1))
+		drm_err(&dev_priv->drm, "PLL %d Power not enabled\n",
+			pll->info->id);
+}
+
+static void icl_pll_enable(struct drm_i915_private *dev_priv,
+			   struct intel_shared_dpll *pll,
+			   i915_reg_t enable_reg)
+{
+	u32 val;
+
+	val = intel_de_read(dev_priv, enable_reg);
+	val |= PLL_ENABLE;
+	intel_de_write(dev_priv, enable_reg, val);
+
+	/* Timeout is actually 600us. */
+	if (intel_de_wait_for_set(dev_priv, enable_reg, PLL_LOCK, 1))
+		drm_err(&dev_priv->drm, "PLL %d not locked\n", pll->info->id);
+}
+
+static void adlp_cmtg_clock_gating_wa(struct drm_i915_private *i915, struct intel_shared_dpll *pll)
+{
+	u32 val;
+
+	if (!IS_ADLP_DISPLAY_STEP(i915, STEP_A0, STEP_B0) ||
+	    pll->info->id != DPLL_ID_ICL_DPLL0)
+		return;
+	/*
+	 * Wa_16011069516:adl-p[a0]
+	 *
+	 * All CMTG regs are unreliable until CMTG clock gating is disabled,
+	 * so we can only assume the default TRANS_CMTG_CHICKEN reg value and
+	 * sanity check this assumption with a double read, which presumably
+	 * returns the correct value even with clock gating on.
+	 *
+	 * Instead of the usual place for workarounds we apply this one here,
+	 * since TRANS_CMTG_CHICKEN is only accessible while DPLL0 is enabled.
+	 */
+	val = intel_de_read(i915, TRANS_CMTG_CHICKEN);
+	val = intel_de_read(i915, TRANS_CMTG_CHICKEN);
+	intel_de_write(i915, TRANS_CMTG_CHICKEN, DISABLE_DPT_CLK_GATING);
+	if (drm_WARN_ON(&i915->drm, val & ~DISABLE_DPT_CLK_GATING))
+		drm_dbg_kms(&i915->drm, "Unexpected flags in TRANS_CMTG_CHICKEN: %08x\n", val);
+}
+
+static void combo_pll_enable(struct drm_i915_private *dev_priv,
+			     struct intel_shared_dpll *pll)
+{
+	i915_reg_t enable_reg = intel_combo_pll_enable_reg(dev_priv, pll);
+
+	if (IS_JSL_EHL(dev_priv) &&
+	    pll->info->id == DPLL_ID_EHL_DPLL4) {
+
+		/*
+		 * We need to disable DC states when this DPLL is enabled.
+		 * This can be done by taking a reference on DPLL4 power
+		 * domain.
+		 */
+		pll->wakeref = intel_display_power_get(dev_priv,
+						       POWER_DOMAIN_DC_OFF);
+	}
+
+	icl_pll_power_enable(dev_priv, pll, enable_reg);
+
+	icl_dpll_write(dev_priv, pll);
+
+	/*
+	 * DVFS pre sequence would be here, but in our driver the cdclk code
+	 * paths should already be setting the appropriate voltage, hence we do
+	 * nothing here.
+	 */
+
+	icl_pll_enable(dev_priv, pll, enable_reg);
+
+	adlp_cmtg_clock_gating_wa(dev_priv, pll);
+
+	/* DVFS post sequence would be here. See the comment above. */
+}
+
+static void tbt_pll_enable(struct drm_i915_private *dev_priv,
+			   struct intel_shared_dpll *pll)
+{
+	icl_pll_power_enable(dev_priv, pll, TBT_PLL_ENABLE);
+
+	icl_dpll_write(dev_priv, pll);
+
+	/*
+	 * DVFS pre sequence would be here, but in our driver the cdclk code
+	 * paths should already be setting the appropriate voltage, hence we do
+	 * nothing here.
+	 */
+
+	icl_pll_enable(dev_priv, pll, TBT_PLL_ENABLE);
+
+	/* DVFS post sequence would be here. See the comment above. */
+}
+
+static void mg_pll_enable(struct drm_i915_private *dev_priv,
+			  struct intel_shared_dpll *pll)
+{
+	i915_reg_t enable_reg = intel_tc_pll_enable_reg(dev_priv, pll);
+
+	icl_pll_power_enable(dev_priv, pll, enable_reg);
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		dkl_pll_write(dev_priv, pll);
+	else
+		icl_mg_pll_write(dev_priv, pll);
+
+	/*
+	 * DVFS pre sequence would be here, but in our driver the cdclk code
+	 * paths should already be setting the appropriate voltage, hence we do
+	 * nothing here.
+	 */
+
+	icl_pll_enable(dev_priv, pll, enable_reg);
+
+	/* DVFS post sequence would be here. See the comment above. */
+}
+
+static void icl_pll_disable(struct drm_i915_private *dev_priv,
+			    struct intel_shared_dpll *pll,
+			    i915_reg_t enable_reg)
+{
+	u32 val;
+
+	/* The first steps are done by intel_ddi_post_disable(). */
+
+	/*
+	 * DVFS pre sequence would be here, but in our driver the cdclk code
+	 * paths should already be setting the appropriate voltage, hence we do
+	 * nothing here.
+	 */
+
+	val = intel_de_read(dev_priv, enable_reg);
+	val &= ~PLL_ENABLE;
+	intel_de_write(dev_priv, enable_reg, val);
+
+	/* Timeout is actually 1us. */
+	if (intel_de_wait_for_clear(dev_priv, enable_reg, PLL_LOCK, 1))
+		drm_err(&dev_priv->drm, "PLL %d locked\n", pll->info->id);
+
+	/* DVFS post sequence would be here. See the comment above. */
+
+	val = intel_de_read(dev_priv, enable_reg);
+	val &= ~PLL_POWER_ENABLE;
+	intel_de_write(dev_priv, enable_reg, val);
+
+	/*
+	 * The spec says we need to "wait" but it also says it should be
+	 * immediate.
+	 */
+	if (intel_de_wait_for_clear(dev_priv, enable_reg, PLL_POWER_STATE, 1))
+		drm_err(&dev_priv->drm, "PLL %d Power not disabled\n",
+			pll->info->id);
+}
+
+static void combo_pll_disable(struct drm_i915_private *dev_priv,
+			      struct intel_shared_dpll *pll)
+{
+	i915_reg_t enable_reg = intel_combo_pll_enable_reg(dev_priv, pll);
+
+	icl_pll_disable(dev_priv, pll, enable_reg);
+
+	if (IS_JSL_EHL(dev_priv) &&
+	    pll->info->id == DPLL_ID_EHL_DPLL4)
+		intel_display_power_put(dev_priv, POWER_DOMAIN_DC_OFF,
+					pll->wakeref);
+}
+
+static void tbt_pll_disable(struct drm_i915_private *dev_priv,
+			    struct intel_shared_dpll *pll)
+{
+	icl_pll_disable(dev_priv, pll, TBT_PLL_ENABLE);
+}
+
+static void mg_pll_disable(struct drm_i915_private *dev_priv,
+			   struct intel_shared_dpll *pll)
+{
+	i915_reg_t enable_reg = intel_tc_pll_enable_reg(dev_priv, pll);
+
+	icl_pll_disable(dev_priv, pll, enable_reg);
+}
+
+static void icl_update_dpll_ref_clks(struct drm_i915_private *i915)
+{
+	/* No SSC ref */
+	i915->display.dpll.ref_clks.nssc = i915->display.cdclk.hw.ref;
+}
+
+static void icl_dump_hw_state(struct drm_i915_private *dev_priv,
+			      const struct intel_dpll_hw_state *hw_state)
+{
+	drm_dbg_kms(&dev_priv->drm,
+		    "dpll_hw_state: cfgcr0: 0x%x, cfgcr1: 0x%x, div0: 0x%x, "
+		    "mg_refclkin_ctl: 0x%x, hg_clktop2_coreclkctl1: 0x%x, "
+		    "mg_clktop2_hsclkctl: 0x%x, mg_pll_div0: 0x%x, "
+		    "mg_pll_div2: 0x%x, mg_pll_lf: 0x%x, "
+		    "mg_pll_frac_lock: 0x%x, mg_pll_ssc: 0x%x, "
+		    "mg_pll_bias: 0x%x, mg_pll_tdc_coldst_bias: 0x%x\n",
+		    hw_state->cfgcr0, hw_state->cfgcr1,
+		    hw_state->div0,
+		    hw_state->mg_refclkin_ctl,
+		    hw_state->mg_clktop2_coreclkctl1,
+		    hw_state->mg_clktop2_hsclkctl,
+		    hw_state->mg_pll_div0,
+		    hw_state->mg_pll_div1,
+		    hw_state->mg_pll_lf,
+		    hw_state->mg_pll_frac_lock,
+		    hw_state->mg_pll_ssc,
+		    hw_state->mg_pll_bias,
+		    hw_state->mg_pll_tdc_coldst_bias);
+}
+
+static const struct intel_shared_dpll_funcs combo_pll_funcs = {
+	.enable = combo_pll_enable,
+	.disable = combo_pll_disable,
+	.get_hw_state = combo_pll_get_hw_state,
+	.get_freq = icl_ddi_combo_pll_get_freq,
+};
+
+static const struct intel_shared_dpll_funcs tbt_pll_funcs = {
+	.enable = tbt_pll_enable,
+	.disable = tbt_pll_disable,
+	.get_hw_state = tbt_pll_get_hw_state,
+	.get_freq = icl_ddi_tbt_pll_get_freq,
+};
+
+static const struct intel_shared_dpll_funcs mg_pll_funcs = {
+	.enable = mg_pll_enable,
+	.disable = mg_pll_disable,
+	.get_hw_state = mg_pll_get_hw_state,
+	.get_freq = icl_ddi_mg_pll_get_freq,
+};
+
+static const struct dpll_info icl_plls[] = {
+	{ "DPLL 0",   &combo_pll_funcs, DPLL_ID_ICL_DPLL0,  0 },
+	{ "DPLL 1",   &combo_pll_funcs, DPLL_ID_ICL_DPLL1,  0 },
+	{ "TBT PLL",  &tbt_pll_funcs, DPLL_ID_ICL_TBTPLL, 0 },
+	{ "MG PLL 1", &mg_pll_funcs, DPLL_ID_ICL_MGPLL1, 0 },
+	{ "MG PLL 2", &mg_pll_funcs, DPLL_ID_ICL_MGPLL2, 0 },
+	{ "MG PLL 3", &mg_pll_funcs, DPLL_ID_ICL_MGPLL3, 0 },
+	{ "MG PLL 4", &mg_pll_funcs, DPLL_ID_ICL_MGPLL4, 0 },
+	{ },
+};
+
+static const struct intel_dpll_mgr icl_pll_mgr = {
+	.dpll_info = icl_plls,
+	.compute_dplls = icl_compute_dplls,
+	.get_dplls = icl_get_dplls,
+	.put_dplls = icl_put_dplls,
+	.update_active_dpll = icl_update_active_dpll,
+	.update_ref_clks = icl_update_dpll_ref_clks,
+	.dump_hw_state = icl_dump_hw_state,
+};
+
+static const struct dpll_info ehl_plls[] = {
+	{ "DPLL 0", &combo_pll_funcs, DPLL_ID_ICL_DPLL0, 0 },
+	{ "DPLL 1", &combo_pll_funcs, DPLL_ID_ICL_DPLL1, 0 },
+	{ "DPLL 4", &combo_pll_funcs, DPLL_ID_EHL_DPLL4, 0 },
+	{ },
+};
+
+static const struct intel_dpll_mgr ehl_pll_mgr = {
+	.dpll_info = ehl_plls,
+	.compute_dplls = icl_compute_dplls,
+	.get_dplls = icl_get_dplls,
+	.put_dplls = icl_put_dplls,
+	.update_ref_clks = icl_update_dpll_ref_clks,
+	.dump_hw_state = icl_dump_hw_state,
+};
+
+static const struct intel_shared_dpll_funcs dkl_pll_funcs = {
+	.enable = mg_pll_enable,
+	.disable = mg_pll_disable,
+	.get_hw_state = dkl_pll_get_hw_state,
+	.get_freq = icl_ddi_mg_pll_get_freq,
+};
+
+static const struct dpll_info tgl_plls[] = {
+	{ "DPLL 0", &combo_pll_funcs, DPLL_ID_ICL_DPLL0,  0 },
+	{ "DPLL 1", &combo_pll_funcs, DPLL_ID_ICL_DPLL1,  0 },
+	{ "TBT PLL",  &tbt_pll_funcs, DPLL_ID_ICL_TBTPLL, 0 },
+	{ "TC PLL 1", &dkl_pll_funcs, DPLL_ID_ICL_MGPLL1, 0 },
+	{ "TC PLL 2", &dkl_pll_funcs, DPLL_ID_ICL_MGPLL2, 0 },
+	{ "TC PLL 3", &dkl_pll_funcs, DPLL_ID_ICL_MGPLL3, 0 },
+	{ "TC PLL 4", &dkl_pll_funcs, DPLL_ID_ICL_MGPLL4, 0 },
+	{ "TC PLL 5", &dkl_pll_funcs, DPLL_ID_TGL_MGPLL5, 0 },
+	{ "TC PLL 6", &dkl_pll_funcs, DPLL_ID_TGL_MGPLL6, 0 },
+	{ },
+};
+
+static const struct intel_dpll_mgr tgl_pll_mgr = {
+	.dpll_info = tgl_plls,
+	.compute_dplls = icl_compute_dplls,
+	.get_dplls = icl_get_dplls,
+	.put_dplls = icl_put_dplls,
+	.update_active_dpll = icl_update_active_dpll,
+	.update_ref_clks = icl_update_dpll_ref_clks,
+	.dump_hw_state = icl_dump_hw_state,
+};
+
+static const struct dpll_info rkl_plls[] = {
+	{ "DPLL 0", &combo_pll_funcs, DPLL_ID_ICL_DPLL0, 0 },
+	{ "DPLL 1", &combo_pll_funcs, DPLL_ID_ICL_DPLL1, 0 },
+	{ "DPLL 4", &combo_pll_funcs, DPLL_ID_EHL_DPLL4, 0 },
+	{ },
+};
+
+static const struct intel_dpll_mgr rkl_pll_mgr = {
+	.dpll_info = rkl_plls,
+	.compute_dplls = icl_compute_dplls,
+	.get_dplls = icl_get_dplls,
+	.put_dplls = icl_put_dplls,
+	.update_ref_clks = icl_update_dpll_ref_clks,
+	.dump_hw_state = icl_dump_hw_state,
+};
+
+static const struct dpll_info dg1_plls[] = {
+	{ "DPLL 0", &combo_pll_funcs, DPLL_ID_DG1_DPLL0, 0 },
+	{ "DPLL 1", &combo_pll_funcs, DPLL_ID_DG1_DPLL1, 0 },
+	{ "DPLL 2", &combo_pll_funcs, DPLL_ID_DG1_DPLL2, 0 },
+	{ "DPLL 3", &combo_pll_funcs, DPLL_ID_DG1_DPLL3, 0 },
+	{ },
+};
+
+static const struct intel_dpll_mgr dg1_pll_mgr = {
+	.dpll_info = dg1_plls,
+	.compute_dplls = icl_compute_dplls,
+	.get_dplls = icl_get_dplls,
+	.put_dplls = icl_put_dplls,
+	.update_ref_clks = icl_update_dpll_ref_clks,
+	.dump_hw_state = icl_dump_hw_state,
+};
+
+static const struct dpll_info adls_plls[] = {
+	{ "DPLL 0", &combo_pll_funcs, DPLL_ID_ICL_DPLL0, 0 },
+	{ "DPLL 1", &combo_pll_funcs, DPLL_ID_ICL_DPLL1, 0 },
+	{ "DPLL 2", &combo_pll_funcs, DPLL_ID_DG1_DPLL2, 0 },
+	{ "DPLL 3", &combo_pll_funcs, DPLL_ID_DG1_DPLL3, 0 },
+	{ },
+};
+
+static const struct intel_dpll_mgr adls_pll_mgr = {
+	.dpll_info = adls_plls,
+	.compute_dplls = icl_compute_dplls,
+	.get_dplls = icl_get_dplls,
+	.put_dplls = icl_put_dplls,
+	.update_ref_clks = icl_update_dpll_ref_clks,
+	.dump_hw_state = icl_dump_hw_state,
+};
+
+static const struct dpll_info adlp_plls[] = {
+	{ "DPLL 0", &combo_pll_funcs, DPLL_ID_ICL_DPLL0,  0 },
+	{ "DPLL 1", &combo_pll_funcs, DPLL_ID_ICL_DPLL1,  0 },
+	{ "TBT PLL",  &tbt_pll_funcs, DPLL_ID_ICL_TBTPLL, 0 },
+	{ "TC PLL 1", &dkl_pll_funcs, DPLL_ID_ICL_MGPLL1, 0 },
+	{ "TC PLL 2", &dkl_pll_funcs, DPLL_ID_ICL_MGPLL2, 0 },
+	{ "TC PLL 3", &dkl_pll_funcs, DPLL_ID_ICL_MGPLL3, 0 },
+	{ "TC PLL 4", &dkl_pll_funcs, DPLL_ID_ICL_MGPLL4, 0 },
+	{ },
+};
+
+static const struct intel_dpll_mgr adlp_pll_mgr = {
+	.dpll_info = adlp_plls,
+	.compute_dplls = icl_compute_dplls,
+	.get_dplls = icl_get_dplls,
+	.put_dplls = icl_put_dplls,
+	.update_active_dpll = icl_update_active_dpll,
+	.update_ref_clks = icl_update_dpll_ref_clks,
+	.dump_hw_state = icl_dump_hw_state,
+};
+
+/**
+ * intel_shared_dpll_init - Initialize shared DPLLs
+ * @dev_priv: i915 device
+ *
+ * Initialize shared DPLLs for @dev_priv.
+ */
+void intel_shared_dpll_init(struct drm_i915_private *dev_priv)
+{
+	const struct intel_dpll_mgr *dpll_mgr = NULL;
+	const struct dpll_info *dpll_info;
+	int i;
+
+	if (IS_DG2(dev_priv))
+		/* No shared DPLLs on DG2; port PLLs are part of the PHY */
+		dpll_mgr = NULL;
+	else if (IS_ALDERLAKE_P(dev_priv))
+		dpll_mgr = &adlp_pll_mgr;
+	else if (IS_ALDERLAKE_S(dev_priv))
+		dpll_mgr = &adls_pll_mgr;
+	else if (IS_DG1(dev_priv))
+		dpll_mgr = &dg1_pll_mgr;
+	else if (IS_ROCKETLAKE(dev_priv))
+		dpll_mgr = &rkl_pll_mgr;
+	else if (DISPLAY_VER(dev_priv) >= 12)
+		dpll_mgr = &tgl_pll_mgr;
+	else if (IS_JSL_EHL(dev_priv))
+		dpll_mgr = &ehl_pll_mgr;
+	else if (DISPLAY_VER(dev_priv) >= 11)
+		dpll_mgr = &icl_pll_mgr;
+	else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		dpll_mgr = &bxt_pll_mgr;
+	else if (DISPLAY_VER(dev_priv) == 9)
+		dpll_mgr = &skl_pll_mgr;
+	else if (HAS_DDI(dev_priv))
+		dpll_mgr = &hsw_pll_mgr;
+	else if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv))
+		dpll_mgr = &pch_pll_mgr;
+
+	if (!dpll_mgr) {
+		dev_priv->display.dpll.num_shared_dpll = 0;
+		return;
+	}
+
+	dpll_info = dpll_mgr->dpll_info;
+
+	for (i = 0; dpll_info[i].name; i++) {
+		if (drm_WARN_ON(&dev_priv->drm,
+				i >= ARRAY_SIZE(dev_priv->display.dpll.shared_dplls)))
+			break;
+
+		drm_WARN_ON(&dev_priv->drm, i != dpll_info[i].id);
+		dev_priv->display.dpll.shared_dplls[i].info = &dpll_info[i];
+	}
+
+	dev_priv->display.dpll.mgr = dpll_mgr;
+	dev_priv->display.dpll.num_shared_dpll = i;
+	mutex_init(&dev_priv->display.dpll.lock);
+}
+
+/**
+ * intel_compute_shared_dplls - compute DPLL state CRTC and encoder combination
+ * @state: atomic state
+ * @crtc: CRTC to compute DPLLs for
+ * @encoder: encoder
+ *
+ * This function computes the DPLL state for the given CRTC and encoder.
+ *
+ * The new configuration in the atomic commit @state is made effective by
+ * calling intel_shared_dpll_swap_state().
+ *
+ * Returns:
+ * 0 on success, negative error code on falure.
+ */
+int intel_compute_shared_dplls(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc,
+			       struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_dpll_mgr *dpll_mgr = dev_priv->display.dpll.mgr;
+
+	if (drm_WARN_ON(&dev_priv->drm, !dpll_mgr))
+		return -EINVAL;
+
+	return dpll_mgr->compute_dplls(state, crtc, encoder);
+}
+
+/**
+ * intel_reserve_shared_dplls - reserve DPLLs for CRTC and encoder combination
+ * @state: atomic state
+ * @crtc: CRTC to reserve DPLLs for
+ * @encoder: encoder
+ *
+ * This function reserves all required DPLLs for the given CRTC and encoder
+ * combination in the current atomic commit @state and the new @crtc atomic
+ * state.
+ *
+ * The new configuration in the atomic commit @state is made effective by
+ * calling intel_shared_dpll_swap_state().
+ *
+ * The reserved DPLLs should be released by calling
+ * intel_release_shared_dplls().
+ *
+ * Returns:
+ * 0 if all required DPLLs were successfully reserved,
+ * negative error code otherwise.
+ */
+int intel_reserve_shared_dplls(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc,
+			       struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_dpll_mgr *dpll_mgr = dev_priv->display.dpll.mgr;
+
+	if (drm_WARN_ON(&dev_priv->drm, !dpll_mgr))
+		return -EINVAL;
+
+	return dpll_mgr->get_dplls(state, crtc, encoder);
+}
+
+/**
+ * intel_release_shared_dplls - end use of DPLLs by CRTC in atomic state
+ * @state: atomic state
+ * @crtc: crtc from which the DPLLs are to be released
+ *
+ * This function releases all DPLLs reserved by intel_reserve_shared_dplls()
+ * from the current atomic commit @state and the old @crtc atomic state.
+ *
+ * The new configuration in the atomic commit @state is made effective by
+ * calling intel_shared_dpll_swap_state().
+ */
+void intel_release_shared_dplls(struct intel_atomic_state *state,
+				struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	const struct intel_dpll_mgr *dpll_mgr = dev_priv->display.dpll.mgr;
+
+	/*
+	 * FIXME: this function is called for every platform having a
+	 * compute_clock hook, even though the platform doesn't yet support
+	 * the shared DPLL framework and intel_reserve_shared_dplls() is not
+	 * called on those.
+	 */
+	if (!dpll_mgr)
+		return;
+
+	dpll_mgr->put_dplls(state, crtc);
+}
+
+/**
+ * intel_update_active_dpll - update the active DPLL for a CRTC/encoder
+ * @state: atomic state
+ * @crtc: the CRTC for which to update the active DPLL
+ * @encoder: encoder determining the type of port DPLL
+ *
+ * Update the active DPLL for the given @crtc/@encoder in @crtc's atomic state,
+ * from the port DPLLs reserved previously by intel_reserve_shared_dplls(). The
+ * DPLL selected will be based on the current mode of the encoder's port.
+ */
+void intel_update_active_dpll(struct intel_atomic_state *state,
+			      struct intel_crtc *crtc,
+			      struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	const struct intel_dpll_mgr *dpll_mgr = dev_priv->display.dpll.mgr;
+
+	if (drm_WARN_ON(&dev_priv->drm, !dpll_mgr))
+		return;
+
+	dpll_mgr->update_active_dpll(state, crtc, encoder);
+}
+
+/**
+ * intel_dpll_get_freq - calculate the DPLL's output frequency
+ * @i915: i915 device
+ * @pll: DPLL for which to calculate the output frequency
+ * @pll_state: DPLL state from which to calculate the output frequency
+ *
+ * Return the output frequency corresponding to @pll's passed in @pll_state.
+ */
+int intel_dpll_get_freq(struct drm_i915_private *i915,
+			const struct intel_shared_dpll *pll,
+			const struct intel_dpll_hw_state *pll_state)
+{
+	if (drm_WARN_ON(&i915->drm, !pll->info->funcs->get_freq))
+		return 0;
+
+	return pll->info->funcs->get_freq(i915, pll, pll_state);
+}
+
+/**
+ * intel_dpll_get_hw_state - readout the DPLL's hardware state
+ * @i915: i915 device
+ * @pll: DPLL for which to calculate the output frequency
+ * @hw_state: DPLL's hardware state
+ *
+ * Read out @pll's hardware state into @hw_state.
+ */
+bool intel_dpll_get_hw_state(struct drm_i915_private *i915,
+			     struct intel_shared_dpll *pll,
+			     struct intel_dpll_hw_state *hw_state)
+{
+	return pll->info->funcs->get_hw_state(i915, pll, hw_state);
+}
+
+static void readout_dpll_hw_state(struct drm_i915_private *i915,
+				  struct intel_shared_dpll *pll)
+{
+	struct intel_crtc *crtc;
+
+	pll->on = intel_dpll_get_hw_state(i915, pll, &pll->state.hw_state);
+
+	if (IS_JSL_EHL(i915) && pll->on &&
+	    pll->info->id == DPLL_ID_EHL_DPLL4) {
+		pll->wakeref = intel_display_power_get(i915,
+						       POWER_DOMAIN_DC_OFF);
+	}
+
+	pll->state.pipe_mask = 0;
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		if (crtc_state->hw.active && crtc_state->shared_dpll == pll)
+			pll->state.pipe_mask |= BIT(crtc->pipe);
+	}
+	pll->active_mask = pll->state.pipe_mask;
+
+	drm_dbg_kms(&i915->drm,
+		    "%s hw state readout: pipe_mask 0x%x, on %i\n",
+		    pll->info->name, pll->state.pipe_mask, pll->on);
+}
+
+void intel_dpll_update_ref_clks(struct drm_i915_private *i915)
+{
+	if (i915->display.dpll.mgr && i915->display.dpll.mgr->update_ref_clks)
+		i915->display.dpll.mgr->update_ref_clks(i915);
+}
+
+void intel_dpll_readout_hw_state(struct drm_i915_private *i915)
+{
+	int i;
+
+	for (i = 0; i < i915->display.dpll.num_shared_dpll; i++)
+		readout_dpll_hw_state(i915, &i915->display.dpll.shared_dplls[i]);
+}
+
+static void sanitize_dpll_state(struct drm_i915_private *i915,
+				struct intel_shared_dpll *pll)
+{
+	if (!pll->on)
+		return;
+
+	adlp_cmtg_clock_gating_wa(i915, pll);
+
+	if (pll->active_mask)
+		return;
+
+	drm_dbg_kms(&i915->drm,
+		    "%s enabled but not in use, disabling\n",
+		    pll->info->name);
+
+	pll->info->funcs->disable(i915, pll);
+	pll->on = false;
+}
+
+void intel_dpll_sanitize_state(struct drm_i915_private *i915)
+{
+	int i;
+
+	for (i = 0; i < i915->display.dpll.num_shared_dpll; i++)
+		sanitize_dpll_state(i915, &i915->display.dpll.shared_dplls[i]);
+}
+
+/**
+ * intel_dpll_dump_hw_state - write hw_state to dmesg
+ * @dev_priv: i915 drm device
+ * @hw_state: hw state to be written to the log
+ *
+ * Write the relevant values in @hw_state to dmesg using drm_dbg_kms.
+ */
+void intel_dpll_dump_hw_state(struct drm_i915_private *dev_priv,
+			      const struct intel_dpll_hw_state *hw_state)
+{
+	if (dev_priv->display.dpll.mgr) {
+		dev_priv->display.dpll.mgr->dump_hw_state(dev_priv, hw_state);
+	} else {
+		/* fallback for platforms that don't use the shared dpll
+		 * infrastructure
+		 */
+		drm_dbg_kms(&dev_priv->drm,
+			    "dpll_hw_state: dpll: 0x%x, dpll_md: 0x%x, "
+			    "fp0: 0x%x, fp1: 0x%x\n",
+			    hw_state->dpll,
+			    hw_state->dpll_md,
+			    hw_state->fp0,
+			    hw_state->fp1);
+	}
+}
+
+static void
+verify_single_dpll_state(struct drm_i915_private *dev_priv,
+			 struct intel_shared_dpll *pll,
+			 struct intel_crtc *crtc,
+			 struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_dpll_hw_state dpll_hw_state;
+	u8 pipe_mask;
+	bool active;
+
+	memset(&dpll_hw_state, 0, sizeof(dpll_hw_state));
+
+	drm_dbg_kms(&dev_priv->drm, "%s\n", pll->info->name);
+
+	active = intel_dpll_get_hw_state(dev_priv, pll, &dpll_hw_state);
+
+	if (!(pll->info->flags & INTEL_DPLL_ALWAYS_ON)) {
+		I915_STATE_WARN(!pll->on && pll->active_mask,
+				"pll in active use but not on in sw tracking\n");
+		I915_STATE_WARN(pll->on && !pll->active_mask,
+				"pll is on but not used by any active pipe\n");
+		I915_STATE_WARN(pll->on != active,
+				"pll on state mismatch (expected %i, found %i)\n",
+				pll->on, active);
+	}
+
+	if (!crtc) {
+		I915_STATE_WARN(pll->active_mask & ~pll->state.pipe_mask,
+				"more active pll users than references: 0x%x vs 0x%x\n",
+				pll->active_mask, pll->state.pipe_mask);
+
+		return;
+	}
+
+	pipe_mask = BIT(crtc->pipe);
+
+	if (new_crtc_state->hw.active)
+		I915_STATE_WARN(!(pll->active_mask & pipe_mask),
+				"pll active mismatch (expected pipe %c in active mask 0x%x)\n",
+				pipe_name(crtc->pipe), pll->active_mask);
+	else
+		I915_STATE_WARN(pll->active_mask & pipe_mask,
+				"pll active mismatch (didn't expect pipe %c in active mask 0x%x)\n",
+				pipe_name(crtc->pipe), pll->active_mask);
+
+	I915_STATE_WARN(!(pll->state.pipe_mask & pipe_mask),
+			"pll enabled crtcs mismatch (expected 0x%x in 0x%x)\n",
+			pipe_mask, pll->state.pipe_mask);
+
+	I915_STATE_WARN(pll->on && memcmp(&pll->state.hw_state,
+					  &dpll_hw_state,
+					  sizeof(dpll_hw_state)),
+			"pll hw state mismatch\n");
+}
+
+void intel_shared_dpll_state_verify(struct intel_crtc *crtc,
+				    struct intel_crtc_state *old_crtc_state,
+				    struct intel_crtc_state *new_crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	if (new_crtc_state->shared_dpll)
+		verify_single_dpll_state(dev_priv, new_crtc_state->shared_dpll,
+					 crtc, new_crtc_state);
+
+	if (old_crtc_state->shared_dpll &&
+	    old_crtc_state->shared_dpll != new_crtc_state->shared_dpll) {
+		u8 pipe_mask = BIT(crtc->pipe);
+		struct intel_shared_dpll *pll = old_crtc_state->shared_dpll;
+
+		I915_STATE_WARN(pll->active_mask & pipe_mask,
+				"pll active mismatch (didn't expect pipe %c in active mask (0x%x))\n",
+				pipe_name(crtc->pipe), pll->active_mask);
+		I915_STATE_WARN(pll->state.pipe_mask & pipe_mask,
+				"pll enabled crtcs mismatch (found %x in enabled mask (0x%x))\n",
+				pipe_name(crtc->pipe), pll->state.pipe_mask);
+	}
+}
+
+void intel_shared_dpll_verify_disabled(struct drm_i915_private *i915)
+{
+	int i;
+
+	for (i = 0; i < i915->display.dpll.num_shared_dpll; i++)
+		verify_single_dpll_state(i915, &i915->display.dpll.shared_dplls[i],
+					 NULL, NULL);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dpll_mgr.h b/drivers/gpu/drm/i915/display/intel_dpll_mgr.h
new file mode 100644
index 000000000..3247dc300
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dpll_mgr.h
@@ -0,0 +1,376 @@
+/*
+ * Copyright © 2012-2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _INTEL_DPLL_MGR_H_
+#define _INTEL_DPLL_MGR_H_
+
+#include <linux/types.h>
+
+#include "intel_wakeref.h"
+
+/*FIXME: Move this to a more appropriate place. */
+#define abs_diff(a, b) ({			\
+	typeof(a) __a = (a);			\
+	typeof(b) __b = (b);			\
+	(void) (&__a == &__b);			\
+	__a > __b ? (__a - __b) : (__b - __a); })
+
+enum tc_port;
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_encoder;
+struct intel_shared_dpll;
+struct intel_shared_dpll_funcs;
+
+/**
+ * enum intel_dpll_id - possible DPLL ids
+ *
+ * Enumeration of possible IDs for a DPLL. Real shared dpll ids must be >= 0.
+ */
+enum intel_dpll_id {
+	/**
+	 * @DPLL_ID_PRIVATE: non-shared dpll in use
+	 */
+	DPLL_ID_PRIVATE = -1,
+
+	/**
+	 * @DPLL_ID_PCH_PLL_A: DPLL A in ILK, SNB and IVB
+	 */
+	DPLL_ID_PCH_PLL_A = 0,
+	/**
+	 * @DPLL_ID_PCH_PLL_B: DPLL B in ILK, SNB and IVB
+	 */
+	DPLL_ID_PCH_PLL_B = 1,
+
+
+	/**
+	 * @DPLL_ID_WRPLL1: HSW and BDW WRPLL1
+	 */
+	DPLL_ID_WRPLL1 = 0,
+	/**
+	 * @DPLL_ID_WRPLL2: HSW and BDW WRPLL2
+	 */
+	DPLL_ID_WRPLL2 = 1,
+	/**
+	 * @DPLL_ID_SPLL: HSW and BDW SPLL
+	 */
+	DPLL_ID_SPLL = 2,
+	/**
+	 * @DPLL_ID_LCPLL_810: HSW and BDW 0.81 GHz LCPLL
+	 */
+	DPLL_ID_LCPLL_810 = 3,
+	/**
+	 * @DPLL_ID_LCPLL_1350: HSW and BDW 1.35 GHz LCPLL
+	 */
+	DPLL_ID_LCPLL_1350 = 4,
+	/**
+	 * @DPLL_ID_LCPLL_2700: HSW and BDW 2.7 GHz LCPLL
+	 */
+	DPLL_ID_LCPLL_2700 = 5,
+
+
+	/**
+	 * @DPLL_ID_SKL_DPLL0: SKL and later DPLL0
+	 */
+	DPLL_ID_SKL_DPLL0 = 0,
+	/**
+	 * @DPLL_ID_SKL_DPLL1: SKL and later DPLL1
+	 */
+	DPLL_ID_SKL_DPLL1 = 1,
+	/**
+	 * @DPLL_ID_SKL_DPLL2: SKL and later DPLL2
+	 */
+	DPLL_ID_SKL_DPLL2 = 2,
+	/**
+	 * @DPLL_ID_SKL_DPLL3: SKL and later DPLL3
+	 */
+	DPLL_ID_SKL_DPLL3 = 3,
+
+
+	/**
+	 * @DPLL_ID_ICL_DPLL0: ICL/TGL combo PHY DPLL0
+	 */
+	DPLL_ID_ICL_DPLL0 = 0,
+	/**
+	 * @DPLL_ID_ICL_DPLL1: ICL/TGL combo PHY DPLL1
+	 */
+	DPLL_ID_ICL_DPLL1 = 1,
+	/**
+	 * @DPLL_ID_EHL_DPLL4: EHL combo PHY DPLL4
+	 */
+	DPLL_ID_EHL_DPLL4 = 2,
+	/**
+	 * @DPLL_ID_ICL_TBTPLL: ICL/TGL TBT PLL
+	 */
+	DPLL_ID_ICL_TBTPLL = 2,
+	/**
+	 * @DPLL_ID_ICL_MGPLL1: ICL MG PLL 1 port 1 (C),
+	 *                      TGL TC PLL 1 port 1 (TC1)
+	 */
+	DPLL_ID_ICL_MGPLL1 = 3,
+	/**
+	 * @DPLL_ID_ICL_MGPLL2: ICL MG PLL 1 port 2 (D)
+	 *                      TGL TC PLL 1 port 2 (TC2)
+	 */
+	DPLL_ID_ICL_MGPLL2 = 4,
+	/**
+	 * @DPLL_ID_ICL_MGPLL3: ICL MG PLL 1 port 3 (E)
+	 *                      TGL TC PLL 1 port 3 (TC3)
+	 */
+	DPLL_ID_ICL_MGPLL3 = 5,
+	/**
+	 * @DPLL_ID_ICL_MGPLL4: ICL MG PLL 1 port 4 (F)
+	 *                      TGL TC PLL 1 port 4 (TC4)
+	 */
+	DPLL_ID_ICL_MGPLL4 = 6,
+	/**
+	 * @DPLL_ID_TGL_MGPLL5: TGL TC PLL port 5 (TC5)
+	 */
+	DPLL_ID_TGL_MGPLL5 = 7,
+	/**
+	 * @DPLL_ID_TGL_MGPLL6: TGL TC PLL port 6 (TC6)
+	 */
+	DPLL_ID_TGL_MGPLL6 = 8,
+
+	/**
+	 * @DPLL_ID_DG1_DPLL0: DG1 combo PHY DPLL0
+	 */
+	DPLL_ID_DG1_DPLL0 = 0,
+	/**
+	 * @DPLL_ID_DG1_DPLL1: DG1 combo PHY DPLL1
+	 */
+	DPLL_ID_DG1_DPLL1 = 1,
+	/**
+	 * @DPLL_ID_DG1_DPLL2: DG1 combo PHY DPLL2
+	 */
+	DPLL_ID_DG1_DPLL2 = 2,
+	/**
+	 * @DPLL_ID_DG1_DPLL3: DG1 combo PHY DPLL3
+	 */
+	DPLL_ID_DG1_DPLL3 = 3,
+};
+
+#define I915_NUM_PLLS 9
+
+enum icl_port_dpll_id {
+	ICL_PORT_DPLL_DEFAULT,
+	ICL_PORT_DPLL_MG_PHY,
+
+	ICL_PORT_DPLL_COUNT,
+};
+
+struct intel_dpll_hw_state {
+	/* i9xx, pch plls */
+	u32 dpll;
+	u32 dpll_md;
+	u32 fp0;
+	u32 fp1;
+
+	/* hsw, bdw */
+	u32 wrpll;
+	u32 spll;
+
+	/* skl */
+	/*
+	 * DPLL_CTRL1 has 6 bits for each each this DPLL. We store those in
+	 * lower part of ctrl1 and they get shifted into position when writing
+	 * the register.  This allows us to easily compare the state to share
+	 * the DPLL.
+	 */
+	u32 ctrl1;
+	/* HDMI only, 0 when used for DP */
+	u32 cfgcr1, cfgcr2;
+
+	/* icl */
+	u32 cfgcr0;
+
+	/* tgl */
+	u32 div0;
+
+	/* bxt */
+	u32 ebb0, ebb4, pll0, pll1, pll2, pll3, pll6, pll8, pll9, pll10, pcsdw12;
+
+	/*
+	 * ICL uses the following, already defined:
+	 * u32 cfgcr0, cfgcr1;
+	 */
+	u32 mg_refclkin_ctl;
+	u32 mg_clktop2_coreclkctl1;
+	u32 mg_clktop2_hsclkctl;
+	u32 mg_pll_div0;
+	u32 mg_pll_div1;
+	u32 mg_pll_lf;
+	u32 mg_pll_frac_lock;
+	u32 mg_pll_ssc;
+	u32 mg_pll_bias;
+	u32 mg_pll_tdc_coldst_bias;
+	u32 mg_pll_bias_mask;
+	u32 mg_pll_tdc_coldst_bias_mask;
+};
+
+/**
+ * struct intel_shared_dpll_state - hold the DPLL atomic state
+ *
+ * This structure holds an atomic state for the DPLL, that can represent
+ * either its current state (in struct &intel_shared_dpll) or a desired
+ * future state which would be applied by an atomic mode set (stored in
+ * a struct &intel_atomic_state).
+ *
+ * See also intel_reserve_shared_dplls() and intel_release_shared_dplls().
+ */
+struct intel_shared_dpll_state {
+	/**
+	 * @pipe_mask: mask of pipes using this DPLL, active or not
+	 */
+	u8 pipe_mask;
+
+	/**
+	 * @hw_state: hardware configuration for the DPLL stored in
+	 * struct &intel_dpll_hw_state.
+	 */
+	struct intel_dpll_hw_state hw_state;
+};
+
+/**
+ * struct dpll_info - display PLL platform specific info
+ */
+struct dpll_info {
+	/**
+	 * @name: DPLL name; used for logging
+	 */
+	const char *name;
+
+	/**
+	 * @funcs: platform specific hooks
+	 */
+	const struct intel_shared_dpll_funcs *funcs;
+
+	/**
+	 * @id: unique indentifier for this DPLL; should match the index in the
+	 * dev_priv->shared_dplls array
+	 */
+	enum intel_dpll_id id;
+
+#define INTEL_DPLL_ALWAYS_ON	(1 << 0)
+	/**
+	 * @flags:
+	 *
+	 * INTEL_DPLL_ALWAYS_ON
+	 *     Inform the state checker that the DPLL is kept enabled even if
+	 *     not in use by any CRTC.
+	 */
+	u32 flags;
+};
+
+/**
+ * struct intel_shared_dpll - display PLL with tracked state and users
+ */
+struct intel_shared_dpll {
+	/**
+	 * @state:
+	 *
+	 * Store the state for the pll, including its hw state
+	 * and CRTCs using it.
+	 */
+	struct intel_shared_dpll_state state;
+
+	/**
+	 * @active_mask: mask of active pipes (i.e. DPMS on) using this DPLL
+	 */
+	u8 active_mask;
+
+	/**
+	 * @on: is the PLL actually active? Disabled during modeset
+	 */
+	bool on;
+
+	/**
+	 * @info: platform specific info
+	 */
+	const struct dpll_info *info;
+
+	/**
+	 * @wakeref: In some platforms a device-level runtime pm reference may
+	 * need to be grabbed to disable DC states while this DPLL is enabled
+	 */
+	intel_wakeref_t wakeref;
+};
+
+#define SKL_DPLL0 0
+#define SKL_DPLL1 1
+#define SKL_DPLL2 2
+#define SKL_DPLL3 3
+
+/* shared dpll functions */
+struct intel_shared_dpll *
+intel_get_shared_dpll_by_id(struct drm_i915_private *dev_priv,
+			    enum intel_dpll_id id);
+enum intel_dpll_id
+intel_get_shared_dpll_id(struct drm_i915_private *dev_priv,
+			 struct intel_shared_dpll *pll);
+void assert_shared_dpll(struct drm_i915_private *dev_priv,
+			struct intel_shared_dpll *pll,
+			bool state);
+#define assert_shared_dpll_enabled(d, p) assert_shared_dpll(d, p, true)
+#define assert_shared_dpll_disabled(d, p) assert_shared_dpll(d, p, false)
+int intel_compute_shared_dplls(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc,
+			       struct intel_encoder *encoder);
+int intel_reserve_shared_dplls(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc,
+			       struct intel_encoder *encoder);
+void intel_release_shared_dplls(struct intel_atomic_state *state,
+				struct intel_crtc *crtc);
+void icl_set_active_port_dpll(struct intel_crtc_state *crtc_state,
+			      enum icl_port_dpll_id port_dpll_id);
+void intel_update_active_dpll(struct intel_atomic_state *state,
+			      struct intel_crtc *crtc,
+			      struct intel_encoder *encoder);
+int intel_dpll_get_freq(struct drm_i915_private *i915,
+			const struct intel_shared_dpll *pll,
+			const struct intel_dpll_hw_state *pll_state);
+bool intel_dpll_get_hw_state(struct drm_i915_private *i915,
+			     struct intel_shared_dpll *pll,
+			     struct intel_dpll_hw_state *hw_state);
+void intel_enable_shared_dpll(const struct intel_crtc_state *crtc_state);
+void intel_disable_shared_dpll(const struct intel_crtc_state *crtc_state);
+void intel_shared_dpll_swap_state(struct intel_atomic_state *state);
+void intel_shared_dpll_init(struct drm_i915_private *dev_priv);
+void intel_dpll_update_ref_clks(struct drm_i915_private *dev_priv);
+void intel_dpll_readout_hw_state(struct drm_i915_private *dev_priv);
+void intel_dpll_sanitize_state(struct drm_i915_private *dev_priv);
+
+void intel_dpll_dump_hw_state(struct drm_i915_private *dev_priv,
+			      const struct intel_dpll_hw_state *hw_state);
+enum intel_dpll_id icl_tc_port_to_pll_id(enum tc_port tc_port);
+bool intel_dpll_is_combophy(enum intel_dpll_id id);
+
+void intel_shared_dpll_state_verify(struct intel_crtc *crtc,
+				    struct intel_crtc_state *old_crtc_state,
+				    struct intel_crtc_state *new_crtc_state);
+void intel_shared_dpll_verify_disabled(struct drm_i915_private *i915);
+
+#endif /* _INTEL_DPLL_MGR_H_ */
diff --git a/drivers/gpu/drm/i915/display/intel_dpt.c b/drivers/gpu/drm/i915/display/intel_dpt.c
new file mode 100644
index 000000000..ea8a08b9c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dpt.c
@@ -0,0 +1,316 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include "gem/i915_gem_domain.h"
+#include "gem/i915_gem_internal.h"
+#include "gt/gen8_ppgtt.h"
+
+#include "i915_drv.h"
+#include "intel_display_types.h"
+#include "intel_dpt.h"
+#include "intel_fb.h"
+
+struct i915_dpt {
+	struct i915_address_space vm;
+
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	void __iomem *iomem;
+};
+
+#define i915_is_dpt(vm) ((vm)->is_dpt)
+
+static inline struct i915_dpt *
+i915_vm_to_dpt(struct i915_address_space *vm)
+{
+	BUILD_BUG_ON(offsetof(struct i915_dpt, vm));
+	GEM_BUG_ON(!i915_is_dpt(vm));
+	return container_of(vm, struct i915_dpt, vm);
+}
+
+#define dpt_total_entries(dpt) ((dpt)->vm.total >> PAGE_SHIFT)
+
+static void gen8_set_pte(void __iomem *addr, gen8_pte_t pte)
+{
+	writeq(pte, addr);
+}
+
+static void dpt_insert_page(struct i915_address_space *vm,
+			    dma_addr_t addr,
+			    u64 offset,
+			    enum i915_cache_level level,
+			    u32 flags)
+{
+	struct i915_dpt *dpt = i915_vm_to_dpt(vm);
+	gen8_pte_t __iomem *base = dpt->iomem;
+
+	gen8_set_pte(base + offset / I915_GTT_PAGE_SIZE,
+		     vm->pte_encode(addr, level, flags));
+}
+
+static void dpt_insert_entries(struct i915_address_space *vm,
+			       struct i915_vma_resource *vma_res,
+			       enum i915_cache_level level,
+			       u32 flags)
+{
+	struct i915_dpt *dpt = i915_vm_to_dpt(vm);
+	gen8_pte_t __iomem *base = dpt->iomem;
+	const gen8_pte_t pte_encode = vm->pte_encode(0, level, flags);
+	struct sgt_iter sgt_iter;
+	dma_addr_t addr;
+	int i;
+
+	/*
+	 * Note that we ignore PTE_READ_ONLY here. The caller must be careful
+	 * not to allow the user to override access to a read only page.
+	 */
+
+	i = vma_res->start / I915_GTT_PAGE_SIZE;
+	for_each_sgt_daddr(addr, sgt_iter, vma_res->bi.pages)
+		gen8_set_pte(&base[i++], pte_encode | addr);
+}
+
+static void dpt_clear_range(struct i915_address_space *vm,
+			    u64 start, u64 length)
+{
+}
+
+static void dpt_bind_vma(struct i915_address_space *vm,
+			 struct i915_vm_pt_stash *stash,
+			 struct i915_vma_resource *vma_res,
+			 enum i915_cache_level cache_level,
+			 u32 flags)
+{
+	u32 pte_flags;
+
+	if (vma_res->bound_flags)
+		return;
+
+	/* Applicable to VLV (gen8+ do not support RO in the GGTT) */
+	pte_flags = 0;
+	if (vm->has_read_only && vma_res->bi.readonly)
+		pte_flags |= PTE_READ_ONLY;
+	if (vma_res->bi.lmem)
+		pte_flags |= PTE_LM;
+
+	vm->insert_entries(vm, vma_res, cache_level, pte_flags);
+
+	vma_res->page_sizes_gtt = I915_GTT_PAGE_SIZE;
+
+	/*
+	 * Without aliasing PPGTT there's no difference between
+	 * GLOBAL/LOCAL_BIND, it's all the same ptes. Hence unconditionally
+	 * upgrade to both bound if we bind either to avoid double-binding.
+	 */
+	vma_res->bound_flags = I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND;
+}
+
+static void dpt_unbind_vma(struct i915_address_space *vm,
+			   struct i915_vma_resource *vma_res)
+{
+	vm->clear_range(vm, vma_res->start, vma_res->vma_size);
+}
+
+static void dpt_cleanup(struct i915_address_space *vm)
+{
+	struct i915_dpt *dpt = i915_vm_to_dpt(vm);
+
+	i915_gem_object_put(dpt->obj);
+}
+
+struct i915_vma *intel_dpt_pin(struct i915_address_space *vm)
+{
+	struct drm_i915_private *i915 = vm->i915;
+	struct i915_dpt *dpt = i915_vm_to_dpt(vm);
+	intel_wakeref_t wakeref;
+	struct i915_vma *vma;
+	void __iomem *iomem;
+	struct i915_gem_ww_ctx ww;
+	u64 pin_flags = 0;
+	int err;
+
+	if (i915_gem_object_is_stolen(dpt->obj))
+		pin_flags |= PIN_MAPPABLE;
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+	atomic_inc(&i915->gpu_error.pending_fb_pin);
+
+	for_i915_gem_ww(&ww, err, true) {
+		err = i915_gem_object_lock(dpt->obj, &ww);
+		if (err)
+			continue;
+
+		vma = i915_gem_object_ggtt_pin_ww(dpt->obj, &ww, NULL, 0, 4096,
+						  pin_flags);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			continue;
+		}
+
+		iomem = i915_vma_pin_iomap(vma);
+		i915_vma_unpin(vma);
+
+		if (IS_ERR(iomem)) {
+			err = PTR_ERR(iomem);
+			continue;
+		}
+
+		dpt->vma = vma;
+		dpt->iomem = iomem;
+
+		i915_vma_get(vma);
+	}
+
+	dpt->obj->mm.dirty = true;
+
+	atomic_dec(&i915->gpu_error.pending_fb_pin);
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+
+	return err ? ERR_PTR(err) : vma;
+}
+
+void intel_dpt_unpin(struct i915_address_space *vm)
+{
+	struct i915_dpt *dpt = i915_vm_to_dpt(vm);
+
+	i915_vma_unpin_iomap(dpt->vma);
+	i915_vma_put(dpt->vma);
+}
+
+/**
+ * intel_dpt_resume - restore the memory mapping for all DPT FBs during system resume
+ * @i915: device instance
+ *
+ * Restore the memory mapping during system resume for all framebuffers which
+ * are mapped to HW via a GGTT->DPT page table. The content of these page
+ * tables are not stored in the hibernation image during S4 and S3RST->S4
+ * transitions, so here we reprogram the PTE entries in those tables.
+ *
+ * This function must be called after the mappings in GGTT have been restored calling
+ * i915_ggtt_resume().
+ */
+void intel_dpt_resume(struct drm_i915_private *i915)
+{
+	struct drm_framebuffer *drm_fb;
+
+	if (!HAS_DISPLAY(i915))
+		return;
+
+	mutex_lock(&i915->drm.mode_config.fb_lock);
+	drm_for_each_fb(drm_fb, &i915->drm) {
+		struct intel_framebuffer *fb = to_intel_framebuffer(drm_fb);
+
+		if (fb->dpt_vm)
+			i915_ggtt_resume_vm(fb->dpt_vm);
+	}
+	mutex_unlock(&i915->drm.mode_config.fb_lock);
+}
+
+/**
+ * intel_dpt_suspend - suspend the memory mapping for all DPT FBs during system suspend
+ * @i915: device instance
+ *
+ * Suspend the memory mapping during system suspend for all framebuffers which
+ * are mapped to HW via a GGTT->DPT page table.
+ *
+ * This function must be called before the mappings in GGTT are suspended calling
+ * i915_ggtt_suspend().
+ */
+void intel_dpt_suspend(struct drm_i915_private *i915)
+{
+	struct drm_framebuffer *drm_fb;
+
+	if (!HAS_DISPLAY(i915))
+		return;
+
+	mutex_lock(&i915->drm.mode_config.fb_lock);
+
+	drm_for_each_fb(drm_fb, &i915->drm) {
+		struct intel_framebuffer *fb = to_intel_framebuffer(drm_fb);
+
+		if (fb->dpt_vm)
+			i915_ggtt_suspend_vm(fb->dpt_vm);
+	}
+
+	mutex_unlock(&i915->drm.mode_config.fb_lock);
+}
+
+struct i915_address_space *
+intel_dpt_create(struct intel_framebuffer *fb)
+{
+	struct drm_gem_object *obj = &intel_fb_obj(&fb->base)->base;
+	struct drm_i915_private *i915 = to_i915(obj->dev);
+	struct drm_i915_gem_object *dpt_obj;
+	struct i915_address_space *vm;
+	struct i915_dpt *dpt;
+	size_t size;
+	int ret;
+
+	if (intel_fb_needs_pot_stride_remap(fb))
+		size = intel_remapped_info_size(&fb->remapped_view.gtt.remapped);
+	else
+		size = DIV_ROUND_UP_ULL(obj->size, I915_GTT_PAGE_SIZE);
+
+	size = round_up(size * sizeof(gen8_pte_t), I915_GTT_PAGE_SIZE);
+
+	dpt_obj = i915_gem_object_create_lmem(i915, size, I915_BO_ALLOC_CONTIGUOUS);
+	if (IS_ERR(dpt_obj) && i915_ggtt_has_aperture(to_gt(i915)->ggtt))
+		dpt_obj = i915_gem_object_create_stolen(i915, size);
+	if (IS_ERR(dpt_obj) && !HAS_LMEM(i915)) {
+		drm_dbg_kms(&i915->drm, "Allocating dpt from smem\n");
+		dpt_obj = i915_gem_object_create_shmem(i915, size);
+	}
+	if (IS_ERR(dpt_obj))
+		return ERR_CAST(dpt_obj);
+
+	ret = i915_gem_object_lock_interruptible(dpt_obj, NULL);
+	if (!ret) {
+		ret = i915_gem_object_set_cache_level(dpt_obj, I915_CACHE_NONE);
+		i915_gem_object_unlock(dpt_obj);
+	}
+	if (ret) {
+		i915_gem_object_put(dpt_obj);
+		return ERR_PTR(ret);
+	}
+
+	dpt = kzalloc(sizeof(*dpt), GFP_KERNEL);
+	if (!dpt) {
+		i915_gem_object_put(dpt_obj);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	vm = &dpt->vm;
+
+	vm->gt = to_gt(i915);
+	vm->i915 = i915;
+	vm->dma = i915->drm.dev;
+	vm->total = (size / sizeof(gen8_pte_t)) * I915_GTT_PAGE_SIZE;
+	vm->is_dpt = true;
+
+	i915_address_space_init(vm, VM_CLASS_DPT);
+
+	vm->insert_page = dpt_insert_page;
+	vm->clear_range = dpt_clear_range;
+	vm->insert_entries = dpt_insert_entries;
+	vm->cleanup = dpt_cleanup;
+
+	vm->vma_ops.bind_vma    = dpt_bind_vma;
+	vm->vma_ops.unbind_vma  = dpt_unbind_vma;
+
+	vm->pte_encode = gen8_ggtt_pte_encode;
+
+	dpt->obj = dpt_obj;
+	dpt->obj->is_dpt = true;
+
+	return &dpt->vm;
+}
+
+void intel_dpt_destroy(struct i915_address_space *vm)
+{
+	struct i915_dpt *dpt = i915_vm_to_dpt(vm);
+
+	dpt->obj->is_dpt = false;
+	i915_vm_put(&dpt->vm);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dpt.h b/drivers/gpu/drm/i915/display/intel_dpt.h
new file mode 100644
index 000000000..e18a9f767
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dpt.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __INTEL_DPT_H__
+#define __INTEL_DPT_H__
+
+struct drm_i915_private;
+
+struct i915_address_space;
+struct i915_vma;
+struct intel_framebuffer;
+
+void intel_dpt_destroy(struct i915_address_space *vm);
+struct i915_vma *intel_dpt_pin(struct i915_address_space *vm);
+void intel_dpt_unpin(struct i915_address_space *vm);
+void intel_dpt_suspend(struct drm_i915_private *i915);
+void intel_dpt_resume(struct drm_i915_private *i915);
+struct i915_address_space *
+intel_dpt_create(struct intel_framebuffer *fb);
+
+#endif /* __INTEL_DPT_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_drrs.c b/drivers/gpu/drm/i915/display/intel_drrs.c
new file mode 100644
index 000000000..7da4a9cbe
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_drrs.c
@@ -0,0 +1,299 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_drrs.h"
+#include "intel_panel.h"
+
+/**
+ * DOC: Display Refresh Rate Switching (DRRS)
+ *
+ * Display Refresh Rate Switching (DRRS) is a power conservation feature
+ * which enables swtching between low and high refresh rates,
+ * dynamically, based on the usage scenario. This feature is applicable
+ * for internal panels.
+ *
+ * Indication that the panel supports DRRS is given by the panel EDID, which
+ * would list multiple refresh rates for one resolution.
+ *
+ * DRRS is of 2 types - static and seamless.
+ * Static DRRS involves changing refresh rate (RR) by doing a full modeset
+ * (may appear as a blink on screen) and is used in dock-undock scenario.
+ * Seamless DRRS involves changing RR without any visual effect to the user
+ * and can be used during normal system usage. This is done by programming
+ * certain registers.
+ *
+ * Support for static/seamless DRRS may be indicated in the VBT based on
+ * inputs from the panel spec.
+ *
+ * DRRS saves power by switching to low RR based on usage scenarios.
+ *
+ * The implementation is based on frontbuffer tracking implementation.  When
+ * there is a disturbance on the screen triggered by user activity or a periodic
+ * system activity, DRRS is disabled (RR is changed to high RR).  When there is
+ * no movement on screen, after a timeout of 1 second, a switch to low RR is
+ * made.
+ *
+ * For integration with frontbuffer tracking code, intel_drrs_invalidate()
+ * and intel_drrs_flush() are called.
+ *
+ * DRRS can be further extended to support other internal panels and also
+ * the scenario of video playback wherein RR is set based on the rate
+ * requested by userspace.
+ */
+
+const char *intel_drrs_type_str(enum drrs_type drrs_type)
+{
+	static const char * const str[] = {
+		[DRRS_TYPE_NONE] = "none",
+		[DRRS_TYPE_STATIC] = "static",
+		[DRRS_TYPE_SEAMLESS] = "seamless",
+	};
+
+	if (drrs_type >= ARRAY_SIZE(str))
+		return "<invalid>";
+
+	return str[drrs_type];
+}
+
+static void
+intel_drrs_set_refresh_rate_pipeconf(struct intel_crtc *crtc,
+				     enum drrs_refresh_rate refresh_rate)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc->drrs.cpu_transcoder;
+	u32 val, bit;
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		bit = PIPECONF_REFRESH_RATE_ALT_VLV;
+	else
+		bit = PIPECONF_REFRESH_RATE_ALT_ILK;
+
+	val = intel_de_read(dev_priv, PIPECONF(cpu_transcoder));
+
+	if (refresh_rate == DRRS_REFRESH_RATE_LOW)
+		val |= bit;
+	else
+		val &= ~bit;
+
+	intel_de_write(dev_priv, PIPECONF(cpu_transcoder), val);
+}
+
+static void
+intel_drrs_set_refresh_rate_m_n(struct intel_crtc *crtc,
+				enum drrs_refresh_rate refresh_rate)
+{
+	intel_cpu_transcoder_set_m1_n1(crtc, crtc->drrs.cpu_transcoder,
+				       refresh_rate == DRRS_REFRESH_RATE_LOW ?
+				       &crtc->drrs.m2_n2 : &crtc->drrs.m_n);
+}
+
+bool intel_drrs_is_active(struct intel_crtc *crtc)
+{
+	return crtc->drrs.cpu_transcoder != INVALID_TRANSCODER;
+}
+
+static void intel_drrs_set_state(struct intel_crtc *crtc,
+				 enum drrs_refresh_rate refresh_rate)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	if (refresh_rate == crtc->drrs.refresh_rate)
+		return;
+
+	if (intel_cpu_transcoder_has_m2_n2(dev_priv, crtc->drrs.cpu_transcoder))
+		intel_drrs_set_refresh_rate_pipeconf(crtc, refresh_rate);
+	else
+		intel_drrs_set_refresh_rate_m_n(crtc, refresh_rate);
+
+	crtc->drrs.refresh_rate = refresh_rate;
+}
+
+static void intel_drrs_schedule_work(struct intel_crtc *crtc)
+{
+	mod_delayed_work(system_wq, &crtc->drrs.work, msecs_to_jiffies(1000));
+}
+
+static unsigned int intel_drrs_frontbuffer_bits(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	unsigned int frontbuffer_bits;
+
+	frontbuffer_bits = INTEL_FRONTBUFFER_ALL_MASK(crtc->pipe);
+
+	for_each_intel_crtc_in_pipe_mask(&i915->drm, crtc,
+					 crtc_state->bigjoiner_pipes)
+		frontbuffer_bits |= INTEL_FRONTBUFFER_ALL_MASK(crtc->pipe);
+
+	return frontbuffer_bits;
+}
+
+/**
+ * intel_drrs_activate - activate DRRS
+ * @crtc_state: the crtc state
+ *
+ * Activates DRRS on the crtc.
+ */
+void intel_drrs_activate(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	if (!crtc_state->has_drrs)
+		return;
+
+	if (!crtc_state->hw.active)
+		return;
+
+	if (intel_crtc_is_bigjoiner_slave(crtc_state))
+		return;
+
+	mutex_lock(&crtc->drrs.mutex);
+
+	crtc->drrs.cpu_transcoder = crtc_state->cpu_transcoder;
+	crtc->drrs.m_n = crtc_state->dp_m_n;
+	crtc->drrs.m2_n2 = crtc_state->dp_m2_n2;
+	crtc->drrs.frontbuffer_bits = intel_drrs_frontbuffer_bits(crtc_state);
+	crtc->drrs.busy_frontbuffer_bits = 0;
+
+	intel_drrs_schedule_work(crtc);
+
+	mutex_unlock(&crtc->drrs.mutex);
+}
+
+/**
+ * intel_drrs_deactivate - deactivate DRRS
+ * @old_crtc_state: the old crtc state
+ *
+ * Deactivates DRRS on the crtc.
+ */
+void intel_drrs_deactivate(const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+
+	if (!old_crtc_state->has_drrs)
+		return;
+
+	if (!old_crtc_state->hw.active)
+		return;
+
+	if (intel_crtc_is_bigjoiner_slave(old_crtc_state))
+		return;
+
+	mutex_lock(&crtc->drrs.mutex);
+
+	if (intel_drrs_is_active(crtc))
+		intel_drrs_set_state(crtc, DRRS_REFRESH_RATE_HIGH);
+
+	crtc->drrs.cpu_transcoder = INVALID_TRANSCODER;
+	crtc->drrs.frontbuffer_bits = 0;
+	crtc->drrs.busy_frontbuffer_bits = 0;
+
+	mutex_unlock(&crtc->drrs.mutex);
+
+	cancel_delayed_work_sync(&crtc->drrs.work);
+}
+
+static void intel_drrs_downclock_work(struct work_struct *work)
+{
+	struct intel_crtc *crtc = container_of(work, typeof(*crtc), drrs.work.work);
+
+	mutex_lock(&crtc->drrs.mutex);
+
+	if (intel_drrs_is_active(crtc) && !crtc->drrs.busy_frontbuffer_bits)
+		intel_drrs_set_state(crtc, DRRS_REFRESH_RATE_LOW);
+
+	mutex_unlock(&crtc->drrs.mutex);
+}
+
+static void intel_drrs_frontbuffer_update(struct drm_i915_private *dev_priv,
+					  unsigned int all_frontbuffer_bits,
+					  bool invalidate)
+{
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		unsigned int frontbuffer_bits;
+
+		mutex_lock(&crtc->drrs.mutex);
+
+		frontbuffer_bits = all_frontbuffer_bits & crtc->drrs.frontbuffer_bits;
+		if (!frontbuffer_bits) {
+			mutex_unlock(&crtc->drrs.mutex);
+			continue;
+		}
+
+		if (invalidate)
+			crtc->drrs.busy_frontbuffer_bits |= frontbuffer_bits;
+		else
+			crtc->drrs.busy_frontbuffer_bits &= ~frontbuffer_bits;
+
+		/* flush/invalidate means busy screen hence upclock */
+		intel_drrs_set_state(crtc, DRRS_REFRESH_RATE_HIGH);
+
+		/*
+		 * flush also means no more activity hence schedule downclock, if all
+		 * other fbs are quiescent too
+		 */
+		if (!crtc->drrs.busy_frontbuffer_bits)
+			intel_drrs_schedule_work(crtc);
+		else
+			cancel_delayed_work(&crtc->drrs.work);
+
+		mutex_unlock(&crtc->drrs.mutex);
+	}
+}
+
+/**
+ * intel_drrs_invalidate - Disable Idleness DRRS
+ * @dev_priv: i915 device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called everytime rendering on the given planes start.
+ * Hence DRRS needs to be Upclocked, i.e. (LOW_RR -> HIGH_RR).
+ *
+ * Dirty frontbuffers relevant to DRRS are tracked in busy_frontbuffer_bits.
+ */
+void intel_drrs_invalidate(struct drm_i915_private *dev_priv,
+			   unsigned int frontbuffer_bits)
+{
+	intel_drrs_frontbuffer_update(dev_priv, frontbuffer_bits, true);
+}
+
+/**
+ * intel_drrs_flush - Restart Idleness DRRS
+ * @dev_priv: i915 device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called every time rendering on the given planes has
+ * completed or flip on a crtc is completed. So DRRS should be upclocked
+ * (LOW_RR -> HIGH_RR). And also Idleness detection should be started again,
+ * if no other planes are dirty.
+ *
+ * Dirty frontbuffers relevant to DRRS are tracked in busy_frontbuffer_bits.
+ */
+void intel_drrs_flush(struct drm_i915_private *dev_priv,
+		      unsigned int frontbuffer_bits)
+{
+	intel_drrs_frontbuffer_update(dev_priv, frontbuffer_bits, false);
+}
+
+/**
+ * intel_crtc_drrs_init - Init DRRS for CRTC
+ * @crtc: crtc
+ *
+ * This function is called only once at driver load to initialize basic
+ * DRRS stuff.
+ *
+ */
+void intel_crtc_drrs_init(struct intel_crtc *crtc)
+{
+	INIT_DELAYED_WORK(&crtc->drrs.work, intel_drrs_downclock_work);
+	mutex_init(&crtc->drrs.mutex);
+	crtc->drrs.cpu_transcoder = INVALID_TRANSCODER;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_drrs.h b/drivers/gpu/drm/i915/display/intel_drrs.h
new file mode 100644
index 000000000..3ad1be1ad
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_drrs.h
@@ -0,0 +1,28 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __INTEL_DRRS_H__
+#define __INTEL_DRRS_H__
+
+#include <linux/types.h>
+
+enum drrs_type;
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_connector;
+
+const char *intel_drrs_type_str(enum drrs_type drrs_type);
+bool intel_drrs_is_active(struct intel_crtc *crtc);
+void intel_drrs_activate(const struct intel_crtc_state *crtc_state);
+void intel_drrs_deactivate(const struct intel_crtc_state *crtc_state);
+void intel_drrs_invalidate(struct drm_i915_private *dev_priv,
+			   unsigned int frontbuffer_bits);
+void intel_drrs_flush(struct drm_i915_private *dev_priv,
+		      unsigned int frontbuffer_bits);
+void intel_crtc_drrs_init(struct intel_crtc *crtc);
+
+#endif /* __INTEL_DRRS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dsb.c b/drivers/gpu/drm/i915/display/intel_dsb.c
new file mode 100644
index 000000000..fc9c3e41c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dsb.c
@@ -0,0 +1,363 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ *
+ */
+
+#include "gem/i915_gem_internal.h"
+
+#include "i915_drv.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dsb.h"
+
+struct i915_vma;
+
+enum dsb_id {
+	INVALID_DSB = -1,
+	DSB1,
+	DSB2,
+	DSB3,
+	MAX_DSB_PER_PIPE
+};
+
+struct intel_dsb {
+	enum dsb_id id;
+	u32 *cmd_buf;
+	struct i915_vma *vma;
+
+	/*
+	 * free_pos will point the first free entry position
+	 * and help in calculating tail of command buffer.
+	 */
+	int free_pos;
+
+	/*
+	 * ins_start_offset will help to store start address of the dsb
+	 * instuction and help in identifying the batch of auto-increment
+	 * register.
+	 */
+	u32 ins_start_offset;
+};
+
+#define DSB_BUF_SIZE    (2 * PAGE_SIZE)
+
+/**
+ * DOC: DSB
+ *
+ * A DSB (Display State Buffer) is a queue of MMIO instructions in the memory
+ * which can be offloaded to DSB HW in Display Controller. DSB HW is a DMA
+ * engine that can be programmed to download the DSB from memory.
+ * It allows driver to batch submit display HW programming. This helps to
+ * reduce loading time and CPU activity, thereby making the context switch
+ * faster. DSB Support added from Gen12 Intel graphics based platform.
+ *
+ * DSB's can access only the pipe, plane, and transcoder Data Island Packet
+ * registers.
+ *
+ * DSB HW can support only register writes (both indexed and direct MMIO
+ * writes). There are no registers reads possible with DSB HW engine.
+ */
+
+/* DSB opcodes. */
+#define DSB_OPCODE_SHIFT		24
+#define DSB_OPCODE_MMIO_WRITE		0x1
+#define DSB_OPCODE_INDEXED_WRITE	0x9
+#define DSB_BYTE_EN			0xF
+#define DSB_BYTE_EN_SHIFT		20
+#define DSB_REG_VALUE_MASK		0xfffff
+
+static bool is_dsb_busy(struct drm_i915_private *i915, enum pipe pipe,
+			enum dsb_id id)
+{
+	return DSB_STATUS & intel_de_read(i915, DSB_CTRL(pipe, id));
+}
+
+static bool intel_dsb_enable_engine(struct drm_i915_private *i915,
+				    enum pipe pipe, enum dsb_id id)
+{
+	u32 dsb_ctrl;
+
+	dsb_ctrl = intel_de_read(i915, DSB_CTRL(pipe, id));
+	if (DSB_STATUS & dsb_ctrl) {
+		drm_dbg_kms(&i915->drm, "DSB engine is busy.\n");
+		return false;
+	}
+
+	dsb_ctrl |= DSB_ENABLE;
+	intel_de_write(i915, DSB_CTRL(pipe, id), dsb_ctrl);
+
+	intel_de_posting_read(i915, DSB_CTRL(pipe, id));
+	return true;
+}
+
+static bool intel_dsb_disable_engine(struct drm_i915_private *i915,
+				     enum pipe pipe, enum dsb_id id)
+{
+	u32 dsb_ctrl;
+
+	dsb_ctrl = intel_de_read(i915, DSB_CTRL(pipe, id));
+	if (DSB_STATUS & dsb_ctrl) {
+		drm_dbg_kms(&i915->drm, "DSB engine is busy.\n");
+		return false;
+	}
+
+	dsb_ctrl &= ~DSB_ENABLE;
+	intel_de_write(i915, DSB_CTRL(pipe, id), dsb_ctrl);
+
+	intel_de_posting_read(i915, DSB_CTRL(pipe, id));
+	return true;
+}
+
+/**
+ * intel_dsb_indexed_reg_write() -Write to the DSB context for auto
+ * increment register.
+ * @crtc_state: intel_crtc_state structure
+ * @reg: register address.
+ * @val: value.
+ *
+ * This function is used for writing register-value pair in command
+ * buffer of DSB for auto-increment register. During command buffer overflow,
+ * a warning is thrown and rest all erroneous condition register programming
+ * is done through mmio write.
+ */
+
+void intel_dsb_indexed_reg_write(const struct intel_crtc_state *crtc_state,
+				 i915_reg_t reg, u32 val)
+{
+	struct intel_dsb *dsb = crtc_state->dsb;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 *buf;
+	u32 reg_val;
+
+	if (!dsb) {
+		intel_de_write_fw(dev_priv, reg, val);
+		return;
+	}
+	buf = dsb->cmd_buf;
+	if (drm_WARN_ON(&dev_priv->drm, dsb->free_pos >= DSB_BUF_SIZE)) {
+		drm_dbg_kms(&dev_priv->drm, "DSB buffer overflow\n");
+		return;
+	}
+
+	/*
+	 * For example the buffer will look like below for 3 dwords for auto
+	 * increment register:
+	 * +--------------------------------------------------------+
+	 * | size = 3 | offset &| value1 | value2 | value3 | zero   |
+	 * |          | opcode  |        |        |        |        |
+	 * +--------------------------------------------------------+
+	 * +          +         +        +        +        +        +
+	 * 0          4         8        12       16       20       24
+	 * Byte
+	 *
+	 * As every instruction is 8 byte aligned the index of dsb instruction
+	 * will start always from even number while dealing with u32 array. If
+	 * we are writing odd no of dwords, Zeros will be added in the end for
+	 * padding.
+	 */
+	reg_val = buf[dsb->ins_start_offset + 1] & DSB_REG_VALUE_MASK;
+	if (reg_val != i915_mmio_reg_offset(reg)) {
+		/* Every instruction should be 8 byte aligned. */
+		dsb->free_pos = ALIGN(dsb->free_pos, 2);
+
+		dsb->ins_start_offset = dsb->free_pos;
+
+		/* Update the size. */
+		buf[dsb->free_pos++] = 1;
+
+		/* Update the opcode and reg. */
+		buf[dsb->free_pos++] = (DSB_OPCODE_INDEXED_WRITE  <<
+					DSB_OPCODE_SHIFT) |
+					i915_mmio_reg_offset(reg);
+
+		/* Update the value. */
+		buf[dsb->free_pos++] = val;
+	} else {
+		/* Update the new value. */
+		buf[dsb->free_pos++] = val;
+
+		/* Update the size. */
+		buf[dsb->ins_start_offset]++;
+	}
+
+	/* if number of data words is odd, then the last dword should be 0.*/
+	if (dsb->free_pos & 0x1)
+		buf[dsb->free_pos] = 0;
+}
+
+/**
+ * intel_dsb_reg_write() -Write to the DSB context for normal
+ * register.
+ * @crtc_state: intel_crtc_state structure
+ * @reg: register address.
+ * @val: value.
+ *
+ * This function is used for writing register-value pair in command
+ * buffer of DSB. During command buffer overflow, a warning  is thrown
+ * and rest all erroneous condition register programming is done
+ * through mmio write.
+ */
+void intel_dsb_reg_write(const struct intel_crtc_state *crtc_state,
+			 i915_reg_t reg, u32 val)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_dsb *dsb;
+	u32 *buf;
+
+	dsb = crtc_state->dsb;
+	if (!dsb) {
+		intel_de_write_fw(dev_priv, reg, val);
+		return;
+	}
+
+	buf = dsb->cmd_buf;
+	if (drm_WARN_ON(&dev_priv->drm, dsb->free_pos >= DSB_BUF_SIZE)) {
+		drm_dbg_kms(&dev_priv->drm, "DSB buffer overflow\n");
+		return;
+	}
+
+	dsb->ins_start_offset = dsb->free_pos;
+	buf[dsb->free_pos++] = val;
+	buf[dsb->free_pos++] = (DSB_OPCODE_MMIO_WRITE  << DSB_OPCODE_SHIFT) |
+			       (DSB_BYTE_EN << DSB_BYTE_EN_SHIFT) |
+			       i915_mmio_reg_offset(reg);
+}
+
+/**
+ * intel_dsb_commit() - Trigger workload execution of DSB.
+ * @crtc_state: intel_crtc_state structure
+ *
+ * This function is used to do actual write to hardware using DSB.
+ * On errors, fall back to MMIO. Also this function help to reset the context.
+ */
+void intel_dsb_commit(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_dsb *dsb = crtc_state->dsb;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	u32 tail;
+
+	if (!(dsb && dsb->free_pos))
+		return;
+
+	if (!intel_dsb_enable_engine(dev_priv, pipe, dsb->id))
+		goto reset;
+
+	if (is_dsb_busy(dev_priv, pipe, dsb->id)) {
+		drm_err(&dev_priv->drm,
+			"HEAD_PTR write failed - dsb engine is busy.\n");
+		goto reset;
+	}
+	intel_de_write(dev_priv, DSB_HEAD(pipe, dsb->id),
+		       i915_ggtt_offset(dsb->vma));
+
+	tail = ALIGN(dsb->free_pos * 4, CACHELINE_BYTES);
+	if (tail > dsb->free_pos * 4)
+		memset(&dsb->cmd_buf[dsb->free_pos], 0,
+		       (tail - dsb->free_pos * 4));
+
+	if (is_dsb_busy(dev_priv, pipe, dsb->id)) {
+		drm_err(&dev_priv->drm,
+			"TAIL_PTR write failed - dsb engine is busy.\n");
+		goto reset;
+	}
+	drm_dbg_kms(&dev_priv->drm,
+		    "DSB execution started - head 0x%x, tail 0x%x\n",
+		    i915_ggtt_offset(dsb->vma), tail);
+	intel_de_write(dev_priv, DSB_TAIL(pipe, dsb->id),
+		       i915_ggtt_offset(dsb->vma) + tail);
+	if (wait_for(!is_dsb_busy(dev_priv, pipe, dsb->id), 1)) {
+		drm_err(&dev_priv->drm,
+			"Timed out waiting for DSB workload completion.\n");
+		goto reset;
+	}
+
+reset:
+	dsb->free_pos = 0;
+	dsb->ins_start_offset = 0;
+	intel_dsb_disable_engine(dev_priv, pipe, dsb->id);
+}
+
+/**
+ * intel_dsb_prepare() - Allocate, pin and map the DSB command buffer.
+ * @crtc_state: intel_crtc_state structure to prepare associated dsb instance.
+ *
+ * This function prepare the command buffer which is used to store dsb
+ * instructions with data.
+ */
+void intel_dsb_prepare(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_dsb *dsb;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u32 *buf;
+	intel_wakeref_t wakeref;
+
+	if (!HAS_DSB(i915))
+		return;
+
+	dsb = kmalloc(sizeof(*dsb), GFP_KERNEL);
+	if (!dsb) {
+		drm_err(&i915->drm, "DSB object creation failed\n");
+		return;
+	}
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	obj = i915_gem_object_create_internal(i915, DSB_BUF_SIZE);
+	if (IS_ERR(obj)) {
+		kfree(dsb);
+		goto out;
+	}
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, 0);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		kfree(dsb);
+		goto out;
+	}
+
+	buf = i915_gem_object_pin_map_unlocked(vma->obj, I915_MAP_WC);
+	if (IS_ERR(buf)) {
+		i915_vma_unpin_and_release(&vma, I915_VMA_RELEASE_MAP);
+		kfree(dsb);
+		goto out;
+	}
+
+	dsb->id = DSB1;
+	dsb->vma = vma;
+	dsb->cmd_buf = buf;
+	dsb->free_pos = 0;
+	dsb->ins_start_offset = 0;
+	crtc_state->dsb = dsb;
+out:
+	if (!crtc_state->dsb)
+		drm_info(&i915->drm,
+			 "DSB queue setup failed, will fallback to MMIO for display HW programming\n");
+
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+}
+
+/**
+ * intel_dsb_cleanup() - To cleanup DSB context.
+ * @crtc_state: intel_crtc_state structure to cleanup associated dsb instance.
+ *
+ * This function cleanup the DSB context by unpinning and releasing
+ * the VMA object associated with it.
+ */
+void intel_dsb_cleanup(struct intel_crtc_state *crtc_state)
+{
+	if (!crtc_state->dsb)
+		return;
+
+	i915_vma_unpin_and_release(&crtc_state->dsb->vma, I915_VMA_RELEASE_MAP);
+	kfree(crtc_state->dsb);
+	crtc_state->dsb = NULL;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dsb.h b/drivers/gpu/drm/i915/display/intel_dsb.h
new file mode 100644
index 000000000..74dd2b334
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dsb.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef _INTEL_DSB_H
+#define _INTEL_DSB_H
+
+#include <linux/types.h>
+
+#include "i915_reg_defs.h"
+
+struct intel_crtc_state;
+
+void intel_dsb_prepare(struct intel_crtc_state *crtc_state);
+void intel_dsb_cleanup(struct intel_crtc_state *crtc_state);
+void intel_dsb_reg_write(const struct intel_crtc_state *crtc_state,
+			 i915_reg_t reg, u32 val);
+void intel_dsb_indexed_reg_write(const struct intel_crtc_state *crtc_state,
+				 i915_reg_t reg, u32 val);
+void intel_dsb_commit(const struct intel_crtc_state *crtc_state);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_dsi.c b/drivers/gpu/drm/i915/display/intel_dsi.c
new file mode 100644
index 000000000..5efdd471a
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dsi.c
@@ -0,0 +1,114 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <drm/drm_mipi_dsi.h>
+
+#include "i915_drv.h"
+#include "intel_dsi.h"
+#include "intel_panel.h"
+
+int intel_dsi_bitrate(const struct intel_dsi *intel_dsi)
+{
+	int bpp = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format);
+
+	if (WARN_ON(bpp < 0))
+		bpp = 16;
+
+	return intel_dsi->pclk * bpp / intel_dsi->lane_count;
+}
+
+int intel_dsi_tlpx_ns(const struct intel_dsi *intel_dsi)
+{
+	switch (intel_dsi->escape_clk_div) {
+	default:
+	case 0:
+		return 50;
+	case 1:
+		return 100;
+	case 2:
+		return 200;
+	}
+}
+
+int intel_dsi_get_modes(struct drm_connector *connector)
+{
+	return intel_panel_get_modes(to_intel_connector(connector));
+}
+
+enum drm_mode_status intel_dsi_mode_valid(struct drm_connector *connector,
+					  struct drm_display_mode *mode)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_fixed_mode(intel_connector, mode);
+	int max_dotclk = to_i915(connector->dev)->max_dotclk_freq;
+	enum drm_mode_status status;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return MODE_NO_DBLESCAN;
+
+	status = intel_panel_mode_valid(intel_connector, mode);
+	if (status != MODE_OK)
+		return status;
+
+	if (fixed_mode->clock > max_dotclk)
+		return MODE_CLOCK_HIGH;
+
+	return intel_mode_valid_max_plane_size(dev_priv, mode, false);
+}
+
+struct intel_dsi_host *intel_dsi_host_init(struct intel_dsi *intel_dsi,
+					   const struct mipi_dsi_host_ops *funcs,
+					   enum port port)
+{
+	struct intel_dsi_host *host;
+	struct mipi_dsi_device *device;
+
+	host = kzalloc(sizeof(*host), GFP_KERNEL);
+	if (!host)
+		return NULL;
+
+	host->base.ops = funcs;
+	host->intel_dsi = intel_dsi;
+	host->port = port;
+
+	/*
+	 * We should call mipi_dsi_host_register(&host->base) here, but we don't
+	 * have a host->dev, and we don't have OF stuff either. So just use the
+	 * dsi framework as a library and hope for the best. Create the dsi
+	 * devices by ourselves here too. Need to be careful though, because we
+	 * don't initialize any of the driver model devices here.
+	 */
+	device = kzalloc(sizeof(*device), GFP_KERNEL);
+	if (!device) {
+		kfree(host);
+		return NULL;
+	}
+
+	device->host = &host->base;
+	host->device = device;
+
+	return host;
+}
+
+enum drm_panel_orientation
+intel_dsi_get_panel_orientation(struct intel_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	enum drm_panel_orientation orientation;
+
+	orientation = connector->panel.vbt.dsi.orientation;
+	if (orientation != DRM_MODE_PANEL_ORIENTATION_UNKNOWN)
+		return orientation;
+
+	orientation = dev_priv->display.vbt.orientation;
+	if (orientation != DRM_MODE_PANEL_ORIENTATION_UNKNOWN)
+		return orientation;
+
+	return DRM_MODE_PANEL_ORIENTATION_NORMAL;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dsi.h b/drivers/gpu/drm/i915/display/intel_dsi.h
new file mode 100644
index 000000000..ce80bd8be
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dsi.h
@@ -0,0 +1,177 @@
+/*
+ * Copyright © 2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef _INTEL_DSI_H
+#define _INTEL_DSI_H
+
+#include <drm/drm_crtc.h>
+#include <drm/drm_mipi_dsi.h>
+
+#include "intel_display_types.h"
+
+#define INTEL_DSI_VIDEO_MODE	0
+#define INTEL_DSI_COMMAND_MODE	1
+
+/* Dual Link support */
+#define DSI_DUAL_LINK_NONE		0
+#define DSI_DUAL_LINK_FRONT_BACK	1
+#define DSI_DUAL_LINK_PIXEL_ALT		2
+
+struct intel_dsi_host;
+
+struct intel_dsi {
+	struct intel_encoder base;
+
+	struct intel_dsi_host *dsi_hosts[I915_MAX_PORTS];
+	intel_wakeref_t io_wakeref[I915_MAX_PORTS];
+
+	/* GPIO Desc for panel and backlight control */
+	struct gpio_desc *gpio_panel;
+	struct gpio_desc *gpio_backlight;
+
+	struct intel_connector *attached_connector;
+
+	/* bit mask of ports (vlv dsi) or phys (icl dsi) being driven */
+	union {
+		u16 ports;	/* VLV DSI */
+		u16 phys;	/* ICL DSI */
+	};
+
+	/* if true, use HS mode, otherwise LP */
+	bool hs;
+
+	/* virtual channel */
+	int channel;
+
+	/* Video mode or command mode */
+	u16 operation_mode;
+
+	/* number of DSI lanes */
+	unsigned int lane_count;
+
+	/* i2c bus associated with the slave device */
+	int i2c_bus_num;
+
+	/*
+	 * video mode pixel format
+	 *
+	 * XXX: consolidate on .format in struct mipi_dsi_device.
+	 */
+	enum mipi_dsi_pixel_format pixel_format;
+
+	/* NON_BURST_SYNC_PULSE, NON_BURST_SYNC_EVENTS, or BURST_MODE */
+	int video_mode;
+
+	/* eot for MIPI_EOT_DISABLE register */
+	u8 eotp_pkt;
+	u8 clock_stop;
+
+	u8 escape_clk_div;
+	u8 dual_link;
+
+	/* RGB or BGR */
+	bool bgr_enabled;
+
+	u8 pixel_overlap;
+	u32 port_bits;
+	u32 bw_timer;
+	u32 dphy_reg;
+
+	/* data lanes dphy timing */
+	u32 dphy_data_lane_reg;
+	u32 video_frmt_cfg_bits;
+	u16 lp_byte_clk;
+
+	/* timeouts in byte clocks */
+	u16 hs_tx_timeout;
+	u16 lp_rx_timeout;
+	u16 turn_arnd_val;
+	u16 rst_timer_val;
+	u16 hs_to_lp_count;
+	u16 clk_lp_to_hs_count;
+	u16 clk_hs_to_lp_count;
+
+	u16 init_count;
+	u32 pclk;
+	u16 burst_mode_ratio;
+
+	/* all delays in ms */
+	u16 backlight_off_delay;
+	u16 backlight_on_delay;
+	u16 panel_on_delay;
+	u16 panel_off_delay;
+	u16 panel_pwr_cycle_delay;
+	ktime_t panel_power_off_time;
+};
+
+struct intel_dsi_host {
+	struct mipi_dsi_host base;
+	struct intel_dsi *intel_dsi;
+	enum port port;
+
+	/* our little hack */
+	struct mipi_dsi_device *device;
+};
+
+static inline struct intel_dsi_host *to_intel_dsi_host(struct mipi_dsi_host *h)
+{
+	return container_of(h, struct intel_dsi_host, base);
+}
+
+#define for_each_dsi_port(__port, __ports_mask) \
+	for_each_port_masked(__port, __ports_mask)
+#define for_each_dsi_phy(__phy, __phys_mask) \
+	for_each_phy_masked(__phy, __phys_mask)
+
+static inline struct intel_dsi *enc_to_intel_dsi(struct intel_encoder *encoder)
+{
+	return container_of(&encoder->base, struct intel_dsi, base.base);
+}
+
+static inline bool is_vid_mode(struct intel_dsi *intel_dsi)
+{
+	return intel_dsi->operation_mode == INTEL_DSI_VIDEO_MODE;
+}
+
+static inline bool is_cmd_mode(struct intel_dsi *intel_dsi)
+{
+	return intel_dsi->operation_mode == INTEL_DSI_COMMAND_MODE;
+}
+
+static inline u16 intel_dsi_encoder_ports(struct intel_encoder *encoder)
+{
+	return enc_to_intel_dsi(encoder)->ports;
+}
+
+int intel_dsi_bitrate(const struct intel_dsi *intel_dsi);
+int intel_dsi_tlpx_ns(const struct intel_dsi *intel_dsi);
+enum drm_panel_orientation
+intel_dsi_get_panel_orientation(struct intel_connector *connector);
+int intel_dsi_get_modes(struct drm_connector *connector);
+enum drm_mode_status intel_dsi_mode_valid(struct drm_connector *connector,
+					  struct drm_display_mode *mode);
+struct intel_dsi_host *intel_dsi_host_init(struct intel_dsi *intel_dsi,
+					   const struct mipi_dsi_host_ops *funcs,
+					   enum port port);
+
+#endif /* _INTEL_DSI_H */
diff --git a/drivers/gpu/drm/i915/display/intel_dsi_dcs_backlight.c b/drivers/gpu/drm/i915/display/intel_dsi_dcs_backlight.c
new file mode 100644
index 000000000..20e466d84
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dsi_dcs_backlight.c
@@ -0,0 +1,200 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Author: Deepak M <m.deepak at intel.com>
+ */
+
+#include <drm/drm_mipi_dsi.h>
+#include <video/mipi_display.h>
+
+#include "i915_drv.h"
+#include "intel_display_types.h"
+#include "intel_dsi.h"
+#include "intel_dsi_dcs_backlight.h"
+
+#define CONTROL_DISPLAY_BCTRL		(1 << 5)
+#define CONTROL_DISPLAY_DD		(1 << 3)
+#define CONTROL_DISPLAY_BL		(1 << 2)
+
+#define POWER_SAVE_OFF			(0 << 0)
+#define POWER_SAVE_LOW			(1 << 0)
+#define POWER_SAVE_MEDIUM		(2 << 0)
+#define POWER_SAVE_HIGH			(3 << 0)
+#define POWER_SAVE_OUTDOOR_MODE		(4 << 0)
+
+#define PANEL_PWM_MAX_VALUE		0xFF
+
+static u32 dcs_get_backlight(struct intel_connector *connector, enum pipe unused)
+{
+	struct intel_encoder *encoder = intel_attached_encoder(connector);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	struct intel_panel *panel = &connector->panel;
+	struct mipi_dsi_device *dsi_device;
+	u8 data[2] = {};
+	enum port port;
+	size_t len = panel->backlight.max > U8_MAX ? 2 : 1;
+
+	for_each_dsi_port(port, panel->vbt.dsi.bl_ports) {
+		dsi_device = intel_dsi->dsi_hosts[port]->device;
+		mipi_dsi_dcs_read(dsi_device, MIPI_DCS_GET_DISPLAY_BRIGHTNESS,
+				  &data, len);
+		break;
+	}
+
+	return (data[1] << 8) | data[0];
+}
+
+static void dcs_set_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(to_intel_encoder(conn_state->best_encoder));
+	struct intel_panel *panel = &to_intel_connector(conn_state->connector)->panel;
+	struct mipi_dsi_device *dsi_device;
+	u8 data[2] = {};
+	enum port port;
+	size_t len = panel->backlight.max > U8_MAX ? 2 : 1;
+	unsigned long mode_flags;
+
+	if (len == 1) {
+		data[0] = level;
+	} else {
+		data[0] = level >> 8;
+		data[1] = level;
+	}
+
+	for_each_dsi_port(port, panel->vbt.dsi.bl_ports) {
+		dsi_device = intel_dsi->dsi_hosts[port]->device;
+		mode_flags = dsi_device->mode_flags;
+		dsi_device->mode_flags &= ~MIPI_DSI_MODE_LPM;
+		mipi_dsi_dcs_write(dsi_device, MIPI_DCS_SET_DISPLAY_BRIGHTNESS,
+				   &data, len);
+		dsi_device->mode_flags = mode_flags;
+	}
+}
+
+static void dcs_disable_backlight(const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(to_intel_encoder(conn_state->best_encoder));
+	struct intel_panel *panel = &to_intel_connector(conn_state->connector)->panel;
+	struct mipi_dsi_device *dsi_device;
+	enum port port;
+
+	dcs_set_backlight(conn_state, 0);
+
+	for_each_dsi_port(port, panel->vbt.dsi.cabc_ports) {
+		u8 cabc = POWER_SAVE_OFF;
+
+		dsi_device = intel_dsi->dsi_hosts[port]->device;
+		mipi_dsi_dcs_write(dsi_device, MIPI_DCS_WRITE_POWER_SAVE,
+				   &cabc, sizeof(cabc));
+	}
+
+	for_each_dsi_port(port, panel->vbt.dsi.bl_ports) {
+		u8 ctrl = 0;
+
+		dsi_device = intel_dsi->dsi_hosts[port]->device;
+
+		mipi_dsi_dcs_read(dsi_device, MIPI_DCS_GET_CONTROL_DISPLAY,
+				  &ctrl, sizeof(ctrl));
+
+		ctrl &= ~CONTROL_DISPLAY_BL;
+		ctrl &= ~CONTROL_DISPLAY_DD;
+		ctrl &= ~CONTROL_DISPLAY_BCTRL;
+
+		mipi_dsi_dcs_write(dsi_device, MIPI_DCS_WRITE_CONTROL_DISPLAY,
+				   &ctrl, sizeof(ctrl));
+	}
+}
+
+static void dcs_enable_backlight(const struct intel_crtc_state *crtc_state,
+				 const struct drm_connector_state *conn_state, u32 level)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(to_intel_encoder(conn_state->best_encoder));
+	struct intel_panel *panel = &to_intel_connector(conn_state->connector)->panel;
+	struct mipi_dsi_device *dsi_device;
+	enum port port;
+
+	for_each_dsi_port(port, panel->vbt.dsi.bl_ports) {
+		u8 ctrl = 0;
+
+		dsi_device = intel_dsi->dsi_hosts[port]->device;
+
+		mipi_dsi_dcs_read(dsi_device, MIPI_DCS_GET_CONTROL_DISPLAY,
+				  &ctrl, sizeof(ctrl));
+
+		ctrl |= CONTROL_DISPLAY_BL;
+		ctrl |= CONTROL_DISPLAY_DD;
+		ctrl |= CONTROL_DISPLAY_BCTRL;
+
+		mipi_dsi_dcs_write(dsi_device, MIPI_DCS_WRITE_CONTROL_DISPLAY,
+				   &ctrl, sizeof(ctrl));
+	}
+
+	for_each_dsi_port(port, panel->vbt.dsi.cabc_ports) {
+		u8 cabc = POWER_SAVE_MEDIUM;
+
+		dsi_device = intel_dsi->dsi_hosts[port]->device;
+		mipi_dsi_dcs_write(dsi_device, MIPI_DCS_WRITE_POWER_SAVE,
+				   &cabc, sizeof(cabc));
+	}
+
+	dcs_set_backlight(conn_state, level);
+}
+
+static int dcs_setup_backlight(struct intel_connector *connector,
+			       enum pipe unused)
+{
+	struct intel_panel *panel = &connector->panel;
+
+	if (panel->vbt.backlight.brightness_precision_bits > 8)
+		panel->backlight.max = (1 << panel->vbt.backlight.brightness_precision_bits) - 1;
+	else
+		panel->backlight.max = PANEL_PWM_MAX_VALUE;
+
+	panel->backlight.level = panel->backlight.max;
+
+	return 0;
+}
+
+static const struct intel_panel_bl_funcs dcs_bl_funcs = {
+	.setup = dcs_setup_backlight,
+	.enable = dcs_enable_backlight,
+	.disable = dcs_disable_backlight,
+	.set = dcs_set_backlight,
+	.get = dcs_get_backlight,
+};
+
+int intel_dsi_dcs_init_backlight_funcs(struct intel_connector *intel_connector)
+{
+	struct drm_device *dev = intel_connector->base.dev;
+	struct intel_encoder *encoder = intel_attached_encoder(intel_connector);
+	struct intel_panel *panel = &intel_connector->panel;
+
+	if (panel->vbt.backlight.type != INTEL_BACKLIGHT_DSI_DCS)
+		return -ENODEV;
+
+	if (drm_WARN_ON(dev, encoder->type != INTEL_OUTPUT_DSI))
+		return -EINVAL;
+
+	panel->backlight.funcs = &dcs_bl_funcs;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dsi_dcs_backlight.h b/drivers/gpu/drm/i915/display/intel_dsi_dcs_backlight.h
new file mode 100644
index 000000000..eb0194784
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dsi_dcs_backlight.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DSI_DCS_BACKLIGHT_H__
+#define __INTEL_DSI_DCS_BACKLIGHT_H__
+
+struct intel_connector;
+
+int intel_dsi_dcs_init_backlight_funcs(struct intel_connector *intel_connector);
+
+#endif /* __INTEL_DSI_DCS_BACKLIGHT_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dsi_vbt.c b/drivers/gpu/drm/i915/display/intel_dsi_vbt.c
new file mode 100644
index 000000000..f102c13cb
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dsi_vbt.c
@@ -0,0 +1,1037 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Author: Shobhit Kumar <shobhit.kumar@intel.com>
+ *
+ */
+
+#include <linux/gpio/consumer.h>
+#include <linux/gpio/machine.h>
+#include <linux/mfd/intel_soc_pmic.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/pinctrl/machine.h>
+#include <linux/slab.h>
+#include <linux/string_helpers.h>
+
+#include <asm/unaligned.h>
+
+#include <drm/drm_crtc.h>
+#include <drm/drm_edid.h>
+
+#include <video/mipi_display.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dsi.h"
+#include "intel_dsi_vbt.h"
+#include "intel_gmbus_regs.h"
+#include "vlv_dsi.h"
+#include "vlv_dsi_regs.h"
+#include "vlv_sideband.h"
+
+#define MIPI_TRANSFER_MODE_SHIFT	0
+#define MIPI_VIRTUAL_CHANNEL_SHIFT	1
+#define MIPI_PORT_SHIFT			3
+
+/* base offsets for gpio pads */
+#define VLV_GPIO_NC_0_HV_DDI0_HPD	0x4130
+#define VLV_GPIO_NC_1_HV_DDI0_DDC_SDA	0x4120
+#define VLV_GPIO_NC_2_HV_DDI0_DDC_SCL	0x4110
+#define VLV_GPIO_NC_3_PANEL0_VDDEN	0x4140
+#define VLV_GPIO_NC_4_PANEL0_BKLTEN	0x4150
+#define VLV_GPIO_NC_5_PANEL0_BKLTCTL	0x4160
+#define VLV_GPIO_NC_6_HV_DDI1_HPD	0x4180
+#define VLV_GPIO_NC_7_HV_DDI1_DDC_SDA	0x4190
+#define VLV_GPIO_NC_8_HV_DDI1_DDC_SCL	0x4170
+#define VLV_GPIO_NC_9_PANEL1_VDDEN	0x4100
+#define VLV_GPIO_NC_10_PANEL1_BKLTEN	0x40E0
+#define VLV_GPIO_NC_11_PANEL1_BKLTCTL	0x40F0
+
+#define VLV_GPIO_PCONF0(base_offset)	(base_offset)
+#define VLV_GPIO_PAD_VAL(base_offset)	((base_offset) + 8)
+
+struct gpio_map {
+	u16 base_offset;
+	bool init;
+};
+
+static struct gpio_map vlv_gpio_table[] = {
+	{ VLV_GPIO_NC_0_HV_DDI0_HPD },
+	{ VLV_GPIO_NC_1_HV_DDI0_DDC_SDA },
+	{ VLV_GPIO_NC_2_HV_DDI0_DDC_SCL },
+	{ VLV_GPIO_NC_3_PANEL0_VDDEN },
+	{ VLV_GPIO_NC_4_PANEL0_BKLTEN },
+	{ VLV_GPIO_NC_5_PANEL0_BKLTCTL },
+	{ VLV_GPIO_NC_6_HV_DDI1_HPD },
+	{ VLV_GPIO_NC_7_HV_DDI1_DDC_SDA },
+	{ VLV_GPIO_NC_8_HV_DDI1_DDC_SCL },
+	{ VLV_GPIO_NC_9_PANEL1_VDDEN },
+	{ VLV_GPIO_NC_10_PANEL1_BKLTEN },
+	{ VLV_GPIO_NC_11_PANEL1_BKLTCTL },
+};
+
+struct i2c_adapter_lookup {
+	u16 slave_addr;
+	struct intel_dsi *intel_dsi;
+	acpi_handle dev_handle;
+};
+
+#define CHV_GPIO_IDX_START_N		0
+#define CHV_GPIO_IDX_START_E		73
+#define CHV_GPIO_IDX_START_SW		100
+#define CHV_GPIO_IDX_START_SE		198
+
+#define CHV_VBT_MAX_PINS_PER_FMLY	15
+
+#define CHV_GPIO_PAD_CFG0(f, i)		(0x4400 + (f) * 0x400 + (i) * 8)
+#define  CHV_GPIO_GPIOEN		(1 << 15)
+#define  CHV_GPIO_GPIOCFG_GPIO		(0 << 8)
+#define  CHV_GPIO_GPIOCFG_GPO		(1 << 8)
+#define  CHV_GPIO_GPIOCFG_GPI		(2 << 8)
+#define  CHV_GPIO_GPIOCFG_HIZ		(3 << 8)
+#define  CHV_GPIO_GPIOTXSTATE(state)	((!!(state)) << 1)
+
+#define CHV_GPIO_PAD_CFG1(f, i)		(0x4400 + (f) * 0x400 + (i) * 8 + 4)
+#define  CHV_GPIO_CFGLOCK		(1 << 31)
+
+/* ICL DSI Display GPIO Pins */
+#define  ICL_GPIO_DDSP_HPD_A		0
+#define  ICL_GPIO_L_VDDEN_1		1
+#define  ICL_GPIO_L_BKLTEN_1		2
+#define  ICL_GPIO_DDPA_CTRLCLK_1	3
+#define  ICL_GPIO_DDPA_CTRLDATA_1	4
+#define  ICL_GPIO_DDSP_HPD_B		5
+#define  ICL_GPIO_L_VDDEN_2		6
+#define  ICL_GPIO_L_BKLTEN_2		7
+#define  ICL_GPIO_DDPA_CTRLCLK_2	8
+#define  ICL_GPIO_DDPA_CTRLDATA_2	9
+
+static enum port intel_dsi_seq_port_to_port(struct intel_dsi *intel_dsi,
+					    u8 seq_port)
+{
+	/*
+	 * If single link DSI is being used on any port, the VBT sequence block
+	 * send packet apparently always has 0 for the port. Just use the port
+	 * we have configured, and ignore the sequence block port.
+	 */
+	if (hweight8(intel_dsi->ports) == 1)
+		return ffs(intel_dsi->ports) - 1;
+
+	if (seq_port) {
+		if (intel_dsi->ports & BIT(PORT_B))
+			return PORT_B;
+		else if (intel_dsi->ports & BIT(PORT_C))
+			return PORT_C;
+	}
+
+	return PORT_A;
+}
+
+static const u8 *mipi_exec_send_packet(struct intel_dsi *intel_dsi,
+				       const u8 *data)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_dsi->base.base.dev);
+	struct mipi_dsi_device *dsi_device;
+	u8 type, flags, seq_port;
+	u16 len;
+	enum port port;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	flags = *data++;
+	type = *data++;
+
+	len = *((u16 *) data);
+	data += 2;
+
+	seq_port = (flags >> MIPI_PORT_SHIFT) & 3;
+
+	port = intel_dsi_seq_port_to_port(intel_dsi, seq_port);
+
+	if (drm_WARN_ON(&dev_priv->drm, !intel_dsi->dsi_hosts[port]))
+		goto out;
+
+	dsi_device = intel_dsi->dsi_hosts[port]->device;
+	if (!dsi_device) {
+		drm_dbg_kms(&dev_priv->drm, "no dsi device for port %c\n",
+			    port_name(port));
+		goto out;
+	}
+
+	if ((flags >> MIPI_TRANSFER_MODE_SHIFT) & 1)
+		dsi_device->mode_flags &= ~MIPI_DSI_MODE_LPM;
+	else
+		dsi_device->mode_flags |= MIPI_DSI_MODE_LPM;
+
+	dsi_device->channel = (flags >> MIPI_VIRTUAL_CHANNEL_SHIFT) & 3;
+
+	switch (type) {
+	case MIPI_DSI_GENERIC_SHORT_WRITE_0_PARAM:
+		mipi_dsi_generic_write(dsi_device, NULL, 0);
+		break;
+	case MIPI_DSI_GENERIC_SHORT_WRITE_1_PARAM:
+		mipi_dsi_generic_write(dsi_device, data, 1);
+		break;
+	case MIPI_DSI_GENERIC_SHORT_WRITE_2_PARAM:
+		mipi_dsi_generic_write(dsi_device, data, 2);
+		break;
+	case MIPI_DSI_GENERIC_READ_REQUEST_0_PARAM:
+	case MIPI_DSI_GENERIC_READ_REQUEST_1_PARAM:
+	case MIPI_DSI_GENERIC_READ_REQUEST_2_PARAM:
+		drm_dbg(&dev_priv->drm,
+			"Generic Read not yet implemented or used\n");
+		break;
+	case MIPI_DSI_GENERIC_LONG_WRITE:
+		mipi_dsi_generic_write(dsi_device, data, len);
+		break;
+	case MIPI_DSI_DCS_SHORT_WRITE:
+		mipi_dsi_dcs_write_buffer(dsi_device, data, 1);
+		break;
+	case MIPI_DSI_DCS_SHORT_WRITE_PARAM:
+		mipi_dsi_dcs_write_buffer(dsi_device, data, 2);
+		break;
+	case MIPI_DSI_DCS_READ:
+		drm_dbg(&dev_priv->drm,
+			"DCS Read not yet implemented or used\n");
+		break;
+	case MIPI_DSI_DCS_LONG_WRITE:
+		mipi_dsi_dcs_write_buffer(dsi_device, data, len);
+		break;
+	}
+
+	if (DISPLAY_VER(dev_priv) < 11)
+		vlv_dsi_wait_for_fifo_empty(intel_dsi, port);
+
+out:
+	data += len;
+
+	return data;
+}
+
+static const u8 *mipi_exec_delay(struct intel_dsi *intel_dsi, const u8 *data)
+{
+	struct drm_i915_private *i915 = to_i915(intel_dsi->base.base.dev);
+	u32 delay = *((const u32 *) data);
+
+	drm_dbg_kms(&i915->drm, "\n");
+
+	usleep_range(delay, delay + 10);
+	data += 4;
+
+	return data;
+}
+
+static void vlv_exec_gpio(struct intel_connector *connector,
+			  u8 gpio_source, u8 gpio_index, bool value)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct gpio_map *map;
+	u16 pconf0, padval;
+	u32 tmp;
+	u8 port;
+
+	if (gpio_index >= ARRAY_SIZE(vlv_gpio_table)) {
+		drm_dbg_kms(&dev_priv->drm, "unknown gpio index %u\n",
+			    gpio_index);
+		return;
+	}
+
+	map = &vlv_gpio_table[gpio_index];
+
+	if (connector->panel.vbt.dsi.seq_version >= 3) {
+		/* XXX: this assumes vlv_gpio_table only has NC GPIOs. */
+		port = IOSF_PORT_GPIO_NC;
+	} else {
+		if (gpio_source == 0) {
+			port = IOSF_PORT_GPIO_NC;
+		} else if (gpio_source == 1) {
+			drm_dbg_kms(&dev_priv->drm, "SC gpio not supported\n");
+			return;
+		} else {
+			drm_dbg_kms(&dev_priv->drm,
+				    "unknown gpio source %u\n", gpio_source);
+			return;
+		}
+	}
+
+	pconf0 = VLV_GPIO_PCONF0(map->base_offset);
+	padval = VLV_GPIO_PAD_VAL(map->base_offset);
+
+	vlv_iosf_sb_get(dev_priv, BIT(VLV_IOSF_SB_GPIO));
+	if (!map->init) {
+		/* FIXME: remove constant below */
+		vlv_iosf_sb_write(dev_priv, port, pconf0, 0x2000CC00);
+		map->init = true;
+	}
+
+	tmp = 0x4 | value;
+	vlv_iosf_sb_write(dev_priv, port, padval, tmp);
+	vlv_iosf_sb_put(dev_priv, BIT(VLV_IOSF_SB_GPIO));
+}
+
+static void chv_exec_gpio(struct intel_connector *connector,
+			  u8 gpio_source, u8 gpio_index, bool value)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	u16 cfg0, cfg1;
+	u16 family_num;
+	u8 port;
+
+	if (connector->panel.vbt.dsi.seq_version >= 3) {
+		if (gpio_index >= CHV_GPIO_IDX_START_SE) {
+			/* XXX: it's unclear whether 255->57 is part of SE. */
+			gpio_index -= CHV_GPIO_IDX_START_SE;
+			port = CHV_IOSF_PORT_GPIO_SE;
+		} else if (gpio_index >= CHV_GPIO_IDX_START_SW) {
+			gpio_index -= CHV_GPIO_IDX_START_SW;
+			port = CHV_IOSF_PORT_GPIO_SW;
+		} else if (gpio_index >= CHV_GPIO_IDX_START_E) {
+			gpio_index -= CHV_GPIO_IDX_START_E;
+			port = CHV_IOSF_PORT_GPIO_E;
+		} else {
+			port = CHV_IOSF_PORT_GPIO_N;
+		}
+	} else {
+		/* XXX: The spec is unclear about CHV GPIO on seq v2 */
+		if (gpio_source != 0) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "unknown gpio source %u\n", gpio_source);
+			return;
+		}
+
+		if (gpio_index >= CHV_GPIO_IDX_START_E) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "invalid gpio index %u for GPIO N\n",
+				    gpio_index);
+			return;
+		}
+
+		port = CHV_IOSF_PORT_GPIO_N;
+	}
+
+	family_num = gpio_index / CHV_VBT_MAX_PINS_PER_FMLY;
+	gpio_index = gpio_index % CHV_VBT_MAX_PINS_PER_FMLY;
+
+	cfg0 = CHV_GPIO_PAD_CFG0(family_num, gpio_index);
+	cfg1 = CHV_GPIO_PAD_CFG1(family_num, gpio_index);
+
+	vlv_iosf_sb_get(dev_priv, BIT(VLV_IOSF_SB_GPIO));
+	vlv_iosf_sb_write(dev_priv, port, cfg1, 0);
+	vlv_iosf_sb_write(dev_priv, port, cfg0,
+			  CHV_GPIO_GPIOEN | CHV_GPIO_GPIOCFG_GPO |
+			  CHV_GPIO_GPIOTXSTATE(value));
+	vlv_iosf_sb_put(dev_priv, BIT(VLV_IOSF_SB_GPIO));
+}
+
+static void bxt_exec_gpio(struct intel_connector *connector,
+			  u8 gpio_source, u8 gpio_index, bool value)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	/* XXX: this table is a quick ugly hack. */
+	static struct gpio_desc *bxt_gpio_table[U8_MAX + 1];
+	struct gpio_desc *gpio_desc = bxt_gpio_table[gpio_index];
+
+	if (!gpio_desc) {
+		gpio_desc = devm_gpiod_get_index(dev_priv->drm.dev,
+						 NULL, gpio_index,
+						 value ? GPIOD_OUT_LOW :
+						 GPIOD_OUT_HIGH);
+
+		if (IS_ERR_OR_NULL(gpio_desc)) {
+			drm_err(&dev_priv->drm,
+				"GPIO index %u request failed (%ld)\n",
+				gpio_index, PTR_ERR(gpio_desc));
+			return;
+		}
+
+		bxt_gpio_table[gpio_index] = gpio_desc;
+	}
+
+	gpiod_set_value(gpio_desc, value);
+}
+
+static void icl_exec_gpio(struct intel_connector *connector,
+			  u8 gpio_source, u8 gpio_index, bool value)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+
+	drm_dbg_kms(&dev_priv->drm, "Skipping ICL GPIO element execution\n");
+}
+
+enum {
+	MIPI_RESET_1 = 0,
+	MIPI_AVDD_EN_1,
+	MIPI_BKLT_EN_1,
+	MIPI_AVEE_EN_1,
+	MIPI_VIO_EN_1,
+	MIPI_RESET_2,
+	MIPI_AVDD_EN_2,
+	MIPI_BKLT_EN_2,
+	MIPI_AVEE_EN_2,
+	MIPI_VIO_EN_2,
+};
+
+static void icl_native_gpio_set_value(struct drm_i915_private *dev_priv,
+				      int gpio, bool value)
+{
+	int index;
+
+	if (drm_WARN_ON(&dev_priv->drm, DISPLAY_VER(dev_priv) == 11 && gpio >= MIPI_RESET_2))
+		return;
+
+	switch (gpio) {
+	case MIPI_RESET_1:
+	case MIPI_RESET_2:
+		index = gpio == MIPI_RESET_1 ? HPD_PORT_A : HPD_PORT_B;
+
+		/*
+		 * Disable HPD to set the pin to output, and set output
+		 * value. The HPD pin should not be enabled for DSI anyway,
+		 * assuming the board design and VBT are sane, and the pin isn't
+		 * used by a non-DSI encoder.
+		 *
+		 * The locking protects against concurrent SHOTPLUG_CTL_DDI
+		 * modifications in irq setup and handling.
+		 */
+		spin_lock_irq(&dev_priv->irq_lock);
+		intel_de_rmw(dev_priv, SHOTPLUG_CTL_DDI,
+			     SHOTPLUG_CTL_DDI_HPD_ENABLE(index) |
+			     SHOTPLUG_CTL_DDI_HPD_OUTPUT_DATA(index),
+			     value ? SHOTPLUG_CTL_DDI_HPD_OUTPUT_DATA(index) : 0);
+		spin_unlock_irq(&dev_priv->irq_lock);
+		break;
+	case MIPI_AVDD_EN_1:
+	case MIPI_AVDD_EN_2:
+		index = gpio == MIPI_AVDD_EN_1 ? 0 : 1;
+
+		intel_de_rmw(dev_priv, PP_CONTROL(index), PANEL_POWER_ON,
+			     value ? PANEL_POWER_ON : 0);
+		break;
+	case MIPI_BKLT_EN_1:
+	case MIPI_BKLT_EN_2:
+		index = gpio == MIPI_BKLT_EN_1 ? 0 : 1;
+
+		intel_de_rmw(dev_priv, PP_CONTROL(index), EDP_BLC_ENABLE,
+			     value ? EDP_BLC_ENABLE : 0);
+		break;
+	case MIPI_AVEE_EN_1:
+	case MIPI_AVEE_EN_2:
+		index = gpio == MIPI_AVEE_EN_1 ? 1 : 2;
+
+		intel_de_rmw(dev_priv, GPIO(dev_priv, index),
+			     GPIO_CLOCK_VAL_OUT,
+			     GPIO_CLOCK_DIR_MASK | GPIO_CLOCK_DIR_OUT |
+			     GPIO_CLOCK_VAL_MASK | (value ? GPIO_CLOCK_VAL_OUT : 0));
+		break;
+	case MIPI_VIO_EN_1:
+	case MIPI_VIO_EN_2:
+		index = gpio == MIPI_VIO_EN_1 ? 1 : 2;
+
+		intel_de_rmw(dev_priv, GPIO(dev_priv, index),
+			     GPIO_DATA_VAL_OUT,
+			     GPIO_DATA_DIR_MASK | GPIO_DATA_DIR_OUT |
+			     GPIO_DATA_VAL_MASK | (value ? GPIO_DATA_VAL_OUT : 0));
+		break;
+	default:
+		MISSING_CASE(gpio);
+	}
+}
+
+static const u8 *mipi_exec_gpio(struct intel_dsi *intel_dsi, const u8 *data)
+{
+	struct drm_device *dev = intel_dsi->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_connector *connector = intel_dsi->attached_connector;
+	u8 gpio_source, gpio_index = 0, gpio_number;
+	bool value;
+	bool native = DISPLAY_VER(dev_priv) >= 11;
+
+	if (connector->panel.vbt.dsi.seq_version >= 3)
+		gpio_index = *data++;
+
+	gpio_number = *data++;
+
+	/* gpio source in sequence v2 only */
+	if (connector->panel.vbt.dsi.seq_version == 2)
+		gpio_source = (*data >> 1) & 3;
+	else
+		gpio_source = 0;
+
+	if (connector->panel.vbt.dsi.seq_version >= 4 && *data & BIT(1))
+		native = false;
+
+	/* pull up/down */
+	value = *data++ & 1;
+
+	drm_dbg_kms(&dev_priv->drm, "GPIO index %u, number %u, source %u, native %s, set to %s\n",
+		    gpio_index, gpio_number, gpio_source, str_yes_no(native), str_on_off(value));
+
+	if (native)
+		icl_native_gpio_set_value(dev_priv, gpio_number, value);
+	else if (DISPLAY_VER(dev_priv) >= 11)
+		icl_exec_gpio(connector, gpio_source, gpio_index, value);
+	else if (IS_VALLEYVIEW(dev_priv))
+		vlv_exec_gpio(connector, gpio_source, gpio_number, value);
+	else if (IS_CHERRYVIEW(dev_priv))
+		chv_exec_gpio(connector, gpio_source, gpio_number, value);
+	else
+		bxt_exec_gpio(connector, gpio_source, gpio_index, value);
+
+	return data;
+}
+
+#ifdef CONFIG_ACPI
+static int i2c_adapter_lookup(struct acpi_resource *ares, void *data)
+{
+	struct i2c_adapter_lookup *lookup = data;
+	struct intel_dsi *intel_dsi = lookup->intel_dsi;
+	struct acpi_resource_i2c_serialbus *sb;
+	struct i2c_adapter *adapter;
+	acpi_handle adapter_handle;
+	acpi_status status;
+
+	if (!i2c_acpi_get_i2c_resource(ares, &sb))
+		return 1;
+
+	if (lookup->slave_addr != sb->slave_address)
+		return 1;
+
+	status = acpi_get_handle(lookup->dev_handle,
+				 sb->resource_source.string_ptr,
+				 &adapter_handle);
+	if (ACPI_FAILURE(status))
+		return 1;
+
+	adapter = i2c_acpi_find_adapter_by_handle(adapter_handle);
+	if (adapter)
+		intel_dsi->i2c_bus_num = adapter->nr;
+
+	return 1;
+}
+
+static void i2c_acpi_find_adapter(struct intel_dsi *intel_dsi,
+				  const u16 slave_addr)
+{
+	struct drm_device *drm_dev = intel_dsi->base.base.dev;
+	struct acpi_device *adev = ACPI_COMPANION(drm_dev->dev);
+	struct i2c_adapter_lookup lookup = {
+		.slave_addr = slave_addr,
+		.intel_dsi = intel_dsi,
+		.dev_handle = acpi_device_handle(adev),
+	};
+	LIST_HEAD(resource_list);
+
+	acpi_dev_get_resources(adev, &resource_list, i2c_adapter_lookup, &lookup);
+	acpi_dev_free_resource_list(&resource_list);
+}
+#else
+static inline void i2c_acpi_find_adapter(struct intel_dsi *intel_dsi,
+					 const u16 slave_addr)
+{
+}
+#endif
+
+static const u8 *mipi_exec_i2c(struct intel_dsi *intel_dsi, const u8 *data)
+{
+	struct drm_i915_private *i915 = to_i915(intel_dsi->base.base.dev);
+	struct i2c_adapter *adapter;
+	struct i2c_msg msg;
+	int ret;
+	u8 vbt_i2c_bus_num = *(data + 2);
+	u16 slave_addr = *(u16 *)(data + 3);
+	u8 reg_offset = *(data + 5);
+	u8 payload_size = *(data + 6);
+	u8 *payload_data;
+
+	if (intel_dsi->i2c_bus_num < 0) {
+		intel_dsi->i2c_bus_num = vbt_i2c_bus_num;
+		i2c_acpi_find_adapter(intel_dsi, slave_addr);
+	}
+
+	adapter = i2c_get_adapter(intel_dsi->i2c_bus_num);
+	if (!adapter) {
+		drm_err(&i915->drm, "Cannot find a valid i2c bus for xfer\n");
+		goto err_bus;
+	}
+
+	payload_data = kzalloc(payload_size + 1, GFP_KERNEL);
+	if (!payload_data)
+		goto err_alloc;
+
+	payload_data[0] = reg_offset;
+	memcpy(&payload_data[1], (data + 7), payload_size);
+
+	msg.addr = slave_addr;
+	msg.flags = 0;
+	msg.len = payload_size + 1;
+	msg.buf = payload_data;
+
+	ret = i2c_transfer(adapter, &msg, 1);
+	if (ret < 0)
+		drm_err(&i915->drm,
+			"Failed to xfer payload of size (%u) to reg (%u)\n",
+			payload_size, reg_offset);
+
+	kfree(payload_data);
+err_alloc:
+	i2c_put_adapter(adapter);
+err_bus:
+	return data + payload_size + 7;
+}
+
+static const u8 *mipi_exec_spi(struct intel_dsi *intel_dsi, const u8 *data)
+{
+	struct drm_i915_private *i915 = to_i915(intel_dsi->base.base.dev);
+
+	drm_dbg_kms(&i915->drm, "Skipping SPI element execution\n");
+
+	return data + *(data + 5) + 6;
+}
+
+static const u8 *mipi_exec_pmic(struct intel_dsi *intel_dsi, const u8 *data)
+{
+	struct drm_i915_private *i915 = to_i915(intel_dsi->base.base.dev);
+#ifdef CONFIG_PMIC_OPREGION
+	u32 value, mask, reg_address;
+	u16 i2c_address;
+	int ret;
+
+	/* byte 0 aka PMIC Flag is reserved */
+	i2c_address	= get_unaligned_le16(data + 1);
+	reg_address	= get_unaligned_le32(data + 3);
+	value		= get_unaligned_le32(data + 7);
+	mask		= get_unaligned_le32(data + 11);
+
+	ret = intel_soc_pmic_exec_mipi_pmic_seq_element(i2c_address,
+							reg_address,
+							value, mask);
+	if (ret)
+		drm_err(&i915->drm, "%s failed, error: %d\n", __func__, ret);
+#else
+	drm_err(&i915->drm,
+		"Your hardware requires CONFIG_PMIC_OPREGION and it is not set\n");
+#endif
+
+	return data + 15;
+}
+
+typedef const u8 * (*fn_mipi_elem_exec)(struct intel_dsi *intel_dsi,
+					const u8 *data);
+static const fn_mipi_elem_exec exec_elem[] = {
+	[MIPI_SEQ_ELEM_SEND_PKT] = mipi_exec_send_packet,
+	[MIPI_SEQ_ELEM_DELAY] = mipi_exec_delay,
+	[MIPI_SEQ_ELEM_GPIO] = mipi_exec_gpio,
+	[MIPI_SEQ_ELEM_I2C] = mipi_exec_i2c,
+	[MIPI_SEQ_ELEM_SPI] = mipi_exec_spi,
+	[MIPI_SEQ_ELEM_PMIC] = mipi_exec_pmic,
+};
+
+/*
+ * MIPI Sequence from VBT #53 parsing logic
+ * We have already separated each seqence during bios parsing
+ * Following is generic execution function for any sequence
+ */
+
+static const char * const seq_name[] = {
+	[MIPI_SEQ_DEASSERT_RESET] = "MIPI_SEQ_DEASSERT_RESET",
+	[MIPI_SEQ_INIT_OTP] = "MIPI_SEQ_INIT_OTP",
+	[MIPI_SEQ_DISPLAY_ON] = "MIPI_SEQ_DISPLAY_ON",
+	[MIPI_SEQ_DISPLAY_OFF]  = "MIPI_SEQ_DISPLAY_OFF",
+	[MIPI_SEQ_ASSERT_RESET] = "MIPI_SEQ_ASSERT_RESET",
+	[MIPI_SEQ_BACKLIGHT_ON] = "MIPI_SEQ_BACKLIGHT_ON",
+	[MIPI_SEQ_BACKLIGHT_OFF] = "MIPI_SEQ_BACKLIGHT_OFF",
+	[MIPI_SEQ_TEAR_ON] = "MIPI_SEQ_TEAR_ON",
+	[MIPI_SEQ_TEAR_OFF] = "MIPI_SEQ_TEAR_OFF",
+	[MIPI_SEQ_POWER_ON] = "MIPI_SEQ_POWER_ON",
+	[MIPI_SEQ_POWER_OFF] = "MIPI_SEQ_POWER_OFF",
+};
+
+static const char *sequence_name(enum mipi_seq seq_id)
+{
+	if (seq_id < ARRAY_SIZE(seq_name) && seq_name[seq_id])
+		return seq_name[seq_id];
+	else
+		return "(unknown)";
+}
+
+static void intel_dsi_vbt_exec(struct intel_dsi *intel_dsi,
+			       enum mipi_seq seq_id)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_dsi->base.base.dev);
+	struct intel_connector *connector = intel_dsi->attached_connector;
+	const u8 *data;
+	fn_mipi_elem_exec mipi_elem_exec;
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			seq_id >= ARRAY_SIZE(connector->panel.vbt.dsi.sequence)))
+		return;
+
+	data = connector->panel.vbt.dsi.sequence[seq_id];
+	if (!data)
+		return;
+
+	drm_WARN_ON(&dev_priv->drm, *data != seq_id);
+
+	drm_dbg_kms(&dev_priv->drm, "Starting MIPI sequence %d - %s\n",
+		    seq_id, sequence_name(seq_id));
+
+	/* Skip Sequence Byte. */
+	data++;
+
+	/* Skip Size of Sequence. */
+	if (connector->panel.vbt.dsi.seq_version >= 3)
+		data += 4;
+
+	while (1) {
+		u8 operation_byte = *data++;
+		u8 operation_size = 0;
+
+		if (operation_byte == MIPI_SEQ_ELEM_END)
+			break;
+
+		if (operation_byte < ARRAY_SIZE(exec_elem))
+			mipi_elem_exec = exec_elem[operation_byte];
+		else
+			mipi_elem_exec = NULL;
+
+		/* Size of Operation. */
+		if (connector->panel.vbt.dsi.seq_version >= 3)
+			operation_size = *data++;
+
+		if (mipi_elem_exec) {
+			const u8 *next = data + operation_size;
+
+			data = mipi_elem_exec(intel_dsi, data);
+
+			/* Consistency check if we have size. */
+			if (operation_size && data != next) {
+				drm_err(&dev_priv->drm,
+					"Inconsistent operation size\n");
+				return;
+			}
+		} else if (operation_size) {
+			/* We have size, skip. */
+			drm_dbg_kms(&dev_priv->drm,
+				    "Unsupported MIPI operation byte %u\n",
+				    operation_byte);
+			data += operation_size;
+		} else {
+			/* No size, can't skip without parsing. */
+			drm_err(&dev_priv->drm,
+				"Unsupported MIPI operation byte %u\n",
+				operation_byte);
+			return;
+		}
+	}
+}
+
+void intel_dsi_vbt_exec_sequence(struct intel_dsi *intel_dsi,
+				 enum mipi_seq seq_id)
+{
+	if (seq_id == MIPI_SEQ_POWER_ON && intel_dsi->gpio_panel)
+		gpiod_set_value_cansleep(intel_dsi->gpio_panel, 1);
+	if (seq_id == MIPI_SEQ_BACKLIGHT_ON && intel_dsi->gpio_backlight)
+		gpiod_set_value_cansleep(intel_dsi->gpio_backlight, 1);
+
+	intel_dsi_vbt_exec(intel_dsi, seq_id);
+
+	if (seq_id == MIPI_SEQ_POWER_OFF && intel_dsi->gpio_panel)
+		gpiod_set_value_cansleep(intel_dsi->gpio_panel, 0);
+	if (seq_id == MIPI_SEQ_BACKLIGHT_OFF && intel_dsi->gpio_backlight)
+		gpiod_set_value_cansleep(intel_dsi->gpio_backlight, 0);
+}
+
+void intel_dsi_log_params(struct intel_dsi *intel_dsi)
+{
+	struct drm_i915_private *i915 = to_i915(intel_dsi->base.base.dev);
+
+	drm_dbg_kms(&i915->drm, "Pclk %d\n", intel_dsi->pclk);
+	drm_dbg_kms(&i915->drm, "Pixel overlap %d\n",
+		    intel_dsi->pixel_overlap);
+	drm_dbg_kms(&i915->drm, "Lane count %d\n", intel_dsi->lane_count);
+	drm_dbg_kms(&i915->drm, "DPHY param reg 0x%x\n", intel_dsi->dphy_reg);
+	drm_dbg_kms(&i915->drm, "Video mode format %s\n",
+		    intel_dsi->video_mode == NON_BURST_SYNC_PULSE ?
+		    "non-burst with sync pulse" :
+		    intel_dsi->video_mode == NON_BURST_SYNC_EVENTS ?
+		    "non-burst with sync events" :
+		    intel_dsi->video_mode == BURST_MODE ?
+		    "burst" : "<unknown>");
+	drm_dbg_kms(&i915->drm, "Burst mode ratio %d\n",
+		    intel_dsi->burst_mode_ratio);
+	drm_dbg_kms(&i915->drm, "Reset timer %d\n", intel_dsi->rst_timer_val);
+	drm_dbg_kms(&i915->drm, "Eot %s\n",
+		    str_enabled_disabled(intel_dsi->eotp_pkt));
+	drm_dbg_kms(&i915->drm, "Clockstop %s\n",
+		    str_enabled_disabled(!intel_dsi->clock_stop));
+	drm_dbg_kms(&i915->drm, "Mode %s\n",
+		    intel_dsi->operation_mode ? "command" : "video");
+	if (intel_dsi->dual_link == DSI_DUAL_LINK_FRONT_BACK)
+		drm_dbg_kms(&i915->drm,
+			    "Dual link: DSI_DUAL_LINK_FRONT_BACK\n");
+	else if (intel_dsi->dual_link == DSI_DUAL_LINK_PIXEL_ALT)
+		drm_dbg_kms(&i915->drm,
+			    "Dual link: DSI_DUAL_LINK_PIXEL_ALT\n");
+	else
+		drm_dbg_kms(&i915->drm, "Dual link: NONE\n");
+	drm_dbg_kms(&i915->drm, "Pixel Format %d\n", intel_dsi->pixel_format);
+	drm_dbg_kms(&i915->drm, "TLPX %d\n", intel_dsi->escape_clk_div);
+	drm_dbg_kms(&i915->drm, "LP RX Timeout 0x%x\n",
+		    intel_dsi->lp_rx_timeout);
+	drm_dbg_kms(&i915->drm, "Turnaround Timeout 0x%x\n",
+		    intel_dsi->turn_arnd_val);
+	drm_dbg_kms(&i915->drm, "Init Count 0x%x\n", intel_dsi->init_count);
+	drm_dbg_kms(&i915->drm, "HS to LP Count 0x%x\n",
+		    intel_dsi->hs_to_lp_count);
+	drm_dbg_kms(&i915->drm, "LP Byte Clock %d\n", intel_dsi->lp_byte_clk);
+	drm_dbg_kms(&i915->drm, "DBI BW Timer 0x%x\n", intel_dsi->bw_timer);
+	drm_dbg_kms(&i915->drm, "LP to HS Clock Count 0x%x\n",
+		    intel_dsi->clk_lp_to_hs_count);
+	drm_dbg_kms(&i915->drm, "HS to LP Clock Count 0x%x\n",
+		    intel_dsi->clk_hs_to_lp_count);
+	drm_dbg_kms(&i915->drm, "BTA %s\n",
+		    str_enabled_disabled(!(intel_dsi->video_frmt_cfg_bits & DISABLE_VIDEO_BTA)));
+}
+
+bool intel_dsi_vbt_init(struct intel_dsi *intel_dsi, u16 panel_id)
+{
+	struct drm_device *dev = intel_dsi->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_connector *connector = intel_dsi->attached_connector;
+	struct mipi_config *mipi_config = connector->panel.vbt.dsi.config;
+	struct mipi_pps_data *pps = connector->panel.vbt.dsi.pps;
+	struct drm_display_mode *mode = connector->panel.vbt.lfp_lvds_vbt_mode;
+	u16 burst_mode_ratio;
+	enum port port;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	intel_dsi->eotp_pkt = mipi_config->eot_pkt_disabled ? 0 : 1;
+	intel_dsi->clock_stop = mipi_config->enable_clk_stop ? 1 : 0;
+	intel_dsi->lane_count = mipi_config->lane_cnt + 1;
+	intel_dsi->pixel_format =
+			pixel_format_from_register_bits(
+				mipi_config->videomode_color_format << 7);
+
+	intel_dsi->dual_link = mipi_config->dual_link;
+	intel_dsi->pixel_overlap = mipi_config->pixel_overlap;
+	intel_dsi->operation_mode = mipi_config->is_cmd_mode;
+	intel_dsi->video_mode = mipi_config->video_transfer_mode;
+	intel_dsi->escape_clk_div = mipi_config->byte_clk_sel;
+	intel_dsi->lp_rx_timeout = mipi_config->lp_rx_timeout;
+	intel_dsi->hs_tx_timeout = mipi_config->hs_tx_timeout;
+	intel_dsi->turn_arnd_val = mipi_config->turn_around_timeout;
+	intel_dsi->rst_timer_val = mipi_config->device_reset_timer;
+	intel_dsi->init_count = mipi_config->master_init_timer;
+	intel_dsi->bw_timer = mipi_config->dbi_bw_timer;
+	intel_dsi->video_frmt_cfg_bits =
+		mipi_config->bta_enabled ? DISABLE_VIDEO_BTA : 0;
+	intel_dsi->bgr_enabled = mipi_config->rgb_flip;
+
+	/* Starting point, adjusted depending on dual link and burst mode */
+	intel_dsi->pclk = mode->clock;
+
+	/* In dual link mode each port needs half of pixel clock */
+	if (intel_dsi->dual_link) {
+		intel_dsi->pclk /= 2;
+
+		/* we can enable pixel_overlap if needed by panel. In this
+		 * case we need to increase the pixelclock for extra pixels
+		 */
+		if (intel_dsi->dual_link == DSI_DUAL_LINK_FRONT_BACK) {
+			intel_dsi->pclk += DIV_ROUND_UP(mode->vtotal * intel_dsi->pixel_overlap * 60, 1000);
+		}
+	}
+
+	/* Burst Mode Ratio
+	 * Target ddr frequency from VBT / non burst ddr freq
+	 * multiply by 100 to preserve remainder
+	 */
+	if (intel_dsi->video_mode == BURST_MODE) {
+		if (mipi_config->target_burst_mode_freq) {
+			u32 bitrate = intel_dsi_bitrate(intel_dsi);
+
+			/*
+			 * Sometimes the VBT contains a slightly lower clock,
+			 * then the bitrate we have calculated, in this case
+			 * just replace it with the calculated bitrate.
+			 */
+			if (mipi_config->target_burst_mode_freq < bitrate &&
+			    intel_fuzzy_clock_check(
+					mipi_config->target_burst_mode_freq,
+					bitrate))
+				mipi_config->target_burst_mode_freq = bitrate;
+
+			if (mipi_config->target_burst_mode_freq < bitrate) {
+				drm_err(&dev_priv->drm,
+					"Burst mode freq is less than computed\n");
+				return false;
+			}
+
+			burst_mode_ratio = DIV_ROUND_UP(
+				mipi_config->target_burst_mode_freq * 100,
+				bitrate);
+
+			intel_dsi->pclk = DIV_ROUND_UP(intel_dsi->pclk * burst_mode_ratio, 100);
+		} else {
+			drm_err(&dev_priv->drm,
+				"Burst mode target is not set\n");
+			return false;
+		}
+	} else
+		burst_mode_ratio = 100;
+
+	intel_dsi->burst_mode_ratio = burst_mode_ratio;
+
+	/* delays in VBT are in unit of 100us, so need to convert
+	 * here in ms
+	 * Delay (100us) * 100 /1000 = Delay / 10 (ms) */
+	intel_dsi->backlight_off_delay = pps->bl_disable_delay / 10;
+	intel_dsi->backlight_on_delay = pps->bl_enable_delay / 10;
+	intel_dsi->panel_on_delay = pps->panel_on_delay / 10;
+	intel_dsi->panel_off_delay = pps->panel_off_delay / 10;
+	intel_dsi->panel_pwr_cycle_delay = pps->panel_power_cycle_delay / 10;
+
+	intel_dsi->i2c_bus_num = -1;
+
+	/* a regular driver would get the device in probe */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		mipi_dsi_attach(intel_dsi->dsi_hosts[port]->device);
+	}
+
+	return true;
+}
+
+/*
+ * On some BYT/CHT devs some sequences are incomplete and we need to manually
+ * control some GPIOs. We need to add a GPIO lookup table before we get these.
+ * If the GOP did not initialize the panel (HDMI inserted) we may need to also
+ * change the pinmux for the SoC's PWM0 pin from GPIO to PWM.
+ */
+static struct gpiod_lookup_table pmic_panel_gpio_table = {
+	/* Intel GFX is consumer */
+	.dev_id = "0000:00:02.0",
+	.table = {
+		/* Panel EN/DISABLE */
+		GPIO_LOOKUP("gpio_crystalcove", 94, "panel", GPIO_ACTIVE_HIGH),
+		{ }
+	},
+};
+
+static struct gpiod_lookup_table soc_panel_gpio_table = {
+	.dev_id = "0000:00:02.0",
+	.table = {
+		GPIO_LOOKUP("INT33FC:01", 10, "backlight", GPIO_ACTIVE_HIGH),
+		GPIO_LOOKUP("INT33FC:01", 11, "panel", GPIO_ACTIVE_HIGH),
+		{ }
+	},
+};
+
+static const struct pinctrl_map soc_pwm_pinctrl_map[] = {
+	PIN_MAP_MUX_GROUP("0000:00:02.0", "soc_pwm0", "INT33FC:00",
+			  "pwm0_grp", "pwm"),
+};
+
+void intel_dsi_vbt_gpio_init(struct intel_dsi *intel_dsi, bool panel_is_on)
+{
+	struct drm_device *dev = intel_dsi->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_connector *connector = intel_dsi->attached_connector;
+	struct mipi_config *mipi_config = connector->panel.vbt.dsi.config;
+	enum gpiod_flags flags = panel_is_on ? GPIOD_OUT_HIGH : GPIOD_OUT_LOW;
+	bool want_backlight_gpio = false;
+	bool want_panel_gpio = false;
+	struct pinctrl *pinctrl;
+	int ret;
+
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    mipi_config->pwm_blc == PPS_BLC_PMIC) {
+		gpiod_add_lookup_table(&pmic_panel_gpio_table);
+		want_panel_gpio = true;
+	}
+
+	if (IS_VALLEYVIEW(dev_priv) && mipi_config->pwm_blc == PPS_BLC_SOC) {
+		gpiod_add_lookup_table(&soc_panel_gpio_table);
+		want_panel_gpio = true;
+		want_backlight_gpio = true;
+
+		/* Ensure PWM0 pin is muxed as PWM instead of GPIO */
+		ret = pinctrl_register_mappings(soc_pwm_pinctrl_map,
+					     ARRAY_SIZE(soc_pwm_pinctrl_map));
+		if (ret)
+			drm_err(&dev_priv->drm,
+				"Failed to register pwm0 pinmux mapping\n");
+
+		pinctrl = devm_pinctrl_get_select(dev->dev, "soc_pwm0");
+		if (IS_ERR(pinctrl))
+			drm_err(&dev_priv->drm,
+				"Failed to set pinmux to PWM\n");
+	}
+
+	if (want_panel_gpio) {
+		intel_dsi->gpio_panel = gpiod_get(dev->dev, "panel", flags);
+		if (IS_ERR(intel_dsi->gpio_panel)) {
+			drm_err(&dev_priv->drm,
+				"Failed to own gpio for panel control\n");
+			intel_dsi->gpio_panel = NULL;
+		}
+	}
+
+	if (want_backlight_gpio) {
+		intel_dsi->gpio_backlight =
+			gpiod_get(dev->dev, "backlight", flags);
+		if (IS_ERR(intel_dsi->gpio_backlight)) {
+			drm_err(&dev_priv->drm,
+				"Failed to own gpio for backlight control\n");
+			intel_dsi->gpio_backlight = NULL;
+		}
+	}
+}
+
+void intel_dsi_vbt_gpio_cleanup(struct intel_dsi *intel_dsi)
+{
+	struct drm_device *dev = intel_dsi->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_connector *connector = intel_dsi->attached_connector;
+	struct mipi_config *mipi_config = connector->panel.vbt.dsi.config;
+
+	if (intel_dsi->gpio_panel) {
+		gpiod_put(intel_dsi->gpio_panel);
+		intel_dsi->gpio_panel = NULL;
+	}
+
+	if (intel_dsi->gpio_backlight) {
+		gpiod_put(intel_dsi->gpio_backlight);
+		intel_dsi->gpio_backlight = NULL;
+	}
+
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    mipi_config->pwm_blc == PPS_BLC_PMIC)
+		gpiod_remove_lookup_table(&pmic_panel_gpio_table);
+
+	if (IS_VALLEYVIEW(dev_priv) && mipi_config->pwm_blc == PPS_BLC_SOC) {
+		pinctrl_unregister_mappings(soc_pwm_pinctrl_map);
+		gpiod_remove_lookup_table(&soc_panel_gpio_table);
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dsi_vbt.h b/drivers/gpu/drm/i915/display/intel_dsi_vbt.h
new file mode 100644
index 000000000..468d873fa
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dsi_vbt.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __INTEL_DSI_VBT_H__
+#define __INTEL_DSI_VBT_H__
+
+#include <linux/types.h>
+
+enum mipi_seq;
+struct intel_dsi;
+
+bool intel_dsi_vbt_init(struct intel_dsi *intel_dsi, u16 panel_id);
+void intel_dsi_vbt_gpio_init(struct intel_dsi *intel_dsi, bool panel_is_on);
+void intel_dsi_vbt_gpio_cleanup(struct intel_dsi *intel_dsi);
+void intel_dsi_vbt_exec_sequence(struct intel_dsi *intel_dsi,
+				 enum mipi_seq seq_id);
+void intel_dsi_log_params(struct intel_dsi *intel_dsi);
+
+#endif /* __INTEL_DSI_VBT_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dvo.c b/drivers/gpu/drm/i915/display/intel_dvo.c
new file mode 100644
index 000000000..511c58907
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dvo.c
@@ -0,0 +1,550 @@
+/*
+ * Copyright 2006 Dave Airlie <airlied@linux.ie>
+ * Copyright © 2006-2007 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Eric Anholt <eric@anholt.net>
+ */
+
+#include <linux/i2c.h>
+#include <linux/slab.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc.h>
+
+#include "i915_drv.h"
+#include "intel_connector.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dvo.h"
+#include "intel_dvo_dev.h"
+#include "intel_gmbus.h"
+#include "intel_panel.h"
+
+#define INTEL_DVO_CHIP_NONE	0
+#define INTEL_DVO_CHIP_LVDS	1
+#define INTEL_DVO_CHIP_TMDS	2
+#define INTEL_DVO_CHIP_TVOUT	4
+#define INTEL_DVO_CHIP_LVDS_NO_FIXED	5
+
+#define SIL164_ADDR	0x38
+#define CH7xxx_ADDR	0x76
+#define TFP410_ADDR	0x38
+#define NS2501_ADDR     0x38
+
+static const struct intel_dvo_device intel_dvo_devices[] = {
+	{
+		.type = INTEL_DVO_CHIP_TMDS,
+		.name = "sil164",
+		.dvo_reg = DVOC,
+		.dvo_srcdim_reg = DVOC_SRCDIM,
+		.slave_addr = SIL164_ADDR,
+		.dev_ops = &sil164_ops,
+	},
+	{
+		.type = INTEL_DVO_CHIP_TMDS,
+		.name = "ch7xxx",
+		.dvo_reg = DVOC,
+		.dvo_srcdim_reg = DVOC_SRCDIM,
+		.slave_addr = CH7xxx_ADDR,
+		.dev_ops = &ch7xxx_ops,
+	},
+	{
+		.type = INTEL_DVO_CHIP_TMDS,
+		.name = "ch7xxx",
+		.dvo_reg = DVOC,
+		.dvo_srcdim_reg = DVOC_SRCDIM,
+		.slave_addr = 0x75, /* For some ch7010 */
+		.dev_ops = &ch7xxx_ops,
+	},
+	{
+		.type = INTEL_DVO_CHIP_LVDS,
+		.name = "ivch",
+		.dvo_reg = DVOA,
+		.dvo_srcdim_reg = DVOA_SRCDIM,
+		.slave_addr = 0x02, /* Might also be 0x44, 0x84, 0xc4 */
+		.dev_ops = &ivch_ops,
+	},
+	{
+		.type = INTEL_DVO_CHIP_TMDS,
+		.name = "tfp410",
+		.dvo_reg = DVOC,
+		.dvo_srcdim_reg = DVOC_SRCDIM,
+		.slave_addr = TFP410_ADDR,
+		.dev_ops = &tfp410_ops,
+	},
+	{
+		.type = INTEL_DVO_CHIP_LVDS,
+		.name = "ch7017",
+		.dvo_reg = DVOC,
+		.dvo_srcdim_reg = DVOC_SRCDIM,
+		.slave_addr = 0x75,
+		.gpio = GMBUS_PIN_DPB,
+		.dev_ops = &ch7017_ops,
+	},
+	{
+		.type = INTEL_DVO_CHIP_LVDS_NO_FIXED,
+		.name = "ns2501",
+		.dvo_reg = DVOB,
+		.dvo_srcdim_reg = DVOB_SRCDIM,
+		.slave_addr = NS2501_ADDR,
+		.dev_ops = &ns2501_ops,
+	},
+};
+
+struct intel_dvo {
+	struct intel_encoder base;
+
+	struct intel_dvo_device dev;
+
+	struct intel_connector *attached_connector;
+
+	bool panel_wants_dither;
+};
+
+static struct intel_dvo *enc_to_dvo(struct intel_encoder *encoder)
+{
+	return container_of(encoder, struct intel_dvo, base);
+}
+
+static struct intel_dvo *intel_attached_dvo(struct intel_connector *connector)
+{
+	return enc_to_dvo(intel_attached_encoder(connector));
+}
+
+static bool intel_dvo_connector_get_hw_state(struct intel_connector *connector)
+{
+	struct drm_device *dev = connector->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_dvo *intel_dvo = intel_attached_dvo(connector);
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, intel_dvo->dev.dvo_reg);
+
+	if (!(tmp & DVO_ENABLE))
+		return false;
+
+	return intel_dvo->dev.dev_ops->get_hw_state(&intel_dvo->dev);
+}
+
+static bool intel_dvo_get_hw_state(struct intel_encoder *encoder,
+				   enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, intel_dvo->dev.dvo_reg);
+
+	*pipe = (tmp & DVO_PIPE_SEL_MASK) >> DVO_PIPE_SEL_SHIFT;
+
+	return tmp & DVO_ENABLE;
+}
+
+static void intel_dvo_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
+	u32 tmp, flags = 0;
+
+	pipe_config->output_types |= BIT(INTEL_OUTPUT_DVO);
+
+	tmp = intel_de_read(dev_priv, intel_dvo->dev.dvo_reg);
+	if (tmp & DVO_HSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+	if (tmp & DVO_VSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	pipe_config->hw.adjusted_mode.flags |= flags;
+
+	pipe_config->hw.adjusted_mode.crtc_clock = pipe_config->port_clock;
+}
+
+static void intel_disable_dvo(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *old_crtc_state,
+			      const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
+	i915_reg_t dvo_reg = intel_dvo->dev.dvo_reg;
+	u32 temp = intel_de_read(dev_priv, dvo_reg);
+
+	intel_dvo->dev.dev_ops->dpms(&intel_dvo->dev, false);
+	intel_de_write(dev_priv, dvo_reg, temp & ~DVO_ENABLE);
+	intel_de_read(dev_priv, dvo_reg);
+}
+
+static void intel_enable_dvo(struct intel_atomic_state *state,
+			     struct intel_encoder *encoder,
+			     const struct intel_crtc_state *pipe_config,
+			     const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
+	i915_reg_t dvo_reg = intel_dvo->dev.dvo_reg;
+	u32 temp = intel_de_read(dev_priv, dvo_reg);
+
+	intel_dvo->dev.dev_ops->mode_set(&intel_dvo->dev,
+					 &pipe_config->hw.mode,
+					 &pipe_config->hw.adjusted_mode);
+
+	intel_de_write(dev_priv, dvo_reg, temp | DVO_ENABLE);
+	intel_de_read(dev_priv, dvo_reg);
+
+	intel_dvo->dev.dev_ops->dpms(&intel_dvo->dev, true);
+}
+
+static enum drm_mode_status
+intel_dvo_mode_valid(struct drm_connector *connector,
+		     struct drm_display_mode *mode)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct intel_dvo *intel_dvo = intel_attached_dvo(intel_connector);
+	struct drm_i915_private *i915 = to_i915(intel_connector->base.dev);
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_fixed_mode(intel_connector, mode);
+	int max_dotclk = to_i915(connector->dev)->max_dotclk_freq;
+	int target_clock = mode->clock;
+	enum drm_mode_status status;
+
+	status = intel_cpu_transcoder_mode_valid(i915, mode);
+	if (status != MODE_OK)
+		return status;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return MODE_NO_DBLESCAN;
+
+	/* XXX: Validate clock range */
+
+	if (fixed_mode) {
+		enum drm_mode_status status;
+
+		status = intel_panel_mode_valid(intel_connector, mode);
+		if (status != MODE_OK)
+			return status;
+
+		target_clock = fixed_mode->clock;
+	}
+
+	if (target_clock > max_dotclk)
+		return MODE_CLOCK_HIGH;
+
+	return intel_dvo->dev.dev_ops->mode_valid(&intel_dvo->dev, mode);
+}
+
+static int intel_dvo_compute_config(struct intel_encoder *encoder,
+				    struct intel_crtc_state *pipe_config,
+				    struct drm_connector_state *conn_state)
+{
+	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_fixed_mode(intel_dvo->attached_connector, adjusted_mode);
+
+	/*
+	 * If we have timings from the BIOS for the panel, put them in
+	 * to the adjusted mode.  The CRTC will be set up for this mode,
+	 * with the panel scaling set up to source from the H/VDisplay
+	 * of the original mode.
+	 */
+	if (fixed_mode) {
+		int ret;
+
+		ret = intel_panel_compute_config(connector, adjusted_mode);
+		if (ret)
+			return ret;
+	}
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+
+	return 0;
+}
+
+static void intel_dvo_pre_enable(struct intel_atomic_state *state,
+				 struct intel_encoder *encoder,
+				 const struct intel_crtc_state *pipe_config,
+				 const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	const struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	struct intel_dvo *intel_dvo = enc_to_dvo(encoder);
+	enum pipe pipe = crtc->pipe;
+	u32 dvo_val;
+	i915_reg_t dvo_reg = intel_dvo->dev.dvo_reg;
+	i915_reg_t dvo_srcdim_reg = intel_dvo->dev.dvo_srcdim_reg;
+
+	/* Save the data order, since I don't know what it should be set to. */
+	dvo_val = intel_de_read(dev_priv, dvo_reg) &
+		  (DVO_PRESERVE_MASK | DVO_DATA_ORDER_GBRG);
+	dvo_val |= DVO_DATA_ORDER_FP | DVO_BORDER_ENABLE |
+		   DVO_BLANK_ACTIVE_HIGH;
+
+	dvo_val |= DVO_PIPE_SEL(pipe);
+	dvo_val |= DVO_PIPE_STALL;
+	if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
+		dvo_val |= DVO_HSYNC_ACTIVE_HIGH;
+	if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
+		dvo_val |= DVO_VSYNC_ACTIVE_HIGH;
+
+	intel_de_write(dev_priv, dvo_srcdim_reg,
+		       (adjusted_mode->crtc_hdisplay << DVO_SRCDIM_HORIZONTAL_SHIFT) | (adjusted_mode->crtc_vdisplay << DVO_SRCDIM_VERTICAL_SHIFT));
+	intel_de_write(dev_priv, dvo_reg, dvo_val);
+}
+
+static enum drm_connector_status
+intel_dvo_detect(struct drm_connector *connector, bool force)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct intel_dvo *intel_dvo = intel_attached_dvo(to_intel_connector(connector));
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	if (!INTEL_DISPLAY_ENABLED(i915))
+		return connector_status_disconnected;
+
+	return intel_dvo->dev.dev_ops->detect(&intel_dvo->dev);
+}
+
+static int intel_dvo_get_modes(struct drm_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	int num_modes;
+
+	/*
+	 * We should probably have an i2c driver get_modes function for those
+	 * devices which will have a fixed set of modes determined by the chip
+	 * (TV-out, for example), but for now with just TMDS and LVDS,
+	 * that's not the case.
+	 */
+	num_modes = intel_ddc_get_modes(connector,
+					intel_gmbus_get_adapter(dev_priv, GMBUS_PIN_DPC));
+	if (num_modes)
+		return num_modes;
+
+	return intel_panel_get_modes(to_intel_connector(connector));
+}
+
+static const struct drm_connector_funcs intel_dvo_connector_funcs = {
+	.detect = intel_dvo_detect,
+	.late_register = intel_connector_register,
+	.early_unregister = intel_connector_unregister,
+	.destroy = intel_connector_destroy,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+};
+
+static const struct drm_connector_helper_funcs intel_dvo_connector_helper_funcs = {
+	.mode_valid = intel_dvo_mode_valid,
+	.get_modes = intel_dvo_get_modes,
+};
+
+static void intel_dvo_enc_destroy(struct drm_encoder *encoder)
+{
+	struct intel_dvo *intel_dvo = enc_to_dvo(to_intel_encoder(encoder));
+
+	if (intel_dvo->dev.dev_ops->destroy)
+		intel_dvo->dev.dev_ops->destroy(&intel_dvo->dev);
+
+	intel_encoder_destroy(encoder);
+}
+
+static const struct drm_encoder_funcs intel_dvo_enc_funcs = {
+	.destroy = intel_dvo_enc_destroy,
+};
+
+static enum port intel_dvo_port(i915_reg_t dvo_reg)
+{
+	if (i915_mmio_reg_equal(dvo_reg, DVOA))
+		return PORT_A;
+	else if (i915_mmio_reg_equal(dvo_reg, DVOB))
+		return PORT_B;
+	else
+		return PORT_C;
+}
+
+void intel_dvo_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *intel_encoder;
+	struct intel_dvo *intel_dvo;
+	struct intel_connector *intel_connector;
+	int i;
+	int encoder_type = DRM_MODE_ENCODER_NONE;
+
+	intel_dvo = kzalloc(sizeof(*intel_dvo), GFP_KERNEL);
+	if (!intel_dvo)
+		return;
+
+	intel_connector = intel_connector_alloc();
+	if (!intel_connector) {
+		kfree(intel_dvo);
+		return;
+	}
+
+	intel_dvo->attached_connector = intel_connector;
+
+	intel_encoder = &intel_dvo->base;
+
+	intel_encoder->disable = intel_disable_dvo;
+	intel_encoder->enable = intel_enable_dvo;
+	intel_encoder->get_hw_state = intel_dvo_get_hw_state;
+	intel_encoder->get_config = intel_dvo_get_config;
+	intel_encoder->compute_config = intel_dvo_compute_config;
+	intel_encoder->pre_enable = intel_dvo_pre_enable;
+	intel_connector->get_hw_state = intel_dvo_connector_get_hw_state;
+
+	/* Now, try to find a controller */
+	for (i = 0; i < ARRAY_SIZE(intel_dvo_devices); i++) {
+		struct drm_connector *connector = &intel_connector->base;
+		const struct intel_dvo_device *dvo = &intel_dvo_devices[i];
+		struct i2c_adapter *i2c;
+		int gpio;
+		bool dvoinit;
+		enum pipe pipe;
+		u32 dpll[I915_MAX_PIPES];
+		enum port port;
+
+		/*
+		 * Allow the I2C driver info to specify the GPIO to be used in
+		 * special cases, but otherwise default to what's defined
+		 * in the spec.
+		 */
+		if (intel_gmbus_is_valid_pin(dev_priv, dvo->gpio))
+			gpio = dvo->gpio;
+		else if (dvo->type == INTEL_DVO_CHIP_LVDS)
+			gpio = GMBUS_PIN_SSC;
+		else
+			gpio = GMBUS_PIN_DPB;
+
+		/*
+		 * Set up the I2C bus necessary for the chip we're probing.
+		 * It appears that everything is on GPIOE except for panels
+		 * on i830 laptops, which are on GPIOB (DVOA).
+		 */
+		i2c = intel_gmbus_get_adapter(dev_priv, gpio);
+
+		intel_dvo->dev = *dvo;
+
+		/*
+		 * GMBUS NAK handling seems to be unstable, hence let the
+		 * transmitter detection run in bit banging mode for now.
+		 */
+		intel_gmbus_force_bit(i2c, true);
+
+		/*
+		 * ns2501 requires the DVO 2x clock before it will
+		 * respond to i2c accesses, so make sure we have
+		 * have the clock enabled before we attempt to
+		 * initialize the device.
+		 */
+		for_each_pipe(dev_priv, pipe) {
+			dpll[pipe] = intel_de_read(dev_priv, DPLL(pipe));
+			intel_de_write(dev_priv, DPLL(pipe),
+				       dpll[pipe] | DPLL_DVO_2X_MODE);
+		}
+
+		dvoinit = dvo->dev_ops->init(&intel_dvo->dev, i2c);
+
+		/* restore the DVO 2x clock state to original */
+		for_each_pipe(dev_priv, pipe) {
+			intel_de_write(dev_priv, DPLL(pipe), dpll[pipe]);
+		}
+
+		intel_gmbus_force_bit(i2c, false);
+
+		if (!dvoinit)
+			continue;
+
+		port = intel_dvo_port(dvo->dvo_reg);
+		drm_encoder_init(&dev_priv->drm, &intel_encoder->base,
+				 &intel_dvo_enc_funcs, encoder_type,
+				 "DVO %c", port_name(port));
+
+		intel_encoder->type = INTEL_OUTPUT_DVO;
+		intel_encoder->power_domain = POWER_DOMAIN_PORT_OTHER;
+		intel_encoder->port = port;
+		intel_encoder->pipe_mask = ~0;
+
+		if (dvo->type != INTEL_DVO_CHIP_LVDS)
+			intel_encoder->cloneable = (1 << INTEL_OUTPUT_ANALOG) |
+				(1 << INTEL_OUTPUT_DVO);
+
+		switch (dvo->type) {
+		case INTEL_DVO_CHIP_TMDS:
+			intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT |
+				DRM_CONNECTOR_POLL_DISCONNECT;
+			drm_connector_init(&dev_priv->drm, connector,
+					   &intel_dvo_connector_funcs,
+					   DRM_MODE_CONNECTOR_DVII);
+			encoder_type = DRM_MODE_ENCODER_TMDS;
+			break;
+		case INTEL_DVO_CHIP_LVDS_NO_FIXED:
+		case INTEL_DVO_CHIP_LVDS:
+			drm_connector_init(&dev_priv->drm, connector,
+					   &intel_dvo_connector_funcs,
+					   DRM_MODE_CONNECTOR_LVDS);
+			encoder_type = DRM_MODE_ENCODER_LVDS;
+			break;
+		}
+
+		drm_connector_helper_add(connector,
+					 &intel_dvo_connector_helper_funcs);
+		connector->display_info.subpixel_order = SubPixelHorizontalRGB;
+		connector->interlace_allowed = false;
+		connector->doublescan_allowed = false;
+
+		intel_connector_attach_encoder(intel_connector, intel_encoder);
+		if (dvo->type == INTEL_DVO_CHIP_LVDS) {
+			/*
+			 * For our LVDS chipsets, we should hopefully be able
+			 * to dig the fixed panel mode out of the BIOS data.
+			 * However, it's in a different format from the BIOS
+			 * data on chipsets with integrated LVDS (stored in AIM
+			 * headers, likely), so for now, just get the current
+			 * mode being output through DVO.
+			 */
+			intel_panel_add_encoder_fixed_mode(intel_connector,
+							   intel_encoder);
+
+			intel_panel_init(intel_connector);
+
+			intel_dvo->panel_wants_dither = true;
+		}
+
+		return;
+	}
+
+	kfree(intel_dvo);
+	kfree(intel_connector);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_dvo.h b/drivers/gpu/drm/i915/display/intel_dvo.h
new file mode 100644
index 000000000..3ed0fdf8e
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dvo.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_DVO_H__
+#define __INTEL_DVO_H__
+
+struct drm_i915_private;
+
+void intel_dvo_init(struct drm_i915_private *dev_priv);
+
+#endif /* __INTEL_DVO_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_dvo_dev.h b/drivers/gpu/drm/i915/display/intel_dvo_dev.h
new file mode 100644
index 000000000..50205f064
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_dvo_dev.h
@@ -0,0 +1,140 @@
+/*
+ * Copyright © 2006 Eric Anholt
+ *
+ * Permission to use, copy, modify, distribute, and sell this software and its
+ * documentation for any purpose is hereby granted without fee, provided that
+ * the above copyright notice appear in all copies and that both that copyright
+ * notice and this permission notice appear in supporting documentation, and
+ * that the name of the copyright holders not be used in advertising or
+ * publicity pertaining to distribution of the software without specific,
+ * written prior permission.  The copyright holders make no representations
+ * about the suitability of this software for any purpose.  It is provided "as
+ * is" without express or implied warranty.
+ *
+ * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
+ * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
+ * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
+ * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
+ * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
+ * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
+ * OF THIS SOFTWARE.
+ */
+
+#ifndef __INTEL_DVO_DEV_H__
+#define __INTEL_DVO_DEV_H__
+
+#include <linux/i2c.h>
+
+#include <drm/drm_crtc.h>
+
+#include "i915_reg_defs.h"
+
+struct intel_dvo_device {
+	const char *name;
+	int type;
+	/* DVOA/B/C output register */
+	i915_reg_t dvo_reg;
+	i915_reg_t dvo_srcdim_reg;
+	/* GPIO register used for i2c bus to control this device */
+	u32 gpio;
+	int slave_addr;
+
+	const struct intel_dvo_dev_ops *dev_ops;
+	void *dev_priv;
+	struct i2c_adapter *i2c_bus;
+};
+
+struct intel_dvo_dev_ops {
+	/*
+	 * Initialize the device at startup time.
+	 * Returns NULL if the device does not exist.
+	 */
+	bool (*init)(struct intel_dvo_device *dvo,
+		     struct i2c_adapter *i2cbus);
+
+	/*
+	 * Called to allow the output a chance to create properties after the
+	 * RandR objects have been created.
+	 */
+	void (*create_resources)(struct intel_dvo_device *dvo);
+
+	/*
+	 * Turn on/off output.
+	 *
+	 * Because none of our dvo drivers support an intermediate power levels,
+	 * we don't expose this in the interfac.
+	 */
+	void (*dpms)(struct intel_dvo_device *dvo, bool enable);
+
+	/*
+	 * Callback for testing a video mode for a given output.
+	 *
+	 * This function should only check for cases where a mode can't
+	 * be supported on the output specifically, and not represent
+	 * generic CRTC limitations.
+	 *
+	 * \return MODE_OK if the mode is valid, or another MODE_* otherwise.
+	 */
+	enum drm_mode_status (*mode_valid)(struct intel_dvo_device *dvo,
+					   struct drm_display_mode *mode);
+
+	/*
+	 * Callback for preparing mode changes on an output
+	 */
+	void (*prepare)(struct intel_dvo_device *dvo);
+
+	/*
+	 * Callback for committing mode changes on an output
+	 */
+	void (*commit)(struct intel_dvo_device *dvo);
+
+	/*
+	 * Callback for setting up a video mode after fixups have been made.
+	 *
+	 * This is only called while the output is disabled.  The dpms callback
+	 * must be all that's necessary for the output, to turn the output on
+	 * after this function is called.
+	 */
+	void (*mode_set)(struct intel_dvo_device *dvo,
+			 const struct drm_display_mode *mode,
+			 const struct drm_display_mode *adjusted_mode);
+
+	/*
+	 * Probe for a connected output, and return detect_status.
+	 */
+	enum drm_connector_status (*detect)(struct intel_dvo_device *dvo);
+
+	/*
+	 * Probe the current hw status, returning true if the connected output
+	 * is active.
+	 */
+	bool (*get_hw_state)(struct intel_dvo_device *dev);
+
+	/**
+	 * Query the device for the modes it provides.
+	 *
+	 * This function may also update MonInfo, mm_width, and mm_height.
+	 *
+	 * \return singly-linked list of modes or NULL if no modes found.
+	 */
+	struct drm_display_mode *(*get_modes)(struct intel_dvo_device *dvo);
+
+	/**
+	 * Clean up driver-specific bits of the output
+	 */
+	void (*destroy) (struct intel_dvo_device *dvo);
+
+	/**
+	 * Debugging hook to dump device registers to log file
+	 */
+	void (*dump_regs)(struct intel_dvo_device *dvo);
+};
+
+extern const struct intel_dvo_dev_ops sil164_ops;
+extern const struct intel_dvo_dev_ops ch7xxx_ops;
+extern const struct intel_dvo_dev_ops ivch_ops;
+extern const struct intel_dvo_dev_ops tfp410_ops;
+extern const struct intel_dvo_dev_ops ch7017_ops;
+extern const struct intel_dvo_dev_ops ns2501_ops;
+
+#endif /* __INTEL_DVO_DEV_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_fb.c b/drivers/gpu/drm/i915/display/intel_fb.c
new file mode 100644
index 000000000..c69a63879
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fb.c
@@ -0,0 +1,2081 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <drm/drm_blend.h>
+#include <drm/drm_framebuffer.h>
+#include <drm/drm_modeset_helper.h>
+
+#include "i915_drv.h"
+#include "intel_display.h"
+#include "intel_display_types.h"
+#include "intel_dpt.h"
+#include "intel_fb.h"
+
+#define check_array_bounds(i915, a, i) drm_WARN_ON(&(i915)->drm, (i) >= ARRAY_SIZE(a))
+
+/*
+ * From the Sky Lake PRM:
+ * "The Color Control Surface (CCS) contains the compression status of
+ *  the cache-line pairs. The compression state of the cache-line pair
+ *  is specified by 2 bits in the CCS. Each CCS cache-line represents
+ *  an area on the main surface of 16 x16 sets of 128 byte Y-tiled
+ *  cache-line-pairs. CCS is always Y tiled."
+ *
+ * Since cache line pairs refers to horizontally adjacent cache lines,
+ * each cache line in the CCS corresponds to an area of 32x16 cache
+ * lines on the main surface. Since each pixel is 4 bytes, this gives
+ * us a ratio of one byte in the CCS for each 8x16 pixels in the
+ * main surface.
+ */
+static const struct drm_format_info skl_ccs_formats[] = {
+	{ .format = DRM_FORMAT_XRGB8888, .depth = 24, .num_planes = 2,
+	  .cpp = { 4, 1, }, .hsub = 8, .vsub = 16, },
+	{ .format = DRM_FORMAT_XBGR8888, .depth = 24, .num_planes = 2,
+	  .cpp = { 4, 1, }, .hsub = 8, .vsub = 16, },
+	{ .format = DRM_FORMAT_ARGB8888, .depth = 32, .num_planes = 2,
+	  .cpp = { 4, 1, }, .hsub = 8, .vsub = 16, .has_alpha = true, },
+	{ .format = DRM_FORMAT_ABGR8888, .depth = 32, .num_planes = 2,
+	  .cpp = { 4, 1, }, .hsub = 8, .vsub = 16, .has_alpha = true, },
+};
+
+/*
+ * Gen-12 compression uses 4 bits of CCS data for each cache line pair in the
+ * main surface. And each 64B CCS cache line represents an area of 4x1 Y-tiles
+ * in the main surface. With 4 byte pixels and each Y-tile having dimensions of
+ * 32x32 pixels, the ratio turns out to 1B in the CCS for every 2x32 pixels in
+ * the main surface.
+ */
+static const struct drm_format_info gen12_ccs_formats[] = {
+	{ .format = DRM_FORMAT_XRGB8888, .depth = 24, .num_planes = 2,
+	  .char_per_block = { 4, 1 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 1, .vsub = 1, },
+	{ .format = DRM_FORMAT_XBGR8888, .depth = 24, .num_planes = 2,
+	  .char_per_block = { 4, 1 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 1, .vsub = 1, },
+	{ .format = DRM_FORMAT_ARGB8888, .depth = 32, .num_planes = 2,
+	  .char_per_block = { 4, 1 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 1, .vsub = 1, .has_alpha = true },
+	{ .format = DRM_FORMAT_ABGR8888, .depth = 32, .num_planes = 2,
+	  .char_per_block = { 4, 1 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 1, .vsub = 1, .has_alpha = true },
+	{ .format = DRM_FORMAT_YUYV, .num_planes = 2,
+	  .char_per_block = { 2, 1 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 2, .vsub = 1, .is_yuv = true },
+	{ .format = DRM_FORMAT_YVYU, .num_planes = 2,
+	  .char_per_block = { 2, 1 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 2, .vsub = 1, .is_yuv = true },
+	{ .format = DRM_FORMAT_UYVY, .num_planes = 2,
+	  .char_per_block = { 2, 1 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 2, .vsub = 1, .is_yuv = true },
+	{ .format = DRM_FORMAT_VYUY, .num_planes = 2,
+	  .char_per_block = { 2, 1 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 2, .vsub = 1, .is_yuv = true },
+	{ .format = DRM_FORMAT_XYUV8888, .num_planes = 2,
+	  .char_per_block = { 4, 1 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 1, .vsub = 1, .is_yuv = true },
+	{ .format = DRM_FORMAT_NV12, .num_planes = 4,
+	  .char_per_block = { 1, 2, 1, 1 }, .block_w = { 1, 1, 4, 4 }, .block_h = { 1, 1, 1, 1 },
+	  .hsub = 2, .vsub = 2, .is_yuv = true },
+	{ .format = DRM_FORMAT_P010, .num_planes = 4,
+	  .char_per_block = { 2, 4, 1, 1 }, .block_w = { 1, 1, 2, 2 }, .block_h = { 1, 1, 1, 1 },
+	  .hsub = 2, .vsub = 2, .is_yuv = true },
+	{ .format = DRM_FORMAT_P012, .num_planes = 4,
+	  .char_per_block = { 2, 4, 1, 1 }, .block_w = { 1, 1, 2, 2 }, .block_h = { 1, 1, 1, 1 },
+	  .hsub = 2, .vsub = 2, .is_yuv = true },
+	{ .format = DRM_FORMAT_P016, .num_planes = 4,
+	  .char_per_block = { 2, 4, 1, 1 }, .block_w = { 1, 1, 2, 2 }, .block_h = { 1, 1, 1, 1 },
+	  .hsub = 2, .vsub = 2, .is_yuv = true },
+};
+
+/*
+ * Same as gen12_ccs_formats[] above, but with additional surface used
+ * to pass Clear Color information in plane 2 with 64 bits of data.
+ */
+static const struct drm_format_info gen12_ccs_cc_formats[] = {
+	{ .format = DRM_FORMAT_XRGB8888, .depth = 24, .num_planes = 3,
+	  .char_per_block = { 4, 1, 0 }, .block_w = { 1, 2, 2 }, .block_h = { 1, 1, 1 },
+	  .hsub = 1, .vsub = 1, },
+	{ .format = DRM_FORMAT_XBGR8888, .depth = 24, .num_planes = 3,
+	  .char_per_block = { 4, 1, 0 }, .block_w = { 1, 2, 2 }, .block_h = { 1, 1, 1 },
+	  .hsub = 1, .vsub = 1, },
+	{ .format = DRM_FORMAT_ARGB8888, .depth = 32, .num_planes = 3,
+	  .char_per_block = { 4, 1, 0 }, .block_w = { 1, 2, 2 }, .block_h = { 1, 1, 1 },
+	  .hsub = 1, .vsub = 1, .has_alpha = true },
+	{ .format = DRM_FORMAT_ABGR8888, .depth = 32, .num_planes = 3,
+	  .char_per_block = { 4, 1, 0 }, .block_w = { 1, 2, 2 }, .block_h = { 1, 1, 1 },
+	  .hsub = 1, .vsub = 1, .has_alpha = true },
+};
+
+static const struct drm_format_info gen12_flat_ccs_cc_formats[] = {
+	{ .format = DRM_FORMAT_XRGB8888, .depth = 24, .num_planes = 2,
+	  .char_per_block = { 4, 0 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 1, .vsub = 1, },
+	{ .format = DRM_FORMAT_XBGR8888, .depth = 24, .num_planes = 2,
+	  .char_per_block = { 4, 0 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 1, .vsub = 1, },
+	{ .format = DRM_FORMAT_ARGB8888, .depth = 32, .num_planes = 2,
+	  .char_per_block = { 4, 0 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 1, .vsub = 1, .has_alpha = true },
+	{ .format = DRM_FORMAT_ABGR8888, .depth = 32, .num_planes = 2,
+	  .char_per_block = { 4, 0 }, .block_w = { 1, 2 }, .block_h = { 1, 1 },
+	  .hsub = 1, .vsub = 1, .has_alpha = true },
+};
+
+struct intel_modifier_desc {
+	u64 modifier;
+	struct {
+		u8 from;
+		u8 until;
+	} display_ver;
+#define DISPLAY_VER_ALL		{ 0, -1 }
+
+	const struct drm_format_info *formats;
+	int format_count;
+#define FORMAT_OVERRIDE(format_list) \
+	.formats = format_list, \
+	.format_count = ARRAY_SIZE(format_list)
+
+	u8 plane_caps;
+
+	struct {
+		u8 cc_planes:3;
+		u8 packed_aux_planes:4;
+		u8 planar_aux_planes:4;
+	} ccs;
+};
+
+#define INTEL_PLANE_CAP_CCS_MASK	(INTEL_PLANE_CAP_CCS_RC | \
+					 INTEL_PLANE_CAP_CCS_RC_CC | \
+					 INTEL_PLANE_CAP_CCS_MC)
+#define INTEL_PLANE_CAP_TILING_MASK	(INTEL_PLANE_CAP_TILING_X | \
+					 INTEL_PLANE_CAP_TILING_Y | \
+					 INTEL_PLANE_CAP_TILING_Yf | \
+					 INTEL_PLANE_CAP_TILING_4)
+#define INTEL_PLANE_CAP_TILING_NONE	0
+
+static const struct intel_modifier_desc intel_modifiers[] = {
+	{
+		.modifier = I915_FORMAT_MOD_4_TILED_DG2_MC_CCS,
+		.display_ver = { 13, 13 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_4 | INTEL_PLANE_CAP_CCS_MC,
+	}, {
+		.modifier = I915_FORMAT_MOD_4_TILED_DG2_RC_CCS_CC,
+		.display_ver = { 13, 13 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_4 | INTEL_PLANE_CAP_CCS_RC_CC,
+
+		.ccs.cc_planes = BIT(1),
+
+		FORMAT_OVERRIDE(gen12_flat_ccs_cc_formats),
+	}, {
+		.modifier = I915_FORMAT_MOD_4_TILED_DG2_RC_CCS,
+		.display_ver = { 13, 13 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_4 | INTEL_PLANE_CAP_CCS_RC,
+	}, {
+		.modifier = I915_FORMAT_MOD_4_TILED,
+		.display_ver = { 13, 13 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_4,
+	}, {
+		.modifier = I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS,
+		.display_ver = { 12, 13 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_Y | INTEL_PLANE_CAP_CCS_MC,
+
+		.ccs.packed_aux_planes = BIT(1),
+		.ccs.planar_aux_planes = BIT(2) | BIT(3),
+
+		FORMAT_OVERRIDE(gen12_ccs_formats),
+	}, {
+		.modifier = I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS,
+		.display_ver = { 12, 13 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_Y | INTEL_PLANE_CAP_CCS_RC,
+
+		.ccs.packed_aux_planes = BIT(1),
+
+		FORMAT_OVERRIDE(gen12_ccs_formats),
+	}, {
+		.modifier = I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC,
+		.display_ver = { 12, 13 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_Y | INTEL_PLANE_CAP_CCS_RC_CC,
+
+		.ccs.cc_planes = BIT(2),
+		.ccs.packed_aux_planes = BIT(1),
+
+		FORMAT_OVERRIDE(gen12_ccs_cc_formats),
+	}, {
+		.modifier = I915_FORMAT_MOD_Yf_TILED_CCS,
+		.display_ver = { 9, 11 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_Yf | INTEL_PLANE_CAP_CCS_RC,
+
+		.ccs.packed_aux_planes = BIT(1),
+
+		FORMAT_OVERRIDE(skl_ccs_formats),
+	}, {
+		.modifier = I915_FORMAT_MOD_Y_TILED_CCS,
+		.display_ver = { 9, 11 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_Y | INTEL_PLANE_CAP_CCS_RC,
+
+		.ccs.packed_aux_planes = BIT(1),
+
+		FORMAT_OVERRIDE(skl_ccs_formats),
+	}, {
+		.modifier = I915_FORMAT_MOD_Yf_TILED,
+		.display_ver = { 9, 11 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_Yf,
+	}, {
+		.modifier = I915_FORMAT_MOD_Y_TILED,
+		.display_ver = { 9, 13 },
+		.plane_caps = INTEL_PLANE_CAP_TILING_Y,
+	}, {
+		.modifier = I915_FORMAT_MOD_X_TILED,
+		.display_ver = DISPLAY_VER_ALL,
+		.plane_caps = INTEL_PLANE_CAP_TILING_X,
+	}, {
+		.modifier = DRM_FORMAT_MOD_LINEAR,
+		.display_ver = DISPLAY_VER_ALL,
+	},
+};
+
+static const struct intel_modifier_desc *lookup_modifier_or_null(u64 modifier)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(intel_modifiers); i++)
+		if (intel_modifiers[i].modifier == modifier)
+			return &intel_modifiers[i];
+
+	return NULL;
+}
+
+static const struct intel_modifier_desc *lookup_modifier(u64 modifier)
+{
+	const struct intel_modifier_desc *md = lookup_modifier_or_null(modifier);
+
+	if (WARN_ON(!md))
+		return &intel_modifiers[0];
+
+	return md;
+}
+
+static const struct drm_format_info *
+lookup_format_info(const struct drm_format_info formats[],
+		   int num_formats, u32 format)
+{
+	int i;
+
+	for (i = 0; i < num_formats; i++) {
+		if (formats[i].format == format)
+			return &formats[i];
+	}
+
+	return NULL;
+}
+
+/**
+ * intel_fb_get_format_info: Get a modifier specific format information
+ * @cmd: FB add command structure
+ *
+ * Returns:
+ * Returns the format information for @cmd->pixel_format specific to @cmd->modifier[0],
+ * or %NULL if the modifier doesn't override the format.
+ */
+const struct drm_format_info *
+intel_fb_get_format_info(const struct drm_mode_fb_cmd2 *cmd)
+{
+	const struct intel_modifier_desc *md = lookup_modifier_or_null(cmd->modifier[0]);
+
+	if (!md || !md->formats)
+		return NULL;
+
+	return lookup_format_info(md->formats, md->format_count, cmd->pixel_format);
+}
+
+static bool plane_caps_contain_any(u8 caps, u8 mask)
+{
+	return caps & mask;
+}
+
+static bool plane_caps_contain_all(u8 caps, u8 mask)
+{
+	return (caps & mask) == mask;
+}
+
+/**
+ * intel_fb_is_ccs_modifier: Check if a modifier is a CCS modifier type
+ * @modifier: Modifier to check
+ *
+ * Returns:
+ * Returns %true if @modifier is a render, render with color clear or
+ * media compression modifier.
+ */
+bool intel_fb_is_ccs_modifier(u64 modifier)
+{
+	return plane_caps_contain_any(lookup_modifier(modifier)->plane_caps,
+				      INTEL_PLANE_CAP_CCS_MASK);
+}
+
+/**
+ * intel_fb_is_rc_ccs_cc_modifier: Check if a modifier is an RC CCS CC modifier type
+ * @modifier: Modifier to check
+ *
+ * Returns:
+ * Returns %true if @modifier is a render with color clear modifier.
+ */
+bool intel_fb_is_rc_ccs_cc_modifier(u64 modifier)
+{
+	return plane_caps_contain_any(lookup_modifier(modifier)->plane_caps,
+				      INTEL_PLANE_CAP_CCS_RC_CC);
+}
+
+/**
+ * intel_fb_is_mc_ccs_modifier: Check if a modifier is an MC CCS modifier type
+ * @modifier: Modifier to check
+ *
+ * Returns:
+ * Returns %true if @modifier is a media compression modifier.
+ */
+bool intel_fb_is_mc_ccs_modifier(u64 modifier)
+{
+	return plane_caps_contain_any(lookup_modifier(modifier)->plane_caps,
+				      INTEL_PLANE_CAP_CCS_MC);
+}
+
+static bool check_modifier_display_ver_range(const struct intel_modifier_desc *md,
+					     u8 display_ver_from, u8 display_ver_until)
+{
+	return md->display_ver.from <= display_ver_until &&
+		display_ver_from <= md->display_ver.until;
+}
+
+static bool plane_has_modifier(struct drm_i915_private *i915,
+			       u8 plane_caps,
+			       const struct intel_modifier_desc *md)
+{
+	if (!IS_DISPLAY_VER(i915, md->display_ver.from, md->display_ver.until))
+		return false;
+
+	if (!plane_caps_contain_all(plane_caps, md->plane_caps))
+		return false;
+
+	return true;
+}
+
+/**
+ * intel_fb_plane_get_modifiers: Get the modifiers for the given platform and plane capabilities
+ * @i915: i915 device instance
+ * @plane_caps: capabilities for the plane the modifiers are queried for
+ *
+ * Returns:
+ * Returns the list of modifiers allowed by the @i915 platform and @plane_caps.
+ * The caller must free the returned buffer.
+ */
+u64 *intel_fb_plane_get_modifiers(struct drm_i915_private *i915,
+				  u8 plane_caps)
+{
+	u64 *list, *p;
+	int count = 1;		/* +1 for invalid modifier terminator */
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(intel_modifiers); i++) {
+		if (plane_has_modifier(i915, plane_caps, &intel_modifiers[i]))
+			count++;
+	}
+
+	list = kmalloc_array(count, sizeof(*list), GFP_KERNEL);
+	if (drm_WARN_ON(&i915->drm, !list))
+		return NULL;
+
+	p = list;
+	for (i = 0; i < ARRAY_SIZE(intel_modifiers); i++) {
+		if (plane_has_modifier(i915, plane_caps, &intel_modifiers[i]))
+			*p++ = intel_modifiers[i].modifier;
+	}
+	*p++ = DRM_FORMAT_MOD_INVALID;
+
+	return list;
+}
+
+/**
+ * intel_fb_plane_supports_modifier: Determine if a modifier is supported by the given plane
+ * @plane: Plane to check the modifier support for
+ * @modifier: The modifier to check the support for
+ *
+ * Returns:
+ * %true if the @modifier is supported on @plane.
+ */
+bool intel_fb_plane_supports_modifier(struct intel_plane *plane, u64 modifier)
+{
+	int i;
+
+	for (i = 0; i < plane->base.modifier_count; i++)
+		if (plane->base.modifiers[i] == modifier)
+			return true;
+
+	return false;
+}
+
+static bool format_is_yuv_semiplanar(const struct intel_modifier_desc *md,
+				     const struct drm_format_info *info)
+{
+	if (!info->is_yuv)
+		return false;
+
+	if (hweight8(md->ccs.planar_aux_planes) == 2)
+		return info->num_planes == 4;
+	else
+		return info->num_planes == 2;
+}
+
+/**
+ * intel_format_info_is_yuv_semiplanar: Check if the given format is YUV semiplanar
+ * @info: format to check
+ * @modifier: modifier used with the format
+ *
+ * Returns:
+ * %true if @info / @modifier is YUV semiplanar.
+ */
+bool intel_format_info_is_yuv_semiplanar(const struct drm_format_info *info,
+					 u64 modifier)
+{
+	return format_is_yuv_semiplanar(lookup_modifier(modifier), info);
+}
+
+static u8 ccs_aux_plane_mask(const struct intel_modifier_desc *md,
+			     const struct drm_format_info *format)
+{
+	if (format_is_yuv_semiplanar(md, format))
+		return md->ccs.planar_aux_planes;
+	else
+		return md->ccs.packed_aux_planes;
+}
+
+/**
+ * intel_fb_is_ccs_aux_plane: Check if a framebuffer color plane is a CCS AUX plane
+ * @fb: Framebuffer
+ * @color_plane: color plane index to check
+ *
+ * Returns:
+ * Returns %true if @fb's color plane at index @color_plane is a CCS AUX plane.
+ */
+bool intel_fb_is_ccs_aux_plane(const struct drm_framebuffer *fb, int color_plane)
+{
+	const struct intel_modifier_desc *md = lookup_modifier(fb->modifier);
+
+	return ccs_aux_plane_mask(md, fb->format) & BIT(color_plane);
+}
+
+/**
+ * intel_fb_is_gen12_ccs_aux_plane: Check if a framebuffer color plane is a GEN12 CCS AUX plane
+ * @fb: Framebuffer
+ * @color_plane: color plane index to check
+ *
+ * Returns:
+ * Returns %true if @fb's color plane at index @color_plane is a GEN12 CCS AUX plane.
+ */
+static bool intel_fb_is_gen12_ccs_aux_plane(const struct drm_framebuffer *fb, int color_plane)
+{
+	const struct intel_modifier_desc *md = lookup_modifier(fb->modifier);
+
+	return check_modifier_display_ver_range(md, 12, 13) &&
+	       ccs_aux_plane_mask(md, fb->format) & BIT(color_plane);
+}
+
+/**
+ * intel_fb_rc_ccs_cc_plane: Get the CCS CC color plane index for a framebuffer
+ * @fb: Framebuffer
+ *
+ * Returns:
+ * Returns the index of the color clear plane for @fb, or -1 if @fb is not a
+ * framebuffer using a render compression/color clear modifier.
+ */
+int intel_fb_rc_ccs_cc_plane(const struct drm_framebuffer *fb)
+{
+	const struct intel_modifier_desc *md = lookup_modifier(fb->modifier);
+
+	if (!md->ccs.cc_planes)
+		return -1;
+
+	drm_WARN_ON_ONCE(fb->dev, hweight8(md->ccs.cc_planes) > 1);
+
+	return ilog2((int)md->ccs.cc_planes);
+}
+
+static bool is_gen12_ccs_cc_plane(const struct drm_framebuffer *fb, int color_plane)
+{
+	return intel_fb_rc_ccs_cc_plane(fb) == color_plane;
+}
+
+static bool is_semiplanar_uv_plane(const struct drm_framebuffer *fb, int color_plane)
+{
+	return intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier) &&
+		color_plane == 1;
+}
+
+bool is_surface_linear(const struct drm_framebuffer *fb, int color_plane)
+{
+	return fb->modifier == DRM_FORMAT_MOD_LINEAR ||
+	       intel_fb_is_gen12_ccs_aux_plane(fb, color_plane) ||
+	       is_gen12_ccs_cc_plane(fb, color_plane);
+}
+
+int main_to_ccs_plane(const struct drm_framebuffer *fb, int main_plane)
+{
+	drm_WARN_ON(fb->dev, !intel_fb_is_ccs_modifier(fb->modifier) ||
+		    (main_plane && main_plane >= fb->format->num_planes / 2));
+
+	return fb->format->num_planes / 2 + main_plane;
+}
+
+int skl_ccs_to_main_plane(const struct drm_framebuffer *fb, int ccs_plane)
+{
+	drm_WARN_ON(fb->dev, !intel_fb_is_ccs_modifier(fb->modifier) ||
+		    ccs_plane < fb->format->num_planes / 2);
+
+	if (is_gen12_ccs_cc_plane(fb, ccs_plane))
+		return 0;
+
+	return ccs_plane - fb->format->num_planes / 2;
+}
+
+static unsigned int gen12_ccs_aux_stride(struct intel_framebuffer *fb, int ccs_plane)
+{
+	int main_plane = skl_ccs_to_main_plane(&fb->base, ccs_plane);
+	unsigned int main_stride = fb->base.pitches[main_plane];
+	unsigned int main_tile_width = intel_tile_width_bytes(&fb->base, main_plane);
+
+	return DIV_ROUND_UP(main_stride, 4 * main_tile_width) * 64;
+}
+
+int skl_main_to_aux_plane(const struct drm_framebuffer *fb, int main_plane)
+{
+	const struct intel_modifier_desc *md = lookup_modifier(fb->modifier);
+	struct drm_i915_private *i915 = to_i915(fb->dev);
+
+	if (md->ccs.packed_aux_planes | md->ccs.planar_aux_planes)
+		return main_to_ccs_plane(fb, main_plane);
+	else if (DISPLAY_VER(i915) < 11 &&
+		 format_is_yuv_semiplanar(md, fb->format))
+		return 1;
+	else
+		return 0;
+}
+
+unsigned int intel_tile_size(const struct drm_i915_private *i915)
+{
+	return DISPLAY_VER(i915) == 2 ? 2048 : 4096;
+}
+
+unsigned int
+intel_tile_width_bytes(const struct drm_framebuffer *fb, int color_plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(fb->dev);
+	unsigned int cpp = fb->format->cpp[color_plane];
+
+	switch (fb->modifier) {
+	case DRM_FORMAT_MOD_LINEAR:
+		return intel_tile_size(dev_priv);
+	case I915_FORMAT_MOD_X_TILED:
+		if (DISPLAY_VER(dev_priv) == 2)
+			return 128;
+		else
+			return 512;
+	case I915_FORMAT_MOD_4_TILED_DG2_RC_CCS:
+	case I915_FORMAT_MOD_4_TILED_DG2_RC_CCS_CC:
+	case I915_FORMAT_MOD_4_TILED_DG2_MC_CCS:
+	case I915_FORMAT_MOD_4_TILED:
+		/*
+		 * Each 4K tile consists of 64B(8*8) subtiles, with
+		 * same shape as Y Tile(i.e 4*16B OWords)
+		 */
+		return 128;
+	case I915_FORMAT_MOD_Y_TILED_CCS:
+		if (intel_fb_is_ccs_aux_plane(fb, color_plane))
+			return 128;
+		fallthrough;
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS:
+		if (intel_fb_is_ccs_aux_plane(fb, color_plane) ||
+		    is_gen12_ccs_cc_plane(fb, color_plane))
+			return 64;
+		fallthrough;
+	case I915_FORMAT_MOD_Y_TILED:
+		if (DISPLAY_VER(dev_priv) == 2 || HAS_128_BYTE_Y_TILING(dev_priv))
+			return 128;
+		else
+			return 512;
+	case I915_FORMAT_MOD_Yf_TILED_CCS:
+		if (intel_fb_is_ccs_aux_plane(fb, color_plane))
+			return 128;
+		fallthrough;
+	case I915_FORMAT_MOD_Yf_TILED:
+		switch (cpp) {
+		case 1:
+			return 64;
+		case 2:
+		case 4:
+			return 128;
+		case 8:
+		case 16:
+			return 256;
+		default:
+			MISSING_CASE(cpp);
+			return cpp;
+		}
+		break;
+	default:
+		MISSING_CASE(fb->modifier);
+		return cpp;
+	}
+}
+
+unsigned int intel_tile_height(const struct drm_framebuffer *fb, int color_plane)
+{
+	return intel_tile_size(to_i915(fb->dev)) /
+		intel_tile_width_bytes(fb, color_plane);
+}
+
+/*
+ * Return the tile dimensions in pixel units, based on the (2 or 4 kbyte) GTT
+ * page tile size.
+ */
+static void intel_tile_dims(const struct drm_framebuffer *fb, int color_plane,
+			    unsigned int *tile_width,
+			    unsigned int *tile_height)
+{
+	unsigned int tile_width_bytes = intel_tile_width_bytes(fb, color_plane);
+	unsigned int cpp = fb->format->cpp[color_plane];
+
+	*tile_width = tile_width_bytes / cpp;
+	*tile_height = intel_tile_height(fb, color_plane);
+}
+
+/*
+ * Return the tile dimensions in pixel units, based on the tile block size.
+ * The block covers the full GTT page sized tile on all tiled surfaces and
+ * it's a 64 byte portion of the tile on TGL+ CCS surfaces.
+ */
+static void intel_tile_block_dims(const struct drm_framebuffer *fb, int color_plane,
+				  unsigned int *tile_width,
+				  unsigned int *tile_height)
+{
+	intel_tile_dims(fb, color_plane, tile_width, tile_height);
+
+	if (intel_fb_is_gen12_ccs_aux_plane(fb, color_plane))
+		*tile_height = 1;
+}
+
+unsigned int intel_tile_row_size(const struct drm_framebuffer *fb, int color_plane)
+{
+	unsigned int tile_width, tile_height;
+
+	intel_tile_dims(fb, color_plane, &tile_width, &tile_height);
+
+	return fb->pitches[color_plane] * tile_height;
+}
+
+unsigned int
+intel_fb_align_height(const struct drm_framebuffer *fb,
+		      int color_plane, unsigned int height)
+{
+	unsigned int tile_height = intel_tile_height(fb, color_plane);
+
+	return ALIGN(height, tile_height);
+}
+
+static unsigned int intel_fb_modifier_to_tiling(u64 fb_modifier)
+{
+	u8 tiling_caps = lookup_modifier(fb_modifier)->plane_caps &
+			 INTEL_PLANE_CAP_TILING_MASK;
+
+	switch (tiling_caps) {
+	case INTEL_PLANE_CAP_TILING_Y:
+		return I915_TILING_Y;
+	case INTEL_PLANE_CAP_TILING_X:
+		return I915_TILING_X;
+	case INTEL_PLANE_CAP_TILING_4:
+	case INTEL_PLANE_CAP_TILING_Yf:
+	case INTEL_PLANE_CAP_TILING_NONE:
+		return I915_TILING_NONE;
+	default:
+		MISSING_CASE(tiling_caps);
+		return I915_TILING_NONE;
+	}
+}
+
+static bool intel_modifier_uses_dpt(struct drm_i915_private *i915, u64 modifier)
+{
+	return DISPLAY_VER(i915) >= 13 && modifier != DRM_FORMAT_MOD_LINEAR;
+}
+
+bool intel_fb_uses_dpt(const struct drm_framebuffer *fb)
+{
+	return fb && intel_modifier_uses_dpt(to_i915(fb->dev), fb->modifier);
+}
+
+unsigned int intel_cursor_alignment(const struct drm_i915_private *i915)
+{
+	if (IS_I830(i915))
+		return 16 * 1024;
+	else if (IS_I85X(i915))
+		return 256;
+	else if (IS_I845G(i915) || IS_I865G(i915))
+		return 32;
+	else
+		return 4 * 1024;
+}
+
+static unsigned int intel_linear_alignment(const struct drm_i915_private *dev_priv)
+{
+	if (DISPLAY_VER(dev_priv) >= 9)
+		return 256 * 1024;
+	else if (IS_I965G(dev_priv) || IS_I965GM(dev_priv) ||
+		 IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		return 128 * 1024;
+	else if (DISPLAY_VER(dev_priv) >= 4)
+		return 4 * 1024;
+	else
+		return 0;
+}
+
+unsigned int intel_surf_alignment(const struct drm_framebuffer *fb,
+				  int color_plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(fb->dev);
+
+	if (intel_fb_uses_dpt(fb))
+		return 512 * 4096;
+
+	/* AUX_DIST needs only 4K alignment */
+	if (intel_fb_is_ccs_aux_plane(fb, color_plane))
+		return 4096;
+
+	if (is_semiplanar_uv_plane(fb, color_plane)) {
+		/*
+		 * TODO: cross-check wrt. the bspec stride in bytes * 64 bytes
+		 * alignment for linear UV planes on all platforms.
+		 */
+		if (DISPLAY_VER(dev_priv) >= 12) {
+			if (fb->modifier == DRM_FORMAT_MOD_LINEAR)
+				return intel_linear_alignment(dev_priv);
+
+			return intel_tile_row_size(fb, color_plane);
+		}
+
+		return 4096;
+	}
+
+	drm_WARN_ON(&dev_priv->drm, color_plane != 0);
+
+	switch (fb->modifier) {
+	case DRM_FORMAT_MOD_LINEAR:
+		return intel_linear_alignment(dev_priv);
+	case I915_FORMAT_MOD_X_TILED:
+		if (HAS_ASYNC_FLIPS(dev_priv))
+			return 256 * 1024;
+		return 0;
+	case I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC:
+		return 16 * 1024;
+	case I915_FORMAT_MOD_Y_TILED_CCS:
+	case I915_FORMAT_MOD_Yf_TILED_CCS:
+	case I915_FORMAT_MOD_Y_TILED:
+	case I915_FORMAT_MOD_4_TILED:
+	case I915_FORMAT_MOD_Yf_TILED:
+		return 1 * 1024 * 1024;
+	case I915_FORMAT_MOD_4_TILED_DG2_RC_CCS:
+	case I915_FORMAT_MOD_4_TILED_DG2_RC_CCS_CC:
+	case I915_FORMAT_MOD_4_TILED_DG2_MC_CCS:
+		return 16 * 1024;
+	default:
+		MISSING_CASE(fb->modifier);
+		return 0;
+	}
+}
+
+void intel_fb_plane_get_subsampling(int *hsub, int *vsub,
+				    const struct drm_framebuffer *fb,
+				    int color_plane)
+{
+	int main_plane;
+
+	if (color_plane == 0) {
+		*hsub = 1;
+		*vsub = 1;
+
+		return;
+	}
+
+	/*
+	 * TODO: Deduct the subsampling from the char block for all CCS
+	 * formats and planes.
+	 */
+	if (!intel_fb_is_gen12_ccs_aux_plane(fb, color_plane)) {
+		*hsub = fb->format->hsub;
+		*vsub = fb->format->vsub;
+
+		return;
+	}
+
+	main_plane = skl_ccs_to_main_plane(fb, color_plane);
+	*hsub = drm_format_info_block_width(fb->format, color_plane) /
+		drm_format_info_block_width(fb->format, main_plane);
+
+	/*
+	 * The min stride check in the core framebuffer_check() function
+	 * assumes that format->hsub applies to every plane except for the
+	 * first plane. That's incorrect for the CCS AUX plane of the first
+	 * plane, but for the above check to pass we must define the block
+	 * width with that subsampling applied to it. Adjust the width here
+	 * accordingly, so we can calculate the actual subsampling factor.
+	 */
+	if (main_plane == 0)
+		*hsub *= fb->format->hsub;
+
+	*vsub = 32;
+}
+
+static void intel_fb_plane_dims(const struct intel_framebuffer *fb, int color_plane, int *w, int *h)
+{
+	int main_plane = intel_fb_is_ccs_aux_plane(&fb->base, color_plane) ?
+			 skl_ccs_to_main_plane(&fb->base, color_plane) : 0;
+	unsigned int main_width = fb->base.width;
+	unsigned int main_height = fb->base.height;
+	int main_hsub, main_vsub;
+	int hsub, vsub;
+
+	intel_fb_plane_get_subsampling(&main_hsub, &main_vsub, &fb->base, main_plane);
+	intel_fb_plane_get_subsampling(&hsub, &vsub, &fb->base, color_plane);
+
+	*w = DIV_ROUND_UP(main_width, main_hsub * hsub);
+	*h = DIV_ROUND_UP(main_height, main_vsub * vsub);
+}
+
+static u32 intel_adjust_tile_offset(int *x, int *y,
+				    unsigned int tile_width,
+				    unsigned int tile_height,
+				    unsigned int tile_size,
+				    unsigned int pitch_tiles,
+				    u32 old_offset,
+				    u32 new_offset)
+{
+	unsigned int pitch_pixels = pitch_tiles * tile_width;
+	unsigned int tiles;
+
+	WARN_ON(old_offset & (tile_size - 1));
+	WARN_ON(new_offset & (tile_size - 1));
+	WARN_ON(new_offset > old_offset);
+
+	tiles = (old_offset - new_offset) / tile_size;
+
+	*y += tiles / pitch_tiles * tile_height;
+	*x += tiles % pitch_tiles * tile_width;
+
+	/* minimize x in case it got needlessly big */
+	*y += *x / pitch_pixels * tile_height;
+	*x %= pitch_pixels;
+
+	return new_offset;
+}
+
+static u32 intel_adjust_linear_offset(int *x, int *y,
+				      unsigned int cpp,
+				      unsigned int pitch,
+				      u32 old_offset,
+				      u32 new_offset)
+{
+	old_offset += *y * pitch + *x * cpp;
+
+	*y = (old_offset - new_offset) / pitch;
+	*x = ((old_offset - new_offset) - *y * pitch) / cpp;
+
+	return new_offset;
+}
+
+static u32 intel_adjust_aligned_offset(int *x, int *y,
+				       const struct drm_framebuffer *fb,
+				       int color_plane,
+				       unsigned int rotation,
+				       unsigned int pitch,
+				       u32 old_offset, u32 new_offset)
+{
+	struct drm_i915_private *i915 = to_i915(fb->dev);
+	unsigned int cpp = fb->format->cpp[color_plane];
+
+	drm_WARN_ON(&i915->drm, new_offset > old_offset);
+
+	if (!is_surface_linear(fb, color_plane)) {
+		unsigned int tile_size, tile_width, tile_height;
+		unsigned int pitch_tiles;
+
+		tile_size = intel_tile_size(i915);
+		intel_tile_dims(fb, color_plane, &tile_width, &tile_height);
+
+		if (drm_rotation_90_or_270(rotation)) {
+			pitch_tiles = pitch / tile_height;
+			swap(tile_width, tile_height);
+		} else {
+			pitch_tiles = pitch / (tile_width * cpp);
+		}
+
+		intel_adjust_tile_offset(x, y, tile_width, tile_height,
+					 tile_size, pitch_tiles,
+					 old_offset, new_offset);
+	} else {
+		intel_adjust_linear_offset(x, y, cpp, pitch,
+					   old_offset, new_offset);
+	}
+
+	return new_offset;
+}
+
+/*
+ * Adjust the tile offset by moving the difference into
+ * the x/y offsets.
+ */
+u32 intel_plane_adjust_aligned_offset(int *x, int *y,
+				      const struct intel_plane_state *state,
+				      int color_plane,
+				      u32 old_offset, u32 new_offset)
+{
+	return intel_adjust_aligned_offset(x, y, state->hw.fb, color_plane,
+					   state->hw.rotation,
+					   state->view.color_plane[color_plane].mapping_stride,
+					   old_offset, new_offset);
+}
+
+/*
+ * Computes the aligned offset to the base tile and adjusts
+ * x, y. bytes per pixel is assumed to be a power-of-two.
+ *
+ * In the 90/270 rotated case, x and y are assumed
+ * to be already rotated to match the rotated GTT view, and
+ * pitch is the tile_height aligned framebuffer height.
+ *
+ * This function is used when computing the derived information
+ * under intel_framebuffer, so using any of that information
+ * here is not allowed. Anything under drm_framebuffer can be
+ * used. This is why the user has to pass in the pitch since it
+ * is specified in the rotated orientation.
+ */
+static u32 intel_compute_aligned_offset(struct drm_i915_private *i915,
+					int *x, int *y,
+					const struct drm_framebuffer *fb,
+					int color_plane,
+					unsigned int pitch,
+					unsigned int rotation,
+					u32 alignment)
+{
+	unsigned int cpp = fb->format->cpp[color_plane];
+	u32 offset, offset_aligned;
+
+	if (!is_surface_linear(fb, color_plane)) {
+		unsigned int tile_size, tile_width, tile_height;
+		unsigned int tile_rows, tiles, pitch_tiles;
+
+		tile_size = intel_tile_size(i915);
+		intel_tile_dims(fb, color_plane, &tile_width, &tile_height);
+
+		if (drm_rotation_90_or_270(rotation)) {
+			pitch_tiles = pitch / tile_height;
+			swap(tile_width, tile_height);
+		} else {
+			pitch_tiles = pitch / (tile_width * cpp);
+		}
+
+		tile_rows = *y / tile_height;
+		*y %= tile_height;
+
+		tiles = *x / tile_width;
+		*x %= tile_width;
+
+		offset = (tile_rows * pitch_tiles + tiles) * tile_size;
+
+		offset_aligned = offset;
+		if (alignment)
+			offset_aligned = rounddown(offset_aligned, alignment);
+
+		intel_adjust_tile_offset(x, y, tile_width, tile_height,
+					 tile_size, pitch_tiles,
+					 offset, offset_aligned);
+	} else {
+		offset = *y * pitch + *x * cpp;
+		offset_aligned = offset;
+		if (alignment) {
+			offset_aligned = rounddown(offset_aligned, alignment);
+			*y = (offset % alignment) / pitch;
+			*x = ((offset % alignment) - *y * pitch) / cpp;
+		} else {
+			*y = *x = 0;
+		}
+	}
+
+	return offset_aligned;
+}
+
+u32 intel_plane_compute_aligned_offset(int *x, int *y,
+				       const struct intel_plane_state *state,
+				       int color_plane)
+{
+	struct intel_plane *intel_plane = to_intel_plane(state->uapi.plane);
+	struct drm_i915_private *i915 = to_i915(intel_plane->base.dev);
+	const struct drm_framebuffer *fb = state->hw.fb;
+	unsigned int rotation = state->hw.rotation;
+	int pitch = state->view.color_plane[color_plane].mapping_stride;
+	u32 alignment;
+
+	if (intel_plane->id == PLANE_CURSOR)
+		alignment = intel_cursor_alignment(i915);
+	else
+		alignment = intel_surf_alignment(fb, color_plane);
+
+	return intel_compute_aligned_offset(i915, x, y, fb, color_plane,
+					    pitch, rotation, alignment);
+}
+
+/* Convert the fb->offset[] into x/y offsets */
+static int intel_fb_offset_to_xy(int *x, int *y,
+				 const struct drm_framebuffer *fb,
+				 int color_plane)
+{
+	struct drm_i915_private *i915 = to_i915(fb->dev);
+	unsigned int height;
+	u32 alignment;
+
+	if (DISPLAY_VER(i915) >= 12 &&
+	    !intel_fb_needs_pot_stride_remap(to_intel_framebuffer(fb)) &&
+	    is_semiplanar_uv_plane(fb, color_plane))
+		alignment = intel_tile_row_size(fb, color_plane);
+	else if (fb->modifier != DRM_FORMAT_MOD_LINEAR)
+		alignment = intel_tile_size(i915);
+	else
+		alignment = 0;
+
+	if (alignment != 0 && fb->offsets[color_plane] % alignment) {
+		drm_dbg_kms(&i915->drm,
+			    "Misaligned offset 0x%08x for color plane %d\n",
+			    fb->offsets[color_plane], color_plane);
+		return -EINVAL;
+	}
+
+	height = drm_framebuffer_plane_height(fb->height, fb, color_plane);
+	height = ALIGN(height, intel_tile_height(fb, color_plane));
+
+	/* Catch potential overflows early */
+	if (add_overflows_t(u32, mul_u32_u32(height, fb->pitches[color_plane]),
+			    fb->offsets[color_plane])) {
+		drm_dbg_kms(&i915->drm,
+			    "Bad offset 0x%08x or pitch %d for color plane %d\n",
+			    fb->offsets[color_plane], fb->pitches[color_plane],
+			    color_plane);
+		return -ERANGE;
+	}
+
+	*x = 0;
+	*y = 0;
+
+	intel_adjust_aligned_offset(x, y,
+				    fb, color_plane, DRM_MODE_ROTATE_0,
+				    fb->pitches[color_plane],
+				    fb->offsets[color_plane], 0);
+
+	return 0;
+}
+
+static int intel_fb_check_ccs_xy(const struct drm_framebuffer *fb, int ccs_plane, int x, int y)
+{
+	struct drm_i915_private *i915 = to_i915(fb->dev);
+	const struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
+	int main_plane;
+	int hsub, vsub;
+	int tile_width, tile_height;
+	int ccs_x, ccs_y;
+	int main_x, main_y;
+
+	if (!intel_fb_is_ccs_aux_plane(fb, ccs_plane))
+		return 0;
+
+	/*
+	 * While all the tile dimensions are based on a 2k or 4k GTT page size
+	 * here the main and CCS coordinates must match only within a (64 byte
+	 * on TGL+) block inside the tile.
+	 */
+	intel_tile_block_dims(fb, ccs_plane, &tile_width, &tile_height);
+	intel_fb_plane_get_subsampling(&hsub, &vsub, fb, ccs_plane);
+
+	tile_width *= hsub;
+	tile_height *= vsub;
+
+	ccs_x = (x * hsub) % tile_width;
+	ccs_y = (y * vsub) % tile_height;
+
+	main_plane = skl_ccs_to_main_plane(fb, ccs_plane);
+	main_x = intel_fb->normal_view.color_plane[main_plane].x % tile_width;
+	main_y = intel_fb->normal_view.color_plane[main_plane].y % tile_height;
+
+	/*
+	 * CCS doesn't have its own x/y offset register, so the intra CCS tile
+	 * x/y offsets must match between CCS and the main surface.
+	 */
+	if (main_x != ccs_x || main_y != ccs_y) {
+		drm_dbg_kms(&i915->drm,
+			      "Bad CCS x/y (main %d,%d ccs %d,%d) full (main %d,%d ccs %d,%d)\n",
+			      main_x, main_y,
+			      ccs_x, ccs_y,
+			      intel_fb->normal_view.color_plane[main_plane].x,
+			      intel_fb->normal_view.color_plane[main_plane].y,
+			      x, y);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static bool intel_plane_can_remap(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int i;
+
+	/* We don't want to deal with remapping with cursors */
+	if (plane->id == PLANE_CURSOR)
+		return false;
+
+	/*
+	 * The display engine limits already match/exceed the
+	 * render engine limits, so not much point in remapping.
+	 * Would also need to deal with the fence POT alignment
+	 * and gen2 2KiB GTT tile size.
+	 */
+	if (DISPLAY_VER(i915) < 4)
+		return false;
+
+	/*
+	 * The new CCS hash mode isn't compatible with remapping as
+	 * the virtual address of the pages affects the compressed data.
+	 */
+	if (intel_fb_is_ccs_modifier(fb->modifier))
+		return false;
+
+	/* Linear needs a page aligned stride for remapping */
+	if (fb->modifier == DRM_FORMAT_MOD_LINEAR) {
+		unsigned int alignment = intel_tile_size(i915) - 1;
+
+		for (i = 0; i < fb->format->num_planes; i++) {
+			if (fb->pitches[i] & alignment)
+				return false;
+		}
+	}
+
+	return true;
+}
+
+bool intel_fb_needs_pot_stride_remap(const struct intel_framebuffer *fb)
+{
+	struct drm_i915_private *i915 = to_i915(fb->base.dev);
+
+	return (IS_ALDERLAKE_P(i915) || DISPLAY_VER(i915) >= 14) &&
+		intel_fb_uses_dpt(&fb->base);
+}
+
+static int intel_fb_pitch(const struct intel_framebuffer *fb, int color_plane, unsigned int rotation)
+{
+	if (drm_rotation_90_or_270(rotation))
+		return fb->rotated_view.color_plane[color_plane].mapping_stride;
+	else if (intel_fb_needs_pot_stride_remap(fb))
+		return fb->remapped_view.color_plane[color_plane].mapping_stride;
+	else
+		return fb->normal_view.color_plane[color_plane].mapping_stride;
+}
+
+static bool intel_plane_needs_remap(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	const struct intel_framebuffer *fb = to_intel_framebuffer(plane_state->hw.fb);
+	unsigned int rotation = plane_state->hw.rotation;
+	u32 stride, max_stride;
+
+	/*
+	 * No remapping for invisible planes since we don't have
+	 * an actual source viewport to remap.
+	 */
+	if (!plane_state->uapi.visible)
+		return false;
+
+	if (!intel_plane_can_remap(plane_state))
+		return false;
+
+	/*
+	 * FIXME: aux plane limits on gen9+ are
+	 * unclear in Bspec, for now no checking.
+	 */
+	stride = intel_fb_pitch(fb, 0, rotation);
+	max_stride = plane->max_stride(plane, fb->base.format->format,
+				       fb->base.modifier, rotation);
+
+	return stride > max_stride;
+}
+
+static int convert_plane_offset_to_xy(const struct intel_framebuffer *fb, int color_plane,
+				      int plane_width, int *x, int *y)
+{
+	struct drm_i915_gem_object *obj = intel_fb_obj(&fb->base);
+	int ret;
+
+	ret = intel_fb_offset_to_xy(x, y, &fb->base, color_plane);
+	if (ret) {
+		drm_dbg_kms(fb->base.dev,
+			    "bad fb plane %d offset: 0x%x\n",
+			    color_plane, fb->base.offsets[color_plane]);
+		return ret;
+	}
+
+	ret = intel_fb_check_ccs_xy(&fb->base, color_plane, *x, *y);
+	if (ret)
+		return ret;
+
+	/*
+	 * The fence (if used) is aligned to the start of the object
+	 * so having the framebuffer wrap around across the edge of the
+	 * fenced region doesn't really work. We have no API to configure
+	 * the fence start offset within the object (nor could we probably
+	 * on gen2/3). So it's just easier if we just require that the
+	 * fb layout agrees with the fence layout. We already check that the
+	 * fb stride matches the fence stride elsewhere.
+	 */
+	if (color_plane == 0 && i915_gem_object_is_tiled(obj) &&
+	    (*x + plane_width) * fb->base.format->cpp[color_plane] > fb->base.pitches[color_plane]) {
+		drm_dbg_kms(fb->base.dev,
+			    "bad fb plane %d offset: 0x%x\n",
+			    color_plane, fb->base.offsets[color_plane]);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static u32 calc_plane_aligned_offset(const struct intel_framebuffer *fb, int color_plane, int *x, int *y)
+{
+	struct drm_i915_private *i915 = to_i915(fb->base.dev);
+	unsigned int tile_size = intel_tile_size(i915);
+	u32 offset;
+
+	offset = intel_compute_aligned_offset(i915, x, y, &fb->base, color_plane,
+					      fb->base.pitches[color_plane],
+					      DRM_MODE_ROTATE_0,
+					      tile_size);
+
+	return offset / tile_size;
+}
+
+struct fb_plane_view_dims {
+	unsigned int width, height;
+	unsigned int tile_width, tile_height;
+};
+
+static void init_plane_view_dims(const struct intel_framebuffer *fb, int color_plane,
+				 unsigned int width, unsigned int height,
+				 struct fb_plane_view_dims *dims)
+{
+	dims->width = width;
+	dims->height = height;
+
+	intel_tile_dims(&fb->base, color_plane, &dims->tile_width, &dims->tile_height);
+}
+
+static unsigned int
+plane_view_src_stride_tiles(const struct intel_framebuffer *fb, int color_plane,
+			    const struct fb_plane_view_dims *dims)
+{
+	return DIV_ROUND_UP(fb->base.pitches[color_plane],
+			    dims->tile_width * fb->base.format->cpp[color_plane]);
+}
+
+static unsigned int
+plane_view_dst_stride_tiles(const struct intel_framebuffer *fb, int color_plane,
+			    unsigned int pitch_tiles)
+{
+	if (intel_fb_needs_pot_stride_remap(fb)) {
+		/*
+		 * ADL_P, the only platform needing a POT stride has a minimum
+		 * of 8 main surface tiles.
+		 */
+		return roundup_pow_of_two(max(pitch_tiles, 8u));
+	} else {
+		return pitch_tiles;
+	}
+}
+
+static unsigned int
+plane_view_scanout_stride(const struct intel_framebuffer *fb, int color_plane,
+			  unsigned int tile_width,
+			  unsigned int src_stride_tiles, unsigned int dst_stride_tiles)
+{
+	struct drm_i915_private *i915 = to_i915(fb->base.dev);
+	unsigned int stride_tiles;
+
+	if ((IS_ALDERLAKE_P(i915) || DISPLAY_VER(i915) >= 14) &&
+	    src_stride_tiles < dst_stride_tiles)
+		stride_tiles = src_stride_tiles;
+	else
+		stride_tiles = dst_stride_tiles;
+
+	return stride_tiles * tile_width * fb->base.format->cpp[color_plane];
+}
+
+static unsigned int
+plane_view_width_tiles(const struct intel_framebuffer *fb, int color_plane,
+		       const struct fb_plane_view_dims *dims,
+		       int x)
+{
+	return DIV_ROUND_UP(x + dims->width, dims->tile_width);
+}
+
+static unsigned int
+plane_view_height_tiles(const struct intel_framebuffer *fb, int color_plane,
+			const struct fb_plane_view_dims *dims,
+			int y)
+{
+	return DIV_ROUND_UP(y + dims->height, dims->tile_height);
+}
+
+static unsigned int
+plane_view_linear_tiles(const struct intel_framebuffer *fb, int color_plane,
+			const struct fb_plane_view_dims *dims,
+			int x, int y)
+{
+	struct drm_i915_private *i915 = to_i915(fb->base.dev);
+	unsigned int size;
+
+	size = (y + dims->height) * fb->base.pitches[color_plane] +
+		x * fb->base.format->cpp[color_plane];
+
+	return DIV_ROUND_UP(size, intel_tile_size(i915));
+}
+
+#define assign_chk_ovf(i915, var, val) ({ \
+	drm_WARN_ON(&(i915)->drm, overflows_type(val, var)); \
+	(var) = (val); \
+})
+
+#define assign_bfld_chk_ovf(i915, var, val) ({ \
+	(var) = (val); \
+	drm_WARN_ON(&(i915)->drm, (var) != (val)); \
+	(var); \
+})
+
+static u32 calc_plane_remap_info(const struct intel_framebuffer *fb, int color_plane,
+				 const struct fb_plane_view_dims *dims,
+				 u32 obj_offset, u32 gtt_offset, int x, int y,
+				 struct intel_fb_view *view)
+{
+	struct drm_i915_private *i915 = to_i915(fb->base.dev);
+	struct intel_remapped_plane_info *remap_info = &view->gtt.remapped.plane[color_plane];
+	struct i915_color_plane_view *color_plane_info = &view->color_plane[color_plane];
+	unsigned int tile_width = dims->tile_width;
+	unsigned int tile_height = dims->tile_height;
+	unsigned int tile_size = intel_tile_size(i915);
+	struct drm_rect r;
+	u32 size = 0;
+
+	assign_bfld_chk_ovf(i915, remap_info->offset, obj_offset);
+
+	if (intel_fb_is_gen12_ccs_aux_plane(&fb->base, color_plane)) {
+		remap_info->linear = 1;
+
+		assign_chk_ovf(i915, remap_info->size,
+			       plane_view_linear_tiles(fb, color_plane, dims, x, y));
+	} else {
+		remap_info->linear = 0;
+
+		assign_chk_ovf(i915, remap_info->src_stride,
+			       plane_view_src_stride_tiles(fb, color_plane, dims));
+		assign_chk_ovf(i915, remap_info->width,
+			       plane_view_width_tiles(fb, color_plane, dims, x));
+		assign_chk_ovf(i915, remap_info->height,
+			       plane_view_height_tiles(fb, color_plane, dims, y));
+	}
+
+	if (view->gtt.type == I915_GTT_VIEW_ROTATED) {
+		drm_WARN_ON(&i915->drm, remap_info->linear);
+		check_array_bounds(i915, view->gtt.rotated.plane, color_plane);
+
+		assign_chk_ovf(i915, remap_info->dst_stride,
+			       plane_view_dst_stride_tiles(fb, color_plane, remap_info->height));
+
+		/* rotate the x/y offsets to match the GTT view */
+		drm_rect_init(&r, x, y, dims->width, dims->height);
+		drm_rect_rotate(&r,
+				remap_info->width * tile_width,
+				remap_info->height * tile_height,
+				DRM_MODE_ROTATE_270);
+
+		color_plane_info->x = r.x1;
+		color_plane_info->y = r.y1;
+
+		color_plane_info->mapping_stride = remap_info->dst_stride * tile_height;
+		color_plane_info->scanout_stride = color_plane_info->mapping_stride;
+
+		size += remap_info->dst_stride * remap_info->width;
+
+		/* rotate the tile dimensions to match the GTT view */
+		swap(tile_width, tile_height);
+	} else {
+		drm_WARN_ON(&i915->drm, view->gtt.type != I915_GTT_VIEW_REMAPPED);
+
+		check_array_bounds(i915, view->gtt.remapped.plane, color_plane);
+
+		if (view->gtt.remapped.plane_alignment) {
+			unsigned int aligned_offset = ALIGN(gtt_offset,
+							    view->gtt.remapped.plane_alignment);
+
+			size += aligned_offset - gtt_offset;
+			gtt_offset = aligned_offset;
+		}
+
+		color_plane_info->x = x;
+		color_plane_info->y = y;
+
+		if (remap_info->linear) {
+			color_plane_info->mapping_stride = fb->base.pitches[color_plane];
+			color_plane_info->scanout_stride = color_plane_info->mapping_stride;
+
+			size += remap_info->size;
+		} else {
+			unsigned int dst_stride;
+
+			/*
+			 * The hardware automagically calculates the CCS AUX surface
+			 * stride from the main surface stride so can't really remap a
+			 * smaller subset (unless we'd remap in whole AUX page units).
+			 */
+			if (intel_fb_needs_pot_stride_remap(fb) &&
+			    intel_fb_is_ccs_modifier(fb->base.modifier))
+				dst_stride = remap_info->src_stride;
+			else
+				dst_stride = remap_info->width;
+
+			dst_stride = plane_view_dst_stride_tiles(fb, color_plane, dst_stride);
+
+			assign_chk_ovf(i915, remap_info->dst_stride, dst_stride);
+			color_plane_info->mapping_stride = dst_stride *
+							   tile_width *
+							   fb->base.format->cpp[color_plane];
+			color_plane_info->scanout_stride =
+				plane_view_scanout_stride(fb, color_plane, tile_width,
+							  remap_info->src_stride,
+							  dst_stride);
+
+			size += dst_stride * remap_info->height;
+		}
+	}
+
+	/*
+	 * We only keep the x/y offsets, so push all of the gtt offset into
+	 * the x/y offsets.  x,y will hold the first pixel of the framebuffer
+	 * plane from the start of the remapped/rotated gtt mapping.
+	 */
+	if (remap_info->linear)
+		intel_adjust_linear_offset(&color_plane_info->x, &color_plane_info->y,
+					   fb->base.format->cpp[color_plane],
+					   color_plane_info->mapping_stride,
+					   gtt_offset * tile_size, 0);
+	else
+		intel_adjust_tile_offset(&color_plane_info->x, &color_plane_info->y,
+					 tile_width, tile_height,
+					 tile_size, remap_info->dst_stride,
+					 gtt_offset * tile_size, 0);
+
+	return size;
+}
+
+#undef assign_chk_ovf
+
+/* Return number of tiles @color_plane needs. */
+static unsigned int
+calc_plane_normal_size(const struct intel_framebuffer *fb, int color_plane,
+		       const struct fb_plane_view_dims *dims,
+		       int x, int y)
+{
+	unsigned int tiles;
+
+	if (is_surface_linear(&fb->base, color_plane)) {
+		tiles = plane_view_linear_tiles(fb, color_plane, dims, x, y);
+	} else {
+		tiles = plane_view_src_stride_tiles(fb, color_plane, dims) *
+			plane_view_height_tiles(fb, color_plane, dims, y);
+		/*
+		 * If the plane isn't horizontally tile aligned,
+		 * we need one more tile.
+		 */
+		if (x != 0)
+			tiles++;
+	}
+
+	return tiles;
+}
+
+static void intel_fb_view_init(struct drm_i915_private *i915, struct intel_fb_view *view,
+			       enum i915_gtt_view_type view_type)
+{
+	memset(view, 0, sizeof(*view));
+	view->gtt.type = view_type;
+
+	if (view_type == I915_GTT_VIEW_REMAPPED &&
+	    (IS_ALDERLAKE_P(i915) || DISPLAY_VER(i915) >= 14))
+		view->gtt.remapped.plane_alignment = SZ_2M / PAGE_SIZE;
+}
+
+bool intel_fb_supports_90_270_rotation(const struct intel_framebuffer *fb)
+{
+	if (DISPLAY_VER(to_i915(fb->base.dev)) >= 13)
+		return false;
+
+	return fb->base.modifier == I915_FORMAT_MOD_Y_TILED ||
+	       fb->base.modifier == I915_FORMAT_MOD_Yf_TILED;
+}
+
+int intel_fill_fb_info(struct drm_i915_private *i915, struct intel_framebuffer *fb)
+{
+	struct drm_i915_gem_object *obj = intel_fb_obj(&fb->base);
+	u32 gtt_offset_rotated = 0;
+	u32 gtt_offset_remapped = 0;
+	unsigned int max_size = 0;
+	int i, num_planes = fb->base.format->num_planes;
+	unsigned int tile_size = intel_tile_size(i915);
+
+	intel_fb_view_init(i915, &fb->normal_view, I915_GTT_VIEW_NORMAL);
+
+	drm_WARN_ON(&i915->drm,
+		    intel_fb_supports_90_270_rotation(fb) &&
+		    intel_fb_needs_pot_stride_remap(fb));
+
+	if (intel_fb_supports_90_270_rotation(fb))
+		intel_fb_view_init(i915, &fb->rotated_view, I915_GTT_VIEW_ROTATED);
+	if (intel_fb_needs_pot_stride_remap(fb))
+		intel_fb_view_init(i915, &fb->remapped_view, I915_GTT_VIEW_REMAPPED);
+
+	for (i = 0; i < num_planes; i++) {
+		struct fb_plane_view_dims view_dims;
+		unsigned int width, height;
+		unsigned int cpp, size;
+		u32 offset;
+		int x, y;
+		int ret;
+
+		/*
+		 * Plane 2 of Render Compression with Clear Color fb modifier
+		 * is consumed by the driver and not passed to DE. Skip the
+		 * arithmetic related to alignment and offset calculation.
+		 */
+		if (is_gen12_ccs_cc_plane(&fb->base, i)) {
+			if (IS_ALIGNED(fb->base.offsets[i], PAGE_SIZE))
+				continue;
+			else
+				return -EINVAL;
+		}
+
+		cpp = fb->base.format->cpp[i];
+		intel_fb_plane_dims(fb, i, &width, &height);
+
+		ret = convert_plane_offset_to_xy(fb, i, width, &x, &y);
+		if (ret)
+			return ret;
+
+		init_plane_view_dims(fb, i, width, height, &view_dims);
+
+		/*
+		 * First pixel of the framebuffer from
+		 * the start of the normal gtt mapping.
+		 */
+		fb->normal_view.color_plane[i].x = x;
+		fb->normal_view.color_plane[i].y = y;
+		fb->normal_view.color_plane[i].mapping_stride = fb->base.pitches[i];
+		fb->normal_view.color_plane[i].scanout_stride =
+			fb->normal_view.color_plane[i].mapping_stride;
+
+		offset = calc_plane_aligned_offset(fb, i, &x, &y);
+
+		if (intel_fb_supports_90_270_rotation(fb))
+			gtt_offset_rotated += calc_plane_remap_info(fb, i, &view_dims,
+								    offset, gtt_offset_rotated, x, y,
+								    &fb->rotated_view);
+
+		if (intel_fb_needs_pot_stride_remap(fb))
+			gtt_offset_remapped += calc_plane_remap_info(fb, i, &view_dims,
+								     offset, gtt_offset_remapped, x, y,
+								     &fb->remapped_view);
+
+		size = calc_plane_normal_size(fb, i, &view_dims, x, y);
+		/* how many tiles in total needed in the bo */
+		max_size = max(max_size, offset + size);
+	}
+
+	if (mul_u32_u32(max_size, tile_size) > obj->base.size) {
+		drm_dbg_kms(&i915->drm,
+			    "fb too big for bo (need %llu bytes, have %zu bytes)\n",
+			    mul_u32_u32(max_size, tile_size), obj->base.size);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void intel_plane_remap_gtt(struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 =
+		to_i915(plane_state->uapi.plane->dev);
+	struct drm_framebuffer *fb = plane_state->hw.fb;
+	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
+	unsigned int rotation = plane_state->hw.rotation;
+	int i, num_planes = fb->format->num_planes;
+	unsigned int src_x, src_y;
+	unsigned int src_w, src_h;
+	u32 gtt_offset = 0;
+
+	intel_fb_view_init(i915, &plane_state->view,
+			   drm_rotation_90_or_270(rotation) ? I915_GTT_VIEW_ROTATED :
+							      I915_GTT_VIEW_REMAPPED);
+
+	src_x = plane_state->uapi.src.x1 >> 16;
+	src_y = plane_state->uapi.src.y1 >> 16;
+	src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+	src_h = drm_rect_height(&plane_state->uapi.src) >> 16;
+
+	drm_WARN_ON(&i915->drm, intel_fb_is_ccs_modifier(fb->modifier));
+
+	/* Make src coordinates relative to the viewport */
+	drm_rect_translate(&plane_state->uapi.src,
+			   -(src_x << 16), -(src_y << 16));
+
+	/* Rotate src coordinates to match rotated GTT view */
+	if (drm_rotation_90_or_270(rotation))
+		drm_rect_rotate(&plane_state->uapi.src,
+				src_w << 16, src_h << 16,
+				DRM_MODE_ROTATE_270);
+
+	for (i = 0; i < num_planes; i++) {
+		unsigned int hsub = i ? fb->format->hsub : 1;
+		unsigned int vsub = i ? fb->format->vsub : 1;
+		struct fb_plane_view_dims view_dims;
+		unsigned int width, height;
+		unsigned int x, y;
+		u32 offset;
+
+		x = src_x / hsub;
+		y = src_y / vsub;
+		width = src_w / hsub;
+		height = src_h / vsub;
+
+		init_plane_view_dims(intel_fb, i, width, height, &view_dims);
+
+		/*
+		 * First pixel of the src viewport from the
+		 * start of the normal gtt mapping.
+		 */
+		x += intel_fb->normal_view.color_plane[i].x;
+		y += intel_fb->normal_view.color_plane[i].y;
+
+		offset = calc_plane_aligned_offset(intel_fb, i, &x, &y);
+
+		gtt_offset += calc_plane_remap_info(intel_fb, i, &view_dims,
+						    offset, gtt_offset, x, y,
+						    &plane_state->view);
+	}
+}
+
+void intel_fb_fill_view(const struct intel_framebuffer *fb, unsigned int rotation,
+			struct intel_fb_view *view)
+{
+	if (drm_rotation_90_or_270(rotation))
+		*view = fb->rotated_view;
+	else if (intel_fb_needs_pot_stride_remap(fb))
+		*view = fb->remapped_view;
+	else
+		*view = fb->normal_view;
+}
+
+static
+u32 intel_fb_max_stride(struct drm_i915_private *dev_priv,
+			u32 pixel_format, u64 modifier)
+{
+	/*
+	 * Arbitrary limit for gen4+ chosen to match the
+	 * render engine max stride.
+	 *
+	 * The new CCS hash mode makes remapping impossible
+	 */
+	if (DISPLAY_VER(dev_priv) < 4 || intel_fb_is_ccs_modifier(modifier) ||
+	    intel_modifier_uses_dpt(dev_priv, modifier))
+		return intel_plane_fb_max_stride(dev_priv, pixel_format, modifier);
+	else if (DISPLAY_VER(dev_priv) >= 7)
+		return 256 * 1024;
+	else
+		return 128 * 1024;
+}
+
+static u32
+intel_fb_stride_alignment(const struct drm_framebuffer *fb, int color_plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(fb->dev);
+	u32 tile_width;
+
+	if (is_surface_linear(fb, color_plane)) {
+		u32 max_stride = intel_plane_fb_max_stride(dev_priv,
+							   fb->format->format,
+							   fb->modifier);
+
+		/*
+		 * To make remapping with linear generally feasible
+		 * we need the stride to be page aligned.
+		 */
+		if (fb->pitches[color_plane] > max_stride &&
+		    !intel_fb_is_ccs_modifier(fb->modifier))
+			return intel_tile_size(dev_priv);
+		else
+			return 64;
+	}
+
+	tile_width = intel_tile_width_bytes(fb, color_plane);
+	if (intel_fb_is_ccs_modifier(fb->modifier)) {
+		/*
+		 * On TGL the surface stride must be 4 tile aligned, mapped by
+		 * one 64 byte cacheline on the CCS AUX surface.
+		 */
+		if (DISPLAY_VER(dev_priv) >= 12)
+			tile_width *= 4;
+		/*
+		 * Display WA #0531: skl,bxt,kbl,glk
+		 *
+		 * Render decompression and plane width > 3840
+		 * combined with horizontal panning requires the
+		 * plane stride to be a multiple of 4. We'll just
+		 * require the entire fb to accommodate that to avoid
+		 * potential runtime errors at plane configuration time.
+		 */
+		else if ((DISPLAY_VER(dev_priv) == 9 || IS_GEMINILAKE(dev_priv)) &&
+			 color_plane == 0 && fb->width > 3840)
+			tile_width *= 4;
+	}
+	return tile_width;
+}
+
+static int intel_plane_check_stride(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+	u32 stride, max_stride;
+
+	/*
+	 * We ignore stride for all invisible planes that
+	 * can be remapped. Otherwise we could end up
+	 * with a false positive when the remapping didn't
+	 * kick in due the plane being invisible.
+	 */
+	if (intel_plane_can_remap(plane_state) &&
+	    !plane_state->uapi.visible)
+		return 0;
+
+	/* FIXME other color planes? */
+	stride = plane_state->view.color_plane[0].mapping_stride;
+	max_stride = plane->max_stride(plane, fb->format->format,
+				       fb->modifier, rotation);
+
+	if (stride > max_stride) {
+		DRM_DEBUG_KMS("[FB:%d] stride (%d) exceeds [PLANE:%d:%s] max stride (%d)\n",
+			      fb->base.id, stride,
+			      plane->base.base.id, plane->base.name, max_stride);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int intel_plane_compute_gtt(struct intel_plane_state *plane_state)
+{
+	const struct intel_framebuffer *fb =
+		to_intel_framebuffer(plane_state->hw.fb);
+	unsigned int rotation = plane_state->hw.rotation;
+
+	if (!fb)
+		return 0;
+
+	if (intel_plane_needs_remap(plane_state)) {
+		intel_plane_remap_gtt(plane_state);
+
+		/*
+		 * Sometimes even remapping can't overcome
+		 * the stride limitations :( Can happen with
+		 * big plane sizes and suitably misaligned
+		 * offsets.
+		 */
+		return intel_plane_check_stride(plane_state);
+	}
+
+	intel_fb_fill_view(fb, rotation, &plane_state->view);
+
+	/* Rotate src coordinates to match rotated GTT view */
+	if (drm_rotation_90_or_270(rotation))
+		drm_rect_rotate(&plane_state->uapi.src,
+				fb->base.width << 16, fb->base.height << 16,
+				DRM_MODE_ROTATE_270);
+
+	return intel_plane_check_stride(plane_state);
+}
+
+static void intel_user_framebuffer_destroy(struct drm_framebuffer *fb)
+{
+	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
+
+	drm_framebuffer_cleanup(fb);
+
+	if (intel_fb_uses_dpt(fb))
+		intel_dpt_destroy(intel_fb->dpt_vm);
+
+	intel_frontbuffer_put(intel_fb->frontbuffer);
+
+	kfree(intel_fb);
+}
+
+static int intel_user_framebuffer_create_handle(struct drm_framebuffer *fb,
+						struct drm_file *file,
+						unsigned int *handle)
+{
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+
+	if (i915_gem_object_is_userptr(obj)) {
+		drm_dbg(&i915->drm,
+			"attempting to use a userptr for a framebuffer, denied\n");
+		return -EINVAL;
+	}
+
+	return drm_gem_handle_create(file, &obj->base, handle);
+}
+
+static int intel_user_framebuffer_dirty(struct drm_framebuffer *fb,
+					struct drm_file *file,
+					unsigned int flags, unsigned int color,
+					struct drm_clip_rect *clips,
+					unsigned int num_clips)
+{
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+
+	i915_gem_object_flush_if_display(obj);
+	intel_frontbuffer_flush(to_intel_frontbuffer(fb), ORIGIN_DIRTYFB);
+
+	return 0;
+}
+
+static const struct drm_framebuffer_funcs intel_fb_funcs = {
+	.destroy = intel_user_framebuffer_destroy,
+	.create_handle = intel_user_framebuffer_create_handle,
+	.dirty = intel_user_framebuffer_dirty,
+};
+
+int intel_framebuffer_init(struct intel_framebuffer *intel_fb,
+			   struct drm_i915_gem_object *obj,
+			   struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
+	struct drm_framebuffer *fb = &intel_fb->base;
+	u32 max_stride;
+	unsigned int tiling, stride;
+	int ret = -EINVAL;
+	int i;
+
+	intel_fb->frontbuffer = intel_frontbuffer_get(obj);
+	if (!intel_fb->frontbuffer)
+		return -ENOMEM;
+
+	i915_gem_object_lock(obj, NULL);
+	tiling = i915_gem_object_get_tiling(obj);
+	stride = i915_gem_object_get_stride(obj);
+	i915_gem_object_unlock(obj);
+
+	if (mode_cmd->flags & DRM_MODE_FB_MODIFIERS) {
+		/*
+		 * If there's a fence, enforce that
+		 * the fb modifier and tiling mode match.
+		 */
+		if (tiling != I915_TILING_NONE &&
+		    tiling != intel_fb_modifier_to_tiling(mode_cmd->modifier[0])) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "tiling_mode doesn't match fb modifier\n");
+			goto err;
+		}
+	} else {
+		if (tiling == I915_TILING_X) {
+			mode_cmd->modifier[0] = I915_FORMAT_MOD_X_TILED;
+		} else if (tiling == I915_TILING_Y) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "No Y tiling for legacy addfb\n");
+			goto err;
+		}
+	}
+
+	if (!drm_any_plane_has_format(&dev_priv->drm,
+				      mode_cmd->pixel_format,
+				      mode_cmd->modifier[0])) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "unsupported pixel format %p4cc / modifier 0x%llx\n",
+			    &mode_cmd->pixel_format, mode_cmd->modifier[0]);
+		goto err;
+	}
+
+	/*
+	 * gen2/3 display engine uses the fence if present,
+	 * so the tiling mode must match the fb modifier exactly.
+	 */
+	if (DISPLAY_VER(dev_priv) < 4 &&
+	    tiling != intel_fb_modifier_to_tiling(mode_cmd->modifier[0])) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "tiling_mode must match fb modifier exactly on gen2/3\n");
+		goto err;
+	}
+
+	max_stride = intel_fb_max_stride(dev_priv, mode_cmd->pixel_format,
+					 mode_cmd->modifier[0]);
+	if (mode_cmd->pitches[0] > max_stride) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "%s pitch (%u) must be at most %d\n",
+			    mode_cmd->modifier[0] != DRM_FORMAT_MOD_LINEAR ?
+			    "tiled" : "linear",
+			    mode_cmd->pitches[0], max_stride);
+		goto err;
+	}
+
+	/*
+	 * If there's a fence, enforce that
+	 * the fb pitch and fence stride match.
+	 */
+	if (tiling != I915_TILING_NONE && mode_cmd->pitches[0] != stride) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "pitch (%d) must match tiling stride (%d)\n",
+			    mode_cmd->pitches[0], stride);
+		goto err;
+	}
+
+	/* FIXME need to adjust LINOFF/TILEOFF accordingly. */
+	if (mode_cmd->offsets[0] != 0) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "plane 0 offset (0x%08x) must be 0\n",
+			    mode_cmd->offsets[0]);
+		goto err;
+	}
+
+	drm_helper_mode_fill_fb_struct(&dev_priv->drm, fb, mode_cmd);
+
+	for (i = 0; i < fb->format->num_planes; i++) {
+		u32 stride_alignment;
+
+		if (mode_cmd->handles[i] != mode_cmd->handles[0]) {
+			drm_dbg_kms(&dev_priv->drm, "bad plane %d handle\n",
+				    i);
+			goto err;
+		}
+
+		stride_alignment = intel_fb_stride_alignment(fb, i);
+		if (fb->pitches[i] & (stride_alignment - 1)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "plane %d pitch (%d) must be at least %u byte aligned\n",
+				    i, fb->pitches[i], stride_alignment);
+			goto err;
+		}
+
+		if (intel_fb_is_gen12_ccs_aux_plane(fb, i)) {
+			int ccs_aux_stride = gen12_ccs_aux_stride(intel_fb, i);
+
+			if (fb->pitches[i] != ccs_aux_stride) {
+				drm_dbg_kms(&dev_priv->drm,
+					    "ccs aux plane %d pitch (%d) must be %d\n",
+					    i,
+					    fb->pitches[i], ccs_aux_stride);
+				goto err;
+			}
+		}
+
+		fb->obj[i] = &obj->base;
+	}
+
+	ret = intel_fill_fb_info(dev_priv, intel_fb);
+	if (ret)
+		goto err;
+
+	if (intel_fb_uses_dpt(fb)) {
+		struct i915_address_space *vm;
+
+		vm = intel_dpt_create(intel_fb);
+		if (IS_ERR(vm)) {
+			ret = PTR_ERR(vm);
+			goto err;
+		}
+
+		intel_fb->dpt_vm = vm;
+	}
+
+	ret = drm_framebuffer_init(&dev_priv->drm, fb, &intel_fb_funcs);
+	if (ret) {
+		drm_err(&dev_priv->drm, "framebuffer init failed %d\n", ret);
+		goto err;
+	}
+
+	return 0;
+
+err:
+	intel_frontbuffer_put(intel_fb->frontbuffer);
+	return ret;
+}
+
+struct drm_framebuffer *
+intel_user_framebuffer_create(struct drm_device *dev,
+			      struct drm_file *filp,
+			      const struct drm_mode_fb_cmd2 *user_mode_cmd)
+{
+	struct drm_framebuffer *fb;
+	struct drm_i915_gem_object *obj;
+	struct drm_mode_fb_cmd2 mode_cmd = *user_mode_cmd;
+	struct drm_i915_private *i915;
+
+	obj = i915_gem_object_lookup(filp, mode_cmd.handles[0]);
+	if (!obj)
+		return ERR_PTR(-ENOENT);
+
+	/* object is backed with LMEM for discrete */
+	i915 = to_i915(obj->base.dev);
+	if (HAS_LMEM(i915) && !i915_gem_object_can_migrate(obj, INTEL_REGION_LMEM_0)) {
+		/* object is "remote", not in local memory */
+		i915_gem_object_put(obj);
+		return ERR_PTR(-EREMOTE);
+	}
+
+	fb = intel_framebuffer_create(obj, &mode_cmd);
+	i915_gem_object_put(obj);
+
+	return fb;
+}
+
+struct drm_framebuffer *
+intel_framebuffer_create(struct drm_i915_gem_object *obj,
+			 struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	struct intel_framebuffer *intel_fb;
+	int ret;
+
+	intel_fb = kzalloc(sizeof(*intel_fb), GFP_KERNEL);
+	if (!intel_fb)
+		return ERR_PTR(-ENOMEM);
+
+	ret = intel_framebuffer_init(intel_fb, obj, mode_cmd);
+	if (ret)
+		goto err;
+
+	return &intel_fb->base;
+
+err:
+	kfree(intel_fb);
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_fb.h b/drivers/gpu/drm/i915/display/intel_fb.h
new file mode 100644
index 000000000..12386f13a
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fb.h
@@ -0,0 +1,96 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020-2021 Intel Corporation
+ */
+
+#ifndef __INTEL_FB_H__
+#define __INTEL_FB_H__
+
+#include <linux/bits.h>
+#include <linux/types.h>
+
+struct drm_device;
+struct drm_file;
+struct drm_framebuffer;
+struct drm_i915_gem_object;
+struct drm_i915_private;
+struct drm_mode_fb_cmd2;
+struct intel_fb_view;
+struct intel_framebuffer;
+struct intel_plane;
+struct intel_plane_state;
+
+#define INTEL_PLANE_CAP_NONE		0
+#define INTEL_PLANE_CAP_CCS_RC		BIT(0)
+#define INTEL_PLANE_CAP_CCS_RC_CC	BIT(1)
+#define INTEL_PLANE_CAP_CCS_MC		BIT(2)
+#define INTEL_PLANE_CAP_TILING_X	BIT(3)
+#define INTEL_PLANE_CAP_TILING_Y	BIT(4)
+#define INTEL_PLANE_CAP_TILING_Yf	BIT(5)
+#define INTEL_PLANE_CAP_TILING_4	BIT(6)
+
+bool intel_fb_is_ccs_modifier(u64 modifier);
+bool intel_fb_is_rc_ccs_cc_modifier(u64 modifier);
+bool intel_fb_is_mc_ccs_modifier(u64 modifier);
+
+bool intel_fb_is_ccs_aux_plane(const struct drm_framebuffer *fb, int color_plane);
+int intel_fb_rc_ccs_cc_plane(const struct drm_framebuffer *fb);
+
+u64 *intel_fb_plane_get_modifiers(struct drm_i915_private *i915,
+				  u8 plane_caps);
+bool intel_fb_plane_supports_modifier(struct intel_plane *plane, u64 modifier);
+
+const struct drm_format_info *
+intel_fb_get_format_info(const struct drm_mode_fb_cmd2 *cmd);
+
+bool
+intel_format_info_is_yuv_semiplanar(const struct drm_format_info *info,
+				    u64 modifier);
+
+bool is_surface_linear(const struct drm_framebuffer *fb, int color_plane);
+
+int main_to_ccs_plane(const struct drm_framebuffer *fb, int main_plane);
+int skl_ccs_to_main_plane(const struct drm_framebuffer *fb, int ccs_plane);
+int skl_main_to_aux_plane(const struct drm_framebuffer *fb, int main_plane);
+
+unsigned int intel_tile_size(const struct drm_i915_private *i915);
+unsigned int intel_tile_width_bytes(const struct drm_framebuffer *fb, int color_plane);
+unsigned int intel_tile_height(const struct drm_framebuffer *fb, int color_plane);
+unsigned int intel_tile_row_size(const struct drm_framebuffer *fb, int color_plane);
+unsigned int intel_fb_align_height(const struct drm_framebuffer *fb,
+				   int color_plane, unsigned int height);
+unsigned int intel_cursor_alignment(const struct drm_i915_private *i915);
+unsigned int intel_surf_alignment(const struct drm_framebuffer *fb,
+				  int color_plane);
+
+void intel_fb_plane_get_subsampling(int *hsub, int *vsub,
+				    const struct drm_framebuffer *fb,
+				    int color_plane);
+
+u32 intel_plane_adjust_aligned_offset(int *x, int *y,
+				      const struct intel_plane_state *state,
+				      int color_plane,
+				      u32 old_offset, u32 new_offset);
+u32 intel_plane_compute_aligned_offset(int *x, int *y,
+				       const struct intel_plane_state *state,
+				       int color_plane);
+
+bool intel_fb_needs_pot_stride_remap(const struct intel_framebuffer *fb);
+bool intel_fb_supports_90_270_rotation(const struct intel_framebuffer *fb);
+
+int intel_fill_fb_info(struct drm_i915_private *i915, struct intel_framebuffer *fb);
+void intel_fb_fill_view(const struct intel_framebuffer *fb, unsigned int rotation,
+			struct intel_fb_view *view);
+int intel_plane_compute_gtt(struct intel_plane_state *plane_state);
+
+int intel_framebuffer_init(struct intel_framebuffer *ifb,
+			   struct drm_i915_gem_object *obj,
+			   struct drm_mode_fb_cmd2 *mode_cmd);
+struct drm_framebuffer *
+intel_user_framebuffer_create(struct drm_device *dev,
+			      struct drm_file *filp,
+			      const struct drm_mode_fb_cmd2 *user_mode_cmd);
+
+bool intel_fb_uses_dpt(const struct drm_framebuffer *fb);
+
+#endif /* __INTEL_FB_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_fb_pin.c b/drivers/gpu/drm/i915/display/intel_fb_pin.c
new file mode 100644
index 000000000..1dddd6abd
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fb_pin.c
@@ -0,0 +1,303 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+/**
+ * DOC: display pinning helpers
+ */
+
+#include "gem/i915_gem_domain.h"
+#include "gem/i915_gem_object.h"
+
+#include "i915_drv.h"
+#include "intel_display_types.h"
+#include "intel_dpt.h"
+#include "intel_fb.h"
+#include "intel_fb_pin.h"
+
+static struct i915_vma *
+intel_pin_fb_obj_dpt(struct drm_framebuffer *fb,
+		     const struct i915_gtt_view *view,
+		     bool uses_fence,
+		     unsigned long *out_flags,
+		     struct i915_address_space *vm)
+{
+	struct drm_device *dev = fb->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	struct i915_gem_ww_ctx ww;
+	struct i915_vma *vma;
+	u32 alignment;
+	int ret;
+
+	/*
+	 * We are not syncing against the binding (and potential migrations)
+	 * below, so this vm must never be async.
+	 */
+	GEM_WARN_ON(vm->bind_async_flags);
+
+	if (WARN_ON(!i915_gem_object_is_framebuffer(obj)))
+		return ERR_PTR(-EINVAL);
+
+	alignment = 4096 * 512;
+
+	atomic_inc(&dev_priv->gpu_error.pending_fb_pin);
+
+	for_i915_gem_ww(&ww, ret, true) {
+		ret = i915_gem_object_lock(obj, &ww);
+		if (ret)
+			continue;
+
+		if (HAS_LMEM(dev_priv)) {
+			unsigned int flags = obj->flags;
+
+			/*
+			 * For this type of buffer we need to able to read from the CPU
+			 * the clear color value found in the buffer, hence we need to
+			 * ensure it is always in the mappable part of lmem, if this is
+			 * a small-bar device.
+			 */
+			if (intel_fb_rc_ccs_cc_plane(fb) >= 0)
+				flags &= ~I915_BO_ALLOC_GPU_ONLY;
+			ret = __i915_gem_object_migrate(obj, &ww, INTEL_REGION_LMEM_0,
+							flags);
+			if (ret)
+				continue;
+		}
+
+		ret = i915_gem_object_set_cache_level(obj, I915_CACHE_NONE);
+		if (ret)
+			continue;
+
+		vma = i915_vma_instance(obj, vm, view);
+		if (IS_ERR(vma)) {
+			ret = PTR_ERR(vma);
+			continue;
+		}
+
+		if (i915_vma_misplaced(vma, 0, alignment, 0)) {
+			ret = i915_vma_unbind(vma);
+			if (ret)
+				continue;
+		}
+
+		ret = i915_vma_pin_ww(vma, &ww, 0, alignment, PIN_GLOBAL);
+		if (ret)
+			continue;
+	}
+	if (ret) {
+		vma = ERR_PTR(ret);
+		goto err;
+	}
+
+	vma->display_alignment = max_t(u64, vma->display_alignment, alignment);
+
+	i915_gem_object_flush_if_display(obj);
+
+	i915_vma_get(vma);
+err:
+	atomic_dec(&dev_priv->gpu_error.pending_fb_pin);
+
+	return vma;
+}
+
+struct i915_vma *
+intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb,
+			   bool phys_cursor,
+			   const struct i915_gtt_view *view,
+			   bool uses_fence,
+			   unsigned long *out_flags)
+{
+	struct drm_device *dev = fb->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	intel_wakeref_t wakeref;
+	struct i915_gem_ww_ctx ww;
+	struct i915_vma *vma;
+	unsigned int pinctl;
+	u32 alignment;
+	int ret;
+
+	if (drm_WARN_ON(dev, !i915_gem_object_is_framebuffer(obj)))
+		return ERR_PTR(-EINVAL);
+
+	if (phys_cursor)
+		alignment = intel_cursor_alignment(dev_priv);
+	else
+		alignment = intel_surf_alignment(fb, 0);
+	if (drm_WARN_ON(dev, alignment && !is_power_of_2(alignment)))
+		return ERR_PTR(-EINVAL);
+
+	/* Note that the w/a also requires 64 PTE of padding following the
+	 * bo. We currently fill all unused PTE with the shadow page and so
+	 * we should always have valid PTE following the scanout preventing
+	 * the VT-d warning.
+	 */
+	if (intel_scanout_needs_vtd_wa(dev_priv) && alignment < 256 * 1024)
+		alignment = 256 * 1024;
+
+	/*
+	 * Global gtt pte registers are special registers which actually forward
+	 * writes to a chunk of system memory. Which means that there is no risk
+	 * that the register values disappear as soon as we call
+	 * intel_runtime_pm_put(), so it is correct to wrap only the
+	 * pin/unpin/fence and not more.
+	 */
+	wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+
+	atomic_inc(&dev_priv->gpu_error.pending_fb_pin);
+
+	/*
+	 * Valleyview is definitely limited to scanning out the first
+	 * 512MiB. Lets presume this behaviour was inherited from the
+	 * g4x display engine and that all earlier gen are similarly
+	 * limited. Testing suggests that it is a little more
+	 * complicated than this. For example, Cherryview appears quite
+	 * happy to scanout from anywhere within its global aperture.
+	 */
+	pinctl = 0;
+	if (HAS_GMCH(dev_priv))
+		pinctl |= PIN_MAPPABLE;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	ret = i915_gem_object_lock(obj, &ww);
+	if (!ret && phys_cursor)
+		ret = i915_gem_object_attach_phys(obj, alignment);
+	else if (!ret && HAS_LMEM(dev_priv))
+		ret = i915_gem_object_migrate(obj, &ww, INTEL_REGION_LMEM_0);
+	/* TODO: Do we need to sync when migration becomes async? */
+	if (!ret)
+		ret = i915_gem_object_pin_pages(obj);
+	if (ret)
+		goto err;
+
+	vma = i915_gem_object_pin_to_display_plane(obj, &ww, alignment,
+						   view, pinctl);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto err_unpin;
+	}
+
+	if (uses_fence && i915_vma_is_map_and_fenceable(vma)) {
+		/*
+		 * Install a fence for tiled scan-out. Pre-i965 always needs a
+		 * fence, whereas 965+ only requires a fence if using
+		 * framebuffer compression.  For simplicity, we always, when
+		 * possible, install a fence as the cost is not that onerous.
+		 *
+		 * If we fail to fence the tiled scanout, then either the
+		 * modeset will reject the change (which is highly unlikely as
+		 * the affected systems, all but one, do not have unmappable
+		 * space) or we will not be able to enable full powersaving
+		 * techniques (also likely not to apply due to various limits
+		 * FBC and the like impose on the size of the buffer, which
+		 * presumably we violated anyway with this unmappable buffer).
+		 * Anyway, it is presumably better to stumble onwards with
+		 * something and try to run the system in a "less than optimal"
+		 * mode that matches the user configuration.
+		 */
+		ret = i915_vma_pin_fence(vma);
+		if (ret != 0 && DISPLAY_VER(dev_priv) < 4) {
+			i915_vma_unpin(vma);
+			goto err_unpin;
+		}
+		ret = 0;
+
+		if (vma->fence)
+			*out_flags |= PLANE_HAS_FENCE;
+	}
+
+	i915_vma_get(vma);
+
+err_unpin:
+	i915_gem_object_unpin_pages(obj);
+err:
+	if (ret == -EDEADLK) {
+		ret = i915_gem_ww_ctx_backoff(&ww);
+		if (!ret)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	if (ret)
+		vma = ERR_PTR(ret);
+
+	atomic_dec(&dev_priv->gpu_error.pending_fb_pin);
+	intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+	return vma;
+}
+
+void intel_unpin_fb_vma(struct i915_vma *vma, unsigned long flags)
+{
+	if (flags & PLANE_HAS_FENCE)
+		i915_vma_unpin_fence(vma);
+	i915_vma_unpin(vma);
+	i915_vma_put(vma);
+}
+
+int intel_plane_pin_fb(struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	struct drm_framebuffer *fb = plane_state->hw.fb;
+	struct i915_vma *vma;
+	bool phys_cursor =
+		plane->id == PLANE_CURSOR &&
+		INTEL_INFO(dev_priv)->display.cursor_needs_physical;
+
+	if (!intel_fb_uses_dpt(fb)) {
+		vma = intel_pin_and_fence_fb_obj(fb, phys_cursor,
+						 &plane_state->view.gtt,
+						 intel_plane_uses_fence(plane_state),
+						 &plane_state->flags);
+		if (IS_ERR(vma))
+			return PTR_ERR(vma);
+
+		plane_state->ggtt_vma = vma;
+	} else {
+		struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
+
+		vma = intel_dpt_pin(intel_fb->dpt_vm);
+		if (IS_ERR(vma))
+			return PTR_ERR(vma);
+
+		plane_state->ggtt_vma = vma;
+
+		vma = intel_pin_fb_obj_dpt(fb, &plane_state->view.gtt, false,
+					   &plane_state->flags, intel_fb->dpt_vm);
+		if (IS_ERR(vma)) {
+			intel_dpt_unpin(intel_fb->dpt_vm);
+			plane_state->ggtt_vma = NULL;
+			return PTR_ERR(vma);
+		}
+
+		plane_state->dpt_vma = vma;
+
+		WARN_ON(plane_state->ggtt_vma == plane_state->dpt_vma);
+	}
+
+	return 0;
+}
+
+void intel_plane_unpin_fb(struct intel_plane_state *old_plane_state)
+{
+	struct drm_framebuffer *fb = old_plane_state->hw.fb;
+	struct i915_vma *vma;
+
+	if (!intel_fb_uses_dpt(fb)) {
+		vma = fetch_and_zero(&old_plane_state->ggtt_vma);
+		if (vma)
+			intel_unpin_fb_vma(vma, old_plane_state->flags);
+	} else {
+		struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
+
+		vma = fetch_and_zero(&old_plane_state->dpt_vma);
+		if (vma)
+			intel_unpin_fb_vma(vma, old_plane_state->flags);
+
+		vma = fetch_and_zero(&old_plane_state->ggtt_vma);
+		if (vma)
+			intel_dpt_unpin(intel_fb->dpt_vm);
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_fb_pin.h b/drivers/gpu/drm/i915/display/intel_fb_pin.h
new file mode 100644
index 000000000..de0efaa25
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fb_pin.h
@@ -0,0 +1,28 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __INTEL_FB_PIN_H__
+#define __INTEL_FB_PIN_H__
+
+#include <linux/types.h>
+
+struct drm_framebuffer;
+struct i915_vma;
+struct intel_plane_state;
+struct i915_gtt_view;
+
+struct i915_vma *
+intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb,
+			   bool phys_cursor,
+			   const struct i915_gtt_view *view,
+			   bool uses_fence,
+			   unsigned long *out_flags);
+
+void intel_unpin_fb_vma(struct i915_vma *vma, unsigned long flags);
+
+int intel_plane_pin_fb(struct intel_plane_state *plane_state);
+void intel_plane_unpin_fb(struct intel_plane_state *old_plane_state);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_fbc.c b/drivers/gpu/drm/i915/display/intel_fbc.c
new file mode 100644
index 000000000..f38175304
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fbc.c
@@ -0,0 +1,1846 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+/**
+ * DOC: Frame Buffer Compression (FBC)
+ *
+ * FBC tries to save memory bandwidth (and so power consumption) by
+ * compressing the amount of memory used by the display. It is total
+ * transparent to user space and completely handled in the kernel.
+ *
+ * The benefits of FBC are mostly visible with solid backgrounds and
+ * variation-less patterns. It comes from keeping the memory footprint small
+ * and having fewer memory pages opened and accessed for refreshing the display.
+ *
+ * i915 is responsible to reserve stolen memory for FBC and configure its
+ * offset on proper registers. The hardware takes care of all
+ * compress/decompress. However there are many known cases where we have to
+ * forcibly disable it to allow proper screen updates.
+ */
+
+#include <linux/string_helpers.h>
+
+#include <drm/drm_blend.h>
+#include <drm/drm_fourcc.h>
+
+#include "i915_drv.h"
+#include "i915_utils.h"
+#include "i915_vgpu.h"
+#include "intel_cdclk.h"
+#include "intel_de.h"
+#include "intel_display_trace.h"
+#include "intel_display_types.h"
+#include "intel_fbc.h"
+#include "intel_frontbuffer.h"
+
+#define for_each_fbc_id(__dev_priv, __fbc_id) \
+	for ((__fbc_id) = INTEL_FBC_A; (__fbc_id) < I915_MAX_FBCS; (__fbc_id)++) \
+		for_each_if(RUNTIME_INFO(__dev_priv)->fbc_mask & BIT(__fbc_id))
+
+#define for_each_intel_fbc(__dev_priv, __fbc, __fbc_id) \
+	for_each_fbc_id((__dev_priv), (__fbc_id)) \
+		for_each_if((__fbc) = (__dev_priv)->display.fbc[(__fbc_id)])
+
+struct intel_fbc_funcs {
+	void (*activate)(struct intel_fbc *fbc);
+	void (*deactivate)(struct intel_fbc *fbc);
+	bool (*is_active)(struct intel_fbc *fbc);
+	bool (*is_compressing)(struct intel_fbc *fbc);
+	void (*nuke)(struct intel_fbc *fbc);
+	void (*program_cfb)(struct intel_fbc *fbc);
+	void (*set_false_color)(struct intel_fbc *fbc, bool enable);
+};
+
+struct intel_fbc_state {
+	struct intel_plane *plane;
+	unsigned int cfb_stride;
+	unsigned int cfb_size;
+	unsigned int fence_y_offset;
+	u16 override_cfb_stride;
+	u16 interval;
+	s8 fence_id;
+};
+
+struct intel_fbc {
+	struct drm_i915_private *i915;
+	const struct intel_fbc_funcs *funcs;
+
+	/*
+	 * This is always the inner lock when overlapping with
+	 * struct_mutex and it's the outer lock when overlapping
+	 * with stolen_lock.
+	 */
+	struct mutex lock;
+	unsigned int busy_bits;
+
+	struct drm_mm_node compressed_fb;
+	struct drm_mm_node compressed_llb;
+
+	enum intel_fbc_id id;
+
+	u8 limit;
+
+	bool false_color;
+
+	bool active;
+	bool activated;
+	bool flip_pending;
+
+	bool underrun_detected;
+	struct work_struct underrun_work;
+
+	/*
+	 * This structure contains everything that's relevant to program the
+	 * hardware registers. When we want to figure out if we need to disable
+	 * and re-enable FBC for a new configuration we just check if there's
+	 * something different in the struct. The genx_fbc_activate functions
+	 * are supposed to read from it in order to program the registers.
+	 */
+	struct intel_fbc_state state;
+	const char *no_fbc_reason;
+};
+
+/* plane stride in pixels */
+static unsigned int intel_fbc_plane_stride(const struct intel_plane_state *plane_state)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int stride;
+
+	stride = plane_state->view.color_plane[0].mapping_stride;
+	if (!drm_rotation_90_or_270(plane_state->hw.rotation))
+		stride /= fb->format->cpp[0];
+
+	return stride;
+}
+
+/* plane stride based cfb stride in bytes, assuming 1:1 compression limit */
+static unsigned int _intel_fbc_cfb_stride(const struct intel_plane_state *plane_state)
+{
+	unsigned int cpp = 4; /* FBC always 4 bytes per pixel */
+
+	return intel_fbc_plane_stride(plane_state) * cpp;
+}
+
+/* minimum acceptable cfb stride in bytes, assuming 1:1 compression limit */
+static unsigned int skl_fbc_min_cfb_stride(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	unsigned int limit = 4; /* 1:4 compression limit is the worst case */
+	unsigned int cpp = 4; /* FBC always 4 bytes per pixel */
+	unsigned int width = drm_rect_width(&plane_state->uapi.src) >> 16;
+	unsigned int height = 4; /* FBC segment is 4 lines */
+	unsigned int stride;
+
+	/* minimum segment stride we can use */
+	stride = width * cpp * height / limit;
+
+	/*
+	 * Wa_16011863758: icl+
+	 * Avoid some hardware segment address miscalculation.
+	 */
+	if (DISPLAY_VER(i915) >= 11)
+		stride += 64;
+
+	/*
+	 * At least some of the platforms require each 4 line segment to
+	 * be 512 byte aligned. Just do it always for simplicity.
+	 */
+	stride = ALIGN(stride, 512);
+
+	/* convert back to single line equivalent with 1:1 compression limit */
+	return stride * limit / height;
+}
+
+/* properly aligned cfb stride in bytes, assuming 1:1 compression limit */
+static unsigned int intel_fbc_cfb_stride(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	unsigned int stride = _intel_fbc_cfb_stride(plane_state);
+
+	/*
+	 * At least some of the platforms require each 4 line segment to
+	 * be 512 byte aligned. Aligning each line to 512 bytes guarantees
+	 * that regardless of the compression limit we choose later.
+	 */
+	if (DISPLAY_VER(i915) >= 9)
+		return max(ALIGN(stride, 512), skl_fbc_min_cfb_stride(plane_state));
+	else
+		return stride;
+}
+
+static unsigned int intel_fbc_cfb_size(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	int lines = drm_rect_height(&plane_state->uapi.src) >> 16;
+
+	if (DISPLAY_VER(i915) == 7)
+		lines = min(lines, 2048);
+	else if (DISPLAY_VER(i915) >= 8)
+		lines = min(lines, 2560);
+
+	return lines * intel_fbc_cfb_stride(plane_state);
+}
+
+static u16 intel_fbc_override_cfb_stride(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	unsigned int stride_aligned = intel_fbc_cfb_stride(plane_state);
+	unsigned int stride = _intel_fbc_cfb_stride(plane_state);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+
+	/*
+	 * Override stride in 64 byte units per 4 line segment.
+	 *
+	 * Gen9 hw miscalculates cfb stride for linear as
+	 * PLANE_STRIDE*512 instead of PLANE_STRIDE*64, so
+	 * we always need to use the override there.
+	 */
+	if (stride != stride_aligned ||
+	    (DISPLAY_VER(i915) == 9 && fb->modifier == DRM_FORMAT_MOD_LINEAR))
+		return stride_aligned * 4 / 64;
+
+	return 0;
+}
+
+static u32 i8xx_fbc_ctl(struct intel_fbc *fbc)
+{
+	const struct intel_fbc_state *fbc_state = &fbc->state;
+	struct drm_i915_private *i915 = fbc->i915;
+	unsigned int cfb_stride;
+	u32 fbc_ctl;
+
+	cfb_stride = fbc_state->cfb_stride / fbc->limit;
+
+	/* FBC_CTL wants 32B or 64B units */
+	if (DISPLAY_VER(i915) == 2)
+		cfb_stride = (cfb_stride / 32) - 1;
+	else
+		cfb_stride = (cfb_stride / 64) - 1;
+
+	fbc_ctl = FBC_CTL_PERIODIC |
+		FBC_CTL_INTERVAL(fbc_state->interval) |
+		FBC_CTL_STRIDE(cfb_stride);
+
+	if (IS_I945GM(i915))
+		fbc_ctl |= FBC_CTL_C3_IDLE; /* 945 needs special SR handling */
+
+	if (fbc_state->fence_id >= 0)
+		fbc_ctl |= FBC_CTL_FENCENO(fbc_state->fence_id);
+
+	return fbc_ctl;
+}
+
+static u32 i965_fbc_ctl2(struct intel_fbc *fbc)
+{
+	const struct intel_fbc_state *fbc_state = &fbc->state;
+	u32 fbc_ctl2;
+
+	fbc_ctl2 = FBC_CTL_FENCE_DBL | FBC_CTL_IDLE_IMM |
+		FBC_CTL_PLANE(fbc_state->plane->i9xx_plane);
+
+	if (fbc_state->fence_id >= 0)
+		fbc_ctl2 |= FBC_CTL_CPU_FENCE_EN;
+
+	return fbc_ctl2;
+}
+
+static void i8xx_fbc_deactivate(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+	u32 fbc_ctl;
+
+	/* Disable compression */
+	fbc_ctl = intel_de_read(i915, FBC_CONTROL);
+	if ((fbc_ctl & FBC_CTL_EN) == 0)
+		return;
+
+	fbc_ctl &= ~FBC_CTL_EN;
+	intel_de_write(i915, FBC_CONTROL, fbc_ctl);
+
+	/* Wait for compressing bit to clear */
+	if (intel_de_wait_for_clear(i915, FBC_STATUS,
+				    FBC_STAT_COMPRESSING, 10)) {
+		drm_dbg_kms(&i915->drm, "FBC idle timed out\n");
+		return;
+	}
+}
+
+static void i8xx_fbc_activate(struct intel_fbc *fbc)
+{
+	const struct intel_fbc_state *fbc_state = &fbc->state;
+	struct drm_i915_private *i915 = fbc->i915;
+	int i;
+
+	/* Clear old tags */
+	for (i = 0; i < (FBC_LL_SIZE / 32) + 1; i++)
+		intel_de_write(i915, FBC_TAG(i), 0);
+
+	if (DISPLAY_VER(i915) == 4) {
+		intel_de_write(i915, FBC_CONTROL2,
+			       i965_fbc_ctl2(fbc));
+		intel_de_write(i915, FBC_FENCE_OFF,
+			       fbc_state->fence_y_offset);
+	}
+
+	intel_de_write(i915, FBC_CONTROL,
+		       FBC_CTL_EN | i8xx_fbc_ctl(fbc));
+}
+
+static bool i8xx_fbc_is_active(struct intel_fbc *fbc)
+{
+	return intel_de_read(fbc->i915, FBC_CONTROL) & FBC_CTL_EN;
+}
+
+static bool i8xx_fbc_is_compressing(struct intel_fbc *fbc)
+{
+	return intel_de_read(fbc->i915, FBC_STATUS) &
+		(FBC_STAT_COMPRESSING | FBC_STAT_COMPRESSED);
+}
+
+static void i8xx_fbc_nuke(struct intel_fbc *fbc)
+{
+	struct intel_fbc_state *fbc_state = &fbc->state;
+	enum i9xx_plane_id i9xx_plane = fbc_state->plane->i9xx_plane;
+	struct drm_i915_private *dev_priv = fbc->i915;
+
+	spin_lock_irq(&dev_priv->uncore.lock);
+	intel_de_write_fw(dev_priv, DSPADDR(i9xx_plane),
+			  intel_de_read_fw(dev_priv, DSPADDR(i9xx_plane)));
+	spin_unlock_irq(&dev_priv->uncore.lock);
+}
+
+static void i8xx_fbc_program_cfb(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+
+	GEM_BUG_ON(range_overflows_end_t(u64, i915->dsm.start,
+					 fbc->compressed_fb.start, U32_MAX));
+	GEM_BUG_ON(range_overflows_end_t(u64, i915->dsm.start,
+					 fbc->compressed_llb.start, U32_MAX));
+
+	intel_de_write(i915, FBC_CFB_BASE,
+		       i915->dsm.start + fbc->compressed_fb.start);
+	intel_de_write(i915, FBC_LL_BASE,
+		       i915->dsm.start + fbc->compressed_llb.start);
+}
+
+static const struct intel_fbc_funcs i8xx_fbc_funcs = {
+	.activate = i8xx_fbc_activate,
+	.deactivate = i8xx_fbc_deactivate,
+	.is_active = i8xx_fbc_is_active,
+	.is_compressing = i8xx_fbc_is_compressing,
+	.nuke = i8xx_fbc_nuke,
+	.program_cfb = i8xx_fbc_program_cfb,
+};
+
+static void i965_fbc_nuke(struct intel_fbc *fbc)
+{
+	struct intel_fbc_state *fbc_state = &fbc->state;
+	enum i9xx_plane_id i9xx_plane = fbc_state->plane->i9xx_plane;
+	struct drm_i915_private *dev_priv = fbc->i915;
+
+	spin_lock_irq(&dev_priv->uncore.lock);
+	intel_de_write_fw(dev_priv, DSPSURF(i9xx_plane),
+			  intel_de_read_fw(dev_priv, DSPSURF(i9xx_plane)));
+	spin_unlock_irq(&dev_priv->uncore.lock);
+}
+
+static const struct intel_fbc_funcs i965_fbc_funcs = {
+	.activate = i8xx_fbc_activate,
+	.deactivate = i8xx_fbc_deactivate,
+	.is_active = i8xx_fbc_is_active,
+	.is_compressing = i8xx_fbc_is_compressing,
+	.nuke = i965_fbc_nuke,
+	.program_cfb = i8xx_fbc_program_cfb,
+};
+
+static u32 g4x_dpfc_ctl_limit(struct intel_fbc *fbc)
+{
+	switch (fbc->limit) {
+	default:
+		MISSING_CASE(fbc->limit);
+		fallthrough;
+	case 1:
+		return DPFC_CTL_LIMIT_1X;
+	case 2:
+		return DPFC_CTL_LIMIT_2X;
+	case 4:
+		return DPFC_CTL_LIMIT_4X;
+	}
+}
+
+static u32 g4x_dpfc_ctl(struct intel_fbc *fbc)
+{
+	const struct intel_fbc_state *fbc_state = &fbc->state;
+	struct drm_i915_private *i915 = fbc->i915;
+	u32 dpfc_ctl;
+
+	dpfc_ctl = g4x_dpfc_ctl_limit(fbc) |
+		DPFC_CTL_PLANE_G4X(fbc_state->plane->i9xx_plane);
+
+	if (IS_G4X(i915))
+		dpfc_ctl |= DPFC_CTL_SR_EN;
+
+	if (fbc_state->fence_id >= 0) {
+		dpfc_ctl |= DPFC_CTL_FENCE_EN_G4X;
+
+		if (DISPLAY_VER(i915) < 6)
+			dpfc_ctl |= DPFC_CTL_FENCENO(fbc_state->fence_id);
+	}
+
+	return dpfc_ctl;
+}
+
+static void g4x_fbc_activate(struct intel_fbc *fbc)
+{
+	const struct intel_fbc_state *fbc_state = &fbc->state;
+	struct drm_i915_private *i915 = fbc->i915;
+
+	intel_de_write(i915, DPFC_FENCE_YOFF,
+		       fbc_state->fence_y_offset);
+
+	intel_de_write(i915, DPFC_CONTROL,
+		       DPFC_CTL_EN | g4x_dpfc_ctl(fbc));
+}
+
+static void g4x_fbc_deactivate(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+	u32 dpfc_ctl;
+
+	/* Disable compression */
+	dpfc_ctl = intel_de_read(i915, DPFC_CONTROL);
+	if (dpfc_ctl & DPFC_CTL_EN) {
+		dpfc_ctl &= ~DPFC_CTL_EN;
+		intel_de_write(i915, DPFC_CONTROL, dpfc_ctl);
+	}
+}
+
+static bool g4x_fbc_is_active(struct intel_fbc *fbc)
+{
+	return intel_de_read(fbc->i915, DPFC_CONTROL) & DPFC_CTL_EN;
+}
+
+static bool g4x_fbc_is_compressing(struct intel_fbc *fbc)
+{
+	return intel_de_read(fbc->i915, DPFC_STATUS) & DPFC_COMP_SEG_MASK;
+}
+
+static void g4x_fbc_program_cfb(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+
+	intel_de_write(i915, DPFC_CB_BASE, fbc->compressed_fb.start);
+}
+
+static const struct intel_fbc_funcs g4x_fbc_funcs = {
+	.activate = g4x_fbc_activate,
+	.deactivate = g4x_fbc_deactivate,
+	.is_active = g4x_fbc_is_active,
+	.is_compressing = g4x_fbc_is_compressing,
+	.nuke = i965_fbc_nuke,
+	.program_cfb = g4x_fbc_program_cfb,
+};
+
+static void ilk_fbc_activate(struct intel_fbc *fbc)
+{
+	struct intel_fbc_state *fbc_state = &fbc->state;
+	struct drm_i915_private *i915 = fbc->i915;
+
+	intel_de_write(i915, ILK_DPFC_FENCE_YOFF(fbc->id),
+		       fbc_state->fence_y_offset);
+
+	intel_de_write(i915, ILK_DPFC_CONTROL(fbc->id),
+		       DPFC_CTL_EN | g4x_dpfc_ctl(fbc));
+}
+
+static void ilk_fbc_deactivate(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+	u32 dpfc_ctl;
+
+	/* Disable compression */
+	dpfc_ctl = intel_de_read(i915, ILK_DPFC_CONTROL(fbc->id));
+	if (dpfc_ctl & DPFC_CTL_EN) {
+		dpfc_ctl &= ~DPFC_CTL_EN;
+		intel_de_write(i915, ILK_DPFC_CONTROL(fbc->id), dpfc_ctl);
+	}
+}
+
+static bool ilk_fbc_is_active(struct intel_fbc *fbc)
+{
+	return intel_de_read(fbc->i915, ILK_DPFC_CONTROL(fbc->id)) & DPFC_CTL_EN;
+}
+
+static bool ilk_fbc_is_compressing(struct intel_fbc *fbc)
+{
+	return intel_de_read(fbc->i915, ILK_DPFC_STATUS(fbc->id)) & DPFC_COMP_SEG_MASK;
+}
+
+static void ilk_fbc_program_cfb(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+
+	intel_de_write(i915, ILK_DPFC_CB_BASE(fbc->id), fbc->compressed_fb.start);
+}
+
+static const struct intel_fbc_funcs ilk_fbc_funcs = {
+	.activate = ilk_fbc_activate,
+	.deactivate = ilk_fbc_deactivate,
+	.is_active = ilk_fbc_is_active,
+	.is_compressing = ilk_fbc_is_compressing,
+	.nuke = i965_fbc_nuke,
+	.program_cfb = ilk_fbc_program_cfb,
+};
+
+static void snb_fbc_program_fence(struct intel_fbc *fbc)
+{
+	const struct intel_fbc_state *fbc_state = &fbc->state;
+	struct drm_i915_private *i915 = fbc->i915;
+	u32 ctl = 0;
+
+	if (fbc_state->fence_id >= 0)
+		ctl = SNB_DPFC_FENCE_EN | SNB_DPFC_FENCENO(fbc_state->fence_id);
+
+	intel_de_write(i915, SNB_DPFC_CTL_SA, ctl);
+	intel_de_write(i915, SNB_DPFC_CPU_FENCE_OFFSET, fbc_state->fence_y_offset);
+}
+
+static void snb_fbc_activate(struct intel_fbc *fbc)
+{
+	snb_fbc_program_fence(fbc);
+
+	ilk_fbc_activate(fbc);
+}
+
+static void snb_fbc_nuke(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+
+	intel_de_write(i915, MSG_FBC_REND_STATE(fbc->id), FBC_REND_NUKE);
+	intel_de_posting_read(i915, MSG_FBC_REND_STATE(fbc->id));
+}
+
+static const struct intel_fbc_funcs snb_fbc_funcs = {
+	.activate = snb_fbc_activate,
+	.deactivate = ilk_fbc_deactivate,
+	.is_active = ilk_fbc_is_active,
+	.is_compressing = ilk_fbc_is_compressing,
+	.nuke = snb_fbc_nuke,
+	.program_cfb = ilk_fbc_program_cfb,
+};
+
+static void glk_fbc_program_cfb_stride(struct intel_fbc *fbc)
+{
+	const struct intel_fbc_state *fbc_state = &fbc->state;
+	struct drm_i915_private *i915 = fbc->i915;
+	u32 val = 0;
+
+	if (fbc_state->override_cfb_stride)
+		val |= FBC_STRIDE_OVERRIDE |
+			FBC_STRIDE(fbc_state->override_cfb_stride / fbc->limit);
+
+	intel_de_write(i915, GLK_FBC_STRIDE(fbc->id), val);
+}
+
+static void skl_fbc_program_cfb_stride(struct intel_fbc *fbc)
+{
+	const struct intel_fbc_state *fbc_state = &fbc->state;
+	struct drm_i915_private *i915 = fbc->i915;
+	u32 val = 0;
+
+	/* Display WA #0529: skl, kbl, bxt. */
+	if (fbc_state->override_cfb_stride)
+		val |= CHICKEN_FBC_STRIDE_OVERRIDE |
+			CHICKEN_FBC_STRIDE(fbc_state->override_cfb_stride / fbc->limit);
+
+	intel_de_rmw(i915, CHICKEN_MISC_4,
+		     CHICKEN_FBC_STRIDE_OVERRIDE |
+		     CHICKEN_FBC_STRIDE_MASK, val);
+}
+
+static u32 ivb_dpfc_ctl(struct intel_fbc *fbc)
+{
+	const struct intel_fbc_state *fbc_state = &fbc->state;
+	struct drm_i915_private *i915 = fbc->i915;
+	u32 dpfc_ctl;
+
+	dpfc_ctl = g4x_dpfc_ctl_limit(fbc);
+
+	if (IS_IVYBRIDGE(i915))
+		dpfc_ctl |= DPFC_CTL_PLANE_IVB(fbc_state->plane->i9xx_plane);
+
+	if (fbc_state->fence_id >= 0)
+		dpfc_ctl |= DPFC_CTL_FENCE_EN_IVB;
+
+	if (fbc->false_color)
+		dpfc_ctl |= DPFC_CTL_FALSE_COLOR;
+
+	return dpfc_ctl;
+}
+
+static void ivb_fbc_activate(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+
+	if (DISPLAY_VER(i915) >= 10)
+		glk_fbc_program_cfb_stride(fbc);
+	else if (DISPLAY_VER(i915) == 9)
+		skl_fbc_program_cfb_stride(fbc);
+
+	if (to_gt(i915)->ggtt->num_fences)
+		snb_fbc_program_fence(fbc);
+
+	intel_de_write(i915, ILK_DPFC_CONTROL(fbc->id),
+		       DPFC_CTL_EN | ivb_dpfc_ctl(fbc));
+}
+
+static bool ivb_fbc_is_compressing(struct intel_fbc *fbc)
+{
+	return intel_de_read(fbc->i915, ILK_DPFC_STATUS2(fbc->id)) & DPFC_COMP_SEG_MASK_IVB;
+}
+
+static void ivb_fbc_set_false_color(struct intel_fbc *fbc,
+				    bool enable)
+{
+	intel_de_rmw(fbc->i915, ILK_DPFC_CONTROL(fbc->id),
+		     DPFC_CTL_FALSE_COLOR, enable ? DPFC_CTL_FALSE_COLOR : 0);
+}
+
+static const struct intel_fbc_funcs ivb_fbc_funcs = {
+	.activate = ivb_fbc_activate,
+	.deactivate = ilk_fbc_deactivate,
+	.is_active = ilk_fbc_is_active,
+	.is_compressing = ivb_fbc_is_compressing,
+	.nuke = snb_fbc_nuke,
+	.program_cfb = ilk_fbc_program_cfb,
+	.set_false_color = ivb_fbc_set_false_color,
+};
+
+static bool intel_fbc_hw_is_active(struct intel_fbc *fbc)
+{
+	return fbc->funcs->is_active(fbc);
+}
+
+static void intel_fbc_hw_activate(struct intel_fbc *fbc)
+{
+	trace_intel_fbc_activate(fbc->state.plane);
+
+	fbc->active = true;
+	fbc->activated = true;
+
+	fbc->funcs->activate(fbc);
+}
+
+static void intel_fbc_hw_deactivate(struct intel_fbc *fbc)
+{
+	trace_intel_fbc_deactivate(fbc->state.plane);
+
+	fbc->active = false;
+
+	fbc->funcs->deactivate(fbc);
+}
+
+static bool intel_fbc_is_compressing(struct intel_fbc *fbc)
+{
+	return fbc->funcs->is_compressing(fbc);
+}
+
+static void intel_fbc_nuke(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+
+	drm_WARN_ON(&i915->drm, fbc->flip_pending);
+
+	trace_intel_fbc_nuke(fbc->state.plane);
+
+	fbc->funcs->nuke(fbc);
+}
+
+static void intel_fbc_activate(struct intel_fbc *fbc)
+{
+	intel_fbc_hw_activate(fbc);
+	intel_fbc_nuke(fbc);
+
+	fbc->no_fbc_reason = NULL;
+}
+
+static void intel_fbc_deactivate(struct intel_fbc *fbc, const char *reason)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+
+	drm_WARN_ON(&i915->drm, !mutex_is_locked(&fbc->lock));
+
+	if (fbc->active)
+		intel_fbc_hw_deactivate(fbc);
+
+	fbc->no_fbc_reason = reason;
+}
+
+static u64 intel_fbc_cfb_base_max(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 5 || IS_G4X(i915))
+		return BIT_ULL(28);
+	else
+		return BIT_ULL(32);
+}
+
+static u64 intel_fbc_stolen_end(struct drm_i915_private *i915)
+{
+	u64 end;
+
+	/* The FBC hardware for BDW/SKL doesn't have access to the stolen
+	 * reserved range size, so it always assumes the maximum (8mb) is used.
+	 * If we enable FBC using a CFB on that memory range we'll get FIFO
+	 * underruns, even if that range is not reserved by the BIOS. */
+	if (IS_BROADWELL(i915) ||
+	    (DISPLAY_VER(i915) == 9 && !IS_BROXTON(i915)))
+		end = resource_size(&i915->dsm) - 8 * 1024 * 1024;
+	else
+		end = U64_MAX;
+
+	return min(end, intel_fbc_cfb_base_max(i915));
+}
+
+static int intel_fbc_min_limit(const struct intel_plane_state *plane_state)
+{
+	return plane_state->hw.fb->format->cpp[0] == 2 ? 2 : 1;
+}
+
+static int intel_fbc_max_limit(struct drm_i915_private *i915)
+{
+	/* WaFbcOnly1to1Ratio:ctg */
+	if (IS_G4X(i915))
+		return 1;
+
+	/*
+	 * FBC2 can only do 1:1, 1:2, 1:4, we limit
+	 * FBC1 to the same out of convenience.
+	 */
+	return 4;
+}
+
+static int find_compression_limit(struct intel_fbc *fbc,
+				  unsigned int size, int min_limit)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+	u64 end = intel_fbc_stolen_end(i915);
+	int ret, limit = min_limit;
+
+	size /= limit;
+
+	/* Try to over-allocate to reduce reallocations and fragmentation. */
+	ret = i915_gem_stolen_insert_node_in_range(i915, &fbc->compressed_fb,
+						   size <<= 1, 4096, 0, end);
+	if (ret == 0)
+		return limit;
+
+	for (; limit <= intel_fbc_max_limit(i915); limit <<= 1) {
+		ret = i915_gem_stolen_insert_node_in_range(i915, &fbc->compressed_fb,
+							   size >>= 1, 4096, 0, end);
+		if (ret == 0)
+			return limit;
+	}
+
+	return 0;
+}
+
+static int intel_fbc_alloc_cfb(struct intel_fbc *fbc,
+			       unsigned int size, int min_limit)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+	int ret;
+
+	drm_WARN_ON(&i915->drm,
+		    drm_mm_node_allocated(&fbc->compressed_fb));
+	drm_WARN_ON(&i915->drm,
+		    drm_mm_node_allocated(&fbc->compressed_llb));
+
+	if (DISPLAY_VER(i915) < 5 && !IS_G4X(i915)) {
+		ret = i915_gem_stolen_insert_node(i915, &fbc->compressed_llb,
+						  4096, 4096);
+		if (ret)
+			goto err;
+	}
+
+	ret = find_compression_limit(fbc, size, min_limit);
+	if (!ret)
+		goto err_llb;
+	else if (ret > min_limit)
+		drm_info_once(&i915->drm,
+			      "Reducing the compressed framebuffer size. This may lead to less power savings than a non-reduced-size. Try to increase stolen memory size if available in BIOS.\n");
+
+	fbc->limit = ret;
+
+	drm_dbg_kms(&i915->drm,
+		    "reserved %llu bytes of contiguous stolen space for FBC, limit: %d\n",
+		    fbc->compressed_fb.size, fbc->limit);
+
+	return 0;
+
+err_llb:
+	if (drm_mm_node_allocated(&fbc->compressed_llb))
+		i915_gem_stolen_remove_node(i915, &fbc->compressed_llb);
+err:
+	if (drm_mm_initialized(&i915->mm.stolen))
+		drm_info_once(&i915->drm, "not enough stolen space for compressed buffer (need %d more bytes), disabling. Hint: you may be able to increase stolen memory size in the BIOS to avoid this.\n", size);
+	return -ENOSPC;
+}
+
+static void intel_fbc_program_cfb(struct intel_fbc *fbc)
+{
+	fbc->funcs->program_cfb(fbc);
+}
+
+static void intel_fbc_program_workarounds(struct intel_fbc *fbc)
+{
+	/* Wa_22014263786:icl,jsl,tgl,dg1,rkl,adls,adlp */
+	if (DISPLAY_VER(fbc->i915) >= 11 && !IS_DG2(fbc->i915))
+		intel_de_rmw(fbc->i915, ILK_DPFC_CHICKEN(fbc->id), 0,
+			     DPFC_CHICKEN_FORCE_SLB_INVALIDATION);
+}
+
+static void __intel_fbc_cleanup_cfb(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+
+	if (WARN_ON(intel_fbc_hw_is_active(fbc)))
+		return;
+
+	if (drm_mm_node_allocated(&fbc->compressed_llb))
+		i915_gem_stolen_remove_node(i915, &fbc->compressed_llb);
+	if (drm_mm_node_allocated(&fbc->compressed_fb))
+		i915_gem_stolen_remove_node(i915, &fbc->compressed_fb);
+}
+
+void intel_fbc_cleanup(struct drm_i915_private *i915)
+{
+	struct intel_fbc *fbc;
+	enum intel_fbc_id fbc_id;
+
+	for_each_intel_fbc(i915, fbc, fbc_id) {
+		mutex_lock(&fbc->lock);
+		__intel_fbc_cleanup_cfb(fbc);
+		mutex_unlock(&fbc->lock);
+
+		kfree(fbc);
+	}
+}
+
+static bool stride_is_valid(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int stride = intel_fbc_plane_stride(plane_state) *
+		fb->format->cpp[0];
+
+	/* This should have been caught earlier. */
+	if (drm_WARN_ON_ONCE(&i915->drm, (stride & (64 - 1)) != 0))
+		return false;
+
+	/* Below are the additional FBC restrictions. */
+	if (stride < 512)
+		return false;
+
+	if (DISPLAY_VER(i915) == 2 || DISPLAY_VER(i915) == 3)
+		return stride == 4096 || stride == 8192;
+
+	if (DISPLAY_VER(i915) == 4 && !IS_G4X(i915) && stride < 2048)
+		return false;
+
+	/* Display WA #1105: skl,bxt,kbl,cfl,glk */
+	if ((DISPLAY_VER(i915) == 9 || IS_GEMINILAKE(i915)) &&
+	    fb->modifier == DRM_FORMAT_MOD_LINEAR && stride & 511)
+		return false;
+
+	if (stride > 16384)
+		return false;
+
+	return true;
+}
+
+static bool pixel_format_is_valid(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+
+	switch (fb->format->format) {
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_XBGR8888:
+		return true;
+	case DRM_FORMAT_XRGB1555:
+	case DRM_FORMAT_RGB565:
+		/* 16bpp not supported on gen2 */
+		if (DISPLAY_VER(i915) == 2)
+			return false;
+		/* WaFbcOnly1to1Ratio:ctg */
+		if (IS_G4X(i915))
+			return false;
+		return true;
+	default:
+		return false;
+	}
+}
+
+static bool rotation_is_valid(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+
+	if (DISPLAY_VER(i915) >= 9 && fb->format->format == DRM_FORMAT_RGB565 &&
+	    drm_rotation_90_or_270(rotation))
+		return false;
+	else if (DISPLAY_VER(i915) <= 4 && !IS_G4X(i915) &&
+		 rotation != DRM_MODE_ROTATE_0)
+		return false;
+
+	return true;
+}
+
+/*
+ * For some reason, the hardware tracking starts looking at whatever we
+ * programmed as the display plane base address register. It does not look at
+ * the X and Y offset registers. That's why we include the src x/y offsets
+ * instead of just looking at the plane size.
+ */
+static bool intel_fbc_hw_tracking_covers_screen(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	unsigned int effective_w, effective_h, max_w, max_h;
+
+	if (DISPLAY_VER(i915) >= 10) {
+		max_w = 5120;
+		max_h = 4096;
+	} else if (DISPLAY_VER(i915) >= 8 || IS_HASWELL(i915)) {
+		max_w = 4096;
+		max_h = 4096;
+	} else if (IS_G4X(i915) || DISPLAY_VER(i915) >= 5) {
+		max_w = 4096;
+		max_h = 2048;
+	} else {
+		max_w = 2048;
+		max_h = 1536;
+	}
+
+	effective_w = plane_state->view.color_plane[0].x +
+		(drm_rect_width(&plane_state->uapi.src) >> 16);
+	effective_h = plane_state->view.color_plane[0].y +
+		(drm_rect_height(&plane_state->uapi.src) >> 16);
+
+	return effective_w <= max_w && effective_h <= max_h;
+}
+
+static bool tiling_is_valid(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+
+	switch (fb->modifier) {
+	case DRM_FORMAT_MOD_LINEAR:
+	case I915_FORMAT_MOD_Y_TILED:
+	case I915_FORMAT_MOD_Yf_TILED:
+		return DISPLAY_VER(i915) >= 9;
+	case I915_FORMAT_MOD_4_TILED:
+	case I915_FORMAT_MOD_X_TILED:
+		return true;
+	default:
+		return false;
+	}
+}
+
+static void intel_fbc_update_state(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc,
+				   struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_plane_state *plane_state =
+		intel_atomic_get_new_plane_state(state, plane);
+	struct intel_fbc *fbc = plane->fbc;
+	struct intel_fbc_state *fbc_state = &fbc->state;
+
+	WARN_ON(plane_state->no_fbc_reason);
+	WARN_ON(fbc_state->plane && fbc_state->plane != plane);
+
+	fbc_state->plane = plane;
+
+	/* FBC1 compression interval: arbitrary choice of 1 second */
+	fbc_state->interval = drm_mode_vrefresh(&crtc_state->hw.adjusted_mode);
+
+	fbc_state->fence_y_offset = intel_plane_fence_y_offset(plane_state);
+
+	drm_WARN_ON(&i915->drm, plane_state->flags & PLANE_HAS_FENCE &&
+		    !plane_state->ggtt_vma->fence);
+
+	if (plane_state->flags & PLANE_HAS_FENCE &&
+	    plane_state->ggtt_vma->fence)
+		fbc_state->fence_id = plane_state->ggtt_vma->fence->id;
+	else
+		fbc_state->fence_id = -1;
+
+	fbc_state->cfb_stride = intel_fbc_cfb_stride(plane_state);
+	fbc_state->cfb_size = intel_fbc_cfb_size(plane_state);
+	fbc_state->override_cfb_stride = intel_fbc_override_cfb_stride(plane_state);
+}
+
+static bool intel_fbc_is_fence_ok(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+
+	/* The use of a CPU fence is one of two ways to detect writes by the
+	 * CPU to the scanout and trigger updates to the FBC.
+	 *
+	 * The other method is by software tracking (see
+	 * intel_fbc_invalidate/flush()), it will manually notify FBC and nuke
+	 * the current compressed buffer and recompress it.
+	 *
+	 * Note that is possible for a tiled surface to be unmappable (and
+	 * so have no fence associated with it) due to aperture constraints
+	 * at the time of pinning.
+	 *
+	 * FIXME with 90/270 degree rotation we should use the fence on
+	 * the normal GTT view (the rotated view doesn't even have a
+	 * fence). Would need changes to the FBC fence Y offset as well.
+	 * For now this will effectively disable FBC with 90/270 degree
+	 * rotation.
+	 */
+	return DISPLAY_VER(i915) >= 9 ||
+		(plane_state->flags & PLANE_HAS_FENCE &&
+		 plane_state->ggtt_vma->fence);
+}
+
+static bool intel_fbc_is_cfb_ok(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct intel_fbc *fbc = plane->fbc;
+
+	return intel_fbc_min_limit(plane_state) <= fbc->limit &&
+		intel_fbc_cfb_size(plane_state) <= fbc->compressed_fb.size * fbc->limit;
+}
+
+static bool intel_fbc_is_ok(const struct intel_plane_state *plane_state)
+{
+	return !plane_state->no_fbc_reason &&
+		intel_fbc_is_fence_ok(plane_state) &&
+		intel_fbc_is_cfb_ok(plane_state);
+}
+
+static int intel_fbc_check_plane(struct intel_atomic_state *state,
+				 struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_plane_state *plane_state =
+		intel_atomic_get_new_plane_state(state, plane);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	struct intel_crtc *crtc = to_intel_crtc(plane_state->hw.crtc);
+	const struct intel_crtc_state *crtc_state;
+	struct intel_fbc *fbc = plane->fbc;
+
+	if (!fbc)
+		return 0;
+
+	if (intel_vgpu_active(i915)) {
+		plane_state->no_fbc_reason = "VGPU active";
+		return 0;
+	}
+
+	if (!i915->params.enable_fbc) {
+		plane_state->no_fbc_reason = "disabled per module param or by default";
+		return 0;
+	}
+
+	if (!plane_state->uapi.visible) {
+		plane_state->no_fbc_reason = "plane not visible";
+		return 0;
+	}
+
+	crtc_state = intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (crtc_state->hw.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE) {
+		plane_state->no_fbc_reason = "interlaced mode not supported";
+		return 0;
+	}
+
+	if (crtc_state->double_wide) {
+		plane_state->no_fbc_reason = "double wide pipe not supported";
+		return 0;
+	}
+
+	/*
+	 * Display 12+ is not supporting FBC with PSR2.
+	 * Recommendation is to keep this combination disabled
+	 * Bspec: 50422 HSD: 14010260002
+	 */
+	if (DISPLAY_VER(i915) >= 12 && crtc_state->has_psr2) {
+		plane_state->no_fbc_reason = "PSR2 enabled";
+		return 0;
+	}
+
+	/* Wa_14016291713 */
+	if (IS_DISPLAY_VER(i915, 12, 13) && crtc_state->has_psr) {
+		plane_state->no_fbc_reason = "PSR1 enabled (Wa_14016291713)";
+		return 0;
+	}
+
+	if (!pixel_format_is_valid(plane_state)) {
+		plane_state->no_fbc_reason = "pixel format not supported";
+		return 0;
+	}
+
+	if (!tiling_is_valid(plane_state)) {
+		plane_state->no_fbc_reason = "tiling not supported";
+		return 0;
+	}
+
+	if (!rotation_is_valid(plane_state)) {
+		plane_state->no_fbc_reason = "rotation not supported";
+		return 0;
+	}
+
+	if (!stride_is_valid(plane_state)) {
+		plane_state->no_fbc_reason = "stride not supported";
+		return 0;
+	}
+
+	if (plane_state->hw.pixel_blend_mode != DRM_MODE_BLEND_PIXEL_NONE &&
+	    fb->format->has_alpha) {
+		plane_state->no_fbc_reason = "per-pixel alpha not supported";
+		return 0;
+	}
+
+	if (!intel_fbc_hw_tracking_covers_screen(plane_state)) {
+		plane_state->no_fbc_reason = "plane size too big";
+		return 0;
+	}
+
+	/*
+	 * Work around a problem on GEN9+ HW, where enabling FBC on a plane
+	 * having a Y offset that isn't divisible by 4 causes FIFO underrun
+	 * and screen flicker.
+	 */
+	if (DISPLAY_VER(i915) >= 9 &&
+	    plane_state->view.color_plane[0].y & 3) {
+		plane_state->no_fbc_reason = "plane start Y offset misaligned";
+		return 0;
+	}
+
+	/* Wa_22010751166: icl, ehl, tgl, dg1, rkl */
+	if (DISPLAY_VER(i915) >= 11 &&
+	    (plane_state->view.color_plane[0].y +
+	     (drm_rect_height(&plane_state->uapi.src) >> 16)) & 3) {
+		plane_state->no_fbc_reason = "plane end Y offset misaligned";
+		return 0;
+	}
+
+	/* WaFbcExceedCdClockThreshold:hsw,bdw */
+	if (IS_HASWELL(i915) || IS_BROADWELL(i915)) {
+		const struct intel_cdclk_state *cdclk_state;
+
+		cdclk_state = intel_atomic_get_cdclk_state(state);
+		if (IS_ERR(cdclk_state))
+			return PTR_ERR(cdclk_state);
+
+		if (crtc_state->pixel_rate >= cdclk_state->logical.cdclk * 95 / 100) {
+			plane_state->no_fbc_reason = "pixel rate too high";
+			return 0;
+		}
+	}
+
+	plane_state->no_fbc_reason = NULL;
+
+	return 0;
+}
+
+
+static bool intel_fbc_can_flip_nuke(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc,
+				    struct intel_plane *plane)
+{
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_plane_state *old_plane_state =
+		intel_atomic_get_old_plane_state(state, plane);
+	const struct intel_plane_state *new_plane_state =
+		intel_atomic_get_new_plane_state(state, plane);
+	const struct drm_framebuffer *old_fb = old_plane_state->hw.fb;
+	const struct drm_framebuffer *new_fb = new_plane_state->hw.fb;
+
+	if (drm_atomic_crtc_needs_modeset(&new_crtc_state->uapi))
+		return false;
+
+	if (!intel_fbc_is_ok(old_plane_state) ||
+	    !intel_fbc_is_ok(new_plane_state))
+		return false;
+
+	if (old_fb->format->format != new_fb->format->format)
+		return false;
+
+	if (old_fb->modifier != new_fb->modifier)
+		return false;
+
+	if (intel_fbc_plane_stride(old_plane_state) !=
+	    intel_fbc_plane_stride(new_plane_state))
+		return false;
+
+	if (intel_fbc_cfb_stride(old_plane_state) !=
+	    intel_fbc_cfb_stride(new_plane_state))
+		return false;
+
+	if (intel_fbc_cfb_size(old_plane_state) !=
+	    intel_fbc_cfb_size(new_plane_state))
+		return false;
+
+	if (intel_fbc_override_cfb_stride(old_plane_state) !=
+	    intel_fbc_override_cfb_stride(new_plane_state))
+		return false;
+
+	return true;
+}
+
+static bool __intel_fbc_pre_update(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc,
+				   struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_fbc *fbc = plane->fbc;
+	bool need_vblank_wait = false;
+
+	fbc->flip_pending = true;
+
+	if (intel_fbc_can_flip_nuke(state, crtc, plane))
+		return need_vblank_wait;
+
+	intel_fbc_deactivate(fbc, "update pending");
+
+	/*
+	 * Display WA #1198: glk+
+	 * Need an extra vblank wait between FBC disable and most plane
+	 * updates. Bspec says this is only needed for plane disable, but
+	 * that is not true. Touching most plane registers will cause the
+	 * corruption to appear. Also SKL/derivatives do not seem to be
+	 * affected.
+	 *
+	 * TODO: could optimize this a bit by sampling the frame
+	 * counter when we disable FBC (if it was already done earlier)
+	 * and skipping the extra vblank wait before the plane update
+	 * if at least one frame has already passed.
+	 */
+	if (fbc->activated && DISPLAY_VER(i915) >= 10)
+		need_vblank_wait = true;
+	fbc->activated = false;
+
+	return need_vblank_wait;
+}
+
+bool intel_fbc_pre_update(struct intel_atomic_state *state,
+			  struct intel_crtc *crtc)
+{
+	const struct intel_plane_state *plane_state;
+	bool need_vblank_wait = false;
+	struct intel_plane *plane;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		struct intel_fbc *fbc = plane->fbc;
+
+		if (!fbc || plane->pipe != crtc->pipe)
+			continue;
+
+		mutex_lock(&fbc->lock);
+
+		if (fbc->state.plane == plane)
+			need_vblank_wait |= __intel_fbc_pre_update(state, crtc, plane);
+
+		mutex_unlock(&fbc->lock);
+	}
+
+	return need_vblank_wait;
+}
+
+static void __intel_fbc_disable(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+	struct intel_plane *plane = fbc->state.plane;
+
+	drm_WARN_ON(&i915->drm, !mutex_is_locked(&fbc->lock));
+	drm_WARN_ON(&i915->drm, fbc->active);
+
+	drm_dbg_kms(&i915->drm, "Disabling FBC on [PLANE:%d:%s]\n",
+		    plane->base.base.id, plane->base.name);
+
+	__intel_fbc_cleanup_cfb(fbc);
+
+	fbc->state.plane = NULL;
+	fbc->flip_pending = false;
+	fbc->busy_bits = 0;
+}
+
+static void __intel_fbc_post_update(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+
+	drm_WARN_ON(&i915->drm, !mutex_is_locked(&fbc->lock));
+
+	if (!fbc->busy_bits)
+		intel_fbc_activate(fbc);
+	else
+		intel_fbc_deactivate(fbc, "frontbuffer write");
+}
+
+void intel_fbc_post_update(struct intel_atomic_state *state,
+			   struct intel_crtc *crtc)
+{
+	const struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		struct intel_fbc *fbc = plane->fbc;
+
+		if (!fbc || plane->pipe != crtc->pipe)
+			continue;
+
+		mutex_lock(&fbc->lock);
+
+		if (fbc->state.plane == plane) {
+			fbc->flip_pending = false;
+			__intel_fbc_post_update(fbc);
+		}
+
+		mutex_unlock(&fbc->lock);
+	}
+}
+
+static unsigned int intel_fbc_get_frontbuffer_bit(struct intel_fbc *fbc)
+{
+	if (fbc->state.plane)
+		return fbc->state.plane->frontbuffer_bit;
+	else
+		return 0;
+}
+
+static void __intel_fbc_invalidate(struct intel_fbc *fbc,
+				   unsigned int frontbuffer_bits,
+				   enum fb_op_origin origin)
+{
+	if (origin == ORIGIN_FLIP || origin == ORIGIN_CURSOR_UPDATE)
+		return;
+
+	mutex_lock(&fbc->lock);
+
+	frontbuffer_bits &= intel_fbc_get_frontbuffer_bit(fbc);
+	if (!frontbuffer_bits)
+		goto out;
+
+	fbc->busy_bits |= frontbuffer_bits;
+	intel_fbc_deactivate(fbc, "frontbuffer write");
+
+out:
+	mutex_unlock(&fbc->lock);
+}
+
+void intel_fbc_invalidate(struct drm_i915_private *i915,
+			  unsigned int frontbuffer_bits,
+			  enum fb_op_origin origin)
+{
+	struct intel_fbc *fbc;
+	enum intel_fbc_id fbc_id;
+
+	for_each_intel_fbc(i915, fbc, fbc_id)
+		__intel_fbc_invalidate(fbc, frontbuffer_bits, origin);
+
+}
+
+static void __intel_fbc_flush(struct intel_fbc *fbc,
+			      unsigned int frontbuffer_bits,
+			      enum fb_op_origin origin)
+{
+	mutex_lock(&fbc->lock);
+
+	frontbuffer_bits &= intel_fbc_get_frontbuffer_bit(fbc);
+	if (!frontbuffer_bits)
+		goto out;
+
+	fbc->busy_bits &= ~frontbuffer_bits;
+
+	if (origin == ORIGIN_FLIP || origin == ORIGIN_CURSOR_UPDATE)
+		goto out;
+
+	if (fbc->busy_bits || fbc->flip_pending)
+		goto out;
+
+	if (fbc->active)
+		intel_fbc_nuke(fbc);
+	else
+		intel_fbc_activate(fbc);
+
+out:
+	mutex_unlock(&fbc->lock);
+}
+
+void intel_fbc_flush(struct drm_i915_private *i915,
+		     unsigned int frontbuffer_bits,
+		     enum fb_op_origin origin)
+{
+	struct intel_fbc *fbc;
+	enum intel_fbc_id fbc_id;
+
+	for_each_intel_fbc(i915, fbc, fbc_id)
+		__intel_fbc_flush(fbc, frontbuffer_bits, origin);
+}
+
+int intel_fbc_atomic_check(struct intel_atomic_state *state)
+{
+	struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		int ret;
+
+		ret = intel_fbc_check_plane(state, plane);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static void __intel_fbc_enable(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc,
+			       struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_plane_state *plane_state =
+		intel_atomic_get_new_plane_state(state, plane);
+	struct intel_fbc *fbc = plane->fbc;
+
+	if (fbc->state.plane) {
+		if (fbc->state.plane != plane)
+			return;
+
+		if (intel_fbc_is_ok(plane_state)) {
+			intel_fbc_update_state(state, crtc, plane);
+			return;
+		}
+
+		__intel_fbc_disable(fbc);
+	}
+
+	drm_WARN_ON(&i915->drm, fbc->active);
+
+	fbc->no_fbc_reason = plane_state->no_fbc_reason;
+	if (fbc->no_fbc_reason)
+		return;
+
+	if (!intel_fbc_is_fence_ok(plane_state)) {
+		fbc->no_fbc_reason = "framebuffer not fenced";
+		return;
+	}
+
+	if (fbc->underrun_detected) {
+		fbc->no_fbc_reason = "FIFO underrun";
+		return;
+	}
+
+	if (intel_fbc_alloc_cfb(fbc, intel_fbc_cfb_size(plane_state),
+				intel_fbc_min_limit(plane_state))) {
+		fbc->no_fbc_reason = "not enough stolen memory";
+		return;
+	}
+
+	drm_dbg_kms(&i915->drm, "Enabling FBC on [PLANE:%d:%s]\n",
+		    plane->base.base.id, plane->base.name);
+	fbc->no_fbc_reason = "FBC enabled but not active yet\n";
+
+	intel_fbc_update_state(state, crtc, plane);
+
+	intel_fbc_program_workarounds(fbc);
+	intel_fbc_program_cfb(fbc);
+}
+
+/**
+ * intel_fbc_disable - disable FBC if it's associated with crtc
+ * @crtc: the CRTC
+ *
+ * This function disables FBC if it's associated with the provided CRTC.
+ */
+void intel_fbc_disable(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_plane *plane;
+
+	for_each_intel_plane(&i915->drm, plane) {
+		struct intel_fbc *fbc = plane->fbc;
+
+		if (!fbc || plane->pipe != crtc->pipe)
+			continue;
+
+		mutex_lock(&fbc->lock);
+		if (fbc->state.plane == plane)
+			__intel_fbc_disable(fbc);
+		mutex_unlock(&fbc->lock);
+	}
+}
+
+void intel_fbc_update(struct intel_atomic_state *state,
+		      struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	int i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		struct intel_fbc *fbc = plane->fbc;
+
+		if (!fbc || plane->pipe != crtc->pipe)
+			continue;
+
+		mutex_lock(&fbc->lock);
+
+		if (crtc_state->update_pipe && plane_state->no_fbc_reason) {
+			if (fbc->state.plane == plane)
+				__intel_fbc_disable(fbc);
+		} else {
+			__intel_fbc_enable(state, crtc, plane);
+		}
+
+		mutex_unlock(&fbc->lock);
+	}
+}
+
+static void intel_fbc_underrun_work_fn(struct work_struct *work)
+{
+	struct intel_fbc *fbc = container_of(work, typeof(*fbc), underrun_work);
+	struct drm_i915_private *i915 = fbc->i915;
+
+	mutex_lock(&fbc->lock);
+
+	/* Maybe we were scheduled twice. */
+	if (fbc->underrun_detected || !fbc->state.plane)
+		goto out;
+
+	drm_dbg_kms(&i915->drm, "Disabling FBC due to FIFO underrun.\n");
+	fbc->underrun_detected = true;
+
+	intel_fbc_deactivate(fbc, "FIFO underrun");
+	if (!fbc->flip_pending)
+		intel_crtc_wait_for_next_vblank(intel_crtc_for_pipe(i915, fbc->state.plane->pipe));
+	__intel_fbc_disable(fbc);
+out:
+	mutex_unlock(&fbc->lock);
+}
+
+static void __intel_fbc_reset_underrun(struct intel_fbc *fbc)
+{
+	struct drm_i915_private *i915 = fbc->i915;
+
+	cancel_work_sync(&fbc->underrun_work);
+
+	mutex_lock(&fbc->lock);
+
+	if (fbc->underrun_detected) {
+		drm_dbg_kms(&i915->drm,
+			    "Re-allowing FBC after fifo underrun\n");
+		fbc->no_fbc_reason = "FIFO underrun cleared";
+	}
+
+	fbc->underrun_detected = false;
+	mutex_unlock(&fbc->lock);
+}
+
+/*
+ * intel_fbc_reset_underrun - reset FBC fifo underrun status.
+ * @i915: the i915 device
+ *
+ * See intel_fbc_handle_fifo_underrun_irq(). For automated testing we
+ * want to re-enable FBC after an underrun to increase test coverage.
+ */
+void intel_fbc_reset_underrun(struct drm_i915_private *i915)
+{
+	struct intel_fbc *fbc;
+	enum intel_fbc_id fbc_id;
+
+	for_each_intel_fbc(i915, fbc, fbc_id)
+		__intel_fbc_reset_underrun(fbc);
+}
+
+static void __intel_fbc_handle_fifo_underrun_irq(struct intel_fbc *fbc)
+{
+	/*
+	 * There's no guarantee that underrun_detected won't be set to true
+	 * right after this check and before the work is scheduled, but that's
+	 * not a problem since we'll check it again under the work function
+	 * while FBC is locked. This check here is just to prevent us from
+	 * unnecessarily scheduling the work, and it relies on the fact that we
+	 * never switch underrun_detect back to false after it's true.
+	 */
+	if (READ_ONCE(fbc->underrun_detected))
+		return;
+
+	schedule_work(&fbc->underrun_work);
+}
+
+/**
+ * intel_fbc_handle_fifo_underrun_irq - disable FBC when we get a FIFO underrun
+ * @i915: i915 device
+ *
+ * Without FBC, most underruns are harmless and don't really cause too many
+ * problems, except for an annoying message on dmesg. With FBC, underruns can
+ * become black screens or even worse, especially when paired with bad
+ * watermarks. So in order for us to be on the safe side, completely disable FBC
+ * in case we ever detect a FIFO underrun on any pipe. An underrun on any pipe
+ * already suggests that watermarks may be bad, so try to be as safe as
+ * possible.
+ *
+ * This function is called from the IRQ handler.
+ */
+void intel_fbc_handle_fifo_underrun_irq(struct drm_i915_private *i915)
+{
+	struct intel_fbc *fbc;
+	enum intel_fbc_id fbc_id;
+
+	for_each_intel_fbc(i915, fbc, fbc_id)
+		__intel_fbc_handle_fifo_underrun_irq(fbc);
+}
+
+/*
+ * The DDX driver changes its behavior depending on the value it reads from
+ * i915.enable_fbc, so sanitize it by translating the default value into either
+ * 0 or 1 in order to allow it to know what's going on.
+ *
+ * Notice that this is done at driver initialization and we still allow user
+ * space to change the value during runtime without sanitizing it again. IGT
+ * relies on being able to change i915.enable_fbc at runtime.
+ */
+static int intel_sanitize_fbc_option(struct drm_i915_private *i915)
+{
+	if (i915->params.enable_fbc >= 0)
+		return !!i915->params.enable_fbc;
+
+	if (!HAS_FBC(i915))
+		return 0;
+
+	if (IS_BROADWELL(i915) || DISPLAY_VER(i915) >= 9)
+		return 1;
+
+	return 0;
+}
+
+static bool need_fbc_vtd_wa(struct drm_i915_private *i915)
+{
+	/* WaFbcTurnOffFbcWhenHyperVisorIsUsed:skl,bxt */
+	if (i915_vtd_active(i915) &&
+	    (IS_SKYLAKE(i915) || IS_BROXTON(i915))) {
+		drm_info(&i915->drm,
+			 "Disabling framebuffer compression (FBC) to prevent screen flicker with VT-d enabled\n");
+		return true;
+	}
+
+	return false;
+}
+
+void intel_fbc_add_plane(struct intel_fbc *fbc, struct intel_plane *plane)
+{
+	plane->fbc = fbc;
+}
+
+static struct intel_fbc *intel_fbc_create(struct drm_i915_private *i915,
+					  enum intel_fbc_id fbc_id)
+{
+	struct intel_fbc *fbc;
+
+	fbc = kzalloc(sizeof(*fbc), GFP_KERNEL);
+	if (!fbc)
+		return NULL;
+
+	fbc->id = fbc_id;
+	fbc->i915 = i915;
+	INIT_WORK(&fbc->underrun_work, intel_fbc_underrun_work_fn);
+	mutex_init(&fbc->lock);
+
+	if (DISPLAY_VER(i915) >= 7)
+		fbc->funcs = &ivb_fbc_funcs;
+	else if (DISPLAY_VER(i915) == 6)
+		fbc->funcs = &snb_fbc_funcs;
+	else if (DISPLAY_VER(i915) == 5)
+		fbc->funcs = &ilk_fbc_funcs;
+	else if (IS_G4X(i915))
+		fbc->funcs = &g4x_fbc_funcs;
+	else if (DISPLAY_VER(i915) == 4)
+		fbc->funcs = &i965_fbc_funcs;
+	else
+		fbc->funcs = &i8xx_fbc_funcs;
+
+	return fbc;
+}
+
+/**
+ * intel_fbc_init - Initialize FBC
+ * @i915: the i915 device
+ *
+ * This function might be called during PM init process.
+ */
+void intel_fbc_init(struct drm_i915_private *i915)
+{
+	enum intel_fbc_id fbc_id;
+
+	if (!drm_mm_initialized(&i915->mm.stolen))
+		RUNTIME_INFO(i915)->fbc_mask = 0;
+
+	if (need_fbc_vtd_wa(i915))
+		RUNTIME_INFO(i915)->fbc_mask = 0;
+
+	i915->params.enable_fbc = intel_sanitize_fbc_option(i915);
+	drm_dbg_kms(&i915->drm, "Sanitized enable_fbc value: %d\n",
+		    i915->params.enable_fbc);
+
+	for_each_fbc_id(i915, fbc_id)
+		i915->display.fbc[fbc_id] = intel_fbc_create(i915, fbc_id);
+}
+
+/**
+ * intel_fbc_sanitize - Sanitize FBC
+ * @i915: the i915 device
+ *
+ * Make sure FBC is initially disabled since we have no
+ * idea eg. into which parts of stolen it might be scribbling
+ * into.
+ */
+void intel_fbc_sanitize(struct drm_i915_private *i915)
+{
+	struct intel_fbc *fbc;
+	enum intel_fbc_id fbc_id;
+
+	for_each_intel_fbc(i915, fbc, fbc_id) {
+		if (intel_fbc_hw_is_active(fbc))
+			intel_fbc_hw_deactivate(fbc);
+	}
+}
+
+static int intel_fbc_debugfs_status_show(struct seq_file *m, void *unused)
+{
+	struct intel_fbc *fbc = m->private;
+	struct drm_i915_private *i915 = fbc->i915;
+	struct intel_plane *plane;
+	intel_wakeref_t wakeref;
+
+	drm_modeset_lock_all(&i915->drm);
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+	mutex_lock(&fbc->lock);
+
+	if (fbc->active) {
+		seq_puts(m, "FBC enabled\n");
+		seq_printf(m, "Compressing: %s\n",
+			   str_yes_no(intel_fbc_is_compressing(fbc)));
+	} else {
+		seq_printf(m, "FBC disabled: %s\n", fbc->no_fbc_reason);
+	}
+
+	for_each_intel_plane(&i915->drm, plane) {
+		const struct intel_plane_state *plane_state =
+			to_intel_plane_state(plane->base.state);
+
+		if (plane->fbc != fbc)
+			continue;
+
+		seq_printf(m, "%c [PLANE:%d:%s]: %s\n",
+			   fbc->state.plane == plane ? '*' : ' ',
+			   plane->base.base.id, plane->base.name,
+			   plane_state->no_fbc_reason ?: "FBC possible");
+	}
+
+	mutex_unlock(&fbc->lock);
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+
+	drm_modeset_unlock_all(&i915->drm);
+
+	return 0;
+}
+
+DEFINE_SHOW_ATTRIBUTE(intel_fbc_debugfs_status);
+
+static int intel_fbc_debugfs_false_color_get(void *data, u64 *val)
+{
+	struct intel_fbc *fbc = data;
+
+	*val = fbc->false_color;
+
+	return 0;
+}
+
+static int intel_fbc_debugfs_false_color_set(void *data, u64 val)
+{
+	struct intel_fbc *fbc = data;
+
+	mutex_lock(&fbc->lock);
+
+	fbc->false_color = val;
+
+	if (fbc->active)
+		fbc->funcs->set_false_color(fbc, fbc->false_color);
+
+	mutex_unlock(&fbc->lock);
+
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(intel_fbc_debugfs_false_color_fops,
+			intel_fbc_debugfs_false_color_get,
+			intel_fbc_debugfs_false_color_set,
+			"%llu\n");
+
+static void intel_fbc_debugfs_add(struct intel_fbc *fbc,
+				  struct dentry *parent)
+{
+	debugfs_create_file("i915_fbc_status", 0444, parent,
+			    fbc, &intel_fbc_debugfs_status_fops);
+
+	if (fbc->funcs->set_false_color)
+		debugfs_create_file("i915_fbc_false_color", 0644, parent,
+				    fbc, &intel_fbc_debugfs_false_color_fops);
+}
+
+void intel_fbc_crtc_debugfs_add(struct intel_crtc *crtc)
+{
+	struct intel_plane *plane = to_intel_plane(crtc->base.primary);
+
+	if (plane->fbc)
+		intel_fbc_debugfs_add(plane->fbc, crtc->base.debugfs_entry);
+}
+
+/* FIXME: remove this once igt is on board with per-crtc stuff */
+void intel_fbc_debugfs_register(struct drm_i915_private *i915)
+{
+	struct drm_minor *minor = i915->drm.primary;
+	struct intel_fbc *fbc;
+
+	fbc = i915->display.fbc[INTEL_FBC_A];
+	if (fbc)
+		intel_fbc_debugfs_add(fbc, minor->debugfs_root);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_fbc.h b/drivers/gpu/drm/i915/display/intel_fbc.h
new file mode 100644
index 000000000..4adb98afe
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fbc.h
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_FBC_H__
+#define __INTEL_FBC_H__
+
+#include <linux/types.h>
+
+enum fb_op_origin;
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_fbc;
+struct intel_plane;
+struct intel_plane_state;
+
+enum intel_fbc_id {
+	INTEL_FBC_A,
+	INTEL_FBC_B,
+
+	I915_MAX_FBCS,
+};
+
+int intel_fbc_atomic_check(struct intel_atomic_state *state);
+bool intel_fbc_pre_update(struct intel_atomic_state *state,
+			  struct intel_crtc *crtc);
+void intel_fbc_post_update(struct intel_atomic_state *state,
+			   struct intel_crtc *crtc);
+void intel_fbc_init(struct drm_i915_private *dev_priv);
+void intel_fbc_cleanup(struct drm_i915_private *dev_priv);
+void intel_fbc_sanitize(struct drm_i915_private *dev_priv);
+void intel_fbc_update(struct intel_atomic_state *state,
+		      struct intel_crtc *crtc);
+void intel_fbc_disable(struct intel_crtc *crtc);
+void intel_fbc_invalidate(struct drm_i915_private *dev_priv,
+			  unsigned int frontbuffer_bits,
+			  enum fb_op_origin origin);
+void intel_fbc_flush(struct drm_i915_private *dev_priv,
+		     unsigned int frontbuffer_bits, enum fb_op_origin origin);
+void intel_fbc_add_plane(struct intel_fbc *fbc, struct intel_plane *plane);
+void intel_fbc_handle_fifo_underrun_irq(struct drm_i915_private *i915);
+void intel_fbc_reset_underrun(struct drm_i915_private *i915);
+void intel_fbc_crtc_debugfs_add(struct intel_crtc *crtc);
+void intel_fbc_debugfs_register(struct drm_i915_private *i915);
+
+#endif /* __INTEL_FBC_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_fbdev.c b/drivers/gpu/drm/i915/display/intel_fbdev.c
new file mode 100644
index 000000000..968915000
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fbdev.c
@@ -0,0 +1,733 @@
+/*
+ * Copyright © 2007 David Airlie
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *     David Airlie
+ */
+
+#include <linux/async.h>
+#include <linux/console.h>
+#include <linux/delay.h>
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/string.h>
+#include <linux/sysrq.h>
+#include <linux/tty.h>
+#include <linux/vga_switcheroo.h>
+
+#include <drm/drm_crtc.h>
+#include <drm/drm_fb_helper.h>
+#include <drm/drm_fourcc.h>
+
+#include "gem/i915_gem_lmem.h"
+
+#include "i915_drv.h"
+#include "intel_display_types.h"
+#include "intel_fb.h"
+#include "intel_fb_pin.h"
+#include "intel_fbdev.h"
+#include "intel_frontbuffer.h"
+
+struct intel_fbdev {
+	struct drm_fb_helper helper;
+	struct intel_framebuffer *fb;
+	struct i915_vma *vma;
+	unsigned long vma_flags;
+	async_cookie_t cookie;
+	int preferred_bpp;
+
+	/* Whether or not fbdev hpd processing is temporarily suspended */
+	bool hpd_suspended: 1;
+	/* Set when a hotplug was received while HPD processing was suspended */
+	bool hpd_waiting: 1;
+
+	/* Protects hpd_suspended */
+	struct mutex hpd_lock;
+};
+
+static struct intel_frontbuffer *to_frontbuffer(struct intel_fbdev *ifbdev)
+{
+	return ifbdev->fb->frontbuffer;
+}
+
+static void intel_fbdev_invalidate(struct intel_fbdev *ifbdev)
+{
+	intel_frontbuffer_invalidate(to_frontbuffer(ifbdev), ORIGIN_CPU);
+}
+
+static int intel_fbdev_set_par(struct fb_info *info)
+{
+	struct drm_fb_helper *fb_helper = info->par;
+	struct intel_fbdev *ifbdev =
+		container_of(fb_helper, struct intel_fbdev, helper);
+	int ret;
+
+	ret = drm_fb_helper_set_par(info);
+	if (ret == 0)
+		intel_fbdev_invalidate(ifbdev);
+
+	return ret;
+}
+
+static int intel_fbdev_blank(int blank, struct fb_info *info)
+{
+	struct drm_fb_helper *fb_helper = info->par;
+	struct intel_fbdev *ifbdev =
+		container_of(fb_helper, struct intel_fbdev, helper);
+	int ret;
+
+	ret = drm_fb_helper_blank(blank, info);
+	if (ret == 0)
+		intel_fbdev_invalidate(ifbdev);
+
+	return ret;
+}
+
+static int intel_fbdev_pan_display(struct fb_var_screeninfo *var,
+				   struct fb_info *info)
+{
+	struct drm_fb_helper *fb_helper = info->par;
+	struct intel_fbdev *ifbdev =
+		container_of(fb_helper, struct intel_fbdev, helper);
+	int ret;
+
+	ret = drm_fb_helper_pan_display(var, info);
+	if (ret == 0)
+		intel_fbdev_invalidate(ifbdev);
+
+	return ret;
+}
+
+static const struct fb_ops intelfb_ops = {
+	.owner = THIS_MODULE,
+	DRM_FB_HELPER_DEFAULT_OPS,
+	.fb_set_par = intel_fbdev_set_par,
+	.fb_fillrect = drm_fb_helper_cfb_fillrect,
+	.fb_copyarea = drm_fb_helper_cfb_copyarea,
+	.fb_imageblit = drm_fb_helper_cfb_imageblit,
+	.fb_pan_display = intel_fbdev_pan_display,
+	.fb_blank = intel_fbdev_blank,
+};
+
+static int intelfb_alloc(struct drm_fb_helper *helper,
+			 struct drm_fb_helper_surface_size *sizes)
+{
+	struct intel_fbdev *ifbdev =
+		container_of(helper, struct intel_fbdev, helper);
+	struct drm_framebuffer *fb;
+	struct drm_device *dev = helper->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_mode_fb_cmd2 mode_cmd = {};
+	struct drm_i915_gem_object *obj;
+	int size;
+
+	/* we don't do packed 24bpp */
+	if (sizes->surface_bpp == 24)
+		sizes->surface_bpp = 32;
+
+	mode_cmd.width = sizes->surface_width;
+	mode_cmd.height = sizes->surface_height;
+
+	mode_cmd.pitches[0] = ALIGN(mode_cmd.width *
+				    DIV_ROUND_UP(sizes->surface_bpp, 8), 64);
+	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
+							  sizes->surface_depth);
+
+	size = mode_cmd.pitches[0] * mode_cmd.height;
+	size = PAGE_ALIGN(size);
+
+	obj = ERR_PTR(-ENODEV);
+	if (HAS_LMEM(dev_priv)) {
+		obj = i915_gem_object_create_lmem(dev_priv, size,
+						  I915_BO_ALLOC_CONTIGUOUS);
+	} else {
+		/*
+		 * If the FB is too big, just don't use it since fbdev is not very
+		 * important and we should probably use that space with FBC or other
+		 * features.
+		 */
+		if (size * 2 < dev_priv->stolen_usable_size)
+			obj = i915_gem_object_create_stolen(dev_priv, size);
+		if (IS_ERR(obj))
+			obj = i915_gem_object_create_shmem(dev_priv, size);
+	}
+
+	if (IS_ERR(obj)) {
+		drm_err(&dev_priv->drm, "failed to allocate framebuffer (%pe)\n", obj);
+		return PTR_ERR(obj);
+	}
+
+	fb = intel_framebuffer_create(obj, &mode_cmd);
+	i915_gem_object_put(obj);
+	if (IS_ERR(fb))
+		return PTR_ERR(fb);
+
+	ifbdev->fb = to_intel_framebuffer(fb);
+	return 0;
+}
+
+static int intelfb_create(struct drm_fb_helper *helper,
+			  struct drm_fb_helper_surface_size *sizes)
+{
+	struct intel_fbdev *ifbdev =
+		container_of(helper, struct intel_fbdev, helper);
+	struct intel_framebuffer *intel_fb = ifbdev->fb;
+	struct drm_device *dev = helper->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	struct i915_ggtt *ggtt = to_gt(dev_priv)->ggtt;
+	const struct i915_gtt_view view = {
+		.type = I915_GTT_VIEW_NORMAL,
+	};
+	intel_wakeref_t wakeref;
+	struct fb_info *info;
+	struct i915_vma *vma;
+	unsigned long flags = 0;
+	bool prealloc = false;
+	void __iomem *vaddr;
+	struct drm_i915_gem_object *obj;
+	struct i915_gem_ww_ctx ww;
+	int ret;
+
+	mutex_lock(&ifbdev->hpd_lock);
+	ret = ifbdev->hpd_suspended ? -EAGAIN : 0;
+	mutex_unlock(&ifbdev->hpd_lock);
+	if (ret)
+		return ret;
+
+	if (intel_fb &&
+	    (sizes->fb_width > intel_fb->base.width ||
+	     sizes->fb_height > intel_fb->base.height)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "BIOS fb too small (%dx%d), we require (%dx%d),"
+			    " releasing it\n",
+			    intel_fb->base.width, intel_fb->base.height,
+			    sizes->fb_width, sizes->fb_height);
+		drm_framebuffer_put(&intel_fb->base);
+		intel_fb = ifbdev->fb = NULL;
+	}
+	if (!intel_fb || drm_WARN_ON(dev, !intel_fb_obj(&intel_fb->base))) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "no BIOS fb, allocating a new one\n");
+		ret = intelfb_alloc(helper, sizes);
+		if (ret)
+			return ret;
+		intel_fb = ifbdev->fb;
+	} else {
+		drm_dbg_kms(&dev_priv->drm, "re-using BIOS fb\n");
+		prealloc = true;
+		sizes->fb_width = intel_fb->base.width;
+		sizes->fb_height = intel_fb->base.height;
+	}
+
+	wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+
+	/* Pin the GGTT vma for our access via info->screen_base.
+	 * This also validates that any existing fb inherited from the
+	 * BIOS is suitable for own access.
+	 */
+	vma = intel_pin_and_fence_fb_obj(&ifbdev->fb->base, false,
+					 &view, false, &flags);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto out_unlock;
+	}
+
+	info = drm_fb_helper_alloc_fbi(helper);
+	if (IS_ERR(info)) {
+		drm_err(&dev_priv->drm, "Failed to allocate fb_info (%pe)\n", info);
+		ret = PTR_ERR(info);
+		goto out_unpin;
+	}
+
+	ifbdev->helper.fb = &ifbdev->fb->base;
+
+	info->fbops = &intelfb_ops;
+
+	/* setup aperture base/size for vesafb takeover */
+	obj = intel_fb_obj(&intel_fb->base);
+	if (i915_gem_object_is_lmem(obj)) {
+		struct intel_memory_region *mem = obj->mm.region;
+
+		info->apertures->ranges[0].base = mem->io_start;
+		info->apertures->ranges[0].size = mem->io_size;
+
+		/* Use fbdev's framebuffer from lmem for discrete */
+		info->fix.smem_start =
+			(unsigned long)(mem->io_start +
+					i915_gem_object_get_dma_address(obj, 0));
+		info->fix.smem_len = obj->base.size;
+	} else {
+		info->apertures->ranges[0].base = ggtt->gmadr.start;
+		info->apertures->ranges[0].size = ggtt->mappable_end;
+
+		/* Our framebuffer is the entirety of fbdev's system memory */
+		info->fix.smem_start =
+			(unsigned long)(ggtt->gmadr.start + vma->node.start);
+		info->fix.smem_len = vma->size;
+	}
+
+	for_i915_gem_ww(&ww, ret, false) {
+		ret = i915_gem_object_lock(vma->obj, &ww);
+
+		if (ret)
+			continue;
+
+		vaddr = i915_vma_pin_iomap(vma);
+		if (IS_ERR(vaddr)) {
+			drm_err(&dev_priv->drm,
+				"Failed to remap framebuffer into virtual memory (%pe)\n", vaddr);
+			ret = PTR_ERR(vaddr);
+			continue;
+		}
+	}
+
+	if (ret)
+		goto out_unpin;
+
+	info->screen_base = vaddr;
+	info->screen_size = vma->size;
+
+	drm_fb_helper_fill_info(info, &ifbdev->helper, sizes);
+
+	/* If the object is shmemfs backed, it will have given us zeroed pages.
+	 * If the object is stolen however, it will be full of whatever
+	 * garbage was left in there.
+	 */
+	if (!i915_gem_object_is_shmem(vma->obj) && !prealloc)
+		memset_io(info->screen_base, 0, info->screen_size);
+
+	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
+
+	drm_dbg_kms(&dev_priv->drm, "allocated %dx%d fb: 0x%08x\n",
+		    ifbdev->fb->base.width, ifbdev->fb->base.height,
+		    i915_ggtt_offset(vma));
+	ifbdev->vma = vma;
+	ifbdev->vma_flags = flags;
+
+	intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+	vga_switcheroo_client_fb_set(pdev, info);
+	return 0;
+
+out_unpin:
+	intel_unpin_fb_vma(vma, flags);
+out_unlock:
+	intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+	return ret;
+}
+
+static const struct drm_fb_helper_funcs intel_fb_helper_funcs = {
+	.fb_probe = intelfb_create,
+};
+
+static void intel_fbdev_destroy(struct intel_fbdev *ifbdev)
+{
+	/* We rely on the object-free to release the VMA pinning for
+	 * the info->screen_base mmaping. Leaking the VMA is simpler than
+	 * trying to rectify all the possible error paths leading here.
+	 */
+
+	drm_fb_helper_fini(&ifbdev->helper);
+
+	if (ifbdev->vma)
+		intel_unpin_fb_vma(ifbdev->vma, ifbdev->vma_flags);
+
+	if (ifbdev->fb)
+		drm_framebuffer_remove(&ifbdev->fb->base);
+
+	kfree(ifbdev);
+}
+
+/*
+ * Build an intel_fbdev struct using a BIOS allocated framebuffer, if possible.
+ * The core display code will have read out the current plane configuration,
+ * so we use that to figure out if there's an object for us to use as the
+ * fb, and if so, we re-use it for the fbdev configuration.
+ *
+ * Note we only support a single fb shared across pipes for boot (mostly for
+ * fbcon), so we just find the biggest and use that.
+ */
+static bool intel_fbdev_init_bios(struct drm_device *dev,
+				  struct intel_fbdev *ifbdev)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct intel_framebuffer *fb = NULL;
+	struct intel_crtc *crtc;
+	unsigned int max_size = 0;
+
+	/* Find the largest fb */
+	for_each_intel_crtc(dev, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct intel_plane *plane =
+			to_intel_plane(crtc->base.primary);
+		struct intel_plane_state *plane_state =
+			to_intel_plane_state(plane->base.state);
+		struct drm_i915_gem_object *obj =
+			intel_fb_obj(plane_state->uapi.fb);
+
+		if (!crtc_state->uapi.active) {
+			drm_dbg_kms(&i915->drm,
+				    "[CRTC:%d:%s] not active, skipping\n",
+				    crtc->base.base.id, crtc->base.name);
+			continue;
+		}
+
+		if (!obj) {
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] no fb, skipping\n",
+				    plane->base.base.id, plane->base.name);
+			continue;
+		}
+
+		if (obj->base.size > max_size) {
+			drm_dbg_kms(&i915->drm,
+				    "found possible fb from [PLANE:%d:%s]\n",
+				    plane->base.base.id, plane->base.name);
+			fb = to_intel_framebuffer(plane_state->uapi.fb);
+			max_size = obj->base.size;
+		}
+	}
+
+	if (!fb) {
+		drm_dbg_kms(&i915->drm,
+			    "no active fbs found, not using BIOS config\n");
+		goto out;
+	}
+
+	/* Now make sure all the pipes will fit into it */
+	for_each_intel_crtc(dev, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct intel_plane *plane =
+			to_intel_plane(crtc->base.primary);
+		unsigned int cur_size;
+
+		if (!crtc_state->uapi.active) {
+			drm_dbg_kms(&i915->drm,
+				    "[CRTC:%d:%s] not active, skipping\n",
+				    crtc->base.base.id, crtc->base.name);
+			continue;
+		}
+
+		drm_dbg_kms(&i915->drm, "checking [PLANE:%d:%s] for BIOS fb\n",
+			    plane->base.base.id, plane->base.name);
+
+		/*
+		 * See if the plane fb we found above will fit on this
+		 * pipe.  Note we need to use the selected fb's pitch and bpp
+		 * rather than the current pipe's, since they differ.
+		 */
+		cur_size = crtc_state->uapi.adjusted_mode.crtc_hdisplay;
+		cur_size = cur_size * fb->base.format->cpp[0];
+		if (fb->base.pitches[0] < cur_size) {
+			drm_dbg_kms(&i915->drm,
+				    "fb not wide enough for [PLANE:%d:%s] (%d vs %d)\n",
+				    plane->base.base.id, plane->base.name,
+				    cur_size, fb->base.pitches[0]);
+			fb = NULL;
+			break;
+		}
+
+		cur_size = crtc_state->uapi.adjusted_mode.crtc_vdisplay;
+		cur_size = intel_fb_align_height(&fb->base, 0, cur_size);
+		cur_size *= fb->base.pitches[0];
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] area: %dx%d, bpp: %d, size: %d\n",
+			    crtc->base.base.id, crtc->base.name,
+			    crtc_state->uapi.adjusted_mode.crtc_hdisplay,
+			    crtc_state->uapi.adjusted_mode.crtc_vdisplay,
+			    fb->base.format->cpp[0] * 8,
+			    cur_size);
+
+		if (cur_size > max_size) {
+			drm_dbg_kms(&i915->drm,
+				    "fb not big enough for [PLANE:%d:%s] (%d vs %d)\n",
+				    plane->base.base.id, plane->base.name,
+				    cur_size, max_size);
+			fb = NULL;
+			break;
+		}
+
+		drm_dbg_kms(&i915->drm,
+			    "fb big enough [PLANE:%d:%s] (%d >= %d)\n",
+			    plane->base.base.id, plane->base.name,
+			    max_size, cur_size);
+	}
+
+	if (!fb) {
+		drm_dbg_kms(&i915->drm,
+			    "BIOS fb not suitable for all pipes, not using\n");
+		goto out;
+	}
+
+	ifbdev->preferred_bpp = fb->base.format->cpp[0] * 8;
+	ifbdev->fb = fb;
+
+	drm_framebuffer_get(&ifbdev->fb->base);
+
+	/* Final pass to check if any active pipes don't have fbs */
+	for_each_intel_crtc(dev, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct intel_plane *plane =
+			to_intel_plane(crtc->base.primary);
+		struct intel_plane_state *plane_state =
+			to_intel_plane_state(plane->base.state);
+
+		if (!crtc_state->uapi.active)
+			continue;
+
+		drm_WARN(dev, !plane_state->uapi.fb,
+			 "re-used BIOS config but lost an fb on [PLANE:%d:%s]\n",
+			 plane->base.base.id, plane->base.name);
+	}
+
+
+	drm_dbg_kms(&i915->drm, "using BIOS fb for initial console\n");
+	return true;
+
+out:
+
+	return false;
+}
+
+static void intel_fbdev_suspend_worker(struct work_struct *work)
+{
+	intel_fbdev_set_suspend(&container_of(work,
+					      struct drm_i915_private,
+					      display.fbdev.suspend_work)->drm,
+				FBINFO_STATE_RUNNING,
+				true);
+}
+
+int intel_fbdev_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_fbdev *ifbdev;
+	int ret;
+
+	if (drm_WARN_ON(dev, !HAS_DISPLAY(dev_priv)))
+		return -ENODEV;
+
+	ifbdev = kzalloc(sizeof(struct intel_fbdev), GFP_KERNEL);
+	if (ifbdev == NULL)
+		return -ENOMEM;
+
+	mutex_init(&ifbdev->hpd_lock);
+	drm_fb_helper_prepare(dev, &ifbdev->helper, &intel_fb_helper_funcs);
+
+	if (!intel_fbdev_init_bios(dev, ifbdev))
+		ifbdev->preferred_bpp = 32;
+
+	ret = drm_fb_helper_init(dev, &ifbdev->helper);
+	if (ret) {
+		kfree(ifbdev);
+		return ret;
+	}
+
+	dev_priv->display.fbdev.fbdev = ifbdev;
+	INIT_WORK(&dev_priv->display.fbdev.suspend_work, intel_fbdev_suspend_worker);
+
+	return 0;
+}
+
+static void intel_fbdev_initial_config(void *data, async_cookie_t cookie)
+{
+	struct intel_fbdev *ifbdev = data;
+
+	/* Due to peculiar init order wrt to hpd handling this is separate. */
+	if (drm_fb_helper_initial_config(&ifbdev->helper,
+					 ifbdev->preferred_bpp))
+		intel_fbdev_unregister(to_i915(ifbdev->helper.dev));
+}
+
+void intel_fbdev_initial_config_async(struct drm_device *dev)
+{
+	struct intel_fbdev *ifbdev = to_i915(dev)->display.fbdev.fbdev;
+
+	if (!ifbdev)
+		return;
+
+	ifbdev->cookie = async_schedule(intel_fbdev_initial_config, ifbdev);
+}
+
+static void intel_fbdev_sync(struct intel_fbdev *ifbdev)
+{
+	if (!ifbdev->cookie)
+		return;
+
+	/* Only serialises with all preceding async calls, hence +1 */
+	async_synchronize_cookie(ifbdev->cookie + 1);
+	ifbdev->cookie = 0;
+}
+
+void intel_fbdev_unregister(struct drm_i915_private *dev_priv)
+{
+	struct intel_fbdev *ifbdev = dev_priv->display.fbdev.fbdev;
+
+	if (!ifbdev)
+		return;
+
+	intel_fbdev_set_suspend(&dev_priv->drm, FBINFO_STATE_SUSPENDED, true);
+
+	if (!current_is_async())
+		intel_fbdev_sync(ifbdev);
+
+	drm_fb_helper_unregister_fbi(&ifbdev->helper);
+}
+
+void intel_fbdev_fini(struct drm_i915_private *dev_priv)
+{
+	struct intel_fbdev *ifbdev = fetch_and_zero(&dev_priv->display.fbdev.fbdev);
+
+	if (!ifbdev)
+		return;
+
+	intel_fbdev_destroy(ifbdev);
+}
+
+/* Suspends/resumes fbdev processing of incoming HPD events. When resuming HPD
+ * processing, fbdev will perform a full connector reprobe if a hotplug event
+ * was received while HPD was suspended.
+ */
+static void intel_fbdev_hpd_set_suspend(struct drm_i915_private *i915, int state)
+{
+	struct intel_fbdev *ifbdev = i915->display.fbdev.fbdev;
+	bool send_hpd = false;
+
+	mutex_lock(&ifbdev->hpd_lock);
+	ifbdev->hpd_suspended = state == FBINFO_STATE_SUSPENDED;
+	send_hpd = !ifbdev->hpd_suspended && ifbdev->hpd_waiting;
+	ifbdev->hpd_waiting = false;
+	mutex_unlock(&ifbdev->hpd_lock);
+
+	if (send_hpd) {
+		drm_dbg_kms(&i915->drm, "Handling delayed fbcon HPD event\n");
+		drm_fb_helper_hotplug_event(&ifbdev->helper);
+	}
+}
+
+void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_fbdev *ifbdev = dev_priv->display.fbdev.fbdev;
+	struct fb_info *info;
+
+	if (!ifbdev)
+		return;
+
+	if (drm_WARN_ON(&dev_priv->drm, !HAS_DISPLAY(dev_priv)))
+		return;
+
+	if (!ifbdev->vma)
+		goto set_suspend;
+
+	info = ifbdev->helper.fbdev;
+
+	if (synchronous) {
+		/* Flush any pending work to turn the console on, and then
+		 * wait to turn it off. It must be synchronous as we are
+		 * about to suspend or unload the driver.
+		 *
+		 * Note that from within the work-handler, we cannot flush
+		 * ourselves, so only flush outstanding work upon suspend!
+		 */
+		if (state != FBINFO_STATE_RUNNING)
+			flush_work(&dev_priv->display.fbdev.suspend_work);
+
+		console_lock();
+	} else {
+		/*
+		 * The console lock can be pretty contented on resume due
+		 * to all the printk activity.  Try to keep it out of the hot
+		 * path of resume if possible.
+		 */
+		drm_WARN_ON(dev, state != FBINFO_STATE_RUNNING);
+		if (!console_trylock()) {
+			/* Don't block our own workqueue as this can
+			 * be run in parallel with other i915.ko tasks.
+			 */
+			schedule_work(&dev_priv->display.fbdev.suspend_work);
+			return;
+		}
+	}
+
+	/* On resume from hibernation: If the object is shmemfs backed, it has
+	 * been restored from swap. If the object is stolen however, it will be
+	 * full of whatever garbage was left in there.
+	 */
+	if (state == FBINFO_STATE_RUNNING &&
+	    !i915_gem_object_is_shmem(intel_fb_obj(&ifbdev->fb->base)))
+		memset_io(info->screen_base, 0, info->screen_size);
+
+	drm_fb_helper_set_suspend(&ifbdev->helper, state);
+	console_unlock();
+
+set_suspend:
+	intel_fbdev_hpd_set_suspend(dev_priv, state);
+}
+
+void intel_fbdev_output_poll_changed(struct drm_device *dev)
+{
+	struct intel_fbdev *ifbdev = to_i915(dev)->display.fbdev.fbdev;
+	bool send_hpd;
+
+	if (!ifbdev)
+		return;
+
+	intel_fbdev_sync(ifbdev);
+
+	mutex_lock(&ifbdev->hpd_lock);
+	send_hpd = !ifbdev->hpd_suspended;
+	ifbdev->hpd_waiting = true;
+	mutex_unlock(&ifbdev->hpd_lock);
+
+	if (send_hpd && (ifbdev->vma || ifbdev->helper.deferred_setup))
+		drm_fb_helper_hotplug_event(&ifbdev->helper);
+}
+
+void intel_fbdev_restore_mode(struct drm_device *dev)
+{
+	struct intel_fbdev *ifbdev = to_i915(dev)->display.fbdev.fbdev;
+
+	if (!ifbdev)
+		return;
+
+	intel_fbdev_sync(ifbdev);
+	if (!ifbdev->vma)
+		return;
+
+	if (drm_fb_helper_restore_fbdev_mode_unlocked(&ifbdev->helper) == 0)
+		intel_fbdev_invalidate(ifbdev);
+}
+
+struct intel_framebuffer *intel_fbdev_framebuffer(struct intel_fbdev *fbdev)
+{
+	if (!fbdev || !fbdev->helper.fb)
+		return NULL;
+
+	return to_intel_framebuffer(fbdev->helper.fb);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_fbdev.h b/drivers/gpu/drm/i915/display/intel_fbdev.h
new file mode 100644
index 000000000..0e95e9472
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fbdev.h
@@ -0,0 +1,60 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_FBDEV_H__
+#define __INTEL_FBDEV_H__
+
+#include <linux/types.h>
+
+struct drm_device;
+struct drm_i915_private;
+struct intel_fbdev;
+struct intel_framebuffer;
+
+#ifdef CONFIG_DRM_FBDEV_EMULATION
+int intel_fbdev_init(struct drm_device *dev);
+void intel_fbdev_initial_config_async(struct drm_device *dev);
+void intel_fbdev_unregister(struct drm_i915_private *dev_priv);
+void intel_fbdev_fini(struct drm_i915_private *dev_priv);
+void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous);
+void intel_fbdev_output_poll_changed(struct drm_device *dev);
+void intel_fbdev_restore_mode(struct drm_device *dev);
+struct intel_framebuffer *intel_fbdev_framebuffer(struct intel_fbdev *fbdev);
+#else
+static inline int intel_fbdev_init(struct drm_device *dev)
+{
+	return 0;
+}
+
+static inline void intel_fbdev_initial_config_async(struct drm_device *dev)
+{
+}
+
+static inline void intel_fbdev_unregister(struct drm_i915_private *dev_priv)
+{
+}
+
+static inline void intel_fbdev_fini(struct drm_i915_private *dev_priv)
+{
+}
+
+static inline void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous)
+{
+}
+
+static inline void intel_fbdev_output_poll_changed(struct drm_device *dev)
+{
+}
+
+static inline void intel_fbdev_restore_mode(struct drm_device *dev)
+{
+}
+static inline struct intel_framebuffer *intel_fbdev_framebuffer(struct intel_fbdev *fbdev)
+{
+	return NULL;
+}
+#endif
+
+#endif /* __INTEL_FBDEV_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_fdi.c b/drivers/gpu/drm/i915/display/intel_fdi.c
new file mode 100644
index 000000000..7f47e5c85
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fdi.c
@@ -0,0 +1,1076 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+
+#include "intel_atomic.h"
+#include "intel_crtc.h"
+#include "intel_ddi.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_fdi.h"
+
+struct intel_fdi_funcs {
+	void (*fdi_link_train)(struct intel_crtc *crtc,
+			       const struct intel_crtc_state *crtc_state);
+};
+
+static void assert_fdi_tx(struct drm_i915_private *dev_priv,
+			  enum pipe pipe, bool state)
+{
+	bool cur_state;
+
+	if (HAS_DDI(dev_priv)) {
+		/*
+		 * DDI does not have a specific FDI_TX register.
+		 *
+		 * FDI is never fed from EDP transcoder
+		 * so pipe->transcoder cast is fine here.
+		 */
+		enum transcoder cpu_transcoder = (enum transcoder)pipe;
+		cur_state = intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder)) & TRANS_DDI_FUNC_ENABLE;
+	} else {
+		cur_state = intel_de_read(dev_priv, FDI_TX_CTL(pipe)) & FDI_TX_ENABLE;
+	}
+	I915_STATE_WARN(cur_state != state,
+			"FDI TX state assertion failure (expected %s, current %s)\n",
+			str_on_off(state), str_on_off(cur_state));
+}
+
+void assert_fdi_tx_enabled(struct drm_i915_private *i915, enum pipe pipe)
+{
+	assert_fdi_tx(i915, pipe, true);
+}
+
+void assert_fdi_tx_disabled(struct drm_i915_private *i915, enum pipe pipe)
+{
+	assert_fdi_tx(i915, pipe, false);
+}
+
+static void assert_fdi_rx(struct drm_i915_private *dev_priv,
+			  enum pipe pipe, bool state)
+{
+	bool cur_state;
+
+	cur_state = intel_de_read(dev_priv, FDI_RX_CTL(pipe)) & FDI_RX_ENABLE;
+	I915_STATE_WARN(cur_state != state,
+			"FDI RX state assertion failure (expected %s, current %s)\n",
+			str_on_off(state), str_on_off(cur_state));
+}
+
+void assert_fdi_rx_enabled(struct drm_i915_private *i915, enum pipe pipe)
+{
+	assert_fdi_rx(i915, pipe, true);
+}
+
+void assert_fdi_rx_disabled(struct drm_i915_private *i915, enum pipe pipe)
+{
+	assert_fdi_rx(i915, pipe, false);
+}
+
+void assert_fdi_tx_pll_enabled(struct drm_i915_private *i915,
+			       enum pipe pipe)
+{
+	bool cur_state;
+
+	/* ILK FDI PLL is always enabled */
+	if (IS_IRONLAKE(i915))
+		return;
+
+	/* On Haswell, DDI ports are responsible for the FDI PLL setup */
+	if (HAS_DDI(i915))
+		return;
+
+	cur_state = intel_de_read(i915, FDI_TX_CTL(pipe)) & FDI_TX_PLL_ENABLE;
+	I915_STATE_WARN(!cur_state, "FDI TX PLL assertion failure, should be active but is disabled\n");
+}
+
+static void assert_fdi_rx_pll(struct drm_i915_private *i915,
+			      enum pipe pipe, bool state)
+{
+	bool cur_state;
+
+	cur_state = intel_de_read(i915, FDI_RX_CTL(pipe)) & FDI_RX_PLL_ENABLE;
+	I915_STATE_WARN(cur_state != state,
+			"FDI RX PLL assertion failure (expected %s, current %s)\n",
+			str_on_off(state), str_on_off(cur_state));
+}
+
+void assert_fdi_rx_pll_enabled(struct drm_i915_private *i915, enum pipe pipe)
+{
+	assert_fdi_rx_pll(i915, pipe, true);
+}
+
+void assert_fdi_rx_pll_disabled(struct drm_i915_private *i915, enum pipe pipe)
+{
+	assert_fdi_rx_pll(i915, pipe, false);
+}
+
+void intel_fdi_link_train(struct intel_crtc *crtc,
+			  const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	dev_priv->display.funcs.fdi->fdi_link_train(crtc, crtc_state);
+}
+
+/* units of 100MHz */
+static int pipe_required_fdi_lanes(struct intel_crtc_state *crtc_state)
+{
+	if (crtc_state->hw.enable && crtc_state->has_pch_encoder)
+		return crtc_state->fdi_lanes;
+
+	return 0;
+}
+
+static int ilk_check_fdi_lanes(struct drm_device *dev, enum pipe pipe,
+			       struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_atomic_state *state = pipe_config->uapi.state;
+	struct intel_crtc *other_crtc;
+	struct intel_crtc_state *other_crtc_state;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "checking fdi config on pipe %c, lanes %i\n",
+		    pipe_name(pipe), pipe_config->fdi_lanes);
+	if (pipe_config->fdi_lanes > 4) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "invalid fdi lane config on pipe %c: %i lanes\n",
+			    pipe_name(pipe), pipe_config->fdi_lanes);
+		return -EINVAL;
+	}
+
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
+		if (pipe_config->fdi_lanes > 2) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "only 2 lanes on haswell, required: %i lanes\n",
+				    pipe_config->fdi_lanes);
+			return -EINVAL;
+		} else {
+			return 0;
+		}
+	}
+
+	if (INTEL_NUM_PIPES(dev_priv) == 2)
+		return 0;
+
+	/* Ivybridge 3 pipe is really complicated */
+	switch (pipe) {
+	case PIPE_A:
+		return 0;
+	case PIPE_B:
+		if (pipe_config->fdi_lanes <= 2)
+			return 0;
+
+		other_crtc = intel_crtc_for_pipe(dev_priv, PIPE_C);
+		other_crtc_state =
+			intel_atomic_get_crtc_state(state, other_crtc);
+		if (IS_ERR(other_crtc_state))
+			return PTR_ERR(other_crtc_state);
+
+		if (pipe_required_fdi_lanes(other_crtc_state) > 0) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "invalid shared fdi lane config on pipe %c: %i lanes\n",
+				    pipe_name(pipe), pipe_config->fdi_lanes);
+			return -EINVAL;
+		}
+		return 0;
+	case PIPE_C:
+		if (pipe_config->fdi_lanes > 2) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "only 2 lanes on pipe %c: required %i lanes\n",
+				    pipe_name(pipe), pipe_config->fdi_lanes);
+			return -EINVAL;
+		}
+
+		other_crtc = intel_crtc_for_pipe(dev_priv, PIPE_B);
+		other_crtc_state =
+			intel_atomic_get_crtc_state(state, other_crtc);
+		if (IS_ERR(other_crtc_state))
+			return PTR_ERR(other_crtc_state);
+
+		if (pipe_required_fdi_lanes(other_crtc_state) > 2) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "fdi link B uses too many lanes to enable link C\n");
+			return -EINVAL;
+		}
+		return 0;
+	default:
+		MISSING_CASE(pipe);
+		return 0;
+	}
+}
+
+void intel_fdi_pll_freq_update(struct drm_i915_private *i915)
+{
+	if (IS_IRONLAKE(i915)) {
+		u32 fdi_pll_clk =
+			intel_de_read(i915, FDI_PLL_BIOS_0) & FDI_PLL_FB_CLOCK_MASK;
+
+		i915->display.fdi.pll_freq = (fdi_pll_clk + 2) * 10000;
+	} else if (IS_SANDYBRIDGE(i915) || IS_IVYBRIDGE(i915)) {
+		i915->display.fdi.pll_freq = 270000;
+	} else {
+		return;
+	}
+
+	drm_dbg(&i915->drm, "FDI PLL freq=%d\n", i915->display.fdi.pll_freq);
+}
+
+int intel_fdi_link_freq(struct drm_i915_private *i915,
+			const struct intel_crtc_state *pipe_config)
+{
+	if (HAS_DDI(i915))
+		return pipe_config->port_clock; /* SPLL */
+	else
+		return i915->display.fdi.pll_freq;
+}
+
+int ilk_fdi_compute_config(struct intel_crtc *crtc,
+			   struct intel_crtc_state *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *i915 = to_i915(dev);
+	const struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	int lane, link_bw, fdi_dotclock, ret;
+	bool needs_recompute = false;
+
+retry:
+	/* FDI is a binary signal running at ~2.7GHz, encoding
+	 * each output octet as 10 bits. The actual frequency
+	 * is stored as a divider into a 100MHz clock, and the
+	 * mode pixel clock is stored in units of 1KHz.
+	 * Hence the bw of each lane in terms of the mode signal
+	 * is:
+	 */
+	link_bw = intel_fdi_link_freq(i915, pipe_config);
+
+	fdi_dotclock = adjusted_mode->crtc_clock;
+
+	lane = ilk_get_lanes_required(fdi_dotclock, link_bw,
+				      pipe_config->pipe_bpp);
+
+	pipe_config->fdi_lanes = lane;
+
+	intel_link_compute_m_n(pipe_config->pipe_bpp, lane, fdi_dotclock,
+			       link_bw, &pipe_config->fdi_m_n, false);
+
+	ret = ilk_check_fdi_lanes(dev, crtc->pipe, pipe_config);
+	if (ret == -EDEADLK)
+		return ret;
+
+	if (ret == -EINVAL && pipe_config->pipe_bpp > 6*3) {
+		pipe_config->pipe_bpp -= 2*3;
+		drm_dbg_kms(&i915->drm,
+			    "fdi link bw constraint, reducing pipe bpp to %i\n",
+			    pipe_config->pipe_bpp);
+		needs_recompute = true;
+		pipe_config->bw_constrained = true;
+
+		goto retry;
+	}
+
+	if (needs_recompute)
+		return -EAGAIN;
+
+	return ret;
+}
+
+static void cpt_set_fdi_bc_bifurcation(struct drm_i915_private *dev_priv, bool enable)
+{
+	u32 temp;
+
+	temp = intel_de_read(dev_priv, SOUTH_CHICKEN1);
+	if (!!(temp & FDI_BC_BIFURCATION_SELECT) == enable)
+		return;
+
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, FDI_RX_CTL(PIPE_B)) &
+		    FDI_RX_ENABLE);
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, FDI_RX_CTL(PIPE_C)) &
+		    FDI_RX_ENABLE);
+
+	temp &= ~FDI_BC_BIFURCATION_SELECT;
+	if (enable)
+		temp |= FDI_BC_BIFURCATION_SELECT;
+
+	drm_dbg_kms(&dev_priv->drm, "%sabling fdi C rx\n",
+		    enable ? "en" : "dis");
+	intel_de_write(dev_priv, SOUTH_CHICKEN1, temp);
+	intel_de_posting_read(dev_priv, SOUTH_CHICKEN1);
+}
+
+static void ivb_update_fdi_bc_bifurcation(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	switch (crtc->pipe) {
+	case PIPE_A:
+		break;
+	case PIPE_B:
+		if (crtc_state->fdi_lanes > 2)
+			cpt_set_fdi_bc_bifurcation(dev_priv, false);
+		else
+			cpt_set_fdi_bc_bifurcation(dev_priv, true);
+
+		break;
+	case PIPE_C:
+		cpt_set_fdi_bc_bifurcation(dev_priv, true);
+
+		break;
+	default:
+		MISSING_CASE(crtc->pipe);
+	}
+}
+
+void intel_fdi_normal_train(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp;
+
+	/* enable normal train */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if (IS_IVYBRIDGE(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_NONE_IVB;
+		temp |= FDI_LINK_TRAIN_NONE_IVB | FDI_TX_ENHANCE_FRAME_ENABLE;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_NONE | FDI_TX_ENHANCE_FRAME_ENABLE;
+	}
+	intel_de_write(dev_priv, reg, temp);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if (HAS_PCH_CPT(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_NORMAL_CPT;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_NONE;
+	}
+	intel_de_write(dev_priv, reg, temp | FDI_RX_ENHANCE_FRAME_ENABLE);
+
+	/* wait one idle pattern time */
+	intel_de_posting_read(dev_priv, reg);
+	udelay(1000);
+
+	/* IVB wants error correction enabled */
+	if (IS_IVYBRIDGE(dev_priv))
+		intel_de_write(dev_priv, reg,
+			       intel_de_read(dev_priv, reg) | FDI_FS_ERRC_ENABLE | FDI_FE_ERRC_ENABLE);
+}
+
+/* The FDI link training functions for ILK/Ibexpeak. */
+static void ilk_fdi_link_train(struct intel_crtc *crtc,
+			       const struct intel_crtc_state *crtc_state)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp, tries;
+
+	/*
+	 * Write the TU size bits before fdi link training, so that error
+	 * detection works.
+	 */
+	intel_de_write(dev_priv, FDI_RX_TUSIZE1(pipe),
+		       intel_de_read(dev_priv, PIPE_DATA_M1(pipe)) & TU_SIZE_MASK);
+
+	/* FDI needs bits from pipe first */
+	assert_transcoder_enabled(dev_priv, crtc_state->cpu_transcoder);
+
+	/* Train 1: umask FDI RX Interrupt symbol_lock and bit_lock bit
+	   for train result */
+	reg = FDI_RX_IMR(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_RX_SYMBOL_LOCK;
+	temp &= ~FDI_RX_BIT_LOCK;
+	intel_de_write(dev_priv, reg, temp);
+	intel_de_read(dev_priv, reg);
+	udelay(150);
+
+	/* enable CPU FDI TX and PCH FDI RX */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_DP_PORT_WIDTH_MASK;
+	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_1;
+	intel_de_write(dev_priv, reg, temp | FDI_TX_ENABLE);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_1;
+	intel_de_write(dev_priv, reg, temp | FDI_RX_ENABLE);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	/* Ironlake workaround, enable clock pointer after FDI enable*/
+	intel_de_write(dev_priv, FDI_RX_CHICKEN(pipe),
+		       FDI_RX_PHASE_SYNC_POINTER_OVR);
+	intel_de_write(dev_priv, FDI_RX_CHICKEN(pipe),
+		       FDI_RX_PHASE_SYNC_POINTER_OVR | FDI_RX_PHASE_SYNC_POINTER_EN);
+
+	reg = FDI_RX_IIR(pipe);
+	for (tries = 0; tries < 5; tries++) {
+		temp = intel_de_read(dev_priv, reg);
+		drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+
+		if ((temp & FDI_RX_BIT_LOCK)) {
+			drm_dbg_kms(&dev_priv->drm, "FDI train 1 done.\n");
+			intel_de_write(dev_priv, reg, temp | FDI_RX_BIT_LOCK);
+			break;
+		}
+	}
+	if (tries == 5)
+		drm_err(&dev_priv->drm, "FDI train 1 fail!\n");
+
+	/* Train 2 */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_2;
+	intel_de_write(dev_priv, reg, temp);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_2;
+	intel_de_write(dev_priv, reg, temp);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	reg = FDI_RX_IIR(pipe);
+	for (tries = 0; tries < 5; tries++) {
+		temp = intel_de_read(dev_priv, reg);
+		drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+
+		if (temp & FDI_RX_SYMBOL_LOCK) {
+			intel_de_write(dev_priv, reg,
+				       temp | FDI_RX_SYMBOL_LOCK);
+			drm_dbg_kms(&dev_priv->drm, "FDI train 2 done.\n");
+			break;
+		}
+	}
+	if (tries == 5)
+		drm_err(&dev_priv->drm, "FDI train 2 fail!\n");
+
+	drm_dbg_kms(&dev_priv->drm, "FDI train done\n");
+
+}
+
+static const int snb_b_fdi_train_param[] = {
+	FDI_LINK_TRAIN_400MV_0DB_SNB_B,
+	FDI_LINK_TRAIN_400MV_6DB_SNB_B,
+	FDI_LINK_TRAIN_600MV_3_5DB_SNB_B,
+	FDI_LINK_TRAIN_800MV_0DB_SNB_B,
+};
+
+/* The FDI link training functions for SNB/Cougarpoint. */
+static void gen6_fdi_link_train(struct intel_crtc *crtc,
+				const struct intel_crtc_state *crtc_state)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp, i, retry;
+
+	/*
+	 * Write the TU size bits before fdi link training, so that error
+	 * detection works.
+	 */
+	intel_de_write(dev_priv, FDI_RX_TUSIZE1(pipe),
+		       intel_de_read(dev_priv, PIPE_DATA_M1(pipe)) & TU_SIZE_MASK);
+
+	/* Train 1: umask FDI RX Interrupt symbol_lock and bit_lock bit
+	   for train result */
+	reg = FDI_RX_IMR(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_RX_SYMBOL_LOCK;
+	temp &= ~FDI_RX_BIT_LOCK;
+	intel_de_write(dev_priv, reg, temp);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	/* enable CPU FDI TX and PCH FDI RX */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_DP_PORT_WIDTH_MASK;
+	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_1;
+	temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
+	/* SNB-B */
+	temp |= FDI_LINK_TRAIN_400MV_0DB_SNB_B;
+	intel_de_write(dev_priv, reg, temp | FDI_TX_ENABLE);
+
+	intel_de_write(dev_priv, FDI_RX_MISC(pipe),
+		       FDI_RX_TP1_TO_TP2_48 | FDI_RX_FDI_DELAY_90);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if (HAS_PCH_CPT(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_PATTERN_1;
+	}
+	intel_de_write(dev_priv, reg, temp | FDI_RX_ENABLE);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	for (i = 0; i < 4; i++) {
+		reg = FDI_TX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
+		temp |= snb_b_fdi_train_param[i];
+		intel_de_write(dev_priv, reg, temp);
+
+		intel_de_posting_read(dev_priv, reg);
+		udelay(500);
+
+		for (retry = 0; retry < 5; retry++) {
+			reg = FDI_RX_IIR(pipe);
+			temp = intel_de_read(dev_priv, reg);
+			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+			if (temp & FDI_RX_BIT_LOCK) {
+				intel_de_write(dev_priv, reg,
+					       temp | FDI_RX_BIT_LOCK);
+				drm_dbg_kms(&dev_priv->drm,
+					    "FDI train 1 done.\n");
+				break;
+			}
+			udelay(50);
+		}
+		if (retry < 5)
+			break;
+	}
+	if (i == 4)
+		drm_err(&dev_priv->drm, "FDI train 1 fail!\n");
+
+	/* Train 2 */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_2;
+	if (IS_SANDYBRIDGE(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
+		/* SNB-B */
+		temp |= FDI_LINK_TRAIN_400MV_0DB_SNB_B;
+	}
+	intel_de_write(dev_priv, reg, temp);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if (HAS_PCH_CPT(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_PATTERN_2_CPT;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_PATTERN_2;
+	}
+	intel_de_write(dev_priv, reg, temp);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	for (i = 0; i < 4; i++) {
+		reg = FDI_TX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
+		temp |= snb_b_fdi_train_param[i];
+		intel_de_write(dev_priv, reg, temp);
+
+		intel_de_posting_read(dev_priv, reg);
+		udelay(500);
+
+		for (retry = 0; retry < 5; retry++) {
+			reg = FDI_RX_IIR(pipe);
+			temp = intel_de_read(dev_priv, reg);
+			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+			if (temp & FDI_RX_SYMBOL_LOCK) {
+				intel_de_write(dev_priv, reg,
+					       temp | FDI_RX_SYMBOL_LOCK);
+				drm_dbg_kms(&dev_priv->drm,
+					    "FDI train 2 done.\n");
+				break;
+			}
+			udelay(50);
+		}
+		if (retry < 5)
+			break;
+	}
+	if (i == 4)
+		drm_err(&dev_priv->drm, "FDI train 2 fail!\n");
+
+	drm_dbg_kms(&dev_priv->drm, "FDI train done.\n");
+}
+
+/* Manual link training for Ivy Bridge A0 parts */
+static void ivb_manual_fdi_link_train(struct intel_crtc *crtc,
+				      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp, i, j;
+
+	ivb_update_fdi_bc_bifurcation(crtc_state);
+
+	/*
+	 * Write the TU size bits before fdi link training, so that error
+	 * detection works.
+	 */
+	intel_de_write(dev_priv, FDI_RX_TUSIZE1(pipe),
+		       intel_de_read(dev_priv, PIPE_DATA_M1(pipe)) & TU_SIZE_MASK);
+
+	/* Train 1: umask FDI RX Interrupt symbol_lock and bit_lock bit
+	   for train result */
+	reg = FDI_RX_IMR(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_RX_SYMBOL_LOCK;
+	temp &= ~FDI_RX_BIT_LOCK;
+	intel_de_write(dev_priv, reg, temp);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(150);
+
+	drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR before link train 0x%x\n",
+		    intel_de_read(dev_priv, FDI_RX_IIR(pipe)));
+
+	/* Try each vswing and preemphasis setting twice before moving on */
+	for (j = 0; j < ARRAY_SIZE(snb_b_fdi_train_param) * 2; j++) {
+		/* disable first in case we need to retry */
+		reg = FDI_TX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~(FDI_LINK_TRAIN_AUTO | FDI_LINK_TRAIN_NONE_IVB);
+		temp &= ~FDI_TX_ENABLE;
+		intel_de_write(dev_priv, reg, temp);
+
+		reg = FDI_RX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_LINK_TRAIN_AUTO;
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp &= ~FDI_RX_ENABLE;
+		intel_de_write(dev_priv, reg, temp);
+
+		/* enable CPU FDI TX and PCH FDI RX */
+		reg = FDI_TX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_DP_PORT_WIDTH_MASK;
+		temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
+		temp |= FDI_LINK_TRAIN_PATTERN_1_IVB;
+		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
+		temp |= snb_b_fdi_train_param[j/2];
+		temp |= FDI_COMPOSITE_SYNC;
+		intel_de_write(dev_priv, reg, temp | FDI_TX_ENABLE);
+
+		intel_de_write(dev_priv, FDI_RX_MISC(pipe),
+			       FDI_RX_TP1_TO_TP2_48 | FDI_RX_FDI_DELAY_90);
+
+		reg = FDI_RX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
+		temp |= FDI_COMPOSITE_SYNC;
+		intel_de_write(dev_priv, reg, temp | FDI_RX_ENABLE);
+
+		intel_de_posting_read(dev_priv, reg);
+		udelay(1); /* should be 0.5us */
+
+		for (i = 0; i < 4; i++) {
+			reg = FDI_RX_IIR(pipe);
+			temp = intel_de_read(dev_priv, reg);
+			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+
+			if (temp & FDI_RX_BIT_LOCK ||
+			    (intel_de_read(dev_priv, reg) & FDI_RX_BIT_LOCK)) {
+				intel_de_write(dev_priv, reg,
+					       temp | FDI_RX_BIT_LOCK);
+				drm_dbg_kms(&dev_priv->drm,
+					    "FDI train 1 done, level %i.\n",
+					    i);
+				break;
+			}
+			udelay(1); /* should be 0.5us */
+		}
+		if (i == 4) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "FDI train 1 fail on vswing %d\n", j / 2);
+			continue;
+		}
+
+		/* Train 2 */
+		reg = FDI_TX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_LINK_TRAIN_NONE_IVB;
+		temp |= FDI_LINK_TRAIN_PATTERN_2_IVB;
+		intel_de_write(dev_priv, reg, temp);
+
+		reg = FDI_RX_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_PATTERN_2_CPT;
+		intel_de_write(dev_priv, reg, temp);
+
+		intel_de_posting_read(dev_priv, reg);
+		udelay(2); /* should be 1.5us */
+
+		for (i = 0; i < 4; i++) {
+			reg = FDI_RX_IIR(pipe);
+			temp = intel_de_read(dev_priv, reg);
+			drm_dbg_kms(&dev_priv->drm, "FDI_RX_IIR 0x%x\n", temp);
+
+			if (temp & FDI_RX_SYMBOL_LOCK ||
+			    (intel_de_read(dev_priv, reg) & FDI_RX_SYMBOL_LOCK)) {
+				intel_de_write(dev_priv, reg,
+					       temp | FDI_RX_SYMBOL_LOCK);
+				drm_dbg_kms(&dev_priv->drm,
+					    "FDI train 2 done, level %i.\n",
+					    i);
+				goto train_done;
+			}
+			udelay(2); /* should be 1.5us */
+		}
+		if (i == 4)
+			drm_dbg_kms(&dev_priv->drm,
+				    "FDI train 2 fail on vswing %d\n", j / 2);
+	}
+
+train_done:
+	drm_dbg_kms(&dev_priv->drm, "FDI train done.\n");
+}
+
+/* Starting with Haswell, different DDI ports can work in FDI mode for
+ * connection to the PCH-located connectors. For this, it is necessary to train
+ * both the DDI port and PCH receiver for the desired DDI buffer settings.
+ *
+ * The recommended port to work in FDI mode is DDI E, which we use here. Also,
+ * please note that when FDI mode is active on DDI E, it shares 2 lines with
+ * DDI A (which is used for eDP)
+ */
+void hsw_fdi_link_train(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 temp, i, rx_ctl_val;
+	int n_entries;
+
+	encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+
+	hsw_prepare_dp_ddi_buffers(encoder, crtc_state);
+
+	/* Set the FDI_RX_MISC pwrdn lanes and the 2 workarounds listed at the
+	 * mode set "sequence for CRT port" document:
+	 * - TP1 to TP2 time with the default value
+	 * - FDI delay to 90h
+	 *
+	 * WaFDIAutoLinkSetTimingOverrride:hsw
+	 */
+	intel_de_write(dev_priv, FDI_RX_MISC(PIPE_A),
+		       FDI_RX_PWRDN_LANE1_VAL(2) | FDI_RX_PWRDN_LANE0_VAL(2) | FDI_RX_TP1_TO_TP2_48 | FDI_RX_FDI_DELAY_90);
+
+	/* Enable the PCH Receiver FDI PLL */
+	rx_ctl_val = dev_priv->display.fdi.rx_config | FDI_RX_ENHANCE_FRAME_ENABLE |
+		     FDI_RX_PLL_ENABLE |
+		     FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
+	intel_de_write(dev_priv, FDI_RX_CTL(PIPE_A), rx_ctl_val);
+	intel_de_posting_read(dev_priv, FDI_RX_CTL(PIPE_A));
+	udelay(220);
+
+	/* Switch from Rawclk to PCDclk */
+	rx_ctl_val |= FDI_PCDCLK;
+	intel_de_write(dev_priv, FDI_RX_CTL(PIPE_A), rx_ctl_val);
+
+	/* Configure Port Clock Select */
+	drm_WARN_ON(&dev_priv->drm, crtc_state->shared_dpll->info->id != DPLL_ID_SPLL);
+	intel_ddi_enable_clock(encoder, crtc_state);
+
+	/* Start the training iterating through available voltages and emphasis,
+	 * testing each value twice. */
+	for (i = 0; i < n_entries * 2; i++) {
+		/* Configure DP_TP_CTL with auto-training */
+		intel_de_write(dev_priv, DP_TP_CTL(PORT_E),
+			       DP_TP_CTL_FDI_AUTOTRAIN |
+			       DP_TP_CTL_ENHANCED_FRAME_ENABLE |
+			       DP_TP_CTL_LINK_TRAIN_PAT1 |
+			       DP_TP_CTL_ENABLE);
+
+		/* Configure and enable DDI_BUF_CTL for DDI E with next voltage.
+		 * DDI E does not support port reversal, the functionality is
+		 * achieved on the PCH side in FDI_RX_CTL, so no need to set the
+		 * port reversal bit */
+		intel_de_write(dev_priv, DDI_BUF_CTL(PORT_E),
+			       DDI_BUF_CTL_ENABLE | ((crtc_state->fdi_lanes - 1) << 1) | DDI_BUF_TRANS_SELECT(i / 2));
+		intel_de_posting_read(dev_priv, DDI_BUF_CTL(PORT_E));
+
+		udelay(600);
+
+		/* Program PCH FDI Receiver TU */
+		intel_de_write(dev_priv, FDI_RX_TUSIZE1(PIPE_A), TU_SIZE(64));
+
+		/* Enable PCH FDI Receiver with auto-training */
+		rx_ctl_val |= FDI_RX_ENABLE | FDI_LINK_TRAIN_AUTO;
+		intel_de_write(dev_priv, FDI_RX_CTL(PIPE_A), rx_ctl_val);
+		intel_de_posting_read(dev_priv, FDI_RX_CTL(PIPE_A));
+
+		/* Wait for FDI receiver lane calibration */
+		udelay(30);
+
+		/* Unset FDI_RX_MISC pwrdn lanes */
+		temp = intel_de_read(dev_priv, FDI_RX_MISC(PIPE_A));
+		temp &= ~(FDI_RX_PWRDN_LANE1_MASK | FDI_RX_PWRDN_LANE0_MASK);
+		intel_de_write(dev_priv, FDI_RX_MISC(PIPE_A), temp);
+		intel_de_posting_read(dev_priv, FDI_RX_MISC(PIPE_A));
+
+		/* Wait for FDI auto training time */
+		udelay(5);
+
+		temp = intel_de_read(dev_priv, DP_TP_STATUS(PORT_E));
+		if (temp & DP_TP_STATUS_AUTOTRAIN_DONE) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "FDI link training done on step %d\n", i);
+			break;
+		}
+
+		/*
+		 * Leave things enabled even if we failed to train FDI.
+		 * Results in less fireworks from the state checker.
+		 */
+		if (i == n_entries * 2 - 1) {
+			drm_err(&dev_priv->drm, "FDI link training failed!\n");
+			break;
+		}
+
+		rx_ctl_val &= ~FDI_RX_ENABLE;
+		intel_de_write(dev_priv, FDI_RX_CTL(PIPE_A), rx_ctl_val);
+		intel_de_posting_read(dev_priv, FDI_RX_CTL(PIPE_A));
+
+		temp = intel_de_read(dev_priv, DDI_BUF_CTL(PORT_E));
+		temp &= ~DDI_BUF_CTL_ENABLE;
+		intel_de_write(dev_priv, DDI_BUF_CTL(PORT_E), temp);
+		intel_de_posting_read(dev_priv, DDI_BUF_CTL(PORT_E));
+
+		/* Disable DP_TP_CTL and FDI_RX_CTL and retry */
+		temp = intel_de_read(dev_priv, DP_TP_CTL(PORT_E));
+		temp &= ~(DP_TP_CTL_ENABLE | DP_TP_CTL_LINK_TRAIN_MASK);
+		temp |= DP_TP_CTL_LINK_TRAIN_PAT1;
+		intel_de_write(dev_priv, DP_TP_CTL(PORT_E), temp);
+		intel_de_posting_read(dev_priv, DP_TP_CTL(PORT_E));
+
+		intel_wait_ddi_buf_idle(dev_priv, PORT_E);
+
+		/* Reset FDI_RX_MISC pwrdn lanes */
+		temp = intel_de_read(dev_priv, FDI_RX_MISC(PIPE_A));
+		temp &= ~(FDI_RX_PWRDN_LANE1_MASK | FDI_RX_PWRDN_LANE0_MASK);
+		temp |= FDI_RX_PWRDN_LANE1_VAL(2) | FDI_RX_PWRDN_LANE0_VAL(2);
+		intel_de_write(dev_priv, FDI_RX_MISC(PIPE_A), temp);
+		intel_de_posting_read(dev_priv, FDI_RX_MISC(PIPE_A));
+	}
+
+	/* Enable normal pixel sending for FDI */
+	intel_de_write(dev_priv, DP_TP_CTL(PORT_E),
+		       DP_TP_CTL_FDI_AUTOTRAIN |
+		       DP_TP_CTL_LINK_TRAIN_NORMAL |
+		       DP_TP_CTL_ENHANCED_FRAME_ENABLE |
+		       DP_TP_CTL_ENABLE);
+}
+
+void hsw_fdi_disable(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 val;
+
+	/*
+	 * Bspec lists this as both step 13 (before DDI_BUF_CTL disable)
+	 * and step 18 (after clearing PORT_CLK_SEL). Based on a BUN,
+	 * step 13 is the correct place for it. Step 18 is where it was
+	 * originally before the BUN.
+	 */
+	val = intel_de_read(dev_priv, FDI_RX_CTL(PIPE_A));
+	val &= ~FDI_RX_ENABLE;
+	intel_de_write(dev_priv, FDI_RX_CTL(PIPE_A), val);
+
+	val = intel_de_read(dev_priv, DDI_BUF_CTL(PORT_E));
+	val &= ~DDI_BUF_CTL_ENABLE;
+	intel_de_write(dev_priv, DDI_BUF_CTL(PORT_E), val);
+
+	intel_wait_ddi_buf_idle(dev_priv, PORT_E);
+
+	intel_ddi_disable_clock(encoder);
+
+	val = intel_de_read(dev_priv, FDI_RX_MISC(PIPE_A));
+	val &= ~(FDI_RX_PWRDN_LANE1_MASK | FDI_RX_PWRDN_LANE0_MASK);
+	val |= FDI_RX_PWRDN_LANE1_VAL(2) | FDI_RX_PWRDN_LANE0_VAL(2);
+	intel_de_write(dev_priv, FDI_RX_MISC(PIPE_A), val);
+
+	val = intel_de_read(dev_priv, FDI_RX_CTL(PIPE_A));
+	val &= ~FDI_PCDCLK;
+	intel_de_write(dev_priv, FDI_RX_CTL(PIPE_A), val);
+
+	val = intel_de_read(dev_priv, FDI_RX_CTL(PIPE_A));
+	val &= ~FDI_RX_PLL_ENABLE;
+	intel_de_write(dev_priv, FDI_RX_CTL(PIPE_A), val);
+}
+
+void ilk_fdi_pll_enable(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp;
+
+	/* enable PCH FDI RX PLL, wait warmup plus DMI latency */
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~(FDI_DP_PORT_WIDTH_MASK | (0x7 << 16));
+	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
+	temp |= (intel_de_read(dev_priv, PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
+	intel_de_write(dev_priv, reg, temp | FDI_RX_PLL_ENABLE);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(200);
+
+	/* Switch from Rawclk to PCDclk */
+	temp = intel_de_read(dev_priv, reg);
+	intel_de_write(dev_priv, reg, temp | FDI_PCDCLK);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(200);
+
+	/* Enable CPU FDI TX PLL, always on for Ironlake */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if ((temp & FDI_TX_PLL_ENABLE) == 0) {
+		intel_de_write(dev_priv, reg, temp | FDI_TX_PLL_ENABLE);
+
+		intel_de_posting_read(dev_priv, reg);
+		udelay(100);
+	}
+}
+
+void ilk_fdi_pll_disable(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp;
+
+	/* Switch from PCDclk to Rawclk */
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	intel_de_write(dev_priv, reg, temp & ~FDI_PCDCLK);
+
+	/* Disable CPU FDI TX PLL */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	intel_de_write(dev_priv, reg, temp & ~FDI_TX_PLL_ENABLE);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(100);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	intel_de_write(dev_priv, reg, temp & ~FDI_RX_PLL_ENABLE);
+
+	/* Wait for the clocks to turn off. */
+	intel_de_posting_read(dev_priv, reg);
+	udelay(100);
+}
+
+void ilk_fdi_disable(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 temp;
+
+	/* disable CPU FDI tx and PCH FDI rx */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	intel_de_write(dev_priv, reg, temp & ~FDI_TX_ENABLE);
+	intel_de_posting_read(dev_priv, reg);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~(0x7 << 16);
+	temp |= (intel_de_read(dev_priv, PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
+	intel_de_write(dev_priv, reg, temp & ~FDI_RX_ENABLE);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(100);
+
+	/* Ironlake workaround, disable clock pointer after downing FDI */
+	if (HAS_PCH_IBX(dev_priv))
+		intel_de_write(dev_priv, FDI_RX_CHICKEN(pipe),
+			       FDI_RX_PHASE_SYNC_POINTER_OVR);
+
+	/* still set train pattern 1 */
+	reg = FDI_TX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_PATTERN_1;
+	intel_de_write(dev_priv, reg, temp);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = intel_de_read(dev_priv, reg);
+	if (HAS_PCH_CPT(dev_priv)) {
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_PATTERN_1_CPT;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_PATTERN_1;
+	}
+	/* BPC in FDI rx is consistent with that in PIPECONF */
+	temp &= ~(0x07 << 16);
+	temp |= (intel_de_read(dev_priv, PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
+	intel_de_write(dev_priv, reg, temp);
+
+	intel_de_posting_read(dev_priv, reg);
+	udelay(100);
+}
+
+static const struct intel_fdi_funcs ilk_funcs = {
+	.fdi_link_train = ilk_fdi_link_train,
+};
+
+static const struct intel_fdi_funcs gen6_funcs = {
+	.fdi_link_train = gen6_fdi_link_train,
+};
+
+static const struct intel_fdi_funcs ivb_funcs = {
+	.fdi_link_train = ivb_manual_fdi_link_train,
+};
+
+void
+intel_fdi_init_hook(struct drm_i915_private *dev_priv)
+{
+	if (IS_IRONLAKE(dev_priv)) {
+		dev_priv->display.funcs.fdi = &ilk_funcs;
+	} else if (IS_SANDYBRIDGE(dev_priv)) {
+		dev_priv->display.funcs.fdi = &gen6_funcs;
+	} else if (IS_IVYBRIDGE(dev_priv)) {
+		/* FIXME: detect B0+ stepping and use auto training */
+		dev_priv->display.funcs.fdi = &ivb_funcs;
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_fdi.h b/drivers/gpu/drm/i915/display/intel_fdi.h
new file mode 100644
index 000000000..1cdb86172
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fdi.h
@@ -0,0 +1,40 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _INTEL_FDI_H_
+#define _INTEL_FDI_H_
+
+enum pipe;
+struct drm_i915_private;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_encoder;
+
+int intel_fdi_link_freq(struct drm_i915_private *i915,
+			const struct intel_crtc_state *pipe_config);
+int ilk_fdi_compute_config(struct intel_crtc *intel_crtc,
+			   struct intel_crtc_state *pipe_config);
+void intel_fdi_normal_train(struct intel_crtc *crtc);
+void ilk_fdi_disable(struct intel_crtc *crtc);
+void ilk_fdi_pll_disable(struct intel_crtc *intel_crtc);
+void ilk_fdi_pll_enable(const struct intel_crtc_state *crtc_state);
+void intel_fdi_init_hook(struct drm_i915_private *dev_priv);
+void hsw_fdi_link_train(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state);
+void hsw_fdi_disable(struct intel_encoder *encoder);
+void intel_fdi_pll_freq_update(struct drm_i915_private *i915);
+
+void intel_fdi_link_train(struct intel_crtc *crtc,
+			  const struct intel_crtc_state *crtc_state);
+
+void assert_fdi_tx_enabled(struct drm_i915_private *i915, enum pipe pipe);
+void assert_fdi_tx_disabled(struct drm_i915_private *i915, enum pipe pipe);
+void assert_fdi_rx_enabled(struct drm_i915_private *i915, enum pipe pipe);
+void assert_fdi_rx_disabled(struct drm_i915_private *i915, enum pipe pipe);
+void assert_fdi_tx_pll_enabled(struct drm_i915_private *i915, enum pipe pipe);
+void assert_fdi_rx_pll_enabled(struct drm_i915_private *i915, enum pipe pipe);
+void assert_fdi_rx_pll_disabled(struct drm_i915_private *i915, enum pipe pipe);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_fifo_underrun.c b/drivers/gpu/drm/i915/display/intel_fifo_underrun.c
new file mode 100644
index 000000000..d636d21fa
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fifo_underrun.c
@@ -0,0 +1,511 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Daniel Vetter <daniel.vetter@ffwll.ch>
+ *
+ */
+
+#include "i915_drv.h"
+#include "intel_de.h"
+#include "intel_display_trace.h"
+#include "intel_display_types.h"
+#include "intel_fbc.h"
+#include "intel_fifo_underrun.h"
+
+/**
+ * DOC: fifo underrun handling
+ *
+ * The i915 driver checks for display fifo underruns using the interrupt signals
+ * provided by the hardware. This is enabled by default and fairly useful to
+ * debug display issues, especially watermark settings.
+ *
+ * If an underrun is detected this is logged into dmesg. To avoid flooding logs
+ * and occupying the cpu underrun interrupts are disabled after the first
+ * occurrence until the next modeset on a given pipe.
+ *
+ * Note that underrun detection on gmch platforms is a bit more ugly since there
+ * is no interrupt (despite that the signalling bit is in the PIPESTAT pipe
+ * interrupt register). Also on some other platforms underrun interrupts are
+ * shared, which means that if we detect an underrun we need to disable underrun
+ * reporting on all pipes.
+ *
+ * The code also supports underrun detection on the PCH transcoder.
+ */
+
+static bool ivb_can_enable_err_int(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_crtc *crtc;
+	enum pipe pipe;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	for_each_pipe(dev_priv, pipe) {
+		crtc = intel_crtc_for_pipe(dev_priv, pipe);
+
+		if (crtc->cpu_fifo_underrun_disabled)
+			return false;
+	}
+
+	return true;
+}
+
+static bool cpt_can_enable_serr_int(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum pipe pipe;
+	struct intel_crtc *crtc;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	for_each_pipe(dev_priv, pipe) {
+		crtc = intel_crtc_for_pipe(dev_priv, pipe);
+
+		if (crtc->pch_fifo_underrun_disabled)
+			return false;
+	}
+
+	return true;
+}
+
+static void i9xx_check_fifo_underruns(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	i915_reg_t reg = PIPESTAT(crtc->pipe);
+	u32 enable_mask;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	if ((intel_de_read(dev_priv, reg) & PIPE_FIFO_UNDERRUN_STATUS) == 0)
+		return;
+
+	enable_mask = i915_pipestat_enable_mask(dev_priv, crtc->pipe);
+	intel_de_write(dev_priv, reg, enable_mask | PIPE_FIFO_UNDERRUN_STATUS);
+	intel_de_posting_read(dev_priv, reg);
+
+	trace_intel_cpu_fifo_underrun(dev_priv, crtc->pipe);
+	drm_err(&dev_priv->drm, "pipe %c underrun\n", pipe_name(crtc->pipe));
+}
+
+static void i9xx_set_fifo_underrun_reporting(struct drm_device *dev,
+					     enum pipe pipe,
+					     bool enable, bool old)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	i915_reg_t reg = PIPESTAT(pipe);
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	if (enable) {
+		u32 enable_mask = i915_pipestat_enable_mask(dev_priv, pipe);
+
+		intel_de_write(dev_priv, reg,
+			       enable_mask | PIPE_FIFO_UNDERRUN_STATUS);
+		intel_de_posting_read(dev_priv, reg);
+	} else {
+		if (old && intel_de_read(dev_priv, reg) & PIPE_FIFO_UNDERRUN_STATUS)
+			drm_err(&dev_priv->drm, "pipe %c underrun\n",
+				pipe_name(pipe));
+	}
+}
+
+static void ilk_set_fifo_underrun_reporting(struct drm_device *dev,
+					    enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 bit = (pipe == PIPE_A) ?
+		DE_PIPEA_FIFO_UNDERRUN : DE_PIPEB_FIFO_UNDERRUN;
+
+	if (enable)
+		ilk_enable_display_irq(dev_priv, bit);
+	else
+		ilk_disable_display_irq(dev_priv, bit);
+}
+
+static void ivb_check_fifo_underruns(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	u32 err_int = intel_de_read(dev_priv, GEN7_ERR_INT);
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	if ((err_int & ERR_INT_FIFO_UNDERRUN(pipe)) == 0)
+		return;
+
+	intel_de_write(dev_priv, GEN7_ERR_INT, ERR_INT_FIFO_UNDERRUN(pipe));
+	intel_de_posting_read(dev_priv, GEN7_ERR_INT);
+
+	trace_intel_cpu_fifo_underrun(dev_priv, pipe);
+	drm_err(&dev_priv->drm, "fifo underrun on pipe %c\n", pipe_name(pipe));
+}
+
+static void ivb_set_fifo_underrun_reporting(struct drm_device *dev,
+					    enum pipe pipe, bool enable,
+					    bool old)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	if (enable) {
+		intel_de_write(dev_priv, GEN7_ERR_INT,
+			       ERR_INT_FIFO_UNDERRUN(pipe));
+
+		if (!ivb_can_enable_err_int(dev))
+			return;
+
+		ilk_enable_display_irq(dev_priv, DE_ERR_INT_IVB);
+	} else {
+		ilk_disable_display_irq(dev_priv, DE_ERR_INT_IVB);
+
+		if (old &&
+		    intel_de_read(dev_priv, GEN7_ERR_INT) & ERR_INT_FIFO_UNDERRUN(pipe)) {
+			drm_err(&dev_priv->drm,
+				"uncleared fifo underrun on pipe %c\n",
+				pipe_name(pipe));
+		}
+	}
+}
+
+static u32
+icl_pipe_status_underrun_mask(struct drm_i915_private *dev_priv)
+{
+	u32 mask = PIPE_STATUS_UNDERRUN;
+
+	if (DISPLAY_VER(dev_priv) >= 13)
+		mask |= PIPE_STATUS_SOFT_UNDERRUN_XELPD |
+			PIPE_STATUS_HARD_UNDERRUN_XELPD |
+			PIPE_STATUS_PORT_UNDERRUN_XELPD;
+
+	return mask;
+}
+
+static void bdw_set_fifo_underrun_reporting(struct drm_device *dev,
+					    enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 mask = gen8_de_pipe_underrun_mask(dev_priv);
+
+	if (enable) {
+		if (DISPLAY_VER(dev_priv) >= 11)
+			intel_de_write(dev_priv, ICL_PIPESTATUS(pipe),
+				       icl_pipe_status_underrun_mask(dev_priv));
+
+		bdw_enable_pipe_irq(dev_priv, pipe, mask);
+	} else {
+		bdw_disable_pipe_irq(dev_priv, pipe, mask);
+	}
+}
+
+static void ibx_set_fifo_underrun_reporting(struct drm_device *dev,
+					    enum pipe pch_transcoder,
+					    bool enable)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 bit = (pch_transcoder == PIPE_A) ?
+		SDE_TRANSA_FIFO_UNDER : SDE_TRANSB_FIFO_UNDER;
+
+	if (enable)
+		ibx_enable_display_interrupt(dev_priv, bit);
+	else
+		ibx_disable_display_interrupt(dev_priv, bit);
+}
+
+static void cpt_check_pch_fifo_underruns(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pch_transcoder = crtc->pipe;
+	u32 serr_int = intel_de_read(dev_priv, SERR_INT);
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	if ((serr_int & SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder)) == 0)
+		return;
+
+	intel_de_write(dev_priv, SERR_INT,
+		       SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder));
+	intel_de_posting_read(dev_priv, SERR_INT);
+
+	trace_intel_pch_fifo_underrun(dev_priv, pch_transcoder);
+	drm_err(&dev_priv->drm, "pch fifo underrun on pch transcoder %c\n",
+		pipe_name(pch_transcoder));
+}
+
+static void cpt_set_fifo_underrun_reporting(struct drm_device *dev,
+					    enum pipe pch_transcoder,
+					    bool enable, bool old)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	if (enable) {
+		intel_de_write(dev_priv, SERR_INT,
+			       SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder));
+
+		if (!cpt_can_enable_serr_int(dev))
+			return;
+
+		ibx_enable_display_interrupt(dev_priv, SDE_ERROR_CPT);
+	} else {
+		ibx_disable_display_interrupt(dev_priv, SDE_ERROR_CPT);
+
+		if (old && intel_de_read(dev_priv, SERR_INT) &
+		    SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder)) {
+			drm_err(&dev_priv->drm,
+				"uncleared pch fifo underrun on pch transcoder %c\n",
+				pipe_name(pch_transcoder));
+		}
+	}
+}
+
+static bool __intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
+						    enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_crtc *crtc = intel_crtc_for_pipe(dev_priv, pipe);
+	bool old;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	old = !crtc->cpu_fifo_underrun_disabled;
+	crtc->cpu_fifo_underrun_disabled = !enable;
+
+	if (HAS_GMCH(dev_priv))
+		i9xx_set_fifo_underrun_reporting(dev, pipe, enable, old);
+	else if (IS_IRONLAKE(dev_priv) || IS_SANDYBRIDGE(dev_priv))
+		ilk_set_fifo_underrun_reporting(dev, pipe, enable);
+	else if (DISPLAY_VER(dev_priv) == 7)
+		ivb_set_fifo_underrun_reporting(dev, pipe, enable, old);
+	else if (DISPLAY_VER(dev_priv) >= 8)
+		bdw_set_fifo_underrun_reporting(dev, pipe, enable);
+
+	return old;
+}
+
+/**
+ * intel_set_cpu_fifo_underrun_reporting - set cpu fifo underrrun reporting state
+ * @dev_priv: i915 device instance
+ * @pipe: (CPU) pipe to set state for
+ * @enable: whether underruns should be reported or not
+ *
+ * This function sets the fifo underrun state for @pipe. It is used in the
+ * modeset code to avoid false positives since on many platforms underruns are
+ * expected when disabling or enabling the pipe.
+ *
+ * Notice that on some platforms disabling underrun reports for one pipe
+ * disables for all due to shared interrupts. Actual reporting is still per-pipe
+ * though.
+ *
+ * Returns the previous state of underrun reporting.
+ */
+bool intel_set_cpu_fifo_underrun_reporting(struct drm_i915_private *dev_priv,
+					   enum pipe pipe, bool enable)
+{
+	unsigned long flags;
+	bool ret;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+	ret = __intel_set_cpu_fifo_underrun_reporting(&dev_priv->drm, pipe,
+						      enable);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+
+	return ret;
+}
+
+/**
+ * intel_set_pch_fifo_underrun_reporting - set PCH fifo underrun reporting state
+ * @dev_priv: i915 device instance
+ * @pch_transcoder: the PCH transcoder (same as pipe on IVB and older)
+ * @enable: whether underruns should be reported or not
+ *
+ * This function makes us disable or enable PCH fifo underruns for a specific
+ * PCH transcoder. Notice that on some PCHs (e.g. CPT/PPT), disabling FIFO
+ * underrun reporting for one transcoder may also disable all the other PCH
+ * error interruts for the other transcoders, due to the fact that there's just
+ * one interrupt mask/enable bit for all the transcoders.
+ *
+ * Returns the previous state of underrun reporting.
+ */
+bool intel_set_pch_fifo_underrun_reporting(struct drm_i915_private *dev_priv,
+					   enum pipe pch_transcoder,
+					   bool enable)
+{
+	struct intel_crtc *crtc =
+		intel_crtc_for_pipe(dev_priv, pch_transcoder);
+	unsigned long flags;
+	bool old;
+
+	/*
+	 * NOTE: Pre-LPT has a fixed cpu pipe -> pch transcoder mapping, but LPT
+	 * has only one pch transcoder A that all pipes can use. To avoid racy
+	 * pch transcoder -> pipe lookups from interrupt code simply store the
+	 * underrun statistics in crtc A. Since we never expose this anywhere
+	 * nor use it outside of the fifo underrun code here using the "wrong"
+	 * crtc on LPT won't cause issues.
+	 */
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+
+	old = !crtc->pch_fifo_underrun_disabled;
+	crtc->pch_fifo_underrun_disabled = !enable;
+
+	if (HAS_PCH_IBX(dev_priv))
+		ibx_set_fifo_underrun_reporting(&dev_priv->drm,
+						pch_transcoder,
+						enable);
+	else
+		cpt_set_fifo_underrun_reporting(&dev_priv->drm,
+						pch_transcoder,
+						enable, old);
+
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	return old;
+}
+
+/**
+ * intel_cpu_fifo_underrun_irq_handler - handle CPU fifo underrun interrupt
+ * @dev_priv: i915 device instance
+ * @pipe: (CPU) pipe to set state for
+ *
+ * This handles a CPU fifo underrun interrupt, generating an underrun warning
+ * into dmesg if underrun reporting is enabled and then disables the underrun
+ * interrupt to avoid an irq storm.
+ */
+void intel_cpu_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
+					 enum pipe pipe)
+{
+	struct intel_crtc *crtc = intel_crtc_for_pipe(dev_priv, pipe);
+	u32 underruns = 0;
+
+	/* We may be called too early in init, thanks BIOS! */
+	if (crtc == NULL)
+		return;
+
+	/* GMCH can't disable fifo underruns, filter them. */
+	if (HAS_GMCH(dev_priv) &&
+	    crtc->cpu_fifo_underrun_disabled)
+		return;
+
+	/*
+	 * Starting with display version 11, the PIPE_STAT register records
+	 * whether an underrun has happened, and on XELPD+, it will also record
+	 * whether the underrun was soft/hard and whether it was triggered by
+	 * the downstream port logic.  We should clear these bits (which use
+	 * write-1-to-clear logic) too.
+	 *
+	 * Note that although the IIR gives us the same underrun and soft/hard
+	 * information, PIPE_STAT is the only place we can find out whether
+	 * the underrun was caused by the downstream port.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		underruns = intel_de_read(dev_priv, ICL_PIPESTATUS(pipe)) &
+			icl_pipe_status_underrun_mask(dev_priv);
+		intel_de_write(dev_priv, ICL_PIPESTATUS(pipe), underruns);
+	}
+
+	if (intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false)) {
+		trace_intel_cpu_fifo_underrun(dev_priv, pipe);
+
+		if (DISPLAY_VER(dev_priv) >= 11)
+			drm_err(&dev_priv->drm, "CPU pipe %c FIFO underrun: %s%s%s%s\n",
+				pipe_name(pipe),
+				underruns & PIPE_STATUS_SOFT_UNDERRUN_XELPD ? "soft," : "",
+				underruns & PIPE_STATUS_HARD_UNDERRUN_XELPD ? "hard," : "",
+				underruns & PIPE_STATUS_PORT_UNDERRUN_XELPD ? "port," : "",
+				underruns & PIPE_STATUS_UNDERRUN ? "transcoder," : "");
+		else
+			drm_err(&dev_priv->drm, "CPU pipe %c FIFO underrun\n", pipe_name(pipe));
+	}
+
+	intel_fbc_handle_fifo_underrun_irq(dev_priv);
+}
+
+/**
+ * intel_pch_fifo_underrun_irq_handler - handle PCH fifo underrun interrupt
+ * @dev_priv: i915 device instance
+ * @pch_transcoder: the PCH transcoder (same as pipe on IVB and older)
+ *
+ * This handles a PCH fifo underrun interrupt, generating an underrun warning
+ * into dmesg if underrun reporting is enabled and then disables the underrun
+ * interrupt to avoid an irq storm.
+ */
+void intel_pch_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
+					 enum pipe pch_transcoder)
+{
+	if (intel_set_pch_fifo_underrun_reporting(dev_priv, pch_transcoder,
+						  false)) {
+		trace_intel_pch_fifo_underrun(dev_priv, pch_transcoder);
+		drm_err(&dev_priv->drm, "PCH transcoder %c FIFO underrun\n",
+			pipe_name(pch_transcoder));
+	}
+}
+
+/**
+ * intel_check_cpu_fifo_underruns - check for CPU fifo underruns immediately
+ * @dev_priv: i915 device instance
+ *
+ * Check for CPU fifo underruns immediately. Useful on IVB/HSW where the shared
+ * error interrupt may have been disabled, and so CPU fifo underruns won't
+ * necessarily raise an interrupt, and on GMCH platforms where underruns never
+ * raise an interrupt.
+ */
+void intel_check_cpu_fifo_underruns(struct drm_i915_private *dev_priv)
+{
+	struct intel_crtc *crtc;
+
+	spin_lock_irq(&dev_priv->irq_lock);
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		if (crtc->cpu_fifo_underrun_disabled)
+			continue;
+
+		if (HAS_GMCH(dev_priv))
+			i9xx_check_fifo_underruns(crtc);
+		else if (DISPLAY_VER(dev_priv) == 7)
+			ivb_check_fifo_underruns(crtc);
+	}
+
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
+
+/**
+ * intel_check_pch_fifo_underruns - check for PCH fifo underruns immediately
+ * @dev_priv: i915 device instance
+ *
+ * Check for PCH fifo underruns immediately. Useful on CPT/PPT where the shared
+ * error interrupt may have been disabled, and so PCH fifo underruns won't
+ * necessarily raise an interrupt.
+ */
+void intel_check_pch_fifo_underruns(struct drm_i915_private *dev_priv)
+{
+	struct intel_crtc *crtc;
+
+	spin_lock_irq(&dev_priv->irq_lock);
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		if (crtc->pch_fifo_underrun_disabled)
+			continue;
+
+		if (HAS_PCH_CPT(dev_priv))
+			cpt_check_pch_fifo_underruns(crtc);
+	}
+
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_fifo_underrun.h b/drivers/gpu/drm/i915/display/intel_fifo_underrun.h
new file mode 100644
index 000000000..e04f22ac1
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_fifo_underrun.h
@@ -0,0 +1,27 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_FIFO_UNDERRUN_H__
+#define __INTEL_FIFO_UNDERRUN_H__
+
+#include <linux/types.h>
+
+#include "intel_display.h"
+
+struct drm_i915_private;
+
+bool intel_set_cpu_fifo_underrun_reporting(struct drm_i915_private *dev_priv,
+					   enum pipe pipe, bool enable);
+bool intel_set_pch_fifo_underrun_reporting(struct drm_i915_private *dev_priv,
+					   enum pipe pch_transcoder,
+					   bool enable);
+void intel_cpu_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
+					 enum pipe pipe);
+void intel_pch_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
+					 enum pipe pch_transcoder);
+void intel_check_cpu_fifo_underruns(struct drm_i915_private *dev_priv);
+void intel_check_pch_fifo_underruns(struct drm_i915_private *dev_priv);
+
+#endif /* __INTEL_FIFO_UNDERRUN_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_frontbuffer.c b/drivers/gpu/drm/i915/display/intel_frontbuffer.c
new file mode 100644
index 000000000..d80e3e8a9
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_frontbuffer.c
@@ -0,0 +1,328 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Daniel Vetter <daniel.vetter@ffwll.ch>
+ */
+
+/**
+ * DOC: frontbuffer tracking
+ *
+ * Many features require us to track changes to the currently active
+ * frontbuffer, especially rendering targeted at the frontbuffer.
+ *
+ * To be able to do so we track frontbuffers using a bitmask for all possible
+ * frontbuffer slots through intel_frontbuffer_track(). The functions in this
+ * file are then called when the contents of the frontbuffer are invalidated,
+ * when frontbuffer rendering has stopped again to flush out all the changes
+ * and when the frontbuffer is exchanged with a flip. Subsystems interested in
+ * frontbuffer changes (e.g. PSR, FBC, DRRS) should directly put their callbacks
+ * into the relevant places and filter for the frontbuffer slots that they are
+ * interested int.
+ *
+ * On a high level there are two types of powersaving features. The first one
+ * work like a special cache (FBC and PSR) and are interested when they should
+ * stop caching and when to restart caching. This is done by placing callbacks
+ * into the invalidate and the flush functions: At invalidate the caching must
+ * be stopped and at flush time it can be restarted. And maybe they need to know
+ * when the frontbuffer changes (e.g. when the hw doesn't initiate an invalidate
+ * and flush on its own) which can be achieved with placing callbacks into the
+ * flip functions.
+ *
+ * The other type of display power saving feature only cares about busyness
+ * (e.g. DRRS). In that case all three (invalidate, flush and flip) indicate
+ * busyness. There is no direct way to detect idleness. Instead an idle timer
+ * work delayed work should be started from the flush and flip functions and
+ * cancelled as soon as busyness is detected.
+ */
+
+#include "i915_drv.h"
+#include "intel_display_trace.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_drrs.h"
+#include "intel_fbc.h"
+#include "intel_frontbuffer.h"
+#include "intel_psr.h"
+
+/**
+ * frontbuffer_flush - flush frontbuffer
+ * @i915: i915 device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ * @origin: which operation caused the flush
+ *
+ * This function gets called every time rendering on the given planes has
+ * completed and frontbuffer caching can be started again. Flushes will get
+ * delayed if they're blocked by some outstanding asynchronous rendering.
+ *
+ * Can be called without any locks held.
+ */
+static void frontbuffer_flush(struct drm_i915_private *i915,
+			      unsigned int frontbuffer_bits,
+			      enum fb_op_origin origin)
+{
+	/* Delay flushing when rings are still busy.*/
+	spin_lock(&i915->display.fb_tracking.lock);
+	frontbuffer_bits &= ~i915->display.fb_tracking.busy_bits;
+	spin_unlock(&i915->display.fb_tracking.lock);
+
+	if (!frontbuffer_bits)
+		return;
+
+	trace_intel_frontbuffer_flush(frontbuffer_bits, origin);
+
+	might_sleep();
+	intel_drrs_flush(i915, frontbuffer_bits);
+	intel_psr_flush(i915, frontbuffer_bits, origin);
+	intel_fbc_flush(i915, frontbuffer_bits, origin);
+}
+
+/**
+ * intel_frontbuffer_flip_prepare - prepare asynchronous frontbuffer flip
+ * @i915: i915 device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called after scheduling a flip on @obj. The actual
+ * frontbuffer flushing will be delayed until completion is signalled with
+ * intel_frontbuffer_flip_complete. If an invalidate happens in between this
+ * flush will be cancelled.
+ *
+ * Can be called without any locks held.
+ */
+void intel_frontbuffer_flip_prepare(struct drm_i915_private *i915,
+				    unsigned frontbuffer_bits)
+{
+	spin_lock(&i915->display.fb_tracking.lock);
+	i915->display.fb_tracking.flip_bits |= frontbuffer_bits;
+	/* Remove stale busy bits due to the old buffer. */
+	i915->display.fb_tracking.busy_bits &= ~frontbuffer_bits;
+	spin_unlock(&i915->display.fb_tracking.lock);
+}
+
+/**
+ * intel_frontbuffer_flip_complete - complete asynchronous frontbuffer flip
+ * @i915: i915 device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called after the flip has been latched and will complete
+ * on the next vblank. It will execute the flush if it hasn't been cancelled yet.
+ *
+ * Can be called without any locks held.
+ */
+void intel_frontbuffer_flip_complete(struct drm_i915_private *i915,
+				     unsigned frontbuffer_bits)
+{
+	spin_lock(&i915->display.fb_tracking.lock);
+	/* Mask any cancelled flips. */
+	frontbuffer_bits &= i915->display.fb_tracking.flip_bits;
+	i915->display.fb_tracking.flip_bits &= ~frontbuffer_bits;
+	spin_unlock(&i915->display.fb_tracking.lock);
+
+	if (frontbuffer_bits)
+		frontbuffer_flush(i915, frontbuffer_bits, ORIGIN_FLIP);
+}
+
+/**
+ * intel_frontbuffer_flip - synchronous frontbuffer flip
+ * @i915: i915 device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ *
+ * This function gets called after scheduling a flip on @obj. This is for
+ * synchronous plane updates which will happen on the next vblank and which will
+ * not get delayed by pending gpu rendering.
+ *
+ * Can be called without any locks held.
+ */
+void intel_frontbuffer_flip(struct drm_i915_private *i915,
+			    unsigned frontbuffer_bits)
+{
+	spin_lock(&i915->display.fb_tracking.lock);
+	/* Remove stale busy bits due to the old buffer. */
+	i915->display.fb_tracking.busy_bits &= ~frontbuffer_bits;
+	spin_unlock(&i915->display.fb_tracking.lock);
+
+	frontbuffer_flush(i915, frontbuffer_bits, ORIGIN_FLIP);
+}
+
+void __intel_fb_invalidate(struct intel_frontbuffer *front,
+			   enum fb_op_origin origin,
+			   unsigned int frontbuffer_bits)
+{
+	struct drm_i915_private *i915 = to_i915(front->obj->base.dev);
+
+	if (origin == ORIGIN_CS) {
+		spin_lock(&i915->display.fb_tracking.lock);
+		i915->display.fb_tracking.busy_bits |= frontbuffer_bits;
+		i915->display.fb_tracking.flip_bits &= ~frontbuffer_bits;
+		spin_unlock(&i915->display.fb_tracking.lock);
+	}
+
+	trace_intel_frontbuffer_invalidate(frontbuffer_bits, origin);
+
+	might_sleep();
+	intel_psr_invalidate(i915, frontbuffer_bits, origin);
+	intel_drrs_invalidate(i915, frontbuffer_bits);
+	intel_fbc_invalidate(i915, frontbuffer_bits, origin);
+}
+
+void __intel_fb_flush(struct intel_frontbuffer *front,
+		      enum fb_op_origin origin,
+		      unsigned int frontbuffer_bits)
+{
+	struct drm_i915_private *i915 = to_i915(front->obj->base.dev);
+
+	if (origin == ORIGIN_CS) {
+		spin_lock(&i915->display.fb_tracking.lock);
+		/* Filter out new bits since rendering started. */
+		frontbuffer_bits &= i915->display.fb_tracking.busy_bits;
+		i915->display.fb_tracking.busy_bits &= ~frontbuffer_bits;
+		spin_unlock(&i915->display.fb_tracking.lock);
+	}
+
+	if (frontbuffer_bits)
+		frontbuffer_flush(i915, frontbuffer_bits, origin);
+}
+
+static int frontbuffer_active(struct i915_active *ref)
+{
+	struct intel_frontbuffer *front =
+		container_of(ref, typeof(*front), write);
+
+	kref_get(&front->ref);
+	return 0;
+}
+
+static void frontbuffer_retire(struct i915_active *ref)
+{
+	struct intel_frontbuffer *front =
+		container_of(ref, typeof(*front), write);
+
+	intel_frontbuffer_flush(front, ORIGIN_CS);
+	intel_frontbuffer_put(front);
+}
+
+static void frontbuffer_release(struct kref *ref)
+	__releases(&to_i915(front->obj->base.dev)->display.fb_tracking.lock)
+{
+	struct intel_frontbuffer *front =
+		container_of(ref, typeof(*front), ref);
+	struct drm_i915_gem_object *obj = front->obj;
+	struct i915_vma *vma;
+
+	drm_WARN_ON(obj->base.dev, atomic_read(&front->bits));
+
+	spin_lock(&obj->vma.lock);
+	for_each_ggtt_vma(vma, obj) {
+		i915_vma_clear_scanout(vma);
+		vma->display_alignment = I915_GTT_MIN_ALIGNMENT;
+	}
+	spin_unlock(&obj->vma.lock);
+
+	RCU_INIT_POINTER(obj->frontbuffer, NULL);
+	spin_unlock(&to_i915(obj->base.dev)->display.fb_tracking.lock);
+
+	i915_active_fini(&front->write);
+
+	i915_gem_object_put(obj);
+	kfree_rcu(front, rcu);
+}
+
+struct intel_frontbuffer *
+intel_frontbuffer_get(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct intel_frontbuffer *front;
+
+	front = __intel_frontbuffer_get(obj);
+	if (front)
+		return front;
+
+	front = kmalloc(sizeof(*front), GFP_KERNEL);
+	if (!front)
+		return NULL;
+
+	front->obj = obj;
+	kref_init(&front->ref);
+	atomic_set(&front->bits, 0);
+	i915_active_init(&front->write,
+			 frontbuffer_active,
+			 frontbuffer_retire,
+			 I915_ACTIVE_RETIRE_SLEEPS);
+
+	spin_lock(&i915->display.fb_tracking.lock);
+	if (rcu_access_pointer(obj->frontbuffer)) {
+		kfree(front);
+		front = rcu_dereference_protected(obj->frontbuffer, true);
+		kref_get(&front->ref);
+	} else {
+		i915_gem_object_get(obj);
+		rcu_assign_pointer(obj->frontbuffer, front);
+	}
+	spin_unlock(&i915->display.fb_tracking.lock);
+
+	return front;
+}
+
+void intel_frontbuffer_put(struct intel_frontbuffer *front)
+{
+	kref_put_lock(&front->ref,
+		      frontbuffer_release,
+		      &to_i915(front->obj->base.dev)->display.fb_tracking.lock);
+}
+
+/**
+ * intel_frontbuffer_track - update frontbuffer tracking
+ * @old: current buffer for the frontbuffer slots
+ * @new: new buffer for the frontbuffer slots
+ * @frontbuffer_bits: bitmask of frontbuffer slots
+ *
+ * This updates the frontbuffer tracking bits @frontbuffer_bits by clearing them
+ * from @old and setting them in @new. Both @old and @new can be NULL.
+ */
+void intel_frontbuffer_track(struct intel_frontbuffer *old,
+			     struct intel_frontbuffer *new,
+			     unsigned int frontbuffer_bits)
+{
+	/*
+	 * Control of individual bits within the mask are guarded by
+	 * the owning plane->mutex, i.e. we can never see concurrent
+	 * manipulation of individual bits. But since the bitfield as a whole
+	 * is updated using RMW, we need to use atomics in order to update
+	 * the bits.
+	 */
+	BUILD_BUG_ON(INTEL_FRONTBUFFER_BITS_PER_PIPE * I915_MAX_PIPES >
+		     BITS_PER_TYPE(atomic_t));
+	BUILD_BUG_ON(INTEL_FRONTBUFFER_BITS_PER_PIPE * I915_MAX_PIPES > 32);
+	BUILD_BUG_ON(I915_MAX_PLANES > INTEL_FRONTBUFFER_BITS_PER_PIPE);
+
+	if (old) {
+		drm_WARN_ON(old->obj->base.dev,
+			    !(atomic_read(&old->bits) & frontbuffer_bits));
+		atomic_andnot(frontbuffer_bits, &old->bits);
+	}
+
+	if (new) {
+		drm_WARN_ON(new->obj->base.dev,
+			    atomic_read(&new->bits) & frontbuffer_bits);
+		atomic_or(frontbuffer_bits, &new->bits);
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_frontbuffer.h b/drivers/gpu/drm/i915/display/intel_frontbuffer.h
new file mode 100644
index 000000000..3c474ed93
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_frontbuffer.h
@@ -0,0 +1,170 @@
+/*
+ * Copyright (c) 2014-2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __INTEL_FRONTBUFFER_H__
+#define __INTEL_FRONTBUFFER_H__
+
+#include <linux/atomic.h>
+#include <linux/bits.h>
+#include <linux/kref.h>
+
+#include "gem/i915_gem_object_types.h"
+#include "i915_active_types.h"
+
+struct drm_i915_private;
+
+enum fb_op_origin {
+	ORIGIN_CPU = 0,
+	ORIGIN_CS,
+	ORIGIN_FLIP,
+	ORIGIN_DIRTYFB,
+	ORIGIN_CURSOR_UPDATE,
+};
+
+struct intel_frontbuffer {
+	struct kref ref;
+	atomic_t bits;
+	struct i915_active write;
+	struct drm_i915_gem_object *obj;
+	struct rcu_head rcu;
+};
+
+/*
+ * Frontbuffer tracking bits. Set in obj->frontbuffer_bits while a gem bo is
+ * considered to be the frontbuffer for the given plane interface-wise. This
+ * doesn't mean that the hw necessarily already scans it out, but that any
+ * rendering (by the cpu or gpu) will land in the frontbuffer eventually.
+ *
+ * We have one bit per pipe and per scanout plane type.
+ */
+#define INTEL_FRONTBUFFER_BITS_PER_PIPE 8
+#define INTEL_FRONTBUFFER(pipe, plane_id) \
+	BIT((plane_id) + INTEL_FRONTBUFFER_BITS_PER_PIPE * (pipe));
+#define INTEL_FRONTBUFFER_OVERLAY(pipe) \
+	BIT(INTEL_FRONTBUFFER_BITS_PER_PIPE - 1 + INTEL_FRONTBUFFER_BITS_PER_PIPE * (pipe))
+#define INTEL_FRONTBUFFER_ALL_MASK(pipe) \
+	GENMASK(INTEL_FRONTBUFFER_BITS_PER_PIPE * ((pipe) + 1) - 1,	\
+		INTEL_FRONTBUFFER_BITS_PER_PIPE * (pipe))
+
+void intel_frontbuffer_flip_prepare(struct drm_i915_private *i915,
+				    unsigned frontbuffer_bits);
+void intel_frontbuffer_flip_complete(struct drm_i915_private *i915,
+				     unsigned frontbuffer_bits);
+void intel_frontbuffer_flip(struct drm_i915_private *i915,
+			    unsigned frontbuffer_bits);
+
+void intel_frontbuffer_put(struct intel_frontbuffer *front);
+
+static inline struct intel_frontbuffer *
+__intel_frontbuffer_get(const struct drm_i915_gem_object *obj)
+{
+	struct intel_frontbuffer *front;
+
+	if (likely(!rcu_access_pointer(obj->frontbuffer)))
+		return NULL;
+
+	rcu_read_lock();
+	do {
+		front = rcu_dereference(obj->frontbuffer);
+		if (!front)
+			break;
+
+		if (unlikely(!kref_get_unless_zero(&front->ref)))
+			continue;
+
+		if (likely(front == rcu_access_pointer(obj->frontbuffer)))
+			break;
+
+		intel_frontbuffer_put(front);
+	} while (1);
+	rcu_read_unlock();
+
+	return front;
+}
+
+struct intel_frontbuffer *
+intel_frontbuffer_get(struct drm_i915_gem_object *obj);
+
+void __intel_fb_invalidate(struct intel_frontbuffer *front,
+			   enum fb_op_origin origin,
+			   unsigned int frontbuffer_bits);
+
+/**
+ * intel_frontbuffer_invalidate - invalidate frontbuffer object
+ * @front: GEM object to invalidate
+ * @origin: which operation caused the invalidation
+ *
+ * This function gets called every time rendering on the given object starts and
+ * frontbuffer caching (fbc, low refresh rate for DRRS, panel self refresh) must
+ * be invalidated. For ORIGIN_CS any subsequent invalidation will be delayed
+ * until the rendering completes or a flip on this frontbuffer plane is
+ * scheduled.
+ */
+static inline bool intel_frontbuffer_invalidate(struct intel_frontbuffer *front,
+						enum fb_op_origin origin)
+{
+	unsigned int frontbuffer_bits;
+
+	if (!front)
+		return false;
+
+	frontbuffer_bits = atomic_read(&front->bits);
+	if (!frontbuffer_bits)
+		return false;
+
+	__intel_fb_invalidate(front, origin, frontbuffer_bits);
+	return true;
+}
+
+void __intel_fb_flush(struct intel_frontbuffer *front,
+		      enum fb_op_origin origin,
+		      unsigned int frontbuffer_bits);
+
+/**
+ * intel_frontbuffer_flush - flush frontbuffer object
+ * @front: GEM object to flush
+ * @origin: which operation caused the flush
+ *
+ * This function gets called every time rendering on the given object has
+ * completed and frontbuffer caching can be started again.
+ */
+static inline void intel_frontbuffer_flush(struct intel_frontbuffer *front,
+					   enum fb_op_origin origin)
+{
+	unsigned int frontbuffer_bits;
+
+	if (!front)
+		return;
+
+	frontbuffer_bits = atomic_read(&front->bits);
+	if (!frontbuffer_bits)
+		return;
+
+	__intel_fb_flush(front, origin, frontbuffer_bits);
+}
+
+void intel_frontbuffer_track(struct intel_frontbuffer *old,
+			     struct intel_frontbuffer *new,
+			     unsigned int frontbuffer_bits);
+
+#endif /* __INTEL_FRONTBUFFER_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_global_state.c b/drivers/gpu/drm/i915/display/intel_global_state.c
new file mode 100644
index 000000000..7a19215ad
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_global_state.c
@@ -0,0 +1,257 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/string.h>
+
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_display_types.h"
+#include "intel_global_state.h"
+
+static void __intel_atomic_global_state_free(struct kref *kref)
+{
+	struct intel_global_state *obj_state =
+		container_of(kref, struct intel_global_state, ref);
+	struct intel_global_obj *obj = obj_state->obj;
+
+	obj->funcs->atomic_destroy_state(obj, obj_state);
+}
+
+static void intel_atomic_global_state_put(struct intel_global_state *obj_state)
+{
+	kref_put(&obj_state->ref, __intel_atomic_global_state_free);
+}
+
+static struct intel_global_state *
+intel_atomic_global_state_get(struct intel_global_state *obj_state)
+{
+	kref_get(&obj_state->ref);
+
+	return obj_state;
+}
+
+void intel_atomic_global_obj_init(struct drm_i915_private *dev_priv,
+				  struct intel_global_obj *obj,
+				  struct intel_global_state *state,
+				  const struct intel_global_state_funcs *funcs)
+{
+	memset(obj, 0, sizeof(*obj));
+
+	state->obj = obj;
+
+	kref_init(&state->ref);
+
+	obj->state = state;
+	obj->funcs = funcs;
+	list_add_tail(&obj->head, &dev_priv->global_obj_list);
+}
+
+void intel_atomic_global_obj_cleanup(struct drm_i915_private *dev_priv)
+{
+	struct intel_global_obj *obj, *next;
+
+	list_for_each_entry_safe(obj, next, &dev_priv->global_obj_list, head) {
+		list_del(&obj->head);
+
+		drm_WARN_ON(&dev_priv->drm, kref_read(&obj->state->ref) != 1);
+		intel_atomic_global_state_put(obj->state);
+	}
+}
+
+static void assert_global_state_write_locked(struct drm_i915_private *dev_priv)
+{
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc)
+		drm_modeset_lock_assert_held(&crtc->base.mutex);
+}
+
+static bool modeset_lock_is_held(struct drm_modeset_acquire_ctx *ctx,
+				 struct drm_modeset_lock *lock)
+{
+	struct drm_modeset_lock *l;
+
+	list_for_each_entry(l, &ctx->locked, head) {
+		if (lock == l)
+			return true;
+	}
+
+	return false;
+}
+
+static void assert_global_state_read_locked(struct intel_atomic_state *state)
+{
+	struct drm_modeset_acquire_ctx *ctx = state->base.acquire_ctx;
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		if (modeset_lock_is_held(ctx, &crtc->base.mutex))
+			return;
+	}
+
+	drm_WARN(&dev_priv->drm, 1, "Global state not read locked\n");
+}
+
+struct intel_global_state *
+intel_atomic_get_global_obj_state(struct intel_atomic_state *state,
+				  struct intel_global_obj *obj)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	int index, num_objs, i;
+	size_t size;
+	struct __intel_global_objs_state *arr;
+	struct intel_global_state *obj_state;
+
+	for (i = 0; i < state->num_global_objs; i++)
+		if (obj == state->global_objs[i].ptr)
+			return state->global_objs[i].state;
+
+	assert_global_state_read_locked(state);
+
+	num_objs = state->num_global_objs + 1;
+	size = sizeof(*state->global_objs) * num_objs;
+	arr = krealloc(state->global_objs, size, GFP_KERNEL);
+	if (!arr)
+		return ERR_PTR(-ENOMEM);
+
+	state->global_objs = arr;
+	index = state->num_global_objs;
+	memset(&state->global_objs[index], 0, sizeof(*state->global_objs));
+
+	obj_state = obj->funcs->atomic_duplicate_state(obj);
+	if (!obj_state)
+		return ERR_PTR(-ENOMEM);
+
+	obj_state->obj = obj;
+	obj_state->changed = false;
+
+	kref_init(&obj_state->ref);
+
+	state->global_objs[index].state = obj_state;
+	state->global_objs[index].old_state =
+		intel_atomic_global_state_get(obj->state);
+	state->global_objs[index].new_state = obj_state;
+	state->global_objs[index].ptr = obj;
+	obj_state->state = state;
+
+	state->num_global_objs = num_objs;
+
+	drm_dbg_atomic(&i915->drm, "Added new global object %p state %p to %p\n",
+		       obj, obj_state, state);
+
+	return obj_state;
+}
+
+struct intel_global_state *
+intel_atomic_get_old_global_obj_state(struct intel_atomic_state *state,
+				      struct intel_global_obj *obj)
+{
+	int i;
+
+	for (i = 0; i < state->num_global_objs; i++)
+		if (obj == state->global_objs[i].ptr)
+			return state->global_objs[i].old_state;
+
+	return NULL;
+}
+
+struct intel_global_state *
+intel_atomic_get_new_global_obj_state(struct intel_atomic_state *state,
+				      struct intel_global_obj *obj)
+{
+	int i;
+
+	for (i = 0; i < state->num_global_objs; i++)
+		if (obj == state->global_objs[i].ptr)
+			return state->global_objs[i].new_state;
+
+	return NULL;
+}
+
+void intel_atomic_swap_global_state(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_global_state *old_obj_state, *new_obj_state;
+	struct intel_global_obj *obj;
+	int i;
+
+	for_each_oldnew_global_obj_in_state(state, obj, old_obj_state,
+					    new_obj_state, i) {
+		drm_WARN_ON(&dev_priv->drm, obj->state != old_obj_state);
+
+		/*
+		 * If the new state wasn't modified (and properly
+		 * locked for write access) we throw it away.
+		 */
+		if (!new_obj_state->changed)
+			continue;
+
+		assert_global_state_write_locked(dev_priv);
+
+		old_obj_state->state = state;
+		new_obj_state->state = NULL;
+
+		state->global_objs[i].state = old_obj_state;
+
+		intel_atomic_global_state_put(obj->state);
+		obj->state = intel_atomic_global_state_get(new_obj_state);
+	}
+}
+
+void intel_atomic_clear_global_state(struct intel_atomic_state *state)
+{
+	int i;
+
+	for (i = 0; i < state->num_global_objs; i++) {
+		intel_atomic_global_state_put(state->global_objs[i].old_state);
+		intel_atomic_global_state_put(state->global_objs[i].new_state);
+
+		state->global_objs[i].ptr = NULL;
+		state->global_objs[i].state = NULL;
+		state->global_objs[i].old_state = NULL;
+		state->global_objs[i].new_state = NULL;
+	}
+	state->num_global_objs = 0;
+}
+
+int intel_atomic_lock_global_state(struct intel_global_state *obj_state)
+{
+	struct intel_atomic_state *state = obj_state->state;
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		int ret;
+
+		ret = drm_modeset_lock(&crtc->base.mutex,
+				       state->base.acquire_ctx);
+		if (ret)
+			return ret;
+	}
+
+	obj_state->changed = true;
+
+	return 0;
+}
+
+int intel_atomic_serialize_global_state(struct intel_global_state *obj_state)
+{
+	struct intel_atomic_state *state = obj_state->state;
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		struct intel_crtc_state *crtc_state;
+
+		crtc_state = intel_atomic_get_crtc_state(&state->base, crtc);
+		if (IS_ERR(crtc_state))
+			return PTR_ERR(crtc_state);
+	}
+
+	obj_state->changed = true;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_global_state.h b/drivers/gpu/drm/i915/display/intel_global_state.h
new file mode 100644
index 000000000..1f16fa307
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_global_state.h
@@ -0,0 +1,90 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __INTEL_GLOBAL_STATE_H__
+#define __INTEL_GLOBAL_STATE_H__
+
+#include <linux/kref.h>
+#include <linux/list.h>
+
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_global_obj;
+struct intel_global_state;
+
+struct intel_global_state_funcs {
+	struct intel_global_state *(*atomic_duplicate_state)(struct intel_global_obj *obj);
+	void (*atomic_destroy_state)(struct intel_global_obj *obj,
+				     struct intel_global_state *state);
+};
+
+struct intel_global_obj {
+	struct list_head head;
+	struct intel_global_state *state;
+	const struct intel_global_state_funcs *funcs;
+};
+
+#define intel_for_each_global_obj(obj, dev_priv) \
+	list_for_each_entry(obj, &(dev_priv)->global_obj_list, head)
+
+#define for_each_new_global_obj_in_state(__state, obj, new_obj_state, __i) \
+	for ((__i) = 0; \
+	     (__i) < (__state)->num_global_objs && \
+		     ((obj) = (__state)->global_objs[__i].ptr, \
+		      (new_obj_state) = (__state)->global_objs[__i].new_state, 1); \
+	     (__i)++) \
+		for_each_if(obj)
+
+#define for_each_old_global_obj_in_state(__state, obj, new_obj_state, __i) \
+	for ((__i) = 0; \
+	     (__i) < (__state)->num_global_objs && \
+		     ((obj) = (__state)->global_objs[__i].ptr, \
+		      (new_obj_state) = (__state)->global_objs[__i].old_state, 1); \
+	     (__i)++) \
+		for_each_if(obj)
+
+#define for_each_oldnew_global_obj_in_state(__state, obj, old_obj_state, new_obj_state, __i) \
+	for ((__i) = 0; \
+	     (__i) < (__state)->num_global_objs && \
+		     ((obj) = (__state)->global_objs[__i].ptr, \
+		      (old_obj_state) = (__state)->global_objs[__i].old_state, \
+		      (new_obj_state) = (__state)->global_objs[__i].new_state, 1); \
+	     (__i)++) \
+		for_each_if(obj)
+
+struct intel_global_state {
+	struct intel_global_obj *obj;
+	struct intel_atomic_state *state;
+	struct kref ref;
+	bool changed;
+};
+
+struct __intel_global_objs_state {
+	struct intel_global_obj *ptr;
+	struct intel_global_state *state, *old_state, *new_state;
+};
+
+void intel_atomic_global_obj_init(struct drm_i915_private *dev_priv,
+				  struct intel_global_obj *obj,
+				  struct intel_global_state *state,
+				  const struct intel_global_state_funcs *funcs);
+void intel_atomic_global_obj_cleanup(struct drm_i915_private *dev_priv);
+
+struct intel_global_state *
+intel_atomic_get_global_obj_state(struct intel_atomic_state *state,
+				  struct intel_global_obj *obj);
+struct intel_global_state *
+intel_atomic_get_old_global_obj_state(struct intel_atomic_state *state,
+				      struct intel_global_obj *obj);
+struct intel_global_state *
+intel_atomic_get_new_global_obj_state(struct intel_atomic_state *state,
+				      struct intel_global_obj *obj);
+
+void intel_atomic_swap_global_state(struct intel_atomic_state *state);
+void intel_atomic_clear_global_state(struct intel_atomic_state *state);
+int intel_atomic_lock_global_state(struct intel_global_state *obj_state);
+int intel_atomic_serialize_global_state(struct intel_global_state *obj_state);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_gmbus.c b/drivers/gpu/drm/i915/display/intel_gmbus.c
new file mode 100644
index 000000000..74443f57f
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_gmbus.c
@@ -0,0 +1,1003 @@
+/*
+ * Copyright (c) 2006 Dave Airlie <airlied@linux.ie>
+ * Copyright © 2006-2008,2010 Intel Corporation
+ *   Jesse Barnes <jesse.barnes@intel.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Eric Anholt <eric@anholt.net>
+ *	Chris Wilson <chris@chris-wilson.co.uk>
+ */
+
+#include <linux/export.h>
+#include <linux/i2c-algo-bit.h>
+#include <linux/i2c.h>
+
+#include <drm/display/drm_hdcp_helper.h>
+
+#include "i915_drv.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_gmbus.h"
+#include "intel_gmbus_regs.h"
+
+struct intel_gmbus {
+	struct i2c_adapter adapter;
+#define GMBUS_FORCE_BIT_RETRY (1U << 31)
+	u32 force_bit;
+	u32 reg0;
+	i915_reg_t gpio_reg;
+	struct i2c_algo_bit_data bit_algo;
+	struct drm_i915_private *i915;
+};
+
+struct gmbus_pin {
+	const char *name;
+	enum i915_gpio gpio;
+};
+
+/* Map gmbus pin pairs to names and registers. */
+static const struct gmbus_pin gmbus_pins[] = {
+	[GMBUS_PIN_SSC] = { "ssc", GPIOB },
+	[GMBUS_PIN_VGADDC] = { "vga", GPIOA },
+	[GMBUS_PIN_PANEL] = { "panel", GPIOC },
+	[GMBUS_PIN_DPC] = { "dpc", GPIOD },
+	[GMBUS_PIN_DPB] = { "dpb", GPIOE },
+	[GMBUS_PIN_DPD] = { "dpd", GPIOF },
+};
+
+static const struct gmbus_pin gmbus_pins_bdw[] = {
+	[GMBUS_PIN_VGADDC] = { "vga", GPIOA },
+	[GMBUS_PIN_DPC] = { "dpc", GPIOD },
+	[GMBUS_PIN_DPB] = { "dpb", GPIOE },
+	[GMBUS_PIN_DPD] = { "dpd", GPIOF },
+};
+
+static const struct gmbus_pin gmbus_pins_skl[] = {
+	[GMBUS_PIN_DPC] = { "dpc", GPIOD },
+	[GMBUS_PIN_DPB] = { "dpb", GPIOE },
+	[GMBUS_PIN_DPD] = { "dpd", GPIOF },
+};
+
+static const struct gmbus_pin gmbus_pins_bxt[] = {
+	[GMBUS_PIN_1_BXT] = { "dpb", GPIOB },
+	[GMBUS_PIN_2_BXT] = { "dpc", GPIOC },
+	[GMBUS_PIN_3_BXT] = { "misc", GPIOD },
+};
+
+static const struct gmbus_pin gmbus_pins_cnp[] = {
+	[GMBUS_PIN_1_BXT] = { "dpb", GPIOB },
+	[GMBUS_PIN_2_BXT] = { "dpc", GPIOC },
+	[GMBUS_PIN_3_BXT] = { "misc", GPIOD },
+	[GMBUS_PIN_4_CNP] = { "dpd", GPIOE },
+};
+
+static const struct gmbus_pin gmbus_pins_icp[] = {
+	[GMBUS_PIN_1_BXT] = { "dpa", GPIOB },
+	[GMBUS_PIN_2_BXT] = { "dpb", GPIOC },
+	[GMBUS_PIN_3_BXT] = { "dpc", GPIOD },
+	[GMBUS_PIN_9_TC1_ICP] = { "tc1", GPIOJ },
+	[GMBUS_PIN_10_TC2_ICP] = { "tc2", GPIOK },
+	[GMBUS_PIN_11_TC3_ICP] = { "tc3", GPIOL },
+	[GMBUS_PIN_12_TC4_ICP] = { "tc4", GPIOM },
+	[GMBUS_PIN_13_TC5_TGP] = { "tc5", GPION },
+	[GMBUS_PIN_14_TC6_TGP] = { "tc6", GPIOO },
+};
+
+static const struct gmbus_pin gmbus_pins_dg1[] = {
+	[GMBUS_PIN_1_BXT] = { "dpa", GPIOB },
+	[GMBUS_PIN_2_BXT] = { "dpb", GPIOC },
+	[GMBUS_PIN_3_BXT] = { "dpc", GPIOD },
+	[GMBUS_PIN_4_CNP] = { "dpd", GPIOE },
+};
+
+static const struct gmbus_pin gmbus_pins_dg2[] = {
+	[GMBUS_PIN_1_BXT] = { "dpa", GPIOB },
+	[GMBUS_PIN_2_BXT] = { "dpb", GPIOC },
+	[GMBUS_PIN_3_BXT] = { "dpc", GPIOD },
+	[GMBUS_PIN_4_CNP] = { "dpd", GPIOE },
+	[GMBUS_PIN_9_TC1_ICP] = { "tc1", GPIOJ },
+};
+
+static const struct gmbus_pin gmbus_pins_mtp[] = {
+	[GMBUS_PIN_1_BXT] = { "dpa", GPIOB },
+	[GMBUS_PIN_2_BXT] = { "dpb", GPIOC },
+	[GMBUS_PIN_3_BXT] = { "dpc", GPIOD },
+	[GMBUS_PIN_4_CNP] = { "dpd", GPIOE },
+	[GMBUS_PIN_5_MTP] = { "dpe", GPIOF },
+	[GMBUS_PIN_9_TC1_ICP] = { "tc1", GPIOJ },
+	[GMBUS_PIN_10_TC2_ICP] = { "tc2", GPIOK },
+	[GMBUS_PIN_11_TC3_ICP] = { "tc3", GPIOL },
+	[GMBUS_PIN_12_TC4_ICP] = { "tc4", GPIOM },
+};
+
+static const struct gmbus_pin *get_gmbus_pin(struct drm_i915_private *i915,
+					     unsigned int pin)
+{
+	const struct gmbus_pin *pins;
+	size_t size;
+
+	if (INTEL_PCH_TYPE(i915) >= PCH_DG2) {
+		pins = gmbus_pins_dg2;
+		size = ARRAY_SIZE(gmbus_pins_dg2);
+	} else if (INTEL_PCH_TYPE(i915) >= PCH_DG1) {
+		pins = gmbus_pins_dg1;
+		size = ARRAY_SIZE(gmbus_pins_dg1);
+	} else if (INTEL_PCH_TYPE(i915) >= PCH_MTP) {
+		pins = gmbus_pins_mtp;
+		size = ARRAY_SIZE(gmbus_pins_mtp);
+	} else if (INTEL_PCH_TYPE(i915) >= PCH_ICP) {
+		pins = gmbus_pins_icp;
+		size = ARRAY_SIZE(gmbus_pins_icp);
+	} else if (HAS_PCH_CNP(i915)) {
+		pins = gmbus_pins_cnp;
+		size = ARRAY_SIZE(gmbus_pins_cnp);
+	} else if (IS_GEMINILAKE(i915) || IS_BROXTON(i915)) {
+		pins = gmbus_pins_bxt;
+		size = ARRAY_SIZE(gmbus_pins_bxt);
+	} else if (DISPLAY_VER(i915) == 9) {
+		pins = gmbus_pins_skl;
+		size = ARRAY_SIZE(gmbus_pins_skl);
+	} else if (IS_BROADWELL(i915)) {
+		pins = gmbus_pins_bdw;
+		size = ARRAY_SIZE(gmbus_pins_bdw);
+	} else {
+		pins = gmbus_pins;
+		size = ARRAY_SIZE(gmbus_pins);
+	}
+
+	if (pin >= size || !pins[pin].name)
+		return NULL;
+
+	return &pins[pin];
+}
+
+bool intel_gmbus_is_valid_pin(struct drm_i915_private *i915, unsigned int pin)
+{
+	return get_gmbus_pin(i915, pin);
+}
+
+/* Intel GPIO access functions */
+
+#define I2C_RISEFALL_TIME 10
+
+static inline struct intel_gmbus *
+to_intel_gmbus(struct i2c_adapter *i2c)
+{
+	return container_of(i2c, struct intel_gmbus, adapter);
+}
+
+void
+intel_gmbus_reset(struct drm_i915_private *i915)
+{
+	intel_de_write(i915, GMBUS0(i915), 0);
+	intel_de_write(i915, GMBUS4(i915), 0);
+}
+
+static void pnv_gmbus_clock_gating(struct drm_i915_private *i915,
+				   bool enable)
+{
+	u32 val;
+
+	/* When using bit bashing for I2C, this bit needs to be set to 1 */
+	val = intel_de_read(i915, DSPCLK_GATE_D(i915));
+	if (!enable)
+		val |= PNV_GMBUSUNIT_CLOCK_GATE_DISABLE;
+	else
+		val &= ~PNV_GMBUSUNIT_CLOCK_GATE_DISABLE;
+	intel_de_write(i915, DSPCLK_GATE_D(i915), val);
+}
+
+static void pch_gmbus_clock_gating(struct drm_i915_private *i915,
+				   bool enable)
+{
+	u32 val;
+
+	val = intel_de_read(i915, SOUTH_DSPCLK_GATE_D);
+	if (!enable)
+		val |= PCH_GMBUSUNIT_CLOCK_GATE_DISABLE;
+	else
+		val &= ~PCH_GMBUSUNIT_CLOCK_GATE_DISABLE;
+	intel_de_write(i915, SOUTH_DSPCLK_GATE_D, val);
+}
+
+static void bxt_gmbus_clock_gating(struct drm_i915_private *i915,
+				   bool enable)
+{
+	u32 val;
+
+	val = intel_de_read(i915, GEN9_CLKGATE_DIS_4);
+	if (!enable)
+		val |= BXT_GMBUS_GATING_DIS;
+	else
+		val &= ~BXT_GMBUS_GATING_DIS;
+	intel_de_write(i915, GEN9_CLKGATE_DIS_4, val);
+}
+
+static u32 get_reserved(struct intel_gmbus *bus)
+{
+	struct drm_i915_private *i915 = bus->i915;
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 reserved = 0;
+
+	/* On most chips, these bits must be preserved in software. */
+	if (!IS_I830(i915) && !IS_I845G(i915))
+		reserved = intel_uncore_read_notrace(uncore, bus->gpio_reg) &
+			   (GPIO_DATA_PULLUP_DISABLE |
+			    GPIO_CLOCK_PULLUP_DISABLE);
+
+	return reserved;
+}
+
+static int get_clock(void *data)
+{
+	struct intel_gmbus *bus = data;
+	struct intel_uncore *uncore = &bus->i915->uncore;
+	u32 reserved = get_reserved(bus);
+
+	intel_uncore_write_notrace(uncore,
+				   bus->gpio_reg,
+				   reserved | GPIO_CLOCK_DIR_MASK);
+	intel_uncore_write_notrace(uncore, bus->gpio_reg, reserved);
+
+	return (intel_uncore_read_notrace(uncore, bus->gpio_reg) &
+		GPIO_CLOCK_VAL_IN) != 0;
+}
+
+static int get_data(void *data)
+{
+	struct intel_gmbus *bus = data;
+	struct intel_uncore *uncore = &bus->i915->uncore;
+	u32 reserved = get_reserved(bus);
+
+	intel_uncore_write_notrace(uncore,
+				   bus->gpio_reg,
+				   reserved | GPIO_DATA_DIR_MASK);
+	intel_uncore_write_notrace(uncore, bus->gpio_reg, reserved);
+
+	return (intel_uncore_read_notrace(uncore, bus->gpio_reg) &
+		GPIO_DATA_VAL_IN) != 0;
+}
+
+static void set_clock(void *data, int state_high)
+{
+	struct intel_gmbus *bus = data;
+	struct intel_uncore *uncore = &bus->i915->uncore;
+	u32 reserved = get_reserved(bus);
+	u32 clock_bits;
+
+	if (state_high)
+		clock_bits = GPIO_CLOCK_DIR_IN | GPIO_CLOCK_DIR_MASK;
+	else
+		clock_bits = GPIO_CLOCK_DIR_OUT | GPIO_CLOCK_DIR_MASK |
+			     GPIO_CLOCK_VAL_MASK;
+
+	intel_uncore_write_notrace(uncore,
+				   bus->gpio_reg,
+				   reserved | clock_bits);
+	intel_uncore_posting_read(uncore, bus->gpio_reg);
+}
+
+static void set_data(void *data, int state_high)
+{
+	struct intel_gmbus *bus = data;
+	struct intel_uncore *uncore = &bus->i915->uncore;
+	u32 reserved = get_reserved(bus);
+	u32 data_bits;
+
+	if (state_high)
+		data_bits = GPIO_DATA_DIR_IN | GPIO_DATA_DIR_MASK;
+	else
+		data_bits = GPIO_DATA_DIR_OUT | GPIO_DATA_DIR_MASK |
+			GPIO_DATA_VAL_MASK;
+
+	intel_uncore_write_notrace(uncore, bus->gpio_reg, reserved | data_bits);
+	intel_uncore_posting_read(uncore, bus->gpio_reg);
+}
+
+static int
+intel_gpio_pre_xfer(struct i2c_adapter *adapter)
+{
+	struct intel_gmbus *bus = to_intel_gmbus(adapter);
+	struct drm_i915_private *i915 = bus->i915;
+
+	intel_gmbus_reset(i915);
+
+	if (IS_PINEVIEW(i915))
+		pnv_gmbus_clock_gating(i915, false);
+
+	set_data(bus, 1);
+	set_clock(bus, 1);
+	udelay(I2C_RISEFALL_TIME);
+	return 0;
+}
+
+static void
+intel_gpio_post_xfer(struct i2c_adapter *adapter)
+{
+	struct intel_gmbus *bus = to_intel_gmbus(adapter);
+	struct drm_i915_private *i915 = bus->i915;
+
+	set_data(bus, 1);
+	set_clock(bus, 1);
+
+	if (IS_PINEVIEW(i915))
+		pnv_gmbus_clock_gating(i915, true);
+}
+
+static void
+intel_gpio_setup(struct intel_gmbus *bus, i915_reg_t gpio_reg)
+{
+	struct i2c_algo_bit_data *algo;
+
+	algo = &bus->bit_algo;
+
+	bus->gpio_reg = gpio_reg;
+	bus->adapter.algo_data = algo;
+	algo->setsda = set_data;
+	algo->setscl = set_clock;
+	algo->getsda = get_data;
+	algo->getscl = get_clock;
+	algo->pre_xfer = intel_gpio_pre_xfer;
+	algo->post_xfer = intel_gpio_post_xfer;
+	algo->udelay = I2C_RISEFALL_TIME;
+	algo->timeout = usecs_to_jiffies(2200);
+	algo->data = bus;
+}
+
+static bool has_gmbus_irq(struct drm_i915_private *i915)
+{
+	/*
+	 * encoder->shutdown() may want to use GMBUS
+	 * after irqs have already been disabled.
+	 */
+	return HAS_GMBUS_IRQ(i915) && intel_irqs_enabled(i915);
+}
+
+static int gmbus_wait(struct drm_i915_private *i915, u32 status, u32 irq_en)
+{
+	DEFINE_WAIT(wait);
+	u32 gmbus2;
+	int ret;
+
+	/* Important: The hw handles only the first bit, so set only one! Since
+	 * we also need to check for NAKs besides the hw ready/idle signal, we
+	 * need to wake up periodically and check that ourselves.
+	 */
+	if (!has_gmbus_irq(i915))
+		irq_en = 0;
+
+	add_wait_queue(&i915->display.gmbus.wait_queue, &wait);
+	intel_de_write_fw(i915, GMBUS4(i915), irq_en);
+
+	status |= GMBUS_SATOER;
+	ret = wait_for_us((gmbus2 = intel_de_read_fw(i915, GMBUS2(i915))) & status,
+			  2);
+	if (ret)
+		ret = wait_for((gmbus2 = intel_de_read_fw(i915, GMBUS2(i915))) & status,
+			       50);
+
+	intel_de_write_fw(i915, GMBUS4(i915), 0);
+	remove_wait_queue(&i915->display.gmbus.wait_queue, &wait);
+
+	if (gmbus2 & GMBUS_SATOER)
+		return -ENXIO;
+
+	return ret;
+}
+
+static int
+gmbus_wait_idle(struct drm_i915_private *i915)
+{
+	DEFINE_WAIT(wait);
+	u32 irq_enable;
+	int ret;
+
+	/* Important: The hw handles only the first bit, so set only one! */
+	irq_enable = 0;
+	if (has_gmbus_irq(i915))
+		irq_enable = GMBUS_IDLE_EN;
+
+	add_wait_queue(&i915->display.gmbus.wait_queue, &wait);
+	intel_de_write_fw(i915, GMBUS4(i915), irq_enable);
+
+	ret = intel_wait_for_register_fw(&i915->uncore,
+					 GMBUS2(i915), GMBUS_ACTIVE, 0,
+					 10);
+
+	intel_de_write_fw(i915, GMBUS4(i915), 0);
+	remove_wait_queue(&i915->display.gmbus.wait_queue, &wait);
+
+	return ret;
+}
+
+static unsigned int gmbus_max_xfer_size(struct drm_i915_private *i915)
+{
+	return DISPLAY_VER(i915) >= 9 ? GEN9_GMBUS_BYTE_COUNT_MAX :
+	       GMBUS_BYTE_COUNT_MAX;
+}
+
+static int
+gmbus_xfer_read_chunk(struct drm_i915_private *i915,
+		      unsigned short addr, u8 *buf, unsigned int len,
+		      u32 gmbus0_reg, u32 gmbus1_index)
+{
+	unsigned int size = len;
+	bool burst_read = len > gmbus_max_xfer_size(i915);
+	bool extra_byte_added = false;
+
+	if (burst_read) {
+		/*
+		 * As per HW Spec, for 512Bytes need to read extra Byte and
+		 * Ignore the extra byte read.
+		 */
+		if (len == 512) {
+			extra_byte_added = true;
+			len++;
+		}
+		size = len % 256 + 256;
+		intel_de_write_fw(i915, GMBUS0(i915),
+				  gmbus0_reg | GMBUS_BYTE_CNT_OVERRIDE);
+	}
+
+	intel_de_write_fw(i915, GMBUS1(i915),
+			  gmbus1_index | GMBUS_CYCLE_WAIT | (size << GMBUS_BYTE_COUNT_SHIFT) | (addr << GMBUS_SLAVE_ADDR_SHIFT) | GMBUS_SLAVE_READ | GMBUS_SW_RDY);
+	while (len) {
+		int ret;
+		u32 val, loop = 0;
+
+		ret = gmbus_wait(i915, GMBUS_HW_RDY, GMBUS_HW_RDY_EN);
+		if (ret)
+			return ret;
+
+		val = intel_de_read_fw(i915, GMBUS3(i915));
+		do {
+			if (extra_byte_added && len == 1)
+				break;
+
+			*buf++ = val & 0xff;
+			val >>= 8;
+		} while (--len && ++loop < 4);
+
+		if (burst_read && len == size - 4)
+			/* Reset the override bit */
+			intel_de_write_fw(i915, GMBUS0(i915), gmbus0_reg);
+	}
+
+	return 0;
+}
+
+/*
+ * HW spec says that 512Bytes in Burst read need special treatment.
+ * But it doesn't talk about other multiple of 256Bytes. And couldn't locate
+ * an I2C slave, which supports such a lengthy burst read too for experiments.
+ *
+ * So until things get clarified on HW support, to avoid the burst read length
+ * in fold of 256Bytes except 512, max burst read length is fixed at 767Bytes.
+ */
+#define INTEL_GMBUS_BURST_READ_MAX_LEN		767U
+
+static int
+gmbus_xfer_read(struct drm_i915_private *i915, struct i2c_msg *msg,
+		u32 gmbus0_reg, u32 gmbus1_index)
+{
+	u8 *buf = msg->buf;
+	unsigned int rx_size = msg->len;
+	unsigned int len;
+	int ret;
+
+	do {
+		if (HAS_GMBUS_BURST_READ(i915))
+			len = min(rx_size, INTEL_GMBUS_BURST_READ_MAX_LEN);
+		else
+			len = min(rx_size, gmbus_max_xfer_size(i915));
+
+		ret = gmbus_xfer_read_chunk(i915, msg->addr, buf, len,
+					    gmbus0_reg, gmbus1_index);
+		if (ret)
+			return ret;
+
+		rx_size -= len;
+		buf += len;
+	} while (rx_size != 0);
+
+	return 0;
+}
+
+static int
+gmbus_xfer_write_chunk(struct drm_i915_private *i915,
+		       unsigned short addr, u8 *buf, unsigned int len,
+		       u32 gmbus1_index)
+{
+	unsigned int chunk_size = len;
+	u32 val, loop;
+
+	val = loop = 0;
+	while (len && loop < 4) {
+		val |= *buf++ << (8 * loop++);
+		len -= 1;
+	}
+
+	intel_de_write_fw(i915, GMBUS3(i915), val);
+	intel_de_write_fw(i915, GMBUS1(i915),
+			  gmbus1_index | GMBUS_CYCLE_WAIT | (chunk_size << GMBUS_BYTE_COUNT_SHIFT) | (addr << GMBUS_SLAVE_ADDR_SHIFT) | GMBUS_SLAVE_WRITE | GMBUS_SW_RDY);
+	while (len) {
+		int ret;
+
+		val = loop = 0;
+		do {
+			val |= *buf++ << (8 * loop);
+		} while (--len && ++loop < 4);
+
+		intel_de_write_fw(i915, GMBUS3(i915), val);
+
+		ret = gmbus_wait(i915, GMBUS_HW_RDY, GMBUS_HW_RDY_EN);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int
+gmbus_xfer_write(struct drm_i915_private *i915, struct i2c_msg *msg,
+		 u32 gmbus1_index)
+{
+	u8 *buf = msg->buf;
+	unsigned int tx_size = msg->len;
+	unsigned int len;
+	int ret;
+
+	do {
+		len = min(tx_size, gmbus_max_xfer_size(i915));
+
+		ret = gmbus_xfer_write_chunk(i915, msg->addr, buf, len,
+					     gmbus1_index);
+		if (ret)
+			return ret;
+
+		buf += len;
+		tx_size -= len;
+	} while (tx_size != 0);
+
+	return 0;
+}
+
+/*
+ * The gmbus controller can combine a 1 or 2 byte write with another read/write
+ * that immediately follows it by using an "INDEX" cycle.
+ */
+static bool
+gmbus_is_index_xfer(struct i2c_msg *msgs, int i, int num)
+{
+	return (i + 1 < num &&
+		msgs[i].addr == msgs[i + 1].addr &&
+		!(msgs[i].flags & I2C_M_RD) &&
+		(msgs[i].len == 1 || msgs[i].len == 2) &&
+		msgs[i + 1].len > 0);
+}
+
+static int
+gmbus_index_xfer(struct drm_i915_private *i915, struct i2c_msg *msgs,
+		 u32 gmbus0_reg)
+{
+	u32 gmbus1_index = 0;
+	u32 gmbus5 = 0;
+	int ret;
+
+	if (msgs[0].len == 2)
+		gmbus5 = GMBUS_2BYTE_INDEX_EN |
+			 msgs[0].buf[1] | (msgs[0].buf[0] << 8);
+	if (msgs[0].len == 1)
+		gmbus1_index = GMBUS_CYCLE_INDEX |
+			       (msgs[0].buf[0] << GMBUS_SLAVE_INDEX_SHIFT);
+
+	/* GMBUS5 holds 16-bit index */
+	if (gmbus5)
+		intel_de_write_fw(i915, GMBUS5(i915), gmbus5);
+
+	if (msgs[1].flags & I2C_M_RD)
+		ret = gmbus_xfer_read(i915, &msgs[1], gmbus0_reg,
+				      gmbus1_index);
+	else
+		ret = gmbus_xfer_write(i915, &msgs[1], gmbus1_index);
+
+	/* Clear GMBUS5 after each index transfer */
+	if (gmbus5)
+		intel_de_write_fw(i915, GMBUS5(i915), 0);
+
+	return ret;
+}
+
+static int
+do_gmbus_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs, int num,
+	      u32 gmbus0_source)
+{
+	struct intel_gmbus *bus = to_intel_gmbus(adapter);
+	struct drm_i915_private *i915 = bus->i915;
+	int i = 0, inc, try = 0;
+	int ret = 0;
+
+	/* Display WA #0868: skl,bxt,kbl,cfl,glk */
+	if (IS_GEMINILAKE(i915) || IS_BROXTON(i915))
+		bxt_gmbus_clock_gating(i915, false);
+	else if (HAS_PCH_SPT(i915) || HAS_PCH_CNP(i915))
+		pch_gmbus_clock_gating(i915, false);
+
+retry:
+	intel_de_write_fw(i915, GMBUS0(i915), gmbus0_source | bus->reg0);
+
+	for (; i < num; i += inc) {
+		inc = 1;
+		if (gmbus_is_index_xfer(msgs, i, num)) {
+			ret = gmbus_index_xfer(i915, &msgs[i],
+					       gmbus0_source | bus->reg0);
+			inc = 2; /* an index transmission is two msgs */
+		} else if (msgs[i].flags & I2C_M_RD) {
+			ret = gmbus_xfer_read(i915, &msgs[i],
+					      gmbus0_source | bus->reg0, 0);
+		} else {
+			ret = gmbus_xfer_write(i915, &msgs[i], 0);
+		}
+
+		if (!ret)
+			ret = gmbus_wait(i915,
+					 GMBUS_HW_WAIT_PHASE, GMBUS_HW_WAIT_EN);
+		if (ret == -ETIMEDOUT)
+			goto timeout;
+		else if (ret)
+			goto clear_err;
+	}
+
+	/* Generate a STOP condition on the bus. Note that gmbus can't generata
+	 * a STOP on the very first cycle. To simplify the code we
+	 * unconditionally generate the STOP condition with an additional gmbus
+	 * cycle. */
+	intel_de_write_fw(i915, GMBUS1(i915), GMBUS_CYCLE_STOP | GMBUS_SW_RDY);
+
+	/* Mark the GMBUS interface as disabled after waiting for idle.
+	 * We will re-enable it at the start of the next xfer,
+	 * till then let it sleep.
+	 */
+	if (gmbus_wait_idle(i915)) {
+		drm_dbg_kms(&i915->drm,
+			    "GMBUS [%s] timed out waiting for idle\n",
+			    adapter->name);
+		ret = -ETIMEDOUT;
+	}
+	intel_de_write_fw(i915, GMBUS0(i915), 0);
+	ret = ret ?: i;
+	goto out;
+
+clear_err:
+	/*
+	 * Wait for bus to IDLE before clearing NAK.
+	 * If we clear the NAK while bus is still active, then it will stay
+	 * active and the next transaction may fail.
+	 *
+	 * If no ACK is received during the address phase of a transaction, the
+	 * adapter must report -ENXIO. It is not clear what to return if no ACK
+	 * is received at other times. But we have to be careful to not return
+	 * spurious -ENXIO because that will prevent i2c and drm edid functions
+	 * from retrying. So return -ENXIO only when gmbus properly quiescents -
+	 * timing out seems to happen when there _is_ a ddc chip present, but
+	 * it's slow responding and only answers on the 2nd retry.
+	 */
+	ret = -ENXIO;
+	if (gmbus_wait_idle(i915)) {
+		drm_dbg_kms(&i915->drm,
+			    "GMBUS [%s] timed out after NAK\n",
+			    adapter->name);
+		ret = -ETIMEDOUT;
+	}
+
+	/* Toggle the Software Clear Interrupt bit. This has the effect
+	 * of resetting the GMBUS controller and so clearing the
+	 * BUS_ERROR raised by the slave's NAK.
+	 */
+	intel_de_write_fw(i915, GMBUS1(i915), GMBUS_SW_CLR_INT);
+	intel_de_write_fw(i915, GMBUS1(i915), 0);
+	intel_de_write_fw(i915, GMBUS0(i915), 0);
+
+	drm_dbg_kms(&i915->drm, "GMBUS [%s] NAK for addr: %04x %c(%d)\n",
+		    adapter->name, msgs[i].addr,
+		    (msgs[i].flags & I2C_M_RD) ? 'r' : 'w', msgs[i].len);
+
+	/*
+	 * Passive adapters sometimes NAK the first probe. Retry the first
+	 * message once on -ENXIO for GMBUS transfers; the bit banging algorithm
+	 * has retries internally. See also the retry loop in
+	 * drm_do_probe_ddc_edid, which bails out on the first -ENXIO.
+	 */
+	if (ret == -ENXIO && i == 0 && try++ == 0) {
+		drm_dbg_kms(&i915->drm,
+			    "GMBUS [%s] NAK on first message, retry\n",
+			    adapter->name);
+		goto retry;
+	}
+
+	goto out;
+
+timeout:
+	drm_dbg_kms(&i915->drm,
+		    "GMBUS [%s] timed out, falling back to bit banging on pin %d\n",
+		    bus->adapter.name, bus->reg0 & 0xff);
+	intel_de_write_fw(i915, GMBUS0(i915), 0);
+
+	/*
+	 * Hardware may not support GMBUS over these pins? Try GPIO bitbanging
+	 * instead. Use EAGAIN to have i2c core retry.
+	 */
+	ret = -EAGAIN;
+
+out:
+	/* Display WA #0868: skl,bxt,kbl,cfl,glk */
+	if (IS_GEMINILAKE(i915) || IS_BROXTON(i915))
+		bxt_gmbus_clock_gating(i915, true);
+	else if (HAS_PCH_SPT(i915) || HAS_PCH_CNP(i915))
+		pch_gmbus_clock_gating(i915, true);
+
+	return ret;
+}
+
+static int
+gmbus_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs, int num)
+{
+	struct intel_gmbus *bus = to_intel_gmbus(adapter);
+	struct drm_i915_private *i915 = bus->i915;
+	intel_wakeref_t wakeref;
+	int ret;
+
+	wakeref = intel_display_power_get(i915, POWER_DOMAIN_GMBUS);
+
+	if (bus->force_bit) {
+		ret = i2c_bit_algo.master_xfer(adapter, msgs, num);
+		if (ret < 0)
+			bus->force_bit &= ~GMBUS_FORCE_BIT_RETRY;
+	} else {
+		ret = do_gmbus_xfer(adapter, msgs, num, 0);
+		if (ret == -EAGAIN)
+			bus->force_bit |= GMBUS_FORCE_BIT_RETRY;
+	}
+
+	intel_display_power_put(i915, POWER_DOMAIN_GMBUS, wakeref);
+
+	return ret;
+}
+
+int intel_gmbus_output_aksv(struct i2c_adapter *adapter)
+{
+	struct intel_gmbus *bus = to_intel_gmbus(adapter);
+	struct drm_i915_private *i915 = bus->i915;
+	u8 cmd = DRM_HDCP_DDC_AKSV;
+	u8 buf[DRM_HDCP_KSV_LEN] = { 0 };
+	struct i2c_msg msgs[] = {
+		{
+			.addr = DRM_HDCP_DDC_ADDR,
+			.flags = 0,
+			.len = sizeof(cmd),
+			.buf = &cmd,
+		},
+		{
+			.addr = DRM_HDCP_DDC_ADDR,
+			.flags = 0,
+			.len = sizeof(buf),
+			.buf = buf,
+		}
+	};
+	intel_wakeref_t wakeref;
+	int ret;
+
+	wakeref = intel_display_power_get(i915, POWER_DOMAIN_GMBUS);
+	mutex_lock(&i915->display.gmbus.mutex);
+
+	/*
+	 * In order to output Aksv to the receiver, use an indexed write to
+	 * pass the i2c command, and tell GMBUS to use the HW-provided value
+	 * instead of sourcing GMBUS3 for the data.
+	 */
+	ret = do_gmbus_xfer(adapter, msgs, ARRAY_SIZE(msgs), GMBUS_AKSV_SELECT);
+
+	mutex_unlock(&i915->display.gmbus.mutex);
+	intel_display_power_put(i915, POWER_DOMAIN_GMBUS, wakeref);
+
+	return ret;
+}
+
+static u32 gmbus_func(struct i2c_adapter *adapter)
+{
+	return i2c_bit_algo.functionality(adapter) &
+		(I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL |
+		/* I2C_FUNC_10BIT_ADDR | */
+		I2C_FUNC_SMBUS_READ_BLOCK_DATA |
+		I2C_FUNC_SMBUS_BLOCK_PROC_CALL);
+}
+
+static const struct i2c_algorithm gmbus_algorithm = {
+	.master_xfer	= gmbus_xfer,
+	.functionality	= gmbus_func
+};
+
+static void gmbus_lock_bus(struct i2c_adapter *adapter,
+			   unsigned int flags)
+{
+	struct intel_gmbus *bus = to_intel_gmbus(adapter);
+	struct drm_i915_private *i915 = bus->i915;
+
+	mutex_lock(&i915->display.gmbus.mutex);
+}
+
+static int gmbus_trylock_bus(struct i2c_adapter *adapter,
+			     unsigned int flags)
+{
+	struct intel_gmbus *bus = to_intel_gmbus(adapter);
+	struct drm_i915_private *i915 = bus->i915;
+
+	return mutex_trylock(&i915->display.gmbus.mutex);
+}
+
+static void gmbus_unlock_bus(struct i2c_adapter *adapter,
+			     unsigned int flags)
+{
+	struct intel_gmbus *bus = to_intel_gmbus(adapter);
+	struct drm_i915_private *i915 = bus->i915;
+
+	mutex_unlock(&i915->display.gmbus.mutex);
+}
+
+static const struct i2c_lock_operations gmbus_lock_ops = {
+	.lock_bus =    gmbus_lock_bus,
+	.trylock_bus = gmbus_trylock_bus,
+	.unlock_bus =  gmbus_unlock_bus,
+};
+
+/**
+ * intel_gmbus_setup - instantiate all Intel i2c GMBuses
+ * @i915: i915 device private
+ */
+int intel_gmbus_setup(struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	unsigned int pin;
+	int ret;
+
+	if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+		i915->display.gmbus.mmio_base = VLV_DISPLAY_BASE;
+	else if (!HAS_GMCH(i915))
+		/*
+		 * Broxton uses the same PCH offsets for South Display Engine,
+		 * even though it doesn't have a PCH.
+		 */
+		i915->display.gmbus.mmio_base = PCH_DISPLAY_BASE;
+
+	mutex_init(&i915->display.gmbus.mutex);
+	init_waitqueue_head(&i915->display.gmbus.wait_queue);
+
+	for (pin = 0; pin < ARRAY_SIZE(i915->display.gmbus.bus); pin++) {
+		const struct gmbus_pin *gmbus_pin;
+		struct intel_gmbus *bus;
+
+		gmbus_pin = get_gmbus_pin(i915, pin);
+		if (!gmbus_pin)
+			continue;
+
+		bus = kzalloc(sizeof(*bus), GFP_KERNEL);
+		if (!bus) {
+			ret = -ENOMEM;
+			goto err;
+		}
+
+		bus->adapter.owner = THIS_MODULE;
+		bus->adapter.class = I2C_CLASS_DDC;
+		snprintf(bus->adapter.name,
+			 sizeof(bus->adapter.name),
+			 "i915 gmbus %s", gmbus_pin->name);
+
+		bus->adapter.dev.parent = &pdev->dev;
+		bus->i915 = i915;
+
+		bus->adapter.algo = &gmbus_algorithm;
+		bus->adapter.lock_ops = &gmbus_lock_ops;
+
+		/*
+		 * We wish to retry with bit banging
+		 * after a timed out GMBUS attempt.
+		 */
+		bus->adapter.retries = 1;
+
+		/* By default use a conservative clock rate */
+		bus->reg0 = pin | GMBUS_RATE_100KHZ;
+
+		/* gmbus seems to be broken on i830 */
+		if (IS_I830(i915))
+			bus->force_bit = 1;
+
+		intel_gpio_setup(bus, GPIO(i915, gmbus_pin->gpio));
+
+		ret = i2c_add_adapter(&bus->adapter);
+		if (ret) {
+			kfree(bus);
+			goto err;
+		}
+
+		i915->display.gmbus.bus[pin] = bus;
+	}
+
+	intel_gmbus_reset(i915);
+
+	return 0;
+
+err:
+	intel_gmbus_teardown(i915);
+
+	return ret;
+}
+
+struct i2c_adapter *intel_gmbus_get_adapter(struct drm_i915_private *i915,
+					    unsigned int pin)
+{
+	if (drm_WARN_ON(&i915->drm, pin >= ARRAY_SIZE(i915->display.gmbus.bus) ||
+			!i915->display.gmbus.bus[pin]))
+		return NULL;
+
+	return &i915->display.gmbus.bus[pin]->adapter;
+}
+
+void intel_gmbus_force_bit(struct i2c_adapter *adapter, bool force_bit)
+{
+	struct intel_gmbus *bus = to_intel_gmbus(adapter);
+	struct drm_i915_private *i915 = bus->i915;
+
+	mutex_lock(&i915->display.gmbus.mutex);
+
+	bus->force_bit += force_bit ? 1 : -1;
+	drm_dbg_kms(&i915->drm,
+		    "%sabling bit-banging on %s. force bit now %d\n",
+		    force_bit ? "en" : "dis", adapter->name,
+		    bus->force_bit);
+
+	mutex_unlock(&i915->display.gmbus.mutex);
+}
+
+bool intel_gmbus_is_forced_bit(struct i2c_adapter *adapter)
+{
+	struct intel_gmbus *bus = to_intel_gmbus(adapter);
+
+	return bus->force_bit;
+}
+
+void intel_gmbus_teardown(struct drm_i915_private *i915)
+{
+	unsigned int pin;
+
+	for (pin = 0; pin < ARRAY_SIZE(i915->display.gmbus.bus); pin++) {
+		struct intel_gmbus *bus;
+
+		bus = i915->display.gmbus.bus[pin];
+		if (!bus)
+			continue;
+
+		i2c_del_adapter(&bus->adapter);
+
+		kfree(bus);
+		i915->display.gmbus.bus[pin] = NULL;
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_gmbus.h b/drivers/gpu/drm/i915/display/intel_gmbus.h
new file mode 100644
index 000000000..20f704bd4
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_gmbus.h
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_GMBUS_H__
+#define __INTEL_GMBUS_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct i2c_adapter;
+
+#define GMBUS_PIN_DISABLED	0
+#define GMBUS_PIN_SSC		1
+#define GMBUS_PIN_VGADDC	2
+#define GMBUS_PIN_PANEL		3
+#define GMBUS_PIN_DPD_CHV	3 /* HDMID_CHV */
+#define GMBUS_PIN_DPC		4 /* HDMIC */
+#define GMBUS_PIN_DPB		5 /* SDVO, HDMIB */
+#define GMBUS_PIN_DPD		6 /* HDMID */
+#define GMBUS_PIN_RESERVED	7 /* 7 reserved */
+#define GMBUS_PIN_1_BXT		1 /* BXT+ (atom) and CNP+ (big core) */
+#define GMBUS_PIN_2_BXT		2
+#define GMBUS_PIN_3_BXT		3
+#define GMBUS_PIN_4_CNP		4
+#define GMBUS_PIN_5_MTP		5
+#define GMBUS_PIN_9_TC1_ICP	9
+#define GMBUS_PIN_10_TC2_ICP	10
+#define GMBUS_PIN_11_TC3_ICP	11
+#define GMBUS_PIN_12_TC4_ICP	12
+#define GMBUS_PIN_13_TC5_TGP	13
+#define GMBUS_PIN_14_TC6_TGP	14
+
+#define GMBUS_NUM_PINS	15 /* including 0 */
+
+int intel_gmbus_setup(struct drm_i915_private *dev_priv);
+void intel_gmbus_teardown(struct drm_i915_private *dev_priv);
+bool intel_gmbus_is_valid_pin(struct drm_i915_private *dev_priv,
+			      unsigned int pin);
+int intel_gmbus_output_aksv(struct i2c_adapter *adapter);
+
+struct i2c_adapter *
+intel_gmbus_get_adapter(struct drm_i915_private *dev_priv, unsigned int pin);
+void intel_gmbus_force_bit(struct i2c_adapter *adapter, bool force_bit);
+bool intel_gmbus_is_forced_bit(struct i2c_adapter *adapter);
+void intel_gmbus_reset(struct drm_i915_private *dev_priv);
+
+#endif /* __INTEL_GMBUS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_gmbus_regs.h b/drivers/gpu/drm/i915/display/intel_gmbus_regs.h
new file mode 100644
index 000000000..53aacbda9
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_gmbus_regs.h
@@ -0,0 +1,81 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_GMBUS_REGS_H__
+#define __INTEL_GMBUS_REGS_H__
+
+#include "i915_reg_defs.h"
+
+#define GMBUS_MMIO_BASE(__i915) ((__i915)->display.gmbus.mmio_base)
+
+#define GPIO(__i915, gpio)	_MMIO(GMBUS_MMIO_BASE(__i915) + 0x5010 + 4 * (gpio))
+#define   GPIO_CLOCK_DIR_MASK		(1 << 0)
+#define   GPIO_CLOCK_DIR_IN		(0 << 1)
+#define   GPIO_CLOCK_DIR_OUT		(1 << 1)
+#define   GPIO_CLOCK_VAL_MASK		(1 << 2)
+#define   GPIO_CLOCK_VAL_OUT		(1 << 3)
+#define   GPIO_CLOCK_VAL_IN		(1 << 4)
+#define   GPIO_CLOCK_PULLUP_DISABLE	(1 << 5)
+#define   GPIO_DATA_DIR_MASK		(1 << 8)
+#define   GPIO_DATA_DIR_IN		(0 << 9)
+#define   GPIO_DATA_DIR_OUT		(1 << 9)
+#define   GPIO_DATA_VAL_MASK		(1 << 10)
+#define   GPIO_DATA_VAL_OUT		(1 << 11)
+#define   GPIO_DATA_VAL_IN		(1 << 12)
+#define   GPIO_DATA_PULLUP_DISABLE	(1 << 13)
+
+/* clock/port select */
+#define GMBUS0(__i915)		_MMIO(GMBUS_MMIO_BASE(__i915) + 0x5100)
+#define   GMBUS_AKSV_SELECT		(1 << 11)
+#define   GMBUS_RATE_100KHZ		(0 << 8)
+#define   GMBUS_RATE_50KHZ		(1 << 8)
+#define   GMBUS_RATE_400KHZ		(2 << 8) /* reserved on Pineview */
+#define   GMBUS_RATE_1MHZ		(3 << 8) /* reserved on Pineview */
+#define   GMBUS_HOLD_EXT		(1 << 7) /* 300ns hold time, rsvd on Pineview */
+#define   GMBUS_BYTE_CNT_OVERRIDE	(1 << 6)
+
+/* command/status */
+#define GMBUS1(__i915)		_MMIO(GMBUS_MMIO_BASE(__i915) + 0x5104)
+#define   GMBUS_SW_CLR_INT		(1 << 31)
+#define   GMBUS_SW_RDY			(1 << 30)
+#define   GMBUS_ENT			(1 << 29) /* enable timeout */
+#define   GMBUS_CYCLE_NONE		(0 << 25)
+#define   GMBUS_CYCLE_WAIT		(1 << 25)
+#define   GMBUS_CYCLE_INDEX		(2 << 25)
+#define   GMBUS_CYCLE_STOP		(4 << 25)
+#define   GMBUS_BYTE_COUNT_SHIFT	16
+#define   GMBUS_BYTE_COUNT_MAX		256U
+#define   GEN9_GMBUS_BYTE_COUNT_MAX	511U
+#define   GMBUS_SLAVE_INDEX_SHIFT	8
+#define   GMBUS_SLAVE_ADDR_SHIFT	1
+#define   GMBUS_SLAVE_READ		(1 << 0)
+#define   GMBUS_SLAVE_WRITE		(0 << 0)
+
+/* status */
+#define GMBUS2(__i915)		_MMIO(GMBUS_MMIO_BASE(__i915) + 0x5108)
+#define   GMBUS_INUSE			(1 << 15)
+#define   GMBUS_HW_WAIT_PHASE		(1 << 14)
+#define   GMBUS_STALL_TIMEOUT		(1 << 13)
+#define   GMBUS_INT			(1 << 12)
+#define   GMBUS_HW_RDY			(1 << 11)
+#define   GMBUS_SATOER			(1 << 10)
+#define   GMBUS_ACTIVE			(1 << 9)
+
+/* data buffer bytes 3-0 */
+#define GMBUS3(__i915)		_MMIO(GMBUS_MMIO_BASE(__i915) + 0x510c)
+
+/* interrupt mask (Pineview+) */
+#define GMBUS4(__i915)		_MMIO(GMBUS_MMIO_BASE(__i915) + 0x5110)
+#define   GMBUS_SLAVE_TIMEOUT_EN	(1 << 4)
+#define   GMBUS_NAK_EN			(1 << 3)
+#define   GMBUS_IDLE_EN			(1 << 2)
+#define   GMBUS_HW_WAIT_EN		(1 << 1)
+#define   GMBUS_HW_RDY_EN		(1 << 0)
+
+/* byte index */
+#define GMBUS5(__i915)		_MMIO(GMBUS_MMIO_BASE(__i915) + 0x5120)
+#define   GMBUS_2BYTE_INDEX_EN		(1 << 31)
+
+#endif /* __INTEL_GMBUS_REGS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_hdcp.c b/drivers/gpu/drm/i915/display/intel_hdcp.c
new file mode 100644
index 000000000..6406fd487
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_hdcp.c
@@ -0,0 +1,2586 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright (C) 2017 Google, Inc.
+ * Copyright _ 2017-2019, Intel Corporation.
+ *
+ * Authors:
+ * Sean Paul <seanpaul@chromium.org>
+ * Ramalingam C <ramalingam.c@intel.com>
+ */
+
+#include <linux/component.h>
+#include <linux/i2c.h>
+#include <linux/random.h>
+
+#include <drm/display/drm_hdcp_helper.h>
+#include <drm/i915_component.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_connector.h"
+#include "intel_de.h"
+#include "intel_display_power.h"
+#include "intel_display_power_well.h"
+#include "intel_display_types.h"
+#include "intel_hdcp.h"
+#include "intel_hdcp_regs.h"
+#include "intel_pcode.h"
+
+#define KEY_LOAD_TRIES	5
+#define HDCP2_LC_RETRY_CNT			3
+
+static int intel_conn_to_vcpi(struct intel_connector *connector)
+{
+	struct drm_dp_mst_topology_mgr *mgr;
+	struct drm_dp_mst_atomic_payload *payload;
+	struct drm_dp_mst_topology_state *mst_state;
+	int vcpi = 0;
+
+	/* For HDMI this is forced to be 0x0. For DP SST also this is 0x0. */
+	if (!connector->port)
+		return 0;
+	mgr = connector->port->mgr;
+
+	drm_modeset_lock(&mgr->base.lock, NULL);
+	mst_state = to_drm_dp_mst_topology_state(mgr->base.state);
+	payload = drm_atomic_get_mst_payload_state(mst_state, connector->port);
+	if (drm_WARN_ON(mgr->dev, !payload))
+		goto out;
+
+	vcpi = payload->vcpi;
+	if (drm_WARN_ON(mgr->dev, vcpi < 0)) {
+		vcpi = 0;
+		goto out;
+	}
+out:
+	drm_modeset_unlock(&mgr->base.lock);
+	return vcpi;
+}
+
+/*
+ * intel_hdcp_required_content_stream selects the most highest common possible HDCP
+ * content_type for all streams in DP MST topology because security f/w doesn't
+ * have any provision to mark content_type for each stream separately, it marks
+ * all available streams with the content_type proivided at the time of port
+ * authentication. This may prohibit the userspace to use type1 content on
+ * HDCP 2.2 capable sink because of other sink are not capable of HDCP 2.2 in
+ * DP MST topology. Though it is not compulsory, security fw should change its
+ * policy to mark different content_types for different streams.
+ */
+static int
+intel_hdcp_required_content_stream(struct intel_digital_port *dig_port)
+{
+	struct drm_connector_list_iter conn_iter;
+	struct intel_digital_port *conn_dig_port;
+	struct intel_connector *connector;
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	bool enforce_type0 = false;
+	int k;
+
+	data->k = 0;
+
+	if (dig_port->hdcp_auth_status)
+		return 0;
+
+	drm_connector_list_iter_begin(&i915->drm, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		if (connector->base.status == connector_status_disconnected)
+			continue;
+
+		if (!intel_encoder_is_mst(intel_attached_encoder(connector)))
+			continue;
+
+		conn_dig_port = intel_attached_dig_port(connector);
+		if (conn_dig_port != dig_port)
+			continue;
+
+		if (!enforce_type0 && !dig_port->hdcp_mst_type1_capable)
+			enforce_type0 = true;
+
+		data->streams[data->k].stream_id = intel_conn_to_vcpi(connector);
+		data->k++;
+
+		/* if there is only one active stream */
+		if (dig_port->dp.active_mst_links <= 1)
+			break;
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	if (drm_WARN_ON(&i915->drm, data->k > INTEL_NUM_PIPES(i915) || data->k == 0))
+		return -EINVAL;
+
+	/*
+	 * Apply common protection level across all streams in DP MST Topology.
+	 * Use highest supported content type for all streams in DP MST Topology.
+	 */
+	for (k = 0; k < data->k; k++)
+		data->streams[k].stream_type =
+			enforce_type0 ? DRM_MODE_HDCP_CONTENT_TYPE0 : DRM_MODE_HDCP_CONTENT_TYPE1;
+
+	return 0;
+}
+
+static int intel_hdcp_prepare_streams(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	int ret;
+
+	if (!intel_encoder_is_mst(intel_attached_encoder(connector))) {
+		data->k = 1;
+		data->streams[0].stream_type = hdcp->content_type;
+	} else {
+		ret = intel_hdcp_required_content_stream(dig_port);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static
+bool intel_hdcp_is_ksv_valid(u8 *ksv)
+{
+	int i, ones = 0;
+	/* KSV has 20 1's and 20 0's */
+	for (i = 0; i < DRM_HDCP_KSV_LEN; i++)
+		ones += hweight8(ksv[i]);
+	if (ones != 20)
+		return false;
+
+	return true;
+}
+
+static
+int intel_hdcp_read_valid_bksv(struct intel_digital_port *dig_port,
+			       const struct intel_hdcp_shim *shim, u8 *bksv)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	int ret, i, tries = 2;
+
+	/* HDCP spec states that we must retry the bksv if it is invalid */
+	for (i = 0; i < tries; i++) {
+		ret = shim->read_bksv(dig_port, bksv);
+		if (ret)
+			return ret;
+		if (intel_hdcp_is_ksv_valid(bksv))
+			break;
+	}
+	if (i == tries) {
+		drm_dbg_kms(&i915->drm, "Bksv is invalid\n");
+		return -ENODEV;
+	}
+
+	return 0;
+}
+
+/* Is HDCP1.4 capable on Platform and Sink */
+bool intel_hdcp_capable(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	const struct intel_hdcp_shim *shim = connector->hdcp.shim;
+	bool capable = false;
+	u8 bksv[5];
+
+	if (!shim)
+		return capable;
+
+	if (shim->hdcp_capable) {
+		shim->hdcp_capable(dig_port, &capable);
+	} else {
+		if (!intel_hdcp_read_valid_bksv(dig_port, shim, bksv))
+			capable = true;
+	}
+
+	return capable;
+}
+
+/* Is HDCP2.2 capable on Platform and Sink */
+bool intel_hdcp2_capable(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	bool capable = false;
+
+	/* I915 support for HDCP2.2 */
+	if (!hdcp->hdcp2_supported)
+		return false;
+
+	/* MEI interface is solid */
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	if (!dev_priv->display.hdcp.comp_added ||  !dev_priv->display.hdcp.master) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return false;
+	}
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	/* Sink's capability for HDCP2.2 */
+	hdcp->shim->hdcp_2_2_capable(dig_port, &capable);
+
+	return capable;
+}
+
+static bool intel_hdcp_in_use(struct drm_i915_private *dev_priv,
+			      enum transcoder cpu_transcoder, enum port port)
+{
+	return intel_de_read(dev_priv,
+	                     HDCP_STATUS(dev_priv, cpu_transcoder, port)) &
+	       HDCP_STATUS_ENC;
+}
+
+static bool intel_hdcp2_in_use(struct drm_i915_private *dev_priv,
+			       enum transcoder cpu_transcoder, enum port port)
+{
+	return intel_de_read(dev_priv,
+	                     HDCP2_STATUS(dev_priv, cpu_transcoder, port)) &
+	       LINK_ENCRYPTION_STATUS;
+}
+
+static int intel_hdcp_poll_ksv_fifo(struct intel_digital_port *dig_port,
+				    const struct intel_hdcp_shim *shim)
+{
+	int ret, read_ret;
+	bool ksv_ready;
+
+	/* Poll for ksv list ready (spec says max time allowed is 5s) */
+	ret = __wait_for(read_ret = shim->read_ksv_ready(dig_port,
+							 &ksv_ready),
+			 read_ret || ksv_ready, 5 * 1000 * 1000, 1000,
+			 100 * 1000);
+	if (ret)
+		return ret;
+	if (read_ret)
+		return read_ret;
+	if (!ksv_ready)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+static bool hdcp_key_loadable(struct drm_i915_private *dev_priv)
+{
+	enum i915_power_well_id id;
+	intel_wakeref_t wakeref;
+	bool enabled = false;
+
+	/*
+	 * On HSW and BDW, Display HW loads the Key as soon as Display resumes.
+	 * On all BXT+, SW can load the keys only when the PW#1 is turned on.
+	 */
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+		id = HSW_DISP_PW_GLOBAL;
+	else
+		id = SKL_DISP_PW_1;
+
+	/* PG1 (power well #1) needs to be enabled */
+	with_intel_runtime_pm(&dev_priv->runtime_pm, wakeref)
+		enabled = intel_display_power_well_is_enabled(dev_priv, id);
+
+	/*
+	 * Another req for hdcp key loadability is enabled state of pll for
+	 * cdclk. Without active crtc we wont land here. So we are assuming that
+	 * cdclk is already on.
+	 */
+
+	return enabled;
+}
+
+static void intel_hdcp_clear_keys(struct drm_i915_private *dev_priv)
+{
+	intel_de_write(dev_priv, HDCP_KEY_CONF, HDCP_CLEAR_KEYS_TRIGGER);
+	intel_de_write(dev_priv, HDCP_KEY_STATUS,
+		       HDCP_KEY_LOAD_DONE | HDCP_KEY_LOAD_STATUS | HDCP_FUSE_IN_PROGRESS | HDCP_FUSE_ERROR | HDCP_FUSE_DONE);
+}
+
+static int intel_hdcp_load_keys(struct drm_i915_private *dev_priv)
+{
+	int ret;
+	u32 val;
+
+	val = intel_de_read(dev_priv, HDCP_KEY_STATUS);
+	if ((val & HDCP_KEY_LOAD_DONE) && (val & HDCP_KEY_LOAD_STATUS))
+		return 0;
+
+	/*
+	 * On HSW and BDW HW loads the HDCP1.4 Key when Display comes
+	 * out of reset. So if Key is not already loaded, its an error state.
+	 */
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+		if (!(intel_de_read(dev_priv, HDCP_KEY_STATUS) & HDCP_KEY_LOAD_DONE))
+			return -ENXIO;
+
+	/*
+	 * Initiate loading the HDCP key from fuses.
+	 *
+	 * BXT+ platforms, HDCP key needs to be loaded by SW. Only display
+	 * version 9 platforms (minus BXT) differ in the key load trigger
+	 * process from other platforms. These platforms use the GT Driver
+	 * Mailbox interface.
+	 */
+	if (DISPLAY_VER(dev_priv) == 9 && !IS_BROXTON(dev_priv)) {
+		ret = snb_pcode_write(&dev_priv->uncore, SKL_PCODE_LOAD_HDCP_KEYS, 1);
+		if (ret) {
+			drm_err(&dev_priv->drm,
+				"Failed to initiate HDCP key load (%d)\n",
+				ret);
+			return ret;
+		}
+	} else {
+		intel_de_write(dev_priv, HDCP_KEY_CONF, HDCP_KEY_LOAD_TRIGGER);
+	}
+
+	/* Wait for the keys to load (500us) */
+	ret = __intel_wait_for_register(&dev_priv->uncore, HDCP_KEY_STATUS,
+					HDCP_KEY_LOAD_DONE, HDCP_KEY_LOAD_DONE,
+					10, 1, &val);
+	if (ret)
+		return ret;
+	else if (!(val & HDCP_KEY_LOAD_STATUS))
+		return -ENXIO;
+
+	/* Send Aksv over to PCH display for use in authentication */
+	intel_de_write(dev_priv, HDCP_KEY_CONF, HDCP_AKSV_SEND_TRIGGER);
+
+	return 0;
+}
+
+/* Returns updated SHA-1 index */
+static int intel_write_sha_text(struct drm_i915_private *dev_priv, u32 sha_text)
+{
+	intel_de_write(dev_priv, HDCP_SHA_TEXT, sha_text);
+	if (intel_de_wait_for_set(dev_priv, HDCP_REP_CTL, HDCP_SHA1_READY, 1)) {
+		drm_err(&dev_priv->drm, "Timed out waiting for SHA1 ready\n");
+		return -ETIMEDOUT;
+	}
+	return 0;
+}
+
+static
+u32 intel_hdcp_get_repeater_ctl(struct drm_i915_private *dev_priv,
+				enum transcoder cpu_transcoder, enum port port)
+{
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		switch (cpu_transcoder) {
+		case TRANSCODER_A:
+			return HDCP_TRANSA_REP_PRESENT |
+			       HDCP_TRANSA_SHA1_M0;
+		case TRANSCODER_B:
+			return HDCP_TRANSB_REP_PRESENT |
+			       HDCP_TRANSB_SHA1_M0;
+		case TRANSCODER_C:
+			return HDCP_TRANSC_REP_PRESENT |
+			       HDCP_TRANSC_SHA1_M0;
+		case TRANSCODER_D:
+			return HDCP_TRANSD_REP_PRESENT |
+			       HDCP_TRANSD_SHA1_M0;
+		default:
+			drm_err(&dev_priv->drm, "Unknown transcoder %d\n",
+				cpu_transcoder);
+			return -EINVAL;
+		}
+	}
+
+	switch (port) {
+	case PORT_A:
+		return HDCP_DDIA_REP_PRESENT | HDCP_DDIA_SHA1_M0;
+	case PORT_B:
+		return HDCP_DDIB_REP_PRESENT | HDCP_DDIB_SHA1_M0;
+	case PORT_C:
+		return HDCP_DDIC_REP_PRESENT | HDCP_DDIC_SHA1_M0;
+	case PORT_D:
+		return HDCP_DDID_REP_PRESENT | HDCP_DDID_SHA1_M0;
+	case PORT_E:
+		return HDCP_DDIE_REP_PRESENT | HDCP_DDIE_SHA1_M0;
+	default:
+		drm_err(&dev_priv->drm, "Unknown port %d\n", port);
+		return -EINVAL;
+	}
+}
+
+static
+int intel_hdcp_validate_v_prime(struct intel_connector *connector,
+				const struct intel_hdcp_shim *shim,
+				u8 *ksv_fifo, u8 num_downstream, u8 *bstatus)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	enum transcoder cpu_transcoder = connector->hdcp.cpu_transcoder;
+	enum port port = dig_port->base.port;
+	u32 vprime, sha_text, sha_leftovers, rep_ctl;
+	int ret, i, j, sha_idx;
+
+	/* Process V' values from the receiver */
+	for (i = 0; i < DRM_HDCP_V_PRIME_NUM_PARTS; i++) {
+		ret = shim->read_v_prime_part(dig_port, i, &vprime);
+		if (ret)
+			return ret;
+		intel_de_write(dev_priv, HDCP_SHA_V_PRIME(i), vprime);
+	}
+
+	/*
+	 * We need to write the concatenation of all device KSVs, BINFO (DP) ||
+	 * BSTATUS (HDMI), and M0 (which is added via HDCP_REP_CTL). This byte
+	 * stream is written via the HDCP_SHA_TEXT register in 32-bit
+	 * increments. Every 64 bytes, we need to write HDCP_REP_CTL again. This
+	 * index will keep track of our progress through the 64 bytes as well as
+	 * helping us work the 40-bit KSVs through our 32-bit register.
+	 *
+	 * NOTE: data passed via HDCP_SHA_TEXT should be big-endian
+	 */
+	sha_idx = 0;
+	sha_text = 0;
+	sha_leftovers = 0;
+	rep_ctl = intel_hdcp_get_repeater_ctl(dev_priv, cpu_transcoder, port);
+	intel_de_write(dev_priv, HDCP_REP_CTL, rep_ctl | HDCP_SHA1_TEXT_32);
+	for (i = 0; i < num_downstream; i++) {
+		unsigned int sha_empty;
+		u8 *ksv = &ksv_fifo[i * DRM_HDCP_KSV_LEN];
+
+		/* Fill up the empty slots in sha_text and write it out */
+		sha_empty = sizeof(sha_text) - sha_leftovers;
+		for (j = 0; j < sha_empty; j++) {
+			u8 off = ((sizeof(sha_text) - j - 1 - sha_leftovers) * 8);
+			sha_text |= ksv[j] << off;
+		}
+
+		ret = intel_write_sha_text(dev_priv, sha_text);
+		if (ret < 0)
+			return ret;
+
+		/* Programming guide writes this every 64 bytes */
+		sha_idx += sizeof(sha_text);
+		if (!(sha_idx % 64))
+			intel_de_write(dev_priv, HDCP_REP_CTL,
+				       rep_ctl | HDCP_SHA1_TEXT_32);
+
+		/* Store the leftover bytes from the ksv in sha_text */
+		sha_leftovers = DRM_HDCP_KSV_LEN - sha_empty;
+		sha_text = 0;
+		for (j = 0; j < sha_leftovers; j++)
+			sha_text |= ksv[sha_empty + j] <<
+					((sizeof(sha_text) - j - 1) * 8);
+
+		/*
+		 * If we still have room in sha_text for more data, continue.
+		 * Otherwise, write it out immediately.
+		 */
+		if (sizeof(sha_text) > sha_leftovers)
+			continue;
+
+		ret = intel_write_sha_text(dev_priv, sha_text);
+		if (ret < 0)
+			return ret;
+		sha_leftovers = 0;
+		sha_text = 0;
+		sha_idx += sizeof(sha_text);
+	}
+
+	/*
+	 * We need to write BINFO/BSTATUS, and M0 now. Depending on how many
+	 * bytes are leftover from the last ksv, we might be able to fit them
+	 * all in sha_text (first 2 cases), or we might need to split them up
+	 * into 2 writes (last 2 cases).
+	 */
+	if (sha_leftovers == 0) {
+		/* Write 16 bits of text, 16 bits of M0 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_16);
+		ret = intel_write_sha_text(dev_priv,
+					   bstatus[0] << 8 | bstatus[1]);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+
+		/* Write 32 bits of M0 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_0);
+		ret = intel_write_sha_text(dev_priv, 0);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+
+		/* Write 16 bits of M0 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_16);
+		ret = intel_write_sha_text(dev_priv, 0);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+
+	} else if (sha_leftovers == 1) {
+		/* Write 24 bits of text, 8 bits of M0 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_24);
+		sha_text |= bstatus[0] << 16 | bstatus[1] << 8;
+		/* Only 24-bits of data, must be in the LSB */
+		sha_text = (sha_text & 0xffffff00) >> 8;
+		ret = intel_write_sha_text(dev_priv, sha_text);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+
+		/* Write 32 bits of M0 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_0);
+		ret = intel_write_sha_text(dev_priv, 0);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+
+		/* Write 24 bits of M0 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_8);
+		ret = intel_write_sha_text(dev_priv, 0);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+
+	} else if (sha_leftovers == 2) {
+		/* Write 32 bits of text */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_32);
+		sha_text |= bstatus[0] << 8 | bstatus[1];
+		ret = intel_write_sha_text(dev_priv, sha_text);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+
+		/* Write 64 bits of M0 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_0);
+		for (i = 0; i < 2; i++) {
+			ret = intel_write_sha_text(dev_priv, 0);
+			if (ret < 0)
+				return ret;
+			sha_idx += sizeof(sha_text);
+		}
+
+		/*
+		 * Terminate the SHA-1 stream by hand. For the other leftover
+		 * cases this is appended by the hardware.
+		 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_32);
+		sha_text = DRM_HDCP_SHA1_TERMINATOR << 24;
+		ret = intel_write_sha_text(dev_priv, sha_text);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+	} else if (sha_leftovers == 3) {
+		/* Write 32 bits of text (filled from LSB) */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_32);
+		sha_text |= bstatus[0];
+		ret = intel_write_sha_text(dev_priv, sha_text);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+
+		/* Write 8 bits of text (filled from LSB), 24 bits of M0 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_8);
+		ret = intel_write_sha_text(dev_priv, bstatus[1]);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+
+		/* Write 32 bits of M0 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_0);
+		ret = intel_write_sha_text(dev_priv, 0);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+
+		/* Write 8 bits of M0 */
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       rep_ctl | HDCP_SHA1_TEXT_24);
+		ret = intel_write_sha_text(dev_priv, 0);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+	} else {
+		drm_dbg_kms(&dev_priv->drm, "Invalid number of leftovers %d\n",
+			    sha_leftovers);
+		return -EINVAL;
+	}
+
+	intel_de_write(dev_priv, HDCP_REP_CTL, rep_ctl | HDCP_SHA1_TEXT_32);
+	/* Fill up to 64-4 bytes with zeros (leave the last write for length) */
+	while ((sha_idx % 64) < (64 - sizeof(sha_text))) {
+		ret = intel_write_sha_text(dev_priv, 0);
+		if (ret < 0)
+			return ret;
+		sha_idx += sizeof(sha_text);
+	}
+
+	/*
+	 * Last write gets the length of the concatenation in bits. That is:
+	 *  - 5 bytes per device
+	 *  - 10 bytes for BINFO/BSTATUS(2), M0(8)
+	 */
+	sha_text = (num_downstream * 5 + 10) * 8;
+	ret = intel_write_sha_text(dev_priv, sha_text);
+	if (ret < 0)
+		return ret;
+
+	/* Tell the HW we're done with the hash and wait for it to ACK */
+	intel_de_write(dev_priv, HDCP_REP_CTL,
+		       rep_ctl | HDCP_SHA1_COMPLETE_HASH);
+	if (intel_de_wait_for_set(dev_priv, HDCP_REP_CTL,
+				  HDCP_SHA1_COMPLETE, 1)) {
+		drm_err(&dev_priv->drm, "Timed out waiting for SHA1 complete\n");
+		return -ETIMEDOUT;
+	}
+	if (!(intel_de_read(dev_priv, HDCP_REP_CTL) & HDCP_SHA1_V_MATCH)) {
+		drm_dbg_kms(&dev_priv->drm, "SHA-1 mismatch, HDCP failed\n");
+		return -ENXIO;
+	}
+
+	return 0;
+}
+
+/* Implements Part 2 of the HDCP authorization procedure */
+static
+int intel_hdcp_auth_downstream(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	const struct intel_hdcp_shim *shim = connector->hdcp.shim;
+	u8 bstatus[2], num_downstream, *ksv_fifo;
+	int ret, i, tries = 3;
+
+	ret = intel_hdcp_poll_ksv_fifo(dig_port, shim);
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "KSV list failed to become ready (%d)\n", ret);
+		return ret;
+	}
+
+	ret = shim->read_bstatus(dig_port, bstatus);
+	if (ret)
+		return ret;
+
+	if (DRM_HDCP_MAX_DEVICE_EXCEEDED(bstatus[0]) ||
+	    DRM_HDCP_MAX_CASCADE_EXCEEDED(bstatus[1])) {
+		drm_dbg_kms(&dev_priv->drm, "Max Topology Limit Exceeded\n");
+		return -EPERM;
+	}
+
+	/*
+	 * When repeater reports 0 device count, HDCP1.4 spec allows disabling
+	 * the HDCP encryption. That implies that repeater can't have its own
+	 * display. As there is no consumption of encrypted content in the
+	 * repeater with 0 downstream devices, we are failing the
+	 * authentication.
+	 */
+	num_downstream = DRM_HDCP_NUM_DOWNSTREAM(bstatus[0]);
+	if (num_downstream == 0) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Repeater with zero downstream devices\n");
+		return -EINVAL;
+	}
+
+	ksv_fifo = kcalloc(DRM_HDCP_KSV_LEN, num_downstream, GFP_KERNEL);
+	if (!ksv_fifo) {
+		drm_dbg_kms(&dev_priv->drm, "Out of mem: ksv_fifo\n");
+		return -ENOMEM;
+	}
+
+	ret = shim->read_ksv_fifo(dig_port, num_downstream, ksv_fifo);
+	if (ret)
+		goto err;
+
+	if (drm_hdcp_check_ksvs_revoked(&dev_priv->drm, ksv_fifo,
+					num_downstream) > 0) {
+		drm_err(&dev_priv->drm, "Revoked Ksv(s) in ksv_fifo\n");
+		ret = -EPERM;
+		goto err;
+	}
+
+	/*
+	 * When V prime mismatches, DP Spec mandates re-read of
+	 * V prime atleast twice.
+	 */
+	for (i = 0; i < tries; i++) {
+		ret = intel_hdcp_validate_v_prime(connector, shim,
+						  ksv_fifo, num_downstream,
+						  bstatus);
+		if (!ret)
+			break;
+	}
+
+	if (i == tries) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "V Prime validation failed.(%d)\n", ret);
+		goto err;
+	}
+
+	drm_dbg_kms(&dev_priv->drm, "HDCP is enabled (%d downstream devices)\n",
+		    num_downstream);
+	ret = 0;
+err:
+	kfree(ksv_fifo);
+	return ret;
+}
+
+/* Implements Part 1 of the HDCP authorization procedure */
+static int intel_hdcp_auth(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	const struct intel_hdcp_shim *shim = hdcp->shim;
+	enum transcoder cpu_transcoder = connector->hdcp.cpu_transcoder;
+	enum port port = dig_port->base.port;
+	unsigned long r0_prime_gen_start;
+	int ret, i, tries = 2;
+	union {
+		u32 reg[2];
+		u8 shim[DRM_HDCP_AN_LEN];
+	} an;
+	union {
+		u32 reg[2];
+		u8 shim[DRM_HDCP_KSV_LEN];
+	} bksv;
+	union {
+		u32 reg;
+		u8 shim[DRM_HDCP_RI_LEN];
+	} ri;
+	bool repeater_present, hdcp_capable;
+
+	/*
+	 * Detects whether the display is HDCP capable. Although we check for
+	 * valid Bksv below, the HDCP over DP spec requires that we check
+	 * whether the display supports HDCP before we write An. For HDMI
+	 * displays, this is not necessary.
+	 */
+	if (shim->hdcp_capable) {
+		ret = shim->hdcp_capable(dig_port, &hdcp_capable);
+		if (ret)
+			return ret;
+		if (!hdcp_capable) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Panel is not HDCP capable\n");
+			return -EINVAL;
+		}
+	}
+
+	/* Initialize An with 2 random values and acquire it */
+	for (i = 0; i < 2; i++)
+		intel_de_write(dev_priv,
+			       HDCP_ANINIT(dev_priv, cpu_transcoder, port),
+			       get_random_u32());
+	intel_de_write(dev_priv, HDCP_CONF(dev_priv, cpu_transcoder, port),
+		       HDCP_CONF_CAPTURE_AN);
+
+	/* Wait for An to be acquired */
+	if (intel_de_wait_for_set(dev_priv,
+				  HDCP_STATUS(dev_priv, cpu_transcoder, port),
+				  HDCP_STATUS_AN_READY, 1)) {
+		drm_err(&dev_priv->drm, "Timed out waiting for An\n");
+		return -ETIMEDOUT;
+	}
+
+	an.reg[0] = intel_de_read(dev_priv,
+				  HDCP_ANLO(dev_priv, cpu_transcoder, port));
+	an.reg[1] = intel_de_read(dev_priv,
+				  HDCP_ANHI(dev_priv, cpu_transcoder, port));
+	ret = shim->write_an_aksv(dig_port, an.shim);
+	if (ret)
+		return ret;
+
+	r0_prime_gen_start = jiffies;
+
+	memset(&bksv, 0, sizeof(bksv));
+
+	ret = intel_hdcp_read_valid_bksv(dig_port, shim, bksv.shim);
+	if (ret < 0)
+		return ret;
+
+	if (drm_hdcp_check_ksvs_revoked(&dev_priv->drm, bksv.shim, 1) > 0) {
+		drm_err(&dev_priv->drm, "BKSV is revoked\n");
+		return -EPERM;
+	}
+
+	intel_de_write(dev_priv, HDCP_BKSVLO(dev_priv, cpu_transcoder, port),
+		       bksv.reg[0]);
+	intel_de_write(dev_priv, HDCP_BKSVHI(dev_priv, cpu_transcoder, port),
+		       bksv.reg[1]);
+
+	ret = shim->repeater_present(dig_port, &repeater_present);
+	if (ret)
+		return ret;
+	if (repeater_present)
+		intel_de_write(dev_priv, HDCP_REP_CTL,
+			       intel_hdcp_get_repeater_ctl(dev_priv, cpu_transcoder, port));
+
+	ret = shim->toggle_signalling(dig_port, cpu_transcoder, true);
+	if (ret)
+		return ret;
+
+	intel_de_write(dev_priv, HDCP_CONF(dev_priv, cpu_transcoder, port),
+		       HDCP_CONF_AUTH_AND_ENC);
+
+	/* Wait for R0 ready */
+	if (wait_for(intel_de_read(dev_priv, HDCP_STATUS(dev_priv, cpu_transcoder, port)) &
+		     (HDCP_STATUS_R0_READY | HDCP_STATUS_ENC), 1)) {
+		drm_err(&dev_priv->drm, "Timed out waiting for R0 ready\n");
+		return -ETIMEDOUT;
+	}
+
+	/*
+	 * Wait for R0' to become available. The spec says 100ms from Aksv, but
+	 * some monitors can take longer than this. We'll set the timeout at
+	 * 300ms just to be sure.
+	 *
+	 * On DP, there's an R0_READY bit available but no such bit
+	 * exists on HDMI. Since the upper-bound is the same, we'll just do
+	 * the stupid thing instead of polling on one and not the other.
+	 */
+	wait_remaining_ms_from_jiffies(r0_prime_gen_start, 300);
+
+	tries = 3;
+
+	/*
+	 * DP HDCP Spec mandates the two more reattempt to read R0, incase
+	 * of R0 mismatch.
+	 */
+	for (i = 0; i < tries; i++) {
+		ri.reg = 0;
+		ret = shim->read_ri_prime(dig_port, ri.shim);
+		if (ret)
+			return ret;
+		intel_de_write(dev_priv,
+			       HDCP_RPRIME(dev_priv, cpu_transcoder, port),
+			       ri.reg);
+
+		/* Wait for Ri prime match */
+		if (!wait_for(intel_de_read(dev_priv, HDCP_STATUS(dev_priv, cpu_transcoder, port)) &
+			      (HDCP_STATUS_RI_MATCH | HDCP_STATUS_ENC), 1))
+			break;
+	}
+
+	if (i == tries) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Timed out waiting for Ri prime match (%x)\n",
+			    intel_de_read(dev_priv, HDCP_STATUS(dev_priv,
+					  cpu_transcoder, port)));
+		return -ETIMEDOUT;
+	}
+
+	/* Wait for encryption confirmation */
+	if (intel_de_wait_for_set(dev_priv,
+				  HDCP_STATUS(dev_priv, cpu_transcoder, port),
+				  HDCP_STATUS_ENC,
+				  HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS)) {
+		drm_err(&dev_priv->drm, "Timed out waiting for encryption\n");
+		return -ETIMEDOUT;
+	}
+
+	/* DP MST Auth Part 1 Step 2.a and Step 2.b */
+	if (shim->stream_encryption) {
+		ret = shim->stream_encryption(connector, true);
+		if (ret) {
+			drm_err(&dev_priv->drm, "[%s:%d] Failed to enable HDCP 1.4 stream enc\n",
+				connector->base.name, connector->base.base.id);
+			return ret;
+		}
+		drm_dbg_kms(&dev_priv->drm, "HDCP 1.4 transcoder: %s stream encrypted\n",
+			    transcoder_name(hdcp->stream_transcoder));
+	}
+
+	if (repeater_present)
+		return intel_hdcp_auth_downstream(connector);
+
+	drm_dbg_kms(&dev_priv->drm, "HDCP is enabled (no repeater present)\n");
+	return 0;
+}
+
+static int _intel_hdcp_disable(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum port port = dig_port->base.port;
+	enum transcoder cpu_transcoder = hdcp->cpu_transcoder;
+	u32 repeater_ctl;
+	int ret;
+
+	drm_dbg_kms(&dev_priv->drm, "[%s:%d] HDCP is being disabled...\n",
+		    connector->base.name, connector->base.base.id);
+
+	if (hdcp->shim->stream_encryption) {
+		ret = hdcp->shim->stream_encryption(connector, false);
+		if (ret) {
+			drm_err(&dev_priv->drm, "[%s:%d] Failed to disable HDCP 1.4 stream enc\n",
+				connector->base.name, connector->base.base.id);
+			return ret;
+		}
+		drm_dbg_kms(&dev_priv->drm, "HDCP 1.4 transcoder: %s stream encryption disabled\n",
+			    transcoder_name(hdcp->stream_transcoder));
+		/*
+		 * If there are other connectors on this port using HDCP,
+		 * don't disable it until it disabled HDCP encryption for
+		 * all connectors in MST topology.
+		 */
+		if (dig_port->num_hdcp_streams > 0)
+			return 0;
+	}
+
+	hdcp->hdcp_encrypted = false;
+	intel_de_write(dev_priv, HDCP_CONF(dev_priv, cpu_transcoder, port), 0);
+	if (intel_de_wait_for_clear(dev_priv,
+				    HDCP_STATUS(dev_priv, cpu_transcoder, port),
+				    ~0, HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS)) {
+		drm_err(&dev_priv->drm,
+			"Failed to disable HDCP, timeout clearing status\n");
+		return -ETIMEDOUT;
+	}
+
+	repeater_ctl = intel_hdcp_get_repeater_ctl(dev_priv, cpu_transcoder,
+						   port);
+	intel_de_write(dev_priv, HDCP_REP_CTL,
+		       intel_de_read(dev_priv, HDCP_REP_CTL) & ~repeater_ctl);
+
+	ret = hdcp->shim->toggle_signalling(dig_port, cpu_transcoder, false);
+	if (ret) {
+		drm_err(&dev_priv->drm, "Failed to disable HDCP signalling\n");
+		return ret;
+	}
+
+	drm_dbg_kms(&dev_priv->drm, "HDCP is disabled\n");
+	return 0;
+}
+
+static int _intel_hdcp_enable(struct intel_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	int i, ret, tries = 3;
+
+	drm_dbg_kms(&dev_priv->drm, "[%s:%d] HDCP is being enabled...\n",
+		    connector->base.name, connector->base.base.id);
+
+	if (!hdcp_key_loadable(dev_priv)) {
+		drm_err(&dev_priv->drm, "HDCP key Load is not possible\n");
+		return -ENXIO;
+	}
+
+	for (i = 0; i < KEY_LOAD_TRIES; i++) {
+		ret = intel_hdcp_load_keys(dev_priv);
+		if (!ret)
+			break;
+		intel_hdcp_clear_keys(dev_priv);
+	}
+	if (ret) {
+		drm_err(&dev_priv->drm, "Could not load HDCP keys, (%d)\n",
+			ret);
+		return ret;
+	}
+
+	/* Incase of authentication failures, HDCP spec expects reauth. */
+	for (i = 0; i < tries; i++) {
+		ret = intel_hdcp_auth(connector);
+		if (!ret) {
+			hdcp->hdcp_encrypted = true;
+			return 0;
+		}
+
+		drm_dbg_kms(&dev_priv->drm, "HDCP Auth failure (%d)\n", ret);
+
+		/* Ensuring HDCP encryption and signalling are stopped. */
+		_intel_hdcp_disable(connector);
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "HDCP authentication failed (%d tries/%d)\n", tries, ret);
+	return ret;
+}
+
+static struct intel_connector *intel_hdcp_to_connector(struct intel_hdcp *hdcp)
+{
+	return container_of(hdcp, struct intel_connector, hdcp);
+}
+
+static void intel_hdcp_update_value(struct intel_connector *connector,
+				    u64 value, bool update_property)
+{
+	struct drm_device *dev = connector->base.dev;
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+
+	drm_WARN_ON(connector->base.dev, !mutex_is_locked(&hdcp->mutex));
+
+	if (hdcp->value == value)
+		return;
+
+	drm_WARN_ON(dev, !mutex_is_locked(&dig_port->hdcp_mutex));
+
+	if (hdcp->value == DRM_MODE_CONTENT_PROTECTION_ENABLED) {
+		if (!drm_WARN_ON(dev, dig_port->num_hdcp_streams == 0))
+			dig_port->num_hdcp_streams--;
+	} else if (value == DRM_MODE_CONTENT_PROTECTION_ENABLED) {
+		dig_port->num_hdcp_streams++;
+	}
+
+	hdcp->value = value;
+	if (update_property) {
+		drm_connector_get(&connector->base);
+		schedule_work(&hdcp->prop_work);
+	}
+}
+
+/* Implements Part 3 of the HDCP authorization procedure */
+static int intel_hdcp_check_link(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum port port = dig_port->base.port;
+	enum transcoder cpu_transcoder;
+	int ret = 0;
+
+	mutex_lock(&hdcp->mutex);
+	mutex_lock(&dig_port->hdcp_mutex);
+
+	cpu_transcoder = hdcp->cpu_transcoder;
+
+	/* Check_link valid only when HDCP1.4 is enabled */
+	if (hdcp->value != DRM_MODE_CONTENT_PROTECTION_ENABLED ||
+	    !hdcp->hdcp_encrypted) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			!intel_hdcp_in_use(dev_priv, cpu_transcoder, port))) {
+		drm_err(&dev_priv->drm,
+			"%s:%d HDCP link stopped encryption,%x\n",
+			connector->base.name, connector->base.base.id,
+			intel_de_read(dev_priv, HDCP_STATUS(dev_priv, cpu_transcoder, port)));
+		ret = -ENXIO;
+		intel_hdcp_update_value(connector,
+					DRM_MODE_CONTENT_PROTECTION_DESIRED,
+					true);
+		goto out;
+	}
+
+	if (hdcp->shim->check_link(dig_port, connector)) {
+		if (hdcp->value != DRM_MODE_CONTENT_PROTECTION_UNDESIRED) {
+			intel_hdcp_update_value(connector,
+				DRM_MODE_CONTENT_PROTECTION_ENABLED, true);
+		}
+		goto out;
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "[%s:%d] HDCP link failed, retrying authentication\n",
+		    connector->base.name, connector->base.base.id);
+
+	ret = _intel_hdcp_disable(connector);
+	if (ret) {
+		drm_err(&dev_priv->drm, "Failed to disable hdcp (%d)\n", ret);
+		intel_hdcp_update_value(connector,
+					DRM_MODE_CONTENT_PROTECTION_DESIRED,
+					true);
+		goto out;
+	}
+
+	ret = _intel_hdcp_enable(connector);
+	if (ret) {
+		drm_err(&dev_priv->drm, "Failed to enable hdcp (%d)\n", ret);
+		intel_hdcp_update_value(connector,
+					DRM_MODE_CONTENT_PROTECTION_DESIRED,
+					true);
+		goto out;
+	}
+
+out:
+	mutex_unlock(&dig_port->hdcp_mutex);
+	mutex_unlock(&hdcp->mutex);
+	return ret;
+}
+
+static void intel_hdcp_prop_work(struct work_struct *work)
+{
+	struct intel_hdcp *hdcp = container_of(work, struct intel_hdcp,
+					       prop_work);
+	struct intel_connector *connector = intel_hdcp_to_connector(hdcp);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+
+	drm_modeset_lock(&dev_priv->drm.mode_config.connection_mutex, NULL);
+	mutex_lock(&hdcp->mutex);
+
+	/*
+	 * This worker is only used to flip between ENABLED/DESIRED. Either of
+	 * those to UNDESIRED is handled by core. If value == UNDESIRED,
+	 * we're running just after hdcp has been disabled, so just exit
+	 */
+	if (hdcp->value != DRM_MODE_CONTENT_PROTECTION_UNDESIRED)
+		drm_hdcp_update_content_protection(&connector->base,
+						   hdcp->value);
+
+	mutex_unlock(&hdcp->mutex);
+	drm_modeset_unlock(&dev_priv->drm.mode_config.connection_mutex);
+
+	drm_connector_put(&connector->base);
+}
+
+bool is_hdcp_supported(struct drm_i915_private *dev_priv, enum port port)
+{
+	return RUNTIME_INFO(dev_priv)->has_hdcp &&
+		(DISPLAY_VER(dev_priv) >= 12 || port < PORT_E);
+}
+
+static int
+hdcp2_prepare_ake_init(struct intel_connector *connector,
+		       struct hdcp2_ake_init *ake_data)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->initiate_hdcp2_session(comp->mei_dev, data, ake_data);
+	if (ret)
+		drm_dbg_kms(&dev_priv->drm, "Prepare_ake_init failed. %d\n",
+			    ret);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int
+hdcp2_verify_rx_cert_prepare_km(struct intel_connector *connector,
+				struct hdcp2_ake_send_cert *rx_cert,
+				bool *paired,
+				struct hdcp2_ake_no_stored_km *ek_pub_km,
+				size_t *msg_sz)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->verify_receiver_cert_prepare_km(comp->mei_dev, data,
+							 rx_cert, paired,
+							 ek_pub_km, msg_sz);
+	if (ret < 0)
+		drm_dbg_kms(&dev_priv->drm, "Verify rx_cert failed. %d\n",
+			    ret);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int hdcp2_verify_hprime(struct intel_connector *connector,
+			       struct hdcp2_ake_send_hprime *rx_hprime)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->verify_hprime(comp->mei_dev, data, rx_hprime);
+	if (ret < 0)
+		drm_dbg_kms(&dev_priv->drm, "Verify hprime failed. %d\n", ret);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int
+hdcp2_store_pairing_info(struct intel_connector *connector,
+			 struct hdcp2_ake_send_pairing_info *pairing_info)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->store_pairing_info(comp->mei_dev, data, pairing_info);
+	if (ret < 0)
+		drm_dbg_kms(&dev_priv->drm, "Store pairing info failed. %d\n",
+			    ret);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int
+hdcp2_prepare_lc_init(struct intel_connector *connector,
+		      struct hdcp2_lc_init *lc_init)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->initiate_locality_check(comp->mei_dev, data, lc_init);
+	if (ret < 0)
+		drm_dbg_kms(&dev_priv->drm, "Prepare lc_init failed. %d\n",
+			    ret);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int
+hdcp2_verify_lprime(struct intel_connector *connector,
+		    struct hdcp2_lc_send_lprime *rx_lprime)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->verify_lprime(comp->mei_dev, data, rx_lprime);
+	if (ret < 0)
+		drm_dbg_kms(&dev_priv->drm, "Verify L_Prime failed. %d\n",
+			    ret);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int hdcp2_prepare_skey(struct intel_connector *connector,
+			      struct hdcp2_ske_send_eks *ske_data)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->get_session_key(comp->mei_dev, data, ske_data);
+	if (ret < 0)
+		drm_dbg_kms(&dev_priv->drm, "Get session key failed. %d\n",
+			    ret);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int
+hdcp2_verify_rep_topology_prepare_ack(struct intel_connector *connector,
+				      struct hdcp2_rep_send_receiverid_list
+								*rep_topology,
+				      struct hdcp2_rep_send_ack *rep_send_ack)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->repeater_check_flow_prepare_ack(comp->mei_dev, data,
+							 rep_topology,
+							 rep_send_ack);
+	if (ret < 0)
+		drm_dbg_kms(&dev_priv->drm,
+			    "Verify rep topology failed. %d\n", ret);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int
+hdcp2_verify_mprime(struct intel_connector *connector,
+		    struct hdcp2_rep_stream_ready *stream_ready)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->verify_mprime(comp->mei_dev, data, stream_ready);
+	if (ret < 0)
+		drm_dbg_kms(&dev_priv->drm, "Verify mprime failed. %d\n", ret);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int hdcp2_authenticate_port(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->enable_hdcp_authentication(comp->mei_dev, data);
+	if (ret < 0)
+		drm_dbg_kms(&dev_priv->drm, "Enable hdcp auth failed. %d\n",
+			    ret);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int hdcp2_close_mei_session(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct i915_hdcp_comp_master *comp;
+	int ret;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	comp = dev_priv->display.hdcp.master;
+
+	if (!comp || !comp->ops) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return -EINVAL;
+	}
+
+	ret = comp->ops->close_hdcp_session(comp->mei_dev,
+					     &dig_port->hdcp_port_data);
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return ret;
+}
+
+static int hdcp2_deauthenticate_port(struct intel_connector *connector)
+{
+	return hdcp2_close_mei_session(connector);
+}
+
+/* Authentication flow starts from here */
+static int hdcp2_authentication_key_exchange(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	union {
+		struct hdcp2_ake_init ake_init;
+		struct hdcp2_ake_send_cert send_cert;
+		struct hdcp2_ake_no_stored_km no_stored_km;
+		struct hdcp2_ake_send_hprime send_hprime;
+		struct hdcp2_ake_send_pairing_info pairing_info;
+	} msgs;
+	const struct intel_hdcp_shim *shim = hdcp->shim;
+	size_t size;
+	int ret;
+
+	/* Init for seq_num */
+	hdcp->seq_num_v = 0;
+	hdcp->seq_num_m = 0;
+
+	ret = hdcp2_prepare_ake_init(connector, &msgs.ake_init);
+	if (ret < 0)
+		return ret;
+
+	ret = shim->write_2_2_msg(dig_port, &msgs.ake_init,
+				  sizeof(msgs.ake_init));
+	if (ret < 0)
+		return ret;
+
+	ret = shim->read_2_2_msg(dig_port, HDCP_2_2_AKE_SEND_CERT,
+				 &msgs.send_cert, sizeof(msgs.send_cert));
+	if (ret < 0)
+		return ret;
+
+	if (msgs.send_cert.rx_caps[0] != HDCP_2_2_RX_CAPS_VERSION_VAL) {
+		drm_dbg_kms(&dev_priv->drm, "cert.rx_caps dont claim HDCP2.2\n");
+		return -EINVAL;
+	}
+
+	hdcp->is_repeater = HDCP_2_2_RX_REPEATER(msgs.send_cert.rx_caps[2]);
+
+	if (drm_hdcp_check_ksvs_revoked(&dev_priv->drm,
+					msgs.send_cert.cert_rx.receiver_id,
+					1) > 0) {
+		drm_err(&dev_priv->drm, "Receiver ID is revoked\n");
+		return -EPERM;
+	}
+
+	/*
+	 * Here msgs.no_stored_km will hold msgs corresponding to the km
+	 * stored also.
+	 */
+	ret = hdcp2_verify_rx_cert_prepare_km(connector, &msgs.send_cert,
+					      &hdcp->is_paired,
+					      &msgs.no_stored_km, &size);
+	if (ret < 0)
+		return ret;
+
+	ret = shim->write_2_2_msg(dig_port, &msgs.no_stored_km, size);
+	if (ret < 0)
+		return ret;
+
+	ret = shim->read_2_2_msg(dig_port, HDCP_2_2_AKE_SEND_HPRIME,
+				 &msgs.send_hprime, sizeof(msgs.send_hprime));
+	if (ret < 0)
+		return ret;
+
+	ret = hdcp2_verify_hprime(connector, &msgs.send_hprime);
+	if (ret < 0)
+		return ret;
+
+	if (!hdcp->is_paired) {
+		/* Pairing is required */
+		ret = shim->read_2_2_msg(dig_port,
+					 HDCP_2_2_AKE_SEND_PAIRING_INFO,
+					 &msgs.pairing_info,
+					 sizeof(msgs.pairing_info));
+		if (ret < 0)
+			return ret;
+
+		ret = hdcp2_store_pairing_info(connector, &msgs.pairing_info);
+		if (ret < 0)
+			return ret;
+		hdcp->is_paired = true;
+	}
+
+	return 0;
+}
+
+static int hdcp2_locality_check(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	union {
+		struct hdcp2_lc_init lc_init;
+		struct hdcp2_lc_send_lprime send_lprime;
+	} msgs;
+	const struct intel_hdcp_shim *shim = hdcp->shim;
+	int tries = HDCP2_LC_RETRY_CNT, ret, i;
+
+	for (i = 0; i < tries; i++) {
+		ret = hdcp2_prepare_lc_init(connector, &msgs.lc_init);
+		if (ret < 0)
+			continue;
+
+		ret = shim->write_2_2_msg(dig_port, &msgs.lc_init,
+				      sizeof(msgs.lc_init));
+		if (ret < 0)
+			continue;
+
+		ret = shim->read_2_2_msg(dig_port,
+					 HDCP_2_2_LC_SEND_LPRIME,
+					 &msgs.send_lprime,
+					 sizeof(msgs.send_lprime));
+		if (ret < 0)
+			continue;
+
+		ret = hdcp2_verify_lprime(connector, &msgs.send_lprime);
+		if (!ret)
+			break;
+	}
+
+	return ret;
+}
+
+static int hdcp2_session_key_exchange(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	struct hdcp2_ske_send_eks send_eks;
+	int ret;
+
+	ret = hdcp2_prepare_skey(connector, &send_eks);
+	if (ret < 0)
+		return ret;
+
+	ret = hdcp->shim->write_2_2_msg(dig_port, &send_eks,
+					sizeof(send_eks));
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+static
+int _hdcp2_propagate_stream_management_info(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	union {
+		struct hdcp2_rep_stream_manage stream_manage;
+		struct hdcp2_rep_stream_ready stream_ready;
+	} msgs;
+	const struct intel_hdcp_shim *shim = hdcp->shim;
+	int ret, streams_size_delta, i;
+
+	if (connector->hdcp.seq_num_m > HDCP_2_2_SEQ_NUM_MAX)
+		return -ERANGE;
+
+	/* Prepare RepeaterAuth_Stream_Manage msg */
+	msgs.stream_manage.msg_id = HDCP_2_2_REP_STREAM_MANAGE;
+	drm_hdcp_cpu_to_be24(msgs.stream_manage.seq_num_m, hdcp->seq_num_m);
+
+	msgs.stream_manage.k = cpu_to_be16(data->k);
+
+	for (i = 0; i < data->k; i++) {
+		msgs.stream_manage.streams[i].stream_id = data->streams[i].stream_id;
+		msgs.stream_manage.streams[i].stream_type = data->streams[i].stream_type;
+	}
+
+	streams_size_delta = (HDCP_2_2_MAX_CONTENT_STREAMS_CNT - data->k) *
+				sizeof(struct hdcp2_streamid_type);
+	/* Send it to Repeater */
+	ret = shim->write_2_2_msg(dig_port, &msgs.stream_manage,
+				  sizeof(msgs.stream_manage) - streams_size_delta);
+	if (ret < 0)
+		goto out;
+
+	ret = shim->read_2_2_msg(dig_port, HDCP_2_2_REP_STREAM_READY,
+				 &msgs.stream_ready, sizeof(msgs.stream_ready));
+	if (ret < 0)
+		goto out;
+
+	data->seq_num_m = hdcp->seq_num_m;
+
+	ret = hdcp2_verify_mprime(connector, &msgs.stream_ready);
+
+out:
+	hdcp->seq_num_m++;
+
+	return ret;
+}
+
+static
+int hdcp2_authenticate_repeater_topology(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	union {
+		struct hdcp2_rep_send_receiverid_list recvid_list;
+		struct hdcp2_rep_send_ack rep_ack;
+	} msgs;
+	const struct intel_hdcp_shim *shim = hdcp->shim;
+	u32 seq_num_v, device_cnt;
+	u8 *rx_info;
+	int ret;
+
+	ret = shim->read_2_2_msg(dig_port, HDCP_2_2_REP_SEND_RECVID_LIST,
+				 &msgs.recvid_list, sizeof(msgs.recvid_list));
+	if (ret < 0)
+		return ret;
+
+	rx_info = msgs.recvid_list.rx_info;
+
+	if (HDCP_2_2_MAX_CASCADE_EXCEEDED(rx_info[1]) ||
+	    HDCP_2_2_MAX_DEVS_EXCEEDED(rx_info[1])) {
+		drm_dbg_kms(&dev_priv->drm, "Topology Max Size Exceeded\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * MST topology is not Type 1 capable if it contains a downstream
+	 * device that is only HDCP 1.x or Legacy HDCP 2.0/2.1 compliant.
+	 */
+	dig_port->hdcp_mst_type1_capable =
+		!HDCP_2_2_HDCP1_DEVICE_CONNECTED(rx_info[1]) &&
+		!HDCP_2_2_HDCP_2_0_REP_CONNECTED(rx_info[1]);
+
+	/* Converting and Storing the seq_num_v to local variable as DWORD */
+	seq_num_v =
+		drm_hdcp_be24_to_cpu((const u8 *)msgs.recvid_list.seq_num_v);
+
+	if (!hdcp->hdcp2_encrypted && seq_num_v) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Non zero Seq_num_v at first RecvId_List msg\n");
+		return -EINVAL;
+	}
+
+	if (seq_num_v < hdcp->seq_num_v) {
+		/* Roll over of the seq_num_v from repeater. Reauthenticate. */
+		drm_dbg_kms(&dev_priv->drm, "Seq_num_v roll over.\n");
+		return -EINVAL;
+	}
+
+	device_cnt = (HDCP_2_2_DEV_COUNT_HI(rx_info[0]) << 4 |
+		      HDCP_2_2_DEV_COUNT_LO(rx_info[1]));
+	if (drm_hdcp_check_ksvs_revoked(&dev_priv->drm,
+					msgs.recvid_list.receiver_ids,
+					device_cnt) > 0) {
+		drm_err(&dev_priv->drm, "Revoked receiver ID(s) is in list\n");
+		return -EPERM;
+	}
+
+	ret = hdcp2_verify_rep_topology_prepare_ack(connector,
+						    &msgs.recvid_list,
+						    &msgs.rep_ack);
+	if (ret < 0)
+		return ret;
+
+	hdcp->seq_num_v = seq_num_v;
+	ret = shim->write_2_2_msg(dig_port, &msgs.rep_ack,
+				  sizeof(msgs.rep_ack));
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+static int hdcp2_authenticate_sink(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	const struct intel_hdcp_shim *shim = hdcp->shim;
+	int ret;
+
+	ret = hdcp2_authentication_key_exchange(connector);
+	if (ret < 0) {
+		drm_dbg_kms(&i915->drm, "AKE Failed. Err : %d\n", ret);
+		return ret;
+	}
+
+	ret = hdcp2_locality_check(connector);
+	if (ret < 0) {
+		drm_dbg_kms(&i915->drm,
+			    "Locality Check failed. Err : %d\n", ret);
+		return ret;
+	}
+
+	ret = hdcp2_session_key_exchange(connector);
+	if (ret < 0) {
+		drm_dbg_kms(&i915->drm, "SKE Failed. Err : %d\n", ret);
+		return ret;
+	}
+
+	if (shim->config_stream_type) {
+		ret = shim->config_stream_type(dig_port,
+					       hdcp->is_repeater,
+					       hdcp->content_type);
+		if (ret < 0)
+			return ret;
+	}
+
+	if (hdcp->is_repeater) {
+		ret = hdcp2_authenticate_repeater_topology(connector);
+		if (ret < 0) {
+			drm_dbg_kms(&i915->drm,
+				    "Repeater Auth Failed. Err: %d\n", ret);
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+static int hdcp2_enable_stream_encryption(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum transcoder cpu_transcoder = hdcp->cpu_transcoder;
+	enum port port = dig_port->base.port;
+	int ret = 0;
+
+	if (!(intel_de_read(dev_priv, HDCP2_STATUS(dev_priv, cpu_transcoder, port)) &
+			    LINK_ENCRYPTION_STATUS)) {
+		drm_err(&dev_priv->drm, "[%s:%d] HDCP 2.2 Link is not encrypted\n",
+			connector->base.name, connector->base.base.id);
+		ret = -EPERM;
+		goto link_recover;
+	}
+
+	if (hdcp->shim->stream_2_2_encryption) {
+		ret = hdcp->shim->stream_2_2_encryption(connector, true);
+		if (ret) {
+			drm_err(&dev_priv->drm, "[%s:%d] Failed to enable HDCP 2.2 stream enc\n",
+				connector->base.name, connector->base.base.id);
+			return ret;
+		}
+		drm_dbg_kms(&dev_priv->drm, "HDCP 2.2 transcoder: %s stream encrypted\n",
+			    transcoder_name(hdcp->stream_transcoder));
+	}
+
+	return 0;
+
+link_recover:
+	if (hdcp2_deauthenticate_port(connector) < 0)
+		drm_dbg_kms(&dev_priv->drm, "Port deauth failed.\n");
+
+	dig_port->hdcp_auth_status = false;
+	data->k = 0;
+
+	return ret;
+}
+
+static int hdcp2_enable_encryption(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum port port = dig_port->base.port;
+	enum transcoder cpu_transcoder = hdcp->cpu_transcoder;
+	int ret;
+
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, HDCP2_STATUS(dev_priv, cpu_transcoder, port)) &
+		    LINK_ENCRYPTION_STATUS);
+	if (hdcp->shim->toggle_signalling) {
+		ret = hdcp->shim->toggle_signalling(dig_port, cpu_transcoder,
+						    true);
+		if (ret) {
+			drm_err(&dev_priv->drm,
+				"Failed to enable HDCP signalling. %d\n",
+				ret);
+			return ret;
+		}
+	}
+
+	if (intel_de_read(dev_priv, HDCP2_STATUS(dev_priv, cpu_transcoder, port)) &
+	    LINK_AUTH_STATUS) {
+		/* Link is Authenticated. Now set for Encryption */
+		intel_de_write(dev_priv,
+			       HDCP2_CTL(dev_priv, cpu_transcoder, port),
+			       intel_de_read(dev_priv, HDCP2_CTL(dev_priv, cpu_transcoder, port)) | CTL_LINK_ENCRYPTION_REQ);
+	}
+
+	ret = intel_de_wait_for_set(dev_priv,
+				    HDCP2_STATUS(dev_priv, cpu_transcoder,
+						 port),
+				    LINK_ENCRYPTION_STATUS,
+				    HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS);
+	dig_port->hdcp_auth_status = true;
+
+	return ret;
+}
+
+static int hdcp2_disable_encryption(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum port port = dig_port->base.port;
+	enum transcoder cpu_transcoder = hdcp->cpu_transcoder;
+	int ret;
+
+	drm_WARN_ON(&dev_priv->drm, !(intel_de_read(dev_priv, HDCP2_STATUS(dev_priv, cpu_transcoder, port)) &
+				      LINK_ENCRYPTION_STATUS));
+
+	intel_de_write(dev_priv, HDCP2_CTL(dev_priv, cpu_transcoder, port),
+		       intel_de_read(dev_priv, HDCP2_CTL(dev_priv, cpu_transcoder, port)) & ~CTL_LINK_ENCRYPTION_REQ);
+
+	ret = intel_de_wait_for_clear(dev_priv,
+				      HDCP2_STATUS(dev_priv, cpu_transcoder,
+						   port),
+				      LINK_ENCRYPTION_STATUS,
+				      HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS);
+	if (ret == -ETIMEDOUT)
+		drm_dbg_kms(&dev_priv->drm, "Disable Encryption Timedout");
+
+	if (hdcp->shim->toggle_signalling) {
+		ret = hdcp->shim->toggle_signalling(dig_port, cpu_transcoder,
+						    false);
+		if (ret) {
+			drm_err(&dev_priv->drm,
+				"Failed to disable HDCP signalling. %d\n",
+				ret);
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+static int
+hdcp2_propagate_stream_management_info(struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	int i, tries = 3, ret;
+
+	if (!connector->hdcp.is_repeater)
+		return 0;
+
+	for (i = 0; i < tries; i++) {
+		ret = _hdcp2_propagate_stream_management_info(connector);
+		if (!ret)
+			break;
+
+		/* Lets restart the auth incase of seq_num_m roll over */
+		if (connector->hdcp.seq_num_m > HDCP_2_2_SEQ_NUM_MAX) {
+			drm_dbg_kms(&i915->drm,
+				    "seq_num_m roll over.(%d)\n", ret);
+			break;
+		}
+
+		drm_dbg_kms(&i915->drm,
+			    "HDCP2 stream management %d of %d Failed.(%d)\n",
+			    i + 1, tries, ret);
+	}
+
+	return ret;
+}
+
+static int hdcp2_authenticate_and_encrypt(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	int ret = 0, i, tries = 3;
+
+	for (i = 0; i < tries && !dig_port->hdcp_auth_status; i++) {
+		ret = hdcp2_authenticate_sink(connector);
+		if (!ret) {
+			ret = intel_hdcp_prepare_streams(connector);
+			if (ret) {
+				drm_dbg_kms(&i915->drm,
+					    "Prepare streams failed.(%d)\n",
+					    ret);
+				break;
+			}
+
+			ret = hdcp2_propagate_stream_management_info(connector);
+			if (ret) {
+				drm_dbg_kms(&i915->drm,
+					    "Stream management failed.(%d)\n",
+					    ret);
+				break;
+			}
+
+			ret = hdcp2_authenticate_port(connector);
+			if (!ret)
+				break;
+			drm_dbg_kms(&i915->drm, "HDCP2 port auth failed.(%d)\n",
+				    ret);
+		}
+
+		/* Clearing the mei hdcp session */
+		drm_dbg_kms(&i915->drm, "HDCP2.2 Auth %d of %d Failed.(%d)\n",
+			    i + 1, tries, ret);
+		if (hdcp2_deauthenticate_port(connector) < 0)
+			drm_dbg_kms(&i915->drm, "Port deauth failed.\n");
+	}
+
+	if (!ret && !dig_port->hdcp_auth_status) {
+		/*
+		 * Ensuring the required 200mSec min time interval between
+		 * Session Key Exchange and encryption.
+		 */
+		msleep(HDCP_2_2_DELAY_BEFORE_ENCRYPTION_EN);
+		ret = hdcp2_enable_encryption(connector);
+		if (ret < 0) {
+			drm_dbg_kms(&i915->drm,
+				    "Encryption Enable Failed.(%d)\n", ret);
+			if (hdcp2_deauthenticate_port(connector) < 0)
+				drm_dbg_kms(&i915->drm, "Port deauth failed.\n");
+		}
+	}
+
+	if (!ret)
+		ret = hdcp2_enable_stream_encryption(connector);
+
+	return ret;
+}
+
+static int _intel_hdcp2_enable(struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	int ret;
+
+	drm_dbg_kms(&i915->drm, "[%s:%d] HDCP2.2 is being enabled. Type: %d\n",
+		    connector->base.name, connector->base.base.id,
+		    hdcp->content_type);
+
+	ret = hdcp2_authenticate_and_encrypt(connector);
+	if (ret) {
+		drm_dbg_kms(&i915->drm, "HDCP2 Type%d  Enabling Failed. (%d)\n",
+			    hdcp->content_type, ret);
+		return ret;
+	}
+
+	drm_dbg_kms(&i915->drm, "[%s:%d] HDCP2.2 is enabled. Type %d\n",
+		    connector->base.name, connector->base.base.id,
+		    hdcp->content_type);
+
+	hdcp->hdcp2_encrypted = true;
+	return 0;
+}
+
+static int
+_intel_hdcp2_disable(struct intel_connector *connector, bool hdcp2_link_recovery)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	int ret;
+
+	drm_dbg_kms(&i915->drm, "[%s:%d] HDCP2.2 is being Disabled\n",
+		    connector->base.name, connector->base.base.id);
+
+	if (hdcp->shim->stream_2_2_encryption) {
+		ret = hdcp->shim->stream_2_2_encryption(connector, false);
+		if (ret) {
+			drm_err(&i915->drm, "[%s:%d] Failed to disable HDCP 2.2 stream enc\n",
+				connector->base.name, connector->base.base.id);
+			return ret;
+		}
+		drm_dbg_kms(&i915->drm, "HDCP 2.2 transcoder: %s stream encryption disabled\n",
+			    transcoder_name(hdcp->stream_transcoder));
+
+		if (dig_port->num_hdcp_streams > 0 && !hdcp2_link_recovery)
+			return 0;
+	}
+
+	ret = hdcp2_disable_encryption(connector);
+
+	if (hdcp2_deauthenticate_port(connector) < 0)
+		drm_dbg_kms(&i915->drm, "Port deauth failed.\n");
+
+	connector->hdcp.hdcp2_encrypted = false;
+	dig_port->hdcp_auth_status = false;
+	data->k = 0;
+
+	return ret;
+}
+
+/* Implements the Link Integrity Check for HDCP2.2 */
+static int intel_hdcp2_check_link(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum port port = dig_port->base.port;
+	enum transcoder cpu_transcoder;
+	int ret = 0;
+
+	mutex_lock(&hdcp->mutex);
+	mutex_lock(&dig_port->hdcp_mutex);
+	cpu_transcoder = hdcp->cpu_transcoder;
+
+	/* hdcp2_check_link is expected only when HDCP2.2 is Enabled */
+	if (hdcp->value != DRM_MODE_CONTENT_PROTECTION_ENABLED ||
+	    !hdcp->hdcp2_encrypted) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			!intel_hdcp2_in_use(dev_priv, cpu_transcoder, port))) {
+		drm_err(&dev_priv->drm,
+			"HDCP2.2 link stopped the encryption, %x\n",
+			intel_de_read(dev_priv, HDCP2_STATUS(dev_priv, cpu_transcoder, port)));
+		ret = -ENXIO;
+		_intel_hdcp2_disable(connector, true);
+		intel_hdcp_update_value(connector,
+					DRM_MODE_CONTENT_PROTECTION_DESIRED,
+					true);
+		goto out;
+	}
+
+	ret = hdcp->shim->check_2_2_link(dig_port, connector);
+	if (ret == HDCP_LINK_PROTECTED) {
+		if (hdcp->value != DRM_MODE_CONTENT_PROTECTION_UNDESIRED) {
+			intel_hdcp_update_value(connector,
+					DRM_MODE_CONTENT_PROTECTION_ENABLED,
+					true);
+		}
+		goto out;
+	}
+
+	if (ret == HDCP_TOPOLOGY_CHANGE) {
+		if (hdcp->value == DRM_MODE_CONTENT_PROTECTION_UNDESIRED)
+			goto out;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "HDCP2.2 Downstream topology change\n");
+		ret = hdcp2_authenticate_repeater_topology(connector);
+		if (!ret) {
+			intel_hdcp_update_value(connector,
+					DRM_MODE_CONTENT_PROTECTION_ENABLED,
+					true);
+			goto out;
+		}
+		drm_dbg_kms(&dev_priv->drm,
+			    "[%s:%d] Repeater topology auth failed.(%d)\n",
+			    connector->base.name, connector->base.base.id,
+			    ret);
+	} else {
+		drm_dbg_kms(&dev_priv->drm,
+			    "[%s:%d] HDCP2.2 link failed, retrying auth\n",
+			    connector->base.name, connector->base.base.id);
+	}
+
+	ret = _intel_hdcp2_disable(connector, true);
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"[%s:%d] Failed to disable hdcp2.2 (%d)\n",
+			connector->base.name, connector->base.base.id, ret);
+		intel_hdcp_update_value(connector,
+				DRM_MODE_CONTENT_PROTECTION_DESIRED, true);
+		goto out;
+	}
+
+	ret = _intel_hdcp2_enable(connector);
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "[%s:%d] Failed to enable hdcp2.2 (%d)\n",
+			    connector->base.name, connector->base.base.id,
+			    ret);
+		intel_hdcp_update_value(connector,
+					DRM_MODE_CONTENT_PROTECTION_DESIRED,
+					true);
+		goto out;
+	}
+
+out:
+	mutex_unlock(&dig_port->hdcp_mutex);
+	mutex_unlock(&hdcp->mutex);
+	return ret;
+}
+
+static void intel_hdcp_check_work(struct work_struct *work)
+{
+	struct intel_hdcp *hdcp = container_of(to_delayed_work(work),
+					       struct intel_hdcp,
+					       check_work);
+	struct intel_connector *connector = intel_hdcp_to_connector(hdcp);
+
+	if (drm_connector_is_unregistered(&connector->base))
+		return;
+
+	if (!intel_hdcp2_check_link(connector))
+		schedule_delayed_work(&hdcp->check_work,
+				      DRM_HDCP2_CHECK_PERIOD_MS);
+	else if (!intel_hdcp_check_link(connector))
+		schedule_delayed_work(&hdcp->check_work,
+				      DRM_HDCP_CHECK_PERIOD_MS);
+}
+
+static int i915_hdcp_component_bind(struct device *i915_kdev,
+				    struct device *mei_kdev, void *data)
+{
+	struct drm_i915_private *dev_priv = kdev_to_i915(i915_kdev);
+
+	drm_dbg(&dev_priv->drm, "I915 HDCP comp bind\n");
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	dev_priv->display.hdcp.master = (struct i915_hdcp_comp_master *)data;
+	dev_priv->display.hdcp.master->mei_dev = mei_kdev;
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	return 0;
+}
+
+static void i915_hdcp_component_unbind(struct device *i915_kdev,
+				       struct device *mei_kdev, void *data)
+{
+	struct drm_i915_private *dev_priv = kdev_to_i915(i915_kdev);
+
+	drm_dbg(&dev_priv->drm, "I915 HDCP comp unbind\n");
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	dev_priv->display.hdcp.master = NULL;
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+}
+
+static const struct component_ops i915_hdcp_component_ops = {
+	.bind   = i915_hdcp_component_bind,
+	.unbind = i915_hdcp_component_unbind,
+};
+
+static enum mei_fw_ddi intel_get_mei_fw_ddi_index(enum port port)
+{
+	switch (port) {
+	case PORT_A:
+		return MEI_DDI_A;
+	case PORT_B ... PORT_F:
+		return (enum mei_fw_ddi)port;
+	default:
+		return MEI_DDI_INVALID_PORT;
+	}
+}
+
+static enum mei_fw_tc intel_get_mei_fw_tc(enum transcoder cpu_transcoder)
+{
+	switch (cpu_transcoder) {
+	case TRANSCODER_A ... TRANSCODER_D:
+		return (enum mei_fw_tc)(cpu_transcoder | 0x10);
+	default: /* eDP, DSI TRANSCODERS are non HDCP capable */
+		return MEI_INVALID_TRANSCODER;
+	}
+}
+
+static int initialize_hdcp_port_data(struct intel_connector *connector,
+				     struct intel_digital_port *dig_port,
+				     const struct intel_hdcp_shim *shim)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct hdcp_port_data *data = &dig_port->hdcp_port_data;
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	enum port port = dig_port->base.port;
+
+	if (DISPLAY_VER(dev_priv) < 12)
+		data->fw_ddi = intel_get_mei_fw_ddi_index(port);
+	else
+		/*
+		 * As per ME FW API expectation, for GEN 12+, fw_ddi is filled
+		 * with zero(INVALID PORT index).
+		 */
+		data->fw_ddi = MEI_DDI_INVALID_PORT;
+
+	/*
+	 * As associated transcoder is set and modified at modeset, here fw_tc
+	 * is initialized to zero (invalid transcoder index). This will be
+	 * retained for <Gen12 forever.
+	 */
+	data->fw_tc = MEI_INVALID_TRANSCODER;
+
+	data->port_type = (u8)HDCP_PORT_TYPE_INTEGRATED;
+	data->protocol = (u8)shim->protocol;
+
+	if (!data->streams)
+		data->streams = kcalloc(INTEL_NUM_PIPES(dev_priv),
+					sizeof(struct hdcp2_streamid_type),
+					GFP_KERNEL);
+	if (!data->streams) {
+		drm_err(&dev_priv->drm, "Out of Memory\n");
+		return -ENOMEM;
+	}
+	/* For SST */
+	data->streams[0].stream_id = 0;
+	data->streams[0].stream_type = hdcp->content_type;
+
+	return 0;
+}
+
+static bool is_hdcp2_supported(struct drm_i915_private *dev_priv)
+{
+	if (!IS_ENABLED(CONFIG_INTEL_MEI_HDCP))
+		return false;
+
+	return (DISPLAY_VER(dev_priv) >= 10 ||
+		IS_KABYLAKE(dev_priv) ||
+		IS_COFFEELAKE(dev_priv) ||
+		IS_COMETLAKE(dev_priv));
+}
+
+void intel_hdcp_component_init(struct drm_i915_private *dev_priv)
+{
+	int ret;
+
+	if (!is_hdcp2_supported(dev_priv))
+		return;
+
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	drm_WARN_ON(&dev_priv->drm, dev_priv->display.hdcp.comp_added);
+
+	dev_priv->display.hdcp.comp_added = true;
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+	ret = component_add_typed(dev_priv->drm.dev, &i915_hdcp_component_ops,
+				  I915_COMPONENT_HDCP);
+	if (ret < 0) {
+		drm_dbg_kms(&dev_priv->drm, "Failed at component add(%d)\n",
+			    ret);
+		mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+		dev_priv->display.hdcp.comp_added = false;
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return;
+	}
+}
+
+static void intel_hdcp2_init(struct intel_connector *connector,
+			     struct intel_digital_port *dig_port,
+			     const struct intel_hdcp_shim *shim)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	int ret;
+
+	ret = initialize_hdcp_port_data(connector, dig_port, shim);
+	if (ret) {
+		drm_dbg_kms(&i915->drm, "Mei hdcp data init failed\n");
+		return;
+	}
+
+	hdcp->hdcp2_supported = true;
+}
+
+int intel_hdcp_init(struct intel_connector *connector,
+		    struct intel_digital_port *dig_port,
+		    const struct intel_hdcp_shim *shim)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	int ret;
+
+	if (!shim)
+		return -EINVAL;
+
+	if (is_hdcp2_supported(dev_priv))
+		intel_hdcp2_init(connector, dig_port, shim);
+
+	ret =
+	drm_connector_attach_content_protection_property(&connector->base,
+							 hdcp->hdcp2_supported);
+	if (ret) {
+		hdcp->hdcp2_supported = false;
+		kfree(dig_port->hdcp_port_data.streams);
+		return ret;
+	}
+
+	hdcp->shim = shim;
+	mutex_init(&hdcp->mutex);
+	INIT_DELAYED_WORK(&hdcp->check_work, intel_hdcp_check_work);
+	INIT_WORK(&hdcp->prop_work, intel_hdcp_prop_work);
+	init_waitqueue_head(&hdcp->cp_irq_queue);
+
+	return 0;
+}
+
+int intel_hdcp_enable(struct intel_connector *connector,
+		      const struct intel_crtc_state *pipe_config, u8 content_type)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	unsigned long check_link_interval = DRM_HDCP_CHECK_PERIOD_MS;
+	int ret = -EINVAL;
+
+	if (!hdcp->shim)
+		return -ENOENT;
+
+	if (!connector->encoder) {
+		drm_err(&dev_priv->drm, "[%s:%d] encoder is not initialized\n",
+			connector->base.name, connector->base.base.id);
+		return -ENODEV;
+	}
+
+	mutex_lock(&hdcp->mutex);
+	mutex_lock(&dig_port->hdcp_mutex);
+	drm_WARN_ON(&dev_priv->drm,
+		    hdcp->value == DRM_MODE_CONTENT_PROTECTION_ENABLED);
+	hdcp->content_type = content_type;
+
+	if (intel_crtc_has_type(pipe_config, INTEL_OUTPUT_DP_MST)) {
+		hdcp->cpu_transcoder = pipe_config->mst_master_transcoder;
+		hdcp->stream_transcoder = pipe_config->cpu_transcoder;
+	} else {
+		hdcp->cpu_transcoder = pipe_config->cpu_transcoder;
+		hdcp->stream_transcoder = INVALID_TRANSCODER;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		dig_port->hdcp_port_data.fw_tc = intel_get_mei_fw_tc(hdcp->cpu_transcoder);
+
+	/*
+	 * Considering that HDCP2.2 is more secure than HDCP1.4, If the setup
+	 * is capable of HDCP2.2, it is preferred to use HDCP2.2.
+	 */
+	if (intel_hdcp2_capable(connector)) {
+		ret = _intel_hdcp2_enable(connector);
+		if (!ret)
+			check_link_interval = DRM_HDCP2_CHECK_PERIOD_MS;
+	}
+
+	/*
+	 * When HDCP2.2 fails and Content Type is not Type1, HDCP1.4 will
+	 * be attempted.
+	 */
+	if (ret && intel_hdcp_capable(connector) &&
+	    hdcp->content_type != DRM_MODE_HDCP_CONTENT_TYPE1) {
+		ret = _intel_hdcp_enable(connector);
+	}
+
+	if (!ret) {
+		schedule_delayed_work(&hdcp->check_work, check_link_interval);
+		intel_hdcp_update_value(connector,
+					DRM_MODE_CONTENT_PROTECTION_ENABLED,
+					true);
+	}
+
+	mutex_unlock(&dig_port->hdcp_mutex);
+	mutex_unlock(&hdcp->mutex);
+	return ret;
+}
+
+int intel_hdcp_disable(struct intel_connector *connector)
+{
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	int ret = 0;
+
+	if (!hdcp->shim)
+		return -ENOENT;
+
+	mutex_lock(&hdcp->mutex);
+	mutex_lock(&dig_port->hdcp_mutex);
+
+	if (hdcp->value == DRM_MODE_CONTENT_PROTECTION_UNDESIRED)
+		goto out;
+
+	intel_hdcp_update_value(connector,
+				DRM_MODE_CONTENT_PROTECTION_UNDESIRED, false);
+	if (hdcp->hdcp2_encrypted)
+		ret = _intel_hdcp2_disable(connector, false);
+	else if (hdcp->hdcp_encrypted)
+		ret = _intel_hdcp_disable(connector);
+
+out:
+	mutex_unlock(&dig_port->hdcp_mutex);
+	mutex_unlock(&hdcp->mutex);
+	cancel_delayed_work_sync(&hdcp->check_work);
+	return ret;
+}
+
+void intel_hdcp_update_pipe(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    const struct drm_connector_state *conn_state)
+{
+	struct intel_connector *connector =
+				to_intel_connector(conn_state->connector);
+	struct intel_hdcp *hdcp = &connector->hdcp;
+	bool content_protection_type_changed, desired_and_not_enabled = false;
+
+	if (!connector->hdcp.shim)
+		return;
+
+	content_protection_type_changed =
+		(conn_state->hdcp_content_type != hdcp->content_type &&
+		 conn_state->content_protection !=
+		 DRM_MODE_CONTENT_PROTECTION_UNDESIRED);
+
+	/*
+	 * During the HDCP encryption session if Type change is requested,
+	 * disable the HDCP and reenable it with new TYPE value.
+	 */
+	if (conn_state->content_protection ==
+	    DRM_MODE_CONTENT_PROTECTION_UNDESIRED ||
+	    content_protection_type_changed)
+		intel_hdcp_disable(connector);
+
+	/*
+	 * Mark the hdcp state as DESIRED after the hdcp disable of type
+	 * change procedure.
+	 */
+	if (content_protection_type_changed) {
+		mutex_lock(&hdcp->mutex);
+		hdcp->value = DRM_MODE_CONTENT_PROTECTION_DESIRED;
+		drm_connector_get(&connector->base);
+		schedule_work(&hdcp->prop_work);
+		mutex_unlock(&hdcp->mutex);
+	}
+
+	if (conn_state->content_protection ==
+	    DRM_MODE_CONTENT_PROTECTION_DESIRED) {
+		mutex_lock(&hdcp->mutex);
+		/* Avoid enabling hdcp, if it already ENABLED */
+		desired_and_not_enabled =
+			hdcp->value != DRM_MODE_CONTENT_PROTECTION_ENABLED;
+		mutex_unlock(&hdcp->mutex);
+		/*
+		 * If HDCP already ENABLED and CP property is DESIRED, schedule
+		 * prop_work to update correct CP property to user space.
+		 */
+		if (!desired_and_not_enabled && !content_protection_type_changed) {
+			drm_connector_get(&connector->base);
+			schedule_work(&hdcp->prop_work);
+		}
+	}
+
+	if (desired_and_not_enabled || content_protection_type_changed)
+		intel_hdcp_enable(connector,
+				  crtc_state,
+				  (u8)conn_state->hdcp_content_type);
+}
+
+void intel_hdcp_component_fini(struct drm_i915_private *dev_priv)
+{
+	mutex_lock(&dev_priv->display.hdcp.comp_mutex);
+	if (!dev_priv->display.hdcp.comp_added) {
+		mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+		return;
+	}
+
+	dev_priv->display.hdcp.comp_added = false;
+	mutex_unlock(&dev_priv->display.hdcp.comp_mutex);
+
+	component_del(dev_priv->drm.dev, &i915_hdcp_component_ops);
+}
+
+void intel_hdcp_cleanup(struct intel_connector *connector)
+{
+	struct intel_hdcp *hdcp = &connector->hdcp;
+
+	if (!hdcp->shim)
+		return;
+
+	/*
+	 * If the connector is registered, it's possible userspace could kick
+	 * off another HDCP enable, which would re-spawn the workers.
+	 */
+	drm_WARN_ON(connector->base.dev,
+		connector->base.registration_state == DRM_CONNECTOR_REGISTERED);
+
+	/*
+	 * Now that the connector is not registered, check_work won't be run,
+	 * but cancel any outstanding instances of it
+	 */
+	cancel_delayed_work_sync(&hdcp->check_work);
+
+	/*
+	 * We don't cancel prop_work in the same way as check_work since it
+	 * requires connection_mutex which could be held while calling this
+	 * function. Instead, we rely on the connector references grabbed before
+	 * scheduling prop_work to ensure the connector is alive when prop_work
+	 * is run. So if we're in the destroy path (which is where this
+	 * function should be called), we're "guaranteed" that prop_work is not
+	 * active (tl;dr This Should Never Happen).
+	 */
+	drm_WARN_ON(connector->base.dev, work_pending(&hdcp->prop_work));
+
+	mutex_lock(&hdcp->mutex);
+	hdcp->shim = NULL;
+	mutex_unlock(&hdcp->mutex);
+}
+
+void intel_hdcp_atomic_check(struct drm_connector *connector,
+			     struct drm_connector_state *old_state,
+			     struct drm_connector_state *new_state)
+{
+	u64 old_cp = old_state->content_protection;
+	u64 new_cp = new_state->content_protection;
+	struct drm_crtc_state *crtc_state;
+
+	if (!new_state->crtc) {
+		/*
+		 * If the connector is being disabled with CP enabled, mark it
+		 * desired so it's re-enabled when the connector is brought back
+		 */
+		if (old_cp == DRM_MODE_CONTENT_PROTECTION_ENABLED)
+			new_state->content_protection =
+				DRM_MODE_CONTENT_PROTECTION_DESIRED;
+		return;
+	}
+
+	crtc_state = drm_atomic_get_new_crtc_state(new_state->state,
+						   new_state->crtc);
+	/*
+	 * Fix the HDCP uapi content protection state in case of modeset.
+	 * FIXME: As per HDCP content protection property uapi doc, an uevent()
+	 * need to be sent if there is transition from ENABLED->DESIRED.
+	 */
+	if (drm_atomic_crtc_needs_modeset(crtc_state) &&
+	    (old_cp == DRM_MODE_CONTENT_PROTECTION_ENABLED &&
+	    new_cp != DRM_MODE_CONTENT_PROTECTION_UNDESIRED))
+		new_state->content_protection =
+			DRM_MODE_CONTENT_PROTECTION_DESIRED;
+
+	/*
+	 * Nothing to do if the state didn't change, or HDCP was activated since
+	 * the last commit. And also no change in hdcp content type.
+	 */
+	if (old_cp == new_cp ||
+	    (old_cp == DRM_MODE_CONTENT_PROTECTION_DESIRED &&
+	     new_cp == DRM_MODE_CONTENT_PROTECTION_ENABLED)) {
+		if (old_state->hdcp_content_type ==
+				new_state->hdcp_content_type)
+			return;
+	}
+
+	crtc_state->mode_changed = true;
+}
+
+/* Handles the CP_IRQ raised from the DP HDCP sink */
+void intel_hdcp_handle_cp_irq(struct intel_connector *connector)
+{
+	struct intel_hdcp *hdcp = &connector->hdcp;
+
+	if (!hdcp->shim)
+		return;
+
+	atomic_inc(&connector->hdcp.cp_irq_count);
+	wake_up_all(&connector->hdcp.cp_irq_queue);
+
+	schedule_delayed_work(&hdcp->check_work, 0);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_hdcp.h b/drivers/gpu/drm/i915/display/intel_hdcp.h
new file mode 100644
index 000000000..8f53b0c7f
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_hdcp.h
@@ -0,0 +1,46 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_HDCP_H__
+#define __INTEL_HDCP_H__
+
+#include <linux/types.h>
+
+#define HDCP_ENCRYPT_STATUS_CHANGE_TIMEOUT_MS	50
+
+struct drm_connector;
+struct drm_connector_state;
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_connector;
+struct intel_crtc_state;
+struct intel_encoder;
+struct intel_hdcp_shim;
+struct intel_digital_port;
+enum port;
+enum transcoder;
+
+void intel_hdcp_atomic_check(struct drm_connector *connector,
+			     struct drm_connector_state *old_state,
+			     struct drm_connector_state *new_state);
+int intel_hdcp_init(struct intel_connector *connector,
+		    struct intel_digital_port *dig_port,
+		    const struct intel_hdcp_shim *hdcp_shim);
+int intel_hdcp_enable(struct intel_connector *connector,
+		      const struct intel_crtc_state *pipe_config, u8 content_type);
+int intel_hdcp_disable(struct intel_connector *connector);
+void intel_hdcp_update_pipe(struct intel_atomic_state *state,
+			    struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    const struct drm_connector_state *conn_state);
+bool is_hdcp_supported(struct drm_i915_private *dev_priv, enum port port);
+bool intel_hdcp_capable(struct intel_connector *connector);
+bool intel_hdcp2_capable(struct intel_connector *connector);
+void intel_hdcp_component_init(struct drm_i915_private *dev_priv);
+void intel_hdcp_component_fini(struct drm_i915_private *dev_priv);
+void intel_hdcp_cleanup(struct intel_connector *connector);
+void intel_hdcp_handle_cp_irq(struct intel_connector *connector);
+
+#endif /* __INTEL_HDCP_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_hdcp_regs.h b/drivers/gpu/drm/i915/display/intel_hdcp_regs.h
new file mode 100644
index 000000000..2a3733e89
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_hdcp_regs.h
@@ -0,0 +1,270 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_HDCP_REGS_H__
+#define __INTEL_HDCP_REGS_H__
+
+#include "i915_reg_defs.h"
+
+/* HDCP Key Registers */
+#define HDCP_KEY_CONF			_MMIO(0x66c00)
+#define  HDCP_AKSV_SEND_TRIGGER		REG_BIT(31)
+#define  HDCP_CLEAR_KEYS_TRIGGER	REG_BIT(30)
+#define  HDCP_KEY_LOAD_TRIGGER		REG_BIT(8)
+#define HDCP_KEY_STATUS			_MMIO(0x66c04)
+#define  HDCP_FUSE_IN_PROGRESS		REG_BIT(7)
+#define  HDCP_FUSE_ERROR		REG_BIT(6)
+#define  HDCP_FUSE_DONE			REG_BIT(5)
+#define  HDCP_KEY_LOAD_STATUS		REG_BIT(1)
+#define  HDCP_KEY_LOAD_DONE		REG_BIT(0)
+#define HDCP_AKSV_LO			_MMIO(0x66c10)
+#define HDCP_AKSV_HI			_MMIO(0x66c14)
+
+/* HDCP Repeater Registers */
+#define HDCP_REP_CTL			_MMIO(0x66d00)
+#define  HDCP_TRANSA_REP_PRESENT	REG_BIT(31)
+#define  HDCP_TRANSB_REP_PRESENT	REG_BIT(30)
+#define  HDCP_TRANSC_REP_PRESENT	REG_BIT(29)
+#define  HDCP_TRANSD_REP_PRESENT	REG_BIT(28)
+#define  HDCP_DDIB_REP_PRESENT		REG_BIT(30)
+#define  HDCP_DDIA_REP_PRESENT		REG_BIT(29)
+#define  HDCP_DDIC_REP_PRESENT		REG_BIT(28)
+#define  HDCP_DDID_REP_PRESENT		REG_BIT(27)
+#define  HDCP_DDIF_REP_PRESENT		REG_BIT(26)
+#define  HDCP_DDIE_REP_PRESENT		REG_BIT(25)
+#define  HDCP_TRANSA_SHA1_M0		(1 << 20)
+#define  HDCP_TRANSB_SHA1_M0		(2 << 20)
+#define  HDCP_TRANSC_SHA1_M0		(3 << 20)
+#define  HDCP_TRANSD_SHA1_M0		(4 << 20)
+#define  HDCP_DDIB_SHA1_M0		(1 << 20)
+#define  HDCP_DDIA_SHA1_M0		(2 << 20)
+#define  HDCP_DDIC_SHA1_M0		(3 << 20)
+#define  HDCP_DDID_SHA1_M0		(4 << 20)
+#define  HDCP_DDIF_SHA1_M0		(5 << 20)
+#define  HDCP_DDIE_SHA1_M0		(6 << 20) /* Bspec says 5? */
+#define  HDCP_SHA1_BUSY			REG_BIT(16)
+#define  HDCP_SHA1_READY		REG_BIT(17)
+#define  HDCP_SHA1_COMPLETE		REG_BIT(18)
+#define  HDCP_SHA1_V_MATCH		REG_BIT(19)
+#define  HDCP_SHA1_TEXT_32		(1 << 1)
+#define  HDCP_SHA1_COMPLETE_HASH	(2 << 1)
+#define  HDCP_SHA1_TEXT_24		(4 << 1)
+#define  HDCP_SHA1_TEXT_16		(5 << 1)
+#define  HDCP_SHA1_TEXT_8		(6 << 1)
+#define  HDCP_SHA1_TEXT_0		(7 << 1)
+#define HDCP_SHA_V_PRIME_H0		_MMIO(0x66d04)
+#define HDCP_SHA_V_PRIME_H1		_MMIO(0x66d08)
+#define HDCP_SHA_V_PRIME_H2		_MMIO(0x66d0C)
+#define HDCP_SHA_V_PRIME_H3		_MMIO(0x66d10)
+#define HDCP_SHA_V_PRIME_H4		_MMIO(0x66d14)
+#define HDCP_SHA_V_PRIME(h)		_MMIO((0x66d04 + (h) * 4))
+#define HDCP_SHA_TEXT			_MMIO(0x66d18)
+
+/* HDCP Auth Registers */
+#define _PORTA_HDCP_AUTHENC		0x66800
+#define _PORTB_HDCP_AUTHENC		0x66500
+#define _PORTC_HDCP_AUTHENC		0x66600
+#define _PORTD_HDCP_AUTHENC		0x66700
+#define _PORTE_HDCP_AUTHENC		0x66A00
+#define _PORTF_HDCP_AUTHENC		0x66900
+#define _PORT_HDCP_AUTHENC(port, x)	_MMIO(_PICK(port, \
+					  _PORTA_HDCP_AUTHENC, \
+					  _PORTB_HDCP_AUTHENC, \
+					  _PORTC_HDCP_AUTHENC, \
+					  _PORTD_HDCP_AUTHENC, \
+					  _PORTE_HDCP_AUTHENC, \
+					  _PORTF_HDCP_AUTHENC) + (x))
+#define PORT_HDCP_CONF(port)		_PORT_HDCP_AUTHENC(port, 0x0)
+#define _TRANSA_HDCP_CONF		0x66400
+#define _TRANSB_HDCP_CONF		0x66500
+#define TRANS_HDCP_CONF(trans)		_MMIO_TRANS(trans, _TRANSA_HDCP_CONF, \
+						    _TRANSB_HDCP_CONF)
+#define HDCP_CONF(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP_CONF(trans) : \
+					 PORT_HDCP_CONF(port))
+
+#define  HDCP_CONF_CAPTURE_AN		REG_BIT(0)
+#define  HDCP_CONF_AUTH_AND_ENC		(REG_BIT(1) | REG_BIT(0))
+#define PORT_HDCP_ANINIT(port)		_PORT_HDCP_AUTHENC(port, 0x4)
+#define _TRANSA_HDCP_ANINIT		0x66404
+#define _TRANSB_HDCP_ANINIT		0x66504
+#define TRANS_HDCP_ANINIT(trans)	_MMIO_TRANS(trans, \
+						    _TRANSA_HDCP_ANINIT, \
+						    _TRANSB_HDCP_ANINIT)
+#define HDCP_ANINIT(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP_ANINIT(trans) : \
+					 PORT_HDCP_ANINIT(port))
+
+#define PORT_HDCP_ANLO(port)		_PORT_HDCP_AUTHENC(port, 0x8)
+#define _TRANSA_HDCP_ANLO		0x66408
+#define _TRANSB_HDCP_ANLO		0x66508
+#define TRANS_HDCP_ANLO(trans)		_MMIO_TRANS(trans, _TRANSA_HDCP_ANLO, \
+						    _TRANSB_HDCP_ANLO)
+#define HDCP_ANLO(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP_ANLO(trans) : \
+					 PORT_HDCP_ANLO(port))
+
+#define PORT_HDCP_ANHI(port)		_PORT_HDCP_AUTHENC(port, 0xC)
+#define _TRANSA_HDCP_ANHI		0x6640C
+#define _TRANSB_HDCP_ANHI		0x6650C
+#define TRANS_HDCP_ANHI(trans)		_MMIO_TRANS(trans, _TRANSA_HDCP_ANHI, \
+						    _TRANSB_HDCP_ANHI)
+#define HDCP_ANHI(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP_ANHI(trans) : \
+					 PORT_HDCP_ANHI(port))
+
+#define PORT_HDCP_BKSVLO(port)		_PORT_HDCP_AUTHENC(port, 0x10)
+#define _TRANSA_HDCP_BKSVLO		0x66410
+#define _TRANSB_HDCP_BKSVLO		0x66510
+#define TRANS_HDCP_BKSVLO(trans)	_MMIO_TRANS(trans, \
+						    _TRANSA_HDCP_BKSVLO, \
+						    _TRANSB_HDCP_BKSVLO)
+#define HDCP_BKSVLO(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP_BKSVLO(trans) : \
+					 PORT_HDCP_BKSVLO(port))
+
+#define PORT_HDCP_BKSVHI(port)		_PORT_HDCP_AUTHENC(port, 0x14)
+#define _TRANSA_HDCP_BKSVHI		0x66414
+#define _TRANSB_HDCP_BKSVHI		0x66514
+#define TRANS_HDCP_BKSVHI(trans)	_MMIO_TRANS(trans, \
+						    _TRANSA_HDCP_BKSVHI, \
+						    _TRANSB_HDCP_BKSVHI)
+#define HDCP_BKSVHI(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP_BKSVHI(trans) : \
+					 PORT_HDCP_BKSVHI(port))
+
+#define PORT_HDCP_RPRIME(port)		_PORT_HDCP_AUTHENC(port, 0x18)
+#define _TRANSA_HDCP_RPRIME		0x66418
+#define _TRANSB_HDCP_RPRIME		0x66518
+#define TRANS_HDCP_RPRIME(trans)	_MMIO_TRANS(trans, \
+						    _TRANSA_HDCP_RPRIME, \
+						    _TRANSB_HDCP_RPRIME)
+#define HDCP_RPRIME(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP_RPRIME(trans) : \
+					 PORT_HDCP_RPRIME(port))
+
+#define PORT_HDCP_STATUS(port)		_PORT_HDCP_AUTHENC(port, 0x1C)
+#define _TRANSA_HDCP_STATUS		0x6641C
+#define _TRANSB_HDCP_STATUS		0x6651C
+#define TRANS_HDCP_STATUS(trans)	_MMIO_TRANS(trans, \
+						    _TRANSA_HDCP_STATUS, \
+						    _TRANSB_HDCP_STATUS)
+#define HDCP_STATUS(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP_STATUS(trans) : \
+					 PORT_HDCP_STATUS(port))
+
+#define  HDCP_STATUS_STREAM_A_ENC	REG_BIT(31)
+#define  HDCP_STATUS_STREAM_B_ENC	REG_BIT(30)
+#define  HDCP_STATUS_STREAM_C_ENC	REG_BIT(29)
+#define  HDCP_STATUS_STREAM_D_ENC	REG_BIT(28)
+#define  HDCP_STATUS_AUTH		REG_BIT(21)
+#define  HDCP_STATUS_ENC		REG_BIT(20)
+#define  HDCP_STATUS_RI_MATCH		REG_BIT(19)
+#define  HDCP_STATUS_R0_READY		REG_BIT(18)
+#define  HDCP_STATUS_AN_READY		REG_BIT(17)
+#define  HDCP_STATUS_CIPHER		REG_BIT(16)
+#define  HDCP_STATUS_FRAME_CNT(x)	(((x) >> 8) & 0xff)
+
+/* HDCP2.2 Registers */
+#define _PORTA_HDCP2_BASE		0x66800
+#define _PORTB_HDCP2_BASE		0x66500
+#define _PORTC_HDCP2_BASE		0x66600
+#define _PORTD_HDCP2_BASE		0x66700
+#define _PORTE_HDCP2_BASE		0x66A00
+#define _PORTF_HDCP2_BASE		0x66900
+#define _PORT_HDCP2_BASE(port, x)	_MMIO(_PICK((port), \
+					  _PORTA_HDCP2_BASE, \
+					  _PORTB_HDCP2_BASE, \
+					  _PORTC_HDCP2_BASE, \
+					  _PORTD_HDCP2_BASE, \
+					  _PORTE_HDCP2_BASE, \
+					  _PORTF_HDCP2_BASE) + (x))
+
+#define PORT_HDCP2_AUTH(port)		_PORT_HDCP2_BASE(port, 0x98)
+#define _TRANSA_HDCP2_AUTH		0x66498
+#define _TRANSB_HDCP2_AUTH		0x66598
+#define TRANS_HDCP2_AUTH(trans)		_MMIO_TRANS(trans, _TRANSA_HDCP2_AUTH, \
+						    _TRANSB_HDCP2_AUTH)
+#define   AUTH_LINK_AUTHENTICATED	REG_BIT(31)
+#define   AUTH_LINK_TYPE		REG_BIT(30)
+#define   AUTH_FORCE_CLR_INPUTCTR	REG_BIT(19)
+#define   AUTH_CLR_KEYS			REG_BIT(18)
+#define HDCP2_AUTH(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP2_AUTH(trans) : \
+					 PORT_HDCP2_AUTH(port))
+
+#define PORT_HDCP2_CTL(port)		_PORT_HDCP2_BASE(port, 0xB0)
+#define _TRANSA_HDCP2_CTL		0x664B0
+#define _TRANSB_HDCP2_CTL		0x665B0
+#define TRANS_HDCP2_CTL(trans)		_MMIO_TRANS(trans, _TRANSA_HDCP2_CTL, \
+						    _TRANSB_HDCP2_CTL)
+#define   CTL_LINK_ENCRYPTION_REQ	REG_BIT(31)
+#define HDCP2_CTL(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP2_CTL(trans) : \
+					 PORT_HDCP2_CTL(port))
+
+#define PORT_HDCP2_STATUS(port)		_PORT_HDCP2_BASE(port, 0xB4)
+#define _TRANSA_HDCP2_STATUS		0x664B4
+#define _TRANSB_HDCP2_STATUS		0x665B4
+#define TRANS_HDCP2_STATUS(trans)	_MMIO_TRANS(trans, \
+						    _TRANSA_HDCP2_STATUS, \
+						    _TRANSB_HDCP2_STATUS)
+#define   LINK_TYPE_STATUS		REG_BIT(22)
+#define   LINK_AUTH_STATUS		REG_BIT(21)
+#define   LINK_ENCRYPTION_STATUS	REG_BIT(20)
+#define HDCP2_STATUS(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP2_STATUS(trans) : \
+					 PORT_HDCP2_STATUS(port))
+
+#define _PIPEA_HDCP2_STREAM_STATUS	0x668C0
+#define _PIPEB_HDCP2_STREAM_STATUS	0x665C0
+#define _PIPEC_HDCP2_STREAM_STATUS	0x666C0
+#define _PIPED_HDCP2_STREAM_STATUS	0x667C0
+#define PIPE_HDCP2_STREAM_STATUS(pipe)		_MMIO(_PICK((pipe), \
+						      _PIPEA_HDCP2_STREAM_STATUS, \
+						      _PIPEB_HDCP2_STREAM_STATUS, \
+						      _PIPEC_HDCP2_STREAM_STATUS, \
+						      _PIPED_HDCP2_STREAM_STATUS))
+
+#define _TRANSA_HDCP2_STREAM_STATUS		0x664C0
+#define _TRANSB_HDCP2_STREAM_STATUS		0x665C0
+#define TRANS_HDCP2_STREAM_STATUS(trans)	_MMIO_TRANS(trans, \
+						    _TRANSA_HDCP2_STREAM_STATUS, \
+						    _TRANSB_HDCP2_STREAM_STATUS)
+#define   STREAM_ENCRYPTION_STATUS	REG_BIT(31)
+#define   STREAM_TYPE_STATUS		REG_BIT(30)
+#define HDCP2_STREAM_STATUS(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP2_STREAM_STATUS(trans) : \
+					 PIPE_HDCP2_STREAM_STATUS(pipe))
+
+#define _PORTA_HDCP2_AUTH_STREAM		0x66F00
+#define _PORTB_HDCP2_AUTH_STREAM		0x66F04
+#define PORT_HDCP2_AUTH_STREAM(port)	_MMIO_PORT(port, \
+						   _PORTA_HDCP2_AUTH_STREAM, \
+						   _PORTB_HDCP2_AUTH_STREAM)
+#define _TRANSA_HDCP2_AUTH_STREAM		0x66F00
+#define _TRANSB_HDCP2_AUTH_STREAM		0x66F04
+#define TRANS_HDCP2_AUTH_STREAM(trans)	_MMIO_TRANS(trans, \
+						    _TRANSA_HDCP2_AUTH_STREAM, \
+						    _TRANSB_HDCP2_AUTH_STREAM)
+#define   AUTH_STREAM_TYPE		REG_BIT(31)
+#define HDCP2_AUTH_STREAM(dev_priv, trans, port) \
+					(GRAPHICS_VER(dev_priv) >= 12 ? \
+					 TRANS_HDCP2_AUTH_STREAM(trans) : \
+					 PORT_HDCP2_AUTH_STREAM(port))
+
+#endif /* __INTEL_HDCP_REGS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_hdmi.c b/drivers/gpu/drm/i915/display/intel_hdmi.c
new file mode 100644
index 000000000..2600019fc
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_hdmi.c
@@ -0,0 +1,3237 @@
+/*
+ * Copyright 2006 Dave Airlie <airlied@linux.ie>
+ * Copyright © 2006-2009 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Eric Anholt <eric@anholt.net>
+ *	Jesse Barnes <jesse.barnes@intel.com>
+ */
+
+#include <linux/delay.h>
+#include <linux/hdmi.h>
+#include <linux/i2c.h>
+#include <linux/slab.h>
+#include <linux/string_helpers.h>
+
+#include <drm/display/drm_hdcp_helper.h>
+#include <drm/display/drm_hdmi_helper.h>
+#include <drm/display/drm_scdc_helper.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_edid.h>
+#include <drm/intel_lpe_audio.h>
+
+#include "i915_debugfs.h"
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_connector.h"
+#include "intel_ddi.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_gmbus.h"
+#include "intel_hdcp.h"
+#include "intel_hdcp_regs.h"
+#include "intel_hdmi.h"
+#include "intel_lspcon.h"
+#include "intel_panel.h"
+#include "intel_snps_phy.h"
+
+static struct drm_i915_private *intel_hdmi_to_i915(struct intel_hdmi *intel_hdmi)
+{
+	return to_i915(hdmi_to_dig_port(intel_hdmi)->base.base.dev);
+}
+
+static void
+assert_hdmi_port_disabled(struct intel_hdmi *intel_hdmi)
+{
+	struct drm_i915_private *dev_priv = intel_hdmi_to_i915(intel_hdmi);
+	u32 enabled_bits;
+
+	enabled_bits = HAS_DDI(dev_priv) ? DDI_BUF_CTL_ENABLE : SDVO_ENABLE;
+
+	drm_WARN(&dev_priv->drm,
+		 intel_de_read(dev_priv, intel_hdmi->hdmi_reg) & enabled_bits,
+		 "HDMI port enabled, expecting disabled\n");
+}
+
+static void
+assert_hdmi_transcoder_func_disabled(struct drm_i915_private *dev_priv,
+				     enum transcoder cpu_transcoder)
+{
+	drm_WARN(&dev_priv->drm,
+		 intel_de_read(dev_priv, TRANS_DDI_FUNC_CTL(cpu_transcoder)) &
+		 TRANS_DDI_FUNC_ENABLE,
+		 "HDMI transcoder function enabled, expecting disabled\n");
+}
+
+static u32 g4x_infoframe_index(unsigned int type)
+{
+	switch (type) {
+	case HDMI_PACKET_TYPE_GAMUT_METADATA:
+		return VIDEO_DIP_SELECT_GAMUT;
+	case HDMI_INFOFRAME_TYPE_AVI:
+		return VIDEO_DIP_SELECT_AVI;
+	case HDMI_INFOFRAME_TYPE_SPD:
+		return VIDEO_DIP_SELECT_SPD;
+	case HDMI_INFOFRAME_TYPE_VENDOR:
+		return VIDEO_DIP_SELECT_VENDOR;
+	default:
+		MISSING_CASE(type);
+		return 0;
+	}
+}
+
+static u32 g4x_infoframe_enable(unsigned int type)
+{
+	switch (type) {
+	case HDMI_PACKET_TYPE_GENERAL_CONTROL:
+		return VIDEO_DIP_ENABLE_GCP;
+	case HDMI_PACKET_TYPE_GAMUT_METADATA:
+		return VIDEO_DIP_ENABLE_GAMUT;
+	case DP_SDP_VSC:
+		return 0;
+	case HDMI_INFOFRAME_TYPE_AVI:
+		return VIDEO_DIP_ENABLE_AVI;
+	case HDMI_INFOFRAME_TYPE_SPD:
+		return VIDEO_DIP_ENABLE_SPD;
+	case HDMI_INFOFRAME_TYPE_VENDOR:
+		return VIDEO_DIP_ENABLE_VENDOR;
+	case HDMI_INFOFRAME_TYPE_DRM:
+		return 0;
+	default:
+		MISSING_CASE(type);
+		return 0;
+	}
+}
+
+static u32 hsw_infoframe_enable(unsigned int type)
+{
+	switch (type) {
+	case HDMI_PACKET_TYPE_GENERAL_CONTROL:
+		return VIDEO_DIP_ENABLE_GCP_HSW;
+	case HDMI_PACKET_TYPE_GAMUT_METADATA:
+		return VIDEO_DIP_ENABLE_GMP_HSW;
+	case DP_SDP_VSC:
+		return VIDEO_DIP_ENABLE_VSC_HSW;
+	case DP_SDP_PPS:
+		return VDIP_ENABLE_PPS;
+	case HDMI_INFOFRAME_TYPE_AVI:
+		return VIDEO_DIP_ENABLE_AVI_HSW;
+	case HDMI_INFOFRAME_TYPE_SPD:
+		return VIDEO_DIP_ENABLE_SPD_HSW;
+	case HDMI_INFOFRAME_TYPE_VENDOR:
+		return VIDEO_DIP_ENABLE_VS_HSW;
+	case HDMI_INFOFRAME_TYPE_DRM:
+		return VIDEO_DIP_ENABLE_DRM_GLK;
+	default:
+		MISSING_CASE(type);
+		return 0;
+	}
+}
+
+static i915_reg_t
+hsw_dip_data_reg(struct drm_i915_private *dev_priv,
+		 enum transcoder cpu_transcoder,
+		 unsigned int type,
+		 int i)
+{
+	switch (type) {
+	case HDMI_PACKET_TYPE_GAMUT_METADATA:
+		return HSW_TVIDEO_DIP_GMP_DATA(cpu_transcoder, i);
+	case DP_SDP_VSC:
+		return HSW_TVIDEO_DIP_VSC_DATA(cpu_transcoder, i);
+	case DP_SDP_PPS:
+		return ICL_VIDEO_DIP_PPS_DATA(cpu_transcoder, i);
+	case HDMI_INFOFRAME_TYPE_AVI:
+		return HSW_TVIDEO_DIP_AVI_DATA(cpu_transcoder, i);
+	case HDMI_INFOFRAME_TYPE_SPD:
+		return HSW_TVIDEO_DIP_SPD_DATA(cpu_transcoder, i);
+	case HDMI_INFOFRAME_TYPE_VENDOR:
+		return HSW_TVIDEO_DIP_VS_DATA(cpu_transcoder, i);
+	case HDMI_INFOFRAME_TYPE_DRM:
+		return GLK_TVIDEO_DIP_DRM_DATA(cpu_transcoder, i);
+	default:
+		MISSING_CASE(type);
+		return INVALID_MMIO_REG;
+	}
+}
+
+static int hsw_dip_data_size(struct drm_i915_private *dev_priv,
+			     unsigned int type)
+{
+	switch (type) {
+	case DP_SDP_VSC:
+		return VIDEO_DIP_VSC_DATA_SIZE;
+	case DP_SDP_PPS:
+		return VIDEO_DIP_PPS_DATA_SIZE;
+	case HDMI_PACKET_TYPE_GAMUT_METADATA:
+		if (DISPLAY_VER(dev_priv) >= 11)
+			return VIDEO_DIP_GMP_DATA_SIZE;
+		else
+			return VIDEO_DIP_DATA_SIZE;
+	default:
+		return VIDEO_DIP_DATA_SIZE;
+	}
+}
+
+static void g4x_write_infoframe(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state,
+				unsigned int type,
+				const void *frame, ssize_t len)
+{
+	const u32 *data = frame;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 val = intel_de_read(dev_priv, VIDEO_DIP_CTL);
+	int i;
+
+	drm_WARN(&dev_priv->drm, !(val & VIDEO_DIP_ENABLE),
+		 "Writing DIP with CTL reg disabled\n");
+
+	val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */
+	val |= g4x_infoframe_index(type);
+
+	val &= ~g4x_infoframe_enable(type);
+
+	intel_de_write(dev_priv, VIDEO_DIP_CTL, val);
+
+	for (i = 0; i < len; i += 4) {
+		intel_de_write(dev_priv, VIDEO_DIP_DATA, *data);
+		data++;
+	}
+	/* Write every possible data byte to force correct ECC calculation. */
+	for (; i < VIDEO_DIP_DATA_SIZE; i += 4)
+		intel_de_write(dev_priv, VIDEO_DIP_DATA, 0);
+
+	val |= g4x_infoframe_enable(type);
+	val &= ~VIDEO_DIP_FREQ_MASK;
+	val |= VIDEO_DIP_FREQ_VSYNC;
+
+	intel_de_write(dev_priv, VIDEO_DIP_CTL, val);
+	intel_de_posting_read(dev_priv, VIDEO_DIP_CTL);
+}
+
+static void g4x_read_infoframe(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state,
+			       unsigned int type,
+			       void *frame, ssize_t len)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 val, *data = frame;
+	int i;
+
+	val = intel_de_read(dev_priv, VIDEO_DIP_CTL);
+
+	val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */
+	val |= g4x_infoframe_index(type);
+
+	intel_de_write(dev_priv, VIDEO_DIP_CTL, val);
+
+	for (i = 0; i < len; i += 4)
+		*data++ = intel_de_read(dev_priv, VIDEO_DIP_DATA);
+}
+
+static u32 g4x_infoframes_enabled(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 val = intel_de_read(dev_priv, VIDEO_DIP_CTL);
+
+	if ((val & VIDEO_DIP_ENABLE) == 0)
+		return 0;
+
+	if ((val & VIDEO_DIP_PORT_MASK) != VIDEO_DIP_PORT(encoder->port))
+		return 0;
+
+	return val & (VIDEO_DIP_ENABLE_AVI |
+		      VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_SPD);
+}
+
+static void ibx_write_infoframe(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state,
+				unsigned int type,
+				const void *frame, ssize_t len)
+{
+	const u32 *data = frame;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	i915_reg_t reg = TVIDEO_DIP_CTL(crtc->pipe);
+	u32 val = intel_de_read(dev_priv, reg);
+	int i;
+
+	drm_WARN(&dev_priv->drm, !(val & VIDEO_DIP_ENABLE),
+		 "Writing DIP with CTL reg disabled\n");
+
+	val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */
+	val |= g4x_infoframe_index(type);
+
+	val &= ~g4x_infoframe_enable(type);
+
+	intel_de_write(dev_priv, reg, val);
+
+	for (i = 0; i < len; i += 4) {
+		intel_de_write(dev_priv, TVIDEO_DIP_DATA(crtc->pipe),
+			       *data);
+		data++;
+	}
+	/* Write every possible data byte to force correct ECC calculation. */
+	for (; i < VIDEO_DIP_DATA_SIZE; i += 4)
+		intel_de_write(dev_priv, TVIDEO_DIP_DATA(crtc->pipe), 0);
+
+	val |= g4x_infoframe_enable(type);
+	val &= ~VIDEO_DIP_FREQ_MASK;
+	val |= VIDEO_DIP_FREQ_VSYNC;
+
+	intel_de_write(dev_priv, reg, val);
+	intel_de_posting_read(dev_priv, reg);
+}
+
+static void ibx_read_infoframe(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state,
+			       unsigned int type,
+			       void *frame, ssize_t len)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	u32 val, *data = frame;
+	int i;
+
+	val = intel_de_read(dev_priv, TVIDEO_DIP_CTL(crtc->pipe));
+
+	val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */
+	val |= g4x_infoframe_index(type);
+
+	intel_de_write(dev_priv, TVIDEO_DIP_CTL(crtc->pipe), val);
+
+	for (i = 0; i < len; i += 4)
+		*data++ = intel_de_read(dev_priv, TVIDEO_DIP_DATA(crtc->pipe));
+}
+
+static u32 ibx_infoframes_enabled(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum pipe pipe = to_intel_crtc(pipe_config->uapi.crtc)->pipe;
+	i915_reg_t reg = TVIDEO_DIP_CTL(pipe);
+	u32 val = intel_de_read(dev_priv, reg);
+
+	if ((val & VIDEO_DIP_ENABLE) == 0)
+		return 0;
+
+	if ((val & VIDEO_DIP_PORT_MASK) != VIDEO_DIP_PORT(encoder->port))
+		return 0;
+
+	return val & (VIDEO_DIP_ENABLE_AVI |
+		      VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT |
+		      VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP);
+}
+
+static void cpt_write_infoframe(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state,
+				unsigned int type,
+				const void *frame, ssize_t len)
+{
+	const u32 *data = frame;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	i915_reg_t reg = TVIDEO_DIP_CTL(crtc->pipe);
+	u32 val = intel_de_read(dev_priv, reg);
+	int i;
+
+	drm_WARN(&dev_priv->drm, !(val & VIDEO_DIP_ENABLE),
+		 "Writing DIP with CTL reg disabled\n");
+
+	val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */
+	val |= g4x_infoframe_index(type);
+
+	/* The DIP control register spec says that we need to update the AVI
+	 * infoframe without clearing its enable bit */
+	if (type != HDMI_INFOFRAME_TYPE_AVI)
+		val &= ~g4x_infoframe_enable(type);
+
+	intel_de_write(dev_priv, reg, val);
+
+	for (i = 0; i < len; i += 4) {
+		intel_de_write(dev_priv, TVIDEO_DIP_DATA(crtc->pipe),
+			       *data);
+		data++;
+	}
+	/* Write every possible data byte to force correct ECC calculation. */
+	for (; i < VIDEO_DIP_DATA_SIZE; i += 4)
+		intel_de_write(dev_priv, TVIDEO_DIP_DATA(crtc->pipe), 0);
+
+	val |= g4x_infoframe_enable(type);
+	val &= ~VIDEO_DIP_FREQ_MASK;
+	val |= VIDEO_DIP_FREQ_VSYNC;
+
+	intel_de_write(dev_priv, reg, val);
+	intel_de_posting_read(dev_priv, reg);
+}
+
+static void cpt_read_infoframe(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state,
+			       unsigned int type,
+			       void *frame, ssize_t len)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	u32 val, *data = frame;
+	int i;
+
+	val = intel_de_read(dev_priv, TVIDEO_DIP_CTL(crtc->pipe));
+
+	val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */
+	val |= g4x_infoframe_index(type);
+
+	intel_de_write(dev_priv, TVIDEO_DIP_CTL(crtc->pipe), val);
+
+	for (i = 0; i < len; i += 4)
+		*data++ = intel_de_read(dev_priv, TVIDEO_DIP_DATA(crtc->pipe));
+}
+
+static u32 cpt_infoframes_enabled(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum pipe pipe = to_intel_crtc(pipe_config->uapi.crtc)->pipe;
+	u32 val = intel_de_read(dev_priv, TVIDEO_DIP_CTL(pipe));
+
+	if ((val & VIDEO_DIP_ENABLE) == 0)
+		return 0;
+
+	return val & (VIDEO_DIP_ENABLE_AVI |
+		      VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT |
+		      VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP);
+}
+
+static void vlv_write_infoframe(struct intel_encoder *encoder,
+				const struct intel_crtc_state *crtc_state,
+				unsigned int type,
+				const void *frame, ssize_t len)
+{
+	const u32 *data = frame;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	i915_reg_t reg = VLV_TVIDEO_DIP_CTL(crtc->pipe);
+	u32 val = intel_de_read(dev_priv, reg);
+	int i;
+
+	drm_WARN(&dev_priv->drm, !(val & VIDEO_DIP_ENABLE),
+		 "Writing DIP with CTL reg disabled\n");
+
+	val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */
+	val |= g4x_infoframe_index(type);
+
+	val &= ~g4x_infoframe_enable(type);
+
+	intel_de_write(dev_priv, reg, val);
+
+	for (i = 0; i < len; i += 4) {
+		intel_de_write(dev_priv,
+			       VLV_TVIDEO_DIP_DATA(crtc->pipe), *data);
+		data++;
+	}
+	/* Write every possible data byte to force correct ECC calculation. */
+	for (; i < VIDEO_DIP_DATA_SIZE; i += 4)
+		intel_de_write(dev_priv,
+			       VLV_TVIDEO_DIP_DATA(crtc->pipe), 0);
+
+	val |= g4x_infoframe_enable(type);
+	val &= ~VIDEO_DIP_FREQ_MASK;
+	val |= VIDEO_DIP_FREQ_VSYNC;
+
+	intel_de_write(dev_priv, reg, val);
+	intel_de_posting_read(dev_priv, reg);
+}
+
+static void vlv_read_infoframe(struct intel_encoder *encoder,
+			       const struct intel_crtc_state *crtc_state,
+			       unsigned int type,
+			       void *frame, ssize_t len)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	u32 val, *data = frame;
+	int i;
+
+	val = intel_de_read(dev_priv, VLV_TVIDEO_DIP_CTL(crtc->pipe));
+
+	val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */
+	val |= g4x_infoframe_index(type);
+
+	intel_de_write(dev_priv, VLV_TVIDEO_DIP_CTL(crtc->pipe), val);
+
+	for (i = 0; i < len; i += 4)
+		*data++ = intel_de_read(dev_priv,
+				        VLV_TVIDEO_DIP_DATA(crtc->pipe));
+}
+
+static u32 vlv_infoframes_enabled(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum pipe pipe = to_intel_crtc(pipe_config->uapi.crtc)->pipe;
+	u32 val = intel_de_read(dev_priv, VLV_TVIDEO_DIP_CTL(pipe));
+
+	if ((val & VIDEO_DIP_ENABLE) == 0)
+		return 0;
+
+	if ((val & VIDEO_DIP_PORT_MASK) != VIDEO_DIP_PORT(encoder->port))
+		return 0;
+
+	return val & (VIDEO_DIP_ENABLE_AVI |
+		      VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT |
+		      VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP);
+}
+
+void hsw_write_infoframe(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *crtc_state,
+			 unsigned int type,
+			 const void *frame, ssize_t len)
+{
+	const u32 *data = frame;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	i915_reg_t ctl_reg = HSW_TVIDEO_DIP_CTL(cpu_transcoder);
+	int data_size;
+	int i;
+	u32 val = intel_de_read(dev_priv, ctl_reg);
+
+	data_size = hsw_dip_data_size(dev_priv, type);
+
+	drm_WARN_ON(&dev_priv->drm, len > data_size);
+
+	val &= ~hsw_infoframe_enable(type);
+	intel_de_write(dev_priv, ctl_reg, val);
+
+	for (i = 0; i < len; i += 4) {
+		intel_de_write(dev_priv,
+			       hsw_dip_data_reg(dev_priv, cpu_transcoder, type, i >> 2),
+			       *data);
+		data++;
+	}
+	/* Write every possible data byte to force correct ECC calculation. */
+	for (; i < data_size; i += 4)
+		intel_de_write(dev_priv,
+			       hsw_dip_data_reg(dev_priv, cpu_transcoder, type, i >> 2),
+			       0);
+
+	/* Wa_14013475917 */
+	if (DISPLAY_VER(dev_priv) == 13 && crtc_state->has_psr &&
+	    type == DP_SDP_VSC)
+		return;
+
+	val |= hsw_infoframe_enable(type);
+	intel_de_write(dev_priv, ctl_reg, val);
+	intel_de_posting_read(dev_priv, ctl_reg);
+}
+
+void hsw_read_infoframe(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state,
+			unsigned int type, void *frame, ssize_t len)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 *data = frame;
+	int i;
+
+	for (i = 0; i < len; i += 4)
+		*data++ = intel_de_read(dev_priv,
+				        hsw_dip_data_reg(dev_priv, cpu_transcoder, type, i >> 2));
+}
+
+static u32 hsw_infoframes_enabled(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 val = intel_de_read(dev_priv,
+				HSW_TVIDEO_DIP_CTL(pipe_config->cpu_transcoder));
+	u32 mask;
+
+	mask = (VIDEO_DIP_ENABLE_VSC_HSW | VIDEO_DIP_ENABLE_AVI_HSW |
+		VIDEO_DIP_ENABLE_GCP_HSW | VIDEO_DIP_ENABLE_VS_HSW |
+		VIDEO_DIP_ENABLE_GMP_HSW | VIDEO_DIP_ENABLE_SPD_HSW);
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		mask |= VIDEO_DIP_ENABLE_DRM_GLK;
+
+	return val & mask;
+}
+
+static const u8 infoframe_type_to_idx[] = {
+	HDMI_PACKET_TYPE_GENERAL_CONTROL,
+	HDMI_PACKET_TYPE_GAMUT_METADATA,
+	DP_SDP_VSC,
+	HDMI_INFOFRAME_TYPE_AVI,
+	HDMI_INFOFRAME_TYPE_SPD,
+	HDMI_INFOFRAME_TYPE_VENDOR,
+	HDMI_INFOFRAME_TYPE_DRM,
+};
+
+u32 intel_hdmi_infoframe_enable(unsigned int type)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(infoframe_type_to_idx); i++) {
+		if (infoframe_type_to_idx[i] == type)
+			return BIT(i);
+	}
+
+	return 0;
+}
+
+u32 intel_hdmi_infoframes_enabled(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	u32 val, ret = 0;
+	int i;
+
+	val = dig_port->infoframes_enabled(encoder, crtc_state);
+
+	/* map from hardware bits to dip idx */
+	for (i = 0; i < ARRAY_SIZE(infoframe_type_to_idx); i++) {
+		unsigned int type = infoframe_type_to_idx[i];
+
+		if (HAS_DDI(dev_priv)) {
+			if (val & hsw_infoframe_enable(type))
+				ret |= BIT(i);
+		} else {
+			if (val & g4x_infoframe_enable(type))
+				ret |= BIT(i);
+		}
+	}
+
+	return ret;
+}
+
+/*
+ * The data we write to the DIP data buffer registers is 1 byte bigger than the
+ * HDMI infoframe size because of an ECC/reserved byte at position 3 (starting
+ * at 0). It's also a byte used by DisplayPort so the same DIP registers can be
+ * used for both technologies.
+ *
+ * DW0: Reserved/ECC/DP | HB2 | HB1 | HB0
+ * DW1:       DB3       | DB2 | DB1 | DB0
+ * DW2:       DB7       | DB6 | DB5 | DB4
+ * DW3: ...
+ *
+ * (HB is Header Byte, DB is Data Byte)
+ *
+ * The hdmi pack() functions don't know about that hardware specific hole so we
+ * trick them by giving an offset into the buffer and moving back the header
+ * bytes by one.
+ */
+static void intel_write_infoframe(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state,
+				  enum hdmi_infoframe_type type,
+				  const union hdmi_infoframe *frame)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	u8 buffer[VIDEO_DIP_DATA_SIZE];
+	ssize_t len;
+
+	if ((crtc_state->infoframes.enable &
+	     intel_hdmi_infoframe_enable(type)) == 0)
+		return;
+
+	if (drm_WARN_ON(encoder->base.dev, frame->any.type != type))
+		return;
+
+	/* see comment above for the reason for this offset */
+	len = hdmi_infoframe_pack_only(frame, buffer + 1, sizeof(buffer) - 1);
+	if (drm_WARN_ON(encoder->base.dev, len < 0))
+		return;
+
+	/* Insert the 'hole' (see big comment above) at position 3 */
+	memmove(&buffer[0], &buffer[1], 3);
+	buffer[3] = 0;
+	len++;
+
+	dig_port->write_infoframe(encoder, crtc_state, type, buffer, len);
+}
+
+void intel_read_infoframe(struct intel_encoder *encoder,
+			  const struct intel_crtc_state *crtc_state,
+			  enum hdmi_infoframe_type type,
+			  union hdmi_infoframe *frame)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	u8 buffer[VIDEO_DIP_DATA_SIZE];
+	int ret;
+
+	if ((crtc_state->infoframes.enable &
+	     intel_hdmi_infoframe_enable(type)) == 0)
+		return;
+
+	dig_port->read_infoframe(encoder, crtc_state,
+				       type, buffer, sizeof(buffer));
+
+	/* Fill the 'hole' (see big comment above) at position 3 */
+	memmove(&buffer[1], &buffer[0], 3);
+
+	/* see comment above for the reason for this offset */
+	ret = hdmi_infoframe_unpack(frame, buffer + 1, sizeof(buffer) - 1);
+	if (ret) {
+		drm_dbg_kms(encoder->base.dev,
+			    "Failed to unpack infoframe type 0x%02x\n", type);
+		return;
+	}
+
+	if (frame->any.type != type)
+		drm_dbg_kms(encoder->base.dev,
+			    "Found the wrong infoframe type 0x%x (expected 0x%02x)\n",
+			    frame->any.type, type);
+}
+
+static bool
+intel_hdmi_compute_avi_infoframe(struct intel_encoder *encoder,
+				 struct intel_crtc_state *crtc_state,
+				 struct drm_connector_state *conn_state)
+{
+	struct hdmi_avi_infoframe *frame = &crtc_state->infoframes.avi.avi;
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	struct drm_connector *connector = conn_state->connector;
+	int ret;
+
+	if (!crtc_state->has_infoframe)
+		return true;
+
+	crtc_state->infoframes.enable |=
+		intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_AVI);
+
+	ret = drm_hdmi_avi_infoframe_from_display_mode(frame, connector,
+						       adjusted_mode);
+	if (ret)
+		return false;
+
+	if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR420)
+		frame->colorspace = HDMI_COLORSPACE_YUV420;
+	else if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR444)
+		frame->colorspace = HDMI_COLORSPACE_YUV444;
+	else
+		frame->colorspace = HDMI_COLORSPACE_RGB;
+
+	drm_hdmi_avi_infoframe_colorimetry(frame, conn_state);
+
+	/* nonsense combination */
+	drm_WARN_ON(encoder->base.dev, crtc_state->limited_color_range &&
+		    crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB);
+
+	if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_RGB) {
+		drm_hdmi_avi_infoframe_quant_range(frame, connector,
+						   adjusted_mode,
+						   crtc_state->limited_color_range ?
+						   HDMI_QUANTIZATION_RANGE_LIMITED :
+						   HDMI_QUANTIZATION_RANGE_FULL);
+	} else {
+		frame->quantization_range = HDMI_QUANTIZATION_RANGE_DEFAULT;
+		frame->ycc_quantization_range = HDMI_YCC_QUANTIZATION_RANGE_LIMITED;
+	}
+
+	drm_hdmi_avi_infoframe_content_type(frame, conn_state);
+
+	/* TODO: handle pixel repetition for YCBCR420 outputs */
+
+	ret = hdmi_avi_infoframe_check(frame);
+	if (drm_WARN_ON(encoder->base.dev, ret))
+		return false;
+
+	return true;
+}
+
+static bool
+intel_hdmi_compute_spd_infoframe(struct intel_encoder *encoder,
+				 struct intel_crtc_state *crtc_state,
+				 struct drm_connector_state *conn_state)
+{
+	struct hdmi_spd_infoframe *frame = &crtc_state->infoframes.spd.spd;
+	int ret;
+
+	if (!crtc_state->has_infoframe)
+		return true;
+
+	crtc_state->infoframes.enable |=
+		intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_SPD);
+
+	ret = hdmi_spd_infoframe_init(frame, "Intel", "Integrated gfx");
+	if (drm_WARN_ON(encoder->base.dev, ret))
+		return false;
+
+	frame->sdi = HDMI_SPD_SDI_PC;
+
+	ret = hdmi_spd_infoframe_check(frame);
+	if (drm_WARN_ON(encoder->base.dev, ret))
+		return false;
+
+	return true;
+}
+
+static bool
+intel_hdmi_compute_hdmi_infoframe(struct intel_encoder *encoder,
+				  struct intel_crtc_state *crtc_state,
+				  struct drm_connector_state *conn_state)
+{
+	struct hdmi_vendor_infoframe *frame =
+		&crtc_state->infoframes.hdmi.vendor.hdmi;
+	const struct drm_display_info *info =
+		&conn_state->connector->display_info;
+	int ret;
+
+	if (!crtc_state->has_infoframe || !info->has_hdmi_infoframe)
+		return true;
+
+	crtc_state->infoframes.enable |=
+		intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_VENDOR);
+
+	ret = drm_hdmi_vendor_infoframe_from_display_mode(frame,
+							  conn_state->connector,
+							  &crtc_state->hw.adjusted_mode);
+	if (drm_WARN_ON(encoder->base.dev, ret))
+		return false;
+
+	ret = hdmi_vendor_infoframe_check(frame);
+	if (drm_WARN_ON(encoder->base.dev, ret))
+		return false;
+
+	return true;
+}
+
+static bool
+intel_hdmi_compute_drm_infoframe(struct intel_encoder *encoder,
+				 struct intel_crtc_state *crtc_state,
+				 struct drm_connector_state *conn_state)
+{
+	struct hdmi_drm_infoframe *frame = &crtc_state->infoframes.drm.drm;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	int ret;
+
+	if (DISPLAY_VER(dev_priv) < 10)
+		return true;
+
+	if (!crtc_state->has_infoframe)
+		return true;
+
+	if (!conn_state->hdr_output_metadata)
+		return true;
+
+	crtc_state->infoframes.enable |=
+		intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_DRM);
+
+	ret = drm_hdmi_infoframe_set_hdr_metadata(frame, conn_state);
+	if (ret < 0) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "couldn't set HDR metadata in infoframe\n");
+		return false;
+	}
+
+	ret = hdmi_drm_infoframe_check(frame);
+	if (drm_WARN_ON(&dev_priv->drm, ret))
+		return false;
+
+	return true;
+}
+
+static void g4x_set_infoframes(struct intel_encoder *encoder,
+			       bool enable,
+			       const struct intel_crtc_state *crtc_state,
+			       const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_hdmi *intel_hdmi = &dig_port->hdmi;
+	i915_reg_t reg = VIDEO_DIP_CTL;
+	u32 val = intel_de_read(dev_priv, reg);
+	u32 port = VIDEO_DIP_PORT(encoder->port);
+
+	assert_hdmi_port_disabled(intel_hdmi);
+
+	/* If the registers were not initialized yet, they might be zeroes,
+	 * which means we're selecting the AVI DIP and we're setting its
+	 * frequency to once. This seems to really confuse the HW and make
+	 * things stop working (the register spec says the AVI always needs to
+	 * be sent every VSync). So here we avoid writing to the register more
+	 * than we need and also explicitly select the AVI DIP and explicitly
+	 * set its frequency to every VSync. Avoiding to write it twice seems to
+	 * be enough to solve the problem, but being defensive shouldn't hurt us
+	 * either. */
+	val |= VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC;
+
+	if (!enable) {
+		if (!(val & VIDEO_DIP_ENABLE))
+			return;
+		if (port != (val & VIDEO_DIP_PORT_MASK)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "video DIP still enabled on port %c\n",
+				    (val & VIDEO_DIP_PORT_MASK) >> 29);
+			return;
+		}
+		val &= ~(VIDEO_DIP_ENABLE | VIDEO_DIP_ENABLE_AVI |
+			 VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_SPD);
+		intel_de_write(dev_priv, reg, val);
+		intel_de_posting_read(dev_priv, reg);
+		return;
+	}
+
+	if (port != (val & VIDEO_DIP_PORT_MASK)) {
+		if (val & VIDEO_DIP_ENABLE) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "video DIP already enabled on port %c\n",
+				    (val & VIDEO_DIP_PORT_MASK) >> 29);
+			return;
+		}
+		val &= ~VIDEO_DIP_PORT_MASK;
+		val |= port;
+	}
+
+	val |= VIDEO_DIP_ENABLE;
+	val &= ~(VIDEO_DIP_ENABLE_AVI |
+		 VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_SPD);
+
+	intel_de_write(dev_priv, reg, val);
+	intel_de_posting_read(dev_priv, reg);
+
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_AVI,
+			      &crtc_state->infoframes.avi);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_SPD,
+			      &crtc_state->infoframes.spd);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_VENDOR,
+			      &crtc_state->infoframes.hdmi);
+}
+
+/*
+ * Determine if default_phase=1 can be indicated in the GCP infoframe.
+ *
+ * From HDMI specification 1.4a:
+ * - The first pixel of each Video Data Period shall always have a pixel packing phase of 0
+ * - The first pixel following each Video Data Period shall have a pixel packing phase of 0
+ * - The PP bits shall be constant for all GCPs and will be equal to the last packing phase
+ * - The first pixel following every transition of HSYNC or VSYNC shall have a pixel packing
+ *   phase of 0
+ */
+static bool gcp_default_phase_possible(int pipe_bpp,
+				       const struct drm_display_mode *mode)
+{
+	unsigned int pixels_per_group;
+
+	switch (pipe_bpp) {
+	case 30:
+		/* 4 pixels in 5 clocks */
+		pixels_per_group = 4;
+		break;
+	case 36:
+		/* 2 pixels in 3 clocks */
+		pixels_per_group = 2;
+		break;
+	case 48:
+		/* 1 pixel in 2 clocks */
+		pixels_per_group = 1;
+		break;
+	default:
+		/* phase information not relevant for 8bpc */
+		return false;
+	}
+
+	return mode->crtc_hdisplay % pixels_per_group == 0 &&
+		mode->crtc_htotal % pixels_per_group == 0 &&
+		mode->crtc_hblank_start % pixels_per_group == 0 &&
+		mode->crtc_hblank_end % pixels_per_group == 0 &&
+		mode->crtc_hsync_start % pixels_per_group == 0 &&
+		mode->crtc_hsync_end % pixels_per_group == 0 &&
+		((mode->flags & DRM_MODE_FLAG_INTERLACE) == 0 ||
+		 mode->crtc_htotal/2 % pixels_per_group == 0);
+}
+
+static bool intel_hdmi_set_gcp_infoframe(struct intel_encoder *encoder,
+					 const struct intel_crtc_state *crtc_state,
+					 const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	i915_reg_t reg;
+
+	if ((crtc_state->infoframes.enable &
+	     intel_hdmi_infoframe_enable(HDMI_PACKET_TYPE_GENERAL_CONTROL)) == 0)
+		return false;
+
+	if (HAS_DDI(dev_priv))
+		reg = HSW_TVIDEO_DIP_GCP(crtc_state->cpu_transcoder);
+	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		reg = VLV_TVIDEO_DIP_GCP(crtc->pipe);
+	else if (HAS_PCH_SPLIT(dev_priv))
+		reg = TVIDEO_DIP_GCP(crtc->pipe);
+	else
+		return false;
+
+	intel_de_write(dev_priv, reg, crtc_state->infoframes.gcp);
+
+	return true;
+}
+
+void intel_hdmi_read_gcp_infoframe(struct intel_encoder *encoder,
+				   struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	i915_reg_t reg;
+
+	if ((crtc_state->infoframes.enable &
+	     intel_hdmi_infoframe_enable(HDMI_PACKET_TYPE_GENERAL_CONTROL)) == 0)
+		return;
+
+	if (HAS_DDI(dev_priv))
+		reg = HSW_TVIDEO_DIP_GCP(crtc_state->cpu_transcoder);
+	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		reg = VLV_TVIDEO_DIP_GCP(crtc->pipe);
+	else if (HAS_PCH_SPLIT(dev_priv))
+		reg = TVIDEO_DIP_GCP(crtc->pipe);
+	else
+		return;
+
+	crtc_state->infoframes.gcp = intel_de_read(dev_priv, reg);
+}
+
+static void intel_hdmi_compute_gcp_infoframe(struct intel_encoder *encoder,
+					     struct intel_crtc_state *crtc_state,
+					     struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (IS_G4X(dev_priv) || !crtc_state->has_infoframe)
+		return;
+
+	crtc_state->infoframes.enable |=
+		intel_hdmi_infoframe_enable(HDMI_PACKET_TYPE_GENERAL_CONTROL);
+
+	/* Indicate color indication for deep color mode */
+	if (crtc_state->pipe_bpp > 24)
+		crtc_state->infoframes.gcp |= GCP_COLOR_INDICATION;
+
+	/* Enable default_phase whenever the display mode is suitably aligned */
+	if (gcp_default_phase_possible(crtc_state->pipe_bpp,
+				       &crtc_state->hw.adjusted_mode))
+		crtc_state->infoframes.gcp |= GCP_DEFAULT_PHASE_ENABLE;
+}
+
+static void ibx_set_infoframes(struct intel_encoder *encoder,
+			       bool enable,
+			       const struct intel_crtc_state *crtc_state,
+			       const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_hdmi *intel_hdmi = &dig_port->hdmi;
+	i915_reg_t reg = TVIDEO_DIP_CTL(crtc->pipe);
+	u32 val = intel_de_read(dev_priv, reg);
+	u32 port = VIDEO_DIP_PORT(encoder->port);
+
+	assert_hdmi_port_disabled(intel_hdmi);
+
+	/* See the big comment in g4x_set_infoframes() */
+	val |= VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC;
+
+	if (!enable) {
+		if (!(val & VIDEO_DIP_ENABLE))
+			return;
+		val &= ~(VIDEO_DIP_ENABLE | VIDEO_DIP_ENABLE_AVI |
+			 VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT |
+			 VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP);
+		intel_de_write(dev_priv, reg, val);
+		intel_de_posting_read(dev_priv, reg);
+		return;
+	}
+
+	if (port != (val & VIDEO_DIP_PORT_MASK)) {
+		drm_WARN(&dev_priv->drm, val & VIDEO_DIP_ENABLE,
+			 "DIP already enabled on port %c\n",
+			 (val & VIDEO_DIP_PORT_MASK) >> 29);
+		val &= ~VIDEO_DIP_PORT_MASK;
+		val |= port;
+	}
+
+	val |= VIDEO_DIP_ENABLE;
+	val &= ~(VIDEO_DIP_ENABLE_AVI |
+		 VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT |
+		 VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP);
+
+	if (intel_hdmi_set_gcp_infoframe(encoder, crtc_state, conn_state))
+		val |= VIDEO_DIP_ENABLE_GCP;
+
+	intel_de_write(dev_priv, reg, val);
+	intel_de_posting_read(dev_priv, reg);
+
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_AVI,
+			      &crtc_state->infoframes.avi);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_SPD,
+			      &crtc_state->infoframes.spd);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_VENDOR,
+			      &crtc_state->infoframes.hdmi);
+}
+
+static void cpt_set_infoframes(struct intel_encoder *encoder,
+			       bool enable,
+			       const struct intel_crtc_state *crtc_state,
+			       const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	i915_reg_t reg = TVIDEO_DIP_CTL(crtc->pipe);
+	u32 val = intel_de_read(dev_priv, reg);
+
+	assert_hdmi_port_disabled(intel_hdmi);
+
+	/* See the big comment in g4x_set_infoframes() */
+	val |= VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC;
+
+	if (!enable) {
+		if (!(val & VIDEO_DIP_ENABLE))
+			return;
+		val &= ~(VIDEO_DIP_ENABLE | VIDEO_DIP_ENABLE_AVI |
+			 VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT |
+			 VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP);
+		intel_de_write(dev_priv, reg, val);
+		intel_de_posting_read(dev_priv, reg);
+		return;
+	}
+
+	/* Set both together, unset both together: see the spec. */
+	val |= VIDEO_DIP_ENABLE | VIDEO_DIP_ENABLE_AVI;
+	val &= ~(VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT |
+		 VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP);
+
+	if (intel_hdmi_set_gcp_infoframe(encoder, crtc_state, conn_state))
+		val |= VIDEO_DIP_ENABLE_GCP;
+
+	intel_de_write(dev_priv, reg, val);
+	intel_de_posting_read(dev_priv, reg);
+
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_AVI,
+			      &crtc_state->infoframes.avi);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_SPD,
+			      &crtc_state->infoframes.spd);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_VENDOR,
+			      &crtc_state->infoframes.hdmi);
+}
+
+static void vlv_set_infoframes(struct intel_encoder *encoder,
+			       bool enable,
+			       const struct intel_crtc_state *crtc_state,
+			       const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	i915_reg_t reg = VLV_TVIDEO_DIP_CTL(crtc->pipe);
+	u32 val = intel_de_read(dev_priv, reg);
+	u32 port = VIDEO_DIP_PORT(encoder->port);
+
+	assert_hdmi_port_disabled(intel_hdmi);
+
+	/* See the big comment in g4x_set_infoframes() */
+	val |= VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC;
+
+	if (!enable) {
+		if (!(val & VIDEO_DIP_ENABLE))
+			return;
+		val &= ~(VIDEO_DIP_ENABLE | VIDEO_DIP_ENABLE_AVI |
+			 VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT |
+			 VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP);
+		intel_de_write(dev_priv, reg, val);
+		intel_de_posting_read(dev_priv, reg);
+		return;
+	}
+
+	if (port != (val & VIDEO_DIP_PORT_MASK)) {
+		drm_WARN(&dev_priv->drm, val & VIDEO_DIP_ENABLE,
+			 "DIP already enabled on port %c\n",
+			 (val & VIDEO_DIP_PORT_MASK) >> 29);
+		val &= ~VIDEO_DIP_PORT_MASK;
+		val |= port;
+	}
+
+	val |= VIDEO_DIP_ENABLE;
+	val &= ~(VIDEO_DIP_ENABLE_AVI |
+		 VIDEO_DIP_ENABLE_VENDOR | VIDEO_DIP_ENABLE_GAMUT |
+		 VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_ENABLE_GCP);
+
+	if (intel_hdmi_set_gcp_infoframe(encoder, crtc_state, conn_state))
+		val |= VIDEO_DIP_ENABLE_GCP;
+
+	intel_de_write(dev_priv, reg, val);
+	intel_de_posting_read(dev_priv, reg);
+
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_AVI,
+			      &crtc_state->infoframes.avi);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_SPD,
+			      &crtc_state->infoframes.spd);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_VENDOR,
+			      &crtc_state->infoframes.hdmi);
+}
+
+static void hsw_set_infoframes(struct intel_encoder *encoder,
+			       bool enable,
+			       const struct intel_crtc_state *crtc_state,
+			       const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	i915_reg_t reg = HSW_TVIDEO_DIP_CTL(crtc_state->cpu_transcoder);
+	u32 val = intel_de_read(dev_priv, reg);
+
+	assert_hdmi_transcoder_func_disabled(dev_priv,
+					     crtc_state->cpu_transcoder);
+
+	val &= ~(VIDEO_DIP_ENABLE_VSC_HSW | VIDEO_DIP_ENABLE_AVI_HSW |
+		 VIDEO_DIP_ENABLE_GCP_HSW | VIDEO_DIP_ENABLE_VS_HSW |
+		 VIDEO_DIP_ENABLE_GMP_HSW | VIDEO_DIP_ENABLE_SPD_HSW |
+		 VIDEO_DIP_ENABLE_DRM_GLK);
+
+	if (!enable) {
+		intel_de_write(dev_priv, reg, val);
+		intel_de_posting_read(dev_priv, reg);
+		return;
+	}
+
+	if (intel_hdmi_set_gcp_infoframe(encoder, crtc_state, conn_state))
+		val |= VIDEO_DIP_ENABLE_GCP_HSW;
+
+	intel_de_write(dev_priv, reg, val);
+	intel_de_posting_read(dev_priv, reg);
+
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_AVI,
+			      &crtc_state->infoframes.avi);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_SPD,
+			      &crtc_state->infoframes.spd);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_VENDOR,
+			      &crtc_state->infoframes.hdmi);
+	intel_write_infoframe(encoder, crtc_state,
+			      HDMI_INFOFRAME_TYPE_DRM,
+			      &crtc_state->infoframes.drm);
+}
+
+void intel_dp_dual_mode_set_tmds_output(struct intel_hdmi *hdmi, bool enable)
+{
+	struct drm_i915_private *dev_priv = intel_hdmi_to_i915(hdmi);
+	struct i2c_adapter *adapter;
+
+	if (hdmi->dp_dual_mode.type < DRM_DP_DUAL_MODE_TYPE2_DVI)
+		return;
+
+	adapter = intel_gmbus_get_adapter(dev_priv, hdmi->ddc_bus);
+
+	drm_dbg_kms(&dev_priv->drm, "%s DP dual mode adaptor TMDS output\n",
+		    enable ? "Enabling" : "Disabling");
+
+	drm_dp_dual_mode_set_tmds_output(&dev_priv->drm, hdmi->dp_dual_mode.type, adapter, enable);
+}
+
+static int intel_hdmi_hdcp_read(struct intel_digital_port *dig_port,
+				unsigned int offset, void *buffer, size_t size)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_hdmi *hdmi = &dig_port->hdmi;
+	struct i2c_adapter *adapter = intel_gmbus_get_adapter(i915,
+							      hdmi->ddc_bus);
+	int ret;
+	u8 start = offset & 0xff;
+	struct i2c_msg msgs[] = {
+		{
+			.addr = DRM_HDCP_DDC_ADDR,
+			.flags = 0,
+			.len = 1,
+			.buf = &start,
+		},
+		{
+			.addr = DRM_HDCP_DDC_ADDR,
+			.flags = I2C_M_RD,
+			.len = size,
+			.buf = buffer
+		}
+	};
+	ret = i2c_transfer(adapter, msgs, ARRAY_SIZE(msgs));
+	if (ret == ARRAY_SIZE(msgs))
+		return 0;
+	return ret >= 0 ? -EIO : ret;
+}
+
+static int intel_hdmi_hdcp_write(struct intel_digital_port *dig_port,
+				 unsigned int offset, void *buffer, size_t size)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_hdmi *hdmi = &dig_port->hdmi;
+	struct i2c_adapter *adapter = intel_gmbus_get_adapter(i915,
+							      hdmi->ddc_bus);
+	int ret;
+	u8 *write_buf;
+	struct i2c_msg msg;
+
+	write_buf = kzalloc(size + 1, GFP_KERNEL);
+	if (!write_buf)
+		return -ENOMEM;
+
+	write_buf[0] = offset & 0xff;
+	memcpy(&write_buf[1], buffer, size);
+
+	msg.addr = DRM_HDCP_DDC_ADDR;
+	msg.flags = 0,
+	msg.len = size + 1,
+	msg.buf = write_buf;
+
+	ret = i2c_transfer(adapter, &msg, 1);
+	if (ret == 1)
+		ret = 0;
+	else if (ret >= 0)
+		ret = -EIO;
+
+	kfree(write_buf);
+	return ret;
+}
+
+static
+int intel_hdmi_hdcp_write_an_aksv(struct intel_digital_port *dig_port,
+				  u8 *an)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_hdmi *hdmi = &dig_port->hdmi;
+	struct i2c_adapter *adapter = intel_gmbus_get_adapter(i915,
+							      hdmi->ddc_bus);
+	int ret;
+
+	ret = intel_hdmi_hdcp_write(dig_port, DRM_HDCP_DDC_AN, an,
+				    DRM_HDCP_AN_LEN);
+	if (ret) {
+		drm_dbg_kms(&i915->drm, "Write An over DDC failed (%d)\n",
+			    ret);
+		return ret;
+	}
+
+	ret = intel_gmbus_output_aksv(adapter);
+	if (ret < 0) {
+		drm_dbg_kms(&i915->drm, "Failed to output aksv (%d)\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static int intel_hdmi_hdcp_read_bksv(struct intel_digital_port *dig_port,
+				     u8 *bksv)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	int ret;
+	ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_BKSV, bksv,
+				   DRM_HDCP_KSV_LEN);
+	if (ret)
+		drm_dbg_kms(&i915->drm, "Read Bksv over DDC failed (%d)\n",
+			    ret);
+	return ret;
+}
+
+static
+int intel_hdmi_hdcp_read_bstatus(struct intel_digital_port *dig_port,
+				 u8 *bstatus)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	int ret;
+	ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_BSTATUS,
+				   bstatus, DRM_HDCP_BSTATUS_LEN);
+	if (ret)
+		drm_dbg_kms(&i915->drm, "Read bstatus over DDC failed (%d)\n",
+			    ret);
+	return ret;
+}
+
+static
+int intel_hdmi_hdcp_repeater_present(struct intel_digital_port *dig_port,
+				     bool *repeater_present)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	int ret;
+	u8 val;
+
+	ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_BCAPS, &val, 1);
+	if (ret) {
+		drm_dbg_kms(&i915->drm, "Read bcaps over DDC failed (%d)\n",
+			    ret);
+		return ret;
+	}
+	*repeater_present = val & DRM_HDCP_DDC_BCAPS_REPEATER_PRESENT;
+	return 0;
+}
+
+static
+int intel_hdmi_hdcp_read_ri_prime(struct intel_digital_port *dig_port,
+				  u8 *ri_prime)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	int ret;
+	ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_RI_PRIME,
+				   ri_prime, DRM_HDCP_RI_LEN);
+	if (ret)
+		drm_dbg_kms(&i915->drm, "Read Ri' over DDC failed (%d)\n",
+			    ret);
+	return ret;
+}
+
+static
+int intel_hdmi_hdcp_read_ksv_ready(struct intel_digital_port *dig_port,
+				   bool *ksv_ready)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	int ret;
+	u8 val;
+
+	ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_BCAPS, &val, 1);
+	if (ret) {
+		drm_dbg_kms(&i915->drm, "Read bcaps over DDC failed (%d)\n",
+			    ret);
+		return ret;
+	}
+	*ksv_ready = val & DRM_HDCP_DDC_BCAPS_KSV_FIFO_READY;
+	return 0;
+}
+
+static
+int intel_hdmi_hdcp_read_ksv_fifo(struct intel_digital_port *dig_port,
+				  int num_downstream, u8 *ksv_fifo)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	int ret;
+	ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_KSV_FIFO,
+				   ksv_fifo, num_downstream * DRM_HDCP_KSV_LEN);
+	if (ret) {
+		drm_dbg_kms(&i915->drm,
+			    "Read ksv fifo over DDC failed (%d)\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static
+int intel_hdmi_hdcp_read_v_prime_part(struct intel_digital_port *dig_port,
+				      int i, u32 *part)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	int ret;
+
+	if (i >= DRM_HDCP_V_PRIME_NUM_PARTS)
+		return -EINVAL;
+
+	ret = intel_hdmi_hdcp_read(dig_port, DRM_HDCP_DDC_V_PRIME(i),
+				   part, DRM_HDCP_V_PRIME_PART_LEN);
+	if (ret)
+		drm_dbg_kms(&i915->drm, "Read V'[%d] over DDC failed (%d)\n",
+			    i, ret);
+	return ret;
+}
+
+static int kbl_repositioning_enc_en_signal(struct intel_connector *connector,
+					   enum transcoder cpu_transcoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct intel_digital_port *dig_port = intel_attached_dig_port(connector);
+	struct intel_crtc *crtc = to_intel_crtc(connector->base.state->crtc);
+	u32 scanline;
+	int ret;
+
+	for (;;) {
+		scanline = intel_de_read(dev_priv, PIPEDSL(crtc->pipe));
+		if (scanline > 100 && scanline < 200)
+			break;
+		usleep_range(25, 50);
+	}
+
+	ret = intel_ddi_toggle_hdcp_bits(&dig_port->base, cpu_transcoder,
+					 false, TRANS_DDI_HDCP_SIGNALLING);
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"Disable HDCP signalling failed (%d)\n", ret);
+		return ret;
+	}
+
+	ret = intel_ddi_toggle_hdcp_bits(&dig_port->base, cpu_transcoder,
+					 true, TRANS_DDI_HDCP_SIGNALLING);
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"Enable HDCP signalling failed (%d)\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static
+int intel_hdmi_hdcp_toggle_signalling(struct intel_digital_port *dig_port,
+				      enum transcoder cpu_transcoder,
+				      bool enable)
+{
+	struct intel_hdmi *hdmi = &dig_port->hdmi;
+	struct intel_connector *connector = hdmi->attached_connector;
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	int ret;
+
+	if (!enable)
+		usleep_range(6, 60); /* Bspec says >= 6us */
+
+	ret = intel_ddi_toggle_hdcp_bits(&dig_port->base,
+					 cpu_transcoder, enable,
+					 TRANS_DDI_HDCP_SIGNALLING);
+	if (ret) {
+		drm_err(&dev_priv->drm, "%s HDCP signalling failed (%d)\n",
+			enable ? "Enable" : "Disable", ret);
+		return ret;
+	}
+
+	/*
+	 * WA: To fix incorrect positioning of the window of
+	 * opportunity and enc_en signalling in KABYLAKE.
+	 */
+	if (IS_KABYLAKE(dev_priv) && enable)
+		return kbl_repositioning_enc_en_signal(connector,
+						       cpu_transcoder);
+
+	return 0;
+}
+
+static
+bool intel_hdmi_hdcp_check_link_once(struct intel_digital_port *dig_port,
+				     struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	enum port port = dig_port->base.port;
+	enum transcoder cpu_transcoder = connector->hdcp.cpu_transcoder;
+	int ret;
+	union {
+		u32 reg;
+		u8 shim[DRM_HDCP_RI_LEN];
+	} ri;
+
+	ret = intel_hdmi_hdcp_read_ri_prime(dig_port, ri.shim);
+	if (ret)
+		return false;
+
+	intel_de_write(i915, HDCP_RPRIME(i915, cpu_transcoder, port), ri.reg);
+
+	/* Wait for Ri prime match */
+	if (wait_for((intel_de_read(i915, HDCP_STATUS(i915, cpu_transcoder, port)) &
+		      (HDCP_STATUS_RI_MATCH | HDCP_STATUS_ENC)) ==
+		     (HDCP_STATUS_RI_MATCH | HDCP_STATUS_ENC), 1)) {
+		drm_dbg_kms(&i915->drm, "Ri' mismatch detected (%x)\n",
+			intel_de_read(i915, HDCP_STATUS(i915, cpu_transcoder,
+							port)));
+		return false;
+	}
+	return true;
+}
+
+static
+bool intel_hdmi_hdcp_check_link(struct intel_digital_port *dig_port,
+				struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	int retry;
+
+	for (retry = 0; retry < 3; retry++)
+		if (intel_hdmi_hdcp_check_link_once(dig_port, connector))
+			return true;
+
+	drm_err(&i915->drm, "Link check failed\n");
+	return false;
+}
+
+struct hdcp2_hdmi_msg_timeout {
+	u8 msg_id;
+	u16 timeout;
+};
+
+static const struct hdcp2_hdmi_msg_timeout hdcp2_msg_timeout[] = {
+	{ HDCP_2_2_AKE_SEND_CERT, HDCP_2_2_CERT_TIMEOUT_MS, },
+	{ HDCP_2_2_AKE_SEND_PAIRING_INFO, HDCP_2_2_PAIRING_TIMEOUT_MS, },
+	{ HDCP_2_2_LC_SEND_LPRIME, HDCP_2_2_HDMI_LPRIME_TIMEOUT_MS, },
+	{ HDCP_2_2_REP_SEND_RECVID_LIST, HDCP_2_2_RECVID_LIST_TIMEOUT_MS, },
+	{ HDCP_2_2_REP_STREAM_READY, HDCP_2_2_STREAM_READY_TIMEOUT_MS, },
+};
+
+static
+int intel_hdmi_hdcp2_read_rx_status(struct intel_digital_port *dig_port,
+				    u8 *rx_status)
+{
+	return intel_hdmi_hdcp_read(dig_port,
+				    HDCP_2_2_HDMI_REG_RXSTATUS_OFFSET,
+				    rx_status,
+				    HDCP_2_2_HDMI_RXSTATUS_LEN);
+}
+
+static int get_hdcp2_msg_timeout(u8 msg_id, bool is_paired)
+{
+	int i;
+
+	if (msg_id == HDCP_2_2_AKE_SEND_HPRIME) {
+		if (is_paired)
+			return HDCP_2_2_HPRIME_PAIRED_TIMEOUT_MS;
+		else
+			return HDCP_2_2_HPRIME_NO_PAIRED_TIMEOUT_MS;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(hdcp2_msg_timeout); i++) {
+		if (hdcp2_msg_timeout[i].msg_id == msg_id)
+			return hdcp2_msg_timeout[i].timeout;
+	}
+
+	return -EINVAL;
+}
+
+static int
+hdcp2_detect_msg_availability(struct intel_digital_port *dig_port,
+			      u8 msg_id, bool *msg_ready,
+			      ssize_t *msg_sz)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	u8 rx_status[HDCP_2_2_HDMI_RXSTATUS_LEN];
+	int ret;
+
+	ret = intel_hdmi_hdcp2_read_rx_status(dig_port, rx_status);
+	if (ret < 0) {
+		drm_dbg_kms(&i915->drm, "rx_status read failed. Err %d\n",
+			    ret);
+		return ret;
+	}
+
+	*msg_sz = ((HDCP_2_2_HDMI_RXSTATUS_MSG_SZ_HI(rx_status[1]) << 8) |
+		  rx_status[0]);
+
+	if (msg_id == HDCP_2_2_REP_SEND_RECVID_LIST)
+		*msg_ready = (HDCP_2_2_HDMI_RXSTATUS_READY(rx_status[1]) &&
+			     *msg_sz);
+	else
+		*msg_ready = *msg_sz;
+
+	return 0;
+}
+
+static ssize_t
+intel_hdmi_hdcp2_wait_for_msg(struct intel_digital_port *dig_port,
+			      u8 msg_id, bool paired)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	bool msg_ready = false;
+	int timeout, ret;
+	ssize_t msg_sz = 0;
+
+	timeout = get_hdcp2_msg_timeout(msg_id, paired);
+	if (timeout < 0)
+		return timeout;
+
+	ret = __wait_for(ret = hdcp2_detect_msg_availability(dig_port,
+							     msg_id, &msg_ready,
+							     &msg_sz),
+			 !ret && msg_ready && msg_sz, timeout * 1000,
+			 1000, 5 * 1000);
+	if (ret)
+		drm_dbg_kms(&i915->drm, "msg_id: %d, ret: %d, timeout: %d\n",
+			    msg_id, ret, timeout);
+
+	return ret ? ret : msg_sz;
+}
+
+static
+int intel_hdmi_hdcp2_write_msg(struct intel_digital_port *dig_port,
+			       void *buf, size_t size)
+{
+	unsigned int offset;
+
+	offset = HDCP_2_2_HDMI_REG_WR_MSG_OFFSET;
+	return intel_hdmi_hdcp_write(dig_port, offset, buf, size);
+}
+
+static
+int intel_hdmi_hdcp2_read_msg(struct intel_digital_port *dig_port,
+			      u8 msg_id, void *buf, size_t size)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_hdmi *hdmi = &dig_port->hdmi;
+	struct intel_hdcp *hdcp = &hdmi->attached_connector->hdcp;
+	unsigned int offset;
+	ssize_t ret;
+
+	ret = intel_hdmi_hdcp2_wait_for_msg(dig_port, msg_id,
+					    hdcp->is_paired);
+	if (ret < 0)
+		return ret;
+
+	/*
+	 * Available msg size should be equal to or lesser than the
+	 * available buffer.
+	 */
+	if (ret > size) {
+		drm_dbg_kms(&i915->drm,
+			    "msg_sz(%zd) is more than exp size(%zu)\n",
+			    ret, size);
+		return -EINVAL;
+	}
+
+	offset = HDCP_2_2_HDMI_REG_RD_MSG_OFFSET;
+	ret = intel_hdmi_hdcp_read(dig_port, offset, buf, ret);
+	if (ret)
+		drm_dbg_kms(&i915->drm, "Failed to read msg_id: %d(%zd)\n",
+			    msg_id, ret);
+
+	return ret;
+}
+
+static
+int intel_hdmi_hdcp2_check_link(struct intel_digital_port *dig_port,
+				struct intel_connector *connector)
+{
+	u8 rx_status[HDCP_2_2_HDMI_RXSTATUS_LEN];
+	int ret;
+
+	ret = intel_hdmi_hdcp2_read_rx_status(dig_port, rx_status);
+	if (ret)
+		return ret;
+
+	/*
+	 * Re-auth request and Link Integrity Failures are represented by
+	 * same bit. i.e reauth_req.
+	 */
+	if (HDCP_2_2_HDMI_RXSTATUS_REAUTH_REQ(rx_status[1]))
+		ret = HDCP_REAUTH_REQUEST;
+	else if (HDCP_2_2_HDMI_RXSTATUS_READY(rx_status[1]))
+		ret = HDCP_TOPOLOGY_CHANGE;
+
+	return ret;
+}
+
+static
+int intel_hdmi_hdcp2_capable(struct intel_digital_port *dig_port,
+			     bool *capable)
+{
+	u8 hdcp2_version;
+	int ret;
+
+	*capable = false;
+	ret = intel_hdmi_hdcp_read(dig_port, HDCP_2_2_HDMI_REG_VER_OFFSET,
+				   &hdcp2_version, sizeof(hdcp2_version));
+	if (!ret && hdcp2_version & HDCP_2_2_HDMI_SUPPORT_MASK)
+		*capable = true;
+
+	return ret;
+}
+
+static const struct intel_hdcp_shim intel_hdmi_hdcp_shim = {
+	.write_an_aksv = intel_hdmi_hdcp_write_an_aksv,
+	.read_bksv = intel_hdmi_hdcp_read_bksv,
+	.read_bstatus = intel_hdmi_hdcp_read_bstatus,
+	.repeater_present = intel_hdmi_hdcp_repeater_present,
+	.read_ri_prime = intel_hdmi_hdcp_read_ri_prime,
+	.read_ksv_ready = intel_hdmi_hdcp_read_ksv_ready,
+	.read_ksv_fifo = intel_hdmi_hdcp_read_ksv_fifo,
+	.read_v_prime_part = intel_hdmi_hdcp_read_v_prime_part,
+	.toggle_signalling = intel_hdmi_hdcp_toggle_signalling,
+	.check_link = intel_hdmi_hdcp_check_link,
+	.write_2_2_msg = intel_hdmi_hdcp2_write_msg,
+	.read_2_2_msg = intel_hdmi_hdcp2_read_msg,
+	.check_2_2_link	= intel_hdmi_hdcp2_check_link,
+	.hdcp_2_2_capable = intel_hdmi_hdcp2_capable,
+	.protocol = HDCP_PROTOCOL_HDMI,
+};
+
+static int intel_hdmi_source_max_tmds_clock(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	int max_tmds_clock, vbt_max_tmds_clock;
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		max_tmds_clock = 594000;
+	else if (DISPLAY_VER(dev_priv) >= 8 || IS_HASWELL(dev_priv))
+		max_tmds_clock = 300000;
+	else if (DISPLAY_VER(dev_priv) >= 5)
+		max_tmds_clock = 225000;
+	else
+		max_tmds_clock = 165000;
+
+	vbt_max_tmds_clock = intel_bios_max_tmds_clock(encoder);
+	if (vbt_max_tmds_clock)
+		max_tmds_clock = min(max_tmds_clock, vbt_max_tmds_clock);
+
+	return max_tmds_clock;
+}
+
+static bool intel_has_hdmi_sink(struct intel_hdmi *hdmi,
+				const struct drm_connector_state *conn_state)
+{
+	return hdmi->has_hdmi_sink &&
+		READ_ONCE(to_intel_digital_connector_state(conn_state)->force_audio) != HDMI_AUDIO_OFF_DVI;
+}
+
+static bool intel_hdmi_is_ycbcr420(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR420;
+}
+
+static int hdmi_port_clock_limit(struct intel_hdmi *hdmi,
+				 bool respect_downstream_limits,
+				 bool has_hdmi_sink)
+{
+	struct intel_encoder *encoder = &hdmi_to_dig_port(hdmi)->base;
+	int max_tmds_clock = intel_hdmi_source_max_tmds_clock(encoder);
+
+	if (respect_downstream_limits) {
+		struct intel_connector *connector = hdmi->attached_connector;
+		const struct drm_display_info *info = &connector->base.display_info;
+
+		if (hdmi->dp_dual_mode.max_tmds_clock)
+			max_tmds_clock = min(max_tmds_clock,
+					     hdmi->dp_dual_mode.max_tmds_clock);
+
+		if (info->max_tmds_clock)
+			max_tmds_clock = min(max_tmds_clock,
+					     info->max_tmds_clock);
+		else if (!has_hdmi_sink)
+			max_tmds_clock = min(max_tmds_clock, 165000);
+	}
+
+	return max_tmds_clock;
+}
+
+static enum drm_mode_status
+hdmi_port_clock_valid(struct intel_hdmi *hdmi,
+		      int clock, bool respect_downstream_limits,
+		      bool has_hdmi_sink)
+{
+	struct drm_i915_private *dev_priv = intel_hdmi_to_i915(hdmi);
+	enum phy phy = intel_port_to_phy(dev_priv, hdmi_to_dig_port(hdmi)->base.port);
+
+	if (clock < 25000)
+		return MODE_CLOCK_LOW;
+	if (clock > hdmi_port_clock_limit(hdmi, respect_downstream_limits,
+					  has_hdmi_sink))
+		return MODE_CLOCK_HIGH;
+
+	/* GLK DPLL can't generate 446-480 MHz */
+	if (IS_GEMINILAKE(dev_priv) && clock > 446666 && clock < 480000)
+		return MODE_CLOCK_RANGE;
+
+	/* BXT/GLK DPLL can't generate 223-240 MHz */
+	if ((IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) &&
+	    clock > 223333 && clock < 240000)
+		return MODE_CLOCK_RANGE;
+
+	/* CHV DPLL can't generate 216-240 MHz */
+	if (IS_CHERRYVIEW(dev_priv) && clock > 216000 && clock < 240000)
+		return MODE_CLOCK_RANGE;
+
+	/* ICL+ combo PHY PLL can't generate 500-533.2 MHz */
+	if (intel_phy_is_combo(dev_priv, phy) && clock > 500000 && clock < 533200)
+		return MODE_CLOCK_RANGE;
+
+	/* ICL+ TC PHY PLL can't generate 500-532.8 MHz */
+	if (intel_phy_is_tc(dev_priv, phy) && clock > 500000 && clock < 532800)
+		return MODE_CLOCK_RANGE;
+
+	/*
+	 * SNPS PHYs' MPLLB table-based programming can only handle a fixed
+	 * set of link rates.
+	 *
+	 * FIXME: We will hopefully get an algorithmic way of programming
+	 * the MPLLB for HDMI in the future.
+	 */
+	if (IS_DG2(dev_priv))
+		return intel_snps_phy_check_hdmi_link_rate(clock);
+
+	return MODE_OK;
+}
+
+int intel_hdmi_tmds_clock(int clock, int bpc, bool ycbcr420_output)
+{
+	/* YCBCR420 TMDS rate requirement is half the pixel clock */
+	if (ycbcr420_output)
+		clock /= 2;
+
+	/*
+	 * Need to adjust the port link by:
+	 *  1.5x for 12bpc
+	 *  1.25x for 10bpc
+	 */
+	return DIV_ROUND_CLOSEST(clock * bpc, 8);
+}
+
+static bool intel_hdmi_source_bpc_possible(struct drm_i915_private *i915, int bpc)
+{
+	switch (bpc) {
+	case 12:
+		return !HAS_GMCH(i915);
+	case 10:
+		return DISPLAY_VER(i915) >= 11;
+	case 8:
+		return true;
+	default:
+		MISSING_CASE(bpc);
+		return false;
+	}
+}
+
+static bool intel_hdmi_sink_bpc_possible(struct drm_connector *connector,
+					 int bpc, bool has_hdmi_sink, bool ycbcr420_output)
+{
+	const struct drm_display_info *info = &connector->display_info;
+	const struct drm_hdmi_info *hdmi = &info->hdmi;
+
+	switch (bpc) {
+	case 12:
+		if (!has_hdmi_sink)
+			return false;
+
+		if (ycbcr420_output)
+			return hdmi->y420_dc_modes & DRM_EDID_YCBCR420_DC_36;
+		else
+			return info->edid_hdmi_rgb444_dc_modes & DRM_EDID_HDMI_DC_36;
+	case 10:
+		if (!has_hdmi_sink)
+			return false;
+
+		if (ycbcr420_output)
+			return hdmi->y420_dc_modes & DRM_EDID_YCBCR420_DC_30;
+		else
+			return info->edid_hdmi_rgb444_dc_modes & DRM_EDID_HDMI_DC_30;
+	case 8:
+		return true;
+	default:
+		MISSING_CASE(bpc);
+		return false;
+	}
+}
+
+static enum drm_mode_status
+intel_hdmi_mode_clock_valid(struct drm_connector *connector, int clock,
+			    bool has_hdmi_sink, bool ycbcr420_output)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct intel_hdmi *hdmi = intel_attached_hdmi(to_intel_connector(connector));
+	enum drm_mode_status status = MODE_OK;
+	int bpc;
+
+	/*
+	 * Try all color depths since valid port clock range
+	 * can have holes. Any mode that can be used with at
+	 * least one color depth is accepted.
+	 */
+	for (bpc = 12; bpc >= 8; bpc -= 2) {
+		int tmds_clock = intel_hdmi_tmds_clock(clock, bpc, ycbcr420_output);
+
+		if (!intel_hdmi_source_bpc_possible(i915, bpc))
+			continue;
+
+		if (!intel_hdmi_sink_bpc_possible(connector, bpc, has_hdmi_sink, ycbcr420_output))
+			continue;
+
+		status = hdmi_port_clock_valid(hdmi, tmds_clock, true, has_hdmi_sink);
+		if (status == MODE_OK)
+			return MODE_OK;
+	}
+
+	/* can never happen */
+	drm_WARN_ON(&i915->drm, status == MODE_OK);
+
+	return status;
+}
+
+static enum drm_mode_status
+intel_hdmi_mode_valid(struct drm_connector *connector,
+		      struct drm_display_mode *mode)
+{
+	struct intel_hdmi *hdmi = intel_attached_hdmi(to_intel_connector(connector));
+	struct drm_i915_private *dev_priv = intel_hdmi_to_i915(hdmi);
+	enum drm_mode_status status;
+	int clock = mode->clock;
+	int max_dotclk = to_i915(connector->dev)->max_dotclk_freq;
+	bool has_hdmi_sink = intel_has_hdmi_sink(hdmi, connector->state);
+	bool ycbcr_420_only;
+
+	status = intel_cpu_transcoder_mode_valid(dev_priv, mode);
+	if (status != MODE_OK)
+		return status;
+
+	if ((mode->flags & DRM_MODE_FLAG_3D_MASK) == DRM_MODE_FLAG_3D_FRAME_PACKING)
+		clock *= 2;
+
+	if (clock > max_dotclk)
+		return MODE_CLOCK_HIGH;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLCLK) {
+		if (!has_hdmi_sink)
+			return MODE_CLOCK_LOW;
+		clock *= 2;
+	}
+
+	/*
+	 * HDMI2.1 requires higher resolution modes like 8k60, 4K120 to be
+	 * enumerated only if FRL is supported. Current platforms do not support
+	 * FRL so prune the higher resolution modes that require doctclock more
+	 * than 600MHz.
+	 */
+	if (clock > 600000)
+		return MODE_CLOCK_HIGH;
+
+	ycbcr_420_only = drm_mode_is_420_only(&connector->display_info, mode);
+
+	status = intel_hdmi_mode_clock_valid(connector, clock, has_hdmi_sink, ycbcr_420_only);
+	if (status != MODE_OK) {
+		if (ycbcr_420_only ||
+		    !connector->ycbcr_420_allowed ||
+		    !drm_mode_is_420_also(&connector->display_info, mode))
+			return status;
+
+		status = intel_hdmi_mode_clock_valid(connector, clock, has_hdmi_sink, true);
+		if (status != MODE_OK)
+			return status;
+	}
+
+	return intel_mode_valid_max_plane_size(dev_priv, mode, false);
+}
+
+bool intel_hdmi_bpc_possible(const struct intel_crtc_state *crtc_state,
+			     int bpc, bool has_hdmi_sink, bool ycbcr420_output)
+{
+	struct drm_atomic_state *state = crtc_state->uapi.state;
+	struct drm_connector_state *connector_state;
+	struct drm_connector *connector;
+	int i;
+
+	for_each_new_connector_in_state(state, connector, connector_state, i) {
+		if (connector_state->crtc != crtc_state->uapi.crtc)
+			continue;
+
+		if (!intel_hdmi_sink_bpc_possible(connector, bpc, has_hdmi_sink, ycbcr420_output))
+			return false;
+	}
+
+	return true;
+}
+
+static bool hdmi_bpc_possible(const struct intel_crtc_state *crtc_state, int bpc)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(crtc_state->uapi.crtc->dev);
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+
+	if (!intel_hdmi_source_bpc_possible(dev_priv, bpc))
+		return false;
+
+	/*
+	 * HDMI deep color affects the clocks, so it's only possible
+	 * when not cloning with other encoder types.
+	 */
+	if (bpc > 8 && crtc_state->output_types != BIT(INTEL_OUTPUT_HDMI))
+		return false;
+
+	/* Display Wa_1405510057:icl,ehl */
+	if (intel_hdmi_is_ycbcr420(crtc_state) &&
+	    bpc == 10 && DISPLAY_VER(dev_priv) == 11 &&
+	    (adjusted_mode->crtc_hblank_end -
+	     adjusted_mode->crtc_hblank_start) % 8 == 2)
+		return false;
+
+	return intel_hdmi_bpc_possible(crtc_state, bpc, crtc_state->has_hdmi_sink,
+				       intel_hdmi_is_ycbcr420(crtc_state));
+}
+
+static int intel_hdmi_compute_bpc(struct intel_encoder *encoder,
+				  struct intel_crtc_state *crtc_state,
+				  int clock, bool respect_downstream_limits)
+{
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	bool ycbcr420_output = intel_hdmi_is_ycbcr420(crtc_state);
+	int bpc;
+
+	/*
+	 * pipe_bpp could already be below 8bpc due to FDI
+	 * bandwidth constraints. HDMI minimum is 8bpc however.
+	 */
+	bpc = max(crtc_state->pipe_bpp / 3, 8);
+
+	/*
+	 * We will never exceed downstream TMDS clock limits while
+	 * attempting deep color. If the user insists on forcing an
+	 * out of spec mode they will have to be satisfied with 8bpc.
+	 */
+	if (!respect_downstream_limits)
+		bpc = 8;
+
+	for (; bpc >= 8; bpc -= 2) {
+		int tmds_clock = intel_hdmi_tmds_clock(clock, bpc, ycbcr420_output);
+
+		if (hdmi_bpc_possible(crtc_state, bpc) &&
+		    hdmi_port_clock_valid(intel_hdmi, tmds_clock,
+					  respect_downstream_limits,
+					  crtc_state->has_hdmi_sink) == MODE_OK)
+			return bpc;
+	}
+
+	return -EINVAL;
+}
+
+static int intel_hdmi_compute_clock(struct intel_encoder *encoder,
+				    struct intel_crtc_state *crtc_state,
+				    bool respect_downstream_limits)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	int bpc, clock = adjusted_mode->crtc_clock;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK)
+		clock *= 2;
+
+	bpc = intel_hdmi_compute_bpc(encoder, crtc_state, clock,
+				     respect_downstream_limits);
+	if (bpc < 0)
+		return bpc;
+
+	crtc_state->port_clock =
+		intel_hdmi_tmds_clock(clock, bpc, intel_hdmi_is_ycbcr420(crtc_state));
+
+	/*
+	 * pipe_bpp could already be below 8bpc due to
+	 * FDI bandwidth constraints. We shouldn't bump it
+	 * back up to the HDMI minimum 8bpc in that case.
+	 */
+	crtc_state->pipe_bpp = min(crtc_state->pipe_bpp, bpc * 3);
+
+	drm_dbg_kms(&i915->drm,
+		    "picking %d bpc for HDMI output (pipe bpp: %d)\n",
+		    bpc, crtc_state->pipe_bpp);
+
+	return 0;
+}
+
+bool intel_hdmi_limited_color_range(const struct intel_crtc_state *crtc_state,
+				    const struct drm_connector_state *conn_state)
+{
+	const struct intel_digital_connector_state *intel_conn_state =
+		to_intel_digital_connector_state(conn_state);
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+
+	/*
+	 * Our YCbCr output is always limited range.
+	 * crtc_state->limited_color_range only applies to RGB,
+	 * and it must never be set for YCbCr or we risk setting
+	 * some conflicting bits in PIPECONF which will mess up
+	 * the colors on the monitor.
+	 */
+	if (crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB)
+		return false;
+
+	if (intel_conn_state->broadcast_rgb == INTEL_BROADCAST_RGB_AUTO) {
+		/* See CEA-861-E - 5.1 Default Encoding Parameters */
+		return crtc_state->has_hdmi_sink &&
+			drm_default_rgb_quant_range(adjusted_mode) ==
+			HDMI_QUANTIZATION_RANGE_LIMITED;
+	} else {
+		return intel_conn_state->broadcast_rgb == INTEL_BROADCAST_RGB_LIMITED;
+	}
+}
+
+static bool intel_hdmi_has_audio(struct intel_encoder *encoder,
+				 const struct intel_crtc_state *crtc_state,
+				 const struct drm_connector_state *conn_state)
+{
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	const struct intel_digital_connector_state *intel_conn_state =
+		to_intel_digital_connector_state(conn_state);
+
+	if (!crtc_state->has_hdmi_sink)
+		return false;
+
+	if (intel_conn_state->force_audio == HDMI_AUDIO_AUTO)
+		return intel_hdmi->has_audio;
+	else
+		return intel_conn_state->force_audio == HDMI_AUDIO_ON;
+}
+
+static enum intel_output_format
+intel_hdmi_output_format(struct intel_connector *connector,
+			 bool ycbcr_420_output)
+{
+	if (connector->base.ycbcr_420_allowed && ycbcr_420_output)
+		return INTEL_OUTPUT_FORMAT_YCBCR420;
+	else
+		return INTEL_OUTPUT_FORMAT_RGB;
+}
+
+static int intel_hdmi_compute_output_format(struct intel_encoder *encoder,
+					    struct intel_crtc_state *crtc_state,
+					    const struct drm_connector_state *conn_state,
+					    bool respect_downstream_limits)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	const struct drm_display_info *info = &connector->base.display_info;
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	bool ycbcr_420_only = drm_mode_is_420_only(info, adjusted_mode);
+	int ret;
+
+	crtc_state->output_format = intel_hdmi_output_format(connector, ycbcr_420_only);
+
+	if (ycbcr_420_only && !intel_hdmi_is_ycbcr420(crtc_state)) {
+		drm_dbg_kms(&i915->drm,
+			    "YCbCr 4:2:0 mode but YCbCr 4:2:0 output not possible. Falling back to RGB.\n");
+		crtc_state->output_format = INTEL_OUTPUT_FORMAT_RGB;
+	}
+
+	ret = intel_hdmi_compute_clock(encoder, crtc_state, respect_downstream_limits);
+	if (ret) {
+		if (intel_hdmi_is_ycbcr420(crtc_state) ||
+		    !connector->base.ycbcr_420_allowed ||
+		    !drm_mode_is_420_also(info, adjusted_mode))
+			return ret;
+
+		crtc_state->output_format = intel_hdmi_output_format(connector, true);
+		ret = intel_hdmi_compute_clock(encoder, crtc_state, respect_downstream_limits);
+	}
+
+	return ret;
+}
+
+int intel_hdmi_compute_config(struct intel_encoder *encoder,
+			      struct intel_crtc_state *pipe_config,
+			      struct drm_connector_state *conn_state)
+{
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	struct drm_connector *connector = conn_state->connector;
+	struct drm_scdc *scdc = &connector->display_info.hdmi.scdc;
+	int ret;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+	pipe_config->has_hdmi_sink = intel_has_hdmi_sink(intel_hdmi,
+							 conn_state);
+
+	if (pipe_config->has_hdmi_sink)
+		pipe_config->has_infoframe = true;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK)
+		pipe_config->pixel_multiplier = 2;
+
+	if (HAS_PCH_SPLIT(dev_priv) && !HAS_DDI(dev_priv))
+		pipe_config->has_pch_encoder = true;
+
+	pipe_config->has_audio =
+		intel_hdmi_has_audio(encoder, pipe_config, conn_state);
+
+	/*
+	 * Try to respect downstream TMDS clock limits first, if
+	 * that fails assume the user might know something we don't.
+	 */
+	ret = intel_hdmi_compute_output_format(encoder, pipe_config, conn_state, true);
+	if (ret)
+		ret = intel_hdmi_compute_output_format(encoder, pipe_config, conn_state, false);
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "unsupported HDMI clock (%d kHz), rejecting mode\n",
+			    pipe_config->hw.adjusted_mode.crtc_clock);
+		return ret;
+	}
+
+	if (intel_hdmi_is_ycbcr420(pipe_config)) {
+		ret = intel_panel_fitting(pipe_config, conn_state);
+		if (ret)
+			return ret;
+	}
+
+	pipe_config->limited_color_range =
+		intel_hdmi_limited_color_range(pipe_config, conn_state);
+
+	if (conn_state->picture_aspect_ratio)
+		adjusted_mode->picture_aspect_ratio =
+			conn_state->picture_aspect_ratio;
+
+	pipe_config->lane_count = 4;
+
+	if (scdc->scrambling.supported && DISPLAY_VER(dev_priv) >= 10) {
+		if (scdc->scrambling.low_rates)
+			pipe_config->hdmi_scrambling = true;
+
+		if (pipe_config->port_clock > 340000) {
+			pipe_config->hdmi_scrambling = true;
+			pipe_config->hdmi_high_tmds_clock_ratio = true;
+		}
+	}
+
+	intel_hdmi_compute_gcp_infoframe(encoder, pipe_config,
+					 conn_state);
+
+	if (!intel_hdmi_compute_avi_infoframe(encoder, pipe_config, conn_state)) {
+		drm_dbg_kms(&dev_priv->drm, "bad AVI infoframe\n");
+		return -EINVAL;
+	}
+
+	if (!intel_hdmi_compute_spd_infoframe(encoder, pipe_config, conn_state)) {
+		drm_dbg_kms(&dev_priv->drm, "bad SPD infoframe\n");
+		return -EINVAL;
+	}
+
+	if (!intel_hdmi_compute_hdmi_infoframe(encoder, pipe_config, conn_state)) {
+		drm_dbg_kms(&dev_priv->drm, "bad HDMI infoframe\n");
+		return -EINVAL;
+	}
+
+	if (!intel_hdmi_compute_drm_infoframe(encoder, pipe_config, conn_state)) {
+		drm_dbg_kms(&dev_priv->drm, "bad DRM infoframe\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+void intel_hdmi_encoder_shutdown(struct intel_encoder *encoder)
+{
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+
+	/*
+	 * Give a hand to buggy BIOSen which forget to turn
+	 * the TMDS output buffers back on after a reboot.
+	 */
+	intel_dp_dual_mode_set_tmds_output(intel_hdmi, true);
+}
+
+static void
+intel_hdmi_unset_edid(struct drm_connector *connector)
+{
+	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(to_intel_connector(connector));
+
+	intel_hdmi->has_hdmi_sink = false;
+	intel_hdmi->has_audio = false;
+
+	intel_hdmi->dp_dual_mode.type = DRM_DP_DUAL_MODE_NONE;
+	intel_hdmi->dp_dual_mode.max_tmds_clock = 0;
+
+	kfree(to_intel_connector(connector)->detect_edid);
+	to_intel_connector(connector)->detect_edid = NULL;
+}
+
+static void
+intel_hdmi_dp_dual_mode_detect(struct drm_connector *connector, bool has_edid)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *hdmi = intel_attached_hdmi(to_intel_connector(connector));
+	enum port port = hdmi_to_dig_port(hdmi)->base.port;
+	struct i2c_adapter *adapter =
+		intel_gmbus_get_adapter(dev_priv, hdmi->ddc_bus);
+	enum drm_dp_dual_mode_type type = drm_dp_dual_mode_detect(&dev_priv->drm, adapter);
+
+	/*
+	 * Type 1 DVI adaptors are not required to implement any
+	 * registers, so we can't always detect their presence.
+	 * Ideally we should be able to check the state of the
+	 * CONFIG1 pin, but no such luck on our hardware.
+	 *
+	 * The only method left to us is to check the VBT to see
+	 * if the port is a dual mode capable DP port. But let's
+	 * only do that when we sucesfully read the EDID, to avoid
+	 * confusing log messages about DP dual mode adaptors when
+	 * there's nothing connected to the port.
+	 */
+	if (type == DRM_DP_DUAL_MODE_UNKNOWN) {
+		/* An overridden EDID imply that we want this port for testing.
+		 * Make sure not to set limits for that port.
+		 */
+		if (has_edid && !connector->override_edid &&
+		    intel_bios_is_port_dp_dual_mode(dev_priv, port)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Assuming DP dual mode adaptor presence based on VBT\n");
+			type = DRM_DP_DUAL_MODE_TYPE1_DVI;
+		} else {
+			type = DRM_DP_DUAL_MODE_NONE;
+		}
+	}
+
+	if (type == DRM_DP_DUAL_MODE_NONE)
+		return;
+
+	hdmi->dp_dual_mode.type = type;
+	hdmi->dp_dual_mode.max_tmds_clock =
+		drm_dp_dual_mode_max_tmds_clock(&dev_priv->drm, type, adapter);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "DP dual mode adaptor (%s) detected (max TMDS clock: %d kHz)\n",
+		    drm_dp_get_dual_mode_type_name(type),
+		    hdmi->dp_dual_mode.max_tmds_clock);
+
+	/* Older VBTs are often buggy and can't be trusted :( Play it safe. */
+	if ((DISPLAY_VER(dev_priv) >= 8 || IS_HASWELL(dev_priv)) &&
+	    !intel_bios_is_port_dp_dual_mode(dev_priv, port)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Ignoring DP dual mode adaptor max TMDS clock for native HDMI port\n");
+		hdmi->dp_dual_mode.max_tmds_clock = 0;
+	}
+}
+
+static bool
+intel_hdmi_set_edid(struct drm_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(to_intel_connector(connector));
+	intel_wakeref_t wakeref;
+	struct edid *edid;
+	bool connected = false;
+	struct i2c_adapter *i2c;
+
+	wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS);
+
+	i2c = intel_gmbus_get_adapter(dev_priv, intel_hdmi->ddc_bus);
+
+	edid = drm_get_edid(connector, i2c);
+
+	if (!edid && !intel_gmbus_is_forced_bit(i2c)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "HDMI GMBUS EDID read failed, retry using GPIO bit-banging\n");
+		intel_gmbus_force_bit(i2c, true);
+		edid = drm_get_edid(connector, i2c);
+		intel_gmbus_force_bit(i2c, false);
+	}
+
+	intel_hdmi_dp_dual_mode_detect(connector, edid != NULL);
+
+	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS, wakeref);
+
+	to_intel_connector(connector)->detect_edid = edid;
+	if (edid && edid->input & DRM_EDID_INPUT_DIGITAL) {
+		intel_hdmi->has_audio = drm_detect_monitor_audio(edid);
+		intel_hdmi->has_hdmi_sink = drm_detect_hdmi_monitor(edid);
+
+		connected = true;
+	}
+
+	cec_notifier_set_phys_addr_from_edid(intel_hdmi->cec_notifier, edid);
+
+	return connected;
+}
+
+static enum drm_connector_status
+intel_hdmi_detect(struct drm_connector *connector, bool force)
+{
+	enum drm_connector_status status = connector_status_disconnected;
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(to_intel_connector(connector));
+	struct intel_encoder *encoder = &hdmi_to_dig_port(intel_hdmi)->base;
+	intel_wakeref_t wakeref;
+
+	drm_dbg_kms(&dev_priv->drm, "[CONNECTOR:%d:%s]\n",
+		    connector->base.id, connector->name);
+
+	if (!INTEL_DISPLAY_ENABLED(dev_priv))
+		return connector_status_disconnected;
+
+	wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS);
+
+	if (DISPLAY_VER(dev_priv) >= 11 &&
+	    !intel_digital_port_connected(encoder))
+		goto out;
+
+	intel_hdmi_unset_edid(connector);
+
+	if (intel_hdmi_set_edid(connector))
+		status = connector_status_connected;
+
+out:
+	intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS, wakeref);
+
+	if (status != connector_status_connected)
+		cec_notifier_phys_addr_invalidate(intel_hdmi->cec_notifier);
+
+	/*
+	 * Make sure the refs for power wells enabled during detect are
+	 * dropped to avoid a new detect cycle triggered by HPD polling.
+	 */
+	intel_display_power_flush_work(dev_priv);
+
+	return status;
+}
+
+static void
+intel_hdmi_force(struct drm_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+
+	drm_dbg_kms(&i915->drm, "[CONNECTOR:%d:%s]\n",
+		    connector->base.id, connector->name);
+
+	intel_hdmi_unset_edid(connector);
+
+	if (connector->status != connector_status_connected)
+		return;
+
+	intel_hdmi_set_edid(connector);
+}
+
+static int intel_hdmi_get_modes(struct drm_connector *connector)
+{
+	struct edid *edid;
+
+	edid = to_intel_connector(connector)->detect_edid;
+	if (edid == NULL)
+		return 0;
+
+	return intel_connector_update_modes(connector, edid);
+}
+
+static struct i2c_adapter *
+intel_hdmi_get_i2c_adapter(struct drm_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	struct intel_hdmi *intel_hdmi = intel_attached_hdmi(to_intel_connector(connector));
+
+	return intel_gmbus_get_adapter(dev_priv, intel_hdmi->ddc_bus);
+}
+
+static void intel_hdmi_create_i2c_symlink(struct drm_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct i2c_adapter *adapter = intel_hdmi_get_i2c_adapter(connector);
+	struct kobject *i2c_kobj = &adapter->dev.kobj;
+	struct kobject *connector_kobj = &connector->kdev->kobj;
+	int ret;
+
+	ret = sysfs_create_link(connector_kobj, i2c_kobj, i2c_kobj->name);
+	if (ret)
+		drm_err(&i915->drm, "Failed to create i2c symlink (%d)\n", ret);
+}
+
+static void intel_hdmi_remove_i2c_symlink(struct drm_connector *connector)
+{
+	struct i2c_adapter *adapter = intel_hdmi_get_i2c_adapter(connector);
+	struct kobject *i2c_kobj = &adapter->dev.kobj;
+	struct kobject *connector_kobj = &connector->kdev->kobj;
+
+	sysfs_remove_link(connector_kobj, i2c_kobj->name);
+}
+
+static int
+intel_hdmi_connector_register(struct drm_connector *connector)
+{
+	int ret;
+
+	ret = intel_connector_register(connector);
+	if (ret)
+		return ret;
+
+	intel_hdmi_create_i2c_symlink(connector);
+
+	return ret;
+}
+
+static void intel_hdmi_connector_unregister(struct drm_connector *connector)
+{
+	struct cec_notifier *n = intel_attached_hdmi(to_intel_connector(connector))->cec_notifier;
+
+	cec_notifier_conn_unregister(n);
+
+	intel_hdmi_remove_i2c_symlink(connector);
+	intel_connector_unregister(connector);
+}
+
+static const struct drm_connector_funcs intel_hdmi_connector_funcs = {
+	.detect = intel_hdmi_detect,
+	.force = intel_hdmi_force,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.atomic_get_property = intel_digital_connector_atomic_get_property,
+	.atomic_set_property = intel_digital_connector_atomic_set_property,
+	.late_register = intel_hdmi_connector_register,
+	.early_unregister = intel_hdmi_connector_unregister,
+	.destroy = intel_connector_destroy,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.atomic_duplicate_state = intel_digital_connector_duplicate_state,
+};
+
+static const struct drm_connector_helper_funcs intel_hdmi_connector_helper_funcs = {
+	.get_modes = intel_hdmi_get_modes,
+	.mode_valid = intel_hdmi_mode_valid,
+	.atomic_check = intel_digital_connector_atomic_check,
+};
+
+static void
+intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+
+	intel_attach_force_audio_property(connector);
+	intel_attach_broadcast_rgb_property(connector);
+	intel_attach_aspect_ratio_property(connector);
+
+	intel_attach_hdmi_colorspace_property(connector);
+	drm_connector_attach_content_type_property(connector);
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		drm_connector_attach_hdr_output_metadata_property(connector);
+
+	if (!HAS_GMCH(dev_priv))
+		drm_connector_attach_max_bpc_property(connector, 8, 12);
+}
+
+/*
+ * intel_hdmi_handle_sink_scrambling: handle sink scrambling/clock ratio setup
+ * @encoder: intel_encoder
+ * @connector: drm_connector
+ * @high_tmds_clock_ratio = bool to indicate if the function needs to set
+ *  or reset the high tmds clock ratio for scrambling
+ * @scrambling: bool to Indicate if the function needs to set or reset
+ *  sink scrambling
+ *
+ * This function handles scrambling on HDMI 2.0 capable sinks.
+ * If required clock rate is > 340 Mhz && scrambling is supported by sink
+ * it enables scrambling. This should be called before enabling the HDMI
+ * 2.0 port, as the sink can choose to disable the scrambling if it doesn't
+ * detect a scrambled clock within 100 ms.
+ *
+ * Returns:
+ * True on success, false on failure.
+ */
+bool intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder,
+				       struct drm_connector *connector,
+				       bool high_tmds_clock_ratio,
+				       bool scrambling)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
+	struct drm_scrambling *sink_scrambling =
+		&connector->display_info.hdmi.scdc.scrambling;
+	struct i2c_adapter *adapter =
+		intel_gmbus_get_adapter(dev_priv, intel_hdmi->ddc_bus);
+
+	if (!sink_scrambling->supported)
+		return true;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "[CONNECTOR:%d:%s] scrambling=%s, TMDS bit clock ratio=1/%d\n",
+		    connector->base.id, connector->name,
+		    str_yes_no(scrambling), high_tmds_clock_ratio ? 40 : 10);
+
+	/* Set TMDS bit clock ratio to 1/40 or 1/10, and enable/disable scrambling */
+	return drm_scdc_set_high_tmds_clock_ratio(adapter,
+						  high_tmds_clock_ratio) &&
+		drm_scdc_set_scrambling(adapter, scrambling);
+}
+
+static u8 chv_port_to_ddc_pin(struct drm_i915_private *dev_priv, enum port port)
+{
+	u8 ddc_pin;
+
+	switch (port) {
+	case PORT_B:
+		ddc_pin = GMBUS_PIN_DPB;
+		break;
+	case PORT_C:
+		ddc_pin = GMBUS_PIN_DPC;
+		break;
+	case PORT_D:
+		ddc_pin = GMBUS_PIN_DPD_CHV;
+		break;
+	default:
+		MISSING_CASE(port);
+		ddc_pin = GMBUS_PIN_DPB;
+		break;
+	}
+	return ddc_pin;
+}
+
+static u8 bxt_port_to_ddc_pin(struct drm_i915_private *dev_priv, enum port port)
+{
+	u8 ddc_pin;
+
+	switch (port) {
+	case PORT_B:
+		ddc_pin = GMBUS_PIN_1_BXT;
+		break;
+	case PORT_C:
+		ddc_pin = GMBUS_PIN_2_BXT;
+		break;
+	default:
+		MISSING_CASE(port);
+		ddc_pin = GMBUS_PIN_1_BXT;
+		break;
+	}
+	return ddc_pin;
+}
+
+static u8 cnp_port_to_ddc_pin(struct drm_i915_private *dev_priv,
+			      enum port port)
+{
+	u8 ddc_pin;
+
+	switch (port) {
+	case PORT_B:
+		ddc_pin = GMBUS_PIN_1_BXT;
+		break;
+	case PORT_C:
+		ddc_pin = GMBUS_PIN_2_BXT;
+		break;
+	case PORT_D:
+		ddc_pin = GMBUS_PIN_4_CNP;
+		break;
+	case PORT_F:
+		ddc_pin = GMBUS_PIN_3_BXT;
+		break;
+	default:
+		MISSING_CASE(port);
+		ddc_pin = GMBUS_PIN_1_BXT;
+		break;
+	}
+	return ddc_pin;
+}
+
+static u8 icl_port_to_ddc_pin(struct drm_i915_private *dev_priv, enum port port)
+{
+	enum phy phy = intel_port_to_phy(dev_priv, port);
+
+	if (intel_phy_is_combo(dev_priv, phy))
+		return GMBUS_PIN_1_BXT + port;
+	else if (intel_phy_is_tc(dev_priv, phy))
+		return GMBUS_PIN_9_TC1_ICP + intel_port_to_tc(dev_priv, port);
+
+	drm_WARN(&dev_priv->drm, 1, "Unknown port:%c\n", port_name(port));
+	return GMBUS_PIN_2_BXT;
+}
+
+static u8 mcc_port_to_ddc_pin(struct drm_i915_private *dev_priv, enum port port)
+{
+	enum phy phy = intel_port_to_phy(dev_priv, port);
+	u8 ddc_pin;
+
+	switch (phy) {
+	case PHY_A:
+		ddc_pin = GMBUS_PIN_1_BXT;
+		break;
+	case PHY_B:
+		ddc_pin = GMBUS_PIN_2_BXT;
+		break;
+	case PHY_C:
+		ddc_pin = GMBUS_PIN_9_TC1_ICP;
+		break;
+	default:
+		MISSING_CASE(phy);
+		ddc_pin = GMBUS_PIN_1_BXT;
+		break;
+	}
+	return ddc_pin;
+}
+
+static u8 rkl_port_to_ddc_pin(struct drm_i915_private *dev_priv, enum port port)
+{
+	enum phy phy = intel_port_to_phy(dev_priv, port);
+
+	WARN_ON(port == PORT_C);
+
+	/*
+	 * Pin mapping for RKL depends on which PCH is present.  With TGP, the
+	 * final two outputs use type-c pins, even though they're actually
+	 * combo outputs.  With CMP, the traditional DDI A-D pins are used for
+	 * all outputs.
+	 */
+	if (INTEL_PCH_TYPE(dev_priv) >= PCH_TGP && phy >= PHY_C)
+		return GMBUS_PIN_9_TC1_ICP + phy - PHY_C;
+
+	return GMBUS_PIN_1_BXT + phy;
+}
+
+static u8 gen9bc_tgp_port_to_ddc_pin(struct drm_i915_private *i915, enum port port)
+{
+	enum phy phy = intel_port_to_phy(i915, port);
+
+	drm_WARN_ON(&i915->drm, port == PORT_A);
+
+	/*
+	 * Pin mapping for GEN9 BC depends on which PCH is present.  With TGP,
+	 * final two outputs use type-c pins, even though they're actually
+	 * combo outputs.  With CMP, the traditional DDI A-D pins are used for
+	 * all outputs.
+	 */
+	if (INTEL_PCH_TYPE(i915) >= PCH_TGP && phy >= PHY_C)
+		return GMBUS_PIN_9_TC1_ICP + phy - PHY_C;
+
+	return GMBUS_PIN_1_BXT + phy;
+}
+
+static u8 dg1_port_to_ddc_pin(struct drm_i915_private *dev_priv, enum port port)
+{
+	return intel_port_to_phy(dev_priv, port) + 1;
+}
+
+static u8 adls_port_to_ddc_pin(struct drm_i915_private *dev_priv, enum port port)
+{
+	enum phy phy = intel_port_to_phy(dev_priv, port);
+
+	WARN_ON(port == PORT_B || port == PORT_C);
+
+	/*
+	 * Pin mapping for ADL-S requires TC pins for all combo phy outputs
+	 * except first combo output.
+	 */
+	if (phy == PHY_A)
+		return GMBUS_PIN_1_BXT;
+
+	return GMBUS_PIN_9_TC1_ICP + phy - PHY_B;
+}
+
+static u8 g4x_port_to_ddc_pin(struct drm_i915_private *dev_priv,
+			      enum port port)
+{
+	u8 ddc_pin;
+
+	switch (port) {
+	case PORT_B:
+		ddc_pin = GMBUS_PIN_DPB;
+		break;
+	case PORT_C:
+		ddc_pin = GMBUS_PIN_DPC;
+		break;
+	case PORT_D:
+		ddc_pin = GMBUS_PIN_DPD;
+		break;
+	default:
+		MISSING_CASE(port);
+		ddc_pin = GMBUS_PIN_DPB;
+		break;
+	}
+	return ddc_pin;
+}
+
+static u8 intel_hdmi_ddc_pin(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum port port = encoder->port;
+	u8 ddc_pin;
+
+	ddc_pin = intel_bios_alternate_ddc_pin(encoder);
+	if (ddc_pin) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Using DDC pin 0x%x for port %c (VBT)\n",
+			    ddc_pin, port_name(port));
+		return ddc_pin;
+	}
+
+	if (IS_ALDERLAKE_S(dev_priv))
+		ddc_pin = adls_port_to_ddc_pin(dev_priv, port);
+	else if (INTEL_PCH_TYPE(dev_priv) >= PCH_DG1)
+		ddc_pin = dg1_port_to_ddc_pin(dev_priv, port);
+	else if (IS_ROCKETLAKE(dev_priv))
+		ddc_pin = rkl_port_to_ddc_pin(dev_priv, port);
+	else if (DISPLAY_VER(dev_priv) == 9 && HAS_PCH_TGP(dev_priv))
+		ddc_pin = gen9bc_tgp_port_to_ddc_pin(dev_priv, port);
+	else if (IS_JSL_EHL(dev_priv) && HAS_PCH_TGP(dev_priv))
+		ddc_pin = mcc_port_to_ddc_pin(dev_priv, port);
+	else if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP)
+		ddc_pin = icl_port_to_ddc_pin(dev_priv, port);
+	else if (HAS_PCH_CNP(dev_priv))
+		ddc_pin = cnp_port_to_ddc_pin(dev_priv, port);
+	else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		ddc_pin = bxt_port_to_ddc_pin(dev_priv, port);
+	else if (IS_CHERRYVIEW(dev_priv))
+		ddc_pin = chv_port_to_ddc_pin(dev_priv, port);
+	else
+		ddc_pin = g4x_port_to_ddc_pin(dev_priv, port);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Using DDC pin 0x%x for port %c (platform default)\n",
+		    ddc_pin, port_name(port));
+
+	return ddc_pin;
+}
+
+void intel_infoframe_init(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(dig_port->base.base.dev);
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		dig_port->write_infoframe = vlv_write_infoframe;
+		dig_port->read_infoframe = vlv_read_infoframe;
+		dig_port->set_infoframes = vlv_set_infoframes;
+		dig_port->infoframes_enabled = vlv_infoframes_enabled;
+	} else if (IS_G4X(dev_priv)) {
+		dig_port->write_infoframe = g4x_write_infoframe;
+		dig_port->read_infoframe = g4x_read_infoframe;
+		dig_port->set_infoframes = g4x_set_infoframes;
+		dig_port->infoframes_enabled = g4x_infoframes_enabled;
+	} else if (HAS_DDI(dev_priv)) {
+		if (intel_bios_is_lspcon_present(dev_priv, dig_port->base.port)) {
+			dig_port->write_infoframe = lspcon_write_infoframe;
+			dig_port->read_infoframe = lspcon_read_infoframe;
+			dig_port->set_infoframes = lspcon_set_infoframes;
+			dig_port->infoframes_enabled = lspcon_infoframes_enabled;
+		} else {
+			dig_port->write_infoframe = hsw_write_infoframe;
+			dig_port->read_infoframe = hsw_read_infoframe;
+			dig_port->set_infoframes = hsw_set_infoframes;
+			dig_port->infoframes_enabled = hsw_infoframes_enabled;
+		}
+	} else if (HAS_PCH_IBX(dev_priv)) {
+		dig_port->write_infoframe = ibx_write_infoframe;
+		dig_port->read_infoframe = ibx_read_infoframe;
+		dig_port->set_infoframes = ibx_set_infoframes;
+		dig_port->infoframes_enabled = ibx_infoframes_enabled;
+	} else {
+		dig_port->write_infoframe = cpt_write_infoframe;
+		dig_port->read_infoframe = cpt_read_infoframe;
+		dig_port->set_infoframes = cpt_set_infoframes;
+		dig_port->infoframes_enabled = cpt_infoframes_enabled;
+	}
+}
+
+void intel_hdmi_init_connector(struct intel_digital_port *dig_port,
+			       struct intel_connector *intel_connector)
+{
+	struct drm_connector *connector = &intel_connector->base;
+	struct intel_hdmi *intel_hdmi = &dig_port->hdmi;
+	struct intel_encoder *intel_encoder = &dig_port->base;
+	struct drm_device *dev = intel_encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct i2c_adapter *ddc;
+	enum port port = intel_encoder->port;
+	struct cec_connector_info conn_info;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Adding HDMI connector on [ENCODER:%d:%s]\n",
+		    intel_encoder->base.base.id, intel_encoder->base.name);
+
+	if (DISPLAY_VER(dev_priv) < 12 && drm_WARN_ON(dev, port == PORT_A))
+		return;
+
+	if (drm_WARN(dev, dig_port->max_lanes < 4,
+		     "Not enough lanes (%d) for HDMI on [ENCODER:%d:%s]\n",
+		     dig_port->max_lanes, intel_encoder->base.base.id,
+		     intel_encoder->base.name))
+		return;
+
+	intel_hdmi->ddc_bus = intel_hdmi_ddc_pin(intel_encoder);
+	ddc = intel_gmbus_get_adapter(dev_priv, intel_hdmi->ddc_bus);
+
+	drm_connector_init_with_ddc(dev, connector,
+				    &intel_hdmi_connector_funcs,
+				    DRM_MODE_CONNECTOR_HDMIA,
+				    ddc);
+	drm_connector_helper_add(connector, &intel_hdmi_connector_helper_funcs);
+
+	connector->interlace_allowed = 1;
+	connector->doublescan_allowed = 0;
+	connector->stereo_allowed = 1;
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		connector->ycbcr_420_allowed = true;
+
+	intel_connector->polled = DRM_CONNECTOR_POLL_HPD;
+
+	if (HAS_DDI(dev_priv))
+		intel_connector->get_hw_state = intel_ddi_connector_get_hw_state;
+	else
+		intel_connector->get_hw_state = intel_connector_get_hw_state;
+
+	intel_hdmi_add_properties(intel_hdmi, connector);
+
+	intel_connector_attach_encoder(intel_connector, intel_encoder);
+	intel_hdmi->attached_connector = intel_connector;
+
+	if (is_hdcp_supported(dev_priv, port)) {
+		int ret = intel_hdcp_init(intel_connector, dig_port,
+					  &intel_hdmi_hdcp_shim);
+		if (ret)
+			drm_dbg_kms(&dev_priv->drm,
+				    "HDCP init failed, skipping.\n");
+	}
+
+	/* For G4X desktop chip, PEG_BAND_GAP_DATA 3:0 must first be written
+	 * 0xd.  Failure to do so will result in spurious interrupts being
+	 * generated on the port when a cable is not attached.
+	 */
+	if (IS_G45(dev_priv)) {
+		u32 temp = intel_de_read(dev_priv, PEG_BAND_GAP_DATA);
+		intel_de_write(dev_priv, PEG_BAND_GAP_DATA,
+		               (temp & ~0xf) | 0xd);
+	}
+
+	cec_fill_conn_info_from_drm(&conn_info, connector);
+
+	intel_hdmi->cec_notifier =
+		cec_notifier_conn_register(dev->dev, port_identifier(port),
+					   &conn_info);
+	if (!intel_hdmi->cec_notifier)
+		drm_dbg_kms(&dev_priv->drm, "CEC notifier get failed\n");
+}
+
+/*
+ * intel_hdmi_dsc_get_slice_height - get the dsc slice_height
+ * @vactive: Vactive of a display mode
+ *
+ * @return: appropriate dsc slice height for a given mode.
+ */
+int intel_hdmi_dsc_get_slice_height(int vactive)
+{
+	int slice_height;
+
+	/*
+	 * Slice Height determination : HDMI2.1 Section 7.7.5.2
+	 * Select smallest slice height >=96, that results in a valid PPS and
+	 * requires minimum padding lines required for final slice.
+	 *
+	 * Assumption : Vactive is even.
+	 */
+	for (slice_height = 96; slice_height <= vactive; slice_height += 2)
+		if (vactive % slice_height == 0)
+			return slice_height;
+
+	return 0;
+}
+
+/*
+ * intel_hdmi_dsc_get_num_slices - get no. of dsc slices based on dsc encoder
+ * and dsc decoder capabilities
+ *
+ * @crtc_state: intel crtc_state
+ * @src_max_slices: maximum slices supported by the DSC encoder
+ * @src_max_slice_width: maximum slice width supported by DSC encoder
+ * @hdmi_max_slices: maximum slices supported by sink DSC decoder
+ * @hdmi_throughput: maximum clock per slice (MHz) supported by HDMI sink
+ *
+ * @return: num of dsc slices that can be supported by the dsc encoder
+ * and decoder.
+ */
+int
+intel_hdmi_dsc_get_num_slices(const struct intel_crtc_state *crtc_state,
+			      int src_max_slices, int src_max_slice_width,
+			      int hdmi_max_slices, int hdmi_throughput)
+{
+/* Pixel rates in KPixels/sec */
+#define HDMI_DSC_PEAK_PIXEL_RATE		2720000
+/*
+ * Rates at which the source and sink are required to process pixels in each
+ * slice, can be two levels: either atleast 340000KHz or atleast 40000KHz.
+ */
+#define HDMI_DSC_MAX_ENC_THROUGHPUT_0		340000
+#define HDMI_DSC_MAX_ENC_THROUGHPUT_1		400000
+
+/* Spec limits the slice width to 2720 pixels */
+#define MAX_HDMI_SLICE_WIDTH			2720
+	int kslice_adjust;
+	int adjusted_clk_khz;
+	int min_slices;
+	int target_slices;
+	int max_throughput; /* max clock freq. in khz per slice */
+	int max_slice_width;
+	int slice_width;
+	int pixel_clock = crtc_state->hw.adjusted_mode.crtc_clock;
+
+	if (!hdmi_throughput)
+		return 0;
+
+	/*
+	 * Slice Width determination : HDMI2.1 Section 7.7.5.1
+	 * kslice_adjust factor for 4:2:0, and 4:2:2 formats is 0.5, where as
+	 * for 4:4:4 is 1.0. Multiplying these factors by 10 and later
+	 * dividing adjusted clock value by 10.
+	 */
+	if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR444 ||
+	    crtc_state->output_format == INTEL_OUTPUT_FORMAT_RGB)
+		kslice_adjust = 10;
+	else
+		kslice_adjust = 5;
+
+	/*
+	 * As per spec, the rate at which the source and the sink process
+	 * the pixels per slice are at two levels: atleast 340Mhz or 400Mhz.
+	 * This depends upon the pixel clock rate and output formats
+	 * (kslice adjust).
+	 * If pixel clock * kslice adjust >= 2720MHz slices can be processed
+	 * at max 340MHz, otherwise they can be processed at max 400MHz.
+	 */
+
+	adjusted_clk_khz = DIV_ROUND_UP(kslice_adjust * pixel_clock, 10);
+
+	if (adjusted_clk_khz <= HDMI_DSC_PEAK_PIXEL_RATE)
+		max_throughput = HDMI_DSC_MAX_ENC_THROUGHPUT_0;
+	else
+		max_throughput = HDMI_DSC_MAX_ENC_THROUGHPUT_1;
+
+	/*
+	 * Taking into account the sink's capability for maximum
+	 * clock per slice (in MHz) as read from HF-VSDB.
+	 */
+	max_throughput = min(max_throughput, hdmi_throughput * 1000);
+
+	min_slices = DIV_ROUND_UP(adjusted_clk_khz, max_throughput);
+	max_slice_width = min(MAX_HDMI_SLICE_WIDTH, src_max_slice_width);
+
+	/*
+	 * Keep on increasing the num of slices/line, starting from min_slices
+	 * per line till we get such a number, for which the slice_width is
+	 * just less than max_slice_width. The slices/line selected should be
+	 * less than or equal to the max horizontal slices that the combination
+	 * of PCON encoder and HDMI decoder can support.
+	 */
+	slice_width = max_slice_width;
+
+	do {
+		if (min_slices <= 1 && src_max_slices >= 1 && hdmi_max_slices >= 1)
+			target_slices = 1;
+		else if (min_slices <= 2 && src_max_slices >= 2 && hdmi_max_slices >= 2)
+			target_slices = 2;
+		else if (min_slices <= 4 && src_max_slices >= 4 && hdmi_max_slices >= 4)
+			target_slices = 4;
+		else if (min_slices <= 8 && src_max_slices >= 8 && hdmi_max_slices >= 8)
+			target_slices = 8;
+		else if (min_slices <= 12 && src_max_slices >= 12 && hdmi_max_slices >= 12)
+			target_slices = 12;
+		else if (min_slices <= 16 && src_max_slices >= 16 && hdmi_max_slices >= 16)
+			target_slices = 16;
+		else
+			return 0;
+
+		slice_width = DIV_ROUND_UP(crtc_state->hw.adjusted_mode.hdisplay, target_slices);
+		if (slice_width >= max_slice_width)
+			min_slices = target_slices + 1;
+	} while (slice_width >= max_slice_width);
+
+	return target_slices;
+}
+
+/*
+ * intel_hdmi_dsc_get_bpp - get the appropriate compressed bits_per_pixel based on
+ * source and sink capabilities.
+ *
+ * @src_fraction_bpp: fractional bpp supported by the source
+ * @slice_width: dsc slice width supported by the source and sink
+ * @num_slices: num of slices supported by the source and sink
+ * @output_format: video output format
+ * @hdmi_all_bpp: sink supports decoding of 1/16th bpp setting
+ * @hdmi_max_chunk_bytes: max bytes in a line of chunks supported by sink
+ *
+ * @return: compressed bits_per_pixel in step of 1/16 of bits_per_pixel
+ */
+int
+intel_hdmi_dsc_get_bpp(int src_fractional_bpp, int slice_width, int num_slices,
+		       int output_format, bool hdmi_all_bpp,
+		       int hdmi_max_chunk_bytes)
+{
+	int max_dsc_bpp, min_dsc_bpp;
+	int target_bytes;
+	bool bpp_found = false;
+	int bpp_decrement_x16;
+	int bpp_target;
+	int bpp_target_x16;
+
+	/*
+	 * Get min bpp and max bpp as per Table 7.23, in HDMI2.1 spec
+	 * Start with the max bpp and keep on decrementing with
+	 * fractional bpp, if supported by PCON DSC encoder
+	 *
+	 * for each bpp we check if no of bytes can be supported by HDMI sink
+	 */
+
+	/* Assuming: bpc as 8*/
+	if (output_format == INTEL_OUTPUT_FORMAT_YCBCR420) {
+		min_dsc_bpp = 6;
+		max_dsc_bpp = 3 * 4; /* 3*bpc/2 */
+	} else if (output_format == INTEL_OUTPUT_FORMAT_YCBCR444 ||
+		   output_format == INTEL_OUTPUT_FORMAT_RGB) {
+		min_dsc_bpp = 8;
+		max_dsc_bpp = 3 * 8; /* 3*bpc */
+	} else {
+		/* Assuming 4:2:2 encoding */
+		min_dsc_bpp = 7;
+		max_dsc_bpp = 2 * 8; /* 2*bpc */
+	}
+
+	/*
+	 * Taking into account if all dsc_all_bpp supported by HDMI2.1 sink
+	 * Section 7.7.34 : Source shall not enable compressed Video
+	 * Transport with bpp_target settings above 12 bpp unless
+	 * DSC_all_bpp is set to 1.
+	 */
+	if (!hdmi_all_bpp)
+		max_dsc_bpp = min(max_dsc_bpp, 12);
+
+	/*
+	 * The Sink has a limit of compressed data in bytes for a scanline,
+	 * as described in max_chunk_bytes field in HFVSDB block of edid.
+	 * The no. of bytes depend on the target bits per pixel that the
+	 * source configures. So we start with the max_bpp and calculate
+	 * the target_chunk_bytes. We keep on decrementing the target_bpp,
+	 * till we get the target_chunk_bytes just less than what the sink's
+	 * max_chunk_bytes, or else till we reach the min_dsc_bpp.
+	 *
+	 * The decrement is according to the fractional support from PCON DSC
+	 * encoder. For fractional BPP we use bpp_target as a multiple of 16.
+	 *
+	 * bpp_target_x16 = bpp_target * 16
+	 * So we need to decrement by {1, 2, 4, 8, 16} for fractional bpps
+	 * {1/16, 1/8, 1/4, 1/2, 1} respectively.
+	 */
+
+	bpp_target = max_dsc_bpp;
+
+	/* src does not support fractional bpp implies decrement by 16 for bppx16 */
+	if (!src_fractional_bpp)
+		src_fractional_bpp = 1;
+	bpp_decrement_x16 = DIV_ROUND_UP(16, src_fractional_bpp);
+	bpp_target_x16 = (bpp_target * 16) - bpp_decrement_x16;
+
+	while (bpp_target_x16 > (min_dsc_bpp * 16)) {
+		int bpp;
+
+		bpp = DIV_ROUND_UP(bpp_target_x16, 16);
+		target_bytes = DIV_ROUND_UP((num_slices * slice_width * bpp), 8);
+		if (target_bytes <= hdmi_max_chunk_bytes) {
+			bpp_found = true;
+			break;
+		}
+		bpp_target_x16 -= bpp_decrement_x16;
+	}
+	if (bpp_found)
+		return bpp_target_x16;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_hdmi.h b/drivers/gpu/drm/i915/display/intel_hdmi.h
new file mode 100644
index 000000000..93f65a917
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_hdmi.h
@@ -0,0 +1,58 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_HDMI_H__
+#define __INTEL_HDMI_H__
+
+#include <linux/hdmi.h>
+#include <linux/types.h>
+
+struct drm_connector;
+struct drm_encoder;
+struct drm_i915_private;
+struct intel_connector;
+struct intel_digital_port;
+struct intel_encoder;
+struct intel_crtc_state;
+struct intel_hdmi;
+struct drm_connector_state;
+union hdmi_infoframe;
+enum port;
+
+void intel_hdmi_init_connector(struct intel_digital_port *dig_port,
+			       struct intel_connector *intel_connector);
+int intel_hdmi_compute_config(struct intel_encoder *encoder,
+			      struct intel_crtc_state *pipe_config,
+			      struct drm_connector_state *conn_state);
+void intel_hdmi_encoder_shutdown(struct intel_encoder *encoder);
+bool intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder,
+				       struct drm_connector *connector,
+				       bool high_tmds_clock_ratio,
+				       bool scrambling);
+void intel_dp_dual_mode_set_tmds_output(struct intel_hdmi *hdmi, bool enable);
+void intel_infoframe_init(struct intel_digital_port *dig_port);
+u32 intel_hdmi_infoframes_enabled(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state);
+u32 intel_hdmi_infoframe_enable(unsigned int type);
+void intel_hdmi_read_gcp_infoframe(struct intel_encoder *encoder,
+				   struct intel_crtc_state *crtc_state);
+void intel_read_infoframe(struct intel_encoder *encoder,
+			  const struct intel_crtc_state *crtc_state,
+			  enum hdmi_infoframe_type type,
+			  union hdmi_infoframe *frame);
+bool intel_hdmi_limited_color_range(const struct intel_crtc_state *crtc_state,
+				    const struct drm_connector_state *conn_state);
+bool intel_hdmi_bpc_possible(const struct intel_crtc_state *crtc_state,
+			     int bpc, bool has_hdmi_sink, bool ycbcr420_output);
+int intel_hdmi_tmds_clock(int clock, int bpc, bool ycbcr420_output);
+int intel_hdmi_dsc_get_bpp(int src_fractional_bpp, int slice_width,
+			   int num_slices, int output_format, bool hdmi_all_bpp,
+			   int hdmi_max_chunk_bytes);
+int intel_hdmi_dsc_get_num_slices(const struct intel_crtc_state *crtc_state,
+				  int src_max_slices, int src_max_slice_width,
+				  int hdmi_max_slices, int hdmi_throughput);
+int intel_hdmi_dsc_get_slice_height(int vactive);
+
+#endif /* __INTEL_HDMI_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_hotplug.c b/drivers/gpu/drm/i915/display/intel_hotplug.c
new file mode 100644
index 000000000..f7a2f485b
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_hotplug.c
@@ -0,0 +1,769 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/kernel.h>
+
+#include "i915_drv.h"
+#include "i915_irq.h"
+#include "intel_display_types.h"
+#include "intel_hotplug.h"
+
+/**
+ * DOC: Hotplug
+ *
+ * Simply put, hotplug occurs when a display is connected to or disconnected
+ * from the system. However, there may be adapters and docking stations and
+ * Display Port short pulses and MST devices involved, complicating matters.
+ *
+ * Hotplug in i915 is handled in many different levels of abstraction.
+ *
+ * The platform dependent interrupt handling code in i915_irq.c enables,
+ * disables, and does preliminary handling of the interrupts. The interrupt
+ * handlers gather the hotplug detect (HPD) information from relevant registers
+ * into a platform independent mask of hotplug pins that have fired.
+ *
+ * The platform independent interrupt handler intel_hpd_irq_handler() in
+ * intel_hotplug.c does hotplug irq storm detection and mitigation, and passes
+ * further processing to appropriate bottom halves (Display Port specific and
+ * regular hotplug).
+ *
+ * The Display Port work function i915_digport_work_func() calls into
+ * intel_dp_hpd_pulse() via hooks, which handles DP short pulses and DP MST long
+ * pulses, with failures and non-MST long pulses triggering regular hotplug
+ * processing on the connector.
+ *
+ * The regular hotplug work function i915_hotplug_work_func() calls connector
+ * detect hooks, and, if connector status changes, triggers sending of hotplug
+ * uevent to userspace via drm_kms_helper_hotplug_event().
+ *
+ * Finally, the userspace is responsible for triggering a modeset upon receiving
+ * the hotplug uevent, disabling or enabling the crtc as needed.
+ *
+ * The hotplug interrupt storm detection and mitigation code keeps track of the
+ * number of interrupts per hotplug pin per a period of time, and if the number
+ * of interrupts exceeds a certain threshold, the interrupt is disabled for a
+ * while before being re-enabled. The intention is to mitigate issues raising
+ * from broken hardware triggering massive amounts of interrupts and grinding
+ * the system to a halt.
+ *
+ * Current implementation expects that hotplug interrupt storm will not be
+ * seen when display port sink is connected, hence on platforms whose DP
+ * callback is handled by i915_digport_work_func reenabling of hpd is not
+ * performed (it was never expected to be disabled in the first place ;) )
+ * this is specific to DP sinks handled by this routine and any other display
+ * such as HDMI or DVI enabled on the same port will have proper logic since
+ * it will use i915_hotplug_work_func where this logic is handled.
+ */
+
+/**
+ * intel_hpd_pin_default - return default pin associated with certain port.
+ * @dev_priv: private driver data pointer
+ * @port: the hpd port to get associated pin
+ *
+ * It is only valid and used by digital port encoder.
+ *
+ * Return pin that is associatade with @port.
+ */
+enum hpd_pin intel_hpd_pin_default(struct drm_i915_private *dev_priv,
+				   enum port port)
+{
+	return HPD_PORT_A + port - PORT_A;
+}
+
+#define HPD_STORM_DETECT_PERIOD		1000
+#define HPD_STORM_REENABLE_DELAY	(2 * 60 * 1000)
+#define HPD_RETRY_DELAY			1000
+
+static enum hpd_pin
+intel_connector_hpd_pin(struct intel_connector *connector)
+{
+	struct intel_encoder *encoder = intel_attached_encoder(connector);
+
+	/*
+	 * MST connectors get their encoder attached dynamically
+	 * so need to make sure we have an encoder here. But since
+	 * MST encoders have their hpd_pin set to HPD_NONE we don't
+	 * have to special case them beyond that.
+	 */
+	return encoder ? encoder->hpd_pin : HPD_NONE;
+}
+
+/**
+ * intel_hpd_irq_storm_detect - gather stats and detect HPD IRQ storm on a pin
+ * @dev_priv: private driver data pointer
+ * @pin: the pin to gather stats on
+ * @long_hpd: whether the HPD IRQ was long or short
+ *
+ * Gather stats about HPD IRQs from the specified @pin, and detect IRQ
+ * storms. Only the pin specific stats and state are changed, the caller is
+ * responsible for further action.
+ *
+ * The number of IRQs that are allowed within @HPD_STORM_DETECT_PERIOD is
+ * stored in @dev_priv->display.hotplug.hpd_storm_threshold which defaults to
+ * @HPD_STORM_DEFAULT_THRESHOLD. Long IRQs count as +10 to this threshold, and
+ * short IRQs count as +1. If this threshold is exceeded, it's considered an
+ * IRQ storm and the IRQ state is set to @HPD_MARK_DISABLED.
+ *
+ * By default, most systems will only count long IRQs towards
+ * &dev_priv->display.hotplug.hpd_storm_threshold. However, some older systems also
+ * suffer from short IRQ storms and must also track these. Because short IRQ
+ * storms are naturally caused by sideband interactions with DP MST devices,
+ * short IRQ detection is only enabled for systems without DP MST support.
+ * Systems which are new enough to support DP MST are far less likely to
+ * suffer from IRQ storms at all, so this is fine.
+ *
+ * The HPD threshold can be controlled through i915_hpd_storm_ctl in debugfs,
+ * and should only be adjusted for automated hotplug testing.
+ *
+ * Return true if an IRQ storm was detected on @pin.
+ */
+static bool intel_hpd_irq_storm_detect(struct drm_i915_private *dev_priv,
+				       enum hpd_pin pin, bool long_hpd)
+{
+	struct intel_hotplug *hpd = &dev_priv->display.hotplug;
+	unsigned long start = hpd->stats[pin].last_jiffies;
+	unsigned long end = start + msecs_to_jiffies(HPD_STORM_DETECT_PERIOD);
+	const int increment = long_hpd ? 10 : 1;
+	const int threshold = hpd->hpd_storm_threshold;
+	bool storm = false;
+
+	if (!threshold ||
+	    (!long_hpd && !dev_priv->display.hotplug.hpd_short_storm_enabled))
+		return false;
+
+	if (!time_in_range(jiffies, start, end)) {
+		hpd->stats[pin].last_jiffies = jiffies;
+		hpd->stats[pin].count = 0;
+	}
+
+	hpd->stats[pin].count += increment;
+	if (hpd->stats[pin].count > threshold) {
+		hpd->stats[pin].state = HPD_MARK_DISABLED;
+		drm_dbg_kms(&dev_priv->drm,
+			    "HPD interrupt storm detected on PIN %d\n", pin);
+		storm = true;
+	} else {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Received HPD interrupt on PIN %d - cnt: %d\n",
+			      pin,
+			      hpd->stats[pin].count);
+	}
+
+	return storm;
+}
+
+static void
+intel_hpd_irq_storm_switch_to_polling(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_connector_list_iter conn_iter;
+	struct intel_connector *connector;
+	bool hpd_disabled = false;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		enum hpd_pin pin;
+
+		if (connector->base.polled != DRM_CONNECTOR_POLL_HPD)
+			continue;
+
+		pin = intel_connector_hpd_pin(connector);
+		if (pin == HPD_NONE ||
+		    dev_priv->display.hotplug.stats[pin].state != HPD_MARK_DISABLED)
+			continue;
+
+		drm_info(&dev_priv->drm,
+			 "HPD interrupt storm detected on connector %s: "
+			 "switching from hotplug detection to polling\n",
+			 connector->base.name);
+
+		dev_priv->display.hotplug.stats[pin].state = HPD_DISABLED;
+		connector->base.polled = DRM_CONNECTOR_POLL_CONNECT |
+			DRM_CONNECTOR_POLL_DISCONNECT;
+		hpd_disabled = true;
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	/* Enable polling and queue hotplug re-enabling. */
+	if (hpd_disabled) {
+		drm_kms_helper_poll_enable(dev);
+		mod_delayed_work(system_wq, &dev_priv->display.hotplug.reenable_work,
+				 msecs_to_jiffies(HPD_STORM_REENABLE_DELAY));
+	}
+}
+
+static void intel_hpd_irq_storm_reenable_work(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, typeof(*dev_priv),
+			     display.hotplug.reenable_work.work);
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_connector_list_iter conn_iter;
+	struct intel_connector *connector;
+	intel_wakeref_t wakeref;
+	enum hpd_pin pin;
+
+	wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+
+	spin_lock_irq(&dev_priv->irq_lock);
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		pin = intel_connector_hpd_pin(connector);
+		if (pin == HPD_NONE ||
+		    dev_priv->display.hotplug.stats[pin].state != HPD_DISABLED)
+			continue;
+
+		if (connector->base.polled != connector->polled)
+			drm_dbg(&dev_priv->drm,
+				"Reenabling HPD on connector %s\n",
+				connector->base.name);
+		connector->base.polled = connector->polled;
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	for_each_hpd_pin(pin) {
+		if (dev_priv->display.hotplug.stats[pin].state == HPD_DISABLED)
+			dev_priv->display.hotplug.stats[pin].state = HPD_ENABLED;
+	}
+
+	intel_hpd_irq_setup(dev_priv);
+
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+}
+
+enum intel_hotplug_state
+intel_encoder_hotplug(struct intel_encoder *encoder,
+		      struct intel_connector *connector)
+{
+	struct drm_device *dev = connector->base.dev;
+	enum drm_connector_status old_status;
+	u64 old_epoch_counter;
+	bool ret = false;
+
+	drm_WARN_ON(dev, !mutex_is_locked(&dev->mode_config.mutex));
+	old_status = connector->base.status;
+	old_epoch_counter = connector->base.epoch_counter;
+
+	connector->base.status =
+		drm_helper_probe_detect(&connector->base, NULL, false);
+
+	if (old_epoch_counter != connector->base.epoch_counter)
+		ret = true;
+
+	if (ret) {
+		drm_dbg_kms(dev, "[CONNECTOR:%d:%s] status updated from %s to %s (epoch counter %llu->%llu)\n",
+			    connector->base.base.id,
+			    connector->base.name,
+			    drm_get_connector_status_name(old_status),
+			    drm_get_connector_status_name(connector->base.status),
+			    old_epoch_counter,
+			    connector->base.epoch_counter);
+		return INTEL_HOTPLUG_CHANGED;
+	}
+	return INTEL_HOTPLUG_UNCHANGED;
+}
+
+static bool intel_encoder_has_hpd_pulse(struct intel_encoder *encoder)
+{
+	return intel_encoder_is_dig_port(encoder) &&
+		enc_to_dig_port(encoder)->hpd_pulse != NULL;
+}
+
+static void i915_digport_work_func(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, struct drm_i915_private, display.hotplug.dig_port_work);
+	u32 long_port_mask, short_port_mask;
+	struct intel_encoder *encoder;
+	u32 old_bits = 0;
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	long_port_mask = dev_priv->display.hotplug.long_port_mask;
+	dev_priv->display.hotplug.long_port_mask = 0;
+	short_port_mask = dev_priv->display.hotplug.short_port_mask;
+	dev_priv->display.hotplug.short_port_mask = 0;
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		struct intel_digital_port *dig_port;
+		enum port port = encoder->port;
+		bool long_hpd, short_hpd;
+		enum irqreturn ret;
+
+		if (!intel_encoder_has_hpd_pulse(encoder))
+			continue;
+
+		long_hpd = long_port_mask & BIT(port);
+		short_hpd = short_port_mask & BIT(port);
+
+		if (!long_hpd && !short_hpd)
+			continue;
+
+		dig_port = enc_to_dig_port(encoder);
+
+		ret = dig_port->hpd_pulse(dig_port, long_hpd);
+		if (ret == IRQ_NONE) {
+			/* fall back to old school hpd */
+			old_bits |= BIT(encoder->hpd_pin);
+		}
+	}
+
+	if (old_bits) {
+		spin_lock_irq(&dev_priv->irq_lock);
+		dev_priv->display.hotplug.event_bits |= old_bits;
+		spin_unlock_irq(&dev_priv->irq_lock);
+		queue_delayed_work(system_wq, &dev_priv->display.hotplug.hotplug_work, 0);
+	}
+}
+
+/**
+ * intel_hpd_trigger_irq - trigger an hpd irq event for a port
+ * @dig_port: digital port
+ *
+ * Trigger an HPD interrupt event for the given port, emulating a short pulse
+ * generated by the sink, and schedule the dig port work to handle it.
+ */
+void intel_hpd_trigger_irq(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	spin_lock_irq(&i915->irq_lock);
+	i915->display.hotplug.short_port_mask |= BIT(dig_port->base.port);
+	spin_unlock_irq(&i915->irq_lock);
+
+	queue_work(i915->display.hotplug.dp_wq, &i915->display.hotplug.dig_port_work);
+}
+
+/*
+ * Handle hotplug events outside the interrupt handler proper.
+ */
+static void i915_hotplug_work_func(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, struct drm_i915_private,
+			     display.hotplug.hotplug_work.work);
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_connector_list_iter conn_iter;
+	struct intel_connector *connector;
+	u32 changed = 0, retry = 0;
+	u32 hpd_event_bits;
+	u32 hpd_retry_bits;
+
+	mutex_lock(&dev->mode_config.mutex);
+	drm_dbg_kms(&dev_priv->drm, "running encoder hotplug functions\n");
+
+	spin_lock_irq(&dev_priv->irq_lock);
+
+	hpd_event_bits = dev_priv->display.hotplug.event_bits;
+	dev_priv->display.hotplug.event_bits = 0;
+	hpd_retry_bits = dev_priv->display.hotplug.retry_bits;
+	dev_priv->display.hotplug.retry_bits = 0;
+
+	/* Enable polling for connectors which had HPD IRQ storms */
+	intel_hpd_irq_storm_switch_to_polling(dev_priv);
+
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		enum hpd_pin pin;
+		u32 hpd_bit;
+
+		pin = intel_connector_hpd_pin(connector);
+		if (pin == HPD_NONE)
+			continue;
+
+		hpd_bit = BIT(pin);
+		if ((hpd_event_bits | hpd_retry_bits) & hpd_bit) {
+			struct intel_encoder *encoder =
+				intel_attached_encoder(connector);
+
+			if (hpd_event_bits & hpd_bit)
+				connector->hotplug_retries = 0;
+			else
+				connector->hotplug_retries++;
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "Connector %s (pin %i) received hotplug event. (retry %d)\n",
+				    connector->base.name, pin,
+				    connector->hotplug_retries);
+
+			switch (encoder->hotplug(encoder, connector)) {
+			case INTEL_HOTPLUG_UNCHANGED:
+				break;
+			case INTEL_HOTPLUG_CHANGED:
+				changed |= hpd_bit;
+				break;
+			case INTEL_HOTPLUG_RETRY:
+				retry |= hpd_bit;
+				break;
+			}
+		}
+	}
+	drm_connector_list_iter_end(&conn_iter);
+	mutex_unlock(&dev->mode_config.mutex);
+
+	if (changed)
+		drm_kms_helper_hotplug_event(dev);
+
+	/* Remove shared HPD pins that have changed */
+	retry &= ~changed;
+	if (retry) {
+		spin_lock_irq(&dev_priv->irq_lock);
+		dev_priv->display.hotplug.retry_bits |= retry;
+		spin_unlock_irq(&dev_priv->irq_lock);
+
+		mod_delayed_work(system_wq, &dev_priv->display.hotplug.hotplug_work,
+				 msecs_to_jiffies(HPD_RETRY_DELAY));
+	}
+}
+
+
+/**
+ * intel_hpd_irq_handler - main hotplug irq handler
+ * @dev_priv: drm_i915_private
+ * @pin_mask: a mask of hpd pins that have triggered the irq
+ * @long_mask: a mask of hpd pins that may be long hpd pulses
+ *
+ * This is the main hotplug irq handler for all platforms. The platform specific
+ * irq handlers call the platform specific hotplug irq handlers, which read and
+ * decode the appropriate registers into bitmasks about hpd pins that have
+ * triggered (@pin_mask), and which of those pins may be long pulses
+ * (@long_mask). The @long_mask is ignored if the port corresponding to the pin
+ * is not a digital port.
+ *
+ * Here, we do hotplug irq storm detection and mitigation, and pass further
+ * processing to appropriate bottom halves.
+ */
+void intel_hpd_irq_handler(struct drm_i915_private *dev_priv,
+			   u32 pin_mask, u32 long_mask)
+{
+	struct intel_encoder *encoder;
+	bool storm_detected = false;
+	bool queue_dig = false, queue_hp = false;
+	u32 long_hpd_pulse_mask = 0;
+	u32 short_hpd_pulse_mask = 0;
+	enum hpd_pin pin;
+
+	if (!pin_mask)
+		return;
+
+	spin_lock(&dev_priv->irq_lock);
+
+	/*
+	 * Determine whether ->hpd_pulse() exists for each pin, and
+	 * whether we have a short or a long pulse. This is needed
+	 * as each pin may have up to two encoders (HDMI and DP) and
+	 * only the one of them (DP) will have ->hpd_pulse().
+	 */
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		enum port port = encoder->port;
+		bool long_hpd;
+
+		pin = encoder->hpd_pin;
+		if (!(BIT(pin) & pin_mask))
+			continue;
+
+		if (!intel_encoder_has_hpd_pulse(encoder))
+			continue;
+
+		long_hpd = long_mask & BIT(pin);
+
+		drm_dbg(&dev_priv->drm,
+			"digital hpd on [ENCODER:%d:%s] - %s\n",
+			encoder->base.base.id, encoder->base.name,
+			long_hpd ? "long" : "short");
+		queue_dig = true;
+
+		if (long_hpd) {
+			long_hpd_pulse_mask |= BIT(pin);
+			dev_priv->display.hotplug.long_port_mask |= BIT(port);
+		} else {
+			short_hpd_pulse_mask |= BIT(pin);
+			dev_priv->display.hotplug.short_port_mask |= BIT(port);
+		}
+	}
+
+	/* Now process each pin just once */
+	for_each_hpd_pin(pin) {
+		bool long_hpd;
+
+		if (!(BIT(pin) & pin_mask))
+			continue;
+
+		if (dev_priv->display.hotplug.stats[pin].state == HPD_DISABLED) {
+			/*
+			 * On GMCH platforms the interrupt mask bits only
+			 * prevent irq generation, not the setting of the
+			 * hotplug bits itself. So only WARN about unexpected
+			 * interrupts on saner platforms.
+			 */
+			drm_WARN_ONCE(&dev_priv->drm, !HAS_GMCH(dev_priv),
+				      "Received HPD interrupt on pin %d although disabled\n",
+				      pin);
+			continue;
+		}
+
+		if (dev_priv->display.hotplug.stats[pin].state != HPD_ENABLED)
+			continue;
+
+		/*
+		 * Delegate to ->hpd_pulse() if one of the encoders for this
+		 * pin has it, otherwise let the hotplug_work deal with this
+		 * pin directly.
+		 */
+		if (((short_hpd_pulse_mask | long_hpd_pulse_mask) & BIT(pin))) {
+			long_hpd = long_hpd_pulse_mask & BIT(pin);
+		} else {
+			dev_priv->display.hotplug.event_bits |= BIT(pin);
+			long_hpd = true;
+			queue_hp = true;
+		}
+
+		if (intel_hpd_irq_storm_detect(dev_priv, pin, long_hpd)) {
+			dev_priv->display.hotplug.event_bits &= ~BIT(pin);
+			storm_detected = true;
+			queue_hp = true;
+		}
+	}
+
+	/*
+	 * Disable any IRQs that storms were detected on. Polling enablement
+	 * happens later in our hotplug work.
+	 */
+	if (storm_detected)
+		intel_hpd_irq_setup(dev_priv);
+	spin_unlock(&dev_priv->irq_lock);
+
+	/*
+	 * Our hotplug handler can grab modeset locks (by calling down into the
+	 * fb helpers). Hence it must not be run on our own dev-priv->wq work
+	 * queue for otherwise the flush_work in the pageflip code will
+	 * deadlock.
+	 */
+	if (queue_dig)
+		queue_work(dev_priv->display.hotplug.dp_wq, &dev_priv->display.hotplug.dig_port_work);
+	if (queue_hp)
+		queue_delayed_work(system_wq, &dev_priv->display.hotplug.hotplug_work, 0);
+}
+
+/**
+ * intel_hpd_init - initializes and enables hpd support
+ * @dev_priv: i915 device instance
+ *
+ * This function enables the hotplug support. It requires that interrupts have
+ * already been enabled with intel_irq_init_hw(). From this point on hotplug and
+ * poll request can run concurrently to other code, so locking rules must be
+ * obeyed.
+ *
+ * This is a separate step from interrupt enabling to simplify the locking rules
+ * in the driver load and resume code.
+ *
+ * Also see: intel_hpd_poll_enable() and intel_hpd_poll_disable().
+ */
+void intel_hpd_init(struct drm_i915_private *dev_priv)
+{
+	int i;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	for_each_hpd_pin(i) {
+		dev_priv->display.hotplug.stats[i].count = 0;
+		dev_priv->display.hotplug.stats[i].state = HPD_ENABLED;
+	}
+
+	/*
+	 * Interrupt setup is already guaranteed to be single-threaded, this is
+	 * just to make the assert_spin_locked checks happy.
+	 */
+	spin_lock_irq(&dev_priv->irq_lock);
+	intel_hpd_irq_setup(dev_priv);
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
+
+static void i915_hpd_poll_init_work(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, struct drm_i915_private,
+			     display.hotplug.poll_init_work);
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_connector_list_iter conn_iter;
+	struct intel_connector *connector;
+	bool enabled;
+
+	mutex_lock(&dev->mode_config.mutex);
+
+	enabled = READ_ONCE(dev_priv->display.hotplug.poll_enabled);
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		enum hpd_pin pin;
+
+		pin = intel_connector_hpd_pin(connector);
+		if (pin == HPD_NONE)
+			continue;
+
+		connector->base.polled = connector->polled;
+
+		if (enabled && connector->base.polled == DRM_CONNECTOR_POLL_HPD)
+			connector->base.polled = DRM_CONNECTOR_POLL_CONNECT |
+				DRM_CONNECTOR_POLL_DISCONNECT;
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	if (enabled)
+		drm_kms_helper_poll_enable(dev);
+
+	mutex_unlock(&dev->mode_config.mutex);
+
+	/*
+	 * We might have missed any hotplugs that happened while we were
+	 * in the middle of disabling polling
+	 */
+	if (!enabled)
+		drm_helper_hpd_irq_event(dev);
+}
+
+/**
+ * intel_hpd_poll_enable - enable polling for connectors with hpd
+ * @dev_priv: i915 device instance
+ *
+ * This function enables polling for all connectors which support HPD.
+ * Under certain conditions HPD may not be functional. On most Intel GPUs,
+ * this happens when we enter runtime suspend.
+ * On Valleyview and Cherryview systems, this also happens when we shut off all
+ * of the powerwells.
+ *
+ * Since this function can get called in contexts where we're already holding
+ * dev->mode_config.mutex, we do the actual hotplug enabling in a seperate
+ * worker.
+ *
+ * Also see: intel_hpd_init() and intel_hpd_poll_disable().
+ */
+void intel_hpd_poll_enable(struct drm_i915_private *dev_priv)
+{
+	if (!HAS_DISPLAY(dev_priv) ||
+	    !INTEL_DISPLAY_ENABLED(dev_priv))
+		return;
+
+	WRITE_ONCE(dev_priv->display.hotplug.poll_enabled, true);
+
+	/*
+	 * We might already be holding dev->mode_config.mutex, so do this in a
+	 * seperate worker
+	 * As well, there's no issue if we race here since we always reschedule
+	 * this worker anyway
+	 */
+	schedule_work(&dev_priv->display.hotplug.poll_init_work);
+}
+
+/**
+ * intel_hpd_poll_disable - disable polling for connectors with hpd
+ * @dev_priv: i915 device instance
+ *
+ * This function disables polling for all connectors which support HPD.
+ * Under certain conditions HPD may not be functional. On most Intel GPUs,
+ * this happens when we enter runtime suspend.
+ * On Valleyview and Cherryview systems, this also happens when we shut off all
+ * of the powerwells.
+ *
+ * Since this function can get called in contexts where we're already holding
+ * dev->mode_config.mutex, we do the actual hotplug enabling in a seperate
+ * worker.
+ *
+ * Also used during driver init to initialize connector->polled
+ * appropriately for all connectors.
+ *
+ * Also see: intel_hpd_init() and intel_hpd_poll_enable().
+ */
+void intel_hpd_poll_disable(struct drm_i915_private *dev_priv)
+{
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	WRITE_ONCE(dev_priv->display.hotplug.poll_enabled, false);
+	schedule_work(&dev_priv->display.hotplug.poll_init_work);
+}
+
+void intel_hpd_init_work(struct drm_i915_private *dev_priv)
+{
+	INIT_DELAYED_WORK(&dev_priv->display.hotplug.hotplug_work,
+			  i915_hotplug_work_func);
+	INIT_WORK(&dev_priv->display.hotplug.dig_port_work, i915_digport_work_func);
+	INIT_WORK(&dev_priv->display.hotplug.poll_init_work, i915_hpd_poll_init_work);
+	INIT_DELAYED_WORK(&dev_priv->display.hotplug.reenable_work,
+			  intel_hpd_irq_storm_reenable_work);
+}
+
+void intel_hpd_cancel_work(struct drm_i915_private *dev_priv)
+{
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	spin_lock_irq(&dev_priv->irq_lock);
+
+	dev_priv->display.hotplug.long_port_mask = 0;
+	dev_priv->display.hotplug.short_port_mask = 0;
+	dev_priv->display.hotplug.event_bits = 0;
+	dev_priv->display.hotplug.retry_bits = 0;
+
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	cancel_work_sync(&dev_priv->display.hotplug.dig_port_work);
+	cancel_delayed_work_sync(&dev_priv->display.hotplug.hotplug_work);
+	cancel_work_sync(&dev_priv->display.hotplug.poll_init_work);
+	cancel_delayed_work_sync(&dev_priv->display.hotplug.reenable_work);
+}
+
+bool intel_hpd_disable(struct drm_i915_private *dev_priv, enum hpd_pin pin)
+{
+	bool ret = false;
+
+	if (pin == HPD_NONE)
+		return false;
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	if (dev_priv->display.hotplug.stats[pin].state == HPD_ENABLED) {
+		dev_priv->display.hotplug.stats[pin].state = HPD_DISABLED;
+		ret = true;
+	}
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	return ret;
+}
+
+void intel_hpd_enable(struct drm_i915_private *dev_priv, enum hpd_pin pin)
+{
+	if (pin == HPD_NONE)
+		return;
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	dev_priv->display.hotplug.stats[pin].state = HPD_ENABLED;
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_hotplug.h b/drivers/gpu/drm/i915/display/intel_hotplug.h
new file mode 100644
index 000000000..b87e95d60
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_hotplug.h
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_HOTPLUG_H__
+#define __INTEL_HOTPLUG_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct intel_connector;
+struct intel_digital_port;
+struct intel_encoder;
+enum port;
+
+void intel_hpd_poll_enable(struct drm_i915_private *dev_priv);
+void intel_hpd_poll_disable(struct drm_i915_private *dev_priv);
+enum intel_hotplug_state intel_encoder_hotplug(struct intel_encoder *encoder,
+					       struct intel_connector *connector);
+void intel_hpd_irq_handler(struct drm_i915_private *dev_priv,
+			   u32 pin_mask, u32 long_mask);
+void intel_hpd_trigger_irq(struct intel_digital_port *dig_port);
+void intel_hpd_init(struct drm_i915_private *dev_priv);
+void intel_hpd_init_work(struct drm_i915_private *dev_priv);
+void intel_hpd_cancel_work(struct drm_i915_private *dev_priv);
+enum hpd_pin intel_hpd_pin_default(struct drm_i915_private *dev_priv,
+				   enum port port);
+bool intel_hpd_disable(struct drm_i915_private *dev_priv, enum hpd_pin pin);
+void intel_hpd_enable(struct drm_i915_private *dev_priv, enum hpd_pin pin);
+
+#endif /* __INTEL_HOTPLUG_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_lpe_audio.c b/drivers/gpu/drm/i915/display/intel_lpe_audio.c
new file mode 100644
index 000000000..dca6003cc
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_lpe_audio.c
@@ -0,0 +1,368 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
+ *    Jerome Anand <jerome.anand@intel.com>
+ *    based on VED patches
+ *
+ */
+
+/**
+ * DOC: LPE Audio integration for HDMI or DP playback
+ *
+ * Motivation:
+ * Atom platforms (e.g. valleyview and cherryTrail) integrates a DMA-based
+ * interface as an alternative to the traditional HDaudio path. While this
+ * mode is unrelated to the LPE aka SST audio engine, the documentation refers
+ * to this mode as LPE so we keep this notation for the sake of consistency.
+ *
+ * The interface is handled by a separate standalone driver maintained in the
+ * ALSA subsystem for simplicity. To minimize the interaction between the two
+ * subsystems, a bridge is setup between the hdmi-lpe-audio and i915:
+ * 1. Create a platform device to share MMIO/IRQ resources
+ * 2. Make the platform device child of i915 device for runtime PM.
+ * 3. Create IRQ chip to forward the LPE audio irqs.
+ * the hdmi-lpe-audio driver probes the lpe audio device and creates a new
+ * sound card
+ *
+ * Threats:
+ * Due to the restriction in Linux platform device model, user need manually
+ * uninstall the hdmi-lpe-audio driver before uninstalling i915 module,
+ * otherwise we might run into use-after-free issues after i915 removes the
+ * platform device: even though hdmi-lpe-audio driver is released, the modules
+ * is still in "installed" status.
+ *
+ * Implementation:
+ * The MMIO/REG platform resources are created according to the registers
+ * specification.
+ * When forwarding LPE audio irqs, the flow control handler selection depends
+ * on the platform, for example on valleyview handle_simple_irq is enough.
+ *
+ */
+
+#include <linux/acpi.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/irq.h>
+#include <linux/pci.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+
+#include <drm/intel_lpe_audio.h>
+
+#include "i915_drv.h"
+#include "intel_de.h"
+#include "intel_lpe_audio.h"
+#include "intel_pci_config.h"
+
+#define HAS_LPE_AUDIO(dev_priv) ((dev_priv)->display.audio.lpe.platdev != NULL)
+
+static struct platform_device *
+lpe_audio_platdev_create(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct pci_dev *pdev = to_pci_dev(dev->dev);
+	struct platform_device_info pinfo = {};
+	struct resource *rsc;
+	struct platform_device *platdev;
+	struct intel_hdmi_lpe_audio_pdata *pdata;
+
+	pdata = kzalloc(sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return ERR_PTR(-ENOMEM);
+
+	rsc = kcalloc(2, sizeof(*rsc), GFP_KERNEL);
+	if (!rsc) {
+		kfree(pdata);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	rsc[0].start    = rsc[0].end = dev_priv->display.audio.lpe.irq;
+	rsc[0].flags    = IORESOURCE_IRQ;
+	rsc[0].name     = "hdmi-lpe-audio-irq";
+
+	rsc[1].start    = pci_resource_start(pdev, GTTMMADR_BAR) +
+		I915_HDMI_LPE_AUDIO_BASE;
+	rsc[1].end      = pci_resource_start(pdev, GTTMMADR_BAR) +
+		I915_HDMI_LPE_AUDIO_BASE + I915_HDMI_LPE_AUDIO_SIZE - 1;
+	rsc[1].flags    = IORESOURCE_MEM;
+	rsc[1].name     = "hdmi-lpe-audio-mmio";
+
+	pinfo.parent = dev->dev;
+	pinfo.name = "hdmi-lpe-audio";
+	pinfo.id = -1;
+	pinfo.res = rsc;
+	pinfo.num_res = 2;
+	pinfo.data = pdata;
+	pinfo.size_data = sizeof(*pdata);
+	pinfo.dma_mask = DMA_BIT_MASK(32);
+
+	pdata->num_pipes = INTEL_NUM_PIPES(dev_priv);
+	pdata->num_ports = IS_CHERRYVIEW(dev_priv) ? 3 : 2; /* B,C,D or B,C */
+	pdata->port[0].pipe = -1;
+	pdata->port[1].pipe = -1;
+	pdata->port[2].pipe = -1;
+	spin_lock_init(&pdata->lpe_audio_slock);
+
+	platdev = platform_device_register_full(&pinfo);
+	kfree(rsc);
+	kfree(pdata);
+
+	if (IS_ERR(platdev)) {
+		drm_err(&dev_priv->drm,
+			"Failed to allocate LPE audio platform device\n");
+		return platdev;
+	}
+
+	pm_runtime_no_callbacks(&platdev->dev);
+
+	return platdev;
+}
+
+static void lpe_audio_platdev_destroy(struct drm_i915_private *dev_priv)
+{
+	/* XXX Note that platform_device_register_full() allocates a dma_mask
+	 * and never frees it. We can't free it here as we cannot guarantee
+	 * this is the last reference (i.e. that the dma_mask will not be
+	 * used after our unregister). So ee choose to leak the sizeof(u64)
+	 * allocation here - it should be fixed in the platform_device rather
+	 * than us fiddle with its internals.
+	 */
+
+	platform_device_unregister(dev_priv->display.audio.lpe.platdev);
+}
+
+static void lpe_audio_irq_unmask(struct irq_data *d)
+{
+}
+
+static void lpe_audio_irq_mask(struct irq_data *d)
+{
+}
+
+static struct irq_chip lpe_audio_irqchip = {
+	.name = "hdmi_lpe_audio_irqchip",
+	.irq_mask = lpe_audio_irq_mask,
+	.irq_unmask = lpe_audio_irq_unmask,
+};
+
+static int lpe_audio_irq_init(struct drm_i915_private *dev_priv)
+{
+	int irq = dev_priv->display.audio.lpe.irq;
+
+	drm_WARN_ON(&dev_priv->drm, !intel_irqs_enabled(dev_priv));
+	irq_set_chip_and_handler_name(irq,
+				&lpe_audio_irqchip,
+				handle_simple_irq,
+				"hdmi_lpe_audio_irq_handler");
+
+	return irq_set_chip_data(irq, dev_priv);
+}
+
+static bool lpe_audio_detect(struct drm_i915_private *dev_priv)
+{
+	int lpe_present = false;
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		static const struct pci_device_id atom_hdaudio_ids[] = {
+			/* Baytrail */
+			{PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x0f04)},
+			/* Braswell */
+			{PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x2284)},
+			{}
+		};
+
+		if (!pci_dev_present(atom_hdaudio_ids)) {
+			drm_info(&dev_priv->drm,
+				 "HDaudio controller not detected, using LPE audio instead\n");
+			lpe_present = true;
+		}
+	}
+	return lpe_present;
+}
+
+static int lpe_audio_setup(struct drm_i915_private *dev_priv)
+{
+	int ret;
+
+	dev_priv->display.audio.lpe.irq = irq_alloc_desc(0);
+	if (dev_priv->display.audio.lpe.irq < 0) {
+		drm_err(&dev_priv->drm, "Failed to allocate IRQ desc: %d\n",
+			dev_priv->display.audio.lpe.irq);
+		ret = dev_priv->display.audio.lpe.irq;
+		goto err;
+	}
+
+	drm_dbg(&dev_priv->drm, "irq = %d\n", dev_priv->display.audio.lpe.irq);
+
+	ret = lpe_audio_irq_init(dev_priv);
+
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"Failed to initialize irqchip for lpe audio: %d\n",
+			ret);
+		goto err_free_irq;
+	}
+
+	dev_priv->display.audio.lpe.platdev = lpe_audio_platdev_create(dev_priv);
+
+	if (IS_ERR(dev_priv->display.audio.lpe.platdev)) {
+		ret = PTR_ERR(dev_priv->display.audio.lpe.platdev);
+		drm_err(&dev_priv->drm,
+			"Failed to create lpe audio platform device: %d\n",
+			ret);
+		goto err_free_irq;
+	}
+
+	/* enable chicken bit; at least this is required for Dell Wyse 3040
+	 * with DP outputs (but only sometimes by some reason!)
+	 */
+	intel_de_write(dev_priv, VLV_AUD_CHICKEN_BIT_REG,
+		       VLV_CHICKEN_BIT_DBG_ENABLE);
+
+	return 0;
+err_free_irq:
+	irq_free_desc(dev_priv->display.audio.lpe.irq);
+err:
+	dev_priv->display.audio.lpe.irq = -1;
+	dev_priv->display.audio.lpe.platdev = NULL;
+	return ret;
+}
+
+/**
+ * intel_lpe_audio_irq_handler() - forwards the LPE audio irq
+ * @dev_priv: the i915 drm device private data
+ *
+ * the LPE Audio irq is forwarded to the irq handler registered by LPE audio
+ * driver.
+ */
+void intel_lpe_audio_irq_handler(struct drm_i915_private *dev_priv)
+{
+	int ret;
+
+	if (!HAS_LPE_AUDIO(dev_priv))
+		return;
+
+	ret = generic_handle_irq(dev_priv->display.audio.lpe.irq);
+	if (ret)
+		drm_err_ratelimited(&dev_priv->drm,
+				    "error handling LPE audio irq: %d\n", ret);
+}
+
+/**
+ * intel_lpe_audio_init() - detect and setup the bridge between HDMI LPE Audio
+ * driver and i915
+ * @dev_priv: the i915 drm device private data
+ *
+ * Return: 0 if successful. non-zero if detection or
+ * llocation/initialization fails
+ */
+int intel_lpe_audio_init(struct drm_i915_private *dev_priv)
+{
+	int ret = -ENODEV;
+
+	if (lpe_audio_detect(dev_priv)) {
+		ret = lpe_audio_setup(dev_priv);
+		if (ret < 0)
+			drm_err(&dev_priv->drm,
+				"failed to setup LPE Audio bridge\n");
+	}
+	return ret;
+}
+
+/**
+ * intel_lpe_audio_teardown() - destroy the bridge between HDMI LPE
+ * audio driver and i915
+ * @dev_priv: the i915 drm device private data
+ *
+ * release all the resources for LPE audio <-> i915 bridge.
+ */
+void intel_lpe_audio_teardown(struct drm_i915_private *dev_priv)
+{
+	if (!HAS_LPE_AUDIO(dev_priv))
+		return;
+
+	lpe_audio_platdev_destroy(dev_priv);
+
+	irq_free_desc(dev_priv->display.audio.lpe.irq);
+
+	dev_priv->display.audio.lpe.irq = -1;
+	dev_priv->display.audio.lpe.platdev = NULL;
+}
+
+/**
+ * intel_lpe_audio_notify() - notify lpe audio event
+ * audio driver and i915
+ * @dev_priv: the i915 drm device private data
+ * @pipe: pipe
+ * @port: port
+ * @eld : ELD data
+ * @ls_clock: Link symbol clock in kHz
+ * @dp_output: Driving a DP output?
+ *
+ * Notify lpe audio driver of eld change.
+ */
+void intel_lpe_audio_notify(struct drm_i915_private *dev_priv,
+			    enum pipe pipe, enum port port,
+			    const void *eld, int ls_clock, bool dp_output)
+{
+	unsigned long irqflags;
+	struct intel_hdmi_lpe_audio_pdata *pdata;
+	struct intel_hdmi_lpe_audio_port_pdata *ppdata;
+	u32 audio_enable;
+
+	if (!HAS_LPE_AUDIO(dev_priv))
+		return;
+
+	pdata = dev_get_platdata(&dev_priv->display.audio.lpe.platdev->dev);
+	ppdata = &pdata->port[port - PORT_B];
+
+	spin_lock_irqsave(&pdata->lpe_audio_slock, irqflags);
+
+	audio_enable = intel_de_read(dev_priv, VLV_AUD_PORT_EN_DBG(port));
+
+	if (eld != NULL) {
+		memcpy(ppdata->eld, eld, HDMI_MAX_ELD_BYTES);
+		ppdata->pipe = pipe;
+		ppdata->ls_clock = ls_clock;
+		ppdata->dp_output = dp_output;
+
+		/* Unmute the amp for both DP and HDMI */
+		intel_de_write(dev_priv, VLV_AUD_PORT_EN_DBG(port),
+			       audio_enable & ~VLV_AMP_MUTE);
+	} else {
+		memset(ppdata->eld, 0, HDMI_MAX_ELD_BYTES);
+		ppdata->pipe = -1;
+		ppdata->ls_clock = 0;
+		ppdata->dp_output = false;
+
+		/* Mute the amp for both DP and HDMI */
+		intel_de_write(dev_priv, VLV_AUD_PORT_EN_DBG(port),
+			       audio_enable | VLV_AMP_MUTE);
+	}
+
+	if (pdata->notify_audio_lpe)
+		pdata->notify_audio_lpe(dev_priv->display.audio.lpe.platdev, port - PORT_B);
+
+	spin_unlock_irqrestore(&pdata->lpe_audio_slock, irqflags);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_lpe_audio.h b/drivers/gpu/drm/i915/display/intel_lpe_audio.h
new file mode 100644
index 000000000..f848c5038
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_lpe_audio.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_LPE_AUDIO_H__
+#define __INTEL_LPE_AUDIO_H__
+
+#include <linux/types.h>
+
+enum pipe;
+enum port;
+struct drm_i915_private;
+
+int  intel_lpe_audio_init(struct drm_i915_private *dev_priv);
+void intel_lpe_audio_teardown(struct drm_i915_private *dev_priv);
+void intel_lpe_audio_irq_handler(struct drm_i915_private *dev_priv);
+void intel_lpe_audio_notify(struct drm_i915_private *dev_priv,
+			    enum pipe pipe, enum port port,
+			    const void *eld, int ls_clock, bool dp_output);
+
+#endif /* __INTEL_LPE_AUDIO_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_lspcon.c b/drivers/gpu/drm/i915/display/intel_lspcon.c
new file mode 100644
index 000000000..15d59de88
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_lspcon.c
@@ -0,0 +1,716 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ *
+ */
+
+#include <drm/display/drm_dp_dual_mode_helper.h>
+#include <drm/display/drm_hdmi_helper.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_edid.h>
+
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_lspcon.h"
+#include "intel_hdmi.h"
+
+/* LSPCON OUI Vendor ID(signatures) */
+#define LSPCON_VENDOR_PARADE_OUI 0x001CF8
+#define LSPCON_VENDOR_MCA_OUI 0x0060AD
+
+#define DPCD_MCA_LSPCON_HDR_STATUS	0x70003
+#define DPCD_PARADE_LSPCON_HDR_STATUS	0x00511
+
+/* AUX addresses to write MCA AVI IF */
+#define LSPCON_MCA_AVI_IF_WRITE_OFFSET 0x5C0
+#define LSPCON_MCA_AVI_IF_CTRL 0x5DF
+#define  LSPCON_MCA_AVI_IF_KICKOFF (1 << 0)
+#define  LSPCON_MCA_AVI_IF_HANDLED (1 << 1)
+
+/* AUX addresses to write Parade AVI IF */
+#define LSPCON_PARADE_AVI_IF_WRITE_OFFSET 0x516
+#define LSPCON_PARADE_AVI_IF_CTRL 0x51E
+#define  LSPCON_PARADE_AVI_IF_KICKOFF (1 << 7)
+#define LSPCON_PARADE_AVI_IF_DATA_SIZE 32
+
+static struct intel_dp *lspcon_to_intel_dp(struct intel_lspcon *lspcon)
+{
+	struct intel_digital_port *dig_port =
+		container_of(lspcon, struct intel_digital_port, lspcon);
+
+	return &dig_port->dp;
+}
+
+static const char *lspcon_mode_name(enum drm_lspcon_mode mode)
+{
+	switch (mode) {
+	case DRM_LSPCON_MODE_PCON:
+		return "PCON";
+	case DRM_LSPCON_MODE_LS:
+		return "LS";
+	case DRM_LSPCON_MODE_INVALID:
+		return "INVALID";
+	default:
+		MISSING_CASE(mode);
+		return "INVALID";
+	}
+}
+
+static bool lspcon_detect_vendor(struct intel_lspcon *lspcon)
+{
+	struct intel_dp *dp = lspcon_to_intel_dp(lspcon);
+	struct drm_i915_private *i915 = dp_to_i915(dp);
+	struct drm_dp_dpcd_ident *ident;
+	u32 vendor_oui;
+
+	if (drm_dp_read_desc(&dp->aux, &dp->desc, drm_dp_is_branch(dp->dpcd))) {
+		drm_err(&i915->drm, "Can't read description\n");
+		return false;
+	}
+
+	ident = &dp->desc.ident;
+	vendor_oui = (ident->oui[0] << 16) | (ident->oui[1] << 8) |
+		      ident->oui[2];
+
+	switch (vendor_oui) {
+	case LSPCON_VENDOR_MCA_OUI:
+		lspcon->vendor = LSPCON_VENDOR_MCA;
+		drm_dbg_kms(&i915->drm, "Vendor: Mega Chips\n");
+		break;
+
+	case LSPCON_VENDOR_PARADE_OUI:
+		lspcon->vendor = LSPCON_VENDOR_PARADE;
+		drm_dbg_kms(&i915->drm, "Vendor: Parade Tech\n");
+		break;
+
+	default:
+		drm_err(&i915->drm, "Invalid/Unknown vendor OUI\n");
+		return false;
+	}
+
+	return true;
+}
+
+static u32 get_hdr_status_reg(struct intel_lspcon *lspcon)
+{
+	if (lspcon->vendor == LSPCON_VENDOR_MCA)
+		return DPCD_MCA_LSPCON_HDR_STATUS;
+	else
+		return DPCD_PARADE_LSPCON_HDR_STATUS;
+}
+
+void lspcon_detect_hdr_capability(struct intel_lspcon *lspcon)
+{
+	struct intel_dp *intel_dp = lspcon_to_intel_dp(lspcon);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 hdr_caps;
+	int ret;
+
+	ret = drm_dp_dpcd_read(&intel_dp->aux, get_hdr_status_reg(lspcon),
+			       &hdr_caps, 1);
+
+	if (ret < 0) {
+		drm_dbg_kms(&i915->drm, "HDR capability detection failed\n");
+		lspcon->hdr_supported = false;
+	} else if (hdr_caps & 0x1) {
+		drm_dbg_kms(&i915->drm, "LSPCON capable of HDR\n");
+		lspcon->hdr_supported = true;
+	}
+}
+
+static enum drm_lspcon_mode lspcon_get_current_mode(struct intel_lspcon *lspcon)
+{
+	struct intel_dp *intel_dp = lspcon_to_intel_dp(lspcon);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	enum drm_lspcon_mode current_mode;
+	struct i2c_adapter *adapter = &intel_dp->aux.ddc;
+
+	if (drm_lspcon_get_mode(intel_dp->aux.drm_dev, adapter, &current_mode)) {
+		drm_dbg_kms(&i915->drm, "Error reading LSPCON mode\n");
+		return DRM_LSPCON_MODE_INVALID;
+	}
+	return current_mode;
+}
+
+static enum drm_lspcon_mode lspcon_wait_mode(struct intel_lspcon *lspcon,
+					     enum drm_lspcon_mode mode)
+{
+	struct intel_dp *intel_dp = lspcon_to_intel_dp(lspcon);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	enum drm_lspcon_mode current_mode;
+
+	current_mode = lspcon_get_current_mode(lspcon);
+	if (current_mode == mode)
+		goto out;
+
+	drm_dbg_kms(&i915->drm, "Waiting for LSPCON mode %s to settle\n",
+		    lspcon_mode_name(mode));
+
+	wait_for((current_mode = lspcon_get_current_mode(lspcon)) == mode, 400);
+	if (current_mode != mode)
+		drm_err(&i915->drm, "LSPCON mode hasn't settled\n");
+
+out:
+	drm_dbg_kms(&i915->drm, "Current LSPCON mode %s\n",
+		    lspcon_mode_name(current_mode));
+
+	return current_mode;
+}
+
+static int lspcon_change_mode(struct intel_lspcon *lspcon,
+			      enum drm_lspcon_mode mode)
+{
+	struct intel_dp *intel_dp = lspcon_to_intel_dp(lspcon);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int err;
+	enum drm_lspcon_mode current_mode;
+	struct i2c_adapter *adapter = &intel_dp->aux.ddc;
+
+	err = drm_lspcon_get_mode(intel_dp->aux.drm_dev, adapter, &current_mode);
+	if (err) {
+		drm_err(&i915->drm, "Error reading LSPCON mode\n");
+		return err;
+	}
+
+	if (current_mode == mode) {
+		drm_dbg_kms(&i915->drm, "Current mode = desired LSPCON mode\n");
+		return 0;
+	}
+
+	err = drm_lspcon_set_mode(intel_dp->aux.drm_dev, adapter, mode);
+	if (err < 0) {
+		drm_err(&i915->drm, "LSPCON mode change failed\n");
+		return err;
+	}
+
+	lspcon->mode = mode;
+	drm_dbg_kms(&i915->drm, "LSPCON mode changed done\n");
+	return 0;
+}
+
+static bool lspcon_wake_native_aux_ch(struct intel_lspcon *lspcon)
+{
+	struct intel_dp *intel_dp = lspcon_to_intel_dp(lspcon);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 rev;
+
+	if (drm_dp_dpcd_readb(&lspcon_to_intel_dp(lspcon)->aux, DP_DPCD_REV,
+			      &rev) != 1) {
+		drm_dbg_kms(&i915->drm, "Native AUX CH down\n");
+		return false;
+	}
+
+	drm_dbg_kms(&i915->drm, "Native AUX CH up, DPCD version: %d.%d\n",
+		    rev >> 4, rev & 0xf);
+
+	return true;
+}
+
+static bool lspcon_probe(struct intel_lspcon *lspcon)
+{
+	int retry;
+	enum drm_dp_dual_mode_type adaptor_type;
+	struct intel_dp *intel_dp = lspcon_to_intel_dp(lspcon);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct i2c_adapter *adapter = &intel_dp->aux.ddc;
+	enum drm_lspcon_mode expected_mode;
+
+	expected_mode = lspcon_wake_native_aux_ch(lspcon) ?
+			DRM_LSPCON_MODE_PCON : DRM_LSPCON_MODE_LS;
+
+	/* Lets probe the adaptor and check its type */
+	for (retry = 0; retry < 6; retry++) {
+		if (retry)
+			usleep_range(500, 1000);
+
+		adaptor_type = drm_dp_dual_mode_detect(intel_dp->aux.drm_dev, adapter);
+		if (adaptor_type == DRM_DP_DUAL_MODE_LSPCON)
+			break;
+	}
+
+	if (adaptor_type != DRM_DP_DUAL_MODE_LSPCON) {
+		drm_dbg_kms(&i915->drm, "No LSPCON detected, found %s\n",
+			    drm_dp_get_dual_mode_type_name(adaptor_type));
+		return false;
+	}
+
+	/* Yay ... got a LSPCON device */
+	drm_dbg_kms(&i915->drm, "LSPCON detected\n");
+	lspcon->mode = lspcon_wait_mode(lspcon, expected_mode);
+
+	/*
+	 * In the SW state machine, lets Put LSPCON in PCON mode only.
+	 * In this way, it will work with both HDMI 1.4 sinks as well as HDMI
+	 * 2.0 sinks.
+	 */
+	if (lspcon->mode != DRM_LSPCON_MODE_PCON) {
+		if (lspcon_change_mode(lspcon, DRM_LSPCON_MODE_PCON) < 0) {
+			drm_err(&i915->drm, "LSPCON mode change to PCON failed\n");
+			return false;
+		}
+	}
+	return true;
+}
+
+static void lspcon_resume_in_pcon_wa(struct intel_lspcon *lspcon)
+{
+	struct intel_dp *intel_dp = lspcon_to_intel_dp(lspcon);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	unsigned long start = jiffies;
+
+	while (1) {
+		if (intel_digital_port_connected(&dig_port->base)) {
+			drm_dbg_kms(&i915->drm, "LSPCON recovering in PCON mode after %u ms\n",
+				    jiffies_to_msecs(jiffies - start));
+			return;
+		}
+
+		if (time_after(jiffies, start + msecs_to_jiffies(1000)))
+			break;
+
+		usleep_range(10000, 15000);
+	}
+
+	drm_dbg_kms(&i915->drm, "LSPCON DP descriptor mismatch after resume\n");
+}
+
+static bool lspcon_parade_fw_ready(struct drm_dp_aux *aux)
+{
+	u8 avi_if_ctrl;
+	u8 retry;
+	ssize_t ret;
+
+	/* Check if LSPCON FW is ready for data */
+	for (retry = 0; retry < 5; retry++) {
+		if (retry)
+			usleep_range(200, 300);
+
+		ret = drm_dp_dpcd_read(aux, LSPCON_PARADE_AVI_IF_CTRL,
+				       &avi_if_ctrl, 1);
+		if (ret < 0) {
+			drm_err(aux->drm_dev, "Failed to read AVI IF control\n");
+			return false;
+		}
+
+		if ((avi_if_ctrl & LSPCON_PARADE_AVI_IF_KICKOFF) == 0)
+			return true;
+	}
+
+	drm_err(aux->drm_dev, "Parade FW not ready to accept AVI IF\n");
+	return false;
+}
+
+static bool _lspcon_parade_write_infoframe_blocks(struct drm_dp_aux *aux,
+						  u8 *avi_buf)
+{
+	u8 avi_if_ctrl;
+	u8 block_count = 0;
+	u8 *data;
+	u16 reg;
+	ssize_t ret;
+
+	while (block_count < 4) {
+		if (!lspcon_parade_fw_ready(aux)) {
+			drm_dbg_kms(aux->drm_dev, "LSPCON FW not ready, block %d\n",
+				    block_count);
+			return false;
+		}
+
+		reg = LSPCON_PARADE_AVI_IF_WRITE_OFFSET;
+		data = avi_buf + block_count * 8;
+		ret = drm_dp_dpcd_write(aux, reg, data, 8);
+		if (ret < 0) {
+			drm_err(aux->drm_dev, "Failed to write AVI IF block %d\n",
+				block_count);
+			return false;
+		}
+
+		/*
+		 * Once a block of data is written, we have to inform the FW
+		 * about this by writing into avi infoframe control register:
+		 * - set the kickoff bit[7] to 1
+		 * - write the block no. to bits[1:0]
+		 */
+		reg = LSPCON_PARADE_AVI_IF_CTRL;
+		avi_if_ctrl = LSPCON_PARADE_AVI_IF_KICKOFF | block_count;
+		ret = drm_dp_dpcd_write(aux, reg, &avi_if_ctrl, 1);
+		if (ret < 0) {
+			drm_err(aux->drm_dev, "Failed to update (0x%x), block %d\n",
+				reg, block_count);
+			return false;
+		}
+
+		block_count++;
+	}
+
+	drm_dbg_kms(aux->drm_dev, "Wrote AVI IF blocks successfully\n");
+	return true;
+}
+
+static bool _lspcon_write_avi_infoframe_parade(struct drm_dp_aux *aux,
+					       const u8 *frame,
+					       ssize_t len)
+{
+	u8 avi_if[LSPCON_PARADE_AVI_IF_DATA_SIZE] = {1, };
+
+	/*
+	 * Parade's frames contains 32 bytes of data, divided
+	 * into 4 frames:
+	 *	Token byte (first byte of first frame, must be non-zero)
+	 *	HB0 to HB2	 from AVI IF (3 bytes header)
+	 *	PB0 to PB27 from AVI IF (28 bytes data)
+	 * So it should look like this
+	 *	first block: | <token> <HB0-HB2> <DB0-DB3> |
+	 *	next 3 blocks: |<DB4-DB11>|<DB12-DB19>|<DB20-DB28>|
+	 */
+
+	if (len > LSPCON_PARADE_AVI_IF_DATA_SIZE - 1) {
+		drm_err(aux->drm_dev, "Invalid length of infoframes\n");
+		return false;
+	}
+
+	memcpy(&avi_if[1], frame, len);
+
+	if (!_lspcon_parade_write_infoframe_blocks(aux, avi_if)) {
+		drm_dbg_kms(aux->drm_dev, "Failed to write infoframe blocks\n");
+		return false;
+	}
+
+	return true;
+}
+
+static bool _lspcon_write_avi_infoframe_mca(struct drm_dp_aux *aux,
+					    const u8 *buffer, ssize_t len)
+{
+	int ret;
+	u32 val = 0;
+	u32 retry;
+	u16 reg;
+	const u8 *data = buffer;
+
+	reg = LSPCON_MCA_AVI_IF_WRITE_OFFSET;
+	while (val < len) {
+		/* DPCD write for AVI IF can fail on a slow FW day, so retry */
+		for (retry = 0; retry < 5; retry++) {
+			ret = drm_dp_dpcd_write(aux, reg, (void *)data, 1);
+			if (ret == 1) {
+				break;
+			} else if (retry < 4) {
+				mdelay(50);
+				continue;
+			} else {
+				drm_err(aux->drm_dev, "DPCD write failed at:0x%x\n", reg);
+				return false;
+			}
+		}
+		val++; reg++; data++;
+	}
+
+	val = 0;
+	reg = LSPCON_MCA_AVI_IF_CTRL;
+	ret = drm_dp_dpcd_read(aux, reg, &val, 1);
+	if (ret < 0) {
+		drm_err(aux->drm_dev, "DPCD read failed, address 0x%x\n", reg);
+		return false;
+	}
+
+	/* Indicate LSPCON chip about infoframe, clear bit 1 and set bit 0 */
+	val &= ~LSPCON_MCA_AVI_IF_HANDLED;
+	val |= LSPCON_MCA_AVI_IF_KICKOFF;
+
+	ret = drm_dp_dpcd_write(aux, reg, &val, 1);
+	if (ret < 0) {
+		drm_err(aux->drm_dev, "DPCD read failed, address 0x%x\n", reg);
+		return false;
+	}
+
+	val = 0;
+	ret = drm_dp_dpcd_read(aux, reg, &val, 1);
+	if (ret < 0) {
+		drm_err(aux->drm_dev, "DPCD read failed, address 0x%x\n", reg);
+		return false;
+	}
+
+	if (val == LSPCON_MCA_AVI_IF_HANDLED)
+		drm_dbg_kms(aux->drm_dev, "AVI IF handled by FW\n");
+
+	return true;
+}
+
+void lspcon_write_infoframe(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    unsigned int type,
+			    const void *frame, ssize_t len)
+{
+	bool ret = true;
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct intel_lspcon *lspcon = enc_to_intel_lspcon(encoder);
+
+	switch (type) {
+	case HDMI_INFOFRAME_TYPE_AVI:
+		if (lspcon->vendor == LSPCON_VENDOR_MCA)
+			ret = _lspcon_write_avi_infoframe_mca(&intel_dp->aux,
+							      frame, len);
+		else
+			ret = _lspcon_write_avi_infoframe_parade(&intel_dp->aux,
+								 frame, len);
+		break;
+	case HDMI_PACKET_TYPE_GAMUT_METADATA:
+		drm_dbg_kms(&i915->drm, "Update HDR metadata for lspcon\n");
+		/* It uses the legacy hsw implementation for the same */
+		hsw_write_infoframe(encoder, crtc_state, type, frame, len);
+		break;
+	default:
+		return;
+	}
+
+	if (!ret) {
+		drm_err(&i915->drm, "Failed to write infoframes\n");
+		return;
+	}
+}
+
+void lspcon_read_infoframe(struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state,
+			   unsigned int type,
+			   void *frame, ssize_t len)
+{
+	/* FIXME implement for AVI Infoframe as well */
+	if (type == HDMI_PACKET_TYPE_GAMUT_METADATA)
+		hsw_read_infoframe(encoder, crtc_state, type,
+				   frame, len);
+}
+
+void lspcon_set_infoframes(struct intel_encoder *encoder,
+			   bool enable,
+			   const struct intel_crtc_state *crtc_state,
+			   const struct drm_connector_state *conn_state)
+{
+	ssize_t ret;
+	union hdmi_infoframe frame;
+	u8 buf[VIDEO_DIP_DATA_SIZE];
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_lspcon *lspcon = &dig_port->lspcon;
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+
+	if (!lspcon->active) {
+		drm_err(&i915->drm, "Writing infoframes while LSPCON disabled ?\n");
+		return;
+	}
+
+	/* FIXME precompute infoframes */
+
+	ret = drm_hdmi_avi_infoframe_from_display_mode(&frame.avi,
+						       conn_state->connector,
+						       adjusted_mode);
+	if (ret < 0) {
+		drm_err(&i915->drm, "couldn't fill AVI infoframe\n");
+		return;
+	}
+
+	/*
+	 * Currently there is no interface defined to
+	 * check user preference between RGB/YCBCR444
+	 * or YCBCR420. So the only possible case for
+	 * YCBCR444 usage is driving YCBCR420 output
+	 * with LSPCON, when pipe is configured for
+	 * YCBCR444 output and LSPCON takes care of
+	 * downsampling it.
+	 */
+	if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR444)
+		frame.avi.colorspace = HDMI_COLORSPACE_YUV420;
+	else
+		frame.avi.colorspace = HDMI_COLORSPACE_RGB;
+
+	/* Set the Colorspace as per the HDMI spec */
+	drm_hdmi_avi_infoframe_colorimetry(&frame.avi, conn_state);
+
+	/* nonsense combination */
+	drm_WARN_ON(encoder->base.dev, crtc_state->limited_color_range &&
+		    crtc_state->output_format != INTEL_OUTPUT_FORMAT_RGB);
+
+	if (crtc_state->output_format == INTEL_OUTPUT_FORMAT_RGB) {
+		drm_hdmi_avi_infoframe_quant_range(&frame.avi,
+						   conn_state->connector,
+						   adjusted_mode,
+						   crtc_state->limited_color_range ?
+						   HDMI_QUANTIZATION_RANGE_LIMITED :
+						   HDMI_QUANTIZATION_RANGE_FULL);
+	} else {
+		frame.avi.quantization_range = HDMI_QUANTIZATION_RANGE_DEFAULT;
+		frame.avi.ycc_quantization_range = HDMI_YCC_QUANTIZATION_RANGE_LIMITED;
+	}
+
+	drm_hdmi_avi_infoframe_content_type(&frame.avi, conn_state);
+
+	ret = hdmi_infoframe_pack(&frame, buf, sizeof(buf));
+	if (ret < 0) {
+		drm_err(&i915->drm, "Failed to pack AVI IF\n");
+		return;
+	}
+
+	dig_port->write_infoframe(encoder, crtc_state, HDMI_INFOFRAME_TYPE_AVI,
+				  buf, ret);
+}
+
+static bool _lspcon_read_avi_infoframe_enabled_mca(struct drm_dp_aux *aux)
+{
+	int ret;
+	u32 val = 0;
+	u16 reg = LSPCON_MCA_AVI_IF_CTRL;
+
+	ret = drm_dp_dpcd_read(aux, reg, &val, 1);
+	if (ret < 0) {
+		drm_err(aux->drm_dev, "DPCD read failed, address 0x%x\n", reg);
+		return false;
+	}
+
+	return val & LSPCON_MCA_AVI_IF_KICKOFF;
+}
+
+static bool _lspcon_read_avi_infoframe_enabled_parade(struct drm_dp_aux *aux)
+{
+	int ret;
+	u32 val = 0;
+	u16 reg = LSPCON_PARADE_AVI_IF_CTRL;
+
+	ret = drm_dp_dpcd_read(aux, reg, &val, 1);
+	if (ret < 0) {
+		drm_err(aux->drm_dev, "DPCD read failed, address 0x%x\n", reg);
+		return false;
+	}
+
+	return val & LSPCON_PARADE_AVI_IF_KICKOFF;
+}
+
+u32 lspcon_infoframes_enabled(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *pipe_config)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_lspcon *lspcon = enc_to_intel_lspcon(encoder);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	bool infoframes_enabled;
+	u32 val = 0;
+	u32 mask, tmp;
+
+	if (lspcon->vendor == LSPCON_VENDOR_MCA)
+		infoframes_enabled = _lspcon_read_avi_infoframe_enabled_mca(&intel_dp->aux);
+	else
+		infoframes_enabled = _lspcon_read_avi_infoframe_enabled_parade(&intel_dp->aux);
+
+	if (infoframes_enabled)
+		val |= intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_AVI);
+
+	if (lspcon->hdr_supported) {
+		tmp = intel_de_read(dev_priv,
+				    HSW_TVIDEO_DIP_CTL(pipe_config->cpu_transcoder));
+		mask = VIDEO_DIP_ENABLE_GMP_HSW;
+
+		if (tmp & mask)
+			val |= intel_hdmi_infoframe_enable(HDMI_PACKET_TYPE_GAMUT_METADATA);
+	}
+
+	return val;
+}
+
+void lspcon_wait_pcon_mode(struct intel_lspcon *lspcon)
+{
+	lspcon_wait_mode(lspcon, DRM_LSPCON_MODE_PCON);
+}
+
+bool lspcon_init(struct intel_digital_port *dig_port)
+{
+	struct intel_dp *intel_dp = &dig_port->dp;
+	struct intel_lspcon *lspcon = &dig_port->lspcon;
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct drm_connector *connector = &intel_dp->attached_connector->base;
+
+	lspcon->active = false;
+	lspcon->mode = DRM_LSPCON_MODE_INVALID;
+
+	if (!lspcon_probe(lspcon)) {
+		drm_err(&i915->drm, "Failed to probe lspcon\n");
+		return false;
+	}
+
+	if (drm_dp_read_dpcd_caps(&intel_dp->aux, intel_dp->dpcd) != 0) {
+		drm_err(&i915->drm, "LSPCON DPCD read failed\n");
+		return false;
+	}
+
+	if (!lspcon_detect_vendor(lspcon)) {
+		drm_err(&i915->drm, "LSPCON vendor detection failed\n");
+		return false;
+	}
+
+	connector->ycbcr_420_allowed = true;
+	lspcon->active = true;
+	drm_dbg_kms(&i915->drm, "Success: LSPCON init\n");
+	return true;
+}
+
+u32 intel_lspcon_infoframes_enabled(struct intel_encoder *encoder,
+				    const struct intel_crtc_state *pipe_config)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+
+	return dig_port->infoframes_enabled(encoder, pipe_config);
+}
+
+void lspcon_resume(struct intel_digital_port *dig_port)
+{
+	struct intel_lspcon *lspcon = &dig_port->lspcon;
+	struct drm_device *dev = dig_port->base.base.dev;
+	struct drm_i915_private *i915 = to_i915(dev);
+	enum drm_lspcon_mode expected_mode;
+
+	if (!intel_bios_is_lspcon_present(i915, dig_port->base.port))
+		return;
+
+	if (!lspcon->active) {
+		if (!lspcon_init(dig_port)) {
+			drm_err(&i915->drm, "LSPCON init failed on port %c\n",
+				port_name(dig_port->base.port));
+			return;
+		}
+	}
+
+	if (lspcon_wake_native_aux_ch(lspcon)) {
+		expected_mode = DRM_LSPCON_MODE_PCON;
+		lspcon_resume_in_pcon_wa(lspcon);
+	} else {
+		expected_mode = DRM_LSPCON_MODE_LS;
+	}
+
+	if (lspcon_wait_mode(lspcon, expected_mode) == DRM_LSPCON_MODE_PCON)
+		return;
+
+	if (lspcon_change_mode(lspcon, DRM_LSPCON_MODE_PCON))
+		drm_err(&i915->drm, "LSPCON resume failed\n");
+	else
+		drm_dbg_kms(&i915->drm, "LSPCON resume success\n");
+}
diff --git a/drivers/gpu/drm/i915/display/intel_lspcon.h b/drivers/gpu/drm/i915/display/intel_lspcon.h
new file mode 100644
index 000000000..e19e10492
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_lspcon.h
@@ -0,0 +1,47 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_LSPCON_H__
+#define __INTEL_LSPCON_H__
+
+#include <linux/types.h>
+
+struct drm_connector;
+struct drm_connector_state;
+struct intel_crtc_state;
+struct intel_digital_port;
+struct intel_encoder;
+struct intel_lspcon;
+
+bool lspcon_init(struct intel_digital_port *dig_port);
+void lspcon_detect_hdr_capability(struct intel_lspcon *lspcon);
+void lspcon_resume(struct intel_digital_port *dig_port);
+void lspcon_wait_pcon_mode(struct intel_lspcon *lspcon);
+void lspcon_write_infoframe(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state,
+			    unsigned int type,
+			    const void *buf, ssize_t len);
+void lspcon_read_infoframe(struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state,
+			   unsigned int type,
+			   void *frame, ssize_t len);
+void lspcon_set_infoframes(struct intel_encoder *encoder,
+			   bool enable,
+			   const struct intel_crtc_state *crtc_state,
+			   const struct drm_connector_state *conn_state);
+u32 lspcon_infoframes_enabled(struct intel_encoder *encoder,
+			      const struct intel_crtc_state *pipe_config);
+u32 intel_lspcon_infoframes_enabled(struct intel_encoder *encoder,
+				    const struct intel_crtc_state *pipe_config);
+void hsw_write_infoframe(struct intel_encoder *encoder,
+			 const struct intel_crtc_state *crtc_state,
+			 unsigned int type,
+			 const void *frame, ssize_t len);
+void hsw_read_infoframe(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state,
+			unsigned int type,
+			void *frame, ssize_t len);
+
+#endif /* __INTEL_LSPCON_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_lvds.c b/drivers/gpu/drm/i915/display/intel_lvds.c
new file mode 100644
index 000000000..40b5d3d3c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_lvds.c
@@ -0,0 +1,1019 @@
+/*
+ * Copyright © 2006-2007 Intel Corporation
+ * Copyright (c) 2006 Dave Airlie <airlied@linux.ie>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Eric Anholt <eric@anholt.net>
+ *      Dave Airlie <airlied@linux.ie>
+ *      Jesse Barnes <jesse.barnes@intel.com>
+ */
+
+#include <acpi/button.h>
+#include <linux/acpi.h>
+#include <linux/dmi.h>
+#include <linux/i2c.h>
+#include <linux/slab.h>
+#include <linux/vga_switcheroo.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_edid.h>
+
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_backlight.h"
+#include "intel_connector.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dpll.h"
+#include "intel_fdi.h"
+#include "intel_gmbus.h"
+#include "intel_lvds.h"
+#include "intel_panel.h"
+
+/* Private structure for the integrated LVDS support */
+struct intel_lvds_pps {
+	/* 100us units */
+	int t1_t2;
+	int t3;
+	int t4;
+	int t5;
+	int tx;
+
+	int divider;
+
+	int port;
+	bool powerdown_on_reset;
+};
+
+struct intel_lvds_encoder {
+	struct intel_encoder base;
+
+	bool is_dual_link;
+	i915_reg_t reg;
+	u32 a3_power;
+
+	struct intel_lvds_pps init_pps;
+	u32 init_lvds_val;
+
+	struct intel_connector *attached_connector;
+};
+
+static struct intel_lvds_encoder *to_lvds_encoder(struct drm_encoder *encoder)
+{
+	return container_of(encoder, struct intel_lvds_encoder, base.base);
+}
+
+bool intel_lvds_port_enabled(struct drm_i915_private *dev_priv,
+			     i915_reg_t lvds_reg, enum pipe *pipe)
+{
+	u32 val;
+
+	val = intel_de_read(dev_priv, lvds_reg);
+
+	/* asserts want to know the pipe even if the port is disabled */
+	if (HAS_PCH_CPT(dev_priv))
+		*pipe = REG_FIELD_GET(LVDS_PIPE_SEL_MASK_CPT, val);
+	else
+		*pipe = REG_FIELD_GET(LVDS_PIPE_SEL_MASK, val);
+
+	return val & LVDS_PORT_EN;
+}
+
+static bool intel_lvds_get_hw_state(struct intel_encoder *encoder,
+				    enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
+	intel_wakeref_t wakeref;
+	bool ret;
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     encoder->power_domain);
+	if (!wakeref)
+		return false;
+
+	ret = intel_lvds_port_enabled(dev_priv, lvds_encoder->reg, pipe);
+
+	intel_display_power_put(dev_priv, encoder->power_domain, wakeref);
+
+	return ret;
+}
+
+static void intel_lvds_get_config(struct intel_encoder *encoder,
+				  struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
+	u32 tmp, flags = 0;
+
+	pipe_config->output_types |= BIT(INTEL_OUTPUT_LVDS);
+
+	tmp = intel_de_read(dev_priv, lvds_encoder->reg);
+	if (tmp & LVDS_HSYNC_POLARITY)
+		flags |= DRM_MODE_FLAG_NHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	if (tmp & LVDS_VSYNC_POLARITY)
+		flags |= DRM_MODE_FLAG_NVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_PVSYNC;
+
+	pipe_config->hw.adjusted_mode.flags |= flags;
+
+	if (DISPLAY_VER(dev_priv) < 5)
+		pipe_config->gmch_pfit.lvds_border_bits =
+			tmp & LVDS_BORDER_ENABLE;
+
+	/* gen2/3 store dither state in pfit control, needs to match */
+	if (DISPLAY_VER(dev_priv) < 4) {
+		tmp = intel_de_read(dev_priv, PFIT_CONTROL);
+
+		pipe_config->gmch_pfit.control |= tmp & PANEL_8TO6_DITHER_ENABLE;
+	}
+
+	pipe_config->hw.adjusted_mode.crtc_clock = pipe_config->port_clock;
+}
+
+static void intel_lvds_pps_get_hw_state(struct drm_i915_private *dev_priv,
+					struct intel_lvds_pps *pps)
+{
+	u32 val;
+
+	pps->powerdown_on_reset = intel_de_read(dev_priv, PP_CONTROL(0)) & PANEL_POWER_RESET;
+
+	val = intel_de_read(dev_priv, PP_ON_DELAYS(0));
+	pps->port = REG_FIELD_GET(PANEL_PORT_SELECT_MASK, val);
+	pps->t1_t2 = REG_FIELD_GET(PANEL_POWER_UP_DELAY_MASK, val);
+	pps->t5 = REG_FIELD_GET(PANEL_LIGHT_ON_DELAY_MASK, val);
+
+	val = intel_de_read(dev_priv, PP_OFF_DELAYS(0));
+	pps->t3 = REG_FIELD_GET(PANEL_POWER_DOWN_DELAY_MASK, val);
+	pps->tx = REG_FIELD_GET(PANEL_LIGHT_OFF_DELAY_MASK, val);
+
+	val = intel_de_read(dev_priv, PP_DIVISOR(0));
+	pps->divider = REG_FIELD_GET(PP_REFERENCE_DIVIDER_MASK, val);
+	val = REG_FIELD_GET(PANEL_POWER_CYCLE_DELAY_MASK, val);
+	/*
+	 * Remove the BSpec specified +1 (100ms) offset that accounts for a
+	 * too short power-cycle delay due to the asynchronous programming of
+	 * the register.
+	 */
+	if (val)
+		val--;
+	/* Convert from 100ms to 100us units */
+	pps->t4 = val * 1000;
+
+	if (DISPLAY_VER(dev_priv) <= 4 &&
+	    pps->t1_t2 == 0 && pps->t5 == 0 && pps->t3 == 0 && pps->tx == 0) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Panel power timings uninitialized, "
+			    "setting defaults\n");
+		/* Set T2 to 40ms and T5 to 200ms in 100 usec units */
+		pps->t1_t2 = 40 * 10;
+		pps->t5 = 200 * 10;
+		/* Set T3 to 35ms and Tx to 200ms in 100 usec units */
+		pps->t3 = 35 * 10;
+		pps->tx = 200 * 10;
+	}
+
+	drm_dbg(&dev_priv->drm, "LVDS PPS:t1+t2 %d t3 %d t4 %d t5 %d tx %d "
+		"divider %d port %d powerdown_on_reset %d\n",
+		pps->t1_t2, pps->t3, pps->t4, pps->t5, pps->tx,
+		pps->divider, pps->port, pps->powerdown_on_reset);
+}
+
+static void intel_lvds_pps_init_hw(struct drm_i915_private *dev_priv,
+				   struct intel_lvds_pps *pps)
+{
+	u32 val;
+
+	val = intel_de_read(dev_priv, PP_CONTROL(0));
+	drm_WARN_ON(&dev_priv->drm,
+		    (val & PANEL_UNLOCK_MASK) != PANEL_UNLOCK_REGS);
+	if (pps->powerdown_on_reset)
+		val |= PANEL_POWER_RESET;
+	intel_de_write(dev_priv, PP_CONTROL(0), val);
+
+	intel_de_write(dev_priv, PP_ON_DELAYS(0),
+		       REG_FIELD_PREP(PANEL_PORT_SELECT_MASK, pps->port) | REG_FIELD_PREP(PANEL_POWER_UP_DELAY_MASK, pps->t1_t2) | REG_FIELD_PREP(PANEL_LIGHT_ON_DELAY_MASK, pps->t5));
+
+	intel_de_write(dev_priv, PP_OFF_DELAYS(0),
+		       REG_FIELD_PREP(PANEL_POWER_DOWN_DELAY_MASK, pps->t3) | REG_FIELD_PREP(PANEL_LIGHT_OFF_DELAY_MASK, pps->tx));
+
+	intel_de_write(dev_priv, PP_DIVISOR(0),
+		       REG_FIELD_PREP(PP_REFERENCE_DIVIDER_MASK, pps->divider) | REG_FIELD_PREP(PANEL_POWER_CYCLE_DELAY_MASK, DIV_ROUND_UP(pps->t4, 1000) + 1));
+}
+
+static void intel_pre_enable_lvds(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *pipe_config,
+				  const struct drm_connector_state *conn_state)
+{
+	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	const struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	enum pipe pipe = crtc->pipe;
+	u32 temp;
+
+	if (HAS_PCH_SPLIT(dev_priv)) {
+		assert_fdi_rx_pll_disabled(dev_priv, pipe);
+		assert_shared_dpll_disabled(dev_priv,
+					    pipe_config->shared_dpll);
+	} else {
+		assert_pll_disabled(dev_priv, pipe);
+	}
+
+	intel_lvds_pps_init_hw(dev_priv, &lvds_encoder->init_pps);
+
+	temp = lvds_encoder->init_lvds_val;
+	temp |= LVDS_PORT_EN | LVDS_A0A2_CLKA_POWER_UP;
+
+	if (HAS_PCH_CPT(dev_priv)) {
+		temp &= ~LVDS_PIPE_SEL_MASK_CPT;
+		temp |= LVDS_PIPE_SEL_CPT(pipe);
+	} else {
+		temp &= ~LVDS_PIPE_SEL_MASK;
+		temp |= LVDS_PIPE_SEL(pipe);
+	}
+
+	/* set the corresponsding LVDS_BORDER bit */
+	temp &= ~LVDS_BORDER_ENABLE;
+	temp |= pipe_config->gmch_pfit.lvds_border_bits;
+
+	/*
+	 * Set the B0-B3 data pairs corresponding to whether we're going to
+	 * set the DPLLs for dual-channel mode or not.
+	 */
+	if (lvds_encoder->is_dual_link)
+		temp |= LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP;
+	else
+		temp &= ~(LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP);
+
+	/*
+	 * It would be nice to set 24 vs 18-bit mode (LVDS_A3_POWER_UP)
+	 * appropriately here, but we need to look more thoroughly into how
+	 * panels behave in the two modes. For now, let's just maintain the
+	 * value we got from the BIOS.
+	 */
+	temp &= ~LVDS_A3_POWER_MASK;
+	temp |= lvds_encoder->a3_power;
+
+	/*
+	 * Set the dithering flag on LVDS as needed, note that there is no
+	 * special lvds dither control bit on pch-split platforms, dithering is
+	 * only controlled through the PIPECONF reg.
+	 */
+	if (DISPLAY_VER(dev_priv) == 4) {
+		/*
+		 * Bspec wording suggests that LVDS port dithering only exists
+		 * for 18bpp panels.
+		 */
+		if (pipe_config->dither && pipe_config->pipe_bpp == 18)
+			temp |= LVDS_ENABLE_DITHER;
+		else
+			temp &= ~LVDS_ENABLE_DITHER;
+	}
+	temp &= ~(LVDS_HSYNC_POLARITY | LVDS_VSYNC_POLARITY);
+	if (adjusted_mode->flags & DRM_MODE_FLAG_NHSYNC)
+		temp |= LVDS_HSYNC_POLARITY;
+	if (adjusted_mode->flags & DRM_MODE_FLAG_NVSYNC)
+		temp |= LVDS_VSYNC_POLARITY;
+
+	intel_de_write(dev_priv, lvds_encoder->reg, temp);
+}
+
+/*
+ * Sets the power state for the panel.
+ */
+static void intel_enable_lvds(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *pipe_config,
+			      const struct drm_connector_state *conn_state)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	intel_de_write(dev_priv, lvds_encoder->reg,
+		       intel_de_read(dev_priv, lvds_encoder->reg) | LVDS_PORT_EN);
+
+	intel_de_write(dev_priv, PP_CONTROL(0),
+		       intel_de_read(dev_priv, PP_CONTROL(0)) | PANEL_POWER_ON);
+	intel_de_posting_read(dev_priv, lvds_encoder->reg);
+
+	if (intel_de_wait_for_set(dev_priv, PP_STATUS(0), PP_ON, 5000))
+		drm_err(&dev_priv->drm,
+			"timed out waiting for panel to power on\n");
+
+	intel_backlight_enable(pipe_config, conn_state);
+}
+
+static void intel_disable_lvds(struct intel_atomic_state *state,
+			       struct intel_encoder *encoder,
+			       const struct intel_crtc_state *old_crtc_state,
+			       const struct drm_connector_state *old_conn_state)
+{
+	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	intel_de_write(dev_priv, PP_CONTROL(0),
+		       intel_de_read(dev_priv, PP_CONTROL(0)) & ~PANEL_POWER_ON);
+	if (intel_de_wait_for_clear(dev_priv, PP_STATUS(0), PP_ON, 1000))
+		drm_err(&dev_priv->drm,
+			"timed out waiting for panel to power off\n");
+
+	intel_de_write(dev_priv, lvds_encoder->reg,
+		       intel_de_read(dev_priv, lvds_encoder->reg) & ~LVDS_PORT_EN);
+	intel_de_posting_read(dev_priv, lvds_encoder->reg);
+}
+
+static void gmch_disable_lvds(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *old_crtc_state,
+			      const struct drm_connector_state *old_conn_state)
+
+{
+	intel_backlight_disable(old_conn_state);
+
+	intel_disable_lvds(state, encoder, old_crtc_state, old_conn_state);
+}
+
+static void pch_disable_lvds(struct intel_atomic_state *state,
+			     struct intel_encoder *encoder,
+			     const struct intel_crtc_state *old_crtc_state,
+			     const struct drm_connector_state *old_conn_state)
+{
+	intel_backlight_disable(old_conn_state);
+}
+
+static void pch_post_disable_lvds(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *old_crtc_state,
+				  const struct drm_connector_state *old_conn_state)
+{
+	intel_disable_lvds(state, encoder, old_crtc_state, old_conn_state);
+}
+
+static void intel_lvds_shutdown(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (intel_de_wait_for_clear(dev_priv, PP_STATUS(0), PP_CYCLE_DELAY_ACTIVE, 5000))
+		drm_err(&dev_priv->drm,
+			"timed out waiting for panel power cycle delay\n");
+}
+
+static enum drm_mode_status
+intel_lvds_mode_valid(struct drm_connector *connector,
+		      struct drm_display_mode *mode)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+	struct drm_i915_private *i915 = to_i915(intel_connector->base.dev);
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_fixed_mode(intel_connector, mode);
+	int max_pixclk = to_i915(connector->dev)->max_dotclk_freq;
+	enum drm_mode_status status;
+
+	status = intel_cpu_transcoder_mode_valid(i915, mode);
+	if (status != MODE_OK)
+		return status;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return MODE_NO_DBLESCAN;
+
+	status = intel_panel_mode_valid(intel_connector, mode);
+	if (status != MODE_OK)
+		return status;
+
+	if (fixed_mode->clock > max_pixclk)
+		return MODE_CLOCK_HIGH;
+
+	return MODE_OK;
+}
+
+static int intel_lvds_compute_config(struct intel_encoder *intel_encoder,
+				     struct intel_crtc_state *pipe_config,
+				     struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
+	struct intel_lvds_encoder *lvds_encoder =
+		to_lvds_encoder(&intel_encoder->base);
+	struct intel_connector *intel_connector =
+		lvds_encoder->attached_connector;
+	struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	unsigned int lvds_bpp;
+	int ret;
+
+	/* Should never happen!! */
+	if (DISPLAY_VER(dev_priv) < 4 && crtc->pipe == 0) {
+		drm_err(&dev_priv->drm, "Can't support LVDS on pipe A\n");
+		return -EINVAL;
+	}
+
+	if (lvds_encoder->a3_power == LVDS_A3_POWER_UP)
+		lvds_bpp = 8*3;
+	else
+		lvds_bpp = 6*3;
+
+	if (lvds_bpp != pipe_config->pipe_bpp && !pipe_config->bw_constrained) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "forcing display bpp (was %d) to LVDS (%d)\n",
+			    pipe_config->pipe_bpp, lvds_bpp);
+		pipe_config->pipe_bpp = lvds_bpp;
+	}
+
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+
+	/*
+	 * We have timings from the BIOS for the panel, put them in
+	 * to the adjusted mode.  The CRTC will be set up for this mode,
+	 * with the panel scaling set up to source from the H/VDisplay
+	 * of the original mode.
+	 */
+	ret = intel_panel_compute_config(intel_connector, adjusted_mode);
+	if (ret)
+		return ret;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	if (HAS_PCH_SPLIT(dev_priv))
+		pipe_config->has_pch_encoder = true;
+
+	ret = intel_panel_fitting(pipe_config, conn_state);
+	if (ret)
+		return ret;
+
+	/*
+	 * XXX: It would be nice to support lower refresh rates on the
+	 * panels to reduce power consumption, and perhaps match the
+	 * user's requested refresh rate.
+	 */
+
+	return 0;
+}
+
+/*
+ * Return the list of DDC modes if available, or the BIOS fixed mode otherwise.
+ */
+static int intel_lvds_get_modes(struct drm_connector *connector)
+{
+	struct intel_connector *intel_connector = to_intel_connector(connector);
+
+	/* use cached edid if we have one */
+	if (!IS_ERR_OR_NULL(intel_connector->edid))
+		return drm_add_edid_modes(connector, intel_connector->edid);
+
+	return intel_panel_get_modes(intel_connector);
+}
+
+static const struct drm_connector_helper_funcs intel_lvds_connector_helper_funcs = {
+	.get_modes = intel_lvds_get_modes,
+	.mode_valid = intel_lvds_mode_valid,
+	.atomic_check = intel_digital_connector_atomic_check,
+};
+
+static const struct drm_connector_funcs intel_lvds_connector_funcs = {
+	.detect = intel_panel_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.atomic_get_property = intel_digital_connector_atomic_get_property,
+	.atomic_set_property = intel_digital_connector_atomic_set_property,
+	.late_register = intel_connector_register,
+	.early_unregister = intel_connector_unregister,
+	.destroy = intel_connector_destroy,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.atomic_duplicate_state = intel_digital_connector_duplicate_state,
+};
+
+static const struct drm_encoder_funcs intel_lvds_enc_funcs = {
+	.destroy = intel_encoder_destroy,
+};
+
+static int intel_no_lvds_dmi_callback(const struct dmi_system_id *id)
+{
+	DRM_INFO("Skipping LVDS initialization for %s\n", id->ident);
+	return 1;
+}
+
+/* These systems claim to have LVDS, but really don't */
+static const struct dmi_system_id intel_no_lvds[] = {
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Apple Mac Mini (Core series)",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Apple"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "Macmini1,1"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Apple Mac Mini (Core 2 series)",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Apple"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "Macmini2,1"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "MSI IM-945GSE-A",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "MSI"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "A9830IMS"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Dell Studio Hybrid",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "Studio Hybrid 140g"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Dell OptiPlex FX170",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex FX170"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "AOpen Mini PC",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "AOpen"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "i965GMx-IF"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "AOpen Mini PC MP915",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "AOpen"),
+			DMI_MATCH(DMI_BOARD_NAME, "i915GMx-F"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "AOpen i915GMm-HFS",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "AOpen"),
+			DMI_MATCH(DMI_BOARD_NAME, "i915GMm-HFS"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+                .ident = "AOpen i45GMx-I",
+                .matches = {
+                        DMI_MATCH(DMI_BOARD_VENDOR, "AOpen"),
+                        DMI_MATCH(DMI_BOARD_NAME, "i45GMx-I"),
+                },
+        },
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Aopen i945GTt-VFA",
+		.matches = {
+			DMI_MATCH(DMI_PRODUCT_VERSION, "AO00001JW"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Clientron U800",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Clientron"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "U800"),
+		},
+	},
+	{
+                .callback = intel_no_lvds_dmi_callback,
+                .ident = "Clientron E830",
+                .matches = {
+                        DMI_MATCH(DMI_SYS_VENDOR, "Clientron"),
+                        DMI_MATCH(DMI_PRODUCT_NAME, "E830"),
+                },
+        },
+        {
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Asus EeeBox PC EB1007",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "ASUSTeK Computer INC."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "EB1007"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Asus AT5NM10T-I",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."),
+			DMI_MATCH(DMI_BOARD_NAME, "AT5NM10T-I"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Hewlett-Packard HP t5740",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
+			DMI_MATCH(DMI_PRODUCT_NAME, " t5740"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Hewlett-Packard t5745",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "hp t5745"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Hewlett-Packard st5747",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "hp st5747"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "MSI Wind Box DC500",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "MICRO-STAR INTERNATIONAL CO., LTD"),
+			DMI_MATCH(DMI_BOARD_NAME, "MS-7469"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Gigabyte GA-D525TUD",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Gigabyte Technology Co., Ltd."),
+			DMI_MATCH(DMI_BOARD_NAME, "D525TUD"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Supermicro X7SPA-H",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Supermicro"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "X7SPA-H"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Fujitsu Esprimo Q900",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "FUJITSU"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "ESPRIMO Q900"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Intel D410PT",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Intel"),
+			DMI_MATCH(DMI_BOARD_NAME, "D410PT"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Intel D425KT",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Intel"),
+			DMI_EXACT_MATCH(DMI_BOARD_NAME, "D425KT"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Intel D510MO",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Intel"),
+			DMI_EXACT_MATCH(DMI_BOARD_NAME, "D510MO"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Intel D525MW",
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "Intel"),
+			DMI_EXACT_MATCH(DMI_BOARD_NAME, "D525MW"),
+		},
+	},
+	{
+		.callback = intel_no_lvds_dmi_callback,
+		.ident = "Radiant P845",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Radiant Systems Inc"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "P845"),
+		},
+	},
+
+	{ }	/* terminating entry */
+};
+
+static int intel_dual_link_lvds_callback(const struct dmi_system_id *id)
+{
+	DRM_INFO("Forcing lvds to dual link mode on %s\n", id->ident);
+	return 1;
+}
+
+static const struct dmi_system_id intel_dual_link_lvds[] = {
+	{
+		.callback = intel_dual_link_lvds_callback,
+		.ident = "Apple MacBook Pro 15\" (2010)",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Apple Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "MacBookPro6,2"),
+		},
+	},
+	{
+		.callback = intel_dual_link_lvds_callback,
+		.ident = "Apple MacBook Pro 15\" (2011)",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Apple Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "MacBookPro8,2"),
+		},
+	},
+	{
+		.callback = intel_dual_link_lvds_callback,
+		.ident = "Apple MacBook Pro 15\" (2012)",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Apple Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "MacBookPro9,1"),
+		},
+	},
+	{ }	/* terminating entry */
+};
+
+struct intel_encoder *intel_get_lvds_encoder(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		if (encoder->type == INTEL_OUTPUT_LVDS)
+			return encoder;
+	}
+
+	return NULL;
+}
+
+bool intel_is_dual_link_lvds(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder = intel_get_lvds_encoder(dev_priv);
+
+	return encoder && to_lvds_encoder(&encoder->base)->is_dual_link;
+}
+
+static bool compute_is_dual_link_lvds(struct intel_lvds_encoder *lvds_encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(lvds_encoder->base.base.dev);
+	struct intel_connector *connector = lvds_encoder->attached_connector;
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_preferred_fixed_mode(connector);
+	unsigned int val;
+
+	/* use the module option value if specified */
+	if (dev_priv->params.lvds_channel_mode > 0)
+		return dev_priv->params.lvds_channel_mode == 2;
+
+	/* single channel LVDS is limited to 112 MHz */
+	if (fixed_mode->clock > 112999)
+		return true;
+
+	if (dmi_check_system(intel_dual_link_lvds))
+		return true;
+
+	/*
+	 * BIOS should set the proper LVDS register value at boot, but
+	 * in reality, it doesn't set the value when the lid is closed;
+	 * we need to check "the value to be set" in VBT when LVDS
+	 * register is uninitialized.
+	 */
+	val = intel_de_read(dev_priv, lvds_encoder->reg);
+	if (HAS_PCH_CPT(dev_priv))
+		val &= ~(LVDS_DETECTED | LVDS_PIPE_SEL_MASK_CPT);
+	else
+		val &= ~(LVDS_DETECTED | LVDS_PIPE_SEL_MASK);
+	if (val == 0)
+		val = connector->panel.vbt.bios_lvds_val;
+
+	return (val & LVDS_CLKB_POWER_MASK) == LVDS_CLKB_POWER_UP;
+}
+
+/**
+ * intel_lvds_init - setup LVDS connectors on this device
+ * @dev_priv: i915 device
+ *
+ * Create the connector, register the LVDS DDC bus, and try to figure out what
+ * modes we can display on the LVDS panel (if present).
+ */
+void intel_lvds_init(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_lvds_encoder *lvds_encoder;
+	struct intel_encoder *intel_encoder;
+	struct intel_connector *intel_connector;
+	struct drm_connector *connector;
+	struct drm_encoder *encoder;
+	struct edid *edid;
+	i915_reg_t lvds_reg;
+	u32 lvds;
+	u8 pin;
+	u32 allowed_scalers;
+
+	/* Skip init on machines we know falsely report LVDS */
+	if (dmi_check_system(intel_no_lvds)) {
+		drm_WARN(dev, !dev_priv->display.vbt.int_lvds_support,
+			 "Useless DMI match. Internal LVDS support disabled by VBT\n");
+		return;
+	}
+
+	if (!dev_priv->display.vbt.int_lvds_support) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Internal LVDS support disabled by VBT\n");
+		return;
+	}
+
+	if (HAS_PCH_SPLIT(dev_priv))
+		lvds_reg = PCH_LVDS;
+	else
+		lvds_reg = LVDS;
+
+	lvds = intel_de_read(dev_priv, lvds_reg);
+
+	if (HAS_PCH_SPLIT(dev_priv)) {
+		if ((lvds & LVDS_DETECTED) == 0)
+			return;
+	}
+
+	pin = GMBUS_PIN_PANEL;
+	if (!intel_bios_is_lvds_present(dev_priv, &pin)) {
+		if ((lvds & LVDS_PORT_EN) == 0) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "LVDS is not present in VBT\n");
+			return;
+		}
+		drm_dbg_kms(&dev_priv->drm,
+			    "LVDS is not present in VBT, but enabled anyway\n");
+	}
+
+	lvds_encoder = kzalloc(sizeof(*lvds_encoder), GFP_KERNEL);
+	if (!lvds_encoder)
+		return;
+
+	intel_connector = intel_connector_alloc();
+	if (!intel_connector) {
+		kfree(lvds_encoder);
+		return;
+	}
+
+	lvds_encoder->attached_connector = intel_connector;
+
+	intel_encoder = &lvds_encoder->base;
+	encoder = &intel_encoder->base;
+	connector = &intel_connector->base;
+	drm_connector_init(dev, &intel_connector->base, &intel_lvds_connector_funcs,
+			   DRM_MODE_CONNECTOR_LVDS);
+
+	drm_encoder_init(dev, &intel_encoder->base, &intel_lvds_enc_funcs,
+			 DRM_MODE_ENCODER_LVDS, "LVDS");
+
+	intel_encoder->enable = intel_enable_lvds;
+	intel_encoder->pre_enable = intel_pre_enable_lvds;
+	intel_encoder->compute_config = intel_lvds_compute_config;
+	if (HAS_PCH_SPLIT(dev_priv)) {
+		intel_encoder->disable = pch_disable_lvds;
+		intel_encoder->post_disable = pch_post_disable_lvds;
+	} else {
+		intel_encoder->disable = gmch_disable_lvds;
+	}
+	intel_encoder->get_hw_state = intel_lvds_get_hw_state;
+	intel_encoder->get_config = intel_lvds_get_config;
+	intel_encoder->update_pipe = intel_backlight_update;
+	intel_encoder->shutdown = intel_lvds_shutdown;
+	intel_connector->get_hw_state = intel_connector_get_hw_state;
+
+	intel_connector_attach_encoder(intel_connector, intel_encoder);
+
+	intel_encoder->type = INTEL_OUTPUT_LVDS;
+	intel_encoder->power_domain = POWER_DOMAIN_PORT_OTHER;
+	intel_encoder->port = PORT_NONE;
+	intel_encoder->cloneable = 0;
+	if (DISPLAY_VER(dev_priv) < 4)
+		intel_encoder->pipe_mask = BIT(PIPE_B);
+	else
+		intel_encoder->pipe_mask = ~0;
+
+	drm_connector_helper_add(connector, &intel_lvds_connector_helper_funcs);
+	connector->display_info.subpixel_order = SubPixelHorizontalRGB;
+	connector->interlace_allowed = false;
+	connector->doublescan_allowed = false;
+
+	lvds_encoder->reg = lvds_reg;
+
+	/* create the scaling mode property */
+	allowed_scalers = BIT(DRM_MODE_SCALE_ASPECT);
+	allowed_scalers |= BIT(DRM_MODE_SCALE_FULLSCREEN);
+	allowed_scalers |= BIT(DRM_MODE_SCALE_CENTER);
+	drm_connector_attach_scaling_mode_property(connector, allowed_scalers);
+	connector->state->scaling_mode = DRM_MODE_SCALE_ASPECT;
+
+	intel_lvds_pps_get_hw_state(dev_priv, &lvds_encoder->init_pps);
+	lvds_encoder->init_lvds_val = lvds;
+
+	/*
+	 * LVDS discovery:
+	 * 1) check for EDID on DDC
+	 * 2) check for VBT data
+	 * 3) check to see if LVDS is already on
+	 *    if none of the above, no panel
+	 */
+
+	/*
+	 * Attempt to get the fixed panel mode from DDC.  Assume that the
+	 * preferred mode is the right one.
+	 */
+	mutex_lock(&dev->mode_config.mutex);
+	if (vga_switcheroo_handler_flags() & VGA_SWITCHEROO_CAN_SWITCH_DDC)
+		edid = drm_get_edid_switcheroo(connector,
+				    intel_gmbus_get_adapter(dev_priv, pin));
+	else
+		edid = drm_get_edid(connector,
+				    intel_gmbus_get_adapter(dev_priv, pin));
+	if (edid) {
+		if (drm_add_edid_modes(connector, edid)) {
+			drm_connector_update_edid_property(connector,
+								edid);
+		} else {
+			kfree(edid);
+			edid = ERR_PTR(-EINVAL);
+		}
+	} else {
+		edid = ERR_PTR(-ENOENT);
+	}
+	intel_connector->edid = edid;
+
+	intel_bios_init_panel_late(dev_priv, &intel_connector->panel, NULL,
+				   IS_ERR(edid) ? NULL : edid);
+
+	/* Try EDID first */
+	intel_panel_add_edid_fixed_modes(intel_connector,
+					 intel_connector->panel.vbt.drrs_type != DRRS_TYPE_NONE);
+
+	/* Failed to get EDID, what about VBT? */
+	if (!intel_panel_preferred_fixed_mode(intel_connector))
+		intel_panel_add_vbt_lfp_fixed_mode(intel_connector);
+
+	/*
+	 * If we didn't get a fixed mode from EDID or VBT, try checking
+	 * if the panel is already turned on.  If so, assume that
+	 * whatever is currently programmed is the correct mode.
+	 */
+	if (!intel_panel_preferred_fixed_mode(intel_connector))
+		intel_panel_add_encoder_fixed_mode(intel_connector, intel_encoder);
+
+	mutex_unlock(&dev->mode_config.mutex);
+
+	/* If we still don't have a mode after all that, give up. */
+	if (!intel_panel_preferred_fixed_mode(intel_connector))
+		goto failed;
+
+	intel_panel_init(intel_connector);
+
+	intel_backlight_setup(intel_connector, INVALID_PIPE);
+
+	lvds_encoder->is_dual_link = compute_is_dual_link_lvds(lvds_encoder);
+	drm_dbg_kms(&dev_priv->drm, "detected %s-link lvds configuration\n",
+		    lvds_encoder->is_dual_link ? "dual" : "single");
+
+	lvds_encoder->a3_power = lvds & LVDS_A3_POWER_MASK;
+
+	return;
+
+failed:
+	drm_dbg_kms(&dev_priv->drm, "No LVDS modes found, disabling.\n");
+	drm_connector_cleanup(connector);
+	drm_encoder_cleanup(encoder);
+	kfree(lvds_encoder);
+	intel_connector_free(intel_connector);
+	return;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_lvds.h b/drivers/gpu/drm/i915/display/intel_lvds.h
new file mode 100644
index 000000000..9d3372dc5
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_lvds.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_LVDS_H__
+#define __INTEL_LVDS_H__
+
+#include <linux/types.h>
+
+#include "i915_reg_defs.h"
+
+enum pipe;
+struct drm_i915_private;
+
+bool intel_lvds_port_enabled(struct drm_i915_private *dev_priv,
+			     i915_reg_t lvds_reg, enum pipe *pipe);
+void intel_lvds_init(struct drm_i915_private *dev_priv);
+struct intel_encoder *intel_get_lvds_encoder(struct drm_i915_private *dev_priv);
+bool intel_is_dual_link_lvds(struct drm_i915_private *dev_priv);
+
+#endif /* __INTEL_LVDS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_modeset_setup.c b/drivers/gpu/drm/i915/display/intel_modeset_setup.c
new file mode 100644
index 000000000..cbfabd58b
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_modeset_setup.c
@@ -0,0 +1,735 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ *
+ * Read out the current hardware modeset state, and sanitize it to the current
+ * state.
+ */
+
+#include <drm/drm_atomic_uapi.h>
+#include <drm/drm_atomic_state_helper.h>
+
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_bw.h"
+#include "intel_color.h"
+#include "intel_crtc.h"
+#include "intel_crtc_state_dump.h"
+#include "intel_ddi.h"
+#include "intel_de.h"
+#include "intel_display.h"
+#include "intel_display_power.h"
+#include "intel_display_types.h"
+#include "intel_modeset_setup.h"
+#include "intel_pch_display.h"
+#include "intel_pm.h"
+#include "skl_watermark.h"
+
+static void intel_crtc_disable_noatomic(struct intel_crtc *crtc,
+					struct drm_modeset_acquire_ctx *ctx)
+{
+	struct intel_encoder *encoder;
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_bw_state *bw_state =
+		to_intel_bw_state(i915->display.bw.obj.state);
+	struct intel_cdclk_state *cdclk_state =
+		to_intel_cdclk_state(i915->display.cdclk.obj.state);
+	struct intel_dbuf_state *dbuf_state =
+		to_intel_dbuf_state(i915->display.dbuf.obj.state);
+	struct intel_crtc_state *crtc_state =
+		to_intel_crtc_state(crtc->base.state);
+	struct intel_plane *plane;
+	struct drm_atomic_state *state;
+	struct intel_crtc_state *temp_crtc_state;
+	enum pipe pipe = crtc->pipe;
+	int ret;
+
+	if (!crtc_state->hw.active)
+		return;
+
+	for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) {
+		const struct intel_plane_state *plane_state =
+			to_intel_plane_state(plane->base.state);
+
+		if (plane_state->uapi.visible)
+			intel_plane_disable_noatomic(crtc, plane);
+	}
+
+	state = drm_atomic_state_alloc(&i915->drm);
+	if (!state) {
+		drm_dbg_kms(&i915->drm,
+			    "failed to disable [CRTC:%d:%s], out of memory",
+			    crtc->base.base.id, crtc->base.name);
+		return;
+	}
+
+	state->acquire_ctx = ctx;
+
+	/* Everything's already locked, -EDEADLK can't happen. */
+	temp_crtc_state = intel_atomic_get_crtc_state(state, crtc);
+	ret = drm_atomic_add_affected_connectors(state, &crtc->base);
+
+	drm_WARN_ON(&i915->drm, IS_ERR(temp_crtc_state) || ret);
+
+	i915->display.funcs.display->crtc_disable(to_intel_atomic_state(state), crtc);
+
+	drm_atomic_state_put(state);
+
+	drm_dbg_kms(&i915->drm,
+		    "[CRTC:%d:%s] hw state adjusted, was enabled, now disabled\n",
+		    crtc->base.base.id, crtc->base.name);
+
+	crtc->active = false;
+	crtc->base.enabled = false;
+
+	drm_WARN_ON(&i915->drm,
+		    drm_atomic_set_mode_for_crtc(&crtc_state->uapi, NULL) < 0);
+	crtc_state->uapi.active = false;
+	crtc_state->uapi.connector_mask = 0;
+	crtc_state->uapi.encoder_mask = 0;
+	intel_crtc_free_hw_state(crtc_state);
+	memset(&crtc_state->hw, 0, sizeof(crtc_state->hw));
+
+	for_each_encoder_on_crtc(&i915->drm, &crtc->base, encoder)
+		encoder->base.crtc = NULL;
+
+	intel_fbc_disable(crtc);
+	intel_update_watermarks(i915);
+	intel_disable_shared_dpll(crtc_state);
+
+	intel_display_power_put_all_in_set(i915, &crtc->enabled_power_domains);
+
+	cdclk_state->min_cdclk[pipe] = 0;
+	cdclk_state->min_voltage_level[pipe] = 0;
+	cdclk_state->active_pipes &= ~BIT(pipe);
+
+	dbuf_state->active_pipes &= ~BIT(pipe);
+
+	bw_state->data_rate[pipe] = 0;
+	bw_state->num_active_planes[pipe] = 0;
+}
+
+static void intel_modeset_update_connector_atomic_state(struct drm_i915_private *i915)
+{
+	struct intel_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+
+	drm_connector_list_iter_begin(&i915->drm, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		struct drm_connector_state *conn_state = connector->base.state;
+		struct intel_encoder *encoder =
+			to_intel_encoder(connector->base.encoder);
+
+		if (conn_state->crtc)
+			drm_connector_put(&connector->base);
+
+		if (encoder) {
+			struct intel_crtc *crtc =
+				to_intel_crtc(encoder->base.crtc);
+			const struct intel_crtc_state *crtc_state =
+				to_intel_crtc_state(crtc->base.state);
+
+			conn_state->best_encoder = &encoder->base;
+			conn_state->crtc = &crtc->base;
+			conn_state->max_bpc = (crtc_state->pipe_bpp ?: 24) / 3;
+
+			drm_connector_get(&connector->base);
+		} else {
+			conn_state->best_encoder = NULL;
+			conn_state->crtc = NULL;
+		}
+	}
+	drm_connector_list_iter_end(&conn_iter);
+}
+
+static void intel_crtc_copy_hw_to_uapi_state(struct intel_crtc_state *crtc_state)
+{
+	if (intel_crtc_is_bigjoiner_slave(crtc_state))
+		return;
+
+	crtc_state->uapi.enable = crtc_state->hw.enable;
+	crtc_state->uapi.active = crtc_state->hw.active;
+	drm_WARN_ON(crtc_state->uapi.crtc->dev,
+		    drm_atomic_set_mode_for_crtc(&crtc_state->uapi, &crtc_state->hw.mode) < 0);
+
+	crtc_state->uapi.adjusted_mode = crtc_state->hw.adjusted_mode;
+	crtc_state->uapi.scaling_filter = crtc_state->hw.scaling_filter;
+
+	drm_property_replace_blob(&crtc_state->uapi.degamma_lut,
+				  crtc_state->hw.degamma_lut);
+	drm_property_replace_blob(&crtc_state->uapi.gamma_lut,
+				  crtc_state->hw.gamma_lut);
+	drm_property_replace_blob(&crtc_state->uapi.ctm,
+				  crtc_state->hw.ctm);
+}
+
+static void
+intel_sanitize_plane_mapping(struct drm_i915_private *i915)
+{
+	struct intel_crtc *crtc;
+
+	if (DISPLAY_VER(i915) >= 4)
+		return;
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_plane *plane =
+			to_intel_plane(crtc->base.primary);
+		struct intel_crtc *plane_crtc;
+		enum pipe pipe;
+
+		if (!plane->get_hw_state(plane, &pipe))
+			continue;
+
+		if (pipe == crtc->pipe)
+			continue;
+
+		drm_dbg_kms(&i915->drm,
+			    "[PLANE:%d:%s] attached to the wrong pipe, disabling plane\n",
+			    plane->base.base.id, plane->base.name);
+
+		plane_crtc = intel_crtc_for_pipe(i915, pipe);
+		intel_plane_disable_noatomic(plane_crtc, plane);
+	}
+}
+
+static bool intel_crtc_has_encoders(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct intel_encoder *encoder;
+
+	for_each_encoder_on_crtc(dev, &crtc->base, encoder)
+		return true;
+
+	return false;
+}
+
+static struct intel_connector *intel_encoder_find_connector(struct intel_encoder *encoder)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct intel_connector *connector;
+
+	for_each_connector_on_encoder(dev, &encoder->base, connector)
+		return connector;
+
+	return NULL;
+}
+
+static void intel_sanitize_fifo_underrun_reporting(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	if (!crtc_state->hw.active && !HAS_GMCH(i915))
+		return;
+
+	/*
+	 * We start out with underrun reporting disabled to avoid races.
+	 * For correct bookkeeping mark this on active crtcs.
+	 *
+	 * Also on gmch platforms we dont have any hardware bits to
+	 * disable the underrun reporting. Which means we need to start
+	 * out with underrun reporting disabled also on inactive pipes,
+	 * since otherwise we'll complain about the garbage we read when
+	 * e.g. coming up after runtime pm.
+	 *
+	 * No protection against concurrent access is required - at
+	 * worst a fifo underrun happens which also sets this to false.
+	 */
+	crtc->cpu_fifo_underrun_disabled = true;
+
+	/*
+	 * We track the PCH trancoder underrun reporting state
+	 * within the crtc. With crtc for pipe A housing the underrun
+	 * reporting state for PCH transcoder A, crtc for pipe B housing
+	 * it for PCH transcoder B, etc. LPT-H has only PCH transcoder A,
+	 * and marking underrun reporting as disabled for the non-existing
+	 * PCH transcoders B and C would prevent enabling the south
+	 * error interrupt (see cpt_can_enable_serr_int()).
+	 */
+	if (intel_has_pch_trancoder(i915, crtc->pipe))
+		crtc->pch_fifo_underrun_disabled = true;
+}
+
+static void intel_sanitize_crtc(struct intel_crtc *crtc,
+				struct drm_modeset_acquire_ctx *ctx)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state);
+
+	if (crtc_state->hw.active) {
+		struct intel_plane *plane;
+
+		/* Disable everything but the primary plane */
+		for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) {
+			const struct intel_plane_state *plane_state =
+				to_intel_plane_state(plane->base.state);
+
+			if (plane_state->uapi.visible &&
+			    plane->base.type != DRM_PLANE_TYPE_PRIMARY)
+				intel_plane_disable_noatomic(crtc, plane);
+		}
+
+		/* Disable any background color/etc. set by the BIOS */
+		intel_color_commit_noarm(crtc_state);
+		intel_color_commit_arm(crtc_state);
+	}
+
+	/*
+	 * Adjust the state of the output pipe according to whether we have
+	 * active connectors/encoders.
+	 */
+	if (crtc_state->hw.active && !intel_crtc_has_encoders(crtc) &&
+	    !intel_crtc_is_bigjoiner_slave(crtc_state))
+		intel_crtc_disable_noatomic(crtc, ctx);
+}
+
+static bool has_bogus_dpll_config(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+
+	/*
+	 * Some SNB BIOSen (eg. ASUS K53SV) are known to misprogram
+	 * the hardware when a high res displays plugged in. DPLL P
+	 * divider is zero, and the pipe timings are bonkers. We'll
+	 * try to disable everything in that case.
+	 *
+	 * FIXME would be nice to be able to sanitize this state
+	 * without several WARNs, but for now let's take the easy
+	 * road.
+	 */
+	return IS_SANDYBRIDGE(i915) &&
+		crtc_state->hw.active &&
+		crtc_state->shared_dpll &&
+		crtc_state->port_clock == 0;
+}
+
+static void intel_sanitize_encoder(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_connector *connector;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
+	struct intel_crtc_state *crtc_state = crtc ?
+		to_intel_crtc_state(crtc->base.state) : NULL;
+
+	/*
+	 * We need to check both for a crtc link (meaning that the encoder is
+	 * active and trying to read from a pipe) and the pipe itself being
+	 * active.
+	 */
+	bool has_active_crtc = crtc_state &&
+		crtc_state->hw.active;
+
+	if (crtc_state && has_bogus_dpll_config(crtc_state)) {
+		drm_dbg_kms(&i915->drm,
+			    "BIOS has misprogrammed the hardware. Disabling pipe %c\n",
+			    pipe_name(crtc->pipe));
+		has_active_crtc = false;
+	}
+
+	connector = intel_encoder_find_connector(encoder);
+	if (connector && !has_active_crtc) {
+		drm_dbg_kms(&i915->drm,
+			    "[ENCODER:%d:%s] has active connectors but no active pipe!\n",
+			    encoder->base.base.id,
+			    encoder->base.name);
+
+		/*
+		 * Connector is active, but has no active pipe. This is fallout
+		 * from our resume register restoring. Disable the encoder
+		 * manually again.
+		 */
+		if (crtc_state) {
+			struct drm_encoder *best_encoder;
+
+			drm_dbg_kms(&i915->drm,
+				    "[ENCODER:%d:%s] manually disabled\n",
+				    encoder->base.base.id,
+				    encoder->base.name);
+
+			/* avoid oopsing in case the hooks consult best_encoder */
+			best_encoder = connector->base.state->best_encoder;
+			connector->base.state->best_encoder = &encoder->base;
+
+			/* FIXME NULL atomic state passed! */
+			if (encoder->disable)
+				encoder->disable(NULL, encoder, crtc_state,
+						 connector->base.state);
+			if (encoder->post_disable)
+				encoder->post_disable(NULL, encoder, crtc_state,
+						      connector->base.state);
+
+			connector->base.state->best_encoder = best_encoder;
+		}
+		encoder->base.crtc = NULL;
+
+		/*
+		 * Inconsistent output/port/pipe state happens presumably due to
+		 * a bug in one of the get_hw_state functions. Or someplace else
+		 * in our code, like the register restore mess on resume. Clamp
+		 * things to off as a safer default.
+		 */
+		connector->base.dpms = DRM_MODE_DPMS_OFF;
+		connector->base.encoder = NULL;
+	}
+
+	/* notify opregion of the sanitized encoder state */
+	intel_opregion_notify_encoder(encoder, connector && has_active_crtc);
+
+	if (HAS_DDI(i915))
+		intel_ddi_sanitize_encoder_pll_mapping(encoder);
+}
+
+/* FIXME read out full plane state for all planes */
+static void readout_plane_state(struct drm_i915_private *i915)
+{
+	struct intel_plane *plane;
+	struct intel_crtc *crtc;
+
+	for_each_intel_plane(&i915->drm, plane) {
+		struct intel_plane_state *plane_state =
+			to_intel_plane_state(plane->base.state);
+		struct intel_crtc_state *crtc_state;
+		enum pipe pipe = PIPE_A;
+		bool visible;
+
+		visible = plane->get_hw_state(plane, &pipe);
+
+		crtc = intel_crtc_for_pipe(i915, pipe);
+		crtc_state = to_intel_crtc_state(crtc->base.state);
+
+		intel_set_plane_visible(crtc_state, plane_state, visible);
+
+		drm_dbg_kms(&i915->drm,
+			    "[PLANE:%d:%s] hw state readout: %s, pipe %c\n",
+			    plane->base.base.id, plane->base.name,
+			    str_enabled_disabled(visible), pipe_name(pipe));
+	}
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		intel_plane_fixup_bitmasks(crtc_state);
+	}
+}
+
+static void intel_modeset_readout_hw_state(struct drm_i915_private *i915)
+{
+	struct intel_cdclk_state *cdclk_state =
+		to_intel_cdclk_state(i915->display.cdclk.obj.state);
+	struct intel_dbuf_state *dbuf_state =
+		to_intel_dbuf_state(i915->display.dbuf.obj.state);
+	enum pipe pipe;
+	struct intel_crtc *crtc;
+	struct intel_encoder *encoder;
+	struct intel_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	u8 active_pipes = 0;
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		__drm_atomic_helper_crtc_destroy_state(&crtc_state->uapi);
+		intel_crtc_free_hw_state(crtc_state);
+		intel_crtc_state_reset(crtc_state, crtc);
+
+		intel_crtc_get_pipe_config(crtc_state);
+
+		crtc_state->hw.enable = crtc_state->hw.active;
+
+		crtc->base.enabled = crtc_state->hw.enable;
+		crtc->active = crtc_state->hw.active;
+
+		if (crtc_state->hw.active)
+			active_pipes |= BIT(crtc->pipe);
+
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] hw state readout: %s\n",
+			    crtc->base.base.id, crtc->base.name,
+			    str_enabled_disabled(crtc_state->hw.active));
+	}
+
+	cdclk_state->active_pipes = active_pipes;
+	dbuf_state->active_pipes = active_pipes;
+
+	readout_plane_state(i915);
+
+	for_each_intel_encoder(&i915->drm, encoder) {
+		struct intel_crtc_state *crtc_state = NULL;
+
+		pipe = 0;
+
+		if (encoder->get_hw_state(encoder, &pipe)) {
+			crtc = intel_crtc_for_pipe(i915, pipe);
+			crtc_state = to_intel_crtc_state(crtc->base.state);
+
+			encoder->base.crtc = &crtc->base;
+			intel_encoder_get_config(encoder, crtc_state);
+
+			/* read out to slave crtc as well for bigjoiner */
+			if (crtc_state->bigjoiner_pipes) {
+				struct intel_crtc *slave_crtc;
+
+				/* encoder should read be linked to bigjoiner master */
+				WARN_ON(intel_crtc_is_bigjoiner_slave(crtc_state));
+
+				for_each_intel_crtc_in_pipe_mask(&i915->drm, slave_crtc,
+								 intel_crtc_bigjoiner_slave_pipes(crtc_state)) {
+					struct intel_crtc_state *slave_crtc_state;
+
+					slave_crtc_state = to_intel_crtc_state(slave_crtc->base.state);
+					intel_encoder_get_config(encoder, slave_crtc_state);
+				}
+			}
+		} else {
+			encoder->base.crtc = NULL;
+		}
+
+		if (encoder->sync_state)
+			encoder->sync_state(encoder, crtc_state);
+
+		drm_dbg_kms(&i915->drm,
+			    "[ENCODER:%d:%s] hw state readout: %s, pipe %c\n",
+			    encoder->base.base.id, encoder->base.name,
+			    str_enabled_disabled(encoder->base.crtc),
+			    pipe_name(pipe));
+	}
+
+	intel_dpll_readout_hw_state(i915);
+
+	drm_connector_list_iter_begin(&i915->drm, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		if (connector->get_hw_state(connector)) {
+			struct intel_crtc_state *crtc_state;
+			struct intel_crtc *crtc;
+
+			connector->base.dpms = DRM_MODE_DPMS_ON;
+
+			encoder = intel_attached_encoder(connector);
+			connector->base.encoder = &encoder->base;
+
+			crtc = to_intel_crtc(encoder->base.crtc);
+			crtc_state = crtc ? to_intel_crtc_state(crtc->base.state) : NULL;
+
+			if (crtc_state && crtc_state->hw.active) {
+				/*
+				 * This has to be done during hardware readout
+				 * because anything calling .crtc_disable may
+				 * rely on the connector_mask being accurate.
+				 */
+				crtc_state->uapi.connector_mask |=
+					drm_connector_mask(&connector->base);
+				crtc_state->uapi.encoder_mask |=
+					drm_encoder_mask(&encoder->base);
+			}
+		} else {
+			connector->base.dpms = DRM_MODE_DPMS_OFF;
+			connector->base.encoder = NULL;
+		}
+		drm_dbg_kms(&i915->drm,
+			    "[CONNECTOR:%d:%s] hw state readout: %s\n",
+			    connector->base.base.id, connector->base.name,
+			    str_enabled_disabled(connector->base.encoder));
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_bw_state *bw_state =
+			to_intel_bw_state(i915->display.bw.obj.state);
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct intel_plane *plane;
+		int min_cdclk = 0;
+
+		if (crtc_state->hw.active) {
+			/*
+			 * The initial mode needs to be set in order to keep
+			 * the atomic core happy. It wants a valid mode if the
+			 * crtc's enabled, so we do the above call.
+			 *
+			 * But we don't set all the derived state fully, hence
+			 * set a flag to indicate that a full recalculation is
+			 * needed on the next commit.
+			 */
+			crtc_state->inherited = true;
+
+			intel_crtc_update_active_timings(crtc_state);
+
+			intel_crtc_copy_hw_to_uapi_state(crtc_state);
+		}
+
+		for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) {
+			const struct intel_plane_state *plane_state =
+				to_intel_plane_state(plane->base.state);
+
+			/*
+			 * FIXME don't have the fb yet, so can't
+			 * use intel_plane_data_rate() :(
+			 */
+			if (plane_state->uapi.visible)
+				crtc_state->data_rate[plane->id] =
+					4 * crtc_state->pixel_rate;
+			/*
+			 * FIXME don't have the fb yet, so can't
+			 * use plane->min_cdclk() :(
+			 */
+			if (plane_state->uapi.visible && plane->min_cdclk) {
+				if (crtc_state->double_wide || DISPLAY_VER(i915) >= 10)
+					crtc_state->min_cdclk[plane->id] =
+						DIV_ROUND_UP(crtc_state->pixel_rate, 2);
+				else
+					crtc_state->min_cdclk[plane->id] =
+						crtc_state->pixel_rate;
+			}
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] min_cdclk %d kHz\n",
+				    plane->base.base.id, plane->base.name,
+				    crtc_state->min_cdclk[plane->id]);
+		}
+
+		if (crtc_state->hw.active) {
+			min_cdclk = intel_crtc_compute_min_cdclk(crtc_state);
+			if (drm_WARN_ON(&i915->drm, min_cdclk < 0))
+				min_cdclk = 0;
+		}
+
+		cdclk_state->min_cdclk[crtc->pipe] = min_cdclk;
+		cdclk_state->min_voltage_level[crtc->pipe] =
+			crtc_state->min_voltage_level;
+
+		intel_bw_crtc_update(bw_state, crtc_state);
+	}
+}
+
+static void
+get_encoder_power_domains(struct drm_i915_private *i915)
+{
+	struct intel_encoder *encoder;
+
+	for_each_intel_encoder(&i915->drm, encoder) {
+		struct intel_crtc_state *crtc_state;
+
+		if (!encoder->get_power_domains)
+			continue;
+
+		/*
+		 * MST-primary and inactive encoders don't have a crtc state
+		 * and neither of these require any power domain references.
+		 */
+		if (!encoder->base.crtc)
+			continue;
+
+		crtc_state = to_intel_crtc_state(encoder->base.crtc->state);
+		encoder->get_power_domains(encoder, crtc_state);
+	}
+}
+
+static void intel_early_display_was(struct drm_i915_private *i915)
+{
+	/*
+	 * Display WA #1185 WaDisableDARBFClkGating:glk,icl,ehl,tgl
+	 * Also known as Wa_14010480278.
+	 */
+	if (IS_DISPLAY_VER(i915, 10, 12))
+		intel_de_write(i915, GEN9_CLKGATE_DIS_0,
+			       intel_de_read(i915, GEN9_CLKGATE_DIS_0) | DARBF_GATING_DIS);
+
+	if (IS_HASWELL(i915)) {
+		/*
+		 * WaRsPkgCStateDisplayPMReq:hsw
+		 * System hang if this isn't done before disabling all planes!
+		 */
+		intel_de_write(i915, CHICKEN_PAR1_1,
+			       intel_de_read(i915, CHICKEN_PAR1_1) | FORCE_ARB_IDLE_PLANES);
+	}
+
+	if (IS_KABYLAKE(i915) || IS_COFFEELAKE(i915) || IS_COMETLAKE(i915)) {
+		/* Display WA #1142:kbl,cfl,cml */
+		intel_de_rmw(i915, CHICKEN_PAR1_1,
+			     KBL_ARB_FILL_SPARE_22, KBL_ARB_FILL_SPARE_22);
+		intel_de_rmw(i915, CHICKEN_MISC_2,
+			     KBL_ARB_FILL_SPARE_13 | KBL_ARB_FILL_SPARE_14,
+			     KBL_ARB_FILL_SPARE_14);
+	}
+}
+
+void intel_modeset_setup_hw_state(struct drm_i915_private *i915,
+				  struct drm_modeset_acquire_ctx *ctx)
+{
+	struct intel_encoder *encoder;
+	struct intel_crtc *crtc;
+	intel_wakeref_t wakeref;
+
+	wakeref = intel_display_power_get(i915, POWER_DOMAIN_INIT);
+
+	intel_early_display_was(i915);
+	intel_modeset_readout_hw_state(i915);
+
+	/* HW state is read out, now we need to sanitize this mess. */
+	get_encoder_power_domains(i915);
+
+	intel_pch_sanitize(i915);
+
+	/*
+	 * intel_sanitize_plane_mapping() may need to do vblank
+	 * waits, so we need vblank interrupts restored beforehand.
+	 */
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		intel_sanitize_fifo_underrun_reporting(crtc_state);
+
+		drm_crtc_vblank_reset(&crtc->base);
+
+		if (crtc_state->hw.active)
+			intel_crtc_vblank_on(crtc_state);
+	}
+
+	intel_fbc_sanitize(i915);
+
+	intel_sanitize_plane_mapping(i915);
+
+	for_each_intel_encoder(&i915->drm, encoder)
+		intel_sanitize_encoder(encoder);
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		intel_sanitize_crtc(crtc, ctx);
+		intel_crtc_state_dump(crtc_state, NULL, "setup_hw_state");
+	}
+
+	intel_modeset_update_connector_atomic_state(i915);
+
+	intel_dpll_sanitize_state(i915);
+
+	if (IS_G4X(i915)) {
+		g4x_wm_get_hw_state(i915);
+		g4x_wm_sanitize(i915);
+	} else if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) {
+		vlv_wm_get_hw_state(i915);
+		vlv_wm_sanitize(i915);
+	} else if (DISPLAY_VER(i915) >= 9) {
+		skl_wm_get_hw_state(i915);
+		skl_wm_sanitize(i915);
+	} else if (HAS_PCH_SPLIT(i915)) {
+		ilk_wm_get_hw_state(i915);
+	}
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct intel_power_domain_mask put_domains;
+
+		intel_modeset_get_crtc_power_domains(crtc_state, &put_domains);
+		if (drm_WARN_ON(&i915->drm, !bitmap_empty(put_domains.bits, POWER_DOMAIN_NUM)))
+			intel_modeset_put_crtc_power_domains(crtc, &put_domains);
+	}
+
+	intel_display_power_put(i915, POWER_DOMAIN_INIT, wakeref);
+
+	intel_power_domains_sanitize_state(i915);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_modeset_setup.h b/drivers/gpu/drm/i915/display/intel_modeset_setup.h
new file mode 100644
index 000000000..3beff67b3
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_modeset_setup.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_MODESET_SETUP_H__
+#define __INTEL_MODESET_SETUP_H__
+
+struct drm_i915_private;
+struct drm_modeset_acquire_ctx;
+
+void intel_modeset_setup_hw_state(struct drm_i915_private *i915,
+				  struct drm_modeset_acquire_ctx *ctx);
+
+#endif /* __INTEL_MODESET_SETUP_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_modeset_verify.c b/drivers/gpu/drm/i915/display/intel_modeset_verify.c
new file mode 100644
index 000000000..0fdcf2e6d
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_modeset_verify.c
@@ -0,0 +1,246 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ *
+ * High level crtc/connector/encoder modeset state verification.
+ */
+
+#include <drm/drm_atomic_state_helper.h>
+
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_crtc.h"
+#include "intel_crtc_state_dump.h"
+#include "intel_display.h"
+#include "intel_display_types.h"
+#include "intel_fdi.h"
+#include "intel_modeset_verify.h"
+#include "intel_snps_phy.h"
+#include "skl_watermark.h"
+
+/*
+ * Cross check the actual hw state with our own modeset state tracking (and its
+ * internal consistency).
+ */
+static void intel_connector_verify_state(struct intel_crtc_state *crtc_state,
+					 struct drm_connector_state *conn_state)
+{
+	struct intel_connector *connector = to_intel_connector(conn_state->connector);
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+
+	drm_dbg_kms(&i915->drm, "[CONNECTOR:%d:%s]\n",
+		    connector->base.base.id, connector->base.name);
+
+	if (connector->get_hw_state(connector)) {
+		struct intel_encoder *encoder = intel_attached_encoder(connector);
+
+		I915_STATE_WARN(!crtc_state,
+				"connector enabled without attached crtc\n");
+
+		if (!crtc_state)
+			return;
+
+		I915_STATE_WARN(!crtc_state->hw.active,
+				"connector is active, but attached crtc isn't\n");
+
+		if (!encoder || encoder->type == INTEL_OUTPUT_DP_MST)
+			return;
+
+		I915_STATE_WARN(conn_state->best_encoder != &encoder->base,
+				"atomic encoder doesn't match attached encoder\n");
+
+		I915_STATE_WARN(conn_state->crtc != encoder->base.crtc,
+				"attached encoder crtc differs from connector crtc\n");
+	} else {
+		I915_STATE_WARN(crtc_state && crtc_state->hw.active,
+				"attached crtc is active, but connector isn't\n");
+		I915_STATE_WARN(!crtc_state && conn_state->best_encoder,
+				"best encoder set without crtc!\n");
+	}
+}
+
+static void
+verify_connector_state(struct intel_atomic_state *state,
+		       struct intel_crtc *crtc)
+{
+	struct drm_connector *connector;
+	struct drm_connector_state *new_conn_state;
+	int i;
+
+	for_each_new_connector_in_state(&state->base, connector, new_conn_state, i) {
+		struct drm_encoder *encoder = connector->encoder;
+		struct intel_crtc_state *crtc_state = NULL;
+
+		if (new_conn_state->crtc != &crtc->base)
+			continue;
+
+		if (crtc)
+			crtc_state = intel_atomic_get_new_crtc_state(state, crtc);
+
+		intel_connector_verify_state(crtc_state, new_conn_state);
+
+		I915_STATE_WARN(new_conn_state->best_encoder != encoder,
+				"connector's atomic encoder doesn't match legacy encoder\n");
+	}
+}
+
+static void intel_pipe_config_sanity_check(struct drm_i915_private *dev_priv,
+					   const struct intel_crtc_state *pipe_config)
+{
+	if (pipe_config->has_pch_encoder) {
+		int fdi_dotclock = intel_dotclock_calculate(intel_fdi_link_freq(dev_priv, pipe_config),
+							    &pipe_config->fdi_m_n);
+		int dotclock = pipe_config->hw.adjusted_mode.crtc_clock;
+
+		/*
+		 * FDI already provided one idea for the dotclock.
+		 * Yell if the encoder disagrees. Allow for slight
+		 * rounding differences.
+		 */
+		drm_WARN(&dev_priv->drm, abs(fdi_dotclock - dotclock) > 1,
+			 "FDI dotclock and encoder dotclock mismatch, fdi: %i, encoder: %i\n",
+			 fdi_dotclock, dotclock);
+	}
+}
+
+static void
+verify_encoder_state(struct drm_i915_private *dev_priv, struct intel_atomic_state *state)
+{
+	struct intel_encoder *encoder;
+	struct drm_connector *connector;
+	struct drm_connector_state *old_conn_state, *new_conn_state;
+	int i;
+
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		bool enabled = false, found = false;
+		enum pipe pipe;
+
+		drm_dbg_kms(&dev_priv->drm, "[ENCODER:%d:%s]\n",
+			    encoder->base.base.id,
+			    encoder->base.name);
+
+		for_each_oldnew_connector_in_state(&state->base, connector, old_conn_state,
+						   new_conn_state, i) {
+			if (old_conn_state->best_encoder == &encoder->base)
+				found = true;
+
+			if (new_conn_state->best_encoder != &encoder->base)
+				continue;
+
+			found = true;
+			enabled = true;
+
+			I915_STATE_WARN(new_conn_state->crtc !=
+					encoder->base.crtc,
+					"connector's crtc doesn't match encoder crtc\n");
+		}
+
+		if (!found)
+			continue;
+
+		I915_STATE_WARN(!!encoder->base.crtc != enabled,
+				"encoder's enabled state mismatch (expected %i, found %i)\n",
+				!!encoder->base.crtc, enabled);
+
+		if (!encoder->base.crtc) {
+			bool active;
+
+			active = encoder->get_hw_state(encoder, &pipe);
+			I915_STATE_WARN(active,
+					"encoder detached but still enabled on pipe %c.\n",
+					pipe_name(pipe));
+		}
+	}
+}
+
+static void
+verify_crtc_state(struct intel_crtc *crtc,
+		  struct intel_crtc_state *old_crtc_state,
+		  struct intel_crtc_state *new_crtc_state)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_encoder *encoder;
+	struct intel_crtc_state *pipe_config = old_crtc_state;
+	struct drm_atomic_state *state = old_crtc_state->uapi.state;
+	struct intel_crtc *master_crtc;
+
+	__drm_atomic_helper_crtc_destroy_state(&old_crtc_state->uapi);
+	intel_crtc_free_hw_state(old_crtc_state);
+	intel_crtc_state_reset(old_crtc_state, crtc);
+	old_crtc_state->uapi.state = state;
+
+	drm_dbg_kms(&dev_priv->drm, "[CRTC:%d:%s]\n", crtc->base.base.id,
+		    crtc->base.name);
+
+	pipe_config->hw.enable = new_crtc_state->hw.enable;
+
+	intel_crtc_get_pipe_config(pipe_config);
+
+	/* we keep both pipes enabled on 830 */
+	if (IS_I830(dev_priv) && pipe_config->hw.active)
+		pipe_config->hw.active = new_crtc_state->hw.active;
+
+	I915_STATE_WARN(new_crtc_state->hw.active != pipe_config->hw.active,
+			"crtc active state doesn't match with hw state (expected %i, found %i)\n",
+			new_crtc_state->hw.active, pipe_config->hw.active);
+
+	I915_STATE_WARN(crtc->active != new_crtc_state->hw.active,
+			"transitional active state does not match atomic hw state (expected %i, found %i)\n",
+			new_crtc_state->hw.active, crtc->active);
+
+	master_crtc = intel_master_crtc(new_crtc_state);
+
+	for_each_encoder_on_crtc(dev, &master_crtc->base, encoder) {
+		enum pipe pipe;
+		bool active;
+
+		active = encoder->get_hw_state(encoder, &pipe);
+		I915_STATE_WARN(active != new_crtc_state->hw.active,
+				"[ENCODER:%i] active %i with crtc active %i\n",
+				encoder->base.base.id, active,
+				new_crtc_state->hw.active);
+
+		I915_STATE_WARN(active && master_crtc->pipe != pipe,
+				"Encoder connected to wrong pipe %c\n",
+				pipe_name(pipe));
+
+		if (active)
+			intel_encoder_get_config(encoder, pipe_config);
+	}
+
+	if (!new_crtc_state->hw.active)
+		return;
+
+	intel_pipe_config_sanity_check(dev_priv, pipe_config);
+
+	if (!intel_pipe_config_compare(new_crtc_state,
+				       pipe_config, false)) {
+		I915_STATE_WARN(1, "pipe state doesn't match!\n");
+		intel_crtc_state_dump(pipe_config, NULL, "hw state");
+		intel_crtc_state_dump(new_crtc_state, NULL, "sw state");
+	}
+}
+
+void intel_modeset_verify_crtc(struct intel_crtc *crtc,
+			       struct intel_atomic_state *state,
+			       struct intel_crtc_state *old_crtc_state,
+			       struct intel_crtc_state *new_crtc_state)
+{
+	if (!intel_crtc_needs_modeset(new_crtc_state) && !new_crtc_state->update_pipe)
+		return;
+
+	intel_wm_state_verify(crtc, new_crtc_state);
+	verify_connector_state(state, crtc);
+	verify_crtc_state(crtc, old_crtc_state, new_crtc_state);
+	intel_shared_dpll_state_verify(crtc, old_crtc_state, new_crtc_state);
+	intel_mpllb_state_verify(state, new_crtc_state);
+}
+
+void intel_modeset_verify_disabled(struct drm_i915_private *dev_priv,
+				   struct intel_atomic_state *state)
+{
+	verify_encoder_state(dev_priv, state);
+	verify_connector_state(state, NULL);
+	intel_shared_dpll_verify_disabled(dev_priv);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_modeset_verify.h b/drivers/gpu/drm/i915/display/intel_modeset_verify.h
new file mode 100644
index 000000000..2d6fbe4f7
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_modeset_verify.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_MODESET_VERIFY_H__
+#define __INTEL_MODESET_VERIFY_H__
+
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+
+void intel_modeset_verify_crtc(struct intel_crtc *crtc,
+			       struct intel_atomic_state *state,
+			       struct intel_crtc_state *old_crtc_state,
+			       struct intel_crtc_state *new_crtc_state);
+void intel_modeset_verify_disabled(struct drm_i915_private *dev_priv,
+				   struct intel_atomic_state *state);
+
+#endif /* __INTEL_MODESET_VERIFY_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_opregion.c b/drivers/gpu/drm/i915/display/intel_opregion.c
new file mode 100644
index 000000000..caa07ef34
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_opregion.c
@@ -0,0 +1,1250 @@
+/*
+ * Copyright 2008 Intel Corporation <hong.liu@intel.com>
+ * Copyright 2008 Red Hat <mjg@redhat.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial
+ * portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NON-INFRINGEMENT.  IN NO EVENT SHALL INTEL AND/OR ITS SUPPLIERS BE
+ * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#include <linux/acpi.h>
+#include <linux/dmi.h>
+#include <linux/firmware.h>
+#include <acpi/video.h>
+
+#include <drm/drm_edid.h>
+
+#include "i915_drv.h"
+#include "intel_acpi.h"
+#include "intel_backlight.h"
+#include "intel_display_types.h"
+#include "intel_opregion.h"
+#include "intel_pci_config.h"
+
+#define OPREGION_HEADER_OFFSET 0
+#define OPREGION_ACPI_OFFSET   0x100
+#define   ACPI_CLID 0x01ac /* current lid state indicator */
+#define   ACPI_CDCK 0x01b0 /* current docking state indicator */
+#define OPREGION_SWSCI_OFFSET  0x200
+#define OPREGION_ASLE_OFFSET   0x300
+#define OPREGION_VBT_OFFSET    0x400
+#define OPREGION_ASLE_EXT_OFFSET	0x1C00
+
+#define OPREGION_SIGNATURE "IntelGraphicsMem"
+#define MBOX_ACPI		BIT(0)	/* Mailbox #1 */
+#define MBOX_SWSCI		BIT(1)	/* Mailbox #2 (obsolete from v2.x) */
+#define MBOX_ASLE		BIT(2)	/* Mailbox #3 */
+#define MBOX_ASLE_EXT		BIT(4)	/* Mailbox #5 */
+#define MBOX_BACKLIGHT		BIT(5)	/* Mailbox #2 (valid from v3.x) */
+
+#define PCON_HEADLESS_SKU	BIT(13)
+
+struct opregion_header {
+	u8 signature[16];
+	u32 size;
+	struct {
+		u8 rsvd;
+		u8 revision;
+		u8 minor;
+		u8 major;
+	}  __packed over;
+	u8 bios_ver[32];
+	u8 vbios_ver[16];
+	u8 driver_ver[16];
+	u32 mboxes;
+	u32 driver_model;
+	u32 pcon;
+	u8 dver[32];
+	u8 rsvd[124];
+} __packed;
+
+/* OpRegion mailbox #1: public ACPI methods */
+struct opregion_acpi {
+	u32 drdy;       /* driver readiness */
+	u32 csts;       /* notification status */
+	u32 cevt;       /* current event */
+	u8 rsvd1[20];
+	u32 didl[8];    /* supported display devices ID list */
+	u32 cpdl[8];    /* currently presented display list */
+	u32 cadl[8];    /* currently active display list */
+	u32 nadl[8];    /* next active devices list */
+	u32 aslp;       /* ASL sleep time-out */
+	u32 tidx;       /* toggle table index */
+	u32 chpd;       /* current hotplug enable indicator */
+	u32 clid;       /* current lid state*/
+	u32 cdck;       /* current docking state */
+	u32 sxsw;       /* Sx state resume */
+	u32 evts;       /* ASL supported events */
+	u32 cnot;       /* current OS notification */
+	u32 nrdy;       /* driver status */
+	u32 did2[7];	/* extended supported display devices ID list */
+	u32 cpd2[7];	/* extended attached display devices list */
+	u8 rsvd2[4];
+} __packed;
+
+/* OpRegion mailbox #2: SWSCI */
+struct opregion_swsci {
+	u32 scic;       /* SWSCI command|status|data */
+	u32 parm;       /* command parameters */
+	u32 dslp;       /* driver sleep time-out */
+	u8 rsvd[244];
+} __packed;
+
+/* OpRegion mailbox #3: ASLE */
+struct opregion_asle {
+	u32 ardy;       /* driver readiness */
+	u32 aslc;       /* ASLE interrupt command */
+	u32 tche;       /* technology enabled indicator */
+	u32 alsi;       /* current ALS illuminance reading */
+	u32 bclp;       /* backlight brightness to set */
+	u32 pfit;       /* panel fitting state */
+	u32 cblv;       /* current brightness level */
+	u16 bclm[20];   /* backlight level duty cycle mapping table */
+	u32 cpfm;       /* current panel fitting mode */
+	u32 epfm;       /* enabled panel fitting modes */
+	u8 plut[74];    /* panel LUT and identifier */
+	u32 pfmb;       /* PWM freq and min brightness */
+	u32 cddv;       /* color correction default values */
+	u32 pcft;       /* power conservation features */
+	u32 srot;       /* supported rotation angles */
+	u32 iuer;       /* IUER events */
+	u64 fdss;
+	u32 fdsp;
+	u32 stat;
+	u64 rvda;	/* Physical (2.0) or relative from opregion (2.1+)
+			 * address of raw VBT data. */
+	u32 rvds;	/* Size of raw vbt data */
+	u8 rsvd[58];
+} __packed;
+
+/* OpRegion mailbox #5: ASLE ext */
+struct opregion_asle_ext {
+	u32 phed;	/* Panel Header */
+	u8 bddc[256];	/* Panel EDID */
+	u8 rsvd[764];
+} __packed;
+
+/* Driver readiness indicator */
+#define ASLE_ARDY_READY		(1 << 0)
+#define ASLE_ARDY_NOT_READY	(0 << 0)
+
+/* ASLE Interrupt Command (ASLC) bits */
+#define ASLC_SET_ALS_ILLUM		(1 << 0)
+#define ASLC_SET_BACKLIGHT		(1 << 1)
+#define ASLC_SET_PFIT			(1 << 2)
+#define ASLC_SET_PWM_FREQ		(1 << 3)
+#define ASLC_SUPPORTED_ROTATION_ANGLES	(1 << 4)
+#define ASLC_BUTTON_ARRAY		(1 << 5)
+#define ASLC_CONVERTIBLE_INDICATOR	(1 << 6)
+#define ASLC_DOCKING_INDICATOR		(1 << 7)
+#define ASLC_ISCT_STATE_CHANGE		(1 << 8)
+#define ASLC_REQ_MSK			0x1ff
+/* response bits */
+#define ASLC_ALS_ILLUM_FAILED		(1 << 10)
+#define ASLC_BACKLIGHT_FAILED		(1 << 12)
+#define ASLC_PFIT_FAILED		(1 << 14)
+#define ASLC_PWM_FREQ_FAILED		(1 << 16)
+#define ASLC_ROTATION_ANGLES_FAILED	(1 << 18)
+#define ASLC_BUTTON_ARRAY_FAILED	(1 << 20)
+#define ASLC_CONVERTIBLE_FAILED		(1 << 22)
+#define ASLC_DOCKING_FAILED		(1 << 24)
+#define ASLC_ISCT_STATE_FAILED		(1 << 26)
+
+/* Technology enabled indicator */
+#define ASLE_TCHE_ALS_EN	(1 << 0)
+#define ASLE_TCHE_BLC_EN	(1 << 1)
+#define ASLE_TCHE_PFIT_EN	(1 << 2)
+#define ASLE_TCHE_PFMB_EN	(1 << 3)
+
+/* ASLE backlight brightness to set */
+#define ASLE_BCLP_VALID                (1<<31)
+#define ASLE_BCLP_MSK          (~(1<<31))
+
+/* ASLE panel fitting request */
+#define ASLE_PFIT_VALID         (1<<31)
+#define ASLE_PFIT_CENTER (1<<0)
+#define ASLE_PFIT_STRETCH_TEXT (1<<1)
+#define ASLE_PFIT_STRETCH_GFX (1<<2)
+
+/* PWM frequency and minimum brightness */
+#define ASLE_PFMB_BRIGHTNESS_MASK (0xff)
+#define ASLE_PFMB_BRIGHTNESS_VALID (1<<8)
+#define ASLE_PFMB_PWM_MASK (0x7ffffe00)
+#define ASLE_PFMB_PWM_VALID (1<<31)
+
+#define ASLE_CBLV_VALID         (1<<31)
+
+/* IUER */
+#define ASLE_IUER_DOCKING		(1 << 7)
+#define ASLE_IUER_CONVERTIBLE		(1 << 6)
+#define ASLE_IUER_ROTATION_LOCK_BTN	(1 << 4)
+#define ASLE_IUER_VOLUME_DOWN_BTN	(1 << 3)
+#define ASLE_IUER_VOLUME_UP_BTN		(1 << 2)
+#define ASLE_IUER_WINDOWS_BTN		(1 << 1)
+#define ASLE_IUER_POWER_BTN		(1 << 0)
+
+#define ASLE_PHED_EDID_VALID_MASK	0x3
+
+/* Software System Control Interrupt (SWSCI) */
+#define SWSCI_SCIC_INDICATOR		(1 << 0)
+#define SWSCI_SCIC_MAIN_FUNCTION_SHIFT	1
+#define SWSCI_SCIC_MAIN_FUNCTION_MASK	(0xf << 1)
+#define SWSCI_SCIC_SUB_FUNCTION_SHIFT	8
+#define SWSCI_SCIC_SUB_FUNCTION_MASK	(0xff << 8)
+#define SWSCI_SCIC_EXIT_PARAMETER_SHIFT	8
+#define SWSCI_SCIC_EXIT_PARAMETER_MASK	(0xff << 8)
+#define SWSCI_SCIC_EXIT_STATUS_SHIFT	5
+#define SWSCI_SCIC_EXIT_STATUS_MASK	(7 << 5)
+#define SWSCI_SCIC_EXIT_STATUS_SUCCESS	1
+
+#define SWSCI_FUNCTION_CODE(main, sub) \
+	((main) << SWSCI_SCIC_MAIN_FUNCTION_SHIFT | \
+	 (sub) << SWSCI_SCIC_SUB_FUNCTION_SHIFT)
+
+/* SWSCI: Get BIOS Data (GBDA) */
+#define SWSCI_GBDA			4
+#define SWSCI_GBDA_SUPPORTED_CALLS	SWSCI_FUNCTION_CODE(SWSCI_GBDA, 0)
+#define SWSCI_GBDA_REQUESTED_CALLBACKS	SWSCI_FUNCTION_CODE(SWSCI_GBDA, 1)
+#define SWSCI_GBDA_BOOT_DISPLAY_PREF	SWSCI_FUNCTION_CODE(SWSCI_GBDA, 4)
+#define SWSCI_GBDA_PANEL_DETAILS	SWSCI_FUNCTION_CODE(SWSCI_GBDA, 5)
+#define SWSCI_GBDA_TV_STANDARD		SWSCI_FUNCTION_CODE(SWSCI_GBDA, 6)
+#define SWSCI_GBDA_INTERNAL_GRAPHICS	SWSCI_FUNCTION_CODE(SWSCI_GBDA, 7)
+#define SWSCI_GBDA_SPREAD_SPECTRUM	SWSCI_FUNCTION_CODE(SWSCI_GBDA, 10)
+
+/* SWSCI: System BIOS Callbacks (SBCB) */
+#define SWSCI_SBCB			6
+#define SWSCI_SBCB_SUPPORTED_CALLBACKS	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 0)
+#define SWSCI_SBCB_INIT_COMPLETION	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 1)
+#define SWSCI_SBCB_PRE_HIRES_SET_MODE	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 3)
+#define SWSCI_SBCB_POST_HIRES_SET_MODE	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 4)
+#define SWSCI_SBCB_DISPLAY_SWITCH	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 5)
+#define SWSCI_SBCB_SET_TV_FORMAT	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 6)
+#define SWSCI_SBCB_ADAPTER_POWER_STATE	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 7)
+#define SWSCI_SBCB_DISPLAY_POWER_STATE	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 8)
+#define SWSCI_SBCB_SET_BOOT_DISPLAY	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 9)
+#define SWSCI_SBCB_SET_PANEL_DETAILS	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 10)
+#define SWSCI_SBCB_SET_INTERNAL_GFX	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 11)
+#define SWSCI_SBCB_POST_HIRES_TO_DOS_FS	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 16)
+#define SWSCI_SBCB_SUSPEND_RESUME	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 17)
+#define SWSCI_SBCB_SET_SPREAD_SPECTRUM	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 18)
+#define SWSCI_SBCB_POST_VBE_PM		SWSCI_FUNCTION_CODE(SWSCI_SBCB, 19)
+#define SWSCI_SBCB_ENABLE_DISABLE_AUDIO	SWSCI_FUNCTION_CODE(SWSCI_SBCB, 21)
+
+#define MAX_DSLP	1500
+
+static int check_swsci_function(struct drm_i915_private *i915, u32 function)
+{
+	struct opregion_swsci *swsci = i915->display.opregion.swsci;
+	u32 main_function, sub_function;
+
+	if (!swsci)
+		return -ENODEV;
+
+	main_function = (function & SWSCI_SCIC_MAIN_FUNCTION_MASK) >>
+		SWSCI_SCIC_MAIN_FUNCTION_SHIFT;
+	sub_function = (function & SWSCI_SCIC_SUB_FUNCTION_MASK) >>
+		SWSCI_SCIC_SUB_FUNCTION_SHIFT;
+
+	/* Check if we can call the function. See swsci_setup for details. */
+	if (main_function == SWSCI_SBCB) {
+		if ((i915->display.opregion.swsci_sbcb_sub_functions &
+		     (1 << sub_function)) == 0)
+			return -EINVAL;
+	} else if (main_function == SWSCI_GBDA) {
+		if ((i915->display.opregion.swsci_gbda_sub_functions &
+		     (1 << sub_function)) == 0)
+			return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int swsci(struct drm_i915_private *dev_priv,
+		 u32 function, u32 parm, u32 *parm_out)
+{
+	struct opregion_swsci *swsci = dev_priv->display.opregion.swsci;
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	u32 scic, dslp;
+	u16 swsci_val;
+	int ret;
+
+	ret = check_swsci_function(dev_priv, function);
+	if (ret)
+		return ret;
+
+	/* Driver sleep timeout in ms. */
+	dslp = swsci->dslp;
+	if (!dslp) {
+		/* The spec says 2ms should be the default, but it's too small
+		 * for some machines. */
+		dslp = 50;
+	} else if (dslp > MAX_DSLP) {
+		/* Hey bios, trust must be earned. */
+		DRM_INFO_ONCE("ACPI BIOS requests an excessive sleep of %u ms, "
+			      "using %u ms instead\n", dslp, MAX_DSLP);
+		dslp = MAX_DSLP;
+	}
+
+	/* The spec tells us to do this, but we are the only user... */
+	scic = swsci->scic;
+	if (scic & SWSCI_SCIC_INDICATOR) {
+		drm_dbg(&dev_priv->drm, "SWSCI request already in progress\n");
+		return -EBUSY;
+	}
+
+	scic = function | SWSCI_SCIC_INDICATOR;
+
+	swsci->parm = parm;
+	swsci->scic = scic;
+
+	/* Ensure SCI event is selected and event trigger is cleared. */
+	pci_read_config_word(pdev, SWSCI, &swsci_val);
+	if (!(swsci_val & SWSCI_SCISEL) || (swsci_val & SWSCI_GSSCIE)) {
+		swsci_val |= SWSCI_SCISEL;
+		swsci_val &= ~SWSCI_GSSCIE;
+		pci_write_config_word(pdev, SWSCI, swsci_val);
+	}
+
+	/* Use event trigger to tell bios to check the mail. */
+	swsci_val |= SWSCI_GSSCIE;
+	pci_write_config_word(pdev, SWSCI, swsci_val);
+
+	/* Poll for the result. */
+#define C (((scic = swsci->scic) & SWSCI_SCIC_INDICATOR) == 0)
+	if (wait_for(C, dslp)) {
+		drm_dbg(&dev_priv->drm, "SWSCI request timed out\n");
+		return -ETIMEDOUT;
+	}
+
+	scic = (scic & SWSCI_SCIC_EXIT_STATUS_MASK) >>
+		SWSCI_SCIC_EXIT_STATUS_SHIFT;
+
+	/* Note: scic == 0 is an error! */
+	if (scic != SWSCI_SCIC_EXIT_STATUS_SUCCESS) {
+		drm_dbg(&dev_priv->drm, "SWSCI request error %u\n", scic);
+		return -EIO;
+	}
+
+	if (parm_out)
+		*parm_out = swsci->parm;
+
+	return 0;
+
+#undef C
+}
+
+#define DISPLAY_TYPE_CRT			0
+#define DISPLAY_TYPE_TV				1
+#define DISPLAY_TYPE_EXTERNAL_FLAT_PANEL	2
+#define DISPLAY_TYPE_INTERNAL_FLAT_PANEL	3
+
+int intel_opregion_notify_encoder(struct intel_encoder *intel_encoder,
+				  bool enable)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
+	u32 parm = 0;
+	u32 type = 0;
+	u32 port;
+	int ret;
+
+	/* don't care about old stuff for now */
+	if (!HAS_DDI(dev_priv))
+		return 0;
+
+	/* Avoid port out of bounds checks if SWSCI isn't there. */
+	ret = check_swsci_function(dev_priv, SWSCI_SBCB_DISPLAY_POWER_STATE);
+	if (ret)
+		return ret;
+
+	if (intel_encoder->type == INTEL_OUTPUT_DSI)
+		port = 0;
+	else
+		port = intel_encoder->port;
+
+	if (port == PORT_E)  {
+		port = 0;
+	} else {
+		parm |= 1 << port;
+		port++;
+	}
+
+	/*
+	 * The port numbering and mapping here is bizarre. The now-obsolete
+	 * swsci spec supports ports numbered [0..4]. Port E is handled as a
+	 * special case, but port F and beyond are not. The functionality is
+	 * supposed to be obsolete for new platforms. Just bail out if the port
+	 * number is out of bounds after mapping.
+	 */
+	if (port > 4) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "[ENCODER:%d:%s] port %c (index %u) out of bounds for display power state notification\n",
+			    intel_encoder->base.base.id, intel_encoder->base.name,
+			    port_name(intel_encoder->port), port);
+		return -EINVAL;
+	}
+
+	if (!enable)
+		parm |= 4 << 8;
+
+	switch (intel_encoder->type) {
+	case INTEL_OUTPUT_ANALOG:
+		type = DISPLAY_TYPE_CRT;
+		break;
+	case INTEL_OUTPUT_DDI:
+	case INTEL_OUTPUT_DP:
+	case INTEL_OUTPUT_HDMI:
+	case INTEL_OUTPUT_DP_MST:
+		type = DISPLAY_TYPE_EXTERNAL_FLAT_PANEL;
+		break;
+	case INTEL_OUTPUT_EDP:
+	case INTEL_OUTPUT_DSI:
+		type = DISPLAY_TYPE_INTERNAL_FLAT_PANEL;
+		break;
+	default:
+		drm_WARN_ONCE(&dev_priv->drm, 1,
+			      "unsupported intel_encoder type %d\n",
+			      intel_encoder->type);
+		return -EINVAL;
+	}
+
+	parm |= type << (16 + port * 3);
+
+	return swsci(dev_priv, SWSCI_SBCB_DISPLAY_POWER_STATE, parm, NULL);
+}
+
+static const struct {
+	pci_power_t pci_power_state;
+	u32 parm;
+} power_state_map[] = {
+	{ PCI_D0,	0x00 },
+	{ PCI_D1,	0x01 },
+	{ PCI_D2,	0x02 },
+	{ PCI_D3hot,	0x04 },
+	{ PCI_D3cold,	0x04 },
+};
+
+int intel_opregion_notify_adapter(struct drm_i915_private *dev_priv,
+				  pci_power_t state)
+{
+	int i;
+
+	if (!HAS_DDI(dev_priv))
+		return 0;
+
+	for (i = 0; i < ARRAY_SIZE(power_state_map); i++) {
+		if (state == power_state_map[i].pci_power_state)
+			return swsci(dev_priv, SWSCI_SBCB_ADAPTER_POWER_STATE,
+				     power_state_map[i].parm, NULL);
+	}
+
+	return -EINVAL;
+}
+
+static u32 asle_set_backlight(struct drm_i915_private *dev_priv, u32 bclp)
+{
+	struct intel_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	struct opregion_asle *asle = dev_priv->display.opregion.asle;
+	struct drm_device *dev = &dev_priv->drm;
+
+	drm_dbg(&dev_priv->drm, "bclp = 0x%08x\n", bclp);
+
+	if (acpi_video_get_backlight_type() == acpi_backlight_native) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "opregion backlight request ignored\n");
+		return 0;
+	}
+
+	if (!(bclp & ASLE_BCLP_VALID))
+		return ASLC_BACKLIGHT_FAILED;
+
+	bclp &= ASLE_BCLP_MSK;
+	if (bclp > 255)
+		return ASLC_BACKLIGHT_FAILED;
+
+	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
+
+	/*
+	 * Update backlight on all connectors that support backlight (usually
+	 * only one).
+	 */
+	drm_dbg_kms(&dev_priv->drm, "updating opregion backlight %d/255\n",
+		    bclp);
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter)
+		intel_backlight_set_acpi(connector->base.state, bclp, 255);
+	drm_connector_list_iter_end(&conn_iter);
+	asle->cblv = DIV_ROUND_UP(bclp * 100, 255) | ASLE_CBLV_VALID;
+
+	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+
+
+	return 0;
+}
+
+static u32 asle_set_als_illum(struct drm_i915_private *dev_priv, u32 alsi)
+{
+	/* alsi is the current ALS reading in lux. 0 indicates below sensor
+	   range, 0xffff indicates above sensor range. 1-0xfffe are valid */
+	drm_dbg(&dev_priv->drm, "Illum is not supported\n");
+	return ASLC_ALS_ILLUM_FAILED;
+}
+
+static u32 asle_set_pwm_freq(struct drm_i915_private *dev_priv, u32 pfmb)
+{
+	drm_dbg(&dev_priv->drm, "PWM freq is not supported\n");
+	return ASLC_PWM_FREQ_FAILED;
+}
+
+static u32 asle_set_pfit(struct drm_i915_private *dev_priv, u32 pfit)
+{
+	/* Panel fitting is currently controlled by the X code, so this is a
+	   noop until modesetting support works fully */
+	drm_dbg(&dev_priv->drm, "Pfit is not supported\n");
+	return ASLC_PFIT_FAILED;
+}
+
+static u32 asle_set_supported_rotation_angles(struct drm_i915_private *dev_priv, u32 srot)
+{
+	drm_dbg(&dev_priv->drm, "SROT is not supported\n");
+	return ASLC_ROTATION_ANGLES_FAILED;
+}
+
+static u32 asle_set_button_array(struct drm_i915_private *dev_priv, u32 iuer)
+{
+	if (!iuer)
+		drm_dbg(&dev_priv->drm,
+			"Button array event is not supported (nothing)\n");
+	if (iuer & ASLE_IUER_ROTATION_LOCK_BTN)
+		drm_dbg(&dev_priv->drm,
+			"Button array event is not supported (rotation lock)\n");
+	if (iuer & ASLE_IUER_VOLUME_DOWN_BTN)
+		drm_dbg(&dev_priv->drm,
+			"Button array event is not supported (volume down)\n");
+	if (iuer & ASLE_IUER_VOLUME_UP_BTN)
+		drm_dbg(&dev_priv->drm,
+			"Button array event is not supported (volume up)\n");
+	if (iuer & ASLE_IUER_WINDOWS_BTN)
+		drm_dbg(&dev_priv->drm,
+			"Button array event is not supported (windows)\n");
+	if (iuer & ASLE_IUER_POWER_BTN)
+		drm_dbg(&dev_priv->drm,
+			"Button array event is not supported (power)\n");
+
+	return ASLC_BUTTON_ARRAY_FAILED;
+}
+
+static u32 asle_set_convertible(struct drm_i915_private *dev_priv, u32 iuer)
+{
+	if (iuer & ASLE_IUER_CONVERTIBLE)
+		drm_dbg(&dev_priv->drm,
+			"Convertible is not supported (clamshell)\n");
+	else
+		drm_dbg(&dev_priv->drm,
+			"Convertible is not supported (slate)\n");
+
+	return ASLC_CONVERTIBLE_FAILED;
+}
+
+static u32 asle_set_docking(struct drm_i915_private *dev_priv, u32 iuer)
+{
+	if (iuer & ASLE_IUER_DOCKING)
+		drm_dbg(&dev_priv->drm, "Docking is not supported (docked)\n");
+	else
+		drm_dbg(&dev_priv->drm,
+			"Docking is not supported (undocked)\n");
+
+	return ASLC_DOCKING_FAILED;
+}
+
+static u32 asle_isct_state(struct drm_i915_private *dev_priv)
+{
+	drm_dbg(&dev_priv->drm, "ISCT is not supported\n");
+	return ASLC_ISCT_STATE_FAILED;
+}
+
+static void asle_work(struct work_struct *work)
+{
+	struct intel_opregion *opregion =
+		container_of(work, struct intel_opregion, asle_work);
+	struct drm_i915_private *dev_priv =
+		container_of(opregion, struct drm_i915_private, display.opregion);
+	struct opregion_asle *asle = dev_priv->display.opregion.asle;
+	u32 aslc_stat = 0;
+	u32 aslc_req;
+
+	if (!asle)
+		return;
+
+	aslc_req = asle->aslc;
+
+	if (!(aslc_req & ASLC_REQ_MSK)) {
+		drm_dbg(&dev_priv->drm,
+			"No request on ASLC interrupt 0x%08x\n", aslc_req);
+		return;
+	}
+
+	if (aslc_req & ASLC_SET_ALS_ILLUM)
+		aslc_stat |= asle_set_als_illum(dev_priv, asle->alsi);
+
+	if (aslc_req & ASLC_SET_BACKLIGHT)
+		aslc_stat |= asle_set_backlight(dev_priv, asle->bclp);
+
+	if (aslc_req & ASLC_SET_PFIT)
+		aslc_stat |= asle_set_pfit(dev_priv, asle->pfit);
+
+	if (aslc_req & ASLC_SET_PWM_FREQ)
+		aslc_stat |= asle_set_pwm_freq(dev_priv, asle->pfmb);
+
+	if (aslc_req & ASLC_SUPPORTED_ROTATION_ANGLES)
+		aslc_stat |= asle_set_supported_rotation_angles(dev_priv,
+							asle->srot);
+
+	if (aslc_req & ASLC_BUTTON_ARRAY)
+		aslc_stat |= asle_set_button_array(dev_priv, asle->iuer);
+
+	if (aslc_req & ASLC_CONVERTIBLE_INDICATOR)
+		aslc_stat |= asle_set_convertible(dev_priv, asle->iuer);
+
+	if (aslc_req & ASLC_DOCKING_INDICATOR)
+		aslc_stat |= asle_set_docking(dev_priv, asle->iuer);
+
+	if (aslc_req & ASLC_ISCT_STATE_CHANGE)
+		aslc_stat |= asle_isct_state(dev_priv);
+
+	asle->aslc = aslc_stat;
+}
+
+void intel_opregion_asle_intr(struct drm_i915_private *dev_priv)
+{
+	if (dev_priv->display.opregion.asle)
+		schedule_work(&dev_priv->display.opregion.asle_work);
+}
+
+#define ACPI_EV_DISPLAY_SWITCH (1<<0)
+#define ACPI_EV_LID            (1<<1)
+#define ACPI_EV_DOCK           (1<<2)
+
+/*
+ * The only video events relevant to opregion are 0x80. These indicate either a
+ * docking event, lid switch or display switch request. In Linux, these are
+ * handled by the dock, button and video drivers.
+ */
+static int intel_opregion_video_event(struct notifier_block *nb,
+				      unsigned long val, void *data)
+{
+	struct intel_opregion *opregion = container_of(nb, struct intel_opregion,
+						       acpi_notifier);
+	struct acpi_bus_event *event = data;
+	struct opregion_acpi *acpi;
+	int ret = NOTIFY_OK;
+
+	if (strcmp(event->device_class, ACPI_VIDEO_CLASS) != 0)
+		return NOTIFY_DONE;
+
+	acpi = opregion->acpi;
+
+	if (event->type == 0x80 && ((acpi->cevt & 1) == 0))
+		ret = NOTIFY_BAD;
+
+	acpi->csts = 0;
+
+	return ret;
+}
+
+/*
+ * Initialise the DIDL field in opregion. This passes a list of devices to
+ * the firmware. Values are defined by section B.4.2 of the ACPI specification
+ * (version 3)
+ */
+
+static void set_did(struct intel_opregion *opregion, int i, u32 val)
+{
+	if (i < ARRAY_SIZE(opregion->acpi->didl)) {
+		opregion->acpi->didl[i] = val;
+	} else {
+		i -= ARRAY_SIZE(opregion->acpi->didl);
+
+		if (WARN_ON(i >= ARRAY_SIZE(opregion->acpi->did2)))
+			return;
+
+		opregion->acpi->did2[i] = val;
+	}
+}
+
+static void intel_didl_outputs(struct drm_i915_private *dev_priv)
+{
+	struct intel_opregion *opregion = &dev_priv->display.opregion;
+	struct intel_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	int i = 0, max_outputs;
+
+	/*
+	 * In theory, did2, the extended didl, gets added at opregion version
+	 * 3.0. In practice, however, we're supposed to set it for earlier
+	 * versions as well, since a BIOS that doesn't understand did2 should
+	 * not look at it anyway. Use a variable so we can tweak this if a need
+	 * arises later.
+	 */
+	max_outputs = ARRAY_SIZE(opregion->acpi->didl) +
+		ARRAY_SIZE(opregion->acpi->did2);
+
+	intel_acpi_device_id_update(dev_priv);
+
+	drm_connector_list_iter_begin(&dev_priv->drm, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		if (i < max_outputs)
+			set_did(opregion, i, connector->acpi_device_id);
+		i++;
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	drm_dbg_kms(&dev_priv->drm, "%d outputs detected\n", i);
+
+	if (i > max_outputs)
+		drm_err(&dev_priv->drm,
+			"More than %d outputs in connector list\n",
+			max_outputs);
+
+	/* If fewer than max outputs, the list must be null terminated */
+	if (i < max_outputs)
+		set_did(opregion, i, 0);
+}
+
+static void intel_setup_cadls(struct drm_i915_private *dev_priv)
+{
+	struct intel_opregion *opregion = &dev_priv->display.opregion;
+	struct intel_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+	int i = 0;
+
+	/*
+	 * Initialize the CADL field from the connector device ids. This is
+	 * essentially the same as copying from the DIDL. Technically, this is
+	 * not always correct as display outputs may exist, but not active. This
+	 * initialization is necessary for some Clevo laptops that check this
+	 * field before processing the brightness and display switching hotkeys.
+	 *
+	 * Note that internal panels should be at the front of the connector
+	 * list already, ensuring they're not left out.
+	 */
+	drm_connector_list_iter_begin(&dev_priv->drm, &conn_iter);
+	for_each_intel_connector_iter(connector, &conn_iter) {
+		if (i >= ARRAY_SIZE(opregion->acpi->cadl))
+			break;
+		opregion->acpi->cadl[i++] = connector->acpi_device_id;
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	/* If fewer than 8 active devices, the list must be null terminated */
+	if (i < ARRAY_SIZE(opregion->acpi->cadl))
+		opregion->acpi->cadl[i] = 0;
+}
+
+static void swsci_setup(struct drm_i915_private *dev_priv)
+{
+	struct intel_opregion *opregion = &dev_priv->display.opregion;
+	bool requested_callbacks = false;
+	u32 tmp;
+
+	/* Sub-function code 0 is okay, let's allow them. */
+	opregion->swsci_gbda_sub_functions = 1;
+	opregion->swsci_sbcb_sub_functions = 1;
+
+	/* We use GBDA to ask for supported GBDA calls. */
+	if (swsci(dev_priv, SWSCI_GBDA_SUPPORTED_CALLS, 0, &tmp) == 0) {
+		/* make the bits match the sub-function codes */
+		tmp <<= 1;
+		opregion->swsci_gbda_sub_functions |= tmp;
+	}
+
+	/*
+	 * We also use GBDA to ask for _requested_ SBCB callbacks. The driver
+	 * must not call interfaces that are not specifically requested by the
+	 * bios.
+	 */
+	if (swsci(dev_priv, SWSCI_GBDA_REQUESTED_CALLBACKS, 0, &tmp) == 0) {
+		/* here, the bits already match sub-function codes */
+		opregion->swsci_sbcb_sub_functions |= tmp;
+		requested_callbacks = true;
+	}
+
+	/*
+	 * But we use SBCB to ask for _supported_ SBCB calls. This does not mean
+	 * the callback is _requested_. But we still can't call interfaces that
+	 * are not requested.
+	 */
+	if (swsci(dev_priv, SWSCI_SBCB_SUPPORTED_CALLBACKS, 0, &tmp) == 0) {
+		/* make the bits match the sub-function codes */
+		u32 low = tmp & 0x7ff;
+		u32 high = tmp & ~0xfff; /* bit 11 is reserved */
+		tmp = (high << 4) | (low << 1) | 1;
+
+		/* best guess what to do with supported wrt requested */
+		if (requested_callbacks) {
+			u32 req = opregion->swsci_sbcb_sub_functions;
+			if ((req & tmp) != req)
+				drm_dbg(&dev_priv->drm,
+					"SWSCI BIOS requested (%08x) SBCB callbacks that are not supported (%08x)\n",
+					req, tmp);
+			/* XXX: for now, trust the requested callbacks */
+			/* opregion->swsci_sbcb_sub_functions &= tmp; */
+		} else {
+			opregion->swsci_sbcb_sub_functions |= tmp;
+		}
+	}
+
+	drm_dbg(&dev_priv->drm,
+		"SWSCI GBDA callbacks %08x, SBCB callbacks %08x\n",
+		opregion->swsci_gbda_sub_functions,
+		opregion->swsci_sbcb_sub_functions);
+}
+
+static int intel_no_opregion_vbt_callback(const struct dmi_system_id *id)
+{
+	DRM_DEBUG_KMS("Falling back to manually reading VBT from "
+		      "VBIOS ROM for %s\n", id->ident);
+	return 1;
+}
+
+static const struct dmi_system_id intel_no_opregion_vbt[] = {
+	{
+		.callback = intel_no_opregion_vbt_callback,
+		.ident = "ThinkCentre A57",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "LENOVO"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "97027RG"),
+		},
+	},
+	{ }
+};
+
+static int intel_load_vbt_firmware(struct drm_i915_private *dev_priv)
+{
+	struct intel_opregion *opregion = &dev_priv->display.opregion;
+	const struct firmware *fw = NULL;
+	const char *name = dev_priv->params.vbt_firmware;
+	int ret;
+
+	if (!name || !*name)
+		return -ENOENT;
+
+	ret = request_firmware(&fw, name, dev_priv->drm.dev);
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"Requesting VBT firmware \"%s\" failed (%d)\n",
+			name, ret);
+		return ret;
+	}
+
+	if (intel_bios_is_valid_vbt(fw->data, fw->size)) {
+		opregion->vbt_firmware = kmemdup(fw->data, fw->size, GFP_KERNEL);
+		if (opregion->vbt_firmware) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Found valid VBT firmware \"%s\"\n", name);
+			opregion->vbt = opregion->vbt_firmware;
+			opregion->vbt_size = fw->size;
+			ret = 0;
+		} else {
+			ret = -ENOMEM;
+		}
+	} else {
+		drm_dbg_kms(&dev_priv->drm, "Invalid VBT firmware \"%s\"\n",
+			    name);
+		ret = -EINVAL;
+	}
+
+	release_firmware(fw);
+
+	return ret;
+}
+
+int intel_opregion_setup(struct drm_i915_private *dev_priv)
+{
+	struct intel_opregion *opregion = &dev_priv->display.opregion;
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	u32 asls, mboxes;
+	char buf[sizeof(OPREGION_SIGNATURE)];
+	int err = 0;
+	void *base;
+	const void *vbt;
+	u32 vbt_size;
+
+	BUILD_BUG_ON(sizeof(struct opregion_header) != 0x100);
+	BUILD_BUG_ON(sizeof(struct opregion_acpi) != 0x100);
+	BUILD_BUG_ON(sizeof(struct opregion_swsci) != 0x100);
+	BUILD_BUG_ON(sizeof(struct opregion_asle) != 0x100);
+	BUILD_BUG_ON(sizeof(struct opregion_asle_ext) != 0x400);
+
+	pci_read_config_dword(pdev, ASLS, &asls);
+	drm_dbg(&dev_priv->drm, "graphic opregion physical addr: 0x%x\n",
+		asls);
+	if (asls == 0) {
+		drm_dbg(&dev_priv->drm, "ACPI OpRegion not supported!\n");
+		return -ENOTSUPP;
+	}
+
+	INIT_WORK(&opregion->asle_work, asle_work);
+
+	base = memremap(asls, OPREGION_SIZE, MEMREMAP_WB);
+	if (!base)
+		return -ENOMEM;
+
+	memcpy(buf, base, sizeof(buf));
+
+	if (memcmp(buf, OPREGION_SIGNATURE, 16)) {
+		drm_dbg(&dev_priv->drm, "opregion signature mismatch\n");
+		err = -EINVAL;
+		goto err_out;
+	}
+	opregion->header = base;
+	opregion->lid_state = base + ACPI_CLID;
+
+	drm_dbg(&dev_priv->drm, "ACPI OpRegion version %u.%u.%u\n",
+		opregion->header->over.major,
+		opregion->header->over.minor,
+		opregion->header->over.revision);
+
+	mboxes = opregion->header->mboxes;
+	if (mboxes & MBOX_ACPI) {
+		drm_dbg(&dev_priv->drm, "Public ACPI methods supported\n");
+		opregion->acpi = base + OPREGION_ACPI_OFFSET;
+		/*
+		 * Indicate we handle monitor hotplug events ourselves so we do
+		 * not need ACPI notifications for them. Disabling these avoids
+		 * triggering the AML code doing the notifation, which may be
+		 * broken as Windows also seems to disable these.
+		 */
+		opregion->acpi->chpd = 1;
+	}
+
+	if (mboxes & MBOX_SWSCI) {
+		u8 major = opregion->header->over.major;
+
+		if (major >= 3) {
+			drm_err(&dev_priv->drm, "SWSCI Mailbox #2 present for opregion v3.x, ignoring\n");
+		} else {
+			if (major >= 2)
+				drm_dbg(&dev_priv->drm, "SWSCI Mailbox #2 present for opregion v2.x\n");
+			drm_dbg(&dev_priv->drm, "SWSCI supported\n");
+			opregion->swsci = base + OPREGION_SWSCI_OFFSET;
+			swsci_setup(dev_priv);
+		}
+	}
+
+	if (mboxes & MBOX_ASLE) {
+		drm_dbg(&dev_priv->drm, "ASLE supported\n");
+		opregion->asle = base + OPREGION_ASLE_OFFSET;
+
+		opregion->asle->ardy = ASLE_ARDY_NOT_READY;
+	}
+
+	if (mboxes & MBOX_ASLE_EXT) {
+		drm_dbg(&dev_priv->drm, "ASLE extension supported\n");
+		opregion->asle_ext = base + OPREGION_ASLE_EXT_OFFSET;
+	}
+
+	if (mboxes & MBOX_BACKLIGHT) {
+		drm_dbg(&dev_priv->drm, "Mailbox #2 for backlight present\n");
+	}
+
+	if (intel_load_vbt_firmware(dev_priv) == 0)
+		goto out;
+
+	if (dmi_check_system(intel_no_opregion_vbt))
+		goto out;
+
+	if (opregion->header->over.major >= 2 && opregion->asle &&
+	    opregion->asle->rvda && opregion->asle->rvds) {
+		resource_size_t rvda = opregion->asle->rvda;
+
+		/*
+		 * opregion 2.0: rvda is the physical VBT address.
+		 *
+		 * opregion 2.1+: rvda is unsigned, relative offset from
+		 * opregion base, and should never point within opregion.
+		 */
+		if (opregion->header->over.major > 2 ||
+		    opregion->header->over.minor >= 1) {
+			drm_WARN_ON(&dev_priv->drm, rvda < OPREGION_SIZE);
+
+			rvda += asls;
+		}
+
+		opregion->rvda = memremap(rvda, opregion->asle->rvds,
+					  MEMREMAP_WB);
+
+		vbt = opregion->rvda;
+		vbt_size = opregion->asle->rvds;
+		if (intel_bios_is_valid_vbt(vbt, vbt_size)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Found valid VBT in ACPI OpRegion (RVDA)\n");
+			opregion->vbt = vbt;
+			opregion->vbt_size = vbt_size;
+			goto out;
+		} else {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Invalid VBT in ACPI OpRegion (RVDA)\n");
+			memunmap(opregion->rvda);
+			opregion->rvda = NULL;
+		}
+	}
+
+	vbt = base + OPREGION_VBT_OFFSET;
+	/*
+	 * The VBT specification says that if the ASLE ext mailbox is not used
+	 * its area is reserved, but on some CHT boards the VBT extends into the
+	 * ASLE ext area. Allow this even though it is against the spec, so we
+	 * do not end up rejecting the VBT on those boards (and end up not
+	 * finding the LCD panel because of this).
+	 */
+	vbt_size = (mboxes & MBOX_ASLE_EXT) ?
+		OPREGION_ASLE_EXT_OFFSET : OPREGION_SIZE;
+	vbt_size -= OPREGION_VBT_OFFSET;
+	if (intel_bios_is_valid_vbt(vbt, vbt_size)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Found valid VBT in ACPI OpRegion (Mailbox #4)\n");
+		opregion->vbt = vbt;
+		opregion->vbt_size = vbt_size;
+	} else {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Invalid VBT in ACPI OpRegion (Mailbox #4)\n");
+	}
+
+out:
+	return 0;
+
+err_out:
+	memunmap(base);
+	return err;
+}
+
+static int intel_use_opregion_panel_type_callback(const struct dmi_system_id *id)
+{
+	DRM_INFO("Using panel type from OpRegion on %s\n", id->ident);
+	return 1;
+}
+
+static const struct dmi_system_id intel_use_opregion_panel_type[] = {
+	{
+		.callback = intel_use_opregion_panel_type_callback,
+		.ident = "Conrac GmbH IX45GM2",
+		.matches = {DMI_MATCH(DMI_SYS_VENDOR, "Conrac GmbH"),
+			    DMI_MATCH(DMI_PRODUCT_NAME, "IX45GM2"),
+		},
+	},
+	{ }
+};
+
+int
+intel_opregion_get_panel_type(struct drm_i915_private *dev_priv)
+{
+	u32 panel_details;
+	int ret;
+
+	ret = swsci(dev_priv, SWSCI_GBDA_PANEL_DETAILS, 0x0, &panel_details);
+	if (ret)
+		return ret;
+
+	ret = (panel_details >> 8) & 0xff;
+	if (ret > 0x10) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Invalid OpRegion panel type 0x%x\n", ret);
+		return -EINVAL;
+	}
+
+	/* fall back to VBT panel type? */
+	if (ret == 0x0) {
+		drm_dbg_kms(&dev_priv->drm, "No panel type in OpRegion\n");
+		return -ENODEV;
+	}
+
+	/*
+	 * So far we know that some machined must use it, others must not use it.
+	 * There doesn't seem to be any way to determine which way to go, except
+	 * via a quirk list :(
+	 */
+	if (!dmi_check_system(intel_use_opregion_panel_type)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Ignoring OpRegion panel type (%d)\n", ret - 1);
+		return -ENODEV;
+	}
+
+	return ret - 1;
+}
+
+/**
+ * intel_opregion_get_edid - Fetch EDID from ACPI OpRegion mailbox #5
+ * @intel_connector: eDP connector
+ *
+ * This reads the ACPI Opregion mailbox #5 to extract the EDID that is passed
+ * to it.
+ *
+ * Returns:
+ * The EDID in the OpRegion, or NULL if there is none or it's invalid.
+ *
+ */
+struct edid *intel_opregion_get_edid(struct intel_connector *intel_connector)
+{
+	struct drm_connector *connector = &intel_connector->base;
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct intel_opregion *opregion = &i915->display.opregion;
+	const void *in_edid;
+	const struct edid *edid;
+	struct edid *new_edid;
+	int len;
+
+	if (!opregion->asle_ext)
+		return NULL;
+
+	in_edid = opregion->asle_ext->bddc;
+
+	/* Validity corresponds to number of 128-byte blocks */
+	len = (opregion->asle_ext->phed & ASLE_PHED_EDID_VALID_MASK) * 128;
+	if (!len || !memchr_inv(in_edid, 0, len))
+		return NULL;
+
+	edid = in_edid;
+
+	if (len < EDID_LENGTH * (1 + edid->extensions)) {
+		drm_dbg_kms(&i915->drm, "Invalid EDID in ACPI OpRegion (Mailbox #5): too short\n");
+		return NULL;
+	}
+	new_edid = drm_edid_duplicate(edid);
+	if (!new_edid)
+		return NULL;
+	if (!drm_edid_is_valid(new_edid)) {
+		kfree(new_edid);
+		drm_dbg_kms(&i915->drm, "Invalid EDID in ACPI OpRegion (Mailbox #5)\n");
+		return NULL;
+	}
+	return new_edid;
+}
+
+bool intel_opregion_headless_sku(struct drm_i915_private *i915)
+{
+	struct intel_opregion *opregion = &i915->display.opregion;
+	struct opregion_header *header = opregion->header;
+
+	if (!header || header->over.major < 2 ||
+	    (header->over.major == 2 && header->over.minor < 3))
+		return false;
+
+	return opregion->header->pcon & PCON_HEADLESS_SKU;
+}
+
+void intel_opregion_register(struct drm_i915_private *i915)
+{
+	struct intel_opregion *opregion = &i915->display.opregion;
+
+	if (!opregion->header)
+		return;
+
+	if (opregion->acpi) {
+		opregion->acpi_notifier.notifier_call =
+			intel_opregion_video_event;
+		register_acpi_notifier(&opregion->acpi_notifier);
+	}
+
+	intel_opregion_resume(i915);
+}
+
+void intel_opregion_resume(struct drm_i915_private *i915)
+{
+	struct intel_opregion *opregion = &i915->display.opregion;
+
+	if (!opregion->header)
+		return;
+
+	if (opregion->acpi) {
+		intel_didl_outputs(i915);
+		intel_setup_cadls(i915);
+
+		/*
+		 * Notify BIOS we are ready to handle ACPI video ext notifs.
+		 * Right now, all the events are handled by the ACPI video
+		 * module. We don't actually need to do anything with them.
+		 */
+		opregion->acpi->csts = 0;
+		opregion->acpi->drdy = 1;
+	}
+
+	if (opregion->asle) {
+		opregion->asle->tche = ASLE_TCHE_BLC_EN;
+		opregion->asle->ardy = ASLE_ARDY_READY;
+	}
+
+	/* Some platforms abuse the _DSM to enable MUX */
+	intel_dsm_get_bios_data_funcs_supported(i915);
+
+	intel_opregion_notify_adapter(i915, PCI_D0);
+}
+
+void intel_opregion_suspend(struct drm_i915_private *i915, pci_power_t state)
+{
+	struct intel_opregion *opregion = &i915->display.opregion;
+
+	if (!opregion->header)
+		return;
+
+	intel_opregion_notify_adapter(i915, state);
+
+	if (opregion->asle)
+		opregion->asle->ardy = ASLE_ARDY_NOT_READY;
+
+	cancel_work_sync(&i915->display.opregion.asle_work);
+
+	if (opregion->acpi)
+		opregion->acpi->drdy = 0;
+}
+
+void intel_opregion_unregister(struct drm_i915_private *i915)
+{
+	struct intel_opregion *opregion = &i915->display.opregion;
+
+	intel_opregion_suspend(i915, PCI_D1);
+
+	if (!opregion->header)
+		return;
+
+	if (opregion->acpi_notifier.notifier_call) {
+		unregister_acpi_notifier(&opregion->acpi_notifier);
+		opregion->acpi_notifier.notifier_call = NULL;
+	}
+
+	/* just clear all opregion memory pointers now */
+	memunmap(opregion->header);
+	if (opregion->rvda) {
+		memunmap(opregion->rvda);
+		opregion->rvda = NULL;
+	}
+	if (opregion->vbt_firmware) {
+		kfree(opregion->vbt_firmware);
+		opregion->vbt_firmware = NULL;
+	}
+	opregion->header = NULL;
+	opregion->acpi = NULL;
+	opregion->swsci = NULL;
+	opregion->asle = NULL;
+	opregion->asle_ext = NULL;
+	opregion->vbt = NULL;
+	opregion->lid_state = NULL;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_opregion.h b/drivers/gpu/drm/i915/display/intel_opregion.h
new file mode 100644
index 000000000..2f261f985
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_opregion.h
@@ -0,0 +1,139 @@
+/*
+ * Copyright © 2008-2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _INTEL_OPREGION_H_
+#define _INTEL_OPREGION_H_
+
+#include <linux/workqueue.h>
+#include <linux/pci.h>
+
+struct drm_i915_private;
+struct intel_connector;
+struct intel_encoder;
+
+struct opregion_header;
+struct opregion_acpi;
+struct opregion_swsci;
+struct opregion_asle;
+struct opregion_asle_ext;
+
+struct intel_opregion {
+	struct opregion_header *header;
+	struct opregion_acpi *acpi;
+	struct opregion_swsci *swsci;
+	u32 swsci_gbda_sub_functions;
+	u32 swsci_sbcb_sub_functions;
+	struct opregion_asle *asle;
+	struct opregion_asle_ext *asle_ext;
+	void *rvda;
+	void *vbt_firmware;
+	const void *vbt;
+	u32 vbt_size;
+	u32 *lid_state;
+	struct work_struct asle_work;
+	struct notifier_block acpi_notifier;
+};
+
+#define OPREGION_SIZE            (8 * 1024)
+
+#ifdef CONFIG_ACPI
+
+int intel_opregion_setup(struct drm_i915_private *dev_priv);
+
+void intel_opregion_register(struct drm_i915_private *dev_priv);
+void intel_opregion_unregister(struct drm_i915_private *dev_priv);
+
+void intel_opregion_resume(struct drm_i915_private *dev_priv);
+void intel_opregion_suspend(struct drm_i915_private *dev_priv,
+			    pci_power_t state);
+
+void intel_opregion_asle_intr(struct drm_i915_private *dev_priv);
+int intel_opregion_notify_encoder(struct intel_encoder *intel_encoder,
+				  bool enable);
+int intel_opregion_notify_adapter(struct drm_i915_private *dev_priv,
+				  pci_power_t state);
+int intel_opregion_get_panel_type(struct drm_i915_private *dev_priv);
+struct edid *intel_opregion_get_edid(struct intel_connector *connector);
+
+bool intel_opregion_headless_sku(struct drm_i915_private *i915);
+
+#else /* CONFIG_ACPI*/
+
+static inline int intel_opregion_setup(struct drm_i915_private *dev_priv)
+{
+	return 0;
+}
+
+static inline void intel_opregion_register(struct drm_i915_private *dev_priv)
+{
+}
+
+static inline void intel_opregion_unregister(struct drm_i915_private *dev_priv)
+{
+}
+
+static inline void intel_opregion_resume(struct drm_i915_private *dev_priv)
+{
+}
+
+static inline void intel_opregion_suspend(struct drm_i915_private *dev_priv,
+					  pci_power_t state)
+{
+}
+
+static inline void intel_opregion_asle_intr(struct drm_i915_private *dev_priv)
+{
+}
+
+static inline int
+intel_opregion_notify_encoder(struct intel_encoder *intel_encoder, bool enable)
+{
+	return 0;
+}
+
+static inline int
+intel_opregion_notify_adapter(struct drm_i915_private *dev, pci_power_t state)
+{
+	return 0;
+}
+
+static inline int intel_opregion_get_panel_type(struct drm_i915_private *dev)
+{
+	return -ENODEV;
+}
+
+static inline struct edid *
+intel_opregion_get_edid(struct intel_connector *connector)
+{
+	return NULL;
+}
+
+static inline bool intel_opregion_headless_sku(struct drm_i915_private *i915)
+{
+	return false;
+}
+
+#endif /* CONFIG_ACPI */
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_overlay.c b/drivers/gpu/drm/i915/display/intel_overlay.c
new file mode 100644
index 000000000..c12bdca8d
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_overlay.c
@@ -0,0 +1,1533 @@
+/*
+ * Copyright © 2009
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Daniel Vetter <daniel@ffwll.ch>
+ *
+ * Derived from Xorg ddx, xf86-video-intel, src/i830_video.c
+ */
+
+#include <drm/drm_fourcc.h>
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_ring.h"
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_frontbuffer.h"
+#include "intel_overlay.h"
+#include "intel_pci_config.h"
+
+/* Limits for overlay size. According to intel doc, the real limits are:
+ * Y width: 4095, UV width (planar): 2047, Y height: 2047,
+ * UV width (planar): * 1023. But the xorg thinks 2048 for height and width. Use
+ * the mininum of both.  */
+#define IMAGE_MAX_WIDTH		2048
+#define IMAGE_MAX_HEIGHT	2046 /* 2 * 1023 */
+/* on 830 and 845 these large limits result in the card hanging */
+#define IMAGE_MAX_WIDTH_LEGACY	1024
+#define IMAGE_MAX_HEIGHT_LEGACY	1088
+
+/* overlay register definitions */
+/* OCMD register */
+#define OCMD_TILED_SURFACE	(0x1<<19)
+#define OCMD_MIRROR_MASK	(0x3<<17)
+#define OCMD_MIRROR_MODE	(0x3<<17)
+#define OCMD_MIRROR_HORIZONTAL	(0x1<<17)
+#define OCMD_MIRROR_VERTICAL	(0x2<<17)
+#define OCMD_MIRROR_BOTH	(0x3<<17)
+#define OCMD_BYTEORDER_MASK	(0x3<<14) /* zero for YUYV or FOURCC YUY2 */
+#define OCMD_UV_SWAP		(0x1<<14) /* YVYU */
+#define OCMD_Y_SWAP		(0x2<<14) /* UYVY or FOURCC UYVY */
+#define OCMD_Y_AND_UV_SWAP	(0x3<<14) /* VYUY */
+#define OCMD_SOURCE_FORMAT_MASK (0xf<<10)
+#define OCMD_RGB_888		(0x1<<10) /* not in i965 Intel docs */
+#define OCMD_RGB_555		(0x2<<10) /* not in i965 Intel docs */
+#define OCMD_RGB_565		(0x3<<10) /* not in i965 Intel docs */
+#define OCMD_YUV_422_PACKED	(0x8<<10)
+#define OCMD_YUV_411_PACKED	(0x9<<10) /* not in i965 Intel docs */
+#define OCMD_YUV_420_PLANAR	(0xc<<10)
+#define OCMD_YUV_422_PLANAR	(0xd<<10)
+#define OCMD_YUV_410_PLANAR	(0xe<<10) /* also 411 */
+#define OCMD_TVSYNCFLIP_PARITY	(0x1<<9)
+#define OCMD_TVSYNCFLIP_ENABLE	(0x1<<7)
+#define OCMD_BUF_TYPE_MASK	(0x1<<5)
+#define OCMD_BUF_TYPE_FRAME	(0x0<<5)
+#define OCMD_BUF_TYPE_FIELD	(0x1<<5)
+#define OCMD_TEST_MODE		(0x1<<4)
+#define OCMD_BUFFER_SELECT	(0x3<<2)
+#define OCMD_BUFFER0		(0x0<<2)
+#define OCMD_BUFFER1		(0x1<<2)
+#define OCMD_FIELD_SELECT	(0x1<<2)
+#define OCMD_FIELD0		(0x0<<1)
+#define OCMD_FIELD1		(0x1<<1)
+#define OCMD_ENABLE		(0x1<<0)
+
+/* OCONFIG register */
+#define OCONF_PIPE_MASK		(0x1<<18)
+#define OCONF_PIPE_A		(0x0<<18)
+#define OCONF_PIPE_B		(0x1<<18)
+#define OCONF_GAMMA2_ENABLE	(0x1<<16)
+#define OCONF_CSC_MODE_BT601	(0x0<<5)
+#define OCONF_CSC_MODE_BT709	(0x1<<5)
+#define OCONF_CSC_BYPASS	(0x1<<4)
+#define OCONF_CC_OUT_8BIT	(0x1<<3)
+#define OCONF_TEST_MODE		(0x1<<2)
+#define OCONF_THREE_LINE_BUFFER	(0x1<<0)
+#define OCONF_TWO_LINE_BUFFER	(0x0<<0)
+
+/* DCLRKM (dst-key) register */
+#define DST_KEY_ENABLE		(0x1<<31)
+#define CLK_RGB24_MASK		0x0
+#define CLK_RGB16_MASK		0x070307
+#define CLK_RGB15_MASK		0x070707
+
+#define RGB30_TO_COLORKEY(c) \
+	((((c) & 0x3fc00000) >> 6) | (((c) & 0x000ff000) >> 4) | (((c) & 0x000003fc) >> 2))
+#define RGB16_TO_COLORKEY(c) \
+	((((c) & 0xf800) << 8) | (((c) & 0x07e0) << 5) | (((c) & 0x001f) << 3))
+#define RGB15_TO_COLORKEY(c) \
+	((((c) & 0x7c00) << 9) | (((c) & 0x03e0) << 6) | (((c) & 0x001f) << 3))
+#define RGB8I_TO_COLORKEY(c) \
+	((((c) & 0xff) << 16) | (((c) & 0xff) << 8) | (((c) & 0xff) << 0))
+
+/* overlay flip addr flag */
+#define OFC_UPDATE		0x1
+
+/* polyphase filter coefficients */
+#define N_HORIZ_Y_TAPS          5
+#define N_VERT_Y_TAPS           3
+#define N_HORIZ_UV_TAPS         3
+#define N_VERT_UV_TAPS          3
+#define N_PHASES                17
+#define MAX_TAPS                5
+
+/* memory bufferd overlay registers */
+struct overlay_registers {
+	u32 OBUF_0Y;
+	u32 OBUF_1Y;
+	u32 OBUF_0U;
+	u32 OBUF_0V;
+	u32 OBUF_1U;
+	u32 OBUF_1V;
+	u32 OSTRIDE;
+	u32 YRGB_VPH;
+	u32 UV_VPH;
+	u32 HORZ_PH;
+	u32 INIT_PHS;
+	u32 DWINPOS;
+	u32 DWINSZ;
+	u32 SWIDTH;
+	u32 SWIDTHSW;
+	u32 SHEIGHT;
+	u32 YRGBSCALE;
+	u32 UVSCALE;
+	u32 OCLRC0;
+	u32 OCLRC1;
+	u32 DCLRKV;
+	u32 DCLRKM;
+	u32 SCLRKVH;
+	u32 SCLRKVL;
+	u32 SCLRKEN;
+	u32 OCONFIG;
+	u32 OCMD;
+	u32 RESERVED1; /* 0x6C */
+	u32 OSTART_0Y;
+	u32 OSTART_1Y;
+	u32 OSTART_0U;
+	u32 OSTART_0V;
+	u32 OSTART_1U;
+	u32 OSTART_1V;
+	u32 OTILEOFF_0Y;
+	u32 OTILEOFF_1Y;
+	u32 OTILEOFF_0U;
+	u32 OTILEOFF_0V;
+	u32 OTILEOFF_1U;
+	u32 OTILEOFF_1V;
+	u32 FASTHSCALE; /* 0xA0 */
+	u32 UVSCALEV; /* 0xA4 */
+	u32 RESERVEDC[(0x200 - 0xA8) / 4]; /* 0xA8 - 0x1FC */
+	u16 Y_VCOEFS[N_VERT_Y_TAPS * N_PHASES]; /* 0x200 */
+	u16 RESERVEDD[0x100 / 2 - N_VERT_Y_TAPS * N_PHASES];
+	u16 Y_HCOEFS[N_HORIZ_Y_TAPS * N_PHASES]; /* 0x300 */
+	u16 RESERVEDE[0x200 / 2 - N_HORIZ_Y_TAPS * N_PHASES];
+	u16 UV_VCOEFS[N_VERT_UV_TAPS * N_PHASES]; /* 0x500 */
+	u16 RESERVEDF[0x100 / 2 - N_VERT_UV_TAPS * N_PHASES];
+	u16 UV_HCOEFS[N_HORIZ_UV_TAPS * N_PHASES]; /* 0x600 */
+	u16 RESERVEDG[0x100 / 2 - N_HORIZ_UV_TAPS * N_PHASES];
+};
+
+struct intel_overlay {
+	struct drm_i915_private *i915;
+	struct intel_context *context;
+	struct intel_crtc *crtc;
+	struct i915_vma *vma;
+	struct i915_vma *old_vma;
+	struct intel_frontbuffer *frontbuffer;
+	bool active;
+	bool pfit_active;
+	u32 pfit_vscale_ratio; /* shifted-point number, (1<<12) == 1.0 */
+	u32 color_key:24;
+	u32 color_key_enabled:1;
+	u32 brightness, contrast, saturation;
+	u32 old_xscale, old_yscale;
+	/* register access */
+	struct drm_i915_gem_object *reg_bo;
+	struct overlay_registers __iomem *regs;
+	u32 flip_addr;
+	/* flip handling */
+	struct i915_active last_flip;
+	void (*flip_complete)(struct intel_overlay *ovl);
+};
+
+static void i830_overlay_clock_gating(struct drm_i915_private *dev_priv,
+				      bool enable)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	u8 val;
+
+	/* WA_OVERLAY_CLKGATE:alm */
+	if (enable)
+		intel_de_write(dev_priv, DSPCLK_GATE_D(dev_priv), 0);
+	else
+		intel_de_write(dev_priv, DSPCLK_GATE_D(dev_priv),
+			       OVRUNIT_CLOCK_GATE_DISABLE);
+
+	/* WA_DISABLE_L2CACHE_CLOCK_GATING:alm */
+	pci_bus_read_config_byte(pdev->bus,
+				 PCI_DEVFN(0, 0), I830_CLOCK_GATE, &val);
+	if (enable)
+		val &= ~I830_L2_CACHE_CLOCK_GATE_DISABLE;
+	else
+		val |= I830_L2_CACHE_CLOCK_GATE_DISABLE;
+	pci_bus_write_config_byte(pdev->bus,
+				  PCI_DEVFN(0, 0), I830_CLOCK_GATE, val);
+}
+
+static struct i915_request *
+alloc_request(struct intel_overlay *overlay, void (*fn)(struct intel_overlay *))
+{
+	struct i915_request *rq;
+	int err;
+
+	overlay->flip_complete = fn;
+
+	rq = i915_request_create(overlay->context);
+	if (IS_ERR(rq))
+		return rq;
+
+	err = i915_active_add_request(&overlay->last_flip, rq);
+	if (err) {
+		i915_request_add(rq);
+		return ERR_PTR(err);
+	}
+
+	return rq;
+}
+
+/* overlay needs to be disable in OCMD reg */
+static int intel_overlay_on(struct intel_overlay *overlay)
+{
+	struct drm_i915_private *dev_priv = overlay->i915;
+	struct i915_request *rq;
+	u32 *cs;
+
+	drm_WARN_ON(&dev_priv->drm, overlay->active);
+
+	rq = alloc_request(overlay, NULL);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	overlay->active = true;
+
+	if (IS_I830(dev_priv))
+		i830_overlay_clock_gating(dev_priv, false);
+
+	*cs++ = MI_OVERLAY_FLIP | MI_OVERLAY_ON;
+	*cs++ = overlay->flip_addr | OFC_UPDATE;
+	*cs++ = MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	i915_request_add(rq);
+
+	return i915_active_wait(&overlay->last_flip);
+}
+
+static void intel_overlay_flip_prepare(struct intel_overlay *overlay,
+				       struct i915_vma *vma)
+{
+	enum pipe pipe = overlay->crtc->pipe;
+	struct intel_frontbuffer *frontbuffer = NULL;
+
+	drm_WARN_ON(&overlay->i915->drm, overlay->old_vma);
+
+	if (vma)
+		frontbuffer = intel_frontbuffer_get(vma->obj);
+
+	intel_frontbuffer_track(overlay->frontbuffer, frontbuffer,
+				INTEL_FRONTBUFFER_OVERLAY(pipe));
+
+	if (overlay->frontbuffer)
+		intel_frontbuffer_put(overlay->frontbuffer);
+	overlay->frontbuffer = frontbuffer;
+
+	intel_frontbuffer_flip_prepare(overlay->i915,
+				       INTEL_FRONTBUFFER_OVERLAY(pipe));
+
+	overlay->old_vma = overlay->vma;
+	if (vma)
+		overlay->vma = i915_vma_get(vma);
+	else
+		overlay->vma = NULL;
+}
+
+/* overlay needs to be enabled in OCMD reg */
+static int intel_overlay_continue(struct intel_overlay *overlay,
+				  struct i915_vma *vma,
+				  bool load_polyphase_filter)
+{
+	struct drm_i915_private *dev_priv = overlay->i915;
+	struct i915_request *rq;
+	u32 flip_addr = overlay->flip_addr;
+	u32 tmp, *cs;
+
+	drm_WARN_ON(&dev_priv->drm, !overlay->active);
+
+	if (load_polyphase_filter)
+		flip_addr |= OFC_UPDATE;
+
+	/* check for underruns */
+	tmp = intel_de_read(dev_priv, DOVSTA);
+	if (tmp & (1 << 17))
+		drm_dbg(&dev_priv->drm, "overlay underrun, DOVSTA: %x\n", tmp);
+
+	rq = alloc_request(overlay, NULL);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	*cs++ = MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE;
+	*cs++ = flip_addr;
+	intel_ring_advance(rq, cs);
+
+	intel_overlay_flip_prepare(overlay, vma);
+	i915_request_add(rq);
+
+	return 0;
+}
+
+static void intel_overlay_release_old_vma(struct intel_overlay *overlay)
+{
+	struct i915_vma *vma;
+
+	vma = fetch_and_zero(&overlay->old_vma);
+	if (drm_WARN_ON(&overlay->i915->drm, !vma))
+		return;
+
+	intel_frontbuffer_flip_complete(overlay->i915,
+					INTEL_FRONTBUFFER_OVERLAY(overlay->crtc->pipe));
+
+	i915_vma_unpin(vma);
+	i915_vma_put(vma);
+}
+
+static void
+intel_overlay_release_old_vid_tail(struct intel_overlay *overlay)
+{
+	intel_overlay_release_old_vma(overlay);
+}
+
+static void intel_overlay_off_tail(struct intel_overlay *overlay)
+{
+	struct drm_i915_private *dev_priv = overlay->i915;
+
+	intel_overlay_release_old_vma(overlay);
+
+	overlay->crtc->overlay = NULL;
+	overlay->crtc = NULL;
+	overlay->active = false;
+
+	if (IS_I830(dev_priv))
+		i830_overlay_clock_gating(dev_priv, true);
+}
+
+static void intel_overlay_last_flip_retire(struct i915_active *active)
+{
+	struct intel_overlay *overlay =
+		container_of(active, typeof(*overlay), last_flip);
+
+	if (overlay->flip_complete)
+		overlay->flip_complete(overlay);
+}
+
+/* overlay needs to be disabled in OCMD reg */
+static int intel_overlay_off(struct intel_overlay *overlay)
+{
+	struct i915_request *rq;
+	u32 *cs, flip_addr = overlay->flip_addr;
+
+	drm_WARN_ON(&overlay->i915->drm, !overlay->active);
+
+	/* According to intel docs the overlay hw may hang (when switching
+	 * off) without loading the filter coeffs. It is however unclear whether
+	 * this applies to the disabling of the overlay or to the switching off
+	 * of the hw. Do it in both cases */
+	flip_addr |= OFC_UPDATE;
+
+	rq = alloc_request(overlay, intel_overlay_off_tail);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 6);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	/* wait for overlay to go idle */
+	*cs++ = MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE;
+	*cs++ = flip_addr;
+	*cs++ = MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP;
+
+	/* turn overlay off */
+	*cs++ = MI_OVERLAY_FLIP | MI_OVERLAY_OFF;
+	*cs++ = flip_addr;
+	*cs++ = MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP;
+
+	intel_ring_advance(rq, cs);
+
+	intel_overlay_flip_prepare(overlay, NULL);
+	i915_request_add(rq);
+
+	return i915_active_wait(&overlay->last_flip);
+}
+
+/* recover from an interruption due to a signal
+ * We have to be careful not to repeat work forever an make forward progess. */
+static int intel_overlay_recover_from_interrupt(struct intel_overlay *overlay)
+{
+	return i915_active_wait(&overlay->last_flip);
+}
+
+/* Wait for pending overlay flip and release old frame.
+ * Needs to be called before the overlay register are changed
+ * via intel_overlay_(un)map_regs
+ */
+static int intel_overlay_release_old_vid(struct intel_overlay *overlay)
+{
+	struct drm_i915_private *dev_priv = overlay->i915;
+	struct i915_request *rq;
+	u32 *cs;
+
+	/*
+	 * Only wait if there is actually an old frame to release to
+	 * guarantee forward progress.
+	 */
+	if (!overlay->old_vma)
+		return 0;
+
+	if (!(intel_de_read(dev_priv, GEN2_ISR) & I915_OVERLAY_PLANE_FLIP_PENDING_INTERRUPT)) {
+		intel_overlay_release_old_vid_tail(overlay);
+		return 0;
+	}
+
+	rq = alloc_request(overlay, intel_overlay_release_old_vid_tail);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	*cs++ = MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	i915_request_add(rq);
+
+	return i915_active_wait(&overlay->last_flip);
+}
+
+void intel_overlay_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_overlay *overlay = dev_priv->display.overlay;
+
+	if (!overlay)
+		return;
+
+	overlay->old_xscale = 0;
+	overlay->old_yscale = 0;
+	overlay->crtc = NULL;
+	overlay->active = false;
+}
+
+static int packed_depth_bytes(u32 format)
+{
+	switch (format & I915_OVERLAY_DEPTH_MASK) {
+	case I915_OVERLAY_YUV422:
+		return 4;
+	case I915_OVERLAY_YUV411:
+		/* return 6; not implemented */
+	default:
+		return -EINVAL;
+	}
+}
+
+static int packed_width_bytes(u32 format, short width)
+{
+	switch (format & I915_OVERLAY_DEPTH_MASK) {
+	case I915_OVERLAY_YUV422:
+		return width << 1;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int uv_hsubsampling(u32 format)
+{
+	switch (format & I915_OVERLAY_DEPTH_MASK) {
+	case I915_OVERLAY_YUV422:
+	case I915_OVERLAY_YUV420:
+		return 2;
+	case I915_OVERLAY_YUV411:
+	case I915_OVERLAY_YUV410:
+		return 4;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int uv_vsubsampling(u32 format)
+{
+	switch (format & I915_OVERLAY_DEPTH_MASK) {
+	case I915_OVERLAY_YUV420:
+	case I915_OVERLAY_YUV410:
+		return 2;
+	case I915_OVERLAY_YUV422:
+	case I915_OVERLAY_YUV411:
+		return 1;
+	default:
+		return -EINVAL;
+	}
+}
+
+static u32 calc_swidthsw(struct drm_i915_private *dev_priv, u32 offset, u32 width)
+{
+	u32 sw;
+
+	if (DISPLAY_VER(dev_priv) == 2)
+		sw = ALIGN((offset & 31) + width, 32);
+	else
+		sw = ALIGN((offset & 63) + width, 64);
+
+	if (sw == 0)
+		return 0;
+
+	return (sw - 32) >> 3;
+}
+
+static const u16 y_static_hcoeffs[N_PHASES][N_HORIZ_Y_TAPS] = {
+	[ 0] = { 0x3000, 0xb4a0, 0x1930, 0x1920, 0xb4a0, },
+	[ 1] = { 0x3000, 0xb500, 0x19d0, 0x1880, 0xb440, },
+	[ 2] = { 0x3000, 0xb540, 0x1a88, 0x2f80, 0xb3e0, },
+	[ 3] = { 0x3000, 0xb580, 0x1b30, 0x2e20, 0xb380, },
+	[ 4] = { 0x3000, 0xb5c0, 0x1bd8, 0x2cc0, 0xb320, },
+	[ 5] = { 0x3020, 0xb5e0, 0x1c60, 0x2b80, 0xb2c0, },
+	[ 6] = { 0x3020, 0xb5e0, 0x1cf8, 0x2a20, 0xb260, },
+	[ 7] = { 0x3020, 0xb5e0, 0x1d80, 0x28e0, 0xb200, },
+	[ 8] = { 0x3020, 0xb5c0, 0x1e08, 0x3f40, 0xb1c0, },
+	[ 9] = { 0x3020, 0xb580, 0x1e78, 0x3ce0, 0xb160, },
+	[10] = { 0x3040, 0xb520, 0x1ed8, 0x3aa0, 0xb120, },
+	[11] = { 0x3040, 0xb4a0, 0x1f30, 0x3880, 0xb0e0, },
+	[12] = { 0x3040, 0xb400, 0x1f78, 0x3680, 0xb0a0, },
+	[13] = { 0x3020, 0xb340, 0x1fb8, 0x34a0, 0xb060, },
+	[14] = { 0x3020, 0xb240, 0x1fe0, 0x32e0, 0xb040, },
+	[15] = { 0x3020, 0xb140, 0x1ff8, 0x3160, 0xb020, },
+	[16] = { 0xb000, 0x3000, 0x0800, 0x3000, 0xb000, },
+};
+
+static const u16 uv_static_hcoeffs[N_PHASES][N_HORIZ_UV_TAPS] = {
+	[ 0] = { 0x3000, 0x1800, 0x1800, },
+	[ 1] = { 0xb000, 0x18d0, 0x2e60, },
+	[ 2] = { 0xb000, 0x1990, 0x2ce0, },
+	[ 3] = { 0xb020, 0x1a68, 0x2b40, },
+	[ 4] = { 0xb040, 0x1b20, 0x29e0, },
+	[ 5] = { 0xb060, 0x1bd8, 0x2880, },
+	[ 6] = { 0xb080, 0x1c88, 0x3e60, },
+	[ 7] = { 0xb0a0, 0x1d28, 0x3c00, },
+	[ 8] = { 0xb0c0, 0x1db8, 0x39e0, },
+	[ 9] = { 0xb0e0, 0x1e40, 0x37e0, },
+	[10] = { 0xb100, 0x1eb8, 0x3620, },
+	[11] = { 0xb100, 0x1f18, 0x34a0, },
+	[12] = { 0xb100, 0x1f68, 0x3360, },
+	[13] = { 0xb0e0, 0x1fa8, 0x3240, },
+	[14] = { 0xb0c0, 0x1fe0, 0x3140, },
+	[15] = { 0xb060, 0x1ff0, 0x30a0, },
+	[16] = { 0x3000, 0x0800, 0x3000, },
+};
+
+static void update_polyphase_filter(struct overlay_registers __iomem *regs)
+{
+	memcpy_toio(regs->Y_HCOEFS, y_static_hcoeffs, sizeof(y_static_hcoeffs));
+	memcpy_toio(regs->UV_HCOEFS, uv_static_hcoeffs,
+		    sizeof(uv_static_hcoeffs));
+}
+
+static bool update_scaling_factors(struct intel_overlay *overlay,
+				   struct overlay_registers __iomem *regs,
+				   struct drm_intel_overlay_put_image *params)
+{
+	/* fixed point with a 12 bit shift */
+	u32 xscale, yscale, xscale_UV, yscale_UV;
+#define FP_SHIFT 12
+#define FRACT_MASK 0xfff
+	bool scale_changed = false;
+	int uv_hscale = uv_hsubsampling(params->flags);
+	int uv_vscale = uv_vsubsampling(params->flags);
+
+	if (params->dst_width > 1)
+		xscale = ((params->src_scan_width - 1) << FP_SHIFT) /
+			params->dst_width;
+	else
+		xscale = 1 << FP_SHIFT;
+
+	if (params->dst_height > 1)
+		yscale = ((params->src_scan_height - 1) << FP_SHIFT) /
+			params->dst_height;
+	else
+		yscale = 1 << FP_SHIFT;
+
+	/*if (params->format & I915_OVERLAY_YUV_PLANAR) {*/
+	xscale_UV = xscale/uv_hscale;
+	yscale_UV = yscale/uv_vscale;
+	/* make the Y scale to UV scale ratio an exact multiply */
+	xscale = xscale_UV * uv_hscale;
+	yscale = yscale_UV * uv_vscale;
+	/*} else {
+	  xscale_UV = 0;
+	  yscale_UV = 0;
+	  }*/
+
+	if (xscale != overlay->old_xscale || yscale != overlay->old_yscale)
+		scale_changed = true;
+	overlay->old_xscale = xscale;
+	overlay->old_yscale = yscale;
+
+	iowrite32(((yscale & FRACT_MASK) << 20) |
+		  ((xscale >> FP_SHIFT)  << 16) |
+		  ((xscale & FRACT_MASK) << 3),
+		 &regs->YRGBSCALE);
+
+	iowrite32(((yscale_UV & FRACT_MASK) << 20) |
+		  ((xscale_UV >> FP_SHIFT)  << 16) |
+		  ((xscale_UV & FRACT_MASK) << 3),
+		 &regs->UVSCALE);
+
+	iowrite32((((yscale    >> FP_SHIFT) << 16) |
+		   ((yscale_UV >> FP_SHIFT) << 0)),
+		 &regs->UVSCALEV);
+
+	if (scale_changed)
+		update_polyphase_filter(regs);
+
+	return scale_changed;
+}
+
+static void update_colorkey(struct intel_overlay *overlay,
+			    struct overlay_registers __iomem *regs)
+{
+	const struct intel_plane_state *state =
+		to_intel_plane_state(overlay->crtc->base.primary->state);
+	u32 key = overlay->color_key;
+	u32 format = 0;
+	u32 flags = 0;
+
+	if (overlay->color_key_enabled)
+		flags |= DST_KEY_ENABLE;
+
+	if (state->uapi.visible)
+		format = state->hw.fb->format->format;
+
+	switch (format) {
+	case DRM_FORMAT_C8:
+		key = RGB8I_TO_COLORKEY(key);
+		flags |= CLK_RGB24_MASK;
+		break;
+	case DRM_FORMAT_XRGB1555:
+		key = RGB15_TO_COLORKEY(key);
+		flags |= CLK_RGB15_MASK;
+		break;
+	case DRM_FORMAT_RGB565:
+		key = RGB16_TO_COLORKEY(key);
+		flags |= CLK_RGB16_MASK;
+		break;
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_XBGR2101010:
+		key = RGB30_TO_COLORKEY(key);
+		flags |= CLK_RGB24_MASK;
+		break;
+	default:
+		flags |= CLK_RGB24_MASK;
+		break;
+	}
+
+	iowrite32(key, &regs->DCLRKV);
+	iowrite32(flags, &regs->DCLRKM);
+}
+
+static u32 overlay_cmd_reg(struct drm_intel_overlay_put_image *params)
+{
+	u32 cmd = OCMD_ENABLE | OCMD_BUF_TYPE_FRAME | OCMD_BUFFER0;
+
+	if (params->flags & I915_OVERLAY_YUV_PLANAR) {
+		switch (params->flags & I915_OVERLAY_DEPTH_MASK) {
+		case I915_OVERLAY_YUV422:
+			cmd |= OCMD_YUV_422_PLANAR;
+			break;
+		case I915_OVERLAY_YUV420:
+			cmd |= OCMD_YUV_420_PLANAR;
+			break;
+		case I915_OVERLAY_YUV411:
+		case I915_OVERLAY_YUV410:
+			cmd |= OCMD_YUV_410_PLANAR;
+			break;
+		}
+	} else { /* YUV packed */
+		switch (params->flags & I915_OVERLAY_DEPTH_MASK) {
+		case I915_OVERLAY_YUV422:
+			cmd |= OCMD_YUV_422_PACKED;
+			break;
+		case I915_OVERLAY_YUV411:
+			cmd |= OCMD_YUV_411_PACKED;
+			break;
+		}
+
+		switch (params->flags & I915_OVERLAY_SWAP_MASK) {
+		case I915_OVERLAY_NO_SWAP:
+			break;
+		case I915_OVERLAY_UV_SWAP:
+			cmd |= OCMD_UV_SWAP;
+			break;
+		case I915_OVERLAY_Y_SWAP:
+			cmd |= OCMD_Y_SWAP;
+			break;
+		case I915_OVERLAY_Y_AND_UV_SWAP:
+			cmd |= OCMD_Y_AND_UV_SWAP;
+			break;
+		}
+	}
+
+	return cmd;
+}
+
+static struct i915_vma *intel_overlay_pin_fb(struct drm_i915_gem_object *new_bo)
+{
+	struct i915_gem_ww_ctx ww;
+	struct i915_vma *vma;
+	int ret;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	ret = i915_gem_object_lock(new_bo, &ww);
+	if (!ret) {
+		vma = i915_gem_object_pin_to_display_plane(new_bo, &ww, 0,
+							   NULL, PIN_MAPPABLE);
+		ret = PTR_ERR_OR_ZERO(vma);
+	}
+	if (ret == -EDEADLK) {
+		ret = i915_gem_ww_ctx_backoff(&ww);
+		if (!ret)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	if (ret)
+		return ERR_PTR(ret);
+
+	return vma;
+}
+
+static int intel_overlay_do_put_image(struct intel_overlay *overlay,
+				      struct drm_i915_gem_object *new_bo,
+				      struct drm_intel_overlay_put_image *params)
+{
+	struct overlay_registers __iomem *regs = overlay->regs;
+	struct drm_i915_private *dev_priv = overlay->i915;
+	u32 swidth, swidthsw, sheight, ostride;
+	enum pipe pipe = overlay->crtc->pipe;
+	bool scale_changed = false;
+	struct i915_vma *vma;
+	int ret, tmp_width;
+
+	drm_WARN_ON(&dev_priv->drm,
+		    !drm_modeset_is_locked(&dev_priv->drm.mode_config.connection_mutex));
+
+	ret = intel_overlay_release_old_vid(overlay);
+	if (ret != 0)
+		return ret;
+
+	atomic_inc(&dev_priv->gpu_error.pending_fb_pin);
+
+	vma = intel_overlay_pin_fb(new_bo);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto out_pin_section;
+	}
+
+	i915_gem_object_flush_frontbuffer(new_bo, ORIGIN_DIRTYFB);
+
+	if (!overlay->active) {
+		const struct intel_crtc_state *crtc_state =
+			overlay->crtc->config;
+		u32 oconfig = 0;
+
+		if (crtc_state->gamma_enable &&
+		    crtc_state->gamma_mode == GAMMA_MODE_MODE_8BIT)
+			oconfig |= OCONF_CC_OUT_8BIT;
+		if (crtc_state->gamma_enable)
+			oconfig |= OCONF_GAMMA2_ENABLE;
+		if (DISPLAY_VER(dev_priv) == 4)
+			oconfig |= OCONF_CSC_MODE_BT709;
+		oconfig |= pipe == 0 ?
+			OCONF_PIPE_A : OCONF_PIPE_B;
+		iowrite32(oconfig, &regs->OCONFIG);
+
+		ret = intel_overlay_on(overlay);
+		if (ret != 0)
+			goto out_unpin;
+	}
+
+	iowrite32(params->dst_y << 16 | params->dst_x, &regs->DWINPOS);
+	iowrite32(params->dst_height << 16 | params->dst_width, &regs->DWINSZ);
+
+	if (params->flags & I915_OVERLAY_YUV_PACKED)
+		tmp_width = packed_width_bytes(params->flags,
+					       params->src_width);
+	else
+		tmp_width = params->src_width;
+
+	swidth = params->src_width;
+	swidthsw = calc_swidthsw(dev_priv, params->offset_Y, tmp_width);
+	sheight = params->src_height;
+	iowrite32(i915_ggtt_offset(vma) + params->offset_Y, &regs->OBUF_0Y);
+	ostride = params->stride_Y;
+
+	if (params->flags & I915_OVERLAY_YUV_PLANAR) {
+		int uv_hscale = uv_hsubsampling(params->flags);
+		int uv_vscale = uv_vsubsampling(params->flags);
+		u32 tmp_U, tmp_V;
+
+		swidth |= (params->src_width / uv_hscale) << 16;
+		sheight |= (params->src_height / uv_vscale) << 16;
+
+		tmp_U = calc_swidthsw(dev_priv, params->offset_U,
+				      params->src_width / uv_hscale);
+		tmp_V = calc_swidthsw(dev_priv, params->offset_V,
+				      params->src_width / uv_hscale);
+		swidthsw |= max(tmp_U, tmp_V) << 16;
+
+		iowrite32(i915_ggtt_offset(vma) + params->offset_U,
+			  &regs->OBUF_0U);
+		iowrite32(i915_ggtt_offset(vma) + params->offset_V,
+			  &regs->OBUF_0V);
+
+		ostride |= params->stride_UV << 16;
+	}
+
+	iowrite32(swidth, &regs->SWIDTH);
+	iowrite32(swidthsw, &regs->SWIDTHSW);
+	iowrite32(sheight, &regs->SHEIGHT);
+	iowrite32(ostride, &regs->OSTRIDE);
+
+	scale_changed = update_scaling_factors(overlay, regs, params);
+
+	update_colorkey(overlay, regs);
+
+	iowrite32(overlay_cmd_reg(params), &regs->OCMD);
+
+	ret = intel_overlay_continue(overlay, vma, scale_changed);
+	if (ret)
+		goto out_unpin;
+
+	return 0;
+
+out_unpin:
+	i915_vma_unpin(vma);
+out_pin_section:
+	atomic_dec(&dev_priv->gpu_error.pending_fb_pin);
+
+	return ret;
+}
+
+int intel_overlay_switch_off(struct intel_overlay *overlay)
+{
+	struct drm_i915_private *dev_priv = overlay->i915;
+	int ret;
+
+	drm_WARN_ON(&dev_priv->drm,
+		    !drm_modeset_is_locked(&dev_priv->drm.mode_config.connection_mutex));
+
+	ret = intel_overlay_recover_from_interrupt(overlay);
+	if (ret != 0)
+		return ret;
+
+	if (!overlay->active)
+		return 0;
+
+	ret = intel_overlay_release_old_vid(overlay);
+	if (ret != 0)
+		return ret;
+
+	iowrite32(0, &overlay->regs->OCMD);
+
+	return intel_overlay_off(overlay);
+}
+
+static int check_overlay_possible_on_crtc(struct intel_overlay *overlay,
+					  struct intel_crtc *crtc)
+{
+	if (!crtc->active)
+		return -EINVAL;
+
+	/* can't use the overlay with double wide pipe */
+	if (crtc->config->double_wide)
+		return -EINVAL;
+
+	return 0;
+}
+
+static void update_pfit_vscale_ratio(struct intel_overlay *overlay)
+{
+	struct drm_i915_private *dev_priv = overlay->i915;
+	u32 pfit_control = intel_de_read(dev_priv, PFIT_CONTROL);
+	u32 ratio;
+
+	/* XXX: This is not the same logic as in the xorg driver, but more in
+	 * line with the intel documentation for the i965
+	 */
+	if (DISPLAY_VER(dev_priv) >= 4) {
+		/* on i965 use the PGM reg to read out the autoscaler values */
+		ratio = intel_de_read(dev_priv, PFIT_PGM_RATIOS) >> PFIT_VERT_SCALE_SHIFT_965;
+	} else {
+		if (pfit_control & VERT_AUTO_SCALE)
+			ratio = intel_de_read(dev_priv, PFIT_AUTO_RATIOS);
+		else
+			ratio = intel_de_read(dev_priv, PFIT_PGM_RATIOS);
+		ratio >>= PFIT_VERT_SCALE_SHIFT;
+	}
+
+	overlay->pfit_vscale_ratio = ratio;
+}
+
+static int check_overlay_dst(struct intel_overlay *overlay,
+			     struct drm_intel_overlay_put_image *rec)
+{
+	const struct intel_crtc_state *crtc_state =
+		overlay->crtc->config;
+	struct drm_rect req, clipped;
+
+	drm_rect_init(&req, rec->dst_x, rec->dst_y,
+		      rec->dst_width, rec->dst_height);
+
+	clipped = req;
+	drm_rect_intersect(&clipped, &crtc_state->pipe_src);
+
+	if (!drm_rect_visible(&clipped) ||
+	    !drm_rect_equals(&clipped, &req))
+		return -EINVAL;
+
+	return 0;
+}
+
+static int check_overlay_scaling(struct drm_intel_overlay_put_image *rec)
+{
+	u32 tmp;
+
+	/* downscaling limit is 8.0 */
+	tmp = ((rec->src_scan_height << 16) / rec->dst_height) >> 16;
+	if (tmp > 7)
+		return -EINVAL;
+
+	tmp = ((rec->src_scan_width << 16) / rec->dst_width) >> 16;
+	if (tmp > 7)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int check_overlay_src(struct drm_i915_private *dev_priv,
+			     struct drm_intel_overlay_put_image *rec,
+			     struct drm_i915_gem_object *new_bo)
+{
+	int uv_hscale = uv_hsubsampling(rec->flags);
+	int uv_vscale = uv_vsubsampling(rec->flags);
+	u32 stride_mask;
+	int depth;
+	u32 tmp;
+
+	/* check src dimensions */
+	if (IS_I845G(dev_priv) || IS_I830(dev_priv)) {
+		if (rec->src_height > IMAGE_MAX_HEIGHT_LEGACY ||
+		    rec->src_width  > IMAGE_MAX_WIDTH_LEGACY)
+			return -EINVAL;
+	} else {
+		if (rec->src_height > IMAGE_MAX_HEIGHT ||
+		    rec->src_width  > IMAGE_MAX_WIDTH)
+			return -EINVAL;
+	}
+
+	/* better safe than sorry, use 4 as the maximal subsampling ratio */
+	if (rec->src_height < N_VERT_Y_TAPS*4 ||
+	    rec->src_width  < N_HORIZ_Y_TAPS*4)
+		return -EINVAL;
+
+	/* check alignment constraints */
+	switch (rec->flags & I915_OVERLAY_TYPE_MASK) {
+	case I915_OVERLAY_RGB:
+		/* not implemented */
+		return -EINVAL;
+
+	case I915_OVERLAY_YUV_PACKED:
+		if (uv_vscale != 1)
+			return -EINVAL;
+
+		depth = packed_depth_bytes(rec->flags);
+		if (depth < 0)
+			return depth;
+
+		/* ignore UV planes */
+		rec->stride_UV = 0;
+		rec->offset_U = 0;
+		rec->offset_V = 0;
+		/* check pixel alignment */
+		if (rec->offset_Y % depth)
+			return -EINVAL;
+		break;
+
+	case I915_OVERLAY_YUV_PLANAR:
+		if (uv_vscale < 0 || uv_hscale < 0)
+			return -EINVAL;
+		/* no offset restrictions for planar formats */
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	if (rec->src_width % uv_hscale)
+		return -EINVAL;
+
+	/* stride checking */
+	if (IS_I830(dev_priv) || IS_I845G(dev_priv))
+		stride_mask = 255;
+	else
+		stride_mask = 63;
+
+	if (rec->stride_Y & stride_mask || rec->stride_UV & stride_mask)
+		return -EINVAL;
+	if (DISPLAY_VER(dev_priv) == 4 && rec->stride_Y < 512)
+		return -EINVAL;
+
+	tmp = (rec->flags & I915_OVERLAY_TYPE_MASK) == I915_OVERLAY_YUV_PLANAR ?
+		4096 : 8192;
+	if (rec->stride_Y > tmp || rec->stride_UV > 2*1024)
+		return -EINVAL;
+
+	/* check buffer dimensions */
+	switch (rec->flags & I915_OVERLAY_TYPE_MASK) {
+	case I915_OVERLAY_RGB:
+	case I915_OVERLAY_YUV_PACKED:
+		/* always 4 Y values per depth pixels */
+		if (packed_width_bytes(rec->flags, rec->src_width) > rec->stride_Y)
+			return -EINVAL;
+
+		tmp = rec->stride_Y*rec->src_height;
+		if (rec->offset_Y + tmp > new_bo->base.size)
+			return -EINVAL;
+		break;
+
+	case I915_OVERLAY_YUV_PLANAR:
+		if (rec->src_width > rec->stride_Y)
+			return -EINVAL;
+		if (rec->src_width/uv_hscale > rec->stride_UV)
+			return -EINVAL;
+
+		tmp = rec->stride_Y * rec->src_height;
+		if (rec->offset_Y + tmp > new_bo->base.size)
+			return -EINVAL;
+
+		tmp = rec->stride_UV * (rec->src_height / uv_vscale);
+		if (rec->offset_U + tmp > new_bo->base.size ||
+		    rec->offset_V + tmp > new_bo->base.size)
+			return -EINVAL;
+		break;
+	}
+
+	return 0;
+}
+
+int intel_overlay_put_image_ioctl(struct drm_device *dev, void *data,
+				  struct drm_file *file_priv)
+{
+	struct drm_intel_overlay_put_image *params = data;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_overlay *overlay;
+	struct drm_crtc *drmmode_crtc;
+	struct intel_crtc *crtc;
+	struct drm_i915_gem_object *new_bo;
+	int ret;
+
+	overlay = dev_priv->display.overlay;
+	if (!overlay) {
+		drm_dbg(&dev_priv->drm, "userspace bug: no overlay\n");
+		return -ENODEV;
+	}
+
+	if (!(params->flags & I915_OVERLAY_ENABLE)) {
+		drm_modeset_lock_all(dev);
+		ret = intel_overlay_switch_off(overlay);
+		drm_modeset_unlock_all(dev);
+
+		return ret;
+	}
+
+	drmmode_crtc = drm_crtc_find(dev, file_priv, params->crtc_id);
+	if (!drmmode_crtc)
+		return -ENOENT;
+	crtc = to_intel_crtc(drmmode_crtc);
+
+	new_bo = i915_gem_object_lookup(file_priv, params->bo_handle);
+	if (!new_bo)
+		return -ENOENT;
+
+	drm_modeset_lock_all(dev);
+
+	if (i915_gem_object_is_tiled(new_bo)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "buffer used for overlay image can not be tiled\n");
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+
+	ret = intel_overlay_recover_from_interrupt(overlay);
+	if (ret != 0)
+		goto out_unlock;
+
+	if (overlay->crtc != crtc) {
+		ret = intel_overlay_switch_off(overlay);
+		if (ret != 0)
+			goto out_unlock;
+
+		ret = check_overlay_possible_on_crtc(overlay, crtc);
+		if (ret != 0)
+			goto out_unlock;
+
+		overlay->crtc = crtc;
+		crtc->overlay = overlay;
+
+		/* line too wide, i.e. one-line-mode */
+		if (drm_rect_width(&crtc->config->pipe_src) > 1024 &&
+		    crtc->config->gmch_pfit.control & PFIT_ENABLE) {
+			overlay->pfit_active = true;
+			update_pfit_vscale_ratio(overlay);
+		} else
+			overlay->pfit_active = false;
+	}
+
+	ret = check_overlay_dst(overlay, params);
+	if (ret != 0)
+		goto out_unlock;
+
+	if (overlay->pfit_active) {
+		params->dst_y = (((u32)params->dst_y << 12) /
+				 overlay->pfit_vscale_ratio);
+		/* shifting right rounds downwards, so add 1 */
+		params->dst_height = (((u32)params->dst_height << 12) /
+				 overlay->pfit_vscale_ratio) + 1;
+	}
+
+	if (params->src_scan_height > params->src_height ||
+	    params->src_scan_width > params->src_width) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+
+	ret = check_overlay_src(dev_priv, params, new_bo);
+	if (ret != 0)
+		goto out_unlock;
+
+	/* Check scaling after src size to prevent a divide-by-zero. */
+	ret = check_overlay_scaling(params);
+	if (ret != 0)
+		goto out_unlock;
+
+	ret = intel_overlay_do_put_image(overlay, new_bo, params);
+	if (ret != 0)
+		goto out_unlock;
+
+	drm_modeset_unlock_all(dev);
+	i915_gem_object_put(new_bo);
+
+	return 0;
+
+out_unlock:
+	drm_modeset_unlock_all(dev);
+	i915_gem_object_put(new_bo);
+
+	return ret;
+}
+
+static void update_reg_attrs(struct intel_overlay *overlay,
+			     struct overlay_registers __iomem *regs)
+{
+	iowrite32((overlay->contrast << 18) | (overlay->brightness & 0xff),
+		  &regs->OCLRC0);
+	iowrite32(overlay->saturation, &regs->OCLRC1);
+}
+
+static bool check_gamma_bounds(u32 gamma1, u32 gamma2)
+{
+	int i;
+
+	if (gamma1 & 0xff000000 || gamma2 & 0xff000000)
+		return false;
+
+	for (i = 0; i < 3; i++) {
+		if (((gamma1 >> i*8) & 0xff) >= ((gamma2 >> i*8) & 0xff))
+			return false;
+	}
+
+	return true;
+}
+
+static bool check_gamma5_errata(u32 gamma5)
+{
+	int i;
+
+	for (i = 0; i < 3; i++) {
+		if (((gamma5 >> i*8) & 0xff) == 0x80)
+			return false;
+	}
+
+	return true;
+}
+
+static int check_gamma(struct drm_intel_overlay_attrs *attrs)
+{
+	if (!check_gamma_bounds(0, attrs->gamma0) ||
+	    !check_gamma_bounds(attrs->gamma0, attrs->gamma1) ||
+	    !check_gamma_bounds(attrs->gamma1, attrs->gamma2) ||
+	    !check_gamma_bounds(attrs->gamma2, attrs->gamma3) ||
+	    !check_gamma_bounds(attrs->gamma3, attrs->gamma4) ||
+	    !check_gamma_bounds(attrs->gamma4, attrs->gamma5) ||
+	    !check_gamma_bounds(attrs->gamma5, 0x00ffffff))
+		return -EINVAL;
+
+	if (!check_gamma5_errata(attrs->gamma5))
+		return -EINVAL;
+
+	return 0;
+}
+
+int intel_overlay_attrs_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *file_priv)
+{
+	struct drm_intel_overlay_attrs *attrs = data;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_overlay *overlay;
+	int ret;
+
+	overlay = dev_priv->display.overlay;
+	if (!overlay) {
+		drm_dbg(&dev_priv->drm, "userspace bug: no overlay\n");
+		return -ENODEV;
+	}
+
+	drm_modeset_lock_all(dev);
+
+	ret = -EINVAL;
+	if (!(attrs->flags & I915_OVERLAY_UPDATE_ATTRS)) {
+		attrs->color_key  = overlay->color_key;
+		attrs->brightness = overlay->brightness;
+		attrs->contrast   = overlay->contrast;
+		attrs->saturation = overlay->saturation;
+
+		if (DISPLAY_VER(dev_priv) != 2) {
+			attrs->gamma0 = intel_de_read(dev_priv, OGAMC0);
+			attrs->gamma1 = intel_de_read(dev_priv, OGAMC1);
+			attrs->gamma2 = intel_de_read(dev_priv, OGAMC2);
+			attrs->gamma3 = intel_de_read(dev_priv, OGAMC3);
+			attrs->gamma4 = intel_de_read(dev_priv, OGAMC4);
+			attrs->gamma5 = intel_de_read(dev_priv, OGAMC5);
+		}
+	} else {
+		if (attrs->brightness < -128 || attrs->brightness > 127)
+			goto out_unlock;
+		if (attrs->contrast > 255)
+			goto out_unlock;
+		if (attrs->saturation > 1023)
+			goto out_unlock;
+
+		overlay->color_key  = attrs->color_key;
+		overlay->brightness = attrs->brightness;
+		overlay->contrast   = attrs->contrast;
+		overlay->saturation = attrs->saturation;
+
+		update_reg_attrs(overlay, overlay->regs);
+
+		if (attrs->flags & I915_OVERLAY_UPDATE_GAMMA) {
+			if (DISPLAY_VER(dev_priv) == 2)
+				goto out_unlock;
+
+			if (overlay->active) {
+				ret = -EBUSY;
+				goto out_unlock;
+			}
+
+			ret = check_gamma(attrs);
+			if (ret)
+				goto out_unlock;
+
+			intel_de_write(dev_priv, OGAMC0, attrs->gamma0);
+			intel_de_write(dev_priv, OGAMC1, attrs->gamma1);
+			intel_de_write(dev_priv, OGAMC2, attrs->gamma2);
+			intel_de_write(dev_priv, OGAMC3, attrs->gamma3);
+			intel_de_write(dev_priv, OGAMC4, attrs->gamma4);
+			intel_de_write(dev_priv, OGAMC5, attrs->gamma5);
+		}
+	}
+	overlay->color_key_enabled = (attrs->flags & I915_OVERLAY_DISABLE_DEST_COLORKEY) == 0;
+
+	ret = 0;
+out_unlock:
+	drm_modeset_unlock_all(dev);
+
+	return ret;
+}
+
+static int get_registers(struct intel_overlay *overlay, bool use_phys)
+{
+	struct drm_i915_private *i915 = overlay->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err;
+
+	obj = i915_gem_object_create_stolen(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_put_bo;
+	}
+
+	if (use_phys)
+		overlay->flip_addr = sg_dma_address(obj->mm.pages->sgl);
+	else
+		overlay->flip_addr = i915_ggtt_offset(vma);
+	overlay->regs = i915_vma_pin_iomap(vma);
+	i915_vma_unpin(vma);
+
+	if (IS_ERR(overlay->regs)) {
+		err = PTR_ERR(overlay->regs);
+		goto err_put_bo;
+	}
+
+	overlay->reg_bo = obj;
+	return 0;
+
+err_put_bo:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+void intel_overlay_setup(struct drm_i915_private *dev_priv)
+{
+	struct intel_overlay *overlay;
+	struct intel_engine_cs *engine;
+	int ret;
+
+	if (!HAS_OVERLAY(dev_priv))
+		return;
+
+	engine = to_gt(dev_priv)->engine[RCS0];
+	if (!engine || !engine->kernel_context)
+		return;
+
+	overlay = kzalloc(sizeof(*overlay), GFP_KERNEL);
+	if (!overlay)
+		return;
+
+	overlay->i915 = dev_priv;
+	overlay->context = engine->kernel_context;
+	overlay->color_key = 0x0101fe;
+	overlay->color_key_enabled = true;
+	overlay->brightness = -19;
+	overlay->contrast = 75;
+	overlay->saturation = 146;
+
+	i915_active_init(&overlay->last_flip,
+			 NULL, intel_overlay_last_flip_retire, 0);
+
+	ret = get_registers(overlay, OVERLAY_NEEDS_PHYSICAL(dev_priv));
+	if (ret)
+		goto out_free;
+
+	memset_io(overlay->regs, 0, sizeof(struct overlay_registers));
+	update_polyphase_filter(overlay->regs);
+	update_reg_attrs(overlay, overlay->regs);
+
+	dev_priv->display.overlay = overlay;
+	drm_info(&dev_priv->drm, "Initialized overlay support.\n");
+	return;
+
+out_free:
+	kfree(overlay);
+}
+
+void intel_overlay_cleanup(struct drm_i915_private *dev_priv)
+{
+	struct intel_overlay *overlay;
+
+	overlay = fetch_and_zero(&dev_priv->display.overlay);
+	if (!overlay)
+		return;
+
+	/*
+	 * The bo's should be free'd by the generic code already.
+	 * Furthermore modesetting teardown happens beforehand so the
+	 * hardware should be off already.
+	 */
+	drm_WARN_ON(&dev_priv->drm, overlay->active);
+
+	i915_gem_object_put(overlay->reg_bo);
+	i915_active_fini(&overlay->last_flip);
+
+	kfree(overlay);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+
+struct intel_overlay_error_state {
+	struct overlay_registers regs;
+	unsigned long base;
+	u32 dovsta;
+	u32 isr;
+};
+
+struct intel_overlay_error_state *
+intel_overlay_capture_error_state(struct drm_i915_private *dev_priv)
+{
+	struct intel_overlay *overlay = dev_priv->display.overlay;
+	struct intel_overlay_error_state *error;
+
+	if (!overlay || !overlay->active)
+		return NULL;
+
+	error = kmalloc(sizeof(*error), GFP_ATOMIC);
+	if (error == NULL)
+		return NULL;
+
+	error->dovsta = intel_de_read(dev_priv, DOVSTA);
+	error->isr = intel_de_read(dev_priv, GEN2_ISR);
+	error->base = overlay->flip_addr;
+
+	memcpy_fromio(&error->regs, overlay->regs, sizeof(error->regs));
+
+	return error;
+}
+
+void
+intel_overlay_print_error_state(struct drm_i915_error_state_buf *m,
+				struct intel_overlay_error_state *error)
+{
+	i915_error_printf(m, "Overlay, status: 0x%08x, interrupt: 0x%08x\n",
+			  error->dovsta, error->isr);
+	i915_error_printf(m, "  Register file at 0x%08lx:\n",
+			  error->base);
+
+#define P(x) i915_error_printf(m, "    " #x ":	0x%08x\n", error->regs.x)
+	P(OBUF_0Y);
+	P(OBUF_1Y);
+	P(OBUF_0U);
+	P(OBUF_0V);
+	P(OBUF_1U);
+	P(OBUF_1V);
+	P(OSTRIDE);
+	P(YRGB_VPH);
+	P(UV_VPH);
+	P(HORZ_PH);
+	P(INIT_PHS);
+	P(DWINPOS);
+	P(DWINSZ);
+	P(SWIDTH);
+	P(SWIDTHSW);
+	P(SHEIGHT);
+	P(YRGBSCALE);
+	P(UVSCALE);
+	P(OCLRC0);
+	P(OCLRC1);
+	P(DCLRKV);
+	P(DCLRKM);
+	P(SCLRKVH);
+	P(SCLRKVL);
+	P(SCLRKEN);
+	P(OCONFIG);
+	P(OCMD);
+	P(OSTART_0Y);
+	P(OSTART_1Y);
+	P(OSTART_0U);
+	P(OSTART_0V);
+	P(OSTART_1U);
+	P(OSTART_1V);
+	P(OTILEOFF_0Y);
+	P(OTILEOFF_1Y);
+	P(OTILEOFF_0U);
+	P(OTILEOFF_0V);
+	P(OTILEOFF_1U);
+	P(OTILEOFF_1V);
+	P(FASTHSCALE);
+	P(UVSCALEV);
+#undef P
+}
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_overlay.h b/drivers/gpu/drm/i915/display/intel_overlay.h
new file mode 100644
index 000000000..a167c28ac
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_overlay.h
@@ -0,0 +1,29 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_OVERLAY_H__
+#define __INTEL_OVERLAY_H__
+
+struct drm_device;
+struct drm_file;
+struct drm_i915_error_state_buf;
+struct drm_i915_private;
+struct intel_overlay;
+struct intel_overlay_error_state;
+
+void intel_overlay_setup(struct drm_i915_private *dev_priv);
+void intel_overlay_cleanup(struct drm_i915_private *dev_priv);
+int intel_overlay_switch_off(struct intel_overlay *overlay);
+int intel_overlay_put_image_ioctl(struct drm_device *dev, void *data,
+				  struct drm_file *file_priv);
+int intel_overlay_attrs_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *file_priv);
+void intel_overlay_reset(struct drm_i915_private *dev_priv);
+struct intel_overlay_error_state *
+intel_overlay_capture_error_state(struct drm_i915_private *dev_priv);
+void intel_overlay_print_error_state(struct drm_i915_error_state_buf *e,
+				     struct intel_overlay_error_state *error);
+
+#endif /* __INTEL_OVERLAY_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_panel.c b/drivers/gpu/drm/i915/display/intel_panel.c
new file mode 100644
index 000000000..b50db0dd2
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_panel.c
@@ -0,0 +1,686 @@
+/*
+ * Copyright © 2006-2010 Intel Corporation
+ * Copyright (c) 2006 Dave Airlie <airlied@linux.ie>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Eric Anholt <eric@anholt.net>
+ *      Dave Airlie <airlied@linux.ie>
+ *      Jesse Barnes <jesse.barnes@intel.com>
+ *      Chris Wilson <chris@chris-wilson.co.uk>
+ */
+
+#include <linux/kernel.h>
+#include <linux/pwm.h>
+
+#include "intel_backlight.h"
+#include "intel_connector.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_drrs.h"
+#include "intel_panel.h"
+#include "intel_quirks.h"
+
+bool intel_panel_use_ssc(struct drm_i915_private *i915)
+{
+	if (i915->params.panel_use_ssc >= 0)
+		return i915->params.panel_use_ssc != 0;
+	return i915->display.vbt.lvds_use_ssc &&
+		!intel_has_quirk(i915, QUIRK_LVDS_SSC_DISABLE);
+}
+
+const struct drm_display_mode *
+intel_panel_preferred_fixed_mode(struct intel_connector *connector)
+{
+	return list_first_entry_or_null(&connector->panel.fixed_modes,
+					struct drm_display_mode, head);
+}
+
+const struct drm_display_mode *
+intel_panel_fixed_mode(struct intel_connector *connector,
+		       const struct drm_display_mode *mode)
+{
+	const struct drm_display_mode *fixed_mode, *best_mode = NULL;
+	int vrefresh = drm_mode_vrefresh(mode);
+
+	/* pick the fixed_mode that is closest in terms of vrefresh */
+	list_for_each_entry(fixed_mode, &connector->panel.fixed_modes, head) {
+		if (!best_mode ||
+		    abs(drm_mode_vrefresh(fixed_mode) - vrefresh) <
+		    abs(drm_mode_vrefresh(best_mode) - vrefresh))
+			best_mode = fixed_mode;
+	}
+
+	return best_mode;
+}
+
+static bool is_alt_drrs_mode(const struct drm_display_mode *mode,
+			     const struct drm_display_mode *preferred_mode)
+{
+	return drm_mode_match(mode, preferred_mode,
+			      DRM_MODE_MATCH_TIMINGS |
+			      DRM_MODE_MATCH_FLAGS |
+			      DRM_MODE_MATCH_3D_FLAGS) &&
+		mode->clock != preferred_mode->clock;
+}
+
+static bool is_alt_fixed_mode(const struct drm_display_mode *mode,
+			      const struct drm_display_mode *preferred_mode)
+{
+	u32 sync_flags = DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NHSYNC |
+		DRM_MODE_FLAG_PVSYNC | DRM_MODE_FLAG_NVSYNC;
+
+	return (mode->flags & ~sync_flags) == (preferred_mode->flags & ~sync_flags) &&
+		mode->hdisplay == preferred_mode->hdisplay &&
+		mode->vdisplay == preferred_mode->vdisplay;
+}
+
+const struct drm_display_mode *
+intel_panel_downclock_mode(struct intel_connector *connector,
+			   const struct drm_display_mode *adjusted_mode)
+{
+	const struct drm_display_mode *fixed_mode, *best_mode = NULL;
+	int min_vrefresh = connector->panel.vbt.seamless_drrs_min_refresh_rate;
+	int max_vrefresh = drm_mode_vrefresh(adjusted_mode);
+
+	/* pick the fixed_mode with the lowest refresh rate */
+	list_for_each_entry(fixed_mode, &connector->panel.fixed_modes, head) {
+		int vrefresh = drm_mode_vrefresh(fixed_mode);
+
+		if (is_alt_drrs_mode(fixed_mode, adjusted_mode) &&
+		    vrefresh >= min_vrefresh && vrefresh < max_vrefresh) {
+			max_vrefresh = vrefresh;
+			best_mode = fixed_mode;
+		}
+	}
+
+	return best_mode;
+}
+
+const struct drm_display_mode *
+intel_panel_highest_mode(struct intel_connector *connector,
+			 const struct drm_display_mode *adjusted_mode)
+{
+	const struct drm_display_mode *fixed_mode, *best_mode = adjusted_mode;
+
+	/* pick the fixed_mode that has the highest clock */
+	list_for_each_entry(fixed_mode, &connector->panel.fixed_modes, head) {
+		if (fixed_mode->clock > best_mode->clock)
+			best_mode = fixed_mode;
+	}
+
+	return best_mode;
+}
+
+int intel_panel_get_modes(struct intel_connector *connector)
+{
+	const struct drm_display_mode *fixed_mode;
+	int num_modes = 0;
+
+	list_for_each_entry(fixed_mode, &connector->panel.fixed_modes, head) {
+		struct drm_display_mode *mode;
+
+		mode = drm_mode_duplicate(connector->base.dev, fixed_mode);
+		if (mode) {
+			drm_mode_probed_add(&connector->base, mode);
+			num_modes++;
+		}
+	}
+
+	return num_modes;
+}
+
+enum drrs_type intel_panel_drrs_type(struct intel_connector *connector)
+{
+	if (list_empty(&connector->panel.fixed_modes) ||
+	    list_is_singular(&connector->panel.fixed_modes))
+		return DRRS_TYPE_NONE;
+
+	return connector->panel.vbt.drrs_type;
+}
+
+int intel_panel_compute_config(struct intel_connector *connector,
+			       struct drm_display_mode *adjusted_mode)
+{
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_fixed_mode(connector, adjusted_mode);
+
+	if (!fixed_mode)
+		return 0;
+
+	/*
+	 * We don't want to lie too much to the user about the refresh
+	 * rate they're going to get. But we have to allow a bit of latitude
+	 * for Xorg since it likes to automagically cook up modes with slightly
+	 * off refresh rates.
+	 */
+	if (abs(drm_mode_vrefresh(adjusted_mode) - drm_mode_vrefresh(fixed_mode)) > 1) {
+		drm_dbg_kms(connector->base.dev,
+			    "[CONNECTOR:%d:%s] Requested mode vrefresh (%d Hz) does not match fixed mode vrefresh (%d Hz)\n",
+			    connector->base.base.id, connector->base.name,
+			    drm_mode_vrefresh(adjusted_mode), drm_mode_vrefresh(fixed_mode));
+
+		return -EINVAL;
+	}
+
+	drm_mode_copy(adjusted_mode, fixed_mode);
+
+	drm_mode_set_crtcinfo(adjusted_mode, 0);
+
+	return 0;
+}
+
+static void intel_panel_add_edid_alt_fixed_modes(struct intel_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	const struct drm_display_mode *preferred_mode =
+		intel_panel_preferred_fixed_mode(connector);
+	struct drm_display_mode *mode, *next;
+
+	list_for_each_entry_safe(mode, next, &connector->base.probed_modes, head) {
+		if (!is_alt_fixed_mode(mode, preferred_mode))
+			continue;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "[CONNECTOR:%d:%s] using alternate EDID fixed mode: " DRM_MODE_FMT "\n",
+			    connector->base.base.id, connector->base.name,
+			    DRM_MODE_ARG(mode));
+
+		list_move_tail(&mode->head, &connector->panel.fixed_modes);
+	}
+}
+
+static void intel_panel_add_edid_preferred_mode(struct intel_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	struct drm_display_mode *scan, *fixed_mode = NULL;
+
+	if (list_empty(&connector->base.probed_modes))
+		return;
+
+	/* make sure the preferred mode is first */
+	list_for_each_entry(scan, &connector->base.probed_modes, head) {
+		if (scan->type & DRM_MODE_TYPE_PREFERRED) {
+			fixed_mode = scan;
+			break;
+		}
+	}
+
+	if (!fixed_mode)
+		fixed_mode = list_first_entry(&connector->base.probed_modes,
+					      typeof(*fixed_mode), head);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "[CONNECTOR:%d:%s] using %s EDID fixed mode: " DRM_MODE_FMT "\n",
+		    connector->base.base.id, connector->base.name,
+		    fixed_mode->type & DRM_MODE_TYPE_PREFERRED ? "preferred" : "first",
+		    DRM_MODE_ARG(fixed_mode));
+
+	fixed_mode->type |= DRM_MODE_TYPE_PREFERRED;
+
+	list_move_tail(&fixed_mode->head, &connector->panel.fixed_modes);
+}
+
+static void intel_panel_destroy_probed_modes(struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct drm_display_mode *mode, *next;
+
+	list_for_each_entry_safe(mode, next, &connector->base.probed_modes, head) {
+		drm_dbg_kms(&i915->drm,
+			    "[CONNECTOR:%d:%s] not using EDID mode: " DRM_MODE_FMT "\n",
+			    connector->base.base.id, connector->base.name,
+			    DRM_MODE_ARG(mode));
+		list_del(&mode->head);
+		drm_mode_destroy(&i915->drm, mode);
+	}
+}
+
+void intel_panel_add_edid_fixed_modes(struct intel_connector *connector,
+				      bool use_alt_fixed_modes)
+{
+	intel_panel_add_edid_preferred_mode(connector);
+	if (intel_panel_preferred_fixed_mode(connector) && use_alt_fixed_modes)
+		intel_panel_add_edid_alt_fixed_modes(connector);
+	intel_panel_destroy_probed_modes(connector);
+}
+
+static void intel_panel_add_fixed_mode(struct intel_connector *connector,
+				       struct drm_display_mode *fixed_mode,
+				       const char *type)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct drm_display_info *info = &connector->base.display_info;
+
+	if (!fixed_mode)
+		return;
+
+	fixed_mode->type |= DRM_MODE_TYPE_PREFERRED | DRM_MODE_TYPE_DRIVER;
+
+	info->width_mm = fixed_mode->width_mm;
+	info->height_mm = fixed_mode->height_mm;
+
+	drm_dbg_kms(&i915->drm, "[CONNECTOR:%d:%s] using %s fixed mode: " DRM_MODE_FMT "\n",
+		    connector->base.base.id, connector->base.name, type,
+		    DRM_MODE_ARG(fixed_mode));
+
+	list_add_tail(&fixed_mode->head, &connector->panel.fixed_modes);
+}
+
+void intel_panel_add_vbt_lfp_fixed_mode(struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	const struct drm_display_mode *mode;
+
+	mode = connector->panel.vbt.lfp_lvds_vbt_mode;
+	if (!mode)
+		return;
+
+	intel_panel_add_fixed_mode(connector,
+				   drm_mode_duplicate(&i915->drm, mode),
+				   "VBT LFP");
+}
+
+void intel_panel_add_vbt_sdvo_fixed_mode(struct intel_connector *connector)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	const struct drm_display_mode *mode;
+
+	mode = connector->panel.vbt.sdvo_lvds_vbt_mode;
+	if (!mode)
+		return;
+
+	intel_panel_add_fixed_mode(connector,
+				   drm_mode_duplicate(&i915->drm, mode),
+				   "VBT SDVO");
+}
+
+void intel_panel_add_encoder_fixed_mode(struct intel_connector *connector,
+					struct intel_encoder *encoder)
+{
+	intel_panel_add_fixed_mode(connector,
+				   intel_encoder_current_mode(encoder),
+				   "current (BIOS)");
+}
+
+/* adjusted_mode has been preset to be the panel's fixed mode */
+static int pch_panel_fitting(struct intel_crtc_state *crtc_state,
+			     const struct drm_connector_state *conn_state)
+{
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	int pipe_src_w = drm_rect_width(&crtc_state->pipe_src);
+	int pipe_src_h = drm_rect_height(&crtc_state->pipe_src);
+	int x, y, width, height;
+
+	/* Native modes don't need fitting */
+	if (adjusted_mode->crtc_hdisplay == pipe_src_w &&
+	    adjusted_mode->crtc_vdisplay == pipe_src_h &&
+	    crtc_state->output_format != INTEL_OUTPUT_FORMAT_YCBCR420)
+		return 0;
+
+	switch (conn_state->scaling_mode) {
+	case DRM_MODE_SCALE_CENTER:
+		width = pipe_src_w;
+		height = pipe_src_h;
+		x = (adjusted_mode->crtc_hdisplay - width + 1)/2;
+		y = (adjusted_mode->crtc_vdisplay - height + 1)/2;
+		break;
+
+	case DRM_MODE_SCALE_ASPECT:
+		/* Scale but preserve the aspect ratio */
+		{
+			u32 scaled_width = adjusted_mode->crtc_hdisplay * pipe_src_h;
+			u32 scaled_height = pipe_src_w * adjusted_mode->crtc_vdisplay;
+			if (scaled_width > scaled_height) { /* pillar */
+				width = scaled_height / pipe_src_h;
+				if (width & 1)
+					width++;
+				x = (adjusted_mode->crtc_hdisplay - width + 1) / 2;
+				y = 0;
+				height = adjusted_mode->crtc_vdisplay;
+			} else if (scaled_width < scaled_height) { /* letter */
+				height = scaled_width / pipe_src_w;
+				if (height & 1)
+				    height++;
+				y = (adjusted_mode->crtc_vdisplay - height + 1) / 2;
+				x = 0;
+				width = adjusted_mode->crtc_hdisplay;
+			} else {
+				x = y = 0;
+				width = adjusted_mode->crtc_hdisplay;
+				height = adjusted_mode->crtc_vdisplay;
+			}
+		}
+		break;
+
+	case DRM_MODE_SCALE_NONE:
+		WARN_ON(adjusted_mode->crtc_hdisplay != pipe_src_w);
+		WARN_ON(adjusted_mode->crtc_vdisplay != pipe_src_h);
+		fallthrough;
+	case DRM_MODE_SCALE_FULLSCREEN:
+		x = y = 0;
+		width = adjusted_mode->crtc_hdisplay;
+		height = adjusted_mode->crtc_vdisplay;
+		break;
+
+	default:
+		MISSING_CASE(conn_state->scaling_mode);
+		return -EINVAL;
+	}
+
+	drm_rect_init(&crtc_state->pch_pfit.dst,
+		      x, y, width, height);
+	crtc_state->pch_pfit.enabled = true;
+
+	return 0;
+}
+
+static void
+centre_horizontally(struct drm_display_mode *adjusted_mode,
+		    int width)
+{
+	u32 border, sync_pos, blank_width, sync_width;
+
+	/* keep the hsync and hblank widths constant */
+	sync_width = adjusted_mode->crtc_hsync_end - adjusted_mode->crtc_hsync_start;
+	blank_width = adjusted_mode->crtc_hblank_end - adjusted_mode->crtc_hblank_start;
+	sync_pos = (blank_width - sync_width + 1) / 2;
+
+	border = (adjusted_mode->crtc_hdisplay - width + 1) / 2;
+	border += border & 1; /* make the border even */
+
+	adjusted_mode->crtc_hdisplay = width;
+	adjusted_mode->crtc_hblank_start = width + border;
+	adjusted_mode->crtc_hblank_end = adjusted_mode->crtc_hblank_start + blank_width;
+
+	adjusted_mode->crtc_hsync_start = adjusted_mode->crtc_hblank_start + sync_pos;
+	adjusted_mode->crtc_hsync_end = adjusted_mode->crtc_hsync_start + sync_width;
+}
+
+static void
+centre_vertically(struct drm_display_mode *adjusted_mode,
+		  int height)
+{
+	u32 border, sync_pos, blank_width, sync_width;
+
+	/* keep the vsync and vblank widths constant */
+	sync_width = adjusted_mode->crtc_vsync_end - adjusted_mode->crtc_vsync_start;
+	blank_width = adjusted_mode->crtc_vblank_end - adjusted_mode->crtc_vblank_start;
+	sync_pos = (blank_width - sync_width + 1) / 2;
+
+	border = (adjusted_mode->crtc_vdisplay - height + 1) / 2;
+
+	adjusted_mode->crtc_vdisplay = height;
+	adjusted_mode->crtc_vblank_start = height + border;
+	adjusted_mode->crtc_vblank_end = adjusted_mode->crtc_vblank_start + blank_width;
+
+	adjusted_mode->crtc_vsync_start = adjusted_mode->crtc_vblank_start + sync_pos;
+	adjusted_mode->crtc_vsync_end = adjusted_mode->crtc_vsync_start + sync_width;
+}
+
+static u32 panel_fitter_scaling(u32 source, u32 target)
+{
+	/*
+	 * Floating point operation is not supported. So the FACTOR
+	 * is defined, which can avoid the floating point computation
+	 * when calculating the panel ratio.
+	 */
+#define ACCURACY 12
+#define FACTOR (1 << ACCURACY)
+	u32 ratio = source * FACTOR / target;
+	return (FACTOR * ratio + FACTOR/2) / FACTOR;
+}
+
+static void i965_scale_aspect(struct intel_crtc_state *crtc_state,
+			      u32 *pfit_control)
+{
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	int pipe_src_w = drm_rect_width(&crtc_state->pipe_src);
+	int pipe_src_h = drm_rect_height(&crtc_state->pipe_src);
+	u32 scaled_width = adjusted_mode->crtc_hdisplay * pipe_src_h;
+	u32 scaled_height = pipe_src_w * adjusted_mode->crtc_vdisplay;
+
+	/* 965+ is easy, it does everything in hw */
+	if (scaled_width > scaled_height)
+		*pfit_control |= PFIT_ENABLE |
+			PFIT_SCALING_PILLAR;
+	else if (scaled_width < scaled_height)
+		*pfit_control |= PFIT_ENABLE |
+			PFIT_SCALING_LETTER;
+	else if (adjusted_mode->crtc_hdisplay != pipe_src_w)
+		*pfit_control |= PFIT_ENABLE | PFIT_SCALING_AUTO;
+}
+
+static void i9xx_scale_aspect(struct intel_crtc_state *crtc_state,
+			      u32 *pfit_control, u32 *pfit_pgm_ratios,
+			      u32 *border)
+{
+	struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	int pipe_src_w = drm_rect_width(&crtc_state->pipe_src);
+	int pipe_src_h = drm_rect_height(&crtc_state->pipe_src);
+	u32 scaled_width = adjusted_mode->crtc_hdisplay * pipe_src_h;
+	u32 scaled_height = pipe_src_w * adjusted_mode->crtc_vdisplay;
+	u32 bits;
+
+	/*
+	 * For earlier chips we have to calculate the scaling
+	 * ratio by hand and program it into the
+	 * PFIT_PGM_RATIO register
+	 */
+	if (scaled_width > scaled_height) { /* pillar */
+		centre_horizontally(adjusted_mode,
+				    scaled_height / pipe_src_h);
+
+		*border = LVDS_BORDER_ENABLE;
+		if (pipe_src_h != adjusted_mode->crtc_vdisplay) {
+			bits = panel_fitter_scaling(pipe_src_h,
+						    adjusted_mode->crtc_vdisplay);
+
+			*pfit_pgm_ratios |= (bits << PFIT_HORIZ_SCALE_SHIFT |
+					     bits << PFIT_VERT_SCALE_SHIFT);
+			*pfit_control |= (PFIT_ENABLE |
+					  VERT_INTERP_BILINEAR |
+					  HORIZ_INTERP_BILINEAR);
+		}
+	} else if (scaled_width < scaled_height) { /* letter */
+		centre_vertically(adjusted_mode,
+				  scaled_width / pipe_src_w);
+
+		*border = LVDS_BORDER_ENABLE;
+		if (pipe_src_w != adjusted_mode->crtc_hdisplay) {
+			bits = panel_fitter_scaling(pipe_src_w,
+						    adjusted_mode->crtc_hdisplay);
+
+			*pfit_pgm_ratios |= (bits << PFIT_HORIZ_SCALE_SHIFT |
+					     bits << PFIT_VERT_SCALE_SHIFT);
+			*pfit_control |= (PFIT_ENABLE |
+					  VERT_INTERP_BILINEAR |
+					  HORIZ_INTERP_BILINEAR);
+		}
+	} else {
+		/* Aspects match, Let hw scale both directions */
+		*pfit_control |= (PFIT_ENABLE |
+				  VERT_AUTO_SCALE | HORIZ_AUTO_SCALE |
+				  VERT_INTERP_BILINEAR |
+				  HORIZ_INTERP_BILINEAR);
+	}
+}
+
+static int gmch_panel_fitting(struct intel_crtc_state *crtc_state,
+			      const struct drm_connector_state *conn_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 pfit_control = 0, pfit_pgm_ratios = 0, border = 0;
+	struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	int pipe_src_w = drm_rect_width(&crtc_state->pipe_src);
+	int pipe_src_h = drm_rect_height(&crtc_state->pipe_src);
+
+	/* Native modes don't need fitting */
+	if (adjusted_mode->crtc_hdisplay == pipe_src_w &&
+	    adjusted_mode->crtc_vdisplay == pipe_src_h)
+		goto out;
+
+	switch (conn_state->scaling_mode) {
+	case DRM_MODE_SCALE_CENTER:
+		/*
+		 * For centered modes, we have to calculate border widths &
+		 * heights and modify the values programmed into the CRTC.
+		 */
+		centre_horizontally(adjusted_mode, pipe_src_w);
+		centre_vertically(adjusted_mode, pipe_src_h);
+		border = LVDS_BORDER_ENABLE;
+		break;
+	case DRM_MODE_SCALE_ASPECT:
+		/* Scale but preserve the aspect ratio */
+		if (DISPLAY_VER(dev_priv) >= 4)
+			i965_scale_aspect(crtc_state, &pfit_control);
+		else
+			i9xx_scale_aspect(crtc_state, &pfit_control,
+					  &pfit_pgm_ratios, &border);
+		break;
+	case DRM_MODE_SCALE_FULLSCREEN:
+		/*
+		 * Full scaling, even if it changes the aspect ratio.
+		 * Fortunately this is all done for us in hw.
+		 */
+		if (pipe_src_h != adjusted_mode->crtc_vdisplay ||
+		    pipe_src_w != adjusted_mode->crtc_hdisplay) {
+			pfit_control |= PFIT_ENABLE;
+			if (DISPLAY_VER(dev_priv) >= 4)
+				pfit_control |= PFIT_SCALING_AUTO;
+			else
+				pfit_control |= (VERT_AUTO_SCALE |
+						 VERT_INTERP_BILINEAR |
+						 HORIZ_AUTO_SCALE |
+						 HORIZ_INTERP_BILINEAR);
+		}
+		break;
+	default:
+		MISSING_CASE(conn_state->scaling_mode);
+		return -EINVAL;
+	}
+
+	/* 965+ wants fuzzy fitting */
+	/* FIXME: handle multiple panels by failing gracefully */
+	if (DISPLAY_VER(dev_priv) >= 4)
+		pfit_control |= PFIT_PIPE(crtc->pipe) | PFIT_FILTER_FUZZY;
+
+out:
+	if ((pfit_control & PFIT_ENABLE) == 0) {
+		pfit_control = 0;
+		pfit_pgm_ratios = 0;
+	}
+
+	/* Make sure pre-965 set dither correctly for 18bpp panels. */
+	if (DISPLAY_VER(dev_priv) < 4 && crtc_state->pipe_bpp == 18)
+		pfit_control |= PANEL_8TO6_DITHER_ENABLE;
+
+	crtc_state->gmch_pfit.control = pfit_control;
+	crtc_state->gmch_pfit.pgm_ratios = pfit_pgm_ratios;
+	crtc_state->gmch_pfit.lvds_border_bits = border;
+
+	return 0;
+}
+
+int intel_panel_fitting(struct intel_crtc_state *crtc_state,
+			const struct drm_connector_state *conn_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	if (HAS_GMCH(i915))
+		return gmch_panel_fitting(crtc_state, conn_state);
+	else
+		return pch_panel_fitting(crtc_state, conn_state);
+}
+
+enum drm_connector_status
+intel_panel_detect(struct drm_connector *connector, bool force)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+
+	if (!INTEL_DISPLAY_ENABLED(i915))
+		return connector_status_disconnected;
+
+	return connector_status_connected;
+}
+
+enum drm_mode_status
+intel_panel_mode_valid(struct intel_connector *connector,
+		       const struct drm_display_mode *mode)
+{
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_fixed_mode(connector, mode);
+
+	if (!fixed_mode)
+		return MODE_OK;
+
+	if (mode->hdisplay != fixed_mode->hdisplay)
+		return MODE_PANEL;
+
+	if (mode->vdisplay != fixed_mode->vdisplay)
+		return MODE_PANEL;
+
+	if (drm_mode_vrefresh(mode) != drm_mode_vrefresh(fixed_mode))
+		return MODE_PANEL;
+
+	return MODE_OK;
+}
+
+void intel_panel_init_alloc(struct intel_connector *connector)
+{
+	struct intel_panel *panel = &connector->panel;
+
+	connector->panel.vbt.panel_type = -1;
+	INIT_LIST_HEAD(&panel->fixed_modes);
+}
+
+int intel_panel_init(struct intel_connector *connector)
+{
+	struct intel_panel *panel = &connector->panel;
+
+	intel_backlight_init_funcs(panel);
+
+	drm_dbg_kms(connector->base.dev,
+		    "[CONNECTOR:%d:%s] DRRS type: %s\n",
+		    connector->base.base.id, connector->base.name,
+		    intel_drrs_type_str(intel_panel_drrs_type(connector)));
+
+	return 0;
+}
+
+void intel_panel_fini(struct intel_connector *connector)
+{
+	struct intel_panel *panel = &connector->panel;
+	struct drm_display_mode *fixed_mode, *next;
+
+	intel_backlight_destroy(panel);
+
+	intel_bios_fini_panel(panel);
+
+	list_for_each_entry_safe(fixed_mode, next, &panel->fixed_modes, head) {
+		list_del(&fixed_mode->head);
+		drm_mode_destroy(connector->base.dev, fixed_mode);
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_panel.h b/drivers/gpu/drm/i915/display/intel_panel.h
new file mode 100644
index 000000000..4b51e1c51
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_panel.h
@@ -0,0 +1,54 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_PANEL_H__
+#define __INTEL_PANEL_H__
+
+#include <linux/types.h>
+
+enum drm_connector_status;
+enum drrs_type;
+struct drm_connector;
+struct drm_connector_state;
+struct drm_display_mode;
+struct drm_i915_private;
+struct intel_connector;
+struct intel_crtc_state;
+struct intel_encoder;
+
+void intel_panel_init_alloc(struct intel_connector *connector);
+int intel_panel_init(struct intel_connector *connector);
+void intel_panel_fini(struct intel_connector *connector);
+enum drm_connector_status
+intel_panel_detect(struct drm_connector *connector, bool force);
+bool intel_panel_use_ssc(struct drm_i915_private *i915);
+const struct drm_display_mode *
+intel_panel_preferred_fixed_mode(struct intel_connector *connector);
+const struct drm_display_mode *
+intel_panel_fixed_mode(struct intel_connector *connector,
+		       const struct drm_display_mode *mode);
+const struct drm_display_mode *
+intel_panel_downclock_mode(struct intel_connector *connector,
+			   const struct drm_display_mode *adjusted_mode);
+const struct drm_display_mode *
+intel_panel_highest_mode(struct intel_connector *connector,
+			 const struct drm_display_mode *adjusted_mode);
+int intel_panel_get_modes(struct intel_connector *connector);
+enum drrs_type intel_panel_drrs_type(struct intel_connector *connector);
+enum drm_mode_status
+intel_panel_mode_valid(struct intel_connector *connector,
+		       const struct drm_display_mode *mode);
+int intel_panel_fitting(struct intel_crtc_state *crtc_state,
+			const struct drm_connector_state *conn_state);
+int intel_panel_compute_config(struct intel_connector *connector,
+			       struct drm_display_mode *adjusted_mode);
+void intel_panel_add_edid_fixed_modes(struct intel_connector *connector,
+				      bool use_alt_fixed_modes);
+void intel_panel_add_vbt_lfp_fixed_mode(struct intel_connector *connector);
+void intel_panel_add_vbt_sdvo_fixed_mode(struct intel_connector *connector);
+void intel_panel_add_encoder_fixed_mode(struct intel_connector *connector,
+					struct intel_encoder *encoder);
+
+#endif /* __INTEL_PANEL_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_pch_display.c b/drivers/gpu/drm/i915/display/intel_pch_display.c
new file mode 100644
index 000000000..837152dca
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_pch_display.c
@@ -0,0 +1,650 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include "g4x_dp.h"
+#include "intel_crt.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_fdi.h"
+#include "intel_lvds.h"
+#include "intel_pch_display.h"
+#include "intel_pch_refclk.h"
+#include "intel_pps.h"
+#include "intel_sdvo.h"
+
+bool intel_has_pch_trancoder(struct drm_i915_private *i915,
+			     enum pipe pch_transcoder)
+{
+	return HAS_PCH_IBX(i915) || HAS_PCH_CPT(i915) ||
+		(HAS_PCH_LPT_H(i915) && pch_transcoder == PIPE_A);
+}
+
+enum pipe intel_crtc_pch_transcoder(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	if (HAS_PCH_LPT(i915))
+		return PIPE_A;
+	else
+		return crtc->pipe;
+}
+
+static void assert_pch_dp_disabled(struct drm_i915_private *dev_priv,
+				   enum pipe pipe, enum port port,
+				   i915_reg_t dp_reg)
+{
+	enum pipe port_pipe;
+	bool state;
+
+	state = g4x_dp_port_enabled(dev_priv, dp_reg, port, &port_pipe);
+
+	I915_STATE_WARN(state && port_pipe == pipe,
+			"PCH DP %c enabled on transcoder %c, should be disabled\n",
+			port_name(port), pipe_name(pipe));
+
+	I915_STATE_WARN(HAS_PCH_IBX(dev_priv) && !state && port_pipe == PIPE_B,
+			"IBX PCH DP %c still using transcoder B\n",
+			port_name(port));
+}
+
+static void assert_pch_hdmi_disabled(struct drm_i915_private *dev_priv,
+				     enum pipe pipe, enum port port,
+				     i915_reg_t hdmi_reg)
+{
+	enum pipe port_pipe;
+	bool state;
+
+	state = intel_sdvo_port_enabled(dev_priv, hdmi_reg, &port_pipe);
+
+	I915_STATE_WARN(state && port_pipe == pipe,
+			"PCH HDMI %c enabled on transcoder %c, should be disabled\n",
+			port_name(port), pipe_name(pipe));
+
+	I915_STATE_WARN(HAS_PCH_IBX(dev_priv) && !state && port_pipe == PIPE_B,
+			"IBX PCH HDMI %c still using transcoder B\n",
+			port_name(port));
+}
+
+static void assert_pch_ports_disabled(struct drm_i915_private *dev_priv,
+				      enum pipe pipe)
+{
+	enum pipe port_pipe;
+
+	assert_pch_dp_disabled(dev_priv, pipe, PORT_B, PCH_DP_B);
+	assert_pch_dp_disabled(dev_priv, pipe, PORT_C, PCH_DP_C);
+	assert_pch_dp_disabled(dev_priv, pipe, PORT_D, PCH_DP_D);
+
+	I915_STATE_WARN(intel_crt_port_enabled(dev_priv, PCH_ADPA, &port_pipe) &&
+			port_pipe == pipe,
+			"PCH VGA enabled on transcoder %c, should be disabled\n",
+			pipe_name(pipe));
+
+	I915_STATE_WARN(intel_lvds_port_enabled(dev_priv, PCH_LVDS, &port_pipe) &&
+			port_pipe == pipe,
+			"PCH LVDS enabled on transcoder %c, should be disabled\n",
+			pipe_name(pipe));
+
+	/* PCH SDVOB multiplex with HDMIB */
+	assert_pch_hdmi_disabled(dev_priv, pipe, PORT_B, PCH_HDMIB);
+	assert_pch_hdmi_disabled(dev_priv, pipe, PORT_C, PCH_HDMIC);
+	assert_pch_hdmi_disabled(dev_priv, pipe, PORT_D, PCH_HDMID);
+}
+
+static void assert_pch_transcoder_disabled(struct drm_i915_private *dev_priv,
+					   enum pipe pipe)
+{
+	u32 val;
+	bool enabled;
+
+	val = intel_de_read(dev_priv, PCH_TRANSCONF(pipe));
+	enabled = !!(val & TRANS_ENABLE);
+	I915_STATE_WARN(enabled,
+			"transcoder assertion failed, should be off on pipe %c but is still active\n",
+			pipe_name(pipe));
+}
+
+static void ibx_sanitize_pch_hdmi_port(struct drm_i915_private *dev_priv,
+				       enum port port, i915_reg_t hdmi_reg)
+{
+	u32 val = intel_de_read(dev_priv, hdmi_reg);
+
+	if (val & SDVO_ENABLE ||
+	    (val & SDVO_PIPE_SEL_MASK) == SDVO_PIPE_SEL(PIPE_A))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Sanitizing transcoder select for HDMI %c\n",
+		    port_name(port));
+
+	val &= ~SDVO_PIPE_SEL_MASK;
+	val |= SDVO_PIPE_SEL(PIPE_A);
+
+	intel_de_write(dev_priv, hdmi_reg, val);
+}
+
+static void ibx_sanitize_pch_dp_port(struct drm_i915_private *dev_priv,
+				     enum port port, i915_reg_t dp_reg)
+{
+	u32 val = intel_de_read(dev_priv, dp_reg);
+
+	if (val & DP_PORT_EN ||
+	    (val & DP_PIPE_SEL_MASK) == DP_PIPE_SEL(PIPE_A))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Sanitizing transcoder select for DP %c\n",
+		    port_name(port));
+
+	val &= ~DP_PIPE_SEL_MASK;
+	val |= DP_PIPE_SEL(PIPE_A);
+
+	intel_de_write(dev_priv, dp_reg, val);
+}
+
+static void ibx_sanitize_pch_ports(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * The BIOS may select transcoder B on some of the PCH
+	 * ports even it doesn't enable the port. This would trip
+	 * assert_pch_dp_disabled() and assert_pch_hdmi_disabled().
+	 * Sanitize the transcoder select bits to prevent that. We
+	 * assume that the BIOS never actually enabled the port,
+	 * because if it did we'd actually have to toggle the port
+	 * on and back off to make the transcoder A select stick
+	 * (see. intel_dp_link_down(), intel_disable_hdmi(),
+	 * intel_disable_sdvo()).
+	 */
+	ibx_sanitize_pch_dp_port(dev_priv, PORT_B, PCH_DP_B);
+	ibx_sanitize_pch_dp_port(dev_priv, PORT_C, PCH_DP_C);
+	ibx_sanitize_pch_dp_port(dev_priv, PORT_D, PCH_DP_D);
+
+	/* PCH SDVOB multiplex with HDMIB */
+	ibx_sanitize_pch_hdmi_port(dev_priv, PORT_B, PCH_HDMIB);
+	ibx_sanitize_pch_hdmi_port(dev_priv, PORT_C, PCH_HDMIC);
+	ibx_sanitize_pch_hdmi_port(dev_priv, PORT_D, PCH_HDMID);
+}
+
+static void intel_pch_transcoder_set_m1_n1(struct intel_crtc *crtc,
+					   const struct intel_link_m_n *m_n)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	intel_set_m_n(dev_priv, m_n,
+		      PCH_TRANS_DATA_M1(pipe), PCH_TRANS_DATA_N1(pipe),
+		      PCH_TRANS_LINK_M1(pipe), PCH_TRANS_LINK_N1(pipe));
+}
+
+static void intel_pch_transcoder_set_m2_n2(struct intel_crtc *crtc,
+					   const struct intel_link_m_n *m_n)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	intel_set_m_n(dev_priv, m_n,
+		      PCH_TRANS_DATA_M2(pipe), PCH_TRANS_DATA_N2(pipe),
+		      PCH_TRANS_LINK_M2(pipe), PCH_TRANS_LINK_N2(pipe));
+}
+
+void intel_pch_transcoder_get_m1_n1(struct intel_crtc *crtc,
+				    struct intel_link_m_n *m_n)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	intel_get_m_n(dev_priv, m_n,
+		      PCH_TRANS_DATA_M1(pipe), PCH_TRANS_DATA_N1(pipe),
+		      PCH_TRANS_LINK_M1(pipe), PCH_TRANS_LINK_N1(pipe));
+}
+
+void intel_pch_transcoder_get_m2_n2(struct intel_crtc *crtc,
+				    struct intel_link_m_n *m_n)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	intel_get_m_n(dev_priv, m_n,
+		      PCH_TRANS_DATA_M2(pipe), PCH_TRANS_DATA_N2(pipe),
+		      PCH_TRANS_LINK_M2(pipe), PCH_TRANS_LINK_N2(pipe));
+}
+
+static void ilk_pch_transcoder_set_timings(const struct intel_crtc_state *crtc_state,
+					   enum pipe pch_transcoder)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	intel_de_write(dev_priv, PCH_TRANS_HTOTAL(pch_transcoder),
+		       intel_de_read(dev_priv, HTOTAL(cpu_transcoder)));
+	intel_de_write(dev_priv, PCH_TRANS_HBLANK(pch_transcoder),
+		       intel_de_read(dev_priv, HBLANK(cpu_transcoder)));
+	intel_de_write(dev_priv, PCH_TRANS_HSYNC(pch_transcoder),
+		       intel_de_read(dev_priv, HSYNC(cpu_transcoder)));
+
+	intel_de_write(dev_priv, PCH_TRANS_VTOTAL(pch_transcoder),
+		       intel_de_read(dev_priv, VTOTAL(cpu_transcoder)));
+	intel_de_write(dev_priv, PCH_TRANS_VBLANK(pch_transcoder),
+		       intel_de_read(dev_priv, VBLANK(cpu_transcoder)));
+	intel_de_write(dev_priv, PCH_TRANS_VSYNC(pch_transcoder),
+		       intel_de_read(dev_priv, VSYNC(cpu_transcoder)));
+	intel_de_write(dev_priv, PCH_TRANS_VSYNCSHIFT(pch_transcoder),
+		       intel_de_read(dev_priv, VSYNCSHIFT(cpu_transcoder)));
+}
+
+static void ilk_enable_pch_transcoder(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 val, pipeconf_val;
+
+	/* Make sure PCH DPLL is enabled */
+	assert_shared_dpll_enabled(dev_priv, crtc_state->shared_dpll);
+
+	/* FDI must be feeding us bits for PCH ports */
+	assert_fdi_tx_enabled(dev_priv, pipe);
+	assert_fdi_rx_enabled(dev_priv, pipe);
+
+	if (HAS_PCH_CPT(dev_priv)) {
+		reg = TRANS_CHICKEN2(pipe);
+		val = intel_de_read(dev_priv, reg);
+		/*
+		 * Workaround: Set the timing override bit
+		 * before enabling the pch transcoder.
+		 */
+		val |= TRANS_CHICKEN2_TIMING_OVERRIDE;
+		/* Configure frame start delay to match the CPU */
+		val &= ~TRANS_CHICKEN2_FRAME_START_DELAY_MASK;
+		val |= TRANS_CHICKEN2_FRAME_START_DELAY(crtc_state->framestart_delay - 1);
+		intel_de_write(dev_priv, reg, val);
+	}
+
+	reg = PCH_TRANSCONF(pipe);
+	val = intel_de_read(dev_priv, reg);
+	pipeconf_val = intel_de_read(dev_priv, PIPECONF(pipe));
+
+	if (HAS_PCH_IBX(dev_priv)) {
+		/* Configure frame start delay to match the CPU */
+		val &= ~TRANS_FRAME_START_DELAY_MASK;
+		val |= TRANS_FRAME_START_DELAY(crtc_state->framestart_delay - 1);
+
+		/*
+		 * Make the BPC in transcoder be consistent with
+		 * that in pipeconf reg. For HDMI we must use 8bpc
+		 * here for both 8bpc and 12bpc.
+		 */
+		val &= ~PIPECONF_BPC_MASK;
+		if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI))
+			val |= PIPECONF_BPC_8;
+		else
+			val |= pipeconf_val & PIPECONF_BPC_MASK;
+	}
+
+	val &= ~TRANS_INTERLACE_MASK;
+	if ((pipeconf_val & PIPECONF_INTERLACE_MASK_ILK) == PIPECONF_INTERLACE_IF_ID_ILK) {
+		if (HAS_PCH_IBX(dev_priv) &&
+		    intel_crtc_has_type(crtc_state, INTEL_OUTPUT_SDVO))
+			val |= TRANS_INTERLACE_LEGACY_VSYNC_IBX;
+		else
+			val |= TRANS_INTERLACE_INTERLACED;
+	} else {
+		val |= TRANS_INTERLACE_PROGRESSIVE;
+	}
+
+	intel_de_write(dev_priv, reg, val | TRANS_ENABLE);
+	if (intel_de_wait_for_set(dev_priv, reg, TRANS_STATE_ENABLE, 100))
+		drm_err(&dev_priv->drm, "failed to enable transcoder %c\n",
+			pipe_name(pipe));
+}
+
+static void ilk_disable_pch_transcoder(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	i915_reg_t reg;
+	u32 val;
+
+	/* FDI relies on the transcoder */
+	assert_fdi_tx_disabled(dev_priv, pipe);
+	assert_fdi_rx_disabled(dev_priv, pipe);
+
+	/* Ports must be off as well */
+	assert_pch_ports_disabled(dev_priv, pipe);
+
+	reg = PCH_TRANSCONF(pipe);
+	val = intel_de_read(dev_priv, reg);
+	val &= ~TRANS_ENABLE;
+	intel_de_write(dev_priv, reg, val);
+	/* wait for PCH transcoder off, transcoder state */
+	if (intel_de_wait_for_clear(dev_priv, reg, TRANS_STATE_ENABLE, 50))
+		drm_err(&dev_priv->drm, "failed to disable transcoder %c\n",
+			pipe_name(pipe));
+
+	if (HAS_PCH_CPT(dev_priv)) {
+		/* Workaround: Clear the timing override chicken bit again. */
+		reg = TRANS_CHICKEN2(pipe);
+		val = intel_de_read(dev_priv, reg);
+		val &= ~TRANS_CHICKEN2_TIMING_OVERRIDE;
+		intel_de_write(dev_priv, reg, val);
+	}
+}
+
+void ilk_pch_pre_enable(struct intel_atomic_state *state,
+			struct intel_crtc *crtc)
+{
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	/*
+	 * Note: FDI PLL enabling _must_ be done before we enable the
+	 * cpu pipes, hence this is separate from all the other fdi/pch
+	 * enabling.
+	 */
+	ilk_fdi_pll_enable(crtc_state);
+}
+
+/*
+ * Enable PCH resources required for PCH ports:
+ *   - PCH PLLs
+ *   - FDI training & RX/TX
+ *   - update transcoder timings
+ *   - DP transcoding bits
+ *   - transcoder
+ */
+void ilk_pch_enable(struct intel_atomic_state *state,
+		    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	enum pipe pipe = crtc->pipe;
+	u32 temp;
+
+	assert_pch_transcoder_disabled(dev_priv, pipe);
+
+	/* For PCH output, training FDI link */
+	intel_fdi_link_train(crtc, crtc_state);
+
+	/*
+	 * We need to program the right clock selection
+	 * before writing the pixel multiplier into the DPLL.
+	 */
+	if (HAS_PCH_CPT(dev_priv)) {
+		u32 sel;
+
+		temp = intel_de_read(dev_priv, PCH_DPLL_SEL);
+		temp |= TRANS_DPLL_ENABLE(pipe);
+		sel = TRANS_DPLLB_SEL(pipe);
+		if (crtc_state->shared_dpll ==
+		    intel_get_shared_dpll_by_id(dev_priv, DPLL_ID_PCH_PLL_B))
+			temp |= sel;
+		else
+			temp &= ~sel;
+		intel_de_write(dev_priv, PCH_DPLL_SEL, temp);
+	}
+
+	/*
+	 * XXX: pch pll's can be enabled any time before we enable the PCH
+	 * transcoder, and we actually should do this to not upset any PCH
+	 * transcoder that already use the clock when we share it.
+	 *
+	 * Note that enable_shared_dpll tries to do the right thing, but
+	 * get_shared_dpll unconditionally resets the pll - we need that
+	 * to have the right LVDS enable sequence.
+	 */
+	intel_enable_shared_dpll(crtc_state);
+
+	/* set transcoder timing, panel must allow it */
+	assert_pps_unlocked(dev_priv, pipe);
+	if (intel_crtc_has_dp_encoder(crtc_state)) {
+		intel_pch_transcoder_set_m1_n1(crtc, &crtc_state->dp_m_n);
+		intel_pch_transcoder_set_m2_n2(crtc, &crtc_state->dp_m2_n2);
+	}
+	ilk_pch_transcoder_set_timings(crtc_state, pipe);
+
+	intel_fdi_normal_train(crtc);
+
+	/* For PCH DP, enable TRANS_DP_CTL */
+	if (HAS_PCH_CPT(dev_priv) &&
+	    intel_crtc_has_dp_encoder(crtc_state)) {
+		const struct drm_display_mode *adjusted_mode =
+			&crtc_state->hw.adjusted_mode;
+		u32 bpc = (intel_de_read(dev_priv, PIPECONF(pipe)) & PIPECONF_BPC_MASK) >> 5;
+		i915_reg_t reg = TRANS_DP_CTL(pipe);
+		enum port port;
+
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~(TRANS_DP_PORT_SEL_MASK |
+			  TRANS_DP_VSYNC_ACTIVE_HIGH |
+			  TRANS_DP_HSYNC_ACTIVE_HIGH |
+			  TRANS_DP_BPC_MASK);
+		temp |= TRANS_DP_OUTPUT_ENABLE;
+		temp |= bpc << 9; /* same format but at 11:9 */
+
+		if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
+			temp |= TRANS_DP_HSYNC_ACTIVE_HIGH;
+		if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
+			temp |= TRANS_DP_VSYNC_ACTIVE_HIGH;
+
+		port = intel_get_crtc_new_encoder(state, crtc_state)->port;
+		drm_WARN_ON(&dev_priv->drm, port < PORT_B || port > PORT_D);
+		temp |= TRANS_DP_PORT_SEL(port);
+
+		intel_de_write(dev_priv, reg, temp);
+	}
+
+	ilk_enable_pch_transcoder(crtc_state);
+}
+
+void ilk_pch_disable(struct intel_atomic_state *state,
+		     struct intel_crtc *crtc)
+{
+	ilk_fdi_disable(crtc);
+}
+
+void ilk_pch_post_disable(struct intel_atomic_state *state,
+			  struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	ilk_disable_pch_transcoder(crtc);
+
+	if (HAS_PCH_CPT(dev_priv)) {
+		i915_reg_t reg;
+		u32 temp;
+
+		/* disable TRANS_DP_CTL */
+		reg = TRANS_DP_CTL(pipe);
+		temp = intel_de_read(dev_priv, reg);
+		temp &= ~(TRANS_DP_OUTPUT_ENABLE |
+			  TRANS_DP_PORT_SEL_MASK);
+		temp |= TRANS_DP_PORT_SEL_NONE;
+		intel_de_write(dev_priv, reg, temp);
+
+		/* disable DPLL_SEL */
+		temp = intel_de_read(dev_priv, PCH_DPLL_SEL);
+		temp &= ~(TRANS_DPLL_ENABLE(pipe) | TRANS_DPLLB_SEL(pipe));
+		intel_de_write(dev_priv, PCH_DPLL_SEL, temp);
+	}
+
+	ilk_fdi_pll_disable(crtc);
+}
+
+static void ilk_pch_clock_get(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	/* read out port_clock from the DPLL */
+	i9xx_crtc_clock_get(crtc, crtc_state);
+
+	/*
+	 * In case there is an active pipe without active ports,
+	 * we may need some idea for the dotclock anyway.
+	 * Calculate one based on the FDI configuration.
+	 */
+	crtc_state->hw.adjusted_mode.crtc_clock =
+		intel_dotclock_calculate(intel_fdi_link_freq(dev_priv, crtc_state),
+					 &crtc_state->fdi_m_n);
+}
+
+void ilk_pch_get_config(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_shared_dpll *pll;
+	enum pipe pipe = crtc->pipe;
+	enum intel_dpll_id pll_id;
+	bool pll_active;
+	u32 tmp;
+
+	if ((intel_de_read(dev_priv, PCH_TRANSCONF(pipe)) & TRANS_ENABLE) == 0)
+		return;
+
+	crtc_state->has_pch_encoder = true;
+
+	tmp = intel_de_read(dev_priv, FDI_RX_CTL(pipe));
+	crtc_state->fdi_lanes = ((FDI_DP_PORT_WIDTH_MASK & tmp) >>
+				 FDI_DP_PORT_WIDTH_SHIFT) + 1;
+
+	intel_cpu_transcoder_get_m1_n1(crtc, crtc_state->cpu_transcoder,
+				       &crtc_state->fdi_m_n);
+
+	if (HAS_PCH_IBX(dev_priv)) {
+		/*
+		 * The pipe->pch transcoder and pch transcoder->pll
+		 * mapping is fixed.
+		 */
+		pll_id = (enum intel_dpll_id) pipe;
+	} else {
+		tmp = intel_de_read(dev_priv, PCH_DPLL_SEL);
+		if (tmp & TRANS_DPLLB_SEL(pipe))
+			pll_id = DPLL_ID_PCH_PLL_B;
+		else
+			pll_id = DPLL_ID_PCH_PLL_A;
+	}
+
+	crtc_state->shared_dpll = intel_get_shared_dpll_by_id(dev_priv, pll_id);
+	pll = crtc_state->shared_dpll;
+
+	pll_active = intel_dpll_get_hw_state(dev_priv, pll,
+					     &crtc_state->dpll_hw_state);
+	drm_WARN_ON(&dev_priv->drm, !pll_active);
+
+	tmp = crtc_state->dpll_hw_state.dpll;
+	crtc_state->pixel_multiplier =
+		((tmp & PLL_REF_SDVO_HDMI_MULTIPLIER_MASK)
+		 >> PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT) + 1;
+
+	ilk_pch_clock_get(crtc_state);
+}
+
+static void lpt_enable_pch_transcoder(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 val, pipeconf_val;
+
+	/* FDI must be feeding us bits for PCH ports */
+	assert_fdi_tx_enabled(dev_priv, (enum pipe) cpu_transcoder);
+	assert_fdi_rx_enabled(dev_priv, PIPE_A);
+
+	val = intel_de_read(dev_priv, TRANS_CHICKEN2(PIPE_A));
+	/* Workaround: set timing override bit. */
+	val |= TRANS_CHICKEN2_TIMING_OVERRIDE;
+	/* Configure frame start delay to match the CPU */
+	val &= ~TRANS_CHICKEN2_FRAME_START_DELAY_MASK;
+	val |= TRANS_CHICKEN2_FRAME_START_DELAY(crtc_state->framestart_delay - 1);
+	intel_de_write(dev_priv, TRANS_CHICKEN2(PIPE_A), val);
+
+	val = TRANS_ENABLE;
+	pipeconf_val = intel_de_read(dev_priv, PIPECONF(cpu_transcoder));
+
+	if ((pipeconf_val & PIPECONF_INTERLACE_MASK_HSW) == PIPECONF_INTERLACE_IF_ID_ILK)
+		val |= TRANS_INTERLACE_INTERLACED;
+	else
+		val |= TRANS_INTERLACE_PROGRESSIVE;
+
+	intel_de_write(dev_priv, LPT_TRANSCONF, val);
+	if (intel_de_wait_for_set(dev_priv, LPT_TRANSCONF,
+				  TRANS_STATE_ENABLE, 100))
+		drm_err(&dev_priv->drm, "Failed to enable PCH transcoder\n");
+}
+
+static void lpt_disable_pch_transcoder(struct drm_i915_private *dev_priv)
+{
+	u32 val;
+
+	val = intel_de_read(dev_priv, LPT_TRANSCONF);
+	val &= ~TRANS_ENABLE;
+	intel_de_write(dev_priv, LPT_TRANSCONF, val);
+	/* wait for PCH transcoder off, transcoder state */
+	if (intel_de_wait_for_clear(dev_priv, LPT_TRANSCONF,
+				    TRANS_STATE_ENABLE, 50))
+		drm_err(&dev_priv->drm, "Failed to disable PCH transcoder\n");
+
+	/* Workaround: clear timing override bit. */
+	val = intel_de_read(dev_priv, TRANS_CHICKEN2(PIPE_A));
+	val &= ~TRANS_CHICKEN2_TIMING_OVERRIDE;
+	intel_de_write(dev_priv, TRANS_CHICKEN2(PIPE_A), val);
+}
+
+void lpt_pch_enable(struct intel_atomic_state *state,
+		    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	assert_pch_transcoder_disabled(dev_priv, PIPE_A);
+
+	lpt_program_iclkip(crtc_state);
+
+	/* Set transcoder timing. */
+	ilk_pch_transcoder_set_timings(crtc_state, PIPE_A);
+
+	lpt_enable_pch_transcoder(crtc_state);
+}
+
+void lpt_pch_disable(struct intel_atomic_state *state,
+		     struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	lpt_disable_pch_transcoder(dev_priv);
+
+	lpt_disable_iclkip(dev_priv);
+}
+
+void lpt_pch_get_config(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 tmp;
+
+	if ((intel_de_read(dev_priv, LPT_TRANSCONF) & TRANS_ENABLE) == 0)
+		return;
+
+	crtc_state->has_pch_encoder = true;
+
+	tmp = intel_de_read(dev_priv, FDI_RX_CTL(PIPE_A));
+	crtc_state->fdi_lanes = ((FDI_DP_PORT_WIDTH_MASK & tmp) >>
+				 FDI_DP_PORT_WIDTH_SHIFT) + 1;
+
+	intel_cpu_transcoder_get_m1_n1(crtc, crtc_state->cpu_transcoder,
+				       &crtc_state->fdi_m_n);
+
+	crtc_state->hw.adjusted_mode.crtc_clock = lpt_get_iclkip(dev_priv);
+}
+
+void intel_pch_sanitize(struct drm_i915_private *i915)
+{
+	if (HAS_PCH_IBX(i915))
+		ibx_sanitize_pch_ports(i915);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_pch_display.h b/drivers/gpu/drm/i915/display/intel_pch_display.h
new file mode 100644
index 000000000..41a63413c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_pch_display.h
@@ -0,0 +1,45 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _INTEL_PCH_DISPLAY_H_
+#define _INTEL_PCH_DISPLAY_H_
+
+#include <linux/types.h>
+
+enum pipe;
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_link_m_n;
+
+bool intel_has_pch_trancoder(struct drm_i915_private *i915,
+			     enum pipe pch_transcoder);
+enum pipe intel_crtc_pch_transcoder(struct intel_crtc *crtc);
+
+void ilk_pch_pre_enable(struct intel_atomic_state *state,
+			struct intel_crtc *crtc);
+void ilk_pch_enable(struct intel_atomic_state *state,
+		    struct intel_crtc *crtc);
+void ilk_pch_disable(struct intel_atomic_state *state,
+		     struct intel_crtc *crtc);
+void ilk_pch_post_disable(struct intel_atomic_state *state,
+			  struct intel_crtc *crtc);
+void ilk_pch_get_config(struct intel_crtc_state *crtc_state);
+
+void lpt_pch_enable(struct intel_atomic_state *state,
+		    struct intel_crtc *crtc);
+void lpt_pch_disable(struct intel_atomic_state *state,
+		     struct intel_crtc *crtc);
+void lpt_pch_get_config(struct intel_crtc_state *crtc_state);
+
+void intel_pch_transcoder_get_m1_n1(struct intel_crtc *crtc,
+				    struct intel_link_m_n *m_n);
+void intel_pch_transcoder_get_m2_n2(struct intel_crtc *crtc,
+				    struct intel_link_m_n *m_n);
+
+void intel_pch_sanitize(struct drm_i915_private *i915);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_pch_refclk.c b/drivers/gpu/drm/i915/display/intel_pch_refclk.c
new file mode 100644
index 000000000..a66097cdc
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_pch_refclk.c
@@ -0,0 +1,679 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_panel.h"
+#include "intel_pch_refclk.h"
+#include "intel_sbi.h"
+
+static void lpt_fdi_reset_mphy(struct drm_i915_private *dev_priv)
+{
+	u32 tmp;
+
+	tmp = intel_de_read(dev_priv, SOUTH_CHICKEN2);
+	tmp |= FDI_MPHY_IOSFSB_RESET_CTL;
+	intel_de_write(dev_priv, SOUTH_CHICKEN2, tmp);
+
+	if (wait_for_us(intel_de_read(dev_priv, SOUTH_CHICKEN2) &
+			FDI_MPHY_IOSFSB_RESET_STATUS, 100))
+		drm_err(&dev_priv->drm, "FDI mPHY reset assert timeout\n");
+
+	tmp = intel_de_read(dev_priv, SOUTH_CHICKEN2);
+	tmp &= ~FDI_MPHY_IOSFSB_RESET_CTL;
+	intel_de_write(dev_priv, SOUTH_CHICKEN2, tmp);
+
+	if (wait_for_us((intel_de_read(dev_priv, SOUTH_CHICKEN2) &
+			 FDI_MPHY_IOSFSB_RESET_STATUS) == 0, 100))
+		drm_err(&dev_priv->drm, "FDI mPHY reset de-assert timeout\n");
+}
+
+/* WaMPhyProgramming:hsw */
+static void lpt_fdi_program_mphy(struct drm_i915_private *dev_priv)
+{
+	u32 tmp;
+
+	lpt_fdi_reset_mphy(dev_priv);
+
+	tmp = intel_sbi_read(dev_priv, 0x8008, SBI_MPHY);
+	tmp &= ~(0xFF << 24);
+	tmp |= (0x12 << 24);
+	intel_sbi_write(dev_priv, 0x8008, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x2008, SBI_MPHY);
+	tmp |= (1 << 11);
+	intel_sbi_write(dev_priv, 0x2008, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x2108, SBI_MPHY);
+	tmp |= (1 << 11);
+	intel_sbi_write(dev_priv, 0x2108, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x206C, SBI_MPHY);
+	tmp |= (1 << 24) | (1 << 21) | (1 << 18);
+	intel_sbi_write(dev_priv, 0x206C, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x216C, SBI_MPHY);
+	tmp |= (1 << 24) | (1 << 21) | (1 << 18);
+	intel_sbi_write(dev_priv, 0x216C, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x2080, SBI_MPHY);
+	tmp &= ~(7 << 13);
+	tmp |= (5 << 13);
+	intel_sbi_write(dev_priv, 0x2080, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x2180, SBI_MPHY);
+	tmp &= ~(7 << 13);
+	tmp |= (5 << 13);
+	intel_sbi_write(dev_priv, 0x2180, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x208C, SBI_MPHY);
+	tmp &= ~0xFF;
+	tmp |= 0x1C;
+	intel_sbi_write(dev_priv, 0x208C, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x218C, SBI_MPHY);
+	tmp &= ~0xFF;
+	tmp |= 0x1C;
+	intel_sbi_write(dev_priv, 0x218C, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x2098, SBI_MPHY);
+	tmp &= ~(0xFF << 16);
+	tmp |= (0x1C << 16);
+	intel_sbi_write(dev_priv, 0x2098, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x2198, SBI_MPHY);
+	tmp &= ~(0xFF << 16);
+	tmp |= (0x1C << 16);
+	intel_sbi_write(dev_priv, 0x2198, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x20C4, SBI_MPHY);
+	tmp |= (1 << 27);
+	intel_sbi_write(dev_priv, 0x20C4, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x21C4, SBI_MPHY);
+	tmp |= (1 << 27);
+	intel_sbi_write(dev_priv, 0x21C4, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x20EC, SBI_MPHY);
+	tmp &= ~(0xF << 28);
+	tmp |= (4 << 28);
+	intel_sbi_write(dev_priv, 0x20EC, tmp, SBI_MPHY);
+
+	tmp = intel_sbi_read(dev_priv, 0x21EC, SBI_MPHY);
+	tmp &= ~(0xF << 28);
+	tmp |= (4 << 28);
+	intel_sbi_write(dev_priv, 0x21EC, tmp, SBI_MPHY);
+}
+
+void lpt_disable_iclkip(struct drm_i915_private *dev_priv)
+{
+	u32 temp;
+
+	intel_de_write(dev_priv, PIXCLK_GATE, PIXCLK_GATE_GATE);
+
+	mutex_lock(&dev_priv->sb_lock);
+
+	temp = intel_sbi_read(dev_priv, SBI_SSCCTL6, SBI_ICLK);
+	temp |= SBI_SSCCTL_DISABLE;
+	intel_sbi_write(dev_priv, SBI_SSCCTL6, temp, SBI_ICLK);
+
+	mutex_unlock(&dev_priv->sb_lock);
+}
+
+struct iclkip_params {
+	u32 iclk_virtual_root_freq;
+	u32 iclk_pi_range;
+	u32 divsel, phaseinc, auxdiv, phasedir, desired_divisor;
+};
+
+static void iclkip_params_init(struct iclkip_params *p)
+{
+	memset(p, 0, sizeof(*p));
+
+	p->iclk_virtual_root_freq = 172800 * 1000;
+	p->iclk_pi_range = 64;
+}
+
+static int lpt_iclkip_freq(struct iclkip_params *p)
+{
+	return DIV_ROUND_CLOSEST(p->iclk_virtual_root_freq,
+				 p->desired_divisor << p->auxdiv);
+}
+
+static void lpt_compute_iclkip(struct iclkip_params *p, int clock)
+{
+	iclkip_params_init(p);
+
+	/* The iCLK virtual clock root frequency is in MHz,
+	 * but the adjusted_mode->crtc_clock in KHz. To get the
+	 * divisors, it is necessary to divide one by another, so we
+	 * convert the virtual clock precision to KHz here for higher
+	 * precision.
+	 */
+	for (p->auxdiv = 0; p->auxdiv < 2; p->auxdiv++) {
+		p->desired_divisor = DIV_ROUND_CLOSEST(p->iclk_virtual_root_freq,
+						       clock << p->auxdiv);
+		p->divsel = (p->desired_divisor / p->iclk_pi_range) - 2;
+		p->phaseinc = p->desired_divisor % p->iclk_pi_range;
+
+		/*
+		 * Near 20MHz is a corner case which is
+		 * out of range for the 7-bit divisor
+		 */
+		if (p->divsel <= 0x7f)
+			break;
+	}
+}
+
+int lpt_iclkip(const struct intel_crtc_state *crtc_state)
+{
+	struct iclkip_params p;
+
+	lpt_compute_iclkip(&p, crtc_state->hw.adjusted_mode.crtc_clock);
+
+	return lpt_iclkip_freq(&p);
+}
+
+/* Program iCLKIP clock to the desired frequency */
+void lpt_program_iclkip(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	int clock = crtc_state->hw.adjusted_mode.crtc_clock;
+	struct iclkip_params p;
+	u32 temp;
+
+	lpt_disable_iclkip(dev_priv);
+
+	lpt_compute_iclkip(&p, clock);
+	drm_WARN_ON(&dev_priv->drm, lpt_iclkip_freq(&p) != clock);
+
+	/* This should not happen with any sane values */
+	drm_WARN_ON(&dev_priv->drm, SBI_SSCDIVINTPHASE_DIVSEL(p.divsel) &
+		    ~SBI_SSCDIVINTPHASE_DIVSEL_MASK);
+	drm_WARN_ON(&dev_priv->drm, SBI_SSCDIVINTPHASE_DIR(p.phasedir) &
+		    ~SBI_SSCDIVINTPHASE_INCVAL_MASK);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "iCLKIP clock: found settings for %dKHz refresh rate: auxdiv=%x, divsel=%x, phasedir=%x, phaseinc=%x\n",
+		    clock, p.auxdiv, p.divsel, p.phasedir, p.phaseinc);
+
+	mutex_lock(&dev_priv->sb_lock);
+
+	/* Program SSCDIVINTPHASE6 */
+	temp = intel_sbi_read(dev_priv, SBI_SSCDIVINTPHASE6, SBI_ICLK);
+	temp &= ~SBI_SSCDIVINTPHASE_DIVSEL_MASK;
+	temp |= SBI_SSCDIVINTPHASE_DIVSEL(p.divsel);
+	temp &= ~SBI_SSCDIVINTPHASE_INCVAL_MASK;
+	temp |= SBI_SSCDIVINTPHASE_INCVAL(p.phaseinc);
+	temp |= SBI_SSCDIVINTPHASE_DIR(p.phasedir);
+	temp |= SBI_SSCDIVINTPHASE_PROPAGATE;
+	intel_sbi_write(dev_priv, SBI_SSCDIVINTPHASE6, temp, SBI_ICLK);
+
+	/* Program SSCAUXDIV */
+	temp = intel_sbi_read(dev_priv, SBI_SSCAUXDIV6, SBI_ICLK);
+	temp &= ~SBI_SSCAUXDIV_FINALDIV2SEL(1);
+	temp |= SBI_SSCAUXDIV_FINALDIV2SEL(p.auxdiv);
+	intel_sbi_write(dev_priv, SBI_SSCAUXDIV6, temp, SBI_ICLK);
+
+	/* Enable modulator and associated divider */
+	temp = intel_sbi_read(dev_priv, SBI_SSCCTL6, SBI_ICLK);
+	temp &= ~SBI_SSCCTL_DISABLE;
+	intel_sbi_write(dev_priv, SBI_SSCCTL6, temp, SBI_ICLK);
+
+	mutex_unlock(&dev_priv->sb_lock);
+
+	/* Wait for initialization time */
+	udelay(24);
+
+	intel_de_write(dev_priv, PIXCLK_GATE, PIXCLK_GATE_UNGATE);
+}
+
+int lpt_get_iclkip(struct drm_i915_private *dev_priv)
+{
+	struct iclkip_params p;
+	u32 temp;
+
+	if ((intel_de_read(dev_priv, PIXCLK_GATE) & PIXCLK_GATE_UNGATE) == 0)
+		return 0;
+
+	iclkip_params_init(&p);
+
+	mutex_lock(&dev_priv->sb_lock);
+
+	temp = intel_sbi_read(dev_priv, SBI_SSCCTL6, SBI_ICLK);
+	if (temp & SBI_SSCCTL_DISABLE) {
+		mutex_unlock(&dev_priv->sb_lock);
+		return 0;
+	}
+
+	temp = intel_sbi_read(dev_priv, SBI_SSCDIVINTPHASE6, SBI_ICLK);
+	p.divsel = (temp & SBI_SSCDIVINTPHASE_DIVSEL_MASK) >>
+		SBI_SSCDIVINTPHASE_DIVSEL_SHIFT;
+	p.phaseinc = (temp & SBI_SSCDIVINTPHASE_INCVAL_MASK) >>
+		SBI_SSCDIVINTPHASE_INCVAL_SHIFT;
+
+	temp = intel_sbi_read(dev_priv, SBI_SSCAUXDIV6, SBI_ICLK);
+	p.auxdiv = (temp & SBI_SSCAUXDIV_FINALDIV2SEL_MASK) >>
+		SBI_SSCAUXDIV_FINALDIV2SEL_SHIFT;
+
+	mutex_unlock(&dev_priv->sb_lock);
+
+	p.desired_divisor = (p.divsel + 2) * p.iclk_pi_range + p.phaseinc;
+
+	return lpt_iclkip_freq(&p);
+}
+
+/* Implements 3 different sequences from BSpec chapter "Display iCLK
+ * Programming" based on the parameters passed:
+ * - Sequence to enable CLKOUT_DP
+ * - Sequence to enable CLKOUT_DP without spread
+ * - Sequence to enable CLKOUT_DP for FDI usage and configure PCH FDI I/O
+ */
+static void lpt_enable_clkout_dp(struct drm_i915_private *dev_priv,
+				 bool with_spread, bool with_fdi)
+{
+	u32 reg, tmp;
+
+	if (drm_WARN(&dev_priv->drm, with_fdi && !with_spread,
+		     "FDI requires downspread\n"))
+		with_spread = true;
+	if (drm_WARN(&dev_priv->drm, HAS_PCH_LPT_LP(dev_priv) &&
+		     with_fdi, "LP PCH doesn't have FDI\n"))
+		with_fdi = false;
+
+	mutex_lock(&dev_priv->sb_lock);
+
+	tmp = intel_sbi_read(dev_priv, SBI_SSCCTL, SBI_ICLK);
+	tmp &= ~SBI_SSCCTL_DISABLE;
+	tmp |= SBI_SSCCTL_PATHALT;
+	intel_sbi_write(dev_priv, SBI_SSCCTL, tmp, SBI_ICLK);
+
+	udelay(24);
+
+	if (with_spread) {
+		tmp = intel_sbi_read(dev_priv, SBI_SSCCTL, SBI_ICLK);
+		tmp &= ~SBI_SSCCTL_PATHALT;
+		intel_sbi_write(dev_priv, SBI_SSCCTL, tmp, SBI_ICLK);
+
+		if (with_fdi)
+			lpt_fdi_program_mphy(dev_priv);
+	}
+
+	reg = HAS_PCH_LPT_LP(dev_priv) ? SBI_GEN0 : SBI_DBUFF0;
+	tmp = intel_sbi_read(dev_priv, reg, SBI_ICLK);
+	tmp |= SBI_GEN0_CFG_BUFFENABLE_DISABLE;
+	intel_sbi_write(dev_priv, reg, tmp, SBI_ICLK);
+
+	mutex_unlock(&dev_priv->sb_lock);
+}
+
+/* Sequence to disable CLKOUT_DP */
+void lpt_disable_clkout_dp(struct drm_i915_private *dev_priv)
+{
+	u32 reg, tmp;
+
+	mutex_lock(&dev_priv->sb_lock);
+
+	reg = HAS_PCH_LPT_LP(dev_priv) ? SBI_GEN0 : SBI_DBUFF0;
+	tmp = intel_sbi_read(dev_priv, reg, SBI_ICLK);
+	tmp &= ~SBI_GEN0_CFG_BUFFENABLE_DISABLE;
+	intel_sbi_write(dev_priv, reg, tmp, SBI_ICLK);
+
+	tmp = intel_sbi_read(dev_priv, SBI_SSCCTL, SBI_ICLK);
+	if (!(tmp & SBI_SSCCTL_DISABLE)) {
+		if (!(tmp & SBI_SSCCTL_PATHALT)) {
+			tmp |= SBI_SSCCTL_PATHALT;
+			intel_sbi_write(dev_priv, SBI_SSCCTL, tmp, SBI_ICLK);
+			udelay(32);
+		}
+		tmp |= SBI_SSCCTL_DISABLE;
+		intel_sbi_write(dev_priv, SBI_SSCCTL, tmp, SBI_ICLK);
+	}
+
+	mutex_unlock(&dev_priv->sb_lock);
+}
+
+#define BEND_IDX(steps) ((50 + (steps)) / 5)
+
+static const u16 sscdivintphase[] = {
+	[BEND_IDX( 50)] = 0x3B23,
+	[BEND_IDX( 45)] = 0x3B23,
+	[BEND_IDX( 40)] = 0x3C23,
+	[BEND_IDX( 35)] = 0x3C23,
+	[BEND_IDX( 30)] = 0x3D23,
+	[BEND_IDX( 25)] = 0x3D23,
+	[BEND_IDX( 20)] = 0x3E23,
+	[BEND_IDX( 15)] = 0x3E23,
+	[BEND_IDX( 10)] = 0x3F23,
+	[BEND_IDX(  5)] = 0x3F23,
+	[BEND_IDX(  0)] = 0x0025,
+	[BEND_IDX( -5)] = 0x0025,
+	[BEND_IDX(-10)] = 0x0125,
+	[BEND_IDX(-15)] = 0x0125,
+	[BEND_IDX(-20)] = 0x0225,
+	[BEND_IDX(-25)] = 0x0225,
+	[BEND_IDX(-30)] = 0x0325,
+	[BEND_IDX(-35)] = 0x0325,
+	[BEND_IDX(-40)] = 0x0425,
+	[BEND_IDX(-45)] = 0x0425,
+	[BEND_IDX(-50)] = 0x0525,
+};
+
+/*
+ * Bend CLKOUT_DP
+ * steps -50 to 50 inclusive, in steps of 5
+ * < 0 slow down the clock, > 0 speed up the clock, 0 == no bend (135MHz)
+ * change in clock period = -(steps / 10) * 5.787 ps
+ */
+static void lpt_bend_clkout_dp(struct drm_i915_private *dev_priv, int steps)
+{
+	u32 tmp;
+	int idx = BEND_IDX(steps);
+
+	if (drm_WARN_ON(&dev_priv->drm, steps % 5 != 0))
+		return;
+
+	if (drm_WARN_ON(&dev_priv->drm, idx >= ARRAY_SIZE(sscdivintphase)))
+		return;
+
+	mutex_lock(&dev_priv->sb_lock);
+
+	if (steps % 10 != 0)
+		tmp = 0xAAAAAAAB;
+	else
+		tmp = 0x00000000;
+	intel_sbi_write(dev_priv, SBI_SSCDITHPHASE, tmp, SBI_ICLK);
+
+	tmp = intel_sbi_read(dev_priv, SBI_SSCDIVINTPHASE, SBI_ICLK);
+	tmp &= 0xffff0000;
+	tmp |= sscdivintphase[idx];
+	intel_sbi_write(dev_priv, SBI_SSCDIVINTPHASE, tmp, SBI_ICLK);
+
+	mutex_unlock(&dev_priv->sb_lock);
+}
+
+#undef BEND_IDX
+
+static bool spll_uses_pch_ssc(struct drm_i915_private *dev_priv)
+{
+	u32 fuse_strap = intel_de_read(dev_priv, FUSE_STRAP);
+	u32 ctl = intel_de_read(dev_priv, SPLL_CTL);
+
+	if ((ctl & SPLL_PLL_ENABLE) == 0)
+		return false;
+
+	if ((ctl & SPLL_REF_MASK) == SPLL_REF_MUXED_SSC &&
+	    (fuse_strap & HSW_CPU_SSC_ENABLE) == 0)
+		return true;
+
+	if (IS_BROADWELL(dev_priv) &&
+	    (ctl & SPLL_REF_MASK) == SPLL_REF_PCH_SSC_BDW)
+		return true;
+
+	return false;
+}
+
+static bool wrpll_uses_pch_ssc(struct drm_i915_private *dev_priv,
+			       enum intel_dpll_id id)
+{
+	u32 fuse_strap = intel_de_read(dev_priv, FUSE_STRAP);
+	u32 ctl = intel_de_read(dev_priv, WRPLL_CTL(id));
+
+	if ((ctl & WRPLL_PLL_ENABLE) == 0)
+		return false;
+
+	if ((ctl & WRPLL_REF_MASK) == WRPLL_REF_PCH_SSC)
+		return true;
+
+	if ((IS_BROADWELL(dev_priv) || IS_HSW_ULT(dev_priv)) &&
+	    (ctl & WRPLL_REF_MASK) == WRPLL_REF_MUXED_SSC_BDW &&
+	    (fuse_strap & HSW_CPU_SSC_ENABLE) == 0)
+		return true;
+
+	return false;
+}
+
+static void lpt_init_pch_refclk(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+	bool has_fdi = false;
+
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		switch (encoder->type) {
+		case INTEL_OUTPUT_ANALOG:
+			has_fdi = true;
+			break;
+		default:
+			break;
+		}
+	}
+
+	/*
+	 * The BIOS may have decided to use the PCH SSC
+	 * reference so we must not disable it until the
+	 * relevant PLLs have stopped relying on it. We'll
+	 * just leave the PCH SSC reference enabled in case
+	 * any active PLL is using it. It will get disabled
+	 * after runtime suspend if we don't have FDI.
+	 *
+	 * TODO: Move the whole reference clock handling
+	 * to the modeset sequence proper so that we can
+	 * actually enable/disable/reconfigure these things
+	 * safely. To do that we need to introduce a real
+	 * clock hierarchy. That would also allow us to do
+	 * clock bending finally.
+	 */
+	dev_priv->pch_ssc_use = 0;
+
+	if (spll_uses_pch_ssc(dev_priv)) {
+		drm_dbg_kms(&dev_priv->drm, "SPLL using PCH SSC\n");
+		dev_priv->pch_ssc_use |= BIT(DPLL_ID_SPLL);
+	}
+
+	if (wrpll_uses_pch_ssc(dev_priv, DPLL_ID_WRPLL1)) {
+		drm_dbg_kms(&dev_priv->drm, "WRPLL1 using PCH SSC\n");
+		dev_priv->pch_ssc_use |= BIT(DPLL_ID_WRPLL1);
+	}
+
+	if (wrpll_uses_pch_ssc(dev_priv, DPLL_ID_WRPLL2)) {
+		drm_dbg_kms(&dev_priv->drm, "WRPLL2 using PCH SSC\n");
+		dev_priv->pch_ssc_use |= BIT(DPLL_ID_WRPLL2);
+	}
+
+	if (dev_priv->pch_ssc_use)
+		return;
+
+	if (has_fdi) {
+		lpt_bend_clkout_dp(dev_priv, 0);
+		lpt_enable_clkout_dp(dev_priv, true, true);
+	} else {
+		lpt_disable_clkout_dp(dev_priv);
+	}
+}
+
+static void ilk_init_pch_refclk(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+	int i;
+	u32 val, final;
+	bool has_lvds = false;
+	bool has_cpu_edp = false;
+	bool has_panel = false;
+	bool has_ck505 = false;
+	bool can_ssc = false;
+	bool using_ssc_source = false;
+
+	/* We need to take the global config into account */
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		switch (encoder->type) {
+		case INTEL_OUTPUT_LVDS:
+			has_panel = true;
+			has_lvds = true;
+			break;
+		case INTEL_OUTPUT_EDP:
+			has_panel = true;
+			if (encoder->port == PORT_A)
+				has_cpu_edp = true;
+			break;
+		default:
+			break;
+		}
+	}
+
+	if (HAS_PCH_IBX(dev_priv)) {
+		has_ck505 = dev_priv->display.vbt.display_clock_mode;
+		can_ssc = has_ck505;
+	} else {
+		has_ck505 = false;
+		can_ssc = true;
+	}
+
+	/* Check if any DPLLs are using the SSC source */
+	for (i = 0; i < dev_priv->display.dpll.num_shared_dpll; i++) {
+		u32 temp = intel_de_read(dev_priv, PCH_DPLL(i));
+
+		if (!(temp & DPLL_VCO_ENABLE))
+			continue;
+
+		if ((temp & PLL_REF_INPUT_MASK) ==
+		    PLLB_REF_INPUT_SPREADSPECTRUMIN) {
+			using_ssc_source = true;
+			break;
+		}
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "has_panel %d has_lvds %d has_ck505 %d using_ssc_source %d\n",
+		    has_panel, has_lvds, has_ck505, using_ssc_source);
+
+	/* Ironlake: try to setup display ref clock before DPLL
+	 * enabling. This is only under driver's control after
+	 * PCH B stepping, previous chipset stepping should be
+	 * ignoring this setting.
+	 */
+	val = intel_de_read(dev_priv, PCH_DREF_CONTROL);
+
+	/* As we must carefully and slowly disable/enable each source in turn,
+	 * compute the final state we want first and check if we need to
+	 * make any changes at all.
+	 */
+	final = val;
+	final &= ~DREF_NONSPREAD_SOURCE_MASK;
+	if (has_ck505)
+		final |= DREF_NONSPREAD_CK505_ENABLE;
+	else
+		final |= DREF_NONSPREAD_SOURCE_ENABLE;
+
+	final &= ~DREF_SSC_SOURCE_MASK;
+	final &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
+	final &= ~DREF_SSC1_ENABLE;
+
+	if (has_panel) {
+		final |= DREF_SSC_SOURCE_ENABLE;
+
+		if (intel_panel_use_ssc(dev_priv) && can_ssc)
+			final |= DREF_SSC1_ENABLE;
+
+		if (has_cpu_edp) {
+			if (intel_panel_use_ssc(dev_priv) && can_ssc)
+				final |= DREF_CPU_SOURCE_OUTPUT_DOWNSPREAD;
+			else
+				final |= DREF_CPU_SOURCE_OUTPUT_NONSPREAD;
+		} else {
+			final |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
+		}
+	} else if (using_ssc_source) {
+		final |= DREF_SSC_SOURCE_ENABLE;
+		final |= DREF_SSC1_ENABLE;
+	}
+
+	if (final == val)
+		return;
+
+	/* Always enable nonspread source */
+	val &= ~DREF_NONSPREAD_SOURCE_MASK;
+
+	if (has_ck505)
+		val |= DREF_NONSPREAD_CK505_ENABLE;
+	else
+		val |= DREF_NONSPREAD_SOURCE_ENABLE;
+
+	if (has_panel) {
+		val &= ~DREF_SSC_SOURCE_MASK;
+		val |= DREF_SSC_SOURCE_ENABLE;
+
+		/* SSC must be turned on before enabling the CPU output  */
+		if (intel_panel_use_ssc(dev_priv) && can_ssc) {
+			drm_dbg_kms(&dev_priv->drm, "Using SSC on panel\n");
+			val |= DREF_SSC1_ENABLE;
+		} else {
+			val &= ~DREF_SSC1_ENABLE;
+		}
+
+		/* Get SSC going before enabling the outputs */
+		intel_de_write(dev_priv, PCH_DREF_CONTROL, val);
+		intel_de_posting_read(dev_priv, PCH_DREF_CONTROL);
+		udelay(200);
+
+		val &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
+
+		/* Enable CPU source on CPU attached eDP */
+		if (has_cpu_edp) {
+			if (intel_panel_use_ssc(dev_priv) && can_ssc) {
+				drm_dbg_kms(&dev_priv->drm,
+					    "Using SSC on eDP\n");
+				val |= DREF_CPU_SOURCE_OUTPUT_DOWNSPREAD;
+			} else {
+				val |= DREF_CPU_SOURCE_OUTPUT_NONSPREAD;
+			}
+		} else {
+			val |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
+		}
+
+		intel_de_write(dev_priv, PCH_DREF_CONTROL, val);
+		intel_de_posting_read(dev_priv, PCH_DREF_CONTROL);
+		udelay(200);
+	} else {
+		drm_dbg_kms(&dev_priv->drm, "Disabling CPU source output\n");
+
+		val &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
+
+		/* Turn off CPU output */
+		val |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
+
+		intel_de_write(dev_priv, PCH_DREF_CONTROL, val);
+		intel_de_posting_read(dev_priv, PCH_DREF_CONTROL);
+		udelay(200);
+
+		if (!using_ssc_source) {
+			drm_dbg_kms(&dev_priv->drm, "Disabling SSC source\n");
+
+			/* Turn off the SSC source */
+			val &= ~DREF_SSC_SOURCE_MASK;
+			val |= DREF_SSC_SOURCE_DISABLE;
+
+			/* Turn off SSC1 */
+			val &= ~DREF_SSC1_ENABLE;
+
+			intel_de_write(dev_priv, PCH_DREF_CONTROL, val);
+			intel_de_posting_read(dev_priv, PCH_DREF_CONTROL);
+			udelay(200);
+		}
+	}
+
+	drm_WARN_ON(&dev_priv->drm, val != final);
+}
+
+/*
+ * Initialize reference clocks when the driver loads
+ */
+void intel_init_pch_refclk(struct drm_i915_private *dev_priv)
+{
+	if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv))
+		ilk_init_pch_refclk(dev_priv);
+	else if (HAS_PCH_LPT(dev_priv))
+		lpt_init_pch_refclk(dev_priv);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_pch_refclk.h b/drivers/gpu/drm/i915/display/intel_pch_refclk.h
new file mode 100644
index 000000000..9bcf56629
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_pch_refclk.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _INTEL_PCH_REFCLK_H_
+#define _INTEL_PCH_REFCLK_H_
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct intel_crtc_state;
+
+void lpt_program_iclkip(const struct intel_crtc_state *crtc_state);
+void lpt_disable_iclkip(struct drm_i915_private *dev_priv);
+int lpt_get_iclkip(struct drm_i915_private *dev_priv);
+int lpt_iclkip(const struct intel_crtc_state *crtc_state);
+
+void intel_init_pch_refclk(struct drm_i915_private *dev_priv);
+void lpt_disable_clkout_dp(struct drm_i915_private *dev_priv);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_pipe_crc.c b/drivers/gpu/drm/i915/display/intel_pipe_crc.c
new file mode 100644
index 000000000..8ac263f47
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_pipe_crc.c
@@ -0,0 +1,672 @@
+/*
+ * Copyright © 2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Author: Damien Lespiau <damien.lespiau@intel.com>
+ *
+ */
+
+#include <linux/circ_buf.h>
+#include <linux/ctype.h>
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+
+#include "intel_atomic.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_pipe_crc.h"
+
+static const char * const pipe_crc_sources[] = {
+	[INTEL_PIPE_CRC_SOURCE_NONE] = "none",
+	[INTEL_PIPE_CRC_SOURCE_PLANE1] = "plane1",
+	[INTEL_PIPE_CRC_SOURCE_PLANE2] = "plane2",
+	[INTEL_PIPE_CRC_SOURCE_PLANE3] = "plane3",
+	[INTEL_PIPE_CRC_SOURCE_PLANE4] = "plane4",
+	[INTEL_PIPE_CRC_SOURCE_PLANE5] = "plane5",
+	[INTEL_PIPE_CRC_SOURCE_PLANE6] = "plane6",
+	[INTEL_PIPE_CRC_SOURCE_PLANE7] = "plane7",
+	[INTEL_PIPE_CRC_SOURCE_PIPE] = "pipe",
+	[INTEL_PIPE_CRC_SOURCE_TV] = "TV",
+	[INTEL_PIPE_CRC_SOURCE_DP_B] = "DP-B",
+	[INTEL_PIPE_CRC_SOURCE_DP_C] = "DP-C",
+	[INTEL_PIPE_CRC_SOURCE_DP_D] = "DP-D",
+	[INTEL_PIPE_CRC_SOURCE_AUTO] = "auto",
+};
+
+static int i8xx_pipe_crc_ctl_reg(enum intel_pipe_crc_source *source,
+				 u32 *val)
+{
+	if (*source == INTEL_PIPE_CRC_SOURCE_AUTO)
+		*source = INTEL_PIPE_CRC_SOURCE_PIPE;
+
+	switch (*source) {
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_INCLUDE_BORDER_I8XX;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		*val = 0;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int i9xx_pipe_crc_auto_source(struct drm_i915_private *dev_priv,
+				     enum pipe pipe,
+				     enum intel_pipe_crc_source *source)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_encoder *encoder;
+	struct intel_crtc *crtc;
+	struct intel_digital_port *dig_port;
+	int ret = 0;
+
+	*source = INTEL_PIPE_CRC_SOURCE_PIPE;
+
+	drm_modeset_lock_all(dev);
+	for_each_intel_encoder(dev, encoder) {
+		if (!encoder->base.crtc)
+			continue;
+
+		crtc = to_intel_crtc(encoder->base.crtc);
+
+		if (crtc->pipe != pipe)
+			continue;
+
+		switch (encoder->type) {
+		case INTEL_OUTPUT_TVOUT:
+			*source = INTEL_PIPE_CRC_SOURCE_TV;
+			break;
+		case INTEL_OUTPUT_DP:
+		case INTEL_OUTPUT_EDP:
+			dig_port = enc_to_dig_port(encoder);
+			switch (dig_port->base.port) {
+			case PORT_B:
+				*source = INTEL_PIPE_CRC_SOURCE_DP_B;
+				break;
+			case PORT_C:
+				*source = INTEL_PIPE_CRC_SOURCE_DP_C;
+				break;
+			case PORT_D:
+				*source = INTEL_PIPE_CRC_SOURCE_DP_D;
+				break;
+			default:
+				drm_WARN(dev, 1, "nonexisting DP port %c\n",
+					 port_name(dig_port->base.port));
+				break;
+			}
+			break;
+		default:
+			break;
+		}
+	}
+	drm_modeset_unlock_all(dev);
+
+	return ret;
+}
+
+static int vlv_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv,
+				enum pipe pipe,
+				enum intel_pipe_crc_source *source,
+				u32 *val)
+{
+	bool need_stable_symbols = false;
+
+	if (*source == INTEL_PIPE_CRC_SOURCE_AUTO) {
+		int ret = i9xx_pipe_crc_auto_source(dev_priv, pipe, source);
+		if (ret)
+			return ret;
+	}
+
+	switch (*source) {
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PIPE_VLV;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_DP_B:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_DP_B_VLV;
+		need_stable_symbols = true;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_DP_C:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_DP_C_VLV;
+		need_stable_symbols = true;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_DP_D:
+		if (!IS_CHERRYVIEW(dev_priv))
+			return -EINVAL;
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_DP_D_VLV;
+		need_stable_symbols = true;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		*val = 0;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	/*
+	 * When the pipe CRC tap point is after the transcoders we need
+	 * to tweak symbol-level features to produce a deterministic series of
+	 * symbols for a given frame. We need to reset those features only once
+	 * a frame (instead of every nth symbol):
+	 *   - DC-balance: used to ensure a better clock recovery from the data
+	 *     link (SDVO)
+	 *   - DisplayPort scrambling: used for EMI reduction
+	 */
+	if (need_stable_symbols) {
+		u32 tmp = intel_de_read(dev_priv, PORT_DFT2_G4X);
+
+		tmp |= DC_BALANCE_RESET_VLV;
+		switch (pipe) {
+		case PIPE_A:
+			tmp |= PIPE_A_SCRAMBLE_RESET;
+			break;
+		case PIPE_B:
+			tmp |= PIPE_B_SCRAMBLE_RESET;
+			break;
+		case PIPE_C:
+			tmp |= PIPE_C_SCRAMBLE_RESET;
+			break;
+		default:
+			return -EINVAL;
+		}
+		intel_de_write(dev_priv, PORT_DFT2_G4X, tmp);
+	}
+
+	return 0;
+}
+
+static int i9xx_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv,
+				 enum pipe pipe,
+				 enum intel_pipe_crc_source *source,
+				 u32 *val)
+{
+	if (*source == INTEL_PIPE_CRC_SOURCE_AUTO) {
+		int ret = i9xx_pipe_crc_auto_source(dev_priv, pipe, source);
+		if (ret)
+			return ret;
+	}
+
+	switch (*source) {
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PIPE_I9XX;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_TV:
+		if (!SUPPORTS_TV(dev_priv))
+			return -EINVAL;
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_TV_PRE;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		*val = 0;
+		break;
+	default:
+		/*
+		 * The DP CRC source doesn't work on g4x.
+		 * It can be made to work to some degree by selecting
+		 * the correct CRC source before the port is enabled,
+		 * and not touching the CRC source bits again until
+		 * the port is disabled. But even then the bits
+		 * eventually get stuck and a reboot is needed to get
+		 * working CRCs on the pipe again. Let's simply
+		 * refuse to use DP CRCs on g4x.
+		 */
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void vlv_undo_pipe_scramble_reset(struct drm_i915_private *dev_priv,
+					 enum pipe pipe)
+{
+	u32 tmp = intel_de_read(dev_priv, PORT_DFT2_G4X);
+
+	switch (pipe) {
+	case PIPE_A:
+		tmp &= ~PIPE_A_SCRAMBLE_RESET;
+		break;
+	case PIPE_B:
+		tmp &= ~PIPE_B_SCRAMBLE_RESET;
+		break;
+	case PIPE_C:
+		tmp &= ~PIPE_C_SCRAMBLE_RESET;
+		break;
+	default:
+		return;
+	}
+	if (!(tmp & PIPE_SCRAMBLE_RESET_MASK))
+		tmp &= ~DC_BALANCE_RESET_VLV;
+	intel_de_write(dev_priv, PORT_DFT2_G4X, tmp);
+}
+
+static int ilk_pipe_crc_ctl_reg(enum intel_pipe_crc_source *source,
+				u32 *val)
+{
+	if (*source == INTEL_PIPE_CRC_SOURCE_AUTO)
+		*source = INTEL_PIPE_CRC_SOURCE_PIPE;
+
+	switch (*source) {
+	case INTEL_PIPE_CRC_SOURCE_PLANE1:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PRIMARY_ILK;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PLANE2:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_SPRITE_ILK;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PIPE_ILK;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		*val = 0;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void
+intel_crtc_crc_setup_workarounds(struct intel_crtc *crtc, bool enable)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *pipe_config;
+	struct drm_atomic_state *state;
+	struct drm_modeset_acquire_ctx ctx;
+	int ret;
+
+	drm_modeset_acquire_init(&ctx, 0);
+
+	state = drm_atomic_state_alloc(&dev_priv->drm);
+	if (!state) {
+		ret = -ENOMEM;
+		goto unlock;
+	}
+
+	state->acquire_ctx = &ctx;
+
+retry:
+	pipe_config = intel_atomic_get_crtc_state(state, crtc);
+	if (IS_ERR(pipe_config)) {
+		ret = PTR_ERR(pipe_config);
+		goto put_state;
+	}
+
+	pipe_config->uapi.mode_changed = pipe_config->has_psr;
+	pipe_config->crc_enabled = enable;
+
+	if (IS_HASWELL(dev_priv) &&
+	    pipe_config->hw.active && crtc->pipe == PIPE_A &&
+	    pipe_config->cpu_transcoder == TRANSCODER_EDP)
+		pipe_config->uapi.mode_changed = true;
+
+	ret = drm_atomic_commit(state);
+
+put_state:
+	if (ret == -EDEADLK) {
+		drm_atomic_state_clear(state);
+		drm_modeset_backoff(&ctx);
+		goto retry;
+	}
+
+	drm_atomic_state_put(state);
+unlock:
+	drm_WARN(&dev_priv->drm, ret,
+		 "Toggling workaround to %i returns %i\n", enable, ret);
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+}
+
+static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv,
+				enum pipe pipe,
+				enum intel_pipe_crc_source *source,
+				u32 *val)
+{
+	if (*source == INTEL_PIPE_CRC_SOURCE_AUTO)
+		*source = INTEL_PIPE_CRC_SOURCE_PIPE;
+
+	switch (*source) {
+	case INTEL_PIPE_CRC_SOURCE_PLANE1:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PRIMARY_IVB;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PLANE2:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_SPRITE_IVB;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PF_IVB;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		*val = 0;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int skl_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv,
+				enum pipe pipe,
+				enum intel_pipe_crc_source *source,
+				u32 *val)
+{
+	if (*source == INTEL_PIPE_CRC_SOURCE_AUTO)
+		*source = INTEL_PIPE_CRC_SOURCE_PIPE;
+
+	switch (*source) {
+	case INTEL_PIPE_CRC_SOURCE_PLANE1:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PLANE_1_SKL;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PLANE2:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PLANE_2_SKL;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PLANE3:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PLANE_3_SKL;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PLANE4:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PLANE_4_SKL;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PLANE5:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PLANE_5_SKL;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PLANE6:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PLANE_6_SKL;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PLANE7:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_PLANE_7_SKL;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+		*val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_DMUX_SKL;
+		break;
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		*val = 0;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int get_new_crc_ctl_reg(struct drm_i915_private *dev_priv,
+			       enum pipe pipe,
+			       enum intel_pipe_crc_source *source, u32 *val)
+{
+	if (DISPLAY_VER(dev_priv) == 2)
+		return i8xx_pipe_crc_ctl_reg(source, val);
+	else if (DISPLAY_VER(dev_priv) < 5)
+		return i9xx_pipe_crc_ctl_reg(dev_priv, pipe, source, val);
+	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		return vlv_pipe_crc_ctl_reg(dev_priv, pipe, source, val);
+	else if (IS_IRONLAKE(dev_priv) || IS_SANDYBRIDGE(dev_priv))
+		return ilk_pipe_crc_ctl_reg(source, val);
+	else if (DISPLAY_VER(dev_priv) < 9)
+		return ivb_pipe_crc_ctl_reg(dev_priv, pipe, source, val);
+	else
+		return skl_pipe_crc_ctl_reg(dev_priv, pipe, source, val);
+}
+
+static int
+display_crc_ctl_parse_source(const char *buf, enum intel_pipe_crc_source *s)
+{
+	int i;
+
+	if (!buf) {
+		*s = INTEL_PIPE_CRC_SOURCE_NONE;
+		return 0;
+	}
+
+	i = match_string(pipe_crc_sources, ARRAY_SIZE(pipe_crc_sources), buf);
+	if (i < 0)
+		return i;
+
+	*s = i;
+	return 0;
+}
+
+void intel_crtc_crc_init(struct intel_crtc *crtc)
+{
+	struct intel_pipe_crc *pipe_crc = &crtc->pipe_crc;
+
+	spin_lock_init(&pipe_crc->lock);
+}
+
+static int i8xx_crc_source_valid(struct drm_i915_private *dev_priv,
+				 const enum intel_pipe_crc_source source)
+{
+	switch (source) {
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int i9xx_crc_source_valid(struct drm_i915_private *dev_priv,
+				 const enum intel_pipe_crc_source source)
+{
+	switch (source) {
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+	case INTEL_PIPE_CRC_SOURCE_TV:
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int vlv_crc_source_valid(struct drm_i915_private *dev_priv,
+				const enum intel_pipe_crc_source source)
+{
+	switch (source) {
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+	case INTEL_PIPE_CRC_SOURCE_DP_B:
+	case INTEL_PIPE_CRC_SOURCE_DP_C:
+	case INTEL_PIPE_CRC_SOURCE_DP_D:
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int ilk_crc_source_valid(struct drm_i915_private *dev_priv,
+				const enum intel_pipe_crc_source source)
+{
+	switch (source) {
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+	case INTEL_PIPE_CRC_SOURCE_PLANE1:
+	case INTEL_PIPE_CRC_SOURCE_PLANE2:
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int ivb_crc_source_valid(struct drm_i915_private *dev_priv,
+				const enum intel_pipe_crc_source source)
+{
+	switch (source) {
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+	case INTEL_PIPE_CRC_SOURCE_PLANE1:
+	case INTEL_PIPE_CRC_SOURCE_PLANE2:
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int skl_crc_source_valid(struct drm_i915_private *dev_priv,
+				const enum intel_pipe_crc_source source)
+{
+	switch (source) {
+	case INTEL_PIPE_CRC_SOURCE_PIPE:
+	case INTEL_PIPE_CRC_SOURCE_PLANE1:
+	case INTEL_PIPE_CRC_SOURCE_PLANE2:
+	case INTEL_PIPE_CRC_SOURCE_PLANE3:
+	case INTEL_PIPE_CRC_SOURCE_PLANE4:
+	case INTEL_PIPE_CRC_SOURCE_PLANE5:
+	case INTEL_PIPE_CRC_SOURCE_PLANE6:
+	case INTEL_PIPE_CRC_SOURCE_PLANE7:
+	case INTEL_PIPE_CRC_SOURCE_NONE:
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int
+intel_is_valid_crc_source(struct drm_i915_private *dev_priv,
+			  const enum intel_pipe_crc_source source)
+{
+	if (DISPLAY_VER(dev_priv) == 2)
+		return i8xx_crc_source_valid(dev_priv, source);
+	else if (DISPLAY_VER(dev_priv) < 5)
+		return i9xx_crc_source_valid(dev_priv, source);
+	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		return vlv_crc_source_valid(dev_priv, source);
+	else if (IS_IRONLAKE(dev_priv) || IS_SANDYBRIDGE(dev_priv))
+		return ilk_crc_source_valid(dev_priv, source);
+	else if (DISPLAY_VER(dev_priv) < 9)
+		return ivb_crc_source_valid(dev_priv, source);
+	else
+		return skl_crc_source_valid(dev_priv, source);
+}
+
+const char *const *intel_crtc_get_crc_sources(struct drm_crtc *crtc,
+					      size_t *count)
+{
+	*count = ARRAY_SIZE(pipe_crc_sources);
+	return pipe_crc_sources;
+}
+
+int intel_crtc_verify_crc_source(struct drm_crtc *crtc, const char *source_name,
+				 size_t *values_cnt)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+	enum intel_pipe_crc_source source;
+
+	if (display_crc_ctl_parse_source(source_name, &source) < 0) {
+		drm_dbg(&dev_priv->drm, "unknown source %s\n", source_name);
+		return -EINVAL;
+	}
+
+	if (source == INTEL_PIPE_CRC_SOURCE_AUTO ||
+	    intel_is_valid_crc_source(dev_priv, source) == 0) {
+		*values_cnt = 5;
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
+int intel_crtc_set_crc_source(struct drm_crtc *_crtc, const char *source_name)
+{
+	struct intel_crtc *crtc = to_intel_crtc(_crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_pipe_crc *pipe_crc = &crtc->pipe_crc;
+	enum intel_display_power_domain power_domain;
+	enum intel_pipe_crc_source source;
+	enum pipe pipe = crtc->pipe;
+	intel_wakeref_t wakeref;
+	u32 val = 0; /* shut up gcc */
+	int ret = 0;
+	bool enable;
+
+	if (display_crc_ctl_parse_source(source_name, &source) < 0) {
+		drm_dbg(&dev_priv->drm, "unknown source %s\n", source_name);
+		return -EINVAL;
+	}
+
+	power_domain = POWER_DOMAIN_PIPE(pipe);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Trying to capture CRC while pipe is off\n");
+		return -EIO;
+	}
+
+	enable = source != INTEL_PIPE_CRC_SOURCE_NONE;
+	if (enable)
+		intel_crtc_crc_setup_workarounds(crtc, true);
+
+	ret = get_new_crc_ctl_reg(dev_priv, pipe, &source, &val);
+	if (ret != 0)
+		goto out;
+
+	pipe_crc->source = source;
+	intel_de_write(dev_priv, PIPE_CRC_CTL(pipe), val);
+	intel_de_posting_read(dev_priv, PIPE_CRC_CTL(pipe));
+
+	if (!source) {
+		if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+			vlv_undo_pipe_scramble_reset(dev_priv, pipe);
+	}
+
+	pipe_crc->skipped = 0;
+
+out:
+	if (!enable)
+		intel_crtc_crc_setup_workarounds(crtc, false);
+
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+
+	return ret;
+}
+
+void intel_crtc_enable_pipe_crc(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_pipe_crc *pipe_crc = &crtc->pipe_crc;
+	enum pipe pipe = crtc->pipe;
+	u32 val = 0;
+
+	if (!crtc->base.crc.opened)
+		return;
+
+	if (get_new_crc_ctl_reg(dev_priv, pipe, &pipe_crc->source, &val) < 0)
+		return;
+
+	/* Don't need pipe_crc->lock here, IRQs are not generated. */
+	pipe_crc->skipped = 0;
+
+	intel_de_write(dev_priv, PIPE_CRC_CTL(pipe), val);
+	intel_de_posting_read(dev_priv, PIPE_CRC_CTL(pipe));
+}
+
+void intel_crtc_disable_pipe_crc(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_pipe_crc *pipe_crc = &crtc->pipe_crc;
+	enum pipe pipe = crtc->pipe;
+
+	/* Swallow crc's until we stop generating them. */
+	spin_lock_irq(&pipe_crc->lock);
+	pipe_crc->skipped = INT_MIN;
+	spin_unlock_irq(&pipe_crc->lock);
+
+	intel_de_write(dev_priv, PIPE_CRC_CTL(pipe), 0);
+	intel_de_posting_read(dev_priv, PIPE_CRC_CTL(pipe));
+	intel_synchronize_irq(dev_priv);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_pipe_crc.h b/drivers/gpu/drm/i915/display/intel_pipe_crc.h
new file mode 100644
index 000000000..43012b189
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_pipe_crc.h
@@ -0,0 +1,38 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_PIPE_CRC_H__
+#define __INTEL_PIPE_CRC_H__
+
+#include <linux/types.h>
+
+struct drm_crtc;
+struct drm_i915_private;
+struct intel_crtc;
+
+#ifdef CONFIG_DEBUG_FS
+void intel_crtc_crc_init(struct intel_crtc *crtc);
+int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name);
+int intel_crtc_verify_crc_source(struct drm_crtc *crtc,
+				 const char *source_name, size_t *values_cnt);
+const char *const *intel_crtc_get_crc_sources(struct drm_crtc *crtc,
+					      size_t *count);
+void intel_crtc_disable_pipe_crc(struct intel_crtc *crtc);
+void intel_crtc_enable_pipe_crc(struct intel_crtc *crtc);
+#else
+static inline void intel_crtc_crc_init(struct intel_crtc *crtc) {}
+#define intel_crtc_set_crc_source NULL
+#define intel_crtc_verify_crc_source NULL
+#define intel_crtc_get_crc_sources NULL
+static inline void intel_crtc_disable_pipe_crc(struct intel_crtc *crtc)
+{
+}
+
+static inline void intel_crtc_enable_pipe_crc(struct intel_crtc *crtc)
+{
+}
+#endif
+
+#endif /* __INTEL_PIPE_CRC_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_plane_initial.c b/drivers/gpu/drm/i915/display/intel_plane_initial.c
new file mode 100644
index 000000000..76be796df
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_plane_initial.c
@@ -0,0 +1,323 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include "gem/i915_gem_region.h"
+#include "i915_drv.h"
+#include "intel_atomic_plane.h"
+#include "intel_display.h"
+#include "intel_display_types.h"
+#include "intel_fb.h"
+#include "intel_plane_initial.h"
+
+static bool
+intel_reuse_initial_plane_obj(struct drm_i915_private *i915,
+			      const struct intel_initial_plane_config *plane_config,
+			      struct drm_framebuffer **fb,
+			      struct i915_vma **vma)
+{
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct intel_plane *plane =
+			to_intel_plane(crtc->base.primary);
+		struct intel_plane_state *plane_state =
+			to_intel_plane_state(plane->base.state);
+
+		if (!crtc_state->uapi.active)
+			continue;
+
+		if (!plane_state->ggtt_vma)
+			continue;
+
+		if (intel_plane_ggtt_offset(plane_state) == plane_config->base) {
+			*fb = plane_state->hw.fb;
+			*vma = plane_state->ggtt_vma;
+			return true;
+		}
+	}
+
+	return false;
+}
+
+static struct i915_vma *
+initial_plane_vma(struct drm_i915_private *i915,
+		  struct intel_initial_plane_config *plane_config)
+{
+	struct intel_memory_region *mem;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	resource_size_t phys_base;
+	u32 base, size;
+	u64 pinctl;
+
+	if (plane_config->size == 0)
+		return NULL;
+
+	base = round_down(plane_config->base, I915_GTT_MIN_ALIGNMENT);
+	if (IS_DGFX(i915)) {
+		gen8_pte_t __iomem *gte = to_gt(i915)->ggtt->gsm;
+		gen8_pte_t pte;
+
+		gte += base / I915_GTT_PAGE_SIZE;
+
+		pte = ioread64(gte);
+		if (!(pte & GEN12_GGTT_PTE_LM)) {
+			drm_err(&i915->drm,
+				"Initial plane programming missing PTE_LM bit\n");
+			return NULL;
+		}
+
+		phys_base = pte & I915_GTT_PAGE_MASK;
+		mem = i915->mm.regions[INTEL_REGION_LMEM_0];
+
+		/*
+		 * We don't currently expect this to ever be placed in the
+		 * stolen portion.
+		 */
+		if (phys_base >= resource_size(&mem->region)) {
+			drm_err(&i915->drm,
+				"Initial plane programming using invalid range, phys_base=%pa\n",
+				&phys_base);
+			return NULL;
+		}
+
+		drm_dbg(&i915->drm,
+			"Using phys_base=%pa, based on initial plane programming\n",
+			&phys_base);
+	} else {
+		phys_base = base;
+		mem = i915->mm.stolen_region;
+	}
+
+	if (!mem)
+		return NULL;
+
+	size = round_up(plane_config->base + plane_config->size,
+			mem->min_page_size);
+	size -= base;
+
+	/*
+	 * If the FB is too big, just don't use it since fbdev is not very
+	 * important and we should probably use that space with FBC or other
+	 * features.
+	 */
+	if (IS_ENABLED(CONFIG_FRAMEBUFFER_CONSOLE) &&
+	    mem == i915->mm.stolen_region &&
+	    size * 2 > i915->stolen_usable_size)
+		return NULL;
+
+	obj = i915_gem_object_create_region_at(mem, phys_base, size, 0);
+	if (IS_ERR(obj))
+		return NULL;
+
+	/*
+	 * Mark it WT ahead of time to avoid changing the
+	 * cache_level during fbdev initialization. The
+	 * unbind there would get stuck waiting for rcu.
+	 */
+	i915_gem_object_set_cache_coherency(obj, HAS_WT(i915) ?
+					    I915_CACHE_WT : I915_CACHE_NONE);
+
+	switch (plane_config->tiling) {
+	case I915_TILING_NONE:
+		break;
+	case I915_TILING_X:
+	case I915_TILING_Y:
+		obj->tiling_and_stride =
+			plane_config->fb->base.pitches[0] |
+			plane_config->tiling;
+		break;
+	default:
+		MISSING_CASE(plane_config->tiling);
+		goto err_obj;
+	}
+
+	vma = i915_vma_instance(obj, &to_gt(i915)->ggtt->vm, NULL);
+	if (IS_ERR(vma))
+		goto err_obj;
+
+	pinctl = PIN_GLOBAL | PIN_OFFSET_FIXED | base;
+	if (HAS_GMCH(i915))
+		pinctl |= PIN_MAPPABLE;
+	if (i915_vma_pin(vma, 0, 0, pinctl))
+		goto err_obj;
+
+	if (i915_gem_object_is_tiled(obj) &&
+	    !i915_vma_is_map_and_fenceable(vma))
+		goto err_obj;
+
+	return vma;
+
+err_obj:
+	i915_gem_object_put(obj);
+	return NULL;
+}
+
+static bool
+intel_alloc_initial_plane_obj(struct intel_crtc *crtc,
+			      struct intel_initial_plane_config *plane_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_mode_fb_cmd2 mode_cmd = { 0 };
+	struct drm_framebuffer *fb = &plane_config->fb->base;
+	struct i915_vma *vma;
+
+	switch (fb->modifier) {
+	case DRM_FORMAT_MOD_LINEAR:
+	case I915_FORMAT_MOD_X_TILED:
+	case I915_FORMAT_MOD_Y_TILED:
+	case I915_FORMAT_MOD_4_TILED:
+		break;
+	default:
+		drm_dbg(&dev_priv->drm,
+			"Unsupported modifier for initial FB: 0x%llx\n",
+			fb->modifier);
+		return false;
+	}
+
+	vma = initial_plane_vma(dev_priv, plane_config);
+	if (!vma)
+		return false;
+
+	mode_cmd.pixel_format = fb->format->format;
+	mode_cmd.width = fb->width;
+	mode_cmd.height = fb->height;
+	mode_cmd.pitches[0] = fb->pitches[0];
+	mode_cmd.modifier[0] = fb->modifier;
+	mode_cmd.flags = DRM_MODE_FB_MODIFIERS;
+
+	if (intel_framebuffer_init(to_intel_framebuffer(fb),
+				   vma->obj, &mode_cmd)) {
+		drm_dbg_kms(&dev_priv->drm, "intel fb init failed\n");
+		goto err_vma;
+	}
+
+	plane_config->vma = vma;
+	return true;
+
+err_vma:
+	i915_vma_put(vma);
+	return false;
+}
+
+static void
+intel_find_initial_plane_obj(struct intel_crtc *crtc,
+			     struct intel_initial_plane_config *plane_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_plane *plane =
+		to_intel_plane(crtc->base.primary);
+	struct intel_plane_state *plane_state =
+		to_intel_plane_state(plane->base.state);
+	struct drm_framebuffer *fb;
+	struct i915_vma *vma;
+
+	/*
+	 * TODO:
+	 *   Disable planes if get_initial_plane_config() failed.
+	 *   Make sure things work if the surface base is not page aligned.
+	 */
+	if (!plane_config->fb)
+		return;
+
+	if (intel_alloc_initial_plane_obj(crtc, plane_config)) {
+		fb = &plane_config->fb->base;
+		vma = plane_config->vma;
+		goto valid_fb;
+	}
+
+	/*
+	 * Failed to alloc the obj, check to see if we should share
+	 * an fb with another CRTC instead
+	 */
+	if (intel_reuse_initial_plane_obj(dev_priv, plane_config, &fb, &vma))
+		goto valid_fb;
+
+	/*
+	 * We've failed to reconstruct the BIOS FB.  Current display state
+	 * indicates that the primary plane is visible, but has a NULL FB,
+	 * which will lead to problems later if we don't fix it up.  The
+	 * simplest solution is to just disable the primary plane now and
+	 * pretend the BIOS never had it enabled.
+	 */
+	intel_plane_disable_noatomic(crtc, plane);
+
+	return;
+
+valid_fb:
+	plane_state->uapi.rotation = plane_config->rotation;
+	intel_fb_fill_view(to_intel_framebuffer(fb),
+			   plane_state->uapi.rotation, &plane_state->view);
+
+	__i915_vma_pin(vma);
+	plane_state->ggtt_vma = i915_vma_get(vma);
+	if (intel_plane_uses_fence(plane_state) &&
+	    i915_vma_pin_fence(vma) == 0 && vma->fence)
+		plane_state->flags |= PLANE_HAS_FENCE;
+
+	plane_state->uapi.src_x = 0;
+	plane_state->uapi.src_y = 0;
+	plane_state->uapi.src_w = fb->width << 16;
+	plane_state->uapi.src_h = fb->height << 16;
+
+	plane_state->uapi.crtc_x = 0;
+	plane_state->uapi.crtc_y = 0;
+	plane_state->uapi.crtc_w = fb->width;
+	plane_state->uapi.crtc_h = fb->height;
+
+	if (plane_config->tiling)
+		dev_priv->preserve_bios_swizzle = true;
+
+	plane_state->uapi.fb = fb;
+	drm_framebuffer_get(fb);
+
+	plane_state->uapi.crtc = &crtc->base;
+	intel_plane_copy_uapi_to_hw_state(plane_state, plane_state, crtc);
+
+	atomic_or(plane->frontbuffer_bit, &to_intel_frontbuffer(fb)->bits);
+}
+
+static void plane_config_fini(struct intel_initial_plane_config *plane_config)
+{
+	if (plane_config->fb) {
+		struct drm_framebuffer *fb = &plane_config->fb->base;
+
+		/* We may only have the stub and not a full framebuffer */
+		if (drm_framebuffer_read_refcount(fb))
+			drm_framebuffer_put(fb);
+		else
+			kfree(fb);
+	}
+
+	if (plane_config->vma)
+		i915_vma_put(plane_config->vma);
+}
+
+void intel_crtc_initial_plane_config(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_initial_plane_config plane_config = {};
+
+	/*
+	 * Note that reserving the BIOS fb up front prevents us
+	 * from stuffing other stolen allocations like the ring
+	 * on top.  This prevents some ugliness at boot time, and
+	 * can even allow for smooth boot transitions if the BIOS
+	 * fb is large enough for the active pipe configuration.
+	 */
+	dev_priv->display.funcs.display->get_initial_plane_config(crtc, &plane_config);
+
+	/*
+	 * If the fb is shared between multiple heads, we'll
+	 * just get the first one.
+	 */
+	intel_find_initial_plane_obj(crtc, &plane_config);
+
+	plane_config_fini(&plane_config);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_plane_initial.h b/drivers/gpu/drm/i915/display/intel_plane_initial.h
new file mode 100644
index 000000000..c7e35ab31
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_plane_initial.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __INTEL_PLANE_INITIAL_H__
+#define __INTEL_PLANE_INITIAL_H__
+
+struct intel_crtc;
+
+void intel_crtc_initial_plane_config(struct intel_crtc *crtc);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_pps.c b/drivers/gpu/drm/i915/display/intel_pps.c
new file mode 100644
index 000000000..21944f5bf
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_pps.c
@@ -0,0 +1,1554 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "g4x_dp.h"
+#include "i915_drv.h"
+#include "intel_de.h"
+#include "intel_display_power_well.h"
+#include "intel_display_types.h"
+#include "intel_dp.h"
+#include "intel_dpll.h"
+#include "intel_lvds.h"
+#include "intel_pps.h"
+#include "intel_quirks.h"
+
+static void vlv_steal_power_sequencer(struct drm_i915_private *dev_priv,
+				      enum pipe pipe);
+
+static void pps_init_delays(struct intel_dp *intel_dp);
+static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd);
+
+intel_wakeref_t intel_pps_lock(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	intel_wakeref_t wakeref;
+
+	/*
+	 * See intel_pps_reset_all() why we need a power domain reference here.
+	 */
+	wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_DISPLAY_CORE);
+	mutex_lock(&dev_priv->display.pps.mutex);
+
+	return wakeref;
+}
+
+intel_wakeref_t intel_pps_unlock(struct intel_dp *intel_dp,
+				 intel_wakeref_t wakeref)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	mutex_unlock(&dev_priv->display.pps.mutex);
+	intel_display_power_put(dev_priv, POWER_DOMAIN_DISPLAY_CORE, wakeref);
+
+	return 0;
+}
+
+static void
+vlv_power_sequencer_kick(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum pipe pipe = intel_dp->pps.pps_pipe;
+	bool pll_enabled, release_cl_override = false;
+	enum dpio_phy phy = DPIO_PHY(pipe);
+	enum dpio_channel ch = vlv_pipe_to_channel(pipe);
+	u32 DP;
+
+	if (drm_WARN(&dev_priv->drm,
+		     intel_de_read(dev_priv, intel_dp->output_reg) & DP_PORT_EN,
+		     "skipping pipe %c power sequencer kick due to [ENCODER:%d:%s] being active\n",
+		     pipe_name(pipe), dig_port->base.base.base.id,
+		     dig_port->base.base.name))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "kicking pipe %c power sequencer for [ENCODER:%d:%s]\n",
+		    pipe_name(pipe), dig_port->base.base.base.id,
+		    dig_port->base.base.name);
+
+	/* Preserve the BIOS-computed detected bit. This is
+	 * supposed to be read-only.
+	 */
+	DP = intel_de_read(dev_priv, intel_dp->output_reg) & DP_DETECTED;
+	DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
+	DP |= DP_PORT_WIDTH(1);
+	DP |= DP_LINK_TRAIN_PAT_1;
+
+	if (IS_CHERRYVIEW(dev_priv))
+		DP |= DP_PIPE_SEL_CHV(pipe);
+	else
+		DP |= DP_PIPE_SEL(pipe);
+
+	pll_enabled = intel_de_read(dev_priv, DPLL(pipe)) & DPLL_VCO_ENABLE;
+
+	/*
+	 * The DPLL for the pipe must be enabled for this to work.
+	 * So enable temporarily it if it's not already enabled.
+	 */
+	if (!pll_enabled) {
+		release_cl_override = IS_CHERRYVIEW(dev_priv) &&
+			!chv_phy_powergate_ch(dev_priv, phy, ch, true);
+
+		if (vlv_force_pll_on(dev_priv, pipe, vlv_get_dpll(dev_priv))) {
+			drm_err(&dev_priv->drm,
+				"Failed to force on pll for pipe %c!\n",
+				pipe_name(pipe));
+			return;
+		}
+	}
+
+	/*
+	 * Similar magic as in intel_dp_enable_port().
+	 * We _must_ do this port enable + disable trick
+	 * to make this power sequencer lock onto the port.
+	 * Otherwise even VDD force bit won't work.
+	 */
+	intel_de_write(dev_priv, intel_dp->output_reg, DP);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+
+	intel_de_write(dev_priv, intel_dp->output_reg, DP | DP_PORT_EN);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+
+	intel_de_write(dev_priv, intel_dp->output_reg, DP & ~DP_PORT_EN);
+	intel_de_posting_read(dev_priv, intel_dp->output_reg);
+
+	if (!pll_enabled) {
+		vlv_force_pll_off(dev_priv, pipe);
+
+		if (release_cl_override)
+			chv_phy_powergate_ch(dev_priv, phy, ch, false);
+	}
+}
+
+static enum pipe vlv_find_free_pps(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+	unsigned int pipes = (1 << PIPE_A) | (1 << PIPE_B);
+
+	/*
+	 * We don't have power sequencer currently.
+	 * Pick one that's not used by other ports.
+	 */
+	for_each_intel_dp(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		if (encoder->type == INTEL_OUTPUT_EDP) {
+			drm_WARN_ON(&dev_priv->drm,
+				    intel_dp->pps.active_pipe != INVALID_PIPE &&
+				    intel_dp->pps.active_pipe !=
+				    intel_dp->pps.pps_pipe);
+
+			if (intel_dp->pps.pps_pipe != INVALID_PIPE)
+				pipes &= ~(1 << intel_dp->pps.pps_pipe);
+		} else {
+			drm_WARN_ON(&dev_priv->drm,
+				    intel_dp->pps.pps_pipe != INVALID_PIPE);
+
+			if (intel_dp->pps.active_pipe != INVALID_PIPE)
+				pipes &= ~(1 << intel_dp->pps.active_pipe);
+		}
+	}
+
+	if (pipes == 0)
+		return INVALID_PIPE;
+
+	return ffs(pipes) - 1;
+}
+
+static enum pipe
+vlv_power_sequencer_pipe(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum pipe pipe;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	/* We should never land here with regular DP ports */
+	drm_WARN_ON(&dev_priv->drm, !intel_dp_is_edp(intel_dp));
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.active_pipe != INVALID_PIPE &&
+		    intel_dp->pps.active_pipe != intel_dp->pps.pps_pipe);
+
+	if (intel_dp->pps.pps_pipe != INVALID_PIPE)
+		return intel_dp->pps.pps_pipe;
+
+	pipe = vlv_find_free_pps(dev_priv);
+
+	/*
+	 * Didn't find one. This should not happen since there
+	 * are two power sequencers and up to two eDP ports.
+	 */
+	if (drm_WARN_ON(&dev_priv->drm, pipe == INVALID_PIPE))
+		pipe = PIPE_A;
+
+	vlv_steal_power_sequencer(dev_priv, pipe);
+	intel_dp->pps.pps_pipe = pipe;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "picked pipe %c power sequencer for [ENCODER:%d:%s]\n",
+		    pipe_name(intel_dp->pps.pps_pipe),
+		    dig_port->base.base.base.id,
+		    dig_port->base.base.name);
+
+	/* init power sequencer on this pipe and port */
+	pps_init_delays(intel_dp);
+	pps_init_registers(intel_dp, true);
+
+	/*
+	 * Even vdd force doesn't work until we've made
+	 * the power sequencer lock in on the port.
+	 */
+	vlv_power_sequencer_kick(intel_dp);
+
+	return intel_dp->pps.pps_pipe;
+}
+
+static int
+bxt_power_sequencer_idx(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_connector *connector = intel_dp->attached_connector;
+	int backlight_controller = connector->panel.vbt.backlight.controller;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	/* We should never land here with regular DP ports */
+	drm_WARN_ON(&dev_priv->drm, !intel_dp_is_edp(intel_dp));
+
+	if (!intel_dp->pps.pps_reset)
+		return backlight_controller;
+
+	intel_dp->pps.pps_reset = false;
+
+	/*
+	 * Only the HW needs to be reprogrammed, the SW state is fixed and
+	 * has been setup during connector init.
+	 */
+	pps_init_registers(intel_dp, false);
+
+	return backlight_controller;
+}
+
+typedef bool (*vlv_pipe_check)(struct drm_i915_private *dev_priv,
+			       enum pipe pipe);
+
+static bool vlv_pipe_has_pp_on(struct drm_i915_private *dev_priv,
+			       enum pipe pipe)
+{
+	return intel_de_read(dev_priv, PP_STATUS(pipe)) & PP_ON;
+}
+
+static bool vlv_pipe_has_vdd_on(struct drm_i915_private *dev_priv,
+				enum pipe pipe)
+{
+	return intel_de_read(dev_priv, PP_CONTROL(pipe)) & EDP_FORCE_VDD;
+}
+
+static bool vlv_pipe_any(struct drm_i915_private *dev_priv,
+			 enum pipe pipe)
+{
+	return true;
+}
+
+static enum pipe
+vlv_initial_pps_pipe(struct drm_i915_private *dev_priv,
+		     enum port port,
+		     vlv_pipe_check pipe_check)
+{
+	enum pipe pipe;
+
+	for (pipe = PIPE_A; pipe <= PIPE_B; pipe++) {
+		u32 port_sel = intel_de_read(dev_priv, PP_ON_DELAYS(pipe)) &
+			PANEL_PORT_SELECT_MASK;
+
+		if (port_sel != PANEL_PORT_SELECT_VLV(port))
+			continue;
+
+		if (!pipe_check(dev_priv, pipe))
+			continue;
+
+		return pipe;
+	}
+
+	return INVALID_PIPE;
+}
+
+static void
+vlv_initial_power_sequencer_setup(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum port port = dig_port->base.port;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	/* try to find a pipe with this port selected */
+	/* first pick one where the panel is on */
+	intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+						      vlv_pipe_has_pp_on);
+	/* didn't find one? pick one where vdd is on */
+	if (intel_dp->pps.pps_pipe == INVALID_PIPE)
+		intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+							      vlv_pipe_has_vdd_on);
+	/* didn't find one? pick one with just the correct port */
+	if (intel_dp->pps.pps_pipe == INVALID_PIPE)
+		intel_dp->pps.pps_pipe = vlv_initial_pps_pipe(dev_priv, port,
+							      vlv_pipe_any);
+
+	/* didn't find one? just let vlv_power_sequencer_pipe() pick one when needed */
+	if (intel_dp->pps.pps_pipe == INVALID_PIPE) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "no initial power sequencer for [ENCODER:%d:%s]\n",
+			    dig_port->base.base.base.id,
+			    dig_port->base.base.name);
+		return;
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "initial power sequencer for [ENCODER:%d:%s]: pipe %c\n",
+		    dig_port->base.base.base.id,
+		    dig_port->base.base.name,
+		    pipe_name(intel_dp->pps.pps_pipe));
+}
+
+void intel_pps_reset_all(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+
+	if (drm_WARN_ON(&dev_priv->drm, !IS_LP(dev_priv)))
+		return;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	/*
+	 * We can't grab pps_mutex here due to deadlock with power_domain
+	 * mutex when power_domain functions are called while holding pps_mutex.
+	 * That also means that in order to use pps_pipe the code needs to
+	 * hold both a power domain reference and pps_mutex, and the power domain
+	 * reference get/put must be done while _not_ holding pps_mutex.
+	 * pps_{lock,unlock}() do these steps in the correct order, so one
+	 * should use them always.
+	 */
+
+	for_each_intel_dp(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		drm_WARN_ON(&dev_priv->drm,
+			    intel_dp->pps.active_pipe != INVALID_PIPE);
+
+		if (encoder->type != INTEL_OUTPUT_EDP)
+			continue;
+
+		if (DISPLAY_VER(dev_priv) >= 9)
+			intel_dp->pps.pps_reset = true;
+		else
+			intel_dp->pps.pps_pipe = INVALID_PIPE;
+	}
+}
+
+struct pps_registers {
+	i915_reg_t pp_ctrl;
+	i915_reg_t pp_stat;
+	i915_reg_t pp_on;
+	i915_reg_t pp_off;
+	i915_reg_t pp_div;
+};
+
+static void intel_pps_get_registers(struct intel_dp *intel_dp,
+				    struct pps_registers *regs)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	int pps_idx = 0;
+
+	memset(regs, 0, sizeof(*regs));
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		pps_idx = bxt_power_sequencer_idx(intel_dp);
+	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		pps_idx = vlv_power_sequencer_pipe(intel_dp);
+
+	regs->pp_ctrl = PP_CONTROL(pps_idx);
+	regs->pp_stat = PP_STATUS(pps_idx);
+	regs->pp_on = PP_ON_DELAYS(pps_idx);
+	regs->pp_off = PP_OFF_DELAYS(pps_idx);
+
+	/* Cycle delay moved from PP_DIVISOR to PP_CONTROL */
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv) ||
+	    INTEL_PCH_TYPE(dev_priv) >= PCH_CNP)
+		regs->pp_div = INVALID_MMIO_REG;
+	else
+		regs->pp_div = PP_DIVISOR(pps_idx);
+}
+
+static i915_reg_t
+_pp_ctrl_reg(struct intel_dp *intel_dp)
+{
+	struct pps_registers regs;
+
+	intel_pps_get_registers(intel_dp, &regs);
+
+	return regs.pp_ctrl;
+}
+
+static i915_reg_t
+_pp_stat_reg(struct intel_dp *intel_dp)
+{
+	struct pps_registers regs;
+
+	intel_pps_get_registers(intel_dp, &regs);
+
+	return regs.pp_stat;
+}
+
+static bool edp_have_panel_power(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    intel_dp->pps.pps_pipe == INVALID_PIPE)
+		return false;
+
+	return (intel_de_read(dev_priv, _pp_stat_reg(intel_dp)) & PP_ON) != 0;
+}
+
+static bool edp_have_panel_vdd(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    intel_dp->pps.pps_pipe == INVALID_PIPE)
+		return false;
+
+	return intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp)) & EDP_FORCE_VDD;
+}
+
+void intel_pps_check_power_unlocked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	if (!edp_have_panel_power(intel_dp) && !edp_have_panel_vdd(intel_dp)) {
+		drm_WARN(&dev_priv->drm, 1,
+			 "eDP powered off while attempting aux channel communication.\n");
+		drm_dbg_kms(&dev_priv->drm, "Status 0x%08x Control 0x%08x\n",
+			    intel_de_read(dev_priv, _pp_stat_reg(intel_dp)),
+			    intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp)));
+	}
+}
+
+#define IDLE_ON_MASK		(PP_ON | PP_SEQUENCE_MASK | 0                     | PP_SEQUENCE_STATE_MASK)
+#define IDLE_ON_VALUE   	(PP_ON | PP_SEQUENCE_NONE | 0                     | PP_SEQUENCE_STATE_ON_IDLE)
+
+#define IDLE_OFF_MASK		(PP_ON | PP_SEQUENCE_MASK | 0                     | 0)
+#define IDLE_OFF_VALUE		(0     | PP_SEQUENCE_NONE | 0                     | 0)
+
+#define IDLE_CYCLE_MASK		(PP_ON | PP_SEQUENCE_MASK | PP_CYCLE_DELAY_ACTIVE | PP_SEQUENCE_STATE_MASK)
+#define IDLE_CYCLE_VALUE	(0     | PP_SEQUENCE_NONE | 0                     | PP_SEQUENCE_STATE_OFF_IDLE)
+
+static void intel_pps_verify_state(struct intel_dp *intel_dp);
+
+static void wait_panel_status(struct intel_dp *intel_dp,
+				       u32 mask,
+				       u32 value)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	i915_reg_t pp_stat_reg, pp_ctrl_reg;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	intel_pps_verify_state(intel_dp);
+
+	pp_stat_reg = _pp_stat_reg(intel_dp);
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "mask %08x value %08x status %08x control %08x\n",
+		    mask, value,
+		    intel_de_read(dev_priv, pp_stat_reg),
+		    intel_de_read(dev_priv, pp_ctrl_reg));
+
+	if (intel_de_wait_for_register(dev_priv, pp_stat_reg,
+				       mask, value, 5000))
+		drm_err(&dev_priv->drm,
+			"Panel status timeout: status %08x control %08x\n",
+			intel_de_read(dev_priv, pp_stat_reg),
+			intel_de_read(dev_priv, pp_ctrl_reg));
+
+	drm_dbg_kms(&dev_priv->drm, "Wait complete\n");
+}
+
+static void wait_panel_on(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	drm_dbg_kms(&i915->drm, "Wait for panel power on\n");
+	wait_panel_status(intel_dp, IDLE_ON_MASK, IDLE_ON_VALUE);
+}
+
+static void wait_panel_off(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	drm_dbg_kms(&i915->drm, "Wait for panel power off time\n");
+	wait_panel_status(intel_dp, IDLE_OFF_MASK, IDLE_OFF_VALUE);
+}
+
+static void wait_panel_power_cycle(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	ktime_t panel_power_on_time;
+	s64 panel_power_off_duration;
+
+	drm_dbg_kms(&i915->drm, "Wait for panel power cycle\n");
+
+	/* take the difference of current time and panel power off time
+	 * and then make panel wait for t11_t12 if needed. */
+	panel_power_on_time = ktime_get_boottime();
+	panel_power_off_duration = ktime_ms_delta(panel_power_on_time, intel_dp->pps.panel_power_off_time);
+
+	/* When we disable the VDD override bit last we have to do the manual
+	 * wait. */
+	if (panel_power_off_duration < (s64)intel_dp->pps.panel_power_cycle_delay)
+		wait_remaining_ms_from_jiffies(jiffies,
+				       intel_dp->pps.panel_power_cycle_delay - panel_power_off_duration);
+
+	wait_panel_status(intel_dp, IDLE_CYCLE_MASK, IDLE_CYCLE_VALUE);
+}
+
+void intel_pps_wait_power_cycle(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref)
+		wait_panel_power_cycle(intel_dp);
+}
+
+static void wait_backlight_on(struct intel_dp *intel_dp)
+{
+	wait_remaining_ms_from_jiffies(intel_dp->pps.last_power_on,
+				       intel_dp->pps.backlight_on_delay);
+}
+
+static void edp_wait_backlight_off(struct intel_dp *intel_dp)
+{
+	wait_remaining_ms_from_jiffies(intel_dp->pps.last_backlight_off,
+				       intel_dp->pps.backlight_off_delay);
+}
+
+/* Read the current pp_control value, unlocking the register if it
+ * is locked
+ */
+
+static  u32 ilk_get_pp_control(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 control;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	control = intel_de_read(dev_priv, _pp_ctrl_reg(intel_dp));
+	if (drm_WARN_ON(&dev_priv->drm, !HAS_DDI(dev_priv) &&
+			(control & PANEL_UNLOCK_MASK) != PANEL_UNLOCK_REGS)) {
+		control &= ~PANEL_UNLOCK_MASK;
+		control |= PANEL_UNLOCK_REGS;
+	}
+	return control;
+}
+
+/*
+ * Must be paired with intel_pps_vdd_off_unlocked().
+ * Must hold pps_mutex around the whole on/off sequence.
+ * Can be nested with intel_pps_vdd_{on,off}() calls.
+ */
+bool intel_pps_vdd_on_unlocked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	u32 pp;
+	i915_reg_t pp_stat_reg, pp_ctrl_reg;
+	bool need_to_disable = !intel_dp->pps.want_panel_vdd;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return false;
+
+	cancel_delayed_work(&intel_dp->pps.panel_vdd_work);
+	intel_dp->pps.want_panel_vdd = true;
+
+	if (edp_have_panel_vdd(intel_dp))
+		return need_to_disable;
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.vdd_wakeref);
+	intel_dp->pps.vdd_wakeref = intel_display_power_get(dev_priv,
+							    intel_aux_power_domain(dig_port));
+
+	drm_dbg_kms(&dev_priv->drm, "Turning [ENCODER:%d:%s] VDD on\n",
+		    dig_port->base.base.base.id,
+		    dig_port->base.base.name);
+
+	if (!edp_have_panel_power(intel_dp))
+		wait_panel_power_cycle(intel_dp);
+
+	pp = ilk_get_pp_control(intel_dp);
+	pp |= EDP_FORCE_VDD;
+
+	pp_stat_reg = _pp_stat_reg(intel_dp);
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+
+	intel_de_write(dev_priv, pp_ctrl_reg, pp);
+	intel_de_posting_read(dev_priv, pp_ctrl_reg);
+	drm_dbg_kms(&dev_priv->drm, "PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
+		    intel_de_read(dev_priv, pp_stat_reg),
+		    intel_de_read(dev_priv, pp_ctrl_reg));
+	/*
+	 * If the panel wasn't on, delay before accessing aux channel
+	 */
+	if (!edp_have_panel_power(intel_dp)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "[ENCODER:%d:%s] panel power wasn't enabled\n",
+			    dig_port->base.base.base.id,
+			    dig_port->base.base.name);
+		msleep(intel_dp->pps.panel_power_up_delay);
+	}
+
+	return need_to_disable;
+}
+
+/*
+ * Must be paired with intel_pps_off().
+ * Nested calls to these functions are not allowed since
+ * we drop the lock. Caller must use some higher level
+ * locking to prevent nested calls from other threads.
+ */
+void intel_pps_vdd_on(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+	bool vdd;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	vdd = false;
+	with_intel_pps_lock(intel_dp, wakeref)
+		vdd = intel_pps_vdd_on_unlocked(intel_dp);
+	I915_STATE_WARN(!vdd, "[ENCODER:%d:%s] VDD already requested on\n",
+			dp_to_dig_port(intel_dp)->base.base.base.id,
+			dp_to_dig_port(intel_dp)->base.base.name);
+}
+
+static void intel_pps_vdd_off_sync_unlocked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port =
+		dp_to_dig_port(intel_dp);
+	u32 pp;
+	i915_reg_t pp_stat_reg, pp_ctrl_reg;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.want_panel_vdd);
+
+	if (!edp_have_panel_vdd(intel_dp))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "Turning [ENCODER:%d:%s] VDD off\n",
+		    dig_port->base.base.base.id,
+		    dig_port->base.base.name);
+
+	pp = ilk_get_pp_control(intel_dp);
+	pp &= ~EDP_FORCE_VDD;
+
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+	pp_stat_reg = _pp_stat_reg(intel_dp);
+
+	intel_de_write(dev_priv, pp_ctrl_reg, pp);
+	intel_de_posting_read(dev_priv, pp_ctrl_reg);
+
+	/* Make sure sequencer is idle before allowing subsequent activity */
+	drm_dbg_kms(&dev_priv->drm, "PP_STATUS: 0x%08x PP_CONTROL: 0x%08x\n",
+		    intel_de_read(dev_priv, pp_stat_reg),
+		    intel_de_read(dev_priv, pp_ctrl_reg));
+
+	if ((pp & PANEL_POWER_ON) == 0)
+		intel_dp->pps.panel_power_off_time = ktime_get_boottime();
+
+	intel_display_power_put(dev_priv,
+				intel_aux_power_domain(dig_port),
+				fetch_and_zero(&intel_dp->pps.vdd_wakeref));
+}
+
+void intel_pps_vdd_off_sync(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	cancel_delayed_work_sync(&intel_dp->pps.panel_vdd_work);
+	/*
+	 * vdd might still be enabled due to the delayed vdd off.
+	 * Make sure vdd is actually turned off here.
+	 */
+	with_intel_pps_lock(intel_dp, wakeref)
+		intel_pps_vdd_off_sync_unlocked(intel_dp);
+}
+
+static void edp_panel_vdd_work(struct work_struct *__work)
+{
+	struct intel_pps *pps = container_of(to_delayed_work(__work),
+					     struct intel_pps, panel_vdd_work);
+	struct intel_dp *intel_dp = container_of(pps, struct intel_dp, pps);
+	intel_wakeref_t wakeref;
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		if (!intel_dp->pps.want_panel_vdd)
+			intel_pps_vdd_off_sync_unlocked(intel_dp);
+	}
+}
+
+static void edp_panel_vdd_schedule_off(struct intel_dp *intel_dp)
+{
+	unsigned long delay;
+
+	/*
+	 * We may not yet know the real power sequencing delays,
+	 * so keep VDD enabled until we're done with init.
+	 */
+	if (intel_dp->pps.initializing)
+		return;
+
+	/*
+	 * Queue the timer to fire a long time from now (relative to the power
+	 * down delay) to keep the panel power up across a sequence of
+	 * operations.
+	 */
+	delay = msecs_to_jiffies(intel_dp->pps.panel_power_cycle_delay * 5);
+	schedule_delayed_work(&intel_dp->pps.panel_vdd_work, delay);
+}
+
+/*
+ * Must be paired with edp_panel_vdd_on().
+ * Must hold pps_mutex around the whole on/off sequence.
+ * Can be nested with intel_pps_vdd_{on,off}() calls.
+ */
+void intel_pps_vdd_off_unlocked(struct intel_dp *intel_dp, bool sync)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	I915_STATE_WARN(!intel_dp->pps.want_panel_vdd, "[ENCODER:%d:%s] VDD not forced on",
+			dp_to_dig_port(intel_dp)->base.base.base.id,
+			dp_to_dig_port(intel_dp)->base.base.name);
+
+	intel_dp->pps.want_panel_vdd = false;
+
+	if (sync)
+		intel_pps_vdd_off_sync_unlocked(intel_dp);
+	else
+		edp_panel_vdd_schedule_off(intel_dp);
+}
+
+void intel_pps_on_unlocked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 pp;
+	i915_reg_t pp_ctrl_reg;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "Turn [ENCODER:%d:%s] panel power on\n",
+		    dp_to_dig_port(intel_dp)->base.base.base.id,
+		    dp_to_dig_port(intel_dp)->base.base.name);
+
+	if (drm_WARN(&dev_priv->drm, edp_have_panel_power(intel_dp),
+		     "[ENCODER:%d:%s] panel power already on\n",
+		     dp_to_dig_port(intel_dp)->base.base.base.id,
+		     dp_to_dig_port(intel_dp)->base.base.name))
+		return;
+
+	wait_panel_power_cycle(intel_dp);
+
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+	pp = ilk_get_pp_control(intel_dp);
+	if (IS_IRONLAKE(dev_priv)) {
+		/* ILK workaround: disable reset around power sequence */
+		pp &= ~PANEL_POWER_RESET;
+		intel_de_write(dev_priv, pp_ctrl_reg, pp);
+		intel_de_posting_read(dev_priv, pp_ctrl_reg);
+	}
+
+	pp |= PANEL_POWER_ON;
+	if (!IS_IRONLAKE(dev_priv))
+		pp |= PANEL_POWER_RESET;
+
+	intel_de_write(dev_priv, pp_ctrl_reg, pp);
+	intel_de_posting_read(dev_priv, pp_ctrl_reg);
+
+	wait_panel_on(intel_dp);
+	intel_dp->pps.last_power_on = jiffies;
+
+	if (IS_IRONLAKE(dev_priv)) {
+		pp |= PANEL_POWER_RESET; /* restore panel reset bit */
+		intel_de_write(dev_priv, pp_ctrl_reg, pp);
+		intel_de_posting_read(dev_priv, pp_ctrl_reg);
+	}
+}
+
+void intel_pps_on(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref)
+		intel_pps_on_unlocked(intel_dp);
+}
+
+void intel_pps_off_unlocked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	u32 pp;
+	i915_reg_t pp_ctrl_reg;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "Turn [ENCODER:%d:%s] panel power off\n",
+		    dig_port->base.base.base.id, dig_port->base.base.name);
+
+	drm_WARN(&dev_priv->drm, !intel_dp->pps.want_panel_vdd,
+		 "Need [ENCODER:%d:%s] VDD to turn off panel\n",
+		 dig_port->base.base.base.id, dig_port->base.base.name);
+
+	pp = ilk_get_pp_control(intel_dp);
+	/* We need to switch off panel power _and_ force vdd, for otherwise some
+	 * panels get very unhappy and cease to work. */
+	pp &= ~(PANEL_POWER_ON | PANEL_POWER_RESET | EDP_FORCE_VDD |
+		EDP_BLC_ENABLE);
+
+	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+
+	intel_dp->pps.want_panel_vdd = false;
+
+	intel_de_write(dev_priv, pp_ctrl_reg, pp);
+	intel_de_posting_read(dev_priv, pp_ctrl_reg);
+
+	wait_panel_off(intel_dp);
+	intel_dp->pps.panel_power_off_time = ktime_get_boottime();
+
+	/* We got a reference when we enabled the VDD. */
+	intel_display_power_put(dev_priv,
+				intel_aux_power_domain(dig_port),
+				fetch_and_zero(&intel_dp->pps.vdd_wakeref));
+}
+
+void intel_pps_off(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref)
+		intel_pps_off_unlocked(intel_dp);
+}
+
+/* Enable backlight in the panel power control. */
+void intel_pps_backlight_on(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	intel_wakeref_t wakeref;
+
+	/*
+	 * If we enable the backlight right away following a panel power
+	 * on, we may see slight flicker as the panel syncs with the eDP
+	 * link.  So delay a bit to make sure the image is solid before
+	 * allowing it to appear.
+	 */
+	wait_backlight_on(intel_dp);
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		i915_reg_t pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+		u32 pp;
+
+		pp = ilk_get_pp_control(intel_dp);
+		pp |= EDP_BLC_ENABLE;
+
+		intel_de_write(dev_priv, pp_ctrl_reg, pp);
+		intel_de_posting_read(dev_priv, pp_ctrl_reg);
+	}
+}
+
+/* Disable backlight in the panel power control. */
+void intel_pps_backlight_off(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		i915_reg_t pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
+		u32 pp;
+
+		pp = ilk_get_pp_control(intel_dp);
+		pp &= ~EDP_BLC_ENABLE;
+
+		intel_de_write(dev_priv, pp_ctrl_reg, pp);
+		intel_de_posting_read(dev_priv, pp_ctrl_reg);
+	}
+
+	intel_dp->pps.last_backlight_off = jiffies;
+	edp_wait_backlight_off(intel_dp);
+}
+
+/*
+ * Hook for controlling the panel power control backlight through the bl_power
+ * sysfs attribute. Take care to handle multiple calls.
+ */
+void intel_pps_backlight_power(struct intel_connector *connector, bool enable)
+{
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_dp *intel_dp = intel_attached_dp(connector);
+	intel_wakeref_t wakeref;
+	bool is_enabled;
+
+	is_enabled = false;
+	with_intel_pps_lock(intel_dp, wakeref)
+		is_enabled = ilk_get_pp_control(intel_dp) & EDP_BLC_ENABLE;
+	if (is_enabled == enable)
+		return;
+
+	drm_dbg_kms(&i915->drm, "panel power control backlight %s\n",
+		    enable ? "enable" : "disable");
+
+	if (enable)
+		intel_pps_backlight_on(intel_dp);
+	else
+		intel_pps_backlight_off(intel_dp);
+}
+
+static void vlv_detach_power_sequencer(struct intel_dp *intel_dp)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
+	enum pipe pipe = intel_dp->pps.pps_pipe;
+	i915_reg_t pp_on_reg = PP_ON_DELAYS(pipe);
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.active_pipe != INVALID_PIPE);
+
+	if (drm_WARN_ON(&dev_priv->drm, pipe != PIPE_A && pipe != PIPE_B))
+		return;
+
+	intel_pps_vdd_off_sync_unlocked(intel_dp);
+
+	/*
+	 * VLV seems to get confused when multiple power sequencers
+	 * have the same port selected (even if only one has power/vdd
+	 * enabled). The failure manifests as vlv_wait_port_ready() failing
+	 * CHV on the other hand doesn't seem to mind having the same port
+	 * selected in multiple power sequencers, but let's clear the
+	 * port select always when logically disconnecting a power sequencer
+	 * from a port.
+	 */
+	drm_dbg_kms(&dev_priv->drm,
+		    "detaching pipe %c power sequencer from [ENCODER:%d:%s]\n",
+		    pipe_name(pipe), dig_port->base.base.base.id,
+		    dig_port->base.base.name);
+	intel_de_write(dev_priv, pp_on_reg, 0);
+	intel_de_posting_read(dev_priv, pp_on_reg);
+
+	intel_dp->pps.pps_pipe = INVALID_PIPE;
+}
+
+static void vlv_steal_power_sequencer(struct drm_i915_private *dev_priv,
+				      enum pipe pipe)
+{
+	struct intel_encoder *encoder;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	for_each_intel_dp(&dev_priv->drm, encoder) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		drm_WARN(&dev_priv->drm, intel_dp->pps.active_pipe == pipe,
+			 "stealing pipe %c power sequencer from active [ENCODER:%d:%s]\n",
+			 pipe_name(pipe), encoder->base.base.id,
+			 encoder->base.name);
+
+		if (intel_dp->pps.pps_pipe != pipe)
+			continue;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "stealing pipe %c power sequencer from [ENCODER:%d:%s]\n",
+			    pipe_name(pipe), encoder->base.base.id,
+			    encoder->base.name);
+
+		/* make sure vdd is off before we steal it */
+		vlv_detach_power_sequencer(intel_dp);
+	}
+}
+
+void vlv_pps_init(struct intel_encoder *encoder,
+		  const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.active_pipe != INVALID_PIPE);
+
+	if (intel_dp->pps.pps_pipe != INVALID_PIPE &&
+	    intel_dp->pps.pps_pipe != crtc->pipe) {
+		/*
+		 * If another power sequencer was being used on this
+		 * port previously make sure to turn off vdd there while
+		 * we still have control of it.
+		 */
+		vlv_detach_power_sequencer(intel_dp);
+	}
+
+	/*
+	 * We may be stealing the power
+	 * sequencer from another port.
+	 */
+	vlv_steal_power_sequencer(dev_priv, crtc->pipe);
+
+	intel_dp->pps.active_pipe = crtc->pipe;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	/* now it's all ours */
+	intel_dp->pps.pps_pipe = crtc->pipe;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "initializing pipe %c power sequencer for [ENCODER:%d:%s]\n",
+		    pipe_name(intel_dp->pps.pps_pipe), encoder->base.base.id,
+		    encoder->base.name);
+
+	/* init power sequencer on this pipe and port */
+	pps_init_delays(intel_dp);
+	pps_init_registers(intel_dp, true);
+}
+
+static void pps_vdd_init(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	if (!edp_have_panel_vdd(intel_dp))
+		return;
+
+	/*
+	 * The VDD bit needs a power domain reference, so if the bit is
+	 * already enabled when we boot or resume, grab this reference and
+	 * schedule a vdd off, so we don't hold on to the reference
+	 * indefinitely.
+	 */
+	drm_dbg_kms(&dev_priv->drm,
+		    "VDD left on by BIOS, adjusting state tracking\n");
+	drm_WARN_ON(&dev_priv->drm, intel_dp->pps.vdd_wakeref);
+	intel_dp->pps.vdd_wakeref = intel_display_power_get(dev_priv,
+							    intel_aux_power_domain(dig_port));
+}
+
+bool intel_pps_have_panel_power_or_vdd(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+	bool have_power = false;
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		have_power = edp_have_panel_power(intel_dp) ||
+			     edp_have_panel_vdd(intel_dp);
+	}
+
+	return have_power;
+}
+
+static void pps_init_timestamps(struct intel_dp *intel_dp)
+{
+	intel_dp->pps.panel_power_off_time = ktime_get_boottime();
+	intel_dp->pps.last_power_on = jiffies;
+	intel_dp->pps.last_backlight_off = jiffies;
+}
+
+static void
+intel_pps_readout_hw_state(struct intel_dp *intel_dp, struct edp_power_seq *seq)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 pp_on, pp_off, pp_ctl;
+	struct pps_registers regs;
+
+	intel_pps_get_registers(intel_dp, &regs);
+
+	pp_ctl = ilk_get_pp_control(intel_dp);
+
+	/* Ensure PPS is unlocked */
+	if (!HAS_DDI(dev_priv))
+		intel_de_write(dev_priv, regs.pp_ctrl, pp_ctl);
+
+	pp_on = intel_de_read(dev_priv, regs.pp_on);
+	pp_off = intel_de_read(dev_priv, regs.pp_off);
+
+	/* Pull timing values out of registers */
+	seq->t1_t3 = REG_FIELD_GET(PANEL_POWER_UP_DELAY_MASK, pp_on);
+	seq->t8 = REG_FIELD_GET(PANEL_LIGHT_ON_DELAY_MASK, pp_on);
+	seq->t9 = REG_FIELD_GET(PANEL_LIGHT_OFF_DELAY_MASK, pp_off);
+	seq->t10 = REG_FIELD_GET(PANEL_POWER_DOWN_DELAY_MASK, pp_off);
+
+	if (i915_mmio_reg_valid(regs.pp_div)) {
+		u32 pp_div;
+
+		pp_div = intel_de_read(dev_priv, regs.pp_div);
+
+		seq->t11_t12 = REG_FIELD_GET(PANEL_POWER_CYCLE_DELAY_MASK, pp_div) * 1000;
+	} else {
+		seq->t11_t12 = REG_FIELD_GET(BXT_POWER_CYCLE_DELAY_MASK, pp_ctl) * 1000;
+	}
+}
+
+static void
+intel_pps_dump_state(struct intel_dp *intel_dp, const char *state_name,
+		     const struct edp_power_seq *seq)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	drm_dbg_kms(&i915->drm, "%s t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n",
+		    state_name,
+		    seq->t1_t3, seq->t8, seq->t9, seq->t10, seq->t11_t12);
+}
+
+static void
+intel_pps_verify_state(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	struct edp_power_seq hw;
+	struct edp_power_seq *sw = &intel_dp->pps.pps_delays;
+
+	intel_pps_readout_hw_state(intel_dp, &hw);
+
+	if (hw.t1_t3 != sw->t1_t3 || hw.t8 != sw->t8 || hw.t9 != sw->t9 ||
+	    hw.t10 != sw->t10 || hw.t11_t12 != sw->t11_t12) {
+		drm_err(&i915->drm, "PPS state mismatch\n");
+		intel_pps_dump_state(intel_dp, "sw", sw);
+		intel_pps_dump_state(intel_dp, "hw", &hw);
+	}
+}
+
+static bool pps_delays_valid(struct edp_power_seq *delays)
+{
+	return delays->t1_t3 || delays->t8 || delays->t9 ||
+		delays->t10 || delays->t11_t12;
+}
+
+static void pps_init_delays_bios(struct intel_dp *intel_dp,
+				 struct edp_power_seq *bios)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	if (!pps_delays_valid(&intel_dp->pps.bios_pps_delays))
+		intel_pps_readout_hw_state(intel_dp, &intel_dp->pps.bios_pps_delays);
+
+	*bios = intel_dp->pps.bios_pps_delays;
+
+	intel_pps_dump_state(intel_dp, "bios", bios);
+}
+
+static void pps_init_delays_vbt(struct intel_dp *intel_dp,
+				struct edp_power_seq *vbt)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_connector *connector = intel_dp->attached_connector;
+
+	*vbt = connector->panel.vbt.edp.pps;
+
+	if (!pps_delays_valid(vbt))
+		return;
+
+	/* On Toshiba Satellite P50-C-18C system the VBT T12 delay
+	 * of 500ms appears to be too short. Ocassionally the panel
+	 * just fails to power back on. Increasing the delay to 800ms
+	 * seems sufficient to avoid this problem.
+	 */
+	if (intel_has_quirk(dev_priv, QUIRK_INCREASE_T12_DELAY)) {
+		vbt->t11_t12 = max_t(u16, vbt->t11_t12, 1300 * 10);
+		drm_dbg_kms(&dev_priv->drm,
+			    "Increasing T12 panel delay as per the quirk to %d\n",
+			    vbt->t11_t12);
+	}
+
+	/* T11_T12 delay is special and actually in units of 100ms, but zero
+	 * based in the hw (so we need to add 100 ms). But the sw vbt
+	 * table multiplies it with 1000 to make it in units of 100usec,
+	 * too. */
+	vbt->t11_t12 += 100 * 10;
+
+	intel_pps_dump_state(intel_dp, "vbt", vbt);
+}
+
+static void pps_init_delays_spec(struct intel_dp *intel_dp,
+				 struct edp_power_seq *spec)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	/* Upper limits from eDP 1.3 spec. Note that we use the clunky units of
+	 * our hw here, which are all in 100usec. */
+	spec->t1_t3 = 210 * 10;
+	spec->t8 = 50 * 10; /* no limit for t8, use t7 instead */
+	spec->t9 = 50 * 10; /* no limit for t9, make it symmetric with t8 */
+	spec->t10 = 500 * 10;
+	/* This one is special and actually in units of 100ms, but zero
+	 * based in the hw (so we need to add 100 ms). But the sw vbt
+	 * table multiplies it with 1000 to make it in units of 100usec,
+	 * too. */
+	spec->t11_t12 = (510 + 100) * 10;
+
+	intel_pps_dump_state(intel_dp, "spec", spec);
+}
+
+static void pps_init_delays(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct edp_power_seq cur, vbt, spec,
+		*final = &intel_dp->pps.pps_delays;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	/* already initialized? */
+	if (pps_delays_valid(final))
+		return;
+
+	pps_init_delays_bios(intel_dp, &cur);
+	pps_init_delays_vbt(intel_dp, &vbt);
+	pps_init_delays_spec(intel_dp, &spec);
+
+	/* Use the max of the register settings and vbt. If both are
+	 * unset, fall back to the spec limits. */
+#define assign_final(field)	final->field = (max(cur.field, vbt.field) == 0 ? \
+				       spec.field : \
+				       max(cur.field, vbt.field))
+	assign_final(t1_t3);
+	assign_final(t8);
+	assign_final(t9);
+	assign_final(t10);
+	assign_final(t11_t12);
+#undef assign_final
+
+#define get_delay(field)	(DIV_ROUND_UP(final->field, 10))
+	intel_dp->pps.panel_power_up_delay = get_delay(t1_t3);
+	intel_dp->pps.backlight_on_delay = get_delay(t8);
+	intel_dp->pps.backlight_off_delay = get_delay(t9);
+	intel_dp->pps.panel_power_down_delay = get_delay(t10);
+	intel_dp->pps.panel_power_cycle_delay = get_delay(t11_t12);
+#undef get_delay
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "panel power up delay %d, power down delay %d, power cycle delay %d\n",
+		    intel_dp->pps.panel_power_up_delay,
+		    intel_dp->pps.panel_power_down_delay,
+		    intel_dp->pps.panel_power_cycle_delay);
+
+	drm_dbg_kms(&dev_priv->drm, "backlight on delay %d, off delay %d\n",
+		    intel_dp->pps.backlight_on_delay,
+		    intel_dp->pps.backlight_off_delay);
+
+	/*
+	 * We override the HW backlight delays to 1 because we do manual waits
+	 * on them. For T8, even BSpec recommends doing it. For T9, if we
+	 * don't do this, we'll end up waiting for the backlight off delay
+	 * twice: once when we do the manual sleep, and once when we disable
+	 * the panel and wait for the PP_STATUS bit to become zero.
+	 */
+	final->t8 = 1;
+	final->t9 = 1;
+
+	/*
+	 * HW has only a 100msec granularity for t11_t12 so round it up
+	 * accordingly.
+	 */
+	final->t11_t12 = roundup(final->t11_t12, 100 * 10);
+}
+
+static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 pp_on, pp_off, port_sel = 0;
+	int div = RUNTIME_INFO(dev_priv)->rawclk_freq / 1000;
+	struct pps_registers regs;
+	enum port port = dp_to_dig_port(intel_dp)->base.port;
+	const struct edp_power_seq *seq = &intel_dp->pps.pps_delays;
+
+	lockdep_assert_held(&dev_priv->display.pps.mutex);
+
+	intel_pps_get_registers(intel_dp, &regs);
+
+	/*
+	 * On some VLV machines the BIOS can leave the VDD
+	 * enabled even on power sequencers which aren't
+	 * hooked up to any port. This would mess up the
+	 * power domain tracking the first time we pick
+	 * one of these power sequencers for use since
+	 * intel_pps_vdd_on_unlocked() would notice that the VDD was
+	 * already on and therefore wouldn't grab the power
+	 * domain reference. Disable VDD first to avoid this.
+	 * This also avoids spuriously turning the VDD on as
+	 * soon as the new power sequencer gets initialized.
+	 */
+	if (force_disable_vdd) {
+		u32 pp = ilk_get_pp_control(intel_dp);
+
+		drm_WARN(&dev_priv->drm, pp & PANEL_POWER_ON,
+			 "Panel power already on\n");
+
+		if (pp & EDP_FORCE_VDD)
+			drm_dbg_kms(&dev_priv->drm,
+				    "VDD already on, disabling first\n");
+
+		pp &= ~EDP_FORCE_VDD;
+
+		intel_de_write(dev_priv, regs.pp_ctrl, pp);
+	}
+
+	pp_on = REG_FIELD_PREP(PANEL_POWER_UP_DELAY_MASK, seq->t1_t3) |
+		REG_FIELD_PREP(PANEL_LIGHT_ON_DELAY_MASK, seq->t8);
+	pp_off = REG_FIELD_PREP(PANEL_LIGHT_OFF_DELAY_MASK, seq->t9) |
+		REG_FIELD_PREP(PANEL_POWER_DOWN_DELAY_MASK, seq->t10);
+
+	/* Haswell doesn't have any port selection bits for the panel
+	 * power sequencer any more. */
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		port_sel = PANEL_PORT_SELECT_VLV(port);
+	} else if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)) {
+		switch (port) {
+		case PORT_A:
+			port_sel = PANEL_PORT_SELECT_DPA;
+			break;
+		case PORT_C:
+			port_sel = PANEL_PORT_SELECT_DPC;
+			break;
+		case PORT_D:
+			port_sel = PANEL_PORT_SELECT_DPD;
+			break;
+		default:
+			MISSING_CASE(port);
+			break;
+		}
+	}
+
+	pp_on |= port_sel;
+
+	intel_de_write(dev_priv, regs.pp_on, pp_on);
+	intel_de_write(dev_priv, regs.pp_off, pp_off);
+
+	/*
+	 * Compute the divisor for the pp clock, simply match the Bspec formula.
+	 */
+	if (i915_mmio_reg_valid(regs.pp_div)) {
+		intel_de_write(dev_priv, regs.pp_div,
+			       REG_FIELD_PREP(PP_REFERENCE_DIVIDER_MASK, (100 * div) / 2 - 1) | REG_FIELD_PREP(PANEL_POWER_CYCLE_DELAY_MASK, DIV_ROUND_UP(seq->t11_t12, 1000)));
+	} else {
+		u32 pp_ctl;
+
+		pp_ctl = intel_de_read(dev_priv, regs.pp_ctrl);
+		pp_ctl &= ~BXT_POWER_CYCLE_DELAY_MASK;
+		pp_ctl |= REG_FIELD_PREP(BXT_POWER_CYCLE_DELAY_MASK, DIV_ROUND_UP(seq->t11_t12, 1000));
+		intel_de_write(dev_priv, regs.pp_ctrl, pp_ctl);
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "panel power sequencer register settings: PP_ON %#x, PP_OFF %#x, PP_DIV %#x\n",
+		    intel_de_read(dev_priv, regs.pp_on),
+		    intel_de_read(dev_priv, regs.pp_off),
+		    i915_mmio_reg_valid(regs.pp_div) ?
+		    intel_de_read(dev_priv, regs.pp_div) :
+		    (intel_de_read(dev_priv, regs.pp_ctrl) & BXT_POWER_CYCLE_DELAY_MASK));
+}
+
+void intel_pps_encoder_reset(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	intel_wakeref_t wakeref;
+
+	if (!intel_dp_is_edp(intel_dp))
+		return;
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		/*
+		 * Reinit the power sequencer also on the resume path, in case
+		 * BIOS did something nasty with it.
+		 */
+		if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+			vlv_initial_power_sequencer_setup(intel_dp);
+
+		pps_init_delays(intel_dp);
+		pps_init_registers(intel_dp, false);
+		pps_vdd_init(intel_dp);
+
+		if (edp_have_panel_vdd(intel_dp))
+			edp_panel_vdd_schedule_off(intel_dp);
+	}
+}
+
+void intel_pps_init(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	intel_wakeref_t wakeref;
+
+	intel_dp->pps.initializing = true;
+	INIT_DELAYED_WORK(&intel_dp->pps.panel_vdd_work, edp_panel_vdd_work);
+
+	pps_init_timestamps(intel_dp);
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+			vlv_initial_power_sequencer_setup(intel_dp);
+
+		pps_init_delays(intel_dp);
+		pps_init_registers(intel_dp, false);
+		pps_vdd_init(intel_dp);
+	}
+}
+
+void intel_pps_init_late(struct intel_dp *intel_dp)
+{
+	intel_wakeref_t wakeref;
+
+	with_intel_pps_lock(intel_dp, wakeref) {
+		/* Reinit delays after per-panel info has been parsed from VBT */
+		memset(&intel_dp->pps.pps_delays, 0, sizeof(intel_dp->pps.pps_delays));
+		pps_init_delays(intel_dp);
+		pps_init_registers(intel_dp, false);
+
+		intel_dp->pps.initializing = false;
+
+		if (edp_have_panel_vdd(intel_dp))
+			edp_panel_vdd_schedule_off(intel_dp);
+	}
+}
+
+void intel_pps_unlock_regs_wa(struct drm_i915_private *dev_priv)
+{
+	int pps_num;
+	int pps_idx;
+
+	if (!HAS_DISPLAY(dev_priv) || HAS_DDI(dev_priv))
+		return;
+	/*
+	 * This w/a is needed at least on CPT/PPT, but to be sure apply it
+	 * everywhere where registers can be write protected.
+	 */
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		pps_num = 2;
+	else
+		pps_num = 1;
+
+	for (pps_idx = 0; pps_idx < pps_num; pps_idx++) {
+		u32 val = intel_de_read(dev_priv, PP_CONTROL(pps_idx));
+
+		val = (val & ~PANEL_UNLOCK_MASK) | PANEL_UNLOCK_REGS;
+		intel_de_write(dev_priv, PP_CONTROL(pps_idx), val);
+	}
+}
+
+void intel_pps_setup(struct drm_i915_private *i915)
+{
+	if (HAS_PCH_SPLIT(i915) || IS_GEMINILAKE(i915) || IS_BROXTON(i915))
+		i915->display.pps.mmio_base = PCH_PPS_BASE;
+	else if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+		i915->display.pps.mmio_base = VLV_PPS_BASE;
+	else
+		i915->display.pps.mmio_base = PPS_BASE;
+}
+
+void assert_pps_unlocked(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	i915_reg_t pp_reg;
+	u32 val;
+	enum pipe panel_pipe = INVALID_PIPE;
+	bool locked = true;
+
+	if (drm_WARN_ON(&dev_priv->drm, HAS_DDI(dev_priv)))
+		return;
+
+	if (HAS_PCH_SPLIT(dev_priv)) {
+		u32 port_sel;
+
+		pp_reg = PP_CONTROL(0);
+		port_sel = intel_de_read(dev_priv, PP_ON_DELAYS(0)) & PANEL_PORT_SELECT_MASK;
+
+		switch (port_sel) {
+		case PANEL_PORT_SELECT_LVDS:
+			intel_lvds_port_enabled(dev_priv, PCH_LVDS, &panel_pipe);
+			break;
+		case PANEL_PORT_SELECT_DPA:
+			g4x_dp_port_enabled(dev_priv, DP_A, PORT_A, &panel_pipe);
+			break;
+		case PANEL_PORT_SELECT_DPC:
+			g4x_dp_port_enabled(dev_priv, PCH_DP_C, PORT_C, &panel_pipe);
+			break;
+		case PANEL_PORT_SELECT_DPD:
+			g4x_dp_port_enabled(dev_priv, PCH_DP_D, PORT_D, &panel_pipe);
+			break;
+		default:
+			MISSING_CASE(port_sel);
+			break;
+		}
+	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		/* presumably write lock depends on pipe, not port select */
+		pp_reg = PP_CONTROL(pipe);
+		panel_pipe = pipe;
+	} else {
+		u32 port_sel;
+
+		pp_reg = PP_CONTROL(0);
+		port_sel = intel_de_read(dev_priv, PP_ON_DELAYS(0)) & PANEL_PORT_SELECT_MASK;
+
+		drm_WARN_ON(&dev_priv->drm,
+			    port_sel != PANEL_PORT_SELECT_LVDS);
+		intel_lvds_port_enabled(dev_priv, LVDS, &panel_pipe);
+	}
+
+	val = intel_de_read(dev_priv, pp_reg);
+	if (!(val & PANEL_POWER_ON) ||
+	    ((val & PANEL_UNLOCK_MASK) == PANEL_UNLOCK_REGS))
+		locked = false;
+
+	I915_STATE_WARN(panel_pipe == pipe && locked,
+			"panel assertion failure, pipe %c regs locked\n",
+			pipe_name(pipe));
+}
diff --git a/drivers/gpu/drm/i915/display/intel_pps.h b/drivers/gpu/drm/i915/display/intel_pps.h
new file mode 100644
index 000000000..a3a56f903
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_pps.h
@@ -0,0 +1,56 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __INTEL_PPS_H__
+#define __INTEL_PPS_H__
+
+#include <linux/types.h>
+
+#include "intel_wakeref.h"
+
+enum pipe;
+struct drm_i915_private;
+struct intel_connector;
+struct intel_crtc_state;
+struct intel_dp;
+struct intel_encoder;
+
+intel_wakeref_t intel_pps_lock(struct intel_dp *intel_dp);
+intel_wakeref_t intel_pps_unlock(struct intel_dp *intel_dp, intel_wakeref_t wakeref);
+
+#define with_intel_pps_lock(dp, wf)						\
+	for ((wf) = intel_pps_lock(dp); (wf); (wf) = intel_pps_unlock((dp), (wf)))
+
+void intel_pps_backlight_on(struct intel_dp *intel_dp);
+void intel_pps_backlight_off(struct intel_dp *intel_dp);
+void intel_pps_backlight_power(struct intel_connector *connector, bool enable);
+
+bool intel_pps_vdd_on_unlocked(struct intel_dp *intel_dp);
+void intel_pps_vdd_off_unlocked(struct intel_dp *intel_dp, bool sync);
+void intel_pps_on_unlocked(struct intel_dp *intel_dp);
+void intel_pps_off_unlocked(struct intel_dp *intel_dp);
+void intel_pps_check_power_unlocked(struct intel_dp *intel_dp);
+
+void intel_pps_vdd_on(struct intel_dp *intel_dp);
+void intel_pps_on(struct intel_dp *intel_dp);
+void intel_pps_off(struct intel_dp *intel_dp);
+void intel_pps_vdd_off_sync(struct intel_dp *intel_dp);
+bool intel_pps_have_panel_power_or_vdd(struct intel_dp *intel_dp);
+void intel_pps_wait_power_cycle(struct intel_dp *intel_dp);
+
+void intel_pps_init(struct intel_dp *intel_dp);
+void intel_pps_init_late(struct intel_dp *intel_dp);
+void intel_pps_encoder_reset(struct intel_dp *intel_dp);
+void intel_pps_reset_all(struct drm_i915_private *i915);
+
+void vlv_pps_init(struct intel_encoder *encoder,
+		  const struct intel_crtc_state *crtc_state);
+
+void intel_pps_unlock_regs_wa(struct drm_i915_private *i915);
+void intel_pps_setup(struct drm_i915_private *i915);
+
+void assert_pps_unlocked(struct drm_i915_private *i915, enum pipe pipe);
+
+#endif /* __INTEL_PPS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c
new file mode 100644
index 000000000..e2d7c0a68
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_psr.c
@@ -0,0 +1,2656 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_damage_helper.h>
+
+#include "display/intel_dp.h"
+
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dp_aux.h"
+#include "intel_hdmi.h"
+#include "intel_psr.h"
+#include "intel_snps_phy.h"
+#include "skl_universal_plane.h"
+
+/**
+ * DOC: Panel Self Refresh (PSR/SRD)
+ *
+ * Since Haswell Display controller supports Panel Self-Refresh on display
+ * panels witch have a remote frame buffer (RFB) implemented according to PSR
+ * spec in eDP1.3. PSR feature allows the display to go to lower standby states
+ * when system is idle but display is on as it eliminates display refresh
+ * request to DDR memory completely as long as the frame buffer for that
+ * display is unchanged.
+ *
+ * Panel Self Refresh must be supported by both Hardware (source) and
+ * Panel (sink).
+ *
+ * PSR saves power by caching the framebuffer in the panel RFB, which allows us
+ * to power down the link and memory controller. For DSI panels the same idea
+ * is called "manual mode".
+ *
+ * The implementation uses the hardware-based PSR support which automatically
+ * enters/exits self-refresh mode. The hardware takes care of sending the
+ * required DP aux message and could even retrain the link (that part isn't
+ * enabled yet though). The hardware also keeps track of any frontbuffer
+ * changes to know when to exit self-refresh mode again. Unfortunately that
+ * part doesn't work too well, hence why the i915 PSR support uses the
+ * software frontbuffer tracking to make sure it doesn't miss a screen
+ * update. For this integration intel_psr_invalidate() and intel_psr_flush()
+ * get called by the frontbuffer tracking code. Note that because of locking
+ * issues the self-refresh re-enable code is done from a work queue, which
+ * must be correctly synchronized/cancelled when shutting down the pipe."
+ *
+ * DC3CO (DC3 clock off)
+ *
+ * On top of PSR2, GEN12 adds a intermediate power savings state that turns
+ * clock off automatically during PSR2 idle state.
+ * The smaller overhead of DC3co entry/exit vs. the overhead of PSR2 deep sleep
+ * entry/exit allows the HW to enter a low-power state even when page flipping
+ * periodically (for instance a 30fps video playback scenario).
+ *
+ * Every time a flips occurs PSR2 will get out of deep sleep state(if it was),
+ * so DC3CO is enabled and tgl_dc3co_disable_work is schedule to run after 6
+ * frames, if no other flip occurs and the function above is executed, DC3CO is
+ * disabled and PSR2 is configured to enter deep sleep, resetting again in case
+ * of another flip.
+ * Front buffer modifications do not trigger DC3CO activation on purpose as it
+ * would bring a lot of complexity and most of the moderns systems will only
+ * use page flips.
+ */
+
+static bool psr_global_enabled(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	switch (intel_dp->psr.debug & I915_PSR_DEBUG_MODE_MASK) {
+	case I915_PSR_DEBUG_DEFAULT:
+		if (i915->params.enable_psr == -1)
+			return connector->panel.vbt.psr.enable;
+		return i915->params.enable_psr;
+	case I915_PSR_DEBUG_DISABLE:
+		return false;
+	default:
+		return true;
+	}
+}
+
+static bool psr2_global_enabled(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+
+	switch (intel_dp->psr.debug & I915_PSR_DEBUG_MODE_MASK) {
+	case I915_PSR_DEBUG_DISABLE:
+	case I915_PSR_DEBUG_FORCE_PSR1:
+		return false;
+	default:
+		if (i915->params.enable_psr == 1)
+			return false;
+		return true;
+	}
+}
+
+static u32 psr_irq_psr_error_bit_get(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	return DISPLAY_VER(dev_priv) >= 12 ? TGL_PSR_ERROR :
+		EDP_PSR_ERROR(intel_dp->psr.transcoder);
+}
+
+static u32 psr_irq_post_exit_bit_get(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	return DISPLAY_VER(dev_priv) >= 12 ? TGL_PSR_POST_EXIT :
+		EDP_PSR_POST_EXIT(intel_dp->psr.transcoder);
+}
+
+static u32 psr_irq_pre_entry_bit_get(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	return DISPLAY_VER(dev_priv) >= 12 ? TGL_PSR_PRE_ENTRY :
+		EDP_PSR_PRE_ENTRY(intel_dp->psr.transcoder);
+}
+
+static u32 psr_irq_mask_get(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	return DISPLAY_VER(dev_priv) >= 12 ? TGL_PSR_MASK :
+		EDP_PSR_MASK(intel_dp->psr.transcoder);
+}
+
+static void psr_irq_control(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	i915_reg_t imr_reg;
+	u32 mask, val;
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		imr_reg = TRANS_PSR_IMR(intel_dp->psr.transcoder);
+	else
+		imr_reg = EDP_PSR_IMR;
+
+	mask = psr_irq_psr_error_bit_get(intel_dp);
+	if (intel_dp->psr.debug & I915_PSR_DEBUG_IRQ)
+		mask |= psr_irq_post_exit_bit_get(intel_dp) |
+			psr_irq_pre_entry_bit_get(intel_dp);
+
+	val = intel_de_read(dev_priv, imr_reg);
+	val &= ~psr_irq_mask_get(intel_dp);
+	val |= ~mask;
+	intel_de_write(dev_priv, imr_reg, val);
+}
+
+static void psr_event_print(struct drm_i915_private *i915,
+			    u32 val, bool psr2_enabled)
+{
+	drm_dbg_kms(&i915->drm, "PSR exit events: 0x%x\n", val);
+	if (val & PSR_EVENT_PSR2_WD_TIMER_EXPIRE)
+		drm_dbg_kms(&i915->drm, "\tPSR2 watchdog timer expired\n");
+	if ((val & PSR_EVENT_PSR2_DISABLED) && psr2_enabled)
+		drm_dbg_kms(&i915->drm, "\tPSR2 disabled\n");
+	if (val & PSR_EVENT_SU_DIRTY_FIFO_UNDERRUN)
+		drm_dbg_kms(&i915->drm, "\tSU dirty FIFO underrun\n");
+	if (val & PSR_EVENT_SU_CRC_FIFO_UNDERRUN)
+		drm_dbg_kms(&i915->drm, "\tSU CRC FIFO underrun\n");
+	if (val & PSR_EVENT_GRAPHICS_RESET)
+		drm_dbg_kms(&i915->drm, "\tGraphics reset\n");
+	if (val & PSR_EVENT_PCH_INTERRUPT)
+		drm_dbg_kms(&i915->drm, "\tPCH interrupt\n");
+	if (val & PSR_EVENT_MEMORY_UP)
+		drm_dbg_kms(&i915->drm, "\tMemory up\n");
+	if (val & PSR_EVENT_FRONT_BUFFER_MODIFY)
+		drm_dbg_kms(&i915->drm, "\tFront buffer modification\n");
+	if (val & PSR_EVENT_WD_TIMER_EXPIRE)
+		drm_dbg_kms(&i915->drm, "\tPSR watchdog timer expired\n");
+	if (val & PSR_EVENT_PIPE_REGISTERS_UPDATE)
+		drm_dbg_kms(&i915->drm, "\tPIPE registers updated\n");
+	if (val & PSR_EVENT_REGISTER_UPDATE)
+		drm_dbg_kms(&i915->drm, "\tRegister updated\n");
+	if (val & PSR_EVENT_HDCP_ENABLE)
+		drm_dbg_kms(&i915->drm, "\tHDCP enabled\n");
+	if (val & PSR_EVENT_KVMR_SESSION_ENABLE)
+		drm_dbg_kms(&i915->drm, "\tKVMR session enabled\n");
+	if (val & PSR_EVENT_VBI_ENABLE)
+		drm_dbg_kms(&i915->drm, "\tVBI enabled\n");
+	if (val & PSR_EVENT_LPSP_MODE_EXIT)
+		drm_dbg_kms(&i915->drm, "\tLPSP mode exited\n");
+	if ((val & PSR_EVENT_PSR_DISABLE) && !psr2_enabled)
+		drm_dbg_kms(&i915->drm, "\tPSR disabled\n");
+}
+
+void intel_psr_irq_handler(struct intel_dp *intel_dp, u32 psr_iir)
+{
+	enum transcoder cpu_transcoder = intel_dp->psr.transcoder;
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	ktime_t time_ns =  ktime_get();
+	i915_reg_t imr_reg;
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		imr_reg = TRANS_PSR_IMR(intel_dp->psr.transcoder);
+	else
+		imr_reg = EDP_PSR_IMR;
+
+	if (psr_iir & psr_irq_pre_entry_bit_get(intel_dp)) {
+		intel_dp->psr.last_entry_attempt = time_ns;
+		drm_dbg_kms(&dev_priv->drm,
+			    "[transcoder %s] PSR entry attempt in 2 vblanks\n",
+			    transcoder_name(cpu_transcoder));
+	}
+
+	if (psr_iir & psr_irq_post_exit_bit_get(intel_dp)) {
+		intel_dp->psr.last_exit = time_ns;
+		drm_dbg_kms(&dev_priv->drm,
+			    "[transcoder %s] PSR exit completed\n",
+			    transcoder_name(cpu_transcoder));
+
+		if (DISPLAY_VER(dev_priv) >= 9) {
+			u32 val = intel_de_read(dev_priv,
+						PSR_EVENT(cpu_transcoder));
+			bool psr2_enabled = intel_dp->psr.psr2_enabled;
+
+			intel_de_write(dev_priv, PSR_EVENT(cpu_transcoder),
+				       val);
+			psr_event_print(dev_priv, val, psr2_enabled);
+		}
+	}
+
+	if (psr_iir & psr_irq_psr_error_bit_get(intel_dp)) {
+		u32 val;
+
+		drm_warn(&dev_priv->drm, "[transcoder %s] PSR aux error\n",
+			 transcoder_name(cpu_transcoder));
+
+		intel_dp->psr.irq_aux_error = true;
+
+		/*
+		 * If this interruption is not masked it will keep
+		 * interrupting so fast that it prevents the scheduled
+		 * work to run.
+		 * Also after a PSR error, we don't want to arm PSR
+		 * again so we don't care about unmask the interruption
+		 * or unset irq_aux_error.
+		 */
+		val = intel_de_read(dev_priv, imr_reg);
+		val |= psr_irq_psr_error_bit_get(intel_dp);
+		intel_de_write(dev_priv, imr_reg, val);
+
+		schedule_work(&intel_dp->psr.work);
+	}
+}
+
+static bool intel_dp_get_alpm_status(struct intel_dp *intel_dp)
+{
+	u8 alpm_caps = 0;
+
+	if (drm_dp_dpcd_readb(&intel_dp->aux, DP_RECEIVER_ALPM_CAP,
+			      &alpm_caps) != 1)
+		return false;
+	return alpm_caps & DP_ALPM_CAP;
+}
+
+static u8 intel_dp_get_sink_sync_latency(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	u8 val = 8; /* assume the worst if we can't read the value */
+
+	if (drm_dp_dpcd_readb(&intel_dp->aux,
+			      DP_SYNCHRONIZATION_LATENCY_IN_SINK, &val) == 1)
+		val &= DP_MAX_RESYNC_FRAME_COUNT_MASK;
+	else
+		drm_dbg_kms(&i915->drm,
+			    "Unable to get sink synchronization latency, assuming 8 frames\n");
+	return val;
+}
+
+static void intel_dp_get_su_granularity(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	ssize_t r;
+	u16 w;
+	u8 y;
+
+	/* If sink don't have specific granularity requirements set legacy ones */
+	if (!(intel_dp->psr_dpcd[1] & DP_PSR2_SU_GRANULARITY_REQUIRED)) {
+		/* As PSR2 HW sends full lines, we do not care about x granularity */
+		w = 4;
+		y = 4;
+		goto exit;
+	}
+
+	r = drm_dp_dpcd_read(&intel_dp->aux, DP_PSR2_SU_X_GRANULARITY, &w, 2);
+	if (r != 2)
+		drm_dbg_kms(&i915->drm,
+			    "Unable to read DP_PSR2_SU_X_GRANULARITY\n");
+	/*
+	 * Spec says that if the value read is 0 the default granularity should
+	 * be used instead.
+	 */
+	if (r != 2 || w == 0)
+		w = 4;
+
+	r = drm_dp_dpcd_read(&intel_dp->aux, DP_PSR2_SU_Y_GRANULARITY, &y, 1);
+	if (r != 1) {
+		drm_dbg_kms(&i915->drm,
+			    "Unable to read DP_PSR2_SU_Y_GRANULARITY\n");
+		y = 4;
+	}
+	if (y == 0)
+		y = 1;
+
+exit:
+	intel_dp->psr.su_w_granularity = w;
+	intel_dp->psr.su_y_granularity = y;
+}
+
+void intel_psr_init_dpcd(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(dp_to_dig_port(intel_dp)->base.base.dev);
+
+	drm_dp_dpcd_read(&intel_dp->aux, DP_PSR_SUPPORT, intel_dp->psr_dpcd,
+			 sizeof(intel_dp->psr_dpcd));
+
+	if (!intel_dp->psr_dpcd[0])
+		return;
+	drm_dbg_kms(&dev_priv->drm, "eDP panel supports PSR version %x\n",
+		    intel_dp->psr_dpcd[0]);
+
+	if (drm_dp_has_quirk(&intel_dp->desc, DP_DPCD_QUIRK_NO_PSR)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR support not currently available for this panel\n");
+		return;
+	}
+
+	if (!(intel_dp->edp_dpcd[1] & DP_EDP_SET_POWER_CAP)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Panel lacks power state control, PSR cannot be enabled\n");
+		return;
+	}
+
+	intel_dp->psr.sink_support = true;
+	intel_dp->psr.sink_sync_latency =
+		intel_dp_get_sink_sync_latency(intel_dp);
+
+	if (DISPLAY_VER(dev_priv) >= 9 &&
+	    (intel_dp->psr_dpcd[0] == DP_PSR2_WITH_Y_COORD_IS_SUPPORTED)) {
+		bool y_req = intel_dp->psr_dpcd[1] &
+			     DP_PSR2_SU_Y_COORDINATE_REQUIRED;
+		bool alpm = intel_dp_get_alpm_status(intel_dp);
+
+		/*
+		 * All panels that supports PSR version 03h (PSR2 +
+		 * Y-coordinate) can handle Y-coordinates in VSC but we are
+		 * only sure that it is going to be used when required by the
+		 * panel. This way panel is capable to do selective update
+		 * without a aux frame sync.
+		 *
+		 * To support PSR version 02h and PSR version 03h without
+		 * Y-coordinate requirement panels we would need to enable
+		 * GTC first.
+		 */
+		intel_dp->psr.sink_psr2_support = y_req && alpm;
+		drm_dbg_kms(&dev_priv->drm, "PSR2 %ssupported\n",
+			    intel_dp->psr.sink_psr2_support ? "" : "not ");
+
+		if (intel_dp->psr.sink_psr2_support) {
+			intel_dp->psr.colorimetry_support =
+				intel_dp_get_colorimetry_status(intel_dp);
+			intel_dp_get_su_granularity(intel_dp);
+		}
+	}
+}
+
+static void intel_psr_enable_sink(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u8 dpcd_val = DP_PSR_ENABLE;
+
+	/* Enable ALPM at sink for psr2 */
+	if (intel_dp->psr.psr2_enabled) {
+		drm_dp_dpcd_writeb(&intel_dp->aux, DP_RECEIVER_ALPM_CONFIG,
+				   DP_ALPM_ENABLE |
+				   DP_ALPM_LOCK_ERROR_IRQ_HPD_ENABLE);
+
+		dpcd_val |= DP_PSR_ENABLE_PSR2 | DP_PSR_IRQ_HPD_WITH_CRC_ERRORS;
+	} else {
+		if (intel_dp->psr.link_standby)
+			dpcd_val |= DP_PSR_MAIN_LINK_ACTIVE;
+
+		if (DISPLAY_VER(dev_priv) >= 8)
+			dpcd_val |= DP_PSR_CRC_VERIFICATION;
+	}
+
+	if (intel_dp->psr.req_psr2_sdp_prior_scanline)
+		dpcd_val |= DP_PSR_SU_REGION_SCANLINE_CAPTURE;
+
+	drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, dpcd_val);
+
+	drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER, DP_SET_POWER_D0);
+}
+
+static u32 intel_psr1_get_tp_time(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 val = 0;
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		val |= EDP_PSR_TP4_TIME_0US;
+
+	if (dev_priv->params.psr_safest_params) {
+		val |= EDP_PSR_TP1_TIME_2500us;
+		val |= EDP_PSR_TP2_TP3_TIME_2500us;
+		goto check_tp3_sel;
+	}
+
+	if (connector->panel.vbt.psr.tp1_wakeup_time_us == 0)
+		val |= EDP_PSR_TP1_TIME_0us;
+	else if (connector->panel.vbt.psr.tp1_wakeup_time_us <= 100)
+		val |= EDP_PSR_TP1_TIME_100us;
+	else if (connector->panel.vbt.psr.tp1_wakeup_time_us <= 500)
+		val |= EDP_PSR_TP1_TIME_500us;
+	else
+		val |= EDP_PSR_TP1_TIME_2500us;
+
+	if (connector->panel.vbt.psr.tp2_tp3_wakeup_time_us == 0)
+		val |= EDP_PSR_TP2_TP3_TIME_0us;
+	else if (connector->panel.vbt.psr.tp2_tp3_wakeup_time_us <= 100)
+		val |= EDP_PSR_TP2_TP3_TIME_100us;
+	else if (connector->panel.vbt.psr.tp2_tp3_wakeup_time_us <= 500)
+		val |= EDP_PSR_TP2_TP3_TIME_500us;
+	else
+		val |= EDP_PSR_TP2_TP3_TIME_2500us;
+
+check_tp3_sel:
+	if (intel_dp_source_supports_tps3(dev_priv) &&
+	    drm_dp_tps3_supported(intel_dp->dpcd))
+		val |= EDP_PSR_TP1_TP3_SEL;
+	else
+		val |= EDP_PSR_TP1_TP2_SEL;
+
+	return val;
+}
+
+static u8 psr_compute_idle_frames(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	int idle_frames;
+
+	/* Let's use 6 as the minimum to cover all known cases including the
+	 * off-by-one issue that HW has in some cases.
+	 */
+	idle_frames = max(6, connector->panel.vbt.psr.idle_frames);
+	idle_frames = max(idle_frames, intel_dp->psr.sink_sync_latency + 1);
+
+	if (drm_WARN_ON(&dev_priv->drm, idle_frames > 0xf))
+		idle_frames = 0xf;
+
+	return idle_frames;
+}
+
+static void hsw_activate_psr1(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 max_sleep_time = 0x1f;
+	u32 val = EDP_PSR_ENABLE;
+
+	val |= psr_compute_idle_frames(intel_dp) << EDP_PSR_IDLE_FRAME_SHIFT;
+
+	val |= max_sleep_time << EDP_PSR_MAX_SLEEP_TIME_SHIFT;
+	if (IS_HASWELL(dev_priv))
+		val |= EDP_PSR_MIN_LINK_ENTRY_TIME_8_LINES;
+
+	if (intel_dp->psr.link_standby)
+		val |= EDP_PSR_LINK_STANDBY;
+
+	val |= intel_psr1_get_tp_time(intel_dp);
+
+	if (DISPLAY_VER(dev_priv) >= 8)
+		val |= EDP_PSR_CRC_ENABLE;
+
+	val |= (intel_de_read(dev_priv, EDP_PSR_CTL(intel_dp->psr.transcoder)) &
+		EDP_PSR_RESTORE_PSR_ACTIVE_CTX_MASK);
+	intel_de_write(dev_priv, EDP_PSR_CTL(intel_dp->psr.transcoder), val);
+}
+
+static u32 intel_psr2_get_tp_time(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 val = 0;
+
+	if (dev_priv->params.psr_safest_params)
+		return EDP_PSR2_TP2_TIME_2500us;
+
+	if (connector->panel.vbt.psr.psr2_tp2_tp3_wakeup_time_us >= 0 &&
+	    connector->panel.vbt.psr.psr2_tp2_tp3_wakeup_time_us <= 50)
+		val |= EDP_PSR2_TP2_TIME_50us;
+	else if (connector->panel.vbt.psr.psr2_tp2_tp3_wakeup_time_us <= 100)
+		val |= EDP_PSR2_TP2_TIME_100us;
+	else if (connector->panel.vbt.psr.psr2_tp2_tp3_wakeup_time_us <= 500)
+		val |= EDP_PSR2_TP2_TIME_500us;
+	else
+		val |= EDP_PSR2_TP2_TIME_2500us;
+
+	return val;
+}
+
+static void hsw_activate_psr2(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 val = EDP_PSR2_ENABLE;
+
+	val |= psr_compute_idle_frames(intel_dp) << EDP_PSR2_IDLE_FRAME_SHIFT;
+
+	if (!IS_ALDERLAKE_P(dev_priv))
+		val |= EDP_SU_TRACK_ENABLE;
+
+	if (DISPLAY_VER(dev_priv) >= 10 && DISPLAY_VER(dev_priv) <= 12)
+		val |= EDP_Y_COORDINATE_ENABLE;
+
+	val |= EDP_PSR2_FRAME_BEFORE_SU(max_t(u8, intel_dp->psr.sink_sync_latency + 1, 2));
+	val |= intel_psr2_get_tp_time(intel_dp);
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		if (intel_dp->psr.io_wake_lines < 9 &&
+		    intel_dp->psr.fast_wake_lines < 9)
+			val |= TGL_EDP_PSR2_BLOCK_COUNT_NUM_2;
+		else
+			val |= TGL_EDP_PSR2_BLOCK_COUNT_NUM_3;
+	}
+
+	/* Wa_22012278275:adl-p */
+	if (IS_ADLP_DISPLAY_STEP(dev_priv, STEP_A0, STEP_E0)) {
+		static const u8 map[] = {
+			2, /* 5 lines */
+			1, /* 6 lines */
+			0, /* 7 lines */
+			3, /* 8 lines */
+			6, /* 9 lines */
+			5, /* 10 lines */
+			4, /* 11 lines */
+			7, /* 12 lines */
+		};
+		/*
+		 * Still using the default IO_BUFFER_WAKE and FAST_WAKE, see
+		 * comments bellow for more information
+		 */
+		u32 tmp;
+
+		tmp = map[intel_dp->psr.io_wake_lines - TGL_EDP_PSR2_IO_BUFFER_WAKE_MIN_LINES];
+		tmp = tmp << TGL_EDP_PSR2_IO_BUFFER_WAKE_SHIFT;
+		val |= tmp;
+
+		tmp = map[intel_dp->psr.fast_wake_lines - TGL_EDP_PSR2_FAST_WAKE_MIN_LINES];
+		tmp = tmp << TGL_EDP_PSR2_FAST_WAKE_MIN_SHIFT;
+		val |= tmp;
+	} else if (DISPLAY_VER(dev_priv) >= 12) {
+		val |= TGL_EDP_PSR2_IO_BUFFER_WAKE(intel_dp->psr.io_wake_lines);
+		val |= TGL_EDP_PSR2_FAST_WAKE(intel_dp->psr.fast_wake_lines);
+	} else if (DISPLAY_VER(dev_priv) >= 9) {
+		val |= EDP_PSR2_IO_BUFFER_WAKE(intel_dp->psr.io_wake_lines);
+		val |= EDP_PSR2_FAST_WAKE(intel_dp->psr.fast_wake_lines);
+	}
+
+	if (intel_dp->psr.req_psr2_sdp_prior_scanline)
+		val |= EDP_PSR2_SU_SDP_SCANLINE;
+
+	if (intel_dp->psr.psr2_sel_fetch_enabled) {
+		u32 tmp;
+
+		/* Wa_1408330847 */
+		if (IS_TGL_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0))
+			intel_de_rmw(dev_priv, CHICKEN_PAR1_1,
+				     DIS_RAM_BYPASS_PSR2_MAN_TRACK,
+				     DIS_RAM_BYPASS_PSR2_MAN_TRACK);
+
+		tmp = intel_de_read(dev_priv, PSR2_MAN_TRK_CTL(intel_dp->psr.transcoder));
+		drm_WARN_ON(&dev_priv->drm, !(tmp & PSR2_MAN_TRK_CTL_ENABLE));
+	} else if (HAS_PSR2_SEL_FETCH(dev_priv)) {
+		intel_de_write(dev_priv,
+			       PSR2_MAN_TRK_CTL(intel_dp->psr.transcoder), 0);
+	}
+
+	/*
+	 * PSR2 HW is incorrectly using EDP_PSR_TP1_TP3_SEL and BSpec is
+	 * recommending keep this bit unset while PSR2 is enabled.
+	 */
+	intel_de_write(dev_priv, EDP_PSR_CTL(intel_dp->psr.transcoder), 0);
+
+	intel_de_write(dev_priv, EDP_PSR2_CTL(intel_dp->psr.transcoder), val);
+}
+
+static bool
+transcoder_has_psr2(struct drm_i915_private *dev_priv, enum transcoder trans)
+{
+	if (IS_ALDERLAKE_P(dev_priv))
+		return trans == TRANSCODER_A || trans == TRANSCODER_B;
+	else if (DISPLAY_VER(dev_priv) >= 12)
+		return trans == TRANSCODER_A;
+	else
+		return trans == TRANSCODER_EDP;
+}
+
+static u32 intel_get_frame_time_us(const struct intel_crtc_state *cstate)
+{
+	if (!cstate || !cstate->hw.active)
+		return 0;
+
+	return DIV_ROUND_UP(1000 * 1000,
+			    drm_mode_vrefresh(&cstate->hw.adjusted_mode));
+}
+
+static void psr2_program_idle_frames(struct intel_dp *intel_dp,
+				     u32 idle_frames)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 val;
+
+	idle_frames <<=  EDP_PSR2_IDLE_FRAME_SHIFT;
+	val = intel_de_read(dev_priv, EDP_PSR2_CTL(intel_dp->psr.transcoder));
+	val &= ~EDP_PSR2_IDLE_FRAME_MASK;
+	val |= idle_frames;
+	intel_de_write(dev_priv, EDP_PSR2_CTL(intel_dp->psr.transcoder), val);
+}
+
+static void tgl_psr2_enable_dc3co(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	psr2_program_idle_frames(intel_dp, 0);
+	intel_display_power_set_target_dc_state(dev_priv, DC_STATE_EN_DC3CO);
+}
+
+static void tgl_psr2_disable_dc3co(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	intel_display_power_set_target_dc_state(dev_priv, DC_STATE_EN_UPTO_DC6);
+	psr2_program_idle_frames(intel_dp, psr_compute_idle_frames(intel_dp));
+}
+
+static void tgl_dc3co_disable_work(struct work_struct *work)
+{
+	struct intel_dp *intel_dp =
+		container_of(work, typeof(*intel_dp), psr.dc3co_work.work);
+
+	mutex_lock(&intel_dp->psr.lock);
+	/* If delayed work is pending, it is not idle */
+	if (delayed_work_pending(&intel_dp->psr.dc3co_work))
+		goto unlock;
+
+	tgl_psr2_disable_dc3co(intel_dp);
+unlock:
+	mutex_unlock(&intel_dp->psr.lock);
+}
+
+static void tgl_disallow_dc3co_on_psr2_exit(struct intel_dp *intel_dp)
+{
+	if (!intel_dp->psr.dc3co_exitline)
+		return;
+
+	cancel_delayed_work(&intel_dp->psr.dc3co_work);
+	/* Before PSR2 exit disallow dc3co*/
+	tgl_psr2_disable_dc3co(intel_dp);
+}
+
+static bool
+dc3co_is_pipe_port_compatible(struct intel_dp *intel_dp,
+			      struct intel_crtc_state *crtc_state)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	enum pipe pipe = to_intel_crtc(crtc_state->uapi.crtc)->pipe;
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	enum port port = dig_port->base.port;
+
+	if (IS_ALDERLAKE_P(dev_priv))
+		return pipe <= PIPE_B && port <= PORT_B;
+	else
+		return pipe == PIPE_A && port == PORT_A;
+}
+
+static void
+tgl_dc3co_exitline_compute_config(struct intel_dp *intel_dp,
+				  struct intel_crtc_state *crtc_state)
+{
+	const u32 crtc_vdisplay = crtc_state->uapi.adjusted_mode.crtc_vdisplay;
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 exit_scanlines;
+
+	/*
+	 * FIXME: Due to the changed sequence of activating/deactivating DC3CO,
+	 * disable DC3CO until the changed dc3co activating/deactivating sequence
+	 * is applied. B.Specs:49196
+	 */
+	return;
+
+	/*
+	 * DMC's DC3CO exit mechanism has an issue with Selective Fecth
+	 * TODO: when the issue is addressed, this restriction should be removed.
+	 */
+	if (crtc_state->enable_psr2_sel_fetch)
+		return;
+
+	if (!(dev_priv->display.dmc.allowed_dc_mask & DC_STATE_EN_DC3CO))
+		return;
+
+	if (!dc3co_is_pipe_port_compatible(intel_dp, crtc_state))
+		return;
+
+	/* Wa_16011303918:adl-p */
+	if (IS_ADLP_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0))
+		return;
+
+	/*
+	 * DC3CO Exit time 200us B.Spec 49196
+	 * PSR2 transcoder Early Exit scanlines = ROUNDUP(200 / line time) + 1
+	 */
+	exit_scanlines =
+		intel_usecs_to_scanlines(&crtc_state->uapi.adjusted_mode, 200) + 1;
+
+	if (drm_WARN_ON(&dev_priv->drm, exit_scanlines > crtc_vdisplay))
+		return;
+
+	crtc_state->dc3co_exitline = crtc_vdisplay - exit_scanlines;
+}
+
+static bool intel_psr2_sel_fetch_config_valid(struct intel_dp *intel_dp,
+					      struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	if (!dev_priv->params.enable_psr2_sel_fetch &&
+	    intel_dp->psr.debug != I915_PSR_DEBUG_ENABLE_SEL_FETCH) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 sel fetch not enabled, disabled by parameter\n");
+		return false;
+	}
+
+	if (crtc_state->uapi.async_flip) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 sel fetch not enabled, async flip enabled\n");
+		return false;
+	}
+
+	/* Wa_14010254185 Wa_14010103792 */
+	if (IS_TGL_DISPLAY_STEP(dev_priv, STEP_A0, STEP_C0)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 sel fetch not enabled, missing the implementation of WAs\n");
+		return false;
+	}
+
+	return crtc_state->enable_psr2_sel_fetch = true;
+}
+
+static bool psr2_granularity_check(struct intel_dp *intel_dp,
+				   struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	const int crtc_hdisplay = crtc_state->hw.adjusted_mode.crtc_hdisplay;
+	const int crtc_vdisplay = crtc_state->hw.adjusted_mode.crtc_vdisplay;
+	u16 y_granularity = 0;
+
+	/* PSR2 HW only send full lines so we only need to validate the width */
+	if (crtc_hdisplay % intel_dp->psr.su_w_granularity)
+		return false;
+
+	if (crtc_vdisplay % intel_dp->psr.su_y_granularity)
+		return false;
+
+	/* HW tracking is only aligned to 4 lines */
+	if (!crtc_state->enable_psr2_sel_fetch)
+		return intel_dp->psr.su_y_granularity == 4;
+
+	/*
+	 * adl_p has 1 line granularity. For other platforms with SW tracking we
+	 * can adjust the y coordinates to match sink requirement if multiple of
+	 * 4.
+	 */
+	if (IS_ALDERLAKE_P(dev_priv))
+		y_granularity = intel_dp->psr.su_y_granularity;
+	else if (intel_dp->psr.su_y_granularity <= 2)
+		y_granularity = 4;
+	else if ((intel_dp->psr.su_y_granularity % 4) == 0)
+		y_granularity = intel_dp->psr.su_y_granularity;
+
+	if (y_granularity == 0 || crtc_vdisplay % y_granularity)
+		return false;
+
+	crtc_state->su_y_granularity = y_granularity;
+	return true;
+}
+
+static bool _compute_psr2_sdp_prior_scanline_indication(struct intel_dp *intel_dp,
+							struct intel_crtc_state *crtc_state)
+{
+	const struct drm_display_mode *adjusted_mode = &crtc_state->uapi.adjusted_mode;
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 hblank_total, hblank_ns, req_ns;
+
+	hblank_total = adjusted_mode->crtc_hblank_end - adjusted_mode->crtc_hblank_start;
+	hblank_ns = div_u64(1000000ULL * hblank_total, adjusted_mode->crtc_clock);
+
+	/* From spec: ((60 / number of lanes) + 11) * 1000 / symbol clock frequency MHz */
+	req_ns = ((60 / crtc_state->lane_count) + 11) * 1000 / (crtc_state->port_clock / 1000);
+
+	if ((hblank_ns - req_ns) > 100)
+		return true;
+
+	/* Not supported <13 / Wa_22012279113:adl-p */
+	if (DISPLAY_VER(dev_priv) <= 13 || intel_dp->edp_dpcd[0] < DP_EDP_14b)
+		return false;
+
+	crtc_state->req_psr2_sdp_prior_scanline = true;
+	return true;
+}
+
+static bool _compute_psr2_wake_times(struct intel_dp *intel_dp,
+				     struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = dp_to_i915(intel_dp);
+	int io_wake_lines, io_wake_time, fast_wake_lines, fast_wake_time;
+	u8 max_wake_lines;
+
+	if (DISPLAY_VER(i915) >= 12) {
+		io_wake_time = 42;
+		/*
+		 * According to Bspec it's 42us, but based on testing
+		 * it is not enough -> use 45 us.
+		 */
+		fast_wake_time = 45;
+		max_wake_lines = 12;
+	} else {
+		io_wake_time = 50;
+		fast_wake_time = 32;
+		max_wake_lines = 8;
+	}
+
+	io_wake_lines = intel_usecs_to_scanlines(
+		&crtc_state->hw.adjusted_mode, io_wake_time);
+	fast_wake_lines = intel_usecs_to_scanlines(
+		&crtc_state->hw.adjusted_mode, fast_wake_time);
+
+	if (io_wake_lines > max_wake_lines ||
+	    fast_wake_lines > max_wake_lines)
+		return false;
+
+	if (i915->params.psr_safest_params)
+		io_wake_lines = fast_wake_lines = max_wake_lines;
+
+	/* According to Bspec lower limit should be set as 7 lines. */
+	intel_dp->psr.io_wake_lines = max(io_wake_lines, 7);
+	intel_dp->psr.fast_wake_lines = max(fast_wake_lines, 7);
+
+	return true;
+}
+
+static bool intel_psr2_config_valid(struct intel_dp *intel_dp,
+				    struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	int crtc_hdisplay = crtc_state->hw.adjusted_mode.crtc_hdisplay;
+	int crtc_vdisplay = crtc_state->hw.adjusted_mode.crtc_vdisplay;
+	int psr_max_h = 0, psr_max_v = 0, max_bpp = 0;
+
+	if (!intel_dp->psr.sink_psr2_support)
+		return false;
+
+	/* JSL and EHL only supports eDP 1.3 */
+	if (IS_JSL_EHL(dev_priv)) {
+		drm_dbg_kms(&dev_priv->drm, "PSR2 not supported by phy\n");
+		return false;
+	}
+
+	/* Wa_16011181250 */
+	if (IS_ROCKETLAKE(dev_priv) || IS_ALDERLAKE_S(dev_priv) ||
+	    IS_DG2(dev_priv)) {
+		drm_dbg_kms(&dev_priv->drm, "PSR2 is defeatured for this platform\n");
+		return false;
+	}
+
+	if (IS_ADLP_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0)) {
+		drm_dbg_kms(&dev_priv->drm, "PSR2 not completely functional in this stepping\n");
+		return false;
+	}
+
+	if (!transcoder_has_psr2(dev_priv, crtc_state->cpu_transcoder)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 not supported in transcoder %s\n",
+			    transcoder_name(crtc_state->cpu_transcoder));
+		return false;
+	}
+
+	if (!psr2_global_enabled(intel_dp)) {
+		drm_dbg_kms(&dev_priv->drm, "PSR2 disabled by flag\n");
+		return false;
+	}
+
+	/*
+	 * DSC and PSR2 cannot be enabled simultaneously. If a requested
+	 * resolution requires DSC to be enabled, priority is given to DSC
+	 * over PSR2.
+	 */
+	if (crtc_state->dsc.compression_enable) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 cannot be enabled since DSC is enabled\n");
+		return false;
+	}
+
+	if (crtc_state->crc_enabled) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 not enabled because it would inhibit pipe CRC calculation\n");
+		return false;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		psr_max_h = 5120;
+		psr_max_v = 3200;
+		max_bpp = 30;
+	} else if (DISPLAY_VER(dev_priv) >= 10) {
+		psr_max_h = 4096;
+		psr_max_v = 2304;
+		max_bpp = 24;
+	} else if (DISPLAY_VER(dev_priv) == 9) {
+		psr_max_h = 3640;
+		psr_max_v = 2304;
+		max_bpp = 24;
+	}
+
+	if (crtc_state->pipe_bpp > max_bpp) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 not enabled, pipe bpp %d > max supported %d\n",
+			    crtc_state->pipe_bpp, max_bpp);
+		return false;
+	}
+
+	/* Wa_16011303918:adl-p */
+	if (crtc_state->vrr.enable &&
+	    IS_ADLP_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 not enabled, not compatible with HW stepping + VRR\n");
+		return false;
+	}
+
+	if (!_compute_psr2_sdp_prior_scanline_indication(intel_dp, crtc_state)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 not enabled, PSR2 SDP indication do not fit in hblank\n");
+		return false;
+	}
+
+	if (!_compute_psr2_wake_times(intel_dp, crtc_state)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 not enabled, Unable to use long enough wake times\n");
+		return false;
+	}
+
+	if (HAS_PSR2_SEL_FETCH(dev_priv)) {
+		if (!intel_psr2_sel_fetch_config_valid(intel_dp, crtc_state) &&
+		    !HAS_PSR_HW_TRACKING(dev_priv)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "PSR2 not enabled, selective fetch not valid and no HW tracking available\n");
+			return false;
+		}
+	}
+
+	/* Wa_2209313811 */
+	if (!crtc_state->enable_psr2_sel_fetch &&
+	    IS_TGL_DISPLAY_STEP(dev_priv, STEP_A0, STEP_C0)) {
+		drm_dbg_kms(&dev_priv->drm, "PSR2 HW tracking is not supported this Display stepping\n");
+		goto unsupported;
+	}
+
+	if (!psr2_granularity_check(intel_dp, crtc_state)) {
+		drm_dbg_kms(&dev_priv->drm, "PSR2 not enabled, SU granularity not compatible\n");
+		goto unsupported;
+	}
+
+	if (!crtc_state->enable_psr2_sel_fetch &&
+	    (crtc_hdisplay > psr_max_h || crtc_vdisplay > psr_max_v)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR2 not enabled, resolution %dx%d > max supported %dx%d\n",
+			    crtc_hdisplay, crtc_vdisplay,
+			    psr_max_h, psr_max_v);
+		goto unsupported;
+	}
+
+	tgl_dc3co_exitline_compute_config(intel_dp, crtc_state);
+	return true;
+
+unsupported:
+	crtc_state->enable_psr2_sel_fetch = false;
+	return false;
+}
+
+void intel_psr_compute_config(struct intel_dp *intel_dp,
+			      struct intel_crtc_state *crtc_state,
+			      struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	int psr_setup_time;
+
+	/*
+	 * Current PSR panels don't work reliably with VRR enabled
+	 * So if VRR is enabled, do not enable PSR.
+	 */
+	if (crtc_state->vrr.enable)
+		return;
+
+	if (!CAN_PSR(intel_dp))
+		return;
+
+	if (!psr_global_enabled(intel_dp)) {
+		drm_dbg_kms(&dev_priv->drm, "PSR disabled by flag\n");
+		return;
+	}
+
+	if (intel_dp->psr.sink_not_reliable) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR sink implementation is not reliable\n");
+		return;
+	}
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR condition failed: Interlaced mode enabled\n");
+		return;
+	}
+
+	psr_setup_time = drm_dp_psr_setup_time(intel_dp->psr_dpcd);
+	if (psr_setup_time < 0) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR condition failed: Invalid PSR setup time (0x%02x)\n",
+			    intel_dp->psr_dpcd[1]);
+		return;
+	}
+
+	if (intel_usecs_to_scanlines(adjusted_mode, psr_setup_time) >
+	    adjusted_mode->crtc_vtotal - adjusted_mode->crtc_vdisplay - 1) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR condition failed: PSR setup time (%d us) too long\n",
+			    psr_setup_time);
+		return;
+	}
+
+	crtc_state->has_psr = true;
+	crtc_state->has_psr2 = intel_psr2_config_valid(intel_dp, crtc_state);
+
+	crtc_state->infoframes.enable |= intel_hdmi_infoframe_enable(DP_SDP_VSC);
+	intel_dp_compute_psr_vsc_sdp(intel_dp, crtc_state, conn_state,
+				     &crtc_state->psr_vsc);
+}
+
+void intel_psr_get_config(struct intel_encoder *encoder,
+			  struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	struct intel_dp *intel_dp;
+	u32 val;
+
+	if (!dig_port)
+		return;
+
+	intel_dp = &dig_port->dp;
+	if (!CAN_PSR(intel_dp))
+		return;
+
+	mutex_lock(&intel_dp->psr.lock);
+	if (!intel_dp->psr.enabled)
+		goto unlock;
+
+	/*
+	 * Not possible to read EDP_PSR/PSR2_CTL registers as it is
+	 * enabled/disabled because of frontbuffer tracking and others.
+	 */
+	pipe_config->has_psr = true;
+	pipe_config->has_psr2 = intel_dp->psr.psr2_enabled;
+	pipe_config->infoframes.enable |= intel_hdmi_infoframe_enable(DP_SDP_VSC);
+
+	if (!intel_dp->psr.psr2_enabled)
+		goto unlock;
+
+	if (HAS_PSR2_SEL_FETCH(dev_priv)) {
+		val = intel_de_read(dev_priv, PSR2_MAN_TRK_CTL(intel_dp->psr.transcoder));
+		if (val & PSR2_MAN_TRK_CTL_ENABLE)
+			pipe_config->enable_psr2_sel_fetch = true;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		val = intel_de_read(dev_priv, EXITLINE(intel_dp->psr.transcoder));
+		val &= EXITLINE_MASK;
+		pipe_config->dc3co_exitline = val;
+	}
+unlock:
+	mutex_unlock(&intel_dp->psr.lock);
+}
+
+static void intel_psr_activate(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	enum transcoder transcoder = intel_dp->psr.transcoder;
+
+	if (transcoder_has_psr2(dev_priv, transcoder))
+		drm_WARN_ON(&dev_priv->drm,
+			    intel_de_read(dev_priv, EDP_PSR2_CTL(transcoder)) & EDP_PSR2_ENABLE);
+
+	drm_WARN_ON(&dev_priv->drm,
+		    intel_de_read(dev_priv, EDP_PSR_CTL(transcoder)) & EDP_PSR_ENABLE);
+	drm_WARN_ON(&dev_priv->drm, intel_dp->psr.active);
+	lockdep_assert_held(&intel_dp->psr.lock);
+
+	/* psr1 and psr2 are mutually exclusive.*/
+	if (intel_dp->psr.psr2_enabled)
+		hsw_activate_psr2(intel_dp);
+	else
+		hsw_activate_psr1(intel_dp);
+
+	intel_dp->psr.active = true;
+}
+
+static u32 wa_16013835468_bit_get(struct intel_dp *intel_dp)
+{
+	switch (intel_dp->psr.pipe) {
+	case PIPE_A:
+		return LATENCY_REPORTING_REMOVED_PIPE_A;
+	case PIPE_B:
+		return LATENCY_REPORTING_REMOVED_PIPE_B;
+	case PIPE_C:
+		return LATENCY_REPORTING_REMOVED_PIPE_C;
+	default:
+		MISSING_CASE(intel_dp->psr.pipe);
+		return 0;
+	}
+}
+
+static void intel_psr_enable_source(struct intel_dp *intel_dp,
+				    const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	enum transcoder cpu_transcoder = intel_dp->psr.transcoder;
+	u32 mask;
+
+	/*
+	 * Per Spec: Avoid continuous PSR exit by masking MEMUP and HPD also
+	 * mask LPSP to avoid dependency on other drivers that might block
+	 * runtime_pm besides preventing  other hw tracking issues now we
+	 * can rely on frontbuffer tracking.
+	 */
+	mask = EDP_PSR_DEBUG_MASK_MEMUP |
+	       EDP_PSR_DEBUG_MASK_HPD |
+	       EDP_PSR_DEBUG_MASK_LPSP |
+	       EDP_PSR_DEBUG_MASK_MAX_SLEEP;
+
+	if (DISPLAY_VER(dev_priv) < 11)
+		mask |= EDP_PSR_DEBUG_MASK_DISP_REG_WRITE;
+
+	intel_de_write(dev_priv, EDP_PSR_DEBUG(intel_dp->psr.transcoder),
+		       mask);
+
+	psr_irq_control(intel_dp);
+
+	if (intel_dp->psr.dc3co_exitline) {
+		u32 val;
+
+		/*
+		 * TODO: if future platforms supports DC3CO in more than one
+		 * transcoder, EXITLINE will need to be unset when disabling PSR
+		 */
+		val = intel_de_read(dev_priv, EXITLINE(cpu_transcoder));
+		val &= ~EXITLINE_MASK;
+		val |= intel_dp->psr.dc3co_exitline << EXITLINE_SHIFT;
+		val |= EXITLINE_ENABLE;
+		intel_de_write(dev_priv, EXITLINE(cpu_transcoder), val);
+	}
+
+	if (HAS_PSR_HW_TRACKING(dev_priv) && HAS_PSR2_SEL_FETCH(dev_priv))
+		intel_de_rmw(dev_priv, CHICKEN_PAR1_1, IGNORE_PSR2_HW_TRACKING,
+			     intel_dp->psr.psr2_sel_fetch_enabled ?
+			     IGNORE_PSR2_HW_TRACKING : 0);
+
+	if (intel_dp->psr.psr2_enabled) {
+		if (DISPLAY_VER(dev_priv) == 9)
+			intel_de_rmw(dev_priv, CHICKEN_TRANS(cpu_transcoder), 0,
+				     PSR2_VSC_ENABLE_PROG_HEADER |
+				     PSR2_ADD_VERTICAL_LINE_COUNT);
+
+		/*
+		 * Wa_16014451276:adlp
+		 * All supported adlp panels have 1-based X granularity, this may
+		 * cause issues if non-supported panels are used.
+		 */
+		if (IS_ALDERLAKE_P(dev_priv))
+			intel_de_rmw(dev_priv, CHICKEN_TRANS(cpu_transcoder), 0,
+				     ADLP_1_BASED_X_GRANULARITY);
+
+		/* Wa_16011168373:adl-p */
+		if (IS_ADLP_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0))
+			intel_de_rmw(dev_priv,
+				     TRANS_SET_CONTEXT_LATENCY(intel_dp->psr.transcoder),
+				     TRANS_SET_CONTEXT_LATENCY_MASK,
+				     TRANS_SET_CONTEXT_LATENCY_VALUE(1));
+
+		/* Wa_16012604467:adlp */
+		if (IS_ALDERLAKE_P(dev_priv))
+			intel_de_rmw(dev_priv, CLKGATE_DIS_MISC, 0,
+				     CLKGATE_DIS_MISC_DMASC_GATING_DIS);
+
+		/* Wa_16013835468:tgl[b0+], dg1 */
+		if (IS_TGL_DISPLAY_STEP(dev_priv, STEP_B0, STEP_FOREVER) ||
+		    IS_DG1(dev_priv)) {
+			u16 vtotal, vblank;
+
+			vtotal = crtc_state->uapi.adjusted_mode.crtc_vtotal -
+				 crtc_state->uapi.adjusted_mode.crtc_vdisplay;
+			vblank = crtc_state->uapi.adjusted_mode.crtc_vblank_end -
+				 crtc_state->uapi.adjusted_mode.crtc_vblank_start;
+			if (vblank > vtotal)
+				intel_de_rmw(dev_priv, GEN8_CHICKEN_DCPR_1, 0,
+					     wa_16013835468_bit_get(intel_dp));
+		}
+	}
+}
+
+static bool psr_interrupt_error_check(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 val;
+
+	/*
+	 * If a PSR error happened and the driver is reloaded, the EDP_PSR_IIR
+	 * will still keep the error set even after the reset done in the
+	 * irq_preinstall and irq_uninstall hooks.
+	 * And enabling in this situation cause the screen to freeze in the
+	 * first time that PSR HW tries to activate so lets keep PSR disabled
+	 * to avoid any rendering problems.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 12)
+		val = intel_de_read(dev_priv,
+				    TRANS_PSR_IIR(intel_dp->psr.transcoder));
+	else
+		val = intel_de_read(dev_priv, EDP_PSR_IIR);
+	val &= psr_irq_psr_error_bit_get(intel_dp);
+	if (val) {
+		intel_dp->psr.sink_not_reliable = true;
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR interruption error set, not enabling PSR\n");
+		return false;
+	}
+
+	return true;
+}
+
+static void intel_psr_enable_locked(struct intel_dp *intel_dp,
+				    const struct intel_crtc_state *crtc_state)
+{
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	enum phy phy = intel_port_to_phy(dev_priv, dig_port->base.port);
+	struct intel_encoder *encoder = &dig_port->base;
+	u32 val;
+
+	drm_WARN_ON(&dev_priv->drm, intel_dp->psr.enabled);
+
+	intel_dp->psr.psr2_enabled = crtc_state->has_psr2;
+	intel_dp->psr.busy_frontbuffer_bits = 0;
+	intel_dp->psr.pipe = to_intel_crtc(crtc_state->uapi.crtc)->pipe;
+	intel_dp->psr.transcoder = crtc_state->cpu_transcoder;
+	/* DC5/DC6 requires at least 6 idle frames */
+	val = usecs_to_jiffies(intel_get_frame_time_us(crtc_state) * 6);
+	intel_dp->psr.dc3co_exit_delay = val;
+	intel_dp->psr.dc3co_exitline = crtc_state->dc3co_exitline;
+	intel_dp->psr.psr2_sel_fetch_enabled = crtc_state->enable_psr2_sel_fetch;
+	intel_dp->psr.psr2_sel_fetch_cff_enabled = false;
+	intel_dp->psr.req_psr2_sdp_prior_scanline =
+		crtc_state->req_psr2_sdp_prior_scanline;
+
+	if (!psr_interrupt_error_check(intel_dp))
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "Enabling PSR%s\n",
+		    intel_dp->psr.psr2_enabled ? "2" : "1");
+	intel_write_dp_vsc_sdp(encoder, crtc_state, &crtc_state->psr_vsc);
+	intel_snps_phy_update_psr_power_state(dev_priv, phy, true);
+	intel_psr_enable_sink(intel_dp);
+	intel_psr_enable_source(intel_dp, crtc_state);
+	intel_dp->psr.enabled = true;
+	intel_dp->psr.paused = false;
+
+	intel_psr_activate(intel_dp);
+}
+
+static void intel_psr_exit(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	u32 val;
+
+	if (!intel_dp->psr.active) {
+		if (transcoder_has_psr2(dev_priv, intel_dp->psr.transcoder)) {
+			val = intel_de_read(dev_priv,
+					    EDP_PSR2_CTL(intel_dp->psr.transcoder));
+			drm_WARN_ON(&dev_priv->drm, val & EDP_PSR2_ENABLE);
+		}
+
+		val = intel_de_read(dev_priv,
+				    EDP_PSR_CTL(intel_dp->psr.transcoder));
+		drm_WARN_ON(&dev_priv->drm, val & EDP_PSR_ENABLE);
+
+		return;
+	}
+
+	if (intel_dp->psr.psr2_enabled) {
+		tgl_disallow_dc3co_on_psr2_exit(intel_dp);
+		val = intel_de_read(dev_priv,
+				    EDP_PSR2_CTL(intel_dp->psr.transcoder));
+		drm_WARN_ON(&dev_priv->drm, !(val & EDP_PSR2_ENABLE));
+		val &= ~EDP_PSR2_ENABLE;
+		intel_de_write(dev_priv,
+			       EDP_PSR2_CTL(intel_dp->psr.transcoder), val);
+	} else {
+		val = intel_de_read(dev_priv,
+				    EDP_PSR_CTL(intel_dp->psr.transcoder));
+		drm_WARN_ON(&dev_priv->drm, !(val & EDP_PSR_ENABLE));
+		val &= ~EDP_PSR_ENABLE;
+		intel_de_write(dev_priv,
+			       EDP_PSR_CTL(intel_dp->psr.transcoder), val);
+	}
+	intel_dp->psr.active = false;
+}
+
+static void intel_psr_wait_exit_locked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	i915_reg_t psr_status;
+	u32 psr_status_mask;
+
+	if (intel_dp->psr.psr2_enabled) {
+		psr_status = EDP_PSR2_STATUS(intel_dp->psr.transcoder);
+		psr_status_mask = EDP_PSR2_STATUS_STATE_MASK;
+	} else {
+		psr_status = EDP_PSR_STATUS(intel_dp->psr.transcoder);
+		psr_status_mask = EDP_PSR_STATUS_STATE_MASK;
+	}
+
+	/* Wait till PSR is idle */
+	if (intel_de_wait_for_clear(dev_priv, psr_status,
+				    psr_status_mask, 2000))
+		drm_err(&dev_priv->drm, "Timed out waiting PSR idle state\n");
+}
+
+static void intel_psr_disable_locked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	enum phy phy = intel_port_to_phy(dev_priv,
+					 dp_to_dig_port(intel_dp)->base.port);
+
+	lockdep_assert_held(&intel_dp->psr.lock);
+
+	if (!intel_dp->psr.enabled)
+		return;
+
+	drm_dbg_kms(&dev_priv->drm, "Disabling PSR%s\n",
+		    intel_dp->psr.psr2_enabled ? "2" : "1");
+
+	intel_psr_exit(intel_dp);
+	intel_psr_wait_exit_locked(intel_dp);
+
+	/* Wa_1408330847 */
+	if (intel_dp->psr.psr2_sel_fetch_enabled &&
+	    IS_TGL_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0))
+		intel_de_rmw(dev_priv, CHICKEN_PAR1_1,
+			     DIS_RAM_BYPASS_PSR2_MAN_TRACK, 0);
+
+	if (intel_dp->psr.psr2_enabled) {
+		/* Wa_16011168373:adl-p */
+		if (IS_ADLP_DISPLAY_STEP(dev_priv, STEP_A0, STEP_B0))
+			intel_de_rmw(dev_priv,
+				     TRANS_SET_CONTEXT_LATENCY(intel_dp->psr.transcoder),
+				     TRANS_SET_CONTEXT_LATENCY_MASK, 0);
+
+		/* Wa_16012604467:adlp */
+		if (IS_ALDERLAKE_P(dev_priv))
+			intel_de_rmw(dev_priv, CLKGATE_DIS_MISC,
+				     CLKGATE_DIS_MISC_DMASC_GATING_DIS, 0);
+
+		/* Wa_16013835468:tgl[b0+], dg1 */
+		if (IS_TGL_DISPLAY_STEP(dev_priv, STEP_B0, STEP_FOREVER) ||
+		    IS_DG1(dev_priv))
+			intel_de_rmw(dev_priv, GEN8_CHICKEN_DCPR_1,
+				     wa_16013835468_bit_get(intel_dp), 0);
+	}
+
+	intel_snps_phy_update_psr_power_state(dev_priv, phy, false);
+
+	/* Disable PSR on Sink */
+	drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, 0);
+
+	if (intel_dp->psr.psr2_enabled)
+		drm_dp_dpcd_writeb(&intel_dp->aux, DP_RECEIVER_ALPM_CONFIG, 0);
+
+	intel_dp->psr.enabled = false;
+	intel_dp->psr.psr2_enabled = false;
+	intel_dp->psr.psr2_sel_fetch_enabled = false;
+	intel_dp->psr.psr2_sel_fetch_cff_enabled = false;
+}
+
+/**
+ * intel_psr_disable - Disable PSR
+ * @intel_dp: Intel DP
+ * @old_crtc_state: old CRTC state
+ *
+ * This function needs to be called before disabling pipe.
+ */
+void intel_psr_disable(struct intel_dp *intel_dp,
+		       const struct intel_crtc_state *old_crtc_state)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	if (!old_crtc_state->has_psr)
+		return;
+
+	if (drm_WARN_ON(&dev_priv->drm, !CAN_PSR(intel_dp)))
+		return;
+
+	mutex_lock(&intel_dp->psr.lock);
+
+	intel_psr_disable_locked(intel_dp);
+
+	mutex_unlock(&intel_dp->psr.lock);
+	cancel_work_sync(&intel_dp->psr.work);
+	cancel_delayed_work_sync(&intel_dp->psr.dc3co_work);
+}
+
+/**
+ * intel_psr_pause - Pause PSR
+ * @intel_dp: Intel DP
+ *
+ * This function need to be called after enabling psr.
+ */
+void intel_psr_pause(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_psr *psr = &intel_dp->psr;
+
+	if (!CAN_PSR(intel_dp))
+		return;
+
+	mutex_lock(&psr->lock);
+
+	if (!psr->enabled) {
+		mutex_unlock(&psr->lock);
+		return;
+	}
+
+	/* If we ever hit this, we will need to add refcount to pause/resume */
+	drm_WARN_ON(&dev_priv->drm, psr->paused);
+
+	intel_psr_exit(intel_dp);
+	intel_psr_wait_exit_locked(intel_dp);
+	psr->paused = true;
+
+	mutex_unlock(&psr->lock);
+
+	cancel_work_sync(&psr->work);
+	cancel_delayed_work_sync(&psr->dc3co_work);
+}
+
+/**
+ * intel_psr_resume - Resume PSR
+ * @intel_dp: Intel DP
+ *
+ * This function need to be called after pausing psr.
+ */
+void intel_psr_resume(struct intel_dp *intel_dp)
+{
+	struct intel_psr *psr = &intel_dp->psr;
+
+	if (!CAN_PSR(intel_dp))
+		return;
+
+	mutex_lock(&psr->lock);
+
+	if (!psr->paused)
+		goto unlock;
+
+	psr->paused = false;
+	intel_psr_activate(intel_dp);
+
+unlock:
+	mutex_unlock(&psr->lock);
+}
+
+static u32 man_trk_ctl_enable_bit_get(struct drm_i915_private *dev_priv)
+{
+	return IS_ALDERLAKE_P(dev_priv) ? 0 : PSR2_MAN_TRK_CTL_ENABLE;
+}
+
+static u32 man_trk_ctl_single_full_frame_bit_get(struct drm_i915_private *dev_priv)
+{
+	return IS_ALDERLAKE_P(dev_priv) ?
+	       ADLP_PSR2_MAN_TRK_CTL_SF_SINGLE_FULL_FRAME :
+	       PSR2_MAN_TRK_CTL_SF_SINGLE_FULL_FRAME;
+}
+
+static u32 man_trk_ctl_partial_frame_bit_get(struct drm_i915_private *dev_priv)
+{
+	return IS_ALDERLAKE_P(dev_priv) ?
+	       ADLP_PSR2_MAN_TRK_CTL_SF_PARTIAL_FRAME_UPDATE :
+	       PSR2_MAN_TRK_CTL_SF_PARTIAL_FRAME_UPDATE;
+}
+
+static u32 man_trk_ctl_continuos_full_frame(struct drm_i915_private *dev_priv)
+{
+	return IS_ALDERLAKE_P(dev_priv) ?
+	       ADLP_PSR2_MAN_TRK_CTL_SF_CONTINUOS_FULL_FRAME :
+	       PSR2_MAN_TRK_CTL_SF_CONTINUOS_FULL_FRAME;
+}
+
+static void psr_force_hw_tracking_exit(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	if (intel_dp->psr.psr2_sel_fetch_enabled)
+		intel_de_write(dev_priv,
+			       PSR2_MAN_TRK_CTL(intel_dp->psr.transcoder),
+			       man_trk_ctl_enable_bit_get(dev_priv) |
+			       man_trk_ctl_partial_frame_bit_get(dev_priv) |
+			       man_trk_ctl_single_full_frame_bit_get(dev_priv));
+
+	/*
+	 * Display WA #0884: skl+
+	 * This documented WA for bxt can be safely applied
+	 * broadly so we can force HW tracking to exit PSR
+	 * instead of disabling and re-enabling.
+	 * Workaround tells us to write 0 to CUR_SURFLIVE_A,
+	 * but it makes more sense write to the current active
+	 * pipe.
+	 *
+	 * This workaround do not exist for platforms with display 10 or newer
+	 * but testing proved that it works for up display 13, for newer
+	 * than that testing will be needed.
+	 */
+	intel_de_write(dev_priv, CURSURFLIVE(intel_dp->psr.pipe), 0);
+}
+
+void intel_psr2_disable_plane_sel_fetch(struct intel_plane *plane,
+					const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	if (!crtc_state->enable_psr2_sel_fetch)
+		return;
+
+	intel_de_write_fw(dev_priv, PLANE_SEL_FETCH_CTL(pipe, plane->id), 0);
+}
+
+void intel_psr2_program_plane_sel_fetch(struct intel_plane *plane,
+					const struct intel_crtc_state *crtc_state,
+					const struct intel_plane_state *plane_state,
+					int color_plane)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	const struct drm_rect *clip;
+	u32 val;
+	int x, y;
+
+	if (!crtc_state->enable_psr2_sel_fetch)
+		return;
+
+	if (plane->id == PLANE_CURSOR) {
+		intel_de_write_fw(dev_priv, PLANE_SEL_FETCH_CTL(pipe, plane->id),
+				  plane_state->ctl);
+		return;
+	}
+
+	clip = &plane_state->psr2_sel_fetch_area;
+
+	val = (clip->y1 + plane_state->uapi.dst.y1) << 16;
+	val |= plane_state->uapi.dst.x1;
+	intel_de_write_fw(dev_priv, PLANE_SEL_FETCH_POS(pipe, plane->id), val);
+
+	x = plane_state->view.color_plane[color_plane].x;
+
+	/*
+	 * From Bspec: UV surface Start Y Position = half of Y plane Y
+	 * start position.
+	 */
+	if (!color_plane)
+		y = plane_state->view.color_plane[color_plane].y + clip->y1;
+	else
+		y = plane_state->view.color_plane[color_plane].y + clip->y1 / 2;
+
+	val = y << 16 | x;
+
+	intel_de_write_fw(dev_priv, PLANE_SEL_FETCH_OFFSET(pipe, plane->id),
+			  val);
+
+	/* Sizes are 0 based */
+	val = (drm_rect_height(clip) - 1) << 16;
+	val |= (drm_rect_width(&plane_state->uapi.src) >> 16) - 1;
+	intel_de_write_fw(dev_priv, PLANE_SEL_FETCH_SIZE(pipe, plane->id), val);
+
+	intel_de_write_fw(dev_priv, PLANE_SEL_FETCH_CTL(pipe, plane->id),
+			  PLANE_SEL_FETCH_CTL_ENABLE);
+}
+
+void intel_psr2_program_trans_man_trk_ctl(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	struct intel_encoder *encoder;
+
+	if (!crtc_state->enable_psr2_sel_fetch)
+		return;
+
+	for_each_intel_encoder_mask_with_psr(&dev_priv->drm, encoder,
+					     crtc_state->uapi.encoder_mask) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		lockdep_assert_held(&intel_dp->psr.lock);
+		if (intel_dp->psr.psr2_sel_fetch_cff_enabled)
+			return;
+		break;
+	}
+
+	intel_de_write(dev_priv, PSR2_MAN_TRK_CTL(crtc_state->cpu_transcoder),
+		       crtc_state->psr2_man_track_ctl);
+}
+
+static void psr2_man_trk_ctl_calc(struct intel_crtc_state *crtc_state,
+				  struct drm_rect *clip, bool full_update)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 val = man_trk_ctl_enable_bit_get(dev_priv);
+
+	/* SF partial frame enable has to be set even on full update */
+	val |= man_trk_ctl_partial_frame_bit_get(dev_priv);
+
+	if (full_update) {
+		/*
+		 * Not applying Wa_14014971508:adlp as we do not support the
+		 * feature that requires this workaround.
+		 */
+		val |= man_trk_ctl_single_full_frame_bit_get(dev_priv);
+		goto exit;
+	}
+
+	if (clip->y1 == -1)
+		goto exit;
+
+	if (IS_ALDERLAKE_P(dev_priv)) {
+		val |= ADLP_PSR2_MAN_TRK_CTL_SU_REGION_START_ADDR(clip->y1);
+		val |= ADLP_PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR(clip->y2 - 1);
+	} else {
+		drm_WARN_ON(crtc_state->uapi.crtc->dev, clip->y1 % 4 || clip->y2 % 4);
+
+		val |= PSR2_MAN_TRK_CTL_SU_REGION_START_ADDR(clip->y1 / 4 + 1);
+		val |= PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR(clip->y2 / 4 + 1);
+	}
+exit:
+	crtc_state->psr2_man_track_ctl = val;
+}
+
+static void clip_area_update(struct drm_rect *overlap_damage_area,
+			     struct drm_rect *damage_area,
+			     struct drm_rect *pipe_src)
+{
+	if (!drm_rect_intersect(damage_area, pipe_src))
+		return;
+
+	if (overlap_damage_area->y1 == -1) {
+		overlap_damage_area->y1 = damage_area->y1;
+		overlap_damage_area->y2 = damage_area->y2;
+		return;
+	}
+
+	if (damage_area->y1 < overlap_damage_area->y1)
+		overlap_damage_area->y1 = damage_area->y1;
+
+	if (damage_area->y2 > overlap_damage_area->y2)
+		overlap_damage_area->y2 = damage_area->y2;
+}
+
+static void intel_psr2_sel_fetch_pipe_alignment(const struct intel_crtc_state *crtc_state,
+						struct drm_rect *pipe_clip)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	const u16 y_alignment = crtc_state->su_y_granularity;
+
+	pipe_clip->y1 -= pipe_clip->y1 % y_alignment;
+	if (pipe_clip->y2 % y_alignment)
+		pipe_clip->y2 = ((pipe_clip->y2 / y_alignment) + 1) * y_alignment;
+
+	if (IS_ALDERLAKE_P(dev_priv) && crtc_state->dsc.compression_enable)
+		drm_warn(&dev_priv->drm, "Missing PSR2 sel fetch alignment with DSC\n");
+}
+
+/*
+ * TODO: Not clear how to handle planes with negative position,
+ * also planes are not updated if they have a negative X
+ * position so for now doing a full update in this cases
+ *
+ * Plane scaling and rotation is not supported by selective fetch and both
+ * properties can change without a modeset, so need to be check at every
+ * atomic commit.
+ */
+static bool psr2_sel_fetch_plane_state_supported(const struct intel_plane_state *plane_state)
+{
+	if (plane_state->uapi.dst.y1 < 0 ||
+	    plane_state->uapi.dst.x1 < 0 ||
+	    plane_state->scaler_id >= 0 ||
+	    plane_state->uapi.rotation != DRM_MODE_ROTATE_0)
+		return false;
+
+	return true;
+}
+
+/*
+ * Check for pipe properties that is not supported by selective fetch.
+ *
+ * TODO: pipe scaling causes a modeset but skl_update_scaler_crtc() is executed
+ * after intel_psr_compute_config(), so for now keeping PSR2 selective fetch
+ * enabled and going to the full update path.
+ */
+static bool psr2_sel_fetch_pipe_state_supported(const struct intel_crtc_state *crtc_state)
+{
+	if (crtc_state->scaler_state.scaler_id >= 0)
+		return false;
+
+	return true;
+}
+
+int intel_psr2_sel_fetch_update(struct intel_atomic_state *state,
+				struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state = intel_atomic_get_new_crtc_state(state, crtc);
+	struct drm_rect pipe_clip = { .x1 = 0, .y1 = -1, .x2 = INT_MAX, .y2 = -1 };
+	struct intel_plane_state *new_plane_state, *old_plane_state;
+	struct intel_plane *plane;
+	bool full_update = false;
+	int i, ret;
+
+	if (!crtc_state->enable_psr2_sel_fetch)
+		return 0;
+
+	if (!psr2_sel_fetch_pipe_state_supported(crtc_state)) {
+		full_update = true;
+		goto skip_sel_fetch_set_loop;
+	}
+
+	/*
+	 * Calculate minimal selective fetch area of each plane and calculate
+	 * the pipe damaged area.
+	 * In the next loop the plane selective fetch area will actually be set
+	 * using whole pipe damaged area.
+	 */
+	for_each_oldnew_intel_plane_in_state(state, plane, old_plane_state,
+					     new_plane_state, i) {
+		struct drm_rect src, damaged_area = { .x1 = 0, .y1 = -1,
+						      .x2 = INT_MAX };
+
+		if (new_plane_state->uapi.crtc != crtc_state->uapi.crtc)
+			continue;
+
+		if (!new_plane_state->uapi.visible &&
+		    !old_plane_state->uapi.visible)
+			continue;
+
+		if (!psr2_sel_fetch_plane_state_supported(new_plane_state)) {
+			full_update = true;
+			break;
+		}
+
+		/*
+		 * If visibility or plane moved, mark the whole plane area as
+		 * damaged as it needs to be complete redraw in the new and old
+		 * position.
+		 */
+		if (new_plane_state->uapi.visible != old_plane_state->uapi.visible ||
+		    !drm_rect_equals(&new_plane_state->uapi.dst,
+				     &old_plane_state->uapi.dst)) {
+			if (old_plane_state->uapi.visible) {
+				damaged_area.y1 = old_plane_state->uapi.dst.y1;
+				damaged_area.y2 = old_plane_state->uapi.dst.y2;
+				clip_area_update(&pipe_clip, &damaged_area,
+						 &crtc_state->pipe_src);
+			}
+
+			if (new_plane_state->uapi.visible) {
+				damaged_area.y1 = new_plane_state->uapi.dst.y1;
+				damaged_area.y2 = new_plane_state->uapi.dst.y2;
+				clip_area_update(&pipe_clip, &damaged_area,
+						 &crtc_state->pipe_src);
+			}
+			continue;
+		} else if (new_plane_state->uapi.alpha != old_plane_state->uapi.alpha) {
+			/* If alpha changed mark the whole plane area as damaged */
+			damaged_area.y1 = new_plane_state->uapi.dst.y1;
+			damaged_area.y2 = new_plane_state->uapi.dst.y2;
+			clip_area_update(&pipe_clip, &damaged_area,
+					 &crtc_state->pipe_src);
+			continue;
+		}
+
+		src = drm_plane_state_src(&new_plane_state->uapi);
+		drm_rect_fp_to_int(&src, &src);
+
+		if (!drm_atomic_helper_damage_merged(&old_plane_state->uapi,
+						     &new_plane_state->uapi, &damaged_area))
+			continue;
+
+		damaged_area.y1 += new_plane_state->uapi.dst.y1 - src.y1;
+		damaged_area.y2 += new_plane_state->uapi.dst.y1 - src.y1;
+		damaged_area.x1 += new_plane_state->uapi.dst.x1 - src.x1;
+		damaged_area.x2 += new_plane_state->uapi.dst.x1 - src.x1;
+
+		clip_area_update(&pipe_clip, &damaged_area, &crtc_state->pipe_src);
+	}
+
+	/*
+	 * TODO: For now we are just using full update in case
+	 * selective fetch area calculation fails. To optimize this we
+	 * should identify cases where this happens and fix the area
+	 * calculation for those.
+	 */
+	if (pipe_clip.y1 == -1) {
+		drm_info_once(&dev_priv->drm,
+			      "Selective fetch area calculation failed in pipe %c\n",
+			      pipe_name(crtc->pipe));
+		full_update = true;
+	}
+
+	if (full_update)
+		goto skip_sel_fetch_set_loop;
+
+	ret = drm_atomic_add_affected_planes(&state->base, &crtc->base);
+	if (ret)
+		return ret;
+
+	intel_psr2_sel_fetch_pipe_alignment(crtc_state, &pipe_clip);
+
+	/*
+	 * Now that we have the pipe damaged area check if it intersect with
+	 * every plane, if it does set the plane selective fetch area.
+	 */
+	for_each_oldnew_intel_plane_in_state(state, plane, old_plane_state,
+					     new_plane_state, i) {
+		struct drm_rect *sel_fetch_area, inter;
+		struct intel_plane *linked = new_plane_state->planar_linked_plane;
+
+		if (new_plane_state->uapi.crtc != crtc_state->uapi.crtc ||
+		    !new_plane_state->uapi.visible)
+			continue;
+
+		inter = pipe_clip;
+		if (!drm_rect_intersect(&inter, &new_plane_state->uapi.dst))
+			continue;
+
+		if (!psr2_sel_fetch_plane_state_supported(new_plane_state)) {
+			full_update = true;
+			break;
+		}
+
+		sel_fetch_area = &new_plane_state->psr2_sel_fetch_area;
+		sel_fetch_area->y1 = inter.y1 - new_plane_state->uapi.dst.y1;
+		sel_fetch_area->y2 = inter.y2 - new_plane_state->uapi.dst.y1;
+		crtc_state->update_planes |= BIT(plane->id);
+
+		/*
+		 * Sel_fetch_area is calculated for UV plane. Use
+		 * same area for Y plane as well.
+		 */
+		if (linked) {
+			struct intel_plane_state *linked_new_plane_state;
+			struct drm_rect *linked_sel_fetch_area;
+
+			linked_new_plane_state = intel_atomic_get_plane_state(state, linked);
+			if (IS_ERR(linked_new_plane_state))
+				return PTR_ERR(linked_new_plane_state);
+
+			linked_sel_fetch_area = &linked_new_plane_state->psr2_sel_fetch_area;
+			linked_sel_fetch_area->y1 = sel_fetch_area->y1;
+			linked_sel_fetch_area->y2 = sel_fetch_area->y2;
+			crtc_state->update_planes |= BIT(linked->id);
+		}
+	}
+
+skip_sel_fetch_set_loop:
+	psr2_man_trk_ctl_calc(crtc_state, &pipe_clip, full_update);
+	return 0;
+}
+
+void intel_psr_pre_plane_update(struct intel_atomic_state *state,
+				struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	const struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_encoder *encoder;
+
+	if (!HAS_PSR(i915))
+		return;
+
+	for_each_intel_encoder_mask_with_psr(state->base.dev, encoder,
+					     old_crtc_state->uapi.encoder_mask) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+		struct intel_psr *psr = &intel_dp->psr;
+		bool needs_to_disable = false;
+
+		mutex_lock(&psr->lock);
+
+		/*
+		 * Reasons to disable:
+		 * - PSR disabled in new state
+		 * - All planes will go inactive
+		 * - Changing between PSR versions
+		 */
+		needs_to_disable |= intel_crtc_needs_modeset(new_crtc_state);
+		needs_to_disable |= !new_crtc_state->has_psr;
+		needs_to_disable |= !new_crtc_state->active_planes;
+		needs_to_disable |= new_crtc_state->has_psr2 != psr->psr2_enabled;
+
+		if (psr->enabled && needs_to_disable)
+			intel_psr_disable_locked(intel_dp);
+
+		mutex_unlock(&psr->lock);
+	}
+}
+
+static void _intel_psr_post_plane_update(const struct intel_atomic_state *state,
+					 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_encoder *encoder;
+
+	if (!crtc_state->has_psr)
+		return;
+
+	for_each_intel_encoder_mask_with_psr(state->base.dev, encoder,
+					     crtc_state->uapi.encoder_mask) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+		struct intel_psr *psr = &intel_dp->psr;
+
+		mutex_lock(&psr->lock);
+
+		if (psr->sink_not_reliable)
+			goto exit;
+
+		drm_WARN_ON(&dev_priv->drm, psr->enabled && !crtc_state->active_planes);
+
+		/* Only enable if there is active planes */
+		if (!psr->enabled && crtc_state->active_planes)
+			intel_psr_enable_locked(intel_dp, crtc_state);
+
+		/* Force a PSR exit when enabling CRC to avoid CRC timeouts */
+		if (crtc_state->crc_enabled && psr->enabled)
+			psr_force_hw_tracking_exit(intel_dp);
+
+exit:
+		mutex_unlock(&psr->lock);
+	}
+}
+
+void intel_psr_post_plane_update(const struct intel_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state;
+	struct intel_crtc *crtc;
+	int i;
+
+	if (!HAS_PSR(dev_priv))
+		return;
+
+	for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i)
+		_intel_psr_post_plane_update(state, crtc_state);
+}
+
+static int _psr2_ready_for_pipe_update_locked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	/*
+	 * Any state lower than EDP_PSR2_STATUS_STATE_DEEP_SLEEP is enough.
+	 * As all higher states has bit 4 of PSR2 state set we can just wait for
+	 * EDP_PSR2_STATUS_STATE_DEEP_SLEEP to be cleared.
+	 */
+	return intel_de_wait_for_clear(dev_priv,
+				       EDP_PSR2_STATUS(intel_dp->psr.transcoder),
+				       EDP_PSR2_STATUS_STATE_DEEP_SLEEP, 50);
+}
+
+static int _psr1_ready_for_pipe_update_locked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	/*
+	 * From bspec: Panel Self Refresh (BDW+)
+	 * Max. time for PSR to idle = Inverse of the refresh rate + 6 ms of
+	 * exit training time + 1.5 ms of aux channel handshake. 50 ms is
+	 * defensive enough to cover everything.
+	 */
+	return intel_de_wait_for_clear(dev_priv,
+				       EDP_PSR_STATUS(intel_dp->psr.transcoder),
+				       EDP_PSR_STATUS_STATE_MASK, 50);
+}
+
+/**
+ * intel_psr_wait_for_idle_locked - wait for PSR be ready for a pipe update
+ * @new_crtc_state: new CRTC state
+ *
+ * This function is expected to be called from pipe_update_start() where it is
+ * not expected to race with PSR enable or disable.
+ */
+void intel_psr_wait_for_idle_locked(const struct intel_crtc_state *new_crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(new_crtc_state->uapi.crtc->dev);
+	struct intel_encoder *encoder;
+
+	if (!new_crtc_state->has_psr)
+		return;
+
+	for_each_intel_encoder_mask_with_psr(&dev_priv->drm, encoder,
+					     new_crtc_state->uapi.encoder_mask) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+		int ret;
+
+		lockdep_assert_held(&intel_dp->psr.lock);
+
+		if (!intel_dp->psr.enabled)
+			continue;
+
+		if (intel_dp->psr.psr2_enabled)
+			ret = _psr2_ready_for_pipe_update_locked(intel_dp);
+		else
+			ret = _psr1_ready_for_pipe_update_locked(intel_dp);
+
+		if (ret)
+			drm_err(&dev_priv->drm, "PSR wait timed out, atomic update may fail\n");
+	}
+}
+
+static bool __psr_wait_for_idle_locked(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	i915_reg_t reg;
+	u32 mask;
+	int err;
+
+	if (!intel_dp->psr.enabled)
+		return false;
+
+	if (intel_dp->psr.psr2_enabled) {
+		reg = EDP_PSR2_STATUS(intel_dp->psr.transcoder);
+		mask = EDP_PSR2_STATUS_STATE_MASK;
+	} else {
+		reg = EDP_PSR_STATUS(intel_dp->psr.transcoder);
+		mask = EDP_PSR_STATUS_STATE_MASK;
+	}
+
+	mutex_unlock(&intel_dp->psr.lock);
+
+	err = intel_de_wait_for_clear(dev_priv, reg, mask, 50);
+	if (err)
+		drm_err(&dev_priv->drm,
+			"Timed out waiting for PSR Idle for re-enable\n");
+
+	/* After the unlocked wait, verify that PSR is still wanted! */
+	mutex_lock(&intel_dp->psr.lock);
+	return err == 0 && intel_dp->psr.enabled;
+}
+
+static int intel_psr_fastset_force(struct drm_i915_private *dev_priv)
+{
+	struct drm_connector_list_iter conn_iter;
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_modeset_acquire_ctx ctx;
+	struct drm_atomic_state *state;
+	struct drm_connector *conn;
+	int err = 0;
+
+	state = drm_atomic_state_alloc(dev);
+	if (!state)
+		return -ENOMEM;
+
+	drm_modeset_acquire_init(&ctx, DRM_MODESET_ACQUIRE_INTERRUPTIBLE);
+	state->acquire_ctx = &ctx;
+
+retry:
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	drm_for_each_connector_iter(conn, &conn_iter) {
+		struct drm_connector_state *conn_state;
+		struct drm_crtc_state *crtc_state;
+
+		if (conn->connector_type != DRM_MODE_CONNECTOR_eDP)
+			continue;
+
+		conn_state = drm_atomic_get_connector_state(state, conn);
+		if (IS_ERR(conn_state)) {
+			err = PTR_ERR(conn_state);
+			break;
+		}
+
+		if (!conn_state->crtc)
+			continue;
+
+		crtc_state = drm_atomic_get_crtc_state(state, conn_state->crtc);
+		if (IS_ERR(crtc_state)) {
+			err = PTR_ERR(crtc_state);
+			break;
+		}
+
+		/* Mark mode as changed to trigger a pipe->update() */
+		crtc_state->mode_changed = true;
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
+	if (err == 0)
+		err = drm_atomic_commit(state);
+
+	if (err == -EDEADLK) {
+		drm_atomic_state_clear(state);
+		err = drm_modeset_backoff(&ctx);
+		if (!err)
+			goto retry;
+	}
+
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+	drm_atomic_state_put(state);
+
+	return err;
+}
+
+int intel_psr_debug_set(struct intel_dp *intel_dp, u64 val)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	const u32 mode = val & I915_PSR_DEBUG_MODE_MASK;
+	u32 old_mode;
+	int ret;
+
+	if (val & ~(I915_PSR_DEBUG_IRQ | I915_PSR_DEBUG_MODE_MASK) ||
+	    mode > I915_PSR_DEBUG_ENABLE_SEL_FETCH) {
+		drm_dbg_kms(&dev_priv->drm, "Invalid debug mask %llx\n", val);
+		return -EINVAL;
+	}
+
+	ret = mutex_lock_interruptible(&intel_dp->psr.lock);
+	if (ret)
+		return ret;
+
+	old_mode = intel_dp->psr.debug & I915_PSR_DEBUG_MODE_MASK;
+	intel_dp->psr.debug = val;
+
+	/*
+	 * Do it right away if it's already enabled, otherwise it will be done
+	 * when enabling the source.
+	 */
+	if (intel_dp->psr.enabled)
+		psr_irq_control(intel_dp);
+
+	mutex_unlock(&intel_dp->psr.lock);
+
+	if (old_mode != mode)
+		ret = intel_psr_fastset_force(dev_priv);
+
+	return ret;
+}
+
+static void intel_psr_handle_irq(struct intel_dp *intel_dp)
+{
+	struct intel_psr *psr = &intel_dp->psr;
+
+	intel_psr_disable_locked(intel_dp);
+	psr->sink_not_reliable = true;
+	/* let's make sure that sink is awaken */
+	drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER, DP_SET_POWER_D0);
+}
+
+static void intel_psr_work(struct work_struct *work)
+{
+	struct intel_dp *intel_dp =
+		container_of(work, typeof(*intel_dp), psr.work);
+
+	mutex_lock(&intel_dp->psr.lock);
+
+	if (!intel_dp->psr.enabled)
+		goto unlock;
+
+	if (READ_ONCE(intel_dp->psr.irq_aux_error))
+		intel_psr_handle_irq(intel_dp);
+
+	/*
+	 * We have to make sure PSR is ready for re-enable
+	 * otherwise it keeps disabled until next full enable/disable cycle.
+	 * PSR might take some time to get fully disabled
+	 * and be ready for re-enable.
+	 */
+	if (!__psr_wait_for_idle_locked(intel_dp))
+		goto unlock;
+
+	/*
+	 * The delayed work can race with an invalidate hence we need to
+	 * recheck. Since psr_flush first clears this and then reschedules we
+	 * won't ever miss a flush when bailing out here.
+	 */
+	if (intel_dp->psr.busy_frontbuffer_bits || intel_dp->psr.active)
+		goto unlock;
+
+	intel_psr_activate(intel_dp);
+unlock:
+	mutex_unlock(&intel_dp->psr.lock);
+}
+
+static void _psr_invalidate_handle(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	if (intel_dp->psr.psr2_sel_fetch_enabled) {
+		u32 val;
+
+		if (intel_dp->psr.psr2_sel_fetch_cff_enabled) {
+			/* Send one update otherwise lag is observed in screen */
+			intel_de_write(dev_priv, CURSURFLIVE(intel_dp->psr.pipe), 0);
+			return;
+		}
+
+		val = man_trk_ctl_enable_bit_get(dev_priv) |
+		      man_trk_ctl_partial_frame_bit_get(dev_priv) |
+		      man_trk_ctl_continuos_full_frame(dev_priv);
+		intel_de_write(dev_priv, PSR2_MAN_TRK_CTL(intel_dp->psr.transcoder), val);
+		intel_de_write(dev_priv, CURSURFLIVE(intel_dp->psr.pipe), 0);
+		intel_dp->psr.psr2_sel_fetch_cff_enabled = true;
+	} else {
+		intel_psr_exit(intel_dp);
+	}
+}
+
+/**
+ * intel_psr_invalidate - Invalidate PSR
+ * @dev_priv: i915 device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ * @origin: which operation caused the invalidate
+ *
+ * Since the hardware frontbuffer tracking has gaps we need to integrate
+ * with the software frontbuffer tracking. This function gets called every
+ * time frontbuffer rendering starts and a buffer gets dirtied. PSR must be
+ * disabled if the frontbuffer mask contains a buffer relevant to PSR.
+ *
+ * Dirty frontbuffers relevant to PSR are tracked in busy_frontbuffer_bits."
+ */
+void intel_psr_invalidate(struct drm_i915_private *dev_priv,
+			  unsigned frontbuffer_bits, enum fb_op_origin origin)
+{
+	struct intel_encoder *encoder;
+
+	if (origin == ORIGIN_FLIP)
+		return;
+
+	for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) {
+		unsigned int pipe_frontbuffer_bits = frontbuffer_bits;
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		mutex_lock(&intel_dp->psr.lock);
+		if (!intel_dp->psr.enabled) {
+			mutex_unlock(&intel_dp->psr.lock);
+			continue;
+		}
+
+		pipe_frontbuffer_bits &=
+			INTEL_FRONTBUFFER_ALL_MASK(intel_dp->psr.pipe);
+		intel_dp->psr.busy_frontbuffer_bits |= pipe_frontbuffer_bits;
+
+		if (pipe_frontbuffer_bits)
+			_psr_invalidate_handle(intel_dp);
+
+		mutex_unlock(&intel_dp->psr.lock);
+	}
+}
+/*
+ * When we will be completely rely on PSR2 S/W tracking in future,
+ * intel_psr_flush() will invalidate and flush the PSR for ORIGIN_FLIP
+ * event also therefore tgl_dc3co_flush_locked() require to be changed
+ * accordingly in future.
+ */
+static void
+tgl_dc3co_flush_locked(struct intel_dp *intel_dp, unsigned int frontbuffer_bits,
+		       enum fb_op_origin origin)
+{
+	if (!intel_dp->psr.dc3co_exitline || !intel_dp->psr.psr2_enabled ||
+	    !intel_dp->psr.active)
+		return;
+
+	/*
+	 * At every frontbuffer flush flip event modified delay of delayed work,
+	 * when delayed work schedules that means display has been idle.
+	 */
+	if (!(frontbuffer_bits &
+	    INTEL_FRONTBUFFER_ALL_MASK(intel_dp->psr.pipe)))
+		return;
+
+	tgl_psr2_enable_dc3co(intel_dp);
+	mod_delayed_work(system_wq, &intel_dp->psr.dc3co_work,
+			 intel_dp->psr.dc3co_exit_delay);
+}
+
+static void _psr_flush_handle(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	if (intel_dp->psr.psr2_sel_fetch_enabled) {
+		if (intel_dp->psr.psr2_sel_fetch_cff_enabled) {
+			/* can we turn CFF off? */
+			if (intel_dp->psr.busy_frontbuffer_bits == 0) {
+				u32 val = man_trk_ctl_enable_bit_get(dev_priv) |
+					  man_trk_ctl_partial_frame_bit_get(dev_priv) |
+					  man_trk_ctl_single_full_frame_bit_get(dev_priv);
+
+				/*
+				 * turn continuous full frame off and do a single
+				 * full frame
+				 */
+				intel_de_write(dev_priv, PSR2_MAN_TRK_CTL(intel_dp->psr.transcoder),
+					       val);
+				intel_de_write(dev_priv, CURSURFLIVE(intel_dp->psr.pipe), 0);
+				intel_dp->psr.psr2_sel_fetch_cff_enabled = false;
+			}
+		} else {
+			/*
+			 * continuous full frame is disabled, only a single full
+			 * frame is required
+			 */
+			psr_force_hw_tracking_exit(intel_dp);
+		}
+	} else {
+		psr_force_hw_tracking_exit(intel_dp);
+
+		if (!intel_dp->psr.active && !intel_dp->psr.busy_frontbuffer_bits)
+			schedule_work(&intel_dp->psr.work);
+	}
+}
+
+/**
+ * intel_psr_flush - Flush PSR
+ * @dev_priv: i915 device
+ * @frontbuffer_bits: frontbuffer plane tracking bits
+ * @origin: which operation caused the flush
+ *
+ * Since the hardware frontbuffer tracking has gaps we need to integrate
+ * with the software frontbuffer tracking. This function gets called every
+ * time frontbuffer rendering has completed and flushed out to memory. PSR
+ * can be enabled again if no other frontbuffer relevant to PSR is dirty.
+ *
+ * Dirty frontbuffers relevant to PSR are tracked in busy_frontbuffer_bits.
+ */
+void intel_psr_flush(struct drm_i915_private *dev_priv,
+		     unsigned frontbuffer_bits, enum fb_op_origin origin)
+{
+	struct intel_encoder *encoder;
+
+	for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) {
+		unsigned int pipe_frontbuffer_bits = frontbuffer_bits;
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		mutex_lock(&intel_dp->psr.lock);
+		if (!intel_dp->psr.enabled) {
+			mutex_unlock(&intel_dp->psr.lock);
+			continue;
+		}
+
+		pipe_frontbuffer_bits &=
+			INTEL_FRONTBUFFER_ALL_MASK(intel_dp->psr.pipe);
+		intel_dp->psr.busy_frontbuffer_bits &= ~pipe_frontbuffer_bits;
+
+		/*
+		 * If the PSR is paused by an explicit intel_psr_paused() call,
+		 * we have to ensure that the PSR is not activated until
+		 * intel_psr_resume() is called.
+		 */
+		if (intel_dp->psr.paused)
+			goto unlock;
+
+		if (origin == ORIGIN_FLIP ||
+		    (origin == ORIGIN_CURSOR_UPDATE &&
+		     !intel_dp->psr.psr2_sel_fetch_enabled)) {
+			tgl_dc3co_flush_locked(intel_dp, frontbuffer_bits, origin);
+			goto unlock;
+		}
+
+		if (pipe_frontbuffer_bits == 0)
+			goto unlock;
+
+		/* By definition flush = invalidate + flush */
+		_psr_flush_handle(intel_dp);
+unlock:
+		mutex_unlock(&intel_dp->psr.lock);
+	}
+}
+
+/**
+ * intel_psr_init - Init basic PSR work and mutex.
+ * @intel_dp: Intel DP
+ *
+ * This function is called after the initializing connector.
+ * (the initializing of connector treats the handling of connector capabilities)
+ * And it initializes basic PSR stuff for each DP Encoder.
+ */
+void intel_psr_init(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+
+	if (!HAS_PSR(dev_priv))
+		return;
+
+	/*
+	 * HSW spec explicitly says PSR is tied to port A.
+	 * BDW+ platforms have a instance of PSR registers per transcoder but
+	 * BDW, GEN9 and GEN11 are not validated by HW team in other transcoder
+	 * than eDP one.
+	 * For now it only supports one instance of PSR for BDW, GEN9 and GEN11.
+	 * So lets keep it hardcoded to PORT_A for BDW, GEN9 and GEN11.
+	 * But GEN12 supports a instance of PSR registers per transcoder.
+	 */
+	if (DISPLAY_VER(dev_priv) < 12 && dig_port->base.port != PORT_A) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR condition failed: Port not supported\n");
+		return;
+	}
+
+	intel_dp->psr.source_support = true;
+
+	/* Set link_standby x link_off defaults */
+	if (DISPLAY_VER(dev_priv) < 12)
+		/* For new platforms up to TGL let's respect VBT back again */
+		intel_dp->psr.link_standby = connector->panel.vbt.psr.full_link;
+
+	INIT_WORK(&intel_dp->psr.work, intel_psr_work);
+	INIT_DELAYED_WORK(&intel_dp->psr.dc3co_work, tgl_dc3co_disable_work);
+	mutex_init(&intel_dp->psr.lock);
+}
+
+static int psr_get_status_and_error_status(struct intel_dp *intel_dp,
+					   u8 *status, u8 *error_status)
+{
+	struct drm_dp_aux *aux = &intel_dp->aux;
+	int ret;
+
+	ret = drm_dp_dpcd_readb(aux, DP_PSR_STATUS, status);
+	if (ret != 1)
+		return ret;
+
+	ret = drm_dp_dpcd_readb(aux, DP_PSR_ERROR_STATUS, error_status);
+	if (ret != 1)
+		return ret;
+
+	*status = *status & DP_PSR_SINK_STATE_MASK;
+
+	return 0;
+}
+
+static void psr_alpm_check(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct drm_dp_aux *aux = &intel_dp->aux;
+	struct intel_psr *psr = &intel_dp->psr;
+	u8 val;
+	int r;
+
+	if (!psr->psr2_enabled)
+		return;
+
+	r = drm_dp_dpcd_readb(aux, DP_RECEIVER_ALPM_STATUS, &val);
+	if (r != 1) {
+		drm_err(&dev_priv->drm, "Error reading ALPM status\n");
+		return;
+	}
+
+	if (val & DP_ALPM_LOCK_TIMEOUT_ERROR) {
+		intel_psr_disable_locked(intel_dp);
+		psr->sink_not_reliable = true;
+		drm_dbg_kms(&dev_priv->drm,
+			    "ALPM lock timeout error, disabling PSR\n");
+
+		/* Clearing error */
+		drm_dp_dpcd_writeb(aux, DP_RECEIVER_ALPM_STATUS, val);
+	}
+}
+
+static void psr_capability_changed_check(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_psr *psr = &intel_dp->psr;
+	u8 val;
+	int r;
+
+	r = drm_dp_dpcd_readb(&intel_dp->aux, DP_PSR_ESI, &val);
+	if (r != 1) {
+		drm_err(&dev_priv->drm, "Error reading DP_PSR_ESI\n");
+		return;
+	}
+
+	if (val & DP_PSR_CAPS_CHANGE) {
+		intel_psr_disable_locked(intel_dp);
+		psr->sink_not_reliable = true;
+		drm_dbg_kms(&dev_priv->drm,
+			    "Sink PSR capability changed, disabling PSR\n");
+
+		/* Clearing it */
+		drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_ESI, val);
+	}
+}
+
+void intel_psr_short_pulse(struct intel_dp *intel_dp)
+{
+	struct drm_i915_private *dev_priv = dp_to_i915(intel_dp);
+	struct intel_psr *psr = &intel_dp->psr;
+	u8 status, error_status;
+	const u8 errors = DP_PSR_RFB_STORAGE_ERROR |
+			  DP_PSR_VSC_SDP_UNCORRECTABLE_ERROR |
+			  DP_PSR_LINK_CRC_ERROR;
+
+	if (!CAN_PSR(intel_dp))
+		return;
+
+	mutex_lock(&psr->lock);
+
+	if (!psr->enabled)
+		goto exit;
+
+	if (psr_get_status_and_error_status(intel_dp, &status, &error_status)) {
+		drm_err(&dev_priv->drm,
+			"Error reading PSR status or error status\n");
+		goto exit;
+	}
+
+	if (status == DP_PSR_SINK_INTERNAL_ERROR || (error_status & errors)) {
+		intel_psr_disable_locked(intel_dp);
+		psr->sink_not_reliable = true;
+	}
+
+	if (status == DP_PSR_SINK_INTERNAL_ERROR && !error_status)
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR sink internal error, disabling PSR\n");
+	if (error_status & DP_PSR_RFB_STORAGE_ERROR)
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR RFB storage error, disabling PSR\n");
+	if (error_status & DP_PSR_VSC_SDP_UNCORRECTABLE_ERROR)
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR VSC SDP uncorrectable error, disabling PSR\n");
+	if (error_status & DP_PSR_LINK_CRC_ERROR)
+		drm_dbg_kms(&dev_priv->drm,
+			    "PSR Link CRC error, disabling PSR\n");
+
+	if (error_status & ~errors)
+		drm_err(&dev_priv->drm,
+			"PSR_ERROR_STATUS unhandled errors %x\n",
+			error_status & ~errors);
+	/* clear status register */
+	drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_ERROR_STATUS, error_status);
+
+	psr_alpm_check(intel_dp);
+	psr_capability_changed_check(intel_dp);
+
+exit:
+	mutex_unlock(&psr->lock);
+}
+
+bool intel_psr_enabled(struct intel_dp *intel_dp)
+{
+	bool ret;
+
+	if (!CAN_PSR(intel_dp))
+		return false;
+
+	mutex_lock(&intel_dp->psr.lock);
+	ret = intel_dp->psr.enabled;
+	mutex_unlock(&intel_dp->psr.lock);
+
+	return ret;
+}
+
+/**
+ * intel_psr_lock - grab PSR lock
+ * @crtc_state: the crtc state
+ *
+ * This is initially meant to be used by around CRTC update, when
+ * vblank sensitive registers are updated and we need grab the lock
+ * before it to avoid vblank evasion.
+ */
+void intel_psr_lock(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	struct intel_encoder *encoder;
+
+	if (!crtc_state->has_psr)
+		return;
+
+	for_each_intel_encoder_mask_with_psr(&i915->drm, encoder,
+					     crtc_state->uapi.encoder_mask) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		mutex_lock(&intel_dp->psr.lock);
+		break;
+	}
+}
+
+/**
+ * intel_psr_unlock - release PSR lock
+ * @crtc_state: the crtc state
+ *
+ * Release the PSR lock that was held during pipe update.
+ */
+void intel_psr_unlock(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	struct intel_encoder *encoder;
+
+	if (!crtc_state->has_psr)
+		return;
+
+	for_each_intel_encoder_mask_with_psr(&i915->drm, encoder,
+					     crtc_state->uapi.encoder_mask) {
+		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+		mutex_unlock(&intel_dp->psr.lock);
+		break;
+	}
+}
diff --git a/drivers/gpu/drm/i915/display/intel_psr.h b/drivers/gpu/drm/i915/display/intel_psr.h
new file mode 100644
index 000000000..2ac3a46cc
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_psr.h
@@ -0,0 +1,61 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_PSR_H__
+#define __INTEL_PSR_H__
+
+#include <linux/types.h>
+
+enum fb_op_origin;
+struct drm_connector;
+struct drm_connector_state;
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_dp;
+struct intel_encoder;
+struct intel_plane;
+struct intel_plane_state;
+
+void intel_psr_init_dpcd(struct intel_dp *intel_dp);
+void intel_psr_pre_plane_update(struct intel_atomic_state *state,
+				struct intel_crtc *crtc);
+void intel_psr_post_plane_update(const struct intel_atomic_state *state);
+void intel_psr_disable(struct intel_dp *intel_dp,
+		       const struct intel_crtc_state *old_crtc_state);
+int intel_psr_debug_set(struct intel_dp *intel_dp, u64 value);
+void intel_psr_invalidate(struct drm_i915_private *dev_priv,
+			  unsigned frontbuffer_bits,
+			  enum fb_op_origin origin);
+void intel_psr_flush(struct drm_i915_private *dev_priv,
+		     unsigned frontbuffer_bits,
+		     enum fb_op_origin origin);
+void intel_psr_init(struct intel_dp *intel_dp);
+void intel_psr_compute_config(struct intel_dp *intel_dp,
+			      struct intel_crtc_state *crtc_state,
+			      struct drm_connector_state *conn_state);
+void intel_psr_get_config(struct intel_encoder *encoder,
+			  struct intel_crtc_state *pipe_config);
+void intel_psr_irq_handler(struct intel_dp *intel_dp, u32 psr_iir);
+void intel_psr_short_pulse(struct intel_dp *intel_dp);
+void intel_psr_wait_for_idle_locked(const struct intel_crtc_state *new_crtc_state);
+bool intel_psr_enabled(struct intel_dp *intel_dp);
+int intel_psr2_sel_fetch_update(struct intel_atomic_state *state,
+				struct intel_crtc *crtc);
+void intel_psr2_program_trans_man_trk_ctl(const struct intel_crtc_state *crtc_state);
+void intel_psr2_program_plane_sel_fetch(struct intel_plane *plane,
+					const struct intel_crtc_state *crtc_state,
+					const struct intel_plane_state *plane_state,
+					int color_plane);
+void intel_psr2_disable_plane_sel_fetch(struct intel_plane *plane,
+					const struct intel_crtc_state *crtc_state);
+void intel_psr_pause(struct intel_dp *intel_dp);
+void intel_psr_resume(struct intel_dp *intel_dp);
+
+void intel_psr_lock(const struct intel_crtc_state *crtc_state);
+void intel_psr_unlock(const struct intel_crtc_state *crtc_state);
+
+#endif /* __INTEL_PSR_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_qp_tables.c b/drivers/gpu/drm/i915/display/intel_qp_tables.c
new file mode 100644
index 000000000..6f8e4ec5c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_qp_tables.c
@@ -0,0 +1,309 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <drm/display/drm_dsc.h>
+
+#include "i915_utils.h"
+#include "intel_qp_tables.h"
+
+/* from BPP 6 to 24 in steps of 0.5 */
+#define RC_RANGE_QP444_8BPC_MAX_NUM_BPP		37
+
+/* from BPP 6 to 30 in steps of 0.5 */
+#define RC_RANGE_QP444_10BPC_MAX_NUM_BPP	49
+
+/* from BPP 6 to 36 in steps of 0.5 */
+#define RC_RANGE_QP444_12BPC_MAX_NUM_BPP	61
+
+/*
+ * These qp tables are as per the C model
+ * and it has the rows pointing to bpps which increment
+ * in steps of 0.5
+ * We do not support fractional bpps as of today,
+ * hence we would skip the fractional bpps during
+ * our references for qp calclulations.
+ */
+static const u8 rc_range_minqp444_8bpc[DSC_NUM_BUF_RANGES][RC_RANGE_QP444_8BPC_MAX_NUM_BPP] = {
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 3, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 3, 3, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 5, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 1, 1, 1, 1, 1, 1, 1, 1,
+	  1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 5, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 1, 1, 1, 1, 1, 1, 1,
+	  1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 },
+	{ 5, 5, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 1, 1, 1, 1, 1,
+	  1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 },
+	{ 5, 5, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2,
+	  2, 2, 2, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0 },
+	{ 5, 5, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2,
+	  2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 0, 0, 0 },
+	{ 6, 5, 5, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
+	  3, 3, 3, 2, 2, 1, 1, 1, 1, 1, 1, 0, 0, 0 },
+	{ 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 3,
+	  3, 3, 3, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 0 },
+	{ 6, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4,
+	  4, 4, 4, 3, 3, 2, 2, 2, 2, 1, 1, 1, 1, 0 },
+	{ 6, 6, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4,
+	  4, 4, 4, 3, 3, 2, 2, 2, 2, 1, 1, 1, 1, 0 },
+	{ 9, 9, 9, 9, 8, 8, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 6, 5,
+	  5, 5, 5, 4, 4, 3, 3, 3, 3, 2, 2, 1, 1, 1 },
+	{ 14, 14, 13, 13, 12, 12, 12, 12, 11, 11, 10, 10, 10, 10, 9, 9, 9, 8, 8,
+	  8, 7, 7, 7, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 3, 3, 3, 3 }
+};
+
+static const u8 rc_range_maxqp444_8bpc[DSC_NUM_BUF_RANGES][RC_RANGE_QP444_8BPC_MAX_NUM_BPP] = {
+	{ 4, 4, 4, 4, 4, 4, 3, 3, 3, 3, 2, 2, 2, 2, 1, 1, 1, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 6, 6, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 2, 2, 1, 1, 1, 1, 1, 1,
+	  1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0 },
+	{ 8, 7, 7, 6, 5, 5, 5, 5, 5, 5, 5, 5, 4, 4, 4, 3, 2, 2, 1, 1, 1, 1, 1,
+	  1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0 },
+	{ 8, 8, 7, 7, 6, 6, 6, 6, 6, 6, 6, 6, 5, 5, 5, 4, 3, 3, 2, 2, 2, 2, 2,
+	  2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0 },
+	{ 9, 8, 8, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 6, 5, 4, 4, 3, 2, 2, 2, 2, 2,
+	  2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 0 },
+	{ 9, 8, 8, 8, 7, 7, 7, 7, 7, 7, 7, 7, 6, 6, 6, 5, 4, 4, 3, 3, 3, 3, 3,
+	  3, 3, 3, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1 },
+	{ 9, 9, 8, 8, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 6, 5, 4, 4, 3, 3, 3, 3, 3,
+	  3, 3, 3, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1 },
+	{ 10, 10, 9, 9, 8, 8, 8, 8, 8, 8, 8, 8, 8, 7, 7, 6, 5, 5, 4, 4, 4, 4, 3,
+	  3, 3, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1 },
+	{ 11, 11, 10, 10, 9, 9, 9, 9, 9, 9, 8, 8, 8, 7, 7, 6, 6, 5, 5, 5, 5, 5,
+	  4, 4, 4, 4, 3, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1 },
+	{ 12, 11, 11, 10, 10, 10, 9, 9, 9, 9, 9, 9, 9, 8, 8, 7, 6, 6, 5, 5, 5,
+	  5, 4, 4, 4, 4, 3, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1 },
+	{ 12, 12, 11, 11, 10, 10, 10, 10, 10, 10, 9, 9, 9, 8, 8, 7, 7, 6, 6, 6,
+	  5, 5, 4, 4, 4, 4, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 1 },
+	{ 12, 12, 12, 11, 11, 11, 10, 10, 10, 10, 9, 9, 9, 9, 8, 8, 8, 7, 7, 7,
+	  6, 6, 5, 5, 5, 5, 4, 4, 3, 3, 3, 3, 2, 2, 2, 2, 1 },
+	{ 12, 12, 12, 12, 11, 11, 11, 11, 11, 10, 10, 9, 9, 9, 8, 8, 8, 7, 7, 7,
+	  6, 6, 5, 5, 5, 5, 4, 4, 3, 3, 3, 3, 2, 2, 2, 2, 1 },
+	{ 13, 13, 13, 13, 12, 12, 11, 11, 11, 11, 10, 10, 10, 10, 9, 9, 8, 8, 8,
+	  8, 7, 7, 6, 6, 6, 6, 5, 5, 4, 4, 4, 4, 3, 3, 2, 2, 2 },
+	{ 15, 15, 14, 14, 13, 13, 13, 13, 12, 12, 11, 11, 11, 11, 10, 10, 10, 9,
+	  9, 9, 8, 8, 8, 8, 7, 7, 6, 6, 6, 6, 5, 5, 5, 4, 4, 4, 4 }
+};
+
+static const u8 rc_range_minqp444_10bpc[DSC_NUM_BUF_RANGES][RC_RANGE_QP444_10BPC_MAX_NUM_BPP] = {
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0 },
+	{ 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1,
+	  1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0 },
+	{ 7, 6, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 4, 4, 4, 3, 3, 3, 3, 3, 3, 2, 2,
+	  2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0 },
+	{ 7, 7, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3,
+	  3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0 },
+	{ 9, 8, 8, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 3,
+	  3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0 },
+	{ 9, 8, 8, 8, 7, 7, 7, 7, 7, 7, 7, 7, 6, 6, 6, 6, 5, 5, 5, 5, 5, 5, 4,
+	  4, 4, 4, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 0, 0, 0,
+	  0, 0, 0 },
+	{ 9, 9, 8, 8, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 6, 6, 5, 5, 5, 5, 5,
+	  4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 1, 1, 1, 1, 0, 0,
+	  0, 0, 0 },
+	{ 9, 9, 8, 8, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 6, 6, 6, 6, 6, 5,
+	  5, 5, 5, 4, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 1, 1, 1, 0,
+	  0, 0, 0 },
+	{ 9, 9, 8, 8, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 6, 6, 6, 6, 5,
+	  5, 5, 5, 5, 5, 5, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 2, 2, 2, 1, 1, 1,
+	  1, 0, 0 },
+	{ 10, 9, 9, 8, 8, 8, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 6, 6, 6, 6, 6,
+	  6, 6, 6, 6, 6, 5, 5, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 1, 1,
+	  1, 1, 0 },
+	{ 10, 10, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 8, 8, 8, 8, 7, 7, 7, 7, 7, 6,
+	  6, 6, 6, 6, 6, 5, 5, 5, 5, 5, 5, 5, 5, 4, 4, 4, 3, 3, 3, 2, 2, 2, 1,
+	  1, 1, 1 },
+	{ 10, 10, 10, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 8, 8, 8, 8, 8,
+	  7, 7, 7, 7, 7, 7, 6, 6, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 3, 3, 3, 3, 2,
+	  2, 1, 1, 1 },
+	{ 10, 10, 10, 10, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 8, 8,
+	  8, 8, 8, 8, 7, 7, 6, 6, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 3, 3, 3, 3, 2,
+	  2, 2, 2, 1 },
+	{ 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 11, 11, 11, 11, 11, 11, 11,
+	  11, 11, 11, 10, 10, 9, 9, 9, 9, 8, 8, 7, 7, 7, 7, 6, 6, 5, 5, 5, 5, 4,
+	  4, 3, 3, 3, 3, 2, 2, 2, 2, 1 },
+	{ 18, 18, 17, 17, 16, 16, 16, 16, 15, 15, 14, 14, 14, 14, 13, 13, 13,
+	  12, 12, 12, 11, 11, 11, 11, 10, 10, 9, 9, 9, 9, 9, 8, 8, 7, 7, 7, 7,
+	  7, 6, 6, 5, 5, 5, 5, 4, 4, 3, 3, 3 }
+};
+
+static const u8 rc_range_maxqp444_10bpc[DSC_NUM_BUF_RANGES][RC_RANGE_QP444_10BPC_MAX_NUM_BPP] = {
+	{ 8, 8, 8, 8, 8, 8, 7, 7, 7, 6, 5, 5, 4, 4, 3, 3, 3, 2, 2, 2, 2, 2, 2,
+	  1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0 },
+	{ 10, 10, 9, 9, 8, 8, 8, 8, 8, 8, 7, 7, 6, 6, 6, 5, 5, 4, 4, 4, 4, 3, 3,
+	  3, 3, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0 },
+	{ 12, 11, 11, 10, 9, 9, 9, 9, 9, 9, 9, 9, 8, 8, 8, 7, 6, 6, 5, 5, 5, 4,
+	  4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 2, 2, 2, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0,
+	  0, 0, 0, 0 },
+	{ 12, 12, 11, 11, 10, 10, 10, 10, 10, 10, 10, 10, 9, 9, 9, 8, 7, 7, 6,
+	  6, 6, 5, 5, 5, 5, 5, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 1, 1, 1, 1,
+	  1, 0, 0, 0, 0, 0, 0 },
+	{ 13, 12, 12, 11, 11, 11, 11, 11, 11, 11, 11, 11, 10, 10, 9, 8, 8, 7,
+	  6, 6, 6, 6, 5, 5, 5, 5, 5, 5, 4, 4, 4, 3, 3, 3, 3, 3, 2, 2, 2, 2, 1,
+	  1, 1, 1, 0, 0, 0, 0, 0 },
+	{ 13, 12, 12, 12, 11, 11, 11, 11, 11, 11, 11, 11, 10, 10, 10, 9, 8, 8,
+	  7, 7, 7, 7, 6, 6, 6, 6, 5, 5, 5, 5, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2,
+	  2, 2, 1, 1, 1, 1, 0, 0 },
+	{ 13, 13, 12, 12, 11, 11, 11, 11, 11, 11, 11, 11, 11, 10, 10, 9, 8, 8,
+	  7, 7, 7, 7, 7, 6, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 2,
+	  2, 2, 2, 1, 1, 1, 1, 1 },
+	{ 14, 14, 13, 13, 12, 12, 12, 12, 12, 12, 12, 12, 12, 11, 11, 10, 9, 9,
+	  8, 8, 8, 8, 7, 7, 7, 7, 6, 6, 6, 5, 5, 5, 4, 4, 4, 4, 4, 4, 3, 3, 3,
+	  3, 2, 2, 2, 1, 1, 1, 1 },
+	{ 15, 15, 14, 14, 13, 13, 13, 13, 13, 13, 12, 12, 12, 11, 11, 10, 10, 9,
+	  9, 9, 9, 9, 8, 8, 8, 8, 7, 7, 6, 5, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 3,
+	  3, 3, 2, 2, 2, 2, 1, 1 },
+	{ 16, 15, 15, 14, 14, 14, 13, 13, 13, 13, 13, 13, 13, 12, 12, 11, 10,
+	  10, 9, 9, 9, 9, 8, 8, 8, 8, 7, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4,
+	  4, 3, 3, 3, 2, 2, 2, 2, 1 },
+	{ 16, 16, 15, 15, 14, 14, 14, 14, 14, 14, 13, 13, 13, 12, 12, 11, 11,
+	  10, 10, 10, 9, 9, 8, 8, 8, 8, 7, 7, 6, 6, 6, 6, 6, 6, 6, 6, 5, 5, 5,
+	  4, 4, 4, 3, 3, 3, 2, 2, 2, 2 },
+	{ 16, 16, 16, 15, 15, 15, 14, 14, 14, 14, 13, 13, 13, 13, 12, 12, 12,
+	  11, 11, 11, 10, 10, 9, 9, 9, 9, 8, 8, 7, 7, 7, 7, 6, 6, 6, 6, 5, 5, 5,
+	  5, 4, 4, 4, 4, 3, 3, 2, 2, 2 },
+	{ 16, 16, 16, 16, 15, 15, 15, 15, 15, 14, 14, 13, 13, 13, 12, 12, 12,
+	  11, 11, 11, 10, 10, 9, 9, 9, 9, 8, 8, 7, 7, 7, 7, 6, 6, 6, 6, 5, 5, 5,
+	  5, 4, 4, 4, 4, 3, 3, 3, 3, 2 },
+	{ 17, 17, 17, 17, 16, 16, 15, 15, 15, 15, 14, 14, 14, 14, 13, 13, 12,
+	  12, 12, 12, 11, 11, 10, 10, 10, 10, 9, 9, 8, 8, 8, 8, 7, 7, 6, 6, 6,
+	  6, 5, 5, 4, 4, 4, 4, 3, 3, 3, 3, 2 },
+	{ 19, 19, 18, 18, 17, 17, 17, 17, 16, 16, 15, 15, 15, 15, 14, 14, 14,
+	  13, 13, 13, 12, 12, 12, 12, 11, 11, 10, 10, 10, 10, 10, 9, 9, 8, 8, 8,
+	  8, 8, 7, 7, 6, 6, 6, 6, 5, 5, 4, 4, 4 }
+};
+
+static const u8 rc_range_minqp444_12bpc[DSC_NUM_BUF_RANGES][RC_RANGE_QP444_12BPC_MAX_NUM_BPP] = {
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2,
+	  1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 11, 10, 10, 9, 8, 8, 8, 8, 8, 8, 8, 8, 7, 7, 7, 6, 5, 5, 4, 4, 4, 3, 3, 3, 3,
+	  3, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 11, 11, 10, 10, 9, 9, 9, 9, 9, 9, 9, 9, 8, 8, 8, 7, 6, 6, 6, 6, 6, 5, 5, 5,
+	  5, 5, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 13, 12, 12, 11, 11, 11, 11, 11, 11, 11, 11, 11, 10, 10, 9, 9, 9, 8, 7, 7, 7,
+	  7, 5, 5, 5, 5, 5, 5, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 1, 1,
+	  1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 13, 12, 12, 12, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 10, 9, 9, 8, 8,
+	  8, 8, 6, 6, 6, 6, 5, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 2, 2, 2,
+	  2, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 13, 13, 12, 12, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 10, 9, 9, 9, 9,
+	  9, 9, 9, 8, 8, 8, 7, 7, 7, 6, 5, 5, 5, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 3, 3,
+	  3, 2, 2, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0 },
+	{ 13, 13, 12, 12, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 10, 10, 10,
+	  10, 10, 10, 9, 9, 9, 9, 8, 8, 8, 7, 7, 7, 6, 6, 6, 6, 6, 6, 5, 5, 5, 5, 4, 4,
+	  4, 3, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 0, 0, 0, 0 },
+	{ 13, 13, 12, 12, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11,
+	  11, 11, 11, 10, 10, 10, 10, 9, 9, 8, 7, 7, 7, 7, 7, 6, 6, 6, 6, 6, 6, 5, 5,
+	  5, 4, 4, 4, 4, 3, 3, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 0, 0 },
+	{ 14, 13, 13, 12, 12, 12, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11,
+	  11, 11, 11, 10, 10, 10, 10, 9, 9, 8, 8, 8, 8, 8, 8, 7, 7, 7, 7, 7, 6, 6, 5,
+	  5, 5, 4, 4, 4, 4, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 1, 1, 0 },
+	{ 14, 14, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 12, 12, 12, 12,
+	  12, 11, 11, 11, 11, 11, 11, 10, 10, 9, 9, 9, 9, 9, 9, 9, 9, 8, 8, 8, 7, 7, 7,
+	  6, 6, 6, 5, 5, 5, 5, 4, 4, 3, 3, 3, 2, 2, 2, 2, 1, 1, 1 },
+	{ 14, 14, 14, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13,
+	  13, 12, 12, 11, 11, 11, 11, 11, 11, 10, 10, 10, 10, 9, 9, 9, 9, 8, 8, 8, 8,
+	  7, 7, 7, 7, 6, 6, 5, 5, 5, 4, 4, 4, 3, 3, 3, 3, 2, 2, 1, 1, 1 },
+	{ 14, 14, 14, 14, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13,
+	  13, 12, 12, 12, 12, 12, 12, 11, 11, 10, 10, 10, 10, 9, 9, 9, 9, 8, 8, 8, 8,
+	  7, 7, 7, 7, 6, 6, 6, 6, 5, 4, 4, 4, 3, 3, 3, 3, 2, 2, 1, 1, 1 },
+	{ 17, 17, 17, 17, 16, 16, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15,
+	  15, 14, 14, 13, 13, 13, 13, 12, 12, 11, 11, 11, 11, 10, 10, 9, 9, 9, 9, 8, 8,
+	  7, 7, 7, 7, 7, 6, 6, 6, 5, 5, 5, 5, 4, 4, 3, 3, 3, 3, 2, 2, 1 },
+	{ 22, 22, 21, 21, 20, 20, 20, 20, 19, 19, 18, 18, 18, 18, 17, 17, 17, 16, 16,
+	  16, 15, 15, 15, 15, 14, 14, 13, 13, 13, 13, 13, 12, 12, 11, 11, 11, 11, 11,
+	  10, 10, 9, 9, 9, 9, 9, 8, 8, 7, 7, 7, 7, 7, 6, 6, 5, 5, 5, 5, 4, 4, 3 }
+};
+
+static const u8 rc_range_maxqp444_12bpc[DSC_NUM_BUF_RANGES][RC_RANGE_QP444_12BPC_MAX_NUM_BPP] = {
+	{ 12, 12, 12, 12, 12, 12, 11, 11, 11, 10, 9, 9, 6, 6, 5, 5, 5, 4, 4, 4, 4, 4,
+	  4, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0,
+	  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 14, 14, 13, 13, 12, 12, 12, 12, 12, 12, 11, 11, 9, 9, 9, 8, 8, 7, 7, 7, 7, 5,
+	  5, 5, 5, 5, 4, 4, 4, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+	  1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 16, 15, 15, 14, 13, 13, 13, 13, 13, 13, 13, 13, 12, 12, 12, 11, 10, 10, 9, 9,
+	  9, 7, 7, 7, 7, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 3, 3, 3, 3, 3, 3, 2, 2, 2, 1, 1,
+	  1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 16, 16, 15, 15, 14, 14, 14, 14, 14, 14, 14, 14, 13, 13, 13, 12, 11, 11, 10,
+	  10, 10, 8, 8, 8, 8, 8, 7, 7, 6, 5, 5, 5, 5, 5, 5, 5, 4, 4, 3, 3, 3, 3, 3, 2,
+	  2, 2, 2, 2, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 17, 16, 16, 15, 15, 15, 15, 15, 15, 15, 15, 15, 14, 14, 13, 12, 12, 11, 10,
+	  10, 10, 10, 8, 8, 8, 8, 8, 8, 7, 7, 7, 6, 6, 5, 5, 5, 4, 4, 4, 4, 3, 3, 3, 3,
+	  2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0 },
+	{ 17, 16, 16, 16, 15, 15, 15, 15, 15, 15, 15, 15, 14, 14, 14, 13, 12, 12, 11,
+	  11, 11, 11, 9, 9, 9, 9, 8, 8, 8, 8, 7, 6, 6, 6, 6, 6, 5, 5, 5, 5, 4, 4, 4, 3,
+	  3, 3, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 0 },
+	{ 17, 17, 16, 16, 15, 15, 15, 15, 15, 15, 15, 15, 15, 14, 14, 13, 12, 12, 11,
+	  11, 11, 11, 11, 10, 10, 10, 9, 9, 9, 8, 7, 7, 7, 7, 7, 7, 7, 6, 6, 6, 5, 5,
+	  5, 5, 4, 4, 4, 3, 3, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 0 },
+	{ 18, 18, 17, 17, 16, 16, 16, 16, 16, 16, 16, 16, 16, 15, 15, 14, 13, 13, 12,
+	  12, 12, 12, 11, 11, 11, 11, 10, 10, 10, 8, 8, 8, 7, 7, 7, 7, 7, 7, 6, 6, 6,
+	  6, 5, 5, 5, 4, 4, 3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 1, 1, 1, 1 },
+	{ 19, 19, 18, 18, 17, 17, 17, 17, 17, 17, 16, 16, 16, 15, 15, 14, 14, 13, 13,
+	  13, 13, 13, 12, 12, 12, 12, 11, 11, 10, 9, 8, 8, 8, 8, 7, 7, 7, 7, 7, 7, 6,
+	  6, 6, 5, 5, 5, 5, 4, 4, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 1, 1 },
+	{ 20, 19, 19, 18, 18, 18, 17, 17, 17, 17, 17, 17, 17, 16, 16, 15, 14, 14, 13,
+	  13, 13, 13, 12, 12, 12, 12, 11, 11, 10, 10, 9, 9, 9, 9, 8, 8, 8, 8, 8, 7, 7,
+	  6, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 2, 2, 1 },
+	{ 20, 20, 19, 19, 18, 18, 18, 18, 18, 18, 17, 17, 17, 16, 16, 15, 15, 14, 14,
+	  14, 13, 13, 12, 12, 12, 12, 11, 11, 10, 10, 10, 10, 10, 10, 10, 10, 9, 9, 9,
+	  8, 8, 8, 7, 7, 7, 6, 6, 6, 6, 5, 5, 4, 4, 4, 3, 3, 3, 3, 2, 2, 2 },
+	{ 20, 20, 20, 19, 19, 19, 18, 18, 18, 18, 17, 17, 17, 17, 16, 16, 16, 15, 15,
+	  15, 14, 14, 13, 13, 13, 13, 12, 12, 11, 11, 11, 11, 10, 10, 10, 10, 9, 9, 9,
+	  9, 8, 8, 8, 8, 7, 7, 6, 6, 6, 5, 5, 5, 4, 4, 4, 4, 3, 3, 2, 2, 2 },
+	{ 20, 20, 20, 20, 19, 19, 19, 19, 19, 18, 18, 17, 17, 17, 16, 16, 16, 15, 15,
+	  15, 14, 14, 13, 13, 13, 13, 12, 12, 11, 11, 11, 11, 10, 10, 10, 10, 9, 9, 9,
+	  9, 8, 8, 8, 8, 7, 7, 7, 7, 6, 5, 5, 5, 4, 4, 4, 4, 3, 3, 2, 2, 2 },
+	{ 21, 21, 21, 21, 20, 20, 19, 19, 19, 19, 18, 18, 18, 18, 17, 17, 16, 16, 16,
+	  16, 15, 15, 14, 14, 14, 14, 13, 13, 12, 12, 12, 12, 11, 11, 10, 10, 10, 10,
+	  9, 9, 8, 8, 8, 8, 8, 7, 7, 7, 6, 6, 6, 6, 5, 5, 4, 4, 4, 4, 3, 3, 2 },
+	{ 23, 23, 22, 22, 21, 21, 21, 21, 20, 20, 19, 19, 19, 19, 18, 18, 18, 17, 17,
+	  17, 16, 16, 16, 16, 15, 15, 14, 14, 14, 14, 14, 13, 13, 12, 12, 12, 12, 12,
+	  11, 11, 10, 10, 10, 10, 10, 9, 9, 8, 8, 8, 8, 8, 7, 7, 6, 6, 6, 6, 5, 5, 4 }
+};
+
+#define PARAM_TABLE(_minmax, _bpc, _row, _col)  do { \
+	if (bpc == (_bpc)) \
+		return rc_range_##_minmax##qp444_##_bpc##bpc[_row][_col]; \
+} while (0)
+
+u8 intel_lookup_range_min_qp(int bpc, int buf_i, int bpp_i)
+{
+	PARAM_TABLE(min, 8, buf_i, bpp_i);
+	PARAM_TABLE(min, 10, buf_i, bpp_i);
+	PARAM_TABLE(min, 12, buf_i, bpp_i);
+
+	MISSING_CASE(bpc);
+	return 0;
+}
+
+u8 intel_lookup_range_max_qp(int bpc, int buf_i, int bpp_i)
+{
+	PARAM_TABLE(max, 8, buf_i, bpp_i);
+	PARAM_TABLE(max, 10, buf_i, bpp_i);
+	PARAM_TABLE(max, 12, buf_i, bpp_i);
+
+	MISSING_CASE(bpc);
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_qp_tables.h b/drivers/gpu/drm/i915/display/intel_qp_tables.h
new file mode 100644
index 000000000..9fb3c36bd
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_qp_tables.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _INTEL_QP_TABLES_H_
+#define _INTEL_QP_TABLES_H_
+
+#include <linux/types.h>
+
+u8 intel_lookup_range_min_qp(int bpc, int buf_i, int bpp_i);
+u8 intel_lookup_range_max_qp(int bpc, int buf_i, int bpp_i);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/intel_quirks.c b/drivers/gpu/drm/i915/display/intel_quirks.c
new file mode 100644
index 000000000..a280448df
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_quirks.c
@@ -0,0 +1,230 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/dmi.h>
+
+#include "i915_drv.h"
+#include "intel_display_types.h"
+#include "intel_quirks.h"
+
+static void intel_set_quirk(struct drm_i915_private *i915, enum intel_quirk_id quirk)
+{
+	i915->display.quirks.mask |= BIT(quirk);
+}
+
+/*
+ * Some machines (Lenovo U160) do not work with SSC on LVDS for some reason
+ */
+static void quirk_ssc_force_disable(struct drm_i915_private *i915)
+{
+	intel_set_quirk(i915, QUIRK_LVDS_SSC_DISABLE);
+	drm_info(&i915->drm, "applying lvds SSC disable quirk\n");
+}
+
+/*
+ * A machine (e.g. Acer Aspire 5734Z) may need to invert the panel backlight
+ * brightness value
+ */
+static void quirk_invert_brightness(struct drm_i915_private *i915)
+{
+	intel_set_quirk(i915, QUIRK_INVERT_BRIGHTNESS);
+	drm_info(&i915->drm, "applying inverted panel brightness quirk\n");
+}
+
+/* Some VBT's incorrectly indicate no backlight is present */
+static void quirk_backlight_present(struct drm_i915_private *i915)
+{
+	intel_set_quirk(i915, QUIRK_BACKLIGHT_PRESENT);
+	drm_info(&i915->drm, "applying backlight present quirk\n");
+}
+
+/* Toshiba Satellite P50-C-18C requires T12 delay to be min 800ms
+ * which is 300 ms greater than eDP spec T12 min.
+ */
+static void quirk_increase_t12_delay(struct drm_i915_private *i915)
+{
+	intel_set_quirk(i915, QUIRK_INCREASE_T12_DELAY);
+	drm_info(&i915->drm, "Applying T12 delay quirk\n");
+}
+
+/*
+ * GeminiLake NUC HDMI outputs require additional off time
+ * this allows the onboard retimer to correctly sync to signal
+ */
+static void quirk_increase_ddi_disabled_time(struct drm_i915_private *i915)
+{
+	intel_set_quirk(i915, QUIRK_INCREASE_DDI_DISABLED_TIME);
+	drm_info(&i915->drm, "Applying Increase DDI Disabled quirk\n");
+}
+
+static void quirk_no_pps_backlight_power_hook(struct drm_i915_private *i915)
+{
+	intel_set_quirk(i915, QUIRK_NO_PPS_BACKLIGHT_POWER_HOOK);
+	drm_info(&i915->drm, "Applying no pps backlight power quirk\n");
+}
+
+struct intel_quirk {
+	int device;
+	int subsystem_vendor;
+	int subsystem_device;
+	void (*hook)(struct drm_i915_private *i915);
+};
+
+/* For systems that don't have a meaningful PCI subdevice/subvendor ID */
+struct intel_dmi_quirk {
+	void (*hook)(struct drm_i915_private *i915);
+	const struct dmi_system_id (*dmi_id_list)[];
+};
+
+static int intel_dmi_reverse_brightness(const struct dmi_system_id *id)
+{
+	DRM_INFO("Backlight polarity reversed on %s\n", id->ident);
+	return 1;
+}
+
+static int intel_dmi_no_pps_backlight(const struct dmi_system_id *id)
+{
+	DRM_INFO("No pps backlight support on %s\n", id->ident);
+	return 1;
+}
+
+static const struct intel_dmi_quirk intel_dmi_quirks[] = {
+	{
+		.dmi_id_list = &(const struct dmi_system_id[]) {
+			{
+				.callback = intel_dmi_reverse_brightness,
+				.ident = "NCR Corporation",
+				.matches = {DMI_MATCH(DMI_SYS_VENDOR, "NCR Corporation"),
+					    DMI_MATCH(DMI_PRODUCT_NAME, ""),
+				},
+			},
+			{
+				.callback = intel_dmi_reverse_brightness,
+				.ident = "Thundersoft TST178 tablet",
+				/* DMI strings are too generic, also match on BIOS date */
+				.matches = {DMI_EXACT_MATCH(DMI_BOARD_VENDOR, "AMI Corporation"),
+					    DMI_EXACT_MATCH(DMI_BOARD_NAME, "Aptio CRB"),
+					    DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "To be filled by O.E.M."),
+					    DMI_EXACT_MATCH(DMI_BIOS_DATE, "04/15/2014"),
+				},
+			},
+			{ }  /* terminating entry */
+		},
+		.hook = quirk_invert_brightness,
+	},
+	{
+		.dmi_id_list = &(const struct dmi_system_id[]) {
+			{
+				.callback = intel_dmi_no_pps_backlight,
+				.ident = "Google Lillipup sku524294",
+				.matches = {DMI_EXACT_MATCH(DMI_BOARD_VENDOR, "Google"),
+					    DMI_EXACT_MATCH(DMI_BOARD_NAME, "Lindar"),
+					    DMI_EXACT_MATCH(DMI_PRODUCT_SKU, "sku524294"),
+				},
+			},
+			{
+				.callback = intel_dmi_no_pps_backlight,
+				.ident = "Google Lillipup sku524295",
+				.matches = {DMI_EXACT_MATCH(DMI_BOARD_VENDOR, "Google"),
+					    DMI_EXACT_MATCH(DMI_BOARD_NAME, "Lindar"),
+					    DMI_EXACT_MATCH(DMI_PRODUCT_SKU, "sku524295"),
+				},
+			},
+			{ }
+		},
+		.hook = quirk_no_pps_backlight_power_hook,
+	},
+};
+
+static struct intel_quirk intel_quirks[] = {
+	/* Lenovo U160 cannot use SSC on LVDS */
+	{ 0x0046, 0x17aa, 0x3920, quirk_ssc_force_disable },
+
+	/* Sony Vaio Y cannot use SSC on LVDS */
+	{ 0x0046, 0x104d, 0x9076, quirk_ssc_force_disable },
+
+	/* Acer Aspire 5734Z must invert backlight brightness */
+	{ 0x2a42, 0x1025, 0x0459, quirk_invert_brightness },
+
+	/* Acer/eMachines G725 */
+	{ 0x2a42, 0x1025, 0x0210, quirk_invert_brightness },
+
+	/* Acer/eMachines e725 */
+	{ 0x2a42, 0x1025, 0x0212, quirk_invert_brightness },
+
+	/* Acer/Packard Bell NCL20 */
+	{ 0x2a42, 0x1025, 0x034b, quirk_invert_brightness },
+
+	/* Acer Aspire 4736Z */
+	{ 0x2a42, 0x1025, 0x0260, quirk_invert_brightness },
+
+	/* Acer Aspire 5336 */
+	{ 0x2a42, 0x1025, 0x048a, quirk_invert_brightness },
+
+	/* Acer C720 and C720P Chromebooks (Celeron 2955U) have backlights */
+	{ 0x0a06, 0x1025, 0x0a11, quirk_backlight_present },
+
+	/* Acer C720 Chromebook (Core i3 4005U) */
+	{ 0x0a16, 0x1025, 0x0a11, quirk_backlight_present },
+
+	/* Apple Macbook 2,1 (Core 2 T7400) */
+	{ 0x27a2, 0x8086, 0x7270, quirk_backlight_present },
+
+	/* Apple Macbook 4,1 */
+	{ 0x2a02, 0x106b, 0x00a1, quirk_backlight_present },
+
+	/* Toshiba CB35 Chromebook (Celeron 2955U) */
+	{ 0x0a06, 0x1179, 0x0a88, quirk_backlight_present },
+
+	/* HP Chromebook 14 (Celeron 2955U) */
+	{ 0x0a06, 0x103c, 0x21ed, quirk_backlight_present },
+
+	/* Dell Chromebook 11 */
+	{ 0x0a06, 0x1028, 0x0a35, quirk_backlight_present },
+
+	/* Dell Chromebook 11 (2015 version) */
+	{ 0x0a16, 0x1028, 0x0a35, quirk_backlight_present },
+
+	/* Toshiba Satellite P50-C-18C */
+	{ 0x191B, 0x1179, 0xF840, quirk_increase_t12_delay },
+
+	/* GeminiLake NUC */
+	{ 0x3185, 0x8086, 0x2072, quirk_increase_ddi_disabled_time },
+	{ 0x3184, 0x8086, 0x2072, quirk_increase_ddi_disabled_time },
+	/* ASRock ITX*/
+	{ 0x3185, 0x1849, 0x2212, quirk_increase_ddi_disabled_time },
+	{ 0x3184, 0x1849, 0x2212, quirk_increase_ddi_disabled_time },
+	/* ECS Liva Q2 */
+	{ 0x3185, 0x1019, 0xa94d, quirk_increase_ddi_disabled_time },
+	{ 0x3184, 0x1019, 0xa94d, quirk_increase_ddi_disabled_time },
+	/* HP Notebook - 14-r206nv */
+	{ 0x0f31, 0x103c, 0x220f, quirk_invert_brightness },
+};
+
+void intel_init_quirks(struct drm_i915_private *i915)
+{
+	struct pci_dev *d = to_pci_dev(i915->drm.dev);
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(intel_quirks); i++) {
+		struct intel_quirk *q = &intel_quirks[i];
+
+		if (d->device == q->device &&
+		    (d->subsystem_vendor == q->subsystem_vendor ||
+		     q->subsystem_vendor == PCI_ANY_ID) &&
+		    (d->subsystem_device == q->subsystem_device ||
+		     q->subsystem_device == PCI_ANY_ID))
+			q->hook(i915);
+	}
+	for (i = 0; i < ARRAY_SIZE(intel_dmi_quirks); i++) {
+		if (dmi_check_system(*intel_dmi_quirks[i].dmi_id_list) != 0)
+			intel_dmi_quirks[i].hook(i915);
+	}
+}
+
+bool intel_has_quirk(struct drm_i915_private *i915, enum intel_quirk_id quirk)
+{
+	return i915->display.quirks.mask & BIT(quirk);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_quirks.h b/drivers/gpu/drm/i915/display/intel_quirks.h
new file mode 100644
index 000000000..10a4d1631
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_quirks.h
@@ -0,0 +1,25 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_QUIRKS_H__
+#define __INTEL_QUIRKS_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+
+enum intel_quirk_id {
+	QUIRK_BACKLIGHT_PRESENT,
+	QUIRK_INCREASE_DDI_DISABLED_TIME,
+	QUIRK_INCREASE_T12_DELAY,
+	QUIRK_INVERT_BRIGHTNESS,
+	QUIRK_LVDS_SSC_DISABLE,
+	QUIRK_NO_PPS_BACKLIGHT_POWER_HOOK,
+};
+
+void intel_init_quirks(struct drm_i915_private *i915);
+bool intel_has_quirk(struct drm_i915_private *i915, enum intel_quirk_id quirk);
+
+#endif /* __INTEL_QUIRKS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_sdvo.c b/drivers/gpu/drm/i915/display/intel_sdvo.c
new file mode 100644
index 000000000..8294dddfd
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_sdvo.c
@@ -0,0 +1,3446 @@
+/*
+ * Copyright 2006 Dave Airlie <airlied@linux.ie>
+ * Copyright © 2006-2007 Intel Corporation
+ *   Jesse Barnes <jesse.barnes@intel.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Eric Anholt <eric@anholt.net>
+ */
+
+#include <linux/delay.h>
+#include <linux/export.h>
+#include <linux/i2c.h>
+#include <linux/slab.h>
+
+#include <drm/display/drm_hdmi_helper.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_edid.h>
+
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_connector.h"
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_fifo_underrun.h"
+#include "intel_gmbus.h"
+#include "intel_hdmi.h"
+#include "intel_hotplug.h"
+#include "intel_panel.h"
+#include "intel_sdvo.h"
+#include "intel_sdvo_regs.h"
+
+#define SDVO_TMDS_MASK (SDVO_OUTPUT_TMDS0 | SDVO_OUTPUT_TMDS1)
+#define SDVO_RGB_MASK  (SDVO_OUTPUT_RGB0 | SDVO_OUTPUT_RGB1)
+#define SDVO_LVDS_MASK (SDVO_OUTPUT_LVDS0 | SDVO_OUTPUT_LVDS1)
+#define SDVO_TV_MASK   (SDVO_OUTPUT_CVBS0 | SDVO_OUTPUT_SVID0 | SDVO_OUTPUT_YPRPB0)
+
+#define SDVO_OUTPUT_MASK (SDVO_TMDS_MASK | SDVO_RGB_MASK | SDVO_LVDS_MASK |\
+			SDVO_TV_MASK)
+
+#define IS_TV(c)	(c->output_flag & SDVO_TV_MASK)
+#define IS_TMDS(c)	(c->output_flag & SDVO_TMDS_MASK)
+#define IS_LVDS(c)	(c->output_flag & SDVO_LVDS_MASK)
+#define IS_TV_OR_LVDS(c) (c->output_flag & (SDVO_TV_MASK | SDVO_LVDS_MASK))
+#define IS_DIGITAL(c) (c->output_flag & (SDVO_TMDS_MASK | SDVO_LVDS_MASK))
+
+
+static const char * const tv_format_names[] = {
+	"NTSC_M"   , "NTSC_J"  , "NTSC_443",
+	"PAL_B"    , "PAL_D"   , "PAL_G"   ,
+	"PAL_H"    , "PAL_I"   , "PAL_M"   ,
+	"PAL_N"    , "PAL_NC"  , "PAL_60"  ,
+	"SECAM_B"  , "SECAM_D" , "SECAM_G" ,
+	"SECAM_K"  , "SECAM_K1", "SECAM_L" ,
+	"SECAM_60"
+};
+
+#define TV_FORMAT_NUM  ARRAY_SIZE(tv_format_names)
+
+struct intel_sdvo {
+	struct intel_encoder base;
+
+	struct i2c_adapter *i2c;
+	u8 slave_addr;
+
+	struct i2c_adapter ddc;
+
+	/* Register for the SDVO device: SDVOB or SDVOC */
+	i915_reg_t sdvo_reg;
+
+	/* Active outputs controlled by this SDVO output */
+	u16 controlled_output;
+
+	/*
+	 * Capabilities of the SDVO device returned by
+	 * intel_sdvo_get_capabilities()
+	 */
+	struct intel_sdvo_caps caps;
+
+	u8 colorimetry_cap;
+
+	/* Pixel clock limitations reported by the SDVO device, in kHz */
+	int pixel_clock_min, pixel_clock_max;
+
+	/*
+	* For multiple function SDVO device,
+	* this is for current attached outputs.
+	*/
+	u16 attached_output;
+
+	/*
+	 * Hotplug activation bits for this device
+	 */
+	u16 hotplug_active;
+
+	enum port port;
+
+	bool has_hdmi_audio;
+
+	/* DDC bus used by this SDVO encoder */
+	u8 ddc_bus;
+
+	/*
+	 * the sdvo flag gets lost in round trip: dtd->adjusted_mode->dtd
+	 */
+	u8 dtd_sdvo_flags;
+};
+
+struct intel_sdvo_connector {
+	struct intel_connector base;
+
+	/* Mark the type of connector */
+	u16 output_flag;
+
+	/* This contains all current supported TV format */
+	u8 tv_format_supported[TV_FORMAT_NUM];
+	int   format_supported_num;
+	struct drm_property *tv_format;
+
+	/* add the property for the SDVO-TV */
+	struct drm_property *left;
+	struct drm_property *right;
+	struct drm_property *top;
+	struct drm_property *bottom;
+	struct drm_property *hpos;
+	struct drm_property *vpos;
+	struct drm_property *contrast;
+	struct drm_property *saturation;
+	struct drm_property *hue;
+	struct drm_property *sharpness;
+	struct drm_property *flicker_filter;
+	struct drm_property *flicker_filter_adaptive;
+	struct drm_property *flicker_filter_2d;
+	struct drm_property *tv_chroma_filter;
+	struct drm_property *tv_luma_filter;
+	struct drm_property *dot_crawl;
+
+	/* add the property for the SDVO-TV/LVDS */
+	struct drm_property *brightness;
+
+	/* this is to get the range of margin.*/
+	u32 max_hscan, max_vscan;
+
+	/**
+	 * This is set if we treat the device as HDMI, instead of DVI.
+	 */
+	bool is_hdmi;
+};
+
+struct intel_sdvo_connector_state {
+	/* base.base: tv.saturation/contrast/hue/brightness */
+	struct intel_digital_connector_state base;
+
+	struct {
+		unsigned overscan_h, overscan_v, hpos, vpos, sharpness;
+		unsigned flicker_filter, flicker_filter_2d, flicker_filter_adaptive;
+		unsigned chroma_filter, luma_filter, dot_crawl;
+	} tv;
+};
+
+static struct intel_sdvo *to_sdvo(struct intel_encoder *encoder)
+{
+	return container_of(encoder, struct intel_sdvo, base);
+}
+
+static struct intel_sdvo *intel_attached_sdvo(struct intel_connector *connector)
+{
+	return to_sdvo(intel_attached_encoder(connector));
+}
+
+static struct intel_sdvo_connector *
+to_intel_sdvo_connector(struct drm_connector *connector)
+{
+	return container_of(connector, struct intel_sdvo_connector, base.base);
+}
+
+#define to_intel_sdvo_connector_state(conn_state) \
+	container_of((conn_state), struct intel_sdvo_connector_state, base.base)
+
+static bool
+intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, u16 flags);
+static bool
+intel_sdvo_tv_create_property(struct intel_sdvo *intel_sdvo,
+			      struct intel_sdvo_connector *intel_sdvo_connector,
+			      int type);
+static bool
+intel_sdvo_create_enhance_property(struct intel_sdvo *intel_sdvo,
+				   struct intel_sdvo_connector *intel_sdvo_connector);
+
+/*
+ * Writes the SDVOB or SDVOC with the given value, but always writes both
+ * SDVOB and SDVOC to work around apparent hardware issues (according to
+ * comments in the BIOS).
+ */
+static void intel_sdvo_write_sdvox(struct intel_sdvo *intel_sdvo, u32 val)
+{
+	struct drm_device *dev = intel_sdvo->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 bval = val, cval = val;
+	int i;
+
+	if (HAS_PCH_SPLIT(dev_priv)) {
+		intel_de_write(dev_priv, intel_sdvo->sdvo_reg, val);
+		intel_de_posting_read(dev_priv, intel_sdvo->sdvo_reg);
+		/*
+		 * HW workaround, need to write this twice for issue
+		 * that may result in first write getting masked.
+		 */
+		if (HAS_PCH_IBX(dev_priv)) {
+			intel_de_write(dev_priv, intel_sdvo->sdvo_reg, val);
+			intel_de_posting_read(dev_priv, intel_sdvo->sdvo_reg);
+		}
+		return;
+	}
+
+	if (intel_sdvo->port == PORT_B)
+		cval = intel_de_read(dev_priv, GEN3_SDVOC);
+	else
+		bval = intel_de_read(dev_priv, GEN3_SDVOB);
+
+	/*
+	 * Write the registers twice for luck. Sometimes,
+	 * writing them only once doesn't appear to 'stick'.
+	 * The BIOS does this too. Yay, magic
+	 */
+	for (i = 0; i < 2; i++) {
+		intel_de_write(dev_priv, GEN3_SDVOB, bval);
+		intel_de_posting_read(dev_priv, GEN3_SDVOB);
+
+		intel_de_write(dev_priv, GEN3_SDVOC, cval);
+		intel_de_posting_read(dev_priv, GEN3_SDVOC);
+	}
+}
+
+static bool intel_sdvo_read_byte(struct intel_sdvo *intel_sdvo, u8 addr, u8 *ch)
+{
+	struct i2c_msg msgs[] = {
+		{
+			.addr = intel_sdvo->slave_addr,
+			.flags = 0,
+			.len = 1,
+			.buf = &addr,
+		},
+		{
+			.addr = intel_sdvo->slave_addr,
+			.flags = I2C_M_RD,
+			.len = 1,
+			.buf = ch,
+		}
+	};
+	int ret;
+
+	if ((ret = i2c_transfer(intel_sdvo->i2c, msgs, 2)) == 2)
+		return true;
+
+	DRM_DEBUG_KMS("i2c transfer returned %d\n", ret);
+	return false;
+}
+
+#define SDVO_CMD_NAME_ENTRY(cmd_) { .cmd = SDVO_CMD_ ## cmd_, .name = #cmd_ }
+
+/** Mapping of command numbers to names, for debug output */
+static const struct {
+	u8 cmd;
+	const char *name;
+} __packed sdvo_cmd_names[] = {
+	SDVO_CMD_NAME_ENTRY(RESET),
+	SDVO_CMD_NAME_ENTRY(GET_DEVICE_CAPS),
+	SDVO_CMD_NAME_ENTRY(GET_FIRMWARE_REV),
+	SDVO_CMD_NAME_ENTRY(GET_TRAINED_INPUTS),
+	SDVO_CMD_NAME_ENTRY(GET_ACTIVE_OUTPUTS),
+	SDVO_CMD_NAME_ENTRY(SET_ACTIVE_OUTPUTS),
+	SDVO_CMD_NAME_ENTRY(GET_IN_OUT_MAP),
+	SDVO_CMD_NAME_ENTRY(SET_IN_OUT_MAP),
+	SDVO_CMD_NAME_ENTRY(GET_ATTACHED_DISPLAYS),
+	SDVO_CMD_NAME_ENTRY(GET_HOT_PLUG_SUPPORT),
+	SDVO_CMD_NAME_ENTRY(SET_ACTIVE_HOT_PLUG),
+	SDVO_CMD_NAME_ENTRY(GET_ACTIVE_HOT_PLUG),
+	SDVO_CMD_NAME_ENTRY(GET_INTERRUPT_EVENT_SOURCE),
+	SDVO_CMD_NAME_ENTRY(SET_TARGET_INPUT),
+	SDVO_CMD_NAME_ENTRY(SET_TARGET_OUTPUT),
+	SDVO_CMD_NAME_ENTRY(GET_INPUT_TIMINGS_PART1),
+	SDVO_CMD_NAME_ENTRY(GET_INPUT_TIMINGS_PART2),
+	SDVO_CMD_NAME_ENTRY(SET_INPUT_TIMINGS_PART1),
+	SDVO_CMD_NAME_ENTRY(SET_INPUT_TIMINGS_PART2),
+	SDVO_CMD_NAME_ENTRY(SET_OUTPUT_TIMINGS_PART1),
+	SDVO_CMD_NAME_ENTRY(SET_OUTPUT_TIMINGS_PART2),
+	SDVO_CMD_NAME_ENTRY(GET_OUTPUT_TIMINGS_PART1),
+	SDVO_CMD_NAME_ENTRY(GET_OUTPUT_TIMINGS_PART2),
+	SDVO_CMD_NAME_ENTRY(CREATE_PREFERRED_INPUT_TIMING),
+	SDVO_CMD_NAME_ENTRY(GET_PREFERRED_INPUT_TIMING_PART1),
+	SDVO_CMD_NAME_ENTRY(GET_PREFERRED_INPUT_TIMING_PART2),
+	SDVO_CMD_NAME_ENTRY(GET_INPUT_PIXEL_CLOCK_RANGE),
+	SDVO_CMD_NAME_ENTRY(GET_OUTPUT_PIXEL_CLOCK_RANGE),
+	SDVO_CMD_NAME_ENTRY(GET_SUPPORTED_CLOCK_RATE_MULTS),
+	SDVO_CMD_NAME_ENTRY(GET_CLOCK_RATE_MULT),
+	SDVO_CMD_NAME_ENTRY(SET_CLOCK_RATE_MULT),
+	SDVO_CMD_NAME_ENTRY(GET_SUPPORTED_TV_FORMATS),
+	SDVO_CMD_NAME_ENTRY(GET_TV_FORMAT),
+	SDVO_CMD_NAME_ENTRY(SET_TV_FORMAT),
+	SDVO_CMD_NAME_ENTRY(GET_SUPPORTED_POWER_STATES),
+	SDVO_CMD_NAME_ENTRY(GET_POWER_STATE),
+	SDVO_CMD_NAME_ENTRY(SET_ENCODER_POWER_STATE),
+	SDVO_CMD_NAME_ENTRY(SET_DISPLAY_POWER_STATE),
+	SDVO_CMD_NAME_ENTRY(SET_CONTROL_BUS_SWITCH),
+	SDVO_CMD_NAME_ENTRY(GET_SDTV_RESOLUTION_SUPPORT),
+	SDVO_CMD_NAME_ENTRY(GET_SCALED_HDTV_RESOLUTION_SUPPORT),
+	SDVO_CMD_NAME_ENTRY(GET_SUPPORTED_ENHANCEMENTS),
+
+	/* Add the op code for SDVO enhancements */
+	SDVO_CMD_NAME_ENTRY(GET_MAX_HPOS),
+	SDVO_CMD_NAME_ENTRY(GET_HPOS),
+	SDVO_CMD_NAME_ENTRY(SET_HPOS),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_VPOS),
+	SDVO_CMD_NAME_ENTRY(GET_VPOS),
+	SDVO_CMD_NAME_ENTRY(SET_VPOS),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_SATURATION),
+	SDVO_CMD_NAME_ENTRY(GET_SATURATION),
+	SDVO_CMD_NAME_ENTRY(SET_SATURATION),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_HUE),
+	SDVO_CMD_NAME_ENTRY(GET_HUE),
+	SDVO_CMD_NAME_ENTRY(SET_HUE),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_CONTRAST),
+	SDVO_CMD_NAME_ENTRY(GET_CONTRAST),
+	SDVO_CMD_NAME_ENTRY(SET_CONTRAST),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_BRIGHTNESS),
+	SDVO_CMD_NAME_ENTRY(GET_BRIGHTNESS),
+	SDVO_CMD_NAME_ENTRY(SET_BRIGHTNESS),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_OVERSCAN_H),
+	SDVO_CMD_NAME_ENTRY(GET_OVERSCAN_H),
+	SDVO_CMD_NAME_ENTRY(SET_OVERSCAN_H),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_OVERSCAN_V),
+	SDVO_CMD_NAME_ENTRY(GET_OVERSCAN_V),
+	SDVO_CMD_NAME_ENTRY(SET_OVERSCAN_V),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_FLICKER_FILTER),
+	SDVO_CMD_NAME_ENTRY(GET_FLICKER_FILTER),
+	SDVO_CMD_NAME_ENTRY(SET_FLICKER_FILTER),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_FLICKER_FILTER_ADAPTIVE),
+	SDVO_CMD_NAME_ENTRY(GET_FLICKER_FILTER_ADAPTIVE),
+	SDVO_CMD_NAME_ENTRY(SET_FLICKER_FILTER_ADAPTIVE),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_FLICKER_FILTER_2D),
+	SDVO_CMD_NAME_ENTRY(GET_FLICKER_FILTER_2D),
+	SDVO_CMD_NAME_ENTRY(SET_FLICKER_FILTER_2D),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_SHARPNESS),
+	SDVO_CMD_NAME_ENTRY(GET_SHARPNESS),
+	SDVO_CMD_NAME_ENTRY(SET_SHARPNESS),
+	SDVO_CMD_NAME_ENTRY(GET_DOT_CRAWL),
+	SDVO_CMD_NAME_ENTRY(SET_DOT_CRAWL),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_TV_CHROMA_FILTER),
+	SDVO_CMD_NAME_ENTRY(GET_TV_CHROMA_FILTER),
+	SDVO_CMD_NAME_ENTRY(SET_TV_CHROMA_FILTER),
+	SDVO_CMD_NAME_ENTRY(GET_MAX_TV_LUMA_FILTER),
+	SDVO_CMD_NAME_ENTRY(GET_TV_LUMA_FILTER),
+	SDVO_CMD_NAME_ENTRY(SET_TV_LUMA_FILTER),
+
+	/* HDMI op code */
+	SDVO_CMD_NAME_ENTRY(GET_SUPP_ENCODE),
+	SDVO_CMD_NAME_ENTRY(GET_ENCODE),
+	SDVO_CMD_NAME_ENTRY(SET_ENCODE),
+	SDVO_CMD_NAME_ENTRY(SET_PIXEL_REPLI),
+	SDVO_CMD_NAME_ENTRY(GET_PIXEL_REPLI),
+	SDVO_CMD_NAME_ENTRY(GET_COLORIMETRY_CAP),
+	SDVO_CMD_NAME_ENTRY(SET_COLORIMETRY),
+	SDVO_CMD_NAME_ENTRY(GET_COLORIMETRY),
+	SDVO_CMD_NAME_ENTRY(GET_AUDIO_ENCRYPT_PREFER),
+	SDVO_CMD_NAME_ENTRY(SET_AUDIO_STAT),
+	SDVO_CMD_NAME_ENTRY(GET_AUDIO_STAT),
+	SDVO_CMD_NAME_ENTRY(GET_HBUF_INDEX),
+	SDVO_CMD_NAME_ENTRY(SET_HBUF_INDEX),
+	SDVO_CMD_NAME_ENTRY(GET_HBUF_INFO),
+	SDVO_CMD_NAME_ENTRY(GET_HBUF_AV_SPLIT),
+	SDVO_CMD_NAME_ENTRY(SET_HBUF_AV_SPLIT),
+	SDVO_CMD_NAME_ENTRY(GET_HBUF_TXRATE),
+	SDVO_CMD_NAME_ENTRY(SET_HBUF_TXRATE),
+	SDVO_CMD_NAME_ENTRY(SET_HBUF_DATA),
+	SDVO_CMD_NAME_ENTRY(GET_HBUF_DATA),
+};
+
+#undef SDVO_CMD_NAME_ENTRY
+
+static const char *sdvo_cmd_name(u8 cmd)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(sdvo_cmd_names); i++) {
+		if (cmd == sdvo_cmd_names[i].cmd)
+			return sdvo_cmd_names[i].name;
+	}
+
+	return NULL;
+}
+
+#define SDVO_NAME(svdo) ((svdo)->port == PORT_B ? "SDVOB" : "SDVOC")
+
+static void intel_sdvo_debug_write(struct intel_sdvo *intel_sdvo, u8 cmd,
+				   const void *args, int args_len)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_sdvo->base.base.dev);
+	const char *cmd_name;
+	int i, pos = 0;
+	char buffer[64];
+
+#define BUF_PRINT(args...) \
+	pos += snprintf(buffer + pos, max_t(int, sizeof(buffer) - pos, 0), args)
+
+	for (i = 0; i < args_len; i++) {
+		BUF_PRINT("%02X ", ((u8 *)args)[i]);
+	}
+	for (; i < 8; i++) {
+		BUF_PRINT("   ");
+	}
+
+	cmd_name = sdvo_cmd_name(cmd);
+	if (cmd_name)
+		BUF_PRINT("(%s)", cmd_name);
+	else
+		BUF_PRINT("(%02X)", cmd);
+
+	drm_WARN_ON(&dev_priv->drm, pos >= sizeof(buffer) - 1);
+#undef BUF_PRINT
+
+	DRM_DEBUG_KMS("%s: W: %02X %s\n", SDVO_NAME(intel_sdvo), cmd, buffer);
+}
+
+static const char * const cmd_status_names[] = {
+	[SDVO_CMD_STATUS_POWER_ON] = "Power on",
+	[SDVO_CMD_STATUS_SUCCESS] = "Success",
+	[SDVO_CMD_STATUS_NOTSUPP] = "Not supported",
+	[SDVO_CMD_STATUS_INVALID_ARG] = "Invalid arg",
+	[SDVO_CMD_STATUS_PENDING] = "Pending",
+	[SDVO_CMD_STATUS_TARGET_NOT_SPECIFIED] = "Target not specified",
+	[SDVO_CMD_STATUS_SCALING_NOT_SUPP] = "Scaling not supported",
+};
+
+static const char *sdvo_cmd_status(u8 status)
+{
+	if (status < ARRAY_SIZE(cmd_status_names))
+		return cmd_status_names[status];
+	else
+		return NULL;
+}
+
+static bool __intel_sdvo_write_cmd(struct intel_sdvo *intel_sdvo, u8 cmd,
+				   const void *args, int args_len,
+				   bool unlocked)
+{
+	u8 *buf, status;
+	struct i2c_msg *msgs;
+	int i, ret = true;
+
+	/* Would be simpler to allocate both in one go ? */
+	buf = kzalloc(args_len * 2 + 2, GFP_KERNEL);
+	if (!buf)
+		return false;
+
+	msgs = kcalloc(args_len + 3, sizeof(*msgs), GFP_KERNEL);
+	if (!msgs) {
+		kfree(buf);
+		return false;
+	}
+
+	intel_sdvo_debug_write(intel_sdvo, cmd, args, args_len);
+
+	for (i = 0; i < args_len; i++) {
+		msgs[i].addr = intel_sdvo->slave_addr;
+		msgs[i].flags = 0;
+		msgs[i].len = 2;
+		msgs[i].buf = buf + 2 *i;
+		buf[2*i + 0] = SDVO_I2C_ARG_0 - i;
+		buf[2*i + 1] = ((u8*)args)[i];
+	}
+	msgs[i].addr = intel_sdvo->slave_addr;
+	msgs[i].flags = 0;
+	msgs[i].len = 2;
+	msgs[i].buf = buf + 2*i;
+	buf[2*i + 0] = SDVO_I2C_OPCODE;
+	buf[2*i + 1] = cmd;
+
+	/* the following two are to read the response */
+	status = SDVO_I2C_CMD_STATUS;
+	msgs[i+1].addr = intel_sdvo->slave_addr;
+	msgs[i+1].flags = 0;
+	msgs[i+1].len = 1;
+	msgs[i+1].buf = &status;
+
+	msgs[i+2].addr = intel_sdvo->slave_addr;
+	msgs[i+2].flags = I2C_M_RD;
+	msgs[i+2].len = 1;
+	msgs[i+2].buf = &status;
+
+	if (unlocked)
+		ret = i2c_transfer(intel_sdvo->i2c, msgs, i+3);
+	else
+		ret = __i2c_transfer(intel_sdvo->i2c, msgs, i+3);
+	if (ret < 0) {
+		DRM_DEBUG_KMS("I2c transfer returned %d\n", ret);
+		ret = false;
+		goto out;
+	}
+	if (ret != i+3) {
+		/* failure in I2C transfer */
+		DRM_DEBUG_KMS("I2c transfer returned %d/%d\n", ret, i+3);
+		ret = false;
+	}
+
+out:
+	kfree(msgs);
+	kfree(buf);
+	return ret;
+}
+
+static bool intel_sdvo_write_cmd(struct intel_sdvo *intel_sdvo, u8 cmd,
+				 const void *args, int args_len)
+{
+	return __intel_sdvo_write_cmd(intel_sdvo, cmd, args, args_len, true);
+}
+
+static bool intel_sdvo_read_response(struct intel_sdvo *intel_sdvo,
+				     void *response, int response_len)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_sdvo->base.base.dev);
+	const char *cmd_status;
+	u8 retry = 15; /* 5 quick checks, followed by 10 long checks */
+	u8 status;
+	int i, pos = 0;
+	char buffer[64];
+
+	buffer[0] = '\0';
+
+	/*
+	 * The documentation states that all commands will be
+	 * processed within 15µs, and that we need only poll
+	 * the status byte a maximum of 3 times in order for the
+	 * command to be complete.
+	 *
+	 * Check 5 times in case the hardware failed to read the docs.
+	 *
+	 * Also beware that the first response by many devices is to
+	 * reply PENDING and stall for time. TVs are notorious for
+	 * requiring longer than specified to complete their replies.
+	 * Originally (in the DDX long ago), the delay was only ever 15ms
+	 * with an additional delay of 30ms applied for TVs added later after
+	 * many experiments. To accommodate both sets of delays, we do a
+	 * sequence of slow checks if the device is falling behind and fails
+	 * to reply within 5*15µs.
+	 */
+	if (!intel_sdvo_read_byte(intel_sdvo,
+				  SDVO_I2C_CMD_STATUS,
+				  &status))
+		goto log_fail;
+
+	while ((status == SDVO_CMD_STATUS_PENDING ||
+		status == SDVO_CMD_STATUS_TARGET_NOT_SPECIFIED) && --retry) {
+		if (retry < 10)
+			msleep(15);
+		else
+			udelay(15);
+
+		if (!intel_sdvo_read_byte(intel_sdvo,
+					  SDVO_I2C_CMD_STATUS,
+					  &status))
+			goto log_fail;
+	}
+
+#define BUF_PRINT(args...) \
+	pos += snprintf(buffer + pos, max_t(int, sizeof(buffer) - pos, 0), args)
+
+	cmd_status = sdvo_cmd_status(status);
+	if (cmd_status)
+		BUF_PRINT("(%s)", cmd_status);
+	else
+		BUF_PRINT("(??? %d)", status);
+
+	if (status != SDVO_CMD_STATUS_SUCCESS)
+		goto log_fail;
+
+	/* Read the command response */
+	for (i = 0; i < response_len; i++) {
+		if (!intel_sdvo_read_byte(intel_sdvo,
+					  SDVO_I2C_RETURN_0 + i,
+					  &((u8 *)response)[i]))
+			goto log_fail;
+		BUF_PRINT(" %02X", ((u8 *)response)[i]);
+	}
+
+	drm_WARN_ON(&dev_priv->drm, pos >= sizeof(buffer) - 1);
+#undef BUF_PRINT
+
+	DRM_DEBUG_KMS("%s: R: %s\n", SDVO_NAME(intel_sdvo), buffer);
+	return true;
+
+log_fail:
+	DRM_DEBUG_KMS("%s: R: ... failed %s\n",
+		      SDVO_NAME(intel_sdvo), buffer);
+	return false;
+}
+
+static int intel_sdvo_get_pixel_multiplier(const struct drm_display_mode *adjusted_mode)
+{
+	if (adjusted_mode->crtc_clock >= 100000)
+		return 1;
+	else if (adjusted_mode->crtc_clock >= 50000)
+		return 2;
+	else
+		return 4;
+}
+
+static bool __intel_sdvo_set_control_bus_switch(struct intel_sdvo *intel_sdvo,
+						u8 ddc_bus)
+{
+	/* This must be the immediately preceding write before the i2c xfer */
+	return __intel_sdvo_write_cmd(intel_sdvo,
+				      SDVO_CMD_SET_CONTROL_BUS_SWITCH,
+				      &ddc_bus, 1, false);
+}
+
+static bool intel_sdvo_set_value(struct intel_sdvo *intel_sdvo, u8 cmd, const void *data, int len)
+{
+	if (!intel_sdvo_write_cmd(intel_sdvo, cmd, data, len))
+		return false;
+
+	return intel_sdvo_read_response(intel_sdvo, NULL, 0);
+}
+
+static bool
+intel_sdvo_get_value(struct intel_sdvo *intel_sdvo, u8 cmd, void *value, int len)
+{
+	if (!intel_sdvo_write_cmd(intel_sdvo, cmd, NULL, 0))
+		return false;
+
+	return intel_sdvo_read_response(intel_sdvo, value, len);
+}
+
+static bool intel_sdvo_set_target_input(struct intel_sdvo *intel_sdvo)
+{
+	struct intel_sdvo_set_target_input_args targets = {0};
+	return intel_sdvo_set_value(intel_sdvo,
+				    SDVO_CMD_SET_TARGET_INPUT,
+				    &targets, sizeof(targets));
+}
+
+/*
+ * Return whether each input is trained.
+ *
+ * This function is making an assumption about the layout of the response,
+ * which should be checked against the docs.
+ */
+static bool intel_sdvo_get_trained_inputs(struct intel_sdvo *intel_sdvo, bool *input_1, bool *input_2)
+{
+	struct intel_sdvo_get_trained_inputs_response response;
+
+	BUILD_BUG_ON(sizeof(response) != 1);
+	if (!intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_TRAINED_INPUTS,
+				  &response, sizeof(response)))
+		return false;
+
+	*input_1 = response.input0_trained;
+	*input_2 = response.input1_trained;
+	return true;
+}
+
+static bool intel_sdvo_set_active_outputs(struct intel_sdvo *intel_sdvo,
+					  u16 outputs)
+{
+	return intel_sdvo_set_value(intel_sdvo,
+				    SDVO_CMD_SET_ACTIVE_OUTPUTS,
+				    &outputs, sizeof(outputs));
+}
+
+static bool intel_sdvo_get_active_outputs(struct intel_sdvo *intel_sdvo,
+					  u16 *outputs)
+{
+	return intel_sdvo_get_value(intel_sdvo,
+				    SDVO_CMD_GET_ACTIVE_OUTPUTS,
+				    outputs, sizeof(*outputs));
+}
+
+static bool intel_sdvo_set_encoder_power_state(struct intel_sdvo *intel_sdvo,
+					       int mode)
+{
+	u8 state = SDVO_ENCODER_STATE_ON;
+
+	switch (mode) {
+	case DRM_MODE_DPMS_ON:
+		state = SDVO_ENCODER_STATE_ON;
+		break;
+	case DRM_MODE_DPMS_STANDBY:
+		state = SDVO_ENCODER_STATE_STANDBY;
+		break;
+	case DRM_MODE_DPMS_SUSPEND:
+		state = SDVO_ENCODER_STATE_SUSPEND;
+		break;
+	case DRM_MODE_DPMS_OFF:
+		state = SDVO_ENCODER_STATE_OFF;
+		break;
+	}
+
+	return intel_sdvo_set_value(intel_sdvo,
+				    SDVO_CMD_SET_ENCODER_POWER_STATE, &state, sizeof(state));
+}
+
+static bool intel_sdvo_get_input_pixel_clock_range(struct intel_sdvo *intel_sdvo,
+						   int *clock_min,
+						   int *clock_max)
+{
+	struct intel_sdvo_pixel_clock_range clocks;
+
+	BUILD_BUG_ON(sizeof(clocks) != 4);
+	if (!intel_sdvo_get_value(intel_sdvo,
+				  SDVO_CMD_GET_INPUT_PIXEL_CLOCK_RANGE,
+				  &clocks, sizeof(clocks)))
+		return false;
+
+	/* Convert the values from units of 10 kHz to kHz. */
+	*clock_min = clocks.min * 10;
+	*clock_max = clocks.max * 10;
+	return true;
+}
+
+static bool intel_sdvo_set_target_output(struct intel_sdvo *intel_sdvo,
+					 u16 outputs)
+{
+	return intel_sdvo_set_value(intel_sdvo,
+				    SDVO_CMD_SET_TARGET_OUTPUT,
+				    &outputs, sizeof(outputs));
+}
+
+static bool intel_sdvo_set_timing(struct intel_sdvo *intel_sdvo, u8 cmd,
+				  struct intel_sdvo_dtd *dtd)
+{
+	return intel_sdvo_set_value(intel_sdvo, cmd, &dtd->part1, sizeof(dtd->part1)) &&
+		intel_sdvo_set_value(intel_sdvo, cmd + 1, &dtd->part2, sizeof(dtd->part2));
+}
+
+static bool intel_sdvo_get_timing(struct intel_sdvo *intel_sdvo, u8 cmd,
+				  struct intel_sdvo_dtd *dtd)
+{
+	return intel_sdvo_get_value(intel_sdvo, cmd, &dtd->part1, sizeof(dtd->part1)) &&
+		intel_sdvo_get_value(intel_sdvo, cmd + 1, &dtd->part2, sizeof(dtd->part2));
+}
+
+static bool intel_sdvo_set_input_timing(struct intel_sdvo *intel_sdvo,
+					 struct intel_sdvo_dtd *dtd)
+{
+	return intel_sdvo_set_timing(intel_sdvo,
+				     SDVO_CMD_SET_INPUT_TIMINGS_PART1, dtd);
+}
+
+static bool intel_sdvo_set_output_timing(struct intel_sdvo *intel_sdvo,
+					 struct intel_sdvo_dtd *dtd)
+{
+	return intel_sdvo_set_timing(intel_sdvo,
+				     SDVO_CMD_SET_OUTPUT_TIMINGS_PART1, dtd);
+}
+
+static bool intel_sdvo_get_input_timing(struct intel_sdvo *intel_sdvo,
+					struct intel_sdvo_dtd *dtd)
+{
+	return intel_sdvo_get_timing(intel_sdvo,
+				     SDVO_CMD_GET_INPUT_TIMINGS_PART1, dtd);
+}
+
+static bool
+intel_sdvo_create_preferred_input_timing(struct intel_sdvo *intel_sdvo,
+					 struct intel_sdvo_connector *intel_sdvo_connector,
+					 const struct drm_display_mode *mode)
+{
+	struct intel_sdvo_preferred_input_timing_args args;
+
+	memset(&args, 0, sizeof(args));
+	args.clock = mode->clock / 10;
+	args.width = mode->hdisplay;
+	args.height = mode->vdisplay;
+	args.interlace = 0;
+
+	if (IS_LVDS(intel_sdvo_connector)) {
+		const struct drm_display_mode *fixed_mode =
+			intel_panel_fixed_mode(&intel_sdvo_connector->base, mode);
+
+		if (fixed_mode->hdisplay != args.width ||
+		    fixed_mode->vdisplay != args.height)
+			args.scaled = 1;
+	}
+
+	return intel_sdvo_set_value(intel_sdvo,
+				    SDVO_CMD_CREATE_PREFERRED_INPUT_TIMING,
+				    &args, sizeof(args));
+}
+
+static bool intel_sdvo_get_preferred_input_timing(struct intel_sdvo *intel_sdvo,
+						  struct intel_sdvo_dtd *dtd)
+{
+	BUILD_BUG_ON(sizeof(dtd->part1) != 8);
+	BUILD_BUG_ON(sizeof(dtd->part2) != 8);
+	return intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART1,
+				    &dtd->part1, sizeof(dtd->part1)) &&
+		intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART2,
+				     &dtd->part2, sizeof(dtd->part2));
+}
+
+static bool intel_sdvo_set_clock_rate_mult(struct intel_sdvo *intel_sdvo, u8 val)
+{
+	return intel_sdvo_set_value(intel_sdvo, SDVO_CMD_SET_CLOCK_RATE_MULT, &val, 1);
+}
+
+static void intel_sdvo_get_dtd_from_mode(struct intel_sdvo_dtd *dtd,
+					 const struct drm_display_mode *mode)
+{
+	u16 width, height;
+	u16 h_blank_len, h_sync_len, v_blank_len, v_sync_len;
+	u16 h_sync_offset, v_sync_offset;
+	int mode_clock;
+
+	memset(dtd, 0, sizeof(*dtd));
+
+	width = mode->hdisplay;
+	height = mode->vdisplay;
+
+	/* do some mode translations */
+	h_blank_len = mode->htotal - mode->hdisplay;
+	h_sync_len = mode->hsync_end - mode->hsync_start;
+
+	v_blank_len = mode->vtotal - mode->vdisplay;
+	v_sync_len = mode->vsync_end - mode->vsync_start;
+
+	h_sync_offset = mode->hsync_start - mode->hdisplay;
+	v_sync_offset = mode->vsync_start - mode->vdisplay;
+
+	mode_clock = mode->clock;
+	mode_clock /= 10;
+	dtd->part1.clock = mode_clock;
+
+	dtd->part1.h_active = width & 0xff;
+	dtd->part1.h_blank = h_blank_len & 0xff;
+	dtd->part1.h_high = (((width >> 8) & 0xf) << 4) |
+		((h_blank_len >> 8) & 0xf);
+	dtd->part1.v_active = height & 0xff;
+	dtd->part1.v_blank = v_blank_len & 0xff;
+	dtd->part1.v_high = (((height >> 8) & 0xf) << 4) |
+		((v_blank_len >> 8) & 0xf);
+
+	dtd->part2.h_sync_off = h_sync_offset & 0xff;
+	dtd->part2.h_sync_width = h_sync_len & 0xff;
+	dtd->part2.v_sync_off_width = (v_sync_offset & 0xf) << 4 |
+		(v_sync_len & 0xf);
+	dtd->part2.sync_off_width_high = ((h_sync_offset & 0x300) >> 2) |
+		((h_sync_len & 0x300) >> 4) | ((v_sync_offset & 0x30) >> 2) |
+		((v_sync_len & 0x30) >> 4);
+
+	dtd->part2.dtd_flags = 0x18;
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		dtd->part2.dtd_flags |= DTD_FLAG_INTERLACE;
+	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+		dtd->part2.dtd_flags |= DTD_FLAG_HSYNC_POSITIVE;
+	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+		dtd->part2.dtd_flags |= DTD_FLAG_VSYNC_POSITIVE;
+
+	dtd->part2.v_sync_off_high = v_sync_offset & 0xc0;
+}
+
+static void intel_sdvo_get_mode_from_dtd(struct drm_display_mode *pmode,
+					 const struct intel_sdvo_dtd *dtd)
+{
+	struct drm_display_mode mode = {};
+
+	mode.hdisplay = dtd->part1.h_active;
+	mode.hdisplay += ((dtd->part1.h_high >> 4) & 0x0f) << 8;
+	mode.hsync_start = mode.hdisplay + dtd->part2.h_sync_off;
+	mode.hsync_start += (dtd->part2.sync_off_width_high & 0xc0) << 2;
+	mode.hsync_end = mode.hsync_start + dtd->part2.h_sync_width;
+	mode.hsync_end += (dtd->part2.sync_off_width_high & 0x30) << 4;
+	mode.htotal = mode.hdisplay + dtd->part1.h_blank;
+	mode.htotal += (dtd->part1.h_high & 0xf) << 8;
+
+	mode.vdisplay = dtd->part1.v_active;
+	mode.vdisplay += ((dtd->part1.v_high >> 4) & 0x0f) << 8;
+	mode.vsync_start = mode.vdisplay;
+	mode.vsync_start += (dtd->part2.v_sync_off_width >> 4) & 0xf;
+	mode.vsync_start += (dtd->part2.sync_off_width_high & 0x0c) << 2;
+	mode.vsync_start += dtd->part2.v_sync_off_high & 0xc0;
+	mode.vsync_end = mode.vsync_start +
+		(dtd->part2.v_sync_off_width & 0xf);
+	mode.vsync_end += (dtd->part2.sync_off_width_high & 0x3) << 4;
+	mode.vtotal = mode.vdisplay + dtd->part1.v_blank;
+	mode.vtotal += (dtd->part1.v_high & 0xf) << 8;
+
+	mode.clock = dtd->part1.clock * 10;
+
+	if (dtd->part2.dtd_flags & DTD_FLAG_INTERLACE)
+		mode.flags |= DRM_MODE_FLAG_INTERLACE;
+	if (dtd->part2.dtd_flags & DTD_FLAG_HSYNC_POSITIVE)
+		mode.flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		mode.flags |= DRM_MODE_FLAG_NHSYNC;
+	if (dtd->part2.dtd_flags & DTD_FLAG_VSYNC_POSITIVE)
+		mode.flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		mode.flags |= DRM_MODE_FLAG_NVSYNC;
+
+	drm_mode_set_crtcinfo(&mode, 0);
+
+	drm_mode_copy(pmode, &mode);
+}
+
+static bool intel_sdvo_check_supp_encode(struct intel_sdvo *intel_sdvo)
+{
+	struct intel_sdvo_encode encode;
+
+	BUILD_BUG_ON(sizeof(encode) != 2);
+	return intel_sdvo_get_value(intel_sdvo,
+				  SDVO_CMD_GET_SUPP_ENCODE,
+				  &encode, sizeof(encode));
+}
+
+static bool intel_sdvo_set_encode(struct intel_sdvo *intel_sdvo,
+				  u8 mode)
+{
+	return intel_sdvo_set_value(intel_sdvo, SDVO_CMD_SET_ENCODE, &mode, 1);
+}
+
+static bool intel_sdvo_set_colorimetry(struct intel_sdvo *intel_sdvo,
+				       u8 mode)
+{
+	return intel_sdvo_set_value(intel_sdvo, SDVO_CMD_SET_COLORIMETRY, &mode, 1);
+}
+
+static bool intel_sdvo_set_pixel_replication(struct intel_sdvo *intel_sdvo,
+					     u8 pixel_repeat)
+{
+	return intel_sdvo_set_value(intel_sdvo, SDVO_CMD_SET_PIXEL_REPLI,
+				    &pixel_repeat, 1);
+}
+
+static bool intel_sdvo_set_audio_state(struct intel_sdvo *intel_sdvo,
+				       u8 audio_state)
+{
+	return intel_sdvo_set_value(intel_sdvo, SDVO_CMD_SET_AUDIO_STAT,
+				    &audio_state, 1);
+}
+
+static bool intel_sdvo_get_hbuf_size(struct intel_sdvo *intel_sdvo,
+				     u8 *hbuf_size)
+{
+	if (!intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_HBUF_INFO,
+				  hbuf_size, 1))
+		return false;
+
+	/* Buffer size is 0 based, hooray! However zero means zero. */
+	if (*hbuf_size)
+		(*hbuf_size)++;
+
+	return true;
+}
+
+#if 0
+static void intel_sdvo_dump_hdmi_buf(struct intel_sdvo *intel_sdvo)
+{
+	int i, j;
+	u8 set_buf_index[2];
+	u8 av_split;
+	u8 buf_size;
+	u8 buf[48];
+	u8 *pos;
+
+	intel_sdvo_get_value(encoder, SDVO_CMD_GET_HBUF_AV_SPLIT, &av_split, 1);
+
+	for (i = 0; i <= av_split; i++) {
+		set_buf_index[0] = i; set_buf_index[1] = 0;
+		intel_sdvo_write_cmd(encoder, SDVO_CMD_SET_HBUF_INDEX,
+				     set_buf_index, 2);
+		intel_sdvo_write_cmd(encoder, SDVO_CMD_GET_HBUF_INFO, NULL, 0);
+		intel_sdvo_read_response(encoder, &buf_size, 1);
+
+		pos = buf;
+		for (j = 0; j <= buf_size; j += 8) {
+			intel_sdvo_write_cmd(encoder, SDVO_CMD_GET_HBUF_DATA,
+					     NULL, 0);
+			intel_sdvo_read_response(encoder, pos, 8);
+			pos += 8;
+		}
+	}
+}
+#endif
+
+static bool intel_sdvo_write_infoframe(struct intel_sdvo *intel_sdvo,
+				       unsigned int if_index, u8 tx_rate,
+				       const u8 *data, unsigned int length)
+{
+	u8 set_buf_index[2] = { if_index, 0 };
+	u8 hbuf_size, tmp[8];
+	int i;
+
+	if (!intel_sdvo_set_value(intel_sdvo,
+				  SDVO_CMD_SET_HBUF_INDEX,
+				  set_buf_index, 2))
+		return false;
+
+	if (!intel_sdvo_get_hbuf_size(intel_sdvo, &hbuf_size))
+		return false;
+
+	DRM_DEBUG_KMS("writing sdvo hbuf: %i, length %u, hbuf_size: %i\n",
+		      if_index, length, hbuf_size);
+
+	if (hbuf_size < length)
+		return false;
+
+	for (i = 0; i < hbuf_size; i += 8) {
+		memset(tmp, 0, 8);
+		if (i < length)
+			memcpy(tmp, data + i, min_t(unsigned, 8, length - i));
+
+		if (!intel_sdvo_set_value(intel_sdvo,
+					  SDVO_CMD_SET_HBUF_DATA,
+					  tmp, 8))
+			return false;
+	}
+
+	return intel_sdvo_set_value(intel_sdvo,
+				    SDVO_CMD_SET_HBUF_TXRATE,
+				    &tx_rate, 1);
+}
+
+static ssize_t intel_sdvo_read_infoframe(struct intel_sdvo *intel_sdvo,
+					 unsigned int if_index,
+					 u8 *data, unsigned int length)
+{
+	u8 set_buf_index[2] = { if_index, 0 };
+	u8 hbuf_size, tx_rate, av_split;
+	int i;
+
+	if (!intel_sdvo_get_value(intel_sdvo,
+				  SDVO_CMD_GET_HBUF_AV_SPLIT,
+				  &av_split, 1))
+		return -ENXIO;
+
+	if (av_split < if_index)
+		return 0;
+
+	if (!intel_sdvo_set_value(intel_sdvo,
+				  SDVO_CMD_SET_HBUF_INDEX,
+				  set_buf_index, 2))
+		return -ENXIO;
+
+	if (!intel_sdvo_get_value(intel_sdvo,
+				  SDVO_CMD_GET_HBUF_TXRATE,
+				  &tx_rate, 1))
+		return -ENXIO;
+
+	if (tx_rate == SDVO_HBUF_TX_DISABLED)
+		return 0;
+
+	if (!intel_sdvo_get_hbuf_size(intel_sdvo, &hbuf_size))
+		return false;
+
+	DRM_DEBUG_KMS("reading sdvo hbuf: %i, length %u, hbuf_size: %i\n",
+		      if_index, length, hbuf_size);
+
+	hbuf_size = min_t(unsigned int, length, hbuf_size);
+
+	for (i = 0; i < hbuf_size; i += 8) {
+		if (!intel_sdvo_write_cmd(intel_sdvo, SDVO_CMD_GET_HBUF_DATA, NULL, 0))
+			return -ENXIO;
+		if (!intel_sdvo_read_response(intel_sdvo, &data[i],
+					      min_t(unsigned int, 8, hbuf_size - i)))
+			return -ENXIO;
+	}
+
+	return hbuf_size;
+}
+
+static bool intel_sdvo_compute_avi_infoframe(struct intel_sdvo *intel_sdvo,
+					     struct intel_crtc_state *crtc_state,
+					     struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_sdvo->base.base.dev);
+	struct hdmi_avi_infoframe *frame = &crtc_state->infoframes.avi.avi;
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	int ret;
+
+	if (!crtc_state->has_hdmi_sink)
+		return true;
+
+	crtc_state->infoframes.enable |=
+		intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_AVI);
+
+	ret = drm_hdmi_avi_infoframe_from_display_mode(frame,
+						       conn_state->connector,
+						       adjusted_mode);
+	if (ret)
+		return false;
+
+	drm_hdmi_avi_infoframe_quant_range(frame,
+					   conn_state->connector,
+					   adjusted_mode,
+					   crtc_state->limited_color_range ?
+					   HDMI_QUANTIZATION_RANGE_LIMITED :
+					   HDMI_QUANTIZATION_RANGE_FULL);
+
+	ret = hdmi_avi_infoframe_check(frame);
+	if (drm_WARN_ON(&dev_priv->drm, ret))
+		return false;
+
+	return true;
+}
+
+static bool intel_sdvo_set_avi_infoframe(struct intel_sdvo *intel_sdvo,
+					 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_sdvo->base.base.dev);
+	u8 sdvo_data[HDMI_INFOFRAME_SIZE(AVI)];
+	const union hdmi_infoframe *frame = &crtc_state->infoframes.avi;
+	ssize_t len;
+
+	if ((crtc_state->infoframes.enable &
+	     intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_AVI)) == 0)
+		return true;
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			frame->any.type != HDMI_INFOFRAME_TYPE_AVI))
+		return false;
+
+	len = hdmi_infoframe_pack_only(frame, sdvo_data, sizeof(sdvo_data));
+	if (drm_WARN_ON(&dev_priv->drm, len < 0))
+		return false;
+
+	return intel_sdvo_write_infoframe(intel_sdvo, SDVO_HBUF_INDEX_AVI_IF,
+					  SDVO_HBUF_TX_VSYNC,
+					  sdvo_data, len);
+}
+
+static void intel_sdvo_get_avi_infoframe(struct intel_sdvo *intel_sdvo,
+					 struct intel_crtc_state *crtc_state)
+{
+	u8 sdvo_data[HDMI_INFOFRAME_SIZE(AVI)];
+	union hdmi_infoframe *frame = &crtc_state->infoframes.avi;
+	ssize_t len;
+	int ret;
+
+	if (!crtc_state->has_hdmi_sink)
+		return;
+
+	len = intel_sdvo_read_infoframe(intel_sdvo, SDVO_HBUF_INDEX_AVI_IF,
+					sdvo_data, sizeof(sdvo_data));
+	if (len < 0) {
+		DRM_DEBUG_KMS("failed to read AVI infoframe\n");
+		return;
+	} else if (len == 0) {
+		return;
+	}
+
+	crtc_state->infoframes.enable |=
+		intel_hdmi_infoframe_enable(HDMI_INFOFRAME_TYPE_AVI);
+
+	ret = hdmi_infoframe_unpack(frame, sdvo_data, len);
+	if (ret) {
+		DRM_DEBUG_KMS("Failed to unpack AVI infoframe\n");
+		return;
+	}
+
+	if (frame->any.type != HDMI_INFOFRAME_TYPE_AVI)
+		DRM_DEBUG_KMS("Found the wrong infoframe type 0x%x (expected 0x%02x)\n",
+			      frame->any.type, HDMI_INFOFRAME_TYPE_AVI);
+}
+
+static bool intel_sdvo_set_tv_format(struct intel_sdvo *intel_sdvo,
+				     const struct drm_connector_state *conn_state)
+{
+	struct intel_sdvo_tv_format format;
+	u32 format_map;
+
+	format_map = 1 << conn_state->tv.mode;
+	memset(&format, 0, sizeof(format));
+	memcpy(&format, &format_map, min(sizeof(format), sizeof(format_map)));
+
+	BUILD_BUG_ON(sizeof(format) != 6);
+	return intel_sdvo_set_value(intel_sdvo,
+				    SDVO_CMD_SET_TV_FORMAT,
+				    &format, sizeof(format));
+}
+
+static bool
+intel_sdvo_set_output_timings_from_mode(struct intel_sdvo *intel_sdvo,
+					const struct drm_display_mode *mode)
+{
+	struct intel_sdvo_dtd output_dtd;
+
+	if (!intel_sdvo_set_target_output(intel_sdvo,
+					  intel_sdvo->attached_output))
+		return false;
+
+	intel_sdvo_get_dtd_from_mode(&output_dtd, mode);
+	if (!intel_sdvo_set_output_timing(intel_sdvo, &output_dtd))
+		return false;
+
+	return true;
+}
+
+/*
+ * Asks the sdvo controller for the preferred input mode given the output mode.
+ * Unfortunately we have to set up the full output mode to do that.
+ */
+static bool
+intel_sdvo_get_preferred_input_mode(struct intel_sdvo *intel_sdvo,
+				    struct intel_sdvo_connector *intel_sdvo_connector,
+				    const struct drm_display_mode *mode,
+				    struct drm_display_mode *adjusted_mode)
+{
+	struct intel_sdvo_dtd input_dtd;
+
+	/* Reset the input timing to the screen. Assume always input 0. */
+	if (!intel_sdvo_set_target_input(intel_sdvo))
+		return false;
+
+	if (!intel_sdvo_create_preferred_input_timing(intel_sdvo,
+						      intel_sdvo_connector,
+						      mode))
+		return false;
+
+	if (!intel_sdvo_get_preferred_input_timing(intel_sdvo,
+						   &input_dtd))
+		return false;
+
+	intel_sdvo_get_mode_from_dtd(adjusted_mode, &input_dtd);
+	intel_sdvo->dtd_sdvo_flags = input_dtd.part2.sdvo_flags;
+
+	return true;
+}
+
+static void i9xx_adjust_sdvo_tv_clock(struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(pipe_config->uapi.crtc->dev);
+	unsigned dotclock = pipe_config->port_clock;
+	struct dpll *clock = &pipe_config->dpll;
+
+	/*
+	 * SDVO TV has fixed PLL values depend on its clock range,
+	 * this mirrors vbios setting.
+	 */
+	if (dotclock >= 100000 && dotclock < 140500) {
+		clock->p1 = 2;
+		clock->p2 = 10;
+		clock->n = 3;
+		clock->m1 = 16;
+		clock->m2 = 8;
+	} else if (dotclock >= 140500 && dotclock <= 200000) {
+		clock->p1 = 1;
+		clock->p2 = 10;
+		clock->n = 6;
+		clock->m1 = 12;
+		clock->m2 = 8;
+	} else {
+		drm_WARN(&dev_priv->drm, 1,
+			 "SDVO TV clock out of range: %i\n", dotclock);
+	}
+
+	pipe_config->clock_set = true;
+}
+
+static bool intel_has_hdmi_sink(struct intel_sdvo_connector *intel_sdvo_connector,
+				const struct drm_connector_state *conn_state)
+{
+	struct drm_connector *connector = conn_state->connector;
+
+	return intel_sdvo_connector->is_hdmi &&
+		connector->display_info.is_hdmi &&
+		READ_ONCE(to_intel_digital_connector_state(conn_state)->force_audio) != HDMI_AUDIO_OFF_DVI;
+}
+
+static bool intel_sdvo_limited_color_range(struct intel_encoder *encoder,
+					   const struct intel_crtc_state *crtc_state,
+					   const struct drm_connector_state *conn_state)
+{
+	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
+
+	if ((intel_sdvo->colorimetry_cap & SDVO_COLORIMETRY_RGB220) == 0)
+		return false;
+
+	return intel_hdmi_limited_color_range(crtc_state, conn_state);
+}
+
+static int intel_sdvo_compute_config(struct intel_encoder *encoder,
+				     struct intel_crtc_state *pipe_config,
+				     struct drm_connector_state *conn_state)
+{
+	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
+	struct intel_sdvo_connector_state *intel_sdvo_state =
+		to_intel_sdvo_connector_state(conn_state);
+	struct intel_sdvo_connector *intel_sdvo_connector =
+		to_intel_sdvo_connector(conn_state->connector);
+	struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	struct drm_display_mode *mode = &pipe_config->hw.mode;
+
+	DRM_DEBUG_KMS("forcing bpc to 8 for SDVO\n");
+	pipe_config->pipe_bpp = 8*3;
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+
+	if (HAS_PCH_SPLIT(to_i915(encoder->base.dev)))
+		pipe_config->has_pch_encoder = true;
+
+	/*
+	 * We need to construct preferred input timings based on our
+	 * output timings.  To do that, we have to set the output
+	 * timings, even though this isn't really the right place in
+	 * the sequence to do it. Oh well.
+	 */
+	if (IS_TV(intel_sdvo_connector)) {
+		if (!intel_sdvo_set_output_timings_from_mode(intel_sdvo, mode))
+			return -EINVAL;
+
+		(void) intel_sdvo_get_preferred_input_mode(intel_sdvo,
+							   intel_sdvo_connector,
+							   mode,
+							   adjusted_mode);
+		pipe_config->sdvo_tv_clock = true;
+	} else if (IS_LVDS(intel_sdvo_connector)) {
+		const struct drm_display_mode *fixed_mode =
+			intel_panel_fixed_mode(&intel_sdvo_connector->base, mode);
+		int ret;
+
+		ret = intel_panel_compute_config(&intel_sdvo_connector->base,
+						 adjusted_mode);
+		if (ret)
+			return ret;
+
+		if (!intel_sdvo_set_output_timings_from_mode(intel_sdvo, fixed_mode))
+			return -EINVAL;
+
+		(void) intel_sdvo_get_preferred_input_mode(intel_sdvo,
+							   intel_sdvo_connector,
+							   mode,
+							   adjusted_mode);
+	}
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	/*
+	 * Make the CRTC code factor in the SDVO pixel multiplier.  The
+	 * SDVO device will factor out the multiplier during mode_set.
+	 */
+	pipe_config->pixel_multiplier =
+		intel_sdvo_get_pixel_multiplier(adjusted_mode);
+
+	pipe_config->has_hdmi_sink = intel_has_hdmi_sink(intel_sdvo_connector, conn_state);
+
+	if (pipe_config->has_hdmi_sink) {
+		if (intel_sdvo_state->base.force_audio == HDMI_AUDIO_AUTO)
+			pipe_config->has_audio = intel_sdvo->has_hdmi_audio;
+		else
+			pipe_config->has_audio =
+				intel_sdvo_state->base.force_audio == HDMI_AUDIO_ON;
+	}
+
+	pipe_config->limited_color_range =
+		intel_sdvo_limited_color_range(encoder, pipe_config,
+					       conn_state);
+
+	/* Clock computation needs to happen after pixel multiplier. */
+	if (IS_TV(intel_sdvo_connector))
+		i9xx_adjust_sdvo_tv_clock(pipe_config);
+
+	if (conn_state->picture_aspect_ratio)
+		adjusted_mode->picture_aspect_ratio =
+			conn_state->picture_aspect_ratio;
+
+	if (!intel_sdvo_compute_avi_infoframe(intel_sdvo,
+					      pipe_config, conn_state)) {
+		DRM_DEBUG_KMS("bad AVI infoframe\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+#define UPDATE_PROPERTY(input, NAME) \
+	do { \
+		val = input; \
+		intel_sdvo_set_value(intel_sdvo, SDVO_CMD_SET_##NAME, &val, sizeof(val)); \
+	} while (0)
+
+static void intel_sdvo_update_props(struct intel_sdvo *intel_sdvo,
+				    const struct intel_sdvo_connector_state *sdvo_state)
+{
+	const struct drm_connector_state *conn_state = &sdvo_state->base.base;
+	struct intel_sdvo_connector *intel_sdvo_conn =
+		to_intel_sdvo_connector(conn_state->connector);
+	u16 val;
+
+	if (intel_sdvo_conn->left)
+		UPDATE_PROPERTY(sdvo_state->tv.overscan_h, OVERSCAN_H);
+
+	if (intel_sdvo_conn->top)
+		UPDATE_PROPERTY(sdvo_state->tv.overscan_v, OVERSCAN_V);
+
+	if (intel_sdvo_conn->hpos)
+		UPDATE_PROPERTY(sdvo_state->tv.hpos, HPOS);
+
+	if (intel_sdvo_conn->vpos)
+		UPDATE_PROPERTY(sdvo_state->tv.vpos, VPOS);
+
+	if (intel_sdvo_conn->saturation)
+		UPDATE_PROPERTY(conn_state->tv.saturation, SATURATION);
+
+	if (intel_sdvo_conn->contrast)
+		UPDATE_PROPERTY(conn_state->tv.contrast, CONTRAST);
+
+	if (intel_sdvo_conn->hue)
+		UPDATE_PROPERTY(conn_state->tv.hue, HUE);
+
+	if (intel_sdvo_conn->brightness)
+		UPDATE_PROPERTY(conn_state->tv.brightness, BRIGHTNESS);
+
+	if (intel_sdvo_conn->sharpness)
+		UPDATE_PROPERTY(sdvo_state->tv.sharpness, SHARPNESS);
+
+	if (intel_sdvo_conn->flicker_filter)
+		UPDATE_PROPERTY(sdvo_state->tv.flicker_filter, FLICKER_FILTER);
+
+	if (intel_sdvo_conn->flicker_filter_2d)
+		UPDATE_PROPERTY(sdvo_state->tv.flicker_filter_2d, FLICKER_FILTER_2D);
+
+	if (intel_sdvo_conn->flicker_filter_adaptive)
+		UPDATE_PROPERTY(sdvo_state->tv.flicker_filter_adaptive, FLICKER_FILTER_ADAPTIVE);
+
+	if (intel_sdvo_conn->tv_chroma_filter)
+		UPDATE_PROPERTY(sdvo_state->tv.chroma_filter, TV_CHROMA_FILTER);
+
+	if (intel_sdvo_conn->tv_luma_filter)
+		UPDATE_PROPERTY(sdvo_state->tv.luma_filter, TV_LUMA_FILTER);
+
+	if (intel_sdvo_conn->dot_crawl)
+		UPDATE_PROPERTY(sdvo_state->tv.dot_crawl, DOT_CRAWL);
+
+#undef UPDATE_PROPERTY
+}
+
+static void intel_sdvo_pre_enable(struct intel_atomic_state *state,
+				  struct intel_encoder *intel_encoder,
+				  const struct intel_crtc_state *crtc_state,
+				  const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	const struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	const struct intel_sdvo_connector_state *sdvo_state =
+		to_intel_sdvo_connector_state(conn_state);
+	struct intel_sdvo_connector *intel_sdvo_connector =
+		to_intel_sdvo_connector(conn_state->connector);
+	const struct drm_display_mode *mode = &crtc_state->hw.mode;
+	struct intel_sdvo *intel_sdvo = to_sdvo(intel_encoder);
+	u32 sdvox;
+	struct intel_sdvo_in_out_map in_out;
+	struct intel_sdvo_dtd input_dtd, output_dtd;
+	int rate;
+
+	intel_sdvo_update_props(intel_sdvo, sdvo_state);
+
+	/*
+	 * First, set the input mapping for the first input to our controlled
+	 * output. This is only correct if we're a single-input device, in
+	 * which case the first input is the output from the appropriate SDVO
+	 * channel on the motherboard.  In a two-input device, the first input
+	 * will be SDVOB and the second SDVOC.
+	 */
+	in_out.in0 = intel_sdvo->attached_output;
+	in_out.in1 = 0;
+
+	intel_sdvo_set_value(intel_sdvo,
+			     SDVO_CMD_SET_IN_OUT_MAP,
+			     &in_out, sizeof(in_out));
+
+	/* Set the output timings to the screen */
+	if (!intel_sdvo_set_target_output(intel_sdvo,
+					  intel_sdvo->attached_output))
+		return;
+
+	/* lvds has a special fixed output timing. */
+	if (IS_LVDS(intel_sdvo_connector)) {
+		const struct drm_display_mode *fixed_mode =
+			intel_panel_fixed_mode(&intel_sdvo_connector->base, mode);
+
+		intel_sdvo_get_dtd_from_mode(&output_dtd, fixed_mode);
+	} else {
+		intel_sdvo_get_dtd_from_mode(&output_dtd, mode);
+	}
+	if (!intel_sdvo_set_output_timing(intel_sdvo, &output_dtd))
+		drm_info(&dev_priv->drm,
+			 "Setting output timings on %s failed\n",
+			 SDVO_NAME(intel_sdvo));
+
+	/* Set the input timing to the screen. Assume always input 0. */
+	if (!intel_sdvo_set_target_input(intel_sdvo))
+		return;
+
+	if (crtc_state->has_hdmi_sink) {
+		intel_sdvo_set_encode(intel_sdvo, SDVO_ENCODE_HDMI);
+		intel_sdvo_set_colorimetry(intel_sdvo,
+					   crtc_state->limited_color_range ?
+					   SDVO_COLORIMETRY_RGB220 :
+					   SDVO_COLORIMETRY_RGB256);
+		intel_sdvo_set_avi_infoframe(intel_sdvo, crtc_state);
+		intel_sdvo_set_pixel_replication(intel_sdvo,
+						 !!(adjusted_mode->flags &
+						    DRM_MODE_FLAG_DBLCLK));
+	} else
+		intel_sdvo_set_encode(intel_sdvo, SDVO_ENCODE_DVI);
+
+	if (IS_TV(intel_sdvo_connector) &&
+	    !intel_sdvo_set_tv_format(intel_sdvo, conn_state))
+		return;
+
+	intel_sdvo_get_dtd_from_mode(&input_dtd, adjusted_mode);
+
+	if (IS_TV(intel_sdvo_connector) || IS_LVDS(intel_sdvo_connector))
+		input_dtd.part2.sdvo_flags = intel_sdvo->dtd_sdvo_flags;
+	if (!intel_sdvo_set_input_timing(intel_sdvo, &input_dtd))
+		drm_info(&dev_priv->drm,
+			 "Setting input timings on %s failed\n",
+			 SDVO_NAME(intel_sdvo));
+
+	switch (crtc_state->pixel_multiplier) {
+	default:
+		drm_WARN(&dev_priv->drm, 1,
+			 "unknown pixel multiplier specified\n");
+		fallthrough;
+	case 1: rate = SDVO_CLOCK_RATE_MULT_1X; break;
+	case 2: rate = SDVO_CLOCK_RATE_MULT_2X; break;
+	case 4: rate = SDVO_CLOCK_RATE_MULT_4X; break;
+	}
+	if (!intel_sdvo_set_clock_rate_mult(intel_sdvo, rate))
+		return;
+
+	/* Set the SDVO control regs. */
+	if (DISPLAY_VER(dev_priv) >= 4) {
+		/* The real mode polarity is set by the SDVO commands, using
+		 * struct intel_sdvo_dtd. */
+		sdvox = SDVO_VSYNC_ACTIVE_HIGH | SDVO_HSYNC_ACTIVE_HIGH;
+		if (DISPLAY_VER(dev_priv) < 5)
+			sdvox |= SDVO_BORDER_ENABLE;
+	} else {
+		sdvox = intel_de_read(dev_priv, intel_sdvo->sdvo_reg);
+		if (intel_sdvo->port == PORT_B)
+			sdvox &= SDVOB_PRESERVE_MASK;
+		else
+			sdvox &= SDVOC_PRESERVE_MASK;
+		sdvox |= (9 << 19) | SDVO_BORDER_ENABLE;
+	}
+
+	if (HAS_PCH_CPT(dev_priv))
+		sdvox |= SDVO_PIPE_SEL_CPT(crtc->pipe);
+	else
+		sdvox |= SDVO_PIPE_SEL(crtc->pipe);
+
+	if (DISPLAY_VER(dev_priv) >= 4) {
+		/* done in crtc_mode_set as the dpll_md reg must be written early */
+	} else if (IS_I945G(dev_priv) || IS_I945GM(dev_priv) ||
+		   IS_G33(dev_priv) || IS_PINEVIEW(dev_priv)) {
+		/* done in crtc_mode_set as it lives inside the dpll register */
+	} else {
+		sdvox |= (crtc_state->pixel_multiplier - 1)
+			<< SDVO_PORT_MULTIPLY_SHIFT;
+	}
+
+	if (input_dtd.part2.sdvo_flags & SDVO_NEED_TO_STALL &&
+	    DISPLAY_VER(dev_priv) < 5)
+		sdvox |= SDVO_STALL_SELECT;
+	intel_sdvo_write_sdvox(intel_sdvo, sdvox);
+}
+
+static bool intel_sdvo_connector_get_hw_state(struct intel_connector *connector)
+{
+	struct intel_sdvo_connector *intel_sdvo_connector =
+		to_intel_sdvo_connector(&connector->base);
+	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(connector);
+	u16 active_outputs = 0;
+
+	intel_sdvo_get_active_outputs(intel_sdvo, &active_outputs);
+
+	return active_outputs & intel_sdvo_connector->output_flag;
+}
+
+bool intel_sdvo_port_enabled(struct drm_i915_private *dev_priv,
+			     i915_reg_t sdvo_reg, enum pipe *pipe)
+{
+	u32 val;
+
+	val = intel_de_read(dev_priv, sdvo_reg);
+
+	/* asserts want to know the pipe even if the port is disabled */
+	if (HAS_PCH_CPT(dev_priv))
+		*pipe = (val & SDVO_PIPE_SEL_MASK_CPT) >> SDVO_PIPE_SEL_SHIFT_CPT;
+	else if (IS_CHERRYVIEW(dev_priv))
+		*pipe = (val & SDVO_PIPE_SEL_MASK_CHV) >> SDVO_PIPE_SEL_SHIFT_CHV;
+	else
+		*pipe = (val & SDVO_PIPE_SEL_MASK) >> SDVO_PIPE_SEL_SHIFT;
+
+	return val & SDVO_ENABLE;
+}
+
+static bool intel_sdvo_get_hw_state(struct intel_encoder *encoder,
+				    enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
+	u16 active_outputs = 0;
+	bool ret;
+
+	intel_sdvo_get_active_outputs(intel_sdvo, &active_outputs);
+
+	ret = intel_sdvo_port_enabled(dev_priv, intel_sdvo->sdvo_reg, pipe);
+
+	return ret || active_outputs;
+}
+
+static void intel_sdvo_get_config(struct intel_encoder *encoder,
+				  struct intel_crtc_state *pipe_config)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
+	struct intel_sdvo_dtd dtd;
+	int encoder_pixel_multiplier = 0;
+	int dotclock;
+	u32 flags = 0, sdvox;
+	u8 val;
+	bool ret;
+
+	pipe_config->output_types |= BIT(INTEL_OUTPUT_SDVO);
+
+	sdvox = intel_de_read(dev_priv, intel_sdvo->sdvo_reg);
+
+	ret = intel_sdvo_get_input_timing(intel_sdvo, &dtd);
+	if (!ret) {
+		/*
+		 * Some sdvo encoders are not spec compliant and don't
+		 * implement the mandatory get_timings function.
+		 */
+		drm_dbg(&dev_priv->drm, "failed to retrieve SDVO DTD\n");
+		pipe_config->quirks |= PIPE_CONFIG_QUIRK_MODE_SYNC_FLAGS;
+	} else {
+		if (dtd.part2.dtd_flags & DTD_FLAG_HSYNC_POSITIVE)
+			flags |= DRM_MODE_FLAG_PHSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NHSYNC;
+
+		if (dtd.part2.dtd_flags & DTD_FLAG_VSYNC_POSITIVE)
+			flags |= DRM_MODE_FLAG_PVSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NVSYNC;
+	}
+
+	pipe_config->hw.adjusted_mode.flags |= flags;
+
+	/*
+	 * pixel multiplier readout is tricky: Only on i915g/gm it is stored in
+	 * the sdvo port register, on all other platforms it is part of the dpll
+	 * state. Since the general pipe state readout happens before the
+	 * encoder->get_config we so already have a valid pixel multplier on all
+	 * other platfroms.
+	 */
+	if (IS_I915G(dev_priv) || IS_I915GM(dev_priv)) {
+		pipe_config->pixel_multiplier =
+			((sdvox & SDVO_PORT_MULTIPLY_MASK)
+			 >> SDVO_PORT_MULTIPLY_SHIFT) + 1;
+	}
+
+	dotclock = pipe_config->port_clock;
+
+	if (pipe_config->pixel_multiplier)
+		dotclock /= pipe_config->pixel_multiplier;
+
+	pipe_config->hw.adjusted_mode.crtc_clock = dotclock;
+
+	/* Cross check the port pixel multiplier with the sdvo encoder state. */
+	if (intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_CLOCK_RATE_MULT,
+				 &val, 1)) {
+		switch (val) {
+		case SDVO_CLOCK_RATE_MULT_1X:
+			encoder_pixel_multiplier = 1;
+			break;
+		case SDVO_CLOCK_RATE_MULT_2X:
+			encoder_pixel_multiplier = 2;
+			break;
+		case SDVO_CLOCK_RATE_MULT_4X:
+			encoder_pixel_multiplier = 4;
+			break;
+		}
+	}
+
+	drm_WARN(dev,
+		 encoder_pixel_multiplier != pipe_config->pixel_multiplier,
+		 "SDVO pixel multiplier mismatch, port: %i, encoder: %i\n",
+		 pipe_config->pixel_multiplier, encoder_pixel_multiplier);
+
+	if (intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_COLORIMETRY,
+				 &val, 1)) {
+		if (val == SDVO_COLORIMETRY_RGB220)
+			pipe_config->limited_color_range = true;
+	}
+
+	if (intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_AUDIO_STAT,
+				 &val, 1)) {
+		u8 mask = SDVO_AUDIO_ELD_VALID | SDVO_AUDIO_PRESENCE_DETECT;
+
+		if ((val & mask) == mask)
+			pipe_config->has_audio = true;
+	}
+
+	if (intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_ENCODE,
+				 &val, 1)) {
+		if (val == SDVO_ENCODE_HDMI)
+			pipe_config->has_hdmi_sink = true;
+	}
+
+	intel_sdvo_get_avi_infoframe(intel_sdvo, pipe_config);
+}
+
+static void intel_sdvo_disable_audio(struct intel_sdvo *intel_sdvo)
+{
+	intel_sdvo_set_audio_state(intel_sdvo, 0);
+}
+
+static void intel_sdvo_enable_audio(struct intel_sdvo *intel_sdvo,
+				    const struct intel_crtc_state *crtc_state,
+				    const struct drm_connector_state *conn_state)
+{
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	struct drm_connector *connector = conn_state->connector;
+	u8 *eld = connector->eld;
+
+	eld[6] = drm_av_sync_delay(connector, adjusted_mode) / 2;
+
+	intel_sdvo_set_audio_state(intel_sdvo, 0);
+
+	intel_sdvo_write_infoframe(intel_sdvo, SDVO_HBUF_INDEX_ELD,
+				   SDVO_HBUF_TX_DISABLED,
+				   eld, drm_eld_size(eld));
+
+	intel_sdvo_set_audio_state(intel_sdvo, SDVO_AUDIO_ELD_VALID |
+				   SDVO_AUDIO_PRESENCE_DETECT);
+}
+
+static void intel_disable_sdvo(struct intel_atomic_state *state,
+			       struct intel_encoder *encoder,
+			       const struct intel_crtc_state *old_crtc_state,
+			       const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	u32 temp;
+
+	if (old_crtc_state->has_audio)
+		intel_sdvo_disable_audio(intel_sdvo);
+
+	intel_sdvo_set_active_outputs(intel_sdvo, 0);
+	if (0)
+		intel_sdvo_set_encoder_power_state(intel_sdvo,
+						   DRM_MODE_DPMS_OFF);
+
+	temp = intel_de_read(dev_priv, intel_sdvo->sdvo_reg);
+
+	temp &= ~SDVO_ENABLE;
+	intel_sdvo_write_sdvox(intel_sdvo, temp);
+
+	/*
+	 * HW workaround for IBX, we need to move the port
+	 * to transcoder A after disabling it to allow the
+	 * matching DP port to be enabled on transcoder A.
+	 */
+	if (HAS_PCH_IBX(dev_priv) && crtc->pipe == PIPE_B) {
+		/*
+		 * We get CPU/PCH FIFO underruns on the other pipe when
+		 * doing the workaround. Sweep them under the rug.
+		 */
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, PIPE_A, false);
+		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, false);
+
+		temp &= ~SDVO_PIPE_SEL_MASK;
+		temp |= SDVO_ENABLE | SDVO_PIPE_SEL(PIPE_A);
+		intel_sdvo_write_sdvox(intel_sdvo, temp);
+
+		temp &= ~SDVO_ENABLE;
+		intel_sdvo_write_sdvox(intel_sdvo, temp);
+
+		intel_wait_for_vblank_if_active(dev_priv, PIPE_A);
+		intel_set_cpu_fifo_underrun_reporting(dev_priv, PIPE_A, true);
+		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, true);
+	}
+}
+
+static void pch_disable_sdvo(struct intel_atomic_state *state,
+			     struct intel_encoder *encoder,
+			     const struct intel_crtc_state *old_crtc_state,
+			     const struct drm_connector_state *old_conn_state)
+{
+}
+
+static void pch_post_disable_sdvo(struct intel_atomic_state *state,
+				  struct intel_encoder *encoder,
+				  const struct intel_crtc_state *old_crtc_state,
+				  const struct drm_connector_state *old_conn_state)
+{
+	intel_disable_sdvo(state, encoder, old_crtc_state, old_conn_state);
+}
+
+static void intel_enable_sdvo(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *pipe_config,
+			      const struct drm_connector_state *conn_state)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	u32 temp;
+	bool input1, input2;
+	int i;
+	bool success;
+
+	temp = intel_de_read(dev_priv, intel_sdvo->sdvo_reg);
+	temp |= SDVO_ENABLE;
+	intel_sdvo_write_sdvox(intel_sdvo, temp);
+
+	for (i = 0; i < 2; i++)
+		intel_crtc_wait_for_next_vblank(crtc);
+
+	success = intel_sdvo_get_trained_inputs(intel_sdvo, &input1, &input2);
+	/*
+	 * Warn if the device reported failure to sync.
+	 *
+	 * A lot of SDVO devices fail to notify of sync, but it's
+	 * a given it the status is a success, we succeeded.
+	 */
+	if (success && !input1) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "First %s output reported failure to "
+			    "sync\n", SDVO_NAME(intel_sdvo));
+	}
+
+	if (0)
+		intel_sdvo_set_encoder_power_state(intel_sdvo,
+						   DRM_MODE_DPMS_ON);
+	intel_sdvo_set_active_outputs(intel_sdvo, intel_sdvo->attached_output);
+
+	if (pipe_config->has_audio)
+		intel_sdvo_enable_audio(intel_sdvo, pipe_config, conn_state);
+}
+
+static enum drm_mode_status
+intel_sdvo_mode_valid(struct drm_connector *connector,
+		      struct drm_display_mode *mode)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(to_intel_connector(connector));
+	struct intel_sdvo_connector *intel_sdvo_connector =
+		to_intel_sdvo_connector(connector);
+	bool has_hdmi_sink = intel_has_hdmi_sink(intel_sdvo_connector, connector->state);
+	int max_dotclk = i915->max_dotclk_freq;
+	enum drm_mode_status status;
+	int clock = mode->clock;
+
+	status = intel_cpu_transcoder_mode_valid(i915, mode);
+	if (status != MODE_OK)
+		return status;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return MODE_NO_DBLESCAN;
+
+	if (clock > max_dotclk)
+		return MODE_CLOCK_HIGH;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLCLK) {
+		if (!has_hdmi_sink)
+			return MODE_CLOCK_LOW;
+		clock *= 2;
+	}
+
+	if (intel_sdvo->pixel_clock_min > clock)
+		return MODE_CLOCK_LOW;
+
+	if (intel_sdvo->pixel_clock_max < clock)
+		return MODE_CLOCK_HIGH;
+
+	if (IS_LVDS(intel_sdvo_connector)) {
+		enum drm_mode_status status;
+
+		status = intel_panel_mode_valid(&intel_sdvo_connector->base, mode);
+		if (status != MODE_OK)
+			return status;
+	}
+
+	return MODE_OK;
+}
+
+static bool intel_sdvo_get_capabilities(struct intel_sdvo *intel_sdvo, struct intel_sdvo_caps *caps)
+{
+	BUILD_BUG_ON(sizeof(*caps) != 8);
+	if (!intel_sdvo_get_value(intel_sdvo,
+				  SDVO_CMD_GET_DEVICE_CAPS,
+				  caps, sizeof(*caps)))
+		return false;
+
+	DRM_DEBUG_KMS("SDVO capabilities:\n"
+		      "  vendor_id: %d\n"
+		      "  device_id: %d\n"
+		      "  device_rev_id: %d\n"
+		      "  sdvo_version_major: %d\n"
+		      "  sdvo_version_minor: %d\n"
+		      "  sdvo_inputs_mask: %d\n"
+		      "  smooth_scaling: %d\n"
+		      "  sharp_scaling: %d\n"
+		      "  up_scaling: %d\n"
+		      "  down_scaling: %d\n"
+		      "  stall_support: %d\n"
+		      "  output_flags: %d\n",
+		      caps->vendor_id,
+		      caps->device_id,
+		      caps->device_rev_id,
+		      caps->sdvo_version_major,
+		      caps->sdvo_version_minor,
+		      caps->sdvo_inputs_mask,
+		      caps->smooth_scaling,
+		      caps->sharp_scaling,
+		      caps->up_scaling,
+		      caps->down_scaling,
+		      caps->stall_support,
+		      caps->output_flags);
+
+	return true;
+}
+
+static u8 intel_sdvo_get_colorimetry_cap(struct intel_sdvo *intel_sdvo)
+{
+	u8 cap;
+
+	if (!intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_COLORIMETRY_CAP,
+				  &cap, sizeof(cap)))
+		return SDVO_COLORIMETRY_RGB256;
+
+	return cap;
+}
+
+static u16 intel_sdvo_get_hotplug_support(struct intel_sdvo *intel_sdvo)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_sdvo->base.base.dev);
+	u16 hotplug;
+
+	if (!I915_HAS_HOTPLUG(dev_priv))
+		return 0;
+
+	/*
+	 * HW Erratum: SDVO Hotplug is broken on all i945G chips, there's noise
+	 * on the line.
+	 */
+	if (IS_I945G(dev_priv) || IS_I945GM(dev_priv))
+		return 0;
+
+	if (!intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_HOT_PLUG_SUPPORT,
+					&hotplug, sizeof(hotplug)))
+		return 0;
+
+	return hotplug;
+}
+
+static void intel_sdvo_enable_hotplug(struct intel_encoder *encoder)
+{
+	struct intel_sdvo *intel_sdvo = to_sdvo(encoder);
+
+	intel_sdvo_write_cmd(intel_sdvo, SDVO_CMD_SET_ACTIVE_HOT_PLUG,
+			     &intel_sdvo->hotplug_active, 2);
+}
+
+static enum intel_hotplug_state
+intel_sdvo_hotplug(struct intel_encoder *encoder,
+		   struct intel_connector *connector)
+{
+	intel_sdvo_enable_hotplug(encoder);
+
+	return intel_encoder_hotplug(encoder, connector);
+}
+
+static bool
+intel_sdvo_multifunc_encoder(struct intel_sdvo *intel_sdvo)
+{
+	/* Is there more than one type of output? */
+	return hweight16(intel_sdvo->caps.output_flags) > 1;
+}
+
+static struct edid *
+intel_sdvo_get_edid(struct drm_connector *connector)
+{
+	struct intel_sdvo *sdvo = intel_attached_sdvo(to_intel_connector(connector));
+	return drm_get_edid(connector, &sdvo->ddc);
+}
+
+/* Mac mini hack -- use the same DDC as the analog connector */
+static struct edid *
+intel_sdvo_get_analog_edid(struct drm_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+
+	return drm_get_edid(connector,
+			    intel_gmbus_get_adapter(dev_priv,
+						    dev_priv->display.vbt.crt_ddc_pin));
+}
+
+static enum drm_connector_status
+intel_sdvo_tmds_sink_detect(struct drm_connector *connector)
+{
+	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(to_intel_connector(connector));
+	struct intel_sdvo_connector *intel_sdvo_connector =
+		to_intel_sdvo_connector(connector);
+	enum drm_connector_status status;
+	struct edid *edid;
+
+	edid = intel_sdvo_get_edid(connector);
+
+	if (edid == NULL && intel_sdvo_multifunc_encoder(intel_sdvo)) {
+		u8 ddc, saved_ddc = intel_sdvo->ddc_bus;
+
+		/*
+		 * Don't use the 1 as the argument of DDC bus switch to get
+		 * the EDID. It is used for SDVO SPD ROM.
+		 */
+		for (ddc = intel_sdvo->ddc_bus >> 1; ddc > 1; ddc >>= 1) {
+			intel_sdvo->ddc_bus = ddc;
+			edid = intel_sdvo_get_edid(connector);
+			if (edid)
+				break;
+		}
+		/*
+		 * If we found the EDID on the other bus,
+		 * assume that is the correct DDC bus.
+		 */
+		if (edid == NULL)
+			intel_sdvo->ddc_bus = saved_ddc;
+	}
+
+	/*
+	 * When there is no edid and no monitor is connected with VGA
+	 * port, try to use the CRT ddc to read the EDID for DVI-connector.
+	 */
+	if (edid == NULL)
+		edid = intel_sdvo_get_analog_edid(connector);
+
+	status = connector_status_unknown;
+	if (edid != NULL) {
+		/* DDC bus is shared, match EDID to connector type */
+		if (edid->input & DRM_EDID_INPUT_DIGITAL) {
+			status = connector_status_connected;
+			if (intel_sdvo_connector->is_hdmi) {
+				intel_sdvo->has_hdmi_audio = drm_detect_monitor_audio(edid);
+			}
+		} else
+			status = connector_status_disconnected;
+		kfree(edid);
+	}
+
+	return status;
+}
+
+static bool
+intel_sdvo_connector_matches_edid(struct intel_sdvo_connector *sdvo,
+				  struct edid *edid)
+{
+	bool monitor_is_digital = !!(edid->input & DRM_EDID_INPUT_DIGITAL);
+	bool connector_is_digital = !!IS_DIGITAL(sdvo);
+
+	DRM_DEBUG_KMS("connector_is_digital? %d, monitor_is_digital? %d\n",
+		      connector_is_digital, monitor_is_digital);
+	return connector_is_digital == monitor_is_digital;
+}
+
+static enum drm_connector_status
+intel_sdvo_detect(struct drm_connector *connector, bool force)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(to_intel_connector(connector));
+	struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector);
+	enum drm_connector_status ret;
+	u16 response;
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	if (!INTEL_DISPLAY_ENABLED(i915))
+		return connector_status_disconnected;
+
+	if (!intel_sdvo_get_value(intel_sdvo,
+				  SDVO_CMD_GET_ATTACHED_DISPLAYS,
+				  &response, 2))
+		return connector_status_unknown;
+
+	DRM_DEBUG_KMS("SDVO response %d %d [%x]\n",
+		      response & 0xff, response >> 8,
+		      intel_sdvo_connector->output_flag);
+
+	if (response == 0)
+		return connector_status_disconnected;
+
+	intel_sdvo->attached_output = response;
+
+	intel_sdvo->has_hdmi_audio = false;
+
+	if ((intel_sdvo_connector->output_flag & response) == 0)
+		ret = connector_status_disconnected;
+	else if (IS_TMDS(intel_sdvo_connector))
+		ret = intel_sdvo_tmds_sink_detect(connector);
+	else {
+		struct edid *edid;
+
+		/* if we have an edid check it matches the connection */
+		edid = intel_sdvo_get_edid(connector);
+		if (edid == NULL)
+			edid = intel_sdvo_get_analog_edid(connector);
+		if (edid != NULL) {
+			if (intel_sdvo_connector_matches_edid(intel_sdvo_connector,
+							      edid))
+				ret = connector_status_connected;
+			else
+				ret = connector_status_disconnected;
+
+			kfree(edid);
+		} else
+			ret = connector_status_connected;
+	}
+
+	return ret;
+}
+
+static int intel_sdvo_get_ddc_modes(struct drm_connector *connector)
+{
+	int num_modes = 0;
+	struct edid *edid;
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/* set the bus switch and get the modes */
+	edid = intel_sdvo_get_edid(connector);
+
+	/*
+	 * Mac mini hack.  On this device, the DVI-I connector shares one DDC
+	 * link between analog and digital outputs. So, if the regular SDVO
+	 * DDC fails, check to see if the analog output is disconnected, in
+	 * which case we'll look there for the digital DDC data.
+	 */
+	if (!edid)
+		edid = intel_sdvo_get_analog_edid(connector);
+
+	if (!edid)
+		return 0;
+
+	if (intel_sdvo_connector_matches_edid(to_intel_sdvo_connector(connector),
+					      edid))
+		num_modes += intel_connector_update_modes(connector, edid);
+
+	kfree(edid);
+
+	return num_modes;
+}
+
+/*
+ * Set of SDVO TV modes.
+ * Note!  This is in reply order (see loop in get_tv_modes).
+ * XXX: all 60Hz refresh?
+ */
+static const struct drm_display_mode sdvo_tv_modes[] = {
+	{ DRM_MODE("320x200", DRM_MODE_TYPE_DRIVER, 5815, 320, 321, 384,
+		   416, 0, 200, 201, 232, 233, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("320x240", DRM_MODE_TYPE_DRIVER, 6814, 320, 321, 384,
+		   416, 0, 240, 241, 272, 273, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("400x300", DRM_MODE_TYPE_DRIVER, 9910, 400, 401, 464,
+		   496, 0, 300, 301, 332, 333, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("640x350", DRM_MODE_TYPE_DRIVER, 16913, 640, 641, 704,
+		   736, 0, 350, 351, 382, 383, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("640x400", DRM_MODE_TYPE_DRIVER, 19121, 640, 641, 704,
+		   736, 0, 400, 401, 432, 433, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 22654, 640, 641, 704,
+		   736, 0, 480, 481, 512, 513, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("704x480", DRM_MODE_TYPE_DRIVER, 24624, 704, 705, 768,
+		   800, 0, 480, 481, 512, 513, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("704x576", DRM_MODE_TYPE_DRIVER, 29232, 704, 705, 768,
+		   800, 0, 576, 577, 608, 609, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("720x350", DRM_MODE_TYPE_DRIVER, 18751, 720, 721, 784,
+		   816, 0, 350, 351, 382, 383, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("720x400", DRM_MODE_TYPE_DRIVER, 21199, 720, 721, 784,
+		   816, 0, 400, 401, 432, 433, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 25116, 720, 721, 784,
+		   816, 0, 480, 481, 512, 513, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("720x540", DRM_MODE_TYPE_DRIVER, 28054, 720, 721, 784,
+		   816, 0, 540, 541, 572, 573, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 29816, 720, 721, 784,
+		   816, 0, 576, 577, 608, 609, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("768x576", DRM_MODE_TYPE_DRIVER, 31570, 768, 769, 832,
+		   864, 0, 576, 577, 608, 609, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 34030, 800, 801, 864,
+		   896, 0, 600, 601, 632, 633, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("832x624", DRM_MODE_TYPE_DRIVER, 36581, 832, 833, 896,
+		   928, 0, 624, 625, 656, 657, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("920x766", DRM_MODE_TYPE_DRIVER, 48707, 920, 921, 984,
+		   1016, 0, 766, 767, 798, 799, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 53827, 1024, 1025, 1088,
+		   1120, 0, 768, 769, 800, 801, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 87265, 1280, 1281, 1344,
+		   1376, 0, 1024, 1025, 1056, 1057, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+};
+
+static int intel_sdvo_get_tv_modes(struct drm_connector *connector)
+{
+	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(to_intel_connector(connector));
+	const struct drm_connector_state *conn_state = connector->state;
+	struct intel_sdvo_sdtv_resolution_request tv_res;
+	u32 reply = 0, format_map = 0;
+	int num_modes = 0;
+	int i;
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
+		      connector->base.id, connector->name);
+
+	/*
+	 * Read the list of supported input resolutions for the selected TV
+	 * format.
+	 */
+	format_map = 1 << conn_state->tv.mode;
+	memcpy(&tv_res, &format_map,
+	       min(sizeof(format_map), sizeof(struct intel_sdvo_sdtv_resolution_request)));
+
+	if (!intel_sdvo_set_target_output(intel_sdvo, intel_sdvo->attached_output))
+		return 0;
+
+	BUILD_BUG_ON(sizeof(tv_res) != 3);
+	if (!intel_sdvo_write_cmd(intel_sdvo,
+				  SDVO_CMD_GET_SDTV_RESOLUTION_SUPPORT,
+				  &tv_res, sizeof(tv_res)))
+		return 0;
+	if (!intel_sdvo_read_response(intel_sdvo, &reply, 3))
+		return 0;
+
+	for (i = 0; i < ARRAY_SIZE(sdvo_tv_modes); i++) {
+		if (reply & (1 << i)) {
+			struct drm_display_mode *nmode;
+			nmode = drm_mode_duplicate(connector->dev,
+						   &sdvo_tv_modes[i]);
+			if (nmode) {
+				drm_mode_probed_add(connector, nmode);
+				num_modes++;
+			}
+		}
+	}
+
+	return num_modes;
+}
+
+static int intel_sdvo_get_lvds_modes(struct drm_connector *connector)
+{
+	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(to_intel_connector(connector));
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	int num_modes = 0;
+
+	drm_dbg_kms(&dev_priv->drm, "[CONNECTOR:%d:%s]\n",
+		    connector->base.id, connector->name);
+
+	num_modes += intel_panel_get_modes(to_intel_connector(connector));
+	num_modes += intel_ddc_get_modes(connector, &intel_sdvo->ddc);
+
+	return num_modes;
+}
+
+static int intel_sdvo_get_modes(struct drm_connector *connector)
+{
+	struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector);
+
+	if (IS_TV(intel_sdvo_connector))
+		return intel_sdvo_get_tv_modes(connector);
+	else if (IS_LVDS(intel_sdvo_connector))
+		return intel_sdvo_get_lvds_modes(connector);
+	else
+		return intel_sdvo_get_ddc_modes(connector);
+}
+
+static int
+intel_sdvo_connector_atomic_get_property(struct drm_connector *connector,
+					 const struct drm_connector_state *state,
+					 struct drm_property *property,
+					 u64 *val)
+{
+	struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector);
+	const struct intel_sdvo_connector_state *sdvo_state = to_intel_sdvo_connector_state((void *)state);
+
+	if (property == intel_sdvo_connector->tv_format) {
+		int i;
+
+		for (i = 0; i < intel_sdvo_connector->format_supported_num; i++)
+			if (state->tv.mode == intel_sdvo_connector->tv_format_supported[i]) {
+				*val = i;
+
+				return 0;
+			}
+
+		drm_WARN_ON(connector->dev, 1);
+		*val = 0;
+	} else if (property == intel_sdvo_connector->top ||
+		   property == intel_sdvo_connector->bottom)
+		*val = intel_sdvo_connector->max_vscan - sdvo_state->tv.overscan_v;
+	else if (property == intel_sdvo_connector->left ||
+		 property == intel_sdvo_connector->right)
+		*val = intel_sdvo_connector->max_hscan - sdvo_state->tv.overscan_h;
+	else if (property == intel_sdvo_connector->hpos)
+		*val = sdvo_state->tv.hpos;
+	else if (property == intel_sdvo_connector->vpos)
+		*val = sdvo_state->tv.vpos;
+	else if (property == intel_sdvo_connector->saturation)
+		*val = state->tv.saturation;
+	else if (property == intel_sdvo_connector->contrast)
+		*val = state->tv.contrast;
+	else if (property == intel_sdvo_connector->hue)
+		*val = state->tv.hue;
+	else if (property == intel_sdvo_connector->brightness)
+		*val = state->tv.brightness;
+	else if (property == intel_sdvo_connector->sharpness)
+		*val = sdvo_state->tv.sharpness;
+	else if (property == intel_sdvo_connector->flicker_filter)
+		*val = sdvo_state->tv.flicker_filter;
+	else if (property == intel_sdvo_connector->flicker_filter_2d)
+		*val = sdvo_state->tv.flicker_filter_2d;
+	else if (property == intel_sdvo_connector->flicker_filter_adaptive)
+		*val = sdvo_state->tv.flicker_filter_adaptive;
+	else if (property == intel_sdvo_connector->tv_chroma_filter)
+		*val = sdvo_state->tv.chroma_filter;
+	else if (property == intel_sdvo_connector->tv_luma_filter)
+		*val = sdvo_state->tv.luma_filter;
+	else if (property == intel_sdvo_connector->dot_crawl)
+		*val = sdvo_state->tv.dot_crawl;
+	else
+		return intel_digital_connector_atomic_get_property(connector, state, property, val);
+
+	return 0;
+}
+
+static int
+intel_sdvo_connector_atomic_set_property(struct drm_connector *connector,
+					 struct drm_connector_state *state,
+					 struct drm_property *property,
+					 u64 val)
+{
+	struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector);
+	struct intel_sdvo_connector_state *sdvo_state = to_intel_sdvo_connector_state(state);
+
+	if (property == intel_sdvo_connector->tv_format) {
+		state->tv.mode = intel_sdvo_connector->tv_format_supported[val];
+
+		if (state->crtc) {
+			struct drm_crtc_state *crtc_state =
+				drm_atomic_get_new_crtc_state(state->state, state->crtc);
+
+			crtc_state->connectors_changed = true;
+		}
+	} else if (property == intel_sdvo_connector->top ||
+		   property == intel_sdvo_connector->bottom)
+		/* Cannot set these independent from each other */
+		sdvo_state->tv.overscan_v = intel_sdvo_connector->max_vscan - val;
+	else if (property == intel_sdvo_connector->left ||
+		 property == intel_sdvo_connector->right)
+		/* Cannot set these independent from each other */
+		sdvo_state->tv.overscan_h = intel_sdvo_connector->max_hscan - val;
+	else if (property == intel_sdvo_connector->hpos)
+		sdvo_state->tv.hpos = val;
+	else if (property == intel_sdvo_connector->vpos)
+		sdvo_state->tv.vpos = val;
+	else if (property == intel_sdvo_connector->saturation)
+		state->tv.saturation = val;
+	else if (property == intel_sdvo_connector->contrast)
+		state->tv.contrast = val;
+	else if (property == intel_sdvo_connector->hue)
+		state->tv.hue = val;
+	else if (property == intel_sdvo_connector->brightness)
+		state->tv.brightness = val;
+	else if (property == intel_sdvo_connector->sharpness)
+		sdvo_state->tv.sharpness = val;
+	else if (property == intel_sdvo_connector->flicker_filter)
+		sdvo_state->tv.flicker_filter = val;
+	else if (property == intel_sdvo_connector->flicker_filter_2d)
+		sdvo_state->tv.flicker_filter_2d = val;
+	else if (property == intel_sdvo_connector->flicker_filter_adaptive)
+		sdvo_state->tv.flicker_filter_adaptive = val;
+	else if (property == intel_sdvo_connector->tv_chroma_filter)
+		sdvo_state->tv.chroma_filter = val;
+	else if (property == intel_sdvo_connector->tv_luma_filter)
+		sdvo_state->tv.luma_filter = val;
+	else if (property == intel_sdvo_connector->dot_crawl)
+		sdvo_state->tv.dot_crawl = val;
+	else
+		return intel_digital_connector_atomic_set_property(connector, state, property, val);
+
+	return 0;
+}
+
+static int
+intel_sdvo_connector_register(struct drm_connector *connector)
+{
+	struct intel_sdvo *sdvo = intel_attached_sdvo(to_intel_connector(connector));
+	int ret;
+
+	ret = intel_connector_register(connector);
+	if (ret)
+		return ret;
+
+	return sysfs_create_link(&connector->kdev->kobj,
+				 &sdvo->ddc.dev.kobj,
+				 sdvo->ddc.dev.kobj.name);
+}
+
+static void
+intel_sdvo_connector_unregister(struct drm_connector *connector)
+{
+	struct intel_sdvo *sdvo = intel_attached_sdvo(to_intel_connector(connector));
+
+	sysfs_remove_link(&connector->kdev->kobj,
+			  sdvo->ddc.dev.kobj.name);
+	intel_connector_unregister(connector);
+}
+
+static struct drm_connector_state *
+intel_sdvo_connector_duplicate_state(struct drm_connector *connector)
+{
+	struct intel_sdvo_connector_state *state;
+
+	state = kmemdup(connector->state, sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return NULL;
+
+	__drm_atomic_helper_connector_duplicate_state(connector, &state->base.base);
+	return &state->base.base;
+}
+
+static const struct drm_connector_funcs intel_sdvo_connector_funcs = {
+	.detect = intel_sdvo_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.atomic_get_property = intel_sdvo_connector_atomic_get_property,
+	.atomic_set_property = intel_sdvo_connector_atomic_set_property,
+	.late_register = intel_sdvo_connector_register,
+	.early_unregister = intel_sdvo_connector_unregister,
+	.destroy = intel_connector_destroy,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.atomic_duplicate_state = intel_sdvo_connector_duplicate_state,
+};
+
+static int intel_sdvo_atomic_check(struct drm_connector *conn,
+				   struct drm_atomic_state *state)
+{
+	struct drm_connector_state *new_conn_state =
+		drm_atomic_get_new_connector_state(state, conn);
+	struct drm_connector_state *old_conn_state =
+		drm_atomic_get_old_connector_state(state, conn);
+	struct intel_sdvo_connector_state *old_state =
+		to_intel_sdvo_connector_state(old_conn_state);
+	struct intel_sdvo_connector_state *new_state =
+		to_intel_sdvo_connector_state(new_conn_state);
+
+	if (new_conn_state->crtc &&
+	    (memcmp(&old_state->tv, &new_state->tv, sizeof(old_state->tv)) ||
+	     memcmp(&old_conn_state->tv, &new_conn_state->tv, sizeof(old_conn_state->tv)))) {
+		struct drm_crtc_state *crtc_state =
+			drm_atomic_get_new_crtc_state(state,
+						      new_conn_state->crtc);
+
+		crtc_state->connectors_changed = true;
+	}
+
+	return intel_digital_connector_atomic_check(conn, state);
+}
+
+static const struct drm_connector_helper_funcs intel_sdvo_connector_helper_funcs = {
+	.get_modes = intel_sdvo_get_modes,
+	.mode_valid = intel_sdvo_mode_valid,
+	.atomic_check = intel_sdvo_atomic_check,
+};
+
+static void intel_sdvo_enc_destroy(struct drm_encoder *encoder)
+{
+	struct intel_sdvo *intel_sdvo = to_sdvo(to_intel_encoder(encoder));
+
+	i2c_del_adapter(&intel_sdvo->ddc);
+	intel_encoder_destroy(encoder);
+}
+
+static const struct drm_encoder_funcs intel_sdvo_enc_funcs = {
+	.destroy = intel_sdvo_enc_destroy,
+};
+
+static void
+intel_sdvo_guess_ddc_bus(struct intel_sdvo *sdvo)
+{
+	u16 mask = 0;
+	unsigned int num_bits;
+
+	/*
+	 * Make a mask of outputs less than or equal to our own priority in the
+	 * list.
+	 */
+	switch (sdvo->controlled_output) {
+	case SDVO_OUTPUT_LVDS1:
+		mask |= SDVO_OUTPUT_LVDS1;
+		fallthrough;
+	case SDVO_OUTPUT_LVDS0:
+		mask |= SDVO_OUTPUT_LVDS0;
+		fallthrough;
+	case SDVO_OUTPUT_TMDS1:
+		mask |= SDVO_OUTPUT_TMDS1;
+		fallthrough;
+	case SDVO_OUTPUT_TMDS0:
+		mask |= SDVO_OUTPUT_TMDS0;
+		fallthrough;
+	case SDVO_OUTPUT_RGB1:
+		mask |= SDVO_OUTPUT_RGB1;
+		fallthrough;
+	case SDVO_OUTPUT_RGB0:
+		mask |= SDVO_OUTPUT_RGB0;
+		break;
+	}
+
+	/* Count bits to find what number we are in the priority list. */
+	mask &= sdvo->caps.output_flags;
+	num_bits = hweight16(mask);
+	/* If more than 3 outputs, default to DDC bus 3 for now. */
+	if (num_bits > 3)
+		num_bits = 3;
+
+	/* Corresponds to SDVO_CONTROL_BUS_DDCx */
+	sdvo->ddc_bus = 1 << num_bits;
+}
+
+/*
+ * Choose the appropriate DDC bus for control bus switch command for this
+ * SDVO output based on the controlled output.
+ *
+ * DDC bus number assignment is in a priority order of RGB outputs, then TMDS
+ * outputs, then LVDS outputs.
+ */
+static void
+intel_sdvo_select_ddc_bus(struct drm_i915_private *dev_priv,
+			  struct intel_sdvo *sdvo)
+{
+	struct sdvo_device_mapping *mapping;
+
+	if (sdvo->port == PORT_B)
+		mapping = &dev_priv->display.vbt.sdvo_mappings[0];
+	else
+		mapping = &dev_priv->display.vbt.sdvo_mappings[1];
+
+	if (mapping->initialized)
+		sdvo->ddc_bus = 1 << ((mapping->ddc_pin & 0xf0) >> 4);
+	else
+		intel_sdvo_guess_ddc_bus(sdvo);
+}
+
+static void
+intel_sdvo_select_i2c_bus(struct drm_i915_private *dev_priv,
+			  struct intel_sdvo *sdvo)
+{
+	struct sdvo_device_mapping *mapping;
+	u8 pin;
+
+	if (sdvo->port == PORT_B)
+		mapping = &dev_priv->display.vbt.sdvo_mappings[0];
+	else
+		mapping = &dev_priv->display.vbt.sdvo_mappings[1];
+
+	if (mapping->initialized &&
+	    intel_gmbus_is_valid_pin(dev_priv, mapping->i2c_pin))
+		pin = mapping->i2c_pin;
+	else
+		pin = GMBUS_PIN_DPB;
+
+	sdvo->i2c = intel_gmbus_get_adapter(dev_priv, pin);
+
+	/*
+	 * With gmbus we should be able to drive sdvo i2c at 2MHz, but somehow
+	 * our code totally fails once we start using gmbus. Hence fall back to
+	 * bit banging for now.
+	 */
+	intel_gmbus_force_bit(sdvo->i2c, true);
+}
+
+/* undo any changes intel_sdvo_select_i2c_bus() did to sdvo->i2c */
+static void
+intel_sdvo_unselect_i2c_bus(struct intel_sdvo *sdvo)
+{
+	intel_gmbus_force_bit(sdvo->i2c, false);
+}
+
+static bool
+intel_sdvo_is_hdmi_connector(struct intel_sdvo *intel_sdvo, int device)
+{
+	return intel_sdvo_check_supp_encode(intel_sdvo);
+}
+
+static u8
+intel_sdvo_get_slave_addr(struct drm_i915_private *dev_priv,
+			  struct intel_sdvo *sdvo)
+{
+	struct sdvo_device_mapping *my_mapping, *other_mapping;
+
+	if (sdvo->port == PORT_B) {
+		my_mapping = &dev_priv->display.vbt.sdvo_mappings[0];
+		other_mapping = &dev_priv->display.vbt.sdvo_mappings[1];
+	} else {
+		my_mapping = &dev_priv->display.vbt.sdvo_mappings[1];
+		other_mapping = &dev_priv->display.vbt.sdvo_mappings[0];
+	}
+
+	/* If the BIOS described our SDVO device, take advantage of it. */
+	if (my_mapping->slave_addr)
+		return my_mapping->slave_addr;
+
+	/*
+	 * If the BIOS only described a different SDVO device, use the
+	 * address that it isn't using.
+	 */
+	if (other_mapping->slave_addr) {
+		if (other_mapping->slave_addr == 0x70)
+			return 0x72;
+		else
+			return 0x70;
+	}
+
+	/*
+	 * No SDVO device info is found for another DVO port,
+	 * so use mapping assumption we had before BIOS parsing.
+	 */
+	if (sdvo->port == PORT_B)
+		return 0x70;
+	else
+		return 0x72;
+}
+
+static int
+intel_sdvo_connector_init(struct intel_sdvo_connector *connector,
+			  struct intel_sdvo *encoder)
+{
+	struct drm_connector *drm_connector;
+	int ret;
+
+	drm_connector = &connector->base.base;
+	ret = drm_connector_init(encoder->base.base.dev,
+			   drm_connector,
+			   &intel_sdvo_connector_funcs,
+			   connector->base.base.connector_type);
+	if (ret < 0)
+		return ret;
+
+	drm_connector_helper_add(drm_connector,
+				 &intel_sdvo_connector_helper_funcs);
+
+	connector->base.base.interlace_allowed = 1;
+	connector->base.base.doublescan_allowed = 0;
+	connector->base.base.display_info.subpixel_order = SubPixelHorizontalRGB;
+	connector->base.get_hw_state = intel_sdvo_connector_get_hw_state;
+
+	intel_connector_attach_encoder(&connector->base, &encoder->base);
+
+	return 0;
+}
+
+static void
+intel_sdvo_add_hdmi_properties(struct intel_sdvo *intel_sdvo,
+			       struct intel_sdvo_connector *connector)
+{
+	intel_attach_force_audio_property(&connector->base.base);
+	if (intel_sdvo->colorimetry_cap & SDVO_COLORIMETRY_RGB220)
+		intel_attach_broadcast_rgb_property(&connector->base.base);
+	intel_attach_aspect_ratio_property(&connector->base.base);
+}
+
+static struct intel_sdvo_connector *intel_sdvo_connector_alloc(void)
+{
+	struct intel_sdvo_connector *sdvo_connector;
+	struct intel_sdvo_connector_state *conn_state;
+
+	sdvo_connector = kzalloc(sizeof(*sdvo_connector), GFP_KERNEL);
+	if (!sdvo_connector)
+		return NULL;
+
+	conn_state = kzalloc(sizeof(*conn_state), GFP_KERNEL);
+	if (!conn_state) {
+		kfree(sdvo_connector);
+		return NULL;
+	}
+
+	__drm_atomic_helper_connector_reset(&sdvo_connector->base.base,
+					    &conn_state->base.base);
+
+	intel_panel_init_alloc(&sdvo_connector->base);
+
+	return sdvo_connector;
+}
+
+static bool
+intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
+{
+	struct drm_encoder *encoder = &intel_sdvo->base.base;
+	struct drm_connector *connector;
+	struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
+	struct intel_connector *intel_connector;
+	struct intel_sdvo_connector *intel_sdvo_connector;
+
+	DRM_DEBUG_KMS("initialising DVI device %d\n", device);
+
+	intel_sdvo_connector = intel_sdvo_connector_alloc();
+	if (!intel_sdvo_connector)
+		return false;
+
+	if (device == 0)
+		intel_sdvo_connector->output_flag = SDVO_OUTPUT_TMDS0;
+	else if (device == 1)
+		intel_sdvo_connector->output_flag = SDVO_OUTPUT_TMDS1;
+
+	intel_connector = &intel_sdvo_connector->base;
+	connector = &intel_connector->base;
+	if (intel_sdvo_get_hotplug_support(intel_sdvo) &
+		intel_sdvo_connector->output_flag) {
+		intel_sdvo->hotplug_active |= intel_sdvo_connector->output_flag;
+		/*
+		 * Some SDVO devices have one-shot hotplug interrupts.
+		 * Ensure that they get re-enabled when an interrupt happens.
+		 */
+		intel_connector->polled = DRM_CONNECTOR_POLL_HPD;
+		intel_encoder->hotplug = intel_sdvo_hotplug;
+		intel_sdvo_enable_hotplug(intel_encoder);
+	} else {
+		intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
+	}
+	encoder->encoder_type = DRM_MODE_ENCODER_TMDS;
+	connector->connector_type = DRM_MODE_CONNECTOR_DVID;
+
+	if (intel_sdvo_is_hdmi_connector(intel_sdvo, device)) {
+		connector->connector_type = DRM_MODE_CONNECTOR_HDMIA;
+		intel_sdvo_connector->is_hdmi = true;
+	}
+
+	if (intel_sdvo_connector_init(intel_sdvo_connector, intel_sdvo) < 0) {
+		kfree(intel_sdvo_connector);
+		return false;
+	}
+
+	if (intel_sdvo_connector->is_hdmi)
+		intel_sdvo_add_hdmi_properties(intel_sdvo, intel_sdvo_connector);
+
+	return true;
+}
+
+static bool
+intel_sdvo_tv_init(struct intel_sdvo *intel_sdvo, int type)
+{
+	struct drm_encoder *encoder = &intel_sdvo->base.base;
+	struct drm_connector *connector;
+	struct intel_connector *intel_connector;
+	struct intel_sdvo_connector *intel_sdvo_connector;
+
+	DRM_DEBUG_KMS("initialising TV type %d\n", type);
+
+	intel_sdvo_connector = intel_sdvo_connector_alloc();
+	if (!intel_sdvo_connector)
+		return false;
+
+	intel_connector = &intel_sdvo_connector->base;
+	connector = &intel_connector->base;
+	encoder->encoder_type = DRM_MODE_ENCODER_TVDAC;
+	connector->connector_type = DRM_MODE_CONNECTOR_SVIDEO;
+
+	intel_sdvo_connector->output_flag = type;
+
+	if (intel_sdvo_connector_init(intel_sdvo_connector, intel_sdvo) < 0) {
+		kfree(intel_sdvo_connector);
+		return false;
+	}
+
+	if (!intel_sdvo_tv_create_property(intel_sdvo, intel_sdvo_connector, type))
+		goto err;
+
+	if (!intel_sdvo_create_enhance_property(intel_sdvo, intel_sdvo_connector))
+		goto err;
+
+	return true;
+
+err:
+	intel_connector_destroy(connector);
+	return false;
+}
+
+static bool
+intel_sdvo_analog_init(struct intel_sdvo *intel_sdvo, int device)
+{
+	struct drm_encoder *encoder = &intel_sdvo->base.base;
+	struct drm_connector *connector;
+	struct intel_connector *intel_connector;
+	struct intel_sdvo_connector *intel_sdvo_connector;
+
+	DRM_DEBUG_KMS("initialising analog device %d\n", device);
+
+	intel_sdvo_connector = intel_sdvo_connector_alloc();
+	if (!intel_sdvo_connector)
+		return false;
+
+	intel_connector = &intel_sdvo_connector->base;
+	connector = &intel_connector->base;
+	intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT;
+	encoder->encoder_type = DRM_MODE_ENCODER_DAC;
+	connector->connector_type = DRM_MODE_CONNECTOR_VGA;
+
+	if (device == 0)
+		intel_sdvo_connector->output_flag = SDVO_OUTPUT_RGB0;
+	else if (device == 1)
+		intel_sdvo_connector->output_flag = SDVO_OUTPUT_RGB1;
+
+	if (intel_sdvo_connector_init(intel_sdvo_connector, intel_sdvo) < 0) {
+		kfree(intel_sdvo_connector);
+		return false;
+	}
+
+	return true;
+}
+
+static bool
+intel_sdvo_lvds_init(struct intel_sdvo *intel_sdvo, int device)
+{
+	struct drm_encoder *encoder = &intel_sdvo->base.base;
+	struct drm_i915_private *i915 = to_i915(encoder->dev);
+	struct drm_connector *connector;
+	struct intel_connector *intel_connector;
+	struct intel_sdvo_connector *intel_sdvo_connector;
+
+	DRM_DEBUG_KMS("initialising LVDS device %d\n", device);
+
+	intel_sdvo_connector = intel_sdvo_connector_alloc();
+	if (!intel_sdvo_connector)
+		return false;
+
+	intel_connector = &intel_sdvo_connector->base;
+	connector = &intel_connector->base;
+	encoder->encoder_type = DRM_MODE_ENCODER_LVDS;
+	connector->connector_type = DRM_MODE_CONNECTOR_LVDS;
+
+	if (device == 0)
+		intel_sdvo_connector->output_flag = SDVO_OUTPUT_LVDS0;
+	else if (device == 1)
+		intel_sdvo_connector->output_flag = SDVO_OUTPUT_LVDS1;
+
+	if (intel_sdvo_connector_init(intel_sdvo_connector, intel_sdvo) < 0) {
+		kfree(intel_sdvo_connector);
+		return false;
+	}
+
+	if (!intel_sdvo_create_enhance_property(intel_sdvo, intel_sdvo_connector))
+		goto err;
+
+	intel_bios_init_panel_late(i915, &intel_connector->panel, NULL, NULL);
+
+	/*
+	 * Fetch modes from VBT. For SDVO prefer the VBT mode since some
+	 * SDVO->LVDS transcoders can't cope with the EDID mode.
+	 */
+	intel_panel_add_vbt_sdvo_fixed_mode(intel_connector);
+
+	if (!intel_panel_preferred_fixed_mode(intel_connector)) {
+		mutex_lock(&i915->drm.mode_config.mutex);
+
+		intel_ddc_get_modes(connector, &intel_sdvo->ddc);
+		intel_panel_add_edid_fixed_modes(intel_connector, false);
+
+		mutex_unlock(&i915->drm.mode_config.mutex);
+	}
+
+	intel_panel_init(intel_connector);
+
+	if (!intel_panel_preferred_fixed_mode(intel_connector))
+		goto err;
+
+	return true;
+
+err:
+	intel_connector_destroy(connector);
+	return false;
+}
+
+static u16 intel_sdvo_filter_output_flags(u16 flags)
+{
+	flags &= SDVO_OUTPUT_MASK;
+
+	/* SDVO requires XXX1 function may not exist unless it has XXX0 function.*/
+	if (!(flags & SDVO_OUTPUT_TMDS0))
+		flags &= ~SDVO_OUTPUT_TMDS1;
+
+	if (!(flags & SDVO_OUTPUT_RGB0))
+		flags &= ~SDVO_OUTPUT_RGB1;
+
+	if (!(flags & SDVO_OUTPUT_LVDS0))
+		flags &= ~SDVO_OUTPUT_LVDS1;
+
+	return flags;
+}
+
+static bool
+intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, u16 flags)
+{
+	struct drm_i915_private *i915 = to_i915(intel_sdvo->base.base.dev);
+
+	flags = intel_sdvo_filter_output_flags(flags);
+
+	intel_sdvo->controlled_output = flags;
+
+	intel_sdvo_select_ddc_bus(i915, intel_sdvo);
+
+	if (flags & SDVO_OUTPUT_TMDS0)
+		if (!intel_sdvo_dvi_init(intel_sdvo, 0))
+			return false;
+
+	if (flags & SDVO_OUTPUT_TMDS1)
+		if (!intel_sdvo_dvi_init(intel_sdvo, 1))
+			return false;
+
+	/* TV has no XXX1 function block */
+	if (flags & SDVO_OUTPUT_SVID0)
+		if (!intel_sdvo_tv_init(intel_sdvo, SDVO_OUTPUT_SVID0))
+			return false;
+
+	if (flags & SDVO_OUTPUT_CVBS0)
+		if (!intel_sdvo_tv_init(intel_sdvo, SDVO_OUTPUT_CVBS0))
+			return false;
+
+	if (flags & SDVO_OUTPUT_YPRPB0)
+		if (!intel_sdvo_tv_init(intel_sdvo, SDVO_OUTPUT_YPRPB0))
+			return false;
+
+	if (flags & SDVO_OUTPUT_RGB0)
+		if (!intel_sdvo_analog_init(intel_sdvo, 0))
+			return false;
+
+	if (flags & SDVO_OUTPUT_RGB1)
+		if (!intel_sdvo_analog_init(intel_sdvo, 1))
+			return false;
+
+	if (flags & SDVO_OUTPUT_LVDS0)
+		if (!intel_sdvo_lvds_init(intel_sdvo, 0))
+			return false;
+
+	if (flags & SDVO_OUTPUT_LVDS1)
+		if (!intel_sdvo_lvds_init(intel_sdvo, 1))
+			return false;
+
+	if (flags == 0) {
+		unsigned char bytes[2];
+
+		memcpy(bytes, &intel_sdvo->caps.output_flags, 2);
+		DRM_DEBUG_KMS("%s: Unknown SDVO output type (0x%02x%02x)\n",
+			      SDVO_NAME(intel_sdvo),
+			      bytes[0], bytes[1]);
+		return false;
+	}
+	intel_sdvo->base.pipe_mask = ~0;
+
+	return true;
+}
+
+static void intel_sdvo_output_cleanup(struct intel_sdvo *intel_sdvo)
+{
+	struct drm_device *dev = intel_sdvo->base.base.dev;
+	struct drm_connector *connector, *tmp;
+
+	list_for_each_entry_safe(connector, tmp,
+				 &dev->mode_config.connector_list, head) {
+		if (intel_attached_encoder(to_intel_connector(connector)) == &intel_sdvo->base) {
+			drm_connector_unregister(connector);
+			intel_connector_destroy(connector);
+		}
+	}
+}
+
+static bool intel_sdvo_tv_create_property(struct intel_sdvo *intel_sdvo,
+					  struct intel_sdvo_connector *intel_sdvo_connector,
+					  int type)
+{
+	struct drm_device *dev = intel_sdvo->base.base.dev;
+	struct intel_sdvo_tv_format format;
+	u32 format_map, i;
+
+	if (!intel_sdvo_set_target_output(intel_sdvo, type))
+		return false;
+
+	BUILD_BUG_ON(sizeof(format) != 6);
+	if (!intel_sdvo_get_value(intel_sdvo,
+				  SDVO_CMD_GET_SUPPORTED_TV_FORMATS,
+				  &format, sizeof(format)))
+		return false;
+
+	memcpy(&format_map, &format, min(sizeof(format_map), sizeof(format)));
+
+	if (format_map == 0)
+		return false;
+
+	intel_sdvo_connector->format_supported_num = 0;
+	for (i = 0 ; i < TV_FORMAT_NUM; i++)
+		if (format_map & (1 << i))
+			intel_sdvo_connector->tv_format_supported[intel_sdvo_connector->format_supported_num++] = i;
+
+
+	intel_sdvo_connector->tv_format =
+			drm_property_create(dev, DRM_MODE_PROP_ENUM,
+					    "mode", intel_sdvo_connector->format_supported_num);
+	if (!intel_sdvo_connector->tv_format)
+		return false;
+
+	for (i = 0; i < intel_sdvo_connector->format_supported_num; i++)
+		drm_property_add_enum(intel_sdvo_connector->tv_format, i,
+				      tv_format_names[intel_sdvo_connector->tv_format_supported[i]]);
+
+	intel_sdvo_connector->base.base.state->tv.mode = intel_sdvo_connector->tv_format_supported[0];
+	drm_object_attach_property(&intel_sdvo_connector->base.base.base,
+				   intel_sdvo_connector->tv_format, 0);
+	return true;
+
+}
+
+#define _ENHANCEMENT(state_assignment, name, NAME) do { \
+	if (enhancements.name) { \
+		if (!intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_MAX_##NAME, &data_value, 4) || \
+		    !intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_##NAME, &response, 2)) \
+			return false; \
+		intel_sdvo_connector->name = \
+			drm_property_create_range(dev, 0, #name, 0, data_value[0]); \
+		if (!intel_sdvo_connector->name) return false; \
+		state_assignment = response; \
+		drm_object_attach_property(&connector->base, \
+					   intel_sdvo_connector->name, 0); \
+		DRM_DEBUG_KMS(#name ": max %d, default %d, current %d\n", \
+			      data_value[0], data_value[1], response); \
+	} \
+} while (0)
+
+#define ENHANCEMENT(state, name, NAME) _ENHANCEMENT((state)->name, name, NAME)
+
+static bool
+intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo,
+				      struct intel_sdvo_connector *intel_sdvo_connector,
+				      struct intel_sdvo_enhancements_reply enhancements)
+{
+	struct drm_device *dev = intel_sdvo->base.base.dev;
+	struct drm_connector *connector = &intel_sdvo_connector->base.base;
+	struct drm_connector_state *conn_state = connector->state;
+	struct intel_sdvo_connector_state *sdvo_state =
+		to_intel_sdvo_connector_state(conn_state);
+	u16 response, data_value[2];
+
+	/* when horizontal overscan is supported, Add the left/right property */
+	if (enhancements.overscan_h) {
+		if (!intel_sdvo_get_value(intel_sdvo,
+					  SDVO_CMD_GET_MAX_OVERSCAN_H,
+					  &data_value, 4))
+			return false;
+
+		if (!intel_sdvo_get_value(intel_sdvo,
+					  SDVO_CMD_GET_OVERSCAN_H,
+					  &response, 2))
+			return false;
+
+		sdvo_state->tv.overscan_h = response;
+
+		intel_sdvo_connector->max_hscan = data_value[0];
+		intel_sdvo_connector->left =
+			drm_property_create_range(dev, 0, "left_margin", 0, data_value[0]);
+		if (!intel_sdvo_connector->left)
+			return false;
+
+		drm_object_attach_property(&connector->base,
+					   intel_sdvo_connector->left, 0);
+
+		intel_sdvo_connector->right =
+			drm_property_create_range(dev, 0, "right_margin", 0, data_value[0]);
+		if (!intel_sdvo_connector->right)
+			return false;
+
+		drm_object_attach_property(&connector->base,
+					      intel_sdvo_connector->right, 0);
+		DRM_DEBUG_KMS("h_overscan: max %d, "
+			      "default %d, current %d\n",
+			      data_value[0], data_value[1], response);
+	}
+
+	if (enhancements.overscan_v) {
+		if (!intel_sdvo_get_value(intel_sdvo,
+					  SDVO_CMD_GET_MAX_OVERSCAN_V,
+					  &data_value, 4))
+			return false;
+
+		if (!intel_sdvo_get_value(intel_sdvo,
+					  SDVO_CMD_GET_OVERSCAN_V,
+					  &response, 2))
+			return false;
+
+		sdvo_state->tv.overscan_v = response;
+
+		intel_sdvo_connector->max_vscan = data_value[0];
+		intel_sdvo_connector->top =
+			drm_property_create_range(dev, 0,
+					    "top_margin", 0, data_value[0]);
+		if (!intel_sdvo_connector->top)
+			return false;
+
+		drm_object_attach_property(&connector->base,
+					   intel_sdvo_connector->top, 0);
+
+		intel_sdvo_connector->bottom =
+			drm_property_create_range(dev, 0,
+					    "bottom_margin", 0, data_value[0]);
+		if (!intel_sdvo_connector->bottom)
+			return false;
+
+		drm_object_attach_property(&connector->base,
+					      intel_sdvo_connector->bottom, 0);
+		DRM_DEBUG_KMS("v_overscan: max %d, "
+			      "default %d, current %d\n",
+			      data_value[0], data_value[1], response);
+	}
+
+	ENHANCEMENT(&sdvo_state->tv, hpos, HPOS);
+	ENHANCEMENT(&sdvo_state->tv, vpos, VPOS);
+	ENHANCEMENT(&conn_state->tv, saturation, SATURATION);
+	ENHANCEMENT(&conn_state->tv, contrast, CONTRAST);
+	ENHANCEMENT(&conn_state->tv, hue, HUE);
+	ENHANCEMENT(&conn_state->tv, brightness, BRIGHTNESS);
+	ENHANCEMENT(&sdvo_state->tv, sharpness, SHARPNESS);
+	ENHANCEMENT(&sdvo_state->tv, flicker_filter, FLICKER_FILTER);
+	ENHANCEMENT(&sdvo_state->tv, flicker_filter_adaptive, FLICKER_FILTER_ADAPTIVE);
+	ENHANCEMENT(&sdvo_state->tv, flicker_filter_2d, FLICKER_FILTER_2D);
+	_ENHANCEMENT(sdvo_state->tv.chroma_filter, tv_chroma_filter, TV_CHROMA_FILTER);
+	_ENHANCEMENT(sdvo_state->tv.luma_filter, tv_luma_filter, TV_LUMA_FILTER);
+
+	if (enhancements.dot_crawl) {
+		if (!intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_DOT_CRAWL, &response, 2))
+			return false;
+
+		sdvo_state->tv.dot_crawl = response & 0x1;
+		intel_sdvo_connector->dot_crawl =
+			drm_property_create_range(dev, 0, "dot_crawl", 0, 1);
+		if (!intel_sdvo_connector->dot_crawl)
+			return false;
+
+		drm_object_attach_property(&connector->base,
+					   intel_sdvo_connector->dot_crawl, 0);
+		DRM_DEBUG_KMS("dot crawl: current %d\n", response);
+	}
+
+	return true;
+}
+
+static bool
+intel_sdvo_create_enhance_property_lvds(struct intel_sdvo *intel_sdvo,
+					struct intel_sdvo_connector *intel_sdvo_connector,
+					struct intel_sdvo_enhancements_reply enhancements)
+{
+	struct drm_device *dev = intel_sdvo->base.base.dev;
+	struct drm_connector *connector = &intel_sdvo_connector->base.base;
+	u16 response, data_value[2];
+
+	ENHANCEMENT(&connector->state->tv, brightness, BRIGHTNESS);
+
+	return true;
+}
+#undef ENHANCEMENT
+#undef _ENHANCEMENT
+
+static bool intel_sdvo_create_enhance_property(struct intel_sdvo *intel_sdvo,
+					       struct intel_sdvo_connector *intel_sdvo_connector)
+{
+	union {
+		struct intel_sdvo_enhancements_reply reply;
+		u16 response;
+	} enhancements;
+
+	BUILD_BUG_ON(sizeof(enhancements) != 2);
+
+	if (!intel_sdvo_get_value(intel_sdvo,
+				  SDVO_CMD_GET_SUPPORTED_ENHANCEMENTS,
+				  &enhancements, sizeof(enhancements)) ||
+	    enhancements.response == 0) {
+		DRM_DEBUG_KMS("No enhancement is supported\n");
+		return true;
+	}
+
+	if (IS_TV(intel_sdvo_connector))
+		return intel_sdvo_create_enhance_property_tv(intel_sdvo, intel_sdvo_connector, enhancements.reply);
+	else if (IS_LVDS(intel_sdvo_connector))
+		return intel_sdvo_create_enhance_property_lvds(intel_sdvo, intel_sdvo_connector, enhancements.reply);
+	else
+		return true;
+}
+
+static int intel_sdvo_ddc_proxy_xfer(struct i2c_adapter *adapter,
+				     struct i2c_msg *msgs,
+				     int num)
+{
+	struct intel_sdvo *sdvo = adapter->algo_data;
+
+	if (!__intel_sdvo_set_control_bus_switch(sdvo, sdvo->ddc_bus))
+		return -EIO;
+
+	return sdvo->i2c->algo->master_xfer(sdvo->i2c, msgs, num);
+}
+
+static u32 intel_sdvo_ddc_proxy_func(struct i2c_adapter *adapter)
+{
+	struct intel_sdvo *sdvo = adapter->algo_data;
+	return sdvo->i2c->algo->functionality(sdvo->i2c);
+}
+
+static const struct i2c_algorithm intel_sdvo_ddc_proxy = {
+	.master_xfer	= intel_sdvo_ddc_proxy_xfer,
+	.functionality	= intel_sdvo_ddc_proxy_func
+};
+
+static void proxy_lock_bus(struct i2c_adapter *adapter,
+			   unsigned int flags)
+{
+	struct intel_sdvo *sdvo = adapter->algo_data;
+	sdvo->i2c->lock_ops->lock_bus(sdvo->i2c, flags);
+}
+
+static int proxy_trylock_bus(struct i2c_adapter *adapter,
+			     unsigned int flags)
+{
+	struct intel_sdvo *sdvo = adapter->algo_data;
+	return sdvo->i2c->lock_ops->trylock_bus(sdvo->i2c, flags);
+}
+
+static void proxy_unlock_bus(struct i2c_adapter *adapter,
+			     unsigned int flags)
+{
+	struct intel_sdvo *sdvo = adapter->algo_data;
+	sdvo->i2c->lock_ops->unlock_bus(sdvo->i2c, flags);
+}
+
+static const struct i2c_lock_operations proxy_lock_ops = {
+	.lock_bus =    proxy_lock_bus,
+	.trylock_bus = proxy_trylock_bus,
+	.unlock_bus =  proxy_unlock_bus,
+};
+
+static bool
+intel_sdvo_init_ddc_proxy(struct intel_sdvo *sdvo,
+			  struct drm_i915_private *dev_priv)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+
+	sdvo->ddc.owner = THIS_MODULE;
+	sdvo->ddc.class = I2C_CLASS_DDC;
+	snprintf(sdvo->ddc.name, I2C_NAME_SIZE, "SDVO DDC proxy");
+	sdvo->ddc.dev.parent = &pdev->dev;
+	sdvo->ddc.algo_data = sdvo;
+	sdvo->ddc.algo = &intel_sdvo_ddc_proxy;
+	sdvo->ddc.lock_ops = &proxy_lock_ops;
+
+	return i2c_add_adapter(&sdvo->ddc) == 0;
+}
+
+static void assert_sdvo_port_valid(const struct drm_i915_private *dev_priv,
+				   enum port port)
+{
+	if (HAS_PCH_SPLIT(dev_priv))
+		drm_WARN_ON(&dev_priv->drm, port != PORT_B);
+	else
+		drm_WARN_ON(&dev_priv->drm, port != PORT_B && port != PORT_C);
+}
+
+bool intel_sdvo_init(struct drm_i915_private *dev_priv,
+		     i915_reg_t sdvo_reg, enum port port)
+{
+	struct intel_encoder *intel_encoder;
+	struct intel_sdvo *intel_sdvo;
+	int i;
+
+	assert_sdvo_port_valid(dev_priv, port);
+
+	intel_sdvo = kzalloc(sizeof(*intel_sdvo), GFP_KERNEL);
+	if (!intel_sdvo)
+		return false;
+
+	intel_sdvo->sdvo_reg = sdvo_reg;
+	intel_sdvo->port = port;
+	intel_sdvo->slave_addr =
+		intel_sdvo_get_slave_addr(dev_priv, intel_sdvo) >> 1;
+	intel_sdvo_select_i2c_bus(dev_priv, intel_sdvo);
+	if (!intel_sdvo_init_ddc_proxy(intel_sdvo, dev_priv))
+		goto err_i2c_bus;
+
+	/* encoder type will be decided later */
+	intel_encoder = &intel_sdvo->base;
+	intel_encoder->type = INTEL_OUTPUT_SDVO;
+	intel_encoder->power_domain = POWER_DOMAIN_PORT_OTHER;
+	intel_encoder->port = port;
+	drm_encoder_init(&dev_priv->drm, &intel_encoder->base,
+			 &intel_sdvo_enc_funcs, 0,
+			 "SDVO %c", port_name(port));
+
+	/* Read the regs to test if we can talk to the device */
+	for (i = 0; i < 0x40; i++) {
+		u8 byte;
+
+		if (!intel_sdvo_read_byte(intel_sdvo, i, &byte)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "No SDVO device found on %s\n",
+				    SDVO_NAME(intel_sdvo));
+			goto err;
+		}
+	}
+
+	intel_encoder->compute_config = intel_sdvo_compute_config;
+	if (HAS_PCH_SPLIT(dev_priv)) {
+		intel_encoder->disable = pch_disable_sdvo;
+		intel_encoder->post_disable = pch_post_disable_sdvo;
+	} else {
+		intel_encoder->disable = intel_disable_sdvo;
+	}
+	intel_encoder->pre_enable = intel_sdvo_pre_enable;
+	intel_encoder->enable = intel_enable_sdvo;
+	intel_encoder->get_hw_state = intel_sdvo_get_hw_state;
+	intel_encoder->get_config = intel_sdvo_get_config;
+
+	/* In default case sdvo lvds is false */
+	if (!intel_sdvo_get_capabilities(intel_sdvo, &intel_sdvo->caps))
+		goto err;
+
+	intel_sdvo->colorimetry_cap =
+		intel_sdvo_get_colorimetry_cap(intel_sdvo);
+
+	if (intel_sdvo_output_setup(intel_sdvo,
+				    intel_sdvo->caps.output_flags) != true) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "SDVO output failed to setup on %s\n",
+			    SDVO_NAME(intel_sdvo));
+		/* Output_setup can leave behind connectors! */
+		goto err_output;
+	}
+
+	/*
+	 * Only enable the hotplug irq if we need it, to work around noisy
+	 * hotplug lines.
+	 */
+	if (intel_sdvo->hotplug_active) {
+		if (intel_sdvo->port == PORT_B)
+			intel_encoder->hpd_pin = HPD_SDVO_B;
+		else
+			intel_encoder->hpd_pin = HPD_SDVO_C;
+	}
+
+	/*
+	 * Cloning SDVO with anything is often impossible, since the SDVO
+	 * encoder can request a special input timing mode. And even if that's
+	 * not the case we have evidence that cloning a plain unscaled mode with
+	 * VGA doesn't really work. Furthermore the cloning flags are way too
+	 * simplistic anyway to express such constraints, so just give up on
+	 * cloning for SDVO encoders.
+	 */
+	intel_sdvo->base.cloneable = 0;
+
+	/* Set the input timing to the screen. Assume always input 0. */
+	if (!intel_sdvo_set_target_input(intel_sdvo))
+		goto err_output;
+
+	if (!intel_sdvo_get_input_pixel_clock_range(intel_sdvo,
+						    &intel_sdvo->pixel_clock_min,
+						    &intel_sdvo->pixel_clock_max))
+		goto err_output;
+
+	drm_dbg_kms(&dev_priv->drm, "%s device VID/DID: %02X:%02X.%02X, "
+			"clock range %dMHz - %dMHz, "
+			"input 1: %c, input 2: %c, "
+			"output 1: %c, output 2: %c\n",
+			SDVO_NAME(intel_sdvo),
+			intel_sdvo->caps.vendor_id, intel_sdvo->caps.device_id,
+			intel_sdvo->caps.device_rev_id,
+			intel_sdvo->pixel_clock_min / 1000,
+			intel_sdvo->pixel_clock_max / 1000,
+			(intel_sdvo->caps.sdvo_inputs_mask & 0x1) ? 'Y' : 'N',
+			(intel_sdvo->caps.sdvo_inputs_mask & 0x2) ? 'Y' : 'N',
+			/* check currently supported outputs */
+			intel_sdvo->caps.output_flags &
+			(SDVO_OUTPUT_TMDS0 | SDVO_OUTPUT_RGB0) ? 'Y' : 'N',
+			intel_sdvo->caps.output_flags &
+			(SDVO_OUTPUT_TMDS1 | SDVO_OUTPUT_RGB1) ? 'Y' : 'N');
+	return true;
+
+err_output:
+	intel_sdvo_output_cleanup(intel_sdvo);
+
+err:
+	drm_encoder_cleanup(&intel_encoder->base);
+	i2c_del_adapter(&intel_sdvo->ddc);
+err_i2c_bus:
+	intel_sdvo_unselect_i2c_bus(intel_sdvo);
+	kfree(intel_sdvo);
+
+	return false;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_sdvo.h b/drivers/gpu/drm/i915/display/intel_sdvo.h
new file mode 100644
index 000000000..2868852c8
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_sdvo.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_SDVO_H__
+#define __INTEL_SDVO_H__
+
+#include <linux/types.h>
+
+#include "i915_reg_defs.h"
+
+struct drm_i915_private;
+enum pipe;
+enum port;
+
+bool intel_sdvo_port_enabled(struct drm_i915_private *dev_priv,
+			     i915_reg_t sdvo_reg, enum pipe *pipe);
+bool intel_sdvo_init(struct drm_i915_private *dev_priv,
+		     i915_reg_t reg, enum port port);
+
+#endif /* __INTEL_SDVO_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_sdvo_regs.h b/drivers/gpu/drm/i915/display/intel_sdvo_regs.h
new file mode 100644
index 000000000..74dc6c042
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_sdvo_regs.h
@@ -0,0 +1,741 @@
+/*
+ * Copyright © 2006-2007 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Eric Anholt <eric@anholt.net>
+ */
+
+#ifndef __INTEL_SDVO_REGS_H__
+#define __INTEL_SDVO_REGS_H__
+
+#include <linux/compiler.h>
+#include <linux/types.h>
+
+/*
+ * SDVO command definitions and structures.
+ */
+
+#define SDVO_OUTPUT_FIRST   (0)
+#define SDVO_OUTPUT_TMDS0   (1 << 0)
+#define SDVO_OUTPUT_RGB0    (1 << 1)
+#define SDVO_OUTPUT_CVBS0   (1 << 2)
+#define SDVO_OUTPUT_SVID0   (1 << 3)
+#define SDVO_OUTPUT_YPRPB0  (1 << 4)
+#define SDVO_OUTPUT_SCART0  (1 << 5)
+#define SDVO_OUTPUT_LVDS0   (1 << 6)
+#define SDVO_OUTPUT_TMDS1   (1 << 8)
+#define SDVO_OUTPUT_RGB1    (1 << 9)
+#define SDVO_OUTPUT_CVBS1   (1 << 10)
+#define SDVO_OUTPUT_SVID1   (1 << 11)
+#define SDVO_OUTPUT_YPRPB1  (1 << 12)
+#define SDVO_OUTPUT_SCART1  (1 << 13)
+#define SDVO_OUTPUT_LVDS1   (1 << 14)
+#define SDVO_OUTPUT_LAST    (14)
+
+struct intel_sdvo_caps {
+	u8 vendor_id;
+	u8 device_id;
+	u8 device_rev_id;
+	u8 sdvo_version_major;
+	u8 sdvo_version_minor;
+	unsigned int sdvo_inputs_mask:2;
+	unsigned int smooth_scaling:1;
+	unsigned int sharp_scaling:1;
+	unsigned int up_scaling:1;
+	unsigned int down_scaling:1;
+	unsigned int stall_support:1;
+	unsigned int pad:1;
+	u16 output_flags;
+} __packed;
+
+/* Note: SDVO detailed timing flags match EDID misc flags. */
+#define DTD_FLAG_HSYNC_POSITIVE (1 << 1)
+#define DTD_FLAG_VSYNC_POSITIVE (1 << 2)
+#define DTD_FLAG_INTERLACE	(1 << 7)
+
+/* This matches the EDID DTD structure, more or less */
+struct intel_sdvo_dtd {
+	struct {
+		u16 clock;	/* pixel clock, in 10kHz units */
+		u8 h_active;	/* lower 8 bits (pixels) */
+		u8 h_blank;	/* lower 8 bits (pixels) */
+		u8 h_high;	/* upper 4 bits each h_active, h_blank */
+		u8 v_active;	/* lower 8 bits (lines) */
+		u8 v_blank;	/* lower 8 bits (lines) */
+		u8 v_high;	/* upper 4 bits each v_active, v_blank */
+	} part1;
+
+	struct {
+		u8 h_sync_off;	/* lower 8 bits, from hblank start */
+		u8 h_sync_width;	/* lower 8 bits (pixels) */
+		/* lower 4 bits each vsync offset, vsync width */
+		u8 v_sync_off_width;
+		/*
+		* 2 high bits of hsync offset, 2 high bits of hsync width,
+		* bits 4-5 of vsync offset, and 2 high bits of vsync width.
+		*/
+		u8 sync_off_width_high;
+		u8 dtd_flags;
+		u8 sdvo_flags;
+		/* bits 6-7 of vsync offset at bits 6-7 */
+		u8 v_sync_off_high;
+		u8 reserved;
+	} part2;
+} __packed;
+
+struct intel_sdvo_pixel_clock_range {
+	u16 min;	/* pixel clock, in 10kHz units */
+	u16 max;	/* pixel clock, in 10kHz units */
+} __packed;
+
+struct intel_sdvo_preferred_input_timing_args {
+	u16 clock;
+	u16 width;
+	u16 height;
+	u8	interlace:1;
+	u8	scaled:1;
+	u8	pad:6;
+} __packed;
+
+/* I2C registers for SDVO */
+#define SDVO_I2C_ARG_0				0x07
+#define SDVO_I2C_ARG_1				0x06
+#define SDVO_I2C_ARG_2				0x05
+#define SDVO_I2C_ARG_3				0x04
+#define SDVO_I2C_ARG_4				0x03
+#define SDVO_I2C_ARG_5				0x02
+#define SDVO_I2C_ARG_6				0x01
+#define SDVO_I2C_ARG_7				0x00
+#define SDVO_I2C_OPCODE				0x08
+#define SDVO_I2C_CMD_STATUS			0x09
+#define SDVO_I2C_RETURN_0			0x0a
+#define SDVO_I2C_RETURN_1			0x0b
+#define SDVO_I2C_RETURN_2			0x0c
+#define SDVO_I2C_RETURN_3			0x0d
+#define SDVO_I2C_RETURN_4			0x0e
+#define SDVO_I2C_RETURN_5			0x0f
+#define SDVO_I2C_RETURN_6			0x10
+#define SDVO_I2C_RETURN_7			0x11
+#define SDVO_I2C_VENDOR_BEGIN			0x20
+
+/* Status results */
+#define SDVO_CMD_STATUS_POWER_ON		0x0
+#define SDVO_CMD_STATUS_SUCCESS			0x1
+#define SDVO_CMD_STATUS_NOTSUPP			0x2
+#define SDVO_CMD_STATUS_INVALID_ARG		0x3
+#define SDVO_CMD_STATUS_PENDING			0x4
+#define SDVO_CMD_STATUS_TARGET_NOT_SPECIFIED	0x5
+#define SDVO_CMD_STATUS_SCALING_NOT_SUPP	0x6
+
+/* SDVO commands, argument/result registers */
+
+#define SDVO_CMD_RESET					0x01
+
+/* Returns a struct intel_sdvo_caps */
+#define SDVO_CMD_GET_DEVICE_CAPS			0x02
+
+#define SDVO_CMD_GET_FIRMWARE_REV			0x86
+# define SDVO_DEVICE_FIRMWARE_MINOR			SDVO_I2C_RETURN_0
+# define SDVO_DEVICE_FIRMWARE_MAJOR			SDVO_I2C_RETURN_1
+# define SDVO_DEVICE_FIRMWARE_PATCH			SDVO_I2C_RETURN_2
+
+/*
+ * Reports which inputs are trained (managed to sync).
+ *
+ * Devices must have trained within 2 vsyncs of a mode change.
+ */
+#define SDVO_CMD_GET_TRAINED_INPUTS			0x03
+struct intel_sdvo_get_trained_inputs_response {
+	unsigned int input0_trained:1;
+	unsigned int input1_trained:1;
+	unsigned int pad:6;
+} __packed;
+
+/* Returns a struct intel_sdvo_output_flags of active outputs. */
+#define SDVO_CMD_GET_ACTIVE_OUTPUTS			0x04
+
+/*
+ * Sets the current set of active outputs.
+ *
+ * Takes a struct intel_sdvo_output_flags.  Must be preceded by a SET_IN_OUT_MAP
+ * on multi-output devices.
+ */
+#define SDVO_CMD_SET_ACTIVE_OUTPUTS			0x05
+
+/*
+ * Returns the current mapping of SDVO inputs to outputs on the device.
+ *
+ * Returns two struct intel_sdvo_output_flags structures.
+ */
+#define SDVO_CMD_GET_IN_OUT_MAP				0x06
+struct intel_sdvo_in_out_map {
+	u16 in0, in1;
+};
+
+/*
+ * Sets the current mapping of SDVO inputs to outputs on the device.
+ *
+ * Takes two struct i380_sdvo_output_flags structures.
+ */
+#define SDVO_CMD_SET_IN_OUT_MAP				0x07
+
+/*
+ * Returns a struct intel_sdvo_output_flags of attached displays.
+ */
+#define SDVO_CMD_GET_ATTACHED_DISPLAYS			0x0b
+
+/*
+ * Returns a struct intel_sdvo_ouptut_flags of displays supporting hot plugging.
+ */
+#define SDVO_CMD_GET_HOT_PLUG_SUPPORT			0x0c
+
+/*
+ * Takes a struct intel_sdvo_output_flags.
+ */
+#define SDVO_CMD_SET_ACTIVE_HOT_PLUG			0x0d
+
+/*
+ * Returns a struct intel_sdvo_output_flags of displays with hot plug
+ * interrupts enabled.
+ */
+#define SDVO_CMD_GET_ACTIVE_HOT_PLUG			0x0e
+
+#define SDVO_CMD_GET_INTERRUPT_EVENT_SOURCE		0x0f
+struct intel_sdvo_get_interrupt_event_source_response {
+	u16 interrupt_status;
+	unsigned int ambient_light_interrupt:1;
+	unsigned int hdmi_audio_encrypt_change:1;
+	unsigned int pad:6;
+} __packed;
+
+/*
+ * Selects which input is affected by future input commands.
+ *
+ * Commands affected include SET_INPUT_TIMINGS_PART[12],
+ * GET_INPUT_TIMINGS_PART[12], GET_PREFERRED_INPUT_TIMINGS_PART[12],
+ * GET_INPUT_PIXEL_CLOCK_RANGE, and CREATE_PREFERRED_INPUT_TIMINGS.
+ */
+#define SDVO_CMD_SET_TARGET_INPUT			0x10
+struct intel_sdvo_set_target_input_args {
+	unsigned int target_1:1;
+	unsigned int pad:7;
+} __packed;
+
+/*
+ * Takes a struct intel_sdvo_output_flags of which outputs are targeted by
+ * future output commands.
+ *
+ * Affected commands inclue SET_OUTPUT_TIMINGS_PART[12],
+ * GET_OUTPUT_TIMINGS_PART[12], and GET_OUTPUT_PIXEL_CLOCK_RANGE.
+ */
+#define SDVO_CMD_SET_TARGET_OUTPUT			0x11
+
+#define SDVO_CMD_GET_INPUT_TIMINGS_PART1		0x12
+#define SDVO_CMD_GET_INPUT_TIMINGS_PART2		0x13
+#define SDVO_CMD_SET_INPUT_TIMINGS_PART1		0x14
+#define SDVO_CMD_SET_INPUT_TIMINGS_PART2		0x15
+#define SDVO_CMD_SET_OUTPUT_TIMINGS_PART1		0x16
+#define SDVO_CMD_SET_OUTPUT_TIMINGS_PART2		0x17
+#define SDVO_CMD_GET_OUTPUT_TIMINGS_PART1		0x18
+#define SDVO_CMD_GET_OUTPUT_TIMINGS_PART2		0x19
+/* Part 1 */
+# define SDVO_DTD_CLOCK_LOW				SDVO_I2C_ARG_0
+# define SDVO_DTD_CLOCK_HIGH				SDVO_I2C_ARG_1
+# define SDVO_DTD_H_ACTIVE				SDVO_I2C_ARG_2
+# define SDVO_DTD_H_BLANK				SDVO_I2C_ARG_3
+# define SDVO_DTD_H_HIGH				SDVO_I2C_ARG_4
+# define SDVO_DTD_V_ACTIVE				SDVO_I2C_ARG_5
+# define SDVO_DTD_V_BLANK				SDVO_I2C_ARG_6
+# define SDVO_DTD_V_HIGH				SDVO_I2C_ARG_7
+/* Part 2 */
+# define SDVO_DTD_HSYNC_OFF				SDVO_I2C_ARG_0
+# define SDVO_DTD_HSYNC_WIDTH				SDVO_I2C_ARG_1
+# define SDVO_DTD_VSYNC_OFF_WIDTH			SDVO_I2C_ARG_2
+# define SDVO_DTD_SYNC_OFF_WIDTH_HIGH			SDVO_I2C_ARG_3
+# define SDVO_DTD_DTD_FLAGS				SDVO_I2C_ARG_4
+# define SDVO_DTD_DTD_FLAG_INTERLACED				(1 << 7)
+# define SDVO_DTD_DTD_FLAG_STEREO_MASK				(3 << 5)
+# define SDVO_DTD_DTD_FLAG_INPUT_MASK				(3 << 3)
+# define SDVO_DTD_DTD_FLAG_SYNC_MASK				(3 << 1)
+# define SDVO_DTD_SDVO_FLAS				SDVO_I2C_ARG_5
+# define SDVO_DTD_SDVO_FLAG_STALL				(1 << 7)
+# define SDVO_DTD_SDVO_FLAG_CENTERED				(0 << 6)
+# define SDVO_DTD_SDVO_FLAG_UPPER_LEFT				(1 << 6)
+# define SDVO_DTD_SDVO_FLAG_SCALING_MASK			(3 << 4)
+# define SDVO_DTD_SDVO_FLAG_SCALING_NONE			(0 << 4)
+# define SDVO_DTD_SDVO_FLAG_SCALING_SHARP			(1 << 4)
+# define SDVO_DTD_SDVO_FLAG_SCALING_SMOOTH			(2 << 4)
+# define SDVO_DTD_VSYNC_OFF_HIGH			SDVO_I2C_ARG_6
+
+/*
+ * Generates a DTD based on the given width, height, and flags.
+ *
+ * This will be supported by any device supporting scaling or interlaced
+ * modes.
+ */
+#define SDVO_CMD_CREATE_PREFERRED_INPUT_TIMING		0x1a
+# define SDVO_PREFERRED_INPUT_TIMING_CLOCK_LOW		SDVO_I2C_ARG_0
+# define SDVO_PREFERRED_INPUT_TIMING_CLOCK_HIGH		SDVO_I2C_ARG_1
+# define SDVO_PREFERRED_INPUT_TIMING_WIDTH_LOW		SDVO_I2C_ARG_2
+# define SDVO_PREFERRED_INPUT_TIMING_WIDTH_HIGH		SDVO_I2C_ARG_3
+# define SDVO_PREFERRED_INPUT_TIMING_HEIGHT_LOW		SDVO_I2C_ARG_4
+# define SDVO_PREFERRED_INPUT_TIMING_HEIGHT_HIGH	SDVO_I2C_ARG_5
+# define SDVO_PREFERRED_INPUT_TIMING_FLAGS		SDVO_I2C_ARG_6
+# define SDVO_PREFERRED_INPUT_TIMING_FLAGS_INTERLACED		(1 << 0)
+# define SDVO_PREFERRED_INPUT_TIMING_FLAGS_SCALED		(1 << 1)
+
+#define SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART1	0x1b
+#define SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART2	0x1c
+
+/* Returns a struct intel_sdvo_pixel_clock_range */
+#define SDVO_CMD_GET_INPUT_PIXEL_CLOCK_RANGE		0x1d
+/* Returns a struct intel_sdvo_pixel_clock_range */
+#define SDVO_CMD_GET_OUTPUT_PIXEL_CLOCK_RANGE		0x1e
+
+/* Returns a byte bitfield containing SDVO_CLOCK_RATE_MULT_* flags */
+#define SDVO_CMD_GET_SUPPORTED_CLOCK_RATE_MULTS		0x1f
+
+/* Returns a byte containing a SDVO_CLOCK_RATE_MULT_* flag */
+#define SDVO_CMD_GET_CLOCK_RATE_MULT			0x20
+/* Takes a byte containing a SDVO_CLOCK_RATE_MULT_* flag */
+#define SDVO_CMD_SET_CLOCK_RATE_MULT			0x21
+# define SDVO_CLOCK_RATE_MULT_1X				(1 << 0)
+# define SDVO_CLOCK_RATE_MULT_2X				(1 << 1)
+# define SDVO_CLOCK_RATE_MULT_4X				(1 << 3)
+
+#define SDVO_CMD_GET_SUPPORTED_TV_FORMATS		0x27
+/* 6 bytes of bit flags for TV formats shared by all TV format functions */
+struct intel_sdvo_tv_format {
+	unsigned int ntsc_m:1;
+	unsigned int ntsc_j:1;
+	unsigned int ntsc_443:1;
+	unsigned int pal_b:1;
+	unsigned int pal_d:1;
+	unsigned int pal_g:1;
+	unsigned int pal_h:1;
+	unsigned int pal_i:1;
+
+	unsigned int pal_m:1;
+	unsigned int pal_n:1;
+	unsigned int pal_nc:1;
+	unsigned int pal_60:1;
+	unsigned int secam_b:1;
+	unsigned int secam_d:1;
+	unsigned int secam_g:1;
+	unsigned int secam_k:1;
+
+	unsigned int secam_k1:1;
+	unsigned int secam_l:1;
+	unsigned int secam_60:1;
+	unsigned int hdtv_std_smpte_240m_1080i_59:1;
+	unsigned int hdtv_std_smpte_240m_1080i_60:1;
+	unsigned int hdtv_std_smpte_260m_1080i_59:1;
+	unsigned int hdtv_std_smpte_260m_1080i_60:1;
+	unsigned int hdtv_std_smpte_274m_1080i_50:1;
+
+	unsigned int hdtv_std_smpte_274m_1080i_59:1;
+	unsigned int hdtv_std_smpte_274m_1080i_60:1;
+	unsigned int hdtv_std_smpte_274m_1080p_23:1;
+	unsigned int hdtv_std_smpte_274m_1080p_24:1;
+	unsigned int hdtv_std_smpte_274m_1080p_25:1;
+	unsigned int hdtv_std_smpte_274m_1080p_29:1;
+	unsigned int hdtv_std_smpte_274m_1080p_30:1;
+	unsigned int hdtv_std_smpte_274m_1080p_50:1;
+
+	unsigned int hdtv_std_smpte_274m_1080p_59:1;
+	unsigned int hdtv_std_smpte_274m_1080p_60:1;
+	unsigned int hdtv_std_smpte_295m_1080i_50:1;
+	unsigned int hdtv_std_smpte_295m_1080p_50:1;
+	unsigned int hdtv_std_smpte_296m_720p_59:1;
+	unsigned int hdtv_std_smpte_296m_720p_60:1;
+	unsigned int hdtv_std_smpte_296m_720p_50:1;
+	unsigned int hdtv_std_smpte_293m_480p_59:1;
+
+	unsigned int hdtv_std_smpte_170m_480i_59:1;
+	unsigned int hdtv_std_iturbt601_576i_50:1;
+	unsigned int hdtv_std_iturbt601_576p_50:1;
+	unsigned int hdtv_std_eia_7702a_480i_60:1;
+	unsigned int hdtv_std_eia_7702a_480p_60:1;
+	unsigned int pad:3;
+} __packed;
+
+#define SDVO_CMD_GET_TV_FORMAT				0x28
+
+#define SDVO_CMD_SET_TV_FORMAT				0x29
+
+/* Returns the resolutiosn that can be used with the given TV format */
+#define SDVO_CMD_GET_SDTV_RESOLUTION_SUPPORT		0x83
+struct intel_sdvo_sdtv_resolution_request {
+	unsigned int ntsc_m:1;
+	unsigned int ntsc_j:1;
+	unsigned int ntsc_443:1;
+	unsigned int pal_b:1;
+	unsigned int pal_d:1;
+	unsigned int pal_g:1;
+	unsigned int pal_h:1;
+	unsigned int pal_i:1;
+
+	unsigned int pal_m:1;
+	unsigned int pal_n:1;
+	unsigned int pal_nc:1;
+	unsigned int pal_60:1;
+	unsigned int secam_b:1;
+	unsigned int secam_d:1;
+	unsigned int secam_g:1;
+	unsigned int secam_k:1;
+
+	unsigned int secam_k1:1;
+	unsigned int secam_l:1;
+	unsigned int secam_60:1;
+	unsigned int pad:5;
+} __packed;
+
+struct intel_sdvo_sdtv_resolution_reply {
+	unsigned int res_320x200:1;
+	unsigned int res_320x240:1;
+	unsigned int res_400x300:1;
+	unsigned int res_640x350:1;
+	unsigned int res_640x400:1;
+	unsigned int res_640x480:1;
+	unsigned int res_704x480:1;
+	unsigned int res_704x576:1;
+
+	unsigned int res_720x350:1;
+	unsigned int res_720x400:1;
+	unsigned int res_720x480:1;
+	unsigned int res_720x540:1;
+	unsigned int res_720x576:1;
+	unsigned int res_768x576:1;
+	unsigned int res_800x600:1;
+	unsigned int res_832x624:1;
+
+	unsigned int res_920x766:1;
+	unsigned int res_1024x768:1;
+	unsigned int res_1280x1024:1;
+	unsigned int pad:5;
+} __packed;
+
+/* Get supported resolution with squire pixel aspect ratio that can be
+   scaled for the requested HDTV format */
+#define SDVO_CMD_GET_SCALED_HDTV_RESOLUTION_SUPPORT		0x85
+
+struct intel_sdvo_hdtv_resolution_request {
+	unsigned int hdtv_std_smpte_240m_1080i_59:1;
+	unsigned int hdtv_std_smpte_240m_1080i_60:1;
+	unsigned int hdtv_std_smpte_260m_1080i_59:1;
+	unsigned int hdtv_std_smpte_260m_1080i_60:1;
+	unsigned int hdtv_std_smpte_274m_1080i_50:1;
+	unsigned int hdtv_std_smpte_274m_1080i_59:1;
+	unsigned int hdtv_std_smpte_274m_1080i_60:1;
+	unsigned int hdtv_std_smpte_274m_1080p_23:1;
+
+	unsigned int hdtv_std_smpte_274m_1080p_24:1;
+	unsigned int hdtv_std_smpte_274m_1080p_25:1;
+	unsigned int hdtv_std_smpte_274m_1080p_29:1;
+	unsigned int hdtv_std_smpte_274m_1080p_30:1;
+	unsigned int hdtv_std_smpte_274m_1080p_50:1;
+	unsigned int hdtv_std_smpte_274m_1080p_59:1;
+	unsigned int hdtv_std_smpte_274m_1080p_60:1;
+	unsigned int hdtv_std_smpte_295m_1080i_50:1;
+
+	unsigned int hdtv_std_smpte_295m_1080p_50:1;
+	unsigned int hdtv_std_smpte_296m_720p_59:1;
+	unsigned int hdtv_std_smpte_296m_720p_60:1;
+	unsigned int hdtv_std_smpte_296m_720p_50:1;
+	unsigned int hdtv_std_smpte_293m_480p_59:1;
+	unsigned int hdtv_std_smpte_170m_480i_59:1;
+	unsigned int hdtv_std_iturbt601_576i_50:1;
+	unsigned int hdtv_std_iturbt601_576p_50:1;
+
+	unsigned int hdtv_std_eia_7702a_480i_60:1;
+	unsigned int hdtv_std_eia_7702a_480p_60:1;
+	unsigned int pad:6;
+} __packed;
+
+struct intel_sdvo_hdtv_resolution_reply {
+	unsigned int res_640x480:1;
+	unsigned int res_800x600:1;
+	unsigned int res_1024x768:1;
+	unsigned int res_1280x960:1;
+	unsigned int res_1400x1050:1;
+	unsigned int res_1600x1200:1;
+	unsigned int res_1920x1440:1;
+	unsigned int res_2048x1536:1;
+
+	unsigned int res_2560x1920:1;
+	unsigned int res_3200x2400:1;
+	unsigned int res_3840x2880:1;
+	unsigned int pad1:5;
+
+	unsigned int res_848x480:1;
+	unsigned int res_1064x600:1;
+	unsigned int res_1280x720:1;
+	unsigned int res_1360x768:1;
+	unsigned int res_1704x960:1;
+	unsigned int res_1864x1050:1;
+	unsigned int res_1920x1080:1;
+	unsigned int res_2128x1200:1;
+
+	unsigned int res_2560x1400:1;
+	unsigned int res_2728x1536:1;
+	unsigned int res_3408x1920:1;
+	unsigned int res_4264x2400:1;
+	unsigned int res_5120x2880:1;
+	unsigned int pad2:3;
+
+	unsigned int res_768x480:1;
+	unsigned int res_960x600:1;
+	unsigned int res_1152x720:1;
+	unsigned int res_1124x768:1;
+	unsigned int res_1536x960:1;
+	unsigned int res_1680x1050:1;
+	unsigned int res_1728x1080:1;
+	unsigned int res_1920x1200:1;
+
+	unsigned int res_2304x1440:1;
+	unsigned int res_2456x1536:1;
+	unsigned int res_3072x1920:1;
+	unsigned int res_3840x2400:1;
+	unsigned int res_4608x2880:1;
+	unsigned int pad3:3;
+
+	unsigned int res_1280x1024:1;
+	unsigned int pad4:7;
+
+	unsigned int res_1280x768:1;
+	unsigned int pad5:7;
+} __packed;
+
+/* Get supported power state returns info for encoder and monitor, rely on
+   last SetTargetInput and SetTargetOutput calls */
+#define SDVO_CMD_GET_SUPPORTED_POWER_STATES		0x2a
+/* Get power state returns info for encoder and monitor, rely on last
+   SetTargetInput and SetTargetOutput calls */
+#define SDVO_CMD_GET_POWER_STATE			0x2b
+#define SDVO_CMD_GET_ENCODER_POWER_STATE		0x2b
+#define SDVO_CMD_SET_ENCODER_POWER_STATE		0x2c
+# define SDVO_ENCODER_STATE_ON					(1 << 0)
+# define SDVO_ENCODER_STATE_STANDBY				(1 << 1)
+# define SDVO_ENCODER_STATE_SUSPEND				(1 << 2)
+# define SDVO_ENCODER_STATE_OFF					(1 << 3)
+# define SDVO_MONITOR_STATE_ON					(1 << 4)
+# define SDVO_MONITOR_STATE_STANDBY				(1 << 5)
+# define SDVO_MONITOR_STATE_SUSPEND				(1 << 6)
+# define SDVO_MONITOR_STATE_OFF					(1 << 7)
+
+#define SDVO_CMD_GET_MAX_PANEL_POWER_SEQUENCING		0x2d
+#define SDVO_CMD_GET_PANEL_POWER_SEQUENCING		0x2e
+#define SDVO_CMD_SET_PANEL_POWER_SEQUENCING		0x2f
+/*
+ * The panel power sequencing parameters are in units of milliseconds.
+ * The high fields are bits 8:9 of the 10-bit values.
+ */
+struct sdvo_panel_power_sequencing {
+	u8 t0;
+	u8 t1;
+	u8 t2;
+	u8 t3;
+	u8 t4;
+
+	unsigned int t0_high:2;
+	unsigned int t1_high:2;
+	unsigned int t2_high:2;
+	unsigned int t3_high:2;
+
+	unsigned int t4_high:2;
+	unsigned int pad:6;
+} __packed;
+
+#define SDVO_CMD_GET_MAX_BACKLIGHT_LEVEL		0x30
+struct sdvo_max_backlight_reply {
+	u8 max_value;
+	u8 default_value;
+} __packed;
+
+#define SDVO_CMD_GET_BACKLIGHT_LEVEL			0x31
+#define SDVO_CMD_SET_BACKLIGHT_LEVEL			0x32
+
+#define SDVO_CMD_GET_AMBIENT_LIGHT			0x33
+struct sdvo_get_ambient_light_reply {
+	u16 trip_low;
+	u16 trip_high;
+	u16 value;
+} __packed;
+#define SDVO_CMD_SET_AMBIENT_LIGHT			0x34
+struct sdvo_set_ambient_light_reply {
+	u16 trip_low;
+	u16 trip_high;
+	unsigned int enable:1;
+	unsigned int pad:7;
+} __packed;
+
+/* Set display power state */
+#define SDVO_CMD_SET_DISPLAY_POWER_STATE		0x7d
+# define SDVO_DISPLAY_STATE_ON				(1 << 0)
+# define SDVO_DISPLAY_STATE_STANDBY			(1 << 1)
+# define SDVO_DISPLAY_STATE_SUSPEND			(1 << 2)
+# define SDVO_DISPLAY_STATE_OFF				(1 << 3)
+
+#define SDVO_CMD_GET_SUPPORTED_ENHANCEMENTS		0x84
+struct intel_sdvo_enhancements_reply {
+	unsigned int flicker_filter:1;
+	unsigned int flicker_filter_adaptive:1;
+	unsigned int flicker_filter_2d:1;
+	unsigned int saturation:1;
+	unsigned int hue:1;
+	unsigned int brightness:1;
+	unsigned int contrast:1;
+	unsigned int overscan_h:1;
+
+	unsigned int overscan_v:1;
+	unsigned int hpos:1;
+	unsigned int vpos:1;
+	unsigned int sharpness:1;
+	unsigned int dot_crawl:1;
+	unsigned int dither:1;
+	unsigned int tv_chroma_filter:1;
+	unsigned int tv_luma_filter:1;
+} __packed;
+
+/* Picture enhancement limits below are dependent on the current TV format,
+ * and thus need to be queried and set after it.
+ */
+#define SDVO_CMD_GET_MAX_FLICKER_FILTER			0x4d
+#define SDVO_CMD_GET_MAX_FLICKER_FILTER_ADAPTIVE	0x7b
+#define SDVO_CMD_GET_MAX_FLICKER_FILTER_2D		0x52
+#define SDVO_CMD_GET_MAX_SATURATION			0x55
+#define SDVO_CMD_GET_MAX_HUE				0x58
+#define SDVO_CMD_GET_MAX_BRIGHTNESS			0x5b
+#define SDVO_CMD_GET_MAX_CONTRAST			0x5e
+#define SDVO_CMD_GET_MAX_OVERSCAN_H			0x61
+#define SDVO_CMD_GET_MAX_OVERSCAN_V			0x64
+#define SDVO_CMD_GET_MAX_HPOS				0x67
+#define SDVO_CMD_GET_MAX_VPOS				0x6a
+#define SDVO_CMD_GET_MAX_SHARPNESS			0x6d
+#define SDVO_CMD_GET_MAX_TV_CHROMA_FILTER		0x74
+#define SDVO_CMD_GET_MAX_TV_LUMA_FILTER			0x77
+struct intel_sdvo_enhancement_limits_reply {
+	u16 max_value;
+	u16 default_value;
+} __packed;
+
+#define SDVO_CMD_GET_LVDS_PANEL_INFORMATION		0x7f
+#define SDVO_CMD_SET_LVDS_PANEL_INFORMATION		0x80
+# define SDVO_LVDS_COLOR_DEPTH_18			(0 << 0)
+# define SDVO_LVDS_COLOR_DEPTH_24			(1 << 0)
+# define SDVO_LVDS_CONNECTOR_SPWG			(0 << 2)
+# define SDVO_LVDS_CONNECTOR_OPENLDI			(1 << 2)
+# define SDVO_LVDS_SINGLE_CHANNEL			(0 << 4)
+# define SDVO_LVDS_DUAL_CHANNEL				(1 << 4)
+
+#define SDVO_CMD_GET_FLICKER_FILTER			0x4e
+#define SDVO_CMD_SET_FLICKER_FILTER			0x4f
+#define SDVO_CMD_GET_FLICKER_FILTER_ADAPTIVE		0x50
+#define SDVO_CMD_SET_FLICKER_FILTER_ADAPTIVE		0x51
+#define SDVO_CMD_GET_FLICKER_FILTER_2D			0x53
+#define SDVO_CMD_SET_FLICKER_FILTER_2D			0x54
+#define SDVO_CMD_GET_SATURATION				0x56
+#define SDVO_CMD_SET_SATURATION				0x57
+#define SDVO_CMD_GET_HUE				0x59
+#define SDVO_CMD_SET_HUE				0x5a
+#define SDVO_CMD_GET_BRIGHTNESS				0x5c
+#define SDVO_CMD_SET_BRIGHTNESS				0x5d
+#define SDVO_CMD_GET_CONTRAST				0x5f
+#define SDVO_CMD_SET_CONTRAST				0x60
+#define SDVO_CMD_GET_OVERSCAN_H				0x62
+#define SDVO_CMD_SET_OVERSCAN_H				0x63
+#define SDVO_CMD_GET_OVERSCAN_V				0x65
+#define SDVO_CMD_SET_OVERSCAN_V				0x66
+#define SDVO_CMD_GET_HPOS				0x68
+#define SDVO_CMD_SET_HPOS				0x69
+#define SDVO_CMD_GET_VPOS				0x6b
+#define SDVO_CMD_SET_VPOS				0x6c
+#define SDVO_CMD_GET_SHARPNESS				0x6e
+#define SDVO_CMD_SET_SHARPNESS				0x6f
+#define SDVO_CMD_GET_TV_CHROMA_FILTER			0x75
+#define SDVO_CMD_SET_TV_CHROMA_FILTER			0x76
+#define SDVO_CMD_GET_TV_LUMA_FILTER			0x78
+#define SDVO_CMD_SET_TV_LUMA_FILTER			0x79
+struct intel_sdvo_enhancements_arg {
+	u16 value;
+} __packed;
+
+#define SDVO_CMD_GET_DOT_CRAWL				0x70
+#define SDVO_CMD_SET_DOT_CRAWL				0x71
+# define SDVO_DOT_CRAWL_ON					(1 << 0)
+# define SDVO_DOT_CRAWL_DEFAULT_ON				(1 << 1)
+
+#define SDVO_CMD_GET_DITHER				0x72
+#define SDVO_CMD_SET_DITHER				0x73
+# define SDVO_DITHER_ON						(1 << 0)
+# define SDVO_DITHER_DEFAULT_ON					(1 << 1)
+
+#define SDVO_CMD_SET_CONTROL_BUS_SWITCH			0x7a
+# define SDVO_CONTROL_BUS_PROM				(1 << 0)
+# define SDVO_CONTROL_BUS_DDC1				(1 << 1)
+# define SDVO_CONTROL_BUS_DDC2				(1 << 2)
+# define SDVO_CONTROL_BUS_DDC3				(1 << 3)
+
+/* HDMI op codes */
+#define SDVO_CMD_GET_SUPP_ENCODE	0x9d
+#define SDVO_CMD_GET_ENCODE		0x9e
+#define SDVO_CMD_SET_ENCODE		0x9f
+  #define SDVO_ENCODE_DVI	0x0
+  #define SDVO_ENCODE_HDMI	0x1
+#define SDVO_CMD_SET_PIXEL_REPLI	0x8b
+#define SDVO_CMD_GET_PIXEL_REPLI	0x8c
+#define SDVO_CMD_GET_COLORIMETRY_CAP	0x8d
+#define SDVO_CMD_SET_COLORIMETRY	0x8e
+  #define SDVO_COLORIMETRY_RGB256	(1 << 0)
+  #define SDVO_COLORIMETRY_RGB220	(1 << 1)
+  #define SDVO_COLORIMETRY_YCrCb422	(1 << 2)
+  #define SDVO_COLORIMETRY_YCrCb444	(1 << 3)
+#define SDVO_CMD_GET_COLORIMETRY	0x8f
+#define SDVO_CMD_GET_AUDIO_ENCRYPT_PREFER 0x90
+#define SDVO_CMD_SET_AUDIO_STAT		0x91
+#define SDVO_CMD_GET_AUDIO_STAT		0x92
+  #define SDVO_AUDIO_ELD_VALID		(1 << 0)
+  #define SDVO_AUDIO_PRESENCE_DETECT	(1 << 1)
+  #define SDVO_AUDIO_CP_READY		(1 << 2)
+#define SDVO_CMD_SET_HBUF_INDEX		0x93
+  #define SDVO_HBUF_INDEX_ELD		0
+  #define SDVO_HBUF_INDEX_AVI_IF	1
+#define SDVO_CMD_GET_HBUF_INDEX		0x94
+#define SDVO_CMD_GET_HBUF_INFO		0x95
+#define SDVO_CMD_SET_HBUF_AV_SPLIT	0x96
+#define SDVO_CMD_GET_HBUF_AV_SPLIT	0x97
+#define SDVO_CMD_SET_HBUF_DATA		0x98
+#define SDVO_CMD_GET_HBUF_DATA		0x99
+#define SDVO_CMD_SET_HBUF_TXRATE	0x9a
+#define SDVO_CMD_GET_HBUF_TXRATE	0x9b
+  #define SDVO_HBUF_TX_DISABLED	(0 << 6)
+  #define SDVO_HBUF_TX_ONCE	(2 << 6)
+  #define SDVO_HBUF_TX_VSYNC	(3 << 6)
+#define SDVO_CMD_GET_AUDIO_TX_INFO	0x9c
+#define SDVO_NEED_TO_STALL  (1 << 7)
+
+struct intel_sdvo_encode {
+	u8 dvi_rev;
+	u8 hdmi_rev;
+} __packed;
+
+#endif /* __INTEL_SDVO_REGS_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_snps_phy.c b/drivers/gpu/drm/i915/display/intel_snps_phy.c
new file mode 100644
index 000000000..3326c79c7
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_snps_phy.c
@@ -0,0 +1,2036 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/util_macros.h>
+
+#include "intel_ddi.h"
+#include "intel_ddi_buf_trans.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_snps_phy.h"
+#include "intel_snps_phy_regs.h"
+
+/**
+ * DOC: Synopsis PHY support
+ *
+ * Synopsis PHYs are primarily programmed by looking up magic register values
+ * in tables rather than calculating the necessary values at runtime.
+ *
+ * Of special note is that the SNPS PHYs include a dedicated port PLL, known as
+ * an "MPLLB."  The MPLLB replaces the shared DPLL functionality used on other
+ * platforms and must be programming directly during the modeset sequence
+ * since it is not handled by the shared DPLL framework as on other platforms.
+ */
+
+void intel_snps_phy_wait_for_calibration(struct drm_i915_private *i915)
+{
+	enum phy phy;
+
+	for_each_phy_masked(phy, ~0) {
+		if (!intel_phy_is_snps(i915, phy))
+			continue;
+
+		/*
+		 * If calibration does not complete successfully, we'll remember
+		 * which phy was affected and skip setup of the corresponding
+		 * output later.
+		 */
+		if (intel_de_wait_for_clear(i915, DG2_PHY_MISC(phy),
+					    DG2_PHY_DP_TX_ACK_MASK, 25))
+			i915->snps_phy_failed_calibration |= BIT(phy);
+	}
+}
+
+void intel_snps_phy_update_psr_power_state(struct drm_i915_private *dev_priv,
+					   enum phy phy, bool enable)
+{
+	u32 val;
+
+	if (!intel_phy_is_snps(dev_priv, phy))
+		return;
+
+	val = REG_FIELD_PREP(SNPS_PHY_TX_REQ_LN_DIS_PWR_STATE_PSR,
+			     enable ? 2 : 3);
+	intel_uncore_rmw(&dev_priv->uncore, SNPS_PHY_TX_REQ(phy),
+			 SNPS_PHY_TX_REQ_LN_DIS_PWR_STATE_PSR, val);
+}
+
+void intel_snps_phy_set_signal_levels(struct intel_encoder *encoder,
+				      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	const struct intel_ddi_buf_trans *trans;
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+	int n_entries, ln;
+
+	trans = encoder->get_buf_trans(encoder, crtc_state, &n_entries);
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !trans))
+		return;
+
+	for (ln = 0; ln < 4; ln++) {
+		int level = intel_ddi_level(encoder, crtc_state, ln);
+		u32 val = 0;
+
+		val |= REG_FIELD_PREP(SNPS_PHY_TX_EQ_MAIN, trans->entries[level].snps.vswing);
+		val |= REG_FIELD_PREP(SNPS_PHY_TX_EQ_PRE, trans->entries[level].snps.pre_cursor);
+		val |= REG_FIELD_PREP(SNPS_PHY_TX_EQ_POST, trans->entries[level].snps.post_cursor);
+
+		intel_de_write(dev_priv, SNPS_PHY_TX_EQ(ln, phy), val);
+	}
+}
+
+/*
+ * Basic DP link rates with 100 MHz reference clock.
+ */
+
+static const struct intel_mpllb_state dg2_dp_rbr_100 = {
+	.clock = 162000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 20) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 65) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 127),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 226),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 5),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 39321) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 3),
+};
+
+static const struct intel_mpllb_state dg2_dp_hbr1_100 = {
+	.clock = 270000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 20) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 65) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 127),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 184),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 1),
+};
+
+static const struct intel_mpllb_state dg2_dp_hbr2_100 = {
+	.clock = 540000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 20) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 65) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 127),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 184),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 1),
+};
+
+static const struct intel_mpllb_state dg2_dp_hbr3_100 = {
+	.clock = 810000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 19) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 65) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 127),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 292),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 1),
+};
+
+static const struct intel_mpllb_state dg2_dp_uhbr10_100 = {
+	.clock = 1000000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 21) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 65) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 127),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV_MULTIPLIER, 8) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_WORD_DIV2_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DP2_MODE, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SHIM_DIV32_CLK_SEL, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 368),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 1),
+
+	/*
+	 * SSC will be enabled, DP UHBR has a minimum SSC requirement.
+	 */
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_PEAK, 58982),
+	.mpllb_sscstep =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_STEPSIZE, 76101),
+};
+
+static const struct intel_mpllb_state dg2_dp_uhbr13_100 = {
+	.clock = 1350000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 5) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 45) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 65) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 127),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV_MULTIPLIER, 8) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_WORD_DIV2_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DP2_MODE, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 508),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 1),
+
+	/*
+	 * SSC will be enabled, DP UHBR has a minimum SSC requirement.
+	 */
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_PEAK, 79626),
+	.mpllb_sscstep =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_STEPSIZE, 102737),
+};
+
+static const struct intel_mpllb_state * const dg2_dp_100_tables[] = {
+	&dg2_dp_rbr_100,
+	&dg2_dp_hbr1_100,
+	&dg2_dp_hbr2_100,
+	&dg2_dp_hbr3_100,
+	&dg2_dp_uhbr10_100,
+	&dg2_dp_uhbr13_100,
+	NULL,
+};
+
+/*
+ * eDP link rates with 100 MHz reference clock.
+ */
+
+static const struct intel_mpllb_state dg2_edp_r216 = {
+	.clock = 216000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 19) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 65) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 127),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 312),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 5),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 52428) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 4),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_PEAK, 50961),
+	.mpllb_sscstep =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_STEPSIZE, 65752),
+};
+
+static const struct intel_mpllb_state dg2_edp_r243 = {
+	.clock = 243000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 20) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 65) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 127),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 356),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 5),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 2),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_PEAK, 57331),
+	.mpllb_sscstep =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_STEPSIZE, 73971),
+};
+
+static const struct intel_mpllb_state dg2_edp_r324 = {
+	.clock = 324000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 20) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 65) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 127),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 226),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 5),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 39321) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 3),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_PEAK, 38221),
+	.mpllb_sscstep =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_STEPSIZE, 49314),
+};
+
+static const struct intel_mpllb_state dg2_edp_r432 = {
+	.clock = 432000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 19) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 65) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 127),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 312),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 5),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 52428) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 4),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_PEAK, 50961),
+	.mpllb_sscstep =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_STEPSIZE, 65752),
+};
+
+static const struct intel_mpllb_state * const dg2_edp_tables[] = {
+	&dg2_dp_rbr_100,
+	&dg2_edp_r216,
+	&dg2_edp_r243,
+	&dg2_dp_hbr1_100,
+	&dg2_edp_r324,
+	&dg2_edp_r432,
+	&dg2_dp_hbr2_100,
+	&dg2_dp_hbr3_100,
+	NULL,
+};
+
+/*
+ * HDMI link rates with 100 MHz reference clock.
+ */
+
+static const struct intel_mpllb_state dg2_hdmi_25_175 = {
+	.clock = 25175,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 5) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 15) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 5) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 128) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 143),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 36663) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 71),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_27_0 = {
+	.clock = 27000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 5) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 15) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 5) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 140) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 5),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 2),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_74_25 = {
+	.clock = 74250,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 15) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 86) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 5),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 2),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_148_5 = {
+	.clock = 148500,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 15) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 86) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 5),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 2),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+/* values in the below table are calculted using the algo */
+static const struct intel_mpllb_state dg2_hdmi_25200 = {
+	.clock = 25200,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 7) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 5) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 128) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 41943) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 2621),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_27027 = {
+	.clock = 27027,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 5) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 140) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 31876) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 46555),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_28320 = {
+	.clock = 28320,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 5) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 148) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 40894) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 30408),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_30240 = {
+	.clock = 30240,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 5) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 160) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 50331) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 42466),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_31500 = {
+	.clock = 31500,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 7) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 68) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 26214),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_36000 = {
+	.clock = 36000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 82) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 39321) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 39320),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_40000 = {
+	.clock = 40000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 96) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 0),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_49500 = {
+	.clock = 49500,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 1),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 126) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 13107) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 13107),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_50000 = {
+	.clock = 50000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 1),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 128) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 0),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_57284 = {
+	.clock = 57284,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 150) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 42886) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 49701),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_58000 = {
+	.clock = 58000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 152) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 52428) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 52427),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_65000 = {
+	.clock = 65000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 7) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 72) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 0),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_71000 = {
+	.clock = 71000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 80) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 52428) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 52427),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_74176 = {
+	.clock = 74176,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 86) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 22334) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 43829),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_75000 = {
+	.clock = 75000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 88) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 0),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_78750 = {
+	.clock = 78750,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 94) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 0),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_85500 = {
+	.clock = 85500,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 104) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 26214),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_88750 = {
+	.clock = 88750,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 7) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 15) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 1),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 110) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 0),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_106500 = {
+	.clock = 106500,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 138) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 13107) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 13107),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_108000 = {
+	.clock = 108000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 140) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 26214),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_115500 = {
+	.clock = 115500,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 152) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 26214),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_119000 = {
+	.clock = 119000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 3) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 158) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 13107) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 13107),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_135000 = {
+	.clock = 135000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 7) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 15) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 76) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 0),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_138500 = {
+	.clock = 138500,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 78) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 26214),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_147160 = {
+	.clock = 147160,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 84) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 56623) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 6815),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_148352 = {
+	.clock = 148352,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 86) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 22334) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 43829),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_154000 = {
+	.clock = 154000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 13) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 90) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 39321) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 39320),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_162000 = {
+	.clock = 162000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 96) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 52428) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 52427),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_209800 = {
+	.clock = 209800,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 7) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 134) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 60293) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 7864),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_262750 = {
+	.clock = 262750,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 7) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 72) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 36044) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 52427),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_267300 = {
+	.clock = 267300,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 7) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 74) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 30146) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 36699),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_268500 = {
+	.clock = 268500,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 7) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 74) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 45875) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 13107),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_296703 = {
+	.clock = 296703,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 86) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 22321) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 36804),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_241500 = {
+	.clock = 241500,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 160) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 39321) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 39320),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_319890 = {
+	.clock = 319890,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 2),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 94) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 64094) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 13631),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_497750 = {
+	.clock = 497750,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 15) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 0),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 166) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 36044) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 52427),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_592000 = {
+	.clock = 592000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 86) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 13107) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 13107),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_593407 = {
+	.clock = 593407,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 0) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 86) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 22328) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 7549),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_297 = {
+	.clock = 297000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 6) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 14) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_TX_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 86) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 65535),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 26214),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state dg2_hdmi_594 = {
+	.clock = 594000,
+	.ref_control =
+		REG_FIELD_PREP(SNPS_PHY_REF_CONTROL_REF_RANGE, 3),
+	.mpllb_cp =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT, 4) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP, 15) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_INT_GS, 64) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_CP_PROP_GS, 124),
+	.mpllb_div =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_DIV5_CLK_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_PMIX_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_V2I, 2) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FREQ_VCO, 3),
+	.mpllb_div2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_REF_CLK_DIV, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_MULTIPLIER, 86) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_HDMI_DIV, 1),
+	.mpllb_fracn1 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_EN, 1) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_DEN, 5),
+	.mpllb_fracn2 =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_QUOT, 26214) |
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_FRACN_REM, 2),
+	.mpllb_sscen =
+		REG_FIELD_PREP(SNPS_PHY_MPLLB_SSC_UP_SPREAD, 1),
+};
+
+static const struct intel_mpllb_state * const dg2_hdmi_tables[] = {
+	&dg2_hdmi_25_175,
+	&dg2_hdmi_27_0,
+	&dg2_hdmi_74_25,
+	&dg2_hdmi_148_5,
+	&dg2_hdmi_297,
+	&dg2_hdmi_594,
+	&dg2_hdmi_25200,
+	&dg2_hdmi_27027,
+	&dg2_hdmi_28320,
+	&dg2_hdmi_30240,
+	&dg2_hdmi_31500,
+	&dg2_hdmi_36000,
+	&dg2_hdmi_40000,
+	&dg2_hdmi_49500,
+	&dg2_hdmi_50000,
+	&dg2_hdmi_57284,
+	&dg2_hdmi_58000,
+	&dg2_hdmi_65000,
+	&dg2_hdmi_71000,
+	&dg2_hdmi_74176,
+	&dg2_hdmi_75000,
+	&dg2_hdmi_78750,
+	&dg2_hdmi_85500,
+	&dg2_hdmi_88750,
+	&dg2_hdmi_106500,
+	&dg2_hdmi_108000,
+	&dg2_hdmi_115500,
+	&dg2_hdmi_119000,
+	&dg2_hdmi_135000,
+	&dg2_hdmi_138500,
+	&dg2_hdmi_147160,
+	&dg2_hdmi_148352,
+	&dg2_hdmi_154000,
+	&dg2_hdmi_162000,
+	&dg2_hdmi_209800,
+	&dg2_hdmi_241500,
+	&dg2_hdmi_262750,
+	&dg2_hdmi_267300,
+	&dg2_hdmi_268500,
+	&dg2_hdmi_296703,
+	&dg2_hdmi_319890,
+	&dg2_hdmi_497750,
+	&dg2_hdmi_592000,
+	&dg2_hdmi_593407,
+	NULL,
+};
+
+static const struct intel_mpllb_state * const *
+intel_mpllb_tables_get(struct intel_crtc_state *crtc_state,
+		       struct intel_encoder *encoder)
+{
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP)) {
+		return dg2_edp_tables;
+	} else if (intel_crtc_has_dp_encoder(crtc_state)) {
+		return dg2_dp_100_tables;
+	} else if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI)) {
+		return dg2_hdmi_tables;
+	}
+
+	MISSING_CASE(encoder->type);
+	return NULL;
+}
+
+int intel_mpllb_calc_state(struct intel_crtc_state *crtc_state,
+			   struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	const struct intel_mpllb_state * const *tables;
+	int i;
+
+	if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI)) {
+		if (intel_snps_phy_check_hdmi_link_rate(crtc_state->port_clock)
+		    != MODE_OK) {
+			/*
+			 * FIXME: Can only support fixed HDMI frequencies
+			 * until we have a proper algorithm under a valid
+			 * license.
+			 */
+			drm_dbg_kms(&i915->drm, "Can't support HDMI link rate %d\n",
+				    crtc_state->port_clock);
+			return -EINVAL;
+		}
+	}
+
+	tables = intel_mpllb_tables_get(crtc_state, encoder);
+	if (!tables)
+		return -EINVAL;
+
+	for (i = 0; tables[i]; i++) {
+		if (crtc_state->port_clock == tables[i]->clock) {
+			crtc_state->mpllb_state = *tables[i];
+			return 0;
+		}
+	}
+
+	return -EINVAL;
+}
+
+void intel_mpllb_enable(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	const struct intel_mpllb_state *pll_state = &crtc_state->mpllb_state;
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+	i915_reg_t enable_reg = (phy <= PHY_D ?
+				 DG2_PLL_ENABLE(phy) : MG_PLL_ENABLE(0));
+
+	/*
+	 * 3. Software programs the following PLL registers for the desired
+	 * frequency.
+	 */
+	intel_de_write(dev_priv, SNPS_PHY_MPLLB_CP(phy), pll_state->mpllb_cp);
+	intel_de_write(dev_priv, SNPS_PHY_MPLLB_DIV(phy), pll_state->mpllb_div);
+	intel_de_write(dev_priv, SNPS_PHY_MPLLB_DIV2(phy), pll_state->mpllb_div2);
+	intel_de_write(dev_priv, SNPS_PHY_MPLLB_SSCEN(phy), pll_state->mpllb_sscen);
+	intel_de_write(dev_priv, SNPS_PHY_MPLLB_SSCSTEP(phy), pll_state->mpllb_sscstep);
+	intel_de_write(dev_priv, SNPS_PHY_MPLLB_FRACN1(phy), pll_state->mpllb_fracn1);
+	intel_de_write(dev_priv, SNPS_PHY_MPLLB_FRACN2(phy), pll_state->mpllb_fracn2);
+
+	/*
+	 * 4. If the frequency will result in a change to the voltage
+	 * requirement, follow the Display Voltage Frequency Switching -
+	 * Sequence Before Frequency Change.
+	 *
+	 * We handle this step in bxt_set_cdclk().
+	 */
+
+	/* 5. Software sets DPLL_ENABLE [PLL Enable] to "1". */
+	intel_uncore_rmw(&dev_priv->uncore, enable_reg, 0, PLL_ENABLE);
+
+	/*
+	 * 9. Software sets SNPS_PHY_MPLLB_DIV dp_mpllb_force_en to "1". This
+	 * will keep the PLL running during the DDI lane programming and any
+	 * typeC DP cable disconnect. Do not set the force before enabling the
+	 * PLL because that will start the PLL before it has sampled the
+	 * divider values.
+	 */
+	intel_de_write(dev_priv, SNPS_PHY_MPLLB_DIV(phy),
+		       pll_state->mpllb_div | SNPS_PHY_MPLLB_FORCE_EN);
+
+	/*
+	 * 10. Software polls on register DPLL_ENABLE [PLL Lock] to confirm PLL
+	 * is locked at new settings. This register bit is sampling PHY
+	 * dp_mpllb_state interface signal.
+	 */
+	if (intel_de_wait_for_set(dev_priv, enable_reg, PLL_LOCK, 5))
+		drm_dbg_kms(&dev_priv->drm, "Port %c PLL not locked\n", phy_name(phy));
+
+	/*
+	 * 11. If the frequency will result in a change to the voltage
+	 * requirement, follow the Display Voltage Frequency Switching -
+	 * Sequence After Frequency Change.
+	 *
+	 * We handle this step in bxt_set_cdclk().
+	 */
+}
+
+void intel_mpllb_disable(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+	i915_reg_t enable_reg = (phy <= PHY_D ?
+				 DG2_PLL_ENABLE(phy) : MG_PLL_ENABLE(0));
+
+	/*
+	 * 1. If the frequency will result in a change to the voltage
+	 * requirement, follow the Display Voltage Frequency Switching -
+	 * Sequence Before Frequency Change.
+	 *
+	 * We handle this step in bxt_set_cdclk().
+	 */
+
+	/* 2. Software programs DPLL_ENABLE [PLL Enable] to "0" */
+	intel_uncore_rmw(&i915->uncore, enable_reg, PLL_ENABLE, 0);
+
+	/*
+	 * 4. Software programs SNPS_PHY_MPLLB_DIV dp_mpllb_force_en to "0".
+	 * This will allow the PLL to stop running.
+	 */
+	intel_uncore_rmw(&i915->uncore, SNPS_PHY_MPLLB_DIV(phy),
+			 SNPS_PHY_MPLLB_FORCE_EN, 0);
+
+	/*
+	 * 5. Software polls DPLL_ENABLE [PLL Lock] for PHY acknowledgment
+	 * (dp_txX_ack) that the new transmitter setting request is completed.
+	 */
+	if (intel_de_wait_for_clear(i915, enable_reg, PLL_LOCK, 5))
+		drm_err(&i915->drm, "Port %c PLL not locked\n", phy_name(phy));
+
+	/*
+	 * 6. If the frequency will result in a change to the voltage
+	 * requirement, follow the Display Voltage Frequency Switching -
+	 * Sequence After Frequency Change.
+	 *
+	 * We handle this step in bxt_set_cdclk().
+	 */
+}
+
+int intel_mpllb_calc_port_clock(struct intel_encoder *encoder,
+				const struct intel_mpllb_state *pll_state)
+{
+	unsigned int frac_quot = 0, frac_rem = 0, frac_den = 1;
+	unsigned int multiplier, tx_clk_div, refclk;
+	bool frac_en;
+
+	if (0)
+		refclk = 38400;
+	else
+		refclk = 100000;
+
+	refclk >>= REG_FIELD_GET(SNPS_PHY_MPLLB_REF_CLK_DIV, pll_state->mpllb_div2) - 1;
+
+	frac_en = REG_FIELD_GET(SNPS_PHY_MPLLB_FRACN_EN, pll_state->mpllb_fracn1);
+
+	if (frac_en) {
+		frac_quot = REG_FIELD_GET(SNPS_PHY_MPLLB_FRACN_QUOT, pll_state->mpllb_fracn2);
+		frac_rem = REG_FIELD_GET(SNPS_PHY_MPLLB_FRACN_REM, pll_state->mpllb_fracn2);
+		frac_den = REG_FIELD_GET(SNPS_PHY_MPLLB_FRACN_DEN, pll_state->mpllb_fracn1);
+	}
+
+	multiplier = REG_FIELD_GET(SNPS_PHY_MPLLB_MULTIPLIER, pll_state->mpllb_div2) / 2 + 16;
+
+	tx_clk_div = REG_FIELD_GET(SNPS_PHY_MPLLB_TX_CLK_DIV, pll_state->mpllb_div);
+
+	return DIV_ROUND_CLOSEST_ULL(mul_u32_u32(refclk, (multiplier << 16) + frac_quot) +
+				     DIV_ROUND_CLOSEST(refclk * frac_rem, frac_den),
+				     10 << (tx_clk_div + 16));
+}
+
+void intel_mpllb_readout_hw_state(struct intel_encoder *encoder,
+				  struct intel_mpllb_state *pll_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum phy phy = intel_port_to_phy(dev_priv, encoder->port);
+
+	pll_state->mpllb_cp = intel_de_read(dev_priv, SNPS_PHY_MPLLB_CP(phy));
+	pll_state->mpllb_div = intel_de_read(dev_priv, SNPS_PHY_MPLLB_DIV(phy));
+	pll_state->mpllb_div2 = intel_de_read(dev_priv, SNPS_PHY_MPLLB_DIV2(phy));
+	pll_state->mpllb_sscen = intel_de_read(dev_priv, SNPS_PHY_MPLLB_SSCEN(phy));
+	pll_state->mpllb_sscstep = intel_de_read(dev_priv, SNPS_PHY_MPLLB_SSCSTEP(phy));
+	pll_state->mpllb_fracn1 = intel_de_read(dev_priv, SNPS_PHY_MPLLB_FRACN1(phy));
+	pll_state->mpllb_fracn2 = intel_de_read(dev_priv, SNPS_PHY_MPLLB_FRACN2(phy));
+
+	/*
+	 * REF_CONTROL is under firmware control and never programmed by the
+	 * driver; we read it only for sanity checking purposes.  The bspec
+	 * only tells us the expected value for one field in this register,
+	 * so we'll only read out those specific bits here.
+	 */
+	pll_state->ref_control = intel_de_read(dev_priv, SNPS_PHY_REF_CONTROL(phy)) &
+		SNPS_PHY_REF_CONTROL_REF_RANGE;
+
+	/*
+	 * MPLLB_DIV is programmed twice, once with the software-computed
+	 * state, then again with the MPLLB_FORCE_EN bit added.  Drop that
+	 * extra bit during readout so that we return the actual expected
+	 * software state.
+	 */
+	pll_state->mpllb_div &= ~SNPS_PHY_MPLLB_FORCE_EN;
+}
+
+int intel_snps_phy_check_hdmi_link_rate(int clock)
+{
+	const struct intel_mpllb_state * const *tables = dg2_hdmi_tables;
+	int i;
+
+	for (i = 0; tables[i]; i++) {
+		if (clock == tables[i]->clock)
+			return MODE_OK;
+	}
+
+	return MODE_CLOCK_RANGE;
+}
+
+void intel_mpllb_state_verify(struct intel_atomic_state *state,
+			      struct intel_crtc_state *new_crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_mpllb_state mpllb_hw_state = { 0 };
+	struct intel_mpllb_state *mpllb_sw_state = &new_crtc_state->mpllb_state;
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct intel_encoder *encoder;
+
+	if (!IS_DG2(i915))
+		return;
+
+	if (!new_crtc_state->hw.active)
+		return;
+
+	encoder = intel_get_crtc_new_encoder(state, new_crtc_state);
+	intel_mpllb_readout_hw_state(encoder, &mpllb_hw_state);
+
+#define MPLLB_CHECK(__name)						\
+	I915_STATE_WARN(mpllb_sw_state->__name != mpllb_hw_state.__name,	\
+			"[CRTC:%d:%s] mismatch in MPLLB: %s (expected 0x%08x, found 0x%08x)", \
+			crtc->base.base.id, crtc->base.name,		\
+			__stringify(__name),				\
+			mpllb_sw_state->__name, mpllb_hw_state.__name)
+
+	MPLLB_CHECK(mpllb_cp);
+	MPLLB_CHECK(mpllb_div);
+	MPLLB_CHECK(mpllb_div2);
+	MPLLB_CHECK(mpllb_fracn1);
+	MPLLB_CHECK(mpllb_fracn2);
+	MPLLB_CHECK(mpllb_sscen);
+	MPLLB_CHECK(mpllb_sscstep);
+
+	/*
+	 * ref_control is handled by the hardware/firemware and never
+	 * programmed by the software, but the proper values are supplied
+	 * in the bspec for verification purposes.
+	 */
+	MPLLB_CHECK(ref_control);
+
+#undef MPLLB_CHECK
+}
diff --git a/drivers/gpu/drm/i915/display/intel_snps_phy.h b/drivers/gpu/drm/i915/display/intel_snps_phy.h
new file mode 100644
index 000000000..557ef820b
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_snps_phy.h
@@ -0,0 +1,38 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_SNPS_PHY_H__
+#define __INTEL_SNPS_PHY_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_crtc_state;
+struct intel_encoder;
+struct intel_mpllb_state;
+enum phy;
+
+void intel_snps_phy_wait_for_calibration(struct drm_i915_private *dev_priv);
+void intel_snps_phy_update_psr_power_state(struct drm_i915_private *dev_priv,
+					   enum phy phy, bool enable);
+
+int intel_mpllb_calc_state(struct intel_crtc_state *crtc_state,
+			   struct intel_encoder *encoder);
+void intel_mpllb_enable(struct intel_encoder *encoder,
+			const struct intel_crtc_state *crtc_state);
+void intel_mpllb_disable(struct intel_encoder *encoder);
+void intel_mpllb_readout_hw_state(struct intel_encoder *encoder,
+				  struct intel_mpllb_state *pll_state);
+int intel_mpllb_calc_port_clock(struct intel_encoder *encoder,
+				const struct intel_mpllb_state *pll_state);
+
+int intel_snps_phy_check_hdmi_link_rate(int clock);
+void intel_snps_phy_set_signal_levels(struct intel_encoder *encoder,
+				      const struct intel_crtc_state *crtc_state);
+void intel_mpllb_state_verify(struct intel_atomic_state *state,
+			      struct intel_crtc_state *new_crtc_state);
+
+#endif /* __INTEL_SNPS_PHY_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_snps_phy_regs.h b/drivers/gpu/drm/i915/display/intel_snps_phy_regs.h
new file mode 100644
index 000000000..0543465aa
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_snps_phy_regs.h
@@ -0,0 +1,75 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_SNPS_PHY_REGS__
+#define __INTEL_SNPS_PHY_REGS__
+
+#include "i915_reg_defs.h"
+
+#define _SNPS_PHY_A_BASE			0x168000
+#define _SNPS_PHY_B_BASE			0x169000
+#define _SNPS_PHY(phy)				_PHY(phy, \
+						     _SNPS_PHY_A_BASE, \
+						     _SNPS_PHY_B_BASE)
+#define _SNPS2(phy, reg)			(_SNPS_PHY(phy) - \
+						 _SNPS_PHY_A_BASE + (reg))
+#define _MMIO_SNPS(phy, reg)			_MMIO(_SNPS2(phy, reg))
+#define _MMIO_SNPS_LN(ln, phy, reg)		_MMIO(_SNPS2(phy, \
+							     (reg) + (ln) * 0x10))
+
+#define SNPS_PHY_MPLLB_CP(phy)			_MMIO_SNPS(phy, 0x168000)
+#define   SNPS_PHY_MPLLB_CP_INT			REG_GENMASK(31, 25)
+#define   SNPS_PHY_MPLLB_CP_INT_GS		REG_GENMASK(23, 17)
+#define   SNPS_PHY_MPLLB_CP_PROP		REG_GENMASK(15, 9)
+#define   SNPS_PHY_MPLLB_CP_PROP_GS		REG_GENMASK(7, 1)
+
+#define SNPS_PHY_MPLLB_DIV(phy)			_MMIO_SNPS(phy, 0x168004)
+#define   SNPS_PHY_MPLLB_FORCE_EN		REG_BIT(31)
+#define   SNPS_PHY_MPLLB_DIV_CLK_EN		REG_BIT(30)
+#define   SNPS_PHY_MPLLB_DIV5_CLK_EN		REG_BIT(29)
+#define   SNPS_PHY_MPLLB_V2I			REG_GENMASK(27, 26)
+#define   SNPS_PHY_MPLLB_FREQ_VCO		REG_GENMASK(25, 24)
+#define   SNPS_PHY_MPLLB_DIV_MULTIPLIER		REG_GENMASK(23, 16)
+#define   SNPS_PHY_MPLLB_PMIX_EN		REG_BIT(10)
+#define   SNPS_PHY_MPLLB_DP2_MODE		REG_BIT(9)
+#define   SNPS_PHY_MPLLB_WORD_DIV2_EN		REG_BIT(8)
+#define   SNPS_PHY_MPLLB_TX_CLK_DIV		REG_GENMASK(7, 5)
+#define   SNPS_PHY_MPLLB_SHIM_DIV32_CLK_SEL	REG_BIT(0)
+
+#define SNPS_PHY_MPLLB_FRACN1(phy)		_MMIO_SNPS(phy, 0x168008)
+#define   SNPS_PHY_MPLLB_FRACN_EN		REG_BIT(31)
+#define   SNPS_PHY_MPLLB_FRACN_CGG_UPDATE_EN	REG_BIT(30)
+#define   SNPS_PHY_MPLLB_FRACN_DEN		REG_GENMASK(15, 0)
+
+#define SNPS_PHY_MPLLB_FRACN2(phy)		_MMIO_SNPS(phy, 0x16800C)
+#define   SNPS_PHY_MPLLB_FRACN_REM		REG_GENMASK(31, 16)
+#define   SNPS_PHY_MPLLB_FRACN_QUOT		REG_GENMASK(15, 0)
+
+#define SNPS_PHY_MPLLB_SSCEN(phy)		_MMIO_SNPS(phy, 0x168014)
+#define   SNPS_PHY_MPLLB_SSC_EN			REG_BIT(31)
+#define   SNPS_PHY_MPLLB_SSC_UP_SPREAD		REG_BIT(30)
+#define   SNPS_PHY_MPLLB_SSC_PEAK		REG_GENMASK(29, 10)
+
+#define SNPS_PHY_MPLLB_SSCSTEP(phy)		_MMIO_SNPS(phy, 0x168018)
+#define   SNPS_PHY_MPLLB_SSC_STEPSIZE		REG_GENMASK(31, 11)
+
+#define SNPS_PHY_MPLLB_DIV2(phy)		_MMIO_SNPS(phy, 0x16801C)
+#define   SNPS_PHY_MPLLB_HDMI_PIXEL_CLK_DIV	REG_GENMASK(19, 18)
+#define   SNPS_PHY_MPLLB_HDMI_DIV		REG_GENMASK(17, 15)
+#define   SNPS_PHY_MPLLB_REF_CLK_DIV		REG_GENMASK(14, 12)
+#define   SNPS_PHY_MPLLB_MULTIPLIER		REG_GENMASK(11, 0)
+
+#define SNPS_PHY_REF_CONTROL(phy)		_MMIO_SNPS(phy, 0x168188)
+#define   SNPS_PHY_REF_CONTROL_REF_RANGE	REG_GENMASK(31, 27)
+
+#define SNPS_PHY_TX_REQ(phy)			_MMIO_SNPS(phy, 0x168200)
+#define   SNPS_PHY_TX_REQ_LN_DIS_PWR_STATE_PSR	REG_GENMASK(31, 30)
+
+#define SNPS_PHY_TX_EQ(ln, phy)			_MMIO_SNPS_LN(ln, phy, 0x168300)
+#define   SNPS_PHY_TX_EQ_MAIN			REG_GENMASK(23, 18)
+#define   SNPS_PHY_TX_EQ_POST			REG_GENMASK(15, 10)
+#define   SNPS_PHY_TX_EQ_PRE			REG_GENMASK(7, 2)
+
+#endif /* __INTEL_SNPS_PHY_REGS__ */
diff --git a/drivers/gpu/drm/i915/display/intel_sprite.c b/drivers/gpu/drm/i915/display/intel_sprite.c
new file mode 100644
index 000000000..7649c50b5
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_sprite.c
@@ -0,0 +1,1848 @@
+/*
+ * Copyright © 2011 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *   Jesse Barnes <jbarnes@virtuousgeek.org>
+ *
+ * New plane/sprite handling.
+ *
+ * The older chips had a separate interface for programming plane related
+ * registers; newer ones are much simpler and we can use the new DRM plane
+ * support.
+ */
+
+#include <linux/string_helpers.h>
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_blend.h>
+#include <drm/drm_color_mgmt.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_damage_helper.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_rect.h>
+
+#include "i915_drv.h"
+#include "i915_vgpu.h"
+#include "i9xx_plane.h"
+#include "intel_atomic_plane.h"
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_fb.h"
+#include "intel_frontbuffer.h"
+#include "intel_sprite.h"
+#include "intel_vrr.h"
+
+int intel_plane_check_src_coordinates(struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	struct drm_rect *src = &plane_state->uapi.src;
+	u32 src_x, src_y, src_w, src_h, hsub, vsub;
+	bool rotated = drm_rotation_90_or_270(plane_state->hw.rotation);
+
+	/*
+	 * FIXME hsub/vsub vs. block size is a mess. Pre-tgl CCS
+	 * abuses hsub/vsub so we can't use them here. But as they
+	 * are limited to 32bpp RGB formats we don't actually need
+	 * to check anything.
+	 */
+	if (fb->modifier == I915_FORMAT_MOD_Y_TILED_CCS ||
+	    fb->modifier == I915_FORMAT_MOD_Yf_TILED_CCS)
+		return 0;
+
+	/*
+	 * Hardware doesn't handle subpixel coordinates.
+	 * Adjust to (macro)pixel boundary, but be careful not to
+	 * increase the source viewport size, because that could
+	 * push the downscaling factor out of bounds.
+	 */
+	src_x = src->x1 >> 16;
+	src_w = drm_rect_width(src) >> 16;
+	src_y = src->y1 >> 16;
+	src_h = drm_rect_height(src) >> 16;
+
+	drm_rect_init(src, src_x << 16, src_y << 16,
+		      src_w << 16, src_h << 16);
+
+	if (fb->format->format == DRM_FORMAT_RGB565 && rotated) {
+		hsub = 2;
+		vsub = 2;
+	} else {
+		hsub = fb->format->hsub;
+		vsub = fb->format->vsub;
+	}
+
+	if (rotated)
+		hsub = vsub = max(hsub, vsub);
+
+	if (src_x % hsub || src_w % hsub) {
+		drm_dbg_kms(&i915->drm, "src x/w (%u, %u) must be a multiple of %u (rotated: %s)\n",
+			    src_x, src_w, hsub, str_yes_no(rotated));
+		return -EINVAL;
+	}
+
+	if (src_y % vsub || src_h % vsub) {
+		drm_dbg_kms(&i915->drm, "src y/h (%u, %u) must be a multiple of %u (rotated: %s)\n",
+			    src_y, src_h, vsub, str_yes_no(rotated));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void i9xx_plane_linear_gamma(u16 gamma[8])
+{
+	/* The points are not evenly spaced. */
+	static const u8 in[8] = { 0, 1, 2, 4, 8, 16, 24, 32 };
+	int i;
+
+	for (i = 0; i < 8; i++)
+		gamma[i] = (in[i] << 8) / 32;
+}
+
+static void
+chv_sprite_update_csc(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	enum plane_id plane_id = plane->id;
+	/*
+	 * |r|   | c0 c1 c2 |   |cr|
+	 * |g| = | c3 c4 c5 | x |y |
+	 * |b|   | c6 c7 c8 |   |cb|
+	 *
+	 * Coefficients are s3.12.
+	 *
+	 * Cb and Cr apparently come in as signed already, and
+	 * we always get full range data in on account of CLRC0/1.
+	 */
+	static const s16 csc_matrix[][9] = {
+		/* BT.601 full range YCbCr -> full range RGB */
+		[DRM_COLOR_YCBCR_BT601] = {
+			 5743, 4096,     0,
+			-2925, 4096, -1410,
+			    0, 4096,  7258,
+		},
+		/* BT.709 full range YCbCr -> full range RGB */
+		[DRM_COLOR_YCBCR_BT709] = {
+			 6450, 4096,     0,
+			-1917, 4096,  -767,
+			    0, 4096,  7601,
+		},
+	};
+	const s16 *csc = csc_matrix[plane_state->hw.color_encoding];
+
+	/* Seems RGB data bypasses the CSC always */
+	if (!fb->format->is_yuv)
+		return;
+
+	intel_de_write_fw(dev_priv, SPCSCYGOFF(plane_id),
+			  SPCSC_OOFF(0) | SPCSC_IOFF(0));
+	intel_de_write_fw(dev_priv, SPCSCCBOFF(plane_id),
+			  SPCSC_OOFF(0) | SPCSC_IOFF(0));
+	intel_de_write_fw(dev_priv, SPCSCCROFF(plane_id),
+			  SPCSC_OOFF(0) | SPCSC_IOFF(0));
+
+	intel_de_write_fw(dev_priv, SPCSCC01(plane_id),
+			  SPCSC_C1(csc[1]) | SPCSC_C0(csc[0]));
+	intel_de_write_fw(dev_priv, SPCSCC23(plane_id),
+			  SPCSC_C1(csc[3]) | SPCSC_C0(csc[2]));
+	intel_de_write_fw(dev_priv, SPCSCC45(plane_id),
+			  SPCSC_C1(csc[5]) | SPCSC_C0(csc[4]));
+	intel_de_write_fw(dev_priv, SPCSCC67(plane_id),
+			  SPCSC_C1(csc[7]) | SPCSC_C0(csc[6]));
+	intel_de_write_fw(dev_priv, SPCSCC8(plane_id), SPCSC_C0(csc[8]));
+
+	intel_de_write_fw(dev_priv, SPCSCYGICLAMP(plane_id),
+			  SPCSC_IMAX(1023) | SPCSC_IMIN(0));
+	intel_de_write_fw(dev_priv, SPCSCCBICLAMP(plane_id),
+			  SPCSC_IMAX(512) | SPCSC_IMIN(-512));
+	intel_de_write_fw(dev_priv, SPCSCCRICLAMP(plane_id),
+			  SPCSC_IMAX(512) | SPCSC_IMIN(-512));
+
+	intel_de_write_fw(dev_priv, SPCSCYGOCLAMP(plane_id),
+			  SPCSC_OMAX(1023) | SPCSC_OMIN(0));
+	intel_de_write_fw(dev_priv, SPCSCCBOCLAMP(plane_id),
+			  SPCSC_OMAX(1023) | SPCSC_OMIN(0));
+	intel_de_write_fw(dev_priv, SPCSCCROCLAMP(plane_id),
+			  SPCSC_OMAX(1023) | SPCSC_OMIN(0));
+}
+
+#define SIN_0 0
+#define COS_0 1
+
+static void
+vlv_sprite_update_clrc(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	enum pipe pipe = plane->pipe;
+	enum plane_id plane_id = plane->id;
+	int contrast, brightness, sh_scale, sh_sin, sh_cos;
+
+	if (fb->format->is_yuv &&
+	    plane_state->hw.color_range == DRM_COLOR_YCBCR_LIMITED_RANGE) {
+		/*
+		 * Expand limited range to full range:
+		 * Contrast is applied first and is used to expand Y range.
+		 * Brightness is applied second and is used to remove the
+		 * offset from Y. Saturation/hue is used to expand CbCr range.
+		 */
+		contrast = DIV_ROUND_CLOSEST(255 << 6, 235 - 16);
+		brightness = -DIV_ROUND_CLOSEST(16 * 255, 235 - 16);
+		sh_scale = DIV_ROUND_CLOSEST(128 << 7, 240 - 128);
+		sh_sin = SIN_0 * sh_scale;
+		sh_cos = COS_0 * sh_scale;
+	} else {
+		/* Pass-through everything. */
+		contrast = 1 << 6;
+		brightness = 0;
+		sh_scale = 1 << 7;
+		sh_sin = SIN_0 * sh_scale;
+		sh_cos = COS_0 * sh_scale;
+	}
+
+	/* FIXME these register are single buffered :( */
+	intel_de_write_fw(dev_priv, SPCLRC0(pipe, plane_id),
+			  SP_CONTRAST(contrast) | SP_BRIGHTNESS(brightness));
+	intel_de_write_fw(dev_priv, SPCLRC1(pipe, plane_id),
+			  SP_SH_SIN(sh_sin) | SP_SH_COS(sh_cos));
+}
+
+static void
+vlv_plane_ratio(const struct intel_crtc_state *crtc_state,
+		const struct intel_plane_state *plane_state,
+		unsigned int *num, unsigned int *den)
+{
+	u8 active_planes = crtc_state->active_planes & ~BIT(PLANE_CURSOR);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int cpp = fb->format->cpp[0];
+
+	/*
+	 * VLV bspec only considers cases where all three planes are
+	 * enabled, and cases where the primary and one sprite is enabled.
+	 * Let's assume the case with just two sprites enabled also
+	 * maps to the latter case.
+	 */
+	if (hweight8(active_planes) == 3) {
+		switch (cpp) {
+		case 8:
+			*num = 11;
+			*den = 8;
+			break;
+		case 4:
+			*num = 18;
+			*den = 16;
+			break;
+		default:
+			*num = 1;
+			*den = 1;
+			break;
+		}
+	} else if (hweight8(active_planes) == 2) {
+		switch (cpp) {
+		case 8:
+			*num = 10;
+			*den = 8;
+			break;
+		case 4:
+			*num = 17;
+			*den = 16;
+			break;
+		default:
+			*num = 1;
+			*den = 1;
+			break;
+		}
+	} else {
+		switch (cpp) {
+		case 8:
+			*num = 10;
+			*den = 8;
+			break;
+		default:
+			*num = 1;
+			*den = 1;
+			break;
+		}
+	}
+}
+
+int vlv_plane_min_cdclk(const struct intel_crtc_state *crtc_state,
+			const struct intel_plane_state *plane_state)
+{
+	unsigned int pixel_rate;
+	unsigned int num, den;
+
+	/*
+	 * Note that crtc_state->pixel_rate accounts for both
+	 * horizontal and vertical panel fitter downscaling factors.
+	 * Pre-HSW bspec tells us to only consider the horizontal
+	 * downscaling factor here. We ignore that and just consider
+	 * both for simplicity.
+	 */
+	pixel_rate = crtc_state->pixel_rate;
+
+	vlv_plane_ratio(crtc_state, plane_state, &num, &den);
+
+	return DIV_ROUND_UP(pixel_rate * num, den);
+}
+
+static u32 vlv_sprite_ctl_crtc(const struct intel_crtc_state *crtc_state)
+{
+	u32 sprctl = 0;
+
+	if (crtc_state->gamma_enable)
+		sprctl |= SP_PIPE_GAMMA_ENABLE;
+
+	return sprctl;
+}
+
+static u32 vlv_sprite_ctl(const struct intel_crtc_state *crtc_state,
+			  const struct intel_plane_state *plane_state)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+	const struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+	u32 sprctl;
+
+	sprctl = SP_ENABLE;
+
+	switch (fb->format->format) {
+	case DRM_FORMAT_YUYV:
+		sprctl |= SP_FORMAT_YUV422 | SP_YUV_ORDER_YUYV;
+		break;
+	case DRM_FORMAT_YVYU:
+		sprctl |= SP_FORMAT_YUV422 | SP_YUV_ORDER_YVYU;
+		break;
+	case DRM_FORMAT_UYVY:
+		sprctl |= SP_FORMAT_YUV422 | SP_YUV_ORDER_UYVY;
+		break;
+	case DRM_FORMAT_VYUY:
+		sprctl |= SP_FORMAT_YUV422 | SP_YUV_ORDER_VYUY;
+		break;
+	case DRM_FORMAT_C8:
+		sprctl |= SP_FORMAT_8BPP;
+		break;
+	case DRM_FORMAT_RGB565:
+		sprctl |= SP_FORMAT_BGR565;
+		break;
+	case DRM_FORMAT_XRGB8888:
+		sprctl |= SP_FORMAT_BGRX8888;
+		break;
+	case DRM_FORMAT_ARGB8888:
+		sprctl |= SP_FORMAT_BGRA8888;
+		break;
+	case DRM_FORMAT_XBGR2101010:
+		sprctl |= SP_FORMAT_RGBX1010102;
+		break;
+	case DRM_FORMAT_ABGR2101010:
+		sprctl |= SP_FORMAT_RGBA1010102;
+		break;
+	case DRM_FORMAT_XRGB2101010:
+		sprctl |= SP_FORMAT_BGRX1010102;
+		break;
+	case DRM_FORMAT_ARGB2101010:
+		sprctl |= SP_FORMAT_BGRA1010102;
+		break;
+	case DRM_FORMAT_XBGR8888:
+		sprctl |= SP_FORMAT_RGBX8888;
+		break;
+	case DRM_FORMAT_ABGR8888:
+		sprctl |= SP_FORMAT_RGBA8888;
+		break;
+	default:
+		MISSING_CASE(fb->format->format);
+		return 0;
+	}
+
+	if (plane_state->hw.color_encoding == DRM_COLOR_YCBCR_BT709)
+		sprctl |= SP_YUV_FORMAT_BT709;
+
+	if (fb->modifier == I915_FORMAT_MOD_X_TILED)
+		sprctl |= SP_TILED;
+
+	if (rotation & DRM_MODE_ROTATE_180)
+		sprctl |= SP_ROTATE_180;
+
+	if (rotation & DRM_MODE_REFLECT_X)
+		sprctl |= SP_MIRROR;
+
+	if (key->flags & I915_SET_COLORKEY_SOURCE)
+		sprctl |= SP_SOURCE_KEY;
+
+	return sprctl;
+}
+
+static void vlv_sprite_update_gamma(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	enum pipe pipe = plane->pipe;
+	enum plane_id plane_id = plane->id;
+	u16 gamma[8];
+	int i;
+
+	/* Seems RGB data bypasses the gamma always */
+	if (!fb->format->is_yuv)
+		return;
+
+	i9xx_plane_linear_gamma(gamma);
+
+	/* FIXME these register are single buffered :( */
+	/* The two end points are implicit (0.0 and 1.0) */
+	for (i = 1; i < 8 - 1; i++)
+		intel_de_write_fw(dev_priv, SPGAMC(pipe, plane_id, i - 1),
+				  gamma[i] << 16 | gamma[i] << 8 | gamma[i]);
+}
+
+static void
+vlv_sprite_update_noarm(struct intel_plane *plane,
+			const struct intel_crtc_state *crtc_state,
+			const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	enum plane_id plane_id = plane->id;
+	int crtc_x = plane_state->uapi.dst.x1;
+	int crtc_y = plane_state->uapi.dst.y1;
+	u32 crtc_w = drm_rect_width(&plane_state->uapi.dst);
+	u32 crtc_h = drm_rect_height(&plane_state->uapi.dst);
+
+	intel_de_write_fw(dev_priv, SPSTRIDE(pipe, plane_id),
+			  plane_state->view.color_plane[0].mapping_stride);
+	intel_de_write_fw(dev_priv, SPPOS(pipe, plane_id),
+			  SP_POS_Y(crtc_y) | SP_POS_X(crtc_x));
+	intel_de_write_fw(dev_priv, SPSIZE(pipe, plane_id),
+			  SP_HEIGHT(crtc_h - 1) | SP_WIDTH(crtc_w - 1));
+}
+
+static void
+vlv_sprite_update_arm(struct intel_plane *plane,
+		      const struct intel_crtc_state *crtc_state,
+		      const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	enum plane_id plane_id = plane->id;
+	const struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+	u32 sprsurf_offset = plane_state->view.color_plane[0].offset;
+	u32 x = plane_state->view.color_plane[0].x;
+	u32 y = plane_state->view.color_plane[0].y;
+	u32 sprctl, linear_offset;
+
+	sprctl = plane_state->ctl | vlv_sprite_ctl_crtc(crtc_state);
+
+	linear_offset = intel_fb_xy_to_linear(x, y, plane_state, 0);
+
+	if (IS_CHERRYVIEW(dev_priv) && pipe == PIPE_B)
+		chv_sprite_update_csc(plane_state);
+
+	if (key->flags) {
+		intel_de_write_fw(dev_priv, SPKEYMINVAL(pipe, plane_id),
+				  key->min_value);
+		intel_de_write_fw(dev_priv, SPKEYMSK(pipe, plane_id),
+				  key->channel_mask);
+		intel_de_write_fw(dev_priv, SPKEYMAXVAL(pipe, plane_id),
+				  key->max_value);
+	}
+
+	intel_de_write_fw(dev_priv, SPCONSTALPHA(pipe, plane_id), 0);
+
+	intel_de_write_fw(dev_priv, SPLINOFF(pipe, plane_id), linear_offset);
+	intel_de_write_fw(dev_priv, SPTILEOFF(pipe, plane_id),
+			  SP_OFFSET_Y(y) | SP_OFFSET_X(x));
+
+	/*
+	 * The control register self-arms if the plane was previously
+	 * disabled. Try to make the plane enable atomic by writing
+	 * the control register just before the surface register.
+	 */
+	intel_de_write_fw(dev_priv, SPCNTR(pipe, plane_id), sprctl);
+	intel_de_write_fw(dev_priv, SPSURF(pipe, plane_id),
+			  intel_plane_ggtt_offset(plane_state) + sprsurf_offset);
+
+	vlv_sprite_update_clrc(plane_state);
+	vlv_sprite_update_gamma(plane_state);
+}
+
+static void
+vlv_sprite_disable_arm(struct intel_plane *plane,
+		       const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	enum plane_id plane_id = plane->id;
+
+	intel_de_write_fw(dev_priv, SPCNTR(pipe, plane_id), 0);
+	intel_de_write_fw(dev_priv, SPSURF(pipe, plane_id), 0);
+}
+
+static bool
+vlv_sprite_get_hw_state(struct intel_plane *plane,
+			enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum intel_display_power_domain power_domain;
+	enum plane_id plane_id = plane->id;
+	intel_wakeref_t wakeref;
+	bool ret;
+
+	power_domain = POWER_DOMAIN_PIPE(plane->pipe);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref)
+		return false;
+
+	ret = intel_de_read(dev_priv, SPCNTR(plane->pipe, plane_id)) & SP_ENABLE;
+
+	*pipe = plane->pipe;
+
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+
+	return ret;
+}
+
+static void ivb_plane_ratio(const struct intel_crtc_state *crtc_state,
+			    const struct intel_plane_state *plane_state,
+			    unsigned int *num, unsigned int *den)
+{
+	u8 active_planes = crtc_state->active_planes & ~BIT(PLANE_CURSOR);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int cpp = fb->format->cpp[0];
+
+	if (hweight8(active_planes) == 2) {
+		switch (cpp) {
+		case 8:
+			*num = 10;
+			*den = 8;
+			break;
+		case 4:
+			*num = 17;
+			*den = 16;
+			break;
+		default:
+			*num = 1;
+			*den = 1;
+			break;
+		}
+	} else {
+		switch (cpp) {
+		case 8:
+			*num = 9;
+			*den = 8;
+			break;
+		default:
+			*num = 1;
+			*den = 1;
+			break;
+		}
+	}
+}
+
+static void ivb_plane_ratio_scaling(const struct intel_crtc_state *crtc_state,
+				    const struct intel_plane_state *plane_state,
+				    unsigned int *num, unsigned int *den)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int cpp = fb->format->cpp[0];
+
+	switch (cpp) {
+	case 8:
+		*num = 12;
+		*den = 8;
+		break;
+	case 4:
+		*num = 19;
+		*den = 16;
+		break;
+	case 2:
+		*num = 33;
+		*den = 32;
+		break;
+	default:
+		*num = 1;
+		*den = 1;
+		break;
+	}
+}
+
+int ivb_plane_min_cdclk(const struct intel_crtc_state *crtc_state,
+			const struct intel_plane_state *plane_state)
+{
+	unsigned int pixel_rate;
+	unsigned int num, den;
+
+	/*
+	 * Note that crtc_state->pixel_rate accounts for both
+	 * horizontal and vertical panel fitter downscaling factors.
+	 * Pre-HSW bspec tells us to only consider the horizontal
+	 * downscaling factor here. We ignore that and just consider
+	 * both for simplicity.
+	 */
+	pixel_rate = crtc_state->pixel_rate;
+
+	ivb_plane_ratio(crtc_state, plane_state, &num, &den);
+
+	return DIV_ROUND_UP(pixel_rate * num, den);
+}
+
+static int ivb_sprite_min_cdclk(const struct intel_crtc_state *crtc_state,
+				const struct intel_plane_state *plane_state)
+{
+	unsigned int src_w, dst_w, pixel_rate;
+	unsigned int num, den;
+
+	/*
+	 * Note that crtc_state->pixel_rate accounts for both
+	 * horizontal and vertical panel fitter downscaling factors.
+	 * Pre-HSW bspec tells us to only consider the horizontal
+	 * downscaling factor here. We ignore that and just consider
+	 * both for simplicity.
+	 */
+	pixel_rate = crtc_state->pixel_rate;
+
+	src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+	dst_w = drm_rect_width(&plane_state->uapi.dst);
+
+	if (src_w != dst_w)
+		ivb_plane_ratio_scaling(crtc_state, plane_state, &num, &den);
+	else
+		ivb_plane_ratio(crtc_state, plane_state, &num, &den);
+
+	/* Horizontal downscaling limits the maximum pixel rate */
+	dst_w = min(src_w, dst_w);
+
+	return DIV_ROUND_UP_ULL(mul_u32_u32(pixel_rate, num * src_w),
+				den * dst_w);
+}
+
+static void hsw_plane_ratio(const struct intel_crtc_state *crtc_state,
+			    const struct intel_plane_state *plane_state,
+			    unsigned int *num, unsigned int *den)
+{
+	u8 active_planes = crtc_state->active_planes & ~BIT(PLANE_CURSOR);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int cpp = fb->format->cpp[0];
+
+	if (hweight8(active_planes) == 2) {
+		switch (cpp) {
+		case 8:
+			*num = 10;
+			*den = 8;
+			break;
+		default:
+			*num = 1;
+			*den = 1;
+			break;
+		}
+	} else {
+		switch (cpp) {
+		case 8:
+			*num = 9;
+			*den = 8;
+			break;
+		default:
+			*num = 1;
+			*den = 1;
+			break;
+		}
+	}
+}
+
+int hsw_plane_min_cdclk(const struct intel_crtc_state *crtc_state,
+			const struct intel_plane_state *plane_state)
+{
+	unsigned int pixel_rate = crtc_state->pixel_rate;
+	unsigned int num, den;
+
+	hsw_plane_ratio(crtc_state, plane_state, &num, &den);
+
+	return DIV_ROUND_UP(pixel_rate * num, den);
+}
+
+static u32 ivb_sprite_ctl_crtc(const struct intel_crtc_state *crtc_state)
+{
+	u32 sprctl = 0;
+
+	if (crtc_state->gamma_enable)
+		sprctl |= SPRITE_PIPE_GAMMA_ENABLE;
+
+	if (crtc_state->csc_enable)
+		sprctl |= SPRITE_PIPE_CSC_ENABLE;
+
+	return sprctl;
+}
+
+static bool ivb_need_sprite_gamma(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+
+	return fb->format->cpp[0] == 8 &&
+		(IS_IVYBRIDGE(dev_priv) || IS_HASWELL(dev_priv));
+}
+
+static u32 ivb_sprite_ctl(const struct intel_crtc_state *crtc_state,
+			  const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+	const struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+	u32 sprctl;
+
+	sprctl = SPRITE_ENABLE;
+
+	if (IS_IVYBRIDGE(dev_priv))
+		sprctl |= SPRITE_TRICKLE_FEED_DISABLE;
+
+	switch (fb->format->format) {
+	case DRM_FORMAT_XBGR8888:
+		sprctl |= SPRITE_FORMAT_RGBX888 | SPRITE_RGB_ORDER_RGBX;
+		break;
+	case DRM_FORMAT_XRGB8888:
+		sprctl |= SPRITE_FORMAT_RGBX888;
+		break;
+	case DRM_FORMAT_XBGR2101010:
+		sprctl |= SPRITE_FORMAT_RGBX101010 | SPRITE_RGB_ORDER_RGBX;
+		break;
+	case DRM_FORMAT_XRGB2101010:
+		sprctl |= SPRITE_FORMAT_RGBX101010;
+		break;
+	case DRM_FORMAT_XBGR16161616F:
+		sprctl |= SPRITE_FORMAT_RGBX161616 | SPRITE_RGB_ORDER_RGBX;
+		break;
+	case DRM_FORMAT_XRGB16161616F:
+		sprctl |= SPRITE_FORMAT_RGBX161616;
+		break;
+	case DRM_FORMAT_YUYV:
+		sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_YUYV;
+		break;
+	case DRM_FORMAT_YVYU:
+		sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_YVYU;
+		break;
+	case DRM_FORMAT_UYVY:
+		sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_UYVY;
+		break;
+	case DRM_FORMAT_VYUY:
+		sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_VYUY;
+		break;
+	default:
+		MISSING_CASE(fb->format->format);
+		return 0;
+	}
+
+	if (!ivb_need_sprite_gamma(plane_state))
+		sprctl |= SPRITE_PLANE_GAMMA_DISABLE;
+
+	if (plane_state->hw.color_encoding == DRM_COLOR_YCBCR_BT709)
+		sprctl |= SPRITE_YUV_TO_RGB_CSC_FORMAT_BT709;
+
+	if (plane_state->hw.color_range == DRM_COLOR_YCBCR_FULL_RANGE)
+		sprctl |= SPRITE_YUV_RANGE_CORRECTION_DISABLE;
+
+	if (fb->modifier == I915_FORMAT_MOD_X_TILED)
+		sprctl |= SPRITE_TILED;
+
+	if (rotation & DRM_MODE_ROTATE_180)
+		sprctl |= SPRITE_ROTATE_180;
+
+	if (key->flags & I915_SET_COLORKEY_DESTINATION)
+		sprctl |= SPRITE_DEST_KEY;
+	else if (key->flags & I915_SET_COLORKEY_SOURCE)
+		sprctl |= SPRITE_SOURCE_KEY;
+
+	return sprctl;
+}
+
+static void ivb_sprite_linear_gamma(const struct intel_plane_state *plane_state,
+				    u16 gamma[18])
+{
+	int scale, i;
+
+	/*
+	 * WaFP16GammaEnabling:ivb,hsw
+	 * "Workaround : When using the 64-bit format, the sprite output
+	 *  on each color channel has one quarter amplitude. It can be
+	 *  brought up to full amplitude by using sprite internal gamma
+	 *  correction, pipe gamma correction, or pipe color space
+	 *  conversion to multiply the sprite output by four."
+	 */
+	scale = 4;
+
+	for (i = 0; i < 16; i++)
+		gamma[i] = min((scale * i << 10) / 16, (1 << 10) - 1);
+
+	gamma[i] = min((scale * i << 10) / 16, 1 << 10);
+	i++;
+
+	gamma[i] = 3 << 10;
+	i++;
+}
+
+static void ivb_sprite_update_gamma(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	u16 gamma[18];
+	int i;
+
+	if (!ivb_need_sprite_gamma(plane_state))
+		return;
+
+	ivb_sprite_linear_gamma(plane_state, gamma);
+
+	/* FIXME these register are single buffered :( */
+	for (i = 0; i < 16; i++)
+		intel_de_write_fw(dev_priv, SPRGAMC(pipe, i),
+				  gamma[i] << 20 | gamma[i] << 10 | gamma[i]);
+
+	intel_de_write_fw(dev_priv, SPRGAMC16(pipe, 0), gamma[i]);
+	intel_de_write_fw(dev_priv, SPRGAMC16(pipe, 1), gamma[i]);
+	intel_de_write_fw(dev_priv, SPRGAMC16(pipe, 2), gamma[i]);
+	i++;
+
+	intel_de_write_fw(dev_priv, SPRGAMC17(pipe, 0), gamma[i]);
+	intel_de_write_fw(dev_priv, SPRGAMC17(pipe, 1), gamma[i]);
+	intel_de_write_fw(dev_priv, SPRGAMC17(pipe, 2), gamma[i]);
+	i++;
+}
+
+static void
+ivb_sprite_update_noarm(struct intel_plane *plane,
+			const struct intel_crtc_state *crtc_state,
+			const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	int crtc_x = plane_state->uapi.dst.x1;
+	int crtc_y = plane_state->uapi.dst.y1;
+	u32 crtc_w = drm_rect_width(&plane_state->uapi.dst);
+	u32 crtc_h = drm_rect_height(&plane_state->uapi.dst);
+	u32 src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+	u32 src_h = drm_rect_height(&plane_state->uapi.src) >> 16;
+	u32 sprscale = 0;
+
+	if (crtc_w != src_w || crtc_h != src_h)
+		sprscale = SPRITE_SCALE_ENABLE |
+			SPRITE_SRC_WIDTH(src_w - 1) |
+			SPRITE_SRC_HEIGHT(src_h - 1);
+
+	intel_de_write_fw(dev_priv, SPRSTRIDE(pipe),
+			  plane_state->view.color_plane[0].mapping_stride);
+	intel_de_write_fw(dev_priv, SPRPOS(pipe),
+			  SPRITE_POS_Y(crtc_y) | SPRITE_POS_X(crtc_x));
+	intel_de_write_fw(dev_priv, SPRSIZE(pipe),
+			  SPRITE_HEIGHT(crtc_h - 1) | SPRITE_WIDTH(crtc_w - 1));
+	if (IS_IVYBRIDGE(dev_priv))
+		intel_de_write_fw(dev_priv, SPRSCALE(pipe), sprscale);
+}
+
+static void
+ivb_sprite_update_arm(struct intel_plane *plane,
+		      const struct intel_crtc_state *crtc_state,
+		      const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	const struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+	u32 sprsurf_offset = plane_state->view.color_plane[0].offset;
+	u32 x = plane_state->view.color_plane[0].x;
+	u32 y = plane_state->view.color_plane[0].y;
+	u32 sprctl, linear_offset;
+
+	sprctl = plane_state->ctl | ivb_sprite_ctl_crtc(crtc_state);
+
+	linear_offset = intel_fb_xy_to_linear(x, y, plane_state, 0);
+
+	if (key->flags) {
+		intel_de_write_fw(dev_priv, SPRKEYVAL(pipe), key->min_value);
+		intel_de_write_fw(dev_priv, SPRKEYMSK(pipe),
+				  key->channel_mask);
+		intel_de_write_fw(dev_priv, SPRKEYMAX(pipe), key->max_value);
+	}
+
+	/* HSW consolidates SPRTILEOFF and SPRLINOFF into a single SPROFFSET
+	 * register */
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
+		intel_de_write_fw(dev_priv, SPROFFSET(pipe),
+				  SPRITE_OFFSET_Y(y) | SPRITE_OFFSET_X(x));
+	} else {
+		intel_de_write_fw(dev_priv, SPRLINOFF(pipe), linear_offset);
+		intel_de_write_fw(dev_priv, SPRTILEOFF(pipe),
+				  SPRITE_OFFSET_Y(y) | SPRITE_OFFSET_X(x));
+	}
+
+	/*
+	 * The control register self-arms if the plane was previously
+	 * disabled. Try to make the plane enable atomic by writing
+	 * the control register just before the surface register.
+	 */
+	intel_de_write_fw(dev_priv, SPRCTL(pipe), sprctl);
+	intel_de_write_fw(dev_priv, SPRSURF(pipe),
+			  intel_plane_ggtt_offset(plane_state) + sprsurf_offset);
+
+	ivb_sprite_update_gamma(plane_state);
+}
+
+static void
+ivb_sprite_disable_arm(struct intel_plane *plane,
+		       const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	intel_de_write_fw(dev_priv, SPRCTL(pipe), 0);
+	/* Disable the scaler */
+	if (IS_IVYBRIDGE(dev_priv))
+		intel_de_write_fw(dev_priv, SPRSCALE(pipe), 0);
+	intel_de_write_fw(dev_priv, SPRSURF(pipe), 0);
+}
+
+static bool
+ivb_sprite_get_hw_state(struct intel_plane *plane,
+			enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum intel_display_power_domain power_domain;
+	intel_wakeref_t wakeref;
+	bool ret;
+
+	power_domain = POWER_DOMAIN_PIPE(plane->pipe);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref)
+		return false;
+
+	ret =  intel_de_read(dev_priv, SPRCTL(plane->pipe)) & SPRITE_ENABLE;
+
+	*pipe = plane->pipe;
+
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+
+	return ret;
+}
+
+static int g4x_sprite_min_cdclk(const struct intel_crtc_state *crtc_state,
+				const struct intel_plane_state *plane_state)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int hscale, pixel_rate;
+	unsigned int limit, decimate;
+
+	/*
+	 * Note that crtc_state->pixel_rate accounts for both
+	 * horizontal and vertical panel fitter downscaling factors.
+	 * Pre-HSW bspec tells us to only consider the horizontal
+	 * downscaling factor here. We ignore that and just consider
+	 * both for simplicity.
+	 */
+	pixel_rate = crtc_state->pixel_rate;
+
+	/* Horizontal downscaling limits the maximum pixel rate */
+	hscale = drm_rect_calc_hscale(&plane_state->uapi.src,
+				      &plane_state->uapi.dst,
+				      0, INT_MAX);
+	hscale = max(hscale, 0x10000u);
+
+	/* Decimation steps at 2x,4x,8x,16x */
+	decimate = ilog2(hscale >> 16);
+	hscale >>= decimate;
+
+	/* Starting limit is 90% of cdclk */
+	limit = 9;
+
+	/* -10% per decimation step */
+	limit -= decimate;
+
+	/* -10% for RGB */
+	if (!fb->format->is_yuv)
+		limit--;
+
+	/*
+	 * We should also do -10% if sprite scaling is enabled
+	 * on the other pipe, but we can't really check for that,
+	 * so we ignore it.
+	 */
+
+	return DIV_ROUND_UP_ULL(mul_u32_u32(pixel_rate, 10 * hscale),
+				limit << 16);
+}
+
+static unsigned int
+g4x_sprite_max_stride(struct intel_plane *plane,
+		      u32 pixel_format, u64 modifier,
+		      unsigned int rotation)
+{
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+
+	/* Limit to 4k pixels to guarantee TILEOFF.x doesn't get too big. */
+	if (modifier == I915_FORMAT_MOD_X_TILED)
+		return min(4096 * cpp, 16 * 1024);
+	else
+		return 16 * 1024;
+}
+
+static unsigned int
+hsw_sprite_max_stride(struct intel_plane *plane,
+		      u32 pixel_format, u64 modifier,
+		      unsigned int rotation)
+{
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+
+	/* Limit to 8k pixels to guarantee OFFSET.x doesn't get too big. */
+	return min(8192 * cpp, 16 * 1024);
+}
+
+static u32 g4x_sprite_ctl_crtc(const struct intel_crtc_state *crtc_state)
+{
+	u32 dvscntr = 0;
+
+	if (crtc_state->gamma_enable)
+		dvscntr |= DVS_PIPE_GAMMA_ENABLE;
+
+	if (crtc_state->csc_enable)
+		dvscntr |= DVS_PIPE_CSC_ENABLE;
+
+	return dvscntr;
+}
+
+static u32 g4x_sprite_ctl(const struct intel_crtc_state *crtc_state,
+			  const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+	const struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+	u32 dvscntr;
+
+	dvscntr = DVS_ENABLE;
+
+	if (IS_SANDYBRIDGE(dev_priv))
+		dvscntr |= DVS_TRICKLE_FEED_DISABLE;
+
+	switch (fb->format->format) {
+	case DRM_FORMAT_XBGR8888:
+		dvscntr |= DVS_FORMAT_RGBX888 | DVS_RGB_ORDER_XBGR;
+		break;
+	case DRM_FORMAT_XRGB8888:
+		dvscntr |= DVS_FORMAT_RGBX888;
+		break;
+	case DRM_FORMAT_XBGR2101010:
+		dvscntr |= DVS_FORMAT_RGBX101010 | DVS_RGB_ORDER_XBGR;
+		break;
+	case DRM_FORMAT_XRGB2101010:
+		dvscntr |= DVS_FORMAT_RGBX101010;
+		break;
+	case DRM_FORMAT_XBGR16161616F:
+		dvscntr |= DVS_FORMAT_RGBX161616 | DVS_RGB_ORDER_XBGR;
+		break;
+	case DRM_FORMAT_XRGB16161616F:
+		dvscntr |= DVS_FORMAT_RGBX161616;
+		break;
+	case DRM_FORMAT_YUYV:
+		dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_YUYV;
+		break;
+	case DRM_FORMAT_YVYU:
+		dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_YVYU;
+		break;
+	case DRM_FORMAT_UYVY:
+		dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_UYVY;
+		break;
+	case DRM_FORMAT_VYUY:
+		dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_VYUY;
+		break;
+	default:
+		MISSING_CASE(fb->format->format);
+		return 0;
+	}
+
+	if (plane_state->hw.color_encoding == DRM_COLOR_YCBCR_BT709)
+		dvscntr |= DVS_YUV_FORMAT_BT709;
+
+	if (plane_state->hw.color_range == DRM_COLOR_YCBCR_FULL_RANGE)
+		dvscntr |= DVS_YUV_RANGE_CORRECTION_DISABLE;
+
+	if (fb->modifier == I915_FORMAT_MOD_X_TILED)
+		dvscntr |= DVS_TILED;
+
+	if (rotation & DRM_MODE_ROTATE_180)
+		dvscntr |= DVS_ROTATE_180;
+
+	if (key->flags & I915_SET_COLORKEY_DESTINATION)
+		dvscntr |= DVS_DEST_KEY;
+	else if (key->flags & I915_SET_COLORKEY_SOURCE)
+		dvscntr |= DVS_SOURCE_KEY;
+
+	return dvscntr;
+}
+
+static void g4x_sprite_update_gamma(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	enum pipe pipe = plane->pipe;
+	u16 gamma[8];
+	int i;
+
+	/* Seems RGB data bypasses the gamma always */
+	if (!fb->format->is_yuv)
+		return;
+
+	i9xx_plane_linear_gamma(gamma);
+
+	/* FIXME these register are single buffered :( */
+	/* The two end points are implicit (0.0 and 1.0) */
+	for (i = 1; i < 8 - 1; i++)
+		intel_de_write_fw(dev_priv, DVSGAMC_G4X(pipe, i - 1),
+				  gamma[i] << 16 | gamma[i] << 8 | gamma[i]);
+}
+
+static void ilk_sprite_linear_gamma(u16 gamma[17])
+{
+	int i;
+
+	for (i = 0; i < 17; i++)
+		gamma[i] = (i << 10) / 16;
+}
+
+static void ilk_sprite_update_gamma(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	enum pipe pipe = plane->pipe;
+	u16 gamma[17];
+	int i;
+
+	/* Seems RGB data bypasses the gamma always */
+	if (!fb->format->is_yuv)
+		return;
+
+	ilk_sprite_linear_gamma(gamma);
+
+	/* FIXME these register are single buffered :( */
+	for (i = 0; i < 16; i++)
+		intel_de_write_fw(dev_priv, DVSGAMC_ILK(pipe, i),
+				  gamma[i] << 20 | gamma[i] << 10 | gamma[i]);
+
+	intel_de_write_fw(dev_priv, DVSGAMCMAX_ILK(pipe, 0), gamma[i]);
+	intel_de_write_fw(dev_priv, DVSGAMCMAX_ILK(pipe, 1), gamma[i]);
+	intel_de_write_fw(dev_priv, DVSGAMCMAX_ILK(pipe, 2), gamma[i]);
+	i++;
+}
+
+static void
+g4x_sprite_update_noarm(struct intel_plane *plane,
+			const struct intel_crtc_state *crtc_state,
+			const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	int crtc_x = plane_state->uapi.dst.x1;
+	int crtc_y = plane_state->uapi.dst.y1;
+	u32 crtc_w = drm_rect_width(&plane_state->uapi.dst);
+	u32 crtc_h = drm_rect_height(&plane_state->uapi.dst);
+	u32 src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+	u32 src_h = drm_rect_height(&plane_state->uapi.src) >> 16;
+	u32 dvsscale = 0;
+
+	if (crtc_w != src_w || crtc_h != src_h)
+		dvsscale = DVS_SCALE_ENABLE |
+			DVS_SRC_WIDTH(src_w - 1) |
+			DVS_SRC_HEIGHT(src_h - 1);
+
+	intel_de_write_fw(dev_priv, DVSSTRIDE(pipe),
+			  plane_state->view.color_plane[0].mapping_stride);
+	intel_de_write_fw(dev_priv, DVSPOS(pipe),
+			  DVS_POS_Y(crtc_y) | DVS_POS_X(crtc_x));
+	intel_de_write_fw(dev_priv, DVSSIZE(pipe),
+			  DVS_HEIGHT(crtc_h - 1) | DVS_WIDTH(crtc_w - 1));
+	intel_de_write_fw(dev_priv, DVSSCALE(pipe), dvsscale);
+}
+
+static void
+g4x_sprite_update_arm(struct intel_plane *plane,
+		      const struct intel_crtc_state *crtc_state,
+		      const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	const struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+	u32 dvssurf_offset = plane_state->view.color_plane[0].offset;
+	u32 x = plane_state->view.color_plane[0].x;
+	u32 y = plane_state->view.color_plane[0].y;
+	u32 dvscntr, linear_offset;
+
+	dvscntr = plane_state->ctl | g4x_sprite_ctl_crtc(crtc_state);
+
+	linear_offset = intel_fb_xy_to_linear(x, y, plane_state, 0);
+
+	if (key->flags) {
+		intel_de_write_fw(dev_priv, DVSKEYVAL(pipe), key->min_value);
+		intel_de_write_fw(dev_priv, DVSKEYMSK(pipe),
+				  key->channel_mask);
+		intel_de_write_fw(dev_priv, DVSKEYMAX(pipe), key->max_value);
+	}
+
+	intel_de_write_fw(dev_priv, DVSLINOFF(pipe), linear_offset);
+	intel_de_write_fw(dev_priv, DVSTILEOFF(pipe), (y << 16) | x);
+
+	/*
+	 * The control register self-arms if the plane was previously
+	 * disabled. Try to make the plane enable atomic by writing
+	 * the control register just before the surface register.
+	 */
+	intel_de_write_fw(dev_priv, DVSCNTR(pipe), dvscntr);
+	intel_de_write_fw(dev_priv, DVSSURF(pipe),
+			  intel_plane_ggtt_offset(plane_state) + dvssurf_offset);
+
+	if (IS_G4X(dev_priv))
+		g4x_sprite_update_gamma(plane_state);
+	else
+		ilk_sprite_update_gamma(plane_state);
+}
+
+static void
+g4x_sprite_disable_arm(struct intel_plane *plane,
+		       const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	intel_de_write_fw(dev_priv, DVSCNTR(pipe), 0);
+	/* Disable the scaler */
+	intel_de_write_fw(dev_priv, DVSSCALE(pipe), 0);
+	intel_de_write_fw(dev_priv, DVSSURF(pipe), 0);
+}
+
+static bool
+g4x_sprite_get_hw_state(struct intel_plane *plane,
+			enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum intel_display_power_domain power_domain;
+	intel_wakeref_t wakeref;
+	bool ret;
+
+	power_domain = POWER_DOMAIN_PIPE(plane->pipe);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref)
+		return false;
+
+	ret = intel_de_read(dev_priv, DVSCNTR(plane->pipe)) & DVS_ENABLE;
+
+	*pipe = plane->pipe;
+
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+
+	return ret;
+}
+
+static bool g4x_fb_scalable(const struct drm_framebuffer *fb)
+{
+	if (!fb)
+		return false;
+
+	switch (fb->format->format) {
+	case DRM_FORMAT_C8:
+	case DRM_FORMAT_XRGB16161616F:
+	case DRM_FORMAT_ARGB16161616F:
+	case DRM_FORMAT_XBGR16161616F:
+	case DRM_FORMAT_ABGR16161616F:
+		return false;
+	default:
+		return true;
+	}
+}
+
+static int
+g4x_sprite_check_scaling(struct intel_crtc_state *crtc_state,
+			 struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	const struct drm_rect *src = &plane_state->uapi.src;
+	const struct drm_rect *dst = &plane_state->uapi.dst;
+	int src_x, src_w, src_h, crtc_w, crtc_h;
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	unsigned int stride = plane_state->view.color_plane[0].mapping_stride;
+	unsigned int cpp = fb->format->cpp[0];
+	unsigned int width_bytes;
+	int min_width, min_height;
+
+	crtc_w = drm_rect_width(dst);
+	crtc_h = drm_rect_height(dst);
+
+	src_x = src->x1 >> 16;
+	src_w = drm_rect_width(src) >> 16;
+	src_h = drm_rect_height(src) >> 16;
+
+	if (src_w == crtc_w && src_h == crtc_h)
+		return 0;
+
+	min_width = 3;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
+		if (src_h & 1) {
+			drm_dbg_kms(&i915->drm, "Source height must be even with interlaced modes\n");
+			return -EINVAL;
+		}
+		min_height = 6;
+	} else {
+		min_height = 3;
+	}
+
+	width_bytes = ((src_x * cpp) & 63) + src_w * cpp;
+
+	if (src_w < min_width || src_h < min_height ||
+	    src_w > 2048 || src_h > 2048) {
+		drm_dbg_kms(&i915->drm, "Source dimensions (%dx%d) exceed hardware limits (%dx%d - %dx%d)\n",
+			    src_w, src_h, min_width, min_height, 2048, 2048);
+		return -EINVAL;
+	}
+
+	if (width_bytes > 4096) {
+		drm_dbg_kms(&i915->drm, "Fetch width (%d) exceeds hardware max with scaling (%u)\n",
+			    width_bytes, 4096);
+		return -EINVAL;
+	}
+
+	if (stride > 4096) {
+		drm_dbg_kms(&i915->drm, "Stride (%u) exceeds hardware max with scaling (%u)\n",
+			    stride, 4096);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+g4x_sprite_check(struct intel_crtc_state *crtc_state,
+		 struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	int min_scale = DRM_PLANE_NO_SCALING;
+	int max_scale = DRM_PLANE_NO_SCALING;
+	int ret;
+
+	if (g4x_fb_scalable(plane_state->hw.fb)) {
+		if (DISPLAY_VER(dev_priv) < 7) {
+			min_scale = 1;
+			max_scale = 16 << 16;
+		} else if (IS_IVYBRIDGE(dev_priv)) {
+			min_scale = 1;
+			max_scale = 2 << 16;
+		}
+	}
+
+	ret = intel_atomic_plane_check_clipping(plane_state, crtc_state,
+						min_scale, max_scale, true);
+	if (ret)
+		return ret;
+
+	ret = i9xx_check_plane_surface(plane_state);
+	if (ret)
+		return ret;
+
+	if (!plane_state->uapi.visible)
+		return 0;
+
+	ret = intel_plane_check_src_coordinates(plane_state);
+	if (ret)
+		return ret;
+
+	ret = g4x_sprite_check_scaling(crtc_state, plane_state);
+	if (ret)
+		return ret;
+
+	if (DISPLAY_VER(dev_priv) >= 7)
+		plane_state->ctl = ivb_sprite_ctl(crtc_state, plane_state);
+	else
+		plane_state->ctl = g4x_sprite_ctl(crtc_state, plane_state);
+
+	return 0;
+}
+
+int chv_plane_check_rotation(const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	unsigned int rotation = plane_state->hw.rotation;
+
+	/* CHV ignores the mirror bit when the rotate bit is set :( */
+	if (IS_CHERRYVIEW(dev_priv) &&
+	    rotation & DRM_MODE_ROTATE_180 &&
+	    rotation & DRM_MODE_REFLECT_X) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Cannot rotate and reflect at the same time\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+vlv_sprite_check(struct intel_crtc_state *crtc_state,
+		 struct intel_plane_state *plane_state)
+{
+	int ret;
+
+	ret = chv_plane_check_rotation(plane_state);
+	if (ret)
+		return ret;
+
+	ret = intel_atomic_plane_check_clipping(plane_state, crtc_state,
+						DRM_PLANE_NO_SCALING,
+						DRM_PLANE_NO_SCALING,
+						true);
+	if (ret)
+		return ret;
+
+	ret = i9xx_check_plane_surface(plane_state);
+	if (ret)
+		return ret;
+
+	if (!plane_state->uapi.visible)
+		return 0;
+
+	ret = intel_plane_check_src_coordinates(plane_state);
+	if (ret)
+		return ret;
+
+	plane_state->ctl = vlv_sprite_ctl(crtc_state, plane_state);
+
+	return 0;
+}
+
+static bool has_dst_key_in_primary_plane(struct drm_i915_private *dev_priv)
+{
+	return DISPLAY_VER(dev_priv) >= 9;
+}
+
+static void intel_plane_set_ckey(struct intel_plane_state *plane_state,
+				 const struct drm_intel_sprite_colorkey *set)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+
+	*key = *set;
+
+	/*
+	 * We want src key enabled on the
+	 * sprite and not on the primary.
+	 */
+	if (plane->id == PLANE_PRIMARY &&
+	    set->flags & I915_SET_COLORKEY_SOURCE)
+		key->flags = 0;
+
+	/*
+	 * On SKL+ we want dst key enabled on
+	 * the primary and not on the sprite.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 9 && plane->id != PLANE_PRIMARY &&
+	    set->flags & I915_SET_COLORKEY_DESTINATION)
+		key->flags = 0;
+}
+
+int intel_sprite_set_colorkey_ioctl(struct drm_device *dev, void *data,
+				    struct drm_file *file_priv)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_intel_sprite_colorkey *set = data;
+	struct drm_plane *plane;
+	struct drm_plane_state *plane_state;
+	struct drm_atomic_state *state;
+	struct drm_modeset_acquire_ctx ctx;
+	int ret = 0;
+
+	/* ignore the pointless "none" flag */
+	set->flags &= ~I915_SET_COLORKEY_NONE;
+
+	if (set->flags & ~(I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE))
+		return -EINVAL;
+
+	/* Make sure we don't try to enable both src & dest simultaneously */
+	if ((set->flags & (I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE)) == (I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE))
+		return -EINVAL;
+
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    set->flags & I915_SET_COLORKEY_DESTINATION)
+		return -EINVAL;
+
+	plane = drm_plane_find(dev, file_priv, set->plane_id);
+	if (!plane || plane->type != DRM_PLANE_TYPE_OVERLAY)
+		return -ENOENT;
+
+	/*
+	 * SKL+ only plane 2 can do destination keying against plane 1.
+	 * Also multiple planes can't do destination keying on the same
+	 * pipe simultaneously.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 9 &&
+	    to_intel_plane(plane)->id >= PLANE_SPRITE1 &&
+	    set->flags & I915_SET_COLORKEY_DESTINATION)
+		return -EINVAL;
+
+	drm_modeset_acquire_init(&ctx, 0);
+
+	state = drm_atomic_state_alloc(plane->dev);
+	if (!state) {
+		ret = -ENOMEM;
+		goto out;
+	}
+	state->acquire_ctx = &ctx;
+
+	while (1) {
+		plane_state = drm_atomic_get_plane_state(state, plane);
+		ret = PTR_ERR_OR_ZERO(plane_state);
+		if (!ret)
+			intel_plane_set_ckey(to_intel_plane_state(plane_state), set);
+
+		/*
+		 * On some platforms we have to configure
+		 * the dst colorkey on the primary plane.
+		 */
+		if (!ret && has_dst_key_in_primary_plane(dev_priv)) {
+			struct intel_crtc *crtc =
+				intel_crtc_for_pipe(dev_priv,
+						    to_intel_plane(plane)->pipe);
+
+			plane_state = drm_atomic_get_plane_state(state,
+								 crtc->base.primary);
+			ret = PTR_ERR_OR_ZERO(plane_state);
+			if (!ret)
+				intel_plane_set_ckey(to_intel_plane_state(plane_state), set);
+		}
+
+		if (!ret)
+			ret = drm_atomic_commit(state);
+
+		if (ret != -EDEADLK)
+			break;
+
+		drm_atomic_state_clear(state);
+		drm_modeset_backoff(&ctx);
+	}
+
+	drm_atomic_state_put(state);
+out:
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+	return ret;
+}
+
+static const u32 g4x_sprite_formats[] = {
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+};
+
+static const u32 snb_sprite_formats[] = {
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_XRGB16161616F,
+	DRM_FORMAT_XBGR16161616F,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+};
+
+static const u32 vlv_sprite_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_ABGR2101010,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+};
+
+static const u32 chv_pipe_b_sprite_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_ARGB2101010,
+	DRM_FORMAT_ABGR2101010,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+};
+
+static bool g4x_sprite_format_mod_supported(struct drm_plane *_plane,
+					    u32 format, u64 modifier)
+{
+	if (!intel_fb_plane_supports_modifier(to_intel_plane(_plane), modifier))
+		return false;
+
+	switch (format) {
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_YVYU:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_VYUY:
+		if (modifier == DRM_FORMAT_MOD_LINEAR ||
+		    modifier == I915_FORMAT_MOD_X_TILED)
+			return true;
+		fallthrough;
+	default:
+		return false;
+	}
+}
+
+static bool snb_sprite_format_mod_supported(struct drm_plane *_plane,
+					    u32 format, u64 modifier)
+{
+	if (!intel_fb_plane_supports_modifier(to_intel_plane(_plane), modifier))
+		return false;
+
+	switch (format) {
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_XBGR2101010:
+	case DRM_FORMAT_XRGB16161616F:
+	case DRM_FORMAT_XBGR16161616F:
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_YVYU:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_VYUY:
+		if (modifier == DRM_FORMAT_MOD_LINEAR ||
+		    modifier == I915_FORMAT_MOD_X_TILED)
+			return true;
+		fallthrough;
+	default:
+		return false;
+	}
+}
+
+static bool vlv_sprite_format_mod_supported(struct drm_plane *_plane,
+					    u32 format, u64 modifier)
+{
+	if (!intel_fb_plane_supports_modifier(to_intel_plane(_plane), modifier))
+		return false;
+
+	switch (format) {
+	case DRM_FORMAT_C8:
+	case DRM_FORMAT_RGB565:
+	case DRM_FORMAT_ABGR8888:
+	case DRM_FORMAT_ARGB8888:
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_XBGR2101010:
+	case DRM_FORMAT_ABGR2101010:
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_ARGB2101010:
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_YVYU:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_VYUY:
+		if (modifier == DRM_FORMAT_MOD_LINEAR ||
+		    modifier == I915_FORMAT_MOD_X_TILED)
+			return true;
+		fallthrough;
+	default:
+		return false;
+	}
+}
+
+static const struct drm_plane_funcs g4x_sprite_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = intel_plane_destroy,
+	.atomic_duplicate_state = intel_plane_duplicate_state,
+	.atomic_destroy_state = intel_plane_destroy_state,
+	.format_mod_supported = g4x_sprite_format_mod_supported,
+};
+
+static const struct drm_plane_funcs snb_sprite_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = intel_plane_destroy,
+	.atomic_duplicate_state = intel_plane_duplicate_state,
+	.atomic_destroy_state = intel_plane_destroy_state,
+	.format_mod_supported = snb_sprite_format_mod_supported,
+};
+
+static const struct drm_plane_funcs vlv_sprite_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = intel_plane_destroy,
+	.atomic_duplicate_state = intel_plane_duplicate_state,
+	.atomic_destroy_state = intel_plane_destroy_state,
+	.format_mod_supported = vlv_sprite_format_mod_supported,
+};
+
+struct intel_plane *
+intel_sprite_plane_create(struct drm_i915_private *dev_priv,
+			  enum pipe pipe, int sprite)
+{
+	struct intel_plane *plane;
+	const struct drm_plane_funcs *plane_funcs;
+	unsigned int supported_rotations;
+	const u64 *modifiers;
+	const u32 *formats;
+	int num_formats;
+	int ret, zpos;
+
+	plane = intel_plane_alloc();
+	if (IS_ERR(plane))
+		return plane;
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		plane->update_noarm = vlv_sprite_update_noarm;
+		plane->update_arm = vlv_sprite_update_arm;
+		plane->disable_arm = vlv_sprite_disable_arm;
+		plane->get_hw_state = vlv_sprite_get_hw_state;
+		plane->check_plane = vlv_sprite_check;
+		plane->max_stride = i965_plane_max_stride;
+		plane->min_cdclk = vlv_plane_min_cdclk;
+
+		if (IS_CHERRYVIEW(dev_priv) && pipe == PIPE_B) {
+			formats = chv_pipe_b_sprite_formats;
+			num_formats = ARRAY_SIZE(chv_pipe_b_sprite_formats);
+		} else {
+			formats = vlv_sprite_formats;
+			num_formats = ARRAY_SIZE(vlv_sprite_formats);
+		}
+
+		plane_funcs = &vlv_sprite_funcs;
+	} else if (DISPLAY_VER(dev_priv) >= 7) {
+		plane->update_noarm = ivb_sprite_update_noarm;
+		plane->update_arm = ivb_sprite_update_arm;
+		plane->disable_arm = ivb_sprite_disable_arm;
+		plane->get_hw_state = ivb_sprite_get_hw_state;
+		plane->check_plane = g4x_sprite_check;
+
+		if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv)) {
+			plane->max_stride = hsw_sprite_max_stride;
+			plane->min_cdclk = hsw_plane_min_cdclk;
+		} else {
+			plane->max_stride = g4x_sprite_max_stride;
+			plane->min_cdclk = ivb_sprite_min_cdclk;
+		}
+
+		formats = snb_sprite_formats;
+		num_formats = ARRAY_SIZE(snb_sprite_formats);
+
+		plane_funcs = &snb_sprite_funcs;
+	} else {
+		plane->update_noarm = g4x_sprite_update_noarm;
+		plane->update_arm = g4x_sprite_update_arm;
+		plane->disable_arm = g4x_sprite_disable_arm;
+		plane->get_hw_state = g4x_sprite_get_hw_state;
+		plane->check_plane = g4x_sprite_check;
+		plane->max_stride = g4x_sprite_max_stride;
+		plane->min_cdclk = g4x_sprite_min_cdclk;
+
+		if (IS_SANDYBRIDGE(dev_priv)) {
+			formats = snb_sprite_formats;
+			num_formats = ARRAY_SIZE(snb_sprite_formats);
+
+			plane_funcs = &snb_sprite_funcs;
+		} else {
+			formats = g4x_sprite_formats;
+			num_formats = ARRAY_SIZE(g4x_sprite_formats);
+
+			plane_funcs = &g4x_sprite_funcs;
+		}
+	}
+
+	if (IS_CHERRYVIEW(dev_priv) && pipe == PIPE_B) {
+		supported_rotations =
+			DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180 |
+			DRM_MODE_REFLECT_X;
+	} else {
+		supported_rotations =
+			DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180;
+	}
+
+	plane->pipe = pipe;
+	plane->id = PLANE_SPRITE0 + sprite;
+	plane->frontbuffer_bit = INTEL_FRONTBUFFER(pipe, plane->id);
+
+	modifiers = intel_fb_plane_get_modifiers(dev_priv, INTEL_PLANE_CAP_TILING_X);
+
+	ret = drm_universal_plane_init(&dev_priv->drm, &plane->base,
+				       0, plane_funcs,
+				       formats, num_formats, modifiers,
+				       DRM_PLANE_TYPE_OVERLAY,
+				       "sprite %c", sprite_name(pipe, sprite));
+	kfree(modifiers);
+
+	if (ret)
+		goto fail;
+
+	drm_plane_create_rotation_property(&plane->base,
+					   DRM_MODE_ROTATE_0,
+					   supported_rotations);
+
+	drm_plane_create_color_properties(&plane->base,
+					  BIT(DRM_COLOR_YCBCR_BT601) |
+					  BIT(DRM_COLOR_YCBCR_BT709),
+					  BIT(DRM_COLOR_YCBCR_LIMITED_RANGE) |
+					  BIT(DRM_COLOR_YCBCR_FULL_RANGE),
+					  DRM_COLOR_YCBCR_BT709,
+					  DRM_COLOR_YCBCR_LIMITED_RANGE);
+
+	zpos = sprite + 1;
+	drm_plane_create_zpos_immutable_property(&plane->base, zpos);
+
+	intel_plane_helper_add(plane);
+
+	return plane;
+
+fail:
+	intel_plane_free(plane);
+
+	return ERR_PTR(ret);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_sprite.h b/drivers/gpu/drm/i915/display/intel_sprite.h
new file mode 100644
index 000000000..4f63e4967
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_sprite.h
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_SPRITE_H__
+#define __INTEL_SPRITE_H__
+
+#include <linux/types.h>
+
+#include "intel_display.h"
+
+struct drm_device;
+struct drm_display_mode;
+struct drm_file;
+struct drm_i915_private;
+struct intel_crtc_state;
+struct intel_plane_state;
+
+/*
+ * FIXME: We should instead only take spinlocks once for the entire update
+ * instead of once per mmio.
+ */
+#if IS_ENABLED(CONFIG_PROVE_LOCKING)
+#define VBLANK_EVASION_TIME_US 250
+#else
+#define VBLANK_EVASION_TIME_US 100
+#endif
+
+struct intel_plane *intel_sprite_plane_create(struct drm_i915_private *dev_priv,
+					      enum pipe pipe, int plane);
+int intel_sprite_set_colorkey_ioctl(struct drm_device *dev, void *data,
+				    struct drm_file *file_priv);
+int intel_plane_check_src_coordinates(struct intel_plane_state *plane_state);
+int chv_plane_check_rotation(const struct intel_plane_state *plane_state);
+
+static inline u8 icl_hdr_plane_mask(void)
+{
+	return BIT(PLANE_PRIMARY) |
+		BIT(PLANE_SPRITE0) | BIT(PLANE_SPRITE1);
+}
+
+int ivb_plane_min_cdclk(const struct intel_crtc_state *crtc_state,
+			const struct intel_plane_state *plane_state);
+int hsw_plane_min_cdclk(const struct intel_crtc_state *crtc_state,
+			const struct intel_plane_state *plane_state);
+int vlv_plane_min_cdclk(const struct intel_crtc_state *crtc_state,
+			const struct intel_plane_state *plane_state);
+
+#endif /* __INTEL_SPRITE_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_tc.c b/drivers/gpu/drm/i915/display/intel_tc.c
new file mode 100644
index 000000000..bda77828d
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_tc.c
@@ -0,0 +1,1007 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_de.h"
+#include "intel_display.h"
+#include "intel_display_power_map.h"
+#include "intel_display_types.h"
+#include "intel_dp_mst.h"
+#include "intel_tc.h"
+#include "intel_tc_phy_regs.h"
+
+static const char *tc_port_mode_name(enum tc_port_mode mode)
+{
+	static const char * const names[] = {
+		[TC_PORT_DISCONNECTED] = "disconnected",
+		[TC_PORT_TBT_ALT] = "tbt-alt",
+		[TC_PORT_DP_ALT] = "dp-alt",
+		[TC_PORT_LEGACY] = "legacy",
+	};
+
+	if (WARN_ON(mode >= ARRAY_SIZE(names)))
+		mode = TC_PORT_DISCONNECTED;
+
+	return names[mode];
+}
+
+static bool intel_tc_port_in_mode(struct intel_digital_port *dig_port,
+				  enum tc_port_mode mode)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	enum phy phy = intel_port_to_phy(i915, dig_port->base.port);
+
+	return intel_phy_is_tc(i915, phy) && dig_port->tc_mode == mode;
+}
+
+bool intel_tc_port_in_tbt_alt_mode(struct intel_digital_port *dig_port)
+{
+	return intel_tc_port_in_mode(dig_port, TC_PORT_TBT_ALT);
+}
+
+bool intel_tc_port_in_dp_alt_mode(struct intel_digital_port *dig_port)
+{
+	return intel_tc_port_in_mode(dig_port, TC_PORT_DP_ALT);
+}
+
+bool intel_tc_port_in_legacy_mode(struct intel_digital_port *dig_port)
+{
+	return intel_tc_port_in_mode(dig_port, TC_PORT_LEGACY);
+}
+
+bool intel_tc_cold_requires_aux_pw(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	return (DISPLAY_VER(i915) == 11 && dig_port->tc_legacy_port) ||
+		IS_ALDERLAKE_P(i915);
+}
+
+static enum intel_display_power_domain
+tc_cold_get_power_domain(struct intel_digital_port *dig_port, enum tc_port_mode mode)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	if (mode == TC_PORT_TBT_ALT || !intel_tc_cold_requires_aux_pw(dig_port))
+		return POWER_DOMAIN_TC_COLD_OFF;
+
+	return intel_display_power_legacy_aux_domain(i915, dig_port->aux_ch);
+}
+
+static intel_wakeref_t
+tc_cold_block_in_mode(struct intel_digital_port *dig_port, enum tc_port_mode mode,
+		      enum intel_display_power_domain *domain)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	*domain = tc_cold_get_power_domain(dig_port, mode);
+
+	return intel_display_power_get(i915, *domain);
+}
+
+static intel_wakeref_t
+tc_cold_block(struct intel_digital_port *dig_port, enum intel_display_power_domain *domain)
+{
+	return tc_cold_block_in_mode(dig_port, dig_port->tc_mode, domain);
+}
+
+static void
+tc_cold_unblock(struct intel_digital_port *dig_port, enum intel_display_power_domain domain,
+		intel_wakeref_t wakeref)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	/*
+	 * wakeref == -1, means some error happened saving save_depot_stack but
+	 * power should still be put down and 0 is a invalid save_depot_stack
+	 * id so can be used to skip it for non TC legacy ports.
+	 */
+	if (wakeref == 0)
+		return;
+
+	intel_display_power_put(i915, domain, wakeref);
+}
+
+static void
+assert_tc_cold_blocked(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	bool enabled;
+
+	enabled = intel_display_power_is_enabled(i915,
+						 tc_cold_get_power_domain(dig_port,
+									  dig_port->tc_mode));
+	drm_WARN_ON(&i915->drm, !enabled);
+}
+
+static enum intel_display_power_domain
+tc_port_power_domain(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	enum tc_port tc_port = intel_port_to_tc(i915, dig_port->base.port);
+
+	return POWER_DOMAIN_PORT_DDI_LANES_TC1 + tc_port - TC_PORT_1;
+}
+
+static void
+assert_tc_port_power_enabled(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	drm_WARN_ON(&i915->drm,
+		    !intel_display_power_is_enabled(i915, tc_port_power_domain(dig_port)));
+}
+
+u32 intel_tc_port_get_lane_mask(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 lane_mask;
+
+	lane_mask = intel_uncore_read(uncore,
+				      PORT_TX_DFLEXDPSP(dig_port->tc_phy_fia));
+
+	drm_WARN_ON(&i915->drm, lane_mask == 0xffffffff);
+	assert_tc_cold_blocked(dig_port);
+
+	lane_mask &= DP_LANE_ASSIGNMENT_MASK(dig_port->tc_phy_fia_idx);
+	return lane_mask >> DP_LANE_ASSIGNMENT_SHIFT(dig_port->tc_phy_fia_idx);
+}
+
+u32 intel_tc_port_get_pin_assignment_mask(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 pin_mask;
+
+	pin_mask = intel_uncore_read(uncore,
+				     PORT_TX_DFLEXPA1(dig_port->tc_phy_fia));
+
+	drm_WARN_ON(&i915->drm, pin_mask == 0xffffffff);
+	assert_tc_cold_blocked(dig_port);
+
+	return (pin_mask & DP_PIN_ASSIGNMENT_MASK(dig_port->tc_phy_fia_idx)) >>
+	       DP_PIN_ASSIGNMENT_SHIFT(dig_port->tc_phy_fia_idx);
+}
+
+int intel_tc_port_fia_max_lane_count(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	intel_wakeref_t wakeref;
+	u32 lane_mask;
+
+	if (dig_port->tc_mode != TC_PORT_DP_ALT)
+		return 4;
+
+	assert_tc_cold_blocked(dig_port);
+
+	lane_mask = 0;
+	with_intel_display_power(i915, POWER_DOMAIN_DISPLAY_CORE, wakeref)
+		lane_mask = intel_tc_port_get_lane_mask(dig_port);
+
+	switch (lane_mask) {
+	default:
+		MISSING_CASE(lane_mask);
+		fallthrough;
+	case 0x1:
+	case 0x2:
+	case 0x4:
+	case 0x8:
+		return 1;
+	case 0x3:
+	case 0xc:
+		return 2;
+	case 0xf:
+		return 4;
+	}
+}
+
+void intel_tc_port_set_fia_lane_count(struct intel_digital_port *dig_port,
+				      int required_lanes)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	bool lane_reversal = dig_port->saved_port_bits & DDI_BUF_PORT_REVERSAL;
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 val;
+
+	drm_WARN_ON(&i915->drm,
+		    lane_reversal && dig_port->tc_mode != TC_PORT_LEGACY);
+
+	assert_tc_cold_blocked(dig_port);
+
+	val = intel_uncore_read(uncore,
+				PORT_TX_DFLEXDPMLE1(dig_port->tc_phy_fia));
+	val &= ~DFLEXDPMLE1_DPMLETC_MASK(dig_port->tc_phy_fia_idx);
+
+	switch (required_lanes) {
+	case 1:
+		val |= lane_reversal ?
+			DFLEXDPMLE1_DPMLETC_ML3(dig_port->tc_phy_fia_idx) :
+			DFLEXDPMLE1_DPMLETC_ML0(dig_port->tc_phy_fia_idx);
+		break;
+	case 2:
+		val |= lane_reversal ?
+			DFLEXDPMLE1_DPMLETC_ML3_2(dig_port->tc_phy_fia_idx) :
+			DFLEXDPMLE1_DPMLETC_ML1_0(dig_port->tc_phy_fia_idx);
+		break;
+	case 4:
+		val |= DFLEXDPMLE1_DPMLETC_ML3_0(dig_port->tc_phy_fia_idx);
+		break;
+	default:
+		MISSING_CASE(required_lanes);
+	}
+
+	intel_uncore_write(uncore,
+			   PORT_TX_DFLEXDPMLE1(dig_port->tc_phy_fia), val);
+}
+
+static void tc_port_fixup_legacy_flag(struct intel_digital_port *dig_port,
+				      u32 live_status_mask)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	u32 valid_hpd_mask;
+
+	if (dig_port->tc_legacy_port)
+		valid_hpd_mask = BIT(TC_PORT_LEGACY);
+	else
+		valid_hpd_mask = BIT(TC_PORT_DP_ALT) |
+				 BIT(TC_PORT_TBT_ALT);
+
+	if (!(live_status_mask & ~valid_hpd_mask))
+		return;
+
+	/* If live status mismatches the VBT flag, trust the live status. */
+	drm_dbg_kms(&i915->drm,
+		    "Port %s: live status %08x mismatches the legacy port flag %08x, fixing flag\n",
+		    dig_port->tc_port_name, live_status_mask, valid_hpd_mask);
+
+	dig_port->tc_legacy_port = !dig_port->tc_legacy_port;
+}
+
+static u32 icl_tc_port_live_status_mask(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 isr_bit = i915->display.hotplug.pch_hpd[dig_port->base.hpd_pin];
+	u32 mask = 0;
+	u32 val;
+
+	val = intel_uncore_read(uncore,
+				PORT_TX_DFLEXDPSP(dig_port->tc_phy_fia));
+
+	if (val == 0xffffffff) {
+		drm_dbg_kms(&i915->drm,
+			    "Port %s: PHY in TCCOLD, nothing connected\n",
+			    dig_port->tc_port_name);
+		return mask;
+	}
+
+	if (val & TC_LIVE_STATE_TBT(dig_port->tc_phy_fia_idx))
+		mask |= BIT(TC_PORT_TBT_ALT);
+	if (val & TC_LIVE_STATE_TC(dig_port->tc_phy_fia_idx))
+		mask |= BIT(TC_PORT_DP_ALT);
+
+	if (intel_uncore_read(uncore, SDEISR) & isr_bit)
+		mask |= BIT(TC_PORT_LEGACY);
+
+	/* The sink can be connected only in a single mode. */
+	if (!drm_WARN_ON_ONCE(&i915->drm, hweight32(mask) > 1))
+		tc_port_fixup_legacy_flag(dig_port, mask);
+
+	return mask;
+}
+
+static u32 adl_tc_port_live_status_mask(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	enum tc_port tc_port = intel_port_to_tc(i915, dig_port->base.port);
+	u32 isr_bit = i915->display.hotplug.pch_hpd[dig_port->base.hpd_pin];
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 val, mask = 0;
+
+	/*
+	 * On ADL-P HW/FW will wake from TCCOLD to complete the read access of
+	 * registers in IOM. Note that this doesn't apply to PHY and FIA
+	 * registers.
+	 */
+	val = intel_uncore_read(uncore, TCSS_DDI_STATUS(tc_port));
+	if (val & TCSS_DDI_STATUS_HPD_LIVE_STATUS_ALT)
+		mask |= BIT(TC_PORT_DP_ALT);
+	if (val & TCSS_DDI_STATUS_HPD_LIVE_STATUS_TBT)
+		mask |= BIT(TC_PORT_TBT_ALT);
+
+	if (intel_uncore_read(uncore, SDEISR) & isr_bit)
+		mask |= BIT(TC_PORT_LEGACY);
+
+	/* The sink can be connected only in a single mode. */
+	if (!drm_WARN_ON(&i915->drm, hweight32(mask) > 1))
+		tc_port_fixup_legacy_flag(dig_port, mask);
+
+	return mask;
+}
+
+static u32 tc_port_live_status_mask(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	if (IS_ALDERLAKE_P(i915))
+		return adl_tc_port_live_status_mask(dig_port);
+
+	return icl_tc_port_live_status_mask(dig_port);
+}
+
+/*
+ * Return the PHY status complete flag indicating that display can acquire the
+ * PHY ownership. The IOM firmware sets this flag when a DP-alt or legacy sink
+ * is connected and it's ready to switch the ownership to display. The flag
+ * will be left cleared when a TBT-alt sink is connected, where the PHY is
+ * owned by the TBT subsystem and so switching the ownership to display is not
+ * required.
+ */
+static bool icl_tc_phy_status_complete(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 val;
+
+	val = intel_uncore_read(uncore,
+				PORT_TX_DFLEXDPPMS(dig_port->tc_phy_fia));
+	if (val == 0xffffffff) {
+		drm_dbg_kms(&i915->drm,
+			    "Port %s: PHY in TCCOLD, assuming not complete\n",
+			    dig_port->tc_port_name);
+		return false;
+	}
+
+	return val & DP_PHY_MODE_STATUS_COMPLETED(dig_port->tc_phy_fia_idx);
+}
+
+/*
+ * Return the PHY status complete flag indicating that display can acquire the
+ * PHY ownership. The IOM firmware sets this flag when it's ready to switch
+ * the ownership to display, regardless of what sink is connected (TBT-alt,
+ * DP-alt, legacy or nothing). For TBT-alt sinks the PHY is owned by the TBT
+ * subsystem and so switching the ownership to display is not required.
+ */
+static bool adl_tc_phy_status_complete(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	enum tc_port tc_port = intel_port_to_tc(i915, dig_port->base.port);
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 val;
+
+	val = intel_uncore_read(uncore, TCSS_DDI_STATUS(tc_port));
+	if (val == 0xffffffff) {
+		drm_dbg_kms(&i915->drm,
+			    "Port %s: PHY in TCCOLD, assuming not complete\n",
+			    dig_port->tc_port_name);
+		return false;
+	}
+
+	return val & TCSS_DDI_STATUS_READY;
+}
+
+static bool tc_phy_status_complete(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	if (IS_ALDERLAKE_P(i915))
+		return adl_tc_phy_status_complete(dig_port);
+
+	return icl_tc_phy_status_complete(dig_port);
+}
+
+static bool icl_tc_phy_take_ownership(struct intel_digital_port *dig_port,
+				      bool take)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 val;
+
+	val = intel_uncore_read(uncore,
+				PORT_TX_DFLEXDPCSSS(dig_port->tc_phy_fia));
+	if (val == 0xffffffff) {
+		drm_dbg_kms(&i915->drm,
+			    "Port %s: PHY in TCCOLD, can't %s ownership\n",
+			    dig_port->tc_port_name, take ? "take" : "release");
+
+		return false;
+	}
+
+	val &= ~DP_PHY_MODE_STATUS_NOT_SAFE(dig_port->tc_phy_fia_idx);
+	if (take)
+		val |= DP_PHY_MODE_STATUS_NOT_SAFE(dig_port->tc_phy_fia_idx);
+
+	intel_uncore_write(uncore,
+			   PORT_TX_DFLEXDPCSSS(dig_port->tc_phy_fia), val);
+
+	return true;
+}
+
+static bool adl_tc_phy_take_ownership(struct intel_digital_port *dig_port,
+				      bool take)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	enum port port = dig_port->base.port;
+	u32 val;
+
+	val = intel_uncore_read(uncore, DDI_BUF_CTL(port));
+	if (take)
+		val |= DDI_BUF_CTL_TC_PHY_OWNERSHIP;
+	else
+		val &= ~DDI_BUF_CTL_TC_PHY_OWNERSHIP;
+	intel_uncore_write(uncore, DDI_BUF_CTL(port), val);
+
+	return true;
+}
+
+static bool tc_phy_take_ownership(struct intel_digital_port *dig_port, bool take)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	if (IS_ALDERLAKE_P(i915))
+		return adl_tc_phy_take_ownership(dig_port, take);
+
+	return icl_tc_phy_take_ownership(dig_port, take);
+}
+
+static bool icl_tc_phy_is_owned(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 val;
+
+	val = intel_uncore_read(uncore,
+				PORT_TX_DFLEXDPCSSS(dig_port->tc_phy_fia));
+	if (val == 0xffffffff) {
+		drm_dbg_kms(&i915->drm,
+			    "Port %s: PHY in TCCOLD, assume not owned\n",
+			    dig_port->tc_port_name);
+		return false;
+	}
+
+	return val & DP_PHY_MODE_STATUS_NOT_SAFE(dig_port->tc_phy_fia_idx);
+}
+
+static bool adl_tc_phy_is_owned(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	enum port port = dig_port->base.port;
+	u32 val;
+
+	val = intel_uncore_read(uncore, DDI_BUF_CTL(port));
+	return val & DDI_BUF_CTL_TC_PHY_OWNERSHIP;
+}
+
+static bool tc_phy_is_owned(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	if (IS_ALDERLAKE_P(i915))
+		return adl_tc_phy_is_owned(dig_port);
+
+	return icl_tc_phy_is_owned(dig_port);
+}
+
+/*
+ * This function implements the first part of the Connect Flow described by our
+ * specification, Gen11 TypeC Programming chapter. The rest of the flow (reading
+ * lanes, EDID, etc) is done as needed in the typical places.
+ *
+ * Unlike the other ports, type-C ports are not available to use as soon as we
+ * get a hotplug. The type-C PHYs can be shared between multiple controllers:
+ * display, USB, etc. As a result, handshaking through FIA is required around
+ * connect and disconnect to cleanly transfer ownership with the controller and
+ * set the type-C power state.
+ */
+static void icl_tc_phy_connect(struct intel_digital_port *dig_port,
+			       int required_lanes)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	u32 live_status_mask;
+	int max_lanes;
+
+	if (!tc_phy_status_complete(dig_port)) {
+		drm_dbg_kms(&i915->drm, "Port %s: PHY not ready\n",
+			    dig_port->tc_port_name);
+		goto out_set_tbt_alt_mode;
+	}
+
+	live_status_mask = tc_port_live_status_mask(dig_port);
+	if (!(live_status_mask & (BIT(TC_PORT_DP_ALT) | BIT(TC_PORT_LEGACY))) &&
+	    !dig_port->tc_legacy_port) {
+		drm_dbg_kms(&i915->drm, "Port %s: PHY ownership not required (live status %02x)\n",
+			    dig_port->tc_port_name, live_status_mask);
+		goto out_set_tbt_alt_mode;
+	}
+
+	if (!tc_phy_take_ownership(dig_port, true) &&
+	    !drm_WARN_ON(&i915->drm, dig_port->tc_legacy_port))
+		goto out_set_tbt_alt_mode;
+
+	max_lanes = intel_tc_port_fia_max_lane_count(dig_port);
+	if (dig_port->tc_legacy_port) {
+		drm_WARN_ON(&i915->drm, max_lanes != 4);
+		dig_port->tc_mode = TC_PORT_LEGACY;
+
+		return;
+	}
+
+	/*
+	 * Now we have to re-check the live state, in case the port recently
+	 * became disconnected. Not necessary for legacy mode.
+	 */
+	if (!(tc_port_live_status_mask(dig_port) & BIT(TC_PORT_DP_ALT))) {
+		drm_dbg_kms(&i915->drm, "Port %s: PHY sudden disconnect\n",
+			    dig_port->tc_port_name);
+		goto out_release_phy;
+	}
+
+	if (max_lanes < required_lanes) {
+		drm_dbg_kms(&i915->drm,
+			    "Port %s: PHY max lanes %d < required lanes %d\n",
+			    dig_port->tc_port_name,
+			    max_lanes, required_lanes);
+		goto out_release_phy;
+	}
+
+	dig_port->tc_mode = TC_PORT_DP_ALT;
+
+	return;
+
+out_release_phy:
+	tc_phy_take_ownership(dig_port, false);
+out_set_tbt_alt_mode:
+	dig_port->tc_mode = TC_PORT_TBT_ALT;
+}
+
+/*
+ * See the comment at the connect function. This implements the Disconnect
+ * Flow.
+ */
+static void icl_tc_phy_disconnect(struct intel_digital_port *dig_port)
+{
+	switch (dig_port->tc_mode) {
+	case TC_PORT_LEGACY:
+	case TC_PORT_DP_ALT:
+		tc_phy_take_ownership(dig_port, false);
+		fallthrough;
+	case TC_PORT_TBT_ALT:
+		dig_port->tc_mode = TC_PORT_DISCONNECTED;
+		fallthrough;
+	case TC_PORT_DISCONNECTED:
+		break;
+	default:
+		MISSING_CASE(dig_port->tc_mode);
+	}
+}
+
+static bool icl_tc_phy_is_connected(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	if (!tc_phy_status_complete(dig_port)) {
+		drm_dbg_kms(&i915->drm, "Port %s: PHY status not complete\n",
+			    dig_port->tc_port_name);
+		return dig_port->tc_mode == TC_PORT_TBT_ALT;
+	}
+
+	/* On ADL-P the PHY complete flag is set in TBT mode as well. */
+	if (IS_ALDERLAKE_P(i915) && dig_port->tc_mode == TC_PORT_TBT_ALT)
+		return true;
+
+	if (!tc_phy_is_owned(dig_port)) {
+		drm_dbg_kms(&i915->drm, "Port %s: PHY not owned\n",
+			    dig_port->tc_port_name);
+
+		return false;
+	}
+
+	return dig_port->tc_mode == TC_PORT_DP_ALT ||
+	       dig_port->tc_mode == TC_PORT_LEGACY;
+}
+
+static enum tc_port_mode
+intel_tc_port_get_current_mode(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	u32 live_status_mask = tc_port_live_status_mask(dig_port);
+	enum tc_port_mode mode;
+
+	if (!tc_phy_is_owned(dig_port) ||
+	    drm_WARN_ON(&i915->drm, !tc_phy_status_complete(dig_port)))
+		return TC_PORT_TBT_ALT;
+
+	mode = dig_port->tc_legacy_port ? TC_PORT_LEGACY : TC_PORT_DP_ALT;
+	if (live_status_mask) {
+		enum tc_port_mode live_mode = fls(live_status_mask) - 1;
+
+		if (!drm_WARN_ON(&i915->drm, live_mode == TC_PORT_TBT_ALT))
+			mode = live_mode;
+	}
+
+	return mode;
+}
+
+static enum tc_port_mode
+intel_tc_port_get_target_mode(struct intel_digital_port *dig_port)
+{
+	u32 live_status_mask = tc_port_live_status_mask(dig_port);
+
+	if (live_status_mask)
+		return fls(live_status_mask) - 1;
+
+	return TC_PORT_TBT_ALT;
+}
+
+static void intel_tc_port_reset_mode(struct intel_digital_port *dig_port,
+				     int required_lanes, bool force_disconnect)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	enum tc_port_mode old_tc_mode = dig_port->tc_mode;
+
+	intel_display_power_flush_work(i915);
+	if (!intel_tc_cold_requires_aux_pw(dig_port)) {
+		enum intel_display_power_domain aux_domain;
+		bool aux_powered;
+
+		aux_domain = intel_aux_power_domain(dig_port);
+		aux_powered = intel_display_power_is_enabled(i915, aux_domain);
+		drm_WARN_ON(&i915->drm, aux_powered);
+	}
+
+	icl_tc_phy_disconnect(dig_port);
+	if (!force_disconnect)
+		icl_tc_phy_connect(dig_port, required_lanes);
+
+	drm_dbg_kms(&i915->drm, "Port %s: TC port mode reset (%s -> %s)\n",
+		    dig_port->tc_port_name,
+		    tc_port_mode_name(old_tc_mode),
+		    tc_port_mode_name(dig_port->tc_mode));
+}
+
+static bool intel_tc_port_needs_reset(struct intel_digital_port *dig_port)
+{
+	return intel_tc_port_get_target_mode(dig_port) != dig_port->tc_mode;
+}
+
+static void intel_tc_port_update_mode(struct intel_digital_port *dig_port,
+				      int required_lanes, bool force_disconnect)
+{
+	enum intel_display_power_domain domain;
+	intel_wakeref_t wref;
+	bool needs_reset = force_disconnect;
+
+	if (!needs_reset) {
+		/* Get power domain required to check the hotplug live status. */
+		wref = tc_cold_block(dig_port, &domain);
+		needs_reset = intel_tc_port_needs_reset(dig_port);
+		tc_cold_unblock(dig_port, domain, wref);
+	}
+
+	if (!needs_reset)
+		return;
+
+	/* Get power domain required for resetting the mode. */
+	wref = tc_cold_block_in_mode(dig_port, TC_PORT_DISCONNECTED, &domain);
+
+	intel_tc_port_reset_mode(dig_port, required_lanes, force_disconnect);
+
+	/* Get power domain matching the new mode after reset. */
+	tc_cold_unblock(dig_port, dig_port->tc_lock_power_domain,
+			fetch_and_zero(&dig_port->tc_lock_wakeref));
+	if (dig_port->tc_mode != TC_PORT_DISCONNECTED)
+		dig_port->tc_lock_wakeref = tc_cold_block(dig_port,
+							  &dig_port->tc_lock_power_domain);
+
+	tc_cold_unblock(dig_port, domain, wref);
+}
+
+static void __intel_tc_port_get_link(struct intel_digital_port *dig_port)
+{
+	dig_port->tc_link_refcount++;
+}
+
+static void __intel_tc_port_put_link(struct intel_digital_port *dig_port)
+{
+	dig_port->tc_link_refcount--;
+}
+
+static bool tc_port_is_enabled(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	assert_tc_port_power_enabled(dig_port);
+
+	return intel_de_read(i915, DDI_BUF_CTL(dig_port->base.port)) &
+	       DDI_BUF_CTL_ENABLE;
+}
+
+/**
+ * intel_tc_port_init_mode: Read out HW state and init the given port's TypeC mode
+ * @dig_port: digital port
+ *
+ * Read out the HW state and initialize the TypeC mode of @dig_port. The mode
+ * will be locked until intel_tc_port_sanitize_mode() is called.
+ */
+void intel_tc_port_init_mode(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	intel_wakeref_t tc_cold_wref;
+	enum intel_display_power_domain domain;
+
+	mutex_lock(&dig_port->tc_lock);
+
+	drm_WARN_ON(&i915->drm, dig_port->tc_mode != TC_PORT_DISCONNECTED);
+	drm_WARN_ON(&i915->drm, dig_port->tc_lock_wakeref);
+	drm_WARN_ON(&i915->drm, dig_port->tc_link_refcount);
+
+	tc_cold_wref = tc_cold_block(dig_port, &domain);
+
+	dig_port->tc_mode = intel_tc_port_get_current_mode(dig_port);
+	/*
+	 * Save the initial mode for the state check in
+	 * intel_tc_port_sanitize_mode().
+	 */
+	dig_port->tc_init_mode = dig_port->tc_mode;
+	dig_port->tc_lock_wakeref = tc_cold_block(dig_port, &dig_port->tc_lock_power_domain);
+
+	/*
+	 * The PHY needs to be connected for AUX to work during HW readout and
+	 * MST topology resume, but the PHY mode can only be changed if the
+	 * port is disabled.
+	 */
+	if (!tc_port_is_enabled(dig_port))
+		intel_tc_port_update_mode(dig_port, 1, false);
+
+	/* Prevent changing dig_port->tc_mode until intel_tc_port_sanitize_mode() is called. */
+	__intel_tc_port_get_link(dig_port);
+
+	tc_cold_unblock(dig_port, domain, tc_cold_wref);
+
+	drm_dbg_kms(&i915->drm, "Port %s: init mode (%s)\n",
+		    dig_port->tc_port_name,
+		    tc_port_mode_name(dig_port->tc_mode));
+
+	mutex_unlock(&dig_port->tc_lock);
+}
+
+/**
+ * intel_tc_port_sanitize_mode: Sanitize the given port's TypeC mode
+ * @dig_port: digital port
+ *
+ * Sanitize @dig_port's TypeC mode wrt. the encoder's state right after driver
+ * loading and system resume:
+ * If the encoder is enabled keep the TypeC mode/PHY connected state locked until
+ * the encoder is disabled.
+ * If the encoder is disabled make sure the PHY is disconnected.
+ */
+void intel_tc_port_sanitize_mode(struct intel_digital_port *dig_port)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	struct intel_encoder *encoder = &dig_port->base;
+	int active_links = 0;
+
+	mutex_lock(&dig_port->tc_lock);
+
+	if (dig_port->dp.is_mst)
+		active_links = intel_dp_mst_encoder_active_links(dig_port);
+	else if (encoder->base.crtc)
+		active_links = to_intel_crtc(encoder->base.crtc)->active;
+
+	drm_WARN_ON(&i915->drm, dig_port->tc_link_refcount != 1);
+	if (active_links) {
+		if (!icl_tc_phy_is_connected(dig_port))
+			drm_dbg_kms(&i915->drm,
+				    "Port %s: PHY disconnected with %d active link(s)\n",
+				    dig_port->tc_port_name, active_links);
+	} else {
+		/*
+		 * TBT-alt is the default mode in any case the PHY ownership is not
+		 * held (regardless of the sink's connected live state), so
+		 * we'll just switch to disconnected mode from it here without
+		 * a note.
+		 */
+		if (dig_port->tc_init_mode != TC_PORT_TBT_ALT)
+			drm_dbg_kms(&i915->drm,
+				    "Port %s: PHY left in %s mode on disabled port, disconnecting it\n",
+				    dig_port->tc_port_name,
+				    tc_port_mode_name(dig_port->tc_init_mode));
+		icl_tc_phy_disconnect(dig_port);
+		__intel_tc_port_put_link(dig_port);
+
+		tc_cold_unblock(dig_port, dig_port->tc_lock_power_domain,
+				fetch_and_zero(&dig_port->tc_lock_wakeref));
+	}
+
+	drm_dbg_kms(&i915->drm, "Port %s: sanitize mode (%s)\n",
+		    dig_port->tc_port_name,
+		    tc_port_mode_name(dig_port->tc_mode));
+
+	mutex_unlock(&dig_port->tc_lock);
+}
+
+/*
+ * The type-C ports are different because even when they are connected, they may
+ * not be available/usable by the graphics driver: see the comment on
+ * icl_tc_phy_connect(). So in our driver instead of adding the additional
+ * concept of "usable" and make everything check for "connected and usable" we
+ * define a port as "connected" when it is not only connected, but also when it
+ * is usable by the rest of the driver. That maintains the old assumption that
+ * connected ports are usable, and avoids exposing to the users objects they
+ * can't really use.
+ */
+bool intel_tc_port_connected(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	bool is_connected;
+
+	intel_tc_port_lock(dig_port);
+
+	is_connected = tc_port_live_status_mask(dig_port) &
+		       BIT(dig_port->tc_mode);
+
+	intel_tc_port_unlock(dig_port);
+
+	return is_connected;
+}
+
+static void __intel_tc_port_lock(struct intel_digital_port *dig_port,
+				 int required_lanes)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+
+	mutex_lock(&dig_port->tc_lock);
+
+	cancel_delayed_work(&dig_port->tc_disconnect_phy_work);
+
+	if (!dig_port->tc_link_refcount)
+		intel_tc_port_update_mode(dig_port, required_lanes,
+					  false);
+
+	drm_WARN_ON(&i915->drm, dig_port->tc_mode == TC_PORT_DISCONNECTED);
+	drm_WARN_ON(&i915->drm, dig_port->tc_mode != TC_PORT_TBT_ALT &&
+				!tc_phy_is_owned(dig_port));
+}
+
+void intel_tc_port_lock(struct intel_digital_port *dig_port)
+{
+	__intel_tc_port_lock(dig_port, 1);
+}
+
+/**
+ * intel_tc_port_disconnect_phy_work: disconnect TypeC PHY from display port
+ * @dig_port: digital port
+ *
+ * Disconnect the given digital port from its TypeC PHY (handing back the
+ * control of the PHY to the TypeC subsystem). This will happen in a delayed
+ * manner after each aux transactions and modeset disables.
+ */
+static void intel_tc_port_disconnect_phy_work(struct work_struct *work)
+{
+	struct intel_digital_port *dig_port =
+		container_of(work, struct intel_digital_port, tc_disconnect_phy_work.work);
+
+	mutex_lock(&dig_port->tc_lock);
+
+	if (!dig_port->tc_link_refcount)
+		intel_tc_port_update_mode(dig_port, 1, true);
+
+	mutex_unlock(&dig_port->tc_lock);
+}
+
+/**
+ * intel_tc_port_flush_work: flush the work disconnecting the PHY
+ * @dig_port: digital port
+ *
+ * Flush the delayed work disconnecting an idle PHY.
+ */
+void intel_tc_port_flush_work(struct intel_digital_port *dig_port)
+{
+	flush_delayed_work(&dig_port->tc_disconnect_phy_work);
+}
+
+void intel_tc_port_unlock(struct intel_digital_port *dig_port)
+{
+	if (!dig_port->tc_link_refcount && dig_port->tc_mode != TC_PORT_DISCONNECTED)
+		queue_delayed_work(system_unbound_wq, &dig_port->tc_disconnect_phy_work,
+				   msecs_to_jiffies(1000));
+
+	mutex_unlock(&dig_port->tc_lock);
+}
+
+bool intel_tc_port_ref_held(struct intel_digital_port *dig_port)
+{
+	return mutex_is_locked(&dig_port->tc_lock) ||
+	       dig_port->tc_link_refcount;
+}
+
+void intel_tc_port_get_link(struct intel_digital_port *dig_port,
+			    int required_lanes)
+{
+	__intel_tc_port_lock(dig_port, required_lanes);
+	__intel_tc_port_get_link(dig_port);
+	intel_tc_port_unlock(dig_port);
+}
+
+void intel_tc_port_put_link(struct intel_digital_port *dig_port)
+{
+	intel_tc_port_lock(dig_port);
+	__intel_tc_port_put_link(dig_port);
+	intel_tc_port_unlock(dig_port);
+
+	/*
+	 * Disconnecting the PHY after the PHY's PLL gets disabled may
+	 * hang the system on ADL-P, so disconnect the PHY here synchronously.
+	 * TODO: remove this once the root cause of the ordering requirement
+	 * is found/fixed.
+	 */
+	intel_tc_port_flush_work(dig_port);
+}
+
+static bool
+tc_has_modular_fia(struct drm_i915_private *i915, struct intel_digital_port *dig_port)
+{
+	enum intel_display_power_domain domain;
+	intel_wakeref_t wakeref;
+	u32 val;
+
+	if (!INTEL_INFO(i915)->display.has_modular_fia)
+		return false;
+
+	mutex_lock(&dig_port->tc_lock);
+	wakeref = tc_cold_block(dig_port, &domain);
+	val = intel_uncore_read(&i915->uncore, PORT_TX_DFLEXDPSP(FIA1));
+	tc_cold_unblock(dig_port, domain, wakeref);
+	mutex_unlock(&dig_port->tc_lock);
+
+	drm_WARN_ON(&i915->drm, val == 0xffffffff);
+
+	return val & MODULAR_FIA_MASK;
+}
+
+static void
+tc_port_load_fia_params(struct drm_i915_private *i915, struct intel_digital_port *dig_port)
+{
+	enum port port = dig_port->base.port;
+	enum tc_port tc_port = intel_port_to_tc(i915, port);
+
+	/*
+	 * Each Modular FIA instance houses 2 TC ports. In SOC that has more
+	 * than two TC ports, there are multiple instances of Modular FIA.
+	 */
+	if (tc_has_modular_fia(i915, dig_port)) {
+		dig_port->tc_phy_fia = tc_port / 2;
+		dig_port->tc_phy_fia_idx = tc_port % 2;
+	} else {
+		dig_port->tc_phy_fia = FIA1;
+		dig_port->tc_phy_fia_idx = tc_port;
+	}
+}
+
+void intel_tc_port_init(struct intel_digital_port *dig_port, bool is_legacy)
+{
+	struct drm_i915_private *i915 = to_i915(dig_port->base.base.dev);
+	enum port port = dig_port->base.port;
+	enum tc_port tc_port = intel_port_to_tc(i915, port);
+
+	if (drm_WARN_ON(&i915->drm, tc_port == TC_PORT_NONE))
+		return;
+
+	snprintf(dig_port->tc_port_name, sizeof(dig_port->tc_port_name),
+		 "%c/TC#%d", port_name(port), tc_port + 1);
+
+	mutex_init(&dig_port->tc_lock);
+	INIT_DELAYED_WORK(&dig_port->tc_disconnect_phy_work, intel_tc_port_disconnect_phy_work);
+	dig_port->tc_legacy_port = is_legacy;
+	dig_port->tc_mode = TC_PORT_DISCONNECTED;
+	dig_port->tc_link_refcount = 0;
+	tc_port_load_fia_params(i915, dig_port);
+
+	intel_tc_port_init_mode(dig_port);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_tc.h b/drivers/gpu/drm/i915/display/intel_tc.h
new file mode 100644
index 000000000..d54082e2d
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_tc.h
@@ -0,0 +1,41 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_TC_H__
+#define __INTEL_TC_H__
+
+#include <linux/mutex.h>
+#include <linux/types.h>
+
+struct intel_digital_port;
+struct intel_encoder;
+
+bool intel_tc_port_in_tbt_alt_mode(struct intel_digital_port *dig_port);
+bool intel_tc_port_in_dp_alt_mode(struct intel_digital_port *dig_port);
+bool intel_tc_port_in_legacy_mode(struct intel_digital_port *dig_port);
+
+bool intel_tc_port_connected(struct intel_encoder *encoder);
+
+u32 intel_tc_port_get_lane_mask(struct intel_digital_port *dig_port);
+u32 intel_tc_port_get_pin_assignment_mask(struct intel_digital_port *dig_port);
+int intel_tc_port_fia_max_lane_count(struct intel_digital_port *dig_port);
+void intel_tc_port_set_fia_lane_count(struct intel_digital_port *dig_port,
+				      int required_lanes);
+
+void intel_tc_port_init_mode(struct intel_digital_port *dig_port);
+void intel_tc_port_sanitize_mode(struct intel_digital_port *dig_port);
+void intel_tc_port_lock(struct intel_digital_port *dig_port);
+void intel_tc_port_unlock(struct intel_digital_port *dig_port);
+void intel_tc_port_flush_work(struct intel_digital_port *dig_port);
+void intel_tc_port_get_link(struct intel_digital_port *dig_port,
+			    int required_lanes);
+void intel_tc_port_put_link(struct intel_digital_port *dig_port);
+bool intel_tc_port_ref_held(struct intel_digital_port *dig_port);
+
+void intel_tc_port_init(struct intel_digital_port *dig_port, bool is_legacy);
+
+bool intel_tc_cold_requires_aux_pw(struct intel_digital_port *dig_port);
+
+#endif /* __INTEL_TC_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_tc_phy_regs.h b/drivers/gpu/drm/i915/display/intel_tc_phy_regs.h
new file mode 100644
index 000000000..5a545086f
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_tc_phy_regs.h
@@ -0,0 +1,280 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_TC_PHY_REGS__
+#define __INTEL_TC_PHY_REGS__
+
+#include "i915_reg_defs.h"
+
+#define MG_PHY_PORT_LN(ln, tc_port, ln0p1, ln0p2, ln1p1) \
+	_MMIO(_PORT(tc_port, ln0p1, ln0p2) + (ln) * ((ln1p1) - (ln0p1)))
+
+#define MG_TX_LINK_PARAMS_TX1LN0_PORT1		0x16812C
+#define MG_TX_LINK_PARAMS_TX1LN1_PORT1		0x16852C
+#define MG_TX_LINK_PARAMS_TX1LN0_PORT2		0x16912C
+#define MG_TX_LINK_PARAMS_TX1LN1_PORT2		0x16952C
+#define MG_TX1_LINK_PARAMS(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_TX_LINK_PARAMS_TX1LN0_PORT1, \
+		       MG_TX_LINK_PARAMS_TX1LN0_PORT2, \
+		       MG_TX_LINK_PARAMS_TX1LN1_PORT1)
+
+#define MG_TX_LINK_PARAMS_TX2LN0_PORT1		0x1680AC
+#define MG_TX_LINK_PARAMS_TX2LN1_PORT1		0x1684AC
+#define MG_TX_LINK_PARAMS_TX2LN0_PORT2		0x1690AC
+#define MG_TX_LINK_PARAMS_TX2LN1_PORT2		0x1694AC
+#define MG_TX2_LINK_PARAMS(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_TX_LINK_PARAMS_TX2LN0_PORT1, \
+		       MG_TX_LINK_PARAMS_TX2LN0_PORT2, \
+		       MG_TX_LINK_PARAMS_TX2LN1_PORT1)
+#define   CRI_USE_FS32			(1 << 5)
+
+#define MG_TX_PISO_READLOAD_TX1LN0_PORT1		0x16814C
+#define MG_TX_PISO_READLOAD_TX1LN1_PORT1		0x16854C
+#define MG_TX_PISO_READLOAD_TX1LN0_PORT2		0x16914C
+#define MG_TX_PISO_READLOAD_TX1LN1_PORT2		0x16954C
+#define MG_TX1_PISO_READLOAD(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_TX_PISO_READLOAD_TX1LN0_PORT1, \
+		       MG_TX_PISO_READLOAD_TX1LN0_PORT2, \
+		       MG_TX_PISO_READLOAD_TX1LN1_PORT1)
+
+#define MG_TX_PISO_READLOAD_TX2LN0_PORT1		0x1680CC
+#define MG_TX_PISO_READLOAD_TX2LN1_PORT1		0x1684CC
+#define MG_TX_PISO_READLOAD_TX2LN0_PORT2		0x1690CC
+#define MG_TX_PISO_READLOAD_TX2LN1_PORT2		0x1694CC
+#define MG_TX2_PISO_READLOAD(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_TX_PISO_READLOAD_TX2LN0_PORT1, \
+		       MG_TX_PISO_READLOAD_TX2LN0_PORT2, \
+		       MG_TX_PISO_READLOAD_TX2LN1_PORT1)
+#define   CRI_CALCINIT					(1 << 1)
+
+#define MG_TX_SWINGCTRL_TX1LN0_PORT1		0x168148
+#define MG_TX_SWINGCTRL_TX1LN1_PORT1		0x168548
+#define MG_TX_SWINGCTRL_TX1LN0_PORT2		0x169148
+#define MG_TX_SWINGCTRL_TX1LN1_PORT2		0x169548
+#define MG_TX1_SWINGCTRL(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_TX_SWINGCTRL_TX1LN0_PORT1, \
+		       MG_TX_SWINGCTRL_TX1LN0_PORT2, \
+		       MG_TX_SWINGCTRL_TX1LN1_PORT1)
+
+#define MG_TX_SWINGCTRL_TX2LN0_PORT1		0x1680C8
+#define MG_TX_SWINGCTRL_TX2LN1_PORT1		0x1684C8
+#define MG_TX_SWINGCTRL_TX2LN0_PORT2		0x1690C8
+#define MG_TX_SWINGCTRL_TX2LN1_PORT2		0x1694C8
+#define MG_TX2_SWINGCTRL(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_TX_SWINGCTRL_TX2LN0_PORT1, \
+		       MG_TX_SWINGCTRL_TX2LN0_PORT2, \
+		       MG_TX_SWINGCTRL_TX2LN1_PORT1)
+#define   CRI_TXDEEMPH_OVERRIDE_17_12(x)		((x) << 0)
+#define   CRI_TXDEEMPH_OVERRIDE_17_12_MASK		(0x3F << 0)
+
+#define MG_TX_DRVCTRL_TX1LN0_TXPORT1			0x168144
+#define MG_TX_DRVCTRL_TX1LN1_TXPORT1			0x168544
+#define MG_TX_DRVCTRL_TX1LN0_TXPORT2			0x169144
+#define MG_TX_DRVCTRL_TX1LN1_TXPORT2			0x169544
+#define MG_TX_DRVCTRL_TX1LN0_TXPORT3			0x16A144
+#define MG_TX_DRVCTRL_TX1LN1_TXPORT3			0x16A544
+#define MG_TX_DRVCTRL_TX1LN0_TXPORT4			0x16B144
+#define MG_TX_DRVCTRL_TX1LN1_TXPORT4			0x16B544
+#define MG_TX1_DRVCTRL(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_TX_DRVCTRL_TX1LN0_TXPORT1, \
+		       MG_TX_DRVCTRL_TX1LN0_TXPORT2, \
+		       MG_TX_DRVCTRL_TX1LN1_TXPORT1)
+
+#define MG_TX_DRVCTRL_TX2LN0_PORT1			0x1680C4
+#define MG_TX_DRVCTRL_TX2LN1_PORT1			0x1684C4
+#define MG_TX_DRVCTRL_TX2LN0_PORT2			0x1690C4
+#define MG_TX_DRVCTRL_TX2LN1_PORT2			0x1694C4
+#define MG_TX2_DRVCTRL(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_TX_DRVCTRL_TX2LN0_PORT1, \
+		       MG_TX_DRVCTRL_TX2LN0_PORT2, \
+		       MG_TX_DRVCTRL_TX2LN1_PORT1)
+#define   CRI_TXDEEMPH_OVERRIDE_11_6(x)			((x) << 24)
+#define   CRI_TXDEEMPH_OVERRIDE_11_6_MASK		(0x3F << 24)
+#define   CRI_TXDEEMPH_OVERRIDE_EN			(1 << 22)
+#define   CRI_TXDEEMPH_OVERRIDE_5_0(x)			((x) << 16)
+#define   CRI_TXDEEMPH_OVERRIDE_5_0_MASK		(0x3F << 16)
+#define   CRI_LOADGEN_SEL(x)				((x) << 12)
+#define   CRI_LOADGEN_SEL_MASK				(0x3 << 12)
+
+#define MG_CLKHUB_LN0_PORT1			0x16839C
+#define MG_CLKHUB_LN1_PORT1			0x16879C
+#define MG_CLKHUB_LN0_PORT2			0x16939C
+#define MG_CLKHUB_LN1_PORT2			0x16979C
+#define MG_CLKHUB(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_CLKHUB_LN0_PORT1, \
+		       MG_CLKHUB_LN0_PORT2, \
+		       MG_CLKHUB_LN1_PORT1)
+#define   CFG_LOW_RATE_LKREN_EN				(1 << 11)
+
+#define MG_TX_DCC_TX1LN0_PORT1			0x168110
+#define MG_TX_DCC_TX1LN1_PORT1			0x168510
+#define MG_TX_DCC_TX1LN0_PORT2			0x169110
+#define MG_TX_DCC_TX1LN1_PORT2			0x169510
+#define MG_TX1_DCC(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_TX_DCC_TX1LN0_PORT1, \
+		       MG_TX_DCC_TX1LN0_PORT2, \
+		       MG_TX_DCC_TX1LN1_PORT1)
+#define MG_TX_DCC_TX2LN0_PORT1			0x168090
+#define MG_TX_DCC_TX2LN1_PORT1			0x168490
+#define MG_TX_DCC_TX2LN0_PORT2			0x169090
+#define MG_TX_DCC_TX2LN1_PORT2			0x169490
+#define MG_TX2_DCC(ln, tc_port) \
+	MG_PHY_PORT_LN(ln, tc_port, MG_TX_DCC_TX2LN0_PORT1, \
+		       MG_TX_DCC_TX2LN0_PORT2, \
+		       MG_TX_DCC_TX2LN1_PORT1)
+#define   CFG_AMI_CK_DIV_OVERRIDE_VAL(x)	((x) << 25)
+#define   CFG_AMI_CK_DIV_OVERRIDE_VAL_MASK	(0x3 << 25)
+#define   CFG_AMI_CK_DIV_OVERRIDE_EN		(1 << 24)
+
+#define MG_DP_MODE_LN0_ACU_PORT1			0x1683A0
+#define MG_DP_MODE_LN1_ACU_PORT1			0x1687A0
+#define MG_DP_MODE_LN0_ACU_PORT2			0x1693A0
+#define MG_DP_MODE_LN1_ACU_PORT2			0x1697A0
+#define MG_DP_MODE(ln, tc_port)	\
+	MG_PHY_PORT_LN(ln, tc_port, MG_DP_MODE_LN0_ACU_PORT1, \
+		       MG_DP_MODE_LN0_ACU_PORT2, \
+		       MG_DP_MODE_LN1_ACU_PORT1)
+#define   MG_DP_MODE_CFG_DP_X2_MODE			(1 << 7)
+#define   MG_DP_MODE_CFG_DP_X1_MODE			(1 << 6)
+
+#define FIA1_BASE			0x163000
+#define FIA2_BASE			0x16E000
+#define FIA3_BASE			0x16F000
+#define _FIA(fia)			_PICK((fia), FIA1_BASE, FIA2_BASE, FIA3_BASE)
+#define _MMIO_FIA(fia, off)		_MMIO(_FIA(fia) + (off))
+
+/* ICL PHY DFLEX registers */
+#define PORT_TX_DFLEXDPMLE1(fia)		_MMIO_FIA((fia),  0x008C0)
+#define   DFLEXDPMLE1_DPMLETC_MASK(idx)		(0xf << (4 * (idx)))
+#define   DFLEXDPMLE1_DPMLETC_ML0(idx)		(1 << (4 * (idx)))
+#define   DFLEXDPMLE1_DPMLETC_ML1_0(idx)	(3 << (4 * (idx)))
+#define   DFLEXDPMLE1_DPMLETC_ML3(idx)		(8 << (4 * (idx)))
+#define   DFLEXDPMLE1_DPMLETC_ML3_2(idx)	(12 << (4 * (idx)))
+#define   DFLEXDPMLE1_DPMLETC_ML3_0(idx)	(15 << (4 * (idx)))
+
+#define _MG_REFCLKIN_CTL_PORT1				0x16892C
+#define _MG_REFCLKIN_CTL_PORT2				0x16992C
+#define   MG_REFCLKIN_CTL_OD_2_MUX(x)			((x) << 8)
+#define   MG_REFCLKIN_CTL_OD_2_MUX_MASK			(0x7 << 8)
+#define MG_REFCLKIN_CTL(tc_port) _MMIO_PORT((tc_port), \
+					    _MG_REFCLKIN_CTL_PORT1, \
+					    _MG_REFCLKIN_CTL_PORT2)
+
+#define _MG_CLKTOP2_CORECLKCTL1_PORT1			0x1688D8
+#define _MG_CLKTOP2_CORECLKCTL1_PORT2			0x1698D8
+#define   MG_CLKTOP2_CORECLKCTL1_B_DIVRATIO(x)		((x) << 16)
+#define   MG_CLKTOP2_CORECLKCTL1_B_DIVRATIO_MASK	(0xff << 16)
+#define   MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO(x)		((x) << 8)
+#define   MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO_MASK	(0xff << 8)
+#define MG_CLKTOP2_CORECLKCTL1(tc_port) _MMIO_PORT((tc_port), \
+						   _MG_CLKTOP2_CORECLKCTL1_PORT1, \
+						   _MG_CLKTOP2_CORECLKCTL1_PORT2)
+
+#define _MG_CLKTOP2_HSCLKCTL_PORT1			0x1688D4
+#define _MG_CLKTOP2_HSCLKCTL_PORT2			0x1698D4
+#define   MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL(x)		((x) << 16)
+#define   MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL_MASK	(0x1 << 16)
+#define   MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL(x)	((x) << 14)
+#define   MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL_MASK	(0x3 << 14)
+#define   MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_MASK		(0x3 << 12)
+#define   MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_2		(0 << 12)
+#define   MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_3		(1 << 12)
+#define   MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_5		(2 << 12)
+#define   MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_7		(3 << 12)
+#define   MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO(x)		((x) << 8)
+#define   MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_SHIFT		8
+#define   MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_MASK		(0xf << 8)
+#define MG_CLKTOP2_HSCLKCTL(tc_port) _MMIO_PORT((tc_port), \
+						_MG_CLKTOP2_HSCLKCTL_PORT1, \
+						_MG_CLKTOP2_HSCLKCTL_PORT2)
+
+#define _MG_PLL_DIV0_PORT1				0x168A00
+#define _MG_PLL_DIV0_PORT2				0x169A00
+#define   MG_PLL_DIV0_FRACNEN_H				(1 << 30)
+#define   MG_PLL_DIV0_FBDIV_FRAC_MASK			(0x3fffff << 8)
+#define   MG_PLL_DIV0_FBDIV_FRAC_SHIFT			8
+#define   MG_PLL_DIV0_FBDIV_FRAC(x)			((x) << 8)
+#define   MG_PLL_DIV0_FBDIV_INT_MASK			(0xff << 0)
+#define   MG_PLL_DIV0_FBDIV_INT(x)			((x) << 0)
+#define MG_PLL_DIV0(tc_port) _MMIO_PORT((tc_port), _MG_PLL_DIV0_PORT1, \
+					_MG_PLL_DIV0_PORT2)
+
+#define _MG_PLL_DIV1_PORT1				0x168A04
+#define _MG_PLL_DIV1_PORT2				0x169A04
+#define   MG_PLL_DIV1_IREF_NDIVRATIO(x)			((x) << 16)
+#define   MG_PLL_DIV1_DITHER_DIV_1			(0 << 12)
+#define   MG_PLL_DIV1_DITHER_DIV_2			(1 << 12)
+#define   MG_PLL_DIV1_DITHER_DIV_4			(2 << 12)
+#define   MG_PLL_DIV1_DITHER_DIV_8			(3 << 12)
+#define   MG_PLL_DIV1_NDIVRATIO(x)			((x) << 4)
+#define   MG_PLL_DIV1_FBPREDIV_MASK			(0xf << 0)
+#define   MG_PLL_DIV1_FBPREDIV(x)			((x) << 0)
+#define MG_PLL_DIV1(tc_port) _MMIO_PORT((tc_port), _MG_PLL_DIV1_PORT1, \
+					_MG_PLL_DIV1_PORT2)
+
+#define _MG_PLL_LF_PORT1				0x168A08
+#define _MG_PLL_LF_PORT2				0x169A08
+#define   MG_PLL_LF_TDCTARGETCNT(x)			((x) << 24)
+#define   MG_PLL_LF_AFCCNTSEL_256			(0 << 20)
+#define   MG_PLL_LF_AFCCNTSEL_512			(1 << 20)
+#define   MG_PLL_LF_GAINCTRL(x)				((x) << 16)
+#define   MG_PLL_LF_INT_COEFF(x)			((x) << 8)
+#define   MG_PLL_LF_PROP_COEFF(x)			((x) << 0)
+#define MG_PLL_LF(tc_port) _MMIO_PORT((tc_port), _MG_PLL_LF_PORT1, \
+				      _MG_PLL_LF_PORT2)
+
+#define _MG_PLL_FRAC_LOCK_PORT1				0x168A0C
+#define _MG_PLL_FRAC_LOCK_PORT2				0x169A0C
+#define   MG_PLL_FRAC_LOCK_TRUELOCK_CRIT_32		(1 << 18)
+#define   MG_PLL_FRAC_LOCK_EARLYLOCK_CRIT_32		(1 << 16)
+#define   MG_PLL_FRAC_LOCK_LOCKTHRESH(x)		((x) << 11)
+#define   MG_PLL_FRAC_LOCK_DCODITHEREN			(1 << 10)
+#define   MG_PLL_FRAC_LOCK_FEEDFWRDCAL_EN		(1 << 8)
+#define   MG_PLL_FRAC_LOCK_FEEDFWRDGAIN(x)		((x) << 0)
+#define MG_PLL_FRAC_LOCK(tc_port) _MMIO_PORT((tc_port), \
+					     _MG_PLL_FRAC_LOCK_PORT1, \
+					     _MG_PLL_FRAC_LOCK_PORT2)
+
+#define _MG_PLL_SSC_PORT1				0x168A10
+#define _MG_PLL_SSC_PORT2				0x169A10
+#define   MG_PLL_SSC_EN					(1 << 28)
+#define   MG_PLL_SSC_TYPE(x)				((x) << 26)
+#define   MG_PLL_SSC_STEPLENGTH(x)			((x) << 16)
+#define   MG_PLL_SSC_STEPNUM(x)				((x) << 10)
+#define   MG_PLL_SSC_FLLEN				(1 << 9)
+#define   MG_PLL_SSC_STEPSIZE(x)			((x) << 0)
+#define MG_PLL_SSC(tc_port) _MMIO_PORT((tc_port), _MG_PLL_SSC_PORT1, \
+				       _MG_PLL_SSC_PORT2)
+
+#define _MG_PLL_BIAS_PORT1				0x168A14
+#define _MG_PLL_BIAS_PORT2				0x169A14
+#define   MG_PLL_BIAS_BIAS_GB_SEL(x)			((x) << 30)
+#define   MG_PLL_BIAS_BIAS_GB_SEL_MASK			(0x3 << 30)
+#define   MG_PLL_BIAS_INIT_DCOAMP(x)			((x) << 24)
+#define   MG_PLL_BIAS_INIT_DCOAMP_MASK			(0x3f << 24)
+#define   MG_PLL_BIAS_BIAS_BONUS(x)			((x) << 16)
+#define   MG_PLL_BIAS_BIAS_BONUS_MASK			(0xff << 16)
+#define   MG_PLL_BIAS_BIASCAL_EN			(1 << 15)
+#define   MG_PLL_BIAS_CTRIM(x)				((x) << 8)
+#define   MG_PLL_BIAS_CTRIM_MASK			(0x1f << 8)
+#define   MG_PLL_BIAS_VREF_RDAC(x)			((x) << 5)
+#define   MG_PLL_BIAS_VREF_RDAC_MASK			(0x7 << 5)
+#define   MG_PLL_BIAS_IREFTRIM(x)			((x) << 0)
+#define   MG_PLL_BIAS_IREFTRIM_MASK			(0x1f << 0)
+#define MG_PLL_BIAS(tc_port) _MMIO_PORT((tc_port), _MG_PLL_BIAS_PORT1, \
+					_MG_PLL_BIAS_PORT2)
+
+#define _MG_PLL_TDC_COLDST_BIAS_PORT1			0x168A18
+#define _MG_PLL_TDC_COLDST_BIAS_PORT2			0x169A18
+#define   MG_PLL_TDC_COLDST_IREFINT_EN			(1 << 27)
+#define   MG_PLL_TDC_COLDST_REFBIAS_START_PULSE_W(x)	((x) << 17)
+#define   MG_PLL_TDC_COLDST_COLDSTART			(1 << 16)
+#define   MG_PLL_TDC_TDCOVCCORR_EN			(1 << 2)
+#define   MG_PLL_TDC_TDCSEL(x)				((x) << 0)
+#define MG_PLL_TDC_COLDST_BIAS(tc_port) _MMIO_PORT((tc_port), \
+						   _MG_PLL_TDC_COLDST_BIAS_PORT1, \
+						   _MG_PLL_TDC_COLDST_BIAS_PORT2)
+
+#endif /* __INTEL_TC_PHY_REGS__ */
diff --git a/drivers/gpu/drm/i915/display/intel_tv.c b/drivers/gpu/drm/i915/display/intel_tv.c
new file mode 100644
index 000000000..fb25be800
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_tv.c
@@ -0,0 +1,2020 @@
+/*
+ * Copyright © 2006-2008 Intel Corporation
+ *   Jesse Barnes <jesse.barnes@intel.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *
+ */
+
+/** @file
+ * Integrated TV-out support for the 915GM and 945GM.
+ */
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_edid.h>
+
+#include "i915_drv.h"
+#include "intel_connector.h"
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dpll.h"
+#include "intel_hotplug.h"
+#include "intel_tv.h"
+
+enum tv_margin {
+	TV_MARGIN_LEFT, TV_MARGIN_TOP,
+	TV_MARGIN_RIGHT, TV_MARGIN_BOTTOM
+};
+
+struct intel_tv {
+	struct intel_encoder base;
+
+	int type;
+};
+
+struct video_levels {
+	u16 blank, black;
+	u8 burst;
+};
+
+struct color_conversion {
+	u16 ry, gy, by, ay;
+	u16 ru, gu, bu, au;
+	u16 rv, gv, bv, av;
+};
+
+static const u32 filter_table[] = {
+	0xB1403000, 0x2E203500, 0x35002E20, 0x3000B140,
+	0x35A0B160, 0x2DC02E80, 0xB1403480, 0xB1603000,
+	0x2EA03640, 0x34002D80, 0x3000B120, 0x36E0B160,
+	0x2D202EF0, 0xB1203380, 0xB1603000, 0x2F303780,
+	0x33002CC0, 0x3000B100, 0x3820B160, 0x2C802F50,
+	0xB10032A0, 0xB1603000, 0x2F9038C0, 0x32202C20,
+	0x3000B0E0, 0x3980B160, 0x2BC02FC0, 0xB0E031C0,
+	0xB1603000, 0x2FF03A20, 0x31602B60, 0xB020B0C0,
+	0x3AE0B160, 0x2B001810, 0xB0C03120, 0xB140B020,
+	0x18283BA0, 0x30C02A80, 0xB020B0A0, 0x3C60B140,
+	0x2A201838, 0xB0A03080, 0xB120B020, 0x18383D20,
+	0x304029C0, 0xB040B080, 0x3DE0B100, 0x29601848,
+	0xB0803000, 0xB100B040, 0x18483EC0, 0xB0402900,
+	0xB040B060, 0x3F80B0C0, 0x28801858, 0xB060B080,
+	0xB0A0B060, 0x18602820, 0xB0A02820, 0x0000B060,
+	0xB1403000, 0x2E203500, 0x35002E20, 0x3000B140,
+	0x35A0B160, 0x2DC02E80, 0xB1403480, 0xB1603000,
+	0x2EA03640, 0x34002D80, 0x3000B120, 0x36E0B160,
+	0x2D202EF0, 0xB1203380, 0xB1603000, 0x2F303780,
+	0x33002CC0, 0x3000B100, 0x3820B160, 0x2C802F50,
+	0xB10032A0, 0xB1603000, 0x2F9038C0, 0x32202C20,
+	0x3000B0E0, 0x3980B160, 0x2BC02FC0, 0xB0E031C0,
+	0xB1603000, 0x2FF03A20, 0x31602B60, 0xB020B0C0,
+	0x3AE0B160, 0x2B001810, 0xB0C03120, 0xB140B020,
+	0x18283BA0, 0x30C02A80, 0xB020B0A0, 0x3C60B140,
+	0x2A201838, 0xB0A03080, 0xB120B020, 0x18383D20,
+	0x304029C0, 0xB040B080, 0x3DE0B100, 0x29601848,
+	0xB0803000, 0xB100B040, 0x18483EC0, 0xB0402900,
+	0xB040B060, 0x3F80B0C0, 0x28801858, 0xB060B080,
+	0xB0A0B060, 0x18602820, 0xB0A02820, 0x0000B060,
+	0x36403000, 0x2D002CC0, 0x30003640, 0x2D0036C0,
+	0x35C02CC0, 0x37403000, 0x2C802D40, 0x30003540,
+	0x2D8037C0, 0x34C02C40, 0x38403000, 0x2BC02E00,
+	0x30003440, 0x2E2038C0, 0x34002B80, 0x39803000,
+	0x2B402E40, 0x30003380, 0x2E603A00, 0x33402B00,
+	0x3A803040, 0x2A802EA0, 0x30403300, 0x2EC03B40,
+	0x32802A40, 0x3C003040, 0x2A002EC0, 0x30803240,
+	0x2EC03C80, 0x320029C0, 0x3D403080, 0x29402F00,
+	0x308031C0, 0x2F203DC0, 0x31802900, 0x3E8030C0,
+	0x28802F40, 0x30C03140, 0x2F203F40, 0x31402840,
+	0x28003100, 0x28002F00, 0x00003100, 0x36403000,
+	0x2D002CC0, 0x30003640, 0x2D0036C0,
+	0x35C02CC0, 0x37403000, 0x2C802D40, 0x30003540,
+	0x2D8037C0, 0x34C02C40, 0x38403000, 0x2BC02E00,
+	0x30003440, 0x2E2038C0, 0x34002B80, 0x39803000,
+	0x2B402E40, 0x30003380, 0x2E603A00, 0x33402B00,
+	0x3A803040, 0x2A802EA0, 0x30403300, 0x2EC03B40,
+	0x32802A40, 0x3C003040, 0x2A002EC0, 0x30803240,
+	0x2EC03C80, 0x320029C0, 0x3D403080, 0x29402F00,
+	0x308031C0, 0x2F203DC0, 0x31802900, 0x3E8030C0,
+	0x28802F40, 0x30C03140, 0x2F203F40, 0x31402840,
+	0x28003100, 0x28002F00, 0x00003100,
+};
+
+/*
+ * Color conversion values have 3 separate fixed point formats:
+ *
+ * 10 bit fields (ay, au)
+ *   1.9 fixed point (b.bbbbbbbbb)
+ * 11 bit fields (ry, by, ru, gu, gv)
+ *   exp.mantissa (ee.mmmmmmmmm)
+ *   ee = 00 = 10^-1 (0.mmmmmmmmm)
+ *   ee = 01 = 10^-2 (0.0mmmmmmmmm)
+ *   ee = 10 = 10^-3 (0.00mmmmmmmmm)
+ *   ee = 11 = 10^-4 (0.000mmmmmmmmm)
+ * 12 bit fields (gy, rv, bu)
+ *   exp.mantissa (eee.mmmmmmmmm)
+ *   eee = 000 = 10^-1 (0.mmmmmmmmm)
+ *   eee = 001 = 10^-2 (0.0mmmmmmmmm)
+ *   eee = 010 = 10^-3 (0.00mmmmmmmmm)
+ *   eee = 011 = 10^-4 (0.000mmmmmmmmm)
+ *   eee = 100 = reserved
+ *   eee = 101 = reserved
+ *   eee = 110 = reserved
+ *   eee = 111 = 10^0 (m.mmmmmmmm) (only usable for 1.0 representation)
+ *
+ * Saturation and contrast are 8 bits, with their own representation:
+ * 8 bit field (saturation, contrast)
+ *   exp.mantissa (ee.mmmmmm)
+ *   ee = 00 = 10^-1 (0.mmmmmm)
+ *   ee = 01 = 10^0 (m.mmmmm)
+ *   ee = 10 = 10^1 (mm.mmmm)
+ *   ee = 11 = 10^2 (mmm.mmm)
+ *
+ * Simple conversion function:
+ *
+ * static u32
+ * float_to_csc_11(float f)
+ * {
+ *     u32 exp;
+ *     u32 mant;
+ *     u32 ret;
+ *
+ *     if (f < 0)
+ *         f = -f;
+ *
+ *     if (f >= 1) {
+ *         exp = 0x7;
+ *	   mant = 1 << 8;
+ *     } else {
+ *         for (exp = 0; exp < 3 && f < 0.5; exp++)
+ *	   f *= 2.0;
+ *         mant = (f * (1 << 9) + 0.5);
+ *         if (mant >= (1 << 9))
+ *             mant = (1 << 9) - 1;
+ *     }
+ *     ret = (exp << 9) | mant;
+ *     return ret;
+ * }
+ */
+
+/*
+ * Behold, magic numbers!  If we plant them they might grow a big
+ * s-video cable to the sky... or something.
+ *
+ * Pre-converted to appropriate hex value.
+ */
+
+/*
+ * PAL & NTSC values for composite & s-video connections
+ */
+static const struct color_conversion ntsc_m_csc_composite = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x0104,
+	.ru = 0x0733, .gu = 0x052d, .bu = 0x05c7, .au = 0x0200,
+	.rv = 0x0340, .gv = 0x030c, .bv = 0x06d0, .av = 0x0200,
+};
+
+static const struct video_levels ntsc_m_levels_composite = {
+	.blank = 225, .black = 267, .burst = 113,
+};
+
+static const struct color_conversion ntsc_m_csc_svideo = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x0133,
+	.ru = 0x076a, .gu = 0x0564, .bu = 0x030d, .au = 0x0200,
+	.rv = 0x037a, .gv = 0x033d, .bv = 0x06f6, .av = 0x0200,
+};
+
+static const struct video_levels ntsc_m_levels_svideo = {
+	.blank = 266, .black = 316, .burst = 133,
+};
+
+static const struct color_conversion ntsc_j_csc_composite = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x0119,
+	.ru = 0x074c, .gu = 0x0546, .bu = 0x05ec, .au = 0x0200,
+	.rv = 0x035a, .gv = 0x0322, .bv = 0x06e1, .av = 0x0200,
+};
+
+static const struct video_levels ntsc_j_levels_composite = {
+	.blank = 225, .black = 225, .burst = 113,
+};
+
+static const struct color_conversion ntsc_j_csc_svideo = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x014c,
+	.ru = 0x0788, .gu = 0x0581, .bu = 0x0322, .au = 0x0200,
+	.rv = 0x0399, .gv = 0x0356, .bv = 0x070a, .av = 0x0200,
+};
+
+static const struct video_levels ntsc_j_levels_svideo = {
+	.blank = 266, .black = 266, .burst = 133,
+};
+
+static const struct color_conversion pal_csc_composite = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x0113,
+	.ru = 0x0745, .gu = 0x053f, .bu = 0x05e1, .au = 0x0200,
+	.rv = 0x0353, .gv = 0x031c, .bv = 0x06dc, .av = 0x0200,
+};
+
+static const struct video_levels pal_levels_composite = {
+	.blank = 237, .black = 237, .burst = 118,
+};
+
+static const struct color_conversion pal_csc_svideo = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x0145,
+	.ru = 0x0780, .gu = 0x0579, .bu = 0x031c, .au = 0x0200,
+	.rv = 0x0390, .gv = 0x034f, .bv = 0x0705, .av = 0x0200,
+};
+
+static const struct video_levels pal_levels_svideo = {
+	.blank = 280, .black = 280, .burst = 139,
+};
+
+static const struct color_conversion pal_m_csc_composite = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x0104,
+	.ru = 0x0733, .gu = 0x052d, .bu = 0x05c7, .au = 0x0200,
+	.rv = 0x0340, .gv = 0x030c, .bv = 0x06d0, .av = 0x0200,
+};
+
+static const struct video_levels pal_m_levels_composite = {
+	.blank = 225, .black = 267, .burst = 113,
+};
+
+static const struct color_conversion pal_m_csc_svideo = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x0133,
+	.ru = 0x076a, .gu = 0x0564, .bu = 0x030d, .au = 0x0200,
+	.rv = 0x037a, .gv = 0x033d, .bv = 0x06f6, .av = 0x0200,
+};
+
+static const struct video_levels pal_m_levels_svideo = {
+	.blank = 266, .black = 316, .burst = 133,
+};
+
+static const struct color_conversion pal_n_csc_composite = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x0104,
+	.ru = 0x0733, .gu = 0x052d, .bu = 0x05c7, .au = 0x0200,
+	.rv = 0x0340, .gv = 0x030c, .bv = 0x06d0, .av = 0x0200,
+};
+
+static const struct video_levels pal_n_levels_composite = {
+	.blank = 225, .black = 267, .burst = 118,
+};
+
+static const struct color_conversion pal_n_csc_svideo = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x0133,
+	.ru = 0x076a, .gu = 0x0564, .bu = 0x030d, .au = 0x0200,
+	.rv = 0x037a, .gv = 0x033d, .bv = 0x06f6, .av = 0x0200,
+};
+
+static const struct video_levels pal_n_levels_svideo = {
+	.blank = 266, .black = 316, .burst = 139,
+};
+
+/*
+ * Component connections
+ */
+static const struct color_conversion sdtv_csc_yprpb = {
+	.ry = 0x0332, .gy = 0x012d, .by = 0x07d3, .ay = 0x0145,
+	.ru = 0x0559, .gu = 0x0353, .bu = 0x0100, .au = 0x0200,
+	.rv = 0x0100, .gv = 0x03ad, .bv = 0x074d, .av = 0x0200,
+};
+
+static const struct color_conversion hdtv_csc_yprpb = {
+	.ry = 0x05b3, .gy = 0x016e, .by = 0x0728, .ay = 0x0145,
+	.ru = 0x07d5, .gu = 0x038b, .bu = 0x0100, .au = 0x0200,
+	.rv = 0x0100, .gv = 0x03d1, .bv = 0x06bc, .av = 0x0200,
+};
+
+static const struct video_levels component_levels = {
+	.blank = 279, .black = 279, .burst = 0,
+};
+
+
+struct tv_mode {
+	const char *name;
+
+	u32 clock;
+	u16 refresh; /* in millihertz (for precision) */
+	u8 oversample;
+	u8 hsync_end;
+	u16 hblank_start, hblank_end, htotal;
+	bool progressive : 1, trilevel_sync : 1, component_only : 1;
+	u8 vsync_start_f1, vsync_start_f2, vsync_len;
+	bool veq_ena : 1;
+	u8 veq_start_f1, veq_start_f2, veq_len;
+	u8 vi_end_f1, vi_end_f2;
+	u16 nbr_end;
+	bool burst_ena : 1;
+	u8 hburst_start, hburst_len;
+	u8 vburst_start_f1;
+	u16 vburst_end_f1;
+	u8 vburst_start_f2;
+	u16 vburst_end_f2;
+	u8 vburst_start_f3;
+	u16 vburst_end_f3;
+	u8 vburst_start_f4;
+	u16 vburst_end_f4;
+	/*
+	 * subcarrier programming
+	 */
+	u16 dda2_size, dda3_size;
+	u8 dda1_inc;
+	u16 dda2_inc, dda3_inc;
+	u32 sc_reset;
+	bool pal_burst : 1;
+	/*
+	 * blank/black levels
+	 */
+	const struct video_levels *composite_levels, *svideo_levels;
+	const struct color_conversion *composite_color, *svideo_color;
+	const u32 *filter_table;
+};
+
+
+/*
+ * Sub carrier DDA
+ *
+ *  I think this works as follows:
+ *
+ *  subcarrier freq = pixel_clock * (dda1_inc + dda2_inc / dda2_size) / 4096
+ *
+ * Presumably, when dda3 is added in, it gets to adjust the dda2_inc value
+ *
+ * So,
+ *  dda1_ideal = subcarrier/pixel * 4096
+ *  dda1_inc = floor (dda1_ideal)
+ *  dda2 = dda1_ideal - dda1_inc
+ *
+ *  then pick a ratio for dda2 that gives the closest approximation. If
+ *  you can't get close enough, you can play with dda3 as well. This
+ *  seems likely to happen when dda2 is small as the jumps would be larger
+ *
+ * To invert this,
+ *
+ *  pixel_clock = subcarrier * 4096 / (dda1_inc + dda2_inc / dda2_size)
+ *
+ * The constants below were all computed using a 107.520MHz clock
+ */
+
+/*
+ * Register programming values for TV modes.
+ *
+ * These values account for -1s required.
+ */
+static const struct tv_mode tv_modes[] = {
+	{
+		.name		= "NTSC-M",
+		.clock		= 108000,
+		.refresh	= 59940,
+		.oversample	= 8,
+		.component_only = false,
+		/* 525 Lines, 60 Fields, 15.734KHz line, Sub-Carrier 3.580MHz */
+
+		.hsync_end	= 64,		    .hblank_end		= 124,
+		.hblank_start	= 836,		    .htotal		= 857,
+
+		.progressive	= false,	    .trilevel_sync = false,
+
+		.vsync_start_f1	= 6,		    .vsync_start_f2	= 7,
+		.vsync_len	= 6,
+
+		.veq_ena	= true,		    .veq_start_f1	= 0,
+		.veq_start_f2	= 1,		    .veq_len		= 18,
+
+		.vi_end_f1	= 20,		    .vi_end_f2		= 21,
+		.nbr_end	= 240,
+
+		.burst_ena	= true,
+		.hburst_start	= 72,		    .hburst_len		= 34,
+		.vburst_start_f1 = 9,		    .vburst_end_f1	= 240,
+		.vburst_start_f2 = 10,		    .vburst_end_f2	= 240,
+		.vburst_start_f3 = 9,		    .vburst_end_f3	= 240,
+		.vburst_start_f4 = 10,		    .vburst_end_f4	= 240,
+
+		/* desired 3.5800000 actual 3.5800000 clock 107.52 */
+		.dda1_inc	=    135,
+		.dda2_inc	=  20800,	    .dda2_size		=  27456,
+		.dda3_inc	=      0,	    .dda3_size		=      0,
+		.sc_reset	= TV_SC_RESET_EVERY_4,
+		.pal_burst	= false,
+
+		.composite_levels = &ntsc_m_levels_composite,
+		.composite_color = &ntsc_m_csc_composite,
+		.svideo_levels  = &ntsc_m_levels_svideo,
+		.svideo_color = &ntsc_m_csc_svideo,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name		= "NTSC-443",
+		.clock		= 108000,
+		.refresh	= 59940,
+		.oversample	= 8,
+		.component_only = false,
+		/* 525 Lines, 60 Fields, 15.734KHz line, Sub-Carrier 4.43MHz */
+		.hsync_end	= 64,		    .hblank_end		= 124,
+		.hblank_start	= 836,		    .htotal		= 857,
+
+		.progressive	= false,	    .trilevel_sync = false,
+
+		.vsync_start_f1 = 6,		    .vsync_start_f2	= 7,
+		.vsync_len	= 6,
+
+		.veq_ena	= true,		    .veq_start_f1	= 0,
+		.veq_start_f2	= 1,		    .veq_len		= 18,
+
+		.vi_end_f1	= 20,		    .vi_end_f2		= 21,
+		.nbr_end	= 240,
+
+		.burst_ena	= true,
+		.hburst_start	= 72,		    .hburst_len		= 34,
+		.vburst_start_f1 = 9,		    .vburst_end_f1	= 240,
+		.vburst_start_f2 = 10,		    .vburst_end_f2	= 240,
+		.vburst_start_f3 = 9,		    .vburst_end_f3	= 240,
+		.vburst_start_f4 = 10,		    .vburst_end_f4	= 240,
+
+		/* desired 4.4336180 actual 4.4336180 clock 107.52 */
+		.dda1_inc       =    168,
+		.dda2_inc       =   4093,       .dda2_size      =  27456,
+		.dda3_inc       =    310,       .dda3_size      =    525,
+		.sc_reset   = TV_SC_RESET_NEVER,
+		.pal_burst  = false,
+
+		.composite_levels = &ntsc_m_levels_composite,
+		.composite_color = &ntsc_m_csc_composite,
+		.svideo_levels  = &ntsc_m_levels_svideo,
+		.svideo_color = &ntsc_m_csc_svideo,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name		= "NTSC-J",
+		.clock		= 108000,
+		.refresh	= 59940,
+		.oversample	= 8,
+		.component_only = false,
+
+		/* 525 Lines, 60 Fields, 15.734KHz line, Sub-Carrier 3.580MHz */
+		.hsync_end	= 64,		    .hblank_end		= 124,
+		.hblank_start = 836,	    .htotal		= 857,
+
+		.progressive	= false,    .trilevel_sync = false,
+
+		.vsync_start_f1	= 6,	    .vsync_start_f2	= 7,
+		.vsync_len	= 6,
+
+		.veq_ena      = true,	    .veq_start_f1	= 0,
+		.veq_start_f2 = 1,	    .veq_len		= 18,
+
+		.vi_end_f1	= 20,		    .vi_end_f2		= 21,
+		.nbr_end	= 240,
+
+		.burst_ena	= true,
+		.hburst_start	= 72,		    .hburst_len		= 34,
+		.vburst_start_f1 = 9,		    .vburst_end_f1	= 240,
+		.vburst_start_f2 = 10,		    .vburst_end_f2	= 240,
+		.vburst_start_f3 = 9,		    .vburst_end_f3	= 240,
+		.vburst_start_f4 = 10,		    .vburst_end_f4	= 240,
+
+		/* desired 3.5800000 actual 3.5800000 clock 107.52 */
+		.dda1_inc	=    135,
+		.dda2_inc	=  20800,	    .dda2_size		=  27456,
+		.dda3_inc	=      0,	    .dda3_size		=      0,
+		.sc_reset	= TV_SC_RESET_EVERY_4,
+		.pal_burst	= false,
+
+		.composite_levels = &ntsc_j_levels_composite,
+		.composite_color = &ntsc_j_csc_composite,
+		.svideo_levels  = &ntsc_j_levels_svideo,
+		.svideo_color = &ntsc_j_csc_svideo,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name		= "PAL-M",
+		.clock		= 108000,
+		.refresh	= 59940,
+		.oversample	= 8,
+		.component_only = false,
+
+		/* 525 Lines, 60 Fields, 15.734KHz line, Sub-Carrier 3.580MHz */
+		.hsync_end	= 64,		  .hblank_end		= 124,
+		.hblank_start = 836,	  .htotal		= 857,
+
+		.progressive	= false,	    .trilevel_sync = false,
+
+		.vsync_start_f1	= 6,		    .vsync_start_f2	= 7,
+		.vsync_len	= 6,
+
+		.veq_ena	= true,		    .veq_start_f1	= 0,
+		.veq_start_f2	= 1,		    .veq_len		= 18,
+
+		.vi_end_f1	= 20,		    .vi_end_f2		= 21,
+		.nbr_end	= 240,
+
+		.burst_ena	= true,
+		.hburst_start	= 72,		    .hburst_len		= 34,
+		.vburst_start_f1 = 9,		    .vburst_end_f1	= 240,
+		.vburst_start_f2 = 10,		    .vburst_end_f2	= 240,
+		.vburst_start_f3 = 9,		    .vburst_end_f3	= 240,
+		.vburst_start_f4 = 10,		    .vburst_end_f4	= 240,
+
+		/* desired 3.5800000 actual 3.5800000 clock 107.52 */
+		.dda1_inc	=    135,
+		.dda2_inc	=  16704,	    .dda2_size		=  27456,
+		.dda3_inc	=      0,	    .dda3_size		=      0,
+		.sc_reset	= TV_SC_RESET_EVERY_8,
+		.pal_burst  = true,
+
+		.composite_levels = &pal_m_levels_composite,
+		.composite_color = &pal_m_csc_composite,
+		.svideo_levels  = &pal_m_levels_svideo,
+		.svideo_color = &pal_m_csc_svideo,
+
+		.filter_table = filter_table,
+	},
+	{
+		/* 625 Lines, 50 Fields, 15.625KHz line, Sub-Carrier 4.434MHz */
+		.name	    = "PAL-N",
+		.clock		= 108000,
+		.refresh	= 50000,
+		.oversample	= 8,
+		.component_only = false,
+
+		.hsync_end	= 64,		    .hblank_end		= 128,
+		.hblank_start = 844,	    .htotal		= 863,
+
+		.progressive  = false,    .trilevel_sync = false,
+
+
+		.vsync_start_f1	= 6,	   .vsync_start_f2	= 7,
+		.vsync_len	= 6,
+
+		.veq_ena	= true,		    .veq_start_f1	= 0,
+		.veq_start_f2	= 1,		    .veq_len		= 18,
+
+		.vi_end_f1	= 24,		    .vi_end_f2		= 25,
+		.nbr_end	= 286,
+
+		.burst_ena	= true,
+		.hburst_start = 73,	    .hburst_len		= 34,
+		.vburst_start_f1 = 8,	    .vburst_end_f1	= 285,
+		.vburst_start_f2 = 8,	    .vburst_end_f2	= 286,
+		.vburst_start_f3 = 9,	    .vburst_end_f3	= 286,
+		.vburst_start_f4 = 9,	    .vburst_end_f4	= 285,
+
+
+		/* desired 4.4336180 actual 4.4336180 clock 107.52 */
+		.dda1_inc       =    135,
+		.dda2_inc       =  23578,       .dda2_size      =  27648,
+		.dda3_inc       =    134,       .dda3_size      =    625,
+		.sc_reset   = TV_SC_RESET_EVERY_8,
+		.pal_burst  = true,
+
+		.composite_levels = &pal_n_levels_composite,
+		.composite_color = &pal_n_csc_composite,
+		.svideo_levels  = &pal_n_levels_svideo,
+		.svideo_color = &pal_n_csc_svideo,
+
+		.filter_table = filter_table,
+	},
+	{
+		/* 625 Lines, 50 Fields, 15.625KHz line, Sub-Carrier 4.434MHz */
+		.name	    = "PAL",
+		.clock		= 108000,
+		.refresh	= 50000,
+		.oversample	= 8,
+		.component_only = false,
+
+		.hsync_end	= 64,		    .hblank_end		= 142,
+		.hblank_start	= 844,	    .htotal		= 863,
+
+		.progressive	= false,    .trilevel_sync = false,
+
+		.vsync_start_f1	= 5,	    .vsync_start_f2	= 6,
+		.vsync_len	= 5,
+
+		.veq_ena	= true,	    .veq_start_f1	= 0,
+		.veq_start_f2	= 1,	    .veq_len		= 15,
+
+		.vi_end_f1	= 24,		    .vi_end_f2		= 25,
+		.nbr_end	= 286,
+
+		.burst_ena	= true,
+		.hburst_start	= 73,		    .hburst_len		= 32,
+		.vburst_start_f1 = 8,		    .vburst_end_f1	= 285,
+		.vburst_start_f2 = 8,		    .vburst_end_f2	= 286,
+		.vburst_start_f3 = 9,		    .vburst_end_f3	= 286,
+		.vburst_start_f4 = 9,		    .vburst_end_f4	= 285,
+
+		/* desired 4.4336180 actual 4.4336180 clock 107.52 */
+		.dda1_inc       =    168,
+		.dda2_inc       =   4122,       .dda2_size      =  27648,
+		.dda3_inc       =     67,       .dda3_size      =    625,
+		.sc_reset   = TV_SC_RESET_EVERY_8,
+		.pal_burst  = true,
+
+		.composite_levels = &pal_levels_composite,
+		.composite_color = &pal_csc_composite,
+		.svideo_levels  = &pal_levels_svideo,
+		.svideo_color = &pal_csc_svideo,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name       = "480p",
+		.clock		= 108000,
+		.refresh	= 59940,
+		.oversample     = 4,
+		.component_only = true,
+
+		.hsync_end      = 64,               .hblank_end         = 122,
+		.hblank_start   = 842,              .htotal             = 857,
+
+		.progressive    = true,		    .trilevel_sync = false,
+
+		.vsync_start_f1 = 12,               .vsync_start_f2     = 12,
+		.vsync_len      = 12,
+
+		.veq_ena        = false,
+
+		.vi_end_f1      = 44,               .vi_end_f2          = 44,
+		.nbr_end        = 479,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name       = "576p",
+		.clock		= 108000,
+		.refresh	= 50000,
+		.oversample     = 4,
+		.component_only = true,
+
+		.hsync_end      = 64,               .hblank_end         = 139,
+		.hblank_start   = 859,              .htotal             = 863,
+
+		.progressive    = true,		    .trilevel_sync = false,
+
+		.vsync_start_f1 = 10,               .vsync_start_f2     = 10,
+		.vsync_len      = 10,
+
+		.veq_ena        = false,
+
+		.vi_end_f1      = 48,               .vi_end_f2          = 48,
+		.nbr_end        = 575,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name       = "720p@60Hz",
+		.clock		= 148500,
+		.refresh	= 60000,
+		.oversample     = 2,
+		.component_only = true,
+
+		.hsync_end      = 80,               .hblank_end         = 300,
+		.hblank_start   = 1580,             .htotal             = 1649,
+
+		.progressive	= true,		    .trilevel_sync = true,
+
+		.vsync_start_f1 = 10,               .vsync_start_f2     = 10,
+		.vsync_len      = 10,
+
+		.veq_ena        = false,
+
+		.vi_end_f1      = 29,               .vi_end_f2          = 29,
+		.nbr_end        = 719,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name       = "720p@50Hz",
+		.clock		= 148500,
+		.refresh	= 50000,
+		.oversample     = 2,
+		.component_only = true,
+
+		.hsync_end      = 80,               .hblank_end         = 300,
+		.hblank_start   = 1580,             .htotal             = 1979,
+
+		.progressive	= true,		    .trilevel_sync = true,
+
+		.vsync_start_f1 = 10,               .vsync_start_f2     = 10,
+		.vsync_len      = 10,
+
+		.veq_ena        = false,
+
+		.vi_end_f1      = 29,               .vi_end_f2          = 29,
+		.nbr_end        = 719,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name       = "1080i@50Hz",
+		.clock		= 148500,
+		.refresh	= 50000,
+		.oversample     = 2,
+		.component_only = true,
+
+		.hsync_end      = 88,               .hblank_end         = 235,
+		.hblank_start   = 2155,             .htotal             = 2639,
+
+		.progressive	= false,	  .trilevel_sync = true,
+
+		.vsync_start_f1 = 4,              .vsync_start_f2     = 5,
+		.vsync_len      = 10,
+
+		.veq_ena	= true,	    .veq_start_f1	= 4,
+		.veq_start_f2   = 4,	    .veq_len		= 10,
+
+
+		.vi_end_f1      = 21,           .vi_end_f2          = 22,
+		.nbr_end        = 539,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name       = "1080i@60Hz",
+		.clock		= 148500,
+		.refresh	= 60000,
+		.oversample     = 2,
+		.component_only = true,
+
+		.hsync_end      = 88,               .hblank_end         = 235,
+		.hblank_start   = 2155,             .htotal             = 2199,
+
+		.progressive	= false,	    .trilevel_sync = true,
+
+		.vsync_start_f1 = 4,               .vsync_start_f2     = 5,
+		.vsync_len      = 10,
+
+		.veq_ena	= true,		    .veq_start_f1	= 4,
+		.veq_start_f2	= 4,		    .veq_len		= 10,
+
+
+		.vi_end_f1      = 21,               .vi_end_f2          = 22,
+		.nbr_end        = 539,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+
+	{
+		.name       = "1080p@30Hz",
+		.clock		= 148500,
+		.refresh	= 30000,
+		.oversample     = 2,
+		.component_only = true,
+
+		.hsync_end      = 88,               .hblank_end         = 235,
+		.hblank_start   = 2155,             .htotal             = 2199,
+
+		.progressive	= true,		    .trilevel_sync = true,
+
+		.vsync_start_f1 = 8,               .vsync_start_f2     = 8,
+		.vsync_len      = 10,
+
+		.veq_ena	= false,	.veq_start_f1	= 0,
+		.veq_start_f2	= 0,		    .veq_len		= 0,
+
+		.vi_end_f1      = 44,               .vi_end_f2          = 44,
+		.nbr_end        = 1079,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+
+	{
+		.name       = "1080p@50Hz",
+		.clock		= 148500,
+		.refresh	= 50000,
+		.oversample     = 1,
+		.component_only = true,
+
+		.hsync_end      = 88,               .hblank_end         = 235,
+		.hblank_start   = 2155,             .htotal             = 2639,
+
+		.progressive	= true,		    .trilevel_sync = true,
+
+		.vsync_start_f1 = 8,               .vsync_start_f2     = 8,
+		.vsync_len      = 10,
+
+		.veq_ena	= false,	.veq_start_f1	= 0,
+		.veq_start_f2	= 0,		    .veq_len		= 0,
+
+		.vi_end_f1      = 44,               .vi_end_f2          = 44,
+		.nbr_end        = 1079,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+
+	{
+		.name       = "1080p@60Hz",
+		.clock		= 148500,
+		.refresh	= 60000,
+		.oversample     = 1,
+		.component_only = true,
+
+		.hsync_end      = 88,               .hblank_end         = 235,
+		.hblank_start   = 2155,             .htotal             = 2199,
+
+		.progressive	= true,		    .trilevel_sync = true,
+
+		.vsync_start_f1 = 8,               .vsync_start_f2     = 8,
+		.vsync_len      = 10,
+
+		.veq_ena	= false,		    .veq_start_f1	= 0,
+		.veq_start_f2	= 0,		    .veq_len		= 0,
+
+		.vi_end_f1      = 44,               .vi_end_f2          = 44,
+		.nbr_end        = 1079,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+};
+
+struct intel_tv_connector_state {
+	struct drm_connector_state base;
+
+	/*
+	 * May need to override the user margins for
+	 * gen3 >1024 wide source vertical centering.
+	 */
+	struct {
+		u16 top, bottom;
+	} margins;
+
+	bool bypass_vfilter;
+};
+
+#define to_intel_tv_connector_state(x) container_of(x, struct intel_tv_connector_state, base)
+
+static struct drm_connector_state *
+intel_tv_connector_duplicate_state(struct drm_connector *connector)
+{
+	struct intel_tv_connector_state *state;
+
+	state = kmemdup(connector->state, sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return NULL;
+
+	__drm_atomic_helper_connector_duplicate_state(connector, &state->base);
+	return &state->base;
+}
+
+static struct intel_tv *enc_to_tv(struct intel_encoder *encoder)
+{
+	return container_of(encoder, struct intel_tv, base);
+}
+
+static struct intel_tv *intel_attached_tv(struct intel_connector *connector)
+{
+	return enc_to_tv(intel_attached_encoder(connector));
+}
+
+static bool
+intel_tv_get_hw_state(struct intel_encoder *encoder, enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 tmp = intel_de_read(dev_priv, TV_CTL);
+
+	*pipe = (tmp & TV_ENC_PIPE_SEL_MASK) >> TV_ENC_PIPE_SEL_SHIFT;
+
+	return tmp & TV_ENC_ENABLE;
+}
+
+static void
+intel_enable_tv(struct intel_atomic_state *state,
+		struct intel_encoder *encoder,
+		const struct intel_crtc_state *pipe_config,
+		const struct drm_connector_state *conn_state)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	/* Prevents vblank waits from timing out in intel_tv_detect_type() */
+	intel_crtc_wait_for_next_vblank(to_intel_crtc(pipe_config->uapi.crtc));
+
+	intel_de_write(dev_priv, TV_CTL,
+		       intel_de_read(dev_priv, TV_CTL) | TV_ENC_ENABLE);
+}
+
+static void
+intel_disable_tv(struct intel_atomic_state *state,
+		 struct intel_encoder *encoder,
+		 const struct intel_crtc_state *old_crtc_state,
+		 const struct drm_connector_state *old_conn_state)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	intel_de_write(dev_priv, TV_CTL,
+		       intel_de_read(dev_priv, TV_CTL) & ~TV_ENC_ENABLE);
+}
+
+static const struct tv_mode *intel_tv_mode_find(const struct drm_connector_state *conn_state)
+{
+	int format = conn_state->tv.mode;
+
+	return &tv_modes[format];
+}
+
+static enum drm_mode_status
+intel_tv_mode_valid(struct drm_connector *connector,
+		    struct drm_display_mode *mode)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	const struct tv_mode *tv_mode = intel_tv_mode_find(connector->state);
+	int max_dotclk = i915->max_dotclk_freq;
+	enum drm_mode_status status;
+
+	status = intel_cpu_transcoder_mode_valid(i915, mode);
+	if (status != MODE_OK)
+		return status;
+
+	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return MODE_NO_DBLESCAN;
+
+	if (mode->clock > max_dotclk)
+		return MODE_CLOCK_HIGH;
+
+	/* Ensure TV refresh is close to desired refresh */
+	if (abs(tv_mode->refresh - drm_mode_vrefresh(mode) * 1000) >= 1000)
+		return MODE_CLOCK_RANGE;
+
+	return MODE_OK;
+}
+
+static int
+intel_tv_mode_vdisplay(const struct tv_mode *tv_mode)
+{
+	if (tv_mode->progressive)
+		return tv_mode->nbr_end + 1;
+	else
+		return 2 * (tv_mode->nbr_end + 1);
+}
+
+static void
+intel_tv_mode_to_mode(struct drm_display_mode *mode,
+		      const struct tv_mode *tv_mode,
+		      int clock)
+{
+	mode->clock = clock / (tv_mode->oversample >> !tv_mode->progressive);
+
+	/*
+	 * tv_mode horizontal timings:
+	 *
+	 * hsync_end
+	 *    | hblank_end
+	 *    |    | hblank_start
+	 *    |    |       | htotal
+	 *    |     _______    |
+	 *     ____/       \___
+	 * \__/                \
+	 */
+	mode->hdisplay =
+		tv_mode->hblank_start - tv_mode->hblank_end;
+	mode->hsync_start = mode->hdisplay +
+		tv_mode->htotal - tv_mode->hblank_start;
+	mode->hsync_end = mode->hsync_start +
+		tv_mode->hsync_end;
+	mode->htotal = tv_mode->htotal + 1;
+
+	/*
+	 * tv_mode vertical timings:
+	 *
+	 * vsync_start
+	 *    | vsync_end
+	 *    |  | vi_end nbr_end
+	 *    |  |    |       |
+	 *    |  |     _______
+	 * \__    ____/       \
+	 *    \__/
+	 */
+	mode->vdisplay = intel_tv_mode_vdisplay(tv_mode);
+	if (tv_mode->progressive) {
+		mode->vsync_start = mode->vdisplay +
+			tv_mode->vsync_start_f1 + 1;
+		mode->vsync_end = mode->vsync_start +
+			tv_mode->vsync_len;
+		mode->vtotal = mode->vdisplay +
+			tv_mode->vi_end_f1 + 1;
+	} else {
+		mode->vsync_start = mode->vdisplay +
+			tv_mode->vsync_start_f1 + 1 +
+			tv_mode->vsync_start_f2 + 1;
+		mode->vsync_end = mode->vsync_start +
+			2 * tv_mode->vsync_len;
+		mode->vtotal = mode->vdisplay +
+			tv_mode->vi_end_f1 + 1 +
+			tv_mode->vi_end_f2 + 1;
+	}
+
+	/* TV has it's own notion of sync and other mode flags, so clear them. */
+	mode->flags = 0;
+
+	snprintf(mode->name, sizeof(mode->name),
+		 "%dx%d%c (%s)",
+		 mode->hdisplay, mode->vdisplay,
+		 tv_mode->progressive ? 'p' : 'i',
+		 tv_mode->name);
+}
+
+static void intel_tv_scale_mode_horiz(struct drm_display_mode *mode,
+				      int hdisplay, int left_margin,
+				      int right_margin)
+{
+	int hsync_start = mode->hsync_start - mode->hdisplay + right_margin;
+	int hsync_end = mode->hsync_end - mode->hdisplay + right_margin;
+	int new_htotal = mode->htotal * hdisplay /
+		(mode->hdisplay - left_margin - right_margin);
+
+	mode->clock = mode->clock * new_htotal / mode->htotal;
+
+	mode->hdisplay = hdisplay;
+	mode->hsync_start = hdisplay + hsync_start * new_htotal / mode->htotal;
+	mode->hsync_end = hdisplay + hsync_end * new_htotal / mode->htotal;
+	mode->htotal = new_htotal;
+}
+
+static void intel_tv_scale_mode_vert(struct drm_display_mode *mode,
+				     int vdisplay, int top_margin,
+				     int bottom_margin)
+{
+	int vsync_start = mode->vsync_start - mode->vdisplay + bottom_margin;
+	int vsync_end = mode->vsync_end - mode->vdisplay + bottom_margin;
+	int new_vtotal = mode->vtotal * vdisplay /
+		(mode->vdisplay - top_margin - bottom_margin);
+
+	mode->clock = mode->clock * new_vtotal / mode->vtotal;
+
+	mode->vdisplay = vdisplay;
+	mode->vsync_start = vdisplay + vsync_start * new_vtotal / mode->vtotal;
+	mode->vsync_end = vdisplay + vsync_end * new_vtotal / mode->vtotal;
+	mode->vtotal = new_vtotal;
+}
+
+static void
+intel_tv_get_config(struct intel_encoder *encoder,
+		    struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct drm_display_mode *adjusted_mode =
+		&pipe_config->hw.adjusted_mode;
+	struct drm_display_mode mode = {};
+	u32 tv_ctl, hctl1, hctl3, vctl1, vctl2, tmp;
+	struct tv_mode tv_mode = {};
+	int hdisplay = adjusted_mode->crtc_hdisplay;
+	int vdisplay = adjusted_mode->crtc_vdisplay;
+	int xsize, ysize, xpos, ypos;
+
+	pipe_config->output_types |= BIT(INTEL_OUTPUT_TVOUT);
+
+	tv_ctl = intel_de_read(dev_priv, TV_CTL);
+	hctl1 = intel_de_read(dev_priv, TV_H_CTL_1);
+	hctl3 = intel_de_read(dev_priv, TV_H_CTL_3);
+	vctl1 = intel_de_read(dev_priv, TV_V_CTL_1);
+	vctl2 = intel_de_read(dev_priv, TV_V_CTL_2);
+
+	tv_mode.htotal = (hctl1 & TV_HTOTAL_MASK) >> TV_HTOTAL_SHIFT;
+	tv_mode.hsync_end = (hctl1 & TV_HSYNC_END_MASK) >> TV_HSYNC_END_SHIFT;
+
+	tv_mode.hblank_start = (hctl3 & TV_HBLANK_START_MASK) >> TV_HBLANK_START_SHIFT;
+	tv_mode.hblank_end = (hctl3 & TV_HSYNC_END_MASK) >> TV_HBLANK_END_SHIFT;
+
+	tv_mode.nbr_end = (vctl1 & TV_NBR_END_MASK) >> TV_NBR_END_SHIFT;
+	tv_mode.vi_end_f1 = (vctl1 & TV_VI_END_F1_MASK) >> TV_VI_END_F1_SHIFT;
+	tv_mode.vi_end_f2 = (vctl1 & TV_VI_END_F2_MASK) >> TV_VI_END_F2_SHIFT;
+
+	tv_mode.vsync_len = (vctl2 & TV_VSYNC_LEN_MASK) >> TV_VSYNC_LEN_SHIFT;
+	tv_mode.vsync_start_f1 = (vctl2 & TV_VSYNC_START_F1_MASK) >> TV_VSYNC_START_F1_SHIFT;
+	tv_mode.vsync_start_f2 = (vctl2 & TV_VSYNC_START_F2_MASK) >> TV_VSYNC_START_F2_SHIFT;
+
+	tv_mode.clock = pipe_config->port_clock;
+
+	tv_mode.progressive = tv_ctl & TV_PROGRESSIVE;
+
+	switch (tv_ctl & TV_OVERSAMPLE_MASK) {
+	case TV_OVERSAMPLE_8X:
+		tv_mode.oversample = 8;
+		break;
+	case TV_OVERSAMPLE_4X:
+		tv_mode.oversample = 4;
+		break;
+	case TV_OVERSAMPLE_2X:
+		tv_mode.oversample = 2;
+		break;
+	default:
+		tv_mode.oversample = 1;
+		break;
+	}
+
+	tmp = intel_de_read(dev_priv, TV_WIN_POS);
+	xpos = tmp >> 16;
+	ypos = tmp & 0xffff;
+
+	tmp = intel_de_read(dev_priv, TV_WIN_SIZE);
+	xsize = tmp >> 16;
+	ysize = tmp & 0xffff;
+
+	intel_tv_mode_to_mode(&mode, &tv_mode, pipe_config->port_clock);
+
+	drm_dbg_kms(&dev_priv->drm, "TV mode: " DRM_MODE_FMT "\n",
+		    DRM_MODE_ARG(&mode));
+
+	intel_tv_scale_mode_horiz(&mode, hdisplay,
+				  xpos, mode.hdisplay - xsize - xpos);
+	intel_tv_scale_mode_vert(&mode, vdisplay,
+				 ypos, mode.vdisplay - ysize - ypos);
+
+	adjusted_mode->crtc_clock = mode.clock;
+	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
+		adjusted_mode->crtc_clock /= 2;
+
+	/* pixel counter doesn't work on i965gm TV output */
+	if (IS_I965GM(dev_priv))
+		pipe_config->mode_flags |=
+			I915_MODE_FLAG_USE_SCANLINE_COUNTER;
+}
+
+static bool intel_tv_source_too_wide(struct drm_i915_private *dev_priv,
+				     int hdisplay)
+{
+	return DISPLAY_VER(dev_priv) == 3 && hdisplay > 1024;
+}
+
+static bool intel_tv_vert_scaling(const struct drm_display_mode *tv_mode,
+				  const struct drm_connector_state *conn_state,
+				  int vdisplay)
+{
+	return tv_mode->crtc_vdisplay -
+		conn_state->tv.margins.top -
+		conn_state->tv.margins.bottom !=
+		vdisplay;
+}
+
+static int
+intel_tv_compute_config(struct intel_encoder *encoder,
+			struct intel_crtc_state *pipe_config,
+			struct drm_connector_state *conn_state)
+{
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(pipe_config->uapi.state);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_tv_connector_state *tv_conn_state =
+		to_intel_tv_connector_state(conn_state);
+	const struct tv_mode *tv_mode = intel_tv_mode_find(conn_state);
+	struct drm_display_mode *adjusted_mode =
+		&pipe_config->hw.adjusted_mode;
+	int hdisplay = adjusted_mode->crtc_hdisplay;
+	int vdisplay = adjusted_mode->crtc_vdisplay;
+	int ret;
+
+	if (!tv_mode)
+		return -EINVAL;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+
+	drm_dbg_kms(&dev_priv->drm, "forcing bpc to 8 for TV\n");
+	pipe_config->pipe_bpp = 8*3;
+
+	pipe_config->port_clock = tv_mode->clock;
+
+	ret = intel_dpll_crtc_compute_clock(state, crtc);
+	if (ret)
+		return ret;
+
+	pipe_config->clock_set = true;
+
+	intel_tv_mode_to_mode(adjusted_mode, tv_mode, pipe_config->port_clock);
+	drm_mode_set_crtcinfo(adjusted_mode, 0);
+
+	if (intel_tv_source_too_wide(dev_priv, hdisplay) ||
+	    !intel_tv_vert_scaling(adjusted_mode, conn_state, vdisplay)) {
+		int extra, top, bottom;
+
+		extra = adjusted_mode->crtc_vdisplay - vdisplay;
+
+		if (extra < 0) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "No vertical scaling for >1024 pixel wide modes\n");
+			return -EINVAL;
+		}
+
+		/* Need to turn off the vertical filter and center the image */
+
+		/* Attempt to maintain the relative sizes of the margins */
+		top = conn_state->tv.margins.top;
+		bottom = conn_state->tv.margins.bottom;
+
+		if (top + bottom)
+			top = extra * top / (top + bottom);
+		else
+			top = extra / 2;
+		bottom = extra - top;
+
+		tv_conn_state->margins.top = top;
+		tv_conn_state->margins.bottom = bottom;
+
+		tv_conn_state->bypass_vfilter = true;
+
+		if (!tv_mode->progressive) {
+			adjusted_mode->clock /= 2;
+			adjusted_mode->crtc_clock /= 2;
+			adjusted_mode->flags |= DRM_MODE_FLAG_INTERLACE;
+		}
+	} else {
+		tv_conn_state->margins.top = conn_state->tv.margins.top;
+		tv_conn_state->margins.bottom = conn_state->tv.margins.bottom;
+
+		tv_conn_state->bypass_vfilter = false;
+	}
+
+	drm_dbg_kms(&dev_priv->drm, "TV mode: " DRM_MODE_FMT "\n",
+		    DRM_MODE_ARG(adjusted_mode));
+
+	/*
+	 * The pipe scanline counter behaviour looks as follows when
+	 * using the TV encoder:
+	 *
+	 * time ->
+	 *
+	 * dsl=vtotal-1       |             |
+	 *                   ||            ||
+	 *               ___| |        ___| |
+	 *              /     |       /     |
+	 *             /      |      /      |
+	 * dsl=0   ___/       |_____/       |
+	 *        | | |  |  | |
+	 *         ^ ^ ^   ^ ^
+	 *         | | |   | pipe vblank/first part of tv vblank
+	 *         | | |   bottom margin
+	 *         | | active
+	 *         | top margin
+	 *         remainder of tv vblank
+	 *
+	 * When the TV encoder is used the pipe wants to run faster
+	 * than expected rate. During the active portion the TV
+	 * encoder stalls the pipe every few lines to keep it in
+	 * check. When the TV encoder reaches the bottom margin the
+	 * pipe simply stops. Once we reach the TV vblank the pipe is
+	 * no longer stalled and it runs at the max rate (apparently
+	 * oversample clock on gen3, cdclk on gen4). Once the pipe
+	 * reaches the pipe vtotal the pipe stops for the remainder
+	 * of the TV vblank/top margin. The pipe starts up again when
+	 * the TV encoder exits the top margin.
+	 *
+	 * To avoid huge hassles for vblank timestamping we scale
+	 * the pipe timings as if the pipe always runs at the average
+	 * rate it maintains during the active period. This also
+	 * gives us a reasonable guesstimate as to the pixel rate.
+	 * Due to the variation in the actual pipe speed the scanline
+	 * counter will give us slightly erroneous results during the
+	 * TV vblank/margins. But since vtotal was selected such that
+	 * it matches the average rate of the pipe during the active
+	 * portion the error shouldn't cause any serious grief to
+	 * vblank timestamps.
+	 *
+	 * For posterity here is the empirically derived formula
+	 * that gives us the maximum length of the pipe vblank
+	 * we can use without causing display corruption. Following
+	 * this would allow us to have a ticking scanline counter
+	 * everywhere except during the bottom margin (there the
+	 * pipe always stops). Ie. this would eliminate the second
+	 * flat portion of the above graph. However this would also
+	 * complicate vblank timestamping as the pipe vtotal would
+	 * no longer match the average rate the pipe runs at during
+	 * the active portion. Hence following this formula seems
+	 * more trouble that it's worth.
+	 *
+	 * if (GRAPHICS_VER(dev_priv) == 4) {
+	 *	num = cdclk * (tv_mode->oversample >> !tv_mode->progressive);
+	 *	den = tv_mode->clock;
+	 * } else {
+	 *	num = tv_mode->oversample >> !tv_mode->progressive;
+	 *	den = 1;
+	 * }
+	 * max_pipe_vblank_len ~=
+	 *	(num * tv_htotal * (tv_vblank_len + top_margin)) /
+	 *	(den * pipe_htotal);
+	 */
+	intel_tv_scale_mode_horiz(adjusted_mode, hdisplay,
+				  conn_state->tv.margins.left,
+				  conn_state->tv.margins.right);
+	intel_tv_scale_mode_vert(adjusted_mode, vdisplay,
+				 tv_conn_state->margins.top,
+				 tv_conn_state->margins.bottom);
+	drm_mode_set_crtcinfo(adjusted_mode, 0);
+	adjusted_mode->name[0] = '\0';
+
+	/* pixel counter doesn't work on i965gm TV output */
+	if (IS_I965GM(dev_priv))
+		pipe_config->mode_flags |=
+			I915_MODE_FLAG_USE_SCANLINE_COUNTER;
+
+	return 0;
+}
+
+static void
+set_tv_mode_timings(struct drm_i915_private *dev_priv,
+		    const struct tv_mode *tv_mode,
+		    bool burst_ena)
+{
+	u32 hctl1, hctl2, hctl3;
+	u32 vctl1, vctl2, vctl3, vctl4, vctl5, vctl6, vctl7;
+
+	hctl1 = (tv_mode->hsync_end << TV_HSYNC_END_SHIFT) |
+		(tv_mode->htotal << TV_HTOTAL_SHIFT);
+
+	hctl2 = (tv_mode->hburst_start << 16) |
+		(tv_mode->hburst_len << TV_HBURST_LEN_SHIFT);
+
+	if (burst_ena)
+		hctl2 |= TV_BURST_ENA;
+
+	hctl3 = (tv_mode->hblank_start << TV_HBLANK_START_SHIFT) |
+		(tv_mode->hblank_end << TV_HBLANK_END_SHIFT);
+
+	vctl1 = (tv_mode->nbr_end << TV_NBR_END_SHIFT) |
+		(tv_mode->vi_end_f1 << TV_VI_END_F1_SHIFT) |
+		(tv_mode->vi_end_f2 << TV_VI_END_F2_SHIFT);
+
+	vctl2 = (tv_mode->vsync_len << TV_VSYNC_LEN_SHIFT) |
+		(tv_mode->vsync_start_f1 << TV_VSYNC_START_F1_SHIFT) |
+		(tv_mode->vsync_start_f2 << TV_VSYNC_START_F2_SHIFT);
+
+	vctl3 = (tv_mode->veq_len << TV_VEQ_LEN_SHIFT) |
+		(tv_mode->veq_start_f1 << TV_VEQ_START_F1_SHIFT) |
+		(tv_mode->veq_start_f2 << TV_VEQ_START_F2_SHIFT);
+
+	if (tv_mode->veq_ena)
+		vctl3 |= TV_EQUAL_ENA;
+
+	vctl4 = (tv_mode->vburst_start_f1 << TV_VBURST_START_F1_SHIFT) |
+		(tv_mode->vburst_end_f1 << TV_VBURST_END_F1_SHIFT);
+
+	vctl5 = (tv_mode->vburst_start_f2 << TV_VBURST_START_F2_SHIFT) |
+		(tv_mode->vburst_end_f2 << TV_VBURST_END_F2_SHIFT);
+
+	vctl6 = (tv_mode->vburst_start_f3 << TV_VBURST_START_F3_SHIFT) |
+		(tv_mode->vburst_end_f3 << TV_VBURST_END_F3_SHIFT);
+
+	vctl7 = (tv_mode->vburst_start_f4 << TV_VBURST_START_F4_SHIFT) |
+		(tv_mode->vburst_end_f4 << TV_VBURST_END_F4_SHIFT);
+
+	intel_de_write(dev_priv, TV_H_CTL_1, hctl1);
+	intel_de_write(dev_priv, TV_H_CTL_2, hctl2);
+	intel_de_write(dev_priv, TV_H_CTL_3, hctl3);
+	intel_de_write(dev_priv, TV_V_CTL_1, vctl1);
+	intel_de_write(dev_priv, TV_V_CTL_2, vctl2);
+	intel_de_write(dev_priv, TV_V_CTL_3, vctl3);
+	intel_de_write(dev_priv, TV_V_CTL_4, vctl4);
+	intel_de_write(dev_priv, TV_V_CTL_5, vctl5);
+	intel_de_write(dev_priv, TV_V_CTL_6, vctl6);
+	intel_de_write(dev_priv, TV_V_CTL_7, vctl7);
+}
+
+static void set_color_conversion(struct drm_i915_private *dev_priv,
+				 const struct color_conversion *color_conversion)
+{
+	if (!color_conversion)
+		return;
+
+	intel_de_write(dev_priv, TV_CSC_Y,
+		       (color_conversion->ry << 16) | color_conversion->gy);
+	intel_de_write(dev_priv, TV_CSC_Y2,
+		       (color_conversion->by << 16) | color_conversion->ay);
+	intel_de_write(dev_priv, TV_CSC_U,
+		       (color_conversion->ru << 16) | color_conversion->gu);
+	intel_de_write(dev_priv, TV_CSC_U2,
+		       (color_conversion->bu << 16) | color_conversion->au);
+	intel_de_write(dev_priv, TV_CSC_V,
+		       (color_conversion->rv << 16) | color_conversion->gv);
+	intel_de_write(dev_priv, TV_CSC_V2,
+		       (color_conversion->bv << 16) | color_conversion->av);
+}
+
+static void intel_tv_pre_enable(struct intel_atomic_state *state,
+				struct intel_encoder *encoder,
+				const struct intel_crtc_state *pipe_config,
+				const struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct intel_tv *intel_tv = enc_to_tv(encoder);
+	const struct intel_tv_connector_state *tv_conn_state =
+		to_intel_tv_connector_state(conn_state);
+	const struct tv_mode *tv_mode = intel_tv_mode_find(conn_state);
+	u32 tv_ctl, tv_filter_ctl;
+	u32 scctl1, scctl2, scctl3;
+	int i, j;
+	const struct video_levels *video_levels;
+	const struct color_conversion *color_conversion;
+	bool burst_ena;
+	int xpos, ypos;
+	unsigned int xsize, ysize;
+
+	if (!tv_mode)
+		return;	/* can't happen (mode_prepare prevents this) */
+
+	tv_ctl = intel_de_read(dev_priv, TV_CTL);
+	tv_ctl &= TV_CTL_SAVE;
+
+	switch (intel_tv->type) {
+	default:
+	case DRM_MODE_CONNECTOR_Unknown:
+	case DRM_MODE_CONNECTOR_Composite:
+		tv_ctl |= TV_ENC_OUTPUT_COMPOSITE;
+		video_levels = tv_mode->composite_levels;
+		color_conversion = tv_mode->composite_color;
+		burst_ena = tv_mode->burst_ena;
+		break;
+	case DRM_MODE_CONNECTOR_Component:
+		tv_ctl |= TV_ENC_OUTPUT_COMPONENT;
+		video_levels = &component_levels;
+		if (tv_mode->burst_ena)
+			color_conversion = &sdtv_csc_yprpb;
+		else
+			color_conversion = &hdtv_csc_yprpb;
+		burst_ena = false;
+		break;
+	case DRM_MODE_CONNECTOR_SVIDEO:
+		tv_ctl |= TV_ENC_OUTPUT_SVIDEO;
+		video_levels = tv_mode->svideo_levels;
+		color_conversion = tv_mode->svideo_color;
+		burst_ena = tv_mode->burst_ena;
+		break;
+	}
+
+	tv_ctl |= TV_ENC_PIPE_SEL(crtc->pipe);
+
+	switch (tv_mode->oversample) {
+	case 8:
+		tv_ctl |= TV_OVERSAMPLE_8X;
+		break;
+	case 4:
+		tv_ctl |= TV_OVERSAMPLE_4X;
+		break;
+	case 2:
+		tv_ctl |= TV_OVERSAMPLE_2X;
+		break;
+	default:
+		tv_ctl |= TV_OVERSAMPLE_NONE;
+		break;
+	}
+
+	if (tv_mode->progressive)
+		tv_ctl |= TV_PROGRESSIVE;
+	if (tv_mode->trilevel_sync)
+		tv_ctl |= TV_TRILEVEL_SYNC;
+	if (tv_mode->pal_burst)
+		tv_ctl |= TV_PAL_BURST;
+
+	scctl1 = 0;
+	if (tv_mode->dda1_inc)
+		scctl1 |= TV_SC_DDA1_EN;
+	if (tv_mode->dda2_inc)
+		scctl1 |= TV_SC_DDA2_EN;
+	if (tv_mode->dda3_inc)
+		scctl1 |= TV_SC_DDA3_EN;
+	scctl1 |= tv_mode->sc_reset;
+	if (video_levels)
+		scctl1 |= video_levels->burst << TV_BURST_LEVEL_SHIFT;
+	scctl1 |= tv_mode->dda1_inc << TV_SCDDA1_INC_SHIFT;
+
+	scctl2 = tv_mode->dda2_size << TV_SCDDA2_SIZE_SHIFT |
+		tv_mode->dda2_inc << TV_SCDDA2_INC_SHIFT;
+
+	scctl3 = tv_mode->dda3_size << TV_SCDDA3_SIZE_SHIFT |
+		tv_mode->dda3_inc << TV_SCDDA3_INC_SHIFT;
+
+	/* Enable two fixes for the chips that need them. */
+	if (IS_I915GM(dev_priv))
+		tv_ctl |= TV_ENC_C0_FIX | TV_ENC_SDP_FIX;
+
+	set_tv_mode_timings(dev_priv, tv_mode, burst_ena);
+
+	intel_de_write(dev_priv, TV_SC_CTL_1, scctl1);
+	intel_de_write(dev_priv, TV_SC_CTL_2, scctl2);
+	intel_de_write(dev_priv, TV_SC_CTL_3, scctl3);
+
+	set_color_conversion(dev_priv, color_conversion);
+
+	if (DISPLAY_VER(dev_priv) >= 4)
+		intel_de_write(dev_priv, TV_CLR_KNOBS, 0x00404000);
+	else
+		intel_de_write(dev_priv, TV_CLR_KNOBS, 0x00606000);
+
+	if (video_levels)
+		intel_de_write(dev_priv, TV_CLR_LEVEL,
+			       ((video_levels->black << TV_BLACK_LEVEL_SHIFT) | (video_levels->blank << TV_BLANK_LEVEL_SHIFT)));
+
+	assert_transcoder_disabled(dev_priv, pipe_config->cpu_transcoder);
+
+	/* Filter ctl must be set before TV_WIN_SIZE */
+	tv_filter_ctl = TV_AUTO_SCALE;
+	if (tv_conn_state->bypass_vfilter)
+		tv_filter_ctl |= TV_V_FILTER_BYPASS;
+	intel_de_write(dev_priv, TV_FILTER_CTL_1, tv_filter_ctl);
+
+	xsize = tv_mode->hblank_start - tv_mode->hblank_end;
+	ysize = intel_tv_mode_vdisplay(tv_mode);
+
+	xpos = conn_state->tv.margins.left;
+	ypos = tv_conn_state->margins.top;
+	xsize -= (conn_state->tv.margins.left +
+		  conn_state->tv.margins.right);
+	ysize -= (tv_conn_state->margins.top +
+		  tv_conn_state->margins.bottom);
+	intel_de_write(dev_priv, TV_WIN_POS, (xpos << 16) | ypos);
+	intel_de_write(dev_priv, TV_WIN_SIZE, (xsize << 16) | ysize);
+
+	j = 0;
+	for (i = 0; i < 60; i++)
+		intel_de_write(dev_priv, TV_H_LUMA(i),
+			       tv_mode->filter_table[j++]);
+	for (i = 0; i < 60; i++)
+		intel_de_write(dev_priv, TV_H_CHROMA(i),
+			       tv_mode->filter_table[j++]);
+	for (i = 0; i < 43; i++)
+		intel_de_write(dev_priv, TV_V_LUMA(i),
+			       tv_mode->filter_table[j++]);
+	for (i = 0; i < 43; i++)
+		intel_de_write(dev_priv, TV_V_CHROMA(i),
+			       tv_mode->filter_table[j++]);
+	intel_de_write(dev_priv, TV_DAC,
+		       intel_de_read(dev_priv, TV_DAC) & TV_DAC_SAVE);
+	intel_de_write(dev_priv, TV_CTL, tv_ctl);
+}
+
+static int
+intel_tv_detect_type(struct intel_tv *intel_tv,
+		      struct drm_connector *connector)
+{
+	struct intel_crtc *crtc = to_intel_crtc(connector->state->crtc);
+	struct drm_device *dev = connector->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 tv_ctl, save_tv_ctl;
+	u32 tv_dac, save_tv_dac;
+	int type;
+
+	/* Disable TV interrupts around load detect or we'll recurse */
+	if (connector->polled & DRM_CONNECTOR_POLL_HPD) {
+		spin_lock_irq(&dev_priv->irq_lock);
+		i915_disable_pipestat(dev_priv, 0,
+				      PIPE_HOTPLUG_INTERRUPT_STATUS |
+				      PIPE_HOTPLUG_TV_INTERRUPT_STATUS);
+		spin_unlock_irq(&dev_priv->irq_lock);
+	}
+
+	save_tv_dac = tv_dac = intel_de_read(dev_priv, TV_DAC);
+	save_tv_ctl = tv_ctl = intel_de_read(dev_priv, TV_CTL);
+
+	/* Poll for TV detection */
+	tv_ctl &= ~(TV_ENC_ENABLE | TV_ENC_PIPE_SEL_MASK | TV_TEST_MODE_MASK);
+	tv_ctl |= TV_TEST_MODE_MONITOR_DETECT;
+	tv_ctl |= TV_ENC_PIPE_SEL(crtc->pipe);
+
+	tv_dac &= ~(TVDAC_SENSE_MASK | DAC_A_MASK | DAC_B_MASK | DAC_C_MASK);
+	tv_dac |= (TVDAC_STATE_CHG_EN |
+		   TVDAC_A_SENSE_CTL |
+		   TVDAC_B_SENSE_CTL |
+		   TVDAC_C_SENSE_CTL |
+		   DAC_CTL_OVERRIDE |
+		   DAC_A_0_7_V |
+		   DAC_B_0_7_V |
+		   DAC_C_0_7_V);
+
+
+	/*
+	 * The TV sense state should be cleared to zero on cantiga platform. Otherwise
+	 * the TV is misdetected. This is hardware requirement.
+	 */
+	if (IS_GM45(dev_priv))
+		tv_dac &= ~(TVDAC_STATE_CHG_EN | TVDAC_A_SENSE_CTL |
+			    TVDAC_B_SENSE_CTL | TVDAC_C_SENSE_CTL);
+
+	intel_de_write(dev_priv, TV_CTL, tv_ctl);
+	intel_de_write(dev_priv, TV_DAC, tv_dac);
+	intel_de_posting_read(dev_priv, TV_DAC);
+
+	intel_crtc_wait_for_next_vblank(crtc);
+
+	type = -1;
+	tv_dac = intel_de_read(dev_priv, TV_DAC);
+	drm_dbg_kms(&dev_priv->drm, "TV detected: %x, %x\n", tv_ctl, tv_dac);
+	/*
+	 *  A B C
+	 *  0 1 1 Composite
+	 *  1 0 X svideo
+	 *  0 0 0 Component
+	 */
+	if ((tv_dac & TVDAC_SENSE_MASK) == (TVDAC_B_SENSE | TVDAC_C_SENSE)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Detected Composite TV connection\n");
+		type = DRM_MODE_CONNECTOR_Composite;
+	} else if ((tv_dac & (TVDAC_A_SENSE|TVDAC_B_SENSE)) == TVDAC_A_SENSE) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Detected S-Video TV connection\n");
+		type = DRM_MODE_CONNECTOR_SVIDEO;
+	} else if ((tv_dac & TVDAC_SENSE_MASK) == 0) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Detected Component TV connection\n");
+		type = DRM_MODE_CONNECTOR_Component;
+	} else {
+		drm_dbg_kms(&dev_priv->drm, "Unrecognised TV connection\n");
+		type = -1;
+	}
+
+	intel_de_write(dev_priv, TV_DAC, save_tv_dac & ~TVDAC_STATE_CHG_EN);
+	intel_de_write(dev_priv, TV_CTL, save_tv_ctl);
+	intel_de_posting_read(dev_priv, TV_CTL);
+
+	/* For unknown reasons the hw barfs if we don't do this vblank wait. */
+	intel_crtc_wait_for_next_vblank(crtc);
+
+	/* Restore interrupt config */
+	if (connector->polled & DRM_CONNECTOR_POLL_HPD) {
+		spin_lock_irq(&dev_priv->irq_lock);
+		i915_enable_pipestat(dev_priv, 0,
+				     PIPE_HOTPLUG_INTERRUPT_STATUS |
+				     PIPE_HOTPLUG_TV_INTERRUPT_STATUS);
+		spin_unlock_irq(&dev_priv->irq_lock);
+	}
+
+	return type;
+}
+
+/*
+ * Here we set accurate tv format according to connector type
+ * i.e Component TV should not be assigned by NTSC or PAL
+ */
+static void intel_tv_find_better_format(struct drm_connector *connector)
+{
+	struct intel_tv *intel_tv = intel_attached_tv(to_intel_connector(connector));
+	const struct tv_mode *tv_mode = intel_tv_mode_find(connector->state);
+	int i;
+
+	/* Component supports everything so we can keep the current mode */
+	if (intel_tv->type == DRM_MODE_CONNECTOR_Component)
+		return;
+
+	/* If the current mode is fine don't change it */
+	if (!tv_mode->component_only)
+		return;
+
+	for (i = 0; i < ARRAY_SIZE(tv_modes); i++) {
+		tv_mode = &tv_modes[i];
+
+		if (!tv_mode->component_only)
+			break;
+	}
+
+	connector->state->tv.mode = i;
+}
+
+static int
+intel_tv_detect(struct drm_connector *connector,
+		struct drm_modeset_acquire_ctx *ctx,
+		bool force)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+	struct intel_tv *intel_tv = intel_attached_tv(to_intel_connector(connector));
+	enum drm_connector_status status;
+	int type;
+
+	drm_dbg_kms(&i915->drm, "[CONNECTOR:%d:%s] force=%d\n",
+		    connector->base.id, connector->name, force);
+
+	if (!INTEL_DISPLAY_ENABLED(i915))
+		return connector_status_disconnected;
+
+	if (force) {
+		struct intel_load_detect_pipe tmp;
+		int ret;
+
+		ret = intel_get_load_detect_pipe(connector, &tmp, ctx);
+		if (ret < 0)
+			return ret;
+
+		if (ret > 0) {
+			type = intel_tv_detect_type(intel_tv, connector);
+			intel_release_load_detect_pipe(connector, &tmp, ctx);
+			status = type < 0 ?
+				connector_status_disconnected :
+				connector_status_connected;
+		} else
+			status = connector_status_unknown;
+
+		if (status == connector_status_connected) {
+			intel_tv->type = type;
+			intel_tv_find_better_format(connector);
+		}
+
+		return status;
+	} else
+		return connector->status;
+}
+
+static const struct input_res {
+	u16 w, h;
+} input_res_table[] = {
+	{ 640, 480 },
+	{ 800, 600 },
+	{ 1024, 768 },
+	{ 1280, 1024 },
+	{ 848, 480 },
+	{ 1280, 720 },
+	{ 1920, 1080 },
+};
+
+/* Choose preferred mode according to line number of TV format */
+static bool
+intel_tv_is_preferred_mode(const struct drm_display_mode *mode,
+			   const struct tv_mode *tv_mode)
+{
+	int vdisplay = intel_tv_mode_vdisplay(tv_mode);
+
+	/* prefer 480 line modes for all SD TV modes */
+	if (vdisplay <= 576)
+		vdisplay = 480;
+
+	return vdisplay == mode->vdisplay;
+}
+
+static void
+intel_tv_set_mode_type(struct drm_display_mode *mode,
+		       const struct tv_mode *tv_mode)
+{
+	mode->type = DRM_MODE_TYPE_DRIVER;
+
+	if (intel_tv_is_preferred_mode(mode, tv_mode))
+		mode->type |= DRM_MODE_TYPE_PREFERRED;
+}
+
+static int
+intel_tv_get_modes(struct drm_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->dev);
+	const struct tv_mode *tv_mode = intel_tv_mode_find(connector->state);
+	int i, count = 0;
+
+	for (i = 0; i < ARRAY_SIZE(input_res_table); i++) {
+		const struct input_res *input = &input_res_table[i];
+		struct drm_display_mode *mode;
+
+		if (input->w > 1024 &&
+		    !tv_mode->progressive &&
+		    !tv_mode->component_only)
+			continue;
+
+		/* no vertical scaling with wide sources on gen3 */
+		if (DISPLAY_VER(dev_priv) == 3 && input->w > 1024 &&
+		    input->h > intel_tv_mode_vdisplay(tv_mode))
+			continue;
+
+		mode = drm_mode_create(connector->dev);
+		if (!mode)
+			continue;
+
+		/*
+		 * We take the TV mode and scale it to look
+		 * like it had the expected h/vdisplay. This
+		 * provides the most information to userspace
+		 * about the actual timings of the mode. We
+		 * do ignore the margins though.
+		 */
+		intel_tv_mode_to_mode(mode, tv_mode, tv_mode->clock);
+		if (count == 0) {
+			drm_dbg_kms(&dev_priv->drm, "TV mode: " DRM_MODE_FMT "\n",
+				    DRM_MODE_ARG(mode));
+		}
+		intel_tv_scale_mode_horiz(mode, input->w, 0, 0);
+		intel_tv_scale_mode_vert(mode, input->h, 0, 0);
+		intel_tv_set_mode_type(mode, tv_mode);
+
+		drm_mode_set_name(mode);
+
+		drm_mode_probed_add(connector, mode);
+		count++;
+	}
+
+	return count;
+}
+
+static const struct drm_connector_funcs intel_tv_connector_funcs = {
+	.late_register = intel_connector_register,
+	.early_unregister = intel_connector_unregister,
+	.destroy = intel_connector_destroy,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.atomic_duplicate_state = intel_tv_connector_duplicate_state,
+};
+
+static int intel_tv_atomic_check(struct drm_connector *connector,
+				 struct drm_atomic_state *state)
+{
+	struct drm_connector_state *new_state;
+	struct drm_crtc_state *new_crtc_state;
+	struct drm_connector_state *old_state;
+
+	new_state = drm_atomic_get_new_connector_state(state, connector);
+	if (!new_state->crtc)
+		return 0;
+
+	old_state = drm_atomic_get_old_connector_state(state, connector);
+	new_crtc_state = drm_atomic_get_new_crtc_state(state, new_state->crtc);
+
+	if (old_state->tv.mode != new_state->tv.mode ||
+	    old_state->tv.margins.left != new_state->tv.margins.left ||
+	    old_state->tv.margins.right != new_state->tv.margins.right ||
+	    old_state->tv.margins.top != new_state->tv.margins.top ||
+	    old_state->tv.margins.bottom != new_state->tv.margins.bottom) {
+		/* Force a modeset. */
+
+		new_crtc_state->connectors_changed = true;
+	}
+
+	return 0;
+}
+
+static const struct drm_connector_helper_funcs intel_tv_connector_helper_funcs = {
+	.detect_ctx = intel_tv_detect,
+	.mode_valid = intel_tv_mode_valid,
+	.get_modes = intel_tv_get_modes,
+	.atomic_check = intel_tv_atomic_check,
+};
+
+static const struct drm_encoder_funcs intel_tv_enc_funcs = {
+	.destroy = intel_encoder_destroy,
+};
+
+void
+intel_tv_init(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct drm_connector *connector;
+	struct intel_tv *intel_tv;
+	struct intel_encoder *intel_encoder;
+	struct intel_connector *intel_connector;
+	u32 tv_dac_on, tv_dac_off, save_tv_dac;
+	const char *tv_format_names[ARRAY_SIZE(tv_modes)];
+	int i, initial_mode = 0;
+	struct drm_connector_state *state;
+
+	if ((intel_de_read(dev_priv, TV_CTL) & TV_FUSE_STATE_MASK) == TV_FUSE_STATE_DISABLED)
+		return;
+
+	if (!intel_bios_is_tv_present(dev_priv)) {
+		drm_dbg_kms(&dev_priv->drm, "Integrated TV is not present.\n");
+		return;
+	}
+
+	/*
+	 * Sanity check the TV output by checking to see if the
+	 * DAC register holds a value
+	 */
+	save_tv_dac = intel_de_read(dev_priv, TV_DAC);
+
+	intel_de_write(dev_priv, TV_DAC, save_tv_dac | TVDAC_STATE_CHG_EN);
+	tv_dac_on = intel_de_read(dev_priv, TV_DAC);
+
+	intel_de_write(dev_priv, TV_DAC, save_tv_dac & ~TVDAC_STATE_CHG_EN);
+	tv_dac_off = intel_de_read(dev_priv, TV_DAC);
+
+	intel_de_write(dev_priv, TV_DAC, save_tv_dac);
+
+	/*
+	 * If the register does not hold the state change enable
+	 * bit, (either as a 0 or a 1), assume it doesn't really
+	 * exist
+	 */
+	if ((tv_dac_on & TVDAC_STATE_CHG_EN) == 0 ||
+	    (tv_dac_off & TVDAC_STATE_CHG_EN) != 0)
+		return;
+
+	intel_tv = kzalloc(sizeof(*intel_tv), GFP_KERNEL);
+	if (!intel_tv) {
+		return;
+	}
+
+	intel_connector = intel_connector_alloc();
+	if (!intel_connector) {
+		kfree(intel_tv);
+		return;
+	}
+
+	intel_encoder = &intel_tv->base;
+	connector = &intel_connector->base;
+	state = connector->state;
+
+	/*
+	 * The documentation, for the older chipsets at least, recommend
+	 * using a polling method rather than hotplug detection for TVs.
+	 * This is because in order to perform the hotplug detection, the PLLs
+	 * for the TV must be kept alive increasing power drain and starving
+	 * bandwidth from other encoders. Notably for instance, it causes
+	 * pipe underruns on Crestline when this encoder is supposedly idle.
+	 *
+	 * More recent chipsets favour HDMI rather than integrated S-Video.
+	 */
+	intel_connector->polled = DRM_CONNECTOR_POLL_CONNECT;
+
+	drm_connector_init(dev, connector, &intel_tv_connector_funcs,
+			   DRM_MODE_CONNECTOR_SVIDEO);
+
+	drm_encoder_init(dev, &intel_encoder->base, &intel_tv_enc_funcs,
+			 DRM_MODE_ENCODER_TVDAC, "TV");
+
+	intel_encoder->compute_config = intel_tv_compute_config;
+	intel_encoder->get_config = intel_tv_get_config;
+	intel_encoder->pre_enable = intel_tv_pre_enable;
+	intel_encoder->enable = intel_enable_tv;
+	intel_encoder->disable = intel_disable_tv;
+	intel_encoder->get_hw_state = intel_tv_get_hw_state;
+	intel_connector->get_hw_state = intel_connector_get_hw_state;
+
+	intel_connector_attach_encoder(intel_connector, intel_encoder);
+
+	intel_encoder->type = INTEL_OUTPUT_TVOUT;
+	intel_encoder->power_domain = POWER_DOMAIN_PORT_OTHER;
+	intel_encoder->port = PORT_NONE;
+	intel_encoder->pipe_mask = ~0;
+	intel_encoder->cloneable = 0;
+	intel_tv->type = DRM_MODE_CONNECTOR_Unknown;
+
+	/* BIOS margin values */
+	state->tv.margins.left = 54;
+	state->tv.margins.top = 36;
+	state->tv.margins.right = 46;
+	state->tv.margins.bottom = 37;
+
+	state->tv.mode = initial_mode;
+
+	drm_connector_helper_add(connector, &intel_tv_connector_helper_funcs);
+	connector->interlace_allowed = false;
+	connector->doublescan_allowed = false;
+
+	/* Create TV properties then attach current values */
+	for (i = 0; i < ARRAY_SIZE(tv_modes); i++) {
+		/* 1080p50/1080p60 not supported on gen3 */
+		if (DISPLAY_VER(dev_priv) == 3 &&
+		    tv_modes[i].oversample == 1)
+			break;
+
+		tv_format_names[i] = tv_modes[i].name;
+	}
+	drm_mode_create_tv_properties(dev, i, tv_format_names);
+
+	drm_object_attach_property(&connector->base, dev->mode_config.tv_mode_property,
+				   state->tv.mode);
+	drm_object_attach_property(&connector->base,
+				   dev->mode_config.tv_left_margin_property,
+				   state->tv.margins.left);
+	drm_object_attach_property(&connector->base,
+				   dev->mode_config.tv_top_margin_property,
+				   state->tv.margins.top);
+	drm_object_attach_property(&connector->base,
+				   dev->mode_config.tv_right_margin_property,
+				   state->tv.margins.right);
+	drm_object_attach_property(&connector->base,
+				   dev->mode_config.tv_bottom_margin_property,
+				   state->tv.margins.bottom);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_tv.h b/drivers/gpu/drm/i915/display/intel_tv.h
new file mode 100644
index 000000000..44518575e
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_tv.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_TV_H__
+#define __INTEL_TV_H__
+
+struct drm_i915_private;
+
+void intel_tv_init(struct drm_i915_private *dev_priv);
+
+#endif /* __INTEL_TV_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_vbt_defs.h b/drivers/gpu/drm/i915/display/intel_vbt_defs.h
new file mode 100644
index 000000000..a9f44abfc
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_vbt_defs.h
@@ -0,0 +1,1067 @@
+/*
+ * Copyright © 2006-2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *
+ */
+
+/*
+ * This information is private to VBT parsing in intel_bios.c.
+ *
+ * Please do NOT include anywhere else.
+ */
+#ifndef _INTEL_BIOS_PRIVATE
+#error "intel_vbt_defs.h is private to intel_bios.c"
+#endif
+
+#ifndef _INTEL_VBT_DEFS_H_
+#define _INTEL_VBT_DEFS_H_
+
+#include "intel_bios.h"
+
+/**
+ * struct vbt_header - VBT Header structure
+ * @signature:		VBT signature, always starts with "$VBT"
+ * @version:		Version of this structure
+ * @header_size:	Size of this structure
+ * @vbt_size:		Size of VBT (VBT Header, BDB Header and data blocks)
+ * @vbt_checksum:	Checksum
+ * @reserved0:		Reserved
+ * @bdb_offset:		Offset of &struct bdb_header from beginning of VBT
+ * @aim_offset:		Offsets of add-in data blocks from beginning of VBT
+ */
+struct vbt_header {
+	u8 signature[20];
+	u16 version;
+	u16 header_size;
+	u16 vbt_size;
+	u8 vbt_checksum;
+	u8 reserved0;
+	u32 bdb_offset;
+	u32 aim_offset[4];
+} __packed;
+
+/**
+ * struct bdb_header - BDB Header structure
+ * @signature:		BDB signature "BIOS_DATA_BLOCK"
+ * @version:		Version of the data block definitions
+ * @header_size:	Size of this structure
+ * @bdb_size:		Size of BDB (BDB Header and data blocks)
+ */
+struct bdb_header {
+	u8 signature[16];
+	u16 version;
+	u16 header_size;
+	u16 bdb_size;
+} __packed;
+
+/*
+ * BDB version number dependencies are documented as:
+ *
+ * <start>+
+ *    indicates the field was introduced in version <start>
+ *    and is still valid
+ *
+ * <start>-<end>
+ *    indicates the field was introduced in version <start>
+ *    and obsoleted in version <end>+1.
+ *
+ * ??? indicates the specific version number is unknown
+ */
+
+/*
+ * There are several types of BIOS data blocks (BDBs), each block has
+ * an ID and size in the first 3 bytes (ID in first, size in next 2).
+ * Known types are listed below.
+ */
+enum bdb_block_id {
+	BDB_GENERAL_FEATURES		= 1,
+	BDB_GENERAL_DEFINITIONS		= 2,
+	BDB_OLD_TOGGLE_LIST		= 3,
+	BDB_MODE_SUPPORT_LIST		= 4,
+	BDB_GENERIC_MODE_TABLE		= 5,
+	BDB_EXT_MMIO_REGS		= 6,
+	BDB_SWF_IO			= 7,
+	BDB_SWF_MMIO			= 8,
+	BDB_PSR				= 9,
+	BDB_MODE_REMOVAL_TABLE		= 10,
+	BDB_CHILD_DEVICE_TABLE		= 11,
+	BDB_DRIVER_FEATURES		= 12,
+	BDB_DRIVER_PERSISTENCE		= 13,
+	BDB_EXT_TABLE_PTRS		= 14,
+	BDB_DOT_CLOCK_OVERRIDE		= 15,
+	BDB_DISPLAY_SELECT		= 16,
+	BDB_DRIVER_ROTATION		= 18,
+	BDB_DISPLAY_REMOVE		= 19,
+	BDB_OEM_CUSTOM			= 20,
+	BDB_EFP_LIST			= 21, /* workarounds for VGA hsync/vsync */
+	BDB_SDVO_LVDS_OPTIONS		= 22,
+	BDB_SDVO_PANEL_DTDS		= 23,
+	BDB_SDVO_LVDS_PNP_IDS		= 24,
+	BDB_SDVO_LVDS_POWER_SEQ		= 25,
+	BDB_TV_OPTIONS			= 26,
+	BDB_EDP				= 27,
+	BDB_LVDS_OPTIONS		= 40,
+	BDB_LVDS_LFP_DATA_PTRS		= 41,
+	BDB_LVDS_LFP_DATA		= 42,
+	BDB_LVDS_BACKLIGHT		= 43,
+	BDB_LFP_POWER			= 44,
+	BDB_MIPI_CONFIG			= 52,
+	BDB_MIPI_SEQUENCE		= 53,
+	BDB_COMPRESSION_PARAMETERS	= 56,
+	BDB_GENERIC_DTD			= 58,
+	BDB_SKIP			= 254, /* VBIOS private block, ignore */
+};
+
+/*
+ * Block 1 - General Bit Definitions
+ */
+
+struct bdb_general_features {
+        /* bits 1 */
+	u8 panel_fitting:2;
+	u8 flexaim:1;
+	u8 msg_enable:1;
+	u8 clear_screen:3;
+	u8 color_flip:1;
+
+        /* bits 2 */
+	u8 download_ext_vbt:1;
+	u8 enable_ssc:1;
+	u8 ssc_freq:1;
+	u8 enable_lfp_on_override:1;
+	u8 disable_ssc_ddt:1;
+	u8 underscan_vga_timings:1;
+	u8 display_clock_mode:1;
+	u8 vbios_hotplug_support:1;
+
+        /* bits 3 */
+	u8 disable_smooth_vision:1;
+	u8 single_dvi:1;
+	u8 rotate_180:1;					/* 181+ */
+	u8 fdi_rx_polarity_inverted:1;
+	u8 vbios_extended_mode:1;				/* 160+ */
+	u8 copy_ilfp_dtd_to_sdvo_lvds_dtd:1;			/* 160+ */
+	u8 panel_best_fit_timing:1;				/* 160+ */
+	u8 ignore_strap_state:1;				/* 160+ */
+
+        /* bits 4 */
+	u8 legacy_monitor_detect;
+
+        /* bits 5 */
+	u8 int_crt_support:1;
+	u8 int_tv_support:1;
+	u8 int_efp_support:1;
+	u8 dp_ssc_enable:1;	/* PCH attached eDP supports SSC */
+	u8 dp_ssc_freq:1;	/* SSC freq for PCH attached eDP */
+	u8 dp_ssc_dongle_supported:1;
+	u8 rsvd11:2; /* finish byte */
+
+	/* bits 6 */
+	u8 tc_hpd_retry_timeout:7;				/* 242+ */
+	u8 rsvd12:1;
+
+	/* bits 7 */
+	u8 afc_startup_config:2;				/* 249+ */
+	u8 rsvd13:6;
+} __packed;
+
+/*
+ * Block 2 - General Bytes Definition
+ */
+
+/* pre-915 */
+#define GPIO_PIN_DVI_LVDS	0x03 /* "DVI/LVDS DDC GPIO pins" */
+#define GPIO_PIN_ADD_I2C	0x05 /* "ADDCARD I2C GPIO pins" */
+#define GPIO_PIN_ADD_DDC	0x04 /* "ADDCARD DDC GPIO pins" */
+#define GPIO_PIN_ADD_DDC_I2C	0x06 /* "ADDCARD DDC/I2C GPIO pins" */
+
+/* Device handle */
+#define DEVICE_HANDLE_CRT	0x0001
+#define DEVICE_HANDLE_EFP1	0x0004
+#define DEVICE_HANDLE_EFP2	0x0040
+#define DEVICE_HANDLE_EFP3	0x0020
+#define DEVICE_HANDLE_EFP4	0x0010 /* 194+ */
+#define DEVICE_HANDLE_EFP5	0x0002 /* 215+ */
+#define DEVICE_HANDLE_EFP6	0x0001 /* 217+ */
+#define DEVICE_HANDLE_EFP7	0x0100 /* 217+ */
+#define DEVICE_HANDLE_EFP8	0x0200 /* 217+ */
+#define DEVICE_HANDLE_LFP1	0x0008
+#define DEVICE_HANDLE_LFP2	0x0080
+
+/* Pre 915 */
+#define DEVICE_TYPE_NONE	0x00
+#define DEVICE_TYPE_CRT		0x01
+#define DEVICE_TYPE_TV		0x09
+#define DEVICE_TYPE_EFP		0x12
+#define DEVICE_TYPE_LFP		0x22
+/* On 915+ */
+#define DEVICE_TYPE_CRT_DPMS		0x6001
+#define DEVICE_TYPE_CRT_DPMS_HOTPLUG	0x4001
+#define DEVICE_TYPE_TV_COMPOSITE	0x0209
+#define DEVICE_TYPE_TV_MACROVISION	0x0289
+#define DEVICE_TYPE_TV_RF_COMPOSITE	0x020c
+#define DEVICE_TYPE_TV_SVIDEO_COMPOSITE	0x0609
+#define DEVICE_TYPE_TV_SCART		0x0209
+#define DEVICE_TYPE_TV_CODEC_HOTPLUG_PWR 0x6009
+#define DEVICE_TYPE_EFP_HOTPLUG_PWR	0x6012
+#define DEVICE_TYPE_EFP_DVI_HOTPLUG_PWR	0x6052
+#define DEVICE_TYPE_EFP_DVI_I		0x6053
+#define DEVICE_TYPE_EFP_DVI_D_DUAL	0x6152
+#define DEVICE_TYPE_EFP_DVI_D_HDCP	0x60d2
+#define DEVICE_TYPE_OPENLDI_HOTPLUG_PWR	0x6062
+#define DEVICE_TYPE_OPENLDI_DUALPIX	0x6162
+#define DEVICE_TYPE_LFP_PANELLINK	0x5012
+#define DEVICE_TYPE_LFP_CMOS_PWR	0x5042
+#define DEVICE_TYPE_LFP_LVDS_PWR	0x5062
+#define DEVICE_TYPE_LFP_LVDS_DUAL	0x5162
+#define DEVICE_TYPE_LFP_LVDS_DUAL_HDCP	0x51e2
+
+/* Add the device class for LFP, TV, HDMI */
+#define DEVICE_TYPE_INT_LFP		0x1022
+#define DEVICE_TYPE_INT_TV		0x1009
+#define DEVICE_TYPE_HDMI		0x60D2
+#define DEVICE_TYPE_DP			0x68C6
+#define DEVICE_TYPE_DP_DUAL_MODE	0x60D6
+#define DEVICE_TYPE_eDP			0x78C6
+
+#define DEVICE_TYPE_CLASS_EXTENSION	(1 << 15)
+#define DEVICE_TYPE_POWER_MANAGEMENT	(1 << 14)
+#define DEVICE_TYPE_HOTPLUG_SIGNALING	(1 << 13)
+#define DEVICE_TYPE_INTERNAL_CONNECTOR	(1 << 12)
+#define DEVICE_TYPE_NOT_HDMI_OUTPUT	(1 << 11)
+#define DEVICE_TYPE_MIPI_OUTPUT		(1 << 10)
+#define DEVICE_TYPE_COMPOSITE_OUTPUT	(1 << 9)
+#define DEVICE_TYPE_DUAL_CHANNEL	(1 << 8)
+#define DEVICE_TYPE_HIGH_SPEED_LINK	(1 << 6)
+#define DEVICE_TYPE_LVDS_SIGNALING	(1 << 5)
+#define DEVICE_TYPE_TMDS_DVI_SIGNALING	(1 << 4)
+#define DEVICE_TYPE_VIDEO_SIGNALING	(1 << 3)
+#define DEVICE_TYPE_DISPLAYPORT_OUTPUT	(1 << 2)
+#define DEVICE_TYPE_DIGITAL_OUTPUT	(1 << 1)
+#define DEVICE_TYPE_ANALOG_OUTPUT	(1 << 0)
+
+#define DEVICE_CFG_NONE		0x00
+#define DEVICE_CFG_12BIT_DVOB	0x01
+#define DEVICE_CFG_12BIT_DVOC	0x02
+#define DEVICE_CFG_24BIT_DVOBC	0x09
+#define DEVICE_CFG_24BIT_DVOCB	0x0a
+#define DEVICE_CFG_DUAL_DVOB	0x11
+#define DEVICE_CFG_DUAL_DVOC	0x12
+#define DEVICE_CFG_DUAL_DVOBC	0x13
+#define DEVICE_CFG_DUAL_LINK_DVOBC	0x19
+#define DEVICE_CFG_DUAL_LINK_DVOCB	0x1a
+
+#define DEVICE_WIRE_NONE	0x00
+#define DEVICE_WIRE_DVOB	0x01
+#define DEVICE_WIRE_DVOC	0x02
+#define DEVICE_WIRE_DVOBC	0x03
+#define DEVICE_WIRE_DVOBB	0x05
+#define DEVICE_WIRE_DVOCC	0x06
+#define DEVICE_WIRE_DVOB_MASTER 0x0d
+#define DEVICE_WIRE_DVOC_MASTER 0x0e
+
+/* dvo_port pre BDB 155 */
+#define DEVICE_PORT_DVOA	0x00 /* none on 845+ */
+#define DEVICE_PORT_DVOB	0x01
+#define DEVICE_PORT_DVOC	0x02
+
+/* dvo_port BDB 155+ */
+#define DVO_PORT_HDMIA		0
+#define DVO_PORT_HDMIB		1
+#define DVO_PORT_HDMIC		2
+#define DVO_PORT_HDMID		3
+#define DVO_PORT_LVDS		4
+#define DVO_PORT_TV		5
+#define DVO_PORT_CRT		6
+#define DVO_PORT_DPB		7
+#define DVO_PORT_DPC		8
+#define DVO_PORT_DPD		9
+#define DVO_PORT_DPA		10
+#define DVO_PORT_DPE		11				/* 193+ */
+#define DVO_PORT_HDMIE		12				/* 193+ */
+#define DVO_PORT_DPF		13				/* N/A */
+#define DVO_PORT_HDMIF		14				/* N/A */
+#define DVO_PORT_DPG		15				/* 217+ */
+#define DVO_PORT_HDMIG		16				/* 217+ */
+#define DVO_PORT_DPH		17				/* 217+ */
+#define DVO_PORT_HDMIH		18				/* 217+ */
+#define DVO_PORT_DPI		19				/* 217+ */
+#define DVO_PORT_HDMII		20				/* 217+ */
+#define DVO_PORT_MIPIA		21				/* 171+ */
+#define DVO_PORT_MIPIB		22				/* 171+ */
+#define DVO_PORT_MIPIC		23				/* 171+ */
+#define DVO_PORT_MIPID		24				/* 171+ */
+
+#define HDMI_MAX_DATA_RATE_PLATFORM	0			/* 204+ */
+#define HDMI_MAX_DATA_RATE_297		1			/* 204+ */
+#define HDMI_MAX_DATA_RATE_165		2			/* 204+ */
+#define HDMI_MAX_DATA_RATE_594		3			/* 249+ */
+#define HDMI_MAX_DATA_RATE_340		4			/* 249+ */
+#define HDMI_MAX_DATA_RATE_300		5			/* 249+ */
+
+#define LEGACY_CHILD_DEVICE_CONFIG_SIZE		33
+
+/* DDC Bus DDI Type 155+ */
+enum vbt_gmbus_ddi {
+	DDC_BUS_DDI_B = 0x1,
+	DDC_BUS_DDI_C,
+	DDC_BUS_DDI_D,
+	DDC_BUS_DDI_F,
+	ICL_DDC_BUS_DDI_A = 0x1,
+	ICL_DDC_BUS_DDI_B,
+	TGL_DDC_BUS_DDI_C,
+	RKL_DDC_BUS_DDI_D = 0x3,
+	RKL_DDC_BUS_DDI_E,
+	ICL_DDC_BUS_PORT_1 = 0x4,
+	ICL_DDC_BUS_PORT_2,
+	ICL_DDC_BUS_PORT_3,
+	ICL_DDC_BUS_PORT_4,
+	TGL_DDC_BUS_PORT_5,
+	TGL_DDC_BUS_PORT_6,
+	ADLS_DDC_BUS_PORT_TC1 = 0x2,
+	ADLS_DDC_BUS_PORT_TC2,
+	ADLS_DDC_BUS_PORT_TC3,
+	ADLS_DDC_BUS_PORT_TC4,
+	ADLP_DDC_BUS_PORT_TC1 = 0x3,
+	ADLP_DDC_BUS_PORT_TC2,
+	ADLP_DDC_BUS_PORT_TC3,
+	ADLP_DDC_BUS_PORT_TC4
+
+};
+
+#define DP_AUX_A 0x40
+#define DP_AUX_B 0x10
+#define DP_AUX_C 0x20
+#define DP_AUX_D 0x30
+#define DP_AUX_E 0x50
+#define DP_AUX_F 0x60
+#define DP_AUX_G 0x70
+#define DP_AUX_H 0x80
+#define DP_AUX_I 0x90
+
+/* DP max link rate 216+ */
+#define BDB_216_VBT_DP_MAX_LINK_RATE_HBR3	0
+#define BDB_216_VBT_DP_MAX_LINK_RATE_HBR2	1
+#define BDB_216_VBT_DP_MAX_LINK_RATE_HBR	2
+#define BDB_216_VBT_DP_MAX_LINK_RATE_LBR	3
+
+/* DP max link rate 230+ */
+#define BDB_230_VBT_DP_MAX_LINK_RATE_DEF	0
+#define BDB_230_VBT_DP_MAX_LINK_RATE_LBR	1
+#define BDB_230_VBT_DP_MAX_LINK_RATE_HBR	2
+#define BDB_230_VBT_DP_MAX_LINK_RATE_HBR2	3
+#define BDB_230_VBT_DP_MAX_LINK_RATE_HBR3	4
+#define BDB_230_VBT_DP_MAX_LINK_RATE_UHBR10	5
+#define BDB_230_VBT_DP_MAX_LINK_RATE_UHBR13P5	6
+#define BDB_230_VBT_DP_MAX_LINK_RATE_UHBR20	7
+
+/*
+ * The child device config, aka the display device data structure, provides a
+ * description of a port and its configuration on the platform.
+ *
+ * The child device config size has been increased, and fields have been added
+ * and their meaning has changed over time. Care must be taken when accessing
+ * basically any of the fields to ensure the correct interpretation for the BDB
+ * version in question.
+ *
+ * When we copy the child device configs to dev_priv->display.vbt.child_dev, we
+ * reserve space for the full structure below, and initialize the tail not
+ * actually present in VBT to zeros. Accessing those fields is fine, as long as
+ * the default zero is taken into account, again according to the BDB version.
+ *
+ * BDB versions 155 and below are considered legacy, and version 155 seems to be
+ * a baseline for some of the VBT documentation. When adding new fields, please
+ * include the BDB version when the field was added, if it's above that.
+ */
+struct child_device_config {
+	u16 handle;
+	u16 device_type; /* See DEVICE_TYPE_* above */
+
+	union {
+		u8  device_id[10]; /* ascii string */
+		struct {
+			u8 i2c_speed;
+			u8 dp_onboard_redriver_preemph:3;	/* 158+ */
+			u8 dp_onboard_redriver_vswing:3;	/* 158+ */
+			u8 dp_onboard_redriver_present:1;	/* 158+ */
+			u8 reserved0:1;
+			u8 dp_ondock_redriver_preemph:3;	/* 158+ */
+			u8 dp_ondock_redriver_vswing:3;		/* 158+ */
+			u8 dp_ondock_redriver_present:1;	/* 158+ */
+			u8 reserved1:1;
+			u8 hdmi_level_shifter_value:5;		/* 158+ */
+			u8 hdmi_max_data_rate:3;		/* 204+ */
+			u16 dtd_buf_ptr;			/* 161+ */
+			u8 edidless_efp:1;			/* 161+ */
+			u8 compression_enable:1;		/* 198+ */
+			u8 compression_method_cps:1;		/* 198+ */
+			u8 ganged_edp:1;			/* 202+ */
+			u8 lttpr_non_transparent:1;		/* 235+ */
+			u8 disable_compression_for_ext_disp:1;	/* 251+ */
+			u8 reserved2:2;
+			u8 compression_structure_index:4;	/* 198+ */
+			u8 reserved3:4;
+			u8 hdmi_max_frl_rate:4;			/* 237+ */
+			u8 hdmi_max_frl_rate_valid:1;		/* 237+ */
+			u8 reserved4:3;				/* 237+ */
+			u8 reserved5;
+		} __packed;
+	} __packed;
+
+	u16 addin_offset;
+	u8 dvo_port; /* See DEVICE_PORT_* and DVO_PORT_* above */
+	u8 i2c_pin;
+	u8 slave_addr;
+	u8 ddc_pin;
+	u16 edid_ptr;
+	u8 dvo_cfg; /* See DEVICE_CFG_* above */
+
+	union {
+		struct {
+			u8 dvo2_port;
+			u8 i2c2_pin;
+			u8 slave2_addr;
+			u8 ddc2_pin;
+		} __packed;
+		struct {
+			u8 efp_routed:1;			/* 158+ */
+			u8 lane_reversal:1;			/* 184+ */
+			u8 lspcon:1;				/* 192+ */
+			u8 iboost:1;				/* 196+ */
+			u8 hpd_invert:1;			/* 196+ */
+			u8 use_vbt_vswing:1;			/* 218+ */
+			u8 dp_max_lane_count:2;			/* 244+ */
+			u8 hdmi_support:1;			/* 158+ */
+			u8 dp_support:1;			/* 158+ */
+			u8 tmds_support:1;			/* 158+ */
+			u8 support_reserved:5;
+			u8 aux_channel;
+			u8 dongle_detect;
+		} __packed;
+	} __packed;
+
+	u8 pipe_cap:2;
+	u8 sdvo_stall:1;					/* 158+ */
+	u8 hpd_status:2;
+	u8 integrated_encoder:1;
+	u8 capabilities_reserved:2;
+	u8 dvo_wiring; /* See DEVICE_WIRE_* above */
+
+	union {
+		u8 dvo2_wiring;
+		u8 mipi_bridge_type;				/* 171+ */
+	} __packed;
+
+	u16 extended_type;
+	u8 dvo_function;
+	u8 dp_usb_type_c:1;					/* 195+ */
+	u8 tbt:1;						/* 209+ */
+	u8 flags2_reserved:2;					/* 195+ */
+	u8 dp_port_trace_length:4;				/* 209+ */
+	u8 dp_gpio_index;					/* 195+ */
+	u16 dp_gpio_pin_num;					/* 195+ */
+	u8 dp_iboost_level:4;					/* 196+ */
+	u8 hdmi_iboost_level:4;					/* 196+ */
+	u8 dp_max_link_rate:3;					/* 216+ */
+	u8 dp_max_link_rate_reserved:5;				/* 216+ */
+} __packed;
+
+struct bdb_general_definitions {
+	/* DDC GPIO */
+	u8 crt_ddc_gmbus_pin;
+
+	/* DPMS bits */
+	u8 dpms_non_acpi:1;
+	u8 skip_boot_crt_detect:1;
+	u8 dpms_aim:1;
+	u8 rsvd1:5; /* finish byte */
+
+	/* boot device bits */
+	u8 boot_display[2];
+	u8 child_dev_size;
+
+	/*
+	 * Device info:
+	 * If TV is present, it'll be at devices[0].
+	 * LVDS will be next, either devices[0] or [1], if present.
+	 * On some platforms the number of device is 6. But could be as few as
+	 * 4 if both TV and LVDS are missing.
+	 * And the device num is related with the size of general definition
+	 * block. It is obtained by using the following formula:
+	 * number = (block_size - sizeof(bdb_general_definitions))/
+	 *	     defs->child_dev_size;
+	 */
+	u8 devices[];
+} __packed;
+
+/*
+ * Block 9 - SRD Feature Block
+ */
+
+struct psr_table {
+	/* Feature bits */
+	u8 full_link:1;						/* 165+ */
+	u8 require_aux_to_wakeup:1;				/* 165+ */
+	u8 feature_bits_rsvd:6;
+
+	/* Wait times */
+	u8 idle_frames:4;					/* 165+ */
+	u8 lines_to_wait:3;					/* 165+ */
+	u8 wait_times_rsvd:1;
+
+	/* TP wake up time in multiple of 100 */
+	u16 tp1_wakeup_time;					/* 165+ */
+	u16 tp2_tp3_wakeup_time;				/* 165+ */
+} __packed;
+
+struct bdb_psr {
+	struct psr_table psr_table[16];
+
+	/* PSR2 TP2/TP3 wakeup time for 16 panels */
+	u32 psr2_tp2_tp3_wakeup_time;				/* 226+ */
+} __packed;
+
+/*
+ * Block 12 - Driver Features Data Block
+ */
+
+#define BDB_DRIVER_FEATURE_NO_LVDS		0
+#define BDB_DRIVER_FEATURE_INT_LVDS		1
+#define BDB_DRIVER_FEATURE_SDVO_LVDS		2
+#define BDB_DRIVER_FEATURE_INT_SDVO_LVDS	3
+
+struct bdb_driver_features {
+	/* Driver bits */
+	u8 boot_dev_algorithm:1;
+	u8 allow_display_switch_dvd:1;
+	u8 allow_display_switch_dos:1;
+	u8 hotplug_dvo:1;
+	u8 dual_view_zoom:1;
+	u8 int15h_hook:1;
+	u8 sprite_in_clone:1;
+	u8 primary_lfp_id:1;
+
+	u16 boot_mode_x;
+	u16 boot_mode_y;
+	u8 boot_mode_bpp;
+	u8 boot_mode_refresh;
+
+	/* Extended Driver Bits 1 */
+	u16 enable_lfp_primary:1;
+	u16 selective_mode_pruning:1;
+	u16 dual_frequency:1;
+	u16 render_clock_freq:1; /* 0: high freq; 1: low freq */
+	u16 nt_clone_support:1;
+	u16 power_scheme_ui:1; /* 0: CUI; 1: 3rd party */
+	u16 sprite_display_assign:1; /* 0: secondary; 1: primary */
+	u16 cui_aspect_scaling:1;
+	u16 preserve_aspect_ratio:1;
+	u16 sdvo_device_power_down:1;
+	u16 crt_hotplug:1;
+	u16 lvds_config:2;
+	u16 tv_hotplug:1;
+	u16 hdmi_config:2;
+
+	/* Driver Flags 1 */
+	u8 static_display:1;					/* 163+ */
+	u8 embedded_platform:1;					/* 163+ */
+	u8 display_subsystem_enable:1;				/* 163+ */
+	u8 reserved0:5;
+
+	u16 legacy_crt_max_x;
+	u16 legacy_crt_max_y;
+	u8 legacy_crt_max_refresh;
+
+	/* Extended Driver Bits 2 */
+	u8 hdmi_termination:1;
+	u8 cea861d_hdmi_support:1;
+	u8 self_refresh_enable:1;
+	u8 reserved1:5;
+
+	u8 custom_vbt_version;					/* 155+ */
+
+	/* Driver Feature Flags */
+	u16 rmpm_enabled:1;					/* 165+ */
+	u16 s2ddt_enabled:1;					/* 165+ */
+	u16 dpst_enabled:1;					/* 165-227 */
+	u16 bltclt_enabled:1;					/* 165+ */
+	u16 adb_enabled:1;					/* 165-227 */
+	u16 drrs_enabled:1;					/* 165-227 */
+	u16 grs_enabled:1;					/* 165+ */
+	u16 gpmt_enabled:1;					/* 165+ */
+	u16 tbt_enabled:1;					/* 165+ */
+	u16 psr_enabled:1;					/* 165-227 */
+	u16 ips_enabled:1;					/* 165+ */
+	u16 dpfs_enabled:1;					/* 165+ */
+	u16 dmrrs_enabled:1;					/* 174-227 */
+	u16 adt_enabled:1;					/* ???-228 */
+	u16 hpd_wake:1;						/* 201-240 */
+	u16 pc_feature_valid:1;
+} __packed;
+
+/*
+ * Block 22 - SDVO LVDS General Options
+ */
+
+struct bdb_sdvo_lvds_options {
+	u8 panel_backlight;
+	u8 h40_set_panel_type;
+	u8 panel_type;
+	u8 ssc_clk_freq;
+	u16 als_low_trip;
+	u16 als_high_trip;
+	u8 sclalarcoeff_tab_row_num;
+	u8 sclalarcoeff_tab_row_size;
+	u8 coefficient[8];
+	u8 panel_misc_bits_1;
+	u8 panel_misc_bits_2;
+	u8 panel_misc_bits_3;
+	u8 panel_misc_bits_4;
+} __packed;
+
+/*
+ * Block 23 - SDVO LVDS Panel DTDs
+ */
+
+struct lvds_dvo_timing {
+	u16 clock;		/**< In 10khz */
+	u8 hactive_lo;
+	u8 hblank_lo;
+	u8 hblank_hi:4;
+	u8 hactive_hi:4;
+	u8 vactive_lo;
+	u8 vblank_lo;
+	u8 vblank_hi:4;
+	u8 vactive_hi:4;
+	u8 hsync_off_lo;
+	u8 hsync_pulse_width_lo;
+	u8 vsync_pulse_width_lo:4;
+	u8 vsync_off_lo:4;
+	u8 vsync_pulse_width_hi:2;
+	u8 vsync_off_hi:2;
+	u8 hsync_pulse_width_hi:2;
+	u8 hsync_off_hi:2;
+	u8 himage_lo;
+	u8 vimage_lo;
+	u8 vimage_hi:4;
+	u8 himage_hi:4;
+	u8 h_border;
+	u8 v_border;
+	u8 rsvd1:3;
+	u8 digital:2;
+	u8 vsync_positive:1;
+	u8 hsync_positive:1;
+	u8 non_interlaced:1;
+} __packed;
+
+struct bdb_sdvo_panel_dtds {
+	struct lvds_dvo_timing dtds[4];
+} __packed;
+
+/*
+ * Block 27 - eDP VBT Block
+ */
+
+#define EDP_18BPP	0
+#define EDP_24BPP	1
+#define EDP_30BPP	2
+#define EDP_RATE_1_62	0
+#define EDP_RATE_2_7	1
+#define EDP_RATE_5_4	2
+#define EDP_LANE_1	0
+#define EDP_LANE_2	1
+#define EDP_LANE_4	3
+#define EDP_PREEMPHASIS_NONE	0
+#define EDP_PREEMPHASIS_3_5dB	1
+#define EDP_PREEMPHASIS_6dB	2
+#define EDP_PREEMPHASIS_9_5dB	3
+#define EDP_VSWING_0_4V		0
+#define EDP_VSWING_0_6V		1
+#define EDP_VSWING_0_8V		2
+#define EDP_VSWING_1_2V		3
+
+
+struct edp_fast_link_params {
+	u8 rate:4;						/* ???-223 */
+	u8 lanes:4;
+	u8 preemphasis:4;
+	u8 vswing:4;
+} __packed;
+
+struct edp_pwm_delays {
+	u16 pwm_on_to_backlight_enable;
+	u16 backlight_disable_to_pwm_off;
+} __packed;
+
+struct edp_full_link_params {
+	u8 preemphasis:4;
+	u8 vswing:4;
+} __packed;
+
+struct edp_apical_params {
+	u32 panel_oui;
+	u32 dpcd_base_address;
+	u32 dpcd_idridix_control_0;
+	u32 dpcd_option_select;
+	u32 dpcd_backlight;
+	u32 ambient_light;
+	u32 backlight_scale;
+} __packed;
+
+struct bdb_edp {
+	struct edp_power_seq power_seqs[16];
+	u32 color_depth;
+	struct edp_fast_link_params fast_link_params[16];
+	u32 sdrrs_msa_timing_delay;
+
+	/* ith bit indicates enabled/disabled for (i+1)th panel */
+	u16 edp_s3d_feature;					/* 162+ */
+	u16 edp_t3_optimization;				/* 165+ */
+	u64 edp_vswing_preemph;					/* 173+ */
+	u16 fast_link_training;					/* 182+ */
+	u16 dpcd_600h_write_required;				/* 185+ */
+	struct edp_pwm_delays pwm_delays[16];			/* 186+ */
+	u16 full_link_params_provided;				/* 199+ */
+	struct edp_full_link_params full_link_params[16];	/* 199+ */
+	u16 apical_enable;					/* 203+ */
+	struct edp_apical_params apical_params[16];		/* 203+ */
+	u16 edp_fast_link_training_rate[16];			/* 224+ */
+	u16 edp_max_port_link_rate[16];				/* 244+ */
+} __packed;
+
+/*
+ * Block 40 - LFP Data Block
+ */
+
+struct bdb_lvds_options {
+	u8 panel_type;
+	u8 panel_type2;						/* 212+ */
+	/* LVDS capabilities, stored in a dword */
+	u8 pfit_mode:2;
+	u8 pfit_text_mode_enhanced:1;
+	u8 pfit_gfx_mode_enhanced:1;
+	u8 pfit_ratio_auto:1;
+	u8 pixel_dither:1;
+	u8 lvds_edid:1;						/* ???-240 */
+	u8 rsvd2:1;
+	u8 rsvd4;
+	/* LVDS Panel channel bits stored here */
+	u32 lvds_panel_channel_bits;
+	/* LVDS SSC (Spread Spectrum Clock) bits stored here. */
+	u16 ssc_bits;
+	u16 ssc_freq;
+	u16 ssc_ddt;
+	/* Panel color depth defined here */
+	u16 panel_color_depth;
+	/* LVDS panel type bits stored here */
+	u32 dps_panel_type_bits;
+	/* LVDS backlight control type bits stored here */
+	u32 blt_control_type_bits;				/* ???-240 */
+
+	u16 lcdvcc_s0_enable;					/* 200+ */
+	u32 rotation;						/* 228+ */
+	u32 position;						/* 240+ */
+} __packed;
+
+/*
+ * Block 41 - LFP Data Table Pointers
+ */
+struct lvds_lfp_data_ptr_table {
+	u16 offset; /* offsets are from start of bdb */
+	u8 table_size;
+} __packed;
+
+/* LFP pointer table contains entries to the struct below */
+struct lvds_lfp_data_ptr {
+	struct lvds_lfp_data_ptr_table fp_timing;
+	struct lvds_lfp_data_ptr_table dvo_timing;
+	struct lvds_lfp_data_ptr_table panel_pnp_id;
+} __packed;
+
+struct bdb_lvds_lfp_data_ptrs {
+	u8 lvds_entries;
+	struct lvds_lfp_data_ptr ptr[16];
+	struct lvds_lfp_data_ptr_table panel_name;		/* (156-163?)+ */
+} __packed;
+
+/*
+ * Block 42 - LFP Data Tables
+ */
+
+/* LFP data has 3 blocks per entry */
+struct lvds_fp_timing {
+	u16 x_res;
+	u16 y_res;
+	u32 lvds_reg;
+	u32 lvds_reg_val;
+	u32 pp_on_reg;
+	u32 pp_on_reg_val;
+	u32 pp_off_reg;
+	u32 pp_off_reg_val;
+	u32 pp_cycle_reg;
+	u32 pp_cycle_reg_val;
+	u32 pfit_reg;
+	u32 pfit_reg_val;
+	u16 terminator;
+} __packed;
+
+struct lvds_pnp_id {
+	u16 mfg_name;
+	u16 product_code;
+	u32 serial;
+	u8 mfg_week;
+	u8 mfg_year;
+} __packed;
+
+/*
+ * For reference only. fp_timing has variable size so
+ * the data must be accessed using the data table pointers.
+ * Do not use this directly!
+ */
+struct lvds_lfp_data_entry {
+	struct lvds_fp_timing fp_timing;
+	struct lvds_dvo_timing dvo_timing;
+	struct lvds_pnp_id pnp_id;
+} __packed;
+
+struct bdb_lvds_lfp_data {
+	struct lvds_lfp_data_entry data[16];
+} __packed;
+
+struct lvds_lfp_panel_name {
+	u8 name[13];
+} __packed;
+
+struct lvds_lfp_black_border {
+	u8 top;		/* 227+ */
+	u8 bottom;	/* 227+ */
+	u8 left;	/* 238+ */
+	u8 right;	/* 238+ */
+} __packed;
+
+struct bdb_lvds_lfp_data_tail {
+	struct lvds_lfp_panel_name panel_name[16];		/* (156-163?)+ */
+	u16 scaling_enable;					/* 187+ */
+	u8 seamless_drrs_min_refresh_rate[16];			/* 188+ */
+	u8 pixel_overlap_count[16];				/* 208+ */
+	struct lvds_lfp_black_border black_border[16];		/* 227+ */
+	u16 dual_lfp_port_sync_enable;				/* 231+ */
+	u16 gpu_dithering_for_banding_artifacts;		/* 245+ */
+} __packed;
+
+/*
+ * Block 43 - LFP Backlight Control Data Block
+ */
+
+#define BDB_BACKLIGHT_TYPE_NONE	0
+#define BDB_BACKLIGHT_TYPE_PWM	2
+
+struct lfp_backlight_data_entry {
+	u8 type:2;
+	u8 active_low_pwm:1;
+	u8 obsolete1:5;
+	u16 pwm_freq_hz;
+	u8 min_brightness;					/* ???-233 */
+	u8 obsolete2;
+	u8 obsolete3;
+} __packed;
+
+struct lfp_backlight_control_method {
+	u8 type:4;
+	u8 controller:4;
+} __packed;
+
+struct lfp_brightness_level {
+	u16 level;
+	u16 reserved;
+} __packed;
+
+#define EXP_BDB_LFP_BL_DATA_SIZE_REV_191 \
+	offsetof(struct bdb_lfp_backlight_data, brightness_level)
+#define EXP_BDB_LFP_BL_DATA_SIZE_REV_234 \
+	offsetof(struct bdb_lfp_backlight_data, brightness_precision_bits)
+
+struct bdb_lfp_backlight_data {
+	u8 entry_size;
+	struct lfp_backlight_data_entry data[16];
+	u8 level[16];							/* ???-233 */
+	struct lfp_backlight_control_method backlight_control[16];
+	struct lfp_brightness_level brightness_level[16];		/* 234+ */
+	struct lfp_brightness_level brightness_min_level[16];		/* 234+ */
+	u8 brightness_precision_bits[16];				/* 236+ */
+	u16 hdr_dpcd_refresh_timeout[16];				/* 239+ */
+} __packed;
+
+/*
+ * Block 44 - LFP Power Conservation Features Block
+ */
+struct lfp_power_features {
+	u8 reserved1:1;
+	u8 power_conservation_pref:3;
+	u8 reserved2:1;
+	u8 lace_enabled_status:1;					/* 210+ */
+	u8 lace_support:1;						/* 210+ */
+	u8 als_enable:1;
+} __packed;
+
+struct als_data_entry {
+	u16 backlight_adjust;
+	u16 lux;
+} __packed;
+
+struct aggressiveness_profile_entry {
+	u8 dpst_aggressiveness : 4;
+	u8 lace_aggressiveness : 4;
+} __packed;
+
+struct aggressiveness_profile2_entry {
+	u8 opst_aggressiveness : 4;
+	u8 elp_aggressiveness : 4;
+} __packed;
+
+struct bdb_lfp_power {
+	struct lfp_power_features features;				/* ???-227 */
+	struct als_data_entry als[5];
+	u8 lace_aggressiveness_profile:3;				/* 210-227 */
+	u8 reserved1:5;
+	u16 dpst;							/* 228+ */
+	u16 psr;							/* 228+ */
+	u16 drrs;							/* 228+ */
+	u16 lace_support;						/* 228+ */
+	u16 adt;							/* 228+ */
+	u16 dmrrs;							/* 228+ */
+	u16 adb;							/* 228+ */
+	u16 lace_enabled_status;					/* 228+ */
+	struct aggressiveness_profile_entry aggressiveness[16];		/* 228+ */
+	u16 hobl;							/* 232+ */
+	u16 vrr_feature_enabled;					/* 233+ */
+	u16 elp;							/* 247+ */
+	u16 opst;							/* 247+ */
+	struct aggressiveness_profile2_entry aggressiveness2[16];	/* 247+ */
+} __packed;
+
+/*
+ * Block 52 - MIPI Configuration Block
+ */
+
+#define MAX_MIPI_CONFIGURATIONS	6
+
+struct bdb_mipi_config {
+	struct mipi_config config[MAX_MIPI_CONFIGURATIONS];		/* 175+ */
+	struct mipi_pps_data pps[MAX_MIPI_CONFIGURATIONS];		/* 177+ */
+	struct edp_pwm_delays pwm_delays[MAX_MIPI_CONFIGURATIONS];	/* 186+ */
+	u8 pmic_i2c_bus_number[MAX_MIPI_CONFIGURATIONS];		/* 190+ */
+} __packed;
+
+/*
+ * Block 53 - MIPI Sequence Block
+ */
+
+struct bdb_mipi_sequence {
+	u8 version;
+	u8 data[]; /* up to 6 variable length blocks */
+} __packed;
+
+/*
+ * Block 56 - Compression Parameters
+ */
+
+#define VBT_RC_BUFFER_BLOCK_SIZE_1KB	0
+#define VBT_RC_BUFFER_BLOCK_SIZE_4KB	1
+#define VBT_RC_BUFFER_BLOCK_SIZE_16KB	2
+#define VBT_RC_BUFFER_BLOCK_SIZE_64KB	3
+
+#define VBT_DSC_LINE_BUFFER_DEPTH(vbt_value)	((vbt_value) + 8) /* bits */
+#define VBT_DSC_MAX_BPP(vbt_value)		(6 + (vbt_value) * 2)
+
+struct dsc_compression_parameters_entry {
+	u8 version_major:4;
+	u8 version_minor:4;
+
+	u8 rc_buffer_block_size:2;
+	u8 reserved1:6;
+
+	/*
+	 * Buffer size in bytes:
+	 *
+	 * 4 ^ rc_buffer_block_size * 1024 * (rc_buffer_size + 1) bytes
+	 */
+	u8 rc_buffer_size;
+	u32 slices_per_line;
+
+	u8 line_buffer_depth:4;
+	u8 reserved2:4;
+
+	/* Flag Bits 1 */
+	u8 block_prediction_enable:1;
+	u8 reserved3:7;
+
+	u8 max_bpp; /* mapping */
+
+	/* Color depth capabilities */
+	u8 reserved4:1;
+	u8 support_8bpc:1;
+	u8 support_10bpc:1;
+	u8 support_12bpc:1;
+	u8 reserved5:4;
+
+	u16 slice_height;
+} __packed;
+
+struct bdb_compression_parameters {
+	u16 entry_size;
+	struct dsc_compression_parameters_entry data[16];
+} __packed;
+
+/*
+ * Block 58 - Generic DTD Block
+ */
+
+struct generic_dtd_entry {
+	u32 pixel_clock;
+	u16 hactive;
+	u16 hblank;
+	u16 hfront_porch;
+	u16 hsync;
+	u16 vactive;
+	u16 vblank;
+	u16 vfront_porch;
+	u16 vsync;
+	u16 width_mm;
+	u16 height_mm;
+
+	/* Flags */
+	u8 rsvd_flags:6;
+	u8 vsync_positive_polarity:1;
+	u8 hsync_positive_polarity:1;
+
+	u8 rsvd[3];
+} __packed;
+
+struct bdb_generic_dtd {
+	u16 gdtd_size;
+	struct generic_dtd_entry dtd[];	/* up to 24 DTD's */
+} __packed;
+
+#endif /* _INTEL_VBT_DEFS_H_ */
diff --git a/drivers/gpu/drm/i915/display/intel_vdsc.c b/drivers/gpu/drm/i915/display/intel_vdsc.c
new file mode 100644
index 000000000..269f97923
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_vdsc.c
@@ -0,0 +1,1218 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2018 Intel Corporation
+ *
+ * Author: Gaurav K Singh <gaurav.k.singh@intel.com>
+ *         Manasi Navare <manasi.d.navare@intel.com>
+ */
+#include <linux/limits.h>
+
+#include <drm/display/drm_dsc_helper.h>
+
+#include "i915_drv.h"
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dsi.h"
+#include "intel_qp_tables.h"
+#include "intel_vdsc.h"
+
+enum ROW_INDEX_BPP {
+	ROW_INDEX_6BPP = 0,
+	ROW_INDEX_8BPP,
+	ROW_INDEX_10BPP,
+	ROW_INDEX_12BPP,
+	ROW_INDEX_15BPP,
+	MAX_ROW_INDEX
+};
+
+enum COLUMN_INDEX_BPC {
+	COLUMN_INDEX_8BPC = 0,
+	COLUMN_INDEX_10BPC,
+	COLUMN_INDEX_12BPC,
+	COLUMN_INDEX_14BPC,
+	COLUMN_INDEX_16BPC,
+	MAX_COLUMN_INDEX
+};
+
+/* From DSC_v1.11 spec, rc_parameter_Set syntax element typically constant */
+static const u16 rc_buf_thresh[] = {
+	896, 1792, 2688, 3584, 4480, 5376, 6272, 6720, 7168, 7616,
+	7744, 7872, 8000, 8064
+};
+
+struct rc_parameters {
+	u16 initial_xmit_delay;
+	u8 first_line_bpg_offset;
+	u16 initial_offset;
+	u8 flatness_min_qp;
+	u8 flatness_max_qp;
+	u8 rc_quant_incr_limit0;
+	u8 rc_quant_incr_limit1;
+	struct drm_dsc_rc_range_parameters rc_range_params[DSC_NUM_BUF_RANGES];
+};
+
+/*
+ * Selected Rate Control Related Parameter Recommended Values
+ * from DSC_v1.11 spec & C Model release: DSC_model_20161212
+ */
+static const struct rc_parameters rc_parameters[][MAX_COLUMN_INDEX] = {
+{
+	/* 6BPP/8BPC */
+	{ 768, 15, 6144, 3, 13, 11, 11, {
+		{ 0, 4, 0 }, { 1, 6, -2 }, { 3, 8, -2 }, { 4, 8, -4 },
+		{ 5, 9, -6 }, { 5, 9, -6 }, { 6, 9, -6 }, { 6, 10, -8 },
+		{ 7, 11, -8 }, { 8, 12, -10 }, { 9, 12, -10 }, { 10, 12, -12 },
+		{ 10, 12, -12 }, { 11, 12, -12 }, { 13, 14, -12 }
+		}
+	},
+	/* 6BPP/10BPC */
+	{ 768, 15, 6144, 7, 17, 15, 15, {
+		{ 0, 8, 0 }, { 3, 10, -2 }, { 7, 12, -2 }, { 8, 12, -4 },
+		{ 9, 13, -6 }, { 9, 13, -6 }, { 10, 13, -6 }, { 10, 14, -8 },
+		{ 11, 15, -8 }, { 12, 16, -10 }, { 13, 16, -10 },
+		{ 14, 16, -12 }, { 14, 16, -12 }, { 15, 16, -12 },
+		{ 17, 18, -12 }
+		}
+	},
+	/* 6BPP/12BPC */
+	{ 768, 15, 6144, 11, 21, 19, 19, {
+		{ 0, 12, 0 }, { 5, 14, -2 }, { 11, 16, -2 }, { 12, 16, -4 },
+		{ 13, 17, -6 }, { 13, 17, -6 }, { 14, 17, -6 }, { 14, 18, -8 },
+		{ 15, 19, -8 }, { 16, 20, -10 }, { 17, 20, -10 },
+		{ 18, 20, -12 }, { 18, 20, -12 }, { 19, 20, -12 },
+		{ 21, 22, -12 }
+		}
+	},
+	/* 6BPP/14BPC */
+	{ 768, 15, 6144, 15, 25, 23, 27, {
+		{ 0, 16, 0 }, { 7, 18, -2 }, { 15, 20, -2 }, { 16, 20, -4 },
+		{ 17, 21, -6 }, { 17, 21, -6 }, { 18, 21, -6 }, { 18, 22, -8 },
+		{ 19, 23, -8 }, { 20, 24, -10 }, { 21, 24, -10 },
+		{ 22, 24, -12 }, { 22, 24, -12 }, { 23, 24, -12 },
+		{ 25, 26, -12 }
+		}
+	},
+	/* 6BPP/16BPC */
+	{ 768, 15, 6144, 19, 29, 27, 27, {
+		{ 0, 20, 0 }, { 9, 22, -2 }, { 19, 24, -2 }, { 20, 24, -4 },
+		{ 21, 25, -6 }, { 21, 25, -6 }, { 22, 25, -6 }, { 22, 26, -8 },
+		{ 23, 27, -8 }, { 24, 28, -10 }, { 25, 28, -10 },
+		{ 26, 28, -12 }, { 26, 28, -12 }, { 27, 28, -12 },
+		{ 29, 30, -12 }
+		}
+	},
+},
+{
+	/* 8BPP/8BPC */
+	{ 512, 12, 6144, 3, 12, 11, 11, {
+		{ 0, 4, 2 }, { 0, 4, 0 }, { 1, 5, 0 }, { 1, 6, -2 },
+		{ 3, 7, -4 }, { 3, 7, -6 }, { 3, 7, -8 }, { 3, 8, -8 },
+		{ 3, 9, -8 }, { 3, 10, -10 }, { 5, 11, -10 }, { 5, 12, -12 },
+		{ 5, 13, -12 }, { 7, 13, -12 }, { 13, 15, -12 }
+		}
+	},
+	/* 8BPP/10BPC */
+	{ 512, 12, 6144, 7, 16, 15, 15, {
+		{ 0, 4, 2 }, { 4, 8, 0 }, { 5, 9, 0 }, { 5, 10, -2 },
+		{ 7, 11, -4 }, { 7, 11, -6 }, { 7, 11, -8 }, { 7, 12, -8 },
+		{ 7, 13, -8 }, { 7, 14, -10 }, { 9, 15, -10 }, { 9, 16, -12 },
+		{ 9, 17, -12 }, { 11, 17, -12 }, { 17, 19, -12 }
+		}
+	},
+	/* 8BPP/12BPC */
+	{ 512, 12, 6144, 11, 20, 19, 19, {
+		{ 0, 12, 2 }, { 4, 12, 0 }, { 9, 13, 0 }, { 9, 14, -2 },
+		{ 11, 15, -4 }, { 11, 15, -6 }, { 11, 15, -8 }, { 11, 16, -8 },
+		{ 11, 17, -8 }, { 11, 18, -10 }, { 13, 19, -10 },
+		{ 13, 20, -12 }, { 13, 21, -12 }, { 15, 21, -12 },
+		{ 21, 23, -12 }
+		}
+	},
+	/* 8BPP/14BPC */
+	{ 512, 12, 6144, 15, 24, 23, 23, {
+		{ 0, 12, 0 }, { 5, 13, 0 }, { 11, 15, 0 }, { 12, 17, -2 },
+		{ 15, 19, -4 }, { 15, 19, -6 }, { 15, 19, -8 }, { 15, 20, -8 },
+		{ 15, 21, -8 }, { 15, 22, -10 }, { 17, 22, -10 },
+		{ 17, 23, -12 }, { 17, 23, -12 }, { 21, 24, -12 },
+		{ 24, 25, -12 }
+		}
+	},
+	/* 8BPP/16BPC */
+	{ 512, 12, 6144, 19, 28, 27, 27, {
+		{ 0, 12, 2 }, { 6, 14, 0 }, { 13, 17, 0 }, { 15, 20, -2 },
+		{ 19, 23, -4 }, { 19, 23, -6 }, { 19, 23, -8 }, { 19, 24, -8 },
+		{ 19, 25, -8 }, { 19, 26, -10 }, { 21, 26, -10 },
+		{ 21, 27, -12 }, { 21, 27, -12 }, { 25, 28, -12 },
+		{ 28, 29, -12 }
+		}
+	},
+},
+{
+	/* 10BPP/8BPC */
+	{ 410, 15, 5632, 3, 12, 11, 11, {
+		{ 0, 3, 2 }, { 0, 4, 0 }, { 1, 5, 0 }, { 2, 6, -2 },
+		{ 3, 7, -4 }, { 3, 7, -6 }, { 3, 7, -8 }, { 3, 8, -8 },
+		{ 3, 9, -8 }, { 3, 9, -10 }, { 5, 10, -10 }, { 5, 10, -10 },
+		{ 5, 11, -12 }, { 7, 11, -12 }, { 11, 12, -12 }
+		}
+	},
+	/* 10BPP/10BPC */
+	{ 410, 15, 5632, 7, 16, 15, 15, {
+		{ 0, 7, 2 }, { 4, 8, 0 }, { 5, 9, 0 }, { 6, 10, -2 },
+		{ 7, 11, -4 }, { 7, 11, -6 }, { 7, 11, -8 }, { 7, 12, -8 },
+		{ 7, 13, -8 }, { 7, 13, -10 }, { 9, 14, -10 }, { 9, 14, -10 },
+		{ 9, 15, -12 }, { 11, 15, -12 }, { 15, 16, -12 }
+		}
+	},
+	/* 10BPP/12BPC */
+	{ 410, 15, 5632, 11, 20, 19, 19, {
+		{ 0, 11, 2 }, { 4, 12, 0 }, { 9, 13, 0 }, { 10, 14, -2 },
+		{ 11, 15, -4 }, { 11, 15, -6 }, { 11, 15, -8 }, { 11, 16, -8 },
+		{ 11, 17, -8 }, { 11, 17, -10 }, { 13, 18, -10 },
+		{ 13, 18, -10 }, { 13, 19, -12 }, { 15, 19, -12 },
+		{ 19, 20, -12 }
+		}
+	},
+	/* 10BPP/14BPC */
+	{ 410, 15, 5632, 15, 24, 23, 23, {
+		{ 0, 11, 2 }, { 5, 13, 0 }, { 11, 15, 0 }, { 13, 18, -2 },
+		{ 15, 19, -4 }, { 15, 19, -6 }, { 15, 19, -8 }, { 15, 20, -8 },
+		{ 15, 21, -8 }, { 15, 21, -10 }, { 17, 22, -10 },
+		{ 17, 22, -10 }, { 17, 23, -12 }, { 19, 23, -12 },
+		{ 23, 24, -12 }
+		}
+	},
+	/* 10BPP/16BPC */
+	{ 410, 15, 5632, 19, 28, 27, 27, {
+		{ 0, 11, 2 }, { 6, 14, 0 }, { 13, 17, 0 }, { 16, 20, -2 },
+		{ 19, 23, -4 }, { 19, 23, -6 }, { 19, 23, -8 }, { 19, 24, -8 },
+		{ 19, 25, -8 }, { 19, 25, -10 }, { 21, 26, -10 },
+		{ 21, 26, -10 }, { 21, 27, -12 }, { 23, 27, -12 },
+		{ 27, 28, -12 }
+		}
+	},
+},
+{
+	/* 12BPP/8BPC */
+	{ 341, 15, 2048, 3, 12, 11, 11, {
+		{ 0, 2, 2 }, { 0, 4, 0 }, { 1, 5, 0 }, { 1, 6, -2 },
+		{ 3, 7, -4 }, { 3, 7, -6 }, { 3, 7, -8 }, { 3, 8, -8 },
+		{ 3, 9, -8 }, { 3, 10, -10 }, { 5, 11, -10 },
+		{ 5, 12, -12 }, { 5, 13, -12 }, { 7, 13, -12 }, { 13, 15, -12 }
+		}
+	},
+	/* 12BPP/10BPC */
+	{ 341, 15, 2048, 7, 16, 15, 15, {
+		{ 0, 2, 2 }, { 2, 5, 0 }, { 3, 7, 0 }, { 4, 8, -2 },
+		{ 6, 9, -4 }, { 7, 10, -6 }, { 7, 11, -8 }, { 7, 12, -8 },
+		{ 7, 13, -8 }, { 7, 14, -10 }, { 9, 15, -10 }, { 9, 16, -12 },
+		{ 9, 17, -12 }, { 11, 17, -12 }, { 17, 19, -12 }
+		}
+	},
+	/* 12BPP/12BPC */
+	{ 341, 15, 2048, 11, 20, 19, 19, {
+		{ 0, 6, 2 }, { 4, 9, 0 }, { 7, 11, 0 }, { 8, 12, -2 },
+		{ 10, 13, -4 }, { 11, 14, -6 }, { 11, 15, -8 }, { 11, 16, -8 },
+		{ 11, 17, -8 }, { 11, 18, -10 }, { 13, 19, -10 },
+		{ 13, 20, -12 }, { 13, 21, -12 }, { 15, 21, -12 },
+		{ 21, 23, -12 }
+		}
+	},
+	/* 12BPP/14BPC */
+	{ 341, 15, 2048, 15, 24, 23, 23, {
+		{ 0, 6, 2 }, { 7, 10, 0 }, { 9, 13, 0 }, { 11, 16, -2 },
+		{ 14, 17, -4 }, { 15, 18, -6 }, { 15, 19, -8 }, { 15, 20, -8 },
+		{ 15, 20, -8 }, { 15, 21, -10 }, { 17, 21, -10 },
+		{ 17, 21, -12 }, { 17, 21, -12 }, { 19, 22, -12 },
+		{ 22, 23, -12 }
+		}
+	},
+	/* 12BPP/16BPC */
+	{ 341, 15, 2048, 19, 28, 27, 27, {
+		{ 0, 6, 2 }, { 6, 11, 0 }, { 11, 15, 0 }, { 14, 18, -2 },
+		{ 18, 21, -4 }, { 19, 22, -6 }, { 19, 23, -8 }, { 19, 24, -8 },
+		{ 19, 24, -8 }, { 19, 25, -10 }, { 21, 25, -10 },
+		{ 21, 25, -12 }, { 21, 25, -12 }, { 23, 26, -12 },
+		{ 26, 27, -12 }
+		}
+	},
+},
+{
+	/* 15BPP/8BPC */
+	{ 273, 15, 2048, 3, 12, 11, 11, {
+		{ 0, 0, 10 }, { 0, 1, 8 }, { 0, 1, 6 }, { 0, 2, 4 },
+		{ 1, 2, 2 }, { 1, 3, 0 }, { 1, 3, -2 }, { 2, 4, -4 },
+		{ 2, 5, -6 }, { 3, 5, -8 }, { 4, 6, -10 }, { 4, 7, -10 },
+		{ 5, 7, -12 }, { 7, 8, -12 }, { 8, 9, -12 }
+		}
+	},
+	/* 15BPP/10BPC */
+	{ 273, 15, 2048, 7, 16, 15, 15, {
+		{ 0, 2, 10 }, { 2, 5, 8 }, { 3, 5, 6 }, { 4, 6, 4 },
+		{ 5, 6, 2 }, { 5, 7, 0 }, { 5, 7, -2 }, { 6, 8, -4 },
+		{ 6, 9, -6 }, { 7, 9, -8 }, { 8, 10, -10 }, { 8, 11, -10 },
+		{ 9, 11, -12 }, { 11, 12, -12 }, { 12, 13, -12 }
+		}
+	},
+	/* 15BPP/12BPC */
+	{ 273, 15, 2048, 11, 20, 19, 19, {
+		{ 0, 4, 10 }, { 2, 7, 8 }, { 4, 9, 6 }, { 6, 11, 4 },
+		{ 9, 11, 2 }, { 9, 11, 0 }, { 9, 12, -2 }, { 10, 12, -4 },
+		{ 11, 13, -6 }, { 11, 13, -8 }, { 12, 14, -10 },
+		{ 13, 15, -10 }, { 13, 15, -12 }, { 15, 16, -12 },
+		{ 16, 17, -12 }
+		}
+	},
+	/* 15BPP/14BPC */
+	{ 273, 15, 2048, 15, 24, 23, 23, {
+		{ 0, 4, 10 }, { 3, 8, 8 }, { 6, 11, 6 }, { 9, 14, 4 },
+		{ 13, 15, 2 }, { 13, 15, 0 }, { 13, 16, -2 }, { 14, 16, -4 },
+		{ 15, 17, -6 }, { 15, 17, -8 }, { 16, 18, -10 },
+		{ 17, 19, -10 }, { 17, 19, -12 }, { 19, 20, -12 },
+		{ 20, 21, -12 }
+		}
+	},
+	/* 15BPP/16BPC */
+	{ 273, 15, 2048, 19, 28, 27, 27, {
+		{ 0, 4, 10 }, { 4, 9, 8 }, { 8, 13, 6 }, { 12, 17, 4 },
+		{ 17, 19, 2 }, { 17, 20, 0 }, { 17, 20, -2 }, { 18, 20, -4 },
+		{ 19, 21, -6 }, { 19, 21, -8 }, { 20, 22, -10 },
+		{ 21, 23, -10 }, { 21, 23, -12 }, { 23, 24, -12 },
+		{ 24, 25, -12 }
+		}
+	}
+}
+
+};
+
+static int get_row_index_for_rc_params(u16 compressed_bpp)
+{
+	switch (compressed_bpp) {
+	case 6:
+		return ROW_INDEX_6BPP;
+	case 8:
+		return ROW_INDEX_8BPP;
+	case 10:
+		return ROW_INDEX_10BPP;
+	case 12:
+		return ROW_INDEX_12BPP;
+	case 15:
+		return ROW_INDEX_15BPP;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int get_column_index_for_rc_params(u8 bits_per_component)
+{
+	switch (bits_per_component) {
+	case 8:
+		return COLUMN_INDEX_8BPC;
+	case 10:
+		return COLUMN_INDEX_10BPC;
+	case 12:
+		return COLUMN_INDEX_12BPC;
+	case 14:
+		return COLUMN_INDEX_14BPC;
+	case 16:
+		return COLUMN_INDEX_16BPC;
+	default:
+		return -EINVAL;
+	}
+}
+
+static const struct rc_parameters *get_rc_params(u16 compressed_bpp,
+						 u8 bits_per_component)
+{
+	int row_index, column_index;
+
+	row_index = get_row_index_for_rc_params(compressed_bpp);
+	if (row_index < 0)
+		return NULL;
+
+	column_index = get_column_index_for_rc_params(bits_per_component);
+	if (column_index < 0)
+		return NULL;
+
+	return &rc_parameters[row_index][column_index];
+}
+
+bool intel_dsc_source_support(const struct intel_crtc_state *crtc_state)
+{
+	const struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	if (!RUNTIME_INFO(i915)->has_dsc)
+		return false;
+
+	if (DISPLAY_VER(i915) >= 12)
+		return true;
+
+	if (DISPLAY_VER(i915) >= 11 && cpu_transcoder != TRANSCODER_A)
+		return true;
+
+	return false;
+}
+
+static bool is_pipe_dsc(struct intel_crtc *crtc, enum transcoder cpu_transcoder)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	if (DISPLAY_VER(i915) >= 12)
+		return true;
+
+	if (cpu_transcoder == TRANSCODER_EDP ||
+	    cpu_transcoder == TRANSCODER_DSI_0 ||
+	    cpu_transcoder == TRANSCODER_DSI_1)
+		return false;
+
+	/* There's no pipe A DSC engine on ICL */
+	drm_WARN_ON(&i915->drm, crtc->pipe == PIPE_A);
+
+	return true;
+}
+
+static void
+calculate_rc_params(struct rc_parameters *rc,
+		    struct drm_dsc_config *vdsc_cfg)
+{
+	int bpc = vdsc_cfg->bits_per_component;
+	int bpp = vdsc_cfg->bits_per_pixel >> 4;
+	static const s8 ofs_und6[] = {
+		0, -2, -2, -4, -6, -6, -8, -8, -8, -10, -10, -12, -12, -12, -12
+	};
+	static const s8 ofs_und8[] = {
+		2, 0, 0, -2, -4, -6, -8, -8, -8, -10, -10, -10, -12, -12, -12
+	};
+	static const s8 ofs_und12[] = {
+		2, 0, 0, -2, -4, -6, -8, -8, -8, -10, -10, -10, -12, -12, -12
+	};
+	static const s8 ofs_und15[] = {
+		10, 8, 6, 4, 2, 0, -2, -4, -6, -8, -10, -10, -12, -12, -12
+	};
+	int qp_bpc_modifier = (bpc - 8) * 2;
+	u32 res, buf_i, bpp_i;
+
+	if (vdsc_cfg->slice_height >= 8)
+		rc->first_line_bpg_offset =
+			12 + DIV_ROUND_UP((9 * min(34, vdsc_cfg->slice_height - 8)), 100);
+	else
+		rc->first_line_bpg_offset = 2 * (vdsc_cfg->slice_height - 1);
+
+	/* Our hw supports only 444 modes as of today */
+	if (bpp >= 12)
+		rc->initial_offset = 2048;
+	else if (bpp >= 10)
+		rc->initial_offset = 5632 - DIV_ROUND_UP(((bpp - 10) * 3584), 2);
+	else if (bpp >= 8)
+		rc->initial_offset = 6144 - DIV_ROUND_UP(((bpp - 8) * 512), 2);
+	else
+		rc->initial_offset = 6144;
+
+	/* initial_xmit_delay = rc_model_size/2/compression_bpp */
+	rc->initial_xmit_delay = DIV_ROUND_UP(DSC_RC_MODEL_SIZE_CONST, 2 * bpp);
+
+	rc->flatness_min_qp = 3 + qp_bpc_modifier;
+	rc->flatness_max_qp = 12 + qp_bpc_modifier;
+
+	rc->rc_quant_incr_limit0 = 11 + qp_bpc_modifier;
+	rc->rc_quant_incr_limit1 = 11 + qp_bpc_modifier;
+
+	bpp_i  = (2 * (bpp - 6));
+	for (buf_i = 0; buf_i < DSC_NUM_BUF_RANGES; buf_i++) {
+		/* Read range_minqp and range_max_qp from qp tables */
+		rc->rc_range_params[buf_i].range_min_qp =
+			intel_lookup_range_min_qp(bpc, buf_i, bpp_i);
+		rc->rc_range_params[buf_i].range_max_qp =
+			intel_lookup_range_max_qp(bpc, buf_i, bpp_i);
+
+		/* Calculate range_bgp_offset */
+		if (bpp <= 6) {
+			rc->rc_range_params[buf_i].range_bpg_offset = ofs_und6[buf_i];
+		} else if (bpp <= 8) {
+			res = DIV_ROUND_UP(((bpp - 6) * (ofs_und8[buf_i] - ofs_und6[buf_i])), 2);
+			rc->rc_range_params[buf_i].range_bpg_offset =
+								ofs_und6[buf_i] + res;
+		} else if (bpp <= 12) {
+			rc->rc_range_params[buf_i].range_bpg_offset =
+								ofs_und8[buf_i];
+		} else if (bpp <= 15) {
+			res = DIV_ROUND_UP(((bpp - 12) * (ofs_und15[buf_i] - ofs_und12[buf_i])), 3);
+			rc->rc_range_params[buf_i].range_bpg_offset =
+								ofs_und12[buf_i] + res;
+		} else {
+			rc->rc_range_params[buf_i].range_bpg_offset =
+								ofs_und15[buf_i];
+		}
+	}
+}
+
+int intel_dsc_compute_params(struct intel_crtc_state *pipe_config)
+{
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct drm_dsc_config *vdsc_cfg = &pipe_config->dsc.config;
+	u16 compressed_bpp = pipe_config->dsc.compressed_bpp;
+	const struct rc_parameters *rc_params;
+	struct rc_parameters *rc = NULL;
+	u8 i = 0;
+
+	vdsc_cfg->pic_width = pipe_config->hw.adjusted_mode.crtc_hdisplay;
+	vdsc_cfg->slice_width = DIV_ROUND_UP(vdsc_cfg->pic_width,
+					     pipe_config->dsc.slice_count);
+
+	/* Gen 11 does not support YCbCr */
+	vdsc_cfg->simple_422 = false;
+	/* Gen 11 does not support VBR */
+	vdsc_cfg->vbr_enable = false;
+
+	/* Gen 11 only supports integral values of bpp */
+	vdsc_cfg->bits_per_pixel = compressed_bpp << 4;
+	vdsc_cfg->bits_per_component = pipe_config->pipe_bpp / 3;
+
+	for (i = 0; i < DSC_NUM_BUF_RANGES - 1; i++) {
+		/*
+		 * six 0s are appended to the lsb of each threshold value
+		 * internally in h/w.
+		 * Only 8 bits are allowed for programming RcBufThreshold
+		 */
+		vdsc_cfg->rc_buf_thresh[i] = rc_buf_thresh[i] >> 6;
+	}
+
+	/*
+	 * For 6bpp, RC Buffer threshold 12 and 13 need a different value
+	 * as per C Model
+	 */
+	if (compressed_bpp == 6) {
+		vdsc_cfg->rc_buf_thresh[12] = 0x7C;
+		vdsc_cfg->rc_buf_thresh[13] = 0x7D;
+	}
+
+	/*
+	 * From XE_LPD onwards we supports compression bpps in steps of 1
+	 * upto uncompressed bpp-1, hence add calculations for all the rc
+	 * parameters
+	 */
+	if (DISPLAY_VER(dev_priv) >= 13) {
+		rc = kmalloc(sizeof(*rc), GFP_KERNEL);
+		if (!rc)
+			return -ENOMEM;
+
+		calculate_rc_params(rc, vdsc_cfg);
+		rc_params = rc;
+	} else {
+		rc_params = get_rc_params(compressed_bpp,
+					  vdsc_cfg->bits_per_component);
+		if (!rc_params)
+			return -EINVAL;
+	}
+
+	vdsc_cfg->first_line_bpg_offset = rc_params->first_line_bpg_offset;
+	vdsc_cfg->initial_xmit_delay = rc_params->initial_xmit_delay;
+	vdsc_cfg->initial_offset = rc_params->initial_offset;
+	vdsc_cfg->flatness_min_qp = rc_params->flatness_min_qp;
+	vdsc_cfg->flatness_max_qp = rc_params->flatness_max_qp;
+	vdsc_cfg->rc_quant_incr_limit0 = rc_params->rc_quant_incr_limit0;
+	vdsc_cfg->rc_quant_incr_limit1 = rc_params->rc_quant_incr_limit1;
+
+	for (i = 0; i < DSC_NUM_BUF_RANGES; i++) {
+		vdsc_cfg->rc_range_params[i].range_min_qp =
+			rc_params->rc_range_params[i].range_min_qp;
+		vdsc_cfg->rc_range_params[i].range_max_qp =
+			rc_params->rc_range_params[i].range_max_qp;
+		/*
+		 * Range BPG Offset uses 2's complement and is only a 6 bits. So
+		 * mask it to get only 6 bits.
+		 */
+		vdsc_cfg->rc_range_params[i].range_bpg_offset =
+			rc_params->rc_range_params[i].range_bpg_offset &
+			DSC_RANGE_BPG_OFFSET_MASK;
+	}
+
+	/*
+	 * BitsPerComponent value determines mux_word_size:
+	 * When BitsPerComponent is less than or 10bpc, muxWordSize will be equal to
+	 * 48 bits otherwise 64
+	 */
+	if (vdsc_cfg->bits_per_component <= 10)
+		vdsc_cfg->mux_word_size = DSC_MUX_WORD_SIZE_8_10_BPC;
+	else
+		vdsc_cfg->mux_word_size = DSC_MUX_WORD_SIZE_12_BPC;
+
+	/* InitialScaleValue is a 6 bit value with 3 fractional bits (U3.3) */
+	vdsc_cfg->initial_scale_value = (vdsc_cfg->rc_model_size << 3) /
+		(vdsc_cfg->rc_model_size - vdsc_cfg->initial_offset);
+
+	kfree(rc);
+
+	return 0;
+}
+
+enum intel_display_power_domain
+intel_dsc_power_domain(struct intel_crtc *crtc, enum transcoder cpu_transcoder)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	/*
+	 * VDSC/joining uses a separate power well, PW2, and requires
+	 * POWER_DOMAIN_TRANSCODER_VDSC_PW2 power domain in two cases:
+	 *
+	 *  - ICL eDP/DSI transcoder
+	 *  - Display version 12 (except RKL) pipe A
+	 *
+	 * For any other pipe, VDSC/joining uses the power well associated with
+	 * the pipe in use. Hence another reference on the pipe power domain
+	 * will suffice. (Except no VDSC/joining on ICL pipe A.)
+	 */
+	if (DISPLAY_VER(i915) == 12 && !IS_ROCKETLAKE(i915) && pipe == PIPE_A)
+		return POWER_DOMAIN_TRANSCODER_VDSC_PW2;
+	else if (is_pipe_dsc(crtc, cpu_transcoder))
+		return POWER_DOMAIN_PIPE(pipe);
+	else
+		return POWER_DOMAIN_TRANSCODER_VDSC_PW2;
+}
+
+static void intel_dsc_pps_configure(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_dsc_config *vdsc_cfg = &crtc_state->dsc.config;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	enum pipe pipe = crtc->pipe;
+	u32 pps_val = 0;
+	u32 rc_buf_thresh_dword[4];
+	u32 rc_range_params_dword[8];
+	u8 num_vdsc_instances = (crtc_state->dsc.dsc_split) ? 2 : 1;
+	int i = 0;
+
+	if (crtc_state->bigjoiner_pipes)
+		num_vdsc_instances *= 2;
+
+	/* Populate PICTURE_PARAMETER_SET_0 registers */
+	pps_val = DSC_VER_MAJ | vdsc_cfg->dsc_version_minor <<
+		DSC_VER_MIN_SHIFT |
+		vdsc_cfg->bits_per_component << DSC_BPC_SHIFT |
+		vdsc_cfg->line_buf_depth << DSC_LINE_BUF_DEPTH_SHIFT;
+	if (vdsc_cfg->dsc_version_minor == 2)
+		pps_val |= DSC_ALT_ICH_SEL;
+	if (vdsc_cfg->block_pred_enable)
+		pps_val |= DSC_BLOCK_PREDICTION;
+	if (vdsc_cfg->convert_rgb)
+		pps_val |= DSC_COLOR_SPACE_CONVERSION;
+	if (vdsc_cfg->simple_422)
+		pps_val |= DSC_422_ENABLE;
+	if (vdsc_cfg->vbr_enable)
+		pps_val |= DSC_VBR_ENABLE;
+	drm_dbg_kms(&dev_priv->drm, "PPS0 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_0,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv, DSCC_PICTURE_PARAMETER_SET_0,
+				       pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_0(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_0(pipe),
+				       pps_val);
+	}
+
+	/* Populate PICTURE_PARAMETER_SET_1 registers */
+	pps_val = 0;
+	pps_val |= DSC_BPP(vdsc_cfg->bits_per_pixel);
+	drm_dbg_kms(&dev_priv->drm, "PPS1 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_1,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv, DSCC_PICTURE_PARAMETER_SET_1,
+				       pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_1(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_1(pipe),
+				       pps_val);
+	}
+
+	/* Populate PICTURE_PARAMETER_SET_2 registers */
+	pps_val = 0;
+	pps_val |= DSC_PIC_HEIGHT(vdsc_cfg->pic_height) |
+		DSC_PIC_WIDTH(vdsc_cfg->pic_width / num_vdsc_instances);
+	drm_dbg_kms(&dev_priv->drm, "PPS2 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_2,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv, DSCC_PICTURE_PARAMETER_SET_2,
+				       pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_2(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_2(pipe),
+				       pps_val);
+	}
+
+	/* Populate PICTURE_PARAMETER_SET_3 registers */
+	pps_val = 0;
+	pps_val |= DSC_SLICE_HEIGHT(vdsc_cfg->slice_height) |
+		DSC_SLICE_WIDTH(vdsc_cfg->slice_width);
+	drm_dbg_kms(&dev_priv->drm, "PPS3 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_3,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv, DSCC_PICTURE_PARAMETER_SET_3,
+				       pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_3(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_3(pipe),
+				       pps_val);
+	}
+
+	/* Populate PICTURE_PARAMETER_SET_4 registers */
+	pps_val = 0;
+	pps_val |= DSC_INITIAL_XMIT_DELAY(vdsc_cfg->initial_xmit_delay) |
+		DSC_INITIAL_DEC_DELAY(vdsc_cfg->initial_dec_delay);
+	drm_dbg_kms(&dev_priv->drm, "PPS4 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_4,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv, DSCC_PICTURE_PARAMETER_SET_4,
+				       pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_4(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_4(pipe),
+				       pps_val);
+	}
+
+	/* Populate PICTURE_PARAMETER_SET_5 registers */
+	pps_val = 0;
+	pps_val |= DSC_SCALE_INC_INT(vdsc_cfg->scale_increment_interval) |
+		DSC_SCALE_DEC_INT(vdsc_cfg->scale_decrement_interval);
+	drm_dbg_kms(&dev_priv->drm, "PPS5 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_5,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv, DSCC_PICTURE_PARAMETER_SET_5,
+				       pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_5(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_5(pipe),
+				       pps_val);
+	}
+
+	/* Populate PICTURE_PARAMETER_SET_6 registers */
+	pps_val = 0;
+	pps_val |= DSC_INITIAL_SCALE_VALUE(vdsc_cfg->initial_scale_value) |
+		DSC_FIRST_LINE_BPG_OFFSET(vdsc_cfg->first_line_bpg_offset) |
+		DSC_FLATNESS_MIN_QP(vdsc_cfg->flatness_min_qp) |
+		DSC_FLATNESS_MAX_QP(vdsc_cfg->flatness_max_qp);
+	drm_dbg_kms(&dev_priv->drm, "PPS6 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_6,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv, DSCC_PICTURE_PARAMETER_SET_6,
+				       pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_6(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_6(pipe),
+				       pps_val);
+	}
+
+	/* Populate PICTURE_PARAMETER_SET_7 registers */
+	pps_val = 0;
+	pps_val |= DSC_SLICE_BPG_OFFSET(vdsc_cfg->slice_bpg_offset) |
+		DSC_NFL_BPG_OFFSET(vdsc_cfg->nfl_bpg_offset);
+	drm_dbg_kms(&dev_priv->drm, "PPS7 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_7,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv, DSCC_PICTURE_PARAMETER_SET_7,
+				       pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_7(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_7(pipe),
+				       pps_val);
+	}
+
+	/* Populate PICTURE_PARAMETER_SET_8 registers */
+	pps_val = 0;
+	pps_val |= DSC_FINAL_OFFSET(vdsc_cfg->final_offset) |
+		DSC_INITIAL_OFFSET(vdsc_cfg->initial_offset);
+	drm_dbg_kms(&dev_priv->drm, "PPS8 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_8,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv, DSCC_PICTURE_PARAMETER_SET_8,
+				       pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_8(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_8(pipe),
+				       pps_val);
+	}
+
+	/* Populate PICTURE_PARAMETER_SET_9 registers */
+	pps_val = 0;
+	pps_val |= DSC_RC_MODEL_SIZE(vdsc_cfg->rc_model_size) |
+		DSC_RC_EDGE_FACTOR(DSC_RC_EDGE_FACTOR_CONST);
+	drm_dbg_kms(&dev_priv->drm, "PPS9 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_9,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv, DSCC_PICTURE_PARAMETER_SET_9,
+				       pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_9(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_9(pipe),
+				       pps_val);
+	}
+
+	/* Populate PICTURE_PARAMETER_SET_10 registers */
+	pps_val = 0;
+	pps_val |= DSC_RC_QUANT_INC_LIMIT0(vdsc_cfg->rc_quant_incr_limit0) |
+		DSC_RC_QUANT_INC_LIMIT1(vdsc_cfg->rc_quant_incr_limit1) |
+		DSC_RC_TARGET_OFF_HIGH(DSC_RC_TGT_OFFSET_HI_CONST) |
+		DSC_RC_TARGET_OFF_LOW(DSC_RC_TGT_OFFSET_LO_CONST);
+	drm_dbg_kms(&dev_priv->drm, "PPS10 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_10,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       DSCC_PICTURE_PARAMETER_SET_10, pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_10(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_10(pipe),
+				       pps_val);
+	}
+
+	/* Populate Picture parameter set 16 */
+	pps_val = 0;
+	pps_val |= DSC_SLICE_CHUNK_SIZE(vdsc_cfg->slice_chunk_size) |
+		DSC_SLICE_PER_LINE((vdsc_cfg->pic_width / num_vdsc_instances) /
+				   vdsc_cfg->slice_width) |
+		DSC_SLICE_ROW_PER_FRAME(vdsc_cfg->pic_height /
+					vdsc_cfg->slice_height);
+	drm_dbg_kms(&dev_priv->drm, "PPS16 = 0x%08x\n", pps_val);
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_PICTURE_PARAMETER_SET_16,
+			       pps_val);
+		/*
+		 * If 2 VDSC instances are needed, configure PPS for second
+		 * VDSC
+		 */
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       DSCC_PICTURE_PARAMETER_SET_16, pps_val);
+	} else {
+		intel_de_write(dev_priv,
+			       ICL_DSC0_PICTURE_PARAMETER_SET_16(pipe),
+			       pps_val);
+		if (crtc_state->dsc.dsc_split)
+			intel_de_write(dev_priv,
+				       ICL_DSC1_PICTURE_PARAMETER_SET_16(pipe),
+				       pps_val);
+	}
+
+	/* Populate the RC_BUF_THRESH registers */
+	memset(rc_buf_thresh_dword, 0, sizeof(rc_buf_thresh_dword));
+	for (i = 0; i < DSC_NUM_BUF_RANGES - 1; i++) {
+		rc_buf_thresh_dword[i / 4] |=
+			(u32)(vdsc_cfg->rc_buf_thresh[i] <<
+			      BITS_PER_BYTE * (i % 4));
+		drm_dbg_kms(&dev_priv->drm, "RC_BUF_THRESH_%d = 0x%08x\n", i,
+			    rc_buf_thresh_dword[i / 4]);
+	}
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_RC_BUF_THRESH_0,
+			       rc_buf_thresh_dword[0]);
+		intel_de_write(dev_priv, DSCA_RC_BUF_THRESH_0_UDW,
+			       rc_buf_thresh_dword[1]);
+		intel_de_write(dev_priv, DSCA_RC_BUF_THRESH_1,
+			       rc_buf_thresh_dword[2]);
+		intel_de_write(dev_priv, DSCA_RC_BUF_THRESH_1_UDW,
+			       rc_buf_thresh_dword[3]);
+		if (crtc_state->dsc.dsc_split) {
+			intel_de_write(dev_priv, DSCC_RC_BUF_THRESH_0,
+				       rc_buf_thresh_dword[0]);
+			intel_de_write(dev_priv, DSCC_RC_BUF_THRESH_0_UDW,
+				       rc_buf_thresh_dword[1]);
+			intel_de_write(dev_priv, DSCC_RC_BUF_THRESH_1,
+				       rc_buf_thresh_dword[2]);
+			intel_de_write(dev_priv, DSCC_RC_BUF_THRESH_1_UDW,
+				       rc_buf_thresh_dword[3]);
+		}
+	} else {
+		intel_de_write(dev_priv, ICL_DSC0_RC_BUF_THRESH_0(pipe),
+			       rc_buf_thresh_dword[0]);
+		intel_de_write(dev_priv, ICL_DSC0_RC_BUF_THRESH_0_UDW(pipe),
+			       rc_buf_thresh_dword[1]);
+		intel_de_write(dev_priv, ICL_DSC0_RC_BUF_THRESH_1(pipe),
+			       rc_buf_thresh_dword[2]);
+		intel_de_write(dev_priv, ICL_DSC0_RC_BUF_THRESH_1_UDW(pipe),
+			       rc_buf_thresh_dword[3]);
+		if (crtc_state->dsc.dsc_split) {
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_BUF_THRESH_0(pipe),
+				       rc_buf_thresh_dword[0]);
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_BUF_THRESH_0_UDW(pipe),
+				       rc_buf_thresh_dword[1]);
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_BUF_THRESH_1(pipe),
+				       rc_buf_thresh_dword[2]);
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_BUF_THRESH_1_UDW(pipe),
+				       rc_buf_thresh_dword[3]);
+		}
+	}
+
+	/* Populate the RC_RANGE_PARAMETERS registers */
+	memset(rc_range_params_dword, 0, sizeof(rc_range_params_dword));
+	for (i = 0; i < DSC_NUM_BUF_RANGES; i++) {
+		rc_range_params_dword[i / 2] |=
+			(u32)(((vdsc_cfg->rc_range_params[i].range_bpg_offset <<
+				RC_BPG_OFFSET_SHIFT) |
+			       (vdsc_cfg->rc_range_params[i].range_max_qp <<
+				RC_MAX_QP_SHIFT) |
+			       (vdsc_cfg->rc_range_params[i].range_min_qp <<
+				RC_MIN_QP_SHIFT)) << 16 * (i % 2));
+		drm_dbg_kms(&dev_priv->drm, "RC_RANGE_PARAM_%d = 0x%08x\n", i,
+			    rc_range_params_dword[i / 2]);
+	}
+	if (!is_pipe_dsc(crtc, cpu_transcoder)) {
+		intel_de_write(dev_priv, DSCA_RC_RANGE_PARAMETERS_0,
+			       rc_range_params_dword[0]);
+		intel_de_write(dev_priv, DSCA_RC_RANGE_PARAMETERS_0_UDW,
+			       rc_range_params_dword[1]);
+		intel_de_write(dev_priv, DSCA_RC_RANGE_PARAMETERS_1,
+			       rc_range_params_dword[2]);
+		intel_de_write(dev_priv, DSCA_RC_RANGE_PARAMETERS_1_UDW,
+			       rc_range_params_dword[3]);
+		intel_de_write(dev_priv, DSCA_RC_RANGE_PARAMETERS_2,
+			       rc_range_params_dword[4]);
+		intel_de_write(dev_priv, DSCA_RC_RANGE_PARAMETERS_2_UDW,
+			       rc_range_params_dword[5]);
+		intel_de_write(dev_priv, DSCA_RC_RANGE_PARAMETERS_3,
+			       rc_range_params_dword[6]);
+		intel_de_write(dev_priv, DSCA_RC_RANGE_PARAMETERS_3_UDW,
+			       rc_range_params_dword[7]);
+		if (crtc_state->dsc.dsc_split) {
+			intel_de_write(dev_priv, DSCC_RC_RANGE_PARAMETERS_0,
+				       rc_range_params_dword[0]);
+			intel_de_write(dev_priv,
+				       DSCC_RC_RANGE_PARAMETERS_0_UDW,
+				       rc_range_params_dword[1]);
+			intel_de_write(dev_priv, DSCC_RC_RANGE_PARAMETERS_1,
+				       rc_range_params_dword[2]);
+			intel_de_write(dev_priv,
+				       DSCC_RC_RANGE_PARAMETERS_1_UDW,
+				       rc_range_params_dword[3]);
+			intel_de_write(dev_priv, DSCC_RC_RANGE_PARAMETERS_2,
+				       rc_range_params_dword[4]);
+			intel_de_write(dev_priv,
+				       DSCC_RC_RANGE_PARAMETERS_2_UDW,
+				       rc_range_params_dword[5]);
+			intel_de_write(dev_priv, DSCC_RC_RANGE_PARAMETERS_3,
+				       rc_range_params_dword[6]);
+			intel_de_write(dev_priv,
+				       DSCC_RC_RANGE_PARAMETERS_3_UDW,
+				       rc_range_params_dword[7]);
+		}
+	} else {
+		intel_de_write(dev_priv, ICL_DSC0_RC_RANGE_PARAMETERS_0(pipe),
+			       rc_range_params_dword[0]);
+		intel_de_write(dev_priv,
+			       ICL_DSC0_RC_RANGE_PARAMETERS_0_UDW(pipe),
+			       rc_range_params_dword[1]);
+		intel_de_write(dev_priv, ICL_DSC0_RC_RANGE_PARAMETERS_1(pipe),
+			       rc_range_params_dword[2]);
+		intel_de_write(dev_priv,
+			       ICL_DSC0_RC_RANGE_PARAMETERS_1_UDW(pipe),
+			       rc_range_params_dword[3]);
+		intel_de_write(dev_priv, ICL_DSC0_RC_RANGE_PARAMETERS_2(pipe),
+			       rc_range_params_dword[4]);
+		intel_de_write(dev_priv,
+			       ICL_DSC0_RC_RANGE_PARAMETERS_2_UDW(pipe),
+			       rc_range_params_dword[5]);
+		intel_de_write(dev_priv, ICL_DSC0_RC_RANGE_PARAMETERS_3(pipe),
+			       rc_range_params_dword[6]);
+		intel_de_write(dev_priv,
+			       ICL_DSC0_RC_RANGE_PARAMETERS_3_UDW(pipe),
+			       rc_range_params_dword[7]);
+		if (crtc_state->dsc.dsc_split) {
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_RANGE_PARAMETERS_0(pipe),
+				       rc_range_params_dword[0]);
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_RANGE_PARAMETERS_0_UDW(pipe),
+				       rc_range_params_dword[1]);
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_RANGE_PARAMETERS_1(pipe),
+				       rc_range_params_dword[2]);
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_RANGE_PARAMETERS_1_UDW(pipe),
+				       rc_range_params_dword[3]);
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_RANGE_PARAMETERS_2(pipe),
+				       rc_range_params_dword[4]);
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_RANGE_PARAMETERS_2_UDW(pipe),
+				       rc_range_params_dword[5]);
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_RANGE_PARAMETERS_3(pipe),
+				       rc_range_params_dword[6]);
+			intel_de_write(dev_priv,
+				       ICL_DSC1_RC_RANGE_PARAMETERS_3_UDW(pipe),
+				       rc_range_params_dword[7]);
+		}
+	}
+}
+
+void intel_dsc_dsi_pps_write(struct intel_encoder *encoder,
+			     const struct intel_crtc_state *crtc_state)
+{
+	const struct drm_dsc_config *vdsc_cfg = &crtc_state->dsc.config;
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	struct mipi_dsi_device *dsi;
+	struct drm_dsc_picture_parameter_set pps;
+	enum port port;
+
+	if (!crtc_state->dsc.compression_enable)
+		return;
+
+	drm_dsc_pps_payload_pack(&pps, vdsc_cfg);
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		dsi = intel_dsi->dsi_hosts[port]->device;
+
+		mipi_dsi_picture_parameter_set(dsi, &pps);
+		mipi_dsi_compression_mode(dsi, true);
+	}
+}
+
+void intel_dsc_dp_pps_write(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state)
+{
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	const struct drm_dsc_config *vdsc_cfg = &crtc_state->dsc.config;
+	struct drm_dsc_pps_infoframe dp_dsc_pps_sdp;
+
+	if (!crtc_state->dsc.compression_enable)
+		return;
+
+	/* Prepare DP SDP PPS header as per DP 1.4 spec, Table 2-123 */
+	drm_dsc_dp_pps_header_init(&dp_dsc_pps_sdp.pps_header);
+
+	/* Fill the PPS payload bytes as per DSC spec 1.2 Table 4-1 */
+	drm_dsc_pps_payload_pack(&dp_dsc_pps_sdp.pps_payload, vdsc_cfg);
+
+	dig_port->write_infoframe(encoder, crtc_state,
+				  DP_SDP_PPS, &dp_dsc_pps_sdp,
+				  sizeof(dp_dsc_pps_sdp));
+}
+
+static i915_reg_t dss_ctl1_reg(struct intel_crtc *crtc, enum transcoder cpu_transcoder)
+{
+	return is_pipe_dsc(crtc, cpu_transcoder) ?
+		ICL_PIPE_DSS_CTL1(crtc->pipe) : DSS_CTL1;
+}
+
+static i915_reg_t dss_ctl2_reg(struct intel_crtc *crtc, enum transcoder cpu_transcoder)
+{
+	return is_pipe_dsc(crtc, cpu_transcoder) ?
+		ICL_PIPE_DSS_CTL2(crtc->pipe) : DSS_CTL2;
+}
+
+void intel_uncompressed_joiner_enable(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 dss_ctl1_val = 0;
+
+	if (crtc_state->bigjoiner_pipes && !crtc_state->dsc.compression_enable) {
+		if (intel_crtc_is_bigjoiner_slave(crtc_state))
+			dss_ctl1_val |= UNCOMPRESSED_JOINER_SLAVE;
+		else
+			dss_ctl1_val |= UNCOMPRESSED_JOINER_MASTER;
+
+		intel_de_write(dev_priv, dss_ctl1_reg(crtc, crtc_state->cpu_transcoder), dss_ctl1_val);
+	}
+}
+
+void intel_dsc_enable(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	u32 dss_ctl1_val = 0;
+	u32 dss_ctl2_val = 0;
+
+	if (!crtc_state->dsc.compression_enable)
+		return;
+
+	intel_dsc_pps_configure(crtc_state);
+
+	dss_ctl2_val |= LEFT_BRANCH_VDSC_ENABLE;
+	if (crtc_state->dsc.dsc_split) {
+		dss_ctl2_val |= RIGHT_BRANCH_VDSC_ENABLE;
+		dss_ctl1_val |= JOINER_ENABLE;
+	}
+	if (crtc_state->bigjoiner_pipes) {
+		dss_ctl1_val |= BIG_JOINER_ENABLE;
+		if (!intel_crtc_is_bigjoiner_slave(crtc_state))
+			dss_ctl1_val |= MASTER_BIG_JOINER_ENABLE;
+	}
+	intel_de_write(dev_priv, dss_ctl1_reg(crtc, crtc_state->cpu_transcoder), dss_ctl1_val);
+	intel_de_write(dev_priv, dss_ctl2_reg(crtc, crtc_state->cpu_transcoder), dss_ctl2_val);
+}
+
+void intel_dsc_disable(const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	/* Disable only if either of them is enabled */
+	if (old_crtc_state->dsc.compression_enable ||
+	    old_crtc_state->bigjoiner_pipes) {
+		intel_de_write(dev_priv, dss_ctl1_reg(crtc, old_crtc_state->cpu_transcoder), 0);
+		intel_de_write(dev_priv, dss_ctl2_reg(crtc, old_crtc_state->cpu_transcoder), 0);
+	}
+}
+
+void intel_dsc_get_config(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct drm_dsc_config *vdsc_cfg = &crtc_state->dsc.config;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	enum pipe pipe = crtc->pipe;
+	enum intel_display_power_domain power_domain;
+	intel_wakeref_t wakeref;
+	u32 dss_ctl1, dss_ctl2, val;
+
+	if (!intel_dsc_source_support(crtc_state))
+		return;
+
+	power_domain = intel_dsc_power_domain(crtc, cpu_transcoder);
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref)
+		return;
+
+	dss_ctl1 = intel_de_read(dev_priv, dss_ctl1_reg(crtc, cpu_transcoder));
+	dss_ctl2 = intel_de_read(dev_priv, dss_ctl2_reg(crtc, cpu_transcoder));
+
+	crtc_state->dsc.compression_enable = dss_ctl2 & LEFT_BRANCH_VDSC_ENABLE;
+	if (!crtc_state->dsc.compression_enable)
+		goto out;
+
+	crtc_state->dsc.dsc_split = (dss_ctl2 & RIGHT_BRANCH_VDSC_ENABLE) &&
+		(dss_ctl1 & JOINER_ENABLE);
+
+	/* FIXME: add more state readout as needed */
+
+	/* PPS1 */
+	if (!is_pipe_dsc(crtc, cpu_transcoder))
+		val = intel_de_read(dev_priv, DSCA_PICTURE_PARAMETER_SET_1);
+	else
+		val = intel_de_read(dev_priv,
+				    ICL_DSC0_PICTURE_PARAMETER_SET_1(pipe));
+	vdsc_cfg->bits_per_pixel = val;
+	crtc_state->dsc.compressed_bpp = vdsc_cfg->bits_per_pixel >> 4;
+out:
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_vdsc.h b/drivers/gpu/drm/i915/display/intel_vdsc.h
new file mode 100644
index 000000000..8763f00fa
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_vdsc.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_VDSC_H__
+#define __INTEL_VDSC_H__
+
+#include <linux/types.h>
+
+enum transcoder;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_encoder;
+
+bool intel_dsc_source_support(const struct intel_crtc_state *crtc_state);
+void intel_uncompressed_joiner_enable(const struct intel_crtc_state *crtc_state);
+void intel_dsc_enable(const struct intel_crtc_state *crtc_state);
+void intel_dsc_disable(const struct intel_crtc_state *crtc_state);
+int intel_dsc_compute_params(struct intel_crtc_state *pipe_config);
+void intel_dsc_get_config(struct intel_crtc_state *crtc_state);
+enum intel_display_power_domain
+intel_dsc_power_domain(struct intel_crtc *crtc, enum transcoder cpu_transcoder);
+struct intel_crtc *intel_dsc_get_bigjoiner_secondary(const struct intel_crtc *primary_crtc);
+void intel_dsc_dsi_pps_write(struct intel_encoder *encoder,
+			     const struct intel_crtc_state *crtc_state);
+void intel_dsc_dp_pps_write(struct intel_encoder *encoder,
+			    const struct intel_crtc_state *crtc_state);
+
+#endif /* __INTEL_VDSC_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_vga.c b/drivers/gpu/drm/i915/display/intel_vga.c
new file mode 100644
index 000000000..b5d058404
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_vga.c
@@ -0,0 +1,167 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/pci.h>
+#include <linux/vgaarb.h>
+
+#include <drm/i915_drm.h>
+#include <video/vga.h>
+
+#include "i915_drv.h"
+#include "intel_de.h"
+#include "intel_vga.h"
+
+static i915_reg_t intel_vga_cntrl_reg(struct drm_i915_private *i915)
+{
+	if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+		return VLV_VGACNTRL;
+	else if (DISPLAY_VER(i915) >= 5)
+		return CPU_VGACNTRL;
+	else
+		return VGACNTRL;
+}
+
+/* Disable the VGA plane that we never use */
+void intel_vga_disable(struct drm_i915_private *dev_priv)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	i915_reg_t vga_reg = intel_vga_cntrl_reg(dev_priv);
+	u8 sr1;
+
+	if (intel_de_read(dev_priv, vga_reg) & VGA_DISP_DISABLE)
+		return;
+
+	/* WaEnableVGAAccessThroughIOPort:ctg,elk,ilk,snb,ivb,vlv,hsw */
+	vga_get_uninterruptible(pdev, VGA_RSRC_LEGACY_IO);
+	outb(0x01, VGA_SEQ_I);
+	sr1 = inb(VGA_SEQ_D);
+	outb(sr1 | VGA_SR01_SCREEN_OFF, VGA_SEQ_D);
+	vga_put(pdev, VGA_RSRC_LEGACY_IO);
+	udelay(300);
+
+	intel_de_write(dev_priv, vga_reg, VGA_DISP_DISABLE);
+	intel_de_posting_read(dev_priv, vga_reg);
+}
+
+void intel_vga_redisable_power_on(struct drm_i915_private *dev_priv)
+{
+	i915_reg_t vga_reg = intel_vga_cntrl_reg(dev_priv);
+
+	if (!(intel_de_read(dev_priv, vga_reg) & VGA_DISP_DISABLE)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Something enabled VGA plane, disabling it\n");
+		intel_vga_disable(dev_priv);
+	}
+}
+
+void intel_vga_redisable(struct drm_i915_private *i915)
+{
+	intel_wakeref_t wakeref;
+
+	/*
+	 * This function can be called both from intel_modeset_setup_hw_state or
+	 * at a very early point in our resume sequence, where the power well
+	 * structures are not yet restored. Since this function is at a very
+	 * paranoid "someone might have enabled VGA while we were not looking"
+	 * level, just check if the power well is enabled instead of trying to
+	 * follow the "don't touch the power well if we don't need it" policy
+	 * the rest of the driver uses.
+	 */
+	wakeref = intel_display_power_get_if_enabled(i915, POWER_DOMAIN_VGA);
+	if (!wakeref)
+		return;
+
+	intel_vga_redisable_power_on(i915);
+
+	intel_display_power_put(i915, POWER_DOMAIN_VGA, wakeref);
+}
+
+void intel_vga_reset_io_mem(struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+
+	/*
+	 * After we re-enable the power well, if we touch VGA register 0x3d5
+	 * we'll get unclaimed register interrupts. This stops after we write
+	 * anything to the VGA MSR register. The vgacon module uses this
+	 * register all the time, so if we unbind our driver and, as a
+	 * consequence, bind vgacon, we'll get stuck in an infinite loop at
+	 * console_unlock(). So make here we touch the VGA MSR register, making
+	 * sure vgacon can keep working normally without triggering interrupts
+	 * and error messages.
+	 */
+	vga_get_uninterruptible(pdev, VGA_RSRC_LEGACY_IO);
+	outb(inb(VGA_MIS_R), VGA_MIS_W);
+	vga_put(pdev, VGA_RSRC_LEGACY_IO);
+}
+
+static int
+intel_vga_set_state(struct drm_i915_private *i915, bool enable_decode)
+{
+	unsigned int reg = DISPLAY_VER(i915) >= 6 ? SNB_GMCH_CTRL : INTEL_GMCH_CTRL;
+	u16 gmch_ctrl;
+
+	if (pci_read_config_word(i915->bridge_dev, reg, &gmch_ctrl)) {
+		drm_err(&i915->drm, "failed to read control word\n");
+		return -EIO;
+	}
+
+	if (!!(gmch_ctrl & INTEL_GMCH_VGA_DISABLE) == !enable_decode)
+		return 0;
+
+	if (enable_decode)
+		gmch_ctrl &= ~INTEL_GMCH_VGA_DISABLE;
+	else
+		gmch_ctrl |= INTEL_GMCH_VGA_DISABLE;
+
+	if (pci_write_config_word(i915->bridge_dev, reg, gmch_ctrl)) {
+		drm_err(&i915->drm, "failed to write control word\n");
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static unsigned int
+intel_vga_set_decode(struct pci_dev *pdev, bool enable_decode)
+{
+	struct drm_i915_private *i915 = pdev_to_i915(pdev);
+
+	intel_vga_set_state(i915, enable_decode);
+
+	if (enable_decode)
+		return VGA_RSRC_LEGACY_IO | VGA_RSRC_LEGACY_MEM |
+		       VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM;
+	else
+		return VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM;
+}
+
+int intel_vga_register(struct drm_i915_private *i915)
+{
+
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	int ret;
+
+	/*
+	 * If we have > 1 VGA cards, then we need to arbitrate access to the
+	 * common VGA resources.
+	 *
+	 * If we are a secondary display controller (!PCI_DISPLAY_CLASS_VGA),
+	 * then we do not take part in VGA arbitration and the
+	 * vga_client_register() fails with -ENODEV.
+	 */
+	ret = vga_client_register(pdev, intel_vga_set_decode);
+	if (ret && ret != -ENODEV)
+		return ret;
+
+	return 0;
+}
+
+void intel_vga_unregister(struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+
+	vga_client_unregister(pdev);
+}
diff --git a/drivers/gpu/drm/i915/display/intel_vga.h b/drivers/gpu/drm/i915/display/intel_vga.h
new file mode 100644
index 000000000..ba5b55b91
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_vga.h
@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_VGA_H__
+#define __INTEL_VGA_H__
+
+struct drm_i915_private;
+
+void intel_vga_reset_io_mem(struct drm_i915_private *i915);
+void intel_vga_disable(struct drm_i915_private *i915);
+void intel_vga_redisable(struct drm_i915_private *i915);
+void intel_vga_redisable_power_on(struct drm_i915_private *i915);
+int intel_vga_register(struct drm_i915_private *i915);
+void intel_vga_unregister(struct drm_i915_private *i915);
+
+#endif /* __INTEL_VGA_H__ */
diff --git a/drivers/gpu/drm/i915/display/intel_vrr.c b/drivers/gpu/drm/i915/display/intel_vrr.c
new file mode 100644
index 000000000..5eac99021
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_vrr.c
@@ -0,0 +1,262 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ *
+ */
+
+#include "i915_drv.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_vrr.h"
+
+bool intel_vrr_is_capable(struct intel_connector *connector)
+{
+	const struct drm_display_info *info = &connector->base.display_info;
+	struct drm_i915_private *i915 = to_i915(connector->base.dev);
+	struct intel_dp *intel_dp;
+
+	/*
+	 * DP Sink is capable of VRR video timings if
+	 * Ignore MSA bit is set in DPCD.
+	 * EDID monitor range also should be atleast 10 for reasonable
+	 * Adaptive Sync or Variable Refresh Rate end user experience.
+	 */
+	switch (connector->base.connector_type) {
+	case DRM_MODE_CONNECTOR_eDP:
+		if (!connector->panel.vbt.vrr)
+			return false;
+		fallthrough;
+	case DRM_MODE_CONNECTOR_DisplayPort:
+		intel_dp = intel_attached_dp(connector);
+
+		if (!drm_dp_sink_can_do_video_without_timing_msa(intel_dp->dpcd))
+			return false;
+
+		break;
+	default:
+		return false;
+	}
+
+	return HAS_VRR(i915) &&
+		info->monitor_range.max_vfreq - info->monitor_range.min_vfreq > 10;
+}
+
+void
+intel_vrr_check_modeset(struct intel_atomic_state *state)
+{
+	int i;
+	struct intel_crtc_state *old_crtc_state, *new_crtc_state;
+	struct intel_crtc *crtc;
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		if (new_crtc_state->uapi.vrr_enabled !=
+		    old_crtc_state->uapi.vrr_enabled)
+			new_crtc_state->uapi.mode_changed = true;
+	}
+}
+
+/*
+ * Without VRR registers get latched at:
+ *  vblank_start
+ *
+ * With VRR the earliest registers can get latched is:
+ *  intel_vrr_vmin_vblank_start(), which if we want to maintain
+ *  the correct min vtotal is >=vblank_start+1
+ *
+ * The latest point registers can get latched is the vmax decision boundary:
+ *  intel_vrr_vmax_vblank_start()
+ *
+ * Between those two points the vblank exit starts (and hence registers get
+ * latched) ASAP after a push is sent.
+ *
+ * framestart_delay is programmable 1-4.
+ */
+static int intel_vrr_vblank_exit_length(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	/* The hw imposes the extra scanline before frame start */
+	if (DISPLAY_VER(i915) >= 13)
+		return crtc_state->vrr.guardband + crtc_state->framestart_delay + 1;
+	else
+		return crtc_state->vrr.pipeline_full + crtc_state->framestart_delay + 1;
+}
+
+int intel_vrr_vmin_vblank_start(const struct intel_crtc_state *crtc_state)
+{
+	/* Min vblank actually determined by flipline that is always >=vmin+1 */
+	return crtc_state->vrr.vmin + 1 - intel_vrr_vblank_exit_length(crtc_state);
+}
+
+int intel_vrr_vmax_vblank_start(const struct intel_crtc_state *crtc_state)
+{
+	return crtc_state->vrr.vmax - intel_vrr_vblank_exit_length(crtc_state);
+}
+
+void
+intel_vrr_compute_config(struct intel_crtc_state *crtc_state,
+			 struct drm_connector_state *conn_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_connector *connector =
+		to_intel_connector(conn_state->connector);
+	struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
+	const struct drm_display_info *info = &connector->base.display_info;
+	int vmin, vmax;
+
+	if (!intel_vrr_is_capable(connector))
+		return;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
+		return;
+
+	if (!crtc_state->uapi.vrr_enabled)
+		return;
+
+	vmin = DIV_ROUND_UP(adjusted_mode->crtc_clock * 1000,
+			    adjusted_mode->crtc_htotal * info->monitor_range.max_vfreq);
+	vmax = adjusted_mode->crtc_clock * 1000 /
+		(adjusted_mode->crtc_htotal * info->monitor_range.min_vfreq);
+
+	vmin = max_t(int, vmin, adjusted_mode->crtc_vtotal);
+	vmax = max_t(int, vmax, adjusted_mode->crtc_vtotal);
+
+	if (vmin >= vmax)
+		return;
+
+	/*
+	 * flipline determines the min vblank length the hardware will
+	 * generate, and flipline>=vmin+1, hence we reduce vmin by one
+	 * to make sure we can get the actual min vblank length.
+	 */
+	crtc_state->vrr.vmin = vmin - 1;
+	crtc_state->vrr.vmax = vmax;
+	crtc_state->vrr.enable = true;
+
+	crtc_state->vrr.flipline = crtc_state->vrr.vmin + 1;
+
+	/*
+	 * For XE_LPD+, we use guardband and pipeline override
+	 * is deprecated.
+	 */
+	if (DISPLAY_VER(i915) >= 13) {
+		/*
+		 * FIXME: Subtract Window2 delay from below value.
+		 *
+		 * Window2 specifies time required to program DSB (Window2) in
+		 * number of scan lines. Assuming 0 for no DSB.
+		 */
+		crtc_state->vrr.guardband =
+			crtc_state->vrr.vmin - adjusted_mode->crtc_vdisplay;
+	} else {
+		/*
+		 * FIXME: s/4/framestart_delay/ to get consistent
+		 * earliest/latest points for register latching regardless
+		 * of the framestart_delay used?
+		 *
+		 * FIXME: this really needs the extra scanline to provide consistent
+		 * behaviour for all framestart_delay values. Otherwise with
+		 * framestart_delay==4 we will end up extending the min vblank by
+		 * one extra line.
+		 */
+		crtc_state->vrr.pipeline_full =
+			min(255, crtc_state->vrr.vmin - adjusted_mode->crtc_vdisplay - 4 - 1);
+	}
+
+	crtc_state->mode_flags |= I915_MODE_FLAG_VRR;
+}
+
+void intel_vrr_enable(struct intel_encoder *encoder,
+		      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 trans_vrr_ctl;
+
+	if (!crtc_state->vrr.enable)
+		return;
+
+	if (DISPLAY_VER(dev_priv) >= 13)
+		trans_vrr_ctl = VRR_CTL_VRR_ENABLE |
+			VRR_CTL_IGN_MAX_SHIFT | VRR_CTL_FLIP_LINE_EN |
+			XELPD_VRR_CTL_VRR_GUARDBAND(crtc_state->vrr.guardband);
+	else
+		trans_vrr_ctl = VRR_CTL_VRR_ENABLE |
+			VRR_CTL_IGN_MAX_SHIFT | VRR_CTL_FLIP_LINE_EN |
+			VRR_CTL_PIPELINE_FULL(crtc_state->vrr.pipeline_full) |
+			VRR_CTL_PIPELINE_FULL_OVERRIDE;
+
+	intel_de_write(dev_priv, TRANS_VRR_VMIN(cpu_transcoder), crtc_state->vrr.vmin - 1);
+	intel_de_write(dev_priv, TRANS_VRR_VMAX(cpu_transcoder), crtc_state->vrr.vmax - 1);
+	intel_de_write(dev_priv, TRANS_VRR_CTL(cpu_transcoder), trans_vrr_ctl);
+	intel_de_write(dev_priv, TRANS_VRR_FLIPLINE(cpu_transcoder), crtc_state->vrr.flipline - 1);
+	intel_de_write(dev_priv, TRANS_PUSH(cpu_transcoder), TRANS_PUSH_EN);
+}
+
+void intel_vrr_send_push(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	if (!crtc_state->vrr.enable)
+		return;
+
+	intel_de_write(dev_priv, TRANS_PUSH(cpu_transcoder),
+		       TRANS_PUSH_EN | TRANS_PUSH_SEND);
+}
+
+bool intel_vrr_is_push_sent(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+
+	if (!crtc_state->vrr.enable)
+		return false;
+
+	return intel_de_read(dev_priv, TRANS_PUSH(cpu_transcoder)) & TRANS_PUSH_SEND;
+}
+
+void intel_vrr_disable(const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = old_crtc_state->cpu_transcoder;
+
+	if (!old_crtc_state->vrr.enable)
+		return;
+
+	intel_de_write(dev_priv, TRANS_VRR_CTL(cpu_transcoder), 0);
+	intel_de_write(dev_priv, TRANS_PUSH(cpu_transcoder), 0);
+}
+
+void intel_vrr_get_config(struct intel_crtc *crtc,
+			  struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
+	u32 trans_vrr_ctl;
+
+	trans_vrr_ctl = intel_de_read(dev_priv, TRANS_VRR_CTL(cpu_transcoder));
+	crtc_state->vrr.enable = trans_vrr_ctl & VRR_CTL_VRR_ENABLE;
+	if (!crtc_state->vrr.enable)
+		return;
+
+	if (DISPLAY_VER(dev_priv) >= 13)
+		crtc_state->vrr.guardband =
+			REG_FIELD_GET(XELPD_VRR_CTL_VRR_GUARDBAND_MASK, trans_vrr_ctl);
+	else
+		if (trans_vrr_ctl & VRR_CTL_PIPELINE_FULL_OVERRIDE)
+			crtc_state->vrr.pipeline_full =
+				REG_FIELD_GET(VRR_CTL_PIPELINE_FULL_MASK, trans_vrr_ctl);
+	if (trans_vrr_ctl & VRR_CTL_FLIP_LINE_EN)
+		crtc_state->vrr.flipline = intel_de_read(dev_priv, TRANS_VRR_FLIPLINE(cpu_transcoder)) + 1;
+	crtc_state->vrr.vmax = intel_de_read(dev_priv, TRANS_VRR_VMAX(cpu_transcoder)) + 1;
+	crtc_state->vrr.vmin = intel_de_read(dev_priv, TRANS_VRR_VMIN(cpu_transcoder)) + 1;
+
+	crtc_state->mode_flags |= I915_MODE_FLAG_VRR;
+}
diff --git a/drivers/gpu/drm/i915/display/intel_vrr.h b/drivers/gpu/drm/i915/display/intel_vrr.h
new file mode 100644
index 000000000..9fda1135b
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/intel_vrr.h
@@ -0,0 +1,33 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_VRR_H__
+#define __INTEL_VRR_H__
+
+#include <linux/types.h>
+
+struct drm_connector_state;
+struct intel_atomic_state;
+struct intel_connector;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_dp;
+struct intel_encoder;
+
+bool intel_vrr_is_capable(struct intel_connector *connector);
+void intel_vrr_check_modeset(struct intel_atomic_state *state);
+void intel_vrr_compute_config(struct intel_crtc_state *crtc_state,
+			      struct drm_connector_state *conn_state);
+void intel_vrr_enable(struct intel_encoder *encoder,
+		      const struct intel_crtc_state *crtc_state);
+void intel_vrr_send_push(const struct intel_crtc_state *crtc_state);
+bool intel_vrr_is_push_sent(const struct intel_crtc_state *crtc_state);
+void intel_vrr_disable(const struct intel_crtc_state *old_crtc_state);
+void intel_vrr_get_config(struct intel_crtc *crtc,
+			  struct intel_crtc_state *crtc_state);
+int intel_vrr_vmax_vblank_start(const struct intel_crtc_state *crtc_state);
+int intel_vrr_vmin_vblank_start(const struct intel_crtc_state *crtc_state);
+
+#endif /* __INTEL_VRR_H__ */
diff --git a/drivers/gpu/drm/i915/display/skl_scaler.c b/drivers/gpu/drm/i915/display/skl_scaler.c
new file mode 100644
index 000000000..0b74f91e8
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/skl_scaler.c
@@ -0,0 +1,847 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_fb.h"
+#include "skl_scaler.h"
+#include "skl_universal_plane.h"
+
+/*
+ * The hardware phase 0.0 refers to the center of the pixel.
+ * We want to start from the top/left edge which is phase
+ * -0.5. That matches how the hardware calculates the scaling
+ * factors (from top-left of the first pixel to bottom-right
+ * of the last pixel, as opposed to the pixel centers).
+ *
+ * For 4:2:0 subsampled chroma planes we obviously have to
+ * adjust that so that the chroma sample position lands in
+ * the right spot.
+ *
+ * Note that for packed YCbCr 4:2:2 formats there is no way to
+ * control chroma siting. The hardware simply replicates the
+ * chroma samples for both of the luma samples, and thus we don't
+ * actually get the expected MPEG2 chroma siting convention :(
+ * The same behaviour is observed on pre-SKL platforms as well.
+ *
+ * Theory behind the formula (note that we ignore sub-pixel
+ * source coordinates):
+ * s = source sample position
+ * d = destination sample position
+ *
+ * Downscaling 4:1:
+ * -0.5
+ * | 0.0
+ * | |     1.5 (initial phase)
+ * | |     |
+ * v v     v
+ * | s | s | s | s |
+ * |       d       |
+ *
+ * Upscaling 1:4:
+ * -0.5
+ * | -0.375 (initial phase)
+ * | |     0.0
+ * | |     |
+ * v v     v
+ * |       s       |
+ * | d | d | d | d |
+ */
+static u16 skl_scaler_calc_phase(int sub, int scale, bool chroma_cosited)
+{
+	int phase = -0x8000;
+	u16 trip = 0;
+
+	if (chroma_cosited)
+		phase += (sub - 1) * 0x8000 / sub;
+
+	phase += scale / (2 * sub);
+
+	/*
+	 * Hardware initial phase limited to [-0.5:1.5].
+	 * Since the max hardware scale factor is 3.0, we
+	 * should never actually excdeed 1.0 here.
+	 */
+	WARN_ON(phase < -0x8000 || phase > 0x18000);
+
+	if (phase < 0)
+		phase = 0x10000 + phase;
+	else
+		trip = PS_PHASE_TRIP;
+
+	return ((phase >> 2) & PS_PHASE_MASK) | trip;
+}
+
+#define SKL_MIN_SRC_W 8
+#define SKL_MAX_SRC_W 4096
+#define SKL_MIN_SRC_H 8
+#define SKL_MAX_SRC_H 4096
+#define SKL_MIN_DST_W 8
+#define SKL_MAX_DST_W 4096
+#define SKL_MIN_DST_H 8
+#define SKL_MAX_DST_H 4096
+#define ICL_MAX_SRC_W 5120
+#define ICL_MAX_SRC_H 4096
+#define ICL_MAX_DST_W 5120
+#define ICL_MAX_DST_H 4096
+#define MTL_MAX_SRC_W 4096
+#define MTL_MAX_SRC_H 8192
+#define MTL_MAX_DST_W 8192
+#define MTL_MAX_DST_H 8192
+#define SKL_MIN_YUV_420_SRC_W 16
+#define SKL_MIN_YUV_420_SRC_H 16
+
+static int
+skl_update_scaler(struct intel_crtc_state *crtc_state, bool force_detach,
+		  unsigned int scaler_user, int *scaler_id,
+		  int src_w, int src_h, int dst_w, int dst_h,
+		  const struct drm_format_info *format,
+		  u64 modifier, bool need_scaler)
+{
+	struct intel_crtc_scaler_state *scaler_state =
+		&crtc_state->scaler_state;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+	int pipe_src_w = drm_rect_width(&crtc_state->pipe_src);
+	int pipe_src_h = drm_rect_height(&crtc_state->pipe_src);
+	int min_src_w, min_src_h, min_dst_w, min_dst_h;
+	int max_src_w, max_src_h, max_dst_w, max_dst_h;
+
+	/*
+	 * Src coordinates are already rotated by 270 degrees for
+	 * the 90/270 degree plane rotation cases (to match the
+	 * GTT mapping), hence no need to account for rotation here.
+	 */
+	if (src_w != dst_w || src_h != dst_h)
+		need_scaler = true;
+
+	/*
+	 * Scaling/fitting not supported in IF-ID mode in GEN9+
+	 * TODO: Interlace fetch mode doesn't support YUV420 planar formats.
+	 * Once NV12 is enabled, handle it here while allocating scaler
+	 * for NV12.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 9 && crtc_state->hw.enable &&
+	    need_scaler && adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Pipe/Plane scaling not supported with IF-ID mode\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * if plane is being disabled or scaler is no more required or force detach
+	 *  - free scaler binded to this plane/crtc
+	 *  - in order to do this, update crtc->scaler_usage
+	 *
+	 * Here scaler state in crtc_state is set free so that
+	 * scaler can be assigned to other user. Actual register
+	 * update to free the scaler is done in plane/panel-fit programming.
+	 * For this purpose crtc/plane_state->scaler_id isn't reset here.
+	 */
+	if (force_detach || !need_scaler) {
+		if (*scaler_id >= 0) {
+			scaler_state->scaler_users &= ~(1 << scaler_user);
+			scaler_state->scalers[*scaler_id].in_use = 0;
+
+			drm_dbg_kms(&dev_priv->drm,
+				    "scaler_user index %u.%u: "
+				    "Staged freeing scaler id %d scaler_users = 0x%x\n",
+				    crtc->pipe, scaler_user, *scaler_id,
+				    scaler_state->scaler_users);
+			*scaler_id = -1;
+		}
+		return 0;
+	}
+
+	if (format && intel_format_info_is_yuv_semiplanar(format, modifier) &&
+	    (src_h < SKL_MIN_YUV_420_SRC_H || src_w < SKL_MIN_YUV_420_SRC_W)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Planar YUV: src dimensions not met\n");
+		return -EINVAL;
+	}
+
+	min_src_w = SKL_MIN_SRC_W;
+	min_src_h = SKL_MIN_SRC_H;
+	min_dst_w = SKL_MIN_DST_W;
+	min_dst_h = SKL_MIN_DST_H;
+
+	if (DISPLAY_VER(dev_priv) < 11) {
+		max_src_w = SKL_MAX_SRC_W;
+		max_src_h = SKL_MAX_SRC_H;
+		max_dst_w = SKL_MAX_DST_W;
+		max_dst_h = SKL_MAX_DST_H;
+	} else if (DISPLAY_VER(dev_priv) < 14) {
+		max_src_w = ICL_MAX_SRC_W;
+		max_src_h = ICL_MAX_SRC_H;
+		max_dst_w = ICL_MAX_DST_W;
+		max_dst_h = ICL_MAX_DST_H;
+	} else {
+		max_src_w = MTL_MAX_SRC_W;
+		max_src_h = MTL_MAX_SRC_H;
+		max_dst_w = MTL_MAX_DST_W;
+		max_dst_h = MTL_MAX_DST_H;
+	}
+
+	/* range checks */
+	if (src_w < min_src_w || src_h < min_src_h ||
+	    dst_w < min_dst_w || dst_h < min_dst_h ||
+	    src_w > max_src_w || src_h > max_src_h ||
+	    dst_w > max_dst_w || dst_h > max_dst_h) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "scaler_user index %u.%u: src %ux%u dst %ux%u "
+			    "size is out of scaler range\n",
+			    crtc->pipe, scaler_user, src_w, src_h,
+			    dst_w, dst_h);
+		return -EINVAL;
+	}
+
+	/*
+	 * The pipe scaler does not use all the bits of PIPESRC, at least
+	 * on the earlier platforms. So even when we're scaling a plane
+	 * the *pipe* source size must not be too large. For simplicity
+	 * we assume the limits match the scaler source size limits. Might
+	 * not be 100% accurate on all platforms, but good enough for now.
+	 */
+	if (pipe_src_w > max_src_w || pipe_src_h > max_src_h) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "scaler_user index %u.%u: pipe src size %ux%u "
+			    "is out of scaler range\n",
+			    crtc->pipe, scaler_user, pipe_src_w, pipe_src_h);
+		return -EINVAL;
+	}
+
+	/* mark this plane as a scaler user in crtc_state */
+	scaler_state->scaler_users |= (1 << scaler_user);
+	drm_dbg_kms(&dev_priv->drm, "scaler_user index %u.%u: "
+		    "staged scaling request for %ux%u->%ux%u scaler_users = 0x%x\n",
+		    crtc->pipe, scaler_user, src_w, src_h, dst_w, dst_h,
+		    scaler_state->scaler_users);
+
+	return 0;
+}
+
+int skl_update_scaler_crtc(struct intel_crtc_state *crtc_state)
+{
+	const struct drm_display_mode *pipe_mode = &crtc_state->hw.pipe_mode;
+	int width, height;
+
+	if (crtc_state->pch_pfit.enabled) {
+		width = drm_rect_width(&crtc_state->pch_pfit.dst);
+		height = drm_rect_height(&crtc_state->pch_pfit.dst);
+	} else {
+		width = pipe_mode->crtc_hdisplay;
+		height = pipe_mode->crtc_vdisplay;
+	}
+	return skl_update_scaler(crtc_state, !crtc_state->hw.active,
+				 SKL_CRTC_INDEX,
+				 &crtc_state->scaler_state.scaler_id,
+				 drm_rect_width(&crtc_state->pipe_src),
+				 drm_rect_height(&crtc_state->pipe_src),
+				 width, height, NULL, 0,
+				 crtc_state->pch_pfit.enabled);
+}
+
+/**
+ * skl_update_scaler_plane - Stages update to scaler state for a given plane.
+ * @crtc_state: crtc's scaler state
+ * @plane_state: atomic plane state to update
+ *
+ * Return
+ *     0 - scaler_usage updated successfully
+ *    error - requested scaling cannot be supported or other error condition
+ */
+int skl_update_scaler_plane(struct intel_crtc_state *crtc_state,
+			    struct intel_plane_state *plane_state)
+{
+	struct intel_plane *intel_plane =
+		to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(intel_plane->base.dev);
+	struct drm_framebuffer *fb = plane_state->hw.fb;
+	int ret;
+	bool force_detach = !fb || !plane_state->uapi.visible;
+	bool need_scaler = false;
+
+	/* Pre-gen11 and SDR planes always need a scaler for planar formats. */
+	if (!icl_is_hdr_plane(dev_priv, intel_plane->id) &&
+	    fb && intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier))
+		need_scaler = true;
+
+	ret = skl_update_scaler(crtc_state, force_detach,
+				drm_plane_index(&intel_plane->base),
+				&plane_state->scaler_id,
+				drm_rect_width(&plane_state->uapi.src) >> 16,
+				drm_rect_height(&plane_state->uapi.src) >> 16,
+				drm_rect_width(&plane_state->uapi.dst),
+				drm_rect_height(&plane_state->uapi.dst),
+				fb ? fb->format : NULL,
+				fb ? fb->modifier : 0,
+				need_scaler);
+
+	if (ret || plane_state->scaler_id < 0)
+		return ret;
+
+	/* check colorkey */
+	if (plane_state->ckey.flags) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "[PLANE:%d:%s] scaling with color key not allowed",
+			    intel_plane->base.base.id,
+			    intel_plane->base.name);
+		return -EINVAL;
+	}
+
+	/* Check src format */
+	switch (fb->format->format) {
+	case DRM_FORMAT_RGB565:
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_ABGR8888:
+	case DRM_FORMAT_ARGB8888:
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_XBGR2101010:
+	case DRM_FORMAT_ARGB2101010:
+	case DRM_FORMAT_ABGR2101010:
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_YVYU:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_VYUY:
+	case DRM_FORMAT_NV12:
+	case DRM_FORMAT_XYUV8888:
+	case DRM_FORMAT_P010:
+	case DRM_FORMAT_P012:
+	case DRM_FORMAT_P016:
+	case DRM_FORMAT_Y210:
+	case DRM_FORMAT_Y212:
+	case DRM_FORMAT_Y216:
+	case DRM_FORMAT_XVYU2101010:
+	case DRM_FORMAT_XVYU12_16161616:
+	case DRM_FORMAT_XVYU16161616:
+		break;
+	case DRM_FORMAT_XBGR16161616F:
+	case DRM_FORMAT_ABGR16161616F:
+	case DRM_FORMAT_XRGB16161616F:
+	case DRM_FORMAT_ARGB16161616F:
+		if (DISPLAY_VER(dev_priv) >= 11)
+			break;
+		fallthrough;
+	default:
+		drm_dbg_kms(&dev_priv->drm,
+			    "[PLANE:%d:%s] FB:%d unsupported scaling format 0x%x\n",
+			    intel_plane->base.base.id, intel_plane->base.name,
+			    fb->base.id, fb->format->format);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int intel_atomic_setup_scaler(struct intel_crtc_scaler_state *scaler_state,
+				     int num_scalers_need, struct intel_crtc *intel_crtc,
+				     const char *name, int idx,
+				     struct intel_plane_state *plane_state,
+				     int *scaler_id)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_crtc->base.dev);
+	int j;
+	u32 mode;
+
+	if (*scaler_id < 0) {
+		/* find a free scaler */
+		for (j = 0; j < intel_crtc->num_scalers; j++) {
+			if (scaler_state->scalers[j].in_use)
+				continue;
+
+			*scaler_id = j;
+			scaler_state->scalers[*scaler_id].in_use = 1;
+			break;
+		}
+	}
+
+	if (drm_WARN(&dev_priv->drm, *scaler_id < 0,
+		     "Cannot find scaler for %s:%d\n", name, idx))
+		return -EINVAL;
+
+	/* set scaler mode */
+	if (plane_state && plane_state->hw.fb &&
+	    plane_state->hw.fb->format->is_yuv &&
+	    plane_state->hw.fb->format->num_planes > 1) {
+		struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+
+		if (DISPLAY_VER(dev_priv) == 9) {
+			mode = SKL_PS_SCALER_MODE_NV12;
+		} else if (icl_is_hdr_plane(dev_priv, plane->id)) {
+			/*
+			 * On gen11+'s HDR planes we only use the scaler for
+			 * scaling. They have a dedicated chroma upsampler, so
+			 * we don't need the scaler to upsample the UV plane.
+			 */
+			mode = PS_SCALER_MODE_NORMAL;
+		} else {
+			struct intel_plane *linked =
+				plane_state->planar_linked_plane;
+
+			mode = PS_SCALER_MODE_PLANAR;
+
+			if (linked)
+				mode |= PS_PLANE_Y_SEL(linked->id);
+		}
+	} else if (DISPLAY_VER(dev_priv) >= 10) {
+		mode = PS_SCALER_MODE_NORMAL;
+	} else if (num_scalers_need == 1 && intel_crtc->num_scalers > 1) {
+		/*
+		 * when only 1 scaler is in use on a pipe with 2 scalers
+		 * scaler 0 operates in high quality (HQ) mode.
+		 * In this case use scaler 0 to take advantage of HQ mode
+		 */
+		scaler_state->scalers[*scaler_id].in_use = 0;
+		*scaler_id = 0;
+		scaler_state->scalers[0].in_use = 1;
+		mode = SKL_PS_SCALER_MODE_HQ;
+	} else {
+		mode = SKL_PS_SCALER_MODE_DYN;
+	}
+
+	/*
+	 * FIXME: we should also check the scaler factors for pfit, so
+	 * this shouldn't be tied directly to planes.
+	 */
+	if (plane_state && plane_state->hw.fb) {
+		const struct drm_framebuffer *fb = plane_state->hw.fb;
+		const struct drm_rect *src = &plane_state->uapi.src;
+		const struct drm_rect *dst = &plane_state->uapi.dst;
+		int hscale, vscale, max_vscale, max_hscale;
+
+		/*
+		 * FIXME: When two scalers are needed, but only one of
+		 * them needs to downscale, we should make sure that
+		 * the one that needs downscaling support is assigned
+		 * as the first scaler, so we don't reject downscaling
+		 * unnecessarily.
+		 */
+
+		if (DISPLAY_VER(dev_priv) >= 14) {
+			/*
+			 * On versions 14 and up, only the first
+			 * scaler supports a vertical scaling factor
+			 * of more than 1.0, while a horizontal
+			 * scaling factor of 3.0 is supported.
+			 */
+			max_hscale = 0x30000 - 1;
+			if (*scaler_id == 0)
+				max_vscale = 0x30000 - 1;
+			else
+				max_vscale = 0x10000;
+
+		} else if (DISPLAY_VER(dev_priv) >= 10 ||
+			   !intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier)) {
+			max_hscale = 0x30000 - 1;
+			max_vscale = 0x30000 - 1;
+		} else {
+			max_hscale = 0x20000 - 1;
+			max_vscale = 0x20000 - 1;
+		}
+
+		/*
+		 * FIXME: We should change the if-else block above to
+		 * support HQ vs dynamic scaler properly.
+		 */
+
+		/* Check if required scaling is within limits */
+		hscale = drm_rect_calc_hscale(src, dst, 1, max_hscale);
+		vscale = drm_rect_calc_vscale(src, dst, 1, max_vscale);
+
+		if (hscale < 0 || vscale < 0) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Scaler %d doesn't support required plane scaling\n",
+				    *scaler_id);
+			drm_rect_debug_print("src: ", src, true);
+			drm_rect_debug_print("dst: ", dst, false);
+
+			return -EINVAL;
+		}
+	}
+
+	drm_dbg_kms(&dev_priv->drm, "Attached scaler id %u.%u to %s:%d\n",
+		    intel_crtc->pipe, *scaler_id, name, idx);
+	scaler_state->scalers[*scaler_id].mode = mode;
+
+	return 0;
+}
+
+/**
+ * intel_atomic_setup_scalers() - setup scalers for crtc per staged requests
+ * @dev_priv: i915 device
+ * @intel_crtc: intel crtc
+ * @crtc_state: incoming crtc_state to validate and setup scalers
+ *
+ * This function sets up scalers based on staged scaling requests for
+ * a @crtc and its planes. It is called from crtc level check path. If request
+ * is a supportable request, it attaches scalers to requested planes and crtc.
+ *
+ * This function takes into account the current scaler(s) in use by any planes
+ * not being part of this atomic state
+ *
+ *  Returns:
+ *         0 - scalers were setup successfully
+ *         error code - otherwise
+ */
+int intel_atomic_setup_scalers(struct drm_i915_private *dev_priv,
+			       struct intel_crtc *intel_crtc,
+			       struct intel_crtc_state *crtc_state)
+{
+	struct drm_plane *plane = NULL;
+	struct intel_plane *intel_plane;
+	struct intel_crtc_scaler_state *scaler_state =
+		&crtc_state->scaler_state;
+	struct drm_atomic_state *drm_state = crtc_state->uapi.state;
+	struct intel_atomic_state *intel_state = to_intel_atomic_state(drm_state);
+	int num_scalers_need;
+	int i;
+
+	num_scalers_need = hweight32(scaler_state->scaler_users);
+
+	/*
+	 * High level flow:
+	 * - staged scaler requests are already in scaler_state->scaler_users
+	 * - check whether staged scaling requests can be supported
+	 * - add planes using scalers that aren't in current transaction
+	 * - assign scalers to requested users
+	 * - as part of plane commit, scalers will be committed
+	 *   (i.e., either attached or detached) to respective planes in hw
+	 * - as part of crtc_commit, scaler will be either attached or detached
+	 *   to crtc in hw
+	 */
+
+	/* fail if required scalers > available scalers */
+	if (num_scalers_need > intel_crtc->num_scalers) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Too many scaling requests %d > %d\n",
+			    num_scalers_need, intel_crtc->num_scalers);
+		return -EINVAL;
+	}
+
+	/* walkthrough scaler_users bits and start assigning scalers */
+	for (i = 0; i < sizeof(scaler_state->scaler_users) * 8; i++) {
+		struct intel_plane_state *plane_state = NULL;
+		int *scaler_id;
+		const char *name;
+		int idx, ret;
+
+		/* skip if scaler not required */
+		if (!(scaler_state->scaler_users & (1 << i)))
+			continue;
+
+		if (i == SKL_CRTC_INDEX) {
+			name = "CRTC";
+			idx = intel_crtc->base.base.id;
+
+			/* panel fitter case: assign as a crtc scaler */
+			scaler_id = &scaler_state->scaler_id;
+		} else {
+			name = "PLANE";
+
+			/* plane scaler case: assign as a plane scaler */
+			/* find the plane that set the bit as scaler_user */
+			plane = drm_state->planes[i].ptr;
+
+			/*
+			 * to enable/disable hq mode, add planes that are using scaler
+			 * into this transaction
+			 */
+			if (!plane) {
+				struct drm_plane_state *state;
+
+				/*
+				 * GLK+ scalers don't have a HQ mode so it
+				 * isn't necessary to change between HQ and dyn mode
+				 * on those platforms.
+				 */
+				if (DISPLAY_VER(dev_priv) >= 10)
+					continue;
+
+				plane = drm_plane_from_index(&dev_priv->drm, i);
+				state = drm_atomic_get_plane_state(drm_state, plane);
+				if (IS_ERR(state)) {
+					drm_dbg_kms(&dev_priv->drm,
+						    "Failed to add [PLANE:%d] to drm_state\n",
+						    plane->base.id);
+					return PTR_ERR(state);
+				}
+			}
+
+			intel_plane = to_intel_plane(plane);
+			idx = plane->base.id;
+
+			/* plane on different crtc cannot be a scaler user of this crtc */
+			if (drm_WARN_ON(&dev_priv->drm,
+					intel_plane->pipe != intel_crtc->pipe))
+				continue;
+
+			plane_state = intel_atomic_get_new_plane_state(intel_state,
+								       intel_plane);
+			scaler_id = &plane_state->scaler_id;
+		}
+
+		ret = intel_atomic_setup_scaler(scaler_state, num_scalers_need,
+						intel_crtc, name, idx,
+						plane_state, scaler_id);
+		if (ret < 0)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int glk_coef_tap(int i)
+{
+	return i % 7;
+}
+
+static u16 glk_nearest_filter_coef(int t)
+{
+	return t == 3 ? 0x0800 : 0x3000;
+}
+
+/*
+ *  Theory behind setting nearest-neighbor integer scaling:
+ *
+ *  17 phase of 7 taps requires 119 coefficients in 60 dwords per set.
+ *  The letter represents the filter tap (D is the center tap) and the number
+ *  represents the coefficient set for a phase (0-16).
+ *
+ *         +------------+------------------------+------------------------+
+ *         |Index value | Data value coeffient 1 | Data value coeffient 2 |
+ *         +------------+------------------------+------------------------+
+ *         |   00h      |          B0            |          A0            |
+ *         +------------+------------------------+------------------------+
+ *         |   01h      |          D0            |          C0            |
+ *         +------------+------------------------+------------------------+
+ *         |   02h      |          F0            |          E0            |
+ *         +------------+------------------------+------------------------+
+ *         |   03h      |          A1            |          G0            |
+ *         +------------+------------------------+------------------------+
+ *         |   04h      |          C1            |          B1            |
+ *         +------------+------------------------+------------------------+
+ *         |   ...      |          ...           |          ...           |
+ *         +------------+------------------------+------------------------+
+ *         |   38h      |          B16           |          A16           |
+ *         +------------+------------------------+------------------------+
+ *         |   39h      |          D16           |          C16           |
+ *         +------------+------------------------+------------------------+
+ *         |   3Ah      |          F16           |          C16           |
+ *         +------------+------------------------+------------------------+
+ *         |   3Bh      |        Reserved        |          G16           |
+ *         +------------+------------------------+------------------------+
+ *
+ *  To enable nearest-neighbor scaling:  program scaler coefficents with
+ *  the center tap (Dxx) values set to 1 and all other values set to 0 as per
+ *  SCALER_COEFFICIENT_FORMAT
+ *
+ */
+
+static void glk_program_nearest_filter_coefs(struct drm_i915_private *dev_priv,
+					     enum pipe pipe, int id, int set)
+{
+	int i;
+
+	intel_de_write_fw(dev_priv, GLK_PS_COEF_INDEX_SET(pipe, id, set),
+			  PS_COEE_INDEX_AUTO_INC);
+
+	for (i = 0; i < 17 * 7; i += 2) {
+		u32 tmp;
+		int t;
+
+		t = glk_coef_tap(i);
+		tmp = glk_nearest_filter_coef(t);
+
+		t = glk_coef_tap(i + 1);
+		tmp |= glk_nearest_filter_coef(t) << 16;
+
+		intel_de_write_fw(dev_priv, GLK_PS_COEF_DATA_SET(pipe, id, set),
+				  tmp);
+	}
+
+	intel_de_write_fw(dev_priv, GLK_PS_COEF_INDEX_SET(pipe, id, set), 0);
+}
+
+static u32 skl_scaler_get_filter_select(enum drm_scaling_filter filter, int set)
+{
+	if (filter == DRM_SCALING_FILTER_NEAREST_NEIGHBOR) {
+		return (PS_FILTER_PROGRAMMED |
+			PS_Y_VERT_FILTER_SELECT(set) |
+			PS_Y_HORZ_FILTER_SELECT(set) |
+			PS_UV_VERT_FILTER_SELECT(set) |
+			PS_UV_HORZ_FILTER_SELECT(set));
+	}
+
+	return PS_FILTER_MEDIUM;
+}
+
+static void skl_scaler_setup_filter(struct drm_i915_private *dev_priv, enum pipe pipe,
+				    int id, int set, enum drm_scaling_filter filter)
+{
+	switch (filter) {
+	case DRM_SCALING_FILTER_DEFAULT:
+		break;
+	case DRM_SCALING_FILTER_NEAREST_NEIGHBOR:
+		glk_program_nearest_filter_coefs(dev_priv, pipe, id, set);
+		break;
+	default:
+		MISSING_CASE(filter);
+	}
+}
+
+void skl_pfit_enable(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_crtc_scaler_state *scaler_state =
+		&crtc_state->scaler_state;
+	const struct drm_rect *dst = &crtc_state->pch_pfit.dst;
+	u16 uv_rgb_hphase, uv_rgb_vphase;
+	enum pipe pipe = crtc->pipe;
+	int width = drm_rect_width(dst);
+	int height = drm_rect_height(dst);
+	int x = dst->x1;
+	int y = dst->y1;
+	int hscale, vscale;
+	struct drm_rect src;
+	int id;
+	u32 ps_ctrl;
+
+	if (!crtc_state->pch_pfit.enabled)
+		return;
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			crtc_state->scaler_state.scaler_id < 0))
+		return;
+
+	drm_rect_init(&src, 0, 0,
+		      drm_rect_width(&crtc_state->pipe_src) << 16,
+		      drm_rect_height(&crtc_state->pipe_src) << 16);
+
+	hscale = drm_rect_calc_hscale(&src, dst, 0, INT_MAX);
+	vscale = drm_rect_calc_vscale(&src, dst, 0, INT_MAX);
+
+	uv_rgb_hphase = skl_scaler_calc_phase(1, hscale, false);
+	uv_rgb_vphase = skl_scaler_calc_phase(1, vscale, false);
+
+	id = scaler_state->scaler_id;
+
+	ps_ctrl = skl_scaler_get_filter_select(crtc_state->hw.scaling_filter, 0);
+	ps_ctrl |=  PS_SCALER_EN | scaler_state->scalers[id].mode;
+
+	skl_scaler_setup_filter(dev_priv, pipe, id, 0,
+				crtc_state->hw.scaling_filter);
+
+	intel_de_write_fw(dev_priv, SKL_PS_CTRL(pipe, id), ps_ctrl);
+
+	intel_de_write_fw(dev_priv, SKL_PS_VPHASE(pipe, id),
+			  PS_Y_PHASE(0) | PS_UV_RGB_PHASE(uv_rgb_vphase));
+	intel_de_write_fw(dev_priv, SKL_PS_HPHASE(pipe, id),
+			  PS_Y_PHASE(0) | PS_UV_RGB_PHASE(uv_rgb_hphase));
+	intel_de_write_fw(dev_priv, SKL_PS_WIN_POS(pipe, id),
+			  x << 16 | y);
+	intel_de_write_fw(dev_priv, SKL_PS_WIN_SZ(pipe, id),
+			  width << 16 | height);
+}
+
+void
+skl_program_plane_scaler(struct intel_plane *plane,
+			 const struct intel_crtc_state *crtc_state,
+			 const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	enum pipe pipe = plane->pipe;
+	int scaler_id = plane_state->scaler_id;
+	const struct intel_scaler *scaler =
+		&crtc_state->scaler_state.scalers[scaler_id];
+	int crtc_x = plane_state->uapi.dst.x1;
+	int crtc_y = plane_state->uapi.dst.y1;
+	u32 crtc_w = drm_rect_width(&plane_state->uapi.dst);
+	u32 crtc_h = drm_rect_height(&plane_state->uapi.dst);
+	u16 y_hphase, uv_rgb_hphase;
+	u16 y_vphase, uv_rgb_vphase;
+	int hscale, vscale;
+	u32 ps_ctrl;
+
+	hscale = drm_rect_calc_hscale(&plane_state->uapi.src,
+				      &plane_state->uapi.dst,
+				      0, INT_MAX);
+	vscale = drm_rect_calc_vscale(&plane_state->uapi.src,
+				      &plane_state->uapi.dst,
+				      0, INT_MAX);
+
+	/* TODO: handle sub-pixel coordinates */
+	if (intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier) &&
+	    !icl_is_hdr_plane(dev_priv, plane->id)) {
+		y_hphase = skl_scaler_calc_phase(1, hscale, false);
+		y_vphase = skl_scaler_calc_phase(1, vscale, false);
+
+		/* MPEG2 chroma siting convention */
+		uv_rgb_hphase = skl_scaler_calc_phase(2, hscale, true);
+		uv_rgb_vphase = skl_scaler_calc_phase(2, vscale, false);
+	} else {
+		/* not used */
+		y_hphase = 0;
+		y_vphase = 0;
+
+		uv_rgb_hphase = skl_scaler_calc_phase(1, hscale, false);
+		uv_rgb_vphase = skl_scaler_calc_phase(1, vscale, false);
+	}
+
+	ps_ctrl = skl_scaler_get_filter_select(plane_state->hw.scaling_filter, 0);
+	ps_ctrl |= PS_SCALER_EN | PS_PLANE_SEL(plane->id) | scaler->mode;
+
+	skl_scaler_setup_filter(dev_priv, pipe, scaler_id, 0,
+				plane_state->hw.scaling_filter);
+
+	intel_de_write_fw(dev_priv, SKL_PS_CTRL(pipe, scaler_id), ps_ctrl);
+	intel_de_write_fw(dev_priv, SKL_PS_VPHASE(pipe, scaler_id),
+			  PS_Y_PHASE(y_vphase) | PS_UV_RGB_PHASE(uv_rgb_vphase));
+	intel_de_write_fw(dev_priv, SKL_PS_HPHASE(pipe, scaler_id),
+			  PS_Y_PHASE(y_hphase) | PS_UV_RGB_PHASE(uv_rgb_hphase));
+	intel_de_write_fw(dev_priv, SKL_PS_WIN_POS(pipe, scaler_id),
+			  (crtc_x << 16) | crtc_y);
+	intel_de_write_fw(dev_priv, SKL_PS_WIN_SZ(pipe, scaler_id),
+			  (crtc_w << 16) | crtc_h);
+}
+
+static void skl_detach_scaler(struct intel_crtc *crtc, int id)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	intel_de_write_fw(dev_priv, SKL_PS_CTRL(crtc->pipe, id), 0);
+	intel_de_write_fw(dev_priv, SKL_PS_WIN_POS(crtc->pipe, id), 0);
+	intel_de_write_fw(dev_priv, SKL_PS_WIN_SZ(crtc->pipe, id), 0);
+}
+
+/*
+ * This function detaches (aka. unbinds) unused scalers in hardware
+ */
+void skl_detach_scalers(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	const struct intel_crtc_scaler_state *scaler_state =
+		&crtc_state->scaler_state;
+	int i;
+
+	/* loop through and disable scalers that aren't in use */
+	for (i = 0; i < crtc->num_scalers; i++) {
+		if (!scaler_state->scalers[i].in_use)
+			skl_detach_scaler(crtc, i);
+	}
+}
+
+void skl_scaler_disable(const struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->uapi.crtc);
+	int i;
+
+	for (i = 0; i < crtc->num_scalers; i++)
+		skl_detach_scaler(crtc, i);
+}
diff --git a/drivers/gpu/drm/i915/display/skl_scaler.h b/drivers/gpu/drm/i915/display/skl_scaler.h
new file mode 100644
index 000000000..f040f6ac0
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/skl_scaler.h
@@ -0,0 +1,35 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+#ifndef INTEL_SCALER_H
+#define INTEL_SCALER_H
+
+#include <linux/types.h>
+
+enum drm_scaling_filter;
+enum pipe;
+struct drm_i915_private;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_plane;
+struct intel_plane_state;
+
+int skl_update_scaler_crtc(struct intel_crtc_state *crtc_state);
+
+int skl_update_scaler_plane(struct intel_crtc_state *crtc_state,
+			    struct intel_plane_state *plane_state);
+
+int intel_atomic_setup_scalers(struct drm_i915_private *dev_priv,
+			       struct intel_crtc *intel_crtc,
+			       struct intel_crtc_state *crtc_state);
+
+void skl_pfit_enable(const struct intel_crtc_state *crtc_state);
+
+void skl_program_plane_scaler(struct intel_plane *plane,
+			      const struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state);
+void skl_detach_scalers(const struct intel_crtc_state *crtc_state);
+void skl_scaler_disable(const struct intel_crtc_state *old_crtc_state);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/skl_universal_plane.c b/drivers/gpu/drm/i915/display/skl_universal_plane.c
new file mode 100644
index 000000000..bc523a3d1
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/skl_universal_plane.c
@@ -0,0 +1,2525 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_blend.h>
+#include <drm/drm_damage_helper.h>
+#include <drm/drm_fourcc.h>
+
+#include "i915_drv.h"
+#include "intel_atomic_plane.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_fb.h"
+#include "intel_fbc.h"
+#include "intel_psr.h"
+#include "intel_sprite.h"
+#include "skl_scaler.h"
+#include "skl_universal_plane.h"
+#include "skl_watermark.h"
+#include "pxp/intel_pxp.h"
+
+static const u32 skl_plane_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_XRGB16161616F,
+	DRM_FORMAT_XBGR16161616F,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+	DRM_FORMAT_XYUV8888,
+};
+
+static const u32 skl_planar_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_XRGB16161616F,
+	DRM_FORMAT_XBGR16161616F,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_XYUV8888,
+};
+
+static const u32 glk_planar_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_XRGB16161616F,
+	DRM_FORMAT_XBGR16161616F,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_XYUV8888,
+	DRM_FORMAT_P010,
+	DRM_FORMAT_P012,
+	DRM_FORMAT_P016,
+};
+
+static const u32 icl_sdr_y_plane_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_ARGB2101010,
+	DRM_FORMAT_ABGR2101010,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+	DRM_FORMAT_Y210,
+	DRM_FORMAT_Y212,
+	DRM_FORMAT_Y216,
+	DRM_FORMAT_XYUV8888,
+	DRM_FORMAT_XVYU2101010,
+	DRM_FORMAT_XVYU12_16161616,
+	DRM_FORMAT_XVYU16161616,
+};
+
+static const u32 icl_sdr_uv_plane_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_ARGB2101010,
+	DRM_FORMAT_ABGR2101010,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_P010,
+	DRM_FORMAT_P012,
+	DRM_FORMAT_P016,
+	DRM_FORMAT_Y210,
+	DRM_FORMAT_Y212,
+	DRM_FORMAT_Y216,
+	DRM_FORMAT_XYUV8888,
+	DRM_FORMAT_XVYU2101010,
+	DRM_FORMAT_XVYU12_16161616,
+	DRM_FORMAT_XVYU16161616,
+};
+
+static const u32 icl_hdr_plane_formats[] = {
+	DRM_FORMAT_C8,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_XRGB2101010,
+	DRM_FORMAT_XBGR2101010,
+	DRM_FORMAT_ARGB2101010,
+	DRM_FORMAT_ABGR2101010,
+	DRM_FORMAT_XRGB16161616F,
+	DRM_FORMAT_XBGR16161616F,
+	DRM_FORMAT_ARGB16161616F,
+	DRM_FORMAT_ABGR16161616F,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_VYUY,
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_P010,
+	DRM_FORMAT_P012,
+	DRM_FORMAT_P016,
+	DRM_FORMAT_Y210,
+	DRM_FORMAT_Y212,
+	DRM_FORMAT_Y216,
+	DRM_FORMAT_XYUV8888,
+	DRM_FORMAT_XVYU2101010,
+	DRM_FORMAT_XVYU12_16161616,
+	DRM_FORMAT_XVYU16161616,
+};
+
+int skl_format_to_fourcc(int format, bool rgb_order, bool alpha)
+{
+	switch (format) {
+	case PLANE_CTL_FORMAT_RGB_565:
+		return DRM_FORMAT_RGB565;
+	case PLANE_CTL_FORMAT_NV12:
+		return DRM_FORMAT_NV12;
+	case PLANE_CTL_FORMAT_XYUV:
+		return DRM_FORMAT_XYUV8888;
+	case PLANE_CTL_FORMAT_P010:
+		return DRM_FORMAT_P010;
+	case PLANE_CTL_FORMAT_P012:
+		return DRM_FORMAT_P012;
+	case PLANE_CTL_FORMAT_P016:
+		return DRM_FORMAT_P016;
+	case PLANE_CTL_FORMAT_Y210:
+		return DRM_FORMAT_Y210;
+	case PLANE_CTL_FORMAT_Y212:
+		return DRM_FORMAT_Y212;
+	case PLANE_CTL_FORMAT_Y216:
+		return DRM_FORMAT_Y216;
+	case PLANE_CTL_FORMAT_Y410:
+		return DRM_FORMAT_XVYU2101010;
+	case PLANE_CTL_FORMAT_Y412:
+		return DRM_FORMAT_XVYU12_16161616;
+	case PLANE_CTL_FORMAT_Y416:
+		return DRM_FORMAT_XVYU16161616;
+	default:
+	case PLANE_CTL_FORMAT_XRGB_8888:
+		if (rgb_order) {
+			if (alpha)
+				return DRM_FORMAT_ABGR8888;
+			else
+				return DRM_FORMAT_XBGR8888;
+		} else {
+			if (alpha)
+				return DRM_FORMAT_ARGB8888;
+			else
+				return DRM_FORMAT_XRGB8888;
+		}
+	case PLANE_CTL_FORMAT_XRGB_2101010:
+		if (rgb_order) {
+			if (alpha)
+				return DRM_FORMAT_ABGR2101010;
+			else
+				return DRM_FORMAT_XBGR2101010;
+		} else {
+			if (alpha)
+				return DRM_FORMAT_ARGB2101010;
+			else
+				return DRM_FORMAT_XRGB2101010;
+		}
+	case PLANE_CTL_FORMAT_XRGB_16161616F:
+		if (rgb_order) {
+			if (alpha)
+				return DRM_FORMAT_ABGR16161616F;
+			else
+				return DRM_FORMAT_XBGR16161616F;
+		} else {
+			if (alpha)
+				return DRM_FORMAT_ARGB16161616F;
+			else
+				return DRM_FORMAT_XRGB16161616F;
+		}
+	}
+}
+
+static u8 icl_nv12_y_plane_mask(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 13 || HAS_D12_PLANE_MINIMIZATION(i915))
+		return BIT(PLANE_SPRITE2) | BIT(PLANE_SPRITE3);
+	else
+		return BIT(PLANE_SPRITE4) | BIT(PLANE_SPRITE5);
+}
+
+bool icl_is_nv12_y_plane(struct drm_i915_private *dev_priv,
+			 enum plane_id plane_id)
+{
+	return DISPLAY_VER(dev_priv) >= 11 &&
+		icl_nv12_y_plane_mask(dev_priv) & BIT(plane_id);
+}
+
+bool icl_is_hdr_plane(struct drm_i915_private *dev_priv, enum plane_id plane_id)
+{
+	return DISPLAY_VER(dev_priv) >= 11 &&
+		icl_hdr_plane_mask() & BIT(plane_id);
+}
+
+static int icl_plane_min_cdclk(const struct intel_crtc_state *crtc_state,
+			       const struct intel_plane_state *plane_state)
+{
+	unsigned int pixel_rate = intel_plane_pixel_rate(crtc_state, plane_state);
+
+	/* two pixels per clock */
+	return DIV_ROUND_UP(pixel_rate, 2);
+}
+
+static void
+glk_plane_ratio(const struct intel_plane_state *plane_state,
+		unsigned int *num, unsigned int *den)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+
+	if (fb->format->cpp[0] == 8) {
+		*num = 10;
+		*den = 8;
+	} else {
+		*num = 1;
+		*den = 1;
+	}
+}
+
+static int glk_plane_min_cdclk(const struct intel_crtc_state *crtc_state,
+			       const struct intel_plane_state *plane_state)
+{
+	unsigned int pixel_rate = intel_plane_pixel_rate(crtc_state, plane_state);
+	unsigned int num, den;
+
+	glk_plane_ratio(plane_state, &num, &den);
+
+	/* two pixels per clock */
+	return DIV_ROUND_UP(pixel_rate * num, 2 * den);
+}
+
+static void
+skl_plane_ratio(const struct intel_plane_state *plane_state,
+		unsigned int *num, unsigned int *den)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+
+	if (fb->format->cpp[0] == 8) {
+		*num = 9;
+		*den = 8;
+	} else {
+		*num = 1;
+		*den = 1;
+	}
+}
+
+static int skl_plane_min_cdclk(const struct intel_crtc_state *crtc_state,
+			       const struct intel_plane_state *plane_state)
+{
+	unsigned int pixel_rate = intel_plane_pixel_rate(crtc_state, plane_state);
+	unsigned int num, den;
+
+	skl_plane_ratio(plane_state, &num, &den);
+
+	return DIV_ROUND_UP(pixel_rate * num, den);
+}
+
+static int skl_plane_max_width(const struct drm_framebuffer *fb,
+			       int color_plane,
+			       unsigned int rotation)
+{
+	int cpp = fb->format->cpp[color_plane];
+
+	switch (fb->modifier) {
+	case DRM_FORMAT_MOD_LINEAR:
+	case I915_FORMAT_MOD_X_TILED:
+		/*
+		 * Validated limit is 4k, but has 5k should
+		 * work apart from the following features:
+		 * - Ytile (already limited to 4k)
+		 * - FP16 (already limited to 4k)
+		 * - render compression (already limited to 4k)
+		 * - KVMR sprite and cursor (don't care)
+		 * - horizontal panning (TODO verify this)
+		 * - pipe and plane scaling (TODO verify this)
+		 */
+		if (cpp == 8)
+			return 4096;
+		else
+			return 5120;
+	case I915_FORMAT_MOD_Y_TILED_CCS:
+	case I915_FORMAT_MOD_Yf_TILED_CCS:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS:
+		/* FIXME AUX plane? */
+	case I915_FORMAT_MOD_Y_TILED:
+	case I915_FORMAT_MOD_Yf_TILED:
+		if (cpp == 8)
+			return 2048;
+		else
+			return 4096;
+	default:
+		MISSING_CASE(fb->modifier);
+		return 2048;
+	}
+}
+
+static int glk_plane_max_width(const struct drm_framebuffer *fb,
+			       int color_plane,
+			       unsigned int rotation)
+{
+	int cpp = fb->format->cpp[color_plane];
+
+	switch (fb->modifier) {
+	case DRM_FORMAT_MOD_LINEAR:
+	case I915_FORMAT_MOD_X_TILED:
+		if (cpp == 8)
+			return 4096;
+		else
+			return 5120;
+	case I915_FORMAT_MOD_Y_TILED_CCS:
+	case I915_FORMAT_MOD_Yf_TILED_CCS:
+		/* FIXME AUX plane? */
+	case I915_FORMAT_MOD_Y_TILED:
+	case I915_FORMAT_MOD_Yf_TILED:
+		if (cpp == 8)
+			return 2048;
+		else
+			return 5120;
+	default:
+		MISSING_CASE(fb->modifier);
+		return 2048;
+	}
+}
+
+static int icl_plane_min_width(const struct drm_framebuffer *fb,
+			       int color_plane,
+			       unsigned int rotation)
+{
+	/* Wa_14011264657, Wa_14011050563: gen11+ */
+	switch (fb->format->format) {
+	case DRM_FORMAT_C8:
+		return 18;
+	case DRM_FORMAT_RGB565:
+		return 10;
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_ARGB8888:
+	case DRM_FORMAT_ABGR8888:
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_XBGR2101010:
+	case DRM_FORMAT_ARGB2101010:
+	case DRM_FORMAT_ABGR2101010:
+	case DRM_FORMAT_XVYU2101010:
+	case DRM_FORMAT_Y212:
+	case DRM_FORMAT_Y216:
+		return 6;
+	case DRM_FORMAT_NV12:
+		return 20;
+	case DRM_FORMAT_P010:
+	case DRM_FORMAT_P012:
+	case DRM_FORMAT_P016:
+		return 12;
+	case DRM_FORMAT_XRGB16161616F:
+	case DRM_FORMAT_XBGR16161616F:
+	case DRM_FORMAT_ARGB16161616F:
+	case DRM_FORMAT_ABGR16161616F:
+	case DRM_FORMAT_XVYU12_16161616:
+	case DRM_FORMAT_XVYU16161616:
+		return 4;
+	default:
+		return 1;
+	}
+}
+
+static int icl_hdr_plane_max_width(const struct drm_framebuffer *fb,
+				   int color_plane,
+				   unsigned int rotation)
+{
+	if (intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier))
+		return 4096;
+	else
+		return 5120;
+}
+
+static int icl_sdr_plane_max_width(const struct drm_framebuffer *fb,
+				   int color_plane,
+				   unsigned int rotation)
+{
+	return 5120;
+}
+
+static int skl_plane_max_height(const struct drm_framebuffer *fb,
+				int color_plane,
+				unsigned int rotation)
+{
+	return 4096;
+}
+
+static int icl_plane_max_height(const struct drm_framebuffer *fb,
+				int color_plane,
+				unsigned int rotation)
+{
+	return 4320;
+}
+
+static unsigned int
+skl_plane_max_stride(struct intel_plane *plane,
+		     u32 pixel_format, u64 modifier,
+		     unsigned int rotation)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	const struct drm_format_info *info = drm_format_info(pixel_format);
+	int cpp = info->cpp[0];
+	int max_horizontal_pixels = 8192;
+	int max_stride_bytes;
+
+	if (DISPLAY_VER(i915) >= 13) {
+		/*
+		 * The stride in bytes must not exceed of the size
+		 * of 128K bytes. For pixel formats of 64bpp will allow
+		 * for a 16K pixel surface.
+		 */
+		max_stride_bytes = 131072;
+		if (cpp == 8)
+			max_horizontal_pixels = 16384;
+		else
+			max_horizontal_pixels = 65536;
+	} else {
+		/*
+		 * "The stride in bytes must not exceed the
+		 * of the size of 8K pixels and 32K bytes."
+		 */
+		max_stride_bytes = 32768;
+	}
+
+	if (drm_rotation_90_or_270(rotation))
+		return min(max_horizontal_pixels, max_stride_bytes / cpp);
+	else
+		return min(max_horizontal_pixels * cpp, max_stride_bytes);
+}
+
+
+/* Preoffset values for YUV to RGB Conversion */
+#define PREOFF_YUV_TO_RGB_HI		0x1800
+#define PREOFF_YUV_TO_RGB_ME		0x0000
+#define PREOFF_YUV_TO_RGB_LO		0x1800
+
+#define  ROFF(x)          (((x) & 0xffff) << 16)
+#define  GOFF(x)          (((x) & 0xffff) << 0)
+#define  BOFF(x)          (((x) & 0xffff) << 16)
+
+/*
+ * Programs the input color space conversion stage for ICL HDR planes.
+ * Note that it is assumed that this stage always happens after YUV
+ * range correction. Thus, the input to this stage is assumed to be
+ * in full-range YCbCr.
+ */
+static void
+icl_program_input_csc(struct intel_plane *plane,
+		      const struct intel_crtc_state *crtc_state,
+		      const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	enum plane_id plane_id = plane->id;
+
+	static const u16 input_csc_matrix[][9] = {
+		/*
+		 * BT.601 full range YCbCr -> full range RGB
+		 * The matrix required is :
+		 * [1.000, 0.000, 1.371,
+		 *  1.000, -0.336, -0.698,
+		 *  1.000, 1.732, 0.0000]
+		 */
+		[DRM_COLOR_YCBCR_BT601] = {
+			0x7AF8, 0x7800, 0x0,
+			0x8B28, 0x7800, 0x9AC0,
+			0x0, 0x7800, 0x7DD8,
+		},
+		/*
+		 * BT.709 full range YCbCr -> full range RGB
+		 * The matrix required is :
+		 * [1.000, 0.000, 1.574,
+		 *  1.000, -0.187, -0.468,
+		 *  1.000, 1.855, 0.0000]
+		 */
+		[DRM_COLOR_YCBCR_BT709] = {
+			0x7C98, 0x7800, 0x0,
+			0x9EF8, 0x7800, 0xAC00,
+			0x0, 0x7800,  0x7ED8,
+		},
+		/*
+		 * BT.2020 full range YCbCr -> full range RGB
+		 * The matrix required is :
+		 * [1.000, 0.000, 1.474,
+		 *  1.000, -0.1645, -0.5713,
+		 *  1.000, 1.8814, 0.0000]
+		 */
+		[DRM_COLOR_YCBCR_BT2020] = {
+			0x7BC8, 0x7800, 0x0,
+			0x8928, 0x7800, 0xAA88,
+			0x0, 0x7800, 0x7F10,
+		},
+	};
+	const u16 *csc = input_csc_matrix[plane_state->hw.color_encoding];
+
+	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_COEFF(pipe, plane_id, 0),
+			  ROFF(csc[0]) | GOFF(csc[1]));
+	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_COEFF(pipe, plane_id, 1),
+			  BOFF(csc[2]));
+	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_COEFF(pipe, plane_id, 2),
+			  ROFF(csc[3]) | GOFF(csc[4]));
+	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_COEFF(pipe, plane_id, 3),
+			  BOFF(csc[5]));
+	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_COEFF(pipe, plane_id, 4),
+			  ROFF(csc[6]) | GOFF(csc[7]));
+	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_COEFF(pipe, plane_id, 5),
+			  BOFF(csc[8]));
+
+	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_PREOFF(pipe, plane_id, 0),
+			  PREOFF_YUV_TO_RGB_HI);
+	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_PREOFF(pipe, plane_id, 1),
+			  PREOFF_YUV_TO_RGB_ME);
+	intel_de_write_fw(dev_priv, PLANE_INPUT_CSC_PREOFF(pipe, plane_id, 2),
+			  PREOFF_YUV_TO_RGB_LO);
+	intel_de_write_fw(dev_priv,
+			  PLANE_INPUT_CSC_POSTOFF(pipe, plane_id, 0), 0x0);
+	intel_de_write_fw(dev_priv,
+			  PLANE_INPUT_CSC_POSTOFF(pipe, plane_id, 1), 0x0);
+	intel_de_write_fw(dev_priv,
+			  PLANE_INPUT_CSC_POSTOFF(pipe, plane_id, 2), 0x0);
+}
+
+static unsigned int skl_plane_stride_mult(const struct drm_framebuffer *fb,
+					  int color_plane, unsigned int rotation)
+{
+	/*
+	 * The stride is either expressed as a multiple of 64 bytes chunks for
+	 * linear buffers or in number of tiles for tiled buffers.
+	 */
+	if (is_surface_linear(fb, color_plane))
+		return 64;
+	else if (drm_rotation_90_or_270(rotation))
+		return intel_tile_height(fb, color_plane);
+	else
+		return intel_tile_width_bytes(fb, color_plane);
+}
+
+static u32 skl_plane_stride(const struct intel_plane_state *plane_state,
+			    int color_plane)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+	u32 stride = plane_state->view.color_plane[color_plane].scanout_stride;
+
+	if (color_plane >= fb->format->num_planes)
+		return 0;
+
+	return stride / skl_plane_stride_mult(fb, color_plane, rotation);
+}
+
+static void
+skl_plane_disable_arm(struct intel_plane *plane,
+		      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe = plane->pipe;
+
+	skl_write_plane_wm(plane, crtc_state);
+
+	intel_de_write_fw(dev_priv, PLANE_CTL(pipe, plane_id), 0);
+	intel_de_write_fw(dev_priv, PLANE_SURF(pipe, plane_id), 0);
+}
+
+static void
+icl_plane_disable_arm(struct intel_plane *plane,
+		      const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe = plane->pipe;
+
+	if (icl_is_hdr_plane(dev_priv, plane_id))
+		intel_de_write_fw(dev_priv, PLANE_CUS_CTL(pipe, plane_id), 0);
+
+	skl_write_plane_wm(plane, crtc_state);
+
+	intel_psr2_disable_plane_sel_fetch(plane, crtc_state);
+	intel_de_write_fw(dev_priv, PLANE_CTL(pipe, plane_id), 0);
+	intel_de_write_fw(dev_priv, PLANE_SURF(pipe, plane_id), 0);
+}
+
+static bool
+skl_plane_get_hw_state(struct intel_plane *plane,
+		       enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum intel_display_power_domain power_domain;
+	enum plane_id plane_id = plane->id;
+	intel_wakeref_t wakeref;
+	bool ret;
+
+	power_domain = POWER_DOMAIN_PIPE(plane->pipe);
+	wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain);
+	if (!wakeref)
+		return false;
+
+	ret = intel_de_read(dev_priv, PLANE_CTL(plane->pipe, plane_id)) & PLANE_CTL_ENABLE;
+
+	*pipe = plane->pipe;
+
+	intel_display_power_put(dev_priv, power_domain, wakeref);
+
+	return ret;
+}
+
+static u32 skl_plane_ctl_format(u32 pixel_format)
+{
+	switch (pixel_format) {
+	case DRM_FORMAT_C8:
+		return PLANE_CTL_FORMAT_INDEXED;
+	case DRM_FORMAT_RGB565:
+		return PLANE_CTL_FORMAT_RGB_565;
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_ABGR8888:
+		return PLANE_CTL_FORMAT_XRGB_8888 | PLANE_CTL_ORDER_RGBX;
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_ARGB8888:
+		return PLANE_CTL_FORMAT_XRGB_8888;
+	case DRM_FORMAT_XBGR2101010:
+	case DRM_FORMAT_ABGR2101010:
+		return PLANE_CTL_FORMAT_XRGB_2101010 | PLANE_CTL_ORDER_RGBX;
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_ARGB2101010:
+		return PLANE_CTL_FORMAT_XRGB_2101010;
+	case DRM_FORMAT_XBGR16161616F:
+	case DRM_FORMAT_ABGR16161616F:
+		return PLANE_CTL_FORMAT_XRGB_16161616F | PLANE_CTL_ORDER_RGBX;
+	case DRM_FORMAT_XRGB16161616F:
+	case DRM_FORMAT_ARGB16161616F:
+		return PLANE_CTL_FORMAT_XRGB_16161616F;
+	case DRM_FORMAT_XYUV8888:
+		return PLANE_CTL_FORMAT_XYUV;
+	case DRM_FORMAT_YUYV:
+		return PLANE_CTL_FORMAT_YUV422 | PLANE_CTL_YUV422_ORDER_YUYV;
+	case DRM_FORMAT_YVYU:
+		return PLANE_CTL_FORMAT_YUV422 | PLANE_CTL_YUV422_ORDER_YVYU;
+	case DRM_FORMAT_UYVY:
+		return PLANE_CTL_FORMAT_YUV422 | PLANE_CTL_YUV422_ORDER_UYVY;
+	case DRM_FORMAT_VYUY:
+		return PLANE_CTL_FORMAT_YUV422 | PLANE_CTL_YUV422_ORDER_VYUY;
+	case DRM_FORMAT_NV12:
+		return PLANE_CTL_FORMAT_NV12;
+	case DRM_FORMAT_P010:
+		return PLANE_CTL_FORMAT_P010;
+	case DRM_FORMAT_P012:
+		return PLANE_CTL_FORMAT_P012;
+	case DRM_FORMAT_P016:
+		return PLANE_CTL_FORMAT_P016;
+	case DRM_FORMAT_Y210:
+		return PLANE_CTL_FORMAT_Y210;
+	case DRM_FORMAT_Y212:
+		return PLANE_CTL_FORMAT_Y212;
+	case DRM_FORMAT_Y216:
+		return PLANE_CTL_FORMAT_Y216;
+	case DRM_FORMAT_XVYU2101010:
+		return PLANE_CTL_FORMAT_Y410;
+	case DRM_FORMAT_XVYU12_16161616:
+		return PLANE_CTL_FORMAT_Y412;
+	case DRM_FORMAT_XVYU16161616:
+		return PLANE_CTL_FORMAT_Y416;
+	default:
+		MISSING_CASE(pixel_format);
+	}
+
+	return 0;
+}
+
+static u32 skl_plane_ctl_alpha(const struct intel_plane_state *plane_state)
+{
+	if (!plane_state->hw.fb->format->has_alpha)
+		return PLANE_CTL_ALPHA_DISABLE;
+
+	switch (plane_state->hw.pixel_blend_mode) {
+	case DRM_MODE_BLEND_PIXEL_NONE:
+		return PLANE_CTL_ALPHA_DISABLE;
+	case DRM_MODE_BLEND_PREMULTI:
+		return PLANE_CTL_ALPHA_SW_PREMULTIPLY;
+	case DRM_MODE_BLEND_COVERAGE:
+		return PLANE_CTL_ALPHA_HW_PREMULTIPLY;
+	default:
+		MISSING_CASE(plane_state->hw.pixel_blend_mode);
+		return PLANE_CTL_ALPHA_DISABLE;
+	}
+}
+
+static u32 glk_plane_color_ctl_alpha(const struct intel_plane_state *plane_state)
+{
+	if (!plane_state->hw.fb->format->has_alpha)
+		return PLANE_COLOR_ALPHA_DISABLE;
+
+	switch (plane_state->hw.pixel_blend_mode) {
+	case DRM_MODE_BLEND_PIXEL_NONE:
+		return PLANE_COLOR_ALPHA_DISABLE;
+	case DRM_MODE_BLEND_PREMULTI:
+		return PLANE_COLOR_ALPHA_SW_PREMULTIPLY;
+	case DRM_MODE_BLEND_COVERAGE:
+		return PLANE_COLOR_ALPHA_HW_PREMULTIPLY;
+	default:
+		MISSING_CASE(plane_state->hw.pixel_blend_mode);
+		return PLANE_COLOR_ALPHA_DISABLE;
+	}
+}
+
+static u32 skl_plane_ctl_tiling(u64 fb_modifier)
+{
+	switch (fb_modifier) {
+	case DRM_FORMAT_MOD_LINEAR:
+		break;
+	case I915_FORMAT_MOD_X_TILED:
+		return PLANE_CTL_TILED_X;
+	case I915_FORMAT_MOD_Y_TILED:
+		return PLANE_CTL_TILED_Y;
+	case I915_FORMAT_MOD_4_TILED:
+		return PLANE_CTL_TILED_4;
+	case I915_FORMAT_MOD_4_TILED_DG2_RC_CCS:
+		return PLANE_CTL_TILED_4 |
+			PLANE_CTL_RENDER_DECOMPRESSION_ENABLE |
+			PLANE_CTL_CLEAR_COLOR_DISABLE;
+	case I915_FORMAT_MOD_4_TILED_DG2_MC_CCS:
+		return PLANE_CTL_TILED_4 |
+			PLANE_CTL_MEDIA_DECOMPRESSION_ENABLE |
+			PLANE_CTL_CLEAR_COLOR_DISABLE;
+	case I915_FORMAT_MOD_4_TILED_DG2_RC_CCS_CC:
+		return PLANE_CTL_TILED_4 | PLANE_CTL_RENDER_DECOMPRESSION_ENABLE;
+	case I915_FORMAT_MOD_Y_TILED_CCS:
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC:
+		return PLANE_CTL_TILED_Y | PLANE_CTL_RENDER_DECOMPRESSION_ENABLE;
+	case I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS:
+		return PLANE_CTL_TILED_Y |
+		       PLANE_CTL_RENDER_DECOMPRESSION_ENABLE |
+		       PLANE_CTL_CLEAR_COLOR_DISABLE;
+	case I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS:
+		return PLANE_CTL_TILED_Y | PLANE_CTL_MEDIA_DECOMPRESSION_ENABLE;
+	case I915_FORMAT_MOD_Yf_TILED:
+		return PLANE_CTL_TILED_YF;
+	case I915_FORMAT_MOD_Yf_TILED_CCS:
+		return PLANE_CTL_TILED_YF | PLANE_CTL_RENDER_DECOMPRESSION_ENABLE;
+	default:
+		MISSING_CASE(fb_modifier);
+	}
+
+	return 0;
+}
+
+static u32 skl_plane_ctl_rotate(unsigned int rotate)
+{
+	switch (rotate) {
+	case DRM_MODE_ROTATE_0:
+		break;
+	/*
+	 * DRM_MODE_ROTATE_ is counter clockwise to stay compatible with Xrandr
+	 * while i915 HW rotation is clockwise, thats why this swapping.
+	 */
+	case DRM_MODE_ROTATE_90:
+		return PLANE_CTL_ROTATE_270;
+	case DRM_MODE_ROTATE_180:
+		return PLANE_CTL_ROTATE_180;
+	case DRM_MODE_ROTATE_270:
+		return PLANE_CTL_ROTATE_90;
+	default:
+		MISSING_CASE(rotate);
+	}
+
+	return 0;
+}
+
+static u32 icl_plane_ctl_flip(unsigned int reflect)
+{
+	switch (reflect) {
+	case 0:
+		break;
+	case DRM_MODE_REFLECT_X:
+		return PLANE_CTL_FLIP_HORIZONTAL;
+	case DRM_MODE_REFLECT_Y:
+	default:
+		MISSING_CASE(reflect);
+	}
+
+	return 0;
+}
+
+static u32 adlp_plane_ctl_arb_slots(const struct intel_plane_state *plane_state)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+
+	if (intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier)) {
+		switch (fb->format->cpp[0]) {
+		case 2:
+			return PLANE_CTL_ARB_SLOTS(1);
+		default:
+			return PLANE_CTL_ARB_SLOTS(0);
+		}
+	} else {
+		switch (fb->format->cpp[0]) {
+		case 8:
+			return PLANE_CTL_ARB_SLOTS(3);
+		case 4:
+			return PLANE_CTL_ARB_SLOTS(1);
+		default:
+			return PLANE_CTL_ARB_SLOTS(0);
+		}
+	}
+}
+
+static u32 skl_plane_ctl_crtc(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	u32 plane_ctl = 0;
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		return plane_ctl;
+
+	if (crtc_state->gamma_enable)
+		plane_ctl |= PLANE_CTL_PIPE_GAMMA_ENABLE;
+
+	if (crtc_state->csc_enable)
+		plane_ctl |= PLANE_CTL_PIPE_CSC_ENABLE;
+
+	return plane_ctl;
+}
+
+static u32 skl_plane_ctl(const struct intel_crtc_state *crtc_state,
+			 const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+	const struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+	u32 plane_ctl;
+
+	plane_ctl = PLANE_CTL_ENABLE;
+
+	if (DISPLAY_VER(dev_priv) < 10) {
+		plane_ctl |= skl_plane_ctl_alpha(plane_state);
+		plane_ctl |= PLANE_CTL_PLANE_GAMMA_DISABLE;
+
+		if (plane_state->hw.color_encoding == DRM_COLOR_YCBCR_BT709)
+			plane_ctl |= PLANE_CTL_YUV_TO_RGB_CSC_FORMAT_BT709;
+
+		if (plane_state->hw.color_range == DRM_COLOR_YCBCR_FULL_RANGE)
+			plane_ctl |= PLANE_CTL_YUV_RANGE_CORRECTION_DISABLE;
+	}
+
+	plane_ctl |= skl_plane_ctl_format(fb->format->format);
+	plane_ctl |= skl_plane_ctl_tiling(fb->modifier);
+	plane_ctl |= skl_plane_ctl_rotate(rotation & DRM_MODE_ROTATE_MASK);
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		plane_ctl |= icl_plane_ctl_flip(rotation &
+						DRM_MODE_REFLECT_MASK);
+
+	if (key->flags & I915_SET_COLORKEY_DESTINATION)
+		plane_ctl |= PLANE_CTL_KEY_ENABLE_DESTINATION;
+	else if (key->flags & I915_SET_COLORKEY_SOURCE)
+		plane_ctl |= PLANE_CTL_KEY_ENABLE_SOURCE;
+
+	/* Wa_22012358565:adl-p */
+	if (DISPLAY_VER(dev_priv) == 13)
+		plane_ctl |= adlp_plane_ctl_arb_slots(plane_state);
+
+	return plane_ctl;
+}
+
+static u32 glk_plane_color_ctl_crtc(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	u32 plane_color_ctl = 0;
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		return plane_color_ctl;
+
+	if (crtc_state->gamma_enable)
+		plane_color_ctl |= PLANE_COLOR_PIPE_GAMMA_ENABLE;
+
+	if (crtc_state->csc_enable)
+		plane_color_ctl |= PLANE_COLOR_PIPE_CSC_ENABLE;
+
+	return plane_color_ctl;
+}
+
+static u32 glk_plane_color_ctl(const struct intel_crtc_state *crtc_state,
+			       const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	u32 plane_color_ctl = 0;
+
+	plane_color_ctl |= PLANE_COLOR_PLANE_GAMMA_DISABLE;
+	plane_color_ctl |= glk_plane_color_ctl_alpha(plane_state);
+
+	if (fb->format->is_yuv && !icl_is_hdr_plane(dev_priv, plane->id)) {
+		switch (plane_state->hw.color_encoding) {
+		case DRM_COLOR_YCBCR_BT709:
+			plane_color_ctl |= PLANE_COLOR_CSC_MODE_YUV709_TO_RGB709;
+			break;
+		case DRM_COLOR_YCBCR_BT2020:
+			plane_color_ctl |=
+				PLANE_COLOR_CSC_MODE_YUV2020_TO_RGB2020;
+			break;
+		default:
+			plane_color_ctl |=
+				PLANE_COLOR_CSC_MODE_YUV601_TO_RGB601;
+		}
+		if (plane_state->hw.color_range == DRM_COLOR_YCBCR_FULL_RANGE)
+			plane_color_ctl |= PLANE_COLOR_YUV_RANGE_CORRECTION_DISABLE;
+	} else if (fb->format->is_yuv) {
+		plane_color_ctl |= PLANE_COLOR_INPUT_CSC_ENABLE;
+		if (plane_state->hw.color_range == DRM_COLOR_YCBCR_FULL_RANGE)
+			plane_color_ctl |= PLANE_COLOR_YUV_RANGE_CORRECTION_DISABLE;
+	}
+
+	if (plane_state->force_black)
+		plane_color_ctl |= PLANE_COLOR_PLANE_CSC_ENABLE;
+
+	return plane_color_ctl;
+}
+
+static u32 skl_surf_address(const struct intel_plane_state *plane_state,
+			    int color_plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	u32 offset = plane_state->view.color_plane[color_plane].offset;
+
+	if (intel_fb_uses_dpt(fb)) {
+		/*
+		 * The DPT object contains only one vma, so the VMA's offset
+		 * within the DPT is always 0.
+		 */
+		drm_WARN_ON(&i915->drm, plane_state->dpt_vma->node.start);
+		drm_WARN_ON(&i915->drm, offset & 0x1fffff);
+		return offset >> 9;
+	} else {
+		drm_WARN_ON(&i915->drm, offset & 0xfff);
+		return offset;
+	}
+}
+
+static u32 skl_plane_surf(const struct intel_plane_state *plane_state,
+			  int color_plane)
+{
+	u32 plane_surf;
+
+	plane_surf = intel_plane_ggtt_offset(plane_state) +
+		skl_surf_address(plane_state, color_plane);
+
+	if (plane_state->decrypt)
+		plane_surf |= PLANE_SURF_DECRYPT;
+
+	return plane_surf;
+}
+
+static u32 skl_plane_aux_dist(const struct intel_plane_state *plane_state,
+			      int color_plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int aux_plane = skl_main_to_aux_plane(fb, color_plane);
+	u32 aux_dist;
+
+	if (!aux_plane)
+		return 0;
+
+	aux_dist = skl_surf_address(plane_state, aux_plane) -
+		skl_surf_address(plane_state, color_plane);
+
+	if (DISPLAY_VER(i915) < 12)
+		aux_dist |= PLANE_AUX_STRIDE(skl_plane_stride(plane_state, aux_plane));
+
+	return aux_dist;
+}
+
+static u32 skl_plane_keyval(const struct intel_plane_state *plane_state)
+{
+	const struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+
+	return key->min_value;
+}
+
+static u32 skl_plane_keymax(const struct intel_plane_state *plane_state)
+{
+	const struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+	u8 alpha = plane_state->hw.alpha >> 8;
+
+	return (key->max_value & 0xffffff) | PLANE_KEYMAX_ALPHA(alpha);
+}
+
+static u32 skl_plane_keymsk(const struct intel_plane_state *plane_state)
+{
+	const struct drm_intel_sprite_colorkey *key = &plane_state->ckey;
+	u8 alpha = plane_state->hw.alpha >> 8;
+	u32 keymsk;
+
+	keymsk = key->channel_mask & 0x7ffffff;
+	if (alpha < 0xff)
+		keymsk |= PLANE_KEYMSK_ALPHA_ENABLE;
+
+	return keymsk;
+}
+
+static void icl_plane_csc_load_black(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe = plane->pipe;
+
+	intel_de_write_fw(i915, PLANE_CSC_COEFF(pipe, plane_id, 0), 0);
+	intel_de_write_fw(i915, PLANE_CSC_COEFF(pipe, plane_id, 1), 0);
+
+	intel_de_write_fw(i915, PLANE_CSC_COEFF(pipe, plane_id, 2), 0);
+	intel_de_write_fw(i915, PLANE_CSC_COEFF(pipe, plane_id, 3), 0);
+
+	intel_de_write_fw(i915, PLANE_CSC_COEFF(pipe, plane_id, 4), 0);
+	intel_de_write_fw(i915, PLANE_CSC_COEFF(pipe, plane_id, 5), 0);
+
+	intel_de_write_fw(i915, PLANE_CSC_PREOFF(pipe, plane_id, 0), 0);
+	intel_de_write_fw(i915, PLANE_CSC_PREOFF(pipe, plane_id, 1), 0);
+	intel_de_write_fw(i915, PLANE_CSC_PREOFF(pipe, plane_id, 2), 0);
+
+	intel_de_write_fw(i915, PLANE_CSC_POSTOFF(pipe, plane_id, 0), 0);
+	intel_de_write_fw(i915, PLANE_CSC_POSTOFF(pipe, plane_id, 1), 0);
+	intel_de_write_fw(i915, PLANE_CSC_POSTOFF(pipe, plane_id, 2), 0);
+}
+
+static int icl_plane_color_plane(const struct intel_plane_state *plane_state)
+{
+	/* Program the UV plane on planar master */
+	if (plane_state->planar_linked_plane && !plane_state->planar_slave)
+		return 1;
+	else
+		return 0;
+}
+
+static void
+skl_plane_update_noarm(struct intel_plane *plane,
+		       const struct intel_crtc_state *crtc_state,
+		       const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe = plane->pipe;
+	u32 stride = skl_plane_stride(plane_state, 0);
+	int crtc_x = plane_state->uapi.dst.x1;
+	int crtc_y = plane_state->uapi.dst.y1;
+	u32 src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+	u32 src_h = drm_rect_height(&plane_state->uapi.src) >> 16;
+
+	/* The scaler will handle the output position */
+	if (plane_state->scaler_id >= 0) {
+		crtc_x = 0;
+		crtc_y = 0;
+	}
+
+	intel_de_write_fw(dev_priv, PLANE_STRIDE(pipe, plane_id),
+			  PLANE_STRIDE_(stride));
+	intel_de_write_fw(dev_priv, PLANE_POS(pipe, plane_id),
+			  PLANE_POS_Y(crtc_y) | PLANE_POS_X(crtc_x));
+	intel_de_write_fw(dev_priv, PLANE_SIZE(pipe, plane_id),
+			  PLANE_HEIGHT(src_h - 1) | PLANE_WIDTH(src_w - 1));
+
+	skl_write_plane_wm(plane, crtc_state);
+}
+
+static void
+skl_plane_update_arm(struct intel_plane *plane,
+		     const struct intel_crtc_state *crtc_state,
+		     const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe = plane->pipe;
+	u32 x = plane_state->view.color_plane[0].x;
+	u32 y = plane_state->view.color_plane[0].y;
+	u32 plane_ctl, plane_color_ctl = 0;
+
+	plane_ctl = plane_state->ctl |
+		skl_plane_ctl_crtc(crtc_state);
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		plane_color_ctl = plane_state->color_ctl |
+			glk_plane_color_ctl_crtc(crtc_state);
+
+	intel_de_write_fw(dev_priv, PLANE_KEYVAL(pipe, plane_id), skl_plane_keyval(plane_state));
+	intel_de_write_fw(dev_priv, PLANE_KEYMSK(pipe, plane_id), skl_plane_keymsk(plane_state));
+	intel_de_write_fw(dev_priv, PLANE_KEYMAX(pipe, plane_id), skl_plane_keymax(plane_state));
+
+	intel_de_write_fw(dev_priv, PLANE_OFFSET(pipe, plane_id),
+			  PLANE_OFFSET_Y(y) | PLANE_OFFSET_X(x));
+
+	intel_de_write_fw(dev_priv, PLANE_AUX_DIST(pipe, plane_id),
+			  skl_plane_aux_dist(plane_state, 0));
+
+	intel_de_write_fw(dev_priv, PLANE_AUX_OFFSET(pipe, plane_id),
+			  PLANE_OFFSET_Y(plane_state->view.color_plane[1].y) |
+			  PLANE_OFFSET_X(plane_state->view.color_plane[1].x));
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		intel_de_write_fw(dev_priv, PLANE_COLOR_CTL(pipe, plane_id), plane_color_ctl);
+
+	/*
+	 * Enable the scaler before the plane so that we don't
+	 * get a catastrophic underrun even if the two operations
+	 * end up happening in two different frames.
+	 *
+	 * TODO: split into noarm+arm pair
+	 */
+	if (plane_state->scaler_id >= 0)
+		skl_program_plane_scaler(plane, crtc_state, plane_state);
+
+	/*
+	 * The control register self-arms if the plane was previously
+	 * disabled. Try to make the plane enable atomic by writing
+	 * the control register just before the surface register.
+	 */
+	intel_de_write_fw(dev_priv, PLANE_CTL(pipe, plane_id), plane_ctl);
+	intel_de_write_fw(dev_priv, PLANE_SURF(pipe, plane_id),
+			  skl_plane_surf(plane_state, 0));
+}
+
+static void
+icl_plane_update_noarm(struct intel_plane *plane,
+		       const struct intel_crtc_state *crtc_state,
+		       const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe = plane->pipe;
+	int color_plane = icl_plane_color_plane(plane_state);
+	u32 stride = skl_plane_stride(plane_state, color_plane);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int crtc_x = plane_state->uapi.dst.x1;
+	int crtc_y = plane_state->uapi.dst.y1;
+	int x = plane_state->view.color_plane[color_plane].x;
+	int y = plane_state->view.color_plane[color_plane].y;
+	int src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+	int src_h = drm_rect_height(&plane_state->uapi.src) >> 16;
+	u32 plane_color_ctl;
+
+	plane_color_ctl = plane_state->color_ctl |
+		glk_plane_color_ctl_crtc(crtc_state);
+
+	/* The scaler will handle the output position */
+	if (plane_state->scaler_id >= 0) {
+		crtc_x = 0;
+		crtc_y = 0;
+	}
+
+	intel_de_write_fw(dev_priv, PLANE_STRIDE(pipe, plane_id),
+			  PLANE_STRIDE_(stride));
+	intel_de_write_fw(dev_priv, PLANE_POS(pipe, plane_id),
+			  PLANE_POS_Y(crtc_y) | PLANE_POS_X(crtc_x));
+	intel_de_write_fw(dev_priv, PLANE_SIZE(pipe, plane_id),
+			  PLANE_HEIGHT(src_h - 1) | PLANE_WIDTH(src_w - 1));
+
+	intel_de_write_fw(dev_priv, PLANE_KEYVAL(pipe, plane_id), skl_plane_keyval(plane_state));
+	intel_de_write_fw(dev_priv, PLANE_KEYMSK(pipe, plane_id), skl_plane_keymsk(plane_state));
+	intel_de_write_fw(dev_priv, PLANE_KEYMAX(pipe, plane_id), skl_plane_keymax(plane_state));
+
+	intel_de_write_fw(dev_priv, PLANE_OFFSET(pipe, plane_id),
+			  PLANE_OFFSET_Y(y) | PLANE_OFFSET_X(x));
+
+	if (intel_fb_is_rc_ccs_cc_modifier(fb->modifier)) {
+		intel_de_write_fw(dev_priv, PLANE_CC_VAL(pipe, plane_id, 0),
+				  lower_32_bits(plane_state->ccval));
+		intel_de_write_fw(dev_priv, PLANE_CC_VAL(pipe, plane_id, 1),
+				  upper_32_bits(plane_state->ccval));
+	}
+
+	/* FLAT CCS doesn't need to program AUX_DIST */
+	if (!HAS_FLAT_CCS(dev_priv))
+		intel_de_write_fw(dev_priv, PLANE_AUX_DIST(pipe, plane_id),
+				  skl_plane_aux_dist(plane_state, color_plane));
+
+	if (icl_is_hdr_plane(dev_priv, plane_id))
+		intel_de_write_fw(dev_priv, PLANE_CUS_CTL(pipe, plane_id),
+				  plane_state->cus_ctl);
+
+	intel_de_write_fw(dev_priv, PLANE_COLOR_CTL(pipe, plane_id), plane_color_ctl);
+
+	if (fb->format->is_yuv && icl_is_hdr_plane(dev_priv, plane_id))
+		icl_program_input_csc(plane, crtc_state, plane_state);
+
+	skl_write_plane_wm(plane, crtc_state);
+
+	/*
+	 * FIXME: pxp session invalidation can hit any time even at time of commit
+	 * or after the commit, display content will be garbage.
+	 */
+	if (plane_state->force_black)
+		icl_plane_csc_load_black(plane);
+
+	intel_psr2_program_plane_sel_fetch(plane, crtc_state, plane_state, color_plane);
+}
+
+static void
+icl_plane_update_arm(struct intel_plane *plane,
+		     const struct intel_crtc_state *crtc_state,
+		     const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe = plane->pipe;
+	int color_plane = icl_plane_color_plane(plane_state);
+	u32 plane_ctl;
+
+	plane_ctl = plane_state->ctl |
+		skl_plane_ctl_crtc(crtc_state);
+
+	/*
+	 * Enable the scaler before the plane so that we don't
+	 * get a catastrophic underrun even if the two operations
+	 * end up happening in two different frames.
+	 *
+	 * TODO: split into noarm+arm pair
+	 */
+	if (plane_state->scaler_id >= 0)
+		skl_program_plane_scaler(plane, crtc_state, plane_state);
+
+	/*
+	 * The control register self-arms if the plane was previously
+	 * disabled. Try to make the plane enable atomic by writing
+	 * the control register just before the surface register.
+	 */
+	intel_de_write_fw(dev_priv, PLANE_CTL(pipe, plane_id), plane_ctl);
+	intel_de_write_fw(dev_priv, PLANE_SURF(pipe, plane_id),
+			  skl_plane_surf(plane_state, color_plane));
+}
+
+static void
+skl_plane_async_flip(struct intel_plane *plane,
+		     const struct intel_crtc_state *crtc_state,
+		     const struct intel_plane_state *plane_state,
+		     bool async_flip)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe = plane->pipe;
+	u32 plane_ctl = plane_state->ctl;
+
+	plane_ctl |= skl_plane_ctl_crtc(crtc_state);
+
+	if (async_flip)
+		plane_ctl |= PLANE_CTL_ASYNC_FLIP;
+
+	intel_de_write_fw(dev_priv, PLANE_CTL(pipe, plane_id), plane_ctl);
+	intel_de_write_fw(dev_priv, PLANE_SURF(pipe, plane_id),
+			  skl_plane_surf(plane_state, 0));
+}
+
+static bool intel_format_is_p01x(u32 format)
+{
+	switch (format) {
+	case DRM_FORMAT_P010:
+	case DRM_FORMAT_P012:
+	case DRM_FORMAT_P016:
+		return true;
+	default:
+		return false;
+	}
+}
+
+static int skl_plane_check_fb(const struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+
+	if (!fb)
+		return 0;
+
+	if (rotation & ~(DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180) &&
+	    intel_fb_is_ccs_modifier(fb->modifier)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "RC support only with 0/180 degree rotation (%x)\n",
+			    rotation);
+		return -EINVAL;
+	}
+
+	if (rotation & DRM_MODE_REFLECT_X &&
+	    fb->modifier == DRM_FORMAT_MOD_LINEAR) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "horizontal flip is not supported with linear surface formats\n");
+		return -EINVAL;
+	}
+
+	if (drm_rotation_90_or_270(rotation)) {
+		if (!intel_fb_supports_90_270_rotation(to_intel_framebuffer(fb))) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Y/Yf tiling required for 90/270!\n");
+			return -EINVAL;
+		}
+
+		/*
+		 * 90/270 is not allowed with RGB64 16:16:16:16 and
+		 * Indexed 8-bit. RGB 16-bit 5:6:5 is allowed gen11 onwards.
+		 */
+		switch (fb->format->format) {
+		case DRM_FORMAT_RGB565:
+			if (DISPLAY_VER(dev_priv) >= 11)
+				break;
+			fallthrough;
+		case DRM_FORMAT_C8:
+		case DRM_FORMAT_XRGB16161616F:
+		case DRM_FORMAT_XBGR16161616F:
+		case DRM_FORMAT_ARGB16161616F:
+		case DRM_FORMAT_ABGR16161616F:
+		case DRM_FORMAT_Y210:
+		case DRM_FORMAT_Y212:
+		case DRM_FORMAT_Y216:
+		case DRM_FORMAT_XVYU12_16161616:
+		case DRM_FORMAT_XVYU16161616:
+			drm_dbg_kms(&dev_priv->drm,
+				    "Unsupported pixel format %p4cc for 90/270!\n",
+				    &fb->format->format);
+			return -EINVAL;
+		default:
+			break;
+		}
+	}
+
+	/* Y-tiling is not supported in IF-ID Interlace mode */
+	if (crtc_state->hw.enable &&
+	    crtc_state->hw.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE &&
+	    fb->modifier != DRM_FORMAT_MOD_LINEAR &&
+	    fb->modifier != I915_FORMAT_MOD_X_TILED) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Y/Yf tiling not supported in IF-ID mode\n");
+		return -EINVAL;
+	}
+
+	/* Wa_1606054188:tgl,adl-s */
+	if ((IS_ALDERLAKE_S(dev_priv) || IS_TIGERLAKE(dev_priv)) &&
+	    plane_state->ckey.flags & I915_SET_COLORKEY_SOURCE &&
+	    intel_format_is_p01x(fb->format->format)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Source color keying not supported with P01x formats\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int skl_plane_check_dst_coordinates(const struct intel_crtc_state *crtc_state,
+					   const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(plane_state->uapi.plane->dev);
+	int crtc_x = plane_state->uapi.dst.x1;
+	int crtc_w = drm_rect_width(&plane_state->uapi.dst);
+	int pipe_src_w = drm_rect_width(&crtc_state->pipe_src);
+
+	/*
+	 * Display WA #1175: glk
+	 * Planes other than the cursor may cause FIFO underflow and display
+	 * corruption if starting less than 4 pixels from the right edge of
+	 * the screen.
+	 * Besides the above WA fix the similar problem, where planes other
+	 * than the cursor ending less than 4 pixels from the left edge of the
+	 * screen may cause FIFO underflow and display corruption.
+	 */
+	if (DISPLAY_VER(dev_priv) == 10 &&
+	    (crtc_x + crtc_w < 4 || crtc_x > pipe_src_w - 4)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "requested plane X %s position %d invalid (valid range %d-%d)\n",
+			    crtc_x + crtc_w < 4 ? "end" : "start",
+			    crtc_x + crtc_w < 4 ? crtc_x + crtc_w : crtc_x,
+			    4, pipe_src_w - 4);
+		return -ERANGE;
+	}
+
+	return 0;
+}
+
+static int skl_plane_check_nv12_rotation(const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane_state->uapi.plane->dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+	int src_w = drm_rect_width(&plane_state->uapi.src) >> 16;
+
+	/* Display WA #1106 */
+	if (intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier) &&
+	    src_w & 3 &&
+	    (rotation == DRM_MODE_ROTATE_270 ||
+	     rotation == (DRM_MODE_REFLECT_X | DRM_MODE_ROTATE_90))) {
+		drm_dbg_kms(&i915->drm, "src width must be multiple of 4 for rotated planar YUV\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int skl_plane_max_scale(struct drm_i915_private *dev_priv,
+			       const struct drm_framebuffer *fb)
+{
+	/*
+	 * We don't yet know the final source width nor
+	 * whether we can use the HQ scaler mode. Assume
+	 * the best case.
+	 * FIXME need to properly check this later.
+	 */
+	if (DISPLAY_VER(dev_priv) >= 10 ||
+	    !intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier))
+		return 0x30000 - 1;
+	else
+		return 0x20000 - 1;
+}
+
+static int intel_plane_min_width(struct intel_plane *plane,
+				 const struct drm_framebuffer *fb,
+				 int color_plane,
+				 unsigned int rotation)
+{
+	if (plane->min_width)
+		return plane->min_width(fb, color_plane, rotation);
+	else
+		return 1;
+}
+
+static int intel_plane_max_width(struct intel_plane *plane,
+				 const struct drm_framebuffer *fb,
+				 int color_plane,
+				 unsigned int rotation)
+{
+	if (plane->max_width)
+		return plane->max_width(fb, color_plane, rotation);
+	else
+		return INT_MAX;
+}
+
+static int intel_plane_max_height(struct intel_plane *plane,
+				  const struct drm_framebuffer *fb,
+				  int color_plane,
+				  unsigned int rotation)
+{
+	if (plane->max_height)
+		return plane->max_height(fb, color_plane, rotation);
+	else
+		return INT_MAX;
+}
+
+static bool
+skl_check_main_ccs_coordinates(struct intel_plane_state *plane_state,
+			       int main_x, int main_y, u32 main_offset,
+			       int ccs_plane)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int aux_x = plane_state->view.color_plane[ccs_plane].x;
+	int aux_y = plane_state->view.color_plane[ccs_plane].y;
+	u32 aux_offset = plane_state->view.color_plane[ccs_plane].offset;
+	u32 alignment = intel_surf_alignment(fb, ccs_plane);
+	int hsub;
+	int vsub;
+
+	intel_fb_plane_get_subsampling(&hsub, &vsub, fb, ccs_plane);
+	while (aux_offset >= main_offset && aux_y <= main_y) {
+		int x, y;
+
+		if (aux_x == main_x && aux_y == main_y)
+			break;
+
+		if (aux_offset == 0)
+			break;
+
+		x = aux_x / hsub;
+		y = aux_y / vsub;
+		aux_offset = intel_plane_adjust_aligned_offset(&x, &y,
+							       plane_state,
+							       ccs_plane,
+							       aux_offset,
+							       aux_offset -
+								alignment);
+		aux_x = x * hsub + aux_x % hsub;
+		aux_y = y * vsub + aux_y % vsub;
+	}
+
+	if (aux_x != main_x || aux_y != main_y)
+		return false;
+
+	plane_state->view.color_plane[ccs_plane].offset = aux_offset;
+	plane_state->view.color_plane[ccs_plane].x = aux_x;
+	plane_state->view.color_plane[ccs_plane].y = aux_y;
+
+	return true;
+}
+
+
+int skl_calc_main_surface_offset(const struct intel_plane_state *plane_state,
+				 int *x, int *y, u32 *offset)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	const int aux_plane = skl_main_to_aux_plane(fb, 0);
+	const u32 aux_offset = plane_state->view.color_plane[aux_plane].offset;
+	const u32 alignment = intel_surf_alignment(fb, 0);
+	const int w = drm_rect_width(&plane_state->uapi.src) >> 16;
+
+	intel_add_fb_offsets(x, y, plane_state, 0);
+	*offset = intel_plane_compute_aligned_offset(x, y, plane_state, 0);
+	if (drm_WARN_ON(&dev_priv->drm, alignment && !is_power_of_2(alignment)))
+		return -EINVAL;
+
+	/*
+	 * AUX surface offset is specified as the distance from the
+	 * main surface offset, and it must be non-negative. Make
+	 * sure that is what we will get.
+	 */
+	if (aux_plane && *offset > aux_offset)
+		*offset = intel_plane_adjust_aligned_offset(x, y, plane_state, 0,
+							    *offset,
+							    aux_offset & ~(alignment - 1));
+
+	/*
+	 * When using an X-tiled surface, the plane blows up
+	 * if the x offset + width exceed the stride.
+	 *
+	 * TODO: linear and Y-tiled seem fine, Yf untested,
+	 */
+	if (fb->modifier == I915_FORMAT_MOD_X_TILED) {
+		int cpp = fb->format->cpp[0];
+
+		while ((*x + w) * cpp > plane_state->view.color_plane[0].mapping_stride) {
+			if (*offset == 0) {
+				drm_dbg_kms(&dev_priv->drm,
+					    "Unable to find suitable display surface offset due to X-tiling\n");
+				return -EINVAL;
+			}
+
+			*offset = intel_plane_adjust_aligned_offset(x, y, plane_state, 0,
+								    *offset,
+								    *offset - alignment);
+		}
+	}
+
+	return 0;
+}
+
+static int skl_check_main_surface(struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	const unsigned int rotation = plane_state->hw.rotation;
+	int x = plane_state->uapi.src.x1 >> 16;
+	int y = plane_state->uapi.src.y1 >> 16;
+	const int w = drm_rect_width(&plane_state->uapi.src) >> 16;
+	const int h = drm_rect_height(&plane_state->uapi.src) >> 16;
+	const int min_width = intel_plane_min_width(plane, fb, 0, rotation);
+	const int max_width = intel_plane_max_width(plane, fb, 0, rotation);
+	const int max_height = intel_plane_max_height(plane, fb, 0, rotation);
+	const int aux_plane = skl_main_to_aux_plane(fb, 0);
+	const u32 alignment = intel_surf_alignment(fb, 0);
+	u32 offset;
+	int ret;
+
+	if (w > max_width || w < min_width || h > max_height || h < 1) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "requested Y/RGB source size %dx%d outside limits (min: %dx1 max: %dx%d)\n",
+			    w, h, min_width, max_width, max_height);
+		return -EINVAL;
+	}
+
+	ret = skl_calc_main_surface_offset(plane_state, &x, &y, &offset);
+	if (ret)
+		return ret;
+
+	/*
+	 * CCS AUX surface doesn't have its own x/y offsets, we must make sure
+	 * they match with the main surface x/y offsets. On DG2
+	 * there's no aux plane on fb so skip this checking.
+	 */
+	if (intel_fb_is_ccs_modifier(fb->modifier) && aux_plane) {
+		while (!skl_check_main_ccs_coordinates(plane_state, x, y,
+						       offset, aux_plane)) {
+			if (offset == 0)
+				break;
+
+			offset = intel_plane_adjust_aligned_offset(&x, &y, plane_state, 0,
+								   offset, offset - alignment);
+		}
+
+		if (x != plane_state->view.color_plane[aux_plane].x ||
+		    y != plane_state->view.color_plane[aux_plane].y) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Unable to find suitable display surface offset due to CCS\n");
+			return -EINVAL;
+		}
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 13)
+		drm_WARN_ON(&dev_priv->drm, x > 65535 || y > 65535);
+	else
+		drm_WARN_ON(&dev_priv->drm, x > 8191 || y > 8191);
+
+	plane_state->view.color_plane[0].offset = offset;
+	plane_state->view.color_plane[0].x = x;
+	plane_state->view.color_plane[0].y = y;
+
+	/*
+	 * Put the final coordinates back so that the src
+	 * coordinate checks will see the right values.
+	 */
+	drm_rect_translate_to(&plane_state->uapi.src,
+			      x << 16, y << 16);
+
+	return 0;
+}
+
+static int skl_check_nv12_aux_surface(struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	unsigned int rotation = plane_state->hw.rotation;
+	int uv_plane = 1;
+	int ccs_plane = intel_fb_is_ccs_modifier(fb->modifier) ?
+			skl_main_to_aux_plane(fb, uv_plane) : 0;
+	int max_width = intel_plane_max_width(plane, fb, uv_plane, rotation);
+	int max_height = intel_plane_max_height(plane, fb, uv_plane, rotation);
+	int x = plane_state->uapi.src.x1 >> 17;
+	int y = plane_state->uapi.src.y1 >> 17;
+	int w = drm_rect_width(&plane_state->uapi.src) >> 17;
+	int h = drm_rect_height(&plane_state->uapi.src) >> 17;
+	u32 offset;
+
+	/* FIXME not quite sure how/if these apply to the chroma plane */
+	if (w > max_width || h > max_height) {
+		drm_dbg_kms(&i915->drm,
+			    "CbCr source size %dx%d too big (limit %dx%d)\n",
+			    w, h, max_width, max_height);
+		return -EINVAL;
+	}
+
+	intel_add_fb_offsets(&x, &y, plane_state, uv_plane);
+	offset = intel_plane_compute_aligned_offset(&x, &y,
+						    plane_state, uv_plane);
+
+	if (ccs_plane) {
+		u32 aux_offset = plane_state->view.color_plane[ccs_plane].offset;
+		u32 alignment = intel_surf_alignment(fb, uv_plane);
+
+		if (offset > aux_offset)
+			offset = intel_plane_adjust_aligned_offset(&x, &y,
+								   plane_state,
+								   uv_plane,
+								   offset,
+								   aux_offset & ~(alignment - 1));
+
+		while (!skl_check_main_ccs_coordinates(plane_state, x, y,
+						       offset, ccs_plane)) {
+			if (offset == 0)
+				break;
+
+			offset = intel_plane_adjust_aligned_offset(&x, &y,
+								   plane_state,
+								   uv_plane,
+								   offset, offset - alignment);
+		}
+
+		if (x != plane_state->view.color_plane[ccs_plane].x ||
+		    y != plane_state->view.color_plane[ccs_plane].y) {
+			drm_dbg_kms(&i915->drm,
+				    "Unable to find suitable display surface offset due to CCS\n");
+			return -EINVAL;
+		}
+	}
+
+	if (DISPLAY_VER(i915) >= 13)
+		drm_WARN_ON(&i915->drm, x > 65535 || y > 65535);
+	else
+		drm_WARN_ON(&i915->drm, x > 8191 || y > 8191);
+
+	plane_state->view.color_plane[uv_plane].offset = offset;
+	plane_state->view.color_plane[uv_plane].x = x;
+	plane_state->view.color_plane[uv_plane].y = y;
+
+	return 0;
+}
+
+static int skl_check_ccs_aux_surface(struct intel_plane_state *plane_state)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int src_x = plane_state->uapi.src.x1 >> 16;
+	int src_y = plane_state->uapi.src.y1 >> 16;
+	u32 offset;
+	int ccs_plane;
+
+	for (ccs_plane = 0; ccs_plane < fb->format->num_planes; ccs_plane++) {
+		int main_hsub, main_vsub;
+		int hsub, vsub;
+		int x, y;
+
+		if (!intel_fb_is_ccs_aux_plane(fb, ccs_plane))
+			continue;
+
+		intel_fb_plane_get_subsampling(&main_hsub, &main_vsub, fb,
+					       skl_ccs_to_main_plane(fb, ccs_plane));
+		intel_fb_plane_get_subsampling(&hsub, &vsub, fb, ccs_plane);
+
+		hsub *= main_hsub;
+		vsub *= main_vsub;
+		x = src_x / hsub;
+		y = src_y / vsub;
+
+		intel_add_fb_offsets(&x, &y, plane_state, ccs_plane);
+
+		offset = intel_plane_compute_aligned_offset(&x, &y,
+							    plane_state,
+							    ccs_plane);
+
+		plane_state->view.color_plane[ccs_plane].offset = offset;
+		plane_state->view.color_plane[ccs_plane].x = (x * hsub + src_x % hsub) / main_hsub;
+		plane_state->view.color_plane[ccs_plane].y = (y * vsub + src_y % vsub) / main_vsub;
+	}
+
+	return 0;
+}
+
+static int skl_check_plane_surface(struct intel_plane_state *plane_state)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int ret;
+
+	ret = intel_plane_compute_gtt(plane_state);
+	if (ret)
+		return ret;
+
+	if (!plane_state->uapi.visible)
+		return 0;
+
+	/*
+	 * Handle the AUX surface first since the main surface setup depends on
+	 * it.
+	 */
+	if (intel_fb_is_ccs_modifier(fb->modifier)) {
+		ret = skl_check_ccs_aux_surface(plane_state);
+		if (ret)
+			return ret;
+	}
+
+	if (intel_format_info_is_yuv_semiplanar(fb->format,
+						fb->modifier)) {
+		ret = skl_check_nv12_aux_surface(plane_state);
+		if (ret)
+			return ret;
+	}
+
+	ret = skl_check_main_surface(plane_state);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static bool skl_fb_scalable(const struct drm_framebuffer *fb)
+{
+	if (!fb)
+		return false;
+
+	switch (fb->format->format) {
+	case DRM_FORMAT_C8:
+		return false;
+	case DRM_FORMAT_XRGB16161616F:
+	case DRM_FORMAT_ARGB16161616F:
+	case DRM_FORMAT_XBGR16161616F:
+	case DRM_FORMAT_ABGR16161616F:
+		return DISPLAY_VER(to_i915(fb->dev)) >= 11;
+	default:
+		return true;
+	}
+}
+
+static bool bo_has_valid_encryption(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+
+	return intel_pxp_key_check(&to_gt(i915)->pxp, obj, false) == 0;
+}
+
+static bool pxp_is_borked(struct drm_i915_gem_object *obj)
+{
+	return i915_gem_object_is_protected(obj) && !bo_has_valid_encryption(obj);
+}
+
+static int skl_plane_check(struct intel_crtc_state *crtc_state,
+			   struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int min_scale = DRM_PLANE_NO_SCALING;
+	int max_scale = DRM_PLANE_NO_SCALING;
+	int ret;
+
+	ret = skl_plane_check_fb(crtc_state, plane_state);
+	if (ret)
+		return ret;
+
+	/* use scaler when colorkey is not required */
+	if (!plane_state->ckey.flags && skl_fb_scalable(fb)) {
+		min_scale = 1;
+		max_scale = skl_plane_max_scale(dev_priv, fb);
+	}
+
+	ret = intel_atomic_plane_check_clipping(plane_state, crtc_state,
+						min_scale, max_scale, true);
+	if (ret)
+		return ret;
+
+	ret = skl_check_plane_surface(plane_state);
+	if (ret)
+		return ret;
+
+	if (!plane_state->uapi.visible)
+		return 0;
+
+	ret = skl_plane_check_dst_coordinates(crtc_state, plane_state);
+	if (ret)
+		return ret;
+
+	ret = intel_plane_check_src_coordinates(plane_state);
+	if (ret)
+		return ret;
+
+	ret = skl_plane_check_nv12_rotation(plane_state);
+	if (ret)
+		return ret;
+
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		plane_state->decrypt = bo_has_valid_encryption(intel_fb_obj(fb));
+		plane_state->force_black = pxp_is_borked(intel_fb_obj(fb));
+	}
+
+	/* HW only has 8 bits pixel precision, disable plane if invisible */
+	if (!(plane_state->hw.alpha >> 8))
+		plane_state->uapi.visible = false;
+
+	plane_state->ctl = skl_plane_ctl(crtc_state, plane_state);
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		plane_state->color_ctl = glk_plane_color_ctl(crtc_state,
+							     plane_state);
+
+	if (intel_format_info_is_yuv_semiplanar(fb->format, fb->modifier) &&
+	    icl_is_hdr_plane(dev_priv, plane->id))
+		/* Enable and use MPEG-2 chroma siting */
+		plane_state->cus_ctl = PLANE_CUS_ENABLE |
+			PLANE_CUS_HPHASE_0 |
+			PLANE_CUS_VPHASE_SIGN_NEGATIVE | PLANE_CUS_VPHASE_0_25;
+	else
+		plane_state->cus_ctl = 0;
+
+	return 0;
+}
+
+static enum intel_fbc_id skl_fbc_id_for_pipe(enum pipe pipe)
+{
+	return pipe - PIPE_A + INTEL_FBC_A;
+}
+
+static bool skl_plane_has_fbc(struct drm_i915_private *dev_priv,
+			      enum intel_fbc_id fbc_id, enum plane_id plane_id)
+{
+	if ((RUNTIME_INFO(dev_priv)->fbc_mask & BIT(fbc_id)) == 0)
+		return false;
+
+	return plane_id == PLANE_PRIMARY;
+}
+
+static struct intel_fbc *skl_plane_fbc(struct drm_i915_private *dev_priv,
+				       enum pipe pipe, enum plane_id plane_id)
+{
+	enum intel_fbc_id fbc_id = skl_fbc_id_for_pipe(pipe);
+
+	if (skl_plane_has_fbc(dev_priv, fbc_id, plane_id))
+		return dev_priv->display.fbc[fbc_id];
+	else
+		return NULL;
+}
+
+static bool skl_plane_has_planar(struct drm_i915_private *dev_priv,
+				 enum pipe pipe, enum plane_id plane_id)
+{
+	/* Display WA #0870: skl, bxt */
+	if (IS_SKYLAKE(dev_priv) || IS_BROXTON(dev_priv))
+		return false;
+
+	if (DISPLAY_VER(dev_priv) == 9 && pipe == PIPE_C)
+		return false;
+
+	if (plane_id != PLANE_PRIMARY && plane_id != PLANE_SPRITE0)
+		return false;
+
+	return true;
+}
+
+static const u32 *skl_get_plane_formats(struct drm_i915_private *dev_priv,
+					enum pipe pipe, enum plane_id plane_id,
+					int *num_formats)
+{
+	if (skl_plane_has_planar(dev_priv, pipe, plane_id)) {
+		*num_formats = ARRAY_SIZE(skl_planar_formats);
+		return skl_planar_formats;
+	} else {
+		*num_formats = ARRAY_SIZE(skl_plane_formats);
+		return skl_plane_formats;
+	}
+}
+
+static const u32 *glk_get_plane_formats(struct drm_i915_private *dev_priv,
+					enum pipe pipe, enum plane_id plane_id,
+					int *num_formats)
+{
+	if (skl_plane_has_planar(dev_priv, pipe, plane_id)) {
+		*num_formats = ARRAY_SIZE(glk_planar_formats);
+		return glk_planar_formats;
+	} else {
+		*num_formats = ARRAY_SIZE(skl_plane_formats);
+		return skl_plane_formats;
+	}
+}
+
+static const u32 *icl_get_plane_formats(struct drm_i915_private *dev_priv,
+					enum pipe pipe, enum plane_id plane_id,
+					int *num_formats)
+{
+	if (icl_is_hdr_plane(dev_priv, plane_id)) {
+		*num_formats = ARRAY_SIZE(icl_hdr_plane_formats);
+		return icl_hdr_plane_formats;
+	} else if (icl_is_nv12_y_plane(dev_priv, plane_id)) {
+		*num_formats = ARRAY_SIZE(icl_sdr_y_plane_formats);
+		return icl_sdr_y_plane_formats;
+	} else {
+		*num_formats = ARRAY_SIZE(icl_sdr_uv_plane_formats);
+		return icl_sdr_uv_plane_formats;
+	}
+}
+
+static bool skl_plane_format_mod_supported(struct drm_plane *_plane,
+					   u32 format, u64 modifier)
+{
+	struct intel_plane *plane = to_intel_plane(_plane);
+
+	if (!intel_fb_plane_supports_modifier(plane, modifier))
+		return false;
+
+	switch (format) {
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_ARGB8888:
+	case DRM_FORMAT_ABGR8888:
+		if (intel_fb_is_ccs_modifier(modifier))
+			return true;
+		fallthrough;
+	case DRM_FORMAT_RGB565:
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_XBGR2101010:
+	case DRM_FORMAT_ARGB2101010:
+	case DRM_FORMAT_ABGR2101010:
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_YVYU:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_VYUY:
+	case DRM_FORMAT_NV12:
+	case DRM_FORMAT_XYUV8888:
+	case DRM_FORMAT_P010:
+	case DRM_FORMAT_P012:
+	case DRM_FORMAT_P016:
+	case DRM_FORMAT_XVYU2101010:
+		if (modifier == I915_FORMAT_MOD_Yf_TILED)
+			return true;
+		fallthrough;
+	case DRM_FORMAT_C8:
+	case DRM_FORMAT_XBGR16161616F:
+	case DRM_FORMAT_ABGR16161616F:
+	case DRM_FORMAT_XRGB16161616F:
+	case DRM_FORMAT_ARGB16161616F:
+	case DRM_FORMAT_Y210:
+	case DRM_FORMAT_Y212:
+	case DRM_FORMAT_Y216:
+	case DRM_FORMAT_XVYU12_16161616:
+	case DRM_FORMAT_XVYU16161616:
+		if (modifier == DRM_FORMAT_MOD_LINEAR ||
+		    modifier == I915_FORMAT_MOD_X_TILED ||
+		    modifier == I915_FORMAT_MOD_Y_TILED)
+			return true;
+		fallthrough;
+	default:
+		return false;
+	}
+}
+
+static bool gen12_plane_format_mod_supported(struct drm_plane *_plane,
+					     u32 format, u64 modifier)
+{
+	struct intel_plane *plane = to_intel_plane(_plane);
+
+	if (!intel_fb_plane_supports_modifier(plane, modifier))
+		return false;
+
+	switch (format) {
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_XBGR8888:
+	case DRM_FORMAT_ARGB8888:
+	case DRM_FORMAT_ABGR8888:
+		if (intel_fb_is_ccs_modifier(modifier))
+			return true;
+		fallthrough;
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_YVYU:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_VYUY:
+	case DRM_FORMAT_NV12:
+	case DRM_FORMAT_XYUV8888:
+	case DRM_FORMAT_P010:
+	case DRM_FORMAT_P012:
+	case DRM_FORMAT_P016:
+		if (intel_fb_is_mc_ccs_modifier(modifier))
+			return true;
+		fallthrough;
+	case DRM_FORMAT_RGB565:
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_XBGR2101010:
+	case DRM_FORMAT_ARGB2101010:
+	case DRM_FORMAT_ABGR2101010:
+	case DRM_FORMAT_XVYU2101010:
+	case DRM_FORMAT_C8:
+	case DRM_FORMAT_XBGR16161616F:
+	case DRM_FORMAT_ABGR16161616F:
+	case DRM_FORMAT_XRGB16161616F:
+	case DRM_FORMAT_ARGB16161616F:
+	case DRM_FORMAT_Y210:
+	case DRM_FORMAT_Y212:
+	case DRM_FORMAT_Y216:
+	case DRM_FORMAT_XVYU12_16161616:
+	case DRM_FORMAT_XVYU16161616:
+		if (!intel_fb_is_ccs_modifier(modifier))
+			return true;
+		fallthrough;
+	default:
+		return false;
+	}
+}
+
+static const struct drm_plane_funcs skl_plane_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = intel_plane_destroy,
+	.atomic_duplicate_state = intel_plane_duplicate_state,
+	.atomic_destroy_state = intel_plane_destroy_state,
+	.format_mod_supported = skl_plane_format_mod_supported,
+};
+
+static const struct drm_plane_funcs gen12_plane_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = intel_plane_destroy,
+	.atomic_duplicate_state = intel_plane_duplicate_state,
+	.atomic_destroy_state = intel_plane_destroy_state,
+	.format_mod_supported = gen12_plane_format_mod_supported,
+};
+
+static void
+skl_plane_enable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	bdw_enable_pipe_irq(i915, pipe, GEN9_PIPE_PLANE_FLIP_DONE(plane->id));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static void
+skl_plane_disable_flip_done(struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+
+	spin_lock_irq(&i915->irq_lock);
+	bdw_disable_pipe_irq(i915, pipe, GEN9_PIPE_PLANE_FLIP_DONE(plane->id));
+	spin_unlock_irq(&i915->irq_lock);
+}
+
+static bool skl_plane_has_rc_ccs(struct drm_i915_private *i915,
+				 enum pipe pipe, enum plane_id plane_id)
+{
+	/* Wa_22011186057 */
+	if (IS_ADLP_DISPLAY_STEP(i915, STEP_A0, STEP_B0))
+		return false;
+
+	if (DISPLAY_VER(i915) >= 11)
+		return true;
+
+	if (IS_GEMINILAKE(i915))
+		return pipe != PIPE_C;
+
+	return pipe != PIPE_C &&
+		(plane_id == PLANE_PRIMARY ||
+		 plane_id == PLANE_SPRITE0);
+}
+
+static bool gen12_plane_has_mc_ccs(struct drm_i915_private *i915,
+				   enum plane_id plane_id)
+{
+	if (DISPLAY_VER(i915) < 12)
+		return false;
+
+	/* Wa_14010477008:tgl[a0..c0],rkl[all],dg1[all] */
+	if (IS_DG1(i915) || IS_ROCKETLAKE(i915) ||
+	    IS_TGL_DISPLAY_STEP(i915, STEP_A0, STEP_D0))
+		return false;
+
+	/* Wa_22011186057 */
+	if (IS_ADLP_DISPLAY_STEP(i915, STEP_A0, STEP_B0))
+		return false;
+
+	/* Wa_14013215631 */
+	if (IS_DG2_DISPLAY_STEP(i915, STEP_A0, STEP_C0))
+		return false;
+
+	return plane_id < PLANE_SPRITE4;
+}
+
+static u8 skl_get_plane_caps(struct drm_i915_private *i915,
+			     enum pipe pipe, enum plane_id plane_id)
+{
+	u8 caps = INTEL_PLANE_CAP_TILING_X;
+
+	if (DISPLAY_VER(i915) < 13 || IS_ALDERLAKE_P(i915))
+		caps |= INTEL_PLANE_CAP_TILING_Y;
+	if (DISPLAY_VER(i915) < 12)
+		caps |= INTEL_PLANE_CAP_TILING_Yf;
+	if (HAS_4TILE(i915))
+		caps |= INTEL_PLANE_CAP_TILING_4;
+
+	if (skl_plane_has_rc_ccs(i915, pipe, plane_id)) {
+		caps |= INTEL_PLANE_CAP_CCS_RC;
+		if (DISPLAY_VER(i915) >= 12)
+			caps |= INTEL_PLANE_CAP_CCS_RC_CC;
+	}
+
+	if (gen12_plane_has_mc_ccs(i915, plane_id))
+		caps |= INTEL_PLANE_CAP_CCS_MC;
+
+	return caps;
+}
+
+struct intel_plane *
+skl_universal_plane_create(struct drm_i915_private *dev_priv,
+			   enum pipe pipe, enum plane_id plane_id)
+{
+	const struct drm_plane_funcs *plane_funcs;
+	struct intel_plane *plane;
+	enum drm_plane_type plane_type;
+	unsigned int supported_rotations;
+	unsigned int supported_csc;
+	const u64 *modifiers;
+	const u32 *formats;
+	int num_formats;
+	int ret;
+
+	plane = intel_plane_alloc();
+	if (IS_ERR(plane))
+		return plane;
+
+	plane->pipe = pipe;
+	plane->id = plane_id;
+	plane->frontbuffer_bit = INTEL_FRONTBUFFER(pipe, plane_id);
+
+	intel_fbc_add_plane(skl_plane_fbc(dev_priv, pipe, plane_id), plane);
+
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		plane->min_width = icl_plane_min_width;
+		if (icl_is_hdr_plane(dev_priv, plane_id))
+			plane->max_width = icl_hdr_plane_max_width;
+		else
+			plane->max_width = icl_sdr_plane_max_width;
+		plane->max_height = icl_plane_max_height;
+		plane->min_cdclk = icl_plane_min_cdclk;
+	} else if (DISPLAY_VER(dev_priv) >= 10) {
+		plane->max_width = glk_plane_max_width;
+		plane->max_height = skl_plane_max_height;
+		plane->min_cdclk = glk_plane_min_cdclk;
+	} else {
+		plane->max_width = skl_plane_max_width;
+		plane->max_height = skl_plane_max_height;
+		plane->min_cdclk = skl_plane_min_cdclk;
+	}
+
+	plane->max_stride = skl_plane_max_stride;
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		plane->update_noarm = icl_plane_update_noarm;
+		plane->update_arm = icl_plane_update_arm;
+		plane->disable_arm = icl_plane_disable_arm;
+	} else {
+		plane->update_noarm = skl_plane_update_noarm;
+		plane->update_arm = skl_plane_update_arm;
+		plane->disable_arm = skl_plane_disable_arm;
+	}
+	plane->get_hw_state = skl_plane_get_hw_state;
+	plane->check_plane = skl_plane_check;
+
+	if (plane_id == PLANE_PRIMARY) {
+		plane->need_async_flip_disable_wa = IS_DISPLAY_VER(dev_priv,
+								   9, 10);
+		plane->async_flip = skl_plane_async_flip;
+		plane->enable_flip_done = skl_plane_enable_flip_done;
+		plane->disable_flip_done = skl_plane_disable_flip_done;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		formats = icl_get_plane_formats(dev_priv, pipe,
+						plane_id, &num_formats);
+	else if (DISPLAY_VER(dev_priv) >= 10)
+		formats = glk_get_plane_formats(dev_priv, pipe,
+						plane_id, &num_formats);
+	else
+		formats = skl_get_plane_formats(dev_priv, pipe,
+						plane_id, &num_formats);
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		plane_funcs = &gen12_plane_funcs;
+	else
+		plane_funcs = &skl_plane_funcs;
+
+	if (plane_id == PLANE_PRIMARY)
+		plane_type = DRM_PLANE_TYPE_PRIMARY;
+	else
+		plane_type = DRM_PLANE_TYPE_OVERLAY;
+
+	modifiers = intel_fb_plane_get_modifiers(dev_priv,
+						 skl_get_plane_caps(dev_priv, pipe, plane_id));
+
+	ret = drm_universal_plane_init(&dev_priv->drm, &plane->base,
+				       0, plane_funcs,
+				       formats, num_formats, modifiers,
+				       plane_type,
+				       "plane %d%c", plane_id + 1,
+				       pipe_name(pipe));
+
+	kfree(modifiers);
+
+	if (ret)
+		goto fail;
+
+	if (DISPLAY_VER(dev_priv) >= 13)
+		supported_rotations = DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_180;
+	else
+		supported_rotations =
+			DRM_MODE_ROTATE_0 | DRM_MODE_ROTATE_90 |
+			DRM_MODE_ROTATE_180 | DRM_MODE_ROTATE_270;
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		supported_rotations |= DRM_MODE_REFLECT_X;
+
+	drm_plane_create_rotation_property(&plane->base,
+					   DRM_MODE_ROTATE_0,
+					   supported_rotations);
+
+	supported_csc = BIT(DRM_COLOR_YCBCR_BT601) | BIT(DRM_COLOR_YCBCR_BT709);
+
+	if (DISPLAY_VER(dev_priv) >= 10)
+		supported_csc |= BIT(DRM_COLOR_YCBCR_BT2020);
+
+	drm_plane_create_color_properties(&plane->base,
+					  supported_csc,
+					  BIT(DRM_COLOR_YCBCR_LIMITED_RANGE) |
+					  BIT(DRM_COLOR_YCBCR_FULL_RANGE),
+					  DRM_COLOR_YCBCR_BT709,
+					  DRM_COLOR_YCBCR_LIMITED_RANGE);
+
+	drm_plane_create_alpha_property(&plane->base);
+	drm_plane_create_blend_mode_property(&plane->base,
+					     BIT(DRM_MODE_BLEND_PIXEL_NONE) |
+					     BIT(DRM_MODE_BLEND_PREMULTI) |
+					     BIT(DRM_MODE_BLEND_COVERAGE));
+
+	drm_plane_create_zpos_immutable_property(&plane->base, plane_id);
+
+	if (DISPLAY_VER(dev_priv) >= 12)
+		drm_plane_enable_fb_damage_clips(&plane->base);
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		drm_plane_create_scaling_filter_property(&plane->base,
+						BIT(DRM_SCALING_FILTER_DEFAULT) |
+						BIT(DRM_SCALING_FILTER_NEAREST_NEIGHBOR));
+
+	intel_plane_helper_add(plane);
+
+	return plane;
+
+fail:
+	intel_plane_free(plane);
+
+	return ERR_PTR(ret);
+}
+
+void
+skl_get_initial_plane_config(struct intel_crtc *crtc,
+			     struct intel_initial_plane_config *plane_config)
+{
+	struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state);
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_plane *plane = to_intel_plane(crtc->base.primary);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe;
+	u32 val, base, offset, stride_mult, tiling, alpha;
+	int fourcc, pixel_format;
+	unsigned int aligned_height;
+	struct drm_framebuffer *fb;
+	struct intel_framebuffer *intel_fb;
+	static_assert(PLANE_CTL_TILED_YF == PLANE_CTL_TILED_4);
+
+	if (!plane->get_hw_state(plane, &pipe))
+		return;
+
+	drm_WARN_ON(dev, pipe != crtc->pipe);
+
+	if (crtc_state->bigjoiner_pipes) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Unsupported bigjoiner configuration for initial FB\n");
+		return;
+	}
+
+	intel_fb = kzalloc(sizeof(*intel_fb), GFP_KERNEL);
+	if (!intel_fb) {
+		drm_dbg_kms(&dev_priv->drm, "failed to alloc fb\n");
+		return;
+	}
+
+	fb = &intel_fb->base;
+
+	fb->dev = dev;
+
+	val = intel_de_read(dev_priv, PLANE_CTL(pipe, plane_id));
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		pixel_format = val & PLANE_CTL_FORMAT_MASK_ICL;
+	else
+		pixel_format = val & PLANE_CTL_FORMAT_MASK_SKL;
+
+	if (DISPLAY_VER(dev_priv) >= 10) {
+		u32 color_ctl;
+
+		color_ctl = intel_de_read(dev_priv, PLANE_COLOR_CTL(pipe, plane_id));
+		alpha = REG_FIELD_GET(PLANE_COLOR_ALPHA_MASK, color_ctl);
+	} else {
+		alpha = REG_FIELD_GET(PLANE_CTL_ALPHA_MASK, val);
+	}
+
+	fourcc = skl_format_to_fourcc(pixel_format,
+				      val & PLANE_CTL_ORDER_RGBX, alpha);
+	fb->format = drm_format_info(fourcc);
+
+	tiling = val & PLANE_CTL_TILED_MASK;
+	switch (tiling) {
+	case PLANE_CTL_TILED_LINEAR:
+		fb->modifier = DRM_FORMAT_MOD_LINEAR;
+		break;
+	case PLANE_CTL_TILED_X:
+		plane_config->tiling = I915_TILING_X;
+		fb->modifier = I915_FORMAT_MOD_X_TILED;
+		break;
+	case PLANE_CTL_TILED_Y:
+		plane_config->tiling = I915_TILING_Y;
+		if (val & PLANE_CTL_RENDER_DECOMPRESSION_ENABLE)
+			if (DISPLAY_VER(dev_priv) >= 12)
+				fb->modifier = I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS;
+			else
+				fb->modifier = I915_FORMAT_MOD_Y_TILED_CCS;
+		else if (val & PLANE_CTL_MEDIA_DECOMPRESSION_ENABLE)
+			fb->modifier = I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS;
+		else
+			fb->modifier = I915_FORMAT_MOD_Y_TILED;
+		break;
+	case PLANE_CTL_TILED_YF: /* aka PLANE_CTL_TILED_4 on XE_LPD+ */
+		if (HAS_4TILE(dev_priv)) {
+			u32 rc_mask = PLANE_CTL_RENDER_DECOMPRESSION_ENABLE |
+				      PLANE_CTL_CLEAR_COLOR_DISABLE;
+
+			if ((val & rc_mask) == rc_mask)
+				fb->modifier = I915_FORMAT_MOD_4_TILED_DG2_RC_CCS;
+			else if (val & PLANE_CTL_MEDIA_DECOMPRESSION_ENABLE)
+				fb->modifier = I915_FORMAT_MOD_4_TILED_DG2_MC_CCS;
+			else if (val & PLANE_CTL_RENDER_DECOMPRESSION_ENABLE)
+				fb->modifier = I915_FORMAT_MOD_4_TILED_DG2_RC_CCS_CC;
+			else
+				fb->modifier = I915_FORMAT_MOD_4_TILED;
+		} else {
+			if (val & PLANE_CTL_RENDER_DECOMPRESSION_ENABLE)
+				fb->modifier = I915_FORMAT_MOD_Yf_TILED_CCS;
+			else
+				fb->modifier = I915_FORMAT_MOD_Yf_TILED;
+		}
+		break;
+	default:
+		MISSING_CASE(tiling);
+		goto error;
+	}
+
+	/*
+	 * DRM_MODE_ROTATE_ is counter clockwise to stay compatible with Xrandr
+	 * while i915 HW rotation is clockwise, thats why this swapping.
+	 */
+	switch (val & PLANE_CTL_ROTATE_MASK) {
+	case PLANE_CTL_ROTATE_0:
+		plane_config->rotation = DRM_MODE_ROTATE_0;
+		break;
+	case PLANE_CTL_ROTATE_90:
+		plane_config->rotation = DRM_MODE_ROTATE_270;
+		break;
+	case PLANE_CTL_ROTATE_180:
+		plane_config->rotation = DRM_MODE_ROTATE_180;
+		break;
+	case PLANE_CTL_ROTATE_270:
+		plane_config->rotation = DRM_MODE_ROTATE_90;
+		break;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 11 && val & PLANE_CTL_FLIP_HORIZONTAL)
+		plane_config->rotation |= DRM_MODE_REFLECT_X;
+
+	/* 90/270 degree rotation would require extra work */
+	if (drm_rotation_90_or_270(plane_config->rotation))
+		goto error;
+
+	base = intel_de_read(dev_priv, PLANE_SURF(pipe, plane_id)) & PLANE_SURF_ADDR_MASK;
+	plane_config->base = base;
+
+	offset = intel_de_read(dev_priv, PLANE_OFFSET(pipe, plane_id));
+
+	val = intel_de_read(dev_priv, PLANE_SIZE(pipe, plane_id));
+	fb->height = REG_FIELD_GET(PLANE_HEIGHT_MASK, val) + 1;
+	fb->width = REG_FIELD_GET(PLANE_WIDTH_MASK, val) + 1;
+
+	val = intel_de_read(dev_priv, PLANE_STRIDE(pipe, plane_id));
+	stride_mult = skl_plane_stride_mult(fb, 0, DRM_MODE_ROTATE_0);
+
+	fb->pitches[0] = REG_FIELD_GET(PLANE_STRIDE__MASK, val) * stride_mult;
+
+	aligned_height = intel_fb_align_height(fb, 0, fb->height);
+
+	plane_config->size = fb->pitches[0] * aligned_height;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "%s/%s with fb: size=%dx%d@%d, offset=%x, pitch %d, size 0x%x\n",
+		    crtc->base.name, plane->base.name, fb->width, fb->height,
+		    fb->format->cpp[0] * 8, base, fb->pitches[0],
+		    plane_config->size);
+
+	plane_config->fb = intel_fb;
+	return;
+
+error:
+	kfree(intel_fb);
+}
diff --git a/drivers/gpu/drm/i915/display/skl_universal_plane.h b/drivers/gpu/drm/i915/display/skl_universal_plane.h
new file mode 100644
index 000000000..351040b64
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/skl_universal_plane.h
@@ -0,0 +1,35 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef _SKL_UNIVERSAL_PLANE_H_
+#define _SKL_UNIVERSAL_PLANE_H_
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct intel_crtc;
+struct intel_initial_plane_config;
+struct intel_plane_state;
+
+enum pipe;
+enum plane_id;
+
+struct intel_plane *
+skl_universal_plane_create(struct drm_i915_private *dev_priv,
+			   enum pipe pipe, enum plane_id plane_id);
+
+void skl_get_initial_plane_config(struct intel_crtc *crtc,
+				  struct intel_initial_plane_config *plane_config);
+
+int skl_format_to_fourcc(int format, bool rgb_order, bool alpha);
+
+int skl_calc_main_surface_offset(const struct intel_plane_state *plane_state,
+				 int *x, int *y, u32 *offset);
+
+bool icl_is_nv12_y_plane(struct drm_i915_private *dev_priv,
+			 enum plane_id plane_id);
+bool icl_is_hdr_plane(struct drm_i915_private *dev_priv, enum plane_id plane_id);
+
+#endif
diff --git a/drivers/gpu/drm/i915/display/skl_watermark.c b/drivers/gpu/drm/i915/display/skl_watermark.c
new file mode 100644
index 000000000..a7adf0247
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/skl_watermark.c
@@ -0,0 +1,3575 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include <drm/drm_blend.h>
+
+#include "intel_atomic.h"
+#include "intel_atomic_plane.h"
+#include "intel_bw.h"
+#include "intel_de.h"
+#include "intel_display.h"
+#include "intel_display_power.h"
+#include "intel_display_types.h"
+#include "intel_fb.h"
+#include "skl_watermark.h"
+
+#include "i915_drv.h"
+#include "i915_fixed.h"
+#include "i915_reg.h"
+#include "intel_pcode.h"
+#include "intel_pm.h"
+
+static void skl_sagv_disable(struct drm_i915_private *i915);
+
+/* Stores plane specific WM parameters */
+struct skl_wm_params {
+	bool x_tiled, y_tiled;
+	bool rc_surface;
+	bool is_planar;
+	u32 width;
+	u8 cpp;
+	u32 plane_pixel_rate;
+	u32 y_min_scanlines;
+	u32 plane_bytes_per_line;
+	uint_fixed_16_16_t plane_blocks_per_line;
+	uint_fixed_16_16_t y_tile_minimum;
+	u32 linetime_us;
+	u32 dbuf_block_size;
+};
+
+u8 intel_enabled_dbuf_slices_mask(struct drm_i915_private *i915)
+{
+	u8 enabled_slices = 0;
+	enum dbuf_slice slice;
+
+	for_each_dbuf_slice(i915, slice) {
+		if (intel_uncore_read(&i915->uncore,
+				      DBUF_CTL_S(slice)) & DBUF_POWER_STATE)
+			enabled_slices |= BIT(slice);
+	}
+
+	return enabled_slices;
+}
+
+/*
+ * FIXME: We still don't have the proper code detect if we need to apply the WA,
+ * so assume we'll always need it in order to avoid underruns.
+ */
+static bool skl_needs_memory_bw_wa(struct drm_i915_private *i915)
+{
+	return DISPLAY_VER(i915) == 9;
+}
+
+static bool
+intel_has_sagv(struct drm_i915_private *i915)
+{
+	return DISPLAY_VER(i915) >= 9 && !IS_LP(i915) &&
+		i915->display.sagv.status != I915_SAGV_NOT_CONTROLLED;
+}
+
+static u32
+intel_sagv_block_time(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 14) {
+		u32 val;
+
+		val = intel_uncore_read(&i915->uncore, MTL_LATENCY_SAGV);
+
+		return REG_FIELD_GET(MTL_LATENCY_QCLK_SAGV, val);
+	} else if (DISPLAY_VER(i915) >= 12) {
+		u32 val = 0;
+		int ret;
+
+		ret = snb_pcode_read(&i915->uncore,
+				     GEN12_PCODE_READ_SAGV_BLOCK_TIME_US,
+				     &val, NULL);
+		if (ret) {
+			drm_dbg_kms(&i915->drm, "Couldn't read SAGV block time!\n");
+			return 0;
+		}
+
+		return val;
+	} else if (DISPLAY_VER(i915) == 11) {
+		return 10;
+	} else if (DISPLAY_VER(i915) == 9 && !IS_LP(i915)) {
+		return 30;
+	} else {
+		return 0;
+	}
+}
+
+static void intel_sagv_init(struct drm_i915_private *i915)
+{
+	if (!intel_has_sagv(i915))
+		i915->display.sagv.status = I915_SAGV_NOT_CONTROLLED;
+
+	/*
+	 * Probe to see if we have working SAGV control.
+	 * For icl+ this was already determined by intel_bw_init_hw().
+	 */
+	if (DISPLAY_VER(i915) < 11)
+		skl_sagv_disable(i915);
+
+	drm_WARN_ON(&i915->drm, i915->display.sagv.status == I915_SAGV_UNKNOWN);
+
+	i915->display.sagv.block_time_us = intel_sagv_block_time(i915);
+
+	drm_dbg_kms(&i915->drm, "SAGV supported: %s, original SAGV block time: %u us\n",
+		    str_yes_no(intel_has_sagv(i915)), i915->display.sagv.block_time_us);
+
+	/* avoid overflow when adding with wm0 latency/etc. */
+	if (drm_WARN(&i915->drm, i915->display.sagv.block_time_us > U16_MAX,
+		     "Excessive SAGV block time %u, ignoring\n",
+		     i915->display.sagv.block_time_us))
+		i915->display.sagv.block_time_us = 0;
+
+	if (!intel_has_sagv(i915))
+		i915->display.sagv.block_time_us = 0;
+}
+
+/*
+ * SAGV dynamically adjusts the system agent voltage and clock frequencies
+ * depending on power and performance requirements. The display engine access
+ * to system memory is blocked during the adjustment time. Because of the
+ * blocking time, having this enabled can cause full system hangs and/or pipe
+ * underruns if we don't meet all of the following requirements:
+ *
+ *  - <= 1 pipe enabled
+ *  - All planes can enable watermarks for latencies >= SAGV engine block time
+ *  - We're not using an interlaced display configuration
+ */
+static void skl_sagv_enable(struct drm_i915_private *i915)
+{
+	int ret;
+
+	if (!intel_has_sagv(i915))
+		return;
+
+	if (i915->display.sagv.status == I915_SAGV_ENABLED)
+		return;
+
+	drm_dbg_kms(&i915->drm, "Enabling SAGV\n");
+	ret = snb_pcode_write(&i915->uncore, GEN9_PCODE_SAGV_CONTROL,
+			      GEN9_SAGV_ENABLE);
+
+	/* We don't need to wait for SAGV when enabling */
+
+	/*
+	 * Some skl systems, pre-release machines in particular,
+	 * don't actually have SAGV.
+	 */
+	if (IS_SKYLAKE(i915) && ret == -ENXIO) {
+		drm_dbg(&i915->drm, "No SAGV found on system, ignoring\n");
+		i915->display.sagv.status = I915_SAGV_NOT_CONTROLLED;
+		return;
+	} else if (ret < 0) {
+		drm_err(&i915->drm, "Failed to enable SAGV\n");
+		return;
+	}
+
+	i915->display.sagv.status = I915_SAGV_ENABLED;
+}
+
+static void skl_sagv_disable(struct drm_i915_private *i915)
+{
+	int ret;
+
+	if (!intel_has_sagv(i915))
+		return;
+
+	if (i915->display.sagv.status == I915_SAGV_DISABLED)
+		return;
+
+	drm_dbg_kms(&i915->drm, "Disabling SAGV\n");
+	/* bspec says to keep retrying for at least 1 ms */
+	ret = skl_pcode_request(&i915->uncore, GEN9_PCODE_SAGV_CONTROL,
+				GEN9_SAGV_DISABLE,
+				GEN9_SAGV_IS_DISABLED, GEN9_SAGV_IS_DISABLED,
+				1);
+	/*
+	 * Some skl systems, pre-release machines in particular,
+	 * don't actually have SAGV.
+	 */
+	if (IS_SKYLAKE(i915) && ret == -ENXIO) {
+		drm_dbg(&i915->drm, "No SAGV found on system, ignoring\n");
+		i915->display.sagv.status = I915_SAGV_NOT_CONTROLLED;
+		return;
+	} else if (ret < 0) {
+		drm_err(&i915->drm, "Failed to disable SAGV (%d)\n", ret);
+		return;
+	}
+
+	i915->display.sagv.status = I915_SAGV_DISABLED;
+}
+
+static void skl_sagv_pre_plane_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_bw_state *new_bw_state =
+		intel_atomic_get_new_bw_state(state);
+
+	if (!new_bw_state)
+		return;
+
+	if (!intel_can_enable_sagv(i915, new_bw_state))
+		skl_sagv_disable(i915);
+}
+
+static void skl_sagv_post_plane_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_bw_state *new_bw_state =
+		intel_atomic_get_new_bw_state(state);
+
+	if (!new_bw_state)
+		return;
+
+	if (intel_can_enable_sagv(i915, new_bw_state))
+		skl_sagv_enable(i915);
+}
+
+static void icl_sagv_pre_plane_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_bw_state *old_bw_state =
+		intel_atomic_get_old_bw_state(state);
+	const struct intel_bw_state *new_bw_state =
+		intel_atomic_get_new_bw_state(state);
+	u16 old_mask, new_mask;
+
+	if (!new_bw_state)
+		return;
+
+	old_mask = old_bw_state->qgv_points_mask;
+	new_mask = old_bw_state->qgv_points_mask | new_bw_state->qgv_points_mask;
+
+	if (old_mask == new_mask)
+		return;
+
+	WARN_ON(!new_bw_state->base.changed);
+
+	drm_dbg_kms(&i915->drm, "Restricting QGV points: 0x%x -> 0x%x\n",
+		    old_mask, new_mask);
+
+	/*
+	 * Restrict required qgv points before updating the configuration.
+	 * According to BSpec we can't mask and unmask qgv points at the same
+	 * time. Also masking should be done before updating the configuration
+	 * and unmasking afterwards.
+	 */
+	icl_pcode_restrict_qgv_points(i915, new_mask);
+}
+
+static void icl_sagv_post_plane_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_bw_state *old_bw_state =
+		intel_atomic_get_old_bw_state(state);
+	const struct intel_bw_state *new_bw_state =
+		intel_atomic_get_new_bw_state(state);
+	u16 old_mask, new_mask;
+
+	if (!new_bw_state)
+		return;
+
+	old_mask = old_bw_state->qgv_points_mask | new_bw_state->qgv_points_mask;
+	new_mask = new_bw_state->qgv_points_mask;
+
+	if (old_mask == new_mask)
+		return;
+
+	WARN_ON(!new_bw_state->base.changed);
+
+	drm_dbg_kms(&i915->drm, "Relaxing QGV points: 0x%x -> 0x%x\n",
+		    old_mask, new_mask);
+
+	/*
+	 * Allow required qgv points after updating the configuration.
+	 * According to BSpec we can't mask and unmask qgv points at the same
+	 * time. Also masking should be done before updating the configuration
+	 * and unmasking afterwards.
+	 */
+	icl_pcode_restrict_qgv_points(i915, new_mask);
+}
+
+void intel_sagv_pre_plane_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+
+	/*
+	 * Just return if we can't control SAGV or don't have it.
+	 * This is different from situation when we have SAGV but just can't
+	 * afford it due to DBuf limitation - in case if SAGV is completely
+	 * disabled in a BIOS, we are not even allowed to send a PCode request,
+	 * as it will throw an error. So have to check it here.
+	 */
+	if (!intel_has_sagv(i915))
+		return;
+
+	if (DISPLAY_VER(i915) >= 11)
+		icl_sagv_pre_plane_update(state);
+	else
+		skl_sagv_pre_plane_update(state);
+}
+
+void intel_sagv_post_plane_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+
+	/*
+	 * Just return if we can't control SAGV or don't have it.
+	 * This is different from situation when we have SAGV but just can't
+	 * afford it due to DBuf limitation - in case if SAGV is completely
+	 * disabled in a BIOS, we are not even allowed to send a PCode request,
+	 * as it will throw an error. So have to check it here.
+	 */
+	if (!intel_has_sagv(i915))
+		return;
+
+	if (DISPLAY_VER(i915) >= 11)
+		icl_sagv_post_plane_update(state);
+	else
+		skl_sagv_post_plane_update(state);
+}
+
+static bool skl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	enum plane_id plane_id;
+	int max_level = INT_MAX;
+
+	if (!intel_has_sagv(i915))
+		return false;
+
+	if (!crtc_state->hw.active)
+		return true;
+
+	if (crtc_state->hw.pipe_mode.flags & DRM_MODE_FLAG_INTERLACE)
+		return false;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		const struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+		int level;
+
+		/* Skip this plane if it's not enabled */
+		if (!wm->wm[0].enable)
+			continue;
+
+		/* Find the highest enabled wm level for this plane */
+		for (level = ilk_wm_max_level(i915);
+		     !wm->wm[level].enable; --level)
+		     { }
+
+		/* Highest common enabled wm level for all planes */
+		max_level = min(level, max_level);
+	}
+
+	/* No enabled planes? */
+	if (max_level == INT_MAX)
+		return true;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		const struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+
+		/*
+		 * All enabled planes must have enabled a common wm level that
+		 * can tolerate memory latencies higher than sagv_block_time_us
+		 */
+		if (wm->wm[0].enable && !wm->wm[max_level].can_sagv)
+			return false;
+	}
+
+	return true;
+}
+
+static bool tgl_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	enum plane_id plane_id;
+
+	if (!crtc_state->hw.active)
+		return true;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		const struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+
+		if (wm->wm[0].enable && !wm->sagv.wm0.enable)
+			return false;
+	}
+
+	return true;
+}
+
+static bool intel_crtc_can_enable_sagv(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+
+	if (DISPLAY_VER(i915) >= 12)
+		return tgl_crtc_can_enable_sagv(crtc_state);
+	else
+		return skl_crtc_can_enable_sagv(crtc_state);
+}
+
+bool intel_can_enable_sagv(struct drm_i915_private *i915,
+			   const struct intel_bw_state *bw_state)
+{
+	if (DISPLAY_VER(i915) < 11 &&
+	    bw_state->active_pipes && !is_power_of_2(bw_state->active_pipes))
+		return false;
+
+	return bw_state->pipe_sagv_reject == 0;
+}
+
+static int intel_compute_sagv_mask(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	int ret;
+	struct intel_crtc *crtc;
+	struct intel_crtc_state *new_crtc_state;
+	struct intel_bw_state *new_bw_state = NULL;
+	const struct intel_bw_state *old_bw_state = NULL;
+	int i;
+
+	for_each_new_intel_crtc_in_state(state, crtc,
+					 new_crtc_state, i) {
+		new_bw_state = intel_atomic_get_bw_state(state);
+		if (IS_ERR(new_bw_state))
+			return PTR_ERR(new_bw_state);
+
+		old_bw_state = intel_atomic_get_old_bw_state(state);
+
+		if (intel_crtc_can_enable_sagv(new_crtc_state))
+			new_bw_state->pipe_sagv_reject &= ~BIT(crtc->pipe);
+		else
+			new_bw_state->pipe_sagv_reject |= BIT(crtc->pipe);
+	}
+
+	if (!new_bw_state)
+		return 0;
+
+	new_bw_state->active_pipes =
+		intel_calc_active_pipes(state, old_bw_state->active_pipes);
+
+	if (new_bw_state->active_pipes != old_bw_state->active_pipes) {
+		ret = intel_atomic_lock_global_state(&new_bw_state->base);
+		if (ret)
+			return ret;
+	}
+
+	if (intel_can_enable_sagv(i915, new_bw_state) !=
+	    intel_can_enable_sagv(i915, old_bw_state)) {
+		ret = intel_atomic_serialize_global_state(&new_bw_state->base);
+		if (ret)
+			return ret;
+	} else if (new_bw_state->pipe_sagv_reject != old_bw_state->pipe_sagv_reject) {
+		ret = intel_atomic_lock_global_state(&new_bw_state->base);
+		if (ret)
+			return ret;
+	}
+
+	for_each_new_intel_crtc_in_state(state, crtc,
+					 new_crtc_state, i) {
+		struct skl_pipe_wm *pipe_wm = &new_crtc_state->wm.skl.optimal;
+
+		/*
+		 * We store use_sagv_wm in the crtc state rather than relying on
+		 * that bw state since we have no convenient way to get at the
+		 * latter from the plane commit hooks (especially in the legacy
+		 * cursor case)
+		 */
+		pipe_wm->use_sagv_wm = !HAS_HW_SAGV_WM(i915) &&
+			DISPLAY_VER(i915) >= 12 &&
+			intel_can_enable_sagv(i915, new_bw_state);
+	}
+
+	return 0;
+}
+
+static u16 skl_ddb_entry_init(struct skl_ddb_entry *entry,
+			      u16 start, u16 end)
+{
+	entry->start = start;
+	entry->end = end;
+
+	return end;
+}
+
+static int intel_dbuf_slice_size(struct drm_i915_private *i915)
+{
+	return INTEL_INFO(i915)->display.dbuf.size /
+		hweight8(INTEL_INFO(i915)->display.dbuf.slice_mask);
+}
+
+static void
+skl_ddb_entry_for_slices(struct drm_i915_private *i915, u8 slice_mask,
+			 struct skl_ddb_entry *ddb)
+{
+	int slice_size = intel_dbuf_slice_size(i915);
+
+	if (!slice_mask) {
+		ddb->start = 0;
+		ddb->end = 0;
+		return;
+	}
+
+	ddb->start = (ffs(slice_mask) - 1) * slice_size;
+	ddb->end = fls(slice_mask) * slice_size;
+
+	WARN_ON(ddb->start >= ddb->end);
+	WARN_ON(ddb->end > INTEL_INFO(i915)->display.dbuf.size);
+}
+
+static unsigned int mbus_ddb_offset(struct drm_i915_private *i915, u8 slice_mask)
+{
+	struct skl_ddb_entry ddb;
+
+	if (slice_mask & (BIT(DBUF_S1) | BIT(DBUF_S2)))
+		slice_mask = BIT(DBUF_S1);
+	else if (slice_mask & (BIT(DBUF_S3) | BIT(DBUF_S4)))
+		slice_mask = BIT(DBUF_S3);
+
+	skl_ddb_entry_for_slices(i915, slice_mask, &ddb);
+
+	return ddb.start;
+}
+
+u32 skl_ddb_dbuf_slice_mask(struct drm_i915_private *i915,
+			    const struct skl_ddb_entry *entry)
+{
+	int slice_size = intel_dbuf_slice_size(i915);
+	enum dbuf_slice start_slice, end_slice;
+	u8 slice_mask = 0;
+
+	if (!skl_ddb_entry_size(entry))
+		return 0;
+
+	start_slice = entry->start / slice_size;
+	end_slice = (entry->end - 1) / slice_size;
+
+	/*
+	 * Per plane DDB entry can in a really worst case be on multiple slices
+	 * but single entry is anyway contigious.
+	 */
+	while (start_slice <= end_slice) {
+		slice_mask |= BIT(start_slice);
+		start_slice++;
+	}
+
+	return slice_mask;
+}
+
+static unsigned int intel_crtc_ddb_weight(const struct intel_crtc_state *crtc_state)
+{
+	const struct drm_display_mode *pipe_mode = &crtc_state->hw.pipe_mode;
+	int hdisplay, vdisplay;
+
+	if (!crtc_state->hw.active)
+		return 0;
+
+	/*
+	 * Watermark/ddb requirement highly depends upon width of the
+	 * framebuffer, So instead of allocating DDB equally among pipes
+	 * distribute DDB based on resolution/width of the display.
+	 */
+	drm_mode_get_hv_timing(pipe_mode, &hdisplay, &vdisplay);
+
+	return hdisplay;
+}
+
+static void intel_crtc_dbuf_weights(const struct intel_dbuf_state *dbuf_state,
+				    enum pipe for_pipe,
+				    unsigned int *weight_start,
+				    unsigned int *weight_end,
+				    unsigned int *weight_total)
+{
+	struct drm_i915_private *i915 =
+		to_i915(dbuf_state->base.state->base.dev);
+	enum pipe pipe;
+
+	*weight_start = 0;
+	*weight_end = 0;
+	*weight_total = 0;
+
+	for_each_pipe(i915, pipe) {
+		int weight = dbuf_state->weight[pipe];
+
+		/*
+		 * Do not account pipes using other slice sets
+		 * luckily as of current BSpec slice sets do not partially
+		 * intersect(pipes share either same one slice or same slice set
+		 * i.e no partial intersection), so it is enough to check for
+		 * equality for now.
+		 */
+		if (dbuf_state->slices[pipe] != dbuf_state->slices[for_pipe])
+			continue;
+
+		*weight_total += weight;
+		if (pipe < for_pipe) {
+			*weight_start += weight;
+			*weight_end += weight;
+		} else if (pipe == for_pipe) {
+			*weight_end += weight;
+		}
+	}
+}
+
+static int
+skl_crtc_allocate_ddb(struct intel_atomic_state *state, struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	unsigned int weight_total, weight_start, weight_end;
+	const struct intel_dbuf_state *old_dbuf_state =
+		intel_atomic_get_old_dbuf_state(state);
+	struct intel_dbuf_state *new_dbuf_state =
+		intel_atomic_get_new_dbuf_state(state);
+	struct intel_crtc_state *crtc_state;
+	struct skl_ddb_entry ddb_slices;
+	enum pipe pipe = crtc->pipe;
+	unsigned int mbus_offset = 0;
+	u32 ddb_range_size;
+	u32 dbuf_slice_mask;
+	u32 start, end;
+	int ret;
+
+	if (new_dbuf_state->weight[pipe] == 0) {
+		skl_ddb_entry_init(&new_dbuf_state->ddb[pipe], 0, 0);
+		goto out;
+	}
+
+	dbuf_slice_mask = new_dbuf_state->slices[pipe];
+
+	skl_ddb_entry_for_slices(i915, dbuf_slice_mask, &ddb_slices);
+	mbus_offset = mbus_ddb_offset(i915, dbuf_slice_mask);
+	ddb_range_size = skl_ddb_entry_size(&ddb_slices);
+
+	intel_crtc_dbuf_weights(new_dbuf_state, pipe,
+				&weight_start, &weight_end, &weight_total);
+
+	start = ddb_range_size * weight_start / weight_total;
+	end = ddb_range_size * weight_end / weight_total;
+
+	skl_ddb_entry_init(&new_dbuf_state->ddb[pipe],
+			   ddb_slices.start - mbus_offset + start,
+			   ddb_slices.start - mbus_offset + end);
+
+out:
+	if (old_dbuf_state->slices[pipe] == new_dbuf_state->slices[pipe] &&
+	    skl_ddb_entry_equal(&old_dbuf_state->ddb[pipe],
+				&new_dbuf_state->ddb[pipe]))
+		return 0;
+
+	ret = intel_atomic_lock_global_state(&new_dbuf_state->base);
+	if (ret)
+		return ret;
+
+	crtc_state = intel_atomic_get_crtc_state(&state->base, crtc);
+	if (IS_ERR(crtc_state))
+		return PTR_ERR(crtc_state);
+
+	/*
+	 * Used for checking overlaps, so we need absolute
+	 * offsets instead of MBUS relative offsets.
+	 */
+	crtc_state->wm.skl.ddb.start = mbus_offset + new_dbuf_state->ddb[pipe].start;
+	crtc_state->wm.skl.ddb.end = mbus_offset + new_dbuf_state->ddb[pipe].end;
+
+	drm_dbg_kms(&i915->drm,
+		    "[CRTC:%d:%s] dbuf slices 0x%x -> 0x%x, ddb (%d - %d) -> (%d - %d), active pipes 0x%x -> 0x%x\n",
+		    crtc->base.base.id, crtc->base.name,
+		    old_dbuf_state->slices[pipe], new_dbuf_state->slices[pipe],
+		    old_dbuf_state->ddb[pipe].start, old_dbuf_state->ddb[pipe].end,
+		    new_dbuf_state->ddb[pipe].start, new_dbuf_state->ddb[pipe].end,
+		    old_dbuf_state->active_pipes, new_dbuf_state->active_pipes);
+
+	return 0;
+}
+
+static int skl_compute_wm_params(const struct intel_crtc_state *crtc_state,
+				 int width, const struct drm_format_info *format,
+				 u64 modifier, unsigned int rotation,
+				 u32 plane_pixel_rate, struct skl_wm_params *wp,
+				 int color_plane);
+
+static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state,
+				 struct intel_plane *plane,
+				 int level,
+				 unsigned int latency,
+				 const struct skl_wm_params *wp,
+				 const struct skl_wm_level *result_prev,
+				 struct skl_wm_level *result /* out */);
+
+static unsigned int
+skl_cursor_allocation(const struct intel_crtc_state *crtc_state,
+		      int num_active)
+{
+	struct intel_plane *plane = to_intel_plane(crtc_state->uapi.crtc->cursor);
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	int level, max_level = ilk_wm_max_level(i915);
+	struct skl_wm_level wm = {};
+	int ret, min_ddb_alloc = 0;
+	struct skl_wm_params wp;
+
+	ret = skl_compute_wm_params(crtc_state, 256,
+				    drm_format_info(DRM_FORMAT_ARGB8888),
+				    DRM_FORMAT_MOD_LINEAR,
+				    DRM_MODE_ROTATE_0,
+				    crtc_state->pixel_rate, &wp, 0);
+	drm_WARN_ON(&i915->drm, ret);
+
+	for (level = 0; level <= max_level; level++) {
+		unsigned int latency = i915->display.wm.skl_latency[level];
+
+		skl_compute_plane_wm(crtc_state, plane, level, latency, &wp, &wm, &wm);
+		if (wm.min_ddb_alloc == U16_MAX)
+			break;
+
+		min_ddb_alloc = wm.min_ddb_alloc;
+	}
+
+	return max(num_active == 1 ? 32 : 8, min_ddb_alloc);
+}
+
+static void skl_ddb_entry_init_from_hw(struct skl_ddb_entry *entry, u32 reg)
+{
+	skl_ddb_entry_init(entry,
+			   REG_FIELD_GET(PLANE_BUF_START_MASK, reg),
+			   REG_FIELD_GET(PLANE_BUF_END_MASK, reg));
+	if (entry->end)
+		entry->end++;
+}
+
+static void
+skl_ddb_get_hw_plane_state(struct drm_i915_private *i915,
+			   const enum pipe pipe,
+			   const enum plane_id plane_id,
+			   struct skl_ddb_entry *ddb,
+			   struct skl_ddb_entry *ddb_y)
+{
+	u32 val;
+
+	/* Cursor doesn't support NV12/planar, so no extra calculation needed */
+	if (plane_id == PLANE_CURSOR) {
+		val = intel_uncore_read(&i915->uncore, CUR_BUF_CFG(pipe));
+		skl_ddb_entry_init_from_hw(ddb, val);
+		return;
+	}
+
+	val = intel_uncore_read(&i915->uncore, PLANE_BUF_CFG(pipe, plane_id));
+	skl_ddb_entry_init_from_hw(ddb, val);
+
+	if (DISPLAY_VER(i915) >= 11)
+		return;
+
+	val = intel_uncore_read(&i915->uncore, PLANE_NV12_BUF_CFG(pipe, plane_id));
+	skl_ddb_entry_init_from_hw(ddb_y, val);
+}
+
+static void skl_pipe_ddb_get_hw_state(struct intel_crtc *crtc,
+				      struct skl_ddb_entry *ddb,
+				      struct skl_ddb_entry *ddb_y)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	enum intel_display_power_domain power_domain;
+	enum pipe pipe = crtc->pipe;
+	intel_wakeref_t wakeref;
+	enum plane_id plane_id;
+
+	power_domain = POWER_DOMAIN_PIPE(pipe);
+	wakeref = intel_display_power_get_if_enabled(i915, power_domain);
+	if (!wakeref)
+		return;
+
+	for_each_plane_id_on_crtc(crtc, plane_id)
+		skl_ddb_get_hw_plane_state(i915, pipe,
+					   plane_id,
+					   &ddb[plane_id],
+					   &ddb_y[plane_id]);
+
+	intel_display_power_put(i915, power_domain, wakeref);
+}
+
+struct dbuf_slice_conf_entry {
+	u8 active_pipes;
+	u8 dbuf_mask[I915_MAX_PIPES];
+	bool join_mbus;
+};
+
+/*
+ * Table taken from Bspec 12716
+ * Pipes do have some preferred DBuf slice affinity,
+ * plus there are some hardcoded requirements on how
+ * those should be distributed for multipipe scenarios.
+ * For more DBuf slices algorithm can get even more messy
+ * and less readable, so decided to use a table almost
+ * as is from BSpec itself - that way it is at least easier
+ * to compare, change and check.
+ */
+static const struct dbuf_slice_conf_entry icl_allowed_dbufs[] =
+/* Autogenerated with igt/tools/intel_dbuf_map tool: */
+{
+	{
+		.active_pipes = BIT(PIPE_A),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_B] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_C] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_C] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1),
+			[PIPE_C] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_B] = BIT(DBUF_S1),
+			[PIPE_C] = BIT(DBUF_S2),
+		},
+	},
+	{}
+};
+
+/*
+ * Table taken from Bspec 49255
+ * Pipes do have some preferred DBuf slice affinity,
+ * plus there are some hardcoded requirements on how
+ * those should be distributed for multipipe scenarios.
+ * For more DBuf slices algorithm can get even more messy
+ * and less readable, so decided to use a table almost
+ * as is from BSpec itself - that way it is at least easier
+ * to compare, change and check.
+ */
+static const struct dbuf_slice_conf_entry tgl_allowed_dbufs[] =
+/* Autogenerated with igt/tools/intel_dbuf_map tool: */
+{
+	{
+		.active_pipes = BIT(PIPE_A),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S2),
+			[PIPE_B] = BIT(DBUF_S1),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_C] = BIT(DBUF_S2) | BIT(DBUF_S1),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_C] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1),
+			[PIPE_C] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_B] = BIT(DBUF_S1),
+			[PIPE_C] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_D] = BIT(DBUF_S2) | BIT(DBUF_S1),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_D] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1),
+			[PIPE_D] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_B] = BIT(DBUF_S1),
+			[PIPE_D] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_C] = BIT(DBUF_S1),
+			[PIPE_D] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_C] = BIT(DBUF_S2),
+			[PIPE_D] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1),
+			[PIPE_C] = BIT(DBUF_S2),
+			[PIPE_D] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_B] = BIT(DBUF_S1),
+			[PIPE_C] = BIT(DBUF_S2),
+			[PIPE_D] = BIT(DBUF_S2),
+		},
+	},
+	{}
+};
+
+static const struct dbuf_slice_conf_entry dg2_allowed_dbufs[] = {
+	{
+		.active_pipes = BIT(PIPE_A),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_B] = BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_B] = BIT(DBUF_S2),
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_D] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_D] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_D] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_B] = BIT(DBUF_S2),
+			[PIPE_D] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_C] = BIT(DBUF_S3),
+			[PIPE_D] = BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_C] = BIT(DBUF_S3),
+			[PIPE_D] = BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_C] = BIT(DBUF_S3),
+			[PIPE_D] = BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1),
+			[PIPE_B] = BIT(DBUF_S2),
+			[PIPE_C] = BIT(DBUF_S3),
+			[PIPE_D] = BIT(DBUF_S4),
+		},
+	},
+	{}
+};
+
+static const struct dbuf_slice_conf_entry adlp_allowed_dbufs[] = {
+	/*
+	 * Keep the join_mbus cases first so check_mbus_joined()
+	 * will prefer them over the !join_mbus cases.
+	 */
+	{
+		.active_pipes = BIT(PIPE_A),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2) | BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+		.join_mbus = true,
+	},
+	{
+		.active_pipes = BIT(PIPE_B),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S1) | BIT(DBUF_S2) | BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+		.join_mbus = true,
+	},
+	{
+		.active_pipes = BIT(PIPE_A),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+		.join_mbus = false,
+	},
+	{
+		.active_pipes = BIT(PIPE_B),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+		.join_mbus = false,
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_B] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S3) | BIT(DBUF_S4),
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_B] = BIT(DBUF_S3) | BIT(DBUF_S4),
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_D] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_D] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S3) | BIT(DBUF_S4),
+			[PIPE_D] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_B] = BIT(DBUF_S3) | BIT(DBUF_S4),
+			[PIPE_D] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+			[PIPE_D] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+			[PIPE_D] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_B] = BIT(DBUF_S3) | BIT(DBUF_S4),
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+			[PIPE_D] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{
+		.active_pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D),
+		.dbuf_mask = {
+			[PIPE_A] = BIT(DBUF_S1) | BIT(DBUF_S2),
+			[PIPE_B] = BIT(DBUF_S3) | BIT(DBUF_S4),
+			[PIPE_C] = BIT(DBUF_S3) | BIT(DBUF_S4),
+			[PIPE_D] = BIT(DBUF_S1) | BIT(DBUF_S2),
+		},
+	},
+	{}
+
+};
+
+static bool check_mbus_joined(u8 active_pipes,
+			      const struct dbuf_slice_conf_entry *dbuf_slices)
+{
+	int i;
+
+	for (i = 0; dbuf_slices[i].active_pipes != 0; i++) {
+		if (dbuf_slices[i].active_pipes == active_pipes)
+			return dbuf_slices[i].join_mbus;
+	}
+	return false;
+}
+
+static bool adlp_check_mbus_joined(u8 active_pipes)
+{
+	return check_mbus_joined(active_pipes, adlp_allowed_dbufs);
+}
+
+static u8 compute_dbuf_slices(enum pipe pipe, u8 active_pipes, bool join_mbus,
+			      const struct dbuf_slice_conf_entry *dbuf_slices)
+{
+	int i;
+
+	for (i = 0; dbuf_slices[i].active_pipes != 0; i++) {
+		if (dbuf_slices[i].active_pipes == active_pipes &&
+		    dbuf_slices[i].join_mbus == join_mbus)
+			return dbuf_slices[i].dbuf_mask[pipe];
+	}
+	return 0;
+}
+
+/*
+ * This function finds an entry with same enabled pipe configuration and
+ * returns correspondent DBuf slice mask as stated in BSpec for particular
+ * platform.
+ */
+static u8 icl_compute_dbuf_slices(enum pipe pipe, u8 active_pipes, bool join_mbus)
+{
+	/*
+	 * FIXME: For ICL this is still a bit unclear as prev BSpec revision
+	 * required calculating "pipe ratio" in order to determine
+	 * if one or two slices can be used for single pipe configurations
+	 * as additional constraint to the existing table.
+	 * However based on recent info, it should be not "pipe ratio"
+	 * but rather ratio between pixel_rate and cdclk with additional
+	 * constants, so for now we are using only table until this is
+	 * clarified. Also this is the reason why crtc_state param is
+	 * still here - we will need it once those additional constraints
+	 * pop up.
+	 */
+	return compute_dbuf_slices(pipe, active_pipes, join_mbus,
+				   icl_allowed_dbufs);
+}
+
+static u8 tgl_compute_dbuf_slices(enum pipe pipe, u8 active_pipes, bool join_mbus)
+{
+	return compute_dbuf_slices(pipe, active_pipes, join_mbus,
+				   tgl_allowed_dbufs);
+}
+
+static u8 adlp_compute_dbuf_slices(enum pipe pipe, u8 active_pipes, bool join_mbus)
+{
+	return compute_dbuf_slices(pipe, active_pipes, join_mbus,
+				   adlp_allowed_dbufs);
+}
+
+static u8 dg2_compute_dbuf_slices(enum pipe pipe, u8 active_pipes, bool join_mbus)
+{
+	return compute_dbuf_slices(pipe, active_pipes, join_mbus,
+				   dg2_allowed_dbufs);
+}
+
+static u8 skl_compute_dbuf_slices(struct intel_crtc *crtc, u8 active_pipes, bool join_mbus)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+
+	if (IS_DG2(i915))
+		return dg2_compute_dbuf_slices(pipe, active_pipes, join_mbus);
+	else if (DISPLAY_VER(i915) >= 13)
+		return adlp_compute_dbuf_slices(pipe, active_pipes, join_mbus);
+	else if (DISPLAY_VER(i915) == 12)
+		return tgl_compute_dbuf_slices(pipe, active_pipes, join_mbus);
+	else if (DISPLAY_VER(i915) == 11)
+		return icl_compute_dbuf_slices(pipe, active_pipes, join_mbus);
+	/*
+	 * For anything else just return one slice yet.
+	 * Should be extended for other platforms.
+	 */
+	return active_pipes & BIT(pipe) ? BIT(DBUF_S1) : 0;
+}
+
+static bool
+use_minimal_wm0_only(const struct intel_crtc_state *crtc_state,
+		     struct intel_plane *plane)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+
+	return DISPLAY_VER(i915) >= 13 &&
+	       crtc_state->uapi.async_flip &&
+	       plane->async_flip;
+}
+
+static u64
+skl_total_relative_data_rate(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	enum plane_id plane_id;
+	u64 data_rate = 0;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		if (plane_id == PLANE_CURSOR)
+			continue;
+
+		data_rate += crtc_state->rel_data_rate[plane_id];
+
+		if (DISPLAY_VER(i915) < 11)
+			data_rate += crtc_state->rel_data_rate_y[plane_id];
+	}
+
+	return data_rate;
+}
+
+static const struct skl_wm_level *
+skl_plane_wm_level(const struct skl_pipe_wm *pipe_wm,
+		   enum plane_id plane_id,
+		   int level)
+{
+	const struct skl_plane_wm *wm = &pipe_wm->planes[plane_id];
+
+	if (level == 0 && pipe_wm->use_sagv_wm)
+		return &wm->sagv.wm0;
+
+	return &wm->wm[level];
+}
+
+static const struct skl_wm_level *
+skl_plane_trans_wm(const struct skl_pipe_wm *pipe_wm,
+		   enum plane_id plane_id)
+{
+	const struct skl_plane_wm *wm = &pipe_wm->planes[plane_id];
+
+	if (pipe_wm->use_sagv_wm)
+		return &wm->sagv.trans_wm;
+
+	return &wm->trans_wm;
+}
+
+/*
+ * We only disable the watermarks for each plane if
+ * they exceed the ddb allocation of said plane. This
+ * is done so that we don't end up touching cursor
+ * watermarks needlessly when some other plane reduces
+ * our max possible watermark level.
+ *
+ * Bspec has this to say about the PLANE_WM enable bit:
+ * "All the watermarks at this level for all enabled
+ *  planes must be enabled before the level will be used."
+ * So this is actually safe to do.
+ */
+static void
+skl_check_wm_level(struct skl_wm_level *wm, const struct skl_ddb_entry *ddb)
+{
+	if (wm->min_ddb_alloc > skl_ddb_entry_size(ddb))
+		memset(wm, 0, sizeof(*wm));
+}
+
+static void
+skl_check_nv12_wm_level(struct skl_wm_level *wm, struct skl_wm_level *uv_wm,
+			const struct skl_ddb_entry *ddb_y, const struct skl_ddb_entry *ddb)
+{
+	if (wm->min_ddb_alloc > skl_ddb_entry_size(ddb_y) ||
+	    uv_wm->min_ddb_alloc > skl_ddb_entry_size(ddb)) {
+		memset(wm, 0, sizeof(*wm));
+		memset(uv_wm, 0, sizeof(*uv_wm));
+	}
+}
+
+static bool icl_need_wm1_wa(struct drm_i915_private *i915,
+			    enum plane_id plane_id)
+{
+	/*
+	 * Wa_1408961008:icl, ehl
+	 * Wa_14012656716:tgl, adl
+	 * Underruns with WM1+ disabled
+	 */
+	return DISPLAY_VER(i915) == 11 ||
+	       (IS_DISPLAY_VER(i915, 12, 13) && plane_id == PLANE_CURSOR);
+}
+
+struct skl_plane_ddb_iter {
+	u64 data_rate;
+	u16 start, size;
+};
+
+static void
+skl_allocate_plane_ddb(struct skl_plane_ddb_iter *iter,
+		       struct skl_ddb_entry *ddb,
+		       const struct skl_wm_level *wm,
+		       u64 data_rate)
+{
+	u16 size, extra = 0;
+
+	if (data_rate) {
+		extra = min_t(u16, iter->size,
+			      DIV64_U64_ROUND_UP(iter->size * data_rate,
+						 iter->data_rate));
+		iter->size -= extra;
+		iter->data_rate -= data_rate;
+	}
+
+	/*
+	 * Keep ddb entry of all disabled planes explicitly zeroed
+	 * to avoid skl_ddb_add_affected_planes() adding them to
+	 * the state when other planes change their allocations.
+	 */
+	size = wm->min_ddb_alloc + extra;
+	if (size)
+		iter->start = skl_ddb_entry_init(ddb, iter->start,
+						 iter->start + size);
+}
+
+static int
+skl_crtc_allocate_plane_ddb(struct intel_atomic_state *state,
+			    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_dbuf_state *dbuf_state =
+		intel_atomic_get_new_dbuf_state(state);
+	const struct skl_ddb_entry *alloc = &dbuf_state->ddb[crtc->pipe];
+	int num_active = hweight8(dbuf_state->active_pipes);
+	struct skl_plane_ddb_iter iter;
+	enum plane_id plane_id;
+	u16 cursor_size;
+	u32 blocks;
+	int level;
+
+	/* Clear the partitioning for disabled planes. */
+	memset(crtc_state->wm.skl.plane_ddb, 0, sizeof(crtc_state->wm.skl.plane_ddb));
+	memset(crtc_state->wm.skl.plane_ddb_y, 0, sizeof(crtc_state->wm.skl.plane_ddb_y));
+
+	if (!crtc_state->hw.active)
+		return 0;
+
+	iter.start = alloc->start;
+	iter.size = skl_ddb_entry_size(alloc);
+	if (iter.size == 0)
+		return 0;
+
+	/* Allocate fixed number of blocks for cursor. */
+	cursor_size = skl_cursor_allocation(crtc_state, num_active);
+	iter.size -= cursor_size;
+	skl_ddb_entry_init(&crtc_state->wm.skl.plane_ddb[PLANE_CURSOR],
+			   alloc->end - cursor_size, alloc->end);
+
+	iter.data_rate = skl_total_relative_data_rate(crtc_state);
+
+	/*
+	 * Find the highest watermark level for which we can satisfy the block
+	 * requirement of active planes.
+	 */
+	for (level = ilk_wm_max_level(i915); level >= 0; level--) {
+		blocks = 0;
+		for_each_plane_id_on_crtc(crtc, plane_id) {
+			const struct skl_plane_wm *wm =
+				&crtc_state->wm.skl.optimal.planes[plane_id];
+
+			if (plane_id == PLANE_CURSOR) {
+				const struct skl_ddb_entry *ddb =
+					&crtc_state->wm.skl.plane_ddb[plane_id];
+
+				if (wm->wm[level].min_ddb_alloc > skl_ddb_entry_size(ddb)) {
+					drm_WARN_ON(&i915->drm,
+						    wm->wm[level].min_ddb_alloc != U16_MAX);
+					blocks = U32_MAX;
+					break;
+				}
+				continue;
+			}
+
+			blocks += wm->wm[level].min_ddb_alloc;
+			blocks += wm->uv_wm[level].min_ddb_alloc;
+		}
+
+		if (blocks <= iter.size) {
+			iter.size -= blocks;
+			break;
+		}
+	}
+
+	if (level < 0) {
+		drm_dbg_kms(&i915->drm,
+			    "Requested display configuration exceeds system DDB limitations");
+		drm_dbg_kms(&i915->drm, "minimum required %d/%d\n",
+			    blocks, iter.size);
+		return -EINVAL;
+	}
+
+	/* avoid the WARN later when we don't allocate any extra DDB */
+	if (iter.data_rate == 0)
+		iter.size = 0;
+
+	/*
+	 * Grant each plane the blocks it requires at the highest achievable
+	 * watermark level, plus an extra share of the leftover blocks
+	 * proportional to its relative data rate.
+	 */
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		struct skl_ddb_entry *ddb =
+			&crtc_state->wm.skl.plane_ddb[plane_id];
+		struct skl_ddb_entry *ddb_y =
+			&crtc_state->wm.skl.plane_ddb_y[plane_id];
+		const struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+
+		if (plane_id == PLANE_CURSOR)
+			continue;
+
+		if (DISPLAY_VER(i915) < 11 &&
+		    crtc_state->nv12_planes & BIT(plane_id)) {
+			skl_allocate_plane_ddb(&iter, ddb_y, &wm->wm[level],
+					       crtc_state->rel_data_rate_y[plane_id]);
+			skl_allocate_plane_ddb(&iter, ddb, &wm->uv_wm[level],
+					       crtc_state->rel_data_rate[plane_id]);
+		} else {
+			skl_allocate_plane_ddb(&iter, ddb, &wm->wm[level],
+					       crtc_state->rel_data_rate[plane_id]);
+		}
+	}
+	drm_WARN_ON(&i915->drm, iter.size != 0 || iter.data_rate != 0);
+
+	/*
+	 * When we calculated watermark values we didn't know how high
+	 * of a level we'd actually be able to hit, so we just marked
+	 * all levels as "enabled."  Go back now and disable the ones
+	 * that aren't actually possible.
+	 */
+	for (level++; level <= ilk_wm_max_level(i915); level++) {
+		for_each_plane_id_on_crtc(crtc, plane_id) {
+			const struct skl_ddb_entry *ddb =
+				&crtc_state->wm.skl.plane_ddb[plane_id];
+			const struct skl_ddb_entry *ddb_y =
+				&crtc_state->wm.skl.plane_ddb_y[plane_id];
+			struct skl_plane_wm *wm =
+				&crtc_state->wm.skl.optimal.planes[plane_id];
+
+			if (DISPLAY_VER(i915) < 11 &&
+			    crtc_state->nv12_planes & BIT(plane_id))
+				skl_check_nv12_wm_level(&wm->wm[level],
+							&wm->uv_wm[level],
+							ddb_y, ddb);
+			else
+				skl_check_wm_level(&wm->wm[level], ddb);
+
+			if (icl_need_wm1_wa(i915, plane_id) &&
+			    level == 1 && !wm->wm[level].enable &&
+			    wm->wm[0].enable) {
+				wm->wm[level].blocks = wm->wm[0].blocks;
+				wm->wm[level].lines = wm->wm[0].lines;
+				wm->wm[level].ignore_lines = wm->wm[0].ignore_lines;
+			}
+		}
+	}
+
+	/*
+	 * Go back and disable the transition and SAGV watermarks
+	 * if it turns out we don't have enough DDB blocks for them.
+	 */
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		const struct skl_ddb_entry *ddb =
+			&crtc_state->wm.skl.plane_ddb[plane_id];
+		const struct skl_ddb_entry *ddb_y =
+			&crtc_state->wm.skl.plane_ddb_y[plane_id];
+		struct skl_plane_wm *wm =
+			&crtc_state->wm.skl.optimal.planes[plane_id];
+
+		if (DISPLAY_VER(i915) < 11 &&
+		    crtc_state->nv12_planes & BIT(plane_id)) {
+			skl_check_wm_level(&wm->trans_wm, ddb_y);
+		} else {
+			WARN_ON(skl_ddb_entry_size(ddb_y));
+
+			skl_check_wm_level(&wm->trans_wm, ddb);
+		}
+
+		skl_check_wm_level(&wm->sagv.wm0, ddb);
+		skl_check_wm_level(&wm->sagv.trans_wm, ddb);
+	}
+
+	return 0;
+}
+
+/*
+ * The max latency should be 257 (max the punit can code is 255 and we add 2us
+ * for the read latency) and cpp should always be <= 8, so that
+ * should allow pixel_rate up to ~2 GHz which seems sufficient since max
+ * 2xcdclk is 1350 MHz and the pixel rate should never exceed that.
+ */
+static uint_fixed_16_16_t
+skl_wm_method1(const struct drm_i915_private *i915, u32 pixel_rate,
+	       u8 cpp, u32 latency, u32 dbuf_block_size)
+{
+	u32 wm_intermediate_val;
+	uint_fixed_16_16_t ret;
+
+	if (latency == 0)
+		return FP_16_16_MAX;
+
+	wm_intermediate_val = latency * pixel_rate * cpp;
+	ret = div_fixed16(wm_intermediate_val, 1000 * dbuf_block_size);
+
+	if (DISPLAY_VER(i915) >= 10)
+		ret = add_fixed16_u32(ret, 1);
+
+	return ret;
+}
+
+static uint_fixed_16_16_t
+skl_wm_method2(u32 pixel_rate, u32 pipe_htotal, u32 latency,
+	       uint_fixed_16_16_t plane_blocks_per_line)
+{
+	u32 wm_intermediate_val;
+	uint_fixed_16_16_t ret;
+
+	if (latency == 0)
+		return FP_16_16_MAX;
+
+	wm_intermediate_val = latency * pixel_rate;
+	wm_intermediate_val = DIV_ROUND_UP(wm_intermediate_val,
+					   pipe_htotal * 1000);
+	ret = mul_u32_fixed16(wm_intermediate_val, plane_blocks_per_line);
+	return ret;
+}
+
+static uint_fixed_16_16_t
+intel_get_linetime_us(const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	u32 pixel_rate;
+	u32 crtc_htotal;
+	uint_fixed_16_16_t linetime_us;
+
+	if (!crtc_state->hw.active)
+		return u32_to_fixed16(0);
+
+	pixel_rate = crtc_state->pixel_rate;
+
+	if (drm_WARN_ON(&i915->drm, pixel_rate == 0))
+		return u32_to_fixed16(0);
+
+	crtc_htotal = crtc_state->hw.pipe_mode.crtc_htotal;
+	linetime_us = div_fixed16(crtc_htotal * 1000, pixel_rate);
+
+	return linetime_us;
+}
+
+static int
+skl_compute_wm_params(const struct intel_crtc_state *crtc_state,
+		      int width, const struct drm_format_info *format,
+		      u64 modifier, unsigned int rotation,
+		      u32 plane_pixel_rate, struct skl_wm_params *wp,
+		      int color_plane)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	u32 interm_pbpl;
+
+	/* only planar format has two planes */
+	if (color_plane == 1 &&
+	    !intel_format_info_is_yuv_semiplanar(format, modifier)) {
+		drm_dbg_kms(&i915->drm,
+			    "Non planar format have single plane\n");
+		return -EINVAL;
+	}
+
+	wp->y_tiled = modifier == I915_FORMAT_MOD_Y_TILED ||
+		      modifier == I915_FORMAT_MOD_4_TILED ||
+		      modifier == I915_FORMAT_MOD_Yf_TILED ||
+		      modifier == I915_FORMAT_MOD_Y_TILED_CCS ||
+		      modifier == I915_FORMAT_MOD_Yf_TILED_CCS ||
+		      modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS ||
+		      modifier == I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS ||
+		      modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC ||
+		      modifier == I915_FORMAT_MOD_4_TILED_DG2_RC_CCS ||
+		      modifier == I915_FORMAT_MOD_4_TILED_DG2_MC_CCS ||
+		      modifier == I915_FORMAT_MOD_4_TILED_DG2_RC_CCS_CC;
+	wp->x_tiled = modifier == I915_FORMAT_MOD_X_TILED;
+	wp->rc_surface = modifier == I915_FORMAT_MOD_Y_TILED_CCS ||
+			 modifier == I915_FORMAT_MOD_Yf_TILED_CCS ||
+			 modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS ||
+			 modifier == I915_FORMAT_MOD_Y_TILED_GEN12_MC_CCS ||
+			 modifier == I915_FORMAT_MOD_Y_TILED_GEN12_RC_CCS_CC ||
+			 modifier == I915_FORMAT_MOD_4_TILED_DG2_RC_CCS ||
+			 modifier == I915_FORMAT_MOD_4_TILED_DG2_MC_CCS ||
+			 modifier == I915_FORMAT_MOD_4_TILED_DG2_RC_CCS_CC;
+	wp->is_planar = intel_format_info_is_yuv_semiplanar(format, modifier);
+
+	wp->width = width;
+	if (color_plane == 1 && wp->is_planar)
+		wp->width /= 2;
+
+	wp->cpp = format->cpp[color_plane];
+	wp->plane_pixel_rate = plane_pixel_rate;
+
+	if (DISPLAY_VER(i915) >= 11 &&
+	    modifier == I915_FORMAT_MOD_Yf_TILED  && wp->cpp == 1)
+		wp->dbuf_block_size = 256;
+	else
+		wp->dbuf_block_size = 512;
+
+	if (drm_rotation_90_or_270(rotation)) {
+		switch (wp->cpp) {
+		case 1:
+			wp->y_min_scanlines = 16;
+			break;
+		case 2:
+			wp->y_min_scanlines = 8;
+			break;
+		case 4:
+			wp->y_min_scanlines = 4;
+			break;
+		default:
+			MISSING_CASE(wp->cpp);
+			return -EINVAL;
+		}
+	} else {
+		wp->y_min_scanlines = 4;
+	}
+
+	if (skl_needs_memory_bw_wa(i915))
+		wp->y_min_scanlines *= 2;
+
+	wp->plane_bytes_per_line = wp->width * wp->cpp;
+	if (wp->y_tiled) {
+		interm_pbpl = DIV_ROUND_UP(wp->plane_bytes_per_line *
+					   wp->y_min_scanlines,
+					   wp->dbuf_block_size);
+
+		if (DISPLAY_VER(i915) >= 10)
+			interm_pbpl++;
+
+		wp->plane_blocks_per_line = div_fixed16(interm_pbpl,
+							wp->y_min_scanlines);
+	} else {
+		interm_pbpl = DIV_ROUND_UP(wp->plane_bytes_per_line,
+					   wp->dbuf_block_size);
+
+		if (!wp->x_tiled || DISPLAY_VER(i915) >= 10)
+			interm_pbpl++;
+
+		wp->plane_blocks_per_line = u32_to_fixed16(interm_pbpl);
+	}
+
+	wp->y_tile_minimum = mul_u32_fixed16(wp->y_min_scanlines,
+					     wp->plane_blocks_per_line);
+
+	wp->linetime_us = fixed16_to_u32_round_up(intel_get_linetime_us(crtc_state));
+
+	return 0;
+}
+
+static int
+skl_compute_plane_wm_params(const struct intel_crtc_state *crtc_state,
+			    const struct intel_plane_state *plane_state,
+			    struct skl_wm_params *wp, int color_plane)
+{
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int width;
+
+	/*
+	 * Src coordinates are already rotated by 270 degrees for
+	 * the 90/270 degree plane rotation cases (to match the
+	 * GTT mapping), hence no need to account for rotation here.
+	 */
+	width = drm_rect_width(&plane_state->uapi.src) >> 16;
+
+	return skl_compute_wm_params(crtc_state, width,
+				     fb->format, fb->modifier,
+				     plane_state->hw.rotation,
+				     intel_plane_pixel_rate(crtc_state, plane_state),
+				     wp, color_plane);
+}
+
+static bool skl_wm_has_lines(struct drm_i915_private *i915, int level)
+{
+	if (DISPLAY_VER(i915) >= 10)
+		return true;
+
+	/* The number of lines are ignored for the level 0 watermark. */
+	return level > 0;
+}
+
+static int skl_wm_max_lines(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 13)
+		return 255;
+	else
+		return 31;
+}
+
+static void skl_compute_plane_wm(const struct intel_crtc_state *crtc_state,
+				 struct intel_plane *plane,
+				 int level,
+				 unsigned int latency,
+				 const struct skl_wm_params *wp,
+				 const struct skl_wm_level *result_prev,
+				 struct skl_wm_level *result /* out */)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	uint_fixed_16_16_t method1, method2;
+	uint_fixed_16_16_t selected_result;
+	u32 blocks, lines, min_ddb_alloc = 0;
+
+	if (latency == 0 ||
+	    (use_minimal_wm0_only(crtc_state, plane) && level > 0)) {
+		/* reject it */
+		result->min_ddb_alloc = U16_MAX;
+		return;
+	}
+
+	/*
+	 * WaIncreaseLatencyIPCEnabled: kbl,cfl
+	 * Display WA #1141: kbl,cfl
+	 */
+	if ((IS_KABYLAKE(i915) || IS_COFFEELAKE(i915) || IS_COMETLAKE(i915)) &&
+	    skl_watermark_ipc_enabled(i915))
+		latency += 4;
+
+	if (skl_needs_memory_bw_wa(i915) && wp->x_tiled)
+		latency += 15;
+
+	method1 = skl_wm_method1(i915, wp->plane_pixel_rate,
+				 wp->cpp, latency, wp->dbuf_block_size);
+	method2 = skl_wm_method2(wp->plane_pixel_rate,
+				 crtc_state->hw.pipe_mode.crtc_htotal,
+				 latency,
+				 wp->plane_blocks_per_line);
+
+	if (wp->y_tiled) {
+		selected_result = max_fixed16(method2, wp->y_tile_minimum);
+	} else {
+		if ((wp->cpp * crtc_state->hw.pipe_mode.crtc_htotal /
+		     wp->dbuf_block_size < 1) &&
+		     (wp->plane_bytes_per_line / wp->dbuf_block_size < 1)) {
+			selected_result = method2;
+		} else if (latency >= wp->linetime_us) {
+			if (DISPLAY_VER(i915) == 9)
+				selected_result = min_fixed16(method1, method2);
+			else
+				selected_result = method2;
+		} else {
+			selected_result = method1;
+		}
+	}
+
+	blocks = fixed16_to_u32_round_up(selected_result) + 1;
+	/*
+	 * Lets have blocks at minimum equivalent to plane_blocks_per_line
+	 * as there will be at minimum one line for lines configuration. This
+	 * is a work around for FIFO underruns observed with resolutions like
+	 * 4k 60 Hz in single channel DRAM configurations.
+	 *
+	 * As per the Bspec 49325, if the ddb allocation can hold at least
+	 * one plane_blocks_per_line, we should have selected method2 in
+	 * the above logic. Assuming that modern versions have enough dbuf
+	 * and method2 guarantees blocks equivalent to at least 1 line,
+	 * select the blocks as plane_blocks_per_line.
+	 *
+	 * TODO: Revisit the logic when we have better understanding on DRAM
+	 * channels' impact on the level 0 memory latency and the relevant
+	 * wm calculations.
+	 */
+	if (skl_wm_has_lines(i915, level))
+		blocks = max(blocks,
+			     fixed16_to_u32_round_up(wp->plane_blocks_per_line));
+	lines = div_round_up_fixed16(selected_result,
+				     wp->plane_blocks_per_line);
+
+	if (DISPLAY_VER(i915) == 9) {
+		/* Display WA #1125: skl,bxt,kbl */
+		if (level == 0 && wp->rc_surface)
+			blocks += fixed16_to_u32_round_up(wp->y_tile_minimum);
+
+		/* Display WA #1126: skl,bxt,kbl */
+		if (level >= 1 && level <= 7) {
+			if (wp->y_tiled) {
+				blocks += fixed16_to_u32_round_up(wp->y_tile_minimum);
+				lines += wp->y_min_scanlines;
+			} else {
+				blocks++;
+			}
+
+			/*
+			 * Make sure result blocks for higher latency levels are
+			 * at least as high as level below the current level.
+			 * Assumption in DDB algorithm optimization for special
+			 * cases. Also covers Display WA #1125 for RC.
+			 */
+			if (result_prev->blocks > blocks)
+				blocks = result_prev->blocks;
+		}
+	}
+
+	if (DISPLAY_VER(i915) >= 11) {
+		if (wp->y_tiled) {
+			int extra_lines;
+
+			if (lines % wp->y_min_scanlines == 0)
+				extra_lines = wp->y_min_scanlines;
+			else
+				extra_lines = wp->y_min_scanlines * 2 -
+					lines % wp->y_min_scanlines;
+
+			min_ddb_alloc = mul_round_up_u32_fixed16(lines + extra_lines,
+								 wp->plane_blocks_per_line);
+		} else {
+			min_ddb_alloc = blocks + DIV_ROUND_UP(blocks, 10);
+		}
+	}
+
+	if (!skl_wm_has_lines(i915, level))
+		lines = 0;
+
+	if (lines > skl_wm_max_lines(i915)) {
+		/* reject it */
+		result->min_ddb_alloc = U16_MAX;
+		return;
+	}
+
+	/*
+	 * If lines is valid, assume we can use this watermark level
+	 * for now.  We'll come back and disable it after we calculate the
+	 * DDB allocation if it turns out we don't actually have enough
+	 * blocks to satisfy it.
+	 */
+	result->blocks = blocks;
+	result->lines = lines;
+	/* Bspec says: value >= plane ddb allocation -> invalid, hence the +1 here */
+	result->min_ddb_alloc = max(min_ddb_alloc, blocks) + 1;
+	result->enable = true;
+
+	if (DISPLAY_VER(i915) < 12 && i915->display.sagv.block_time_us)
+		result->can_sagv = latency >= i915->display.sagv.block_time_us;
+}
+
+static void
+skl_compute_wm_levels(const struct intel_crtc_state *crtc_state,
+		      struct intel_plane *plane,
+		      const struct skl_wm_params *wm_params,
+		      struct skl_wm_level *levels)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	int level, max_level = ilk_wm_max_level(i915);
+	struct skl_wm_level *result_prev = &levels[0];
+
+	for (level = 0; level <= max_level; level++) {
+		struct skl_wm_level *result = &levels[level];
+		unsigned int latency = i915->display.wm.skl_latency[level];
+
+		skl_compute_plane_wm(crtc_state, plane, level, latency,
+				     wm_params, result_prev, result);
+
+		result_prev = result;
+	}
+}
+
+static void tgl_compute_sagv_wm(const struct intel_crtc_state *crtc_state,
+				struct intel_plane *plane,
+				const struct skl_wm_params *wm_params,
+				struct skl_plane_wm *plane_wm)
+{
+	struct drm_i915_private *i915 = to_i915(crtc_state->uapi.crtc->dev);
+	struct skl_wm_level *sagv_wm = &plane_wm->sagv.wm0;
+	struct skl_wm_level *levels = plane_wm->wm;
+	unsigned int latency = 0;
+
+	if (i915->display.sagv.block_time_us)
+		latency = i915->display.sagv.block_time_us + i915->display.wm.skl_latency[0];
+
+	skl_compute_plane_wm(crtc_state, plane, 0, latency,
+			     wm_params, &levels[0],
+			     sagv_wm);
+}
+
+static void skl_compute_transition_wm(struct drm_i915_private *i915,
+				      struct skl_wm_level *trans_wm,
+				      const struct skl_wm_level *wm0,
+				      const struct skl_wm_params *wp)
+{
+	u16 trans_min, trans_amount, trans_y_tile_min;
+	u16 wm0_blocks, trans_offset, blocks;
+
+	/* Transition WM don't make any sense if ipc is disabled */
+	if (!skl_watermark_ipc_enabled(i915))
+		return;
+
+	/*
+	 * WaDisableTWM:skl,kbl,cfl,bxt
+	 * Transition WM are not recommended by HW team for GEN9
+	 */
+	if (DISPLAY_VER(i915) == 9)
+		return;
+
+	if (DISPLAY_VER(i915) >= 11)
+		trans_min = 4;
+	else
+		trans_min = 14;
+
+	/* Display WA #1140: glk,cnl */
+	if (DISPLAY_VER(i915) == 10)
+		trans_amount = 0;
+	else
+		trans_amount = 10; /* This is configurable amount */
+
+	trans_offset = trans_min + trans_amount;
+
+	/*
+	 * The spec asks for Selected Result Blocks for wm0 (the real value),
+	 * not Result Blocks (the integer value). Pay attention to the capital
+	 * letters. The value wm_l0->blocks is actually Result Blocks, but
+	 * since Result Blocks is the ceiling of Selected Result Blocks plus 1,
+	 * and since we later will have to get the ceiling of the sum in the
+	 * transition watermarks calculation, we can just pretend Selected
+	 * Result Blocks is Result Blocks minus 1 and it should work for the
+	 * current platforms.
+	 */
+	wm0_blocks = wm0->blocks - 1;
+
+	if (wp->y_tiled) {
+		trans_y_tile_min =
+			(u16)mul_round_up_u32_fixed16(2, wp->y_tile_minimum);
+		blocks = max(wm0_blocks, trans_y_tile_min) + trans_offset;
+	} else {
+		blocks = wm0_blocks + trans_offset;
+	}
+	blocks++;
+
+	/*
+	 * Just assume we can enable the transition watermark.  After
+	 * computing the DDB we'll come back and disable it if that
+	 * assumption turns out to be false.
+	 */
+	trans_wm->blocks = blocks;
+	trans_wm->min_ddb_alloc = max_t(u16, wm0->min_ddb_alloc, blocks + 1);
+	trans_wm->enable = true;
+}
+
+static int skl_build_plane_wm_single(struct intel_crtc_state *crtc_state,
+				     const struct intel_plane_state *plane_state,
+				     struct intel_plane *plane, int color_plane)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct skl_plane_wm *wm = &crtc_state->wm.skl.raw.planes[plane->id];
+	struct skl_wm_params wm_params;
+	int ret;
+
+	ret = skl_compute_plane_wm_params(crtc_state, plane_state,
+					  &wm_params, color_plane);
+	if (ret)
+		return ret;
+
+	skl_compute_wm_levels(crtc_state, plane, &wm_params, wm->wm);
+
+	skl_compute_transition_wm(i915, &wm->trans_wm,
+				  &wm->wm[0], &wm_params);
+
+	if (DISPLAY_VER(i915) >= 12) {
+		tgl_compute_sagv_wm(crtc_state, plane, &wm_params, wm);
+
+		skl_compute_transition_wm(i915, &wm->sagv.trans_wm,
+					  &wm->sagv.wm0, &wm_params);
+	}
+
+	return 0;
+}
+
+static int skl_build_plane_wm_uv(struct intel_crtc_state *crtc_state,
+				 const struct intel_plane_state *plane_state,
+				 struct intel_plane *plane)
+{
+	struct skl_plane_wm *wm = &crtc_state->wm.skl.raw.planes[plane->id];
+	struct skl_wm_params wm_params;
+	int ret;
+
+	wm->is_planar = true;
+
+	/* uv plane watermarks must also be validated for NV12/Planar */
+	ret = skl_compute_plane_wm_params(crtc_state, plane_state,
+					  &wm_params, 1);
+	if (ret)
+		return ret;
+
+	skl_compute_wm_levels(crtc_state, plane, &wm_params, wm->uv_wm);
+
+	return 0;
+}
+
+static int skl_build_plane_wm(struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	enum plane_id plane_id = plane->id;
+	struct skl_plane_wm *wm = &crtc_state->wm.skl.raw.planes[plane_id];
+	const struct drm_framebuffer *fb = plane_state->hw.fb;
+	int ret;
+
+	memset(wm, 0, sizeof(*wm));
+
+	if (!intel_wm_plane_visible(crtc_state, plane_state))
+		return 0;
+
+	ret = skl_build_plane_wm_single(crtc_state, plane_state,
+					plane, 0);
+	if (ret)
+		return ret;
+
+	if (fb->format->is_yuv && fb->format->num_planes > 1) {
+		ret = skl_build_plane_wm_uv(crtc_state, plane_state,
+					    plane);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int icl_build_plane_wm(struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	enum plane_id plane_id = plane->id;
+	struct skl_plane_wm *wm = &crtc_state->wm.skl.raw.planes[plane_id];
+	int ret;
+
+	/* Watermarks calculated in master */
+	if (plane_state->planar_slave)
+		return 0;
+
+	memset(wm, 0, sizeof(*wm));
+
+	if (plane_state->planar_linked_plane) {
+		const struct drm_framebuffer *fb = plane_state->hw.fb;
+
+		drm_WARN_ON(&i915->drm,
+			    !intel_wm_plane_visible(crtc_state, plane_state));
+		drm_WARN_ON(&i915->drm, !fb->format->is_yuv ||
+			    fb->format->num_planes == 1);
+
+		ret = skl_build_plane_wm_single(crtc_state, plane_state,
+						plane_state->planar_linked_plane, 0);
+		if (ret)
+			return ret;
+
+		ret = skl_build_plane_wm_single(crtc_state, plane_state,
+						plane, 1);
+		if (ret)
+			return ret;
+	} else if (intel_wm_plane_visible(crtc_state, plane_state)) {
+		ret = skl_build_plane_wm_single(crtc_state, plane_state,
+						plane, 0);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int skl_build_pipe_wm(struct intel_atomic_state *state,
+			     struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	int ret, i;
+
+	for_each_new_intel_plane_in_state(state, plane, plane_state, i) {
+		/*
+		 * FIXME should perhaps check {old,new}_plane_crtc->hw.crtc
+		 * instead but we don't populate that correctly for NV12 Y
+		 * planes so for now hack this.
+		 */
+		if (plane->pipe != crtc->pipe)
+			continue;
+
+		if (DISPLAY_VER(i915) >= 11)
+			ret = icl_build_plane_wm(crtc_state, plane_state);
+		else
+			ret = skl_build_plane_wm(crtc_state, plane_state);
+		if (ret)
+			return ret;
+	}
+
+	crtc_state->wm.skl.optimal = crtc_state->wm.skl.raw;
+
+	return 0;
+}
+
+static void skl_ddb_entry_write(struct drm_i915_private *i915,
+				i915_reg_t reg,
+				const struct skl_ddb_entry *entry)
+{
+	if (entry->end)
+		intel_de_write_fw(i915, reg,
+				  PLANE_BUF_END(entry->end - 1) |
+				  PLANE_BUF_START(entry->start));
+	else
+		intel_de_write_fw(i915, reg, 0);
+}
+
+static void skl_write_wm_level(struct drm_i915_private *i915,
+			       i915_reg_t reg,
+			       const struct skl_wm_level *level)
+{
+	u32 val = 0;
+
+	if (level->enable)
+		val |= PLANE_WM_EN;
+	if (level->ignore_lines)
+		val |= PLANE_WM_IGNORE_LINES;
+	val |= REG_FIELD_PREP(PLANE_WM_BLOCKS_MASK, level->blocks);
+	val |= REG_FIELD_PREP(PLANE_WM_LINES_MASK, level->lines);
+
+	intel_de_write_fw(i915, reg, val);
+}
+
+void skl_write_plane_wm(struct intel_plane *plane,
+			const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	int level, max_level = ilk_wm_max_level(i915);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe = plane->pipe;
+	const struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal;
+	const struct skl_ddb_entry *ddb =
+		&crtc_state->wm.skl.plane_ddb[plane_id];
+	const struct skl_ddb_entry *ddb_y =
+		&crtc_state->wm.skl.plane_ddb_y[plane_id];
+
+	for (level = 0; level <= max_level; level++)
+		skl_write_wm_level(i915, PLANE_WM(pipe, plane_id, level),
+				   skl_plane_wm_level(pipe_wm, plane_id, level));
+
+	skl_write_wm_level(i915, PLANE_WM_TRANS(pipe, plane_id),
+			   skl_plane_trans_wm(pipe_wm, plane_id));
+
+	if (HAS_HW_SAGV_WM(i915)) {
+		const struct skl_plane_wm *wm = &pipe_wm->planes[plane_id];
+
+		skl_write_wm_level(i915, PLANE_WM_SAGV(pipe, plane_id),
+				   &wm->sagv.wm0);
+		skl_write_wm_level(i915, PLANE_WM_SAGV_TRANS(pipe, plane_id),
+				   &wm->sagv.trans_wm);
+	}
+
+	skl_ddb_entry_write(i915,
+			    PLANE_BUF_CFG(pipe, plane_id), ddb);
+
+	if (DISPLAY_VER(i915) < 11)
+		skl_ddb_entry_write(i915,
+				    PLANE_NV12_BUF_CFG(pipe, plane_id), ddb_y);
+}
+
+void skl_write_cursor_wm(struct intel_plane *plane,
+			 const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	int level, max_level = ilk_wm_max_level(i915);
+	enum plane_id plane_id = plane->id;
+	enum pipe pipe = plane->pipe;
+	const struct skl_pipe_wm *pipe_wm = &crtc_state->wm.skl.optimal;
+	const struct skl_ddb_entry *ddb =
+		&crtc_state->wm.skl.plane_ddb[plane_id];
+
+	for (level = 0; level <= max_level; level++)
+		skl_write_wm_level(i915, CUR_WM(pipe, level),
+				   skl_plane_wm_level(pipe_wm, plane_id, level));
+
+	skl_write_wm_level(i915, CUR_WM_TRANS(pipe),
+			   skl_plane_trans_wm(pipe_wm, plane_id));
+
+	if (HAS_HW_SAGV_WM(i915)) {
+		const struct skl_plane_wm *wm = &pipe_wm->planes[plane_id];
+
+		skl_write_wm_level(i915, CUR_WM_SAGV(pipe),
+				   &wm->sagv.wm0);
+		skl_write_wm_level(i915, CUR_WM_SAGV_TRANS(pipe),
+				   &wm->sagv.trans_wm);
+	}
+
+	skl_ddb_entry_write(i915, CUR_BUF_CFG(pipe), ddb);
+}
+
+static bool skl_wm_level_equals(const struct skl_wm_level *l1,
+				const struct skl_wm_level *l2)
+{
+	return l1->enable == l2->enable &&
+		l1->ignore_lines == l2->ignore_lines &&
+		l1->lines == l2->lines &&
+		l1->blocks == l2->blocks;
+}
+
+static bool skl_plane_wm_equals(struct drm_i915_private *i915,
+				const struct skl_plane_wm *wm1,
+				const struct skl_plane_wm *wm2)
+{
+	int level, max_level = ilk_wm_max_level(i915);
+
+	for (level = 0; level <= max_level; level++) {
+		/*
+		 * We don't check uv_wm as the hardware doesn't actually
+		 * use it. It only gets used for calculating the required
+		 * ddb allocation.
+		 */
+		if (!skl_wm_level_equals(&wm1->wm[level], &wm2->wm[level]))
+			return false;
+	}
+
+	return skl_wm_level_equals(&wm1->trans_wm, &wm2->trans_wm) &&
+		skl_wm_level_equals(&wm1->sagv.wm0, &wm2->sagv.wm0) &&
+		skl_wm_level_equals(&wm1->sagv.trans_wm, &wm2->sagv.trans_wm);
+}
+
+static bool skl_ddb_entries_overlap(const struct skl_ddb_entry *a,
+				    const struct skl_ddb_entry *b)
+{
+	return a->start < b->end && b->start < a->end;
+}
+
+static void skl_ddb_entry_union(struct skl_ddb_entry *a,
+				const struct skl_ddb_entry *b)
+{
+	if (a->end && b->end) {
+		a->start = min(a->start, b->start);
+		a->end = max(a->end, b->end);
+	} else if (b->end) {
+		a->start = b->start;
+		a->end = b->end;
+	}
+}
+
+bool skl_ddb_allocation_overlaps(const struct skl_ddb_entry *ddb,
+				 const struct skl_ddb_entry *entries,
+				 int num_entries, int ignore_idx)
+{
+	int i;
+
+	for (i = 0; i < num_entries; i++) {
+		if (i != ignore_idx &&
+		    skl_ddb_entries_overlap(ddb, &entries[i]))
+			return true;
+	}
+
+	return false;
+}
+
+static int
+skl_ddb_add_affected_planes(const struct intel_crtc_state *old_crtc_state,
+			    struct intel_crtc_state *new_crtc_state)
+{
+	struct intel_atomic_state *state = to_intel_atomic_state(new_crtc_state->uapi.state);
+	struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->uapi.crtc);
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct intel_plane *plane;
+
+	for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) {
+		struct intel_plane_state *plane_state;
+		enum plane_id plane_id = plane->id;
+
+		if (skl_ddb_entry_equal(&old_crtc_state->wm.skl.plane_ddb[plane_id],
+					&new_crtc_state->wm.skl.plane_ddb[plane_id]) &&
+		    skl_ddb_entry_equal(&old_crtc_state->wm.skl.plane_ddb_y[plane_id],
+					&new_crtc_state->wm.skl.plane_ddb_y[plane_id]))
+			continue;
+
+		plane_state = intel_atomic_get_plane_state(state, plane);
+		if (IS_ERR(plane_state))
+			return PTR_ERR(plane_state);
+
+		new_crtc_state->update_planes |= BIT(plane_id);
+	}
+
+	return 0;
+}
+
+static u8 intel_dbuf_enabled_slices(const struct intel_dbuf_state *dbuf_state)
+{
+	struct drm_i915_private *i915 = to_i915(dbuf_state->base.state->base.dev);
+	u8 enabled_slices;
+	enum pipe pipe;
+
+	/*
+	 * FIXME: For now we always enable slice S1 as per
+	 * the Bspec display initialization sequence.
+	 */
+	enabled_slices = BIT(DBUF_S1);
+
+	for_each_pipe(i915, pipe)
+		enabled_slices |= dbuf_state->slices[pipe];
+
+	return enabled_slices;
+}
+
+static int
+skl_compute_ddb(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_dbuf_state *old_dbuf_state;
+	struct intel_dbuf_state *new_dbuf_state = NULL;
+	const struct intel_crtc_state *old_crtc_state;
+	struct intel_crtc_state *new_crtc_state;
+	struct intel_crtc *crtc;
+	int ret, i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		new_dbuf_state = intel_atomic_get_dbuf_state(state);
+		if (IS_ERR(new_dbuf_state))
+			return PTR_ERR(new_dbuf_state);
+
+		old_dbuf_state = intel_atomic_get_old_dbuf_state(state);
+		break;
+	}
+
+	if (!new_dbuf_state)
+		return 0;
+
+	new_dbuf_state->active_pipes =
+		intel_calc_active_pipes(state, old_dbuf_state->active_pipes);
+
+	if (old_dbuf_state->active_pipes != new_dbuf_state->active_pipes) {
+		ret = intel_atomic_lock_global_state(&new_dbuf_state->base);
+		if (ret)
+			return ret;
+	}
+
+	if (HAS_MBUS_JOINING(i915))
+		new_dbuf_state->joined_mbus =
+			adlp_check_mbus_joined(new_dbuf_state->active_pipes);
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		enum pipe pipe = crtc->pipe;
+
+		new_dbuf_state->slices[pipe] =
+			skl_compute_dbuf_slices(crtc, new_dbuf_state->active_pipes,
+						new_dbuf_state->joined_mbus);
+
+		if (old_dbuf_state->slices[pipe] == new_dbuf_state->slices[pipe])
+			continue;
+
+		ret = intel_atomic_lock_global_state(&new_dbuf_state->base);
+		if (ret)
+			return ret;
+	}
+
+	new_dbuf_state->enabled_slices = intel_dbuf_enabled_slices(new_dbuf_state);
+
+	if (old_dbuf_state->enabled_slices != new_dbuf_state->enabled_slices ||
+	    old_dbuf_state->joined_mbus != new_dbuf_state->joined_mbus) {
+		ret = intel_atomic_serialize_global_state(&new_dbuf_state->base);
+		if (ret)
+			return ret;
+
+		if (old_dbuf_state->joined_mbus != new_dbuf_state->joined_mbus) {
+			/* TODO: Implement vblank synchronized MBUS joining changes */
+			ret = intel_modeset_all_pipes(state);
+			if (ret)
+				return ret;
+		}
+
+		drm_dbg_kms(&i915->drm,
+			    "Enabled dbuf slices 0x%x -> 0x%x (total dbuf slices 0x%x), mbus joined? %s->%s\n",
+			    old_dbuf_state->enabled_slices,
+			    new_dbuf_state->enabled_slices,
+			    INTEL_INFO(i915)->display.dbuf.slice_mask,
+			    str_yes_no(old_dbuf_state->joined_mbus),
+			    str_yes_no(new_dbuf_state->joined_mbus));
+	}
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		enum pipe pipe = crtc->pipe;
+
+		new_dbuf_state->weight[pipe] = intel_crtc_ddb_weight(new_crtc_state);
+
+		if (old_dbuf_state->weight[pipe] == new_dbuf_state->weight[pipe])
+			continue;
+
+		ret = intel_atomic_lock_global_state(&new_dbuf_state->base);
+		if (ret)
+			return ret;
+	}
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		ret = skl_crtc_allocate_ddb(state, crtc);
+		if (ret)
+			return ret;
+	}
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		ret = skl_crtc_allocate_plane_ddb(state, crtc);
+		if (ret)
+			return ret;
+
+		ret = skl_ddb_add_affected_planes(old_crtc_state,
+						  new_crtc_state);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static char enast(bool enable)
+{
+	return enable ? '*' : ' ';
+}
+
+static void
+skl_print_wm_changes(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_crtc_state *old_crtc_state;
+	const struct intel_crtc_state *new_crtc_state;
+	struct intel_plane *plane;
+	struct intel_crtc *crtc;
+	int i;
+
+	if (!drm_debug_enabled(DRM_UT_KMS))
+		return;
+
+	for_each_oldnew_intel_crtc_in_state(state, crtc, old_crtc_state,
+					    new_crtc_state, i) {
+		const struct skl_pipe_wm *old_pipe_wm, *new_pipe_wm;
+
+		old_pipe_wm = &old_crtc_state->wm.skl.optimal;
+		new_pipe_wm = &new_crtc_state->wm.skl.optimal;
+
+		for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) {
+			enum plane_id plane_id = plane->id;
+			const struct skl_ddb_entry *old, *new;
+
+			old = &old_crtc_state->wm.skl.plane_ddb[plane_id];
+			new = &new_crtc_state->wm.skl.plane_ddb[plane_id];
+
+			if (skl_ddb_entry_equal(old, new))
+				continue;
+
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] ddb (%4d - %4d) -> (%4d - %4d), size %4d -> %4d\n",
+				    plane->base.base.id, plane->base.name,
+				    old->start, old->end, new->start, new->end,
+				    skl_ddb_entry_size(old), skl_ddb_entry_size(new));
+		}
+
+		for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) {
+			enum plane_id plane_id = plane->id;
+			const struct skl_plane_wm *old_wm, *new_wm;
+
+			old_wm = &old_pipe_wm->planes[plane_id];
+			new_wm = &new_pipe_wm->planes[plane_id];
+
+			if (skl_plane_wm_equals(i915, old_wm, new_wm))
+				continue;
+
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s]   level %cwm0,%cwm1,%cwm2,%cwm3,%cwm4,%cwm5,%cwm6,%cwm7,%ctwm,%cswm,%cstwm"
+				    " -> %cwm0,%cwm1,%cwm2,%cwm3,%cwm4,%cwm5,%cwm6,%cwm7,%ctwm,%cswm,%cstwm\n",
+				    plane->base.base.id, plane->base.name,
+				    enast(old_wm->wm[0].enable), enast(old_wm->wm[1].enable),
+				    enast(old_wm->wm[2].enable), enast(old_wm->wm[3].enable),
+				    enast(old_wm->wm[4].enable), enast(old_wm->wm[5].enable),
+				    enast(old_wm->wm[6].enable), enast(old_wm->wm[7].enable),
+				    enast(old_wm->trans_wm.enable),
+				    enast(old_wm->sagv.wm0.enable),
+				    enast(old_wm->sagv.trans_wm.enable),
+				    enast(new_wm->wm[0].enable), enast(new_wm->wm[1].enable),
+				    enast(new_wm->wm[2].enable), enast(new_wm->wm[3].enable),
+				    enast(new_wm->wm[4].enable), enast(new_wm->wm[5].enable),
+				    enast(new_wm->wm[6].enable), enast(new_wm->wm[7].enable),
+				    enast(new_wm->trans_wm.enable),
+				    enast(new_wm->sagv.wm0.enable),
+				    enast(new_wm->sagv.trans_wm.enable));
+
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s]   lines %c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%4d"
+				      " -> %c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%3d,%c%4d\n",
+				    plane->base.base.id, plane->base.name,
+				    enast(old_wm->wm[0].ignore_lines), old_wm->wm[0].lines,
+				    enast(old_wm->wm[1].ignore_lines), old_wm->wm[1].lines,
+				    enast(old_wm->wm[2].ignore_lines), old_wm->wm[2].lines,
+				    enast(old_wm->wm[3].ignore_lines), old_wm->wm[3].lines,
+				    enast(old_wm->wm[4].ignore_lines), old_wm->wm[4].lines,
+				    enast(old_wm->wm[5].ignore_lines), old_wm->wm[5].lines,
+				    enast(old_wm->wm[6].ignore_lines), old_wm->wm[6].lines,
+				    enast(old_wm->wm[7].ignore_lines), old_wm->wm[7].lines,
+				    enast(old_wm->trans_wm.ignore_lines), old_wm->trans_wm.lines,
+				    enast(old_wm->sagv.wm0.ignore_lines), old_wm->sagv.wm0.lines,
+				    enast(old_wm->sagv.trans_wm.ignore_lines), old_wm->sagv.trans_wm.lines,
+				    enast(new_wm->wm[0].ignore_lines), new_wm->wm[0].lines,
+				    enast(new_wm->wm[1].ignore_lines), new_wm->wm[1].lines,
+				    enast(new_wm->wm[2].ignore_lines), new_wm->wm[2].lines,
+				    enast(new_wm->wm[3].ignore_lines), new_wm->wm[3].lines,
+				    enast(new_wm->wm[4].ignore_lines), new_wm->wm[4].lines,
+				    enast(new_wm->wm[5].ignore_lines), new_wm->wm[5].lines,
+				    enast(new_wm->wm[6].ignore_lines), new_wm->wm[6].lines,
+				    enast(new_wm->wm[7].ignore_lines), new_wm->wm[7].lines,
+				    enast(new_wm->trans_wm.ignore_lines), new_wm->trans_wm.lines,
+				    enast(new_wm->sagv.wm0.ignore_lines), new_wm->sagv.wm0.lines,
+				    enast(new_wm->sagv.trans_wm.ignore_lines), new_wm->sagv.trans_wm.lines);
+
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s]  blocks %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%5d"
+				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%5d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->wm[0].blocks, old_wm->wm[1].blocks,
+				    old_wm->wm[2].blocks, old_wm->wm[3].blocks,
+				    old_wm->wm[4].blocks, old_wm->wm[5].blocks,
+				    old_wm->wm[6].blocks, old_wm->wm[7].blocks,
+				    old_wm->trans_wm.blocks,
+				    old_wm->sagv.wm0.blocks,
+				    old_wm->sagv.trans_wm.blocks,
+				    new_wm->wm[0].blocks, new_wm->wm[1].blocks,
+				    new_wm->wm[2].blocks, new_wm->wm[3].blocks,
+				    new_wm->wm[4].blocks, new_wm->wm[5].blocks,
+				    new_wm->wm[6].blocks, new_wm->wm[7].blocks,
+				    new_wm->trans_wm.blocks,
+				    new_wm->sagv.wm0.blocks,
+				    new_wm->sagv.trans_wm.blocks);
+
+			drm_dbg_kms(&i915->drm,
+				    "[PLANE:%d:%s] min_ddb %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%5d"
+				    " -> %4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%4d,%5d\n",
+				    plane->base.base.id, plane->base.name,
+				    old_wm->wm[0].min_ddb_alloc, old_wm->wm[1].min_ddb_alloc,
+				    old_wm->wm[2].min_ddb_alloc, old_wm->wm[3].min_ddb_alloc,
+				    old_wm->wm[4].min_ddb_alloc, old_wm->wm[5].min_ddb_alloc,
+				    old_wm->wm[6].min_ddb_alloc, old_wm->wm[7].min_ddb_alloc,
+				    old_wm->trans_wm.min_ddb_alloc,
+				    old_wm->sagv.wm0.min_ddb_alloc,
+				    old_wm->sagv.trans_wm.min_ddb_alloc,
+				    new_wm->wm[0].min_ddb_alloc, new_wm->wm[1].min_ddb_alloc,
+				    new_wm->wm[2].min_ddb_alloc, new_wm->wm[3].min_ddb_alloc,
+				    new_wm->wm[4].min_ddb_alloc, new_wm->wm[5].min_ddb_alloc,
+				    new_wm->wm[6].min_ddb_alloc, new_wm->wm[7].min_ddb_alloc,
+				    new_wm->trans_wm.min_ddb_alloc,
+				    new_wm->sagv.wm0.min_ddb_alloc,
+				    new_wm->sagv.trans_wm.min_ddb_alloc);
+		}
+	}
+}
+
+static bool skl_plane_selected_wm_equals(struct intel_plane *plane,
+					 const struct skl_pipe_wm *old_pipe_wm,
+					 const struct skl_pipe_wm *new_pipe_wm)
+{
+	struct drm_i915_private *i915 = to_i915(plane->base.dev);
+	int level, max_level = ilk_wm_max_level(i915);
+
+	for (level = 0; level <= max_level; level++) {
+		/*
+		 * We don't check uv_wm as the hardware doesn't actually
+		 * use it. It only gets used for calculating the required
+		 * ddb allocation.
+		 */
+		if (!skl_wm_level_equals(skl_plane_wm_level(old_pipe_wm, plane->id, level),
+					 skl_plane_wm_level(new_pipe_wm, plane->id, level)))
+			return false;
+	}
+
+	if (HAS_HW_SAGV_WM(i915)) {
+		const struct skl_plane_wm *old_wm = &old_pipe_wm->planes[plane->id];
+		const struct skl_plane_wm *new_wm = &new_pipe_wm->planes[plane->id];
+
+		if (!skl_wm_level_equals(&old_wm->sagv.wm0, &new_wm->sagv.wm0) ||
+		    !skl_wm_level_equals(&old_wm->sagv.trans_wm, &new_wm->sagv.trans_wm))
+			return false;
+	}
+
+	return skl_wm_level_equals(skl_plane_trans_wm(old_pipe_wm, plane->id),
+				   skl_plane_trans_wm(new_pipe_wm, plane->id));
+}
+
+/*
+ * To make sure the cursor watermark registers are always consistent
+ * with our computed state the following scenario needs special
+ * treatment:
+ *
+ * 1. enable cursor
+ * 2. move cursor entirely offscreen
+ * 3. disable cursor
+ *
+ * Step 2. does call .disable_plane() but does not zero the watermarks
+ * (since we consider an offscreen cursor still active for the purposes
+ * of watermarks). Step 3. would not normally call .disable_plane()
+ * because the actual plane visibility isn't changing, and we don't
+ * deallocate the cursor ddb until the pipe gets disabled. So we must
+ * force step 3. to call .disable_plane() to update the watermark
+ * registers properly.
+ *
+ * Other planes do not suffer from this issues as their watermarks are
+ * calculated based on the actual plane visibility. The only time this
+ * can trigger for the other planes is during the initial readout as the
+ * default value of the watermarks registers is not zero.
+ */
+static int skl_wm_add_affected_planes(struct intel_atomic_state *state,
+				      struct intel_crtc *crtc)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_plane *plane;
+
+	for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) {
+		struct intel_plane_state *plane_state;
+		enum plane_id plane_id = plane->id;
+
+		/*
+		 * Force a full wm update for every plane on modeset.
+		 * Required because the reset value of the wm registers
+		 * is non-zero, whereas we want all disabled planes to
+		 * have zero watermarks. So if we turn off the relevant
+		 * power well the hardware state will go out of sync
+		 * with the software state.
+		 */
+		if (!drm_atomic_crtc_needs_modeset(&new_crtc_state->uapi) &&
+		    skl_plane_selected_wm_equals(plane,
+						 &old_crtc_state->wm.skl.optimal,
+						 &new_crtc_state->wm.skl.optimal))
+			continue;
+
+		plane_state = intel_atomic_get_plane_state(state, plane);
+		if (IS_ERR(plane_state))
+			return PTR_ERR(plane_state);
+
+		new_crtc_state->update_planes |= BIT(plane_id);
+	}
+
+	return 0;
+}
+
+static int
+skl_compute_wm(struct intel_atomic_state *state)
+{
+	struct intel_crtc *crtc;
+	struct intel_crtc_state *new_crtc_state;
+	int ret, i;
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		ret = skl_build_pipe_wm(state, crtc);
+		if (ret)
+			return ret;
+	}
+
+	ret = skl_compute_ddb(state);
+	if (ret)
+		return ret;
+
+	ret = intel_compute_sagv_mask(state);
+	if (ret)
+		return ret;
+
+	/*
+	 * skl_compute_ddb() will have adjusted the final watermarks
+	 * based on how much ddb is available. Now we can actually
+	 * check if the final watermarks changed.
+	 */
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		ret = skl_wm_add_affected_planes(state, crtc);
+		if (ret)
+			return ret;
+	}
+
+	skl_print_wm_changes(state);
+
+	return 0;
+}
+
+static void skl_wm_level_from_reg_val(u32 val, struct skl_wm_level *level)
+{
+	level->enable = val & PLANE_WM_EN;
+	level->ignore_lines = val & PLANE_WM_IGNORE_LINES;
+	level->blocks = REG_FIELD_GET(PLANE_WM_BLOCKS_MASK, val);
+	level->lines = REG_FIELD_GET(PLANE_WM_LINES_MASK, val);
+}
+
+static void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc,
+				     struct skl_pipe_wm *out)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	int level, max_level;
+	enum plane_id plane_id;
+	u32 val;
+
+	max_level = ilk_wm_max_level(i915);
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		struct skl_plane_wm *wm = &out->planes[plane_id];
+
+		for (level = 0; level <= max_level; level++) {
+			if (plane_id != PLANE_CURSOR)
+				val = intel_uncore_read(&i915->uncore, PLANE_WM(pipe, plane_id, level));
+			else
+				val = intel_uncore_read(&i915->uncore, CUR_WM(pipe, level));
+
+			skl_wm_level_from_reg_val(val, &wm->wm[level]);
+		}
+
+		if (plane_id != PLANE_CURSOR)
+			val = intel_uncore_read(&i915->uncore, PLANE_WM_TRANS(pipe, plane_id));
+		else
+			val = intel_uncore_read(&i915->uncore, CUR_WM_TRANS(pipe));
+
+		skl_wm_level_from_reg_val(val, &wm->trans_wm);
+
+		if (HAS_HW_SAGV_WM(i915)) {
+			if (plane_id != PLANE_CURSOR)
+				val = intel_uncore_read(&i915->uncore,
+							PLANE_WM_SAGV(pipe, plane_id));
+			else
+				val = intel_uncore_read(&i915->uncore,
+							CUR_WM_SAGV(pipe));
+
+			skl_wm_level_from_reg_val(val, &wm->sagv.wm0);
+
+			if (plane_id != PLANE_CURSOR)
+				val = intel_uncore_read(&i915->uncore,
+							PLANE_WM_SAGV_TRANS(pipe, plane_id));
+			else
+				val = intel_uncore_read(&i915->uncore,
+							CUR_WM_SAGV_TRANS(pipe));
+
+			skl_wm_level_from_reg_val(val, &wm->sagv.trans_wm);
+		} else if (DISPLAY_VER(i915) >= 12) {
+			wm->sagv.wm0 = wm->wm[0];
+			wm->sagv.trans_wm = wm->trans_wm;
+		}
+	}
+}
+
+void skl_wm_get_hw_state(struct drm_i915_private *i915)
+{
+	struct intel_dbuf_state *dbuf_state =
+		to_intel_dbuf_state(i915->display.dbuf.obj.state);
+	struct intel_crtc *crtc;
+
+	if (HAS_MBUS_JOINING(i915))
+		dbuf_state->joined_mbus = intel_de_read(i915, MBUS_CTL) & MBUS_JOIN;
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		enum pipe pipe = crtc->pipe;
+		unsigned int mbus_offset;
+		enum plane_id plane_id;
+		u8 slices;
+
+		memset(&crtc_state->wm.skl.optimal, 0,
+		       sizeof(crtc_state->wm.skl.optimal));
+		if (crtc_state->hw.active)
+			skl_pipe_wm_get_hw_state(crtc, &crtc_state->wm.skl.optimal);
+		crtc_state->wm.skl.raw = crtc_state->wm.skl.optimal;
+
+		memset(&dbuf_state->ddb[pipe], 0, sizeof(dbuf_state->ddb[pipe]));
+
+		for_each_plane_id_on_crtc(crtc, plane_id) {
+			struct skl_ddb_entry *ddb =
+				&crtc_state->wm.skl.plane_ddb[plane_id];
+			struct skl_ddb_entry *ddb_y =
+				&crtc_state->wm.skl.plane_ddb_y[plane_id];
+
+			if (!crtc_state->hw.active)
+				continue;
+
+			skl_ddb_get_hw_plane_state(i915, crtc->pipe,
+						   plane_id, ddb, ddb_y);
+
+			skl_ddb_entry_union(&dbuf_state->ddb[pipe], ddb);
+			skl_ddb_entry_union(&dbuf_state->ddb[pipe], ddb_y);
+		}
+
+		dbuf_state->weight[pipe] = intel_crtc_ddb_weight(crtc_state);
+
+		/*
+		 * Used for checking overlaps, so we need absolute
+		 * offsets instead of MBUS relative offsets.
+		 */
+		slices = skl_compute_dbuf_slices(crtc, dbuf_state->active_pipes,
+						 dbuf_state->joined_mbus);
+		mbus_offset = mbus_ddb_offset(i915, slices);
+		crtc_state->wm.skl.ddb.start = mbus_offset + dbuf_state->ddb[pipe].start;
+		crtc_state->wm.skl.ddb.end = mbus_offset + dbuf_state->ddb[pipe].end;
+
+		/* The slices actually used by the planes on the pipe */
+		dbuf_state->slices[pipe] =
+			skl_ddb_dbuf_slice_mask(i915, &crtc_state->wm.skl.ddb);
+
+		drm_dbg_kms(&i915->drm,
+			    "[CRTC:%d:%s] dbuf slices 0x%x, ddb (%d - %d), active pipes 0x%x, mbus joined: %s\n",
+			    crtc->base.base.id, crtc->base.name,
+			    dbuf_state->slices[pipe], dbuf_state->ddb[pipe].start,
+			    dbuf_state->ddb[pipe].end, dbuf_state->active_pipes,
+			    str_yes_no(dbuf_state->joined_mbus));
+	}
+
+	dbuf_state->enabled_slices = i915->display.dbuf.enabled_slices;
+}
+
+static bool skl_dbuf_is_misconfigured(struct drm_i915_private *i915)
+{
+	const struct intel_dbuf_state *dbuf_state =
+		to_intel_dbuf_state(i915->display.dbuf.obj.state);
+	struct skl_ddb_entry entries[I915_MAX_PIPES] = {};
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		const struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		entries[crtc->pipe] = crtc_state->wm.skl.ddb;
+	}
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		const struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		u8 slices;
+
+		slices = skl_compute_dbuf_slices(crtc, dbuf_state->active_pipes,
+						 dbuf_state->joined_mbus);
+		if (dbuf_state->slices[crtc->pipe] & ~slices)
+			return true;
+
+		if (skl_ddb_allocation_overlaps(&crtc_state->wm.skl.ddb, entries,
+						I915_MAX_PIPES, crtc->pipe))
+			return true;
+	}
+
+	return false;
+}
+
+void skl_wm_sanitize(struct drm_i915_private *i915)
+{
+	struct intel_crtc *crtc;
+
+	/*
+	 * On TGL/RKL (at least) the BIOS likes to assign the planes
+	 * to the wrong DBUF slices. This will cause an infinite loop
+	 * in skl_commit_modeset_enables() as it can't find a way to
+	 * transition between the old bogus DBUF layout to the new
+	 * proper DBUF layout without DBUF allocation overlaps between
+	 * the planes (which cannot be allowed or else the hardware
+	 * may hang). If we detect a bogus DBUF layout just turn off
+	 * all the planes so that skl_commit_modeset_enables() can
+	 * simply ignore them.
+	 */
+	if (!skl_dbuf_is_misconfigured(i915))
+		return;
+
+	drm_dbg_kms(&i915->drm, "BIOS has misprogrammed the DBUF, disabling all planes\n");
+
+	for_each_intel_crtc(&i915->drm, crtc) {
+		struct intel_plane *plane = to_intel_plane(crtc->base.primary);
+		const struct intel_plane_state *plane_state =
+			to_intel_plane_state(plane->base.state);
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		if (plane_state->uapi.visible)
+			intel_plane_disable_noatomic(crtc, plane);
+
+		drm_WARN_ON(&i915->drm, crtc_state->active_planes != 0);
+
+		memset(&crtc_state->wm.skl.ddb, 0, sizeof(crtc_state->wm.skl.ddb));
+	}
+}
+
+void intel_wm_state_verify(struct intel_crtc *crtc,
+			   struct intel_crtc_state *new_crtc_state)
+{
+	struct drm_i915_private *i915 = to_i915(crtc->base.dev);
+	struct skl_hw_state {
+		struct skl_ddb_entry ddb[I915_MAX_PLANES];
+		struct skl_ddb_entry ddb_y[I915_MAX_PLANES];
+		struct skl_pipe_wm wm;
+	} *hw;
+	const struct skl_pipe_wm *sw_wm = &new_crtc_state->wm.skl.optimal;
+	int level, max_level = ilk_wm_max_level(i915);
+	struct intel_plane *plane;
+	u8 hw_enabled_slices;
+
+	if (DISPLAY_VER(i915) < 9 || !new_crtc_state->hw.active)
+		return;
+
+	hw = kzalloc(sizeof(*hw), GFP_KERNEL);
+	if (!hw)
+		return;
+
+	skl_pipe_wm_get_hw_state(crtc, &hw->wm);
+
+	skl_pipe_ddb_get_hw_state(crtc, hw->ddb, hw->ddb_y);
+
+	hw_enabled_slices = intel_enabled_dbuf_slices_mask(i915);
+
+	if (DISPLAY_VER(i915) >= 11 &&
+	    hw_enabled_slices != i915->display.dbuf.enabled_slices)
+		drm_err(&i915->drm,
+			"mismatch in DBUF Slices (expected 0x%x, got 0x%x)\n",
+			i915->display.dbuf.enabled_slices,
+			hw_enabled_slices);
+
+	for_each_intel_plane_on_crtc(&i915->drm, crtc, plane) {
+		const struct skl_ddb_entry *hw_ddb_entry, *sw_ddb_entry;
+		const struct skl_wm_level *hw_wm_level, *sw_wm_level;
+
+		/* Watermarks */
+		for (level = 0; level <= max_level; level++) {
+			hw_wm_level = &hw->wm.planes[plane->id].wm[level];
+			sw_wm_level = skl_plane_wm_level(sw_wm, plane->id, level);
+
+			if (skl_wm_level_equals(hw_wm_level, sw_wm_level))
+				continue;
+
+			drm_err(&i915->drm,
+				"[PLANE:%d:%s] mismatch in WM%d (expected e=%d b=%u l=%u, got e=%d b=%u l=%u)\n",
+				plane->base.base.id, plane->base.name, level,
+				sw_wm_level->enable,
+				sw_wm_level->blocks,
+				sw_wm_level->lines,
+				hw_wm_level->enable,
+				hw_wm_level->blocks,
+				hw_wm_level->lines);
+		}
+
+		hw_wm_level = &hw->wm.planes[plane->id].trans_wm;
+		sw_wm_level = skl_plane_trans_wm(sw_wm, plane->id);
+
+		if (!skl_wm_level_equals(hw_wm_level, sw_wm_level)) {
+			drm_err(&i915->drm,
+				"[PLANE:%d:%s] mismatch in trans WM (expected e=%d b=%u l=%u, got e=%d b=%u l=%u)\n",
+				plane->base.base.id, plane->base.name,
+				sw_wm_level->enable,
+				sw_wm_level->blocks,
+				sw_wm_level->lines,
+				hw_wm_level->enable,
+				hw_wm_level->blocks,
+				hw_wm_level->lines);
+		}
+
+		hw_wm_level = &hw->wm.planes[plane->id].sagv.wm0;
+		sw_wm_level = &sw_wm->planes[plane->id].sagv.wm0;
+
+		if (HAS_HW_SAGV_WM(i915) &&
+		    !skl_wm_level_equals(hw_wm_level, sw_wm_level)) {
+			drm_err(&i915->drm,
+				"[PLANE:%d:%s] mismatch in SAGV WM (expected e=%d b=%u l=%u, got e=%d b=%u l=%u)\n",
+				plane->base.base.id, plane->base.name,
+				sw_wm_level->enable,
+				sw_wm_level->blocks,
+				sw_wm_level->lines,
+				hw_wm_level->enable,
+				hw_wm_level->blocks,
+				hw_wm_level->lines);
+		}
+
+		hw_wm_level = &hw->wm.planes[plane->id].sagv.trans_wm;
+		sw_wm_level = &sw_wm->planes[plane->id].sagv.trans_wm;
+
+		if (HAS_HW_SAGV_WM(i915) &&
+		    !skl_wm_level_equals(hw_wm_level, sw_wm_level)) {
+			drm_err(&i915->drm,
+				"[PLANE:%d:%s] mismatch in SAGV trans WM (expected e=%d b=%u l=%u, got e=%d b=%u l=%u)\n",
+				plane->base.base.id, plane->base.name,
+				sw_wm_level->enable,
+				sw_wm_level->blocks,
+				sw_wm_level->lines,
+				hw_wm_level->enable,
+				hw_wm_level->blocks,
+				hw_wm_level->lines);
+		}
+
+		/* DDB */
+		hw_ddb_entry = &hw->ddb[PLANE_CURSOR];
+		sw_ddb_entry = &new_crtc_state->wm.skl.plane_ddb[PLANE_CURSOR];
+
+		if (!skl_ddb_entry_equal(hw_ddb_entry, sw_ddb_entry)) {
+			drm_err(&i915->drm,
+				"[PLANE:%d:%s] mismatch in DDB (expected (%u,%u), found (%u,%u))\n",
+				plane->base.base.id, plane->base.name,
+				sw_ddb_entry->start, sw_ddb_entry->end,
+				hw_ddb_entry->start, hw_ddb_entry->end);
+		}
+	}
+
+	kfree(hw);
+}
+
+bool skl_watermark_ipc_enabled(struct drm_i915_private *i915)
+{
+	return i915->display.wm.ipc_enabled;
+}
+
+void skl_watermark_ipc_update(struct drm_i915_private *i915)
+{
+	if (!HAS_IPC(i915))
+		return;
+
+	intel_uncore_rmw(&i915->uncore, DISP_ARB_CTL2, DISP_IPC_ENABLE,
+			 skl_watermark_ipc_enabled(i915) ? DISP_IPC_ENABLE : 0);
+}
+
+static bool skl_watermark_ipc_can_enable(struct drm_i915_private *i915)
+{
+	/* Display WA #0477 WaDisableIPC: skl */
+	if (IS_SKYLAKE(i915))
+		return false;
+
+	/* Display WA #1141: SKL:all KBL:all CFL */
+	if (IS_KABYLAKE(i915) ||
+	    IS_COFFEELAKE(i915) ||
+	    IS_COMETLAKE(i915))
+		return i915->dram_info.symmetric_memory;
+
+	return true;
+}
+
+void skl_watermark_ipc_init(struct drm_i915_private *i915)
+{
+	if (!HAS_IPC(i915))
+		return;
+
+	i915->display.wm.ipc_enabled = skl_watermark_ipc_can_enable(i915);
+
+	skl_watermark_ipc_update(i915);
+}
+
+static void
+adjust_wm_latency(struct drm_i915_private *i915,
+		  u16 wm[], int max_level, int read_latency)
+{
+	bool wm_lv_0_adjust_needed = i915->dram_info.wm_lv_0_adjust_needed;
+	int i, level;
+
+	/*
+	 * If a level n (n > 1) has a 0us latency, all levels m (m >= n)
+	 * need to be disabled. We make sure to sanitize the values out
+	 * of the punit to satisfy this requirement.
+	 */
+	for (level = 1; level <= max_level; level++) {
+		if (wm[level] == 0) {
+			for (i = level + 1; i <= max_level; i++)
+				wm[i] = 0;
+
+			max_level = level - 1;
+			break;
+		}
+	}
+
+	/*
+	 * WaWmMemoryReadLatency
+	 *
+	 * punit doesn't take into account the read latency so we need
+	 * to add proper adjustement to each valid level we retrieve
+	 * from the punit when level 0 response data is 0us.
+	 */
+	if (wm[0] == 0) {
+		for (level = 0; level <= max_level; level++)
+			wm[level] += read_latency;
+	}
+
+	/*
+	 * WA Level-0 adjustment for 16GB DIMMs: SKL+
+	 * If we could not get dimm info enable this WA to prevent from
+	 * any underrun. If not able to get Dimm info assume 16GB dimm
+	 * to avoid any underrun.
+	 */
+	if (wm_lv_0_adjust_needed)
+		wm[0] += 1;
+}
+
+static void mtl_read_wm_latency(struct drm_i915_private *i915, u16 wm[])
+{
+	struct intel_uncore *uncore = &i915->uncore;
+	int max_level = ilk_wm_max_level(i915);
+	u32 val;
+
+	val = intel_uncore_read(uncore, MTL_LATENCY_LP0_LP1);
+	wm[0] = REG_FIELD_GET(MTL_LATENCY_LEVEL_EVEN_MASK, val);
+	wm[1] = REG_FIELD_GET(MTL_LATENCY_LEVEL_ODD_MASK, val);
+
+	val = intel_uncore_read(uncore, MTL_LATENCY_LP2_LP3);
+	wm[2] = REG_FIELD_GET(MTL_LATENCY_LEVEL_EVEN_MASK, val);
+	wm[3] = REG_FIELD_GET(MTL_LATENCY_LEVEL_ODD_MASK, val);
+
+	val = intel_uncore_read(uncore, MTL_LATENCY_LP4_LP5);
+	wm[4] = REG_FIELD_GET(MTL_LATENCY_LEVEL_EVEN_MASK, val);
+	wm[5] = REG_FIELD_GET(MTL_LATENCY_LEVEL_ODD_MASK, val);
+
+	adjust_wm_latency(i915, wm, max_level, 6);
+}
+
+static void skl_read_wm_latency(struct drm_i915_private *i915, u16 wm[])
+{
+	int max_level = ilk_wm_max_level(i915);
+	int read_latency = DISPLAY_VER(i915) >= 12 ? 3 : 2;
+	int mult = IS_DG2(i915) ? 2 : 1;
+	u32 val;
+	int ret;
+
+	/* read the first set of memory latencies[0:3] */
+	val = 0; /* data0 to be programmed to 0 for first set */
+	ret = snb_pcode_read(&i915->uncore, GEN9_PCODE_READ_MEM_LATENCY, &val, NULL);
+	if (ret) {
+		drm_err(&i915->drm, "SKL Mailbox read error = %d\n", ret);
+		return;
+	}
+
+	wm[0] = REG_FIELD_GET(GEN9_MEM_LATENCY_LEVEL_0_4_MASK, val) * mult;
+	wm[1] = REG_FIELD_GET(GEN9_MEM_LATENCY_LEVEL_1_5_MASK, val) * mult;
+	wm[2] = REG_FIELD_GET(GEN9_MEM_LATENCY_LEVEL_2_6_MASK, val) * mult;
+	wm[3] = REG_FIELD_GET(GEN9_MEM_LATENCY_LEVEL_3_7_MASK, val) * mult;
+
+	/* read the second set of memory latencies[4:7] */
+	val = 1; /* data0 to be programmed to 1 for second set */
+	ret = snb_pcode_read(&i915->uncore, GEN9_PCODE_READ_MEM_LATENCY, &val, NULL);
+	if (ret) {
+		drm_err(&i915->drm, "SKL Mailbox read error = %d\n", ret);
+		return;
+	}
+
+	wm[4] = REG_FIELD_GET(GEN9_MEM_LATENCY_LEVEL_0_4_MASK, val) * mult;
+	wm[5] = REG_FIELD_GET(GEN9_MEM_LATENCY_LEVEL_1_5_MASK, val) * mult;
+	wm[6] = REG_FIELD_GET(GEN9_MEM_LATENCY_LEVEL_2_6_MASK, val) * mult;
+	wm[7] = REG_FIELD_GET(GEN9_MEM_LATENCY_LEVEL_3_7_MASK, val) * mult;
+
+	adjust_wm_latency(i915, wm, max_level, read_latency);
+}
+
+static void skl_setup_wm_latency(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 14)
+		mtl_read_wm_latency(i915, i915->display.wm.skl_latency);
+	else
+		skl_read_wm_latency(i915, i915->display.wm.skl_latency);
+
+	intel_print_wm_latency(i915, "Gen9 Plane", i915->display.wm.skl_latency);
+}
+
+static const struct intel_wm_funcs skl_wm_funcs = {
+	.compute_global_watermarks = skl_compute_wm,
+};
+
+void skl_wm_init(struct drm_i915_private *i915)
+{
+	intel_sagv_init(i915);
+
+	skl_setup_wm_latency(i915);
+
+	i915->display.funcs.wm = &skl_wm_funcs;
+}
+
+static struct intel_global_state *intel_dbuf_duplicate_state(struct intel_global_obj *obj)
+{
+	struct intel_dbuf_state *dbuf_state;
+
+	dbuf_state = kmemdup(obj->state, sizeof(*dbuf_state), GFP_KERNEL);
+	if (!dbuf_state)
+		return NULL;
+
+	return &dbuf_state->base;
+}
+
+static void intel_dbuf_destroy_state(struct intel_global_obj *obj,
+				     struct intel_global_state *state)
+{
+	kfree(state);
+}
+
+static const struct intel_global_state_funcs intel_dbuf_funcs = {
+	.atomic_duplicate_state = intel_dbuf_duplicate_state,
+	.atomic_destroy_state = intel_dbuf_destroy_state,
+};
+
+struct intel_dbuf_state *
+intel_atomic_get_dbuf_state(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	struct intel_global_state *dbuf_state;
+
+	dbuf_state = intel_atomic_get_global_obj_state(state, &i915->display.dbuf.obj);
+	if (IS_ERR(dbuf_state))
+		return ERR_CAST(dbuf_state);
+
+	return to_intel_dbuf_state(dbuf_state);
+}
+
+int intel_dbuf_init(struct drm_i915_private *i915)
+{
+	struct intel_dbuf_state *dbuf_state;
+
+	dbuf_state = kzalloc(sizeof(*dbuf_state), GFP_KERNEL);
+	if (!dbuf_state)
+		return -ENOMEM;
+
+	intel_atomic_global_obj_init(i915, &i915->display.dbuf.obj,
+				     &dbuf_state->base, &intel_dbuf_funcs);
+
+	return 0;
+}
+
+/*
+ * Configure MBUS_CTL and all DBUF_CTL_S of each slice to join_mbus state before
+ * update the request state of all DBUS slices.
+ */
+static void update_mbus_pre_enable(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	u32 mbus_ctl, dbuf_min_tracker_val;
+	enum dbuf_slice slice;
+	const struct intel_dbuf_state *dbuf_state =
+		intel_atomic_get_new_dbuf_state(state);
+
+	if (!HAS_MBUS_JOINING(i915))
+		return;
+
+	/*
+	 * TODO: Implement vblank synchronized MBUS joining changes.
+	 * Must be properly coordinated with dbuf reprogramming.
+	 */
+	if (dbuf_state->joined_mbus) {
+		mbus_ctl = MBUS_HASHING_MODE_1x4 | MBUS_JOIN |
+			MBUS_JOIN_PIPE_SELECT_NONE;
+		dbuf_min_tracker_val = DBUF_MIN_TRACKER_STATE_SERVICE(3);
+	} else {
+		mbus_ctl = MBUS_HASHING_MODE_2x2 |
+			MBUS_JOIN_PIPE_SELECT_NONE;
+		dbuf_min_tracker_val = DBUF_MIN_TRACKER_STATE_SERVICE(1);
+	}
+
+	intel_de_rmw(i915, MBUS_CTL,
+		     MBUS_HASHING_MODE_MASK | MBUS_JOIN |
+		     MBUS_JOIN_PIPE_SELECT_MASK, mbus_ctl);
+
+	for_each_dbuf_slice(i915, slice)
+		intel_de_rmw(i915, DBUF_CTL_S(slice),
+			     DBUF_MIN_TRACKER_STATE_SERVICE_MASK,
+			     dbuf_min_tracker_val);
+}
+
+void intel_dbuf_pre_plane_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_dbuf_state *new_dbuf_state =
+		intel_atomic_get_new_dbuf_state(state);
+	const struct intel_dbuf_state *old_dbuf_state =
+		intel_atomic_get_old_dbuf_state(state);
+
+	if (!new_dbuf_state ||
+	    (new_dbuf_state->enabled_slices == old_dbuf_state->enabled_slices &&
+	     new_dbuf_state->joined_mbus == old_dbuf_state->joined_mbus))
+		return;
+
+	WARN_ON(!new_dbuf_state->base.changed);
+
+	update_mbus_pre_enable(state);
+	gen9_dbuf_slices_update(i915,
+				old_dbuf_state->enabled_slices |
+				new_dbuf_state->enabled_slices);
+}
+
+void intel_dbuf_post_plane_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_dbuf_state *new_dbuf_state =
+		intel_atomic_get_new_dbuf_state(state);
+	const struct intel_dbuf_state *old_dbuf_state =
+		intel_atomic_get_old_dbuf_state(state);
+
+	if (!new_dbuf_state ||
+	    (new_dbuf_state->enabled_slices == old_dbuf_state->enabled_slices &&
+	     new_dbuf_state->joined_mbus == old_dbuf_state->joined_mbus))
+		return;
+
+	WARN_ON(!new_dbuf_state->base.changed);
+
+	gen9_dbuf_slices_update(i915,
+				new_dbuf_state->enabled_slices);
+}
+
+static bool xelpdp_is_only_pipe_per_dbuf_bank(enum pipe pipe, u8 active_pipes)
+{
+	switch (pipe) {
+	case PIPE_A:
+		return !(active_pipes & BIT(PIPE_D));
+	case PIPE_D:
+		return !(active_pipes & BIT(PIPE_A));
+	case PIPE_B:
+		return !(active_pipes & BIT(PIPE_C));
+	case PIPE_C:
+		return !(active_pipes & BIT(PIPE_B));
+	default: /* to suppress compiler warning */
+		MISSING_CASE(pipe);
+		break;
+	}
+
+	return false;
+}
+
+void intel_mbus_dbox_update(struct intel_atomic_state *state)
+{
+	struct drm_i915_private *i915 = to_i915(state->base.dev);
+	const struct intel_dbuf_state *new_dbuf_state, *old_dbuf_state;
+	const struct intel_crtc_state *new_crtc_state;
+	const struct intel_crtc *crtc;
+	u32 val = 0;
+	int i;
+
+	if (DISPLAY_VER(i915) < 11)
+		return;
+
+	new_dbuf_state = intel_atomic_get_new_dbuf_state(state);
+	old_dbuf_state = intel_atomic_get_old_dbuf_state(state);
+	if (!new_dbuf_state ||
+	    (new_dbuf_state->joined_mbus == old_dbuf_state->joined_mbus &&
+	     new_dbuf_state->active_pipes == old_dbuf_state->active_pipes))
+		return;
+
+	if (DISPLAY_VER(i915) >= 14)
+		val |= MBUS_DBOX_I_CREDIT(2);
+
+	if (DISPLAY_VER(i915) >= 12) {
+		val |= MBUS_DBOX_B2B_TRANSACTIONS_MAX(16);
+		val |= MBUS_DBOX_B2B_TRANSACTIONS_DELAY(1);
+		val |= MBUS_DBOX_REGULATE_B2B_TRANSACTIONS_EN;
+	}
+
+	if (DISPLAY_VER(i915) >= 14)
+		val |= new_dbuf_state->joined_mbus ? MBUS_DBOX_A_CREDIT(12) :
+						     MBUS_DBOX_A_CREDIT(8);
+	else if (IS_ALDERLAKE_P(i915))
+		/* Wa_22010947358:adl-p */
+		val |= new_dbuf_state->joined_mbus ? MBUS_DBOX_A_CREDIT(6) :
+						     MBUS_DBOX_A_CREDIT(4);
+	else
+		val |= MBUS_DBOX_A_CREDIT(2);
+
+	if (DISPLAY_VER(i915) >= 14) {
+		val |= MBUS_DBOX_B_CREDIT(0xA);
+	} else if (IS_ALDERLAKE_P(i915)) {
+		val |= MBUS_DBOX_BW_CREDIT(2);
+		val |= MBUS_DBOX_B_CREDIT(8);
+	} else if (DISPLAY_VER(i915) >= 12) {
+		val |= MBUS_DBOX_BW_CREDIT(2);
+		val |= MBUS_DBOX_B_CREDIT(12);
+	} else {
+		val |= MBUS_DBOX_BW_CREDIT(1);
+		val |= MBUS_DBOX_B_CREDIT(8);
+	}
+
+	for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
+		u32 pipe_val = val;
+
+		if (!new_crtc_state->hw.active)
+			continue;
+
+		if (DISPLAY_VER(i915) >= 14) {
+			if (xelpdp_is_only_pipe_per_dbuf_bank(crtc->pipe,
+							      new_dbuf_state->active_pipes))
+				pipe_val |= MBUS_DBOX_BW_8CREDITS_MTL;
+			else
+				pipe_val |= MBUS_DBOX_BW_4CREDITS_MTL;
+		}
+
+		intel_de_write(i915, PIPE_MBUS_DBOX_CTL(crtc->pipe), pipe_val);
+	}
+}
+
+static int skl_watermark_ipc_status_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *i915 = m->private;
+
+	seq_printf(m, "Isochronous Priority Control: %s\n",
+		   str_yes_no(skl_watermark_ipc_enabled(i915)));
+	return 0;
+}
+
+static int skl_watermark_ipc_status_open(struct inode *inode, struct file *file)
+{
+	struct drm_i915_private *i915 = inode->i_private;
+
+	return single_open(file, skl_watermark_ipc_status_show, i915);
+}
+
+static ssize_t skl_watermark_ipc_status_write(struct file *file,
+					      const char __user *ubuf,
+					      size_t len, loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct drm_i915_private *i915 = m->private;
+	intel_wakeref_t wakeref;
+	bool enable;
+	int ret;
+
+	ret = kstrtobool_from_user(ubuf, len, &enable);
+	if (ret < 0)
+		return ret;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		if (!skl_watermark_ipc_enabled(i915) && enable)
+			drm_info(&i915->drm,
+				 "Enabling IPC: WM will be proper only after next commit\n");
+		i915->display.wm.ipc_enabled = enable;
+		skl_watermark_ipc_update(i915);
+	}
+
+	return len;
+}
+
+static const struct file_operations skl_watermark_ipc_status_fops = {
+	.owner = THIS_MODULE,
+	.open = skl_watermark_ipc_status_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = skl_watermark_ipc_status_write
+};
+
+void skl_watermark_ipc_debugfs_register(struct drm_i915_private *i915)
+{
+	struct drm_minor *minor = i915->drm.primary;
+
+	if (!HAS_IPC(i915))
+		return;
+
+	debugfs_create_file("i915_ipc_status", 0644, minor->debugfs_root, i915,
+			    &skl_watermark_ipc_status_fops);
+}
diff --git a/drivers/gpu/drm/i915/display/skl_watermark.h b/drivers/gpu/drm/i915/display/skl_watermark.h
new file mode 100644
index 000000000..7a5a4e67c
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/skl_watermark.h
@@ -0,0 +1,80 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __SKL_WATERMARK_H__
+#define __SKL_WATERMARK_H__
+
+#include <linux/types.h>
+
+#include "intel_display.h"
+#include "intel_global_state.h"
+#include "intel_pm_types.h"
+
+struct drm_i915_private;
+struct intel_atomic_state;
+struct intel_bw_state;
+struct intel_crtc;
+struct intel_crtc_state;
+struct intel_plane;
+
+u8 intel_enabled_dbuf_slices_mask(struct drm_i915_private *i915);
+
+void intel_sagv_pre_plane_update(struct intel_atomic_state *state);
+void intel_sagv_post_plane_update(struct intel_atomic_state *state);
+bool intel_can_enable_sagv(struct drm_i915_private *i915,
+			   const struct intel_bw_state *bw_state);
+
+u32 skl_ddb_dbuf_slice_mask(struct drm_i915_private *i915,
+			    const struct skl_ddb_entry *entry);
+
+void skl_write_plane_wm(struct intel_plane *plane,
+			const struct intel_crtc_state *crtc_state);
+void skl_write_cursor_wm(struct intel_plane *plane,
+			 const struct intel_crtc_state *crtc_state);
+
+bool skl_ddb_allocation_overlaps(const struct skl_ddb_entry *ddb,
+				 const struct skl_ddb_entry *entries,
+				 int num_entries, int ignore_idx);
+
+void skl_wm_get_hw_state(struct drm_i915_private *i915);
+void skl_wm_sanitize(struct drm_i915_private *i915);
+
+void intel_wm_state_verify(struct intel_crtc *crtc,
+			   struct intel_crtc_state *new_crtc_state);
+
+void skl_watermark_ipc_init(struct drm_i915_private *i915);
+void skl_watermark_ipc_update(struct drm_i915_private *i915);
+bool skl_watermark_ipc_enabled(struct drm_i915_private *i915);
+void skl_watermark_ipc_debugfs_register(struct drm_i915_private *i915);
+
+void skl_wm_init(struct drm_i915_private *i915);
+
+struct intel_dbuf_state {
+	struct intel_global_state base;
+
+	struct skl_ddb_entry ddb[I915_MAX_PIPES];
+	unsigned int weight[I915_MAX_PIPES];
+	u8 slices[I915_MAX_PIPES];
+	u8 enabled_slices;
+	u8 active_pipes;
+	bool joined_mbus;
+};
+
+struct intel_dbuf_state *
+intel_atomic_get_dbuf_state(struct intel_atomic_state *state);
+
+#define to_intel_dbuf_state(x) container_of((x), struct intel_dbuf_state, base)
+#define intel_atomic_get_old_dbuf_state(state) \
+	to_intel_dbuf_state(intel_atomic_get_old_global_obj_state(state, &to_i915(state->base.dev)->display.dbuf.obj))
+#define intel_atomic_get_new_dbuf_state(state) \
+	to_intel_dbuf_state(intel_atomic_get_new_global_obj_state(state, &to_i915(state->base.dev)->display.dbuf.obj))
+
+int intel_dbuf_init(struct drm_i915_private *i915);
+void intel_dbuf_pre_plane_update(struct intel_atomic_state *state);
+void intel_dbuf_post_plane_update(struct intel_atomic_state *state);
+void intel_mbus_dbox_update(struct intel_atomic_state *state);
+
+#endif /* __SKL_WATERMARK_H__ */
+
diff --git a/drivers/gpu/drm/i915/display/vlv_dsi.c b/drivers/gpu/drm/i915/display/vlv_dsi.c
new file mode 100644
index 000000000..114088ca5
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/vlv_dsi.c
@@ -0,0 +1,2015 @@
+/*
+ * Copyright © 2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Author: Jani Nikula <jani.nikula@intel.com>
+ */
+
+#include <linux/slab.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_mipi_dsi.h>
+
+#include "i915_drv.h"
+#include "intel_atomic.h"
+#include "intel_backlight.h"
+#include "intel_connector.h"
+#include "intel_crtc.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dsi.h"
+#include "intel_dsi_vbt.h"
+#include "intel_fifo_underrun.h"
+#include "intel_panel.h"
+#include "skl_scaler.h"
+#include "vlv_dsi.h"
+#include "vlv_dsi_pll.h"
+#include "vlv_dsi_regs.h"
+#include "vlv_sideband.h"
+
+/* return pixels in terms of txbyteclkhs */
+static u16 txbyteclkhs(u16 pixels, int bpp, int lane_count,
+		       u16 burst_mode_ratio)
+{
+	return DIV_ROUND_UP(DIV_ROUND_UP(pixels * bpp * burst_mode_ratio,
+					 8 * 100), lane_count);
+}
+
+/* return pixels equvalent to txbyteclkhs */
+static u16 pixels_from_txbyteclkhs(u16 clk_hs, int bpp, int lane_count,
+			u16 burst_mode_ratio)
+{
+	return DIV_ROUND_UP((clk_hs * lane_count * 8 * 100),
+						(bpp * burst_mode_ratio));
+}
+
+enum mipi_dsi_pixel_format pixel_format_from_register_bits(u32 fmt)
+{
+	/* It just so happens the VBT matches register contents. */
+	switch (fmt) {
+	case VID_MODE_FORMAT_RGB888:
+		return MIPI_DSI_FMT_RGB888;
+	case VID_MODE_FORMAT_RGB666:
+		return MIPI_DSI_FMT_RGB666;
+	case VID_MODE_FORMAT_RGB666_PACKED:
+		return MIPI_DSI_FMT_RGB666_PACKED;
+	case VID_MODE_FORMAT_RGB565:
+		return MIPI_DSI_FMT_RGB565;
+	default:
+		MISSING_CASE(fmt);
+		return MIPI_DSI_FMT_RGB666;
+	}
+}
+
+void vlv_dsi_wait_for_fifo_empty(struct intel_dsi *intel_dsi, enum port port)
+{
+	struct drm_encoder *encoder = &intel_dsi->base.base;
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 mask;
+
+	mask = LP_CTRL_FIFO_EMPTY | HS_CTRL_FIFO_EMPTY |
+		LP_DATA_FIFO_EMPTY | HS_DATA_FIFO_EMPTY;
+
+	if (intel_de_wait_for_set(dev_priv, MIPI_GEN_FIFO_STAT(port),
+				  mask, 100))
+		drm_err(&dev_priv->drm, "DPI FIFOs are not empty\n");
+}
+
+static void write_data(struct drm_i915_private *dev_priv,
+		       i915_reg_t reg,
+		       const u8 *data, u32 len)
+{
+	u32 i, j;
+
+	for (i = 0; i < len; i += 4) {
+		u32 val = 0;
+
+		for (j = 0; j < min_t(u32, len - i, 4); j++)
+			val |= *data++ << 8 * j;
+
+		intel_de_write(dev_priv, reg, val);
+	}
+}
+
+static void read_data(struct drm_i915_private *dev_priv,
+		      i915_reg_t reg,
+		      u8 *data, u32 len)
+{
+	u32 i, j;
+
+	for (i = 0; i < len; i += 4) {
+		u32 val = intel_de_read(dev_priv, reg);
+
+		for (j = 0; j < min_t(u32, len - i, 4); j++)
+			*data++ = val >> 8 * j;
+	}
+}
+
+static ssize_t intel_dsi_host_transfer(struct mipi_dsi_host *host,
+				       const struct mipi_dsi_msg *msg)
+{
+	struct intel_dsi_host *intel_dsi_host = to_intel_dsi_host(host);
+	struct drm_device *dev = intel_dsi_host->intel_dsi->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	enum port port = intel_dsi_host->port;
+	struct mipi_dsi_packet packet;
+	ssize_t ret;
+	const u8 *header, *data;
+	i915_reg_t data_reg, ctrl_reg;
+	u32 data_mask, ctrl_mask;
+
+	ret = mipi_dsi_create_packet(&packet, msg);
+	if (ret < 0)
+		return ret;
+
+	header = packet.header;
+	data = packet.payload;
+
+	if (msg->flags & MIPI_DSI_MSG_USE_LPM) {
+		data_reg = MIPI_LP_GEN_DATA(port);
+		data_mask = LP_DATA_FIFO_FULL;
+		ctrl_reg = MIPI_LP_GEN_CTRL(port);
+		ctrl_mask = LP_CTRL_FIFO_FULL;
+	} else {
+		data_reg = MIPI_HS_GEN_DATA(port);
+		data_mask = HS_DATA_FIFO_FULL;
+		ctrl_reg = MIPI_HS_GEN_CTRL(port);
+		ctrl_mask = HS_CTRL_FIFO_FULL;
+	}
+
+	/* note: this is never true for reads */
+	if (packet.payload_length) {
+		if (intel_de_wait_for_clear(dev_priv, MIPI_GEN_FIFO_STAT(port),
+					    data_mask, 50))
+			drm_err(&dev_priv->drm,
+				"Timeout waiting for HS/LP DATA FIFO !full\n");
+
+		write_data(dev_priv, data_reg, packet.payload,
+			   packet.payload_length);
+	}
+
+	if (msg->rx_len) {
+		intel_de_write(dev_priv, MIPI_INTR_STAT(port),
+			       GEN_READ_DATA_AVAIL);
+	}
+
+	if (intel_de_wait_for_clear(dev_priv, MIPI_GEN_FIFO_STAT(port),
+				    ctrl_mask, 50)) {
+		drm_err(&dev_priv->drm,
+			"Timeout waiting for HS/LP CTRL FIFO !full\n");
+	}
+
+	intel_de_write(dev_priv, ctrl_reg,
+		       header[2] << 16 | header[1] << 8 | header[0]);
+
+	/* ->rx_len is set only for reads */
+	if (msg->rx_len) {
+		data_mask = GEN_READ_DATA_AVAIL;
+		if (intel_de_wait_for_set(dev_priv, MIPI_INTR_STAT(port),
+					  data_mask, 50))
+			drm_err(&dev_priv->drm,
+				"Timeout waiting for read data.\n");
+
+		read_data(dev_priv, data_reg, msg->rx_buf, msg->rx_len);
+	}
+
+	/* XXX: fix for reads and writes */
+	return 4 + packet.payload_length;
+}
+
+static int intel_dsi_host_attach(struct mipi_dsi_host *host,
+				 struct mipi_dsi_device *dsi)
+{
+	return 0;
+}
+
+static int intel_dsi_host_detach(struct mipi_dsi_host *host,
+				 struct mipi_dsi_device *dsi)
+{
+	return 0;
+}
+
+static const struct mipi_dsi_host_ops intel_dsi_host_ops = {
+	.attach = intel_dsi_host_attach,
+	.detach = intel_dsi_host_detach,
+	.transfer = intel_dsi_host_transfer,
+};
+
+/*
+ * send a video mode command
+ *
+ * XXX: commands with data in MIPI_DPI_DATA?
+ */
+static int dpi_send_cmd(struct intel_dsi *intel_dsi, u32 cmd, bool hs,
+			enum port port)
+{
+	struct drm_encoder *encoder = &intel_dsi->base.base;
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 mask;
+
+	/* XXX: pipe, hs */
+	if (hs)
+		cmd &= ~DPI_LP_MODE;
+	else
+		cmd |= DPI_LP_MODE;
+
+	/* clear bit */
+	intel_de_write(dev_priv, MIPI_INTR_STAT(port), SPL_PKT_SENT_INTERRUPT);
+
+	/* XXX: old code skips write if control unchanged */
+	if (cmd == intel_de_read(dev_priv, MIPI_DPI_CONTROL(port)))
+		drm_dbg_kms(&dev_priv->drm,
+			    "Same special packet %02x twice in a row.\n", cmd);
+
+	intel_de_write(dev_priv, MIPI_DPI_CONTROL(port), cmd);
+
+	mask = SPL_PKT_SENT_INTERRUPT;
+	if (intel_de_wait_for_set(dev_priv, MIPI_INTR_STAT(port), mask, 100))
+		drm_err(&dev_priv->drm,
+			"Video mode command 0x%08x send failed.\n", cmd);
+
+	return 0;
+}
+
+static void band_gap_reset(struct drm_i915_private *dev_priv)
+{
+	vlv_flisdsi_get(dev_priv);
+
+	vlv_flisdsi_write(dev_priv, 0x08, 0x0001);
+	vlv_flisdsi_write(dev_priv, 0x0F, 0x0005);
+	vlv_flisdsi_write(dev_priv, 0x0F, 0x0025);
+	udelay(150);
+	vlv_flisdsi_write(dev_priv, 0x0F, 0x0000);
+	vlv_flisdsi_write(dev_priv, 0x08, 0x0000);
+
+	vlv_flisdsi_put(dev_priv);
+}
+
+static int intel_dsi_compute_config(struct intel_encoder *encoder,
+				    struct intel_crtc_state *pipe_config,
+				    struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = container_of(encoder, struct intel_dsi,
+						   base);
+	struct intel_connector *intel_connector = intel_dsi->attached_connector;
+	struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	int ret;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+	pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB;
+
+	ret = intel_panel_compute_config(intel_connector, adjusted_mode);
+	if (ret)
+		return ret;
+
+	ret = intel_panel_fitting(pipe_config, conn_state);
+	if (ret)
+		return ret;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN)
+		return -EINVAL;
+
+	/* DSI uses short packets for sync events, so clear mode flags for DSI */
+	adjusted_mode->flags = 0;
+
+	if (intel_dsi->pixel_format == MIPI_DSI_FMT_RGB888)
+		pipe_config->pipe_bpp = 24;
+	else
+		pipe_config->pipe_bpp = 18;
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		/* Enable Frame time stamp based scanline reporting */
+		pipe_config->mode_flags |=
+			I915_MODE_FLAG_GET_SCANLINE_FROM_TIMESTAMP;
+
+		/* Dual link goes to DSI transcoder A. */
+		if (intel_dsi->ports == BIT(PORT_C))
+			pipe_config->cpu_transcoder = TRANSCODER_DSI_C;
+		else
+			pipe_config->cpu_transcoder = TRANSCODER_DSI_A;
+
+		ret = bxt_dsi_pll_compute(encoder, pipe_config);
+		if (ret)
+			return -EINVAL;
+	} else {
+		ret = vlv_dsi_pll_compute(encoder, pipe_config);
+		if (ret)
+			return -EINVAL;
+	}
+
+	pipe_config->clock_set = true;
+
+	return 0;
+}
+
+static bool glk_dsi_enable_io(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 tmp;
+	bool cold_boot = false;
+
+	/* Set the MIPI mode
+	 * If MIPI_Mode is off, then writing to LP_Wake bit is not reflecting.
+	 * Power ON MIPI IO first and then write into IO reset and LP wake bits
+	 */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = intel_de_read(dev_priv, MIPI_CTRL(port));
+		intel_de_write(dev_priv, MIPI_CTRL(port),
+			       tmp | GLK_MIPIIO_ENABLE);
+	}
+
+	/* Put the IO into reset */
+	tmp = intel_de_read(dev_priv, MIPI_CTRL(PORT_A));
+	tmp &= ~GLK_MIPIIO_RESET_RELEASED;
+	intel_de_write(dev_priv, MIPI_CTRL(PORT_A), tmp);
+
+	/* Program LP Wake */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = intel_de_read(dev_priv, MIPI_CTRL(port));
+		if (!(intel_de_read(dev_priv, MIPI_DEVICE_READY(port)) & DEVICE_READY))
+			tmp &= ~GLK_LP_WAKE;
+		else
+			tmp |= GLK_LP_WAKE;
+		intel_de_write(dev_priv, MIPI_CTRL(port), tmp);
+	}
+
+	/* Wait for Pwr ACK */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_de_wait_for_set(dev_priv, MIPI_CTRL(port),
+					  GLK_MIPIIO_PORT_POWERED, 20))
+			drm_err(&dev_priv->drm, "MIPIO port is powergated\n");
+	}
+
+	/* Check for cold boot scenario */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		cold_boot |=
+			!(intel_de_read(dev_priv, MIPI_DEVICE_READY(port)) & DEVICE_READY);
+	}
+
+	return cold_boot;
+}
+
+static void glk_dsi_device_ready(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 val;
+
+	/* Wait for MIPI PHY status bit to set */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_de_wait_for_set(dev_priv, MIPI_CTRL(port),
+					  GLK_PHY_STATUS_PORT_READY, 20))
+			drm_err(&dev_priv->drm, "PHY is not ON\n");
+	}
+
+	/* Get IO out of reset */
+	val = intel_de_read(dev_priv, MIPI_CTRL(PORT_A));
+	intel_de_write(dev_priv, MIPI_CTRL(PORT_A),
+		       val | GLK_MIPIIO_RESET_RELEASED);
+
+	/* Get IO out of Low power state*/
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (!(intel_de_read(dev_priv, MIPI_DEVICE_READY(port)) & DEVICE_READY)) {
+			val = intel_de_read(dev_priv, MIPI_DEVICE_READY(port));
+			val &= ~ULPS_STATE_MASK;
+			val |= DEVICE_READY;
+			intel_de_write(dev_priv, MIPI_DEVICE_READY(port), val);
+			usleep_range(10, 15);
+		} else {
+			/* Enter ULPS */
+			val = intel_de_read(dev_priv, MIPI_DEVICE_READY(port));
+			val &= ~ULPS_STATE_MASK;
+			val |= (ULPS_STATE_ENTER | DEVICE_READY);
+			intel_de_write(dev_priv, MIPI_DEVICE_READY(port), val);
+
+			/* Wait for ULPS active */
+			if (intel_de_wait_for_clear(dev_priv, MIPI_CTRL(port),
+						    GLK_ULPS_NOT_ACTIVE, 20))
+				drm_err(&dev_priv->drm, "ULPS not active\n");
+
+			/* Exit ULPS */
+			val = intel_de_read(dev_priv, MIPI_DEVICE_READY(port));
+			val &= ~ULPS_STATE_MASK;
+			val |= (ULPS_STATE_EXIT | DEVICE_READY);
+			intel_de_write(dev_priv, MIPI_DEVICE_READY(port), val);
+
+			/* Enter Normal Mode */
+			val = intel_de_read(dev_priv, MIPI_DEVICE_READY(port));
+			val &= ~ULPS_STATE_MASK;
+			val |= (ULPS_STATE_NORMAL_OPERATION | DEVICE_READY);
+			intel_de_write(dev_priv, MIPI_DEVICE_READY(port), val);
+
+			val = intel_de_read(dev_priv, MIPI_CTRL(port));
+			val &= ~GLK_LP_WAKE;
+			intel_de_write(dev_priv, MIPI_CTRL(port), val);
+		}
+	}
+
+	/* Wait for Stop state */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_de_wait_for_set(dev_priv, MIPI_CTRL(port),
+					  GLK_DATA_LANE_STOP_STATE, 20))
+			drm_err(&dev_priv->drm,
+				"Date lane not in STOP state\n");
+	}
+
+	/* Wait for AFE LATCH */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_de_wait_for_set(dev_priv, BXT_MIPI_PORT_CTRL(port),
+					  AFE_LATCHOUT, 20))
+			drm_err(&dev_priv->drm,
+				"D-PHY not entering LP-11 state\n");
+	}
+}
+
+static void bxt_dsi_device_ready(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 val;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	/* Enable MIPI PHY transparent latch */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		val = intel_de_read(dev_priv, BXT_MIPI_PORT_CTRL(port));
+		intel_de_write(dev_priv, BXT_MIPI_PORT_CTRL(port),
+			       val | LP_OUTPUT_HOLD);
+		usleep_range(2000, 2500);
+	}
+
+	/* Clear ULPS and set device ready */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		val = intel_de_read(dev_priv, MIPI_DEVICE_READY(port));
+		val &= ~ULPS_STATE_MASK;
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port), val);
+		usleep_range(2000, 2500);
+		val |= DEVICE_READY;
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port), val);
+	}
+}
+
+static void vlv_dsi_device_ready(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 val;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	vlv_flisdsi_get(dev_priv);
+	/* program rcomp for compliance, reduce from 50 ohms to 45 ohms
+	 * needed everytime after power gate */
+	vlv_flisdsi_write(dev_priv, 0x04, 0x0004);
+	vlv_flisdsi_put(dev_priv);
+
+	/* bandgap reset is needed after everytime we do power gate */
+	band_gap_reset(dev_priv);
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port),
+			       ULPS_STATE_ENTER);
+		usleep_range(2500, 3000);
+
+		/* Enable MIPI PHY transparent latch
+		 * Common bit for both MIPI Port A & MIPI Port C
+		 * No similar bit in MIPI Port C reg
+		 */
+		val = intel_de_read(dev_priv, MIPI_PORT_CTRL(PORT_A));
+		intel_de_write(dev_priv, MIPI_PORT_CTRL(PORT_A),
+			       val | LP_OUTPUT_HOLD);
+		usleep_range(1000, 1500);
+
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port),
+			       ULPS_STATE_EXIT);
+		usleep_range(2500, 3000);
+
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port),
+			       DEVICE_READY);
+		usleep_range(2500, 3000);
+	}
+}
+
+static void intel_dsi_device_ready(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (IS_GEMINILAKE(dev_priv))
+		glk_dsi_device_ready(encoder);
+	else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		bxt_dsi_device_ready(encoder);
+	else
+		vlv_dsi_device_ready(encoder);
+}
+
+static void glk_dsi_enter_low_power_mode(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 val;
+
+	/* Enter ULPS */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		val = intel_de_read(dev_priv, MIPI_DEVICE_READY(port));
+		val &= ~ULPS_STATE_MASK;
+		val |= (ULPS_STATE_ENTER | DEVICE_READY);
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port), val);
+	}
+
+	/* Wait for MIPI PHY status bit to unset */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_de_wait_for_clear(dev_priv, MIPI_CTRL(port),
+					    GLK_PHY_STATUS_PORT_READY, 20))
+			drm_err(&dev_priv->drm, "PHY is not turning OFF\n");
+	}
+
+	/* Wait for Pwr ACK bit to unset */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_de_wait_for_clear(dev_priv, MIPI_CTRL(port),
+					    GLK_MIPIIO_PORT_POWERED, 20))
+			drm_err(&dev_priv->drm,
+				"MIPI IO Port is not powergated\n");
+	}
+}
+
+static void glk_dsi_disable_mipi_io(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 tmp;
+
+	/* Put the IO into reset */
+	tmp = intel_de_read(dev_priv, MIPI_CTRL(PORT_A));
+	tmp &= ~GLK_MIPIIO_RESET_RELEASED;
+	intel_de_write(dev_priv, MIPI_CTRL(PORT_A), tmp);
+
+	/* Wait for MIPI PHY status bit to unset */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_de_wait_for_clear(dev_priv, MIPI_CTRL(port),
+					    GLK_PHY_STATUS_PORT_READY, 20))
+			drm_err(&dev_priv->drm, "PHY is not turning OFF\n");
+	}
+
+	/* Clear MIPI mode */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = intel_de_read(dev_priv, MIPI_CTRL(port));
+		tmp &= ~GLK_MIPIIO_ENABLE;
+		intel_de_write(dev_priv, MIPI_CTRL(port), tmp);
+	}
+}
+
+static void glk_dsi_clear_device_ready(struct intel_encoder *encoder)
+{
+	glk_dsi_enter_low_power_mode(encoder);
+	glk_dsi_disable_mipi_io(encoder);
+}
+
+static void vlv_dsi_clear_device_ready(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+	for_each_dsi_port(port, intel_dsi->ports) {
+		/* Common bit for both MIPI Port A & MIPI Port C on VLV/CHV */
+		i915_reg_t port_ctrl = IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv) ?
+			BXT_MIPI_PORT_CTRL(port) : MIPI_PORT_CTRL(PORT_A);
+		u32 val;
+
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port),
+			       DEVICE_READY | ULPS_STATE_ENTER);
+		usleep_range(2000, 2500);
+
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port),
+			       DEVICE_READY | ULPS_STATE_EXIT);
+		usleep_range(2000, 2500);
+
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port),
+			       DEVICE_READY | ULPS_STATE_ENTER);
+		usleep_range(2000, 2500);
+
+		/*
+		 * On VLV/CHV, wait till Clock lanes are in LP-00 state for MIPI
+		 * Port A only. MIPI Port C has no similar bit for checking.
+		 */
+		if ((IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv) || port == PORT_A) &&
+		    intel_de_wait_for_clear(dev_priv, port_ctrl,
+					    AFE_LATCHOUT, 30))
+			drm_err(&dev_priv->drm, "DSI LP not going Low\n");
+
+		/* Disable MIPI PHY transparent latch */
+		val = intel_de_read(dev_priv, port_ctrl);
+		intel_de_write(dev_priv, port_ctrl, val & ~LP_OUTPUT_HOLD);
+		usleep_range(1000, 1500);
+
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port), 0x00);
+		usleep_range(2000, 2500);
+	}
+}
+
+static void intel_dsi_port_enable(struct intel_encoder *encoder,
+				  const struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+
+	if (intel_dsi->dual_link == DSI_DUAL_LINK_FRONT_BACK) {
+		u32 temp;
+		if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+			for_each_dsi_port(port, intel_dsi->ports) {
+				temp = intel_de_read(dev_priv,
+						     MIPI_CTRL(port));
+				temp &= ~BXT_PIXEL_OVERLAP_CNT_MASK |
+					intel_dsi->pixel_overlap <<
+					BXT_PIXEL_OVERLAP_CNT_SHIFT;
+				intel_de_write(dev_priv, MIPI_CTRL(port),
+					       temp);
+			}
+		} else {
+			temp = intel_de_read(dev_priv, VLV_CHICKEN_3);
+			temp &= ~PIXEL_OVERLAP_CNT_MASK |
+					intel_dsi->pixel_overlap <<
+					PIXEL_OVERLAP_CNT_SHIFT;
+			intel_de_write(dev_priv, VLV_CHICKEN_3, temp);
+		}
+	}
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		i915_reg_t port_ctrl = IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv) ?
+			BXT_MIPI_PORT_CTRL(port) : MIPI_PORT_CTRL(port);
+		u32 temp;
+
+		temp = intel_de_read(dev_priv, port_ctrl);
+
+		temp &= ~LANE_CONFIGURATION_MASK;
+		temp &= ~DUAL_LINK_MODE_MASK;
+
+		if (intel_dsi->ports == (BIT(PORT_A) | BIT(PORT_C))) {
+			temp |= (intel_dsi->dual_link - 1)
+						<< DUAL_LINK_MODE_SHIFT;
+			if (IS_BROXTON(dev_priv))
+				temp |= LANE_CONFIGURATION_DUAL_LINK_A;
+			else
+				temp |= crtc->pipe ?
+					LANE_CONFIGURATION_DUAL_LINK_B :
+					LANE_CONFIGURATION_DUAL_LINK_A;
+		}
+
+		if (intel_dsi->pixel_format != MIPI_DSI_FMT_RGB888)
+			temp |= DITHERING_ENABLE;
+
+		/* assert ip_tg_enable signal */
+		intel_de_write(dev_priv, port_ctrl, temp | DPI_ENABLE);
+		intel_de_posting_read(dev_priv, port_ctrl);
+	}
+}
+
+static void intel_dsi_port_disable(struct intel_encoder *encoder)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		i915_reg_t port_ctrl = IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv) ?
+			BXT_MIPI_PORT_CTRL(port) : MIPI_PORT_CTRL(port);
+		u32 temp;
+
+		/* de-assert ip_tg_enable signal */
+		temp = intel_de_read(dev_priv, port_ctrl);
+		intel_de_write(dev_priv, port_ctrl, temp & ~DPI_ENABLE);
+		intel_de_posting_read(dev_priv, port_ctrl);
+	}
+}
+
+static void intel_dsi_wait_panel_power_cycle(struct intel_dsi *intel_dsi)
+{
+	ktime_t panel_power_on_time;
+	s64 panel_power_off_duration;
+
+	panel_power_on_time = ktime_get_boottime();
+	panel_power_off_duration = ktime_ms_delta(panel_power_on_time,
+						  intel_dsi->panel_power_off_time);
+
+	if (panel_power_off_duration < (s64)intel_dsi->panel_pwr_cycle_delay)
+		msleep(intel_dsi->panel_pwr_cycle_delay - panel_power_off_duration);
+}
+
+static void intel_dsi_prepare(struct intel_encoder *intel_encoder,
+			      const struct intel_crtc_state *pipe_config);
+static void intel_dsi_unprepare(struct intel_encoder *encoder);
+
+/*
+ * Panel enable/disable sequences from the VBT spec.
+ *
+ * Note the spec has AssertReset / DeassertReset swapped from their
+ * usual naming. We use the normal names to avoid confusion (so below
+ * they are swapped compared to the spec).
+ *
+ * Steps starting with MIPI refer to VBT sequences, note that for v2
+ * VBTs several steps which have a VBT in v2 are expected to be handled
+ * directly by the driver, by directly driving gpios for example.
+ *
+ * v2 video mode seq         v3 video mode seq         command mode seq
+ * - power on                - MIPIPanelPowerOn        - power on
+ * - wait t1+t2                                        - wait t1+t2
+ * - MIPIDeassertResetPin    - MIPIDeassertResetPin    - MIPIDeassertResetPin
+ * - io lines to lp-11       - io lines to lp-11       - io lines to lp-11
+ * - MIPISendInitialDcsCmds  - MIPISendInitialDcsCmds  - MIPISendInitialDcsCmds
+ *                                                     - MIPITearOn
+ *                                                     - MIPIDisplayOn
+ * - turn on DPI             - turn on DPI             - set pipe to dsr mode
+ * - MIPIDisplayOn           - MIPIDisplayOn
+ * - wait t5                                           - wait t5
+ * - backlight on            - MIPIBacklightOn         - backlight on
+ * ...                       ...                       ... issue mem cmds ...
+ * - backlight off           - MIPIBacklightOff        - backlight off
+ * - wait t6                                           - wait t6
+ * - MIPIDisplayOff
+ * - turn off DPI            - turn off DPI            - disable pipe dsr mode
+ *                                                     - MIPITearOff
+ *                           - MIPIDisplayOff          - MIPIDisplayOff
+ * - io lines to lp-00       - io lines to lp-00       - io lines to lp-00
+ * - MIPIAssertResetPin      - MIPIAssertResetPin      - MIPIAssertResetPin
+ * - wait t3                                           - wait t3
+ * - power off               - MIPIPanelPowerOff       - power off
+ * - wait t4                                           - wait t4
+ */
+
+/*
+ * DSI port enable has to be done before pipe and plane enable, so we do it in
+ * the pre_enable hook instead of the enable hook.
+ */
+static void intel_dsi_pre_enable(struct intel_atomic_state *state,
+				 struct intel_encoder *encoder,
+				 const struct intel_crtc_state *pipe_config,
+				 const struct drm_connector_state *conn_state)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	enum port port;
+	u32 val;
+	bool glk_cold_boot = false;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	intel_dsi_wait_panel_power_cycle(intel_dsi);
+
+	intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true);
+
+	/*
+	 * The BIOS may leave the PLL in a wonky state where it doesn't
+	 * lock. It needs to be fully powered down to fix it.
+	 */
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		bxt_dsi_pll_disable(encoder);
+		bxt_dsi_pll_enable(encoder, pipe_config);
+	} else {
+		vlv_dsi_pll_disable(encoder);
+		vlv_dsi_pll_enable(encoder, pipe_config);
+	}
+
+	if (IS_BROXTON(dev_priv)) {
+		/* Add MIPI IO reset programming for modeset */
+		val = intel_de_read(dev_priv, BXT_P_CR_GT_DISP_PWRON);
+		intel_de_write(dev_priv, BXT_P_CR_GT_DISP_PWRON,
+			       val | MIPIO_RST_CTRL);
+
+		/* Power up DSI regulator */
+		intel_de_write(dev_priv, BXT_P_DSI_REGULATOR_CFG, STAP_SELECT);
+		intel_de_write(dev_priv, BXT_P_DSI_REGULATOR_TX_CTRL, 0);
+	}
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		u32 val;
+
+		/* Disable DPOunit clock gating, can stall pipe */
+		val = intel_de_read(dev_priv, DSPCLK_GATE_D(dev_priv));
+		val |= DPOUNIT_CLOCK_GATE_DISABLE;
+		intel_de_write(dev_priv, DSPCLK_GATE_D(dev_priv), val);
+	}
+
+	if (!IS_GEMINILAKE(dev_priv))
+		intel_dsi_prepare(encoder, pipe_config);
+
+	/* Give the panel time to power-on and then deassert its reset */
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_POWER_ON);
+	msleep(intel_dsi->panel_on_delay);
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DEASSERT_RESET);
+
+	if (IS_GEMINILAKE(dev_priv)) {
+		glk_cold_boot = glk_dsi_enable_io(encoder);
+
+		/* Prepare port in cold boot(s3/s4) scenario */
+		if (glk_cold_boot)
+			intel_dsi_prepare(encoder, pipe_config);
+	}
+
+	/* Put device in ready state (LP-11) */
+	intel_dsi_device_ready(encoder);
+
+	/* Prepare port in normal boot scenario */
+	if (IS_GEMINILAKE(dev_priv) && !glk_cold_boot)
+		intel_dsi_prepare(encoder, pipe_config);
+
+	/* Send initialization commands in LP mode */
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_INIT_OTP);
+
+	/*
+	 * Enable port in pre-enable phase itself because as per hw team
+	 * recommendation, port should be enabled before plane & pipe
+	 */
+	if (is_cmd_mode(intel_dsi)) {
+		for_each_dsi_port(port, intel_dsi->ports)
+			intel_de_write(dev_priv,
+				       MIPI_MAX_RETURN_PKT_SIZE(port), 8 * 4);
+		intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_TEAR_ON);
+		intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DISPLAY_ON);
+	} else {
+		msleep(20); /* XXX */
+		for_each_dsi_port(port, intel_dsi->ports)
+			dpi_send_cmd(intel_dsi, TURN_ON, false, port);
+		msleep(100);
+
+		intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DISPLAY_ON);
+
+		intel_dsi_port_enable(encoder, pipe_config);
+	}
+
+	intel_backlight_enable(pipe_config, conn_state);
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_BACKLIGHT_ON);
+}
+
+static void bxt_dsi_enable(struct intel_atomic_state *state,
+			   struct intel_encoder *encoder,
+			   const struct intel_crtc_state *crtc_state,
+			   const struct drm_connector_state *conn_state)
+{
+	drm_WARN_ON(state->base.dev, crtc_state->has_pch_encoder);
+
+	intel_crtc_vblank_on(crtc_state);
+}
+
+/*
+ * DSI port disable has to be done after pipe and plane disable, so we do it in
+ * the post_disable hook.
+ */
+static void intel_dsi_disable(struct intel_atomic_state *state,
+			      struct intel_encoder *encoder,
+			      const struct intel_crtc_state *old_crtc_state,
+			      const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+
+	drm_dbg_kms(&i915->drm, "\n");
+
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_BACKLIGHT_OFF);
+	intel_backlight_disable(old_conn_state);
+
+	/*
+	 * According to the spec we should send SHUTDOWN before
+	 * MIPI_SEQ_DISPLAY_OFF only for v3+ VBTs, but field testing
+	 * has shown that the v3 sequence works for v2 VBTs too
+	 */
+	if (is_vid_mode(intel_dsi)) {
+		/* Send Shutdown command to the panel in LP mode */
+		for_each_dsi_port(port, intel_dsi->ports)
+			dpi_send_cmd(intel_dsi, SHUTDOWN, false, port);
+		msleep(10);
+	}
+}
+
+static void intel_dsi_clear_device_ready(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (IS_GEMINILAKE(dev_priv))
+		glk_dsi_clear_device_ready(encoder);
+	else
+		vlv_dsi_clear_device_ready(encoder);
+}
+
+static void intel_dsi_post_disable(struct intel_atomic_state *state,
+				   struct intel_encoder *encoder,
+				   const struct intel_crtc_state *old_crtc_state,
+				   const struct drm_connector_state *old_conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 val;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		intel_crtc_vblank_off(old_crtc_state);
+
+		skl_scaler_disable(old_crtc_state);
+	}
+
+	if (is_vid_mode(intel_dsi)) {
+		for_each_dsi_port(port, intel_dsi->ports)
+			vlv_dsi_wait_for_fifo_empty(intel_dsi, port);
+
+		intel_dsi_port_disable(encoder);
+		usleep_range(2000, 5000);
+	}
+
+	intel_dsi_unprepare(encoder);
+
+	/*
+	 * if disable packets are sent before sending shutdown packet then in
+	 * some next enable sequence send turn on packet error is observed
+	 */
+	if (is_cmd_mode(intel_dsi))
+		intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_TEAR_OFF);
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_DISPLAY_OFF);
+
+	/* Transition to LP-00 */
+	intel_dsi_clear_device_ready(encoder);
+
+	if (IS_BROXTON(dev_priv)) {
+		/* Power down DSI regulator to save power */
+		intel_de_write(dev_priv, BXT_P_DSI_REGULATOR_CFG, STAP_SELECT);
+		intel_de_write(dev_priv, BXT_P_DSI_REGULATOR_TX_CTRL,
+			       HS_IO_CTRL_SELECT);
+
+		/* Add MIPI IO reset programming for modeset */
+		val = intel_de_read(dev_priv, BXT_P_CR_GT_DISP_PWRON);
+		intel_de_write(dev_priv, BXT_P_CR_GT_DISP_PWRON,
+			       val & ~MIPIO_RST_CTRL);
+	}
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		bxt_dsi_pll_disable(encoder);
+	} else {
+		u32 val;
+
+		vlv_dsi_pll_disable(encoder);
+
+		val = intel_de_read(dev_priv, DSPCLK_GATE_D(dev_priv));
+		val &= ~DPOUNIT_CLOCK_GATE_DISABLE;
+		intel_de_write(dev_priv, DSPCLK_GATE_D(dev_priv), val);
+	}
+
+	/* Assert reset */
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_ASSERT_RESET);
+
+	msleep(intel_dsi->panel_off_delay);
+	intel_dsi_vbt_exec_sequence(intel_dsi, MIPI_SEQ_POWER_OFF);
+
+	intel_dsi->panel_power_off_time = ktime_get_boottime();
+}
+
+static void intel_dsi_shutdown(struct intel_encoder *encoder)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+
+	intel_dsi_wait_panel_power_cycle(intel_dsi);
+}
+
+static bool intel_dsi_get_hw_state(struct intel_encoder *encoder,
+				   enum pipe *pipe)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	intel_wakeref_t wakeref;
+	enum port port;
+	bool active = false;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	wakeref = intel_display_power_get_if_enabled(dev_priv,
+						     encoder->power_domain);
+	if (!wakeref)
+		return false;
+
+	/*
+	 * On Broxton the PLL needs to be enabled with a valid divider
+	 * configuration, otherwise accessing DSI registers will hang the
+	 * machine. See BSpec North Display Engine registers/MIPI[BXT].
+	 */
+	if ((IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) &&
+	    !bxt_dsi_pll_is_enabled(dev_priv))
+		goto out_put_power;
+
+	/* XXX: this only works for one DSI output */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		i915_reg_t ctrl_reg = IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv) ?
+			BXT_MIPI_PORT_CTRL(port) : MIPI_PORT_CTRL(port);
+		bool enabled = intel_de_read(dev_priv, ctrl_reg) & DPI_ENABLE;
+
+		/*
+		 * Due to some hardware limitations on VLV/CHV, the DPI enable
+		 * bit in port C control register does not get set. As a
+		 * workaround, check pipe B conf instead.
+		 */
+		if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+		    port == PORT_C)
+			enabled = intel_de_read(dev_priv, PIPECONF(PIPE_B)) & PIPECONF_ENABLE;
+
+		/* Try command mode if video mode not enabled */
+		if (!enabled) {
+			u32 tmp = intel_de_read(dev_priv,
+						MIPI_DSI_FUNC_PRG(port));
+			enabled = tmp & CMD_MODE_DATA_WIDTH_MASK;
+		}
+
+		if (!enabled)
+			continue;
+
+		if (!(intel_de_read(dev_priv, MIPI_DEVICE_READY(port)) & DEVICE_READY))
+			continue;
+
+		if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+			u32 tmp = intel_de_read(dev_priv, MIPI_CTRL(port));
+			tmp &= BXT_PIPE_SELECT_MASK;
+			tmp >>= BXT_PIPE_SELECT_SHIFT;
+
+			if (drm_WARN_ON(&dev_priv->drm, tmp > PIPE_C))
+				continue;
+
+			*pipe = tmp;
+		} else {
+			*pipe = port == PORT_A ? PIPE_A : PIPE_B;
+		}
+
+		active = true;
+		break;
+	}
+
+out_put_power:
+	intel_display_power_put(dev_priv, encoder->power_domain, wakeref);
+
+	return active;
+}
+
+static void bxt_dsi_get_pipe_config(struct intel_encoder *encoder,
+				    struct intel_crtc_state *pipe_config)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_display_mode *adjusted_mode =
+					&pipe_config->hw.adjusted_mode;
+	struct drm_display_mode *adjusted_mode_sw;
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	unsigned int lane_count = intel_dsi->lane_count;
+	unsigned int bpp, fmt;
+	enum port port;
+	u16 hactive, hfp, hsync, hbp, vfp, vsync, vbp;
+	u16 hfp_sw, hsync_sw, hbp_sw;
+	u16 crtc_htotal_sw, crtc_hsync_start_sw, crtc_hsync_end_sw,
+				crtc_hblank_start_sw, crtc_hblank_end_sw;
+
+	/* FIXME: hw readout should not depend on SW state */
+	adjusted_mode_sw = &crtc->config->hw.adjusted_mode;
+
+	/*
+	 * Atleast one port is active as encoder->get_config called only if
+	 * encoder->get_hw_state() returns true.
+	 */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_de_read(dev_priv, BXT_MIPI_PORT_CTRL(port)) & DPI_ENABLE)
+			break;
+	}
+
+	fmt = intel_de_read(dev_priv, MIPI_DSI_FUNC_PRG(port)) & VID_MODE_FORMAT_MASK;
+	bpp = mipi_dsi_pixel_format_to_bpp(
+			pixel_format_from_register_bits(fmt));
+
+	pipe_config->pipe_bpp = bdw_get_pipemisc_bpp(crtc);
+
+	/* Enable Frame time stamo based scanline reporting */
+	pipe_config->mode_flags |=
+		I915_MODE_FLAG_GET_SCANLINE_FROM_TIMESTAMP;
+
+	/* In terms of pixels */
+	adjusted_mode->crtc_hdisplay =
+				intel_de_read(dev_priv,
+				              BXT_MIPI_TRANS_HACTIVE(port));
+	adjusted_mode->crtc_vdisplay =
+				intel_de_read(dev_priv,
+				              BXT_MIPI_TRANS_VACTIVE(port));
+	adjusted_mode->crtc_vtotal =
+				intel_de_read(dev_priv,
+				              BXT_MIPI_TRANS_VTOTAL(port));
+
+	hactive = adjusted_mode->crtc_hdisplay;
+	hfp = intel_de_read(dev_priv, MIPI_HFP_COUNT(port));
+
+	/*
+	 * Meaningful for video mode non-burst sync pulse mode only,
+	 * can be zero for non-burst sync events and burst modes
+	 */
+	hsync = intel_de_read(dev_priv, MIPI_HSYNC_PADDING_COUNT(port));
+	hbp = intel_de_read(dev_priv, MIPI_HBP_COUNT(port));
+
+	/* harizontal values are in terms of high speed byte clock */
+	hfp = pixels_from_txbyteclkhs(hfp, bpp, lane_count,
+						intel_dsi->burst_mode_ratio);
+	hsync = pixels_from_txbyteclkhs(hsync, bpp, lane_count,
+						intel_dsi->burst_mode_ratio);
+	hbp = pixels_from_txbyteclkhs(hbp, bpp, lane_count,
+						intel_dsi->burst_mode_ratio);
+
+	if (intel_dsi->dual_link) {
+		hfp *= 2;
+		hsync *= 2;
+		hbp *= 2;
+	}
+
+	/* vertical values are in terms of lines */
+	vfp = intel_de_read(dev_priv, MIPI_VFP_COUNT(port));
+	vsync = intel_de_read(dev_priv, MIPI_VSYNC_PADDING_COUNT(port));
+	vbp = intel_de_read(dev_priv, MIPI_VBP_COUNT(port));
+
+	adjusted_mode->crtc_htotal = hactive + hfp + hsync + hbp;
+	adjusted_mode->crtc_hsync_start = hfp + adjusted_mode->crtc_hdisplay;
+	adjusted_mode->crtc_hsync_end = hsync + adjusted_mode->crtc_hsync_start;
+	adjusted_mode->crtc_hblank_start = adjusted_mode->crtc_hdisplay;
+	adjusted_mode->crtc_hblank_end = adjusted_mode->crtc_htotal;
+
+	adjusted_mode->crtc_vsync_start = vfp + adjusted_mode->crtc_vdisplay;
+	adjusted_mode->crtc_vsync_end = vsync + adjusted_mode->crtc_vsync_start;
+	adjusted_mode->crtc_vblank_start = adjusted_mode->crtc_vdisplay;
+	adjusted_mode->crtc_vblank_end = adjusted_mode->crtc_vtotal;
+
+	/*
+	 * In BXT DSI there is no regs programmed with few horizontal timings
+	 * in Pixels but txbyteclkhs.. So retrieval process adds some
+	 * ROUND_UP ERRORS in the process of PIXELS<==>txbyteclkhs.
+	 * Actually here for the given adjusted_mode, we are calculating the
+	 * value programmed to the port and then back to the horizontal timing
+	 * param in pixels. This is the expected value, including roundup errors
+	 * And if that is same as retrieved value from port, then
+	 * (HW state) adjusted_mode's horizontal timings are corrected to
+	 * match with SW state to nullify the errors.
+	 */
+	/* Calculating the value programmed to the Port register */
+	hfp_sw = adjusted_mode_sw->crtc_hsync_start -
+					adjusted_mode_sw->crtc_hdisplay;
+	hsync_sw = adjusted_mode_sw->crtc_hsync_end -
+					adjusted_mode_sw->crtc_hsync_start;
+	hbp_sw = adjusted_mode_sw->crtc_htotal -
+					adjusted_mode_sw->crtc_hsync_end;
+
+	if (intel_dsi->dual_link) {
+		hfp_sw /= 2;
+		hsync_sw /= 2;
+		hbp_sw /= 2;
+	}
+
+	hfp_sw = txbyteclkhs(hfp_sw, bpp, lane_count,
+						intel_dsi->burst_mode_ratio);
+	hsync_sw = txbyteclkhs(hsync_sw, bpp, lane_count,
+			    intel_dsi->burst_mode_ratio);
+	hbp_sw = txbyteclkhs(hbp_sw, bpp, lane_count,
+						intel_dsi->burst_mode_ratio);
+
+	/* Reverse calculating the adjusted mode parameters from port reg vals*/
+	hfp_sw = pixels_from_txbyteclkhs(hfp_sw, bpp, lane_count,
+						intel_dsi->burst_mode_ratio);
+	hsync_sw = pixels_from_txbyteclkhs(hsync_sw, bpp, lane_count,
+						intel_dsi->burst_mode_ratio);
+	hbp_sw = pixels_from_txbyteclkhs(hbp_sw, bpp, lane_count,
+						intel_dsi->burst_mode_ratio);
+
+	if (intel_dsi->dual_link) {
+		hfp_sw *= 2;
+		hsync_sw *= 2;
+		hbp_sw *= 2;
+	}
+
+	crtc_htotal_sw = adjusted_mode_sw->crtc_hdisplay + hfp_sw +
+							hsync_sw + hbp_sw;
+	crtc_hsync_start_sw = hfp_sw + adjusted_mode_sw->crtc_hdisplay;
+	crtc_hsync_end_sw = hsync_sw + crtc_hsync_start_sw;
+	crtc_hblank_start_sw = adjusted_mode_sw->crtc_hdisplay;
+	crtc_hblank_end_sw = crtc_htotal_sw;
+
+	if (adjusted_mode->crtc_htotal == crtc_htotal_sw)
+		adjusted_mode->crtc_htotal = adjusted_mode_sw->crtc_htotal;
+
+	if (adjusted_mode->crtc_hsync_start == crtc_hsync_start_sw)
+		adjusted_mode->crtc_hsync_start =
+					adjusted_mode_sw->crtc_hsync_start;
+
+	if (adjusted_mode->crtc_hsync_end == crtc_hsync_end_sw)
+		adjusted_mode->crtc_hsync_end =
+					adjusted_mode_sw->crtc_hsync_end;
+
+	if (adjusted_mode->crtc_hblank_start == crtc_hblank_start_sw)
+		adjusted_mode->crtc_hblank_start =
+					adjusted_mode_sw->crtc_hblank_start;
+
+	if (adjusted_mode->crtc_hblank_end == crtc_hblank_end_sw)
+		adjusted_mode->crtc_hblank_end =
+					adjusted_mode_sw->crtc_hblank_end;
+}
+
+static void intel_dsi_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_state *pipe_config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	u32 pclk;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	pipe_config->output_types |= BIT(INTEL_OUTPUT_DSI);
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		bxt_dsi_get_pipe_config(encoder, pipe_config);
+		pclk = bxt_dsi_get_pclk(encoder, pipe_config);
+	} else {
+		pclk = vlv_dsi_get_pclk(encoder, pipe_config);
+	}
+
+	pipe_config->port_clock = pclk;
+
+	/* FIXME definitely not right for burst/cmd mode/pixel overlap */
+	pipe_config->hw.adjusted_mode.crtc_clock = pclk;
+	if (intel_dsi->dual_link)
+		pipe_config->hw.adjusted_mode.crtc_clock *= 2;
+}
+
+/* return txclkesc cycles in terms of divider and duration in us */
+static u16 txclkesc(u32 divider, unsigned int us)
+{
+	switch (divider) {
+	case ESCAPE_CLOCK_DIVIDER_1:
+	default:
+		return 20 * us;
+	case ESCAPE_CLOCK_DIVIDER_2:
+		return 10 * us;
+	case ESCAPE_CLOCK_DIVIDER_4:
+		return 5 * us;
+	}
+}
+
+static void set_dsi_timings(struct drm_encoder *encoder,
+			    const struct drm_display_mode *adjusted_mode)
+{
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(to_intel_encoder(encoder));
+	enum port port;
+	unsigned int bpp = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format);
+	unsigned int lane_count = intel_dsi->lane_count;
+
+	u16 hactive, hfp, hsync, hbp, vfp, vsync, vbp;
+
+	hactive = adjusted_mode->crtc_hdisplay;
+	hfp = adjusted_mode->crtc_hsync_start - adjusted_mode->crtc_hdisplay;
+	hsync = adjusted_mode->crtc_hsync_end - adjusted_mode->crtc_hsync_start;
+	hbp = adjusted_mode->crtc_htotal - adjusted_mode->crtc_hsync_end;
+
+	if (intel_dsi->dual_link) {
+		hactive /= 2;
+		if (intel_dsi->dual_link == DSI_DUAL_LINK_FRONT_BACK)
+			hactive += intel_dsi->pixel_overlap;
+		hfp /= 2;
+		hsync /= 2;
+		hbp /= 2;
+	}
+
+	vfp = adjusted_mode->crtc_vsync_start - adjusted_mode->crtc_vdisplay;
+	vsync = adjusted_mode->crtc_vsync_end - adjusted_mode->crtc_vsync_start;
+	vbp = adjusted_mode->crtc_vtotal - adjusted_mode->crtc_vsync_end;
+
+	/* horizontal values are in terms of high speed byte clock */
+	hactive = txbyteclkhs(hactive, bpp, lane_count,
+			      intel_dsi->burst_mode_ratio);
+	hfp = txbyteclkhs(hfp, bpp, lane_count, intel_dsi->burst_mode_ratio);
+	hsync = txbyteclkhs(hsync, bpp, lane_count,
+			    intel_dsi->burst_mode_ratio);
+	hbp = txbyteclkhs(hbp, bpp, lane_count, intel_dsi->burst_mode_ratio);
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+			/*
+			 * Program hdisplay and vdisplay on MIPI transcoder.
+			 * This is different from calculated hactive and
+			 * vactive, as they are calculated per channel basis,
+			 * whereas these values should be based on resolution.
+			 */
+			intel_de_write(dev_priv, BXT_MIPI_TRANS_HACTIVE(port),
+				       adjusted_mode->crtc_hdisplay);
+			intel_de_write(dev_priv, BXT_MIPI_TRANS_VACTIVE(port),
+				       adjusted_mode->crtc_vdisplay);
+			intel_de_write(dev_priv, BXT_MIPI_TRANS_VTOTAL(port),
+				       adjusted_mode->crtc_vtotal);
+		}
+
+		intel_de_write(dev_priv, MIPI_HACTIVE_AREA_COUNT(port),
+			       hactive);
+		intel_de_write(dev_priv, MIPI_HFP_COUNT(port), hfp);
+
+		/* meaningful for video mode non-burst sync pulse mode only,
+		 * can be zero for non-burst sync events and burst modes */
+		intel_de_write(dev_priv, MIPI_HSYNC_PADDING_COUNT(port),
+			       hsync);
+		intel_de_write(dev_priv, MIPI_HBP_COUNT(port), hbp);
+
+		/* vertical values are in terms of lines */
+		intel_de_write(dev_priv, MIPI_VFP_COUNT(port), vfp);
+		intel_de_write(dev_priv, MIPI_VSYNC_PADDING_COUNT(port),
+			       vsync);
+		intel_de_write(dev_priv, MIPI_VBP_COUNT(port), vbp);
+	}
+}
+
+static u32 pixel_format_to_reg(enum mipi_dsi_pixel_format fmt)
+{
+	switch (fmt) {
+	case MIPI_DSI_FMT_RGB888:
+		return VID_MODE_FORMAT_RGB888;
+	case MIPI_DSI_FMT_RGB666:
+		return VID_MODE_FORMAT_RGB666;
+	case MIPI_DSI_FMT_RGB666_PACKED:
+		return VID_MODE_FORMAT_RGB666_PACKED;
+	case MIPI_DSI_FMT_RGB565:
+		return VID_MODE_FORMAT_RGB565;
+	default:
+		MISSING_CASE(fmt);
+		return VID_MODE_FORMAT_RGB666;
+	}
+}
+
+static void intel_dsi_prepare(struct intel_encoder *intel_encoder,
+			      const struct intel_crtc_state *pipe_config)
+{
+	struct drm_encoder *encoder = &intel_encoder->base;
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_crtc *crtc = to_intel_crtc(pipe_config->uapi.crtc);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(to_intel_encoder(encoder));
+	const struct drm_display_mode *adjusted_mode = &pipe_config->hw.adjusted_mode;
+	enum port port;
+	unsigned int bpp = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format);
+	u32 val, tmp;
+	u16 mode_hdisplay;
+
+	drm_dbg_kms(&dev_priv->drm, "pipe %c\n", pipe_name(crtc->pipe));
+
+	mode_hdisplay = adjusted_mode->crtc_hdisplay;
+
+	if (intel_dsi->dual_link) {
+		mode_hdisplay /= 2;
+		if (intel_dsi->dual_link == DSI_DUAL_LINK_FRONT_BACK)
+			mode_hdisplay += intel_dsi->pixel_overlap;
+	}
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+			/*
+			 * escape clock divider, 20MHz, shared for A and C.
+			 * device ready must be off when doing this! txclkesc?
+			 */
+			tmp = intel_de_read(dev_priv, MIPI_CTRL(PORT_A));
+			tmp &= ~ESCAPE_CLOCK_DIVIDER_MASK;
+			intel_de_write(dev_priv, MIPI_CTRL(PORT_A),
+				       tmp | ESCAPE_CLOCK_DIVIDER_1);
+
+			/* read request priority is per pipe */
+			tmp = intel_de_read(dev_priv, MIPI_CTRL(port));
+			tmp &= ~READ_REQUEST_PRIORITY_MASK;
+			intel_de_write(dev_priv, MIPI_CTRL(port),
+				       tmp | READ_REQUEST_PRIORITY_HIGH);
+		} else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+			enum pipe pipe = crtc->pipe;
+
+			tmp = intel_de_read(dev_priv, MIPI_CTRL(port));
+			tmp &= ~BXT_PIPE_SELECT_MASK;
+
+			tmp |= BXT_PIPE_SELECT(pipe);
+			intel_de_write(dev_priv, MIPI_CTRL(port), tmp);
+		}
+
+		/* XXX: why here, why like this? handling in irq handler?! */
+		intel_de_write(dev_priv, MIPI_INTR_STAT(port), 0xffffffff);
+		intel_de_write(dev_priv, MIPI_INTR_EN(port), 0xffffffff);
+
+		intel_de_write(dev_priv, MIPI_DPHY_PARAM(port),
+			       intel_dsi->dphy_reg);
+
+		intel_de_write(dev_priv, MIPI_DPI_RESOLUTION(port),
+			       adjusted_mode->crtc_vdisplay << VERTICAL_ADDRESS_SHIFT | mode_hdisplay << HORIZONTAL_ADDRESS_SHIFT);
+	}
+
+	set_dsi_timings(encoder, adjusted_mode);
+
+	val = intel_dsi->lane_count << DATA_LANES_PRG_REG_SHIFT;
+	if (is_cmd_mode(intel_dsi)) {
+		val |= intel_dsi->channel << CMD_MODE_CHANNEL_NUMBER_SHIFT;
+		val |= CMD_MODE_DATA_WIDTH_8_BIT; /* XXX */
+	} else {
+		val |= intel_dsi->channel << VID_MODE_CHANNEL_NUMBER_SHIFT;
+		val |= pixel_format_to_reg(intel_dsi->pixel_format);
+	}
+
+	tmp = 0;
+	if (intel_dsi->eotp_pkt == 0)
+		tmp |= EOT_DISABLE;
+	if (intel_dsi->clock_stop)
+		tmp |= CLOCKSTOP;
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+		tmp |= BXT_DPHY_DEFEATURE_EN;
+		if (!is_cmd_mode(intel_dsi))
+			tmp |= BXT_DEFEATURE_DPI_FIFO_CTR;
+	}
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		intel_de_write(dev_priv, MIPI_DSI_FUNC_PRG(port), val);
+
+		/* timeouts for recovery. one frame IIUC. if counter expires,
+		 * EOT and stop state. */
+
+		/*
+		 * In burst mode, value greater than one DPI line Time in byte
+		 * clock (txbyteclkhs) To timeout this timer 1+ of the above
+		 * said value is recommended.
+		 *
+		 * In non-burst mode, Value greater than one DPI frame time in
+		 * byte clock(txbyteclkhs) To timeout this timer 1+ of the above
+		 * said value is recommended.
+		 *
+		 * In DBI only mode, value greater than one DBI frame time in
+		 * byte clock(txbyteclkhs) To timeout this timer 1+ of the above
+		 * said value is recommended.
+		 */
+
+		if (is_vid_mode(intel_dsi) &&
+			intel_dsi->video_mode == BURST_MODE) {
+			intel_de_write(dev_priv, MIPI_HS_TX_TIMEOUT(port),
+				       txbyteclkhs(adjusted_mode->crtc_htotal, bpp, intel_dsi->lane_count, intel_dsi->burst_mode_ratio) + 1);
+		} else {
+			intel_de_write(dev_priv, MIPI_HS_TX_TIMEOUT(port),
+				       txbyteclkhs(adjusted_mode->crtc_vtotal * adjusted_mode->crtc_htotal, bpp, intel_dsi->lane_count, intel_dsi->burst_mode_ratio) + 1);
+		}
+		intel_de_write(dev_priv, MIPI_LP_RX_TIMEOUT(port),
+			       intel_dsi->lp_rx_timeout);
+		intel_de_write(dev_priv, MIPI_TURN_AROUND_TIMEOUT(port),
+			       intel_dsi->turn_arnd_val);
+		intel_de_write(dev_priv, MIPI_DEVICE_RESET_TIMER(port),
+			       intel_dsi->rst_timer_val);
+
+		/* dphy stuff */
+
+		/* in terms of low power clock */
+		intel_de_write(dev_priv, MIPI_INIT_COUNT(port),
+			       txclkesc(intel_dsi->escape_clk_div, 100));
+
+		if ((IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) &&
+		    !intel_dsi->dual_link) {
+			/*
+			 * BXT spec says write MIPI_INIT_COUNT for
+			 * both the ports, even if only one is
+			 * getting used. So write the other port
+			 * if not in dual link mode.
+			 */
+			intel_de_write(dev_priv,
+				       MIPI_INIT_COUNT(port == PORT_A ? PORT_C : PORT_A),
+				       intel_dsi->init_count);
+		}
+
+		/* recovery disables */
+		intel_de_write(dev_priv, MIPI_EOT_DISABLE(port), tmp);
+
+		/* in terms of low power clock */
+		intel_de_write(dev_priv, MIPI_INIT_COUNT(port),
+			       intel_dsi->init_count);
+
+		/* in terms of txbyteclkhs. actual high to low switch +
+		 * MIPI_STOP_STATE_STALL * MIPI_LP_BYTECLK.
+		 *
+		 * XXX: write MIPI_STOP_STATE_STALL?
+		 */
+		intel_de_write(dev_priv, MIPI_HIGH_LOW_SWITCH_COUNT(port),
+			       intel_dsi->hs_to_lp_count);
+
+		/* XXX: low power clock equivalence in terms of byte clock.
+		 * the number of byte clocks occupied in one low power clock.
+		 * based on txbyteclkhs and txclkesc.
+		 * txclkesc time / txbyteclk time * (105 + MIPI_STOP_STATE_STALL
+		 * ) / 105.???
+		 */
+		intel_de_write(dev_priv, MIPI_LP_BYTECLK(port),
+			       intel_dsi->lp_byte_clk);
+
+		if (IS_GEMINILAKE(dev_priv)) {
+			intel_de_write(dev_priv, MIPI_TLPX_TIME_COUNT(port),
+				       intel_dsi->lp_byte_clk);
+			/* Shadow of DPHY reg */
+			intel_de_write(dev_priv, MIPI_CLK_LANE_TIMING(port),
+				       intel_dsi->dphy_reg);
+		}
+
+		/* the bw essential for transmitting 16 long packets containing
+		 * 252 bytes meant for dcs write memory command is programmed in
+		 * this register in terms of byte clocks. based on dsi transfer
+		 * rate and the number of lanes configured the time taken to
+		 * transmit 16 long packets in a dsi stream varies. */
+		intel_de_write(dev_priv, MIPI_DBI_BW_CTRL(port),
+			       intel_dsi->bw_timer);
+
+		intel_de_write(dev_priv, MIPI_CLK_LANE_SWITCH_TIME_CNT(port),
+			       intel_dsi->clk_lp_to_hs_count << LP_HS_SSW_CNT_SHIFT | intel_dsi->clk_hs_to_lp_count << HS_LP_PWR_SW_CNT_SHIFT);
+
+		if (is_vid_mode(intel_dsi)) {
+			u32 fmt = intel_dsi->video_frmt_cfg_bits | IP_TG_CONFIG;
+
+			/*
+			 * Some panels might have resolution which is not a
+			 * multiple of 64 like 1366 x 768. Enable RANDOM
+			 * resolution support for such panels by default.
+			 */
+			fmt |= RANDOM_DPI_DISPLAY_RESOLUTION;
+
+			switch (intel_dsi->video_mode) {
+			default:
+				MISSING_CASE(intel_dsi->video_mode);
+				fallthrough;
+			case NON_BURST_SYNC_EVENTS:
+				fmt |= VIDEO_MODE_NON_BURST_WITH_SYNC_EVENTS;
+				break;
+			case NON_BURST_SYNC_PULSE:
+				fmt |= VIDEO_MODE_NON_BURST_WITH_SYNC_PULSE;
+				break;
+			case BURST_MODE:
+				fmt |= VIDEO_MODE_BURST;
+				break;
+			}
+
+			intel_de_write(dev_priv, MIPI_VIDEO_MODE_FORMAT(port), fmt);
+		}
+	}
+}
+
+static void intel_dsi_unprepare(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 val;
+
+	if (IS_GEMINILAKE(dev_priv))
+		return;
+
+	for_each_dsi_port(port, intel_dsi->ports) {
+		/* Panel commands can be sent when clock is in LP11 */
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port), 0x0);
+
+		if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+			bxt_dsi_reset_clocks(encoder, port);
+		else
+			vlv_dsi_reset_clocks(encoder, port);
+		intel_de_write(dev_priv, MIPI_EOT_DISABLE(port), CLOCKSTOP);
+
+		val = intel_de_read(dev_priv, MIPI_DSI_FUNC_PRG(port));
+		val &= ~VID_MODE_FORMAT_MASK;
+		intel_de_write(dev_priv, MIPI_DSI_FUNC_PRG(port), val);
+
+		intel_de_write(dev_priv, MIPI_DEVICE_READY(port), 0x1);
+	}
+}
+
+static void intel_dsi_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(to_intel_encoder(encoder));
+
+	intel_dsi_vbt_gpio_cleanup(intel_dsi);
+	intel_encoder_destroy(encoder);
+}
+
+static const struct drm_encoder_funcs intel_dsi_funcs = {
+	.destroy = intel_dsi_encoder_destroy,
+};
+
+static enum drm_mode_status vlv_dsi_mode_valid(struct drm_connector *connector,
+					       struct drm_display_mode *mode)
+{
+	struct drm_i915_private *i915 = to_i915(connector->dev);
+
+	if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) {
+		enum drm_mode_status status;
+
+		status = intel_cpu_transcoder_mode_valid(i915, mode);
+		if (status != MODE_OK)
+			return status;
+	}
+
+	return intel_dsi_mode_valid(connector, mode);
+}
+
+static const struct drm_connector_helper_funcs intel_dsi_connector_helper_funcs = {
+	.get_modes = intel_dsi_get_modes,
+	.mode_valid = vlv_dsi_mode_valid,
+	.atomic_check = intel_digital_connector_atomic_check,
+};
+
+static const struct drm_connector_funcs intel_dsi_connector_funcs = {
+	.detect = intel_panel_detect,
+	.late_register = intel_connector_register,
+	.early_unregister = intel_connector_unregister,
+	.destroy = intel_connector_destroy,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.atomic_get_property = intel_digital_connector_atomic_get_property,
+	.atomic_set_property = intel_digital_connector_atomic_set_property,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.atomic_duplicate_state = intel_digital_connector_duplicate_state,
+};
+
+static void vlv_dsi_add_properties(struct intel_connector *connector)
+{
+	struct drm_i915_private *dev_priv = to_i915(connector->base.dev);
+	const struct drm_display_mode *fixed_mode =
+		intel_panel_preferred_fixed_mode(connector);
+	u32 allowed_scalers;
+
+	allowed_scalers = BIT(DRM_MODE_SCALE_ASPECT) | BIT(DRM_MODE_SCALE_FULLSCREEN);
+	if (!HAS_GMCH(dev_priv))
+		allowed_scalers |= BIT(DRM_MODE_SCALE_CENTER);
+
+	drm_connector_attach_scaling_mode_property(&connector->base,
+						   allowed_scalers);
+
+	connector->base.state->scaling_mode = DRM_MODE_SCALE_ASPECT;
+
+	drm_connector_set_panel_orientation_with_quirk(&connector->base,
+						       intel_dsi_get_panel_orientation(connector),
+						       fixed_mode->hdisplay,
+						       fixed_mode->vdisplay);
+}
+
+#define NS_KHZ_RATIO		1000000
+
+#define PREPARE_CNT_MAX		0x3F
+#define EXIT_ZERO_CNT_MAX	0x3F
+#define CLK_ZERO_CNT_MAX	0xFF
+#define TRAIL_CNT_MAX		0x1F
+
+static void vlv_dphy_param_init(struct intel_dsi *intel_dsi)
+{
+	struct drm_device *dev = intel_dsi->base.base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_connector *connector = intel_dsi->attached_connector;
+	struct mipi_config *mipi_config = connector->panel.vbt.dsi.config;
+	u32 tlpx_ns, extra_byte_count, tlpx_ui;
+	u32 ui_num, ui_den;
+	u32 prepare_cnt, exit_zero_cnt, clk_zero_cnt, trail_cnt;
+	u32 ths_prepare_ns, tclk_trail_ns;
+	u32 tclk_prepare_clkzero, ths_prepare_hszero;
+	u32 lp_to_hs_switch, hs_to_lp_switch;
+	u32 mul;
+
+	tlpx_ns = intel_dsi_tlpx_ns(intel_dsi);
+
+	switch (intel_dsi->lane_count) {
+	case 1:
+	case 2:
+		extra_byte_count = 2;
+		break;
+	case 3:
+		extra_byte_count = 4;
+		break;
+	case 4:
+	default:
+		extra_byte_count = 3;
+		break;
+	}
+
+	/* in Kbps */
+	ui_num = NS_KHZ_RATIO;
+	ui_den = intel_dsi_bitrate(intel_dsi);
+
+	tclk_prepare_clkzero = mipi_config->tclk_prepare_clkzero;
+	ths_prepare_hszero = mipi_config->ths_prepare_hszero;
+
+	/*
+	 * B060
+	 * LP byte clock = TLPX/ (8UI)
+	 */
+	intel_dsi->lp_byte_clk = DIV_ROUND_UP(tlpx_ns * ui_den, 8 * ui_num);
+
+	/* DDR clock period = 2 * UI
+	 * UI(sec) = 1/(bitrate * 10^3) (bitrate is in KHZ)
+	 * UI(nsec) = 10^6 / bitrate
+	 * DDR clock period (nsec) = 2 * UI = (2 * 10^6)/ bitrate
+	 * DDR clock count  = ns_value / DDR clock period
+	 *
+	 * For GEMINILAKE dphy_param_reg will be programmed in terms of
+	 * HS byte clock count for other platform in HS ddr clock count
+	 */
+	mul = IS_GEMINILAKE(dev_priv) ? 8 : 2;
+	ths_prepare_ns = max(mipi_config->ths_prepare,
+			     mipi_config->tclk_prepare);
+
+	/* prepare count */
+	prepare_cnt = DIV_ROUND_UP(ths_prepare_ns * ui_den, ui_num * mul);
+
+	if (prepare_cnt > PREPARE_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm, "prepare count too high %u\n",
+			    prepare_cnt);
+		prepare_cnt = PREPARE_CNT_MAX;
+	}
+
+	/* exit zero count */
+	exit_zero_cnt = DIV_ROUND_UP(
+				(ths_prepare_hszero - ths_prepare_ns) * ui_den,
+				ui_num * mul
+				);
+
+	/*
+	 * Exit zero is unified val ths_zero and ths_exit
+	 * minimum value for ths_exit = 110ns
+	 * min (exit_zero_cnt * 2) = 110/UI
+	 * exit_zero_cnt = 55/UI
+	 */
+	if (exit_zero_cnt < (55 * ui_den / ui_num) && (55 * ui_den) % ui_num)
+		exit_zero_cnt += 1;
+
+	if (exit_zero_cnt > EXIT_ZERO_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm, "exit zero count too high %u\n",
+			    exit_zero_cnt);
+		exit_zero_cnt = EXIT_ZERO_CNT_MAX;
+	}
+
+	/* clk zero count */
+	clk_zero_cnt = DIV_ROUND_UP(
+				(tclk_prepare_clkzero -	ths_prepare_ns)
+				* ui_den, ui_num * mul);
+
+	if (clk_zero_cnt > CLK_ZERO_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm, "clock zero count too high %u\n",
+			    clk_zero_cnt);
+		clk_zero_cnt = CLK_ZERO_CNT_MAX;
+	}
+
+	/* trail count */
+	tclk_trail_ns = max(mipi_config->tclk_trail, mipi_config->ths_trail);
+	trail_cnt = DIV_ROUND_UP(tclk_trail_ns * ui_den, ui_num * mul);
+
+	if (trail_cnt > TRAIL_CNT_MAX) {
+		drm_dbg_kms(&dev_priv->drm, "trail count too high %u\n",
+			    trail_cnt);
+		trail_cnt = TRAIL_CNT_MAX;
+	}
+
+	/* B080 */
+	intel_dsi->dphy_reg = exit_zero_cnt << 24 | trail_cnt << 16 |
+						clk_zero_cnt << 8 | prepare_cnt;
+
+	/*
+	 * LP to HS switch count = 4TLPX + PREP_COUNT * mul + EXIT_ZERO_COUNT *
+	 *					mul + 10UI + Extra Byte Count
+	 *
+	 * HS to LP switch count = THS-TRAIL + 2TLPX + Extra Byte Count
+	 * Extra Byte Count is calculated according to number of lanes.
+	 * High Low Switch Count is the Max of LP to HS and
+	 * HS to LP switch count
+	 *
+	 */
+	tlpx_ui = DIV_ROUND_UP(tlpx_ns * ui_den, ui_num);
+
+	/* B044 */
+	/* FIXME:
+	 * The comment above does not match with the code */
+	lp_to_hs_switch = DIV_ROUND_UP(4 * tlpx_ui + prepare_cnt * mul +
+						exit_zero_cnt * mul + 10, 8);
+
+	hs_to_lp_switch = DIV_ROUND_UP(mipi_config->ths_trail + 2 * tlpx_ui, 8);
+
+	intel_dsi->hs_to_lp_count = max(lp_to_hs_switch, hs_to_lp_switch);
+	intel_dsi->hs_to_lp_count += extra_byte_count;
+
+	/* B088 */
+	/* LP -> HS for clock lanes
+	 * LP clk sync + LP11 + LP01 + tclk_prepare + tclk_zero +
+	 *						extra byte count
+	 * 2TPLX + 1TLPX + 1 TPLX(in ns) + prepare_cnt * 2 + clk_zero_cnt *
+	 *					2(in UI) + extra byte count
+	 * In byteclks = (4TLPX + prepare_cnt * 2 + clk_zero_cnt *2 (in UI)) /
+	 *					8 + extra byte count
+	 */
+	intel_dsi->clk_lp_to_hs_count =
+		DIV_ROUND_UP(
+			4 * tlpx_ui + prepare_cnt * 2 +
+			clk_zero_cnt * 2,
+			8);
+
+	intel_dsi->clk_lp_to_hs_count += extra_byte_count;
+
+	/* HS->LP for Clock Lanes
+	 * Low Power clock synchronisations + 1Tx byteclk + tclk_trail +
+	 *						Extra byte count
+	 * 2TLPX + 8UI + (trail_count*2)(in UI) + Extra byte count
+	 * In byteclks = (2*TLpx(in UI) + trail_count*2 +8)(in UI)/8 +
+	 *						Extra byte count
+	 */
+	intel_dsi->clk_hs_to_lp_count =
+		DIV_ROUND_UP(2 * tlpx_ui + trail_cnt * 2 + 8,
+			8);
+	intel_dsi->clk_hs_to_lp_count += extra_byte_count;
+
+	intel_dsi_log_params(intel_dsi);
+}
+
+void vlv_dsi_init(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_dsi *intel_dsi;
+	struct intel_encoder *intel_encoder;
+	struct drm_encoder *encoder;
+	struct intel_connector *intel_connector;
+	struct drm_connector *connector;
+	struct drm_display_mode *current_mode;
+	enum port port;
+	enum pipe pipe;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	/* There is no detection method for MIPI so rely on VBT */
+	if (!intel_bios_is_dsi_present(dev_priv, &port))
+		return;
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		dev_priv->display.dsi.mmio_base = BXT_MIPI_BASE;
+	else
+		dev_priv->display.dsi.mmio_base = VLV_MIPI_BASE;
+
+	intel_dsi = kzalloc(sizeof(*intel_dsi), GFP_KERNEL);
+	if (!intel_dsi)
+		return;
+
+	intel_connector = intel_connector_alloc();
+	if (!intel_connector) {
+		kfree(intel_dsi);
+		return;
+	}
+
+	intel_encoder = &intel_dsi->base;
+	encoder = &intel_encoder->base;
+	intel_dsi->attached_connector = intel_connector;
+
+	connector = &intel_connector->base;
+
+	drm_encoder_init(dev, encoder, &intel_dsi_funcs, DRM_MODE_ENCODER_DSI,
+			 "DSI %c", port_name(port));
+
+	intel_encoder->compute_config = intel_dsi_compute_config;
+	intel_encoder->pre_enable = intel_dsi_pre_enable;
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		intel_encoder->enable = bxt_dsi_enable;
+	intel_encoder->disable = intel_dsi_disable;
+	intel_encoder->post_disable = intel_dsi_post_disable;
+	intel_encoder->get_hw_state = intel_dsi_get_hw_state;
+	intel_encoder->get_config = intel_dsi_get_config;
+	intel_encoder->update_pipe = intel_backlight_update;
+	intel_encoder->shutdown = intel_dsi_shutdown;
+
+	intel_connector->get_hw_state = intel_connector_get_hw_state;
+
+	intel_encoder->port = port;
+	intel_encoder->type = INTEL_OUTPUT_DSI;
+	intel_encoder->power_domain = POWER_DOMAIN_PORT_DSI;
+	intel_encoder->cloneable = 0;
+
+	/*
+	 * On BYT/CHV, pipe A maps to MIPI DSI port A, pipe B maps to MIPI DSI
+	 * port C. BXT isn't limited like this.
+	 */
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		intel_encoder->pipe_mask = ~0;
+	else if (port == PORT_A)
+		intel_encoder->pipe_mask = BIT(PIPE_A);
+	else
+		intel_encoder->pipe_mask = BIT(PIPE_B);
+
+	intel_dsi->panel_power_off_time = ktime_get_boottime();
+
+	intel_bios_init_panel_late(dev_priv, &intel_connector->panel, NULL, NULL);
+
+	if (intel_connector->panel.vbt.dsi.config->dual_link)
+		intel_dsi->ports = BIT(PORT_A) | BIT(PORT_C);
+	else
+		intel_dsi->ports = BIT(port);
+
+	if (drm_WARN_ON(&dev_priv->drm, intel_connector->panel.vbt.dsi.bl_ports & ~intel_dsi->ports))
+		intel_connector->panel.vbt.dsi.bl_ports &= intel_dsi->ports;
+
+	if (drm_WARN_ON(&dev_priv->drm, intel_connector->panel.vbt.dsi.cabc_ports & ~intel_dsi->ports))
+		intel_connector->panel.vbt.dsi.cabc_ports &= intel_dsi->ports;
+
+	/* Create a DSI host (and a device) for each port. */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		struct intel_dsi_host *host;
+
+		host = intel_dsi_host_init(intel_dsi, &intel_dsi_host_ops,
+					   port);
+		if (!host)
+			goto err;
+
+		intel_dsi->dsi_hosts[port] = host;
+	}
+
+	if (!intel_dsi_vbt_init(intel_dsi, MIPI_DSI_GENERIC_PANEL_ID)) {
+		drm_dbg_kms(&dev_priv->drm, "no device found\n");
+		goto err;
+	}
+
+	/* Use clock read-back from current hw-state for fastboot */
+	current_mode = intel_encoder_current_mode(intel_encoder);
+	if (current_mode) {
+		drm_dbg_kms(&dev_priv->drm, "Calculated pclk %d GOP %d\n",
+			    intel_dsi->pclk, current_mode->clock);
+		if (intel_fuzzy_clock_check(intel_dsi->pclk,
+					    current_mode->clock)) {
+			drm_dbg_kms(&dev_priv->drm, "Using GOP pclk\n");
+			intel_dsi->pclk = current_mode->clock;
+		}
+
+		kfree(current_mode);
+	}
+
+	vlv_dphy_param_init(intel_dsi);
+
+	intel_dsi_vbt_gpio_init(intel_dsi,
+				intel_dsi_get_hw_state(intel_encoder, &pipe));
+
+	drm_connector_init(dev, connector, &intel_dsi_connector_funcs,
+			   DRM_MODE_CONNECTOR_DSI);
+
+	drm_connector_helper_add(connector, &intel_dsi_connector_helper_funcs);
+
+	connector->display_info.subpixel_order = SubPixelHorizontalRGB; /*XXX*/
+	connector->interlace_allowed = false;
+	connector->doublescan_allowed = false;
+
+	intel_connector_attach_encoder(intel_connector, intel_encoder);
+
+	mutex_lock(&dev->mode_config.mutex);
+	intel_panel_add_vbt_lfp_fixed_mode(intel_connector);
+	mutex_unlock(&dev->mode_config.mutex);
+
+	if (!intel_panel_preferred_fixed_mode(intel_connector)) {
+		drm_dbg_kms(&dev_priv->drm, "no fixed mode\n");
+		goto err_cleanup_connector;
+	}
+
+	intel_panel_init(intel_connector);
+
+	intel_backlight_setup(intel_connector, INVALID_PIPE);
+
+	vlv_dsi_add_properties(intel_connector);
+
+	return;
+
+err_cleanup_connector:
+	drm_connector_cleanup(&intel_connector->base);
+err:
+	drm_encoder_cleanup(&intel_encoder->base);
+	kfree(intel_dsi);
+	kfree(intel_connector);
+}
diff --git a/drivers/gpu/drm/i915/display/vlv_dsi.h b/drivers/gpu/drm/i915/display/vlv_dsi.h
new file mode 100644
index 000000000..0c2b279df
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/vlv_dsi.h
@@ -0,0 +1,19 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __VLV_DSI_H__
+#define __VLV_DSI_H__
+
+#include <linux/types.h>
+
+enum port;
+struct drm_i915_private;
+struct intel_dsi;
+
+void vlv_dsi_wait_for_fifo_empty(struct intel_dsi *intel_dsi, enum port port);
+enum mipi_dsi_pixel_format pixel_format_from_register_bits(u32 fmt);
+void vlv_dsi_init(struct drm_i915_private *dev_priv);
+
+#endif /* __VLV_DSI_H__ */
diff --git a/drivers/gpu/drm/i915/display/vlv_dsi_pll.c b/drivers/gpu/drm/i915/display/vlv_dsi_pll.c
new file mode 100644
index 000000000..af7402127
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/vlv_dsi_pll.c
@@ -0,0 +1,624 @@
+/*
+ * Copyright © 2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *	Shobhit Kumar <shobhit.kumar@intel.com>
+ *	Yogesh Mohan Marimuthu <yogesh.mohan.marimuthu@intel.com>
+ */
+
+#include <linux/kernel.h>
+#include <linux/string_helpers.h>
+
+#include "i915_drv.h"
+#include "intel_de.h"
+#include "intel_display_types.h"
+#include "intel_dsi.h"
+#include "vlv_dsi_pll.h"
+#include "vlv_dsi_pll_regs.h"
+#include "vlv_sideband.h"
+
+static const u16 lfsr_converts[] = {
+	426, 469, 234, 373, 442, 221, 110, 311, 411,		/* 62 - 70 */
+	461, 486, 243, 377, 188, 350, 175, 343, 427, 213,	/* 71 - 80 */
+	106, 53, 282, 397, 454, 227, 113, 56, 284, 142,		/* 81 - 90 */
+	71, 35, 273, 136, 324, 418, 465, 488, 500, 506		/* 91 - 100 */
+};
+
+/* Get DSI clock from pixel clock */
+static u32 dsi_clk_from_pclk(u32 pclk, enum mipi_dsi_pixel_format fmt,
+			     int lane_count)
+{
+	u32 dsi_clk_khz;
+	u32 bpp = mipi_dsi_pixel_format_to_bpp(fmt);
+
+	/* DSI data rate = pixel clock * bits per pixel / lane count
+	   pixel clock is converted from KHz to Hz */
+	dsi_clk_khz = DIV_ROUND_CLOSEST(pclk * bpp, lane_count);
+
+	return dsi_clk_khz;
+}
+
+static int dsi_calc_mnp(struct drm_i915_private *dev_priv,
+			struct intel_crtc_state *config,
+			int target_dsi_clk)
+{
+	unsigned int m_min, m_max, p_min = 2, p_max = 6;
+	unsigned int m, n, p;
+	unsigned int calc_m, calc_p;
+	int delta, ref_clk;
+
+	/* target_dsi_clk is expected in kHz */
+	if (target_dsi_clk < 300000 || target_dsi_clk > 1150000) {
+		drm_err(&dev_priv->drm, "DSI CLK Out of Range\n");
+		return -ECHRNG;
+	}
+
+	if (IS_CHERRYVIEW(dev_priv)) {
+		ref_clk = 100000;
+		n = 4;
+		m_min = 70;
+		m_max = 96;
+	} else {
+		ref_clk = 25000;
+		n = 1;
+		m_min = 62;
+		m_max = 92;
+	}
+
+	calc_p = p_min;
+	calc_m = m_min;
+	delta = abs(target_dsi_clk - (m_min * ref_clk) / (p_min * n));
+
+	for (m = m_min; m <= m_max && delta; m++) {
+		for (p = p_min; p <= p_max && delta; p++) {
+			/*
+			 * Find the optimal m and p divisors with minimal delta
+			 * +/- the required clock
+			 */
+			int calc_dsi_clk = (m * ref_clk) / (p * n);
+			int d = abs(target_dsi_clk - calc_dsi_clk);
+			if (d < delta) {
+				delta = d;
+				calc_m = m;
+				calc_p = p;
+			}
+		}
+	}
+
+	/* register has log2(N1), this works fine for powers of two */
+	config->dsi_pll.ctrl = 1 << (DSI_PLL_P1_POST_DIV_SHIFT + calc_p - 2);
+	config->dsi_pll.div =
+		(ffs(n) - 1) << DSI_PLL_N1_DIV_SHIFT |
+		(u32)lfsr_converts[calc_m - 62] << DSI_PLL_M1_DIV_SHIFT;
+
+	return 0;
+}
+
+static int vlv_dsi_pclk(struct intel_encoder *encoder,
+			struct intel_crtc_state *config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	int bpp = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format);
+	u32 dsi_clock;
+	u32 pll_ctl, pll_div;
+	u32 m = 0, p = 0, n;
+	int refclk = IS_CHERRYVIEW(dev_priv) ? 100000 : 25000;
+	int i;
+
+	pll_ctl = config->dsi_pll.ctrl;
+	pll_div = config->dsi_pll.div;
+
+	/* mask out other bits and extract the P1 divisor */
+	pll_ctl &= DSI_PLL_P1_POST_DIV_MASK;
+	pll_ctl = pll_ctl >> (DSI_PLL_P1_POST_DIV_SHIFT - 2);
+
+	/* N1 divisor */
+	n = (pll_div & DSI_PLL_N1_DIV_MASK) >> DSI_PLL_N1_DIV_SHIFT;
+	n = 1 << n; /* register has log2(N1) */
+
+	/* mask out the other bits and extract the M1 divisor */
+	pll_div &= DSI_PLL_M1_DIV_MASK;
+	pll_div = pll_div >> DSI_PLL_M1_DIV_SHIFT;
+
+	while (pll_ctl) {
+		pll_ctl = pll_ctl >> 1;
+		p++;
+	}
+	p--;
+
+	if (!p) {
+		drm_err(&dev_priv->drm, "wrong P1 divisor\n");
+		return 0;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(lfsr_converts); i++) {
+		if (lfsr_converts[i] == pll_div)
+			break;
+	}
+
+	if (i == ARRAY_SIZE(lfsr_converts)) {
+		drm_err(&dev_priv->drm, "wrong m_seed programmed\n");
+		return 0;
+	}
+
+	m = i + 62;
+
+	dsi_clock = (m * refclk) / (p * n);
+
+	return DIV_ROUND_CLOSEST(dsi_clock * intel_dsi->lane_count, bpp);
+}
+
+/*
+ * XXX: The muxing and gating is hard coded for now. Need to add support for
+ * sharing PLLs with two DSI outputs.
+ */
+int vlv_dsi_pll_compute(struct intel_encoder *encoder,
+			struct intel_crtc_state *config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	int pclk, dsi_clk, ret;
+
+	dsi_clk = dsi_clk_from_pclk(intel_dsi->pclk, intel_dsi->pixel_format,
+				    intel_dsi->lane_count);
+
+	ret = dsi_calc_mnp(dev_priv, config, dsi_clk);
+	if (ret) {
+		drm_dbg_kms(&dev_priv->drm, "dsi_calc_mnp failed\n");
+		return ret;
+	}
+
+	if (intel_dsi->ports & (1 << PORT_A))
+		config->dsi_pll.ctrl |= DSI_PLL_CLK_GATE_DSI0_DSIPLL;
+
+	if (intel_dsi->ports & (1 << PORT_C))
+		config->dsi_pll.ctrl |= DSI_PLL_CLK_GATE_DSI1_DSIPLL;
+
+	config->dsi_pll.ctrl |= DSI_PLL_VCO_EN;
+
+	drm_dbg_kms(&dev_priv->drm, "dsi pll div %08x, ctrl %08x\n",
+		    config->dsi_pll.div, config->dsi_pll.ctrl);
+
+	pclk = vlv_dsi_pclk(encoder, config);
+	config->port_clock = pclk;
+
+	/* FIXME definitely not right for burst/cmd mode/pixel overlap */
+	config->hw.adjusted_mode.crtc_clock = pclk;
+	if (intel_dsi->dual_link)
+		config->hw.adjusted_mode.crtc_clock *= 2;
+
+	return 0;
+}
+
+void vlv_dsi_pll_enable(struct intel_encoder *encoder,
+			const struct intel_crtc_state *config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	vlv_cck_get(dev_priv);
+
+	vlv_cck_write(dev_priv, CCK_REG_DSI_PLL_CONTROL, 0);
+	vlv_cck_write(dev_priv, CCK_REG_DSI_PLL_DIVIDER, config->dsi_pll.div);
+	vlv_cck_write(dev_priv, CCK_REG_DSI_PLL_CONTROL,
+		      config->dsi_pll.ctrl & ~DSI_PLL_VCO_EN);
+
+	/* wait at least 0.5 us after ungating before enabling VCO,
+	 * allow hrtimer subsystem optimization by relaxing timing
+	 */
+	usleep_range(10, 50);
+
+	vlv_cck_write(dev_priv, CCK_REG_DSI_PLL_CONTROL, config->dsi_pll.ctrl);
+
+	if (wait_for(vlv_cck_read(dev_priv, CCK_REG_DSI_PLL_CONTROL) &
+						DSI_PLL_LOCK, 20)) {
+
+		vlv_cck_put(dev_priv);
+		drm_err(&dev_priv->drm, "DSI PLL lock failed\n");
+		return;
+	}
+	vlv_cck_put(dev_priv);
+
+	drm_dbg_kms(&dev_priv->drm, "DSI PLL locked\n");
+}
+
+void vlv_dsi_pll_disable(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 tmp;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	vlv_cck_get(dev_priv);
+
+	tmp = vlv_cck_read(dev_priv, CCK_REG_DSI_PLL_CONTROL);
+	tmp &= ~DSI_PLL_VCO_EN;
+	tmp |= DSI_PLL_LDO_GATE;
+	vlv_cck_write(dev_priv, CCK_REG_DSI_PLL_CONTROL, tmp);
+
+	vlv_cck_put(dev_priv);
+}
+
+bool bxt_dsi_pll_is_enabled(struct drm_i915_private *dev_priv)
+{
+	bool enabled;
+	u32 val;
+	u32 mask;
+
+	mask = BXT_DSI_PLL_DO_ENABLE | BXT_DSI_PLL_LOCKED;
+	val = intel_de_read(dev_priv, BXT_DSI_PLL_ENABLE);
+	enabled = (val & mask) == mask;
+
+	if (!enabled)
+		return false;
+
+	/*
+	 * Dividers must be programmed with valid values. As per BSEPC, for
+	 * GEMINLAKE only PORT A divider values are checked while for BXT
+	 * both divider values are validated. Check this here for
+	 * paranoia, since BIOS is known to misconfigure PLLs in this way at
+	 * times, and since accessing DSI registers with invalid dividers
+	 * causes a system hang.
+	 */
+	val = intel_de_read(dev_priv, BXT_DSI_PLL_CTL);
+	if (IS_GEMINILAKE(dev_priv)) {
+		if (!(val & BXT_DSIA_16X_MASK)) {
+			drm_dbg(&dev_priv->drm,
+				"Invalid PLL divider (%08x)\n", val);
+			enabled = false;
+		}
+	} else {
+		if (!(val & BXT_DSIA_16X_MASK) || !(val & BXT_DSIC_16X_MASK)) {
+			drm_dbg(&dev_priv->drm,
+				"Invalid PLL divider (%08x)\n", val);
+			enabled = false;
+		}
+	}
+
+	return enabled;
+}
+
+void bxt_dsi_pll_disable(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 val;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	val = intel_de_read(dev_priv, BXT_DSI_PLL_ENABLE);
+	val &= ~BXT_DSI_PLL_DO_ENABLE;
+	intel_de_write(dev_priv, BXT_DSI_PLL_ENABLE, val);
+
+	/*
+	 * PLL lock should deassert within 200us.
+	 * Wait up to 1ms before timing out.
+	 */
+	if (intel_de_wait_for_clear(dev_priv, BXT_DSI_PLL_ENABLE,
+				    BXT_DSI_PLL_LOCKED, 1))
+		drm_err(&dev_priv->drm,
+			"Timeout waiting for PLL lock deassertion\n");
+}
+
+u32 vlv_dsi_get_pclk(struct intel_encoder *encoder,
+		     struct intel_crtc_state *config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 pll_ctl, pll_div;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	vlv_cck_get(dev_priv);
+	pll_ctl = vlv_cck_read(dev_priv, CCK_REG_DSI_PLL_CONTROL);
+	pll_div = vlv_cck_read(dev_priv, CCK_REG_DSI_PLL_DIVIDER);
+	vlv_cck_put(dev_priv);
+
+	config->dsi_pll.ctrl = pll_ctl & ~DSI_PLL_LOCK;
+	config->dsi_pll.div = pll_div;
+
+	return vlv_dsi_pclk(encoder, config);
+}
+
+static int bxt_dsi_pclk(struct intel_encoder *encoder,
+			const struct intel_crtc_state *config)
+{
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	int bpp = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format);
+	u32 dsi_ratio, dsi_clk;
+
+	dsi_ratio = config->dsi_pll.ctrl & BXT_DSI_PLL_RATIO_MASK;
+	dsi_clk = (dsi_ratio * BXT_REF_CLOCK_KHZ) / 2;
+
+	return DIV_ROUND_CLOSEST(dsi_clk * intel_dsi->lane_count, bpp);
+}
+
+u32 bxt_dsi_get_pclk(struct intel_encoder *encoder,
+		     struct intel_crtc_state *config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	u32 pclk;
+
+	config->dsi_pll.ctrl = intel_de_read(dev_priv, BXT_DSI_PLL_CTL);
+
+	pclk = bxt_dsi_pclk(encoder, config);
+
+	drm_dbg(&dev_priv->drm, "Calculated pclk=%u\n", pclk);
+	return pclk;
+}
+
+void vlv_dsi_reset_clocks(struct intel_encoder *encoder, enum port port)
+{
+	u32 temp;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+
+	temp = intel_de_read(dev_priv, MIPI_CTRL(port));
+	temp &= ~ESCAPE_CLOCK_DIVIDER_MASK;
+	intel_de_write(dev_priv, MIPI_CTRL(port),
+		       temp | intel_dsi->escape_clk_div << ESCAPE_CLOCK_DIVIDER_SHIFT);
+}
+
+static void glk_dsi_program_esc_clock(struct drm_device *dev,
+				   const struct intel_crtc_state *config)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 dsi_rate = 0;
+	u32 pll_ratio = 0;
+	u32 ddr_clk = 0;
+	u32 div1_value = 0;
+	u32 div2_value = 0;
+	u32 txesc1_div = 0;
+	u32 txesc2_div = 0;
+
+	pll_ratio = config->dsi_pll.ctrl & BXT_DSI_PLL_RATIO_MASK;
+
+	dsi_rate = (BXT_REF_CLOCK_KHZ * pll_ratio) / 2;
+
+	ddr_clk = dsi_rate / 2;
+
+	/* Variable divider value */
+	div1_value = DIV_ROUND_CLOSEST(ddr_clk, 20000);
+
+	/* Calculate TXESC1 divider */
+	if (div1_value <= 10)
+		txesc1_div = div1_value;
+	else if ((div1_value > 10) && (div1_value <= 20))
+		txesc1_div = DIV_ROUND_UP(div1_value, 2);
+	else if ((div1_value > 20) && (div1_value <= 30))
+		txesc1_div = DIV_ROUND_UP(div1_value, 4);
+	else if ((div1_value > 30) && (div1_value <= 40))
+		txesc1_div = DIV_ROUND_UP(div1_value, 6);
+	else if ((div1_value > 40) && (div1_value <= 50))
+		txesc1_div = DIV_ROUND_UP(div1_value, 8);
+	else
+		txesc1_div = 10;
+
+	/* Calculate TXESC2 divider */
+	div2_value = DIV_ROUND_UP(div1_value, txesc1_div);
+
+	txesc2_div = min_t(u32, div2_value, 10);
+
+	intel_de_write(dev_priv, MIPIO_TXESC_CLK_DIV1,
+		       (1 << (txesc1_div - 1)) & GLK_TX_ESC_CLK_DIV1_MASK);
+	intel_de_write(dev_priv, MIPIO_TXESC_CLK_DIV2,
+		       (1 << (txesc2_div - 1)) & GLK_TX_ESC_CLK_DIV2_MASK);
+}
+
+/* Program BXT Mipi clocks and dividers */
+static void bxt_dsi_program_clocks(struct drm_device *dev, enum port port,
+				   const struct intel_crtc_state *config)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 tmp;
+	u32 dsi_rate = 0;
+	u32 pll_ratio = 0;
+	u32 rx_div;
+	u32 tx_div;
+	u32 rx_div_upper;
+	u32 rx_div_lower;
+	u32 mipi_8by3_divider;
+
+	/* Clear old configurations */
+	tmp = intel_de_read(dev_priv, BXT_MIPI_CLOCK_CTL);
+	tmp &= ~(BXT_MIPI_TX_ESCLK_FIXDIV_MASK(port));
+	tmp &= ~(BXT_MIPI_RX_ESCLK_UPPER_FIXDIV_MASK(port));
+	tmp &= ~(BXT_MIPI_8X_BY3_DIVIDER_MASK(port));
+	tmp &= ~(BXT_MIPI_RX_ESCLK_LOWER_FIXDIV_MASK(port));
+
+	/* Get the current DSI rate(actual) */
+	pll_ratio = config->dsi_pll.ctrl & BXT_DSI_PLL_RATIO_MASK;
+	dsi_rate = (BXT_REF_CLOCK_KHZ * pll_ratio) / 2;
+
+	/*
+	 * tx clock should be <= 20MHz and the div value must be
+	 * subtracted by 1 as per bspec
+	 */
+	tx_div = DIV_ROUND_UP(dsi_rate, 20000) - 1;
+	/*
+	 * rx clock should be <= 150MHz and the div value must be
+	 * subtracted by 1 as per bspec
+	 */
+	rx_div = DIV_ROUND_UP(dsi_rate, 150000) - 1;
+
+	/*
+	 * rx divider value needs to be updated in the
+	 * two differnt bit fields in the register hence splitting the
+	 * rx divider value accordingly
+	 */
+	rx_div_lower = rx_div & RX_DIVIDER_BIT_1_2;
+	rx_div_upper = (rx_div & RX_DIVIDER_BIT_3_4) >> 2;
+
+	mipi_8by3_divider = 0x2;
+
+	tmp |= BXT_MIPI_8X_BY3_DIVIDER(port, mipi_8by3_divider);
+	tmp |= BXT_MIPI_TX_ESCLK_DIVIDER(port, tx_div);
+	tmp |= BXT_MIPI_RX_ESCLK_LOWER_DIVIDER(port, rx_div_lower);
+	tmp |= BXT_MIPI_RX_ESCLK_UPPER_DIVIDER(port, rx_div_upper);
+
+	intel_de_write(dev_priv, BXT_MIPI_CLOCK_CTL, tmp);
+}
+
+int bxt_dsi_pll_compute(struct intel_encoder *encoder,
+			struct intel_crtc_state *config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	u8 dsi_ratio, dsi_ratio_min, dsi_ratio_max;
+	u32 dsi_clk;
+	int pclk;
+
+	dsi_clk = dsi_clk_from_pclk(intel_dsi->pclk, intel_dsi->pixel_format,
+				    intel_dsi->lane_count);
+
+	/*
+	 * From clock diagram, to get PLL ratio divider, divide double of DSI
+	 * link rate (i.e., 2*8x=16x frequency value) by ref clock. Make sure to
+	 * round 'up' the result
+	 */
+	dsi_ratio = DIV_ROUND_UP(dsi_clk * 2, BXT_REF_CLOCK_KHZ);
+
+	if (IS_BROXTON(dev_priv)) {
+		dsi_ratio_min = BXT_DSI_PLL_RATIO_MIN;
+		dsi_ratio_max = BXT_DSI_PLL_RATIO_MAX;
+	} else {
+		dsi_ratio_min = GLK_DSI_PLL_RATIO_MIN;
+		dsi_ratio_max = GLK_DSI_PLL_RATIO_MAX;
+	}
+
+	if (dsi_ratio < dsi_ratio_min || dsi_ratio > dsi_ratio_max) {
+		drm_err(&dev_priv->drm,
+			"Can't get a suitable ratio from DSI PLL ratios\n");
+		return -ECHRNG;
+	} else
+		drm_dbg_kms(&dev_priv->drm, "DSI PLL calculation is Done!!\n");
+
+	/*
+	 * Program DSI ratio and Select MIPIC and MIPIA PLL output as 8x
+	 * Spec says both have to be programmed, even if one is not getting
+	 * used. Configure MIPI_CLOCK_CTL dividers in modeset
+	 */
+	config->dsi_pll.ctrl = dsi_ratio | BXT_DSIA_16X_BY2 | BXT_DSIC_16X_BY2;
+
+	/* As per recommendation from hardware team,
+	 * Prog PVD ratio =1 if dsi ratio <= 50
+	 */
+	if (IS_BROXTON(dev_priv) && dsi_ratio <= 50)
+		config->dsi_pll.ctrl |= BXT_DSI_PLL_PVD_RATIO_1;
+
+	pclk = bxt_dsi_pclk(encoder, config);
+	config->port_clock = pclk;
+
+	/* FIXME definitely not right for burst/cmd mode/pixel overlap */
+	config->hw.adjusted_mode.crtc_clock = pclk;
+	if (intel_dsi->dual_link)
+		config->hw.adjusted_mode.crtc_clock *= 2;
+
+	return 0;
+}
+
+void bxt_dsi_pll_enable(struct intel_encoder *encoder,
+			const struct intel_crtc_state *config)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(encoder);
+	enum port port;
+	u32 val;
+
+	drm_dbg_kms(&dev_priv->drm, "\n");
+
+	/* Configure PLL vales */
+	intel_de_write(dev_priv, BXT_DSI_PLL_CTL, config->dsi_pll.ctrl);
+	intel_de_posting_read(dev_priv, BXT_DSI_PLL_CTL);
+
+	/* Program TX, RX, Dphy clocks */
+	if (IS_BROXTON(dev_priv)) {
+		for_each_dsi_port(port, intel_dsi->ports)
+			bxt_dsi_program_clocks(encoder->base.dev, port, config);
+	} else {
+		glk_dsi_program_esc_clock(encoder->base.dev, config);
+	}
+
+	/* Enable DSI PLL */
+	val = intel_de_read(dev_priv, BXT_DSI_PLL_ENABLE);
+	val |= BXT_DSI_PLL_DO_ENABLE;
+	intel_de_write(dev_priv, BXT_DSI_PLL_ENABLE, val);
+
+	/* Timeout and fail if PLL not locked */
+	if (intel_de_wait_for_set(dev_priv, BXT_DSI_PLL_ENABLE,
+				  BXT_DSI_PLL_LOCKED, 1)) {
+		drm_err(&dev_priv->drm,
+			"Timed out waiting for DSI PLL to lock\n");
+		return;
+	}
+
+	drm_dbg_kms(&dev_priv->drm, "DSI PLL locked\n");
+}
+
+void bxt_dsi_reset_clocks(struct intel_encoder *encoder, enum port port)
+{
+	u32 tmp;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	/* Clear old configurations */
+	if (IS_BROXTON(dev_priv)) {
+		tmp = intel_de_read(dev_priv, BXT_MIPI_CLOCK_CTL);
+		tmp &= ~(BXT_MIPI_TX_ESCLK_FIXDIV_MASK(port));
+		tmp &= ~(BXT_MIPI_RX_ESCLK_UPPER_FIXDIV_MASK(port));
+		tmp &= ~(BXT_MIPI_8X_BY3_DIVIDER_MASK(port));
+		tmp &= ~(BXT_MIPI_RX_ESCLK_LOWER_FIXDIV_MASK(port));
+		intel_de_write(dev_priv, BXT_MIPI_CLOCK_CTL, tmp);
+	} else {
+		tmp = intel_de_read(dev_priv, MIPIO_TXESC_CLK_DIV1);
+		tmp &= ~GLK_TX_ESC_CLK_DIV1_MASK;
+		intel_de_write(dev_priv, MIPIO_TXESC_CLK_DIV1, tmp);
+
+		tmp = intel_de_read(dev_priv, MIPIO_TXESC_CLK_DIV2);
+		tmp &= ~GLK_TX_ESC_CLK_DIV2_MASK;
+		intel_de_write(dev_priv, MIPIO_TXESC_CLK_DIV2, tmp);
+	}
+	intel_de_write(dev_priv, MIPI_EOT_DISABLE(port), CLOCKSTOP);
+}
+
+static void assert_dsi_pll(struct drm_i915_private *i915, bool state)
+{
+	bool cur_state;
+
+	vlv_cck_get(i915);
+	cur_state = vlv_cck_read(i915, CCK_REG_DSI_PLL_CONTROL) & DSI_PLL_VCO_EN;
+	vlv_cck_put(i915);
+
+	I915_STATE_WARN(cur_state != state,
+			"DSI PLL state assertion failure (expected %s, current %s)\n",
+			str_on_off(state), str_on_off(cur_state));
+}
+
+void assert_dsi_pll_enabled(struct drm_i915_private *i915)
+{
+	assert_dsi_pll(i915, true);
+}
+
+void assert_dsi_pll_disabled(struct drm_i915_private *i915)
+{
+	assert_dsi_pll(i915, false);
+}
diff --git a/drivers/gpu/drm/i915/display/vlv_dsi_pll.h b/drivers/gpu/drm/i915/display/vlv_dsi_pll.h
new file mode 100644
index 000000000..ab9291ad1
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/vlv_dsi_pll.h
@@ -0,0 +1,38 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __VLV_DSI_PLL_H__
+#define __VLV_DSI_PLL_H__
+
+#include <linux/types.h>
+
+enum port;
+struct drm_i915_private;
+struct intel_crtc_state;
+struct intel_encoder;
+
+int vlv_dsi_pll_compute(struct intel_encoder *encoder,
+			struct intel_crtc_state *config);
+void vlv_dsi_pll_enable(struct intel_encoder *encoder,
+			const struct intel_crtc_state *config);
+void vlv_dsi_pll_disable(struct intel_encoder *encoder);
+u32 vlv_dsi_get_pclk(struct intel_encoder *encoder,
+		     struct intel_crtc_state *config);
+void vlv_dsi_reset_clocks(struct intel_encoder *encoder, enum port port);
+
+bool bxt_dsi_pll_is_enabled(struct drm_i915_private *dev_priv);
+int bxt_dsi_pll_compute(struct intel_encoder *encoder,
+			struct intel_crtc_state *config);
+void bxt_dsi_pll_enable(struct intel_encoder *encoder,
+			const struct intel_crtc_state *config);
+void bxt_dsi_pll_disable(struct intel_encoder *encoder);
+u32 bxt_dsi_get_pclk(struct intel_encoder *encoder,
+		     struct intel_crtc_state *config);
+void bxt_dsi_reset_clocks(struct intel_encoder *encoder, enum port port);
+
+void assert_dsi_pll_enabled(struct drm_i915_private *i915);
+void assert_dsi_pll_disabled(struct drm_i915_private *i915);
+
+#endif /* __VLV_DSI_PLL_H__ */
diff --git a/drivers/gpu/drm/i915/display/vlv_dsi_pll_regs.h b/drivers/gpu/drm/i915/display/vlv_dsi_pll_regs.h
new file mode 100644
index 000000000..45590e14e
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/vlv_dsi_pll_regs.h
@@ -0,0 +1,109 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __VLV_DSI_PLL_REGS_H__
+#define __VLV_DSI_PLL_REGS_H__
+
+#include "vlv_dsi_regs.h"
+
+#define MIPIO_TXESC_CLK_DIV1			_MMIO(0x160004)
+#define  GLK_TX_ESC_CLK_DIV1_MASK			0x3FF
+#define MIPIO_TXESC_CLK_DIV2			_MMIO(0x160008)
+#define  GLK_TX_ESC_CLK_DIV2_MASK			0x3FF
+
+#define BXT_MAX_VAR_OUTPUT_KHZ			39500
+
+#define BXT_MIPI_CLOCK_CTL			_MMIO(0x46090)
+#define  BXT_MIPI1_DIV_SHIFT			26
+#define  BXT_MIPI2_DIV_SHIFT			10
+#define  BXT_MIPI_DIV_SHIFT(port)		\
+			_MIPI_PORT(port, BXT_MIPI1_DIV_SHIFT, \
+					BXT_MIPI2_DIV_SHIFT)
+
+/* TX control divider to select actual TX clock output from (8x/var) */
+#define  BXT_MIPI1_TX_ESCLK_SHIFT		26
+#define  BXT_MIPI2_TX_ESCLK_SHIFT		10
+#define  BXT_MIPI_TX_ESCLK_SHIFT(port)		\
+			_MIPI_PORT(port, BXT_MIPI1_TX_ESCLK_SHIFT, \
+					BXT_MIPI2_TX_ESCLK_SHIFT)
+#define  BXT_MIPI1_TX_ESCLK_FIXDIV_MASK		(0x3F << 26)
+#define  BXT_MIPI2_TX_ESCLK_FIXDIV_MASK		(0x3F << 10)
+#define  BXT_MIPI_TX_ESCLK_FIXDIV_MASK(port)	\
+			_MIPI_PORT(port, BXT_MIPI1_TX_ESCLK_FIXDIV_MASK, \
+					BXT_MIPI2_TX_ESCLK_FIXDIV_MASK)
+#define  BXT_MIPI_TX_ESCLK_DIVIDER(port, val)	\
+		(((val) & 0x3F) << BXT_MIPI_TX_ESCLK_SHIFT(port))
+/* RX upper control divider to select actual RX clock output from 8x */
+#define  BXT_MIPI1_RX_ESCLK_UPPER_SHIFT		21
+#define  BXT_MIPI2_RX_ESCLK_UPPER_SHIFT		5
+#define  BXT_MIPI_RX_ESCLK_UPPER_SHIFT(port)		\
+			_MIPI_PORT(port, BXT_MIPI1_RX_ESCLK_UPPER_SHIFT, \
+					BXT_MIPI2_RX_ESCLK_UPPER_SHIFT)
+#define  BXT_MIPI1_RX_ESCLK_UPPER_FIXDIV_MASK		(3 << 21)
+#define  BXT_MIPI2_RX_ESCLK_UPPER_FIXDIV_MASK		(3 << 5)
+#define  BXT_MIPI_RX_ESCLK_UPPER_FIXDIV_MASK(port)	\
+			_MIPI_PORT(port, BXT_MIPI1_RX_ESCLK_UPPER_FIXDIV_MASK, \
+					BXT_MIPI2_RX_ESCLK_UPPER_FIXDIV_MASK)
+#define  BXT_MIPI_RX_ESCLK_UPPER_DIVIDER(port, val)	\
+		(((val) & 3) << BXT_MIPI_RX_ESCLK_UPPER_SHIFT(port))
+/* 8/3X divider to select the actual 8/3X clock output from 8x */
+#define  BXT_MIPI1_8X_BY3_SHIFT                19
+#define  BXT_MIPI2_8X_BY3_SHIFT                3
+#define  BXT_MIPI_8X_BY3_SHIFT(port)          \
+			_MIPI_PORT(port, BXT_MIPI1_8X_BY3_SHIFT, \
+					BXT_MIPI2_8X_BY3_SHIFT)
+#define  BXT_MIPI1_8X_BY3_DIVIDER_MASK         (3 << 19)
+#define  BXT_MIPI2_8X_BY3_DIVIDER_MASK         (3 << 3)
+#define  BXT_MIPI_8X_BY3_DIVIDER_MASK(port)    \
+			_MIPI_PORT(port, BXT_MIPI1_8X_BY3_DIVIDER_MASK, \
+						BXT_MIPI2_8X_BY3_DIVIDER_MASK)
+#define  BXT_MIPI_8X_BY3_DIVIDER(port, val)    \
+			(((val) & 3) << BXT_MIPI_8X_BY3_SHIFT(port))
+/* RX lower control divider to select actual RX clock output from 8x */
+#define  BXT_MIPI1_RX_ESCLK_LOWER_SHIFT		16
+#define  BXT_MIPI2_RX_ESCLK_LOWER_SHIFT		0
+#define  BXT_MIPI_RX_ESCLK_LOWER_SHIFT(port)		\
+			_MIPI_PORT(port, BXT_MIPI1_RX_ESCLK_LOWER_SHIFT, \
+					BXT_MIPI2_RX_ESCLK_LOWER_SHIFT)
+#define  BXT_MIPI1_RX_ESCLK_LOWER_FIXDIV_MASK		(3 << 16)
+#define  BXT_MIPI2_RX_ESCLK_LOWER_FIXDIV_MASK		(3 << 0)
+#define  BXT_MIPI_RX_ESCLK_LOWER_FIXDIV_MASK(port)	\
+			_MIPI_PORT(port, BXT_MIPI1_RX_ESCLK_LOWER_FIXDIV_MASK, \
+					BXT_MIPI2_RX_ESCLK_LOWER_FIXDIV_MASK)
+#define  BXT_MIPI_RX_ESCLK_LOWER_DIVIDER(port, val)	\
+		(((val) & 3) << BXT_MIPI_RX_ESCLK_LOWER_SHIFT(port))
+
+#define RX_DIVIDER_BIT_1_2                     0x3
+#define RX_DIVIDER_BIT_3_4                     0xC
+
+#define BXT_DSI_PLL_CTL			_MMIO(0x161000)
+#define  BXT_DSI_PLL_PVD_RATIO_SHIFT	16
+#define  BXT_DSI_PLL_PVD_RATIO_MASK	(3 << BXT_DSI_PLL_PVD_RATIO_SHIFT)
+#define  BXT_DSI_PLL_PVD_RATIO_1	(1 << BXT_DSI_PLL_PVD_RATIO_SHIFT)
+#define  BXT_DSIC_16X_BY1		(0 << 10)
+#define  BXT_DSIC_16X_BY2		(1 << 10)
+#define  BXT_DSIC_16X_BY3		(2 << 10)
+#define  BXT_DSIC_16X_BY4		(3 << 10)
+#define  BXT_DSIC_16X_MASK		(3 << 10)
+#define  BXT_DSIA_16X_BY1		(0 << 8)
+#define  BXT_DSIA_16X_BY2		(1 << 8)
+#define  BXT_DSIA_16X_BY3		(2 << 8)
+#define  BXT_DSIA_16X_BY4		(3 << 8)
+#define  BXT_DSIA_16X_MASK		(3 << 8)
+#define  BXT_DSI_FREQ_SEL_SHIFT		8
+#define  BXT_DSI_FREQ_SEL_MASK		(0xF << BXT_DSI_FREQ_SEL_SHIFT)
+
+#define BXT_DSI_PLL_RATIO_MAX		0x7D
+#define BXT_DSI_PLL_RATIO_MIN		0x22
+#define GLK_DSI_PLL_RATIO_MAX		0x6F
+#define GLK_DSI_PLL_RATIO_MIN		0x22
+#define BXT_DSI_PLL_RATIO_MASK		0xFF
+#define BXT_REF_CLOCK_KHZ		19200
+
+#define BXT_DSI_PLL_ENABLE		_MMIO(0x46080)
+#define  BXT_DSI_PLL_DO_ENABLE		(1 << 31)
+#define  BXT_DSI_PLL_LOCKED		(1 << 30)
+
+#endif /* __VLV_DSI_PLL_REGS_H__ */
diff --git a/drivers/gpu/drm/i915/display/vlv_dsi_regs.h b/drivers/gpu/drm/i915/display/vlv_dsi_regs.h
new file mode 100644
index 000000000..e065b8f2e
--- /dev/null
+++ b/drivers/gpu/drm/i915/display/vlv_dsi_regs.h
@@ -0,0 +1,482 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __VLV_DSI_REGS_H__
+#define __VLV_DSI_REGS_H__
+
+#include "i915_reg_defs.h"
+
+#define VLV_MIPI_BASE			VLV_DISPLAY_BASE
+#define BXT_MIPI_BASE			0x60000
+
+#define _MIPI_MMIO_BASE(__i915) ((__i915)->display.dsi.mmio_base)
+
+#define _MIPI_PORT(port, a, c)	(((port) == PORT_A) ? a : c)	/* ports A and C only */
+#define _MMIO_MIPI(port, a, c)	_MMIO(_MIPI_PORT(port, a, c))
+
+/* BXT MIPI mode configure */
+#define  _BXT_MIPIA_TRANS_HACTIVE			0x6B0F8
+#define  _BXT_MIPIC_TRANS_HACTIVE			0x6B8F8
+#define  BXT_MIPI_TRANS_HACTIVE(tc)	_MMIO_MIPI(tc, \
+		_BXT_MIPIA_TRANS_HACTIVE, _BXT_MIPIC_TRANS_HACTIVE)
+
+#define  _BXT_MIPIA_TRANS_VACTIVE			0x6B0FC
+#define  _BXT_MIPIC_TRANS_VACTIVE			0x6B8FC
+#define  BXT_MIPI_TRANS_VACTIVE(tc)	_MMIO_MIPI(tc, \
+		_BXT_MIPIA_TRANS_VACTIVE, _BXT_MIPIC_TRANS_VACTIVE)
+
+#define  _BXT_MIPIA_TRANS_VTOTAL			0x6B100
+#define  _BXT_MIPIC_TRANS_VTOTAL			0x6B900
+#define  BXT_MIPI_TRANS_VTOTAL(tc)	_MMIO_MIPI(tc, \
+		_BXT_MIPIA_TRANS_VTOTAL, _BXT_MIPIC_TRANS_VTOTAL)
+
+#define BXT_P_DSI_REGULATOR_CFG			_MMIO(0x160020)
+#define  STAP_SELECT					(1 << 0)
+
+#define BXT_P_DSI_REGULATOR_TX_CTRL		_MMIO(0x160054)
+#define  HS_IO_CTRL_SELECT				(1 << 0)
+
+#define _MIPIA_PORT_CTRL			(VLV_DISPLAY_BASE + 0x61190)
+#define _MIPIC_PORT_CTRL			(VLV_DISPLAY_BASE + 0x61700)
+#define MIPI_PORT_CTRL(port)	_MMIO_MIPI(port, _MIPIA_PORT_CTRL, _MIPIC_PORT_CTRL)
+
+ /* BXT port control */
+#define _BXT_MIPIA_PORT_CTRL				0x6B0C0
+#define _BXT_MIPIC_PORT_CTRL				0x6B8C0
+#define BXT_MIPI_PORT_CTRL(tc)	_MMIO_MIPI(tc, _BXT_MIPIA_PORT_CTRL, _BXT_MIPIC_PORT_CTRL)
+
+#define  DPI_ENABLE					(1 << 31) /* A + C */
+#define  MIPIA_MIPI4DPHY_DELAY_COUNT_SHIFT		27
+#define  MIPIA_MIPI4DPHY_DELAY_COUNT_MASK		(0xf << 27)
+#define  DUAL_LINK_MODE_SHIFT				26
+#define  DUAL_LINK_MODE_MASK				(1 << 26)
+#define  DUAL_LINK_MODE_FRONT_BACK			(0 << 26)
+#define  DUAL_LINK_MODE_PIXEL_ALTERNATIVE		(1 << 26)
+#define  DITHERING_ENABLE				(1 << 25) /* A + C */
+#define  FLOPPED_HSTX					(1 << 23)
+#define  DE_INVERT					(1 << 19) /* XXX */
+#define  MIPIA_FLISDSI_DELAY_COUNT_SHIFT		18
+#define  MIPIA_FLISDSI_DELAY_COUNT_MASK			(0xf << 18)
+#define  AFE_LATCHOUT					(1 << 17)
+#define  LP_OUTPUT_HOLD					(1 << 16)
+#define  MIPIC_FLISDSI_DELAY_COUNT_HIGH_SHIFT		15
+#define  MIPIC_FLISDSI_DELAY_COUNT_HIGH_MASK		(1 << 15)
+#define  MIPIC_MIPI4DPHY_DELAY_COUNT_SHIFT		11
+#define  MIPIC_MIPI4DPHY_DELAY_COUNT_MASK		(0xf << 11)
+#define  CSB_SHIFT					9
+#define  CSB_MASK					(3 << 9)
+#define  CSB_20MHZ					(0 << 9)
+#define  CSB_10MHZ					(1 << 9)
+#define  CSB_40MHZ					(2 << 9)
+#define  BANDGAP_MASK					(1 << 8)
+#define  BANDGAP_PNW_CIRCUIT				(0 << 8)
+#define  BANDGAP_LNC_CIRCUIT				(1 << 8)
+#define  MIPIC_FLISDSI_DELAY_COUNT_LOW_SHIFT		5
+#define  MIPIC_FLISDSI_DELAY_COUNT_LOW_MASK		(7 << 5)
+#define  TEARING_EFFECT_DELAY				(1 << 4) /* A + C */
+#define  TEARING_EFFECT_SHIFT				2 /* A + C */
+#define  TEARING_EFFECT_MASK				(3 << 2)
+#define  TEARING_EFFECT_OFF				(0 << 2)
+#define  TEARING_EFFECT_DSI				(1 << 2)
+#define  TEARING_EFFECT_GPIO				(2 << 2)
+#define  LANE_CONFIGURATION_SHIFT			0
+#define  LANE_CONFIGURATION_MASK			(3 << 0)
+#define  LANE_CONFIGURATION_4LANE			(0 << 0)
+#define  LANE_CONFIGURATION_DUAL_LINK_A			(1 << 0)
+#define  LANE_CONFIGURATION_DUAL_LINK_B			(2 << 0)
+
+#define _MIPIA_TEARING_CTRL			(VLV_DISPLAY_BASE + 0x61194)
+#define _MIPIC_TEARING_CTRL			(VLV_DISPLAY_BASE + 0x61704)
+#define MIPI_TEARING_CTRL(port)			_MMIO_MIPI(port, _MIPIA_TEARING_CTRL, _MIPIC_TEARING_CTRL)
+#define  TEARING_EFFECT_DELAY_SHIFT			0
+#define  TEARING_EFFECT_DELAY_MASK			(0xffff << 0)
+
+/* XXX: all bits reserved */
+#define _MIPIA_AUTOPWG			(VLV_DISPLAY_BASE + 0x611a0)
+
+/* MIPI DSI Controller and D-PHY registers */
+
+#define _MIPIA_DEVICE_READY		(_MIPI_MMIO_BASE(dev_priv) + 0xb000)
+#define _MIPIC_DEVICE_READY		(_MIPI_MMIO_BASE(dev_priv) + 0xb800)
+#define MIPI_DEVICE_READY(port)		_MMIO_MIPI(port, _MIPIA_DEVICE_READY, _MIPIC_DEVICE_READY)
+#define  BUS_POSSESSION					(1 << 3) /* set to give bus to receiver */
+#define  ULPS_STATE_MASK				(3 << 1)
+#define  ULPS_STATE_ENTER				(2 << 1)
+#define  ULPS_STATE_EXIT				(1 << 1)
+#define  ULPS_STATE_NORMAL_OPERATION			(0 << 1)
+#define  DEVICE_READY					(1 << 0)
+
+#define _MIPIA_INTR_STAT		(_MIPI_MMIO_BASE(dev_priv) + 0xb004)
+#define _MIPIC_INTR_STAT		(_MIPI_MMIO_BASE(dev_priv) + 0xb804)
+#define MIPI_INTR_STAT(port)		_MMIO_MIPI(port, _MIPIA_INTR_STAT, _MIPIC_INTR_STAT)
+#define _MIPIA_INTR_EN			(_MIPI_MMIO_BASE(dev_priv) + 0xb008)
+#define _MIPIC_INTR_EN			(_MIPI_MMIO_BASE(dev_priv) + 0xb808)
+#define MIPI_INTR_EN(port)		_MMIO_MIPI(port, _MIPIA_INTR_EN, _MIPIC_INTR_EN)
+#define  TEARING_EFFECT					(1 << 31)
+#define  SPL_PKT_SENT_INTERRUPT				(1 << 30)
+#define  GEN_READ_DATA_AVAIL				(1 << 29)
+#define  LP_GENERIC_WR_FIFO_FULL			(1 << 28)
+#define  HS_GENERIC_WR_FIFO_FULL			(1 << 27)
+#define  RX_PROT_VIOLATION				(1 << 26)
+#define  RX_INVALID_TX_LENGTH				(1 << 25)
+#define  ACK_WITH_NO_ERROR				(1 << 24)
+#define  TURN_AROUND_ACK_TIMEOUT			(1 << 23)
+#define  LP_RX_TIMEOUT					(1 << 22)
+#define  HS_TX_TIMEOUT					(1 << 21)
+#define  DPI_FIFO_UNDERRUN				(1 << 20)
+#define  LOW_CONTENTION					(1 << 19)
+#define  HIGH_CONTENTION				(1 << 18)
+#define  TXDSI_VC_ID_INVALID				(1 << 17)
+#define  TXDSI_DATA_TYPE_NOT_RECOGNISED			(1 << 16)
+#define  TXCHECKSUM_ERROR				(1 << 15)
+#define  TXECC_MULTIBIT_ERROR				(1 << 14)
+#define  TXECC_SINGLE_BIT_ERROR				(1 << 13)
+#define  TXFALSE_CONTROL_ERROR				(1 << 12)
+#define  RXDSI_VC_ID_INVALID				(1 << 11)
+#define  RXDSI_DATA_TYPE_NOT_REGOGNISED			(1 << 10)
+#define  RXCHECKSUM_ERROR				(1 << 9)
+#define  RXECC_MULTIBIT_ERROR				(1 << 8)
+#define  RXECC_SINGLE_BIT_ERROR				(1 << 7)
+#define  RXFALSE_CONTROL_ERROR				(1 << 6)
+#define  RXHS_RECEIVE_TIMEOUT_ERROR			(1 << 5)
+#define  RX_LP_TX_SYNC_ERROR				(1 << 4)
+#define  RXEXCAPE_MODE_ENTRY_ERROR			(1 << 3)
+#define  RXEOT_SYNC_ERROR				(1 << 2)
+#define  RXSOT_SYNC_ERROR				(1 << 1)
+#define  RXSOT_ERROR					(1 << 0)
+
+#define _MIPIA_DSI_FUNC_PRG		(_MIPI_MMIO_BASE(dev_priv) + 0xb00c)
+#define _MIPIC_DSI_FUNC_PRG		(_MIPI_MMIO_BASE(dev_priv) + 0xb80c)
+#define MIPI_DSI_FUNC_PRG(port)		_MMIO_MIPI(port, _MIPIA_DSI_FUNC_PRG, _MIPIC_DSI_FUNC_PRG)
+#define  CMD_MODE_DATA_WIDTH_MASK			(7 << 13)
+#define  CMD_MODE_NOT_SUPPORTED				(0 << 13)
+#define  CMD_MODE_DATA_WIDTH_16_BIT			(1 << 13)
+#define  CMD_MODE_DATA_WIDTH_9_BIT			(2 << 13)
+#define  CMD_MODE_DATA_WIDTH_8_BIT			(3 << 13)
+#define  CMD_MODE_DATA_WIDTH_OPTION1			(4 << 13)
+#define  CMD_MODE_DATA_WIDTH_OPTION2			(5 << 13)
+#define  VID_MODE_FORMAT_MASK				(0xf << 7)
+#define  VID_MODE_NOT_SUPPORTED				(0 << 7)
+#define  VID_MODE_FORMAT_RGB565				(1 << 7)
+#define  VID_MODE_FORMAT_RGB666_PACKED			(2 << 7)
+#define  VID_MODE_FORMAT_RGB666				(3 << 7)
+#define  VID_MODE_FORMAT_RGB888				(4 << 7)
+#define  CMD_MODE_CHANNEL_NUMBER_SHIFT			5
+#define  CMD_MODE_CHANNEL_NUMBER_MASK			(3 << 5)
+#define  VID_MODE_CHANNEL_NUMBER_SHIFT			3
+#define  VID_MODE_CHANNEL_NUMBER_MASK			(3 << 3)
+#define  DATA_LANES_PRG_REG_SHIFT			0
+#define  DATA_LANES_PRG_REG_MASK			(7 << 0)
+
+#define _MIPIA_HS_TX_TIMEOUT		(_MIPI_MMIO_BASE(dev_priv) + 0xb010)
+#define _MIPIC_HS_TX_TIMEOUT		(_MIPI_MMIO_BASE(dev_priv) + 0xb810)
+#define MIPI_HS_TX_TIMEOUT(port)	_MMIO_MIPI(port, _MIPIA_HS_TX_TIMEOUT, _MIPIC_HS_TX_TIMEOUT)
+#define  HIGH_SPEED_TX_TIMEOUT_COUNTER_MASK		0xffffff
+
+#define _MIPIA_LP_RX_TIMEOUT		(_MIPI_MMIO_BASE(dev_priv) + 0xb014)
+#define _MIPIC_LP_RX_TIMEOUT		(_MIPI_MMIO_BASE(dev_priv) + 0xb814)
+#define MIPI_LP_RX_TIMEOUT(port)	_MMIO_MIPI(port, _MIPIA_LP_RX_TIMEOUT, _MIPIC_LP_RX_TIMEOUT)
+#define  LOW_POWER_RX_TIMEOUT_COUNTER_MASK		0xffffff
+
+#define _MIPIA_TURN_AROUND_TIMEOUT	(_MIPI_MMIO_BASE(dev_priv) + 0xb018)
+#define _MIPIC_TURN_AROUND_TIMEOUT	(_MIPI_MMIO_BASE(dev_priv) + 0xb818)
+#define MIPI_TURN_AROUND_TIMEOUT(port)	_MMIO_MIPI(port, _MIPIA_TURN_AROUND_TIMEOUT, _MIPIC_TURN_AROUND_TIMEOUT)
+#define  TURN_AROUND_TIMEOUT_MASK			0x3f
+
+#define _MIPIA_DEVICE_RESET_TIMER	(_MIPI_MMIO_BASE(dev_priv) + 0xb01c)
+#define _MIPIC_DEVICE_RESET_TIMER	(_MIPI_MMIO_BASE(dev_priv) + 0xb81c)
+#define MIPI_DEVICE_RESET_TIMER(port)	_MMIO_MIPI(port, _MIPIA_DEVICE_RESET_TIMER, _MIPIC_DEVICE_RESET_TIMER)
+#define  DEVICE_RESET_TIMER_MASK			0xffff
+
+#define _MIPIA_DPI_RESOLUTION		(_MIPI_MMIO_BASE(dev_priv) + 0xb020)
+#define _MIPIC_DPI_RESOLUTION		(_MIPI_MMIO_BASE(dev_priv) + 0xb820)
+#define MIPI_DPI_RESOLUTION(port)	_MMIO_MIPI(port, _MIPIA_DPI_RESOLUTION, _MIPIC_DPI_RESOLUTION)
+#define  VERTICAL_ADDRESS_SHIFT				16
+#define  VERTICAL_ADDRESS_MASK				(0xffff << 16)
+#define  HORIZONTAL_ADDRESS_SHIFT			0
+#define  HORIZONTAL_ADDRESS_MASK			0xffff
+
+#define _MIPIA_DBI_FIFO_THROTTLE	(_MIPI_MMIO_BASE(dev_priv) + 0xb024)
+#define _MIPIC_DBI_FIFO_THROTTLE	(_MIPI_MMIO_BASE(dev_priv) + 0xb824)
+#define MIPI_DBI_FIFO_THROTTLE(port)	_MMIO_MIPI(port, _MIPIA_DBI_FIFO_THROTTLE, _MIPIC_DBI_FIFO_THROTTLE)
+#define  DBI_FIFO_EMPTY_HALF				(0 << 0)
+#define  DBI_FIFO_EMPTY_QUARTER				(1 << 0)
+#define  DBI_FIFO_EMPTY_7_LOCATIONS			(2 << 0)
+
+/* regs below are bits 15:0 */
+#define _MIPIA_HSYNC_PADDING_COUNT	(_MIPI_MMIO_BASE(dev_priv) + 0xb028)
+#define _MIPIC_HSYNC_PADDING_COUNT	(_MIPI_MMIO_BASE(dev_priv) + 0xb828)
+#define MIPI_HSYNC_PADDING_COUNT(port)	_MMIO_MIPI(port, _MIPIA_HSYNC_PADDING_COUNT, _MIPIC_HSYNC_PADDING_COUNT)
+
+#define _MIPIA_HBP_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb02c)
+#define _MIPIC_HBP_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb82c)
+#define MIPI_HBP_COUNT(port)		_MMIO_MIPI(port, _MIPIA_HBP_COUNT, _MIPIC_HBP_COUNT)
+
+#define _MIPIA_HFP_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb030)
+#define _MIPIC_HFP_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb830)
+#define MIPI_HFP_COUNT(port)		_MMIO_MIPI(port, _MIPIA_HFP_COUNT, _MIPIC_HFP_COUNT)
+
+#define _MIPIA_HACTIVE_AREA_COUNT	(_MIPI_MMIO_BASE(dev_priv) + 0xb034)
+#define _MIPIC_HACTIVE_AREA_COUNT	(_MIPI_MMIO_BASE(dev_priv) + 0xb834)
+#define MIPI_HACTIVE_AREA_COUNT(port)	_MMIO_MIPI(port, _MIPIA_HACTIVE_AREA_COUNT, _MIPIC_HACTIVE_AREA_COUNT)
+
+#define _MIPIA_VSYNC_PADDING_COUNT	(_MIPI_MMIO_BASE(dev_priv) + 0xb038)
+#define _MIPIC_VSYNC_PADDING_COUNT	(_MIPI_MMIO_BASE(dev_priv) + 0xb838)
+#define MIPI_VSYNC_PADDING_COUNT(port)	_MMIO_MIPI(port, _MIPIA_VSYNC_PADDING_COUNT, _MIPIC_VSYNC_PADDING_COUNT)
+
+#define _MIPIA_VBP_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb03c)
+#define _MIPIC_VBP_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb83c)
+#define MIPI_VBP_COUNT(port)		_MMIO_MIPI(port, _MIPIA_VBP_COUNT, _MIPIC_VBP_COUNT)
+
+#define _MIPIA_VFP_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb040)
+#define _MIPIC_VFP_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb840)
+#define MIPI_VFP_COUNT(port)		_MMIO_MIPI(port, _MIPIA_VFP_COUNT, _MIPIC_VFP_COUNT)
+
+#define _MIPIA_HIGH_LOW_SWITCH_COUNT	(_MIPI_MMIO_BASE(dev_priv) + 0xb044)
+#define _MIPIC_HIGH_LOW_SWITCH_COUNT	(_MIPI_MMIO_BASE(dev_priv) + 0xb844)
+#define MIPI_HIGH_LOW_SWITCH_COUNT(port)	_MMIO_MIPI(port,	_MIPIA_HIGH_LOW_SWITCH_COUNT, _MIPIC_HIGH_LOW_SWITCH_COUNT)
+
+#define _MIPIA_DPI_CONTROL		(_MIPI_MMIO_BASE(dev_priv) + 0xb048)
+#define _MIPIC_DPI_CONTROL		(_MIPI_MMIO_BASE(dev_priv) + 0xb848)
+#define MIPI_DPI_CONTROL(port)		_MMIO_MIPI(port, _MIPIA_DPI_CONTROL, _MIPIC_DPI_CONTROL)
+#define  DPI_LP_MODE					(1 << 6)
+#define  BACKLIGHT_OFF					(1 << 5)
+#define  BACKLIGHT_ON					(1 << 4)
+#define  COLOR_MODE_OFF					(1 << 3)
+#define  COLOR_MODE_ON					(1 << 2)
+#define  TURN_ON					(1 << 1)
+#define  SHUTDOWN					(1 << 0)
+
+#define _MIPIA_DPI_DATA			(_MIPI_MMIO_BASE(dev_priv) + 0xb04c)
+#define _MIPIC_DPI_DATA			(_MIPI_MMIO_BASE(dev_priv) + 0xb84c)
+#define MIPI_DPI_DATA(port)		_MMIO_MIPI(port, _MIPIA_DPI_DATA, _MIPIC_DPI_DATA)
+#define  COMMAND_BYTE_SHIFT				0
+#define  COMMAND_BYTE_MASK				(0x3f << 0)
+
+#define _MIPIA_INIT_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb050)
+#define _MIPIC_INIT_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb850)
+#define MIPI_INIT_COUNT(port)		_MMIO_MIPI(port, _MIPIA_INIT_COUNT, _MIPIC_INIT_COUNT)
+#define  MASTER_INIT_TIMER_SHIFT			0
+#define  MASTER_INIT_TIMER_MASK				(0xffff << 0)
+
+#define _MIPIA_MAX_RETURN_PKT_SIZE	(_MIPI_MMIO_BASE(dev_priv) + 0xb054)
+#define _MIPIC_MAX_RETURN_PKT_SIZE	(_MIPI_MMIO_BASE(dev_priv) + 0xb854)
+#define MIPI_MAX_RETURN_PKT_SIZE(port)	_MMIO_MIPI(port, \
+			_MIPIA_MAX_RETURN_PKT_SIZE, _MIPIC_MAX_RETURN_PKT_SIZE)
+#define  MAX_RETURN_PKT_SIZE_SHIFT			0
+#define  MAX_RETURN_PKT_SIZE_MASK			(0x3ff << 0)
+
+#define _MIPIA_VIDEO_MODE_FORMAT	(_MIPI_MMIO_BASE(dev_priv) + 0xb058)
+#define _MIPIC_VIDEO_MODE_FORMAT	(_MIPI_MMIO_BASE(dev_priv) + 0xb858)
+#define MIPI_VIDEO_MODE_FORMAT(port)	_MMIO_MIPI(port, _MIPIA_VIDEO_MODE_FORMAT, _MIPIC_VIDEO_MODE_FORMAT)
+#define  RANDOM_DPI_DISPLAY_RESOLUTION			(1 << 4)
+#define  DISABLE_VIDEO_BTA				(1 << 3)
+#define  IP_TG_CONFIG					(1 << 2)
+#define  VIDEO_MODE_NON_BURST_WITH_SYNC_PULSE		(1 << 0)
+#define  VIDEO_MODE_NON_BURST_WITH_SYNC_EVENTS		(2 << 0)
+#define  VIDEO_MODE_BURST				(3 << 0)
+
+#define _MIPIA_EOT_DISABLE		(_MIPI_MMIO_BASE(dev_priv) + 0xb05c)
+#define _MIPIC_EOT_DISABLE		(_MIPI_MMIO_BASE(dev_priv) + 0xb85c)
+#define MIPI_EOT_DISABLE(port)		_MMIO_MIPI(port, _MIPIA_EOT_DISABLE, _MIPIC_EOT_DISABLE)
+#define  BXT_DEFEATURE_DPI_FIFO_CTR			(1 << 9)
+#define  BXT_DPHY_DEFEATURE_EN				(1 << 8)
+#define  LP_RX_TIMEOUT_ERROR_RECOVERY_DISABLE		(1 << 7)
+#define  HS_RX_TIMEOUT_ERROR_RECOVERY_DISABLE		(1 << 6)
+#define  LOW_CONTENTION_RECOVERY_DISABLE		(1 << 5)
+#define  HIGH_CONTENTION_RECOVERY_DISABLE		(1 << 4)
+#define  TXDSI_TYPE_NOT_RECOGNISED_ERROR_RECOVERY_DISABLE (1 << 3)
+#define  TXECC_MULTIBIT_ERROR_RECOVERY_DISABLE		(1 << 2)
+#define  CLOCKSTOP					(1 << 1)
+#define  EOT_DISABLE					(1 << 0)
+
+#define _MIPIA_LP_BYTECLK		(_MIPI_MMIO_BASE(dev_priv) + 0xb060)
+#define _MIPIC_LP_BYTECLK		(_MIPI_MMIO_BASE(dev_priv) + 0xb860)
+#define MIPI_LP_BYTECLK(port)		_MMIO_MIPI(port, _MIPIA_LP_BYTECLK, _MIPIC_LP_BYTECLK)
+#define  LP_BYTECLK_SHIFT				0
+#define  LP_BYTECLK_MASK				(0xffff << 0)
+
+#define _MIPIA_TLPX_TIME_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb0a4)
+#define _MIPIC_TLPX_TIME_COUNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb8a4)
+#define MIPI_TLPX_TIME_COUNT(port)	 _MMIO_MIPI(port, _MIPIA_TLPX_TIME_COUNT, _MIPIC_TLPX_TIME_COUNT)
+
+#define _MIPIA_CLK_LANE_TIMING		(_MIPI_MMIO_BASE(dev_priv) + 0xb098)
+#define _MIPIC_CLK_LANE_TIMING		(_MIPI_MMIO_BASE(dev_priv) + 0xb898)
+#define MIPI_CLK_LANE_TIMING(port)	 _MMIO_MIPI(port, _MIPIA_CLK_LANE_TIMING, _MIPIC_CLK_LANE_TIMING)
+
+/* bits 31:0 */
+#define _MIPIA_LP_GEN_DATA		(_MIPI_MMIO_BASE(dev_priv) + 0xb064)
+#define _MIPIC_LP_GEN_DATA		(_MIPI_MMIO_BASE(dev_priv) + 0xb864)
+#define MIPI_LP_GEN_DATA(port)		_MMIO_MIPI(port, _MIPIA_LP_GEN_DATA, _MIPIC_LP_GEN_DATA)
+
+/* bits 31:0 */
+#define _MIPIA_HS_GEN_DATA		(_MIPI_MMIO_BASE(dev_priv) + 0xb068)
+#define _MIPIC_HS_GEN_DATA		(_MIPI_MMIO_BASE(dev_priv) + 0xb868)
+#define MIPI_HS_GEN_DATA(port)		_MMIO_MIPI(port, _MIPIA_HS_GEN_DATA, _MIPIC_HS_GEN_DATA)
+
+#define _MIPIA_LP_GEN_CTRL		(_MIPI_MMIO_BASE(dev_priv) + 0xb06c)
+#define _MIPIC_LP_GEN_CTRL		(_MIPI_MMIO_BASE(dev_priv) + 0xb86c)
+#define MIPI_LP_GEN_CTRL(port)		_MMIO_MIPI(port, _MIPIA_LP_GEN_CTRL, _MIPIC_LP_GEN_CTRL)
+#define _MIPIA_HS_GEN_CTRL		(_MIPI_MMIO_BASE(dev_priv) + 0xb070)
+#define _MIPIC_HS_GEN_CTRL		(_MIPI_MMIO_BASE(dev_priv) + 0xb870)
+#define MIPI_HS_GEN_CTRL(port)		_MMIO_MIPI(port, _MIPIA_HS_GEN_CTRL, _MIPIC_HS_GEN_CTRL)
+#define  LONG_PACKET_WORD_COUNT_SHIFT			8
+#define  LONG_PACKET_WORD_COUNT_MASK			(0xffff << 8)
+#define  SHORT_PACKET_PARAM_SHIFT			8
+#define  SHORT_PACKET_PARAM_MASK			(0xffff << 8)
+#define  VIRTUAL_CHANNEL_SHIFT				6
+#define  VIRTUAL_CHANNEL_MASK				(3 << 6)
+#define  DATA_TYPE_SHIFT				0
+#define  DATA_TYPE_MASK					(0x3f << 0)
+/* data type values, see include/video/mipi_display.h */
+
+#define _MIPIA_GEN_FIFO_STAT		(_MIPI_MMIO_BASE(dev_priv) + 0xb074)
+#define _MIPIC_GEN_FIFO_STAT		(_MIPI_MMIO_BASE(dev_priv) + 0xb874)
+#define MIPI_GEN_FIFO_STAT(port)	_MMIO_MIPI(port, _MIPIA_GEN_FIFO_STAT, _MIPIC_GEN_FIFO_STAT)
+#define  DPI_FIFO_EMPTY					(1 << 28)
+#define  DBI_FIFO_EMPTY					(1 << 27)
+#define  LP_CTRL_FIFO_EMPTY				(1 << 26)
+#define  LP_CTRL_FIFO_HALF_EMPTY			(1 << 25)
+#define  LP_CTRL_FIFO_FULL				(1 << 24)
+#define  HS_CTRL_FIFO_EMPTY				(1 << 18)
+#define  HS_CTRL_FIFO_HALF_EMPTY			(1 << 17)
+#define  HS_CTRL_FIFO_FULL				(1 << 16)
+#define  LP_DATA_FIFO_EMPTY				(1 << 10)
+#define  LP_DATA_FIFO_HALF_EMPTY			(1 << 9)
+#define  LP_DATA_FIFO_FULL				(1 << 8)
+#define  HS_DATA_FIFO_EMPTY				(1 << 2)
+#define  HS_DATA_FIFO_HALF_EMPTY			(1 << 1)
+#define  HS_DATA_FIFO_FULL				(1 << 0)
+
+#define _MIPIA_HS_LS_DBI_ENABLE		(_MIPI_MMIO_BASE(dev_priv) + 0xb078)
+#define _MIPIC_HS_LS_DBI_ENABLE		(_MIPI_MMIO_BASE(dev_priv) + 0xb878)
+#define MIPI_HS_LP_DBI_ENABLE(port)	_MMIO_MIPI(port, _MIPIA_HS_LS_DBI_ENABLE, _MIPIC_HS_LS_DBI_ENABLE)
+#define  DBI_HS_LP_MODE_MASK				(1 << 0)
+#define  DBI_LP_MODE					(1 << 0)
+#define  DBI_HS_MODE					(0 << 0)
+
+#define _MIPIA_DPHY_PARAM		(_MIPI_MMIO_BASE(dev_priv) + 0xb080)
+#define _MIPIC_DPHY_PARAM		(_MIPI_MMIO_BASE(dev_priv) + 0xb880)
+#define MIPI_DPHY_PARAM(port)		_MMIO_MIPI(port, _MIPIA_DPHY_PARAM, _MIPIC_DPHY_PARAM)
+#define  EXIT_ZERO_COUNT_SHIFT				24
+#define  EXIT_ZERO_COUNT_MASK				(0x3f << 24)
+#define  TRAIL_COUNT_SHIFT				16
+#define  TRAIL_COUNT_MASK				(0x1f << 16)
+#define  CLK_ZERO_COUNT_SHIFT				8
+#define  CLK_ZERO_COUNT_MASK				(0xff << 8)
+#define  PREPARE_COUNT_SHIFT				0
+#define  PREPARE_COUNT_MASK				(0x3f << 0)
+
+#define _MIPIA_DBI_BW_CTRL		(_MIPI_MMIO_BASE(dev_priv) + 0xb084)
+#define _MIPIC_DBI_BW_CTRL		(_MIPI_MMIO_BASE(dev_priv) + 0xb884)
+#define MIPI_DBI_BW_CTRL(port)		_MMIO_MIPI(port, _MIPIA_DBI_BW_CTRL, _MIPIC_DBI_BW_CTRL)
+
+#define _MIPIA_CLK_LANE_SWITCH_TIME_CNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb088)
+#define _MIPIC_CLK_LANE_SWITCH_TIME_CNT		(_MIPI_MMIO_BASE(dev_priv) + 0xb888)
+#define MIPI_CLK_LANE_SWITCH_TIME_CNT(port)	_MMIO_MIPI(port, _MIPIA_CLK_LANE_SWITCH_TIME_CNT, _MIPIC_CLK_LANE_SWITCH_TIME_CNT)
+#define  LP_HS_SSW_CNT_SHIFT				16
+#define  LP_HS_SSW_CNT_MASK				(0xffff << 16)
+#define  HS_LP_PWR_SW_CNT_SHIFT				0
+#define  HS_LP_PWR_SW_CNT_MASK				(0xffff << 0)
+
+#define _MIPIA_STOP_STATE_STALL		(_MIPI_MMIO_BASE(dev_priv) + 0xb08c)
+#define _MIPIC_STOP_STATE_STALL		(_MIPI_MMIO_BASE(dev_priv) + 0xb88c)
+#define MIPI_STOP_STATE_STALL(port)	_MMIO_MIPI(port, _MIPIA_STOP_STATE_STALL, _MIPIC_STOP_STATE_STALL)
+#define  STOP_STATE_STALL_COUNTER_SHIFT			0
+#define  STOP_STATE_STALL_COUNTER_MASK			(0xff << 0)
+
+#define _MIPIA_INTR_STAT_REG_1		(_MIPI_MMIO_BASE(dev_priv) + 0xb090)
+#define _MIPIC_INTR_STAT_REG_1		(_MIPI_MMIO_BASE(dev_priv) + 0xb890)
+#define MIPI_INTR_STAT_REG_1(port)	_MMIO_MIPI(port, _MIPIA_INTR_STAT_REG_1, _MIPIC_INTR_STAT_REG_1)
+#define _MIPIA_INTR_EN_REG_1		(_MIPI_MMIO_BASE(dev_priv) + 0xb094)
+#define _MIPIC_INTR_EN_REG_1		(_MIPI_MMIO_BASE(dev_priv) + 0xb894)
+#define MIPI_INTR_EN_REG_1(port)	_MMIO_MIPI(port, _MIPIA_INTR_EN_REG_1, _MIPIC_INTR_EN_REG_1)
+#define  RX_CONTENTION_DETECTED				(1 << 0)
+
+/* XXX: only pipe A ?!? */
+#define MIPIA_DBI_TYPEC_CTRL		(_MIPI_MMIO_BASE(dev_priv) + 0xb100)
+#define  DBI_TYPEC_ENABLE				(1 << 31)
+#define  DBI_TYPEC_WIP					(1 << 30)
+#define  DBI_TYPEC_OPTION_SHIFT				28
+#define  DBI_TYPEC_OPTION_MASK				(3 << 28)
+#define  DBI_TYPEC_FREQ_SHIFT				24
+#define  DBI_TYPEC_FREQ_MASK				(0xf << 24)
+#define  DBI_TYPEC_OVERRIDE				(1 << 8)
+#define  DBI_TYPEC_OVERRIDE_COUNTER_SHIFT		0
+#define  DBI_TYPEC_OVERRIDE_COUNTER_MASK		(0xff << 0)
+
+/* MIPI adapter registers */
+
+#define _MIPIA_CTRL			(_MIPI_MMIO_BASE(dev_priv) + 0xb104)
+#define _MIPIC_CTRL			(_MIPI_MMIO_BASE(dev_priv) + 0xb904)
+#define MIPI_CTRL(port)			_MMIO_MIPI(port, _MIPIA_CTRL, _MIPIC_CTRL)
+#define  ESCAPE_CLOCK_DIVIDER_SHIFT			5 /* A only */
+#define  ESCAPE_CLOCK_DIVIDER_MASK			(3 << 5)
+#define  ESCAPE_CLOCK_DIVIDER_1				(0 << 5)
+#define  ESCAPE_CLOCK_DIVIDER_2				(1 << 5)
+#define  ESCAPE_CLOCK_DIVIDER_4				(2 << 5)
+#define  READ_REQUEST_PRIORITY_SHIFT			3
+#define  READ_REQUEST_PRIORITY_MASK			(3 << 3)
+#define  READ_REQUEST_PRIORITY_LOW			(0 << 3)
+#define  READ_REQUEST_PRIORITY_HIGH			(3 << 3)
+#define  RGB_FLIP_TO_BGR				(1 << 2)
+
+#define  BXT_PIPE_SELECT_SHIFT				7
+#define  BXT_PIPE_SELECT_MASK				(7 << 7)
+#define  BXT_PIPE_SELECT(pipe)				((pipe) << 7)
+#define  GLK_PHY_STATUS_PORT_READY			(1 << 31) /* RO */
+#define  GLK_ULPS_NOT_ACTIVE				(1 << 30) /* RO */
+#define  GLK_MIPIIO_RESET_RELEASED			(1 << 28)
+#define  GLK_CLOCK_LANE_STOP_STATE			(1 << 27) /* RO */
+#define  GLK_DATA_LANE_STOP_STATE			(1 << 26) /* RO */
+#define  GLK_LP_WAKE					(1 << 22)
+#define  GLK_LP11_LOW_PWR_MODE				(1 << 21)
+#define  GLK_LP00_LOW_PWR_MODE				(1 << 20)
+#define  GLK_FIREWALL_ENABLE				(1 << 16)
+#define  BXT_PIXEL_OVERLAP_CNT_MASK			(0xf << 10)
+#define  BXT_PIXEL_OVERLAP_CNT_SHIFT			10
+#define  BXT_DSC_ENABLE					(1 << 3)
+#define  BXT_RGB_FLIP					(1 << 2)
+#define  GLK_MIPIIO_PORT_POWERED			(1 << 1) /* RO */
+#define  GLK_MIPIIO_ENABLE				(1 << 0)
+
+#define _MIPIA_DATA_ADDRESS		(_MIPI_MMIO_BASE(dev_priv) + 0xb108)
+#define _MIPIC_DATA_ADDRESS		(_MIPI_MMIO_BASE(dev_priv) + 0xb908)
+#define MIPI_DATA_ADDRESS(port)		_MMIO_MIPI(port, _MIPIA_DATA_ADDRESS, _MIPIC_DATA_ADDRESS)
+#define  DATA_MEM_ADDRESS_SHIFT				5
+#define  DATA_MEM_ADDRESS_MASK				(0x7ffffff << 5)
+#define  DATA_VALID					(1 << 0)
+
+#define _MIPIA_DATA_LENGTH		(_MIPI_MMIO_BASE(dev_priv) + 0xb10c)
+#define _MIPIC_DATA_LENGTH		(_MIPI_MMIO_BASE(dev_priv) + 0xb90c)
+#define MIPI_DATA_LENGTH(port)		_MMIO_MIPI(port, _MIPIA_DATA_LENGTH, _MIPIC_DATA_LENGTH)
+#define  DATA_LENGTH_SHIFT				0
+#define  DATA_LENGTH_MASK				(0xfffff << 0)
+
+#define _MIPIA_COMMAND_ADDRESS		(_MIPI_MMIO_BASE(dev_priv) + 0xb110)
+#define _MIPIC_COMMAND_ADDRESS		(_MIPI_MMIO_BASE(dev_priv) + 0xb910)
+#define MIPI_COMMAND_ADDRESS(port)	_MMIO_MIPI(port, _MIPIA_COMMAND_ADDRESS, _MIPIC_COMMAND_ADDRESS)
+#define  COMMAND_MEM_ADDRESS_SHIFT			5
+#define  COMMAND_MEM_ADDRESS_MASK			(0x7ffffff << 5)
+#define  AUTO_PWG_ENABLE				(1 << 2)
+#define  MEMORY_WRITE_DATA_FROM_PIPE_RENDERING		(1 << 1)
+#define  COMMAND_VALID					(1 << 0)
+
+#define _MIPIA_COMMAND_LENGTH		(_MIPI_MMIO_BASE(dev_priv) + 0xb114)
+#define _MIPIC_COMMAND_LENGTH		(_MIPI_MMIO_BASE(dev_priv) + 0xb914)
+#define MIPI_COMMAND_LENGTH(port)	_MMIO_MIPI(port, _MIPIA_COMMAND_LENGTH, _MIPIC_COMMAND_LENGTH)
+#define  COMMAND_LENGTH_SHIFT(n)			(8 * (n)) /* n: 0...3 */
+#define  COMMAND_LENGTH_MASK(n)				(0xff << (8 * (n)))
+
+#define _MIPIA_READ_DATA_RETURN0	(_MIPI_MMIO_BASE(dev_priv) + 0xb118)
+#define _MIPIC_READ_DATA_RETURN0	(_MIPI_MMIO_BASE(dev_priv) + 0xb918)
+#define MIPI_READ_DATA_RETURN(port, n) _MMIO(_MIPI(port, _MIPIA_READ_DATA_RETURN0, _MIPIC_READ_DATA_RETURN0) + 4 * (n)) /* n: 0...7 */
+
+#define _MIPIA_READ_DATA_VALID		(_MIPI_MMIO_BASE(dev_priv) + 0xb138)
+#define _MIPIC_READ_DATA_VALID		(_MIPI_MMIO_BASE(dev_priv) + 0xb938)
+#define MIPI_READ_DATA_VALID(port)	_MMIO_MIPI(port, _MIPIA_READ_DATA_VALID, _MIPIC_READ_DATA_VALID)
+#define  READ_DATA_VALID(n)				(1 << (n))
+
+#endif /* __VLV_DSI_REGS_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_busy.c b/drivers/gpu/drm/i915/gem/i915_gem_busy.c
new file mode 100644
index 000000000..ddda46824
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_busy.c
@@ -0,0 +1,164 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2014-2016 Intel Corporation
+ */
+
+#include <linux/dma-fence-array.h>
+
+#include "gt/intel_engine.h"
+
+#include "i915_gem_ioctls.h"
+#include "i915_gem_object.h"
+
+static __always_inline u32 __busy_read_flag(u16 id)
+{
+	if (id == (u16)I915_ENGINE_CLASS_INVALID)
+		return 0xffff0000u;
+
+	GEM_BUG_ON(id >= 16);
+	return 0x10000u << id;
+}
+
+static __always_inline u32 __busy_write_id(u16 id)
+{
+	/*
+	 * The uABI guarantees an active writer is also amongst the read
+	 * engines. This would be true if we accessed the activity tracking
+	 * under the lock, but as we perform the lookup of the object and
+	 * its activity locklessly we can not guarantee that the last_write
+	 * being active implies that we have set the same engine flag from
+	 * last_read - hence we always set both read and write busy for
+	 * last_write.
+	 */
+	if (id == (u16)I915_ENGINE_CLASS_INVALID)
+		return 0xffffffffu;
+
+	return (id + 1) | __busy_read_flag(id);
+}
+
+static __always_inline unsigned int
+__busy_set_if_active(struct dma_fence *fence, u32 (*flag)(u16 id))
+{
+	const struct i915_request *rq;
+
+	/*
+	 * We have to check the current hw status of the fence as the uABI
+	 * guarantees forward progress. We could rely on the idle worker
+	 * to eventually flush us, but to minimise latency just ask the
+	 * hardware.
+	 *
+	 * Note we only report on the status of native fences and we currently
+	 * have two native fences:
+	 *
+	 * 1. A composite fence (dma_fence_array) constructed of i915 requests
+	 * created during a parallel submission. In this case we deconstruct the
+	 * composite fence into individual i915 requests and check the status of
+	 * each request.
+	 *
+	 * 2. A single i915 request.
+	 */
+	if (dma_fence_is_array(fence)) {
+		struct dma_fence_array *array = to_dma_fence_array(fence);
+		struct dma_fence **child = array->fences;
+		unsigned int nchild = array->num_fences;
+
+		do {
+			struct dma_fence *current_fence = *child++;
+
+			/* Not an i915 fence, can't be busy per above */
+			if (!dma_fence_is_i915(current_fence) ||
+			    !test_bit(I915_FENCE_FLAG_COMPOSITE,
+				      &current_fence->flags)) {
+				return 0;
+			}
+
+			rq = to_request(current_fence);
+			if (!i915_request_completed(rq))
+				return flag(rq->engine->uabi_class);
+		} while (--nchild);
+
+		/* All requests in array complete, not busy */
+		return 0;
+	} else {
+		if (!dma_fence_is_i915(fence))
+			return 0;
+
+		rq = to_request(fence);
+		if (i915_request_completed(rq))
+			return 0;
+
+		/* Beware type-expansion follies! */
+		BUILD_BUG_ON(!typecheck(u16, rq->engine->uabi_class));
+		return flag(rq->engine->uabi_class);
+	}
+}
+
+static __always_inline unsigned int
+busy_check_reader(struct dma_fence *fence)
+{
+	return __busy_set_if_active(fence, __busy_read_flag);
+}
+
+static __always_inline unsigned int
+busy_check_writer(struct dma_fence *fence)
+{
+	if (!fence)
+		return 0;
+
+	return __busy_set_if_active(fence, __busy_write_id);
+}
+
+int
+i915_gem_busy_ioctl(struct drm_device *dev, void *data,
+		    struct drm_file *file)
+{
+	struct drm_i915_gem_busy *args = data;
+	struct drm_i915_gem_object *obj;
+	struct dma_resv_iter cursor;
+	struct dma_fence *fence;
+	int err;
+
+	err = -ENOENT;
+	rcu_read_lock();
+	obj = i915_gem_object_lookup_rcu(file, args->handle);
+	if (!obj)
+		goto out;
+
+	/*
+	 * A discrepancy here is that we do not report the status of
+	 * non-i915 fences, i.e. even though we may report the object as idle,
+	 * a call to set-domain may still stall waiting for foreign rendering.
+	 * This also means that wait-ioctl may report an object as busy,
+	 * where busy-ioctl considers it idle.
+	 *
+	 * We trade the ability to warn of foreign fences to report on which
+	 * i915 engines are active for the object.
+	 *
+	 * Alternatively, we can trade that extra information on read/write
+	 * activity with
+	 *	args->busy =
+	 *		!dma_resv_test_signaled(obj->resv, DMA_RESV_USAGE_READ);
+	 * to report the overall busyness. This is what the wait-ioctl does.
+	 *
+	 */
+	args->busy = 0;
+	dma_resv_iter_begin(&cursor, obj->base.resv, DMA_RESV_USAGE_READ);
+	dma_resv_for_each_fence_unlocked(&cursor, fence) {
+		if (dma_resv_iter_is_restarted(&cursor))
+			args->busy = 0;
+
+		if (dma_resv_iter_usage(&cursor) <= DMA_RESV_USAGE_WRITE)
+			/* Translate the write fences to the READ *and* WRITE engine */
+			args->busy |= busy_check_writer(fence);
+		else
+			/* Translate read fences to READ set of engines */
+			args->busy |= busy_check_reader(fence);
+	}
+	dma_resv_iter_end(&cursor);
+
+	err = 0;
+out:
+	rcu_read_unlock();
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_clflush.c b/drivers/gpu/drm/i915/gem/i915_gem_clflush.c
new file mode 100644
index 000000000..0512afdd2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_clflush.c
@@ -0,0 +1,138 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include <drm/drm_cache.h>
+
+#include "display/intel_frontbuffer.h"
+
+#include "i915_drv.h"
+#include "i915_gem_clflush.h"
+#include "i915_sw_fence_work.h"
+#include "i915_trace.h"
+
+struct clflush {
+	struct dma_fence_work base;
+	struct drm_i915_gem_object *obj;
+};
+
+static void __do_clflush(struct drm_i915_gem_object *obj)
+{
+	GEM_BUG_ON(!i915_gem_object_has_pages(obj));
+	drm_clflush_sg(obj->mm.pages);
+
+	i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
+}
+
+static void clflush_work(struct dma_fence_work *base)
+{
+	struct clflush *clflush = container_of(base, typeof(*clflush), base);
+
+	__do_clflush(clflush->obj);
+}
+
+static void clflush_release(struct dma_fence_work *base)
+{
+	struct clflush *clflush = container_of(base, typeof(*clflush), base);
+
+	i915_gem_object_unpin_pages(clflush->obj);
+	i915_gem_object_put(clflush->obj);
+}
+
+static const struct dma_fence_work_ops clflush_ops = {
+	.name = "clflush",
+	.work = clflush_work,
+	.release = clflush_release,
+};
+
+static struct clflush *clflush_work_create(struct drm_i915_gem_object *obj)
+{
+	struct clflush *clflush;
+
+	GEM_BUG_ON(!obj->cache_dirty);
+
+	clflush = kmalloc(sizeof(*clflush), GFP_KERNEL);
+	if (!clflush)
+		return NULL;
+
+	if (__i915_gem_object_get_pages(obj) < 0) {
+		kfree(clflush);
+		return NULL;
+	}
+
+	dma_fence_work_init(&clflush->base, &clflush_ops);
+	clflush->obj = i915_gem_object_get(obj); /* obj <-> clflush cycle */
+
+	return clflush;
+}
+
+bool i915_gem_clflush_object(struct drm_i915_gem_object *obj,
+			     unsigned int flags)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct clflush *clflush;
+
+	assert_object_held(obj);
+
+	if (IS_DGFX(i915)) {
+		WARN_ON_ONCE(obj->cache_dirty);
+		return false;
+	}
+
+	/*
+	 * Stolen memory is always coherent with the GPU as it is explicitly
+	 * marked as wc by the system, or the system is cache-coherent.
+	 * Similarly, we only access struct pages through the CPU cache, so
+	 * anything not backed by physical memory we consider to be always
+	 * coherent and not need clflushing.
+	 */
+	if (!i915_gem_object_has_struct_page(obj)) {
+		obj->cache_dirty = false;
+		return false;
+	}
+
+	/* If the GPU is snooping the contents of the CPU cache,
+	 * we do not need to manually clear the CPU cache lines.  However,
+	 * the caches are only snooped when the render cache is
+	 * flushed/invalidated.  As we always have to emit invalidations
+	 * and flushes when moving into and out of the RENDER domain, correct
+	 * snooping behaviour occurs naturally as the result of our domain
+	 * tracking.
+	 */
+	if (!(flags & I915_CLFLUSH_FORCE) &&
+	    obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ)
+		return false;
+
+	trace_i915_gem_object_clflush(obj);
+
+	clflush = NULL;
+	if (!(flags & I915_CLFLUSH_SYNC) &&
+	    dma_resv_reserve_fences(obj->base.resv, 1) == 0)
+		clflush = clflush_work_create(obj);
+	if (clflush) {
+		i915_sw_fence_await_reservation(&clflush->base.chain,
+						obj->base.resv, NULL, true,
+						i915_fence_timeout(i915),
+						I915_FENCE_GFP);
+		dma_resv_add_fence(obj->base.resv, &clflush->base.dma,
+				   DMA_RESV_USAGE_KERNEL);
+		dma_fence_work_commit(&clflush->base);
+		/*
+		 * We must have successfully populated the pages(since we are
+		 * holding a pin on the pages as per the flush worker) to reach
+		 * this point, which must mean we have already done the required
+		 * flush-on-acquire, hence resetting cache_dirty here should be
+		 * safe.
+		 */
+		obj->cache_dirty = false;
+	} else if (obj->mm.pages) {
+		__do_clflush(obj);
+		obj->cache_dirty = false;
+	} else {
+		GEM_BUG_ON(obj->write_domain != I915_GEM_DOMAIN_CPU);
+	}
+
+	return true;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_clflush.h b/drivers/gpu/drm/i915/gem/i915_gem_clflush.h
new file mode 100644
index 000000000..e6c382973
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_clflush.h
@@ -0,0 +1,20 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __I915_GEM_CLFLUSH_H__
+#define __I915_GEM_CLFLUSH_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct drm_i915_gem_object;
+
+bool i915_gem_clflush_object(struct drm_i915_gem_object *obj,
+			     unsigned int flags);
+#define I915_CLFLUSH_FORCE BIT(0)
+#define I915_CLFLUSH_SYNC BIT(1)
+
+#endif /* __I915_GEM_CLFLUSH_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
new file mode 100644
index 000000000..5e1b11db7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
@@ -0,0 +1,2597 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2011-2012 Intel Corporation
+ */
+
+/*
+ * This file implements HW context support. On gen5+ a HW context consists of an
+ * opaque GPU object which is referenced at times of context saves and restores.
+ * With RC6 enabled, the context is also referenced as the GPU enters and exists
+ * from RC6 (GPU has it's own internal power context, except on gen5). Though
+ * something like a context does exist for the media ring, the code only
+ * supports contexts for the render ring.
+ *
+ * In software, there is a distinction between contexts created by the user,
+ * and the default HW context. The default HW context is used by GPU clients
+ * that do not request setup of their own hardware context. The default
+ * context's state is never restored to help prevent programming errors. This
+ * would happen if a client ran and piggy-backed off another clients GPU state.
+ * The default context only exists to give the GPU some offset to load as the
+ * current to invoke a save of the context we actually care about. In fact, the
+ * code could likely be constructed, albeit in a more complicated fashion, to
+ * never use the default context, though that limits the driver's ability to
+ * swap out, and/or destroy other contexts.
+ *
+ * All other contexts are created as a request by the GPU client. These contexts
+ * store GPU state, and thus allow GPU clients to not re-emit state (and
+ * potentially query certain state) at any time. The kernel driver makes
+ * certain that the appropriate commands are inserted.
+ *
+ * The context life cycle is semi-complicated in that context BOs may live
+ * longer than the context itself because of the way the hardware, and object
+ * tracking works. Below is a very crude representation of the state machine
+ * describing the context life.
+ *                                         refcount     pincount     active
+ * S0: initial state                          0            0           0
+ * S1: context created                        1            0           0
+ * S2: context is currently running           2            1           X
+ * S3: GPU referenced, but not current        2            0           1
+ * S4: context is current, but destroyed      1            1           0
+ * S5: like S3, but destroyed                 1            0           1
+ *
+ * The most common (but not all) transitions:
+ * S0->S1: client creates a context
+ * S1->S2: client submits execbuf with context
+ * S2->S3: other clients submits execbuf with context
+ * S3->S1: context object was retired
+ * S3->S2: clients submits another execbuf
+ * S2->S4: context destroy called with current context
+ * S3->S5->S0: destroy path
+ * S4->S5->S0: destroy path on current context
+ *
+ * There are two confusing terms used above:
+ *  The "current context" means the context which is currently running on the
+ *  GPU. The GPU has loaded its state already and has stored away the gtt
+ *  offset of the BO. The GPU is not actively referencing the data at this
+ *  offset, but it will on the next context switch. The only way to avoid this
+ *  is to do a GPU reset.
+ *
+ *  An "active context' is one which was previously the "current context" and is
+ *  on the active list waiting for the next context switch to occur. Until this
+ *  happens, the object must remain at the same gtt offset. It is therefore
+ *  possible to destroy a context, but it is still active.
+ *
+ */
+
+#include <linux/highmem.h>
+#include <linux/log2.h>
+#include <linux/nospec.h>
+
+#include <drm/drm_cache.h>
+#include <drm/drm_syncobj.h>
+
+#include "gt/gen6_ppgtt.h"
+#include "gt/intel_context.h"
+#include "gt/intel_context_param.h"
+#include "gt/intel_engine_heartbeat.h"
+#include "gt/intel_engine_user.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_ring.h"
+
+#include "pxp/intel_pxp.h"
+
+#include "i915_file_private.h"
+#include "i915_gem_context.h"
+#include "i915_trace.h"
+#include "i915_user_extensions.h"
+
+#define ALL_L3_SLICES(dev) (1 << NUM_L3_SLICES(dev)) - 1
+
+static struct kmem_cache *slab_luts;
+
+struct i915_lut_handle *i915_lut_handle_alloc(void)
+{
+	return kmem_cache_alloc(slab_luts, GFP_KERNEL);
+}
+
+void i915_lut_handle_free(struct i915_lut_handle *lut)
+{
+	return kmem_cache_free(slab_luts, lut);
+}
+
+static void lut_close(struct i915_gem_context *ctx)
+{
+	struct radix_tree_iter iter;
+	void __rcu **slot;
+
+	mutex_lock(&ctx->lut_mutex);
+	rcu_read_lock();
+	radix_tree_for_each_slot(slot, &ctx->handles_vma, &iter, 0) {
+		struct i915_vma *vma = rcu_dereference_raw(*slot);
+		struct drm_i915_gem_object *obj = vma->obj;
+		struct i915_lut_handle *lut;
+
+		if (!kref_get_unless_zero(&obj->base.refcount))
+			continue;
+
+		spin_lock(&obj->lut_lock);
+		list_for_each_entry(lut, &obj->lut_list, obj_link) {
+			if (lut->ctx != ctx)
+				continue;
+
+			if (lut->handle != iter.index)
+				continue;
+
+			list_del(&lut->obj_link);
+			break;
+		}
+		spin_unlock(&obj->lut_lock);
+
+		if (&lut->obj_link != &obj->lut_list) {
+			i915_lut_handle_free(lut);
+			radix_tree_iter_delete(&ctx->handles_vma, &iter, slot);
+			i915_vma_close(vma);
+			i915_gem_object_put(obj);
+		}
+
+		i915_gem_object_put(obj);
+	}
+	rcu_read_unlock();
+	mutex_unlock(&ctx->lut_mutex);
+}
+
+static struct intel_context *
+lookup_user_engine(struct i915_gem_context *ctx,
+		   unsigned long flags,
+		   const struct i915_engine_class_instance *ci)
+#define LOOKUP_USER_INDEX BIT(0)
+{
+	int idx;
+
+	if (!!(flags & LOOKUP_USER_INDEX) != i915_gem_context_user_engines(ctx))
+		return ERR_PTR(-EINVAL);
+
+	if (!i915_gem_context_user_engines(ctx)) {
+		struct intel_engine_cs *engine;
+
+		engine = intel_engine_lookup_user(ctx->i915,
+						  ci->engine_class,
+						  ci->engine_instance);
+		if (!engine)
+			return ERR_PTR(-EINVAL);
+
+		idx = engine->legacy_idx;
+	} else {
+		idx = ci->engine_instance;
+	}
+
+	return i915_gem_context_get_engine(ctx, idx);
+}
+
+static int validate_priority(struct drm_i915_private *i915,
+			     const struct drm_i915_gem_context_param *args)
+{
+	s64 priority = args->value;
+
+	if (args->size)
+		return -EINVAL;
+
+	if (!(i915->caps.scheduler & I915_SCHEDULER_CAP_PRIORITY))
+		return -ENODEV;
+
+	if (priority > I915_CONTEXT_MAX_USER_PRIORITY ||
+	    priority < I915_CONTEXT_MIN_USER_PRIORITY)
+		return -EINVAL;
+
+	if (priority > I915_CONTEXT_DEFAULT_PRIORITY &&
+	    !capable(CAP_SYS_NICE))
+		return -EPERM;
+
+	return 0;
+}
+
+static void proto_context_close(struct drm_i915_private *i915,
+				struct i915_gem_proto_context *pc)
+{
+	int i;
+
+	if (pc->pxp_wakeref)
+		intel_runtime_pm_put(&i915->runtime_pm, pc->pxp_wakeref);
+	if (pc->vm)
+		i915_vm_put(pc->vm);
+	if (pc->user_engines) {
+		for (i = 0; i < pc->num_user_engines; i++)
+			kfree(pc->user_engines[i].siblings);
+		kfree(pc->user_engines);
+	}
+	kfree(pc);
+}
+
+static int proto_context_set_persistence(struct drm_i915_private *i915,
+					 struct i915_gem_proto_context *pc,
+					 bool persist)
+{
+	if (persist) {
+		/*
+		 * Only contexts that are short-lived [that will expire or be
+		 * reset] are allowed to survive past termination. We require
+		 * hangcheck to ensure that the persistent requests are healthy.
+		 */
+		if (!i915->params.enable_hangcheck)
+			return -EINVAL;
+
+		pc->user_flags |= BIT(UCONTEXT_PERSISTENCE);
+	} else {
+		/* To cancel a context we use "preempt-to-idle" */
+		if (!(i915->caps.scheduler & I915_SCHEDULER_CAP_PREEMPTION))
+			return -ENODEV;
+
+		/*
+		 * If the cancel fails, we then need to reset, cleanly!
+		 *
+		 * If the per-engine reset fails, all hope is lost! We resort
+		 * to a full GPU reset in that unlikely case, but realistically
+		 * if the engine could not reset, the full reset does not fare
+		 * much better. The damage has been done.
+		 *
+		 * However, if we cannot reset an engine by itself, we cannot
+		 * cleanup a hanging persistent context without causing
+		 * colateral damage, and we should not pretend we can by
+		 * exposing the interface.
+		 */
+		if (!intel_has_reset_engine(to_gt(i915)))
+			return -ENODEV;
+
+		pc->user_flags &= ~BIT(UCONTEXT_PERSISTENCE);
+	}
+
+	return 0;
+}
+
+static int proto_context_set_protected(struct drm_i915_private *i915,
+				       struct i915_gem_proto_context *pc,
+				       bool protected)
+{
+	int ret = 0;
+
+	if (!protected) {
+		pc->uses_protected_content = false;
+	} else if (!intel_pxp_is_enabled(&to_gt(i915)->pxp)) {
+		ret = -ENODEV;
+	} else if ((pc->user_flags & BIT(UCONTEXT_RECOVERABLE)) ||
+		   !(pc->user_flags & BIT(UCONTEXT_BANNABLE))) {
+		ret = -EPERM;
+	} else {
+		pc->uses_protected_content = true;
+
+		/*
+		 * protected context usage requires the PXP session to be up,
+		 * which in turn requires the device to be active.
+		 */
+		pc->pxp_wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+		if (!intel_pxp_is_active(&to_gt(i915)->pxp))
+			ret = intel_pxp_start(&to_gt(i915)->pxp);
+	}
+
+	return ret;
+}
+
+static struct i915_gem_proto_context *
+proto_context_create(struct drm_i915_private *i915, unsigned int flags)
+{
+	struct i915_gem_proto_context *pc, *err;
+
+	pc = kzalloc(sizeof(*pc), GFP_KERNEL);
+	if (!pc)
+		return ERR_PTR(-ENOMEM);
+
+	pc->num_user_engines = -1;
+	pc->user_engines = NULL;
+	pc->user_flags = BIT(UCONTEXT_BANNABLE) |
+			 BIT(UCONTEXT_RECOVERABLE);
+	if (i915->params.enable_hangcheck)
+		pc->user_flags |= BIT(UCONTEXT_PERSISTENCE);
+	pc->sched.priority = I915_PRIORITY_NORMAL;
+
+	if (flags & I915_CONTEXT_CREATE_FLAGS_SINGLE_TIMELINE) {
+		if (!HAS_EXECLISTS(i915)) {
+			err = ERR_PTR(-EINVAL);
+			goto proto_close;
+		}
+		pc->single_timeline = true;
+	}
+
+	return pc;
+
+proto_close:
+	proto_context_close(i915, pc);
+	return err;
+}
+
+static int proto_context_register_locked(struct drm_i915_file_private *fpriv,
+					 struct i915_gem_proto_context *pc,
+					 u32 *id)
+{
+	int ret;
+	void *old;
+
+	lockdep_assert_held(&fpriv->proto_context_lock);
+
+	ret = xa_alloc(&fpriv->context_xa, id, NULL, xa_limit_32b, GFP_KERNEL);
+	if (ret)
+		return ret;
+
+	old = xa_store(&fpriv->proto_context_xa, *id, pc, GFP_KERNEL);
+	if (xa_is_err(old)) {
+		xa_erase(&fpriv->context_xa, *id);
+		return xa_err(old);
+	}
+	WARN_ON(old);
+
+	return 0;
+}
+
+static int proto_context_register(struct drm_i915_file_private *fpriv,
+				  struct i915_gem_proto_context *pc,
+				  u32 *id)
+{
+	int ret;
+
+	mutex_lock(&fpriv->proto_context_lock);
+	ret = proto_context_register_locked(fpriv, pc, id);
+	mutex_unlock(&fpriv->proto_context_lock);
+
+	return ret;
+}
+
+static struct i915_address_space *
+i915_gem_vm_lookup(struct drm_i915_file_private *file_priv, u32 id)
+{
+	struct i915_address_space *vm;
+
+	xa_lock(&file_priv->vm_xa);
+	vm = xa_load(&file_priv->vm_xa, id);
+	if (vm)
+		kref_get(&vm->ref);
+	xa_unlock(&file_priv->vm_xa);
+
+	return vm;
+}
+
+static int set_proto_ctx_vm(struct drm_i915_file_private *fpriv,
+			    struct i915_gem_proto_context *pc,
+			    const struct drm_i915_gem_context_param *args)
+{
+	struct drm_i915_private *i915 = fpriv->dev_priv;
+	struct i915_address_space *vm;
+
+	if (args->size)
+		return -EINVAL;
+
+	if (!HAS_FULL_PPGTT(i915))
+		return -ENODEV;
+
+	if (upper_32_bits(args->value))
+		return -ENOENT;
+
+	vm = i915_gem_vm_lookup(fpriv, args->value);
+	if (!vm)
+		return -ENOENT;
+
+	if (pc->vm)
+		i915_vm_put(pc->vm);
+	pc->vm = vm;
+
+	return 0;
+}
+
+struct set_proto_ctx_engines {
+	struct drm_i915_private *i915;
+	unsigned num_engines;
+	struct i915_gem_proto_engine *engines;
+};
+
+static int
+set_proto_ctx_engines_balance(struct i915_user_extension __user *base,
+			      void *data)
+{
+	struct i915_context_engines_load_balance __user *ext =
+		container_of_user(base, typeof(*ext), base);
+	const struct set_proto_ctx_engines *set = data;
+	struct drm_i915_private *i915 = set->i915;
+	struct intel_engine_cs **siblings;
+	u16 num_siblings, idx;
+	unsigned int n;
+	int err;
+
+	if (!HAS_EXECLISTS(i915))
+		return -ENODEV;
+
+	if (get_user(idx, &ext->engine_index))
+		return -EFAULT;
+
+	if (idx >= set->num_engines) {
+		drm_dbg(&i915->drm, "Invalid placement value, %d >= %d\n",
+			idx, set->num_engines);
+		return -EINVAL;
+	}
+
+	idx = array_index_nospec(idx, set->num_engines);
+	if (set->engines[idx].type != I915_GEM_ENGINE_TYPE_INVALID) {
+		drm_dbg(&i915->drm,
+			"Invalid placement[%d], already occupied\n", idx);
+		return -EEXIST;
+	}
+
+	if (get_user(num_siblings, &ext->num_siblings))
+		return -EFAULT;
+
+	err = check_user_mbz(&ext->flags);
+	if (err)
+		return err;
+
+	err = check_user_mbz(&ext->mbz64);
+	if (err)
+		return err;
+
+	if (num_siblings == 0)
+		return 0;
+
+	siblings = kmalloc_array(num_siblings, sizeof(*siblings), GFP_KERNEL);
+	if (!siblings)
+		return -ENOMEM;
+
+	for (n = 0; n < num_siblings; n++) {
+		struct i915_engine_class_instance ci;
+
+		if (copy_from_user(&ci, &ext->engines[n], sizeof(ci))) {
+			err = -EFAULT;
+			goto err_siblings;
+		}
+
+		siblings[n] = intel_engine_lookup_user(i915,
+						       ci.engine_class,
+						       ci.engine_instance);
+		if (!siblings[n]) {
+			drm_dbg(&i915->drm,
+				"Invalid sibling[%d]: { class:%d, inst:%d }\n",
+				n, ci.engine_class, ci.engine_instance);
+			err = -EINVAL;
+			goto err_siblings;
+		}
+	}
+
+	if (num_siblings == 1) {
+		set->engines[idx].type = I915_GEM_ENGINE_TYPE_PHYSICAL;
+		set->engines[idx].engine = siblings[0];
+		kfree(siblings);
+	} else {
+		set->engines[idx].type = I915_GEM_ENGINE_TYPE_BALANCED;
+		set->engines[idx].num_siblings = num_siblings;
+		set->engines[idx].siblings = siblings;
+	}
+
+	return 0;
+
+err_siblings:
+	kfree(siblings);
+
+	return err;
+}
+
+static int
+set_proto_ctx_engines_bond(struct i915_user_extension __user *base, void *data)
+{
+	struct i915_context_engines_bond __user *ext =
+		container_of_user(base, typeof(*ext), base);
+	const struct set_proto_ctx_engines *set = data;
+	struct drm_i915_private *i915 = set->i915;
+	struct i915_engine_class_instance ci;
+	struct intel_engine_cs *master;
+	u16 idx, num_bonds;
+	int err, n;
+
+	if (GRAPHICS_VER(i915) >= 12 && !IS_TIGERLAKE(i915) &&
+	    !IS_ROCKETLAKE(i915) && !IS_ALDERLAKE_S(i915)) {
+		drm_dbg(&i915->drm,
+			"Bonding not supported on this platform\n");
+		return -ENODEV;
+	}
+
+	if (get_user(idx, &ext->virtual_index))
+		return -EFAULT;
+
+	if (idx >= set->num_engines) {
+		drm_dbg(&i915->drm,
+			"Invalid index for virtual engine: %d >= %d\n",
+			idx, set->num_engines);
+		return -EINVAL;
+	}
+
+	idx = array_index_nospec(idx, set->num_engines);
+	if (set->engines[idx].type == I915_GEM_ENGINE_TYPE_INVALID) {
+		drm_dbg(&i915->drm, "Invalid engine at %d\n", idx);
+		return -EINVAL;
+	}
+
+	if (set->engines[idx].type != I915_GEM_ENGINE_TYPE_PHYSICAL) {
+		drm_dbg(&i915->drm,
+			"Bonding with virtual engines not allowed\n");
+		return -EINVAL;
+	}
+
+	err = check_user_mbz(&ext->flags);
+	if (err)
+		return err;
+
+	for (n = 0; n < ARRAY_SIZE(ext->mbz64); n++) {
+		err = check_user_mbz(&ext->mbz64[n]);
+		if (err)
+			return err;
+	}
+
+	if (copy_from_user(&ci, &ext->master, sizeof(ci)))
+		return -EFAULT;
+
+	master = intel_engine_lookup_user(i915,
+					  ci.engine_class,
+					  ci.engine_instance);
+	if (!master) {
+		drm_dbg(&i915->drm,
+			"Unrecognised master engine: { class:%u, instance:%u }\n",
+			ci.engine_class, ci.engine_instance);
+		return -EINVAL;
+	}
+
+	if (intel_engine_uses_guc(master)) {
+		DRM_DEBUG("bonding extension not supported with GuC submission");
+		return -ENODEV;
+	}
+
+	if (get_user(num_bonds, &ext->num_bonds))
+		return -EFAULT;
+
+	for (n = 0; n < num_bonds; n++) {
+		struct intel_engine_cs *bond;
+
+		if (copy_from_user(&ci, &ext->engines[n], sizeof(ci)))
+			return -EFAULT;
+
+		bond = intel_engine_lookup_user(i915,
+						ci.engine_class,
+						ci.engine_instance);
+		if (!bond) {
+			drm_dbg(&i915->drm,
+				"Unrecognised engine[%d] for bonding: { class:%d, instance: %d }\n",
+				n, ci.engine_class, ci.engine_instance);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static int
+set_proto_ctx_engines_parallel_submit(struct i915_user_extension __user *base,
+				      void *data)
+{
+	struct i915_context_engines_parallel_submit __user *ext =
+		container_of_user(base, typeof(*ext), base);
+	const struct set_proto_ctx_engines *set = data;
+	struct drm_i915_private *i915 = set->i915;
+	struct i915_engine_class_instance prev_engine;
+	u64 flags;
+	int err = 0, n, i, j;
+	u16 slot, width, num_siblings;
+	struct intel_engine_cs **siblings = NULL;
+	intel_engine_mask_t prev_mask;
+
+	if (get_user(slot, &ext->engine_index))
+		return -EFAULT;
+
+	if (get_user(width, &ext->width))
+		return -EFAULT;
+
+	if (get_user(num_siblings, &ext->num_siblings))
+		return -EFAULT;
+
+	if (!intel_uc_uses_guc_submission(&to_gt(i915)->uc) &&
+	    num_siblings != 1) {
+		drm_dbg(&i915->drm, "Only 1 sibling (%d) supported in non-GuC mode\n",
+			num_siblings);
+		return -EINVAL;
+	}
+
+	if (slot >= set->num_engines) {
+		drm_dbg(&i915->drm, "Invalid placement value, %d >= %d\n",
+			slot, set->num_engines);
+		return -EINVAL;
+	}
+
+	if (set->engines[slot].type != I915_GEM_ENGINE_TYPE_INVALID) {
+		drm_dbg(&i915->drm,
+			"Invalid placement[%d], already occupied\n", slot);
+		return -EINVAL;
+	}
+
+	if (get_user(flags, &ext->flags))
+		return -EFAULT;
+
+	if (flags) {
+		drm_dbg(&i915->drm, "Unknown flags 0x%02llx", flags);
+		return -EINVAL;
+	}
+
+	for (n = 0; n < ARRAY_SIZE(ext->mbz64); n++) {
+		err = check_user_mbz(&ext->mbz64[n]);
+		if (err)
+			return err;
+	}
+
+	if (width < 2) {
+		drm_dbg(&i915->drm, "Width (%d) < 2\n", width);
+		return -EINVAL;
+	}
+
+	if (num_siblings < 1) {
+		drm_dbg(&i915->drm, "Number siblings (%d) < 1\n",
+			num_siblings);
+		return -EINVAL;
+	}
+
+	siblings = kmalloc_array(num_siblings * width,
+				 sizeof(*siblings),
+				 GFP_KERNEL);
+	if (!siblings)
+		return -ENOMEM;
+
+	/* Create contexts / engines */
+	for (i = 0; i < width; ++i) {
+		intel_engine_mask_t current_mask = 0;
+
+		for (j = 0; j < num_siblings; ++j) {
+			struct i915_engine_class_instance ci;
+
+			n = i * num_siblings + j;
+			if (copy_from_user(&ci, &ext->engines[n], sizeof(ci))) {
+				err = -EFAULT;
+				goto out_err;
+			}
+
+			siblings[n] =
+				intel_engine_lookup_user(i915, ci.engine_class,
+							 ci.engine_instance);
+			if (!siblings[n]) {
+				drm_dbg(&i915->drm,
+					"Invalid sibling[%d]: { class:%d, inst:%d }\n",
+					n, ci.engine_class, ci.engine_instance);
+				err = -EINVAL;
+				goto out_err;
+			}
+
+			/*
+			 * We don't support breadcrumb handshake on these
+			 * classes
+			 */
+			if (siblings[n]->class == RENDER_CLASS ||
+			    siblings[n]->class == COMPUTE_CLASS) {
+				err = -EINVAL;
+				goto out_err;
+			}
+
+			if (n) {
+				if (prev_engine.engine_class !=
+				    ci.engine_class) {
+					drm_dbg(&i915->drm,
+						"Mismatched class %d, %d\n",
+						prev_engine.engine_class,
+						ci.engine_class);
+					err = -EINVAL;
+					goto out_err;
+				}
+			}
+
+			prev_engine = ci;
+			current_mask |= siblings[n]->logical_mask;
+		}
+
+		if (i > 0) {
+			if (current_mask != prev_mask << 1) {
+				drm_dbg(&i915->drm,
+					"Non contiguous logical mask 0x%x, 0x%x\n",
+					prev_mask, current_mask);
+				err = -EINVAL;
+				goto out_err;
+			}
+		}
+		prev_mask = current_mask;
+	}
+
+	set->engines[slot].type = I915_GEM_ENGINE_TYPE_PARALLEL;
+	set->engines[slot].num_siblings = num_siblings;
+	set->engines[slot].width = width;
+	set->engines[slot].siblings = siblings;
+
+	return 0;
+
+out_err:
+	kfree(siblings);
+
+	return err;
+}
+
+static const i915_user_extension_fn set_proto_ctx_engines_extensions[] = {
+	[I915_CONTEXT_ENGINES_EXT_LOAD_BALANCE] = set_proto_ctx_engines_balance,
+	[I915_CONTEXT_ENGINES_EXT_BOND] = set_proto_ctx_engines_bond,
+	[I915_CONTEXT_ENGINES_EXT_PARALLEL_SUBMIT] =
+		set_proto_ctx_engines_parallel_submit,
+};
+
+static int set_proto_ctx_engines(struct drm_i915_file_private *fpriv,
+			         struct i915_gem_proto_context *pc,
+			         const struct drm_i915_gem_context_param *args)
+{
+	struct drm_i915_private *i915 = fpriv->dev_priv;
+	struct set_proto_ctx_engines set = { .i915 = i915 };
+	struct i915_context_param_engines __user *user =
+		u64_to_user_ptr(args->value);
+	unsigned int n;
+	u64 extensions;
+	int err;
+
+	if (pc->num_user_engines >= 0) {
+		drm_dbg(&i915->drm, "Cannot set engines twice");
+		return -EINVAL;
+	}
+
+	if (args->size < sizeof(*user) ||
+	    !IS_ALIGNED(args->size - sizeof(*user), sizeof(*user->engines))) {
+		drm_dbg(&i915->drm, "Invalid size for engine array: %d\n",
+			args->size);
+		return -EINVAL;
+	}
+
+	set.num_engines = (args->size - sizeof(*user)) / sizeof(*user->engines);
+	/* RING_MASK has no shift so we can use it directly here */
+	if (set.num_engines > I915_EXEC_RING_MASK + 1)
+		return -EINVAL;
+
+	set.engines = kmalloc_array(set.num_engines, sizeof(*set.engines), GFP_KERNEL);
+	if (!set.engines)
+		return -ENOMEM;
+
+	for (n = 0; n < set.num_engines; n++) {
+		struct i915_engine_class_instance ci;
+		struct intel_engine_cs *engine;
+
+		if (copy_from_user(&ci, &user->engines[n], sizeof(ci))) {
+			kfree(set.engines);
+			return -EFAULT;
+		}
+
+		memset(&set.engines[n], 0, sizeof(set.engines[n]));
+
+		if (ci.engine_class == (u16)I915_ENGINE_CLASS_INVALID &&
+		    ci.engine_instance == (u16)I915_ENGINE_CLASS_INVALID_NONE)
+			continue;
+
+		engine = intel_engine_lookup_user(i915,
+						  ci.engine_class,
+						  ci.engine_instance);
+		if (!engine) {
+			drm_dbg(&i915->drm,
+				"Invalid engine[%d]: { class:%d, instance:%d }\n",
+				n, ci.engine_class, ci.engine_instance);
+			kfree(set.engines);
+			return -ENOENT;
+		}
+
+		set.engines[n].type = I915_GEM_ENGINE_TYPE_PHYSICAL;
+		set.engines[n].engine = engine;
+	}
+
+	err = -EFAULT;
+	if (!get_user(extensions, &user->extensions))
+		err = i915_user_extensions(u64_to_user_ptr(extensions),
+					   set_proto_ctx_engines_extensions,
+					   ARRAY_SIZE(set_proto_ctx_engines_extensions),
+					   &set);
+	if (err) {
+		kfree(set.engines);
+		return err;
+	}
+
+	pc->num_user_engines = set.num_engines;
+	pc->user_engines = set.engines;
+
+	return 0;
+}
+
+static int set_proto_ctx_sseu(struct drm_i915_file_private *fpriv,
+			      struct i915_gem_proto_context *pc,
+			      struct drm_i915_gem_context_param *args)
+{
+	struct drm_i915_private *i915 = fpriv->dev_priv;
+	struct drm_i915_gem_context_param_sseu user_sseu;
+	struct intel_sseu *sseu;
+	int ret;
+
+	if (args->size < sizeof(user_sseu))
+		return -EINVAL;
+
+	if (GRAPHICS_VER(i915) != 11)
+		return -ENODEV;
+
+	if (copy_from_user(&user_sseu, u64_to_user_ptr(args->value),
+			   sizeof(user_sseu)))
+		return -EFAULT;
+
+	if (user_sseu.rsvd)
+		return -EINVAL;
+
+	if (user_sseu.flags & ~(I915_CONTEXT_SSEU_FLAG_ENGINE_INDEX))
+		return -EINVAL;
+
+	if (!!(user_sseu.flags & I915_CONTEXT_SSEU_FLAG_ENGINE_INDEX) != (pc->num_user_engines >= 0))
+		return -EINVAL;
+
+	if (pc->num_user_engines >= 0) {
+		int idx = user_sseu.engine.engine_instance;
+		struct i915_gem_proto_engine *pe;
+
+		if (idx >= pc->num_user_engines)
+			return -EINVAL;
+
+		idx = array_index_nospec(idx, pc->num_user_engines);
+		pe = &pc->user_engines[idx];
+
+		/* Only render engine supports RPCS configuration. */
+		if (pe->engine->class != RENDER_CLASS)
+			return -EINVAL;
+
+		sseu = &pe->sseu;
+	} else {
+		/* Only render engine supports RPCS configuration. */
+		if (user_sseu.engine.engine_class != I915_ENGINE_CLASS_RENDER)
+			return -EINVAL;
+
+		/* There is only one render engine */
+		if (user_sseu.engine.engine_instance != 0)
+			return -EINVAL;
+
+		sseu = &pc->legacy_rcs_sseu;
+	}
+
+	ret = i915_gem_user_to_context_sseu(to_gt(i915), &user_sseu, sseu);
+	if (ret)
+		return ret;
+
+	args->size = sizeof(user_sseu);
+
+	return 0;
+}
+
+static int set_proto_ctx_param(struct drm_i915_file_private *fpriv,
+			       struct i915_gem_proto_context *pc,
+			       struct drm_i915_gem_context_param *args)
+{
+	int ret = 0;
+
+	switch (args->param) {
+	case I915_CONTEXT_PARAM_NO_ERROR_CAPTURE:
+		if (args->size)
+			ret = -EINVAL;
+		else if (args->value)
+			pc->user_flags |= BIT(UCONTEXT_NO_ERROR_CAPTURE);
+		else
+			pc->user_flags &= ~BIT(UCONTEXT_NO_ERROR_CAPTURE);
+		break;
+
+	case I915_CONTEXT_PARAM_BANNABLE:
+		if (args->size)
+			ret = -EINVAL;
+		else if (!capable(CAP_SYS_ADMIN) && !args->value)
+			ret = -EPERM;
+		else if (args->value)
+			pc->user_flags |= BIT(UCONTEXT_BANNABLE);
+		else if (pc->uses_protected_content)
+			ret = -EPERM;
+		else
+			pc->user_flags &= ~BIT(UCONTEXT_BANNABLE);
+		break;
+
+	case I915_CONTEXT_PARAM_RECOVERABLE:
+		if (args->size)
+			ret = -EINVAL;
+		else if (!args->value)
+			pc->user_flags &= ~BIT(UCONTEXT_RECOVERABLE);
+		else if (pc->uses_protected_content)
+			ret = -EPERM;
+		else
+			pc->user_flags |= BIT(UCONTEXT_RECOVERABLE);
+		break;
+
+	case I915_CONTEXT_PARAM_PRIORITY:
+		ret = validate_priority(fpriv->dev_priv, args);
+		if (!ret)
+			pc->sched.priority = args->value;
+		break;
+
+	case I915_CONTEXT_PARAM_SSEU:
+		ret = set_proto_ctx_sseu(fpriv, pc, args);
+		break;
+
+	case I915_CONTEXT_PARAM_VM:
+		ret = set_proto_ctx_vm(fpriv, pc, args);
+		break;
+
+	case I915_CONTEXT_PARAM_ENGINES:
+		ret = set_proto_ctx_engines(fpriv, pc, args);
+		break;
+
+	case I915_CONTEXT_PARAM_PERSISTENCE:
+		if (args->size)
+			ret = -EINVAL;
+		else
+			ret = proto_context_set_persistence(fpriv->dev_priv, pc,
+							    args->value);
+		break;
+
+	case I915_CONTEXT_PARAM_PROTECTED_CONTENT:
+		ret = proto_context_set_protected(fpriv->dev_priv, pc,
+						  args->value);
+		break;
+
+	case I915_CONTEXT_PARAM_NO_ZEROMAP:
+	case I915_CONTEXT_PARAM_BAN_PERIOD:
+	case I915_CONTEXT_PARAM_RINGSIZE:
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+static int intel_context_set_gem(struct intel_context *ce,
+				 struct i915_gem_context *ctx,
+				 struct intel_sseu sseu)
+{
+	int ret = 0;
+
+	GEM_BUG_ON(rcu_access_pointer(ce->gem_context));
+	RCU_INIT_POINTER(ce->gem_context, ctx);
+
+	GEM_BUG_ON(intel_context_is_pinned(ce));
+	ce->ring_size = SZ_16K;
+
+	i915_vm_put(ce->vm);
+	ce->vm = i915_gem_context_get_eb_vm(ctx);
+
+	if (ctx->sched.priority >= I915_PRIORITY_NORMAL &&
+	    intel_engine_has_timeslices(ce->engine) &&
+	    intel_engine_has_semaphores(ce->engine))
+		__set_bit(CONTEXT_USE_SEMAPHORES, &ce->flags);
+
+	if (CONFIG_DRM_I915_REQUEST_TIMEOUT &&
+	    ctx->i915->params.request_timeout_ms) {
+		unsigned int timeout_ms = ctx->i915->params.request_timeout_ms;
+
+		intel_context_set_watchdog_us(ce, (u64)timeout_ms * 1000);
+	}
+
+	/* A valid SSEU has no zero fields */
+	if (sseu.slice_mask && !WARN_ON(ce->engine->class != RENDER_CLASS))
+		ret = intel_context_reconfigure_sseu(ce, sseu);
+
+	return ret;
+}
+
+static void __unpin_engines(struct i915_gem_engines *e, unsigned int count)
+{
+	while (count--) {
+		struct intel_context *ce = e->engines[count], *child;
+
+		if (!ce || !test_bit(CONTEXT_PERMA_PIN, &ce->flags))
+			continue;
+
+		for_each_child(ce, child)
+			intel_context_unpin(child);
+		intel_context_unpin(ce);
+	}
+}
+
+static void unpin_engines(struct i915_gem_engines *e)
+{
+	__unpin_engines(e, e->num_engines);
+}
+
+static void __free_engines(struct i915_gem_engines *e, unsigned int count)
+{
+	while (count--) {
+		if (!e->engines[count])
+			continue;
+
+		intel_context_put(e->engines[count]);
+	}
+	kfree(e);
+}
+
+static void free_engines(struct i915_gem_engines *e)
+{
+	__free_engines(e, e->num_engines);
+}
+
+static void free_engines_rcu(struct rcu_head *rcu)
+{
+	struct i915_gem_engines *engines =
+		container_of(rcu, struct i915_gem_engines, rcu);
+
+	i915_sw_fence_fini(&engines->fence);
+	free_engines(engines);
+}
+
+static void accumulate_runtime(struct i915_drm_client *client,
+			       struct i915_gem_engines *engines)
+{
+	struct i915_gem_engines_iter it;
+	struct intel_context *ce;
+
+	if (!client)
+		return;
+
+	/* Transfer accumulated runtime to the parent GEM context. */
+	for_each_gem_engine(ce, engines, it) {
+		unsigned int class = ce->engine->uabi_class;
+
+		GEM_BUG_ON(class >= ARRAY_SIZE(client->past_runtime));
+		atomic64_add(intel_context_get_total_runtime_ns(ce),
+			     &client->past_runtime[class]);
+	}
+}
+
+static int
+engines_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
+{
+	struct i915_gem_engines *engines =
+		container_of(fence, typeof(*engines), fence);
+	struct i915_gem_context *ctx = engines->ctx;
+
+	switch (state) {
+	case FENCE_COMPLETE:
+		if (!list_empty(&engines->link)) {
+			unsigned long flags;
+
+			spin_lock_irqsave(&ctx->stale.lock, flags);
+			list_del(&engines->link);
+			spin_unlock_irqrestore(&ctx->stale.lock, flags);
+		}
+		accumulate_runtime(ctx->client, engines);
+		i915_gem_context_put(ctx);
+
+		break;
+
+	case FENCE_FREE:
+		init_rcu_head(&engines->rcu);
+		call_rcu(&engines->rcu, free_engines_rcu);
+		break;
+	}
+
+	return NOTIFY_DONE;
+}
+
+static struct i915_gem_engines *alloc_engines(unsigned int count)
+{
+	struct i915_gem_engines *e;
+
+	e = kzalloc(struct_size(e, engines, count), GFP_KERNEL);
+	if (!e)
+		return NULL;
+
+	i915_sw_fence_init(&e->fence, engines_notify);
+	return e;
+}
+
+static struct i915_gem_engines *default_engines(struct i915_gem_context *ctx,
+						struct intel_sseu rcs_sseu)
+{
+	const struct intel_gt *gt = to_gt(ctx->i915);
+	struct intel_engine_cs *engine;
+	struct i915_gem_engines *e, *err;
+	enum intel_engine_id id;
+
+	e = alloc_engines(I915_NUM_ENGINES);
+	if (!e)
+		return ERR_PTR(-ENOMEM);
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct intel_sseu sseu = {};
+		int ret;
+
+		if (engine->legacy_idx == INVALID_ENGINE)
+			continue;
+
+		GEM_BUG_ON(engine->legacy_idx >= I915_NUM_ENGINES);
+		GEM_BUG_ON(e->engines[engine->legacy_idx]);
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = ERR_CAST(ce);
+			goto free_engines;
+		}
+
+		e->engines[engine->legacy_idx] = ce;
+		e->num_engines = max(e->num_engines, engine->legacy_idx + 1);
+
+		if (engine->class == RENDER_CLASS)
+			sseu = rcs_sseu;
+
+		ret = intel_context_set_gem(ce, ctx, sseu);
+		if (ret) {
+			err = ERR_PTR(ret);
+			goto free_engines;
+		}
+
+	}
+
+	return e;
+
+free_engines:
+	free_engines(e);
+	return err;
+}
+
+static int perma_pin_contexts(struct intel_context *ce)
+{
+	struct intel_context *child;
+	int i = 0, j = 0, ret;
+
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+
+	ret = intel_context_pin(ce);
+	if (unlikely(ret))
+		return ret;
+
+	for_each_child(ce, child) {
+		ret = intel_context_pin(child);
+		if (unlikely(ret))
+			goto unwind;
+		++i;
+	}
+
+	set_bit(CONTEXT_PERMA_PIN, &ce->flags);
+
+	return 0;
+
+unwind:
+	intel_context_unpin(ce);
+	for_each_child(ce, child) {
+		if (j++ < i)
+			intel_context_unpin(child);
+		else
+			break;
+	}
+
+	return ret;
+}
+
+static struct i915_gem_engines *user_engines(struct i915_gem_context *ctx,
+					     unsigned int num_engines,
+					     struct i915_gem_proto_engine *pe)
+{
+	struct i915_gem_engines *e, *err;
+	unsigned int n;
+
+	e = alloc_engines(num_engines);
+	if (!e)
+		return ERR_PTR(-ENOMEM);
+	e->num_engines = num_engines;
+
+	for (n = 0; n < num_engines; n++) {
+		struct intel_context *ce, *child;
+		int ret;
+
+		switch (pe[n].type) {
+		case I915_GEM_ENGINE_TYPE_PHYSICAL:
+			ce = intel_context_create(pe[n].engine);
+			break;
+
+		case I915_GEM_ENGINE_TYPE_BALANCED:
+			ce = intel_engine_create_virtual(pe[n].siblings,
+							 pe[n].num_siblings, 0);
+			break;
+
+		case I915_GEM_ENGINE_TYPE_PARALLEL:
+			ce = intel_engine_create_parallel(pe[n].siblings,
+							  pe[n].num_siblings,
+							  pe[n].width);
+			break;
+
+		case I915_GEM_ENGINE_TYPE_INVALID:
+		default:
+			GEM_WARN_ON(pe[n].type != I915_GEM_ENGINE_TYPE_INVALID);
+			continue;
+		}
+
+		if (IS_ERR(ce)) {
+			err = ERR_CAST(ce);
+			goto free_engines;
+		}
+
+		e->engines[n] = ce;
+
+		ret = intel_context_set_gem(ce, ctx, pe->sseu);
+		if (ret) {
+			err = ERR_PTR(ret);
+			goto free_engines;
+		}
+		for_each_child(ce, child) {
+			ret = intel_context_set_gem(child, ctx, pe->sseu);
+			if (ret) {
+				err = ERR_PTR(ret);
+				goto free_engines;
+			}
+		}
+
+		/*
+		 * XXX: Must be done after calling intel_context_set_gem as that
+		 * function changes the ring size. The ring is allocated when
+		 * the context is pinned. If the ring size is changed after
+		 * allocation we have a mismatch of the ring size and will cause
+		 * the context to hang. Presumably with a bit of reordering we
+		 * could move the perma-pin step to the backend function
+		 * intel_engine_create_parallel.
+		 */
+		if (pe[n].type == I915_GEM_ENGINE_TYPE_PARALLEL) {
+			ret = perma_pin_contexts(ce);
+			if (ret) {
+				err = ERR_PTR(ret);
+				goto free_engines;
+			}
+		}
+	}
+
+	return e;
+
+free_engines:
+	free_engines(e);
+	return err;
+}
+
+static void i915_gem_context_release_work(struct work_struct *work)
+{
+	struct i915_gem_context *ctx = container_of(work, typeof(*ctx),
+						    release_work);
+	struct i915_address_space *vm;
+
+	trace_i915_context_free(ctx);
+	GEM_BUG_ON(!i915_gem_context_is_closed(ctx));
+
+	spin_lock(&ctx->i915->gem.contexts.lock);
+	list_del(&ctx->link);
+	spin_unlock(&ctx->i915->gem.contexts.lock);
+
+	if (ctx->syncobj)
+		drm_syncobj_put(ctx->syncobj);
+
+	vm = ctx->vm;
+	if (vm)
+		i915_vm_put(vm);
+
+	if (ctx->pxp_wakeref)
+		intel_runtime_pm_put(&ctx->i915->runtime_pm, ctx->pxp_wakeref);
+
+	if (ctx->client)
+		i915_drm_client_put(ctx->client);
+
+	mutex_destroy(&ctx->engines_mutex);
+	mutex_destroy(&ctx->lut_mutex);
+
+	put_pid(ctx->pid);
+	mutex_destroy(&ctx->mutex);
+
+	kfree_rcu(ctx, rcu);
+}
+
+void i915_gem_context_release(struct kref *ref)
+{
+	struct i915_gem_context *ctx = container_of(ref, typeof(*ctx), ref);
+
+	queue_work(ctx->i915->wq, &ctx->release_work);
+}
+
+static inline struct i915_gem_engines *
+__context_engines_static(const struct i915_gem_context *ctx)
+{
+	return rcu_dereference_protected(ctx->engines, true);
+}
+
+static void __reset_context(struct i915_gem_context *ctx,
+			    struct intel_engine_cs *engine)
+{
+	intel_gt_handle_error(engine->gt, engine->mask, 0,
+			      "context closure in %s", ctx->name);
+}
+
+static bool __cancel_engine(struct intel_engine_cs *engine)
+{
+	/*
+	 * Send a "high priority pulse" down the engine to cause the
+	 * current request to be momentarily preempted. (If it fails to
+	 * be preempted, it will be reset). As we have marked our context
+	 * as banned, any incomplete request, including any running, will
+	 * be skipped following the preemption.
+	 *
+	 * If there is no hangchecking (one of the reasons why we try to
+	 * cancel the context) and no forced preemption, there may be no
+	 * means by which we reset the GPU and evict the persistent hog.
+	 * Ergo if we are unable to inject a preemptive pulse that can
+	 * kill the banned context, we fallback to doing a local reset
+	 * instead.
+	 */
+	return intel_engine_pulse(engine) == 0;
+}
+
+static struct intel_engine_cs *active_engine(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = NULL;
+	struct i915_request *rq;
+
+	if (intel_context_has_inflight(ce))
+		return intel_context_inflight(ce);
+
+	if (!ce->timeline)
+		return NULL;
+
+	/*
+	 * rq->link is only SLAB_TYPESAFE_BY_RCU, we need to hold a reference
+	 * to the request to prevent it being transferred to a new timeline
+	 * (and onto a new timeline->requests list).
+	 */
+	rcu_read_lock();
+	list_for_each_entry_reverse(rq, &ce->timeline->requests, link) {
+		bool found;
+
+		/* timeline is already completed upto this point? */
+		if (!i915_request_get_rcu(rq))
+			break;
+
+		/* Check with the backend if the request is inflight */
+		found = true;
+		if (likely(rcu_access_pointer(rq->timeline) == ce->timeline))
+			found = i915_request_active_engine(rq, &engine);
+
+		i915_request_put(rq);
+		if (found)
+			break;
+	}
+	rcu_read_unlock();
+
+	return engine;
+}
+
+static void
+kill_engines(struct i915_gem_engines *engines, bool exit, bool persistent)
+{
+	struct i915_gem_engines_iter it;
+	struct intel_context *ce;
+
+	/*
+	 * Map the user's engine back to the actual engines; one virtual
+	 * engine will be mapped to multiple engines, and using ctx->engine[]
+	 * the same engine may be have multiple instances in the user's map.
+	 * However, we only care about pending requests, so only include
+	 * engines on which there are incomplete requests.
+	 */
+	for_each_gem_engine(ce, engines, it) {
+		struct intel_engine_cs *engine;
+
+		if ((exit || !persistent) && intel_context_revoke(ce))
+			continue; /* Already marked. */
+
+		/*
+		 * Check the current active state of this context; if we
+		 * are currently executing on the GPU we need to evict
+		 * ourselves. On the other hand, if we haven't yet been
+		 * submitted to the GPU or if everything is complete,
+		 * we have nothing to do.
+		 */
+		engine = active_engine(ce);
+
+		/* First attempt to gracefully cancel the context */
+		if (engine && !__cancel_engine(engine) && (exit || !persistent))
+			/*
+			 * If we are unable to send a preemptive pulse to bump
+			 * the context from the GPU, we have to resort to a full
+			 * reset. We hope the collateral damage is worth it.
+			 */
+			__reset_context(engines->ctx, engine);
+	}
+}
+
+static void kill_context(struct i915_gem_context *ctx)
+{
+	struct i915_gem_engines *pos, *next;
+
+	spin_lock_irq(&ctx->stale.lock);
+	GEM_BUG_ON(!i915_gem_context_is_closed(ctx));
+	list_for_each_entry_safe(pos, next, &ctx->stale.engines, link) {
+		if (!i915_sw_fence_await(&pos->fence)) {
+			list_del_init(&pos->link);
+			continue;
+		}
+
+		spin_unlock_irq(&ctx->stale.lock);
+
+		kill_engines(pos, !ctx->i915->params.enable_hangcheck,
+			     i915_gem_context_is_persistent(ctx));
+
+		spin_lock_irq(&ctx->stale.lock);
+		GEM_BUG_ON(i915_sw_fence_signaled(&pos->fence));
+		list_safe_reset_next(pos, next, link);
+		list_del_init(&pos->link); /* decouple from FENCE_COMPLETE */
+
+		i915_sw_fence_complete(&pos->fence);
+	}
+	spin_unlock_irq(&ctx->stale.lock);
+}
+
+static void engines_idle_release(struct i915_gem_context *ctx,
+				 struct i915_gem_engines *engines)
+{
+	struct i915_gem_engines_iter it;
+	struct intel_context *ce;
+
+	INIT_LIST_HEAD(&engines->link);
+
+	engines->ctx = i915_gem_context_get(ctx);
+
+	for_each_gem_engine(ce, engines, it) {
+		int err;
+
+		/* serialises with execbuf */
+		set_bit(CONTEXT_CLOSED_BIT, &ce->flags);
+		if (!intel_context_pin_if_active(ce))
+			continue;
+
+		/* Wait until context is finally scheduled out and retired */
+		err = i915_sw_fence_await_active(&engines->fence,
+						 &ce->active,
+						 I915_ACTIVE_AWAIT_BARRIER);
+		intel_context_unpin(ce);
+		if (err)
+			goto kill;
+	}
+
+	spin_lock_irq(&ctx->stale.lock);
+	if (!i915_gem_context_is_closed(ctx))
+		list_add_tail(&engines->link, &ctx->stale.engines);
+	spin_unlock_irq(&ctx->stale.lock);
+
+kill:
+	if (list_empty(&engines->link)) /* raced, already closed */
+		kill_engines(engines, true,
+			     i915_gem_context_is_persistent(ctx));
+
+	i915_sw_fence_commit(&engines->fence);
+}
+
+static void set_closed_name(struct i915_gem_context *ctx)
+{
+	char *s;
+
+	/* Replace '[]' with '<>' to indicate closed in debug prints */
+
+	s = strrchr(ctx->name, '[');
+	if (!s)
+		return;
+
+	*s = '<';
+
+	s = strchr(s + 1, ']');
+	if (s)
+		*s = '>';
+}
+
+static void context_close(struct i915_gem_context *ctx)
+{
+	struct i915_drm_client *client;
+
+	/* Flush any concurrent set_engines() */
+	mutex_lock(&ctx->engines_mutex);
+	unpin_engines(__context_engines_static(ctx));
+	engines_idle_release(ctx, rcu_replace_pointer(ctx->engines, NULL, 1));
+	i915_gem_context_set_closed(ctx);
+	mutex_unlock(&ctx->engines_mutex);
+
+	mutex_lock(&ctx->mutex);
+
+	set_closed_name(ctx);
+
+	/*
+	 * The LUT uses the VMA as a backpointer to unref the object,
+	 * so we need to clear the LUT before we close all the VMA (inside
+	 * the ppgtt).
+	 */
+	lut_close(ctx);
+
+	ctx->file_priv = ERR_PTR(-EBADF);
+
+	client = ctx->client;
+	if (client) {
+		spin_lock(&client->ctx_lock);
+		list_del_rcu(&ctx->client_link);
+		spin_unlock(&client->ctx_lock);
+	}
+
+	mutex_unlock(&ctx->mutex);
+
+	/*
+	 * If the user has disabled hangchecking, we can not be sure that
+	 * the batches will ever complete after the context is closed,
+	 * keeping the context and all resources pinned forever. So in this
+	 * case we opt to forcibly kill off all remaining requests on
+	 * context close.
+	 */
+	kill_context(ctx);
+
+	i915_gem_context_put(ctx);
+}
+
+static int __context_set_persistence(struct i915_gem_context *ctx, bool state)
+{
+	if (i915_gem_context_is_persistent(ctx) == state)
+		return 0;
+
+	if (state) {
+		/*
+		 * Only contexts that are short-lived [that will expire or be
+		 * reset] are allowed to survive past termination. We require
+		 * hangcheck to ensure that the persistent requests are healthy.
+		 */
+		if (!ctx->i915->params.enable_hangcheck)
+			return -EINVAL;
+
+		i915_gem_context_set_persistence(ctx);
+	} else {
+		/* To cancel a context we use "preempt-to-idle" */
+		if (!(ctx->i915->caps.scheduler & I915_SCHEDULER_CAP_PREEMPTION))
+			return -ENODEV;
+
+		/*
+		 * If the cancel fails, we then need to reset, cleanly!
+		 *
+		 * If the per-engine reset fails, all hope is lost! We resort
+		 * to a full GPU reset in that unlikely case, but realistically
+		 * if the engine could not reset, the full reset does not fare
+		 * much better. The damage has been done.
+		 *
+		 * However, if we cannot reset an engine by itself, we cannot
+		 * cleanup a hanging persistent context without causing
+		 * colateral damage, and we should not pretend we can by
+		 * exposing the interface.
+		 */
+		if (!intel_has_reset_engine(to_gt(ctx->i915)))
+			return -ENODEV;
+
+		i915_gem_context_clear_persistence(ctx);
+	}
+
+	return 0;
+}
+
+static struct i915_gem_context *
+i915_gem_create_context(struct drm_i915_private *i915,
+			const struct i915_gem_proto_context *pc)
+{
+	struct i915_gem_context *ctx;
+	struct i915_address_space *vm = NULL;
+	struct i915_gem_engines *e;
+	int err;
+	int i;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return ERR_PTR(-ENOMEM);
+
+	kref_init(&ctx->ref);
+	ctx->i915 = i915;
+	ctx->sched = pc->sched;
+	mutex_init(&ctx->mutex);
+	INIT_LIST_HEAD(&ctx->link);
+	INIT_WORK(&ctx->release_work, i915_gem_context_release_work);
+
+	spin_lock_init(&ctx->stale.lock);
+	INIT_LIST_HEAD(&ctx->stale.engines);
+
+	if (pc->vm) {
+		vm = i915_vm_get(pc->vm);
+	} else if (HAS_FULL_PPGTT(i915)) {
+		struct i915_ppgtt *ppgtt;
+
+		ppgtt = i915_ppgtt_create(to_gt(i915), 0);
+		if (IS_ERR(ppgtt)) {
+			drm_dbg(&i915->drm, "PPGTT setup failed (%ld)\n",
+				PTR_ERR(ppgtt));
+			err = PTR_ERR(ppgtt);
+			goto err_ctx;
+		}
+		vm = &ppgtt->vm;
+	}
+	if (vm)
+		ctx->vm = vm;
+
+	mutex_init(&ctx->engines_mutex);
+	if (pc->num_user_engines >= 0) {
+		i915_gem_context_set_user_engines(ctx);
+		e = user_engines(ctx, pc->num_user_engines, pc->user_engines);
+	} else {
+		i915_gem_context_clear_user_engines(ctx);
+		e = default_engines(ctx, pc->legacy_rcs_sseu);
+	}
+	if (IS_ERR(e)) {
+		err = PTR_ERR(e);
+		goto err_vm;
+	}
+	RCU_INIT_POINTER(ctx->engines, e);
+
+	INIT_RADIX_TREE(&ctx->handles_vma, GFP_KERNEL);
+	mutex_init(&ctx->lut_mutex);
+
+	/* NB: Mark all slices as needing a remap so that when the context first
+	 * loads it will restore whatever remap state already exists. If there
+	 * is no remap info, it will be a NOP. */
+	ctx->remap_slice = ALL_L3_SLICES(i915);
+
+	ctx->user_flags = pc->user_flags;
+
+	for (i = 0; i < ARRAY_SIZE(ctx->hang_timestamp); i++)
+		ctx->hang_timestamp[i] = jiffies - CONTEXT_FAST_HANG_JIFFIES;
+
+	if (pc->single_timeline) {
+		err = drm_syncobj_create(&ctx->syncobj,
+					 DRM_SYNCOBJ_CREATE_SIGNALED,
+					 NULL);
+		if (err)
+			goto err_engines;
+	}
+
+	if (pc->uses_protected_content) {
+		ctx->pxp_wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+		ctx->uses_protected_content = true;
+	}
+
+	trace_i915_context_create(ctx);
+
+	return ctx;
+
+err_engines:
+	free_engines(e);
+err_vm:
+	if (ctx->vm)
+		i915_vm_put(ctx->vm);
+err_ctx:
+	kfree(ctx);
+	return ERR_PTR(err);
+}
+
+static void init_contexts(struct i915_gem_contexts *gc)
+{
+	spin_lock_init(&gc->lock);
+	INIT_LIST_HEAD(&gc->list);
+}
+
+void i915_gem_init__contexts(struct drm_i915_private *i915)
+{
+	init_contexts(&i915->gem.contexts);
+}
+
+/*
+ * Note that this implicitly consumes the ctx reference, by placing
+ * the ctx in the context_xa.
+ */
+static void gem_context_register(struct i915_gem_context *ctx,
+				 struct drm_i915_file_private *fpriv,
+				 u32 id)
+{
+	struct drm_i915_private *i915 = ctx->i915;
+	void *old;
+
+	ctx->file_priv = fpriv;
+
+	ctx->pid = get_task_pid(current, PIDTYPE_PID);
+	ctx->client = i915_drm_client_get(fpriv->client);
+
+	snprintf(ctx->name, sizeof(ctx->name), "%s[%d]",
+		 current->comm, pid_nr(ctx->pid));
+
+	spin_lock(&ctx->client->ctx_lock);
+	list_add_tail_rcu(&ctx->client_link, &ctx->client->ctx_list);
+	spin_unlock(&ctx->client->ctx_lock);
+
+	spin_lock(&i915->gem.contexts.lock);
+	list_add_tail(&ctx->link, &i915->gem.contexts.list);
+	spin_unlock(&i915->gem.contexts.lock);
+
+	/* And finally expose ourselves to userspace via the idr */
+	old = xa_store(&fpriv->context_xa, id, ctx, GFP_KERNEL);
+	WARN_ON(old);
+}
+
+int i915_gem_context_open(struct drm_i915_private *i915,
+			  struct drm_file *file)
+{
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct i915_gem_proto_context *pc;
+	struct i915_gem_context *ctx;
+	int err;
+
+	mutex_init(&file_priv->proto_context_lock);
+	xa_init_flags(&file_priv->proto_context_xa, XA_FLAGS_ALLOC);
+
+	/* 0 reserved for the default context */
+	xa_init_flags(&file_priv->context_xa, XA_FLAGS_ALLOC1);
+
+	/* 0 reserved for invalid/unassigned ppgtt */
+	xa_init_flags(&file_priv->vm_xa, XA_FLAGS_ALLOC1);
+
+	pc = proto_context_create(i915, 0);
+	if (IS_ERR(pc)) {
+		err = PTR_ERR(pc);
+		goto err;
+	}
+
+	ctx = i915_gem_create_context(i915, pc);
+	proto_context_close(i915, pc);
+	if (IS_ERR(ctx)) {
+		err = PTR_ERR(ctx);
+		goto err;
+	}
+
+	gem_context_register(ctx, file_priv, 0);
+
+	return 0;
+
+err:
+	xa_destroy(&file_priv->vm_xa);
+	xa_destroy(&file_priv->context_xa);
+	xa_destroy(&file_priv->proto_context_xa);
+	mutex_destroy(&file_priv->proto_context_lock);
+	return err;
+}
+
+void i915_gem_context_close(struct drm_file *file)
+{
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct i915_gem_proto_context *pc;
+	struct i915_address_space *vm;
+	struct i915_gem_context *ctx;
+	unsigned long idx;
+
+	xa_for_each(&file_priv->proto_context_xa, idx, pc)
+		proto_context_close(file_priv->dev_priv, pc);
+	xa_destroy(&file_priv->proto_context_xa);
+	mutex_destroy(&file_priv->proto_context_lock);
+
+	xa_for_each(&file_priv->context_xa, idx, ctx)
+		context_close(ctx);
+	xa_destroy(&file_priv->context_xa);
+
+	xa_for_each(&file_priv->vm_xa, idx, vm)
+		i915_vm_put(vm);
+	xa_destroy(&file_priv->vm_xa);
+}
+
+int i915_gem_vm_create_ioctl(struct drm_device *dev, void *data,
+			     struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_vm_control *args = data;
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct i915_ppgtt *ppgtt;
+	u32 id;
+	int err;
+
+	if (!HAS_FULL_PPGTT(i915))
+		return -ENODEV;
+
+	if (args->flags)
+		return -EINVAL;
+
+	ppgtt = i915_ppgtt_create(to_gt(i915), 0);
+	if (IS_ERR(ppgtt))
+		return PTR_ERR(ppgtt);
+
+	if (args->extensions) {
+		err = i915_user_extensions(u64_to_user_ptr(args->extensions),
+					   NULL, 0,
+					   ppgtt);
+		if (err)
+			goto err_put;
+	}
+
+	err = xa_alloc(&file_priv->vm_xa, &id, &ppgtt->vm,
+		       xa_limit_32b, GFP_KERNEL);
+	if (err)
+		goto err_put;
+
+	GEM_BUG_ON(id == 0); /* reserved for invalid/unassigned ppgtt */
+	args->vm_id = id;
+	return 0;
+
+err_put:
+	i915_vm_put(&ppgtt->vm);
+	return err;
+}
+
+int i915_gem_vm_destroy_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *file)
+{
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct drm_i915_gem_vm_control *args = data;
+	struct i915_address_space *vm;
+
+	if (args->flags)
+		return -EINVAL;
+
+	if (args->extensions)
+		return -EINVAL;
+
+	vm = xa_erase(&file_priv->vm_xa, args->vm_id);
+	if (!vm)
+		return -ENOENT;
+
+	i915_vm_put(vm);
+	return 0;
+}
+
+static int get_ppgtt(struct drm_i915_file_private *file_priv,
+		     struct i915_gem_context *ctx,
+		     struct drm_i915_gem_context_param *args)
+{
+	struct i915_address_space *vm;
+	int err;
+	u32 id;
+
+	if (!i915_gem_context_has_full_ppgtt(ctx))
+		return -ENODEV;
+
+	vm = ctx->vm;
+	GEM_BUG_ON(!vm);
+
+	/*
+	 * Get a reference for the allocated handle.  Once the handle is
+	 * visible in the vm_xa table, userspace could try to close it
+	 * from under our feet, so we need to hold the extra reference
+	 * first.
+	 */
+	i915_vm_get(vm);
+
+	err = xa_alloc(&file_priv->vm_xa, &id, vm, xa_limit_32b, GFP_KERNEL);
+	if (err) {
+		i915_vm_put(vm);
+		return err;
+	}
+
+	GEM_BUG_ON(id == 0); /* reserved for invalid/unassigned ppgtt */
+	args->value = id;
+	args->size = 0;
+
+	return err;
+}
+
+int
+i915_gem_user_to_context_sseu(struct intel_gt *gt,
+			      const struct drm_i915_gem_context_param_sseu *user,
+			      struct intel_sseu *context)
+{
+	const struct sseu_dev_info *device = &gt->info.sseu;
+	struct drm_i915_private *i915 = gt->i915;
+	unsigned int dev_subslice_mask = intel_sseu_get_hsw_subslices(device, 0);
+
+	/* No zeros in any field. */
+	if (!user->slice_mask || !user->subslice_mask ||
+	    !user->min_eus_per_subslice || !user->max_eus_per_subslice)
+		return -EINVAL;
+
+	/* Max > min. */
+	if (user->max_eus_per_subslice < user->min_eus_per_subslice)
+		return -EINVAL;
+
+	/*
+	 * Some future proofing on the types since the uAPI is wider than the
+	 * current internal implementation.
+	 */
+	if (overflows_type(user->slice_mask, context->slice_mask) ||
+	    overflows_type(user->subslice_mask, context->subslice_mask) ||
+	    overflows_type(user->min_eus_per_subslice,
+			   context->min_eus_per_subslice) ||
+	    overflows_type(user->max_eus_per_subslice,
+			   context->max_eus_per_subslice))
+		return -EINVAL;
+
+	/* Check validity against hardware. */
+	if (user->slice_mask & ~device->slice_mask)
+		return -EINVAL;
+
+	if (user->subslice_mask & ~dev_subslice_mask)
+		return -EINVAL;
+
+	if (user->max_eus_per_subslice > device->max_eus_per_subslice)
+		return -EINVAL;
+
+	context->slice_mask = user->slice_mask;
+	context->subslice_mask = user->subslice_mask;
+	context->min_eus_per_subslice = user->min_eus_per_subslice;
+	context->max_eus_per_subslice = user->max_eus_per_subslice;
+
+	/* Part specific restrictions. */
+	if (GRAPHICS_VER(i915) == 11) {
+		unsigned int hw_s = hweight8(device->slice_mask);
+		unsigned int hw_ss_per_s = hweight8(dev_subslice_mask);
+		unsigned int req_s = hweight8(context->slice_mask);
+		unsigned int req_ss = hweight8(context->subslice_mask);
+
+		/*
+		 * Only full subslice enablement is possible if more than one
+		 * slice is turned on.
+		 */
+		if (req_s > 1 && req_ss != hw_ss_per_s)
+			return -EINVAL;
+
+		/*
+		 * If more than four (SScount bitfield limit) subslices are
+		 * requested then the number has to be even.
+		 */
+		if (req_ss > 4 && (req_ss & 1))
+			return -EINVAL;
+
+		/*
+		 * If only one slice is enabled and subslice count is below the
+		 * device full enablement, it must be at most half of the all
+		 * available subslices.
+		 */
+		if (req_s == 1 && req_ss < hw_ss_per_s &&
+		    req_ss > (hw_ss_per_s / 2))
+			return -EINVAL;
+
+		/* ABI restriction - VME use case only. */
+
+		/* All slices or one slice only. */
+		if (req_s != 1 && req_s != hw_s)
+			return -EINVAL;
+
+		/*
+		 * Half subslices or full enablement only when one slice is
+		 * enabled.
+		 */
+		if (req_s == 1 &&
+		    (req_ss != hw_ss_per_s && req_ss != (hw_ss_per_s / 2)))
+			return -EINVAL;
+
+		/* No EU configuration changes. */
+		if ((user->min_eus_per_subslice !=
+		     device->max_eus_per_subslice) ||
+		    (user->max_eus_per_subslice !=
+		     device->max_eus_per_subslice))
+			return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int set_sseu(struct i915_gem_context *ctx,
+		    struct drm_i915_gem_context_param *args)
+{
+	struct drm_i915_private *i915 = ctx->i915;
+	struct drm_i915_gem_context_param_sseu user_sseu;
+	struct intel_context *ce;
+	struct intel_sseu sseu;
+	unsigned long lookup;
+	int ret;
+
+	if (args->size < sizeof(user_sseu))
+		return -EINVAL;
+
+	if (GRAPHICS_VER(i915) != 11)
+		return -ENODEV;
+
+	if (copy_from_user(&user_sseu, u64_to_user_ptr(args->value),
+			   sizeof(user_sseu)))
+		return -EFAULT;
+
+	if (user_sseu.rsvd)
+		return -EINVAL;
+
+	if (user_sseu.flags & ~(I915_CONTEXT_SSEU_FLAG_ENGINE_INDEX))
+		return -EINVAL;
+
+	lookup = 0;
+	if (user_sseu.flags & I915_CONTEXT_SSEU_FLAG_ENGINE_INDEX)
+		lookup |= LOOKUP_USER_INDEX;
+
+	ce = lookup_user_engine(ctx, lookup, &user_sseu.engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	/* Only render engine supports RPCS configuration. */
+	if (ce->engine->class != RENDER_CLASS) {
+		ret = -ENODEV;
+		goto out_ce;
+	}
+
+	ret = i915_gem_user_to_context_sseu(ce->engine->gt, &user_sseu, &sseu);
+	if (ret)
+		goto out_ce;
+
+	ret = intel_context_reconfigure_sseu(ce, sseu);
+	if (ret)
+		goto out_ce;
+
+	args->size = sizeof(user_sseu);
+
+out_ce:
+	intel_context_put(ce);
+	return ret;
+}
+
+static int
+set_persistence(struct i915_gem_context *ctx,
+		const struct drm_i915_gem_context_param *args)
+{
+	if (args->size)
+		return -EINVAL;
+
+	return __context_set_persistence(ctx, args->value);
+}
+
+static int set_priority(struct i915_gem_context *ctx,
+			const struct drm_i915_gem_context_param *args)
+{
+	struct i915_gem_engines_iter it;
+	struct intel_context *ce;
+	int err;
+
+	err = validate_priority(ctx->i915, args);
+	if (err)
+		return err;
+
+	ctx->sched.priority = args->value;
+
+	for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it) {
+		if (!intel_engine_has_timeslices(ce->engine))
+			continue;
+
+		if (ctx->sched.priority >= I915_PRIORITY_NORMAL &&
+		    intel_engine_has_semaphores(ce->engine))
+			intel_context_set_use_semaphores(ce);
+		else
+			intel_context_clear_use_semaphores(ce);
+	}
+	i915_gem_context_unlock_engines(ctx);
+
+	return 0;
+}
+
+static int get_protected(struct i915_gem_context *ctx,
+			 struct drm_i915_gem_context_param *args)
+{
+	args->size = 0;
+	args->value = i915_gem_context_uses_protected_content(ctx);
+
+	return 0;
+}
+
+static int ctx_setparam(struct drm_i915_file_private *fpriv,
+			struct i915_gem_context *ctx,
+			struct drm_i915_gem_context_param *args)
+{
+	int ret = 0;
+
+	switch (args->param) {
+	case I915_CONTEXT_PARAM_NO_ERROR_CAPTURE:
+		if (args->size)
+			ret = -EINVAL;
+		else if (args->value)
+			i915_gem_context_set_no_error_capture(ctx);
+		else
+			i915_gem_context_clear_no_error_capture(ctx);
+		break;
+
+	case I915_CONTEXT_PARAM_BANNABLE:
+		if (args->size)
+			ret = -EINVAL;
+		else if (!capable(CAP_SYS_ADMIN) && !args->value)
+			ret = -EPERM;
+		else if (args->value)
+			i915_gem_context_set_bannable(ctx);
+		else if (i915_gem_context_uses_protected_content(ctx))
+			ret = -EPERM; /* can't clear this for protected contexts */
+		else
+			i915_gem_context_clear_bannable(ctx);
+		break;
+
+	case I915_CONTEXT_PARAM_RECOVERABLE:
+		if (args->size)
+			ret = -EINVAL;
+		else if (!args->value)
+			i915_gem_context_clear_recoverable(ctx);
+		else if (i915_gem_context_uses_protected_content(ctx))
+			ret = -EPERM; /* can't set this for protected contexts */
+		else
+			i915_gem_context_set_recoverable(ctx);
+		break;
+
+	case I915_CONTEXT_PARAM_PRIORITY:
+		ret = set_priority(ctx, args);
+		break;
+
+	case I915_CONTEXT_PARAM_SSEU:
+		ret = set_sseu(ctx, args);
+		break;
+
+	case I915_CONTEXT_PARAM_PERSISTENCE:
+		ret = set_persistence(ctx, args);
+		break;
+
+	case I915_CONTEXT_PARAM_PROTECTED_CONTENT:
+	case I915_CONTEXT_PARAM_NO_ZEROMAP:
+	case I915_CONTEXT_PARAM_BAN_PERIOD:
+	case I915_CONTEXT_PARAM_RINGSIZE:
+	case I915_CONTEXT_PARAM_VM:
+	case I915_CONTEXT_PARAM_ENGINES:
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+struct create_ext {
+	struct i915_gem_proto_context *pc;
+	struct drm_i915_file_private *fpriv;
+};
+
+static int create_setparam(struct i915_user_extension __user *ext, void *data)
+{
+	struct drm_i915_gem_context_create_ext_setparam local;
+	const struct create_ext *arg = data;
+
+	if (copy_from_user(&local, ext, sizeof(local)))
+		return -EFAULT;
+
+	if (local.param.ctx_id)
+		return -EINVAL;
+
+	return set_proto_ctx_param(arg->fpriv, arg->pc, &local.param);
+}
+
+static int invalid_ext(struct i915_user_extension __user *ext, void *data)
+{
+	return -EINVAL;
+}
+
+static const i915_user_extension_fn create_extensions[] = {
+	[I915_CONTEXT_CREATE_EXT_SETPARAM] = create_setparam,
+	[I915_CONTEXT_CREATE_EXT_CLONE] = invalid_ext,
+};
+
+static bool client_is_banned(struct drm_i915_file_private *file_priv)
+{
+	return atomic_read(&file_priv->ban_score) >= I915_CLIENT_SCORE_BANNED;
+}
+
+static inline struct i915_gem_context *
+__context_lookup(struct drm_i915_file_private *file_priv, u32 id)
+{
+	struct i915_gem_context *ctx;
+
+	rcu_read_lock();
+	ctx = xa_load(&file_priv->context_xa, id);
+	if (ctx && !kref_get_unless_zero(&ctx->ref))
+		ctx = NULL;
+	rcu_read_unlock();
+
+	return ctx;
+}
+
+static struct i915_gem_context *
+finalize_create_context_locked(struct drm_i915_file_private *file_priv,
+			       struct i915_gem_proto_context *pc, u32 id)
+{
+	struct i915_gem_context *ctx;
+	void *old;
+
+	lockdep_assert_held(&file_priv->proto_context_lock);
+
+	ctx = i915_gem_create_context(file_priv->dev_priv, pc);
+	if (IS_ERR(ctx))
+		return ctx;
+
+	/*
+	 * One for the xarray and one for the caller.  We need to grab
+	 * the reference *prior* to making the ctx visble to userspace
+	 * in gem_context_register(), as at any point after that
+	 * userspace can try to race us with another thread destroying
+	 * the context under our feet.
+	 */
+	i915_gem_context_get(ctx);
+
+	gem_context_register(ctx, file_priv, id);
+
+	old = xa_erase(&file_priv->proto_context_xa, id);
+	GEM_BUG_ON(old != pc);
+	proto_context_close(file_priv->dev_priv, pc);
+
+	return ctx;
+}
+
+struct i915_gem_context *
+i915_gem_context_lookup(struct drm_i915_file_private *file_priv, u32 id)
+{
+	struct i915_gem_proto_context *pc;
+	struct i915_gem_context *ctx;
+
+	ctx = __context_lookup(file_priv, id);
+	if (ctx)
+		return ctx;
+
+	mutex_lock(&file_priv->proto_context_lock);
+	/* Try one more time under the lock */
+	ctx = __context_lookup(file_priv, id);
+	if (!ctx) {
+		pc = xa_load(&file_priv->proto_context_xa, id);
+		if (!pc)
+			ctx = ERR_PTR(-ENOENT);
+		else
+			ctx = finalize_create_context_locked(file_priv, pc, id);
+	}
+	mutex_unlock(&file_priv->proto_context_lock);
+
+	return ctx;
+}
+
+int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
+				  struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_context_create_ext *args = data;
+	struct create_ext ext_data;
+	int ret;
+	u32 id;
+
+	if (!DRIVER_CAPS(i915)->has_logical_contexts)
+		return -ENODEV;
+
+	if (args->flags & I915_CONTEXT_CREATE_FLAGS_UNKNOWN)
+		return -EINVAL;
+
+	ret = intel_gt_terminally_wedged(to_gt(i915));
+	if (ret)
+		return ret;
+
+	ext_data.fpriv = file->driver_priv;
+	if (client_is_banned(ext_data.fpriv)) {
+		drm_dbg(&i915->drm,
+			"client %s[%d] banned from creating ctx\n",
+			current->comm, task_pid_nr(current));
+		return -EIO;
+	}
+
+	ext_data.pc = proto_context_create(i915, args->flags);
+	if (IS_ERR(ext_data.pc))
+		return PTR_ERR(ext_data.pc);
+
+	if (args->flags & I915_CONTEXT_CREATE_FLAGS_USE_EXTENSIONS) {
+		ret = i915_user_extensions(u64_to_user_ptr(args->extensions),
+					   create_extensions,
+					   ARRAY_SIZE(create_extensions),
+					   &ext_data);
+		if (ret)
+			goto err_pc;
+	}
+
+	if (GRAPHICS_VER(i915) > 12) {
+		struct i915_gem_context *ctx;
+
+		/* Get ourselves a context ID */
+		ret = xa_alloc(&ext_data.fpriv->context_xa, &id, NULL,
+			       xa_limit_32b, GFP_KERNEL);
+		if (ret)
+			goto err_pc;
+
+		ctx = i915_gem_create_context(i915, ext_data.pc);
+		if (IS_ERR(ctx)) {
+			ret = PTR_ERR(ctx);
+			goto err_pc;
+		}
+
+		proto_context_close(i915, ext_data.pc);
+		gem_context_register(ctx, ext_data.fpriv, id);
+	} else {
+		ret = proto_context_register(ext_data.fpriv, ext_data.pc, &id);
+		if (ret < 0)
+			goto err_pc;
+	}
+
+	args->ctx_id = id;
+	drm_dbg(&i915->drm, "HW context %d created\n", args->ctx_id);
+
+	return 0;
+
+err_pc:
+	proto_context_close(i915, ext_data.pc);
+	return ret;
+}
+
+int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
+				   struct drm_file *file)
+{
+	struct drm_i915_gem_context_destroy *args = data;
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct i915_gem_proto_context *pc;
+	struct i915_gem_context *ctx;
+
+	if (args->pad != 0)
+		return -EINVAL;
+
+	if (!args->ctx_id)
+		return -ENOENT;
+
+	/* We need to hold the proto-context lock here to prevent races
+	 * with finalize_create_context_locked().
+	 */
+	mutex_lock(&file_priv->proto_context_lock);
+	ctx = xa_erase(&file_priv->context_xa, args->ctx_id);
+	pc = xa_erase(&file_priv->proto_context_xa, args->ctx_id);
+	mutex_unlock(&file_priv->proto_context_lock);
+
+	if (!ctx && !pc)
+		return -ENOENT;
+	GEM_WARN_ON(ctx && pc);
+
+	if (pc)
+		proto_context_close(file_priv->dev_priv, pc);
+
+	if (ctx)
+		context_close(ctx);
+
+	return 0;
+}
+
+static int get_sseu(struct i915_gem_context *ctx,
+		    struct drm_i915_gem_context_param *args)
+{
+	struct drm_i915_gem_context_param_sseu user_sseu;
+	struct intel_context *ce;
+	unsigned long lookup;
+	int err;
+
+	if (args->size == 0)
+		goto out;
+	else if (args->size < sizeof(user_sseu))
+		return -EINVAL;
+
+	if (copy_from_user(&user_sseu, u64_to_user_ptr(args->value),
+			   sizeof(user_sseu)))
+		return -EFAULT;
+
+	if (user_sseu.rsvd)
+		return -EINVAL;
+
+	if (user_sseu.flags & ~(I915_CONTEXT_SSEU_FLAG_ENGINE_INDEX))
+		return -EINVAL;
+
+	lookup = 0;
+	if (user_sseu.flags & I915_CONTEXT_SSEU_FLAG_ENGINE_INDEX)
+		lookup |= LOOKUP_USER_INDEX;
+
+	ce = lookup_user_engine(ctx, lookup, &user_sseu.engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	err = intel_context_lock_pinned(ce); /* serialises with set_sseu */
+	if (err) {
+		intel_context_put(ce);
+		return err;
+	}
+
+	user_sseu.slice_mask = ce->sseu.slice_mask;
+	user_sseu.subslice_mask = ce->sseu.subslice_mask;
+	user_sseu.min_eus_per_subslice = ce->sseu.min_eus_per_subslice;
+	user_sseu.max_eus_per_subslice = ce->sseu.max_eus_per_subslice;
+
+	intel_context_unlock_pinned(ce);
+	intel_context_put(ce);
+
+	if (copy_to_user(u64_to_user_ptr(args->value), &user_sseu,
+			 sizeof(user_sseu)))
+		return -EFAULT;
+
+out:
+	args->size = sizeof(user_sseu);
+
+	return 0;
+}
+
+int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data,
+				    struct drm_file *file)
+{
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct drm_i915_gem_context_param *args = data;
+	struct i915_gem_context *ctx;
+	struct i915_address_space *vm;
+	int ret = 0;
+
+	ctx = i915_gem_context_lookup(file_priv, args->ctx_id);
+	if (IS_ERR(ctx))
+		return PTR_ERR(ctx);
+
+	switch (args->param) {
+	case I915_CONTEXT_PARAM_GTT_SIZE:
+		args->size = 0;
+		vm = i915_gem_context_get_eb_vm(ctx);
+		args->value = vm->total;
+		i915_vm_put(vm);
+
+		break;
+
+	case I915_CONTEXT_PARAM_NO_ERROR_CAPTURE:
+		args->size = 0;
+		args->value = i915_gem_context_no_error_capture(ctx);
+		break;
+
+	case I915_CONTEXT_PARAM_BANNABLE:
+		args->size = 0;
+		args->value = i915_gem_context_is_bannable(ctx);
+		break;
+
+	case I915_CONTEXT_PARAM_RECOVERABLE:
+		args->size = 0;
+		args->value = i915_gem_context_is_recoverable(ctx);
+		break;
+
+	case I915_CONTEXT_PARAM_PRIORITY:
+		args->size = 0;
+		args->value = ctx->sched.priority;
+		break;
+
+	case I915_CONTEXT_PARAM_SSEU:
+		ret = get_sseu(ctx, args);
+		break;
+
+	case I915_CONTEXT_PARAM_VM:
+		ret = get_ppgtt(file_priv, ctx, args);
+		break;
+
+	case I915_CONTEXT_PARAM_PERSISTENCE:
+		args->size = 0;
+		args->value = i915_gem_context_is_persistent(ctx);
+		break;
+
+	case I915_CONTEXT_PARAM_PROTECTED_CONTENT:
+		ret = get_protected(ctx, args);
+		break;
+
+	case I915_CONTEXT_PARAM_NO_ZEROMAP:
+	case I915_CONTEXT_PARAM_BAN_PERIOD:
+	case I915_CONTEXT_PARAM_ENGINES:
+	case I915_CONTEXT_PARAM_RINGSIZE:
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	i915_gem_context_put(ctx);
+	return ret;
+}
+
+int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data,
+				    struct drm_file *file)
+{
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct drm_i915_gem_context_param *args = data;
+	struct i915_gem_proto_context *pc;
+	struct i915_gem_context *ctx;
+	int ret = 0;
+
+	mutex_lock(&file_priv->proto_context_lock);
+	ctx = __context_lookup(file_priv, args->ctx_id);
+	if (!ctx) {
+		pc = xa_load(&file_priv->proto_context_xa, args->ctx_id);
+		if (pc) {
+			/* Contexts should be finalized inside
+			 * GEM_CONTEXT_CREATE starting with graphics
+			 * version 13.
+			 */
+			WARN_ON(GRAPHICS_VER(file_priv->dev_priv) > 12);
+			ret = set_proto_ctx_param(file_priv, pc, args);
+		} else {
+			ret = -ENOENT;
+		}
+	}
+	mutex_unlock(&file_priv->proto_context_lock);
+
+	if (ctx) {
+		ret = ctx_setparam(file_priv, ctx, args);
+		i915_gem_context_put(ctx);
+	}
+
+	return ret;
+}
+
+int i915_gem_context_reset_stats_ioctl(struct drm_device *dev,
+				       void *data, struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_reset_stats *args = data;
+	struct i915_gem_context *ctx;
+
+	if (args->flags || args->pad)
+		return -EINVAL;
+
+	ctx = i915_gem_context_lookup(file->driver_priv, args->ctx_id);
+	if (IS_ERR(ctx))
+		return PTR_ERR(ctx);
+
+	/*
+	 * We opt for unserialised reads here. This may result in tearing
+	 * in the extremely unlikely event of a GPU hang on this context
+	 * as we are querying them. If we need that extra layer of protection,
+	 * we should wrap the hangstats with a seqlock.
+	 */
+
+	if (capable(CAP_SYS_ADMIN))
+		args->reset_count = i915_reset_count(&i915->gpu_error);
+	else
+		args->reset_count = 0;
+
+	args->batch_active = atomic_read(&ctx->guilty_count);
+	args->batch_pending = atomic_read(&ctx->active_count);
+
+	i915_gem_context_put(ctx);
+	return 0;
+}
+
+/* GEM context-engines iterator: for_each_gem_engine() */
+struct intel_context *
+i915_gem_engines_iter_next(struct i915_gem_engines_iter *it)
+{
+	const struct i915_gem_engines *e = it->engines;
+	struct intel_context *ctx;
+
+	if (unlikely(!e))
+		return NULL;
+
+	do {
+		if (it->idx >= e->num_engines)
+			return NULL;
+
+		ctx = e->engines[it->idx++];
+	} while (!ctx);
+
+	return ctx;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_context.c"
+#include "selftests/i915_gem_context.c"
+#endif
+
+void i915_gem_context_module_exit(void)
+{
+	kmem_cache_destroy(slab_luts);
+}
+
+int __init i915_gem_context_module_init(void)
+{
+	slab_luts = KMEM_CACHE(i915_lut_handle, 0);
+	if (!slab_luts)
+		return -ENOMEM;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.h b/drivers/gpu/drm/i915/gem/i915_gem_context.h
new file mode 100644
index 000000000..e5b0f66ea
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.h
@@ -0,0 +1,248 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __I915_GEM_CONTEXT_H__
+#define __I915_GEM_CONTEXT_H__
+
+#include "i915_gem_context_types.h"
+
+#include "gt/intel_context.h"
+
+#include "i915_drv.h"
+#include "i915_gem.h"
+#include "i915_scheduler.h"
+#include "intel_device_info.h"
+
+struct drm_device;
+struct drm_file;
+
+static inline bool i915_gem_context_is_closed(const struct i915_gem_context *ctx)
+{
+	return test_bit(CONTEXT_CLOSED, &ctx->flags);
+}
+
+static inline void i915_gem_context_set_closed(struct i915_gem_context *ctx)
+{
+	GEM_BUG_ON(i915_gem_context_is_closed(ctx));
+	set_bit(CONTEXT_CLOSED, &ctx->flags);
+}
+
+static inline bool i915_gem_context_no_error_capture(const struct i915_gem_context *ctx)
+{
+	return test_bit(UCONTEXT_NO_ERROR_CAPTURE, &ctx->user_flags);
+}
+
+static inline void i915_gem_context_set_no_error_capture(struct i915_gem_context *ctx)
+{
+	set_bit(UCONTEXT_NO_ERROR_CAPTURE, &ctx->user_flags);
+}
+
+static inline void i915_gem_context_clear_no_error_capture(struct i915_gem_context *ctx)
+{
+	clear_bit(UCONTEXT_NO_ERROR_CAPTURE, &ctx->user_flags);
+}
+
+static inline bool i915_gem_context_is_bannable(const struct i915_gem_context *ctx)
+{
+	return test_bit(UCONTEXT_BANNABLE, &ctx->user_flags);
+}
+
+static inline void i915_gem_context_set_bannable(struct i915_gem_context *ctx)
+{
+	set_bit(UCONTEXT_BANNABLE, &ctx->user_flags);
+}
+
+static inline void i915_gem_context_clear_bannable(struct i915_gem_context *ctx)
+{
+	clear_bit(UCONTEXT_BANNABLE, &ctx->user_flags);
+}
+
+static inline bool i915_gem_context_is_recoverable(const struct i915_gem_context *ctx)
+{
+	return test_bit(UCONTEXT_RECOVERABLE, &ctx->user_flags);
+}
+
+static inline void i915_gem_context_set_recoverable(struct i915_gem_context *ctx)
+{
+	set_bit(UCONTEXT_RECOVERABLE, &ctx->user_flags);
+}
+
+static inline void i915_gem_context_clear_recoverable(struct i915_gem_context *ctx)
+{
+	clear_bit(UCONTEXT_RECOVERABLE, &ctx->user_flags);
+}
+
+static inline bool i915_gem_context_is_persistent(const struct i915_gem_context *ctx)
+{
+	return test_bit(UCONTEXT_PERSISTENCE, &ctx->user_flags);
+}
+
+static inline void i915_gem_context_set_persistence(struct i915_gem_context *ctx)
+{
+	set_bit(UCONTEXT_PERSISTENCE, &ctx->user_flags);
+}
+
+static inline void i915_gem_context_clear_persistence(struct i915_gem_context *ctx)
+{
+	clear_bit(UCONTEXT_PERSISTENCE, &ctx->user_flags);
+}
+
+static inline bool
+i915_gem_context_user_engines(const struct i915_gem_context *ctx)
+{
+	return test_bit(CONTEXT_USER_ENGINES, &ctx->flags);
+}
+
+static inline void
+i915_gem_context_set_user_engines(struct i915_gem_context *ctx)
+{
+	set_bit(CONTEXT_USER_ENGINES, &ctx->flags);
+}
+
+static inline void
+i915_gem_context_clear_user_engines(struct i915_gem_context *ctx)
+{
+	clear_bit(CONTEXT_USER_ENGINES, &ctx->flags);
+}
+
+static inline bool
+i915_gem_context_uses_protected_content(const struct i915_gem_context *ctx)
+{
+	return ctx->uses_protected_content;
+}
+
+/* i915_gem_context.c */
+void i915_gem_init__contexts(struct drm_i915_private *i915);
+
+int i915_gem_context_open(struct drm_i915_private *i915,
+			  struct drm_file *file);
+void i915_gem_context_close(struct drm_file *file);
+
+void i915_gem_context_release(struct kref *ctx_ref);
+
+int i915_gem_vm_create_ioctl(struct drm_device *dev, void *data,
+			     struct drm_file *file);
+int i915_gem_vm_destroy_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *file);
+
+int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
+				  struct drm_file *file);
+int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
+				   struct drm_file *file);
+int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data,
+				    struct drm_file *file_priv);
+int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data,
+				    struct drm_file *file_priv);
+int i915_gem_context_reset_stats_ioctl(struct drm_device *dev, void *data,
+				       struct drm_file *file);
+
+struct i915_gem_context *
+i915_gem_context_lookup(struct drm_i915_file_private *file_priv, u32 id);
+
+static inline struct i915_gem_context *
+i915_gem_context_get(struct i915_gem_context *ctx)
+{
+	kref_get(&ctx->ref);
+	return ctx;
+}
+
+static inline void i915_gem_context_put(struct i915_gem_context *ctx)
+{
+	kref_put(&ctx->ref, i915_gem_context_release);
+}
+
+static inline struct i915_address_space *
+i915_gem_context_vm(struct i915_gem_context *ctx)
+{
+	return rcu_dereference_protected(ctx->vm, lockdep_is_held(&ctx->mutex));
+}
+
+static inline bool i915_gem_context_has_full_ppgtt(struct i915_gem_context *ctx)
+{
+	GEM_BUG_ON(!!ctx->vm != HAS_FULL_PPGTT(ctx->i915));
+
+	return !!ctx->vm;
+}
+
+static inline struct i915_address_space *
+i915_gem_context_get_eb_vm(struct i915_gem_context *ctx)
+{
+	struct i915_address_space *vm;
+
+	vm = ctx->vm;
+	if (!vm)
+		vm = &to_gt(ctx->i915)->ggtt->vm;
+	vm = i915_vm_get(vm);
+
+	return vm;
+}
+
+static inline struct i915_gem_engines *
+i915_gem_context_engines(struct i915_gem_context *ctx)
+{
+	return rcu_dereference_protected(ctx->engines,
+					 lockdep_is_held(&ctx->engines_mutex));
+}
+
+static inline struct i915_gem_engines *
+i915_gem_context_lock_engines(struct i915_gem_context *ctx)
+	__acquires(&ctx->engines_mutex)
+{
+	mutex_lock(&ctx->engines_mutex);
+	return i915_gem_context_engines(ctx);
+}
+
+static inline void
+i915_gem_context_unlock_engines(struct i915_gem_context *ctx)
+	__releases(&ctx->engines_mutex)
+{
+	mutex_unlock(&ctx->engines_mutex);
+}
+
+static inline struct intel_context *
+i915_gem_context_get_engine(struct i915_gem_context *ctx, unsigned int idx)
+{
+	struct intel_context *ce;
+
+	rcu_read_lock(); {
+		struct i915_gem_engines *e = rcu_dereference(ctx->engines);
+		if (unlikely(!e)) /* context was closed! */
+			ce = ERR_PTR(-ENOENT);
+		else if (likely(idx < e->num_engines && e->engines[idx]))
+			ce = intel_context_get(e->engines[idx]);
+		else
+			ce = ERR_PTR(-EINVAL);
+	} rcu_read_unlock();
+
+	return ce;
+}
+
+static inline void
+i915_gem_engines_iter_init(struct i915_gem_engines_iter *it,
+			   struct i915_gem_engines *engines)
+{
+	it->engines = engines;
+	it->idx = 0;
+}
+
+struct intel_context *
+i915_gem_engines_iter_next(struct i915_gem_engines_iter *it);
+
+#define for_each_gem_engine(ce, engines, it) \
+	for (i915_gem_engines_iter_init(&(it), (engines)); \
+	     ((ce) = i915_gem_engines_iter_next(&(it)));)
+
+void i915_gem_context_module_exit(void);
+int i915_gem_context_module_init(void);
+
+struct i915_lut_handle *i915_lut_handle_alloc(void);
+void i915_lut_handle_free(struct i915_lut_handle *lut);
+
+int i915_gem_user_to_context_sseu(struct intel_gt *gt,
+				  const struct drm_i915_gem_context_param_sseu *user,
+				  struct intel_sseu *context);
+
+#endif /* !__I915_GEM_CONTEXT_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context_types.h b/drivers/gpu/drm/i915/gem/i915_gem_context_types.h
new file mode 100644
index 000000000..cb78214a7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context_types.h
@@ -0,0 +1,419 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_GEM_CONTEXT_TYPES_H__
+#define __I915_GEM_CONTEXT_TYPES_H__
+
+#include <linux/atomic.h>
+#include <linux/list.h>
+#include <linux/llist.h>
+#include <linux/kref.h>
+#include <linux/mutex.h>
+#include <linux/radix-tree.h>
+#include <linux/rbtree.h>
+#include <linux/rcupdate.h>
+#include <linux/types.h>
+
+#include "gt/intel_context_types.h"
+
+#include "i915_scheduler.h"
+#include "i915_sw_fence.h"
+
+struct pid;
+
+struct drm_i915_private;
+struct drm_i915_file_private;
+struct i915_address_space;
+struct intel_timeline;
+struct intel_ring;
+
+/**
+ * struct i915_gem_engines - A set of engines
+ */
+struct i915_gem_engines {
+	union {
+		/** @link: Link in i915_gem_context::stale::engines */
+		struct list_head link;
+
+		/** @rcu: RCU to use when freeing */
+		struct rcu_head rcu;
+	};
+
+	/** @fence: Fence used for delayed destruction of engines */
+	struct i915_sw_fence fence;
+
+	/** @ctx: i915_gem_context backpointer */
+	struct i915_gem_context *ctx;
+
+	/** @num_engines: Number of engines in this set */
+	unsigned int num_engines;
+
+	/** @engines: Array of engines */
+	struct intel_context *engines[];
+};
+
+/**
+ * struct i915_gem_engines_iter - Iterator for an i915_gem_engines set
+ */
+struct i915_gem_engines_iter {
+	/** @idx: Index into i915_gem_engines::engines */
+	unsigned int idx;
+
+	/** @engines: Engine set being iterated */
+	const struct i915_gem_engines *engines;
+};
+
+/**
+ * enum i915_gem_engine_type - Describes the type of an i915_gem_proto_engine
+ */
+enum i915_gem_engine_type {
+	/** @I915_GEM_ENGINE_TYPE_INVALID: An invalid engine */
+	I915_GEM_ENGINE_TYPE_INVALID = 0,
+
+	/** @I915_GEM_ENGINE_TYPE_PHYSICAL: A single physical engine */
+	I915_GEM_ENGINE_TYPE_PHYSICAL,
+
+	/** @I915_GEM_ENGINE_TYPE_BALANCED: A load-balanced engine set */
+	I915_GEM_ENGINE_TYPE_BALANCED,
+
+	/** @I915_GEM_ENGINE_TYPE_PARALLEL: A parallel engine set */
+	I915_GEM_ENGINE_TYPE_PARALLEL,
+};
+
+/**
+ * struct i915_gem_proto_engine - prototype engine
+ *
+ * This struct describes an engine that a context may contain.  Engines
+ * have four types:
+ *
+ *  - I915_GEM_ENGINE_TYPE_INVALID: Invalid engines can be created but they
+ *    show up as a NULL in i915_gem_engines::engines[i] and any attempt to
+ *    use them by the user results in -EINVAL.  They are also useful during
+ *    proto-context construction because the client may create invalid
+ *    engines and then set them up later as virtual engines.
+ *
+ *  - I915_GEM_ENGINE_TYPE_PHYSICAL: A single physical engine, described by
+ *    i915_gem_proto_engine::engine.
+ *
+ *  - I915_GEM_ENGINE_TYPE_BALANCED: A load-balanced engine set, described
+ *    i915_gem_proto_engine::num_siblings and i915_gem_proto_engine::siblings.
+ *
+ *  - I915_GEM_ENGINE_TYPE_PARALLEL: A parallel submission engine set, described
+ *    i915_gem_proto_engine::width, i915_gem_proto_engine::num_siblings, and
+ *    i915_gem_proto_engine::siblings.
+ */
+struct i915_gem_proto_engine {
+	/** @type: Type of this engine */
+	enum i915_gem_engine_type type;
+
+	/** @engine: Engine, for physical */
+	struct intel_engine_cs *engine;
+
+	/** @num_siblings: Number of balanced or parallel siblings */
+	unsigned int num_siblings;
+
+	/** @width: Width of each sibling */
+	unsigned int width;
+
+	/** @siblings: Balanced siblings or num_siblings * width for parallel */
+	struct intel_engine_cs **siblings;
+
+	/** @sseu: Client-set SSEU parameters */
+	struct intel_sseu sseu;
+};
+
+/**
+ * struct i915_gem_proto_context - prototype context
+ *
+ * The struct i915_gem_proto_context represents the creation parameters for
+ * a struct i915_gem_context.  This is used to gather parameters provided
+ * either through creation flags or via SET_CONTEXT_PARAM so that, when we
+ * create the final i915_gem_context, those parameters can be immutable.
+ *
+ * The context uAPI allows for two methods of setting context parameters:
+ * SET_CONTEXT_PARAM and CONTEXT_CREATE_EXT_SETPARAM.  The former is
+ * allowed to be called at any time while the later happens as part of
+ * GEM_CONTEXT_CREATE.  When these were initially added, Currently,
+ * everything settable via one is settable via the other.  While some
+ * params are fairly simple and setting them on a live context is harmless
+ * such the context priority, others are far trickier such as the VM or the
+ * set of engines.  To avoid some truly nasty race conditions, we don't
+ * allow setting the VM or the set of engines on live contexts.
+ *
+ * The way we dealt with this without breaking older userspace that sets
+ * the VM or engine set via SET_CONTEXT_PARAM is to delay the creation of
+ * the actual context until after the client is done configuring it with
+ * SET_CONTEXT_PARAM.  From the perspective of the client, it has the same
+ * u32 context ID the whole time.  From the perspective of i915, however,
+ * it's an i915_gem_proto_context right up until the point where we attempt
+ * to do something which the proto-context can't handle at which point the
+ * real context gets created.
+ *
+ * This is accomplished via a little xarray dance.  When GEM_CONTEXT_CREATE
+ * is called, we create a proto-context, reserve a slot in context_xa but
+ * leave it NULL, the proto-context in the corresponding slot in
+ * proto_context_xa.  Then, whenever we go to look up a context, we first
+ * check context_xa.  If it's there, we return the i915_gem_context and
+ * we're done.  If it's not, we look in proto_context_xa and, if we find it
+ * there, we create the actual context and kill the proto-context.
+ *
+ * At the time we made this change (April, 2021), we did a fairly complete
+ * audit of existing userspace to ensure this wouldn't break anything:
+ *
+ *  - Mesa/i965 didn't use the engines or VM APIs at all
+ *
+ *  - Mesa/ANV used the engines API but via CONTEXT_CREATE_EXT_SETPARAM and
+ *    didn't use the VM API.
+ *
+ *  - Mesa/iris didn't use the engines or VM APIs at all
+ *
+ *  - The open-source compute-runtime didn't yet use the engines API but
+ *    did use the VM API via SET_CONTEXT_PARAM.  However, CONTEXT_SETPARAM
+ *    was always the second ioctl on that context, immediately following
+ *    GEM_CONTEXT_CREATE.
+ *
+ *  - The media driver sets engines and bonding/balancing via
+ *    SET_CONTEXT_PARAM.  However, CONTEXT_SETPARAM to set the VM was
+ *    always the second ioctl on that context, immediately following
+ *    GEM_CONTEXT_CREATE and setting engines immediately followed that.
+ *
+ * In order for this dance to work properly, any modification to an
+ * i915_gem_proto_context that is exposed to the client via
+ * drm_i915_file_private::proto_context_xa must be guarded by
+ * drm_i915_file_private::proto_context_lock.  The exception is when a
+ * proto-context has not yet been exposed such as when handling
+ * CONTEXT_CREATE_SET_PARAM during GEM_CONTEXT_CREATE.
+ */
+struct i915_gem_proto_context {
+	/** @vm: See &i915_gem_context.vm */
+	struct i915_address_space *vm;
+
+	/** @user_flags: See &i915_gem_context.user_flags */
+	unsigned long user_flags;
+
+	/** @sched: See &i915_gem_context.sched */
+	struct i915_sched_attr sched;
+
+	/** @num_user_engines: Number of user-specified engines or -1 */
+	int num_user_engines;
+
+	/** @user_engines: User-specified engines */
+	struct i915_gem_proto_engine *user_engines;
+
+	/** @legacy_rcs_sseu: Client-set SSEU parameters for the legacy RCS */
+	struct intel_sseu legacy_rcs_sseu;
+
+	/** @single_timeline: See See &i915_gem_context.syncobj */
+	bool single_timeline;
+
+	/** @uses_protected_content: See &i915_gem_context.uses_protected_content */
+	bool uses_protected_content;
+
+	/** @pxp_wakeref: See &i915_gem_context.pxp_wakeref */
+	intel_wakeref_t pxp_wakeref;
+};
+
+/**
+ * struct i915_gem_context - client state
+ *
+ * The struct i915_gem_context represents the combined view of the driver and
+ * logical hardware state for a particular client.
+ */
+struct i915_gem_context {
+	/** @i915: i915 device backpointer */
+	struct drm_i915_private *i915;
+
+	/** @file_priv: owning file descriptor */
+	struct drm_i915_file_private *file_priv;
+
+	/**
+	 * @engines: User defined engines for this context
+	 *
+	 * Various uAPI offer the ability to lookup up an
+	 * index from this array to select an engine operate on.
+	 *
+	 * Multiple logically distinct instances of the same engine
+	 * may be defined in the array, as well as composite virtual
+	 * engines.
+	 *
+	 * Execbuf uses the I915_EXEC_RING_MASK as an index into this
+	 * array to select which HW context + engine to execute on. For
+	 * the default array, the user_ring_map[] is used to translate
+	 * the legacy uABI onto the approprate index (e.g. both
+	 * I915_EXEC_DEFAULT and I915_EXEC_RENDER select the same
+	 * context, and I915_EXEC_BSD is weird). For a use defined
+	 * array, execbuf uses I915_EXEC_RING_MASK as a plain index.
+	 *
+	 * User defined by I915_CONTEXT_PARAM_ENGINE (when the
+	 * CONTEXT_USER_ENGINES flag is set).
+	 */
+	struct i915_gem_engines __rcu *engines;
+
+	/** @engines_mutex: guards writes to engines */
+	struct mutex engines_mutex;
+
+	/**
+	 * @syncobj: Shared timeline syncobj
+	 *
+	 * When the SHARED_TIMELINE flag is set on context creation, we
+	 * emulate a single timeline across all engines using this syncobj.
+	 * For every execbuffer2 call, this syncobj is used as both an in-
+	 * and out-fence.  Unlike the real intel_timeline, this doesn't
+	 * provide perfect atomic in-order guarantees if the client races
+	 * with itself by calling execbuffer2 twice concurrently.  However,
+	 * if userspace races with itself, that's not likely to yield well-
+	 * defined results anyway so we choose to not care.
+	 */
+	struct drm_syncobj *syncobj;
+
+	/**
+	 * @vm: unique address space (GTT)
+	 *
+	 * In full-ppgtt mode, each context has its own address space ensuring
+	 * complete seperation of one client from all others.
+	 *
+	 * In other modes, this is a NULL pointer with the expectation that
+	 * the caller uses the shared global GTT.
+	 */
+	struct i915_address_space *vm;
+
+	/**
+	 * @pid: process id of creator
+	 *
+	 * Note that who created the context may not be the principle user,
+	 * as the context may be shared across a local socket. However,
+	 * that should only affect the default context, all contexts created
+	 * explicitly by the client are expected to be isolated.
+	 */
+	struct pid *pid;
+
+	/** @link: place with &drm_i915_private.context_list */
+	struct list_head link;
+
+	/** @client: struct i915_drm_client */
+	struct i915_drm_client *client;
+
+	/** @client_link: for linking onto &i915_drm_client.ctx_list */
+	struct list_head client_link;
+
+	/**
+	 * @ref: reference count
+	 *
+	 * A reference to a context is held by both the client who created it
+	 * and on each request submitted to the hardware using the request
+	 * (to ensure the hardware has access to the state until it has
+	 * finished all pending writes). See i915_gem_context_get() and
+	 * i915_gem_context_put() for access.
+	 */
+	struct kref ref;
+
+	/**
+	 * @release_work:
+	 *
+	 * Work item for deferred cleanup, since i915_gem_context_put() tends to
+	 * be called from hardirq context.
+	 *
+	 * FIXME: The only real reason for this is &i915_gem_engines.fence, all
+	 * other callers are from process context and need at most some mild
+	 * shuffling to pull the i915_gem_context_put() call out of a spinlock.
+	 */
+	struct work_struct release_work;
+
+	/**
+	 * @rcu: rcu_head for deferred freeing.
+	 */
+	struct rcu_head rcu;
+
+	/**
+	 * @user_flags: small set of booleans controlled by the user
+	 */
+	unsigned long user_flags;
+#define UCONTEXT_NO_ERROR_CAPTURE	1
+#define UCONTEXT_BANNABLE		2
+#define UCONTEXT_RECOVERABLE		3
+#define UCONTEXT_PERSISTENCE		4
+
+	/**
+	 * @flags: small set of booleans
+	 */
+	unsigned long flags;
+#define CONTEXT_CLOSED			0
+#define CONTEXT_USER_ENGINES		1
+
+	/**
+	 * @uses_protected_content: context uses PXP-encrypted objects.
+	 *
+	 * This flag can only be set at ctx creation time and it's immutable for
+	 * the lifetime of the context. See I915_CONTEXT_PARAM_PROTECTED_CONTENT
+	 * in uapi/drm/i915_drm.h for more info on setting restrictions and
+	 * expected behaviour of marked contexts.
+	 */
+	bool uses_protected_content;
+
+	/**
+	 * @pxp_wakeref: wakeref to keep the device awake when PXP is in use
+	 *
+	 * PXP sessions are invalidated when the device is suspended, which in
+	 * turns invalidates all contexts and objects using it. To keep the
+	 * flow simple, we keep the device awake when contexts using PXP objects
+	 * are in use. It is expected that the userspace application only uses
+	 * PXP when the display is on, so taking a wakeref here shouldn't worsen
+	 * our power metrics.
+	 */
+	intel_wakeref_t pxp_wakeref;
+
+	/** @mutex: guards everything that isn't engines or handles_vma */
+	struct mutex mutex;
+
+	/** @sched: scheduler parameters */
+	struct i915_sched_attr sched;
+
+	/** @guilty_count: How many times this context has caused a GPU hang. */
+	atomic_t guilty_count;
+	/**
+	 * @active_count: How many times this context was active during a GPU
+	 * hang, but did not cause it.
+	 */
+	atomic_t active_count;
+
+	/**
+	 * @hang_timestamp: The last time(s) this context caused a GPU hang
+	 */
+	unsigned long hang_timestamp[2];
+#define CONTEXT_FAST_HANG_JIFFIES (120 * HZ) /* 3 hangs within 120s? Banned! */
+
+	/** @remap_slice: Bitmask of cache lines that need remapping */
+	u8 remap_slice;
+
+	/**
+	 * @handles_vma: rbtree to look up our context specific obj/vma for
+	 * the user handle. (user handles are per fd, but the binding is
+	 * per vm, which may be one per context or shared with the global GTT)
+	 */
+	struct radix_tree_root handles_vma;
+
+	/** @lut_mutex: Locks handles_vma */
+	struct mutex lut_mutex;
+
+	/**
+	 * @name: arbitrary name, used for user debug
+	 *
+	 * A name is constructed for the context from the creator's process
+	 * name, pid and user handle in order to uniquely identify the
+	 * context in messages.
+	 */
+	char name[TASK_COMM_LEN + 8];
+
+	/** @stale: tracks stale engines to be destroyed */
+	struct {
+		/** @lock: guards engines */
+		spinlock_t lock;
+		/** @engines: list of stale engines */
+		struct list_head engines;
+	} stale;
+};
+
+#endif /* __I915_GEM_CONTEXT_TYPES_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_create.c b/drivers/gpu/drm/i915/gem/i915_gem_create.c
new file mode 100644
index 000000000..33673fe7e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_create.c
@@ -0,0 +1,456 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <drm/drm_fourcc.h>
+
+#include "gem/i915_gem_ioctls.h"
+#include "gem/i915_gem_lmem.h"
+#include "gem/i915_gem_region.h"
+#include "pxp/intel_pxp.h"
+
+#include "i915_drv.h"
+#include "i915_gem_create.h"
+#include "i915_trace.h"
+#include "i915_user_extensions.h"
+
+static u32 object_max_page_size(struct intel_memory_region **placements,
+				unsigned int n_placements)
+{
+	u32 max_page_size = 0;
+	int i;
+
+	for (i = 0; i < n_placements; i++) {
+		struct intel_memory_region *mr = placements[i];
+
+		GEM_BUG_ON(!is_power_of_2(mr->min_page_size));
+		max_page_size = max_t(u32, max_page_size, mr->min_page_size);
+	}
+
+	GEM_BUG_ON(!max_page_size);
+	return max_page_size;
+}
+
+static int object_set_placements(struct drm_i915_gem_object *obj,
+				 struct intel_memory_region **placements,
+				 unsigned int n_placements)
+{
+	struct intel_memory_region **arr;
+	unsigned int i;
+
+	GEM_BUG_ON(!n_placements);
+
+	/*
+	 * For the common case of one memory region, skip storing an
+	 * allocated array and just point at the region directly.
+	 */
+	if (n_placements == 1) {
+		struct intel_memory_region *mr = placements[0];
+		struct drm_i915_private *i915 = mr->i915;
+
+		obj->mm.placements = &i915->mm.regions[mr->id];
+		obj->mm.n_placements = 1;
+	} else {
+		arr = kmalloc_array(n_placements,
+				    sizeof(struct intel_memory_region *),
+				    GFP_KERNEL);
+		if (!arr)
+			return -ENOMEM;
+
+		for (i = 0; i < n_placements; i++)
+			arr[i] = placements[i];
+
+		obj->mm.placements = arr;
+		obj->mm.n_placements = n_placements;
+	}
+
+	return 0;
+}
+
+static int i915_gem_publish(struct drm_i915_gem_object *obj,
+			    struct drm_file *file,
+			    u64 *size_p,
+			    u32 *handle_p)
+{
+	u64 size = obj->base.size;
+	int ret;
+
+	ret = drm_gem_handle_create(file, &obj->base, handle_p);
+	/* drop reference from allocate - handle holds it now */
+	i915_gem_object_put(obj);
+	if (ret)
+		return ret;
+
+	*size_p = size;
+	return 0;
+}
+
+static struct drm_i915_gem_object *
+__i915_gem_object_create_user_ext(struct drm_i915_private *i915, u64 size,
+				  struct intel_memory_region **placements,
+				  unsigned int n_placements,
+				  unsigned int ext_flags)
+{
+	struct intel_memory_region *mr = placements[0];
+	struct drm_i915_gem_object *obj;
+	unsigned int flags;
+	int ret;
+
+	i915_gem_flush_free_objects(i915);
+
+	size = round_up(size, object_max_page_size(placements, n_placements));
+	if (size == 0)
+		return ERR_PTR(-EINVAL);
+
+	/* For most of the ABI (e.g. mmap) we think in system pages */
+	GEM_BUG_ON(!IS_ALIGNED(size, PAGE_SIZE));
+
+	if (i915_gem_object_size_2big(size))
+		return ERR_PTR(-E2BIG);
+
+	obj = i915_gem_object_alloc();
+	if (!obj)
+		return ERR_PTR(-ENOMEM);
+
+	ret = object_set_placements(obj, placements, n_placements);
+	if (ret)
+		goto object_free;
+
+	/*
+	 * I915_BO_ALLOC_USER will make sure the object is cleared before
+	 * any user access.
+	 */
+	flags = I915_BO_ALLOC_USER;
+
+	ret = mr->ops->init_object(mr, obj, I915_BO_INVALID_OFFSET, size, 0, flags);
+	if (ret)
+		goto object_free;
+
+	GEM_BUG_ON(size != obj->base.size);
+
+	/* Add any flag set by create_ext options */
+	obj->flags |= ext_flags;
+
+	trace_i915_gem_object_create(obj);
+	return obj;
+
+object_free:
+	if (obj->mm.n_placements > 1)
+		kfree(obj->mm.placements);
+	i915_gem_object_free(obj);
+	return ERR_PTR(ret);
+}
+
+/**
+ * Creates a new object using the same path as DRM_I915_GEM_CREATE_EXT
+ * @i915: i915 private
+ * @size: size of the buffer, in bytes
+ * @placements: possible placement regions, in priority order
+ * @n_placements: number of possible placement regions
+ *
+ * This function is exposed primarily for selftests and does very little
+ * error checking.  It is assumed that the set of placement regions has
+ * already been verified to be valid.
+ */
+struct drm_i915_gem_object *
+__i915_gem_object_create_user(struct drm_i915_private *i915, u64 size,
+			      struct intel_memory_region **placements,
+			      unsigned int n_placements)
+{
+	return __i915_gem_object_create_user_ext(i915, size, placements,
+						 n_placements, 0);
+}
+
+int
+i915_gem_dumb_create(struct drm_file *file,
+		     struct drm_device *dev,
+		     struct drm_mode_create_dumb *args)
+{
+	struct drm_i915_gem_object *obj;
+	struct intel_memory_region *mr;
+	enum intel_memory_type mem_type;
+	int cpp = DIV_ROUND_UP(args->bpp, 8);
+	u32 format;
+
+	switch (cpp) {
+	case 1:
+		format = DRM_FORMAT_C8;
+		break;
+	case 2:
+		format = DRM_FORMAT_RGB565;
+		break;
+	case 4:
+		format = DRM_FORMAT_XRGB8888;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	/* have to work out size/pitch and return them */
+	args->pitch = ALIGN(args->width * cpp, 64);
+
+	/* align stride to page size so that we can remap */
+	if (args->pitch > intel_plane_fb_max_stride(to_i915(dev), format,
+						    DRM_FORMAT_MOD_LINEAR))
+		args->pitch = ALIGN(args->pitch, 4096);
+
+	if (args->pitch < args->width)
+		return -EINVAL;
+
+	args->size = mul_u32_u32(args->pitch, args->height);
+
+	mem_type = INTEL_MEMORY_SYSTEM;
+	if (HAS_LMEM(to_i915(dev)))
+		mem_type = INTEL_MEMORY_LOCAL;
+
+	mr = intel_memory_region_by_type(to_i915(dev), mem_type);
+
+	obj = __i915_gem_object_create_user(to_i915(dev), args->size, &mr, 1);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	return i915_gem_publish(obj, file, &args->size, &args->handle);
+}
+
+/**
+ * Creates a new mm object and returns a handle to it.
+ * @dev: drm device pointer
+ * @data: ioctl data blob
+ * @file: drm file pointer
+ */
+int
+i915_gem_create_ioctl(struct drm_device *dev, void *data,
+		      struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_create *args = data;
+	struct drm_i915_gem_object *obj;
+	struct intel_memory_region *mr;
+
+	mr = intel_memory_region_by_type(i915, INTEL_MEMORY_SYSTEM);
+
+	obj = __i915_gem_object_create_user(i915, args->size, &mr, 1);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	return i915_gem_publish(obj, file, &args->size, &args->handle);
+}
+
+struct create_ext {
+	struct drm_i915_private *i915;
+	struct intel_memory_region *placements[INTEL_REGION_UNKNOWN];
+	unsigned int n_placements;
+	unsigned int placement_mask;
+	unsigned long flags;
+};
+
+static void repr_placements(char *buf, size_t size,
+			    struct intel_memory_region **placements,
+			    int n_placements)
+{
+	int i;
+
+	buf[0] = '\0';
+
+	for (i = 0; i < n_placements; i++) {
+		struct intel_memory_region *mr = placements[i];
+		int r;
+
+		r = snprintf(buf, size, "\n  %s -> { class: %d, inst: %d }",
+			     mr->name, mr->type, mr->instance);
+		if (r >= size)
+			return;
+
+		buf += r;
+		size -= r;
+	}
+}
+
+static int set_placements(struct drm_i915_gem_create_ext_memory_regions *args,
+			  struct create_ext *ext_data)
+{
+	struct drm_i915_private *i915 = ext_data->i915;
+	struct drm_i915_gem_memory_class_instance __user *uregions =
+		u64_to_user_ptr(args->regions);
+	struct intel_memory_region *placements[INTEL_REGION_UNKNOWN];
+	u32 mask;
+	int i, ret = 0;
+
+	if (args->pad) {
+		drm_dbg(&i915->drm, "pad should be zero\n");
+		ret = -EINVAL;
+	}
+
+	if (!args->num_regions) {
+		drm_dbg(&i915->drm, "num_regions is zero\n");
+		ret = -EINVAL;
+	}
+
+	BUILD_BUG_ON(ARRAY_SIZE(i915->mm.regions) != ARRAY_SIZE(placements));
+	BUILD_BUG_ON(ARRAY_SIZE(ext_data->placements) != ARRAY_SIZE(placements));
+	if (args->num_regions > ARRAY_SIZE(i915->mm.regions)) {
+		drm_dbg(&i915->drm, "num_regions is too large\n");
+		ret = -EINVAL;
+	}
+
+	if (ret)
+		return ret;
+
+	mask = 0;
+	for (i = 0; i < args->num_regions; i++) {
+		struct drm_i915_gem_memory_class_instance region;
+		struct intel_memory_region *mr;
+
+		if (copy_from_user(&region, uregions, sizeof(region)))
+			return -EFAULT;
+
+		mr = intel_memory_region_lookup(i915,
+						region.memory_class,
+						region.memory_instance);
+		if (!mr || mr->private) {
+			drm_dbg(&i915->drm, "Device is missing region { class: %d, inst: %d } at index = %d\n",
+				region.memory_class, region.memory_instance, i);
+			ret = -EINVAL;
+			goto out_dump;
+		}
+
+		if (mask & BIT(mr->id)) {
+			drm_dbg(&i915->drm, "Found duplicate placement %s -> { class: %d, inst: %d } at index = %d\n",
+				mr->name, region.memory_class,
+				region.memory_instance, i);
+			ret = -EINVAL;
+			goto out_dump;
+		}
+
+		placements[i] = mr;
+		mask |= BIT(mr->id);
+
+		++uregions;
+	}
+
+	if (ext_data->n_placements) {
+		ret = -EINVAL;
+		goto out_dump;
+	}
+
+	ext_data->n_placements = args->num_regions;
+	for (i = 0; i < args->num_regions; i++)
+		ext_data->placements[i] = placements[i];
+
+	ext_data->placement_mask = mask;
+	return 0;
+
+out_dump:
+	if (1) {
+		char buf[256];
+
+		if (ext_data->n_placements) {
+			repr_placements(buf,
+					sizeof(buf),
+					ext_data->placements,
+					ext_data->n_placements);
+			drm_dbg(&i915->drm,
+				"Placements were already set in previous EXT. Existing placements: %s\n",
+				buf);
+		}
+
+		repr_placements(buf, sizeof(buf), placements, i);
+		drm_dbg(&i915->drm, "New placements(so far validated): %s\n", buf);
+	}
+
+	return ret;
+}
+
+static int ext_set_placements(struct i915_user_extension __user *base,
+			      void *data)
+{
+	struct drm_i915_gem_create_ext_memory_regions ext;
+
+	if (copy_from_user(&ext, base, sizeof(ext)))
+		return -EFAULT;
+
+	return set_placements(&ext, data);
+}
+
+static int ext_set_protected(struct i915_user_extension __user *base, void *data)
+{
+	struct drm_i915_gem_create_ext_protected_content ext;
+	struct create_ext *ext_data = data;
+
+	if (copy_from_user(&ext, base, sizeof(ext)))
+		return -EFAULT;
+
+	if (ext.flags)
+		return -EINVAL;
+
+	if (!intel_pxp_is_enabled(&to_gt(ext_data->i915)->pxp))
+		return -ENODEV;
+
+	ext_data->flags |= I915_BO_PROTECTED;
+
+	return 0;
+}
+
+static const i915_user_extension_fn create_extensions[] = {
+	[I915_GEM_CREATE_EXT_MEMORY_REGIONS] = ext_set_placements,
+	[I915_GEM_CREATE_EXT_PROTECTED_CONTENT] = ext_set_protected,
+};
+
+/**
+ * Creates a new mm object and returns a handle to it.
+ * @dev: drm device pointer
+ * @data: ioctl data blob
+ * @file: drm file pointer
+ */
+int
+i915_gem_create_ext_ioctl(struct drm_device *dev, void *data,
+			  struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_create_ext *args = data;
+	struct create_ext ext_data = { .i915 = i915 };
+	struct drm_i915_gem_object *obj;
+	int ret;
+
+	if (args->flags & ~I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS)
+		return -EINVAL;
+
+	ret = i915_user_extensions(u64_to_user_ptr(args->extensions),
+				   create_extensions,
+				   ARRAY_SIZE(create_extensions),
+				   &ext_data);
+	if (ret)
+		return ret;
+
+	if (!ext_data.n_placements) {
+		ext_data.placements[0] =
+			intel_memory_region_by_type(i915, INTEL_MEMORY_SYSTEM);
+		ext_data.n_placements = 1;
+	}
+
+	if (args->flags & I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS) {
+		if (ext_data.n_placements == 1)
+			return -EINVAL;
+
+		/*
+		 * We always need to be able to spill to system memory, if we
+		 * can't place in the mappable part of LMEM.
+		 */
+		if (!(ext_data.placement_mask & BIT(INTEL_REGION_SMEM)))
+			return -EINVAL;
+	} else {
+		if (ext_data.n_placements > 1 ||
+		    ext_data.placements[0]->type != INTEL_MEMORY_SYSTEM)
+			ext_data.flags |= I915_BO_ALLOC_GPU_ONLY;
+	}
+
+	obj = __i915_gem_object_create_user_ext(i915, args->size,
+						ext_data.placements,
+						ext_data.n_placements,
+						ext_data.flags);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	return i915_gem_publish(obj, file, &args->size, &args->handle);
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_create.h b/drivers/gpu/drm/i915/gem/i915_gem_create.h
new file mode 100644
index 000000000..9536aa906
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_create.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_GEM_CREATE_H__
+#define __I915_GEM_CREATE_H__
+
+struct drm_file;
+struct drm_device;
+struct drm_mode_create_dumb;
+
+int i915_gem_dumb_create(struct drm_file *file_priv,
+			 struct drm_device *dev,
+			 struct drm_mode_create_dumb *args);
+
+#endif /* __I915_GEM_CREATE_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c
new file mode 100644
index 000000000..824971a1c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.c
@@ -0,0 +1,349 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright 2012 Red Hat Inc
+ */
+
+#include <linux/dma-buf.h>
+#include <linux/highmem.h>
+#include <linux/dma-resv.h>
+#include <linux/module.h>
+
+#include <asm/smp.h>
+
+#include "gem/i915_gem_dmabuf.h"
+#include "i915_drv.h"
+#include "i915_gem_object.h"
+#include "i915_scatterlist.h"
+
+MODULE_IMPORT_NS(DMA_BUF);
+
+I915_SELFTEST_DECLARE(static bool force_different_devices;)
+
+static struct drm_i915_gem_object *dma_buf_to_obj(struct dma_buf *buf)
+{
+	return to_intel_bo(buf->priv);
+}
+
+static struct sg_table *i915_gem_map_dma_buf(struct dma_buf_attachment *attachment,
+					     enum dma_data_direction dir)
+{
+	struct drm_i915_gem_object *obj = dma_buf_to_obj(attachment->dmabuf);
+	struct sg_table *st;
+	struct scatterlist *src, *dst;
+	int ret, i;
+
+	/* Copy sg so that we make an independent mapping */
+	st = kmalloc(sizeof(struct sg_table), GFP_KERNEL);
+	if (st == NULL) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	ret = sg_alloc_table(st, obj->mm.pages->orig_nents, GFP_KERNEL);
+	if (ret)
+		goto err_free;
+
+	src = obj->mm.pages->sgl;
+	dst = st->sgl;
+	for (i = 0; i < obj->mm.pages->orig_nents; i++) {
+		sg_set_page(dst, sg_page(src), src->length, 0);
+		dst = sg_next(dst);
+		src = sg_next(src);
+	}
+
+	ret = dma_map_sgtable(attachment->dev, st, dir, DMA_ATTR_SKIP_CPU_SYNC);
+	if (ret)
+		goto err_free_sg;
+
+	return st;
+
+err_free_sg:
+	sg_free_table(st);
+err_free:
+	kfree(st);
+err:
+	return ERR_PTR(ret);
+}
+
+static int i915_gem_dmabuf_vmap(struct dma_buf *dma_buf,
+				struct iosys_map *map)
+{
+	struct drm_i915_gem_object *obj = dma_buf_to_obj(dma_buf);
+	void *vaddr;
+
+	vaddr = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+	if (IS_ERR(vaddr))
+		return PTR_ERR(vaddr);
+
+	iosys_map_set_vaddr(map, vaddr);
+
+	return 0;
+}
+
+static void i915_gem_dmabuf_vunmap(struct dma_buf *dma_buf,
+				   struct iosys_map *map)
+{
+	struct drm_i915_gem_object *obj = dma_buf_to_obj(dma_buf);
+
+	i915_gem_object_flush_map(obj);
+	i915_gem_object_unpin_map(obj);
+}
+
+static int i915_gem_dmabuf_mmap(struct dma_buf *dma_buf, struct vm_area_struct *vma)
+{
+	struct drm_i915_gem_object *obj = dma_buf_to_obj(dma_buf);
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	int ret;
+
+	if (obj->base.size < vma->vm_end - vma->vm_start)
+		return -EINVAL;
+
+	if (HAS_LMEM(i915))
+		return drm_gem_prime_mmap(&obj->base, vma);
+
+	if (!obj->base.filp)
+		return -ENODEV;
+
+	ret = call_mmap(obj->base.filp, vma);
+	if (ret)
+		return ret;
+
+	vma_set_file(vma, obj->base.filp);
+
+	return 0;
+}
+
+static int i915_gem_begin_cpu_access(struct dma_buf *dma_buf, enum dma_data_direction direction)
+{
+	struct drm_i915_gem_object *obj = dma_buf_to_obj(dma_buf);
+	bool write = (direction == DMA_BIDIRECTIONAL || direction == DMA_TO_DEVICE);
+	struct i915_gem_ww_ctx ww;
+	int err;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	err = i915_gem_object_lock(obj, &ww);
+	if (!err)
+		err = i915_gem_object_pin_pages(obj);
+	if (!err) {
+		err = i915_gem_object_set_to_cpu_domain(obj, write);
+		i915_gem_object_unpin_pages(obj);
+	}
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	return err;
+}
+
+static int i915_gem_end_cpu_access(struct dma_buf *dma_buf, enum dma_data_direction direction)
+{
+	struct drm_i915_gem_object *obj = dma_buf_to_obj(dma_buf);
+	struct i915_gem_ww_ctx ww;
+	int err;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	err = i915_gem_object_lock(obj, &ww);
+	if (!err)
+		err = i915_gem_object_pin_pages(obj);
+	if (!err) {
+		err = i915_gem_object_set_to_gtt_domain(obj, false);
+		i915_gem_object_unpin_pages(obj);
+	}
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	return err;
+}
+
+static int i915_gem_dmabuf_attach(struct dma_buf *dmabuf,
+				  struct dma_buf_attachment *attach)
+{
+	struct drm_i915_gem_object *obj = dma_buf_to_obj(dmabuf);
+	struct i915_gem_ww_ctx ww;
+	int err;
+
+	if (!i915_gem_object_can_migrate(obj, INTEL_REGION_SMEM))
+		return -EOPNOTSUPP;
+
+	for_i915_gem_ww(&ww, err, true) {
+		err = i915_gem_object_lock(obj, &ww);
+		if (err)
+			continue;
+
+		err = i915_gem_object_migrate(obj, &ww, INTEL_REGION_SMEM);
+		if (err)
+			continue;
+
+		err = i915_gem_object_wait_migration(obj, 0);
+		if (err)
+			continue;
+
+		err = i915_gem_object_pin_pages(obj);
+	}
+
+	return err;
+}
+
+static void i915_gem_dmabuf_detach(struct dma_buf *dmabuf,
+				   struct dma_buf_attachment *attach)
+{
+	struct drm_i915_gem_object *obj = dma_buf_to_obj(dmabuf);
+
+	i915_gem_object_unpin_pages(obj);
+}
+
+static const struct dma_buf_ops i915_dmabuf_ops =  {
+	.attach = i915_gem_dmabuf_attach,
+	.detach = i915_gem_dmabuf_detach,
+	.map_dma_buf = i915_gem_map_dma_buf,
+	.unmap_dma_buf = drm_gem_unmap_dma_buf,
+	.release = drm_gem_dmabuf_release,
+	.mmap = i915_gem_dmabuf_mmap,
+	.vmap = i915_gem_dmabuf_vmap,
+	.vunmap = i915_gem_dmabuf_vunmap,
+	.begin_cpu_access = i915_gem_begin_cpu_access,
+	.end_cpu_access = i915_gem_end_cpu_access,
+};
+
+struct dma_buf *i915_gem_prime_export(struct drm_gem_object *gem_obj, int flags)
+{
+	struct drm_i915_gem_object *obj = to_intel_bo(gem_obj);
+	DEFINE_DMA_BUF_EXPORT_INFO(exp_info);
+
+	exp_info.ops = &i915_dmabuf_ops;
+	exp_info.size = gem_obj->size;
+	exp_info.flags = flags;
+	exp_info.priv = gem_obj;
+	exp_info.resv = obj->base.resv;
+
+	if (obj->ops->dmabuf_export) {
+		int ret = obj->ops->dmabuf_export(obj);
+		if (ret)
+			return ERR_PTR(ret);
+	}
+
+	return drm_gem_dmabuf_export(gem_obj->dev, &exp_info);
+}
+
+static int i915_gem_object_get_pages_dmabuf(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct sg_table *pages;
+	unsigned int sg_page_sizes;
+
+	assert_object_held(obj);
+
+	pages = dma_buf_map_attachment(obj->base.import_attach,
+				       DMA_BIDIRECTIONAL);
+	if (IS_ERR(pages))
+		return PTR_ERR(pages);
+
+	/*
+	 * DG1 is special here since it still snoops transactions even with
+	 * CACHE_NONE. This is not the case with other HAS_SNOOP platforms. We
+	 * might need to revisit this as we add new discrete platforms.
+	 *
+	 * XXX: Consider doing a vmap flush or something, where possible.
+	 * Currently we just do a heavy handed wbinvd_on_all_cpus() here since
+	 * the underlying sg_table might not even point to struct pages, so we
+	 * can't just call drm_clflush_sg or similar, like we do elsewhere in
+	 * the driver.
+	 */
+	if (i915_gem_object_can_bypass_llc(obj) ||
+	    (!HAS_LLC(i915) && !IS_DG1(i915)))
+		wbinvd_on_all_cpus();
+
+	sg_page_sizes = i915_sg_dma_sizes(pages->sgl);
+	__i915_gem_object_set_pages(obj, pages, sg_page_sizes);
+
+	return 0;
+}
+
+static void i915_gem_object_put_pages_dmabuf(struct drm_i915_gem_object *obj,
+					     struct sg_table *pages)
+{
+	dma_buf_unmap_attachment(obj->base.import_attach, pages,
+				 DMA_BIDIRECTIONAL);
+}
+
+static const struct drm_i915_gem_object_ops i915_gem_object_dmabuf_ops = {
+	.name = "i915_gem_object_dmabuf",
+	.get_pages = i915_gem_object_get_pages_dmabuf,
+	.put_pages = i915_gem_object_put_pages_dmabuf,
+};
+
+struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev,
+					     struct dma_buf *dma_buf)
+{
+	static struct lock_class_key lock_class;
+	struct dma_buf_attachment *attach;
+	struct drm_i915_gem_object *obj;
+	int ret;
+
+	/* is this one of own objects? */
+	if (dma_buf->ops == &i915_dmabuf_ops) {
+		obj = dma_buf_to_obj(dma_buf);
+		/* is it from our device? */
+		if (obj->base.dev == dev &&
+		    !I915_SELFTEST_ONLY(force_different_devices)) {
+			/*
+			 * Importing dmabuf exported from out own gem increases
+			 * refcount on gem itself instead of f_count of dmabuf.
+			 */
+			return &i915_gem_object_get(obj)->base;
+		}
+	}
+
+	if (i915_gem_object_size_2big(dma_buf->size))
+		return ERR_PTR(-E2BIG);
+
+	/* need to attach */
+	attach = dma_buf_attach(dma_buf, dev->dev);
+	if (IS_ERR(attach))
+		return ERR_CAST(attach);
+
+	get_dma_buf(dma_buf);
+
+	obj = i915_gem_object_alloc();
+	if (obj == NULL) {
+		ret = -ENOMEM;
+		goto fail_detach;
+	}
+
+	drm_gem_private_object_init(dev, &obj->base, dma_buf->size);
+	i915_gem_object_init(obj, &i915_gem_object_dmabuf_ops, &lock_class,
+			     I915_BO_ALLOC_USER);
+	obj->base.import_attach = attach;
+	obj->base.resv = dma_buf->resv;
+
+	/* We use GTT as shorthand for a coherent domain, one that is
+	 * neither in the GPU cache nor in the CPU cache, where all
+	 * writes are immediately visible in memory. (That's not strictly
+	 * true, but it's close! There are internal buffers such as the
+	 * write-combined buffer or a delay through the chipset for GTT
+	 * writes that do require us to treat GTT as a separate cache domain.)
+	 */
+	obj->read_domains = I915_GEM_DOMAIN_GTT;
+	obj->write_domain = 0;
+
+	return &obj->base;
+
+fail_detach:
+	dma_buf_detach(dma_buf, attach);
+	dma_buf_put(dma_buf);
+
+	return ERR_PTR(ret);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_dmabuf.c"
+#include "selftests/i915_gem_dmabuf.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.h b/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.h
new file mode 100644
index 000000000..6e0405d47
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_dmabuf.h
@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __I915_GEM_DMABUF_H__
+#define __I915_GEM_DMABUF_H__
+
+struct drm_gem_object;
+struct drm_device;
+struct dma_buf;
+
+struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev,
+					     struct dma_buf *dma_buf);
+
+struct dma_buf *i915_gem_prime_export(struct drm_gem_object *gem_obj, int flags);
+
+#endif /* __I915_GEM_DMABUF_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_domain.c b/drivers/gpu/drm/i915/gem/i915_gem_domain.c
new file mode 100644
index 000000000..d44a152ce
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_domain.c
@@ -0,0 +1,740 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2014-2016 Intel Corporation
+ */
+
+#include "display/intel_frontbuffer.h"
+#include "gt/intel_gt.h"
+
+#include "i915_drv.h"
+#include "i915_gem_clflush.h"
+#include "i915_gem_domain.h"
+#include "i915_gem_gtt.h"
+#include "i915_gem_ioctls.h"
+#include "i915_gem_lmem.h"
+#include "i915_gem_mman.h"
+#include "i915_gem_object.h"
+#include "i915_vma.h"
+
+static bool gpu_write_needs_clflush(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+
+	if (IS_DGFX(i915))
+		return false;
+
+	return !(obj->cache_level == I915_CACHE_NONE ||
+		 obj->cache_level == I915_CACHE_WT);
+}
+
+bool i915_gem_cpu_write_needs_clflush(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+
+	if (obj->cache_dirty)
+		return false;
+
+	if (IS_DGFX(i915))
+		return false;
+
+	if (!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_WRITE))
+		return true;
+
+	/* Currently in use by HW (display engine)? Keep flushed. */
+	return i915_gem_object_is_framebuffer(obj);
+}
+
+static void
+flush_write_domain(struct drm_i915_gem_object *obj, unsigned int flush_domains)
+{
+	struct i915_vma *vma;
+
+	assert_object_held(obj);
+
+	if (!(obj->write_domain & flush_domains))
+		return;
+
+	switch (obj->write_domain) {
+	case I915_GEM_DOMAIN_GTT:
+		spin_lock(&obj->vma.lock);
+		for_each_ggtt_vma(vma, obj) {
+			if (i915_vma_unset_ggtt_write(vma))
+				intel_gt_flush_ggtt_writes(vma->vm->gt);
+		}
+		spin_unlock(&obj->vma.lock);
+
+		i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
+		break;
+
+	case I915_GEM_DOMAIN_WC:
+		wmb();
+		break;
+
+	case I915_GEM_DOMAIN_CPU:
+		i915_gem_clflush_object(obj, I915_CLFLUSH_SYNC);
+		break;
+
+	case I915_GEM_DOMAIN_RENDER:
+		if (gpu_write_needs_clflush(obj))
+			obj->cache_dirty = true;
+		break;
+	}
+
+	obj->write_domain = 0;
+}
+
+static void __i915_gem_object_flush_for_display(struct drm_i915_gem_object *obj)
+{
+	/*
+	 * We manually flush the CPU domain so that we can override and
+	 * force the flush for the display, and perform it asyncrhonously.
+	 */
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
+	if (obj->cache_dirty)
+		i915_gem_clflush_object(obj, I915_CLFLUSH_FORCE);
+	obj->write_domain = 0;
+}
+
+void i915_gem_object_flush_if_display(struct drm_i915_gem_object *obj)
+{
+	if (!i915_gem_object_is_framebuffer(obj))
+		return;
+
+	i915_gem_object_lock(obj, NULL);
+	__i915_gem_object_flush_for_display(obj);
+	i915_gem_object_unlock(obj);
+}
+
+void i915_gem_object_flush_if_display_locked(struct drm_i915_gem_object *obj)
+{
+	if (i915_gem_object_is_framebuffer(obj))
+		__i915_gem_object_flush_for_display(obj);
+}
+
+/**
+ * Moves a single object to the WC read, and possibly write domain.
+ * @obj: object to act on
+ * @write: ask for write access or read only
+ *
+ * This function returns when the move is complete, including waiting on
+ * flushes to occur.
+ */
+int
+i915_gem_object_set_to_wc_domain(struct drm_i915_gem_object *obj, bool write)
+{
+	int ret;
+
+	assert_object_held(obj);
+
+	ret = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE |
+				   (write ? I915_WAIT_ALL : 0),
+				   MAX_SCHEDULE_TIMEOUT);
+	if (ret)
+		return ret;
+
+	if (obj->write_domain == I915_GEM_DOMAIN_WC)
+		return 0;
+
+	/* Flush and acquire obj->pages so that we are coherent through
+	 * direct access in memory with previous cached writes through
+	 * shmemfs and that our cache domain tracking remains valid.
+	 * For example, if the obj->filp was moved to swap without us
+	 * being notified and releasing the pages, we would mistakenly
+	 * continue to assume that the obj remained out of the CPU cached
+	 * domain.
+	 */
+	ret = i915_gem_object_pin_pages(obj);
+	if (ret)
+		return ret;
+
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_WC);
+
+	/* Serialise direct access to this object with the barriers for
+	 * coherent writes from the GPU, by effectively invalidating the
+	 * WC domain upon first access.
+	 */
+	if ((obj->read_domains & I915_GEM_DOMAIN_WC) == 0)
+		mb();
+
+	/* It should now be out of any other write domains, and we can update
+	 * the domain values for our changes.
+	 */
+	GEM_BUG_ON((obj->write_domain & ~I915_GEM_DOMAIN_WC) != 0);
+	obj->read_domains |= I915_GEM_DOMAIN_WC;
+	if (write) {
+		obj->read_domains = I915_GEM_DOMAIN_WC;
+		obj->write_domain = I915_GEM_DOMAIN_WC;
+		obj->mm.dirty = true;
+	}
+
+	i915_gem_object_unpin_pages(obj);
+	return 0;
+}
+
+/**
+ * Moves a single object to the GTT read, and possibly write domain.
+ * @obj: object to act on
+ * @write: ask for write access or read only
+ *
+ * This function returns when the move is complete, including waiting on
+ * flushes to occur.
+ */
+int
+i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
+{
+	int ret;
+
+	assert_object_held(obj);
+
+	ret = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE |
+				   (write ? I915_WAIT_ALL : 0),
+				   MAX_SCHEDULE_TIMEOUT);
+	if (ret)
+		return ret;
+
+	if (obj->write_domain == I915_GEM_DOMAIN_GTT)
+		return 0;
+
+	/* Flush and acquire obj->pages so that we are coherent through
+	 * direct access in memory with previous cached writes through
+	 * shmemfs and that our cache domain tracking remains valid.
+	 * For example, if the obj->filp was moved to swap without us
+	 * being notified and releasing the pages, we would mistakenly
+	 * continue to assume that the obj remained out of the CPU cached
+	 * domain.
+	 */
+	ret = i915_gem_object_pin_pages(obj);
+	if (ret)
+		return ret;
+
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_GTT);
+
+	/* Serialise direct access to this object with the barriers for
+	 * coherent writes from the GPU, by effectively invalidating the
+	 * GTT domain upon first access.
+	 */
+	if ((obj->read_domains & I915_GEM_DOMAIN_GTT) == 0)
+		mb();
+
+	/* It should now be out of any other write domains, and we can update
+	 * the domain values for our changes.
+	 */
+	GEM_BUG_ON((obj->write_domain & ~I915_GEM_DOMAIN_GTT) != 0);
+	obj->read_domains |= I915_GEM_DOMAIN_GTT;
+	if (write) {
+		struct i915_vma *vma;
+
+		obj->read_domains = I915_GEM_DOMAIN_GTT;
+		obj->write_domain = I915_GEM_DOMAIN_GTT;
+		obj->mm.dirty = true;
+
+		spin_lock(&obj->vma.lock);
+		for_each_ggtt_vma(vma, obj)
+			if (i915_vma_is_bound(vma, I915_VMA_GLOBAL_BIND))
+				i915_vma_set_ggtt_write(vma);
+		spin_unlock(&obj->vma.lock);
+	}
+
+	i915_gem_object_unpin_pages(obj);
+	return 0;
+}
+
+/**
+ * Changes the cache-level of an object across all VMA.
+ * @obj: object to act on
+ * @cache_level: new cache level to set for the object
+ *
+ * After this function returns, the object will be in the new cache-level
+ * across all GTT and the contents of the backing storage will be coherent,
+ * with respect to the new cache-level. In order to keep the backing storage
+ * coherent for all users, we only allow a single cache level to be set
+ * globally on the object and prevent it from being changed whilst the
+ * hardware is reading from the object. That is if the object is currently
+ * on the scanout it will be set to uncached (or equivalent display
+ * cache coherency) and all non-MOCS GPU access will also be uncached so
+ * that all direct access to the scanout remains coherent.
+ */
+int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
+				    enum i915_cache_level cache_level)
+{
+	int ret;
+
+	if (obj->cache_level == cache_level)
+		return 0;
+
+	ret = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE |
+				   I915_WAIT_ALL,
+				   MAX_SCHEDULE_TIMEOUT);
+	if (ret)
+		return ret;
+
+	/* Always invalidate stale cachelines */
+	if (obj->cache_level != cache_level) {
+		i915_gem_object_set_cache_coherency(obj, cache_level);
+		obj->cache_dirty = true;
+	}
+
+	/* The cache-level will be applied when each vma is rebound. */
+	return i915_gem_object_unbind(obj,
+				      I915_GEM_OBJECT_UNBIND_ACTIVE |
+				      I915_GEM_OBJECT_UNBIND_BARRIER);
+}
+
+int i915_gem_get_caching_ioctl(struct drm_device *dev, void *data,
+			       struct drm_file *file)
+{
+	struct drm_i915_gem_caching *args = data;
+	struct drm_i915_gem_object *obj;
+	int err = 0;
+
+	if (IS_DGFX(to_i915(dev)))
+		return -ENODEV;
+
+	rcu_read_lock();
+	obj = i915_gem_object_lookup_rcu(file, args->handle);
+	if (!obj) {
+		err = -ENOENT;
+		goto out;
+	}
+
+	switch (obj->cache_level) {
+	case I915_CACHE_LLC:
+	case I915_CACHE_L3_LLC:
+		args->caching = I915_CACHING_CACHED;
+		break;
+
+	case I915_CACHE_WT:
+		args->caching = I915_CACHING_DISPLAY;
+		break;
+
+	default:
+		args->caching = I915_CACHING_NONE;
+		break;
+	}
+out:
+	rcu_read_unlock();
+	return err;
+}
+
+int i915_gem_set_caching_ioctl(struct drm_device *dev, void *data,
+			       struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_caching *args = data;
+	struct drm_i915_gem_object *obj;
+	enum i915_cache_level level;
+	int ret = 0;
+
+	if (IS_DGFX(i915))
+		return -ENODEV;
+
+	switch (args->caching) {
+	case I915_CACHING_NONE:
+		level = I915_CACHE_NONE;
+		break;
+	case I915_CACHING_CACHED:
+		/*
+		 * Due to a HW issue on BXT A stepping, GPU stores via a
+		 * snooped mapping may leave stale data in a corresponding CPU
+		 * cacheline, whereas normally such cachelines would get
+		 * invalidated.
+		 */
+		if (!HAS_LLC(i915) && !HAS_SNOOP(i915))
+			return -ENODEV;
+
+		level = I915_CACHE_LLC;
+		break;
+	case I915_CACHING_DISPLAY:
+		level = HAS_WT(i915) ? I915_CACHE_WT : I915_CACHE_NONE;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	obj = i915_gem_object_lookup(file, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	/*
+	 * The caching mode of proxy object is handled by its generator, and
+	 * not allowed to be changed by userspace.
+	 */
+	if (i915_gem_object_is_proxy(obj)) {
+		/*
+		 * Silently allow cached for userptr; the vulkan driver
+		 * sets all objects to cached
+		 */
+		if (!i915_gem_object_is_userptr(obj) ||
+		    args->caching != I915_CACHING_CACHED)
+			ret = -ENXIO;
+
+		goto out;
+	}
+
+	ret = i915_gem_object_lock_interruptible(obj, NULL);
+	if (ret)
+		goto out;
+
+	ret = i915_gem_object_set_cache_level(obj, level);
+	i915_gem_object_unlock(obj);
+
+out:
+	i915_gem_object_put(obj);
+	return ret;
+}
+
+/*
+ * Prepare buffer for display plane (scanout, cursors, etc). Can be called from
+ * an uninterruptible phase (modesetting) and allows any flushes to be pipelined
+ * (for pageflips). We only flush the caches while preparing the buffer for
+ * display, the callers are responsible for frontbuffer flush.
+ */
+struct i915_vma *
+i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj,
+				     struct i915_gem_ww_ctx *ww,
+				     u32 alignment,
+				     const struct i915_gtt_view *view,
+				     unsigned int flags)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_vma *vma;
+	int ret;
+
+	/* Frame buffer must be in LMEM */
+	if (HAS_LMEM(i915) && !i915_gem_object_is_lmem(obj))
+		return ERR_PTR(-EINVAL);
+
+	/*
+	 * The display engine is not coherent with the LLC cache on gen6.  As
+	 * a result, we make sure that the pinning that is about to occur is
+	 * done with uncached PTEs. This is lowest common denominator for all
+	 * chipsets.
+	 *
+	 * However for gen6+, we could do better by using the GFDT bit instead
+	 * of uncaching, which would allow us to flush all the LLC-cached data
+	 * with that bit in the PTE to main memory with just one PIPE_CONTROL.
+	 */
+	ret = i915_gem_object_set_cache_level(obj,
+					      HAS_WT(i915) ?
+					      I915_CACHE_WT : I915_CACHE_NONE);
+	if (ret)
+		return ERR_PTR(ret);
+
+	/*
+	 * As the user may map the buffer once pinned in the display plane
+	 * (e.g. libkms for the bootup splash), we have to ensure that we
+	 * always use map_and_fenceable for all scanout buffers. However,
+	 * it may simply be too big to fit into mappable, in which case
+	 * put it anyway and hope that userspace can cope (but always first
+	 * try to preserve the existing ABI).
+	 */
+	vma = ERR_PTR(-ENOSPC);
+	if ((flags & PIN_MAPPABLE) == 0 &&
+	    (!view || view->type == I915_GTT_VIEW_NORMAL))
+		vma = i915_gem_object_ggtt_pin_ww(obj, ww, view, 0, alignment,
+						  flags | PIN_MAPPABLE |
+						  PIN_NONBLOCK);
+	if (IS_ERR(vma) && vma != ERR_PTR(-EDEADLK))
+		vma = i915_gem_object_ggtt_pin_ww(obj, ww, view, 0,
+						  alignment, flags);
+	if (IS_ERR(vma))
+		return vma;
+
+	vma->display_alignment = max_t(u64, vma->display_alignment, alignment);
+	i915_vma_mark_scanout(vma);
+
+	i915_gem_object_flush_if_display_locked(obj);
+
+	return vma;
+}
+
+/**
+ * Moves a single object to the CPU read, and possibly write domain.
+ * @obj: object to act on
+ * @write: requesting write or read-only access
+ *
+ * This function returns when the move is complete, including waiting on
+ * flushes to occur.
+ */
+int
+i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_object *obj, bool write)
+{
+	int ret;
+
+	assert_object_held(obj);
+
+	ret = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE |
+				   (write ? I915_WAIT_ALL : 0),
+				   MAX_SCHEDULE_TIMEOUT);
+	if (ret)
+		return ret;
+
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
+
+	/* Flush the CPU cache if it's still invalid. */
+	if ((obj->read_domains & I915_GEM_DOMAIN_CPU) == 0) {
+		i915_gem_clflush_object(obj, I915_CLFLUSH_SYNC);
+		obj->read_domains |= I915_GEM_DOMAIN_CPU;
+	}
+
+	/* It should now be out of any other write domains, and we can update
+	 * the domain values for our changes.
+	 */
+	GEM_BUG_ON(obj->write_domain & ~I915_GEM_DOMAIN_CPU);
+
+	/* If we're writing through the CPU, then the GPU read domains will
+	 * need to be invalidated at next use.
+	 */
+	if (write)
+		__start_cpu_write(obj);
+
+	return 0;
+}
+
+/**
+ * Called when user space prepares to use an object with the CPU, either
+ * through the mmap ioctl's mapping or a GTT mapping.
+ * @dev: drm device
+ * @data: ioctl data blob
+ * @file: drm file
+ */
+int
+i915_gem_set_domain_ioctl(struct drm_device *dev, void *data,
+			  struct drm_file *file)
+{
+	struct drm_i915_gem_set_domain *args = data;
+	struct drm_i915_gem_object *obj;
+	u32 read_domains = args->read_domains;
+	u32 write_domain = args->write_domain;
+	int err;
+
+	if (IS_DGFX(to_i915(dev)))
+		return -ENODEV;
+
+	/* Only handle setting domains to types used by the CPU. */
+	if ((write_domain | read_domains) & I915_GEM_GPU_DOMAINS)
+		return -EINVAL;
+
+	/*
+	 * Having something in the write domain implies it's in the read
+	 * domain, and only that read domain.  Enforce that in the request.
+	 */
+	if (write_domain && read_domains != write_domain)
+		return -EINVAL;
+
+	if (!read_domains)
+		return 0;
+
+	obj = i915_gem_object_lookup(file, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	/*
+	 * Try to flush the object off the GPU without holding the lock.
+	 * We will repeat the flush holding the lock in the normal manner
+	 * to catch cases where we are gazumped.
+	 */
+	err = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE |
+				   I915_WAIT_PRIORITY |
+				   (write_domain ? I915_WAIT_ALL : 0),
+				   MAX_SCHEDULE_TIMEOUT);
+	if (err)
+		goto out;
+
+	if (i915_gem_object_is_userptr(obj)) {
+		/*
+		 * Try to grab userptr pages, iris uses set_domain to check
+		 * userptr validity
+		 */
+		err = i915_gem_object_userptr_validate(obj);
+		if (!err)
+			err = i915_gem_object_wait(obj,
+						   I915_WAIT_INTERRUPTIBLE |
+						   I915_WAIT_PRIORITY |
+						   (write_domain ? I915_WAIT_ALL : 0),
+						   MAX_SCHEDULE_TIMEOUT);
+		goto out;
+	}
+
+	/*
+	 * Proxy objects do not control access to the backing storage, ergo
+	 * they cannot be used as a means to manipulate the cache domain
+	 * tracking for that backing storage. The proxy object is always
+	 * considered to be outside of any cache domain.
+	 */
+	if (i915_gem_object_is_proxy(obj)) {
+		err = -ENXIO;
+		goto out;
+	}
+
+	err = i915_gem_object_lock_interruptible(obj, NULL);
+	if (err)
+		goto out;
+
+	/*
+	 * Flush and acquire obj->pages so that we are coherent through
+	 * direct access in memory with previous cached writes through
+	 * shmemfs and that our cache domain tracking remains valid.
+	 * For example, if the obj->filp was moved to swap without us
+	 * being notified and releasing the pages, we would mistakenly
+	 * continue to assume that the obj remained out of the CPU cached
+	 * domain.
+	 */
+	err = i915_gem_object_pin_pages(obj);
+	if (err)
+		goto out_unlock;
+
+	/*
+	 * Already in the desired write domain? Nothing for us to do!
+	 *
+	 * We apply a little bit of cunning here to catch a broader set of
+	 * no-ops. If obj->write_domain is set, we must be in the same
+	 * obj->read_domains, and only that domain. Therefore, if that
+	 * obj->write_domain matches the request read_domains, we are
+	 * already in the same read/write domain and can skip the operation,
+	 * without having to further check the requested write_domain.
+	 */
+	if (READ_ONCE(obj->write_domain) == read_domains)
+		goto out_unpin;
+
+	if (read_domains & I915_GEM_DOMAIN_WC)
+		err = i915_gem_object_set_to_wc_domain(obj, write_domain);
+	else if (read_domains & I915_GEM_DOMAIN_GTT)
+		err = i915_gem_object_set_to_gtt_domain(obj, write_domain);
+	else
+		err = i915_gem_object_set_to_cpu_domain(obj, write_domain);
+
+out_unpin:
+	i915_gem_object_unpin_pages(obj);
+
+out_unlock:
+	i915_gem_object_unlock(obj);
+
+	if (!err && write_domain)
+		i915_gem_object_invalidate_frontbuffer(obj, ORIGIN_CPU);
+
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+/*
+ * Pins the specified object's pages and synchronizes the object with
+ * GPU accesses. Sets needs_clflush to non-zero if the caller should
+ * flush the object from the CPU cache.
+ */
+int i915_gem_object_prepare_read(struct drm_i915_gem_object *obj,
+				 unsigned int *needs_clflush)
+{
+	int ret;
+
+	*needs_clflush = 0;
+	if (!i915_gem_object_has_struct_page(obj))
+		return -ENODEV;
+
+	assert_object_held(obj);
+
+	ret = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE,
+				   MAX_SCHEDULE_TIMEOUT);
+	if (ret)
+		return ret;
+
+	ret = i915_gem_object_pin_pages(obj);
+	if (ret)
+		return ret;
+
+	if (obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ ||
+	    !static_cpu_has(X86_FEATURE_CLFLUSH)) {
+		ret = i915_gem_object_set_to_cpu_domain(obj, false);
+		if (ret)
+			goto err_unpin;
+		else
+			goto out;
+	}
+
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
+
+	/* If we're not in the cpu read domain, set ourself into the gtt
+	 * read domain and manually flush cachelines (if required). This
+	 * optimizes for the case when the gpu will dirty the data
+	 * anyway again before the next pread happens.
+	 */
+	if (!obj->cache_dirty &&
+	    !(obj->read_domains & I915_GEM_DOMAIN_CPU))
+		*needs_clflush = CLFLUSH_BEFORE;
+
+out:
+	/* return with the pages pinned */
+	return 0;
+
+err_unpin:
+	i915_gem_object_unpin_pages(obj);
+	return ret;
+}
+
+int i915_gem_object_prepare_write(struct drm_i915_gem_object *obj,
+				  unsigned int *needs_clflush)
+{
+	int ret;
+
+	*needs_clflush = 0;
+	if (!i915_gem_object_has_struct_page(obj))
+		return -ENODEV;
+
+	assert_object_held(obj);
+
+	ret = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE |
+				   I915_WAIT_ALL,
+				   MAX_SCHEDULE_TIMEOUT);
+	if (ret)
+		return ret;
+
+	ret = i915_gem_object_pin_pages(obj);
+	if (ret)
+		return ret;
+
+	if (obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_WRITE ||
+	    !static_cpu_has(X86_FEATURE_CLFLUSH)) {
+		ret = i915_gem_object_set_to_cpu_domain(obj, true);
+		if (ret)
+			goto err_unpin;
+		else
+			goto out;
+	}
+
+	flush_write_domain(obj, ~I915_GEM_DOMAIN_CPU);
+
+	/* If we're not in the cpu write domain, set ourself into the
+	 * gtt write domain and manually flush cachelines (as required).
+	 * This optimizes for the case when the gpu will use the data
+	 * right away and we therefore have to clflush anyway.
+	 */
+	if (!obj->cache_dirty) {
+		*needs_clflush |= CLFLUSH_AFTER;
+
+		/*
+		 * Same trick applies to invalidate partially written
+		 * cachelines read before writing.
+		 */
+		if (!(obj->read_domains & I915_GEM_DOMAIN_CPU))
+			*needs_clflush |= CLFLUSH_BEFORE;
+	}
+
+out:
+	i915_gem_object_invalidate_frontbuffer(obj, ORIGIN_CPU);
+	obj->mm.dirty = true;
+	/* return with the pages pinned */
+	return 0;
+
+err_unpin:
+	i915_gem_object_unpin_pages(obj);
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_domain.h b/drivers/gpu/drm/i915/gem/i915_gem_domain.h
new file mode 100644
index 000000000..9622df962
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_domain.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __I915_GEM_DOMAIN_H__
+#define __I915_GEM_DOMAIN_H__
+
+struct drm_i915_gem_object;
+enum i915_cache_level;
+
+int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
+				    enum i915_cache_level cache_level);
+
+#endif /* __I915_GEM_DOMAIN_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
new file mode 100644
index 000000000..0a123bb44
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
@@ -0,0 +1,3623 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2008,2010 Intel Corporation
+ */
+
+#include <linux/dma-resv.h>
+#include <linux/highmem.h>
+#include <linux/sync_file.h>
+#include <linux/uaccess.h>
+
+#include <drm/drm_syncobj.h>
+
+#include "display/intel_frontbuffer.h"
+
+#include "gem/i915_gem_ioctls.h"
+#include "gt/intel_context.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_buffer_pool.h"
+#include "gt/intel_gt_pm.h"
+#include "gt/intel_ring.h"
+
+#include "pxp/intel_pxp.h"
+
+#include "i915_cmd_parser.h"
+#include "i915_drv.h"
+#include "i915_file_private.h"
+#include "i915_gem_clflush.h"
+#include "i915_gem_context.h"
+#include "i915_gem_evict.h"
+#include "i915_gem_ioctls.h"
+#include "i915_trace.h"
+#include "i915_user_extensions.h"
+
+struct eb_vma {
+	struct i915_vma *vma;
+	unsigned int flags;
+
+	/** This vma's place in the execbuf reservation list */
+	struct drm_i915_gem_exec_object2 *exec;
+	struct list_head bind_link;
+	struct list_head reloc_link;
+
+	struct hlist_node node;
+	u32 handle;
+};
+
+enum {
+	FORCE_CPU_RELOC = 1,
+	FORCE_GTT_RELOC,
+	FORCE_GPU_RELOC,
+#define DBG_FORCE_RELOC 0 /* choose one of the above! */
+};
+
+/* __EXEC_OBJECT_NO_RESERVE is BIT(31), defined in i915_vma.h */
+#define __EXEC_OBJECT_HAS_PIN		BIT(30)
+#define __EXEC_OBJECT_HAS_FENCE		BIT(29)
+#define __EXEC_OBJECT_USERPTR_INIT	BIT(28)
+#define __EXEC_OBJECT_NEEDS_MAP		BIT(27)
+#define __EXEC_OBJECT_NEEDS_BIAS	BIT(26)
+#define __EXEC_OBJECT_INTERNAL_FLAGS	(~0u << 26) /* all of the above + */
+#define __EXEC_OBJECT_RESERVED (__EXEC_OBJECT_HAS_PIN | __EXEC_OBJECT_HAS_FENCE)
+
+#define __EXEC_HAS_RELOC	BIT(31)
+#define __EXEC_ENGINE_PINNED	BIT(30)
+#define __EXEC_USERPTR_USED	BIT(29)
+#define __EXEC_INTERNAL_FLAGS	(~0u << 29)
+#define UPDATE			PIN_OFFSET_FIXED
+
+#define BATCH_OFFSET_BIAS (256*1024)
+
+#define __I915_EXEC_ILLEGAL_FLAGS \
+	(__I915_EXEC_UNKNOWN_FLAGS | \
+	 I915_EXEC_CONSTANTS_MASK  | \
+	 I915_EXEC_RESOURCE_STREAMER)
+
+/* Catch emission of unexpected errors for CI! */
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)
+#undef EINVAL
+#define EINVAL ({ \
+	DRM_DEBUG_DRIVER("EINVAL at %s:%d\n", __func__, __LINE__); \
+	22; \
+})
+#endif
+
+/**
+ * DOC: User command execution
+ *
+ * Userspace submits commands to be executed on the GPU as an instruction
+ * stream within a GEM object we call a batchbuffer. This instructions may
+ * refer to other GEM objects containing auxiliary state such as kernels,
+ * samplers, render targets and even secondary batchbuffers. Userspace does
+ * not know where in the GPU memory these objects reside and so before the
+ * batchbuffer is passed to the GPU for execution, those addresses in the
+ * batchbuffer and auxiliary objects are updated. This is known as relocation,
+ * or patching. To try and avoid having to relocate each object on the next
+ * execution, userspace is told the location of those objects in this pass,
+ * but this remains just a hint as the kernel may choose a new location for
+ * any object in the future.
+ *
+ * At the level of talking to the hardware, submitting a batchbuffer for the
+ * GPU to execute is to add content to a buffer from which the HW
+ * command streamer is reading.
+ *
+ * 1. Add a command to load the HW context. For Logical Ring Contexts, i.e.
+ *    Execlists, this command is not placed on the same buffer as the
+ *    remaining items.
+ *
+ * 2. Add a command to invalidate caches to the buffer.
+ *
+ * 3. Add a batchbuffer start command to the buffer; the start command is
+ *    essentially a token together with the GPU address of the batchbuffer
+ *    to be executed.
+ *
+ * 4. Add a pipeline flush to the buffer.
+ *
+ * 5. Add a memory write command to the buffer to record when the GPU
+ *    is done executing the batchbuffer. The memory write writes the
+ *    global sequence number of the request, ``i915_request::global_seqno``;
+ *    the i915 driver uses the current value in the register to determine
+ *    if the GPU has completed the batchbuffer.
+ *
+ * 6. Add a user interrupt command to the buffer. This command instructs
+ *    the GPU to issue an interrupt when the command, pipeline flush and
+ *    memory write are completed.
+ *
+ * 7. Inform the hardware of the additional commands added to the buffer
+ *    (by updating the tail pointer).
+ *
+ * Processing an execbuf ioctl is conceptually split up into a few phases.
+ *
+ * 1. Validation - Ensure all the pointers, handles and flags are valid.
+ * 2. Reservation - Assign GPU address space for every object
+ * 3. Relocation - Update any addresses to point to the final locations
+ * 4. Serialisation - Order the request with respect to its dependencies
+ * 5. Construction - Construct a request to execute the batchbuffer
+ * 6. Submission (at some point in the future execution)
+ *
+ * Reserving resources for the execbuf is the most complicated phase. We
+ * neither want to have to migrate the object in the address space, nor do
+ * we want to have to update any relocations pointing to this object. Ideally,
+ * we want to leave the object where it is and for all the existing relocations
+ * to match. If the object is given a new address, or if userspace thinks the
+ * object is elsewhere, we have to parse all the relocation entries and update
+ * the addresses. Userspace can set the I915_EXEC_NORELOC flag to hint that
+ * all the target addresses in all of its objects match the value in the
+ * relocation entries and that they all match the presumed offsets given by the
+ * list of execbuffer objects. Using this knowledge, we know that if we haven't
+ * moved any buffers, all the relocation entries are valid and we can skip
+ * the update. (If userspace is wrong, the likely outcome is an impromptu GPU
+ * hang.) The requirement for using I915_EXEC_NO_RELOC are:
+ *
+ *      The addresses written in the objects must match the corresponding
+ *      reloc.presumed_offset which in turn must match the corresponding
+ *      execobject.offset.
+ *
+ *      Any render targets written to in the batch must be flagged with
+ *      EXEC_OBJECT_WRITE.
+ *
+ *      To avoid stalling, execobject.offset should match the current
+ *      address of that object within the active context.
+ *
+ * The reservation is done is multiple phases. First we try and keep any
+ * object already bound in its current location - so as long as meets the
+ * constraints imposed by the new execbuffer. Any object left unbound after the
+ * first pass is then fitted into any available idle space. If an object does
+ * not fit, all objects are removed from the reservation and the process rerun
+ * after sorting the objects into a priority order (more difficult to fit
+ * objects are tried first). Failing that, the entire VM is cleared and we try
+ * to fit the execbuf once last time before concluding that it simply will not
+ * fit.
+ *
+ * A small complication to all of this is that we allow userspace not only to
+ * specify an alignment and a size for the object in the address space, but
+ * we also allow userspace to specify the exact offset. This objects are
+ * simpler to place (the location is known a priori) all we have to do is make
+ * sure the space is available.
+ *
+ * Once all the objects are in place, patching up the buried pointers to point
+ * to the final locations is a fairly simple job of walking over the relocation
+ * entry arrays, looking up the right address and rewriting the value into
+ * the object. Simple! ... The relocation entries are stored in user memory
+ * and so to access them we have to copy them into a local buffer. That copy
+ * has to avoid taking any pagefaults as they may lead back to a GEM object
+ * requiring the struct_mutex (i.e. recursive deadlock). So once again we split
+ * the relocation into multiple passes. First we try to do everything within an
+ * atomic context (avoid the pagefaults) which requires that we never wait. If
+ * we detect that we may wait, or if we need to fault, then we have to fallback
+ * to a slower path. The slowpath has to drop the mutex. (Can you hear alarm
+ * bells yet?) Dropping the mutex means that we lose all the state we have
+ * built up so far for the execbuf and we must reset any global data. However,
+ * we do leave the objects pinned in their final locations - which is a
+ * potential issue for concurrent execbufs. Once we have left the mutex, we can
+ * allocate and copy all the relocation entries into a large array at our
+ * leisure, reacquire the mutex, reclaim all the objects and other state and
+ * then proceed to update any incorrect addresses with the objects.
+ *
+ * As we process the relocation entries, we maintain a record of whether the
+ * object is being written to. Using NORELOC, we expect userspace to provide
+ * this information instead. We also check whether we can skip the relocation
+ * by comparing the expected value inside the relocation entry with the target's
+ * final address. If they differ, we have to map the current object and rewrite
+ * the 4 or 8 byte pointer within.
+ *
+ * Serialising an execbuf is quite simple according to the rules of the GEM
+ * ABI. Execution within each context is ordered by the order of submission.
+ * Writes to any GEM object are in order of submission and are exclusive. Reads
+ * from a GEM object are unordered with respect to other reads, but ordered by
+ * writes. A write submitted after a read cannot occur before the read, and
+ * similarly any read submitted after a write cannot occur before the write.
+ * Writes are ordered between engines such that only one write occurs at any
+ * time (completing any reads beforehand) - using semaphores where available
+ * and CPU serialisation otherwise. Other GEM access obey the same rules, any
+ * write (either via mmaps using set-domain, or via pwrite) must flush all GPU
+ * reads before starting, and any read (either using set-domain or pread) must
+ * flush all GPU writes before starting. (Note we only employ a barrier before,
+ * we currently rely on userspace not concurrently starting a new execution
+ * whilst reading or writing to an object. This may be an advantage or not
+ * depending on how much you trust userspace not to shoot themselves in the
+ * foot.) Serialisation may just result in the request being inserted into
+ * a DAG awaiting its turn, but most simple is to wait on the CPU until
+ * all dependencies are resolved.
+ *
+ * After all of that, is just a matter of closing the request and handing it to
+ * the hardware (well, leaving it in a queue to be executed). However, we also
+ * offer the ability for batchbuffers to be run with elevated privileges so
+ * that they access otherwise hidden registers. (Used to adjust L3 cache etc.)
+ * Before any batch is given extra privileges we first must check that it
+ * contains no nefarious instructions, we check that each instruction is from
+ * our whitelist and all registers are also from an allowed list. We first
+ * copy the user's batchbuffer to a shadow (so that the user doesn't have
+ * access to it, either by the CPU or GPU as we scan it) and then parse each
+ * instruction. If everything is ok, we set a flag telling the hardware to run
+ * the batchbuffer in trusted mode, otherwise the ioctl is rejected.
+ */
+
+struct eb_fence {
+	struct drm_syncobj *syncobj; /* Use with ptr_mask_bits() */
+	struct dma_fence *dma_fence;
+	u64 value;
+	struct dma_fence_chain *chain_fence;
+};
+
+struct i915_execbuffer {
+	struct drm_i915_private *i915; /** i915 backpointer */
+	struct drm_file *file; /** per-file lookup tables and limits */
+	struct drm_i915_gem_execbuffer2 *args; /** ioctl parameters */
+	struct drm_i915_gem_exec_object2 *exec; /** ioctl execobj[] */
+	struct eb_vma *vma;
+
+	struct intel_gt *gt; /* gt for the execbuf */
+	struct intel_context *context; /* logical state for the request */
+	struct i915_gem_context *gem_context; /** caller's context */
+
+	/** our requests to build */
+	struct i915_request *requests[MAX_ENGINE_INSTANCE + 1];
+	/** identity of the batch obj/vma */
+	struct eb_vma *batches[MAX_ENGINE_INSTANCE + 1];
+	struct i915_vma *trampoline; /** trampoline used for chaining */
+
+	/** used for excl fence in dma_resv objects when > 1 BB submitted */
+	struct dma_fence *composite_fence;
+
+	/** actual size of execobj[] as we may extend it for the cmdparser */
+	unsigned int buffer_count;
+
+	/* number of batches in execbuf IOCTL */
+	unsigned int num_batches;
+
+	/** list of vma not yet bound during reservation phase */
+	struct list_head unbound;
+
+	/** list of vma that have execobj.relocation_count */
+	struct list_head relocs;
+
+	struct i915_gem_ww_ctx ww;
+
+	/**
+	 * Track the most recently used object for relocations, as we
+	 * frequently have to perform multiple relocations within the same
+	 * obj/page
+	 */
+	struct reloc_cache {
+		struct drm_mm_node node; /** temporary GTT binding */
+		unsigned long vaddr; /** Current kmap address */
+		unsigned long page; /** Currently mapped page index */
+		unsigned int graphics_ver; /** Cached value of GRAPHICS_VER */
+		bool use_64bit_reloc : 1;
+		bool has_llc : 1;
+		bool has_fence : 1;
+		bool needs_unfenced : 1;
+	} reloc_cache;
+
+	u64 invalid_flags; /** Set of execobj.flags that are invalid */
+
+	/** Length of batch within object */
+	u64 batch_len[MAX_ENGINE_INSTANCE + 1];
+	u32 batch_start_offset; /** Location within object of batch */
+	u32 batch_flags; /** Flags composed for emit_bb_start() */
+	struct intel_gt_buffer_pool_node *batch_pool; /** pool node for batch buffer */
+
+	/**
+	 * Indicate either the size of the hastable used to resolve
+	 * relocation handles, or if negative that we are using a direct
+	 * index into the execobj[].
+	 */
+	int lut_size;
+	struct hlist_head *buckets; /** ht for relocation handles */
+
+	struct eb_fence *fences;
+	unsigned long num_fences;
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+	struct i915_capture_list *capture_lists[MAX_ENGINE_INSTANCE + 1];
+#endif
+};
+
+static int eb_parse(struct i915_execbuffer *eb);
+static int eb_pin_engine(struct i915_execbuffer *eb, bool throttle);
+static void eb_unpin_engine(struct i915_execbuffer *eb);
+static void eb_capture_release(struct i915_execbuffer *eb);
+
+static inline bool eb_use_cmdparser(const struct i915_execbuffer *eb)
+{
+	return intel_engine_requires_cmd_parser(eb->context->engine) ||
+		(intel_engine_using_cmd_parser(eb->context->engine) &&
+		 eb->args->batch_len);
+}
+
+static int eb_create(struct i915_execbuffer *eb)
+{
+	if (!(eb->args->flags & I915_EXEC_HANDLE_LUT)) {
+		unsigned int size = 1 + ilog2(eb->buffer_count);
+
+		/*
+		 * Without a 1:1 association between relocation handles and
+		 * the execobject[] index, we instead create a hashtable.
+		 * We size it dynamically based on available memory, starting
+		 * first with 1:1 assocative hash and scaling back until
+		 * the allocation succeeds.
+		 *
+		 * Later on we use a positive lut_size to indicate we are
+		 * using this hashtable, and a negative value to indicate a
+		 * direct lookup.
+		 */
+		do {
+			gfp_t flags;
+
+			/* While we can still reduce the allocation size, don't
+			 * raise a warning and allow the allocation to fail.
+			 * On the last pass though, we want to try as hard
+			 * as possible to perform the allocation and warn
+			 * if it fails.
+			 */
+			flags = GFP_KERNEL;
+			if (size > 1)
+				flags |= __GFP_NORETRY | __GFP_NOWARN;
+
+			eb->buckets = kzalloc(sizeof(struct hlist_head) << size,
+					      flags);
+			if (eb->buckets)
+				break;
+		} while (--size);
+
+		if (unlikely(!size))
+			return -ENOMEM;
+
+		eb->lut_size = size;
+	} else {
+		eb->lut_size = -eb->buffer_count;
+	}
+
+	return 0;
+}
+
+static bool
+eb_vma_misplaced(const struct drm_i915_gem_exec_object2 *entry,
+		 const struct i915_vma *vma,
+		 unsigned int flags)
+{
+	if (vma->node.size < entry->pad_to_size)
+		return true;
+
+	if (entry->alignment && !IS_ALIGNED(vma->node.start, entry->alignment))
+		return true;
+
+	if (flags & EXEC_OBJECT_PINNED &&
+	    vma->node.start != entry->offset)
+		return true;
+
+	if (flags & __EXEC_OBJECT_NEEDS_BIAS &&
+	    vma->node.start < BATCH_OFFSET_BIAS)
+		return true;
+
+	if (!(flags & EXEC_OBJECT_SUPPORTS_48B_ADDRESS) &&
+	    (vma->node.start + vma->node.size + 4095) >> 32)
+		return true;
+
+	if (flags & __EXEC_OBJECT_NEEDS_MAP &&
+	    !i915_vma_is_map_and_fenceable(vma))
+		return true;
+
+	return false;
+}
+
+static u64 eb_pin_flags(const struct drm_i915_gem_exec_object2 *entry,
+			unsigned int exec_flags)
+{
+	u64 pin_flags = 0;
+
+	if (exec_flags & EXEC_OBJECT_NEEDS_GTT)
+		pin_flags |= PIN_GLOBAL;
+
+	/*
+	 * Wa32bitGeneralStateOffset & Wa32bitInstructionBaseOffset,
+	 * limit address to the first 4GBs for unflagged objects.
+	 */
+	if (!(exec_flags & EXEC_OBJECT_SUPPORTS_48B_ADDRESS))
+		pin_flags |= PIN_ZONE_4G;
+
+	if (exec_flags & __EXEC_OBJECT_NEEDS_MAP)
+		pin_flags |= PIN_MAPPABLE;
+
+	if (exec_flags & EXEC_OBJECT_PINNED)
+		pin_flags |= entry->offset | PIN_OFFSET_FIXED;
+	else if (exec_flags & __EXEC_OBJECT_NEEDS_BIAS)
+		pin_flags |= BATCH_OFFSET_BIAS | PIN_OFFSET_BIAS;
+
+	return pin_flags;
+}
+
+static inline int
+eb_pin_vma(struct i915_execbuffer *eb,
+	   const struct drm_i915_gem_exec_object2 *entry,
+	   struct eb_vma *ev)
+{
+	struct i915_vma *vma = ev->vma;
+	u64 pin_flags;
+	int err;
+
+	if (vma->node.size)
+		pin_flags = vma->node.start;
+	else
+		pin_flags = entry->offset & PIN_OFFSET_MASK;
+
+	pin_flags |= PIN_USER | PIN_NOEVICT | PIN_OFFSET_FIXED | PIN_VALIDATE;
+	if (unlikely(ev->flags & EXEC_OBJECT_NEEDS_GTT))
+		pin_flags |= PIN_GLOBAL;
+
+	/* Attempt to reuse the current location if available */
+	err = i915_vma_pin_ww(vma, &eb->ww, 0, 0, pin_flags);
+	if (err == -EDEADLK)
+		return err;
+
+	if (unlikely(err)) {
+		if (entry->flags & EXEC_OBJECT_PINNED)
+			return err;
+
+		/* Failing that pick any _free_ space if suitable */
+		err = i915_vma_pin_ww(vma, &eb->ww,
+					     entry->pad_to_size,
+					     entry->alignment,
+					     eb_pin_flags(entry, ev->flags) |
+					     PIN_USER | PIN_NOEVICT | PIN_VALIDATE);
+		if (unlikely(err))
+			return err;
+	}
+
+	if (unlikely(ev->flags & EXEC_OBJECT_NEEDS_FENCE)) {
+		err = i915_vma_pin_fence(vma);
+		if (unlikely(err))
+			return err;
+
+		if (vma->fence)
+			ev->flags |= __EXEC_OBJECT_HAS_FENCE;
+	}
+
+	ev->flags |= __EXEC_OBJECT_HAS_PIN;
+	if (eb_vma_misplaced(entry, vma, ev->flags))
+		return -EBADSLT;
+
+	return 0;
+}
+
+static inline void
+eb_unreserve_vma(struct eb_vma *ev)
+{
+	if (unlikely(ev->flags & __EXEC_OBJECT_HAS_FENCE))
+		__i915_vma_unpin_fence(ev->vma);
+
+	ev->flags &= ~__EXEC_OBJECT_RESERVED;
+}
+
+static int
+eb_validate_vma(struct i915_execbuffer *eb,
+		struct drm_i915_gem_exec_object2 *entry,
+		struct i915_vma *vma)
+{
+	/* Relocations are disallowed for all platforms after TGL-LP.  This
+	 * also covers all platforms with local memory.
+	 */
+	if (entry->relocation_count &&
+	    GRAPHICS_VER(eb->i915) >= 12 && !IS_TIGERLAKE(eb->i915))
+		return -EINVAL;
+
+	if (unlikely(entry->flags & eb->invalid_flags))
+		return -EINVAL;
+
+	if (unlikely(entry->alignment &&
+		     !is_power_of_2_u64(entry->alignment)))
+		return -EINVAL;
+
+	/*
+	 * Offset can be used as input (EXEC_OBJECT_PINNED), reject
+	 * any non-page-aligned or non-canonical addresses.
+	 */
+	if (unlikely(entry->flags & EXEC_OBJECT_PINNED &&
+		     entry->offset != gen8_canonical_addr(entry->offset & I915_GTT_PAGE_MASK)))
+		return -EINVAL;
+
+	/* pad_to_size was once a reserved field, so sanitize it */
+	if (entry->flags & EXEC_OBJECT_PAD_TO_SIZE) {
+		if (unlikely(offset_in_page(entry->pad_to_size)))
+			return -EINVAL;
+	} else {
+		entry->pad_to_size = 0;
+	}
+	/*
+	 * From drm_mm perspective address space is continuous,
+	 * so from this point we're always using non-canonical
+	 * form internally.
+	 */
+	entry->offset = gen8_noncanonical_addr(entry->offset);
+
+	if (!eb->reloc_cache.has_fence) {
+		entry->flags &= ~EXEC_OBJECT_NEEDS_FENCE;
+	} else {
+		if ((entry->flags & EXEC_OBJECT_NEEDS_FENCE ||
+		     eb->reloc_cache.needs_unfenced) &&
+		    i915_gem_object_is_tiled(vma->obj))
+			entry->flags |= EXEC_OBJECT_NEEDS_GTT | __EXEC_OBJECT_NEEDS_MAP;
+	}
+
+	return 0;
+}
+
+static inline bool
+is_batch_buffer(struct i915_execbuffer *eb, unsigned int buffer_idx)
+{
+	return eb->args->flags & I915_EXEC_BATCH_FIRST ?
+		buffer_idx < eb->num_batches :
+		buffer_idx >= eb->args->buffer_count - eb->num_batches;
+}
+
+static int
+eb_add_vma(struct i915_execbuffer *eb,
+	   unsigned int *current_batch,
+	   unsigned int i,
+	   struct i915_vma *vma)
+{
+	struct drm_i915_private *i915 = eb->i915;
+	struct drm_i915_gem_exec_object2 *entry = &eb->exec[i];
+	struct eb_vma *ev = &eb->vma[i];
+
+	ev->vma = vma;
+	ev->exec = entry;
+	ev->flags = entry->flags;
+
+	if (eb->lut_size > 0) {
+		ev->handle = entry->handle;
+		hlist_add_head(&ev->node,
+			       &eb->buckets[hash_32(entry->handle,
+						    eb->lut_size)]);
+	}
+
+	if (entry->relocation_count)
+		list_add_tail(&ev->reloc_link, &eb->relocs);
+
+	/*
+	 * SNA is doing fancy tricks with compressing batch buffers, which leads
+	 * to negative relocation deltas. Usually that works out ok since the
+	 * relocate address is still positive, except when the batch is placed
+	 * very low in the GTT. Ensure this doesn't happen.
+	 *
+	 * Note that actual hangs have only been observed on gen7, but for
+	 * paranoia do it everywhere.
+	 */
+	if (is_batch_buffer(eb, i)) {
+		if (entry->relocation_count &&
+		    !(ev->flags & EXEC_OBJECT_PINNED))
+			ev->flags |= __EXEC_OBJECT_NEEDS_BIAS;
+		if (eb->reloc_cache.has_fence)
+			ev->flags |= EXEC_OBJECT_NEEDS_FENCE;
+
+		eb->batches[*current_batch] = ev;
+
+		if (unlikely(ev->flags & EXEC_OBJECT_WRITE)) {
+			drm_dbg(&i915->drm,
+				"Attempting to use self-modifying batch buffer\n");
+			return -EINVAL;
+		}
+
+		if (range_overflows_t(u64,
+				      eb->batch_start_offset,
+				      eb->args->batch_len,
+				      ev->vma->size)) {
+			drm_dbg(&i915->drm, "Attempting to use out-of-bounds batch\n");
+			return -EINVAL;
+		}
+
+		if (eb->args->batch_len == 0)
+			eb->batch_len[*current_batch] = ev->vma->size -
+				eb->batch_start_offset;
+		else
+			eb->batch_len[*current_batch] = eb->args->batch_len;
+		if (unlikely(eb->batch_len[*current_batch] == 0)) { /* impossible! */
+			drm_dbg(&i915->drm, "Invalid batch length\n");
+			return -EINVAL;
+		}
+
+		++*current_batch;
+	}
+
+	return 0;
+}
+
+static inline int use_cpu_reloc(const struct reloc_cache *cache,
+				const struct drm_i915_gem_object *obj)
+{
+	if (!i915_gem_object_has_struct_page(obj))
+		return false;
+
+	if (DBG_FORCE_RELOC == FORCE_CPU_RELOC)
+		return true;
+
+	if (DBG_FORCE_RELOC == FORCE_GTT_RELOC)
+		return false;
+
+	return (cache->has_llc ||
+		obj->cache_dirty ||
+		obj->cache_level != I915_CACHE_NONE);
+}
+
+static int eb_reserve_vma(struct i915_execbuffer *eb,
+			  struct eb_vma *ev,
+			  u64 pin_flags)
+{
+	struct drm_i915_gem_exec_object2 *entry = ev->exec;
+	struct i915_vma *vma = ev->vma;
+	int err;
+
+	if (drm_mm_node_allocated(&vma->node) &&
+	    eb_vma_misplaced(entry, vma, ev->flags)) {
+		err = i915_vma_unbind(vma);
+		if (err)
+			return err;
+	}
+
+	err = i915_vma_pin_ww(vma, &eb->ww,
+			   entry->pad_to_size, entry->alignment,
+			   eb_pin_flags(entry, ev->flags) | pin_flags);
+	if (err)
+		return err;
+
+	if (entry->offset != vma->node.start) {
+		entry->offset = vma->node.start | UPDATE;
+		eb->args->flags |= __EXEC_HAS_RELOC;
+	}
+
+	if (unlikely(ev->flags & EXEC_OBJECT_NEEDS_FENCE)) {
+		err = i915_vma_pin_fence(vma);
+		if (unlikely(err))
+			return err;
+
+		if (vma->fence)
+			ev->flags |= __EXEC_OBJECT_HAS_FENCE;
+	}
+
+	ev->flags |= __EXEC_OBJECT_HAS_PIN;
+	GEM_BUG_ON(eb_vma_misplaced(entry, vma, ev->flags));
+
+	return 0;
+}
+
+static bool eb_unbind(struct i915_execbuffer *eb, bool force)
+{
+	const unsigned int count = eb->buffer_count;
+	unsigned int i;
+	struct list_head last;
+	bool unpinned = false;
+
+	/* Resort *all* the objects into priority order */
+	INIT_LIST_HEAD(&eb->unbound);
+	INIT_LIST_HEAD(&last);
+
+	for (i = 0; i < count; i++) {
+		struct eb_vma *ev = &eb->vma[i];
+		unsigned int flags = ev->flags;
+
+		if (!force && flags & EXEC_OBJECT_PINNED &&
+		    flags & __EXEC_OBJECT_HAS_PIN)
+			continue;
+
+		unpinned = true;
+		eb_unreserve_vma(ev);
+
+		if (flags & EXEC_OBJECT_PINNED)
+			/* Pinned must have their slot */
+			list_add(&ev->bind_link, &eb->unbound);
+		else if (flags & __EXEC_OBJECT_NEEDS_MAP)
+			/* Map require the lowest 256MiB (aperture) */
+			list_add_tail(&ev->bind_link, &eb->unbound);
+		else if (!(flags & EXEC_OBJECT_SUPPORTS_48B_ADDRESS))
+			/* Prioritise 4GiB region for restricted bo */
+			list_add(&ev->bind_link, &last);
+		else
+			list_add_tail(&ev->bind_link, &last);
+	}
+
+	list_splice_tail(&last, &eb->unbound);
+	return unpinned;
+}
+
+static int eb_reserve(struct i915_execbuffer *eb)
+{
+	struct eb_vma *ev;
+	unsigned int pass;
+	int err = 0;
+	bool unpinned;
+
+	/*
+	 * We have one more buffers that we couldn't bind, which could be due to
+	 * various reasons. To resolve this we have 4 passes, with every next
+	 * level turning the screws tighter:
+	 *
+	 * 0. Unbind all objects that do not match the GTT constraints for the
+	 * execbuffer (fenceable, mappable, alignment etc). Bind all new
+	 * objects.  This avoids unnecessary unbinding of later objects in order
+	 * to make room for the earlier objects *unless* we need to defragment.
+	 *
+	 * 1. Reorder the buffers, where objects with the most restrictive
+	 * placement requirements go first (ignoring fixed location buffers for
+	 * now).  For example, objects needing the mappable aperture (the first
+	 * 256M of GTT), should go first vs objects that can be placed just
+	 * about anywhere. Repeat the previous pass.
+	 *
+	 * 2. Consider buffers that are pinned at a fixed location. Also try to
+	 * evict the entire VM this time, leaving only objects that we were
+	 * unable to lock. Try again to bind the buffers. (still using the new
+	 * buffer order).
+	 *
+	 * 3. We likely have object lock contention for one or more stubborn
+	 * objects in the VM, for which we need to evict to make forward
+	 * progress (perhaps we are fighting the shrinker?). When evicting the
+	 * VM this time around, anything that we can't lock we now track using
+	 * the busy_bo, using the full lock (after dropping the vm->mutex to
+	 * prevent deadlocks), instead of trylock. We then continue to evict the
+	 * VM, this time with the stubborn object locked, which we can now
+	 * hopefully unbind (if still bound in the VM). Repeat until the VM is
+	 * evicted. Finally we should be able bind everything.
+	 */
+	for (pass = 0; pass <= 3; pass++) {
+		int pin_flags = PIN_USER | PIN_VALIDATE;
+
+		if (pass == 0)
+			pin_flags |= PIN_NONBLOCK;
+
+		if (pass >= 1)
+			unpinned = eb_unbind(eb, pass >= 2);
+
+		if (pass == 2) {
+			err = mutex_lock_interruptible(&eb->context->vm->mutex);
+			if (!err) {
+				err = i915_gem_evict_vm(eb->context->vm, &eb->ww, NULL);
+				mutex_unlock(&eb->context->vm->mutex);
+			}
+			if (err)
+				return err;
+		}
+
+		if (pass == 3) {
+retry:
+			err = mutex_lock_interruptible(&eb->context->vm->mutex);
+			if (!err) {
+				struct drm_i915_gem_object *busy_bo = NULL;
+
+				err = i915_gem_evict_vm(eb->context->vm, &eb->ww, &busy_bo);
+				mutex_unlock(&eb->context->vm->mutex);
+				if (err && busy_bo) {
+					err = i915_gem_object_lock(busy_bo, &eb->ww);
+					i915_gem_object_put(busy_bo);
+					if (!err)
+						goto retry;
+				}
+			}
+			if (err)
+				return err;
+		}
+
+		list_for_each_entry(ev, &eb->unbound, bind_link) {
+			err = eb_reserve_vma(eb, ev, pin_flags);
+			if (err)
+				break;
+		}
+
+		if (err != -ENOSPC)
+			break;
+	}
+
+	return err;
+}
+
+static int eb_select_context(struct i915_execbuffer *eb)
+{
+	struct i915_gem_context *ctx;
+
+	ctx = i915_gem_context_lookup(eb->file->driver_priv, eb->args->rsvd1);
+	if (unlikely(IS_ERR(ctx)))
+		return PTR_ERR(ctx);
+
+	eb->gem_context = ctx;
+	if (i915_gem_context_has_full_ppgtt(ctx))
+		eb->invalid_flags |= EXEC_OBJECT_NEEDS_GTT;
+
+	return 0;
+}
+
+static int __eb_add_lut(struct i915_execbuffer *eb,
+			u32 handle, struct i915_vma *vma)
+{
+	struct i915_gem_context *ctx = eb->gem_context;
+	struct i915_lut_handle *lut;
+	int err;
+
+	lut = i915_lut_handle_alloc();
+	if (unlikely(!lut))
+		return -ENOMEM;
+
+	i915_vma_get(vma);
+	if (!atomic_fetch_inc(&vma->open_count))
+		i915_vma_reopen(vma);
+	lut->handle = handle;
+	lut->ctx = ctx;
+
+	/* Check that the context hasn't been closed in the meantime */
+	err = -EINTR;
+	if (!mutex_lock_interruptible(&ctx->lut_mutex)) {
+		if (likely(!i915_gem_context_is_closed(ctx)))
+			err = radix_tree_insert(&ctx->handles_vma, handle, vma);
+		else
+			err = -ENOENT;
+		if (err == 0) { /* And nor has this handle */
+			struct drm_i915_gem_object *obj = vma->obj;
+
+			spin_lock(&obj->lut_lock);
+			if (idr_find(&eb->file->object_idr, handle) == obj) {
+				list_add(&lut->obj_link, &obj->lut_list);
+			} else {
+				radix_tree_delete(&ctx->handles_vma, handle);
+				err = -ENOENT;
+			}
+			spin_unlock(&obj->lut_lock);
+		}
+		mutex_unlock(&ctx->lut_mutex);
+	}
+	if (unlikely(err))
+		goto err;
+
+	return 0;
+
+err:
+	i915_vma_close(vma);
+	i915_vma_put(vma);
+	i915_lut_handle_free(lut);
+	return err;
+}
+
+static struct i915_vma *eb_lookup_vma(struct i915_execbuffer *eb, u32 handle)
+{
+	struct i915_address_space *vm = eb->context->vm;
+
+	do {
+		struct drm_i915_gem_object *obj;
+		struct i915_vma *vma;
+		int err;
+
+		rcu_read_lock();
+		vma = radix_tree_lookup(&eb->gem_context->handles_vma, handle);
+		if (likely(vma && vma->vm == vm))
+			vma = i915_vma_tryget(vma);
+		rcu_read_unlock();
+		if (likely(vma))
+			return vma;
+
+		obj = i915_gem_object_lookup(eb->file, handle);
+		if (unlikely(!obj))
+			return ERR_PTR(-ENOENT);
+
+		/*
+		 * If the user has opted-in for protected-object tracking, make
+		 * sure the object encryption can be used.
+		 * We only need to do this when the object is first used with
+		 * this context, because the context itself will be banned when
+		 * the protected objects become invalid.
+		 */
+		if (i915_gem_context_uses_protected_content(eb->gem_context) &&
+		    i915_gem_object_is_protected(obj)) {
+			err = intel_pxp_key_check(&vm->gt->pxp, obj, true);
+			if (err) {
+				i915_gem_object_put(obj);
+				return ERR_PTR(err);
+			}
+		}
+
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			i915_gem_object_put(obj);
+			return vma;
+		}
+
+		err = __eb_add_lut(eb, handle, vma);
+		if (likely(!err))
+			return vma;
+
+		i915_gem_object_put(obj);
+		if (err != -EEXIST)
+			return ERR_PTR(err);
+	} while (1);
+}
+
+static int eb_lookup_vmas(struct i915_execbuffer *eb)
+{
+	unsigned int i, current_batch = 0;
+	int err = 0;
+
+	INIT_LIST_HEAD(&eb->relocs);
+
+	for (i = 0; i < eb->buffer_count; i++) {
+		struct i915_vma *vma;
+
+		vma = eb_lookup_vma(eb, eb->exec[i].handle);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto err;
+		}
+
+		err = eb_validate_vma(eb, &eb->exec[i], vma);
+		if (unlikely(err)) {
+			i915_vma_put(vma);
+			goto err;
+		}
+
+		err = eb_add_vma(eb, &current_batch, i, vma);
+		if (err)
+			return err;
+
+		if (i915_gem_object_is_userptr(vma->obj)) {
+			err = i915_gem_object_userptr_submit_init(vma->obj);
+			if (err) {
+				if (i + 1 < eb->buffer_count) {
+					/*
+					 * Execbuffer code expects last vma entry to be NULL,
+					 * since we already initialized this entry,
+					 * set the next value to NULL or we mess up
+					 * cleanup handling.
+					 */
+					eb->vma[i + 1].vma = NULL;
+				}
+
+				return err;
+			}
+
+			eb->vma[i].flags |= __EXEC_OBJECT_USERPTR_INIT;
+			eb->args->flags |= __EXEC_USERPTR_USED;
+		}
+	}
+
+	return 0;
+
+err:
+	eb->vma[i].vma = NULL;
+	return err;
+}
+
+static int eb_lock_vmas(struct i915_execbuffer *eb)
+{
+	unsigned int i;
+	int err;
+
+	for (i = 0; i < eb->buffer_count; i++) {
+		struct eb_vma *ev = &eb->vma[i];
+		struct i915_vma *vma = ev->vma;
+
+		err = i915_gem_object_lock(vma->obj, &eb->ww);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int eb_validate_vmas(struct i915_execbuffer *eb)
+{
+	unsigned int i;
+	int err;
+
+	INIT_LIST_HEAD(&eb->unbound);
+
+	err = eb_lock_vmas(eb);
+	if (err)
+		return err;
+
+	for (i = 0; i < eb->buffer_count; i++) {
+		struct drm_i915_gem_exec_object2 *entry = &eb->exec[i];
+		struct eb_vma *ev = &eb->vma[i];
+		struct i915_vma *vma = ev->vma;
+
+		err = eb_pin_vma(eb, entry, ev);
+		if (err == -EDEADLK)
+			return err;
+
+		if (!err) {
+			if (entry->offset != vma->node.start) {
+				entry->offset = vma->node.start | UPDATE;
+				eb->args->flags |= __EXEC_HAS_RELOC;
+			}
+		} else {
+			eb_unreserve_vma(ev);
+
+			list_add_tail(&ev->bind_link, &eb->unbound);
+			if (drm_mm_node_allocated(&vma->node)) {
+				err = i915_vma_unbind(vma);
+				if (err)
+					return err;
+			}
+		}
+
+		/* Reserve enough slots to accommodate composite fences */
+		err = dma_resv_reserve_fences(vma->obj->base.resv, eb->num_batches);
+		if (err)
+			return err;
+
+		GEM_BUG_ON(drm_mm_node_allocated(&vma->node) &&
+			   eb_vma_misplaced(&eb->exec[i], vma, ev->flags));
+	}
+
+	if (!list_empty(&eb->unbound))
+		return eb_reserve(eb);
+
+	return 0;
+}
+
+static struct eb_vma *
+eb_get_vma(const struct i915_execbuffer *eb, unsigned long handle)
+{
+	if (eb->lut_size < 0) {
+		if (handle >= -eb->lut_size)
+			return NULL;
+		return &eb->vma[handle];
+	} else {
+		struct hlist_head *head;
+		struct eb_vma *ev;
+
+		head = &eb->buckets[hash_32(handle, eb->lut_size)];
+		hlist_for_each_entry(ev, head, node) {
+			if (ev->handle == handle)
+				return ev;
+		}
+		return NULL;
+	}
+}
+
+static void eb_release_vmas(struct i915_execbuffer *eb, bool final)
+{
+	const unsigned int count = eb->buffer_count;
+	unsigned int i;
+
+	for (i = 0; i < count; i++) {
+		struct eb_vma *ev = &eb->vma[i];
+		struct i915_vma *vma = ev->vma;
+
+		if (!vma)
+			break;
+
+		eb_unreserve_vma(ev);
+
+		if (final)
+			i915_vma_put(vma);
+	}
+
+	eb_capture_release(eb);
+	eb_unpin_engine(eb);
+}
+
+static void eb_destroy(const struct i915_execbuffer *eb)
+{
+	if (eb->lut_size > 0)
+		kfree(eb->buckets);
+}
+
+static inline u64
+relocation_target(const struct drm_i915_gem_relocation_entry *reloc,
+		  const struct i915_vma *target)
+{
+	return gen8_canonical_addr((int)reloc->delta + target->node.start);
+}
+
+static void reloc_cache_init(struct reloc_cache *cache,
+			     struct drm_i915_private *i915)
+{
+	cache->page = -1;
+	cache->vaddr = 0;
+	/* Must be a variable in the struct to allow GCC to unroll. */
+	cache->graphics_ver = GRAPHICS_VER(i915);
+	cache->has_llc = HAS_LLC(i915);
+	cache->use_64bit_reloc = HAS_64BIT_RELOC(i915);
+	cache->has_fence = cache->graphics_ver < 4;
+	cache->needs_unfenced = INTEL_INFO(i915)->unfenced_needs_alignment;
+	cache->node.flags = 0;
+}
+
+static inline void *unmask_page(unsigned long p)
+{
+	return (void *)(uintptr_t)(p & PAGE_MASK);
+}
+
+static inline unsigned int unmask_flags(unsigned long p)
+{
+	return p & ~PAGE_MASK;
+}
+
+#define KMAP 0x4 /* after CLFLUSH_FLAGS */
+
+static inline struct i915_ggtt *cache_to_ggtt(struct reloc_cache *cache)
+{
+	struct drm_i915_private *i915 =
+		container_of(cache, struct i915_execbuffer, reloc_cache)->i915;
+	return to_gt(i915)->ggtt;
+}
+
+static void reloc_cache_unmap(struct reloc_cache *cache)
+{
+	void *vaddr;
+
+	if (!cache->vaddr)
+		return;
+
+	vaddr = unmask_page(cache->vaddr);
+	if (cache->vaddr & KMAP)
+		kunmap_atomic(vaddr);
+	else
+		io_mapping_unmap_atomic((void __iomem *)vaddr);
+}
+
+static void reloc_cache_remap(struct reloc_cache *cache,
+			      struct drm_i915_gem_object *obj)
+{
+	void *vaddr;
+
+	if (!cache->vaddr)
+		return;
+
+	if (cache->vaddr & KMAP) {
+		struct page *page = i915_gem_object_get_page(obj, cache->page);
+
+		vaddr = kmap_atomic(page);
+		cache->vaddr = unmask_flags(cache->vaddr) |
+			(unsigned long)vaddr;
+	} else {
+		struct i915_ggtt *ggtt = cache_to_ggtt(cache);
+		unsigned long offset;
+
+		offset = cache->node.start;
+		if (!drm_mm_node_allocated(&cache->node))
+			offset += cache->page << PAGE_SHIFT;
+
+		cache->vaddr = (unsigned long)
+			io_mapping_map_atomic_wc(&ggtt->iomap, offset);
+	}
+}
+
+static void reloc_cache_reset(struct reloc_cache *cache, struct i915_execbuffer *eb)
+{
+	void *vaddr;
+
+	if (!cache->vaddr)
+		return;
+
+	vaddr = unmask_page(cache->vaddr);
+	if (cache->vaddr & KMAP) {
+		struct drm_i915_gem_object *obj =
+			(struct drm_i915_gem_object *)cache->node.mm;
+		if (cache->vaddr & CLFLUSH_AFTER)
+			mb();
+
+		kunmap_atomic(vaddr);
+		i915_gem_object_finish_access(obj);
+	} else {
+		struct i915_ggtt *ggtt = cache_to_ggtt(cache);
+
+		intel_gt_flush_ggtt_writes(ggtt->vm.gt);
+		io_mapping_unmap_atomic((void __iomem *)vaddr);
+
+		if (drm_mm_node_allocated(&cache->node)) {
+			ggtt->vm.clear_range(&ggtt->vm,
+					     cache->node.start,
+					     cache->node.size);
+			mutex_lock(&ggtt->vm.mutex);
+			drm_mm_remove_node(&cache->node);
+			mutex_unlock(&ggtt->vm.mutex);
+		} else {
+			i915_vma_unpin((struct i915_vma *)cache->node.mm);
+		}
+	}
+
+	cache->vaddr = 0;
+	cache->page = -1;
+}
+
+static void *reloc_kmap(struct drm_i915_gem_object *obj,
+			struct reloc_cache *cache,
+			unsigned long pageno)
+{
+	void *vaddr;
+	struct page *page;
+
+	if (cache->vaddr) {
+		kunmap_atomic(unmask_page(cache->vaddr));
+	} else {
+		unsigned int flushes;
+		int err;
+
+		err = i915_gem_object_prepare_write(obj, &flushes);
+		if (err)
+			return ERR_PTR(err);
+
+		BUILD_BUG_ON(KMAP & CLFLUSH_FLAGS);
+		BUILD_BUG_ON((KMAP | CLFLUSH_FLAGS) & PAGE_MASK);
+
+		cache->vaddr = flushes | KMAP;
+		cache->node.mm = (void *)obj;
+		if (flushes)
+			mb();
+	}
+
+	page = i915_gem_object_get_page(obj, pageno);
+	if (!obj->mm.dirty)
+		set_page_dirty(page);
+
+	vaddr = kmap_atomic(page);
+	cache->vaddr = unmask_flags(cache->vaddr) | (unsigned long)vaddr;
+	cache->page = pageno;
+
+	return vaddr;
+}
+
+static void *reloc_iomap(struct i915_vma *batch,
+			 struct i915_execbuffer *eb,
+			 unsigned long page)
+{
+	struct drm_i915_gem_object *obj = batch->obj;
+	struct reloc_cache *cache = &eb->reloc_cache;
+	struct i915_ggtt *ggtt = cache_to_ggtt(cache);
+	unsigned long offset;
+	void *vaddr;
+
+	if (cache->vaddr) {
+		intel_gt_flush_ggtt_writes(ggtt->vm.gt);
+		io_mapping_unmap_atomic((void __force __iomem *) unmask_page(cache->vaddr));
+	} else {
+		struct i915_vma *vma = ERR_PTR(-ENODEV);
+		int err;
+
+		if (i915_gem_object_is_tiled(obj))
+			return ERR_PTR(-EINVAL);
+
+		if (use_cpu_reloc(cache, obj))
+			return NULL;
+
+		err = i915_gem_object_set_to_gtt_domain(obj, true);
+		if (err)
+			return ERR_PTR(err);
+
+		/*
+		 * i915_gem_object_ggtt_pin_ww may attempt to remove the batch
+		 * VMA from the object list because we no longer pin.
+		 *
+		 * Only attempt to pin the batch buffer to ggtt if the current batch
+		 * is not inside ggtt, or the batch buffer is not misplaced.
+		 */
+		if (!i915_is_ggtt(batch->vm) ||
+		    !i915_vma_misplaced(batch, 0, 0, PIN_MAPPABLE)) {
+			vma = i915_gem_object_ggtt_pin_ww(obj, &eb->ww, NULL, 0, 0,
+							  PIN_MAPPABLE |
+							  PIN_NONBLOCK /* NOWARN */ |
+							  PIN_NOEVICT);
+		}
+
+		if (vma == ERR_PTR(-EDEADLK))
+			return vma;
+
+		if (IS_ERR(vma)) {
+			memset(&cache->node, 0, sizeof(cache->node));
+			mutex_lock(&ggtt->vm.mutex);
+			err = drm_mm_insert_node_in_range
+				(&ggtt->vm.mm, &cache->node,
+				 PAGE_SIZE, 0, I915_COLOR_UNEVICTABLE,
+				 0, ggtt->mappable_end,
+				 DRM_MM_INSERT_LOW);
+			mutex_unlock(&ggtt->vm.mutex);
+			if (err) /* no inactive aperture space, use cpu reloc */
+				return NULL;
+		} else {
+			cache->node.start = vma->node.start;
+			cache->node.mm = (void *)vma;
+		}
+	}
+
+	offset = cache->node.start;
+	if (drm_mm_node_allocated(&cache->node)) {
+		ggtt->vm.insert_page(&ggtt->vm,
+				     i915_gem_object_get_dma_address(obj, page),
+				     offset, I915_CACHE_NONE, 0);
+	} else {
+		offset += page << PAGE_SHIFT;
+	}
+
+	vaddr = (void __force *)io_mapping_map_atomic_wc(&ggtt->iomap,
+							 offset);
+	cache->page = page;
+	cache->vaddr = (unsigned long)vaddr;
+
+	return vaddr;
+}
+
+static void *reloc_vaddr(struct i915_vma *vma,
+			 struct i915_execbuffer *eb,
+			 unsigned long page)
+{
+	struct reloc_cache *cache = &eb->reloc_cache;
+	void *vaddr;
+
+	if (cache->page == page) {
+		vaddr = unmask_page(cache->vaddr);
+	} else {
+		vaddr = NULL;
+		if ((cache->vaddr & KMAP) == 0)
+			vaddr = reloc_iomap(vma, eb, page);
+		if (!vaddr)
+			vaddr = reloc_kmap(vma->obj, cache, page);
+	}
+
+	return vaddr;
+}
+
+static void clflush_write32(u32 *addr, u32 value, unsigned int flushes)
+{
+	if (unlikely(flushes & (CLFLUSH_BEFORE | CLFLUSH_AFTER))) {
+		if (flushes & CLFLUSH_BEFORE)
+			drm_clflush_virt_range(addr, sizeof(*addr));
+
+		*addr = value;
+
+		/*
+		 * Writes to the same cacheline are serialised by the CPU
+		 * (including clflush). On the write path, we only require
+		 * that it hits memory in an orderly fashion and place
+		 * mb barriers at the start and end of the relocation phase
+		 * to ensure ordering of clflush wrt to the system.
+		 */
+		if (flushes & CLFLUSH_AFTER)
+			drm_clflush_virt_range(addr, sizeof(*addr));
+	} else
+		*addr = value;
+}
+
+static u64
+relocate_entry(struct i915_vma *vma,
+	       const struct drm_i915_gem_relocation_entry *reloc,
+	       struct i915_execbuffer *eb,
+	       const struct i915_vma *target)
+{
+	u64 target_addr = relocation_target(reloc, target);
+	u64 offset = reloc->offset;
+	bool wide = eb->reloc_cache.use_64bit_reloc;
+	void *vaddr;
+
+repeat:
+	vaddr = reloc_vaddr(vma, eb,
+			    offset >> PAGE_SHIFT);
+	if (IS_ERR(vaddr))
+		return PTR_ERR(vaddr);
+
+	GEM_BUG_ON(!IS_ALIGNED(offset, sizeof(u32)));
+	clflush_write32(vaddr + offset_in_page(offset),
+			lower_32_bits(target_addr),
+			eb->reloc_cache.vaddr);
+
+	if (wide) {
+		offset += sizeof(u32);
+		target_addr >>= 32;
+		wide = false;
+		goto repeat;
+	}
+
+	return target->node.start | UPDATE;
+}
+
+static u64
+eb_relocate_entry(struct i915_execbuffer *eb,
+		  struct eb_vma *ev,
+		  const struct drm_i915_gem_relocation_entry *reloc)
+{
+	struct drm_i915_private *i915 = eb->i915;
+	struct eb_vma *target;
+	int err;
+
+	/* we've already hold a reference to all valid objects */
+	target = eb_get_vma(eb, reloc->target_handle);
+	if (unlikely(!target))
+		return -ENOENT;
+
+	/* Validate that the target is in a valid r/w GPU domain */
+	if (unlikely(reloc->write_domain & (reloc->write_domain - 1))) {
+		drm_dbg(&i915->drm, "reloc with multiple write domains: "
+			  "target %d offset %d "
+			  "read %08x write %08x",
+			  reloc->target_handle,
+			  (int) reloc->offset,
+			  reloc->read_domains,
+			  reloc->write_domain);
+		return -EINVAL;
+	}
+	if (unlikely((reloc->write_domain | reloc->read_domains)
+		     & ~I915_GEM_GPU_DOMAINS)) {
+		drm_dbg(&i915->drm, "reloc with read/write non-GPU domains: "
+			  "target %d offset %d "
+			  "read %08x write %08x",
+			  reloc->target_handle,
+			  (int) reloc->offset,
+			  reloc->read_domains,
+			  reloc->write_domain);
+		return -EINVAL;
+	}
+
+	if (reloc->write_domain) {
+		target->flags |= EXEC_OBJECT_WRITE;
+
+		/*
+		 * Sandybridge PPGTT errata: We need a global gtt mapping
+		 * for MI and pipe_control writes because the gpu doesn't
+		 * properly redirect them through the ppgtt for non_secure
+		 * batchbuffers.
+		 */
+		if (reloc->write_domain == I915_GEM_DOMAIN_INSTRUCTION &&
+		    GRAPHICS_VER(eb->i915) == 6 &&
+		    !i915_vma_is_bound(target->vma, I915_VMA_GLOBAL_BIND)) {
+			struct i915_vma *vma = target->vma;
+
+			reloc_cache_unmap(&eb->reloc_cache);
+			mutex_lock(&vma->vm->mutex);
+			err = i915_vma_bind(target->vma,
+					    target->vma->obj->cache_level,
+					    PIN_GLOBAL, NULL, NULL);
+			mutex_unlock(&vma->vm->mutex);
+			reloc_cache_remap(&eb->reloc_cache, ev->vma->obj);
+			if (err)
+				return err;
+		}
+	}
+
+	/*
+	 * If the relocation already has the right value in it, no
+	 * more work needs to be done.
+	 */
+	if (!DBG_FORCE_RELOC &&
+	    gen8_canonical_addr(target->vma->node.start) == reloc->presumed_offset)
+		return 0;
+
+	/* Check that the relocation address is valid... */
+	if (unlikely(reloc->offset >
+		     ev->vma->size - (eb->reloc_cache.use_64bit_reloc ? 8 : 4))) {
+		drm_dbg(&i915->drm, "Relocation beyond object bounds: "
+			  "target %d offset %d size %d.\n",
+			  reloc->target_handle,
+			  (int)reloc->offset,
+			  (int)ev->vma->size);
+		return -EINVAL;
+	}
+	if (unlikely(reloc->offset & 3)) {
+		drm_dbg(&i915->drm, "Relocation not 4-byte aligned: "
+			  "target %d offset %d.\n",
+			  reloc->target_handle,
+			  (int)reloc->offset);
+		return -EINVAL;
+	}
+
+	/*
+	 * If we write into the object, we need to force the synchronisation
+	 * barrier, either with an asynchronous clflush or if we executed the
+	 * patching using the GPU (though that should be serialised by the
+	 * timeline). To be completely sure, and since we are required to
+	 * do relocations we are already stalling, disable the user's opt
+	 * out of our synchronisation.
+	 */
+	ev->flags &= ~EXEC_OBJECT_ASYNC;
+
+	/* and update the user's relocation entry */
+	return relocate_entry(ev->vma, reloc, eb, target->vma);
+}
+
+static int eb_relocate_vma(struct i915_execbuffer *eb, struct eb_vma *ev)
+{
+#define N_RELOC(x) ((x) / sizeof(struct drm_i915_gem_relocation_entry))
+	struct drm_i915_gem_relocation_entry stack[N_RELOC(512)];
+	const struct drm_i915_gem_exec_object2 *entry = ev->exec;
+	struct drm_i915_gem_relocation_entry __user *urelocs =
+		u64_to_user_ptr(entry->relocs_ptr);
+	unsigned long remain = entry->relocation_count;
+
+	if (unlikely(remain > N_RELOC(ULONG_MAX)))
+		return -EINVAL;
+
+	/*
+	 * We must check that the entire relocation array is safe
+	 * to read. However, if the array is not writable the user loses
+	 * the updated relocation values.
+	 */
+	if (unlikely(!access_ok(urelocs, remain * sizeof(*urelocs))))
+		return -EFAULT;
+
+	do {
+		struct drm_i915_gem_relocation_entry *r = stack;
+		unsigned int count =
+			min_t(unsigned long, remain, ARRAY_SIZE(stack));
+		unsigned int copied;
+
+		/*
+		 * This is the fast path and we cannot handle a pagefault
+		 * whilst holding the struct mutex lest the user pass in the
+		 * relocations contained within a mmaped bo. For in such a case
+		 * we, the page fault handler would call i915_gem_fault() and
+		 * we would try to acquire the struct mutex again. Obviously
+		 * this is bad and so lockdep complains vehemently.
+		 */
+		pagefault_disable();
+		copied = __copy_from_user_inatomic(r, urelocs, count * sizeof(r[0]));
+		pagefault_enable();
+		if (unlikely(copied)) {
+			remain = -EFAULT;
+			goto out;
+		}
+
+		remain -= count;
+		do {
+			u64 offset = eb_relocate_entry(eb, ev, r);
+
+			if (likely(offset == 0)) {
+			} else if ((s64)offset < 0) {
+				remain = (int)offset;
+				goto out;
+			} else {
+				/*
+				 * Note that reporting an error now
+				 * leaves everything in an inconsistent
+				 * state as we have *already* changed
+				 * the relocation value inside the
+				 * object. As we have not changed the
+				 * reloc.presumed_offset or will not
+				 * change the execobject.offset, on the
+				 * call we may not rewrite the value
+				 * inside the object, leaving it
+				 * dangling and causing a GPU hang. Unless
+				 * userspace dynamically rebuilds the
+				 * relocations on each execbuf rather than
+				 * presume a static tree.
+				 *
+				 * We did previously check if the relocations
+				 * were writable (access_ok), an error now
+				 * would be a strange race with mprotect,
+				 * having already demonstrated that we
+				 * can read from this userspace address.
+				 */
+				offset = gen8_canonical_addr(offset & ~UPDATE);
+				__put_user(offset,
+					   &urelocs[r - stack].presumed_offset);
+			}
+		} while (r++, --count);
+		urelocs += ARRAY_SIZE(stack);
+	} while (remain);
+out:
+	reloc_cache_reset(&eb->reloc_cache, eb);
+	return remain;
+}
+
+static int
+eb_relocate_vma_slow(struct i915_execbuffer *eb, struct eb_vma *ev)
+{
+	const struct drm_i915_gem_exec_object2 *entry = ev->exec;
+	struct drm_i915_gem_relocation_entry *relocs =
+		u64_to_ptr(typeof(*relocs), entry->relocs_ptr);
+	unsigned int i;
+	int err;
+
+	for (i = 0; i < entry->relocation_count; i++) {
+		u64 offset = eb_relocate_entry(eb, ev, &relocs[i]);
+
+		if ((s64)offset < 0) {
+			err = (int)offset;
+			goto err;
+		}
+	}
+	err = 0;
+err:
+	reloc_cache_reset(&eb->reloc_cache, eb);
+	return err;
+}
+
+static int check_relocations(const struct drm_i915_gem_exec_object2 *entry)
+{
+	const char __user *addr, *end;
+	unsigned long size;
+	char __maybe_unused c;
+
+	size = entry->relocation_count;
+	if (size == 0)
+		return 0;
+
+	if (size > N_RELOC(ULONG_MAX))
+		return -EINVAL;
+
+	addr = u64_to_user_ptr(entry->relocs_ptr);
+	size *= sizeof(struct drm_i915_gem_relocation_entry);
+	if (!access_ok(addr, size))
+		return -EFAULT;
+
+	end = addr + size;
+	for (; addr < end; addr += PAGE_SIZE) {
+		int err = __get_user(c, addr);
+		if (err)
+			return err;
+	}
+	return __get_user(c, end - 1);
+}
+
+static int eb_copy_relocations(const struct i915_execbuffer *eb)
+{
+	struct drm_i915_gem_relocation_entry *relocs;
+	const unsigned int count = eb->buffer_count;
+	unsigned int i;
+	int err;
+
+	for (i = 0; i < count; i++) {
+		const unsigned int nreloc = eb->exec[i].relocation_count;
+		struct drm_i915_gem_relocation_entry __user *urelocs;
+		unsigned long size;
+		unsigned long copied;
+
+		if (nreloc == 0)
+			continue;
+
+		err = check_relocations(&eb->exec[i]);
+		if (err)
+			goto err;
+
+		urelocs = u64_to_user_ptr(eb->exec[i].relocs_ptr);
+		size = nreloc * sizeof(*relocs);
+
+		relocs = kvmalloc_array(size, 1, GFP_KERNEL);
+		if (!relocs) {
+			err = -ENOMEM;
+			goto err;
+		}
+
+		/* copy_from_user is limited to < 4GiB */
+		copied = 0;
+		do {
+			unsigned int len =
+				min_t(u64, BIT_ULL(31), size - copied);
+
+			if (__copy_from_user((char *)relocs + copied,
+					     (char __user *)urelocs + copied,
+					     len))
+				goto end;
+
+			copied += len;
+		} while (copied < size);
+
+		/*
+		 * As we do not update the known relocation offsets after
+		 * relocating (due to the complexities in lock handling),
+		 * we need to mark them as invalid now so that we force the
+		 * relocation processing next time. Just in case the target
+		 * object is evicted and then rebound into its old
+		 * presumed_offset before the next execbuffer - if that
+		 * happened we would make the mistake of assuming that the
+		 * relocations were valid.
+		 */
+		if (!user_access_begin(urelocs, size))
+			goto end;
+
+		for (copied = 0; copied < nreloc; copied++)
+			unsafe_put_user(-1,
+					&urelocs[copied].presumed_offset,
+					end_user);
+		user_access_end();
+
+		eb->exec[i].relocs_ptr = (uintptr_t)relocs;
+	}
+
+	return 0;
+
+end_user:
+	user_access_end();
+end:
+	kvfree(relocs);
+	err = -EFAULT;
+err:
+	while (i--) {
+		relocs = u64_to_ptr(typeof(*relocs), eb->exec[i].relocs_ptr);
+		if (eb->exec[i].relocation_count)
+			kvfree(relocs);
+	}
+	return err;
+}
+
+static int eb_prefault_relocations(const struct i915_execbuffer *eb)
+{
+	const unsigned int count = eb->buffer_count;
+	unsigned int i;
+
+	for (i = 0; i < count; i++) {
+		int err;
+
+		err = check_relocations(&eb->exec[i]);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int eb_reinit_userptr(struct i915_execbuffer *eb)
+{
+	const unsigned int count = eb->buffer_count;
+	unsigned int i;
+	int ret;
+
+	if (likely(!(eb->args->flags & __EXEC_USERPTR_USED)))
+		return 0;
+
+	for (i = 0; i < count; i++) {
+		struct eb_vma *ev = &eb->vma[i];
+
+		if (!i915_gem_object_is_userptr(ev->vma->obj))
+			continue;
+
+		ret = i915_gem_object_userptr_submit_init(ev->vma->obj);
+		if (ret)
+			return ret;
+
+		ev->flags |= __EXEC_OBJECT_USERPTR_INIT;
+	}
+
+	return 0;
+}
+
+static noinline int eb_relocate_parse_slow(struct i915_execbuffer *eb)
+{
+	bool have_copy = false;
+	struct eb_vma *ev;
+	int err = 0;
+
+repeat:
+	if (signal_pending(current)) {
+		err = -ERESTARTSYS;
+		goto out;
+	}
+
+	/* We may process another execbuffer during the unlock... */
+	eb_release_vmas(eb, false);
+	i915_gem_ww_ctx_fini(&eb->ww);
+
+	/*
+	 * We take 3 passes through the slowpatch.
+	 *
+	 * 1 - we try to just prefault all the user relocation entries and
+	 * then attempt to reuse the atomic pagefault disabled fast path again.
+	 *
+	 * 2 - we copy the user entries to a local buffer here outside of the
+	 * local and allow ourselves to wait upon any rendering before
+	 * relocations
+	 *
+	 * 3 - we already have a local copy of the relocation entries, but
+	 * were interrupted (EAGAIN) whilst waiting for the objects, try again.
+	 */
+	if (!err) {
+		err = eb_prefault_relocations(eb);
+	} else if (!have_copy) {
+		err = eb_copy_relocations(eb);
+		have_copy = err == 0;
+	} else {
+		cond_resched();
+		err = 0;
+	}
+
+	if (!err)
+		err = eb_reinit_userptr(eb);
+
+	i915_gem_ww_ctx_init(&eb->ww, true);
+	if (err)
+		goto out;
+
+	/* reacquire the objects */
+repeat_validate:
+	err = eb_pin_engine(eb, false);
+	if (err)
+		goto err;
+
+	err = eb_validate_vmas(eb);
+	if (err)
+		goto err;
+
+	GEM_BUG_ON(!eb->batches[0]);
+
+	list_for_each_entry(ev, &eb->relocs, reloc_link) {
+		if (!have_copy) {
+			err = eb_relocate_vma(eb, ev);
+			if (err)
+				break;
+		} else {
+			err = eb_relocate_vma_slow(eb, ev);
+			if (err)
+				break;
+		}
+	}
+
+	if (err == -EDEADLK)
+		goto err;
+
+	if (err && !have_copy)
+		goto repeat;
+
+	if (err)
+		goto err;
+
+	/* as last step, parse the command buffer */
+	err = eb_parse(eb);
+	if (err)
+		goto err;
+
+	/*
+	 * Leave the user relocations as are, this is the painfully slow path,
+	 * and we want to avoid the complication of dropping the lock whilst
+	 * having buffers reserved in the aperture and so causing spurious
+	 * ENOSPC for random operations.
+	 */
+
+err:
+	if (err == -EDEADLK) {
+		eb_release_vmas(eb, false);
+		err = i915_gem_ww_ctx_backoff(&eb->ww);
+		if (!err)
+			goto repeat_validate;
+	}
+
+	if (err == -EAGAIN)
+		goto repeat;
+
+out:
+	if (have_copy) {
+		const unsigned int count = eb->buffer_count;
+		unsigned int i;
+
+		for (i = 0; i < count; i++) {
+			const struct drm_i915_gem_exec_object2 *entry =
+				&eb->exec[i];
+			struct drm_i915_gem_relocation_entry *relocs;
+
+			if (!entry->relocation_count)
+				continue;
+
+			relocs = u64_to_ptr(typeof(*relocs), entry->relocs_ptr);
+			kvfree(relocs);
+		}
+	}
+
+	return err;
+}
+
+static int eb_relocate_parse(struct i915_execbuffer *eb)
+{
+	int err;
+	bool throttle = true;
+
+retry:
+	err = eb_pin_engine(eb, throttle);
+	if (err) {
+		if (err != -EDEADLK)
+			return err;
+
+		goto err;
+	}
+
+	/* only throttle once, even if we didn't need to throttle */
+	throttle = false;
+
+	err = eb_validate_vmas(eb);
+	if (err == -EAGAIN)
+		goto slow;
+	else if (err)
+		goto err;
+
+	/* The objects are in their final locations, apply the relocations. */
+	if (eb->args->flags & __EXEC_HAS_RELOC) {
+		struct eb_vma *ev;
+
+		list_for_each_entry(ev, &eb->relocs, reloc_link) {
+			err = eb_relocate_vma(eb, ev);
+			if (err)
+				break;
+		}
+
+		if (err == -EDEADLK)
+			goto err;
+		else if (err)
+			goto slow;
+	}
+
+	if (!err)
+		err = eb_parse(eb);
+
+err:
+	if (err == -EDEADLK) {
+		eb_release_vmas(eb, false);
+		err = i915_gem_ww_ctx_backoff(&eb->ww);
+		if (!err)
+			goto retry;
+	}
+
+	return err;
+
+slow:
+	err = eb_relocate_parse_slow(eb);
+	if (err)
+		/*
+		 * If the user expects the execobject.offset and
+		 * reloc.presumed_offset to be an exact match,
+		 * as for using NO_RELOC, then we cannot update
+		 * the execobject.offset until we have completed
+		 * relocation.
+		 */
+		eb->args->flags &= ~__EXEC_HAS_RELOC;
+
+	return err;
+}
+
+/*
+ * Using two helper loops for the order of which requests / batches are created
+ * and added the to backend. Requests are created in order from the parent to
+ * the last child. Requests are added in the reverse order, from the last child
+ * to parent. This is done for locking reasons as the timeline lock is acquired
+ * during request creation and released when the request is added to the
+ * backend. To make lockdep happy (see intel_context_timeline_lock) this must be
+ * the ordering.
+ */
+#define for_each_batch_create_order(_eb, _i) \
+	for ((_i) = 0; (_i) < (_eb)->num_batches; ++(_i))
+#define for_each_batch_add_order(_eb, _i) \
+	BUILD_BUG_ON(!typecheck(int, _i)); \
+	for ((_i) = (_eb)->num_batches - 1; (_i) >= 0; --(_i))
+
+static struct i915_request *
+eb_find_first_request_added(struct i915_execbuffer *eb)
+{
+	int i;
+
+	for_each_batch_add_order(eb, i)
+		if (eb->requests[i])
+			return eb->requests[i];
+
+	GEM_BUG_ON("Request not found");
+
+	return NULL;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+
+/* Stage with GFP_KERNEL allocations before we enter the signaling critical path */
+static int eb_capture_stage(struct i915_execbuffer *eb)
+{
+	const unsigned int count = eb->buffer_count;
+	unsigned int i = count, j;
+
+	while (i--) {
+		struct eb_vma *ev = &eb->vma[i];
+		struct i915_vma *vma = ev->vma;
+		unsigned int flags = ev->flags;
+
+		if (!(flags & EXEC_OBJECT_CAPTURE))
+			continue;
+
+		if (i915_gem_context_is_recoverable(eb->gem_context) &&
+		    (IS_DGFX(eb->i915) || GRAPHICS_VER_FULL(eb->i915) > IP_VER(12, 0)))
+			return -EINVAL;
+
+		for_each_batch_create_order(eb, j) {
+			struct i915_capture_list *capture;
+
+			capture = kmalloc(sizeof(*capture), GFP_KERNEL);
+			if (!capture)
+				continue;
+
+			capture->next = eb->capture_lists[j];
+			capture->vma_res = i915_vma_resource_get(vma->resource);
+			eb->capture_lists[j] = capture;
+		}
+	}
+
+	return 0;
+}
+
+/* Commit once we're in the critical path */
+static void eb_capture_commit(struct i915_execbuffer *eb)
+{
+	unsigned int j;
+
+	for_each_batch_create_order(eb, j) {
+		struct i915_request *rq = eb->requests[j];
+
+		if (!rq)
+			break;
+
+		rq->capture_list = eb->capture_lists[j];
+		eb->capture_lists[j] = NULL;
+	}
+}
+
+/*
+ * Release anything that didn't get committed due to errors.
+ * The capture_list will otherwise be freed at request retire.
+ */
+static void eb_capture_release(struct i915_execbuffer *eb)
+{
+	unsigned int j;
+
+	for_each_batch_create_order(eb, j) {
+		if (eb->capture_lists[j]) {
+			i915_request_free_capture_list(eb->capture_lists[j]);
+			eb->capture_lists[j] = NULL;
+		}
+	}
+}
+
+static void eb_capture_list_clear(struct i915_execbuffer *eb)
+{
+	memset(eb->capture_lists, 0, sizeof(eb->capture_lists));
+}
+
+#else
+
+static int eb_capture_stage(struct i915_execbuffer *eb)
+{
+	return 0;
+}
+
+static void eb_capture_commit(struct i915_execbuffer *eb)
+{
+}
+
+static void eb_capture_release(struct i915_execbuffer *eb)
+{
+}
+
+static void eb_capture_list_clear(struct i915_execbuffer *eb)
+{
+}
+
+#endif
+
+static int eb_move_to_gpu(struct i915_execbuffer *eb)
+{
+	const unsigned int count = eb->buffer_count;
+	unsigned int i = count;
+	int err = 0, j;
+
+	while (i--) {
+		struct eb_vma *ev = &eb->vma[i];
+		struct i915_vma *vma = ev->vma;
+		unsigned int flags = ev->flags;
+		struct drm_i915_gem_object *obj = vma->obj;
+
+		assert_vma_held(vma);
+
+		/*
+		 * If the GPU is not _reading_ through the CPU cache, we need
+		 * to make sure that any writes (both previous GPU writes from
+		 * before a change in snooping levels and normal CPU writes)
+		 * caught in that cache are flushed to main memory.
+		 *
+		 * We want to say
+		 *   obj->cache_dirty &&
+		 *   !(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ)
+		 * but gcc's optimiser doesn't handle that as well and emits
+		 * two jumps instead of one. Maybe one day...
+		 *
+		 * FIXME: There is also sync flushing in set_pages(), which
+		 * serves a different purpose(some of the time at least).
+		 *
+		 * We should consider:
+		 *
+		 *   1. Rip out the async flush code.
+		 *
+		 *   2. Or make the sync flushing use the async clflush path
+		 *   using mandatory fences underneath. Currently the below
+		 *   async flush happens after we bind the object.
+		 */
+		if (unlikely(obj->cache_dirty & ~obj->cache_coherent)) {
+			if (i915_gem_clflush_object(obj, 0))
+				flags &= ~EXEC_OBJECT_ASYNC;
+		}
+
+		/* We only need to await on the first request */
+		if (err == 0 && !(flags & EXEC_OBJECT_ASYNC)) {
+			err = i915_request_await_object
+				(eb_find_first_request_added(eb), obj,
+				 flags & EXEC_OBJECT_WRITE);
+		}
+
+		for_each_batch_add_order(eb, j) {
+			if (err)
+				break;
+			if (!eb->requests[j])
+				continue;
+
+			err = _i915_vma_move_to_active(vma, eb->requests[j],
+						       j ? NULL :
+						       eb->composite_fence ?
+						       eb->composite_fence :
+						       &eb->requests[j]->fence,
+						       flags | __EXEC_OBJECT_NO_RESERVE);
+		}
+	}
+
+#ifdef CONFIG_MMU_NOTIFIER
+	if (!err && (eb->args->flags & __EXEC_USERPTR_USED)) {
+		read_lock(&eb->i915->mm.notifier_lock);
+
+		/*
+		 * count is always at least 1, otherwise __EXEC_USERPTR_USED
+		 * could not have been set
+		 */
+		for (i = 0; i < count; i++) {
+			struct eb_vma *ev = &eb->vma[i];
+			struct drm_i915_gem_object *obj = ev->vma->obj;
+
+			if (!i915_gem_object_is_userptr(obj))
+				continue;
+
+			err = i915_gem_object_userptr_submit_done(obj);
+			if (err)
+				break;
+		}
+
+		read_unlock(&eb->i915->mm.notifier_lock);
+	}
+#endif
+
+	if (unlikely(err))
+		goto err_skip;
+
+	/* Unconditionally flush any chipset caches (for streaming writes). */
+	intel_gt_chipset_flush(eb->gt);
+	eb_capture_commit(eb);
+
+	return 0;
+
+err_skip:
+	for_each_batch_create_order(eb, j) {
+		if (!eb->requests[j])
+			break;
+
+		i915_request_set_error_once(eb->requests[j], err);
+	}
+	return err;
+}
+
+static int i915_gem_check_execbuffer(struct drm_i915_gem_execbuffer2 *exec)
+{
+	if (exec->flags & __I915_EXEC_ILLEGAL_FLAGS)
+		return -EINVAL;
+
+	/* Kernel clipping was a DRI1 misfeature */
+	if (!(exec->flags & (I915_EXEC_FENCE_ARRAY |
+			     I915_EXEC_USE_EXTENSIONS))) {
+		if (exec->num_cliprects || exec->cliprects_ptr)
+			return -EINVAL;
+	}
+
+	if (exec->DR4 == 0xffffffff) {
+		DRM_DEBUG("UXA submitting garbage DR4, fixing up\n");
+		exec->DR4 = 0;
+	}
+	if (exec->DR1 || exec->DR4)
+		return -EINVAL;
+
+	if ((exec->batch_start_offset | exec->batch_len) & 0x7)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int i915_reset_gen7_sol_offsets(struct i915_request *rq)
+{
+	u32 *cs;
+	int i;
+
+	if (GRAPHICS_VER(rq->engine->i915) != 7 || rq->engine->id != RCS0) {
+		drm_dbg(&rq->engine->i915->drm, "sol reset is gen7/rcs only\n");
+		return -EINVAL;
+	}
+
+	cs = intel_ring_begin(rq, 4 * 2 + 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(4);
+	for (i = 0; i < 4; i++) {
+		*cs++ = i915_mmio_reg_offset(GEN7_SO_WRITE_OFFSET(i));
+		*cs++ = 0;
+	}
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static struct i915_vma *
+shadow_batch_pin(struct i915_execbuffer *eb,
+		 struct drm_i915_gem_object *obj,
+		 struct i915_address_space *vm,
+		 unsigned int flags)
+{
+	struct i915_vma *vma;
+	int err;
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma))
+		return vma;
+
+	err = i915_vma_pin_ww(vma, &eb->ww, 0, 0, flags | PIN_VALIDATE);
+	if (err)
+		return ERR_PTR(err);
+
+	return vma;
+}
+
+static struct i915_vma *eb_dispatch_secure(struct i915_execbuffer *eb, struct i915_vma *vma)
+{
+	/*
+	 * snb/ivb/vlv conflate the "batch in ppgtt" bit with the "non-secure
+	 * batch" bit. Hence we need to pin secure batches into the global gtt.
+	 * hsw should have this fixed, but bdw mucks it up again. */
+	if (eb->batch_flags & I915_DISPATCH_SECURE)
+		return i915_gem_object_ggtt_pin_ww(vma->obj, &eb->ww, NULL, 0, 0, PIN_VALIDATE);
+
+	return NULL;
+}
+
+static int eb_parse(struct i915_execbuffer *eb)
+{
+	struct drm_i915_private *i915 = eb->i915;
+	struct intel_gt_buffer_pool_node *pool = eb->batch_pool;
+	struct i915_vma *shadow, *trampoline, *batch;
+	unsigned long len;
+	int err;
+
+	if (!eb_use_cmdparser(eb)) {
+		batch = eb_dispatch_secure(eb, eb->batches[0]->vma);
+		if (IS_ERR(batch))
+			return PTR_ERR(batch);
+
+		goto secure_batch;
+	}
+
+	if (intel_context_is_parallel(eb->context))
+		return -EINVAL;
+
+	len = eb->batch_len[0];
+	if (!CMDPARSER_USES_GGTT(eb->i915)) {
+		/*
+		 * ppGTT backed shadow buffers must be mapped RO, to prevent
+		 * post-scan tampering
+		 */
+		if (!eb->context->vm->has_read_only) {
+			drm_dbg(&i915->drm,
+				"Cannot prevent post-scan tampering without RO capable vm\n");
+			return -EINVAL;
+		}
+	} else {
+		len += I915_CMD_PARSER_TRAMPOLINE_SIZE;
+	}
+	if (unlikely(len < eb->batch_len[0])) /* last paranoid check of overflow */
+		return -EINVAL;
+
+	if (!pool) {
+		pool = intel_gt_get_buffer_pool(eb->gt, len,
+						I915_MAP_WB);
+		if (IS_ERR(pool))
+			return PTR_ERR(pool);
+		eb->batch_pool = pool;
+	}
+
+	err = i915_gem_object_lock(pool->obj, &eb->ww);
+	if (err)
+		return err;
+
+	shadow = shadow_batch_pin(eb, pool->obj, eb->context->vm, PIN_USER);
+	if (IS_ERR(shadow))
+		return PTR_ERR(shadow);
+
+	intel_gt_buffer_pool_mark_used(pool);
+	i915_gem_object_set_readonly(shadow->obj);
+	shadow->private = pool;
+
+	trampoline = NULL;
+	if (CMDPARSER_USES_GGTT(eb->i915)) {
+		trampoline = shadow;
+
+		shadow = shadow_batch_pin(eb, pool->obj,
+					  &eb->gt->ggtt->vm,
+					  PIN_GLOBAL);
+		if (IS_ERR(shadow))
+			return PTR_ERR(shadow);
+
+		shadow->private = pool;
+
+		eb->batch_flags |= I915_DISPATCH_SECURE;
+	}
+
+	batch = eb_dispatch_secure(eb, shadow);
+	if (IS_ERR(batch))
+		return PTR_ERR(batch);
+
+	err = dma_resv_reserve_fences(shadow->obj->base.resv, 1);
+	if (err)
+		return err;
+
+	err = intel_engine_cmd_parser(eb->context->engine,
+				      eb->batches[0]->vma,
+				      eb->batch_start_offset,
+				      eb->batch_len[0],
+				      shadow, trampoline);
+	if (err)
+		return err;
+
+	eb->batches[0] = &eb->vma[eb->buffer_count++];
+	eb->batches[0]->vma = i915_vma_get(shadow);
+	eb->batches[0]->flags = __EXEC_OBJECT_HAS_PIN;
+
+	eb->trampoline = trampoline;
+	eb->batch_start_offset = 0;
+
+secure_batch:
+	if (batch) {
+		if (intel_context_is_parallel(eb->context))
+			return -EINVAL;
+
+		eb->batches[0] = &eb->vma[eb->buffer_count++];
+		eb->batches[0]->flags = __EXEC_OBJECT_HAS_PIN;
+		eb->batches[0]->vma = i915_vma_get(batch);
+	}
+	return 0;
+}
+
+static int eb_request_submit(struct i915_execbuffer *eb,
+			     struct i915_request *rq,
+			     struct i915_vma *batch,
+			     u64 batch_len)
+{
+	int err;
+
+	if (intel_context_nopreempt(rq->context))
+		__set_bit(I915_FENCE_FLAG_NOPREEMPT, &rq->fence.flags);
+
+	if (eb->args->flags & I915_EXEC_GEN7_SOL_RESET) {
+		err = i915_reset_gen7_sol_offsets(rq);
+		if (err)
+			return err;
+	}
+
+	/*
+	 * After we completed waiting for other engines (using HW semaphores)
+	 * then we can signal that this request/batch is ready to run. This
+	 * allows us to determine if the batch is still waiting on the GPU
+	 * or actually running by checking the breadcrumb.
+	 */
+	if (rq->context->engine->emit_init_breadcrumb) {
+		err = rq->context->engine->emit_init_breadcrumb(rq);
+		if (err)
+			return err;
+	}
+
+	err = rq->context->engine->emit_bb_start(rq,
+						 batch->node.start +
+						 eb->batch_start_offset,
+						 batch_len,
+						 eb->batch_flags);
+	if (err)
+		return err;
+
+	if (eb->trampoline) {
+		GEM_BUG_ON(intel_context_is_parallel(rq->context));
+		GEM_BUG_ON(eb->batch_start_offset);
+		err = rq->context->engine->emit_bb_start(rq,
+							 eb->trampoline->node.start +
+							 batch_len, 0, 0);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int eb_submit(struct i915_execbuffer *eb)
+{
+	unsigned int i;
+	int err;
+
+	err = eb_move_to_gpu(eb);
+
+	for_each_batch_create_order(eb, i) {
+		if (!eb->requests[i])
+			break;
+
+		trace_i915_request_queue(eb->requests[i], eb->batch_flags);
+		if (!err)
+			err = eb_request_submit(eb, eb->requests[i],
+						eb->batches[i]->vma,
+						eb->batch_len[i]);
+	}
+
+	return err;
+}
+
+static int num_vcs_engines(struct drm_i915_private *i915)
+{
+	return hweight_long(VDBOX_MASK(to_gt(i915)));
+}
+
+/*
+ * Find one BSD ring to dispatch the corresponding BSD command.
+ * The engine index is returned.
+ */
+static unsigned int
+gen8_dispatch_bsd_engine(struct drm_i915_private *dev_priv,
+			 struct drm_file *file)
+{
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+
+	/* Check whether the file_priv has already selected one ring. */
+	if ((int)file_priv->bsd_engine < 0)
+		file_priv->bsd_engine =
+			prandom_u32_max(num_vcs_engines(dev_priv));
+
+	return file_priv->bsd_engine;
+}
+
+static const enum intel_engine_id user_ring_map[] = {
+	[I915_EXEC_DEFAULT]	= RCS0,
+	[I915_EXEC_RENDER]	= RCS0,
+	[I915_EXEC_BLT]		= BCS0,
+	[I915_EXEC_BSD]		= VCS0,
+	[I915_EXEC_VEBOX]	= VECS0
+};
+
+static struct i915_request *eb_throttle(struct i915_execbuffer *eb, struct intel_context *ce)
+{
+	struct intel_ring *ring = ce->ring;
+	struct intel_timeline *tl = ce->timeline;
+	struct i915_request *rq;
+
+	/*
+	 * Completely unscientific finger-in-the-air estimates for suitable
+	 * maximum user request size (to avoid blocking) and then backoff.
+	 */
+	if (intel_ring_update_space(ring) >= PAGE_SIZE)
+		return NULL;
+
+	/*
+	 * Find a request that after waiting upon, there will be at least half
+	 * the ring available. The hysteresis allows us to compete for the
+	 * shared ring and should mean that we sleep less often prior to
+	 * claiming our resources, but not so long that the ring completely
+	 * drains before we can submit our next request.
+	 */
+	list_for_each_entry(rq, &tl->requests, link) {
+		if (rq->ring != ring)
+			continue;
+
+		if (__intel_ring_space(rq->postfix,
+				       ring->emit, ring->size) > ring->size / 2)
+			break;
+	}
+	if (&rq->link == &tl->requests)
+		return NULL; /* weird, we will check again later for real */
+
+	return i915_request_get(rq);
+}
+
+static int eb_pin_timeline(struct i915_execbuffer *eb, struct intel_context *ce,
+			   bool throttle)
+{
+	struct intel_timeline *tl;
+	struct i915_request *rq = NULL;
+
+	/*
+	 * Take a local wakeref for preparing to dispatch the execbuf as
+	 * we expect to access the hardware fairly frequently in the
+	 * process, and require the engine to be kept awake between accesses.
+	 * Upon dispatch, we acquire another prolonged wakeref that we hold
+	 * until the timeline is idle, which in turn releases the wakeref
+	 * taken on the engine, and the parent device.
+	 */
+	tl = intel_context_timeline_lock(ce);
+	if (IS_ERR(tl))
+		return PTR_ERR(tl);
+
+	intel_context_enter(ce);
+	if (throttle)
+		rq = eb_throttle(eb, ce);
+	intel_context_timeline_unlock(tl);
+
+	if (rq) {
+		bool nonblock = eb->file->filp->f_flags & O_NONBLOCK;
+		long timeout = nonblock ? 0 : MAX_SCHEDULE_TIMEOUT;
+
+		if (i915_request_wait(rq, I915_WAIT_INTERRUPTIBLE,
+				      timeout) < 0) {
+			i915_request_put(rq);
+
+			/*
+			 * Error path, cannot use intel_context_timeline_lock as
+			 * that is user interruptable and this clean up step
+			 * must be done.
+			 */
+			mutex_lock(&ce->timeline->mutex);
+			intel_context_exit(ce);
+			mutex_unlock(&ce->timeline->mutex);
+
+			if (nonblock)
+				return -EWOULDBLOCK;
+			else
+				return -EINTR;
+		}
+		i915_request_put(rq);
+	}
+
+	return 0;
+}
+
+static int eb_pin_engine(struct i915_execbuffer *eb, bool throttle)
+{
+	struct intel_context *ce = eb->context, *child;
+	int err;
+	int i = 0, j = 0;
+
+	GEM_BUG_ON(eb->args->flags & __EXEC_ENGINE_PINNED);
+
+	if (unlikely(intel_context_is_banned(ce)))
+		return -EIO;
+
+	/*
+	 * Pinning the contexts may generate requests in order to acquire
+	 * GGTT space, so do this first before we reserve a seqno for
+	 * ourselves.
+	 */
+	err = intel_context_pin_ww(ce, &eb->ww);
+	if (err)
+		return err;
+	for_each_child(ce, child) {
+		err = intel_context_pin_ww(child, &eb->ww);
+		GEM_BUG_ON(err);	/* perma-pinned should incr a counter */
+	}
+
+	for_each_child(ce, child) {
+		err = eb_pin_timeline(eb, child, throttle);
+		if (err)
+			goto unwind;
+		++i;
+	}
+	err = eb_pin_timeline(eb, ce, throttle);
+	if (err)
+		goto unwind;
+
+	eb->args->flags |= __EXEC_ENGINE_PINNED;
+	return 0;
+
+unwind:
+	for_each_child(ce, child) {
+		if (j++ < i) {
+			mutex_lock(&child->timeline->mutex);
+			intel_context_exit(child);
+			mutex_unlock(&child->timeline->mutex);
+		}
+	}
+	for_each_child(ce, child)
+		intel_context_unpin(child);
+	intel_context_unpin(ce);
+	return err;
+}
+
+static void eb_unpin_engine(struct i915_execbuffer *eb)
+{
+	struct intel_context *ce = eb->context, *child;
+
+	if (!(eb->args->flags & __EXEC_ENGINE_PINNED))
+		return;
+
+	eb->args->flags &= ~__EXEC_ENGINE_PINNED;
+
+	for_each_child(ce, child) {
+		mutex_lock(&child->timeline->mutex);
+		intel_context_exit(child);
+		mutex_unlock(&child->timeline->mutex);
+
+		intel_context_unpin(child);
+	}
+
+	mutex_lock(&ce->timeline->mutex);
+	intel_context_exit(ce);
+	mutex_unlock(&ce->timeline->mutex);
+
+	intel_context_unpin(ce);
+}
+
+static unsigned int
+eb_select_legacy_ring(struct i915_execbuffer *eb)
+{
+	struct drm_i915_private *i915 = eb->i915;
+	struct drm_i915_gem_execbuffer2 *args = eb->args;
+	unsigned int user_ring_id = args->flags & I915_EXEC_RING_MASK;
+
+	if (user_ring_id != I915_EXEC_BSD &&
+	    (args->flags & I915_EXEC_BSD_MASK)) {
+		drm_dbg(&i915->drm,
+			"execbuf with non bsd ring but with invalid "
+			"bsd dispatch flags: %d\n", (int)(args->flags));
+		return -1;
+	}
+
+	if (user_ring_id == I915_EXEC_BSD && num_vcs_engines(i915) > 1) {
+		unsigned int bsd_idx = args->flags & I915_EXEC_BSD_MASK;
+
+		if (bsd_idx == I915_EXEC_BSD_DEFAULT) {
+			bsd_idx = gen8_dispatch_bsd_engine(i915, eb->file);
+		} else if (bsd_idx >= I915_EXEC_BSD_RING1 &&
+			   bsd_idx <= I915_EXEC_BSD_RING2) {
+			bsd_idx >>= I915_EXEC_BSD_SHIFT;
+			bsd_idx--;
+		} else {
+			drm_dbg(&i915->drm,
+				"execbuf with unknown bsd ring: %u\n",
+				bsd_idx);
+			return -1;
+		}
+
+		return _VCS(bsd_idx);
+	}
+
+	if (user_ring_id >= ARRAY_SIZE(user_ring_map)) {
+		drm_dbg(&i915->drm, "execbuf with unknown ring: %u\n",
+			user_ring_id);
+		return -1;
+	}
+
+	return user_ring_map[user_ring_id];
+}
+
+static int
+eb_select_engine(struct i915_execbuffer *eb)
+{
+	struct intel_context *ce, *child;
+	unsigned int idx;
+	int err;
+
+	if (i915_gem_context_user_engines(eb->gem_context))
+		idx = eb->args->flags & I915_EXEC_RING_MASK;
+	else
+		idx = eb_select_legacy_ring(eb);
+
+	ce = i915_gem_context_get_engine(eb->gem_context, idx);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	if (intel_context_is_parallel(ce)) {
+		if (eb->buffer_count < ce->parallel.number_children + 1) {
+			intel_context_put(ce);
+			return -EINVAL;
+		}
+		if (eb->batch_start_offset || eb->args->batch_len) {
+			intel_context_put(ce);
+			return -EINVAL;
+		}
+	}
+	eb->num_batches = ce->parallel.number_children + 1;
+
+	for_each_child(ce, child)
+		intel_context_get(child);
+	intel_gt_pm_get(ce->engine->gt);
+
+	if (!test_bit(CONTEXT_ALLOC_BIT, &ce->flags)) {
+		err = intel_context_alloc_state(ce);
+		if (err)
+			goto err;
+	}
+	for_each_child(ce, child) {
+		if (!test_bit(CONTEXT_ALLOC_BIT, &child->flags)) {
+			err = intel_context_alloc_state(child);
+			if (err)
+				goto err;
+		}
+	}
+
+	/*
+	 * ABI: Before userspace accesses the GPU (e.g. execbuffer), report
+	 * EIO if the GPU is already wedged.
+	 */
+	err = intel_gt_terminally_wedged(ce->engine->gt);
+	if (err)
+		goto err;
+
+	if (!i915_vm_tryget(ce->vm)) {
+		err = -ENOENT;
+		goto err;
+	}
+
+	eb->context = ce;
+	eb->gt = ce->engine->gt;
+
+	/*
+	 * Make sure engine pool stays alive even if we call intel_context_put
+	 * during ww handling. The pool is destroyed when last pm reference
+	 * is dropped, which breaks our -EDEADLK handling.
+	 */
+	return err;
+
+err:
+	intel_gt_pm_put(ce->engine->gt);
+	for_each_child(ce, child)
+		intel_context_put(child);
+	intel_context_put(ce);
+	return err;
+}
+
+static void
+eb_put_engine(struct i915_execbuffer *eb)
+{
+	struct intel_context *child;
+
+	i915_vm_put(eb->context->vm);
+	intel_gt_pm_put(eb->gt);
+	for_each_child(eb->context, child)
+		intel_context_put(child);
+	intel_context_put(eb->context);
+}
+
+static void
+__free_fence_array(struct eb_fence *fences, unsigned int n)
+{
+	while (n--) {
+		drm_syncobj_put(ptr_mask_bits(fences[n].syncobj, 2));
+		dma_fence_put(fences[n].dma_fence);
+		dma_fence_chain_free(fences[n].chain_fence);
+	}
+	kvfree(fences);
+}
+
+static int
+add_timeline_fence_array(struct i915_execbuffer *eb,
+			 const struct drm_i915_gem_execbuffer_ext_timeline_fences *timeline_fences)
+{
+	struct drm_i915_gem_exec_fence __user *user_fences;
+	u64 __user *user_values;
+	struct eb_fence *f;
+	u64 nfences;
+	int err = 0;
+
+	nfences = timeline_fences->fence_count;
+	if (!nfences)
+		return 0;
+
+	/* Check multiplication overflow for access_ok() and kvmalloc_array() */
+	BUILD_BUG_ON(sizeof(size_t) > sizeof(unsigned long));
+	if (nfences > min_t(unsigned long,
+			    ULONG_MAX / sizeof(*user_fences),
+			    SIZE_MAX / sizeof(*f)) - eb->num_fences)
+		return -EINVAL;
+
+	user_fences = u64_to_user_ptr(timeline_fences->handles_ptr);
+	if (!access_ok(user_fences, nfences * sizeof(*user_fences)))
+		return -EFAULT;
+
+	user_values = u64_to_user_ptr(timeline_fences->values_ptr);
+	if (!access_ok(user_values, nfences * sizeof(*user_values)))
+		return -EFAULT;
+
+	f = krealloc(eb->fences,
+		     (eb->num_fences + nfences) * sizeof(*f),
+		     __GFP_NOWARN | GFP_KERNEL);
+	if (!f)
+		return -ENOMEM;
+
+	eb->fences = f;
+	f += eb->num_fences;
+
+	BUILD_BUG_ON(~(ARCH_KMALLOC_MINALIGN - 1) &
+		     ~__I915_EXEC_FENCE_UNKNOWN_FLAGS);
+
+	while (nfences--) {
+		struct drm_i915_gem_exec_fence user_fence;
+		struct drm_syncobj *syncobj;
+		struct dma_fence *fence = NULL;
+		u64 point;
+
+		if (__copy_from_user(&user_fence,
+				     user_fences++,
+				     sizeof(user_fence)))
+			return -EFAULT;
+
+		if (user_fence.flags & __I915_EXEC_FENCE_UNKNOWN_FLAGS)
+			return -EINVAL;
+
+		if (__get_user(point, user_values++))
+			return -EFAULT;
+
+		syncobj = drm_syncobj_find(eb->file, user_fence.handle);
+		if (!syncobj) {
+			DRM_DEBUG("Invalid syncobj handle provided\n");
+			return -ENOENT;
+		}
+
+		fence = drm_syncobj_fence_get(syncobj);
+
+		if (!fence && user_fence.flags &&
+		    !(user_fence.flags & I915_EXEC_FENCE_SIGNAL)) {
+			DRM_DEBUG("Syncobj handle has no fence\n");
+			drm_syncobj_put(syncobj);
+			return -EINVAL;
+		}
+
+		if (fence)
+			err = dma_fence_chain_find_seqno(&fence, point);
+
+		if (err && !(user_fence.flags & I915_EXEC_FENCE_SIGNAL)) {
+			DRM_DEBUG("Syncobj handle missing requested point %llu\n", point);
+			dma_fence_put(fence);
+			drm_syncobj_put(syncobj);
+			return err;
+		}
+
+		/*
+		 * A point might have been signaled already and
+		 * garbage collected from the timeline. In this case
+		 * just ignore the point and carry on.
+		 */
+		if (!fence && !(user_fence.flags & I915_EXEC_FENCE_SIGNAL)) {
+			drm_syncobj_put(syncobj);
+			continue;
+		}
+
+		/*
+		 * For timeline syncobjs we need to preallocate chains for
+		 * later signaling.
+		 */
+		if (point != 0 && user_fence.flags & I915_EXEC_FENCE_SIGNAL) {
+			/*
+			 * Waiting and signaling the same point (when point !=
+			 * 0) would break the timeline.
+			 */
+			if (user_fence.flags & I915_EXEC_FENCE_WAIT) {
+				DRM_DEBUG("Trying to wait & signal the same timeline point.\n");
+				dma_fence_put(fence);
+				drm_syncobj_put(syncobj);
+				return -EINVAL;
+			}
+
+			f->chain_fence = dma_fence_chain_alloc();
+			if (!f->chain_fence) {
+				drm_syncobj_put(syncobj);
+				dma_fence_put(fence);
+				return -ENOMEM;
+			}
+		} else {
+			f->chain_fence = NULL;
+		}
+
+		f->syncobj = ptr_pack_bits(syncobj, user_fence.flags, 2);
+		f->dma_fence = fence;
+		f->value = point;
+		f++;
+		eb->num_fences++;
+	}
+
+	return 0;
+}
+
+static int add_fence_array(struct i915_execbuffer *eb)
+{
+	struct drm_i915_gem_execbuffer2 *args = eb->args;
+	struct drm_i915_gem_exec_fence __user *user;
+	unsigned long num_fences = args->num_cliprects;
+	struct eb_fence *f;
+
+	if (!(args->flags & I915_EXEC_FENCE_ARRAY))
+		return 0;
+
+	if (!num_fences)
+		return 0;
+
+	/* Check multiplication overflow for access_ok() and kvmalloc_array() */
+	BUILD_BUG_ON(sizeof(size_t) > sizeof(unsigned long));
+	if (num_fences > min_t(unsigned long,
+			       ULONG_MAX / sizeof(*user),
+			       SIZE_MAX / sizeof(*f) - eb->num_fences))
+		return -EINVAL;
+
+	user = u64_to_user_ptr(args->cliprects_ptr);
+	if (!access_ok(user, num_fences * sizeof(*user)))
+		return -EFAULT;
+
+	f = krealloc(eb->fences,
+		     (eb->num_fences + num_fences) * sizeof(*f),
+		     __GFP_NOWARN | GFP_KERNEL);
+	if (!f)
+		return -ENOMEM;
+
+	eb->fences = f;
+	f += eb->num_fences;
+	while (num_fences--) {
+		struct drm_i915_gem_exec_fence user_fence;
+		struct drm_syncobj *syncobj;
+		struct dma_fence *fence = NULL;
+
+		if (__copy_from_user(&user_fence, user++, sizeof(user_fence)))
+			return -EFAULT;
+
+		if (user_fence.flags & __I915_EXEC_FENCE_UNKNOWN_FLAGS)
+			return -EINVAL;
+
+		syncobj = drm_syncobj_find(eb->file, user_fence.handle);
+		if (!syncobj) {
+			DRM_DEBUG("Invalid syncobj handle provided\n");
+			return -ENOENT;
+		}
+
+		if (user_fence.flags & I915_EXEC_FENCE_WAIT) {
+			fence = drm_syncobj_fence_get(syncobj);
+			if (!fence) {
+				DRM_DEBUG("Syncobj handle has no fence\n");
+				drm_syncobj_put(syncobj);
+				return -EINVAL;
+			}
+		}
+
+		BUILD_BUG_ON(~(ARCH_KMALLOC_MINALIGN - 1) &
+			     ~__I915_EXEC_FENCE_UNKNOWN_FLAGS);
+
+		f->syncobj = ptr_pack_bits(syncobj, user_fence.flags, 2);
+		f->dma_fence = fence;
+		f->value = 0;
+		f->chain_fence = NULL;
+		f++;
+		eb->num_fences++;
+	}
+
+	return 0;
+}
+
+static void put_fence_array(struct eb_fence *fences, int num_fences)
+{
+	if (fences)
+		__free_fence_array(fences, num_fences);
+}
+
+static int
+await_fence_array(struct i915_execbuffer *eb,
+		  struct i915_request *rq)
+{
+	unsigned int n;
+	int err;
+
+	for (n = 0; n < eb->num_fences; n++) {
+		struct drm_syncobj *syncobj;
+		unsigned int flags;
+
+		syncobj = ptr_unpack_bits(eb->fences[n].syncobj, &flags, 2);
+
+		if (!eb->fences[n].dma_fence)
+			continue;
+
+		err = i915_request_await_dma_fence(rq, eb->fences[n].dma_fence);
+		if (err < 0)
+			return err;
+	}
+
+	return 0;
+}
+
+static void signal_fence_array(const struct i915_execbuffer *eb,
+			       struct dma_fence * const fence)
+{
+	unsigned int n;
+
+	for (n = 0; n < eb->num_fences; n++) {
+		struct drm_syncobj *syncobj;
+		unsigned int flags;
+
+		syncobj = ptr_unpack_bits(eb->fences[n].syncobj, &flags, 2);
+		if (!(flags & I915_EXEC_FENCE_SIGNAL))
+			continue;
+
+		if (eb->fences[n].chain_fence) {
+			drm_syncobj_add_point(syncobj,
+					      eb->fences[n].chain_fence,
+					      fence,
+					      eb->fences[n].value);
+			/*
+			 * The chain's ownership is transferred to the
+			 * timeline.
+			 */
+			eb->fences[n].chain_fence = NULL;
+		} else {
+			drm_syncobj_replace_fence(syncobj, fence);
+		}
+	}
+}
+
+static int
+parse_timeline_fences(struct i915_user_extension __user *ext, void *data)
+{
+	struct i915_execbuffer *eb = data;
+	struct drm_i915_gem_execbuffer_ext_timeline_fences timeline_fences;
+
+	if (copy_from_user(&timeline_fences, ext, sizeof(timeline_fences)))
+		return -EFAULT;
+
+	return add_timeline_fence_array(eb, &timeline_fences);
+}
+
+static void retire_requests(struct intel_timeline *tl, struct i915_request *end)
+{
+	struct i915_request *rq, *rn;
+
+	list_for_each_entry_safe(rq, rn, &tl->requests, link)
+		if (rq == end || !i915_request_retire(rq))
+			break;
+}
+
+static int eb_request_add(struct i915_execbuffer *eb, struct i915_request *rq,
+			  int err, bool last_parallel)
+{
+	struct intel_timeline * const tl = i915_request_timeline(rq);
+	struct i915_sched_attr attr = {};
+	struct i915_request *prev;
+
+	lockdep_assert_held(&tl->mutex);
+	lockdep_unpin_lock(&tl->mutex, rq->cookie);
+
+	trace_i915_request_add(rq);
+
+	prev = __i915_request_commit(rq);
+
+	/* Check that the context wasn't destroyed before submission */
+	if (likely(!intel_context_is_closed(eb->context))) {
+		attr = eb->gem_context->sched;
+	} else {
+		/* Serialise with context_close via the add_to_timeline */
+		i915_request_set_error_once(rq, -ENOENT);
+		__i915_request_skip(rq);
+		err = -ENOENT; /* override any transient errors */
+	}
+
+	if (intel_context_is_parallel(eb->context)) {
+		if (err) {
+			__i915_request_skip(rq);
+			set_bit(I915_FENCE_FLAG_SKIP_PARALLEL,
+				&rq->fence.flags);
+		}
+		if (last_parallel)
+			set_bit(I915_FENCE_FLAG_SUBMIT_PARALLEL,
+				&rq->fence.flags);
+	}
+
+	__i915_request_queue(rq, &attr);
+
+	/* Try to clean up the client's timeline after submitting the request */
+	if (prev)
+		retire_requests(tl, prev);
+
+	mutex_unlock(&tl->mutex);
+
+	return err;
+}
+
+static int eb_requests_add(struct i915_execbuffer *eb, int err)
+{
+	int i;
+
+	/*
+	 * We iterate in reverse order of creation to release timeline mutexes in
+	 * same order.
+	 */
+	for_each_batch_add_order(eb, i) {
+		struct i915_request *rq = eb->requests[i];
+
+		if (!rq)
+			continue;
+		err |= eb_request_add(eb, rq, err, i == 0);
+	}
+
+	return err;
+}
+
+static const i915_user_extension_fn execbuf_extensions[] = {
+	[DRM_I915_GEM_EXECBUFFER_EXT_TIMELINE_FENCES] = parse_timeline_fences,
+};
+
+static int
+parse_execbuf2_extensions(struct drm_i915_gem_execbuffer2 *args,
+			  struct i915_execbuffer *eb)
+{
+	if (!(args->flags & I915_EXEC_USE_EXTENSIONS))
+		return 0;
+
+	/* The execbuf2 extension mechanism reuses cliprects_ptr. So we cannot
+	 * have another flag also using it at the same time.
+	 */
+	if (eb->args->flags & I915_EXEC_FENCE_ARRAY)
+		return -EINVAL;
+
+	if (args->num_cliprects != 0)
+		return -EINVAL;
+
+	return i915_user_extensions(u64_to_user_ptr(args->cliprects_ptr),
+				    execbuf_extensions,
+				    ARRAY_SIZE(execbuf_extensions),
+				    eb);
+}
+
+static void eb_requests_get(struct i915_execbuffer *eb)
+{
+	unsigned int i;
+
+	for_each_batch_create_order(eb, i) {
+		if (!eb->requests[i])
+			break;
+
+		i915_request_get(eb->requests[i]);
+	}
+}
+
+static void eb_requests_put(struct i915_execbuffer *eb)
+{
+	unsigned int i;
+
+	for_each_batch_create_order(eb, i) {
+		if (!eb->requests[i])
+			break;
+
+		i915_request_put(eb->requests[i]);
+	}
+}
+
+static struct sync_file *
+eb_composite_fence_create(struct i915_execbuffer *eb, int out_fence_fd)
+{
+	struct sync_file *out_fence = NULL;
+	struct dma_fence_array *fence_array;
+	struct dma_fence **fences;
+	unsigned int i;
+
+	GEM_BUG_ON(!intel_context_is_parent(eb->context));
+
+	fences = kmalloc_array(eb->num_batches, sizeof(*fences), GFP_KERNEL);
+	if (!fences)
+		return ERR_PTR(-ENOMEM);
+
+	for_each_batch_create_order(eb, i) {
+		fences[i] = &eb->requests[i]->fence;
+		__set_bit(I915_FENCE_FLAG_COMPOSITE,
+			  &eb->requests[i]->fence.flags);
+	}
+
+	fence_array = dma_fence_array_create(eb->num_batches,
+					     fences,
+					     eb->context->parallel.fence_context,
+					     eb->context->parallel.seqno++,
+					     false);
+	if (!fence_array) {
+		kfree(fences);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	/* Move ownership to the dma_fence_array created above */
+	for_each_batch_create_order(eb, i)
+		dma_fence_get(fences[i]);
+
+	if (out_fence_fd != -1) {
+		out_fence = sync_file_create(&fence_array->base);
+		/* sync_file now owns fence_arry, drop creation ref */
+		dma_fence_put(&fence_array->base);
+		if (!out_fence)
+			return ERR_PTR(-ENOMEM);
+	}
+
+	eb->composite_fence = &fence_array->base;
+
+	return out_fence;
+}
+
+static struct sync_file *
+eb_fences_add(struct i915_execbuffer *eb, struct i915_request *rq,
+	      struct dma_fence *in_fence, int out_fence_fd)
+{
+	struct sync_file *out_fence = NULL;
+	int err;
+
+	if (unlikely(eb->gem_context->syncobj)) {
+		struct dma_fence *fence;
+
+		fence = drm_syncobj_fence_get(eb->gem_context->syncobj);
+		err = i915_request_await_dma_fence(rq, fence);
+		dma_fence_put(fence);
+		if (err)
+			return ERR_PTR(err);
+	}
+
+	if (in_fence) {
+		if (eb->args->flags & I915_EXEC_FENCE_SUBMIT)
+			err = i915_request_await_execution(rq, in_fence);
+		else
+			err = i915_request_await_dma_fence(rq, in_fence);
+		if (err < 0)
+			return ERR_PTR(err);
+	}
+
+	if (eb->fences) {
+		err = await_fence_array(eb, rq);
+		if (err)
+			return ERR_PTR(err);
+	}
+
+	if (intel_context_is_parallel(eb->context)) {
+		out_fence = eb_composite_fence_create(eb, out_fence_fd);
+		if (IS_ERR(out_fence))
+			return ERR_PTR(-ENOMEM);
+	} else if (out_fence_fd != -1) {
+		out_fence = sync_file_create(&rq->fence);
+		if (!out_fence)
+			return ERR_PTR(-ENOMEM);
+	}
+
+	return out_fence;
+}
+
+static struct intel_context *
+eb_find_context(struct i915_execbuffer *eb, unsigned int context_number)
+{
+	struct intel_context *child;
+
+	if (likely(context_number == 0))
+		return eb->context;
+
+	for_each_child(eb->context, child)
+		if (!--context_number)
+			return child;
+
+	GEM_BUG_ON("Context not found");
+
+	return NULL;
+}
+
+static struct sync_file *
+eb_requests_create(struct i915_execbuffer *eb, struct dma_fence *in_fence,
+		   int out_fence_fd)
+{
+	struct sync_file *out_fence = NULL;
+	unsigned int i;
+
+	for_each_batch_create_order(eb, i) {
+		/* Allocate a request for this batch buffer nice and early. */
+		eb->requests[i] = i915_request_create(eb_find_context(eb, i));
+		if (IS_ERR(eb->requests[i])) {
+			out_fence = ERR_CAST(eb->requests[i]);
+			eb->requests[i] = NULL;
+			return out_fence;
+		}
+
+		/*
+		 * Only the first request added (committed to backend) has to
+		 * take the in fences into account as all subsequent requests
+		 * will have fences inserted inbetween them.
+		 */
+		if (i + 1 == eb->num_batches) {
+			out_fence = eb_fences_add(eb, eb->requests[i],
+						  in_fence, out_fence_fd);
+			if (IS_ERR(out_fence))
+				return out_fence;
+		}
+
+		/*
+		 * Not really on stack, but we don't want to call
+		 * kfree on the batch_snapshot when we put it, so use the
+		 * _onstack interface.
+		 */
+		if (eb->batches[i]->vma)
+			eb->requests[i]->batch_res =
+				i915_vma_resource_get(eb->batches[i]->vma->resource);
+		if (eb->batch_pool) {
+			GEM_BUG_ON(intel_context_is_parallel(eb->context));
+			intel_gt_buffer_pool_mark_active(eb->batch_pool,
+							 eb->requests[i]);
+		}
+	}
+
+	return out_fence;
+}
+
+static int
+i915_gem_do_execbuffer(struct drm_device *dev,
+		       struct drm_file *file,
+		       struct drm_i915_gem_execbuffer2 *args,
+		       struct drm_i915_gem_exec_object2 *exec)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct i915_execbuffer eb;
+	struct dma_fence *in_fence = NULL;
+	struct sync_file *out_fence = NULL;
+	int out_fence_fd = -1;
+	int err;
+
+	BUILD_BUG_ON(__EXEC_INTERNAL_FLAGS & ~__I915_EXEC_ILLEGAL_FLAGS);
+	BUILD_BUG_ON(__EXEC_OBJECT_INTERNAL_FLAGS &
+		     ~__EXEC_OBJECT_UNKNOWN_FLAGS);
+
+	eb.i915 = i915;
+	eb.file = file;
+	eb.args = args;
+	if (DBG_FORCE_RELOC || !(args->flags & I915_EXEC_NO_RELOC))
+		args->flags |= __EXEC_HAS_RELOC;
+
+	eb.exec = exec;
+	eb.vma = (struct eb_vma *)(exec + args->buffer_count + 1);
+	eb.vma[0].vma = NULL;
+	eb.batch_pool = NULL;
+
+	eb.invalid_flags = __EXEC_OBJECT_UNKNOWN_FLAGS;
+	reloc_cache_init(&eb.reloc_cache, eb.i915);
+
+	eb.buffer_count = args->buffer_count;
+	eb.batch_start_offset = args->batch_start_offset;
+	eb.trampoline = NULL;
+
+	eb.fences = NULL;
+	eb.num_fences = 0;
+
+	eb_capture_list_clear(&eb);
+
+	memset(eb.requests, 0, sizeof(struct i915_request *) *
+	       ARRAY_SIZE(eb.requests));
+	eb.composite_fence = NULL;
+
+	eb.batch_flags = 0;
+	if (args->flags & I915_EXEC_SECURE) {
+		if (GRAPHICS_VER(i915) >= 11)
+			return -ENODEV;
+
+		/* Return -EPERM to trigger fallback code on old binaries. */
+		if (!HAS_SECURE_BATCHES(i915))
+			return -EPERM;
+
+		if (!drm_is_current_master(file) || !capable(CAP_SYS_ADMIN))
+			return -EPERM;
+
+		eb.batch_flags |= I915_DISPATCH_SECURE;
+	}
+	if (args->flags & I915_EXEC_IS_PINNED)
+		eb.batch_flags |= I915_DISPATCH_PINNED;
+
+	err = parse_execbuf2_extensions(args, &eb);
+	if (err)
+		goto err_ext;
+
+	err = add_fence_array(&eb);
+	if (err)
+		goto err_ext;
+
+#define IN_FENCES (I915_EXEC_FENCE_IN | I915_EXEC_FENCE_SUBMIT)
+	if (args->flags & IN_FENCES) {
+		if ((args->flags & IN_FENCES) == IN_FENCES)
+			return -EINVAL;
+
+		in_fence = sync_file_get_fence(lower_32_bits(args->rsvd2));
+		if (!in_fence) {
+			err = -EINVAL;
+			goto err_ext;
+		}
+	}
+#undef IN_FENCES
+
+	if (args->flags & I915_EXEC_FENCE_OUT) {
+		out_fence_fd = get_unused_fd_flags(O_CLOEXEC);
+		if (out_fence_fd < 0) {
+			err = out_fence_fd;
+			goto err_in_fence;
+		}
+	}
+
+	err = eb_create(&eb);
+	if (err)
+		goto err_out_fence;
+
+	GEM_BUG_ON(!eb.lut_size);
+
+	err = eb_select_context(&eb);
+	if (unlikely(err))
+		goto err_destroy;
+
+	err = eb_select_engine(&eb);
+	if (unlikely(err))
+		goto err_context;
+
+	err = eb_lookup_vmas(&eb);
+	if (err) {
+		eb_release_vmas(&eb, true);
+		goto err_engine;
+	}
+
+	i915_gem_ww_ctx_init(&eb.ww, true);
+
+	err = eb_relocate_parse(&eb);
+	if (err) {
+		/*
+		 * If the user expects the execobject.offset and
+		 * reloc.presumed_offset to be an exact match,
+		 * as for using NO_RELOC, then we cannot update
+		 * the execobject.offset until we have completed
+		 * relocation.
+		 */
+		args->flags &= ~__EXEC_HAS_RELOC;
+		goto err_vma;
+	}
+
+	ww_acquire_done(&eb.ww.ctx);
+	err = eb_capture_stage(&eb);
+	if (err)
+		goto err_vma;
+
+	out_fence = eb_requests_create(&eb, in_fence, out_fence_fd);
+	if (IS_ERR(out_fence)) {
+		err = PTR_ERR(out_fence);
+		out_fence = NULL;
+		if (eb.requests[0])
+			goto err_request;
+		else
+			goto err_vma;
+	}
+
+	err = eb_submit(&eb);
+
+err_request:
+	eb_requests_get(&eb);
+	err = eb_requests_add(&eb, err);
+
+	if (eb.fences)
+		signal_fence_array(&eb, eb.composite_fence ?
+				   eb.composite_fence :
+				   &eb.requests[0]->fence);
+
+	if (unlikely(eb.gem_context->syncobj)) {
+		drm_syncobj_replace_fence(eb.gem_context->syncobj,
+					  eb.composite_fence ?
+					  eb.composite_fence :
+					  &eb.requests[0]->fence);
+	}
+
+	if (out_fence) {
+		if (err == 0) {
+			fd_install(out_fence_fd, out_fence->file);
+			args->rsvd2 &= GENMASK_ULL(31, 0); /* keep in-fence */
+			args->rsvd2 |= (u64)out_fence_fd << 32;
+			out_fence_fd = -1;
+		} else {
+			fput(out_fence->file);
+		}
+	}
+
+	if (!out_fence && eb.composite_fence)
+		dma_fence_put(eb.composite_fence);
+
+	eb_requests_put(&eb);
+
+err_vma:
+	eb_release_vmas(&eb, true);
+	WARN_ON(err == -EDEADLK);
+	i915_gem_ww_ctx_fini(&eb.ww);
+
+	if (eb.batch_pool)
+		intel_gt_buffer_pool_put(eb.batch_pool);
+err_engine:
+	eb_put_engine(&eb);
+err_context:
+	i915_gem_context_put(eb.gem_context);
+err_destroy:
+	eb_destroy(&eb);
+err_out_fence:
+	if (out_fence_fd != -1)
+		put_unused_fd(out_fence_fd);
+err_in_fence:
+	dma_fence_put(in_fence);
+err_ext:
+	put_fence_array(eb.fences, eb.num_fences);
+	return err;
+}
+
+static size_t eb_element_size(void)
+{
+	return sizeof(struct drm_i915_gem_exec_object2) + sizeof(struct eb_vma);
+}
+
+static bool check_buffer_count(size_t count)
+{
+	const size_t sz = eb_element_size();
+
+	/*
+	 * When using LUT_HANDLE, we impose a limit of INT_MAX for the lookup
+	 * array size (see eb_create()). Otherwise, we can accept an array as
+	 * large as can be addressed (though use large arrays at your peril)!
+	 */
+
+	return !(count < 1 || count > INT_MAX || count > SIZE_MAX / sz - 1);
+}
+
+int
+i915_gem_execbuffer2_ioctl(struct drm_device *dev, void *data,
+			   struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_execbuffer2 *args = data;
+	struct drm_i915_gem_exec_object2 *exec2_list;
+	const size_t count = args->buffer_count;
+	int err;
+
+	if (!check_buffer_count(count)) {
+		drm_dbg(&i915->drm, "execbuf2 with %zd buffers\n", count);
+		return -EINVAL;
+	}
+
+	err = i915_gem_check_execbuffer(args);
+	if (err)
+		return err;
+
+	/* Allocate extra slots for use by the command parser */
+	exec2_list = kvmalloc_array(count + 2, eb_element_size(),
+				    __GFP_NOWARN | GFP_KERNEL);
+	if (exec2_list == NULL) {
+		drm_dbg(&i915->drm, "Failed to allocate exec list for %zd buffers\n",
+			count);
+		return -ENOMEM;
+	}
+	if (copy_from_user(exec2_list,
+			   u64_to_user_ptr(args->buffers_ptr),
+			   sizeof(*exec2_list) * count)) {
+		drm_dbg(&i915->drm, "copy %zd exec entries failed\n", count);
+		kvfree(exec2_list);
+		return -EFAULT;
+	}
+
+	err = i915_gem_do_execbuffer(dev, file, args, exec2_list);
+
+	/*
+	 * Now that we have begun execution of the batchbuffer, we ignore
+	 * any new error after this point. Also given that we have already
+	 * updated the associated relocations, we try to write out the current
+	 * object locations irrespective of any error.
+	 */
+	if (args->flags & __EXEC_HAS_RELOC) {
+		struct drm_i915_gem_exec_object2 __user *user_exec_list =
+			u64_to_user_ptr(args->buffers_ptr);
+		unsigned int i;
+
+		/* Copy the new buffer offsets back to the user's exec list. */
+		/*
+		 * Note: count * sizeof(*user_exec_list) does not overflow,
+		 * because we checked 'count' in check_buffer_count().
+		 *
+		 * And this range already got effectively checked earlier
+		 * when we did the "copy_from_user()" above.
+		 */
+		if (!user_write_access_begin(user_exec_list,
+					     count * sizeof(*user_exec_list)))
+			goto end;
+
+		for (i = 0; i < args->buffer_count; i++) {
+			if (!(exec2_list[i].offset & UPDATE))
+				continue;
+
+			exec2_list[i].offset =
+				gen8_canonical_addr(exec2_list[i].offset & PIN_OFFSET_MASK);
+			unsafe_put_user(exec2_list[i].offset,
+					&user_exec_list[i].offset,
+					end_user);
+		}
+end_user:
+		user_write_access_end();
+end:;
+	}
+
+	args->flags &= ~__I915_EXEC_UNKNOWN_FLAGS;
+	kvfree(exec2_list);
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_internal.c b/drivers/gpu/drm/i915/gem/i915_gem_internal.c
new file mode 100644
index 000000000..629acb403
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_internal.c
@@ -0,0 +1,199 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2014-2016 Intel Corporation
+ */
+
+#include <linux/scatterlist.h>
+#include <linux/slab.h>
+
+#include "i915_drv.h"
+#include "i915_gem.h"
+#include "i915_gem_internal.h"
+#include "i915_gem_object.h"
+#include "i915_scatterlist.h"
+#include "i915_utils.h"
+
+#define QUIET (__GFP_NORETRY | __GFP_NOWARN)
+#define MAYFAIL (__GFP_RETRY_MAYFAIL | __GFP_NOWARN)
+
+static void internal_free_pages(struct sg_table *st)
+{
+	struct scatterlist *sg;
+
+	for (sg = st->sgl; sg; sg = __sg_next(sg)) {
+		if (sg_page(sg))
+			__free_pages(sg_page(sg), get_order(sg->length));
+	}
+
+	sg_free_table(st);
+	kfree(st);
+}
+
+static int i915_gem_object_get_pages_internal(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct sg_table *st;
+	struct scatterlist *sg;
+	unsigned int sg_page_sizes;
+	unsigned int npages;
+	int max_order = MAX_ORDER;
+	unsigned int max_segment;
+	gfp_t gfp;
+
+	max_segment = i915_sg_segment_size(i915->drm.dev) >> PAGE_SHIFT;
+	max_order = min(max_order, get_order(max_segment));
+
+	gfp = GFP_KERNEL | __GFP_HIGHMEM | __GFP_RECLAIMABLE;
+	if (IS_I965GM(i915) || IS_I965G(i915)) {
+		/* 965gm cannot relocate objects above 4GiB. */
+		gfp &= ~__GFP_HIGHMEM;
+		gfp |= __GFP_DMA32;
+	}
+
+create_st:
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (!st)
+		return -ENOMEM;
+
+	npages = obj->base.size / PAGE_SIZE;
+	if (sg_alloc_table(st, npages, GFP_KERNEL)) {
+		kfree(st);
+		return -ENOMEM;
+	}
+
+	sg = st->sgl;
+	st->nents = 0;
+	sg_page_sizes = 0;
+
+	do {
+		int order = min(fls(npages) - 1, max_order);
+		struct page *page;
+
+		do {
+			page = alloc_pages(gfp | (order ? QUIET : MAYFAIL),
+					   order);
+			if (page)
+				break;
+			if (!order--)
+				goto err;
+
+			/* Limit subsequent allocations as well */
+			max_order = order;
+		} while (1);
+
+		sg_set_page(sg, page, PAGE_SIZE << order, 0);
+		sg_page_sizes |= PAGE_SIZE << order;
+		st->nents++;
+
+		npages -= 1 << order;
+		if (!npages) {
+			sg_mark_end(sg);
+			break;
+		}
+
+		sg = __sg_next(sg);
+	} while (1);
+
+	if (i915_gem_gtt_prepare_pages(obj, st)) {
+		/* Failed to dma-map try again with single page sg segments */
+		if (get_order(st->sgl->length)) {
+			internal_free_pages(st);
+			max_order = 0;
+			goto create_st;
+		}
+		goto err;
+	}
+
+	__i915_gem_object_set_pages(obj, st, sg_page_sizes);
+
+	return 0;
+
+err:
+	sg_set_page(sg, NULL, 0, 0);
+	sg_mark_end(sg);
+	internal_free_pages(st);
+
+	return -ENOMEM;
+}
+
+static void i915_gem_object_put_pages_internal(struct drm_i915_gem_object *obj,
+					       struct sg_table *pages)
+{
+	i915_gem_gtt_finish_pages(obj, pages);
+	internal_free_pages(pages);
+
+	obj->mm.dirty = false;
+
+	__start_cpu_write(obj);
+}
+
+static const struct drm_i915_gem_object_ops i915_gem_object_internal_ops = {
+	.name = "i915_gem_object_internal",
+	.flags = I915_GEM_OBJECT_IS_SHRINKABLE,
+	.get_pages = i915_gem_object_get_pages_internal,
+	.put_pages = i915_gem_object_put_pages_internal,
+};
+
+struct drm_i915_gem_object *
+__i915_gem_object_create_internal(struct drm_i915_private *i915,
+				  const struct drm_i915_gem_object_ops *ops,
+				  phys_addr_t size)
+{
+	static struct lock_class_key lock_class;
+	struct drm_i915_gem_object *obj;
+	unsigned int cache_level;
+
+	GEM_BUG_ON(!size);
+	GEM_BUG_ON(!IS_ALIGNED(size, PAGE_SIZE));
+
+	if (overflows_type(size, obj->base.size))
+		return ERR_PTR(-E2BIG);
+
+	obj = i915_gem_object_alloc();
+	if (!obj)
+		return ERR_PTR(-ENOMEM);
+
+	drm_gem_private_object_init(&i915->drm, &obj->base, size);
+	i915_gem_object_init(obj, ops, &lock_class, 0);
+	obj->mem_flags |= I915_BO_FLAG_STRUCT_PAGE;
+
+	/*
+	 * Mark the object as volatile, such that the pages are marked as
+	 * dontneed whilst they are still pinned. As soon as they are unpinned
+	 * they are allowed to be reaped by the shrinker, and the caller is
+	 * expected to repopulate - the contents of this object are only valid
+	 * whilst active and pinned.
+	 */
+	i915_gem_object_set_volatile(obj);
+
+	obj->read_domains = I915_GEM_DOMAIN_CPU;
+	obj->write_domain = I915_GEM_DOMAIN_CPU;
+
+	cache_level = HAS_LLC(i915) ? I915_CACHE_LLC : I915_CACHE_NONE;
+	i915_gem_object_set_cache_coherency(obj, cache_level);
+
+	return obj;
+}
+
+/**
+ * i915_gem_object_create_internal: create an object with volatile pages
+ * @i915: the i915 device
+ * @size: the size in bytes of backing storage to allocate for the object
+ *
+ * Creates a new object that wraps some internal memory for private use.
+ * This object is not backed by swappable storage, and as such its contents
+ * are volatile and only valid whilst pinned. If the object is reaped by the
+ * shrinker, its pages and data will be discarded. Equally, it is not a full
+ * GEM object and so not valid for access from userspace. This makes it useful
+ * for hardware interfaces like ringbuffers (which are pinned from the time
+ * the request is written to the time the hardware stops accessing it), but
+ * not for contexts (which need to be preserved when not active for later
+ * reuse). Note that it is not cleared upon allocation.
+ */
+struct drm_i915_gem_object *
+i915_gem_object_create_internal(struct drm_i915_private *i915,
+				phys_addr_t size)
+{
+	return __i915_gem_object_create_internal(i915, &i915_gem_object_internal_ops, size);
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_internal.h b/drivers/gpu/drm/i915/gem/i915_gem_internal.h
new file mode 100644
index 000000000..6664e0611
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_internal.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __I915_GEM_INTERNAL_H__
+#define __I915_GEM_INTERNAL_H__
+
+#include <linux/types.h>
+
+struct drm_i915_gem_object;
+struct drm_i915_gem_object_ops;
+struct drm_i915_private;
+
+struct drm_i915_gem_object *
+i915_gem_object_create_internal(struct drm_i915_private *i915,
+				phys_addr_t size);
+struct drm_i915_gem_object *
+__i915_gem_object_create_internal(struct drm_i915_private *i915,
+				  const struct drm_i915_gem_object_ops *ops,
+				  phys_addr_t size);
+
+#endif /* __I915_GEM_INTERNAL_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h b/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h
new file mode 100644
index 000000000..28d6526e3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ioctls.h
@@ -0,0 +1,52 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef I915_GEM_IOCTLS_H
+#define I915_GEM_IOCTLS_H
+
+struct drm_device;
+struct drm_file;
+
+int i915_gem_busy_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file);
+int i915_gem_create_ioctl(struct drm_device *dev, void *data,
+			  struct drm_file *file);
+int i915_gem_create_ext_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *file);
+int i915_gem_execbuffer2_ioctl(struct drm_device *dev, void *data,
+			       struct drm_file *file);
+int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
+				struct drm_file *file);
+int i915_gem_get_caching_ioctl(struct drm_device *dev, void *data,
+			       struct drm_file *file);
+int i915_gem_get_tiling_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *file);
+int i915_gem_madvise_ioctl(struct drm_device *dev, void *data,
+			   struct drm_file *file);
+int i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file);
+int i915_gem_mmap_offset_ioctl(struct drm_device *dev, void *data,
+			       struct drm_file *file);
+int i915_gem_pread_ioctl(struct drm_device *dev, void *data,
+			 struct drm_file *file);
+int i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
+			  struct drm_file *file);
+int i915_gem_set_caching_ioctl(struct drm_device *dev, void *data,
+			       struct drm_file *file);
+int i915_gem_set_domain_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *file);
+int i915_gem_set_tiling_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *file);
+int i915_gem_sw_finish_ioctl(struct drm_device *dev, void *data,
+			     struct drm_file *file);
+int i915_gem_throttle_ioctl(struct drm_device *dev, void *data,
+			    struct drm_file *file);
+int i915_gem_userptr_ioctl(struct drm_device *dev, void *data,
+			   struct drm_file *file);
+int i915_gem_wait_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_lmem.c b/drivers/gpu/drm/i915/gem/i915_gem_lmem.c
new file mode 100644
index 000000000..3198b64ad
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_lmem.c
@@ -0,0 +1,143 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <uapi/drm/i915_drm.h>
+
+#include "intel_memory_region.h"
+#include "gem/i915_gem_region.h"
+#include "gem/i915_gem_lmem.h"
+#include "i915_drv.h"
+
+void __iomem *
+i915_gem_object_lmem_io_map(struct drm_i915_gem_object *obj,
+			    unsigned long n,
+			    unsigned long size)
+{
+	resource_size_t offset;
+
+	GEM_BUG_ON(!i915_gem_object_is_contiguous(obj));
+
+	offset = i915_gem_object_get_dma_address(obj, n);
+	offset -= obj->mm.region->region.start;
+
+	return io_mapping_map_wc(&obj->mm.region->iomap, offset, size);
+}
+
+/**
+ * i915_gem_object_is_lmem - Whether the object is resident in
+ * lmem
+ * @obj: The object to check.
+ *
+ * Even if an object is allowed to migrate and change memory region,
+ * this function checks whether it will always be present in lmem when
+ * valid *or* if that's not the case, whether it's currently resident in lmem.
+ * For migratable and evictable objects, the latter only makes sense when
+ * the object is locked.
+ *
+ * Return: Whether the object migratable but resident in lmem, or not
+ * migratable and will be present in lmem when valid.
+ */
+bool i915_gem_object_is_lmem(struct drm_i915_gem_object *obj)
+{
+	struct intel_memory_region *mr = READ_ONCE(obj->mm.region);
+
+#ifdef CONFIG_LOCKDEP
+	if (i915_gem_object_migratable(obj) &&
+	    i915_gem_object_evictable(obj))
+		assert_object_held(obj);
+#endif
+	return mr && (mr->type == INTEL_MEMORY_LOCAL ||
+		      mr->type == INTEL_MEMORY_STOLEN_LOCAL);
+}
+
+/**
+ * __i915_gem_object_is_lmem - Whether the object is resident in
+ * lmem while in the fence signaling critical path.
+ * @obj: The object to check.
+ *
+ * This function is intended to be called from within the fence signaling
+ * path where the fence, or a pin, keeps the object from being migrated. For
+ * example during gpu reset or similar.
+ *
+ * Return: Whether the object is resident in lmem.
+ */
+bool __i915_gem_object_is_lmem(struct drm_i915_gem_object *obj)
+{
+	struct intel_memory_region *mr = READ_ONCE(obj->mm.region);
+
+#ifdef CONFIG_LOCKDEP
+	GEM_WARN_ON(dma_resv_test_signaled(obj->base.resv, DMA_RESV_USAGE_BOOKKEEP) &&
+		    i915_gem_object_evictable(obj));
+#endif
+	return mr && (mr->type == INTEL_MEMORY_LOCAL ||
+		      mr->type == INTEL_MEMORY_STOLEN_LOCAL);
+}
+
+/**
+ * __i915_gem_object_create_lmem_with_ps - Create lmem object and force the
+ * minimum page size for the backing pages.
+ * @i915: The i915 instance.
+ * @size: The size in bytes for the object. Note that we need to round the size
+ * up depending on the @page_size. The final object size can be fished out from
+ * the drm GEM object.
+ * @page_size: The requested minimum page size in bytes for this object. This is
+ * useful if we need something bigger than the regions min_page_size due to some
+ * hw restriction, or in some very specialised cases where it needs to be
+ * smaller, where the internal fragmentation cost is too great when rounding up
+ * the object size.
+ * @flags: The optional BO allocation flags.
+ *
+ * Note that this interface assumes you know what you are doing when forcing the
+ * @page_size. If this is smaller than the regions min_page_size then it can
+ * never be inserted into any GTT, otherwise it might lead to undefined
+ * behaviour.
+ *
+ * Return: The object pointer, which might be an ERR_PTR in the case of failure.
+ */
+struct drm_i915_gem_object *
+__i915_gem_object_create_lmem_with_ps(struct drm_i915_private *i915,
+				      resource_size_t size,
+				      resource_size_t page_size,
+				      unsigned int flags)
+{
+	return i915_gem_object_create_region(i915->mm.regions[INTEL_REGION_LMEM_0],
+					     size, page_size, flags);
+}
+
+struct drm_i915_gem_object *
+i915_gem_object_create_lmem_from_data(struct drm_i915_private *i915,
+				      const void *data, size_t size)
+{
+	struct drm_i915_gem_object *obj;
+	void *map;
+
+	obj = i915_gem_object_create_lmem(i915,
+					  round_up(size, PAGE_SIZE),
+					  I915_BO_ALLOC_CONTIGUOUS);
+	if (IS_ERR(obj))
+		return obj;
+
+	map = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(map)) {
+		i915_gem_object_put(obj);
+		return map;
+	}
+
+	memcpy(map, data, size);
+
+	i915_gem_object_flush_map(obj);
+	__i915_gem_object_release_map(obj);
+
+	return obj;
+}
+
+struct drm_i915_gem_object *
+i915_gem_object_create_lmem(struct drm_i915_private *i915,
+			    resource_size_t size,
+			    unsigned int flags)
+{
+	return i915_gem_object_create_region(i915->mm.regions[INTEL_REGION_LMEM_0],
+					     size, 0, flags);
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_lmem.h b/drivers/gpu/drm/i915/gem/i915_gem_lmem.h
new file mode 100644
index 000000000..5a7a14e85
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_lmem.h
@@ -0,0 +1,38 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_GEM_LMEM_H
+#define __I915_GEM_LMEM_H
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct drm_i915_gem_object;
+struct intel_memory_region;
+
+void __iomem *
+i915_gem_object_lmem_io_map(struct drm_i915_gem_object *obj,
+			    unsigned long n,
+			    unsigned long size);
+
+bool i915_gem_object_is_lmem(struct drm_i915_gem_object *obj);
+
+bool __i915_gem_object_is_lmem(struct drm_i915_gem_object *obj);
+
+struct drm_i915_gem_object *
+i915_gem_object_create_lmem_from_data(struct drm_i915_private *i915,
+				      const void *data, size_t size);
+
+struct drm_i915_gem_object *
+__i915_gem_object_create_lmem_with_ps(struct drm_i915_private *i915,
+				      resource_size_t size,
+				      resource_size_t page_size,
+				      unsigned int flags);
+struct drm_i915_gem_object *
+i915_gem_object_create_lmem(struct drm_i915_private *i915,
+			    resource_size_t size,
+			    unsigned int flags);
+
+#endif /* !__I915_GEM_LMEM_H */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/i915_gem_mman.c
new file mode 100644
index 000000000..d7e30d889
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.c
@@ -0,0 +1,1049 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2014-2016 Intel Corporation
+ */
+
+#include <linux/anon_inodes.h>
+#include <linux/mman.h>
+#include <linux/pfn_t.h>
+#include <linux/sizes.h>
+
+#include <drm/drm_cache.h>
+
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_requests.h"
+
+#include "i915_drv.h"
+#include "i915_gem_evict.h"
+#include "i915_gem_gtt.h"
+#include "i915_gem_ioctls.h"
+#include "i915_gem_object.h"
+#include "i915_gem_mman.h"
+#include "i915_mm.h"
+#include "i915_trace.h"
+#include "i915_user_extensions.h"
+#include "i915_gem_ttm.h"
+#include "i915_vma.h"
+
+static inline bool
+__vma_matches(struct vm_area_struct *vma, struct file *filp,
+	      unsigned long addr, unsigned long size)
+{
+	if (vma->vm_file != filp)
+		return false;
+
+	return vma->vm_start == addr &&
+	       (vma->vm_end - vma->vm_start) == PAGE_ALIGN(size);
+}
+
+/**
+ * i915_gem_mmap_ioctl - Maps the contents of an object, returning the address
+ *			 it is mapped to.
+ * @dev: drm device
+ * @data: ioctl data blob
+ * @file: drm file
+ *
+ * While the mapping holds a reference on the contents of the object, it doesn't
+ * imply a ref on the object itself.
+ *
+ * IMPORTANT:
+ *
+ * DRM driver writers who look a this function as an example for how to do GEM
+ * mmap support, please don't implement mmap support like here. The modern way
+ * to implement DRM mmap support is with an mmap offset ioctl (like
+ * i915_gem_mmap_gtt) and then using the mmap syscall on the DRM fd directly.
+ * That way debug tooling like valgrind will understand what's going on, hiding
+ * the mmap call in a driver private ioctl will break that. The i915 driver only
+ * does cpu mmaps this way because we didn't know better.
+ */
+int
+i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
+		    struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_mmap *args = data;
+	struct drm_i915_gem_object *obj;
+	unsigned long addr;
+
+	/*
+	 * mmap ioctl is disallowed for all discrete platforms,
+	 * and for all platforms with GRAPHICS_VER > 12.
+	 */
+	if (IS_DGFX(i915) || GRAPHICS_VER_FULL(i915) > IP_VER(12, 0))
+		return -EOPNOTSUPP;
+
+	if (args->flags & ~(I915_MMAP_WC))
+		return -EINVAL;
+
+	if (args->flags & I915_MMAP_WC && !pat_enabled())
+		return -ENODEV;
+
+	obj = i915_gem_object_lookup(file, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	/* prime objects have no backing filp to GEM mmap
+	 * pages from.
+	 */
+	if (!obj->base.filp) {
+		addr = -ENXIO;
+		goto err;
+	}
+
+	if (range_overflows(args->offset, args->size, (u64)obj->base.size)) {
+		addr = -EINVAL;
+		goto err;
+	}
+
+	addr = vm_mmap(obj->base.filp, 0, args->size,
+		       PROT_READ | PROT_WRITE, MAP_SHARED,
+		       args->offset);
+	if (IS_ERR_VALUE(addr))
+		goto err;
+
+	if (args->flags & I915_MMAP_WC) {
+		struct mm_struct *mm = current->mm;
+		struct vm_area_struct *vma;
+
+		if (mmap_write_lock_killable(mm)) {
+			addr = -EINTR;
+			goto err;
+		}
+		vma = find_vma(mm, addr);
+		if (vma && __vma_matches(vma, obj->base.filp, addr, args->size))
+			vma->vm_page_prot =
+				pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
+		else
+			addr = -ENOMEM;
+		mmap_write_unlock(mm);
+		if (IS_ERR_VALUE(addr))
+			goto err;
+	}
+	i915_gem_object_put(obj);
+
+	args->addr_ptr = (u64)addr;
+	return 0;
+
+err:
+	i915_gem_object_put(obj);
+	return addr;
+}
+
+static unsigned int tile_row_pages(const struct drm_i915_gem_object *obj)
+{
+	return i915_gem_object_get_tile_row_size(obj) >> PAGE_SHIFT;
+}
+
+/**
+ * i915_gem_mmap_gtt_version - report the current feature set for GTT mmaps
+ *
+ * A history of the GTT mmap interface:
+ *
+ * 0 - Everything had to fit into the GTT. Both parties of a memcpy had to
+ *     aligned and suitable for fencing, and still fit into the available
+ *     mappable space left by the pinned display objects. A classic problem
+ *     we called the page-fault-of-doom where we would ping-pong between
+ *     two objects that could not fit inside the GTT and so the memcpy
+ *     would page one object in at the expense of the other between every
+ *     single byte.
+ *
+ * 1 - Objects can be any size, and have any compatible fencing (X Y, or none
+ *     as set via i915_gem_set_tiling() [DRM_I915_GEM_SET_TILING]). If the
+ *     object is too large for the available space (or simply too large
+ *     for the mappable aperture!), a view is created instead and faulted
+ *     into userspace. (This view is aligned and sized appropriately for
+ *     fenced access.)
+ *
+ * 2 - Recognise WC as a separate cache domain so that we can flush the
+ *     delayed writes via GTT before performing direct access via WC.
+ *
+ * 3 - Remove implicit set-domain(GTT) and synchronisation on initial
+ *     pagefault; swapin remains transparent.
+ *
+ * 4 - Support multiple fault handlers per object depending on object's
+ *     backing storage (a.k.a. MMAP_OFFSET).
+ *
+ * Restrictions:
+ *
+ *  * snoopable objects cannot be accessed via the GTT. It can cause machine
+ *    hangs on some architectures, corruption on others. An attempt to service
+ *    a GTT page fault from a snoopable object will generate a SIGBUS.
+ *
+ *  * the object must be able to fit into RAM (physical memory, though no
+ *    limited to the mappable aperture).
+ *
+ *
+ * Caveats:
+ *
+ *  * a new GTT page fault will synchronize rendering from the GPU and flush
+ *    all data to system memory. Subsequent access will not be synchronized.
+ *
+ *  * all mappings are revoked on runtime device suspend.
+ *
+ *  * there are only 8, 16 or 32 fence registers to share between all users
+ *    (older machines require fence register for display and blitter access
+ *    as well). Contention of the fence registers will cause the previous users
+ *    to be unmapped and any new access will generate new page faults.
+ *
+ *  * running out of memory while servicing a fault may generate a SIGBUS,
+ *    rather than the expected SIGSEGV.
+ */
+int i915_gem_mmap_gtt_version(void)
+{
+	return 4;
+}
+
+static inline struct i915_gtt_view
+compute_partial_view(const struct drm_i915_gem_object *obj,
+		     pgoff_t page_offset,
+		     unsigned int chunk)
+{
+	struct i915_gtt_view view;
+
+	if (i915_gem_object_is_tiled(obj))
+		chunk = roundup(chunk, tile_row_pages(obj) ?: 1);
+
+	view.type = I915_GTT_VIEW_PARTIAL;
+	view.partial.offset = rounddown(page_offset, chunk);
+	view.partial.size =
+		min_t(unsigned int, chunk,
+		      (obj->base.size >> PAGE_SHIFT) - view.partial.offset);
+
+	/* If the partial covers the entire object, just create a normal VMA. */
+	if (chunk >= obj->base.size >> PAGE_SHIFT)
+		view.type = I915_GTT_VIEW_NORMAL;
+
+	return view;
+}
+
+static vm_fault_t i915_error_to_vmf_fault(int err)
+{
+	switch (err) {
+	default:
+		WARN_ONCE(err, "unhandled error in %s: %i\n", __func__, err);
+		fallthrough;
+	case -EIO: /* shmemfs failure from swap device */
+	case -EFAULT: /* purged object */
+	case -ENODEV: /* bad object, how did you get here! */
+	case -ENXIO: /* unable to access backing store (on device) */
+		return VM_FAULT_SIGBUS;
+
+	case -ENOMEM: /* our allocation failure */
+		return VM_FAULT_OOM;
+
+	case 0:
+	case -EAGAIN:
+	case -ENOSPC: /* transient failure to evict? */
+	case -ENOBUFS: /* temporarily out of fences? */
+	case -ERESTARTSYS:
+	case -EINTR:
+	case -EBUSY:
+		/*
+		 * EBUSY is ok: this just means that another thread
+		 * already did the job.
+		 */
+		return VM_FAULT_NOPAGE;
+	}
+}
+
+static vm_fault_t vm_fault_cpu(struct vm_fault *vmf)
+{
+	struct vm_area_struct *area = vmf->vma;
+	struct i915_mmap_offset *mmo = area->vm_private_data;
+	struct drm_i915_gem_object *obj = mmo->obj;
+	resource_size_t iomap;
+	int err;
+
+	/* Sanity check that we allow writing into this object */
+	if (unlikely(i915_gem_object_is_readonly(obj) &&
+		     area->vm_flags & VM_WRITE))
+		return VM_FAULT_SIGBUS;
+
+	if (i915_gem_object_lock_interruptible(obj, NULL))
+		return VM_FAULT_NOPAGE;
+
+	err = i915_gem_object_pin_pages(obj);
+	if (err)
+		goto out;
+
+	iomap = -1;
+	if (!i915_gem_object_has_struct_page(obj)) {
+		iomap = obj->mm.region->iomap.base;
+		iomap -= obj->mm.region->region.start;
+	}
+
+	/* PTEs are revoked in obj->ops->put_pages() */
+	err = remap_io_sg(area,
+			  area->vm_start, area->vm_end - area->vm_start,
+			  obj->mm.pages->sgl, iomap);
+
+	if (area->vm_flags & VM_WRITE) {
+		GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+		obj->mm.dirty = true;
+	}
+
+	i915_gem_object_unpin_pages(obj);
+
+out:
+	i915_gem_object_unlock(obj);
+	return i915_error_to_vmf_fault(err);
+}
+
+static vm_fault_t vm_fault_gtt(struct vm_fault *vmf)
+{
+#define MIN_CHUNK_PAGES (SZ_1M >> PAGE_SHIFT)
+	struct vm_area_struct *area = vmf->vma;
+	struct i915_mmap_offset *mmo = area->vm_private_data;
+	struct drm_i915_gem_object *obj = mmo->obj;
+	struct drm_device *dev = obj->base.dev;
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct intel_runtime_pm *rpm = &i915->runtime_pm;
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	bool write = area->vm_flags & VM_WRITE;
+	struct i915_gem_ww_ctx ww;
+	intel_wakeref_t wakeref;
+	struct i915_vma *vma;
+	pgoff_t page_offset;
+	int srcu;
+	int ret;
+
+	/* We don't use vmf->pgoff since that has the fake offset */
+	page_offset = (vmf->address - area->vm_start) >> PAGE_SHIFT;
+
+	trace_i915_gem_object_fault(obj, page_offset, true, write);
+
+	wakeref = intel_runtime_pm_get(rpm);
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	ret = i915_gem_object_lock(obj, &ww);
+	if (ret)
+		goto err_rpm;
+
+	/* Sanity check that we allow writing into this object */
+	if (i915_gem_object_is_readonly(obj) && write) {
+		ret = -EFAULT;
+		goto err_rpm;
+	}
+
+	ret = i915_gem_object_pin_pages(obj);
+	if (ret)
+		goto err_rpm;
+
+	ret = intel_gt_reset_trylock(ggtt->vm.gt, &srcu);
+	if (ret)
+		goto err_pages;
+
+	/* Now pin it into the GTT as needed */
+	vma = i915_gem_object_ggtt_pin_ww(obj, &ww, NULL, 0, 0,
+					  PIN_MAPPABLE |
+					  PIN_NONBLOCK /* NOWARN */ |
+					  PIN_NOEVICT);
+	if (IS_ERR(vma) && vma != ERR_PTR(-EDEADLK)) {
+		/* Use a partial view if it is bigger than available space */
+		struct i915_gtt_view view =
+			compute_partial_view(obj, page_offset, MIN_CHUNK_PAGES);
+		unsigned int flags;
+
+		flags = PIN_MAPPABLE | PIN_NOSEARCH;
+		if (view.type == I915_GTT_VIEW_NORMAL)
+			flags |= PIN_NONBLOCK; /* avoid warnings for pinned */
+
+		/*
+		 * Userspace is now writing through an untracked VMA, abandon
+		 * all hope that the hardware is able to track future writes.
+		 */
+
+		vma = i915_gem_object_ggtt_pin_ww(obj, &ww, &view, 0, 0, flags);
+		if (IS_ERR(vma) && vma != ERR_PTR(-EDEADLK)) {
+			flags = PIN_MAPPABLE;
+			view.type = I915_GTT_VIEW_PARTIAL;
+			vma = i915_gem_object_ggtt_pin_ww(obj, &ww, &view, 0, 0, flags);
+		}
+
+		/*
+		 * The entire mappable GGTT is pinned? Unexpected!
+		 * Try to evict the object we locked too, as normally we skip it
+		 * due to lack of short term pinning inside execbuf.
+		 */
+		if (vma == ERR_PTR(-ENOSPC)) {
+			ret = mutex_lock_interruptible(&ggtt->vm.mutex);
+			if (!ret) {
+				ret = i915_gem_evict_vm(&ggtt->vm, &ww, NULL);
+				mutex_unlock(&ggtt->vm.mutex);
+			}
+			if (ret)
+				goto err_reset;
+			vma = i915_gem_object_ggtt_pin_ww(obj, &ww, &view, 0, 0, flags);
+		}
+	}
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto err_reset;
+	}
+
+	/* Access to snoopable pages through the GTT is incoherent. */
+	if (obj->cache_level != I915_CACHE_NONE && !HAS_LLC(i915)) {
+		ret = -EFAULT;
+		goto err_unpin;
+	}
+
+	ret = i915_vma_pin_fence(vma);
+	if (ret)
+		goto err_unpin;
+
+	/* Finally, remap it using the new GTT offset */
+	ret = remap_io_mapping(area,
+			       area->vm_start + (vma->gtt_view.partial.offset << PAGE_SHIFT),
+			       (ggtt->gmadr.start + vma->node.start) >> PAGE_SHIFT,
+			       min_t(u64, vma->size, area->vm_end - area->vm_start),
+			       &ggtt->iomap);
+	if (ret)
+		goto err_fence;
+
+	assert_rpm_wakelock_held(rpm);
+
+	/* Mark as being mmapped into userspace for later revocation */
+	mutex_lock(&to_gt(i915)->ggtt->vm.mutex);
+	if (!i915_vma_set_userfault(vma) && !obj->userfault_count++)
+		list_add(&obj->userfault_link, &to_gt(i915)->ggtt->userfault_list);
+	mutex_unlock(&to_gt(i915)->ggtt->vm.mutex);
+
+	/* Track the mmo associated with the fenced vma */
+	vma->mmo = mmo;
+
+	if (CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND)
+		intel_wakeref_auto(&i915->runtime_pm.userfault_wakeref,
+				   msecs_to_jiffies_timeout(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND));
+
+	if (write) {
+		GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+		i915_vma_set_ggtt_write(vma);
+		obj->mm.dirty = true;
+	}
+
+err_fence:
+	i915_vma_unpin_fence(vma);
+err_unpin:
+	__i915_vma_unpin(vma);
+err_reset:
+	intel_gt_reset_unlock(ggtt->vm.gt, srcu);
+err_pages:
+	i915_gem_object_unpin_pages(obj);
+err_rpm:
+	if (ret == -EDEADLK) {
+		ret = i915_gem_ww_ctx_backoff(&ww);
+		if (!ret)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	intel_runtime_pm_put(rpm, wakeref);
+	return i915_error_to_vmf_fault(ret);
+}
+
+static int
+vm_access(struct vm_area_struct *area, unsigned long addr,
+	  void *buf, int len, int write)
+{
+	struct i915_mmap_offset *mmo = area->vm_private_data;
+	struct drm_i915_gem_object *obj = mmo->obj;
+	struct i915_gem_ww_ctx ww;
+	void *vaddr;
+	int err = 0;
+
+	if (i915_gem_object_is_readonly(obj) && write)
+		return -EACCES;
+
+	addr -= area->vm_start;
+	if (range_overflows_t(u64, addr, len, obj->base.size))
+		return -EINVAL;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	err = i915_gem_object_lock(obj, &ww);
+	if (err)
+		goto out;
+
+	/* As this is primarily for debugging, let's focus on simplicity */
+	vaddr = i915_gem_object_pin_map(obj, I915_MAP_FORCE_WC);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto out;
+	}
+
+	if (write) {
+		memcpy(vaddr + addr, buf, len);
+		__i915_gem_object_flush_map(obj, addr, len);
+	} else {
+		memcpy(buf, vaddr + addr, len);
+	}
+
+	i915_gem_object_unpin_map(obj);
+out:
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+
+	if (err)
+		return err;
+
+	return len;
+}
+
+void __i915_gem_object_release_mmap_gtt(struct drm_i915_gem_object *obj)
+{
+	struct i915_vma *vma;
+
+	GEM_BUG_ON(!obj->userfault_count);
+
+	for_each_ggtt_vma(vma, obj)
+		i915_vma_revoke_mmap(vma);
+
+	GEM_BUG_ON(obj->userfault_count);
+}
+
+/*
+ * It is vital that we remove the page mapping if we have mapped a tiled
+ * object through the GTT and then lose the fence register due to
+ * resource pressure. Similarly if the object has been moved out of the
+ * aperture, than pages mapped into userspace must be revoked. Removing the
+ * mapping will then trigger a page fault on the next user access, allowing
+ * fixup by vm_fault_gtt().
+ */
+void i915_gem_object_release_mmap_gtt(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	intel_wakeref_t wakeref;
+
+	/*
+	 * Serialisation between user GTT access and our code depends upon
+	 * revoking the CPU's PTE whilst the mutex is held. The next user
+	 * pagefault then has to wait until we release the mutex.
+	 *
+	 * Note that RPM complicates somewhat by adding an additional
+	 * requirement that operations to the GGTT be made holding the RPM
+	 * wakeref.
+	 */
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+	mutex_lock(&to_gt(i915)->ggtt->vm.mutex);
+
+	if (!obj->userfault_count)
+		goto out;
+
+	__i915_gem_object_release_mmap_gtt(obj);
+
+	/*
+	 * Ensure that the CPU's PTE are revoked and there are not outstanding
+	 * memory transactions from userspace before we return. The TLB
+	 * flushing implied above by changing the PTE above *should* be
+	 * sufficient, an extra barrier here just provides us with a bit
+	 * of paranoid documentation about our requirement to serialise
+	 * memory writes before touching registers / GSM.
+	 */
+	wmb();
+
+out:
+	mutex_unlock(&to_gt(i915)->ggtt->vm.mutex);
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+}
+
+void i915_gem_object_runtime_pm_release_mmap_offset(struct drm_i915_gem_object *obj)
+{
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+	struct ttm_device *bdev = bo->bdev;
+
+	drm_vma_node_unmap(&bo->base.vma_node, bdev->dev_mapping);
+
+	/*
+	 * We have exclusive access here via runtime suspend. All other callers
+	 * must first grab the rpm wakeref.
+	 */
+	GEM_BUG_ON(!obj->userfault_count);
+	list_del(&obj->userfault_link);
+	obj->userfault_count = 0;
+}
+
+void i915_gem_object_release_mmap_offset(struct drm_i915_gem_object *obj)
+{
+	struct i915_mmap_offset *mmo, *mn;
+
+	if (obj->ops->unmap_virtual)
+		obj->ops->unmap_virtual(obj);
+
+	spin_lock(&obj->mmo.lock);
+	rbtree_postorder_for_each_entry_safe(mmo, mn,
+					     &obj->mmo.offsets, offset) {
+		/*
+		 * vma_node_unmap for GTT mmaps handled already in
+		 * __i915_gem_object_release_mmap_gtt
+		 */
+		if (mmo->mmap_type == I915_MMAP_TYPE_GTT)
+			continue;
+
+		spin_unlock(&obj->mmo.lock);
+		drm_vma_node_unmap(&mmo->vma_node,
+				   obj->base.dev->anon_inode->i_mapping);
+		spin_lock(&obj->mmo.lock);
+	}
+	spin_unlock(&obj->mmo.lock);
+}
+
+static struct i915_mmap_offset *
+lookup_mmo(struct drm_i915_gem_object *obj,
+	   enum i915_mmap_type mmap_type)
+{
+	struct rb_node *rb;
+
+	spin_lock(&obj->mmo.lock);
+	rb = obj->mmo.offsets.rb_node;
+	while (rb) {
+		struct i915_mmap_offset *mmo =
+			rb_entry(rb, typeof(*mmo), offset);
+
+		if (mmo->mmap_type == mmap_type) {
+			spin_unlock(&obj->mmo.lock);
+			return mmo;
+		}
+
+		if (mmo->mmap_type < mmap_type)
+			rb = rb->rb_right;
+		else
+			rb = rb->rb_left;
+	}
+	spin_unlock(&obj->mmo.lock);
+
+	return NULL;
+}
+
+static struct i915_mmap_offset *
+insert_mmo(struct drm_i915_gem_object *obj, struct i915_mmap_offset *mmo)
+{
+	struct rb_node *rb, **p;
+
+	spin_lock(&obj->mmo.lock);
+	rb = NULL;
+	p = &obj->mmo.offsets.rb_node;
+	while (*p) {
+		struct i915_mmap_offset *pos;
+
+		rb = *p;
+		pos = rb_entry(rb, typeof(*pos), offset);
+
+		if (pos->mmap_type == mmo->mmap_type) {
+			spin_unlock(&obj->mmo.lock);
+			drm_vma_offset_remove(obj->base.dev->vma_offset_manager,
+					      &mmo->vma_node);
+			kfree(mmo);
+			return pos;
+		}
+
+		if (pos->mmap_type < mmo->mmap_type)
+			p = &rb->rb_right;
+		else
+			p = &rb->rb_left;
+	}
+	rb_link_node(&mmo->offset, rb, p);
+	rb_insert_color(&mmo->offset, &obj->mmo.offsets);
+	spin_unlock(&obj->mmo.lock);
+
+	return mmo;
+}
+
+static struct i915_mmap_offset *
+mmap_offset_attach(struct drm_i915_gem_object *obj,
+		   enum i915_mmap_type mmap_type,
+		   struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_mmap_offset *mmo;
+	int err;
+
+	GEM_BUG_ON(obj->ops->mmap_offset || obj->ops->mmap_ops);
+
+	mmo = lookup_mmo(obj, mmap_type);
+	if (mmo)
+		goto out;
+
+	mmo = kmalloc(sizeof(*mmo), GFP_KERNEL);
+	if (!mmo)
+		return ERR_PTR(-ENOMEM);
+
+	mmo->obj = obj;
+	mmo->mmap_type = mmap_type;
+	drm_vma_node_reset(&mmo->vma_node);
+
+	err = drm_vma_offset_add(obj->base.dev->vma_offset_manager,
+				 &mmo->vma_node, obj->base.size / PAGE_SIZE);
+	if (likely(!err))
+		goto insert;
+
+	/* Attempt to reap some mmap space from dead objects */
+	err = intel_gt_retire_requests_timeout(to_gt(i915), MAX_SCHEDULE_TIMEOUT,
+					       NULL);
+	if (err)
+		goto err;
+
+	i915_gem_drain_freed_objects(i915);
+	err = drm_vma_offset_add(obj->base.dev->vma_offset_manager,
+				 &mmo->vma_node, obj->base.size / PAGE_SIZE);
+	if (err)
+		goto err;
+
+insert:
+	mmo = insert_mmo(obj, mmo);
+	GEM_BUG_ON(lookup_mmo(obj, mmap_type) != mmo);
+out:
+	if (file)
+		drm_vma_node_allow_once(&mmo->vma_node, file);
+	return mmo;
+
+err:
+	kfree(mmo);
+	return ERR_PTR(err);
+}
+
+static int
+__assign_mmap_offset(struct drm_i915_gem_object *obj,
+		     enum i915_mmap_type mmap_type,
+		     u64 *offset, struct drm_file *file)
+{
+	struct i915_mmap_offset *mmo;
+
+	if (i915_gem_object_never_mmap(obj))
+		return -ENODEV;
+
+	if (obj->ops->mmap_offset)  {
+		if (mmap_type != I915_MMAP_TYPE_FIXED)
+			return -ENODEV;
+
+		*offset = obj->ops->mmap_offset(obj);
+		return 0;
+	}
+
+	if (mmap_type == I915_MMAP_TYPE_FIXED)
+		return -ENODEV;
+
+	if (mmap_type != I915_MMAP_TYPE_GTT &&
+	    !i915_gem_object_has_struct_page(obj) &&
+	    !i915_gem_object_has_iomem(obj))
+		return -ENODEV;
+
+	mmo = mmap_offset_attach(obj, mmap_type, file);
+	if (IS_ERR(mmo))
+		return PTR_ERR(mmo);
+
+	*offset = drm_vma_node_offset_addr(&mmo->vma_node);
+	return 0;
+}
+
+static int
+__assign_mmap_offset_handle(struct drm_file *file,
+			    u32 handle,
+			    enum i915_mmap_type mmap_type,
+			    u64 *offset)
+{
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	obj = i915_gem_object_lookup(file, handle);
+	if (!obj)
+		return -ENOENT;
+
+	err = i915_gem_object_lock_interruptible(obj, NULL);
+	if (err)
+		goto out_put;
+	err = __assign_mmap_offset(obj, mmap_type, offset, file);
+	i915_gem_object_unlock(obj);
+out_put:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+int
+i915_gem_dumb_mmap_offset(struct drm_file *file,
+			  struct drm_device *dev,
+			  u32 handle,
+			  u64 *offset)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	enum i915_mmap_type mmap_type;
+
+	if (HAS_LMEM(to_i915(dev)))
+		mmap_type = I915_MMAP_TYPE_FIXED;
+	else if (pat_enabled())
+		mmap_type = I915_MMAP_TYPE_WC;
+	else if (!i915_ggtt_has_aperture(to_gt(i915)->ggtt))
+		return -ENODEV;
+	else
+		mmap_type = I915_MMAP_TYPE_GTT;
+
+	return __assign_mmap_offset_handle(file, handle, mmap_type, offset);
+}
+
+/**
+ * i915_gem_mmap_offset_ioctl - prepare an object for GTT mmap'ing
+ * @dev: DRM device
+ * @data: GTT mapping ioctl data
+ * @file: GEM object info
+ *
+ * Simply returns the fake offset to userspace so it can mmap it.
+ * The mmap call will end up in drm_gem_mmap(), which will set things
+ * up so we can get faults in the handler above.
+ *
+ * The fault handler will take care of binding the object into the GTT
+ * (since it may have been evicted to make room for something), allocating
+ * a fence register, and mapping the appropriate aperture address into
+ * userspace.
+ */
+int
+i915_gem_mmap_offset_ioctl(struct drm_device *dev, void *data,
+			   struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_mmap_offset *args = data;
+	enum i915_mmap_type type;
+	int err;
+
+	/*
+	 * Historically we failed to check args.pad and args.offset
+	 * and so we cannot use those fields for user input and we cannot
+	 * add -EINVAL for them as the ABI is fixed, i.e. old userspace
+	 * may be feeding in garbage in those fields.
+	 *
+	 * if (args->pad) return -EINVAL; is verbotten!
+	 */
+
+	err = i915_user_extensions(u64_to_user_ptr(args->extensions),
+				   NULL, 0, NULL);
+	if (err)
+		return err;
+
+	switch (args->flags) {
+	case I915_MMAP_OFFSET_GTT:
+		if (!i915_ggtt_has_aperture(to_gt(i915)->ggtt))
+			return -ENODEV;
+		type = I915_MMAP_TYPE_GTT;
+		break;
+
+	case I915_MMAP_OFFSET_WC:
+		if (!pat_enabled())
+			return -ENODEV;
+		type = I915_MMAP_TYPE_WC;
+		break;
+
+	case I915_MMAP_OFFSET_WB:
+		type = I915_MMAP_TYPE_WB;
+		break;
+
+	case I915_MMAP_OFFSET_UC:
+		if (!pat_enabled())
+			return -ENODEV;
+		type = I915_MMAP_TYPE_UC;
+		break;
+
+	case I915_MMAP_OFFSET_FIXED:
+		type = I915_MMAP_TYPE_FIXED;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return __assign_mmap_offset_handle(file, args->handle, type, &args->offset);
+}
+
+static void vm_open(struct vm_area_struct *vma)
+{
+	struct i915_mmap_offset *mmo = vma->vm_private_data;
+	struct drm_i915_gem_object *obj = mmo->obj;
+
+	GEM_BUG_ON(!obj);
+	i915_gem_object_get(obj);
+}
+
+static void vm_close(struct vm_area_struct *vma)
+{
+	struct i915_mmap_offset *mmo = vma->vm_private_data;
+	struct drm_i915_gem_object *obj = mmo->obj;
+
+	GEM_BUG_ON(!obj);
+	i915_gem_object_put(obj);
+}
+
+static const struct vm_operations_struct vm_ops_gtt = {
+	.fault = vm_fault_gtt,
+	.access = vm_access,
+	.open = vm_open,
+	.close = vm_close,
+};
+
+static const struct vm_operations_struct vm_ops_cpu = {
+	.fault = vm_fault_cpu,
+	.access = vm_access,
+	.open = vm_open,
+	.close = vm_close,
+};
+
+static int singleton_release(struct inode *inode, struct file *file)
+{
+	struct drm_i915_private *i915 = file->private_data;
+
+	cmpxchg(&i915->gem.mmap_singleton, file, NULL);
+	drm_dev_put(&i915->drm);
+
+	return 0;
+}
+
+static const struct file_operations singleton_fops = {
+	.owner = THIS_MODULE,
+	.release = singleton_release,
+};
+
+static struct file *mmap_singleton(struct drm_i915_private *i915)
+{
+	struct file *file;
+
+	rcu_read_lock();
+	file = READ_ONCE(i915->gem.mmap_singleton);
+	if (file && !get_file_rcu(file))
+		file = NULL;
+	rcu_read_unlock();
+	if (file)
+		return file;
+
+	file = anon_inode_getfile("i915.gem", &singleton_fops, i915, O_RDWR);
+	if (IS_ERR(file))
+		return file;
+
+	/* Everyone shares a single global address space */
+	file->f_mapping = i915->drm.anon_inode->i_mapping;
+
+	smp_store_mb(i915->gem.mmap_singleton, file);
+	drm_dev_get(&i915->drm);
+
+	return file;
+}
+
+/*
+ * This overcomes the limitation in drm_gem_mmap's assignment of a
+ * drm_gem_object as the vma->vm_private_data. Since we need to
+ * be able to resolve multiple mmap offsets which could be tied
+ * to a single gem object.
+ */
+int i915_gem_mmap(struct file *filp, struct vm_area_struct *vma)
+{
+	struct drm_vma_offset_node *node;
+	struct drm_file *priv = filp->private_data;
+	struct drm_device *dev = priv->minor->dev;
+	struct drm_i915_gem_object *obj = NULL;
+	struct i915_mmap_offset *mmo = NULL;
+	struct file *anon;
+
+	if (drm_dev_is_unplugged(dev))
+		return -ENODEV;
+
+	rcu_read_lock();
+	drm_vma_offset_lock_lookup(dev->vma_offset_manager);
+	node = drm_vma_offset_exact_lookup_locked(dev->vma_offset_manager,
+						  vma->vm_pgoff,
+						  vma_pages(vma));
+	if (node && drm_vma_node_is_allowed(node, priv)) {
+		/*
+		 * Skip 0-refcnted objects as it is in the process of being
+		 * destroyed and will be invalid when the vma manager lock
+		 * is released.
+		 */
+		if (!node->driver_private) {
+			mmo = container_of(node, struct i915_mmap_offset, vma_node);
+			obj = i915_gem_object_get_rcu(mmo->obj);
+
+			GEM_BUG_ON(obj && obj->ops->mmap_ops);
+		} else {
+			obj = i915_gem_object_get_rcu
+				(container_of(node, struct drm_i915_gem_object,
+					      base.vma_node));
+
+			GEM_BUG_ON(obj && !obj->ops->mmap_ops);
+		}
+	}
+	drm_vma_offset_unlock_lookup(dev->vma_offset_manager);
+	rcu_read_unlock();
+	if (!obj)
+		return node ? -EACCES : -EINVAL;
+
+	if (i915_gem_object_is_readonly(obj)) {
+		if (vma->vm_flags & VM_WRITE) {
+			i915_gem_object_put(obj);
+			return -EINVAL;
+		}
+		vma->vm_flags &= ~VM_MAYWRITE;
+	}
+
+	anon = mmap_singleton(to_i915(dev));
+	if (IS_ERR(anon)) {
+		i915_gem_object_put(obj);
+		return PTR_ERR(anon);
+	}
+
+	vma->vm_flags |= VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP | VM_IO;
+
+	/*
+	 * We keep the ref on mmo->obj, not vm_file, but we require
+	 * vma->vm_file->f_mapping, see vma_link(), for later revocation.
+	 * Our userspace is accustomed to having per-file resource cleanup
+	 * (i.e. contexts, objects and requests) on their close(fd), which
+	 * requires avoiding extraneous references to their filp, hence why
+	 * we prefer to use an anonymous file for their mmaps.
+	 */
+	vma_set_file(vma, anon);
+	/* Drop the initial creation reference, the vma is now holding one. */
+	fput(anon);
+
+	if (obj->ops->mmap_ops) {
+		vma->vm_page_prot = pgprot_decrypted(vm_get_page_prot(vma->vm_flags));
+		vma->vm_ops = obj->ops->mmap_ops;
+		vma->vm_private_data = node->driver_private;
+		return 0;
+	}
+
+	vma->vm_private_data = mmo;
+
+	switch (mmo->mmap_type) {
+	case I915_MMAP_TYPE_WC:
+		vma->vm_page_prot =
+			pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
+		vma->vm_ops = &vm_ops_cpu;
+		break;
+
+	case I915_MMAP_TYPE_FIXED:
+		GEM_WARN_ON(1);
+		fallthrough;
+	case I915_MMAP_TYPE_WB:
+		vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
+		vma->vm_ops = &vm_ops_cpu;
+		break;
+
+	case I915_MMAP_TYPE_UC:
+		vma->vm_page_prot =
+			pgprot_noncached(vm_get_page_prot(vma->vm_flags));
+		vma->vm_ops = &vm_ops_cpu;
+		break;
+
+	case I915_MMAP_TYPE_GTT:
+		vma->vm_page_prot =
+			pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
+		vma->vm_ops = &vm_ops_gtt;
+		break;
+	}
+	vma->vm_page_prot = pgprot_decrypted(vma->vm_page_prot);
+
+	return 0;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/i915_gem_mman.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.h b/drivers/gpu/drm/i915/gem/i915_gem_mman.h
new file mode 100644
index 000000000..1fa91b303
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.h
@@ -0,0 +1,33 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_GEM_MMAN_H__
+#define __I915_GEM_MMAN_H__
+
+#include <linux/mm_types.h>
+#include <linux/types.h>
+
+struct drm_device;
+struct drm_file;
+struct drm_i915_gem_object;
+struct file;
+struct i915_mmap_offset;
+struct mutex;
+
+int i915_gem_mmap_gtt_version(void);
+int i915_gem_mmap(struct file *filp, struct vm_area_struct *vma);
+
+int i915_gem_dumb_mmap_offset(struct drm_file *file_priv,
+			      struct drm_device *dev,
+			      u32 handle, u64 *offset);
+
+void __i915_gem_object_release_mmap_gtt(struct drm_i915_gem_object *obj);
+void i915_gem_object_release_mmap_gtt(struct drm_i915_gem_object *obj);
+
+void i915_gem_object_runtime_pm_release_mmap_offset(struct drm_i915_gem_object *obj);
+void i915_gem_object_release_mmap_offset(struct drm_i915_gem_object *obj);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c b/drivers/gpu/drm/i915/gem/i915_gem_object.c
new file mode 100644
index 000000000..a40bc17ac
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c
@@ -0,0 +1,876 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/highmem.h>
+#include <linux/sched/mm.h>
+
+#include <drm/drm_cache.h>
+
+#include "display/intel_frontbuffer.h"
+#include "pxp/intel_pxp.h"
+
+#include "i915_drv.h"
+#include "i915_file_private.h"
+#include "i915_gem_clflush.h"
+#include "i915_gem_context.h"
+#include "i915_gem_dmabuf.h"
+#include "i915_gem_mman.h"
+#include "i915_gem_object.h"
+#include "i915_gem_ttm.h"
+#include "i915_memcpy.h"
+#include "i915_trace.h"
+
+static struct kmem_cache *slab_objects;
+
+static const struct drm_gem_object_funcs i915_gem_object_funcs;
+
+struct drm_i915_gem_object *i915_gem_object_alloc(void)
+{
+	struct drm_i915_gem_object *obj;
+
+	obj = kmem_cache_zalloc(slab_objects, GFP_KERNEL);
+	if (!obj)
+		return NULL;
+	obj->base.funcs = &i915_gem_object_funcs;
+
+	return obj;
+}
+
+void i915_gem_object_free(struct drm_i915_gem_object *obj)
+{
+	return kmem_cache_free(slab_objects, obj);
+}
+
+void i915_gem_object_init(struct drm_i915_gem_object *obj,
+			  const struct drm_i915_gem_object_ops *ops,
+			  struct lock_class_key *key, unsigned flags)
+{
+	/*
+	 * A gem object is embedded both in a struct ttm_buffer_object :/ and
+	 * in a drm_i915_gem_object. Make sure they are aliased.
+	 */
+	BUILD_BUG_ON(offsetof(typeof(*obj), base) !=
+		     offsetof(typeof(*obj), __do_not_access.base));
+
+	spin_lock_init(&obj->vma.lock);
+	INIT_LIST_HEAD(&obj->vma.list);
+
+	INIT_LIST_HEAD(&obj->mm.link);
+
+	INIT_LIST_HEAD(&obj->lut_list);
+	spin_lock_init(&obj->lut_lock);
+
+	spin_lock_init(&obj->mmo.lock);
+	obj->mmo.offsets = RB_ROOT;
+
+	init_rcu_head(&obj->rcu);
+
+	obj->ops = ops;
+	GEM_BUG_ON(flags & ~I915_BO_ALLOC_FLAGS);
+	obj->flags = flags;
+
+	obj->mm.madv = I915_MADV_WILLNEED;
+	INIT_RADIX_TREE(&obj->mm.get_page.radix, GFP_KERNEL | __GFP_NOWARN);
+	mutex_init(&obj->mm.get_page.lock);
+	INIT_RADIX_TREE(&obj->mm.get_dma_page.radix, GFP_KERNEL | __GFP_NOWARN);
+	mutex_init(&obj->mm.get_dma_page.lock);
+}
+
+/**
+ * __i915_gem_object_fini - Clean up a GEM object initialization
+ * @obj: The gem object to cleanup
+ *
+ * This function cleans up gem object fields that are set up by
+ * drm_gem_private_object_init() and i915_gem_object_init().
+ * It's primarily intended as a helper for backends that need to
+ * clean up the gem object in separate steps.
+ */
+void __i915_gem_object_fini(struct drm_i915_gem_object *obj)
+{
+	mutex_destroy(&obj->mm.get_page.lock);
+	mutex_destroy(&obj->mm.get_dma_page.lock);
+	dma_resv_fini(&obj->base._resv);
+}
+
+/**
+ * i915_gem_object_set_cache_coherency - Mark up the object's coherency levels
+ * for a given cache_level
+ * @obj: #drm_i915_gem_object
+ * @cache_level: cache level
+ */
+void i915_gem_object_set_cache_coherency(struct drm_i915_gem_object *obj,
+					 unsigned int cache_level)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+
+	obj->cache_level = cache_level;
+
+	if (cache_level != I915_CACHE_NONE)
+		obj->cache_coherent = (I915_BO_CACHE_COHERENT_FOR_READ |
+				       I915_BO_CACHE_COHERENT_FOR_WRITE);
+	else if (HAS_LLC(i915))
+		obj->cache_coherent = I915_BO_CACHE_COHERENT_FOR_READ;
+	else
+		obj->cache_coherent = 0;
+
+	obj->cache_dirty =
+		!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_WRITE) &&
+		!IS_DGFX(i915);
+}
+
+bool i915_gem_object_can_bypass_llc(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+
+	/*
+	 * This is purely from a security perspective, so we simply don't care
+	 * about non-userspace objects being able to bypass the LLC.
+	 */
+	if (!(obj->flags & I915_BO_ALLOC_USER))
+		return false;
+
+	/*
+	 * EHL and JSL add the 'Bypass LLC' MOCS entry, which should make it
+	 * possible for userspace to bypass the GTT caching bits set by the
+	 * kernel, as per the given object cache_level. This is troublesome
+	 * since the heavy flush we apply when first gathering the pages is
+	 * skipped if the kernel thinks the object is coherent with the GPU. As
+	 * a result it might be possible to bypass the cache and read the
+	 * contents of the page directly, which could be stale data. If it's
+	 * just a case of userspace shooting themselves in the foot then so be
+	 * it, but since i915 takes the stance of always zeroing memory before
+	 * handing it to userspace, we need to prevent this.
+	 */
+	return IS_JSL_EHL(i915);
+}
+
+static void i915_gem_close_object(struct drm_gem_object *gem, struct drm_file *file)
+{
+	struct drm_i915_gem_object *obj = to_intel_bo(gem);
+	struct drm_i915_file_private *fpriv = file->driver_priv;
+	struct i915_lut_handle bookmark = {};
+	struct i915_mmap_offset *mmo, *mn;
+	struct i915_lut_handle *lut, *ln;
+	LIST_HEAD(close);
+
+	spin_lock(&obj->lut_lock);
+	list_for_each_entry_safe(lut, ln, &obj->lut_list, obj_link) {
+		struct i915_gem_context *ctx = lut->ctx;
+
+		if (ctx && ctx->file_priv == fpriv) {
+			i915_gem_context_get(ctx);
+			list_move(&lut->obj_link, &close);
+		}
+
+		/* Break long locks, and carefully continue on from this spot */
+		if (&ln->obj_link != &obj->lut_list) {
+			list_add_tail(&bookmark.obj_link, &ln->obj_link);
+			if (cond_resched_lock(&obj->lut_lock))
+				list_safe_reset_next(&bookmark, ln, obj_link);
+			__list_del_entry(&bookmark.obj_link);
+		}
+	}
+	spin_unlock(&obj->lut_lock);
+
+	spin_lock(&obj->mmo.lock);
+	rbtree_postorder_for_each_entry_safe(mmo, mn, &obj->mmo.offsets, offset)
+		drm_vma_node_revoke(&mmo->vma_node, file);
+	spin_unlock(&obj->mmo.lock);
+
+	list_for_each_entry_safe(lut, ln, &close, obj_link) {
+		struct i915_gem_context *ctx = lut->ctx;
+		struct i915_vma *vma;
+
+		/*
+		 * We allow the process to have multiple handles to the same
+		 * vma, in the same fd namespace, by virtue of flink/open.
+		 */
+
+		mutex_lock(&ctx->lut_mutex);
+		vma = radix_tree_delete(&ctx->handles_vma, lut->handle);
+		if (vma) {
+			GEM_BUG_ON(vma->obj != obj);
+			GEM_BUG_ON(!atomic_read(&vma->open_count));
+			i915_vma_close(vma);
+		}
+		mutex_unlock(&ctx->lut_mutex);
+
+		i915_gem_context_put(lut->ctx);
+		i915_lut_handle_free(lut);
+		i915_gem_object_put(obj);
+	}
+}
+
+void __i915_gem_free_object_rcu(struct rcu_head *head)
+{
+	struct drm_i915_gem_object *obj =
+		container_of(head, typeof(*obj), rcu);
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+
+	i915_gem_object_free(obj);
+
+	GEM_BUG_ON(!atomic_read(&i915->mm.free_count));
+	atomic_dec(&i915->mm.free_count);
+}
+
+static void __i915_gem_object_free_mmaps(struct drm_i915_gem_object *obj)
+{
+	/* Skip serialisation and waking the device if known to be not used. */
+
+	if (obj->userfault_count && !IS_DGFX(to_i915(obj->base.dev)))
+		i915_gem_object_release_mmap_gtt(obj);
+
+	if (!RB_EMPTY_ROOT(&obj->mmo.offsets)) {
+		struct i915_mmap_offset *mmo, *mn;
+
+		i915_gem_object_release_mmap_offset(obj);
+
+		rbtree_postorder_for_each_entry_safe(mmo, mn,
+						     &obj->mmo.offsets,
+						     offset) {
+			drm_vma_offset_remove(obj->base.dev->vma_offset_manager,
+					      &mmo->vma_node);
+			kfree(mmo);
+		}
+		obj->mmo.offsets = RB_ROOT;
+	}
+}
+
+/**
+ * __i915_gem_object_pages_fini - Clean up pages use of a gem object
+ * @obj: The gem object to clean up
+ *
+ * This function cleans up usage of the object mm.pages member. It
+ * is intended for backends that need to clean up a gem object in
+ * separate steps and needs to be called when the object is idle before
+ * the object's backing memory is freed.
+ */
+void __i915_gem_object_pages_fini(struct drm_i915_gem_object *obj)
+{
+	assert_object_held_shared(obj);
+
+	if (!list_empty(&obj->vma.list)) {
+		struct i915_vma *vma;
+
+		spin_lock(&obj->vma.lock);
+		while ((vma = list_first_entry_or_null(&obj->vma.list,
+						       struct i915_vma,
+						       obj_link))) {
+			GEM_BUG_ON(vma->obj != obj);
+			spin_unlock(&obj->vma.lock);
+
+			i915_vma_destroy(vma);
+
+			spin_lock(&obj->vma.lock);
+		}
+		spin_unlock(&obj->vma.lock);
+	}
+
+	__i915_gem_object_free_mmaps(obj);
+
+	atomic_set(&obj->mm.pages_pin_count, 0);
+	__i915_gem_object_put_pages(obj);
+	GEM_BUG_ON(i915_gem_object_has_pages(obj));
+}
+
+void __i915_gem_free_object(struct drm_i915_gem_object *obj)
+{
+	trace_i915_gem_object_destroy(obj);
+
+	GEM_BUG_ON(!list_empty(&obj->lut_list));
+
+	bitmap_free(obj->bit_17);
+
+	if (obj->base.import_attach)
+		drm_prime_gem_destroy(&obj->base, NULL);
+
+	drm_gem_free_mmap_offset(&obj->base);
+
+	if (obj->ops->release)
+		obj->ops->release(obj);
+
+	if (obj->mm.n_placements > 1)
+		kfree(obj->mm.placements);
+
+	if (obj->shares_resv_from)
+		i915_vm_resv_put(obj->shares_resv_from);
+
+	__i915_gem_object_fini(obj);
+}
+
+static void __i915_gem_free_objects(struct drm_i915_private *i915,
+				    struct llist_node *freed)
+{
+	struct drm_i915_gem_object *obj, *on;
+
+	llist_for_each_entry_safe(obj, on, freed, freed) {
+		might_sleep();
+		if (obj->ops->delayed_free) {
+			obj->ops->delayed_free(obj);
+			continue;
+		}
+
+		__i915_gem_object_pages_fini(obj);
+		__i915_gem_free_object(obj);
+
+		/* But keep the pointer alive for RCU-protected lookups */
+		call_rcu(&obj->rcu, __i915_gem_free_object_rcu);
+		cond_resched();
+	}
+}
+
+void i915_gem_flush_free_objects(struct drm_i915_private *i915)
+{
+	struct llist_node *freed = llist_del_all(&i915->mm.free_list);
+
+	if (unlikely(freed))
+		__i915_gem_free_objects(i915, freed);
+}
+
+static void __i915_gem_free_work(struct work_struct *work)
+{
+	struct drm_i915_private *i915 =
+		container_of(work, struct drm_i915_private, mm.free_work);
+
+	i915_gem_flush_free_objects(i915);
+}
+
+static void i915_gem_free_object(struct drm_gem_object *gem_obj)
+{
+	struct drm_i915_gem_object *obj = to_intel_bo(gem_obj);
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+
+	GEM_BUG_ON(i915_gem_object_is_framebuffer(obj));
+
+	/*
+	 * Before we free the object, make sure any pure RCU-only
+	 * read-side critical sections are complete, e.g.
+	 * i915_gem_busy_ioctl(). For the corresponding synchronized
+	 * lookup see i915_gem_object_lookup_rcu().
+	 */
+	atomic_inc(&i915->mm.free_count);
+
+	/*
+	 * Since we require blocking on struct_mutex to unbind the freed
+	 * object from the GPU before releasing resources back to the
+	 * system, we can not do that directly from the RCU callback (which may
+	 * be a softirq context), but must instead then defer that work onto a
+	 * kthread. We use the RCU callback rather than move the freed object
+	 * directly onto the work queue so that we can mix between using the
+	 * worker and performing frees directly from subsequent allocations for
+	 * crude but effective memory throttling.
+	 */
+
+	if (llist_add(&obj->freed, &i915->mm.free_list))
+		queue_work(i915->wq, &i915->mm.free_work);
+}
+
+void __i915_gem_object_flush_frontbuffer(struct drm_i915_gem_object *obj,
+					 enum fb_op_origin origin)
+{
+	struct intel_frontbuffer *front;
+
+	front = __intel_frontbuffer_get(obj);
+	if (front) {
+		intel_frontbuffer_flush(front, origin);
+		intel_frontbuffer_put(front);
+	}
+}
+
+void __i915_gem_object_invalidate_frontbuffer(struct drm_i915_gem_object *obj,
+					      enum fb_op_origin origin)
+{
+	struct intel_frontbuffer *front;
+
+	front = __intel_frontbuffer_get(obj);
+	if (front) {
+		intel_frontbuffer_invalidate(front, origin);
+		intel_frontbuffer_put(front);
+	}
+}
+
+static void
+i915_gem_object_read_from_page_kmap(struct drm_i915_gem_object *obj, u64 offset, void *dst, int size)
+{
+	void *src_map;
+	void *src_ptr;
+
+	src_map = kmap_atomic(i915_gem_object_get_page(obj, offset >> PAGE_SHIFT));
+
+	src_ptr = src_map + offset_in_page(offset);
+	if (!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
+		drm_clflush_virt_range(src_ptr, size);
+	memcpy(dst, src_ptr, size);
+
+	kunmap_atomic(src_map);
+}
+
+static void
+i915_gem_object_read_from_page_iomap(struct drm_i915_gem_object *obj, u64 offset, void *dst, int size)
+{
+	void __iomem *src_map;
+	void __iomem *src_ptr;
+	dma_addr_t dma = i915_gem_object_get_dma_address(obj, offset >> PAGE_SHIFT);
+
+	src_map = io_mapping_map_wc(&obj->mm.region->iomap,
+				    dma - obj->mm.region->region.start,
+				    PAGE_SIZE);
+
+	src_ptr = src_map + offset_in_page(offset);
+	if (!i915_memcpy_from_wc(dst, (void __force *)src_ptr, size))
+		memcpy_fromio(dst, src_ptr, size);
+
+	io_mapping_unmap(src_map);
+}
+
+/**
+ * i915_gem_object_read_from_page - read data from the page of a GEM object
+ * @obj: GEM object to read from
+ * @offset: offset within the object
+ * @dst: buffer to store the read data
+ * @size: size to read
+ *
+ * Reads data from @obj at the specified offset. The requested region to read
+ * from can't cross a page boundary. The caller must ensure that @obj pages
+ * are pinned and that @obj is synced wrt. any related writes.
+ *
+ * Return: %0 on success or -ENODEV if the type of @obj's backing store is
+ * unsupported.
+ */
+int i915_gem_object_read_from_page(struct drm_i915_gem_object *obj, u64 offset, void *dst, int size)
+{
+	GEM_BUG_ON(offset >= obj->base.size);
+	GEM_BUG_ON(offset_in_page(offset) > PAGE_SIZE - size);
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+
+	if (i915_gem_object_has_struct_page(obj))
+		i915_gem_object_read_from_page_kmap(obj, offset, dst, size);
+	else if (i915_gem_object_has_iomem(obj))
+		i915_gem_object_read_from_page_iomap(obj, offset, dst, size);
+	else
+		return -ENODEV;
+
+	return 0;
+}
+
+/**
+ * i915_gem_object_evictable - Whether object is likely evictable after unbind.
+ * @obj: The object to check
+ *
+ * This function checks whether the object is likely unvictable after unbind.
+ * If the object is not locked when checking, the result is only advisory.
+ * If the object is locked when checking, and the function returns true,
+ * then an eviction should indeed be possible. But since unlocked vma
+ * unpinning and unbinding is currently possible, the object can actually
+ * become evictable even if this function returns false.
+ *
+ * Return: true if the object may be evictable. False otherwise.
+ */
+bool i915_gem_object_evictable(struct drm_i915_gem_object *obj)
+{
+	struct i915_vma *vma;
+	int pin_count = atomic_read(&obj->mm.pages_pin_count);
+
+	if (!pin_count)
+		return true;
+
+	spin_lock(&obj->vma.lock);
+	list_for_each_entry(vma, &obj->vma.list, obj_link) {
+		if (i915_vma_is_pinned(vma)) {
+			spin_unlock(&obj->vma.lock);
+			return false;
+		}
+		if (atomic_read(&vma->pages_count))
+			pin_count--;
+	}
+	spin_unlock(&obj->vma.lock);
+	GEM_WARN_ON(pin_count < 0);
+
+	return pin_count == 0;
+}
+
+/**
+ * i915_gem_object_migratable - Whether the object is migratable out of the
+ * current region.
+ * @obj: Pointer to the object.
+ *
+ * Return: Whether the object is allowed to be resident in other
+ * regions than the current while pages are present.
+ */
+bool i915_gem_object_migratable(struct drm_i915_gem_object *obj)
+{
+	struct intel_memory_region *mr = READ_ONCE(obj->mm.region);
+
+	if (!mr)
+		return false;
+
+	return obj->mm.n_placements > 1;
+}
+
+/**
+ * i915_gem_object_has_struct_page - Whether the object is page-backed
+ * @obj: The object to query.
+ *
+ * This function should only be called while the object is locked or pinned,
+ * otherwise the page backing may change under the caller.
+ *
+ * Return: True if page-backed, false otherwise.
+ */
+bool i915_gem_object_has_struct_page(const struct drm_i915_gem_object *obj)
+{
+#ifdef CONFIG_LOCKDEP
+	if (IS_DGFX(to_i915(obj->base.dev)) &&
+	    i915_gem_object_evictable((void __force *)obj))
+		assert_object_held_shared(obj);
+#endif
+	return obj->mem_flags & I915_BO_FLAG_STRUCT_PAGE;
+}
+
+/**
+ * i915_gem_object_has_iomem - Whether the object is iomem-backed
+ * @obj: The object to query.
+ *
+ * This function should only be called while the object is locked or pinned,
+ * otherwise the iomem backing may change under the caller.
+ *
+ * Return: True if iomem-backed, false otherwise.
+ */
+bool i915_gem_object_has_iomem(const struct drm_i915_gem_object *obj)
+{
+#ifdef CONFIG_LOCKDEP
+	if (IS_DGFX(to_i915(obj->base.dev)) &&
+	    i915_gem_object_evictable((void __force *)obj))
+		assert_object_held_shared(obj);
+#endif
+	return obj->mem_flags & I915_BO_FLAG_IOMEM;
+}
+
+/**
+ * i915_gem_object_can_migrate - Whether an object likely can be migrated
+ *
+ * @obj: The object to migrate
+ * @id: The region intended to migrate to
+ *
+ * Check whether the object backend supports migration to the
+ * given region. Note that pinning may affect the ability to migrate as
+ * returned by this function.
+ *
+ * This function is primarily intended as a helper for checking the
+ * possibility to migrate objects and might be slightly less permissive
+ * than i915_gem_object_migrate() when it comes to objects with the
+ * I915_BO_ALLOC_USER flag set.
+ *
+ * Return: true if migration is possible, false otherwise.
+ */
+bool i915_gem_object_can_migrate(struct drm_i915_gem_object *obj,
+				 enum intel_region_id id)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	unsigned int num_allowed = obj->mm.n_placements;
+	struct intel_memory_region *mr;
+	unsigned int i;
+
+	GEM_BUG_ON(id >= INTEL_REGION_UNKNOWN);
+	GEM_BUG_ON(obj->mm.madv != I915_MADV_WILLNEED);
+
+	mr = i915->mm.regions[id];
+	if (!mr)
+		return false;
+
+	if (!IS_ALIGNED(obj->base.size, mr->min_page_size))
+		return false;
+
+	if (obj->mm.region == mr)
+		return true;
+
+	if (!i915_gem_object_evictable(obj))
+		return false;
+
+	if (!obj->ops->migrate)
+		return false;
+
+	if (!(obj->flags & I915_BO_ALLOC_USER))
+		return true;
+
+	if (num_allowed == 0)
+		return false;
+
+	for (i = 0; i < num_allowed; ++i) {
+		if (mr == obj->mm.placements[i])
+			return true;
+	}
+
+	return false;
+}
+
+/**
+ * i915_gem_object_migrate - Migrate an object to the desired region id
+ * @obj: The object to migrate.
+ * @ww: An optional struct i915_gem_ww_ctx. If NULL, the backend may
+ * not be successful in evicting other objects to make room for this object.
+ * @id: The region id to migrate to.
+ *
+ * Attempt to migrate the object to the desired memory region. The
+ * object backend must support migration and the object may not be
+ * pinned, (explicitly pinned pages or pinned vmas). The object must
+ * be locked.
+ * On successful completion, the object will have pages pointing to
+ * memory in the new region, but an async migration task may not have
+ * completed yet, and to accomplish that, i915_gem_object_wait_migration()
+ * must be called.
+ *
+ * Note: the @ww parameter is not used yet, but included to make sure
+ * callers put some effort into obtaining a valid ww ctx if one is
+ * available.
+ *
+ * Return: 0 on success. Negative error code on failure. In particular may
+ * return -ENXIO on lack of region space, -EDEADLK for deadlock avoidance
+ * if @ww is set, -EINTR or -ERESTARTSYS if signal pending, and
+ * -EBUSY if the object is pinned.
+ */
+int i915_gem_object_migrate(struct drm_i915_gem_object *obj,
+			    struct i915_gem_ww_ctx *ww,
+			    enum intel_region_id id)
+{
+	return __i915_gem_object_migrate(obj, ww, id, obj->flags);
+}
+
+/**
+ * __i915_gem_object_migrate - Migrate an object to the desired region id, with
+ * control of the extra flags
+ * @obj: The object to migrate.
+ * @ww: An optional struct i915_gem_ww_ctx. If NULL, the backend may
+ * not be successful in evicting other objects to make room for this object.
+ * @id: The region id to migrate to.
+ * @flags: The object flags. Normally just obj->flags.
+ *
+ * Attempt to migrate the object to the desired memory region. The
+ * object backend must support migration and the object may not be
+ * pinned, (explicitly pinned pages or pinned vmas). The object must
+ * be locked.
+ * On successful completion, the object will have pages pointing to
+ * memory in the new region, but an async migration task may not have
+ * completed yet, and to accomplish that, i915_gem_object_wait_migration()
+ * must be called.
+ *
+ * Note: the @ww parameter is not used yet, but included to make sure
+ * callers put some effort into obtaining a valid ww ctx if one is
+ * available.
+ *
+ * Return: 0 on success. Negative error code on failure. In particular may
+ * return -ENXIO on lack of region space, -EDEADLK for deadlock avoidance
+ * if @ww is set, -EINTR or -ERESTARTSYS if signal pending, and
+ * -EBUSY if the object is pinned.
+ */
+int __i915_gem_object_migrate(struct drm_i915_gem_object *obj,
+			      struct i915_gem_ww_ctx *ww,
+			      enum intel_region_id id,
+			      unsigned int flags)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct intel_memory_region *mr;
+
+	GEM_BUG_ON(id >= INTEL_REGION_UNKNOWN);
+	GEM_BUG_ON(obj->mm.madv != I915_MADV_WILLNEED);
+	assert_object_held(obj);
+
+	mr = i915->mm.regions[id];
+	GEM_BUG_ON(!mr);
+
+	if (!i915_gem_object_can_migrate(obj, id))
+		return -EINVAL;
+
+	if (!obj->ops->migrate) {
+		if (GEM_WARN_ON(obj->mm.region != mr))
+			return -EINVAL;
+		return 0;
+	}
+
+	return obj->ops->migrate(obj, mr, flags);
+}
+
+/**
+ * i915_gem_object_placement_possible - Check whether the object can be
+ * placed at certain memory type
+ * @obj: Pointer to the object
+ * @type: The memory type to check
+ *
+ * Return: True if the object can be placed in @type. False otherwise.
+ */
+bool i915_gem_object_placement_possible(struct drm_i915_gem_object *obj,
+					enum intel_memory_type type)
+{
+	unsigned int i;
+
+	if (!obj->mm.n_placements) {
+		switch (type) {
+		case INTEL_MEMORY_LOCAL:
+			return i915_gem_object_has_iomem(obj);
+		case INTEL_MEMORY_SYSTEM:
+			return i915_gem_object_has_pages(obj);
+		default:
+			/* Ignore stolen for now */
+			GEM_BUG_ON(1);
+			return false;
+		}
+	}
+
+	for (i = 0; i < obj->mm.n_placements; i++) {
+		if (obj->mm.placements[i]->type == type)
+			return true;
+	}
+
+	return false;
+}
+
+/**
+ * i915_gem_object_needs_ccs_pages - Check whether the object requires extra
+ * pages when placed in system-memory, in order to save and later restore the
+ * flat-CCS aux state when the object is moved between local-memory and
+ * system-memory
+ * @obj: Pointer to the object
+ *
+ * Return: True if the object needs extra ccs pages. False otherwise.
+ */
+bool i915_gem_object_needs_ccs_pages(struct drm_i915_gem_object *obj)
+{
+	bool lmem_placement = false;
+	int i;
+
+	if (!HAS_FLAT_CCS(to_i915(obj->base.dev)))
+		return false;
+
+	if (obj->flags & I915_BO_ALLOC_CCS_AUX)
+		return true;
+
+	for (i = 0; i < obj->mm.n_placements; i++) {
+		/* Compression is not allowed for the objects with smem placement */
+		if (obj->mm.placements[i]->type == INTEL_MEMORY_SYSTEM)
+			return false;
+		if (!lmem_placement &&
+		    obj->mm.placements[i]->type == INTEL_MEMORY_LOCAL)
+			lmem_placement = true;
+	}
+
+	return lmem_placement;
+}
+
+void i915_gem_init__objects(struct drm_i915_private *i915)
+{
+	INIT_WORK(&i915->mm.free_work, __i915_gem_free_work);
+}
+
+void i915_objects_module_exit(void)
+{
+	kmem_cache_destroy(slab_objects);
+}
+
+int __init i915_objects_module_init(void)
+{
+	slab_objects = KMEM_CACHE(drm_i915_gem_object, SLAB_HWCACHE_ALIGN);
+	if (!slab_objects)
+		return -ENOMEM;
+
+	return 0;
+}
+
+static const struct drm_gem_object_funcs i915_gem_object_funcs = {
+	.free = i915_gem_free_object,
+	.close = i915_gem_close_object,
+	.export = i915_gem_prime_export,
+};
+
+/**
+ * i915_gem_object_get_moving_fence - Get the object's moving fence if any
+ * @obj: The object whose moving fence to get.
+ * @fence: The resulting fence
+ *
+ * A non-signaled moving fence means that there is an async operation
+ * pending on the object that needs to be waited on before setting up
+ * any GPU- or CPU PTEs to the object's pages.
+ *
+ * Return: Negative error code or 0 for success.
+ */
+int i915_gem_object_get_moving_fence(struct drm_i915_gem_object *obj,
+				     struct dma_fence **fence)
+{
+	return dma_resv_get_singleton(obj->base.resv, DMA_RESV_USAGE_KERNEL,
+				      fence);
+}
+
+/**
+ * i915_gem_object_wait_moving_fence - Wait for the object's moving fence if any
+ * @obj: The object whose moving fence to wait for.
+ * @intr: Whether to wait interruptible.
+ *
+ * If the moving fence signaled without an error, it is detached from the
+ * object and put.
+ *
+ * Return: 0 if successful, -ERESTARTSYS if the wait was interrupted,
+ * negative error code if the async operation represented by the
+ * moving fence failed.
+ */
+int i915_gem_object_wait_moving_fence(struct drm_i915_gem_object *obj,
+				      bool intr)
+{
+	long ret;
+
+	assert_object_held(obj);
+
+	ret = dma_resv_wait_timeout(obj->base. resv, DMA_RESV_USAGE_KERNEL,
+				    intr, MAX_SCHEDULE_TIMEOUT);
+	if (!ret)
+		ret = -ETIME;
+	else if (ret > 0 && i915_gem_object_has_unknown_state(obj))
+		ret = -EIO;
+
+	return ret < 0 ? ret : 0;
+}
+
+/**
+ * i915_gem_object_has_unknown_state - Return true if the object backing pages are
+ * in an unknown_state. This means that userspace must NEVER be allowed to touch
+ * the pages, with either the GPU or CPU.
+ *
+ * ONLY valid to be called after ensuring that all kernel fences have signalled
+ * (in particular the fence for moving/clearing the object).
+ */
+bool i915_gem_object_has_unknown_state(struct drm_i915_gem_object *obj)
+{
+	/*
+	 * The below barrier pairs with the dma_fence_signal() in
+	 * __memcpy_work(). We should only sample the unknown_state after all
+	 * the kernel fences have signalled.
+	 */
+	smp_rmb();
+	return obj->mm.unknown_state;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/huge_gem_object.c"
+#include "selftests/huge_pages.c"
+#include "selftests/i915_gem_migrate.c"
+#include "selftests/i915_gem_object.c"
+#include "selftests/i915_gem_coherency.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.h b/drivers/gpu/drm/i915/gem/i915_gem_object.h
new file mode 100644
index 000000000..ea951e2f5
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object.h
@@ -0,0 +1,654 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __I915_GEM_OBJECT_H__
+#define __I915_GEM_OBJECT_H__
+
+#include <drm/drm_gem.h>
+#include <drm/drm_file.h>
+#include <drm/drm_device.h>
+
+#include "display/intel_frontbuffer.h"
+#include "intel_memory_region.h"
+#include "i915_gem_object_types.h"
+#include "i915_gem_gtt.h"
+#include "i915_gem_ww.h"
+#include "i915_vma_types.h"
+
+enum intel_region_id;
+
+/*
+ * XXX: There is a prevalence of the assumption that we fit the
+ * object's page count inside a 32bit _signed_ variable. Let's document
+ * this and catch if we ever need to fix it. In the meantime, if you do
+ * spot such a local variable, please consider fixing!
+ *
+ * Aside from our own locals (for which we have no excuse!):
+ * - sg_table embeds unsigned int for num_pages
+ * - get_user_pages*() mixed ints with longs
+ */
+#define GEM_CHECK_SIZE_OVERFLOW(sz) \
+	GEM_WARN_ON((sz) >> PAGE_SHIFT > INT_MAX)
+
+static inline bool i915_gem_object_size_2big(u64 size)
+{
+	struct drm_i915_gem_object *obj;
+
+	if (GEM_CHECK_SIZE_OVERFLOW(size))
+		return true;
+
+	if (overflows_type(size, obj->base.size))
+		return true;
+
+	return false;
+}
+
+void i915_gem_init__objects(struct drm_i915_private *i915);
+
+void i915_objects_module_exit(void);
+int i915_objects_module_init(void);
+
+struct drm_i915_gem_object *i915_gem_object_alloc(void);
+void i915_gem_object_free(struct drm_i915_gem_object *obj);
+
+void i915_gem_object_init(struct drm_i915_gem_object *obj,
+			  const struct drm_i915_gem_object_ops *ops,
+			  struct lock_class_key *key,
+			  unsigned alloc_flags);
+
+void __i915_gem_object_fini(struct drm_i915_gem_object *obj);
+
+struct drm_i915_gem_object *
+i915_gem_object_create_shmem(struct drm_i915_private *i915,
+			     resource_size_t size);
+struct drm_i915_gem_object *
+i915_gem_object_create_shmem_from_data(struct drm_i915_private *i915,
+				       const void *data, resource_size_t size);
+struct drm_i915_gem_object *
+__i915_gem_object_create_user(struct drm_i915_private *i915, u64 size,
+			      struct intel_memory_region **placements,
+			      unsigned int n_placements);
+
+extern const struct drm_i915_gem_object_ops i915_gem_shmem_ops;
+
+void __i915_gem_object_release_shmem(struct drm_i915_gem_object *obj,
+				     struct sg_table *pages,
+				     bool needs_clflush);
+
+int i915_gem_object_pwrite_phys(struct drm_i915_gem_object *obj,
+				const struct drm_i915_gem_pwrite *args);
+int i915_gem_object_pread_phys(struct drm_i915_gem_object *obj,
+			       const struct drm_i915_gem_pread *args);
+
+int i915_gem_object_attach_phys(struct drm_i915_gem_object *obj, int align);
+void i915_gem_object_put_pages_shmem(struct drm_i915_gem_object *obj,
+				     struct sg_table *pages);
+void i915_gem_object_put_pages_phys(struct drm_i915_gem_object *obj,
+				    struct sg_table *pages);
+
+void i915_gem_flush_free_objects(struct drm_i915_private *i915);
+
+struct sg_table *
+__i915_gem_object_unset_pages(struct drm_i915_gem_object *obj);
+
+/**
+ * i915_gem_object_lookup_rcu - look up a temporary GEM object from its handle
+ * @filp: DRM file private date
+ * @handle: userspace handle
+ *
+ * Returns:
+ *
+ * A pointer to the object named by the handle if such exists on @filp, NULL
+ * otherwise. This object is only valid whilst under the RCU read lock, and
+ * note carefully the object may be in the process of being destroyed.
+ */
+static inline struct drm_i915_gem_object *
+i915_gem_object_lookup_rcu(struct drm_file *file, u32 handle)
+{
+#ifdef CONFIG_LOCKDEP
+	WARN_ON(debug_locks && !lock_is_held(&rcu_lock_map));
+#endif
+	return idr_find(&file->object_idr, handle);
+}
+
+static inline struct drm_i915_gem_object *
+i915_gem_object_get_rcu(struct drm_i915_gem_object *obj)
+{
+	if (obj && !kref_get_unless_zero(&obj->base.refcount))
+		obj = NULL;
+
+	return obj;
+}
+
+static inline struct drm_i915_gem_object *
+i915_gem_object_lookup(struct drm_file *file, u32 handle)
+{
+	struct drm_i915_gem_object *obj;
+
+	rcu_read_lock();
+	obj = i915_gem_object_lookup_rcu(file, handle);
+	obj = i915_gem_object_get_rcu(obj);
+	rcu_read_unlock();
+
+	return obj;
+}
+
+__deprecated
+struct drm_gem_object *
+drm_gem_object_lookup(struct drm_file *file, u32 handle);
+
+__attribute__((nonnull))
+static inline struct drm_i915_gem_object *
+i915_gem_object_get(struct drm_i915_gem_object *obj)
+{
+	drm_gem_object_get(&obj->base);
+	return obj;
+}
+
+__attribute__((nonnull))
+static inline void
+i915_gem_object_put(struct drm_i915_gem_object *obj)
+{
+	__drm_gem_object_put(&obj->base);
+}
+
+#define assert_object_held(obj) dma_resv_assert_held((obj)->base.resv)
+
+/*
+ * If more than one potential simultaneous locker, assert held.
+ */
+static inline void assert_object_held_shared(const struct drm_i915_gem_object *obj)
+{
+	/*
+	 * Note mm list lookup is protected by
+	 * kref_get_unless_zero().
+	 */
+	if (IS_ENABLED(CONFIG_LOCKDEP) &&
+	    kref_read(&obj->base.refcount) > 0)
+		assert_object_held(obj);
+}
+
+static inline int __i915_gem_object_lock(struct drm_i915_gem_object *obj,
+					 struct i915_gem_ww_ctx *ww,
+					 bool intr)
+{
+	int ret;
+
+	if (intr)
+		ret = dma_resv_lock_interruptible(obj->base.resv, ww ? &ww->ctx : NULL);
+	else
+		ret = dma_resv_lock(obj->base.resv, ww ? &ww->ctx : NULL);
+
+	if (!ret && ww) {
+		i915_gem_object_get(obj);
+		list_add_tail(&obj->obj_link, &ww->obj_list);
+	}
+	if (ret == -EALREADY)
+		ret = 0;
+
+	if (ret == -EDEADLK) {
+		i915_gem_object_get(obj);
+		ww->contended = obj;
+	}
+
+	return ret;
+}
+
+static inline int i915_gem_object_lock(struct drm_i915_gem_object *obj,
+				       struct i915_gem_ww_ctx *ww)
+{
+	return __i915_gem_object_lock(obj, ww, ww && ww->intr);
+}
+
+static inline int i915_gem_object_lock_interruptible(struct drm_i915_gem_object *obj,
+						     struct i915_gem_ww_ctx *ww)
+{
+	WARN_ON(ww && !ww->intr);
+	return __i915_gem_object_lock(obj, ww, true);
+}
+
+static inline bool i915_gem_object_trylock(struct drm_i915_gem_object *obj,
+					   struct i915_gem_ww_ctx *ww)
+{
+	if (!ww)
+		return dma_resv_trylock(obj->base.resv);
+	else
+		return ww_mutex_trylock(&obj->base.resv->lock, &ww->ctx);
+}
+
+static inline void i915_gem_object_unlock(struct drm_i915_gem_object *obj)
+{
+	if (obj->ops->adjust_lru)
+		obj->ops->adjust_lru(obj);
+
+	dma_resv_unlock(obj->base.resv);
+}
+
+static inline void
+i915_gem_object_set_readonly(struct drm_i915_gem_object *obj)
+{
+	obj->flags |= I915_BO_READONLY;
+}
+
+static inline bool
+i915_gem_object_is_readonly(const struct drm_i915_gem_object *obj)
+{
+	return obj->flags & I915_BO_READONLY;
+}
+
+static inline bool
+i915_gem_object_is_contiguous(const struct drm_i915_gem_object *obj)
+{
+	return obj->flags & I915_BO_ALLOC_CONTIGUOUS;
+}
+
+static inline bool
+i915_gem_object_is_volatile(const struct drm_i915_gem_object *obj)
+{
+	return obj->flags & I915_BO_ALLOC_VOLATILE;
+}
+
+static inline void
+i915_gem_object_set_volatile(struct drm_i915_gem_object *obj)
+{
+	obj->flags |= I915_BO_ALLOC_VOLATILE;
+}
+
+static inline bool
+i915_gem_object_has_tiling_quirk(struct drm_i915_gem_object *obj)
+{
+	return test_bit(I915_TILING_QUIRK_BIT, &obj->flags);
+}
+
+static inline void
+i915_gem_object_set_tiling_quirk(struct drm_i915_gem_object *obj)
+{
+	set_bit(I915_TILING_QUIRK_BIT, &obj->flags);
+}
+
+static inline void
+i915_gem_object_clear_tiling_quirk(struct drm_i915_gem_object *obj)
+{
+	clear_bit(I915_TILING_QUIRK_BIT, &obj->flags);
+}
+
+static inline bool
+i915_gem_object_is_protected(const struct drm_i915_gem_object *obj)
+{
+	return obj->flags & I915_BO_PROTECTED;
+}
+
+static inline bool
+i915_gem_object_type_has(const struct drm_i915_gem_object *obj,
+			 unsigned long flags)
+{
+	return obj->ops->flags & flags;
+}
+
+bool i915_gem_object_has_struct_page(const struct drm_i915_gem_object *obj);
+
+bool i915_gem_object_has_iomem(const struct drm_i915_gem_object *obj);
+
+static inline bool
+i915_gem_object_is_shrinkable(const struct drm_i915_gem_object *obj)
+{
+	return i915_gem_object_type_has(obj, I915_GEM_OBJECT_IS_SHRINKABLE);
+}
+
+static inline bool
+i915_gem_object_has_self_managed_shrink_list(const struct drm_i915_gem_object *obj)
+{
+	return i915_gem_object_type_has(obj, I915_GEM_OBJECT_SELF_MANAGED_SHRINK_LIST);
+}
+
+static inline bool
+i915_gem_object_is_proxy(const struct drm_i915_gem_object *obj)
+{
+	return i915_gem_object_type_has(obj, I915_GEM_OBJECT_IS_PROXY);
+}
+
+static inline bool
+i915_gem_object_never_mmap(const struct drm_i915_gem_object *obj)
+{
+	return i915_gem_object_type_has(obj, I915_GEM_OBJECT_NO_MMAP);
+}
+
+static inline bool
+i915_gem_object_is_framebuffer(const struct drm_i915_gem_object *obj)
+{
+	return READ_ONCE(obj->frontbuffer) || obj->is_dpt;
+}
+
+static inline unsigned int
+i915_gem_object_get_tiling(const struct drm_i915_gem_object *obj)
+{
+	return obj->tiling_and_stride & TILING_MASK;
+}
+
+static inline bool
+i915_gem_object_is_tiled(const struct drm_i915_gem_object *obj)
+{
+	return i915_gem_object_get_tiling(obj) != I915_TILING_NONE;
+}
+
+static inline unsigned int
+i915_gem_object_get_stride(const struct drm_i915_gem_object *obj)
+{
+	return obj->tiling_and_stride & STRIDE_MASK;
+}
+
+static inline unsigned int
+i915_gem_tile_height(unsigned int tiling)
+{
+	GEM_BUG_ON(!tiling);
+	return tiling == I915_TILING_Y ? 32 : 8;
+}
+
+static inline unsigned int
+i915_gem_object_get_tile_height(const struct drm_i915_gem_object *obj)
+{
+	return i915_gem_tile_height(i915_gem_object_get_tiling(obj));
+}
+
+static inline unsigned int
+i915_gem_object_get_tile_row_size(const struct drm_i915_gem_object *obj)
+{
+	return (i915_gem_object_get_stride(obj) *
+		i915_gem_object_get_tile_height(obj));
+}
+
+int i915_gem_object_set_tiling(struct drm_i915_gem_object *obj,
+			       unsigned int tiling, unsigned int stride);
+
+struct scatterlist *
+__i915_gem_object_get_sg(struct drm_i915_gem_object *obj,
+			 struct i915_gem_object_page_iter *iter,
+			 unsigned int n,
+			 unsigned int *offset, bool dma);
+
+static inline struct scatterlist *
+i915_gem_object_get_sg(struct drm_i915_gem_object *obj,
+		       unsigned int n,
+		       unsigned int *offset)
+{
+	return __i915_gem_object_get_sg(obj, &obj->mm.get_page, n, offset, false);
+}
+
+static inline struct scatterlist *
+i915_gem_object_get_sg_dma(struct drm_i915_gem_object *obj,
+			   unsigned int n,
+			   unsigned int *offset)
+{
+	return __i915_gem_object_get_sg(obj, &obj->mm.get_dma_page, n, offset, true);
+}
+
+struct page *
+i915_gem_object_get_page(struct drm_i915_gem_object *obj,
+			 unsigned int n);
+
+struct page *
+i915_gem_object_get_dirty_page(struct drm_i915_gem_object *obj,
+			       unsigned int n);
+
+dma_addr_t
+i915_gem_object_get_dma_address_len(struct drm_i915_gem_object *obj,
+				    unsigned long n,
+				    unsigned int *len);
+
+dma_addr_t
+i915_gem_object_get_dma_address(struct drm_i915_gem_object *obj,
+				unsigned long n);
+
+void __i915_gem_object_set_pages(struct drm_i915_gem_object *obj,
+				 struct sg_table *pages,
+				 unsigned int sg_page_sizes);
+
+int ____i915_gem_object_get_pages(struct drm_i915_gem_object *obj);
+int __i915_gem_object_get_pages(struct drm_i915_gem_object *obj);
+
+static inline int __must_check
+i915_gem_object_pin_pages(struct drm_i915_gem_object *obj)
+{
+	assert_object_held(obj);
+
+	if (atomic_inc_not_zero(&obj->mm.pages_pin_count))
+		return 0;
+
+	return __i915_gem_object_get_pages(obj);
+}
+
+int i915_gem_object_pin_pages_unlocked(struct drm_i915_gem_object *obj);
+
+static inline bool
+i915_gem_object_has_pages(struct drm_i915_gem_object *obj)
+{
+	return !IS_ERR_OR_NULL(READ_ONCE(obj->mm.pages));
+}
+
+static inline void
+__i915_gem_object_pin_pages(struct drm_i915_gem_object *obj)
+{
+	GEM_BUG_ON(!i915_gem_object_has_pages(obj));
+
+	atomic_inc(&obj->mm.pages_pin_count);
+}
+
+static inline bool
+i915_gem_object_has_pinned_pages(struct drm_i915_gem_object *obj)
+{
+	return atomic_read(&obj->mm.pages_pin_count);
+}
+
+static inline void
+__i915_gem_object_unpin_pages(struct drm_i915_gem_object *obj)
+{
+	GEM_BUG_ON(!i915_gem_object_has_pages(obj));
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+
+	atomic_dec(&obj->mm.pages_pin_count);
+}
+
+static inline void
+i915_gem_object_unpin_pages(struct drm_i915_gem_object *obj)
+{
+	__i915_gem_object_unpin_pages(obj);
+}
+
+int __i915_gem_object_put_pages(struct drm_i915_gem_object *obj);
+int i915_gem_object_truncate(struct drm_i915_gem_object *obj);
+
+/**
+ * i915_gem_object_pin_map - return a contiguous mapping of the entire object
+ * @obj: the object to map into kernel address space
+ * @type: the type of mapping, used to select pgprot_t
+ *
+ * Calls i915_gem_object_pin_pages() to prevent reaping of the object's
+ * pages and then returns a contiguous mapping of the backing storage into
+ * the kernel address space. Based on the @type of mapping, the PTE will be
+ * set to either WriteBack or WriteCombine (via pgprot_t).
+ *
+ * The caller is responsible for calling i915_gem_object_unpin_map() when the
+ * mapping is no longer required.
+ *
+ * Returns the pointer through which to access the mapped object, or an
+ * ERR_PTR() on error.
+ */
+void *__must_check i915_gem_object_pin_map(struct drm_i915_gem_object *obj,
+					   enum i915_map_type type);
+
+void *__must_check i915_gem_object_pin_map_unlocked(struct drm_i915_gem_object *obj,
+						    enum i915_map_type type);
+
+void __i915_gem_object_flush_map(struct drm_i915_gem_object *obj,
+				 unsigned long offset,
+				 unsigned long size);
+static inline void i915_gem_object_flush_map(struct drm_i915_gem_object *obj)
+{
+	__i915_gem_object_flush_map(obj, 0, obj->base.size);
+}
+
+/**
+ * i915_gem_object_unpin_map - releases an earlier mapping
+ * @obj: the object to unmap
+ *
+ * After pinning the object and mapping its pages, once you are finished
+ * with your access, call i915_gem_object_unpin_map() to release the pin
+ * upon the mapping. Once the pin count reaches zero, that mapping may be
+ * removed.
+ */
+static inline void i915_gem_object_unpin_map(struct drm_i915_gem_object *obj)
+{
+	i915_gem_object_unpin_pages(obj);
+}
+
+void __i915_gem_object_release_map(struct drm_i915_gem_object *obj);
+
+int i915_gem_object_prepare_read(struct drm_i915_gem_object *obj,
+				 unsigned int *needs_clflush);
+int i915_gem_object_prepare_write(struct drm_i915_gem_object *obj,
+				  unsigned int *needs_clflush);
+#define CLFLUSH_BEFORE	BIT(0)
+#define CLFLUSH_AFTER	BIT(1)
+#define CLFLUSH_FLAGS	(CLFLUSH_BEFORE | CLFLUSH_AFTER)
+
+static inline void
+i915_gem_object_finish_access(struct drm_i915_gem_object *obj)
+{
+	i915_gem_object_unpin_pages(obj);
+}
+
+int i915_gem_object_get_moving_fence(struct drm_i915_gem_object *obj,
+				     struct dma_fence **fence);
+int i915_gem_object_wait_moving_fence(struct drm_i915_gem_object *obj,
+				      bool intr);
+bool i915_gem_object_has_unknown_state(struct drm_i915_gem_object *obj);
+
+void i915_gem_object_set_cache_coherency(struct drm_i915_gem_object *obj,
+					 unsigned int cache_level);
+bool i915_gem_object_can_bypass_llc(struct drm_i915_gem_object *obj);
+void i915_gem_object_flush_if_display(struct drm_i915_gem_object *obj);
+void i915_gem_object_flush_if_display_locked(struct drm_i915_gem_object *obj);
+bool i915_gem_cpu_write_needs_clflush(struct drm_i915_gem_object *obj);
+
+int __must_check
+i915_gem_object_set_to_wc_domain(struct drm_i915_gem_object *obj, bool write);
+int __must_check
+i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write);
+int __must_check
+i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_object *obj, bool write);
+struct i915_vma * __must_check
+i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj,
+				     struct i915_gem_ww_ctx *ww,
+				     u32 alignment,
+				     const struct i915_gtt_view *view,
+				     unsigned int flags);
+
+void i915_gem_object_make_unshrinkable(struct drm_i915_gem_object *obj);
+void i915_gem_object_make_shrinkable(struct drm_i915_gem_object *obj);
+void __i915_gem_object_make_shrinkable(struct drm_i915_gem_object *obj);
+void __i915_gem_object_make_purgeable(struct drm_i915_gem_object *obj);
+void i915_gem_object_make_purgeable(struct drm_i915_gem_object *obj);
+
+static inline void __start_cpu_write(struct drm_i915_gem_object *obj)
+{
+	obj->read_domains = I915_GEM_DOMAIN_CPU;
+	obj->write_domain = I915_GEM_DOMAIN_CPU;
+	if (i915_gem_cpu_write_needs_clflush(obj))
+		obj->cache_dirty = true;
+}
+
+void i915_gem_fence_wait_priority(struct dma_fence *fence,
+				  const struct i915_sched_attr *attr);
+
+int i915_gem_object_wait(struct drm_i915_gem_object *obj,
+			 unsigned int flags,
+			 long timeout);
+int i915_gem_object_wait_priority(struct drm_i915_gem_object *obj,
+				  unsigned int flags,
+				  const struct i915_sched_attr *attr);
+
+void __i915_gem_object_flush_frontbuffer(struct drm_i915_gem_object *obj,
+					 enum fb_op_origin origin);
+void __i915_gem_object_invalidate_frontbuffer(struct drm_i915_gem_object *obj,
+					      enum fb_op_origin origin);
+
+static inline void
+i915_gem_object_flush_frontbuffer(struct drm_i915_gem_object *obj,
+				  enum fb_op_origin origin)
+{
+	if (unlikely(rcu_access_pointer(obj->frontbuffer)))
+		__i915_gem_object_flush_frontbuffer(obj, origin);
+}
+
+static inline void
+i915_gem_object_invalidate_frontbuffer(struct drm_i915_gem_object *obj,
+				       enum fb_op_origin origin)
+{
+	if (unlikely(rcu_access_pointer(obj->frontbuffer)))
+		__i915_gem_object_invalidate_frontbuffer(obj, origin);
+}
+
+int i915_gem_object_read_from_page(struct drm_i915_gem_object *obj, u64 offset, void *dst, int size);
+
+bool i915_gem_object_is_shmem(const struct drm_i915_gem_object *obj);
+
+void __i915_gem_free_object_rcu(struct rcu_head *head);
+
+void __i915_gem_object_pages_fini(struct drm_i915_gem_object *obj);
+
+void __i915_gem_free_object(struct drm_i915_gem_object *obj);
+
+bool i915_gem_object_evictable(struct drm_i915_gem_object *obj);
+
+bool i915_gem_object_migratable(struct drm_i915_gem_object *obj);
+
+int i915_gem_object_migrate(struct drm_i915_gem_object *obj,
+			    struct i915_gem_ww_ctx *ww,
+			    enum intel_region_id id);
+int __i915_gem_object_migrate(struct drm_i915_gem_object *obj,
+			      struct i915_gem_ww_ctx *ww,
+			      enum intel_region_id id,
+			      unsigned int flags);
+
+bool i915_gem_object_can_migrate(struct drm_i915_gem_object *obj,
+				 enum intel_region_id id);
+
+int i915_gem_object_wait_migration(struct drm_i915_gem_object *obj,
+				   unsigned int flags);
+
+bool i915_gem_object_placement_possible(struct drm_i915_gem_object *obj,
+					enum intel_memory_type type);
+
+bool i915_gem_object_needs_ccs_pages(struct drm_i915_gem_object *obj);
+
+int shmem_sg_alloc_table(struct drm_i915_private *i915, struct sg_table *st,
+			 size_t size, struct intel_memory_region *mr,
+			 struct address_space *mapping,
+			 unsigned int max_segment);
+void shmem_sg_free_table(struct sg_table *st, struct address_space *mapping,
+			 bool dirty, bool backup);
+void __shmem_writeback(size_t size, struct address_space *mapping);
+
+#ifdef CONFIG_MMU_NOTIFIER
+static inline bool
+i915_gem_object_is_userptr(struct drm_i915_gem_object *obj)
+{
+	return obj->userptr.notifier.mm;
+}
+
+int i915_gem_object_userptr_submit_init(struct drm_i915_gem_object *obj);
+int i915_gem_object_userptr_submit_done(struct drm_i915_gem_object *obj);
+int i915_gem_object_userptr_validate(struct drm_i915_gem_object *obj);
+#else
+static inline bool i915_gem_object_is_userptr(struct drm_i915_gem_object *obj) { return false; }
+
+static inline int i915_gem_object_userptr_submit_init(struct drm_i915_gem_object *obj) { GEM_BUG_ON(1); return -ENODEV; }
+static inline int i915_gem_object_userptr_submit_done(struct drm_i915_gem_object *obj) { GEM_BUG_ON(1); return -ENODEV; }
+static inline int i915_gem_object_userptr_validate(struct drm_i915_gem_object *obj) { GEM_BUG_ON(1); return -ENODEV; }
+
+#endif
+
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h
new file mode 100644
index 000000000..1d0d8ee9d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h
@@ -0,0 +1,678 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __I915_GEM_OBJECT_TYPES_H__
+#define __I915_GEM_OBJECT_TYPES_H__
+
+#include <linux/mmu_notifier.h>
+
+#include <drm/drm_gem.h>
+#include <drm/ttm/ttm_bo_api.h>
+#include <uapi/drm/i915_drm.h>
+
+#include "i915_active.h"
+#include "i915_selftest.h"
+#include "i915_vma_resource.h"
+
+struct drm_i915_gem_object;
+struct intel_fronbuffer;
+struct intel_memory_region;
+
+/*
+ * struct i915_lut_handle tracks the fast lookups from handle to vma used
+ * for execbuf. Although we use a radixtree for that mapping, in order to
+ * remove them as the object or context is closed, we need a secondary list
+ * and a translation entry (i915_lut_handle).
+ */
+struct i915_lut_handle {
+	struct list_head obj_link;
+	struct i915_gem_context *ctx;
+	u32 handle;
+};
+
+struct drm_i915_gem_object_ops {
+	unsigned int flags;
+#define I915_GEM_OBJECT_IS_SHRINKABLE			BIT(1)
+/* Skip the shrinker management in set_pages/unset_pages */
+#define I915_GEM_OBJECT_SELF_MANAGED_SHRINK_LIST	BIT(2)
+#define I915_GEM_OBJECT_IS_PROXY			BIT(3)
+#define I915_GEM_OBJECT_NO_MMAP				BIT(4)
+
+	/* Interface between the GEM object and its backing storage.
+	 * get_pages() is called once prior to the use of the associated set
+	 * of pages before to binding them into the GTT, and put_pages() is
+	 * called after we no longer need them. As we expect there to be
+	 * associated cost with migrating pages between the backing storage
+	 * and making them available for the GPU (e.g. clflush), we may hold
+	 * onto the pages after they are no longer referenced by the GPU
+	 * in case they may be used again shortly (for example migrating the
+	 * pages to a different memory domain within the GTT). put_pages()
+	 * will therefore most likely be called when the object itself is
+	 * being released or under memory pressure (where we attempt to
+	 * reap pages for the shrinker).
+	 */
+	int (*get_pages)(struct drm_i915_gem_object *obj);
+	void (*put_pages)(struct drm_i915_gem_object *obj,
+			  struct sg_table *pages);
+	int (*truncate)(struct drm_i915_gem_object *obj);
+	/**
+	 * shrink - Perform further backend specific actions to facilate
+	 * shrinking.
+	 * @obj: The gem object
+	 * @flags: Extra flags to control shrinking behaviour in the backend
+	 *
+	 * Possible values for @flags:
+	 *
+	 * I915_GEM_OBJECT_SHRINK_WRITEBACK - Try to perform writeback of the
+	 * backing pages, if supported.
+	 *
+	 * I915_GEM_OBJECT_SHRINK_NO_GPU_WAIT - Don't wait for the object to
+	 * idle.  Active objects can be considered later. The TTM backend for
+	 * example might have aync migrations going on, which don't use any
+	 * i915_vma to track the active GTT binding, and hence having an unbound
+	 * object might not be enough.
+	 */
+#define I915_GEM_OBJECT_SHRINK_WRITEBACK   BIT(0)
+#define I915_GEM_OBJECT_SHRINK_NO_GPU_WAIT BIT(1)
+	int (*shrink)(struct drm_i915_gem_object *obj, unsigned int flags);
+
+	int (*pread)(struct drm_i915_gem_object *obj,
+		     const struct drm_i915_gem_pread *arg);
+	int (*pwrite)(struct drm_i915_gem_object *obj,
+		      const struct drm_i915_gem_pwrite *arg);
+	u64 (*mmap_offset)(struct drm_i915_gem_object *obj);
+	void (*unmap_virtual)(struct drm_i915_gem_object *obj);
+
+	int (*dmabuf_export)(struct drm_i915_gem_object *obj);
+
+	/**
+	 * adjust_lru - notify that the madvise value was updated
+	 * @obj: The gem object
+	 *
+	 * The madvise value may have been updated, or object was recently
+	 * referenced so act accordingly (Perhaps changing an LRU list etc).
+	 */
+	void (*adjust_lru)(struct drm_i915_gem_object *obj);
+
+	/**
+	 * delayed_free - Override the default delayed free implementation
+	 */
+	void (*delayed_free)(struct drm_i915_gem_object *obj);
+
+	/**
+	 * migrate - Migrate object to a different region either for
+	 * pinning or for as long as the object lock is held.
+	 */
+	int (*migrate)(struct drm_i915_gem_object *obj,
+		       struct intel_memory_region *mr,
+		       unsigned int flags);
+
+	void (*release)(struct drm_i915_gem_object *obj);
+
+	const struct vm_operations_struct *mmap_ops;
+	const char *name; /* friendly name for debug, e.g. lockdep classes */
+};
+
+/**
+ * enum i915_cache_level - The supported GTT caching values for system memory
+ * pages.
+ *
+ * These translate to some special GTT PTE bits when binding pages into some
+ * address space. It also determines whether an object, or rather its pages are
+ * coherent with the GPU, when also reading or writing through the CPU cache
+ * with those pages.
+ *
+ * Userspace can also control this through struct drm_i915_gem_caching.
+ */
+enum i915_cache_level {
+	/**
+	 * @I915_CACHE_NONE:
+	 *
+	 * GPU access is not coherent with the CPU cache. If the cache is dirty
+	 * and we need the underlying pages to be coherent with some later GPU
+	 * access then we need to manually flush the pages.
+	 *
+	 * On shared LLC platforms reads and writes through the CPU cache are
+	 * still coherent even with this setting. See also
+	 * &drm_i915_gem_object.cache_coherent for more details. Due to this we
+	 * should only ever use uncached for scanout surfaces, otherwise we end
+	 * up over-flushing in some places.
+	 *
+	 * This is the default on non-LLC platforms.
+	 */
+	I915_CACHE_NONE = 0,
+	/**
+	 * @I915_CACHE_LLC:
+	 *
+	 * GPU access is coherent with the CPU cache. If the cache is dirty,
+	 * then the GPU will ensure that access remains coherent, when both
+	 * reading and writing through the CPU cache. GPU writes can dirty the
+	 * CPU cache.
+	 *
+	 * Not used for scanout surfaces.
+	 *
+	 * Applies to both platforms with shared LLC(HAS_LLC), and snooping
+	 * based platforms(HAS_SNOOP).
+	 *
+	 * This is the default on shared LLC platforms.  The only exception is
+	 * scanout objects, where the display engine is not coherent with the
+	 * CPU cache. For such objects I915_CACHE_NONE or I915_CACHE_WT is
+	 * automatically applied by the kernel in pin_for_display, if userspace
+	 * has not done so already.
+	 */
+	I915_CACHE_LLC,
+	/**
+	 * @I915_CACHE_L3_LLC:
+	 *
+	 * Explicitly enable the Gfx L3 cache, with coherent LLC.
+	 *
+	 * The Gfx L3 sits between the domain specific caches, e.g
+	 * sampler/render caches, and the larger LLC. LLC is coherent with the
+	 * GPU, but L3 is only visible to the GPU, so likely needs to be flushed
+	 * when the workload completes.
+	 *
+	 * Not used for scanout surfaces.
+	 *
+	 * Only exposed on some gen7 + GGTT. More recent hardware has dropped
+	 * this explicit setting, where it should now be enabled by default.
+	 */
+	I915_CACHE_L3_LLC,
+	/**
+	 * @I915_CACHE_WT:
+	 *
+	 * Write-through. Used for scanout surfaces.
+	 *
+	 * The GPU can utilise the caches, while still having the display engine
+	 * be coherent with GPU writes, as a result we don't need to flush the
+	 * CPU caches when moving out of the render domain. This is the default
+	 * setting chosen by the kernel, if supported by the HW, otherwise we
+	 * fallback to I915_CACHE_NONE. On the CPU side writes through the CPU
+	 * cache still need to be flushed, to remain coherent with the display
+	 * engine.
+	 */
+	I915_CACHE_WT,
+};
+
+enum i915_map_type {
+	I915_MAP_WB = 0,
+	I915_MAP_WC,
+#define I915_MAP_OVERRIDE BIT(31)
+	I915_MAP_FORCE_WB = I915_MAP_WB | I915_MAP_OVERRIDE,
+	I915_MAP_FORCE_WC = I915_MAP_WC | I915_MAP_OVERRIDE,
+};
+
+enum i915_mmap_type {
+	I915_MMAP_TYPE_GTT = 0,
+	I915_MMAP_TYPE_WC,
+	I915_MMAP_TYPE_WB,
+	I915_MMAP_TYPE_UC,
+	I915_MMAP_TYPE_FIXED,
+};
+
+struct i915_mmap_offset {
+	struct drm_vma_offset_node vma_node;
+	struct drm_i915_gem_object *obj;
+	enum i915_mmap_type mmap_type;
+
+	struct rb_node offset;
+};
+
+struct i915_gem_object_page_iter {
+	struct scatterlist *sg_pos;
+	unsigned int sg_idx; /* in pages, but 32bit eek! */
+
+	struct radix_tree_root radix;
+	struct mutex lock; /* protects this cache */
+};
+
+struct drm_i915_gem_object {
+	/*
+	 * We might have reason to revisit the below since it wastes
+	 * a lot of space for non-ttm gem objects.
+	 * In any case, always use the accessors for the ttm_buffer_object
+	 * when accessing it.
+	 */
+	union {
+		struct drm_gem_object base;
+		struct ttm_buffer_object __do_not_access;
+	};
+
+	const struct drm_i915_gem_object_ops *ops;
+
+	struct {
+		/**
+		 * @vma.lock: protect the list/tree of vmas
+		 */
+		spinlock_t lock;
+
+		/**
+		 * @vma.list: List of VMAs backed by this object
+		 *
+		 * The VMA on this list are ordered by type, all GGTT vma are
+		 * placed at the head and all ppGTT vma are placed at the tail.
+		 * The different types of GGTT vma are unordered between
+		 * themselves, use the @vma.tree (which has a defined order
+		 * between all VMA) to quickly find an exact match.
+		 */
+		struct list_head list;
+
+		/**
+		 * @vma.tree: Ordered tree of VMAs backed by this object
+		 *
+		 * All VMA created for this object are placed in the @vma.tree
+		 * for fast retrieval via a binary search in
+		 * i915_vma_instance(). They are also added to @vma.list for
+		 * easy iteration.
+		 */
+		struct rb_root tree;
+	} vma;
+
+	/**
+	 * @lut_list: List of vma lookup entries in use for this object.
+	 *
+	 * If this object is closed, we need to remove all of its VMA from
+	 * the fast lookup index in associated contexts; @lut_list provides
+	 * this translation from object to context->handles_vma.
+	 */
+	struct list_head lut_list;
+	spinlock_t lut_lock; /* guards lut_list */
+
+	/**
+	 * @obj_link: Link into @i915_gem_ww_ctx.obj_list
+	 *
+	 * When we lock this object through i915_gem_object_lock() with a
+	 * context, we add it to the list to ensure we can unlock everything
+	 * when i915_gem_ww_ctx_backoff() or i915_gem_ww_ctx_fini() are called.
+	 */
+	struct list_head obj_link;
+	/**
+	 * @shared_resv_from: The object shares the resv from this vm.
+	 */
+	struct i915_address_space *shares_resv_from;
+
+	union {
+		struct rcu_head rcu;
+		struct llist_node freed;
+	};
+
+	/**
+	 * Whether the object is currently in the GGTT or any other supported
+	 * fake offset mmap backed by lmem.
+	 */
+	unsigned int userfault_count;
+	struct list_head userfault_link;
+
+	struct {
+		spinlock_t lock; /* Protects access to mmo offsets */
+		struct rb_root offsets;
+	} mmo;
+
+	I915_SELFTEST_DECLARE(struct list_head st_link);
+
+	unsigned long flags;
+#define I915_BO_ALLOC_CONTIGUOUS  BIT(0)
+#define I915_BO_ALLOC_VOLATILE    BIT(1)
+#define I915_BO_ALLOC_CPU_CLEAR   BIT(2)
+#define I915_BO_ALLOC_USER        BIT(3)
+/* Object is allowed to lose its contents on suspend / resume, even if pinned */
+#define I915_BO_ALLOC_PM_VOLATILE BIT(4)
+/* Object needs to be restored early using memcpy during resume */
+#define I915_BO_ALLOC_PM_EARLY    BIT(5)
+/*
+ * Object is likely never accessed by the CPU. This will prioritise the BO to be
+ * allocated in the non-mappable portion of lmem. This is merely a hint, and if
+ * dealing with userspace objects the CPU fault handler is free to ignore this.
+ */
+#define I915_BO_ALLOC_GPU_ONLY	  BIT(6)
+#define I915_BO_ALLOC_CCS_AUX	  BIT(7)
+#define I915_BO_ALLOC_FLAGS (I915_BO_ALLOC_CONTIGUOUS | \
+			     I915_BO_ALLOC_VOLATILE | \
+			     I915_BO_ALLOC_CPU_CLEAR | \
+			     I915_BO_ALLOC_USER | \
+			     I915_BO_ALLOC_PM_VOLATILE | \
+			     I915_BO_ALLOC_PM_EARLY | \
+			     I915_BO_ALLOC_GPU_ONLY | \
+			     I915_BO_ALLOC_CCS_AUX)
+#define I915_BO_READONLY          BIT(8)
+#define I915_TILING_QUIRK_BIT     9 /* unknown swizzling; do not release! */
+#define I915_BO_PROTECTED         BIT(10)
+	/**
+	 * @mem_flags - Mutable placement-related flags
+	 *
+	 * These are flags that indicate specifics of the memory region
+	 * the object is currently in. As such they are only stable
+	 * either under the object lock or if the object is pinned.
+	 */
+	unsigned int mem_flags;
+#define I915_BO_FLAG_STRUCT_PAGE BIT(0) /* Object backed by struct pages */
+#define I915_BO_FLAG_IOMEM       BIT(1) /* Object backed by IO memory */
+	/**
+	 * @cache_level: The desired GTT caching level.
+	 *
+	 * See enum i915_cache_level for possible values, along with what
+	 * each does.
+	 */
+	unsigned int cache_level:3;
+	/**
+	 * @cache_coherent:
+	 *
+	 * Track whether the pages are coherent with the GPU if reading or
+	 * writing through the CPU caches. The largely depends on the
+	 * @cache_level setting.
+	 *
+	 * On platforms which don't have the shared LLC(HAS_SNOOP), like on Atom
+	 * platforms, coherency must be explicitly requested with some special
+	 * GTT caching bits(see enum i915_cache_level). When enabling coherency
+	 * it does come at a performance and power cost on such platforms. On
+	 * the flip side the kernel does not need to manually flush any buffers
+	 * which need to be coherent with the GPU, if the object is not coherent
+	 * i.e @cache_coherent is zero.
+	 *
+	 * On platforms that share the LLC with the CPU(HAS_LLC), all GT memory
+	 * access will automatically snoop the CPU caches(even with CACHE_NONE).
+	 * The one exception is when dealing with the display engine, like with
+	 * scanout surfaces. To handle this the kernel will always flush the
+	 * surface out of the CPU caches when preparing it for scanout.  Also
+	 * note that since scanout surfaces are only ever read by the display
+	 * engine we only need to care about flushing any writes through the CPU
+	 * cache, reads on the other hand will always be coherent.
+	 *
+	 * Something strange here is why @cache_coherent is not a simple
+	 * boolean, i.e coherent vs non-coherent. The reasoning for this is back
+	 * to the display engine not being fully coherent. As a result scanout
+	 * surfaces will either be marked as I915_CACHE_NONE or I915_CACHE_WT.
+	 * In the case of seeing I915_CACHE_NONE the kernel makes the assumption
+	 * that this is likely a scanout surface, and will set @cache_coherent
+	 * as only I915_BO_CACHE_COHERENT_FOR_READ, on platforms with the shared
+	 * LLC. The kernel uses this to always flush writes through the CPU
+	 * cache as early as possible, where it can, in effect keeping
+	 * @cache_dirty clean, so we can potentially avoid stalling when
+	 * flushing the surface just before doing the scanout.  This does mean
+	 * we might unnecessarily flush non-scanout objects in some places, but
+	 * the default assumption is that all normal objects should be using
+	 * I915_CACHE_LLC, at least on platforms with the shared LLC.
+	 *
+	 * Supported values:
+	 *
+	 * I915_BO_CACHE_COHERENT_FOR_READ:
+	 *
+	 * On shared LLC platforms, we use this for special scanout surfaces,
+	 * where the display engine is not coherent with the CPU cache. As such
+	 * we need to ensure we flush any writes before doing the scanout. As an
+	 * optimisation we try to flush any writes as early as possible to avoid
+	 * stalling later.
+	 *
+	 * Thus for scanout surfaces using I915_CACHE_NONE, on shared LLC
+	 * platforms, we use:
+	 *
+	 *	cache_coherent = I915_BO_CACHE_COHERENT_FOR_READ
+	 *
+	 * While for normal objects that are fully coherent, including special
+	 * scanout surfaces marked as I915_CACHE_WT, we use:
+	 *
+	 *	cache_coherent = I915_BO_CACHE_COHERENT_FOR_READ |
+	 *			 I915_BO_CACHE_COHERENT_FOR_WRITE
+	 *
+	 * And then for objects that are not coherent at all we use:
+	 *
+	 *	cache_coherent = 0
+	 *
+	 * I915_BO_CACHE_COHERENT_FOR_WRITE:
+	 *
+	 * When writing through the CPU cache, the GPU is still coherent. Note
+	 * that this also implies I915_BO_CACHE_COHERENT_FOR_READ.
+	 */
+#define I915_BO_CACHE_COHERENT_FOR_READ BIT(0)
+#define I915_BO_CACHE_COHERENT_FOR_WRITE BIT(1)
+	unsigned int cache_coherent:2;
+
+	/**
+	 * @cache_dirty:
+	 *
+	 * Track if we are we dirty with writes through the CPU cache for this
+	 * object. As a result reading directly from main memory might yield
+	 * stale data.
+	 *
+	 * This also ties into whether the kernel is tracking the object as
+	 * coherent with the GPU, as per @cache_coherent, as it determines if
+	 * flushing might be needed at various points.
+	 *
+	 * Another part of @cache_dirty is managing flushing when first
+	 * acquiring the pages for system memory, at this point the pages are
+	 * considered foreign, so the default assumption is that the cache is
+	 * dirty, for example the page zeroing done by the kernel might leave
+	 * writes though the CPU cache, or swapping-in, while the actual data in
+	 * main memory is potentially stale.  Note that this is a potential
+	 * security issue when dealing with userspace objects and zeroing. Now,
+	 * whether we actually need apply the big sledgehammer of flushing all
+	 * the pages on acquire depends on if @cache_coherent is marked as
+	 * I915_BO_CACHE_COHERENT_FOR_WRITE, i.e that the GPU will be coherent
+	 * for both reads and writes though the CPU cache.
+	 *
+	 * Note that on shared LLC platforms we still apply the heavy flush for
+	 * I915_CACHE_NONE objects, under the assumption that this is going to
+	 * be used for scanout.
+	 *
+	 * Update: On some hardware there is now also the 'Bypass LLC' MOCS
+	 * entry, which defeats our @cache_coherent tracking, since userspace
+	 * can freely bypass the CPU cache when touching the pages with the GPU,
+	 * where the kernel is completely unaware. On such platform we need
+	 * apply the sledgehammer-on-acquire regardless of the @cache_coherent.
+	 *
+	 * Special care is taken on non-LLC platforms, to prevent potential
+	 * information leak. The driver currently ensures:
+	 *
+	 *   1. All userspace objects, by default, have @cache_level set as
+	 *   I915_CACHE_NONE. The only exception is userptr objects, where we
+	 *   instead force I915_CACHE_LLC, but we also don't allow userspace to
+	 *   ever change the @cache_level for such objects. Another special case
+	 *   is dma-buf, which doesn't rely on @cache_dirty,  but there we
+	 *   always do a forced flush when acquiring the pages, if there is a
+	 *   chance that the pages can be read directly from main memory with
+	 *   the GPU.
+	 *
+	 *   2. All I915_CACHE_NONE objects have @cache_dirty initially true.
+	 *
+	 *   3. All swapped-out objects(i.e shmem) have @cache_dirty set to
+	 *   true.
+	 *
+	 *   4. The @cache_dirty is never freely reset before the initial
+	 *   flush, even if userspace adjusts the @cache_level through the
+	 *   i915_gem_set_caching_ioctl.
+	 *
+	 *   5. All @cache_dirty objects(including swapped-in) are initially
+	 *   flushed with a synchronous call to drm_clflush_sg in
+	 *   __i915_gem_object_set_pages. The @cache_dirty can be freely reset
+	 *   at this point. All further asynchronous clfushes are never security
+	 *   critical, i.e userspace is free to race against itself.
+	 */
+	unsigned int cache_dirty:1;
+
+	/* @is_dpt: Object houses a display page table (DPT) */
+	unsigned int is_dpt:1;
+
+	/**
+	 * @read_domains: Read memory domains.
+	 *
+	 * These monitor which caches contain read/write data related to the
+	 * object. When transitioning from one set of domains to another,
+	 * the driver is called to ensure that caches are suitably flushed and
+	 * invalidated.
+	 */
+	u16 read_domains;
+
+	/**
+	 * @write_domain: Corresponding unique write memory domain.
+	 */
+	u16 write_domain;
+
+	struct intel_frontbuffer __rcu *frontbuffer;
+
+	/** Current tiling stride for the object, if it's tiled. */
+	unsigned int tiling_and_stride;
+#define FENCE_MINIMUM_STRIDE 128 /* See i915_tiling_ok() */
+#define TILING_MASK (FENCE_MINIMUM_STRIDE - 1)
+#define STRIDE_MASK (~TILING_MASK)
+
+	struct {
+		/*
+		 * Protects the pages and their use. Do not use directly, but
+		 * instead go through the pin/unpin interfaces.
+		 */
+		atomic_t pages_pin_count;
+
+		/**
+		 * @shrink_pin: Prevents the pages from being made visible to
+		 * the shrinker, while the shrink_pin is non-zero. Most users
+		 * should pretty much never have to care about this, outside of
+		 * some special use cases.
+		 *
+		 * By default most objects will start out as visible to the
+		 * shrinker(if I915_GEM_OBJECT_IS_SHRINKABLE) as soon as the
+		 * backing pages are attached to the object, like in
+		 * __i915_gem_object_set_pages(). They will then be removed the
+		 * shrinker list once the pages are released.
+		 *
+		 * The @shrink_pin is incremented by calling
+		 * i915_gem_object_make_unshrinkable(), which will also remove
+		 * the object from the shrinker list, if the pin count was zero.
+		 *
+		 * Callers will then typically call
+		 * i915_gem_object_make_shrinkable() or
+		 * i915_gem_object_make_purgeable() to decrement the pin count,
+		 * and make the pages visible again.
+		 */
+		atomic_t shrink_pin;
+
+		/**
+		 * @ttm_shrinkable: True when the object is using shmem pages
+		 * underneath. Protected by the object lock.
+		 */
+		bool ttm_shrinkable;
+
+		/**
+		 * @unknown_state: Indicate that the object is effectively
+		 * borked. This is write-once and set if we somehow encounter a
+		 * fatal error when moving/clearing the pages, and we are not
+		 * able to fallback to memcpy/memset, like on small-BAR systems.
+		 * The GPU should also be wedged (or in the process) at this
+		 * point.
+		 *
+		 * Only valid to read this after acquiring the dma-resv lock and
+		 * waiting for all DMA_RESV_USAGE_KERNEL fences to be signalled,
+		 * or if we otherwise know that the moving fence has signalled,
+		 * and we are certain the pages underneath are valid for
+		 * immediate access (under normal operation), like just prior to
+		 * binding the object or when setting up the CPU fault handler.
+		 * See i915_gem_object_has_unknown_state();
+		 */
+		bool unknown_state;
+
+		/**
+		 * Priority list of potential placements for this object.
+		 */
+		struct intel_memory_region **placements;
+		int n_placements;
+
+		/**
+		 * Memory region for this object.
+		 */
+		struct intel_memory_region *region;
+
+		/**
+		 * Memory manager resource allocated for this object. Only
+		 * needed for the mock region.
+		 */
+		struct ttm_resource *res;
+
+		/**
+		 * Element within memory_region->objects or region->purgeable
+		 * if the object is marked as DONTNEED. Access is protected by
+		 * region->obj_lock.
+		 */
+		struct list_head region_link;
+
+		struct i915_refct_sgt *rsgt;
+		struct sg_table *pages;
+		void *mapping;
+
+		struct i915_page_sizes page_sizes;
+
+		I915_SELFTEST_DECLARE(unsigned int page_mask);
+
+		struct i915_gem_object_page_iter get_page;
+		struct i915_gem_object_page_iter get_dma_page;
+
+		/**
+		 * Element within i915->mm.shrink_list or i915->mm.purge_list,
+		 * locked by i915->mm.obj_lock.
+		 */
+		struct list_head link;
+
+		/**
+		 * Advice: are the backing pages purgeable?
+		 */
+		unsigned int madv:2;
+
+		/**
+		 * This is set if the object has been written to since the
+		 * pages were last acquired.
+		 */
+		bool dirty:1;
+
+		u32 tlb;
+	} mm;
+
+	struct {
+		struct i915_refct_sgt *cached_io_rsgt;
+		struct i915_gem_object_page_iter get_io_page;
+		struct drm_i915_gem_object *backup;
+		bool created:1;
+	} ttm;
+
+	/*
+	 * Record which PXP key instance this object was created against (if
+	 * any), so we can use it to determine if the encryption is valid by
+	 * comparing against the current key instance.
+	 */
+	u32 pxp_key_instance;
+
+	/** Record of address bit 17 of each page at last unbind. */
+	unsigned long *bit_17;
+
+	union {
+#ifdef CONFIG_MMU_NOTIFIER
+		struct i915_gem_userptr {
+			uintptr_t ptr;
+			unsigned long notifier_seq;
+
+			struct mmu_interval_notifier notifier;
+			struct page **pvec;
+			int page_ref;
+		} userptr;
+#endif
+
+		struct drm_mm_node *stolen;
+
+		resource_size_t bo_offset;
+
+		unsigned long scratch;
+		u64 encode;
+
+		void *gvt_info;
+	};
+};
+
+static inline struct drm_i915_gem_object *
+to_intel_bo(struct drm_gem_object *gem)
+{
+	/* Assert that to_intel_bo(NULL) == NULL */
+	BUILD_BUG_ON(offsetof(struct drm_i915_gem_object, base));
+
+	return container_of(gem, struct drm_i915_gem_object, base);
+}
+
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pages.c b/drivers/gpu/drm/i915/gem/i915_gem_pages.c
new file mode 100644
index 000000000..4df50b049
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_pages.c
@@ -0,0 +1,674 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2014-2016 Intel Corporation
+ */
+
+#include <drm/drm_cache.h>
+
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_pm.h"
+
+#include "i915_drv.h"
+#include "i915_gem_object.h"
+#include "i915_scatterlist.h"
+#include "i915_gem_lmem.h"
+#include "i915_gem_mman.h"
+
+void __i915_gem_object_set_pages(struct drm_i915_gem_object *obj,
+				 struct sg_table *pages,
+				 unsigned int sg_page_sizes)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	unsigned long supported = RUNTIME_INFO(i915)->page_sizes;
+	bool shrinkable;
+	int i;
+
+	assert_object_held_shared(obj);
+
+	if (i915_gem_object_is_volatile(obj))
+		obj->mm.madv = I915_MADV_DONTNEED;
+
+	/* Make the pages coherent with the GPU (flushing any swapin). */
+	if (obj->cache_dirty) {
+		WARN_ON_ONCE(IS_DGFX(i915));
+		obj->write_domain = 0;
+		if (i915_gem_object_has_struct_page(obj))
+			drm_clflush_sg(pages);
+		obj->cache_dirty = false;
+	}
+
+	obj->mm.get_page.sg_pos = pages->sgl;
+	obj->mm.get_page.sg_idx = 0;
+	obj->mm.get_dma_page.sg_pos = pages->sgl;
+	obj->mm.get_dma_page.sg_idx = 0;
+
+	obj->mm.pages = pages;
+
+	GEM_BUG_ON(!sg_page_sizes);
+	obj->mm.page_sizes.phys = sg_page_sizes;
+
+	/*
+	 * Calculate the supported page-sizes which fit into the given
+	 * sg_page_sizes. This will give us the page-sizes which we may be able
+	 * to use opportunistically when later inserting into the GTT. For
+	 * example if phys=2G, then in theory we should be able to use 1G, 2M,
+	 * 64K or 4K pages, although in practice this will depend on a number of
+	 * other factors.
+	 */
+	obj->mm.page_sizes.sg = 0;
+	for_each_set_bit(i, &supported, ilog2(I915_GTT_MAX_PAGE_SIZE) + 1) {
+		if (obj->mm.page_sizes.phys & ~0u << i)
+			obj->mm.page_sizes.sg |= BIT(i);
+	}
+	GEM_BUG_ON(!HAS_PAGE_SIZES(i915, obj->mm.page_sizes.sg));
+
+	shrinkable = i915_gem_object_is_shrinkable(obj);
+
+	if (i915_gem_object_is_tiled(obj) &&
+	    i915->gem_quirks & GEM_QUIRK_PIN_SWIZZLED_PAGES) {
+		GEM_BUG_ON(i915_gem_object_has_tiling_quirk(obj));
+		i915_gem_object_set_tiling_quirk(obj);
+		GEM_BUG_ON(!list_empty(&obj->mm.link));
+		atomic_inc(&obj->mm.shrink_pin);
+		shrinkable = false;
+	}
+
+	if (shrinkable && !i915_gem_object_has_self_managed_shrink_list(obj)) {
+		struct list_head *list;
+		unsigned long flags;
+
+		assert_object_held(obj);
+		spin_lock_irqsave(&i915->mm.obj_lock, flags);
+
+		i915->mm.shrink_count++;
+		i915->mm.shrink_memory += obj->base.size;
+
+		if (obj->mm.madv != I915_MADV_WILLNEED)
+			list = &i915->mm.purge_list;
+		else
+			list = &i915->mm.shrink_list;
+		list_add_tail(&obj->mm.link, list);
+
+		atomic_set(&obj->mm.shrink_pin, 0);
+		spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+	}
+}
+
+int ____i915_gem_object_get_pages(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	int err;
+
+	assert_object_held_shared(obj);
+
+	if (unlikely(obj->mm.madv != I915_MADV_WILLNEED)) {
+		drm_dbg(&i915->drm,
+			"Attempting to obtain a purgeable object\n");
+		return -EFAULT;
+	}
+
+	err = obj->ops->get_pages(obj);
+	GEM_BUG_ON(!err && !i915_gem_object_has_pages(obj));
+
+	return err;
+}
+
+/* Ensure that the associated pages are gathered from the backing storage
+ * and pinned into our object. i915_gem_object_pin_pages() may be called
+ * multiple times before they are released by a single call to
+ * i915_gem_object_unpin_pages() - once the pages are no longer referenced
+ * either as a result of memory pressure (reaping pages under the shrinker)
+ * or as the object is itself released.
+ */
+int __i915_gem_object_get_pages(struct drm_i915_gem_object *obj)
+{
+	int err;
+
+	assert_object_held(obj);
+
+	assert_object_held_shared(obj);
+
+	if (unlikely(!i915_gem_object_has_pages(obj))) {
+		GEM_BUG_ON(i915_gem_object_has_pinned_pages(obj));
+
+		err = ____i915_gem_object_get_pages(obj);
+		if (err)
+			return err;
+
+		smp_mb__before_atomic();
+	}
+	atomic_inc(&obj->mm.pages_pin_count);
+
+	return 0;
+}
+
+int i915_gem_object_pin_pages_unlocked(struct drm_i915_gem_object *obj)
+{
+	struct i915_gem_ww_ctx ww;
+	int err;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	err = i915_gem_object_lock(obj, &ww);
+	if (!err)
+		err = i915_gem_object_pin_pages(obj);
+
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	return err;
+}
+
+/* Immediately discard the backing storage */
+int i915_gem_object_truncate(struct drm_i915_gem_object *obj)
+{
+	if (obj->ops->truncate)
+		return obj->ops->truncate(obj);
+
+	return 0;
+}
+
+static void __i915_gem_object_reset_page_iter(struct drm_i915_gem_object *obj)
+{
+	struct radix_tree_iter iter;
+	void __rcu **slot;
+
+	rcu_read_lock();
+	radix_tree_for_each_slot(slot, &obj->mm.get_page.radix, &iter, 0)
+		radix_tree_delete(&obj->mm.get_page.radix, iter.index);
+	radix_tree_for_each_slot(slot, &obj->mm.get_dma_page.radix, &iter, 0)
+		radix_tree_delete(&obj->mm.get_dma_page.radix, iter.index);
+	rcu_read_unlock();
+}
+
+static void unmap_object(struct drm_i915_gem_object *obj, void *ptr)
+{
+	if (is_vmalloc_addr(ptr))
+		vunmap(ptr);
+}
+
+static void flush_tlb_invalidate(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct intel_gt *gt = to_gt(i915);
+
+	if (!obj->mm.tlb)
+		return;
+
+	intel_gt_invalidate_tlb(gt, obj->mm.tlb);
+	obj->mm.tlb = 0;
+}
+
+struct sg_table *
+__i915_gem_object_unset_pages(struct drm_i915_gem_object *obj)
+{
+	struct sg_table *pages;
+
+	assert_object_held_shared(obj);
+
+	pages = fetch_and_zero(&obj->mm.pages);
+	if (IS_ERR_OR_NULL(pages))
+		return pages;
+
+	if (i915_gem_object_is_volatile(obj))
+		obj->mm.madv = I915_MADV_WILLNEED;
+
+	if (!i915_gem_object_has_self_managed_shrink_list(obj))
+		i915_gem_object_make_unshrinkable(obj);
+
+	if (obj->mm.mapping) {
+		unmap_object(obj, page_mask_bits(obj->mm.mapping));
+		obj->mm.mapping = NULL;
+	}
+
+	__i915_gem_object_reset_page_iter(obj);
+	obj->mm.page_sizes.phys = obj->mm.page_sizes.sg = 0;
+
+	flush_tlb_invalidate(obj);
+
+	return pages;
+}
+
+int __i915_gem_object_put_pages(struct drm_i915_gem_object *obj)
+{
+	struct sg_table *pages;
+
+	if (i915_gem_object_has_pinned_pages(obj))
+		return -EBUSY;
+
+	/* May be called by shrinker from within get_pages() (on another bo) */
+	assert_object_held_shared(obj);
+
+	i915_gem_object_release_mmap_offset(obj);
+
+	/*
+	 * ->put_pages might need to allocate memory for the bit17 swizzle
+	 * array, hence protect them from being reaped by removing them from gtt
+	 * lists early.
+	 */
+	pages = __i915_gem_object_unset_pages(obj);
+
+	/*
+	 * XXX Temporary hijinx to avoid updating all backends to handle
+	 * NULL pages. In the future, when we have more asynchronous
+	 * get_pages backends we should be better able to handle the
+	 * cancellation of the async task in a more uniform manner.
+	 */
+	if (!IS_ERR_OR_NULL(pages))
+		obj->ops->put_pages(obj, pages);
+
+	return 0;
+}
+
+/* The 'mapping' part of i915_gem_object_pin_map() below */
+static void *i915_gem_object_map_page(struct drm_i915_gem_object *obj,
+				      enum i915_map_type type)
+{
+	unsigned long n_pages = obj->base.size >> PAGE_SHIFT, i;
+	struct page *stack[32], **pages = stack, *page;
+	struct sgt_iter iter;
+	pgprot_t pgprot;
+	void *vaddr;
+
+	switch (type) {
+	default:
+		MISSING_CASE(type);
+		fallthrough;	/* to use PAGE_KERNEL anyway */
+	case I915_MAP_WB:
+		/*
+		 * On 32b, highmem using a finite set of indirect PTE (i.e.
+		 * vmap) to provide virtual mappings of the high pages.
+		 * As these are finite, map_new_virtual() must wait for some
+		 * other kmap() to finish when it runs out. If we map a large
+		 * number of objects, there is no method for it to tell us
+		 * to release the mappings, and we deadlock.
+		 *
+		 * However, if we make an explicit vmap of the page, that
+		 * uses a larger vmalloc arena, and also has the ability
+		 * to tell us to release unwanted mappings. Most importantly,
+		 * it will fail and propagate an error instead of waiting
+		 * forever.
+		 *
+		 * So if the page is beyond the 32b boundary, make an explicit
+		 * vmap.
+		 */
+		if (n_pages == 1 && !PageHighMem(sg_page(obj->mm.pages->sgl)))
+			return page_address(sg_page(obj->mm.pages->sgl));
+		pgprot = PAGE_KERNEL;
+		break;
+	case I915_MAP_WC:
+		pgprot = pgprot_writecombine(PAGE_KERNEL_IO);
+		break;
+	}
+
+	if (n_pages > ARRAY_SIZE(stack)) {
+		/* Too big for stack -- allocate temporary array instead */
+		pages = kvmalloc_array(n_pages, sizeof(*pages), GFP_KERNEL);
+		if (!pages)
+			return ERR_PTR(-ENOMEM);
+	}
+
+	i = 0;
+	for_each_sgt_page(page, iter, obj->mm.pages)
+		pages[i++] = page;
+	vaddr = vmap(pages, n_pages, 0, pgprot);
+	if (pages != stack)
+		kvfree(pages);
+
+	return vaddr ?: ERR_PTR(-ENOMEM);
+}
+
+static void *i915_gem_object_map_pfn(struct drm_i915_gem_object *obj,
+				     enum i915_map_type type)
+{
+	resource_size_t iomap = obj->mm.region->iomap.base -
+		obj->mm.region->region.start;
+	unsigned long n_pfn = obj->base.size >> PAGE_SHIFT;
+	unsigned long stack[32], *pfns = stack, i;
+	struct sgt_iter iter;
+	dma_addr_t addr;
+	void *vaddr;
+
+	GEM_BUG_ON(type != I915_MAP_WC);
+
+	if (n_pfn > ARRAY_SIZE(stack)) {
+		/* Too big for stack -- allocate temporary array instead */
+		pfns = kvmalloc_array(n_pfn, sizeof(*pfns), GFP_KERNEL);
+		if (!pfns)
+			return ERR_PTR(-ENOMEM);
+	}
+
+	i = 0;
+	for_each_sgt_daddr(addr, iter, obj->mm.pages)
+		pfns[i++] = (iomap + addr) >> PAGE_SHIFT;
+	vaddr = vmap_pfn(pfns, n_pfn, pgprot_writecombine(PAGE_KERNEL_IO));
+	if (pfns != stack)
+		kvfree(pfns);
+
+	return vaddr ?: ERR_PTR(-ENOMEM);
+}
+
+/* get, pin, and map the pages of the object into kernel space */
+void *i915_gem_object_pin_map(struct drm_i915_gem_object *obj,
+			      enum i915_map_type type)
+{
+	enum i915_map_type has_type;
+	bool pinned;
+	void *ptr;
+	int err;
+
+	if (!i915_gem_object_has_struct_page(obj) &&
+	    !i915_gem_object_has_iomem(obj))
+		return ERR_PTR(-ENXIO);
+
+	if (WARN_ON_ONCE(obj->flags & I915_BO_ALLOC_GPU_ONLY))
+		return ERR_PTR(-EINVAL);
+
+	assert_object_held(obj);
+
+	pinned = !(type & I915_MAP_OVERRIDE);
+	type &= ~I915_MAP_OVERRIDE;
+
+	if (!atomic_inc_not_zero(&obj->mm.pages_pin_count)) {
+		if (unlikely(!i915_gem_object_has_pages(obj))) {
+			GEM_BUG_ON(i915_gem_object_has_pinned_pages(obj));
+
+			err = ____i915_gem_object_get_pages(obj);
+			if (err)
+				return ERR_PTR(err);
+
+			smp_mb__before_atomic();
+		}
+		atomic_inc(&obj->mm.pages_pin_count);
+		pinned = false;
+	}
+	GEM_BUG_ON(!i915_gem_object_has_pages(obj));
+
+	/*
+	 * For discrete our CPU mappings needs to be consistent in order to
+	 * function correctly on !x86. When mapping things through TTM, we use
+	 * the same rules to determine the caching type.
+	 *
+	 * The caching rules, starting from DG1:
+	 *
+	 *	- If the object can be placed in device local-memory, then the
+	 *	  pages should be allocated and mapped as write-combined only.
+	 *
+	 *	- Everything else is always allocated and mapped as write-back,
+	 *	  with the guarantee that everything is also coherent with the
+	 *	  GPU.
+	 *
+	 * Internal users of lmem are already expected to get this right, so no
+	 * fudging needed there.
+	 */
+	if (i915_gem_object_placement_possible(obj, INTEL_MEMORY_LOCAL)) {
+		if (type != I915_MAP_WC && !obj->mm.n_placements) {
+			ptr = ERR_PTR(-ENODEV);
+			goto err_unpin;
+		}
+
+		type = I915_MAP_WC;
+	} else if (IS_DGFX(to_i915(obj->base.dev))) {
+		type = I915_MAP_WB;
+	}
+
+	ptr = page_unpack_bits(obj->mm.mapping, &has_type);
+	if (ptr && has_type != type) {
+		if (pinned) {
+			ptr = ERR_PTR(-EBUSY);
+			goto err_unpin;
+		}
+
+		unmap_object(obj, ptr);
+
+		ptr = obj->mm.mapping = NULL;
+	}
+
+	if (!ptr) {
+		err = i915_gem_object_wait_moving_fence(obj, true);
+		if (err) {
+			ptr = ERR_PTR(err);
+			goto err_unpin;
+		}
+
+		if (GEM_WARN_ON(type == I915_MAP_WC && !pat_enabled()))
+			ptr = ERR_PTR(-ENODEV);
+		else if (i915_gem_object_has_struct_page(obj))
+			ptr = i915_gem_object_map_page(obj, type);
+		else
+			ptr = i915_gem_object_map_pfn(obj, type);
+		if (IS_ERR(ptr))
+			goto err_unpin;
+
+		obj->mm.mapping = page_pack_bits(ptr, type);
+	}
+
+	return ptr;
+
+err_unpin:
+	atomic_dec(&obj->mm.pages_pin_count);
+	return ptr;
+}
+
+void *i915_gem_object_pin_map_unlocked(struct drm_i915_gem_object *obj,
+				       enum i915_map_type type)
+{
+	void *ret;
+
+	i915_gem_object_lock(obj, NULL);
+	ret = i915_gem_object_pin_map(obj, type);
+	i915_gem_object_unlock(obj);
+
+	return ret;
+}
+
+void __i915_gem_object_flush_map(struct drm_i915_gem_object *obj,
+				 unsigned long offset,
+				 unsigned long size)
+{
+	enum i915_map_type has_type;
+	void *ptr;
+
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+	GEM_BUG_ON(range_overflows_t(typeof(obj->base.size),
+				     offset, size, obj->base.size));
+
+	wmb(); /* let all previous writes be visible to coherent partners */
+	obj->mm.dirty = true;
+
+	if (obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_WRITE)
+		return;
+
+	ptr = page_unpack_bits(obj->mm.mapping, &has_type);
+	if (has_type == I915_MAP_WC)
+		return;
+
+	drm_clflush_virt_range(ptr + offset, size);
+	if (size == obj->base.size) {
+		obj->write_domain &= ~I915_GEM_DOMAIN_CPU;
+		obj->cache_dirty = false;
+	}
+}
+
+void __i915_gem_object_release_map(struct drm_i915_gem_object *obj)
+{
+	GEM_BUG_ON(!obj->mm.mapping);
+
+	/*
+	 * We allow removing the mapping from underneath pinned pages!
+	 *
+	 * Furthermore, since this is an unsafe operation reserved only
+	 * for construction time manipulation, we ignore locking prudence.
+	 */
+	unmap_object(obj, page_mask_bits(fetch_and_zero(&obj->mm.mapping)));
+
+	i915_gem_object_unpin_map(obj);
+}
+
+struct scatterlist *
+__i915_gem_object_get_sg(struct drm_i915_gem_object *obj,
+			 struct i915_gem_object_page_iter *iter,
+			 unsigned int n,
+			 unsigned int *offset,
+			 bool dma)
+{
+	struct scatterlist *sg;
+	unsigned int idx, count;
+
+	might_sleep();
+	GEM_BUG_ON(n >= obj->base.size >> PAGE_SHIFT);
+	if (!i915_gem_object_has_pinned_pages(obj))
+		assert_object_held(obj);
+
+	/* As we iterate forward through the sg, we record each entry in a
+	 * radixtree for quick repeated (backwards) lookups. If we have seen
+	 * this index previously, we will have an entry for it.
+	 *
+	 * Initial lookup is O(N), but this is amortized to O(1) for
+	 * sequential page access (where each new request is consecutive
+	 * to the previous one). Repeated lookups are O(lg(obj->base.size)),
+	 * i.e. O(1) with a large constant!
+	 */
+	if (n < READ_ONCE(iter->sg_idx))
+		goto lookup;
+
+	mutex_lock(&iter->lock);
+
+	/* We prefer to reuse the last sg so that repeated lookup of this
+	 * (or the subsequent) sg are fast - comparing against the last
+	 * sg is faster than going through the radixtree.
+	 */
+
+	sg = iter->sg_pos;
+	idx = iter->sg_idx;
+	count = dma ? __sg_dma_page_count(sg) : __sg_page_count(sg);
+
+	while (idx + count <= n) {
+		void *entry;
+		unsigned long i;
+		int ret;
+
+		/* If we cannot allocate and insert this entry, or the
+		 * individual pages from this range, cancel updating the
+		 * sg_idx so that on this lookup we are forced to linearly
+		 * scan onwards, but on future lookups we will try the
+		 * insertion again (in which case we need to be careful of
+		 * the error return reporting that we have already inserted
+		 * this index).
+		 */
+		ret = radix_tree_insert(&iter->radix, idx, sg);
+		if (ret && ret != -EEXIST)
+			goto scan;
+
+		entry = xa_mk_value(idx);
+		for (i = 1; i < count; i++) {
+			ret = radix_tree_insert(&iter->radix, idx + i, entry);
+			if (ret && ret != -EEXIST)
+				goto scan;
+		}
+
+		idx += count;
+		sg = ____sg_next(sg);
+		count = dma ? __sg_dma_page_count(sg) : __sg_page_count(sg);
+	}
+
+scan:
+	iter->sg_pos = sg;
+	iter->sg_idx = idx;
+
+	mutex_unlock(&iter->lock);
+
+	if (unlikely(n < idx)) /* insertion completed by another thread */
+		goto lookup;
+
+	/* In case we failed to insert the entry into the radixtree, we need
+	 * to look beyond the current sg.
+	 */
+	while (idx + count <= n) {
+		idx += count;
+		sg = ____sg_next(sg);
+		count = dma ? __sg_dma_page_count(sg) : __sg_page_count(sg);
+	}
+
+	*offset = n - idx;
+	return sg;
+
+lookup:
+	rcu_read_lock();
+
+	sg = radix_tree_lookup(&iter->radix, n);
+	GEM_BUG_ON(!sg);
+
+	/* If this index is in the middle of multi-page sg entry,
+	 * the radix tree will contain a value entry that points
+	 * to the start of that range. We will return the pointer to
+	 * the base page and the offset of this page within the
+	 * sg entry's range.
+	 */
+	*offset = 0;
+	if (unlikely(xa_is_value(sg))) {
+		unsigned long base = xa_to_value(sg);
+
+		sg = radix_tree_lookup(&iter->radix, base);
+		GEM_BUG_ON(!sg);
+
+		*offset = n - base;
+	}
+
+	rcu_read_unlock();
+
+	return sg;
+}
+
+struct page *
+i915_gem_object_get_page(struct drm_i915_gem_object *obj, unsigned int n)
+{
+	struct scatterlist *sg;
+	unsigned int offset;
+
+	GEM_BUG_ON(!i915_gem_object_has_struct_page(obj));
+
+	sg = i915_gem_object_get_sg(obj, n, &offset);
+	return nth_page(sg_page(sg), offset);
+}
+
+/* Like i915_gem_object_get_page(), but mark the returned page dirty */
+struct page *
+i915_gem_object_get_dirty_page(struct drm_i915_gem_object *obj,
+			       unsigned int n)
+{
+	struct page *page;
+
+	page = i915_gem_object_get_page(obj, n);
+	if (!obj->mm.dirty)
+		set_page_dirty(page);
+
+	return page;
+}
+
+dma_addr_t
+i915_gem_object_get_dma_address_len(struct drm_i915_gem_object *obj,
+				    unsigned long n,
+				    unsigned int *len)
+{
+	struct scatterlist *sg;
+	unsigned int offset;
+
+	sg = i915_gem_object_get_sg_dma(obj, n, &offset);
+
+	if (len)
+		*len = sg_dma_len(sg) - (offset << PAGE_SHIFT);
+
+	return sg_dma_address(sg) + (offset << PAGE_SHIFT);
+}
+
+dma_addr_t
+i915_gem_object_get_dma_address(struct drm_i915_gem_object *obj,
+				unsigned long n)
+{
+	return i915_gem_object_get_dma_address_len(obj, n, NULL);
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_phys.c b/drivers/gpu/drm/i915/gem/i915_gem_phys.c
new file mode 100644
index 000000000..0d0e46dae
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_phys.c
@@ -0,0 +1,259 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2014-2016 Intel Corporation
+ */
+
+#include <linux/highmem.h>
+#include <linux/shmem_fs.h>
+#include <linux/swap.h>
+
+#include <drm/drm_cache.h>
+
+#include "gt/intel_gt.h"
+#include "i915_drv.h"
+#include "i915_gem_object.h"
+#include "i915_gem_region.h"
+#include "i915_gem_tiling.h"
+#include "i915_scatterlist.h"
+
+static int i915_gem_object_get_pages_phys(struct drm_i915_gem_object *obj)
+{
+	struct address_space *mapping = obj->base.filp->f_mapping;
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct scatterlist *sg;
+	struct sg_table *st;
+	dma_addr_t dma;
+	void *vaddr;
+	void *dst;
+	int i;
+
+	if (GEM_WARN_ON(i915_gem_object_needs_bit17_swizzle(obj)))
+		return -EINVAL;
+
+	/*
+	 * Always aligning to the object size, allows a single allocation
+	 * to handle all possible callers, and given typical object sizes,
+	 * the alignment of the buddy allocation will naturally match.
+	 */
+	vaddr = dma_alloc_coherent(obj->base.dev->dev,
+				   roundup_pow_of_two(obj->base.size),
+				   &dma, GFP_KERNEL);
+	if (!vaddr)
+		return -ENOMEM;
+
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (!st)
+		goto err_pci;
+
+	if (sg_alloc_table(st, 1, GFP_KERNEL))
+		goto err_st;
+
+	sg = st->sgl;
+	sg->offset = 0;
+	sg->length = obj->base.size;
+
+	sg_assign_page(sg, (struct page *)vaddr);
+	sg_dma_address(sg) = dma;
+	sg_dma_len(sg) = obj->base.size;
+
+	dst = vaddr;
+	for (i = 0; i < obj->base.size / PAGE_SIZE; i++) {
+		struct page *page;
+		void *src;
+
+		page = shmem_read_mapping_page(mapping, i);
+		if (IS_ERR(page))
+			goto err_st;
+
+		src = kmap_atomic(page);
+		memcpy(dst, src, PAGE_SIZE);
+		drm_clflush_virt_range(dst, PAGE_SIZE);
+		kunmap_atomic(src);
+
+		put_page(page);
+		dst += PAGE_SIZE;
+	}
+
+	intel_gt_chipset_flush(to_gt(i915));
+
+	/* We're no longer struct page backed */
+	obj->mem_flags &= ~I915_BO_FLAG_STRUCT_PAGE;
+	__i915_gem_object_set_pages(obj, st, sg->length);
+
+	return 0;
+
+err_st:
+	kfree(st);
+err_pci:
+	dma_free_coherent(obj->base.dev->dev,
+			  roundup_pow_of_two(obj->base.size),
+			  vaddr, dma);
+	return -ENOMEM;
+}
+
+void
+i915_gem_object_put_pages_phys(struct drm_i915_gem_object *obj,
+			       struct sg_table *pages)
+{
+	dma_addr_t dma = sg_dma_address(pages->sgl);
+	void *vaddr = sg_page(pages->sgl);
+
+	__i915_gem_object_release_shmem(obj, pages, false);
+
+	if (obj->mm.dirty) {
+		struct address_space *mapping = obj->base.filp->f_mapping;
+		void *src = vaddr;
+		int i;
+
+		for (i = 0; i < obj->base.size / PAGE_SIZE; i++) {
+			struct page *page;
+			char *dst;
+
+			page = shmem_read_mapping_page(mapping, i);
+			if (IS_ERR(page))
+				continue;
+
+			dst = kmap_atomic(page);
+			drm_clflush_virt_range(src, PAGE_SIZE);
+			memcpy(dst, src, PAGE_SIZE);
+			kunmap_atomic(dst);
+
+			set_page_dirty(page);
+			if (obj->mm.madv == I915_MADV_WILLNEED)
+				mark_page_accessed(page);
+			put_page(page);
+
+			src += PAGE_SIZE;
+		}
+		obj->mm.dirty = false;
+	}
+
+	sg_free_table(pages);
+	kfree(pages);
+
+	dma_free_coherent(obj->base.dev->dev,
+			  roundup_pow_of_two(obj->base.size),
+			  vaddr, dma);
+}
+
+int i915_gem_object_pwrite_phys(struct drm_i915_gem_object *obj,
+				const struct drm_i915_gem_pwrite *args)
+{
+	void *vaddr = sg_page(obj->mm.pages->sgl) + args->offset;
+	char __user *user_data = u64_to_user_ptr(args->data_ptr);
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	int err;
+
+	err = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE |
+				   I915_WAIT_ALL,
+				   MAX_SCHEDULE_TIMEOUT);
+	if (err)
+		return err;
+
+	/*
+	 * We manually control the domain here and pretend that it
+	 * remains coherent i.e. in the GTT domain, like shmem_pwrite.
+	 */
+	i915_gem_object_invalidate_frontbuffer(obj, ORIGIN_CPU);
+
+	if (copy_from_user(vaddr, user_data, args->size))
+		return -EFAULT;
+
+	drm_clflush_virt_range(vaddr, args->size);
+	intel_gt_chipset_flush(to_gt(i915));
+
+	i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
+	return 0;
+}
+
+int i915_gem_object_pread_phys(struct drm_i915_gem_object *obj,
+			       const struct drm_i915_gem_pread *args)
+{
+	void *vaddr = sg_page(obj->mm.pages->sgl) + args->offset;
+	char __user *user_data = u64_to_user_ptr(args->data_ptr);
+	int err;
+
+	err = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE,
+				   MAX_SCHEDULE_TIMEOUT);
+	if (err)
+		return err;
+
+	drm_clflush_virt_range(vaddr, args->size);
+	if (copy_to_user(user_data, vaddr, args->size))
+		return -EFAULT;
+
+	return 0;
+}
+
+static int i915_gem_object_shmem_to_phys(struct drm_i915_gem_object *obj)
+{
+	struct sg_table *pages;
+	int err;
+
+	pages = __i915_gem_object_unset_pages(obj);
+
+	err = i915_gem_object_get_pages_phys(obj);
+	if (err)
+		goto err_xfer;
+
+	/* Perma-pin (until release) the physical set of pages */
+	__i915_gem_object_pin_pages(obj);
+
+	if (!IS_ERR_OR_NULL(pages))
+		i915_gem_object_put_pages_shmem(obj, pages);
+
+	i915_gem_object_release_memory_region(obj);
+	return 0;
+
+err_xfer:
+	if (!IS_ERR_OR_NULL(pages)) {
+		unsigned int sg_page_sizes = i915_sg_dma_sizes(pages->sgl);
+
+		__i915_gem_object_set_pages(obj, pages, sg_page_sizes);
+	}
+	return err;
+}
+
+int i915_gem_object_attach_phys(struct drm_i915_gem_object *obj, int align)
+{
+	int err;
+
+	assert_object_held(obj);
+
+	if (align > obj->base.size)
+		return -EINVAL;
+
+	if (!i915_gem_object_is_shmem(obj))
+		return -EINVAL;
+
+	if (!i915_gem_object_has_struct_page(obj))
+		return 0;
+
+	err = i915_gem_object_unbind(obj, I915_GEM_OBJECT_UNBIND_ACTIVE);
+	if (err)
+		return err;
+
+	if (obj->mm.madv != I915_MADV_WILLNEED)
+		return -EFAULT;
+
+	if (i915_gem_object_has_tiling_quirk(obj))
+		return -EFAULT;
+
+	if (obj->mm.mapping || i915_gem_object_has_pinned_pages(obj))
+		return -EBUSY;
+
+	if (unlikely(obj->mm.madv != I915_MADV_WILLNEED)) {
+		drm_dbg(obj->base.dev,
+			"Attempting to obtain a purgeable object\n");
+		return -EFAULT;
+	}
+
+	return i915_gem_object_shmem_to_phys(obj);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/i915_gem_phys.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pm.c b/drivers/gpu/drm/i915/gem/i915_gem_pm.c
new file mode 100644
index 000000000..8d30db5e6
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_pm.c
@@ -0,0 +1,231 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "gem/i915_gem_pm.h"
+#include "gem/i915_gem_ttm_pm.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_pm.h"
+#include "gt/intel_gt_requests.h"
+
+#include "i915_driver.h"
+#include "i915_drv.h"
+
+#if defined(CONFIG_X86)
+#include <asm/smp.h>
+#else
+#define wbinvd_on_all_cpus() \
+	pr_warn(DRIVER_NAME ": Missing cache flush in %s\n", __func__)
+#endif
+
+void i915_gem_suspend(struct drm_i915_private *i915)
+{
+	GEM_TRACE("%s\n", dev_name(i915->drm.dev));
+
+	intel_wakeref_auto(&i915->runtime_pm.userfault_wakeref, 0);
+	flush_workqueue(i915->wq);
+
+	/*
+	 * We have to flush all the executing contexts to main memory so
+	 * that they can saved in the hibernation image. To ensure the last
+	 * context image is coherent, we have to switch away from it. That
+	 * leaves the i915->kernel_context still active when
+	 * we actually suspend, and its image in memory may not match the GPU
+	 * state. Fortunately, the kernel_context is disposable and we do
+	 * not rely on its state.
+	 */
+	intel_gt_suspend_prepare(to_gt(i915));
+
+	i915_gem_drain_freed_objects(i915);
+}
+
+static int lmem_restore(struct drm_i915_private *i915, u32 flags)
+{
+	struct intel_memory_region *mr;
+	int ret = 0, id;
+
+	for_each_memory_region(mr, i915, id) {
+		if (mr->type == INTEL_MEMORY_LOCAL) {
+			ret = i915_ttm_restore_region(mr, flags);
+			if (ret)
+				break;
+		}
+	}
+
+	return ret;
+}
+
+static int lmem_suspend(struct drm_i915_private *i915, u32 flags)
+{
+	struct intel_memory_region *mr;
+	int ret = 0, id;
+
+	for_each_memory_region(mr, i915, id) {
+		if (mr->type == INTEL_MEMORY_LOCAL) {
+			ret = i915_ttm_backup_region(mr, flags);
+			if (ret)
+				break;
+		}
+	}
+
+	return ret;
+}
+
+static void lmem_recover(struct drm_i915_private *i915)
+{
+	struct intel_memory_region *mr;
+	int id;
+
+	for_each_memory_region(mr, i915, id)
+		if (mr->type == INTEL_MEMORY_LOCAL)
+			i915_ttm_recover_region(mr);
+}
+
+int i915_gem_backup_suspend(struct drm_i915_private *i915)
+{
+	int ret;
+
+	/* Opportunistically try to evict unpinned objects */
+	ret = lmem_suspend(i915, I915_TTM_BACKUP_ALLOW_GPU);
+	if (ret)
+		goto out_recover;
+
+	i915_gem_suspend(i915);
+
+	/*
+	 * More objects may have become unpinned as requests were
+	 * retired. Now try to evict again. The gt may be wedged here
+	 * in which case we automatically fall back to memcpy.
+	 * We allow also backing up pinned objects that have not been
+	 * marked for early recover, and that may contain, for example,
+	 * page-tables for the migrate context.
+	 */
+	ret = lmem_suspend(i915, I915_TTM_BACKUP_ALLOW_GPU |
+			   I915_TTM_BACKUP_PINNED);
+	if (ret)
+		goto out_recover;
+
+	/*
+	 * Remaining objects are backed up using memcpy once we've stopped
+	 * using the migrate context.
+	 */
+	ret = lmem_suspend(i915, I915_TTM_BACKUP_PINNED);
+	if (ret)
+		goto out_recover;
+
+	return 0;
+
+out_recover:
+	lmem_recover(i915);
+
+	return ret;
+}
+
+void i915_gem_suspend_late(struct drm_i915_private *i915)
+{
+	struct drm_i915_gem_object *obj;
+	struct list_head *phases[] = {
+		&i915->mm.shrink_list,
+		&i915->mm.purge_list,
+		NULL
+	}, **phase;
+	unsigned long flags;
+	bool flush = false;
+
+	/*
+	 * Neither the BIOS, ourselves or any other kernel
+	 * expects the system to be in execlists mode on startup,
+	 * so we need to reset the GPU back to legacy mode. And the only
+	 * known way to disable logical contexts is through a GPU reset.
+	 *
+	 * So in order to leave the system in a known default configuration,
+	 * always reset the GPU upon unload and suspend. Afterwards we then
+	 * clean up the GEM state tracking, flushing off the requests and
+	 * leaving the system in a known idle state.
+	 *
+	 * Note that is of the upmost importance that the GPU is idle and
+	 * all stray writes are flushed *before* we dismantle the backing
+	 * storage for the pinned objects.
+	 *
+	 * However, since we are uncertain that resetting the GPU on older
+	 * machines is a good idea, we don't - just in case it leaves the
+	 * machine in an unusable condition.
+	 */
+
+	intel_gt_suspend_late(to_gt(i915));
+
+	spin_lock_irqsave(&i915->mm.obj_lock, flags);
+	for (phase = phases; *phase; phase++) {
+		list_for_each_entry(obj, *phase, mm.link) {
+			if (!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
+				flush |= (obj->read_domains & I915_GEM_DOMAIN_CPU) == 0;
+			__start_cpu_write(obj); /* presume auto-hibernate */
+		}
+	}
+	spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+	if (flush)
+		wbinvd_on_all_cpus();
+}
+
+int i915_gem_freeze(struct drm_i915_private *i915)
+{
+	/* Discard all purgeable objects, let userspace recover those as
+	 * required after resuming.
+	 */
+	i915_gem_shrink_all(i915);
+
+	return 0;
+}
+
+int i915_gem_freeze_late(struct drm_i915_private *i915)
+{
+	struct drm_i915_gem_object *obj;
+	intel_wakeref_t wakeref;
+
+	/*
+	 * Called just before we write the hibernation image.
+	 *
+	 * We need to update the domain tracking to reflect that the CPU
+	 * will be accessing all the pages to create and restore from the
+	 * hibernation, and so upon restoration those pages will be in the
+	 * CPU domain.
+	 *
+	 * To make sure the hibernation image contains the latest state,
+	 * we update that state just before writing out the image.
+	 *
+	 * To try and reduce the hibernation image, we manually shrink
+	 * the objects as well, see i915_gem_freeze()
+	 */
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
+		i915_gem_shrink(NULL, i915, -1UL, NULL, ~0);
+	i915_gem_drain_freed_objects(i915);
+
+	wbinvd_on_all_cpus();
+	list_for_each_entry(obj, &i915->mm.shrink_list, mm.link)
+		__start_cpu_write(obj);
+
+	return 0;
+}
+
+void i915_gem_resume(struct drm_i915_private *i915)
+{
+	int ret;
+
+	GEM_TRACE("%s\n", dev_name(i915->drm.dev));
+
+	ret = lmem_restore(i915, 0);
+	GEM_WARN_ON(ret);
+
+	/*
+	 * As we didn't flush the kernel context before suspend, we cannot
+	 * guarantee that the context image is complete. So let's just reset
+	 * it and start again.
+	 */
+	intel_gt_resume(to_gt(i915));
+
+	ret = lmem_restore(i915, I915_TTM_BACKUP_ALLOW_GPU);
+	GEM_WARN_ON(ret);
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pm.h b/drivers/gpu/drm/i915/gem/i915_gem_pm.h
new file mode 100644
index 000000000..bedf1e959
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_pm.h
@@ -0,0 +1,26 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_GEM_PM_H__
+#define __I915_GEM_PM_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct work_struct;
+
+void i915_gem_resume(struct drm_i915_private *i915);
+
+void i915_gem_idle_work_handler(struct work_struct *work);
+
+void i915_gem_suspend(struct drm_i915_private *i915);
+void i915_gem_suspend_late(struct drm_i915_private *i915);
+int i915_gem_backup_suspend(struct drm_i915_private *i915);
+
+int i915_gem_freeze(struct drm_i915_private *i915);
+int i915_gem_freeze_late(struct drm_i915_private *i915);
+
+#endif /* __I915_GEM_PM_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_region.c b/drivers/gpu/drm/i915/gem/i915_gem_region.c
new file mode 100644
index 000000000..a4fb577ec
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_region.c
@@ -0,0 +1,208 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <uapi/drm/i915_drm.h>
+
+#include "intel_memory_region.h"
+#include "i915_gem_region.h"
+#include "i915_drv.h"
+#include "i915_trace.h"
+
+void i915_gem_object_init_memory_region(struct drm_i915_gem_object *obj,
+					struct intel_memory_region *mem)
+{
+	obj->mm.region = mem;
+
+	mutex_lock(&mem->objects.lock);
+	list_add(&obj->mm.region_link, &mem->objects.list);
+	mutex_unlock(&mem->objects.lock);
+}
+
+void i915_gem_object_release_memory_region(struct drm_i915_gem_object *obj)
+{
+	struct intel_memory_region *mem = obj->mm.region;
+
+	mutex_lock(&mem->objects.lock);
+	list_del(&obj->mm.region_link);
+	mutex_unlock(&mem->objects.lock);
+}
+
+static struct drm_i915_gem_object *
+__i915_gem_object_create_region(struct intel_memory_region *mem,
+				resource_size_t offset,
+				resource_size_t size,
+				resource_size_t page_size,
+				unsigned int flags)
+{
+	struct drm_i915_gem_object *obj;
+	resource_size_t default_page_size;
+	int err;
+
+	/*
+	 * NB: Our use of resource_size_t for the size stems from using struct
+	 * resource for the mem->region. We might need to revisit this in the
+	 * future.
+	 */
+
+	GEM_BUG_ON(flags & ~I915_BO_ALLOC_FLAGS);
+
+	if (WARN_ON_ONCE(flags & I915_BO_ALLOC_GPU_ONLY &&
+			 (flags & I915_BO_ALLOC_CPU_CLEAR ||
+			  flags & I915_BO_ALLOC_PM_EARLY)))
+		return ERR_PTR(-EINVAL);
+
+	if (!mem)
+		return ERR_PTR(-ENODEV);
+
+	default_page_size = mem->min_page_size;
+	if (page_size)
+		default_page_size = page_size;
+
+	/* We should be able to fit a page within an sg entry */
+	GEM_BUG_ON(overflows_type(default_page_size, u32));
+	GEM_BUG_ON(!is_power_of_2_u64(default_page_size));
+	GEM_BUG_ON(default_page_size < PAGE_SIZE);
+
+	size = round_up(size, default_page_size);
+
+	if (default_page_size == size)
+		flags |= I915_BO_ALLOC_CONTIGUOUS;
+
+	GEM_BUG_ON(!size);
+	GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_MIN_ALIGNMENT));
+
+	if (i915_gem_object_size_2big(size))
+		return ERR_PTR(-E2BIG);
+
+	obj = i915_gem_object_alloc();
+	if (!obj)
+		return ERR_PTR(-ENOMEM);
+
+	/*
+	 * Anything smaller than the min_page_size can't be freely inserted into
+	 * the GTT, due to alignemnt restrictions. For such special objects,
+	 * make sure we force memcpy based suspend-resume. In the future we can
+	 * revisit this, either by allowing special mis-aligned objects in the
+	 * migration path, or by mapping all of LMEM upfront using cheap 1G
+	 * GTT entries.
+	 */
+	if (default_page_size < mem->min_page_size)
+		flags |= I915_BO_ALLOC_PM_EARLY;
+
+	err = mem->ops->init_object(mem, obj, offset, size, page_size, flags);
+	if (err)
+		goto err_object_free;
+
+	trace_i915_gem_object_create(obj);
+	return obj;
+
+err_object_free:
+	i915_gem_object_free(obj);
+	return ERR_PTR(err);
+}
+
+struct drm_i915_gem_object *
+i915_gem_object_create_region(struct intel_memory_region *mem,
+			      resource_size_t size,
+			      resource_size_t page_size,
+			      unsigned int flags)
+{
+	return __i915_gem_object_create_region(mem, I915_BO_INVALID_OFFSET,
+					       size, page_size, flags);
+}
+
+struct drm_i915_gem_object *
+i915_gem_object_create_region_at(struct intel_memory_region *mem,
+				 resource_size_t offset,
+				 resource_size_t size,
+				 unsigned int flags)
+{
+	GEM_BUG_ON(offset == I915_BO_INVALID_OFFSET);
+
+	if (GEM_WARN_ON(!IS_ALIGNED(size, mem->min_page_size)) ||
+	    GEM_WARN_ON(!IS_ALIGNED(offset, mem->min_page_size)))
+		return ERR_PTR(-EINVAL);
+
+	if (range_overflows(offset, size, resource_size(&mem->region)))
+		return ERR_PTR(-EINVAL);
+
+	if (!(flags & I915_BO_ALLOC_GPU_ONLY) &&
+	    offset + size > mem->io_size &&
+	    !i915_ggtt_has_aperture(to_gt(mem->i915)->ggtt))
+		return ERR_PTR(-ENOSPC);
+
+	return __i915_gem_object_create_region(mem, offset, size, 0,
+					       flags | I915_BO_ALLOC_CONTIGUOUS);
+}
+
+/**
+ * i915_gem_process_region - Iterate over all objects of a region using ops
+ * to process and optionally skip objects
+ * @mr: The memory region
+ * @apply: ops and private data
+ *
+ * This function can be used to iterate over the regions object list,
+ * checking whether to skip objects, and, if not, lock the objects and
+ * process them using the supplied ops. Note that this function temporarily
+ * removes objects from the region list while iterating, so that if run
+ * concurrently with itself may not iterate over all objects.
+ *
+ * Return: 0 if successful, negative error code on failure.
+ */
+int i915_gem_process_region(struct intel_memory_region *mr,
+			    struct i915_gem_apply_to_region *apply)
+{
+	const struct i915_gem_apply_to_region_ops *ops = apply->ops;
+	struct drm_i915_gem_object *obj;
+	struct list_head still_in_list;
+	int ret = 0;
+
+	/*
+	 * In the future, a non-NULL apply->ww could mean the caller is
+	 * already in a locking transaction and provides its own context.
+	 */
+	GEM_WARN_ON(apply->ww);
+
+	INIT_LIST_HEAD(&still_in_list);
+	mutex_lock(&mr->objects.lock);
+	for (;;) {
+		struct i915_gem_ww_ctx ww;
+
+		obj = list_first_entry_or_null(&mr->objects.list, typeof(*obj),
+					       mm.region_link);
+		if (!obj)
+			break;
+
+		list_move_tail(&obj->mm.region_link, &still_in_list);
+		if (!kref_get_unless_zero(&obj->base.refcount))
+			continue;
+
+		/*
+		 * Note: Someone else might be migrating the object at this
+		 * point. The object's region is not stable until we lock
+		 * the object.
+		 */
+		mutex_unlock(&mr->objects.lock);
+		apply->ww = &ww;
+		for_i915_gem_ww(&ww, ret, apply->interruptible) {
+			ret = i915_gem_object_lock(obj, apply->ww);
+			if (ret)
+				continue;
+
+			if (obj->mm.region == mr)
+				ret = ops->process_obj(apply, obj);
+			/* Implicit object unlock */
+		}
+
+		i915_gem_object_put(obj);
+		mutex_lock(&mr->objects.lock);
+		if (ret)
+			break;
+	}
+	list_splice_tail(&still_in_list, &mr->objects.list);
+	mutex_unlock(&mr->objects.lock);
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_region.h b/drivers/gpu/drm/i915/gem/i915_gem_region.h
new file mode 100644
index 000000000..2dfcc41c0
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_region.h
@@ -0,0 +1,69 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_GEM_REGION_H__
+#define __I915_GEM_REGION_H__
+
+#include <linux/types.h>
+
+struct intel_memory_region;
+struct drm_i915_gem_object;
+struct sg_table;
+
+struct i915_gem_apply_to_region;
+
+#define I915_BO_INVALID_OFFSET ((resource_size_t)-1)
+
+/**
+ * struct i915_gem_apply_to_region_ops - ops to use when iterating over all
+ * region objects.
+ */
+struct i915_gem_apply_to_region_ops {
+	/**
+	 * process_obj - Process the current object
+	 * @apply: Embed this for private data.
+	 * @obj: The current object.
+	 *
+	 * Note that if this function is part of a ww transaction, and
+	 * if returns -EDEADLK for one of the objects, it may be
+	 * rerun for that same object in the same pass.
+	 */
+	int (*process_obj)(struct i915_gem_apply_to_region *apply,
+			   struct drm_i915_gem_object *obj);
+};
+
+/**
+ * struct i915_gem_apply_to_region - Argument to the struct
+ * i915_gem_apply_to_region_ops functions.
+ * @ops: The ops for the operation.
+ * @ww: Locking context used for the transaction.
+ * @interruptible: Whether to perform object locking interruptible.
+ *
+ * This structure is intended to be embedded in a private struct if needed
+ */
+struct i915_gem_apply_to_region {
+	const struct i915_gem_apply_to_region_ops *ops;
+	struct i915_gem_ww_ctx *ww;
+	u32 interruptible:1;
+};
+
+void i915_gem_object_init_memory_region(struct drm_i915_gem_object *obj,
+					struct intel_memory_region *mem);
+void i915_gem_object_release_memory_region(struct drm_i915_gem_object *obj);
+
+struct drm_i915_gem_object *
+i915_gem_object_create_region(struct intel_memory_region *mem,
+			      resource_size_t size,
+			      resource_size_t page_size,
+			      unsigned int flags);
+struct drm_i915_gem_object *
+i915_gem_object_create_region_at(struct intel_memory_region *mem,
+				 resource_size_t offset,
+				 resource_size_t size,
+				 unsigned int flags);
+
+int i915_gem_process_region(struct intel_memory_region *mr,
+			    struct i915_gem_apply_to_region *apply);
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
new file mode 100644
index 000000000..b7eae3aee
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
@@ -0,0 +1,704 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2014-2016 Intel Corporation
+ */
+
+#include <linux/pagevec.h>
+#include <linux/shmem_fs.h>
+#include <linux/swap.h>
+
+#include <drm/drm_cache.h>
+
+#include "gem/i915_gem_region.h"
+#include "i915_drv.h"
+#include "i915_gem_object.h"
+#include "i915_gem_tiling.h"
+#include "i915_gemfs.h"
+#include "i915_scatterlist.h"
+#include "i915_trace.h"
+
+/*
+ * Move pages to appropriate lru and release the pagevec, decrementing the
+ * ref count of those pages.
+ */
+static void check_release_pagevec(struct pagevec *pvec)
+{
+	check_move_unevictable_pages(pvec);
+	__pagevec_release(pvec);
+	cond_resched();
+}
+
+void shmem_sg_free_table(struct sg_table *st, struct address_space *mapping,
+			 bool dirty, bool backup)
+{
+	struct sgt_iter sgt_iter;
+	struct pagevec pvec;
+	struct page *page;
+
+	mapping_clear_unevictable(mapping);
+
+	pagevec_init(&pvec);
+	for_each_sgt_page(page, sgt_iter, st) {
+		if (dirty)
+			set_page_dirty(page);
+
+		if (backup)
+			mark_page_accessed(page);
+
+		if (!pagevec_add(&pvec, page))
+			check_release_pagevec(&pvec);
+	}
+	if (pagevec_count(&pvec))
+		check_release_pagevec(&pvec);
+
+	sg_free_table(st);
+}
+
+int shmem_sg_alloc_table(struct drm_i915_private *i915, struct sg_table *st,
+			 size_t size, struct intel_memory_region *mr,
+			 struct address_space *mapping,
+			 unsigned int max_segment)
+{
+	const unsigned long page_count = size / PAGE_SIZE;
+	unsigned long i;
+	struct scatterlist *sg;
+	struct page *page;
+	unsigned long last_pfn = 0;	/* suppress gcc warning */
+	gfp_t noreclaim;
+	int ret;
+
+	/*
+	 * If there's no chance of allocating enough pages for the whole
+	 * object, bail early.
+	 */
+	if (size > resource_size(&mr->region))
+		return -ENOMEM;
+
+	if (sg_alloc_table(st, page_count, GFP_KERNEL | __GFP_NOWARN))
+		return -ENOMEM;
+
+	/*
+	 * Get the list of pages out of our struct file.  They'll be pinned
+	 * at this point until we release them.
+	 *
+	 * Fail silently without starting the shrinker
+	 */
+	mapping_set_unevictable(mapping);
+	noreclaim = mapping_gfp_constraint(mapping, ~__GFP_RECLAIM);
+	noreclaim |= __GFP_NORETRY | __GFP_NOWARN;
+
+	sg = st->sgl;
+	st->nents = 0;
+	for (i = 0; i < page_count; i++) {
+		const unsigned int shrink[] = {
+			I915_SHRINK_BOUND | I915_SHRINK_UNBOUND,
+			0,
+		}, *s = shrink;
+		gfp_t gfp = noreclaim;
+
+		do {
+			cond_resched();
+			page = shmem_read_mapping_page_gfp(mapping, i, gfp);
+			if (!IS_ERR(page))
+				break;
+
+			if (!*s) {
+				ret = PTR_ERR(page);
+				goto err_sg;
+			}
+
+			i915_gem_shrink(NULL, i915, 2 * page_count, NULL, *s++);
+
+			/*
+			 * We've tried hard to allocate the memory by reaping
+			 * our own buffer, now let the real VM do its job and
+			 * go down in flames if truly OOM.
+			 *
+			 * However, since graphics tend to be disposable,
+			 * defer the oom here by reporting the ENOMEM back
+			 * to userspace.
+			 */
+			if (!*s) {
+				/* reclaim and warn, but no oom */
+				gfp = mapping_gfp_mask(mapping);
+
+				/*
+				 * Our bo are always dirty and so we require
+				 * kswapd to reclaim our pages (direct reclaim
+				 * does not effectively begin pageout of our
+				 * buffers on its own). However, direct reclaim
+				 * only waits for kswapd when under allocation
+				 * congestion. So as a result __GFP_RECLAIM is
+				 * unreliable and fails to actually reclaim our
+				 * dirty pages -- unless you try over and over
+				 * again with !__GFP_NORETRY. However, we still
+				 * want to fail this allocation rather than
+				 * trigger the out-of-memory killer and for
+				 * this we want __GFP_RETRY_MAYFAIL.
+				 */
+				gfp |= __GFP_RETRY_MAYFAIL | __GFP_NOWARN;
+			}
+		} while (1);
+
+		if (!i ||
+		    sg->length >= max_segment ||
+		    page_to_pfn(page) != last_pfn + 1) {
+			if (i)
+				sg = sg_next(sg);
+
+			st->nents++;
+			sg_set_page(sg, page, PAGE_SIZE, 0);
+		} else {
+			sg->length += PAGE_SIZE;
+		}
+		last_pfn = page_to_pfn(page);
+
+		/* Check that the i965g/gm workaround works. */
+		GEM_BUG_ON(gfp & __GFP_DMA32 && last_pfn >= 0x00100000UL);
+	}
+	if (sg) /* loop terminated early; short sg table */
+		sg_mark_end(sg);
+
+	/* Trim unused sg entries to avoid wasting memory. */
+	i915_sg_trim(st);
+
+	return 0;
+err_sg:
+	sg_mark_end(sg);
+	if (sg != st->sgl) {
+		shmem_sg_free_table(st, mapping, false, false);
+	} else {
+		mapping_clear_unevictable(mapping);
+		sg_free_table(st);
+	}
+
+	/*
+	 * shmemfs first checks if there is enough memory to allocate the page
+	 * and reports ENOSPC should there be insufficient, along with the usual
+	 * ENOMEM for a genuine allocation failure.
+	 *
+	 * We use ENOSPC in our driver to mean that we have run out of aperture
+	 * space and so want to translate the error from shmemfs back to our
+	 * usual understanding of ENOMEM.
+	 */
+	if (ret == -ENOSPC)
+		ret = -ENOMEM;
+
+	return ret;
+}
+
+static int shmem_get_pages(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct intel_memory_region *mem = obj->mm.region;
+	struct address_space *mapping = obj->base.filp->f_mapping;
+	const unsigned long page_count = obj->base.size / PAGE_SIZE;
+	unsigned int max_segment = i915_sg_segment_size(i915->drm.dev);
+	struct sg_table *st;
+	struct sgt_iter sgt_iter;
+	struct page *page;
+	int ret;
+
+	/*
+	 * Assert that the object is not currently in any GPU domain. As it
+	 * wasn't in the GTT, there shouldn't be any way it could have been in
+	 * a GPU cache
+	 */
+	GEM_BUG_ON(obj->read_domains & I915_GEM_GPU_DOMAINS);
+	GEM_BUG_ON(obj->write_domain & I915_GEM_GPU_DOMAINS);
+
+rebuild_st:
+	st = kmalloc(sizeof(*st), GFP_KERNEL | __GFP_NOWARN);
+	if (!st)
+		return -ENOMEM;
+
+	ret = shmem_sg_alloc_table(i915, st, obj->base.size, mem, mapping,
+				   max_segment);
+	if (ret)
+		goto err_st;
+
+	ret = i915_gem_gtt_prepare_pages(obj, st);
+	if (ret) {
+		/*
+		 * DMA remapping failed? One possible cause is that
+		 * it could not reserve enough large entries, asking
+		 * for PAGE_SIZE chunks instead may be helpful.
+		 */
+		if (max_segment > PAGE_SIZE) {
+			for_each_sgt_page(page, sgt_iter, st)
+				put_page(page);
+			sg_free_table(st);
+			kfree(st);
+
+			max_segment = PAGE_SIZE;
+			goto rebuild_st;
+		} else {
+			dev_warn(i915->drm.dev,
+				 "Failed to DMA remap %lu pages\n",
+				 page_count);
+			goto err_pages;
+		}
+	}
+
+	if (i915_gem_object_needs_bit17_swizzle(obj))
+		i915_gem_object_do_bit_17_swizzle(obj, st);
+
+	if (i915_gem_object_can_bypass_llc(obj))
+		obj->cache_dirty = true;
+
+	__i915_gem_object_set_pages(obj, st, i915_sg_dma_sizes(st->sgl));
+
+	return 0;
+
+err_pages:
+	shmem_sg_free_table(st, mapping, false, false);
+	/*
+	 * shmemfs first checks if there is enough memory to allocate the page
+	 * and reports ENOSPC should there be insufficient, along with the usual
+	 * ENOMEM for a genuine allocation failure.
+	 *
+	 * We use ENOSPC in our driver to mean that we have run out of aperture
+	 * space and so want to translate the error from shmemfs back to our
+	 * usual understanding of ENOMEM.
+	 */
+err_st:
+	if (ret == -ENOSPC)
+		ret = -ENOMEM;
+
+	kfree(st);
+
+	return ret;
+}
+
+static int
+shmem_truncate(struct drm_i915_gem_object *obj)
+{
+	/*
+	 * Our goal here is to return as much of the memory as
+	 * is possible back to the system as we are called from OOM.
+	 * To do this we must instruct the shmfs to drop all of its
+	 * backing pages, *now*.
+	 */
+	shmem_truncate_range(file_inode(obj->base.filp), 0, (loff_t)-1);
+	obj->mm.madv = __I915_MADV_PURGED;
+	obj->mm.pages = ERR_PTR(-EFAULT);
+
+	return 0;
+}
+
+void __shmem_writeback(size_t size, struct address_space *mapping)
+{
+	struct writeback_control wbc = {
+		.sync_mode = WB_SYNC_NONE,
+		.nr_to_write = SWAP_CLUSTER_MAX,
+		.range_start = 0,
+		.range_end = LLONG_MAX,
+		.for_reclaim = 1,
+	};
+	unsigned long i;
+
+	/*
+	 * Leave mmapings intact (GTT will have been revoked on unbinding,
+	 * leaving only CPU mmapings around) and add those pages to the LRU
+	 * instead of invoking writeback so they are aged and paged out
+	 * as normal.
+	 */
+
+	/* Begin writeback on each dirty page */
+	for (i = 0; i < size >> PAGE_SHIFT; i++) {
+		struct page *page;
+
+		page = find_lock_page(mapping, i);
+		if (!page)
+			continue;
+
+		if (!page_mapped(page) && clear_page_dirty_for_io(page)) {
+			int ret;
+
+			SetPageReclaim(page);
+			ret = mapping->a_ops->writepage(page, &wbc);
+			if (!PageWriteback(page))
+				ClearPageReclaim(page);
+			if (!ret)
+				goto put;
+		}
+		unlock_page(page);
+put:
+		put_page(page);
+	}
+}
+
+static void
+shmem_writeback(struct drm_i915_gem_object *obj)
+{
+	__shmem_writeback(obj->base.size, obj->base.filp->f_mapping);
+}
+
+static int shmem_shrink(struct drm_i915_gem_object *obj, unsigned int flags)
+{
+	switch (obj->mm.madv) {
+	case I915_MADV_DONTNEED:
+		return i915_gem_object_truncate(obj);
+	case __I915_MADV_PURGED:
+		return 0;
+	}
+
+	if (flags & I915_GEM_OBJECT_SHRINK_WRITEBACK)
+		shmem_writeback(obj);
+
+	return 0;
+}
+
+void
+__i915_gem_object_release_shmem(struct drm_i915_gem_object *obj,
+				struct sg_table *pages,
+				bool needs_clflush)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+
+	GEM_BUG_ON(obj->mm.madv == __I915_MADV_PURGED);
+
+	if (obj->mm.madv == I915_MADV_DONTNEED)
+		obj->mm.dirty = false;
+
+	if (needs_clflush &&
+	    (obj->read_domains & I915_GEM_DOMAIN_CPU) == 0 &&
+	    !(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
+		drm_clflush_sg(pages);
+
+	__start_cpu_write(obj);
+	/*
+	 * On non-LLC igfx platforms, force the flush-on-acquire if this is ever
+	 * swapped-in. Our async flush path is not trust worthy enough yet(and
+	 * happens in the wrong order), and with some tricks it's conceivable
+	 * for userspace to change the cache-level to I915_CACHE_NONE after the
+	 * pages are swapped-in, and since execbuf binds the object before doing
+	 * the async flush, we have a race window.
+	 */
+	if (!HAS_LLC(i915) && !IS_DGFX(i915))
+		obj->cache_dirty = true;
+}
+
+void i915_gem_object_put_pages_shmem(struct drm_i915_gem_object *obj, struct sg_table *pages)
+{
+	__i915_gem_object_release_shmem(obj, pages, true);
+
+	i915_gem_gtt_finish_pages(obj, pages);
+
+	if (i915_gem_object_needs_bit17_swizzle(obj))
+		i915_gem_object_save_bit_17_swizzle(obj, pages);
+
+	shmem_sg_free_table(pages, file_inode(obj->base.filp)->i_mapping,
+			    obj->mm.dirty, obj->mm.madv == I915_MADV_WILLNEED);
+	kfree(pages);
+	obj->mm.dirty = false;
+}
+
+static void
+shmem_put_pages(struct drm_i915_gem_object *obj, struct sg_table *pages)
+{
+	if (likely(i915_gem_object_has_struct_page(obj)))
+		i915_gem_object_put_pages_shmem(obj, pages);
+	else
+		i915_gem_object_put_pages_phys(obj, pages);
+}
+
+static int
+shmem_pwrite(struct drm_i915_gem_object *obj,
+	     const struct drm_i915_gem_pwrite *arg)
+{
+	struct address_space *mapping = obj->base.filp->f_mapping;
+	const struct address_space_operations *aops = mapping->a_ops;
+	char __user *user_data = u64_to_user_ptr(arg->data_ptr);
+	u64 remain, offset;
+	unsigned int pg;
+
+	/* Caller already validated user args */
+	GEM_BUG_ON(!access_ok(user_data, arg->size));
+
+	if (!i915_gem_object_has_struct_page(obj))
+		return i915_gem_object_pwrite_phys(obj, arg);
+
+	/*
+	 * Before we instantiate/pin the backing store for our use, we
+	 * can prepopulate the shmemfs filp efficiently using a write into
+	 * the pagecache. We avoid the penalty of instantiating all the
+	 * pages, important if the user is just writing to a few and never
+	 * uses the object on the GPU, and using a direct write into shmemfs
+	 * allows it to avoid the cost of retrieving a page (either swapin
+	 * or clearing-before-use) before it is overwritten.
+	 */
+	if (i915_gem_object_has_pages(obj))
+		return -ENODEV;
+
+	if (obj->mm.madv != I915_MADV_WILLNEED)
+		return -EFAULT;
+
+	/*
+	 * Before the pages are instantiated the object is treated as being
+	 * in the CPU domain. The pages will be clflushed as required before
+	 * use, and we can freely write into the pages directly. If userspace
+	 * races pwrite with any other operation; corruption will ensue -
+	 * that is userspace's prerogative!
+	 */
+
+	remain = arg->size;
+	offset = arg->offset;
+	pg = offset_in_page(offset);
+
+	do {
+		unsigned int len, unwritten;
+		struct page *page;
+		void *data, *vaddr;
+		int err;
+		char c;
+
+		len = PAGE_SIZE - pg;
+		if (len > remain)
+			len = remain;
+
+		/* Prefault the user page to reduce potential recursion */
+		err = __get_user(c, user_data);
+		if (err)
+			return err;
+
+		err = __get_user(c, user_data + len - 1);
+		if (err)
+			return err;
+
+		err = aops->write_begin(obj->base.filp, mapping, offset, len,
+					&page, &data);
+		if (err < 0)
+			return err;
+
+		vaddr = kmap_atomic(page);
+		unwritten = __copy_from_user_inatomic(vaddr + pg,
+						      user_data,
+						      len);
+		kunmap_atomic(vaddr);
+
+		err = aops->write_end(obj->base.filp, mapping, offset, len,
+				      len - unwritten, page, data);
+		if (err < 0)
+			return err;
+
+		/* We don't handle -EFAULT, leave it to the caller to check */
+		if (unwritten)
+			return -ENODEV;
+
+		remain -= len;
+		user_data += len;
+		offset += len;
+		pg = 0;
+	} while (remain);
+
+	return 0;
+}
+
+static int
+shmem_pread(struct drm_i915_gem_object *obj,
+	    const struct drm_i915_gem_pread *arg)
+{
+	if (!i915_gem_object_has_struct_page(obj))
+		return i915_gem_object_pread_phys(obj, arg);
+
+	return -ENODEV;
+}
+
+static void shmem_release(struct drm_i915_gem_object *obj)
+{
+	if (i915_gem_object_has_struct_page(obj))
+		i915_gem_object_release_memory_region(obj);
+
+	fput(obj->base.filp);
+}
+
+const struct drm_i915_gem_object_ops i915_gem_shmem_ops = {
+	.name = "i915_gem_object_shmem",
+	.flags = I915_GEM_OBJECT_IS_SHRINKABLE,
+
+	.get_pages = shmem_get_pages,
+	.put_pages = shmem_put_pages,
+	.truncate = shmem_truncate,
+	.shrink = shmem_shrink,
+
+	.pwrite = shmem_pwrite,
+	.pread = shmem_pread,
+
+	.release = shmem_release,
+};
+
+static int __create_shmem(struct drm_i915_private *i915,
+			  struct drm_gem_object *obj,
+			  resource_size_t size)
+{
+	unsigned long flags = VM_NORESERVE;
+	struct file *filp;
+
+	drm_gem_private_object_init(&i915->drm, obj, size);
+
+	if (i915->mm.gemfs)
+		filp = shmem_file_setup_with_mnt(i915->mm.gemfs, "i915", size,
+						 flags);
+	else
+		filp = shmem_file_setup("i915", size, flags);
+	if (IS_ERR(filp))
+		return PTR_ERR(filp);
+
+	obj->filp = filp;
+	return 0;
+}
+
+static int shmem_object_init(struct intel_memory_region *mem,
+			     struct drm_i915_gem_object *obj,
+			     resource_size_t offset,
+			     resource_size_t size,
+			     resource_size_t page_size,
+			     unsigned int flags)
+{
+	static struct lock_class_key lock_class;
+	struct drm_i915_private *i915 = mem->i915;
+	struct address_space *mapping;
+	unsigned int cache_level;
+	gfp_t mask;
+	int ret;
+
+	ret = __create_shmem(i915, &obj->base, size);
+	if (ret)
+		return ret;
+
+	mask = GFP_HIGHUSER | __GFP_RECLAIMABLE;
+	if (IS_I965GM(i915) || IS_I965G(i915)) {
+		/* 965gm cannot relocate objects above 4GiB. */
+		mask &= ~__GFP_HIGHMEM;
+		mask |= __GFP_DMA32;
+	}
+
+	mapping = obj->base.filp->f_mapping;
+	mapping_set_gfp_mask(mapping, mask);
+	GEM_BUG_ON(!(mapping_gfp_mask(mapping) & __GFP_RECLAIM));
+
+	i915_gem_object_init(obj, &i915_gem_shmem_ops, &lock_class, flags);
+	obj->mem_flags |= I915_BO_FLAG_STRUCT_PAGE;
+	obj->write_domain = I915_GEM_DOMAIN_CPU;
+	obj->read_domains = I915_GEM_DOMAIN_CPU;
+
+	if (HAS_LLC(i915))
+		/* On some devices, we can have the GPU use the LLC (the CPU
+		 * cache) for about a 10% performance improvement
+		 * compared to uncached.  Graphics requests other than
+		 * display scanout are coherent with the CPU in
+		 * accessing this cache.  This means in this mode we
+		 * don't need to clflush on the CPU side, and on the
+		 * GPU side we only need to flush internal caches to
+		 * get data visible to the CPU.
+		 *
+		 * However, we maintain the display planes as UC, and so
+		 * need to rebind when first used as such.
+		 */
+		cache_level = I915_CACHE_LLC;
+	else
+		cache_level = I915_CACHE_NONE;
+
+	i915_gem_object_set_cache_coherency(obj, cache_level);
+
+	i915_gem_object_init_memory_region(obj, mem);
+
+	return 0;
+}
+
+struct drm_i915_gem_object *
+i915_gem_object_create_shmem(struct drm_i915_private *i915,
+			     resource_size_t size)
+{
+	return i915_gem_object_create_region(i915->mm.regions[INTEL_REGION_SMEM],
+					     size, 0, 0);
+}
+
+/* Allocate a new GEM object and fill it with the supplied data */
+struct drm_i915_gem_object *
+i915_gem_object_create_shmem_from_data(struct drm_i915_private *dev_priv,
+				       const void *data, resource_size_t size)
+{
+	struct drm_i915_gem_object *obj;
+	struct file *file;
+	const struct address_space_operations *aops;
+	resource_size_t offset;
+	int err;
+
+	GEM_WARN_ON(IS_DGFX(dev_priv));
+	obj = i915_gem_object_create_shmem(dev_priv, round_up(size, PAGE_SIZE));
+	if (IS_ERR(obj))
+		return obj;
+
+	GEM_BUG_ON(obj->write_domain != I915_GEM_DOMAIN_CPU);
+
+	file = obj->base.filp;
+	aops = file->f_mapping->a_ops;
+	offset = 0;
+	do {
+		unsigned int len = min_t(typeof(size), size, PAGE_SIZE);
+		struct page *page;
+		void *pgdata, *vaddr;
+
+		err = aops->write_begin(file, file->f_mapping, offset, len,
+					&page, &pgdata);
+		if (err < 0)
+			goto fail;
+
+		vaddr = kmap(page);
+		memcpy(vaddr, data, len);
+		kunmap(page);
+
+		err = aops->write_end(file, file->f_mapping, offset, len, len,
+				      page, pgdata);
+		if (err < 0)
+			goto fail;
+
+		size -= len;
+		data += len;
+		offset += len;
+	} while (size);
+
+	return obj;
+
+fail:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static int init_shmem(struct intel_memory_region *mem)
+{
+	i915_gemfs_init(mem->i915);
+	intel_memory_region_set_name(mem, "system");
+
+	return 0; /* We have fallback to the kernel mnt if gemfs init failed. */
+}
+
+static int release_shmem(struct intel_memory_region *mem)
+{
+	i915_gemfs_fini(mem->i915);
+	return 0;
+}
+
+static const struct intel_memory_region_ops shmem_region_ops = {
+	.init = init_shmem,
+	.release = release_shmem,
+	.init_object = shmem_object_init,
+};
+
+struct intel_memory_region *i915_gem_shmem_setup(struct drm_i915_private *i915,
+						 u16 type, u16 instance)
+{
+	return intel_memory_region_create(i915, 0,
+					  totalram_pages() << PAGE_SHIFT,
+					  PAGE_SIZE, 0, 0,
+					  type, instance,
+					  &shmem_region_ops);
+}
+
+bool i915_gem_object_is_shmem(const struct drm_i915_gem_object *obj)
+{
+	return obj->ops == &i915_gem_shmem_ops;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c
new file mode 100644
index 000000000..8dc5c8874
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c
@@ -0,0 +1,587 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2008-2015 Intel Corporation
+ */
+
+#include <linux/oom.h>
+#include <linux/sched/mm.h>
+#include <linux/shmem_fs.h>
+#include <linux/slab.h>
+#include <linux/swap.h>
+#include <linux/pci.h>
+#include <linux/dma-buf.h>
+#include <linux/vmalloc.h>
+
+#include "gt/intel_gt_requests.h"
+
+#include "i915_trace.h"
+
+static bool swap_available(void)
+{
+	return get_nr_swap_pages() > 0;
+}
+
+static bool can_release_pages(struct drm_i915_gem_object *obj)
+{
+	/* Consider only shrinkable ojects. */
+	if (!i915_gem_object_is_shrinkable(obj))
+		return false;
+
+	/*
+	 * We can only return physical pages to the system if we can either
+	 * discard the contents (because the user has marked them as being
+	 * purgeable) or if we can move their contents out to swap.
+	 */
+	return swap_available() || obj->mm.madv == I915_MADV_DONTNEED;
+}
+
+static bool drop_pages(struct drm_i915_gem_object *obj,
+		       unsigned long shrink, bool trylock_vm)
+{
+	unsigned long flags;
+
+	flags = 0;
+	if (shrink & I915_SHRINK_ACTIVE)
+		flags |= I915_GEM_OBJECT_UNBIND_ACTIVE;
+	if (!(shrink & I915_SHRINK_BOUND))
+		flags |= I915_GEM_OBJECT_UNBIND_TEST;
+	if (trylock_vm)
+		flags |= I915_GEM_OBJECT_UNBIND_VM_TRYLOCK;
+
+	if (i915_gem_object_unbind(obj, flags) == 0)
+		return true;
+
+	return false;
+}
+
+static int try_to_writeback(struct drm_i915_gem_object *obj, unsigned int flags)
+{
+	if (obj->ops->shrink) {
+		unsigned int shrink_flags = 0;
+
+		if (!(flags & I915_SHRINK_ACTIVE))
+			shrink_flags |= I915_GEM_OBJECT_SHRINK_NO_GPU_WAIT;
+
+		if (flags & I915_SHRINK_WRITEBACK)
+			shrink_flags |= I915_GEM_OBJECT_SHRINK_WRITEBACK;
+
+		return obj->ops->shrink(obj, shrink_flags);
+	}
+
+	return 0;
+}
+
+/**
+ * i915_gem_shrink - Shrink buffer object caches
+ * @ww: i915 gem ww acquire ctx, or NULL
+ * @i915: i915 device
+ * @target: amount of memory to make available, in pages
+ * @nr_scanned: optional output for number of pages scanned (incremental)
+ * @shrink: control flags for selecting cache types
+ *
+ * This function is the main interface to the shrinker. It will try to release
+ * up to @target pages of main memory backing storage from buffer objects.
+ * Selection of the specific caches can be done with @flags. This is e.g. useful
+ * when purgeable objects should be removed from caches preferentially.
+ *
+ * Note that it's not guaranteed that released amount is actually available as
+ * free system memory - the pages might still be in-used to due to other reasons
+ * (like cpu mmaps) or the mm core has reused them before we could grab them.
+ * Therefore code that needs to explicitly shrink buffer objects caches (e.g. to
+ * avoid deadlocks in memory reclaim) must fall back to i915_gem_shrink_all().
+ *
+ * Also note that any kind of pinning (both per-vma address space pins and
+ * backing storage pins at the buffer object level) result in the shrinker code
+ * having to skip the object.
+ *
+ * Returns:
+ * The number of pages of backing storage actually released.
+ */
+unsigned long
+i915_gem_shrink(struct i915_gem_ww_ctx *ww,
+		struct drm_i915_private *i915,
+		unsigned long target,
+		unsigned long *nr_scanned,
+		unsigned int shrink)
+{
+	const struct {
+		struct list_head *list;
+		unsigned int bit;
+	} phases[] = {
+		{ &i915->mm.purge_list, ~0u },
+		{
+			&i915->mm.shrink_list,
+			I915_SHRINK_BOUND | I915_SHRINK_UNBOUND
+		},
+		{ NULL, 0 },
+	}, *phase;
+	intel_wakeref_t wakeref = 0;
+	unsigned long count = 0;
+	unsigned long scanned = 0;
+	int err = 0;
+
+	/* CHV + VTD workaround use stop_machine(); need to trylock vm->mutex */
+	bool trylock_vm = !ww && intel_vm_no_concurrent_access_wa(i915);
+
+	trace_i915_gem_shrink(i915, target, shrink);
+
+	/*
+	 * Unbinding of objects will require HW access; Let us not wake the
+	 * device just to recover a little memory. If absolutely necessary,
+	 * we will force the wake during oom-notifier.
+	 */
+	if (shrink & I915_SHRINK_BOUND) {
+		wakeref = intel_runtime_pm_get_if_in_use(&i915->runtime_pm);
+		if (!wakeref)
+			shrink &= ~I915_SHRINK_BOUND;
+	}
+
+	/*
+	 * When shrinking the active list, we should also consider active
+	 * contexts. Active contexts are pinned until they are retired, and
+	 * so can not be simply unbound to retire and unpin their pages. To
+	 * shrink the contexts, we must wait until the gpu is idle and
+	 * completed its switch to the kernel context. In short, we do
+	 * not have a good mechanism for idling a specific context, but
+	 * what we can do is give them a kick so that we do not keep idle
+	 * contexts around longer than is necessary.
+	 */
+	if (shrink & I915_SHRINK_ACTIVE)
+		/* Retire requests to unpin all idle contexts */
+		intel_gt_retire_requests(to_gt(i915));
+
+	/*
+	 * As we may completely rewrite the (un)bound list whilst unbinding
+	 * (due to retiring requests) we have to strictly process only
+	 * one element of the list at the time, and recheck the list
+	 * on every iteration.
+	 *
+	 * In particular, we must hold a reference whilst removing the
+	 * object as we may end up waiting for and/or retiring the objects.
+	 * This might release the final reference (held by the active list)
+	 * and result in the object being freed from under us. This is
+	 * similar to the precautions the eviction code must take whilst
+	 * removing objects.
+	 *
+	 * Also note that although these lists do not hold a reference to
+	 * the object we can safely grab one here: The final object
+	 * unreferencing and the bound_list are both protected by the
+	 * dev->struct_mutex and so we won't ever be able to observe an
+	 * object on the bound_list with a reference count equals 0.
+	 */
+	for (phase = phases; phase->list; phase++) {
+		struct list_head still_in_list;
+		struct drm_i915_gem_object *obj;
+		unsigned long flags;
+
+		if ((shrink & phase->bit) == 0)
+			continue;
+
+		INIT_LIST_HEAD(&still_in_list);
+
+		/*
+		 * We serialize our access to unreferenced objects through
+		 * the use of the struct_mutex. While the objects are not
+		 * yet freed (due to RCU then a workqueue) we still want
+		 * to be able to shrink their pages, so they remain on
+		 * the unbound/bound list until actually freed.
+		 */
+		spin_lock_irqsave(&i915->mm.obj_lock, flags);
+		while (count < target &&
+		       (obj = list_first_entry_or_null(phase->list,
+						       typeof(*obj),
+						       mm.link))) {
+			list_move_tail(&obj->mm.link, &still_in_list);
+
+			if (shrink & I915_SHRINK_VMAPS &&
+			    !is_vmalloc_addr(obj->mm.mapping))
+				continue;
+
+			if (!(shrink & I915_SHRINK_ACTIVE) &&
+			    i915_gem_object_is_framebuffer(obj))
+				continue;
+
+			if (!can_release_pages(obj))
+				continue;
+
+			if (!kref_get_unless_zero(&obj->base.refcount))
+				continue;
+
+			spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+
+			/* May arrive from get_pages on another bo */
+			if (!ww) {
+				if (!i915_gem_object_trylock(obj, NULL))
+					goto skip;
+			} else {
+				err = i915_gem_object_lock(obj, ww);
+				if (err)
+					goto skip;
+			}
+
+			if (drop_pages(obj, shrink, trylock_vm) &&
+			    !__i915_gem_object_put_pages(obj) &&
+			    !try_to_writeback(obj, shrink))
+				count += obj->base.size >> PAGE_SHIFT;
+
+			if (!ww)
+				i915_gem_object_unlock(obj);
+
+			scanned += obj->base.size >> PAGE_SHIFT;
+skip:
+			i915_gem_object_put(obj);
+
+			spin_lock_irqsave(&i915->mm.obj_lock, flags);
+			if (err)
+				break;
+		}
+		list_splice_tail(&still_in_list, phase->list);
+		spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+		if (err)
+			break;
+	}
+
+	if (shrink & I915_SHRINK_BOUND)
+		intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+
+	if (err)
+		return err;
+
+	if (nr_scanned)
+		*nr_scanned += scanned;
+	return count;
+}
+
+/**
+ * i915_gem_shrink_all - Shrink buffer object caches completely
+ * @i915: i915 device
+ *
+ * This is a simple wraper around i915_gem_shrink() to aggressively shrink all
+ * caches completely. It also first waits for and retires all outstanding
+ * requests to also be able to release backing storage for active objects.
+ *
+ * This should only be used in code to intentionally quiescent the gpu or as a
+ * last-ditch effort when memory seems to have run out.
+ *
+ * Returns:
+ * The number of pages of backing storage actually released.
+ */
+unsigned long i915_gem_shrink_all(struct drm_i915_private *i915)
+{
+	intel_wakeref_t wakeref;
+	unsigned long freed = 0;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		freed = i915_gem_shrink(NULL, i915, -1UL, NULL,
+					I915_SHRINK_BOUND |
+					I915_SHRINK_UNBOUND);
+	}
+
+	return freed;
+}
+
+static unsigned long
+i915_gem_shrinker_count(struct shrinker *shrinker, struct shrink_control *sc)
+{
+	struct drm_i915_private *i915 =
+		container_of(shrinker, struct drm_i915_private, mm.shrinker);
+	unsigned long num_objects;
+	unsigned long count;
+
+	count = READ_ONCE(i915->mm.shrink_memory) >> PAGE_SHIFT;
+	num_objects = READ_ONCE(i915->mm.shrink_count);
+
+	/*
+	 * Update our preferred vmscan batch size for the next pass.
+	 * Our rough guess for an effective batch size is roughly 2
+	 * available GEM objects worth of pages. That is we don't want
+	 * the shrinker to fire, until it is worth the cost of freeing an
+	 * entire GEM object.
+	 */
+	if (num_objects) {
+		unsigned long avg = 2 * count / num_objects;
+
+		i915->mm.shrinker.batch =
+			max((i915->mm.shrinker.batch + avg) >> 1,
+			    128ul /* default SHRINK_BATCH */);
+	}
+
+	return count;
+}
+
+static unsigned long
+i915_gem_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc)
+{
+	struct drm_i915_private *i915 =
+		container_of(shrinker, struct drm_i915_private, mm.shrinker);
+	unsigned long freed;
+
+	sc->nr_scanned = 0;
+
+	freed = i915_gem_shrink(NULL, i915,
+				sc->nr_to_scan,
+				&sc->nr_scanned,
+				I915_SHRINK_BOUND |
+				I915_SHRINK_UNBOUND);
+	if (sc->nr_scanned < sc->nr_to_scan && current_is_kswapd()) {
+		intel_wakeref_t wakeref;
+
+		with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+			freed += i915_gem_shrink(NULL, i915,
+						 sc->nr_to_scan - sc->nr_scanned,
+						 &sc->nr_scanned,
+						 I915_SHRINK_ACTIVE |
+						 I915_SHRINK_BOUND |
+						 I915_SHRINK_UNBOUND |
+						 I915_SHRINK_WRITEBACK);
+		}
+	}
+
+	return sc->nr_scanned ? freed : SHRINK_STOP;
+}
+
+static int
+i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
+{
+	struct drm_i915_private *i915 =
+		container_of(nb, struct drm_i915_private, mm.oom_notifier);
+	struct drm_i915_gem_object *obj;
+	unsigned long unevictable, available, freed_pages;
+	intel_wakeref_t wakeref;
+	unsigned long flags;
+
+	freed_pages = 0;
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
+		freed_pages += i915_gem_shrink(NULL, i915, -1UL, NULL,
+					       I915_SHRINK_BOUND |
+					       I915_SHRINK_UNBOUND |
+					       I915_SHRINK_WRITEBACK);
+
+	/* Because we may be allocating inside our own driver, we cannot
+	 * assert that there are no objects with pinned pages that are not
+	 * being pointed to by hardware.
+	 */
+	available = unevictable = 0;
+	spin_lock_irqsave(&i915->mm.obj_lock, flags);
+	list_for_each_entry(obj, &i915->mm.shrink_list, mm.link) {
+		if (!can_release_pages(obj))
+			unevictable += obj->base.size >> PAGE_SHIFT;
+		else
+			available += obj->base.size >> PAGE_SHIFT;
+	}
+	spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+
+	if (freed_pages || available)
+		pr_info("Purging GPU memory, %lu pages freed, "
+			"%lu pages still pinned, %lu pages left available.\n",
+			freed_pages, unevictable, available);
+
+	*(unsigned long *)ptr += freed_pages;
+	return NOTIFY_DONE;
+}
+
+static int
+i915_gem_shrinker_vmap(struct notifier_block *nb, unsigned long event, void *ptr)
+{
+	struct drm_i915_private *i915 =
+		container_of(nb, struct drm_i915_private, mm.vmap_notifier);
+	struct i915_vma *vma, *next;
+	unsigned long freed_pages = 0;
+	intel_wakeref_t wakeref;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
+		freed_pages += i915_gem_shrink(NULL, i915, -1UL, NULL,
+					       I915_SHRINK_BOUND |
+					       I915_SHRINK_UNBOUND |
+					       I915_SHRINK_VMAPS);
+
+	/* We also want to clear any cached iomaps as they wrap vmap */
+	mutex_lock(&to_gt(i915)->ggtt->vm.mutex);
+	list_for_each_entry_safe(vma, next,
+				 &to_gt(i915)->ggtt->vm.bound_list, vm_link) {
+		unsigned long count = vma->node.size >> PAGE_SHIFT;
+		struct drm_i915_gem_object *obj = vma->obj;
+
+		if (!vma->iomap || i915_vma_is_active(vma))
+			continue;
+
+		if (!i915_gem_object_trylock(obj, NULL))
+			continue;
+
+		if (__i915_vma_unbind(vma) == 0)
+			freed_pages += count;
+
+		i915_gem_object_unlock(obj);
+	}
+	mutex_unlock(&to_gt(i915)->ggtt->vm.mutex);
+
+	*(unsigned long *)ptr += freed_pages;
+	return NOTIFY_DONE;
+}
+
+void i915_gem_driver_register__shrinker(struct drm_i915_private *i915)
+{
+	i915->mm.shrinker.scan_objects = i915_gem_shrinker_scan;
+	i915->mm.shrinker.count_objects = i915_gem_shrinker_count;
+	i915->mm.shrinker.seeks = DEFAULT_SEEKS;
+	i915->mm.shrinker.batch = 4096;
+	drm_WARN_ON(&i915->drm, register_shrinker(&i915->mm.shrinker,
+						  "drm-i915_gem"));
+
+	i915->mm.oom_notifier.notifier_call = i915_gem_shrinker_oom;
+	drm_WARN_ON(&i915->drm, register_oom_notifier(&i915->mm.oom_notifier));
+
+	i915->mm.vmap_notifier.notifier_call = i915_gem_shrinker_vmap;
+	drm_WARN_ON(&i915->drm,
+		    register_vmap_purge_notifier(&i915->mm.vmap_notifier));
+}
+
+void i915_gem_driver_unregister__shrinker(struct drm_i915_private *i915)
+{
+	drm_WARN_ON(&i915->drm,
+		    unregister_vmap_purge_notifier(&i915->mm.vmap_notifier));
+	drm_WARN_ON(&i915->drm,
+		    unregister_oom_notifier(&i915->mm.oom_notifier));
+	unregister_shrinker(&i915->mm.shrinker);
+}
+
+void i915_gem_shrinker_taints_mutex(struct drm_i915_private *i915,
+				    struct mutex *mutex)
+{
+	if (!IS_ENABLED(CONFIG_LOCKDEP))
+		return;
+
+	fs_reclaim_acquire(GFP_KERNEL);
+
+	mutex_acquire(&mutex->dep_map, 0, 0, _RET_IP_);
+	mutex_release(&mutex->dep_map, _RET_IP_);
+
+	fs_reclaim_release(GFP_KERNEL);
+}
+
+#define obj_to_i915(obj__) to_i915((obj__)->base.dev)
+
+/**
+ * i915_gem_object_make_unshrinkable - Hide the object from the shrinker. By
+ * default all object types that support shrinking(see IS_SHRINKABLE), will also
+ * make the object visible to the shrinker after allocating the system memory
+ * pages.
+ * @obj: The GEM object.
+ *
+ * This is typically used for special kernel internal objects that can't be
+ * easily processed by the shrinker, like if they are perma-pinned.
+ */
+void i915_gem_object_make_unshrinkable(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = obj_to_i915(obj);
+	unsigned long flags;
+
+	/*
+	 * We can only be called while the pages are pinned or when
+	 * the pages are released. If pinned, we should only be called
+	 * from a single caller under controlled conditions; and on release
+	 * only one caller may release us. Neither the two may cross.
+	 */
+	if (atomic_add_unless(&obj->mm.shrink_pin, 1, 0))
+		return;
+
+	spin_lock_irqsave(&i915->mm.obj_lock, flags);
+	if (!atomic_fetch_inc(&obj->mm.shrink_pin) &&
+	    !list_empty(&obj->mm.link)) {
+		list_del_init(&obj->mm.link);
+		i915->mm.shrink_count--;
+		i915->mm.shrink_memory -= obj->base.size;
+	}
+	spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+}
+
+static void ___i915_gem_object_make_shrinkable(struct drm_i915_gem_object *obj,
+					       struct list_head *head)
+{
+	struct drm_i915_private *i915 = obj_to_i915(obj);
+	unsigned long flags;
+
+	if (!i915_gem_object_is_shrinkable(obj))
+		return;
+
+	if (atomic_add_unless(&obj->mm.shrink_pin, -1, 1))
+		return;
+
+	spin_lock_irqsave(&i915->mm.obj_lock, flags);
+	GEM_BUG_ON(!kref_read(&obj->base.refcount));
+	if (atomic_dec_and_test(&obj->mm.shrink_pin)) {
+		GEM_BUG_ON(!list_empty(&obj->mm.link));
+
+		list_add_tail(&obj->mm.link, head);
+		i915->mm.shrink_count++;
+		i915->mm.shrink_memory += obj->base.size;
+
+	}
+	spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+}
+
+/**
+ * __i915_gem_object_make_shrinkable - Move the object to the tail of the
+ * shrinkable list. Objects on this list might be swapped out. Used with
+ * WILLNEED objects.
+ * @obj: The GEM object.
+ *
+ * DO NOT USE. This is intended to be called on very special objects that don't
+ * yet have mm.pages, but are guaranteed to have potentially reclaimable pages
+ * underneath.
+ */
+void __i915_gem_object_make_shrinkable(struct drm_i915_gem_object *obj)
+{
+	___i915_gem_object_make_shrinkable(obj,
+					   &obj_to_i915(obj)->mm.shrink_list);
+}
+
+/**
+ * __i915_gem_object_make_purgeable - Move the object to the tail of the
+ * purgeable list. Objects on this list might be swapped out. Used with
+ * DONTNEED objects.
+ * @obj: The GEM object.
+ *
+ * DO NOT USE. This is intended to be called on very special objects that don't
+ * yet have mm.pages, but are guaranteed to have potentially reclaimable pages
+ * underneath.
+ */
+void __i915_gem_object_make_purgeable(struct drm_i915_gem_object *obj)
+{
+	___i915_gem_object_make_shrinkable(obj,
+					   &obj_to_i915(obj)->mm.purge_list);
+}
+
+/**
+ * i915_gem_object_make_shrinkable - Move the object to the tail of the
+ * shrinkable list. Objects on this list might be swapped out. Used with
+ * WILLNEED objects.
+ * @obj: The GEM object.
+ *
+ * MUST only be called on objects which have backing pages.
+ *
+ * MUST be balanced with previous call to i915_gem_object_make_unshrinkable().
+ */
+void i915_gem_object_make_shrinkable(struct drm_i915_gem_object *obj)
+{
+	GEM_BUG_ON(!i915_gem_object_has_pages(obj));
+	__i915_gem_object_make_shrinkable(obj);
+}
+
+/**
+ * i915_gem_object_make_purgeable - Move the object to the tail of the purgeable
+ * list. Used with DONTNEED objects. Unlike with shrinkable objects, the
+ * shrinker will attempt to discard the backing pages, instead of trying to swap
+ * them out.
+ * @obj: The GEM object.
+ *
+ * MUST only be called on objects which have backing pages.
+ *
+ * MUST be balanced with previous call to i915_gem_object_make_unshrinkable().
+ */
+void i915_gem_object_make_purgeable(struct drm_i915_gem_object *obj)
+{
+	GEM_BUG_ON(!i915_gem_object_has_pages(obj));
+	__i915_gem_object_make_purgeable(obj);
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shrinker.h b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.h
new file mode 100644
index 000000000..8512470f6
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.h
@@ -0,0 +1,33 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_GEM_SHRINKER_H__
+#define __I915_GEM_SHRINKER_H__
+
+#include <linux/bits.h>
+
+struct drm_i915_private;
+struct i915_gem_ww_ctx;
+struct mutex;
+
+/* i915_gem_shrinker.c */
+unsigned long i915_gem_shrink(struct i915_gem_ww_ctx *ww,
+			      struct drm_i915_private *i915,
+			      unsigned long target,
+			      unsigned long *nr_scanned,
+			      unsigned flags);
+#define I915_SHRINK_UNBOUND	BIT(0)
+#define I915_SHRINK_BOUND	BIT(1)
+#define I915_SHRINK_ACTIVE	BIT(2)
+#define I915_SHRINK_VMAPS	BIT(3)
+#define I915_SHRINK_WRITEBACK	BIT(4)
+
+unsigned long i915_gem_shrink_all(struct drm_i915_private *i915);
+void i915_gem_driver_register__shrinker(struct drm_i915_private *i915);
+void i915_gem_driver_unregister__shrinker(struct drm_i915_private *i915);
+void i915_gem_shrinker_taints_mutex(struct drm_i915_private *i915,
+				    struct mutex *mutex);
+
+#endif /* __I915_GEM_SHRINKER_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_stolen.c b/drivers/gpu/drm/i915/gem/i915_gem_stolen.c
new file mode 100644
index 000000000..acc561c0f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_stolen.c
@@ -0,0 +1,890 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2008-2012 Intel Corporation
+ */
+
+#include <linux/errno.h>
+#include <linux/mutex.h>
+
+#include <drm/drm_mm.h>
+#include <drm/i915_drm.h>
+
+#include "gem/i915_gem_lmem.h"
+#include "gem/i915_gem_region.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_mcr.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_region_lmem.h"
+#include "i915_drv.h"
+#include "i915_gem_stolen.h"
+#include "i915_pci.h"
+#include "i915_reg.h"
+#include "i915_utils.h"
+#include "i915_vgpu.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pci_config.h"
+
+/*
+ * The BIOS typically reserves some of the system's memory for the exclusive
+ * use of the integrated graphics. This memory is no longer available for
+ * use by the OS and so the user finds that his system has less memory
+ * available than he put in. We refer to this memory as stolen.
+ *
+ * The BIOS will allocate its framebuffer from the stolen memory. Our
+ * goal is try to reuse that object for our own fbcon which must always
+ * be available for panics. Anything else we can reuse the stolen memory
+ * for is a boon.
+ */
+
+int i915_gem_stolen_insert_node_in_range(struct drm_i915_private *i915,
+					 struct drm_mm_node *node, u64 size,
+					 unsigned alignment, u64 start, u64 end)
+{
+	int ret;
+
+	if (!drm_mm_initialized(&i915->mm.stolen))
+		return -ENODEV;
+
+	/* WaSkipStolenMemoryFirstPage:bdw+ */
+	if (GRAPHICS_VER(i915) >= 8 && start < 4096)
+		start = 4096;
+
+	mutex_lock(&i915->mm.stolen_lock);
+	ret = drm_mm_insert_node_in_range(&i915->mm.stolen, node,
+					  size, alignment, 0,
+					  start, end, DRM_MM_INSERT_BEST);
+	mutex_unlock(&i915->mm.stolen_lock);
+
+	return ret;
+}
+
+int i915_gem_stolen_insert_node(struct drm_i915_private *i915,
+				struct drm_mm_node *node, u64 size,
+				unsigned alignment)
+{
+	return i915_gem_stolen_insert_node_in_range(i915, node,
+						    size, alignment,
+						    I915_GEM_STOLEN_BIAS,
+						    U64_MAX);
+}
+
+void i915_gem_stolen_remove_node(struct drm_i915_private *i915,
+				 struct drm_mm_node *node)
+{
+	mutex_lock(&i915->mm.stolen_lock);
+	drm_mm_remove_node(node);
+	mutex_unlock(&i915->mm.stolen_lock);
+}
+
+static int i915_adjust_stolen(struct drm_i915_private *i915,
+			      struct resource *dsm)
+{
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	struct intel_uncore *uncore = ggtt->vm.gt->uncore;
+	struct resource *r;
+
+	if (dsm->start == 0 || dsm->end <= dsm->start)
+		return -EINVAL;
+
+	/*
+	 * TODO: We have yet too encounter the case where the GTT wasn't at the
+	 * end of stolen. With that assumption we could simplify this.
+	 */
+
+	/* Make sure we don't clobber the GTT if it's within stolen memory */
+	if (GRAPHICS_VER(i915) <= 4 &&
+	    !IS_G33(i915) && !IS_PINEVIEW(i915) && !IS_G4X(i915)) {
+		struct resource stolen[2] = {*dsm, *dsm};
+		struct resource ggtt_res;
+		resource_size_t ggtt_start;
+
+		ggtt_start = intel_uncore_read(uncore, PGTBL_CTL);
+		if (GRAPHICS_VER(i915) == 4)
+			ggtt_start = (ggtt_start & PGTBL_ADDRESS_LO_MASK) |
+				     (ggtt_start & PGTBL_ADDRESS_HI_MASK) << 28;
+		else
+			ggtt_start &= PGTBL_ADDRESS_LO_MASK;
+
+		ggtt_res =
+			(struct resource) DEFINE_RES_MEM(ggtt_start,
+							 ggtt_total_entries(ggtt) * 4);
+
+		if (ggtt_res.start >= stolen[0].start && ggtt_res.start < stolen[0].end)
+			stolen[0].end = ggtt_res.start;
+		if (ggtt_res.end > stolen[1].start && ggtt_res.end <= stolen[1].end)
+			stolen[1].start = ggtt_res.end;
+
+		/* Pick the larger of the two chunks */
+		if (resource_size(&stolen[0]) > resource_size(&stolen[1]))
+			*dsm = stolen[0];
+		else
+			*dsm = stolen[1];
+
+		if (stolen[0].start != stolen[1].start ||
+		    stolen[0].end != stolen[1].end) {
+			drm_dbg(&i915->drm,
+				"GTT within stolen memory at %pR\n",
+				&ggtt_res);
+			drm_dbg(&i915->drm, "Stolen memory adjusted to %pR\n",
+				dsm);
+		}
+	}
+
+	/*
+	 * With stolen lmem, we don't need to check if the address range
+	 * overlaps with the non-stolen system memory range, since lmem is local
+	 * to the gpu.
+	 */
+	if (HAS_LMEM(i915))
+		return 0;
+
+	/*
+	 * Verify that nothing else uses this physical address. Stolen
+	 * memory should be reserved by the BIOS and hidden from the
+	 * kernel. So if the region is already marked as busy, something
+	 * is seriously wrong.
+	 */
+	r = devm_request_mem_region(i915->drm.dev, dsm->start,
+				    resource_size(dsm),
+				    "Graphics Stolen Memory");
+	if (r == NULL) {
+		/*
+		 * One more attempt but this time requesting region from
+		 * start + 1, as we have seen that this resolves the region
+		 * conflict with the PCI Bus.
+		 * This is a BIOS w/a: Some BIOS wrap stolen in the root
+		 * PCI bus, but have an off-by-one error. Hence retry the
+		 * reservation starting from 1 instead of 0.
+		 * There's also BIOS with off-by-one on the other end.
+		 */
+		r = devm_request_mem_region(i915->drm.dev, dsm->start + 1,
+					    resource_size(dsm) - 2,
+					    "Graphics Stolen Memory");
+		/*
+		 * GEN3 firmware likes to smash pci bridges into the stolen
+		 * range. Apparently this works.
+		 */
+		if (!r && GRAPHICS_VER(i915) != 3) {
+			drm_err(&i915->drm,
+				"conflict detected with stolen region: %pR\n",
+				dsm);
+
+			return -EBUSY;
+		}
+	}
+
+	return 0;
+}
+
+static void i915_gem_cleanup_stolen(struct drm_i915_private *i915)
+{
+	if (!drm_mm_initialized(&i915->mm.stolen))
+		return;
+
+	drm_mm_takedown(&i915->mm.stolen);
+}
+
+static void g4x_get_stolen_reserved(struct drm_i915_private *i915,
+				    struct intel_uncore *uncore,
+				    resource_size_t *base,
+				    resource_size_t *size)
+{
+	u32 reg_val = intel_uncore_read(uncore,
+					IS_GM45(i915) ?
+					CTG_STOLEN_RESERVED :
+					ELK_STOLEN_RESERVED);
+	resource_size_t stolen_top = i915->dsm.end + 1;
+
+	drm_dbg(&i915->drm, "%s_STOLEN_RESERVED = %08x\n",
+		IS_GM45(i915) ? "CTG" : "ELK", reg_val);
+
+	if ((reg_val & G4X_STOLEN_RESERVED_ENABLE) == 0)
+		return;
+
+	/*
+	 * Whether ILK really reuses the ELK register for this is unclear.
+	 * Let's see if we catch anyone with this supposedly enabled on ILK.
+	 */
+	drm_WARN(&i915->drm, GRAPHICS_VER(i915) == 5,
+		 "ILK stolen reserved found? 0x%08x\n",
+		 reg_val);
+
+	if (!(reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK))
+		return;
+
+	*base = (reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK) << 16;
+	drm_WARN_ON(&i915->drm,
+		    (reg_val & G4X_STOLEN_RESERVED_ADDR1_MASK) < *base);
+
+	*size = stolen_top - *base;
+}
+
+static void gen6_get_stolen_reserved(struct drm_i915_private *i915,
+				     struct intel_uncore *uncore,
+				     resource_size_t *base,
+				     resource_size_t *size)
+{
+	u32 reg_val = intel_uncore_read(uncore, GEN6_STOLEN_RESERVED);
+
+	drm_dbg(&i915->drm, "GEN6_STOLEN_RESERVED = %08x\n", reg_val);
+
+	if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE))
+		return;
+
+	*base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK;
+
+	switch (reg_val & GEN6_STOLEN_RESERVED_SIZE_MASK) {
+	case GEN6_STOLEN_RESERVED_1M:
+		*size = 1024 * 1024;
+		break;
+	case GEN6_STOLEN_RESERVED_512K:
+		*size = 512 * 1024;
+		break;
+	case GEN6_STOLEN_RESERVED_256K:
+		*size = 256 * 1024;
+		break;
+	case GEN6_STOLEN_RESERVED_128K:
+		*size = 128 * 1024;
+		break;
+	default:
+		*size = 1024 * 1024;
+		MISSING_CASE(reg_val & GEN6_STOLEN_RESERVED_SIZE_MASK);
+	}
+}
+
+static void vlv_get_stolen_reserved(struct drm_i915_private *i915,
+				    struct intel_uncore *uncore,
+				    resource_size_t *base,
+				    resource_size_t *size)
+{
+	u32 reg_val = intel_uncore_read(uncore, GEN6_STOLEN_RESERVED);
+	resource_size_t stolen_top = i915->dsm.end + 1;
+
+	drm_dbg(&i915->drm, "GEN6_STOLEN_RESERVED = %08x\n", reg_val);
+
+	if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE))
+		return;
+
+	switch (reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK) {
+	default:
+		MISSING_CASE(reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK);
+		fallthrough;
+	case GEN7_STOLEN_RESERVED_1M:
+		*size = 1024 * 1024;
+		break;
+	}
+
+	/*
+	 * On vlv, the ADDR_MASK portion is left as 0 and HW deduces the
+	 * reserved location as (top - size).
+	 */
+	*base = stolen_top - *size;
+}
+
+static void gen7_get_stolen_reserved(struct drm_i915_private *i915,
+				     struct intel_uncore *uncore,
+				     resource_size_t *base,
+				     resource_size_t *size)
+{
+	u32 reg_val = intel_uncore_read(uncore, GEN6_STOLEN_RESERVED);
+
+	drm_dbg(&i915->drm, "GEN6_STOLEN_RESERVED = %08x\n", reg_val);
+
+	if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE))
+		return;
+
+	*base = reg_val & GEN7_STOLEN_RESERVED_ADDR_MASK;
+
+	switch (reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK) {
+	case GEN7_STOLEN_RESERVED_1M:
+		*size = 1024 * 1024;
+		break;
+	case GEN7_STOLEN_RESERVED_256K:
+		*size = 256 * 1024;
+		break;
+	default:
+		*size = 1024 * 1024;
+		MISSING_CASE(reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK);
+	}
+}
+
+static void chv_get_stolen_reserved(struct drm_i915_private *i915,
+				    struct intel_uncore *uncore,
+				    resource_size_t *base,
+				    resource_size_t *size)
+{
+	u32 reg_val = intel_uncore_read(uncore, GEN6_STOLEN_RESERVED);
+
+	drm_dbg(&i915->drm, "GEN6_STOLEN_RESERVED = %08x\n", reg_val);
+
+	if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE))
+		return;
+
+	*base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK;
+
+	switch (reg_val & GEN8_STOLEN_RESERVED_SIZE_MASK) {
+	case GEN8_STOLEN_RESERVED_1M:
+		*size = 1024 * 1024;
+		break;
+	case GEN8_STOLEN_RESERVED_2M:
+		*size = 2 * 1024 * 1024;
+		break;
+	case GEN8_STOLEN_RESERVED_4M:
+		*size = 4 * 1024 * 1024;
+		break;
+	case GEN8_STOLEN_RESERVED_8M:
+		*size = 8 * 1024 * 1024;
+		break;
+	default:
+		*size = 8 * 1024 * 1024;
+		MISSING_CASE(reg_val & GEN8_STOLEN_RESERVED_SIZE_MASK);
+	}
+}
+
+static void bdw_get_stolen_reserved(struct drm_i915_private *i915,
+				    struct intel_uncore *uncore,
+				    resource_size_t *base,
+				    resource_size_t *size)
+{
+	u32 reg_val = intel_uncore_read(uncore, GEN6_STOLEN_RESERVED);
+	resource_size_t stolen_top = i915->dsm.end + 1;
+
+	drm_dbg(&i915->drm, "GEN6_STOLEN_RESERVED = %08x\n", reg_val);
+
+	if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE))
+		return;
+
+	if (!(reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK))
+		return;
+
+	*base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK;
+	*size = stolen_top - *base;
+}
+
+static void icl_get_stolen_reserved(struct drm_i915_private *i915,
+				    struct intel_uncore *uncore,
+				    resource_size_t *base,
+				    resource_size_t *size)
+{
+	u64 reg_val = intel_uncore_read64(uncore, GEN6_STOLEN_RESERVED);
+
+	drm_dbg(&i915->drm, "GEN6_STOLEN_RESERVED = 0x%016llx\n", reg_val);
+
+	*base = reg_val & GEN11_STOLEN_RESERVED_ADDR_MASK;
+
+	switch (reg_val & GEN8_STOLEN_RESERVED_SIZE_MASK) {
+	case GEN8_STOLEN_RESERVED_1M:
+		*size = 1024 * 1024;
+		break;
+	case GEN8_STOLEN_RESERVED_2M:
+		*size = 2 * 1024 * 1024;
+		break;
+	case GEN8_STOLEN_RESERVED_4M:
+		*size = 4 * 1024 * 1024;
+		break;
+	case GEN8_STOLEN_RESERVED_8M:
+		*size = 8 * 1024 * 1024;
+		break;
+	default:
+		*size = 8 * 1024 * 1024;
+		MISSING_CASE(reg_val & GEN8_STOLEN_RESERVED_SIZE_MASK);
+	}
+}
+
+static int i915_gem_init_stolen(struct intel_memory_region *mem)
+{
+	struct drm_i915_private *i915 = mem->i915;
+	struct intel_uncore *uncore = &i915->uncore;
+	resource_size_t reserved_base, stolen_top;
+	resource_size_t reserved_total, reserved_size;
+
+	mutex_init(&i915->mm.stolen_lock);
+
+	if (intel_vgpu_active(i915)) {
+		drm_notice(&i915->drm,
+			   "%s, disabling use of stolen memory\n",
+			   "iGVT-g active");
+		return 0;
+	}
+
+	if (i915_vtd_active(i915) && GRAPHICS_VER(i915) < 8) {
+		drm_notice(&i915->drm,
+			   "%s, disabling use of stolen memory\n",
+			   "DMAR active");
+		return 0;
+	}
+
+	if (resource_size(&mem->region) == 0)
+		return 0;
+
+	i915->dsm = mem->region;
+
+	if (i915_adjust_stolen(i915, &i915->dsm))
+		return 0;
+
+	GEM_BUG_ON(i915->dsm.start == 0);
+	GEM_BUG_ON(i915->dsm.end <= i915->dsm.start);
+
+	stolen_top = i915->dsm.end + 1;
+	reserved_base = stolen_top;
+	reserved_size = 0;
+
+	if (GRAPHICS_VER(i915) >= 11) {
+		icl_get_stolen_reserved(i915, uncore,
+					&reserved_base, &reserved_size);
+	} else if (GRAPHICS_VER(i915) >= 8) {
+		if (IS_LP(i915))
+			chv_get_stolen_reserved(i915, uncore,
+						&reserved_base, &reserved_size);
+		else
+			bdw_get_stolen_reserved(i915, uncore,
+						&reserved_base, &reserved_size);
+	} else if (GRAPHICS_VER(i915) >= 7) {
+		if (IS_VALLEYVIEW(i915))
+			vlv_get_stolen_reserved(i915, uncore,
+						&reserved_base, &reserved_size);
+		else
+			gen7_get_stolen_reserved(i915, uncore,
+						 &reserved_base, &reserved_size);
+	} else if (GRAPHICS_VER(i915) >= 6) {
+		gen6_get_stolen_reserved(i915, uncore,
+					 &reserved_base, &reserved_size);
+	} else if (GRAPHICS_VER(i915) >= 5 || IS_G4X(i915)) {
+		g4x_get_stolen_reserved(i915, uncore,
+					&reserved_base, &reserved_size);
+	}
+
+	/*
+	 * Our expectation is that the reserved space is at the top of the
+	 * stolen region and *never* at the bottom. If we see !reserved_base,
+	 * it likely means we failed to read the registers correctly.
+	 */
+	if (!reserved_base) {
+		drm_err(&i915->drm,
+			"inconsistent reservation %pa + %pa; ignoring\n",
+			&reserved_base, &reserved_size);
+		reserved_base = stolen_top;
+		reserved_size = 0;
+	}
+
+	i915->dsm_reserved =
+		(struct resource)DEFINE_RES_MEM(reserved_base, reserved_size);
+
+	if (!resource_contains(&i915->dsm, &i915->dsm_reserved)) {
+		drm_err(&i915->drm,
+			"Stolen reserved area %pR outside stolen memory %pR\n",
+			&i915->dsm_reserved, &i915->dsm);
+		return 0;
+	}
+
+	/* Exclude the reserved region from driver use */
+	mem->region.end = reserved_base - 1;
+	mem->io_size = min(mem->io_size, resource_size(&mem->region));
+
+	/* It is possible for the reserved area to end before the end of stolen
+	 * memory, so just consider the start. */
+	reserved_total = stolen_top - reserved_base;
+
+	i915->stolen_usable_size =
+		resource_size(&i915->dsm) - reserved_total;
+
+	drm_dbg(&i915->drm,
+		"Memory reserved for graphics device: %lluK, usable: %lluK\n",
+		(u64)resource_size(&i915->dsm) >> 10,
+		(u64)i915->stolen_usable_size >> 10);
+
+	if (i915->stolen_usable_size == 0)
+		return 0;
+
+	/* Basic memrange allocator for stolen space. */
+	drm_mm_init(&i915->mm.stolen, 0, i915->stolen_usable_size);
+
+	return 0;
+}
+
+static void dbg_poison(struct i915_ggtt *ggtt,
+		       dma_addr_t addr, resource_size_t size,
+		       u8 x)
+{
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)
+	if (!drm_mm_node_allocated(&ggtt->error_capture))
+		return;
+
+	if (ggtt->vm.bind_async_flags & I915_VMA_GLOBAL_BIND)
+		return; /* beware stop_machine() inversion */
+
+	GEM_BUG_ON(!IS_ALIGNED(size, PAGE_SIZE));
+
+	mutex_lock(&ggtt->error_mutex);
+	while (size) {
+		void __iomem *s;
+
+		ggtt->vm.insert_page(&ggtt->vm, addr,
+				     ggtt->error_capture.start,
+				     I915_CACHE_NONE, 0);
+		mb();
+
+		s = io_mapping_map_wc(&ggtt->iomap,
+				      ggtt->error_capture.start,
+				      PAGE_SIZE);
+		memset_io(s, x, PAGE_SIZE);
+		io_mapping_unmap(s);
+
+		addr += PAGE_SIZE;
+		size -= PAGE_SIZE;
+	}
+	mb();
+	ggtt->vm.clear_range(&ggtt->vm, ggtt->error_capture.start, PAGE_SIZE);
+	mutex_unlock(&ggtt->error_mutex);
+#endif
+}
+
+static struct sg_table *
+i915_pages_create_for_stolen(struct drm_device *dev,
+			     resource_size_t offset, resource_size_t size)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct sg_table *st;
+	struct scatterlist *sg;
+
+	GEM_BUG_ON(range_overflows(offset, size, resource_size(&i915->dsm)));
+
+	/* We hide that we have no struct page backing our stolen object
+	 * by wrapping the contiguous physical allocation with a fake
+	 * dma mapping in a single scatterlist.
+	 */
+
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (st == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	if (sg_alloc_table(st, 1, GFP_KERNEL)) {
+		kfree(st);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	sg = st->sgl;
+	sg->offset = 0;
+	sg->length = size;
+
+	sg_dma_address(sg) = (dma_addr_t)i915->dsm.start + offset;
+	sg_dma_len(sg) = size;
+
+	return st;
+}
+
+static int i915_gem_object_get_pages_stolen(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct sg_table *pages =
+		i915_pages_create_for_stolen(obj->base.dev,
+					     obj->stolen->start,
+					     obj->stolen->size);
+	if (IS_ERR(pages))
+		return PTR_ERR(pages);
+
+	dbg_poison(to_gt(i915)->ggtt,
+		   sg_dma_address(pages->sgl),
+		   sg_dma_len(pages->sgl),
+		   POISON_INUSE);
+
+	__i915_gem_object_set_pages(obj, pages, obj->stolen->size);
+
+	return 0;
+}
+
+static void i915_gem_object_put_pages_stolen(struct drm_i915_gem_object *obj,
+					     struct sg_table *pages)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	/* Should only be called from i915_gem_object_release_stolen() */
+
+	dbg_poison(to_gt(i915)->ggtt,
+		   sg_dma_address(pages->sgl),
+		   sg_dma_len(pages->sgl),
+		   POISON_FREE);
+
+	sg_free_table(pages);
+	kfree(pages);
+}
+
+static void
+i915_gem_object_release_stolen(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct drm_mm_node *stolen = fetch_and_zero(&obj->stolen);
+
+	GEM_BUG_ON(!stolen);
+	i915_gem_stolen_remove_node(i915, stolen);
+	kfree(stolen);
+
+	i915_gem_object_release_memory_region(obj);
+}
+
+static const struct drm_i915_gem_object_ops i915_gem_object_stolen_ops = {
+	.name = "i915_gem_object_stolen",
+	.get_pages = i915_gem_object_get_pages_stolen,
+	.put_pages = i915_gem_object_put_pages_stolen,
+	.release = i915_gem_object_release_stolen,
+};
+
+static int __i915_gem_object_create_stolen(struct intel_memory_region *mem,
+					   struct drm_i915_gem_object *obj,
+					   struct drm_mm_node *stolen)
+{
+	static struct lock_class_key lock_class;
+	unsigned int cache_level;
+	unsigned int flags;
+	int err;
+
+	/*
+	 * Stolen objects are always physically contiguous since we just
+	 * allocate one big block underneath using the drm_mm range allocator.
+	 */
+	flags = I915_BO_ALLOC_CONTIGUOUS;
+
+	drm_gem_private_object_init(&mem->i915->drm, &obj->base, stolen->size);
+	i915_gem_object_init(obj, &i915_gem_object_stolen_ops, &lock_class, flags);
+
+	obj->stolen = stolen;
+	obj->read_domains = I915_GEM_DOMAIN_CPU | I915_GEM_DOMAIN_GTT;
+	cache_level = HAS_LLC(mem->i915) ? I915_CACHE_LLC : I915_CACHE_NONE;
+	i915_gem_object_set_cache_coherency(obj, cache_level);
+
+	if (WARN_ON(!i915_gem_object_trylock(obj, NULL)))
+		return -EBUSY;
+
+	i915_gem_object_init_memory_region(obj, mem);
+
+	err = i915_gem_object_pin_pages(obj);
+	if (err)
+		i915_gem_object_release_memory_region(obj);
+	i915_gem_object_unlock(obj);
+
+	return err;
+}
+
+static int _i915_gem_object_stolen_init(struct intel_memory_region *mem,
+					struct drm_i915_gem_object *obj,
+					resource_size_t offset,
+					resource_size_t size,
+					resource_size_t page_size,
+					unsigned int flags)
+{
+	struct drm_i915_private *i915 = mem->i915;
+	struct drm_mm_node *stolen;
+	int ret;
+
+	if (!drm_mm_initialized(&i915->mm.stolen))
+		return -ENODEV;
+
+	if (size == 0)
+		return -EINVAL;
+
+	/*
+	 * With discrete devices, where we lack a mappable aperture there is no
+	 * possible way to ever access this memory on the CPU side.
+	 */
+	if (mem->type == INTEL_MEMORY_STOLEN_LOCAL && !mem->io_size &&
+	    !(flags & I915_BO_ALLOC_GPU_ONLY))
+		return -ENOSPC;
+
+	stolen = kzalloc(sizeof(*stolen), GFP_KERNEL);
+	if (!stolen)
+		return -ENOMEM;
+
+	if (offset != I915_BO_INVALID_OFFSET) {
+		drm_dbg(&i915->drm,
+			"creating preallocated stolen object: stolen_offset=%pa, size=%pa\n",
+			&offset, &size);
+
+		stolen->start = offset;
+		stolen->size = size;
+		mutex_lock(&i915->mm.stolen_lock);
+		ret = drm_mm_reserve_node(&i915->mm.stolen, stolen);
+		mutex_unlock(&i915->mm.stolen_lock);
+	} else {
+		ret = i915_gem_stolen_insert_node(i915, stolen, size,
+						  mem->min_page_size);
+	}
+	if (ret)
+		goto err_free;
+
+	ret = __i915_gem_object_create_stolen(mem, obj, stolen);
+	if (ret)
+		goto err_remove;
+
+	return 0;
+
+err_remove:
+	i915_gem_stolen_remove_node(i915, stolen);
+err_free:
+	kfree(stolen);
+	return ret;
+}
+
+struct drm_i915_gem_object *
+i915_gem_object_create_stolen(struct drm_i915_private *i915,
+			      resource_size_t size)
+{
+	return i915_gem_object_create_region(i915->mm.stolen_region, size, 0, 0);
+}
+
+static int init_stolen_smem(struct intel_memory_region *mem)
+{
+	/*
+	 * Initialise stolen early so that we may reserve preallocated
+	 * objects for the BIOS to KMS transition.
+	 */
+	return i915_gem_init_stolen(mem);
+}
+
+static int release_stolen_smem(struct intel_memory_region *mem)
+{
+	i915_gem_cleanup_stolen(mem->i915);
+	return 0;
+}
+
+static const struct intel_memory_region_ops i915_region_stolen_smem_ops = {
+	.init = init_stolen_smem,
+	.release = release_stolen_smem,
+	.init_object = _i915_gem_object_stolen_init,
+};
+
+static int init_stolen_lmem(struct intel_memory_region *mem)
+{
+	int err;
+
+	if (GEM_WARN_ON(resource_size(&mem->region) == 0))
+		return -ENODEV;
+
+	/*
+	 * TODO: For stolen lmem we mostly just care about populating the dsm
+	 * related bits and setting up the drm_mm allocator for the range.
+	 * Perhaps split up i915_gem_init_stolen() for this.
+	 */
+	err = i915_gem_init_stolen(mem);
+	if (err)
+		return err;
+
+	if (mem->io_size && !io_mapping_init_wc(&mem->iomap,
+						mem->io_start,
+						mem->io_size)) {
+		err = -EIO;
+		goto err_cleanup;
+	}
+
+	return 0;
+
+err_cleanup:
+	i915_gem_cleanup_stolen(mem->i915);
+	return err;
+}
+
+static int release_stolen_lmem(struct intel_memory_region *mem)
+{
+	if (mem->io_size)
+		io_mapping_fini(&mem->iomap);
+	i915_gem_cleanup_stolen(mem->i915);
+	return 0;
+}
+
+static const struct intel_memory_region_ops i915_region_stolen_lmem_ops = {
+	.init = init_stolen_lmem,
+	.release = release_stolen_lmem,
+	.init_object = _i915_gem_object_stolen_init,
+};
+
+struct intel_memory_region *
+i915_gem_stolen_lmem_setup(struct drm_i915_private *i915, u16 type,
+			   u16 instance)
+{
+	struct intel_uncore *uncore = &i915->uncore;
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	resource_size_t dsm_size, dsm_base, lmem_size;
+	struct intel_memory_region *mem;
+	resource_size_t io_start, io_size;
+	resource_size_t min_page_size;
+
+	if (WARN_ON_ONCE(instance))
+		return ERR_PTR(-ENODEV);
+
+	if (!i915_pci_resource_valid(pdev, GEN12_LMEM_BAR))
+		return ERR_PTR(-ENXIO);
+
+	/* Use DSM base address instead for stolen memory */
+	dsm_base = intel_uncore_read64(uncore, GEN12_DSMBASE);
+	if (IS_DG1(uncore->i915)) {
+		lmem_size = pci_resource_len(pdev, GEN12_LMEM_BAR);
+		if (WARN_ON(lmem_size < dsm_base))
+			return ERR_PTR(-ENODEV);
+	} else {
+		resource_size_t lmem_range;
+
+		lmem_range = intel_gt_mcr_read_any(&i915->gt0, XEHP_TILE0_ADDR_RANGE) & 0xFFFF;
+		lmem_size = lmem_range >> XEHP_TILE_LMEM_RANGE_SHIFT;
+		lmem_size *= SZ_1G;
+	}
+
+	dsm_size = lmem_size - dsm_base;
+	if (pci_resource_len(pdev, GEN12_LMEM_BAR) < lmem_size) {
+		io_start = 0;
+		io_size = 0;
+	} else {
+		io_start = pci_resource_start(pdev, GEN12_LMEM_BAR) + dsm_base;
+		io_size = dsm_size;
+	}
+
+	min_page_size = HAS_64K_PAGES(i915) ? I915_GTT_PAGE_SIZE_64K :
+						I915_GTT_PAGE_SIZE_4K;
+
+	mem = intel_memory_region_create(i915, dsm_base, dsm_size,
+					 min_page_size,
+					 io_start, io_size,
+					 type, instance,
+					 &i915_region_stolen_lmem_ops);
+	if (IS_ERR(mem))
+		return mem;
+
+	/*
+	 * TODO: consider creating common helper to just print all the
+	 * interesting stuff from intel_memory_region, which we can use for all
+	 * our probed regions.
+	 */
+
+	drm_dbg(&i915->drm, "Stolen Local memory IO start: %pa\n",
+		&mem->io_start);
+	drm_dbg(&i915->drm, "Stolen Local DSM base: %pa\n", &dsm_base);
+
+	intel_memory_region_set_name(mem, "stolen-local");
+
+	mem->private = true;
+
+	return mem;
+}
+
+struct intel_memory_region*
+i915_gem_stolen_smem_setup(struct drm_i915_private *i915, u16 type,
+			   u16 instance)
+{
+	struct intel_memory_region *mem;
+
+	mem = intel_memory_region_create(i915,
+					 intel_graphics_stolen_res.start,
+					 resource_size(&intel_graphics_stolen_res),
+					 PAGE_SIZE, 0, 0, type, instance,
+					 &i915_region_stolen_smem_ops);
+	if (IS_ERR(mem))
+		return mem;
+
+	intel_memory_region_set_name(mem, "stolen-system");
+
+	mem->private = true;
+	return mem;
+}
+
+bool i915_gem_object_is_stolen(const struct drm_i915_gem_object *obj)
+{
+	return obj->ops == &i915_gem_object_stolen_ops;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_stolen.h b/drivers/gpu/drm/i915/gem/i915_gem_stolen.h
new file mode 100644
index 000000000..d5005a39d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_stolen.h
@@ -0,0 +1,39 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_GEM_STOLEN_H__
+#define __I915_GEM_STOLEN_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct drm_mm_node;
+struct drm_i915_gem_object;
+
+int i915_gem_stolen_insert_node(struct drm_i915_private *dev_priv,
+				struct drm_mm_node *node, u64 size,
+				unsigned alignment);
+int i915_gem_stolen_insert_node_in_range(struct drm_i915_private *dev_priv,
+					 struct drm_mm_node *node, u64 size,
+					 unsigned alignment, u64 start,
+					 u64 end);
+void i915_gem_stolen_remove_node(struct drm_i915_private *dev_priv,
+				 struct drm_mm_node *node);
+struct intel_memory_region *
+i915_gem_stolen_smem_setup(struct drm_i915_private *i915, u16 type,
+			   u16 instance);
+struct intel_memory_region *
+i915_gem_stolen_lmem_setup(struct drm_i915_private *i915, u16 type,
+			   u16 instance);
+
+struct drm_i915_gem_object *
+i915_gem_object_create_stolen(struct drm_i915_private *dev_priv,
+			      resource_size_t size);
+
+bool i915_gem_object_is_stolen(const struct drm_i915_gem_object *obj);
+
+#define I915_GEM_STOLEN_BIAS SZ_128K
+
+#endif /* __I915_GEM_STOLEN_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_throttle.c b/drivers/gpu/drm/i915/gem/i915_gem_throttle.c
new file mode 100644
index 000000000..af85d0c28
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_throttle.c
@@ -0,0 +1,102 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2014-2016 Intel Corporation
+ */
+
+#include <linux/jiffies.h>
+
+#include <drm/drm_file.h>
+
+#include "i915_drv.h"
+#include "i915_file_private.h"
+#include "i915_gem_context.h"
+#include "i915_gem_ioctls.h"
+#include "i915_gem_object.h"
+
+/*
+ * 20ms is a fairly arbitrary limit (greater than the average frame time)
+ * chosen to prevent the CPU getting more than a frame ahead of the GPU
+ * (when using lax throttling for the frontbuffer). We also use it to
+ * offer free GPU waitboosts for severely congested workloads.
+ */
+#define DRM_I915_THROTTLE_JIFFIES msecs_to_jiffies(20)
+
+/*
+ * Throttle our rendering by waiting until the ring has completed our requests
+ * emitted over 20 msec ago.
+ *
+ * Note that if we were to use the current jiffies each time around the loop,
+ * we wouldn't escape the function with any frames outstanding if the time to
+ * render a frame was over 20ms.
+ *
+ * This should get us reasonable parallelism between CPU and GPU but also
+ * relatively low latency when blocking on a particular request to finish.
+ */
+int
+i915_gem_throttle_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file)
+{
+	const unsigned long recent_enough = jiffies - DRM_I915_THROTTLE_JIFFIES;
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct i915_gem_context *ctx;
+	unsigned long idx;
+	long ret;
+
+	/* ABI: return -EIO if already wedged */
+	ret = intel_gt_terminally_wedged(to_gt(i915));
+	if (ret)
+		return ret;
+
+	rcu_read_lock();
+	xa_for_each(&file_priv->context_xa, idx, ctx) {
+		struct i915_gem_engines_iter it;
+		struct intel_context *ce;
+
+		if (!kref_get_unless_zero(&ctx->ref))
+			continue;
+		rcu_read_unlock();
+
+		for_each_gem_engine(ce,
+				    i915_gem_context_lock_engines(ctx),
+				    it) {
+			struct i915_request *rq, *target = NULL;
+
+			if (!ce->timeline)
+				continue;
+
+			mutex_lock(&ce->timeline->mutex);
+			list_for_each_entry_reverse(rq,
+						    &ce->timeline->requests,
+						    link) {
+				if (i915_request_completed(rq))
+					break;
+
+				if (time_after(rq->emitted_jiffies,
+					       recent_enough))
+					continue;
+
+				target = i915_request_get(rq);
+				break;
+			}
+			mutex_unlock(&ce->timeline->mutex);
+			if (!target)
+				continue;
+
+			ret = i915_request_wait(target,
+						I915_WAIT_INTERRUPTIBLE,
+						MAX_SCHEDULE_TIMEOUT);
+			i915_request_put(target);
+			if (ret < 0)
+				break;
+		}
+		i915_gem_context_unlock_engines(ctx);
+		i915_gem_context_put(ctx);
+
+		rcu_read_lock();
+	}
+	rcu_read_unlock();
+
+	return ret < 0 ? ret : 0;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_tiling.c b/drivers/gpu/drm/i915/gem/i915_gem_tiling.c
new file mode 100644
index 000000000..bc21b1c23
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_tiling.c
@@ -0,0 +1,472 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2008 Intel Corporation
+ */
+
+#include <linux/string.h>
+#include <linux/bitops.h>
+
+#include "i915_drv.h"
+#include "i915_gem.h"
+#include "i915_gem_ioctls.h"
+#include "i915_gem_mman.h"
+#include "i915_gem_object.h"
+#include "i915_gem_tiling.h"
+#include "i915_reg.h"
+
+/**
+ * DOC: buffer object tiling
+ *
+ * i915_gem_set_tiling_ioctl() and i915_gem_get_tiling_ioctl() is the userspace
+ * interface to declare fence register requirements.
+ *
+ * In principle GEM doesn't care at all about the internal data layout of an
+ * object, and hence it also doesn't care about tiling or swizzling. There's two
+ * exceptions:
+ *
+ * - For X and Y tiling the hardware provides detilers for CPU access, so called
+ *   fences. Since there's only a limited amount of them the kernel must manage
+ *   these, and therefore userspace must tell the kernel the object tiling if it
+ *   wants to use fences for detiling.
+ * - On gen3 and gen4 platforms have a swizzling pattern for tiled objects which
+ *   depends upon the physical page frame number. When swapping such objects the
+ *   page frame number might change and the kernel must be able to fix this up
+ *   and hence now the tiling. Note that on a subset of platforms with
+ *   asymmetric memory channel population the swizzling pattern changes in an
+ *   unknown way, and for those the kernel simply forbids swapping completely.
+ *
+ * Since neither of this applies for new tiling layouts on modern platforms like
+ * W, Ys and Yf tiling GEM only allows object tiling to be set to X or Y tiled.
+ * Anything else can be handled in userspace entirely without the kernel's
+ * invovlement.
+ */
+
+/**
+ * i915_gem_fence_size - required global GTT size for a fence
+ * @i915: i915 device
+ * @size: object size
+ * @tiling: tiling mode
+ * @stride: tiling stride
+ *
+ * Return the required global GTT size for a fence (view of a tiled object),
+ * taking into account potential fence register mapping.
+ */
+u32 i915_gem_fence_size(struct drm_i915_private *i915,
+			u32 size, unsigned int tiling, unsigned int stride)
+{
+	u32 ggtt_size;
+
+	GEM_BUG_ON(!size);
+
+	if (tiling == I915_TILING_NONE)
+		return size;
+
+	GEM_BUG_ON(!stride);
+
+	if (GRAPHICS_VER(i915) >= 4) {
+		stride *= i915_gem_tile_height(tiling);
+		GEM_BUG_ON(!IS_ALIGNED(stride, I965_FENCE_PAGE));
+		return roundup(size, stride);
+	}
+
+	/* Previous chips need a power-of-two fence region when tiling */
+	if (GRAPHICS_VER(i915) == 3)
+		ggtt_size = 1024*1024;
+	else
+		ggtt_size = 512*1024;
+
+	while (ggtt_size < size)
+		ggtt_size <<= 1;
+
+	return ggtt_size;
+}
+
+/**
+ * i915_gem_fence_alignment - required global GTT alignment for a fence
+ * @i915: i915 device
+ * @size: object size
+ * @tiling: tiling mode
+ * @stride: tiling stride
+ *
+ * Return the required global GTT alignment for a fence (a view of a tiled
+ * object), taking into account potential fence register mapping.
+ */
+u32 i915_gem_fence_alignment(struct drm_i915_private *i915, u32 size,
+			     unsigned int tiling, unsigned int stride)
+{
+	GEM_BUG_ON(!size);
+
+	/*
+	 * Minimum alignment is 4k (GTT page size), but might be greater
+	 * if a fence register is needed for the object.
+	 */
+	if (tiling == I915_TILING_NONE)
+		return I915_GTT_MIN_ALIGNMENT;
+
+	if (GRAPHICS_VER(i915) >= 4)
+		return I965_FENCE_PAGE;
+
+	/*
+	 * Previous chips need to be aligned to the size of the smallest
+	 * fence register that can contain the object.
+	 */
+	return i915_gem_fence_size(i915, size, tiling, stride);
+}
+
+/* Check pitch constraints for all chips & tiling formats */
+static bool
+i915_tiling_ok(struct drm_i915_gem_object *obj,
+	       unsigned int tiling, unsigned int stride)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	unsigned int tile_width;
+
+	/* Linear is always fine */
+	if (tiling == I915_TILING_NONE)
+		return true;
+
+	if (tiling > I915_TILING_LAST)
+		return false;
+
+	/* check maximum stride & object size */
+	/* i965+ stores the end address of the gtt mapping in the fence
+	 * reg, so dont bother to check the size */
+	if (GRAPHICS_VER(i915) >= 7) {
+		if (stride / 128 > GEN7_FENCE_MAX_PITCH_VAL)
+			return false;
+	} else if (GRAPHICS_VER(i915) >= 4) {
+		if (stride / 128 > I965_FENCE_MAX_PITCH_VAL)
+			return false;
+	} else {
+		if (stride > 8192)
+			return false;
+
+		if (!is_power_of_2(stride))
+			return false;
+	}
+
+	if (GRAPHICS_VER(i915) == 2 ||
+	    (tiling == I915_TILING_Y && HAS_128_BYTE_Y_TILING(i915)))
+		tile_width = 128;
+	else
+		tile_width = 512;
+
+	if (!stride || !IS_ALIGNED(stride, tile_width))
+		return false;
+
+	return true;
+}
+
+static bool i915_vma_fence_prepare(struct i915_vma *vma,
+				   int tiling_mode, unsigned int stride)
+{
+	struct drm_i915_private *i915 = vma->vm->i915;
+	u32 size, alignment;
+
+	if (!i915_vma_is_map_and_fenceable(vma))
+		return true;
+
+	size = i915_gem_fence_size(i915, vma->size, tiling_mode, stride);
+	if (vma->node.size < size)
+		return false;
+
+	alignment = i915_gem_fence_alignment(i915, vma->size, tiling_mode, stride);
+	if (!IS_ALIGNED(vma->node.start, alignment))
+		return false;
+
+	return true;
+}
+
+/* Make the current GTT allocation valid for the change in tiling. */
+static int
+i915_gem_object_fence_prepare(struct drm_i915_gem_object *obj,
+			      int tiling_mode, unsigned int stride)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	struct i915_vma *vma, *vn;
+	LIST_HEAD(unbind);
+	int ret = 0;
+
+	if (tiling_mode == I915_TILING_NONE)
+		return 0;
+
+	mutex_lock(&ggtt->vm.mutex);
+
+	spin_lock(&obj->vma.lock);
+	for_each_ggtt_vma(vma, obj) {
+		GEM_BUG_ON(vma->vm != &ggtt->vm);
+
+		if (i915_vma_fence_prepare(vma, tiling_mode, stride))
+			continue;
+
+		list_move(&vma->vm_link, &unbind);
+	}
+	spin_unlock(&obj->vma.lock);
+
+	list_for_each_entry_safe(vma, vn, &unbind, vm_link) {
+		ret = __i915_vma_unbind(vma);
+		if (ret) {
+			/* Restore the remaining vma on an error */
+			list_splice(&unbind, &ggtt->vm.bound_list);
+			break;
+		}
+	}
+
+	mutex_unlock(&ggtt->vm.mutex);
+
+	return ret;
+}
+
+bool i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+
+	return to_gt(i915)->ggtt->bit_6_swizzle_x == I915_BIT_6_SWIZZLE_9_10_17 &&
+		i915_gem_object_is_tiled(obj);
+}
+
+int
+i915_gem_object_set_tiling(struct drm_i915_gem_object *obj,
+			   unsigned int tiling, unsigned int stride)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_vma *vma;
+	int err;
+
+	/* Make sure we don't cross-contaminate obj->tiling_and_stride */
+	BUILD_BUG_ON(I915_TILING_LAST & STRIDE_MASK);
+
+	GEM_BUG_ON(!i915_tiling_ok(obj, tiling, stride));
+	GEM_BUG_ON(!stride ^ (tiling == I915_TILING_NONE));
+
+	if ((tiling | stride) == obj->tiling_and_stride)
+		return 0;
+
+	if (i915_gem_object_is_framebuffer(obj))
+		return -EBUSY;
+
+	/* We need to rebind the object if its current allocation
+	 * no longer meets the alignment restrictions for its new
+	 * tiling mode. Otherwise we can just leave it alone, but
+	 * need to ensure that any fence register is updated before
+	 * the next fenced (either through the GTT or by the BLT unit
+	 * on older GPUs) access.
+	 *
+	 * After updating the tiling parameters, we then flag whether
+	 * we need to update an associated fence register. Note this
+	 * has to also include the unfenced register the GPU uses
+	 * whilst executing a fenced command for an untiled object.
+	 */
+
+	i915_gem_object_lock(obj, NULL);
+	if (i915_gem_object_is_framebuffer(obj)) {
+		i915_gem_object_unlock(obj);
+		return -EBUSY;
+	}
+
+	err = i915_gem_object_fence_prepare(obj, tiling, stride);
+	if (err) {
+		i915_gem_object_unlock(obj);
+		return err;
+	}
+
+	/* If the memory has unknown (i.e. varying) swizzling, we pin the
+	 * pages to prevent them being swapped out and causing corruption
+	 * due to the change in swizzling.
+	 */
+	if (i915_gem_object_has_pages(obj) &&
+	    obj->mm.madv == I915_MADV_WILLNEED &&
+	    i915->gem_quirks & GEM_QUIRK_PIN_SWIZZLED_PAGES) {
+		if (tiling == I915_TILING_NONE) {
+			GEM_BUG_ON(!i915_gem_object_has_tiling_quirk(obj));
+			i915_gem_object_clear_tiling_quirk(obj);
+			i915_gem_object_make_shrinkable(obj);
+		}
+		if (!i915_gem_object_is_tiled(obj)) {
+			GEM_BUG_ON(i915_gem_object_has_tiling_quirk(obj));
+			i915_gem_object_make_unshrinkable(obj);
+			i915_gem_object_set_tiling_quirk(obj);
+		}
+	}
+
+	spin_lock(&obj->vma.lock);
+	for_each_ggtt_vma(vma, obj) {
+		vma->fence_size =
+			i915_gem_fence_size(i915, vma->size, tiling, stride);
+		vma->fence_alignment =
+			i915_gem_fence_alignment(i915,
+						 vma->size, tiling, stride);
+
+		if (vma->fence)
+			vma->fence->dirty = true;
+	}
+	spin_unlock(&obj->vma.lock);
+
+	obj->tiling_and_stride = tiling | stride;
+
+	/* Try to preallocate memory required to save swizzling on put-pages */
+	if (i915_gem_object_needs_bit17_swizzle(obj)) {
+		if (!obj->bit_17) {
+			obj->bit_17 = bitmap_zalloc(obj->base.size >> PAGE_SHIFT,
+						    GFP_KERNEL);
+		}
+	} else {
+		bitmap_free(obj->bit_17);
+		obj->bit_17 = NULL;
+	}
+
+	i915_gem_object_unlock(obj);
+
+	/* Force the fence to be reacquired for GTT access */
+	i915_gem_object_release_mmap_gtt(obj);
+
+	return 0;
+}
+
+/**
+ * i915_gem_set_tiling_ioctl - IOCTL handler to set tiling mode
+ * @dev: DRM device
+ * @data: data pointer for the ioctl
+ * @file: DRM file for the ioctl call
+ *
+ * Sets the tiling mode of an object, returning the required swizzling of
+ * bit 6 of addresses in the object.
+ *
+ * Called by the user via ioctl.
+ *
+ * Returns:
+ * Zero on success, negative errno on failure.
+ */
+int
+i915_gem_set_tiling_ioctl(struct drm_device *dev, void *data,
+			  struct drm_file *file)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_i915_gem_set_tiling *args = data;
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	if (!to_gt(dev_priv)->ggtt->num_fences)
+		return -EOPNOTSUPP;
+
+	obj = i915_gem_object_lookup(file, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	/*
+	 * The tiling mode of proxy objects is handled by its generator, and
+	 * not allowed to be changed by userspace.
+	 */
+	if (i915_gem_object_is_proxy(obj)) {
+		err = -ENXIO;
+		goto err;
+	}
+
+	if (!i915_tiling_ok(obj, args->tiling_mode, args->stride)) {
+		err = -EINVAL;
+		goto err;
+	}
+
+	if (args->tiling_mode == I915_TILING_NONE) {
+		args->swizzle_mode = I915_BIT_6_SWIZZLE_NONE;
+		args->stride = 0;
+	} else {
+		if (args->tiling_mode == I915_TILING_X)
+			args->swizzle_mode = to_gt(dev_priv)->ggtt->bit_6_swizzle_x;
+		else
+			args->swizzle_mode = to_gt(dev_priv)->ggtt->bit_6_swizzle_y;
+
+		/* Hide bit 17 swizzling from the user.  This prevents old Mesa
+		 * from aborting the application on sw fallbacks to bit 17,
+		 * and we use the pread/pwrite bit17 paths to swizzle for it.
+		 * If there was a user that was relying on the swizzle
+		 * information for drm_intel_bo_map()ed reads/writes this would
+		 * break it, but we don't have any of those.
+		 */
+		if (args->swizzle_mode == I915_BIT_6_SWIZZLE_9_17)
+			args->swizzle_mode = I915_BIT_6_SWIZZLE_9;
+		if (args->swizzle_mode == I915_BIT_6_SWIZZLE_9_10_17)
+			args->swizzle_mode = I915_BIT_6_SWIZZLE_9_10;
+
+		/* If we can't handle the swizzling, make it untiled. */
+		if (args->swizzle_mode == I915_BIT_6_SWIZZLE_UNKNOWN) {
+			args->tiling_mode = I915_TILING_NONE;
+			args->swizzle_mode = I915_BIT_6_SWIZZLE_NONE;
+			args->stride = 0;
+		}
+	}
+
+	err = i915_gem_object_set_tiling(obj, args->tiling_mode, args->stride);
+
+	/* We have to maintain this existing ABI... */
+	args->stride = i915_gem_object_get_stride(obj);
+	args->tiling_mode = i915_gem_object_get_tiling(obj);
+
+err:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+/**
+ * i915_gem_get_tiling_ioctl - IOCTL handler to get tiling mode
+ * @dev: DRM device
+ * @data: data pointer for the ioctl
+ * @file: DRM file for the ioctl call
+ *
+ * Returns the current tiling mode and required bit 6 swizzling for the object.
+ *
+ * Called by the user via ioctl.
+ *
+ * Returns:
+ * Zero on success, negative errno on failure.
+ */
+int
+i915_gem_get_tiling_ioctl(struct drm_device *dev, void *data,
+			  struct drm_file *file)
+{
+	struct drm_i915_gem_get_tiling *args = data;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_i915_gem_object *obj;
+	int err = -ENOENT;
+
+	if (!to_gt(dev_priv)->ggtt->num_fences)
+		return -EOPNOTSUPP;
+
+	rcu_read_lock();
+	obj = i915_gem_object_lookup_rcu(file, args->handle);
+	if (obj) {
+		args->tiling_mode =
+			READ_ONCE(obj->tiling_and_stride) & TILING_MASK;
+		err = 0;
+	}
+	rcu_read_unlock();
+	if (unlikely(err))
+		return err;
+
+	switch (args->tiling_mode) {
+	case I915_TILING_X:
+		args->swizzle_mode = to_gt(dev_priv)->ggtt->bit_6_swizzle_x;
+		break;
+	case I915_TILING_Y:
+		args->swizzle_mode = to_gt(dev_priv)->ggtt->bit_6_swizzle_y;
+		break;
+	default:
+	case I915_TILING_NONE:
+		args->swizzle_mode = I915_BIT_6_SWIZZLE_NONE;
+		break;
+	}
+
+	/* Hide bit 17 from the user -- see comment in i915_gem_set_tiling */
+	if (dev_priv->gem_quirks & GEM_QUIRK_PIN_SWIZZLED_PAGES)
+		args->phys_swizzle_mode = I915_BIT_6_SWIZZLE_UNKNOWN;
+	else
+		args->phys_swizzle_mode = args->swizzle_mode;
+	if (args->swizzle_mode == I915_BIT_6_SWIZZLE_9_17)
+		args->swizzle_mode = I915_BIT_6_SWIZZLE_9;
+	if (args->swizzle_mode == I915_BIT_6_SWIZZLE_9_10_17)
+		args->swizzle_mode = I915_BIT_6_SWIZZLE_9_10;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_tiling.h b/drivers/gpu/drm/i915/gem/i915_gem_tiling.h
new file mode 100644
index 000000000..6bd5751ab
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_tiling.h
@@ -0,0 +1,20 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __I915_GEM_TILING_H__
+#define __I915_GEM_TILING_H__
+
+#include <linux/types.h>
+
+struct drm_i915_gem_object;
+struct drm_i915_private;
+
+bool i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj);
+u32 i915_gem_fence_size(struct drm_i915_private *i915, u32 size,
+			unsigned int tiling, unsigned int stride);
+u32 i915_gem_fence_alignment(struct drm_i915_private *i915, u32 size,
+			     unsigned int tiling, unsigned int stride);
+
+#endif /* __I915_GEM_TILING_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
new file mode 100644
index 000000000..be4c081e7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
@@ -0,0 +1,1293 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/shmem_fs.h>
+
+#include <drm/ttm/ttm_bo_driver.h>
+#include <drm/ttm/ttm_placement.h>
+#include <drm/drm_buddy.h>
+
+#include "i915_drv.h"
+#include "i915_ttm_buddy_manager.h"
+#include "intel_memory_region.h"
+#include "intel_region_ttm.h"
+
+#include "gem/i915_gem_mman.h"
+#include "gem/i915_gem_object.h"
+#include "gem/i915_gem_region.h"
+#include "gem/i915_gem_ttm.h"
+#include "gem/i915_gem_ttm_move.h"
+#include "gem/i915_gem_ttm_pm.h"
+#include "gt/intel_gpu_commands.h"
+
+#define I915_TTM_PRIO_PURGE     0
+#define I915_TTM_PRIO_NO_PAGES  1
+#define I915_TTM_PRIO_HAS_PAGES 2
+#define I915_TTM_PRIO_NEEDS_CPU_ACCESS 3
+
+/*
+ * Size of struct ttm_place vector in on-stack struct ttm_placement allocs
+ */
+#define I915_TTM_MAX_PLACEMENTS INTEL_REGION_UNKNOWN
+
+/**
+ * struct i915_ttm_tt - TTM page vector with additional private information
+ * @ttm: The base TTM page vector.
+ * @dev: The struct device used for dma mapping and unmapping.
+ * @cached_rsgt: The cached scatter-gather table.
+ * @is_shmem: Set if using shmem.
+ * @filp: The shmem file, if using shmem backend.
+ *
+ * Note that DMA may be going on right up to the point where the page-
+ * vector is unpopulated in delayed destroy. Hence keep the
+ * scatter-gather table mapped and cached up to that point. This is
+ * different from the cached gem object io scatter-gather table which
+ * doesn't have an associated dma mapping.
+ */
+struct i915_ttm_tt {
+	struct ttm_tt ttm;
+	struct device *dev;
+	struct i915_refct_sgt cached_rsgt;
+
+	bool is_shmem;
+	struct file *filp;
+};
+
+static const struct ttm_place sys_placement_flags = {
+	.fpfn = 0,
+	.lpfn = 0,
+	.mem_type = I915_PL_SYSTEM,
+	.flags = 0,
+};
+
+static struct ttm_placement i915_sys_placement = {
+	.num_placement = 1,
+	.placement = &sys_placement_flags,
+	.num_busy_placement = 1,
+	.busy_placement = &sys_placement_flags,
+};
+
+/**
+ * i915_ttm_sys_placement - Return the struct ttm_placement to be
+ * used for an object in system memory.
+ *
+ * Rather than making the struct extern, use this
+ * function.
+ *
+ * Return: A pointer to a static variable for sys placement.
+ */
+struct ttm_placement *i915_ttm_sys_placement(void)
+{
+	return &i915_sys_placement;
+}
+
+static int i915_ttm_err_to_gem(int err)
+{
+	/* Fastpath */
+	if (likely(!err))
+		return 0;
+
+	switch (err) {
+	case -EBUSY:
+		/*
+		 * TTM likes to convert -EDEADLK to -EBUSY, and wants us to
+		 * restart the operation, since we don't record the contending
+		 * lock. We use -EAGAIN to restart.
+		 */
+		return -EAGAIN;
+	case -ENOSPC:
+		/*
+		 * Memory type / region is full, and we can't evict.
+		 * Except possibly system, that returns -ENOMEM;
+		 */
+		return -ENXIO;
+	default:
+		break;
+	}
+
+	return err;
+}
+
+static enum ttm_caching
+i915_ttm_select_tt_caching(const struct drm_i915_gem_object *obj)
+{
+	/*
+	 * Objects only allowed in system get cached cpu-mappings, or when
+	 * evicting lmem-only buffers to system for swapping. Other objects get
+	 * WC mapping for now. Even if in system.
+	 */
+	if (obj->mm.n_placements <= 1)
+		return ttm_cached;
+
+	return ttm_write_combined;
+}
+
+static void
+i915_ttm_place_from_region(const struct intel_memory_region *mr,
+			   struct ttm_place *place,
+			   resource_size_t offset,
+			   resource_size_t size,
+			   unsigned int flags)
+{
+	memset(place, 0, sizeof(*place));
+	place->mem_type = intel_region_to_ttm_type(mr);
+
+	if (mr->type == INTEL_MEMORY_SYSTEM)
+		return;
+
+	if (flags & I915_BO_ALLOC_CONTIGUOUS)
+		place->flags |= TTM_PL_FLAG_CONTIGUOUS;
+	if (offset != I915_BO_INVALID_OFFSET) {
+		place->fpfn = offset >> PAGE_SHIFT;
+		place->lpfn = place->fpfn + (size >> PAGE_SHIFT);
+	} else if (mr->io_size && mr->io_size < mr->total) {
+		if (flags & I915_BO_ALLOC_GPU_ONLY) {
+			place->flags |= TTM_PL_FLAG_TOPDOWN;
+		} else {
+			place->fpfn = 0;
+			place->lpfn = mr->io_size >> PAGE_SHIFT;
+		}
+	}
+}
+
+static void
+i915_ttm_placement_from_obj(const struct drm_i915_gem_object *obj,
+			    struct ttm_place *requested,
+			    struct ttm_place *busy,
+			    struct ttm_placement *placement)
+{
+	unsigned int num_allowed = obj->mm.n_placements;
+	unsigned int flags = obj->flags;
+	unsigned int i;
+
+	placement->num_placement = 1;
+	i915_ttm_place_from_region(num_allowed ? obj->mm.placements[0] :
+				   obj->mm.region, requested, obj->bo_offset,
+				   obj->base.size, flags);
+
+	/* Cache this on object? */
+	placement->num_busy_placement = num_allowed;
+	for (i = 0; i < placement->num_busy_placement; ++i)
+		i915_ttm_place_from_region(obj->mm.placements[i], busy + i,
+					   obj->bo_offset, obj->base.size, flags);
+
+	if (num_allowed == 0) {
+		*busy = *requested;
+		placement->num_busy_placement = 1;
+	}
+
+	placement->placement = requested;
+	placement->busy_placement = busy;
+}
+
+static int i915_ttm_tt_shmem_populate(struct ttm_device *bdev,
+				      struct ttm_tt *ttm,
+				      struct ttm_operation_ctx *ctx)
+{
+	struct drm_i915_private *i915 = container_of(bdev, typeof(*i915), bdev);
+	struct intel_memory_region *mr = i915->mm.regions[INTEL_MEMORY_SYSTEM];
+	struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm);
+	const unsigned int max_segment = i915_sg_segment_size(i915->drm.dev);
+	const size_t size = (size_t)ttm->num_pages << PAGE_SHIFT;
+	struct file *filp = i915_tt->filp;
+	struct sgt_iter sgt_iter;
+	struct sg_table *st;
+	struct page *page;
+	unsigned long i;
+	int err;
+
+	if (!filp) {
+		struct address_space *mapping;
+		gfp_t mask;
+
+		filp = shmem_file_setup("i915-shmem-tt", size, VM_NORESERVE);
+		if (IS_ERR(filp))
+			return PTR_ERR(filp);
+
+		mask = GFP_HIGHUSER | __GFP_RECLAIMABLE;
+
+		mapping = filp->f_mapping;
+		mapping_set_gfp_mask(mapping, mask);
+		GEM_BUG_ON(!(mapping_gfp_mask(mapping) & __GFP_RECLAIM));
+
+		i915_tt->filp = filp;
+	}
+
+	st = &i915_tt->cached_rsgt.table;
+	err = shmem_sg_alloc_table(i915, st, size, mr, filp->f_mapping,
+				   max_segment);
+	if (err)
+		return err;
+
+	err = dma_map_sgtable(i915_tt->dev, st, DMA_BIDIRECTIONAL,
+			      DMA_ATTR_SKIP_CPU_SYNC);
+	if (err)
+		goto err_free_st;
+
+	i = 0;
+	for_each_sgt_page(page, sgt_iter, st)
+		ttm->pages[i++] = page;
+
+	if (ttm->page_flags & TTM_TT_FLAG_SWAPPED)
+		ttm->page_flags &= ~TTM_TT_FLAG_SWAPPED;
+
+	return 0;
+
+err_free_st:
+	shmem_sg_free_table(st, filp->f_mapping, false, false);
+
+	return err;
+}
+
+static void i915_ttm_tt_shmem_unpopulate(struct ttm_tt *ttm)
+{
+	struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm);
+	bool backup = ttm->page_flags & TTM_TT_FLAG_SWAPPED;
+	struct sg_table *st = &i915_tt->cached_rsgt.table;
+
+	shmem_sg_free_table(st, file_inode(i915_tt->filp)->i_mapping,
+			    backup, backup);
+}
+
+static void i915_ttm_tt_release(struct kref *ref)
+{
+	struct i915_ttm_tt *i915_tt =
+		container_of(ref, typeof(*i915_tt), cached_rsgt.kref);
+	struct sg_table *st = &i915_tt->cached_rsgt.table;
+
+	GEM_WARN_ON(st->sgl);
+
+	kfree(i915_tt);
+}
+
+static const struct i915_refct_sgt_ops tt_rsgt_ops = {
+	.release = i915_ttm_tt_release
+};
+
+static struct ttm_tt *i915_ttm_tt_create(struct ttm_buffer_object *bo,
+					 uint32_t page_flags)
+{
+	struct drm_i915_private *i915 = container_of(bo->bdev, typeof(*i915),
+						     bdev);
+	struct ttm_resource_manager *man =
+		ttm_manager_type(bo->bdev, bo->resource->mem_type);
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+	unsigned long ccs_pages = 0;
+	enum ttm_caching caching;
+	struct i915_ttm_tt *i915_tt;
+	int ret;
+
+	if (i915_ttm_is_ghost_object(bo))
+		return NULL;
+
+	i915_tt = kzalloc(sizeof(*i915_tt), GFP_KERNEL);
+	if (!i915_tt)
+		return NULL;
+
+	if (obj->flags & I915_BO_ALLOC_CPU_CLEAR &&
+	    man->use_tt)
+		page_flags |= TTM_TT_FLAG_ZERO_ALLOC;
+
+	caching = i915_ttm_select_tt_caching(obj);
+	if (i915_gem_object_is_shrinkable(obj) && caching == ttm_cached) {
+		page_flags |= TTM_TT_FLAG_EXTERNAL |
+			      TTM_TT_FLAG_EXTERNAL_MAPPABLE;
+		i915_tt->is_shmem = true;
+	}
+
+	if (i915_gem_object_needs_ccs_pages(obj))
+		ccs_pages = DIV_ROUND_UP(DIV_ROUND_UP(bo->base.size,
+						      NUM_BYTES_PER_CCS_BYTE),
+					 PAGE_SIZE);
+
+	ret = ttm_tt_init(&i915_tt->ttm, bo, page_flags, caching, ccs_pages);
+	if (ret)
+		goto err_free;
+
+	__i915_refct_sgt_init(&i915_tt->cached_rsgt, bo->base.size,
+			      &tt_rsgt_ops);
+
+	i915_tt->dev = obj->base.dev->dev;
+
+	return &i915_tt->ttm;
+
+err_free:
+	kfree(i915_tt);
+	return NULL;
+}
+
+static int i915_ttm_tt_populate(struct ttm_device *bdev,
+				struct ttm_tt *ttm,
+				struct ttm_operation_ctx *ctx)
+{
+	struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm);
+
+	if (i915_tt->is_shmem)
+		return i915_ttm_tt_shmem_populate(bdev, ttm, ctx);
+
+	return ttm_pool_alloc(&bdev->pool, ttm, ctx);
+}
+
+static void i915_ttm_tt_unpopulate(struct ttm_device *bdev, struct ttm_tt *ttm)
+{
+	struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm);
+	struct sg_table *st = &i915_tt->cached_rsgt.table;
+
+	if (st->sgl)
+		dma_unmap_sgtable(i915_tt->dev, st, DMA_BIDIRECTIONAL, 0);
+
+	if (i915_tt->is_shmem) {
+		i915_ttm_tt_shmem_unpopulate(ttm);
+	} else {
+		sg_free_table(st);
+		ttm_pool_free(&bdev->pool, ttm);
+	}
+}
+
+static void i915_ttm_tt_destroy(struct ttm_device *bdev, struct ttm_tt *ttm)
+{
+	struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm);
+
+	if (i915_tt->filp)
+		fput(i915_tt->filp);
+
+	ttm_tt_fini(ttm);
+	i915_refct_sgt_put(&i915_tt->cached_rsgt);
+}
+
+static bool i915_ttm_eviction_valuable(struct ttm_buffer_object *bo,
+				       const struct ttm_place *place)
+{
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+
+	if (i915_ttm_is_ghost_object(bo))
+		return false;
+
+	/*
+	 * EXTERNAL objects should never be swapped out by TTM, instead we need
+	 * to handle that ourselves. TTM will already skip such objects for us,
+	 * but we would like to avoid grabbing locks for no good reason.
+	 */
+	if (bo->ttm && bo->ttm->page_flags & TTM_TT_FLAG_EXTERNAL)
+		return false;
+
+	/* Will do for now. Our pinned objects are still on TTM's LRU lists */
+	if (!i915_gem_object_evictable(obj))
+		return false;
+
+	return ttm_bo_eviction_valuable(bo, place);
+}
+
+static void i915_ttm_evict_flags(struct ttm_buffer_object *bo,
+				 struct ttm_placement *placement)
+{
+	*placement = i915_sys_placement;
+}
+
+/**
+ * i915_ttm_free_cached_io_rsgt - Free object cached LMEM information
+ * @obj: The GEM object
+ * This function frees any LMEM-related information that is cached on
+ * the object. For example the radix tree for fast page lookup and the
+ * cached refcounted sg-table
+ */
+void i915_ttm_free_cached_io_rsgt(struct drm_i915_gem_object *obj)
+{
+	struct radix_tree_iter iter;
+	void __rcu **slot;
+
+	if (!obj->ttm.cached_io_rsgt)
+		return;
+
+	rcu_read_lock();
+	radix_tree_for_each_slot(slot, &obj->ttm.get_io_page.radix, &iter, 0)
+		radix_tree_delete(&obj->ttm.get_io_page.radix, iter.index);
+	rcu_read_unlock();
+
+	i915_refct_sgt_put(obj->ttm.cached_io_rsgt);
+	obj->ttm.cached_io_rsgt = NULL;
+}
+
+/**
+ * i915_ttm_purge - Clear an object of its memory
+ * @obj: The object
+ *
+ * This function is called to clear an object of it's memory when it is
+ * marked as not needed anymore.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int i915_ttm_purge(struct drm_i915_gem_object *obj)
+{
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+	struct i915_ttm_tt *i915_tt =
+		container_of(bo->ttm, typeof(*i915_tt), ttm);
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false,
+	};
+	struct ttm_placement place = {};
+	int ret;
+
+	if (obj->mm.madv == __I915_MADV_PURGED)
+		return 0;
+
+	ret = ttm_bo_validate(bo, &place, &ctx);
+	if (ret)
+		return ret;
+
+	if (bo->ttm && i915_tt->filp) {
+		/*
+		 * The below fput(which eventually calls shmem_truncate) might
+		 * be delayed by worker, so when directly called to purge the
+		 * pages(like by the shrinker) we should try to be more
+		 * aggressive and release the pages immediately.
+		 */
+		shmem_truncate_range(file_inode(i915_tt->filp),
+				     0, (loff_t)-1);
+		fput(fetch_and_zero(&i915_tt->filp));
+	}
+
+	obj->write_domain = 0;
+	obj->read_domains = 0;
+	i915_ttm_adjust_gem_after_move(obj);
+	i915_ttm_free_cached_io_rsgt(obj);
+	obj->mm.madv = __I915_MADV_PURGED;
+
+	return 0;
+}
+
+static int i915_ttm_shrink(struct drm_i915_gem_object *obj, unsigned int flags)
+{
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+	struct i915_ttm_tt *i915_tt =
+		container_of(bo->ttm, typeof(*i915_tt), ttm);
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = flags & I915_GEM_OBJECT_SHRINK_NO_GPU_WAIT,
+	};
+	struct ttm_placement place = {};
+	int ret;
+
+	if (!bo->ttm || bo->resource->mem_type != TTM_PL_SYSTEM)
+		return 0;
+
+	GEM_BUG_ON(!i915_tt->is_shmem);
+
+	if (!i915_tt->filp)
+		return 0;
+
+	ret = ttm_bo_wait_ctx(bo, &ctx);
+	if (ret)
+		return ret;
+
+	switch (obj->mm.madv) {
+	case I915_MADV_DONTNEED:
+		return i915_ttm_purge(obj);
+	case __I915_MADV_PURGED:
+		return 0;
+	}
+
+	if (bo->ttm->page_flags & TTM_TT_FLAG_SWAPPED)
+		return 0;
+
+	bo->ttm->page_flags |= TTM_TT_FLAG_SWAPPED;
+	ret = ttm_bo_validate(bo, &place, &ctx);
+	if (ret) {
+		bo->ttm->page_flags &= ~TTM_TT_FLAG_SWAPPED;
+		return ret;
+	}
+
+	if (flags & I915_GEM_OBJECT_SHRINK_WRITEBACK)
+		__shmem_writeback(obj->base.size, i915_tt->filp->f_mapping);
+
+	return 0;
+}
+
+static void i915_ttm_delete_mem_notify(struct ttm_buffer_object *bo)
+{
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+
+	if (bo->resource && !i915_ttm_is_ghost_object(bo)) {
+		__i915_gem_object_pages_fini(obj);
+		i915_ttm_free_cached_io_rsgt(obj);
+	}
+}
+
+static struct i915_refct_sgt *i915_ttm_tt_get_st(struct ttm_tt *ttm)
+{
+	struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm);
+	struct sg_table *st;
+	int ret;
+
+	if (i915_tt->cached_rsgt.table.sgl)
+		return i915_refct_sgt_get(&i915_tt->cached_rsgt);
+
+	st = &i915_tt->cached_rsgt.table;
+	ret = sg_alloc_table_from_pages_segment(st,
+			ttm->pages, ttm->num_pages,
+			0, (unsigned long)ttm->num_pages << PAGE_SHIFT,
+			i915_sg_segment_size(i915_tt->dev), GFP_KERNEL);
+	if (ret) {
+		st->sgl = NULL;
+		return ERR_PTR(ret);
+	}
+
+	ret = dma_map_sgtable(i915_tt->dev, st, DMA_BIDIRECTIONAL, 0);
+	if (ret) {
+		sg_free_table(st);
+		return ERR_PTR(ret);
+	}
+
+	return i915_refct_sgt_get(&i915_tt->cached_rsgt);
+}
+
+/**
+ * i915_ttm_resource_get_st - Get a refcounted sg-table pointing to the
+ * resource memory
+ * @obj: The GEM object used for sg-table caching
+ * @res: The struct ttm_resource for which an sg-table is requested.
+ *
+ * This function returns a refcounted sg-table representing the memory
+ * pointed to by @res. If @res is the object's current resource it may also
+ * cache the sg_table on the object or attempt to access an already cached
+ * sg-table. The refcounted sg-table needs to be put when no-longer in use.
+ *
+ * Return: A valid pointer to a struct i915_refct_sgt or error pointer on
+ * failure.
+ */
+struct i915_refct_sgt *
+i915_ttm_resource_get_st(struct drm_i915_gem_object *obj,
+			 struct ttm_resource *res)
+{
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+	u32 page_alignment;
+
+	if (!i915_ttm_gtt_binds_lmem(res))
+		return i915_ttm_tt_get_st(bo->ttm);
+
+	page_alignment = bo->page_alignment << PAGE_SHIFT;
+	if (!page_alignment)
+		page_alignment = obj->mm.region->min_page_size;
+
+	/*
+	 * If CPU mapping differs, we need to add the ttm_tt pages to
+	 * the resulting st. Might make sense for GGTT.
+	 */
+	GEM_WARN_ON(!i915_ttm_cpu_maps_iomem(res));
+	if (bo->resource == res) {
+		if (!obj->ttm.cached_io_rsgt) {
+			struct i915_refct_sgt *rsgt;
+
+			rsgt = intel_region_ttm_resource_to_rsgt(obj->mm.region,
+								 res,
+								 page_alignment);
+			if (IS_ERR(rsgt))
+				return rsgt;
+
+			obj->ttm.cached_io_rsgt = rsgt;
+		}
+		return i915_refct_sgt_get(obj->ttm.cached_io_rsgt);
+	}
+
+	return intel_region_ttm_resource_to_rsgt(obj->mm.region, res,
+						 page_alignment);
+}
+
+static int i915_ttm_truncate(struct drm_i915_gem_object *obj)
+{
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+	int err;
+
+	WARN_ON_ONCE(obj->mm.madv == I915_MADV_WILLNEED);
+
+	err = ttm_bo_wait(bo, true, false);
+	if (err)
+		return err;
+
+	err = i915_ttm_move_notify(bo);
+	if (err)
+		return err;
+
+	return i915_ttm_purge(obj);
+}
+
+static void i915_ttm_swap_notify(struct ttm_buffer_object *bo)
+{
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+	int ret;
+
+	if (i915_ttm_is_ghost_object(bo))
+		return;
+
+	ret = i915_ttm_move_notify(bo);
+	GEM_WARN_ON(ret);
+	GEM_WARN_ON(obj->ttm.cached_io_rsgt);
+	if (!ret && obj->mm.madv != I915_MADV_WILLNEED)
+		i915_ttm_purge(obj);
+}
+
+/**
+ * i915_ttm_resource_mappable - Return true if the ttm resource is CPU
+ * accessible.
+ * @res: The TTM resource to check.
+ *
+ * This is interesting on small-BAR systems where we may encounter lmem objects
+ * that can't be accessed via the CPU.
+ */
+bool i915_ttm_resource_mappable(struct ttm_resource *res)
+{
+	struct i915_ttm_buddy_resource *bman_res = to_ttm_buddy_resource(res);
+
+	if (!i915_ttm_cpu_maps_iomem(res))
+		return true;
+
+	return bman_res->used_visible_size == bman_res->base.num_pages;
+}
+
+static int i915_ttm_io_mem_reserve(struct ttm_device *bdev, struct ttm_resource *mem)
+{
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(mem->bo);
+	bool unknown_state;
+
+	if (i915_ttm_is_ghost_object(mem->bo))
+		return -EINVAL;
+
+	if (!kref_get_unless_zero(&obj->base.refcount))
+		return -EINVAL;
+
+	assert_object_held(obj);
+
+	unknown_state = i915_gem_object_has_unknown_state(obj);
+	i915_gem_object_put(obj);
+	if (unknown_state)
+		return -EINVAL;
+
+	if (!i915_ttm_cpu_maps_iomem(mem))
+		return 0;
+
+	if (!i915_ttm_resource_mappable(mem))
+		return -EINVAL;
+
+	mem->bus.caching = ttm_write_combined;
+	mem->bus.is_iomem = true;
+
+	return 0;
+}
+
+static unsigned long i915_ttm_io_mem_pfn(struct ttm_buffer_object *bo,
+					 unsigned long page_offset)
+{
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+	struct scatterlist *sg;
+	unsigned long base;
+	unsigned int ofs;
+
+	GEM_BUG_ON(i915_ttm_is_ghost_object(bo));
+	GEM_WARN_ON(bo->ttm);
+
+	base = obj->mm.region->iomap.base - obj->mm.region->region.start;
+	sg = __i915_gem_object_get_sg(obj, &obj->ttm.get_io_page, page_offset, &ofs, true);
+
+	return ((base + sg_dma_address(sg)) >> PAGE_SHIFT) + ofs;
+}
+
+/*
+ * All callbacks need to take care not to downcast a struct ttm_buffer_object
+ * without checking its subclass, since it might be a TTM ghost object.
+ */
+static struct ttm_device_funcs i915_ttm_bo_driver = {
+	.ttm_tt_create = i915_ttm_tt_create,
+	.ttm_tt_populate = i915_ttm_tt_populate,
+	.ttm_tt_unpopulate = i915_ttm_tt_unpopulate,
+	.ttm_tt_destroy = i915_ttm_tt_destroy,
+	.eviction_valuable = i915_ttm_eviction_valuable,
+	.evict_flags = i915_ttm_evict_flags,
+	.move = i915_ttm_move,
+	.swap_notify = i915_ttm_swap_notify,
+	.delete_mem_notify = i915_ttm_delete_mem_notify,
+	.io_mem_reserve = i915_ttm_io_mem_reserve,
+	.io_mem_pfn = i915_ttm_io_mem_pfn,
+};
+
+/**
+ * i915_ttm_driver - Return a pointer to the TTM device funcs
+ *
+ * Return: Pointer to statically allocated TTM device funcs.
+ */
+struct ttm_device_funcs *i915_ttm_driver(void)
+{
+	return &i915_ttm_bo_driver;
+}
+
+static int __i915_ttm_get_pages(struct drm_i915_gem_object *obj,
+				struct ttm_placement *placement)
+{
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false,
+	};
+	int real_num_busy;
+	int ret;
+
+	/* First try only the requested placement. No eviction. */
+	real_num_busy = fetch_and_zero(&placement->num_busy_placement);
+	ret = ttm_bo_validate(bo, placement, &ctx);
+	if (ret) {
+		ret = i915_ttm_err_to_gem(ret);
+		/*
+		 * Anything that wants to restart the operation gets to
+		 * do that.
+		 */
+		if (ret == -EDEADLK || ret == -EINTR || ret == -ERESTARTSYS ||
+		    ret == -EAGAIN)
+			return ret;
+
+		/*
+		 * If the initial attempt fails, allow all accepted placements,
+		 * evicting if necessary.
+		 */
+		placement->num_busy_placement = real_num_busy;
+		ret = ttm_bo_validate(bo, placement, &ctx);
+		if (ret)
+			return i915_ttm_err_to_gem(ret);
+	}
+
+	if (bo->ttm && !ttm_tt_is_populated(bo->ttm)) {
+		ret = ttm_tt_populate(bo->bdev, bo->ttm, &ctx);
+		if (ret)
+			return ret;
+
+		i915_ttm_adjust_domains_after_move(obj);
+		i915_ttm_adjust_gem_after_move(obj);
+	}
+
+	if (!i915_gem_object_has_pages(obj)) {
+		struct i915_refct_sgt *rsgt =
+			i915_ttm_resource_get_st(obj, bo->resource);
+
+		if (IS_ERR(rsgt))
+			return PTR_ERR(rsgt);
+
+		GEM_BUG_ON(obj->mm.rsgt);
+		obj->mm.rsgt = rsgt;
+		__i915_gem_object_set_pages(obj, &rsgt->table,
+					    i915_sg_dma_sizes(rsgt->table.sgl));
+	}
+
+	GEM_BUG_ON(bo->ttm && ((obj->base.size >> PAGE_SHIFT) < bo->ttm->num_pages));
+	i915_ttm_adjust_lru(obj);
+	return ret;
+}
+
+static int i915_ttm_get_pages(struct drm_i915_gem_object *obj)
+{
+	struct ttm_place requested, busy[I915_TTM_MAX_PLACEMENTS];
+	struct ttm_placement placement;
+
+	GEM_BUG_ON(obj->mm.n_placements > I915_TTM_MAX_PLACEMENTS);
+
+	/* Move to the requested placement. */
+	i915_ttm_placement_from_obj(obj, &requested, busy, &placement);
+
+	return __i915_ttm_get_pages(obj, &placement);
+}
+
+/**
+ * DOC: Migration vs eviction
+ *
+ * GEM migration may not be the same as TTM migration / eviction. If
+ * the TTM core decides to evict an object it may be evicted to a
+ * TTM memory type that is not in the object's allowable GEM regions, or
+ * in fact theoretically to a TTM memory type that doesn't correspond to
+ * a GEM memory region. In that case the object's GEM region is not
+ * updated, and the data is migrated back to the GEM region at
+ * get_pages time. TTM may however set up CPU ptes to the object even
+ * when it is evicted.
+ * Gem forced migration using the i915_ttm_migrate() op, is allowed even
+ * to regions that are not in the object's list of allowable placements.
+ */
+static int __i915_ttm_migrate(struct drm_i915_gem_object *obj,
+			      struct intel_memory_region *mr,
+			      unsigned int flags)
+{
+	struct ttm_place requested;
+	struct ttm_placement placement;
+	int ret;
+
+	i915_ttm_place_from_region(mr, &requested, obj->bo_offset,
+				   obj->base.size, flags);
+	placement.num_placement = 1;
+	placement.num_busy_placement = 1;
+	placement.placement = &requested;
+	placement.busy_placement = &requested;
+
+	ret = __i915_ttm_get_pages(obj, &placement);
+	if (ret)
+		return ret;
+
+	/*
+	 * Reinitialize the region bindings. This is primarily
+	 * required for objects where the new region is not in
+	 * its allowable placements.
+	 */
+	if (obj->mm.region != mr) {
+		i915_gem_object_release_memory_region(obj);
+		i915_gem_object_init_memory_region(obj, mr);
+	}
+
+	return 0;
+}
+
+static int i915_ttm_migrate(struct drm_i915_gem_object *obj,
+			    struct intel_memory_region *mr,
+			    unsigned int flags)
+{
+	return __i915_ttm_migrate(obj, mr, flags);
+}
+
+static void i915_ttm_put_pages(struct drm_i915_gem_object *obj,
+			       struct sg_table *st)
+{
+	/*
+	 * We're currently not called from a shrinker, so put_pages()
+	 * typically means the object is about to destroyed, or called
+	 * from move_notify(). So just avoid doing much for now.
+	 * If the object is not destroyed next, The TTM eviction logic
+	 * and shrinkers will move it out if needed.
+	 */
+
+	if (obj->mm.rsgt)
+		i915_refct_sgt_put(fetch_and_zero(&obj->mm.rsgt));
+}
+
+/**
+ * i915_ttm_adjust_lru - Adjust an object's position on relevant LRU lists.
+ * @obj: The object
+ */
+void i915_ttm_adjust_lru(struct drm_i915_gem_object *obj)
+{
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+	struct i915_ttm_tt *i915_tt =
+		container_of(bo->ttm, typeof(*i915_tt), ttm);
+	bool shrinkable =
+		bo->ttm && i915_tt->filp && ttm_tt_is_populated(bo->ttm);
+
+	/*
+	 * Don't manipulate the TTM LRUs while in TTM bo destruction.
+	 * We're called through i915_ttm_delete_mem_notify().
+	 */
+	if (!kref_read(&bo->kref))
+		return;
+
+	/*
+	 * We skip managing the shrinker LRU in set_pages() and just manage
+	 * everything here. This does at least solve the issue with having
+	 * temporary shmem mappings(like with evicted lmem) not being visible to
+	 * the shrinker. Only our shmem objects are shrinkable, everything else
+	 * we keep as unshrinkable.
+	 *
+	 * To make sure everything plays nice we keep an extra shrink pin in TTM
+	 * if the underlying pages are not currently shrinkable. Once we release
+	 * our pin, like when the pages are moved to shmem, the pages will then
+	 * be added to the shrinker LRU, assuming the caller isn't also holding
+	 * a pin.
+	 *
+	 * TODO: consider maybe also bumping the shrinker list here when we have
+	 * already unpinned it, which should give us something more like an LRU.
+	 *
+	 * TODO: There is a small window of opportunity for this function to
+	 * get called from eviction after we've dropped the last GEM refcount,
+	 * but before the TTM deleted flag is set on the object. Avoid
+	 * adjusting the shrinker list in such cases, since the object is
+	 * not available to the shrinker anyway due to its zero refcount.
+	 * To fix this properly we should move to a TTM shrinker LRU list for
+	 * these objects.
+	 */
+	if (kref_get_unless_zero(&obj->base.refcount)) {
+		if (shrinkable != obj->mm.ttm_shrinkable) {
+			if (shrinkable) {
+				if (obj->mm.madv == I915_MADV_WILLNEED)
+					__i915_gem_object_make_shrinkable(obj);
+				else
+					__i915_gem_object_make_purgeable(obj);
+			} else {
+				i915_gem_object_make_unshrinkable(obj);
+			}
+
+			obj->mm.ttm_shrinkable = shrinkable;
+		}
+		i915_gem_object_put(obj);
+	}
+
+	/*
+	 * Put on the correct LRU list depending on the MADV status
+	 */
+	spin_lock(&bo->bdev->lru_lock);
+	if (shrinkable) {
+		/* Try to keep shmem_tt from being considered for shrinking. */
+		bo->priority = TTM_MAX_BO_PRIORITY - 1;
+	} else if (obj->mm.madv != I915_MADV_WILLNEED) {
+		bo->priority = I915_TTM_PRIO_PURGE;
+	} else if (!i915_gem_object_has_pages(obj)) {
+		bo->priority = I915_TTM_PRIO_NO_PAGES;
+	} else {
+		struct ttm_resource_manager *man =
+			ttm_manager_type(bo->bdev, bo->resource->mem_type);
+
+		/*
+		 * If we need to place an LMEM resource which doesn't need CPU
+		 * access then we should try not to victimize mappable objects
+		 * first, since we likely end up stealing more of the mappable
+		 * portion. And likewise when we try to find space for a mappble
+		 * object, we know not to ever victimize objects that don't
+		 * occupy any mappable pages.
+		 */
+		if (i915_ttm_cpu_maps_iomem(bo->resource) &&
+		    i915_ttm_buddy_man_visible_size(man) < man->size &&
+		    !(obj->flags & I915_BO_ALLOC_GPU_ONLY))
+			bo->priority = I915_TTM_PRIO_NEEDS_CPU_ACCESS;
+		else
+			bo->priority = I915_TTM_PRIO_HAS_PAGES;
+	}
+
+	ttm_bo_move_to_lru_tail(bo);
+	spin_unlock(&bo->bdev->lru_lock);
+}
+
+/*
+ * TTM-backed gem object destruction requires some clarification.
+ * Basically we have two possibilities here. We can either rely on the
+ * i915 delayed destruction and put the TTM object when the object
+ * is idle. This would be detected by TTM which would bypass the
+ * TTM delayed destroy handling. The other approach is to put the TTM
+ * object early and rely on the TTM destroyed handling, and then free
+ * the leftover parts of the GEM object once TTM's destroyed list handling is
+ * complete. For now, we rely on the latter for two reasons:
+ * a) TTM can evict an object even when it's on the delayed destroy list,
+ * which in theory allows for complete eviction.
+ * b) There is work going on in TTM to allow freeing an object even when
+ * it's not idle, and using the TTM destroyed list handling could help us
+ * benefit from that.
+ */
+static void i915_ttm_delayed_free(struct drm_i915_gem_object *obj)
+{
+	GEM_BUG_ON(!obj->ttm.created);
+
+	ttm_bo_put(i915_gem_to_ttm(obj));
+}
+
+static vm_fault_t vm_fault_ttm(struct vm_fault *vmf)
+{
+	struct vm_area_struct *area = vmf->vma;
+	struct ttm_buffer_object *bo = area->vm_private_data;
+	struct drm_device *dev = bo->base.dev;
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+	intel_wakeref_t wakeref = 0;
+	vm_fault_t ret;
+	int idx;
+
+	if (i915_ttm_is_ghost_object(bo))
+		return VM_FAULT_SIGBUS;
+
+	/* Sanity check that we allow writing into this object */
+	if (unlikely(i915_gem_object_is_readonly(obj) &&
+		     area->vm_flags & VM_WRITE))
+		return VM_FAULT_SIGBUS;
+
+	ret = ttm_bo_vm_reserve(bo, vmf);
+	if (ret)
+		return ret;
+
+	if (obj->mm.madv != I915_MADV_WILLNEED) {
+		dma_resv_unlock(bo->base.resv);
+		return VM_FAULT_SIGBUS;
+	}
+
+	if (!i915_ttm_resource_mappable(bo->resource)) {
+		int err = -ENODEV;
+		int i;
+
+		for (i = 0; i < obj->mm.n_placements; i++) {
+			struct intel_memory_region *mr = obj->mm.placements[i];
+			unsigned int flags;
+
+			if (!mr->io_size && mr->type != INTEL_MEMORY_SYSTEM)
+				continue;
+
+			flags = obj->flags;
+			flags &= ~I915_BO_ALLOC_GPU_ONLY;
+			err = __i915_ttm_migrate(obj, mr, flags);
+			if (!err)
+				break;
+		}
+
+		if (err) {
+			drm_dbg(dev, "Unable to make resource CPU accessible\n");
+			dma_resv_unlock(bo->base.resv);
+			ret = VM_FAULT_SIGBUS;
+			goto out_rpm;
+		}
+	}
+
+	if (i915_ttm_cpu_maps_iomem(bo->resource))
+		wakeref = intel_runtime_pm_get(&to_i915(obj->base.dev)->runtime_pm);
+
+	if (drm_dev_enter(dev, &idx)) {
+		ret = ttm_bo_vm_fault_reserved(vmf, vmf->vma->vm_page_prot,
+					       TTM_BO_VM_NUM_PREFAULT);
+		drm_dev_exit(idx);
+	} else {
+		ret = ttm_bo_vm_dummy_page(vmf, vmf->vma->vm_page_prot);
+	}
+
+	if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
+		goto out_rpm;
+
+	/*
+	 * ttm_bo_vm_reserve() already has dma_resv_lock.
+	 * userfault_count is protected by dma_resv lock and rpm wakeref.
+	 */
+	if (ret == VM_FAULT_NOPAGE && wakeref && !obj->userfault_count) {
+		obj->userfault_count = 1;
+		spin_lock(&to_i915(obj->base.dev)->runtime_pm.lmem_userfault_lock);
+		list_add(&obj->userfault_link, &to_i915(obj->base.dev)->runtime_pm.lmem_userfault_list);
+		spin_unlock(&to_i915(obj->base.dev)->runtime_pm.lmem_userfault_lock);
+	}
+
+	if (wakeref & CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND)
+		intel_wakeref_auto(&to_i915(obj->base.dev)->runtime_pm.userfault_wakeref,
+				   msecs_to_jiffies_timeout(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND));
+
+	i915_ttm_adjust_lru(obj);
+
+	dma_resv_unlock(bo->base.resv);
+
+out_rpm:
+	if (wakeref)
+		intel_runtime_pm_put(&to_i915(obj->base.dev)->runtime_pm, wakeref);
+
+	return ret;
+}
+
+static int
+vm_access_ttm(struct vm_area_struct *area, unsigned long addr,
+	      void *buf, int len, int write)
+{
+	struct drm_i915_gem_object *obj =
+		i915_ttm_to_gem(area->vm_private_data);
+
+	if (i915_gem_object_is_readonly(obj) && write)
+		return -EACCES;
+
+	return ttm_bo_vm_access(area, addr, buf, len, write);
+}
+
+static void ttm_vm_open(struct vm_area_struct *vma)
+{
+	struct drm_i915_gem_object *obj =
+		i915_ttm_to_gem(vma->vm_private_data);
+
+	GEM_BUG_ON(i915_ttm_is_ghost_object(vma->vm_private_data));
+	i915_gem_object_get(obj);
+}
+
+static void ttm_vm_close(struct vm_area_struct *vma)
+{
+	struct drm_i915_gem_object *obj =
+		i915_ttm_to_gem(vma->vm_private_data);
+
+	GEM_BUG_ON(i915_ttm_is_ghost_object(vma->vm_private_data));
+	i915_gem_object_put(obj);
+}
+
+static const struct vm_operations_struct vm_ops_ttm = {
+	.fault = vm_fault_ttm,
+	.access = vm_access_ttm,
+	.open = ttm_vm_open,
+	.close = ttm_vm_close,
+};
+
+static u64 i915_ttm_mmap_offset(struct drm_i915_gem_object *obj)
+{
+	/* The ttm_bo must be allocated with I915_BO_ALLOC_USER */
+	GEM_BUG_ON(!drm_mm_node_allocated(&obj->base.vma_node.vm_node));
+
+	return drm_vma_node_offset_addr(&obj->base.vma_node);
+}
+
+static void i915_ttm_unmap_virtual(struct drm_i915_gem_object *obj)
+{
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+	intel_wakeref_t wakeref = 0;
+
+	assert_object_held_shared(obj);
+
+	if (i915_ttm_cpu_maps_iomem(bo->resource)) {
+		wakeref = intel_runtime_pm_get(&to_i915(obj->base.dev)->runtime_pm);
+
+		/* userfault_count is protected by obj lock and rpm wakeref. */
+		if (obj->userfault_count) {
+			spin_lock(&to_i915(obj->base.dev)->runtime_pm.lmem_userfault_lock);
+			list_del(&obj->userfault_link);
+			spin_unlock(&to_i915(obj->base.dev)->runtime_pm.lmem_userfault_lock);
+			obj->userfault_count = 0;
+		}
+	}
+
+	ttm_bo_unmap_virtual(i915_gem_to_ttm(obj));
+
+	if (wakeref)
+		intel_runtime_pm_put(&to_i915(obj->base.dev)->runtime_pm, wakeref);
+}
+
+static const struct drm_i915_gem_object_ops i915_gem_ttm_obj_ops = {
+	.name = "i915_gem_object_ttm",
+	.flags = I915_GEM_OBJECT_IS_SHRINKABLE |
+		 I915_GEM_OBJECT_SELF_MANAGED_SHRINK_LIST,
+
+	.get_pages = i915_ttm_get_pages,
+	.put_pages = i915_ttm_put_pages,
+	.truncate = i915_ttm_truncate,
+	.shrink = i915_ttm_shrink,
+
+	.adjust_lru = i915_ttm_adjust_lru,
+	.delayed_free = i915_ttm_delayed_free,
+	.migrate = i915_ttm_migrate,
+
+	.mmap_offset = i915_ttm_mmap_offset,
+	.unmap_virtual = i915_ttm_unmap_virtual,
+	.mmap_ops = &vm_ops_ttm,
+};
+
+void i915_ttm_bo_destroy(struct ttm_buffer_object *bo)
+{
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+
+	i915_gem_object_release_memory_region(obj);
+	mutex_destroy(&obj->ttm.get_io_page.lock);
+
+	if (obj->ttm.created) {
+		/*
+		 * We freely manage the shrinker LRU outide of the mm.pages life
+		 * cycle. As a result when destroying the object we should be
+		 * extra paranoid and ensure we remove it from the LRU, before
+		 * we free the object.
+		 *
+		 * Touching the ttm_shrinkable outside of the object lock here
+		 * should be safe now that the last GEM object ref was dropped.
+		 */
+		if (obj->mm.ttm_shrinkable)
+			i915_gem_object_make_unshrinkable(obj);
+
+		i915_ttm_backup_free(obj);
+
+		/* This releases all gem object bindings to the backend. */
+		__i915_gem_free_object(obj);
+
+		call_rcu(&obj->rcu, __i915_gem_free_object_rcu);
+	} else {
+		__i915_gem_object_fini(obj);
+	}
+}
+
+/**
+ * __i915_gem_ttm_object_init - Initialize a ttm-backed i915 gem object
+ * @mem: The initial memory region for the object.
+ * @obj: The gem object.
+ * @size: Object size in bytes.
+ * @flags: gem object flags.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int __i915_gem_ttm_object_init(struct intel_memory_region *mem,
+			       struct drm_i915_gem_object *obj,
+			       resource_size_t offset,
+			       resource_size_t size,
+			       resource_size_t page_size,
+			       unsigned int flags)
+{
+	static struct lock_class_key lock_class;
+	struct drm_i915_private *i915 = mem->i915;
+	struct ttm_operation_ctx ctx = {
+		.interruptible = true,
+		.no_wait_gpu = false,
+	};
+	enum ttm_bo_type bo_type;
+	int ret;
+
+	drm_gem_private_object_init(&i915->drm, &obj->base, size);
+	i915_gem_object_init(obj, &i915_gem_ttm_obj_ops, &lock_class, flags);
+
+	obj->bo_offset = offset;
+
+	/* Don't put on a region list until we're either locked or fully initialized. */
+	obj->mm.region = mem;
+	INIT_LIST_HEAD(&obj->mm.region_link);
+
+	INIT_RADIX_TREE(&obj->ttm.get_io_page.radix, GFP_KERNEL | __GFP_NOWARN);
+	mutex_init(&obj->ttm.get_io_page.lock);
+	bo_type = (obj->flags & I915_BO_ALLOC_USER) ? ttm_bo_type_device :
+		ttm_bo_type_kernel;
+
+	obj->base.vma_node.driver_private = i915_gem_to_ttm(obj);
+
+	/* Forcing the page size is kernel internal only */
+	GEM_BUG_ON(page_size && obj->mm.n_placements);
+
+	/*
+	 * Keep an extra shrink pin to prevent the object from being made
+	 * shrinkable too early. If the ttm_tt is ever allocated in shmem, we
+	 * drop the pin. The TTM backend manages the shrinker LRU itself,
+	 * outside of the normal mm.pages life cycle.
+	 */
+	i915_gem_object_make_unshrinkable(obj);
+
+	/*
+	 * If this function fails, it will call the destructor, but
+	 * our caller still owns the object. So no freeing in the
+	 * destructor until obj->ttm.created is true.
+	 * Similarly, in delayed_destroy, we can't call ttm_bo_put()
+	 * until successful initialization.
+	 */
+	ret = ttm_bo_init_reserved(&i915->bdev, i915_gem_to_ttm(obj), bo_type,
+				   &i915_sys_placement, page_size >> PAGE_SHIFT,
+				   &ctx, NULL, NULL, i915_ttm_bo_destroy);
+	if (ret)
+		return i915_ttm_err_to_gem(ret);
+
+	obj->ttm.created = true;
+	i915_gem_object_release_memory_region(obj);
+	i915_gem_object_init_memory_region(obj, mem);
+	i915_ttm_adjust_domains_after_move(obj);
+	i915_ttm_adjust_gem_after_move(obj);
+	i915_gem_object_unlock(obj);
+
+	return 0;
+}
+
+static const struct intel_memory_region_ops ttm_system_region_ops = {
+	.init_object = __i915_gem_ttm_object_init,
+	.release = intel_region_ttm_fini,
+};
+
+struct intel_memory_region *
+i915_gem_ttm_system_setup(struct drm_i915_private *i915,
+			  u16 type, u16 instance)
+{
+	struct intel_memory_region *mr;
+
+	mr = intel_memory_region_create(i915, 0,
+					totalram_pages() << PAGE_SHIFT,
+					PAGE_SIZE, 0, 0,
+					type, instance,
+					&ttm_system_region_ops);
+	if (IS_ERR(mr))
+		return mr;
+
+	intel_memory_region_set_name(mr, "system-ttm");
+	return mr;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.h b/drivers/gpu/drm/i915/gem/i915_gem_ttm.h
new file mode 100644
index 000000000..2a94a99ef
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.h
@@ -0,0 +1,106 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+#ifndef _I915_GEM_TTM_H_
+#define _I915_GEM_TTM_H_
+
+#include <drm/ttm/ttm_placement.h>
+
+#include "gem/i915_gem_object_types.h"
+
+/**
+ * i915_gem_to_ttm - Convert a struct drm_i915_gem_object to a
+ * struct ttm_buffer_object.
+ * @obj: Pointer to the gem object.
+ *
+ * Return: Pointer to the embedded struct ttm_buffer_object.
+ */
+static inline struct ttm_buffer_object *
+i915_gem_to_ttm(struct drm_i915_gem_object *obj)
+{
+	return &obj->__do_not_access;
+}
+
+/*
+ * i915 ttm gem object destructor. Internal use only.
+ */
+void i915_ttm_bo_destroy(struct ttm_buffer_object *bo);
+
+/**
+ * i915_ttm_is_ghost_object - Check if the ttm bo is a ghost object.
+ * @bo: Pointer to the ttm buffer object
+ *
+ * Return: True if the ttm bo is not a i915 object but a ghost ttm object,
+ * False otherwise.
+ */
+static inline bool i915_ttm_is_ghost_object(struct ttm_buffer_object *bo)
+{
+	return bo->destroy != i915_ttm_bo_destroy;
+}
+
+/**
+ * i915_ttm_to_gem - Convert a struct ttm_buffer_object to an embedding
+ * struct drm_i915_gem_object.
+ *
+ * Return: Pointer to the embedding struct ttm_buffer_object.
+ */
+static inline struct drm_i915_gem_object *
+i915_ttm_to_gem(struct ttm_buffer_object *bo)
+{
+	return container_of(bo, struct drm_i915_gem_object, __do_not_access);
+}
+
+int __i915_gem_ttm_object_init(struct intel_memory_region *mem,
+			       struct drm_i915_gem_object *obj,
+			       resource_size_t offset,
+			       resource_size_t size,
+			       resource_size_t page_size,
+			       unsigned int flags);
+
+/* Internal I915 TTM declarations and definitions below. */
+
+#define I915_PL_LMEM0 TTM_PL_PRIV
+#define I915_PL_SYSTEM TTM_PL_SYSTEM
+#define I915_PL_STOLEN TTM_PL_VRAM
+#define I915_PL_GGTT TTM_PL_TT
+
+struct ttm_placement *i915_ttm_sys_placement(void);
+
+void i915_ttm_free_cached_io_rsgt(struct drm_i915_gem_object *obj);
+
+struct i915_refct_sgt *
+i915_ttm_resource_get_st(struct drm_i915_gem_object *obj,
+			 struct ttm_resource *res);
+
+void i915_ttm_adjust_lru(struct drm_i915_gem_object *obj);
+
+int i915_ttm_purge(struct drm_i915_gem_object *obj);
+
+/**
+ * i915_ttm_gtt_binds_lmem - Should the memory be viewed as LMEM by the GTT?
+ * @mem: struct ttm_resource representing the memory.
+ *
+ * Return: true if memory should be viewed as LMEM for GTT binding purposes,
+ * false otherwise.
+ */
+static inline bool i915_ttm_gtt_binds_lmem(struct ttm_resource *mem)
+{
+	return mem->mem_type != I915_PL_SYSTEM;
+}
+
+/**
+ * i915_ttm_cpu_maps_iomem - Should the memory be viewed as IOMEM by the CPU?
+ * @mem: struct ttm_resource representing the memory.
+ *
+ * Return: true if memory should be viewed as IOMEM for CPU mapping purposes.
+ */
+static inline bool i915_ttm_cpu_maps_iomem(struct ttm_resource *mem)
+{
+	/* Once / if we support GGTT, this is also false for cached ttm_tts */
+	return mem->mem_type != I915_PL_SYSTEM;
+}
+
+bool i915_ttm_resource_mappable(struct ttm_resource *res);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
new file mode 100644
index 000000000..f59f812dc
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
@@ -0,0 +1,705 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <drm/ttm/ttm_bo_driver.h>
+
+#include "i915_deps.h"
+#include "i915_drv.h"
+#include "intel_memory_region.h"
+#include "intel_region_ttm.h"
+
+#include "gem/i915_gem_object.h"
+#include "gem/i915_gem_region.h"
+#include "gem/i915_gem_ttm.h"
+#include "gem/i915_gem_ttm_move.h"
+
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_migrate.h"
+
+/**
+ * DOC: Selftest failure modes for failsafe migration:
+ *
+ * For fail_gpu_migration, the gpu blit scheduled is always a clear blit
+ * rather than a copy blit, and then we force the failure paths as if
+ * the blit fence returned an error.
+ *
+ * For fail_work_allocation we fail the kmalloc of the async worker, we
+ * sync the gpu blit. If it then fails, or fail_gpu_migration is set to
+ * true, then a memcpy operation is performed sync.
+ */
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+static bool fail_gpu_migration;
+static bool fail_work_allocation;
+static bool ban_memcpy;
+
+void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
+					bool work_allocation)
+{
+	fail_gpu_migration = gpu_migration;
+	fail_work_allocation = work_allocation;
+}
+
+void i915_ttm_migrate_set_ban_memcpy(bool ban)
+{
+	ban_memcpy = ban;
+}
+#endif
+
+static enum i915_cache_level
+i915_ttm_cache_level(struct drm_i915_private *i915, struct ttm_resource *res,
+		     struct ttm_tt *ttm)
+{
+	return ((HAS_LLC(i915) || HAS_SNOOP(i915)) &&
+		!i915_ttm_gtt_binds_lmem(res) &&
+		ttm->caching == ttm_cached) ? I915_CACHE_LLC :
+		I915_CACHE_NONE;
+}
+
+static struct intel_memory_region *
+i915_ttm_region(struct ttm_device *bdev, int ttm_mem_type)
+{
+	struct drm_i915_private *i915 = container_of(bdev, typeof(*i915), bdev);
+
+	/* There's some room for optimization here... */
+	GEM_BUG_ON(ttm_mem_type != I915_PL_SYSTEM &&
+		   ttm_mem_type < I915_PL_LMEM0);
+	if (ttm_mem_type == I915_PL_SYSTEM)
+		return intel_memory_region_lookup(i915, INTEL_MEMORY_SYSTEM,
+						  0);
+
+	return intel_memory_region_lookup(i915, INTEL_MEMORY_LOCAL,
+					  ttm_mem_type - I915_PL_LMEM0);
+}
+
+/**
+ * i915_ttm_adjust_domains_after_move - Adjust the GEM domains after a
+ * TTM move
+ * @obj: The gem object
+ */
+void i915_ttm_adjust_domains_after_move(struct drm_i915_gem_object *obj)
+{
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+
+	if (i915_ttm_cpu_maps_iomem(bo->resource) || bo->ttm->caching != ttm_cached) {
+		obj->write_domain = I915_GEM_DOMAIN_WC;
+		obj->read_domains = I915_GEM_DOMAIN_WC;
+	} else {
+		obj->write_domain = I915_GEM_DOMAIN_CPU;
+		obj->read_domains = I915_GEM_DOMAIN_CPU;
+	}
+}
+
+/**
+ * i915_ttm_adjust_gem_after_move - Adjust the GEM state after a TTM move
+ * @obj: The gem object
+ *
+ * Adjusts the GEM object's region, mem_flags and cache coherency after a
+ * TTM move.
+ */
+void i915_ttm_adjust_gem_after_move(struct drm_i915_gem_object *obj)
+{
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+	unsigned int cache_level;
+	unsigned int i;
+
+	/*
+	 * If object was moved to an allowable region, update the object
+	 * region to consider it migrated. Note that if it's currently not
+	 * in an allowable region, it's evicted and we don't update the
+	 * object region.
+	 */
+	if (intel_region_to_ttm_type(obj->mm.region) != bo->resource->mem_type) {
+		for (i = 0; i < obj->mm.n_placements; ++i) {
+			struct intel_memory_region *mr = obj->mm.placements[i];
+
+			if (intel_region_to_ttm_type(mr) == bo->resource->mem_type &&
+			    mr != obj->mm.region) {
+				i915_gem_object_release_memory_region(obj);
+				i915_gem_object_init_memory_region(obj, mr);
+				break;
+			}
+		}
+	}
+
+	obj->mem_flags &= ~(I915_BO_FLAG_STRUCT_PAGE | I915_BO_FLAG_IOMEM);
+
+	obj->mem_flags |= i915_ttm_cpu_maps_iomem(bo->resource) ? I915_BO_FLAG_IOMEM :
+		I915_BO_FLAG_STRUCT_PAGE;
+
+	cache_level = i915_ttm_cache_level(to_i915(bo->base.dev), bo->resource,
+					   bo->ttm);
+	i915_gem_object_set_cache_coherency(obj, cache_level);
+}
+
+/**
+ * i915_ttm_move_notify - Prepare an object for move
+ * @bo: The ttm buffer object.
+ *
+ * This function prepares an object for move by removing all GPU bindings,
+ * removing all CPU mapings and finally releasing the pages sg-table.
+ *
+ * Return: 0 if successful, negative error code on error.
+ */
+int i915_ttm_move_notify(struct ttm_buffer_object *bo)
+{
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+	int ret;
+
+	/*
+	 * Note: The async unbinding here will actually transform the
+	 * blocking wait for unbind into a wait before finally submitting
+	 * evict / migration blit and thus stall the migration timeline
+	 * which may not be good for overall throughput. We should make
+	 * sure we await the unbind fences *after* the migration blit
+	 * instead of *before* as we currently do.
+	 */
+	ret = i915_gem_object_unbind(obj, I915_GEM_OBJECT_UNBIND_ACTIVE |
+				     I915_GEM_OBJECT_UNBIND_ASYNC);
+	if (ret)
+		return ret;
+
+	ret = __i915_gem_object_put_pages(obj);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static struct dma_fence *i915_ttm_accel_move(struct ttm_buffer_object *bo,
+					     bool clear,
+					     struct ttm_resource *dst_mem,
+					     struct ttm_tt *dst_ttm,
+					     struct sg_table *dst_st,
+					     const struct i915_deps *deps)
+{
+	struct drm_i915_private *i915 = container_of(bo->bdev, typeof(*i915),
+						     bdev);
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+	struct i915_request *rq;
+	struct ttm_tt *src_ttm = bo->ttm;
+	enum i915_cache_level src_level, dst_level;
+	int ret;
+
+	if (!to_gt(i915)->migrate.context || intel_gt_is_wedged(to_gt(i915)))
+		return ERR_PTR(-EINVAL);
+
+	/* With fail_gpu_migration, we always perform a GPU clear. */
+	if (I915_SELFTEST_ONLY(fail_gpu_migration))
+		clear = true;
+
+	dst_level = i915_ttm_cache_level(i915, dst_mem, dst_ttm);
+	if (clear) {
+		if (bo->type == ttm_bo_type_kernel &&
+		    !I915_SELFTEST_ONLY(fail_gpu_migration))
+			return ERR_PTR(-EINVAL);
+
+		intel_engine_pm_get(to_gt(i915)->migrate.context->engine);
+		ret = intel_context_migrate_clear(to_gt(i915)->migrate.context, deps,
+						  dst_st->sgl, dst_level,
+						  i915_ttm_gtt_binds_lmem(dst_mem),
+						  0, &rq);
+	} else {
+		struct i915_refct_sgt *src_rsgt =
+			i915_ttm_resource_get_st(obj, bo->resource);
+
+		if (IS_ERR(src_rsgt))
+			return ERR_CAST(src_rsgt);
+
+		src_level = i915_ttm_cache_level(i915, bo->resource, src_ttm);
+		intel_engine_pm_get(to_gt(i915)->migrate.context->engine);
+		ret = intel_context_migrate_copy(to_gt(i915)->migrate.context,
+						 deps, src_rsgt->table.sgl,
+						 src_level,
+						 i915_ttm_gtt_binds_lmem(bo->resource),
+						 dst_st->sgl, dst_level,
+						 i915_ttm_gtt_binds_lmem(dst_mem),
+						 &rq);
+
+		i915_refct_sgt_put(src_rsgt);
+	}
+
+	intel_engine_pm_put(to_gt(i915)->migrate.context->engine);
+
+	if (ret && rq) {
+		i915_request_wait(rq, 0, MAX_SCHEDULE_TIMEOUT);
+		i915_request_put(rq);
+	}
+
+	return ret ? ERR_PTR(ret) : &rq->fence;
+}
+
+/**
+ * struct i915_ttm_memcpy_arg - argument for the bo memcpy functionality.
+ * @_dst_iter: Storage space for the destination kmap iterator.
+ * @_src_iter: Storage space for the source kmap iterator.
+ * @dst_iter: Pointer to the destination kmap iterator.
+ * @src_iter: Pointer to the source kmap iterator.
+ * @clear: Whether to clear instead of copy.
+ * @src_rsgt: Refcounted scatter-gather list of source memory.
+ * @dst_rsgt: Refcounted scatter-gather list of destination memory.
+ */
+struct i915_ttm_memcpy_arg {
+	union {
+		struct ttm_kmap_iter_tt tt;
+		struct ttm_kmap_iter_iomap io;
+	} _dst_iter,
+	_src_iter;
+	struct ttm_kmap_iter *dst_iter;
+	struct ttm_kmap_iter *src_iter;
+	unsigned long num_pages;
+	bool clear;
+	struct i915_refct_sgt *src_rsgt;
+	struct i915_refct_sgt *dst_rsgt;
+};
+
+/**
+ * struct i915_ttm_memcpy_work - Async memcpy worker under a dma-fence.
+ * @fence: The dma-fence.
+ * @work: The work struct use for the memcpy work.
+ * @lock: The fence lock. Not used to protect anything else ATM.
+ * @irq_work: Low latency worker to signal the fence since it can't be done
+ * from the callback for lockdep reasons.
+ * @cb: Callback for the accelerated migration fence.
+ * @arg: The argument for the memcpy functionality.
+ * @i915: The i915 pointer.
+ * @obj: The GEM object.
+ * @memcpy_allowed: Instead of processing the @arg, and falling back to memcpy
+ * or memset, we wedge the device and set the @obj unknown_state, to prevent
+ * further access to the object with the CPU or GPU.  On some devices we might
+ * only be permitted to use the blitter engine for such operations.
+ */
+struct i915_ttm_memcpy_work {
+	struct dma_fence fence;
+	struct work_struct work;
+	spinlock_t lock;
+	struct irq_work irq_work;
+	struct dma_fence_cb cb;
+	struct i915_ttm_memcpy_arg arg;
+	struct drm_i915_private *i915;
+	struct drm_i915_gem_object *obj;
+	bool memcpy_allowed;
+};
+
+static void i915_ttm_move_memcpy(struct i915_ttm_memcpy_arg *arg)
+{
+	ttm_move_memcpy(arg->clear, arg->num_pages,
+			arg->dst_iter, arg->src_iter);
+}
+
+static void i915_ttm_memcpy_init(struct i915_ttm_memcpy_arg *arg,
+				 struct ttm_buffer_object *bo, bool clear,
+				 struct ttm_resource *dst_mem,
+				 struct ttm_tt *dst_ttm,
+				 struct i915_refct_sgt *dst_rsgt)
+{
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+	struct intel_memory_region *dst_reg, *src_reg;
+
+	dst_reg = i915_ttm_region(bo->bdev, dst_mem->mem_type);
+	src_reg = i915_ttm_region(bo->bdev, bo->resource->mem_type);
+	GEM_BUG_ON(!dst_reg || !src_reg);
+
+	arg->dst_iter = !i915_ttm_cpu_maps_iomem(dst_mem) ?
+		ttm_kmap_iter_tt_init(&arg->_dst_iter.tt, dst_ttm) :
+		ttm_kmap_iter_iomap_init(&arg->_dst_iter.io, &dst_reg->iomap,
+					 &dst_rsgt->table, dst_reg->region.start);
+
+	arg->src_iter = !i915_ttm_cpu_maps_iomem(bo->resource) ?
+		ttm_kmap_iter_tt_init(&arg->_src_iter.tt, bo->ttm) :
+		ttm_kmap_iter_iomap_init(&arg->_src_iter.io, &src_reg->iomap,
+					 &obj->ttm.cached_io_rsgt->table,
+					 src_reg->region.start);
+	arg->clear = clear;
+	arg->num_pages = bo->base.size >> PAGE_SHIFT;
+
+	arg->dst_rsgt = i915_refct_sgt_get(dst_rsgt);
+	arg->src_rsgt = clear ? NULL :
+		i915_ttm_resource_get_st(obj, bo->resource);
+}
+
+static void i915_ttm_memcpy_release(struct i915_ttm_memcpy_arg *arg)
+{
+	i915_refct_sgt_put(arg->src_rsgt);
+	i915_refct_sgt_put(arg->dst_rsgt);
+}
+
+static void __memcpy_work(struct work_struct *work)
+{
+	struct i915_ttm_memcpy_work *copy_work =
+		container_of(work, typeof(*copy_work), work);
+	struct i915_ttm_memcpy_arg *arg = &copy_work->arg;
+	bool cookie;
+
+	/*
+	 * FIXME: We need to take a closer look here. We should be able to plonk
+	 * this into the fence critical section.
+	 */
+	if (!copy_work->memcpy_allowed) {
+		struct intel_gt *gt;
+		unsigned int id;
+
+		for_each_gt(gt, copy_work->i915, id)
+			intel_gt_set_wedged(gt);
+	}
+
+	cookie = dma_fence_begin_signalling();
+
+	if (copy_work->memcpy_allowed) {
+		i915_ttm_move_memcpy(arg);
+	} else {
+		/*
+		 * Prevent further use of the object. Any future GTT binding or
+		 * CPU access is not allowed once we signal the fence. Outside
+		 * of the fence critical section, we then also then wedge the gpu
+		 * to indicate the device is not functional.
+		 *
+		 * The below dma_fence_signal() is our write-memory-barrier.
+		 */
+		copy_work->obj->mm.unknown_state = true;
+	}
+
+	dma_fence_end_signalling(cookie);
+
+	dma_fence_signal(&copy_work->fence);
+
+	i915_ttm_memcpy_release(arg);
+	i915_gem_object_put(copy_work->obj);
+	dma_fence_put(&copy_work->fence);
+}
+
+static void __memcpy_irq_work(struct irq_work *irq_work)
+{
+	struct i915_ttm_memcpy_work *copy_work =
+		container_of(irq_work, typeof(*copy_work), irq_work);
+	struct i915_ttm_memcpy_arg *arg = &copy_work->arg;
+
+	dma_fence_signal(&copy_work->fence);
+	i915_ttm_memcpy_release(arg);
+	i915_gem_object_put(copy_work->obj);
+	dma_fence_put(&copy_work->fence);
+}
+
+static void __memcpy_cb(struct dma_fence *fence, struct dma_fence_cb *cb)
+{
+	struct i915_ttm_memcpy_work *copy_work =
+		container_of(cb, typeof(*copy_work), cb);
+
+	if (unlikely(fence->error || I915_SELFTEST_ONLY(fail_gpu_migration))) {
+		INIT_WORK(&copy_work->work, __memcpy_work);
+		queue_work(system_unbound_wq, &copy_work->work);
+	} else {
+		init_irq_work(&copy_work->irq_work, __memcpy_irq_work);
+		irq_work_queue(&copy_work->irq_work);
+	}
+}
+
+static const char *get_driver_name(struct dma_fence *fence)
+{
+	return "i915_ttm_memcpy_work";
+}
+
+static const char *get_timeline_name(struct dma_fence *fence)
+{
+	return "unbound";
+}
+
+static const struct dma_fence_ops dma_fence_memcpy_ops = {
+	.get_driver_name = get_driver_name,
+	.get_timeline_name = get_timeline_name,
+};
+
+static struct dma_fence *
+i915_ttm_memcpy_work_arm(struct i915_ttm_memcpy_work *work,
+			 struct dma_fence *dep)
+{
+	int ret;
+
+	spin_lock_init(&work->lock);
+	dma_fence_init(&work->fence, &dma_fence_memcpy_ops, &work->lock, 0, 0);
+	dma_fence_get(&work->fence);
+	ret = dma_fence_add_callback(dep, &work->cb, __memcpy_cb);
+	if (ret) {
+		if (ret != -ENOENT)
+			dma_fence_wait(dep, false);
+
+		return ERR_PTR(I915_SELFTEST_ONLY(fail_gpu_migration) ? -EINVAL :
+			       dep->error);
+	}
+
+	return &work->fence;
+}
+
+static bool i915_ttm_memcpy_allowed(struct ttm_buffer_object *bo,
+				    struct ttm_resource *dst_mem)
+{
+	if (i915_gem_object_needs_ccs_pages(i915_ttm_to_gem(bo)))
+		return false;
+
+	if (!(i915_ttm_resource_mappable(bo->resource) &&
+	      i915_ttm_resource_mappable(dst_mem)))
+		return false;
+
+	return I915_SELFTEST_ONLY(ban_memcpy) ? false : true;
+}
+
+static struct dma_fence *
+__i915_ttm_move(struct ttm_buffer_object *bo,
+		const struct ttm_operation_ctx *ctx, bool clear,
+		struct ttm_resource *dst_mem, struct ttm_tt *dst_ttm,
+		struct i915_refct_sgt *dst_rsgt, bool allow_accel,
+		const struct i915_deps *move_deps)
+{
+	const bool memcpy_allowed = i915_ttm_memcpy_allowed(bo, dst_mem);
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+	struct drm_i915_private *i915 = to_i915(bo->base.dev);
+	struct i915_ttm_memcpy_work *copy_work = NULL;
+	struct i915_ttm_memcpy_arg _arg, *arg = &_arg;
+	struct dma_fence *fence = ERR_PTR(-EINVAL);
+
+	if (allow_accel) {
+		fence = i915_ttm_accel_move(bo, clear, dst_mem, dst_ttm,
+					    &dst_rsgt->table, move_deps);
+
+		/*
+		 * We only need to intercept the error when moving to lmem.
+		 * When moving to system, TTM or shmem will provide us with
+		 * cleared pages.
+		 */
+		if (!IS_ERR(fence) && !i915_ttm_gtt_binds_lmem(dst_mem) &&
+		    !I915_SELFTEST_ONLY(fail_gpu_migration ||
+					fail_work_allocation))
+			goto out;
+	}
+
+	/* If we've scheduled gpu migration. Try to arm error intercept. */
+	if (!IS_ERR(fence)) {
+		struct dma_fence *dep = fence;
+
+		if (!I915_SELFTEST_ONLY(fail_work_allocation))
+			copy_work = kzalloc(sizeof(*copy_work), GFP_KERNEL);
+
+		if (copy_work) {
+			copy_work->i915 = i915;
+			copy_work->memcpy_allowed = memcpy_allowed;
+			copy_work->obj = i915_gem_object_get(obj);
+			arg = &copy_work->arg;
+			if (memcpy_allowed)
+				i915_ttm_memcpy_init(arg, bo, clear, dst_mem,
+						     dst_ttm, dst_rsgt);
+
+			fence = i915_ttm_memcpy_work_arm(copy_work, dep);
+		} else {
+			dma_fence_wait(dep, false);
+			fence = ERR_PTR(I915_SELFTEST_ONLY(fail_gpu_migration) ?
+					-EINVAL : fence->error);
+		}
+		dma_fence_put(dep);
+
+		if (!IS_ERR(fence))
+			goto out;
+	} else {
+		int err = PTR_ERR(fence);
+
+		if (err == -EINTR || err == -ERESTARTSYS || err == -EAGAIN)
+			return fence;
+
+		if (move_deps) {
+			err = i915_deps_sync(move_deps, ctx);
+			if (err)
+				return ERR_PTR(err);
+		}
+	}
+
+	/* Error intercept failed or no accelerated migration to start with */
+
+	if (memcpy_allowed) {
+		if (!copy_work)
+			i915_ttm_memcpy_init(arg, bo, clear, dst_mem, dst_ttm,
+					     dst_rsgt);
+		i915_ttm_move_memcpy(arg);
+		i915_ttm_memcpy_release(arg);
+	}
+	if (copy_work)
+		i915_gem_object_put(copy_work->obj);
+	kfree(copy_work);
+
+	return memcpy_allowed ? NULL : ERR_PTR(-EIO);
+out:
+	if (!fence && copy_work) {
+		i915_ttm_memcpy_release(arg);
+		i915_gem_object_put(copy_work->obj);
+		kfree(copy_work);
+	}
+
+	return fence;
+}
+
+/**
+ * i915_ttm_move - The TTM move callback used by i915.
+ * @bo: The buffer object.
+ * @evict: Whether this is an eviction.
+ * @dst_mem: The destination ttm resource.
+ * @hop: If we need multihop, what temporary memory type to move to.
+ *
+ * Return: 0 if successful, negative error code otherwise.
+ */
+int i915_ttm_move(struct ttm_buffer_object *bo, bool evict,
+		  struct ttm_operation_ctx *ctx,
+		  struct ttm_resource *dst_mem,
+		  struct ttm_place *hop)
+{
+	struct drm_i915_gem_object *obj = i915_ttm_to_gem(bo);
+	struct ttm_resource_manager *dst_man =
+		ttm_manager_type(bo->bdev, dst_mem->mem_type);
+	struct dma_fence *migration_fence = NULL;
+	struct ttm_tt *ttm = bo->ttm;
+	struct i915_refct_sgt *dst_rsgt;
+	bool clear;
+	int ret;
+
+	if (GEM_WARN_ON(i915_ttm_is_ghost_object(bo))) {
+		ttm_bo_move_null(bo, dst_mem);
+		return 0;
+	}
+
+	ret = i915_ttm_move_notify(bo);
+	if (ret)
+		return ret;
+
+	if (obj->mm.madv != I915_MADV_WILLNEED) {
+		i915_ttm_purge(obj);
+		ttm_resource_free(bo, &dst_mem);
+		return 0;
+	}
+
+	/* Populate ttm with pages if needed. Typically system memory. */
+	if (ttm && (dst_man->use_tt || (ttm->page_flags & TTM_TT_FLAG_SWAPPED))) {
+		ret = ttm_tt_populate(bo->bdev, ttm, ctx);
+		if (ret)
+			return ret;
+	}
+
+	dst_rsgt = i915_ttm_resource_get_st(obj, dst_mem);
+	if (IS_ERR(dst_rsgt))
+		return PTR_ERR(dst_rsgt);
+
+	clear = !i915_ttm_cpu_maps_iomem(bo->resource) && (!ttm || !ttm_tt_is_populated(ttm));
+	if (!(clear && ttm && !(ttm->page_flags & TTM_TT_FLAG_ZERO_ALLOC))) {
+		struct i915_deps deps;
+
+		i915_deps_init(&deps, GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN);
+		ret = i915_deps_add_resv(&deps, bo->base.resv, ctx);
+		if (ret) {
+			i915_refct_sgt_put(dst_rsgt);
+			return ret;
+		}
+
+		migration_fence = __i915_ttm_move(bo, ctx, clear, dst_mem, ttm,
+						  dst_rsgt, true, &deps);
+		i915_deps_fini(&deps);
+	}
+
+	/* We can possibly get an -ERESTARTSYS here */
+	if (IS_ERR(migration_fence)) {
+		i915_refct_sgt_put(dst_rsgt);
+		return PTR_ERR(migration_fence);
+	}
+
+	if (migration_fence) {
+		if (I915_SELFTEST_ONLY(evict && fail_gpu_migration))
+			ret = -EIO; /* never feed non-migrate fences into ttm */
+		else
+			ret = ttm_bo_move_accel_cleanup(bo, migration_fence, evict,
+							true, dst_mem);
+		if (ret) {
+			dma_fence_wait(migration_fence, false);
+			ttm_bo_move_sync_cleanup(bo, dst_mem);
+		}
+		dma_fence_put(migration_fence);
+	} else {
+		ttm_bo_move_sync_cleanup(bo, dst_mem);
+	}
+
+	i915_ttm_adjust_domains_after_move(obj);
+	i915_ttm_free_cached_io_rsgt(obj);
+
+	if (i915_ttm_gtt_binds_lmem(dst_mem) || i915_ttm_cpu_maps_iomem(dst_mem)) {
+		obj->ttm.cached_io_rsgt = dst_rsgt;
+		obj->ttm.get_io_page.sg_pos = dst_rsgt->table.sgl;
+		obj->ttm.get_io_page.sg_idx = 0;
+	} else {
+		i915_refct_sgt_put(dst_rsgt);
+	}
+
+	i915_ttm_adjust_lru(obj);
+	i915_ttm_adjust_gem_after_move(obj);
+	return 0;
+}
+
+/**
+ * i915_gem_obj_copy_ttm - Copy the contents of one ttm-based gem object to
+ * another
+ * @dst: The destination object
+ * @src: The source object
+ * @allow_accel: Allow using the blitter. Otherwise TTM memcpy is used.
+ * @intr: Whether to perform waits interruptible:
+ *
+ * Note: The caller is responsible for assuring that the underlying
+ * TTM objects are populated if needed and locked.
+ *
+ * Return: Zero on success. Negative error code on error. If @intr == true,
+ * then it may return -ERESTARTSYS or -EINTR.
+ */
+int i915_gem_obj_copy_ttm(struct drm_i915_gem_object *dst,
+			  struct drm_i915_gem_object *src,
+			  bool allow_accel, bool intr)
+{
+	struct ttm_buffer_object *dst_bo = i915_gem_to_ttm(dst);
+	struct ttm_buffer_object *src_bo = i915_gem_to_ttm(src);
+	struct ttm_operation_ctx ctx = {
+		.interruptible = intr,
+	};
+	struct i915_refct_sgt *dst_rsgt;
+	struct dma_fence *copy_fence;
+	struct i915_deps deps;
+	int ret;
+
+	assert_object_held(dst);
+	assert_object_held(src);
+	i915_deps_init(&deps, GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN);
+
+	ret = dma_resv_reserve_fences(src_bo->base.resv, 1);
+	if (ret)
+		return ret;
+
+	ret = dma_resv_reserve_fences(dst_bo->base.resv, 1);
+	if (ret)
+		return ret;
+
+	ret = i915_deps_add_resv(&deps, dst_bo->base.resv, &ctx);
+	if (ret)
+		return ret;
+
+	ret = i915_deps_add_resv(&deps, src_bo->base.resv, &ctx);
+	if (ret)
+		return ret;
+
+	dst_rsgt = i915_ttm_resource_get_st(dst, dst_bo->resource);
+	copy_fence = __i915_ttm_move(src_bo, &ctx, false, dst_bo->resource,
+				     dst_bo->ttm, dst_rsgt, allow_accel,
+				     &deps);
+
+	i915_deps_fini(&deps);
+	i915_refct_sgt_put(dst_rsgt);
+	if (IS_ERR_OR_NULL(copy_fence))
+		return PTR_ERR_OR_ZERO(copy_fence);
+
+	dma_resv_add_fence(dst_bo->base.resv, copy_fence, DMA_RESV_USAGE_WRITE);
+	dma_resv_add_fence(src_bo->base.resv, copy_fence, DMA_RESV_USAGE_READ);
+	dma_fence_put(copy_fence);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
new file mode 100644
index 000000000..8a5d5ab0c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
@@ -0,0 +1,42 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+#ifndef _I915_GEM_TTM_MOVE_H_
+#define _I915_GEM_TTM_MOVE_H_
+
+#include <linux/types.h>
+
+#include "i915_selftest.h"
+
+struct ttm_buffer_object;
+struct ttm_operation_ctx;
+struct ttm_place;
+struct ttm_resource;
+struct ttm_tt;
+
+struct drm_i915_gem_object;
+struct i915_refct_sgt;
+
+int i915_ttm_move_notify(struct ttm_buffer_object *bo);
+
+I915_SELFTEST_DECLARE(void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
+							      bool work_allocation));
+I915_SELFTEST_DECLARE(void i915_ttm_migrate_set_ban_memcpy(bool ban));
+
+int i915_gem_obj_copy_ttm(struct drm_i915_gem_object *dst,
+			  struct drm_i915_gem_object *src,
+			  bool allow_accel, bool intr);
+
+/* Internal I915 TTM declarations and definitions below. */
+
+int i915_ttm_move(struct ttm_buffer_object *bo, bool evict,
+		  struct ttm_operation_ctx *ctx,
+		  struct ttm_resource *dst_mem,
+		  struct ttm_place *hop);
+
+void i915_ttm_adjust_domains_after_move(struct drm_i915_gem_object *obj);
+
+void i915_ttm_adjust_gem_after_move(struct drm_i915_gem_object *obj);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.c
new file mode 100644
index 000000000..7e67742bc
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.c
@@ -0,0 +1,230 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <drm/ttm/ttm_placement.h>
+#include <drm/ttm/ttm_tt.h>
+
+#include "i915_drv.h"
+#include "intel_memory_region.h"
+#include "intel_region_ttm.h"
+
+#include "gem/i915_gem_region.h"
+#include "gem/i915_gem_ttm.h"
+#include "gem/i915_gem_ttm_move.h"
+#include "gem/i915_gem_ttm_pm.h"
+
+/**
+ * i915_ttm_backup_free - Free any backup attached to this object
+ * @obj: The object whose backup is to be freed.
+ */
+void i915_ttm_backup_free(struct drm_i915_gem_object *obj)
+{
+	if (obj->ttm.backup) {
+		i915_gem_object_put(obj->ttm.backup);
+		obj->ttm.backup = NULL;
+	}
+}
+
+/**
+ * struct i915_gem_ttm_pm_apply - Apply-to-region subclass for restore
+ * @base: The i915_gem_apply_to_region we derive from.
+ * @allow_gpu: Whether using the gpu blitter is allowed.
+ * @backup_pinned: On backup, backup also pinned objects.
+ */
+struct i915_gem_ttm_pm_apply {
+	struct i915_gem_apply_to_region base;
+	bool allow_gpu : 1;
+	bool backup_pinned : 1;
+};
+
+static int i915_ttm_backup(struct i915_gem_apply_to_region *apply,
+			   struct drm_i915_gem_object *obj)
+{
+	struct i915_gem_ttm_pm_apply *pm_apply =
+		container_of(apply, typeof(*pm_apply), base);
+	struct ttm_buffer_object *bo = i915_gem_to_ttm(obj);
+	struct ttm_buffer_object *backup_bo;
+	struct drm_i915_private *i915 =
+		container_of(bo->bdev, typeof(*i915), bdev);
+	struct drm_i915_gem_object *backup;
+	struct ttm_operation_ctx ctx = {};
+	unsigned int flags;
+	int err = 0;
+
+	if (bo->resource->mem_type == I915_PL_SYSTEM || obj->ttm.backup)
+		return 0;
+
+	if (pm_apply->allow_gpu && i915_gem_object_evictable(obj))
+		return ttm_bo_validate(bo, i915_ttm_sys_placement(), &ctx);
+
+	if (!pm_apply->backup_pinned ||
+	    (pm_apply->allow_gpu && (obj->flags & I915_BO_ALLOC_PM_EARLY)))
+		return 0;
+
+	if (obj->flags & I915_BO_ALLOC_PM_VOLATILE)
+		return 0;
+
+	/*
+	 * It seems that we might have some framebuffers still pinned at this
+	 * stage, but for such objects we might also need to deal with the CCS
+	 * aux state. Make sure we force the save/restore of the CCS state,
+	 * otherwise we might observe display corruption, when returning from
+	 * suspend.
+	 */
+	flags = 0;
+	if (i915_gem_object_needs_ccs_pages(obj)) {
+		WARN_ON_ONCE(!i915_gem_object_is_framebuffer(obj));
+		WARN_ON_ONCE(!pm_apply->allow_gpu);
+
+		flags = I915_BO_ALLOC_CCS_AUX;
+	}
+	backup = i915_gem_object_create_region(i915->mm.regions[INTEL_REGION_SMEM],
+					       obj->base.size, 0, flags);
+	if (IS_ERR(backup))
+		return PTR_ERR(backup);
+
+	err = i915_gem_object_lock(backup, apply->ww);
+	if (err)
+		goto out_no_lock;
+
+	backup_bo = i915_gem_to_ttm(backup);
+	err = ttm_tt_populate(backup_bo->bdev, backup_bo->ttm, &ctx);
+	if (err)
+		goto out_no_populate;
+
+	err = i915_gem_obj_copy_ttm(backup, obj, pm_apply->allow_gpu, false);
+	if (err) {
+		drm_err(&i915->drm,
+			"Unable to copy from device to system memory, err:%pe\n",
+			ERR_PTR(err));
+		goto out_no_populate;
+	}
+	ttm_bo_wait_ctx(backup_bo, &ctx);
+
+	obj->ttm.backup = backup;
+	return 0;
+
+out_no_populate:
+	i915_gem_ww_unlock_single(backup);
+out_no_lock:
+	i915_gem_object_put(backup);
+
+	return err;
+}
+
+static int i915_ttm_recover(struct i915_gem_apply_to_region *apply,
+			    struct drm_i915_gem_object *obj)
+{
+	i915_ttm_backup_free(obj);
+	return 0;
+}
+
+/**
+ * i915_ttm_recover_region - Free the backup of all objects of a region
+ * @mr: The memory region
+ *
+ * Checks all objects of a region if there is backup attached and if so
+ * frees that backup. Typically this is called to recover after a partially
+ * performed backup.
+ */
+void i915_ttm_recover_region(struct intel_memory_region *mr)
+{
+	static const struct i915_gem_apply_to_region_ops recover_ops = {
+		.process_obj = i915_ttm_recover,
+	};
+	struct i915_gem_apply_to_region apply = {.ops = &recover_ops};
+	int ret;
+
+	ret = i915_gem_process_region(mr, &apply);
+	GEM_WARN_ON(ret);
+}
+
+/**
+ * i915_ttm_backup_region - Back up all objects of a region to smem.
+ * @mr: The memory region
+ * @allow_gpu: Whether to allow the gpu blitter for this backup.
+ * @backup_pinned: Backup also pinned objects.
+ *
+ * Loops over all objects of a region and either evicts them if they are
+ * evictable or backs them up using a backup object if they are pinned.
+ *
+ * Return: Zero on success. Negative error code on error.
+ */
+int i915_ttm_backup_region(struct intel_memory_region *mr, u32 flags)
+{
+	static const struct i915_gem_apply_to_region_ops backup_ops = {
+		.process_obj = i915_ttm_backup,
+	};
+	struct i915_gem_ttm_pm_apply pm_apply = {
+		.base = {.ops = &backup_ops},
+		.allow_gpu = flags & I915_TTM_BACKUP_ALLOW_GPU,
+		.backup_pinned = flags & I915_TTM_BACKUP_PINNED,
+	};
+
+	return i915_gem_process_region(mr, &pm_apply.base);
+}
+
+static int i915_ttm_restore(struct i915_gem_apply_to_region *apply,
+			    struct drm_i915_gem_object *obj)
+{
+	struct i915_gem_ttm_pm_apply *pm_apply =
+		container_of(apply, typeof(*pm_apply), base);
+	struct drm_i915_gem_object *backup = obj->ttm.backup;
+	struct ttm_buffer_object *backup_bo = i915_gem_to_ttm(backup);
+	struct ttm_operation_ctx ctx = {};
+	int err;
+
+	if (!backup)
+		return 0;
+
+	if (!pm_apply->allow_gpu && !(obj->flags & I915_BO_ALLOC_PM_EARLY))
+		return 0;
+
+	err = i915_gem_object_lock(backup, apply->ww);
+	if (err)
+		return err;
+
+	/* Content may have been swapped. */
+	err = ttm_tt_populate(backup_bo->bdev, backup_bo->ttm, &ctx);
+	if (!err) {
+		err = i915_gem_obj_copy_ttm(obj, backup, pm_apply->allow_gpu,
+					    false);
+		GEM_WARN_ON(err);
+		ttm_bo_wait_ctx(backup_bo, &ctx);
+
+		obj->ttm.backup = NULL;
+		err = 0;
+	}
+
+	i915_gem_ww_unlock_single(backup);
+
+	if (!err)
+		i915_gem_object_put(backup);
+
+	return err;
+}
+
+/**
+ * i915_ttm_restore_region - Restore backed-up objects of a region from smem.
+ * @mr: The memory region
+ * @allow_gpu: Whether to allow the gpu blitter to recover.
+ *
+ * Loops over all objects of a region and if they are backed-up, restores
+ * them from smem.
+ *
+ * Return: Zero on success. Negative error code on error.
+ */
+int i915_ttm_restore_region(struct intel_memory_region *mr, u32 flags)
+{
+	static const struct i915_gem_apply_to_region_ops restore_ops = {
+		.process_obj = i915_ttm_restore,
+	};
+	struct i915_gem_ttm_pm_apply pm_apply = {
+		.base = {.ops = &restore_ops},
+		.allow_gpu = flags & I915_TTM_BACKUP_ALLOW_GPU,
+	};
+
+	return i915_gem_process_region(mr, &pm_apply.base);
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.h b/drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.h
new file mode 100644
index 000000000..25ed67a31
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _I915_GEM_TTM_PM_H_
+#define _I915_GEM_TTM_PM_H_
+
+#include <linux/types.h>
+
+struct intel_memory_region;
+struct drm_i915_gem_object;
+
+#define I915_TTM_BACKUP_ALLOW_GPU BIT(0)
+#define I915_TTM_BACKUP_PINNED    BIT(1)
+
+int i915_ttm_backup_region(struct intel_memory_region *mr, u32 flags);
+
+void i915_ttm_recover_region(struct intel_memory_region *mr);
+
+int i915_ttm_restore_region(struct intel_memory_region *mr, u32 flags);
+
+/* Internal I915 TTM functions below. */
+void i915_ttm_backup_free(struct drm_i915_gem_object *obj);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_userptr.c b/drivers/gpu/drm/i915/gem/i915_gem_userptr.c
new file mode 100644
index 000000000..ba14b18d6
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_userptr.c
@@ -0,0 +1,593 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2012-2014 Intel Corporation
+ *
+  * Based on amdgpu_mn, which bears the following notice:
+ *
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ */
+/*
+ * Authors:
+ *    Christian König <christian.koenig@amd.com>
+ */
+
+#include <linux/mmu_context.h>
+#include <linux/mempolicy.h>
+#include <linux/swap.h>
+#include <linux/sched/mm.h>
+
+#include "i915_drv.h"
+#include "i915_gem_ioctls.h"
+#include "i915_gem_object.h"
+#include "i915_gem_userptr.h"
+#include "i915_scatterlist.h"
+
+#ifdef CONFIG_MMU_NOTIFIER
+
+/**
+ * i915_gem_userptr_invalidate - callback to notify about mm change
+ *
+ * @mni: the range (mm) is about to update
+ * @range: details on the invalidation
+ * @cur_seq: Value to pass to mmu_interval_set_seq()
+ *
+ * Block for operations on BOs to finish and mark pages as accessed and
+ * potentially dirty.
+ */
+static bool i915_gem_userptr_invalidate(struct mmu_interval_notifier *mni,
+					const struct mmu_notifier_range *range,
+					unsigned long cur_seq)
+{
+	struct drm_i915_gem_object *obj = container_of(mni, struct drm_i915_gem_object, userptr.notifier);
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	long r;
+
+	if (!mmu_notifier_range_blockable(range))
+		return false;
+
+	write_lock(&i915->mm.notifier_lock);
+
+	mmu_interval_set_seq(mni, cur_seq);
+
+	write_unlock(&i915->mm.notifier_lock);
+
+	/*
+	 * We don't wait when the process is exiting. This is valid
+	 * because the object will be cleaned up anyway.
+	 *
+	 * This is also temporarily required as a hack, because we
+	 * cannot currently force non-consistent batch buffers to preempt
+	 * and reschedule by waiting on it, hanging processes on exit.
+	 */
+	if (current->flags & PF_EXITING)
+		return true;
+
+	/* we will unbind on next submission, still have userptr pins */
+	r = dma_resv_wait_timeout(obj->base.resv, DMA_RESV_USAGE_BOOKKEEP, false,
+				  MAX_SCHEDULE_TIMEOUT);
+	if (r <= 0)
+		drm_err(&i915->drm, "(%ld) failed to wait for idle\n", r);
+
+	return true;
+}
+
+static const struct mmu_interval_notifier_ops i915_gem_userptr_notifier_ops = {
+	.invalidate = i915_gem_userptr_invalidate,
+};
+
+static int
+i915_gem_userptr_init__mmu_notifier(struct drm_i915_gem_object *obj)
+{
+	return mmu_interval_notifier_insert(&obj->userptr.notifier, current->mm,
+					    obj->userptr.ptr, obj->base.size,
+					    &i915_gem_userptr_notifier_ops);
+}
+
+static void i915_gem_object_userptr_drop_ref(struct drm_i915_gem_object *obj)
+{
+	struct page **pvec = NULL;
+
+	assert_object_held_shared(obj);
+
+	if (!--obj->userptr.page_ref) {
+		pvec = obj->userptr.pvec;
+		obj->userptr.pvec = NULL;
+	}
+	GEM_BUG_ON(obj->userptr.page_ref < 0);
+
+	if (pvec) {
+		const unsigned long num_pages = obj->base.size >> PAGE_SHIFT;
+
+		unpin_user_pages(pvec, num_pages);
+		kvfree(pvec);
+	}
+}
+
+static int i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj)
+{
+	const unsigned long num_pages = obj->base.size >> PAGE_SHIFT;
+	unsigned int max_segment = i915_sg_segment_size(obj->base.dev->dev);
+	struct sg_table *st;
+	unsigned int sg_page_sizes;
+	struct page **pvec;
+	int ret;
+
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (!st)
+		return -ENOMEM;
+
+	if (!obj->userptr.page_ref) {
+		ret = -EAGAIN;
+		goto err_free;
+	}
+
+	obj->userptr.page_ref++;
+	pvec = obj->userptr.pvec;
+
+alloc_table:
+	ret = sg_alloc_table_from_pages_segment(st, pvec, num_pages, 0,
+						num_pages << PAGE_SHIFT,
+						max_segment, GFP_KERNEL);
+	if (ret)
+		goto err;
+
+	ret = i915_gem_gtt_prepare_pages(obj, st);
+	if (ret) {
+		sg_free_table(st);
+
+		if (max_segment > PAGE_SIZE) {
+			max_segment = PAGE_SIZE;
+			goto alloc_table;
+		}
+
+		goto err;
+	}
+
+	WARN_ON_ONCE(!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_WRITE));
+	if (i915_gem_object_can_bypass_llc(obj))
+		obj->cache_dirty = true;
+
+	sg_page_sizes = i915_sg_dma_sizes(st->sgl);
+	__i915_gem_object_set_pages(obj, st, sg_page_sizes);
+
+	return 0;
+
+err:
+	i915_gem_object_userptr_drop_ref(obj);
+err_free:
+	kfree(st);
+	return ret;
+}
+
+static void
+i915_gem_userptr_put_pages(struct drm_i915_gem_object *obj,
+			   struct sg_table *pages)
+{
+	struct sgt_iter sgt_iter;
+	struct page *page;
+
+	if (!pages)
+		return;
+
+	__i915_gem_object_release_shmem(obj, pages, true);
+	i915_gem_gtt_finish_pages(obj, pages);
+
+	/*
+	 * We always mark objects as dirty when they are used by the GPU,
+	 * just in case. However, if we set the vma as being read-only we know
+	 * that the object will never have been written to.
+	 */
+	if (i915_gem_object_is_readonly(obj))
+		obj->mm.dirty = false;
+
+	for_each_sgt_page(page, sgt_iter, pages) {
+		if (obj->mm.dirty && trylock_page(page)) {
+			/*
+			 * As this may not be anonymous memory (e.g. shmem)
+			 * but exist on a real mapping, we have to lock
+			 * the page in order to dirty it -- holding
+			 * the page reference is not sufficient to
+			 * prevent the inode from being truncated.
+			 * Play safe and take the lock.
+			 *
+			 * However...!
+			 *
+			 * The mmu-notifier can be invalidated for a
+			 * migrate_folio, that is alreadying holding the lock
+			 * on the folio. Such a try_to_unmap() will result
+			 * in us calling put_pages() and so recursively try
+			 * to lock the page. We avoid that deadlock with
+			 * a trylock_page() and in exchange we risk missing
+			 * some page dirtying.
+			 */
+			set_page_dirty(page);
+			unlock_page(page);
+		}
+
+		mark_page_accessed(page);
+	}
+	obj->mm.dirty = false;
+
+	sg_free_table(pages);
+	kfree(pages);
+
+	i915_gem_object_userptr_drop_ref(obj);
+}
+
+static int i915_gem_object_userptr_unbind(struct drm_i915_gem_object *obj)
+{
+	struct sg_table *pages;
+	int err;
+
+	err = i915_gem_object_unbind(obj, I915_GEM_OBJECT_UNBIND_ACTIVE);
+	if (err)
+		return err;
+
+	if (GEM_WARN_ON(i915_gem_object_has_pinned_pages(obj)))
+		return -EBUSY;
+
+	assert_object_held(obj);
+
+	pages = __i915_gem_object_unset_pages(obj);
+	if (!IS_ERR_OR_NULL(pages))
+		i915_gem_userptr_put_pages(obj, pages);
+
+	return err;
+}
+
+int i915_gem_object_userptr_submit_init(struct drm_i915_gem_object *obj)
+{
+	const unsigned long num_pages = obj->base.size >> PAGE_SHIFT;
+	struct page **pvec;
+	unsigned int gup_flags = 0;
+	unsigned long notifier_seq;
+	int pinned, ret;
+
+	if (obj->userptr.notifier.mm != current->mm)
+		return -EFAULT;
+
+	notifier_seq = mmu_interval_read_begin(&obj->userptr.notifier);
+
+	ret = i915_gem_object_lock_interruptible(obj, NULL);
+	if (ret)
+		return ret;
+
+	if (notifier_seq == obj->userptr.notifier_seq && obj->userptr.pvec) {
+		i915_gem_object_unlock(obj);
+		return 0;
+	}
+
+	ret = i915_gem_object_userptr_unbind(obj);
+	i915_gem_object_unlock(obj);
+	if (ret)
+		return ret;
+
+	pvec = kvmalloc_array(num_pages, sizeof(struct page *), GFP_KERNEL);
+	if (!pvec)
+		return -ENOMEM;
+
+	if (!i915_gem_object_is_readonly(obj))
+		gup_flags |= FOLL_WRITE;
+
+	pinned = ret = 0;
+	while (pinned < num_pages) {
+		ret = pin_user_pages_fast(obj->userptr.ptr + pinned * PAGE_SIZE,
+					  num_pages - pinned, gup_flags,
+					  &pvec[pinned]);
+		if (ret < 0)
+			goto out;
+
+		pinned += ret;
+	}
+	ret = 0;
+
+	ret = i915_gem_object_lock_interruptible(obj, NULL);
+	if (ret)
+		goto out;
+
+	if (mmu_interval_read_retry(&obj->userptr.notifier,
+		!obj->userptr.page_ref ? notifier_seq :
+		obj->userptr.notifier_seq)) {
+		ret = -EAGAIN;
+		goto out_unlock;
+	}
+
+	if (!obj->userptr.page_ref++) {
+		obj->userptr.pvec = pvec;
+		obj->userptr.notifier_seq = notifier_seq;
+		pvec = NULL;
+		ret = ____i915_gem_object_get_pages(obj);
+	}
+
+	obj->userptr.page_ref--;
+
+out_unlock:
+	i915_gem_object_unlock(obj);
+
+out:
+	if (pvec) {
+		unpin_user_pages(pvec, pinned);
+		kvfree(pvec);
+	}
+
+	return ret;
+}
+
+int i915_gem_object_userptr_submit_done(struct drm_i915_gem_object *obj)
+{
+	if (mmu_interval_read_retry(&obj->userptr.notifier,
+				    obj->userptr.notifier_seq)) {
+		/* We collided with the mmu notifier, need to retry */
+
+		return -EAGAIN;
+	}
+
+	return 0;
+}
+
+int i915_gem_object_userptr_validate(struct drm_i915_gem_object *obj)
+{
+	int err;
+
+	err = i915_gem_object_userptr_submit_init(obj);
+	if (err)
+		return err;
+
+	err = i915_gem_object_lock_interruptible(obj, NULL);
+	if (!err) {
+		/*
+		 * Since we only check validity, not use the pages,
+		 * it doesn't matter if we collide with the mmu notifier,
+		 * and -EAGAIN handling is not required.
+		 */
+		err = i915_gem_object_pin_pages(obj);
+		if (!err)
+			i915_gem_object_unpin_pages(obj);
+
+		i915_gem_object_unlock(obj);
+	}
+
+	return err;
+}
+
+static void
+i915_gem_userptr_release(struct drm_i915_gem_object *obj)
+{
+	GEM_WARN_ON(obj->userptr.page_ref);
+
+	mmu_interval_notifier_remove(&obj->userptr.notifier);
+	obj->userptr.notifier.mm = NULL;
+}
+
+static int
+i915_gem_userptr_dmabuf_export(struct drm_i915_gem_object *obj)
+{
+	drm_dbg(obj->base.dev, "Exporting userptr no longer allowed\n");
+
+	return -EINVAL;
+}
+
+static int
+i915_gem_userptr_pwrite(struct drm_i915_gem_object *obj,
+			const struct drm_i915_gem_pwrite *args)
+{
+	drm_dbg(obj->base.dev, "pwrite to userptr no longer allowed\n");
+
+	return -EINVAL;
+}
+
+static int
+i915_gem_userptr_pread(struct drm_i915_gem_object *obj,
+		       const struct drm_i915_gem_pread *args)
+{
+	drm_dbg(obj->base.dev, "pread from userptr no longer allowed\n");
+
+	return -EINVAL;
+}
+
+static const struct drm_i915_gem_object_ops i915_gem_userptr_ops = {
+	.name = "i915_gem_object_userptr",
+	.flags = I915_GEM_OBJECT_IS_SHRINKABLE |
+		 I915_GEM_OBJECT_NO_MMAP |
+		 I915_GEM_OBJECT_IS_PROXY,
+	.get_pages = i915_gem_userptr_get_pages,
+	.put_pages = i915_gem_userptr_put_pages,
+	.dmabuf_export = i915_gem_userptr_dmabuf_export,
+	.pwrite = i915_gem_userptr_pwrite,
+	.pread = i915_gem_userptr_pread,
+	.release = i915_gem_userptr_release,
+};
+
+#endif
+
+static int
+probe_range(struct mm_struct *mm, unsigned long addr, unsigned long len)
+{
+	VMA_ITERATOR(vmi, mm, addr);
+	struct vm_area_struct *vma;
+	unsigned long end = addr + len;
+
+	mmap_read_lock(mm);
+	for_each_vma_range(vmi, vma, end) {
+		/* Check for holes, note that we also update the addr below */
+		if (vma->vm_start > addr)
+			break;
+
+		if (vma->vm_flags & (VM_PFNMAP | VM_MIXEDMAP))
+			break;
+
+		addr = vma->vm_end;
+	}
+	mmap_read_unlock(mm);
+
+	if (vma || addr < end)
+		return -EFAULT;
+	return 0;
+}
+
+/*
+ * Creates a new mm object that wraps some normal memory from the process
+ * context - user memory.
+ *
+ * We impose several restrictions upon the memory being mapped
+ * into the GPU.
+ * 1. It must be page aligned (both start/end addresses, i.e ptr and size).
+ * 2. It must be normal system memory, not a pointer into another map of IO
+ *    space (e.g. it must not be a GTT mmapping of another object).
+ * 3. We only allow a bo as large as we could in theory map into the GTT,
+ *    that is we limit the size to the total size of the GTT.
+ * 4. The bo is marked as being snoopable. The backing pages are left
+ *    accessible directly by the CPU, but reads and writes by the GPU may
+ *    incur the cost of a snoop (unless you have an LLC architecture).
+ *
+ * Synchronisation between multiple users and the GPU is left to userspace
+ * through the normal set-domain-ioctl. The kernel will enforce that the
+ * GPU relinquishes the VMA before it is returned back to the system
+ * i.e. upon free(), munmap() or process termination. However, the userspace
+ * malloc() library may not immediately relinquish the VMA after free() and
+ * instead reuse it whilst the GPU is still reading and writing to the VMA.
+ * Caveat emptor.
+ *
+ * Also note, that the object created here is not currently a "first class"
+ * object, in that several ioctls are banned. These are the CPU access
+ * ioctls: mmap(), pwrite and pread. In practice, you are expected to use
+ * direct access via your pointer rather than use those ioctls. Another
+ * restriction is that we do not allow userptr surfaces to be pinned to the
+ * hardware and so we reject any attempt to create a framebuffer out of a
+ * userptr.
+ *
+ * If you think this is a good interface to use to pass GPU memory between
+ * drivers, please use dma-buf instead. In fact, wherever possible use
+ * dma-buf instead.
+ */
+int
+i915_gem_userptr_ioctl(struct drm_device *dev,
+		       void *data,
+		       struct drm_file *file)
+{
+	static struct lock_class_key __maybe_unused lock_class;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_i915_gem_userptr *args = data;
+	struct drm_i915_gem_object __maybe_unused *obj;
+	int __maybe_unused ret;
+	u32 __maybe_unused handle;
+
+	if (!HAS_LLC(dev_priv) && !HAS_SNOOP(dev_priv)) {
+		/* We cannot support coherent userptr objects on hw without
+		 * LLC and broken snooping.
+		 */
+		return -ENODEV;
+	}
+
+	if (args->flags & ~(I915_USERPTR_READ_ONLY |
+			    I915_USERPTR_UNSYNCHRONIZED |
+			    I915_USERPTR_PROBE))
+		return -EINVAL;
+
+	if (i915_gem_object_size_2big(args->user_size))
+		return -E2BIG;
+
+	if (!args->user_size)
+		return -EINVAL;
+
+	if (offset_in_page(args->user_ptr | args->user_size))
+		return -EINVAL;
+
+	if (!access_ok((char __user *)(unsigned long)args->user_ptr, args->user_size))
+		return -EFAULT;
+
+	if (args->flags & I915_USERPTR_UNSYNCHRONIZED)
+		return -ENODEV;
+
+	if (args->flags & I915_USERPTR_READ_ONLY) {
+		/*
+		 * On almost all of the older hw, we cannot tell the GPU that
+		 * a page is readonly.
+		 */
+		if (!to_gt(dev_priv)->vm->has_read_only)
+			return -ENODEV;
+	}
+
+	if (args->flags & I915_USERPTR_PROBE) {
+		/*
+		 * Check that the range pointed to represents real struct
+		 * pages and not iomappings (at this moment in time!)
+		 */
+		ret = probe_range(current->mm, args->user_ptr, args->user_size);
+		if (ret)
+			return ret;
+	}
+
+#ifdef CONFIG_MMU_NOTIFIER
+	obj = i915_gem_object_alloc();
+	if (obj == NULL)
+		return -ENOMEM;
+
+	drm_gem_private_object_init(dev, &obj->base, args->user_size);
+	i915_gem_object_init(obj, &i915_gem_userptr_ops, &lock_class,
+			     I915_BO_ALLOC_USER);
+	obj->mem_flags = I915_BO_FLAG_STRUCT_PAGE;
+	obj->read_domains = I915_GEM_DOMAIN_CPU;
+	obj->write_domain = I915_GEM_DOMAIN_CPU;
+	i915_gem_object_set_cache_coherency(obj, I915_CACHE_LLC);
+
+	obj->userptr.ptr = args->user_ptr;
+	obj->userptr.notifier_seq = ULONG_MAX;
+	if (args->flags & I915_USERPTR_READ_ONLY)
+		i915_gem_object_set_readonly(obj);
+
+	/* And keep a pointer to the current->mm for resolving the user pages
+	 * at binding. This means that we need to hook into the mmu_notifier
+	 * in order to detect if the mmu is destroyed.
+	 */
+	ret = i915_gem_userptr_init__mmu_notifier(obj);
+	if (ret == 0)
+		ret = drm_gem_handle_create(file, &obj->base, &handle);
+
+	/* drop reference from allocate - handle holds it now */
+	i915_gem_object_put(obj);
+	if (ret)
+		return ret;
+
+	args->handle = handle;
+	return 0;
+#else
+	return -ENODEV;
+#endif
+}
+
+int i915_gem_init_userptr(struct drm_i915_private *dev_priv)
+{
+#ifdef CONFIG_MMU_NOTIFIER
+	rwlock_init(&dev_priv->mm.notifier_lock);
+#endif
+
+	return 0;
+}
+
+void i915_gem_cleanup_userptr(struct drm_i915_private *dev_priv)
+{
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_userptr.h b/drivers/gpu/drm/i915/gem/i915_gem_userptr.h
new file mode 100644
index 000000000..8dadb2f84
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_userptr.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_GEM_USERPTR_H__
+#define __I915_GEM_USERPTR_H__
+
+struct drm_i915_private;
+
+int i915_gem_init_userptr(struct drm_i915_private *dev_priv);
+void i915_gem_cleanup_userptr(struct drm_i915_private *dev_priv);
+
+#endif /* __I915_GEM_USERPTR_H__ */
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_wait.c b/drivers/gpu/drm/i915/gem/i915_gem_wait.c
new file mode 100644
index 000000000..e6e01c2a7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gem_wait.c
@@ -0,0 +1,295 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include <linux/dma-fence-array.h>
+#include <linux/dma-fence-chain.h>
+#include <linux/jiffies.h>
+
+#include "gt/intel_engine.h"
+#include "gt/intel_rps.h"
+
+#include "i915_gem_ioctls.h"
+#include "i915_gem_object.h"
+
+static long
+i915_gem_object_wait_fence(struct dma_fence *fence,
+			   unsigned int flags,
+			   long timeout)
+{
+	BUILD_BUG_ON(I915_WAIT_INTERRUPTIBLE != 0x1);
+
+	if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags))
+		return timeout;
+
+	if (dma_fence_is_i915(fence))
+		return i915_request_wait_timeout(to_request(fence), flags, timeout);
+
+	return dma_fence_wait_timeout(fence,
+				      flags & I915_WAIT_INTERRUPTIBLE,
+				      timeout);
+}
+
+static void
+i915_gem_object_boost(struct dma_resv *resv, unsigned int flags)
+{
+	struct dma_resv_iter cursor;
+	struct dma_fence *fence;
+
+	/*
+	 * Prescan all fences for potential boosting before we begin waiting.
+	 *
+	 * When we wait, we wait on outstanding fences serially. If the
+	 * dma-resv contains a sequence such as 1:1, 1:2 instead of a reduced
+	 * form 1:2, then as we look at each wait in turn we see that each
+	 * request is currently executing and not worthy of boosting. But if
+	 * we only happen to look at the final fence in the sequence (because
+	 * of request coalescing or splitting between read/write arrays by
+	 * the iterator), then we would boost. As such our decision to boost
+	 * or not is delicately balanced on the order we wait on fences.
+	 *
+	 * So instead of looking for boosts sequentially, look for all boosts
+	 * upfront and then wait on the outstanding fences.
+	 */
+
+	dma_resv_iter_begin(&cursor, resv,
+			    dma_resv_usage_rw(flags & I915_WAIT_ALL));
+	dma_resv_for_each_fence_unlocked(&cursor, fence)
+		if (dma_fence_is_i915(fence) &&
+		    !i915_request_started(to_request(fence)))
+			intel_rps_boost(to_request(fence));
+	dma_resv_iter_end(&cursor);
+}
+
+static long
+i915_gem_object_wait_reservation(struct dma_resv *resv,
+				 unsigned int flags,
+				 long timeout)
+{
+	struct dma_resv_iter cursor;
+	struct dma_fence *fence;
+	long ret = timeout ?: 1;
+
+	i915_gem_object_boost(resv, flags);
+
+	dma_resv_iter_begin(&cursor, resv,
+			    dma_resv_usage_rw(flags & I915_WAIT_ALL));
+	dma_resv_for_each_fence_unlocked(&cursor, fence) {
+		ret = i915_gem_object_wait_fence(fence, flags, timeout);
+		if (ret <= 0)
+			break;
+
+		if (timeout)
+			timeout = ret;
+	}
+	dma_resv_iter_end(&cursor);
+
+	return ret;
+}
+
+static void fence_set_priority(struct dma_fence *fence,
+			       const struct i915_sched_attr *attr)
+{
+	struct i915_request *rq;
+	struct intel_engine_cs *engine;
+
+	if (dma_fence_is_signaled(fence) || !dma_fence_is_i915(fence))
+		return;
+
+	rq = to_request(fence);
+	engine = rq->engine;
+
+	rcu_read_lock(); /* RCU serialisation for set-wedged protection */
+	if (engine->sched_engine->schedule)
+		engine->sched_engine->schedule(rq, attr);
+	rcu_read_unlock();
+}
+
+static inline bool __dma_fence_is_chain(const struct dma_fence *fence)
+{
+	return fence->ops == &dma_fence_chain_ops;
+}
+
+void i915_gem_fence_wait_priority(struct dma_fence *fence,
+				  const struct i915_sched_attr *attr)
+{
+	if (dma_fence_is_signaled(fence))
+		return;
+
+	local_bh_disable();
+
+	/* Recurse once into a fence-array */
+	if (dma_fence_is_array(fence)) {
+		struct dma_fence_array *array = to_dma_fence_array(fence);
+		int i;
+
+		for (i = 0; i < array->num_fences; i++)
+			fence_set_priority(array->fences[i], attr);
+	} else if (__dma_fence_is_chain(fence)) {
+		struct dma_fence *iter;
+
+		/* The chain is ordered; if we boost the last, we boost all */
+		dma_fence_chain_for_each(iter, fence) {
+			fence_set_priority(to_dma_fence_chain(iter)->fence,
+					   attr);
+			break;
+		}
+		dma_fence_put(iter);
+	} else {
+		fence_set_priority(fence, attr);
+	}
+
+	local_bh_enable(); /* kick the tasklets if queues were reprioritised */
+}
+
+int
+i915_gem_object_wait_priority(struct drm_i915_gem_object *obj,
+			      unsigned int flags,
+			      const struct i915_sched_attr *attr)
+{
+	struct dma_resv_iter cursor;
+	struct dma_fence *fence;
+
+	dma_resv_iter_begin(&cursor, obj->base.resv,
+			    dma_resv_usage_rw(flags & I915_WAIT_ALL));
+	dma_resv_for_each_fence_unlocked(&cursor, fence)
+		i915_gem_fence_wait_priority(fence, attr);
+	dma_resv_iter_end(&cursor);
+	return 0;
+}
+
+/**
+ * Waits for rendering to the object to be completed
+ * @obj: i915 gem object
+ * @flags: how to wait (under a lock, for all rendering or just for writes etc)
+ * @timeout: how long to wait
+ */
+int
+i915_gem_object_wait(struct drm_i915_gem_object *obj,
+		     unsigned int flags,
+		     long timeout)
+{
+	might_sleep();
+	GEM_BUG_ON(timeout < 0);
+
+	timeout = i915_gem_object_wait_reservation(obj->base.resv,
+						   flags, timeout);
+
+	if (timeout < 0)
+		return timeout;
+
+	return !timeout ? -ETIME : 0;
+}
+
+static inline unsigned long nsecs_to_jiffies_timeout(const u64 n)
+{
+	/* nsecs_to_jiffies64() does not guard against overflow */
+	if (NSEC_PER_SEC % HZ &&
+	    div_u64(n, NSEC_PER_SEC) >= MAX_JIFFY_OFFSET / HZ)
+		return MAX_JIFFY_OFFSET;
+
+	return min_t(u64, MAX_JIFFY_OFFSET, nsecs_to_jiffies64(n) + 1);
+}
+
+static unsigned long to_wait_timeout(s64 timeout_ns)
+{
+	if (timeout_ns < 0)
+		return MAX_SCHEDULE_TIMEOUT;
+
+	if (timeout_ns == 0)
+		return 0;
+
+	return nsecs_to_jiffies_timeout(timeout_ns);
+}
+
+/**
+ * i915_gem_wait_ioctl - implements DRM_IOCTL_I915_GEM_WAIT
+ * @dev: drm device pointer
+ * @data: ioctl data blob
+ * @file: drm file pointer
+ *
+ * Returns 0 if successful, else an error is returned with the remaining time in
+ * the timeout parameter.
+ *  -ETIME: object is still busy after timeout
+ *  -ERESTARTSYS: signal interrupted the wait
+ *  -ENONENT: object doesn't exist
+ * Also possible, but rare:
+ *  -EAGAIN: incomplete, restart syscall
+ *  -ENOMEM: damn
+ *  -ENODEV: Internal IRQ fail
+ *  -E?: The add request failed
+ *
+ * The wait ioctl with a timeout of 0 reimplements the busy ioctl. With any
+ * non-zero timeout parameter the wait ioctl will wait for the given number of
+ * nanoseconds on an object becoming unbusy. Since the wait itself does so
+ * without holding struct_mutex the object may become re-busied before this
+ * function completes. A similar but shorter * race condition exists in the busy
+ * ioctl
+ */
+int
+i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
+{
+	struct drm_i915_gem_wait *args = data;
+	struct drm_i915_gem_object *obj;
+	ktime_t start;
+	long ret;
+
+	if (args->flags != 0)
+		return -EINVAL;
+
+	obj = i915_gem_object_lookup(file, args->bo_handle);
+	if (!obj)
+		return -ENOENT;
+
+	start = ktime_get();
+
+	ret = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE |
+				   I915_WAIT_PRIORITY |
+				   I915_WAIT_ALL,
+				   to_wait_timeout(args->timeout_ns));
+
+	if (args->timeout_ns > 0) {
+		args->timeout_ns -= ktime_to_ns(ktime_sub(ktime_get(), start));
+		if (args->timeout_ns < 0)
+			args->timeout_ns = 0;
+
+		/*
+		 * Apparently ktime isn't accurate enough and occasionally has a
+		 * bit of mismatch in the jiffies<->nsecs<->ktime loop. So patch
+		 * things up to make the test happy. We allow up to 1 jiffy.
+		 *
+		 * This is a regression from the timespec->ktime conversion.
+		 */
+		if (ret == -ETIME && !nsecs_to_jiffies(args->timeout_ns))
+			args->timeout_ns = 0;
+
+		/* Asked to wait beyond the jiffie/scheduler precision? */
+		if (ret == -ETIME && args->timeout_ns)
+			ret = -EAGAIN;
+	}
+
+	i915_gem_object_put(obj);
+	return ret;
+}
+
+/**
+ * i915_gem_object_wait_migration - Sync an accelerated migration operation
+ * @obj: The migrating object.
+ * @flags: waiting flags. Currently supports only I915_WAIT_INTERRUPTIBLE.
+ *
+ * Wait for any pending async migration operation on the object,
+ * whether it's explicitly (i915_gem_object_migrate()) or implicitly
+ * (swapin, initial clearing) initiated.
+ *
+ * Return: 0 if successful, -ERESTARTSYS if a signal was hit during waiting.
+ */
+int i915_gem_object_wait_migration(struct drm_i915_gem_object *obj,
+				   unsigned int flags)
+{
+	might_sleep();
+
+	return i915_gem_object_wait_moving_fence(obj, !!(flags & I915_WAIT_INTERRUPTIBLE));
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gemfs.c b/drivers/gpu/drm/i915/gem/i915_gemfs.c
new file mode 100644
index 000000000..46b9a17d6
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gemfs.c
@@ -0,0 +1,60 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2017 Intel Corporation
+ */
+
+#include <linux/fs.h>
+#include <linux/mount.h>
+
+#include "i915_drv.h"
+#include "i915_gemfs.h"
+#include "i915_utils.h"
+
+void i915_gemfs_init(struct drm_i915_private *i915)
+{
+	char huge_opt[] = "huge=within_size"; /* r/w */
+	struct file_system_type *type;
+	struct vfsmount *gemfs;
+
+	/*
+	 * By creating our own shmemfs mountpoint, we can pass in
+	 * mount flags that better match our usecase.
+	 *
+	 * One example, although it is probably better with a per-file
+	 * control, is selecting huge page allocations ("huge=within_size").
+	 * However, we only do so on platforms which benefit from it, or to
+	 * offset the overhead of iommu lookups, where with latter it is a net
+	 * win even on platforms which would otherwise see some performance
+	 * regressions such a slow reads issue on Broadwell and Skylake.
+	 */
+
+	if (GRAPHICS_VER(i915) < 11 && !i915_vtd_active(i915))
+		return;
+
+	if (!IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE))
+		goto err;
+
+	type = get_fs_type("tmpfs");
+	if (!type)
+		goto err;
+
+	gemfs = vfs_kern_mount(type, SB_KERNMOUNT, type->name, huge_opt);
+	if (IS_ERR(gemfs))
+		goto err;
+
+	i915->mm.gemfs = gemfs;
+	drm_info(&i915->drm, "Using Transparent Hugepages\n");
+	return;
+
+err:
+	drm_notice(&i915->drm,
+		   "Transparent Hugepage support is recommended for optimal performance%s\n",
+		   GRAPHICS_VER(i915) >= 11 ? " on this platform!" :
+					      " when IOMMU is enabled!");
+}
+
+void i915_gemfs_fini(struct drm_i915_private *i915)
+{
+	kern_unmount(i915->mm.gemfs);
+}
diff --git a/drivers/gpu/drm/i915/gem/i915_gemfs.h b/drivers/gpu/drm/i915/gem/i915_gemfs.h
new file mode 100644
index 000000000..5d835e44c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/i915_gemfs.h
@@ -0,0 +1,15 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2017 Intel Corporation
+ */
+
+#ifndef __I915_GEMFS_H__
+#define __I915_GEMFS_H__
+
+struct drm_i915_private;
+
+void i915_gemfs_init(struct drm_i915_private *i915);
+void i915_gemfs_fini(struct drm_i915_private *i915);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gem/selftests/huge_gem_object.c b/drivers/gpu/drm/i915/gem/selftests/huge_gem_object.c
new file mode 100644
index 000000000..f963b8e1e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/huge_gem_object.c
@@ -0,0 +1,127 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include "i915_scatterlist.h"
+
+#include "huge_gem_object.h"
+
+static void huge_free_pages(struct drm_i915_gem_object *obj,
+			    struct sg_table *pages)
+{
+	unsigned long nreal = obj->scratch / PAGE_SIZE;
+	struct sgt_iter sgt_iter;
+	struct page *page;
+
+	for_each_sgt_page(page, sgt_iter, pages) {
+		__free_page(page);
+		if (!--nreal)
+			break;
+	}
+
+	sg_free_table(pages);
+	kfree(pages);
+}
+
+static int huge_get_pages(struct drm_i915_gem_object *obj)
+{
+#define GFP (GFP_KERNEL | __GFP_NOWARN | __GFP_RETRY_MAYFAIL)
+	const unsigned long nreal = obj->scratch / PAGE_SIZE;
+	const unsigned long npages = obj->base.size / PAGE_SIZE;
+	struct scatterlist *sg, *src, *end;
+	struct sg_table *pages;
+	unsigned long n;
+
+	pages = kmalloc(sizeof(*pages), GFP);
+	if (!pages)
+		return -ENOMEM;
+
+	if (sg_alloc_table(pages, npages, GFP)) {
+		kfree(pages);
+		return -ENOMEM;
+	}
+
+	sg = pages->sgl;
+	for (n = 0; n < nreal; n++) {
+		struct page *page;
+
+		page = alloc_page(GFP | __GFP_HIGHMEM);
+		if (!page) {
+			sg_mark_end(sg);
+			goto err;
+		}
+
+		sg_set_page(sg, page, PAGE_SIZE, 0);
+		sg = __sg_next(sg);
+	}
+	if (nreal < npages) {
+		for (end = sg, src = pages->sgl; sg; sg = __sg_next(sg)) {
+			sg_set_page(sg, sg_page(src), PAGE_SIZE, 0);
+			src = __sg_next(src);
+			if (src == end)
+				src = pages->sgl;
+		}
+	}
+
+	if (i915_gem_gtt_prepare_pages(obj, pages))
+		goto err;
+
+	__i915_gem_object_set_pages(obj, pages, PAGE_SIZE);
+
+	return 0;
+
+err:
+	huge_free_pages(obj, pages);
+	return -ENOMEM;
+#undef GFP
+}
+
+static void huge_put_pages(struct drm_i915_gem_object *obj,
+			   struct sg_table *pages)
+{
+	i915_gem_gtt_finish_pages(obj, pages);
+	huge_free_pages(obj, pages);
+
+	obj->mm.dirty = false;
+}
+
+static const struct drm_i915_gem_object_ops huge_ops = {
+	.name = "huge-gem",
+	.get_pages = huge_get_pages,
+	.put_pages = huge_put_pages,
+};
+
+struct drm_i915_gem_object *
+huge_gem_object(struct drm_i915_private *i915,
+		phys_addr_t phys_size,
+		dma_addr_t dma_size)
+{
+	static struct lock_class_key lock_class;
+	struct drm_i915_gem_object *obj;
+	unsigned int cache_level;
+
+	GEM_BUG_ON(!phys_size || phys_size > dma_size);
+	GEM_BUG_ON(!IS_ALIGNED(phys_size, PAGE_SIZE));
+	GEM_BUG_ON(!IS_ALIGNED(dma_size, I915_GTT_PAGE_SIZE));
+
+	if (overflows_type(dma_size, obj->base.size))
+		return ERR_PTR(-E2BIG);
+
+	obj = i915_gem_object_alloc();
+	if (!obj)
+		return ERR_PTR(-ENOMEM);
+
+	drm_gem_private_object_init(&i915->drm, &obj->base, dma_size);
+	i915_gem_object_init(obj, &huge_ops, &lock_class, 0);
+	obj->mem_flags |= I915_BO_FLAG_STRUCT_PAGE;
+
+	obj->read_domains = I915_GEM_DOMAIN_CPU;
+	obj->write_domain = I915_GEM_DOMAIN_CPU;
+	cache_level = HAS_LLC(i915) ? I915_CACHE_LLC : I915_CACHE_NONE;
+	i915_gem_object_set_cache_coherency(obj, cache_level);
+	obj->scratch = phys_size;
+
+	return obj;
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/huge_gem_object.h b/drivers/gpu/drm/i915/gem/selftests/huge_gem_object.h
new file mode 100644
index 000000000..b8cf31b7b
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/huge_gem_object.h
@@ -0,0 +1,33 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __HUGE_GEM_OBJECT_H
+#define __HUGE_GEM_OBJECT_H
+
+#include <linux/types.h>
+
+#include "gem/i915_gem_object_types.h"
+
+struct drm_i915_private;
+
+struct drm_i915_gem_object *
+huge_gem_object(struct drm_i915_private *i915,
+		phys_addr_t phys_size,
+		dma_addr_t dma_size);
+
+static inline phys_addr_t
+huge_gem_object_phys_size(struct drm_i915_gem_object *obj)
+{
+	return obj->scratch;
+}
+
+static inline dma_addr_t
+huge_gem_object_dma_size(struct drm_i915_gem_object *obj)
+{
+	return obj->base.size;
+}
+
+#endif /* !__HUGE_GEM_OBJECT_H */
diff --git a/drivers/gpu/drm/i915/gem/selftests/huge_pages.c b/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
new file mode 100644
index 000000000..02fe7ea8c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
@@ -0,0 +1,1819 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2017 Intel Corporation
+ */
+
+#include <linux/prime_numbers.h>
+#include <linux/string_helpers.h>
+#include <linux/swap.h>
+
+#include "i915_selftest.h"
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_lmem.h"
+#include "gem/i915_gem_pm.h"
+#include "gem/i915_gem_region.h"
+
+#include "gt/intel_gt.h"
+
+#include "igt_gem_utils.h"
+#include "mock_context.h"
+
+#include "selftests/mock_drm.h"
+#include "selftests/mock_gem_device.h"
+#include "selftests/mock_region.h"
+#include "selftests/i915_random.h"
+
+static struct i915_gem_context *hugepage_ctx(struct drm_i915_private *i915,
+					     struct file *file)
+{
+	struct i915_gem_context *ctx = live_context(i915, file);
+	struct i915_address_space *vm;
+
+	if (IS_ERR(ctx))
+		return ctx;
+
+	vm = ctx->vm;
+	if (vm)
+		WRITE_ONCE(vm->scrub_64K, true);
+
+	return ctx;
+}
+
+static const unsigned int page_sizes[] = {
+	I915_GTT_PAGE_SIZE_2M,
+	I915_GTT_PAGE_SIZE_64K,
+	I915_GTT_PAGE_SIZE_4K,
+};
+
+static unsigned int get_largest_page_size(struct drm_i915_private *i915,
+					  u64 rem)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(page_sizes); ++i) {
+		unsigned int page_size = page_sizes[i];
+
+		if (HAS_PAGE_SIZES(i915, page_size) && rem >= page_size)
+			return page_size;
+	}
+
+	return 0;
+}
+
+static void huge_pages_free_pages(struct sg_table *st)
+{
+	struct scatterlist *sg;
+
+	for (sg = st->sgl; sg; sg = __sg_next(sg)) {
+		if (sg_page(sg))
+			__free_pages(sg_page(sg), get_order(sg->length));
+	}
+
+	sg_free_table(st);
+	kfree(st);
+}
+
+static int get_huge_pages(struct drm_i915_gem_object *obj)
+{
+#define GFP (GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY)
+	unsigned int page_mask = obj->mm.page_mask;
+	struct sg_table *st;
+	struct scatterlist *sg;
+	unsigned int sg_page_sizes;
+	u64 rem;
+
+	st = kmalloc(sizeof(*st), GFP);
+	if (!st)
+		return -ENOMEM;
+
+	if (sg_alloc_table(st, obj->base.size >> PAGE_SHIFT, GFP)) {
+		kfree(st);
+		return -ENOMEM;
+	}
+
+	rem = obj->base.size;
+	sg = st->sgl;
+	st->nents = 0;
+	sg_page_sizes = 0;
+
+	/*
+	 * Our goal here is simple, we want to greedily fill the object from
+	 * largest to smallest page-size, while ensuring that we use *every*
+	 * page-size as per the given page-mask.
+	 */
+	do {
+		unsigned int bit = ilog2(page_mask);
+		unsigned int page_size = BIT(bit);
+		int order = get_order(page_size);
+
+		do {
+			struct page *page;
+
+			GEM_BUG_ON(order >= MAX_ORDER);
+			page = alloc_pages(GFP | __GFP_ZERO, order);
+			if (!page)
+				goto err;
+
+			sg_set_page(sg, page, page_size, 0);
+			sg_page_sizes |= page_size;
+			st->nents++;
+
+			rem -= page_size;
+			if (!rem) {
+				sg_mark_end(sg);
+				break;
+			}
+
+			sg = __sg_next(sg);
+		} while ((rem - ((page_size-1) & page_mask)) >= page_size);
+
+		page_mask &= (page_size-1);
+	} while (page_mask);
+
+	if (i915_gem_gtt_prepare_pages(obj, st))
+		goto err;
+
+	GEM_BUG_ON(sg_page_sizes != obj->mm.page_mask);
+	__i915_gem_object_set_pages(obj, st, sg_page_sizes);
+
+	return 0;
+
+err:
+	sg_set_page(sg, NULL, 0, 0);
+	sg_mark_end(sg);
+	huge_pages_free_pages(st);
+
+	return -ENOMEM;
+}
+
+static void put_huge_pages(struct drm_i915_gem_object *obj,
+			   struct sg_table *pages)
+{
+	i915_gem_gtt_finish_pages(obj, pages);
+	huge_pages_free_pages(pages);
+
+	obj->mm.dirty = false;
+
+	__start_cpu_write(obj);
+}
+
+static const struct drm_i915_gem_object_ops huge_page_ops = {
+	.name = "huge-gem",
+	.flags = I915_GEM_OBJECT_IS_SHRINKABLE,
+	.get_pages = get_huge_pages,
+	.put_pages = put_huge_pages,
+};
+
+static struct drm_i915_gem_object *
+huge_pages_object(struct drm_i915_private *i915,
+		  u64 size,
+		  unsigned int page_mask)
+{
+	static struct lock_class_key lock_class;
+	struct drm_i915_gem_object *obj;
+	unsigned int cache_level;
+
+	GEM_BUG_ON(!size);
+	GEM_BUG_ON(!IS_ALIGNED(size, BIT(__ffs(page_mask))));
+
+	if (size >> PAGE_SHIFT > INT_MAX)
+		return ERR_PTR(-E2BIG);
+
+	if (overflows_type(size, obj->base.size))
+		return ERR_PTR(-E2BIG);
+
+	obj = i915_gem_object_alloc();
+	if (!obj)
+		return ERR_PTR(-ENOMEM);
+
+	drm_gem_private_object_init(&i915->drm, &obj->base, size);
+	i915_gem_object_init(obj, &huge_page_ops, &lock_class, 0);
+	obj->mem_flags |= I915_BO_FLAG_STRUCT_PAGE;
+	i915_gem_object_set_volatile(obj);
+
+	obj->write_domain = I915_GEM_DOMAIN_CPU;
+	obj->read_domains = I915_GEM_DOMAIN_CPU;
+
+	cache_level = HAS_LLC(i915) ? I915_CACHE_LLC : I915_CACHE_NONE;
+	i915_gem_object_set_cache_coherency(obj, cache_level);
+
+	obj->mm.page_mask = page_mask;
+
+	return obj;
+}
+
+static int fake_get_huge_pages(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	const u64 max_len = rounddown_pow_of_two(UINT_MAX);
+	struct sg_table *st;
+	struct scatterlist *sg;
+	unsigned int sg_page_sizes;
+	u64 rem;
+
+	st = kmalloc(sizeof(*st), GFP);
+	if (!st)
+		return -ENOMEM;
+
+	if (sg_alloc_table(st, obj->base.size >> PAGE_SHIFT, GFP)) {
+		kfree(st);
+		return -ENOMEM;
+	}
+
+	/* Use optimal page sized chunks to fill in the sg table */
+	rem = obj->base.size;
+	sg = st->sgl;
+	st->nents = 0;
+	sg_page_sizes = 0;
+	do {
+		unsigned int page_size = get_largest_page_size(i915, rem);
+		unsigned int len = min(page_size * div_u64(rem, page_size),
+				       max_len);
+
+		GEM_BUG_ON(!page_size);
+
+		sg->offset = 0;
+		sg->length = len;
+		sg_dma_len(sg) = len;
+		sg_dma_address(sg) = page_size;
+
+		sg_page_sizes |= len;
+
+		st->nents++;
+
+		rem -= len;
+		if (!rem) {
+			sg_mark_end(sg);
+			break;
+		}
+
+		sg = sg_next(sg);
+	} while (1);
+
+	i915_sg_trim(st);
+
+	__i915_gem_object_set_pages(obj, st, sg_page_sizes);
+
+	return 0;
+}
+
+static int fake_get_huge_pages_single(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct sg_table *st;
+	struct scatterlist *sg;
+	unsigned int page_size;
+
+	st = kmalloc(sizeof(*st), GFP);
+	if (!st)
+		return -ENOMEM;
+
+	if (sg_alloc_table(st, 1, GFP)) {
+		kfree(st);
+		return -ENOMEM;
+	}
+
+	sg = st->sgl;
+	st->nents = 1;
+
+	page_size = get_largest_page_size(i915, obj->base.size);
+	GEM_BUG_ON(!page_size);
+
+	sg->offset = 0;
+	sg->length = obj->base.size;
+	sg_dma_len(sg) = obj->base.size;
+	sg_dma_address(sg) = page_size;
+
+	__i915_gem_object_set_pages(obj, st, sg->length);
+
+	return 0;
+#undef GFP
+}
+
+static void fake_free_huge_pages(struct drm_i915_gem_object *obj,
+				 struct sg_table *pages)
+{
+	sg_free_table(pages);
+	kfree(pages);
+}
+
+static void fake_put_huge_pages(struct drm_i915_gem_object *obj,
+				struct sg_table *pages)
+{
+	fake_free_huge_pages(obj, pages);
+	obj->mm.dirty = false;
+}
+
+static const struct drm_i915_gem_object_ops fake_ops = {
+	.name = "fake-gem",
+	.flags = I915_GEM_OBJECT_IS_SHRINKABLE,
+	.get_pages = fake_get_huge_pages,
+	.put_pages = fake_put_huge_pages,
+};
+
+static const struct drm_i915_gem_object_ops fake_ops_single = {
+	.name = "fake-gem",
+	.flags = I915_GEM_OBJECT_IS_SHRINKABLE,
+	.get_pages = fake_get_huge_pages_single,
+	.put_pages = fake_put_huge_pages,
+};
+
+static struct drm_i915_gem_object *
+fake_huge_pages_object(struct drm_i915_private *i915, u64 size, bool single)
+{
+	static struct lock_class_key lock_class;
+	struct drm_i915_gem_object *obj;
+
+	GEM_BUG_ON(!size);
+	GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE));
+
+	if (size >> PAGE_SHIFT > UINT_MAX)
+		return ERR_PTR(-E2BIG);
+
+	if (overflows_type(size, obj->base.size))
+		return ERR_PTR(-E2BIG);
+
+	obj = i915_gem_object_alloc();
+	if (!obj)
+		return ERR_PTR(-ENOMEM);
+
+	drm_gem_private_object_init(&i915->drm, &obj->base, size);
+
+	if (single)
+		i915_gem_object_init(obj, &fake_ops_single, &lock_class, 0);
+	else
+		i915_gem_object_init(obj, &fake_ops, &lock_class, 0);
+
+	i915_gem_object_set_volatile(obj);
+
+	obj->write_domain = I915_GEM_DOMAIN_CPU;
+	obj->read_domains = I915_GEM_DOMAIN_CPU;
+	obj->cache_level = I915_CACHE_NONE;
+
+	return obj;
+}
+
+static int igt_check_page_sizes(struct i915_vma *vma)
+{
+	struct drm_i915_private *i915 = vma->vm->i915;
+	unsigned int supported = RUNTIME_INFO(i915)->page_sizes;
+	struct drm_i915_gem_object *obj = vma->obj;
+	int err;
+
+	/* We have to wait for the async bind to complete before our asserts */
+	err = i915_vma_sync(vma);
+	if (err)
+		return err;
+
+	if (!HAS_PAGE_SIZES(i915, vma->page_sizes.sg)) {
+		pr_err("unsupported page_sizes.sg=%u, supported=%u\n",
+		       vma->page_sizes.sg & ~supported, supported);
+		err = -EINVAL;
+	}
+
+	if (!HAS_PAGE_SIZES(i915, vma->resource->page_sizes_gtt)) {
+		pr_err("unsupported page_sizes.gtt=%u, supported=%u\n",
+		       vma->resource->page_sizes_gtt & ~supported, supported);
+		err = -EINVAL;
+	}
+
+	if (vma->page_sizes.phys != obj->mm.page_sizes.phys) {
+		pr_err("vma->page_sizes.phys(%u) != obj->mm.page_sizes.phys(%u)\n",
+		       vma->page_sizes.phys, obj->mm.page_sizes.phys);
+		err = -EINVAL;
+	}
+
+	if (vma->page_sizes.sg != obj->mm.page_sizes.sg) {
+		pr_err("vma->page_sizes.sg(%u) != obj->mm.page_sizes.sg(%u)\n",
+		       vma->page_sizes.sg, obj->mm.page_sizes.sg);
+		err = -EINVAL;
+	}
+
+	/*
+	 * The dma-api is like a box of chocolates when it comes to the
+	 * alignment of dma addresses, however for LMEM we have total control
+	 * and so can guarantee alignment, likewise when we allocate our blocks
+	 * they should appear in descending order, and if we know that we align
+	 * to the largest page size for the GTT address, we should be able to
+	 * assert that if we see 2M physical pages then we should also get 2M
+	 * GTT pages. If we don't then something might be wrong in our
+	 * construction of the backing pages.
+	 *
+	 * Maintaining alignment is required to utilise huge pages in the ppGGT.
+	 */
+	if (i915_gem_object_is_lmem(obj) &&
+	    IS_ALIGNED(vma->node.start, SZ_2M) &&
+	    vma->page_sizes.sg & SZ_2M &&
+	    vma->resource->page_sizes_gtt < SZ_2M) {
+		pr_err("gtt pages mismatch for LMEM, expected 2M GTT pages, sg(%u), gtt(%u)\n",
+		       vma->page_sizes.sg, vma->resource->page_sizes_gtt);
+		err = -EINVAL;
+	}
+
+	return err;
+}
+
+static int igt_mock_exhaust_device_supported_pages(void *arg)
+{
+	struct i915_ppgtt *ppgtt = arg;
+	struct drm_i915_private *i915 = ppgtt->vm.i915;
+	unsigned int saved_mask = RUNTIME_INFO(i915)->page_sizes;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int i, j, single;
+	int err;
+
+	/*
+	 * Sanity check creating objects with every valid page support
+	 * combination for our mock device.
+	 */
+
+	for (i = 1; i < BIT(ARRAY_SIZE(page_sizes)); i++) {
+		unsigned int combination = SZ_4K; /* Required for ppGTT */
+
+		for (j = 0; j < ARRAY_SIZE(page_sizes); j++) {
+			if (i & BIT(j))
+				combination |= page_sizes[j];
+		}
+
+		RUNTIME_INFO(i915)->page_sizes = combination;
+
+		for (single = 0; single <= 1; ++single) {
+			obj = fake_huge_pages_object(i915, combination, !!single);
+			if (IS_ERR(obj)) {
+				err = PTR_ERR(obj);
+				goto out_device;
+			}
+
+			if (obj->base.size != combination) {
+				pr_err("obj->base.size=%zu, expected=%u\n",
+				       obj->base.size, combination);
+				err = -EINVAL;
+				goto out_put;
+			}
+
+			vma = i915_vma_instance(obj, &ppgtt->vm, NULL);
+			if (IS_ERR(vma)) {
+				err = PTR_ERR(vma);
+				goto out_put;
+			}
+
+			err = i915_vma_pin(vma, 0, 0, PIN_USER);
+			if (err)
+				goto out_put;
+
+			err = igt_check_page_sizes(vma);
+
+			if (vma->page_sizes.sg != combination) {
+				pr_err("page_sizes.sg=%u, expected=%u\n",
+				       vma->page_sizes.sg, combination);
+				err = -EINVAL;
+			}
+
+			i915_vma_unpin(vma);
+			i915_gem_object_put(obj);
+
+			if (err)
+				goto out_device;
+		}
+	}
+
+	goto out_device;
+
+out_put:
+	i915_gem_object_put(obj);
+out_device:
+	RUNTIME_INFO(i915)->page_sizes = saved_mask;
+
+	return err;
+}
+
+static int igt_mock_memory_region_huge_pages(void *arg)
+{
+	const unsigned int flags[] = { 0, I915_BO_ALLOC_CONTIGUOUS };
+	struct i915_ppgtt *ppgtt = arg;
+	struct drm_i915_private *i915 = ppgtt->vm.i915;
+	unsigned long supported = RUNTIME_INFO(i915)->page_sizes;
+	struct intel_memory_region *mem;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int bit;
+	int err = 0;
+
+	mem = mock_region_create(i915, 0, SZ_2G, I915_GTT_PAGE_SIZE_4K, 0, 0);
+	if (IS_ERR(mem)) {
+		pr_err("%s failed to create memory region\n", __func__);
+		return PTR_ERR(mem);
+	}
+
+	for_each_set_bit(bit, &supported, ilog2(I915_GTT_MAX_PAGE_SIZE) + 1) {
+		unsigned int page_size = BIT(bit);
+		resource_size_t phys;
+		int i;
+
+		for (i = 0; i < ARRAY_SIZE(flags); ++i) {
+			obj = i915_gem_object_create_region(mem,
+							    page_size, page_size,
+							    flags[i]);
+			if (IS_ERR(obj)) {
+				err = PTR_ERR(obj);
+				goto out_region;
+			}
+
+			vma = i915_vma_instance(obj, &ppgtt->vm, NULL);
+			if (IS_ERR(vma)) {
+				err = PTR_ERR(vma);
+				goto out_put;
+			}
+
+			err = i915_vma_pin(vma, 0, 0, PIN_USER);
+			if (err)
+				goto out_put;
+
+			err = igt_check_page_sizes(vma);
+			if (err)
+				goto out_unpin;
+
+			phys = i915_gem_object_get_dma_address(obj, 0);
+			if (!IS_ALIGNED(phys, page_size)) {
+				pr_err("%s addr misaligned(%pa) page_size=%u\n",
+				       __func__, &phys, page_size);
+				err = -EINVAL;
+				goto out_unpin;
+			}
+
+			if (vma->resource->page_sizes_gtt != page_size) {
+				pr_err("%s page_sizes.gtt=%u, expected=%u\n",
+				       __func__, vma->resource->page_sizes_gtt,
+				       page_size);
+				err = -EINVAL;
+				goto out_unpin;
+			}
+
+			i915_vma_unpin(vma);
+			__i915_gem_object_put_pages(obj);
+			i915_gem_object_put(obj);
+		}
+	}
+
+	goto out_region;
+
+out_unpin:
+	i915_vma_unpin(vma);
+out_put:
+	i915_gem_object_put(obj);
+out_region:
+	intel_memory_region_destroy(mem);
+	return err;
+}
+
+static int igt_mock_ppgtt_misaligned_dma(void *arg)
+{
+	struct i915_ppgtt *ppgtt = arg;
+	struct drm_i915_private *i915 = ppgtt->vm.i915;
+	unsigned long supported = RUNTIME_INFO(i915)->page_sizes;
+	struct drm_i915_gem_object *obj;
+	int bit;
+	int err;
+
+	/*
+	 * Sanity check dma misalignment for huge pages -- the dma addresses we
+	 * insert into the paging structures need to always respect the page
+	 * size alignment.
+	 */
+
+	bit = ilog2(I915_GTT_PAGE_SIZE_64K);
+
+	for_each_set_bit_from(bit, &supported,
+			      ilog2(I915_GTT_MAX_PAGE_SIZE) + 1) {
+		IGT_TIMEOUT(end_time);
+		unsigned int page_size = BIT(bit);
+		unsigned int flags = PIN_USER | PIN_OFFSET_FIXED;
+		unsigned int offset;
+		unsigned int size =
+			round_up(page_size, I915_GTT_PAGE_SIZE_2M) << 1;
+		struct i915_vma *vma;
+
+		obj = fake_huge_pages_object(i915, size, true);
+		if (IS_ERR(obj))
+			return PTR_ERR(obj);
+
+		if (obj->base.size != size) {
+			pr_err("obj->base.size=%zu, expected=%u\n",
+			       obj->base.size, size);
+			err = -EINVAL;
+			goto out_put;
+		}
+
+		err = i915_gem_object_pin_pages_unlocked(obj);
+		if (err)
+			goto out_put;
+
+		/* Force the page size for this object */
+		obj->mm.page_sizes.sg = page_size;
+
+		vma = i915_vma_instance(obj, &ppgtt->vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out_unpin;
+		}
+
+		err = i915_vma_pin(vma, 0, 0, flags);
+		if (err)
+			goto out_unpin;
+
+
+		err = igt_check_page_sizes(vma);
+
+		if (vma->resource->page_sizes_gtt != page_size) {
+			pr_err("page_sizes.gtt=%u, expected %u\n",
+			       vma->resource->page_sizes_gtt, page_size);
+			err = -EINVAL;
+		}
+
+		i915_vma_unpin(vma);
+
+		if (err)
+			goto out_unpin;
+
+		/*
+		 * Try all the other valid offsets until the next
+		 * boundary -- should always fall back to using 4K
+		 * pages.
+		 */
+		for (offset = 4096; offset < page_size; offset += 4096) {
+			err = i915_vma_unbind_unlocked(vma);
+			if (err)
+				goto out_unpin;
+
+			err = i915_vma_pin(vma, 0, 0, flags | offset);
+			if (err)
+				goto out_unpin;
+
+			err = igt_check_page_sizes(vma);
+
+			if (vma->resource->page_sizes_gtt != I915_GTT_PAGE_SIZE_4K) {
+				pr_err("page_sizes.gtt=%u, expected %llu\n",
+				       vma->resource->page_sizes_gtt,
+				       I915_GTT_PAGE_SIZE_4K);
+				err = -EINVAL;
+			}
+
+			i915_vma_unpin(vma);
+
+			if (err)
+				goto out_unpin;
+
+			if (igt_timeout(end_time,
+					"%s timed out at offset %x with page-size %x\n",
+					__func__, offset, page_size))
+				break;
+		}
+
+		i915_gem_object_lock(obj, NULL);
+		i915_gem_object_unpin_pages(obj);
+		__i915_gem_object_put_pages(obj);
+		i915_gem_object_unlock(obj);
+		i915_gem_object_put(obj);
+	}
+
+	return 0;
+
+out_unpin:
+	i915_gem_object_lock(obj, NULL);
+	i915_gem_object_unpin_pages(obj);
+	i915_gem_object_unlock(obj);
+out_put:
+	i915_gem_object_put(obj);
+
+	return err;
+}
+
+static void close_object_list(struct list_head *objects,
+			      struct i915_ppgtt *ppgtt)
+{
+	struct drm_i915_gem_object *obj, *on;
+
+	list_for_each_entry_safe(obj, on, objects, st_link) {
+		list_del(&obj->st_link);
+		i915_gem_object_lock(obj, NULL);
+		i915_gem_object_unpin_pages(obj);
+		__i915_gem_object_put_pages(obj);
+		i915_gem_object_unlock(obj);
+		i915_gem_object_put(obj);
+	}
+}
+
+static int igt_mock_ppgtt_huge_fill(void *arg)
+{
+	struct i915_ppgtt *ppgtt = arg;
+	struct drm_i915_private *i915 = ppgtt->vm.i915;
+	unsigned long max_pages = ppgtt->vm.total >> PAGE_SHIFT;
+	unsigned long page_num;
+	bool single = false;
+	LIST_HEAD(objects);
+	IGT_TIMEOUT(end_time);
+	int err = -ENODEV;
+
+	for_each_prime_number_from(page_num, 1, max_pages) {
+		struct drm_i915_gem_object *obj;
+		u64 size = page_num << PAGE_SHIFT;
+		struct i915_vma *vma;
+		unsigned int expected_gtt = 0;
+		int i;
+
+		obj = fake_huge_pages_object(i915, size, single);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			break;
+		}
+
+		if (obj->base.size != size) {
+			pr_err("obj->base.size=%zd, expected=%llu\n",
+			       obj->base.size, size);
+			i915_gem_object_put(obj);
+			err = -EINVAL;
+			break;
+		}
+
+		err = i915_gem_object_pin_pages_unlocked(obj);
+		if (err) {
+			i915_gem_object_put(obj);
+			break;
+		}
+
+		list_add(&obj->st_link, &objects);
+
+		vma = i915_vma_instance(obj, &ppgtt->vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			break;
+		}
+
+		err = i915_vma_pin(vma, 0, 0, PIN_USER);
+		if (err)
+			break;
+
+		err = igt_check_page_sizes(vma);
+		if (err) {
+			i915_vma_unpin(vma);
+			break;
+		}
+
+		/*
+		 * Figure out the expected gtt page size knowing that we go from
+		 * largest to smallest page size sg chunks, and that we align to
+		 * the largest page size.
+		 */
+		for (i = 0; i < ARRAY_SIZE(page_sizes); ++i) {
+			unsigned int page_size = page_sizes[i];
+
+			if (HAS_PAGE_SIZES(i915, page_size) &&
+			    size >= page_size) {
+				expected_gtt |= page_size;
+				size &= page_size-1;
+			}
+		}
+
+		GEM_BUG_ON(!expected_gtt);
+		GEM_BUG_ON(size);
+
+		if (expected_gtt & I915_GTT_PAGE_SIZE_4K)
+			expected_gtt &= ~I915_GTT_PAGE_SIZE_64K;
+
+		i915_vma_unpin(vma);
+
+		if (vma->page_sizes.sg & I915_GTT_PAGE_SIZE_64K) {
+			if (!IS_ALIGNED(vma->node.start,
+					I915_GTT_PAGE_SIZE_2M)) {
+				pr_err("node.start(%llx) not aligned to 2M\n",
+				       vma->node.start);
+				err = -EINVAL;
+				break;
+			}
+
+			if (!IS_ALIGNED(vma->node.size,
+					I915_GTT_PAGE_SIZE_2M)) {
+				pr_err("node.size(%llx) not aligned to 2M\n",
+				       vma->node.size);
+				err = -EINVAL;
+				break;
+			}
+		}
+
+		if (vma->resource->page_sizes_gtt != expected_gtt) {
+			pr_err("gtt=%u, expected=%u, size=%zd, single=%s\n",
+			       vma->resource->page_sizes_gtt, expected_gtt,
+			       obj->base.size, str_yes_no(!!single));
+			err = -EINVAL;
+			break;
+		}
+
+		if (igt_timeout(end_time,
+				"%s timed out at size %zd\n",
+				__func__, obj->base.size))
+			break;
+
+		single = !single;
+	}
+
+	close_object_list(&objects, ppgtt);
+
+	if (err == -ENOMEM || err == -ENOSPC)
+		err = 0;
+
+	return err;
+}
+
+static int igt_mock_ppgtt_64K(void *arg)
+{
+	struct i915_ppgtt *ppgtt = arg;
+	struct drm_i915_private *i915 = ppgtt->vm.i915;
+	struct drm_i915_gem_object *obj;
+	const struct object_info {
+		unsigned int size;
+		unsigned int gtt;
+		unsigned int offset;
+	} objects[] = {
+		/* Cases with forced padding/alignment */
+		{
+			.size = SZ_64K,
+			.gtt = I915_GTT_PAGE_SIZE_64K,
+			.offset = 0,
+		},
+		{
+			.size = SZ_64K + SZ_4K,
+			.gtt = I915_GTT_PAGE_SIZE_4K,
+			.offset = 0,
+		},
+		{
+			.size = SZ_64K - SZ_4K,
+			.gtt = I915_GTT_PAGE_SIZE_4K,
+			.offset = 0,
+		},
+		{
+			.size = SZ_2M,
+			.gtt = I915_GTT_PAGE_SIZE_64K,
+			.offset = 0,
+		},
+		{
+			.size = SZ_2M - SZ_4K,
+			.gtt = I915_GTT_PAGE_SIZE_4K,
+			.offset = 0,
+		},
+		{
+			.size = SZ_2M + SZ_4K,
+			.gtt = I915_GTT_PAGE_SIZE_64K | I915_GTT_PAGE_SIZE_4K,
+			.offset = 0,
+		},
+		{
+			.size = SZ_2M + SZ_64K,
+			.gtt = I915_GTT_PAGE_SIZE_64K,
+			.offset = 0,
+		},
+		{
+			.size = SZ_2M - SZ_64K,
+			.gtt = I915_GTT_PAGE_SIZE_64K,
+			.offset = 0,
+		},
+		/* Try without any forced padding/alignment */
+		{
+			.size = SZ_64K,
+			.offset = SZ_2M,
+			.gtt = I915_GTT_PAGE_SIZE_4K,
+		},
+		{
+			.size = SZ_128K,
+			.offset = SZ_2M - SZ_64K,
+			.gtt = I915_GTT_PAGE_SIZE_4K,
+		},
+	};
+	struct i915_vma *vma;
+	int i, single;
+	int err;
+
+	/*
+	 * Sanity check some of the trickiness with 64K pages -- either we can
+	 * safely mark the whole page-table(2M block) as 64K, or we have to
+	 * always fallback to 4K.
+	 */
+
+	if (!HAS_PAGE_SIZES(i915, I915_GTT_PAGE_SIZE_64K))
+		return 0;
+
+	for (i = 0; i < ARRAY_SIZE(objects); ++i) {
+		unsigned int size = objects[i].size;
+		unsigned int expected_gtt = objects[i].gtt;
+		unsigned int offset = objects[i].offset;
+		unsigned int flags = PIN_USER;
+
+		for (single = 0; single <= 1; single++) {
+			obj = fake_huge_pages_object(i915, size, !!single);
+			if (IS_ERR(obj))
+				return PTR_ERR(obj);
+
+			err = i915_gem_object_pin_pages_unlocked(obj);
+			if (err)
+				goto out_object_put;
+
+			/*
+			 * Disable 2M pages -- We only want to use 64K/4K pages
+			 * for this test.
+			 */
+			obj->mm.page_sizes.sg &= ~I915_GTT_PAGE_SIZE_2M;
+
+			vma = i915_vma_instance(obj, &ppgtt->vm, NULL);
+			if (IS_ERR(vma)) {
+				err = PTR_ERR(vma);
+				goto out_object_unpin;
+			}
+
+			if (offset)
+				flags |= PIN_OFFSET_FIXED | offset;
+
+			err = i915_vma_pin(vma, 0, 0, flags);
+			if (err)
+				goto out_object_unpin;
+
+			err = igt_check_page_sizes(vma);
+			if (err)
+				goto out_vma_unpin;
+
+			if (!offset && vma->page_sizes.sg & I915_GTT_PAGE_SIZE_64K) {
+				if (!IS_ALIGNED(vma->node.start,
+						I915_GTT_PAGE_SIZE_2M)) {
+					pr_err("node.start(%llx) not aligned to 2M\n",
+					       vma->node.start);
+					err = -EINVAL;
+					goto out_vma_unpin;
+				}
+
+				if (!IS_ALIGNED(vma->node.size,
+						I915_GTT_PAGE_SIZE_2M)) {
+					pr_err("node.size(%llx) not aligned to 2M\n",
+					       vma->node.size);
+					err = -EINVAL;
+					goto out_vma_unpin;
+				}
+			}
+
+			if (vma->resource->page_sizes_gtt != expected_gtt) {
+				pr_err("gtt=%u, expected=%u, i=%d, single=%s\n",
+				       vma->resource->page_sizes_gtt,
+				       expected_gtt, i, str_yes_no(!!single));
+				err = -EINVAL;
+				goto out_vma_unpin;
+			}
+
+			i915_vma_unpin(vma);
+			i915_gem_object_lock(obj, NULL);
+			i915_gem_object_unpin_pages(obj);
+			__i915_gem_object_put_pages(obj);
+			i915_gem_object_unlock(obj);
+			i915_gem_object_put(obj);
+
+			i915_gem_drain_freed_objects(i915);
+		}
+	}
+
+	return 0;
+
+out_vma_unpin:
+	i915_vma_unpin(vma);
+out_object_unpin:
+	i915_gem_object_lock(obj, NULL);
+	i915_gem_object_unpin_pages(obj);
+	i915_gem_object_unlock(obj);
+out_object_put:
+	i915_gem_object_put(obj);
+
+	return err;
+}
+
+static int gpu_write(struct intel_context *ce,
+		     struct i915_vma *vma,
+		     u32 dw,
+		     u32 val)
+{
+	int err;
+
+	i915_gem_object_lock(vma->obj, NULL);
+	err = i915_gem_object_set_to_gtt_domain(vma->obj, true);
+	i915_gem_object_unlock(vma->obj);
+	if (err)
+		return err;
+
+	return igt_gpu_fill_dw(ce, vma, dw * sizeof(u32),
+			       vma->size >> PAGE_SHIFT, val);
+}
+
+static int
+__cpu_check_shmem(struct drm_i915_gem_object *obj, u32 dword, u32 val)
+{
+	unsigned int needs_flush;
+	unsigned long n;
+	int err;
+
+	i915_gem_object_lock(obj, NULL);
+	err = i915_gem_object_prepare_read(obj, &needs_flush);
+	if (err)
+		goto err_unlock;
+
+	for (n = 0; n < obj->base.size >> PAGE_SHIFT; ++n) {
+		u32 *ptr = kmap_atomic(i915_gem_object_get_page(obj, n));
+
+		if (needs_flush & CLFLUSH_BEFORE)
+			drm_clflush_virt_range(ptr, PAGE_SIZE);
+
+		if (ptr[dword] != val) {
+			pr_err("n=%lu ptr[%u]=%u, val=%u\n",
+			       n, dword, ptr[dword], val);
+			kunmap_atomic(ptr);
+			err = -EINVAL;
+			break;
+		}
+
+		kunmap_atomic(ptr);
+	}
+
+	i915_gem_object_finish_access(obj);
+err_unlock:
+	i915_gem_object_unlock(obj);
+
+	return err;
+}
+
+static int __cpu_check_vmap(struct drm_i915_gem_object *obj, u32 dword, u32 val)
+{
+	unsigned long n = obj->base.size >> PAGE_SHIFT;
+	u32 *ptr;
+	int err;
+
+	err = i915_gem_object_wait(obj, 0, MAX_SCHEDULE_TIMEOUT);
+	if (err)
+		return err;
+
+	ptr = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(ptr))
+		return PTR_ERR(ptr);
+
+	ptr += dword;
+	while (n--) {
+		if (*ptr != val) {
+			pr_err("base[%u]=%08x, val=%08x\n",
+			       dword, *ptr, val);
+			err = -EINVAL;
+			break;
+		}
+
+		ptr += PAGE_SIZE / sizeof(*ptr);
+	}
+
+	i915_gem_object_unpin_map(obj);
+	return err;
+}
+
+static int cpu_check(struct drm_i915_gem_object *obj, u32 dword, u32 val)
+{
+	if (i915_gem_object_has_struct_page(obj))
+		return __cpu_check_shmem(obj, dword, val);
+	else
+		return __cpu_check_vmap(obj, dword, val);
+}
+
+static int __igt_write_huge(struct intel_context *ce,
+			    struct drm_i915_gem_object *obj,
+			    u64 size, u64 offset,
+			    u32 dword, u32 val)
+{
+	unsigned int flags = PIN_USER | PIN_OFFSET_FIXED;
+	struct i915_vma *vma;
+	int err;
+
+	vma = i915_vma_instance(obj, ce->vm, NULL);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	err = i915_vma_pin(vma, size, 0, flags | offset);
+	if (err) {
+		/*
+		 * The ggtt may have some pages reserved so
+		 * refrain from erroring out.
+		 */
+		if (err == -ENOSPC && i915_is_ggtt(ce->vm))
+			err = 0;
+
+		return err;
+	}
+
+	err = igt_check_page_sizes(vma);
+	if (err)
+		goto out_vma_unpin;
+
+	err = gpu_write(ce, vma, dword, val);
+	if (err) {
+		pr_err("gpu-write failed at offset=%llx\n", offset);
+		goto out_vma_unpin;
+	}
+
+	err = cpu_check(obj, dword, val);
+	if (err) {
+		pr_err("cpu-check failed at offset=%llx\n", offset);
+		goto out_vma_unpin;
+	}
+
+out_vma_unpin:
+	i915_vma_unpin(vma);
+	return err;
+}
+
+static int igt_write_huge(struct drm_i915_private *i915,
+			  struct drm_i915_gem_object *obj)
+{
+	struct i915_gem_engines *engines;
+	struct i915_gem_engines_iter it;
+	struct intel_context *ce;
+	I915_RND_STATE(prng);
+	IGT_TIMEOUT(end_time);
+	unsigned int max_page_size;
+	unsigned int count;
+	struct i915_gem_context *ctx;
+	struct file *file;
+	u64 max;
+	u64 num;
+	u64 size;
+	int *order;
+	int i, n;
+	int err = 0;
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	ctx = hugepage_ctx(i915, file);
+	if (IS_ERR(ctx)) {
+		err = PTR_ERR(ctx);
+		goto out;
+	}
+
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+
+	size = obj->base.size;
+	if (obj->mm.page_sizes.sg & I915_GTT_PAGE_SIZE_64K)
+		size = round_up(size, I915_GTT_PAGE_SIZE_2M);
+
+	n = 0;
+	count = 0;
+	max = U64_MAX;
+	for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it) {
+		count++;
+		if (!intel_engine_can_store_dword(ce->engine))
+			continue;
+
+		max = min(max, ce->vm->total);
+		n++;
+	}
+	i915_gem_context_unlock_engines(ctx);
+	if (!n)
+		goto out;
+
+	/*
+	 * To keep things interesting when alternating between engines in our
+	 * randomized order, lets also make feeding to the same engine a few
+	 * times in succession a possibility by enlarging the permutation array.
+	 */
+	order = i915_random_order(count * count, &prng);
+	if (!order) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	max_page_size = rounddown_pow_of_two(obj->mm.page_sizes.sg);
+	max = div_u64(max - size, max_page_size);
+
+	/*
+	 * Try various offsets in an ascending/descending fashion until we
+	 * timeout -- we want to avoid issues hidden by effectively always using
+	 * offset = 0.
+	 */
+	i = 0;
+	engines = i915_gem_context_lock_engines(ctx);
+	for_each_prime_number_from(num, 0, max) {
+		u64 offset_low = num * max_page_size;
+		u64 offset_high = (max - num) * max_page_size;
+		u32 dword = offset_in_page(num) / 4;
+		struct intel_context *ce;
+
+		ce = engines->engines[order[i] % engines->num_engines];
+		i = (i + 1) % (count * count);
+		if (!ce || !intel_engine_can_store_dword(ce->engine))
+			continue;
+
+		/*
+		 * In order to utilize 64K pages we need to both pad the vma
+		 * size and ensure the vma offset is at the start of the pt
+		 * boundary, however to improve coverage we opt for testing both
+		 * aligned and unaligned offsets.
+		 */
+		if (obj->mm.page_sizes.sg & I915_GTT_PAGE_SIZE_64K)
+			offset_low = round_down(offset_low,
+						I915_GTT_PAGE_SIZE_2M);
+
+		err = __igt_write_huge(ce, obj, size, offset_low,
+				       dword, num + 1);
+		if (err)
+			break;
+
+		err = __igt_write_huge(ce, obj, size, offset_high,
+				       dword, num + 1);
+		if (err)
+			break;
+
+		if (igt_timeout(end_time,
+				"%s timed out on %s, offset_low=%llx offset_high=%llx, max_page_size=%x\n",
+				__func__, ce->engine->name, offset_low, offset_high,
+				max_page_size))
+			break;
+	}
+	i915_gem_context_unlock_engines(ctx);
+
+	kfree(order);
+
+out:
+	fput(file);
+	return err;
+}
+
+typedef struct drm_i915_gem_object *
+(*igt_create_fn)(struct drm_i915_private *i915, u32 size, u32 flags);
+
+static inline bool igt_can_allocate_thp(struct drm_i915_private *i915)
+{
+	return i915->mm.gemfs && has_transparent_hugepage();
+}
+
+static struct drm_i915_gem_object *
+igt_create_shmem(struct drm_i915_private *i915, u32 size, u32 flags)
+{
+	if (!igt_can_allocate_thp(i915)) {
+		pr_info("%s missing THP support, skipping\n", __func__);
+		return ERR_PTR(-ENODEV);
+	}
+
+	return i915_gem_object_create_shmem(i915, size);
+}
+
+static struct drm_i915_gem_object *
+igt_create_internal(struct drm_i915_private *i915, u32 size, u32 flags)
+{
+	return i915_gem_object_create_internal(i915, size);
+}
+
+static struct drm_i915_gem_object *
+igt_create_system(struct drm_i915_private *i915, u32 size, u32 flags)
+{
+	return huge_pages_object(i915, size, size);
+}
+
+static struct drm_i915_gem_object *
+igt_create_local(struct drm_i915_private *i915, u32 size, u32 flags)
+{
+	return i915_gem_object_create_lmem(i915, size, flags);
+}
+
+static u32 igt_random_size(struct rnd_state *prng,
+			   u32 min_page_size,
+			   u32 max_page_size)
+{
+	u64 mask;
+	u32 size;
+
+	GEM_BUG_ON(!is_power_of_2(min_page_size));
+	GEM_BUG_ON(!is_power_of_2(max_page_size));
+	GEM_BUG_ON(min_page_size < PAGE_SIZE);
+	GEM_BUG_ON(min_page_size > max_page_size);
+
+	mask = ((max_page_size << 1ULL) - 1) & PAGE_MASK;
+	size = prandom_u32_state(prng) & mask;
+	if (size < min_page_size)
+		size |= min_page_size;
+
+	return size;
+}
+
+static int igt_ppgtt_smoke_huge(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	I915_RND_STATE(prng);
+	struct {
+		igt_create_fn fn;
+		u32 min;
+		u32 max;
+	} backends[] = {
+		{ igt_create_internal, SZ_64K, SZ_2M,  },
+		{ igt_create_shmem,    SZ_64K, SZ_32M, },
+		{ igt_create_local,    SZ_64K, SZ_1G,  },
+	};
+	int err;
+	int i;
+
+	/*
+	 * Sanity check that the HW uses huge pages correctly through our
+	 * various backends -- ensure that our writes land in the right place.
+	 */
+
+	for (i = 0; i < ARRAY_SIZE(backends); ++i) {
+		u32 min = backends[i].min;
+		u32 max = backends[i].max;
+		u32 size = max;
+
+try_again:
+		size = igt_random_size(&prng, min, rounddown_pow_of_two(size));
+
+		obj = backends[i].fn(i915, size, 0);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			if (err == -E2BIG) {
+				size >>= 1;
+				goto try_again;
+			} else if (err == -ENODEV) {
+				err = 0;
+				continue;
+			}
+
+			return err;
+		}
+
+		err = i915_gem_object_pin_pages_unlocked(obj);
+		if (err) {
+			if (err == -ENXIO || err == -E2BIG || err == -ENOMEM) {
+				i915_gem_object_put(obj);
+				size >>= 1;
+				goto try_again;
+			}
+			goto out_put;
+		}
+
+		if (obj->mm.page_sizes.phys < min) {
+			pr_info("%s unable to allocate huge-page(s) with size=%u, i=%d\n",
+				__func__, size, i);
+			err = -ENOMEM;
+			goto out_unpin;
+		}
+
+		err = igt_write_huge(i915, obj);
+		if (err) {
+			pr_err("%s write-huge failed with size=%u, i=%d\n",
+			       __func__, size, i);
+		}
+out_unpin:
+		i915_gem_object_lock(obj, NULL);
+		i915_gem_object_unpin_pages(obj);
+		__i915_gem_object_put_pages(obj);
+		i915_gem_object_unlock(obj);
+out_put:
+		i915_gem_object_put(obj);
+
+		if (err == -ENOMEM || err == -ENXIO)
+			err = 0;
+
+		if (err)
+			break;
+
+		cond_resched();
+	}
+
+	return err;
+}
+
+static int igt_ppgtt_sanity_check(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	unsigned int supported = RUNTIME_INFO(i915)->page_sizes;
+	struct {
+		igt_create_fn fn;
+		unsigned int flags;
+	} backends[] = {
+		{ igt_create_system, 0,                        },
+		{ igt_create_local,  0,                        },
+		{ igt_create_local,  I915_BO_ALLOC_CONTIGUOUS, },
+	};
+	struct {
+		u32 size;
+		u32 pages;
+	} combos[] = {
+		{ SZ_64K,		SZ_64K		},
+		{ SZ_2M,		SZ_2M		},
+		{ SZ_2M,		SZ_64K		},
+		{ SZ_2M - SZ_64K,	SZ_64K		},
+		{ SZ_2M - SZ_4K,	SZ_64K | SZ_4K	},
+		{ SZ_2M + SZ_4K,	SZ_64K | SZ_4K	},
+		{ SZ_2M + SZ_4K,	SZ_2M  | SZ_4K	},
+		{ SZ_2M + SZ_64K,	SZ_2M  | SZ_64K },
+	};
+	int i, j;
+	int err;
+
+	if (supported == I915_GTT_PAGE_SIZE_4K)
+		return 0;
+
+	/*
+	 * Sanity check that the HW behaves with a limited set of combinations.
+	 * We already have a bunch of randomised testing, which should give us
+	 * a decent amount of variation between runs, however we should keep
+	 * this to limit the chances of introducing a temporary regression, by
+	 * testing the most obvious cases that might make something blow up.
+	 */
+
+	for (i = 0; i < ARRAY_SIZE(backends); ++i) {
+		for (j = 0; j < ARRAY_SIZE(combos); ++j) {
+			struct drm_i915_gem_object *obj;
+			u32 size = combos[j].size;
+			u32 pages = combos[j].pages;
+
+			obj = backends[i].fn(i915, size, backends[i].flags);
+			if (IS_ERR(obj)) {
+				err = PTR_ERR(obj);
+				if (err == -ENODEV) {
+					pr_info("Device lacks local memory, skipping\n");
+					err = 0;
+					break;
+				}
+
+				return err;
+			}
+
+			err = i915_gem_object_pin_pages_unlocked(obj);
+			if (err) {
+				i915_gem_object_put(obj);
+				goto out;
+			}
+
+			GEM_BUG_ON(pages > obj->base.size);
+			pages = pages & supported;
+
+			if (pages)
+				obj->mm.page_sizes.sg = pages;
+
+			err = igt_write_huge(i915, obj);
+
+			i915_gem_object_lock(obj, NULL);
+			i915_gem_object_unpin_pages(obj);
+			__i915_gem_object_put_pages(obj);
+			i915_gem_object_unlock(obj);
+			i915_gem_object_put(obj);
+
+			if (err) {
+				pr_err("%s write-huge failed with size=%u pages=%u i=%d, j=%d\n",
+				       __func__, size, pages, i, j);
+				goto out;
+			}
+		}
+
+		cond_resched();
+	}
+
+out:
+	if (err == -ENOMEM)
+		err = 0;
+
+	return err;
+}
+
+static int igt_ppgtt_compact(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	/*
+	 * Simple test to catch issues with compact 64K pages -- since the pt is
+	 * compacted to 256B that gives us 32 entries per pt, however since the
+	 * backing page for the pt is 4K, any extra entries we might incorrectly
+	 * write out should be ignored by the HW. If ever hit such a case this
+	 * test should catch it since some of our writes would land in scratch.
+	 */
+
+	if (!HAS_64K_PAGES(i915)) {
+		pr_info("device lacks compact 64K page support, skipping\n");
+		return 0;
+	}
+
+	if (!HAS_LMEM(i915)) {
+		pr_info("device lacks LMEM support, skipping\n");
+		return 0;
+	}
+
+	/* We want the range to cover multiple page-table boundaries. */
+	obj = i915_gem_object_create_lmem(i915, SZ_4M, 0);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	err = i915_gem_object_pin_pages_unlocked(obj);
+	if (err)
+		goto out_put;
+
+	if (obj->mm.page_sizes.phys < I915_GTT_PAGE_SIZE_64K) {
+		pr_info("LMEM compact unable to allocate huge-page(s)\n");
+		goto out_unpin;
+	}
+
+	/*
+	 * Disable 2M GTT pages by forcing the page-size to 64K for the GTT
+	 * insertion.
+	 */
+	obj->mm.page_sizes.sg = I915_GTT_PAGE_SIZE_64K;
+
+	err = igt_write_huge(i915, obj);
+	if (err)
+		pr_err("LMEM compact write-huge failed\n");
+
+out_unpin:
+	i915_gem_object_unpin_pages(obj);
+out_put:
+	i915_gem_object_put(obj);
+
+	if (err == -ENOMEM)
+		err = 0;
+
+	return err;
+}
+
+static int igt_tmpfs_fallback(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_address_space *vm;
+	struct i915_gem_context *ctx;
+	struct vfsmount *gemfs = i915->mm.gemfs;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	struct file *file;
+	u32 *vaddr;
+	int err = 0;
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	ctx = hugepage_ctx(i915, file);
+	if (IS_ERR(ctx)) {
+		err = PTR_ERR(ctx);
+		goto out;
+	}
+	vm = i915_gem_context_get_eb_vm(ctx);
+
+	/*
+	 * Make sure that we don't burst into a ball of flames upon falling back
+	 * to tmpfs, which we rely on if on the off-chance we encouter a failure
+	 * when setting up gemfs.
+	 */
+
+	i915->mm.gemfs = NULL;
+
+	obj = i915_gem_object_create_shmem(i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto out_restore;
+	}
+
+	vaddr = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto out_put;
+	}
+	*vaddr = 0xdeadbeaf;
+
+	__i915_gem_object_flush_map(obj, 0, 64);
+	i915_gem_object_unpin_map(obj);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto out_put;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto out_put;
+
+	err = igt_check_page_sizes(vma);
+
+	i915_vma_unpin(vma);
+out_put:
+	i915_gem_object_put(obj);
+out_restore:
+	i915->mm.gemfs = gemfs;
+
+	i915_vm_put(vm);
+out:
+	fput(file);
+	return err;
+}
+
+static int igt_shrink_thp(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_address_space *vm;
+	struct i915_gem_context *ctx;
+	struct drm_i915_gem_object *obj;
+	struct i915_gem_engines_iter it;
+	struct intel_context *ce;
+	struct i915_vma *vma;
+	struct file *file;
+	unsigned int flags = PIN_USER;
+	unsigned int n;
+	intel_wakeref_t wf;
+	bool should_swap;
+	int err;
+
+	if (!igt_can_allocate_thp(i915)) {
+		pr_info("missing THP support, skipping\n");
+		return 0;
+	}
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	ctx = hugepage_ctx(i915, file);
+	if (IS_ERR(ctx)) {
+		err = PTR_ERR(ctx);
+		goto out;
+	}
+	vm = i915_gem_context_get_eb_vm(ctx);
+
+	/*
+	 * Sanity check shrinking huge-paged object -- make sure nothing blows
+	 * up.
+	 */
+
+	obj = i915_gem_object_create_shmem(i915, SZ_2M);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto out_vm;
+	}
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto out_put;
+	}
+
+	wf = intel_runtime_pm_get(&i915->runtime_pm); /* active shrink */
+
+	err = i915_vma_pin(vma, 0, 0, flags);
+	if (err)
+		goto out_wf;
+
+	if (obj->mm.page_sizes.phys < I915_GTT_PAGE_SIZE_2M) {
+		pr_info("failed to allocate THP, finishing test early\n");
+		goto out_unpin;
+	}
+
+	err = igt_check_page_sizes(vma);
+	if (err)
+		goto out_unpin;
+
+	n = 0;
+
+	for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it) {
+		if (!intel_engine_can_store_dword(ce->engine))
+			continue;
+
+		err = gpu_write(ce, vma, n++, 0xdeadbeaf);
+		if (err)
+			break;
+	}
+	i915_gem_context_unlock_engines(ctx);
+	/*
+	 * Nuke everything *before* we unpin the pages so we can be reasonably
+	 * sure that when later checking get_nr_swap_pages() that some random
+	 * leftover object doesn't steal the remaining swap space.
+	 */
+	i915_gem_shrink(NULL, i915, -1UL, NULL,
+			I915_SHRINK_BOUND |
+			I915_SHRINK_UNBOUND |
+			I915_SHRINK_ACTIVE);
+	i915_vma_unpin(vma);
+	if (err)
+		goto out_wf;
+
+	/*
+	 * Now that the pages are *unpinned* shrinking should invoke
+	 * shmem to truncate our pages, if we have available swap.
+	 */
+	should_swap = get_nr_swap_pages() > 0;
+	i915_gem_shrink(NULL, i915, -1UL, NULL,
+			I915_SHRINK_BOUND |
+			I915_SHRINK_UNBOUND |
+			I915_SHRINK_ACTIVE |
+			I915_SHRINK_WRITEBACK);
+	if (should_swap == i915_gem_object_has_pages(obj)) {
+		pr_err("unexpected pages mismatch, should_swap=%s\n",
+		       str_yes_no(should_swap));
+		err = -EINVAL;
+		goto out_wf;
+	}
+
+	if (should_swap == (obj->mm.page_sizes.sg || obj->mm.page_sizes.phys)) {
+		pr_err("unexpected residual page-size bits, should_swap=%s\n",
+		       str_yes_no(should_swap));
+		err = -EINVAL;
+		goto out_wf;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, flags);
+	if (err)
+		goto out_wf;
+
+	while (n--) {
+		err = cpu_check(obj, n, 0xdeadbeaf);
+		if (err)
+			break;
+	}
+
+out_unpin:
+	i915_vma_unpin(vma);
+out_wf:
+	intel_runtime_pm_put(&i915->runtime_pm, wf);
+out_put:
+	i915_gem_object_put(obj);
+out_vm:
+	i915_vm_put(vm);
+out:
+	fput(file);
+	return err;
+}
+
+int i915_gem_huge_page_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_mock_exhaust_device_supported_pages),
+		SUBTEST(igt_mock_memory_region_huge_pages),
+		SUBTEST(igt_mock_ppgtt_misaligned_dma),
+		SUBTEST(igt_mock_ppgtt_huge_fill),
+		SUBTEST(igt_mock_ppgtt_64K),
+	};
+	struct drm_i915_private *dev_priv;
+	struct i915_ppgtt *ppgtt;
+	int err;
+
+	dev_priv = mock_gem_device();
+	if (!dev_priv)
+		return -ENOMEM;
+
+	/* Pretend to be a device which supports the 48b PPGTT */
+	RUNTIME_INFO(dev_priv)->ppgtt_type = INTEL_PPGTT_FULL;
+	RUNTIME_INFO(dev_priv)->ppgtt_size = 48;
+
+	ppgtt = i915_ppgtt_create(to_gt(dev_priv), 0);
+	if (IS_ERR(ppgtt)) {
+		err = PTR_ERR(ppgtt);
+		goto out_unlock;
+	}
+
+	if (!i915_vm_is_4lvl(&ppgtt->vm)) {
+		pr_err("failed to create 48b PPGTT\n");
+		err = -EINVAL;
+		goto out_put;
+	}
+
+	/* If we were ever hit this then it's time to mock the 64K scratch */
+	if (!i915_vm_has_scratch_64K(&ppgtt->vm)) {
+		pr_err("PPGTT missing 64K scratch page\n");
+		err = -EINVAL;
+		goto out_put;
+	}
+
+	err = i915_subtests(tests, ppgtt);
+
+out_put:
+	i915_vm_put(&ppgtt->vm);
+out_unlock:
+	mock_destroy_device(dev_priv);
+	return err;
+}
+
+int i915_gem_huge_page_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_shrink_thp),
+		SUBTEST(igt_tmpfs_fallback),
+		SUBTEST(igt_ppgtt_smoke_huge),
+		SUBTEST(igt_ppgtt_sanity_check),
+		SUBTEST(igt_ppgtt_compact),
+	};
+
+	if (!HAS_PPGTT(i915)) {
+		pr_info("PPGTT not supported, skipping live-selftests\n");
+		return 0;
+	}
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_client_blt.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_client_blt.c
new file mode 100644
index 000000000..9a6a6b5b7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_client_blt.c
@@ -0,0 +1,765 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "i915_selftest.h"
+
+#include "gt/intel_context.h"
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_engine_user.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_regs.h"
+#include "gem/i915_gem_lmem.h"
+
+#include "selftests/igt_flush_test.h"
+#include "selftests/mock_drm.h"
+#include "selftests/i915_random.h"
+#include "huge_gem_object.h"
+#include "mock_context.h"
+
+#define OW_SIZE 16                      /* in bytes */
+#define F_SUBTILE_SIZE 64               /* in bytes */
+#define F_TILE_WIDTH 128                /* in bytes */
+#define F_TILE_HEIGHT 32                /* in pixels */
+#define F_SUBTILE_WIDTH  OW_SIZE        /* in bytes */
+#define F_SUBTILE_HEIGHT 4              /* in pixels */
+
+static int linear_x_y_to_ftiled_pos(int x, int y, u32 stride, int bpp)
+{
+	int tile_base;
+	int tile_x, tile_y;
+	int swizzle, subtile;
+	int pixel_size = bpp / 8;
+	int pos;
+
+	/*
+	 * Subtile remapping for F tile. Note that map[a]==b implies map[b]==a
+	 * so we can use the same table to tile and until.
+	 */
+	static const u8 f_subtile_map[] = {
+		 0,  1,  2,  3,  8,  9, 10, 11,
+		 4,  5,  6,  7, 12, 13, 14, 15,
+		16, 17, 18, 19, 24, 25, 26, 27,
+		20, 21, 22, 23, 28, 29, 30, 31,
+		32, 33, 34, 35, 40, 41, 42, 43,
+		36, 37, 38, 39, 44, 45, 46, 47,
+		48, 49, 50, 51, 56, 57, 58, 59,
+		52, 53, 54, 55, 60, 61, 62, 63
+	};
+
+	x *= pixel_size;
+	/*
+	 * Where does the 4k tile start (in bytes)?  This is the same for Y and
+	 * F so we can use the Y-tile algorithm to get to that point.
+	 */
+	tile_base =
+		y / F_TILE_HEIGHT * stride * F_TILE_HEIGHT +
+		x / F_TILE_WIDTH * 4096;
+
+	/* Find pixel within tile */
+	tile_x = x % F_TILE_WIDTH;
+	tile_y = y % F_TILE_HEIGHT;
+
+	/* And figure out the subtile within the 4k tile */
+	subtile = tile_y / F_SUBTILE_HEIGHT * 8 + tile_x / F_SUBTILE_WIDTH;
+
+	/* Swizzle the subtile number according to the bspec diagram */
+	swizzle = f_subtile_map[subtile];
+
+	/* Calculate new position */
+	pos = tile_base +
+		swizzle * F_SUBTILE_SIZE +
+		tile_y % F_SUBTILE_HEIGHT * OW_SIZE +
+		tile_x % F_SUBTILE_WIDTH;
+
+	GEM_BUG_ON(!IS_ALIGNED(pos, pixel_size));
+
+	return pos / pixel_size * 4;
+}
+
+enum client_tiling {
+	CLIENT_TILING_LINEAR,
+	CLIENT_TILING_X,
+	CLIENT_TILING_Y,
+	CLIENT_TILING_4,
+	CLIENT_NUM_TILING_TYPES
+};
+
+#define WIDTH 512
+#define HEIGHT 32
+
+struct blit_buffer {
+	struct i915_vma *vma;
+	u32 start_val;
+	enum client_tiling tiling;
+};
+
+struct tiled_blits {
+	struct intel_context *ce;
+	struct blit_buffer buffers[3];
+	struct blit_buffer scratch;
+	struct i915_vma *batch;
+	u64 hole;
+	u64 align;
+	u32 width;
+	u32 height;
+};
+
+static bool supports_x_tiling(const struct drm_i915_private *i915)
+{
+	int gen = GRAPHICS_VER(i915);
+
+	if (gen < 12)
+		return true;
+
+	if (!HAS_LMEM(i915) || IS_DG1(i915))
+		return false;
+
+	return true;
+}
+
+static bool fast_blit_ok(const struct blit_buffer *buf)
+{
+	int gen = GRAPHICS_VER(buf->vma->vm->i915);
+
+	if (gen < 9)
+		return false;
+
+	if (gen < 12)
+		return true;
+
+	/* filter out platforms with unsupported X-tile support in fastblit */
+	if (buf->tiling == CLIENT_TILING_X && !supports_x_tiling(buf->vma->vm->i915))
+		return false;
+
+	return true;
+}
+
+static int prepare_blit(const struct tiled_blits *t,
+			struct blit_buffer *dst,
+			struct blit_buffer *src,
+			struct drm_i915_gem_object *batch)
+{
+	const int ver = GRAPHICS_VER(to_i915(batch->base.dev));
+	bool use_64b_reloc = ver >= 8;
+	u32 src_pitch, dst_pitch;
+	u32 cmd, *cs;
+
+	cs = i915_gem_object_pin_map_unlocked(batch, I915_MAP_WC);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	if (fast_blit_ok(dst) && fast_blit_ok(src)) {
+		struct intel_gt *gt = t->ce->engine->gt;
+		u32 src_tiles = 0, dst_tiles = 0;
+		u32 src_4t = 0, dst_4t = 0;
+
+		/* Need to program BLIT_CCTL if it is not done previously
+		 * before using XY_FAST_COPY_BLT
+		 */
+		*cs++ = MI_LOAD_REGISTER_IMM(1);
+		*cs++ = i915_mmio_reg_offset(BLIT_CCTL(t->ce->engine->mmio_base));
+		*cs++ = (BLIT_CCTL_SRC_MOCS(gt->mocs.uc_index) |
+			 BLIT_CCTL_DST_MOCS(gt->mocs.uc_index));
+
+		src_pitch = t->width; /* in dwords */
+		if (src->tiling == CLIENT_TILING_4) {
+			src_tiles = XY_FAST_COPY_BLT_D0_SRC_TILE_MODE(YMAJOR);
+			src_4t = XY_FAST_COPY_BLT_D1_SRC_TILE4;
+		} else if (src->tiling == CLIENT_TILING_Y) {
+			src_tiles = XY_FAST_COPY_BLT_D0_SRC_TILE_MODE(YMAJOR);
+		} else if (src->tiling == CLIENT_TILING_X) {
+			src_tiles = XY_FAST_COPY_BLT_D0_SRC_TILE_MODE(TILE_X);
+		} else {
+			src_pitch *= 4; /* in bytes */
+		}
+
+		dst_pitch = t->width; /* in dwords */
+		if (dst->tiling == CLIENT_TILING_4) {
+			dst_tiles = XY_FAST_COPY_BLT_D0_DST_TILE_MODE(YMAJOR);
+			dst_4t = XY_FAST_COPY_BLT_D1_DST_TILE4;
+		} else if (dst->tiling == CLIENT_TILING_Y) {
+			dst_tiles = XY_FAST_COPY_BLT_D0_DST_TILE_MODE(YMAJOR);
+		} else if (dst->tiling == CLIENT_TILING_X) {
+			dst_tiles = XY_FAST_COPY_BLT_D0_DST_TILE_MODE(TILE_X);
+		} else {
+			dst_pitch *= 4; /* in bytes */
+		}
+
+		*cs++ = GEN9_XY_FAST_COPY_BLT_CMD | (10 - 2) |
+			src_tiles | dst_tiles;
+		*cs++ = src_4t | dst_4t | BLT_DEPTH_32 | dst_pitch;
+		*cs++ = 0;
+		*cs++ = t->height << 16 | t->width;
+		*cs++ = lower_32_bits(dst->vma->node.start);
+		*cs++ = upper_32_bits(dst->vma->node.start);
+		*cs++ = 0;
+		*cs++ = src_pitch;
+		*cs++ = lower_32_bits(src->vma->node.start);
+		*cs++ = upper_32_bits(src->vma->node.start);
+	} else {
+		if (ver >= 6) {
+			*cs++ = MI_LOAD_REGISTER_IMM(1);
+			*cs++ = i915_mmio_reg_offset(BCS_SWCTRL);
+			cmd = (BCS_SRC_Y | BCS_DST_Y) << 16;
+			if (src->tiling == CLIENT_TILING_Y)
+				cmd |= BCS_SRC_Y;
+			if (dst->tiling == CLIENT_TILING_Y)
+				cmd |= BCS_DST_Y;
+			*cs++ = cmd;
+
+			cmd = MI_FLUSH_DW;
+			if (ver >= 8)
+				cmd++;
+			*cs++ = cmd;
+			*cs++ = 0;
+			*cs++ = 0;
+			*cs++ = 0;
+		}
+
+		cmd = XY_SRC_COPY_BLT_CMD | BLT_WRITE_RGBA | (8 - 2);
+		if (ver >= 8)
+			cmd += 2;
+
+		src_pitch = t->width * 4;
+		if (src->tiling) {
+			cmd |= XY_SRC_COPY_BLT_SRC_TILED;
+			src_pitch /= 4;
+		}
+
+		dst_pitch = t->width * 4;
+		if (dst->tiling) {
+			cmd |= XY_SRC_COPY_BLT_DST_TILED;
+			dst_pitch /= 4;
+		}
+
+		*cs++ = cmd;
+		*cs++ = BLT_DEPTH_32 | BLT_ROP_SRC_COPY | dst_pitch;
+		*cs++ = 0;
+		*cs++ = t->height << 16 | t->width;
+		*cs++ = lower_32_bits(dst->vma->node.start);
+		if (use_64b_reloc)
+			*cs++ = upper_32_bits(dst->vma->node.start);
+		*cs++ = 0;
+		*cs++ = src_pitch;
+		*cs++ = lower_32_bits(src->vma->node.start);
+		if (use_64b_reloc)
+			*cs++ = upper_32_bits(src->vma->node.start);
+	}
+
+	*cs++ = MI_BATCH_BUFFER_END;
+
+	i915_gem_object_flush_map(batch);
+	i915_gem_object_unpin_map(batch);
+
+	return 0;
+}
+
+static void tiled_blits_destroy_buffers(struct tiled_blits *t)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(t->buffers); i++)
+		i915_vma_put(t->buffers[i].vma);
+
+	i915_vma_put(t->scratch.vma);
+	i915_vma_put(t->batch);
+}
+
+static struct i915_vma *
+__create_vma(struct tiled_blits *t, size_t size, bool lmem)
+{
+	struct drm_i915_private *i915 = t->ce->vm->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+
+	if (lmem)
+		obj = i915_gem_object_create_lmem(i915, size, 0);
+	else
+		obj = i915_gem_object_create_shmem(i915, size);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, t->ce->vm, NULL);
+	if (IS_ERR(vma))
+		i915_gem_object_put(obj);
+
+	return vma;
+}
+
+static struct i915_vma *create_vma(struct tiled_blits *t, bool lmem)
+{
+	return __create_vma(t, PAGE_ALIGN(t->width * t->height * 4), lmem);
+}
+
+static int tiled_blits_create_buffers(struct tiled_blits *t,
+				      int width, int height,
+				      struct rnd_state *prng)
+{
+	struct drm_i915_private *i915 = t->ce->engine->i915;
+	int i;
+
+	t->width = width;
+	t->height = height;
+
+	t->batch = __create_vma(t, PAGE_SIZE, false);
+	if (IS_ERR(t->batch))
+		return PTR_ERR(t->batch);
+
+	t->scratch.vma = create_vma(t, false);
+	if (IS_ERR(t->scratch.vma)) {
+		i915_vma_put(t->batch);
+		return PTR_ERR(t->scratch.vma);
+	}
+
+	for (i = 0; i < ARRAY_SIZE(t->buffers); i++) {
+		struct i915_vma *vma;
+
+		vma = create_vma(t, HAS_LMEM(i915) && i % 2);
+		if (IS_ERR(vma)) {
+			tiled_blits_destroy_buffers(t);
+			return PTR_ERR(vma);
+		}
+
+		t->buffers[i].vma = vma;
+		t->buffers[i].tiling =
+			i915_prandom_u32_max_state(CLIENT_NUM_TILING_TYPES, prng);
+
+		/* Platforms support either TileY or Tile4, not both */
+		if (HAS_4TILE(i915) && t->buffers[i].tiling == CLIENT_TILING_Y)
+			t->buffers[i].tiling = CLIENT_TILING_4;
+		else if (!HAS_4TILE(i915) && t->buffers[i].tiling == CLIENT_TILING_4)
+			t->buffers[i].tiling = CLIENT_TILING_Y;
+	}
+
+	return 0;
+}
+
+static void fill_scratch(struct tiled_blits *t, u32 *vaddr, u32 val)
+{
+	int i;
+
+	t->scratch.start_val = val;
+	for (i = 0; i < t->width * t->height; i++)
+		vaddr[i] = val++;
+
+	i915_gem_object_flush_map(t->scratch.vma->obj);
+}
+
+static u64 swizzle_bit(unsigned int bit, u64 offset)
+{
+	return (offset & BIT_ULL(bit)) >> (bit - 6);
+}
+
+static u64 tiled_offset(const struct intel_gt *gt,
+			u64 v,
+			unsigned int stride,
+			enum client_tiling tiling,
+			int x_pos, int y_pos)
+{
+	unsigned int swizzle;
+	u64 x, y;
+
+	if (tiling == CLIENT_TILING_LINEAR)
+		return v;
+
+	y = div64_u64_rem(v, stride, &x);
+
+	if (tiling == CLIENT_TILING_4) {
+		v = linear_x_y_to_ftiled_pos(x_pos, y_pos, stride, 32);
+
+		/* no swizzling for f-tiling */
+		swizzle = I915_BIT_6_SWIZZLE_NONE;
+	} else if (tiling == CLIENT_TILING_X) {
+		v = div64_u64_rem(y, 8, &y) * stride * 8;
+		v += y * 512;
+		v += div64_u64_rem(x, 512, &x) << 12;
+		v += x;
+
+		swizzle = gt->ggtt->bit_6_swizzle_x;
+	} else {
+		const unsigned int ytile_span = 16;
+		const unsigned int ytile_height = 512;
+
+		v = div64_u64_rem(y, 32, &y) * stride * 32;
+		v += y * ytile_span;
+		v += div64_u64_rem(x, ytile_span, &x) * ytile_height;
+		v += x;
+
+		swizzle = gt->ggtt->bit_6_swizzle_y;
+	}
+
+	switch (swizzle) {
+	case I915_BIT_6_SWIZZLE_9:
+		v ^= swizzle_bit(9, v);
+		break;
+	case I915_BIT_6_SWIZZLE_9_10:
+		v ^= swizzle_bit(9, v) ^ swizzle_bit(10, v);
+		break;
+	case I915_BIT_6_SWIZZLE_9_11:
+		v ^= swizzle_bit(9, v) ^ swizzle_bit(11, v);
+		break;
+	case I915_BIT_6_SWIZZLE_9_10_11:
+		v ^= swizzle_bit(9, v) ^ swizzle_bit(10, v) ^ swizzle_bit(11, v);
+		break;
+	}
+
+	return v;
+}
+
+static const char *repr_tiling(enum client_tiling tiling)
+{
+	switch (tiling) {
+	case CLIENT_TILING_LINEAR: return "linear";
+	case CLIENT_TILING_X: return "X";
+	case CLIENT_TILING_Y: return "Y";
+	case CLIENT_TILING_4: return "F";
+	default: return "unknown";
+	}
+}
+
+static int verify_buffer(const struct tiled_blits *t,
+			 struct blit_buffer *buf,
+			 struct rnd_state *prng)
+{
+	const u32 *vaddr;
+	int ret = 0;
+	int x, y, p;
+
+	x = i915_prandom_u32_max_state(t->width, prng);
+	y = i915_prandom_u32_max_state(t->height, prng);
+	p = y * t->width + x;
+
+	vaddr = i915_gem_object_pin_map_unlocked(buf->vma->obj, I915_MAP_WC);
+	if (IS_ERR(vaddr))
+		return PTR_ERR(vaddr);
+
+	if (vaddr[0] != buf->start_val) {
+		ret = -EINVAL;
+	} else {
+		u64 v = tiled_offset(buf->vma->vm->gt,
+				     p * 4, t->width * 4,
+				     buf->tiling, x, y);
+
+		if (vaddr[v / sizeof(*vaddr)] != buf->start_val + p)
+			ret = -EINVAL;
+	}
+	if (ret) {
+		pr_err("Invalid %s tiling detected at (%d, %d), start_val %x\n",
+		       repr_tiling(buf->tiling),
+		       x, y, buf->start_val);
+		igt_hexdump(vaddr, 4096);
+	}
+
+	i915_gem_object_unpin_map(buf->vma->obj);
+	return ret;
+}
+
+static int move_to_active(struct i915_vma *vma,
+			  struct i915_request *rq,
+			  unsigned int flags)
+{
+	int err;
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, false);
+	if (err == 0)
+		err = i915_vma_move_to_active(vma, rq, flags);
+	i915_vma_unlock(vma);
+
+	return err;
+}
+
+static int pin_buffer(struct i915_vma *vma, u64 addr)
+{
+	int err;
+
+	if (drm_mm_node_allocated(&vma->node) && vma->node.start != addr) {
+		err = i915_vma_unbind_unlocked(vma);
+		if (err)
+			return err;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER | PIN_OFFSET_FIXED | addr);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static int
+tiled_blit(struct tiled_blits *t,
+	   struct blit_buffer *dst, u64 dst_addr,
+	   struct blit_buffer *src, u64 src_addr)
+{
+	struct i915_request *rq;
+	int err;
+
+	err = pin_buffer(src->vma, src_addr);
+	if (err) {
+		pr_err("Cannot pin src @ %llx\n", src_addr);
+		return err;
+	}
+
+	err = pin_buffer(dst->vma, dst_addr);
+	if (err) {
+		pr_err("Cannot pin dst @ %llx\n", dst_addr);
+		goto err_src;
+	}
+
+	err = i915_vma_pin(t->batch, 0, 0, PIN_USER | PIN_HIGH);
+	if (err) {
+		pr_err("cannot pin batch\n");
+		goto err_dst;
+	}
+
+	err = prepare_blit(t, dst, src, t->batch->obj);
+	if (err)
+		goto err_bb;
+
+	rq = intel_context_create_request(t->ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_bb;
+	}
+
+	err = move_to_active(t->batch, rq, 0);
+	if (!err)
+		err = move_to_active(src->vma, rq, 0);
+	if (!err)
+		err = move_to_active(dst->vma, rq, 0);
+	if (!err)
+		err = rq->engine->emit_bb_start(rq,
+						t->batch->node.start,
+						t->batch->node.size,
+						0);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (i915_request_wait(rq, 0, HZ / 2) < 0)
+		err = -ETIME;
+	i915_request_put(rq);
+
+	dst->start_val = src->start_val;
+err_bb:
+	i915_vma_unpin(t->batch);
+err_dst:
+	i915_vma_unpin(dst->vma);
+err_src:
+	i915_vma_unpin(src->vma);
+	return err;
+}
+
+static struct tiled_blits *
+tiled_blits_create(struct intel_engine_cs *engine, struct rnd_state *prng)
+{
+	struct drm_mm_node hole;
+	struct tiled_blits *t;
+	u64 hole_size;
+	int err;
+
+	t = kzalloc(sizeof(*t), GFP_KERNEL);
+	if (!t)
+		return ERR_PTR(-ENOMEM);
+
+	t->ce = intel_context_create(engine);
+	if (IS_ERR(t->ce)) {
+		err = PTR_ERR(t->ce);
+		goto err_free;
+	}
+
+	t->align = i915_vm_min_alignment(t->ce->vm, INTEL_MEMORY_LOCAL);
+	t->align = max(t->align,
+		       i915_vm_min_alignment(t->ce->vm, INTEL_MEMORY_SYSTEM));
+
+	hole_size = 2 * round_up(WIDTH * HEIGHT * 4, t->align);
+	hole_size *= 2; /* room to maneuver */
+	hole_size += 2 * t->align; /* padding on either side */
+
+	mutex_lock(&t->ce->vm->mutex);
+	memset(&hole, 0, sizeof(hole));
+	err = drm_mm_insert_node_in_range(&t->ce->vm->mm, &hole,
+					  hole_size, t->align,
+					  I915_COLOR_UNEVICTABLE,
+					  0, U64_MAX,
+					  DRM_MM_INSERT_BEST);
+	if (!err)
+		drm_mm_remove_node(&hole);
+	mutex_unlock(&t->ce->vm->mutex);
+	if (err) {
+		err = -ENODEV;
+		goto err_put;
+	}
+
+	t->hole = hole.start + t->align;
+	pr_info("Using hole at %llx\n", t->hole);
+
+	err = tiled_blits_create_buffers(t, WIDTH, HEIGHT, prng);
+	if (err)
+		goto err_put;
+
+	return t;
+
+err_put:
+	intel_context_put(t->ce);
+err_free:
+	kfree(t);
+	return ERR_PTR(err);
+}
+
+static void tiled_blits_destroy(struct tiled_blits *t)
+{
+	tiled_blits_destroy_buffers(t);
+
+	intel_context_put(t->ce);
+	kfree(t);
+}
+
+static int tiled_blits_prepare(struct tiled_blits *t,
+			       struct rnd_state *prng)
+{
+	u64 offset = round_up(t->width * t->height * 4, t->align);
+	u32 *map;
+	int err;
+	int i;
+
+	map = i915_gem_object_pin_map_unlocked(t->scratch.vma->obj, I915_MAP_WC);
+	if (IS_ERR(map))
+		return PTR_ERR(map);
+
+	/* Use scratch to fill objects */
+	for (i = 0; i < ARRAY_SIZE(t->buffers); i++) {
+		fill_scratch(t, map, prandom_u32_state(prng));
+		GEM_BUG_ON(verify_buffer(t, &t->scratch, prng));
+
+		err = tiled_blit(t,
+				 &t->buffers[i], t->hole + offset,
+				 &t->scratch, t->hole);
+		if (err == 0)
+			err = verify_buffer(t, &t->buffers[i], prng);
+		if (err) {
+			pr_err("Failed to create buffer %d\n", i);
+			break;
+		}
+	}
+
+	i915_gem_object_unpin_map(t->scratch.vma->obj);
+	return err;
+}
+
+static int tiled_blits_bounce(struct tiled_blits *t, struct rnd_state *prng)
+{
+	u64 offset = round_up(t->width * t->height * 4, 2 * t->align);
+	int err;
+
+	/* We want to check position invariant tiling across GTT eviction */
+
+	err = tiled_blit(t,
+			 &t->buffers[1], t->hole + offset / 2,
+			 &t->buffers[0], t->hole + 2 * offset);
+	if (err)
+		return err;
+
+	/* Simulating GTT eviction of the same buffer / layout */
+	t->buffers[2].tiling = t->buffers[0].tiling;
+
+	/* Reposition so that we overlap the old addresses, and slightly off */
+	err = tiled_blit(t,
+			 &t->buffers[2], t->hole + t->align,
+			 &t->buffers[1], t->hole + 3 * offset / 2);
+	if (err)
+		return err;
+
+	err = verify_buffer(t, &t->buffers[2], prng);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static int __igt_client_tiled_blits(struct intel_engine_cs *engine,
+				    struct rnd_state *prng)
+{
+	struct tiled_blits *t;
+	int err;
+
+	t = tiled_blits_create(engine, prng);
+	if (IS_ERR(t))
+		return PTR_ERR(t);
+
+	err = tiled_blits_prepare(t, prng);
+	if (err)
+		goto out;
+
+	err = tiled_blits_bounce(t, prng);
+	if (err)
+		goto out;
+
+out:
+	tiled_blits_destroy(t);
+	return err;
+}
+
+static bool has_bit17_swizzle(int sw)
+{
+	return (sw == I915_BIT_6_SWIZZLE_9_10_17 ||
+		sw == I915_BIT_6_SWIZZLE_9_17);
+}
+
+static bool bad_swizzling(struct drm_i915_private *i915)
+{
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+
+	if (i915->gem_quirks & GEM_QUIRK_PIN_SWIZZLED_PAGES)
+		return true;
+
+	if (has_bit17_swizzle(ggtt->bit_6_swizzle_x) ||
+	    has_bit17_swizzle(ggtt->bit_6_swizzle_y))
+		return true;
+
+	return false;
+}
+
+static int igt_client_tiled_blits(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	I915_RND_STATE(prng);
+	int inst = 0;
+
+	/* Test requires explicit BLT tiling controls */
+	if (GRAPHICS_VER(i915) < 4)
+		return 0;
+
+	if (bad_swizzling(i915)) /* Requires sane (sub-page) swizzling */
+		return 0;
+
+	do {
+		struct intel_engine_cs *engine;
+		int err;
+
+		engine = intel_engine_lookup_user(i915,
+						  I915_ENGINE_CLASS_COPY,
+						  inst++);
+		if (!engine)
+			return 0;
+
+		err = __igt_client_tiled_blits(engine, &prng);
+		if (err == -ENODEV)
+			err = 0;
+		if (err)
+			return err;
+	} while (1);
+}
+
+int i915_gem_client_blt_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_client_tiled_blits),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c
new file mode 100644
index 000000000..a666d7e61
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c
@@ -0,0 +1,438 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2017 Intel Corporation
+ */
+
+#include <linux/prime_numbers.h>
+
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_pm.h"
+#include "gt/intel_ring.h"
+
+#include "i915_selftest.h"
+#include "selftests/i915_random.h"
+
+struct context {
+	struct drm_i915_gem_object *obj;
+	struct intel_engine_cs *engine;
+};
+
+static int cpu_set(struct context *ctx, unsigned long offset, u32 v)
+{
+	unsigned int needs_clflush;
+	struct page *page;
+	void *map;
+	u32 *cpu;
+	int err;
+
+	i915_gem_object_lock(ctx->obj, NULL);
+	err = i915_gem_object_prepare_write(ctx->obj, &needs_clflush);
+	if (err)
+		goto out;
+
+	page = i915_gem_object_get_page(ctx->obj, offset >> PAGE_SHIFT);
+	map = kmap_atomic(page);
+	cpu = map + offset_in_page(offset);
+
+	if (needs_clflush & CLFLUSH_BEFORE)
+		drm_clflush_virt_range(cpu, sizeof(*cpu));
+
+	*cpu = v;
+
+	if (needs_clflush & CLFLUSH_AFTER)
+		drm_clflush_virt_range(cpu, sizeof(*cpu));
+
+	kunmap_atomic(map);
+	i915_gem_object_finish_access(ctx->obj);
+
+out:
+	i915_gem_object_unlock(ctx->obj);
+	return err;
+}
+
+static int cpu_get(struct context *ctx, unsigned long offset, u32 *v)
+{
+	unsigned int needs_clflush;
+	struct page *page;
+	void *map;
+	u32 *cpu;
+	int err;
+
+	i915_gem_object_lock(ctx->obj, NULL);
+	err = i915_gem_object_prepare_read(ctx->obj, &needs_clflush);
+	if (err)
+		goto out;
+
+	page = i915_gem_object_get_page(ctx->obj, offset >> PAGE_SHIFT);
+	map = kmap_atomic(page);
+	cpu = map + offset_in_page(offset);
+
+	if (needs_clflush & CLFLUSH_BEFORE)
+		drm_clflush_virt_range(cpu, sizeof(*cpu));
+
+	*v = *cpu;
+
+	kunmap_atomic(map);
+	i915_gem_object_finish_access(ctx->obj);
+
+out:
+	i915_gem_object_unlock(ctx->obj);
+	return err;
+}
+
+static int gtt_set(struct context *ctx, unsigned long offset, u32 v)
+{
+	struct i915_vma *vma;
+	u32 __iomem *map;
+	int err = 0;
+
+	i915_gem_object_lock(ctx->obj, NULL);
+	err = i915_gem_object_set_to_gtt_domain(ctx->obj, true);
+	i915_gem_object_unlock(ctx->obj);
+	if (err)
+		return err;
+
+	vma = i915_gem_object_ggtt_pin(ctx->obj, NULL, 0, 0, PIN_MAPPABLE);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	intel_gt_pm_get(vma->vm->gt);
+
+	map = i915_vma_pin_iomap(vma);
+	i915_vma_unpin(vma);
+	if (IS_ERR(map)) {
+		err = PTR_ERR(map);
+		goto out_rpm;
+	}
+
+	iowrite32(v, &map[offset / sizeof(*map)]);
+	i915_vma_unpin_iomap(vma);
+
+out_rpm:
+	intel_gt_pm_put(vma->vm->gt);
+	return err;
+}
+
+static int gtt_get(struct context *ctx, unsigned long offset, u32 *v)
+{
+	struct i915_vma *vma;
+	u32 __iomem *map;
+	int err = 0;
+
+	i915_gem_object_lock(ctx->obj, NULL);
+	err = i915_gem_object_set_to_gtt_domain(ctx->obj, false);
+	i915_gem_object_unlock(ctx->obj);
+	if (err)
+		return err;
+
+	vma = i915_gem_object_ggtt_pin(ctx->obj, NULL, 0, 0, PIN_MAPPABLE);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	intel_gt_pm_get(vma->vm->gt);
+
+	map = i915_vma_pin_iomap(vma);
+	i915_vma_unpin(vma);
+	if (IS_ERR(map)) {
+		err = PTR_ERR(map);
+		goto out_rpm;
+	}
+
+	*v = ioread32(&map[offset / sizeof(*map)]);
+	i915_vma_unpin_iomap(vma);
+
+out_rpm:
+	intel_gt_pm_put(vma->vm->gt);
+	return err;
+}
+
+static int wc_set(struct context *ctx, unsigned long offset, u32 v)
+{
+	u32 *map;
+	int err;
+
+	i915_gem_object_lock(ctx->obj, NULL);
+	err = i915_gem_object_set_to_wc_domain(ctx->obj, true);
+	i915_gem_object_unlock(ctx->obj);
+	if (err)
+		return err;
+
+	map = i915_gem_object_pin_map_unlocked(ctx->obj, I915_MAP_WC);
+	if (IS_ERR(map))
+		return PTR_ERR(map);
+
+	map[offset / sizeof(*map)] = v;
+
+	__i915_gem_object_flush_map(ctx->obj, offset, sizeof(*map));
+	i915_gem_object_unpin_map(ctx->obj);
+
+	return 0;
+}
+
+static int wc_get(struct context *ctx, unsigned long offset, u32 *v)
+{
+	u32 *map;
+	int err;
+
+	i915_gem_object_lock(ctx->obj, NULL);
+	err = i915_gem_object_set_to_wc_domain(ctx->obj, false);
+	i915_gem_object_unlock(ctx->obj);
+	if (err)
+		return err;
+
+	map = i915_gem_object_pin_map_unlocked(ctx->obj, I915_MAP_WC);
+	if (IS_ERR(map))
+		return PTR_ERR(map);
+
+	*v = map[offset / sizeof(*map)];
+	i915_gem_object_unpin_map(ctx->obj);
+
+	return 0;
+}
+
+static int gpu_set(struct context *ctx, unsigned long offset, u32 v)
+{
+	struct i915_request *rq;
+	struct i915_vma *vma;
+	u32 *cs;
+	int err;
+
+	vma = i915_gem_object_ggtt_pin(ctx->obj, NULL, 0, 0, 0);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	i915_gem_object_lock(ctx->obj, NULL);
+	err = i915_gem_object_set_to_gtt_domain(ctx->obj, true);
+	if (err)
+		goto out_unlock;
+
+	rq = intel_engine_create_kernel_request(ctx->engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_unpin;
+	}
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto out_rq;
+	}
+
+	if (GRAPHICS_VER(ctx->engine->i915) >= 8) {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | 1 << 22;
+		*cs++ = lower_32_bits(i915_ggtt_offset(vma) + offset);
+		*cs++ = upper_32_bits(i915_ggtt_offset(vma) + offset);
+		*cs++ = v;
+	} else if (GRAPHICS_VER(ctx->engine->i915) >= 4) {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = 0;
+		*cs++ = i915_ggtt_offset(vma) + offset;
+		*cs++ = v;
+	} else {
+		*cs++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL;
+		*cs++ = i915_ggtt_offset(vma) + offset;
+		*cs++ = v;
+		*cs++ = MI_NOOP;
+	}
+	intel_ring_advance(rq, cs);
+
+	err = i915_request_await_object(rq, vma->obj, true);
+	if (err == 0)
+		err = i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE);
+
+out_rq:
+	i915_request_add(rq);
+out_unpin:
+	i915_vma_unpin(vma);
+out_unlock:
+	i915_gem_object_unlock(ctx->obj);
+
+	return err;
+}
+
+static bool always_valid(struct context *ctx)
+{
+	return true;
+}
+
+static bool needs_fence_registers(struct context *ctx)
+{
+	struct intel_gt *gt = ctx->engine->gt;
+
+	if (intel_gt_is_wedged(gt))
+		return false;
+
+	return gt->ggtt->num_fences;
+}
+
+static bool needs_mi_store_dword(struct context *ctx)
+{
+	if (intel_gt_is_wedged(ctx->engine->gt))
+		return false;
+
+	return intel_engine_can_store_dword(ctx->engine);
+}
+
+static const struct igt_coherency_mode {
+	const char *name;
+	int (*set)(struct context *ctx, unsigned long offset, u32 v);
+	int (*get)(struct context *ctx, unsigned long offset, u32 *v);
+	bool (*valid)(struct context *ctx);
+} igt_coherency_mode[] = {
+	{ "cpu", cpu_set, cpu_get, always_valid },
+	{ "gtt", gtt_set, gtt_get, needs_fence_registers },
+	{ "wc", wc_set, wc_get, always_valid },
+	{ "gpu", gpu_set, NULL, needs_mi_store_dword },
+	{ },
+};
+
+static struct intel_engine_cs *
+random_engine(struct drm_i915_private *i915, struct rnd_state *prng)
+{
+	struct intel_engine_cs *engine;
+	unsigned int count;
+
+	count = 0;
+	for_each_uabi_engine(engine, i915)
+		count++;
+
+	count = i915_prandom_u32_max_state(count, prng);
+	for_each_uabi_engine(engine, i915)
+		if (count-- == 0)
+			return engine;
+
+	return NULL;
+}
+
+static int igt_gem_coherency(void *arg)
+{
+	const unsigned int ncachelines = PAGE_SIZE/64;
+	struct drm_i915_private *i915 = arg;
+	const struct igt_coherency_mode *read, *write, *over;
+	unsigned long count, n;
+	u32 *offsets, *values;
+	I915_RND_STATE(prng);
+	struct context ctx;
+	int err = 0;
+
+	/*
+	 * We repeatedly write, overwrite and read from a sequence of
+	 * cachelines in order to try and detect incoherency (unflushed writes
+	 * from either the CPU or GPU). Each setter/getter uses our cache
+	 * domain API which should prevent incoherency.
+	 */
+
+	offsets = kmalloc_array(ncachelines, 2*sizeof(u32), GFP_KERNEL);
+	if (!offsets)
+		return -ENOMEM;
+	for (count = 0; count < ncachelines; count++)
+		offsets[count] = count * 64 + 4 * (count % 16);
+
+	values = offsets + ncachelines;
+
+	ctx.engine = random_engine(i915, &prng);
+	if (!ctx.engine) {
+		err = -ENODEV;
+		goto out_free;
+	}
+	pr_info("%s: using %s\n", __func__, ctx.engine->name);
+	intel_engine_pm_get(ctx.engine);
+
+	for (over = igt_coherency_mode; over->name; over++) {
+		if (!over->set)
+			continue;
+
+		if (!over->valid(&ctx))
+			continue;
+
+		for (write = igt_coherency_mode; write->name; write++) {
+			if (!write->set)
+				continue;
+
+			if (!write->valid(&ctx))
+				continue;
+
+			for (read = igt_coherency_mode; read->name; read++) {
+				if (!read->get)
+					continue;
+
+				if (!read->valid(&ctx))
+					continue;
+
+				for_each_prime_number_from(count, 1, ncachelines) {
+					ctx.obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+					if (IS_ERR(ctx.obj)) {
+						err = PTR_ERR(ctx.obj);
+						goto out_pm;
+					}
+
+					i915_random_reorder(offsets, ncachelines, &prng);
+					for (n = 0; n < count; n++)
+						values[n] = prandom_u32_state(&prng);
+
+					for (n = 0; n < count; n++) {
+						err = over->set(&ctx, offsets[n], ~values[n]);
+						if (err) {
+							pr_err("Failed to set stale value[%ld/%ld] in object using %s, err=%d\n",
+							       n, count, over->name, err);
+							goto put_object;
+						}
+					}
+
+					for (n = 0; n < count; n++) {
+						err = write->set(&ctx, offsets[n], values[n]);
+						if (err) {
+							pr_err("Failed to set value[%ld/%ld] in object using %s, err=%d\n",
+							       n, count, write->name, err);
+							goto put_object;
+						}
+					}
+
+					for (n = 0; n < count; n++) {
+						u32 found;
+
+						err = read->get(&ctx, offsets[n], &found);
+						if (err) {
+							pr_err("Failed to get value[%ld/%ld] in object using %s, err=%d\n",
+							       n, count, read->name, err);
+							goto put_object;
+						}
+
+						if (found != values[n]) {
+							pr_err("Value[%ld/%ld] mismatch, (overwrite with %s) wrote [%s] %x read [%s] %x (inverse %x), at offset %x\n",
+							       n, count, over->name,
+							       write->name, values[n],
+							       read->name, found,
+							       ~values[n], offsets[n]);
+							err = -EINVAL;
+							goto put_object;
+						}
+					}
+
+					i915_gem_object_put(ctx.obj);
+				}
+			}
+		}
+	}
+out_pm:
+	intel_engine_pm_put(ctx.engine);
+out_free:
+	kfree(offsets);
+	return err;
+
+put_object:
+	i915_gem_object_put(ctx.obj);
+	goto out_pm;
+}
+
+int i915_gem_coherency_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_gem_coherency),
+	};
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c
new file mode 100644
index 000000000..a4858be12
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c
@@ -0,0 +1,1920 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2017 Intel Corporation
+ */
+
+#include <linux/prime_numbers.h>
+#include <linux/string_helpers.h>
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_requests.h"
+#include "gt/intel_reset.h"
+#include "i915_selftest.h"
+
+#include "gem/selftests/igt_gem_utils.h"
+#include "selftests/i915_random.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/igt_live_test.h"
+#include "selftests/igt_reset.h"
+#include "selftests/igt_spinner.h"
+#include "selftests/mock_drm.h"
+#include "selftests/mock_gem_device.h"
+
+#include "huge_gem_object.h"
+#include "igt_gem_utils.h"
+
+#define DW_PER_PAGE (PAGE_SIZE / sizeof(u32))
+
+static int live_nop_switch(void *arg)
+{
+	const unsigned int nctx = 1024;
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	struct i915_gem_context **ctx;
+	struct igt_live_test t;
+	struct file *file;
+	unsigned long n;
+	int err = -ENODEV;
+
+	/*
+	 * Create as many contexts as we can feasibly get away with
+	 * and check we can switch between them rapidly.
+	 *
+	 * Serves as very simple stress test for submission and HW switching
+	 * between contexts.
+	 */
+
+	if (!DRIVER_CAPS(i915)->has_logical_contexts)
+		return 0;
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	ctx = kcalloc(nctx, sizeof(*ctx), GFP_KERNEL);
+	if (!ctx) {
+		err = -ENOMEM;
+		goto out_file;
+	}
+
+	for (n = 0; n < nctx; n++) {
+		ctx[n] = live_context(i915, file);
+		if (IS_ERR(ctx[n])) {
+			err = PTR_ERR(ctx[n]);
+			goto out_file;
+		}
+	}
+
+	for_each_uabi_engine(engine, i915) {
+		struct i915_request *rq = NULL;
+		unsigned long end_time, prime;
+		ktime_t times[2] = {};
+
+		times[0] = ktime_get_raw();
+		for (n = 0; n < nctx; n++) {
+			struct i915_request *this;
+
+			this = igt_request_alloc(ctx[n], engine);
+			if (IS_ERR(this)) {
+				err = PTR_ERR(this);
+				goto out_file;
+			}
+			if (rq) {
+				i915_request_await_dma_fence(this, &rq->fence);
+				i915_request_put(rq);
+			}
+			rq = i915_request_get(this);
+			i915_request_add(this);
+		}
+		if (i915_request_wait(rq, 0, 10 * HZ) < 0) {
+			pr_err("Failed to populated %d contexts\n", nctx);
+			intel_gt_set_wedged(to_gt(i915));
+			i915_request_put(rq);
+			err = -EIO;
+			goto out_file;
+		}
+		i915_request_put(rq);
+
+		times[1] = ktime_get_raw();
+
+		pr_info("Populated %d contexts on %s in %lluns\n",
+			nctx, engine->name, ktime_to_ns(times[1] - times[0]));
+
+		err = igt_live_test_begin(&t, i915, __func__, engine->name);
+		if (err)
+			goto out_file;
+
+		end_time = jiffies + i915_selftest.timeout_jiffies;
+		for_each_prime_number_from(prime, 2, 8192) {
+			times[1] = ktime_get_raw();
+
+			rq = NULL;
+			for (n = 0; n < prime; n++) {
+				struct i915_request *this;
+
+				this = igt_request_alloc(ctx[n % nctx], engine);
+				if (IS_ERR(this)) {
+					err = PTR_ERR(this);
+					goto out_file;
+				}
+
+				if (rq) { /* Force submission order */
+					i915_request_await_dma_fence(this, &rq->fence);
+					i915_request_put(rq);
+				}
+
+				/*
+				 * This space is left intentionally blank.
+				 *
+				 * We do not actually want to perform any
+				 * action with this request, we just want
+				 * to measure the latency in allocation
+				 * and submission of our breadcrumbs -
+				 * ensuring that the bare request is sufficient
+				 * for the system to work (i.e. proper HEAD
+				 * tracking of the rings, interrupt handling,
+				 * etc). It also gives us the lowest bounds
+				 * for latency.
+				 */
+
+				rq = i915_request_get(this);
+				i915_request_add(this);
+			}
+			GEM_BUG_ON(!rq);
+			if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+				pr_err("Switching between %ld contexts timed out\n",
+				       prime);
+				intel_gt_set_wedged(to_gt(i915));
+				i915_request_put(rq);
+				break;
+			}
+			i915_request_put(rq);
+
+			times[1] = ktime_sub(ktime_get_raw(), times[1]);
+			if (prime == 2)
+				times[0] = times[1];
+
+			if (__igt_timeout(end_time, NULL))
+				break;
+		}
+
+		err = igt_live_test_end(&t);
+		if (err)
+			goto out_file;
+
+		pr_info("Switch latencies on %s: 1 = %lluns, %lu = %lluns\n",
+			engine->name,
+			ktime_to_ns(times[0]),
+			prime - 1, div64_u64(ktime_to_ns(times[1]), prime - 1));
+	}
+
+out_file:
+	fput(file);
+	return err;
+}
+
+struct parallel_switch {
+	struct kthread_worker *worker;
+	struct kthread_work work;
+	struct intel_context *ce[2];
+	int result;
+};
+
+static void __live_parallel_switch1(struct kthread_work *work)
+{
+	struct parallel_switch *arg =
+		container_of(work, typeof(*arg), work);
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+
+	count = 0;
+	arg->result = 0;
+	do {
+		struct i915_request *rq = NULL;
+		int n;
+
+		for (n = 0; !arg->result && n < ARRAY_SIZE(arg->ce); n++) {
+			struct i915_request *prev = rq;
+
+			rq = i915_request_create(arg->ce[n]);
+			if (IS_ERR(rq)) {
+				i915_request_put(prev);
+				arg->result = PTR_ERR(rq);
+				break;
+			}
+
+			i915_request_get(rq);
+			if (prev) {
+				arg->result =
+					i915_request_await_dma_fence(rq,
+								     &prev->fence);
+				i915_request_put(prev);
+			}
+
+			i915_request_add(rq);
+		}
+
+		if (IS_ERR_OR_NULL(rq))
+			break;
+
+		if (i915_request_wait(rq, 0, HZ) < 0)
+			arg->result = -ETIME;
+
+		i915_request_put(rq);
+
+		count++;
+	} while (!arg->result && !__igt_timeout(end_time, NULL));
+
+	pr_info("%s: %lu switches (sync) <%d>\n",
+		arg->ce[0]->engine->name, count, arg->result);
+}
+
+static void __live_parallel_switchN(struct kthread_work *work)
+{
+	struct parallel_switch *arg =
+		container_of(work, typeof(*arg), work);
+	struct i915_request *rq = NULL;
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+	int n;
+
+	count = 0;
+	arg->result = 0;
+	do {
+		for (n = 0; !arg->result && n < ARRAY_SIZE(arg->ce); n++) {
+			struct i915_request *prev = rq;
+
+			rq = i915_request_create(arg->ce[n]);
+			if (IS_ERR(rq)) {
+				i915_request_put(prev);
+				arg->result = PTR_ERR(rq);
+				break;
+			}
+
+			i915_request_get(rq);
+			if (prev) {
+				arg->result =
+					i915_request_await_dma_fence(rq,
+								     &prev->fence);
+				i915_request_put(prev);
+			}
+
+			i915_request_add(rq);
+		}
+
+		count++;
+	} while (!arg->result && !__igt_timeout(end_time, NULL));
+
+	if (!IS_ERR_OR_NULL(rq))
+		i915_request_put(rq);
+
+	pr_info("%s: %lu switches (many) <%d>\n",
+		arg->ce[0]->engine->name, count, arg->result);
+}
+
+static int live_parallel_switch(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	static void (* const func[])(struct kthread_work *) = {
+		__live_parallel_switch1,
+		__live_parallel_switchN,
+		NULL,
+	};
+	struct parallel_switch *data = NULL;
+	struct i915_gem_engines *engines;
+	struct i915_gem_engines_iter it;
+	void (* const *fn)(struct kthread_work *);
+	struct i915_gem_context *ctx;
+	struct intel_context *ce;
+	struct file *file;
+	int n, m, count;
+	int err = 0;
+
+	/*
+	 * Check we can process switches on all engines simultaneously.
+	 */
+
+	if (!DRIVER_CAPS(i915)->has_logical_contexts)
+		return 0;
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	ctx = live_context(i915, file);
+	if (IS_ERR(ctx)) {
+		err = PTR_ERR(ctx);
+		goto out_file;
+	}
+
+	engines = i915_gem_context_lock_engines(ctx);
+	count = engines->num_engines;
+
+	data = kcalloc(count, sizeof(*data), GFP_KERNEL);
+	if (!data) {
+		i915_gem_context_unlock_engines(ctx);
+		err = -ENOMEM;
+		goto out_file;
+	}
+
+	m = 0; /* Use the first context as our template for the engines */
+	for_each_gem_engine(ce, engines, it) {
+		err = intel_context_pin(ce);
+		if (err) {
+			i915_gem_context_unlock_engines(ctx);
+			goto out;
+		}
+		data[m++].ce[0] = intel_context_get(ce);
+	}
+	i915_gem_context_unlock_engines(ctx);
+
+	/* Clone the same set of engines into the other contexts */
+	for (n = 1; n < ARRAY_SIZE(data->ce); n++) {
+		ctx = live_context(i915, file);
+		if (IS_ERR(ctx)) {
+			err = PTR_ERR(ctx);
+			goto out;
+		}
+
+		for (m = 0; m < count; m++) {
+			if (!data[m].ce[0])
+				continue;
+
+			ce = intel_context_create(data[m].ce[0]->engine);
+			if (IS_ERR(ce)) {
+				err = PTR_ERR(ce);
+				goto out;
+			}
+
+			err = intel_context_pin(ce);
+			if (err) {
+				intel_context_put(ce);
+				goto out;
+			}
+
+			data[m].ce[n] = ce;
+		}
+	}
+
+	for (n = 0; n < count; n++) {
+		struct kthread_worker *worker;
+
+		if (!data[n].ce[0])
+			continue;
+
+		worker = kthread_create_worker(0, "igt/parallel:%s",
+					       data[n].ce[0]->engine->name);
+		if (IS_ERR(worker)) {
+			err = PTR_ERR(worker);
+			goto out;
+		}
+
+		data[n].worker = worker;
+	}
+
+	for (fn = func; !err && *fn; fn++) {
+		struct igt_live_test t;
+
+		err = igt_live_test_begin(&t, i915, __func__, "");
+		if (err)
+			break;
+
+		for (n = 0; n < count; n++) {
+			if (!data[n].ce[0])
+				continue;
+
+			data[n].result = 0;
+			kthread_init_work(&data[n].work, *fn);
+			kthread_queue_work(data[n].worker, &data[n].work);
+		}
+
+		for (n = 0; n < count; n++) {
+			if (data[n].ce[0]) {
+				kthread_flush_work(&data[n].work);
+				if (data[n].result && !err)
+					err = data[n].result;
+			}
+		}
+
+		if (igt_live_test_end(&t)) {
+			err = err ?: -EIO;
+			break;
+		}
+	}
+
+out:
+	for (n = 0; n < count; n++) {
+		for (m = 0; m < ARRAY_SIZE(data->ce); m++) {
+			if (!data[n].ce[m])
+				continue;
+
+			intel_context_unpin(data[n].ce[m]);
+			intel_context_put(data[n].ce[m]);
+		}
+
+		if (data[n].worker)
+			kthread_destroy_worker(data[n].worker);
+	}
+	kfree(data);
+out_file:
+	fput(file);
+	return err;
+}
+
+static unsigned long real_page_count(struct drm_i915_gem_object *obj)
+{
+	return huge_gem_object_phys_size(obj) >> PAGE_SHIFT;
+}
+
+static unsigned long fake_page_count(struct drm_i915_gem_object *obj)
+{
+	return huge_gem_object_dma_size(obj) >> PAGE_SHIFT;
+}
+
+static int gpu_fill(struct intel_context *ce,
+		    struct drm_i915_gem_object *obj,
+		    unsigned int dw)
+{
+	struct i915_vma *vma;
+	int err;
+
+	GEM_BUG_ON(obj->base.size > ce->vm->total);
+	GEM_BUG_ON(!intel_engine_can_store_dword(ce->engine));
+
+	vma = i915_vma_instance(obj, ce->vm, NULL);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	err = i915_vma_pin(vma, 0, 0, PIN_HIGH | PIN_USER);
+	if (err)
+		return err;
+
+	/*
+	 * Within the GTT the huge objects maps every page onto
+	 * its 1024 real pages (using phys_pfn = dma_pfn % 1024).
+	 * We set the nth dword within the page using the nth
+	 * mapping via the GTT - this should exercise the GTT mapping
+	 * whilst checking that each context provides a unique view
+	 * into the object.
+	 */
+	err = igt_gpu_fill_dw(ce, vma,
+			      (dw * real_page_count(obj)) << PAGE_SHIFT |
+			      (dw * sizeof(u32)),
+			      real_page_count(obj),
+			      dw);
+	i915_vma_unpin(vma);
+
+	return err;
+}
+
+static int cpu_fill(struct drm_i915_gem_object *obj, u32 value)
+{
+	const bool has_llc = HAS_LLC(to_i915(obj->base.dev));
+	unsigned int n, m, need_flush;
+	int err;
+
+	i915_gem_object_lock(obj, NULL);
+	err = i915_gem_object_prepare_write(obj, &need_flush);
+	if (err)
+		goto out;
+
+	for (n = 0; n < real_page_count(obj); n++) {
+		u32 *map;
+
+		map = kmap_atomic(i915_gem_object_get_page(obj, n));
+		for (m = 0; m < DW_PER_PAGE; m++)
+			map[m] = value;
+		if (!has_llc)
+			drm_clflush_virt_range(map, PAGE_SIZE);
+		kunmap_atomic(map);
+	}
+
+	i915_gem_object_finish_access(obj);
+	obj->read_domains = I915_GEM_DOMAIN_GTT | I915_GEM_DOMAIN_CPU;
+	obj->write_domain = 0;
+out:
+	i915_gem_object_unlock(obj);
+	return err;
+}
+
+static noinline int cpu_check(struct drm_i915_gem_object *obj,
+			      unsigned int idx, unsigned int max)
+{
+	unsigned int n, m, needs_flush;
+	int err;
+
+	i915_gem_object_lock(obj, NULL);
+	err = i915_gem_object_prepare_read(obj, &needs_flush);
+	if (err)
+		goto out_unlock;
+
+	for (n = 0; n < real_page_count(obj); n++) {
+		u32 *map;
+
+		map = kmap_atomic(i915_gem_object_get_page(obj, n));
+		if (needs_flush & CLFLUSH_BEFORE)
+			drm_clflush_virt_range(map, PAGE_SIZE);
+
+		for (m = 0; m < max; m++) {
+			if (map[m] != m) {
+				pr_err("%pS: Invalid value at object %d page %d/%ld, offset %d/%d: found %x expected %x\n",
+				       __builtin_return_address(0), idx,
+				       n, real_page_count(obj), m, max,
+				       map[m], m);
+				err = -EINVAL;
+				goto out_unmap;
+			}
+		}
+
+		for (; m < DW_PER_PAGE; m++) {
+			if (map[m] != STACK_MAGIC) {
+				pr_err("%pS: Invalid value at object %d page %d, offset %d: found %x expected %x (uninitialised)\n",
+				       __builtin_return_address(0), idx, n, m,
+				       map[m], STACK_MAGIC);
+				err = -EINVAL;
+				goto out_unmap;
+			}
+		}
+
+out_unmap:
+		kunmap_atomic(map);
+		if (err)
+			break;
+	}
+
+	i915_gem_object_finish_access(obj);
+out_unlock:
+	i915_gem_object_unlock(obj);
+	return err;
+}
+
+static int file_add_object(struct file *file, struct drm_i915_gem_object *obj)
+{
+	int err;
+
+	GEM_BUG_ON(obj->base.handle_count);
+
+	/* tie the object to the drm_file for easy reaping */
+	err = idr_alloc(&to_drm_file(file)->object_idr,
+			&obj->base, 1, 0, GFP_KERNEL);
+	if (err < 0)
+		return err;
+
+	i915_gem_object_get(obj);
+	obj->base.handle_count++;
+	return 0;
+}
+
+static struct drm_i915_gem_object *
+create_test_object(struct i915_address_space *vm,
+		   struct file *file,
+		   struct list_head *objects)
+{
+	struct drm_i915_gem_object *obj;
+	u64 size;
+	int err;
+
+	/* Keep in GEM's good graces */
+	intel_gt_retire_requests(vm->gt);
+
+	size = min(vm->total / 2, 1024ull * DW_PER_PAGE * PAGE_SIZE);
+	size = round_down(size, DW_PER_PAGE * PAGE_SIZE);
+
+	obj = huge_gem_object(vm->i915, DW_PER_PAGE * PAGE_SIZE, size);
+	if (IS_ERR(obj))
+		return obj;
+
+	err = file_add_object(file, obj);
+	i915_gem_object_put(obj);
+	if (err)
+		return ERR_PTR(err);
+
+	err = cpu_fill(obj, STACK_MAGIC);
+	if (err) {
+		pr_err("Failed to fill object with cpu, err=%d\n",
+		       err);
+		return ERR_PTR(err);
+	}
+
+	list_add_tail(&obj->st_link, objects);
+	return obj;
+}
+
+static unsigned long max_dwords(struct drm_i915_gem_object *obj)
+{
+	unsigned long npages = fake_page_count(obj);
+
+	GEM_BUG_ON(!IS_ALIGNED(npages, DW_PER_PAGE));
+	return npages / DW_PER_PAGE;
+}
+
+static void throttle_release(struct i915_request **q, int count)
+{
+	int i;
+
+	for (i = 0; i < count; i++) {
+		if (IS_ERR_OR_NULL(q[i]))
+			continue;
+
+		i915_request_put(fetch_and_zero(&q[i]));
+	}
+}
+
+static int throttle(struct intel_context *ce,
+		    struct i915_request **q, int count)
+{
+	int i;
+
+	if (!IS_ERR_OR_NULL(q[0])) {
+		if (i915_request_wait(q[0],
+				      I915_WAIT_INTERRUPTIBLE,
+				      MAX_SCHEDULE_TIMEOUT) < 0)
+			return -EINTR;
+
+		i915_request_put(q[0]);
+	}
+
+	for (i = 0; i < count - 1; i++)
+		q[i] = q[i + 1];
+
+	q[i] = intel_context_create_request(ce);
+	if (IS_ERR(q[i]))
+		return PTR_ERR(q[i]);
+
+	i915_request_get(q[i]);
+	i915_request_add(q[i]);
+
+	return 0;
+}
+
+static int igt_ctx_exec(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	int err = -ENODEV;
+
+	/*
+	 * Create a few different contexts (with different mm) and write
+	 * through each ctx/mm using the GPU making sure those writes end
+	 * up in the expected pages of our obj.
+	 */
+
+	if (!DRIVER_CAPS(i915)->has_logical_contexts)
+		return 0;
+
+	for_each_uabi_engine(engine, i915) {
+		struct drm_i915_gem_object *obj = NULL;
+		unsigned long ncontexts, ndwords, dw;
+		struct i915_request *tq[5] = {};
+		struct igt_live_test t;
+		IGT_TIMEOUT(end_time);
+		LIST_HEAD(objects);
+		struct file *file;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		if (!engine->context_size)
+			continue; /* No logical context support in HW */
+
+		file = mock_file(i915);
+		if (IS_ERR(file))
+			return PTR_ERR(file);
+
+		err = igt_live_test_begin(&t, i915, __func__, engine->name);
+		if (err)
+			goto out_file;
+
+		ncontexts = 0;
+		ndwords = 0;
+		dw = 0;
+		while (!time_after(jiffies, end_time)) {
+			struct i915_gem_context *ctx;
+			struct intel_context *ce;
+
+			ctx = kernel_context(i915, NULL);
+			if (IS_ERR(ctx)) {
+				err = PTR_ERR(ctx);
+				goto out_file;
+			}
+
+			ce = i915_gem_context_get_engine(ctx, engine->legacy_idx);
+			GEM_BUG_ON(IS_ERR(ce));
+
+			if (!obj) {
+				obj = create_test_object(ce->vm, file, &objects);
+				if (IS_ERR(obj)) {
+					err = PTR_ERR(obj);
+					intel_context_put(ce);
+					kernel_context_close(ctx);
+					goto out_file;
+				}
+			}
+
+			err = gpu_fill(ce, obj, dw);
+			if (err) {
+				pr_err("Failed to fill dword %lu [%lu/%lu] with gpu (%s) [full-ppgtt? %s], err=%d\n",
+				       ndwords, dw, max_dwords(obj),
+				       engine->name,
+				       str_yes_no(i915_gem_context_has_full_ppgtt(ctx)),
+				       err);
+				intel_context_put(ce);
+				kernel_context_close(ctx);
+				goto out_file;
+			}
+
+			err = throttle(ce, tq, ARRAY_SIZE(tq));
+			if (err) {
+				intel_context_put(ce);
+				kernel_context_close(ctx);
+				goto out_file;
+			}
+
+			if (++dw == max_dwords(obj)) {
+				obj = NULL;
+				dw = 0;
+			}
+
+			ndwords++;
+			ncontexts++;
+
+			intel_context_put(ce);
+			kernel_context_close(ctx);
+		}
+
+		pr_info("Submitted %lu contexts to %s, filling %lu dwords\n",
+			ncontexts, engine->name, ndwords);
+
+		ncontexts = dw = 0;
+		list_for_each_entry(obj, &objects, st_link) {
+			unsigned int rem =
+				min_t(unsigned int, ndwords - dw, max_dwords(obj));
+
+			err = cpu_check(obj, ncontexts++, rem);
+			if (err)
+				break;
+
+			dw += rem;
+		}
+
+out_file:
+		throttle_release(tq, ARRAY_SIZE(tq));
+		if (igt_live_test_end(&t))
+			err = -EIO;
+
+		fput(file);
+		if (err)
+			return err;
+
+		i915_gem_drain_freed_objects(i915);
+	}
+
+	return 0;
+}
+
+static int igt_shared_ctx_exec(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_request *tq[5] = {};
+	struct i915_gem_context *parent;
+	struct intel_engine_cs *engine;
+	struct igt_live_test t;
+	struct file *file;
+	int err = 0;
+
+	/*
+	 * Create a few different contexts with the same mm and write
+	 * through each ctx using the GPU making sure those writes end
+	 * up in the expected pages of our obj.
+	 */
+	if (!DRIVER_CAPS(i915)->has_logical_contexts)
+		return 0;
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	parent = live_context(i915, file);
+	if (IS_ERR(parent)) {
+		err = PTR_ERR(parent);
+		goto out_file;
+	}
+
+	if (!parent->vm) { /* not full-ppgtt; nothing to share */
+		err = 0;
+		goto out_file;
+	}
+
+	err = igt_live_test_begin(&t, i915, __func__, "");
+	if (err)
+		goto out_file;
+
+	for_each_uabi_engine(engine, i915) {
+		unsigned long ncontexts, ndwords, dw;
+		struct drm_i915_gem_object *obj = NULL;
+		IGT_TIMEOUT(end_time);
+		LIST_HEAD(objects);
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		dw = 0;
+		ndwords = 0;
+		ncontexts = 0;
+		while (!time_after(jiffies, end_time)) {
+			struct i915_gem_context *ctx;
+			struct intel_context *ce;
+
+			ctx = kernel_context(i915, parent->vm);
+			if (IS_ERR(ctx)) {
+				err = PTR_ERR(ctx);
+				goto out_test;
+			}
+
+			ce = i915_gem_context_get_engine(ctx, engine->legacy_idx);
+			GEM_BUG_ON(IS_ERR(ce));
+
+			if (!obj) {
+				obj = create_test_object(parent->vm,
+							 file, &objects);
+				if (IS_ERR(obj)) {
+					err = PTR_ERR(obj);
+					intel_context_put(ce);
+					kernel_context_close(ctx);
+					goto out_test;
+				}
+			}
+
+			err = gpu_fill(ce, obj, dw);
+			if (err) {
+				pr_err("Failed to fill dword %lu [%lu/%lu] with gpu (%s) [full-ppgtt? %s], err=%d\n",
+				       ndwords, dw, max_dwords(obj),
+				       engine->name,
+				       str_yes_no(i915_gem_context_has_full_ppgtt(ctx)),
+				       err);
+				intel_context_put(ce);
+				kernel_context_close(ctx);
+				goto out_test;
+			}
+
+			err = throttle(ce, tq, ARRAY_SIZE(tq));
+			if (err) {
+				intel_context_put(ce);
+				kernel_context_close(ctx);
+				goto out_test;
+			}
+
+			if (++dw == max_dwords(obj)) {
+				obj = NULL;
+				dw = 0;
+			}
+
+			ndwords++;
+			ncontexts++;
+
+			intel_context_put(ce);
+			kernel_context_close(ctx);
+		}
+		pr_info("Submitted %lu contexts to %s, filling %lu dwords\n",
+			ncontexts, engine->name, ndwords);
+
+		ncontexts = dw = 0;
+		list_for_each_entry(obj, &objects, st_link) {
+			unsigned int rem =
+				min_t(unsigned int, ndwords - dw, max_dwords(obj));
+
+			err = cpu_check(obj, ncontexts++, rem);
+			if (err)
+				goto out_test;
+
+			dw += rem;
+		}
+
+		i915_gem_drain_freed_objects(i915);
+	}
+out_test:
+	throttle_release(tq, ARRAY_SIZE(tq));
+	if (igt_live_test_end(&t))
+		err = -EIO;
+out_file:
+	fput(file);
+	return err;
+}
+
+static int rpcs_query_batch(struct drm_i915_gem_object *rpcs,
+			    struct i915_vma *vma,
+			    struct intel_engine_cs *engine)
+{
+	u32 *cmd;
+
+	GEM_BUG_ON(GRAPHICS_VER(vma->vm->i915) < 8);
+
+	cmd = i915_gem_object_pin_map(rpcs, I915_MAP_WB);
+	if (IS_ERR(cmd))
+		return PTR_ERR(cmd);
+
+	*cmd++ = MI_STORE_REGISTER_MEM_GEN8;
+	*cmd++ = i915_mmio_reg_offset(GEN8_R_PWR_CLK_STATE(engine->mmio_base));
+	*cmd++ = lower_32_bits(vma->node.start);
+	*cmd++ = upper_32_bits(vma->node.start);
+	*cmd = MI_BATCH_BUFFER_END;
+
+	__i915_gem_object_flush_map(rpcs, 0, 64);
+	i915_gem_object_unpin_map(rpcs);
+
+	intel_gt_chipset_flush(vma->vm->gt);
+
+	return 0;
+}
+
+static int
+emit_rpcs_query(struct drm_i915_gem_object *obj,
+		struct intel_context *ce,
+		struct i915_request **rq_out)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_request *rq;
+	struct i915_gem_ww_ctx ww;
+	struct i915_vma *batch;
+	struct i915_vma *vma;
+	struct drm_i915_gem_object *rpcs;
+	int err;
+
+	GEM_BUG_ON(!intel_engine_can_store_dword(ce->engine));
+
+	if (GRAPHICS_VER(i915) < 8)
+		return -EINVAL;
+
+	vma = i915_vma_instance(obj, ce->vm, NULL);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	rpcs = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(rpcs))
+		return PTR_ERR(rpcs);
+
+	batch = i915_vma_instance(rpcs, ce->vm, NULL);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		goto err_put;
+	}
+
+	i915_gem_ww_ctx_init(&ww, false);
+retry:
+	err = i915_gem_object_lock(obj, &ww);
+	if (!err)
+		err = i915_gem_object_lock(rpcs, &ww);
+	if (!err)
+		err = i915_gem_object_set_to_gtt_domain(obj, false);
+	if (!err)
+		err = i915_vma_pin_ww(vma, &ww, 0, 0, PIN_USER);
+	if (err)
+		goto err_put;
+
+	err = i915_vma_pin_ww(batch, &ww, 0, 0, PIN_USER);
+	if (err)
+		goto err_vma;
+
+	err = rpcs_query_batch(rpcs, vma, ce->engine);
+	if (err)
+		goto err_batch;
+
+	rq = i915_request_create(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_batch;
+	}
+
+	err = i915_request_await_object(rq, batch->obj, false);
+	if (err == 0)
+		err = i915_vma_move_to_active(batch, rq, 0);
+	if (err)
+		goto skip_request;
+
+	err = i915_request_await_object(rq, vma->obj, true);
+	if (err == 0)
+		err = i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE);
+	if (err)
+		goto skip_request;
+
+	if (rq->engine->emit_init_breadcrumb) {
+		err = rq->engine->emit_init_breadcrumb(rq);
+		if (err)
+			goto skip_request;
+	}
+
+	err = rq->engine->emit_bb_start(rq,
+					batch->node.start, batch->node.size,
+					0);
+	if (err)
+		goto skip_request;
+
+	*rq_out = i915_request_get(rq);
+
+skip_request:
+	if (err)
+		i915_request_set_error_once(rq, err);
+	i915_request_add(rq);
+err_batch:
+	i915_vma_unpin(batch);
+err_vma:
+	i915_vma_unpin(vma);
+err_put:
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	i915_gem_object_put(rpcs);
+	return err;
+}
+
+#define TEST_IDLE	BIT(0)
+#define TEST_BUSY	BIT(1)
+#define TEST_RESET	BIT(2)
+
+static int
+__sseu_prepare(const char *name,
+	       unsigned int flags,
+	       struct intel_context *ce,
+	       struct igt_spinner **spin)
+{
+	struct i915_request *rq;
+	int ret;
+
+	*spin = NULL;
+	if (!(flags & (TEST_BUSY | TEST_RESET)))
+		return 0;
+
+	*spin = kzalloc(sizeof(**spin), GFP_KERNEL);
+	if (!*spin)
+		return -ENOMEM;
+
+	ret = igt_spinner_init(*spin, ce->engine->gt);
+	if (ret)
+		goto err_free;
+
+	rq = igt_spinner_create_request(*spin, ce, MI_NOOP);
+	if (IS_ERR(rq)) {
+		ret = PTR_ERR(rq);
+		goto err_fini;
+	}
+
+	i915_request_add(rq);
+
+	if (!igt_wait_for_spinner(*spin, rq)) {
+		pr_err("%s: Spinner failed to start!\n", name);
+		ret = -ETIMEDOUT;
+		goto err_end;
+	}
+
+	return 0;
+
+err_end:
+	igt_spinner_end(*spin);
+err_fini:
+	igt_spinner_fini(*spin);
+err_free:
+	kfree(fetch_and_zero(spin));
+	return ret;
+}
+
+static int
+__read_slice_count(struct intel_context *ce,
+		   struct drm_i915_gem_object *obj,
+		   struct igt_spinner *spin,
+		   u32 *rpcs)
+{
+	struct i915_request *rq = NULL;
+	u32 s_mask, s_shift;
+	unsigned int cnt;
+	u32 *buf, val;
+	long ret;
+
+	ret = emit_rpcs_query(obj, ce, &rq);
+	if (ret)
+		return ret;
+
+	if (spin)
+		igt_spinner_end(spin);
+
+	ret = i915_request_wait(rq, 0, MAX_SCHEDULE_TIMEOUT);
+	i915_request_put(rq);
+	if (ret < 0)
+		return ret;
+
+	buf = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+	if (IS_ERR(buf)) {
+		ret = PTR_ERR(buf);
+		return ret;
+	}
+
+	if (GRAPHICS_VER(ce->engine->i915) >= 11) {
+		s_mask = GEN11_RPCS_S_CNT_MASK;
+		s_shift = GEN11_RPCS_S_CNT_SHIFT;
+	} else {
+		s_mask = GEN8_RPCS_S_CNT_MASK;
+		s_shift = GEN8_RPCS_S_CNT_SHIFT;
+	}
+
+	val = *buf;
+	cnt = (val & s_mask) >> s_shift;
+	*rpcs = val;
+
+	i915_gem_object_unpin_map(obj);
+
+	return cnt;
+}
+
+static int
+__check_rpcs(const char *name, u32 rpcs, int slices, unsigned int expected,
+	     const char *prefix, const char *suffix)
+{
+	if (slices == expected)
+		return 0;
+
+	if (slices < 0) {
+		pr_err("%s: %s read slice count failed with %d%s\n",
+		       name, prefix, slices, suffix);
+		return slices;
+	}
+
+	pr_err("%s: %s slice count %d is not %u%s\n",
+	       name, prefix, slices, expected, suffix);
+
+	pr_info("RPCS=0x%x; %u%sx%u%s\n",
+		rpcs, slices,
+		(rpcs & GEN8_RPCS_S_CNT_ENABLE) ? "*" : "",
+		(rpcs & GEN8_RPCS_SS_CNT_MASK) >> GEN8_RPCS_SS_CNT_SHIFT,
+		(rpcs & GEN8_RPCS_SS_CNT_ENABLE) ? "*" : "");
+
+	return -EINVAL;
+}
+
+static int
+__sseu_finish(const char *name,
+	      unsigned int flags,
+	      struct intel_context *ce,
+	      struct drm_i915_gem_object *obj,
+	      unsigned int expected,
+	      struct igt_spinner *spin)
+{
+	unsigned int slices = hweight32(ce->engine->sseu.slice_mask);
+	u32 rpcs = 0;
+	int ret = 0;
+
+	if (flags & TEST_RESET) {
+		ret = intel_engine_reset(ce->engine, "sseu");
+		if (ret)
+			goto out;
+	}
+
+	ret = __read_slice_count(ce, obj,
+				 flags & TEST_RESET ? NULL : spin, &rpcs);
+	ret = __check_rpcs(name, rpcs, ret, expected, "Context", "!");
+	if (ret)
+		goto out;
+
+	ret = __read_slice_count(ce->engine->kernel_context, obj, NULL, &rpcs);
+	ret = __check_rpcs(name, rpcs, ret, slices, "Kernel context", "!");
+
+out:
+	if (spin)
+		igt_spinner_end(spin);
+
+	if ((flags & TEST_IDLE) && ret == 0) {
+		ret = igt_flush_test(ce->engine->i915);
+		if (ret)
+			return ret;
+
+		ret = __read_slice_count(ce, obj, NULL, &rpcs);
+		ret = __check_rpcs(name, rpcs, ret, expected,
+				   "Context", " after idle!");
+	}
+
+	return ret;
+}
+
+static int
+__sseu_test(const char *name,
+	    unsigned int flags,
+	    struct intel_context *ce,
+	    struct drm_i915_gem_object *obj,
+	    struct intel_sseu sseu)
+{
+	struct igt_spinner *spin = NULL;
+	int ret;
+
+	intel_engine_pm_get(ce->engine);
+
+	ret = __sseu_prepare(name, flags, ce, &spin);
+	if (ret)
+		goto out_pm;
+
+	ret = intel_context_reconfigure_sseu(ce, sseu);
+	if (ret)
+		goto out_spin;
+
+	ret = __sseu_finish(name, flags, ce, obj,
+			    hweight32(sseu.slice_mask), spin);
+
+out_spin:
+	if (spin) {
+		igt_spinner_end(spin);
+		igt_spinner_fini(spin);
+		kfree(spin);
+	}
+out_pm:
+	intel_engine_pm_put(ce->engine);
+	return ret;
+}
+
+static int
+__igt_ctx_sseu(struct drm_i915_private *i915,
+	       const char *name,
+	       unsigned int flags)
+{
+	struct drm_i915_gem_object *obj;
+	int inst = 0;
+	int ret = 0;
+
+	if (GRAPHICS_VER(i915) < 9)
+		return 0;
+
+	if (flags & TEST_RESET)
+		igt_global_reset_lock(to_gt(i915));
+
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		ret = PTR_ERR(obj);
+		goto out_unlock;
+	}
+
+	do {
+		struct intel_engine_cs *engine;
+		struct intel_context *ce;
+		struct intel_sseu pg_sseu;
+
+		engine = intel_engine_lookup_user(i915,
+						  I915_ENGINE_CLASS_RENDER,
+						  inst++);
+		if (!engine)
+			break;
+
+		if (hweight32(engine->sseu.slice_mask) < 2)
+			continue;
+
+		if (!engine->gt->info.sseu.has_slice_pg)
+			continue;
+
+		/*
+		 * Gen11 VME friendly power-gated configuration with
+		 * half enabled sub-slices.
+		 */
+		pg_sseu = engine->sseu;
+		pg_sseu.slice_mask = 1;
+		pg_sseu.subslice_mask =
+			~(~0 << (hweight32(engine->sseu.subslice_mask) / 2));
+
+		pr_info("%s: SSEU subtest '%s', flags=%x, def_slices=%u, pg_slices=%u\n",
+			engine->name, name, flags,
+			hweight32(engine->sseu.slice_mask),
+			hweight32(pg_sseu.slice_mask));
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			ret = PTR_ERR(ce);
+			goto out_put;
+		}
+
+		ret = intel_context_pin(ce);
+		if (ret)
+			goto out_ce;
+
+		/* First set the default mask. */
+		ret = __sseu_test(name, flags, ce, obj, engine->sseu);
+		if (ret)
+			goto out_unpin;
+
+		/* Then set a power-gated configuration. */
+		ret = __sseu_test(name, flags, ce, obj, pg_sseu);
+		if (ret)
+			goto out_unpin;
+
+		/* Back to defaults. */
+		ret = __sseu_test(name, flags, ce, obj, engine->sseu);
+		if (ret)
+			goto out_unpin;
+
+		/* One last power-gated configuration for the road. */
+		ret = __sseu_test(name, flags, ce, obj, pg_sseu);
+		if (ret)
+			goto out_unpin;
+
+out_unpin:
+		intel_context_unpin(ce);
+out_ce:
+		intel_context_put(ce);
+	} while (!ret);
+
+	if (igt_flush_test(i915))
+		ret = -EIO;
+
+out_put:
+	i915_gem_object_put(obj);
+
+out_unlock:
+	if (flags & TEST_RESET)
+		igt_global_reset_unlock(to_gt(i915));
+
+	if (ret)
+		pr_err("%s: Failed with %d!\n", name, ret);
+
+	return ret;
+}
+
+static int igt_ctx_sseu(void *arg)
+{
+	struct {
+		const char *name;
+		unsigned int flags;
+	} *phase, phases[] = {
+		{ .name = "basic", .flags = 0 },
+		{ .name = "idle", .flags = TEST_IDLE },
+		{ .name = "busy", .flags = TEST_BUSY },
+		{ .name = "busy-reset", .flags = TEST_BUSY | TEST_RESET },
+		{ .name = "busy-idle", .flags = TEST_BUSY | TEST_IDLE },
+		{ .name = "reset-idle", .flags = TEST_RESET | TEST_IDLE },
+	};
+	unsigned int i;
+	int ret = 0;
+
+	for (i = 0, phase = phases; ret == 0 && i < ARRAY_SIZE(phases);
+	     i++, phase++)
+		ret = __igt_ctx_sseu(arg, phase->name, phase->flags);
+
+	return ret;
+}
+
+static int igt_ctx_readonly(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	unsigned long idx, ndwords, dw, num_engines;
+	struct drm_i915_gem_object *obj = NULL;
+	struct i915_request *tq[5] = {};
+	struct i915_gem_engines_iter it;
+	struct i915_address_space *vm;
+	struct i915_gem_context *ctx;
+	struct intel_context *ce;
+	struct igt_live_test t;
+	I915_RND_STATE(prng);
+	IGT_TIMEOUT(end_time);
+	LIST_HEAD(objects);
+	struct file *file;
+	int err = -ENODEV;
+
+	/*
+	 * Create a few read-only objects (with the occasional writable object)
+	 * and try to write into these object checking that the GPU discards
+	 * any write to a read-only object.
+	 */
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	err = igt_live_test_begin(&t, i915, __func__, "");
+	if (err)
+		goto out_file;
+
+	ctx = live_context(i915, file);
+	if (IS_ERR(ctx)) {
+		err = PTR_ERR(ctx);
+		goto out_file;
+	}
+
+	vm = ctx->vm ?: &to_gt(i915)->ggtt->alias->vm;
+	if (!vm || !vm->has_read_only) {
+		err = 0;
+		goto out_file;
+	}
+
+	num_engines = 0;
+	for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it)
+		if (intel_engine_can_store_dword(ce->engine))
+			num_engines++;
+	i915_gem_context_unlock_engines(ctx);
+
+	ndwords = 0;
+	dw = 0;
+	while (!time_after(jiffies, end_time)) {
+		for_each_gem_engine(ce,
+				    i915_gem_context_lock_engines(ctx), it) {
+			if (!intel_engine_can_store_dword(ce->engine))
+				continue;
+
+			if (!obj) {
+				obj = create_test_object(ce->vm, file, &objects);
+				if (IS_ERR(obj)) {
+					err = PTR_ERR(obj);
+					i915_gem_context_unlock_engines(ctx);
+					goto out_file;
+				}
+
+				if (prandom_u32_state(&prng) & 1)
+					i915_gem_object_set_readonly(obj);
+			}
+
+			err = gpu_fill(ce, obj, dw);
+			if (err) {
+				pr_err("Failed to fill dword %lu [%lu/%lu] with gpu (%s) [full-ppgtt? %s], err=%d\n",
+				       ndwords, dw, max_dwords(obj),
+				       ce->engine->name,
+				       str_yes_no(i915_gem_context_has_full_ppgtt(ctx)),
+				       err);
+				i915_gem_context_unlock_engines(ctx);
+				goto out_file;
+			}
+
+			err = throttle(ce, tq, ARRAY_SIZE(tq));
+			if (err) {
+				i915_gem_context_unlock_engines(ctx);
+				goto out_file;
+			}
+
+			if (++dw == max_dwords(obj)) {
+				obj = NULL;
+				dw = 0;
+			}
+			ndwords++;
+		}
+		i915_gem_context_unlock_engines(ctx);
+	}
+	pr_info("Submitted %lu dwords (across %lu engines)\n",
+		ndwords, num_engines);
+
+	dw = 0;
+	idx = 0;
+	list_for_each_entry(obj, &objects, st_link) {
+		unsigned int rem =
+			min_t(unsigned int, ndwords - dw, max_dwords(obj));
+		unsigned int num_writes;
+
+		num_writes = rem;
+		if (i915_gem_object_is_readonly(obj))
+			num_writes = 0;
+
+		err = cpu_check(obj, idx++, num_writes);
+		if (err)
+			break;
+
+		dw += rem;
+	}
+
+out_file:
+	throttle_release(tq, ARRAY_SIZE(tq));
+	if (igt_live_test_end(&t))
+		err = -EIO;
+
+	fput(file);
+	return err;
+}
+
+static int check_scratch(struct i915_address_space *vm, u64 offset)
+{
+	struct drm_mm_node *node;
+
+	mutex_lock(&vm->mutex);
+	node = __drm_mm_interval_first(&vm->mm,
+				       offset, offset + sizeof(u32) - 1);
+	mutex_unlock(&vm->mutex);
+	if (!node || node->start > offset)
+		return 0;
+
+	GEM_BUG_ON(offset >= node->start + node->size);
+
+	pr_err("Target offset 0x%08x_%08x overlaps with a node in the mm!\n",
+	       upper_32_bits(offset), lower_32_bits(offset));
+	return -EINVAL;
+}
+
+static int write_to_scratch(struct i915_gem_context *ctx,
+			    struct intel_engine_cs *engine,
+			    struct drm_i915_gem_object *obj,
+			    u64 offset, u32 value)
+{
+	struct drm_i915_private *i915 = ctx->i915;
+	struct i915_address_space *vm;
+	struct i915_request *rq;
+	struct i915_vma *vma;
+	u32 *cmd;
+	int err;
+
+	GEM_BUG_ON(offset < I915_GTT_PAGE_SIZE);
+
+	err = check_scratch(ctx->vm, offset);
+	if (err)
+		return err;
+
+	cmd = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+	if (IS_ERR(cmd))
+		return PTR_ERR(cmd);
+
+	*cmd++ = MI_STORE_DWORD_IMM_GEN4;
+	if (GRAPHICS_VER(i915) >= 8) {
+		*cmd++ = lower_32_bits(offset);
+		*cmd++ = upper_32_bits(offset);
+	} else {
+		*cmd++ = 0;
+		*cmd++ = offset;
+	}
+	*cmd++ = value;
+	*cmd = MI_BATCH_BUFFER_END;
+	__i915_gem_object_flush_map(obj, 0, 64);
+	i915_gem_object_unpin_map(obj);
+
+	intel_gt_chipset_flush(engine->gt);
+
+	vm = i915_gem_context_get_eb_vm(ctx);
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto out_vm;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER | PIN_OFFSET_FIXED);
+	if (err)
+		goto out_vm;
+
+	rq = igt_request_alloc(ctx, engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_unpin;
+	}
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, false);
+	if (err == 0)
+		err = i915_vma_move_to_active(vma, rq, 0);
+	i915_vma_unlock(vma);
+	if (err)
+		goto skip_request;
+
+	if (rq->engine->emit_init_breadcrumb) {
+		err = rq->engine->emit_init_breadcrumb(rq);
+		if (err)
+			goto skip_request;
+	}
+
+	err = engine->emit_bb_start(rq, vma->node.start, vma->node.size, 0);
+	if (err)
+		goto skip_request;
+
+	i915_vma_unpin(vma);
+
+	i915_request_add(rq);
+
+	goto out_vm;
+skip_request:
+	i915_request_set_error_once(rq, err);
+	i915_request_add(rq);
+err_unpin:
+	i915_vma_unpin(vma);
+out_vm:
+	i915_vm_put(vm);
+
+	if (!err)
+		err = i915_gem_object_wait(obj, 0, MAX_SCHEDULE_TIMEOUT);
+
+	return err;
+}
+
+static int read_from_scratch(struct i915_gem_context *ctx,
+			     struct intel_engine_cs *engine,
+			     struct drm_i915_gem_object *obj,
+			     u64 offset, u32 *value)
+{
+	struct drm_i915_private *i915 = ctx->i915;
+	struct i915_address_space *vm;
+	const u32 result = 0x100;
+	struct i915_request *rq;
+	struct i915_vma *vma;
+	unsigned int flags;
+	u32 *cmd;
+	int err;
+
+	GEM_BUG_ON(offset < I915_GTT_PAGE_SIZE);
+
+	err = check_scratch(ctx->vm, offset);
+	if (err)
+		return err;
+
+	if (GRAPHICS_VER(i915) >= 8) {
+		const u32 GPR0 = engine->mmio_base + 0x600;
+
+		vm = i915_gem_context_get_eb_vm(ctx);
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out_vm;
+		}
+
+		err = i915_vma_pin(vma, 0, 0, PIN_USER | PIN_OFFSET_FIXED);
+		if (err)
+			goto out_vm;
+
+		cmd = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+		if (IS_ERR(cmd)) {
+			err = PTR_ERR(cmd);
+			goto err_unpin;
+		}
+
+		memset(cmd, POISON_INUSE, PAGE_SIZE);
+		*cmd++ = MI_LOAD_REGISTER_MEM_GEN8;
+		*cmd++ = GPR0;
+		*cmd++ = lower_32_bits(offset);
+		*cmd++ = upper_32_bits(offset);
+		*cmd++ = MI_STORE_REGISTER_MEM_GEN8;
+		*cmd++ = GPR0;
+		*cmd++ = result;
+		*cmd++ = 0;
+		*cmd = MI_BATCH_BUFFER_END;
+
+		i915_gem_object_flush_map(obj);
+		i915_gem_object_unpin_map(obj);
+
+		flags = 0;
+	} else {
+		const u32 reg = engine->mmio_base + 0x420;
+
+		/* hsw: register access even to 3DPRIM! is protected */
+		vm = i915_vm_get(&engine->gt->ggtt->vm);
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out_vm;
+		}
+
+		err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+		if (err)
+			goto out_vm;
+
+		cmd = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+		if (IS_ERR(cmd)) {
+			err = PTR_ERR(cmd);
+			goto err_unpin;
+		}
+
+		memset(cmd, POISON_INUSE, PAGE_SIZE);
+		*cmd++ = MI_LOAD_REGISTER_MEM;
+		*cmd++ = reg;
+		*cmd++ = offset;
+		*cmd++ = MI_STORE_REGISTER_MEM | MI_USE_GGTT;
+		*cmd++ = reg;
+		*cmd++ = vma->node.start + result;
+		*cmd = MI_BATCH_BUFFER_END;
+
+		i915_gem_object_flush_map(obj);
+		i915_gem_object_unpin_map(obj);
+
+		flags = I915_DISPATCH_SECURE;
+	}
+
+	intel_gt_chipset_flush(engine->gt);
+
+	rq = igt_request_alloc(ctx, engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_unpin;
+	}
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, true);
+	if (err == 0)
+		err = i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE);
+	i915_vma_unlock(vma);
+	if (err)
+		goto skip_request;
+
+	if (rq->engine->emit_init_breadcrumb) {
+		err = rq->engine->emit_init_breadcrumb(rq);
+		if (err)
+			goto skip_request;
+	}
+
+	err = engine->emit_bb_start(rq, vma->node.start, vma->node.size, flags);
+	if (err)
+		goto skip_request;
+
+	i915_vma_unpin(vma);
+
+	i915_request_add(rq);
+
+	i915_gem_object_lock(obj, NULL);
+	err = i915_gem_object_set_to_cpu_domain(obj, false);
+	i915_gem_object_unlock(obj);
+	if (err)
+		goto out_vm;
+
+	cmd = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+	if (IS_ERR(cmd)) {
+		err = PTR_ERR(cmd);
+		goto out_vm;
+	}
+
+	*value = cmd[result / sizeof(*cmd)];
+	i915_gem_object_unpin_map(obj);
+
+	goto out_vm;
+skip_request:
+	i915_request_set_error_once(rq, err);
+	i915_request_add(rq);
+err_unpin:
+	i915_vma_unpin(vma);
+out_vm:
+	i915_vm_put(vm);
+
+	if (!err)
+		err = i915_gem_object_wait(obj, 0, MAX_SCHEDULE_TIMEOUT);
+
+	return err;
+}
+
+static int check_scratch_page(struct i915_gem_context *ctx, u32 *out)
+{
+	struct i915_address_space *vm;
+	u32 *vaddr;
+	int err = 0;
+
+	vm = ctx->vm;
+	if (!vm)
+		return -ENODEV;
+
+	if (!vm->scratch[0]) {
+		pr_err("No scratch page!\n");
+		return -EINVAL;
+	}
+
+	vaddr = __px_vaddr(vm->scratch[0]);
+
+	memcpy(out, vaddr, sizeof(*out));
+	if (memchr_inv(vaddr, *out, PAGE_SIZE)) {
+		pr_err("Inconsistent initial state of scratch page!\n");
+		err = -EINVAL;
+	}
+
+	return err;
+}
+
+static int igt_vm_isolation(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_gem_context *ctx_a, *ctx_b;
+	struct drm_i915_gem_object *obj_a, *obj_b;
+	unsigned long num_engines, count;
+	struct intel_engine_cs *engine;
+	struct igt_live_test t;
+	I915_RND_STATE(prng);
+	struct file *file;
+	u64 vm_total;
+	u32 expected;
+	int err;
+
+	if (GRAPHICS_VER(i915) < 7)
+		return 0;
+
+	/*
+	 * The simple goal here is that a write into one context is not
+	 * observed in a second (separate page tables and scratch).
+	 */
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	err = igt_live_test_begin(&t, i915, __func__, "");
+	if (err)
+		goto out_file;
+
+	ctx_a = live_context(i915, file);
+	if (IS_ERR(ctx_a)) {
+		err = PTR_ERR(ctx_a);
+		goto out_file;
+	}
+
+	ctx_b = live_context(i915, file);
+	if (IS_ERR(ctx_b)) {
+		err = PTR_ERR(ctx_b);
+		goto out_file;
+	}
+
+	/* We can only test vm isolation, if the vm are distinct */
+	if (ctx_a->vm == ctx_b->vm)
+		goto out_file;
+
+	/* Read the initial state of the scratch page */
+	err = check_scratch_page(ctx_a, &expected);
+	if (err)
+		goto out_file;
+
+	err = check_scratch_page(ctx_b, &expected);
+	if (err)
+		goto out_file;
+
+	vm_total = ctx_a->vm->total;
+	GEM_BUG_ON(ctx_b->vm->total != vm_total);
+
+	obj_a = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj_a)) {
+		err = PTR_ERR(obj_a);
+		goto out_file;
+	}
+
+	obj_b = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj_b)) {
+		err = PTR_ERR(obj_b);
+		goto put_a;
+	}
+
+	count = 0;
+	num_engines = 0;
+	for_each_uabi_engine(engine, i915) {
+		IGT_TIMEOUT(end_time);
+		unsigned long this = 0;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		/* Not all engines have their own GPR! */
+		if (GRAPHICS_VER(i915) < 8 && engine->class != RENDER_CLASS)
+			continue;
+
+		while (!__igt_timeout(end_time, NULL)) {
+			u32 value = 0xc5c5c5c5;
+			u64 offset;
+
+			/* Leave enough space at offset 0 for the batch */
+			offset = igt_random_offset(&prng,
+						   I915_GTT_PAGE_SIZE, vm_total,
+						   sizeof(u32), alignof_dword);
+
+			err = write_to_scratch(ctx_a, engine, obj_a,
+					       offset, 0xdeadbeef);
+			if (err == 0)
+				err = read_from_scratch(ctx_b, engine, obj_b,
+							offset, &value);
+			if (err)
+				goto put_b;
+
+			if (value != expected) {
+				pr_err("%s: Read %08x from scratch (offset 0x%08x_%08x), after %lu reads!\n",
+				       engine->name, value,
+				       upper_32_bits(offset),
+				       lower_32_bits(offset),
+				       this);
+				err = -EINVAL;
+				goto put_b;
+			}
+
+			this++;
+		}
+		count += this;
+		num_engines++;
+	}
+	pr_info("Checked %lu scratch offsets across %lu engines\n",
+		count, num_engines);
+
+put_b:
+	i915_gem_object_put(obj_b);
+put_a:
+	i915_gem_object_put(obj_a);
+out_file:
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	fput(file);
+	return err;
+}
+
+int i915_gem_context_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_nop_switch),
+		SUBTEST(live_parallel_switch),
+		SUBTEST(igt_ctx_exec),
+		SUBTEST(igt_ctx_readonly),
+		SUBTEST(igt_ctx_sseu),
+		SUBTEST(igt_shared_ctx_exec),
+		SUBTEST(igt_vm_isolation),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c
new file mode 100644
index 000000000..51ed824b0
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c
@@ -0,0 +1,480 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_selftest.h"
+
+#include "mock_dmabuf.h"
+#include "selftests/mock_gem_device.h"
+
+static int igt_dmabuf_export(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct dma_buf *dmabuf;
+
+	obj = i915_gem_object_create_shmem(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	dmabuf = i915_gem_prime_export(&obj->base, 0);
+	i915_gem_object_put(obj);
+	if (IS_ERR(dmabuf)) {
+		pr_err("i915_gem_prime_export failed with err=%d\n",
+		       (int)PTR_ERR(dmabuf));
+		return PTR_ERR(dmabuf);
+	}
+
+	dma_buf_put(dmabuf);
+	return 0;
+}
+
+static int igt_dmabuf_import_self(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj, *import_obj;
+	struct drm_gem_object *import;
+	struct dma_buf *dmabuf;
+	int err;
+
+	obj = i915_gem_object_create_shmem(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	dmabuf = i915_gem_prime_export(&obj->base, 0);
+	if (IS_ERR(dmabuf)) {
+		pr_err("i915_gem_prime_export failed with err=%d\n",
+		       (int)PTR_ERR(dmabuf));
+		err = PTR_ERR(dmabuf);
+		goto out;
+	}
+
+	import = i915_gem_prime_import(&i915->drm, dmabuf);
+	if (IS_ERR(import)) {
+		pr_err("i915_gem_prime_import failed with err=%d\n",
+		       (int)PTR_ERR(import));
+		err = PTR_ERR(import);
+		goto out_dmabuf;
+	}
+	import_obj = to_intel_bo(import);
+
+	if (import != &obj->base) {
+		pr_err("i915_gem_prime_import created a new object!\n");
+		err = -EINVAL;
+		goto out_import;
+	}
+
+	i915_gem_object_lock(import_obj, NULL);
+	err = __i915_gem_object_get_pages(import_obj);
+	i915_gem_object_unlock(import_obj);
+	if (err) {
+		pr_err("Same object dma-buf get_pages failed!\n");
+		goto out_import;
+	}
+
+	err = 0;
+out_import:
+	i915_gem_object_put(import_obj);
+out_dmabuf:
+	dma_buf_put(dmabuf);
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static int igt_dmabuf_import_same_driver_lmem(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_memory_region *lmem = i915->mm.regions[INTEL_REGION_LMEM_0];
+	struct drm_i915_gem_object *obj;
+	struct drm_gem_object *import;
+	struct dma_buf *dmabuf;
+	int err;
+
+	if (!lmem)
+		return 0;
+
+	force_different_devices = true;
+
+	obj = __i915_gem_object_create_user(i915, PAGE_SIZE, &lmem, 1);
+	if (IS_ERR(obj)) {
+		pr_err("__i915_gem_object_create_user failed with err=%ld\n",
+		       PTR_ERR(obj));
+		err = PTR_ERR(obj);
+		goto out_ret;
+	}
+
+	dmabuf = i915_gem_prime_export(&obj->base, 0);
+	if (IS_ERR(dmabuf)) {
+		pr_err("i915_gem_prime_export failed with err=%ld\n",
+		       PTR_ERR(dmabuf));
+		err = PTR_ERR(dmabuf);
+		goto out;
+	}
+
+	/*
+	 * We expect an import of an LMEM-only object to fail with
+	 * -EOPNOTSUPP because it can't be migrated to SMEM.
+	 */
+	import = i915_gem_prime_import(&i915->drm, dmabuf);
+	if (!IS_ERR(import)) {
+		drm_gem_object_put(import);
+		pr_err("i915_gem_prime_import succeeded when it shouldn't have\n");
+		err = -EINVAL;
+	} else if (PTR_ERR(import) != -EOPNOTSUPP) {
+		pr_err("i915_gem_prime_import failed with the wrong err=%ld\n",
+		       PTR_ERR(import));
+		err = PTR_ERR(import);
+	} else {
+		err = 0;
+	}
+
+	dma_buf_put(dmabuf);
+out:
+	i915_gem_object_put(obj);
+out_ret:
+	force_different_devices = false;
+	return err;
+}
+
+static int igt_dmabuf_import_same_driver(struct drm_i915_private *i915,
+					 struct intel_memory_region **regions,
+					 unsigned int num_regions)
+{
+	struct drm_i915_gem_object *obj, *import_obj;
+	struct drm_gem_object *import;
+	struct dma_buf *dmabuf;
+	struct dma_buf_attachment *import_attach;
+	struct sg_table *st;
+	long timeout;
+	int err;
+
+	force_different_devices = true;
+
+	obj = __i915_gem_object_create_user(i915, PAGE_SIZE,
+					    regions, num_regions);
+	if (IS_ERR(obj)) {
+		pr_err("__i915_gem_object_create_user failed with err=%ld\n",
+		       PTR_ERR(obj));
+		err = PTR_ERR(obj);
+		goto out_ret;
+	}
+
+	dmabuf = i915_gem_prime_export(&obj->base, 0);
+	if (IS_ERR(dmabuf)) {
+		pr_err("i915_gem_prime_export failed with err=%ld\n",
+		       PTR_ERR(dmabuf));
+		err = PTR_ERR(dmabuf);
+		goto out;
+	}
+
+	import = i915_gem_prime_import(&i915->drm, dmabuf);
+	if (IS_ERR(import)) {
+		pr_err("i915_gem_prime_import failed with err=%ld\n",
+		       PTR_ERR(import));
+		err = PTR_ERR(import);
+		goto out_dmabuf;
+	}
+	import_obj = to_intel_bo(import);
+
+	if (import == &obj->base) {
+		pr_err("i915_gem_prime_import reused gem object!\n");
+		err = -EINVAL;
+		goto out_import;
+	}
+
+	i915_gem_object_lock(import_obj, NULL);
+	err = __i915_gem_object_get_pages(import_obj);
+	if (err) {
+		pr_err("Different objects dma-buf get_pages failed!\n");
+		i915_gem_object_unlock(import_obj);
+		goto out_import;
+	}
+
+	/*
+	 * If the exported object is not in system memory, something
+	 * weird is going on. TODO: When p2p is supported, this is no
+	 * longer considered weird.
+	 */
+	if (obj->mm.region != i915->mm.regions[INTEL_REGION_SMEM]) {
+		pr_err("Exported dma-buf is not in system memory\n");
+		err = -EINVAL;
+	}
+
+	i915_gem_object_unlock(import_obj);
+
+	/* Now try a fake an importer */
+	import_attach = dma_buf_attach(dmabuf, obj->base.dev->dev);
+	if (IS_ERR(import_attach)) {
+		err = PTR_ERR(import_attach);
+		goto out_import;
+	}
+
+	st = dma_buf_map_attachment(import_attach, DMA_BIDIRECTIONAL);
+	if (IS_ERR(st)) {
+		err = PTR_ERR(st);
+		goto out_detach;
+	}
+
+	timeout = dma_resv_wait_timeout(dmabuf->resv, DMA_RESV_USAGE_WRITE,
+					true, 5 * HZ);
+	if (!timeout) {
+		pr_err("dmabuf wait for exclusive fence timed out.\n");
+		timeout = -ETIME;
+	}
+	err = timeout > 0 ? 0 : timeout;
+	dma_buf_unmap_attachment(import_attach, st, DMA_BIDIRECTIONAL);
+out_detach:
+	dma_buf_detach(dmabuf, import_attach);
+out_import:
+	i915_gem_object_put(import_obj);
+out_dmabuf:
+	dma_buf_put(dmabuf);
+out:
+	i915_gem_object_put(obj);
+out_ret:
+	force_different_devices = false;
+	return err;
+}
+
+static int igt_dmabuf_import_same_driver_smem(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_memory_region *smem = i915->mm.regions[INTEL_REGION_SMEM];
+
+	return igt_dmabuf_import_same_driver(i915, &smem, 1);
+}
+
+static int igt_dmabuf_import_same_driver_lmem_smem(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_memory_region *regions[2];
+
+	if (!i915->mm.regions[INTEL_REGION_LMEM_0])
+		return 0;
+
+	regions[0] = i915->mm.regions[INTEL_REGION_LMEM_0];
+	regions[1] = i915->mm.regions[INTEL_REGION_SMEM];
+	return igt_dmabuf_import_same_driver(i915, regions, 2);
+}
+
+static int igt_dmabuf_import(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct dma_buf *dmabuf;
+	void *obj_map, *dma_map;
+	struct iosys_map map;
+	u32 pattern[] = { 0, 0xaa, 0xcc, 0x55, 0xff };
+	int err, i;
+
+	dmabuf = mock_dmabuf(1);
+	if (IS_ERR(dmabuf))
+		return PTR_ERR(dmabuf);
+
+	obj = to_intel_bo(i915_gem_prime_import(&i915->drm, dmabuf));
+	if (IS_ERR(obj)) {
+		pr_err("i915_gem_prime_import failed with err=%d\n",
+		       (int)PTR_ERR(obj));
+		err = PTR_ERR(obj);
+		goto out_dmabuf;
+	}
+
+	if (obj->base.dev != &i915->drm) {
+		pr_err("i915_gem_prime_import created a non-i915 object!\n");
+		err = -EINVAL;
+		goto out_obj;
+	}
+
+	if (obj->base.size != PAGE_SIZE) {
+		pr_err("i915_gem_prime_import is wrong size found %lld, expected %ld\n",
+		       (long long)obj->base.size, PAGE_SIZE);
+		err = -EINVAL;
+		goto out_obj;
+	}
+
+	err = dma_buf_vmap(dmabuf, &map);
+	dma_map = err ? NULL : map.vaddr;
+	if (!dma_map) {
+		pr_err("dma_buf_vmap failed\n");
+		err = -ENOMEM;
+		goto out_obj;
+	}
+
+	if (0) { /* Can not yet map dmabuf */
+		obj_map = i915_gem_object_pin_map(obj, I915_MAP_WB);
+		if (IS_ERR(obj_map)) {
+			err = PTR_ERR(obj_map);
+			pr_err("i915_gem_object_pin_map failed with err=%d\n", err);
+			goto out_dma_map;
+		}
+
+		for (i = 0; i < ARRAY_SIZE(pattern); i++) {
+			memset(dma_map, pattern[i], PAGE_SIZE);
+			if (memchr_inv(obj_map, pattern[i], PAGE_SIZE)) {
+				err = -EINVAL;
+				pr_err("imported vmap not all set to %x!\n", pattern[i]);
+				i915_gem_object_unpin_map(obj);
+				goto out_dma_map;
+			}
+		}
+
+		for (i = 0; i < ARRAY_SIZE(pattern); i++) {
+			memset(obj_map, pattern[i], PAGE_SIZE);
+			if (memchr_inv(dma_map, pattern[i], PAGE_SIZE)) {
+				err = -EINVAL;
+				pr_err("exported vmap not all set to %x!\n", pattern[i]);
+				i915_gem_object_unpin_map(obj);
+				goto out_dma_map;
+			}
+		}
+
+		i915_gem_object_unpin_map(obj);
+	}
+
+	err = 0;
+out_dma_map:
+	dma_buf_vunmap(dmabuf, &map);
+out_obj:
+	i915_gem_object_put(obj);
+out_dmabuf:
+	dma_buf_put(dmabuf);
+	return err;
+}
+
+static int igt_dmabuf_import_ownership(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct dma_buf *dmabuf;
+	struct iosys_map map;
+	void *ptr;
+	int err;
+
+	dmabuf = mock_dmabuf(1);
+	if (IS_ERR(dmabuf))
+		return PTR_ERR(dmabuf);
+
+	err = dma_buf_vmap(dmabuf, &map);
+	ptr = err ? NULL : map.vaddr;
+	if (!ptr) {
+		pr_err("dma_buf_vmap failed\n");
+		err = -ENOMEM;
+		goto err_dmabuf;
+	}
+
+	memset(ptr, 0xc5, PAGE_SIZE);
+	dma_buf_vunmap(dmabuf, &map);
+
+	obj = to_intel_bo(i915_gem_prime_import(&i915->drm, dmabuf));
+	if (IS_ERR(obj)) {
+		pr_err("i915_gem_prime_import failed with err=%d\n",
+		       (int)PTR_ERR(obj));
+		err = PTR_ERR(obj);
+		goto err_dmabuf;
+	}
+
+	dma_buf_put(dmabuf);
+
+	err = i915_gem_object_pin_pages_unlocked(obj);
+	if (err) {
+		pr_err("i915_gem_object_pin_pages failed with err=%d\n", err);
+		goto out_obj;
+	}
+
+	err = 0;
+	i915_gem_object_unpin_pages(obj);
+out_obj:
+	i915_gem_object_put(obj);
+	return err;
+
+err_dmabuf:
+	dma_buf_put(dmabuf);
+	return err;
+}
+
+static int igt_dmabuf_export_vmap(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct dma_buf *dmabuf;
+	struct iosys_map map;
+	void *ptr;
+	int err;
+
+	obj = i915_gem_object_create_shmem(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	dmabuf = i915_gem_prime_export(&obj->base, 0);
+	if (IS_ERR(dmabuf)) {
+		pr_err("i915_gem_prime_export failed with err=%d\n",
+		       (int)PTR_ERR(dmabuf));
+		err = PTR_ERR(dmabuf);
+		goto err_obj;
+	}
+	i915_gem_object_put(obj);
+
+	err = dma_buf_vmap(dmabuf, &map);
+	ptr = err ? NULL : map.vaddr;
+	if (!ptr) {
+		pr_err("dma_buf_vmap failed\n");
+		err = -ENOMEM;
+		goto out;
+	}
+
+	if (memchr_inv(ptr, 0, dmabuf->size)) {
+		pr_err("Exported object not initialiased to zero!\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	memset(ptr, 0xc5, dmabuf->size);
+
+	err = 0;
+	dma_buf_vunmap(dmabuf, &map);
+out:
+	dma_buf_put(dmabuf);
+	return err;
+
+err_obj:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+int i915_gem_dmabuf_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_dmabuf_export),
+		SUBTEST(igt_dmabuf_import_self),
+		SUBTEST(igt_dmabuf_import),
+		SUBTEST(igt_dmabuf_import_ownership),
+		SUBTEST(igt_dmabuf_export_vmap),
+	};
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	err = i915_subtests(tests, i915);
+
+	mock_destroy_device(i915);
+	return err;
+}
+
+int i915_gem_dmabuf_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_dmabuf_export),
+		SUBTEST(igt_dmabuf_import_same_driver_lmem),
+		SUBTEST(igt_dmabuf_import_same_driver_smem),
+		SUBTEST(igt_dmabuf_import_same_driver_lmem_smem),
+	};
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_migrate.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_migrate.c
new file mode 100644
index 000000000..fe6c37fd7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_migrate.c
@@ -0,0 +1,528 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020-2021 Intel Corporation
+ */
+
+#include "gt/intel_migrate.h"
+#include "gt/intel_gpu_commands.h"
+#include "gem/i915_gem_ttm_move.h"
+
+#include "i915_deps.h"
+
+#include "selftests/igt_reset.h"
+#include "selftests/igt_spinner.h"
+
+static int igt_fill_check_buffer(struct drm_i915_gem_object *obj,
+				 bool fill)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	unsigned int i, count = obj->base.size / sizeof(u32);
+	enum i915_map_type map_type =
+		i915_coherent_map_type(i915, obj, false);
+	u32 *cur;
+	int err = 0;
+
+	assert_object_held(obj);
+	cur = i915_gem_object_pin_map(obj, map_type);
+	if (IS_ERR(cur))
+		return PTR_ERR(cur);
+
+	if (fill)
+		for (i = 0; i < count; ++i)
+			*cur++ = i;
+	else
+		for (i = 0; i < count; ++i)
+			if (*cur++ != i) {
+				pr_err("Object content mismatch at location %d of %d\n", i, count);
+				err = -EINVAL;
+				break;
+			}
+
+	i915_gem_object_unpin_map(obj);
+
+	return err;
+}
+
+static int igt_create_migrate(struct intel_gt *gt, enum intel_region_id src,
+			      enum intel_region_id dst)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_memory_region *src_mr = i915->mm.regions[src];
+	struct intel_memory_region *dst_mr = i915->mm.regions[dst];
+	struct drm_i915_gem_object *obj;
+	struct i915_gem_ww_ctx ww;
+	int err = 0;
+
+	GEM_BUG_ON(!src_mr);
+	GEM_BUG_ON(!dst_mr);
+
+	/* Switch object backing-store on create */
+	obj = i915_gem_object_create_region(src_mr, dst_mr->min_page_size, 0, 0);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	for_i915_gem_ww(&ww, err, true) {
+		err = i915_gem_object_lock(obj, &ww);
+		if (err)
+			continue;
+
+		err = igt_fill_check_buffer(obj, true);
+		if (err)
+			continue;
+
+		err = i915_gem_object_migrate(obj, &ww, dst);
+		if (err)
+			continue;
+
+		err = i915_gem_object_pin_pages(obj);
+		if (err)
+			continue;
+
+		if (i915_gem_object_can_migrate(obj, src))
+			err = -EINVAL;
+
+		i915_gem_object_unpin_pages(obj);
+		err = i915_gem_object_wait_migration(obj, true);
+		if (err)
+			continue;
+
+		err = igt_fill_check_buffer(obj, false);
+	}
+	i915_gem_object_put(obj);
+
+	return err;
+}
+
+static int igt_smem_create_migrate(void *arg)
+{
+	return igt_create_migrate(arg, INTEL_REGION_LMEM_0, INTEL_REGION_SMEM);
+}
+
+static int igt_lmem_create_migrate(void *arg)
+{
+	return igt_create_migrate(arg, INTEL_REGION_SMEM, INTEL_REGION_LMEM_0);
+}
+
+static int igt_same_create_migrate(void *arg)
+{
+	return igt_create_migrate(arg, INTEL_REGION_LMEM_0, INTEL_REGION_LMEM_0);
+}
+
+static int lmem_pages_migrate_one(struct i915_gem_ww_ctx *ww,
+				  struct drm_i915_gem_object *obj,
+				  struct i915_vma *vma,
+				  bool silent_migrate)
+{
+	int err;
+
+	err = i915_gem_object_lock(obj, ww);
+	if (err)
+		return err;
+
+	if (vma) {
+		err = i915_vma_pin_ww(vma, ww, obj->base.size, 0,
+				      0UL | PIN_OFFSET_FIXED |
+				      PIN_USER);
+		if (err) {
+			if (err != -EINTR && err != ERESTARTSYS &&
+			    err != -EDEADLK)
+				pr_err("Failed to pin vma.\n");
+			return err;
+		}
+
+		i915_vma_unpin(vma);
+	}
+
+	/*
+	 * Migration will implicitly unbind (asynchronously) any bound
+	 * vmas.
+	 */
+	if (i915_gem_object_is_lmem(obj)) {
+		err = i915_gem_object_migrate(obj, ww, INTEL_REGION_SMEM);
+		if (err) {
+			if (!silent_migrate)
+				pr_err("Object failed migration to smem\n");
+			if (err)
+				return err;
+		}
+
+		if (i915_gem_object_is_lmem(obj)) {
+			pr_err("object still backed by lmem\n");
+			err = -EINVAL;
+		}
+
+		if (!i915_gem_object_has_struct_page(obj)) {
+			pr_err("object not backed by struct page\n");
+			err = -EINVAL;
+		}
+
+	} else {
+		err = i915_gem_object_migrate(obj, ww, INTEL_REGION_LMEM_0);
+		if (err) {
+			if (!silent_migrate)
+				pr_err("Object failed migration to lmem\n");
+			if (err)
+				return err;
+		}
+
+		if (i915_gem_object_has_struct_page(obj)) {
+			pr_err("object still backed by struct page\n");
+			err = -EINVAL;
+		}
+
+		if (!i915_gem_object_is_lmem(obj)) {
+			pr_err("object not backed by lmem\n");
+			err = -EINVAL;
+		}
+	}
+
+	return err;
+}
+
+static int __igt_lmem_pages_migrate(struct intel_gt *gt,
+				    struct i915_address_space *vm,
+				    struct i915_deps *deps,
+				    struct igt_spinner *spin,
+				    struct dma_fence *spin_fence,
+				    bool borked_migrate)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma = NULL;
+	struct i915_gem_ww_ctx ww;
+	struct i915_request *rq;
+	int err;
+	int i;
+
+	/* From LMEM to shmem and back again */
+
+	obj = i915_gem_object_create_lmem(i915, SZ_2M, 0);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	if (vm) {
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out_put;
+		}
+	}
+
+	/* Initial GPU fill, sync, CPU initialization. */
+	for_i915_gem_ww(&ww, err, true) {
+		err = i915_gem_object_lock(obj, &ww);
+		if (err)
+			continue;
+
+		err = ____i915_gem_object_get_pages(obj);
+		if (err)
+			continue;
+
+		err = intel_migrate_clear(&gt->migrate, &ww, deps,
+					  obj->mm.pages->sgl, obj->cache_level,
+					  i915_gem_object_is_lmem(obj),
+					  0xdeadbeaf, &rq);
+		if (rq) {
+			err = dma_resv_reserve_fences(obj->base.resv, 1);
+			if (!err)
+				dma_resv_add_fence(obj->base.resv, &rq->fence,
+						   DMA_RESV_USAGE_KERNEL);
+			i915_request_put(rq);
+		}
+		if (err)
+			continue;
+
+		if (!vma) {
+			err = igt_fill_check_buffer(obj, true);
+			if (err)
+				continue;
+		}
+	}
+	if (err)
+		goto out_put;
+
+	/*
+	 * Migrate to and from smem without explicitly syncing.
+	 * Finalize with data in smem for fast readout.
+	 */
+	for (i = 1; i <= 5; ++i) {
+		for_i915_gem_ww(&ww, err, true)
+			err = lmem_pages_migrate_one(&ww, obj, vma,
+						     borked_migrate);
+		if (err)
+			goto out_put;
+	}
+
+	err = i915_gem_object_lock_interruptible(obj, NULL);
+	if (err)
+		goto out_put;
+
+	if (spin) {
+		if (dma_fence_is_signaled(spin_fence)) {
+			pr_err("Spinner was terminated by hangcheck.\n");
+			err = -EBUSY;
+			goto out_unlock;
+		}
+		igt_spinner_end(spin);
+	}
+
+	/* Finally sync migration and check content. */
+	err = i915_gem_object_wait_migration(obj, true);
+	if (err)
+		goto out_unlock;
+
+	if (vma) {
+		err = i915_vma_wait_for_bind(vma);
+		if (err)
+			goto out_unlock;
+	} else {
+		err = igt_fill_check_buffer(obj, false);
+	}
+
+out_unlock:
+	i915_gem_object_unlock(obj);
+out_put:
+	i915_gem_object_put(obj);
+
+	return err;
+}
+
+static int igt_lmem_pages_failsafe_migrate(void *arg)
+{
+	int fail_gpu, fail_alloc, ban_memcpy, ret;
+	struct intel_gt *gt = arg;
+
+	for (fail_gpu = 0; fail_gpu < 2; ++fail_gpu) {
+		for (fail_alloc = 0; fail_alloc < 2; ++fail_alloc) {
+			for (ban_memcpy = 0; ban_memcpy < 2; ++ban_memcpy) {
+				pr_info("Simulated failure modes: gpu: %d, alloc:%d, ban_memcpy: %d\n",
+					fail_gpu, fail_alloc, ban_memcpy);
+				i915_ttm_migrate_set_ban_memcpy(ban_memcpy);
+				i915_ttm_migrate_set_failure_modes(fail_gpu,
+								   fail_alloc);
+				ret = __igt_lmem_pages_migrate(gt, NULL, NULL,
+							       NULL, NULL,
+							       ban_memcpy &&
+							       fail_gpu);
+
+				if (ban_memcpy && fail_gpu) {
+					struct intel_gt *__gt;
+					unsigned int id;
+
+					if (ret != -EIO) {
+						pr_err("expected -EIO, got (%d)\n", ret);
+						ret = -EINVAL;
+					} else {
+						ret = 0;
+					}
+
+					for_each_gt(__gt, gt->i915, id) {
+						intel_wakeref_t wakeref;
+						bool wedged;
+
+						mutex_lock(&__gt->reset.mutex);
+						wedged = test_bit(I915_WEDGED, &__gt->reset.flags);
+						mutex_unlock(&__gt->reset.mutex);
+
+						if (fail_gpu && !fail_alloc) {
+							if (!wedged) {
+								pr_err("gt(%u) not wedged\n", id);
+								ret = -EINVAL;
+								continue;
+							}
+						} else if (wedged) {
+							pr_err("gt(%u) incorrectly wedged\n", id);
+							ret = -EINVAL;
+						} else {
+							continue;
+						}
+
+						wakeref = intel_runtime_pm_get(__gt->uncore->rpm);
+						igt_global_reset_lock(__gt);
+						intel_gt_reset(__gt, ALL_ENGINES, NULL);
+						igt_global_reset_unlock(__gt);
+						intel_runtime_pm_put(__gt->uncore->rpm, wakeref);
+					}
+					if (ret)
+						goto out_err;
+				}
+			}
+		}
+	}
+
+out_err:
+	i915_ttm_migrate_set_failure_modes(false, false);
+	i915_ttm_migrate_set_ban_memcpy(false);
+	return ret;
+}
+
+/*
+ * This subtest tests that unbinding at migration is indeed performed
+ * async. We launch a spinner and a number of migrations depending on
+ * that spinner to have terminated. Before each migration we bind a
+ * vma, which should then be async unbound by the migration operation.
+ * If we are able to schedule migrations without blocking while the
+ * spinner is still running, those unbinds are indeed async and non-
+ * blocking.
+ *
+ * Note that each async bind operation is awaiting the previous migration
+ * due to the moving fence resulting from the migration.
+ */
+static int igt_async_migrate(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct i915_ppgtt *ppgtt;
+	struct igt_spinner spin;
+	int err;
+
+	ppgtt = i915_ppgtt_create(gt, 0);
+	if (IS_ERR(ppgtt))
+		return PTR_ERR(ppgtt);
+
+	if (igt_spinner_init(&spin, gt)) {
+		err = -ENOMEM;
+		goto out_spin;
+	}
+
+	for_each_engine(engine, gt, id) {
+		struct ttm_operation_ctx ctx = {
+			.interruptible = true
+		};
+		struct dma_fence *spin_fence;
+		struct intel_context *ce;
+		struct i915_request *rq;
+		struct i915_deps deps;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out_ce;
+		}
+
+		/*
+		 * Use MI_NOOP, making the spinner non-preemptible. If there
+		 * is a code path where we fail async operation due to the
+		 * running spinner, we will block and fail to end the
+		 * spinner resulting in a deadlock. But with a non-
+		 * preemptible spinner, hangcheck will terminate the spinner
+		 * for us, and we will later detect that and fail the test.
+		 */
+		rq = igt_spinner_create_request(&spin, ce, MI_NOOP);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_ce;
+		}
+
+		i915_deps_init(&deps, GFP_KERNEL);
+		err = i915_deps_add_dependency(&deps, &rq->fence, &ctx);
+		spin_fence = dma_fence_get(&rq->fence);
+		i915_request_add(rq);
+		if (err)
+			goto out_ce;
+
+		err = __igt_lmem_pages_migrate(gt, &ppgtt->vm, &deps, &spin,
+					       spin_fence, false);
+		i915_deps_fini(&deps);
+		dma_fence_put(spin_fence);
+		if (err)
+			goto out_ce;
+	}
+
+out_ce:
+	igt_spinner_fini(&spin);
+out_spin:
+	i915_vm_put(&ppgtt->vm);
+
+	return err;
+}
+
+/*
+ * Setting ASYNC_FAIL_ALLOC to 2 will simulate memory allocation failure while
+ * arming the migration error check and block async migration. This
+ * will cause us to deadlock and hangcheck will terminate the spinner
+ * causing the test to fail.
+ */
+#define ASYNC_FAIL_ALLOC 1
+static int igt_lmem_async_migrate(void *arg)
+{
+	int fail_gpu, fail_alloc, ban_memcpy, ret;
+	struct intel_gt *gt = arg;
+
+	for (fail_gpu = 0; fail_gpu < 2; ++fail_gpu) {
+		for (fail_alloc = 0; fail_alloc < ASYNC_FAIL_ALLOC; ++fail_alloc) {
+			for (ban_memcpy = 0; ban_memcpy < 2; ++ban_memcpy) {
+				pr_info("Simulated failure modes: gpu: %d, alloc: %d, ban_memcpy: %d\n",
+					fail_gpu, fail_alloc, ban_memcpy);
+				i915_ttm_migrate_set_ban_memcpy(ban_memcpy);
+				i915_ttm_migrate_set_failure_modes(fail_gpu,
+								   fail_alloc);
+				ret = igt_async_migrate(gt);
+
+				if (fail_gpu && ban_memcpy) {
+					struct intel_gt *__gt;
+					unsigned int id;
+
+					if (ret != -EIO) {
+						pr_err("expected -EIO, got (%d)\n", ret);
+						ret = -EINVAL;
+					} else {
+						ret = 0;
+					}
+
+					for_each_gt(__gt, gt->i915, id) {
+						intel_wakeref_t wakeref;
+						bool wedged;
+
+						mutex_lock(&__gt->reset.mutex);
+						wedged = test_bit(I915_WEDGED, &__gt->reset.flags);
+						mutex_unlock(&__gt->reset.mutex);
+
+						if (fail_gpu && !fail_alloc) {
+							if (!wedged) {
+								pr_err("gt(%u) not wedged\n", id);
+								ret = -EINVAL;
+								continue;
+							}
+						} else if (wedged) {
+							pr_err("gt(%u) incorrectly wedged\n", id);
+							ret = -EINVAL;
+						} else {
+							continue;
+						}
+
+						wakeref = intel_runtime_pm_get(__gt->uncore->rpm);
+						igt_global_reset_lock(__gt);
+						intel_gt_reset(__gt, ALL_ENGINES, NULL);
+						igt_global_reset_unlock(__gt);
+						intel_runtime_pm_put(__gt->uncore->rpm, wakeref);
+					}
+				}
+				if (ret)
+					goto out_err;
+			}
+		}
+	}
+
+out_err:
+	i915_ttm_migrate_set_failure_modes(false, false);
+	i915_ttm_migrate_set_ban_memcpy(false);
+	return ret;
+}
+
+int i915_gem_migrate_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_smem_create_migrate),
+		SUBTEST(igt_lmem_create_migrate),
+		SUBTEST(igt_same_create_migrate),
+		SUBTEST(igt_lmem_pages_failsafe_migrate),
+		SUBTEST(igt_lmem_async_migrate),
+	};
+
+	if (!HAS_LMEM(i915))
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c
new file mode 100644
index 000000000..b73c91aa5
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c
@@ -0,0 +1,1848 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include <linux/highmem.h>
+#include <linux/prime_numbers.h>
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_region.h"
+#include "gem/i915_gem_ttm.h"
+#include "gem/i915_gem_ttm_move.h"
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_pm.h"
+#include "gt/intel_migrate.h"
+#include "i915_ttm_buddy_manager.h"
+
+#include "huge_gem_object.h"
+#include "i915_selftest.h"
+#include "selftests/i915_random.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/igt_reset.h"
+#include "selftests/igt_mmap.h"
+
+struct tile {
+	unsigned int width;
+	unsigned int height;
+	unsigned int stride;
+	unsigned int size;
+	unsigned int tiling;
+	unsigned int swizzle;
+};
+
+static u64 swizzle_bit(unsigned int bit, u64 offset)
+{
+	return (offset & BIT_ULL(bit)) >> (bit - 6);
+}
+
+static u64 tiled_offset(const struct tile *tile, u64 v)
+{
+	u64 x, y;
+
+	if (tile->tiling == I915_TILING_NONE)
+		return v;
+
+	y = div64_u64_rem(v, tile->stride, &x);
+	v = div64_u64_rem(y, tile->height, &y) * tile->stride * tile->height;
+
+	if (tile->tiling == I915_TILING_X) {
+		v += y * tile->width;
+		v += div64_u64_rem(x, tile->width, &x) << tile->size;
+		v += x;
+	} else if (tile->width == 128) {
+		const unsigned int ytile_span = 16;
+		const unsigned int ytile_height = 512;
+
+		v += y * ytile_span;
+		v += div64_u64_rem(x, ytile_span, &x) * ytile_height;
+		v += x;
+	} else {
+		const unsigned int ytile_span = 32;
+		const unsigned int ytile_height = 256;
+
+		v += y * ytile_span;
+		v += div64_u64_rem(x, ytile_span, &x) * ytile_height;
+		v += x;
+	}
+
+	switch (tile->swizzle) {
+	case I915_BIT_6_SWIZZLE_9:
+		v ^= swizzle_bit(9, v);
+		break;
+	case I915_BIT_6_SWIZZLE_9_10:
+		v ^= swizzle_bit(9, v) ^ swizzle_bit(10, v);
+		break;
+	case I915_BIT_6_SWIZZLE_9_11:
+		v ^= swizzle_bit(9, v) ^ swizzle_bit(11, v);
+		break;
+	case I915_BIT_6_SWIZZLE_9_10_11:
+		v ^= swizzle_bit(9, v) ^ swizzle_bit(10, v) ^ swizzle_bit(11, v);
+		break;
+	}
+
+	return v;
+}
+
+static int check_partial_mapping(struct drm_i915_gem_object *obj,
+				 const struct tile *tile,
+				 struct rnd_state *prng)
+{
+	const unsigned long npages = obj->base.size / PAGE_SIZE;
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_gtt_view view;
+	struct i915_vma *vma;
+	unsigned long page;
+	u32 __iomem *io;
+	struct page *p;
+	unsigned int n;
+	u64 offset;
+	u32 *cpu;
+	int err;
+
+	err = i915_gem_object_set_tiling(obj, tile->tiling, tile->stride);
+	if (err) {
+		pr_err("Failed to set tiling mode=%u, stride=%u, err=%d\n",
+		       tile->tiling, tile->stride, err);
+		return err;
+	}
+
+	GEM_BUG_ON(i915_gem_object_get_tiling(obj) != tile->tiling);
+	GEM_BUG_ON(i915_gem_object_get_stride(obj) != tile->stride);
+
+	i915_gem_object_lock(obj, NULL);
+	err = i915_gem_object_set_to_gtt_domain(obj, true);
+	i915_gem_object_unlock(obj);
+	if (err) {
+		pr_err("Failed to flush to GTT write domain; err=%d\n", err);
+		return err;
+	}
+
+	page = i915_prandom_u32_max_state(npages, prng);
+	view = compute_partial_view(obj, page, MIN_CHUNK_PAGES);
+
+	vma = i915_gem_object_ggtt_pin(obj, &view, 0, 0, PIN_MAPPABLE);
+	if (IS_ERR(vma)) {
+		pr_err("Failed to pin partial view: offset=%lu; err=%d\n",
+		       page, (int)PTR_ERR(vma));
+		return PTR_ERR(vma);
+	}
+
+	n = page - view.partial.offset;
+	GEM_BUG_ON(n >= view.partial.size);
+
+	io = i915_vma_pin_iomap(vma);
+	i915_vma_unpin(vma);
+	if (IS_ERR(io)) {
+		pr_err("Failed to iomap partial view: offset=%lu; err=%d\n",
+		       page, (int)PTR_ERR(io));
+		err = PTR_ERR(io);
+		goto out;
+	}
+
+	iowrite32(page, io + n * PAGE_SIZE / sizeof(*io));
+	i915_vma_unpin_iomap(vma);
+
+	offset = tiled_offset(tile, page << PAGE_SHIFT);
+	if (offset >= obj->base.size)
+		goto out;
+
+	intel_gt_flush_ggtt_writes(to_gt(i915));
+
+	p = i915_gem_object_get_page(obj, offset >> PAGE_SHIFT);
+	cpu = kmap(p) + offset_in_page(offset);
+	drm_clflush_virt_range(cpu, sizeof(*cpu));
+	if (*cpu != (u32)page) {
+		pr_err("Partial view for %lu [%u] (offset=%llu, size=%u [%llu, row size %u], fence=%d, tiling=%d, stride=%d) misalignment, expected write to page (%llu + %u [0x%llx]) of 0x%x, found 0x%x\n",
+		       page, n,
+		       view.partial.offset,
+		       view.partial.size,
+		       vma->size >> PAGE_SHIFT,
+		       tile->tiling ? tile_row_pages(obj) : 0,
+		       vma->fence ? vma->fence->id : -1, tile->tiling, tile->stride,
+		       offset >> PAGE_SHIFT,
+		       (unsigned int)offset_in_page(offset),
+		       offset,
+		       (u32)page, *cpu);
+		err = -EINVAL;
+	}
+	*cpu = 0;
+	drm_clflush_virt_range(cpu, sizeof(*cpu));
+	kunmap(p);
+
+out:
+	i915_gem_object_lock(obj, NULL);
+	i915_vma_destroy(vma);
+	i915_gem_object_unlock(obj);
+	return err;
+}
+
+static int check_partial_mappings(struct drm_i915_gem_object *obj,
+				  const struct tile *tile,
+				  unsigned long end_time)
+{
+	const unsigned int nreal = obj->scratch / PAGE_SIZE;
+	const unsigned long npages = obj->base.size / PAGE_SIZE;
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_vma *vma;
+	unsigned long page;
+	int err;
+
+	err = i915_gem_object_set_tiling(obj, tile->tiling, tile->stride);
+	if (err) {
+		pr_err("Failed to set tiling mode=%u, stride=%u, err=%d\n",
+		       tile->tiling, tile->stride, err);
+		return err;
+	}
+
+	GEM_BUG_ON(i915_gem_object_get_tiling(obj) != tile->tiling);
+	GEM_BUG_ON(i915_gem_object_get_stride(obj) != tile->stride);
+
+	i915_gem_object_lock(obj, NULL);
+	err = i915_gem_object_set_to_gtt_domain(obj, true);
+	i915_gem_object_unlock(obj);
+	if (err) {
+		pr_err("Failed to flush to GTT write domain; err=%d\n", err);
+		return err;
+	}
+
+	for_each_prime_number_from(page, 1, npages) {
+		struct i915_gtt_view view =
+			compute_partial_view(obj, page, MIN_CHUNK_PAGES);
+		u32 __iomem *io;
+		struct page *p;
+		unsigned int n;
+		u64 offset;
+		u32 *cpu;
+
+		GEM_BUG_ON(view.partial.size > nreal);
+		cond_resched();
+
+		vma = i915_gem_object_ggtt_pin(obj, &view, 0, 0, PIN_MAPPABLE);
+		if (IS_ERR(vma)) {
+			pr_err("Failed to pin partial view: offset=%lu; err=%d\n",
+			       page, (int)PTR_ERR(vma));
+			return PTR_ERR(vma);
+		}
+
+		n = page - view.partial.offset;
+		GEM_BUG_ON(n >= view.partial.size);
+
+		io = i915_vma_pin_iomap(vma);
+		i915_vma_unpin(vma);
+		if (IS_ERR(io)) {
+			pr_err("Failed to iomap partial view: offset=%lu; err=%d\n",
+			       page, (int)PTR_ERR(io));
+			return PTR_ERR(io);
+		}
+
+		iowrite32(page, io + n * PAGE_SIZE / sizeof(*io));
+		i915_vma_unpin_iomap(vma);
+
+		offset = tiled_offset(tile, page << PAGE_SHIFT);
+		if (offset >= obj->base.size)
+			continue;
+
+		intel_gt_flush_ggtt_writes(to_gt(i915));
+
+		p = i915_gem_object_get_page(obj, offset >> PAGE_SHIFT);
+		cpu = kmap(p) + offset_in_page(offset);
+		drm_clflush_virt_range(cpu, sizeof(*cpu));
+		if (*cpu != (u32)page) {
+			pr_err("Partial view for %lu [%u] (offset=%llu, size=%u [%llu, row size %u], fence=%d, tiling=%d, stride=%d) misalignment, expected write to page (%llu + %u [0x%llx]) of 0x%x, found 0x%x\n",
+			       page, n,
+			       view.partial.offset,
+			       view.partial.size,
+			       vma->size >> PAGE_SHIFT,
+			       tile->tiling ? tile_row_pages(obj) : 0,
+			       vma->fence ? vma->fence->id : -1, tile->tiling, tile->stride,
+			       offset >> PAGE_SHIFT,
+			       (unsigned int)offset_in_page(offset),
+			       offset,
+			       (u32)page, *cpu);
+			err = -EINVAL;
+		}
+		*cpu = 0;
+		drm_clflush_virt_range(cpu, sizeof(*cpu));
+		kunmap(p);
+		if (err)
+			return err;
+
+		i915_gem_object_lock(obj, NULL);
+		i915_vma_destroy(vma);
+		i915_gem_object_unlock(obj);
+
+		if (igt_timeout(end_time,
+				"%s: timed out after tiling=%d stride=%d\n",
+				__func__, tile->tiling, tile->stride))
+			return -EINTR;
+	}
+
+	return 0;
+}
+
+static unsigned int
+setup_tile_size(struct tile *tile, struct drm_i915_private *i915)
+{
+	if (GRAPHICS_VER(i915) <= 2) {
+		tile->height = 16;
+		tile->width = 128;
+		tile->size = 11;
+	} else if (tile->tiling == I915_TILING_Y &&
+		   HAS_128_BYTE_Y_TILING(i915)) {
+		tile->height = 32;
+		tile->width = 128;
+		tile->size = 12;
+	} else {
+		tile->height = 8;
+		tile->width = 512;
+		tile->size = 12;
+	}
+
+	if (GRAPHICS_VER(i915) < 4)
+		return 8192 / tile->width;
+	else if (GRAPHICS_VER(i915) < 7)
+		return 128 * I965_FENCE_MAX_PITCH_VAL / tile->width;
+	else
+		return 128 * GEN7_FENCE_MAX_PITCH_VAL / tile->width;
+}
+
+static int igt_partial_tiling(void *arg)
+{
+	const unsigned int nreal = 1 << 12; /* largest tile row x2 */
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	intel_wakeref_t wakeref;
+	int tiling;
+	int err;
+
+	if (!i915_ggtt_has_aperture(to_gt(i915)->ggtt))
+		return 0;
+
+	/* We want to check the page mapping and fencing of a large object
+	 * mmapped through the GTT. The object we create is larger than can
+	 * possibly be mmaped as a whole, and so we must use partial GGTT vma.
+	 * We then check that a write through each partial GGTT vma ends up
+	 * in the right set of pages within the object, and with the expected
+	 * tiling, which we verify by manual swizzling.
+	 */
+
+	obj = huge_gem_object(i915,
+			      nreal << PAGE_SHIFT,
+			      (1 + next_prime_number(to_gt(i915)->ggtt->vm.total >> PAGE_SHIFT)) << PAGE_SHIFT);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	err = i915_gem_object_pin_pages_unlocked(obj);
+	if (err) {
+		pr_err("Failed to allocate %u pages (%lu total), err=%d\n",
+		       nreal, obj->base.size / PAGE_SIZE, err);
+		goto out;
+	}
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	if (1) {
+		IGT_TIMEOUT(end);
+		struct tile tile;
+
+		tile.height = 1;
+		tile.width = 1;
+		tile.size = 0;
+		tile.stride = 0;
+		tile.swizzle = I915_BIT_6_SWIZZLE_NONE;
+		tile.tiling = I915_TILING_NONE;
+
+		err = check_partial_mappings(obj, &tile, end);
+		if (err && err != -EINTR)
+			goto out_unlock;
+	}
+
+	for (tiling = I915_TILING_X; tiling <= I915_TILING_Y; tiling++) {
+		IGT_TIMEOUT(end);
+		unsigned int max_pitch;
+		unsigned int pitch;
+		struct tile tile;
+
+		if (i915->gem_quirks & GEM_QUIRK_PIN_SWIZZLED_PAGES)
+			/*
+			 * The swizzling pattern is actually unknown as it
+			 * varies based on physical address of each page.
+			 * See i915_gem_detect_bit_6_swizzle().
+			 */
+			break;
+
+		tile.tiling = tiling;
+		switch (tiling) {
+		case I915_TILING_X:
+			tile.swizzle = to_gt(i915)->ggtt->bit_6_swizzle_x;
+			break;
+		case I915_TILING_Y:
+			tile.swizzle = to_gt(i915)->ggtt->bit_6_swizzle_y;
+			break;
+		}
+
+		GEM_BUG_ON(tile.swizzle == I915_BIT_6_SWIZZLE_UNKNOWN);
+		if (tile.swizzle == I915_BIT_6_SWIZZLE_9_17 ||
+		    tile.swizzle == I915_BIT_6_SWIZZLE_9_10_17)
+			continue;
+
+		max_pitch = setup_tile_size(&tile, i915);
+
+		for (pitch = max_pitch; pitch; pitch >>= 1) {
+			tile.stride = tile.width * pitch;
+			err = check_partial_mappings(obj, &tile, end);
+			if (err == -EINTR)
+				goto next_tiling;
+			if (err)
+				goto out_unlock;
+
+			if (pitch > 2 && GRAPHICS_VER(i915) >= 4) {
+				tile.stride = tile.width * (pitch - 1);
+				err = check_partial_mappings(obj, &tile, end);
+				if (err == -EINTR)
+					goto next_tiling;
+				if (err)
+					goto out_unlock;
+			}
+
+			if (pitch < max_pitch && GRAPHICS_VER(i915) >= 4) {
+				tile.stride = tile.width * (pitch + 1);
+				err = check_partial_mappings(obj, &tile, end);
+				if (err == -EINTR)
+					goto next_tiling;
+				if (err)
+					goto out_unlock;
+			}
+		}
+
+		if (GRAPHICS_VER(i915) >= 4) {
+			for_each_prime_number(pitch, max_pitch) {
+				tile.stride = tile.width * pitch;
+				err = check_partial_mappings(obj, &tile, end);
+				if (err == -EINTR)
+					goto next_tiling;
+				if (err)
+					goto out_unlock;
+			}
+		}
+
+next_tiling: ;
+	}
+
+out_unlock:
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+	i915_gem_object_unpin_pages(obj);
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static int igt_smoke_tiling(void *arg)
+{
+	const unsigned int nreal = 1 << 12; /* largest tile row x2 */
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	intel_wakeref_t wakeref;
+	I915_RND_STATE(prng);
+	unsigned long count;
+	IGT_TIMEOUT(end);
+	int err;
+
+	if (!i915_ggtt_has_aperture(to_gt(i915)->ggtt))
+		return 0;
+
+	/*
+	 * igt_partial_tiling() does an exhastive check of partial tiling
+	 * chunking, but will undoubtably run out of time. Here, we do a
+	 * randomised search and hope over many runs of 1s with different
+	 * seeds we will do a thorough check.
+	 *
+	 * Remember to look at the st_seed if we see a flip-flop in BAT!
+	 */
+
+	if (i915->gem_quirks & GEM_QUIRK_PIN_SWIZZLED_PAGES)
+		return 0;
+
+	obj = huge_gem_object(i915,
+			      nreal << PAGE_SHIFT,
+			      (1 + next_prime_number(to_gt(i915)->ggtt->vm.total >> PAGE_SHIFT)) << PAGE_SHIFT);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	err = i915_gem_object_pin_pages_unlocked(obj);
+	if (err) {
+		pr_err("Failed to allocate %u pages (%lu total), err=%d\n",
+		       nreal, obj->base.size / PAGE_SIZE, err);
+		goto out;
+	}
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	count = 0;
+	do {
+		struct tile tile;
+
+		tile.tiling =
+			i915_prandom_u32_max_state(I915_TILING_Y + 1, &prng);
+		switch (tile.tiling) {
+		case I915_TILING_NONE:
+			tile.height = 1;
+			tile.width = 1;
+			tile.size = 0;
+			tile.stride = 0;
+			tile.swizzle = I915_BIT_6_SWIZZLE_NONE;
+			break;
+
+		case I915_TILING_X:
+			tile.swizzle = to_gt(i915)->ggtt->bit_6_swizzle_x;
+			break;
+		case I915_TILING_Y:
+			tile.swizzle = to_gt(i915)->ggtt->bit_6_swizzle_y;
+			break;
+		}
+
+		if (tile.swizzle == I915_BIT_6_SWIZZLE_9_17 ||
+		    tile.swizzle == I915_BIT_6_SWIZZLE_9_10_17)
+			continue;
+
+		if (tile.tiling != I915_TILING_NONE) {
+			unsigned int max_pitch = setup_tile_size(&tile, i915);
+
+			tile.stride =
+				i915_prandom_u32_max_state(max_pitch, &prng);
+			tile.stride = (1 + tile.stride) * tile.width;
+			if (GRAPHICS_VER(i915) < 4)
+				tile.stride = rounddown_pow_of_two(tile.stride);
+		}
+
+		err = check_partial_mapping(obj, &tile, &prng);
+		if (err)
+			break;
+
+		count++;
+	} while (!__igt_timeout(end, NULL));
+
+	pr_info("%s: Completed %lu trials\n", __func__, count);
+
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+	i915_gem_object_unpin_pages(obj);
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static int make_obj_busy(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct intel_engine_cs *engine;
+
+	for_each_uabi_engine(engine, i915) {
+		struct i915_request *rq;
+		struct i915_vma *vma;
+		struct i915_gem_ww_ctx ww;
+		int err;
+
+		vma = i915_vma_instance(obj, &engine->gt->ggtt->vm, NULL);
+		if (IS_ERR(vma))
+			return PTR_ERR(vma);
+
+		i915_gem_ww_ctx_init(&ww, false);
+retry:
+		err = i915_gem_object_lock(obj, &ww);
+		if (!err)
+			err = i915_vma_pin_ww(vma, &ww, 0, 0, PIN_USER);
+		if (err)
+			goto err;
+
+		rq = intel_engine_create_kernel_request(engine);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_unpin;
+		}
+
+		err = i915_request_await_object(rq, vma->obj, true);
+		if (err == 0)
+			err = i915_vma_move_to_active(vma, rq,
+						      EXEC_OBJECT_WRITE);
+
+		i915_request_add(rq);
+err_unpin:
+		i915_vma_unpin(vma);
+err:
+		if (err == -EDEADLK) {
+			err = i915_gem_ww_ctx_backoff(&ww);
+			if (!err)
+				goto retry;
+		}
+		i915_gem_ww_ctx_fini(&ww);
+		if (err)
+			return err;
+	}
+
+	i915_gem_object_put(obj); /* leave it only alive via its active ref */
+	return 0;
+}
+
+static enum i915_mmap_type default_mapping(struct drm_i915_private *i915)
+{
+	if (HAS_LMEM(i915))
+		return I915_MMAP_TYPE_FIXED;
+
+	return I915_MMAP_TYPE_GTT;
+}
+
+static struct drm_i915_gem_object *
+create_sys_or_internal(struct drm_i915_private *i915,
+		       unsigned long size)
+{
+	if (HAS_LMEM(i915)) {
+		struct intel_memory_region *sys_region =
+			i915->mm.regions[INTEL_REGION_SMEM];
+
+		return __i915_gem_object_create_user(i915, size, &sys_region, 1);
+	}
+
+	return i915_gem_object_create_internal(i915, size);
+}
+
+static bool assert_mmap_offset(struct drm_i915_private *i915,
+			       unsigned long size,
+			       int expected)
+{
+	struct drm_i915_gem_object *obj;
+	u64 offset;
+	int ret;
+
+	obj = create_sys_or_internal(i915, size);
+	if (IS_ERR(obj))
+		return expected && expected == PTR_ERR(obj);
+
+	ret = __assign_mmap_offset(obj, default_mapping(i915), &offset, NULL);
+	i915_gem_object_put(obj);
+
+	return ret == expected;
+}
+
+static void disable_retire_worker(struct drm_i915_private *i915)
+{
+	i915_gem_driver_unregister__shrinker(i915);
+	intel_gt_pm_get(to_gt(i915));
+	cancel_delayed_work_sync(&to_gt(i915)->requests.retire_work);
+}
+
+static void restore_retire_worker(struct drm_i915_private *i915)
+{
+	igt_flush_test(i915);
+	intel_gt_pm_put(to_gt(i915));
+	i915_gem_driver_register__shrinker(i915);
+}
+
+static void mmap_offset_lock(struct drm_i915_private *i915)
+	__acquires(&i915->drm.vma_offset_manager->vm_lock)
+{
+	write_lock(&i915->drm.vma_offset_manager->vm_lock);
+}
+
+static void mmap_offset_unlock(struct drm_i915_private *i915)
+	__releases(&i915->drm.vma_offset_manager->vm_lock)
+{
+	write_unlock(&i915->drm.vma_offset_manager->vm_lock);
+}
+
+static int igt_mmap_offset_exhaustion(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_mm *mm = &i915->drm.vma_offset_manager->vm_addr_space_mm;
+	struct drm_i915_gem_object *obj;
+	struct drm_mm_node *hole, *next;
+	int loop, err = 0;
+	u64 offset;
+	int enospc = HAS_LMEM(i915) ? -ENXIO : -ENOSPC;
+
+	/* Disable background reaper */
+	disable_retire_worker(i915);
+	GEM_BUG_ON(!to_gt(i915)->awake);
+	intel_gt_retire_requests(to_gt(i915));
+	i915_gem_drain_freed_objects(i915);
+
+	/* Trim the device mmap space to only a page */
+	mmap_offset_lock(i915);
+	loop = 1; /* PAGE_SIZE units */
+	list_for_each_entry_safe(hole, next, &mm->hole_stack, hole_stack) {
+		struct drm_mm_node *resv;
+
+		resv = kzalloc(sizeof(*resv), GFP_NOWAIT);
+		if (!resv) {
+			err = -ENOMEM;
+			goto out_park;
+		}
+
+		resv->start = drm_mm_hole_node_start(hole) + loop;
+		resv->size = hole->hole_size - loop;
+		resv->color = -1ul;
+		loop = 0;
+
+		if (!resv->size) {
+			kfree(resv);
+			continue;
+		}
+
+		pr_debug("Reserving hole [%llx + %llx]\n",
+			 resv->start, resv->size);
+
+		err = drm_mm_reserve_node(mm, resv);
+		if (err) {
+			pr_err("Failed to trim VMA manager, err=%d\n", err);
+			kfree(resv);
+			goto out_park;
+		}
+	}
+	GEM_BUG_ON(!list_is_singular(&mm->hole_stack));
+	mmap_offset_unlock(i915);
+
+	/* Just fits! */
+	if (!assert_mmap_offset(i915, PAGE_SIZE, 0)) {
+		pr_err("Unable to insert object into single page hole\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	/* Too large */
+	if (!assert_mmap_offset(i915, 2 * PAGE_SIZE, enospc)) {
+		pr_err("Unexpectedly succeeded in inserting too large object into single page hole\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	/* Fill the hole, further allocation attempts should then fail */
+	obj = create_sys_or_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		pr_err("Unable to create object for reclaimed hole\n");
+		goto out;
+	}
+
+	err = __assign_mmap_offset(obj, default_mapping(i915), &offset, NULL);
+	if (err) {
+		pr_err("Unable to insert object into reclaimed hole\n");
+		goto err_obj;
+	}
+
+	if (!assert_mmap_offset(i915, PAGE_SIZE, enospc)) {
+		pr_err("Unexpectedly succeeded in inserting object into no holes!\n");
+		err = -EINVAL;
+		goto err_obj;
+	}
+
+	i915_gem_object_put(obj);
+
+	/* Now fill with busy dead objects that we expect to reap */
+	for (loop = 0; loop < 3; loop++) {
+		if (intel_gt_is_wedged(to_gt(i915)))
+			break;
+
+		obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto out;
+		}
+
+		err = make_obj_busy(obj);
+		if (err) {
+			pr_err("[loop %d] Failed to busy the object\n", loop);
+			goto err_obj;
+		}
+	}
+
+out:
+	mmap_offset_lock(i915);
+out_park:
+	drm_mm_for_each_node_safe(hole, next, mm) {
+		if (hole->color != -1ul)
+			continue;
+
+		drm_mm_remove_node(hole);
+		kfree(hole);
+	}
+	mmap_offset_unlock(i915);
+	restore_retire_worker(i915);
+	return err;
+err_obj:
+	i915_gem_object_put(obj);
+	goto out;
+}
+
+static int gtt_set(struct drm_i915_gem_object *obj)
+{
+	struct i915_vma *vma;
+	void __iomem *map;
+	int err = 0;
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	intel_gt_pm_get(vma->vm->gt);
+	map = i915_vma_pin_iomap(vma);
+	i915_vma_unpin(vma);
+	if (IS_ERR(map)) {
+		err = PTR_ERR(map);
+		goto out;
+	}
+
+	memset_io(map, POISON_INUSE, obj->base.size);
+	i915_vma_unpin_iomap(vma);
+
+out:
+	intel_gt_pm_put(vma->vm->gt);
+	return err;
+}
+
+static int gtt_check(struct drm_i915_gem_object *obj)
+{
+	struct i915_vma *vma;
+	void __iomem *map;
+	int err = 0;
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	intel_gt_pm_get(vma->vm->gt);
+	map = i915_vma_pin_iomap(vma);
+	i915_vma_unpin(vma);
+	if (IS_ERR(map)) {
+		err = PTR_ERR(map);
+		goto out;
+	}
+
+	if (memchr_inv((void __force *)map, POISON_FREE, obj->base.size)) {
+		pr_err("%s: Write via mmap did not land in backing store (GTT)\n",
+		       obj->mm.region->name);
+		err = -EINVAL;
+	}
+	i915_vma_unpin_iomap(vma);
+
+out:
+	intel_gt_pm_put(vma->vm->gt);
+	return err;
+}
+
+static int wc_set(struct drm_i915_gem_object *obj)
+{
+	void *vaddr;
+
+	vaddr = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(vaddr))
+		return PTR_ERR(vaddr);
+
+	memset(vaddr, POISON_INUSE, obj->base.size);
+	i915_gem_object_flush_map(obj);
+	i915_gem_object_unpin_map(obj);
+
+	return 0;
+}
+
+static int wc_check(struct drm_i915_gem_object *obj)
+{
+	void *vaddr;
+	int err = 0;
+
+	vaddr = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(vaddr))
+		return PTR_ERR(vaddr);
+
+	if (memchr_inv(vaddr, POISON_FREE, obj->base.size)) {
+		pr_err("%s: Write via mmap did not land in backing store (WC)\n",
+		       obj->mm.region->name);
+		err = -EINVAL;
+	}
+	i915_gem_object_unpin_map(obj);
+
+	return err;
+}
+
+static bool can_mmap(struct drm_i915_gem_object *obj, enum i915_mmap_type type)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	bool no_map;
+
+	if (obj->ops->mmap_offset)
+		return type == I915_MMAP_TYPE_FIXED;
+	else if (type == I915_MMAP_TYPE_FIXED)
+		return false;
+
+	if (type == I915_MMAP_TYPE_GTT &&
+	    !i915_ggtt_has_aperture(to_gt(i915)->ggtt))
+		return false;
+
+	i915_gem_object_lock(obj, NULL);
+	no_map = (type != I915_MMAP_TYPE_GTT &&
+		  !i915_gem_object_has_struct_page(obj) &&
+		  !i915_gem_object_has_iomem(obj));
+	i915_gem_object_unlock(obj);
+
+	return !no_map;
+}
+
+#define expand32(x) (((x) << 0) | ((x) << 8) | ((x) << 16) | ((x) << 24))
+static int __igt_mmap(struct drm_i915_private *i915,
+		      struct drm_i915_gem_object *obj,
+		      enum i915_mmap_type type)
+{
+	struct vm_area_struct *area;
+	unsigned long addr;
+	int err, i;
+	u64 offset;
+
+	if (!can_mmap(obj, type))
+		return 0;
+
+	err = wc_set(obj);
+	if (err == -ENXIO)
+		err = gtt_set(obj);
+	if (err)
+		return err;
+
+	err = __assign_mmap_offset(obj, type, &offset, NULL);
+	if (err)
+		return err;
+
+	addr = igt_mmap_offset(i915, offset, obj->base.size, PROT_WRITE, MAP_SHARED);
+	if (IS_ERR_VALUE(addr))
+		return addr;
+
+	pr_debug("igt_mmap(%s, %d) @ %lx\n", obj->mm.region->name, type, addr);
+
+	mmap_read_lock(current->mm);
+	area = vma_lookup(current->mm, addr);
+	mmap_read_unlock(current->mm);
+	if (!area) {
+		pr_err("%s: Did not create a vm_area_struct for the mmap\n",
+		       obj->mm.region->name);
+		err = -EINVAL;
+		goto out_unmap;
+	}
+
+	for (i = 0; i < obj->base.size / sizeof(u32); i++) {
+		u32 __user *ux = u64_to_user_ptr((u64)(addr + i * sizeof(*ux)));
+		u32 x;
+
+		if (get_user(x, ux)) {
+			pr_err("%s: Unable to read from mmap, offset:%zd\n",
+			       obj->mm.region->name, i * sizeof(x));
+			err = -EFAULT;
+			goto out_unmap;
+		}
+
+		if (x != expand32(POISON_INUSE)) {
+			pr_err("%s: Read incorrect value from mmap, offset:%zd, found:%x, expected:%x\n",
+			       obj->mm.region->name,
+			       i * sizeof(x), x, expand32(POISON_INUSE));
+			err = -EINVAL;
+			goto out_unmap;
+		}
+
+		x = expand32(POISON_FREE);
+		if (put_user(x, ux)) {
+			pr_err("%s: Unable to write to mmap, offset:%zd\n",
+			       obj->mm.region->name, i * sizeof(x));
+			err = -EFAULT;
+			goto out_unmap;
+		}
+	}
+
+	if (type == I915_MMAP_TYPE_GTT)
+		intel_gt_flush_ggtt_writes(to_gt(i915));
+
+	err = wc_check(obj);
+	if (err == -ENXIO)
+		err = gtt_check(obj);
+out_unmap:
+	vm_munmap(addr, obj->base.size);
+	return err;
+}
+
+static int igt_mmap(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_memory_region *mr;
+	enum intel_region_id id;
+
+	for_each_memory_region(mr, i915, id) {
+		unsigned long sizes[] = {
+			PAGE_SIZE,
+			mr->min_page_size,
+			SZ_4M,
+		};
+		int i;
+
+		if (mr->private)
+			continue;
+
+		for (i = 0; i < ARRAY_SIZE(sizes); i++) {
+			struct drm_i915_gem_object *obj;
+			int err;
+
+			obj = __i915_gem_object_create_user(i915, sizes[i], &mr, 1);
+			if (obj == ERR_PTR(-ENODEV))
+				continue;
+
+			if (IS_ERR(obj))
+				return PTR_ERR(obj);
+
+			err = __igt_mmap(i915, obj, I915_MMAP_TYPE_GTT);
+			if (err == 0)
+				err = __igt_mmap(i915, obj, I915_MMAP_TYPE_WC);
+			if (err == 0)
+				err = __igt_mmap(i915, obj, I915_MMAP_TYPE_FIXED);
+
+			i915_gem_object_put(obj);
+			if (err)
+				return err;
+		}
+	}
+
+	return 0;
+}
+
+static void igt_close_objects(struct drm_i915_private *i915,
+			      struct list_head *objects)
+{
+	struct drm_i915_gem_object *obj, *on;
+
+	list_for_each_entry_safe(obj, on, objects, st_link) {
+		i915_gem_object_lock(obj, NULL);
+		if (i915_gem_object_has_pinned_pages(obj))
+			i915_gem_object_unpin_pages(obj);
+		/* No polluting the memory region between tests */
+		__i915_gem_object_put_pages(obj);
+		i915_gem_object_unlock(obj);
+		list_del(&obj->st_link);
+		i915_gem_object_put(obj);
+	}
+
+	cond_resched();
+
+	i915_gem_drain_freed_objects(i915);
+}
+
+static void igt_make_evictable(struct list_head *objects)
+{
+	struct drm_i915_gem_object *obj;
+
+	list_for_each_entry(obj, objects, st_link) {
+		i915_gem_object_lock(obj, NULL);
+		if (i915_gem_object_has_pinned_pages(obj))
+			i915_gem_object_unpin_pages(obj);
+		i915_gem_object_unlock(obj);
+	}
+
+	cond_resched();
+}
+
+static int igt_fill_mappable(struct intel_memory_region *mr,
+			     struct list_head *objects)
+{
+	u64 size, total;
+	int err;
+
+	total = 0;
+	size = mr->io_size;
+	do {
+		struct drm_i915_gem_object *obj;
+
+		obj = i915_gem_object_create_region(mr, size, 0, 0);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto err_close;
+		}
+
+		list_add(&obj->st_link, objects);
+
+		err = i915_gem_object_pin_pages_unlocked(obj);
+		if (err) {
+			if (err != -ENXIO && err != -ENOMEM)
+				goto err_close;
+
+			if (size == mr->min_page_size) {
+				err = 0;
+				break;
+			}
+
+			size >>= 1;
+			continue;
+		}
+
+		total += obj->base.size;
+	} while (1);
+
+	pr_info("%s filled=%lluMiB\n", __func__, total >> 20);
+	return 0;
+
+err_close:
+	igt_close_objects(mr->i915, objects);
+	return err;
+}
+
+static int ___igt_mmap_migrate(struct drm_i915_private *i915,
+			       struct drm_i915_gem_object *obj,
+			       unsigned long addr,
+			       bool unfaultable)
+{
+	struct vm_area_struct *area;
+	int err = 0, i;
+
+	pr_info("igt_mmap(%s, %d) @ %lx\n",
+		obj->mm.region->name, I915_MMAP_TYPE_FIXED, addr);
+
+	mmap_read_lock(current->mm);
+	area = vma_lookup(current->mm, addr);
+	mmap_read_unlock(current->mm);
+	if (!area) {
+		pr_err("%s: Did not create a vm_area_struct for the mmap\n",
+		       obj->mm.region->name);
+		err = -EINVAL;
+		goto out_unmap;
+	}
+
+	for (i = 0; i < obj->base.size / sizeof(u32); i++) {
+		u32 __user *ux = u64_to_user_ptr((u64)(addr + i * sizeof(*ux)));
+		u32 x;
+
+		if (get_user(x, ux)) {
+			err = -EFAULT;
+			if (!unfaultable) {
+				pr_err("%s: Unable to read from mmap, offset:%zd\n",
+				       obj->mm.region->name, i * sizeof(x));
+				goto out_unmap;
+			}
+
+			continue;
+		}
+
+		if (unfaultable) {
+			pr_err("%s: Faulted unmappable memory\n",
+			       obj->mm.region->name);
+			err = -EINVAL;
+			goto out_unmap;
+		}
+
+		if (x != expand32(POISON_INUSE)) {
+			pr_err("%s: Read incorrect value from mmap, offset:%zd, found:%x, expected:%x\n",
+			       obj->mm.region->name,
+			       i * sizeof(x), x, expand32(POISON_INUSE));
+			err = -EINVAL;
+			goto out_unmap;
+		}
+
+		x = expand32(POISON_FREE);
+		if (put_user(x, ux)) {
+			pr_err("%s: Unable to write to mmap, offset:%zd\n",
+			       obj->mm.region->name, i * sizeof(x));
+			err = -EFAULT;
+			goto out_unmap;
+		}
+	}
+
+	if (unfaultable) {
+		if (err == -EFAULT)
+			err = 0;
+	} else {
+		obj->flags &= ~I915_BO_ALLOC_GPU_ONLY;
+		err = wc_check(obj);
+	}
+out_unmap:
+	vm_munmap(addr, obj->base.size);
+	return err;
+}
+
+#define IGT_MMAP_MIGRATE_TOPDOWN     (1 << 0)
+#define IGT_MMAP_MIGRATE_FILL        (1 << 1)
+#define IGT_MMAP_MIGRATE_EVICTABLE   (1 << 2)
+#define IGT_MMAP_MIGRATE_UNFAULTABLE (1 << 3)
+#define IGT_MMAP_MIGRATE_FAIL_GPU    (1 << 4)
+static int __igt_mmap_migrate(struct intel_memory_region **placements,
+			      int n_placements,
+			      struct intel_memory_region *expected_mr,
+			      unsigned int flags)
+{
+	struct drm_i915_private *i915 = placements[0]->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_request *rq = NULL;
+	unsigned long addr;
+	LIST_HEAD(objects);
+	u64 offset;
+	int err;
+
+	obj = __i915_gem_object_create_user(i915, PAGE_SIZE,
+					    placements,
+					    n_placements);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	if (flags & IGT_MMAP_MIGRATE_TOPDOWN)
+		obj->flags |= I915_BO_ALLOC_GPU_ONLY;
+
+	err = __assign_mmap_offset(obj, I915_MMAP_TYPE_FIXED, &offset, NULL);
+	if (err)
+		goto out_put;
+
+	/*
+	 * This will eventually create a GEM context, due to opening dummy drm
+	 * file, which needs a tiny amount of mappable device memory for the top
+	 * level paging structures(and perhaps scratch), so make sure we
+	 * allocate early, to avoid tears.
+	 */
+	addr = igt_mmap_offset(i915, offset, obj->base.size,
+			       PROT_WRITE, MAP_SHARED);
+	if (IS_ERR_VALUE(addr)) {
+		err = addr;
+		goto out_put;
+	}
+
+	if (flags & IGT_MMAP_MIGRATE_FILL) {
+		err = igt_fill_mappable(placements[0], &objects);
+		if (err)
+			goto out_put;
+	}
+
+	err = i915_gem_object_lock(obj, NULL);
+	if (err)
+		goto out_put;
+
+	err = i915_gem_object_pin_pages(obj);
+	if (err) {
+		i915_gem_object_unlock(obj);
+		goto out_put;
+	}
+
+	err = intel_context_migrate_clear(to_gt(i915)->migrate.context, NULL,
+					  obj->mm.pages->sgl, obj->cache_level,
+					  i915_gem_object_is_lmem(obj),
+					  expand32(POISON_INUSE), &rq);
+	i915_gem_object_unpin_pages(obj);
+	if (rq) {
+		err = dma_resv_reserve_fences(obj->base.resv, 1);
+		if (!err)
+			dma_resv_add_fence(obj->base.resv, &rq->fence,
+					   DMA_RESV_USAGE_KERNEL);
+		i915_request_put(rq);
+	}
+	i915_gem_object_unlock(obj);
+	if (err)
+		goto out_put;
+
+	if (flags & IGT_MMAP_MIGRATE_EVICTABLE)
+		igt_make_evictable(&objects);
+
+	if (flags & IGT_MMAP_MIGRATE_FAIL_GPU) {
+		err = i915_gem_object_lock(obj, NULL);
+		if (err)
+			goto out_put;
+
+		/*
+		 * Ensure we only simulate the gpu failuire when faulting the
+		 * pages.
+		 */
+		err = i915_gem_object_wait_moving_fence(obj, true);
+		i915_gem_object_unlock(obj);
+		if (err)
+			goto out_put;
+		i915_ttm_migrate_set_failure_modes(true, false);
+	}
+
+	err = ___igt_mmap_migrate(i915, obj, addr,
+				  flags & IGT_MMAP_MIGRATE_UNFAULTABLE);
+
+	if (!err && obj->mm.region != expected_mr) {
+		pr_err("%s region mismatch %s\n", __func__, expected_mr->name);
+		err = -EINVAL;
+	}
+
+	if (flags & IGT_MMAP_MIGRATE_FAIL_GPU) {
+		struct intel_gt *gt;
+		unsigned int id;
+
+		i915_ttm_migrate_set_failure_modes(false, false);
+
+		for_each_gt(gt, i915, id) {
+			intel_wakeref_t wakeref;
+			bool wedged;
+
+			mutex_lock(&gt->reset.mutex);
+			wedged = test_bit(I915_WEDGED, &gt->reset.flags);
+			mutex_unlock(&gt->reset.mutex);
+			if (!wedged) {
+				pr_err("gt(%u) not wedged\n", id);
+				err = -EINVAL;
+				continue;
+			}
+
+			wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+			igt_global_reset_lock(gt);
+			intel_gt_reset(gt, ALL_ENGINES, NULL);
+			igt_global_reset_unlock(gt);
+			intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+		}
+
+		if (!i915_gem_object_has_unknown_state(obj)) {
+			pr_err("object missing unknown_state\n");
+			err = -EINVAL;
+		}
+	}
+
+out_put:
+	i915_gem_object_put(obj);
+	igt_close_objects(i915, &objects);
+	return err;
+}
+
+static int igt_mmap_migrate(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_memory_region *system = i915->mm.regions[INTEL_REGION_SMEM];
+	struct intel_memory_region *mr;
+	enum intel_region_id id;
+
+	for_each_memory_region(mr, i915, id) {
+		struct intel_memory_region *mixed[] = { mr, system };
+		struct intel_memory_region *single[] = { mr };
+		struct ttm_resource_manager *man = mr->region_private;
+		resource_size_t saved_io_size;
+		int err;
+
+		if (mr->private)
+			continue;
+
+		if (!mr->io_size)
+			continue;
+
+		/*
+		 * For testing purposes let's force small BAR, if not already
+		 * present.
+		 */
+		saved_io_size = mr->io_size;
+		if (mr->io_size == mr->total) {
+			resource_size_t io_size = mr->io_size;
+
+			io_size = rounddown_pow_of_two(io_size >> 1);
+			if (io_size < PAGE_SIZE)
+				continue;
+
+			mr->io_size = io_size;
+			i915_ttm_buddy_man_force_visible_size(man,
+							      io_size >> PAGE_SHIFT);
+		}
+
+		/*
+		 * Allocate in the mappable portion, should be no suprises here.
+		 */
+		err = __igt_mmap_migrate(mixed, ARRAY_SIZE(mixed), mr, 0);
+		if (err)
+			goto out_io_size;
+
+		/*
+		 * Allocate in the non-mappable portion, but force migrating to
+		 * the mappable portion on fault (LMEM -> LMEM)
+		 */
+		err = __igt_mmap_migrate(single, ARRAY_SIZE(single), mr,
+					 IGT_MMAP_MIGRATE_TOPDOWN |
+					 IGT_MMAP_MIGRATE_FILL |
+					 IGT_MMAP_MIGRATE_EVICTABLE);
+		if (err)
+			goto out_io_size;
+
+		/*
+		 * Allocate in the non-mappable portion, but force spilling into
+		 * system memory on fault (LMEM -> SMEM)
+		 */
+		err = __igt_mmap_migrate(mixed, ARRAY_SIZE(mixed), system,
+					 IGT_MMAP_MIGRATE_TOPDOWN |
+					 IGT_MMAP_MIGRATE_FILL);
+		if (err)
+			goto out_io_size;
+
+		/*
+		 * Allocate in the non-mappable portion, but since the mappable
+		 * portion is already full, and we can't spill to system memory,
+		 * then we should expect the fault to fail.
+		 */
+		err = __igt_mmap_migrate(single, ARRAY_SIZE(single), mr,
+					 IGT_MMAP_MIGRATE_TOPDOWN |
+					 IGT_MMAP_MIGRATE_FILL |
+					 IGT_MMAP_MIGRATE_UNFAULTABLE);
+		if (err)
+			goto out_io_size;
+
+		/*
+		 * Allocate in the non-mappable portion, but force migrating to
+		 * the mappable portion on fault (LMEM -> LMEM). We then also
+		 * simulate a gpu error when moving the pages when faulting the
+		 * pages, which should result in wedging the gpu and returning
+		 * SIGBUS in the fault handler, since we can't fallback to
+		 * memcpy.
+		 */
+		err = __igt_mmap_migrate(single, ARRAY_SIZE(single), mr,
+					 IGT_MMAP_MIGRATE_TOPDOWN |
+					 IGT_MMAP_MIGRATE_FILL |
+					 IGT_MMAP_MIGRATE_EVICTABLE |
+					 IGT_MMAP_MIGRATE_FAIL_GPU |
+					 IGT_MMAP_MIGRATE_UNFAULTABLE);
+out_io_size:
+		mr->io_size = saved_io_size;
+		i915_ttm_buddy_man_force_visible_size(man,
+						      mr->io_size >> PAGE_SHIFT);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static const char *repr_mmap_type(enum i915_mmap_type type)
+{
+	switch (type) {
+	case I915_MMAP_TYPE_GTT: return "gtt";
+	case I915_MMAP_TYPE_WB: return "wb";
+	case I915_MMAP_TYPE_WC: return "wc";
+	case I915_MMAP_TYPE_UC: return "uc";
+	case I915_MMAP_TYPE_FIXED: return "fixed";
+	default: return "unknown";
+	}
+}
+
+static bool can_access(struct drm_i915_gem_object *obj)
+{
+	bool access;
+
+	i915_gem_object_lock(obj, NULL);
+	access = i915_gem_object_has_struct_page(obj) ||
+		i915_gem_object_has_iomem(obj);
+	i915_gem_object_unlock(obj);
+
+	return access;
+}
+
+static int __igt_mmap_access(struct drm_i915_private *i915,
+			     struct drm_i915_gem_object *obj,
+			     enum i915_mmap_type type)
+{
+	unsigned long __user *ptr;
+	unsigned long A, B;
+	unsigned long x, y;
+	unsigned long addr;
+	int err;
+	u64 offset;
+
+	memset(&A, 0xAA, sizeof(A));
+	memset(&B, 0xBB, sizeof(B));
+
+	if (!can_mmap(obj, type) || !can_access(obj))
+		return 0;
+
+	err = __assign_mmap_offset(obj, type, &offset, NULL);
+	if (err)
+		return err;
+
+	addr = igt_mmap_offset(i915, offset, obj->base.size, PROT_WRITE, MAP_SHARED);
+	if (IS_ERR_VALUE(addr))
+		return addr;
+	ptr = (unsigned long __user *)addr;
+
+	err = __put_user(A, ptr);
+	if (err) {
+		pr_err("%s(%s): failed to write into user mmap\n",
+		       obj->mm.region->name, repr_mmap_type(type));
+		goto out_unmap;
+	}
+
+	intel_gt_flush_ggtt_writes(to_gt(i915));
+
+	err = access_process_vm(current, addr, &x, sizeof(x), 0);
+	if (err != sizeof(x)) {
+		pr_err("%s(%s): access_process_vm() read failed\n",
+		       obj->mm.region->name, repr_mmap_type(type));
+		goto out_unmap;
+	}
+
+	err = access_process_vm(current, addr, &B, sizeof(B), FOLL_WRITE);
+	if (err != sizeof(B)) {
+		pr_err("%s(%s): access_process_vm() write failed\n",
+		       obj->mm.region->name, repr_mmap_type(type));
+		goto out_unmap;
+	}
+
+	intel_gt_flush_ggtt_writes(to_gt(i915));
+
+	err = __get_user(y, ptr);
+	if (err) {
+		pr_err("%s(%s): failed to read from user mmap\n",
+		       obj->mm.region->name, repr_mmap_type(type));
+		goto out_unmap;
+	}
+
+	if (x != A || y != B) {
+		pr_err("%s(%s): failed to read/write values, found (%lx, %lx)\n",
+		       obj->mm.region->name, repr_mmap_type(type),
+		       x, y);
+		err = -EINVAL;
+		goto out_unmap;
+	}
+
+out_unmap:
+	vm_munmap(addr, obj->base.size);
+	return err;
+}
+
+static int igt_mmap_access(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_memory_region *mr;
+	enum intel_region_id id;
+
+	for_each_memory_region(mr, i915, id) {
+		struct drm_i915_gem_object *obj;
+		int err;
+
+		if (mr->private)
+			continue;
+
+		obj = __i915_gem_object_create_user(i915, PAGE_SIZE, &mr, 1);
+		if (obj == ERR_PTR(-ENODEV))
+			continue;
+
+		if (IS_ERR(obj))
+			return PTR_ERR(obj);
+
+		err = __igt_mmap_access(i915, obj, I915_MMAP_TYPE_GTT);
+		if (err == 0)
+			err = __igt_mmap_access(i915, obj, I915_MMAP_TYPE_WB);
+		if (err == 0)
+			err = __igt_mmap_access(i915, obj, I915_MMAP_TYPE_WC);
+		if (err == 0)
+			err = __igt_mmap_access(i915, obj, I915_MMAP_TYPE_UC);
+		if (err == 0)
+			err = __igt_mmap_access(i915, obj, I915_MMAP_TYPE_FIXED);
+
+		i915_gem_object_put(obj);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int __igt_mmap_gpu(struct drm_i915_private *i915,
+			  struct drm_i915_gem_object *obj,
+			  enum i915_mmap_type type)
+{
+	struct intel_engine_cs *engine;
+	unsigned long addr;
+	u32 __user *ux;
+	u32 bbe;
+	int err;
+	u64 offset;
+
+	/*
+	 * Verify that the mmap access into the backing store aligns with
+	 * that of the GPU, i.e. that mmap is indeed writing into the same
+	 * page as being read by the GPU.
+	 */
+
+	if (!can_mmap(obj, type))
+		return 0;
+
+	err = wc_set(obj);
+	if (err == -ENXIO)
+		err = gtt_set(obj);
+	if (err)
+		return err;
+
+	err = __assign_mmap_offset(obj, type, &offset, NULL);
+	if (err)
+		return err;
+
+	addr = igt_mmap_offset(i915, offset, obj->base.size, PROT_WRITE, MAP_SHARED);
+	if (IS_ERR_VALUE(addr))
+		return addr;
+
+	ux = u64_to_user_ptr((u64)addr);
+	bbe = MI_BATCH_BUFFER_END;
+	if (put_user(bbe, ux)) {
+		pr_err("%s: Unable to write to mmap\n", obj->mm.region->name);
+		err = -EFAULT;
+		goto out_unmap;
+	}
+
+	if (type == I915_MMAP_TYPE_GTT)
+		intel_gt_flush_ggtt_writes(to_gt(i915));
+
+	for_each_uabi_engine(engine, i915) {
+		struct i915_request *rq;
+		struct i915_vma *vma;
+		struct i915_gem_ww_ctx ww;
+
+		vma = i915_vma_instance(obj, engine->kernel_context->vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out_unmap;
+		}
+
+		i915_gem_ww_ctx_init(&ww, false);
+retry:
+		err = i915_gem_object_lock(obj, &ww);
+		if (!err)
+			err = i915_vma_pin_ww(vma, &ww, 0, 0, PIN_USER);
+		if (err)
+			goto out_ww;
+
+		rq = i915_request_create(engine->kernel_context);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_unpin;
+		}
+
+		err = i915_request_await_object(rq, vma->obj, false);
+		if (err == 0)
+			err = i915_vma_move_to_active(vma, rq, 0);
+
+		err = engine->emit_bb_start(rq, vma->node.start, 0, 0);
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+			struct drm_printer p =
+				drm_info_printer(engine->i915->drm.dev);
+
+			pr_err("%s(%s, %s): Failed to execute batch\n",
+			       __func__, engine->name, obj->mm.region->name);
+			intel_engine_dump(engine, &p,
+					  "%s\n", engine->name);
+
+			intel_gt_set_wedged(engine->gt);
+			err = -EIO;
+		}
+		i915_request_put(rq);
+
+out_unpin:
+		i915_vma_unpin(vma);
+out_ww:
+		if (err == -EDEADLK) {
+			err = i915_gem_ww_ctx_backoff(&ww);
+			if (!err)
+				goto retry;
+		}
+		i915_gem_ww_ctx_fini(&ww);
+		if (err)
+			goto out_unmap;
+	}
+
+out_unmap:
+	vm_munmap(addr, obj->base.size);
+	return err;
+}
+
+static int igt_mmap_gpu(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_memory_region *mr;
+	enum intel_region_id id;
+
+	for_each_memory_region(mr, i915, id) {
+		struct drm_i915_gem_object *obj;
+		int err;
+
+		if (mr->private)
+			continue;
+
+		obj = __i915_gem_object_create_user(i915, PAGE_SIZE, &mr, 1);
+		if (obj == ERR_PTR(-ENODEV))
+			continue;
+
+		if (IS_ERR(obj))
+			return PTR_ERR(obj);
+
+		err = __igt_mmap_gpu(i915, obj, I915_MMAP_TYPE_GTT);
+		if (err == 0)
+			err = __igt_mmap_gpu(i915, obj, I915_MMAP_TYPE_WC);
+		if (err == 0)
+			err = __igt_mmap_gpu(i915, obj, I915_MMAP_TYPE_FIXED);
+
+		i915_gem_object_put(obj);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int check_present_pte(pte_t *pte, unsigned long addr, void *data)
+{
+	if (!pte_present(*pte) || pte_none(*pte)) {
+		pr_err("missing PTE:%lx\n",
+		       (addr - (unsigned long)data) >> PAGE_SHIFT);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int check_absent_pte(pte_t *pte, unsigned long addr, void *data)
+{
+	if (pte_present(*pte) && !pte_none(*pte)) {
+		pr_err("present PTE:%lx; expected to be revoked\n",
+		       (addr - (unsigned long)data) >> PAGE_SHIFT);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int check_present(unsigned long addr, unsigned long len)
+{
+	return apply_to_page_range(current->mm, addr, len,
+				   check_present_pte, (void *)addr);
+}
+
+static int check_absent(unsigned long addr, unsigned long len)
+{
+	return apply_to_page_range(current->mm, addr, len,
+				   check_absent_pte, (void *)addr);
+}
+
+static int prefault_range(u64 start, u64 len)
+{
+	const char __user *addr, *end;
+	char __maybe_unused c;
+	int err;
+
+	addr = u64_to_user_ptr(start);
+	end = addr + len;
+
+	for (; addr < end; addr += PAGE_SIZE) {
+		err = __get_user(c, addr);
+		if (err)
+			return err;
+	}
+
+	return __get_user(c, end - 1);
+}
+
+static int __igt_mmap_revoke(struct drm_i915_private *i915,
+			     struct drm_i915_gem_object *obj,
+			     enum i915_mmap_type type)
+{
+	unsigned long addr;
+	int err;
+	u64 offset;
+
+	if (!can_mmap(obj, type))
+		return 0;
+
+	err = __assign_mmap_offset(obj, type, &offset, NULL);
+	if (err)
+		return err;
+
+	addr = igt_mmap_offset(i915, offset, obj->base.size, PROT_WRITE, MAP_SHARED);
+	if (IS_ERR_VALUE(addr))
+		return addr;
+
+	err = prefault_range(addr, obj->base.size);
+	if (err)
+		goto out_unmap;
+
+	err = check_present(addr, obj->base.size);
+	if (err) {
+		pr_err("%s: was not present\n", obj->mm.region->name);
+		goto out_unmap;
+	}
+
+	/*
+	 * After unbinding the object from the GGTT, its address may be reused
+	 * for other objects. Ergo we have to revoke the previous mmap PTE
+	 * access as it no longer points to the same object.
+	 */
+	i915_gem_object_lock(obj, NULL);
+	err = i915_gem_object_unbind(obj, I915_GEM_OBJECT_UNBIND_ACTIVE);
+	i915_gem_object_unlock(obj);
+	if (err) {
+		pr_err("Failed to unbind object!\n");
+		goto out_unmap;
+	}
+
+	if (type != I915_MMAP_TYPE_GTT) {
+		i915_gem_object_lock(obj, NULL);
+		__i915_gem_object_put_pages(obj);
+		i915_gem_object_unlock(obj);
+		if (i915_gem_object_has_pages(obj)) {
+			pr_err("Failed to put-pages object!\n");
+			err = -EINVAL;
+			goto out_unmap;
+		}
+	}
+
+	err = check_absent(addr, obj->base.size);
+	if (err) {
+		pr_err("%s: was not absent\n", obj->mm.region->name);
+		goto out_unmap;
+	}
+
+out_unmap:
+	vm_munmap(addr, obj->base.size);
+	return err;
+}
+
+static int igt_mmap_revoke(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_memory_region *mr;
+	enum intel_region_id id;
+
+	for_each_memory_region(mr, i915, id) {
+		struct drm_i915_gem_object *obj;
+		int err;
+
+		if (mr->private)
+			continue;
+
+		obj = __i915_gem_object_create_user(i915, PAGE_SIZE, &mr, 1);
+		if (obj == ERR_PTR(-ENODEV))
+			continue;
+
+		if (IS_ERR(obj))
+			return PTR_ERR(obj);
+
+		err = __igt_mmap_revoke(i915, obj, I915_MMAP_TYPE_GTT);
+		if (err == 0)
+			err = __igt_mmap_revoke(i915, obj, I915_MMAP_TYPE_WC);
+		if (err == 0)
+			err = __igt_mmap_revoke(i915, obj, I915_MMAP_TYPE_FIXED);
+
+		i915_gem_object_put(obj);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int i915_gem_mman_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_partial_tiling),
+		SUBTEST(igt_smoke_tiling),
+		SUBTEST(igt_mmap_offset_exhaustion),
+		SUBTEST(igt_mmap),
+		SUBTEST(igt_mmap_migrate),
+		SUBTEST(igt_mmap_access),
+		SUBTEST(igt_mmap_revoke),
+		SUBTEST(igt_mmap_gpu),
+	};
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_object.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_object.c
new file mode 100644
index 000000000..bdf5bb40c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_object.c
@@ -0,0 +1,99 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include "i915_selftest.h"
+
+#include "huge_gem_object.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/mock_gem_device.h"
+
+static int igt_gem_object(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	/* Basic test to ensure we can create an object */
+
+	obj = i915_gem_object_create_shmem(i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		pr_err("i915_gem_object_create failed, err=%d\n", err);
+		goto out;
+	}
+
+	err = 0;
+	i915_gem_object_put(obj);
+out:
+	return err;
+}
+
+static int igt_gem_huge(void *arg)
+{
+	const unsigned int nreal = 509; /* just to be awkward */
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	unsigned int n;
+	int err;
+
+	/* Basic sanitycheck of our huge fake object allocation */
+
+	obj = huge_gem_object(i915,
+			      nreal * PAGE_SIZE,
+			      to_gt(i915)->ggtt->vm.total + PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	err = i915_gem_object_pin_pages_unlocked(obj);
+	if (err) {
+		pr_err("Failed to allocate %u pages (%lu total), err=%d\n",
+		       nreal, obj->base.size / PAGE_SIZE, err);
+		goto out;
+	}
+
+	for (n = 0; n < obj->base.size / PAGE_SIZE; n++) {
+		if (i915_gem_object_get_page(obj, n) !=
+		    i915_gem_object_get_page(obj, n % nreal)) {
+			pr_err("Page lookup mismatch at index %u [%u]\n",
+			       n, n % nreal);
+			err = -EINVAL;
+			goto out_unpin;
+		}
+	}
+
+out_unpin:
+	i915_gem_object_unpin_pages(obj);
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+int i915_gem_object_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_gem_object),
+	};
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	err = i915_subtests(tests, i915);
+
+	mock_destroy_device(i915);
+	return err;
+}
+
+int i915_gem_object_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_gem_huge),
+	};
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_phys.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_phys.c
new file mode 100644
index 000000000..d43d8dae0
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_phys.c
@@ -0,0 +1,87 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include "i915_selftest.h"
+
+#include "selftests/mock_gem_device.h"
+
+static int mock_phys_object(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	/* Create an object and bind it to a contiguous set of physical pages,
+	 * i.e. exercise the i915_gem_object_phys API.
+	 */
+
+	obj = i915_gem_object_create_shmem(i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		pr_err("i915_gem_object_create failed, err=%d\n", err);
+		goto out;
+	}
+
+	i915_gem_object_lock(obj, NULL);
+	if (!i915_gem_object_has_struct_page(obj)) {
+		i915_gem_object_unlock(obj);
+		err = -EINVAL;
+		pr_err("shmem has no struct page\n");
+		goto out_obj;
+	}
+
+	err = i915_gem_object_attach_phys(obj, PAGE_SIZE);
+	i915_gem_object_unlock(obj);
+	if (err) {
+		pr_err("i915_gem_object_attach_phys failed, err=%d\n", err);
+		goto out_obj;
+	}
+
+	if (i915_gem_object_has_struct_page(obj)) {
+		pr_err("i915_gem_object_attach_phys did not create a phys object\n");
+		err = -EINVAL;
+		goto out_obj;
+	}
+
+	if (!atomic_read(&obj->mm.pages_pin_count)) {
+		pr_err("i915_gem_object_attach_phys did not pin its phys pages\n");
+		err = -EINVAL;
+		goto out_obj;
+	}
+
+	/* Make the object dirty so that put_pages must do copy back the data */
+	i915_gem_object_lock(obj, NULL);
+	err = i915_gem_object_set_to_gtt_domain(obj, true);
+	i915_gem_object_unlock(obj);
+	if (err) {
+		pr_err("i915_gem_object_set_to_gtt_domain failed with err=%d\n",
+		       err);
+		goto out_obj;
+	}
+
+out_obj:
+	i915_gem_object_put(obj);
+out:
+	return err;
+}
+
+int i915_gem_phys_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(mock_phys_object),
+	};
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	err = i915_subtests(tests, i915);
+
+	mock_destroy_device(i915);
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/igt_gem_utils.c b/drivers/gpu/drm/i915/gem/selftests/igt_gem_utils.c
new file mode 100644
index 000000000..3c55e77b0
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/igt_gem_utils.c
@@ -0,0 +1,164 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include "igt_gem_utils.h"
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_context.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt.h"
+#include "i915_vma.h"
+#include "i915_drv.h"
+
+#include "i915_request.h"
+
+struct i915_request *
+igt_request_alloc(struct i915_gem_context *ctx, struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+
+	/*
+	 * Pinning the contexts may generate requests in order to acquire
+	 * GGTT space, so do this first before we reserve a seqno for
+	 * ourselves.
+	 */
+	ce = i915_gem_context_get_engine(ctx, engine->legacy_idx);
+	if (IS_ERR(ce))
+		return ERR_CAST(ce);
+
+	rq = intel_context_create_request(ce);
+	intel_context_put(ce);
+
+	return rq;
+}
+
+struct i915_vma *
+igt_emit_store_dw(struct i915_vma *vma,
+		  u64 offset,
+		  unsigned long count,
+		  u32 val)
+{
+	struct drm_i915_gem_object *obj;
+	const int ver = GRAPHICS_VER(vma->vm->i915);
+	unsigned long n, size;
+	u32 *cmd;
+	int err;
+
+	size = (4 * count + 1) * sizeof(u32);
+	size = round_up(size, PAGE_SIZE);
+	obj = i915_gem_object_create_internal(vma->vm->i915, size);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	cmd = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(cmd)) {
+		err = PTR_ERR(cmd);
+		goto err;
+	}
+
+	GEM_BUG_ON(offset + (count - 1) * PAGE_SIZE > vma->node.size);
+	offset += vma->node.start;
+
+	for (n = 0; n < count; n++) {
+		if (ver >= 8) {
+			*cmd++ = MI_STORE_DWORD_IMM_GEN4;
+			*cmd++ = lower_32_bits(offset);
+			*cmd++ = upper_32_bits(offset);
+			*cmd++ = val;
+		} else if (ver >= 4) {
+			*cmd++ = MI_STORE_DWORD_IMM_GEN4 |
+				(ver < 6 ? MI_USE_GGTT : 0);
+			*cmd++ = 0;
+			*cmd++ = offset;
+			*cmd++ = val;
+		} else {
+			*cmd++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL;
+			*cmd++ = offset;
+			*cmd++ = val;
+		}
+		offset += PAGE_SIZE;
+	}
+	*cmd = MI_BATCH_BUFFER_END;
+
+	i915_gem_object_flush_map(obj);
+	i915_gem_object_unpin_map(obj);
+
+	intel_gt_chipset_flush(vma->vm->gt);
+
+	vma = i915_vma_instance(obj, vma->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto err;
+
+	return vma;
+
+err:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+int igt_gpu_fill_dw(struct intel_context *ce,
+		    struct i915_vma *vma, u64 offset,
+		    unsigned long count, u32 val)
+{
+	struct i915_request *rq;
+	struct i915_vma *batch;
+	unsigned int flags;
+	int err;
+
+	GEM_BUG_ON(!intel_engine_can_store_dword(ce->engine));
+	GEM_BUG_ON(!i915_vma_is_pinned(vma));
+
+	batch = igt_emit_store_dw(vma, offset, count, val);
+	if (IS_ERR(batch))
+		return PTR_ERR(batch);
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_batch;
+	}
+
+	i915_vma_lock(batch);
+	err = i915_request_await_object(rq, batch->obj, false);
+	if (err == 0)
+		err = i915_vma_move_to_active(batch, rq, 0);
+	i915_vma_unlock(batch);
+	if (err)
+		goto skip_request;
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, true);
+	if (err == 0)
+		err = i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE);
+	i915_vma_unlock(vma);
+	if (err)
+		goto skip_request;
+
+	flags = 0;
+	if (GRAPHICS_VER(ce->vm->i915) <= 5)
+		flags |= I915_DISPATCH_SECURE;
+
+	err = rq->engine->emit_bb_start(rq,
+					batch->node.start, batch->node.size,
+					flags);
+
+skip_request:
+	if (err)
+		i915_request_set_error_once(rq, err);
+	i915_request_add(rq);
+err_batch:
+	i915_vma_unpin_and_release(&batch, 0);
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/igt_gem_utils.h b/drivers/gpu/drm/i915/gem/selftests/igt_gem_utils.h
new file mode 100644
index 000000000..4221cf84d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/igt_gem_utils.h
@@ -0,0 +1,32 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef __IGT_GEM_UTILS_H__
+#define __IGT_GEM_UTILS_H__
+
+#include <linux/types.h>
+
+struct i915_request;
+struct i915_gem_context;
+struct i915_vma;
+
+struct intel_context;
+struct intel_engine_cs;
+
+struct i915_request *
+igt_request_alloc(struct i915_gem_context *ctx, struct intel_engine_cs *engine);
+
+struct i915_vma *
+igt_emit_store_dw(struct i915_vma *vma,
+		  u64 offset,
+		  unsigned long count,
+		  u32 val);
+
+int igt_gpu_fill_dw(struct intel_context *ce,
+		    struct i915_vma *vma, u64 offset,
+		    unsigned long count, u32 val);
+
+#endif /* __IGT_GEM_UTILS_H__ */
diff --git a/drivers/gpu/drm/i915/gem/selftests/mock_context.c b/drivers/gpu/drm/i915/gem/selftests/mock_context.c
new file mode 100644
index 000000000..8ac6726ec
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/mock_context.c
@@ -0,0 +1,180 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include "i915_file_private.h"
+#include "mock_context.h"
+#include "selftests/mock_drm.h"
+#include "selftests/mock_gtt.h"
+
+struct i915_gem_context *
+mock_context(struct drm_i915_private *i915,
+	     const char *name)
+{
+	struct i915_gem_context *ctx;
+	struct i915_gem_engines *e;
+	struct intel_sseu null_sseu = {};
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return NULL;
+
+	kref_init(&ctx->ref);
+	INIT_LIST_HEAD(&ctx->link);
+	ctx->i915 = i915;
+	INIT_WORK(&ctx->release_work, i915_gem_context_release_work);
+
+	mutex_init(&ctx->mutex);
+
+	spin_lock_init(&ctx->stale.lock);
+	INIT_LIST_HEAD(&ctx->stale.engines);
+
+	i915_gem_context_set_persistence(ctx);
+
+	if (name) {
+		struct i915_ppgtt *ppgtt;
+
+		strncpy(ctx->name, name, sizeof(ctx->name) - 1);
+
+		ppgtt = mock_ppgtt(i915, name);
+		if (!ppgtt)
+			goto err_free;
+
+		ctx->vm = &ppgtt->vm;
+	}
+
+	mutex_init(&ctx->engines_mutex);
+	e = default_engines(ctx, null_sseu);
+	if (IS_ERR(e))
+		goto err_vm;
+	RCU_INIT_POINTER(ctx->engines, e);
+
+	INIT_RADIX_TREE(&ctx->handles_vma, GFP_KERNEL);
+	mutex_init(&ctx->lut_mutex);
+
+	return ctx;
+
+err_vm:
+	if (ctx->vm)
+		i915_vm_put(ctx->vm);
+err_free:
+	kfree(ctx);
+	return NULL;
+}
+
+void mock_context_close(struct i915_gem_context *ctx)
+{
+	context_close(ctx);
+}
+
+void mock_init_contexts(struct drm_i915_private *i915)
+{
+	init_contexts(&i915->gem.contexts);
+}
+
+struct i915_gem_context *
+live_context(struct drm_i915_private *i915, struct file *file)
+{
+	struct drm_i915_file_private *fpriv = to_drm_file(file)->driver_priv;
+	struct i915_gem_proto_context *pc;
+	struct i915_gem_context *ctx;
+	int err;
+	u32 id;
+
+	pc = proto_context_create(i915, 0);
+	if (IS_ERR(pc))
+		return ERR_CAST(pc);
+
+	ctx = i915_gem_create_context(i915, pc);
+	proto_context_close(i915, pc);
+	if (IS_ERR(ctx))
+		return ctx;
+
+	i915_gem_context_set_no_error_capture(ctx);
+
+	err = xa_alloc(&fpriv->context_xa, &id, NULL, xa_limit_32b, GFP_KERNEL);
+	if (err < 0)
+		goto err_ctx;
+
+	gem_context_register(ctx, fpriv, id);
+
+	return ctx;
+
+err_ctx:
+	context_close(ctx);
+	return ERR_PTR(err);
+}
+
+struct i915_gem_context *
+live_context_for_engine(struct intel_engine_cs *engine, struct file *file)
+{
+	struct i915_gem_engines *engines;
+	struct i915_gem_context *ctx;
+	struct intel_sseu null_sseu = {};
+	struct intel_context *ce;
+
+	engines = alloc_engines(1);
+	if (!engines)
+		return ERR_PTR(-ENOMEM);
+
+	ctx = live_context(engine->i915, file);
+	if (IS_ERR(ctx)) {
+		__free_engines(engines, 0);
+		return ctx;
+	}
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce)) {
+		__free_engines(engines, 0);
+		return ERR_CAST(ce);
+	}
+
+	intel_context_set_gem(ce, ctx, null_sseu);
+	engines->engines[0] = ce;
+	engines->num_engines = 1;
+
+	mutex_lock(&ctx->engines_mutex);
+	i915_gem_context_set_user_engines(ctx);
+	engines = rcu_replace_pointer(ctx->engines, engines, 1);
+	mutex_unlock(&ctx->engines_mutex);
+
+	engines_idle_release(ctx, engines);
+
+	return ctx;
+}
+
+struct i915_gem_context *
+kernel_context(struct drm_i915_private *i915,
+	       struct i915_address_space *vm)
+{
+	struct i915_gem_context *ctx;
+	struct i915_gem_proto_context *pc;
+
+	pc = proto_context_create(i915, 0);
+	if (IS_ERR(pc))
+		return ERR_CAST(pc);
+
+	if (vm) {
+		if (pc->vm)
+			i915_vm_put(pc->vm);
+		pc->vm = i915_vm_get(vm);
+	}
+
+	ctx = i915_gem_create_context(i915, pc);
+	proto_context_close(i915, pc);
+	if (IS_ERR(ctx))
+		return ctx;
+
+	i915_gem_context_clear_bannable(ctx);
+	i915_gem_context_set_persistence(ctx);
+	i915_gem_context_set_no_error_capture(ctx);
+
+	return ctx;
+}
+
+void kernel_context_close(struct i915_gem_context *ctx)
+{
+	context_close(ctx);
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/mock_context.h b/drivers/gpu/drm/i915/gem/selftests/mock_context.h
new file mode 100644
index 000000000..7a02fd9b5
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/mock_context.h
@@ -0,0 +1,33 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __MOCK_CONTEXT_H
+#define __MOCK_CONTEXT_H
+
+struct file;
+struct drm_i915_private;
+struct intel_engine_cs;
+struct i915_address_space;
+
+void mock_init_contexts(struct drm_i915_private *i915);
+
+struct i915_gem_context *
+mock_context(struct drm_i915_private *i915,
+	     const char *name);
+
+void mock_context_close(struct i915_gem_context *ctx);
+
+struct i915_gem_context *
+live_context(struct drm_i915_private *i915, struct file *file);
+
+struct i915_gem_context *
+live_context_for_engine(struct intel_engine_cs *engine, struct file *file);
+
+struct i915_gem_context *kernel_context(struct drm_i915_private *i915,
+					struct i915_address_space *vm);
+void kernel_context_close(struct i915_gem_context *ctx);
+
+#endif /* !__MOCK_CONTEXT_H */
diff --git a/drivers/gpu/drm/i915/gem/selftests/mock_dmabuf.c b/drivers/gpu/drm/i915/gem/selftests/mock_dmabuf.c
new file mode 100644
index 000000000..b2a5882b8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/mock_dmabuf.c
@@ -0,0 +1,133 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include "mock_dmabuf.h"
+
+static struct sg_table *mock_map_dma_buf(struct dma_buf_attachment *attachment,
+					 enum dma_data_direction dir)
+{
+	struct mock_dmabuf *mock = to_mock(attachment->dmabuf);
+	struct sg_table *st;
+	struct scatterlist *sg;
+	int i, err;
+
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (!st)
+		return ERR_PTR(-ENOMEM);
+
+	err = sg_alloc_table(st, mock->npages, GFP_KERNEL);
+	if (err)
+		goto err_free;
+
+	sg = st->sgl;
+	for (i = 0; i < mock->npages; i++) {
+		sg_set_page(sg, mock->pages[i], PAGE_SIZE, 0);
+		sg = sg_next(sg);
+	}
+
+	err = dma_map_sgtable(attachment->dev, st, dir, 0);
+	if (err)
+		goto err_st;
+
+	return st;
+
+err_st:
+	sg_free_table(st);
+err_free:
+	kfree(st);
+	return ERR_PTR(err);
+}
+
+static void mock_unmap_dma_buf(struct dma_buf_attachment *attachment,
+			       struct sg_table *st,
+			       enum dma_data_direction dir)
+{
+	dma_unmap_sgtable(attachment->dev, st, dir, 0);
+	sg_free_table(st);
+	kfree(st);
+}
+
+static void mock_dmabuf_release(struct dma_buf *dma_buf)
+{
+	struct mock_dmabuf *mock = to_mock(dma_buf);
+	int i;
+
+	for (i = 0; i < mock->npages; i++)
+		put_page(mock->pages[i]);
+
+	kfree(mock);
+}
+
+static int mock_dmabuf_vmap(struct dma_buf *dma_buf, struct iosys_map *map)
+{
+	struct mock_dmabuf *mock = to_mock(dma_buf);
+	void *vaddr;
+
+	vaddr = vm_map_ram(mock->pages, mock->npages, 0);
+	if (!vaddr)
+		return -ENOMEM;
+	iosys_map_set_vaddr(map, vaddr);
+
+	return 0;
+}
+
+static void mock_dmabuf_vunmap(struct dma_buf *dma_buf, struct iosys_map *map)
+{
+	struct mock_dmabuf *mock = to_mock(dma_buf);
+
+	vm_unmap_ram(map->vaddr, mock->npages);
+}
+
+static int mock_dmabuf_mmap(struct dma_buf *dma_buf, struct vm_area_struct *vma)
+{
+	return -ENODEV;
+}
+
+static const struct dma_buf_ops mock_dmabuf_ops =  {
+	.map_dma_buf = mock_map_dma_buf,
+	.unmap_dma_buf = mock_unmap_dma_buf,
+	.release = mock_dmabuf_release,
+	.mmap = mock_dmabuf_mmap,
+	.vmap = mock_dmabuf_vmap,
+	.vunmap = mock_dmabuf_vunmap,
+};
+
+static struct dma_buf *mock_dmabuf(int npages)
+{
+	struct mock_dmabuf *mock;
+	DEFINE_DMA_BUF_EXPORT_INFO(exp_info);
+	struct dma_buf *dmabuf;
+	int i;
+
+	mock = kmalloc(sizeof(*mock) + npages * sizeof(struct page *),
+		       GFP_KERNEL);
+	if (!mock)
+		return ERR_PTR(-ENOMEM);
+
+	mock->npages = npages;
+	for (i = 0; i < npages; i++) {
+		mock->pages[i] = alloc_page(GFP_KERNEL);
+		if (!mock->pages[i])
+			goto err;
+	}
+
+	exp_info.ops = &mock_dmabuf_ops;
+	exp_info.size = npages * PAGE_SIZE;
+	exp_info.flags = O_CLOEXEC;
+	exp_info.priv = mock;
+
+	dmabuf = dma_buf_export(&exp_info);
+	if (IS_ERR(dmabuf))
+		goto err;
+
+	return dmabuf;
+
+err:
+	while (i--)
+		put_page(mock->pages[i]);
+	kfree(mock);
+	return ERR_PTR(-ENOMEM);
+}
diff --git a/drivers/gpu/drm/i915/gem/selftests/mock_dmabuf.h b/drivers/gpu/drm/i915/gem/selftests/mock_dmabuf.h
new file mode 100644
index 000000000..22818bbb1
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/mock_dmabuf.h
@@ -0,0 +1,22 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __MOCK_DMABUF_H__
+#define __MOCK_DMABUF_H__
+
+#include <linux/dma-buf.h>
+
+struct mock_dmabuf {
+	int npages;
+	struct page *pages[];
+};
+
+static inline struct mock_dmabuf *to_mock(struct dma_buf *buf)
+{
+	return buf->priv;
+}
+
+#endif /* !__MOCK_DMABUF_H__ */
diff --git a/drivers/gpu/drm/i915/gem/selftests/mock_gem_object.h b/drivers/gpu/drm/i915/gem/selftests/mock_gem_object.h
new file mode 100644
index 000000000..688511afa
--- /dev/null
+++ b/drivers/gpu/drm/i915/gem/selftests/mock_gem_object.h
@@ -0,0 +1,16 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __MOCK_GEM_OBJECT_H__
+#define __MOCK_GEM_OBJECT_H__
+
+#include "gem/i915_gem_object_types.h"
+
+struct mock_object {
+	struct drm_i915_gem_object base;
+};
+
+#endif /* !__MOCK_GEM_OBJECT_H__ */
diff --git a/drivers/gpu/drm/i915/gt/gen2_engine_cs.c b/drivers/gpu/drm/i915/gt/gen2_engine_cs.c
new file mode 100644
index 000000000..1c82caf52
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen2_engine_cs.c
@@ -0,0 +1,331 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "gen2_engine_cs.h"
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_engine.h"
+#include "intel_engine_regs.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt.h"
+#include "intel_gt_irq.h"
+#include "intel_ring.h"
+
+int gen2_emit_flush(struct i915_request *rq, u32 mode)
+{
+	unsigned int num_store_dw = 12;
+	u32 cmd, *cs;
+
+	cmd = MI_FLUSH;
+	if (mode & EMIT_INVALIDATE)
+		cmd |= MI_READ_FLUSH;
+
+	cs = intel_ring_begin(rq, 2 + 4 * num_store_dw);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = cmd;
+	while (num_store_dw--) {
+		*cs++ = MI_STORE_DWORD_INDEX;
+		*cs++ = I915_GEM_HWS_SCRATCH * sizeof(u32);
+		*cs++ = 0;
+		*cs++ = MI_FLUSH | MI_NO_WRITE_FLUSH;
+	}
+	*cs++ = cmd;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen4_emit_flush_rcs(struct i915_request *rq, u32 mode)
+{
+	u32 cmd, *cs;
+	int i;
+
+	/*
+	 * read/write caches:
+	 *
+	 * I915_GEM_DOMAIN_RENDER is always invalidated, but is
+	 * only flushed if MI_NO_WRITE_FLUSH is unset.  On 965, it is
+	 * also flushed at 2d versus 3d pipeline switches.
+	 *
+	 * read-only caches:
+	 *
+	 * I915_GEM_DOMAIN_SAMPLER is flushed on pre-965 if
+	 * MI_READ_FLUSH is set, and is always flushed on 965.
+	 *
+	 * I915_GEM_DOMAIN_COMMAND may not exist?
+	 *
+	 * I915_GEM_DOMAIN_INSTRUCTION, which exists on 965, is
+	 * invalidated when MI_EXE_FLUSH is set.
+	 *
+	 * I915_GEM_DOMAIN_VERTEX, which exists on 965, is
+	 * invalidated with every MI_FLUSH.
+	 *
+	 * TLBs:
+	 *
+	 * On 965, TLBs associated with I915_GEM_DOMAIN_COMMAND
+	 * and I915_GEM_DOMAIN_CPU in are invalidated at PTE write and
+	 * I915_GEM_DOMAIN_RENDER and I915_GEM_DOMAIN_SAMPLER
+	 * are flushed at any MI_FLUSH.
+	 */
+
+	cmd = MI_FLUSH;
+	if (mode & EMIT_INVALIDATE) {
+		cmd |= MI_EXE_FLUSH;
+		if (IS_G4X(rq->engine->i915) || GRAPHICS_VER(rq->engine->i915) == 5)
+			cmd |= MI_INVALIDATE_ISP;
+	}
+
+	i = 2;
+	if (mode & EMIT_INVALIDATE)
+		i += 20;
+
+	cs = intel_ring_begin(rq, i);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = cmd;
+
+	/*
+	 * A random delay to let the CS invalidate take effect? Without this
+	 * delay, the GPU relocation path fails as the CS does not see
+	 * the updated contents. Just as important, if we apply the flushes
+	 * to the EMIT_FLUSH branch (i.e. immediately after the relocation
+	 * write and before the invalidate on the next batch), the relocations
+	 * still fail. This implies that is a delay following invalidation
+	 * that is required to reset the caches as opposed to a delay to
+	 * ensure the memory is written.
+	 */
+	if (mode & EMIT_INVALIDATE) {
+		*cs++ = GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE;
+		*cs++ = intel_gt_scratch_offset(rq->engine->gt,
+						INTEL_GT_SCRATCH_FIELD_DEFAULT) |
+			PIPE_CONTROL_GLOBAL_GTT;
+		*cs++ = 0;
+		*cs++ = 0;
+
+		for (i = 0; i < 12; i++)
+			*cs++ = MI_FLUSH;
+
+		*cs++ = GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE;
+		*cs++ = intel_gt_scratch_offset(rq->engine->gt,
+						INTEL_GT_SCRATCH_FIELD_DEFAULT) |
+			PIPE_CONTROL_GLOBAL_GTT;
+		*cs++ = 0;
+		*cs++ = 0;
+	}
+
+	*cs++ = cmd;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen4_emit_flush_vcs(struct i915_request *rq, u32 mode)
+{
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_FLUSH;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static u32 *__gen2_emit_breadcrumb(struct i915_request *rq, u32 *cs,
+				   int flush, int post)
+{
+	GEM_BUG_ON(i915_request_active_timeline(rq)->hwsp_ggtt != rq->engine->status_page.vma);
+	GEM_BUG_ON(offset_in_page(rq->hwsp_seqno) != I915_GEM_HWS_SEQNO_ADDR);
+
+	*cs++ = MI_FLUSH;
+
+	while (flush--) {
+		*cs++ = MI_STORE_DWORD_INDEX;
+		*cs++ = I915_GEM_HWS_SCRATCH * sizeof(u32);
+		*cs++ = rq->fence.seqno;
+	}
+
+	while (post--) {
+		*cs++ = MI_STORE_DWORD_INDEX;
+		*cs++ = I915_GEM_HWS_SEQNO_ADDR;
+		*cs++ = rq->fence.seqno;
+	}
+
+	*cs++ = MI_USER_INTERRUPT;
+
+	rq->tail = intel_ring_offset(rq, cs);
+	assert_ring_tail_valid(rq->ring, rq->tail);
+
+	return cs;
+}
+
+u32 *gen3_emit_breadcrumb(struct i915_request *rq, u32 *cs)
+{
+	return __gen2_emit_breadcrumb(rq, cs, 16, 8);
+}
+
+u32 *gen5_emit_breadcrumb(struct i915_request *rq, u32 *cs)
+{
+	return __gen2_emit_breadcrumb(rq, cs, 8, 8);
+}
+
+/* Just userspace ABI convention to limit the wa batch bo to a resonable size */
+#define I830_BATCH_LIMIT SZ_256K
+#define I830_TLB_ENTRIES (2)
+#define I830_WA_SIZE max(I830_TLB_ENTRIES * SZ_4K, I830_BATCH_LIMIT)
+int i830_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 len,
+		       unsigned int dispatch_flags)
+{
+	u32 *cs, cs_offset =
+		intel_gt_scratch_offset(rq->engine->gt,
+					INTEL_GT_SCRATCH_FIELD_DEFAULT);
+
+	GEM_BUG_ON(rq->engine->gt->scratch->size < I830_WA_SIZE);
+
+	cs = intel_ring_begin(rq, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/* Evict the invalid PTE TLBs */
+	*cs++ = COLOR_BLT_CMD | BLT_WRITE_RGBA;
+	*cs++ = BLT_DEPTH_32 | BLT_ROP_COLOR_COPY | 4096;
+	*cs++ = I830_TLB_ENTRIES << 16 | 4; /* load each page */
+	*cs++ = cs_offset;
+	*cs++ = 0xdeadbeef;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	if ((dispatch_flags & I915_DISPATCH_PINNED) == 0) {
+		if (len > I830_BATCH_LIMIT)
+			return -ENOSPC;
+
+		cs = intel_ring_begin(rq, 6 + 2);
+		if (IS_ERR(cs))
+			return PTR_ERR(cs);
+
+		/*
+		 * Blit the batch (which has now all relocs applied) to the
+		 * stable batch scratch bo area (so that the CS never
+		 * stumbles over its tlb invalidation bug) ...
+		 */
+		*cs++ = SRC_COPY_BLT_CMD | BLT_WRITE_RGBA | (6 - 2);
+		*cs++ = BLT_DEPTH_32 | BLT_ROP_SRC_COPY | 4096;
+		*cs++ = DIV_ROUND_UP(len, 4096) << 16 | 4096;
+		*cs++ = cs_offset;
+		*cs++ = 4096;
+		*cs++ = offset;
+
+		*cs++ = MI_FLUSH;
+		*cs++ = MI_NOOP;
+		intel_ring_advance(rq, cs);
+
+		/* ... and execute it. */
+		offset = cs_offset;
+	}
+
+	if (!(dispatch_flags & I915_DISPATCH_SECURE))
+		offset |= MI_BATCH_NON_SECURE;
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT;
+	*cs++ = offset;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen3_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 len,
+		       unsigned int dispatch_flags)
+{
+	u32 *cs;
+
+	if (!(dispatch_flags & I915_DISPATCH_SECURE))
+		offset |= MI_BATCH_NON_SECURE;
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT;
+	*cs++ = offset;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen4_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 length,
+		       unsigned int dispatch_flags)
+{
+	u32 security;
+	u32 *cs;
+
+	security = MI_BATCH_NON_SECURE_I965;
+	if (dispatch_flags & I915_DISPATCH_SECURE)
+		security = 0;
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT | security;
+	*cs++ = offset;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+void gen2_irq_enable(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	i915->irq_mask &= ~engine->irq_enable_mask;
+	intel_uncore_write16(&i915->uncore, GEN2_IMR, i915->irq_mask);
+	ENGINE_POSTING_READ16(engine, RING_IMR);
+}
+
+void gen2_irq_disable(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	i915->irq_mask |= engine->irq_enable_mask;
+	intel_uncore_write16(&i915->uncore, GEN2_IMR, i915->irq_mask);
+}
+
+void gen3_irq_enable(struct intel_engine_cs *engine)
+{
+	engine->i915->irq_mask &= ~engine->irq_enable_mask;
+	intel_uncore_write(engine->uncore, GEN2_IMR, engine->i915->irq_mask);
+	intel_uncore_posting_read_fw(engine->uncore, GEN2_IMR);
+}
+
+void gen3_irq_disable(struct intel_engine_cs *engine)
+{
+	engine->i915->irq_mask |= engine->irq_enable_mask;
+	intel_uncore_write(engine->uncore, GEN2_IMR, engine->i915->irq_mask);
+}
+
+void gen5_irq_enable(struct intel_engine_cs *engine)
+{
+	gen5_gt_enable_irq(engine->gt, engine->irq_enable_mask);
+}
+
+void gen5_irq_disable(struct intel_engine_cs *engine)
+{
+	gen5_gt_disable_irq(engine->gt, engine->irq_enable_mask);
+}
diff --git a/drivers/gpu/drm/i915/gt/gen2_engine_cs.h b/drivers/gpu/drm/i915/gt/gen2_engine_cs.h
new file mode 100644
index 000000000..a5cd64a65
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen2_engine_cs.h
@@ -0,0 +1,38 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __GEN2_ENGINE_CS_H__
+#define __GEN2_ENGINE_CS_H__
+
+#include <linux/types.h>
+
+struct i915_request;
+struct intel_engine_cs;
+
+int gen2_emit_flush(struct i915_request *rq, u32 mode);
+int gen4_emit_flush_rcs(struct i915_request *rq, u32 mode);
+int gen4_emit_flush_vcs(struct i915_request *rq, u32 mode);
+
+u32 *gen3_emit_breadcrumb(struct i915_request *rq, u32 *cs);
+u32 *gen5_emit_breadcrumb(struct i915_request *rq, u32 *cs);
+
+int i830_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 len,
+		       unsigned int dispatch_flags);
+int gen3_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 len,
+		       unsigned int dispatch_flags);
+int gen4_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 length,
+		       unsigned int dispatch_flags);
+
+void gen2_irq_enable(struct intel_engine_cs *engine);
+void gen2_irq_disable(struct intel_engine_cs *engine);
+void gen3_irq_enable(struct intel_engine_cs *engine);
+void gen3_irq_disable(struct intel_engine_cs *engine);
+void gen5_irq_enable(struct intel_engine_cs *engine);
+void gen5_irq_disable(struct intel_engine_cs *engine);
+
+#endif /* __GEN2_ENGINE_CS_H__ */
diff --git a/drivers/gpu/drm/i915/gt/gen6_engine_cs.c b/drivers/gpu/drm/i915/gt/gen6_engine_cs.c
new file mode 100644
index 000000000..5e65550b4
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen6_engine_cs.c
@@ -0,0 +1,456 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "gen6_engine_cs.h"
+#include "intel_engine.h"
+#include "intel_engine_regs.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt.h"
+#include "intel_gt_irq.h"
+#include "intel_gt_pm_irq.h"
+#include "intel_ring.h"
+
+#define HWS_SCRATCH_ADDR	(I915_GEM_HWS_SCRATCH * sizeof(u32))
+
+/*
+ * Emits a PIPE_CONTROL with a non-zero post-sync operation, for
+ * implementing two workarounds on gen6.  From section 1.4.7.1
+ * "PIPE_CONTROL" of the Sandy Bridge PRM volume 2 part 1:
+ *
+ * [DevSNB-C+{W/A}] Before any depth stall flush (including those
+ * produced by non-pipelined state commands), software needs to first
+ * send a PIPE_CONTROL with no bits set except Post-Sync Operation !=
+ * 0.
+ *
+ * [Dev-SNB{W/A}]: Before a PIPE_CONTROL with Write Cache Flush Enable
+ * =1, a PIPE_CONTROL with any non-zero post-sync-op is required.
+ *
+ * And the workaround for these two requires this workaround first:
+ *
+ * [Dev-SNB{W/A}]: Pipe-control with CS-stall bit set must be sent
+ * BEFORE the pipe-control with a post-sync op and no write-cache
+ * flushes.
+ *
+ * And this last workaround is tricky because of the requirements on
+ * that bit.  From section 1.4.7.2.3 "Stall" of the Sandy Bridge PRM
+ * volume 2 part 1:
+ *
+ *     "1 of the following must also be set:
+ *      - Render Target Cache Flush Enable ([12] of DW1)
+ *      - Depth Cache Flush Enable ([0] of DW1)
+ *      - Stall at Pixel Scoreboard ([1] of DW1)
+ *      - Depth Stall ([13] of DW1)
+ *      - Post-Sync Operation ([13] of DW1)
+ *      - Notify Enable ([8] of DW1)"
+ *
+ * The cache flushes require the workaround flush that triggered this
+ * one, so we can't use it.  Depth stall would trigger the same.
+ * Post-sync nonzero is what triggered this second workaround, so we
+ * can't use that one either.  Notify enable is IRQs, which aren't
+ * really our business.  That leaves only stall at scoreboard.
+ */
+static int
+gen6_emit_post_sync_nonzero_flush(struct i915_request *rq)
+{
+	u32 scratch_addr =
+		intel_gt_scratch_offset(rq->engine->gt,
+					INTEL_GT_SCRATCH_FIELD_RENDER_FLUSH);
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = GFX_OP_PIPE_CONTROL(5);
+	*cs++ = PIPE_CONTROL_CS_STALL | PIPE_CONTROL_STALL_AT_SCOREBOARD;
+	*cs++ = scratch_addr | PIPE_CONTROL_GLOBAL_GTT;
+	*cs++ = 0; /* low dword */
+	*cs++ = 0; /* high dword */
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	cs = intel_ring_begin(rq, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = GFX_OP_PIPE_CONTROL(5);
+	*cs++ = PIPE_CONTROL_QW_WRITE;
+	*cs++ = scratch_addr | PIPE_CONTROL_GLOBAL_GTT;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen6_emit_flush_rcs(struct i915_request *rq, u32 mode)
+{
+	u32 scratch_addr =
+		intel_gt_scratch_offset(rq->engine->gt,
+					INTEL_GT_SCRATCH_FIELD_RENDER_FLUSH);
+	u32 *cs, flags = 0;
+	int ret;
+
+	/* Force SNB workarounds for PIPE_CONTROL flushes */
+	ret = gen6_emit_post_sync_nonzero_flush(rq);
+	if (ret)
+		return ret;
+
+	/*
+	 * Just flush everything.  Experiments have shown that reducing the
+	 * number of bits based on the write domains has little performance
+	 * impact. And when rearranging requests, the order of flushes is
+	 * unknown.
+	 */
+	if (mode & EMIT_FLUSH) {
+		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
+		/*
+		 * Ensure that any following seqno writes only happen
+		 * when the render cache is indeed flushed.
+		 */
+		flags |= PIPE_CONTROL_CS_STALL;
+	}
+	if (mode & EMIT_INVALIDATE) {
+		flags |= PIPE_CONTROL_TLB_INVALIDATE;
+		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+		/*
+		 * TLB invalidate requires a post-sync write.
+		 */
+		flags |= PIPE_CONTROL_QW_WRITE | PIPE_CONTROL_CS_STALL;
+	}
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = flags;
+	*cs++ = scratch_addr | PIPE_CONTROL_GLOBAL_GTT;
+	*cs++ = 0;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+u32 *gen6_emit_breadcrumb_rcs(struct i915_request *rq, u32 *cs)
+{
+	/* First we do the gen6_emit_post_sync_nonzero_flush w/a */
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = PIPE_CONTROL_CS_STALL | PIPE_CONTROL_STALL_AT_SCOREBOARD;
+	*cs++ = 0;
+	*cs++ = 0;
+
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = PIPE_CONTROL_QW_WRITE;
+	*cs++ = intel_gt_scratch_offset(rq->engine->gt,
+					INTEL_GT_SCRATCH_FIELD_DEFAULT) |
+		PIPE_CONTROL_GLOBAL_GTT;
+	*cs++ = 0;
+
+	/* Finally we can flush and with it emit the breadcrumb */
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = (PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
+		 PIPE_CONTROL_DEPTH_CACHE_FLUSH |
+		 PIPE_CONTROL_DC_FLUSH_ENABLE |
+		 PIPE_CONTROL_QW_WRITE |
+		 PIPE_CONTROL_CS_STALL);
+	*cs++ = i915_request_active_seqno(rq) |
+		PIPE_CONTROL_GLOBAL_GTT;
+	*cs++ = rq->fence.seqno;
+
+	*cs++ = MI_USER_INTERRUPT;
+	*cs++ = MI_NOOP;
+
+	rq->tail = intel_ring_offset(rq, cs);
+	assert_ring_tail_valid(rq->ring, rq->tail);
+
+	return cs;
+}
+
+static int mi_flush_dw(struct i915_request *rq, u32 flags)
+{
+	u32 cmd, *cs;
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	cmd = MI_FLUSH_DW;
+
+	/*
+	 * We always require a command barrier so that subsequent
+	 * commands, such as breadcrumb interrupts, are strictly ordered
+	 * wrt the contents of the write cache being flushed to memory
+	 * (and thus being coherent from the CPU).
+	 */
+	cmd |= MI_FLUSH_DW_STORE_INDEX | MI_FLUSH_DW_OP_STOREDW;
+
+	/*
+	 * Bspec vol 1c.3 - blitter engine command streamer:
+	 * "If ENABLED, all TLBs will be invalidated once the flush
+	 * operation is complete. This bit is only valid when the
+	 * Post-Sync Operation field is a value of 1h or 3h."
+	 */
+	cmd |= flags;
+
+	*cs++ = cmd;
+	*cs++ = HWS_SCRATCH_ADDR | MI_FLUSH_DW_USE_GTT;
+	*cs++ = 0;
+	*cs++ = MI_NOOP;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int gen6_flush_dw(struct i915_request *rq, u32 mode, u32 invflags)
+{
+	return mi_flush_dw(rq, mode & EMIT_INVALIDATE ? invflags : 0);
+}
+
+int gen6_emit_flush_xcs(struct i915_request *rq, u32 mode)
+{
+	return gen6_flush_dw(rq, mode, MI_INVALIDATE_TLB);
+}
+
+int gen6_emit_flush_vcs(struct i915_request *rq, u32 mode)
+{
+	return gen6_flush_dw(rq, mode, MI_INVALIDATE_TLB | MI_INVALIDATE_BSD);
+}
+
+int gen6_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 len,
+		       unsigned int dispatch_flags)
+{
+	u32 security;
+	u32 *cs;
+
+	security = MI_BATCH_NON_SECURE_I965;
+	if (dispatch_flags & I915_DISPATCH_SECURE)
+		security = 0;
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	cs = __gen6_emit_bb_start(cs, offset, security);
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int
+hsw_emit_bb_start(struct i915_request *rq,
+		  u64 offset, u32 len,
+		  unsigned int dispatch_flags)
+{
+	u32 security;
+	u32 *cs;
+
+	security = MI_BATCH_PPGTT_HSW | MI_BATCH_NON_SECURE_HSW;
+	if (dispatch_flags & I915_DISPATCH_SECURE)
+		security = 0;
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	cs = __gen6_emit_bb_start(cs, offset, security);
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int gen7_stall_cs(struct i915_request *rq)
+{
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = PIPE_CONTROL_CS_STALL | PIPE_CONTROL_STALL_AT_SCOREBOARD;
+	*cs++ = 0;
+	*cs++ = 0;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen7_emit_flush_rcs(struct i915_request *rq, u32 mode)
+{
+	u32 scratch_addr =
+		intel_gt_scratch_offset(rq->engine->gt,
+					INTEL_GT_SCRATCH_FIELD_RENDER_FLUSH);
+	u32 *cs, flags = 0;
+
+	/*
+	 * Ensure that any following seqno writes only happen when the render
+	 * cache is indeed flushed.
+	 *
+	 * Workaround: 4th PIPE_CONTROL command (except the ones with only
+	 * read-cache invalidate bits set) must have the CS_STALL bit set. We
+	 * don't try to be clever and just set it unconditionally.
+	 */
+	flags |= PIPE_CONTROL_CS_STALL;
+
+	/*
+	 * CS_STALL suggests at least a post-sync write.
+	 */
+	flags |= PIPE_CONTROL_QW_WRITE;
+	flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
+
+	/*
+	 * Just flush everything.  Experiments have shown that reducing the
+	 * number of bits based on the write domains has little performance
+	 * impact.
+	 */
+	if (mode & EMIT_FLUSH) {
+		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
+		flags |= PIPE_CONTROL_FLUSH_ENABLE;
+	}
+	if (mode & EMIT_INVALIDATE) {
+		flags |= PIPE_CONTROL_TLB_INVALIDATE;
+		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_MEDIA_STATE_CLEAR;
+
+		/*
+		 * Workaround: we must issue a pipe_control with CS-stall bit
+		 * set before a pipe_control command that has the state cache
+		 * invalidate bit set.
+		 */
+		gen7_stall_cs(rq);
+	}
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = flags;
+	*cs++ = scratch_addr;
+	*cs++ = 0;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+u32 *gen7_emit_breadcrumb_rcs(struct i915_request *rq, u32 *cs)
+{
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = (PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
+		 PIPE_CONTROL_DEPTH_CACHE_FLUSH |
+		 PIPE_CONTROL_DC_FLUSH_ENABLE |
+		 PIPE_CONTROL_FLUSH_ENABLE |
+		 PIPE_CONTROL_QW_WRITE |
+		 PIPE_CONTROL_GLOBAL_GTT_IVB |
+		 PIPE_CONTROL_CS_STALL);
+	*cs++ = i915_request_active_seqno(rq);
+	*cs++ = rq->fence.seqno;
+
+	*cs++ = MI_USER_INTERRUPT;
+	*cs++ = MI_NOOP;
+
+	rq->tail = intel_ring_offset(rq, cs);
+	assert_ring_tail_valid(rq->ring, rq->tail);
+
+	return cs;
+}
+
+u32 *gen6_emit_breadcrumb_xcs(struct i915_request *rq, u32 *cs)
+{
+	GEM_BUG_ON(i915_request_active_timeline(rq)->hwsp_ggtt != rq->engine->status_page.vma);
+	GEM_BUG_ON(offset_in_page(rq->hwsp_seqno) != I915_GEM_HWS_SEQNO_ADDR);
+
+	*cs++ = MI_FLUSH_DW | MI_FLUSH_DW_OP_STOREDW | MI_FLUSH_DW_STORE_INDEX;
+	*cs++ = I915_GEM_HWS_SEQNO_ADDR | MI_FLUSH_DW_USE_GTT;
+	*cs++ = rq->fence.seqno;
+
+	*cs++ = MI_USER_INTERRUPT;
+
+	rq->tail = intel_ring_offset(rq, cs);
+	assert_ring_tail_valid(rq->ring, rq->tail);
+
+	return cs;
+}
+
+#define GEN7_XCS_WA 32
+u32 *gen7_emit_breadcrumb_xcs(struct i915_request *rq, u32 *cs)
+{
+	int i;
+
+	GEM_BUG_ON(i915_request_active_timeline(rq)->hwsp_ggtt != rq->engine->status_page.vma);
+	GEM_BUG_ON(offset_in_page(rq->hwsp_seqno) != I915_GEM_HWS_SEQNO_ADDR);
+
+	*cs++ = MI_FLUSH_DW | MI_INVALIDATE_TLB |
+		MI_FLUSH_DW_OP_STOREDW | MI_FLUSH_DW_STORE_INDEX;
+	*cs++ = I915_GEM_HWS_SEQNO_ADDR | MI_FLUSH_DW_USE_GTT;
+	*cs++ = rq->fence.seqno;
+
+	for (i = 0; i < GEN7_XCS_WA; i++) {
+		*cs++ = MI_STORE_DWORD_INDEX;
+		*cs++ = I915_GEM_HWS_SEQNO_ADDR;
+		*cs++ = rq->fence.seqno;
+	}
+
+	*cs++ = MI_FLUSH_DW;
+	*cs++ = 0;
+	*cs++ = 0;
+
+	*cs++ = MI_USER_INTERRUPT;
+	*cs++ = MI_NOOP;
+
+	rq->tail = intel_ring_offset(rq, cs);
+	assert_ring_tail_valid(rq->ring, rq->tail);
+
+	return cs;
+}
+#undef GEN7_XCS_WA
+
+void gen6_irq_enable(struct intel_engine_cs *engine)
+{
+	ENGINE_WRITE(engine, RING_IMR,
+		     ~(engine->irq_enable_mask | engine->irq_keep_mask));
+
+	/* Flush/delay to ensure the RING_IMR is active before the GT IMR */
+	ENGINE_POSTING_READ(engine, RING_IMR);
+
+	gen5_gt_enable_irq(engine->gt, engine->irq_enable_mask);
+}
+
+void gen6_irq_disable(struct intel_engine_cs *engine)
+{
+	ENGINE_WRITE(engine, RING_IMR, ~engine->irq_keep_mask);
+	gen5_gt_disable_irq(engine->gt, engine->irq_enable_mask);
+}
+
+void hsw_irq_enable_vecs(struct intel_engine_cs *engine)
+{
+	ENGINE_WRITE(engine, RING_IMR, ~engine->irq_enable_mask);
+
+	/* Flush/delay to ensure the RING_IMR is active before the GT IMR */
+	ENGINE_POSTING_READ(engine, RING_IMR);
+
+	gen6_gt_pm_unmask_irq(engine->gt, engine->irq_enable_mask);
+}
+
+void hsw_irq_disable_vecs(struct intel_engine_cs *engine)
+{
+	ENGINE_WRITE(engine, RING_IMR, ~0);
+	gen6_gt_pm_mask_irq(engine->gt, engine->irq_enable_mask);
+}
diff --git a/drivers/gpu/drm/i915/gt/gen6_engine_cs.h b/drivers/gpu/drm/i915/gt/gen6_engine_cs.h
new file mode 100644
index 000000000..76c6bc9f3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen6_engine_cs.h
@@ -0,0 +1,39 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __GEN6_ENGINE_CS_H__
+#define __GEN6_ENGINE_CS_H__
+
+#include <linux/types.h>
+
+#include "intel_gpu_commands.h"
+
+struct i915_request;
+struct intel_engine_cs;
+
+int gen6_emit_flush_rcs(struct i915_request *rq, u32 mode);
+int gen6_emit_flush_vcs(struct i915_request *rq, u32 mode);
+int gen6_emit_flush_xcs(struct i915_request *rq, u32 mode);
+u32 *gen6_emit_breadcrumb_rcs(struct i915_request *rq, u32 *cs);
+u32 *gen6_emit_breadcrumb_xcs(struct i915_request *rq, u32 *cs);
+
+int gen7_emit_flush_rcs(struct i915_request *rq, u32 mode);
+u32 *gen7_emit_breadcrumb_rcs(struct i915_request *rq, u32 *cs);
+u32 *gen7_emit_breadcrumb_xcs(struct i915_request *rq, u32 *cs);
+
+int gen6_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 len,
+		       unsigned int dispatch_flags);
+int hsw_emit_bb_start(struct i915_request *rq,
+		      u64 offset, u32 len,
+		      unsigned int dispatch_flags);
+
+void gen6_irq_enable(struct intel_engine_cs *engine);
+void gen6_irq_disable(struct intel_engine_cs *engine);
+
+void hsw_irq_enable_vecs(struct intel_engine_cs *engine);
+void hsw_irq_disable_vecs(struct intel_engine_cs *engine);
+
+#endif /* __GEN6_ENGINE_CS_H__ */
diff --git a/drivers/gpu/drm/i915/gt/gen6_ppgtt.c b/drivers/gpu/drm/i915/gt/gen6_ppgtt.c
new file mode 100644
index 000000000..5aaacc53f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen6_ppgtt.c
@@ -0,0 +1,467 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/log2.h>
+
+#include "gem/i915_gem_internal.h"
+
+#include "gen6_ppgtt.h"
+#include "i915_scatterlist.h"
+#include "i915_trace.h"
+#include "i915_vgpu.h"
+#include "intel_gt_regs.h"
+#include "intel_engine_regs.h"
+#include "intel_gt.h"
+
+/* Write pde (index) from the page directory @pd to the page table @pt */
+static void gen6_write_pde(const struct gen6_ppgtt *ppgtt,
+			   const unsigned int pde,
+			   const struct i915_page_table *pt)
+{
+	dma_addr_t addr = pt ? px_dma(pt) : px_dma(ppgtt->base.vm.scratch[1]);
+
+	/* Caller needs to make sure the write completes if necessary */
+	iowrite32(GEN6_PDE_ADDR_ENCODE(addr) | GEN6_PDE_VALID,
+		  ppgtt->pd_addr + pde);
+}
+
+void gen7_ppgtt_enable(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+	u32 ecochk;
+
+	intel_uncore_rmw(uncore, GAC_ECO_BITS, 0, ECOBITS_PPGTT_CACHE64B);
+
+	ecochk = intel_uncore_read(uncore, GAM_ECOCHK);
+	if (IS_HASWELL(i915)) {
+		ecochk |= ECOCHK_PPGTT_WB_HSW;
+	} else {
+		ecochk |= ECOCHK_PPGTT_LLC_IVB;
+		ecochk &= ~ECOCHK_PPGTT_GFDT_IVB;
+	}
+	intel_uncore_write(uncore, GAM_ECOCHK, ecochk);
+}
+
+void gen6_ppgtt_enable(struct intel_gt *gt)
+{
+	struct intel_uncore *uncore = gt->uncore;
+
+	intel_uncore_rmw(uncore,
+			 GAC_ECO_BITS,
+			 0,
+			 ECOBITS_SNB_BIT | ECOBITS_PPGTT_CACHE64B);
+
+	intel_uncore_rmw(uncore,
+			 GAB_CTL,
+			 0,
+			 GAB_CTL_CONT_AFTER_PAGEFAULT);
+
+	intel_uncore_rmw(uncore,
+			 GAM_ECOCHK,
+			 0,
+			 ECOCHK_SNB_BIT | ECOCHK_PPGTT_CACHE64B);
+
+	if (HAS_PPGTT(uncore->i915)) /* may be disabled for VT-d */
+		intel_uncore_write(uncore,
+				   GFX_MODE,
+				   _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
+}
+
+/* PPGTT support for Sandybdrige/Gen6 and later */
+static void gen6_ppgtt_clear_range(struct i915_address_space *vm,
+				   u64 start, u64 length)
+{
+	struct gen6_ppgtt * const ppgtt = to_gen6_ppgtt(i915_vm_to_ppgtt(vm));
+	const unsigned int first_entry = start / I915_GTT_PAGE_SIZE;
+	const gen6_pte_t scratch_pte = vm->scratch[0]->encode;
+	unsigned int pde = first_entry / GEN6_PTES;
+	unsigned int pte = first_entry % GEN6_PTES;
+	unsigned int num_entries = length / I915_GTT_PAGE_SIZE;
+
+	while (num_entries) {
+		struct i915_page_table * const pt =
+			i915_pt_entry(ppgtt->base.pd, pde++);
+		const unsigned int count = min(num_entries, GEN6_PTES - pte);
+		gen6_pte_t *vaddr;
+
+		num_entries -= count;
+
+		GEM_BUG_ON(count > atomic_read(&pt->used));
+		if (!atomic_sub_return(count, &pt->used))
+			ppgtt->scan_for_unused_pt = true;
+
+		/*
+		 * Note that the hw doesn't support removing PDE on the fly
+		 * (they are cached inside the context with no means to
+		 * invalidate the cache), so we can only reset the PTE
+		 * entries back to scratch.
+		 */
+
+		vaddr = px_vaddr(pt);
+		memset32(vaddr + pte, scratch_pte, count);
+
+		pte = 0;
+	}
+}
+
+static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
+				      struct i915_vma_resource *vma_res,
+				      enum i915_cache_level cache_level,
+				      u32 flags)
+{
+	struct i915_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
+	struct i915_page_directory * const pd = ppgtt->pd;
+	unsigned int first_entry = vma_res->start / I915_GTT_PAGE_SIZE;
+	unsigned int act_pt = first_entry / GEN6_PTES;
+	unsigned int act_pte = first_entry % GEN6_PTES;
+	const u32 pte_encode = vm->pte_encode(0, cache_level, flags);
+	struct sgt_dma iter = sgt_dma(vma_res);
+	gen6_pte_t *vaddr;
+
+	GEM_BUG_ON(!pd->entry[act_pt]);
+
+	vaddr = px_vaddr(i915_pt_entry(pd, act_pt));
+	do {
+		GEM_BUG_ON(sg_dma_len(iter.sg) < I915_GTT_PAGE_SIZE);
+		vaddr[act_pte] = pte_encode | GEN6_PTE_ADDR_ENCODE(iter.dma);
+
+		iter.dma += I915_GTT_PAGE_SIZE;
+		if (iter.dma == iter.max) {
+			iter.sg = __sg_next(iter.sg);
+			if (!iter.sg || sg_dma_len(iter.sg) == 0)
+				break;
+
+			iter.dma = sg_dma_address(iter.sg);
+			iter.max = iter.dma + sg_dma_len(iter.sg);
+		}
+
+		if (++act_pte == GEN6_PTES) {
+			vaddr = px_vaddr(i915_pt_entry(pd, ++act_pt));
+			act_pte = 0;
+		}
+	} while (1);
+
+	vma_res->page_sizes_gtt = I915_GTT_PAGE_SIZE;
+}
+
+static void gen6_flush_pd(struct gen6_ppgtt *ppgtt, u64 start, u64 end)
+{
+	struct i915_page_directory * const pd = ppgtt->base.pd;
+	struct i915_page_table *pt;
+	unsigned int pde;
+
+	start = round_down(start, SZ_64K);
+	end = round_up(end, SZ_64K) - start;
+
+	mutex_lock(&ppgtt->flush);
+
+	gen6_for_each_pde(pt, pd, start, end, pde)
+		gen6_write_pde(ppgtt, pde, pt);
+
+	mb();
+	ioread32(ppgtt->pd_addr + pde - 1);
+	gen6_ggtt_invalidate(ppgtt->base.vm.gt->ggtt);
+	mb();
+
+	mutex_unlock(&ppgtt->flush);
+}
+
+static void gen6_alloc_va_range(struct i915_address_space *vm,
+				struct i915_vm_pt_stash *stash,
+				u64 start, u64 length)
+{
+	struct gen6_ppgtt *ppgtt = to_gen6_ppgtt(i915_vm_to_ppgtt(vm));
+	struct i915_page_directory * const pd = ppgtt->base.pd;
+	struct i915_page_table *pt;
+	bool flush = false;
+	u64 from = start;
+	unsigned int pde;
+
+	spin_lock(&pd->lock);
+	gen6_for_each_pde(pt, pd, start, length, pde) {
+		const unsigned int count = gen6_pte_count(start, length);
+
+		if (!pt) {
+			spin_unlock(&pd->lock);
+
+			pt = stash->pt[0];
+			__i915_gem_object_pin_pages(pt->base);
+
+			fill32_px(pt, vm->scratch[0]->encode);
+
+			spin_lock(&pd->lock);
+			if (!pd->entry[pde]) {
+				stash->pt[0] = pt->stash;
+				atomic_set(&pt->used, 0);
+				pd->entry[pde] = pt;
+			} else {
+				pt = pd->entry[pde];
+			}
+
+			flush = true;
+		}
+
+		atomic_add(count, &pt->used);
+	}
+	spin_unlock(&pd->lock);
+
+	if (flush && i915_vma_is_bound(ppgtt->vma, I915_VMA_GLOBAL_BIND)) {
+		intel_wakeref_t wakeref;
+
+		with_intel_runtime_pm(&vm->i915->runtime_pm, wakeref)
+			gen6_flush_pd(ppgtt, from, start);
+	}
+}
+
+static int gen6_ppgtt_init_scratch(struct gen6_ppgtt *ppgtt)
+{
+	struct i915_address_space * const vm = &ppgtt->base.vm;
+	int ret;
+
+	ret = setup_scratch_page(vm);
+	if (ret)
+		return ret;
+
+	vm->scratch[0]->encode =
+		vm->pte_encode(px_dma(vm->scratch[0]),
+			       I915_CACHE_NONE, PTE_READ_ONLY);
+
+	vm->scratch[1] = vm->alloc_pt_dma(vm, I915_GTT_PAGE_SIZE_4K);
+	if (IS_ERR(vm->scratch[1])) {
+		ret = PTR_ERR(vm->scratch[1]);
+		goto err_scratch0;
+	}
+
+	ret = map_pt_dma(vm, vm->scratch[1]);
+	if (ret)
+		goto err_scratch1;
+
+	fill32_px(vm->scratch[1], vm->scratch[0]->encode);
+
+	return 0;
+
+err_scratch1:
+	i915_gem_object_put(vm->scratch[1]);
+err_scratch0:
+	i915_gem_object_put(vm->scratch[0]);
+	vm->scratch[0] = NULL;
+	return ret;
+}
+
+static void gen6_ppgtt_free_pd(struct gen6_ppgtt *ppgtt)
+{
+	struct i915_page_directory * const pd = ppgtt->base.pd;
+	struct i915_page_table *pt;
+	u32 pde;
+
+	gen6_for_all_pdes(pt, pd, pde)
+		if (pt)
+			free_pt(&ppgtt->base.vm, pt);
+}
+
+static void gen6_ppgtt_cleanup(struct i915_address_space *vm)
+{
+	struct gen6_ppgtt *ppgtt = to_gen6_ppgtt(i915_vm_to_ppgtt(vm));
+
+	gen6_ppgtt_free_pd(ppgtt);
+	free_scratch(vm);
+
+	if (ppgtt->base.pd)
+		free_pd(&ppgtt->base.vm, ppgtt->base.pd);
+
+	mutex_destroy(&ppgtt->flush);
+}
+
+static void pd_vma_bind(struct i915_address_space *vm,
+			struct i915_vm_pt_stash *stash,
+			struct i915_vma_resource *vma_res,
+			enum i915_cache_level cache_level,
+			u32 unused)
+{
+	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
+	struct gen6_ppgtt *ppgtt = vma_res->private;
+	u32 ggtt_offset = vma_res->start / I915_GTT_PAGE_SIZE;
+
+	ppgtt->pp_dir = ggtt_offset * sizeof(gen6_pte_t) << 10;
+	ppgtt->pd_addr = (gen6_pte_t __iomem *)ggtt->gsm + ggtt_offset;
+
+	gen6_flush_pd(ppgtt, 0, ppgtt->base.vm.total);
+}
+
+static void pd_vma_unbind(struct i915_address_space *vm,
+			  struct i915_vma_resource *vma_res)
+{
+	struct gen6_ppgtt *ppgtt = vma_res->private;
+	struct i915_page_directory * const pd = ppgtt->base.pd;
+	struct i915_page_table *pt;
+	unsigned int pde;
+
+	if (!ppgtt->scan_for_unused_pt)
+		return;
+
+	/* Free all no longer used page tables */
+	gen6_for_all_pdes(pt, ppgtt->base.pd, pde) {
+		if (!pt || atomic_read(&pt->used))
+			continue;
+
+		free_pt(&ppgtt->base.vm, pt);
+		pd->entry[pde] = NULL;
+	}
+
+	ppgtt->scan_for_unused_pt = false;
+}
+
+static const struct i915_vma_ops pd_vma_ops = {
+	.bind_vma = pd_vma_bind,
+	.unbind_vma = pd_vma_unbind,
+};
+
+int gen6_ppgtt_pin(struct i915_ppgtt *base, struct i915_gem_ww_ctx *ww)
+{
+	struct gen6_ppgtt *ppgtt = to_gen6_ppgtt(base);
+	int err;
+
+	GEM_BUG_ON(!kref_read(&ppgtt->base.vm.ref));
+
+	/*
+	 * Workaround the limited maximum vma->pin_count and the aliasing_ppgtt
+	 * which will be pinned into every active context.
+	 * (When vma->pin_count becomes atomic, I expect we will naturally
+	 * need a larger, unpacked, type and kill this redundancy.)
+	 */
+	if (atomic_add_unless(&ppgtt->pin_count, 1, 0))
+		return 0;
+
+	/* grab the ppgtt resv to pin the object */
+	err = i915_vm_lock_objects(&ppgtt->base.vm, ww);
+	if (err)
+		return err;
+
+	/*
+	 * PPGTT PDEs reside in the GGTT and consists of 512 entries. The
+	 * allocator works in address space sizes, so it's multiplied by page
+	 * size. We allocate at the top of the GTT to avoid fragmentation.
+	 */
+	if (!atomic_read(&ppgtt->pin_count)) {
+		err = i915_ggtt_pin(ppgtt->vma, ww, GEN6_PD_ALIGN, PIN_HIGH);
+
+		GEM_BUG_ON(ppgtt->vma->fence);
+		clear_bit(I915_VMA_CAN_FENCE_BIT, __i915_vma_flags(ppgtt->vma));
+	}
+	if (!err)
+		atomic_inc(&ppgtt->pin_count);
+
+	return err;
+}
+
+static int pd_dummy_obj_get_pages(struct drm_i915_gem_object *obj)
+{
+	obj->mm.pages = ZERO_SIZE_PTR;
+	return 0;
+}
+
+static void pd_dummy_obj_put_pages(struct drm_i915_gem_object *obj,
+				   struct sg_table *pages)
+{
+}
+
+static const struct drm_i915_gem_object_ops pd_dummy_obj_ops = {
+	.name = "pd_dummy_obj",
+	.get_pages = pd_dummy_obj_get_pages,
+	.put_pages = pd_dummy_obj_put_pages,
+};
+
+static struct i915_page_directory *
+gen6_alloc_top_pd(struct gen6_ppgtt *ppgtt)
+{
+	struct i915_ggtt * const ggtt = ppgtt->base.vm.gt->ggtt;
+	struct i915_page_directory *pd;
+	int err;
+
+	pd = __alloc_pd(I915_PDES);
+	if (unlikely(!pd))
+		return ERR_PTR(-ENOMEM);
+
+	pd->pt.base = __i915_gem_object_create_internal(ppgtt->base.vm.gt->i915,
+							&pd_dummy_obj_ops,
+							I915_PDES * SZ_4K);
+	if (IS_ERR(pd->pt.base)) {
+		err = PTR_ERR(pd->pt.base);
+		pd->pt.base = NULL;
+		goto err_pd;
+	}
+
+	pd->pt.base->base.resv = i915_vm_resv_get(&ppgtt->base.vm);
+	pd->pt.base->shares_resv_from = &ppgtt->base.vm;
+
+	ppgtt->vma = i915_vma_instance(pd->pt.base, &ggtt->vm, NULL);
+	if (IS_ERR(ppgtt->vma)) {
+		err = PTR_ERR(ppgtt->vma);
+		ppgtt->vma = NULL;
+		goto err_pd;
+	}
+
+	/* The dummy object we create is special, override ops.. */
+	ppgtt->vma->ops = &pd_vma_ops;
+	ppgtt->vma->private = ppgtt;
+	return pd;
+
+err_pd:
+	free_pd(&ppgtt->base.vm, pd);
+	return ERR_PTR(err);
+}
+
+void gen6_ppgtt_unpin(struct i915_ppgtt *base)
+{
+	struct gen6_ppgtt *ppgtt = to_gen6_ppgtt(base);
+
+	GEM_BUG_ON(!atomic_read(&ppgtt->pin_count));
+	if (atomic_dec_and_test(&ppgtt->pin_count))
+		i915_vma_unpin(ppgtt->vma);
+}
+
+struct i915_ppgtt *gen6_ppgtt_create(struct intel_gt *gt)
+{
+	struct i915_ggtt * const ggtt = gt->ggtt;
+	struct gen6_ppgtt *ppgtt;
+	int err;
+
+	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+	if (!ppgtt)
+		return ERR_PTR(-ENOMEM);
+
+	mutex_init(&ppgtt->flush);
+
+	ppgtt_init(&ppgtt->base, gt, 0);
+	ppgtt->base.vm.pd_shift = ilog2(SZ_4K * SZ_4K / sizeof(gen6_pte_t));
+	ppgtt->base.vm.top = 1;
+
+	ppgtt->base.vm.bind_async_flags = I915_VMA_LOCAL_BIND;
+	ppgtt->base.vm.allocate_va_range = gen6_alloc_va_range;
+	ppgtt->base.vm.clear_range = gen6_ppgtt_clear_range;
+	ppgtt->base.vm.insert_entries = gen6_ppgtt_insert_entries;
+	ppgtt->base.vm.cleanup = gen6_ppgtt_cleanup;
+
+	ppgtt->base.vm.alloc_pt_dma = alloc_pt_dma;
+	ppgtt->base.vm.alloc_scratch_dma = alloc_pt_dma;
+	ppgtt->base.vm.pte_encode = ggtt->vm.pte_encode;
+
+	err = gen6_ppgtt_init_scratch(ppgtt);
+	if (err)
+		goto err_put;
+
+	ppgtt->base.pd = gen6_alloc_top_pd(ppgtt);
+	if (IS_ERR(ppgtt->base.pd)) {
+		err = PTR_ERR(ppgtt->base.pd);
+		goto err_put;
+	}
+
+	return &ppgtt->base;
+
+err_put:
+	i915_vm_put(&ppgtt->base.vm);
+	return ERR_PTR(err);
+}
diff --git a/drivers/gpu/drm/i915/gt/gen6_ppgtt.h b/drivers/gpu/drm/i915/gt/gen6_ppgtt.h
new file mode 100644
index 000000000..5e5cf2ec3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen6_ppgtt.h
@@ -0,0 +1,77 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __GEN6_PPGTT_H__
+#define __GEN6_PPGTT_H__
+
+#include "intel_gtt.h"
+
+struct i915_gem_ww_ctx;
+
+struct gen6_ppgtt {
+	struct i915_ppgtt base;
+
+	struct mutex flush;
+	struct i915_vma *vma;
+	gen6_pte_t __iomem *pd_addr;
+	u32 pp_dir;
+
+	atomic_t pin_count;
+
+	bool scan_for_unused_pt;
+};
+
+static inline u32 gen6_pte_index(u32 addr)
+{
+	return i915_pte_index(addr, GEN6_PDE_SHIFT);
+}
+
+static inline u32 gen6_pte_count(u32 addr, u32 length)
+{
+	return i915_pte_count(addr, length, GEN6_PDE_SHIFT);
+}
+
+static inline u32 gen6_pde_index(u32 addr)
+{
+	return i915_pde_index(addr, GEN6_PDE_SHIFT);
+}
+
+#define __to_gen6_ppgtt(base) container_of(base, struct gen6_ppgtt, base)
+
+static inline struct gen6_ppgtt *to_gen6_ppgtt(struct i915_ppgtt *base)
+{
+	BUILD_BUG_ON(offsetof(struct gen6_ppgtt, base));
+	return __to_gen6_ppgtt(base);
+}
+
+/*
+ * gen6_for_each_pde() iterates over every pde from start until start+length.
+ * If start and start+length are not perfectly divisible, the macro will round
+ * down and up as needed. Start=0 and length=2G effectively iterates over
+ * every PDE in the system. The macro modifies ALL its parameters except 'pd',
+ * so each of the other parameters should preferably be a simple variable, or
+ * at most an lvalue with no side-effects!
+ */
+#define gen6_for_each_pde(pt, pd, start, length, iter)			\
+	for (iter = gen6_pde_index(start);				\
+	     length > 0 && iter < I915_PDES &&				\
+		     (pt = i915_pt_entry(pd, iter), true);		\
+	     ({ u32 temp = ALIGN(start + 1, 1 << GEN6_PDE_SHIFT);	\
+		    temp = min(temp - start, length);			\
+		    start += temp; length -= temp; }), ++iter)
+
+#define gen6_for_all_pdes(pt, pd, iter)					\
+	for (iter = 0;							\
+	     iter < I915_PDES &&					\
+		     (pt = i915_pt_entry(pd, iter), true);		\
+	     ++iter)
+
+int gen6_ppgtt_pin(struct i915_ppgtt *base, struct i915_gem_ww_ctx *ww);
+void gen6_ppgtt_unpin(struct i915_ppgtt *base);
+void gen6_ppgtt_enable(struct intel_gt *gt);
+void gen7_ppgtt_enable(struct intel_gt *gt);
+struct i915_ppgtt *gen6_ppgtt_create(struct intel_gt *gt);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/gen6_renderstate.c b/drivers/gpu/drm/i915/gt/gen6_renderstate.c
new file mode 100644
index 000000000..555e83f3a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen6_renderstate.c
@@ -0,0 +1,297 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Generated by: intel-gpu-tools-1.8-220-g01153e7
+ */
+
+#include "intel_renderstate.h"
+
+static const u32 gen6_null_state_relocs[] = {
+	0x00000020,
+	0x00000024,
+	0x0000002c,
+	0x000001e0,
+	0x000001e4,
+	-1,
+};
+
+static const u32 gen6_null_state_batch[] = {
+	0x69040000,
+	0x790d0001,
+	0x00000000,
+	0x00000000,
+	0x78180000,
+	0x00000001,
+	0x61010008,
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000001,
+	0x00000000,
+	0x00000001,
+	0x61020000,
+	0x00000000,
+	0x78050001,
+	0x00000018,
+	0x00000000,
+	0x780d1002,
+	0x00000000,
+	0x00000000,
+	0x00000420,
+	0x78150003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78100004,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78160003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78110005,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78120002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78170003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79050005,
+	0xe0040000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79100000,
+	0x00000000,
+	0x79000002,
+	0xffffffff,
+	0x00000000,
+	0x00000000,
+	0x780e0002,
+	0x00000441,
+	0x00000401,
+	0x00000401,
+	0x78021002,
+	0x00000000,
+	0x00000000,
+	0x00000400,
+	0x78130012,
+	0x00400810,
+	0x00000000,
+	0x20000000,
+	0x04000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78140007,
+	0x00000280,
+	0x08080000,
+	0x00000000,
+	0x00060000,
+	0x4e080002,
+	0x00100400,
+	0x00000000,
+	0x00000000,
+	0x78090005,
+	0x02000000,
+	0x22220000,
+	0x02f60000,
+	0x11330000,
+	0x02850004,
+	0x11220000,
+	0x78011002,
+	0x00000000,
+	0x00000000,
+	0x00000200,
+	0x78080003,
+	0x00002000,
+	0x00000448,	 /* reloc */
+	0x00000448,	 /* reloc */
+	0x00000000,
+	0x05000000,	 /* cmds end */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000220,	 /* state start */
+	0x00000240,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0060005a,
+	0x204077be,
+	0x000000c0,
+	0x008d0040,
+	0x0060005a,
+	0x206077be,
+	0x000000c0,
+	0x008d0080,
+	0x0060005a,
+	0x208077be,
+	0x000000d0,
+	0x008d0040,
+	0x0060005a,
+	0x20a077be,
+	0x000000d0,
+	0x008d0080,
+	0x00000201,
+	0x20080061,
+	0x00000000,
+	0x00000000,
+	0x00600001,
+	0x20200022,
+	0x008d0000,
+	0x00000000,
+	0x02800031,
+	0x21c01cc9,
+	0x00000020,
+	0x0a8a0001,
+	0x00600001,
+	0x204003be,
+	0x008d01c0,
+	0x00000000,
+	0x00600001,
+	0x206003be,
+	0x008d01e0,
+	0x00000000,
+	0x00600001,
+	0x208003be,
+	0x008d0200,
+	0x00000000,
+	0x00600001,
+	0x20a003be,
+	0x008d0220,
+	0x00000000,
+	0x00600001,
+	0x20c003be,
+	0x008d0240,
+	0x00000000,
+	0x00600001,
+	0x20e003be,
+	0x008d0260,
+	0x00000000,
+	0x00600001,
+	0x210003be,
+	0x008d0280,
+	0x00000000,
+	0x00600001,
+	0x212003be,
+	0x008d02a0,
+	0x00000000,
+	0x05800031,
+	0x24001cc8,
+	0x00000040,
+	0x90019000,
+	0x0000007e,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0000007e,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0000007e,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0000007e,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0000007e,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0000007e,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0000007e,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0000007e,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x30000000,
+	0x00000124,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0xf99a130c,
+	0x799a130c,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x80000031,
+	0x00000003,
+	0x00000000,	 /* state end */
+};
+
+RO_RENDERSTATE(6);
diff --git a/drivers/gpu/drm/i915/gt/gen7_renderclear.c b/drivers/gpu/drm/i915/gt/gen7_renderclear.c
new file mode 100644
index 000000000..317efb145
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen7_renderclear.c
@@ -0,0 +1,453 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "gen7_renderclear.h"
+#include "i915_drv.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt_regs.h"
+
+#define GT3_INLINE_DATA_DELAYS 0x1E00
+#define batch_advance(Y, CS) GEM_BUG_ON((Y)->end != (CS))
+
+struct cb_kernel {
+	const void *data;
+	u32 size;
+};
+
+#define CB_KERNEL(name) { .data = (name), .size = sizeof(name) }
+
+#include "ivb_clear_kernel.c"
+static const struct cb_kernel cb_kernel_ivb = CB_KERNEL(ivb_clear_kernel);
+
+#include "hsw_clear_kernel.c"
+static const struct cb_kernel cb_kernel_hsw = CB_KERNEL(hsw_clear_kernel);
+
+struct batch_chunk {
+	struct i915_vma *vma;
+	u32 offset;
+	u32 *start;
+	u32 *end;
+	u32 max_items;
+};
+
+struct batch_vals {
+	u32 max_threads;
+	u32 state_start;
+	u32 surface_start;
+	u32 surface_height;
+	u32 surface_width;
+	u32 size;
+};
+
+static int num_primitives(const struct batch_vals *bv)
+{
+	/*
+	 * We need to saturate the GPU with work in order to dispatch
+	 * a shader on every HW thread, and clear the thread-local registers.
+	 * In short, we have to dispatch work faster than the shaders can
+	 * run in order to fill the EU and occupy each HW thread.
+	 */
+	return bv->max_threads;
+}
+
+static void
+batch_get_defaults(struct drm_i915_private *i915, struct batch_vals *bv)
+{
+	if (IS_HASWELL(i915)) {
+		switch (INTEL_INFO(i915)->gt) {
+		default:
+		case 1:
+			bv->max_threads = 70;
+			break;
+		case 2:
+			bv->max_threads = 140;
+			break;
+		case 3:
+			bv->max_threads = 280;
+			break;
+		}
+		bv->surface_height = 16 * 16;
+		bv->surface_width = 32 * 2 * 16;
+	} else {
+		switch (INTEL_INFO(i915)->gt) {
+		default:
+		case 1: /* including vlv */
+			bv->max_threads = 36;
+			break;
+		case 2:
+			bv->max_threads = 128;
+			break;
+		}
+		bv->surface_height = 16 * 8;
+		bv->surface_width = 32 * 16;
+	}
+	bv->state_start = round_up(SZ_1K + num_primitives(bv) * 64, SZ_4K);
+	bv->surface_start = bv->state_start + SZ_4K;
+	bv->size = bv->surface_start + bv->surface_height * bv->surface_width;
+}
+
+static void batch_init(struct batch_chunk *bc,
+		       struct i915_vma *vma,
+		       u32 *start, u32 offset, u32 max_bytes)
+{
+	bc->vma = vma;
+	bc->offset = offset;
+	bc->start = start + bc->offset / sizeof(*bc->start);
+	bc->end = bc->start;
+	bc->max_items = max_bytes / sizeof(*bc->start);
+}
+
+static u32 batch_offset(const struct batch_chunk *bc, u32 *cs)
+{
+	return (cs - bc->start) * sizeof(*bc->start) + bc->offset;
+}
+
+static u32 batch_addr(const struct batch_chunk *bc)
+{
+	return bc->vma->node.start;
+}
+
+static void batch_add(struct batch_chunk *bc, const u32 d)
+{
+	GEM_BUG_ON((bc->end - bc->start) >= bc->max_items);
+	*bc->end++ = d;
+}
+
+static u32 *batch_alloc_items(struct batch_chunk *bc, u32 align, u32 items)
+{
+	u32 *map;
+
+	if (align) {
+		u32 *end = PTR_ALIGN(bc->end, align);
+
+		memset32(bc->end, 0, end - bc->end);
+		bc->end = end;
+	}
+
+	map = bc->end;
+	bc->end += items;
+
+	return map;
+}
+
+static u32 *batch_alloc_bytes(struct batch_chunk *bc, u32 align, u32 bytes)
+{
+	GEM_BUG_ON(!IS_ALIGNED(bytes, sizeof(*bc->start)));
+	return batch_alloc_items(bc, align, bytes / sizeof(*bc->start));
+}
+
+static u32
+gen7_fill_surface_state(struct batch_chunk *state,
+			const u32 dst_offset,
+			const struct batch_vals *bv)
+{
+	u32 surface_h = bv->surface_height;
+	u32 surface_w = bv->surface_width;
+	u32 *cs = batch_alloc_items(state, 32, 8);
+	u32 offset = batch_offset(state, cs);
+
+#define SURFACE_2D 1
+#define SURFACEFORMAT_B8G8R8A8_UNORM 0x0C0
+#define RENDER_CACHE_READ_WRITE 1
+
+	*cs++ = SURFACE_2D << 29 |
+		(SURFACEFORMAT_B8G8R8A8_UNORM << 18) |
+		(RENDER_CACHE_READ_WRITE << 8);
+
+	*cs++ = batch_addr(state) + dst_offset;
+
+	*cs++ = ((surface_h / 4 - 1) << 16) | (surface_w / 4 - 1);
+	*cs++ = surface_w;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+#define SHADER_CHANNELS(r, g, b, a) \
+	(((r) << 25) | ((g) << 22) | ((b) << 19) | ((a) << 16))
+	*cs++ = SHADER_CHANNELS(4, 5, 6, 7);
+	batch_advance(state, cs);
+
+	return offset;
+}
+
+static u32
+gen7_fill_binding_table(struct batch_chunk *state,
+			const struct batch_vals *bv)
+{
+	u32 surface_start =
+		gen7_fill_surface_state(state, bv->surface_start, bv);
+	u32 *cs = batch_alloc_items(state, 32, 8);
+	u32 offset = batch_offset(state, cs);
+
+	*cs++ = surface_start - state->offset;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+	batch_advance(state, cs);
+
+	return offset;
+}
+
+static u32
+gen7_fill_kernel_data(struct batch_chunk *state,
+		      const u32 *data,
+		      const u32 size)
+{
+	return batch_offset(state,
+			    memcpy(batch_alloc_bytes(state, 64, size),
+				   data, size));
+}
+
+static u32
+gen7_fill_interface_descriptor(struct batch_chunk *state,
+			       const struct batch_vals *bv,
+			       const struct cb_kernel *kernel,
+			       unsigned int count)
+{
+	u32 kernel_offset =
+		gen7_fill_kernel_data(state, kernel->data, kernel->size);
+	u32 binding_table = gen7_fill_binding_table(state, bv);
+	u32 *cs = batch_alloc_items(state, 32, 8 * count);
+	u32 offset = batch_offset(state, cs);
+
+	*cs++ = kernel_offset;
+	*cs++ = (1 << 7) | (1 << 13);
+	*cs++ = 0;
+	*cs++ = (binding_table - state->offset) | 1;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+
+	/* 1 - 63dummy idds */
+	memset32(cs, 0x00, (count - 1) * 8);
+	batch_advance(state, cs + (count - 1) * 8);
+
+	return offset;
+}
+
+static void
+gen7_emit_state_base_address(struct batch_chunk *batch,
+			     u32 surface_state_base)
+{
+	u32 *cs = batch_alloc_items(batch, 0, 10);
+
+	*cs++ = STATE_BASE_ADDRESS | (10 - 2);
+	/* general */
+	*cs++ = batch_addr(batch) | BASE_ADDRESS_MODIFY;
+	/* surface */
+	*cs++ = (batch_addr(batch) + surface_state_base) | BASE_ADDRESS_MODIFY;
+	/* dynamic */
+	*cs++ = batch_addr(batch) | BASE_ADDRESS_MODIFY;
+	/* indirect */
+	*cs++ = batch_addr(batch) | BASE_ADDRESS_MODIFY;
+	/* instruction */
+	*cs++ = batch_addr(batch) | BASE_ADDRESS_MODIFY;
+
+	/* general/dynamic/indirect/instruction access Bound */
+	*cs++ = 0;
+	*cs++ = BASE_ADDRESS_MODIFY;
+	*cs++ = 0;
+	*cs++ = BASE_ADDRESS_MODIFY;
+	batch_advance(batch, cs);
+}
+
+static void
+gen7_emit_vfe_state(struct batch_chunk *batch,
+		    const struct batch_vals *bv,
+		    u32 urb_size, u32 curbe_size,
+		    u32 mode)
+{
+	u32 threads = bv->max_threads - 1;
+	u32 *cs = batch_alloc_items(batch, 32, 8);
+
+	*cs++ = MEDIA_VFE_STATE | (8 - 2);
+
+	/* scratch buffer */
+	*cs++ = 0;
+
+	/* number of threads & urb entries for GPGPU vs Media Mode */
+	*cs++ = threads << 16 | 1 << 8 | mode << 2;
+
+	*cs++ = 0;
+
+	/* urb entry size & curbe size in 256 bits unit */
+	*cs++ = urb_size << 16 | curbe_size;
+
+	/* scoreboard */
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+	batch_advance(batch, cs);
+}
+
+static void
+gen7_emit_interface_descriptor_load(struct batch_chunk *batch,
+				    const u32 interface_descriptor,
+				    unsigned int count)
+{
+	u32 *cs = batch_alloc_items(batch, 8, 4);
+
+	*cs++ = MEDIA_INTERFACE_DESCRIPTOR_LOAD | (4 - 2);
+	*cs++ = 0;
+	*cs++ = count * 8 * sizeof(*cs);
+
+	/*
+	 * interface descriptor address - it is relative to the dynamics base
+	 * address
+	 */
+	*cs++ = interface_descriptor;
+	batch_advance(batch, cs);
+}
+
+static void
+gen7_emit_media_object(struct batch_chunk *batch,
+		       unsigned int media_object_index)
+{
+	unsigned int x_offset = (media_object_index % 16) * 64;
+	unsigned int y_offset = (media_object_index / 16) * 16;
+	unsigned int pkt = 6 + 3;
+	u32 *cs;
+
+	cs = batch_alloc_items(batch, 8, pkt);
+
+	*cs++ = MEDIA_OBJECT | (pkt - 2);
+
+	/* interface descriptor offset */
+	*cs++ = 0;
+
+	/* without indirect data */
+	*cs++ = 0;
+	*cs++ = 0;
+
+	/* scoreboard */
+	*cs++ = 0;
+	*cs++ = 0;
+
+	/* inline */
+	*cs++ = y_offset << 16 | x_offset;
+	*cs++ = 0;
+	*cs++ = GT3_INLINE_DATA_DELAYS;
+
+	batch_advance(batch, cs);
+}
+
+static void gen7_emit_pipeline_flush(struct batch_chunk *batch)
+{
+	u32 *cs = batch_alloc_items(batch, 0, 4);
+
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
+		PIPE_CONTROL_DEPTH_CACHE_FLUSH |
+		PIPE_CONTROL_DC_FLUSH_ENABLE |
+		PIPE_CONTROL_CS_STALL;
+	*cs++ = 0;
+	*cs++ = 0;
+
+	batch_advance(batch, cs);
+}
+
+static void gen7_emit_pipeline_invalidate(struct batch_chunk *batch)
+{
+	u32 *cs = batch_alloc_items(batch, 0, 10);
+
+	/* ivb: Stall before STATE_CACHE_INVALIDATE */
+	*cs++ = GFX_OP_PIPE_CONTROL(5);
+	*cs++ = PIPE_CONTROL_STALL_AT_SCOREBOARD |
+		PIPE_CONTROL_CS_STALL;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+
+	*cs++ = GFX_OP_PIPE_CONTROL(5);
+	*cs++ = PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+
+	batch_advance(batch, cs);
+}
+
+static void emit_batch(struct i915_vma * const vma,
+		       u32 *start,
+		       const struct batch_vals *bv)
+{
+	struct drm_i915_private *i915 = vma->vm->i915;
+	const unsigned int desc_count = 1;
+	const unsigned int urb_size = 1;
+	struct batch_chunk cmds, state;
+	u32 descriptors;
+	unsigned int i;
+
+	batch_init(&cmds, vma, start, 0, bv->state_start);
+	batch_init(&state, vma, start, bv->state_start, SZ_4K);
+
+	descriptors = gen7_fill_interface_descriptor(&state, bv,
+						     IS_HASWELL(i915) ?
+						     &cb_kernel_hsw :
+						     &cb_kernel_ivb,
+						     desc_count);
+
+	/* Reset inherited context registers */
+	gen7_emit_pipeline_flush(&cmds);
+	gen7_emit_pipeline_invalidate(&cmds);
+	batch_add(&cmds, MI_LOAD_REGISTER_IMM(2));
+	batch_add(&cmds, i915_mmio_reg_offset(CACHE_MODE_0_GEN7));
+	batch_add(&cmds, 0xffff0000 |
+			((IS_IVB_GT1(i915) || IS_VALLEYVIEW(i915)) ?
+			 HIZ_RAW_STALL_OPT_DISABLE :
+			 0));
+	batch_add(&cmds, i915_mmio_reg_offset(CACHE_MODE_1));
+	batch_add(&cmds, 0xffff0000 | PIXEL_SUBSPAN_COLLECT_OPT_DISABLE);
+	gen7_emit_pipeline_invalidate(&cmds);
+	gen7_emit_pipeline_flush(&cmds);
+
+	/* Switch to the media pipeline and our base address */
+	gen7_emit_pipeline_invalidate(&cmds);
+	batch_add(&cmds, PIPELINE_SELECT | PIPELINE_SELECT_MEDIA);
+	batch_add(&cmds, MI_NOOP);
+	gen7_emit_pipeline_invalidate(&cmds);
+
+	gen7_emit_pipeline_flush(&cmds);
+	gen7_emit_state_base_address(&cmds, descriptors);
+	gen7_emit_pipeline_invalidate(&cmds);
+
+	/* Set the clear-residual kernel state */
+	gen7_emit_vfe_state(&cmds, bv, urb_size - 1, 0, 0);
+	gen7_emit_interface_descriptor_load(&cmds, descriptors, desc_count);
+
+	/* Execute the kernel on all HW threads */
+	for (i = 0; i < num_primitives(bv); i++)
+		gen7_emit_media_object(&cmds, i);
+
+	batch_add(&cmds, MI_BATCH_BUFFER_END);
+}
+
+int gen7_setup_clear_gpr_bb(struct intel_engine_cs * const engine,
+			    struct i915_vma * const vma)
+{
+	struct batch_vals bv;
+	u32 *batch;
+
+	batch_get_defaults(engine->i915, &bv);
+	if (!vma)
+		return bv.size;
+
+	GEM_BUG_ON(vma->obj->base.size < bv.size);
+
+	batch = i915_gem_object_pin_map(vma->obj, I915_MAP_WC);
+	if (IS_ERR(batch))
+		return PTR_ERR(batch);
+
+	emit_batch(vma, memset(batch, 0, bv.size), &bv);
+
+	i915_gem_object_flush_map(vma->obj);
+	__i915_gem_object_release_map(vma->obj);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gt/gen7_renderclear.h b/drivers/gpu/drm/i915/gt/gen7_renderclear.h
new file mode 100644
index 000000000..bb100748e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen7_renderclear.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __GEN7_RENDERCLEAR_H__
+#define __GEN7_RENDERCLEAR_H__
+
+struct intel_engine_cs;
+struct i915_vma;
+
+int gen7_setup_clear_gpr_bb(struct intel_engine_cs * const engine,
+			    struct i915_vma * const vma);
+
+#endif /* __GEN7_RENDERCLEAR_H__ */
diff --git a/drivers/gpu/drm/i915/gt/gen7_renderstate.c b/drivers/gpu/drm/i915/gt/gen7_renderstate.c
new file mode 100644
index 000000000..c36e84d30
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen7_renderstate.c
@@ -0,0 +1,261 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Generated by: intel-gpu-tools-1.8-220-g01153e7
+ */
+
+#include "intel_renderstate.h"
+
+static const u32 gen7_null_state_relocs[] = {
+	0x0000000c,
+	0x00000010,
+	0x00000018,
+	0x000001ec,
+	-1,
+};
+
+static const u32 gen7_null_state_batch[] = {
+	0x69040000,
+	0x61010008,
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000001,
+	0x00000000,
+	0x00000001,
+	0x790d0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78180000,
+	0x00000001,
+	0x79160000,
+	0x00000008,
+	0x78300000,
+	0x02010040,
+	0x78310000,
+	0x04000000,
+	0x78320000,
+	0x04000000,
+	0x78330000,
+	0x02000000,
+	0x78100004,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781b0005,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781c0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781d0004,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78110005,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78120002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78210000,
+	0x00000000,
+	0x78130005,
+	0x00000000,
+	0x20000000,
+	0x04000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78140001,
+	0x20000800,
+	0x00000000,
+	0x781e0001,
+	0x00000000,
+	0x00000000,
+	0x78050005,
+	0xe0040000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78040001,
+	0x00000000,
+	0x00000000,
+	0x78240000,
+	0x00000240,
+	0x78230000,
+	0x00000260,
+	0x782f0000,
+	0x00000280,
+	0x781f000c,
+	0x00400810,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78200006,
+	0x000002c0,
+	0x08080000,
+	0x00000000,
+	0x28000402,
+	0x00060000,
+	0x00000000,
+	0x00000000,
+	0x78090005,
+	0x02000000,
+	0x22220000,
+	0x02f60000,
+	0x11230000,
+	0x02f60004,
+	0x11230000,
+	0x78080003,
+	0x00006008,
+	0x00000340,	 /* reloc */
+	0xffffffff,
+	0x00000000,
+	0x782a0000,
+	0x00000360,
+	0x79000002,
+	0xffffffff,
+	0x00000000,
+	0x00000000,
+	0x7b000005,
+	0x0000000f,
+	0x00000003,
+	0x00000000,
+	0x00000001,
+	0x00000000,
+	0x00000000,
+	0x05000000,	 /* cmds end */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000031,	 /* state start */
+	0x00000003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0xf99a130c,
+	0x799a130c,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000492,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0080005a,
+	0x2e2077bd,
+	0x000000c0,
+	0x008d0040,
+	0x0080005a,
+	0x2e6077bd,
+	0x000000d0,
+	0x008d0040,
+	0x02800031,
+	0x21801fa9,
+	0x008d0e20,
+	0x08840001,
+	0x00800001,
+	0x2e2003bd,
+	0x008d0180,
+	0x00000000,
+	0x00800001,
+	0x2e6003bd,
+	0x008d01c0,
+	0x00000000,
+	0x00800001,
+	0x2ea003bd,
+	0x008d0200,
+	0x00000000,
+	0x00800001,
+	0x2ee003bd,
+	0x008d0240,
+	0x00000000,
+	0x05800031,
+	0x20001fa8,
+	0x008d0e20,
+	0x90031000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000380,
+	0x000003a0,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,	 /* state end */
+};
+
+RO_RENDERSTATE(7);
diff --git a/drivers/gpu/drm/i915/gt/gen8_engine_cs.c b/drivers/gpu/drm/i915/gt/gen8_engine_cs.c
new file mode 100644
index 000000000..efc22f9b1
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen8_engine_cs.c
@@ -0,0 +1,781 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#include "gen8_engine_cs.h"
+#include "i915_drv.h"
+#include "intel_engine_regs.h"
+#include "intel_gpu_commands.h"
+#include "intel_lrc.h"
+#include "intel_ring.h"
+
+int gen8_emit_flush_rcs(struct i915_request *rq, u32 mode)
+{
+	bool vf_flush_wa = false, dc_flush_wa = false;
+	u32 *cs, flags = 0;
+	int len;
+
+	flags |= PIPE_CONTROL_CS_STALL;
+
+	if (mode & EMIT_FLUSH) {
+		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
+		flags |= PIPE_CONTROL_FLUSH_ENABLE;
+	}
+
+	if (mode & EMIT_INVALIDATE) {
+		flags |= PIPE_CONTROL_TLB_INVALIDATE;
+		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_QW_WRITE;
+		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
+
+		/*
+		 * On GEN9: before VF_CACHE_INVALIDATE we need to emit a NULL
+		 * pipe control.
+		 */
+		if (GRAPHICS_VER(rq->engine->i915) == 9)
+			vf_flush_wa = true;
+
+		/* WaForGAMHang:kbl */
+		if (IS_KBL_GRAPHICS_STEP(rq->engine->i915, 0, STEP_C0))
+			dc_flush_wa = true;
+	}
+
+	len = 6;
+
+	if (vf_flush_wa)
+		len += 6;
+
+	if (dc_flush_wa)
+		len += 12;
+
+	cs = intel_ring_begin(rq, len);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	if (vf_flush_wa)
+		cs = gen8_emit_pipe_control(cs, 0, 0);
+
+	if (dc_flush_wa)
+		cs = gen8_emit_pipe_control(cs, PIPE_CONTROL_DC_FLUSH_ENABLE,
+					    0);
+
+	cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
+
+	if (dc_flush_wa)
+		cs = gen8_emit_pipe_control(cs, PIPE_CONTROL_CS_STALL, 0);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen8_emit_flush_xcs(struct i915_request *rq, u32 mode)
+{
+	u32 cmd, *cs;
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	cmd = MI_FLUSH_DW + 1;
+
+	/*
+	 * We always require a command barrier so that subsequent
+	 * commands, such as breadcrumb interrupts, are strictly ordered
+	 * wrt the contents of the write cache being flushed to memory
+	 * (and thus being coherent from the CPU).
+	 */
+	cmd |= MI_FLUSH_DW_STORE_INDEX | MI_FLUSH_DW_OP_STOREDW;
+
+	if (mode & EMIT_INVALIDATE) {
+		cmd |= MI_INVALIDATE_TLB;
+		if (rq->engine->class == VIDEO_DECODE_CLASS)
+			cmd |= MI_INVALIDATE_BSD;
+	}
+
+	*cs++ = cmd;
+	*cs++ = LRC_PPHWSP_SCRATCH_ADDR;
+	*cs++ = 0; /* upper addr */
+	*cs++ = 0; /* value */
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen11_emit_flush_rcs(struct i915_request *rq, u32 mode)
+{
+	if (mode & EMIT_FLUSH) {
+		u32 *cs;
+		u32 flags = 0;
+
+		flags |= PIPE_CONTROL_CS_STALL;
+
+		flags |= PIPE_CONTROL_TILE_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
+		flags |= PIPE_CONTROL_FLUSH_ENABLE;
+		flags |= PIPE_CONTROL_QW_WRITE;
+		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
+
+		cs = intel_ring_begin(rq, 6);
+		if (IS_ERR(cs))
+			return PTR_ERR(cs);
+
+		cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
+		intel_ring_advance(rq, cs);
+	}
+
+	if (mode & EMIT_INVALIDATE) {
+		u32 *cs;
+		u32 flags = 0;
+
+		flags |= PIPE_CONTROL_CS_STALL;
+
+		flags |= PIPE_CONTROL_COMMAND_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TLB_INVALIDATE;
+		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_QW_WRITE;
+		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
+
+		cs = intel_ring_begin(rq, 6);
+		if (IS_ERR(cs))
+			return PTR_ERR(cs);
+
+		cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
+		intel_ring_advance(rq, cs);
+	}
+
+	return 0;
+}
+
+static u32 preparser_disable(bool state)
+{
+	return MI_ARB_CHECK | 1 << 8 | state;
+}
+
+static i915_reg_t gen12_get_aux_inv_reg(struct intel_engine_cs *engine)
+{
+	switch (engine->id) {
+	case RCS0:
+		return GEN12_CCS_AUX_INV;
+	case BCS0:
+		return GEN12_BCS0_AUX_INV;
+	case VCS0:
+		return GEN12_VD0_AUX_INV;
+	case VCS2:
+		return GEN12_VD2_AUX_INV;
+	case VECS0:
+		return GEN12_VE0_AUX_INV;
+	case CCS0:
+		return GEN12_CCS0_AUX_INV;
+	default:
+		return INVALID_MMIO_REG;
+	}
+}
+
+static bool gen12_needs_ccs_aux_inv(struct intel_engine_cs *engine)
+{
+	i915_reg_t reg = gen12_get_aux_inv_reg(engine);
+
+	if (IS_PONTEVECCHIO(engine->i915))
+		return false;
+
+	/*
+	 * So far platforms supported by i915 having flat ccs do not require
+	 * AUX invalidation. Check also whether the engine requires it.
+	 */
+	return i915_mmio_reg_valid(reg) && !HAS_FLAT_CCS(engine->i915);
+}
+
+u32 *gen12_emit_aux_table_inv(struct intel_engine_cs *engine, u32 *cs)
+{
+	i915_reg_t inv_reg = gen12_get_aux_inv_reg(engine);
+	u32 gsi_offset = engine->gt->uncore->gsi_offset;
+
+	if (!gen12_needs_ccs_aux_inv(engine))
+		return cs;
+
+	*cs++ = MI_LOAD_REGISTER_IMM(1) | MI_LRI_MMIO_REMAP_EN;
+	*cs++ = i915_mmio_reg_offset(inv_reg) + gsi_offset;
+	*cs++ = AUX_INV;
+
+	*cs++ = MI_SEMAPHORE_WAIT_TOKEN |
+		MI_SEMAPHORE_REGISTER_POLL |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_EQ_SDD;
+	*cs++ = 0;
+	*cs++ = i915_mmio_reg_offset(inv_reg) + gsi_offset;
+	*cs++ = 0;
+	*cs++ = 0;
+
+	return cs;
+}
+
+int gen12_emit_flush_rcs(struct i915_request *rq, u32 mode)
+{
+	struct intel_engine_cs *engine = rq->engine;
+
+	/*
+	 * On Aux CCS platforms the invalidation of the Aux
+	 * table requires quiescing memory traffic beforehand
+	 */
+	if (mode & EMIT_FLUSH || gen12_needs_ccs_aux_inv(engine)) {
+		u32 flags = 0;
+		u32 *cs;
+
+		/*
+		 * L3 fabric flush is needed for AUX CCS invalidation
+		 * which happens as part of pipe-control so we can
+		 * ignore PIPE_CONTROL_FLUSH_L3. Also PIPE_CONTROL_FLUSH_L3
+		 * deals with Protected Memory which is not needed for
+		 * AUX CCS invalidation and lead to unwanted side effects.
+		 */
+		if (mode & EMIT_FLUSH)
+			flags |= PIPE_CONTROL_FLUSH_L3;
+
+		flags |= PIPE_CONTROL_TILE_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
+		flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
+		/* Wa_1409600907:tgl,adl-p */
+		flags |= PIPE_CONTROL_DEPTH_STALL;
+		flags |= PIPE_CONTROL_DC_FLUSH_ENABLE;
+		flags |= PIPE_CONTROL_FLUSH_ENABLE;
+
+		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
+		flags |= PIPE_CONTROL_QW_WRITE;
+
+		flags |= PIPE_CONTROL_CS_STALL;
+
+		if (!HAS_3D_PIPELINE(engine->i915))
+			flags &= ~PIPE_CONTROL_3D_ARCH_FLAGS;
+		else if (engine->class == COMPUTE_CLASS)
+			flags &= ~PIPE_CONTROL_3D_ENGINE_FLAGS;
+
+		cs = intel_ring_begin(rq, 6);
+		if (IS_ERR(cs))
+			return PTR_ERR(cs);
+
+		cs = gen12_emit_pipe_control(cs,
+					     PIPE_CONTROL0_HDC_PIPELINE_FLUSH,
+					     flags, LRC_PPHWSP_SCRATCH_ADDR);
+		intel_ring_advance(rq, cs);
+	}
+
+	if (mode & EMIT_INVALIDATE) {
+		u32 flags = 0;
+		u32 *cs, count;
+
+		flags |= PIPE_CONTROL_COMMAND_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TLB_INVALIDATE;
+		flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
+		flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
+
+		flags |= PIPE_CONTROL_STORE_DATA_INDEX;
+		flags |= PIPE_CONTROL_QW_WRITE;
+
+		flags |= PIPE_CONTROL_CS_STALL;
+
+		if (!HAS_3D_PIPELINE(engine->i915))
+			flags &= ~PIPE_CONTROL_3D_ARCH_FLAGS;
+		else if (engine->class == COMPUTE_CLASS)
+			flags &= ~PIPE_CONTROL_3D_ENGINE_FLAGS;
+
+		count = 8;
+		if (gen12_needs_ccs_aux_inv(rq->engine))
+			count += 8;
+
+		cs = intel_ring_begin(rq, count);
+		if (IS_ERR(cs))
+			return PTR_ERR(cs);
+
+		/*
+		 * Prevent the pre-parser from skipping past the TLB
+		 * invalidate and loading a stale page for the batch
+		 * buffer / request payload.
+		 */
+		*cs++ = preparser_disable(true);
+
+		cs = gen8_emit_pipe_control(cs, flags, LRC_PPHWSP_SCRATCH_ADDR);
+
+		cs = gen12_emit_aux_table_inv(engine, cs);
+
+		*cs++ = preparser_disable(false);
+		intel_ring_advance(rq, cs);
+	}
+
+	return 0;
+}
+
+int gen12_emit_flush_xcs(struct i915_request *rq, u32 mode)
+{
+	u32 cmd = 4;
+	u32 *cs;
+
+	if (mode & EMIT_INVALIDATE) {
+		cmd += 2;
+
+		if (gen12_needs_ccs_aux_inv(rq->engine))
+			cmd += 8;
+	}
+
+	cs = intel_ring_begin(rq, cmd);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	if (mode & EMIT_INVALIDATE)
+		*cs++ = preparser_disable(true);
+
+	cmd = MI_FLUSH_DW + 1;
+
+	/*
+	 * We always require a command barrier so that subsequent
+	 * commands, such as breadcrumb interrupts, are strictly ordered
+	 * wrt the contents of the write cache being flushed to memory
+	 * (and thus being coherent from the CPU).
+	 */
+	cmd |= MI_FLUSH_DW_STORE_INDEX | MI_FLUSH_DW_OP_STOREDW;
+
+	if (mode & EMIT_INVALIDATE) {
+		cmd |= MI_INVALIDATE_TLB;
+		if (rq->engine->class == VIDEO_DECODE_CLASS)
+			cmd |= MI_INVALIDATE_BSD;
+	}
+
+	*cs++ = cmd;
+	*cs++ = LRC_PPHWSP_SCRATCH_ADDR;
+	*cs++ = 0; /* upper addr */
+	*cs++ = 0; /* value */
+
+	cs = gen12_emit_aux_table_inv(rq->engine, cs);
+
+	if (mode & EMIT_INVALIDATE)
+		*cs++ = preparser_disable(false);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static u32 preempt_address(struct intel_engine_cs *engine)
+{
+	return (i915_ggtt_offset(engine->status_page.vma) +
+		I915_GEM_HWS_PREEMPT_ADDR);
+}
+
+static u32 hwsp_offset(const struct i915_request *rq)
+{
+	const struct intel_timeline *tl;
+
+	/* Before the request is executed, the timeline is fixed */
+	tl = rcu_dereference_protected(rq->timeline,
+				       !i915_request_signaled(rq));
+
+	/* See the comment in i915_request_active_seqno(). */
+	return page_mask_bits(tl->hwsp_offset) + offset_in_page(rq->hwsp_seqno);
+}
+
+int gen8_emit_init_breadcrumb(struct i915_request *rq)
+{
+	u32 *cs;
+
+	GEM_BUG_ON(i915_request_has_initial_breadcrumb(rq));
+	if (!i915_request_timeline(rq)->has_initial_breadcrumb)
+		return 0;
+
+	cs = intel_ring_begin(rq, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = hwsp_offset(rq);
+	*cs++ = 0;
+	*cs++ = rq->fence.seqno - 1;
+
+	/*
+	 * Check if we have been preempted before we even get started.
+	 *
+	 * After this point i915_request_started() reports true, even if
+	 * we get preempted and so are no longer running.
+	 *
+	 * i915_request_started() is used during preemption processing
+	 * to decide if the request is currently inside the user payload
+	 * or spinning on a kernel semaphore (or earlier). For no-preemption
+	 * requests, we do allow preemption on the semaphore before the user
+	 * payload, but do not allow preemption once the request is started.
+	 *
+	 * i915_request_started() is similarly used during GPU hangs to
+	 * determine if the user's payload was guilty, and if so, the
+	 * request is banned. Before the request is started, it is assumed
+	 * to be unharmed and an innocent victim of another's hang.
+	 */
+	*cs++ = MI_NOOP;
+	*cs++ = MI_ARB_CHECK;
+
+	intel_ring_advance(rq, cs);
+
+	/* Record the updated position of the request's payload */
+	rq->infix = intel_ring_offset(rq, cs);
+
+	__set_bit(I915_FENCE_FLAG_INITIAL_BREADCRUMB, &rq->fence.flags);
+
+	return 0;
+}
+
+static int __gen125_emit_bb_start(struct i915_request *rq,
+				  u64 offset, u32 len,
+				  const unsigned int flags,
+				  u32 arb)
+{
+	struct intel_context *ce = rq->context;
+	u32 wa_offset = lrc_indirect_bb(ce);
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 12);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_ARB_ON_OFF | arb;
+
+	*cs++ = MI_LOAD_REGISTER_MEM_GEN8 |
+		MI_SRM_LRM_GLOBAL_GTT |
+		MI_LRI_LRM_CS_MMIO;
+	*cs++ = i915_mmio_reg_offset(RING_PREDICATE_RESULT(0));
+	*cs++ = wa_offset + DG2_PREDICATE_RESULT_WA;
+	*cs++ = 0;
+
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 |
+		(flags & I915_DISPATCH_SECURE ? 0 : BIT(8));
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+
+	/* Fixup stray MI_SET_PREDICATE as it prevents us executing the ring */
+	*cs++ = MI_BATCH_BUFFER_START_GEN8;
+	*cs++ = wa_offset + DG2_PREDICATE_RESULT_BB;
+	*cs++ = 0;
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen125_emit_bb_start_noarb(struct i915_request *rq,
+			       u64 offset, u32 len,
+			       const unsigned int flags)
+{
+	return __gen125_emit_bb_start(rq, offset, len, flags, MI_ARB_DISABLE);
+}
+
+int gen125_emit_bb_start(struct i915_request *rq,
+			 u64 offset, u32 len,
+			 const unsigned int flags)
+{
+	return __gen125_emit_bb_start(rq, offset, len, flags, MI_ARB_ENABLE);
+}
+
+int gen8_emit_bb_start_noarb(struct i915_request *rq,
+			     u64 offset, u32 len,
+			     const unsigned int flags)
+{
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/*
+	 * WaDisableCtxRestoreArbitration:bdw,chv
+	 *
+	 * We don't need to perform MI_ARB_ENABLE as often as we do (in
+	 * particular all the gen that do not need the w/a at all!), if we
+	 * took care to make sure that on every switch into this context
+	 * (both ordinary and for preemption) that arbitrartion was enabled
+	 * we would be fine.  However, for gen8 there is another w/a that
+	 * requires us to not preempt inside GPGPU execution, so we keep
+	 * arbitration disabled for gen8 batches. Arbitration will be
+	 * re-enabled before we close the request
+	 * (engine->emit_fini_breadcrumb).
+	 */
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
+	/* FIXME(BDW+): Address space and security selectors. */
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 |
+		(flags & I915_DISPATCH_SECURE ? 0 : BIT(8));
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+int gen8_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 len,
+		       const unsigned int flags)
+{
+	u32 *cs;
+
+	if (unlikely(i915_request_has_nopreempt(rq)))
+		return gen8_emit_bb_start_noarb(rq, offset, len, flags);
+
+	cs = intel_ring_begin(rq, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 |
+		(flags & I915_DISPATCH_SECURE ? 0 : BIT(8));
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+	*cs++ = MI_NOOP;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static void assert_request_valid(struct i915_request *rq)
+{
+	struct intel_ring *ring __maybe_unused = rq->ring;
+
+	/* Can we unwind this request without appearing to go forwards? */
+	GEM_BUG_ON(intel_ring_direction(ring, rq->wa_tail, rq->head) <= 0);
+}
+
+/*
+ * Reserve space for 2 NOOPs at the end of each request to be
+ * used as a workaround for not being allowed to do lite
+ * restore with HEAD==TAIL (WaIdleLiteRestore).
+ */
+static u32 *gen8_emit_wa_tail(struct i915_request *rq, u32 *cs)
+{
+	/* Ensure there's always at least one preemption point per-request. */
+	*cs++ = MI_ARB_CHECK;
+	*cs++ = MI_NOOP;
+	rq->wa_tail = intel_ring_offset(rq, cs);
+
+	/* Check that entire request is less than half the ring */
+	assert_request_valid(rq);
+
+	return cs;
+}
+
+static u32 *emit_preempt_busywait(struct i915_request *rq, u32 *cs)
+{
+	*cs++ = MI_ARB_CHECK; /* trigger IDLE->ACTIVE first */
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_EQ_SDD;
+	*cs++ = 0;
+	*cs++ = preempt_address(rq->engine);
+	*cs++ = 0;
+	*cs++ = MI_NOOP;
+
+	return cs;
+}
+
+static __always_inline u32*
+gen8_emit_fini_breadcrumb_tail(struct i915_request *rq, u32 *cs)
+{
+	*cs++ = MI_USER_INTERRUPT;
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	if (intel_engine_has_semaphores(rq->engine) &&
+	    !intel_uc_uses_guc_submission(&rq->engine->gt->uc))
+		cs = emit_preempt_busywait(rq, cs);
+
+	rq->tail = intel_ring_offset(rq, cs);
+	assert_ring_tail_valid(rq->ring, rq->tail);
+
+	return gen8_emit_wa_tail(rq, cs);
+}
+
+static u32 *emit_xcs_breadcrumb(struct i915_request *rq, u32 *cs)
+{
+	return gen8_emit_ggtt_write(cs, rq->fence.seqno, hwsp_offset(rq), 0);
+}
+
+u32 *gen8_emit_fini_breadcrumb_xcs(struct i915_request *rq, u32 *cs)
+{
+	return gen8_emit_fini_breadcrumb_tail(rq, emit_xcs_breadcrumb(rq, cs));
+}
+
+u32 *gen8_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs)
+{
+	cs = gen8_emit_pipe_control(cs,
+				    PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
+				    PIPE_CONTROL_DEPTH_CACHE_FLUSH |
+				    PIPE_CONTROL_DC_FLUSH_ENABLE,
+				    0);
+
+	/* XXX flush+write+CS_STALL all in one upsets gem_concurrent_blt:kbl */
+	cs = gen8_emit_ggtt_write_rcs(cs,
+				      rq->fence.seqno,
+				      hwsp_offset(rq),
+				      PIPE_CONTROL_FLUSH_ENABLE |
+				      PIPE_CONTROL_CS_STALL);
+
+	return gen8_emit_fini_breadcrumb_tail(rq, cs);
+}
+
+u32 *gen11_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs)
+{
+	cs = gen8_emit_ggtt_write_rcs(cs,
+				      rq->fence.seqno,
+				      hwsp_offset(rq),
+				      PIPE_CONTROL_CS_STALL |
+				      PIPE_CONTROL_TILE_CACHE_FLUSH |
+				      PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
+				      PIPE_CONTROL_DEPTH_CACHE_FLUSH |
+				      PIPE_CONTROL_DC_FLUSH_ENABLE |
+				      PIPE_CONTROL_FLUSH_ENABLE);
+
+	return gen8_emit_fini_breadcrumb_tail(rq, cs);
+}
+
+/*
+ * Note that the CS instruction pre-parser will not stall on the breadcrumb
+ * flush and will continue pre-fetching the instructions after it before the
+ * memory sync is completed. On pre-gen12 HW, the pre-parser will stop at
+ * BB_START/END instructions, so, even though we might pre-fetch the pre-amble
+ * of the next request before the memory has been flushed, we're guaranteed that
+ * we won't access the batch itself too early.
+ * However, on gen12+ the parser can pre-fetch across the BB_START/END commands,
+ * so, if the current request is modifying an instruction in the next request on
+ * the same intel_context, we might pre-fetch and then execute the pre-update
+ * instruction. To avoid this, the users of self-modifying code should either
+ * disable the parser around the code emitting the memory writes, via a new flag
+ * added to MI_ARB_CHECK, or emit the writes from a different intel_context. For
+ * the in-kernel use-cases we've opted to use a separate context, see
+ * reloc_gpu() as an example.
+ * All the above applies only to the instructions themselves. Non-inline data
+ * used by the instructions is not pre-fetched.
+ */
+
+static u32 *gen12_emit_preempt_busywait(struct i915_request *rq, u32 *cs)
+{
+	*cs++ = MI_ARB_CHECK; /* trigger IDLE->ACTIVE first */
+	*cs++ = MI_SEMAPHORE_WAIT_TOKEN |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_EQ_SDD;
+	*cs++ = 0;
+	*cs++ = preempt_address(rq->engine);
+	*cs++ = 0;
+	*cs++ = 0;
+
+	return cs;
+}
+
+/* Wa_14014475959:dg2 */
+#define CCS_SEMAPHORE_PPHWSP_OFFSET	0x540
+static u32 ccs_semaphore_offset(struct i915_request *rq)
+{
+	return i915_ggtt_offset(rq->context->state) +
+		(LRC_PPHWSP_PN * PAGE_SIZE) + CCS_SEMAPHORE_PPHWSP_OFFSET;
+}
+
+/* Wa_14014475959:dg2 */
+static u32 *ccs_emit_wa_busywait(struct i915_request *rq, u32 *cs)
+{
+	int i;
+
+	*cs++ = MI_ATOMIC_INLINE | MI_ATOMIC_GLOBAL_GTT | MI_ATOMIC_CS_STALL |
+		MI_ATOMIC_MOVE;
+	*cs++ = ccs_semaphore_offset(rq);
+	*cs++ = 0;
+	*cs++ = 1;
+
+	/*
+	 * When MI_ATOMIC_INLINE_DATA set this command must be 11 DW + (1 NOP)
+	 * to align. 4 DWs above + 8 filler DWs here.
+	 */
+	for (i = 0; i < 8; ++i)
+		*cs++ = 0;
+
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_EQ_SDD;
+	*cs++ = 0;
+	*cs++ = ccs_semaphore_offset(rq);
+	*cs++ = 0;
+
+	return cs;
+}
+
+static __always_inline u32*
+gen12_emit_fini_breadcrumb_tail(struct i915_request *rq, u32 *cs)
+{
+	*cs++ = MI_USER_INTERRUPT;
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	if (intel_engine_has_semaphores(rq->engine) &&
+	    !intel_uc_uses_guc_submission(&rq->engine->gt->uc))
+		cs = gen12_emit_preempt_busywait(rq, cs);
+
+	/* Wa_14014475959:dg2 */
+	if (intel_engine_uses_wa_hold_ccs_switchout(rq->engine))
+		cs = ccs_emit_wa_busywait(rq, cs);
+
+	rq->tail = intel_ring_offset(rq, cs);
+	assert_ring_tail_valid(rq->ring, rq->tail);
+
+	return gen8_emit_wa_tail(rq, cs);
+}
+
+u32 *gen12_emit_fini_breadcrumb_xcs(struct i915_request *rq, u32 *cs)
+{
+	/* XXX Stalling flush before seqno write; post-sync not */
+	cs = emit_xcs_breadcrumb(rq, __gen8_emit_flush_dw(cs, 0, 0, 0));
+	return gen12_emit_fini_breadcrumb_tail(rq, cs);
+}
+
+u32 *gen12_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs)
+{
+	struct drm_i915_private *i915 = rq->engine->i915;
+	u32 flags = (PIPE_CONTROL_CS_STALL |
+		     PIPE_CONTROL_TILE_CACHE_FLUSH |
+		     PIPE_CONTROL_FLUSH_L3 |
+		     PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
+		     PIPE_CONTROL_DEPTH_CACHE_FLUSH |
+		     PIPE_CONTROL_DC_FLUSH_ENABLE |
+		     PIPE_CONTROL_FLUSH_ENABLE);
+
+	if (GRAPHICS_VER(i915) == 12 && GRAPHICS_VER_FULL(i915) < IP_VER(12, 50))
+		/* Wa_1409600907 */
+		flags |= PIPE_CONTROL_DEPTH_STALL;
+
+	if (!HAS_3D_PIPELINE(rq->engine->i915))
+		flags &= ~PIPE_CONTROL_3D_ARCH_FLAGS;
+	else if (rq->engine->class == COMPUTE_CLASS)
+		flags &= ~PIPE_CONTROL_3D_ENGINE_FLAGS;
+
+	cs = gen12_emit_ggtt_write_rcs(cs,
+				       rq->fence.seqno,
+				       hwsp_offset(rq),
+				       PIPE_CONTROL0_HDC_PIPELINE_FLUSH,
+				       flags);
+
+	return gen12_emit_fini_breadcrumb_tail(rq, cs);
+}
diff --git a/drivers/gpu/drm/i915/gt/gen8_engine_cs.h b/drivers/gpu/drm/i915/gt/gen8_engine_cs.h
new file mode 100644
index 000000000..651eb786e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen8_engine_cs.h
@@ -0,0 +1,138 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#ifndef __GEN8_ENGINE_CS_H__
+#define __GEN8_ENGINE_CS_H__
+
+#include <linux/string.h>
+#include <linux/types.h>
+
+#include "i915_gem.h" /* GEM_BUG_ON */
+#include "intel_gt_regs.h"
+#include "intel_gpu_commands.h"
+
+struct intel_engine_cs;
+struct intel_gt;
+struct i915_request;
+
+int gen8_emit_flush_rcs(struct i915_request *rq, u32 mode);
+int gen11_emit_flush_rcs(struct i915_request *rq, u32 mode);
+int gen12_emit_flush_rcs(struct i915_request *rq, u32 mode);
+
+int gen8_emit_flush_xcs(struct i915_request *rq, u32 mode);
+int gen12_emit_flush_xcs(struct i915_request *rq, u32 mode);
+
+int gen8_emit_init_breadcrumb(struct i915_request *rq);
+
+int gen8_emit_bb_start_noarb(struct i915_request *rq,
+			     u64 offset, u32 len,
+			     const unsigned int flags);
+int gen8_emit_bb_start(struct i915_request *rq,
+		       u64 offset, u32 len,
+		       const unsigned int flags);
+
+int gen125_emit_bb_start_noarb(struct i915_request *rq,
+			       u64 offset, u32 len,
+			       const unsigned int flags);
+int gen125_emit_bb_start(struct i915_request *rq,
+			 u64 offset, u32 len,
+			 const unsigned int flags);
+
+u32 *gen8_emit_fini_breadcrumb_xcs(struct i915_request *rq, u32 *cs);
+u32 *gen12_emit_fini_breadcrumb_xcs(struct i915_request *rq, u32 *cs);
+
+u32 *gen8_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs);
+u32 *gen11_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs);
+u32 *gen12_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs);
+
+u32 *gen12_emit_aux_table_inv(struct intel_engine_cs *engine, u32 *cs);
+
+static inline u32 *
+__gen8_emit_pipe_control(u32 *batch, u32 flags0, u32 flags1, u32 offset)
+{
+	memset(batch, 0, 6 * sizeof(u32));
+
+	batch[0] = GFX_OP_PIPE_CONTROL(6) | flags0;
+	batch[1] = flags1;
+	batch[2] = offset;
+
+	return batch + 6;
+}
+
+static inline u32 *gen8_emit_pipe_control(u32 *batch, u32 flags, u32 offset)
+{
+	return __gen8_emit_pipe_control(batch, 0, flags, offset);
+}
+
+static inline u32 *gen12_emit_pipe_control(u32 *batch, u32 flags0, u32 flags1, u32 offset)
+{
+	return __gen8_emit_pipe_control(batch, flags0, flags1, offset);
+}
+
+static inline u32 *
+__gen8_emit_write_rcs(u32 *cs, u32 value, u32 offset, u32 flags0, u32 flags1)
+{
+	*cs++ = GFX_OP_PIPE_CONTROL(6) | flags0;
+	*cs++ = flags1 | PIPE_CONTROL_QW_WRITE;
+	*cs++ = offset;
+	*cs++ = 0;
+	*cs++ = value;
+	*cs++ = 0; /* We're thrashing one extra dword. */
+
+	return cs;
+}
+
+static inline u32*
+gen8_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
+{
+	/* We're using qword write, offset should be aligned to 8 bytes. */
+	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
+
+	return __gen8_emit_write_rcs(cs,
+				     value,
+				     gtt_offset,
+				     0,
+				     flags | PIPE_CONTROL_GLOBAL_GTT_IVB);
+}
+
+static inline u32*
+gen12_emit_ggtt_write_rcs(u32 *cs, u32 value, u32 gtt_offset, u32 flags0, u32 flags1)
+{
+	/* We're using qword write, offset should be aligned to 8 bytes. */
+	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
+
+	return __gen8_emit_write_rcs(cs,
+				     value,
+				     gtt_offset,
+				     flags0,
+				     flags1 | PIPE_CONTROL_GLOBAL_GTT_IVB);
+}
+
+static inline u32 *
+__gen8_emit_flush_dw(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
+{
+	*cs++ = (MI_FLUSH_DW + 1) | flags;
+	*cs++ = gtt_offset;
+	*cs++ = 0;
+	*cs++ = value;
+
+	return cs;
+}
+
+static inline u32 *
+gen8_emit_ggtt_write(u32 *cs, u32 value, u32 gtt_offset, u32 flags)
+{
+	/* w/a: bit 5 needs to be zero for MI_FLUSH_DW address. */
+	GEM_BUG_ON(gtt_offset & (1 << 5));
+	/* Offset should be aligned to 8 bytes for both (QW/DW) write types */
+	GEM_BUG_ON(!IS_ALIGNED(gtt_offset, 8));
+
+	return __gen8_emit_flush_dw(cs,
+				    value,
+				    gtt_offset | MI_FLUSH_DW_USE_GTT,
+				    flags | MI_FLUSH_DW_OP_STOREDW);
+}
+
+#endif /* __GEN8_ENGINE_CS_H__ */
diff --git a/drivers/gpu/drm/i915/gt/gen8_ppgtt.c b/drivers/gpu/drm/i915/gt/gen8_ppgtt.c
new file mode 100644
index 000000000..2128b7a72
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen8_ppgtt.c
@@ -0,0 +1,994 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/log2.h>
+
+#include "gem/i915_gem_lmem.h"
+
+#include "gen8_ppgtt.h"
+#include "i915_scatterlist.h"
+#include "i915_trace.h"
+#include "i915_pvinfo.h"
+#include "i915_vgpu.h"
+#include "intel_gt.h"
+#include "intel_gtt.h"
+
+static u64 gen8_pde_encode(const dma_addr_t addr,
+			   const enum i915_cache_level level)
+{
+	u64 pde = addr | GEN8_PAGE_PRESENT | GEN8_PAGE_RW;
+
+	if (level != I915_CACHE_NONE)
+		pde |= PPAT_CACHED_PDE;
+	else
+		pde |= PPAT_UNCACHED;
+
+	return pde;
+}
+
+static u64 gen8_pte_encode(dma_addr_t addr,
+			   enum i915_cache_level level,
+			   u32 flags)
+{
+	gen8_pte_t pte = addr | GEN8_PAGE_PRESENT | GEN8_PAGE_RW;
+
+	if (unlikely(flags & PTE_READ_ONLY))
+		pte &= ~GEN8_PAGE_RW;
+
+	if (flags & PTE_LM)
+		pte |= GEN12_PPGTT_PTE_LM;
+
+	switch (level) {
+	case I915_CACHE_NONE:
+		pte |= PPAT_UNCACHED;
+		break;
+	case I915_CACHE_WT:
+		pte |= PPAT_DISPLAY_ELLC;
+		break;
+	default:
+		pte |= PPAT_CACHED;
+		break;
+	}
+
+	return pte;
+}
+
+static void gen8_ppgtt_notify_vgt(struct i915_ppgtt *ppgtt, bool create)
+{
+	struct drm_i915_private *i915 = ppgtt->vm.i915;
+	struct intel_uncore *uncore = ppgtt->vm.gt->uncore;
+	enum vgt_g2v_type msg;
+	int i;
+
+	if (create)
+		atomic_inc(px_used(ppgtt->pd)); /* never remove */
+	else
+		atomic_dec(px_used(ppgtt->pd));
+
+	mutex_lock(&i915->vgpu.lock);
+
+	if (i915_vm_is_4lvl(&ppgtt->vm)) {
+		const u64 daddr = px_dma(ppgtt->pd);
+
+		intel_uncore_write(uncore,
+				   vgtif_reg(pdp[0].lo), lower_32_bits(daddr));
+		intel_uncore_write(uncore,
+				   vgtif_reg(pdp[0].hi), upper_32_bits(daddr));
+
+		msg = create ?
+			VGT_G2V_PPGTT_L4_PAGE_TABLE_CREATE :
+			VGT_G2V_PPGTT_L4_PAGE_TABLE_DESTROY;
+	} else {
+		for (i = 0; i < GEN8_3LVL_PDPES; i++) {
+			const u64 daddr = i915_page_dir_dma_addr(ppgtt, i);
+
+			intel_uncore_write(uncore,
+					   vgtif_reg(pdp[i].lo),
+					   lower_32_bits(daddr));
+			intel_uncore_write(uncore,
+					   vgtif_reg(pdp[i].hi),
+					   upper_32_bits(daddr));
+		}
+
+		msg = create ?
+			VGT_G2V_PPGTT_L3_PAGE_TABLE_CREATE :
+			VGT_G2V_PPGTT_L3_PAGE_TABLE_DESTROY;
+	}
+
+	/* g2v_notify atomically (via hv trap) consumes the message packet. */
+	intel_uncore_write(uncore, vgtif_reg(g2v_notify), msg);
+
+	mutex_unlock(&i915->vgpu.lock);
+}
+
+/* Index shifts into the pagetable are offset by GEN8_PTE_SHIFT [12] */
+#define GEN8_PAGE_SIZE (SZ_4K) /* page and page-directory sizes are the same */
+#define GEN8_PTE_SHIFT (ilog2(GEN8_PAGE_SIZE))
+#define GEN8_PDES (GEN8_PAGE_SIZE / sizeof(u64))
+#define gen8_pd_shift(lvl) ((lvl) * ilog2(GEN8_PDES))
+#define gen8_pd_index(i, lvl) i915_pde_index((i), gen8_pd_shift(lvl))
+#define __gen8_pte_shift(lvl) (GEN8_PTE_SHIFT + gen8_pd_shift(lvl))
+#define __gen8_pte_index(a, lvl) i915_pde_index((a), __gen8_pte_shift(lvl))
+
+#define as_pd(x) container_of((x), typeof(struct i915_page_directory), pt)
+
+static unsigned int
+gen8_pd_range(u64 start, u64 end, int lvl, unsigned int *idx)
+{
+	const int shift = gen8_pd_shift(lvl);
+	const u64 mask = ~0ull << gen8_pd_shift(lvl + 1);
+
+	GEM_BUG_ON(start >= end);
+	end += ~mask >> gen8_pd_shift(1);
+
+	*idx = i915_pde_index(start, shift);
+	if ((start ^ end) & mask)
+		return GEN8_PDES - *idx;
+	else
+		return i915_pde_index(end, shift) - *idx;
+}
+
+static bool gen8_pd_contains(u64 start, u64 end, int lvl)
+{
+	const u64 mask = ~0ull << gen8_pd_shift(lvl + 1);
+
+	GEM_BUG_ON(start >= end);
+	return (start ^ end) & mask && (start & ~mask) == 0;
+}
+
+static unsigned int gen8_pt_count(u64 start, u64 end)
+{
+	GEM_BUG_ON(start >= end);
+	if ((start ^ end) >> gen8_pd_shift(1))
+		return GEN8_PDES - (start & (GEN8_PDES - 1));
+	else
+		return end - start;
+}
+
+static unsigned int gen8_pd_top_count(const struct i915_address_space *vm)
+{
+	unsigned int shift = __gen8_pte_shift(vm->top);
+
+	return (vm->total + (1ull << shift) - 1) >> shift;
+}
+
+static struct i915_page_directory *
+gen8_pdp_for_page_index(struct i915_address_space * const vm, const u64 idx)
+{
+	struct i915_ppgtt * const ppgtt = i915_vm_to_ppgtt(vm);
+
+	if (vm->top == 2)
+		return ppgtt->pd;
+	else
+		return i915_pd_entry(ppgtt->pd, gen8_pd_index(idx, vm->top));
+}
+
+static struct i915_page_directory *
+gen8_pdp_for_page_address(struct i915_address_space * const vm, const u64 addr)
+{
+	return gen8_pdp_for_page_index(vm, addr >> GEN8_PTE_SHIFT);
+}
+
+static void __gen8_ppgtt_cleanup(struct i915_address_space *vm,
+				 struct i915_page_directory *pd,
+				 int count, int lvl)
+{
+	if (lvl) {
+		void **pde = pd->entry;
+
+		do {
+			if (!*pde)
+				continue;
+
+			__gen8_ppgtt_cleanup(vm, *pde, GEN8_PDES, lvl - 1);
+		} while (pde++, --count);
+	}
+
+	free_px(vm, &pd->pt, lvl);
+}
+
+static void gen8_ppgtt_cleanup(struct i915_address_space *vm)
+{
+	struct i915_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
+
+	if (intel_vgpu_active(vm->i915))
+		gen8_ppgtt_notify_vgt(ppgtt, false);
+
+	if (ppgtt->pd)
+		__gen8_ppgtt_cleanup(vm, ppgtt->pd,
+				     gen8_pd_top_count(vm), vm->top);
+
+	free_scratch(vm);
+}
+
+static u64 __gen8_ppgtt_clear(struct i915_address_space * const vm,
+			      struct i915_page_directory * const pd,
+			      u64 start, const u64 end, int lvl)
+{
+	const struct drm_i915_gem_object * const scratch = vm->scratch[lvl];
+	unsigned int idx, len;
+
+	GEM_BUG_ON(end > vm->total >> GEN8_PTE_SHIFT);
+
+	len = gen8_pd_range(start, end, lvl--, &idx);
+	DBG("%s(%p):{ lvl:%d, start:%llx, end:%llx, idx:%d, len:%d, used:%d }\n",
+	    __func__, vm, lvl + 1, start, end,
+	    idx, len, atomic_read(px_used(pd)));
+	GEM_BUG_ON(!len || len >= atomic_read(px_used(pd)));
+
+	do {
+		struct i915_page_table *pt = pd->entry[idx];
+
+		if (atomic_fetch_inc(&pt->used) >> gen8_pd_shift(1) &&
+		    gen8_pd_contains(start, end, lvl)) {
+			DBG("%s(%p):{ lvl:%d, idx:%d, start:%llx, end:%llx } removing pd\n",
+			    __func__, vm, lvl + 1, idx, start, end);
+			clear_pd_entry(pd, idx, scratch);
+			__gen8_ppgtt_cleanup(vm, as_pd(pt), I915_PDES, lvl);
+			start += (u64)I915_PDES << gen8_pd_shift(lvl);
+			continue;
+		}
+
+		if (lvl) {
+			start = __gen8_ppgtt_clear(vm, as_pd(pt),
+						   start, end, lvl);
+		} else {
+			unsigned int count;
+			unsigned int pte = gen8_pd_index(start, 0);
+			unsigned int num_ptes;
+			u64 *vaddr;
+
+			count = gen8_pt_count(start, end);
+			DBG("%s(%p):{ lvl:%d, start:%llx, end:%llx, idx:%d, len:%d, used:%d } removing pte\n",
+			    __func__, vm, lvl, start, end,
+			    gen8_pd_index(start, 0), count,
+			    atomic_read(&pt->used));
+			GEM_BUG_ON(!count || count >= atomic_read(&pt->used));
+
+			num_ptes = count;
+			if (pt->is_compact) {
+				GEM_BUG_ON(num_ptes % 16);
+				GEM_BUG_ON(pte % 16);
+				num_ptes /= 16;
+				pte /= 16;
+			}
+
+			vaddr = px_vaddr(pt);
+			memset64(vaddr + pte,
+				 vm->scratch[0]->encode,
+				 num_ptes);
+
+			atomic_sub(count, &pt->used);
+			start += count;
+		}
+
+		if (release_pd_entry(pd, idx, pt, scratch))
+			free_px(vm, pt, lvl);
+	} while (idx++, --len);
+
+	return start;
+}
+
+static void gen8_ppgtt_clear(struct i915_address_space *vm,
+			     u64 start, u64 length)
+{
+	GEM_BUG_ON(!IS_ALIGNED(start, BIT_ULL(GEN8_PTE_SHIFT)));
+	GEM_BUG_ON(!IS_ALIGNED(length, BIT_ULL(GEN8_PTE_SHIFT)));
+	GEM_BUG_ON(range_overflows(start, length, vm->total));
+
+	start >>= GEN8_PTE_SHIFT;
+	length >>= GEN8_PTE_SHIFT;
+	GEM_BUG_ON(length == 0);
+
+	__gen8_ppgtt_clear(vm, i915_vm_to_ppgtt(vm)->pd,
+			   start, start + length, vm->top);
+}
+
+static void __gen8_ppgtt_alloc(struct i915_address_space * const vm,
+			       struct i915_vm_pt_stash *stash,
+			       struct i915_page_directory * const pd,
+			       u64 * const start, const u64 end, int lvl)
+{
+	unsigned int idx, len;
+
+	GEM_BUG_ON(end > vm->total >> GEN8_PTE_SHIFT);
+
+	len = gen8_pd_range(*start, end, lvl--, &idx);
+	DBG("%s(%p):{ lvl:%d, start:%llx, end:%llx, idx:%d, len:%d, used:%d }\n",
+	    __func__, vm, lvl + 1, *start, end,
+	    idx, len, atomic_read(px_used(pd)));
+	GEM_BUG_ON(!len || (idx + len - 1) >> gen8_pd_shift(1));
+
+	spin_lock(&pd->lock);
+	GEM_BUG_ON(!atomic_read(px_used(pd))); /* Must be pinned! */
+	do {
+		struct i915_page_table *pt = pd->entry[idx];
+
+		if (!pt) {
+			spin_unlock(&pd->lock);
+
+			DBG("%s(%p):{ lvl:%d, idx:%d } allocating new tree\n",
+			    __func__, vm, lvl + 1, idx);
+
+			pt = stash->pt[!!lvl];
+			__i915_gem_object_pin_pages(pt->base);
+
+			fill_px(pt, vm->scratch[lvl]->encode);
+
+			spin_lock(&pd->lock);
+			if (likely(!pd->entry[idx])) {
+				stash->pt[!!lvl] = pt->stash;
+				atomic_set(&pt->used, 0);
+				set_pd_entry(pd, idx, pt);
+			} else {
+				pt = pd->entry[idx];
+			}
+		}
+
+		if (lvl) {
+			atomic_inc(&pt->used);
+			spin_unlock(&pd->lock);
+
+			__gen8_ppgtt_alloc(vm, stash,
+					   as_pd(pt), start, end, lvl);
+
+			spin_lock(&pd->lock);
+			atomic_dec(&pt->used);
+			GEM_BUG_ON(!atomic_read(&pt->used));
+		} else {
+			unsigned int count = gen8_pt_count(*start, end);
+
+			DBG("%s(%p):{ lvl:%d, start:%llx, end:%llx, idx:%d, len:%d, used:%d } inserting pte\n",
+			    __func__, vm, lvl, *start, end,
+			    gen8_pd_index(*start, 0), count,
+			    atomic_read(&pt->used));
+
+			atomic_add(count, &pt->used);
+			/* All other pdes may be simultaneously removed */
+			GEM_BUG_ON(atomic_read(&pt->used) > NALLOC * I915_PDES);
+			*start += count;
+		}
+	} while (idx++, --len);
+	spin_unlock(&pd->lock);
+}
+
+static void gen8_ppgtt_alloc(struct i915_address_space *vm,
+			     struct i915_vm_pt_stash *stash,
+			     u64 start, u64 length)
+{
+	GEM_BUG_ON(!IS_ALIGNED(start, BIT_ULL(GEN8_PTE_SHIFT)));
+	GEM_BUG_ON(!IS_ALIGNED(length, BIT_ULL(GEN8_PTE_SHIFT)));
+	GEM_BUG_ON(range_overflows(start, length, vm->total));
+
+	start >>= GEN8_PTE_SHIFT;
+	length >>= GEN8_PTE_SHIFT;
+	GEM_BUG_ON(length == 0);
+
+	__gen8_ppgtt_alloc(vm, stash, i915_vm_to_ppgtt(vm)->pd,
+			   &start, start + length, vm->top);
+}
+
+static void __gen8_ppgtt_foreach(struct i915_address_space *vm,
+				 struct i915_page_directory *pd,
+				 u64 *start, u64 end, int lvl,
+				 void (*fn)(struct i915_address_space *vm,
+					    struct i915_page_table *pt,
+					    void *data),
+				 void *data)
+{
+	unsigned int idx, len;
+
+	len = gen8_pd_range(*start, end, lvl--, &idx);
+
+	spin_lock(&pd->lock);
+	do {
+		struct i915_page_table *pt = pd->entry[idx];
+
+		atomic_inc(&pt->used);
+		spin_unlock(&pd->lock);
+
+		if (lvl) {
+			__gen8_ppgtt_foreach(vm, as_pd(pt), start, end, lvl,
+					     fn, data);
+		} else {
+			fn(vm, pt, data);
+			*start += gen8_pt_count(*start, end);
+		}
+
+		spin_lock(&pd->lock);
+		atomic_dec(&pt->used);
+	} while (idx++, --len);
+	spin_unlock(&pd->lock);
+}
+
+static void gen8_ppgtt_foreach(struct i915_address_space *vm,
+			       u64 start, u64 length,
+			       void (*fn)(struct i915_address_space *vm,
+					  struct i915_page_table *pt,
+					  void *data),
+			       void *data)
+{
+	start >>= GEN8_PTE_SHIFT;
+	length >>= GEN8_PTE_SHIFT;
+
+	__gen8_ppgtt_foreach(vm, i915_vm_to_ppgtt(vm)->pd,
+			     &start, start + length, vm->top,
+			     fn, data);
+}
+
+static __always_inline u64
+gen8_ppgtt_insert_pte(struct i915_ppgtt *ppgtt,
+		      struct i915_page_directory *pdp,
+		      struct sgt_dma *iter,
+		      u64 idx,
+		      enum i915_cache_level cache_level,
+		      u32 flags)
+{
+	struct i915_page_directory *pd;
+	const gen8_pte_t pte_encode = gen8_pte_encode(0, cache_level, flags);
+	gen8_pte_t *vaddr;
+
+	pd = i915_pd_entry(pdp, gen8_pd_index(idx, 2));
+	vaddr = px_vaddr(i915_pt_entry(pd, gen8_pd_index(idx, 1)));
+	do {
+		GEM_BUG_ON(sg_dma_len(iter->sg) < I915_GTT_PAGE_SIZE);
+		vaddr[gen8_pd_index(idx, 0)] = pte_encode | iter->dma;
+
+		iter->dma += I915_GTT_PAGE_SIZE;
+		if (iter->dma >= iter->max) {
+			iter->sg = __sg_next(iter->sg);
+			if (!iter->sg || sg_dma_len(iter->sg) == 0) {
+				idx = 0;
+				break;
+			}
+
+			iter->dma = sg_dma_address(iter->sg);
+			iter->max = iter->dma + sg_dma_len(iter->sg);
+		}
+
+		if (gen8_pd_index(++idx, 0) == 0) {
+			if (gen8_pd_index(idx, 1) == 0) {
+				/* Limited by sg length for 3lvl */
+				if (gen8_pd_index(idx, 2) == 0)
+					break;
+
+				pd = pdp->entry[gen8_pd_index(idx, 2)];
+			}
+
+			drm_clflush_virt_range(vaddr, PAGE_SIZE);
+			vaddr = px_vaddr(i915_pt_entry(pd, gen8_pd_index(idx, 1)));
+		}
+	} while (1);
+	drm_clflush_virt_range(vaddr, PAGE_SIZE);
+
+	return idx;
+}
+
+static void
+xehpsdv_ppgtt_insert_huge(struct i915_address_space *vm,
+			  struct i915_vma_resource *vma_res,
+			  struct sgt_dma *iter,
+			  enum i915_cache_level cache_level,
+			  u32 flags)
+{
+	const gen8_pte_t pte_encode = vm->pte_encode(0, cache_level, flags);
+	unsigned int rem = sg_dma_len(iter->sg);
+	u64 start = vma_res->start;
+
+	GEM_BUG_ON(!i915_vm_is_4lvl(vm));
+
+	do {
+		struct i915_page_directory * const pdp =
+			gen8_pdp_for_page_address(vm, start);
+		struct i915_page_directory * const pd =
+			i915_pd_entry(pdp, __gen8_pte_index(start, 2));
+		struct i915_page_table *pt =
+			i915_pt_entry(pd, __gen8_pte_index(start, 1));
+		gen8_pte_t encode = pte_encode;
+		unsigned int page_size;
+		gen8_pte_t *vaddr;
+		u16 index, max;
+
+		max = I915_PDES;
+
+		if (vma_res->bi.page_sizes.sg & I915_GTT_PAGE_SIZE_2M &&
+		    IS_ALIGNED(iter->dma, I915_GTT_PAGE_SIZE_2M) &&
+		    rem >= I915_GTT_PAGE_SIZE_2M &&
+		    !__gen8_pte_index(start, 0)) {
+			index = __gen8_pte_index(start, 1);
+			encode |= GEN8_PDE_PS_2M;
+			page_size = I915_GTT_PAGE_SIZE_2M;
+
+			vaddr = px_vaddr(pd);
+		} else {
+			if (encode & GEN12_PPGTT_PTE_LM) {
+				GEM_BUG_ON(__gen8_pte_index(start, 0) % 16);
+				GEM_BUG_ON(rem < I915_GTT_PAGE_SIZE_64K);
+				GEM_BUG_ON(!IS_ALIGNED(iter->dma,
+						       I915_GTT_PAGE_SIZE_64K));
+
+				index = __gen8_pte_index(start, 0) / 16;
+				page_size = I915_GTT_PAGE_SIZE_64K;
+
+				max /= 16;
+
+				vaddr = px_vaddr(pd);
+				vaddr[__gen8_pte_index(start, 1)] |= GEN12_PDE_64K;
+
+				pt->is_compact = true;
+			} else {
+				GEM_BUG_ON(pt->is_compact);
+				index =  __gen8_pte_index(start, 0);
+				page_size = I915_GTT_PAGE_SIZE;
+			}
+
+			vaddr = px_vaddr(pt);
+		}
+
+		do {
+			GEM_BUG_ON(rem < page_size);
+			vaddr[index++] = encode | iter->dma;
+
+			start += page_size;
+			iter->dma += page_size;
+			rem -= page_size;
+			if (iter->dma >= iter->max) {
+				iter->sg = __sg_next(iter->sg);
+				if (!iter->sg)
+					break;
+
+				rem = sg_dma_len(iter->sg);
+				if (!rem)
+					break;
+
+				iter->dma = sg_dma_address(iter->sg);
+				iter->max = iter->dma + rem;
+
+				if (unlikely(!IS_ALIGNED(iter->dma, page_size)))
+					break;
+			}
+		} while (rem >= page_size && index < max);
+
+		vma_res->page_sizes_gtt |= page_size;
+	} while (iter->sg && sg_dma_len(iter->sg));
+}
+
+static void gen8_ppgtt_insert_huge(struct i915_address_space *vm,
+				   struct i915_vma_resource *vma_res,
+				   struct sgt_dma *iter,
+				   enum i915_cache_level cache_level,
+				   u32 flags)
+{
+	const gen8_pte_t pte_encode = gen8_pte_encode(0, cache_level, flags);
+	unsigned int rem = sg_dma_len(iter->sg);
+	u64 start = vma_res->start;
+
+	GEM_BUG_ON(!i915_vm_is_4lvl(vm));
+
+	do {
+		struct i915_page_directory * const pdp =
+			gen8_pdp_for_page_address(vm, start);
+		struct i915_page_directory * const pd =
+			i915_pd_entry(pdp, __gen8_pte_index(start, 2));
+		gen8_pte_t encode = pte_encode;
+		unsigned int maybe_64K = -1;
+		unsigned int page_size;
+		gen8_pte_t *vaddr;
+		u16 index;
+
+		if (vma_res->bi.page_sizes.sg & I915_GTT_PAGE_SIZE_2M &&
+		    IS_ALIGNED(iter->dma, I915_GTT_PAGE_SIZE_2M) &&
+		    rem >= I915_GTT_PAGE_SIZE_2M &&
+		    !__gen8_pte_index(start, 0)) {
+			index = __gen8_pte_index(start, 1);
+			encode |= GEN8_PDE_PS_2M;
+			page_size = I915_GTT_PAGE_SIZE_2M;
+
+			vaddr = px_vaddr(pd);
+		} else {
+			struct i915_page_table *pt =
+				i915_pt_entry(pd, __gen8_pte_index(start, 1));
+
+			index = __gen8_pte_index(start, 0);
+			page_size = I915_GTT_PAGE_SIZE;
+
+			if (!index &&
+			    vma_res->bi.page_sizes.sg & I915_GTT_PAGE_SIZE_64K &&
+			    IS_ALIGNED(iter->dma, I915_GTT_PAGE_SIZE_64K) &&
+			    (IS_ALIGNED(rem, I915_GTT_PAGE_SIZE_64K) ||
+			     rem >= (I915_PDES - index) * I915_GTT_PAGE_SIZE))
+				maybe_64K = __gen8_pte_index(start, 1);
+
+			vaddr = px_vaddr(pt);
+		}
+
+		do {
+			GEM_BUG_ON(sg_dma_len(iter->sg) < page_size);
+			vaddr[index++] = encode | iter->dma;
+
+			start += page_size;
+			iter->dma += page_size;
+			rem -= page_size;
+			if (iter->dma >= iter->max) {
+				iter->sg = __sg_next(iter->sg);
+				if (!iter->sg)
+					break;
+
+				rem = sg_dma_len(iter->sg);
+				if (!rem)
+					break;
+
+				iter->dma = sg_dma_address(iter->sg);
+				iter->max = iter->dma + rem;
+
+				if (maybe_64K != -1 && index < I915_PDES &&
+				    !(IS_ALIGNED(iter->dma, I915_GTT_PAGE_SIZE_64K) &&
+				      (IS_ALIGNED(rem, I915_GTT_PAGE_SIZE_64K) ||
+				       rem >= (I915_PDES - index) * I915_GTT_PAGE_SIZE)))
+					maybe_64K = -1;
+
+				if (unlikely(!IS_ALIGNED(iter->dma, page_size)))
+					break;
+			}
+		} while (rem >= page_size && index < I915_PDES);
+
+		drm_clflush_virt_range(vaddr, PAGE_SIZE);
+
+		/*
+		 * Is it safe to mark the 2M block as 64K? -- Either we have
+		 * filled whole page-table with 64K entries, or filled part of
+		 * it and have reached the end of the sg table and we have
+		 * enough padding.
+		 */
+		if (maybe_64K != -1 &&
+		    (index == I915_PDES ||
+		     (i915_vm_has_scratch_64K(vm) &&
+		      !iter->sg && IS_ALIGNED(vma_res->start +
+					      vma_res->node_size,
+					      I915_GTT_PAGE_SIZE_2M)))) {
+			vaddr = px_vaddr(pd);
+			vaddr[maybe_64K] |= GEN8_PDE_IPS_64K;
+			drm_clflush_virt_range(vaddr, PAGE_SIZE);
+			page_size = I915_GTT_PAGE_SIZE_64K;
+
+			/*
+			 * We write all 4K page entries, even when using 64K
+			 * pages. In order to verify that the HW isn't cheating
+			 * by using the 4K PTE instead of the 64K PTE, we want
+			 * to remove all the surplus entries. If the HW skipped
+			 * the 64K PTE, it will read/write into the scratch page
+			 * instead - which we detect as missing results during
+			 * selftests.
+			 */
+			if (I915_SELFTEST_ONLY(vm->scrub_64K)) {
+				u16 i;
+
+				encode = vm->scratch[0]->encode;
+				vaddr = px_vaddr(i915_pt_entry(pd, maybe_64K));
+
+				for (i = 1; i < index; i += 16)
+					memset64(vaddr + i, encode, 15);
+
+				drm_clflush_virt_range(vaddr, PAGE_SIZE);
+			}
+		}
+
+		vma_res->page_sizes_gtt |= page_size;
+	} while (iter->sg && sg_dma_len(iter->sg));
+}
+
+static void gen8_ppgtt_insert(struct i915_address_space *vm,
+			      struct i915_vma_resource *vma_res,
+			      enum i915_cache_level cache_level,
+			      u32 flags)
+{
+	struct i915_ppgtt * const ppgtt = i915_vm_to_ppgtt(vm);
+	struct sgt_dma iter = sgt_dma(vma_res);
+
+	if (vma_res->bi.page_sizes.sg > I915_GTT_PAGE_SIZE) {
+		if (HAS_64K_PAGES(vm->i915))
+			xehpsdv_ppgtt_insert_huge(vm, vma_res, &iter, cache_level, flags);
+		else
+			gen8_ppgtt_insert_huge(vm, vma_res, &iter, cache_level, flags);
+	} else  {
+		u64 idx = vma_res->start >> GEN8_PTE_SHIFT;
+
+		do {
+			struct i915_page_directory * const pdp =
+				gen8_pdp_for_page_index(vm, idx);
+
+			idx = gen8_ppgtt_insert_pte(ppgtt, pdp, &iter, idx,
+						    cache_level, flags);
+		} while (idx);
+
+		vma_res->page_sizes_gtt = I915_GTT_PAGE_SIZE;
+	}
+}
+
+static void gen8_ppgtt_insert_entry(struct i915_address_space *vm,
+				    dma_addr_t addr,
+				    u64 offset,
+				    enum i915_cache_level level,
+				    u32 flags)
+{
+	u64 idx = offset >> GEN8_PTE_SHIFT;
+	struct i915_page_directory * const pdp =
+		gen8_pdp_for_page_index(vm, idx);
+	struct i915_page_directory *pd =
+		i915_pd_entry(pdp, gen8_pd_index(idx, 2));
+	struct i915_page_table *pt = i915_pt_entry(pd, gen8_pd_index(idx, 1));
+	gen8_pte_t *vaddr;
+
+	GEM_BUG_ON(pt->is_compact);
+
+	vaddr = px_vaddr(pt);
+	vaddr[gen8_pd_index(idx, 0)] = gen8_pte_encode(addr, level, flags);
+	drm_clflush_virt_range(&vaddr[gen8_pd_index(idx, 0)], sizeof(*vaddr));
+}
+
+static void __xehpsdv_ppgtt_insert_entry_lm(struct i915_address_space *vm,
+					    dma_addr_t addr,
+					    u64 offset,
+					    enum i915_cache_level level,
+					    u32 flags)
+{
+	u64 idx = offset >> GEN8_PTE_SHIFT;
+	struct i915_page_directory * const pdp =
+		gen8_pdp_for_page_index(vm, idx);
+	struct i915_page_directory *pd =
+		i915_pd_entry(pdp, gen8_pd_index(idx, 2));
+	struct i915_page_table *pt = i915_pt_entry(pd, gen8_pd_index(idx, 1));
+	gen8_pte_t *vaddr;
+
+	GEM_BUG_ON(!IS_ALIGNED(addr, SZ_64K));
+	GEM_BUG_ON(!IS_ALIGNED(offset, SZ_64K));
+
+	if (!pt->is_compact) {
+		vaddr = px_vaddr(pd);
+		vaddr[gen8_pd_index(idx, 1)] |= GEN12_PDE_64K;
+		pt->is_compact = true;
+	}
+
+	vaddr = px_vaddr(pt);
+	vaddr[gen8_pd_index(idx, 0) / 16] = gen8_pte_encode(addr, level, flags);
+}
+
+static void xehpsdv_ppgtt_insert_entry(struct i915_address_space *vm,
+				       dma_addr_t addr,
+				       u64 offset,
+				       enum i915_cache_level level,
+				       u32 flags)
+{
+	if (flags & PTE_LM)
+		return __xehpsdv_ppgtt_insert_entry_lm(vm, addr, offset,
+						       level, flags);
+
+	return gen8_ppgtt_insert_entry(vm, addr, offset, level, flags);
+}
+
+static int gen8_init_scratch(struct i915_address_space *vm)
+{
+	u32 pte_flags;
+	int ret;
+	int i;
+
+	/*
+	 * If everybody agrees to not to write into the scratch page,
+	 * we can reuse it for all vm, keeping contexts and processes separate.
+	 */
+	if (vm->has_read_only && vm->gt->vm && !i915_is_ggtt(vm->gt->vm)) {
+		struct i915_address_space *clone = vm->gt->vm;
+
+		GEM_BUG_ON(!clone->has_read_only);
+
+		vm->scratch_order = clone->scratch_order;
+		for (i = 0; i <= vm->top; i++)
+			vm->scratch[i] = i915_gem_object_get(clone->scratch[i]);
+
+		return 0;
+	}
+
+	ret = setup_scratch_page(vm);
+	if (ret)
+		return ret;
+
+	pte_flags = vm->has_read_only;
+	if (i915_gem_object_is_lmem(vm->scratch[0]))
+		pte_flags |= PTE_LM;
+
+	vm->scratch[0]->encode =
+		gen8_pte_encode(px_dma(vm->scratch[0]),
+				I915_CACHE_NONE, pte_flags);
+
+	for (i = 1; i <= vm->top; i++) {
+		struct drm_i915_gem_object *obj;
+
+		obj = vm->alloc_pt_dma(vm, I915_GTT_PAGE_SIZE_4K);
+		if (IS_ERR(obj)) {
+			ret = PTR_ERR(obj);
+			goto free_scratch;
+		}
+
+		ret = map_pt_dma(vm, obj);
+		if (ret) {
+			i915_gem_object_put(obj);
+			goto free_scratch;
+		}
+
+		fill_px(obj, vm->scratch[i - 1]->encode);
+		obj->encode = gen8_pde_encode(px_dma(obj), I915_CACHE_NONE);
+
+		vm->scratch[i] = obj;
+	}
+
+	return 0;
+
+free_scratch:
+	while (i--)
+		i915_gem_object_put(vm->scratch[i]);
+	vm->scratch[0] = NULL;
+	return ret;
+}
+
+static int gen8_preallocate_top_level_pdp(struct i915_ppgtt *ppgtt)
+{
+	struct i915_address_space *vm = &ppgtt->vm;
+	struct i915_page_directory *pd = ppgtt->pd;
+	unsigned int idx;
+
+	GEM_BUG_ON(vm->top != 2);
+	GEM_BUG_ON(gen8_pd_top_count(vm) != GEN8_3LVL_PDPES);
+
+	for (idx = 0; idx < GEN8_3LVL_PDPES; idx++) {
+		struct i915_page_directory *pde;
+		int err;
+
+		pde = alloc_pd(vm);
+		if (IS_ERR(pde))
+			return PTR_ERR(pde);
+
+		err = map_pt_dma(vm, pde->pt.base);
+		if (err) {
+			free_pd(vm, pde);
+			return err;
+		}
+
+		fill_px(pde, vm->scratch[1]->encode);
+		set_pd_entry(pd, idx, pde);
+		atomic_inc(px_used(pde)); /* keep pinned */
+	}
+	wmb();
+
+	return 0;
+}
+
+static struct i915_page_directory *
+gen8_alloc_top_pd(struct i915_address_space *vm)
+{
+	const unsigned int count = gen8_pd_top_count(vm);
+	struct i915_page_directory *pd;
+	int err;
+
+	GEM_BUG_ON(count > I915_PDES);
+
+	pd = __alloc_pd(count);
+	if (unlikely(!pd))
+		return ERR_PTR(-ENOMEM);
+
+	pd->pt.base = vm->alloc_pt_dma(vm, I915_GTT_PAGE_SIZE_4K);
+	if (IS_ERR(pd->pt.base)) {
+		err = PTR_ERR(pd->pt.base);
+		pd->pt.base = NULL;
+		goto err_pd;
+	}
+
+	err = map_pt_dma(vm, pd->pt.base);
+	if (err)
+		goto err_pd;
+
+	fill_page_dma(px_base(pd), vm->scratch[vm->top]->encode, count);
+	atomic_inc(px_used(pd)); /* mark as pinned */
+	return pd;
+
+err_pd:
+	free_pd(vm, pd);
+	return ERR_PTR(err);
+}
+
+/*
+ * GEN8 legacy ppgtt programming is accomplished through a max 4 PDP registers
+ * with a net effect resembling a 2-level page table in normal x86 terms. Each
+ * PDP represents 1GB of memory 4 * 512 * 512 * 4096 = 4GB legacy 32b address
+ * space.
+ *
+ */
+struct i915_ppgtt *gen8_ppgtt_create(struct intel_gt *gt,
+				     unsigned long lmem_pt_obj_flags)
+{
+	struct i915_page_directory *pd;
+	struct i915_ppgtt *ppgtt;
+	int err;
+
+	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+	if (!ppgtt)
+		return ERR_PTR(-ENOMEM);
+
+	ppgtt_init(ppgtt, gt, lmem_pt_obj_flags);
+	ppgtt->vm.top = i915_vm_is_4lvl(&ppgtt->vm) ? 3 : 2;
+	ppgtt->vm.pd_shift = ilog2(SZ_4K * SZ_4K / sizeof(gen8_pte_t));
+
+	/*
+	 * From bdw, there is hw support for read-only pages in the PPGTT.
+	 *
+	 * Gen11 has HSDES#:1807136187 unresolved. Disable ro support
+	 * for now.
+	 *
+	 * Gen12 has inherited the same read-only fault issue from gen11.
+	 */
+	ppgtt->vm.has_read_only = !IS_GRAPHICS_VER(gt->i915, 11, 12);
+
+	if (HAS_LMEM(gt->i915)) {
+		ppgtt->vm.alloc_pt_dma = alloc_pt_lmem;
+
+		/*
+		 * On some platforms the hw has dropped support for 4K GTT pages
+		 * when dealing with LMEM, and due to the design of 64K GTT
+		 * pages in the hw, we can only mark the *entire* page-table as
+		 * operating in 64K GTT mode, since the enable bit is still on
+		 * the pde, and not the pte. And since we still need to allow
+		 * 4K GTT pages for SMEM objects, we can't have a "normal" 4K
+		 * page-table with scratch pointing to LMEM, since that's
+		 * undefined from the hw pov. The simplest solution is to just
+		 * move the 64K scratch page to SMEM on such platforms and call
+		 * it a day, since that should work for all configurations.
+		 */
+		if (HAS_64K_PAGES(gt->i915))
+			ppgtt->vm.alloc_scratch_dma = alloc_pt_dma;
+		else
+			ppgtt->vm.alloc_scratch_dma = alloc_pt_lmem;
+	} else {
+		ppgtt->vm.alloc_pt_dma = alloc_pt_dma;
+		ppgtt->vm.alloc_scratch_dma = alloc_pt_dma;
+	}
+
+	ppgtt->vm.pte_encode = gen8_pte_encode;
+
+	ppgtt->vm.bind_async_flags = I915_VMA_LOCAL_BIND;
+	ppgtt->vm.insert_entries = gen8_ppgtt_insert;
+	if (HAS_64K_PAGES(gt->i915))
+		ppgtt->vm.insert_page = xehpsdv_ppgtt_insert_entry;
+	else
+		ppgtt->vm.insert_page = gen8_ppgtt_insert_entry;
+	ppgtt->vm.allocate_va_range = gen8_ppgtt_alloc;
+	ppgtt->vm.clear_range = gen8_ppgtt_clear;
+	ppgtt->vm.foreach = gen8_ppgtt_foreach;
+	ppgtt->vm.cleanup = gen8_ppgtt_cleanup;
+
+	err = gen8_init_scratch(&ppgtt->vm);
+	if (err)
+		goto err_put;
+
+	pd = gen8_alloc_top_pd(&ppgtt->vm);
+	if (IS_ERR(pd)) {
+		err = PTR_ERR(pd);
+		goto err_put;
+	}
+	ppgtt->pd = pd;
+
+	if (!i915_vm_is_4lvl(&ppgtt->vm)) {
+		err = gen8_preallocate_top_level_pdp(ppgtt);
+		if (err)
+			goto err_put;
+	}
+
+	if (intel_vgpu_active(gt->i915))
+		gen8_ppgtt_notify_vgt(ppgtt, true);
+
+	return ppgtt;
+
+err_put:
+	i915_vm_put(&ppgtt->vm);
+	return ERR_PTR(err);
+}
diff --git a/drivers/gpu/drm/i915/gt/gen8_ppgtt.h b/drivers/gpu/drm/i915/gt/gen8_ppgtt.h
new file mode 100644
index 000000000..f541d1926
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen8_ppgtt.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __GEN8_PPGTT_H__
+#define __GEN8_PPGTT_H__
+
+#include <linux/kernel.h>
+
+struct i915_address_space;
+struct intel_gt;
+enum i915_cache_level;
+
+struct i915_ppgtt *gen8_ppgtt_create(struct intel_gt *gt,
+				     unsigned long lmem_pt_obj_flags);
+
+u64 gen8_ggtt_pte_encode(dma_addr_t addr,
+			 enum i915_cache_level level,
+			 u32 flags);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/gen8_renderstate.c b/drivers/gpu/drm/i915/gt/gen8_renderstate.c
new file mode 100644
index 000000000..ef9d7b0dd
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen8_renderstate.c
@@ -0,0 +1,965 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Generated by: intel-gpu-tools-1.8-220-g01153e7
+ */
+
+#include "intel_renderstate.h"
+
+static const u32 gen8_null_state_relocs[] = {
+	0x00000798,
+	0x000007a4,
+	0x000007ac,
+	0x000007bc,
+	-1,
+};
+
+static const u32 gen8_null_state_batch[] = {
+	0x7a000004,
+	0x01000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x69040000,
+	0x78140000,
+	0x04000000,
+	0x7820000a,
+	0x00000000,
+	0x00000000,
+	0x80000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78130002,
+	0x00000000,
+	0x00000000,
+	0x02001808,
+	0x781f0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78510009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78100007,
+	0x00000000,
+	0x00000000,
+	0x00010000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781b0007,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000800,
+	0x00000000,
+	0x78110008,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781e0003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781d0007,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78120002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78500003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781c0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x780c0000,
+	0x00000000,
+	0x78520003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78300000,
+	0x08010040,
+	0x78310000,
+	0x1e000000,
+	0x78320000,
+	0x1e000000,
+	0x78330000,
+	0x1e000000,
+	0x79190002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x791a0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x791b0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79120000,
+	0x00000000,
+	0x79130000,
+	0x00000000,
+	0x79140000,
+	0x00000000,
+	0x79150000,
+	0x00000000,
+	0x79160000,
+	0x00000000,
+	0x78150009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78190009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781a0009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78160009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78170009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78490001,
+	0x00000000,
+	0x00000000,
+	0x784a0000,
+	0x00000000,
+	0x784b0000,
+	0x00000004,
+	0x79170101,
+	0x00000000,
+	0x00000080,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x20000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x40000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x60000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x6101000e,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000001,
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00001001,
+	0x00001001,
+	0x00000001,
+	0x00001001,
+	0x61020001,
+	0x00000000,
+	0x00000000,
+	0x79000002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78050006,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x40000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x80000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0xc0000000,
+	0x00000000,
+	0x00000000,
+	0x79080001,
+	0x00000000,
+	0x00000000,
+	0x790a0001,
+	0x00000000,
+	0x00000000,
+	0x78060003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78070003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78040001,
+	0x00000000,
+	0x00000000,
+	0x79110000,
+	0x00000000,
+	0x780d0000,
+	0x00000000,
+	0x79060000,
+	0x00000000,
+	0x7907001f,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x7902000f,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x790c000f,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x780a0003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78080083,
+	0x00004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x04004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x08004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x10004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x14004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x18004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x1c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x20004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x24004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x28004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x2c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x30004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x34004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x38004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x3c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x40004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x44004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x48004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x4c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x50004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x54004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x58004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x5c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x60004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x64004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x68004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x6c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x70004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x74004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x7c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x80004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78090043,
+	0x02000000,
+	0x22220000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x680b0001,
+	0x78260000,
+	0x00000000,
+	0x78270000,
+	0x00000000,
+	0x78280000,
+	0x00000000,
+	0x78290000,
+	0x00000000,
+	0x782a0000,
+	0x00000000,
+	0x780e0000,
+	0x00000dc1,
+	0x78240000,
+	0x00000e01,
+	0x784f0000,
+	0x80000100,
+	0x784d0000,
+	0x40000000,
+	0x782b0000,
+	0x00000000,
+	0x782c0000,
+	0x00000000,
+	0x782d0000,
+	0x00000000,
+	0x782e0000,
+	0x00000000,
+	0x782f0000,
+	0x00000000,
+	0x780f0000,
+	0x00000000,
+	0x78230000,
+	0x00000e60,
+	0x78210000,
+	0x00000e80,
+	0x7b000005,
+	0x00000004,
+	0x00000001,
+	0x00000000,
+	0x00000001,
+	0x00000000,
+	0x00000000,
+	0x05000000,	 /* cmds end */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,	 /* state start */
+	0x00000000,
+	0x3f800000,
+	0x3f800000,
+	0x3f800000,
+	0x3f800000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,	 /* state end */
+};
+
+RO_RENDERSTATE(8);
diff --git a/drivers/gpu/drm/i915/gt/gen9_renderstate.c b/drivers/gpu/drm/i915/gt/gen9_renderstate.c
new file mode 100644
index 000000000..428b724de
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/gen9_renderstate.c
@@ -0,0 +1,981 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Generated by: intel-gpu-tools-1.19-177-g68e2eab2
+ */
+
+#include "intel_renderstate.h"
+
+static const u32 gen9_null_state_relocs[] = {
+	0x000007a8,
+	0x000007b4,
+	0x000007bc,
+	0x000007cc,
+	-1,
+};
+
+static const u32 gen9_null_state_batch[] = {
+	0x7a000004,
+	0x01000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x69040300,
+	0x78140000,
+	0x04000000,
+	0x7820000a,
+	0x00000000,
+	0x00000000,
+	0x80000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78130002,
+	0x00000000,
+	0x00000000,
+	0x02001808,
+	0x781f0004,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78510009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78100007,
+	0x00000000,
+	0x00000000,
+	0x00010000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781b0007,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000800,
+	0x00000000,
+	0x78110008,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781e0003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781d0009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78120002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78500003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781c0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x780c0000,
+	0x00000000,
+	0x78520003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78300000,
+	0x08010040,
+	0x78310000,
+	0x1e000000,
+	0x78320000,
+	0x1e000000,
+	0x78330000,
+	0x1e000000,
+	0x79190002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x791a0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x791b0002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79120000,
+	0x00000000,
+	0x79130000,
+	0x00000000,
+	0x79140000,
+	0x00000000,
+	0x79150000,
+	0x00000000,
+	0x79160000,
+	0x00000000,
+	0x78150009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78190009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x781a0009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78160009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78170009,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78490001,
+	0x00000000,
+	0x00000000,
+	0x784a0000,
+	0x00000000,
+	0x784b0000,
+	0x00000004,
+	0x79170101,
+	0x00000000,
+	0x00000080,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x20000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x40000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79180006,
+	0x60000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x61010011,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00000001,
+	0x00000000,
+	0x00000001,	 /* reloc */
+	0x00000000,
+	0x00001001,
+	0x00001001,
+	0x00000001,
+	0x00001001,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x61020001,
+	0x00000000,
+	0x00000000,
+	0x79000002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78050006,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x40000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0x80000000,
+	0x00000000,
+	0x00000000,
+	0x79040002,
+	0xc0000000,
+	0x00000000,
+	0x00000000,
+	0x79080001,
+	0x00000000,
+	0x00000000,
+	0x790a0001,
+	0x00000000,
+	0x00000000,
+	0x78060003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78070003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78040001,
+	0x00000000,
+	0x00000000,
+	0x79110000,
+	0x00000000,
+	0x780d0000,
+	0x00000000,
+	0x79060000,
+	0x00000000,
+	0x7907001f,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x7902000f,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x790c000f,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x780a0003,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78080083,
+	0x00004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x04004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x08004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x0c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x10004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x14004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x18004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x1c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x20004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x24004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x28004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x2c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x30004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x34004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x38004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x3c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x40004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x44004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x48004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x4c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x50004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x54004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x58004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x5c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x60004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x64004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x68004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x6c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x70004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x74004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x7c004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x80004000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78090043,
+	0x02000000,
+	0x22220000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x78550003,
+	0x0000000f,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x680b0001,
+	0x780e0000,
+	0x00000e01,
+	0x78240000,
+	0x00000e41,
+	0x784f0000,
+	0x80000100,
+	0x784d0000,
+	0x40000000,
+	0x782b0000,
+	0x00000000,
+	0x782c0000,
+	0x00000000,
+	0x782d0000,
+	0x00000000,
+	0x782e0000,
+	0x00000000,
+	0x782f0000,
+	0x00000000,
+	0x780f0000,
+	0x00000000,
+	0x78230000,
+	0x00000ea0,
+	0x78210000,
+	0x00000ec0,
+	0x78260000,
+	0x00000000,
+	0x78270000,
+	0x00000000,
+	0x78280000,
+	0x00000000,
+	0x78290000,
+	0x00000000,
+	0x782a0000,
+	0x00000000,
+	0x7b000005,
+	0x00000004,
+	0x00000001,
+	0x00000000,
+	0x00000001,
+	0x00000000,
+	0x00000000,
+	0x05000000,	 /* cmds end */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,	 /* state start */
+	0x00000000,
+	0x3f800000,
+	0x3f800000,
+	0x3f800000,
+	0x3f800000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,	 /* state end */
+};
+
+RO_RENDERSTATE(9);
diff --git a/drivers/gpu/drm/i915/gt/hsw_clear_kernel.c b/drivers/gpu/drm/i915/gt/hsw_clear_kernel.c
new file mode 100644
index 000000000..b47f9d4a0
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/hsw_clear_kernel.c
@@ -0,0 +1,61 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ *
+ * Generated by: IGT Gpu Tools on Fri 21 Feb 2020 05:30:13 AM UTC
+ */
+
+static const u32 hsw_clear_kernel[] = {
+	0x00000001, 0x26020128, 0x00000024, 0x00000000,
+	0x00000040, 0x20280c21, 0x00000028, 0x00000001,
+	0x01000010, 0x20000c20, 0x0000002c, 0x00000000,
+	0x00010220, 0x34001c00, 0x00001400, 0x00000160,
+	0x00600001, 0x20600061, 0x00000000, 0x00000000,
+	0x00000008, 0x20601c85, 0x00000e00, 0x0000000c,
+	0x00000005, 0x20601ca5, 0x00000060, 0x00000001,
+	0x00000008, 0x20641c85, 0x00000e00, 0x0000000d,
+	0x00000005, 0x20641ca5, 0x00000064, 0x00000003,
+	0x00000041, 0x207424a5, 0x00000064, 0x00000034,
+	0x00000040, 0x206014a5, 0x00000060, 0x00000074,
+	0x00000008, 0x20681c85, 0x00000e00, 0x00000008,
+	0x00000005, 0x20681ca5, 0x00000068, 0x0000000f,
+	0x00000041, 0x20701ca5, 0x00000060, 0x00000010,
+	0x00000040, 0x206814a5, 0x00000068, 0x00000070,
+	0x00600001, 0x20a00061, 0x00000000, 0x00000000,
+	0x00000005, 0x206c1c85, 0x00000e00, 0x00000007,
+	0x00000041, 0x206c1ca5, 0x0000006c, 0x00000004,
+	0x00600001, 0x20800021, 0x008d0000, 0x00000000,
+	0x00000001, 0x20800021, 0x0000006c, 0x00000000,
+	0x00000001, 0x20840021, 0x00000068, 0x00000000,
+	0x00000001, 0x20880061, 0x00000000, 0x00000003,
+	0x00000005, 0x208c0d21, 0x00000086, 0xffffffff,
+	0x05600032, 0x20a00fa1, 0x008d0080, 0x02190001,
+	0x00000040, 0x20a01ca5, 0x000000a0, 0x00000001,
+	0x05600032, 0x20a00fa1, 0x008d0080, 0x040a8001,
+	0x02000040, 0x20281c21, 0x00000028, 0xffffffff,
+	0x00010220, 0x34001c00, 0x00001400, 0xffffffe0,
+	0x00000001, 0x26020128, 0x00000024, 0x00000000,
+	0x00000001, 0x220010e4, 0x00000000, 0x00000000,
+	0x00000001, 0x220831ec, 0x00000000, 0x007f007f,
+	0x00600001, 0x20400021, 0x008d0000, 0x00000000,
+	0x00600001, 0x2fe00021, 0x008d0000, 0x00000000,
+	0x00200001, 0x20400121, 0x00450020, 0x00000000,
+	0x00000001, 0x20480061, 0x00000000, 0x000f000f,
+	0x00000005, 0x204c0d21, 0x00000046, 0xffffffef,
+	0x00800001, 0x20600061, 0x00000000, 0x00000000,
+	0x00800001, 0x20800061, 0x00000000, 0x00000000,
+	0x00800001, 0x20a00061, 0x00000000, 0x00000000,
+	0x00800001, 0x20c00061, 0x00000000, 0x00000000,
+	0x00800001, 0x20e00061, 0x00000000, 0x00000000,
+	0x00800001, 0x21000061, 0x00000000, 0x00000000,
+	0x00800001, 0x21200061, 0x00000000, 0x00000000,
+	0x00800001, 0x21400061, 0x00000000, 0x00000000,
+	0x05600032, 0x20000fa0, 0x008d0040, 0x120a8000,
+	0x00000040, 0x20402d21, 0x00000020, 0x00100010,
+	0x05600032, 0x20000fa0, 0x008d0040, 0x120a8000,
+	0x02000040, 0x22083d8c, 0x00000208, 0xffffffff,
+	0x00800001, 0xa0000109, 0x00000602, 0x00000000,
+	0x00000040, 0x22001c84, 0x00000200, 0x00000020,
+	0x00010220, 0x34001c00, 0x00001400, 0xffffffc0,
+	0x07600032, 0x20000fa0, 0x008d0fe0, 0x82000010,
+};
diff --git a/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
new file mode 100644
index 000000000..ecc990ec1
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
@@ -0,0 +1,520 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2015-2021 Intel Corporation
+ */
+
+#include <linux/kthread.h>
+#include <linux/string_helpers.h>
+#include <trace/events/dma_fence.h>
+#include <uapi/linux/sched/types.h>
+
+#include "i915_drv.h"
+#include "i915_trace.h"
+#include "intel_breadcrumbs.h"
+#include "intel_context.h"
+#include "intel_engine_pm.h"
+#include "intel_gt_pm.h"
+#include "intel_gt_requests.h"
+
+static bool irq_enable(struct intel_breadcrumbs *b)
+{
+	return intel_engine_irq_enable(b->irq_engine);
+}
+
+static void irq_disable(struct intel_breadcrumbs *b)
+{
+	intel_engine_irq_disable(b->irq_engine);
+}
+
+static void __intel_breadcrumbs_arm_irq(struct intel_breadcrumbs *b)
+{
+	/*
+	 * Since we are waiting on a request, the GPU should be busy
+	 * and should have its own rpm reference.
+	 */
+	if (GEM_WARN_ON(!intel_gt_pm_get_if_awake(b->irq_engine->gt)))
+		return;
+
+	/*
+	 * The breadcrumb irq will be disarmed on the interrupt after the
+	 * waiters are signaled. This gives us a single interrupt window in
+	 * which we can add a new waiter and avoid the cost of re-enabling
+	 * the irq.
+	 */
+	WRITE_ONCE(b->irq_armed, true);
+
+	/* Requests may have completed before we could enable the interrupt. */
+	if (!b->irq_enabled++ && b->irq_enable(b))
+		irq_work_queue(&b->irq_work);
+}
+
+static void intel_breadcrumbs_arm_irq(struct intel_breadcrumbs *b)
+{
+	if (!b->irq_engine)
+		return;
+
+	spin_lock(&b->irq_lock);
+	if (!b->irq_armed)
+		__intel_breadcrumbs_arm_irq(b);
+	spin_unlock(&b->irq_lock);
+}
+
+static void __intel_breadcrumbs_disarm_irq(struct intel_breadcrumbs *b)
+{
+	GEM_BUG_ON(!b->irq_enabled);
+	if (!--b->irq_enabled)
+		b->irq_disable(b);
+
+	WRITE_ONCE(b->irq_armed, false);
+	intel_gt_pm_put_async(b->irq_engine->gt);
+}
+
+static void intel_breadcrumbs_disarm_irq(struct intel_breadcrumbs *b)
+{
+	spin_lock(&b->irq_lock);
+	if (b->irq_armed)
+		__intel_breadcrumbs_disarm_irq(b);
+	spin_unlock(&b->irq_lock);
+}
+
+static void add_signaling_context(struct intel_breadcrumbs *b,
+				  struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->signal_lock);
+
+	spin_lock(&b->signalers_lock);
+	list_add_rcu(&ce->signal_link, &b->signalers);
+	spin_unlock(&b->signalers_lock);
+}
+
+static bool remove_signaling_context(struct intel_breadcrumbs *b,
+				     struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->signal_lock);
+
+	if (!list_empty(&ce->signals))
+		return false;
+
+	spin_lock(&b->signalers_lock);
+	list_del_rcu(&ce->signal_link);
+	spin_unlock(&b->signalers_lock);
+
+	return true;
+}
+
+__maybe_unused static bool
+check_signal_order(struct intel_context *ce, struct i915_request *rq)
+{
+	if (rq->context != ce)
+		return false;
+
+	if (!list_is_last(&rq->signal_link, &ce->signals) &&
+	    i915_seqno_passed(rq->fence.seqno,
+			      list_next_entry(rq, signal_link)->fence.seqno))
+		return false;
+
+	if (!list_is_first(&rq->signal_link, &ce->signals) &&
+	    i915_seqno_passed(list_prev_entry(rq, signal_link)->fence.seqno,
+			      rq->fence.seqno))
+		return false;
+
+	return true;
+}
+
+static bool
+__dma_fence_signal(struct dma_fence *fence)
+{
+	return !test_and_set_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags);
+}
+
+static void
+__dma_fence_signal__timestamp(struct dma_fence *fence, ktime_t timestamp)
+{
+	fence->timestamp = timestamp;
+	set_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &fence->flags);
+	trace_dma_fence_signaled(fence);
+}
+
+static void
+__dma_fence_signal__notify(struct dma_fence *fence,
+			   const struct list_head *list)
+{
+	struct dma_fence_cb *cur, *tmp;
+
+	lockdep_assert_held(fence->lock);
+
+	list_for_each_entry_safe(cur, tmp, list, node) {
+		INIT_LIST_HEAD(&cur->node);
+		cur->func(fence, cur);
+	}
+}
+
+static void add_retire(struct intel_breadcrumbs *b, struct intel_timeline *tl)
+{
+	if (b->irq_engine)
+		intel_engine_add_retire(b->irq_engine, tl);
+}
+
+static struct llist_node *
+slist_add(struct llist_node *node, struct llist_node *head)
+{
+	node->next = head;
+	return node;
+}
+
+static void signal_irq_work(struct irq_work *work)
+{
+	struct intel_breadcrumbs *b = container_of(work, typeof(*b), irq_work);
+	const ktime_t timestamp = ktime_get();
+	struct llist_node *signal, *sn;
+	struct intel_context *ce;
+
+	signal = NULL;
+	if (unlikely(!llist_empty(&b->signaled_requests)))
+		signal = llist_del_all(&b->signaled_requests);
+
+	/*
+	 * Keep the irq armed until the interrupt after all listeners are gone.
+	 *
+	 * Enabling/disabling the interrupt is rather costly, roughly a couple
+	 * of hundred microseconds. If we are proactive and enable/disable
+	 * the interrupt around every request that wants a breadcrumb, we
+	 * quickly drown in the extra orders of magnitude of latency imposed
+	 * on request submission.
+	 *
+	 * So we try to be lazy, and keep the interrupts enabled until no
+	 * more listeners appear within a breadcrumb interrupt interval (that
+	 * is until a request completes that no one cares about). The
+	 * observation is that listeners come in batches, and will often
+	 * listen to a bunch of requests in succession. Though note on icl+,
+	 * interrupts are always enabled due to concerns with rc6 being
+	 * dysfunctional with per-engine interrupt masking.
+	 *
+	 * We also try to avoid raising too many interrupts, as they may
+	 * be generated by userspace batches and it is unfortunately rather
+	 * too easy to drown the CPU under a flood of GPU interrupts. Thus
+	 * whenever no one appears to be listening, we turn off the interrupts.
+	 * Fewer interrupts should conserve power -- at the very least, fewer
+	 * interrupt draw less ire from other users of the system and tools
+	 * like powertop.
+	 */
+	if (!signal && READ_ONCE(b->irq_armed) && list_empty(&b->signalers))
+		intel_breadcrumbs_disarm_irq(b);
+
+	rcu_read_lock();
+	atomic_inc(&b->signaler_active);
+	list_for_each_entry_rcu(ce, &b->signalers, signal_link) {
+		struct i915_request *rq;
+
+		list_for_each_entry_rcu(rq, &ce->signals, signal_link) {
+			bool release;
+
+			if (!__i915_request_is_complete(rq))
+				break;
+
+			if (!test_and_clear_bit(I915_FENCE_FLAG_SIGNAL,
+						&rq->fence.flags))
+				break;
+
+			/*
+			 * Queue for execution after dropping the signaling
+			 * spinlock as the callback chain may end up adding
+			 * more signalers to the same context or engine.
+			 */
+			spin_lock(&ce->signal_lock);
+			list_del_rcu(&rq->signal_link);
+			release = remove_signaling_context(b, ce);
+			spin_unlock(&ce->signal_lock);
+			if (release) {
+				if (intel_timeline_is_last(ce->timeline, rq))
+					add_retire(b, ce->timeline);
+				intel_context_put(ce);
+			}
+
+			if (__dma_fence_signal(&rq->fence))
+				/* We own signal_node now, xfer to local list */
+				signal = slist_add(&rq->signal_node, signal);
+			else
+				i915_request_put(rq);
+		}
+	}
+	atomic_dec(&b->signaler_active);
+	rcu_read_unlock();
+
+	llist_for_each_safe(signal, sn, signal) {
+		struct i915_request *rq =
+			llist_entry(signal, typeof(*rq), signal_node);
+		struct list_head cb_list;
+
+		if (rq->engine->sched_engine->retire_inflight_request_prio)
+			rq->engine->sched_engine->retire_inflight_request_prio(rq);
+
+		spin_lock(&rq->lock);
+		list_replace(&rq->fence.cb_list, &cb_list);
+		__dma_fence_signal__timestamp(&rq->fence, timestamp);
+		__dma_fence_signal__notify(&rq->fence, &cb_list);
+		spin_unlock(&rq->lock);
+
+		i915_request_put(rq);
+	}
+
+	if (!READ_ONCE(b->irq_armed) && !list_empty(&b->signalers))
+		intel_breadcrumbs_arm_irq(b);
+}
+
+struct intel_breadcrumbs *
+intel_breadcrumbs_create(struct intel_engine_cs *irq_engine)
+{
+	struct intel_breadcrumbs *b;
+
+	b = kzalloc(sizeof(*b), GFP_KERNEL);
+	if (!b)
+		return NULL;
+
+	kref_init(&b->ref);
+
+	spin_lock_init(&b->signalers_lock);
+	INIT_LIST_HEAD(&b->signalers);
+	init_llist_head(&b->signaled_requests);
+
+	spin_lock_init(&b->irq_lock);
+	init_irq_work(&b->irq_work, signal_irq_work);
+
+	b->irq_engine = irq_engine;
+	b->irq_enable = irq_enable;
+	b->irq_disable = irq_disable;
+
+	return b;
+}
+
+void intel_breadcrumbs_reset(struct intel_breadcrumbs *b)
+{
+	unsigned long flags;
+
+	if (!b->irq_engine)
+		return;
+
+	spin_lock_irqsave(&b->irq_lock, flags);
+
+	if (b->irq_enabled)
+		b->irq_enable(b);
+	else
+		b->irq_disable(b);
+
+	spin_unlock_irqrestore(&b->irq_lock, flags);
+}
+
+void __intel_breadcrumbs_park(struct intel_breadcrumbs *b)
+{
+	if (!READ_ONCE(b->irq_armed))
+		return;
+
+	/* Kick the work once more to drain the signalers, and disarm the irq */
+	irq_work_sync(&b->irq_work);
+	while (READ_ONCE(b->irq_armed) && !atomic_read(&b->active)) {
+		local_irq_disable();
+		signal_irq_work(&b->irq_work);
+		local_irq_enable();
+		cond_resched();
+	}
+}
+
+void intel_breadcrumbs_free(struct kref *kref)
+{
+	struct intel_breadcrumbs *b = container_of(kref, typeof(*b), ref);
+
+	irq_work_sync(&b->irq_work);
+	GEM_BUG_ON(!list_empty(&b->signalers));
+	GEM_BUG_ON(b->irq_armed);
+
+	kfree(b);
+}
+
+static void irq_signal_request(struct i915_request *rq,
+			       struct intel_breadcrumbs *b)
+{
+	if (!__dma_fence_signal(&rq->fence))
+		return;
+
+	i915_request_get(rq);
+	if (llist_add(&rq->signal_node, &b->signaled_requests))
+		irq_work_queue(&b->irq_work);
+}
+
+static void insert_breadcrumb(struct i915_request *rq)
+{
+	struct intel_breadcrumbs *b = READ_ONCE(rq->engine)->breadcrumbs;
+	struct intel_context *ce = rq->context;
+	struct list_head *pos;
+
+	if (test_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags))
+		return;
+
+	/*
+	 * If the request is already completed, we can transfer it
+	 * straight onto a signaled list, and queue the irq worker for
+	 * its signal completion.
+	 */
+	if (__i915_request_is_complete(rq)) {
+		irq_signal_request(rq, b);
+		return;
+	}
+
+	if (list_empty(&ce->signals)) {
+		intel_context_get(ce);
+		add_signaling_context(b, ce);
+		pos = &ce->signals;
+	} else {
+		/*
+		 * We keep the seqno in retirement order, so we can break
+		 * inside intel_engine_signal_breadcrumbs as soon as we've
+		 * passed the last completed request (or seen a request that
+		 * hasn't event started). We could walk the timeline->requests,
+		 * but keeping a separate signalers_list has the advantage of
+		 * hopefully being much smaller than the full list and so
+		 * provides faster iteration and detection when there are no
+		 * more interrupts required for this context.
+		 *
+		 * We typically expect to add new signalers in order, so we
+		 * start looking for our insertion point from the tail of
+		 * the list.
+		 */
+		list_for_each_prev(pos, &ce->signals) {
+			struct i915_request *it =
+				list_entry(pos, typeof(*it), signal_link);
+
+			if (i915_seqno_passed(rq->fence.seqno, it->fence.seqno))
+				break;
+		}
+	}
+
+	i915_request_get(rq);
+	list_add_rcu(&rq->signal_link, pos);
+	GEM_BUG_ON(!check_signal_order(ce, rq));
+	GEM_BUG_ON(test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags));
+	set_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags);
+
+	/*
+	 * Defer enabling the interrupt to after HW submission and recheck
+	 * the request as it may have completed and raised the interrupt as
+	 * we were attaching it into the lists.
+	 */
+	if (!b->irq_armed || __i915_request_is_complete(rq))
+		irq_work_queue(&b->irq_work);
+}
+
+bool i915_request_enable_breadcrumb(struct i915_request *rq)
+{
+	struct intel_context *ce = rq->context;
+
+	/* Serialises with i915_request_retire() using rq->lock */
+	if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags))
+		return true;
+
+	/*
+	 * Peek at i915_request_submit()/i915_request_unsubmit() status.
+	 *
+	 * If the request is not yet active (and not signaled), we will
+	 * attach the breadcrumb later.
+	 */
+	if (!test_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags))
+		return true;
+
+	spin_lock(&ce->signal_lock);
+	if (test_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags))
+		insert_breadcrumb(rq);
+	spin_unlock(&ce->signal_lock);
+
+	return true;
+}
+
+void i915_request_cancel_breadcrumb(struct i915_request *rq)
+{
+	struct intel_breadcrumbs *b = READ_ONCE(rq->engine)->breadcrumbs;
+	struct intel_context *ce = rq->context;
+	bool release;
+
+	spin_lock(&ce->signal_lock);
+	if (!test_and_clear_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags)) {
+		spin_unlock(&ce->signal_lock);
+		return;
+	}
+
+	list_del_rcu(&rq->signal_link);
+	release = remove_signaling_context(b, ce);
+	spin_unlock(&ce->signal_lock);
+	if (release)
+		intel_context_put(ce);
+
+	if (__i915_request_is_complete(rq))
+		irq_signal_request(rq, b);
+
+	i915_request_put(rq);
+}
+
+void intel_context_remove_breadcrumbs(struct intel_context *ce,
+				      struct intel_breadcrumbs *b)
+{
+	struct i915_request *rq, *rn;
+	bool release = false;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ce->signal_lock, flags);
+
+	if (list_empty(&ce->signals))
+		goto unlock;
+
+	list_for_each_entry_safe(rq, rn, &ce->signals, signal_link) {
+		GEM_BUG_ON(!__i915_request_is_complete(rq));
+		if (!test_and_clear_bit(I915_FENCE_FLAG_SIGNAL,
+					&rq->fence.flags))
+			continue;
+
+		list_del_rcu(&rq->signal_link);
+		irq_signal_request(rq, b);
+		i915_request_put(rq);
+	}
+	release = remove_signaling_context(b, ce);
+
+unlock:
+	spin_unlock_irqrestore(&ce->signal_lock, flags);
+	if (release)
+		intel_context_put(ce);
+
+	while (atomic_read(&b->signaler_active))
+		cpu_relax();
+}
+
+static void print_signals(struct intel_breadcrumbs *b, struct drm_printer *p)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+
+	drm_printf(p, "Signals:\n");
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(ce, &b->signalers, signal_link) {
+		list_for_each_entry_rcu(rq, &ce->signals, signal_link)
+			drm_printf(p, "\t[%llx:%llx%s] @ %dms\n",
+				   rq->fence.context, rq->fence.seqno,
+				   __i915_request_is_complete(rq) ? "!" :
+				   __i915_request_has_started(rq) ? "*" :
+				   "",
+				   jiffies_to_msecs(jiffies - rq->emitted_jiffies));
+	}
+	rcu_read_unlock();
+}
+
+void intel_engine_print_breadcrumbs(struct intel_engine_cs *engine,
+				    struct drm_printer *p)
+{
+	struct intel_breadcrumbs *b;
+
+	b = engine->breadcrumbs;
+	if (!b)
+		return;
+
+	drm_printf(p, "IRQ: %s\n", str_enabled_disabled(b->irq_armed));
+	if (!list_empty(&b->signalers))
+		print_signals(b, p);
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_breadcrumbs.h b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.h
new file mode 100644
index 000000000..be0d4f379
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_breadcrumbs.h
@@ -0,0 +1,63 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_BREADCRUMBS__
+#define __INTEL_BREADCRUMBS__
+
+#include <linux/atomic.h>
+#include <linux/irq_work.h>
+
+#include "intel_breadcrumbs_types.h"
+
+struct drm_printer;
+struct i915_request;
+struct intel_breadcrumbs;
+
+struct intel_breadcrumbs *
+intel_breadcrumbs_create(struct intel_engine_cs *irq_engine);
+void intel_breadcrumbs_free(struct kref *kref);
+
+void intel_breadcrumbs_reset(struct intel_breadcrumbs *b);
+void __intel_breadcrumbs_park(struct intel_breadcrumbs *b);
+
+static inline void intel_breadcrumbs_unpark(struct intel_breadcrumbs *b)
+{
+	atomic_inc(&b->active);
+}
+
+static inline void intel_breadcrumbs_park(struct intel_breadcrumbs *b)
+{
+	if (atomic_dec_and_test(&b->active))
+		__intel_breadcrumbs_park(b);
+}
+
+static inline void
+intel_engine_signal_breadcrumbs(struct intel_engine_cs *engine)
+{
+	irq_work_queue(&engine->breadcrumbs->irq_work);
+}
+
+void intel_engine_print_breadcrumbs(struct intel_engine_cs *engine,
+				    struct drm_printer *p);
+
+bool i915_request_enable_breadcrumb(struct i915_request *request);
+void i915_request_cancel_breadcrumb(struct i915_request *request);
+
+void intel_context_remove_breadcrumbs(struct intel_context *ce,
+				      struct intel_breadcrumbs *b);
+
+static inline struct intel_breadcrumbs *
+intel_breadcrumbs_get(struct intel_breadcrumbs *b)
+{
+	kref_get(&b->ref);
+	return b;
+}
+
+static inline void intel_breadcrumbs_put(struct intel_breadcrumbs *b)
+{
+	kref_put(&b->ref, intel_breadcrumbs_free);
+}
+
+#endif /* __INTEL_BREADCRUMBS__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h b/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h
new file mode 100644
index 000000000..72dfd3748
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h
@@ -0,0 +1,55 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_BREADCRUMBS_TYPES__
+#define __INTEL_BREADCRUMBS_TYPES__
+
+#include <linux/irq_work.h>
+#include <linux/kref.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/types.h>
+
+#include "intel_engine_types.h"
+
+/*
+ * Rather than have every client wait upon all user interrupts,
+ * with the herd waking after every interrupt and each doing the
+ * heavyweight seqno dance, we delegate the task (of being the
+ * bottom-half of the user interrupt) to the first client. After
+ * every interrupt, we wake up one client, who does the heavyweight
+ * coherent seqno read and either goes back to sleep (if incomplete),
+ * or wakes up all the completed clients in parallel, before then
+ * transferring the bottom-half status to the next client in the queue.
+ *
+ * Compared to walking the entire list of waiters in a single dedicated
+ * bottom-half, we reduce the latency of the first waiter by avoiding
+ * a context switch, but incur additional coherent seqno reads when
+ * following the chain of request breadcrumbs. Since it is most likely
+ * that we have a single client waiting on each seqno, then reducing
+ * the overhead of waking that client is much preferred.
+ */
+struct intel_breadcrumbs {
+	struct kref ref;
+	atomic_t active;
+
+	spinlock_t signalers_lock; /* protects the list of signalers */
+	struct list_head signalers;
+	struct llist_head signaled_requests;
+	atomic_t signaler_active;
+
+	spinlock_t irq_lock; /* protects the interrupt from hardirq context */
+	struct irq_work irq_work; /* for use from inside irq_lock */
+	unsigned int irq_enabled;
+	bool irq_armed;
+
+	/* Not all breadcrumbs are attached to physical HW */
+	intel_engine_mask_t	engine_mask;
+	struct intel_engine_cs *irq_engine;
+	bool	(*irq_enable)(struct intel_breadcrumbs *b);
+	void	(*irq_disable)(struct intel_breadcrumbs *b);
+};
+
+#endif /* __INTEL_BREADCRUMBS_TYPES__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c
new file mode 100644
index 000000000..2aa63ec52
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_context.c
@@ -0,0 +1,631 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_pm.h"
+
+#include "i915_drv.h"
+#include "i915_trace.h"
+
+#include "intel_context.h"
+#include "intel_engine.h"
+#include "intel_engine_pm.h"
+#include "intel_ring.h"
+
+static struct kmem_cache *slab_ce;
+
+static struct intel_context *intel_context_alloc(void)
+{
+	return kmem_cache_zalloc(slab_ce, GFP_KERNEL);
+}
+
+static void rcu_context_free(struct rcu_head *rcu)
+{
+	struct intel_context *ce = container_of(rcu, typeof(*ce), rcu);
+
+	trace_intel_context_free(ce);
+	kmem_cache_free(slab_ce, ce);
+}
+
+void intel_context_free(struct intel_context *ce)
+{
+	call_rcu(&ce->rcu, rcu_context_free);
+}
+
+struct intel_context *
+intel_context_create(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+
+	ce = intel_context_alloc();
+	if (!ce)
+		return ERR_PTR(-ENOMEM);
+
+	intel_context_init(ce, engine);
+	trace_intel_context_create(ce);
+	return ce;
+}
+
+int intel_context_alloc_state(struct intel_context *ce)
+{
+	int err = 0;
+
+	if (mutex_lock_interruptible(&ce->pin_mutex))
+		return -EINTR;
+
+	if (!test_bit(CONTEXT_ALLOC_BIT, &ce->flags)) {
+		if (intel_context_is_banned(ce)) {
+			err = -EIO;
+			goto unlock;
+		}
+
+		err = ce->ops->alloc(ce);
+		if (unlikely(err))
+			goto unlock;
+
+		set_bit(CONTEXT_ALLOC_BIT, &ce->flags);
+	}
+
+unlock:
+	mutex_unlock(&ce->pin_mutex);
+	return err;
+}
+
+static int intel_context_active_acquire(struct intel_context *ce)
+{
+	int err;
+
+	__i915_active_acquire(&ce->active);
+
+	if (intel_context_is_barrier(ce) || intel_engine_uses_guc(ce->engine) ||
+	    intel_context_is_parallel(ce))
+		return 0;
+
+	/* Preallocate tracking nodes */
+	err = i915_active_acquire_preallocate_barrier(&ce->active,
+						      ce->engine);
+	if (err)
+		i915_active_release(&ce->active);
+
+	return err;
+}
+
+static void intel_context_active_release(struct intel_context *ce)
+{
+	/* Nodes preallocated in intel_context_active() */
+	i915_active_acquire_barrier(&ce->active);
+	i915_active_release(&ce->active);
+}
+
+static int __context_pin_state(struct i915_vma *vma, struct i915_gem_ww_ctx *ww)
+{
+	unsigned int bias = i915_ggtt_pin_bias(vma) | PIN_OFFSET_BIAS;
+	int err;
+
+	err = i915_ggtt_pin(vma, ww, 0, bias | PIN_HIGH);
+	if (err)
+		return err;
+
+	err = i915_active_acquire(&vma->active);
+	if (err)
+		goto err_unpin;
+
+	/*
+	 * And mark it as a globally pinned object to let the shrinker know
+	 * it cannot reclaim the object until we release it.
+	 */
+	i915_vma_make_unshrinkable(vma);
+	vma->obj->mm.dirty = true;
+
+	return 0;
+
+err_unpin:
+	i915_vma_unpin(vma);
+	return err;
+}
+
+static void __context_unpin_state(struct i915_vma *vma)
+{
+	i915_vma_make_shrinkable(vma);
+	i915_active_release(&vma->active);
+	__i915_vma_unpin(vma);
+}
+
+static int __ring_active(struct intel_ring *ring,
+			 struct i915_gem_ww_ctx *ww)
+{
+	int err;
+
+	err = intel_ring_pin(ring, ww);
+	if (err)
+		return err;
+
+	err = i915_active_acquire(&ring->vma->active);
+	if (err)
+		goto err_pin;
+
+	return 0;
+
+err_pin:
+	intel_ring_unpin(ring);
+	return err;
+}
+
+static void __ring_retire(struct intel_ring *ring)
+{
+	i915_active_release(&ring->vma->active);
+	intel_ring_unpin(ring);
+}
+
+static int intel_context_pre_pin(struct intel_context *ce,
+				 struct i915_gem_ww_ctx *ww)
+{
+	int err;
+
+	CE_TRACE(ce, "active\n");
+
+	err = __ring_active(ce->ring, ww);
+	if (err)
+		return err;
+
+	err = intel_timeline_pin(ce->timeline, ww);
+	if (err)
+		goto err_ring;
+
+	if (!ce->state)
+		return 0;
+
+	err = __context_pin_state(ce->state, ww);
+	if (err)
+		goto err_timeline;
+
+
+	return 0;
+
+err_timeline:
+	intel_timeline_unpin(ce->timeline);
+err_ring:
+	__ring_retire(ce->ring);
+	return err;
+}
+
+static void intel_context_post_unpin(struct intel_context *ce)
+{
+	if (ce->state)
+		__context_unpin_state(ce->state);
+
+	intel_timeline_unpin(ce->timeline);
+	__ring_retire(ce->ring);
+}
+
+int __intel_context_do_pin_ww(struct intel_context *ce,
+			      struct i915_gem_ww_ctx *ww)
+{
+	bool handoff = false;
+	void *vaddr;
+	int err = 0;
+
+	if (unlikely(!test_bit(CONTEXT_ALLOC_BIT, &ce->flags))) {
+		err = intel_context_alloc_state(ce);
+		if (err)
+			return err;
+	}
+
+	/*
+	 * We always pin the context/ring/timeline here, to ensure a pin
+	 * refcount for __intel_context_active(), which prevent a lock
+	 * inversion of ce->pin_mutex vs dma_resv_lock().
+	 */
+
+	err = i915_gem_object_lock(ce->timeline->hwsp_ggtt->obj, ww);
+	if (!err)
+		err = i915_gem_object_lock(ce->ring->vma->obj, ww);
+	if (!err && ce->state)
+		err = i915_gem_object_lock(ce->state->obj, ww);
+	if (!err)
+		err = intel_context_pre_pin(ce, ww);
+	if (err)
+		return err;
+
+	err = ce->ops->pre_pin(ce, ww, &vaddr);
+	if (err)
+		goto err_ctx_unpin;
+
+	err = i915_active_acquire(&ce->active);
+	if (err)
+		goto err_post_unpin;
+
+	err = mutex_lock_interruptible(&ce->pin_mutex);
+	if (err)
+		goto err_release;
+
+	intel_engine_pm_might_get(ce->engine);
+
+	if (unlikely(intel_context_is_closed(ce))) {
+		err = -ENOENT;
+		goto err_unlock;
+	}
+
+	if (likely(!atomic_add_unless(&ce->pin_count, 1, 0))) {
+		err = intel_context_active_acquire(ce);
+		if (unlikely(err))
+			goto err_unlock;
+
+		err = ce->ops->pin(ce, vaddr);
+		if (err) {
+			intel_context_active_release(ce);
+			goto err_unlock;
+		}
+
+		CE_TRACE(ce, "pin ring:{start:%08x, head:%04x, tail:%04x}\n",
+			 i915_ggtt_offset(ce->ring->vma),
+			 ce->ring->head, ce->ring->tail);
+
+		handoff = true;
+		smp_mb__before_atomic(); /* flush pin before it is visible */
+		atomic_inc(&ce->pin_count);
+	}
+
+	GEM_BUG_ON(!intel_context_is_pinned(ce)); /* no overflow! */
+
+	trace_intel_context_do_pin(ce);
+
+err_unlock:
+	mutex_unlock(&ce->pin_mutex);
+err_release:
+	i915_active_release(&ce->active);
+err_post_unpin:
+	if (!handoff)
+		ce->ops->post_unpin(ce);
+err_ctx_unpin:
+	intel_context_post_unpin(ce);
+
+	/*
+	 * Unlock the hwsp_ggtt object since it's shared.
+	 * In principle we can unlock all the global state locked above
+	 * since it's pinned and doesn't need fencing, and will
+	 * thus remain resident until it is explicitly unpinned.
+	 */
+	i915_gem_ww_unlock_single(ce->timeline->hwsp_ggtt->obj);
+
+	return err;
+}
+
+int __intel_context_do_pin(struct intel_context *ce)
+{
+	struct i915_gem_ww_ctx ww;
+	int err;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	err = __intel_context_do_pin_ww(ce, &ww);
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	return err;
+}
+
+void __intel_context_do_unpin(struct intel_context *ce, int sub)
+{
+	if (!atomic_sub_and_test(sub, &ce->pin_count))
+		return;
+
+	CE_TRACE(ce, "unpin\n");
+	ce->ops->unpin(ce);
+	ce->ops->post_unpin(ce);
+
+	/*
+	 * Once released, we may asynchronously drop the active reference.
+	 * As that may be the only reference keeping the context alive,
+	 * take an extra now so that it is not freed before we finish
+	 * dereferencing it.
+	 */
+	intel_context_get(ce);
+	intel_context_active_release(ce);
+	trace_intel_context_do_unpin(ce);
+	intel_context_put(ce);
+}
+
+static void __intel_context_retire(struct i915_active *active)
+{
+	struct intel_context *ce = container_of(active, typeof(*ce), active);
+
+	CE_TRACE(ce, "retire runtime: { total:%lluns, avg:%lluns }\n",
+		 intel_context_get_total_runtime_ns(ce),
+		 intel_context_get_avg_runtime_ns(ce));
+
+	set_bit(CONTEXT_VALID_BIT, &ce->flags);
+	intel_context_post_unpin(ce);
+	intel_context_put(ce);
+}
+
+static int __intel_context_active(struct i915_active *active)
+{
+	struct intel_context *ce = container_of(active, typeof(*ce), active);
+
+	intel_context_get(ce);
+
+	/* everything should already be activated by intel_context_pre_pin() */
+	GEM_WARN_ON(!i915_active_acquire_if_busy(&ce->ring->vma->active));
+	__intel_ring_pin(ce->ring);
+
+	__intel_timeline_pin(ce->timeline);
+
+	if (ce->state) {
+		GEM_WARN_ON(!i915_active_acquire_if_busy(&ce->state->active));
+		__i915_vma_pin(ce->state);
+		i915_vma_make_unshrinkable(ce->state);
+	}
+
+	return 0;
+}
+
+static int
+sw_fence_dummy_notify(struct i915_sw_fence *sf,
+		      enum i915_sw_fence_notify state)
+{
+	return NOTIFY_DONE;
+}
+
+void
+intel_context_init(struct intel_context *ce, struct intel_engine_cs *engine)
+{
+	GEM_BUG_ON(!engine->cops);
+	GEM_BUG_ON(!engine->gt->vm);
+
+	kref_init(&ce->ref);
+
+	ce->engine = engine;
+	ce->ops = engine->cops;
+	ce->sseu = engine->sseu;
+	ce->ring = NULL;
+	ce->ring_size = SZ_4K;
+
+	ewma_runtime_init(&ce->stats.runtime.avg);
+
+	ce->vm = i915_vm_get(engine->gt->vm);
+
+	/* NB ce->signal_link/lock is used under RCU */
+	spin_lock_init(&ce->signal_lock);
+	INIT_LIST_HEAD(&ce->signals);
+
+	mutex_init(&ce->pin_mutex);
+
+	spin_lock_init(&ce->guc_state.lock);
+	INIT_LIST_HEAD(&ce->guc_state.fences);
+	INIT_LIST_HEAD(&ce->guc_state.requests);
+
+	ce->guc_id.id = GUC_INVALID_CONTEXT_ID;
+	INIT_LIST_HEAD(&ce->guc_id.link);
+
+	INIT_LIST_HEAD(&ce->destroyed_link);
+
+	INIT_LIST_HEAD(&ce->parallel.child_list);
+
+	/*
+	 * Initialize fence to be complete as this is expected to be complete
+	 * unless there is a pending schedule disable outstanding.
+	 */
+	i915_sw_fence_init(&ce->guc_state.blocked,
+			   sw_fence_dummy_notify);
+	i915_sw_fence_commit(&ce->guc_state.blocked);
+
+	i915_active_init(&ce->active,
+			 __intel_context_active, __intel_context_retire, 0);
+}
+
+void intel_context_fini(struct intel_context *ce)
+{
+	struct intel_context *child, *next;
+
+	if (ce->timeline)
+		intel_timeline_put(ce->timeline);
+	i915_vm_put(ce->vm);
+
+	/* Need to put the creation ref for the children */
+	if (intel_context_is_parent(ce))
+		for_each_child_safe(ce, child, next)
+			intel_context_put(child);
+
+	mutex_destroy(&ce->pin_mutex);
+	i915_active_fini(&ce->active);
+	i915_sw_fence_fini(&ce->guc_state.blocked);
+}
+
+void i915_context_module_exit(void)
+{
+	kmem_cache_destroy(slab_ce);
+}
+
+int __init i915_context_module_init(void)
+{
+	slab_ce = KMEM_CACHE(intel_context, SLAB_HWCACHE_ALIGN);
+	if (!slab_ce)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void intel_context_enter_engine(struct intel_context *ce)
+{
+	intel_engine_pm_get(ce->engine);
+	intel_timeline_enter(ce->timeline);
+}
+
+void intel_context_exit_engine(struct intel_context *ce)
+{
+	intel_timeline_exit(ce->timeline);
+	intel_engine_pm_put(ce->engine);
+}
+
+int intel_context_prepare_remote_request(struct intel_context *ce,
+					 struct i915_request *rq)
+{
+	struct intel_timeline *tl = ce->timeline;
+	int err;
+
+	/* Only suitable for use in remotely modifying this context */
+	GEM_BUG_ON(rq->context == ce);
+
+	if (rcu_access_pointer(rq->timeline) != tl) { /* timeline sharing! */
+		/* Queue this switch after current activity by this context. */
+		err = i915_active_fence_set(&tl->last_request, rq);
+		if (err)
+			return err;
+	}
+
+	/*
+	 * Guarantee context image and the timeline remains pinned until the
+	 * modifying request is retired by setting the ce activity tracker.
+	 *
+	 * But we only need to take one pin on the account of it. Or in other
+	 * words transfer the pinned ce object to tracked active request.
+	 */
+	GEM_BUG_ON(i915_active_is_idle(&ce->active));
+	return i915_active_add_request(&ce->active, rq);
+}
+
+struct i915_request *intel_context_create_request(struct intel_context *ce)
+{
+	struct i915_gem_ww_ctx ww;
+	struct i915_request *rq;
+	int err;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	err = intel_context_pin_ww(ce, &ww);
+	if (!err) {
+		rq = i915_request_create(ce);
+		intel_context_unpin(ce);
+	} else if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+		rq = ERR_PTR(err);
+	} else {
+		rq = ERR_PTR(err);
+	}
+
+	i915_gem_ww_ctx_fini(&ww);
+
+	if (IS_ERR(rq))
+		return rq;
+
+	/*
+	 * timeline->mutex should be the inner lock, but is used as outer lock.
+	 * Hack around this to shut up lockdep in selftests..
+	 */
+	lockdep_unpin_lock(&ce->timeline->mutex, rq->cookie);
+	mutex_release(&ce->timeline->mutex.dep_map, _RET_IP_);
+	mutex_acquire(&ce->timeline->mutex.dep_map, SINGLE_DEPTH_NESTING, 0, _RET_IP_);
+	rq->cookie = lockdep_pin_lock(&ce->timeline->mutex);
+
+	return rq;
+}
+
+struct i915_request *intel_context_get_active_request(struct intel_context *ce)
+{
+	struct intel_context *parent = intel_context_to_parent(ce);
+	struct i915_request *rq, *active = NULL;
+	unsigned long flags;
+
+	GEM_BUG_ON(!intel_engine_uses_guc(ce->engine));
+
+	/*
+	 * We search the parent list to find an active request on the submitted
+	 * context. The parent list contains the requests for all the contexts
+	 * in the relationship so we have to do a compare of each request's
+	 * context.
+	 */
+	spin_lock_irqsave(&parent->guc_state.lock, flags);
+	list_for_each_entry_reverse(rq, &parent->guc_state.requests,
+				    sched.link) {
+		if (rq->context != ce)
+			continue;
+		if (i915_request_completed(rq))
+			break;
+
+		active = rq;
+	}
+	if (active)
+		active = i915_request_get_rcu(active);
+	spin_unlock_irqrestore(&parent->guc_state.lock, flags);
+
+	return active;
+}
+
+void intel_context_bind_parent_child(struct intel_context *parent,
+				     struct intel_context *child)
+{
+	/*
+	 * Callers responsibility to validate that this function is used
+	 * correctly but we use GEM_BUG_ON here ensure that they do.
+	 */
+	GEM_BUG_ON(intel_context_is_pinned(parent));
+	GEM_BUG_ON(intel_context_is_child(parent));
+	GEM_BUG_ON(intel_context_is_pinned(child));
+	GEM_BUG_ON(intel_context_is_child(child));
+	GEM_BUG_ON(intel_context_is_parent(child));
+
+	parent->parallel.child_index = parent->parallel.number_children++;
+	list_add_tail(&child->parallel.child_link,
+		      &parent->parallel.child_list);
+	child->parallel.parent = parent;
+}
+
+u64 intel_context_get_total_runtime_ns(const struct intel_context *ce)
+{
+	u64 total, active;
+
+	total = ce->stats.runtime.total;
+	if (ce->ops->flags & COPS_RUNTIME_CYCLES)
+		total *= ce->engine->gt->clock_period_ns;
+
+	active = READ_ONCE(ce->stats.active);
+	if (active)
+		active = intel_context_clock() - active;
+
+	return total + active;
+}
+
+u64 intel_context_get_avg_runtime_ns(struct intel_context *ce)
+{
+	u64 avg = ewma_runtime_read(&ce->stats.runtime.avg);
+
+	if (ce->ops->flags & COPS_RUNTIME_CYCLES)
+		avg *= ce->engine->gt->clock_period_ns;
+
+	return avg;
+}
+
+bool intel_context_ban(struct intel_context *ce, struct i915_request *rq)
+{
+	bool ret = intel_context_set_banned(ce);
+
+	trace_intel_context_ban(ce);
+
+	if (ce->ops->revoke)
+		ce->ops->revoke(ce, rq,
+				INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS);
+
+	return ret;
+}
+
+bool intel_context_revoke(struct intel_context *ce)
+{
+	bool ret = intel_context_set_exiting(ce);
+
+	if (ce->ops->revoke)
+		ce->ops->revoke(ce, NULL, ce->engine->props.preempt_timeout_ms);
+
+	return ret;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_context.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_context.h b/drivers/gpu/drm/i915/gt/intel_context.h
new file mode 100644
index 000000000..4ab6c8ddd
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_context.h
@@ -0,0 +1,372 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_CONTEXT_H__
+#define __INTEL_CONTEXT_H__
+
+#include <linux/bitops.h>
+#include <linux/lockdep.h>
+#include <linux/types.h>
+
+#include "i915_active.h"
+#include "i915_drv.h"
+#include "intel_context_types.h"
+#include "intel_engine_types.h"
+#include "intel_ring_types.h"
+#include "intel_timeline_types.h"
+#include "i915_trace.h"
+
+#define CE_TRACE(ce, fmt, ...) do {					\
+	const struct intel_context *ce__ = (ce);			\
+	ENGINE_TRACE(ce__->engine, "context:%llx " fmt,			\
+		     ce__->timeline->fence_context,			\
+		     ##__VA_ARGS__);					\
+} while (0)
+
+#define INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS (1)
+
+struct i915_gem_ww_ctx;
+
+void intel_context_init(struct intel_context *ce,
+			struct intel_engine_cs *engine);
+void intel_context_fini(struct intel_context *ce);
+
+void i915_context_module_exit(void);
+int i915_context_module_init(void);
+
+struct intel_context *
+intel_context_create(struct intel_engine_cs *engine);
+
+int intel_context_alloc_state(struct intel_context *ce);
+
+void intel_context_free(struct intel_context *ce);
+
+int intel_context_reconfigure_sseu(struct intel_context *ce,
+				   const struct intel_sseu sseu);
+
+#define PARENT_SCRATCH_SIZE	PAGE_SIZE
+
+static inline bool intel_context_is_child(struct intel_context *ce)
+{
+	return !!ce->parallel.parent;
+}
+
+static inline bool intel_context_is_parent(struct intel_context *ce)
+{
+	return !!ce->parallel.number_children;
+}
+
+static inline bool intel_context_is_pinned(struct intel_context *ce);
+
+static inline struct intel_context *
+intel_context_to_parent(struct intel_context *ce)
+{
+	if (intel_context_is_child(ce)) {
+		/*
+		 * The parent holds ref count to the child so it is always safe
+		 * for the parent to access the child, but the child has a
+		 * pointer to the parent without a ref. To ensure this is safe
+		 * the child should only access the parent pointer while the
+		 * parent is pinned.
+		 */
+		GEM_BUG_ON(!intel_context_is_pinned(ce->parallel.parent));
+
+		return ce->parallel.parent;
+	} else {
+		return ce;
+	}
+}
+
+static inline bool intel_context_is_parallel(struct intel_context *ce)
+{
+	return intel_context_is_child(ce) || intel_context_is_parent(ce);
+}
+
+void intel_context_bind_parent_child(struct intel_context *parent,
+				     struct intel_context *child);
+
+#define for_each_child(parent, ce)\
+	list_for_each_entry(ce, &(parent)->parallel.child_list,\
+			    parallel.child_link)
+#define for_each_child_safe(parent, ce, cn)\
+	list_for_each_entry_safe(ce, cn, &(parent)->parallel.child_list,\
+				 parallel.child_link)
+
+/**
+ * intel_context_lock_pinned - Stablises the 'pinned' status of the HW context
+ * @ce - the context
+ *
+ * Acquire a lock on the pinned status of the HW context, such that the context
+ * can neither be bound to the GPU or unbound whilst the lock is held, i.e.
+ * intel_context_is_pinned() remains stable.
+ */
+static inline int intel_context_lock_pinned(struct intel_context *ce)
+	__acquires(ce->pin_mutex)
+{
+	return mutex_lock_interruptible(&ce->pin_mutex);
+}
+
+/**
+ * intel_context_is_pinned - Reports the 'pinned' status
+ * @ce - the context
+ *
+ * While in use by the GPU, the context, along with its ring and page
+ * tables is pinned into memory and the GTT.
+ *
+ * Returns: true if the context is currently pinned for use by the GPU.
+ */
+static inline bool
+intel_context_is_pinned(struct intel_context *ce)
+{
+	return atomic_read(&ce->pin_count);
+}
+
+static inline void intel_context_cancel_request(struct intel_context *ce,
+						struct i915_request *rq)
+{
+	GEM_BUG_ON(!ce->ops->cancel_request);
+	return ce->ops->cancel_request(ce, rq);
+}
+
+/**
+ * intel_context_unlock_pinned - Releases the earlier locking of 'pinned' status
+ * @ce - the context
+ *
+ * Releases the lock earlier acquired by intel_context_unlock_pinned().
+ */
+static inline void intel_context_unlock_pinned(struct intel_context *ce)
+	__releases(ce->pin_mutex)
+{
+	mutex_unlock(&ce->pin_mutex);
+}
+
+int __intel_context_do_pin(struct intel_context *ce);
+int __intel_context_do_pin_ww(struct intel_context *ce,
+			      struct i915_gem_ww_ctx *ww);
+
+static inline bool intel_context_pin_if_active(struct intel_context *ce)
+{
+	return atomic_inc_not_zero(&ce->pin_count);
+}
+
+static inline int intel_context_pin(struct intel_context *ce)
+{
+	if (likely(intel_context_pin_if_active(ce)))
+		return 0;
+
+	return __intel_context_do_pin(ce);
+}
+
+static inline int intel_context_pin_ww(struct intel_context *ce,
+				       struct i915_gem_ww_ctx *ww)
+{
+	if (likely(intel_context_pin_if_active(ce)))
+		return 0;
+
+	return __intel_context_do_pin_ww(ce, ww);
+}
+
+static inline void __intel_context_pin(struct intel_context *ce)
+{
+	GEM_BUG_ON(!intel_context_is_pinned(ce));
+	atomic_inc(&ce->pin_count);
+}
+
+void __intel_context_do_unpin(struct intel_context *ce, int sub);
+
+static inline void intel_context_sched_disable_unpin(struct intel_context *ce)
+{
+	__intel_context_do_unpin(ce, 2);
+}
+
+static inline void intel_context_unpin(struct intel_context *ce)
+{
+	if (!ce->ops->sched_disable) {
+		__intel_context_do_unpin(ce, 1);
+	} else {
+		/*
+		 * Move ownership of this pin to the scheduling disable which is
+		 * an async operation. When that operation completes the above
+		 * intel_context_sched_disable_unpin is called potentially
+		 * unpinning the context.
+		 */
+		while (!atomic_add_unless(&ce->pin_count, -1, 1)) {
+			if (atomic_cmpxchg(&ce->pin_count, 1, 2) == 1) {
+				ce->ops->sched_disable(ce);
+				break;
+			}
+		}
+	}
+}
+
+void intel_context_enter_engine(struct intel_context *ce);
+void intel_context_exit_engine(struct intel_context *ce);
+
+static inline void intel_context_enter(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->timeline->mutex);
+	if (!ce->active_count++)
+		ce->ops->enter(ce);
+}
+
+static inline void intel_context_mark_active(struct intel_context *ce)
+{
+	lockdep_assert(lockdep_is_held(&ce->timeline->mutex) ||
+		       test_bit(CONTEXT_IS_PARKING, &ce->flags));
+	++ce->active_count;
+}
+
+static inline void intel_context_exit(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->timeline->mutex);
+	GEM_BUG_ON(!ce->active_count);
+	if (!--ce->active_count)
+		ce->ops->exit(ce);
+}
+
+static inline struct intel_context *intel_context_get(struct intel_context *ce)
+{
+	kref_get(&ce->ref);
+	return ce;
+}
+
+static inline void intel_context_put(struct intel_context *ce)
+{
+	kref_put(&ce->ref, ce->ops->destroy);
+}
+
+static inline struct intel_timeline *__must_check
+intel_context_timeline_lock(struct intel_context *ce)
+	__acquires(&ce->timeline->mutex)
+{
+	struct intel_timeline *tl = ce->timeline;
+	int err;
+
+	if (intel_context_is_parent(ce))
+		err = mutex_lock_interruptible_nested(&tl->mutex, 0);
+	else if (intel_context_is_child(ce))
+		err = mutex_lock_interruptible_nested(&tl->mutex,
+						      ce->parallel.child_index + 1);
+	else
+		err = mutex_lock_interruptible(&tl->mutex);
+	if (err)
+		return ERR_PTR(err);
+
+	return tl;
+}
+
+static inline void intel_context_timeline_unlock(struct intel_timeline *tl)
+	__releases(&tl->mutex)
+{
+	mutex_unlock(&tl->mutex);
+}
+
+int intel_context_prepare_remote_request(struct intel_context *ce,
+					 struct i915_request *rq);
+
+struct i915_request *intel_context_create_request(struct intel_context *ce);
+
+struct i915_request *intel_context_get_active_request(struct intel_context *ce);
+
+static inline bool intel_context_is_barrier(const struct intel_context *ce)
+{
+	return test_bit(CONTEXT_BARRIER_BIT, &ce->flags);
+}
+
+static inline bool intel_context_is_closed(const struct intel_context *ce)
+{
+	return test_bit(CONTEXT_CLOSED_BIT, &ce->flags);
+}
+
+static inline bool intel_context_has_inflight(const struct intel_context *ce)
+{
+	return test_bit(COPS_HAS_INFLIGHT_BIT, &ce->ops->flags);
+}
+
+static inline bool intel_context_use_semaphores(const struct intel_context *ce)
+{
+	return test_bit(CONTEXT_USE_SEMAPHORES, &ce->flags);
+}
+
+static inline void intel_context_set_use_semaphores(struct intel_context *ce)
+{
+	set_bit(CONTEXT_USE_SEMAPHORES, &ce->flags);
+}
+
+static inline void intel_context_clear_use_semaphores(struct intel_context *ce)
+{
+	clear_bit(CONTEXT_USE_SEMAPHORES, &ce->flags);
+}
+
+static inline bool intel_context_is_banned(const struct intel_context *ce)
+{
+	return test_bit(CONTEXT_BANNED, &ce->flags);
+}
+
+static inline bool intel_context_set_banned(struct intel_context *ce)
+{
+	return test_and_set_bit(CONTEXT_BANNED, &ce->flags);
+}
+
+bool intel_context_ban(struct intel_context *ce, struct i915_request *rq);
+
+static inline bool intel_context_is_schedulable(const struct intel_context *ce)
+{
+	return !test_bit(CONTEXT_EXITING, &ce->flags) &&
+	       !test_bit(CONTEXT_BANNED, &ce->flags);
+}
+
+static inline bool intel_context_is_exiting(const struct intel_context *ce)
+{
+	return test_bit(CONTEXT_EXITING, &ce->flags);
+}
+
+static inline bool intel_context_set_exiting(struct intel_context *ce)
+{
+	return test_and_set_bit(CONTEXT_EXITING, &ce->flags);
+}
+
+bool intel_context_revoke(struct intel_context *ce);
+
+static inline bool
+intel_context_force_single_submission(const struct intel_context *ce)
+{
+	return test_bit(CONTEXT_FORCE_SINGLE_SUBMISSION, &ce->flags);
+}
+
+static inline void
+intel_context_set_single_submission(struct intel_context *ce)
+{
+	__set_bit(CONTEXT_FORCE_SINGLE_SUBMISSION, &ce->flags);
+}
+
+static inline bool
+intel_context_nopreempt(const struct intel_context *ce)
+{
+	return test_bit(CONTEXT_NOPREEMPT, &ce->flags);
+}
+
+static inline void
+intel_context_set_nopreempt(struct intel_context *ce)
+{
+	set_bit(CONTEXT_NOPREEMPT, &ce->flags);
+}
+
+static inline void
+intel_context_clear_nopreempt(struct intel_context *ce)
+{
+	clear_bit(CONTEXT_NOPREEMPT, &ce->flags);
+}
+
+u64 intel_context_get_total_runtime_ns(const struct intel_context *ce);
+u64 intel_context_get_avg_runtime_ns(struct intel_context *ce);
+
+static inline u64 intel_context_clock(void)
+{
+	/* As we mix CS cycles with CPU clocks, use the raw monotonic clock. */
+	return ktime_get_raw_fast_ns();
+}
+
+#endif /* __INTEL_CONTEXT_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_context_param.h b/drivers/gpu/drm/i915/gt/intel_context_param.h
new file mode 100644
index 000000000..0c69cb42d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_context_param.h
@@ -0,0 +1,19 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_CONTEXT_PARAM_H
+#define INTEL_CONTEXT_PARAM_H
+
+#include <linux/types.h>
+
+#include "intel_context.h"
+
+static inline void
+intel_context_set_watchdog_us(struct intel_context *ce, u64 timeout_us)
+{
+	ce->watchdog.timeout_us = timeout_us;
+}
+
+#endif /* INTEL_CONTEXT_PARAM_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_context_sseu.c b/drivers/gpu/drm/i915/gt/intel_context_sseu.c
new file mode 100644
index 000000000..ece16c2b5
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_context_sseu.c
@@ -0,0 +1,97 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_vma.h"
+#include "intel_context.h"
+#include "intel_engine_pm.h"
+#include "intel_gpu_commands.h"
+#include "intel_lrc.h"
+#include "intel_lrc_reg.h"
+#include "intel_ring.h"
+#include "intel_sseu.h"
+
+static int gen8_emit_rpcs_config(struct i915_request *rq,
+				 const struct intel_context *ce,
+				 const struct intel_sseu sseu)
+{
+	u64 offset;
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	offset = i915_ggtt_offset(ce->state) +
+		 LRC_STATE_OFFSET + CTX_R_PWR_CLK_STATE * 4;
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+	*cs++ = intel_sseu_make_rpcs(rq->engine->gt, &sseu);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int
+gen8_modify_rpcs(struct intel_context *ce, const struct intel_sseu sseu)
+{
+	struct i915_request *rq;
+	int ret;
+
+	lockdep_assert_held(&ce->pin_mutex);
+
+	/*
+	 * If the context is not idle, we have to submit an ordered request to
+	 * modify its context image via the kernel context (writing to our own
+	 * image, or into the registers directory, does not stick). Pristine
+	 * and idle contexts will be configured on pinning.
+	 */
+	if (!intel_context_pin_if_active(ce))
+		return 0;
+
+	rq = intel_engine_create_kernel_request(ce->engine);
+	if (IS_ERR(rq)) {
+		ret = PTR_ERR(rq);
+		goto out_unpin;
+	}
+
+	/* Serialise with the remote context */
+	ret = intel_context_prepare_remote_request(ce, rq);
+	if (ret == 0)
+		ret = gen8_emit_rpcs_config(rq, ce, sseu);
+
+	i915_request_add(rq);
+out_unpin:
+	intel_context_unpin(ce);
+	return ret;
+}
+
+int
+intel_context_reconfigure_sseu(struct intel_context *ce,
+			       const struct intel_sseu sseu)
+{
+	int ret;
+
+	GEM_BUG_ON(GRAPHICS_VER(ce->engine->i915) < 8);
+
+	ret = intel_context_lock_pinned(ce);
+	if (ret)
+		return ret;
+
+	/* Nothing to do if unmodified. */
+	if (!memcmp(&ce->sseu, &sseu, sizeof(sseu)))
+		goto unlock;
+
+	ret = gen8_modify_rpcs(ce, sseu);
+	if (!ret)
+		ce->sseu = sseu;
+
+unlock:
+	intel_context_unlock_pinned(ce);
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h b/drivers/gpu/drm/i915/gt/intel_context_types.h
new file mode 100644
index 000000000..04eacae1a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_context_types.h
@@ -0,0 +1,316 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_CONTEXT_TYPES__
+#define __INTEL_CONTEXT_TYPES__
+
+#include <linux/average.h>
+#include <linux/kref.h>
+#include <linux/list.h>
+#include <linux/mutex.h>
+#include <linux/types.h>
+
+#include "i915_active_types.h"
+#include "i915_sw_fence.h"
+#include "i915_utils.h"
+#include "intel_engine_types.h"
+#include "intel_sseu.h"
+
+#include "uc/intel_guc_fwif.h"
+
+#define CONTEXT_REDZONE POISON_INUSE
+DECLARE_EWMA(runtime, 3, 8);
+
+struct i915_gem_context;
+struct i915_gem_ww_ctx;
+struct i915_vma;
+struct intel_breadcrumbs;
+struct intel_context;
+struct intel_ring;
+
+struct intel_context_ops {
+	unsigned long flags;
+#define COPS_HAS_INFLIGHT_BIT 0
+#define COPS_HAS_INFLIGHT BIT(COPS_HAS_INFLIGHT_BIT)
+
+#define COPS_RUNTIME_CYCLES_BIT 1
+#define COPS_RUNTIME_CYCLES BIT(COPS_RUNTIME_CYCLES_BIT)
+
+	int (*alloc)(struct intel_context *ce);
+
+	void (*revoke)(struct intel_context *ce, struct i915_request *rq,
+		       unsigned int preempt_timeout_ms);
+
+	int (*pre_pin)(struct intel_context *ce, struct i915_gem_ww_ctx *ww, void **vaddr);
+	int (*pin)(struct intel_context *ce, void *vaddr);
+	void (*unpin)(struct intel_context *ce);
+	void (*post_unpin)(struct intel_context *ce);
+
+	void (*cancel_request)(struct intel_context *ce,
+			       struct i915_request *rq);
+
+	void (*enter)(struct intel_context *ce);
+	void (*exit)(struct intel_context *ce);
+
+	void (*sched_disable)(struct intel_context *ce);
+
+	void (*reset)(struct intel_context *ce);
+	void (*destroy)(struct kref *kref);
+
+	/* virtual/parallel engine/context interface */
+	struct intel_context *(*create_virtual)(struct intel_engine_cs **engine,
+						unsigned int count,
+						unsigned long flags);
+	struct intel_context *(*create_parallel)(struct intel_engine_cs **engines,
+						 unsigned int num_siblings,
+						 unsigned int width);
+	struct intel_engine_cs *(*get_sibling)(struct intel_engine_cs *engine,
+					       unsigned int sibling);
+};
+
+struct intel_context {
+	/*
+	 * Note: Some fields may be accessed under RCU.
+	 *
+	 * Unless otherwise noted a field can safely be assumed to be protected
+	 * by strong reference counting.
+	 */
+	union {
+		struct kref ref; /* no kref_get_unless_zero()! */
+		struct rcu_head rcu;
+	};
+
+	struct intel_engine_cs *engine;
+	struct intel_engine_cs *inflight;
+#define __intel_context_inflight(engine) ptr_mask_bits(engine, 3)
+#define __intel_context_inflight_count(engine) ptr_unmask_bits(engine, 3)
+#define intel_context_inflight(ce) \
+	__intel_context_inflight(READ_ONCE((ce)->inflight))
+#define intel_context_inflight_count(ce) \
+	__intel_context_inflight_count(READ_ONCE((ce)->inflight))
+
+	struct i915_address_space *vm;
+	struct i915_gem_context __rcu *gem_context;
+
+	/*
+	 * @signal_lock protects the list of requests that need signaling,
+	 * @signals. While there are any requests that need signaling,
+	 * we add the context to the breadcrumbs worker, and remove it
+	 * upon completion/cancellation of the last request.
+	 */
+	struct list_head signal_link; /* Accessed under RCU */
+	struct list_head signals; /* Guarded by signal_lock */
+	spinlock_t signal_lock; /* protects signals, the list of requests */
+
+	struct i915_vma *state;
+	u32 ring_size;
+	struct intel_ring *ring;
+	struct intel_timeline *timeline;
+
+	unsigned long flags;
+#define CONTEXT_BARRIER_BIT		0
+#define CONTEXT_ALLOC_BIT		1
+#define CONTEXT_INIT_BIT		2
+#define CONTEXT_VALID_BIT		3
+#define CONTEXT_CLOSED_BIT		4
+#define CONTEXT_USE_SEMAPHORES		5
+#define CONTEXT_BANNED			6
+#define CONTEXT_FORCE_SINGLE_SUBMISSION	7
+#define CONTEXT_NOPREEMPT		8
+#define CONTEXT_LRCA_DIRTY		9
+#define CONTEXT_GUC_INIT		10
+#define CONTEXT_PERMA_PIN		11
+#define CONTEXT_IS_PARKING		12
+#define CONTEXT_EXITING			13
+
+	struct {
+		u64 timeout_us;
+	} watchdog;
+
+	u32 *lrc_reg_state;
+	union {
+		struct {
+			u32 lrca;
+			u32 ccid;
+		};
+		u64 desc;
+	} lrc;
+	u32 tag; /* cookie passed to HW to track this context on submission */
+
+	/** stats: Context GPU engine busyness tracking. */
+	struct intel_context_stats {
+		u64 active;
+
+		/* Time on GPU as tracked by the hw. */
+		struct {
+			struct ewma_runtime avg;
+			u64 total;
+			u32 last;
+			I915_SELFTEST_DECLARE(u32 num_underflow);
+			I915_SELFTEST_DECLARE(u32 max_underflow);
+		} runtime;
+	} stats;
+
+	unsigned int active_count; /* protected by timeline->mutex */
+
+	atomic_t pin_count;
+	struct mutex pin_mutex; /* guards pinning and associated on-gpuing */
+
+	/**
+	 * active: Active tracker for the rq activity (inc. external) on this
+	 * intel_context object.
+	 */
+	struct i915_active active;
+
+	const struct intel_context_ops *ops;
+
+	/** sseu: Control eu/slice partitioning */
+	struct intel_sseu sseu;
+
+	/**
+	 * pinned_contexts_link: List link for the engine's pinned contexts.
+	 * This is only used if this is a perma-pinned kernel context and
+	 * the list is assumed to only be manipulated during driver load
+	 * or unload time so no mutex protection currently.
+	 */
+	struct list_head pinned_contexts_link;
+
+	u8 wa_bb_page; /* if set, page num reserved for context workarounds */
+
+	struct {
+		/** @lock: protects everything in guc_state */
+		spinlock_t lock;
+		/**
+		 * @sched_state: scheduling state of this context using GuC
+		 * submission
+		 */
+		u32 sched_state;
+		/*
+		 * @fences: maintains a list of requests that are currently
+		 * being fenced until a GuC operation completes
+		 */
+		struct list_head fences;
+		/**
+		 * @blocked: fence used to signal when the blocking of a
+		 * context's submissions is complete.
+		 */
+		struct i915_sw_fence blocked;
+		/** @number_committed_requests: number of committed requests */
+		int number_committed_requests;
+		/** @requests: list of active requests on this context */
+		struct list_head requests;
+		/** @prio: the context's current guc priority */
+		u8 prio;
+		/**
+		 * @prio_count: a counter of the number requests in flight in
+		 * each priority bucket
+		 */
+		u32 prio_count[GUC_CLIENT_PRIORITY_NUM];
+	} guc_state;
+
+	struct {
+		/**
+		 * @id: handle which is used to uniquely identify this context
+		 * with the GuC, protected by guc->submission_state.lock
+		 */
+		u16 id;
+		/**
+		 * @ref: the number of references to the guc_id, when
+		 * transitioning in and out of zero protected by
+		 * guc->submission_state.lock
+		 */
+		atomic_t ref;
+		/**
+		 * @link: in guc->guc_id_list when the guc_id has no refs but is
+		 * still valid, protected by guc->submission_state.lock
+		 */
+		struct list_head link;
+	} guc_id;
+
+	/**
+	 * @destroyed_link: link in guc->submission_state.destroyed_contexts, in
+	 * list when context is pending to be destroyed (deregistered with the
+	 * GuC), protected by guc->submission_state.lock
+	 */
+	struct list_head destroyed_link;
+
+	/** @parallel: sub-structure for parallel submission members */
+	struct {
+		union {
+			/**
+			 * @child_list: parent's list of children
+			 * contexts, no protection as immutable after context
+			 * creation
+			 */
+			struct list_head child_list;
+			/**
+			 * @child_link: child's link into parent's list of
+			 * children
+			 */
+			struct list_head child_link;
+		};
+		/** @parent: pointer to parent if child */
+		struct intel_context *parent;
+		/**
+		 * @last_rq: last request submitted on a parallel context, used
+		 * to insert submit fences between requests in the parallel
+		 * context
+		 */
+		struct i915_request *last_rq;
+		/**
+		 * @fence_context: fence context composite fence when doing
+		 * parallel submission
+		 */
+		u64 fence_context;
+		/**
+		 * @seqno: seqno for composite fence when doing parallel
+		 * submission
+		 */
+		u32 seqno;
+		/** @number_children: number of children if parent */
+		u8 number_children;
+		/** @child_index: index into child_list if child */
+		u8 child_index;
+		/** @guc: GuC specific members for parallel submission */
+		struct {
+			/** @wqi_head: cached head pointer in work queue */
+			u16 wqi_head;
+			/** @wqi_tail: cached tail pointer in work queue */
+			u16 wqi_tail;
+			/** @wq_head: pointer to the actual head in work queue */
+			u32 *wq_head;
+			/** @wq_tail: pointer to the actual head in work queue */
+			u32 *wq_tail;
+			/** @wq_status: pointer to the status in work queue */
+			u32 *wq_status;
+
+			/**
+			 * @parent_page: page in context state (ce->state) used
+			 * by parent for work queue, process descriptor
+			 */
+			u8 parent_page;
+		} guc;
+	} parallel;
+
+#ifdef CONFIG_DRM_I915_SELFTEST
+	/**
+	 * @drop_schedule_enable: Force drop of schedule enable G2H for selftest
+	 */
+	bool drop_schedule_enable;
+
+	/**
+	 * @drop_schedule_disable: Force drop of schedule disable G2H for
+	 * selftest
+	 */
+	bool drop_schedule_disable;
+
+	/**
+	 * @drop_deregister: Force drop of deregister G2H for selftest
+	 */
+	bool drop_deregister;
+#endif
+};
+
+#endif /* __INTEL_CONTEXT_TYPES__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_engine.h b/drivers/gpu/drm/i915/gt/intel_engine.h
new file mode 100644
index 000000000..7a4504ea3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine.h
@@ -0,0 +1,357 @@
+/* SPDX-License-Identifier: MIT */
+#ifndef _INTEL_RINGBUFFER_H_
+#define _INTEL_RINGBUFFER_H_
+
+#include <asm/cacheflush.h>
+#include <drm/drm_util.h>
+#include <drm/drm_cache.h>
+
+#include <linux/hashtable.h>
+#include <linux/irq_work.h>
+#include <linux/random.h>
+#include <linux/seqlock.h>
+
+#include "i915_pmu.h"
+#include "i915_request.h"
+#include "i915_selftest.h"
+#include "intel_engine_types.h"
+#include "intel_gt_types.h"
+#include "intel_timeline.h"
+#include "intel_workarounds.h"
+
+struct drm_printer;
+struct intel_context;
+struct intel_gt;
+struct lock_class_key;
+
+/* Early gen2 devices have a cacheline of just 32 bytes, using 64 is overkill,
+ * but keeps the logic simple. Indeed, the whole purpose of this macro is just
+ * to give some inclination as to some of the magic values used in the various
+ * workarounds!
+ */
+#define CACHELINE_BYTES 64
+#define CACHELINE_DWORDS (CACHELINE_BYTES / sizeof(u32))
+
+#define ENGINE_TRACE(e, fmt, ...) do {					\
+	const struct intel_engine_cs *e__ __maybe_unused = (e);		\
+	GEM_TRACE("%s %s: " fmt,					\
+		  dev_name(e__->i915->drm.dev), e__->name,		\
+		  ##__VA_ARGS__);					\
+} while (0)
+
+/*
+ * The register defines to be used with the following macros need to accept a
+ * base param, e.g:
+ *
+ * REG_FOO(base) _MMIO((base) + <relative offset>)
+ * ENGINE_READ(engine, REG_FOO);
+ *
+ * register arrays are to be defined and accessed as follows:
+ *
+ * REG_BAR(base, i) _MMIO((base) + <relative offset> + (i) * <shift>)
+ * ENGINE_READ_IDX(engine, REG_BAR, i)
+ */
+
+#define __ENGINE_REG_OP(op__, engine__, ...) \
+	intel_uncore_##op__((engine__)->uncore, __VA_ARGS__)
+
+#define __ENGINE_READ_OP(op__, engine__, reg__) \
+	__ENGINE_REG_OP(op__, (engine__), reg__((engine__)->mmio_base))
+
+#define ENGINE_READ16(...)	__ENGINE_READ_OP(read16, __VA_ARGS__)
+#define ENGINE_READ(...)	__ENGINE_READ_OP(read, __VA_ARGS__)
+#define ENGINE_READ_FW(...)	__ENGINE_READ_OP(read_fw, __VA_ARGS__)
+#define ENGINE_POSTING_READ(...) __ENGINE_READ_OP(posting_read_fw, __VA_ARGS__)
+#define ENGINE_POSTING_READ16(...) __ENGINE_READ_OP(posting_read16, __VA_ARGS__)
+
+#define ENGINE_READ64(engine__, lower_reg__, upper_reg__) \
+	__ENGINE_REG_OP(read64_2x32, (engine__), \
+			lower_reg__((engine__)->mmio_base), \
+			upper_reg__((engine__)->mmio_base))
+
+#define ENGINE_READ_IDX(engine__, reg__, idx__) \
+	__ENGINE_REG_OP(read, (engine__), reg__((engine__)->mmio_base, (idx__)))
+
+#define __ENGINE_WRITE_OP(op__, engine__, reg__, val__) \
+	__ENGINE_REG_OP(op__, (engine__), reg__((engine__)->mmio_base), (val__))
+
+#define ENGINE_WRITE16(...)	__ENGINE_WRITE_OP(write16, __VA_ARGS__)
+#define ENGINE_WRITE(...)	__ENGINE_WRITE_OP(write, __VA_ARGS__)
+#define ENGINE_WRITE_FW(...)	__ENGINE_WRITE_OP(write_fw, __VA_ARGS__)
+
+#define GEN6_RING_FAULT_REG_READ(engine__) \
+	intel_uncore_read((engine__)->uncore, RING_FAULT_REG(engine__))
+
+#define GEN6_RING_FAULT_REG_POSTING_READ(engine__) \
+	intel_uncore_posting_read((engine__)->uncore, RING_FAULT_REG(engine__))
+
+#define GEN6_RING_FAULT_REG_RMW(engine__, clear__, set__) \
+({ \
+	u32 __val; \
+\
+	__val = intel_uncore_read((engine__)->uncore, \
+				  RING_FAULT_REG(engine__)); \
+	__val &= ~(clear__); \
+	__val |= (set__); \
+	intel_uncore_write((engine__)->uncore, RING_FAULT_REG(engine__), \
+			   __val); \
+})
+
+/* seqno size is actually only a uint32, but since we plan to use MI_FLUSH_DW to
+ * do the writes, and that must have qw aligned offsets, simply pretend it's 8b.
+ */
+
+static inline unsigned int
+execlists_num_ports(const struct intel_engine_execlists * const execlists)
+{
+	return execlists->port_mask + 1;
+}
+
+static inline struct i915_request *
+execlists_active(const struct intel_engine_execlists *execlists)
+{
+	struct i915_request * const *cur, * const *old, *active;
+
+	cur = READ_ONCE(execlists->active);
+	smp_rmb(); /* pairs with overwrite protection in process_csb() */
+	do {
+		old = cur;
+
+		active = READ_ONCE(*cur);
+		cur = READ_ONCE(execlists->active);
+
+		smp_rmb(); /* and complete the seqlock retry */
+	} while (unlikely(cur != old));
+
+	return active;
+}
+
+struct i915_request *
+execlists_unwind_incomplete_requests(struct intel_engine_execlists *execlists);
+
+static inline u32
+intel_read_status_page(const struct intel_engine_cs *engine, int reg)
+{
+	/* Ensure that the compiler doesn't optimize away the load. */
+	return READ_ONCE(engine->status_page.addr[reg]);
+}
+
+static inline void
+intel_write_status_page(struct intel_engine_cs *engine, int reg, u32 value)
+{
+	/* Writing into the status page should be done sparingly. Since
+	 * we do when we are uncertain of the device state, we take a bit
+	 * of extra paranoia to try and ensure that the HWS takes the value
+	 * we give and that it doesn't end up trapped inside the CPU!
+	 */
+	drm_clflush_virt_range(&engine->status_page.addr[reg], sizeof(value));
+	WRITE_ONCE(engine->status_page.addr[reg], value);
+	drm_clflush_virt_range(&engine->status_page.addr[reg], sizeof(value));
+}
+
+/*
+ * Reads a dword out of the status page, which is written to from the command
+ * queue by automatic updates, MI_REPORT_HEAD, MI_STORE_DATA_INDEX, or
+ * MI_STORE_DATA_IMM.
+ *
+ * The following dwords have a reserved meaning:
+ * 0x00: ISR copy, updated when an ISR bit not set in the HWSTAM changes.
+ * 0x04: ring 0 head pointer
+ * 0x05: ring 1 head pointer (915-class)
+ * 0x06: ring 2 head pointer (915-class)
+ * 0x10-0x1b: Context status DWords (GM45)
+ * 0x1f: Last written status offset. (GM45)
+ * 0x20-0x2f: Reserved (Gen6+)
+ *
+ * The area from dword 0x30 to 0x3ff is available for driver usage.
+ */
+#define I915_GEM_HWS_PREEMPT		0x32
+#define I915_GEM_HWS_PREEMPT_ADDR	(I915_GEM_HWS_PREEMPT * sizeof(u32))
+#define I915_GEM_HWS_SEQNO		0x40
+#define I915_GEM_HWS_SEQNO_ADDR		(I915_GEM_HWS_SEQNO * sizeof(u32))
+#define I915_GEM_HWS_MIGRATE		(0x42 * sizeof(u32))
+#define I915_GEM_HWS_PXP		0x60
+#define I915_GEM_HWS_PXP_ADDR		(I915_GEM_HWS_PXP * sizeof(u32))
+#define I915_GEM_HWS_SCRATCH		0x80
+
+#define I915_HWS_CSB_BUF0_INDEX		0x10
+#define I915_HWS_CSB_WRITE_INDEX	0x1f
+#define ICL_HWS_CSB_WRITE_INDEX		0x2f
+#define INTEL_HWS_CSB_WRITE_INDEX(__i915) \
+	(GRAPHICS_VER(__i915) >= 11 ? ICL_HWS_CSB_WRITE_INDEX : I915_HWS_CSB_WRITE_INDEX)
+
+void intel_engine_stop(struct intel_engine_cs *engine);
+void intel_engine_cleanup(struct intel_engine_cs *engine);
+
+int intel_engines_init_mmio(struct intel_gt *gt);
+int intel_engines_init(struct intel_gt *gt);
+
+void intel_engine_free_request_pool(struct intel_engine_cs *engine);
+
+void intel_engines_release(struct intel_gt *gt);
+void intel_engines_free(struct intel_gt *gt);
+
+int intel_engine_init_common(struct intel_engine_cs *engine);
+void intel_engine_cleanup_common(struct intel_engine_cs *engine);
+
+int intel_engine_resume(struct intel_engine_cs *engine);
+
+int intel_ring_submission_setup(struct intel_engine_cs *engine);
+
+int intel_engine_stop_cs(struct intel_engine_cs *engine);
+void intel_engine_cancel_stop_cs(struct intel_engine_cs *engine);
+
+void intel_engine_wait_for_pending_mi_fw(struct intel_engine_cs *engine);
+
+void intel_engine_set_hwsp_writemask(struct intel_engine_cs *engine, u32 mask);
+
+u64 intel_engine_get_active_head(const struct intel_engine_cs *engine);
+u64 intel_engine_get_last_batch_head(const struct intel_engine_cs *engine);
+
+void intel_engine_get_instdone(const struct intel_engine_cs *engine,
+			       struct intel_instdone *instdone);
+
+void intel_engine_init_execlists(struct intel_engine_cs *engine);
+
+bool intel_engine_irq_enable(struct intel_engine_cs *engine);
+void intel_engine_irq_disable(struct intel_engine_cs *engine);
+
+static inline void __intel_engine_reset(struct intel_engine_cs *engine,
+					bool stalled)
+{
+	if (engine->reset.rewind)
+		engine->reset.rewind(engine, stalled);
+	engine->serial++; /* contexts lost */
+}
+
+bool intel_engines_are_idle(struct intel_gt *gt);
+bool intel_engine_is_idle(struct intel_engine_cs *engine);
+
+void __intel_engine_flush_submission(struct intel_engine_cs *engine, bool sync);
+static inline void intel_engine_flush_submission(struct intel_engine_cs *engine)
+{
+	__intel_engine_flush_submission(engine, true);
+}
+
+void intel_engines_reset_default_submission(struct intel_gt *gt);
+
+bool intel_engine_can_store_dword(struct intel_engine_cs *engine);
+
+__printf(3, 4)
+void intel_engine_dump(struct intel_engine_cs *engine,
+		       struct drm_printer *m,
+		       const char *header, ...);
+void intel_engine_dump_active_requests(struct list_head *requests,
+				       struct i915_request *hung_rq,
+				       struct drm_printer *m);
+
+ktime_t intel_engine_get_busy_time(struct intel_engine_cs *engine,
+				   ktime_t *now);
+
+void intel_engine_get_hung_entity(struct intel_engine_cs *engine,
+				  struct intel_context **ce, struct i915_request **rq);
+
+u32 intel_engine_context_size(struct intel_gt *gt, u8 class);
+struct intel_context *
+intel_engine_create_pinned_context(struct intel_engine_cs *engine,
+				   struct i915_address_space *vm,
+				   unsigned int ring_size,
+				   unsigned int hwsp,
+				   struct lock_class_key *key,
+				   const char *name);
+
+void intel_engine_destroy_pinned_context(struct intel_context *ce);
+
+void xehp_enable_ccs_engines(struct intel_engine_cs *engine);
+
+#define ENGINE_PHYSICAL	0
+#define ENGINE_MOCK	1
+#define ENGINE_VIRTUAL	2
+
+static inline bool intel_engine_uses_guc(const struct intel_engine_cs *engine)
+{
+	return engine->gt->submission_method >= INTEL_SUBMISSION_GUC;
+}
+
+static inline bool
+intel_engine_has_preempt_reset(const struct intel_engine_cs *engine)
+{
+	if (!CONFIG_DRM_I915_PREEMPT_TIMEOUT)
+		return false;
+
+	return intel_engine_has_preemption(engine);
+}
+
+#define FORCE_VIRTUAL	BIT(0)
+struct intel_context *
+intel_engine_create_virtual(struct intel_engine_cs **siblings,
+			    unsigned int count, unsigned long flags);
+
+static inline struct intel_context *
+intel_engine_create_parallel(struct intel_engine_cs **engines,
+			     unsigned int num_engines,
+			     unsigned int width)
+{
+	GEM_BUG_ON(!engines[0]->cops->create_parallel);
+	return engines[0]->cops->create_parallel(engines, num_engines, width);
+}
+
+static inline bool
+intel_virtual_engine_has_heartbeat(const struct intel_engine_cs *engine)
+{
+	/*
+	 * For non-GuC submission we expect the back-end to look at the
+	 * heartbeat status of the actual physical engine that the work
+	 * has been (or is being) scheduled on, so we should only reach
+	 * here with GuC submission enabled.
+	 */
+	GEM_BUG_ON(!intel_engine_uses_guc(engine));
+
+	return intel_guc_virtual_engine_has_heartbeat(engine);
+}
+
+static inline bool
+intel_engine_has_heartbeat(const struct intel_engine_cs *engine)
+{
+	if (!CONFIG_DRM_I915_HEARTBEAT_INTERVAL)
+		return false;
+
+	if (intel_engine_is_virtual(engine))
+		return intel_virtual_engine_has_heartbeat(engine);
+	else
+		return READ_ONCE(engine->props.heartbeat_interval_ms);
+}
+
+static inline struct intel_engine_cs *
+intel_engine_get_sibling(struct intel_engine_cs *engine, unsigned int sibling)
+{
+	GEM_BUG_ON(!intel_engine_is_virtual(engine));
+	return engine->cops->get_sibling(engine, sibling);
+}
+
+static inline void
+intel_engine_set_hung_context(struct intel_engine_cs *engine,
+			      struct intel_context *ce)
+{
+	engine->hung_ce = ce;
+}
+
+static inline void
+intel_engine_clear_hung_context(struct intel_engine_cs *engine)
+{
+	intel_engine_set_hung_context(engine, NULL);
+}
+
+static inline struct intel_context *
+intel_engine_get_hung_context(struct intel_engine_cs *engine)
+{
+	return engine->hung_ce;
+}
+
+u64 intel_clamp_heartbeat_interval_ms(struct intel_engine_cs *engine, u64 value);
+u64 intel_clamp_max_busywait_duration_ns(struct intel_engine_cs *engine, u64 value);
+u64 intel_clamp_preempt_timeout_ms(struct intel_engine_cs *engine, u64 value);
+u64 intel_clamp_stop_timeout_ms(struct intel_engine_cs *engine, u64 value);
+u64 intel_clamp_timeslice_duration_ms(struct intel_engine_cs *engine, u64 value);
+
+#endif /* _INTEL_RINGBUFFER_H_ */
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
new file mode 100644
index 000000000..07967adce
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
@@ -0,0 +1,2392 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+
+#include <drm/drm_print.h>
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_internal.h"
+#include "gt/intel_gt_regs.h"
+
+#include "i915_cmd_parser.h"
+#include "i915_drv.h"
+#include "intel_breadcrumbs.h"
+#include "intel_context.h"
+#include "intel_engine.h"
+#include "intel_engine_pm.h"
+#include "intel_engine_regs.h"
+#include "intel_engine_user.h"
+#include "intel_execlists_submission.h"
+#include "intel_gt.h"
+#include "intel_gt_mcr.h"
+#include "intel_gt_pm.h"
+#include "intel_gt_requests.h"
+#include "intel_lrc.h"
+#include "intel_lrc_reg.h"
+#include "intel_reset.h"
+#include "intel_ring.h"
+#include "uc/intel_guc_submission.h"
+
+/* Haswell does have the CXT_SIZE register however it does not appear to be
+ * valid. Now, docs explain in dwords what is in the context object. The full
+ * size is 70720 bytes, however, the power context and execlist context will
+ * never be saved (power context is stored elsewhere, and execlists don't work
+ * on HSW) - so the final size, including the extra state required for the
+ * Resource Streamer, is 66944 bytes, which rounds to 17 pages.
+ */
+#define HSW_CXT_TOTAL_SIZE		(17 * PAGE_SIZE)
+
+#define DEFAULT_LR_CONTEXT_RENDER_SIZE	(22 * PAGE_SIZE)
+#define GEN8_LR_CONTEXT_RENDER_SIZE	(20 * PAGE_SIZE)
+#define GEN9_LR_CONTEXT_RENDER_SIZE	(22 * PAGE_SIZE)
+#define GEN11_LR_CONTEXT_RENDER_SIZE	(14 * PAGE_SIZE)
+
+#define GEN8_LR_CONTEXT_OTHER_SIZE	( 2 * PAGE_SIZE)
+
+#define MAX_MMIO_BASES 3
+struct engine_info {
+	u8 class;
+	u8 instance;
+	/* mmio bases table *must* be sorted in reverse graphics_ver order */
+	struct engine_mmio_base {
+		u32 graphics_ver : 8;
+		u32 base : 24;
+	} mmio_bases[MAX_MMIO_BASES];
+};
+
+static const struct engine_info intel_engines[] = {
+	[RCS0] = {
+		.class = RENDER_CLASS,
+		.instance = 0,
+		.mmio_bases = {
+			{ .graphics_ver = 1, .base = RENDER_RING_BASE }
+		},
+	},
+	[BCS0] = {
+		.class = COPY_ENGINE_CLASS,
+		.instance = 0,
+		.mmio_bases = {
+			{ .graphics_ver = 6, .base = BLT_RING_BASE }
+		},
+	},
+	[BCS1] = {
+		.class = COPY_ENGINE_CLASS,
+		.instance = 1,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHPC_BCS1_RING_BASE }
+		},
+	},
+	[BCS2] = {
+		.class = COPY_ENGINE_CLASS,
+		.instance = 2,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHPC_BCS2_RING_BASE }
+		},
+	},
+	[BCS3] = {
+		.class = COPY_ENGINE_CLASS,
+		.instance = 3,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHPC_BCS3_RING_BASE }
+		},
+	},
+	[BCS4] = {
+		.class = COPY_ENGINE_CLASS,
+		.instance = 4,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHPC_BCS4_RING_BASE }
+		},
+	},
+	[BCS5] = {
+		.class = COPY_ENGINE_CLASS,
+		.instance = 5,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHPC_BCS5_RING_BASE }
+		},
+	},
+	[BCS6] = {
+		.class = COPY_ENGINE_CLASS,
+		.instance = 6,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHPC_BCS6_RING_BASE }
+		},
+	},
+	[BCS7] = {
+		.class = COPY_ENGINE_CLASS,
+		.instance = 7,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHPC_BCS7_RING_BASE }
+		},
+	},
+	[BCS8] = {
+		.class = COPY_ENGINE_CLASS,
+		.instance = 8,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHPC_BCS8_RING_BASE }
+		},
+	},
+	[VCS0] = {
+		.class = VIDEO_DECODE_CLASS,
+		.instance = 0,
+		.mmio_bases = {
+			{ .graphics_ver = 11, .base = GEN11_BSD_RING_BASE },
+			{ .graphics_ver = 6, .base = GEN6_BSD_RING_BASE },
+			{ .graphics_ver = 4, .base = BSD_RING_BASE }
+		},
+	},
+	[VCS1] = {
+		.class = VIDEO_DECODE_CLASS,
+		.instance = 1,
+		.mmio_bases = {
+			{ .graphics_ver = 11, .base = GEN11_BSD2_RING_BASE },
+			{ .graphics_ver = 8, .base = GEN8_BSD2_RING_BASE }
+		},
+	},
+	[VCS2] = {
+		.class = VIDEO_DECODE_CLASS,
+		.instance = 2,
+		.mmio_bases = {
+			{ .graphics_ver = 11, .base = GEN11_BSD3_RING_BASE }
+		},
+	},
+	[VCS3] = {
+		.class = VIDEO_DECODE_CLASS,
+		.instance = 3,
+		.mmio_bases = {
+			{ .graphics_ver = 11, .base = GEN11_BSD4_RING_BASE }
+		},
+	},
+	[VCS4] = {
+		.class = VIDEO_DECODE_CLASS,
+		.instance = 4,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHP_BSD5_RING_BASE }
+		},
+	},
+	[VCS5] = {
+		.class = VIDEO_DECODE_CLASS,
+		.instance = 5,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHP_BSD6_RING_BASE }
+		},
+	},
+	[VCS6] = {
+		.class = VIDEO_DECODE_CLASS,
+		.instance = 6,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHP_BSD7_RING_BASE }
+		},
+	},
+	[VCS7] = {
+		.class = VIDEO_DECODE_CLASS,
+		.instance = 7,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHP_BSD8_RING_BASE }
+		},
+	},
+	[VECS0] = {
+		.class = VIDEO_ENHANCEMENT_CLASS,
+		.instance = 0,
+		.mmio_bases = {
+			{ .graphics_ver = 11, .base = GEN11_VEBOX_RING_BASE },
+			{ .graphics_ver = 7, .base = VEBOX_RING_BASE }
+		},
+	},
+	[VECS1] = {
+		.class = VIDEO_ENHANCEMENT_CLASS,
+		.instance = 1,
+		.mmio_bases = {
+			{ .graphics_ver = 11, .base = GEN11_VEBOX2_RING_BASE }
+		},
+	},
+	[VECS2] = {
+		.class = VIDEO_ENHANCEMENT_CLASS,
+		.instance = 2,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHP_VEBOX3_RING_BASE }
+		},
+	},
+	[VECS3] = {
+		.class = VIDEO_ENHANCEMENT_CLASS,
+		.instance = 3,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = XEHP_VEBOX4_RING_BASE }
+		},
+	},
+	[CCS0] = {
+		.class = COMPUTE_CLASS,
+		.instance = 0,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = GEN12_COMPUTE0_RING_BASE }
+		}
+	},
+	[CCS1] = {
+		.class = COMPUTE_CLASS,
+		.instance = 1,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = GEN12_COMPUTE1_RING_BASE }
+		}
+	},
+	[CCS2] = {
+		.class = COMPUTE_CLASS,
+		.instance = 2,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = GEN12_COMPUTE2_RING_BASE }
+		}
+	},
+	[CCS3] = {
+		.class = COMPUTE_CLASS,
+		.instance = 3,
+		.mmio_bases = {
+			{ .graphics_ver = 12, .base = GEN12_COMPUTE3_RING_BASE }
+		}
+	},
+};
+
+/**
+ * intel_engine_context_size() - return the size of the context for an engine
+ * @gt: the gt
+ * @class: engine class
+ *
+ * Each engine class may require a different amount of space for a context
+ * image.
+ *
+ * Return: size (in bytes) of an engine class specific context image
+ *
+ * Note: this size includes the HWSP, which is part of the context image
+ * in LRC mode, but does not include the "shared data page" used with
+ * GuC submission. The caller should account for this if using the GuC.
+ */
+u32 intel_engine_context_size(struct intel_gt *gt, u8 class)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	u32 cxt_size;
+
+	BUILD_BUG_ON(I915_GTT_PAGE_SIZE != PAGE_SIZE);
+
+	switch (class) {
+	case COMPUTE_CLASS:
+		fallthrough;
+	case RENDER_CLASS:
+		switch (GRAPHICS_VER(gt->i915)) {
+		default:
+			MISSING_CASE(GRAPHICS_VER(gt->i915));
+			return DEFAULT_LR_CONTEXT_RENDER_SIZE;
+		case 12:
+		case 11:
+			return GEN11_LR_CONTEXT_RENDER_SIZE;
+		case 9:
+			return GEN9_LR_CONTEXT_RENDER_SIZE;
+		case 8:
+			return GEN8_LR_CONTEXT_RENDER_SIZE;
+		case 7:
+			if (IS_HASWELL(gt->i915))
+				return HSW_CXT_TOTAL_SIZE;
+
+			cxt_size = intel_uncore_read(uncore, GEN7_CXT_SIZE);
+			return round_up(GEN7_CXT_TOTAL_SIZE(cxt_size) * 64,
+					PAGE_SIZE);
+		case 6:
+			cxt_size = intel_uncore_read(uncore, CXT_SIZE);
+			return round_up(GEN6_CXT_TOTAL_SIZE(cxt_size) * 64,
+					PAGE_SIZE);
+		case 5:
+		case 4:
+			/*
+			 * There is a discrepancy here between the size reported
+			 * by the register and the size of the context layout
+			 * in the docs. Both are described as authorative!
+			 *
+			 * The discrepancy is on the order of a few cachelines,
+			 * but the total is under one page (4k), which is our
+			 * minimum allocation anyway so it should all come
+			 * out in the wash.
+			 */
+			cxt_size = intel_uncore_read(uncore, CXT_SIZE) + 1;
+			drm_dbg(&gt->i915->drm,
+				"graphics_ver = %d CXT_SIZE = %d bytes [0x%08x]\n",
+				GRAPHICS_VER(gt->i915), cxt_size * 64,
+				cxt_size - 1);
+			return round_up(cxt_size * 64, PAGE_SIZE);
+		case 3:
+		case 2:
+		/* For the special day when i810 gets merged. */
+		case 1:
+			return 0;
+		}
+		break;
+	default:
+		MISSING_CASE(class);
+		fallthrough;
+	case VIDEO_DECODE_CLASS:
+	case VIDEO_ENHANCEMENT_CLASS:
+	case COPY_ENGINE_CLASS:
+		if (GRAPHICS_VER(gt->i915) < 8)
+			return 0;
+		return GEN8_LR_CONTEXT_OTHER_SIZE;
+	}
+}
+
+static u32 __engine_mmio_base(struct drm_i915_private *i915,
+			      const struct engine_mmio_base *bases)
+{
+	int i;
+
+	for (i = 0; i < MAX_MMIO_BASES; i++)
+		if (GRAPHICS_VER(i915) >= bases[i].graphics_ver)
+			break;
+
+	GEM_BUG_ON(i == MAX_MMIO_BASES);
+	GEM_BUG_ON(!bases[i].base);
+
+	return bases[i].base;
+}
+
+static void __sprint_engine_name(struct intel_engine_cs *engine)
+{
+	/*
+	 * Before we know what the uABI name for this engine will be,
+	 * we still would like to keep track of this engine in the debug logs.
+	 * We throw in a ' here as a reminder that this isn't its final name.
+	 */
+	GEM_WARN_ON(snprintf(engine->name, sizeof(engine->name), "%s'%u",
+			     intel_engine_class_repr(engine->class),
+			     engine->instance) >= sizeof(engine->name));
+}
+
+void intel_engine_set_hwsp_writemask(struct intel_engine_cs *engine, u32 mask)
+{
+	/*
+	 * Though they added more rings on g4x/ilk, they did not add
+	 * per-engine HWSTAM until gen6.
+	 */
+	if (GRAPHICS_VER(engine->i915) < 6 && engine->class != RENDER_CLASS)
+		return;
+
+	if (GRAPHICS_VER(engine->i915) >= 3)
+		ENGINE_WRITE(engine, RING_HWSTAM, mask);
+	else
+		ENGINE_WRITE16(engine, RING_HWSTAM, mask);
+}
+
+static void intel_engine_sanitize_mmio(struct intel_engine_cs *engine)
+{
+	/* Mask off all writes into the unknown HWSP */
+	intel_engine_set_hwsp_writemask(engine, ~0u);
+}
+
+static void nop_irq_handler(struct intel_engine_cs *engine, u16 iir)
+{
+	GEM_DEBUG_WARN_ON(iir);
+}
+
+static u32 get_reset_domain(u8 ver, enum intel_engine_id id)
+{
+	u32 reset_domain;
+
+	if (ver >= 11) {
+		static const u32 engine_reset_domains[] = {
+			[RCS0]  = GEN11_GRDOM_RENDER,
+			[BCS0]  = GEN11_GRDOM_BLT,
+			[BCS1]  = XEHPC_GRDOM_BLT1,
+			[BCS2]  = XEHPC_GRDOM_BLT2,
+			[BCS3]  = XEHPC_GRDOM_BLT3,
+			[BCS4]  = XEHPC_GRDOM_BLT4,
+			[BCS5]  = XEHPC_GRDOM_BLT5,
+			[BCS6]  = XEHPC_GRDOM_BLT6,
+			[BCS7]  = XEHPC_GRDOM_BLT7,
+			[BCS8]  = XEHPC_GRDOM_BLT8,
+			[VCS0]  = GEN11_GRDOM_MEDIA,
+			[VCS1]  = GEN11_GRDOM_MEDIA2,
+			[VCS2]  = GEN11_GRDOM_MEDIA3,
+			[VCS3]  = GEN11_GRDOM_MEDIA4,
+			[VCS4]  = GEN11_GRDOM_MEDIA5,
+			[VCS5]  = GEN11_GRDOM_MEDIA6,
+			[VCS6]  = GEN11_GRDOM_MEDIA7,
+			[VCS7]  = GEN11_GRDOM_MEDIA8,
+			[VECS0] = GEN11_GRDOM_VECS,
+			[VECS1] = GEN11_GRDOM_VECS2,
+			[VECS2] = GEN11_GRDOM_VECS3,
+			[VECS3] = GEN11_GRDOM_VECS4,
+			[CCS0]  = GEN11_GRDOM_RENDER,
+			[CCS1]  = GEN11_GRDOM_RENDER,
+			[CCS2]  = GEN11_GRDOM_RENDER,
+			[CCS3]  = GEN11_GRDOM_RENDER,
+		};
+		GEM_BUG_ON(id >= ARRAY_SIZE(engine_reset_domains) ||
+			   !engine_reset_domains[id]);
+		reset_domain = engine_reset_domains[id];
+	} else {
+		static const u32 engine_reset_domains[] = {
+			[RCS0]  = GEN6_GRDOM_RENDER,
+			[BCS0]  = GEN6_GRDOM_BLT,
+			[VCS0]  = GEN6_GRDOM_MEDIA,
+			[VCS1]  = GEN8_GRDOM_MEDIA2,
+			[VECS0] = GEN6_GRDOM_VECS,
+		};
+		GEM_BUG_ON(id >= ARRAY_SIZE(engine_reset_domains) ||
+			   !engine_reset_domains[id]);
+		reset_domain = engine_reset_domains[id];
+	}
+
+	return reset_domain;
+}
+
+static int intel_engine_setup(struct intel_gt *gt, enum intel_engine_id id,
+			      u8 logical_instance)
+{
+	const struct engine_info *info = &intel_engines[id];
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_engine_cs *engine;
+	u8 guc_class;
+
+	BUILD_BUG_ON(MAX_ENGINE_CLASS >= BIT(GEN11_ENGINE_CLASS_WIDTH));
+	BUILD_BUG_ON(MAX_ENGINE_INSTANCE >= BIT(GEN11_ENGINE_INSTANCE_WIDTH));
+	BUILD_BUG_ON(I915_MAX_VCS > (MAX_ENGINE_INSTANCE + 1));
+	BUILD_BUG_ON(I915_MAX_VECS > (MAX_ENGINE_INSTANCE + 1));
+
+	if (GEM_DEBUG_WARN_ON(id >= ARRAY_SIZE(gt->engine)))
+		return -EINVAL;
+
+	if (GEM_DEBUG_WARN_ON(info->class > MAX_ENGINE_CLASS))
+		return -EINVAL;
+
+	if (GEM_DEBUG_WARN_ON(info->instance > MAX_ENGINE_INSTANCE))
+		return -EINVAL;
+
+	if (GEM_DEBUG_WARN_ON(gt->engine_class[info->class][info->instance]))
+		return -EINVAL;
+
+	engine = kzalloc(sizeof(*engine), GFP_KERNEL);
+	if (!engine)
+		return -ENOMEM;
+
+	BUILD_BUG_ON(BITS_PER_TYPE(engine->mask) < I915_NUM_ENGINES);
+
+	INIT_LIST_HEAD(&engine->pinned_contexts_list);
+	engine->id = id;
+	engine->legacy_idx = INVALID_ENGINE;
+	engine->mask = BIT(id);
+	engine->reset_domain = get_reset_domain(GRAPHICS_VER(gt->i915),
+						id);
+	engine->i915 = i915;
+	engine->gt = gt;
+	engine->uncore = gt->uncore;
+	guc_class = engine_class_to_guc_class(info->class);
+	engine->guc_id = MAKE_GUC_ID(guc_class, info->instance);
+	engine->mmio_base = __engine_mmio_base(i915, info->mmio_bases);
+
+	engine->irq_handler = nop_irq_handler;
+
+	engine->class = info->class;
+	engine->instance = info->instance;
+	engine->logical_mask = BIT(logical_instance);
+	__sprint_engine_name(engine);
+
+	if ((engine->class == COMPUTE_CLASS && !RCS_MASK(engine->gt) &&
+	     __ffs(CCS_MASK(engine->gt)) == engine->instance) ||
+	     engine->class == RENDER_CLASS)
+		engine->flags |= I915_ENGINE_FIRST_RENDER_COMPUTE;
+
+	/* features common between engines sharing EUs */
+	if (engine->class == RENDER_CLASS || engine->class == COMPUTE_CLASS) {
+		engine->flags |= I915_ENGINE_HAS_RCS_REG_STATE;
+		engine->flags |= I915_ENGINE_HAS_EU_PRIORITY;
+	}
+
+	engine->props.heartbeat_interval_ms =
+		CONFIG_DRM_I915_HEARTBEAT_INTERVAL;
+	engine->props.max_busywait_duration_ns =
+		CONFIG_DRM_I915_MAX_REQUEST_BUSYWAIT;
+	engine->props.preempt_timeout_ms =
+		CONFIG_DRM_I915_PREEMPT_TIMEOUT;
+	engine->props.stop_timeout_ms =
+		CONFIG_DRM_I915_STOP_TIMEOUT;
+	engine->props.timeslice_duration_ms =
+		CONFIG_DRM_I915_TIMESLICE_DURATION;
+
+	/* Override to uninterruptible for OpenCL workloads. */
+	if (GRAPHICS_VER(i915) == 12 && (engine->flags & I915_ENGINE_HAS_RCS_REG_STATE))
+		engine->props.preempt_timeout_ms = 0;
+
+	/* Cap properties according to any system limits */
+#define CLAMP_PROP(field) \
+	do { \
+		u64 clamp = intel_clamp_##field(engine, engine->props.field); \
+		if (clamp != engine->props.field) { \
+			drm_notice(&engine->i915->drm, \
+				   "Warning, clamping %s to %lld to prevent overflow\n", \
+				   #field, clamp); \
+			engine->props.field = clamp; \
+		} \
+	} while (0)
+
+	CLAMP_PROP(heartbeat_interval_ms);
+	CLAMP_PROP(max_busywait_duration_ns);
+	CLAMP_PROP(preempt_timeout_ms);
+	CLAMP_PROP(stop_timeout_ms);
+	CLAMP_PROP(timeslice_duration_ms);
+
+#undef CLAMP_PROP
+
+	engine->defaults = engine->props; /* never to change again */
+
+	engine->context_size = intel_engine_context_size(gt, engine->class);
+	if (WARN_ON(engine->context_size > BIT(20)))
+		engine->context_size = 0;
+	if (engine->context_size)
+		DRIVER_CAPS(i915)->has_logical_contexts = true;
+
+	ewma__engine_latency_init(&engine->latency);
+
+	ATOMIC_INIT_NOTIFIER_HEAD(&engine->context_status_notifier);
+
+	/* Scrub mmio state on takeover */
+	intel_engine_sanitize_mmio(engine);
+
+	gt->engine_class[info->class][info->instance] = engine;
+	gt->engine[id] = engine;
+
+	return 0;
+}
+
+u64 intel_clamp_heartbeat_interval_ms(struct intel_engine_cs *engine, u64 value)
+{
+	value = min_t(u64, value, jiffies_to_msecs(MAX_SCHEDULE_TIMEOUT));
+
+	return value;
+}
+
+u64 intel_clamp_max_busywait_duration_ns(struct intel_engine_cs *engine, u64 value)
+{
+	value = min(value, jiffies_to_nsecs(2));
+
+	return value;
+}
+
+u64 intel_clamp_preempt_timeout_ms(struct intel_engine_cs *engine, u64 value)
+{
+	/*
+	 * NB: The GuC API only supports 32bit values. However, the limit is further
+	 * reduced due to internal calculations which would otherwise overflow.
+	 */
+	if (intel_guc_submission_is_wanted(&engine->gt->uc.guc))
+		value = min_t(u64, value, guc_policy_max_preempt_timeout_ms());
+
+	value = min_t(u64, value, jiffies_to_msecs(MAX_SCHEDULE_TIMEOUT));
+
+	return value;
+}
+
+u64 intel_clamp_stop_timeout_ms(struct intel_engine_cs *engine, u64 value)
+{
+	value = min_t(u64, value, jiffies_to_msecs(MAX_SCHEDULE_TIMEOUT));
+
+	return value;
+}
+
+u64 intel_clamp_timeslice_duration_ms(struct intel_engine_cs *engine, u64 value)
+{
+	/*
+	 * NB: The GuC API only supports 32bit values. However, the limit is further
+	 * reduced due to internal calculations which would otherwise overflow.
+	 */
+	if (intel_guc_submission_is_wanted(&engine->gt->uc.guc))
+		value = min_t(u64, value, guc_policy_max_exec_quantum_ms());
+
+	value = min_t(u64, value, jiffies_to_msecs(MAX_SCHEDULE_TIMEOUT));
+
+	return value;
+}
+
+static void __setup_engine_capabilities(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	if (engine->class == VIDEO_DECODE_CLASS) {
+		/*
+		 * HEVC support is present on first engine instance
+		 * before Gen11 and on all instances afterwards.
+		 */
+		if (GRAPHICS_VER(i915) >= 11 ||
+		    (GRAPHICS_VER(i915) >= 9 && engine->instance == 0))
+			engine->uabi_capabilities |=
+				I915_VIDEO_CLASS_CAPABILITY_HEVC;
+
+		/*
+		 * SFC block is present only on even logical engine
+		 * instances.
+		 */
+		if ((GRAPHICS_VER(i915) >= 11 &&
+		     (engine->gt->info.vdbox_sfc_access &
+		      BIT(engine->instance))) ||
+		    (GRAPHICS_VER(i915) >= 9 && engine->instance == 0))
+			engine->uabi_capabilities |=
+				I915_VIDEO_AND_ENHANCE_CLASS_CAPABILITY_SFC;
+	} else if (engine->class == VIDEO_ENHANCEMENT_CLASS) {
+		if (GRAPHICS_VER(i915) >= 9 &&
+		    engine->gt->info.sfc_mask & BIT(engine->instance))
+			engine->uabi_capabilities |=
+				I915_VIDEO_AND_ENHANCE_CLASS_CAPABILITY_SFC;
+	}
+}
+
+static void intel_setup_engine_capabilities(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id)
+		__setup_engine_capabilities(engine);
+}
+
+/**
+ * intel_engines_release() - free the resources allocated for Command Streamers
+ * @gt: pointer to struct intel_gt
+ */
+void intel_engines_release(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * Before we release the resources held by engine, we must be certain
+	 * that the HW is no longer accessing them -- having the GPU scribble
+	 * to or read from a page being used for something else causes no end
+	 * of fun.
+	 *
+	 * The GPU should be reset by this point, but assume the worst just
+	 * in case we aborted before completely initialising the engines.
+	 */
+	GEM_BUG_ON(intel_gt_pm_is_awake(gt));
+	if (!INTEL_INFO(gt->i915)->gpu_reset_clobbers_display)
+		__intel_gt_reset(gt, ALL_ENGINES);
+
+	/* Decouple the backend; but keep the layout for late GPU resets */
+	for_each_engine(engine, gt, id) {
+		if (!engine->release)
+			continue;
+
+		intel_wakeref_wait_for_idle(&engine->wakeref);
+		GEM_BUG_ON(intel_engine_pm_is_awake(engine));
+
+		engine->release(engine);
+		engine->release = NULL;
+
+		memset(&engine->reset, 0, sizeof(engine->reset));
+	}
+}
+
+void intel_engine_free_request_pool(struct intel_engine_cs *engine)
+{
+	if (!engine->request_pool)
+		return;
+
+	kmem_cache_free(i915_request_slab_cache(), engine->request_pool);
+}
+
+void intel_engines_free(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/* Free the requests! dma-resv keeps fences around for an eternity */
+	rcu_barrier();
+
+	for_each_engine(engine, gt, id) {
+		intel_engine_free_request_pool(engine);
+		kfree(engine);
+		gt->engine[id] = NULL;
+	}
+}
+
+static
+bool gen11_vdbox_has_sfc(struct intel_gt *gt,
+			 unsigned int physical_vdbox,
+			 unsigned int logical_vdbox, u16 vdbox_mask)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	/*
+	 * In Gen11, only even numbered logical VDBOXes are hooked
+	 * up to an SFC (Scaler & Format Converter) unit.
+	 * In Gen12, Even numbered physical instance always are connected
+	 * to an SFC. Odd numbered physical instances have SFC only if
+	 * previous even instance is fused off.
+	 *
+	 * Starting with Xe_HP, there's also a dedicated SFC_ENABLE field
+	 * in the fuse register that tells us whether a specific SFC is present.
+	 */
+	if ((gt->info.sfc_mask & BIT(physical_vdbox / 2)) == 0)
+		return false;
+	else if (MEDIA_VER(i915) >= 12)
+		return (physical_vdbox % 2 == 0) ||
+			!(BIT(physical_vdbox - 1) & vdbox_mask);
+	else if (MEDIA_VER(i915) == 11)
+		return logical_vdbox % 2 == 0;
+
+	return false;
+}
+
+static void engine_mask_apply_media_fuses(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	unsigned int logical_vdbox = 0;
+	unsigned int i;
+	u32 media_fuse, fuse1;
+	u16 vdbox_mask;
+	u16 vebox_mask;
+
+	if (MEDIA_VER(gt->i915) < 11)
+		return;
+
+	/*
+	 * On newer platforms the fusing register is called 'enable' and has
+	 * enable semantics, while on older platforms it is called 'disable'
+	 * and bits have disable semantices.
+	 */
+	media_fuse = intel_uncore_read(gt->uncore, GEN11_GT_VEBOX_VDBOX_DISABLE);
+	if (MEDIA_VER_FULL(i915) < IP_VER(12, 50))
+		media_fuse = ~media_fuse;
+
+	vdbox_mask = media_fuse & GEN11_GT_VDBOX_DISABLE_MASK;
+	vebox_mask = (media_fuse & GEN11_GT_VEBOX_DISABLE_MASK) >>
+		      GEN11_GT_VEBOX_DISABLE_SHIFT;
+
+	if (MEDIA_VER_FULL(i915) >= IP_VER(12, 50)) {
+		fuse1 = intel_uncore_read(gt->uncore, HSW_PAVP_FUSE1);
+		gt->info.sfc_mask = REG_FIELD_GET(XEHP_SFC_ENABLE_MASK, fuse1);
+	} else {
+		gt->info.sfc_mask = ~0;
+	}
+
+	for (i = 0; i < I915_MAX_VCS; i++) {
+		if (!HAS_ENGINE(gt, _VCS(i))) {
+			vdbox_mask &= ~BIT(i);
+			continue;
+		}
+
+		if (!(BIT(i) & vdbox_mask)) {
+			gt->info.engine_mask &= ~BIT(_VCS(i));
+			drm_dbg(&i915->drm, "vcs%u fused off\n", i);
+			continue;
+		}
+
+		if (gen11_vdbox_has_sfc(gt, i, logical_vdbox, vdbox_mask))
+			gt->info.vdbox_sfc_access |= BIT(i);
+		logical_vdbox++;
+	}
+	drm_dbg(&i915->drm, "vdbox enable: %04x, instances: %04lx\n",
+		vdbox_mask, VDBOX_MASK(gt));
+	GEM_BUG_ON(vdbox_mask != VDBOX_MASK(gt));
+
+	for (i = 0; i < I915_MAX_VECS; i++) {
+		if (!HAS_ENGINE(gt, _VECS(i))) {
+			vebox_mask &= ~BIT(i);
+			continue;
+		}
+
+		if (!(BIT(i) & vebox_mask)) {
+			gt->info.engine_mask &= ~BIT(_VECS(i));
+			drm_dbg(&i915->drm, "vecs%u fused off\n", i);
+		}
+	}
+	drm_dbg(&i915->drm, "vebox enable: %04x, instances: %04lx\n",
+		vebox_mask, VEBOX_MASK(gt));
+	GEM_BUG_ON(vebox_mask != VEBOX_MASK(gt));
+}
+
+static void engine_mask_apply_compute_fuses(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_gt_info *info = &gt->info;
+	int ss_per_ccs = info->sseu.max_subslices / I915_MAX_CCS;
+	unsigned long ccs_mask;
+	unsigned int i;
+
+	if (GRAPHICS_VER(i915) < 11)
+		return;
+
+	if (hweight32(CCS_MASK(gt)) <= 1)
+		return;
+
+	ccs_mask = intel_slicemask_from_xehp_dssmask(info->sseu.compute_subslice_mask,
+						     ss_per_ccs);
+	/*
+	 * If all DSS in a quadrant are fused off, the corresponding CCS
+	 * engine is not available for use.
+	 */
+	for_each_clear_bit(i, &ccs_mask, I915_MAX_CCS) {
+		info->engine_mask &= ~BIT(_CCS(i));
+		drm_dbg(&i915->drm, "ccs%u fused off\n", i);
+	}
+}
+
+static void engine_mask_apply_copy_fuses(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_gt_info *info = &gt->info;
+	unsigned long meml3_mask;
+	unsigned long quad;
+
+	if (!(GRAPHICS_VER_FULL(i915) >= IP_VER(12, 60) &&
+	      GRAPHICS_VER_FULL(i915) < IP_VER(12, 70)))
+		return;
+
+	meml3_mask = intel_uncore_read(gt->uncore, GEN10_MIRROR_FUSE3);
+	meml3_mask = REG_FIELD_GET(GEN12_MEML3_EN_MASK, meml3_mask);
+
+	/*
+	 * Link Copy engines may be fused off according to meml3_mask. Each
+	 * bit is a quad that houses 2 Link Copy and two Sub Copy engines.
+	 */
+	for_each_clear_bit(quad, &meml3_mask, GEN12_MAX_MSLICES) {
+		unsigned int instance = quad * 2 + 1;
+		intel_engine_mask_t mask = GENMASK(_BCS(instance + 1),
+						   _BCS(instance));
+
+		if (mask & info->engine_mask) {
+			drm_dbg(&i915->drm, "bcs%u fused off\n", instance);
+			drm_dbg(&i915->drm, "bcs%u fused off\n", instance + 1);
+
+			info->engine_mask &= ~mask;
+		}
+	}
+}
+
+/*
+ * Determine which engines are fused off in our particular hardware.
+ * Note that we have a catch-22 situation where we need to be able to access
+ * the blitter forcewake domain to read the engine fuses, but at the same time
+ * we need to know which engines are available on the system to know which
+ * forcewake domains are present. We solve this by intializing the forcewake
+ * domains based on the full engine mask in the platform capabilities before
+ * calling this function and pruning the domains for fused-off engines
+ * afterwards.
+ */
+static intel_engine_mask_t init_engine_mask(struct intel_gt *gt)
+{
+	struct intel_gt_info *info = &gt->info;
+
+	GEM_BUG_ON(!info->engine_mask);
+
+	engine_mask_apply_media_fuses(gt);
+	engine_mask_apply_compute_fuses(gt);
+	engine_mask_apply_copy_fuses(gt);
+
+	return info->engine_mask;
+}
+
+static void populate_logical_ids(struct intel_gt *gt, u8 *logical_ids,
+				 u8 class, const u8 *map, u8 num_instances)
+{
+	int i, j;
+	u8 current_logical_id = 0;
+
+	for (j = 0; j < num_instances; ++j) {
+		for (i = 0; i < ARRAY_SIZE(intel_engines); ++i) {
+			if (!HAS_ENGINE(gt, i) ||
+			    intel_engines[i].class != class)
+				continue;
+
+			if (intel_engines[i].instance == map[j]) {
+				logical_ids[intel_engines[i].instance] =
+					current_logical_id++;
+				break;
+			}
+		}
+	}
+}
+
+static void setup_logical_ids(struct intel_gt *gt, u8 *logical_ids, u8 class)
+{
+	/*
+	 * Logical to physical mapping is needed for proper support
+	 * to split-frame feature.
+	 */
+	if (MEDIA_VER(gt->i915) >= 11 && class == VIDEO_DECODE_CLASS) {
+		const u8 map[] = { 0, 2, 4, 6, 1, 3, 5, 7 };
+
+		populate_logical_ids(gt, logical_ids, class,
+				     map, ARRAY_SIZE(map));
+	} else {
+		int i;
+		u8 map[MAX_ENGINE_INSTANCE + 1];
+
+		for (i = 0; i < MAX_ENGINE_INSTANCE + 1; ++i)
+			map[i] = i;
+		populate_logical_ids(gt, logical_ids, class,
+				     map, ARRAY_SIZE(map));
+	}
+}
+
+/**
+ * intel_engines_init_mmio() - allocate and prepare the Engine Command Streamers
+ * @gt: pointer to struct intel_gt
+ *
+ * Return: non-zero if the initialization failed.
+ */
+int intel_engines_init_mmio(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	const unsigned int engine_mask = init_engine_mask(gt);
+	unsigned int mask = 0;
+	unsigned int i, class;
+	u8 logical_ids[MAX_ENGINE_INSTANCE + 1];
+	int err;
+
+	drm_WARN_ON(&i915->drm, engine_mask == 0);
+	drm_WARN_ON(&i915->drm, engine_mask &
+		    GENMASK(BITS_PER_TYPE(mask) - 1, I915_NUM_ENGINES));
+
+	if (i915_inject_probe_failure(i915))
+		return -ENODEV;
+
+	for (class = 0; class < MAX_ENGINE_CLASS + 1; ++class) {
+		setup_logical_ids(gt, logical_ids, class);
+
+		for (i = 0; i < ARRAY_SIZE(intel_engines); ++i) {
+			u8 instance = intel_engines[i].instance;
+
+			if (intel_engines[i].class != class ||
+			    !HAS_ENGINE(gt, i))
+				continue;
+
+			err = intel_engine_setup(gt, i,
+						 logical_ids[instance]);
+			if (err)
+				goto cleanup;
+
+			mask |= BIT(i);
+		}
+	}
+
+	/*
+	 * Catch failures to update intel_engines table when the new engines
+	 * are added to the driver by a warning and disabling the forgotten
+	 * engines.
+	 */
+	if (drm_WARN_ON(&i915->drm, mask != engine_mask))
+		gt->info.engine_mask = mask;
+
+	gt->info.num_engines = hweight32(mask);
+
+	intel_gt_check_and_clear_faults(gt);
+
+	intel_setup_engine_capabilities(gt);
+
+	intel_uncore_prune_engine_fw_domains(gt->uncore, gt);
+
+	return 0;
+
+cleanup:
+	intel_engines_free(gt);
+	return err;
+}
+
+void intel_engine_init_execlists(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+
+	execlists->port_mask = 1;
+	GEM_BUG_ON(!is_power_of_2(execlists_num_ports(execlists)));
+	GEM_BUG_ON(execlists_num_ports(execlists) > EXECLIST_MAX_PORTS);
+
+	memset(execlists->pending, 0, sizeof(execlists->pending));
+	execlists->active =
+		memset(execlists->inflight, 0, sizeof(execlists->inflight));
+}
+
+static void cleanup_status_page(struct intel_engine_cs *engine)
+{
+	struct i915_vma *vma;
+
+	/* Prevent writes into HWSP after returning the page to the system */
+	intel_engine_set_hwsp_writemask(engine, ~0u);
+
+	vma = fetch_and_zero(&engine->status_page.vma);
+	if (!vma)
+		return;
+
+	if (!HWS_NEEDS_PHYSICAL(engine->i915))
+		i915_vma_unpin(vma);
+
+	i915_gem_object_unpin_map(vma->obj);
+	i915_gem_object_put(vma->obj);
+}
+
+static int pin_ggtt_status_page(struct intel_engine_cs *engine,
+				struct i915_gem_ww_ctx *ww,
+				struct i915_vma *vma)
+{
+	unsigned int flags;
+
+	if (!HAS_LLC(engine->i915) && i915_ggtt_has_aperture(engine->gt->ggtt))
+		/*
+		 * On g33, we cannot place HWS above 256MiB, so
+		 * restrict its pinning to the low mappable arena.
+		 * Though this restriction is not documented for
+		 * gen4, gen5, or byt, they also behave similarly
+		 * and hang if the HWS is placed at the top of the
+		 * GTT. To generalise, it appears that all !llc
+		 * platforms have issues with us placing the HWS
+		 * above the mappable region (even though we never
+		 * actually map it).
+		 */
+		flags = PIN_MAPPABLE;
+	else
+		flags = PIN_HIGH;
+
+	return i915_ggtt_pin(vma, ww, 0, flags);
+}
+
+static int init_status_page(struct intel_engine_cs *engine)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_gem_ww_ctx ww;
+	struct i915_vma *vma;
+	void *vaddr;
+	int ret;
+
+	INIT_LIST_HEAD(&engine->status_page.timelines);
+
+	/*
+	 * Though the HWS register does support 36bit addresses, historically
+	 * we have had hangs and corruption reported due to wild writes if
+	 * the HWS is placed above 4G. We only allow objects to be allocated
+	 * in GFP_DMA32 for i965, and no earlier physical address users had
+	 * access to more than 4G.
+	 */
+	obj = i915_gem_object_create_internal(engine->i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		drm_err(&engine->i915->drm,
+			"Failed to allocate status page\n");
+		return PTR_ERR(obj);
+	}
+
+	i915_gem_object_set_cache_coherency(obj, I915_CACHE_LLC);
+
+	vma = i915_vma_instance(obj, &engine->gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto err_put;
+	}
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	ret = i915_gem_object_lock(obj, &ww);
+	if (!ret && !HWS_NEEDS_PHYSICAL(engine->i915))
+		ret = pin_ggtt_status_page(engine, &ww, vma);
+	if (ret)
+		goto err;
+
+	vaddr = i915_gem_object_pin_map(obj, I915_MAP_WB);
+	if (IS_ERR(vaddr)) {
+		ret = PTR_ERR(vaddr);
+		goto err_unpin;
+	}
+
+	engine->status_page.addr = memset(vaddr, 0, PAGE_SIZE);
+	engine->status_page.vma = vma;
+
+err_unpin:
+	if (ret)
+		i915_vma_unpin(vma);
+err:
+	if (ret == -EDEADLK) {
+		ret = i915_gem_ww_ctx_backoff(&ww);
+		if (!ret)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+err_put:
+	if (ret)
+		i915_gem_object_put(obj);
+	return ret;
+}
+
+static int engine_setup_common(struct intel_engine_cs *engine)
+{
+	int err;
+
+	init_llist_head(&engine->barrier_tasks);
+
+	err = init_status_page(engine);
+	if (err)
+		return err;
+
+	engine->breadcrumbs = intel_breadcrumbs_create(engine);
+	if (!engine->breadcrumbs) {
+		err = -ENOMEM;
+		goto err_status;
+	}
+
+	engine->sched_engine = i915_sched_engine_create(ENGINE_PHYSICAL);
+	if (!engine->sched_engine) {
+		err = -ENOMEM;
+		goto err_sched_engine;
+	}
+	engine->sched_engine->private_data = engine;
+
+	err = intel_engine_init_cmd_parser(engine);
+	if (err)
+		goto err_cmd_parser;
+
+	intel_engine_init_execlists(engine);
+	intel_engine_init__pm(engine);
+	intel_engine_init_retire(engine);
+
+	/* Use the whole device by default */
+	engine->sseu =
+		intel_sseu_from_device_info(&engine->gt->info.sseu);
+
+	intel_engine_init_workarounds(engine);
+	intel_engine_init_whitelist(engine);
+	intel_engine_init_ctx_wa(engine);
+
+	if (GRAPHICS_VER(engine->i915) >= 12)
+		engine->flags |= I915_ENGINE_HAS_RELATIVE_MMIO;
+
+	return 0;
+
+err_cmd_parser:
+	i915_sched_engine_put(engine->sched_engine);
+err_sched_engine:
+	intel_breadcrumbs_put(engine->breadcrumbs);
+err_status:
+	cleanup_status_page(engine);
+	return err;
+}
+
+struct measure_breadcrumb {
+	struct i915_request rq;
+	struct intel_ring ring;
+	u32 cs[2048];
+};
+
+static int measure_breadcrumb_dw(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct measure_breadcrumb *frame;
+	int dw;
+
+	GEM_BUG_ON(!engine->gt->scratch);
+
+	frame = kzalloc(sizeof(*frame), GFP_KERNEL);
+	if (!frame)
+		return -ENOMEM;
+
+	frame->rq.engine = engine;
+	frame->rq.context = ce;
+	rcu_assign_pointer(frame->rq.timeline, ce->timeline);
+	frame->rq.hwsp_seqno = ce->timeline->hwsp_seqno;
+
+	frame->ring.vaddr = frame->cs;
+	frame->ring.size = sizeof(frame->cs);
+	frame->ring.wrap =
+		BITS_PER_TYPE(frame->ring.size) - ilog2(frame->ring.size);
+	frame->ring.effective_size = frame->ring.size;
+	intel_ring_update_space(&frame->ring);
+	frame->rq.ring = &frame->ring;
+
+	mutex_lock(&ce->timeline->mutex);
+	spin_lock_irq(&engine->sched_engine->lock);
+
+	dw = engine->emit_fini_breadcrumb(&frame->rq, frame->cs) - frame->cs;
+
+	spin_unlock_irq(&engine->sched_engine->lock);
+	mutex_unlock(&ce->timeline->mutex);
+
+	GEM_BUG_ON(dw & 1); /* RING_TAIL must be qword aligned */
+
+	kfree(frame);
+	return dw;
+}
+
+struct intel_context *
+intel_engine_create_pinned_context(struct intel_engine_cs *engine,
+				   struct i915_address_space *vm,
+				   unsigned int ring_size,
+				   unsigned int hwsp,
+				   struct lock_class_key *key,
+				   const char *name)
+{
+	struct intel_context *ce;
+	int err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return ce;
+
+	__set_bit(CONTEXT_BARRIER_BIT, &ce->flags);
+	ce->timeline = page_pack_bits(NULL, hwsp);
+	ce->ring = NULL;
+	ce->ring_size = ring_size;
+
+	i915_vm_put(ce->vm);
+	ce->vm = i915_vm_get(vm);
+
+	err = intel_context_pin(ce); /* perma-pin so it is always available */
+	if (err) {
+		intel_context_put(ce);
+		return ERR_PTR(err);
+	}
+
+	list_add_tail(&ce->pinned_contexts_link, &engine->pinned_contexts_list);
+
+	/*
+	 * Give our perma-pinned kernel timelines a separate lockdep class,
+	 * so that we can use them from within the normal user timelines
+	 * should we need to inject GPU operations during their request
+	 * construction.
+	 */
+	lockdep_set_class_and_name(&ce->timeline->mutex, key, name);
+
+	return ce;
+}
+
+void intel_engine_destroy_pinned_context(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct i915_vma *hwsp = engine->status_page.vma;
+
+	GEM_BUG_ON(ce->timeline->hwsp_ggtt != hwsp);
+
+	mutex_lock(&hwsp->vm->mutex);
+	list_del(&ce->timeline->engine_link);
+	mutex_unlock(&hwsp->vm->mutex);
+
+	list_del(&ce->pinned_contexts_link);
+	intel_context_unpin(ce);
+	intel_context_put(ce);
+}
+
+static struct intel_context *
+create_kernel_context(struct intel_engine_cs *engine)
+{
+	static struct lock_class_key kernel;
+
+	return intel_engine_create_pinned_context(engine, engine->gt->vm, SZ_4K,
+						  I915_GEM_HWS_SEQNO_ADDR,
+						  &kernel, "kernel_context");
+}
+
+/**
+ * intel_engines_init_common - initialize cengine state which might require hw access
+ * @engine: Engine to initialize.
+ *
+ * Initializes @engine@ structure members shared between legacy and execlists
+ * submission modes which do require hardware access.
+ *
+ * Typcally done at later stages of submission mode specific engine setup.
+ *
+ * Returns zero on success or an error code on failure.
+ */
+static int engine_init_common(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+	int ret;
+
+	engine->set_default_submission(engine);
+
+	/*
+	 * We may need to do things with the shrinker which
+	 * require us to immediately switch back to the default
+	 * context. This can cause a problem as pinning the
+	 * default context also requires GTT space which may not
+	 * be available. To avoid this we always pin the default
+	 * context.
+	 */
+	ce = create_kernel_context(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	ret = measure_breadcrumb_dw(ce);
+	if (ret < 0)
+		goto err_context;
+
+	engine->emit_fini_breadcrumb_dw = ret;
+	engine->kernel_context = ce;
+
+	return 0;
+
+err_context:
+	intel_engine_destroy_pinned_context(ce);
+	return ret;
+}
+
+int intel_engines_init(struct intel_gt *gt)
+{
+	int (*setup)(struct intel_engine_cs *engine);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err;
+
+	if (intel_uc_uses_guc_submission(&gt->uc)) {
+		gt->submission_method = INTEL_SUBMISSION_GUC;
+		setup = intel_guc_submission_setup;
+	} else if (HAS_EXECLISTS(gt->i915)) {
+		gt->submission_method = INTEL_SUBMISSION_ELSP;
+		setup = intel_execlists_submission_setup;
+	} else {
+		gt->submission_method = INTEL_SUBMISSION_RING;
+		setup = intel_ring_submission_setup;
+	}
+
+	for_each_engine(engine, gt, id) {
+		err = engine_setup_common(engine);
+		if (err)
+			return err;
+
+		err = setup(engine);
+		if (err) {
+			intel_engine_cleanup_common(engine);
+			return err;
+		}
+
+		/* The backend should now be responsible for cleanup */
+		GEM_BUG_ON(engine->release == NULL);
+
+		err = engine_init_common(engine);
+		if (err)
+			return err;
+
+		intel_engine_add_user(engine);
+	}
+
+	return 0;
+}
+
+/**
+ * intel_engines_cleanup_common - cleans up the engine state created by
+ *                                the common initiailizers.
+ * @engine: Engine to cleanup.
+ *
+ * This cleans up everything created by the common helpers.
+ */
+void intel_engine_cleanup_common(struct intel_engine_cs *engine)
+{
+	GEM_BUG_ON(!list_empty(&engine->sched_engine->requests));
+
+	i915_sched_engine_put(engine->sched_engine);
+	intel_breadcrumbs_put(engine->breadcrumbs);
+
+	intel_engine_fini_retire(engine);
+	intel_engine_cleanup_cmd_parser(engine);
+
+	if (engine->default_state)
+		fput(engine->default_state);
+
+	if (engine->kernel_context)
+		intel_engine_destroy_pinned_context(engine->kernel_context);
+
+	GEM_BUG_ON(!llist_empty(&engine->barrier_tasks));
+	cleanup_status_page(engine);
+
+	intel_wa_list_free(&engine->ctx_wa_list);
+	intel_wa_list_free(&engine->wa_list);
+	intel_wa_list_free(&engine->whitelist);
+}
+
+/**
+ * intel_engine_resume - re-initializes the HW state of the engine
+ * @engine: Engine to resume.
+ *
+ * Returns zero on success or an error code on failure.
+ */
+int intel_engine_resume(struct intel_engine_cs *engine)
+{
+	intel_engine_apply_workarounds(engine);
+	intel_engine_apply_whitelist(engine);
+
+	return engine->resume(engine);
+}
+
+u64 intel_engine_get_active_head(const struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	u64 acthd;
+
+	if (GRAPHICS_VER(i915) >= 8)
+		acthd = ENGINE_READ64(engine, RING_ACTHD, RING_ACTHD_UDW);
+	else if (GRAPHICS_VER(i915) >= 4)
+		acthd = ENGINE_READ(engine, RING_ACTHD);
+	else
+		acthd = ENGINE_READ(engine, ACTHD);
+
+	return acthd;
+}
+
+u64 intel_engine_get_last_batch_head(const struct intel_engine_cs *engine)
+{
+	u64 bbaddr;
+
+	if (GRAPHICS_VER(engine->i915) >= 8)
+		bbaddr = ENGINE_READ64(engine, RING_BBADDR, RING_BBADDR_UDW);
+	else
+		bbaddr = ENGINE_READ(engine, RING_BBADDR);
+
+	return bbaddr;
+}
+
+static unsigned long stop_timeout(const struct intel_engine_cs *engine)
+{
+	if (in_atomic() || irqs_disabled()) /* inside atomic preempt-reset? */
+		return 0;
+
+	/*
+	 * If we are doing a normal GPU reset, we can take our time and allow
+	 * the engine to quiesce. We've stopped submission to the engine, and
+	 * if we wait long enough an innocent context should complete and
+	 * leave the engine idle. So they should not be caught unaware by
+	 * the forthcoming GPU reset (which usually follows the stop_cs)!
+	 */
+	return READ_ONCE(engine->props.stop_timeout_ms);
+}
+
+static int __intel_engine_stop_cs(struct intel_engine_cs *engine,
+				  int fast_timeout_us,
+				  int slow_timeout_ms)
+{
+	struct intel_uncore *uncore = engine->uncore;
+	const i915_reg_t mode = RING_MI_MODE(engine->mmio_base);
+	int err;
+
+	intel_uncore_write_fw(uncore, mode, _MASKED_BIT_ENABLE(STOP_RING));
+
+	/*
+	 * Wa_22011802037 : gen11, gen12, Prior to doing a reset, ensure CS is
+	 * stopped, set ring stop bit and prefetch disable bit to halt CS
+	 */
+	if (IS_GRAPHICS_VER(engine->i915, 11, 12))
+		intel_uncore_write_fw(uncore, RING_MODE_GEN7(engine->mmio_base),
+				      _MASKED_BIT_ENABLE(GEN12_GFX_PREFETCH_DISABLE));
+
+	err = __intel_wait_for_register_fw(engine->uncore, mode,
+					   MODE_IDLE, MODE_IDLE,
+					   fast_timeout_us,
+					   slow_timeout_ms,
+					   NULL);
+
+	/* A final mmio read to let GPU writes be hopefully flushed to memory */
+	intel_uncore_posting_read_fw(uncore, mode);
+	return err;
+}
+
+int intel_engine_stop_cs(struct intel_engine_cs *engine)
+{
+	int err = 0;
+
+	if (GRAPHICS_VER(engine->i915) < 3)
+		return -ENODEV;
+
+	ENGINE_TRACE(engine, "\n");
+	/*
+	 * TODO: Find out why occasionally stopping the CS times out. Seen
+	 * especially with gem_eio tests.
+	 *
+	 * Occasionally trying to stop the cs times out, but does not adversely
+	 * affect functionality. The timeout is set as a config parameter that
+	 * defaults to 100ms. In most cases the follow up operation is to wait
+	 * for pending MI_FORCE_WAKES. The assumption is that this timeout is
+	 * sufficient for any pending MI_FORCEWAKEs to complete. Once root
+	 * caused, the caller must check and handle the return from this
+	 * function.
+	 */
+	if (__intel_engine_stop_cs(engine, 1000, stop_timeout(engine))) {
+		ENGINE_TRACE(engine,
+			     "timed out on STOP_RING -> IDLE; HEAD:%04x, TAIL:%04x\n",
+			     ENGINE_READ_FW(engine, RING_HEAD) & HEAD_ADDR,
+			     ENGINE_READ_FW(engine, RING_TAIL) & TAIL_ADDR);
+
+		/*
+		 * Sometimes we observe that the idle flag is not
+		 * set even though the ring is empty. So double
+		 * check before giving up.
+		 */
+		if ((ENGINE_READ_FW(engine, RING_HEAD) & HEAD_ADDR) !=
+		    (ENGINE_READ_FW(engine, RING_TAIL) & TAIL_ADDR))
+			err = -ETIMEDOUT;
+	}
+
+	return err;
+}
+
+void intel_engine_cancel_stop_cs(struct intel_engine_cs *engine)
+{
+	ENGINE_TRACE(engine, "\n");
+
+	ENGINE_WRITE_FW(engine, RING_MI_MODE, _MASKED_BIT_DISABLE(STOP_RING));
+}
+
+static u32 __cs_pending_mi_force_wakes(struct intel_engine_cs *engine)
+{
+	static const i915_reg_t _reg[I915_NUM_ENGINES] = {
+		[RCS0] = MSG_IDLE_CS,
+		[BCS0] = MSG_IDLE_BCS,
+		[VCS0] = MSG_IDLE_VCS0,
+		[VCS1] = MSG_IDLE_VCS1,
+		[VCS2] = MSG_IDLE_VCS2,
+		[VCS3] = MSG_IDLE_VCS3,
+		[VCS4] = MSG_IDLE_VCS4,
+		[VCS5] = MSG_IDLE_VCS5,
+		[VCS6] = MSG_IDLE_VCS6,
+		[VCS7] = MSG_IDLE_VCS7,
+		[VECS0] = MSG_IDLE_VECS0,
+		[VECS1] = MSG_IDLE_VECS1,
+		[VECS2] = MSG_IDLE_VECS2,
+		[VECS3] = MSG_IDLE_VECS3,
+		[CCS0] = MSG_IDLE_CS,
+		[CCS1] = MSG_IDLE_CS,
+		[CCS2] = MSG_IDLE_CS,
+		[CCS3] = MSG_IDLE_CS,
+	};
+	u32 val;
+
+	if (!_reg[engine->id].reg) {
+		drm_err(&engine->i915->drm,
+			"MSG IDLE undefined for engine id %u\n", engine->id);
+		return 0;
+	}
+
+	val = intel_uncore_read(engine->uncore, _reg[engine->id]);
+
+	/* bits[29:25] & bits[13:9] >> shift */
+	return (val & (val >> 16) & MSG_IDLE_FW_MASK) >> MSG_IDLE_FW_SHIFT;
+}
+
+static void __gpm_wait_for_fw_complete(struct intel_gt *gt, u32 fw_mask)
+{
+	int ret;
+
+	/* Ensure GPM receives fw up/down after CS is stopped */
+	udelay(1);
+
+	/* Wait for forcewake request to complete in GPM */
+	ret =  __intel_wait_for_register_fw(gt->uncore,
+					    GEN9_PWRGT_DOMAIN_STATUS,
+					    fw_mask, fw_mask, 5000, 0, NULL);
+
+	/* Ensure CS receives fw ack from GPM */
+	udelay(1);
+
+	if (ret)
+		GT_TRACE(gt, "Failed to complete pending forcewake %d\n", ret);
+}
+
+/*
+ * Wa_22011802037:gen12: In addition to stopping the cs, we need to wait for any
+ * pending MI_FORCE_WAKEUP requests that the CS has initiated to complete. The
+ * pending status is indicated by bits[13:9] (masked by bits[29:25]) in the
+ * MSG_IDLE register. There's one MSG_IDLE register per reset domain. Since we
+ * are concerned only with the gt reset here, we use a logical OR of pending
+ * forcewakeups from all reset domains and then wait for them to complete by
+ * querying PWRGT_DOMAIN_STATUS.
+ */
+void intel_engine_wait_for_pending_mi_fw(struct intel_engine_cs *engine)
+{
+	u32 fw_pending = __cs_pending_mi_force_wakes(engine);
+
+	if (fw_pending)
+		__gpm_wait_for_fw_complete(engine->gt, fw_pending);
+}
+
+/* NB: please notice the memset */
+void intel_engine_get_instdone(const struct intel_engine_cs *engine,
+			       struct intel_instdone *instdone)
+{
+	struct drm_i915_private *i915 = engine->i915;
+	struct intel_uncore *uncore = engine->uncore;
+	u32 mmio_base = engine->mmio_base;
+	int slice;
+	int subslice;
+	int iter;
+
+	memset(instdone, 0, sizeof(*instdone));
+
+	if (GRAPHICS_VER(i915) >= 8) {
+		instdone->instdone =
+			intel_uncore_read(uncore, RING_INSTDONE(mmio_base));
+
+		if (engine->id != RCS0)
+			return;
+
+		instdone->slice_common =
+			intel_uncore_read(uncore, GEN7_SC_INSTDONE);
+		if (GRAPHICS_VER(i915) >= 12) {
+			instdone->slice_common_extra[0] =
+				intel_uncore_read(uncore, GEN12_SC_INSTDONE_EXTRA);
+			instdone->slice_common_extra[1] =
+				intel_uncore_read(uncore, GEN12_SC_INSTDONE_EXTRA2);
+		}
+
+		for_each_ss_steering(iter, engine->gt, slice, subslice) {
+			instdone->sampler[slice][subslice] =
+				intel_gt_mcr_read(engine->gt,
+						  GEN7_SAMPLER_INSTDONE,
+						  slice, subslice);
+			instdone->row[slice][subslice] =
+				intel_gt_mcr_read(engine->gt,
+						  GEN7_ROW_INSTDONE,
+						  slice, subslice);
+		}
+
+		if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 55)) {
+			for_each_ss_steering(iter, engine->gt, slice, subslice)
+				instdone->geom_svg[slice][subslice] =
+					intel_gt_mcr_read(engine->gt,
+							  XEHPG_INSTDONE_GEOM_SVG,
+							  slice, subslice);
+		}
+	} else if (GRAPHICS_VER(i915) >= 7) {
+		instdone->instdone =
+			intel_uncore_read(uncore, RING_INSTDONE(mmio_base));
+
+		if (engine->id != RCS0)
+			return;
+
+		instdone->slice_common =
+			intel_uncore_read(uncore, GEN7_SC_INSTDONE);
+		instdone->sampler[0][0] =
+			intel_uncore_read(uncore, GEN7_SAMPLER_INSTDONE);
+		instdone->row[0][0] =
+			intel_uncore_read(uncore, GEN7_ROW_INSTDONE);
+	} else if (GRAPHICS_VER(i915) >= 4) {
+		instdone->instdone =
+			intel_uncore_read(uncore, RING_INSTDONE(mmio_base));
+		if (engine->id == RCS0)
+			/* HACK: Using the wrong struct member */
+			instdone->slice_common =
+				intel_uncore_read(uncore, GEN4_INSTDONE1);
+	} else {
+		instdone->instdone = intel_uncore_read(uncore, GEN2_INSTDONE);
+	}
+}
+
+static bool ring_is_idle(struct intel_engine_cs *engine)
+{
+	bool idle = true;
+
+	if (I915_SELFTEST_ONLY(!engine->mmio_base))
+		return true;
+
+	if (!intel_engine_pm_get_if_awake(engine))
+		return true;
+
+	/* First check that no commands are left in the ring */
+	if ((ENGINE_READ(engine, RING_HEAD) & HEAD_ADDR) !=
+	    (ENGINE_READ(engine, RING_TAIL) & TAIL_ADDR))
+		idle = false;
+
+	/* No bit for gen2, so assume the CS parser is idle */
+	if (GRAPHICS_VER(engine->i915) > 2 &&
+	    !(ENGINE_READ(engine, RING_MI_MODE) & MODE_IDLE))
+		idle = false;
+
+	intel_engine_pm_put(engine);
+
+	return idle;
+}
+
+void __intel_engine_flush_submission(struct intel_engine_cs *engine, bool sync)
+{
+	struct tasklet_struct *t = &engine->sched_engine->tasklet;
+
+	if (!t->callback)
+		return;
+
+	local_bh_disable();
+	if (tasklet_trylock(t)) {
+		/* Must wait for any GPU reset in progress. */
+		if (__tasklet_is_enabled(t))
+			t->callback(t);
+		tasklet_unlock(t);
+	}
+	local_bh_enable();
+
+	/* Synchronise and wait for the tasklet on another CPU */
+	if (sync)
+		tasklet_unlock_wait(t);
+}
+
+/**
+ * intel_engine_is_idle() - Report if the engine has finished process all work
+ * @engine: the intel_engine_cs
+ *
+ * Return true if there are no requests pending, nothing left to be submitted
+ * to hardware, and that the engine is idle.
+ */
+bool intel_engine_is_idle(struct intel_engine_cs *engine)
+{
+	/* More white lies, if wedged, hw state is inconsistent */
+	if (intel_gt_is_wedged(engine->gt))
+		return true;
+
+	if (!intel_engine_pm_is_awake(engine))
+		return true;
+
+	/* Waiting to drain ELSP? */
+	intel_synchronize_hardirq(engine->i915);
+	intel_engine_flush_submission(engine);
+
+	/* ELSP is empty, but there are ready requests? E.g. after reset */
+	if (!i915_sched_engine_is_empty(engine->sched_engine))
+		return false;
+
+	/* Ring stopped? */
+	return ring_is_idle(engine);
+}
+
+bool intel_engines_are_idle(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * If the driver is wedged, HW state may be very inconsistent and
+	 * report that it is still busy, even though we have stopped using it.
+	 */
+	if (intel_gt_is_wedged(gt))
+		return true;
+
+	/* Already parked (and passed an idleness test); must still be idle */
+	if (!READ_ONCE(gt->awake))
+		return true;
+
+	for_each_engine(engine, gt, id) {
+		if (!intel_engine_is_idle(engine))
+			return false;
+	}
+
+	return true;
+}
+
+bool intel_engine_irq_enable(struct intel_engine_cs *engine)
+{
+	if (!engine->irq_enable)
+		return false;
+
+	/* Caller disables interrupts */
+	spin_lock(engine->gt->irq_lock);
+	engine->irq_enable(engine);
+	spin_unlock(engine->gt->irq_lock);
+
+	return true;
+}
+
+void intel_engine_irq_disable(struct intel_engine_cs *engine)
+{
+	if (!engine->irq_disable)
+		return;
+
+	/* Caller disables interrupts */
+	spin_lock(engine->gt->irq_lock);
+	engine->irq_disable(engine);
+	spin_unlock(engine->gt->irq_lock);
+}
+
+void intel_engines_reset_default_submission(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id) {
+		if (engine->sanitize)
+			engine->sanitize(engine);
+
+		engine->set_default_submission(engine);
+	}
+}
+
+bool intel_engine_can_store_dword(struct intel_engine_cs *engine)
+{
+	switch (GRAPHICS_VER(engine->i915)) {
+	case 2:
+		return false; /* uses physical not virtual addresses */
+	case 3:
+		/* maybe only uses physical not virtual addresses */
+		return !(IS_I915G(engine->i915) || IS_I915GM(engine->i915));
+	case 4:
+		return !IS_I965G(engine->i915); /* who knows! */
+	case 6:
+		return engine->class != VIDEO_DECODE_CLASS; /* b0rked */
+	default:
+		return true;
+	}
+}
+
+static struct intel_timeline *get_timeline(struct i915_request *rq)
+{
+	struct intel_timeline *tl;
+
+	/*
+	 * Even though we are holding the engine->sched_engine->lock here, there
+	 * is no control over the submission queue per-se and we are
+	 * inspecting the active state at a random point in time, with an
+	 * unknown queue. Play safe and make sure the timeline remains valid.
+	 * (Only being used for pretty printing, one extra kref shouldn't
+	 * cause a camel stampede!)
+	 */
+	rcu_read_lock();
+	tl = rcu_dereference(rq->timeline);
+	if (!kref_get_unless_zero(&tl->kref))
+		tl = NULL;
+	rcu_read_unlock();
+
+	return tl;
+}
+
+static int print_ring(char *buf, int sz, struct i915_request *rq)
+{
+	int len = 0;
+
+	if (!i915_request_signaled(rq)) {
+		struct intel_timeline *tl = get_timeline(rq);
+
+		len = scnprintf(buf, sz,
+				"ring:{start:%08x, hwsp:%08x, seqno:%08x, runtime:%llums}, ",
+				i915_ggtt_offset(rq->ring->vma),
+				tl ? tl->hwsp_offset : 0,
+				hwsp_seqno(rq),
+				DIV_ROUND_CLOSEST_ULL(intel_context_get_total_runtime_ns(rq->context),
+						      1000 * 1000));
+
+		if (tl)
+			intel_timeline_put(tl);
+	}
+
+	return len;
+}
+
+static void hexdump(struct drm_printer *m, const void *buf, size_t len)
+{
+	const size_t rowsize = 8 * sizeof(u32);
+	const void *prev = NULL;
+	bool skip = false;
+	size_t pos;
+
+	for (pos = 0; pos < len; pos += rowsize) {
+		char line[128];
+
+		if (prev && !memcmp(prev, buf + pos, rowsize)) {
+			if (!skip) {
+				drm_printf(m, "*\n");
+				skip = true;
+			}
+			continue;
+		}
+
+		WARN_ON_ONCE(hex_dump_to_buffer(buf + pos, len - pos,
+						rowsize, sizeof(u32),
+						line, sizeof(line),
+						false) >= sizeof(line));
+		drm_printf(m, "[%04zx] %s\n", pos, line);
+
+		prev = buf + pos;
+		skip = false;
+	}
+}
+
+static const char *repr_timer(const struct timer_list *t)
+{
+	if (!READ_ONCE(t->expires))
+		return "inactive";
+
+	if (timer_pending(t))
+		return "active";
+
+	return "expired";
+}
+
+static void intel_engine_print_registers(struct intel_engine_cs *engine,
+					 struct drm_printer *m)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	u64 addr;
+
+	if (engine->id == RENDER_CLASS && IS_GRAPHICS_VER(dev_priv, 4, 7))
+		drm_printf(m, "\tCCID: 0x%08x\n", ENGINE_READ(engine, CCID));
+	if (HAS_EXECLISTS(dev_priv)) {
+		drm_printf(m, "\tEL_STAT_HI: 0x%08x\n",
+			   ENGINE_READ(engine, RING_EXECLIST_STATUS_HI));
+		drm_printf(m, "\tEL_STAT_LO: 0x%08x\n",
+			   ENGINE_READ(engine, RING_EXECLIST_STATUS_LO));
+	}
+	drm_printf(m, "\tRING_START: 0x%08x\n",
+		   ENGINE_READ(engine, RING_START));
+	drm_printf(m, "\tRING_HEAD:  0x%08x\n",
+		   ENGINE_READ(engine, RING_HEAD) & HEAD_ADDR);
+	drm_printf(m, "\tRING_TAIL:  0x%08x\n",
+		   ENGINE_READ(engine, RING_TAIL) & TAIL_ADDR);
+	drm_printf(m, "\tRING_CTL:   0x%08x%s\n",
+		   ENGINE_READ(engine, RING_CTL),
+		   ENGINE_READ(engine, RING_CTL) & (RING_WAIT | RING_WAIT_SEMAPHORE) ? " [waiting]" : "");
+	if (GRAPHICS_VER(engine->i915) > 2) {
+		drm_printf(m, "\tRING_MODE:  0x%08x%s\n",
+			   ENGINE_READ(engine, RING_MI_MODE),
+			   ENGINE_READ(engine, RING_MI_MODE) & (MODE_IDLE) ? " [idle]" : "");
+	}
+
+	if (GRAPHICS_VER(dev_priv) >= 6) {
+		drm_printf(m, "\tRING_IMR:   0x%08x\n",
+			   ENGINE_READ(engine, RING_IMR));
+		drm_printf(m, "\tRING_ESR:   0x%08x\n",
+			   ENGINE_READ(engine, RING_ESR));
+		drm_printf(m, "\tRING_EMR:   0x%08x\n",
+			   ENGINE_READ(engine, RING_EMR));
+		drm_printf(m, "\tRING_EIR:   0x%08x\n",
+			   ENGINE_READ(engine, RING_EIR));
+	}
+
+	addr = intel_engine_get_active_head(engine);
+	drm_printf(m, "\tACTHD:  0x%08x_%08x\n",
+		   upper_32_bits(addr), lower_32_bits(addr));
+	addr = intel_engine_get_last_batch_head(engine);
+	drm_printf(m, "\tBBADDR: 0x%08x_%08x\n",
+		   upper_32_bits(addr), lower_32_bits(addr));
+	if (GRAPHICS_VER(dev_priv) >= 8)
+		addr = ENGINE_READ64(engine, RING_DMA_FADD, RING_DMA_FADD_UDW);
+	else if (GRAPHICS_VER(dev_priv) >= 4)
+		addr = ENGINE_READ(engine, RING_DMA_FADD);
+	else
+		addr = ENGINE_READ(engine, DMA_FADD_I8XX);
+	drm_printf(m, "\tDMA_FADDR: 0x%08x_%08x\n",
+		   upper_32_bits(addr), lower_32_bits(addr));
+	if (GRAPHICS_VER(dev_priv) >= 4) {
+		drm_printf(m, "\tIPEIR: 0x%08x\n",
+			   ENGINE_READ(engine, RING_IPEIR));
+		drm_printf(m, "\tIPEHR: 0x%08x\n",
+			   ENGINE_READ(engine, RING_IPEHR));
+	} else {
+		drm_printf(m, "\tIPEIR: 0x%08x\n", ENGINE_READ(engine, IPEIR));
+		drm_printf(m, "\tIPEHR: 0x%08x\n", ENGINE_READ(engine, IPEHR));
+	}
+
+	if (HAS_EXECLISTS(dev_priv) && !intel_engine_uses_guc(engine)) {
+		struct i915_request * const *port, *rq;
+		const u32 *hws =
+			&engine->status_page.addr[I915_HWS_CSB_BUF0_INDEX];
+		const u8 num_entries = execlists->csb_size;
+		unsigned int idx;
+		u8 read, write;
+
+		drm_printf(m, "\tExeclist tasklet queued? %s (%s), preempt? %s, timeslice? %s\n",
+			   str_yes_no(test_bit(TASKLET_STATE_SCHED, &engine->sched_engine->tasklet.state)),
+			   str_enabled_disabled(!atomic_read(&engine->sched_engine->tasklet.count)),
+			   repr_timer(&engine->execlists.preempt),
+			   repr_timer(&engine->execlists.timer));
+
+		read = execlists->csb_head;
+		write = READ_ONCE(*execlists->csb_write);
+
+		drm_printf(m, "\tExeclist status: 0x%08x %08x; CSB read:%d, write:%d, entries:%d\n",
+			   ENGINE_READ(engine, RING_EXECLIST_STATUS_LO),
+			   ENGINE_READ(engine, RING_EXECLIST_STATUS_HI),
+			   read, write, num_entries);
+
+		if (read >= num_entries)
+			read = 0;
+		if (write >= num_entries)
+			write = 0;
+		if (read > write)
+			write += num_entries;
+		while (read < write) {
+			idx = ++read % num_entries;
+			drm_printf(m, "\tExeclist CSB[%d]: 0x%08x, context: %d\n",
+				   idx, hws[idx * 2], hws[idx * 2 + 1]);
+		}
+
+		i915_sched_engine_active_lock_bh(engine->sched_engine);
+		rcu_read_lock();
+		for (port = execlists->active; (rq = *port); port++) {
+			char hdr[160];
+			int len;
+
+			len = scnprintf(hdr, sizeof(hdr),
+					"\t\tActive[%d]:  ccid:%08x%s%s, ",
+					(int)(port - execlists->active),
+					rq->context->lrc.ccid,
+					intel_context_is_closed(rq->context) ? "!" : "",
+					intel_context_is_banned(rq->context) ? "*" : "");
+			len += print_ring(hdr + len, sizeof(hdr) - len, rq);
+			scnprintf(hdr + len, sizeof(hdr) - len, "rq: ");
+			i915_request_show(m, rq, hdr, 0);
+		}
+		for (port = execlists->pending; (rq = *port); port++) {
+			char hdr[160];
+			int len;
+
+			len = scnprintf(hdr, sizeof(hdr),
+					"\t\tPending[%d]: ccid:%08x%s%s, ",
+					(int)(port - execlists->pending),
+					rq->context->lrc.ccid,
+					intel_context_is_closed(rq->context) ? "!" : "",
+					intel_context_is_banned(rq->context) ? "*" : "");
+			len += print_ring(hdr + len, sizeof(hdr) - len, rq);
+			scnprintf(hdr + len, sizeof(hdr) - len, "rq: ");
+			i915_request_show(m, rq, hdr, 0);
+		}
+		rcu_read_unlock();
+		i915_sched_engine_active_unlock_bh(engine->sched_engine);
+	} else if (GRAPHICS_VER(dev_priv) > 6) {
+		drm_printf(m, "\tPP_DIR_BASE: 0x%08x\n",
+			   ENGINE_READ(engine, RING_PP_DIR_BASE));
+		drm_printf(m, "\tPP_DIR_BASE_READ: 0x%08x\n",
+			   ENGINE_READ(engine, RING_PP_DIR_BASE_READ));
+		drm_printf(m, "\tPP_DIR_DCLV: 0x%08x\n",
+			   ENGINE_READ(engine, RING_PP_DIR_DCLV));
+	}
+}
+
+static void print_request_ring(struct drm_printer *m, struct i915_request *rq)
+{
+	struct i915_vma_resource *vma_res = rq->batch_res;
+	void *ring;
+	int size;
+
+	drm_printf(m,
+		   "[head %04x, postfix %04x, tail %04x, batch 0x%08x_%08x]:\n",
+		   rq->head, rq->postfix, rq->tail,
+		   vma_res ? upper_32_bits(vma_res->start) : ~0u,
+		   vma_res ? lower_32_bits(vma_res->start) : ~0u);
+
+	size = rq->tail - rq->head;
+	if (rq->tail < rq->head)
+		size += rq->ring->size;
+
+	ring = kmalloc(size, GFP_ATOMIC);
+	if (ring) {
+		const void *vaddr = rq->ring->vaddr;
+		unsigned int head = rq->head;
+		unsigned int len = 0;
+
+		if (rq->tail < head) {
+			len = rq->ring->size - head;
+			memcpy(ring, vaddr + head, len);
+			head = 0;
+		}
+		memcpy(ring + len, vaddr + head, size - len);
+
+		hexdump(m, ring, size);
+		kfree(ring);
+	}
+}
+
+static unsigned long read_ul(void *p, size_t x)
+{
+	return *(unsigned long *)(p + x);
+}
+
+static void print_properties(struct intel_engine_cs *engine,
+			     struct drm_printer *m)
+{
+	static const struct pmap {
+		size_t offset;
+		const char *name;
+	} props[] = {
+#define P(x) { \
+	.offset = offsetof(typeof(engine->props), x), \
+	.name = #x \
+}
+		P(heartbeat_interval_ms),
+		P(max_busywait_duration_ns),
+		P(preempt_timeout_ms),
+		P(stop_timeout_ms),
+		P(timeslice_duration_ms),
+
+		{},
+#undef P
+	};
+	const struct pmap *p;
+
+	drm_printf(m, "\tProperties:\n");
+	for (p = props; p->name; p++)
+		drm_printf(m, "\t\t%s: %lu [default %lu]\n",
+			   p->name,
+			   read_ul(&engine->props, p->offset),
+			   read_ul(&engine->defaults, p->offset));
+}
+
+static void engine_dump_request(struct i915_request *rq, struct drm_printer *m, const char *msg)
+{
+	struct intel_timeline *tl = get_timeline(rq);
+
+	i915_request_show(m, rq, msg, 0);
+
+	drm_printf(m, "\t\tring->start:  0x%08x\n",
+		   i915_ggtt_offset(rq->ring->vma));
+	drm_printf(m, "\t\tring->head:   0x%08x\n",
+		   rq->ring->head);
+	drm_printf(m, "\t\tring->tail:   0x%08x\n",
+		   rq->ring->tail);
+	drm_printf(m, "\t\tring->emit:   0x%08x\n",
+		   rq->ring->emit);
+	drm_printf(m, "\t\tring->space:  0x%08x\n",
+		   rq->ring->space);
+
+	if (tl) {
+		drm_printf(m, "\t\tring->hwsp:   0x%08x\n",
+			   tl->hwsp_offset);
+		intel_timeline_put(tl);
+	}
+
+	print_request_ring(m, rq);
+
+	if (rq->context->lrc_reg_state) {
+		drm_printf(m, "Logical Ring Context:\n");
+		hexdump(m, rq->context->lrc_reg_state, PAGE_SIZE);
+	}
+}
+
+void intel_engine_dump_active_requests(struct list_head *requests,
+				       struct i915_request *hung_rq,
+				       struct drm_printer *m)
+{
+	struct i915_request *rq;
+	const char *msg;
+	enum i915_request_state state;
+
+	list_for_each_entry(rq, requests, sched.link) {
+		if (rq == hung_rq)
+			continue;
+
+		state = i915_test_request_state(rq);
+		if (state < I915_REQUEST_QUEUED)
+			continue;
+
+		if (state == I915_REQUEST_ACTIVE)
+			msg = "\t\tactive on engine";
+		else
+			msg = "\t\tactive in queue";
+
+		engine_dump_request(rq, m, msg);
+	}
+}
+
+static void engine_dump_active_requests(struct intel_engine_cs *engine,
+					struct drm_printer *m)
+{
+	struct intel_context *hung_ce = NULL;
+	struct i915_request *hung_rq = NULL;
+
+	/*
+	 * No need for an engine->irq_seqno_barrier() before the seqno reads.
+	 * The GPU is still running so requests are still executing and any
+	 * hardware reads will be out of date by the time they are reported.
+	 * But the intention here is just to report an instantaneous snapshot
+	 * so that's fine.
+	 */
+	intel_engine_get_hung_entity(engine, &hung_ce, &hung_rq);
+
+	drm_printf(m, "\tRequests:\n");
+
+	if (hung_rq)
+		engine_dump_request(hung_rq, m, "\t\thung");
+	else if (hung_ce)
+		drm_printf(m, "\t\tGot hung ce but no hung rq!\n");
+
+	if (intel_uc_uses_guc_submission(&engine->gt->uc))
+		intel_guc_dump_active_requests(engine, hung_rq, m);
+	else
+		intel_execlists_dump_active_requests(engine, hung_rq, m);
+
+	if (hung_rq)
+		i915_request_put(hung_rq);
+}
+
+void intel_engine_dump(struct intel_engine_cs *engine,
+		       struct drm_printer *m,
+		       const char *header, ...)
+{
+	struct i915_gpu_error * const error = &engine->i915->gpu_error;
+	struct i915_request *rq;
+	intel_wakeref_t wakeref;
+	ktime_t dummy;
+
+	if (header) {
+		va_list ap;
+
+		va_start(ap, header);
+		drm_vprintf(m, header, &ap);
+		va_end(ap);
+	}
+
+	if (intel_gt_is_wedged(engine->gt))
+		drm_printf(m, "*** WEDGED ***\n");
+
+	drm_printf(m, "\tAwake? %d\n", atomic_read(&engine->wakeref.count));
+	drm_printf(m, "\tBarriers?: %s\n",
+		   str_yes_no(!llist_empty(&engine->barrier_tasks)));
+	drm_printf(m, "\tLatency: %luus\n",
+		   ewma__engine_latency_read(&engine->latency));
+	if (intel_engine_supports_stats(engine))
+		drm_printf(m, "\tRuntime: %llums\n",
+			   ktime_to_ms(intel_engine_get_busy_time(engine,
+								  &dummy)));
+	drm_printf(m, "\tForcewake: %x domains, %d active\n",
+		   engine->fw_domain, READ_ONCE(engine->fw_active));
+
+	rcu_read_lock();
+	rq = READ_ONCE(engine->heartbeat.systole);
+	if (rq)
+		drm_printf(m, "\tHeartbeat: %d ms ago\n",
+			   jiffies_to_msecs(jiffies - rq->emitted_jiffies));
+	rcu_read_unlock();
+	drm_printf(m, "\tReset count: %d (global %d)\n",
+		   i915_reset_engine_count(error, engine),
+		   i915_reset_count(error));
+	print_properties(engine, m);
+
+	engine_dump_active_requests(engine, m);
+
+	drm_printf(m, "\tMMIO base:  0x%08x\n", engine->mmio_base);
+	wakeref = intel_runtime_pm_get_if_in_use(engine->uncore->rpm);
+	if (wakeref) {
+		intel_engine_print_registers(engine, m);
+		intel_runtime_pm_put(engine->uncore->rpm, wakeref);
+	} else {
+		drm_printf(m, "\tDevice is asleep; skipping register dump\n");
+	}
+
+	intel_execlists_show_requests(engine, m, i915_request_show, 8);
+
+	drm_printf(m, "HWSP:\n");
+	hexdump(m, engine->status_page.addr, PAGE_SIZE);
+
+	drm_printf(m, "Idle? %s\n", str_yes_no(intel_engine_is_idle(engine)));
+
+	intel_engine_print_breadcrumbs(engine, m);
+}
+
+/**
+ * intel_engine_get_busy_time() - Return current accumulated engine busyness
+ * @engine: engine to report on
+ * @now: monotonic timestamp of sampling
+ *
+ * Returns accumulated time @engine was busy since engine stats were enabled.
+ */
+ktime_t intel_engine_get_busy_time(struct intel_engine_cs *engine, ktime_t *now)
+{
+	return engine->busyness(engine, now);
+}
+
+struct intel_context *
+intel_engine_create_virtual(struct intel_engine_cs **siblings,
+			    unsigned int count, unsigned long flags)
+{
+	if (count == 0)
+		return ERR_PTR(-EINVAL);
+
+	if (count == 1 && !(flags & FORCE_VIRTUAL))
+		return intel_context_create(siblings[0]);
+
+	GEM_BUG_ON(!siblings[0]->cops->create_virtual);
+	return siblings[0]->cops->create_virtual(siblings, count, flags);
+}
+
+static struct i915_request *engine_execlist_find_hung_request(struct intel_engine_cs *engine)
+{
+	struct i915_request *request, *active = NULL;
+
+	/*
+	 * This search does not work in GuC submission mode. However, the GuC
+	 * will report the hanging context directly to the driver itself. So
+	 * the driver should never get here when in GuC mode.
+	 */
+	GEM_BUG_ON(intel_uc_uses_guc_submission(&engine->gt->uc));
+
+	/*
+	 * We are called by the error capture, reset and to dump engine
+	 * state at random points in time. In particular, note that neither is
+	 * crucially ordered with an interrupt. After a hang, the GPU is dead
+	 * and we assume that no more writes can happen (we waited long enough
+	 * for all writes that were in transaction to be flushed) - adding an
+	 * extra delay for a recent interrupt is pointless. Hence, we do
+	 * not need an engine->irq_seqno_barrier() before the seqno reads.
+	 * At all other times, we must assume the GPU is still running, but
+	 * we only care about the snapshot of this moment.
+	 */
+	lockdep_assert_held(&engine->sched_engine->lock);
+
+	rcu_read_lock();
+	request = execlists_active(&engine->execlists);
+	if (request) {
+		struct intel_timeline *tl = request->context->timeline;
+
+		list_for_each_entry_from_reverse(request, &tl->requests, link) {
+			if (__i915_request_is_complete(request))
+				break;
+
+			active = request;
+		}
+	}
+	rcu_read_unlock();
+	if (active)
+		return active;
+
+	list_for_each_entry(request, &engine->sched_engine->requests,
+			    sched.link) {
+		if (i915_test_request_state(request) != I915_REQUEST_ACTIVE)
+			continue;
+
+		active = request;
+		break;
+	}
+
+	return active;
+}
+
+void intel_engine_get_hung_entity(struct intel_engine_cs *engine,
+				  struct intel_context **ce, struct i915_request **rq)
+{
+	unsigned long flags;
+
+	*ce = intel_engine_get_hung_context(engine);
+	if (*ce) {
+		intel_engine_clear_hung_context(engine);
+
+		*rq = intel_context_get_active_request(*ce);
+		return;
+	}
+
+	/*
+	 * Getting here with GuC enabled means it is a forced error capture
+	 * with no actual hang. So, no need to attempt the execlist search.
+	 */
+	if (intel_uc_uses_guc_submission(&engine->gt->uc))
+		return;
+
+	spin_lock_irqsave(&engine->sched_engine->lock, flags);
+	*rq = engine_execlist_find_hung_request(engine);
+	if (*rq)
+		*rq = i915_request_get_rcu(*rq);
+	spin_unlock_irqrestore(&engine->sched_engine->lock, flags);
+}
+
+void xehp_enable_ccs_engines(struct intel_engine_cs *engine)
+{
+	/*
+	 * If there are any non-fused-off CCS engines, we need to enable CCS
+	 * support in the RCU_MODE register.  This only needs to be done once,
+	 * so for simplicity we'll take care of this in the RCS engine's
+	 * resume handler; since the RCS and all CCS engines belong to the
+	 * same reset domain and are reset together, this will also take care
+	 * of re-applying the setting after i915-triggered resets.
+	 */
+	if (!CCS_MASK(engine->gt))
+		return;
+
+	intel_uncore_write(engine->uncore, GEN12_RCU_MODE,
+			   _MASKED_BIT_ENABLE(GEN12_RCU_MODE_CCS_ENABLE));
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "mock_engine.c"
+#include "selftest_engine.c"
+#include "selftest_engine_cs.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
new file mode 100644
index 000000000..a3698f611
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
@@ -0,0 +1,373 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_request.h"
+
+#include "intel_context.h"
+#include "intel_engine_heartbeat.h"
+#include "intel_engine_pm.h"
+#include "intel_engine.h"
+#include "intel_gt.h"
+#include "intel_reset.h"
+
+/*
+ * While the engine is active, we send a periodic pulse along the engine
+ * to check on its health and to flush any idle-barriers. If that request
+ * is stuck, and we fail to preempt it, we declare the engine hung and
+ * issue a reset -- in the hope that restores progress.
+ */
+
+static bool next_heartbeat(struct intel_engine_cs *engine)
+{
+	long delay;
+
+	delay = READ_ONCE(engine->props.heartbeat_interval_ms);
+	if (!delay)
+		return false;
+
+	delay = msecs_to_jiffies_timeout(delay);
+	if (delay >= HZ)
+		delay = round_jiffies_up_relative(delay);
+	mod_delayed_work(system_highpri_wq, &engine->heartbeat.work, delay + 1);
+
+	return true;
+}
+
+static struct i915_request *
+heartbeat_create(struct intel_context *ce, gfp_t gfp)
+{
+	struct i915_request *rq;
+
+	intel_context_enter(ce);
+	rq = __i915_request_create(ce, gfp);
+	intel_context_exit(ce);
+
+	return rq;
+}
+
+static void idle_pulse(struct intel_engine_cs *engine, struct i915_request *rq)
+{
+	engine->wakeref_serial = READ_ONCE(engine->serial) + 1;
+	i915_request_add_active_barriers(rq);
+	if (!engine->heartbeat.systole && intel_engine_has_heartbeat(engine))
+		engine->heartbeat.systole = i915_request_get(rq);
+}
+
+static void heartbeat_commit(struct i915_request *rq,
+			     const struct i915_sched_attr *attr)
+{
+	idle_pulse(rq->engine, rq);
+
+	__i915_request_commit(rq);
+	__i915_request_queue(rq, attr);
+}
+
+static void show_heartbeat(const struct i915_request *rq,
+			   struct intel_engine_cs *engine)
+{
+	struct drm_printer p = drm_debug_printer("heartbeat");
+
+	if (!rq) {
+		intel_engine_dump(engine, &p,
+				  "%s heartbeat not ticking\n",
+				  engine->name);
+	} else {
+		intel_engine_dump(engine, &p,
+				  "%s heartbeat {seqno:%llx:%lld, prio:%d} not ticking\n",
+				  engine->name,
+				  rq->fence.context,
+				  rq->fence.seqno,
+				  rq->sched.attr.priority);
+	}
+}
+
+static void
+reset_engine(struct intel_engine_cs *engine, struct i915_request *rq)
+{
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		show_heartbeat(rq, engine);
+
+	if (intel_engine_uses_guc(engine))
+		/*
+		 * GuC itself is toast or GuC's hang detection
+		 * is disabled. Either way, need to find the
+		 * hang culprit manually.
+		 */
+		intel_guc_find_hung_context(engine);
+
+	intel_gt_handle_error(engine->gt, engine->mask,
+			      I915_ERROR_CAPTURE,
+			      "stopped heartbeat on %s",
+			      engine->name);
+}
+
+static void heartbeat(struct work_struct *wrk)
+{
+	struct i915_sched_attr attr = { .priority = I915_PRIORITY_MIN };
+	struct intel_engine_cs *engine =
+		container_of(wrk, typeof(*engine), heartbeat.work.work);
+	struct intel_context *ce = engine->kernel_context;
+	struct i915_request *rq;
+	unsigned long serial;
+
+	/* Just in case everything has gone horribly wrong, give it a kick */
+	intel_engine_flush_submission(engine);
+
+	rq = engine->heartbeat.systole;
+	if (rq && i915_request_completed(rq)) {
+		i915_request_put(rq);
+		engine->heartbeat.systole = NULL;
+	}
+
+	if (!intel_engine_pm_get_if_awake(engine))
+		return;
+
+	if (intel_gt_is_wedged(engine->gt))
+		goto out;
+
+	if (i915_sched_engine_disabled(engine->sched_engine)) {
+		reset_engine(engine, engine->heartbeat.systole);
+		goto out;
+	}
+
+	if (engine->heartbeat.systole) {
+		long delay = READ_ONCE(engine->props.heartbeat_interval_ms);
+
+		/* Safeguard against too-fast worker invocations */
+		if (!time_after(jiffies,
+				rq->emitted_jiffies + msecs_to_jiffies(delay)))
+			goto out;
+
+		if (!i915_sw_fence_signaled(&rq->submit)) {
+			/*
+			 * Not yet submitted, system is stalled.
+			 *
+			 * This more often happens for ring submission,
+			 * where all contexts are funnelled into a common
+			 * ringbuffer. If one context is blocked on an
+			 * external fence, not only is it not submitted,
+			 * but all other contexts, including the kernel
+			 * context are stuck waiting for the signal.
+			 */
+		} else if (engine->sched_engine->schedule &&
+			   rq->sched.attr.priority < I915_PRIORITY_BARRIER) {
+			/*
+			 * Gradually raise the priority of the heartbeat to
+			 * give high priority work [which presumably desires
+			 * low latency and no jitter] the chance to naturally
+			 * complete before being preempted.
+			 */
+			attr.priority = 0;
+			if (rq->sched.attr.priority >= attr.priority)
+				attr.priority = I915_PRIORITY_HEARTBEAT;
+			if (rq->sched.attr.priority >= attr.priority)
+				attr.priority = I915_PRIORITY_BARRIER;
+
+			local_bh_disable();
+			engine->sched_engine->schedule(rq, &attr);
+			local_bh_enable();
+		} else {
+			reset_engine(engine, rq);
+		}
+
+		rq->emitted_jiffies = jiffies;
+		goto out;
+	}
+
+	serial = READ_ONCE(engine->serial);
+	if (engine->wakeref_serial == serial)
+		goto out;
+
+	if (!mutex_trylock(&ce->timeline->mutex)) {
+		/* Unable to lock the kernel timeline, is the engine stuck? */
+		if (xchg(&engine->heartbeat.blocked, serial) == serial)
+			intel_gt_handle_error(engine->gt, engine->mask,
+					      I915_ERROR_CAPTURE,
+					      "no heartbeat on %s",
+					      engine->name);
+		goto out;
+	}
+
+	rq = heartbeat_create(ce, GFP_NOWAIT | __GFP_NOWARN);
+	if (IS_ERR(rq))
+		goto unlock;
+
+	heartbeat_commit(rq, &attr);
+
+unlock:
+	mutex_unlock(&ce->timeline->mutex);
+out:
+	if (!engine->i915->params.enable_hangcheck || !next_heartbeat(engine))
+		i915_request_put(fetch_and_zero(&engine->heartbeat.systole));
+	intel_engine_pm_put(engine);
+}
+
+void intel_engine_unpark_heartbeat(struct intel_engine_cs *engine)
+{
+	if (!CONFIG_DRM_I915_HEARTBEAT_INTERVAL)
+		return;
+
+	next_heartbeat(engine);
+}
+
+void intel_engine_park_heartbeat(struct intel_engine_cs *engine)
+{
+	if (cancel_delayed_work(&engine->heartbeat.work))
+		i915_request_put(fetch_and_zero(&engine->heartbeat.systole));
+}
+
+void intel_gt_unpark_heartbeats(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id)
+		if (intel_engine_pm_is_awake(engine))
+			intel_engine_unpark_heartbeat(engine);
+}
+
+void intel_gt_park_heartbeats(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id)
+		intel_engine_park_heartbeat(engine);
+}
+
+void intel_engine_init_heartbeat(struct intel_engine_cs *engine)
+{
+	INIT_DELAYED_WORK(&engine->heartbeat.work, heartbeat);
+}
+
+static int __intel_engine_pulse(struct intel_engine_cs *engine)
+{
+	struct i915_sched_attr attr = { .priority = I915_PRIORITY_BARRIER };
+	struct intel_context *ce = engine->kernel_context;
+	struct i915_request *rq;
+
+	lockdep_assert_held(&ce->timeline->mutex);
+	GEM_BUG_ON(!intel_engine_has_preemption(engine));
+	GEM_BUG_ON(!intel_engine_pm_is_awake(engine));
+
+	rq = heartbeat_create(ce, GFP_NOWAIT | __GFP_NOWARN);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	__set_bit(I915_FENCE_FLAG_SENTINEL, &rq->fence.flags);
+
+	heartbeat_commit(rq, &attr);
+	GEM_BUG_ON(rq->sched.attr.priority < I915_PRIORITY_BARRIER);
+
+	return 0;
+}
+
+static unsigned long set_heartbeat(struct intel_engine_cs *engine,
+				   unsigned long delay)
+{
+	unsigned long old;
+
+	old = xchg(&engine->props.heartbeat_interval_ms, delay);
+	if (delay)
+		intel_engine_unpark_heartbeat(engine);
+	else
+		intel_engine_park_heartbeat(engine);
+
+	return old;
+}
+
+int intel_engine_set_heartbeat(struct intel_engine_cs *engine,
+			       unsigned long delay)
+{
+	struct intel_context *ce = engine->kernel_context;
+	int err = 0;
+
+	if (!delay && !intel_engine_has_preempt_reset(engine))
+		return -ENODEV;
+
+	intel_engine_pm_get(engine);
+
+	err = mutex_lock_interruptible(&ce->timeline->mutex);
+	if (err)
+		goto out_rpm;
+
+	if (delay != engine->props.heartbeat_interval_ms) {
+		unsigned long saved = set_heartbeat(engine, delay);
+
+		/* recheck current execution */
+		if (intel_engine_has_preemption(engine)) {
+			err = __intel_engine_pulse(engine);
+			if (err)
+				set_heartbeat(engine, saved);
+		}
+	}
+
+	mutex_unlock(&ce->timeline->mutex);
+
+out_rpm:
+	intel_engine_pm_put(engine);
+	return err;
+}
+
+int intel_engine_pulse(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce = engine->kernel_context;
+	int err;
+
+	if (!intel_engine_has_preemption(engine))
+		return -ENODEV;
+
+	if (!intel_engine_pm_get_if_awake(engine))
+		return 0;
+
+	err = -EINTR;
+	if (!mutex_lock_interruptible(&ce->timeline->mutex)) {
+		err = __intel_engine_pulse(engine);
+		mutex_unlock(&ce->timeline->mutex);
+	}
+
+	intel_engine_flush_submission(engine);
+	intel_engine_pm_put(engine);
+	return err;
+}
+
+int intel_engine_flush_barriers(struct intel_engine_cs *engine)
+{
+	struct i915_sched_attr attr = { .priority = I915_PRIORITY_MIN };
+	struct intel_context *ce = engine->kernel_context;
+	struct i915_request *rq;
+	int err;
+
+	if (llist_empty(&engine->barrier_tasks))
+		return 0;
+
+	if (!intel_engine_pm_get_if_awake(engine))
+		return 0;
+
+	if (mutex_lock_interruptible(&ce->timeline->mutex)) {
+		err = -EINTR;
+		goto out_rpm;
+	}
+
+	rq = heartbeat_create(ce, GFP_KERNEL);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_unlock;
+	}
+
+	heartbeat_commit(rq, &attr);
+
+	err = 0;
+out_unlock:
+	mutex_unlock(&ce->timeline->mutex);
+out_rpm:
+	intel_engine_pm_put(engine);
+	return err;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_engine_heartbeat.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.h b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.h
new file mode 100644
index 000000000..5da6d809a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_ENGINE_HEARTBEAT_H
+#define INTEL_ENGINE_HEARTBEAT_H
+
+struct intel_engine_cs;
+struct intel_gt;
+
+void intel_engine_init_heartbeat(struct intel_engine_cs *engine);
+
+int intel_engine_set_heartbeat(struct intel_engine_cs *engine,
+			       unsigned long delay);
+
+void intel_engine_park_heartbeat(struct intel_engine_cs *engine);
+void intel_engine_unpark_heartbeat(struct intel_engine_cs *engine);
+
+void intel_gt_park_heartbeats(struct intel_gt *gt);
+void intel_gt_unpark_heartbeats(struct intel_gt *gt);
+
+int intel_engine_pulse(struct intel_engine_cs *engine);
+int intel_engine_flush_barriers(struct intel_engine_cs *engine);
+
+#endif /* INTEL_ENGINE_HEARTBEAT_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_pm.c b/drivers/gpu/drm/i915/gt/intel_engine_pm.c
new file mode 100644
index 000000000..b0a4a2dbe
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_pm.c
@@ -0,0 +1,305 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "i915_drv.h"
+
+#include "intel_breadcrumbs.h"
+#include "intel_context.h"
+#include "intel_engine.h"
+#include "intel_engine_heartbeat.h"
+#include "intel_engine_pm.h"
+#include "intel_gt.h"
+#include "intel_gt_pm.h"
+#include "intel_rc6.h"
+#include "intel_ring.h"
+#include "shmem_utils.h"
+
+static void dbg_poison_ce(struct intel_context *ce)
+{
+	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		return;
+
+	if (ce->state) {
+		struct drm_i915_gem_object *obj = ce->state->obj;
+		int type = i915_coherent_map_type(ce->engine->i915, obj, true);
+		void *map;
+
+		if (!i915_gem_object_trylock(obj, NULL))
+			return;
+
+		map = i915_gem_object_pin_map(obj, type);
+		if (!IS_ERR(map)) {
+			memset(map, CONTEXT_REDZONE, obj->base.size);
+			i915_gem_object_flush_map(obj);
+			i915_gem_object_unpin_map(obj);
+		}
+		i915_gem_object_unlock(obj);
+	}
+}
+
+static int __engine_unpark(struct intel_wakeref *wf)
+{
+	struct intel_engine_cs *engine =
+		container_of(wf, typeof(*engine), wakeref);
+	struct intel_context *ce;
+
+	ENGINE_TRACE(engine, "\n");
+
+	intel_gt_pm_get(engine->gt);
+
+	/* Discard stale context state from across idling */
+	ce = engine->kernel_context;
+	if (ce) {
+		GEM_BUG_ON(test_bit(CONTEXT_VALID_BIT, &ce->flags));
+
+		/* Flush all pending HW writes before we touch the context */
+		while (unlikely(intel_context_inflight(ce)))
+			intel_engine_flush_submission(engine);
+
+		/* First poison the image to verify we never fully trust it */
+		dbg_poison_ce(ce);
+
+		/* Scrub the context image after our loss of control */
+		ce->ops->reset(ce);
+
+		CE_TRACE(ce, "reset { seqno:%x, *hwsp:%x, ring:%x }\n",
+			 ce->timeline->seqno,
+			 READ_ONCE(*ce->timeline->hwsp_seqno),
+			 ce->ring->emit);
+		GEM_BUG_ON(ce->timeline->seqno !=
+			   READ_ONCE(*ce->timeline->hwsp_seqno));
+	}
+
+	if (engine->unpark)
+		engine->unpark(engine);
+
+	intel_breadcrumbs_unpark(engine->breadcrumbs);
+	intel_engine_unpark_heartbeat(engine);
+	return 0;
+}
+
+static void duration(struct dma_fence *fence, struct dma_fence_cb *cb)
+{
+	struct i915_request *rq = to_request(fence);
+
+	ewma__engine_latency_add(&rq->engine->latency,
+				 ktime_us_delta(rq->fence.timestamp,
+						rq->duration.emitted));
+}
+
+static void
+__queue_and_release_pm(struct i915_request *rq,
+		       struct intel_timeline *tl,
+		       struct intel_engine_cs *engine)
+{
+	struct intel_gt_timelines *timelines = &engine->gt->timelines;
+
+	ENGINE_TRACE(engine, "parking\n");
+
+	/*
+	 * We have to serialise all potential retirement paths with our
+	 * submission, as we don't want to underflow either the
+	 * engine->wakeref.counter or our timeline->active_count.
+	 *
+	 * Equally, we cannot allow a new submission to start until
+	 * after we finish queueing, nor could we allow that submitter
+	 * to retire us before we are ready!
+	 */
+	spin_lock(&timelines->lock);
+
+	/* Let intel_gt_retire_requests() retire us (acquired under lock) */
+	if (!atomic_fetch_inc(&tl->active_count))
+		list_add_tail(&tl->link, &timelines->active_list);
+
+	/* Hand the request over to HW and so engine_retire() */
+	__i915_request_queue_bh(rq);
+
+	/* Let new submissions commence (and maybe retire this timeline) */
+	__intel_wakeref_defer_park(&engine->wakeref);
+
+	spin_unlock(&timelines->lock);
+}
+
+static bool switch_to_kernel_context(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce = engine->kernel_context;
+	struct i915_request *rq;
+	bool result = true;
+
+	/*
+	 * This is execlist specific behaviour intended to ensure the GPU is
+	 * idle by switching to a known 'safe' context. With GuC submission, the
+	 * same idle guarantee is achieved by other means (disabling
+	 * scheduling). Further, switching to a 'safe' context has no effect
+	 * with GuC submission as the scheduler can just switch back again.
+	 *
+	 * FIXME: Move this backend scheduler specific behaviour into the
+	 * scheduler backend.
+	 */
+	if (intel_engine_uses_guc(engine))
+		return true;
+
+	/* GPU is pointing to the void, as good as in the kernel context. */
+	if (intel_gt_is_wedged(engine->gt))
+		return true;
+
+	GEM_BUG_ON(!intel_context_is_barrier(ce));
+	GEM_BUG_ON(ce->timeline->hwsp_ggtt != engine->status_page.vma);
+
+	/* Already inside the kernel context, safe to power down. */
+	if (engine->wakeref_serial == engine->serial)
+		return true;
+
+	/*
+	 * Note, we do this without taking the timeline->mutex. We cannot
+	 * as we may be called while retiring the kernel context and so
+	 * already underneath the timeline->mutex. Instead we rely on the
+	 * exclusive property of the __engine_park that prevents anyone
+	 * else from creating a request on this engine. This also requires
+	 * that the ring is empty and we avoid any waits while constructing
+	 * the context, as they assume protection by the timeline->mutex.
+	 * This should hold true as we can only park the engine after
+	 * retiring the last request, thus all rings should be empty and
+	 * all timelines idle.
+	 *
+	 * For unlocking, there are 2 other parties and the GPU who have a
+	 * stake here.
+	 *
+	 * A new gpu user will be waiting on the engine-pm to start their
+	 * engine_unpark. New waiters are predicated on engine->wakeref.count
+	 * and so intel_wakeref_defer_park() acts like a mutex_unlock of the
+	 * engine->wakeref.
+	 *
+	 * The other party is intel_gt_retire_requests(), which is walking the
+	 * list of active timelines looking for completions. Meanwhile as soon
+	 * as we call __i915_request_queue(), the GPU may complete our request.
+	 * Ergo, if we put ourselves on the timelines.active_list
+	 * (se intel_timeline_enter()) before we increment the
+	 * engine->wakeref.count, we may see the request completion and retire
+	 * it causing an underflow of the engine->wakeref.
+	 */
+	set_bit(CONTEXT_IS_PARKING, &ce->flags);
+	GEM_BUG_ON(atomic_read(&ce->timeline->active_count) < 0);
+
+	rq = __i915_request_create(ce, GFP_NOWAIT);
+	if (IS_ERR(rq))
+		/* Context switch failed, hope for the best! Maybe reset? */
+		goto out_unlock;
+
+	/* Check again on the next retirement. */
+	engine->wakeref_serial = engine->serial + 1;
+	i915_request_add_active_barriers(rq);
+
+	/* Install ourselves as a preemption barrier */
+	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+	if (likely(!__i915_request_commit(rq))) { /* engine should be idle! */
+		/*
+		 * Use an interrupt for precise measurement of duration,
+		 * otherwise we rely on someone else retiring all the requests
+		 * which may delay the signaling (i.e. we will likely wait
+		 * until the background request retirement running every
+		 * second or two).
+		 */
+		BUILD_BUG_ON(sizeof(rq->duration) > sizeof(rq->submitq));
+		dma_fence_add_callback(&rq->fence, &rq->duration.cb, duration);
+		rq->duration.emitted = ktime_get();
+	}
+
+	/* Expose ourselves to the world */
+	__queue_and_release_pm(rq, ce->timeline, engine);
+
+	result = false;
+out_unlock:
+	clear_bit(CONTEXT_IS_PARKING, &ce->flags);
+	return result;
+}
+
+static void call_idle_barriers(struct intel_engine_cs *engine)
+{
+	struct llist_node *node, *next;
+
+	llist_for_each_safe(node, next, llist_del_all(&engine->barrier_tasks)) {
+		struct dma_fence_cb *cb =
+			container_of((struct list_head *)node,
+				     typeof(*cb), node);
+
+		cb->func(ERR_PTR(-EAGAIN), cb);
+	}
+}
+
+static int __engine_park(struct intel_wakeref *wf)
+{
+	struct intel_engine_cs *engine =
+		container_of(wf, typeof(*engine), wakeref);
+
+	engine->saturated = 0;
+
+	/*
+	 * If one and only one request is completed between pm events,
+	 * we know that we are inside the kernel context and it is
+	 * safe to power down. (We are paranoid in case that runtime
+	 * suspend causes corruption to the active context image, and
+	 * want to avoid that impacting userspace.)
+	 */
+	if (!switch_to_kernel_context(engine))
+		return -EBUSY;
+
+	ENGINE_TRACE(engine, "parked\n");
+
+	call_idle_barriers(engine); /* cleanup after wedging */
+
+	intel_engine_park_heartbeat(engine);
+	intel_breadcrumbs_park(engine->breadcrumbs);
+
+	/* Must be reset upon idling, or we may miss the busy wakeup. */
+	GEM_BUG_ON(engine->sched_engine->queue_priority_hint != INT_MIN);
+
+	if (engine->park)
+		engine->park(engine);
+
+	/* While gt calls i915_vma_parked(), we have to break the lock cycle */
+	intel_gt_pm_put_async(engine->gt);
+	return 0;
+}
+
+static const struct intel_wakeref_ops wf_ops = {
+	.get = __engine_unpark,
+	.put = __engine_park,
+};
+
+void intel_engine_init__pm(struct intel_engine_cs *engine)
+{
+	struct intel_runtime_pm *rpm = engine->uncore->rpm;
+
+	intel_wakeref_init(&engine->wakeref, rpm, &wf_ops);
+	intel_engine_init_heartbeat(engine);
+}
+
+/**
+ * intel_engine_reset_pinned_contexts - Reset the pinned contexts of
+ * an engine.
+ * @engine: The engine whose pinned contexts we want to reset.
+ *
+ * Typically the pinned context LMEM images lose or get their content
+ * corrupted on suspend. This function resets their images.
+ */
+void intel_engine_reset_pinned_contexts(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+
+	list_for_each_entry(ce, &engine->pinned_contexts_list,
+			    pinned_contexts_link) {
+		/* kernel context gets reset at __engine_unpark() */
+		if (ce == engine->kernel_context)
+			continue;
+
+		dbg_poison_ce(ce);
+		ce->ops->reset(ce);
+	}
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_engine_pm.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_pm.h b/drivers/gpu/drm/i915/gt/intel_engine_pm.h
new file mode 100644
index 000000000..d68675925
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_pm.h
@@ -0,0 +1,111 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_ENGINE_PM_H
+#define INTEL_ENGINE_PM_H
+
+#include "i915_drv.h"
+#include "i915_request.h"
+#include "intel_engine_types.h"
+#include "intel_wakeref.h"
+#include "intel_gt_pm.h"
+
+static inline bool
+intel_engine_pm_is_awake(const struct intel_engine_cs *engine)
+{
+	return intel_wakeref_is_active(&engine->wakeref);
+}
+
+static inline void __intel_engine_pm_get(struct intel_engine_cs *engine)
+{
+	__intel_wakeref_get(&engine->wakeref);
+}
+
+static inline void intel_engine_pm_get(struct intel_engine_cs *engine)
+{
+	intel_wakeref_get(&engine->wakeref);
+}
+
+static inline bool intel_engine_pm_get_if_awake(struct intel_engine_cs *engine)
+{
+	return intel_wakeref_get_if_active(&engine->wakeref);
+}
+
+static inline void intel_engine_pm_might_get(struct intel_engine_cs *engine)
+{
+	if (!intel_engine_is_virtual(engine)) {
+		intel_wakeref_might_get(&engine->wakeref);
+	} else {
+		struct intel_gt *gt = engine->gt;
+		struct intel_engine_cs *tengine;
+		intel_engine_mask_t tmp, mask = engine->mask;
+
+		for_each_engine_masked(tengine, gt, mask, tmp)
+			intel_wakeref_might_get(&tengine->wakeref);
+	}
+	intel_gt_pm_might_get(engine->gt);
+}
+
+static inline void intel_engine_pm_put(struct intel_engine_cs *engine)
+{
+	intel_wakeref_put(&engine->wakeref);
+}
+
+static inline void intel_engine_pm_put_async(struct intel_engine_cs *engine)
+{
+	intel_wakeref_put_async(&engine->wakeref);
+}
+
+static inline void intel_engine_pm_put_delay(struct intel_engine_cs *engine,
+					     unsigned long delay)
+{
+	intel_wakeref_put_delay(&engine->wakeref, delay);
+}
+
+static inline void intel_engine_pm_flush(struct intel_engine_cs *engine)
+{
+	intel_wakeref_unlock_wait(&engine->wakeref);
+}
+
+static inline void intel_engine_pm_might_put(struct intel_engine_cs *engine)
+{
+	if (!intel_engine_is_virtual(engine)) {
+		intel_wakeref_might_put(&engine->wakeref);
+	} else {
+		struct intel_gt *gt = engine->gt;
+		struct intel_engine_cs *tengine;
+		intel_engine_mask_t tmp, mask = engine->mask;
+
+		for_each_engine_masked(tengine, gt, mask, tmp)
+			intel_wakeref_might_put(&tengine->wakeref);
+	}
+	intel_gt_pm_might_put(engine->gt);
+}
+
+static inline struct i915_request *
+intel_engine_create_kernel_request(struct intel_engine_cs *engine)
+{
+	struct i915_request *rq;
+
+	/*
+	 * The engine->kernel_context is special as it is used inside
+	 * the engine-pm barrier (see __engine_park()), circumventing
+	 * the usual mutexes and relying on the engine-pm barrier
+	 * instead. So whenever we use the engine->kernel_context
+	 * outside of the barrier, we must manually handle the
+	 * engine wakeref to serialise with the use inside.
+	 */
+	intel_engine_pm_get(engine);
+	rq = i915_request_create(engine->kernel_context);
+	intel_engine_pm_put(engine);
+
+	return rq;
+}
+
+void intel_engine_init__pm(struct intel_engine_cs *engine);
+
+void intel_engine_reset_pinned_contexts(struct intel_engine_cs *engine);
+
+#endif /* INTEL_ENGINE_PM_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_regs.h b/drivers/gpu/drm/i915/gt/intel_engine_regs.h
new file mode 100644
index 000000000..fe1a0d5fd
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_regs.h
@@ -0,0 +1,258 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_ENGINE_REGS__
+#define __INTEL_ENGINE_REGS__
+
+#include "i915_reg_defs.h"
+
+#define RING_EXCC(base)				_MMIO((base) + 0x28)
+#define RING_TAIL(base)				_MMIO((base) + 0x30)
+#define   TAIL_ADDR				0x001FFFF8
+#define RING_HEAD(base)				_MMIO((base) + 0x34)
+#define   HEAD_WRAP_COUNT			0xFFE00000
+#define   HEAD_WRAP_ONE				0x00200000
+#define   HEAD_ADDR				0x001FFFFC
+#define RING_START(base)			_MMIO((base) + 0x38)
+#define RING_CTL(base)				_MMIO((base) + 0x3c)
+#define   RING_CTL_SIZE(size)			((size) - PAGE_SIZE) /* in bytes -> pages */
+#define   RING_NR_PAGES				0x001FF000
+#define   RING_REPORT_MASK			0x00000006
+#define   RING_REPORT_64K			0x00000002
+#define   RING_REPORT_128K			0x00000004
+#define   RING_NO_REPORT			0x00000000
+#define   RING_VALID_MASK			0x00000001
+#define   RING_VALID				0x00000001
+#define   RING_INVALID				0x00000000
+#define   RING_WAIT_I8XX			(1 << 0) /* gen2, PRBx_HEAD */
+#define   RING_WAIT				(1 << 11) /* gen3+, PRBx_CTL */
+#define   RING_WAIT_SEMAPHORE			(1 << 10) /* gen6+ */
+#define RING_SYNC_0(base)			_MMIO((base) + 0x40)
+#define RING_SYNC_1(base)			_MMIO((base) + 0x44)
+#define RING_SYNC_2(base)			_MMIO((base) + 0x48)
+#define GEN6_RVSYNC				(RING_SYNC_0(RENDER_RING_BASE))
+#define GEN6_RBSYNC				(RING_SYNC_1(RENDER_RING_BASE))
+#define GEN6_RVESYNC				(RING_SYNC_2(RENDER_RING_BASE))
+#define GEN6_VBSYNC				(RING_SYNC_0(GEN6_BSD_RING_BASE))
+#define GEN6_VRSYNC				(RING_SYNC_1(GEN6_BSD_RING_BASE))
+#define GEN6_VVESYNC				(RING_SYNC_2(GEN6_BSD_RING_BASE))
+#define GEN6_BRSYNC				(RING_SYNC_0(BLT_RING_BASE))
+#define GEN6_BVSYNC				(RING_SYNC_1(BLT_RING_BASE))
+#define GEN6_BVESYNC				(RING_SYNC_2(BLT_RING_BASE))
+#define GEN6_VEBSYNC				(RING_SYNC_0(VEBOX_RING_BASE))
+#define GEN6_VERSYNC				(RING_SYNC_1(VEBOX_RING_BASE))
+#define GEN6_VEVSYNC				(RING_SYNC_2(VEBOX_RING_BASE))
+#define RING_PSMI_CTL(base)			_MMIO((base) + 0x50)
+#define   GEN8_RC_SEMA_IDLE_MSG_DISABLE		REG_BIT(12)
+#define   GEN8_FF_DOP_CLOCK_GATE_DISABLE	REG_BIT(10)
+#define   GEN12_WAIT_FOR_EVENT_POWER_DOWN_DISABLE REG_BIT(7)
+#define   GEN6_BSD_GO_INDICATOR			REG_BIT(4)
+#define   GEN6_BSD_SLEEP_INDICATOR		REG_BIT(3)
+#define   GEN6_BSD_SLEEP_FLUSH_DISABLE		REG_BIT(2)
+#define   GEN6_PSMI_SLEEP_MSG_DISABLE		REG_BIT(0)
+#define RING_MAX_IDLE(base)			_MMIO((base) + 0x54)
+#define  PWRCTX_MAXCNT(base)			_MMIO((base) + 0x54)
+#define    IDLE_TIME_MASK			0xFFFFF
+#define RING_ACTHD_UDW(base)			_MMIO((base) + 0x5c)
+#define RING_DMA_FADD_UDW(base)			_MMIO((base) + 0x60) /* gen8+ */
+#define RING_IPEIR(base)			_MMIO((base) + 0x64)
+#define RING_IPEHR(base)			_MMIO((base) + 0x68)
+#define RING_INSTDONE(base)			_MMIO((base) + 0x6c)
+#define RING_INSTPS(base)			_MMIO((base) + 0x70)
+#define RING_DMA_FADD(base)			_MMIO((base) + 0x78)
+#define RING_ACTHD(base)			_MMIO((base) + 0x74)
+#define RING_HWS_PGA(base)			_MMIO((base) + 0x80)
+#define RING_CMD_BUF_CCTL(base)			_MMIO((base) + 0x84)
+#define IPEIR(base)				_MMIO((base) + 0x88)
+#define IPEHR(base)				_MMIO((base) + 0x8c)
+#define RING_ID(base)				_MMIO((base) + 0x8c)
+#define RING_NOPID(base)			_MMIO((base) + 0x94)
+#define RING_HWSTAM(base)			_MMIO((base) + 0x98)
+#define RING_MI_MODE(base)			_MMIO((base) + 0x9c)
+#define   ASYNC_FLIP_PERF_DISABLE		REG_BIT(14)
+#define   MI_FLUSH_ENABLE			REG_BIT(12)
+#define   TGL_NESTED_BB_EN			REG_BIT(12)
+#define   MODE_IDLE				REG_BIT(9)
+#define   STOP_RING				REG_BIT(8)
+#define   VS_TIMER_DISPATCH			REG_BIT(6)
+#define RING_IMR(base)				_MMIO((base) + 0xa8)
+#define RING_EIR(base)				_MMIO((base) + 0xb0)
+#define RING_EMR(base)				_MMIO((base) + 0xb4)
+#define RING_ESR(base)				_MMIO((base) + 0xb8)
+#define RING_INSTPM(base)			_MMIO((base) + 0xc0)
+#define RING_CMD_CCTL(base)			_MMIO((base) + 0xc4)
+#define ACTHD(base)				_MMIO((base) + 0xc8)
+#define GEN8_R_PWR_CLK_STATE(base)		_MMIO((base) + 0xc8)
+#define   GEN8_RPCS_ENABLE			(1 << 31)
+#define   GEN8_RPCS_S_CNT_ENABLE		(1 << 18)
+#define   GEN8_RPCS_S_CNT_SHIFT			15
+#define   GEN8_RPCS_S_CNT_MASK			(0x7 << GEN8_RPCS_S_CNT_SHIFT)
+#define   GEN11_RPCS_S_CNT_SHIFT		12
+#define   GEN11_RPCS_S_CNT_MASK			(0x3f << GEN11_RPCS_S_CNT_SHIFT)
+#define   GEN8_RPCS_SS_CNT_ENABLE		(1 << 11)
+#define   GEN8_RPCS_SS_CNT_SHIFT		8
+#define   GEN8_RPCS_SS_CNT_MASK			(0x7 << GEN8_RPCS_SS_CNT_SHIFT)
+#define   GEN8_RPCS_EU_MAX_SHIFT		4
+#define   GEN8_RPCS_EU_MAX_MASK			(0xf << GEN8_RPCS_EU_MAX_SHIFT)
+#define   GEN8_RPCS_EU_MIN_SHIFT		0
+#define   GEN8_RPCS_EU_MIN_MASK			(0xf << GEN8_RPCS_EU_MIN_SHIFT)
+
+#define RING_RESET_CTL(base)			_MMIO((base) + 0xd0)
+#define   RESET_CTL_CAT_ERROR			REG_BIT(2)
+#define   RESET_CTL_READY_TO_RESET		REG_BIT(1)
+#define   RESET_CTL_REQUEST_RESET		REG_BIT(0)
+#define DMA_FADD_I8XX(base)			_MMIO((base) + 0xd0)
+#define RING_BBSTATE(base)			_MMIO((base) + 0x110)
+#define   RING_BB_PPGTT				(1 << 5)
+#define RING_SBBADDR(base)			_MMIO((base) + 0x114) /* hsw+ */
+#define RING_SBBSTATE(base)			_MMIO((base) + 0x118) /* hsw+ */
+#define RING_SBBADDR_UDW(base)			_MMIO((base) + 0x11c) /* gen8+ */
+#define RING_BBADDR(base)			_MMIO((base) + 0x140)
+#define RING_BB_OFFSET(base)			_MMIO((base) + 0x158)
+#define RING_BBADDR_UDW(base)			_MMIO((base) + 0x168) /* gen8+ */
+#define CCID(base)				_MMIO((base) + 0x180)
+#define   CCID_EN				BIT(0)
+#define   CCID_EXTENDED_STATE_RESTORE		BIT(2)
+#define   CCID_EXTENDED_STATE_SAVE		BIT(3)
+#define RING_BB_PER_CTX_PTR(base)		_MMIO((base) + 0x1c0) /* gen8+ */
+#define RING_INDIRECT_CTX(base)			_MMIO((base) + 0x1c4) /* gen8+ */
+#define RING_INDIRECT_CTX_OFFSET(base)		_MMIO((base) + 0x1c8) /* gen8+ */
+#define ECOSKPD(base)				_MMIO((base) + 0x1d0)
+#define   ECO_CONSTANT_BUFFER_SR_DISABLE	REG_BIT(4)
+#define   ECO_GATING_CX_ONLY			REG_BIT(3)
+#define   GEN6_BLITTER_FBC_NOTIFY		REG_BIT(3)
+#define   ECO_FLIP_DONE				REG_BIT(0)
+#define   GEN6_BLITTER_LOCK_SHIFT		16
+
+#define BLIT_CCTL(base)				_MMIO((base) + 0x204)
+#define   BLIT_CCTL_DST_MOCS_MASK		REG_GENMASK(14, 8)
+#define   BLIT_CCTL_SRC_MOCS_MASK		REG_GENMASK(6, 0)
+#define   BLIT_CCTL_MASK (BLIT_CCTL_DST_MOCS_MASK | \
+			  BLIT_CCTL_SRC_MOCS_MASK)
+#define   BLIT_CCTL_MOCS(dst, src)				       \
+		(REG_FIELD_PREP(BLIT_CCTL_DST_MOCS_MASK, (dst) << 1) | \
+		 REG_FIELD_PREP(BLIT_CCTL_SRC_MOCS_MASK, (src) << 1))
+
+#define RING_CSCMDOP(base)			_MMIO((base) + 0x20c)
+
+/*
+ * CMD_CCTL read/write fields take a MOCS value and _not_ a table index.
+ * The lsb of each can be considered a separate enabling bit for encryption.
+ * 6:0 == default MOCS value for reads  =>  6:1 == table index for reads.
+ * 13:7 == default MOCS value for writes => 13:8 == table index for writes.
+ * 15:14 == Reserved => 31:30 are set to 0.
+ */
+#define CMD_CCTL_WRITE_OVERRIDE_MASK REG_GENMASK(13, 7)
+#define CMD_CCTL_READ_OVERRIDE_MASK REG_GENMASK(6, 0)
+#define CMD_CCTL_MOCS_MASK (CMD_CCTL_WRITE_OVERRIDE_MASK | \
+			    CMD_CCTL_READ_OVERRIDE_MASK)
+#define CMD_CCTL_MOCS_OVERRIDE(write, read)				      \
+		(REG_FIELD_PREP(CMD_CCTL_WRITE_OVERRIDE_MASK, (write) << 1) | \
+		 REG_FIELD_PREP(CMD_CCTL_READ_OVERRIDE_MASK, (read) << 1))
+
+#define RING_PREDICATE_RESULT(base)		_MMIO((base) + 0x3b8) /* gen12+ */
+
+#define MI_PREDICATE_RESULT_2(base)		_MMIO((base) + 0x3bc)
+#define   LOWER_SLICE_ENABLED			(1 << 0)
+#define   LOWER_SLICE_DISABLED			(0 << 0)
+#define MI_PREDICATE_SRC0(base)			_MMIO((base) + 0x400)
+#define MI_PREDICATE_SRC0_UDW(base)		_MMIO((base) + 0x400 + 4)
+#define MI_PREDICATE_SRC1(base)			_MMIO((base) + 0x408)
+#define MI_PREDICATE_SRC1_UDW(base)		_MMIO((base) + 0x408 + 4)
+#define MI_PREDICATE_DATA(base)			_MMIO((base) + 0x410)
+#define MI_PREDICATE_RESULT(base)		_MMIO((base) + 0x418)
+#define MI_PREDICATE_RESULT_1(base)		_MMIO((base) + 0x41c)
+
+#define RING_PP_DIR_DCLV(base)			_MMIO((base) + 0x220)
+#define   PP_DIR_DCLV_2G			0xffffffff
+#define RING_PP_DIR_BASE(base)			_MMIO((base) + 0x228)
+#define RING_ELSP(base)				_MMIO((base) + 0x230)
+#define RING_EXECLIST_STATUS_LO(base)		_MMIO((base) + 0x234)
+#define RING_EXECLIST_STATUS_HI(base)		_MMIO((base) + 0x234 + 4)
+#define RING_CONTEXT_CONTROL(base)		_MMIO((base) + 0x244)
+#define	  CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT	REG_BIT(0)
+#define   CTX_CTRL_RS_CTX_ENABLE		REG_BIT(1)
+#define	  CTX_CTRL_ENGINE_CTX_SAVE_INHIBIT	REG_BIT(2)
+#define	  CTX_CTRL_INHIBIT_SYN_CTX_SWITCH	REG_BIT(3)
+#define	  GEN12_CTX_CTRL_OAR_CONTEXT_ENABLE	REG_BIT(8)
+#define RING_CTX_SR_CTL(base)			_MMIO((base) + 0x244)
+#define RING_SEMA_WAIT_POLL(base)		_MMIO((base) + 0x24c)
+#define GEN8_RING_PDP_UDW(base, n)		_MMIO((base) + 0x270 + (n) * 8 + 4)
+#define GEN8_RING_PDP_LDW(base, n)		_MMIO((base) + 0x270 + (n) * 8)
+#define RING_MODE_GEN7(base)			_MMIO((base) + 0x29c)
+#define   GFX_RUN_LIST_ENABLE			(1 << 15)
+#define   GFX_INTERRUPT_STEERING		(1 << 14)
+#define   GFX_TLB_INVALIDATE_EXPLICIT		(1 << 13)
+#define   GFX_SURFACE_FAULT_ENABLE		(1 << 12)
+#define   GFX_REPLAY_MODE			(1 << 11)
+#define   GFX_PSMI_GRANULARITY			(1 << 10)
+#define   GEN12_GFX_PREFETCH_DISABLE		REG_BIT(10)
+#define   GFX_PPGTT_ENABLE			(1 << 9)
+#define   GEN8_GFX_PPGTT_48B			(1 << 7)
+#define   GFX_FORWARD_VBLANK_MASK		(3 << 5)
+#define   GFX_FORWARD_VBLANK_NEVER		(0 << 5)
+#define   GFX_FORWARD_VBLANK_ALWAYS		(1 << 5)
+#define   GFX_FORWARD_VBLANK_COND		(2 << 5)
+#define   GEN11_GFX_DISABLE_LEGACY_MODE		(1 << 3)
+#define RING_TIMESTAMP(base)			_MMIO((base) + 0x358)
+#define RING_TIMESTAMP_UDW(base)		_MMIO((base) + 0x358 + 4)
+#define RING_CONTEXT_STATUS_PTR(base)		_MMIO((base) + 0x3a0)
+#define RING_CTX_TIMESTAMP(base)		_MMIO((base) + 0x3a8) /* gen8+ */
+#define RING_PREDICATE_RESULT(base)		_MMIO((base) + 0x3b8)
+#define RING_FORCE_TO_NONPRIV(base, i)		_MMIO(((base) + 0x4D0) + (i) * 4)
+#define   RING_FORCE_TO_NONPRIV_DENY		REG_BIT(30)
+#define   RING_FORCE_TO_NONPRIV_ADDRESS_MASK	REG_GENMASK(25, 2)
+#define   RING_FORCE_TO_NONPRIV_ACCESS_RW	(0 << 28)    /* CFL+ & Gen11+ */
+#define   RING_FORCE_TO_NONPRIV_ACCESS_RD	(1 << 28)
+#define   RING_FORCE_TO_NONPRIV_ACCESS_WR	(2 << 28)
+#define   RING_FORCE_TO_NONPRIV_ACCESS_INVALID	(3 << 28)
+#define   RING_FORCE_TO_NONPRIV_ACCESS_MASK	(3 << 28)
+#define   RING_FORCE_TO_NONPRIV_RANGE_1		(0 << 0)     /* CFL+ & Gen11+ */
+#define   RING_FORCE_TO_NONPRIV_RANGE_4		(1 << 0)
+#define   RING_FORCE_TO_NONPRIV_RANGE_16	(2 << 0)
+#define   RING_FORCE_TO_NONPRIV_RANGE_64	(3 << 0)
+#define   RING_FORCE_TO_NONPRIV_RANGE_MASK	(3 << 0)
+#define   RING_FORCE_TO_NONPRIV_MASK_VALID	\
+	(RING_FORCE_TO_NONPRIV_RANGE_MASK | \
+	 RING_FORCE_TO_NONPRIV_ACCESS_MASK | \
+	 RING_FORCE_TO_NONPRIV_DENY)
+#define   RING_MAX_NONPRIV_SLOTS  12
+
+#define RING_EXECLIST_SQ_CONTENTS(base)		_MMIO((base) + 0x510)
+#define RING_PP_DIR_BASE_READ(base)		_MMIO((base) + 0x518)
+#define RING_EXECLIST_CONTROL(base)		_MMIO((base) + 0x550)
+#define	  EL_CTRL_LOAD				REG_BIT(0)
+
+/* There are 16 64-bit CS General Purpose Registers per-engine on Gen8+ */
+#define GEN8_RING_CS_GPR(base, n)		_MMIO((base) + 0x600 + (n) * 8)
+#define GEN8_RING_CS_GPR_UDW(base, n)		_MMIO((base) + 0x600 + (n) * 8 + 4)
+
+#define GEN11_VCS_SFC_FORCED_LOCK(base)		_MMIO((base) + 0x88c)
+#define   GEN11_VCS_SFC_FORCED_LOCK_BIT		(1 << 0)
+#define GEN11_VCS_SFC_LOCK_STATUS(base)		_MMIO((base) + 0x890)
+#define   GEN11_VCS_SFC_USAGE_BIT		(1 << 0)
+#define   GEN11_VCS_SFC_LOCK_ACK_BIT		(1 << 1)
+
+#define GEN11_VECS_SFC_FORCED_LOCK(base)	_MMIO((base) + 0x201c)
+#define   GEN11_VECS_SFC_FORCED_LOCK_BIT	(1 << 0)
+#define GEN11_VECS_SFC_LOCK_ACK(base)		_MMIO((base) + 0x2018)
+#define   GEN11_VECS_SFC_LOCK_ACK_BIT		(1 << 0)
+#define GEN11_VECS_SFC_USAGE(base)		_MMIO((base) + 0x2014)
+#define   GEN11_VECS_SFC_USAGE_BIT		(1 << 0)
+
+#define RING_HWS_PGA_GEN6(base)	_MMIO((base) + 0x2080)
+
+#define GEN12_HCP_SFC_LOCK_STATUS(base)		_MMIO((base) + 0x2914)
+#define   GEN12_HCP_SFC_LOCK_ACK_BIT		REG_BIT(1)
+#define   GEN12_HCP_SFC_USAGE_BIT		REG_BIT(0)
+
+#define VDBOX_CGCTL3F10(base)			_MMIO((base) + 0x3f10)
+#define   IECPUNIT_CLKGATE_DIS			REG_BIT(22)
+
+#define VDBOX_CGCTL3F18(base)			_MMIO((base) + 0x3f18)
+#define   ALNUNIT_CLKGATE_DIS			REG_BIT(13)
+
+
+#endif /* __INTEL_ENGINE_REGS__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_stats.h b/drivers/gpu/drm/i915/gt/intel_engine_stats.h
new file mode 100644
index 000000000..8e762d683
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_stats.h
@@ -0,0 +1,61 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __INTEL_ENGINE_STATS_H__
+#define __INTEL_ENGINE_STATS_H__
+
+#include <linux/atomic.h>
+#include <linux/ktime.h>
+#include <linux/seqlock.h>
+
+#include "i915_gem.h" /* GEM_BUG_ON */
+#include "intel_engine.h"
+
+static inline void intel_engine_context_in(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists_stats *stats = &engine->stats.execlists;
+	unsigned long flags;
+
+	if (stats->active) {
+		stats->active++;
+		return;
+	}
+
+	/* The writer is serialised; but the pmu reader may be from hardirq */
+	local_irq_save(flags);
+	write_seqcount_begin(&stats->lock);
+
+	stats->start = ktime_get();
+	stats->active++;
+
+	write_seqcount_end(&stats->lock);
+	local_irq_restore(flags);
+
+	GEM_BUG_ON(!stats->active);
+}
+
+static inline void intel_engine_context_out(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists_stats *stats = &engine->stats.execlists;
+	unsigned long flags;
+
+	GEM_BUG_ON(!stats->active);
+	if (stats->active > 1) {
+		stats->active--;
+		return;
+	}
+
+	local_irq_save(flags);
+	write_seqcount_begin(&stats->lock);
+
+	stats->active--;
+	stats->total = ktime_add(stats->total,
+				 ktime_sub(ktime_get(), stats->start));
+
+	write_seqcount_end(&stats->lock);
+	local_irq_restore(flags);
+}
+
+#endif /* __INTEL_ENGINE_STATS_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h
new file mode 100644
index 000000000..107f465a2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h
@@ -0,0 +1,666 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_ENGINE_TYPES__
+#define __INTEL_ENGINE_TYPES__
+
+#include <linux/average.h>
+#include <linux/hashtable.h>
+#include <linux/irq_work.h>
+#include <linux/kref.h>
+#include <linux/list.h>
+#include <linux/llist.h>
+#include <linux/rbtree.h>
+#include <linux/timer.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+
+#include "i915_gem.h"
+#include "i915_pmu.h"
+#include "i915_priolist_types.h"
+#include "i915_selftest.h"
+#include "intel_sseu.h"
+#include "intel_timeline_types.h"
+#include "intel_uncore.h"
+#include "intel_wakeref.h"
+#include "intel_workarounds_types.h"
+
+/* HW Engine class + instance */
+#define RENDER_CLASS		0
+#define VIDEO_DECODE_CLASS	1
+#define VIDEO_ENHANCEMENT_CLASS	2
+#define COPY_ENGINE_CLASS	3
+#define OTHER_CLASS		4
+#define COMPUTE_CLASS		5
+#define MAX_ENGINE_CLASS	5
+#define MAX_ENGINE_INSTANCE	8
+
+#define I915_MAX_SLICES	3
+#define I915_MAX_SUBSLICES 8
+
+#define I915_CMD_HASH_ORDER 9
+
+struct dma_fence;
+struct drm_i915_gem_object;
+struct drm_i915_reg_table;
+struct i915_gem_context;
+struct i915_request;
+struct i915_sched_attr;
+struct i915_sched_engine;
+struct intel_gt;
+struct intel_ring;
+struct intel_uncore;
+struct intel_breadcrumbs;
+
+typedef u32 intel_engine_mask_t;
+#define ALL_ENGINES ((intel_engine_mask_t)~0ul)
+#define VIRTUAL_ENGINES BIT(BITS_PER_TYPE(intel_engine_mask_t) - 1)
+
+struct intel_hw_status_page {
+	struct list_head timelines;
+	struct i915_vma *vma;
+	u32 *addr;
+};
+
+struct intel_instdone {
+	u32 instdone;
+	/* The following exist only in the RCS engine */
+	u32 slice_common;
+	u32 slice_common_extra[2];
+	u32 sampler[GEN_MAX_GSLICES][I915_MAX_SUBSLICES];
+	u32 row[GEN_MAX_GSLICES][I915_MAX_SUBSLICES];
+
+	/* Added in XeHPG */
+	u32 geom_svg[GEN_MAX_GSLICES][I915_MAX_SUBSLICES];
+};
+
+/*
+ * we use a single page to load ctx workarounds so all of these
+ * values are referred in terms of dwords
+ *
+ * struct i915_wa_ctx_bb:
+ *  offset: specifies batch starting position, also helpful in case
+ *    if we want to have multiple batches at different offsets based on
+ *    some criteria. It is not a requirement at the moment but provides
+ *    an option for future use.
+ *  size: size of the batch in DWORDS
+ */
+struct i915_ctx_workarounds {
+	struct i915_wa_ctx_bb {
+		u32 offset;
+		u32 size;
+	} indirect_ctx, per_ctx;
+	struct i915_vma *vma;
+};
+
+#define I915_MAX_VCS	8
+#define I915_MAX_VECS	4
+#define I915_MAX_SFC	(I915_MAX_VCS / 2)
+#define I915_MAX_CCS	4
+#define I915_MAX_RCS	1
+#define I915_MAX_BCS	9
+
+/*
+ * Engine IDs definitions.
+ * Keep instances of the same type engine together.
+ */
+enum intel_engine_id {
+	RCS0 = 0,
+	BCS0,
+	BCS1,
+	BCS2,
+	BCS3,
+	BCS4,
+	BCS5,
+	BCS6,
+	BCS7,
+	BCS8,
+#define _BCS(n) (BCS0 + (n))
+	VCS0,
+	VCS1,
+	VCS2,
+	VCS3,
+	VCS4,
+	VCS5,
+	VCS6,
+	VCS7,
+#define _VCS(n) (VCS0 + (n))
+	VECS0,
+	VECS1,
+	VECS2,
+	VECS3,
+#define _VECS(n) (VECS0 + (n))
+	CCS0,
+	CCS1,
+	CCS2,
+	CCS3,
+#define _CCS(n) (CCS0 + (n))
+	I915_NUM_ENGINES
+#define INVALID_ENGINE ((enum intel_engine_id)-1)
+};
+
+/* A simple estimator for the round-trip latency of an engine */
+DECLARE_EWMA(_engine_latency, 6, 4)
+
+struct st_preempt_hang {
+	struct completion completion;
+	unsigned int count;
+};
+
+/**
+ * struct intel_engine_execlists - execlist submission queue and port state
+ *
+ * The struct intel_engine_execlists represents the combined logical state of
+ * driver and the hardware state for execlist mode of submission.
+ */
+struct intel_engine_execlists {
+	/**
+	 * @timer: kick the current context if its timeslice expires
+	 */
+	struct timer_list timer;
+
+	/**
+	 * @preempt: reset the current context if it fails to give way
+	 */
+	struct timer_list preempt;
+
+	/**
+	 * @preempt_target: active request at the time of the preemption request
+	 *
+	 * We force a preemption to occur if the pending contexts have not
+	 * been promoted to active upon receipt of the CS ack event within
+	 * the timeout. This timeout maybe chosen based on the target,
+	 * using a very short timeout if the context is no longer schedulable.
+	 * That short timeout may not be applicable to other contexts, so
+	 * if a context switch should happen within before the preemption
+	 * timeout, we may shoot early at an innocent context. To prevent this,
+	 * we record which context was active at the time of the preemption
+	 * request and only reset that context upon the timeout.
+	 */
+	const struct i915_request *preempt_target;
+
+	/**
+	 * @ccid: identifier for contexts submitted to this engine
+	 */
+	u32 ccid;
+
+	/**
+	 * @yield: CCID at the time of the last semaphore-wait interrupt.
+	 *
+	 * Instead of leaving a semaphore busy-spinning on an engine, we would
+	 * like to switch to another ready context, i.e. yielding the semaphore
+	 * timeslice.
+	 */
+	u32 yield;
+
+	/**
+	 * @error_interrupt: CS Master EIR
+	 *
+	 * The CS generates an interrupt when it detects an error. We capture
+	 * the first error interrupt, record the EIR and schedule the tasklet.
+	 * In the tasklet, we process the pending CS events to ensure we have
+	 * the guilty request, and then reset the engine.
+	 *
+	 * Low 16b are used by HW, with the upper 16b used as the enabling mask.
+	 * Reserve the upper 16b for tracking internal errors.
+	 */
+	u32 error_interrupt;
+#define ERROR_CSB	BIT(31)
+#define ERROR_PREEMPT	BIT(30)
+
+	/**
+	 * @reset_ccid: Active CCID [EXECLISTS_STATUS_HI] at the time of reset
+	 */
+	u32 reset_ccid;
+
+	/**
+	 * @submit_reg: gen-specific execlist submission register
+	 * set to the ExecList Submission Port (elsp) register pre-Gen11 and to
+	 * the ExecList Submission Queue Contents register array for Gen11+
+	 */
+	u32 __iomem *submit_reg;
+
+	/**
+	 * @ctrl_reg: the enhanced execlists control register, used to load the
+	 * submit queue on the HW and to request preemptions to idle
+	 */
+	u32 __iomem *ctrl_reg;
+
+#define EXECLIST_MAX_PORTS 2
+	/**
+	 * @active: the currently known context executing on HW
+	 */
+	struct i915_request * const *active;
+	/**
+	 * @inflight: the set of contexts submitted and acknowleged by HW
+	 *
+	 * The set of inflight contexts is managed by reading CS events
+	 * from the HW. On a context-switch event (not preemption), we
+	 * know the HW has transitioned from port0 to port1, and we
+	 * advance our inflight/active tracking accordingly.
+	 */
+	struct i915_request *inflight[EXECLIST_MAX_PORTS + 1 /* sentinel */];
+	/**
+	 * @pending: the next set of contexts submitted to ELSP
+	 *
+	 * We store the array of contexts that we submit to HW (via ELSP) and
+	 * promote them to the inflight array once HW has signaled the
+	 * preemption or idle-to-active event.
+	 */
+	struct i915_request *pending[EXECLIST_MAX_PORTS + 1];
+
+	/**
+	 * @port_mask: number of execlist ports - 1
+	 */
+	unsigned int port_mask;
+
+	/**
+	 * @virtual: Queue of requets on a virtual engine, sorted by priority.
+	 * Each RB entry is a struct i915_priolist containing a list of requests
+	 * of the same priority.
+	 */
+	struct rb_root_cached virtual;
+
+	/**
+	 * @csb_write: control register for Context Switch buffer
+	 *
+	 * Note this register may be either mmio or HWSP shadow.
+	 */
+	u32 *csb_write;
+
+	/**
+	 * @csb_status: status array for Context Switch buffer
+	 *
+	 * Note these register may be either mmio or HWSP shadow.
+	 */
+	u64 *csb_status;
+
+	/**
+	 * @csb_size: context status buffer FIFO size
+	 */
+	u8 csb_size;
+
+	/**
+	 * @csb_head: context status buffer head
+	 */
+	u8 csb_head;
+
+	I915_SELFTEST_DECLARE(struct st_preempt_hang preempt_hang;)
+};
+
+#define INTEL_ENGINE_CS_MAX_NAME 8
+
+struct intel_engine_execlists_stats {
+	/**
+	 * @active: Number of contexts currently scheduled in.
+	 */
+	unsigned int active;
+
+	/**
+	 * @lock: Lock protecting the below fields.
+	 */
+	seqcount_t lock;
+
+	/**
+	 * @total: Total time this engine was busy.
+	 *
+	 * Accumulated time not counting the most recent block in cases where
+	 * engine is currently busy (active > 0).
+	 */
+	ktime_t total;
+
+	/**
+	 * @start: Timestamp of the last idle to active transition.
+	 *
+	 * Idle is defined as active == 0, active is active > 0.
+	 */
+	ktime_t start;
+};
+
+struct intel_engine_guc_stats {
+	/**
+	 * @running: Active state of the engine when busyness was last sampled.
+	 */
+	bool running;
+
+	/**
+	 * @prev_total: Previous value of total runtime clock cycles.
+	 */
+	u32 prev_total;
+
+	/**
+	 * @total_gt_clks: Total gt clock cycles this engine was busy.
+	 */
+	u64 total_gt_clks;
+
+	/**
+	 * @start_gt_clk: GT clock time of last idle to active transition.
+	 */
+	u64 start_gt_clk;
+};
+
+struct intel_engine_cs {
+	struct drm_i915_private *i915;
+	struct intel_gt *gt;
+	struct intel_uncore *uncore;
+	char name[INTEL_ENGINE_CS_MAX_NAME];
+
+	enum intel_engine_id id;
+	enum intel_engine_id legacy_idx;
+
+	unsigned int guc_id;
+
+	intel_engine_mask_t mask;
+	u32 reset_domain;
+	/**
+	 * @logical_mask: logical mask of engine, reported to user space via
+	 * query IOCTL and used to communicate with the GuC in logical space.
+	 * The logical instance of a physical engine can change based on product
+	 * and fusing.
+	 */
+	intel_engine_mask_t logical_mask;
+
+	u8 class;
+	u8 instance;
+
+	u16 uabi_class;
+	u16 uabi_instance;
+
+	u32 uabi_capabilities;
+	u32 context_size;
+	u32 mmio_base;
+
+	/*
+	 * Some w/a require forcewake to be held (which prevents RC6) while
+	 * a particular engine is active. If so, we set fw_domain to which
+	 * domains need to be held for the duration of request activity,
+	 * and 0 if none. We try to limit the duration of the hold as much
+	 * as possible.
+	 */
+	enum forcewake_domains fw_domain;
+	unsigned int fw_active;
+
+	unsigned long context_tag;
+
+	struct rb_node uabi_node;
+
+	struct intel_sseu sseu;
+
+	struct i915_sched_engine *sched_engine;
+
+	/* keep a request in reserve for a [pm] barrier under oom */
+	struct i915_request *request_pool;
+
+	struct intel_context *hung_ce;
+
+	struct llist_head barrier_tasks;
+
+	struct intel_context *kernel_context; /* pinned */
+
+	/**
+	 * pinned_contexts_list: List of pinned contexts. This list is only
+	 * assumed to be manipulated during driver load- or unload time and
+	 * does therefore not have any additional protection.
+	 */
+	struct list_head pinned_contexts_list;
+
+	intel_engine_mask_t saturated; /* submitting semaphores too late? */
+
+	struct {
+		struct delayed_work work;
+		struct i915_request *systole;
+		unsigned long blocked;
+	} heartbeat;
+
+	unsigned long serial;
+
+	unsigned long wakeref_serial;
+	struct intel_wakeref wakeref;
+	struct file *default_state;
+
+	struct {
+		struct intel_ring *ring;
+		struct intel_timeline *timeline;
+	} legacy;
+
+	/*
+	 * We track the average duration of the idle pulse on parking the
+	 * engine to keep an estimate of the how the fast the engine is
+	 * under ideal conditions.
+	 */
+	struct ewma__engine_latency latency;
+
+	/* Keep track of all the seqno used, a trail of breadcrumbs */
+	struct intel_breadcrumbs *breadcrumbs;
+
+	struct intel_engine_pmu {
+		/**
+		 * @enable: Bitmask of enable sample events on this engine.
+		 *
+		 * Bits correspond to sample event types, for instance
+		 * I915_SAMPLE_QUEUED is bit 0 etc.
+		 */
+		u32 enable;
+		/**
+		 * @enable_count: Reference count for the enabled samplers.
+		 *
+		 * Index number corresponds to @enum drm_i915_pmu_engine_sample.
+		 */
+		unsigned int enable_count[I915_ENGINE_SAMPLE_COUNT];
+		/**
+		 * @sample: Counter values for sampling events.
+		 *
+		 * Our internal timer stores the current counters in this field.
+		 *
+		 * Index number corresponds to @enum drm_i915_pmu_engine_sample.
+		 */
+		struct i915_pmu_sample sample[I915_ENGINE_SAMPLE_COUNT];
+	} pmu;
+
+	struct intel_hw_status_page status_page;
+	struct i915_ctx_workarounds wa_ctx;
+	struct i915_wa_list ctx_wa_list;
+	struct i915_wa_list wa_list;
+	struct i915_wa_list whitelist;
+
+	u32             irq_keep_mask; /* always keep these interrupts */
+	u32		irq_enable_mask; /* bitmask to enable ring interrupt */
+	void		(*irq_enable)(struct intel_engine_cs *engine);
+	void		(*irq_disable)(struct intel_engine_cs *engine);
+	void		(*irq_handler)(struct intel_engine_cs *engine, u16 iir);
+
+	void		(*sanitize)(struct intel_engine_cs *engine);
+	int		(*resume)(struct intel_engine_cs *engine);
+
+	struct {
+		void (*prepare)(struct intel_engine_cs *engine);
+
+		void (*rewind)(struct intel_engine_cs *engine, bool stalled);
+		void (*cancel)(struct intel_engine_cs *engine);
+
+		void (*finish)(struct intel_engine_cs *engine);
+	} reset;
+
+	void		(*park)(struct intel_engine_cs *engine);
+	void		(*unpark)(struct intel_engine_cs *engine);
+
+	void		(*bump_serial)(struct intel_engine_cs *engine);
+
+	void		(*set_default_submission)(struct intel_engine_cs *engine);
+
+	const struct intel_context_ops *cops;
+
+	int		(*request_alloc)(struct i915_request *rq);
+
+	int		(*emit_flush)(struct i915_request *request, u32 mode);
+#define EMIT_INVALIDATE	BIT(0)
+#define EMIT_FLUSH	BIT(1)
+#define EMIT_BARRIER	(EMIT_INVALIDATE | EMIT_FLUSH)
+	int		(*emit_bb_start)(struct i915_request *rq,
+					 u64 offset, u32 length,
+					 unsigned int dispatch_flags);
+#define I915_DISPATCH_SECURE BIT(0)
+#define I915_DISPATCH_PINNED BIT(1)
+	int		 (*emit_init_breadcrumb)(struct i915_request *rq);
+	u32		*(*emit_fini_breadcrumb)(struct i915_request *rq,
+						 u32 *cs);
+	unsigned int	emit_fini_breadcrumb_dw;
+
+	/* Pass the request to the hardware queue (e.g. directly into
+	 * the legacy ringbuffer or to the end of an execlist).
+	 *
+	 * This is called from an atomic context with irqs disabled; must
+	 * be irq safe.
+	 */
+	void		(*submit_request)(struct i915_request *rq);
+
+	void		(*release)(struct intel_engine_cs *engine);
+
+	/*
+	 * Add / remove request from engine active tracking
+	 */
+	void		(*add_active_request)(struct i915_request *rq);
+	void		(*remove_active_request)(struct i915_request *rq);
+
+	/*
+	 * Get engine busyness and the time at which the busyness was sampled.
+	 */
+	ktime_t		(*busyness)(struct intel_engine_cs *engine,
+				    ktime_t *now);
+
+	struct intel_engine_execlists execlists;
+
+	/*
+	 * Keep track of completed timelines on this engine for early
+	 * retirement with the goal of quickly enabling powersaving as
+	 * soon as the engine is idle.
+	 */
+	struct intel_timeline *retire;
+	struct work_struct retire_work;
+
+	/* status_notifier: list of callbacks for context-switch changes */
+	struct atomic_notifier_head context_status_notifier;
+
+#define I915_ENGINE_USING_CMD_PARSER BIT(0)
+#define I915_ENGINE_SUPPORTS_STATS   BIT(1)
+#define I915_ENGINE_HAS_PREEMPTION   BIT(2)
+#define I915_ENGINE_HAS_SEMAPHORES   BIT(3)
+#define I915_ENGINE_HAS_TIMESLICES   BIT(4)
+#define I915_ENGINE_IS_VIRTUAL       BIT(5)
+#define I915_ENGINE_HAS_RELATIVE_MMIO BIT(6)
+#define I915_ENGINE_REQUIRES_CMD_PARSER BIT(7)
+#define I915_ENGINE_WANT_FORCED_PREEMPTION BIT(8)
+#define I915_ENGINE_HAS_RCS_REG_STATE  BIT(9)
+#define I915_ENGINE_HAS_EU_PRIORITY    BIT(10)
+#define I915_ENGINE_FIRST_RENDER_COMPUTE BIT(11)
+#define I915_ENGINE_USES_WA_HOLD_CCS_SWITCHOUT BIT(12)
+	unsigned int flags;
+
+	/*
+	 * Table of commands the command parser needs to know about
+	 * for this engine.
+	 */
+	DECLARE_HASHTABLE(cmd_hash, I915_CMD_HASH_ORDER);
+
+	/*
+	 * Table of registers allowed in commands that read/write registers.
+	 */
+	const struct drm_i915_reg_table *reg_tables;
+	int reg_table_count;
+
+	/*
+	 * Returns the bitmask for the length field of the specified command.
+	 * Return 0 for an unrecognized/invalid command.
+	 *
+	 * If the command parser finds an entry for a command in the engine's
+	 * cmd_tables, it gets the command's length based on the table entry.
+	 * If not, it calls this function to determine the per-engine length
+	 * field encoding for the command (i.e. different opcode ranges use
+	 * certain bits to encode the command length in the header).
+	 */
+	u32 (*get_cmd_length_mask)(u32 cmd_header);
+
+	struct {
+		union {
+			struct intel_engine_execlists_stats execlists;
+			struct intel_engine_guc_stats guc;
+		};
+
+		/**
+		 * @rps: Utilisation at last RPS sampling.
+		 */
+		ktime_t rps;
+	} stats;
+
+	struct {
+		unsigned long heartbeat_interval_ms;
+		unsigned long max_busywait_duration_ns;
+		unsigned long preempt_timeout_ms;
+		unsigned long stop_timeout_ms;
+		unsigned long timeslice_duration_ms;
+	} props, defaults;
+
+	I915_SELFTEST_DECLARE(struct fault_attr reset_timeout);
+};
+
+static inline bool
+intel_engine_using_cmd_parser(const struct intel_engine_cs *engine)
+{
+	return engine->flags & I915_ENGINE_USING_CMD_PARSER;
+}
+
+static inline bool
+intel_engine_requires_cmd_parser(const struct intel_engine_cs *engine)
+{
+	return engine->flags & I915_ENGINE_REQUIRES_CMD_PARSER;
+}
+
+static inline bool
+intel_engine_supports_stats(const struct intel_engine_cs *engine)
+{
+	return engine->flags & I915_ENGINE_SUPPORTS_STATS;
+}
+
+static inline bool
+intel_engine_has_preemption(const struct intel_engine_cs *engine)
+{
+	return engine->flags & I915_ENGINE_HAS_PREEMPTION;
+}
+
+static inline bool
+intel_engine_has_semaphores(const struct intel_engine_cs *engine)
+{
+	return engine->flags & I915_ENGINE_HAS_SEMAPHORES;
+}
+
+static inline bool
+intel_engine_has_timeslices(const struct intel_engine_cs *engine)
+{
+	if (!CONFIG_DRM_I915_TIMESLICE_DURATION)
+		return false;
+
+	return engine->flags & I915_ENGINE_HAS_TIMESLICES;
+}
+
+static inline bool
+intel_engine_is_virtual(const struct intel_engine_cs *engine)
+{
+	return engine->flags & I915_ENGINE_IS_VIRTUAL;
+}
+
+static inline bool
+intel_engine_has_relative_mmio(const struct intel_engine_cs * const engine)
+{
+	return engine->flags & I915_ENGINE_HAS_RELATIVE_MMIO;
+}
+
+/* Wa_14014475959:dg2 */
+static inline bool
+intel_engine_uses_wa_hold_ccs_switchout(struct intel_engine_cs *engine)
+{
+	return engine->flags & I915_ENGINE_USES_WA_HOLD_CCS_SWITCHOUT;
+}
+
+#endif /* __INTEL_ENGINE_TYPES_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_user.c b/drivers/gpu/drm/i915/gt/intel_engine_user.c
new file mode 100644
index 000000000..46a174f8a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_user.c
@@ -0,0 +1,308 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/list.h>
+#include <linux/list_sort.h>
+#include <linux/llist.h>
+
+#include "i915_drv.h"
+#include "intel_engine.h"
+#include "intel_engine_user.h"
+#include "intel_gt.h"
+#include "uc/intel_guc_submission.h"
+
+struct intel_engine_cs *
+intel_engine_lookup_user(struct drm_i915_private *i915, u8 class, u8 instance)
+{
+	struct rb_node *p = i915->uabi_engines.rb_node;
+
+	while (p) {
+		struct intel_engine_cs *it =
+			rb_entry(p, typeof(*it), uabi_node);
+
+		if (class < it->uabi_class)
+			p = p->rb_left;
+		else if (class > it->uabi_class ||
+			 instance > it->uabi_instance)
+			p = p->rb_right;
+		else if (instance < it->uabi_instance)
+			p = p->rb_left;
+		else
+			return it;
+	}
+
+	return NULL;
+}
+
+void intel_engine_add_user(struct intel_engine_cs *engine)
+{
+	llist_add((struct llist_node *)&engine->uabi_node,
+		  (struct llist_head *)&engine->i915->uabi_engines);
+}
+
+static const u8 uabi_classes[] = {
+	[RENDER_CLASS] = I915_ENGINE_CLASS_RENDER,
+	[COPY_ENGINE_CLASS] = I915_ENGINE_CLASS_COPY,
+	[VIDEO_DECODE_CLASS] = I915_ENGINE_CLASS_VIDEO,
+	[VIDEO_ENHANCEMENT_CLASS] = I915_ENGINE_CLASS_VIDEO_ENHANCE,
+	[COMPUTE_CLASS] = I915_ENGINE_CLASS_COMPUTE,
+};
+
+static int engine_cmp(void *priv, const struct list_head *A,
+		      const struct list_head *B)
+{
+	const struct intel_engine_cs *a =
+		container_of((struct rb_node *)A, typeof(*a), uabi_node);
+	const struct intel_engine_cs *b =
+		container_of((struct rb_node *)B, typeof(*b), uabi_node);
+
+	if (uabi_classes[a->class] < uabi_classes[b->class])
+		return -1;
+	if (uabi_classes[a->class] > uabi_classes[b->class])
+		return 1;
+
+	if (a->instance < b->instance)
+		return -1;
+	if (a->instance > b->instance)
+		return 1;
+
+	return 0;
+}
+
+static struct llist_node *get_engines(struct drm_i915_private *i915)
+{
+	return llist_del_all((struct llist_head *)&i915->uabi_engines);
+}
+
+static void sort_engines(struct drm_i915_private *i915,
+			 struct list_head *engines)
+{
+	struct llist_node *pos, *next;
+
+	llist_for_each_safe(pos, next, get_engines(i915)) {
+		struct intel_engine_cs *engine =
+			container_of((struct rb_node *)pos, typeof(*engine),
+				     uabi_node);
+		list_add((struct list_head *)&engine->uabi_node, engines);
+	}
+	list_sort(NULL, engines, engine_cmp);
+}
+
+static void set_scheduler_caps(struct drm_i915_private *i915)
+{
+	static const struct {
+		u8 engine;
+		u8 sched;
+	} map[] = {
+#define MAP(x, y) { ilog2(I915_ENGINE_##x), ilog2(I915_SCHEDULER_CAP_##y) }
+		MAP(HAS_PREEMPTION, PREEMPTION),
+		MAP(HAS_SEMAPHORES, SEMAPHORES),
+		MAP(SUPPORTS_STATS, ENGINE_BUSY_STATS),
+#undef MAP
+	};
+	struct intel_engine_cs *engine;
+	u32 enabled, disabled;
+
+	enabled = 0;
+	disabled = 0;
+	for_each_uabi_engine(engine, i915) { /* all engines must agree! */
+		int i;
+
+		if (engine->sched_engine->schedule)
+			enabled |= (I915_SCHEDULER_CAP_ENABLED |
+				    I915_SCHEDULER_CAP_PRIORITY);
+		else
+			disabled |= (I915_SCHEDULER_CAP_ENABLED |
+				     I915_SCHEDULER_CAP_PRIORITY);
+
+		if (intel_uc_uses_guc_submission(&to_gt(i915)->uc))
+			enabled |= I915_SCHEDULER_CAP_STATIC_PRIORITY_MAP;
+
+		for (i = 0; i < ARRAY_SIZE(map); i++) {
+			if (engine->flags & BIT(map[i].engine))
+				enabled |= BIT(map[i].sched);
+			else
+				disabled |= BIT(map[i].sched);
+		}
+	}
+
+	i915->caps.scheduler = enabled & ~disabled;
+	if (!(i915->caps.scheduler & I915_SCHEDULER_CAP_ENABLED))
+		i915->caps.scheduler = 0;
+}
+
+const char *intel_engine_class_repr(u8 class)
+{
+	static const char * const uabi_names[] = {
+		[RENDER_CLASS] = "rcs",
+		[COPY_ENGINE_CLASS] = "bcs",
+		[VIDEO_DECODE_CLASS] = "vcs",
+		[VIDEO_ENHANCEMENT_CLASS] = "vecs",
+		[COMPUTE_CLASS] = "ccs",
+	};
+
+	if (class >= ARRAY_SIZE(uabi_names) || !uabi_names[class])
+		return "xxx";
+
+	return uabi_names[class];
+}
+
+struct legacy_ring {
+	struct intel_gt *gt;
+	u8 class;
+	u8 instance;
+};
+
+static int legacy_ring_idx(const struct legacy_ring *ring)
+{
+	static const struct {
+		u8 base, max;
+	} map[] = {
+		[RENDER_CLASS] = { RCS0, 1 },
+		[COPY_ENGINE_CLASS] = { BCS0, 1 },
+		[VIDEO_DECODE_CLASS] = { VCS0, I915_MAX_VCS },
+		[VIDEO_ENHANCEMENT_CLASS] = { VECS0, I915_MAX_VECS },
+		[COMPUTE_CLASS] = { CCS0, I915_MAX_CCS },
+	};
+
+	if (GEM_DEBUG_WARN_ON(ring->class >= ARRAY_SIZE(map)))
+		return INVALID_ENGINE;
+
+	if (GEM_DEBUG_WARN_ON(ring->instance >= map[ring->class].max))
+		return INVALID_ENGINE;
+
+	return map[ring->class].base + ring->instance;
+}
+
+static void add_legacy_ring(struct legacy_ring *ring,
+			    struct intel_engine_cs *engine)
+{
+	if (engine->gt != ring->gt || engine->class != ring->class) {
+		ring->gt = engine->gt;
+		ring->class = engine->class;
+		ring->instance = 0;
+	}
+
+	engine->legacy_idx = legacy_ring_idx(ring);
+	if (engine->legacy_idx != INVALID_ENGINE)
+		ring->instance++;
+}
+
+void intel_engines_driver_register(struct drm_i915_private *i915)
+{
+	struct legacy_ring ring = {};
+	struct list_head *it, *next;
+	struct rb_node **p, *prev;
+	LIST_HEAD(engines);
+
+	sort_engines(i915, &engines);
+
+	prev = NULL;
+	p = &i915->uabi_engines.rb_node;
+	list_for_each_safe(it, next, &engines) {
+		struct intel_engine_cs *engine =
+			container_of((struct rb_node *)it, typeof(*engine),
+				     uabi_node);
+		char old[sizeof(engine->name)];
+
+		if (intel_gt_has_unrecoverable_error(engine->gt))
+			continue; /* ignore incomplete engines */
+
+		GEM_BUG_ON(engine->class >= ARRAY_SIZE(uabi_classes));
+		engine->uabi_class = uabi_classes[engine->class];
+
+		GEM_BUG_ON(engine->uabi_class >=
+			   ARRAY_SIZE(i915->engine_uabi_class_count));
+		engine->uabi_instance =
+			i915->engine_uabi_class_count[engine->uabi_class]++;
+
+		/* Replace the internal name with the final user facing name */
+		memcpy(old, engine->name, sizeof(engine->name));
+		scnprintf(engine->name, sizeof(engine->name), "%s%u",
+			  intel_engine_class_repr(engine->class),
+			  engine->uabi_instance);
+		DRM_DEBUG_DRIVER("renamed %s to %s\n", old, engine->name);
+
+		rb_link_node(&engine->uabi_node, prev, p);
+		rb_insert_color(&engine->uabi_node, &i915->uabi_engines);
+
+		GEM_BUG_ON(intel_engine_lookup_user(i915,
+						    engine->uabi_class,
+						    engine->uabi_instance) != engine);
+
+		/* Fix up the mapping to match default execbuf::user_map[] */
+		add_legacy_ring(&ring, engine);
+
+		prev = &engine->uabi_node;
+		p = &prev->rb_right;
+	}
+
+	if (IS_ENABLED(CONFIG_DRM_I915_SELFTESTS) &&
+	    IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)) {
+		struct intel_engine_cs *engine;
+		unsigned int isolation;
+		int class, inst;
+		int errors = 0;
+
+		for (class = 0; class < ARRAY_SIZE(i915->engine_uabi_class_count); class++) {
+			for (inst = 0; inst < i915->engine_uabi_class_count[class]; inst++) {
+				engine = intel_engine_lookup_user(i915,
+								  class, inst);
+				if (!engine) {
+					pr_err("UABI engine not found for { class:%d, instance:%d }\n",
+					       class, inst);
+					errors++;
+					continue;
+				}
+
+				if (engine->uabi_class != class ||
+				    engine->uabi_instance != inst) {
+					pr_err("Wrong UABI engine:%s { class:%d, instance:%d } found for { class:%d, instance:%d }\n",
+					       engine->name,
+					       engine->uabi_class,
+					       engine->uabi_instance,
+					       class, inst);
+					errors++;
+					continue;
+				}
+			}
+		}
+
+		/*
+		 * Make sure that classes with multiple engine instances all
+		 * share the same basic configuration.
+		 */
+		isolation = intel_engines_has_context_isolation(i915);
+		for_each_uabi_engine(engine, i915) {
+			unsigned int bit = BIT(engine->uabi_class);
+			unsigned int expected = engine->default_state ? bit : 0;
+
+			if ((isolation & bit) != expected) {
+				pr_err("mismatching default context state for class %d on engine %s\n",
+				       engine->uabi_class, engine->name);
+				errors++;
+			}
+		}
+
+		if (drm_WARN(&i915->drm, errors,
+			     "Invalid UABI engine mapping found"))
+			i915->uabi_engines = RB_ROOT;
+	}
+
+	set_scheduler_caps(i915);
+}
+
+unsigned int intel_engines_has_context_isolation(struct drm_i915_private *i915)
+{
+	struct intel_engine_cs *engine;
+	unsigned int which;
+
+	which = 0;
+	for_each_uabi_engine(engine, i915)
+		if (engine->default_state)
+			which |= BIT(engine->uabi_class);
+
+	return which;
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_user.h b/drivers/gpu/drm/i915/gt/intel_engine_user.h
new file mode 100644
index 000000000..3dc7e8ab9
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_engine_user.h
@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_ENGINE_USER_H
+#define INTEL_ENGINE_USER_H
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct intel_engine_cs;
+
+struct intel_engine_cs *
+intel_engine_lookup_user(struct drm_i915_private *i915, u8 class, u8 instance);
+
+unsigned int intel_engines_has_context_isolation(struct drm_i915_private *i915);
+
+void intel_engine_add_user(struct intel_engine_cs *engine);
+void intel_engines_driver_register(struct drm_i915_private *i915);
+
+const char *intel_engine_class_repr(u8 class);
+
+#endif /* INTEL_ENGINE_USER_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
new file mode 100644
index 000000000..f903ee1ce
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
@@ -0,0 +1,4186 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+/**
+ * DOC: Logical Rings, Logical Ring Contexts and Execlists
+ *
+ * Motivation:
+ * GEN8 brings an expansion of the HW contexts: "Logical Ring Contexts".
+ * These expanded contexts enable a number of new abilities, especially
+ * "Execlists" (also implemented in this file).
+ *
+ * One of the main differences with the legacy HW contexts is that logical
+ * ring contexts incorporate many more things to the context's state, like
+ * PDPs or ringbuffer control registers:
+ *
+ * The reason why PDPs are included in the context is straightforward: as
+ * PPGTTs (per-process GTTs) are actually per-context, having the PDPs
+ * contained there mean you don't need to do a ppgtt->switch_mm yourself,
+ * instead, the GPU will do it for you on the context switch.
+ *
+ * But, what about the ringbuffer control registers (head, tail, etc..)?
+ * shouldn't we just need a set of those per engine command streamer? This is
+ * where the name "Logical Rings" starts to make sense: by virtualizing the
+ * rings, the engine cs shifts to a new "ring buffer" with every context
+ * switch. When you want to submit a workload to the GPU you: A) choose your
+ * context, B) find its appropriate virtualized ring, C) write commands to it
+ * and then, finally, D) tell the GPU to switch to that context.
+ *
+ * Instead of the legacy MI_SET_CONTEXT, the way you tell the GPU to switch
+ * to a contexts is via a context execution list, ergo "Execlists".
+ *
+ * LRC implementation:
+ * Regarding the creation of contexts, we have:
+ *
+ * - One global default context.
+ * - One local default context for each opened fd.
+ * - One local extra context for each context create ioctl call.
+ *
+ * Now that ringbuffers belong per-context (and not per-engine, like before)
+ * and that contexts are uniquely tied to a given engine (and not reusable,
+ * like before) we need:
+ *
+ * - One ringbuffer per-engine inside each context.
+ * - One backing object per-engine inside each context.
+ *
+ * The global default context starts its life with these new objects fully
+ * allocated and populated. The local default context for each opened fd is
+ * more complex, because we don't know at creation time which engine is going
+ * to use them. To handle this, we have implemented a deferred creation of LR
+ * contexts:
+ *
+ * The local context starts its life as a hollow or blank holder, that only
+ * gets populated for a given engine once we receive an execbuffer. If later
+ * on we receive another execbuffer ioctl for the same context but a different
+ * engine, we allocate/populate a new ringbuffer and context backing object and
+ * so on.
+ *
+ * Finally, regarding local contexts created using the ioctl call: as they are
+ * only allowed with the render ring, we can allocate & populate them right
+ * away (no need to defer anything, at least for now).
+ *
+ * Execlists implementation:
+ * Execlists are the new method by which, on gen8+ hardware, workloads are
+ * submitted for execution (as opposed to the legacy, ringbuffer-based, method).
+ * This method works as follows:
+ *
+ * When a request is committed, its commands (the BB start and any leading or
+ * trailing commands, like the seqno breadcrumbs) are placed in the ringbuffer
+ * for the appropriate context. The tail pointer in the hardware context is not
+ * updated at this time, but instead, kept by the driver in the ringbuffer
+ * structure. A structure representing this request is added to a request queue
+ * for the appropriate engine: this structure contains a copy of the context's
+ * tail after the request was written to the ring buffer and a pointer to the
+ * context itself.
+ *
+ * If the engine's request queue was empty before the request was added, the
+ * queue is processed immediately. Otherwise the queue will be processed during
+ * a context switch interrupt. In any case, elements on the queue will get sent
+ * (in pairs) to the GPU's ExecLists Submit Port (ELSP, for short) with a
+ * globally unique 20-bits submission ID.
+ *
+ * When execution of a request completes, the GPU updates the context status
+ * buffer with a context complete event and generates a context switch interrupt.
+ * During the interrupt handling, the driver examines the events in the buffer:
+ * for each context complete event, if the announced ID matches that on the head
+ * of the request queue, then that request is retired and removed from the queue.
+ *
+ * After processing, if any requests were retired and the queue is not empty
+ * then a new execution list can be submitted. The two requests at the front of
+ * the queue are next to be submitted but since a context may not occur twice in
+ * an execution list, if subsequent requests have the same ID as the first then
+ * the two requests must be combined. This is done simply by discarding requests
+ * at the head of the queue until either only one requests is left (in which case
+ * we use a NULL second context) or the first two requests have unique IDs.
+ *
+ * By always executing the first two requests in the queue the driver ensures
+ * that the GPU is kept as busy as possible. In the case where a single context
+ * completes but a second context is still executing, the request for this second
+ * context will be at the head of the queue when we remove the first one. This
+ * request will then be resubmitted along with a new request for a different context,
+ * which will cause the hardware to continue executing the second request and queue
+ * the new request (the GPU detects the condition of a context getting preempted
+ * with the same context and optimizes the context switch flow by not doing
+ * preemption, but just sampling the new tail pointer).
+ *
+ */
+#include <linux/interrupt.h>
+#include <linux/string_helpers.h>
+
+#include "i915_drv.h"
+#include "i915_trace.h"
+#include "i915_vgpu.h"
+#include "gen8_engine_cs.h"
+#include "intel_breadcrumbs.h"
+#include "intel_context.h"
+#include "intel_engine_heartbeat.h"
+#include "intel_engine_pm.h"
+#include "intel_engine_regs.h"
+#include "intel_engine_stats.h"
+#include "intel_execlists_submission.h"
+#include "intel_gt.h"
+#include "intel_gt_irq.h"
+#include "intel_gt_pm.h"
+#include "intel_gt_regs.h"
+#include "intel_gt_requests.h"
+#include "intel_lrc.h"
+#include "intel_lrc_reg.h"
+#include "intel_mocs.h"
+#include "intel_reset.h"
+#include "intel_ring.h"
+#include "intel_workarounds.h"
+#include "shmem_utils.h"
+
+#define RING_EXECLIST_QFULL		(1 << 0x2)
+#define RING_EXECLIST1_VALID		(1 << 0x3)
+#define RING_EXECLIST0_VALID		(1 << 0x4)
+#define RING_EXECLIST_ACTIVE_STATUS	(3 << 0xE)
+#define RING_EXECLIST1_ACTIVE		(1 << 0x11)
+#define RING_EXECLIST0_ACTIVE		(1 << 0x12)
+
+#define GEN8_CTX_STATUS_IDLE_ACTIVE	(1 << 0)
+#define GEN8_CTX_STATUS_PREEMPTED	(1 << 1)
+#define GEN8_CTX_STATUS_ELEMENT_SWITCH	(1 << 2)
+#define GEN8_CTX_STATUS_ACTIVE_IDLE	(1 << 3)
+#define GEN8_CTX_STATUS_COMPLETE	(1 << 4)
+#define GEN8_CTX_STATUS_LITE_RESTORE	(1 << 15)
+
+#define GEN8_CTX_STATUS_COMPLETED_MASK \
+	 (GEN8_CTX_STATUS_COMPLETE | GEN8_CTX_STATUS_PREEMPTED)
+
+#define GEN12_CTX_STATUS_SWITCHED_TO_NEW_QUEUE	(0x1) /* lower csb dword */
+#define GEN12_CTX_SWITCH_DETAIL(csb_dw)	((csb_dw) & 0xF) /* upper csb dword */
+#define GEN12_CSB_SW_CTX_ID_MASK		GENMASK(25, 15)
+#define GEN12_IDLE_CTX_ID		0x7FF
+#define GEN12_CSB_CTX_VALID(csb_dw) \
+	(FIELD_GET(GEN12_CSB_SW_CTX_ID_MASK, csb_dw) != GEN12_IDLE_CTX_ID)
+
+#define XEHP_CTX_STATUS_SWITCHED_TO_NEW_QUEUE	BIT(1) /* upper csb dword */
+#define XEHP_CSB_SW_CTX_ID_MASK			GENMASK(31, 10)
+#define XEHP_IDLE_CTX_ID			0xFFFF
+#define XEHP_CSB_CTX_VALID(csb_dw) \
+	(FIELD_GET(XEHP_CSB_SW_CTX_ID_MASK, csb_dw) != XEHP_IDLE_CTX_ID)
+
+/* Typical size of the average request (2 pipecontrols and a MI_BB) */
+#define EXECLISTS_REQUEST_SIZE 64 /* bytes */
+
+struct virtual_engine {
+	struct intel_engine_cs base;
+	struct intel_context context;
+	struct rcu_work rcu;
+
+	/*
+	 * We allow only a single request through the virtual engine at a time
+	 * (each request in the timeline waits for the completion fence of
+	 * the previous before being submitted). By restricting ourselves to
+	 * only submitting a single request, each request is placed on to a
+	 * physical to maximise load spreading (by virtue of the late greedy
+	 * scheduling -- each real engine takes the next available request
+	 * upon idling).
+	 */
+	struct i915_request *request;
+
+	/*
+	 * We keep a rbtree of available virtual engines inside each physical
+	 * engine, sorted by priority. Here we preallocate the nodes we need
+	 * for the virtual engine, indexed by physical_engine->id.
+	 */
+	struct ve_node {
+		struct rb_node rb;
+		int prio;
+	} nodes[I915_NUM_ENGINES];
+
+	/* And finally, which physical engines this virtual engine maps onto. */
+	unsigned int num_siblings;
+	struct intel_engine_cs *siblings[];
+};
+
+static struct virtual_engine *to_virtual_engine(struct intel_engine_cs *engine)
+{
+	GEM_BUG_ON(!intel_engine_is_virtual(engine));
+	return container_of(engine, struct virtual_engine, base);
+}
+
+static struct intel_context *
+execlists_create_virtual(struct intel_engine_cs **siblings, unsigned int count,
+			 unsigned long flags);
+
+static struct i915_request *
+__active_request(const struct intel_timeline * const tl,
+		 struct i915_request *rq,
+		 int error)
+{
+	struct i915_request *active = rq;
+
+	list_for_each_entry_from_reverse(rq, &tl->requests, link) {
+		if (__i915_request_is_complete(rq))
+			break;
+
+		if (error) {
+			i915_request_set_error_once(rq, error);
+			__i915_request_skip(rq);
+		}
+		active = rq;
+	}
+
+	return active;
+}
+
+static struct i915_request *
+active_request(const struct intel_timeline * const tl, struct i915_request *rq)
+{
+	return __active_request(tl, rq, 0);
+}
+
+static void ring_set_paused(const struct intel_engine_cs *engine, int state)
+{
+	/*
+	 * We inspect HWS_PREEMPT with a semaphore inside
+	 * engine->emit_fini_breadcrumb. If the dword is true,
+	 * the ring is paused as the semaphore will busywait
+	 * until the dword is false.
+	 */
+	engine->status_page.addr[I915_GEM_HWS_PREEMPT] = state;
+	if (state)
+		wmb();
+}
+
+static struct i915_priolist *to_priolist(struct rb_node *rb)
+{
+	return rb_entry(rb, struct i915_priolist, node);
+}
+
+static int rq_prio(const struct i915_request *rq)
+{
+	return READ_ONCE(rq->sched.attr.priority);
+}
+
+static int effective_prio(const struct i915_request *rq)
+{
+	int prio = rq_prio(rq);
+
+	/*
+	 * If this request is special and must not be interrupted at any
+	 * cost, so be it. Note we are only checking the most recent request
+	 * in the context and so may be masking an earlier vip request. It
+	 * is hoped that under the conditions where nopreempt is used, this
+	 * will not matter (i.e. all requests to that context will be
+	 * nopreempt for as long as desired).
+	 */
+	if (i915_request_has_nopreempt(rq))
+		prio = I915_PRIORITY_UNPREEMPTABLE;
+
+	return prio;
+}
+
+static int queue_prio(const struct i915_sched_engine *sched_engine)
+{
+	struct rb_node *rb;
+
+	rb = rb_first_cached(&sched_engine->queue);
+	if (!rb)
+		return INT_MIN;
+
+	return to_priolist(rb)->priority;
+}
+
+static int virtual_prio(const struct intel_engine_execlists *el)
+{
+	struct rb_node *rb = rb_first_cached(&el->virtual);
+
+	return rb ? rb_entry(rb, struct ve_node, rb)->prio : INT_MIN;
+}
+
+static bool need_preempt(const struct intel_engine_cs *engine,
+			 const struct i915_request *rq)
+{
+	int last_prio;
+
+	if (!intel_engine_has_semaphores(engine))
+		return false;
+
+	/*
+	 * Check if the current priority hint merits a preemption attempt.
+	 *
+	 * We record the highest value priority we saw during rescheduling
+	 * prior to this dequeue, therefore we know that if it is strictly
+	 * less than the current tail of ESLP[0], we do not need to force
+	 * a preempt-to-idle cycle.
+	 *
+	 * However, the priority hint is a mere hint that we may need to
+	 * preempt. If that hint is stale or we may be trying to preempt
+	 * ourselves, ignore the request.
+	 *
+	 * More naturally we would write
+	 *      prio >= max(0, last);
+	 * except that we wish to prevent triggering preemption at the same
+	 * priority level: the task that is running should remain running
+	 * to preserve FIFO ordering of dependencies.
+	 */
+	last_prio = max(effective_prio(rq), I915_PRIORITY_NORMAL - 1);
+	if (engine->sched_engine->queue_priority_hint <= last_prio)
+		return false;
+
+	/*
+	 * Check against the first request in ELSP[1], it will, thanks to the
+	 * power of PI, be the highest priority of that context.
+	 */
+	if (!list_is_last(&rq->sched.link, &engine->sched_engine->requests) &&
+	    rq_prio(list_next_entry(rq, sched.link)) > last_prio)
+		return true;
+
+	/*
+	 * If the inflight context did not trigger the preemption, then maybe
+	 * it was the set of queued requests? Pick the highest priority in
+	 * the queue (the first active priolist) and see if it deserves to be
+	 * running instead of ELSP[0].
+	 *
+	 * The highest priority request in the queue can not be either
+	 * ELSP[0] or ELSP[1] as, thanks again to PI, if it was the same
+	 * context, it's priority would not exceed ELSP[0] aka last_prio.
+	 */
+	return max(virtual_prio(&engine->execlists),
+		   queue_prio(engine->sched_engine)) > last_prio;
+}
+
+__maybe_unused static bool
+assert_priority_queue(const struct i915_request *prev,
+		      const struct i915_request *next)
+{
+	/*
+	 * Without preemption, the prev may refer to the still active element
+	 * which we refuse to let go.
+	 *
+	 * Even with preemption, there are times when we think it is better not
+	 * to preempt and leave an ostensibly lower priority request in flight.
+	 */
+	if (i915_request_is_active(prev))
+		return true;
+
+	return rq_prio(prev) >= rq_prio(next);
+}
+
+static struct i915_request *
+__unwind_incomplete_requests(struct intel_engine_cs *engine)
+{
+	struct i915_request *rq, *rn, *active = NULL;
+	struct list_head *pl;
+	int prio = I915_PRIORITY_INVALID;
+
+	lockdep_assert_held(&engine->sched_engine->lock);
+
+	list_for_each_entry_safe_reverse(rq, rn,
+					 &engine->sched_engine->requests,
+					 sched.link) {
+		if (__i915_request_is_complete(rq)) {
+			list_del_init(&rq->sched.link);
+			continue;
+		}
+
+		__i915_request_unsubmit(rq);
+
+		GEM_BUG_ON(rq_prio(rq) == I915_PRIORITY_INVALID);
+		if (rq_prio(rq) != prio) {
+			prio = rq_prio(rq);
+			pl = i915_sched_lookup_priolist(engine->sched_engine,
+							prio);
+		}
+		GEM_BUG_ON(i915_sched_engine_is_empty(engine->sched_engine));
+
+		list_move(&rq->sched.link, pl);
+		set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+
+		/* Check in case we rollback so far we wrap [size/2] */
+		if (intel_ring_direction(rq->ring,
+					 rq->tail,
+					 rq->ring->tail + 8) > 0)
+			rq->context->lrc.desc |= CTX_DESC_FORCE_RESTORE;
+
+		active = rq;
+	}
+
+	return active;
+}
+
+struct i915_request *
+execlists_unwind_incomplete_requests(struct intel_engine_execlists *execlists)
+{
+	struct intel_engine_cs *engine =
+		container_of(execlists, typeof(*engine), execlists);
+
+	return __unwind_incomplete_requests(engine);
+}
+
+static void
+execlists_context_status_change(struct i915_request *rq, unsigned long status)
+{
+	/*
+	 * Only used when GVT-g is enabled now. When GVT-g is disabled,
+	 * The compiler should eliminate this function as dead-code.
+	 */
+	if (!IS_ENABLED(CONFIG_DRM_I915_GVT))
+		return;
+
+	atomic_notifier_call_chain(&rq->engine->context_status_notifier,
+				   status, rq);
+}
+
+static void reset_active(struct i915_request *rq,
+			 struct intel_engine_cs *engine)
+{
+	struct intel_context * const ce = rq->context;
+	u32 head;
+
+	/*
+	 * The executing context has been cancelled. We want to prevent
+	 * further execution along this context and propagate the error on
+	 * to anything depending on its results.
+	 *
+	 * In __i915_request_submit(), we apply the -EIO and remove the
+	 * requests' payloads for any banned requests. But first, we must
+	 * rewind the context back to the start of the incomplete request so
+	 * that we do not jump back into the middle of the batch.
+	 *
+	 * We preserve the breadcrumbs and semaphores of the incomplete
+	 * requests so that inter-timeline dependencies (i.e other timelines)
+	 * remain correctly ordered. And we defer to __i915_request_submit()
+	 * so that all asynchronous waits are correctly handled.
+	 */
+	ENGINE_TRACE(engine, "{ reset rq=%llx:%lld }\n",
+		     rq->fence.context, rq->fence.seqno);
+
+	/* On resubmission of the active request, payload will be scrubbed */
+	if (__i915_request_is_complete(rq))
+		head = rq->tail;
+	else
+		head = __active_request(ce->timeline, rq, -EIO)->head;
+	head = intel_ring_wrap(ce->ring, head);
+
+	/* Scrub the context image to prevent replaying the previous batch */
+	lrc_init_regs(ce, engine, true);
+
+	/* We've switched away, so this should be a no-op, but intent matters */
+	ce->lrc.lrca = lrc_update_regs(ce, engine, head);
+}
+
+static bool bad_request(const struct i915_request *rq)
+{
+	return rq->fence.error && i915_request_started(rq);
+}
+
+static struct intel_engine_cs *
+__execlists_schedule_in(struct i915_request *rq)
+{
+	struct intel_engine_cs * const engine = rq->engine;
+	struct intel_context * const ce = rq->context;
+
+	intel_context_get(ce);
+
+	if (unlikely(intel_context_is_closed(ce) &&
+		     !intel_engine_has_heartbeat(engine)))
+		intel_context_set_exiting(ce);
+
+	if (unlikely(!intel_context_is_schedulable(ce) || bad_request(rq)))
+		reset_active(rq, engine);
+
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		lrc_check_regs(ce, engine, "before");
+
+	if (ce->tag) {
+		/* Use a fixed tag for OA and friends */
+		GEM_BUG_ON(ce->tag <= BITS_PER_LONG);
+		ce->lrc.ccid = ce->tag;
+	} else if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 50)) {
+		/* We don't need a strict matching tag, just different values */
+		unsigned int tag = ffs(READ_ONCE(engine->context_tag));
+
+		GEM_BUG_ON(tag == 0 || tag >= BITS_PER_LONG);
+		clear_bit(tag - 1, &engine->context_tag);
+		ce->lrc.ccid = tag << (XEHP_SW_CTX_ID_SHIFT - 32);
+
+		BUILD_BUG_ON(BITS_PER_LONG > GEN12_MAX_CONTEXT_HW_ID);
+
+	} else {
+		/* We don't need a strict matching tag, just different values */
+		unsigned int tag = __ffs(engine->context_tag);
+
+		GEM_BUG_ON(tag >= BITS_PER_LONG);
+		__clear_bit(tag, &engine->context_tag);
+		ce->lrc.ccid = (1 + tag) << (GEN11_SW_CTX_ID_SHIFT - 32);
+
+		BUILD_BUG_ON(BITS_PER_LONG > GEN12_MAX_CONTEXT_HW_ID);
+	}
+
+	ce->lrc.ccid |= engine->execlists.ccid;
+
+	__intel_gt_pm_get(engine->gt);
+	if (engine->fw_domain && !engine->fw_active++)
+		intel_uncore_forcewake_get(engine->uncore, engine->fw_domain);
+	execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_IN);
+	intel_engine_context_in(engine);
+
+	CE_TRACE(ce, "schedule-in, ccid:%x\n", ce->lrc.ccid);
+
+	return engine;
+}
+
+static void execlists_schedule_in(struct i915_request *rq, int idx)
+{
+	struct intel_context * const ce = rq->context;
+	struct intel_engine_cs *old;
+
+	GEM_BUG_ON(!intel_engine_pm_is_awake(rq->engine));
+	trace_i915_request_in(rq, idx);
+
+	old = ce->inflight;
+	if (!old)
+		old = __execlists_schedule_in(rq);
+	WRITE_ONCE(ce->inflight, ptr_inc(old));
+
+	GEM_BUG_ON(intel_context_inflight(ce) != rq->engine);
+}
+
+static void
+resubmit_virtual_request(struct i915_request *rq, struct virtual_engine *ve)
+{
+	struct intel_engine_cs *engine = rq->engine;
+
+	spin_lock_irq(&engine->sched_engine->lock);
+
+	clear_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+	WRITE_ONCE(rq->engine, &ve->base);
+	ve->base.submit_request(rq);
+
+	spin_unlock_irq(&engine->sched_engine->lock);
+}
+
+static void kick_siblings(struct i915_request *rq, struct intel_context *ce)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+	struct intel_engine_cs *engine = rq->engine;
+
+	/*
+	 * After this point, the rq may be transferred to a new sibling, so
+	 * before we clear ce->inflight make sure that the context has been
+	 * removed from the b->signalers and furthermore we need to make sure
+	 * that the concurrent iterator in signal_irq_work is no longer
+	 * following ce->signal_link.
+	 */
+	if (!list_empty(&ce->signals))
+		intel_context_remove_breadcrumbs(ce, engine->breadcrumbs);
+
+	/*
+	 * This engine is now too busy to run this virtual request, so
+	 * see if we can find an alternative engine for it to execute on.
+	 * Once a request has become bonded to this engine, we treat it the
+	 * same as other native request.
+	 */
+	if (i915_request_in_priority_queue(rq) &&
+	    rq->execution_mask != engine->mask)
+		resubmit_virtual_request(rq, ve);
+
+	if (READ_ONCE(ve->request))
+		tasklet_hi_schedule(&ve->base.sched_engine->tasklet);
+}
+
+static void __execlists_schedule_out(struct i915_request * const rq,
+				     struct intel_context * const ce)
+{
+	struct intel_engine_cs * const engine = rq->engine;
+	unsigned int ccid;
+
+	/*
+	 * NB process_csb() is not under the engine->sched_engine->lock and hence
+	 * schedule_out can race with schedule_in meaning that we should
+	 * refrain from doing non-trivial work here.
+	 */
+
+	CE_TRACE(ce, "schedule-out, ccid:%x\n", ce->lrc.ccid);
+	GEM_BUG_ON(ce->inflight != engine);
+
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		lrc_check_regs(ce, engine, "after");
+
+	/*
+	 * If we have just completed this context, the engine may now be
+	 * idle and we want to re-enter powersaving.
+	 */
+	if (intel_timeline_is_last(ce->timeline, rq) &&
+	    __i915_request_is_complete(rq))
+		intel_engine_add_retire(engine, ce->timeline);
+
+	ccid = ce->lrc.ccid;
+	if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 50)) {
+		ccid >>= XEHP_SW_CTX_ID_SHIFT - 32;
+		ccid &= XEHP_MAX_CONTEXT_HW_ID;
+	} else {
+		ccid >>= GEN11_SW_CTX_ID_SHIFT - 32;
+		ccid &= GEN12_MAX_CONTEXT_HW_ID;
+	}
+
+	if (ccid < BITS_PER_LONG) {
+		GEM_BUG_ON(ccid == 0);
+		GEM_BUG_ON(test_bit(ccid - 1, &engine->context_tag));
+		__set_bit(ccid - 1, &engine->context_tag);
+	}
+	intel_engine_context_out(engine);
+	execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_OUT);
+	if (engine->fw_domain && !--engine->fw_active)
+		intel_uncore_forcewake_put(engine->uncore, engine->fw_domain);
+	intel_gt_pm_put_async(engine->gt);
+
+	/*
+	 * If this is part of a virtual engine, its next request may
+	 * have been blocked waiting for access to the active context.
+	 * We have to kick all the siblings again in case we need to
+	 * switch (e.g. the next request is not runnable on this
+	 * engine). Hopefully, we will already have submitted the next
+	 * request before the tasklet runs and do not need to rebuild
+	 * each virtual tree and kick everyone again.
+	 */
+	if (ce->engine != engine)
+		kick_siblings(rq, ce);
+
+	WRITE_ONCE(ce->inflight, NULL);
+	intel_context_put(ce);
+}
+
+static inline void execlists_schedule_out(struct i915_request *rq)
+{
+	struct intel_context * const ce = rq->context;
+
+	trace_i915_request_out(rq);
+
+	GEM_BUG_ON(!ce->inflight);
+	ce->inflight = ptr_dec(ce->inflight);
+	if (!__intel_context_inflight_count(ce->inflight))
+		__execlists_schedule_out(rq, ce);
+
+	i915_request_put(rq);
+}
+
+static u32 map_i915_prio_to_lrc_desc_prio(int prio)
+{
+	if (prio > I915_PRIORITY_NORMAL)
+		return GEN12_CTX_PRIORITY_HIGH;
+	else if (prio < I915_PRIORITY_NORMAL)
+		return GEN12_CTX_PRIORITY_LOW;
+	else
+		return GEN12_CTX_PRIORITY_NORMAL;
+}
+
+static u64 execlists_update_context(struct i915_request *rq)
+{
+	struct intel_context *ce = rq->context;
+	u64 desc;
+	u32 tail, prev;
+
+	desc = ce->lrc.desc;
+	if (rq->engine->flags & I915_ENGINE_HAS_EU_PRIORITY)
+		desc |= map_i915_prio_to_lrc_desc_prio(rq_prio(rq));
+
+	/*
+	 * WaIdleLiteRestore:bdw,skl
+	 *
+	 * We should never submit the context with the same RING_TAIL twice
+	 * just in case we submit an empty ring, which confuses the HW.
+	 *
+	 * We append a couple of NOOPs (gen8_emit_wa_tail) after the end of
+	 * the normal request to be able to always advance the RING_TAIL on
+	 * subsequent resubmissions (for lite restore). Should that fail us,
+	 * and we try and submit the same tail again, force the context
+	 * reload.
+	 *
+	 * If we need to return to a preempted context, we need to skip the
+	 * lite-restore and force it to reload the RING_TAIL. Otherwise, the
+	 * HW has a tendency to ignore us rewinding the TAIL to the end of
+	 * an earlier request.
+	 */
+	GEM_BUG_ON(ce->lrc_reg_state[CTX_RING_TAIL] != rq->ring->tail);
+	prev = rq->ring->tail;
+	tail = intel_ring_set_tail(rq->ring, rq->tail);
+	if (unlikely(intel_ring_direction(rq->ring, tail, prev) <= 0))
+		desc |= CTX_DESC_FORCE_RESTORE;
+	ce->lrc_reg_state[CTX_RING_TAIL] = tail;
+	rq->tail = rq->wa_tail;
+
+	/*
+	 * Make sure the context image is complete before we submit it to HW.
+	 *
+	 * Ostensibly, writes (including the WCB) should be flushed prior to
+	 * an uncached write such as our mmio register access, the empirical
+	 * evidence (esp. on Braswell) suggests that the WC write into memory
+	 * may not be visible to the HW prior to the completion of the UC
+	 * register write and that we may begin execution from the context
+	 * before its image is complete leading to invalid PD chasing.
+	 */
+	wmb();
+
+	ce->lrc.desc &= ~CTX_DESC_FORCE_RESTORE;
+	return desc;
+}
+
+static void write_desc(struct intel_engine_execlists *execlists, u64 desc, u32 port)
+{
+	if (execlists->ctrl_reg) {
+		writel(lower_32_bits(desc), execlists->submit_reg + port * 2);
+		writel(upper_32_bits(desc), execlists->submit_reg + port * 2 + 1);
+	} else {
+		writel(upper_32_bits(desc), execlists->submit_reg);
+		writel(lower_32_bits(desc), execlists->submit_reg);
+	}
+}
+
+static __maybe_unused char *
+dump_port(char *buf, int buflen, const char *prefix, struct i915_request *rq)
+{
+	if (!rq)
+		return "";
+
+	snprintf(buf, buflen, "%sccid:%x %llx:%lld%s prio %d",
+		 prefix,
+		 rq->context->lrc.ccid,
+		 rq->fence.context, rq->fence.seqno,
+		 __i915_request_is_complete(rq) ? "!" :
+		 __i915_request_has_started(rq) ? "*" :
+		 "",
+		 rq_prio(rq));
+
+	return buf;
+}
+
+static __maybe_unused noinline void
+trace_ports(const struct intel_engine_execlists *execlists,
+	    const char *msg,
+	    struct i915_request * const *ports)
+{
+	const struct intel_engine_cs *engine =
+		container_of(execlists, typeof(*engine), execlists);
+	char __maybe_unused p0[40], p1[40];
+
+	if (!ports[0])
+		return;
+
+	ENGINE_TRACE(engine, "%s { %s%s }\n", msg,
+		     dump_port(p0, sizeof(p0), "", ports[0]),
+		     dump_port(p1, sizeof(p1), ", ", ports[1]));
+}
+
+static bool
+reset_in_progress(const struct intel_engine_cs *engine)
+{
+	return unlikely(!__tasklet_is_enabled(&engine->sched_engine->tasklet));
+}
+
+static __maybe_unused noinline bool
+assert_pending_valid(const struct intel_engine_execlists *execlists,
+		     const char *msg)
+{
+	struct intel_engine_cs *engine =
+		container_of(execlists, typeof(*engine), execlists);
+	struct i915_request * const *port, *rq, *prev = NULL;
+	struct intel_context *ce = NULL;
+	u32 ccid = -1;
+
+	trace_ports(execlists, msg, execlists->pending);
+
+	/* We may be messing around with the lists during reset, lalala */
+	if (reset_in_progress(engine))
+		return true;
+
+	if (!execlists->pending[0]) {
+		GEM_TRACE_ERR("%s: Nothing pending for promotion!\n",
+			      engine->name);
+		return false;
+	}
+
+	if (execlists->pending[execlists_num_ports(execlists)]) {
+		GEM_TRACE_ERR("%s: Excess pending[%d] for promotion!\n",
+			      engine->name, execlists_num_ports(execlists));
+		return false;
+	}
+
+	for (port = execlists->pending; (rq = *port); port++) {
+		unsigned long flags;
+		bool ok = true;
+
+		GEM_BUG_ON(!kref_read(&rq->fence.refcount));
+		GEM_BUG_ON(!i915_request_is_active(rq));
+
+		if (ce == rq->context) {
+			GEM_TRACE_ERR("%s: Dup context:%llx in pending[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			return false;
+		}
+		ce = rq->context;
+
+		if (ccid == ce->lrc.ccid) {
+			GEM_TRACE_ERR("%s: Dup ccid:%x context:%llx in pending[%zd]\n",
+				      engine->name,
+				      ccid, ce->timeline->fence_context,
+				      port - execlists->pending);
+			return false;
+		}
+		ccid = ce->lrc.ccid;
+
+		/*
+		 * Sentinels are supposed to be the last request so they flush
+		 * the current execution off the HW. Check that they are the only
+		 * request in the pending submission.
+		 *
+		 * NB: Due to the async nature of preempt-to-busy and request
+		 * cancellation we need to handle the case where request
+		 * becomes a sentinel in parallel to CSB processing.
+		 */
+		if (prev && i915_request_has_sentinel(prev) &&
+		    !READ_ONCE(prev->fence.error)) {
+			GEM_TRACE_ERR("%s: context:%llx after sentinel in pending[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			return false;
+		}
+		prev = rq;
+
+		/*
+		 * We want virtual requests to only be in the first slot so
+		 * that they are never stuck behind a hog and can be immediately
+		 * transferred onto the next idle engine.
+		 */
+		if (rq->execution_mask != engine->mask &&
+		    port != execlists->pending) {
+			GEM_TRACE_ERR("%s: virtual engine:%llx not in prime position[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			return false;
+		}
+
+		/* Hold tightly onto the lock to prevent concurrent retires! */
+		if (!spin_trylock_irqsave(&rq->lock, flags))
+			continue;
+
+		if (__i915_request_is_complete(rq))
+			goto unlock;
+
+		if (i915_active_is_idle(&ce->active) &&
+		    !intel_context_is_barrier(ce)) {
+			GEM_TRACE_ERR("%s: Inactive context:%llx in pending[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			ok = false;
+			goto unlock;
+		}
+
+		if (!i915_vma_is_pinned(ce->state)) {
+			GEM_TRACE_ERR("%s: Unpinned context:%llx in pending[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			ok = false;
+			goto unlock;
+		}
+
+		if (!i915_vma_is_pinned(ce->ring->vma)) {
+			GEM_TRACE_ERR("%s: Unpinned ring:%llx in pending[%zd]\n",
+				      engine->name,
+				      ce->timeline->fence_context,
+				      port - execlists->pending);
+			ok = false;
+			goto unlock;
+		}
+
+unlock:
+		spin_unlock_irqrestore(&rq->lock, flags);
+		if (!ok)
+			return false;
+	}
+
+	return ce;
+}
+
+static void execlists_submit_ports(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists *execlists = &engine->execlists;
+	unsigned int n;
+
+	GEM_BUG_ON(!assert_pending_valid(execlists, "submit"));
+
+	/*
+	 * We can skip acquiring intel_runtime_pm_get() here as it was taken
+	 * on our behalf by the request (see i915_gem_mark_busy()) and it will
+	 * not be relinquished until the device is idle (see
+	 * i915_gem_idle_work_handler()). As a precaution, we make sure
+	 * that all ELSP are drained i.e. we have processed the CSB,
+	 * before allowing ourselves to idle and calling intel_runtime_pm_put().
+	 */
+	GEM_BUG_ON(!intel_engine_pm_is_awake(engine));
+
+	/*
+	 * ELSQ note: the submit queue is not cleared after being submitted
+	 * to the HW so we need to make sure we always clean it up. This is
+	 * currently ensured by the fact that we always write the same number
+	 * of elsq entries, keep this in mind before changing the loop below.
+	 */
+	for (n = execlists_num_ports(execlists); n--; ) {
+		struct i915_request *rq = execlists->pending[n];
+
+		write_desc(execlists,
+			   rq ? execlists_update_context(rq) : 0,
+			   n);
+	}
+
+	/* we need to manually load the submit queue */
+	if (execlists->ctrl_reg)
+		writel(EL_CTRL_LOAD, execlists->ctrl_reg);
+}
+
+static bool ctx_single_port_submission(const struct intel_context *ce)
+{
+	return (IS_ENABLED(CONFIG_DRM_I915_GVT) &&
+		intel_context_force_single_submission(ce));
+}
+
+static bool can_merge_ctx(const struct intel_context *prev,
+			  const struct intel_context *next)
+{
+	if (prev != next)
+		return false;
+
+	if (ctx_single_port_submission(prev))
+		return false;
+
+	return true;
+}
+
+static unsigned long i915_request_flags(const struct i915_request *rq)
+{
+	return READ_ONCE(rq->fence.flags);
+}
+
+static bool can_merge_rq(const struct i915_request *prev,
+			 const struct i915_request *next)
+{
+	GEM_BUG_ON(prev == next);
+	GEM_BUG_ON(!assert_priority_queue(prev, next));
+
+	/*
+	 * We do not submit known completed requests. Therefore if the next
+	 * request is already completed, we can pretend to merge it in
+	 * with the previous context (and we will skip updating the ELSP
+	 * and tracking). Thus hopefully keeping the ELSP full with active
+	 * contexts, despite the best efforts of preempt-to-busy to confuse
+	 * us.
+	 */
+	if (__i915_request_is_complete(next))
+		return true;
+
+	if (unlikely((i915_request_flags(prev) | i915_request_flags(next)) &
+		     (BIT(I915_FENCE_FLAG_NOPREEMPT) |
+		      BIT(I915_FENCE_FLAG_SENTINEL))))
+		return false;
+
+	if (!can_merge_ctx(prev->context, next->context))
+		return false;
+
+	GEM_BUG_ON(i915_seqno_passed(prev->fence.seqno, next->fence.seqno));
+	return true;
+}
+
+static bool virtual_matches(const struct virtual_engine *ve,
+			    const struct i915_request *rq,
+			    const struct intel_engine_cs *engine)
+{
+	const struct intel_engine_cs *inflight;
+
+	if (!rq)
+		return false;
+
+	if (!(rq->execution_mask & engine->mask)) /* We peeked too soon! */
+		return false;
+
+	/*
+	 * We track when the HW has completed saving the context image
+	 * (i.e. when we have seen the final CS event switching out of
+	 * the context) and must not overwrite the context image before
+	 * then. This restricts us to only using the active engine
+	 * while the previous virtualized request is inflight (so
+	 * we reuse the register offsets). This is a very small
+	 * hystersis on the greedy seelction algorithm.
+	 */
+	inflight = intel_context_inflight(&ve->context);
+	if (inflight && inflight != engine)
+		return false;
+
+	return true;
+}
+
+static struct virtual_engine *
+first_virtual_engine(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists *el = &engine->execlists;
+	struct rb_node *rb = rb_first_cached(&el->virtual);
+
+	while (rb) {
+		struct virtual_engine *ve =
+			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
+		struct i915_request *rq = READ_ONCE(ve->request);
+
+		/* lazily cleanup after another engine handled rq */
+		if (!rq || !virtual_matches(ve, rq, engine)) {
+			rb_erase_cached(rb, &el->virtual);
+			RB_CLEAR_NODE(rb);
+			rb = rb_first_cached(&el->virtual);
+			continue;
+		}
+
+		return ve;
+	}
+
+	return NULL;
+}
+
+static void virtual_xfer_context(struct virtual_engine *ve,
+				 struct intel_engine_cs *engine)
+{
+	unsigned int n;
+
+	if (likely(engine == ve->siblings[0]))
+		return;
+
+	GEM_BUG_ON(READ_ONCE(ve->context.inflight));
+	if (!intel_engine_has_relative_mmio(engine))
+		lrc_update_offsets(&ve->context, engine);
+
+	/*
+	 * Move the bound engine to the top of the list for
+	 * future execution. We then kick this tasklet first
+	 * before checking others, so that we preferentially
+	 * reuse this set of bound registers.
+	 */
+	for (n = 1; n < ve->num_siblings; n++) {
+		if (ve->siblings[n] == engine) {
+			swap(ve->siblings[n], ve->siblings[0]);
+			break;
+		}
+	}
+}
+
+static void defer_request(struct i915_request *rq, struct list_head * const pl)
+{
+	LIST_HEAD(list);
+
+	/*
+	 * We want to move the interrupted request to the back of
+	 * the round-robin list (i.e. its priority level), but
+	 * in doing so, we must then move all requests that were in
+	 * flight and were waiting for the interrupted request to
+	 * be run after it again.
+	 */
+	do {
+		struct i915_dependency *p;
+
+		GEM_BUG_ON(i915_request_is_active(rq));
+		list_move_tail(&rq->sched.link, pl);
+
+		for_each_waiter(p, rq) {
+			struct i915_request *w =
+				container_of(p->waiter, typeof(*w), sched);
+
+			if (p->flags & I915_DEPENDENCY_WEAK)
+				continue;
+
+			/* Leave semaphores spinning on the other engines */
+			if (w->engine != rq->engine)
+				continue;
+
+			/* No waiter should start before its signaler */
+			GEM_BUG_ON(i915_request_has_initial_breadcrumb(w) &&
+				   __i915_request_has_started(w) &&
+				   !__i915_request_is_complete(rq));
+
+			if (!i915_request_is_ready(w))
+				continue;
+
+			if (rq_prio(w) < rq_prio(rq))
+				continue;
+
+			GEM_BUG_ON(rq_prio(w) > rq_prio(rq));
+			GEM_BUG_ON(i915_request_is_active(w));
+			list_move_tail(&w->sched.link, &list);
+		}
+
+		rq = list_first_entry_or_null(&list, typeof(*rq), sched.link);
+	} while (rq);
+}
+
+static void defer_active(struct intel_engine_cs *engine)
+{
+	struct i915_request *rq;
+
+	rq = __unwind_incomplete_requests(engine);
+	if (!rq)
+		return;
+
+	defer_request(rq, i915_sched_lookup_priolist(engine->sched_engine,
+						     rq_prio(rq)));
+}
+
+static bool
+timeslice_yield(const struct intel_engine_execlists *el,
+		const struct i915_request *rq)
+{
+	/*
+	 * Once bitten, forever smitten!
+	 *
+	 * If the active context ever busy-waited on a semaphore,
+	 * it will be treated as a hog until the end of its timeslice (i.e.
+	 * until it is scheduled out and replaced by a new submission,
+	 * possibly even its own lite-restore). The HW only sends an interrupt
+	 * on the first miss, and we do know if that semaphore has been
+	 * signaled, or even if it is now stuck on another semaphore. Play
+	 * safe, yield if it might be stuck -- it will be given a fresh
+	 * timeslice in the near future.
+	 */
+	return rq->context->lrc.ccid == READ_ONCE(el->yield);
+}
+
+static bool needs_timeslice(const struct intel_engine_cs *engine,
+			    const struct i915_request *rq)
+{
+	if (!intel_engine_has_timeslices(engine))
+		return false;
+
+	/* If not currently active, or about to switch, wait for next event */
+	if (!rq || __i915_request_is_complete(rq))
+		return false;
+
+	/* We do not need to start the timeslice until after the ACK */
+	if (READ_ONCE(engine->execlists.pending[0]))
+		return false;
+
+	/* If ELSP[1] is occupied, always check to see if worth slicing */
+	if (!list_is_last_rcu(&rq->sched.link,
+			      &engine->sched_engine->requests)) {
+		ENGINE_TRACE(engine, "timeslice required for second inflight context\n");
+		return true;
+	}
+
+	/* Otherwise, ELSP[0] is by itself, but may be waiting in the queue */
+	if (!i915_sched_engine_is_empty(engine->sched_engine)) {
+		ENGINE_TRACE(engine, "timeslice required for queue\n");
+		return true;
+	}
+
+	if (!RB_EMPTY_ROOT(&engine->execlists.virtual.rb_root)) {
+		ENGINE_TRACE(engine, "timeslice required for virtual\n");
+		return true;
+	}
+
+	return false;
+}
+
+static bool
+timeslice_expired(struct intel_engine_cs *engine, const struct i915_request *rq)
+{
+	const struct intel_engine_execlists *el = &engine->execlists;
+
+	if (i915_request_has_nopreempt(rq) && __i915_request_has_started(rq))
+		return false;
+
+	if (!needs_timeslice(engine, rq))
+		return false;
+
+	return timer_expired(&el->timer) || timeslice_yield(el, rq);
+}
+
+static unsigned long timeslice(const struct intel_engine_cs *engine)
+{
+	return READ_ONCE(engine->props.timeslice_duration_ms);
+}
+
+static void start_timeslice(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists *el = &engine->execlists;
+	unsigned long duration;
+
+	/* Disable the timer if there is nothing to switch to */
+	duration = 0;
+	if (needs_timeslice(engine, *el->active)) {
+		/* Avoid continually prolonging an active timeslice */
+		if (timer_active(&el->timer)) {
+			/*
+			 * If we just submitted a new ELSP after an old
+			 * context, that context may have already consumed
+			 * its timeslice, so recheck.
+			 */
+			if (!timer_pending(&el->timer))
+				tasklet_hi_schedule(&engine->sched_engine->tasklet);
+			return;
+		}
+
+		duration = timeslice(engine);
+	}
+
+	set_timer_ms(&el->timer, duration);
+}
+
+static void record_preemption(struct intel_engine_execlists *execlists)
+{
+	(void)I915_SELFTEST_ONLY(execlists->preempt_hang.count++);
+}
+
+static unsigned long active_preempt_timeout(struct intel_engine_cs *engine,
+					    const struct i915_request *rq)
+{
+	if (!rq)
+		return 0;
+
+	/* Only allow ourselves to force reset the currently active context */
+	engine->execlists.preempt_target = rq;
+
+	/* Force a fast reset for terminated contexts (ignoring sysfs!) */
+	if (unlikely(intel_context_is_banned(rq->context) || bad_request(rq)))
+		return INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS;
+
+	return READ_ONCE(engine->props.preempt_timeout_ms);
+}
+
+static void set_preempt_timeout(struct intel_engine_cs *engine,
+				const struct i915_request *rq)
+{
+	if (!intel_engine_has_preempt_reset(engine))
+		return;
+
+	set_timer_ms(&engine->execlists.preempt,
+		     active_preempt_timeout(engine, rq));
+}
+
+static bool completed(const struct i915_request *rq)
+{
+	if (i915_request_has_sentinel(rq))
+		return false;
+
+	return __i915_request_is_complete(rq);
+}
+
+static void execlists_dequeue(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	struct i915_sched_engine * const sched_engine = engine->sched_engine;
+	struct i915_request **port = execlists->pending;
+	struct i915_request ** const last_port = port + execlists->port_mask;
+	struct i915_request *last, * const *active;
+	struct virtual_engine *ve;
+	struct rb_node *rb;
+	bool submit = false;
+
+	/*
+	 * Hardware submission is through 2 ports. Conceptually each port
+	 * has a (RING_START, RING_HEAD, RING_TAIL) tuple. RING_START is
+	 * static for a context, and unique to each, so we only execute
+	 * requests belonging to a single context from each ring. RING_HEAD
+	 * is maintained by the CS in the context image, it marks the place
+	 * where it got up to last time, and through RING_TAIL we tell the CS
+	 * where we want to execute up to this time.
+	 *
+	 * In this list the requests are in order of execution. Consecutive
+	 * requests from the same context are adjacent in the ringbuffer. We
+	 * can combine these requests into a single RING_TAIL update:
+	 *
+	 *              RING_HEAD...req1...req2
+	 *                                    ^- RING_TAIL
+	 * since to execute req2 the CS must first execute req1.
+	 *
+	 * Our goal then is to point each port to the end of a consecutive
+	 * sequence of requests as being the most optimal (fewest wake ups
+	 * and context switches) submission.
+	 */
+
+	spin_lock(&sched_engine->lock);
+
+	/*
+	 * If the queue is higher priority than the last
+	 * request in the currently active context, submit afresh.
+	 * We will resubmit again afterwards in case we need to split
+	 * the active context to interject the preemption request,
+	 * i.e. we will retrigger preemption following the ack in case
+	 * of trouble.
+	 *
+	 */
+	active = execlists->active;
+	while ((last = *active) && completed(last))
+		active++;
+
+	if (last) {
+		if (need_preempt(engine, last)) {
+			ENGINE_TRACE(engine,
+				     "preempting last=%llx:%lld, prio=%d, hint=%d\n",
+				     last->fence.context,
+				     last->fence.seqno,
+				     last->sched.attr.priority,
+				     sched_engine->queue_priority_hint);
+			record_preemption(execlists);
+
+			/*
+			 * Don't let the RING_HEAD advance past the breadcrumb
+			 * as we unwind (and until we resubmit) so that we do
+			 * not accidentally tell it to go backwards.
+			 */
+			ring_set_paused(engine, 1);
+
+			/*
+			 * Note that we have not stopped the GPU at this point,
+			 * so we are unwinding the incomplete requests as they
+			 * remain inflight and so by the time we do complete
+			 * the preemption, some of the unwound requests may
+			 * complete!
+			 */
+			__unwind_incomplete_requests(engine);
+
+			last = NULL;
+		} else if (timeslice_expired(engine, last)) {
+			ENGINE_TRACE(engine,
+				     "expired:%s last=%llx:%lld, prio=%d, hint=%d, yield?=%s\n",
+				     str_yes_no(timer_expired(&execlists->timer)),
+				     last->fence.context, last->fence.seqno,
+				     rq_prio(last),
+				     sched_engine->queue_priority_hint,
+				     str_yes_no(timeslice_yield(execlists, last)));
+
+			/*
+			 * Consume this timeslice; ensure we start a new one.
+			 *
+			 * The timeslice expired, and we will unwind the
+			 * running contexts and recompute the next ELSP.
+			 * If that submit will be the same pair of contexts
+			 * (due to dependency ordering), we will skip the
+			 * submission. If we don't cancel the timer now,
+			 * we will see that the timer has expired and
+			 * reschedule the tasklet; continually until the
+			 * next context switch or other preemption event.
+			 *
+			 * Since we have decided to reschedule based on
+			 * consumption of this timeslice, if we submit the
+			 * same context again, grant it a full timeslice.
+			 */
+			cancel_timer(&execlists->timer);
+			ring_set_paused(engine, 1);
+			defer_active(engine);
+
+			/*
+			 * Unlike for preemption, if we rewind and continue
+			 * executing the same context as previously active,
+			 * the order of execution will remain the same and
+			 * the tail will only advance. We do not need to
+			 * force a full context restore, as a lite-restore
+			 * is sufficient to resample the monotonic TAIL.
+			 *
+			 * If we switch to any other context, similarly we
+			 * will not rewind TAIL of current context, and
+			 * normal save/restore will preserve state and allow
+			 * us to later continue executing the same request.
+			 */
+			last = NULL;
+		} else {
+			/*
+			 * Otherwise if we already have a request pending
+			 * for execution after the current one, we can
+			 * just wait until the next CS event before
+			 * queuing more. In either case we will force a
+			 * lite-restore preemption event, but if we wait
+			 * we hopefully coalesce several updates into a single
+			 * submission.
+			 */
+			if (active[1]) {
+				/*
+				 * Even if ELSP[1] is occupied and not worthy
+				 * of timeslices, our queue might be.
+				 */
+				spin_unlock(&sched_engine->lock);
+				return;
+			}
+		}
+	}
+
+	/* XXX virtual is always taking precedence */
+	while ((ve = first_virtual_engine(engine))) {
+		struct i915_request *rq;
+
+		spin_lock(&ve->base.sched_engine->lock);
+
+		rq = ve->request;
+		if (unlikely(!virtual_matches(ve, rq, engine)))
+			goto unlock; /* lost the race to a sibling */
+
+		GEM_BUG_ON(rq->engine != &ve->base);
+		GEM_BUG_ON(rq->context != &ve->context);
+
+		if (unlikely(rq_prio(rq) < queue_prio(sched_engine))) {
+			spin_unlock(&ve->base.sched_engine->lock);
+			break;
+		}
+
+		if (last && !can_merge_rq(last, rq)) {
+			spin_unlock(&ve->base.sched_engine->lock);
+			spin_unlock(&engine->sched_engine->lock);
+			return; /* leave this for another sibling */
+		}
+
+		ENGINE_TRACE(engine,
+			     "virtual rq=%llx:%lld%s, new engine? %s\n",
+			     rq->fence.context,
+			     rq->fence.seqno,
+			     __i915_request_is_complete(rq) ? "!" :
+			     __i915_request_has_started(rq) ? "*" :
+			     "",
+			     str_yes_no(engine != ve->siblings[0]));
+
+		WRITE_ONCE(ve->request, NULL);
+		WRITE_ONCE(ve->base.sched_engine->queue_priority_hint, INT_MIN);
+
+		rb = &ve->nodes[engine->id].rb;
+		rb_erase_cached(rb, &execlists->virtual);
+		RB_CLEAR_NODE(rb);
+
+		GEM_BUG_ON(!(rq->execution_mask & engine->mask));
+		WRITE_ONCE(rq->engine, engine);
+
+		if (__i915_request_submit(rq)) {
+			/*
+			 * Only after we confirm that we will submit
+			 * this request (i.e. it has not already
+			 * completed), do we want to update the context.
+			 *
+			 * This serves two purposes. It avoids
+			 * unnecessary work if we are resubmitting an
+			 * already completed request after timeslicing.
+			 * But more importantly, it prevents us altering
+			 * ve->siblings[] on an idle context, where
+			 * we may be using ve->siblings[] in
+			 * virtual_context_enter / virtual_context_exit.
+			 */
+			virtual_xfer_context(ve, engine);
+			GEM_BUG_ON(ve->siblings[0] != engine);
+
+			submit = true;
+			last = rq;
+		}
+
+		i915_request_put(rq);
+unlock:
+		spin_unlock(&ve->base.sched_engine->lock);
+
+		/*
+		 * Hmm, we have a bunch of virtual engine requests,
+		 * but the first one was already completed (thanks
+		 * preempt-to-busy!). Keep looking at the veng queue
+		 * until we have no more relevant requests (i.e.
+		 * the normal submit queue has higher priority).
+		 */
+		if (submit)
+			break;
+	}
+
+	while ((rb = rb_first_cached(&sched_engine->queue))) {
+		struct i915_priolist *p = to_priolist(rb);
+		struct i915_request *rq, *rn;
+
+		priolist_for_each_request_consume(rq, rn, p) {
+			bool merge = true;
+
+			/*
+			 * Can we combine this request with the current port?
+			 * It has to be the same context/ringbuffer and not
+			 * have any exceptions (e.g. GVT saying never to
+			 * combine contexts).
+			 *
+			 * If we can combine the requests, we can execute both
+			 * by updating the RING_TAIL to point to the end of the
+			 * second request, and so we never need to tell the
+			 * hardware about the first.
+			 */
+			if (last && !can_merge_rq(last, rq)) {
+				/*
+				 * If we are on the second port and cannot
+				 * combine this request with the last, then we
+				 * are done.
+				 */
+				if (port == last_port)
+					goto done;
+
+				/*
+				 * We must not populate both ELSP[] with the
+				 * same LRCA, i.e. we must submit 2 different
+				 * contexts if we submit 2 ELSP.
+				 */
+				if (last->context == rq->context)
+					goto done;
+
+				if (i915_request_has_sentinel(last))
+					goto done;
+
+				/*
+				 * We avoid submitting virtual requests into
+				 * the secondary ports so that we can migrate
+				 * the request immediately to another engine
+				 * rather than wait for the primary request.
+				 */
+				if (rq->execution_mask != engine->mask)
+					goto done;
+
+				/*
+				 * If GVT overrides us we only ever submit
+				 * port[0], leaving port[1] empty. Note that we
+				 * also have to be careful that we don't queue
+				 * the same context (even though a different
+				 * request) to the second port.
+				 */
+				if (ctx_single_port_submission(last->context) ||
+				    ctx_single_port_submission(rq->context))
+					goto done;
+
+				merge = false;
+			}
+
+			if (__i915_request_submit(rq)) {
+				if (!merge) {
+					*port++ = i915_request_get(last);
+					last = NULL;
+				}
+
+				GEM_BUG_ON(last &&
+					   !can_merge_ctx(last->context,
+							  rq->context));
+				GEM_BUG_ON(last &&
+					   i915_seqno_passed(last->fence.seqno,
+							     rq->fence.seqno));
+
+				submit = true;
+				last = rq;
+			}
+		}
+
+		rb_erase_cached(&p->node, &sched_engine->queue);
+		i915_priolist_free(p);
+	}
+done:
+	*port++ = i915_request_get(last);
+
+	/*
+	 * Here be a bit of magic! Or sleight-of-hand, whichever you prefer.
+	 *
+	 * We choose the priority hint such that if we add a request of greater
+	 * priority than this, we kick the submission tasklet to decide on
+	 * the right order of submitting the requests to hardware. We must
+	 * also be prepared to reorder requests as they are in-flight on the
+	 * HW. We derive the priority hint then as the first "hole" in
+	 * the HW submission ports and if there are no available slots,
+	 * the priority of the lowest executing request, i.e. last.
+	 *
+	 * When we do receive a higher priority request ready to run from the
+	 * user, see queue_request(), the priority hint is bumped to that
+	 * request triggering preemption on the next dequeue (or subsequent
+	 * interrupt for secondary ports).
+	 */
+	sched_engine->queue_priority_hint = queue_prio(sched_engine);
+	i915_sched_engine_reset_on_empty(sched_engine);
+	spin_unlock(&sched_engine->lock);
+
+	/*
+	 * We can skip poking the HW if we ended up with exactly the same set
+	 * of requests as currently running, e.g. trying to timeslice a pair
+	 * of ordered contexts.
+	 */
+	if (submit &&
+	    memcmp(active,
+		   execlists->pending,
+		   (port - execlists->pending) * sizeof(*port))) {
+		*port = NULL;
+		while (port-- != execlists->pending)
+			execlists_schedule_in(*port, port - execlists->pending);
+
+		WRITE_ONCE(execlists->yield, -1);
+		set_preempt_timeout(engine, *active);
+		execlists_submit_ports(engine);
+	} else {
+		ring_set_paused(engine, 0);
+		while (port-- != execlists->pending)
+			i915_request_put(*port);
+		*execlists->pending = NULL;
+	}
+}
+
+static void execlists_dequeue_irq(struct intel_engine_cs *engine)
+{
+	local_irq_disable(); /* Suspend interrupts across request submission */
+	execlists_dequeue(engine);
+	local_irq_enable(); /* flush irq_work (e.g. breadcrumb enabling) */
+}
+
+static void clear_ports(struct i915_request **ports, int count)
+{
+	memset_p((void **)ports, NULL, count);
+}
+
+static void
+copy_ports(struct i915_request **dst, struct i915_request **src, int count)
+{
+	/* A memcpy_p() would be very useful here! */
+	while (count--)
+		WRITE_ONCE(*dst++, *src++); /* avoid write tearing */
+}
+
+static struct i915_request **
+cancel_port_requests(struct intel_engine_execlists * const execlists,
+		     struct i915_request **inactive)
+{
+	struct i915_request * const *port;
+
+	for (port = execlists->pending; *port; port++)
+		*inactive++ = *port;
+	clear_ports(execlists->pending, ARRAY_SIZE(execlists->pending));
+
+	/* Mark the end of active before we overwrite *active */
+	for (port = xchg(&execlists->active, execlists->pending); *port; port++)
+		*inactive++ = *port;
+	clear_ports(execlists->inflight, ARRAY_SIZE(execlists->inflight));
+
+	smp_wmb(); /* complete the seqlock for execlists_active() */
+	WRITE_ONCE(execlists->active, execlists->inflight);
+
+	/* Having cancelled all outstanding process_csb(), stop their timers */
+	GEM_BUG_ON(execlists->pending[0]);
+	cancel_timer(&execlists->timer);
+	cancel_timer(&execlists->preempt);
+
+	return inactive;
+}
+
+/*
+ * Starting with Gen12, the status has a new format:
+ *
+ *     bit  0:     switched to new queue
+ *     bit  1:     reserved
+ *     bit  2:     semaphore wait mode (poll or signal), only valid when
+ *                 switch detail is set to "wait on semaphore"
+ *     bits 3-5:   engine class
+ *     bits 6-11:  engine instance
+ *     bits 12-14: reserved
+ *     bits 15-25: sw context id of the lrc the GT switched to
+ *     bits 26-31: sw counter of the lrc the GT switched to
+ *     bits 32-35: context switch detail
+ *                  - 0: ctx complete
+ *                  - 1: wait on sync flip
+ *                  - 2: wait on vblank
+ *                  - 3: wait on scanline
+ *                  - 4: wait on semaphore
+ *                  - 5: context preempted (not on SEMAPHORE_WAIT or
+ *                       WAIT_FOR_EVENT)
+ *     bit  36:    reserved
+ *     bits 37-43: wait detail (for switch detail 1 to 4)
+ *     bits 44-46: reserved
+ *     bits 47-57: sw context id of the lrc the GT switched away from
+ *     bits 58-63: sw counter of the lrc the GT switched away from
+ *
+ * Xe_HP csb shuffles things around compared to TGL:
+ *
+ *     bits 0-3:   context switch detail (same possible values as TGL)
+ *     bits 4-9:   engine instance
+ *     bits 10-25: sw context id of the lrc the GT switched to
+ *     bits 26-31: sw counter of the lrc the GT switched to
+ *     bit  32:    semaphore wait mode (poll or signal), Only valid when
+ *                 switch detail is set to "wait on semaphore"
+ *     bit  33:    switched to new queue
+ *     bits 34-41: wait detail (for switch detail 1 to 4)
+ *     bits 42-57: sw context id of the lrc the GT switched away from
+ *     bits 58-63: sw counter of the lrc the GT switched away from
+ */
+static inline bool
+__gen12_csb_parse(bool ctx_to_valid, bool ctx_away_valid, bool new_queue,
+		  u8 switch_detail)
+{
+	/*
+	 * The context switch detail is not guaranteed to be 5 when a preemption
+	 * occurs, so we can't just check for that. The check below works for
+	 * all the cases we care about, including preemptions of WAIT
+	 * instructions and lite-restore. Preempt-to-idle via the CTRL register
+	 * would require some extra handling, but we don't support that.
+	 */
+	if (!ctx_away_valid || new_queue) {
+		GEM_BUG_ON(!ctx_to_valid);
+		return true;
+	}
+
+	/*
+	 * switch detail = 5 is covered by the case above and we do not expect a
+	 * context switch on an unsuccessful wait instruction since we always
+	 * use polling mode.
+	 */
+	GEM_BUG_ON(switch_detail);
+	return false;
+}
+
+static bool xehp_csb_parse(const u64 csb)
+{
+	return __gen12_csb_parse(XEHP_CSB_CTX_VALID(lower_32_bits(csb)), /* cxt to */
+				 XEHP_CSB_CTX_VALID(upper_32_bits(csb)), /* cxt away */
+				 upper_32_bits(csb) & XEHP_CTX_STATUS_SWITCHED_TO_NEW_QUEUE,
+				 GEN12_CTX_SWITCH_DETAIL(lower_32_bits(csb)));
+}
+
+static bool gen12_csb_parse(const u64 csb)
+{
+	return __gen12_csb_parse(GEN12_CSB_CTX_VALID(lower_32_bits(csb)), /* cxt to */
+				 GEN12_CSB_CTX_VALID(upper_32_bits(csb)), /* cxt away */
+				 lower_32_bits(csb) & GEN12_CTX_STATUS_SWITCHED_TO_NEW_QUEUE,
+				 GEN12_CTX_SWITCH_DETAIL(upper_32_bits(csb)));
+}
+
+static bool gen8_csb_parse(const u64 csb)
+{
+	return csb & (GEN8_CTX_STATUS_IDLE_ACTIVE | GEN8_CTX_STATUS_PREEMPTED);
+}
+
+static noinline u64
+wa_csb_read(const struct intel_engine_cs *engine, u64 * const csb)
+{
+	u64 entry;
+
+	/*
+	 * Reading from the HWSP has one particular advantage: we can detect
+	 * a stale entry. Since the write into HWSP is broken, we have no reason
+	 * to trust the HW at all, the mmio entry may equally be unordered, so
+	 * we prefer the path that is self-checking and as a last resort,
+	 * return the mmio value.
+	 *
+	 * tgl,dg1:HSDES#22011327657
+	 */
+	preempt_disable();
+	if (wait_for_atomic_us((entry = READ_ONCE(*csb)) != -1, 10)) {
+		int idx = csb - engine->execlists.csb_status;
+		int status;
+
+		status = GEN8_EXECLISTS_STATUS_BUF;
+		if (idx >= 6) {
+			status = GEN11_EXECLISTS_STATUS_BUF2;
+			idx -= 6;
+		}
+		status += sizeof(u64) * idx;
+
+		entry = intel_uncore_read64(engine->uncore,
+					    _MMIO(engine->mmio_base + status));
+	}
+	preempt_enable();
+
+	return entry;
+}
+
+static u64 csb_read(const struct intel_engine_cs *engine, u64 * const csb)
+{
+	u64 entry = READ_ONCE(*csb);
+
+	/*
+	 * Unfortunately, the GPU does not always serialise its write
+	 * of the CSB entries before its write of the CSB pointer, at least
+	 * from the perspective of the CPU, using what is known as a Global
+	 * Observation Point. We may read a new CSB tail pointer, but then
+	 * read the stale CSB entries, causing us to misinterpret the
+	 * context-switch events, and eventually declare the GPU hung.
+	 *
+	 * icl:HSDES#1806554093
+	 * tgl:HSDES#22011248461
+	 */
+	if (unlikely(entry == -1))
+		entry = wa_csb_read(engine, csb);
+
+	/* Consume this entry so that we can spot its future reuse. */
+	WRITE_ONCE(*csb, -1);
+
+	/* ELSP is an implicit wmb() before the GPU wraps and overwrites csb */
+	return entry;
+}
+
+static void new_timeslice(struct intel_engine_execlists *el)
+{
+	/* By cancelling, we will start afresh in start_timeslice() */
+	cancel_timer(&el->timer);
+}
+
+static struct i915_request **
+process_csb(struct intel_engine_cs *engine, struct i915_request **inactive)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	u64 * const buf = execlists->csb_status;
+	const u8 num_entries = execlists->csb_size;
+	struct i915_request **prev;
+	u8 head, tail;
+
+	/*
+	 * As we modify our execlists state tracking we require exclusive
+	 * access. Either we are inside the tasklet, or the tasklet is disabled
+	 * and we assume that is only inside the reset paths and so serialised.
+	 */
+	GEM_BUG_ON(!tasklet_is_locked(&engine->sched_engine->tasklet) &&
+		   !reset_in_progress(engine));
+
+	/*
+	 * Note that csb_write, csb_status may be either in HWSP or mmio.
+	 * When reading from the csb_write mmio register, we have to be
+	 * careful to only use the GEN8_CSB_WRITE_PTR portion, which is
+	 * the low 4bits. As it happens we know the next 4bits are always
+	 * zero and so we can simply masked off the low u8 of the register
+	 * and treat it identically to reading from the HWSP (without having
+	 * to use explicit shifting and masking, and probably bifurcating
+	 * the code to handle the legacy mmio read).
+	 */
+	head = execlists->csb_head;
+	tail = READ_ONCE(*execlists->csb_write);
+	if (unlikely(head == tail))
+		return inactive;
+
+	/*
+	 * We will consume all events from HW, or at least pretend to.
+	 *
+	 * The sequence of events from the HW is deterministic, and derived
+	 * from our writes to the ELSP, with a smidgen of variability for
+	 * the arrival of the asynchronous requests wrt to the inflight
+	 * execution. If the HW sends an event that does not correspond with
+	 * the one we are expecting, we have to abandon all hope as we lose
+	 * all tracking of what the engine is actually executing. We will
+	 * only detect we are out of sequence with the HW when we get an
+	 * 'impossible' event because we have already drained our own
+	 * preemption/promotion queue. If this occurs, we know that we likely
+	 * lost track of execution earlier and must unwind and restart, the
+	 * simplest way is by stop processing the event queue and force the
+	 * engine to reset.
+	 */
+	execlists->csb_head = tail;
+	ENGINE_TRACE(engine, "cs-irq head=%d, tail=%d\n", head, tail);
+
+	/*
+	 * Hopefully paired with a wmb() in HW!
+	 *
+	 * We must complete the read of the write pointer before any reads
+	 * from the CSB, so that we do not see stale values. Without an rmb
+	 * (lfence) the HW may speculatively perform the CSB[] reads *before*
+	 * we perform the READ_ONCE(*csb_write).
+	 */
+	rmb();
+
+	/* Remember who was last running under the timer */
+	prev = inactive;
+	*prev = NULL;
+
+	do {
+		bool promote;
+		u64 csb;
+
+		if (++head == num_entries)
+			head = 0;
+
+		/*
+		 * We are flying near dragons again.
+		 *
+		 * We hold a reference to the request in execlist_port[]
+		 * but no more than that. We are operating in softirq
+		 * context and so cannot hold any mutex or sleep. That
+		 * prevents us stopping the requests we are processing
+		 * in port[] from being retired simultaneously (the
+		 * breadcrumb will be complete before we see the
+		 * context-switch). As we only hold the reference to the
+		 * request, any pointer chasing underneath the request
+		 * is subject to a potential use-after-free. Thus we
+		 * store all of the bookkeeping within port[] as
+		 * required, and avoid using unguarded pointers beneath
+		 * request itself. The same applies to the atomic
+		 * status notifier.
+		 */
+
+		csb = csb_read(engine, buf + head);
+		ENGINE_TRACE(engine, "csb[%d]: status=0x%08x:0x%08x\n",
+			     head, upper_32_bits(csb), lower_32_bits(csb));
+
+		if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 50))
+			promote = xehp_csb_parse(csb);
+		else if (GRAPHICS_VER(engine->i915) >= 12)
+			promote = gen12_csb_parse(csb);
+		else
+			promote = gen8_csb_parse(csb);
+		if (promote) {
+			struct i915_request * const *old = execlists->active;
+
+			if (GEM_WARN_ON(!*execlists->pending)) {
+				execlists->error_interrupt |= ERROR_CSB;
+				break;
+			}
+
+			ring_set_paused(engine, 0);
+
+			/* Point active to the new ELSP; prevent overwriting */
+			WRITE_ONCE(execlists->active, execlists->pending);
+			smp_wmb(); /* notify execlists_active() */
+
+			/* cancel old inflight, prepare for switch */
+			trace_ports(execlists, "preempted", old);
+			while (*old)
+				*inactive++ = *old++;
+
+			/* switch pending to inflight */
+			GEM_BUG_ON(!assert_pending_valid(execlists, "promote"));
+			copy_ports(execlists->inflight,
+				   execlists->pending,
+				   execlists_num_ports(execlists));
+			smp_wmb(); /* complete the seqlock */
+			WRITE_ONCE(execlists->active, execlists->inflight);
+
+			/* XXX Magic delay for tgl */
+			ENGINE_POSTING_READ(engine, RING_CONTEXT_STATUS_PTR);
+
+			WRITE_ONCE(execlists->pending[0], NULL);
+		} else {
+			if (GEM_WARN_ON(!*execlists->active)) {
+				execlists->error_interrupt |= ERROR_CSB;
+				break;
+			}
+
+			/* port0 completed, advanced to port1 */
+			trace_ports(execlists, "completed", execlists->active);
+
+			/*
+			 * We rely on the hardware being strongly
+			 * ordered, that the breadcrumb write is
+			 * coherent (visible from the CPU) before the
+			 * user interrupt is processed. One might assume
+			 * that the breadcrumb write being before the
+			 * user interrupt and the CS event for the context
+			 * switch would therefore be before the CS event
+			 * itself...
+			 */
+			if (GEM_SHOW_DEBUG() &&
+			    !__i915_request_is_complete(*execlists->active)) {
+				struct i915_request *rq = *execlists->active;
+				const u32 *regs __maybe_unused =
+					rq->context->lrc_reg_state;
+
+				ENGINE_TRACE(engine,
+					     "context completed before request!\n");
+				ENGINE_TRACE(engine,
+					     "ring:{start:0x%08x, head:%04x, tail:%04x, ctl:%08x, mode:%08x}\n",
+					     ENGINE_READ(engine, RING_START),
+					     ENGINE_READ(engine, RING_HEAD) & HEAD_ADDR,
+					     ENGINE_READ(engine, RING_TAIL) & TAIL_ADDR,
+					     ENGINE_READ(engine, RING_CTL),
+					     ENGINE_READ(engine, RING_MI_MODE));
+				ENGINE_TRACE(engine,
+					     "rq:{start:%08x, head:%04x, tail:%04x, seqno:%llx:%d, hwsp:%d}, ",
+					     i915_ggtt_offset(rq->ring->vma),
+					     rq->head, rq->tail,
+					     rq->fence.context,
+					     lower_32_bits(rq->fence.seqno),
+					     hwsp_seqno(rq));
+				ENGINE_TRACE(engine,
+					     "ctx:{start:%08x, head:%04x, tail:%04x}, ",
+					     regs[CTX_RING_START],
+					     regs[CTX_RING_HEAD],
+					     regs[CTX_RING_TAIL]);
+			}
+
+			*inactive++ = *execlists->active++;
+
+			GEM_BUG_ON(execlists->active - execlists->inflight >
+				   execlists_num_ports(execlists));
+		}
+	} while (head != tail);
+
+	/*
+	 * Gen11 has proven to fail wrt global observation point between
+	 * entry and tail update, failing on the ordering and thus
+	 * we see an old entry in the context status buffer.
+	 *
+	 * Forcibly evict out entries for the next gpu csb update,
+	 * to increase the odds that we get a fresh entries with non
+	 * working hardware. The cost for doing so comes out mostly with
+	 * the wash as hardware, working or not, will need to do the
+	 * invalidation before.
+	 */
+	drm_clflush_virt_range(&buf[0], num_entries * sizeof(buf[0]));
+
+	/*
+	 * We assume that any event reflects a change in context flow
+	 * and merits a fresh timeslice. We reinstall the timer after
+	 * inspecting the queue to see if we need to resumbit.
+	 */
+	if (*prev != *execlists->active) { /* elide lite-restores */
+		struct intel_context *prev_ce = NULL, *active_ce = NULL;
+
+		/*
+		 * Note the inherent discrepancy between the HW runtime,
+		 * recorded as part of the context switch, and the CPU
+		 * adjustment for active contexts. We have to hope that
+		 * the delay in processing the CS event is very small
+		 * and consistent. It works to our advantage to have
+		 * the CPU adjustment _undershoot_ (i.e. start later than)
+		 * the CS timestamp so we never overreport the runtime
+		 * and correct overselves later when updating from HW.
+		 */
+		if (*prev)
+			prev_ce = (*prev)->context;
+		if (*execlists->active)
+			active_ce = (*execlists->active)->context;
+		if (prev_ce != active_ce) {
+			if (prev_ce)
+				lrc_runtime_stop(prev_ce);
+			if (active_ce)
+				lrc_runtime_start(active_ce);
+		}
+		new_timeslice(execlists);
+	}
+
+	return inactive;
+}
+
+static void post_process_csb(struct i915_request **port,
+			     struct i915_request **last)
+{
+	while (port != last)
+		execlists_schedule_out(*port++);
+}
+
+static void __execlists_hold(struct i915_request *rq)
+{
+	LIST_HEAD(list);
+
+	do {
+		struct i915_dependency *p;
+
+		if (i915_request_is_active(rq))
+			__i915_request_unsubmit(rq);
+
+		clear_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+		list_move_tail(&rq->sched.link,
+			       &rq->engine->sched_engine->hold);
+		i915_request_set_hold(rq);
+		RQ_TRACE(rq, "on hold\n");
+
+		for_each_waiter(p, rq) {
+			struct i915_request *w =
+				container_of(p->waiter, typeof(*w), sched);
+
+			if (p->flags & I915_DEPENDENCY_WEAK)
+				continue;
+
+			/* Leave semaphores spinning on the other engines */
+			if (w->engine != rq->engine)
+				continue;
+
+			if (!i915_request_is_ready(w))
+				continue;
+
+			if (__i915_request_is_complete(w))
+				continue;
+
+			if (i915_request_on_hold(w))
+				continue;
+
+			list_move_tail(&w->sched.link, &list);
+		}
+
+		rq = list_first_entry_or_null(&list, typeof(*rq), sched.link);
+	} while (rq);
+}
+
+static bool execlists_hold(struct intel_engine_cs *engine,
+			   struct i915_request *rq)
+{
+	if (i915_request_on_hold(rq))
+		return false;
+
+	spin_lock_irq(&engine->sched_engine->lock);
+
+	if (__i915_request_is_complete(rq)) { /* too late! */
+		rq = NULL;
+		goto unlock;
+	}
+
+	/*
+	 * Transfer this request onto the hold queue to prevent it
+	 * being resumbitted to HW (and potentially completed) before we have
+	 * released it. Since we may have already submitted following
+	 * requests, we need to remove those as well.
+	 */
+	GEM_BUG_ON(i915_request_on_hold(rq));
+	GEM_BUG_ON(rq->engine != engine);
+	__execlists_hold(rq);
+	GEM_BUG_ON(list_empty(&engine->sched_engine->hold));
+
+unlock:
+	spin_unlock_irq(&engine->sched_engine->lock);
+	return rq;
+}
+
+static bool hold_request(const struct i915_request *rq)
+{
+	struct i915_dependency *p;
+	bool result = false;
+
+	/*
+	 * If one of our ancestors is on hold, we must also be on hold,
+	 * otherwise we will bypass it and execute before it.
+	 */
+	rcu_read_lock();
+	for_each_signaler(p, rq) {
+		const struct i915_request *s =
+			container_of(p->signaler, typeof(*s), sched);
+
+		if (s->engine != rq->engine)
+			continue;
+
+		result = i915_request_on_hold(s);
+		if (result)
+			break;
+	}
+	rcu_read_unlock();
+
+	return result;
+}
+
+static void __execlists_unhold(struct i915_request *rq)
+{
+	LIST_HEAD(list);
+
+	do {
+		struct i915_dependency *p;
+
+		RQ_TRACE(rq, "hold release\n");
+
+		GEM_BUG_ON(!i915_request_on_hold(rq));
+		GEM_BUG_ON(!i915_sw_fence_signaled(&rq->submit));
+
+		i915_request_clear_hold(rq);
+		list_move_tail(&rq->sched.link,
+			       i915_sched_lookup_priolist(rq->engine->sched_engine,
+							  rq_prio(rq)));
+		set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+
+		/* Also release any children on this engine that are ready */
+		for_each_waiter(p, rq) {
+			struct i915_request *w =
+				container_of(p->waiter, typeof(*w), sched);
+
+			if (p->flags & I915_DEPENDENCY_WEAK)
+				continue;
+
+			if (w->engine != rq->engine)
+				continue;
+
+			if (!i915_request_on_hold(w))
+				continue;
+
+			/* Check that no other parents are also on hold */
+			if (hold_request(w))
+				continue;
+
+			list_move_tail(&w->sched.link, &list);
+		}
+
+		rq = list_first_entry_or_null(&list, typeof(*rq), sched.link);
+	} while (rq);
+}
+
+static void execlists_unhold(struct intel_engine_cs *engine,
+			     struct i915_request *rq)
+{
+	spin_lock_irq(&engine->sched_engine->lock);
+
+	/*
+	 * Move this request back to the priority queue, and all of its
+	 * children and grandchildren that were suspended along with it.
+	 */
+	__execlists_unhold(rq);
+
+	if (rq_prio(rq) > engine->sched_engine->queue_priority_hint) {
+		engine->sched_engine->queue_priority_hint = rq_prio(rq);
+		tasklet_hi_schedule(&engine->sched_engine->tasklet);
+	}
+
+	spin_unlock_irq(&engine->sched_engine->lock);
+}
+
+struct execlists_capture {
+	struct work_struct work;
+	struct i915_request *rq;
+	struct i915_gpu_coredump *error;
+};
+
+static void execlists_capture_work(struct work_struct *work)
+{
+	struct execlists_capture *cap = container_of(work, typeof(*cap), work);
+	const gfp_t gfp = __GFP_KSWAPD_RECLAIM | __GFP_RETRY_MAYFAIL |
+		__GFP_NOWARN;
+	struct intel_engine_cs *engine = cap->rq->engine;
+	struct intel_gt_coredump *gt = cap->error->gt;
+	struct intel_engine_capture_vma *vma;
+
+	/* Compress all the objects attached to the request, slow! */
+	vma = intel_engine_coredump_add_request(gt->engine, cap->rq, gfp);
+	if (vma) {
+		struct i915_vma_compress *compress =
+			i915_vma_capture_prepare(gt);
+
+		intel_engine_coredump_add_vma(gt->engine, vma, compress);
+		i915_vma_capture_finish(gt, compress);
+	}
+
+	gt->simulated = gt->engine->simulated;
+	cap->error->simulated = gt->simulated;
+
+	/* Publish the error state, and announce it to the world */
+	i915_error_state_store(cap->error);
+	i915_gpu_coredump_put(cap->error);
+
+	/* Return this request and all that depend upon it for signaling */
+	execlists_unhold(engine, cap->rq);
+	i915_request_put(cap->rq);
+
+	kfree(cap);
+}
+
+static struct execlists_capture *capture_regs(struct intel_engine_cs *engine)
+{
+	const gfp_t gfp = GFP_ATOMIC | __GFP_NOWARN;
+	struct execlists_capture *cap;
+
+	cap = kmalloc(sizeof(*cap), gfp);
+	if (!cap)
+		return NULL;
+
+	cap->error = i915_gpu_coredump_alloc(engine->i915, gfp);
+	if (!cap->error)
+		goto err_cap;
+
+	cap->error->gt = intel_gt_coredump_alloc(engine->gt, gfp, CORE_DUMP_FLAG_NONE);
+	if (!cap->error->gt)
+		goto err_gpu;
+
+	cap->error->gt->engine = intel_engine_coredump_alloc(engine, gfp, CORE_DUMP_FLAG_NONE);
+	if (!cap->error->gt->engine)
+		goto err_gt;
+
+	cap->error->gt->engine->hung = true;
+
+	return cap;
+
+err_gt:
+	kfree(cap->error->gt);
+err_gpu:
+	kfree(cap->error);
+err_cap:
+	kfree(cap);
+	return NULL;
+}
+
+static struct i915_request *
+active_context(struct intel_engine_cs *engine, u32 ccid)
+{
+	const struct intel_engine_execlists * const el = &engine->execlists;
+	struct i915_request * const *port, *rq;
+
+	/*
+	 * Use the most recent result from process_csb(), but just in case
+	 * we trigger an error (via interrupt) before the first CS event has
+	 * been written, peek at the next submission.
+	 */
+
+	for (port = el->active; (rq = *port); port++) {
+		if (rq->context->lrc.ccid == ccid) {
+			ENGINE_TRACE(engine,
+				     "ccid:%x found at active:%zd\n",
+				     ccid, port - el->active);
+			return rq;
+		}
+	}
+
+	for (port = el->pending; (rq = *port); port++) {
+		if (rq->context->lrc.ccid == ccid) {
+			ENGINE_TRACE(engine,
+				     "ccid:%x found at pending:%zd\n",
+				     ccid, port - el->pending);
+			return rq;
+		}
+	}
+
+	ENGINE_TRACE(engine, "ccid:%x not found\n", ccid);
+	return NULL;
+}
+
+static u32 active_ccid(struct intel_engine_cs *engine)
+{
+	return ENGINE_READ_FW(engine, RING_EXECLIST_STATUS_HI);
+}
+
+static void execlists_capture(struct intel_engine_cs *engine)
+{
+	struct execlists_capture *cap;
+
+	if (!IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR))
+		return;
+
+	/*
+	 * We need to _quickly_ capture the engine state before we reset.
+	 * We are inside an atomic section (softirq) here and we are delaying
+	 * the forced preemption event.
+	 */
+	cap = capture_regs(engine);
+	if (!cap)
+		return;
+
+	spin_lock_irq(&engine->sched_engine->lock);
+	cap->rq = active_context(engine, active_ccid(engine));
+	if (cap->rq) {
+		cap->rq = active_request(cap->rq->context->timeline, cap->rq);
+		cap->rq = i915_request_get_rcu(cap->rq);
+	}
+	spin_unlock_irq(&engine->sched_engine->lock);
+	if (!cap->rq)
+		goto err_free;
+
+	/*
+	 * Remove the request from the execlists queue, and take ownership
+	 * of the request. We pass it to our worker who will _slowly_ compress
+	 * all the pages the _user_ requested for debugging their batch, after
+	 * which we return it to the queue for signaling.
+	 *
+	 * By removing them from the execlists queue, we also remove the
+	 * requests from being processed by __unwind_incomplete_requests()
+	 * during the intel_engine_reset(), and so they will *not* be replayed
+	 * afterwards.
+	 *
+	 * Note that because we have not yet reset the engine at this point,
+	 * it is possible for the request that we have identified as being
+	 * guilty, did in fact complete and we will then hit an arbitration
+	 * point allowing the outstanding preemption to succeed. The likelihood
+	 * of that is very low (as capturing of the engine registers should be
+	 * fast enough to run inside an irq-off atomic section!), so we will
+	 * simply hold that request accountable for being non-preemptible
+	 * long enough to force the reset.
+	 */
+	if (!execlists_hold(engine, cap->rq))
+		goto err_rq;
+
+	INIT_WORK(&cap->work, execlists_capture_work);
+	schedule_work(&cap->work);
+	return;
+
+err_rq:
+	i915_request_put(cap->rq);
+err_free:
+	i915_gpu_coredump_put(cap->error);
+	kfree(cap);
+}
+
+static void execlists_reset(struct intel_engine_cs *engine, const char *msg)
+{
+	const unsigned int bit = I915_RESET_ENGINE + engine->id;
+	unsigned long *lock = &engine->gt->reset.flags;
+
+	if (!intel_has_reset_engine(engine->gt))
+		return;
+
+	if (test_and_set_bit(bit, lock))
+		return;
+
+	ENGINE_TRACE(engine, "reset for %s\n", msg);
+
+	/* Mark this tasklet as disabled to avoid waiting for it to complete */
+	tasklet_disable_nosync(&engine->sched_engine->tasklet);
+
+	ring_set_paused(engine, 1); /* Freeze the current request in place */
+	execlists_capture(engine);
+	intel_engine_reset(engine, msg);
+
+	tasklet_enable(&engine->sched_engine->tasklet);
+	clear_and_wake_up_bit(bit, lock);
+}
+
+static bool preempt_timeout(const struct intel_engine_cs *const engine)
+{
+	const struct timer_list *t = &engine->execlists.preempt;
+
+	if (!CONFIG_DRM_I915_PREEMPT_TIMEOUT)
+		return false;
+
+	if (!timer_expired(t))
+		return false;
+
+	return engine->execlists.pending[0];
+}
+
+/*
+ * Check the unread Context Status Buffers and manage the submission of new
+ * contexts to the ELSP accordingly.
+ */
+static void execlists_submission_tasklet(struct tasklet_struct *t)
+{
+	struct i915_sched_engine *sched_engine =
+		from_tasklet(sched_engine, t, tasklet);
+	struct intel_engine_cs * const engine = sched_engine->private_data;
+	struct i915_request *post[2 * EXECLIST_MAX_PORTS];
+	struct i915_request **inactive;
+
+	rcu_read_lock();
+	inactive = process_csb(engine, post);
+	GEM_BUG_ON(inactive - post > ARRAY_SIZE(post));
+
+	if (unlikely(preempt_timeout(engine))) {
+		const struct i915_request *rq = *engine->execlists.active;
+
+		/*
+		 * If after the preempt-timeout expired, we are still on the
+		 * same active request/context as before we initiated the
+		 * preemption, reset the engine.
+		 *
+		 * However, if we have processed a CS event to switch contexts,
+		 * but not yet processed the CS event for the pending
+		 * preemption, reset the timer allowing the new context to
+		 * gracefully exit.
+		 */
+		cancel_timer(&engine->execlists.preempt);
+		if (rq == engine->execlists.preempt_target)
+			engine->execlists.error_interrupt |= ERROR_PREEMPT;
+		else
+			set_timer_ms(&engine->execlists.preempt,
+				     active_preempt_timeout(engine, rq));
+	}
+
+	if (unlikely(READ_ONCE(engine->execlists.error_interrupt))) {
+		const char *msg;
+
+		/* Generate the error message in priority wrt to the user! */
+		if (engine->execlists.error_interrupt & GENMASK(15, 0))
+			msg = "CS error"; /* thrown by a user payload */
+		else if (engine->execlists.error_interrupt & ERROR_CSB)
+			msg = "invalid CSB event";
+		else if (engine->execlists.error_interrupt & ERROR_PREEMPT)
+			msg = "preemption time out";
+		else
+			msg = "internal error";
+
+		engine->execlists.error_interrupt = 0;
+		execlists_reset(engine, msg);
+	}
+
+	if (!engine->execlists.pending[0]) {
+		execlists_dequeue_irq(engine);
+		start_timeslice(engine);
+	}
+
+	post_process_csb(post, inactive);
+	rcu_read_unlock();
+}
+
+static void execlists_irq_handler(struct intel_engine_cs *engine, u16 iir)
+{
+	bool tasklet = false;
+
+	if (unlikely(iir & GT_CS_MASTER_ERROR_INTERRUPT)) {
+		u32 eir;
+
+		/* Upper 16b are the enabling mask, rsvd for internal errors */
+		eir = ENGINE_READ(engine, RING_EIR) & GENMASK(15, 0);
+		ENGINE_TRACE(engine, "CS error: %x\n", eir);
+
+		/* Disable the error interrupt until after the reset */
+		if (likely(eir)) {
+			ENGINE_WRITE(engine, RING_EMR, ~0u);
+			ENGINE_WRITE(engine, RING_EIR, eir);
+			WRITE_ONCE(engine->execlists.error_interrupt, eir);
+			tasklet = true;
+		}
+	}
+
+	if (iir & GT_WAIT_SEMAPHORE_INTERRUPT) {
+		WRITE_ONCE(engine->execlists.yield,
+			   ENGINE_READ_FW(engine, RING_EXECLIST_STATUS_HI));
+		ENGINE_TRACE(engine, "semaphore yield: %08x\n",
+			     engine->execlists.yield);
+		if (del_timer(&engine->execlists.timer))
+			tasklet = true;
+	}
+
+	if (iir & GT_CONTEXT_SWITCH_INTERRUPT)
+		tasklet = true;
+
+	if (iir & GT_RENDER_USER_INTERRUPT)
+		intel_engine_signal_breadcrumbs(engine);
+
+	if (tasklet)
+		tasklet_hi_schedule(&engine->sched_engine->tasklet);
+}
+
+static void __execlists_kick(struct intel_engine_execlists *execlists)
+{
+	struct intel_engine_cs *engine =
+		container_of(execlists, typeof(*engine), execlists);
+
+	/* Kick the tasklet for some interrupt coalescing and reset handling */
+	tasklet_hi_schedule(&engine->sched_engine->tasklet);
+}
+
+#define execlists_kick(t, member) \
+	__execlists_kick(container_of(t, struct intel_engine_execlists, member))
+
+static void execlists_timeslice(struct timer_list *timer)
+{
+	execlists_kick(timer, timer);
+}
+
+static void execlists_preempt(struct timer_list *timer)
+{
+	execlists_kick(timer, preempt);
+}
+
+static void queue_request(struct intel_engine_cs *engine,
+			  struct i915_request *rq)
+{
+	GEM_BUG_ON(!list_empty(&rq->sched.link));
+	list_add_tail(&rq->sched.link,
+		      i915_sched_lookup_priolist(engine->sched_engine,
+						 rq_prio(rq)));
+	set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+}
+
+static bool submit_queue(struct intel_engine_cs *engine,
+			 const struct i915_request *rq)
+{
+	struct i915_sched_engine *sched_engine = engine->sched_engine;
+
+	if (rq_prio(rq) <= sched_engine->queue_priority_hint)
+		return false;
+
+	sched_engine->queue_priority_hint = rq_prio(rq);
+	return true;
+}
+
+static bool ancestor_on_hold(const struct intel_engine_cs *engine,
+			     const struct i915_request *rq)
+{
+	GEM_BUG_ON(i915_request_on_hold(rq));
+	return !list_empty(&engine->sched_engine->hold) && hold_request(rq);
+}
+
+static void execlists_submit_request(struct i915_request *request)
+{
+	struct intel_engine_cs *engine = request->engine;
+	unsigned long flags;
+
+	/* Will be called from irq-context when using foreign fences. */
+	spin_lock_irqsave(&engine->sched_engine->lock, flags);
+
+	if (unlikely(ancestor_on_hold(engine, request))) {
+		RQ_TRACE(request, "ancestor on hold\n");
+		list_add_tail(&request->sched.link,
+			      &engine->sched_engine->hold);
+		i915_request_set_hold(request);
+	} else {
+		queue_request(engine, request);
+
+		GEM_BUG_ON(i915_sched_engine_is_empty(engine->sched_engine));
+		GEM_BUG_ON(list_empty(&request->sched.link));
+
+		if (submit_queue(engine, request))
+			__execlists_kick(&engine->execlists);
+	}
+
+	spin_unlock_irqrestore(&engine->sched_engine->lock, flags);
+}
+
+static int
+__execlists_context_pre_pin(struct intel_context *ce,
+			    struct intel_engine_cs *engine,
+			    struct i915_gem_ww_ctx *ww, void **vaddr)
+{
+	int err;
+
+	err = lrc_pre_pin(ce, engine, ww, vaddr);
+	if (err)
+		return err;
+
+	if (!__test_and_set_bit(CONTEXT_INIT_BIT, &ce->flags)) {
+		lrc_init_state(ce, engine, *vaddr);
+
+		__i915_gem_object_flush_map(ce->state->obj, 0, engine->context_size);
+	}
+
+	return 0;
+}
+
+static int execlists_context_pre_pin(struct intel_context *ce,
+				     struct i915_gem_ww_ctx *ww,
+				     void **vaddr)
+{
+	return __execlists_context_pre_pin(ce, ce->engine, ww, vaddr);
+}
+
+static int execlists_context_pin(struct intel_context *ce, void *vaddr)
+{
+	return lrc_pin(ce, ce->engine, vaddr);
+}
+
+static int execlists_context_alloc(struct intel_context *ce)
+{
+	return lrc_alloc(ce, ce->engine);
+}
+
+static void execlists_context_cancel_request(struct intel_context *ce,
+					     struct i915_request *rq)
+{
+	struct intel_engine_cs *engine = NULL;
+
+	i915_request_active_engine(rq, &engine);
+
+	if (engine && intel_engine_pulse(engine))
+		intel_gt_handle_error(engine->gt, engine->mask, 0,
+				      "request cancellation by %s",
+				      current->comm);
+}
+
+static struct intel_context *
+execlists_create_parallel(struct intel_engine_cs **engines,
+			  unsigned int num_siblings,
+			  unsigned int width)
+{
+	struct intel_context *parent = NULL, *ce, *err;
+	int i;
+
+	GEM_BUG_ON(num_siblings != 1);
+
+	for (i = 0; i < width; ++i) {
+		ce = intel_context_create(engines[i]);
+		if (IS_ERR(ce)) {
+			err = ce;
+			goto unwind;
+		}
+
+		if (i == 0)
+			parent = ce;
+		else
+			intel_context_bind_parent_child(parent, ce);
+	}
+
+	parent->parallel.fence_context = dma_fence_context_alloc(1);
+
+	intel_context_set_nopreempt(parent);
+	for_each_child(parent, ce)
+		intel_context_set_nopreempt(ce);
+
+	return parent;
+
+unwind:
+	if (parent)
+		intel_context_put(parent);
+	return err;
+}
+
+static const struct intel_context_ops execlists_context_ops = {
+	.flags = COPS_HAS_INFLIGHT | COPS_RUNTIME_CYCLES,
+
+	.alloc = execlists_context_alloc,
+
+	.cancel_request = execlists_context_cancel_request,
+
+	.pre_pin = execlists_context_pre_pin,
+	.pin = execlists_context_pin,
+	.unpin = lrc_unpin,
+	.post_unpin = lrc_post_unpin,
+
+	.enter = intel_context_enter_engine,
+	.exit = intel_context_exit_engine,
+
+	.reset = lrc_reset,
+	.destroy = lrc_destroy,
+
+	.create_parallel = execlists_create_parallel,
+	.create_virtual = execlists_create_virtual,
+};
+
+static int emit_pdps(struct i915_request *rq)
+{
+	const struct intel_engine_cs * const engine = rq->engine;
+	struct i915_ppgtt * const ppgtt = i915_vm_to_ppgtt(rq->context->vm);
+	int err, i;
+	u32 *cs;
+
+	GEM_BUG_ON(intel_vgpu_active(rq->engine->i915));
+
+	/*
+	 * Beware ye of the dragons, this sequence is magic!
+	 *
+	 * Small changes to this sequence can cause anything from
+	 * GPU hangs to forcewake errors and machine lockups!
+	 */
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	/* Flush any residual operations from the context load */
+	err = engine->emit_flush(rq, EMIT_FLUSH);
+	if (err)
+		return err;
+
+	/* Magic required to prevent forcewake errors! */
+	err = engine->emit_flush(rq, EMIT_INVALIDATE);
+	if (err)
+		return err;
+
+	cs = intel_ring_begin(rq, 4 * GEN8_3LVL_PDPES + 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/* Ensure the LRI have landed before we invalidate & continue */
+	*cs++ = MI_LOAD_REGISTER_IMM(2 * GEN8_3LVL_PDPES) | MI_LRI_FORCE_POSTED;
+	for (i = GEN8_3LVL_PDPES; i--; ) {
+		const dma_addr_t pd_daddr = i915_page_dir_dma_addr(ppgtt, i);
+		u32 base = engine->mmio_base;
+
+		*cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_UDW(base, i));
+		*cs++ = upper_32_bits(pd_daddr);
+		*cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_LDW(base, i));
+		*cs++ = lower_32_bits(pd_daddr);
+	}
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	intel_ring_advance(rq, cs);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int execlists_request_alloc(struct i915_request *request)
+{
+	int ret;
+
+	GEM_BUG_ON(!intel_context_is_pinned(request->context));
+
+	/*
+	 * Flush enough space to reduce the likelihood of waiting after
+	 * we start building the request - in which case we will just
+	 * have to repeat work.
+	 */
+	request->reserved_space += EXECLISTS_REQUEST_SIZE;
+
+	/*
+	 * Note that after this point, we have committed to using
+	 * this request as it is being used to both track the
+	 * state of engine initialisation and liveness of the
+	 * golden renderstate above. Think twice before you try
+	 * to cancel/unwind this request now.
+	 */
+
+	if (!i915_vm_is_4lvl(request->context->vm)) {
+		ret = emit_pdps(request);
+		if (ret)
+			return ret;
+	}
+
+	/* Unconditionally invalidate GPU caches and TLBs. */
+	ret = request->engine->emit_flush(request, EMIT_INVALIDATE);
+	if (ret)
+		return ret;
+
+	request->reserved_space -= EXECLISTS_REQUEST_SIZE;
+	return 0;
+}
+
+static void reset_csb_pointers(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	const unsigned int reset_value = execlists->csb_size - 1;
+
+	ring_set_paused(engine, 0);
+
+	/*
+	 * Sometimes Icelake forgets to reset its pointers on a GPU reset.
+	 * Bludgeon them with a mmio update to be sure.
+	 */
+	ENGINE_WRITE(engine, RING_CONTEXT_STATUS_PTR,
+		     0xffff << 16 | reset_value << 8 | reset_value);
+	ENGINE_POSTING_READ(engine, RING_CONTEXT_STATUS_PTR);
+
+	/*
+	 * After a reset, the HW starts writing into CSB entry [0]. We
+	 * therefore have to set our HEAD pointer back one entry so that
+	 * the *first* entry we check is entry 0. To complicate this further,
+	 * as we don't wait for the first interrupt after reset, we have to
+	 * fake the HW write to point back to the last entry so that our
+	 * inline comparison of our cached head position against the last HW
+	 * write works even before the first interrupt.
+	 */
+	execlists->csb_head = reset_value;
+	WRITE_ONCE(*execlists->csb_write, reset_value);
+	wmb(); /* Make sure this is visible to HW (paranoia?) */
+
+	/* Check that the GPU does indeed update the CSB entries! */
+	memset(execlists->csb_status, -1, (reset_value + 1) * sizeof(u64));
+	drm_clflush_virt_range(execlists->csb_status,
+			       execlists->csb_size *
+			       sizeof(execlists->csb_status));
+
+	/* Once more for luck and our trusty paranoia */
+	ENGINE_WRITE(engine, RING_CONTEXT_STATUS_PTR,
+		     0xffff << 16 | reset_value << 8 | reset_value);
+	ENGINE_POSTING_READ(engine, RING_CONTEXT_STATUS_PTR);
+
+	GEM_BUG_ON(READ_ONCE(*execlists->csb_write) != reset_value);
+}
+
+static void sanitize_hwsp(struct intel_engine_cs *engine)
+{
+	struct intel_timeline *tl;
+
+	list_for_each_entry(tl, &engine->status_page.timelines, engine_link)
+		intel_timeline_reset_seqno(tl);
+}
+
+static void execlists_sanitize(struct intel_engine_cs *engine)
+{
+	GEM_BUG_ON(execlists_active(&engine->execlists));
+
+	/*
+	 * Poison residual state on resume, in case the suspend didn't!
+	 *
+	 * We have to assume that across suspend/resume (or other loss
+	 * of control) that the contents of our pinned buffers has been
+	 * lost, replaced by garbage. Since this doesn't always happen,
+	 * let's poison such state so that we more quickly spot when
+	 * we falsely assume it has been preserved.
+	 */
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		memset(engine->status_page.addr, POISON_INUSE, PAGE_SIZE);
+
+	reset_csb_pointers(engine);
+
+	/*
+	 * The kernel_context HWSP is stored in the status_page. As above,
+	 * that may be lost on resume/initialisation, and so we need to
+	 * reset the value in the HWSP.
+	 */
+	sanitize_hwsp(engine);
+
+	/* And scrub the dirty cachelines for the HWSP */
+	drm_clflush_virt_range(engine->status_page.addr, PAGE_SIZE);
+
+	intel_engine_reset_pinned_contexts(engine);
+}
+
+static void enable_error_interrupt(struct intel_engine_cs *engine)
+{
+	u32 status;
+
+	engine->execlists.error_interrupt = 0;
+	ENGINE_WRITE(engine, RING_EMR, ~0u);
+	ENGINE_WRITE(engine, RING_EIR, ~0u); /* clear all existing errors */
+
+	status = ENGINE_READ(engine, RING_ESR);
+	if (unlikely(status)) {
+		drm_err(&engine->i915->drm,
+			"engine '%s' resumed still in error: %08x\n",
+			engine->name, status);
+		__intel_gt_reset(engine->gt, engine->mask);
+	}
+
+	/*
+	 * On current gen8+, we have 2 signals to play with
+	 *
+	 * - I915_ERROR_INSTUCTION (bit 0)
+	 *
+	 *    Generate an error if the command parser encounters an invalid
+	 *    instruction
+	 *
+	 *    This is a fatal error.
+	 *
+	 * - CP_PRIV (bit 2)
+	 *
+	 *    Generate an error on privilege violation (where the CP replaces
+	 *    the instruction with a no-op). This also fires for writes into
+	 *    read-only scratch pages.
+	 *
+	 *    This is a non-fatal error, parsing continues.
+	 *
+	 * * there are a few others defined for odd HW that we do not use
+	 *
+	 * Since CP_PRIV fires for cases where we have chosen to ignore the
+	 * error (as the HW is validating and suppressing the mistakes), we
+	 * only unmask the instruction error bit.
+	 */
+	ENGINE_WRITE(engine, RING_EMR, ~I915_ERROR_INSTRUCTION);
+}
+
+static void enable_execlists(struct intel_engine_cs *engine)
+{
+	u32 mode;
+
+	assert_forcewakes_active(engine->uncore, FORCEWAKE_ALL);
+
+	intel_engine_set_hwsp_writemask(engine, ~0u); /* HWSTAM */
+
+	if (GRAPHICS_VER(engine->i915) >= 11)
+		mode = _MASKED_BIT_ENABLE(GEN11_GFX_DISABLE_LEGACY_MODE);
+	else
+		mode = _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE);
+	ENGINE_WRITE_FW(engine, RING_MODE_GEN7, mode);
+
+	ENGINE_WRITE_FW(engine, RING_MI_MODE, _MASKED_BIT_DISABLE(STOP_RING));
+
+	ENGINE_WRITE_FW(engine,
+			RING_HWS_PGA,
+			i915_ggtt_offset(engine->status_page.vma));
+	ENGINE_POSTING_READ(engine, RING_HWS_PGA);
+
+	enable_error_interrupt(engine);
+}
+
+static int execlists_resume(struct intel_engine_cs *engine)
+{
+	intel_mocs_init_engine(engine);
+	intel_breadcrumbs_reset(engine->breadcrumbs);
+
+	enable_execlists(engine);
+
+	if (engine->flags & I915_ENGINE_FIRST_RENDER_COMPUTE)
+		xehp_enable_ccs_engines(engine);
+
+	return 0;
+}
+
+static void execlists_reset_prepare(struct intel_engine_cs *engine)
+{
+	ENGINE_TRACE(engine, "depth<-%d\n",
+		     atomic_read(&engine->sched_engine->tasklet.count));
+
+	/*
+	 * Prevent request submission to the hardware until we have
+	 * completed the reset in i915_gem_reset_finish(). If a request
+	 * is completed by one engine, it may then queue a request
+	 * to a second via its execlists->tasklet *just* as we are
+	 * calling engine->resume() and also writing the ELSP.
+	 * Turning off the execlists->tasklet until the reset is over
+	 * prevents the race.
+	 */
+	__tasklet_disable_sync_once(&engine->sched_engine->tasklet);
+	GEM_BUG_ON(!reset_in_progress(engine));
+
+	/*
+	 * We stop engines, otherwise we might get failed reset and a
+	 * dead gpu (on elk). Also as modern gpu as kbl can suffer
+	 * from system hang if batchbuffer is progressing when
+	 * the reset is issued, regardless of READY_TO_RESET ack.
+	 * Thus assume it is best to stop engines on all gens
+	 * where we have a gpu reset.
+	 *
+	 * WaKBLVECSSemaphoreWaitPoll:kbl (on ALL_ENGINES)
+	 *
+	 * FIXME: Wa for more modern gens needs to be validated
+	 */
+	ring_set_paused(engine, 1);
+	intel_engine_stop_cs(engine);
+
+	/*
+	 * Wa_22011802037:gen11/gen12: In addition to stopping the cs, we need
+	 * to wait for any pending mi force wakeups
+	 */
+	if (IS_GRAPHICS_VER(engine->i915, 11, 12))
+		intel_engine_wait_for_pending_mi_fw(engine);
+
+	engine->execlists.reset_ccid = active_ccid(engine);
+}
+
+static struct i915_request **
+reset_csb(struct intel_engine_cs *engine, struct i915_request **inactive)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+
+	drm_clflush_virt_range(execlists->csb_write,
+			       sizeof(execlists->csb_write[0]));
+
+	inactive = process_csb(engine, inactive); /* drain preemption events */
+
+	/* Following the reset, we need to reload the CSB read/write pointers */
+	reset_csb_pointers(engine);
+
+	return inactive;
+}
+
+static void
+execlists_reset_active(struct intel_engine_cs *engine, bool stalled)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+	u32 head;
+
+	/*
+	 * Save the currently executing context, even if we completed
+	 * its request, it was still running at the time of the
+	 * reset and will have been clobbered.
+	 */
+	rq = active_context(engine, engine->execlists.reset_ccid);
+	if (!rq)
+		return;
+
+	ce = rq->context;
+	GEM_BUG_ON(!i915_vma_is_pinned(ce->state));
+
+	if (__i915_request_is_complete(rq)) {
+		/* Idle context; tidy up the ring so we can restart afresh */
+		head = intel_ring_wrap(ce->ring, rq->tail);
+		goto out_replay;
+	}
+
+	/* We still have requests in-flight; the engine should be active */
+	GEM_BUG_ON(!intel_engine_pm_is_awake(engine));
+
+	/* Context has requests still in-flight; it should not be idle! */
+	GEM_BUG_ON(i915_active_is_idle(&ce->active));
+
+	rq = active_request(ce->timeline, rq);
+	head = intel_ring_wrap(ce->ring, rq->head);
+	GEM_BUG_ON(head == ce->ring->tail);
+
+	/*
+	 * If this request hasn't started yet, e.g. it is waiting on a
+	 * semaphore, we need to avoid skipping the request or else we
+	 * break the signaling chain. However, if the context is corrupt
+	 * the request will not restart and we will be stuck with a wedged
+	 * device. It is quite often the case that if we issue a reset
+	 * while the GPU is loading the context image, that the context
+	 * image becomes corrupt.
+	 *
+	 * Otherwise, if we have not started yet, the request should replay
+	 * perfectly and we do not need to flag the result as being erroneous.
+	 */
+	if (!__i915_request_has_started(rq))
+		goto out_replay;
+
+	/*
+	 * If the request was innocent, we leave the request in the ELSP
+	 * and will try to replay it on restarting. The context image may
+	 * have been corrupted by the reset, in which case we may have
+	 * to service a new GPU hang, but more likely we can continue on
+	 * without impact.
+	 *
+	 * If the request was guilty, we presume the context is corrupt
+	 * and have to at least restore the RING register in the context
+	 * image back to the expected values to skip over the guilty request.
+	 */
+	__i915_request_reset(rq, stalled);
+
+	/*
+	 * We want a simple context + ring to execute the breadcrumb update.
+	 * We cannot rely on the context being intact across the GPU hang,
+	 * so clear it and rebuild just what we need for the breadcrumb.
+	 * All pending requests for this context will be zapped, and any
+	 * future request will be after userspace has had the opportunity
+	 * to recreate its own state.
+	 */
+out_replay:
+	ENGINE_TRACE(engine, "replay {head:%04x, tail:%04x}\n",
+		     head, ce->ring->tail);
+	lrc_reset_regs(ce, engine);
+	ce->lrc.lrca = lrc_update_regs(ce, engine, head);
+}
+
+static void execlists_reset_csb(struct intel_engine_cs *engine, bool stalled)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	struct i915_request *post[2 * EXECLIST_MAX_PORTS];
+	struct i915_request **inactive;
+
+	rcu_read_lock();
+	inactive = reset_csb(engine, post);
+
+	execlists_reset_active(engine, true);
+
+	inactive = cancel_port_requests(execlists, inactive);
+	post_process_csb(post, inactive);
+	rcu_read_unlock();
+}
+
+static void execlists_reset_rewind(struct intel_engine_cs *engine, bool stalled)
+{
+	unsigned long flags;
+
+	ENGINE_TRACE(engine, "\n");
+
+	/* Process the csb, find the guilty context and throw away */
+	execlists_reset_csb(engine, stalled);
+
+	/* Push back any incomplete requests for replay after the reset. */
+	rcu_read_lock();
+	spin_lock_irqsave(&engine->sched_engine->lock, flags);
+	__unwind_incomplete_requests(engine);
+	spin_unlock_irqrestore(&engine->sched_engine->lock, flags);
+	rcu_read_unlock();
+}
+
+static void nop_submission_tasklet(struct tasklet_struct *t)
+{
+	struct i915_sched_engine *sched_engine =
+		from_tasklet(sched_engine, t, tasklet);
+	struct intel_engine_cs * const engine = sched_engine->private_data;
+
+	/* The driver is wedged; don't process any more events. */
+	WRITE_ONCE(engine->sched_engine->queue_priority_hint, INT_MIN);
+}
+
+static void execlists_reset_cancel(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	struct i915_sched_engine * const sched_engine = engine->sched_engine;
+	struct i915_request *rq, *rn;
+	struct rb_node *rb;
+	unsigned long flags;
+
+	ENGINE_TRACE(engine, "\n");
+
+	/*
+	 * Before we call engine->cancel_requests(), we should have exclusive
+	 * access to the submission state. This is arranged for us by the
+	 * caller disabling the interrupt generation, the tasklet and other
+	 * threads that may then access the same state, giving us a free hand
+	 * to reset state. However, we still need to let lockdep be aware that
+	 * we know this state may be accessed in hardirq context, so we
+	 * disable the irq around this manipulation and we want to keep
+	 * the spinlock focused on its duties and not accidentally conflate
+	 * coverage to the submission's irq state. (Similarly, although we
+	 * shouldn't need to disable irq around the manipulation of the
+	 * submission's irq state, we also wish to remind ourselves that
+	 * it is irq state.)
+	 */
+	execlists_reset_csb(engine, true);
+
+	rcu_read_lock();
+	spin_lock_irqsave(&engine->sched_engine->lock, flags);
+
+	/* Mark all executing requests as skipped. */
+	list_for_each_entry(rq, &engine->sched_engine->requests, sched.link)
+		i915_request_put(i915_request_mark_eio(rq));
+	intel_engine_signal_breadcrumbs(engine);
+
+	/* Flush the queued requests to the timeline list (for retiring). */
+	while ((rb = rb_first_cached(&sched_engine->queue))) {
+		struct i915_priolist *p = to_priolist(rb);
+
+		priolist_for_each_request_consume(rq, rn, p) {
+			if (i915_request_mark_eio(rq)) {
+				__i915_request_submit(rq);
+				i915_request_put(rq);
+			}
+		}
+
+		rb_erase_cached(&p->node, &sched_engine->queue);
+		i915_priolist_free(p);
+	}
+
+	/* On-hold requests will be flushed to timeline upon their release */
+	list_for_each_entry(rq, &sched_engine->hold, sched.link)
+		i915_request_put(i915_request_mark_eio(rq));
+
+	/* Cancel all attached virtual engines */
+	while ((rb = rb_first_cached(&execlists->virtual))) {
+		struct virtual_engine *ve =
+			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
+
+		rb_erase_cached(rb, &execlists->virtual);
+		RB_CLEAR_NODE(rb);
+
+		spin_lock(&ve->base.sched_engine->lock);
+		rq = fetch_and_zero(&ve->request);
+		if (rq) {
+			if (i915_request_mark_eio(rq)) {
+				rq->engine = engine;
+				__i915_request_submit(rq);
+				i915_request_put(rq);
+			}
+			i915_request_put(rq);
+
+			ve->base.sched_engine->queue_priority_hint = INT_MIN;
+		}
+		spin_unlock(&ve->base.sched_engine->lock);
+	}
+
+	/* Remaining _unready_ requests will be nop'ed when submitted */
+
+	sched_engine->queue_priority_hint = INT_MIN;
+	sched_engine->queue = RB_ROOT_CACHED;
+
+	GEM_BUG_ON(__tasklet_is_enabled(&engine->sched_engine->tasklet));
+	engine->sched_engine->tasklet.callback = nop_submission_tasklet;
+
+	spin_unlock_irqrestore(&engine->sched_engine->lock, flags);
+	rcu_read_unlock();
+}
+
+static void execlists_reset_finish(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+
+	/*
+	 * After a GPU reset, we may have requests to replay. Do so now while
+	 * we still have the forcewake to be sure that the GPU is not allowed
+	 * to sleep before we restart and reload a context.
+	 *
+	 * If the GPU reset fails, the engine may still be alive with requests
+	 * inflight. We expect those to complete, or for the device to be
+	 * reset as the next level of recovery, and as a final resort we
+	 * will declare the device wedged.
+	 */
+	GEM_BUG_ON(!reset_in_progress(engine));
+
+	/* And kick in case we missed a new request submission. */
+	if (__tasklet_enable(&engine->sched_engine->tasklet))
+		__execlists_kick(execlists);
+
+	ENGINE_TRACE(engine, "depth->%d\n",
+		     atomic_read(&engine->sched_engine->tasklet.count));
+}
+
+static void gen8_logical_ring_enable_irq(struct intel_engine_cs *engine)
+{
+	ENGINE_WRITE(engine, RING_IMR,
+		     ~(engine->irq_enable_mask | engine->irq_keep_mask));
+	ENGINE_POSTING_READ(engine, RING_IMR);
+}
+
+static void gen8_logical_ring_disable_irq(struct intel_engine_cs *engine)
+{
+	ENGINE_WRITE(engine, RING_IMR, ~engine->irq_keep_mask);
+}
+
+static void execlists_park(struct intel_engine_cs *engine)
+{
+	cancel_timer(&engine->execlists.timer);
+	cancel_timer(&engine->execlists.preempt);
+}
+
+static void add_to_engine(struct i915_request *rq)
+{
+	lockdep_assert_held(&rq->engine->sched_engine->lock);
+	list_move_tail(&rq->sched.link, &rq->engine->sched_engine->requests);
+}
+
+static void remove_from_engine(struct i915_request *rq)
+{
+	struct intel_engine_cs *engine, *locked;
+
+	/*
+	 * Virtual engines complicate acquiring the engine timeline lock,
+	 * as their rq->engine pointer is not stable until under that
+	 * engine lock. The simple ploy we use is to take the lock then
+	 * check that the rq still belongs to the newly locked engine.
+	 */
+	locked = READ_ONCE(rq->engine);
+	spin_lock_irq(&locked->sched_engine->lock);
+	while (unlikely(locked != (engine = READ_ONCE(rq->engine)))) {
+		spin_unlock(&locked->sched_engine->lock);
+		spin_lock(&engine->sched_engine->lock);
+		locked = engine;
+	}
+	list_del_init(&rq->sched.link);
+
+	clear_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+	clear_bit(I915_FENCE_FLAG_HOLD, &rq->fence.flags);
+
+	/* Prevent further __await_execution() registering a cb, then flush */
+	set_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags);
+
+	spin_unlock_irq(&locked->sched_engine->lock);
+
+	i915_request_notify_execute_cb_imm(rq);
+}
+
+static bool can_preempt(struct intel_engine_cs *engine)
+{
+	if (GRAPHICS_VER(engine->i915) > 8)
+		return true;
+
+	/* GPGPU on bdw requires extra w/a; not implemented */
+	return engine->class != RENDER_CLASS;
+}
+
+static void kick_execlists(const struct i915_request *rq, int prio)
+{
+	struct intel_engine_cs *engine = rq->engine;
+	struct i915_sched_engine *sched_engine = engine->sched_engine;
+	const struct i915_request *inflight;
+
+	/*
+	 * We only need to kick the tasklet once for the high priority
+	 * new context we add into the queue.
+	 */
+	if (prio <= sched_engine->queue_priority_hint)
+		return;
+
+	rcu_read_lock();
+
+	/* Nothing currently active? We're overdue for a submission! */
+	inflight = execlists_active(&engine->execlists);
+	if (!inflight)
+		goto unlock;
+
+	/*
+	 * If we are already the currently executing context, don't
+	 * bother evaluating if we should preempt ourselves.
+	 */
+	if (inflight->context == rq->context)
+		goto unlock;
+
+	ENGINE_TRACE(engine,
+		     "bumping queue-priority-hint:%d for rq:%llx:%lld, inflight:%llx:%lld prio %d\n",
+		     prio,
+		     rq->fence.context, rq->fence.seqno,
+		     inflight->fence.context, inflight->fence.seqno,
+		     inflight->sched.attr.priority);
+
+	sched_engine->queue_priority_hint = prio;
+
+	/*
+	 * Allow preemption of low -> normal -> high, but we do
+	 * not allow low priority tasks to preempt other low priority
+	 * tasks under the impression that latency for low priority
+	 * tasks does not matter (as much as background throughput),
+	 * so kiss.
+	 */
+	if (prio >= max(I915_PRIORITY_NORMAL, rq_prio(inflight)))
+		tasklet_hi_schedule(&sched_engine->tasklet);
+
+unlock:
+	rcu_read_unlock();
+}
+
+static void execlists_set_default_submission(struct intel_engine_cs *engine)
+{
+	engine->submit_request = execlists_submit_request;
+	engine->sched_engine->schedule = i915_schedule;
+	engine->sched_engine->kick_backend = kick_execlists;
+	engine->sched_engine->tasklet.callback = execlists_submission_tasklet;
+}
+
+static void execlists_shutdown(struct intel_engine_cs *engine)
+{
+	/* Synchronise with residual timers and any softirq they raise */
+	del_timer_sync(&engine->execlists.timer);
+	del_timer_sync(&engine->execlists.preempt);
+	tasklet_kill(&engine->sched_engine->tasklet);
+}
+
+static void execlists_release(struct intel_engine_cs *engine)
+{
+	engine->sanitize = NULL; /* no longer in control, nothing to sanitize */
+
+	execlists_shutdown(engine);
+
+	intel_engine_cleanup_common(engine);
+	lrc_fini_wa_ctx(engine);
+}
+
+static ktime_t __execlists_engine_busyness(struct intel_engine_cs *engine,
+					   ktime_t *now)
+{
+	struct intel_engine_execlists_stats *stats = &engine->stats.execlists;
+	ktime_t total = stats->total;
+
+	/*
+	 * If the engine is executing something at the moment
+	 * add it to the total.
+	 */
+	*now = ktime_get();
+	if (READ_ONCE(stats->active))
+		total = ktime_add(total, ktime_sub(*now, stats->start));
+
+	return total;
+}
+
+static ktime_t execlists_engine_busyness(struct intel_engine_cs *engine,
+					 ktime_t *now)
+{
+	struct intel_engine_execlists_stats *stats = &engine->stats.execlists;
+	unsigned int seq;
+	ktime_t total;
+
+	do {
+		seq = read_seqcount_begin(&stats->lock);
+		total = __execlists_engine_busyness(engine, now);
+	} while (read_seqcount_retry(&stats->lock, seq));
+
+	return total;
+}
+
+static void
+logical_ring_default_vfuncs(struct intel_engine_cs *engine)
+{
+	/* Default vfuncs which can be overridden by each engine. */
+
+	engine->resume = execlists_resume;
+
+	engine->cops = &execlists_context_ops;
+	engine->request_alloc = execlists_request_alloc;
+	engine->add_active_request = add_to_engine;
+	engine->remove_active_request = remove_from_engine;
+
+	engine->reset.prepare = execlists_reset_prepare;
+	engine->reset.rewind = execlists_reset_rewind;
+	engine->reset.cancel = execlists_reset_cancel;
+	engine->reset.finish = execlists_reset_finish;
+
+	engine->park = execlists_park;
+	engine->unpark = NULL;
+
+	engine->emit_flush = gen8_emit_flush_xcs;
+	engine->emit_init_breadcrumb = gen8_emit_init_breadcrumb;
+	engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb_xcs;
+	if (GRAPHICS_VER(engine->i915) >= 12) {
+		engine->emit_fini_breadcrumb = gen12_emit_fini_breadcrumb_xcs;
+		engine->emit_flush = gen12_emit_flush_xcs;
+	}
+	engine->set_default_submission = execlists_set_default_submission;
+
+	if (GRAPHICS_VER(engine->i915) < 11) {
+		engine->irq_enable = gen8_logical_ring_enable_irq;
+		engine->irq_disable = gen8_logical_ring_disable_irq;
+	} else {
+		/*
+		 * TODO: On Gen11 interrupt masks need to be clear
+		 * to allow C6 entry. Keep interrupts enabled at
+		 * and take the hit of generating extra interrupts
+		 * until a more refined solution exists.
+		 */
+	}
+	intel_engine_set_irq_handler(engine, execlists_irq_handler);
+
+	engine->flags |= I915_ENGINE_SUPPORTS_STATS;
+	if (!intel_vgpu_active(engine->i915)) {
+		engine->flags |= I915_ENGINE_HAS_SEMAPHORES;
+		if (can_preempt(engine)) {
+			engine->flags |= I915_ENGINE_HAS_PREEMPTION;
+			if (CONFIG_DRM_I915_TIMESLICE_DURATION)
+				engine->flags |= I915_ENGINE_HAS_TIMESLICES;
+		}
+	}
+
+	if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 50)) {
+		if (intel_engine_has_preemption(engine))
+			engine->emit_bb_start = gen125_emit_bb_start;
+		else
+			engine->emit_bb_start = gen125_emit_bb_start_noarb;
+	} else {
+		if (intel_engine_has_preemption(engine))
+			engine->emit_bb_start = gen8_emit_bb_start;
+		else
+			engine->emit_bb_start = gen8_emit_bb_start_noarb;
+	}
+
+	engine->busyness = execlists_engine_busyness;
+}
+
+static void logical_ring_default_irqs(struct intel_engine_cs *engine)
+{
+	unsigned int shift = 0;
+
+	if (GRAPHICS_VER(engine->i915) < 11) {
+		const u8 irq_shifts[] = {
+			[RCS0]  = GEN8_RCS_IRQ_SHIFT,
+			[BCS0]  = GEN8_BCS_IRQ_SHIFT,
+			[VCS0]  = GEN8_VCS0_IRQ_SHIFT,
+			[VCS1]  = GEN8_VCS1_IRQ_SHIFT,
+			[VECS0] = GEN8_VECS_IRQ_SHIFT,
+		};
+
+		shift = irq_shifts[engine->id];
+	}
+
+	engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT << shift;
+	engine->irq_keep_mask = GT_CONTEXT_SWITCH_INTERRUPT << shift;
+	engine->irq_keep_mask |= GT_CS_MASTER_ERROR_INTERRUPT << shift;
+	engine->irq_keep_mask |= GT_WAIT_SEMAPHORE_INTERRUPT << shift;
+}
+
+static void rcs_submission_override(struct intel_engine_cs *engine)
+{
+	switch (GRAPHICS_VER(engine->i915)) {
+	case 12:
+		engine->emit_flush = gen12_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen12_emit_fini_breadcrumb_rcs;
+		break;
+	case 11:
+		engine->emit_flush = gen11_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen11_emit_fini_breadcrumb_rcs;
+		break;
+	default:
+		engine->emit_flush = gen8_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb_rcs;
+		break;
+	}
+}
+
+int intel_execlists_submission_setup(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists * const execlists = &engine->execlists;
+	struct drm_i915_private *i915 = engine->i915;
+	struct intel_uncore *uncore = engine->uncore;
+	u32 base = engine->mmio_base;
+
+	tasklet_setup(&engine->sched_engine->tasklet, execlists_submission_tasklet);
+	timer_setup(&engine->execlists.timer, execlists_timeslice, 0);
+	timer_setup(&engine->execlists.preempt, execlists_preempt, 0);
+
+	logical_ring_default_vfuncs(engine);
+	logical_ring_default_irqs(engine);
+
+	seqcount_init(&engine->stats.execlists.lock);
+
+	if (engine->flags & I915_ENGINE_HAS_RCS_REG_STATE)
+		rcs_submission_override(engine);
+
+	lrc_init_wa_ctx(engine);
+
+	if (HAS_LOGICAL_RING_ELSQ(i915)) {
+		execlists->submit_reg = uncore->regs +
+			i915_mmio_reg_offset(RING_EXECLIST_SQ_CONTENTS(base));
+		execlists->ctrl_reg = uncore->regs +
+			i915_mmio_reg_offset(RING_EXECLIST_CONTROL(base));
+
+		engine->fw_domain = intel_uncore_forcewake_for_reg(engine->uncore,
+				    RING_EXECLIST_CONTROL(engine->mmio_base),
+				    FW_REG_WRITE);
+	} else {
+		execlists->submit_reg = uncore->regs +
+			i915_mmio_reg_offset(RING_ELSP(base));
+	}
+
+	execlists->csb_status =
+		(u64 *)&engine->status_page.addr[I915_HWS_CSB_BUF0_INDEX];
+
+	execlists->csb_write =
+		&engine->status_page.addr[INTEL_HWS_CSB_WRITE_INDEX(i915)];
+
+	if (GRAPHICS_VER(i915) < 11)
+		execlists->csb_size = GEN8_CSB_ENTRIES;
+	else
+		execlists->csb_size = GEN11_CSB_ENTRIES;
+
+	engine->context_tag = GENMASK(BITS_PER_LONG - 2, 0);
+	if (GRAPHICS_VER(engine->i915) >= 11 &&
+	    GRAPHICS_VER_FULL(engine->i915) < IP_VER(12, 50)) {
+		execlists->ccid |= engine->instance << (GEN11_ENGINE_INSTANCE_SHIFT - 32);
+		execlists->ccid |= engine->class << (GEN11_ENGINE_CLASS_SHIFT - 32);
+	}
+
+	/* Finally, take ownership and responsibility for cleanup! */
+	engine->sanitize = execlists_sanitize;
+	engine->release = execlists_release;
+
+	return 0;
+}
+
+static struct list_head *virtual_queue(struct virtual_engine *ve)
+{
+	return &ve->base.sched_engine->default_priolist.requests;
+}
+
+static void rcu_virtual_context_destroy(struct work_struct *wrk)
+{
+	struct virtual_engine *ve =
+		container_of(wrk, typeof(*ve), rcu.work);
+	unsigned int n;
+
+	GEM_BUG_ON(ve->context.inflight);
+
+	/* Preempt-to-busy may leave a stale request behind. */
+	if (unlikely(ve->request)) {
+		struct i915_request *old;
+
+		spin_lock_irq(&ve->base.sched_engine->lock);
+
+		old = fetch_and_zero(&ve->request);
+		if (old) {
+			GEM_BUG_ON(!__i915_request_is_complete(old));
+			__i915_request_submit(old);
+			i915_request_put(old);
+		}
+
+		spin_unlock_irq(&ve->base.sched_engine->lock);
+	}
+
+	/*
+	 * Flush the tasklet in case it is still running on another core.
+	 *
+	 * This needs to be done before we remove ourselves from the siblings'
+	 * rbtrees as in the case it is running in parallel, it may reinsert
+	 * the rb_node into a sibling.
+	 */
+	tasklet_kill(&ve->base.sched_engine->tasklet);
+
+	/* Decouple ourselves from the siblings, no more access allowed. */
+	for (n = 0; n < ve->num_siblings; n++) {
+		struct intel_engine_cs *sibling = ve->siblings[n];
+		struct rb_node *node = &ve->nodes[sibling->id].rb;
+
+		if (RB_EMPTY_NODE(node))
+			continue;
+
+		spin_lock_irq(&sibling->sched_engine->lock);
+
+		/* Detachment is lazily performed in the sched_engine->tasklet */
+		if (!RB_EMPTY_NODE(node))
+			rb_erase_cached(node, &sibling->execlists.virtual);
+
+		spin_unlock_irq(&sibling->sched_engine->lock);
+	}
+	GEM_BUG_ON(__tasklet_is_scheduled(&ve->base.sched_engine->tasklet));
+	GEM_BUG_ON(!list_empty(virtual_queue(ve)));
+
+	lrc_fini(&ve->context);
+	intel_context_fini(&ve->context);
+
+	if (ve->base.breadcrumbs)
+		intel_breadcrumbs_put(ve->base.breadcrumbs);
+	if (ve->base.sched_engine)
+		i915_sched_engine_put(ve->base.sched_engine);
+	intel_engine_free_request_pool(&ve->base);
+
+	kfree(ve);
+}
+
+static void virtual_context_destroy(struct kref *kref)
+{
+	struct virtual_engine *ve =
+		container_of(kref, typeof(*ve), context.ref);
+
+	GEM_BUG_ON(!list_empty(&ve->context.signals));
+
+	/*
+	 * When destroying the virtual engine, we have to be aware that
+	 * it may still be in use from an hardirq/softirq context causing
+	 * the resubmission of a completed request (background completion
+	 * due to preempt-to-busy). Before we can free the engine, we need
+	 * to flush the submission code and tasklets that are still potentially
+	 * accessing the engine. Flushing the tasklets requires process context,
+	 * and since we can guard the resubmit onto the engine with an RCU read
+	 * lock, we can delegate the free of the engine to an RCU worker.
+	 */
+	INIT_RCU_WORK(&ve->rcu, rcu_virtual_context_destroy);
+	queue_rcu_work(system_wq, &ve->rcu);
+}
+
+static void virtual_engine_initial_hint(struct virtual_engine *ve)
+{
+	int swp;
+
+	/*
+	 * Pick a random sibling on starting to help spread the load around.
+	 *
+	 * New contexts are typically created with exactly the same order
+	 * of siblings, and often started in batches. Due to the way we iterate
+	 * the array of sibling when submitting requests, sibling[0] is
+	 * prioritised for dequeuing. If we make sure that sibling[0] is fairly
+	 * randomised across the system, we also help spread the load by the
+	 * first engine we inspect being different each time.
+	 *
+	 * NB This does not force us to execute on this engine, it will just
+	 * typically be the first we inspect for submission.
+	 */
+	swp = prandom_u32_max(ve->num_siblings);
+	if (swp)
+		swap(ve->siblings[swp], ve->siblings[0]);
+}
+
+static int virtual_context_alloc(struct intel_context *ce)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+
+	return lrc_alloc(ce, ve->siblings[0]);
+}
+
+static int virtual_context_pre_pin(struct intel_context *ce,
+				   struct i915_gem_ww_ctx *ww,
+				   void **vaddr)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+
+	 /* Note: we must use a real engine class for setting up reg state */
+	return __execlists_context_pre_pin(ce, ve->siblings[0], ww, vaddr);
+}
+
+static int virtual_context_pin(struct intel_context *ce, void *vaddr)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+
+	return lrc_pin(ce, ve->siblings[0], vaddr);
+}
+
+static void virtual_context_enter(struct intel_context *ce)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+	unsigned int n;
+
+	for (n = 0; n < ve->num_siblings; n++)
+		intel_engine_pm_get(ve->siblings[n]);
+
+	intel_timeline_enter(ce->timeline);
+}
+
+static void virtual_context_exit(struct intel_context *ce)
+{
+	struct virtual_engine *ve = container_of(ce, typeof(*ve), context);
+	unsigned int n;
+
+	intel_timeline_exit(ce->timeline);
+
+	for (n = 0; n < ve->num_siblings; n++)
+		intel_engine_pm_put(ve->siblings[n]);
+}
+
+static struct intel_engine_cs *
+virtual_get_sibling(struct intel_engine_cs *engine, unsigned int sibling)
+{
+	struct virtual_engine *ve = to_virtual_engine(engine);
+
+	if (sibling >= ve->num_siblings)
+		return NULL;
+
+	return ve->siblings[sibling];
+}
+
+static const struct intel_context_ops virtual_context_ops = {
+	.flags = COPS_HAS_INFLIGHT | COPS_RUNTIME_CYCLES,
+
+	.alloc = virtual_context_alloc,
+
+	.cancel_request = execlists_context_cancel_request,
+
+	.pre_pin = virtual_context_pre_pin,
+	.pin = virtual_context_pin,
+	.unpin = lrc_unpin,
+	.post_unpin = lrc_post_unpin,
+
+	.enter = virtual_context_enter,
+	.exit = virtual_context_exit,
+
+	.destroy = virtual_context_destroy,
+
+	.get_sibling = virtual_get_sibling,
+};
+
+static intel_engine_mask_t virtual_submission_mask(struct virtual_engine *ve)
+{
+	struct i915_request *rq;
+	intel_engine_mask_t mask;
+
+	rq = READ_ONCE(ve->request);
+	if (!rq)
+		return 0;
+
+	/* The rq is ready for submission; rq->execution_mask is now stable. */
+	mask = rq->execution_mask;
+	if (unlikely(!mask)) {
+		/* Invalid selection, submit to a random engine in error */
+		i915_request_set_error_once(rq, -ENODEV);
+		mask = ve->siblings[0]->mask;
+	}
+
+	ENGINE_TRACE(&ve->base, "rq=%llx:%lld, mask=%x, prio=%d\n",
+		     rq->fence.context, rq->fence.seqno,
+		     mask, ve->base.sched_engine->queue_priority_hint);
+
+	return mask;
+}
+
+static void virtual_submission_tasklet(struct tasklet_struct *t)
+{
+	struct i915_sched_engine *sched_engine =
+		from_tasklet(sched_engine, t, tasklet);
+	struct virtual_engine * const ve =
+		(struct virtual_engine *)sched_engine->private_data;
+	const int prio = READ_ONCE(sched_engine->queue_priority_hint);
+	intel_engine_mask_t mask;
+	unsigned int n;
+
+	rcu_read_lock();
+	mask = virtual_submission_mask(ve);
+	rcu_read_unlock();
+	if (unlikely(!mask))
+		return;
+
+	for (n = 0; n < ve->num_siblings; n++) {
+		struct intel_engine_cs *sibling = READ_ONCE(ve->siblings[n]);
+		struct ve_node * const node = &ve->nodes[sibling->id];
+		struct rb_node **parent, *rb;
+		bool first;
+
+		if (!READ_ONCE(ve->request))
+			break; /* already handled by a sibling's tasklet */
+
+		spin_lock_irq(&sibling->sched_engine->lock);
+
+		if (unlikely(!(mask & sibling->mask))) {
+			if (!RB_EMPTY_NODE(&node->rb)) {
+				rb_erase_cached(&node->rb,
+						&sibling->execlists.virtual);
+				RB_CLEAR_NODE(&node->rb);
+			}
+
+			goto unlock_engine;
+		}
+
+		if (unlikely(!RB_EMPTY_NODE(&node->rb))) {
+			/*
+			 * Cheat and avoid rebalancing the tree if we can
+			 * reuse this node in situ.
+			 */
+			first = rb_first_cached(&sibling->execlists.virtual) ==
+				&node->rb;
+			if (prio == node->prio || (prio > node->prio && first))
+				goto submit_engine;
+
+			rb_erase_cached(&node->rb, &sibling->execlists.virtual);
+		}
+
+		rb = NULL;
+		first = true;
+		parent = &sibling->execlists.virtual.rb_root.rb_node;
+		while (*parent) {
+			struct ve_node *other;
+
+			rb = *parent;
+			other = rb_entry(rb, typeof(*other), rb);
+			if (prio > other->prio) {
+				parent = &rb->rb_left;
+			} else {
+				parent = &rb->rb_right;
+				first = false;
+			}
+		}
+
+		rb_link_node(&node->rb, rb, parent);
+		rb_insert_color_cached(&node->rb,
+				       &sibling->execlists.virtual,
+				       first);
+
+submit_engine:
+		GEM_BUG_ON(RB_EMPTY_NODE(&node->rb));
+		node->prio = prio;
+		if (first && prio > sibling->sched_engine->queue_priority_hint)
+			tasklet_hi_schedule(&sibling->sched_engine->tasklet);
+
+unlock_engine:
+		spin_unlock_irq(&sibling->sched_engine->lock);
+
+		if (intel_context_inflight(&ve->context))
+			break;
+	}
+}
+
+static void virtual_submit_request(struct i915_request *rq)
+{
+	struct virtual_engine *ve = to_virtual_engine(rq->engine);
+	unsigned long flags;
+
+	ENGINE_TRACE(&ve->base, "rq=%llx:%lld\n",
+		     rq->fence.context,
+		     rq->fence.seqno);
+
+	GEM_BUG_ON(ve->base.submit_request != virtual_submit_request);
+
+	spin_lock_irqsave(&ve->base.sched_engine->lock, flags);
+
+	/* By the time we resubmit a request, it may be completed */
+	if (__i915_request_is_complete(rq)) {
+		__i915_request_submit(rq);
+		goto unlock;
+	}
+
+	if (ve->request) { /* background completion from preempt-to-busy */
+		GEM_BUG_ON(!__i915_request_is_complete(ve->request));
+		__i915_request_submit(ve->request);
+		i915_request_put(ve->request);
+	}
+
+	ve->base.sched_engine->queue_priority_hint = rq_prio(rq);
+	ve->request = i915_request_get(rq);
+
+	GEM_BUG_ON(!list_empty(virtual_queue(ve)));
+	list_move_tail(&rq->sched.link, virtual_queue(ve));
+
+	tasklet_hi_schedule(&ve->base.sched_engine->tasklet);
+
+unlock:
+	spin_unlock_irqrestore(&ve->base.sched_engine->lock, flags);
+}
+
+static struct intel_context *
+execlists_create_virtual(struct intel_engine_cs **siblings, unsigned int count,
+			 unsigned long flags)
+{
+	struct virtual_engine *ve;
+	unsigned int n;
+	int err;
+
+	ve = kzalloc(struct_size(ve, siblings, count), GFP_KERNEL);
+	if (!ve)
+		return ERR_PTR(-ENOMEM);
+
+	ve->base.i915 = siblings[0]->i915;
+	ve->base.gt = siblings[0]->gt;
+	ve->base.uncore = siblings[0]->uncore;
+	ve->base.id = -1;
+
+	ve->base.class = OTHER_CLASS;
+	ve->base.uabi_class = I915_ENGINE_CLASS_INVALID;
+	ve->base.instance = I915_ENGINE_CLASS_INVALID_VIRTUAL;
+	ve->base.uabi_instance = I915_ENGINE_CLASS_INVALID_VIRTUAL;
+
+	/*
+	 * The decision on whether to submit a request using semaphores
+	 * depends on the saturated state of the engine. We only compute
+	 * this during HW submission of the request, and we need for this
+	 * state to be globally applied to all requests being submitted
+	 * to this engine. Virtual engines encompass more than one physical
+	 * engine and so we cannot accurately tell in advance if one of those
+	 * engines is already saturated and so cannot afford to use a semaphore
+	 * and be pessimized in priority for doing so -- if we are the only
+	 * context using semaphores after all other clients have stopped, we
+	 * will be starved on the saturated system. Such a global switch for
+	 * semaphores is less than ideal, but alas is the current compromise.
+	 */
+	ve->base.saturated = ALL_ENGINES;
+
+	snprintf(ve->base.name, sizeof(ve->base.name), "virtual");
+
+	intel_engine_init_execlists(&ve->base);
+
+	ve->base.sched_engine = i915_sched_engine_create(ENGINE_VIRTUAL);
+	if (!ve->base.sched_engine) {
+		err = -ENOMEM;
+		goto err_put;
+	}
+	ve->base.sched_engine->private_data = &ve->base;
+
+	ve->base.cops = &virtual_context_ops;
+	ve->base.request_alloc = execlists_request_alloc;
+
+	ve->base.sched_engine->schedule = i915_schedule;
+	ve->base.sched_engine->kick_backend = kick_execlists;
+	ve->base.submit_request = virtual_submit_request;
+
+	INIT_LIST_HEAD(virtual_queue(ve));
+	tasklet_setup(&ve->base.sched_engine->tasklet, virtual_submission_tasklet);
+
+	intel_context_init(&ve->context, &ve->base);
+
+	ve->base.breadcrumbs = intel_breadcrumbs_create(NULL);
+	if (!ve->base.breadcrumbs) {
+		err = -ENOMEM;
+		goto err_put;
+	}
+
+	for (n = 0; n < count; n++) {
+		struct intel_engine_cs *sibling = siblings[n];
+
+		GEM_BUG_ON(!is_power_of_2(sibling->mask));
+		if (sibling->mask & ve->base.mask) {
+			DRM_DEBUG("duplicate %s entry in load balancer\n",
+				  sibling->name);
+			err = -EINVAL;
+			goto err_put;
+		}
+
+		/*
+		 * The virtual engine implementation is tightly coupled to
+		 * the execlists backend -- we push out request directly
+		 * into a tree inside each physical engine. We could support
+		 * layering if we handle cloning of the requests and
+		 * submitting a copy into each backend.
+		 */
+		if (sibling->sched_engine->tasklet.callback !=
+		    execlists_submission_tasklet) {
+			err = -ENODEV;
+			goto err_put;
+		}
+
+		GEM_BUG_ON(RB_EMPTY_NODE(&ve->nodes[sibling->id].rb));
+		RB_CLEAR_NODE(&ve->nodes[sibling->id].rb);
+
+		ve->siblings[ve->num_siblings++] = sibling;
+		ve->base.mask |= sibling->mask;
+		ve->base.logical_mask |= sibling->logical_mask;
+
+		/*
+		 * All physical engines must be compatible for their emission
+		 * functions (as we build the instructions during request
+		 * construction and do not alter them before submission
+		 * on the physical engine). We use the engine class as a guide
+		 * here, although that could be refined.
+		 */
+		if (ve->base.class != OTHER_CLASS) {
+			if (ve->base.class != sibling->class) {
+				DRM_DEBUG("invalid mixing of engine class, sibling %d, already %d\n",
+					  sibling->class, ve->base.class);
+				err = -EINVAL;
+				goto err_put;
+			}
+			continue;
+		}
+
+		ve->base.class = sibling->class;
+		ve->base.uabi_class = sibling->uabi_class;
+		snprintf(ve->base.name, sizeof(ve->base.name),
+			 "v%dx%d", ve->base.class, count);
+		ve->base.context_size = sibling->context_size;
+
+		ve->base.add_active_request = sibling->add_active_request;
+		ve->base.remove_active_request = sibling->remove_active_request;
+		ve->base.emit_bb_start = sibling->emit_bb_start;
+		ve->base.emit_flush = sibling->emit_flush;
+		ve->base.emit_init_breadcrumb = sibling->emit_init_breadcrumb;
+		ve->base.emit_fini_breadcrumb = sibling->emit_fini_breadcrumb;
+		ve->base.emit_fini_breadcrumb_dw =
+			sibling->emit_fini_breadcrumb_dw;
+
+		ve->base.flags = sibling->flags;
+	}
+
+	ve->base.flags |= I915_ENGINE_IS_VIRTUAL;
+
+	virtual_engine_initial_hint(ve);
+	return &ve->context;
+
+err_put:
+	intel_context_put(&ve->context);
+	return ERR_PTR(err);
+}
+
+void intel_execlists_show_requests(struct intel_engine_cs *engine,
+				   struct drm_printer *m,
+				   void (*show_request)(struct drm_printer *m,
+							const struct i915_request *rq,
+							const char *prefix,
+							int indent),
+				   unsigned int max)
+{
+	const struct intel_engine_execlists *execlists = &engine->execlists;
+	struct i915_sched_engine *sched_engine = engine->sched_engine;
+	struct i915_request *rq, *last;
+	unsigned long flags;
+	unsigned int count;
+	struct rb_node *rb;
+
+	spin_lock_irqsave(&sched_engine->lock, flags);
+
+	last = NULL;
+	count = 0;
+	list_for_each_entry(rq, &sched_engine->requests, sched.link) {
+		if (count++ < max - 1)
+			show_request(m, rq, "\t\t", 0);
+		else
+			last = rq;
+	}
+	if (last) {
+		if (count > max) {
+			drm_printf(m,
+				   "\t\t...skipping %d executing requests...\n",
+				   count - max);
+		}
+		show_request(m, last, "\t\t", 0);
+	}
+
+	if (sched_engine->queue_priority_hint != INT_MIN)
+		drm_printf(m, "\t\tQueue priority hint: %d\n",
+			   READ_ONCE(sched_engine->queue_priority_hint));
+
+	last = NULL;
+	count = 0;
+	for (rb = rb_first_cached(&sched_engine->queue); rb; rb = rb_next(rb)) {
+		struct i915_priolist *p = rb_entry(rb, typeof(*p), node);
+
+		priolist_for_each_request(rq, p) {
+			if (count++ < max - 1)
+				show_request(m, rq, "\t\t", 0);
+			else
+				last = rq;
+		}
+	}
+	if (last) {
+		if (count > max) {
+			drm_printf(m,
+				   "\t\t...skipping %d queued requests...\n",
+				   count - max);
+		}
+		show_request(m, last, "\t\t", 0);
+	}
+
+	last = NULL;
+	count = 0;
+	for (rb = rb_first_cached(&execlists->virtual); rb; rb = rb_next(rb)) {
+		struct virtual_engine *ve =
+			rb_entry(rb, typeof(*ve), nodes[engine->id].rb);
+		struct i915_request *rq = READ_ONCE(ve->request);
+
+		if (rq) {
+			if (count++ < max - 1)
+				show_request(m, rq, "\t\t", 0);
+			else
+				last = rq;
+		}
+	}
+	if (last) {
+		if (count > max) {
+			drm_printf(m,
+				   "\t\t...skipping %d virtual requests...\n",
+				   count - max);
+		}
+		show_request(m, last, "\t\t", 0);
+	}
+
+	spin_unlock_irqrestore(&sched_engine->lock, flags);
+}
+
+static unsigned long list_count(struct list_head *list)
+{
+	struct list_head *pos;
+	unsigned long count = 0;
+
+	list_for_each(pos, list)
+		count++;
+
+	return count;
+}
+
+void intel_execlists_dump_active_requests(struct intel_engine_cs *engine,
+					  struct i915_request *hung_rq,
+					  struct drm_printer *m)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&engine->sched_engine->lock, flags);
+
+	intel_engine_dump_active_requests(&engine->sched_engine->requests, hung_rq, m);
+
+	drm_printf(m, "\tOn hold?: %lu\n",
+		   list_count(&engine->sched_engine->hold));
+
+	spin_unlock_irqrestore(&engine->sched_engine->lock, flags);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_execlists.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.h b/drivers/gpu/drm/i915/gt/intel_execlists_submission.h
new file mode 100644
index 000000000..d2c7d45ea
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.h
@@ -0,0 +1,42 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#ifndef __INTEL_EXECLISTS_SUBMISSION_H__
+#define __INTEL_EXECLISTS_SUBMISSION_H__
+
+#include <linux/llist.h>
+#include <linux/types.h>
+
+struct drm_printer;
+
+struct i915_request;
+struct intel_context;
+struct intel_engine_cs;
+struct intel_gt;
+
+enum {
+	INTEL_CONTEXT_SCHEDULE_IN = 0,
+	INTEL_CONTEXT_SCHEDULE_OUT,
+	INTEL_CONTEXT_SCHEDULE_PREEMPTED,
+};
+
+int intel_execlists_submission_setup(struct intel_engine_cs *engine);
+
+void intel_execlists_show_requests(struct intel_engine_cs *engine,
+				   struct drm_printer *m,
+				   void (*show_request)(struct drm_printer *m,
+							const struct i915_request *rq,
+							const char *prefix,
+							int indent),
+				   unsigned int max);
+
+void intel_execlists_dump_active_requests(struct intel_engine_cs *engine,
+					  struct i915_request *hung_rq,
+					  struct drm_printer *m);
+
+bool
+intel_engine_in_execlists_submission_mode(const struct intel_engine_cs *engine);
+
+#endif /* __INTEL_EXECLISTS_SUBMISSION_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt.c b/drivers/gpu/drm/i915/gt/intel_ggtt.c
new file mode 100644
index 000000000..a6d0463b1
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_ggtt.c
@@ -0,0 +1,1330 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <asm/set_memory.h>
+#include <asm/smp.h>
+#include <linux/types.h>
+#include <linux/stop_machine.h>
+
+#include <drm/i915_drm.h>
+#include <drm/intel-gtt.h>
+
+#include "gem/i915_gem_lmem.h"
+
+#include "intel_ggtt_gmch.h"
+#include "intel_gt.h"
+#include "intel_gt_regs.h"
+#include "intel_pci_config.h"
+#include "i915_drv.h"
+#include "i915_pci.h"
+#include "i915_scatterlist.h"
+#include "i915_utils.h"
+#include "i915_vgpu.h"
+
+#include "intel_gtt.h"
+#include "gen8_ppgtt.h"
+
+static inline bool suspend_retains_ptes(struct i915_address_space *vm)
+{
+	return GRAPHICS_VER(vm->i915) >= 8 &&
+		!HAS_LMEM(vm->i915) &&
+		vm->is_ggtt;
+}
+
+static void i915_ggtt_color_adjust(const struct drm_mm_node *node,
+				   unsigned long color,
+				   u64 *start,
+				   u64 *end)
+{
+	if (i915_node_color_differs(node, color))
+		*start += I915_GTT_PAGE_SIZE;
+
+	/*
+	 * Also leave a space between the unallocated reserved node after the
+	 * GTT and any objects within the GTT, i.e. we use the color adjustment
+	 * to insert a guard page to prevent prefetches crossing over the
+	 * GTT boundary.
+	 */
+	node = list_next_entry(node, node_list);
+	if (node->color != color)
+		*end -= I915_GTT_PAGE_SIZE;
+}
+
+static int ggtt_init_hw(struct i915_ggtt *ggtt)
+{
+	struct drm_i915_private *i915 = ggtt->vm.i915;
+
+	i915_address_space_init(&ggtt->vm, VM_CLASS_GGTT);
+
+	ggtt->vm.is_ggtt = true;
+
+	/* Only VLV supports read-only GGTT mappings */
+	ggtt->vm.has_read_only = IS_VALLEYVIEW(i915);
+
+	if (!HAS_LLC(i915) && !HAS_PPGTT(i915))
+		ggtt->vm.mm.color_adjust = i915_ggtt_color_adjust;
+
+	if (ggtt->mappable_end) {
+		if (!io_mapping_init_wc(&ggtt->iomap,
+					ggtt->gmadr.start,
+					ggtt->mappable_end)) {
+			ggtt->vm.cleanup(&ggtt->vm);
+			return -EIO;
+		}
+
+		ggtt->mtrr = arch_phys_wc_add(ggtt->gmadr.start,
+					      ggtt->mappable_end);
+	}
+
+	intel_ggtt_init_fences(ggtt);
+
+	return 0;
+}
+
+/**
+ * i915_ggtt_init_hw - Initialize GGTT hardware
+ * @i915: i915 device
+ */
+int i915_ggtt_init_hw(struct drm_i915_private *i915)
+{
+	int ret;
+
+	/*
+	 * Note that we use page colouring to enforce a guard page at the
+	 * end of the address space. This is required as the CS may prefetch
+	 * beyond the end of the batch buffer, across the page boundary,
+	 * and beyond the end of the GTT if we do not provide a guard.
+	 */
+	ret = ggtt_init_hw(to_gt(i915)->ggtt);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+/*
+ * Return the value of the last GGTT pte cast to an u64, if
+ * the system is supposed to retain ptes across resume. 0 otherwise.
+ */
+static u64 read_last_pte(struct i915_address_space *vm)
+{
+	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
+	gen8_pte_t __iomem *ptep;
+
+	if (!suspend_retains_ptes(vm))
+		return 0;
+
+	GEM_BUG_ON(GRAPHICS_VER(vm->i915) < 8);
+	ptep = (typeof(ptep))ggtt->gsm + (ggtt_total_entries(ggtt) - 1);
+	return readq(ptep);
+}
+
+/**
+ * i915_ggtt_suspend_vm - Suspend the memory mappings for a GGTT or DPT VM
+ * @vm: The VM to suspend the mappings for
+ *
+ * Suspend the memory mappings for all objects mapped to HW via the GGTT or a
+ * DPT page table.
+ */
+void i915_ggtt_suspend_vm(struct i915_address_space *vm)
+{
+	struct i915_vma *vma, *vn;
+	int save_skip_rewrite;
+
+	drm_WARN_ON(&vm->i915->drm, !vm->is_ggtt && !vm->is_dpt);
+
+retry:
+	i915_gem_drain_freed_objects(vm->i915);
+
+	mutex_lock(&vm->mutex);
+
+	/*
+	 * Skip rewriting PTE on VMA unbind.
+	 * FIXME: Use an argument to i915_vma_unbind() instead?
+	 */
+	save_skip_rewrite = vm->skip_pte_rewrite;
+	vm->skip_pte_rewrite = true;
+
+	list_for_each_entry_safe(vma, vn, &vm->bound_list, vm_link) {
+		struct drm_i915_gem_object *obj = vma->obj;
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+
+		if (i915_vma_is_pinned(vma) || !i915_vma_is_bound(vma, I915_VMA_GLOBAL_BIND))
+			continue;
+
+		/* unlikely to race when GPU is idle, so no worry about slowpath.. */
+		if (WARN_ON(!i915_gem_object_trylock(obj, NULL))) {
+			/*
+			 * No dead objects should appear here, GPU should be
+			 * completely idle, and userspace suspended
+			 */
+			i915_gem_object_get(obj);
+
+			mutex_unlock(&vm->mutex);
+
+			i915_gem_object_lock(obj, NULL);
+			GEM_WARN_ON(i915_vma_unbind(vma));
+			i915_gem_object_unlock(obj);
+			i915_gem_object_put(obj);
+
+			vm->skip_pte_rewrite = save_skip_rewrite;
+			goto retry;
+		}
+
+		if (!i915_vma_is_bound(vma, I915_VMA_GLOBAL_BIND)) {
+			i915_vma_wait_for_bind(vma);
+
+			__i915_vma_evict(vma, false);
+			drm_mm_remove_node(&vma->node);
+		}
+
+		i915_gem_object_unlock(obj);
+	}
+
+	if (!suspend_retains_ptes(vm))
+		vm->clear_range(vm, 0, vm->total);
+	else
+		i915_vm_to_ggtt(vm)->probed_pte = read_last_pte(vm);
+
+	vm->skip_pte_rewrite = save_skip_rewrite;
+
+	mutex_unlock(&vm->mutex);
+}
+
+void i915_ggtt_suspend(struct i915_ggtt *ggtt)
+{
+	i915_ggtt_suspend_vm(&ggtt->vm);
+	ggtt->invalidate(ggtt);
+
+	intel_gt_check_and_clear_faults(ggtt->vm.gt);
+}
+
+void gen6_ggtt_invalidate(struct i915_ggtt *ggtt)
+{
+	struct intel_uncore *uncore = ggtt->vm.gt->uncore;
+
+	spin_lock_irq(&uncore->lock);
+	intel_uncore_write_fw(uncore, GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN);
+	intel_uncore_read_fw(uncore, GFX_FLSH_CNTL_GEN6);
+	spin_unlock_irq(&uncore->lock);
+}
+
+static void gen8_ggtt_invalidate(struct i915_ggtt *ggtt)
+{
+	struct intel_uncore *uncore = ggtt->vm.gt->uncore;
+
+	/*
+	 * Note that as an uncached mmio write, this will flush the
+	 * WCB of the writes into the GGTT before it triggers the invalidate.
+	 */
+	intel_uncore_write_fw(uncore, GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN);
+}
+
+static void guc_ggtt_invalidate(struct i915_ggtt *ggtt)
+{
+	struct intel_uncore *uncore = ggtt->vm.gt->uncore;
+	struct drm_i915_private *i915 = ggtt->vm.i915;
+
+	gen8_ggtt_invalidate(ggtt);
+
+	if (GRAPHICS_VER(i915) >= 12)
+		intel_uncore_write_fw(uncore, GEN12_GUC_TLB_INV_CR,
+				      GEN12_GUC_TLB_INV_CR_INVALIDATE);
+	else
+		intel_uncore_write_fw(uncore, GEN8_GTCR, GEN8_GTCR_INVALIDATE);
+}
+
+u64 gen8_ggtt_pte_encode(dma_addr_t addr,
+			 enum i915_cache_level level,
+			 u32 flags)
+{
+	gen8_pte_t pte = addr | GEN8_PAGE_PRESENT;
+
+	if (flags & PTE_LM)
+		pte |= GEN12_GGTT_PTE_LM;
+
+	return pte;
+}
+
+static void gen8_set_pte(void __iomem *addr, gen8_pte_t pte)
+{
+	writeq(pte, addr);
+}
+
+static void gen8_ggtt_insert_page(struct i915_address_space *vm,
+				  dma_addr_t addr,
+				  u64 offset,
+				  enum i915_cache_level level,
+				  u32 flags)
+{
+	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
+	gen8_pte_t __iomem *pte =
+		(gen8_pte_t __iomem *)ggtt->gsm + offset / I915_GTT_PAGE_SIZE;
+
+	gen8_set_pte(pte, gen8_ggtt_pte_encode(addr, level, flags));
+
+	ggtt->invalidate(ggtt);
+}
+
+static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
+				     struct i915_vma_resource *vma_res,
+				     enum i915_cache_level level,
+				     u32 flags)
+{
+	const gen8_pte_t pte_encode = gen8_ggtt_pte_encode(0, level, flags);
+	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
+	gen8_pte_t __iomem *gte;
+	gen8_pte_t __iomem *end;
+	struct sgt_iter iter;
+	dma_addr_t addr;
+
+	/*
+	 * Note that we ignore PTE_READ_ONLY here. The caller must be careful
+	 * not to allow the user to override access to a read only page.
+	 */
+
+	gte = (gen8_pte_t __iomem *)ggtt->gsm;
+	gte += vma_res->start / I915_GTT_PAGE_SIZE;
+	end = gte + vma_res->node_size / I915_GTT_PAGE_SIZE;
+
+	for_each_sgt_daddr(addr, iter, vma_res->bi.pages)
+		gen8_set_pte(gte++, pte_encode | addr);
+	GEM_BUG_ON(gte > end);
+
+	/* Fill the allocated but "unused" space beyond the end of the buffer */
+	while (gte < end)
+		gen8_set_pte(gte++, vm->scratch[0]->encode);
+
+	/*
+	 * We want to flush the TLBs only after we're certain all the PTE
+	 * updates have finished.
+	 */
+	ggtt->invalidate(ggtt);
+}
+
+static void gen6_ggtt_insert_page(struct i915_address_space *vm,
+				  dma_addr_t addr,
+				  u64 offset,
+				  enum i915_cache_level level,
+				  u32 flags)
+{
+	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
+	gen6_pte_t __iomem *pte =
+		(gen6_pte_t __iomem *)ggtt->gsm + offset / I915_GTT_PAGE_SIZE;
+
+	iowrite32(vm->pte_encode(addr, level, flags), pte);
+
+	ggtt->invalidate(ggtt);
+}
+
+/*
+ * Binds an object into the global gtt with the specified cache level.
+ * The object will be accessible to the GPU via commands whose operands
+ * reference offsets within the global GTT as well as accessible by the GPU
+ * through the GMADR mapped BAR (i915->mm.gtt->gtt).
+ */
+static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
+				     struct i915_vma_resource *vma_res,
+				     enum i915_cache_level level,
+				     u32 flags)
+{
+	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
+	gen6_pte_t __iomem *gte;
+	gen6_pte_t __iomem *end;
+	struct sgt_iter iter;
+	dma_addr_t addr;
+
+	gte = (gen6_pte_t __iomem *)ggtt->gsm;
+	gte += vma_res->start / I915_GTT_PAGE_SIZE;
+	end = gte + vma_res->node_size / I915_GTT_PAGE_SIZE;
+
+	for_each_sgt_daddr(addr, iter, vma_res->bi.pages)
+		iowrite32(vm->pte_encode(addr, level, flags), gte++);
+	GEM_BUG_ON(gte > end);
+
+	/* Fill the allocated but "unused" space beyond the end of the buffer */
+	while (gte < end)
+		iowrite32(vm->scratch[0]->encode, gte++);
+
+	/*
+	 * We want to flush the TLBs only after we're certain all the PTE
+	 * updates have finished.
+	 */
+	ggtt->invalidate(ggtt);
+}
+
+static void nop_clear_range(struct i915_address_space *vm,
+			    u64 start, u64 length)
+{
+}
+
+static void gen8_ggtt_clear_range(struct i915_address_space *vm,
+				  u64 start, u64 length)
+{
+	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
+	unsigned int first_entry = start / I915_GTT_PAGE_SIZE;
+	unsigned int num_entries = length / I915_GTT_PAGE_SIZE;
+	const gen8_pte_t scratch_pte = vm->scratch[0]->encode;
+	gen8_pte_t __iomem *gtt_base =
+		(gen8_pte_t __iomem *)ggtt->gsm + first_entry;
+	const int max_entries = ggtt_total_entries(ggtt) - first_entry;
+	int i;
+
+	if (WARN(num_entries > max_entries,
+		 "First entry = %d; Num entries = %d (max=%d)\n",
+		 first_entry, num_entries, max_entries))
+		num_entries = max_entries;
+
+	for (i = 0; i < num_entries; i++)
+		gen8_set_pte(&gtt_base[i], scratch_pte);
+}
+
+static void bxt_vtd_ggtt_wa(struct i915_address_space *vm)
+{
+	/*
+	 * Make sure the internal GAM fifo has been cleared of all GTT
+	 * writes before exiting stop_machine(). This guarantees that
+	 * any aperture accesses waiting to start in another process
+	 * cannot back up behind the GTT writes causing a hang.
+	 * The register can be any arbitrary GAM register.
+	 */
+	intel_uncore_posting_read_fw(vm->gt->uncore, GFX_FLSH_CNTL_GEN6);
+}
+
+struct insert_page {
+	struct i915_address_space *vm;
+	dma_addr_t addr;
+	u64 offset;
+	enum i915_cache_level level;
+};
+
+static int bxt_vtd_ggtt_insert_page__cb(void *_arg)
+{
+	struct insert_page *arg = _arg;
+
+	gen8_ggtt_insert_page(arg->vm, arg->addr, arg->offset, arg->level, 0);
+	bxt_vtd_ggtt_wa(arg->vm);
+
+	return 0;
+}
+
+static void bxt_vtd_ggtt_insert_page__BKL(struct i915_address_space *vm,
+					  dma_addr_t addr,
+					  u64 offset,
+					  enum i915_cache_level level,
+					  u32 unused)
+{
+	struct insert_page arg = { vm, addr, offset, level };
+
+	stop_machine(bxt_vtd_ggtt_insert_page__cb, &arg, NULL);
+}
+
+struct insert_entries {
+	struct i915_address_space *vm;
+	struct i915_vma_resource *vma_res;
+	enum i915_cache_level level;
+	u32 flags;
+};
+
+static int bxt_vtd_ggtt_insert_entries__cb(void *_arg)
+{
+	struct insert_entries *arg = _arg;
+
+	gen8_ggtt_insert_entries(arg->vm, arg->vma_res, arg->level, arg->flags);
+	bxt_vtd_ggtt_wa(arg->vm);
+
+	return 0;
+}
+
+static void bxt_vtd_ggtt_insert_entries__BKL(struct i915_address_space *vm,
+					     struct i915_vma_resource *vma_res,
+					     enum i915_cache_level level,
+					     u32 flags)
+{
+	struct insert_entries arg = { vm, vma_res, level, flags };
+
+	stop_machine(bxt_vtd_ggtt_insert_entries__cb, &arg, NULL);
+}
+
+static void gen6_ggtt_clear_range(struct i915_address_space *vm,
+				  u64 start, u64 length)
+{
+	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
+	unsigned int first_entry = start / I915_GTT_PAGE_SIZE;
+	unsigned int num_entries = length / I915_GTT_PAGE_SIZE;
+	gen6_pte_t scratch_pte, __iomem *gtt_base =
+		(gen6_pte_t __iomem *)ggtt->gsm + first_entry;
+	const int max_entries = ggtt_total_entries(ggtt) - first_entry;
+	int i;
+
+	if (WARN(num_entries > max_entries,
+		 "First entry = %d; Num entries = %d (max=%d)\n",
+		 first_entry, num_entries, max_entries))
+		num_entries = max_entries;
+
+	scratch_pte = vm->scratch[0]->encode;
+	for (i = 0; i < num_entries; i++)
+		iowrite32(scratch_pte, &gtt_base[i]);
+}
+
+void intel_ggtt_bind_vma(struct i915_address_space *vm,
+			 struct i915_vm_pt_stash *stash,
+			 struct i915_vma_resource *vma_res,
+			 enum i915_cache_level cache_level,
+			 u32 flags)
+{
+	u32 pte_flags;
+
+	if (vma_res->bound_flags & (~flags & I915_VMA_BIND_MASK))
+		return;
+
+	vma_res->bound_flags |= flags;
+
+	/* Applicable to VLV (gen8+ do not support RO in the GGTT) */
+	pte_flags = 0;
+	if (vma_res->bi.readonly)
+		pte_flags |= PTE_READ_ONLY;
+	if (vma_res->bi.lmem)
+		pte_flags |= PTE_LM;
+
+	vm->insert_entries(vm, vma_res, cache_level, pte_flags);
+	vma_res->page_sizes_gtt = I915_GTT_PAGE_SIZE;
+}
+
+void intel_ggtt_unbind_vma(struct i915_address_space *vm,
+			   struct i915_vma_resource *vma_res)
+{
+	vm->clear_range(vm, vma_res->start, vma_res->vma_size);
+}
+
+/*
+ * Reserve the top of the GuC address space for firmware images. Addresses
+ * beyond GUC_GGTT_TOP in the GuC address space are inaccessible by GuC,
+ * which makes for a suitable range to hold GuC/HuC firmware images if the
+ * size of the GGTT is 4G. However, on a 32-bit platform the size of the GGTT
+ * is limited to 2G, which is less than GUC_GGTT_TOP, but we reserve a chunk
+ * of the same size anyway, which is far more than needed, to keep the logic
+ * in uc_fw_ggtt_offset() simple.
+ */
+#define GUC_TOP_RESERVE_SIZE (SZ_4G - GUC_GGTT_TOP)
+
+static int ggtt_reserve_guc_top(struct i915_ggtt *ggtt)
+{
+	u64 offset;
+	int ret;
+
+	if (!intel_uc_uses_guc(&ggtt->vm.gt->uc))
+		return 0;
+
+	GEM_BUG_ON(ggtt->vm.total <= GUC_TOP_RESERVE_SIZE);
+	offset = ggtt->vm.total - GUC_TOP_RESERVE_SIZE;
+
+	ret = i915_gem_gtt_reserve(&ggtt->vm, NULL, &ggtt->uc_fw,
+				   GUC_TOP_RESERVE_SIZE, offset,
+				   I915_COLOR_UNEVICTABLE, PIN_NOEVICT);
+	if (ret)
+		drm_dbg(&ggtt->vm.i915->drm,
+			"Failed to reserve top of GGTT for GuC\n");
+
+	return ret;
+}
+
+static void ggtt_release_guc_top(struct i915_ggtt *ggtt)
+{
+	if (drm_mm_node_allocated(&ggtt->uc_fw))
+		drm_mm_remove_node(&ggtt->uc_fw);
+}
+
+static void cleanup_init_ggtt(struct i915_ggtt *ggtt)
+{
+	ggtt_release_guc_top(ggtt);
+	if (drm_mm_node_allocated(&ggtt->error_capture))
+		drm_mm_remove_node(&ggtt->error_capture);
+	mutex_destroy(&ggtt->error_mutex);
+}
+
+static int init_ggtt(struct i915_ggtt *ggtt)
+{
+	/*
+	 * Let GEM Manage all of the aperture.
+	 *
+	 * However, leave one page at the end still bound to the scratch page.
+	 * There are a number of places where the hardware apparently prefetches
+	 * past the end of the object, and we've seen multiple hangs with the
+	 * GPU head pointer stuck in a batchbuffer bound at the last page of the
+	 * aperture.  One page should be enough to keep any prefetching inside
+	 * of the aperture.
+	 */
+	unsigned long hole_start, hole_end;
+	struct drm_mm_node *entry;
+	int ret;
+
+	ggtt->pte_lost = true;
+
+	/*
+	 * GuC requires all resources that we're sharing with it to be placed in
+	 * non-WOPCM memory. If GuC is not present or not in use we still need a
+	 * small bias as ring wraparound at offset 0 sometimes hangs. No idea
+	 * why.
+	 */
+	ggtt->pin_bias = max_t(u32, I915_GTT_PAGE_SIZE,
+			       intel_wopcm_guc_size(&ggtt->vm.i915->wopcm));
+
+	ret = intel_vgt_balloon(ggtt);
+	if (ret)
+		return ret;
+
+	mutex_init(&ggtt->error_mutex);
+	if (ggtt->mappable_end) {
+		/*
+		 * Reserve a mappable slot for our lockless error capture.
+		 *
+		 * We strongly prefer taking address 0x0 in order to protect
+		 * other critical buffers against accidental overwrites,
+		 * as writing to address 0 is a very common mistake.
+		 *
+		 * Since 0 may already be in use by the system (e.g. the BIOS
+		 * framebuffer), we let the reservation fail quietly and hope
+		 * 0 remains reserved always.
+		 *
+		 * If we fail to reserve 0, and then fail to find any space
+		 * for an error-capture, remain silent. We can afford not
+		 * to reserve an error_capture node as we have fallback
+		 * paths, and we trust that 0 will remain reserved. However,
+		 * the only likely reason for failure to insert is a driver
+		 * bug, which we expect to cause other failures...
+		 */
+		ggtt->error_capture.size = I915_GTT_PAGE_SIZE;
+		ggtt->error_capture.color = I915_COLOR_UNEVICTABLE;
+		if (drm_mm_reserve_node(&ggtt->vm.mm, &ggtt->error_capture))
+			drm_mm_insert_node_in_range(&ggtt->vm.mm,
+						    &ggtt->error_capture,
+						    ggtt->error_capture.size, 0,
+						    ggtt->error_capture.color,
+						    0, ggtt->mappable_end,
+						    DRM_MM_INSERT_LOW);
+	}
+	if (drm_mm_node_allocated(&ggtt->error_capture))
+		drm_dbg(&ggtt->vm.i915->drm,
+			"Reserved GGTT:[%llx, %llx] for use by error capture\n",
+			ggtt->error_capture.start,
+			ggtt->error_capture.start + ggtt->error_capture.size);
+
+	/*
+	 * The upper portion of the GuC address space has a sizeable hole
+	 * (several MB) that is inaccessible by GuC. Reserve this range within
+	 * GGTT as it can comfortably hold GuC/HuC firmware images.
+	 */
+	ret = ggtt_reserve_guc_top(ggtt);
+	if (ret)
+		goto err;
+
+	/* Clear any non-preallocated blocks */
+	drm_mm_for_each_hole(entry, &ggtt->vm.mm, hole_start, hole_end) {
+		drm_dbg(&ggtt->vm.i915->drm,
+			"clearing unused GTT space: [%lx, %lx]\n",
+			hole_start, hole_end);
+		ggtt->vm.clear_range(&ggtt->vm, hole_start,
+				     hole_end - hole_start);
+	}
+
+	/* And finally clear the reserved guard page */
+	ggtt->vm.clear_range(&ggtt->vm, ggtt->vm.total - PAGE_SIZE, PAGE_SIZE);
+
+	return 0;
+
+err:
+	cleanup_init_ggtt(ggtt);
+	return ret;
+}
+
+static void aliasing_gtt_bind_vma(struct i915_address_space *vm,
+				  struct i915_vm_pt_stash *stash,
+				  struct i915_vma_resource *vma_res,
+				  enum i915_cache_level cache_level,
+				  u32 flags)
+{
+	u32 pte_flags;
+
+	/* Currently applicable only to VLV */
+	pte_flags = 0;
+	if (vma_res->bi.readonly)
+		pte_flags |= PTE_READ_ONLY;
+
+	if (flags & I915_VMA_LOCAL_BIND)
+		ppgtt_bind_vma(&i915_vm_to_ggtt(vm)->alias->vm,
+			       stash, vma_res, cache_level, flags);
+
+	if (flags & I915_VMA_GLOBAL_BIND)
+		vm->insert_entries(vm, vma_res, cache_level, pte_flags);
+
+	vma_res->bound_flags |= flags;
+}
+
+static void aliasing_gtt_unbind_vma(struct i915_address_space *vm,
+				    struct i915_vma_resource *vma_res)
+{
+	if (vma_res->bound_flags & I915_VMA_GLOBAL_BIND)
+		vm->clear_range(vm, vma_res->start, vma_res->vma_size);
+
+	if (vma_res->bound_flags & I915_VMA_LOCAL_BIND)
+		ppgtt_unbind_vma(&i915_vm_to_ggtt(vm)->alias->vm, vma_res);
+}
+
+static int init_aliasing_ppgtt(struct i915_ggtt *ggtt)
+{
+	struct i915_vm_pt_stash stash = {};
+	struct i915_ppgtt *ppgtt;
+	int err;
+
+	ppgtt = i915_ppgtt_create(ggtt->vm.gt, 0);
+	if (IS_ERR(ppgtt))
+		return PTR_ERR(ppgtt);
+
+	if (GEM_WARN_ON(ppgtt->vm.total < ggtt->vm.total)) {
+		err = -ENODEV;
+		goto err_ppgtt;
+	}
+
+	err = i915_vm_alloc_pt_stash(&ppgtt->vm, &stash, ggtt->vm.total);
+	if (err)
+		goto err_ppgtt;
+
+	i915_gem_object_lock(ppgtt->vm.scratch[0], NULL);
+	err = i915_vm_map_pt_stash(&ppgtt->vm, &stash);
+	i915_gem_object_unlock(ppgtt->vm.scratch[0]);
+	if (err)
+		goto err_stash;
+
+	/*
+	 * Note we only pre-allocate as far as the end of the global
+	 * GTT. On 48b / 4-level page-tables, the difference is very,
+	 * very significant! We have to preallocate as GVT/vgpu does
+	 * not like the page directory disappearing.
+	 */
+	ppgtt->vm.allocate_va_range(&ppgtt->vm, &stash, 0, ggtt->vm.total);
+
+	ggtt->alias = ppgtt;
+	ggtt->vm.bind_async_flags |= ppgtt->vm.bind_async_flags;
+
+	GEM_BUG_ON(ggtt->vm.vma_ops.bind_vma != intel_ggtt_bind_vma);
+	ggtt->vm.vma_ops.bind_vma = aliasing_gtt_bind_vma;
+
+	GEM_BUG_ON(ggtt->vm.vma_ops.unbind_vma != intel_ggtt_unbind_vma);
+	ggtt->vm.vma_ops.unbind_vma = aliasing_gtt_unbind_vma;
+
+	i915_vm_free_pt_stash(&ppgtt->vm, &stash);
+	return 0;
+
+err_stash:
+	i915_vm_free_pt_stash(&ppgtt->vm, &stash);
+err_ppgtt:
+	i915_vm_put(&ppgtt->vm);
+	return err;
+}
+
+static void fini_aliasing_ppgtt(struct i915_ggtt *ggtt)
+{
+	struct i915_ppgtt *ppgtt;
+
+	ppgtt = fetch_and_zero(&ggtt->alias);
+	if (!ppgtt)
+		return;
+
+	i915_vm_put(&ppgtt->vm);
+
+	ggtt->vm.vma_ops.bind_vma   = intel_ggtt_bind_vma;
+	ggtt->vm.vma_ops.unbind_vma = intel_ggtt_unbind_vma;
+}
+
+int i915_init_ggtt(struct drm_i915_private *i915)
+{
+	int ret;
+
+	ret = init_ggtt(to_gt(i915)->ggtt);
+	if (ret)
+		return ret;
+
+	if (INTEL_PPGTT(i915) == INTEL_PPGTT_ALIASING) {
+		ret = init_aliasing_ppgtt(to_gt(i915)->ggtt);
+		if (ret)
+			cleanup_init_ggtt(to_gt(i915)->ggtt);
+	}
+
+	return 0;
+}
+
+static void ggtt_cleanup_hw(struct i915_ggtt *ggtt)
+{
+	struct i915_vma *vma, *vn;
+
+	flush_workqueue(ggtt->vm.i915->wq);
+	i915_gem_drain_freed_objects(ggtt->vm.i915);
+
+	mutex_lock(&ggtt->vm.mutex);
+
+	ggtt->vm.skip_pte_rewrite = true;
+
+	list_for_each_entry_safe(vma, vn, &ggtt->vm.bound_list, vm_link) {
+		struct drm_i915_gem_object *obj = vma->obj;
+		bool trylock;
+
+		trylock = i915_gem_object_trylock(obj, NULL);
+		WARN_ON(!trylock);
+
+		WARN_ON(__i915_vma_unbind(vma));
+		if (trylock)
+			i915_gem_object_unlock(obj);
+	}
+
+	if (drm_mm_node_allocated(&ggtt->error_capture))
+		drm_mm_remove_node(&ggtt->error_capture);
+	mutex_destroy(&ggtt->error_mutex);
+
+	ggtt_release_guc_top(ggtt);
+	intel_vgt_deballoon(ggtt);
+
+	ggtt->vm.cleanup(&ggtt->vm);
+
+	mutex_unlock(&ggtt->vm.mutex);
+	i915_address_space_fini(&ggtt->vm);
+
+	arch_phys_wc_del(ggtt->mtrr);
+
+	if (ggtt->iomap.size)
+		io_mapping_fini(&ggtt->iomap);
+}
+
+/**
+ * i915_ggtt_driver_release - Clean up GGTT hardware initialization
+ * @i915: i915 device
+ */
+void i915_ggtt_driver_release(struct drm_i915_private *i915)
+{
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+
+	fini_aliasing_ppgtt(ggtt);
+
+	intel_ggtt_fini_fences(ggtt);
+	ggtt_cleanup_hw(ggtt);
+}
+
+/**
+ * i915_ggtt_driver_late_release - Cleanup of GGTT that needs to be done after
+ * all free objects have been drained.
+ * @i915: i915 device
+ */
+void i915_ggtt_driver_late_release(struct drm_i915_private *i915)
+{
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+
+	GEM_WARN_ON(kref_read(&ggtt->vm.resv_ref) != 1);
+	dma_resv_fini(&ggtt->vm._resv);
+}
+
+static unsigned int gen6_get_total_gtt_size(u16 snb_gmch_ctl)
+{
+	snb_gmch_ctl >>= SNB_GMCH_GGMS_SHIFT;
+	snb_gmch_ctl &= SNB_GMCH_GGMS_MASK;
+	return snb_gmch_ctl << 20;
+}
+
+static unsigned int gen8_get_total_gtt_size(u16 bdw_gmch_ctl)
+{
+	bdw_gmch_ctl >>= BDW_GMCH_GGMS_SHIFT;
+	bdw_gmch_ctl &= BDW_GMCH_GGMS_MASK;
+	if (bdw_gmch_ctl)
+		bdw_gmch_ctl = 1 << bdw_gmch_ctl;
+
+#ifdef CONFIG_X86_32
+	/* Limit 32b platforms to a 2GB GGTT: 4 << 20 / pte size * I915_GTT_PAGE_SIZE */
+	if (bdw_gmch_ctl > 4)
+		bdw_gmch_ctl = 4;
+#endif
+
+	return bdw_gmch_ctl << 20;
+}
+
+static unsigned int chv_get_total_gtt_size(u16 gmch_ctrl)
+{
+	gmch_ctrl >>= SNB_GMCH_GGMS_SHIFT;
+	gmch_ctrl &= SNB_GMCH_GGMS_MASK;
+
+	if (gmch_ctrl)
+		return 1 << (20 + gmch_ctrl);
+
+	return 0;
+}
+
+static unsigned int gen6_gttmmadr_size(struct drm_i915_private *i915)
+{
+	/*
+	 * GEN6: GTTMMADR size is 4MB and GTTADR starts at 2MB offset
+	 * GEN8: GTTMMADR size is 16MB and GTTADR starts at 8MB offset
+	 */
+	GEM_BUG_ON(GRAPHICS_VER(i915) < 6);
+	return (GRAPHICS_VER(i915) < 8) ? SZ_4M : SZ_16M;
+}
+
+static unsigned int gen6_gttadr_offset(struct drm_i915_private *i915)
+{
+	return gen6_gttmmadr_size(i915) / 2;
+}
+
+static int ggtt_probe_common(struct i915_ggtt *ggtt, u64 size)
+{
+	struct drm_i915_private *i915 = ggtt->vm.i915;
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	phys_addr_t phys_addr;
+	u32 pte_flags;
+	int ret;
+
+	GEM_WARN_ON(pci_resource_len(pdev, GTTMMADR_BAR) != gen6_gttmmadr_size(i915));
+	phys_addr = pci_resource_start(pdev, GTTMMADR_BAR) + gen6_gttadr_offset(i915);
+
+	/*
+	 * On BXT+/ICL+ writes larger than 64 bit to the GTT pagetable range
+	 * will be dropped. For WC mappings in general we have 64 byte burst
+	 * writes when the WC buffer is flushed, so we can't use it, but have to
+	 * resort to an uncached mapping. The WC issue is easily caught by the
+	 * readback check when writing GTT PTE entries.
+	 */
+	if (IS_GEN9_LP(i915) || GRAPHICS_VER(i915) >= 11)
+		ggtt->gsm = ioremap(phys_addr, size);
+	else
+		ggtt->gsm = ioremap_wc(phys_addr, size);
+	if (!ggtt->gsm) {
+		drm_err(&i915->drm, "Failed to map the ggtt page table\n");
+		return -ENOMEM;
+	}
+
+	kref_init(&ggtt->vm.resv_ref);
+	ret = setup_scratch_page(&ggtt->vm);
+	if (ret) {
+		drm_err(&i915->drm, "Scratch setup failed\n");
+		/* iounmap will also get called at remove, but meh */
+		iounmap(ggtt->gsm);
+		return ret;
+	}
+
+	pte_flags = 0;
+	if (i915_gem_object_is_lmem(ggtt->vm.scratch[0]))
+		pte_flags |= PTE_LM;
+
+	ggtt->vm.scratch[0]->encode =
+		ggtt->vm.pte_encode(px_dma(ggtt->vm.scratch[0]),
+				    I915_CACHE_NONE, pte_flags);
+
+	return 0;
+}
+
+static void gen6_gmch_remove(struct i915_address_space *vm)
+{
+	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
+
+	iounmap(ggtt->gsm);
+	free_scratch(vm);
+}
+
+static struct resource pci_resource(struct pci_dev *pdev, int bar)
+{
+	return (struct resource)DEFINE_RES_MEM(pci_resource_start(pdev, bar),
+					       pci_resource_len(pdev, bar));
+}
+
+static int gen8_gmch_probe(struct i915_ggtt *ggtt)
+{
+	struct drm_i915_private *i915 = ggtt->vm.i915;
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	unsigned int size;
+	u16 snb_gmch_ctl;
+
+	if (!HAS_LMEM(i915)) {
+		if (!i915_pci_resource_valid(pdev, GTT_APERTURE_BAR))
+			return -ENXIO;
+
+		ggtt->gmadr = pci_resource(pdev, GTT_APERTURE_BAR);
+		ggtt->mappable_end = resource_size(&ggtt->gmadr);
+	}
+
+	pci_read_config_word(pdev, SNB_GMCH_CTRL, &snb_gmch_ctl);
+	if (IS_CHERRYVIEW(i915))
+		size = chv_get_total_gtt_size(snb_gmch_ctl);
+	else
+		size = gen8_get_total_gtt_size(snb_gmch_ctl);
+
+	ggtt->vm.alloc_pt_dma = alloc_pt_dma;
+	ggtt->vm.alloc_scratch_dma = alloc_pt_dma;
+	ggtt->vm.lmem_pt_obj_flags = I915_BO_ALLOC_PM_EARLY;
+
+	ggtt->vm.total = (size / sizeof(gen8_pte_t)) * I915_GTT_PAGE_SIZE;
+	ggtt->vm.cleanup = gen6_gmch_remove;
+	ggtt->vm.insert_page = gen8_ggtt_insert_page;
+	ggtt->vm.clear_range = nop_clear_range;
+	if (intel_scanout_needs_vtd_wa(i915))
+		ggtt->vm.clear_range = gen8_ggtt_clear_range;
+
+	ggtt->vm.insert_entries = gen8_ggtt_insert_entries;
+
+	/*
+	 * Serialize GTT updates with aperture access on BXT if VT-d is on,
+	 * and always on CHV.
+	 */
+	if (intel_vm_no_concurrent_access_wa(i915)) {
+		ggtt->vm.insert_entries = bxt_vtd_ggtt_insert_entries__BKL;
+		ggtt->vm.insert_page    = bxt_vtd_ggtt_insert_page__BKL;
+
+		/*
+		 * Calling stop_machine() version of GGTT update function
+		 * at error capture/reset path will raise lockdep warning.
+		 * Allow calling gen8_ggtt_insert_* directly at reset path
+		 * which is safe from parallel GGTT updates.
+		 */
+		ggtt->vm.raw_insert_page = gen8_ggtt_insert_page;
+		ggtt->vm.raw_insert_entries = gen8_ggtt_insert_entries;
+
+		ggtt->vm.bind_async_flags =
+			I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND;
+	}
+
+	ggtt->invalidate = gen8_ggtt_invalidate;
+
+	ggtt->vm.vma_ops.bind_vma    = intel_ggtt_bind_vma;
+	ggtt->vm.vma_ops.unbind_vma  = intel_ggtt_unbind_vma;
+
+	ggtt->vm.pte_encode = gen8_ggtt_pte_encode;
+
+	setup_private_pat(ggtt->vm.gt->uncore);
+
+	return ggtt_probe_common(ggtt, size);
+}
+
+static u64 snb_pte_encode(dma_addr_t addr,
+			  enum i915_cache_level level,
+			  u32 flags)
+{
+	gen6_pte_t pte = GEN6_PTE_ADDR_ENCODE(addr) | GEN6_PTE_VALID;
+
+	switch (level) {
+	case I915_CACHE_L3_LLC:
+	case I915_CACHE_LLC:
+		pte |= GEN6_PTE_CACHE_LLC;
+		break;
+	case I915_CACHE_NONE:
+		pte |= GEN6_PTE_UNCACHED;
+		break;
+	default:
+		MISSING_CASE(level);
+	}
+
+	return pte;
+}
+
+static u64 ivb_pte_encode(dma_addr_t addr,
+			  enum i915_cache_level level,
+			  u32 flags)
+{
+	gen6_pte_t pte = GEN6_PTE_ADDR_ENCODE(addr) | GEN6_PTE_VALID;
+
+	switch (level) {
+	case I915_CACHE_L3_LLC:
+		pte |= GEN7_PTE_CACHE_L3_LLC;
+		break;
+	case I915_CACHE_LLC:
+		pte |= GEN6_PTE_CACHE_LLC;
+		break;
+	case I915_CACHE_NONE:
+		pte |= GEN6_PTE_UNCACHED;
+		break;
+	default:
+		MISSING_CASE(level);
+	}
+
+	return pte;
+}
+
+static u64 byt_pte_encode(dma_addr_t addr,
+			  enum i915_cache_level level,
+			  u32 flags)
+{
+	gen6_pte_t pte = GEN6_PTE_ADDR_ENCODE(addr) | GEN6_PTE_VALID;
+
+	if (!(flags & PTE_READ_ONLY))
+		pte |= BYT_PTE_WRITEABLE;
+
+	if (level != I915_CACHE_NONE)
+		pte |= BYT_PTE_SNOOPED_BY_CPU_CACHES;
+
+	return pte;
+}
+
+static u64 hsw_pte_encode(dma_addr_t addr,
+			  enum i915_cache_level level,
+			  u32 flags)
+{
+	gen6_pte_t pte = HSW_PTE_ADDR_ENCODE(addr) | GEN6_PTE_VALID;
+
+	if (level != I915_CACHE_NONE)
+		pte |= HSW_WB_LLC_AGE3;
+
+	return pte;
+}
+
+static u64 iris_pte_encode(dma_addr_t addr,
+			   enum i915_cache_level level,
+			   u32 flags)
+{
+	gen6_pte_t pte = HSW_PTE_ADDR_ENCODE(addr) | GEN6_PTE_VALID;
+
+	switch (level) {
+	case I915_CACHE_NONE:
+		break;
+	case I915_CACHE_WT:
+		pte |= HSW_WT_ELLC_LLC_AGE3;
+		break;
+	default:
+		pte |= HSW_WB_ELLC_LLC_AGE3;
+		break;
+	}
+
+	return pte;
+}
+
+static int gen6_gmch_probe(struct i915_ggtt *ggtt)
+{
+	struct drm_i915_private *i915 = ggtt->vm.i915;
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	unsigned int size;
+	u16 snb_gmch_ctl;
+
+	if (!i915_pci_resource_valid(pdev, GTT_APERTURE_BAR))
+		return -ENXIO;
+
+	ggtt->gmadr = pci_resource(pdev, GTT_APERTURE_BAR);
+	ggtt->mappable_end = resource_size(&ggtt->gmadr);
+
+	/*
+	 * 64/512MB is the current min/max we actually know of, but this is
+	 * just a coarse sanity check.
+	 */
+	if (ggtt->mappable_end < (64 << 20) ||
+	    ggtt->mappable_end > (512 << 20)) {
+		drm_err(&i915->drm, "Unknown GMADR size (%pa)\n",
+			&ggtt->mappable_end);
+		return -ENXIO;
+	}
+
+	pci_read_config_word(pdev, SNB_GMCH_CTRL, &snb_gmch_ctl);
+
+	size = gen6_get_total_gtt_size(snb_gmch_ctl);
+	ggtt->vm.total = (size / sizeof(gen6_pte_t)) * I915_GTT_PAGE_SIZE;
+
+	ggtt->vm.alloc_pt_dma = alloc_pt_dma;
+	ggtt->vm.alloc_scratch_dma = alloc_pt_dma;
+
+	ggtt->vm.clear_range = nop_clear_range;
+	if (!HAS_FULL_PPGTT(i915) || intel_scanout_needs_vtd_wa(i915))
+		ggtt->vm.clear_range = gen6_ggtt_clear_range;
+	ggtt->vm.insert_page = gen6_ggtt_insert_page;
+	ggtt->vm.insert_entries = gen6_ggtt_insert_entries;
+	ggtt->vm.cleanup = gen6_gmch_remove;
+
+	ggtt->invalidate = gen6_ggtt_invalidate;
+
+	if (HAS_EDRAM(i915))
+		ggtt->vm.pte_encode = iris_pte_encode;
+	else if (IS_HASWELL(i915))
+		ggtt->vm.pte_encode = hsw_pte_encode;
+	else if (IS_VALLEYVIEW(i915))
+		ggtt->vm.pte_encode = byt_pte_encode;
+	else if (GRAPHICS_VER(i915) >= 7)
+		ggtt->vm.pte_encode = ivb_pte_encode;
+	else
+		ggtt->vm.pte_encode = snb_pte_encode;
+
+	ggtt->vm.vma_ops.bind_vma    = intel_ggtt_bind_vma;
+	ggtt->vm.vma_ops.unbind_vma  = intel_ggtt_unbind_vma;
+
+	return ggtt_probe_common(ggtt, size);
+}
+
+static int ggtt_probe_hw(struct i915_ggtt *ggtt, struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	int ret;
+
+	ggtt->vm.gt = gt;
+	ggtt->vm.i915 = i915;
+	ggtt->vm.dma = i915->drm.dev;
+	dma_resv_init(&ggtt->vm._resv);
+
+	if (GRAPHICS_VER(i915) >= 8)
+		ret = gen8_gmch_probe(ggtt);
+	else if (GRAPHICS_VER(i915) >= 6)
+		ret = gen6_gmch_probe(ggtt);
+	else
+		ret = intel_ggtt_gmch_probe(ggtt);
+
+	if (ret) {
+		dma_resv_fini(&ggtt->vm._resv);
+		return ret;
+	}
+
+	if ((ggtt->vm.total - 1) >> 32) {
+		drm_err(&i915->drm,
+			"We never expected a Global GTT with more than 32bits"
+			" of address space! Found %lldM!\n",
+			ggtt->vm.total >> 20);
+		ggtt->vm.total = 1ULL << 32;
+		ggtt->mappable_end =
+			min_t(u64, ggtt->mappable_end, ggtt->vm.total);
+	}
+
+	if (ggtt->mappable_end > ggtt->vm.total) {
+		drm_err(&i915->drm,
+			"mappable aperture extends past end of GGTT,"
+			" aperture=%pa, total=%llx\n",
+			&ggtt->mappable_end, ggtt->vm.total);
+		ggtt->mappable_end = ggtt->vm.total;
+	}
+
+	/* GMADR is the PCI mmio aperture into the global GTT. */
+	drm_dbg(&i915->drm, "GGTT size = %lluM\n", ggtt->vm.total >> 20);
+	drm_dbg(&i915->drm, "GMADR size = %lluM\n",
+		(u64)ggtt->mappable_end >> 20);
+	drm_dbg(&i915->drm, "DSM size = %lluM\n",
+		(u64)resource_size(&intel_graphics_stolen_res) >> 20);
+
+	return 0;
+}
+
+/**
+ * i915_ggtt_probe_hw - Probe GGTT hardware location
+ * @i915: i915 device
+ */
+int i915_ggtt_probe_hw(struct drm_i915_private *i915)
+{
+	int ret;
+
+	ret = ggtt_probe_hw(to_gt(i915)->ggtt, to_gt(i915));
+	if (ret)
+		return ret;
+
+	if (i915_vtd_active(i915))
+		drm_info(&i915->drm, "VT-d active for gfx access\n");
+
+	return 0;
+}
+
+int i915_ggtt_enable_hw(struct drm_i915_private *i915)
+{
+	if (GRAPHICS_VER(i915) < 6)
+		return intel_ggtt_gmch_enable_hw(i915);
+
+	return 0;
+}
+
+void i915_ggtt_enable_guc(struct i915_ggtt *ggtt)
+{
+	GEM_BUG_ON(ggtt->invalidate != gen8_ggtt_invalidate);
+
+	ggtt->invalidate = guc_ggtt_invalidate;
+
+	ggtt->invalidate(ggtt);
+}
+
+void i915_ggtt_disable_guc(struct i915_ggtt *ggtt)
+{
+	/* XXX Temporary pardon for error unload */
+	if (ggtt->invalidate == gen8_ggtt_invalidate)
+		return;
+
+	/* We should only be called after i915_ggtt_enable_guc() */
+	GEM_BUG_ON(ggtt->invalidate != guc_ggtt_invalidate);
+
+	ggtt->invalidate = gen8_ggtt_invalidate;
+
+	ggtt->invalidate(ggtt);
+}
+
+/**
+ * i915_ggtt_resume_vm - Restore the memory mappings for a GGTT or DPT VM
+ * @vm: The VM to restore the mappings for
+ *
+ * Restore the memory mappings for all objects mapped to HW via the GGTT or a
+ * DPT page table.
+ *
+ * Returns %true if restoring the mapping for any object that was in a write
+ * domain before suspend.
+ */
+bool i915_ggtt_resume_vm(struct i915_address_space *vm)
+{
+	struct i915_vma *vma;
+	bool write_domain_objs = false;
+	bool retained_ptes;
+
+	drm_WARN_ON(&vm->i915->drm, !vm->is_ggtt && !vm->is_dpt);
+
+	/*
+	 * First fill our portion of the GTT with scratch pages if
+	 * they were not retained across suspend.
+	 */
+	retained_ptes = suspend_retains_ptes(vm) &&
+		!i915_vm_to_ggtt(vm)->pte_lost &&
+		!GEM_WARN_ON(i915_vm_to_ggtt(vm)->probed_pte != read_last_pte(vm));
+
+	if (!retained_ptes)
+		vm->clear_range(vm, 0, vm->total);
+
+	/* clflush objects bound into the GGTT and rebind them. */
+	list_for_each_entry(vma, &vm->bound_list, vm_link) {
+		struct drm_i915_gem_object *obj = vma->obj;
+		unsigned int was_bound =
+			atomic_read(&vma->flags) & I915_VMA_BIND_MASK;
+
+		GEM_BUG_ON(!was_bound);
+		if (!retained_ptes) {
+			/*
+			 * Clear the bound flags of the vma resource to allow
+			 * ptes to be repopulated.
+			 */
+			vma->resource->bound_flags = 0;
+			vma->ops->bind_vma(vm, NULL, vma->resource,
+					   obj ? obj->cache_level : 0,
+					   was_bound);
+		}
+		if (obj) { /* only used during resume => exclusive access */
+			write_domain_objs |= fetch_and_zero(&obj->write_domain);
+			obj->read_domains |= I915_GEM_DOMAIN_GTT;
+		}
+	}
+
+	return write_domain_objs;
+}
+
+void i915_ggtt_resume(struct i915_ggtt *ggtt)
+{
+	bool flush;
+
+	intel_gt_check_and_clear_faults(ggtt->vm.gt);
+
+	flush = i915_ggtt_resume_vm(&ggtt->vm);
+
+	ggtt->invalidate(ggtt);
+
+	if (flush)
+		wbinvd_on_all_cpus();
+
+	if (GRAPHICS_VER(ggtt->vm.i915) >= 8)
+		setup_private_pat(ggtt->vm.gt->uncore);
+
+	intel_ggtt_restore_fences(ggtt);
+}
+
+void i915_ggtt_mark_pte_lost(struct drm_i915_private *i915, bool val)
+{
+	to_gt(i915)->ggtt->pte_lost = val;
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c b/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c
new file mode 100644
index 000000000..ea775e601
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c
@@ -0,0 +1,926 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2008-2015 Intel Corporation
+ */
+
+#include <linux/highmem.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "i915_scatterlist.h"
+#include "i915_pvinfo.h"
+#include "i915_vgpu.h"
+#include "intel_gt_regs.h"
+#include "intel_mchbar_regs.h"
+
+/**
+ * DOC: fence register handling
+ *
+ * Important to avoid confusions: "fences" in the i915 driver are not execution
+ * fences used to track command completion but hardware detiler objects which
+ * wrap a given range of the global GTT. Each platform has only a fairly limited
+ * set of these objects.
+ *
+ * Fences are used to detile GTT memory mappings. They're also connected to the
+ * hardware frontbuffer render tracking and hence interact with frontbuffer
+ * compression. Furthermore on older platforms fences are required for tiled
+ * objects used by the display engine. They can also be used by the render
+ * engine - they're required for blitter commands and are optional for render
+ * commands. But on gen4+ both display (with the exception of fbc) and rendering
+ * have their own tiling state bits and don't need fences.
+ *
+ * Also note that fences only support X and Y tiling and hence can't be used for
+ * the fancier new tiling formats like W, Ys and Yf.
+ *
+ * Finally note that because fences are such a restricted resource they're
+ * dynamically associated with objects. Furthermore fence state is committed to
+ * the hardware lazily to avoid unnecessary stalls on gen2/3. Therefore code must
+ * explicitly call i915_gem_object_get_fence() to synchronize fencing status
+ * for cpu access. Also note that some code wants an unfenced view, for those
+ * cases the fence can be removed forcefully with i915_gem_object_put_fence().
+ *
+ * Internally these functions will synchronize with userspace access by removing
+ * CPU ptes into GTT mmaps (not the GTT ptes themselves) as needed.
+ */
+
+#define pipelined 0
+
+static struct drm_i915_private *fence_to_i915(struct i915_fence_reg *fence)
+{
+	return fence->ggtt->vm.i915;
+}
+
+static struct intel_uncore *fence_to_uncore(struct i915_fence_reg *fence)
+{
+	return fence->ggtt->vm.gt->uncore;
+}
+
+static void i965_write_fence_reg(struct i915_fence_reg *fence)
+{
+	i915_reg_t fence_reg_lo, fence_reg_hi;
+	int fence_pitch_shift;
+	u64 val;
+
+	if (GRAPHICS_VER(fence_to_i915(fence)) >= 6) {
+		fence_reg_lo = FENCE_REG_GEN6_LO(fence->id);
+		fence_reg_hi = FENCE_REG_GEN6_HI(fence->id);
+		fence_pitch_shift = GEN6_FENCE_PITCH_SHIFT;
+
+	} else {
+		fence_reg_lo = FENCE_REG_965_LO(fence->id);
+		fence_reg_hi = FENCE_REG_965_HI(fence->id);
+		fence_pitch_shift = I965_FENCE_PITCH_SHIFT;
+	}
+
+	val = 0;
+	if (fence->tiling) {
+		unsigned int stride = fence->stride;
+
+		GEM_BUG_ON(!IS_ALIGNED(stride, 128));
+
+		val = fence->start + fence->size - I965_FENCE_PAGE;
+		val <<= 32;
+		val |= fence->start;
+		val |= (u64)((stride / 128) - 1) << fence_pitch_shift;
+		if (fence->tiling == I915_TILING_Y)
+			val |= BIT(I965_FENCE_TILING_Y_SHIFT);
+		val |= I965_FENCE_REG_VALID;
+	}
+
+	if (!pipelined) {
+		struct intel_uncore *uncore = fence_to_uncore(fence);
+
+		/*
+		 * To w/a incoherency with non-atomic 64-bit register updates,
+		 * we split the 64-bit update into two 32-bit writes. In order
+		 * for a partial fence not to be evaluated between writes, we
+		 * precede the update with write to turn off the fence register,
+		 * and only enable the fence as the last step.
+		 *
+		 * For extra levels of paranoia, we make sure each step lands
+		 * before applying the next step.
+		 */
+		intel_uncore_write_fw(uncore, fence_reg_lo, 0);
+		intel_uncore_posting_read_fw(uncore, fence_reg_lo);
+
+		intel_uncore_write_fw(uncore, fence_reg_hi, upper_32_bits(val));
+		intel_uncore_write_fw(uncore, fence_reg_lo, lower_32_bits(val));
+		intel_uncore_posting_read_fw(uncore, fence_reg_lo);
+	}
+}
+
+static void i915_write_fence_reg(struct i915_fence_reg *fence)
+{
+	u32 val;
+
+	val = 0;
+	if (fence->tiling) {
+		unsigned int stride = fence->stride;
+		unsigned int tiling = fence->tiling;
+		bool is_y_tiled = tiling == I915_TILING_Y;
+
+		if (is_y_tiled && HAS_128_BYTE_Y_TILING(fence_to_i915(fence)))
+			stride /= 128;
+		else
+			stride /= 512;
+		GEM_BUG_ON(!is_power_of_2(stride));
+
+		val = fence->start;
+		if (is_y_tiled)
+			val |= BIT(I830_FENCE_TILING_Y_SHIFT);
+		val |= I915_FENCE_SIZE_BITS(fence->size);
+		val |= ilog2(stride) << I830_FENCE_PITCH_SHIFT;
+
+		val |= I830_FENCE_REG_VALID;
+	}
+
+	if (!pipelined) {
+		struct intel_uncore *uncore = fence_to_uncore(fence);
+		i915_reg_t reg = FENCE_REG(fence->id);
+
+		intel_uncore_write_fw(uncore, reg, val);
+		intel_uncore_posting_read_fw(uncore, reg);
+	}
+}
+
+static void i830_write_fence_reg(struct i915_fence_reg *fence)
+{
+	u32 val;
+
+	val = 0;
+	if (fence->tiling) {
+		unsigned int stride = fence->stride;
+
+		val = fence->start;
+		if (fence->tiling == I915_TILING_Y)
+			val |= BIT(I830_FENCE_TILING_Y_SHIFT);
+		val |= I830_FENCE_SIZE_BITS(fence->size);
+		val |= ilog2(stride / 128) << I830_FENCE_PITCH_SHIFT;
+		val |= I830_FENCE_REG_VALID;
+	}
+
+	if (!pipelined) {
+		struct intel_uncore *uncore = fence_to_uncore(fence);
+		i915_reg_t reg = FENCE_REG(fence->id);
+
+		intel_uncore_write_fw(uncore, reg, val);
+		intel_uncore_posting_read_fw(uncore, reg);
+	}
+}
+
+static void fence_write(struct i915_fence_reg *fence)
+{
+	struct drm_i915_private *i915 = fence_to_i915(fence);
+
+	/*
+	 * Previous access through the fence register is marshalled by
+	 * the mb() inside the fault handlers (i915_gem_release_mmaps)
+	 * and explicitly managed for internal users.
+	 */
+
+	if (GRAPHICS_VER(i915) == 2)
+		i830_write_fence_reg(fence);
+	else if (GRAPHICS_VER(i915) == 3)
+		i915_write_fence_reg(fence);
+	else
+		i965_write_fence_reg(fence);
+
+	/*
+	 * Access through the fenced region afterwards is
+	 * ordered by the posting reads whilst writing the registers.
+	 */
+}
+
+static bool gpu_uses_fence_registers(struct i915_fence_reg *fence)
+{
+	return GRAPHICS_VER(fence_to_i915(fence)) < 4;
+}
+
+static int fence_update(struct i915_fence_reg *fence,
+			struct i915_vma *vma)
+{
+	struct i915_ggtt *ggtt = fence->ggtt;
+	struct intel_uncore *uncore = fence_to_uncore(fence);
+	intel_wakeref_t wakeref;
+	struct i915_vma *old;
+	int ret;
+
+	fence->tiling = 0;
+	if (vma) {
+		GEM_BUG_ON(!i915_gem_object_get_stride(vma->obj) ||
+			   !i915_gem_object_get_tiling(vma->obj));
+
+		if (!i915_vma_is_map_and_fenceable(vma))
+			return -EINVAL;
+
+		if (gpu_uses_fence_registers(fence)) {
+			/* implicit 'unfenced' GPU blits */
+			ret = i915_vma_sync(vma);
+			if (ret)
+				return ret;
+		}
+
+		fence->start = vma->node.start;
+		fence->size = vma->fence_size;
+		fence->stride = i915_gem_object_get_stride(vma->obj);
+		fence->tiling = i915_gem_object_get_tiling(vma->obj);
+	}
+	WRITE_ONCE(fence->dirty, false);
+
+	old = xchg(&fence->vma, NULL);
+	if (old) {
+		/* XXX Ideally we would move the waiting to outside the mutex */
+		ret = i915_active_wait(&fence->active);
+		if (ret) {
+			fence->vma = old;
+			return ret;
+		}
+
+		i915_vma_flush_writes(old);
+
+		/*
+		 * Ensure that all userspace CPU access is completed before
+		 * stealing the fence.
+		 */
+		if (old != vma) {
+			GEM_BUG_ON(old->fence != fence);
+			i915_vma_revoke_mmap(old);
+			old->fence = NULL;
+		}
+
+		list_move(&fence->link, &ggtt->fence_list);
+	}
+
+	/*
+	 * We only need to update the register itself if the device is awake.
+	 * If the device is currently powered down, we will defer the write
+	 * to the runtime resume, see intel_ggtt_restore_fences().
+	 *
+	 * This only works for removing the fence register, on acquisition
+	 * the caller must hold the rpm wakeref. The fence register must
+	 * be cleared before we can use any other fences to ensure that
+	 * the new fences do not overlap the elided clears, confusing HW.
+	 */
+	wakeref = intel_runtime_pm_get_if_in_use(uncore->rpm);
+	if (!wakeref) {
+		GEM_BUG_ON(vma);
+		return 0;
+	}
+
+	WRITE_ONCE(fence->vma, vma);
+	fence_write(fence);
+
+	if (vma) {
+		vma->fence = fence;
+		list_move_tail(&fence->link, &ggtt->fence_list);
+	}
+
+	intel_runtime_pm_put(uncore->rpm, wakeref);
+	return 0;
+}
+
+/**
+ * i915_vma_revoke_fence - force-remove fence for a VMA
+ * @vma: vma to map linearly (not through a fence reg)
+ *
+ * This function force-removes any fence from the given object, which is useful
+ * if the kernel wants to do untiled GTT access.
+ */
+void i915_vma_revoke_fence(struct i915_vma *vma)
+{
+	struct i915_fence_reg *fence = vma->fence;
+	intel_wakeref_t wakeref;
+
+	lockdep_assert_held(&vma->vm->mutex);
+	if (!fence)
+		return;
+
+	GEM_BUG_ON(fence->vma != vma);
+	GEM_BUG_ON(!i915_active_is_idle(&fence->active));
+	GEM_BUG_ON(atomic_read(&fence->pin_count));
+
+	fence->tiling = 0;
+	WRITE_ONCE(fence->vma, NULL);
+	vma->fence = NULL;
+
+	/*
+	 * Skip the write to HW if and only if the device is currently
+	 * suspended.
+	 *
+	 * If the driver does not currently hold a wakeref (if_in_use == 0),
+	 * the device may currently be runtime suspended, or it may be woken
+	 * up before the suspend takes place. If the device is not suspended
+	 * (powered down) and we skip clearing the fence register, the HW is
+	 * left in an undefined state where we may end up with multiple
+	 * registers overlapping.
+	 */
+	with_intel_runtime_pm_if_active(fence_to_uncore(fence)->rpm, wakeref)
+		fence_write(fence);
+}
+
+static bool fence_is_active(const struct i915_fence_reg *fence)
+{
+	return fence->vma && i915_vma_is_active(fence->vma);
+}
+
+static struct i915_fence_reg *fence_find(struct i915_ggtt *ggtt)
+{
+	struct i915_fence_reg *active = NULL;
+	struct i915_fence_reg *fence, *fn;
+
+	list_for_each_entry_safe(fence, fn, &ggtt->fence_list, link) {
+		GEM_BUG_ON(fence->vma && fence->vma->fence != fence);
+
+		if (fence == active) /* now seen this fence twice */
+			active = ERR_PTR(-EAGAIN);
+
+		/* Prefer idle fences so we do not have to wait on the GPU */
+		if (active != ERR_PTR(-EAGAIN) && fence_is_active(fence)) {
+			if (!active)
+				active = fence;
+
+			list_move_tail(&fence->link, &ggtt->fence_list);
+			continue;
+		}
+
+		if (atomic_read(&fence->pin_count))
+			continue;
+
+		return fence;
+	}
+
+	/* Wait for completion of pending flips which consume fences */
+	if (intel_has_pending_fb_unpin(ggtt->vm.i915))
+		return ERR_PTR(-EAGAIN);
+
+	return ERR_PTR(-ENOBUFS);
+}
+
+int __i915_vma_pin_fence(struct i915_vma *vma)
+{
+	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vma->vm);
+	struct i915_fence_reg *fence;
+	struct i915_vma *set = i915_gem_object_is_tiled(vma->obj) ? vma : NULL;
+	int err;
+
+	lockdep_assert_held(&vma->vm->mutex);
+
+	/* Just update our place in the LRU if our fence is getting reused. */
+	if (vma->fence) {
+		fence = vma->fence;
+		GEM_BUG_ON(fence->vma != vma);
+		atomic_inc(&fence->pin_count);
+		if (!fence->dirty) {
+			list_move_tail(&fence->link, &ggtt->fence_list);
+			return 0;
+		}
+	} else if (set) {
+		fence = fence_find(ggtt);
+		if (IS_ERR(fence))
+			return PTR_ERR(fence);
+
+		GEM_BUG_ON(atomic_read(&fence->pin_count));
+		atomic_inc(&fence->pin_count);
+	} else {
+		return 0;
+	}
+
+	err = fence_update(fence, set);
+	if (err)
+		goto out_unpin;
+
+	GEM_BUG_ON(fence->vma != set);
+	GEM_BUG_ON(vma->fence != (set ? fence : NULL));
+
+	if (set)
+		return 0;
+
+out_unpin:
+	atomic_dec(&fence->pin_count);
+	return err;
+}
+
+/**
+ * i915_vma_pin_fence - set up fencing for a vma
+ * @vma: vma to map through a fence reg
+ *
+ * When mapping objects through the GTT, userspace wants to be able to write
+ * to them without having to worry about swizzling if the object is tiled.
+ * This function walks the fence regs looking for a free one for @obj,
+ * stealing one if it can't find any.
+ *
+ * It then sets up the reg based on the object's properties: address, pitch
+ * and tiling format.
+ *
+ * For an untiled surface, this removes any existing fence.
+ *
+ * Returns:
+ *
+ * 0 on success, negative error code on failure.
+ */
+int i915_vma_pin_fence(struct i915_vma *vma)
+{
+	int err;
+
+	if (!vma->fence && !i915_gem_object_is_tiled(vma->obj))
+		return 0;
+
+	/*
+	 * Note that we revoke fences on runtime suspend. Therefore the user
+	 * must keep the device awake whilst using the fence.
+	 */
+	assert_rpm_wakelock_held(vma->vm->gt->uncore->rpm);
+	GEM_BUG_ON(!i915_vma_is_ggtt(vma));
+
+	err = mutex_lock_interruptible(&vma->vm->mutex);
+	if (err)
+		return err;
+
+	err = __i915_vma_pin_fence(vma);
+	mutex_unlock(&vma->vm->mutex);
+
+	return err;
+}
+
+/**
+ * i915_reserve_fence - Reserve a fence for vGPU
+ * @ggtt: Global GTT
+ *
+ * This function walks the fence regs looking for a free one and remove
+ * it from the fence_list. It is used to reserve fence for vGPU to use.
+ */
+struct i915_fence_reg *i915_reserve_fence(struct i915_ggtt *ggtt)
+{
+	struct i915_fence_reg *fence;
+	int count;
+	int ret;
+
+	lockdep_assert_held(&ggtt->vm.mutex);
+
+	/* Keep at least one fence available for the display engine. */
+	count = 0;
+	list_for_each_entry(fence, &ggtt->fence_list, link)
+		count += !atomic_read(&fence->pin_count);
+	if (count <= 1)
+		return ERR_PTR(-ENOSPC);
+
+	fence = fence_find(ggtt);
+	if (IS_ERR(fence))
+		return fence;
+
+	if (fence->vma) {
+		/* Force-remove fence from VMA */
+		ret = fence_update(fence, NULL);
+		if (ret)
+			return ERR_PTR(ret);
+	}
+
+	list_del(&fence->link);
+
+	return fence;
+}
+
+/**
+ * i915_unreserve_fence - Reclaim a reserved fence
+ * @fence: the fence reg
+ *
+ * This function add a reserved fence register from vGPU to the fence_list.
+ */
+void i915_unreserve_fence(struct i915_fence_reg *fence)
+{
+	struct i915_ggtt *ggtt = fence->ggtt;
+
+	lockdep_assert_held(&ggtt->vm.mutex);
+
+	list_add(&fence->link, &ggtt->fence_list);
+}
+
+/**
+ * intel_ggtt_restore_fences - restore fence state
+ * @ggtt: Global GTT
+ *
+ * Restore the hw fence state to match the software tracking again, to be called
+ * after a gpu reset and on resume. Note that on runtime suspend we only cancel
+ * the fences, to be reacquired by the user later.
+ */
+void intel_ggtt_restore_fences(struct i915_ggtt *ggtt)
+{
+	int i;
+
+	for (i = 0; i < ggtt->num_fences; i++)
+		fence_write(&ggtt->fence_regs[i]);
+}
+
+/**
+ * DOC: tiling swizzling details
+ *
+ * The idea behind tiling is to increase cache hit rates by rearranging
+ * pixel data so that a group of pixel accesses are in the same cacheline.
+ * Performance improvement from doing this on the back/depth buffer are on
+ * the order of 30%.
+ *
+ * Intel architectures make this somewhat more complicated, though, by
+ * adjustments made to addressing of data when the memory is in interleaved
+ * mode (matched pairs of DIMMS) to improve memory bandwidth.
+ * For interleaved memory, the CPU sends every sequential 64 bytes
+ * to an alternate memory channel so it can get the bandwidth from both.
+ *
+ * The GPU also rearranges its accesses for increased bandwidth to interleaved
+ * memory, and it matches what the CPU does for non-tiled.  However, when tiled
+ * it does it a little differently, since one walks addresses not just in the
+ * X direction but also Y.  So, along with alternating channels when bit
+ * 6 of the address flips, it also alternates when other bits flip --  Bits 9
+ * (every 512 bytes, an X tile scanline) and 10 (every two X tile scanlines)
+ * are common to both the 915 and 965-class hardware.
+ *
+ * The CPU also sometimes XORs in higher bits as well, to improve
+ * bandwidth doing strided access like we do so frequently in graphics.  This
+ * is called "Channel XOR Randomization" in the MCH documentation.  The result
+ * is that the CPU is XORing in either bit 11 or bit 17 to bit 6 of its address
+ * decode.
+ *
+ * All of this bit 6 XORing has an effect on our memory management,
+ * as we need to make sure that the 3d driver can correctly address object
+ * contents.
+ *
+ * If we don't have interleaved memory, all tiling is safe and no swizzling is
+ * required.
+ *
+ * When bit 17 is XORed in, we simply refuse to tile at all.  Bit
+ * 17 is not just a page offset, so as we page an object out and back in,
+ * individual pages in it will have different bit 17 addresses, resulting in
+ * each 64 bytes being swapped with its neighbor!
+ *
+ * Otherwise, if interleaved, we have to tell the 3d driver what the address
+ * swizzling it needs to do is, since it's writing with the CPU to the pages
+ * (bit 6 and potentially bit 11 XORed in), and the GPU is reading from the
+ * pages (bit 6, 9, and 10 XORed in), resulting in a cumulative bit swizzling
+ * required by the CPU of XORing in bit 6, 9, 10, and potentially 11, in order
+ * to match what the GPU expects.
+ */
+
+/**
+ * detect_bit_6_swizzle - detect bit 6 swizzling pattern
+ * @ggtt: Global GGTT
+ *
+ * Detects bit 6 swizzling of address lookup between IGD access and CPU
+ * access through main memory.
+ */
+static void detect_bit_6_swizzle(struct i915_ggtt *ggtt)
+{
+	struct intel_uncore *uncore = ggtt->vm.gt->uncore;
+	struct drm_i915_private *i915 = ggtt->vm.i915;
+	u32 swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN;
+	u32 swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN;
+
+	if (GRAPHICS_VER(i915) >= 8 || IS_VALLEYVIEW(i915)) {
+		/*
+		 * On BDW+, swizzling is not used. We leave the CPU memory
+		 * controller in charge of optimizing memory accesses without
+		 * the extra address manipulation GPU side.
+		 *
+		 * VLV and CHV don't have GPU swizzling.
+		 */
+		swizzle_x = I915_BIT_6_SWIZZLE_NONE;
+		swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+	} else if (GRAPHICS_VER(i915) >= 6) {
+		if (i915->preserve_bios_swizzle) {
+			if (intel_uncore_read(uncore, DISP_ARB_CTL) &
+			    DISP_TILE_SURFACE_SWIZZLING) {
+				swizzle_x = I915_BIT_6_SWIZZLE_9_10;
+				swizzle_y = I915_BIT_6_SWIZZLE_9;
+			} else {
+				swizzle_x = I915_BIT_6_SWIZZLE_NONE;
+				swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+			}
+		} else {
+			u32 dimm_c0, dimm_c1;
+
+			dimm_c0 = intel_uncore_read(uncore, MAD_DIMM_C0);
+			dimm_c1 = intel_uncore_read(uncore, MAD_DIMM_C1);
+			dimm_c0 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK;
+			dimm_c1 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK;
+			/*
+			 * Enable swizzling when the channels are populated
+			 * with identically sized dimms. We don't need to check
+			 * the 3rd channel because no cpu with gpu attached
+			 * ships in that configuration. Also, swizzling only
+			 * makes sense for 2 channels anyway.
+			 */
+			if (dimm_c0 == dimm_c1) {
+				swizzle_x = I915_BIT_6_SWIZZLE_9_10;
+				swizzle_y = I915_BIT_6_SWIZZLE_9;
+			} else {
+				swizzle_x = I915_BIT_6_SWIZZLE_NONE;
+				swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+			}
+		}
+	} else if (GRAPHICS_VER(i915) == 5) {
+		/*
+		 * On Ironlake whatever DRAM config, GPU always do
+		 * same swizzling setup.
+		 */
+		swizzle_x = I915_BIT_6_SWIZZLE_9_10;
+		swizzle_y = I915_BIT_6_SWIZZLE_9;
+	} else if (GRAPHICS_VER(i915) == 2) {
+		/*
+		 * As far as we know, the 865 doesn't have these bit 6
+		 * swizzling issues.
+		 */
+		swizzle_x = I915_BIT_6_SWIZZLE_NONE;
+		swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+	} else if (IS_G45(i915) || IS_I965G(i915) || IS_G33(i915)) {
+		/*
+		 * The 965, G33, and newer, have a very flexible memory
+		 * configuration.  It will enable dual-channel mode
+		 * (interleaving) on as much memory as it can, and the GPU
+		 * will additionally sometimes enable different bit 6
+		 * swizzling for tiled objects from the CPU.
+		 *
+		 * Here's what I found on the G965:
+		 *    slot fill         memory size  swizzling
+		 * 0A   0B   1A   1B    1-ch   2-ch
+		 * 512  0    0    0     512    0     O
+		 * 512  0    512  0     16     1008  X
+		 * 512  0    0    512   16     1008  X
+		 * 0    512  0    512   16     1008  X
+		 * 1024 1024 1024 0     2048   1024  O
+		 *
+		 * We could probably detect this based on either the DRB
+		 * matching, which was the case for the swizzling required in
+		 * the table above, or from the 1-ch value being less than
+		 * the minimum size of a rank.
+		 *
+		 * Reports indicate that the swizzling actually
+		 * varies depending upon page placement inside the
+		 * channels, i.e. we see swizzled pages where the
+		 * banks of memory are paired and unswizzled on the
+		 * uneven portion, so leave that as unknown.
+		 */
+		if (intel_uncore_read16(uncore, C0DRB3_BW) ==
+		    intel_uncore_read16(uncore, C1DRB3_BW)) {
+			swizzle_x = I915_BIT_6_SWIZZLE_9_10;
+			swizzle_y = I915_BIT_6_SWIZZLE_9;
+		}
+	} else {
+		u32 dcc = intel_uncore_read(uncore, DCC);
+
+		/*
+		 * On 9xx chipsets, channel interleave by the CPU is
+		 * determined by DCC.  For single-channel, neither the CPU
+		 * nor the GPU do swizzling.  For dual channel interleaved,
+		 * the GPU's interleave is bit 9 and 10 for X tiled, and bit
+		 * 9 for Y tiled.  The CPU's interleave is independent, and
+		 * can be based on either bit 11 (haven't seen this yet) or
+		 * bit 17 (common).
+		 */
+		switch (dcc & DCC_ADDRESSING_MODE_MASK) {
+		case DCC_ADDRESSING_MODE_SINGLE_CHANNEL:
+		case DCC_ADDRESSING_MODE_DUAL_CHANNEL_ASYMMETRIC:
+			swizzle_x = I915_BIT_6_SWIZZLE_NONE;
+			swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+			break;
+		case DCC_ADDRESSING_MODE_DUAL_CHANNEL_INTERLEAVED:
+			if (dcc & DCC_CHANNEL_XOR_DISABLE) {
+				/*
+				 * This is the base swizzling by the GPU for
+				 * tiled buffers.
+				 */
+				swizzle_x = I915_BIT_6_SWIZZLE_9_10;
+				swizzle_y = I915_BIT_6_SWIZZLE_9;
+			} else if ((dcc & DCC_CHANNEL_XOR_BIT_17) == 0) {
+				/* Bit 11 swizzling by the CPU in addition. */
+				swizzle_x = I915_BIT_6_SWIZZLE_9_10_11;
+				swizzle_y = I915_BIT_6_SWIZZLE_9_11;
+			} else {
+				/* Bit 17 swizzling by the CPU in addition. */
+				swizzle_x = I915_BIT_6_SWIZZLE_9_10_17;
+				swizzle_y = I915_BIT_6_SWIZZLE_9_17;
+			}
+			break;
+		}
+
+		/* check for L-shaped memory aka modified enhanced addressing */
+		if (GRAPHICS_VER(i915) == 4 &&
+		    !(intel_uncore_read(uncore, DCC2) & DCC2_MODIFIED_ENHANCED_DISABLE)) {
+			swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN;
+			swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN;
+		}
+
+		if (dcc == 0xffffffff) {
+			drm_err(&i915->drm, "Couldn't read from MCHBAR.  "
+				  "Disabling tiling.\n");
+			swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN;
+			swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN;
+		}
+	}
+
+	if (swizzle_x == I915_BIT_6_SWIZZLE_UNKNOWN ||
+	    swizzle_y == I915_BIT_6_SWIZZLE_UNKNOWN) {
+		/*
+		 * Userspace likes to explode if it sees unknown swizzling,
+		 * so lie. We will finish the lie when reporting through
+		 * the get-tiling-ioctl by reporting the physical swizzle
+		 * mode as unknown instead.
+		 *
+		 * As we don't strictly know what the swizzling is, it may be
+		 * bit17 dependent, and so we need to also prevent the pages
+		 * from being moved.
+		 */
+		i915->gem_quirks |= GEM_QUIRK_PIN_SWIZZLED_PAGES;
+		swizzle_x = I915_BIT_6_SWIZZLE_NONE;
+		swizzle_y = I915_BIT_6_SWIZZLE_NONE;
+	}
+
+	to_gt(i915)->ggtt->bit_6_swizzle_x = swizzle_x;
+	to_gt(i915)->ggtt->bit_6_swizzle_y = swizzle_y;
+}
+
+/*
+ * Swap every 64 bytes of this page around, to account for it having a new
+ * bit 17 of its physical address and therefore being interpreted differently
+ * by the GPU.
+ */
+static void swizzle_page(struct page *page)
+{
+	char temp[64];
+	char *vaddr;
+	int i;
+
+	vaddr = kmap(page);
+
+	for (i = 0; i < PAGE_SIZE; i += 128) {
+		memcpy(temp, &vaddr[i], 64);
+		memcpy(&vaddr[i], &vaddr[i + 64], 64);
+		memcpy(&vaddr[i + 64], temp, 64);
+	}
+
+	kunmap(page);
+}
+
+/**
+ * i915_gem_object_do_bit_17_swizzle - fixup bit 17 swizzling
+ * @obj: i915 GEM buffer object
+ * @pages: the scattergather list of physical pages
+ *
+ * This function fixes up the swizzling in case any page frame number for this
+ * object has changed in bit 17 since that state has been saved with
+ * i915_gem_object_save_bit_17_swizzle().
+ *
+ * This is called when pinning backing storage again, since the kernel is free
+ * to move unpinned backing storage around (either by directly moving pages or
+ * by swapping them out and back in again).
+ */
+void
+i915_gem_object_do_bit_17_swizzle(struct drm_i915_gem_object *obj,
+				  struct sg_table *pages)
+{
+	struct sgt_iter sgt_iter;
+	struct page *page;
+	int i;
+
+	if (obj->bit_17 == NULL)
+		return;
+
+	i = 0;
+	for_each_sgt_page(page, sgt_iter, pages) {
+		char new_bit_17 = page_to_phys(page) >> 17;
+
+		if ((new_bit_17 & 0x1) != (test_bit(i, obj->bit_17) != 0)) {
+			swizzle_page(page);
+			set_page_dirty(page);
+		}
+
+		i++;
+	}
+}
+
+/**
+ * i915_gem_object_save_bit_17_swizzle - save bit 17 swizzling
+ * @obj: i915 GEM buffer object
+ * @pages: the scattergather list of physical pages
+ *
+ * This function saves the bit 17 of each page frame number so that swizzling
+ * can be fixed up later on with i915_gem_object_do_bit_17_swizzle(). This must
+ * be called before the backing storage can be unpinned.
+ */
+void
+i915_gem_object_save_bit_17_swizzle(struct drm_i915_gem_object *obj,
+				    struct sg_table *pages)
+{
+	const unsigned int page_count = obj->base.size >> PAGE_SHIFT;
+	struct sgt_iter sgt_iter;
+	struct page *page;
+	int i;
+
+	if (obj->bit_17 == NULL) {
+		obj->bit_17 = bitmap_zalloc(page_count, GFP_KERNEL);
+		if (obj->bit_17 == NULL) {
+			DRM_ERROR("Failed to allocate memory for bit 17 "
+				  "record\n");
+			return;
+		}
+	}
+
+	i = 0;
+
+	for_each_sgt_page(page, sgt_iter, pages) {
+		if (page_to_phys(page) & (1 << 17))
+			__set_bit(i, obj->bit_17);
+		else
+			__clear_bit(i, obj->bit_17);
+		i++;
+	}
+}
+
+void intel_ggtt_init_fences(struct i915_ggtt *ggtt)
+{
+	struct drm_i915_private *i915 = ggtt->vm.i915;
+	struct intel_uncore *uncore = ggtt->vm.gt->uncore;
+	int num_fences;
+	int i;
+
+	INIT_LIST_HEAD(&ggtt->fence_list);
+	INIT_LIST_HEAD(&ggtt->userfault_list);
+
+	detect_bit_6_swizzle(ggtt);
+
+	if (!i915_ggtt_has_aperture(ggtt))
+		num_fences = 0;
+	else if (GRAPHICS_VER(i915) >= 7 &&
+		 !(IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)))
+		num_fences = 32;
+	else if (GRAPHICS_VER(i915) >= 4 ||
+		 IS_I945G(i915) || IS_I945GM(i915) ||
+		 IS_G33(i915) || IS_PINEVIEW(i915))
+		num_fences = 16;
+	else
+		num_fences = 8;
+
+	if (intel_vgpu_active(i915))
+		num_fences = intel_uncore_read(uncore,
+					       vgtif_reg(avail_rs.fence_num));
+	ggtt->fence_regs = kcalloc(num_fences,
+				   sizeof(*ggtt->fence_regs),
+				   GFP_KERNEL);
+	if (!ggtt->fence_regs)
+		num_fences = 0;
+
+	/* Initialize fence registers to zero */
+	for (i = 0; i < num_fences; i++) {
+		struct i915_fence_reg *fence = &ggtt->fence_regs[i];
+
+		i915_active_init(&fence->active, NULL, NULL, 0);
+		fence->ggtt = ggtt;
+		fence->id = i;
+		list_add_tail(&fence->link, &ggtt->fence_list);
+	}
+	ggtt->num_fences = num_fences;
+
+	intel_ggtt_restore_fences(ggtt);
+}
+
+void intel_ggtt_fini_fences(struct i915_ggtt *ggtt)
+{
+	int i;
+
+	for (i = 0; i < ggtt->num_fences; i++) {
+		struct i915_fence_reg *fence = &ggtt->fence_regs[i];
+
+		i915_active_fini(&fence->active);
+	}
+
+	kfree(ggtt->fence_regs);
+}
+
+void intel_gt_init_swizzling(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+
+	if (GRAPHICS_VER(i915) < 5 ||
+	    to_gt(i915)->ggtt->bit_6_swizzle_x == I915_BIT_6_SWIZZLE_NONE)
+		return;
+
+	intel_uncore_rmw(uncore, DISP_ARB_CTL, 0, DISP_TILE_SURFACE_SWIZZLING);
+
+	if (GRAPHICS_VER(i915) == 5)
+		return;
+
+	intel_uncore_rmw(uncore, TILECTL, 0, TILECTL_SWZCTL);
+
+	if (GRAPHICS_VER(i915) == 6)
+		intel_uncore_write(uncore,
+				   ARB_MODE,
+				   _MASKED_BIT_ENABLE(ARB_MODE_SWIZZLE_SNB));
+	else if (GRAPHICS_VER(i915) == 7)
+		intel_uncore_write(uncore,
+				   ARB_MODE,
+				   _MASKED_BIT_ENABLE(ARB_MODE_SWIZZLE_IVB));
+	else if (GRAPHICS_VER(i915) == 8)
+		intel_uncore_write(uncore,
+				   GAMTARBMODE,
+				   _MASKED_BIT_ENABLE(ARB_MODE_SWIZZLE_BDW));
+	else
+		MISSING_CASE(GRAPHICS_VER(i915));
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.h b/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.h
new file mode 100644
index 000000000..25340be5e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.h
@@ -0,0 +1,59 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __INTEL_GGTT_FENCING_H__
+#define __INTEL_GGTT_FENCING_H__
+
+#include <linux/list.h>
+#include <linux/types.h>
+
+#include "i915_active.h"
+
+struct drm_i915_gem_object;
+struct i915_ggtt;
+struct i915_vma;
+struct intel_gt;
+struct sg_table;
+
+#define I965_FENCE_PAGE 4096UL
+
+struct i915_fence_reg {
+	struct list_head link;
+	struct i915_ggtt *ggtt;
+	struct i915_vma *vma;
+	atomic_t pin_count;
+	struct i915_active active;
+	int id;
+	/**
+	 * Whether the tiling parameters for the currently
+	 * associated fence register have changed. Note that
+	 * for the purposes of tracking tiling changes we also
+	 * treat the unfenced register, the register slot that
+	 * the object occupies whilst it executes a fenced
+	 * command (such as BLT on gen2/3), as a "fence".
+	 */
+	bool dirty;
+	u32 start;
+	u32 size;
+	u32 tiling;
+	u32 stride;
+};
+
+struct i915_fence_reg *i915_reserve_fence(struct i915_ggtt *ggtt);
+void i915_unreserve_fence(struct i915_fence_reg *fence);
+
+void intel_ggtt_restore_fences(struct i915_ggtt *ggtt);
+
+void i915_gem_object_do_bit_17_swizzle(struct drm_i915_gem_object *obj,
+				       struct sg_table *pages);
+void i915_gem_object_save_bit_17_swizzle(struct drm_i915_gem_object *obj,
+					 struct sg_table *pages);
+
+void intel_ggtt_init_fences(struct i915_ggtt *ggtt);
+void intel_ggtt_fini_fences(struct i915_ggtt *ggtt);
+
+void intel_gt_init_swizzling(struct intel_gt *gt);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt_gmch.c b/drivers/gpu/drm/i915/gt/intel_ggtt_gmch.c
new file mode 100644
index 000000000..4e2163a1a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_ggtt_gmch.c
@@ -0,0 +1,132 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include "intel_ggtt_gmch.h"
+
+#include <drm/intel-gtt.h>
+#include <drm/i915_drm.h>
+
+#include <linux/agp_backend.h>
+
+#include "i915_drv.h"
+#include "i915_utils.h"
+#include "intel_gtt.h"
+#include "intel_gt_regs.h"
+#include "intel_gt.h"
+
+static void gmch_ggtt_insert_page(struct i915_address_space *vm,
+				  dma_addr_t addr,
+				  u64 offset,
+				  enum i915_cache_level cache_level,
+				  u32 unused)
+{
+	unsigned int flags = (cache_level == I915_CACHE_NONE) ?
+		AGP_USER_MEMORY : AGP_USER_CACHED_MEMORY;
+
+	intel_gmch_gtt_insert_page(addr, offset >> PAGE_SHIFT, flags);
+}
+
+static void gmch_ggtt_insert_entries(struct i915_address_space *vm,
+				     struct i915_vma_resource *vma_res,
+				     enum i915_cache_level cache_level,
+				     u32 unused)
+{
+	unsigned int flags = (cache_level == I915_CACHE_NONE) ?
+		AGP_USER_MEMORY : AGP_USER_CACHED_MEMORY;
+
+	intel_gmch_gtt_insert_sg_entries(vma_res->bi.pages, vma_res->start >> PAGE_SHIFT,
+					 flags);
+}
+
+static void gmch_ggtt_invalidate(struct i915_ggtt *ggtt)
+{
+	intel_gmch_gtt_flush();
+}
+
+static void gmch_ggtt_clear_range(struct i915_address_space *vm,
+				  u64 start, u64 length)
+{
+	intel_gmch_gtt_clear_range(start >> PAGE_SHIFT, length >> PAGE_SHIFT);
+}
+
+static void gmch_ggtt_remove(struct i915_address_space *vm)
+{
+	intel_gmch_remove();
+}
+
+/*
+ * Certain Gen5 chipsets require idling the GPU before unmapping anything from
+ * the GTT when VT-d is enabled.
+ */
+static bool needs_idle_maps(struct drm_i915_private *i915)
+{
+	/*
+	 * Query intel_iommu to see if we need the workaround. Presumably that
+	 * was loaded first.
+	 */
+	if (!i915_vtd_active(i915))
+		return false;
+
+	if (GRAPHICS_VER(i915) == 5 && IS_MOBILE(i915))
+		return true;
+
+	return false;
+}
+
+int intel_ggtt_gmch_probe(struct i915_ggtt *ggtt)
+{
+	struct drm_i915_private *i915 = ggtt->vm.i915;
+	phys_addr_t gmadr_base;
+	int ret;
+
+	ret = intel_gmch_probe(i915->bridge_dev, to_pci_dev(i915->drm.dev), NULL);
+	if (!ret) {
+		drm_err(&i915->drm, "failed to set up gmch\n");
+		return -EIO;
+	}
+
+	intel_gmch_gtt_get(&ggtt->vm.total, &gmadr_base, &ggtt->mappable_end);
+
+	ggtt->gmadr =
+		(struct resource)DEFINE_RES_MEM(gmadr_base, ggtt->mappable_end);
+
+	ggtt->vm.alloc_pt_dma = alloc_pt_dma;
+	ggtt->vm.alloc_scratch_dma = alloc_pt_dma;
+
+	if (needs_idle_maps(i915)) {
+		drm_notice(&i915->drm,
+			   "Flushing DMA requests before IOMMU unmaps; performance may be degraded\n");
+		ggtt->do_idle_maps = true;
+	}
+
+	ggtt->vm.insert_page = gmch_ggtt_insert_page;
+	ggtt->vm.insert_entries = gmch_ggtt_insert_entries;
+	ggtt->vm.clear_range = gmch_ggtt_clear_range;
+	ggtt->vm.cleanup = gmch_ggtt_remove;
+
+	ggtt->invalidate = gmch_ggtt_invalidate;
+
+	ggtt->vm.vma_ops.bind_vma    = intel_ggtt_bind_vma;
+	ggtt->vm.vma_ops.unbind_vma  = intel_ggtt_unbind_vma;
+
+	if (unlikely(ggtt->do_idle_maps))
+		drm_notice(&i915->drm,
+			   "Applying Ironlake quirks for intel_iommu\n");
+
+	return 0;
+}
+
+int intel_ggtt_gmch_enable_hw(struct drm_i915_private *i915)
+{
+	if (!intel_gmch_enable_gtt())
+		return -EIO;
+
+	return 0;
+}
+
+void intel_ggtt_gmch_flush(void)
+{
+	intel_gmch_gtt_flush();
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt_gmch.h b/drivers/gpu/drm/i915/gt/intel_ggtt_gmch.h
new file mode 100644
index 000000000..370bf321b
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_ggtt_gmch.h
@@ -0,0 +1,27 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_GGTT_GMCH_H__
+#define __INTEL_GGTT_GMCH_H__
+
+#include "intel_gtt.h"
+
+/* For x86 platforms */
+#if IS_ENABLED(CONFIG_X86)
+
+void intel_ggtt_gmch_flush(void);
+int intel_ggtt_gmch_enable_hw(struct drm_i915_private *i915);
+int intel_ggtt_gmch_probe(struct i915_ggtt *ggtt);
+
+/* Stubs for non-x86 platforms */
+#else
+
+static inline void intel_ggtt_gmch_flush(void) { }
+static inline int intel_ggtt_gmch_enable_hw(struct drm_i915_private *i915) { return -ENODEV; }
+static inline int intel_ggtt_gmch_probe(struct i915_ggtt *ggtt) { return -ENODEV; }
+
+#endif
+
+#endif /* __INTEL_GGTT_GMCH_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_gpu_commands.h b/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
new file mode 100644
index 000000000..25ea5f8a4
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gpu_commands.h
@@ -0,0 +1,459 @@
+/* SPDX-License-Identifier: MIT*/
+/*
+ * Copyright © 2003-2018 Intel Corporation
+ */
+
+#ifndef _INTEL_GPU_COMMANDS_H_
+#define _INTEL_GPU_COMMANDS_H_
+
+#include <linux/bitops.h>
+
+/*
+ * Target address alignments required for GPU access e.g.
+ * MI_STORE_DWORD_IMM.
+ */
+#define alignof_dword 4
+#define alignof_qword 8
+
+/*
+ * Instruction field definitions used by the command parser
+ */
+#define INSTR_CLIENT_SHIFT      29
+#define   INSTR_MI_CLIENT       0x0
+#define   INSTR_BC_CLIENT       0x2
+#define   INSTR_RC_CLIENT       0x3
+#define INSTR_SUBCLIENT_SHIFT   27
+#define INSTR_SUBCLIENT_MASK    0x18000000
+#define   INSTR_MEDIA_SUBCLIENT 0x2
+#define INSTR_26_TO_24_MASK	0x7000000
+#define   INSTR_26_TO_24_SHIFT	24
+
+#define __INSTR(client) ((client) << INSTR_CLIENT_SHIFT)
+
+/*
+ * Memory interface instructions used by the kernel
+ */
+#define MI_INSTR(opcode, flags) \
+	(__INSTR(INSTR_MI_CLIENT) | (opcode) << 23 | (flags))
+/* Many MI commands use bit 22 of the header dword for GGTT vs PPGTT */
+#define  MI_GLOBAL_GTT    (1<<22)
+
+#define MI_NOOP			MI_INSTR(0, 0)
+#define MI_SET_PREDICATE	MI_INSTR(0x01, 0)
+#define   MI_SET_PREDICATE_DISABLE	(0 << 0)
+#define MI_USER_INTERRUPT	MI_INSTR(0x02, 0)
+#define MI_WAIT_FOR_EVENT       MI_INSTR(0x03, 0)
+#define   MI_WAIT_FOR_OVERLAY_FLIP	(1<<16)
+#define   MI_WAIT_FOR_PLANE_B_FLIP      (1<<6)
+#define   MI_WAIT_FOR_PLANE_A_FLIP      (1<<2)
+#define   MI_WAIT_FOR_PLANE_A_SCANLINES (1<<1)
+#define MI_FLUSH		MI_INSTR(0x04, 0)
+#define   MI_READ_FLUSH		(1 << 0)
+#define   MI_EXE_FLUSH		(1 << 1)
+#define   MI_NO_WRITE_FLUSH	(1 << 2)
+#define   MI_SCENE_COUNT	(1 << 3) /* just increment scene count */
+#define   MI_END_SCENE		(1 << 4) /* flush binner and incr scene count */
+#define   MI_INVALIDATE_ISP	(1 << 5) /* invalidate indirect state pointers */
+#define MI_REPORT_HEAD		MI_INSTR(0x07, 0)
+#define MI_ARB_ON_OFF		MI_INSTR(0x08, 0)
+#define   MI_ARB_ENABLE			(1<<0)
+#define   MI_ARB_DISABLE		(0<<0)
+#define MI_BATCH_BUFFER_END	MI_INSTR(0x0a, 0)
+#define MI_SUSPEND_FLUSH	MI_INSTR(0x0b, 0)
+#define   MI_SUSPEND_FLUSH_EN	(1<<0)
+#define MI_SET_APPID		MI_INSTR(0x0e, 0)
+#define   MI_SET_APPID_SESSION_ID(x)	((x) << 0)
+#define MI_OVERLAY_FLIP		MI_INSTR(0x11, 0)
+#define   MI_OVERLAY_CONTINUE	(0x0<<21)
+#define   MI_OVERLAY_ON		(0x1<<21)
+#define   MI_OVERLAY_OFF	(0x2<<21)
+#define MI_LOAD_SCAN_LINES_INCL MI_INSTR(0x12, 0)
+#define MI_DISPLAY_FLIP		MI_INSTR(0x14, 2)
+#define MI_DISPLAY_FLIP_I915	MI_INSTR(0x14, 1)
+#define   MI_DISPLAY_FLIP_PLANE(n) ((n) << 20)
+/* IVB has funny definitions for which plane to flip. */
+#define   MI_DISPLAY_FLIP_IVB_PLANE_A  (0 << 19)
+#define   MI_DISPLAY_FLIP_IVB_PLANE_B  (1 << 19)
+#define   MI_DISPLAY_FLIP_IVB_SPRITE_A (2 << 19)
+#define   MI_DISPLAY_FLIP_IVB_SPRITE_B (3 << 19)
+#define   MI_DISPLAY_FLIP_IVB_PLANE_C  (4 << 19)
+#define   MI_DISPLAY_FLIP_IVB_SPRITE_C (5 << 19)
+/* SKL ones */
+#define   MI_DISPLAY_FLIP_SKL_PLANE_1_A	(0 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_1_B	(1 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_1_C	(2 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_2_A	(4 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_2_B	(5 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_2_C	(6 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_3_A	(7 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_3_B	(8 << 8)
+#define   MI_DISPLAY_FLIP_SKL_PLANE_3_C	(9 << 8)
+#define MI_SEMAPHORE_MBOX	MI_INSTR(0x16, 1) /* gen6, gen7 */
+#define   MI_SEMAPHORE_GLOBAL_GTT    (1<<22)
+#define   MI_SEMAPHORE_UPDATE	    (1<<21)
+#define   MI_SEMAPHORE_COMPARE	    (1<<20)
+#define   MI_SEMAPHORE_REGISTER	    (1<<18)
+#define   MI_SEMAPHORE_SYNC_VR	    (0<<16) /* RCS  wait for VCS  (RVSYNC) */
+#define   MI_SEMAPHORE_SYNC_VER	    (1<<16) /* RCS  wait for VECS (RVESYNC) */
+#define   MI_SEMAPHORE_SYNC_BR	    (2<<16) /* RCS  wait for BCS  (RBSYNC) */
+#define   MI_SEMAPHORE_SYNC_BV	    (0<<16) /* VCS  wait for BCS  (VBSYNC) */
+#define   MI_SEMAPHORE_SYNC_VEV	    (1<<16) /* VCS  wait for VECS (VVESYNC) */
+#define   MI_SEMAPHORE_SYNC_RV	    (2<<16) /* VCS  wait for RCS  (VRSYNC) */
+#define   MI_SEMAPHORE_SYNC_RB	    (0<<16) /* BCS  wait for RCS  (BRSYNC) */
+#define   MI_SEMAPHORE_SYNC_VEB	    (1<<16) /* BCS  wait for VECS (BVESYNC) */
+#define   MI_SEMAPHORE_SYNC_VB	    (2<<16) /* BCS  wait for VCS  (BVSYNC) */
+#define   MI_SEMAPHORE_SYNC_BVE	    (0<<16) /* VECS wait for BCS  (VEBSYNC) */
+#define   MI_SEMAPHORE_SYNC_VVE	    (1<<16) /* VECS wait for VCS  (VEVSYNC) */
+#define   MI_SEMAPHORE_SYNC_RVE	    (2<<16) /* VECS wait for RCS  (VERSYNC) */
+#define   MI_SEMAPHORE_SYNC_INVALID (3<<16)
+#define   MI_SEMAPHORE_SYNC_MASK    (3<<16)
+#define MI_SET_CONTEXT		MI_INSTR(0x18, 0)
+#define   MI_MM_SPACE_GTT		(1<<8)
+#define   MI_MM_SPACE_PHYSICAL		(0<<8)
+#define   MI_SAVE_EXT_STATE_EN		(1<<3)
+#define   MI_RESTORE_EXT_STATE_EN	(1<<2)
+#define   MI_FORCE_RESTORE		(1<<1)
+#define   MI_RESTORE_INHIBIT		(1<<0)
+#define   HSW_MI_RS_SAVE_STATE_EN       (1<<3)
+#define   HSW_MI_RS_RESTORE_STATE_EN    (1<<2)
+#define MI_SEMAPHORE_SIGNAL	MI_INSTR(0x1b, 0) /* GEN8+ */
+#define   MI_SEMAPHORE_TARGET(engine)	((engine)<<15)
+#define MI_SEMAPHORE_WAIT	MI_INSTR(0x1c, 2) /* GEN8+ */
+#define MI_SEMAPHORE_WAIT_TOKEN	MI_INSTR(0x1c, 3) /* GEN12+ */
+#define   MI_SEMAPHORE_REGISTER_POLL	(1 << 16)
+#define   MI_SEMAPHORE_POLL		(1 << 15)
+#define   MI_SEMAPHORE_SAD_GT_SDD	(0 << 12)
+#define   MI_SEMAPHORE_SAD_GTE_SDD	(1 << 12)
+#define   MI_SEMAPHORE_SAD_LT_SDD	(2 << 12)
+#define   MI_SEMAPHORE_SAD_LTE_SDD	(3 << 12)
+#define   MI_SEMAPHORE_SAD_EQ_SDD	(4 << 12)
+#define   MI_SEMAPHORE_SAD_NEQ_SDD	(5 << 12)
+#define   MI_SEMAPHORE_TOKEN_MASK	REG_GENMASK(9, 5)
+#define   MI_SEMAPHORE_TOKEN_SHIFT	5
+#define MI_STORE_DATA_IMM	MI_INSTR(0x20, 0)
+#define MI_STORE_DWORD_IMM	MI_INSTR(0x20, 1)
+#define MI_STORE_DWORD_IMM_GEN4	MI_INSTR(0x20, 2)
+#define MI_STORE_QWORD_IMM_GEN8 (MI_INSTR(0x20, 3) | REG_BIT(21))
+#define   MI_MEM_VIRTUAL	(1 << 22) /* 945,g33,965 */
+#define   MI_USE_GGTT		(1 << 22) /* g4x+ */
+#define MI_STORE_DWORD_INDEX	MI_INSTR(0x21, 1)
+#define MI_ATOMIC		MI_INSTR(0x2f, 1)
+#define MI_ATOMIC_INLINE	(MI_INSTR(0x2f, 9) | MI_ATOMIC_INLINE_DATA)
+#define   MI_ATOMIC_GLOBAL_GTT		(1 << 22)
+#define   MI_ATOMIC_INLINE_DATA		(1 << 18)
+#define   MI_ATOMIC_CS_STALL		(1 << 17)
+#define	  MI_ATOMIC_MOVE		(0x4 << 8)
+
+/*
+ * Official intel docs are somewhat sloppy concerning MI_LOAD_REGISTER_IMM:
+ * - Always issue a MI_NOOP _before_ the MI_LOAD_REGISTER_IMM - otherwise hw
+ *   simply ignores the register load under certain conditions.
+ * - One can actually load arbitrary many arbitrary registers: Simply issue x
+ *   address/value pairs. Don't overdue it, though, x <= 2^4 must hold!
+ */
+#define MI_LOAD_REGISTER_IMM(x)	MI_INSTR(0x22, 2*(x)-1)
+/* Gen11+. addr = base + (ctx_restore ? offset & GENMASK(12,2) : offset) */
+#define   MI_LRI_LRM_CS_MMIO		REG_BIT(19)
+#define   MI_LRI_MMIO_REMAP_EN		REG_BIT(17)
+#define   MI_LRI_FORCE_POSTED		(1<<12)
+#define MI_LOAD_REGISTER_IMM_MAX_REGS (126)
+#define MI_STORE_REGISTER_MEM        MI_INSTR(0x24, 1)
+#define MI_STORE_REGISTER_MEM_GEN8   MI_INSTR(0x24, 2)
+#define   MI_SRM_LRM_GLOBAL_GTT		(1<<22)
+#define MI_FLUSH_DW		MI_INSTR(0x26, 1) /* for GEN6 */
+#define   MI_FLUSH_DW_PROTECTED_MEM_EN	(1 << 22)
+#define   MI_FLUSH_DW_STORE_INDEX	(1<<21)
+#define   MI_INVALIDATE_TLB		(1<<18)
+#define   MI_FLUSH_DW_CCS		(1<<16)
+#define   MI_FLUSH_DW_OP_STOREDW	(1<<14)
+#define   MI_FLUSH_DW_OP_MASK		(3<<14)
+#define   MI_FLUSH_DW_LLC		(1<<9)
+#define   MI_FLUSH_DW_NOTIFY		(1<<8)
+#define   MI_INVALIDATE_BSD		(1<<7)
+#define   MI_FLUSH_DW_USE_GTT		(1<<2)
+#define   MI_FLUSH_DW_USE_PPGTT		(0<<2)
+#define MI_LOAD_REGISTER_MEM	   MI_INSTR(0x29, 1)
+#define MI_LOAD_REGISTER_MEM_GEN8  MI_INSTR(0x29, 2)
+#define MI_LOAD_REGISTER_REG    MI_INSTR(0x2A, 1)
+#define   MI_LRR_SOURCE_CS_MMIO		REG_BIT(18)
+#define MI_BATCH_BUFFER		MI_INSTR(0x30, 1)
+#define   MI_BATCH_NON_SECURE		(1)
+/* for snb/ivb/vlv this also means "batch in ppgtt" when ppgtt is enabled. */
+#define   MI_BATCH_NON_SECURE_I965	(1<<8)
+#define   MI_BATCH_PPGTT_HSW		(1<<8)
+#define   MI_BATCH_NON_SECURE_HSW	(1<<13)
+#define MI_BATCH_BUFFER_START	MI_INSTR(0x31, 0)
+#define   MI_BATCH_GTT		    (2<<6) /* aliased with (1<<7) on gen4 */
+#define MI_BATCH_BUFFER_START_GEN8	MI_INSTR(0x31, 1)
+#define   MI_BATCH_RESOURCE_STREAMER REG_BIT(10)
+#define   MI_BATCH_PREDICATE         REG_BIT(15) /* HSW+ on RCS only*/
+
+/*
+ * 3D instructions used by the kernel
+ */
+#define GFX_INSTR(opcode, flags) ((0x3 << 29) | ((opcode) << 24) | (flags))
+
+#define GEN9_MEDIA_POOL_STATE     ((0x3 << 29) | (0x2 << 27) | (0x5 << 16) | 4)
+#define   GEN9_MEDIA_POOL_ENABLE  (1 << 31)
+#define GFX_OP_RASTER_RULES    ((0x3<<29)|(0x7<<24))
+#define GFX_OP_SCISSOR         ((0x3<<29)|(0x1c<<24)|(0x10<<19))
+#define   SC_UPDATE_SCISSOR       (0x1<<1)
+#define   SC_ENABLE_MASK          (0x1<<0)
+#define   SC_ENABLE               (0x1<<0)
+#define GFX_OP_LOAD_INDIRECT   ((0x3<<29)|(0x1d<<24)|(0x7<<16))
+#define GFX_OP_SCISSOR_INFO    ((0x3<<29)|(0x1d<<24)|(0x81<<16)|(0x1))
+#define   SCI_YMIN_MASK      (0xffff<<16)
+#define   SCI_XMIN_MASK      (0xffff<<0)
+#define   SCI_YMAX_MASK      (0xffff<<16)
+#define   SCI_XMAX_MASK      (0xffff<<0)
+#define GFX_OP_SCISSOR_ENABLE	 ((0x3<<29)|(0x1c<<24)|(0x10<<19))
+#define GFX_OP_SCISSOR_RECT	 ((0x3<<29)|(0x1d<<24)|(0x81<<16)|1)
+#define GFX_OP_COLOR_FACTOR      ((0x3<<29)|(0x1d<<24)|(0x1<<16)|0x0)
+#define GFX_OP_STIPPLE           ((0x3<<29)|(0x1d<<24)|(0x83<<16))
+#define GFX_OP_MAP_INFO          ((0x3<<29)|(0x1d<<24)|0x4)
+#define GFX_OP_DESTBUFFER_VARS   ((0x3<<29)|(0x1d<<24)|(0x85<<16)|0x0)
+#define GFX_OP_DESTBUFFER_INFO	 ((0x3<<29)|(0x1d<<24)|(0x8e<<16)|1)
+#define GFX_OP_DRAWRECT_INFO     ((0x3<<29)|(0x1d<<24)|(0x80<<16)|(0x3))
+#define GFX_OP_DRAWRECT_INFO_I965  ((0x7900<<16)|0x2)
+
+#define XY_CTRL_SURF_INSTR_SIZE		5
+#define MI_FLUSH_DW_SIZE		3
+#define XY_CTRL_SURF_COPY_BLT		((2 << 29) | (0x48 << 22) | 3)
+#define   SRC_ACCESS_TYPE_SHIFT		21
+#define   DST_ACCESS_TYPE_SHIFT		20
+#define   CCS_SIZE_MASK			0x3FF
+#define   CCS_SIZE_SHIFT		8
+#define   XY_CTRL_SURF_MOCS_MASK	GENMASK(31, 25)
+#define   NUM_CCS_BYTES_PER_BLOCK	256
+#define   NUM_BYTES_PER_CCS_BYTE	256
+#define   NUM_CCS_BLKS_PER_XFER		1024
+#define   INDIRECT_ACCESS		0
+#define   DIRECT_ACCESS			1
+
+#define COLOR_BLT_CMD			(2 << 29 | 0x40 << 22 | (5 - 2))
+#define XY_COLOR_BLT_CMD		(2 << 29 | 0x50 << 22)
+#define XY_FAST_COLOR_BLT_CMD		(2 << 29 | 0x44 << 22)
+#define   XY_FAST_COLOR_BLT_DEPTH_32	(2 << 19)
+#define   XY_FAST_COLOR_BLT_DW		16
+#define   XY_FAST_COLOR_BLT_MOCS_MASK	GENMASK(27, 21)
+#define   XY_FAST_COLOR_BLT_MEM_TYPE_SHIFT 31
+
+#define   XY_FAST_COPY_BLT_D0_SRC_TILING_MASK     REG_GENMASK(21, 20)
+#define   XY_FAST_COPY_BLT_D0_DST_TILING_MASK     REG_GENMASK(14, 13)
+#define   XY_FAST_COPY_BLT_D0_SRC_TILE_MODE(mode)  \
+	REG_FIELD_PREP(XY_FAST_COPY_BLT_D0_SRC_TILING_MASK, mode)
+#define   XY_FAST_COPY_BLT_D0_DST_TILE_MODE(mode)  \
+	REG_FIELD_PREP(XY_FAST_COPY_BLT_D0_DST_TILING_MASK, mode)
+#define     LINEAR				0
+#define     TILE_X				0x1
+#define     XMAJOR				0x1
+#define     YMAJOR				0x2
+#define     TILE_64			0x3
+#define   XY_FAST_COPY_BLT_D1_SRC_TILE4	REG_BIT(31)
+#define   XY_FAST_COPY_BLT_D1_DST_TILE4	REG_BIT(30)
+#define BLIT_CCTL_SRC_MOCS_MASK  REG_GENMASK(6, 0)
+#define BLIT_CCTL_DST_MOCS_MASK  REG_GENMASK(14, 8)
+/* Note:  MOCS value = (index << 1) */
+#define BLIT_CCTL_SRC_MOCS(idx) \
+	REG_FIELD_PREP(BLIT_CCTL_SRC_MOCS_MASK, (idx) << 1)
+#define BLIT_CCTL_DST_MOCS(idx) \
+	REG_FIELD_PREP(BLIT_CCTL_DST_MOCS_MASK, (idx) << 1)
+
+#define SRC_COPY_BLT_CMD		(2 << 29 | 0x43 << 22)
+#define GEN9_XY_FAST_COPY_BLT_CMD	(2 << 29 | 0x42 << 22)
+#define XY_SRC_COPY_BLT_CMD		(2 << 29 | 0x53 << 22)
+#define XY_MONO_SRC_COPY_IMM_BLT	(2 << 29 | 0x71 << 22 | 5)
+#define   BLT_WRITE_A			(2<<20)
+#define   BLT_WRITE_RGB			(1<<20)
+#define   BLT_WRITE_RGBA		(BLT_WRITE_RGB | BLT_WRITE_A)
+#define   BLT_DEPTH_8			(0<<24)
+#define   BLT_DEPTH_16_565		(1<<24)
+#define   BLT_DEPTH_16_1555		(2<<24)
+#define   BLT_DEPTH_32			(3<<24)
+#define   BLT_ROP_SRC_COPY		(0xcc<<16)
+#define   BLT_ROP_COLOR_COPY		(0xf0<<16)
+#define XY_SRC_COPY_BLT_SRC_TILED	(1<<15) /* 965+ only */
+#define XY_SRC_COPY_BLT_DST_TILED	(1<<11) /* 965+ only */
+#define CMD_OP_DISPLAYBUFFER_INFO ((0x0<<29)|(0x14<<23)|2)
+#define   ASYNC_FLIP                (1<<22)
+#define   DISPLAY_PLANE_A           (0<<20)
+#define   DISPLAY_PLANE_B           (1<<20)
+#define GFX_OP_PIPE_CONTROL(len)	((0x3<<29)|(0x3<<27)|(0x2<<24)|((len)-2))
+#define   PIPE_CONTROL_COMMAND_CACHE_INVALIDATE		(1<<29) /* gen11+ */
+#define   PIPE_CONTROL_TILE_CACHE_FLUSH			(1<<28) /* gen11+ */
+#define   PIPE_CONTROL_FLUSH_L3				(1<<27)
+#define   PIPE_CONTROL_AMFS_FLUSH			(1<<25) /* gen12+ */
+#define   PIPE_CONTROL_GLOBAL_GTT_IVB			(1<<24) /* gen7+ */
+#define   PIPE_CONTROL_MMIO_WRITE			(1<<23)
+#define   PIPE_CONTROL_STORE_DATA_INDEX			(1<<21)
+#define   PIPE_CONTROL_CS_STALL				(1<<20)
+#define   PIPE_CONTROL_GLOBAL_SNAPSHOT_RESET		(1<<19)
+#define   PIPE_CONTROL_TLB_INVALIDATE			(1<<18)
+#define   PIPE_CONTROL_PSD_SYNC				(1<<17) /* gen11+ */
+#define   PIPE_CONTROL_MEDIA_STATE_CLEAR		(1<<16)
+#define   PIPE_CONTROL_WRITE_TIMESTAMP			(3<<14)
+#define   PIPE_CONTROL_QW_WRITE				(1<<14)
+#define   PIPE_CONTROL_POST_SYNC_OP_MASK                (3<<14)
+#define   PIPE_CONTROL_DEPTH_STALL			(1<<13)
+#define   PIPE_CONTROL_WRITE_FLUSH			(1<<12)
+#define   PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH	(1<<12) /* gen6+ */
+#define   PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE	(1<<11) /* MBZ on ILK */
+#define   PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE		(1<<10) /* GM45+ only */
+#define   PIPE_CONTROL_INDIRECT_STATE_DISABLE		(1<<9)
+#define   PIPE_CONTROL0_HDC_PIPELINE_FLUSH		REG_BIT(9)  /* gen12 */
+#define   PIPE_CONTROL_NOTIFY				(1<<8)
+#define   PIPE_CONTROL_FLUSH_ENABLE			(1<<7) /* gen7+ */
+#define   PIPE_CONTROL_DC_FLUSH_ENABLE			(1<<5)
+#define   PIPE_CONTROL_VF_CACHE_INVALIDATE		(1<<4)
+#define   PIPE_CONTROL_CONST_CACHE_INVALIDATE		(1<<3)
+#define   PIPE_CONTROL_STATE_CACHE_INVALIDATE		(1<<2)
+#define   PIPE_CONTROL_STALL_AT_SCOREBOARD		(1<<1)
+#define   PIPE_CONTROL_DEPTH_CACHE_FLUSH		(1<<0)
+#define   PIPE_CONTROL_GLOBAL_GTT (1<<2) /* in addr dword */
+
+/*
+ * 3D-related flags that can't be set on _engines_ that lack access to the 3D
+ * pipeline (i.e., CCS engines).
+ */
+#define PIPE_CONTROL_3D_ENGINE_FLAGS (\
+		PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH | \
+		PIPE_CONTROL_DEPTH_CACHE_FLUSH | \
+		PIPE_CONTROL_TILE_CACHE_FLUSH | \
+		PIPE_CONTROL_DEPTH_STALL | \
+		PIPE_CONTROL_STALL_AT_SCOREBOARD | \
+		PIPE_CONTROL_PSD_SYNC | \
+		PIPE_CONTROL_AMFS_FLUSH | \
+		PIPE_CONTROL_VF_CACHE_INVALIDATE | \
+		PIPE_CONTROL_GLOBAL_SNAPSHOT_RESET)
+
+/* 3D-related flags that can't be set on _platforms_ that lack a 3D pipeline */
+#define PIPE_CONTROL_3D_ARCH_FLAGS ( \
+		PIPE_CONTROL_3D_ENGINE_FLAGS | \
+		PIPE_CONTROL_INDIRECT_STATE_DISABLE | \
+		PIPE_CONTROL_FLUSH_ENABLE | \
+		PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE | \
+		PIPE_CONTROL_DC_FLUSH_ENABLE)
+
+#define MI_MATH(x)			MI_INSTR(0x1a, (x) - 1)
+#define MI_MATH_INSTR(opcode, op1, op2) ((opcode) << 20 | (op1) << 10 | (op2))
+/* Opcodes for MI_MATH_INSTR */
+#define   MI_MATH_NOOP			MI_MATH_INSTR(0x000, 0x0, 0x0)
+#define   MI_MATH_LOAD(op1, op2)	MI_MATH_INSTR(0x080, op1, op2)
+#define   MI_MATH_LOADINV(op1, op2)	MI_MATH_INSTR(0x480, op1, op2)
+#define   MI_MATH_LOAD0(op1)		MI_MATH_INSTR(0x081, op1)
+#define   MI_MATH_LOAD1(op1)		MI_MATH_INSTR(0x481, op1)
+#define   MI_MATH_ADD			MI_MATH_INSTR(0x100, 0x0, 0x0)
+#define   MI_MATH_SUB			MI_MATH_INSTR(0x101, 0x0, 0x0)
+#define   MI_MATH_AND			MI_MATH_INSTR(0x102, 0x0, 0x0)
+#define   MI_MATH_OR			MI_MATH_INSTR(0x103, 0x0, 0x0)
+#define   MI_MATH_XOR			MI_MATH_INSTR(0x104, 0x0, 0x0)
+#define   MI_MATH_STORE(op1, op2)	MI_MATH_INSTR(0x180, op1, op2)
+#define   MI_MATH_STOREINV(op1, op2)	MI_MATH_INSTR(0x580, op1, op2)
+/* Registers used as operands in MI_MATH_INSTR */
+#define   MI_MATH_REG(x)		(x)
+#define   MI_MATH_REG_SRCA		0x20
+#define   MI_MATH_REG_SRCB		0x21
+#define   MI_MATH_REG_ACCU		0x31
+#define   MI_MATH_REG_ZF		0x32
+#define   MI_MATH_REG_CF		0x33
+
+/*
+ * Media instructions used by the kernel
+ */
+#define MEDIA_INSTR(pipe, op, sub_op, flags) \
+	(__INSTR(INSTR_RC_CLIENT) | (pipe) << INSTR_SUBCLIENT_SHIFT | \
+	(op) << INSTR_26_TO_24_SHIFT | (sub_op) << 16 | (flags))
+
+#define MFX_WAIT				MEDIA_INSTR(1, 0, 0, 0)
+#define  MFX_WAIT_DW0_MFX_SYNC_CONTROL_FLAG	REG_BIT(8)
+#define  MFX_WAIT_DW0_PXP_SYNC_CONTROL_FLAG	REG_BIT(9)
+
+#define CRYPTO_KEY_EXCHANGE			MEDIA_INSTR(2, 6, 9, 0)
+
+/*
+ * Commands used only by the command parser
+ */
+#define MI_SET_PREDICATE        MI_INSTR(0x01, 0)
+#define MI_ARB_CHECK            MI_INSTR(0x05, 0)
+#define MI_RS_CONTROL           MI_INSTR(0x06, 0)
+#define MI_URB_ATOMIC_ALLOC     MI_INSTR(0x09, 0)
+#define MI_PREDICATE            MI_INSTR(0x0C, 0)
+#define MI_RS_CONTEXT           MI_INSTR(0x0F, 0)
+#define MI_TOPOLOGY_FILTER      MI_INSTR(0x0D, 0)
+#define MI_LOAD_SCAN_LINES_EXCL MI_INSTR(0x13, 0)
+#define MI_URB_CLEAR            MI_INSTR(0x19, 0)
+#define MI_UPDATE_GTT           MI_INSTR(0x23, 0)
+#define MI_CLFLUSH              MI_INSTR(0x27, 0)
+#define MI_REPORT_PERF_COUNT    MI_INSTR(0x28, 0)
+#define   MI_REPORT_PERF_COUNT_GGTT (1<<0)
+#define MI_RS_STORE_DATA_IMM    MI_INSTR(0x2B, 0)
+#define MI_LOAD_URB_MEM         MI_INSTR(0x2C, 0)
+#define MI_STORE_URB_MEM        MI_INSTR(0x2D, 0)
+#define MI_CONDITIONAL_BATCH_BUFFER_END MI_INSTR(0x36, 0)
+
+#define STATE_BASE_ADDRESS \
+	((0x3 << 29) | (0x0 << 27) | (0x1 << 24) | (0x1 << 16))
+#define BASE_ADDRESS_MODIFY		REG_BIT(0)
+#define PIPELINE_SELECT \
+	((0x3 << 29) | (0x1 << 27) | (0x1 << 24) | (0x4 << 16))
+#define PIPELINE_SELECT_MEDIA	       REG_BIT(0)
+#define GFX_OP_3DSTATE_VF_STATISTICS \
+	((0x3 << 29) | (0x1 << 27) | (0x0 << 24) | (0xB << 16))
+#define MEDIA_VFE_STATE \
+	((0x3 << 29) | (0x2 << 27) | (0x0 << 24) | (0x0 << 16))
+#define  MEDIA_VFE_STATE_MMIO_ACCESS_MASK (0x18)
+#define MEDIA_INTERFACE_DESCRIPTOR_LOAD \
+	((0x3 << 29) | (0x2 << 27) | (0x0 << 24) | (0x2 << 16))
+#define MEDIA_OBJECT \
+	((0x3 << 29) | (0x2 << 27) | (0x1 << 24) | (0x0 << 16))
+#define GPGPU_OBJECT                   ((0x3<<29)|(0x2<<27)|(0x1<<24)|(0x4<<16))
+#define GPGPU_WALKER                   ((0x3<<29)|(0x2<<27)|(0x1<<24)|(0x5<<16))
+#define GFX_OP_3DSTATE_DX9_CONSTANTF_VS \
+	((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x39<<16))
+#define GFX_OP_3DSTATE_DX9_CONSTANTF_PS \
+	((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x3A<<16))
+#define GFX_OP_3DSTATE_SO_DECL_LIST \
+	((0x3<<29)|(0x3<<27)|(0x1<<24)|(0x17<<16))
+
+#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_VS \
+	((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x43<<16))
+#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_GS \
+	((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x44<<16))
+#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_HS \
+	((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x45<<16))
+#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_DS \
+	((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x46<<16))
+#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_PS \
+	((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x47<<16))
+
+#define COLOR_BLT     ((0x2<<29)|(0x40<<22))
+#define SRC_COPY_BLT  ((0x2<<29)|(0x43<<22))
+
+/*
+ * Used to convert any address to canonical form.
+ * Starting from gen8, some commands (e.g. STATE_BASE_ADDRESS,
+ * MI_LOAD_REGISTER_MEM and others, see Broadwell PRM Vol2a) require the
+ * addresses to be in a canonical form:
+ * "GraphicsAddress[63:48] are ignored by the HW and assumed to be in correct
+ * canonical form [63:48] == [47]."
+ */
+#define GEN8_HIGH_ADDRESS_BIT 47
+static inline u64 gen8_canonical_addr(u64 address)
+{
+	return sign_extend64(address, GEN8_HIGH_ADDRESS_BIT);
+}
+
+static inline u64 gen8_noncanonical_addr(u64 address)
+{
+	return address & GENMASK_ULL(GEN8_HIGH_ADDRESS_BIT, 0);
+}
+
+static inline u32 *__gen6_emit_bb_start(u32 *cs, u32 addr, unsigned int flags)
+{
+	*cs++ = MI_BATCH_BUFFER_START | flags;
+	*cs++ = addr;
+
+	return cs;
+}
+
+#endif /* _INTEL_GPU_COMMANDS_H_ */
diff --git a/drivers/gpu/drm/i915/gt/intel_gsc.c b/drivers/gpu/drm/i915/gt/intel_gsc.c
new file mode 100644
index 000000000..7af6db319
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gsc.c
@@ -0,0 +1,310 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright(c) 2019-2022, Intel Corporation. All rights reserved.
+ */
+
+#include <linux/irq.h>
+#include <linux/mei_aux.h>
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "gem/i915_gem_region.h"
+#include "gt/intel_gsc.h"
+#include "gt/intel_gt.h"
+
+#define GSC_BAR_LENGTH  0x00000FFC
+
+static void gsc_irq_mask(struct irq_data *d)
+{
+	/* generic irq handling */
+}
+
+static void gsc_irq_unmask(struct irq_data *d)
+{
+	/* generic irq handling */
+}
+
+static struct irq_chip gsc_irq_chip = {
+	.name = "gsc_irq_chip",
+	.irq_mask = gsc_irq_mask,
+	.irq_unmask = gsc_irq_unmask,
+};
+
+static int gsc_irq_init(int irq)
+{
+	irq_set_chip_and_handler_name(irq, &gsc_irq_chip,
+				      handle_simple_irq, "gsc_irq_handler");
+
+	return irq_set_chip_data(irq, NULL);
+}
+
+static int
+gsc_ext_om_alloc(struct intel_gsc *gsc, struct intel_gsc_intf *intf, size_t size)
+{
+	struct intel_gt *gt = gsc_to_gt(gsc);
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	obj = i915_gem_object_create_lmem(gt->i915, size,
+					  I915_BO_ALLOC_CONTIGUOUS |
+					  I915_BO_ALLOC_CPU_CLEAR);
+	if (IS_ERR(obj)) {
+		drm_err(&gt->i915->drm, "Failed to allocate gsc memory\n");
+		return PTR_ERR(obj);
+	}
+
+	err = i915_gem_object_pin_pages_unlocked(obj);
+	if (err) {
+		drm_err(&gt->i915->drm, "Failed to pin pages for gsc memory\n");
+		goto out_put;
+	}
+
+	intf->gem_obj = obj;
+
+	return 0;
+
+out_put:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static void gsc_ext_om_destroy(struct intel_gsc_intf *intf)
+{
+	struct drm_i915_gem_object *obj = fetch_and_zero(&intf->gem_obj);
+
+	if (!obj)
+		return;
+
+	if (i915_gem_object_has_pinned_pages(obj))
+		i915_gem_object_unpin_pages(obj);
+
+	i915_gem_object_put(obj);
+}
+
+struct gsc_def {
+	const char *name;
+	unsigned long bar;
+	size_t bar_size;
+	bool use_polling;
+	bool slow_firmware;
+	size_t lmem_size;
+};
+
+/* gsc resources and definitions (HECI1 and HECI2) */
+static const struct gsc_def gsc_def_dg1[] = {
+	{
+		/* HECI1 not yet implemented. */
+	},
+	{
+		.name = "mei-gscfi",
+		.bar = DG1_GSC_HECI2_BASE,
+		.bar_size = GSC_BAR_LENGTH,
+	}
+};
+
+static const struct gsc_def gsc_def_xehpsdv[] = {
+	{
+		/* HECI1 not enabled on the device. */
+	},
+	{
+		.name = "mei-gscfi",
+		.bar = DG1_GSC_HECI2_BASE,
+		.bar_size = GSC_BAR_LENGTH,
+		.use_polling = true,
+		.slow_firmware = true,
+	}
+};
+
+static const struct gsc_def gsc_def_dg2[] = {
+	{
+		.name = "mei-gsc",
+		.bar = DG2_GSC_HECI1_BASE,
+		.bar_size = GSC_BAR_LENGTH,
+		.lmem_size = SZ_4M,
+	},
+	{
+		.name = "mei-gscfi",
+		.bar = DG2_GSC_HECI2_BASE,
+		.bar_size = GSC_BAR_LENGTH,
+	}
+};
+
+static void gsc_release_dev(struct device *dev)
+{
+	struct auxiliary_device *aux_dev = to_auxiliary_dev(dev);
+	struct mei_aux_device *adev = auxiliary_dev_to_mei_aux_dev(aux_dev);
+
+	kfree(adev);
+}
+
+static void gsc_destroy_one(struct drm_i915_private *i915,
+			    struct intel_gsc *gsc, unsigned int intf_id)
+{
+	struct intel_gsc_intf *intf = &gsc->intf[intf_id];
+
+	if (intf->adev) {
+		auxiliary_device_delete(&intf->adev->aux_dev);
+		auxiliary_device_uninit(&intf->adev->aux_dev);
+		intf->adev = NULL;
+	}
+
+	if (intf->irq >= 0)
+		irq_free_desc(intf->irq);
+	intf->irq = -1;
+
+	gsc_ext_om_destroy(intf);
+}
+
+static void gsc_init_one(struct drm_i915_private *i915, struct intel_gsc *gsc,
+			 unsigned int intf_id)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	struct mei_aux_device *adev;
+	struct auxiliary_device *aux_dev;
+	const struct gsc_def *def;
+	struct intel_gsc_intf *intf = &gsc->intf[intf_id];
+	int ret;
+
+	intf->irq = -1;
+	intf->id = intf_id;
+
+	if (intf_id == 0 && !HAS_HECI_PXP(i915))
+		return;
+
+	if (IS_DG1(i915)) {
+		def = &gsc_def_dg1[intf_id];
+	} else if (IS_XEHPSDV(i915)) {
+		def = &gsc_def_xehpsdv[intf_id];
+	} else if (IS_DG2(i915)) {
+		def = &gsc_def_dg2[intf_id];
+	} else {
+		drm_warn_once(&i915->drm, "Unknown platform\n");
+		return;
+	}
+
+	if (!def->name) {
+		drm_warn_once(&i915->drm, "HECI%d is not implemented!\n", intf_id + 1);
+		return;
+	}
+
+	/* skip irq initialization */
+	if (def->use_polling)
+		goto add_device;
+
+	intf->irq = irq_alloc_desc(0);
+	if (intf->irq < 0) {
+		drm_err(&i915->drm, "gsc irq error %d\n", intf->irq);
+		goto fail;
+	}
+
+	ret = gsc_irq_init(intf->irq);
+	if (ret < 0) {
+		drm_err(&i915->drm, "gsc irq init failed %d\n", ret);
+		goto fail;
+	}
+
+add_device:
+	adev = kzalloc(sizeof(*adev), GFP_KERNEL);
+	if (!adev)
+		goto fail;
+
+	if (def->lmem_size) {
+		drm_dbg(&i915->drm, "setting up GSC lmem\n");
+
+		if (gsc_ext_om_alloc(gsc, intf, def->lmem_size)) {
+			drm_err(&i915->drm, "setting up gsc extended operational memory failed\n");
+			kfree(adev);
+			goto fail;
+		}
+
+		adev->ext_op_mem.start = i915_gem_object_get_dma_address(intf->gem_obj, 0);
+		adev->ext_op_mem.end = adev->ext_op_mem.start + def->lmem_size;
+	}
+
+	adev->irq = intf->irq;
+	adev->bar.parent = &pdev->resource[0];
+	adev->bar.start = def->bar + pdev->resource[0].start;
+	adev->bar.end = adev->bar.start + def->bar_size - 1;
+	adev->bar.flags = IORESOURCE_MEM;
+	adev->bar.desc = IORES_DESC_NONE;
+	adev->slow_firmware = def->slow_firmware;
+
+	aux_dev = &adev->aux_dev;
+	aux_dev->name = def->name;
+	aux_dev->id = (pci_domain_nr(pdev->bus) << 16) |
+		      PCI_DEVID(pdev->bus->number, pdev->devfn);
+	aux_dev->dev.parent = &pdev->dev;
+	aux_dev->dev.release = gsc_release_dev;
+
+	ret = auxiliary_device_init(aux_dev);
+	if (ret < 0) {
+		drm_err(&i915->drm, "gsc aux init failed %d\n", ret);
+		kfree(adev);
+		goto fail;
+	}
+
+	ret = auxiliary_device_add(aux_dev);
+	if (ret < 0) {
+		drm_err(&i915->drm, "gsc aux add failed %d\n", ret);
+		/* adev will be freed with the put_device() and .release sequence */
+		auxiliary_device_uninit(aux_dev);
+		goto fail;
+	}
+	intf->adev = adev;
+
+	return;
+fail:
+	gsc_destroy_one(i915, gsc, intf->id);
+}
+
+static void gsc_irq_handler(struct intel_gt *gt, unsigned int intf_id)
+{
+	int ret;
+
+	if (intf_id >= INTEL_GSC_NUM_INTERFACES) {
+		drm_warn_once(&gt->i915->drm, "GSC irq: intf_id %d is out of range", intf_id);
+		return;
+	}
+
+	if (!HAS_HECI_GSC(gt->i915)) {
+		drm_warn_once(&gt->i915->drm, "GSC irq: not supported");
+		return;
+	}
+
+	if (gt->gsc.intf[intf_id].irq < 0)
+		return;
+
+	ret = generic_handle_irq(gt->gsc.intf[intf_id].irq);
+	if (ret)
+		drm_err_ratelimited(&gt->i915->drm, "error handling GSC irq: %d\n", ret);
+}
+
+void intel_gsc_irq_handler(struct intel_gt *gt, u32 iir)
+{
+	if (iir & GSC_IRQ_INTF(0))
+		gsc_irq_handler(gt, 0);
+	if (iir & GSC_IRQ_INTF(1))
+		gsc_irq_handler(gt, 1);
+}
+
+void intel_gsc_init(struct intel_gsc *gsc, struct drm_i915_private *i915)
+{
+	unsigned int i;
+
+	if (!HAS_HECI_GSC(i915))
+		return;
+
+	for (i = 0; i < INTEL_GSC_NUM_INTERFACES; i++)
+		gsc_init_one(i915, gsc, i);
+}
+
+void intel_gsc_fini(struct intel_gsc *gsc)
+{
+	struct intel_gt *gt = gsc_to_gt(gsc);
+	unsigned int i;
+
+	if (!HAS_HECI_GSC(gt->i915))
+		return;
+
+	for (i = 0; i < INTEL_GSC_NUM_INTERFACES; i++)
+		gsc_destroy_one(gt->i915, gsc, i);
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gsc.h b/drivers/gpu/drm/i915/gt/intel_gsc.h
new file mode 100644
index 000000000..fcac1775e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gsc.h
@@ -0,0 +1,40 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright(c) 2019-2022, Intel Corporation. All rights reserved.
+ */
+#ifndef __INTEL_GSC_DEV_H__
+#define __INTEL_GSC_DEV_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct intel_gt;
+struct mei_aux_device;
+
+#define INTEL_GSC_NUM_INTERFACES 2
+/*
+ * The HECI1 bit corresponds to bit15 and HECI2 to bit14.
+ * The reason for this is to allow growth for more interfaces in the future.
+ */
+#define GSC_IRQ_INTF(_x)  BIT(15 - (_x))
+
+/**
+ * struct intel_gsc - graphics security controller
+ *
+ * @gem_obj: scratch memory GSC operations
+ * @intf : gsc interface
+ */
+struct intel_gsc {
+	struct intel_gsc_intf {
+		struct mei_aux_device *adev;
+		struct drm_i915_gem_object *gem_obj;
+		int irq;
+		unsigned int id;
+	} intf[INTEL_GSC_NUM_INTERFACES];
+};
+
+void intel_gsc_init(struct intel_gsc *gsc, struct drm_i915_private *dev_priv);
+void intel_gsc_fini(struct intel_gsc *gsc);
+void intel_gsc_irq_handler(struct intel_gt *gt, u32 iir);
+
+#endif /* __INTEL_GSC_DEV_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt.c b/drivers/gpu/drm/i915/gt/intel_gt.c
new file mode 100644
index 000000000..91a005c46
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt.c
@@ -0,0 +1,1095 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <drm/drm_managed.h>
+#include <drm/intel-gtt.h>
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_lmem.h"
+#include "pxp/intel_pxp.h"
+
+#include "i915_drv.h"
+#include "i915_perf_oa_regs.h"
+#include "intel_context.h"
+#include "intel_engine_pm.h"
+#include "intel_engine_regs.h"
+#include "intel_ggtt_gmch.h"
+#include "intel_gt.h"
+#include "intel_gt_buffer_pool.h"
+#include "intel_gt_clock_utils.h"
+#include "intel_gt_debugfs.h"
+#include "intel_gt_mcr.h"
+#include "intel_gt_pm.h"
+#include "intel_gt_regs.h"
+#include "intel_gt_requests.h"
+#include "intel_migrate.h"
+#include "intel_mocs.h"
+#include "intel_pci_config.h"
+#include "intel_pm.h"
+#include "intel_rc6.h"
+#include "intel_renderstate.h"
+#include "intel_rps.h"
+#include "intel_sa_media.h"
+#include "intel_gt_sysfs.h"
+#include "intel_uncore.h"
+#include "shmem_utils.h"
+
+void intel_gt_common_init_early(struct intel_gt *gt)
+{
+	spin_lock_init(gt->irq_lock);
+
+	INIT_LIST_HEAD(&gt->closed_vma);
+	spin_lock_init(&gt->closed_lock);
+
+	init_llist_head(&gt->watchdog.list);
+	INIT_WORK(&gt->watchdog.work, intel_gt_watchdog_work);
+
+	intel_gt_init_buffer_pool(gt);
+	intel_gt_init_reset(gt);
+	intel_gt_init_requests(gt);
+	intel_gt_init_timelines(gt);
+	mutex_init(&gt->tlb.invalidate_lock);
+	seqcount_mutex_init(&gt->tlb.seqno, &gt->tlb.invalidate_lock);
+	intel_gt_pm_init_early(gt);
+
+	intel_uc_init_early(&gt->uc);
+	intel_rps_init_early(&gt->rps);
+}
+
+/* Preliminary initialization of Tile 0 */
+int intel_root_gt_init_early(struct drm_i915_private *i915)
+{
+	struct intel_gt *gt = to_gt(i915);
+
+	gt->i915 = i915;
+	gt->uncore = &i915->uncore;
+	gt->irq_lock = drmm_kzalloc(&i915->drm, sizeof(*gt->irq_lock), GFP_KERNEL);
+	if (!gt->irq_lock)
+		return -ENOMEM;
+
+	intel_gt_common_init_early(gt);
+
+	return 0;
+}
+
+static int intel_gt_probe_lmem(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	unsigned int instance = gt->info.id;
+	int id = INTEL_REGION_LMEM_0 + instance;
+	struct intel_memory_region *mem;
+	int err;
+
+	mem = intel_gt_setup_lmem(gt);
+	if (IS_ERR(mem)) {
+		err = PTR_ERR(mem);
+		if (err == -ENODEV)
+			return 0;
+
+		drm_err(&i915->drm,
+			"Failed to setup region(%d) type=%d\n",
+			err, INTEL_MEMORY_LOCAL);
+		return err;
+	}
+
+	mem->id = id;
+	mem->instance = instance;
+
+	intel_memory_region_set_name(mem, "local%u", mem->instance);
+
+	GEM_BUG_ON(!HAS_REGION(i915, id));
+	GEM_BUG_ON(i915->mm.regions[id]);
+	i915->mm.regions[id] = mem;
+
+	return 0;
+}
+
+int intel_gt_assign_ggtt(struct intel_gt *gt)
+{
+	gt->ggtt = drmm_kzalloc(&gt->i915->drm, sizeof(*gt->ggtt), GFP_KERNEL);
+
+	return gt->ggtt ? 0 : -ENOMEM;
+}
+
+int intel_gt_init_mmio(struct intel_gt *gt)
+{
+	intel_gt_init_clock_frequency(gt);
+
+	intel_uc_init_mmio(&gt->uc);
+	intel_sseu_info_init(gt);
+	intel_gt_mcr_init(gt);
+
+	return intel_engines_init_mmio(gt);
+}
+
+static void init_unused_ring(struct intel_gt *gt, u32 base)
+{
+	struct intel_uncore *uncore = gt->uncore;
+
+	intel_uncore_write(uncore, RING_CTL(base), 0);
+	intel_uncore_write(uncore, RING_HEAD(base), 0);
+	intel_uncore_write(uncore, RING_TAIL(base), 0);
+	intel_uncore_write(uncore, RING_START(base), 0);
+}
+
+static void init_unused_rings(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	if (IS_I830(i915)) {
+		init_unused_ring(gt, PRB1_BASE);
+		init_unused_ring(gt, SRB0_BASE);
+		init_unused_ring(gt, SRB1_BASE);
+		init_unused_ring(gt, SRB2_BASE);
+		init_unused_ring(gt, SRB3_BASE);
+	} else if (GRAPHICS_VER(i915) == 2) {
+		init_unused_ring(gt, SRB0_BASE);
+		init_unused_ring(gt, SRB1_BASE);
+	} else if (GRAPHICS_VER(i915) == 3) {
+		init_unused_ring(gt, PRB1_BASE);
+		init_unused_ring(gt, PRB2_BASE);
+	}
+}
+
+int intel_gt_init_hw(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+	int ret;
+
+	gt->last_init_time = ktime_get();
+
+	/* Double layer security blanket, see i915_gem_init() */
+	intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+
+	if (HAS_EDRAM(i915) && GRAPHICS_VER(i915) < 9)
+		intel_uncore_rmw(uncore, HSW_IDICR, 0, IDIHASHMSK(0xf));
+
+	if (IS_HASWELL(i915))
+		intel_uncore_write(uncore,
+				   HSW_MI_PREDICATE_RESULT_2,
+				   IS_HSW_GT3(i915) ?
+				   LOWER_SLICE_ENABLED : LOWER_SLICE_DISABLED);
+
+	/* Apply the GT workarounds... */
+	intel_gt_apply_workarounds(gt);
+	/* ...and determine whether they are sticking. */
+	intel_gt_verify_workarounds(gt, "init");
+
+	intel_gt_init_swizzling(gt);
+
+	/*
+	 * At least 830 can leave some of the unused rings
+	 * "active" (ie. head != tail) after resume which
+	 * will prevent c3 entry. Makes sure all unused rings
+	 * are totally idle.
+	 */
+	init_unused_rings(gt);
+
+	ret = i915_ppgtt_init_hw(gt);
+	if (ret) {
+		DRM_ERROR("Enabling PPGTT failed (%d)\n", ret);
+		goto out;
+	}
+
+	/* We can't enable contexts until all firmware is loaded */
+	ret = intel_uc_init_hw(&gt->uc);
+	if (ret) {
+		i915_probe_error(i915, "Enabling uc failed (%d)\n", ret);
+		goto out;
+	}
+
+	intel_mocs_init(gt);
+
+out:
+	intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
+	return ret;
+}
+
+static void rmw_set(struct intel_uncore *uncore, i915_reg_t reg, u32 set)
+{
+	intel_uncore_rmw(uncore, reg, 0, set);
+}
+
+static void rmw_clear(struct intel_uncore *uncore, i915_reg_t reg, u32 clr)
+{
+	intel_uncore_rmw(uncore, reg, clr, 0);
+}
+
+static void clear_register(struct intel_uncore *uncore, i915_reg_t reg)
+{
+	intel_uncore_rmw(uncore, reg, 0, 0);
+}
+
+static void gen6_clear_engine_error_register(struct intel_engine_cs *engine)
+{
+	GEN6_RING_FAULT_REG_RMW(engine, RING_FAULT_VALID, 0);
+	GEN6_RING_FAULT_REG_POSTING_READ(engine);
+}
+
+void
+intel_gt_clear_error_registers(struct intel_gt *gt,
+			       intel_engine_mask_t engine_mask)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+	u32 eir;
+
+	if (GRAPHICS_VER(i915) != 2)
+		clear_register(uncore, PGTBL_ER);
+
+	if (GRAPHICS_VER(i915) < 4)
+		clear_register(uncore, IPEIR(RENDER_RING_BASE));
+	else
+		clear_register(uncore, IPEIR_I965);
+
+	clear_register(uncore, EIR);
+	eir = intel_uncore_read(uncore, EIR);
+	if (eir) {
+		/*
+		 * some errors might have become stuck,
+		 * mask them.
+		 */
+		DRM_DEBUG_DRIVER("EIR stuck: 0x%08x, masking\n", eir);
+		rmw_set(uncore, EMR, eir);
+		intel_uncore_write(uncore, GEN2_IIR,
+				   I915_MASTER_ERROR_INTERRUPT);
+	}
+
+	if (GRAPHICS_VER(i915) >= 12) {
+		rmw_clear(uncore, GEN12_RING_FAULT_REG, RING_FAULT_VALID);
+		intel_uncore_posting_read(uncore, GEN12_RING_FAULT_REG);
+	} else if (GRAPHICS_VER(i915) >= 8) {
+		rmw_clear(uncore, GEN8_RING_FAULT_REG, RING_FAULT_VALID);
+		intel_uncore_posting_read(uncore, GEN8_RING_FAULT_REG);
+	} else if (GRAPHICS_VER(i915) >= 6) {
+		struct intel_engine_cs *engine;
+		enum intel_engine_id id;
+
+		for_each_engine_masked(engine, gt, engine_mask, id)
+			gen6_clear_engine_error_register(engine);
+	}
+}
+
+static void gen6_check_faults(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	u32 fault;
+
+	for_each_engine(engine, gt, id) {
+		fault = GEN6_RING_FAULT_REG_READ(engine);
+		if (fault & RING_FAULT_VALID) {
+			drm_dbg(&engine->i915->drm, "Unexpected fault\n"
+				"\tAddr: 0x%08lx\n"
+				"\tAddress space: %s\n"
+				"\tSource ID: %d\n"
+				"\tType: %d\n",
+				fault & PAGE_MASK,
+				fault & RING_FAULT_GTTSEL_MASK ?
+				"GGTT" : "PPGTT",
+				RING_FAULT_SRCID(fault),
+				RING_FAULT_FAULT_TYPE(fault));
+		}
+	}
+}
+
+static void gen8_check_faults(struct intel_gt *gt)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	i915_reg_t fault_reg, fault_data0_reg, fault_data1_reg;
+	u32 fault;
+
+	if (GRAPHICS_VER(gt->i915) >= 12) {
+		fault_reg = GEN12_RING_FAULT_REG;
+		fault_data0_reg = GEN12_FAULT_TLB_DATA0;
+		fault_data1_reg = GEN12_FAULT_TLB_DATA1;
+	} else {
+		fault_reg = GEN8_RING_FAULT_REG;
+		fault_data0_reg = GEN8_FAULT_TLB_DATA0;
+		fault_data1_reg = GEN8_FAULT_TLB_DATA1;
+	}
+
+	fault = intel_uncore_read(uncore, fault_reg);
+	if (fault & RING_FAULT_VALID) {
+		u32 fault_data0, fault_data1;
+		u64 fault_addr;
+
+		fault_data0 = intel_uncore_read(uncore, fault_data0_reg);
+		fault_data1 = intel_uncore_read(uncore, fault_data1_reg);
+
+		fault_addr = ((u64)(fault_data1 & FAULT_VA_HIGH_BITS) << 44) |
+			     ((u64)fault_data0 << 12);
+
+		drm_dbg(&uncore->i915->drm, "Unexpected fault\n"
+			"\tAddr: 0x%08x_%08x\n"
+			"\tAddress space: %s\n"
+			"\tEngine ID: %d\n"
+			"\tSource ID: %d\n"
+			"\tType: %d\n",
+			upper_32_bits(fault_addr), lower_32_bits(fault_addr),
+			fault_data1 & FAULT_GTT_SEL ? "GGTT" : "PPGTT",
+			GEN8_RING_FAULT_ENGINE_ID(fault),
+			RING_FAULT_SRCID(fault),
+			RING_FAULT_FAULT_TYPE(fault));
+	}
+}
+
+void intel_gt_check_and_clear_faults(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	/* From GEN8 onwards we only have one 'All Engine Fault Register' */
+	if (GRAPHICS_VER(i915) >= 8)
+		gen8_check_faults(gt);
+	else if (GRAPHICS_VER(i915) >= 6)
+		gen6_check_faults(gt);
+	else
+		return;
+
+	intel_gt_clear_error_registers(gt, ALL_ENGINES);
+}
+
+void intel_gt_flush_ggtt_writes(struct intel_gt *gt)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	intel_wakeref_t wakeref;
+
+	/*
+	 * No actual flushing is required for the GTT write domain for reads
+	 * from the GTT domain. Writes to it "immediately" go to main memory
+	 * as far as we know, so there's no chipset flush. It also doesn't
+	 * land in the GPU render cache.
+	 *
+	 * However, we do have to enforce the order so that all writes through
+	 * the GTT land before any writes to the device, such as updates to
+	 * the GATT itself.
+	 *
+	 * We also have to wait a bit for the writes to land from the GTT.
+	 * An uncached read (i.e. mmio) seems to be ideal for the round-trip
+	 * timing. This issue has only been observed when switching quickly
+	 * between GTT writes and CPU reads from inside the kernel on recent hw,
+	 * and it appears to only affect discrete GTT blocks (i.e. on LLC
+	 * system agents we cannot reproduce this behaviour, until Cannonlake
+	 * that was!).
+	 */
+
+	wmb();
+
+	if (INTEL_INFO(gt->i915)->has_coherent_ggtt)
+		return;
+
+	intel_gt_chipset_flush(gt);
+
+	with_intel_runtime_pm_if_in_use(uncore->rpm, wakeref) {
+		unsigned long flags;
+
+		spin_lock_irqsave(&uncore->lock, flags);
+		intel_uncore_posting_read_fw(uncore,
+					     RING_HEAD(RENDER_RING_BASE));
+		spin_unlock_irqrestore(&uncore->lock, flags);
+	}
+}
+
+void intel_gt_chipset_flush(struct intel_gt *gt)
+{
+	wmb();
+	if (GRAPHICS_VER(gt->i915) < 6)
+		intel_ggtt_gmch_flush();
+}
+
+void intel_gt_driver_register(struct intel_gt *gt)
+{
+	intel_gsc_init(&gt->gsc, gt->i915);
+
+	intel_rps_driver_register(&gt->rps);
+
+	intel_gt_debugfs_register(gt);
+	intel_gt_sysfs_register(gt);
+}
+
+static int intel_gt_init_scratch(struct intel_gt *gt, unsigned int size)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int ret;
+
+	obj = i915_gem_object_create_lmem(i915, size,
+					  I915_BO_ALLOC_VOLATILE |
+					  I915_BO_ALLOC_GPU_ONLY);
+	if (IS_ERR(obj))
+		obj = i915_gem_object_create_stolen(i915, size);
+	if (IS_ERR(obj))
+		obj = i915_gem_object_create_internal(i915, size);
+	if (IS_ERR(obj)) {
+		drm_err(&i915->drm, "Failed to allocate scratch page\n");
+		return PTR_ERR(obj);
+	}
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto err_unref;
+	}
+
+	ret = i915_ggtt_pin(vma, NULL, 0, PIN_HIGH);
+	if (ret)
+		goto err_unref;
+
+	gt->scratch = i915_vma_make_unshrinkable(vma);
+
+	return 0;
+
+err_unref:
+	i915_gem_object_put(obj);
+	return ret;
+}
+
+static void intel_gt_fini_scratch(struct intel_gt *gt)
+{
+	i915_vma_unpin_and_release(&gt->scratch, 0);
+}
+
+static struct i915_address_space *kernel_vm(struct intel_gt *gt)
+{
+	if (INTEL_PPGTT(gt->i915) > INTEL_PPGTT_ALIASING)
+		return &i915_ppgtt_create(gt, I915_BO_ALLOC_PM_EARLY)->vm;
+	else
+		return i915_vm_get(&gt->ggtt->vm);
+}
+
+static int __engines_record_defaults(struct intel_gt *gt)
+{
+	struct i915_request *requests[I915_NUM_ENGINES] = {};
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * As we reset the gpu during very early sanitisation, the current
+	 * register state on the GPU should reflect its defaults values.
+	 * We load a context onto the hw (with restore-inhibit), then switch
+	 * over to a second context to save that default register state. We
+	 * can then prime every new context with that state so they all start
+	 * from the same default HW values.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		struct intel_renderstate so;
+		struct intel_context *ce;
+		struct i915_request *rq;
+
+		/* We must be able to switch to something! */
+		GEM_BUG_ON(!engine->kernel_context);
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out;
+		}
+
+		err = intel_renderstate_init(&so, ce);
+		if (err)
+			goto err;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_fini;
+		}
+
+		err = intel_engine_emit_ctx_wa(rq);
+		if (err)
+			goto err_rq;
+
+		err = intel_renderstate_emit(&so, rq);
+		if (err)
+			goto err_rq;
+
+err_rq:
+		requests[id] = i915_request_get(rq);
+		i915_request_add(rq);
+err_fini:
+		intel_renderstate_fini(&so, ce);
+err:
+		if (err) {
+			intel_context_put(ce);
+			goto out;
+		}
+	}
+
+	/* Flush the default context image to memory, and enable powersaving. */
+	if (intel_gt_wait_for_idle(gt, I915_GEM_IDLE_TIMEOUT) == -ETIME) {
+		err = -EIO;
+		goto out;
+	}
+
+	for (id = 0; id < ARRAY_SIZE(requests); id++) {
+		struct i915_request *rq;
+		struct file *state;
+
+		rq = requests[id];
+		if (!rq)
+			continue;
+
+		if (rq->fence.error) {
+			err = -EIO;
+			goto out;
+		}
+
+		GEM_BUG_ON(!test_bit(CONTEXT_ALLOC_BIT, &rq->context->flags));
+		if (!rq->context->state)
+			continue;
+
+		/* Keep a copy of the state's backing pages; free the obj */
+		state = shmem_create_from_object(rq->context->state->obj);
+		if (IS_ERR(state)) {
+			err = PTR_ERR(state);
+			goto out;
+		}
+		rq->engine->default_state = state;
+	}
+
+out:
+	/*
+	 * If we have to abandon now, we expect the engines to be idle
+	 * and ready to be torn-down. The quickest way we can accomplish
+	 * this is by declaring ourselves wedged.
+	 */
+	if (err)
+		intel_gt_set_wedged(gt);
+
+	for (id = 0; id < ARRAY_SIZE(requests); id++) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+
+		rq = requests[id];
+		if (!rq)
+			continue;
+
+		ce = rq->context;
+		i915_request_put(rq);
+		intel_context_put(ce);
+	}
+	return err;
+}
+
+static int __engines_verify_workarounds(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		if (intel_engine_verify_workarounds(engine, "load"))
+			err = -EIO;
+	}
+
+	/* Flush and restore the kernel context for safety */
+	if (intel_gt_wait_for_idle(gt, I915_GEM_IDLE_TIMEOUT) == -ETIME)
+		err = -EIO;
+
+	return err;
+}
+
+static void __intel_gt_disable(struct intel_gt *gt)
+{
+	intel_gt_set_wedged_on_fini(gt);
+
+	intel_gt_suspend_prepare(gt);
+	intel_gt_suspend_late(gt);
+
+	GEM_BUG_ON(intel_gt_pm_is_awake(gt));
+}
+
+int intel_gt_wait_for_idle(struct intel_gt *gt, long timeout)
+{
+	long remaining_timeout;
+
+	/* If the device is asleep, we have no requests outstanding */
+	if (!intel_gt_pm_is_awake(gt))
+		return 0;
+
+	while ((timeout = intel_gt_retire_requests_timeout(gt, timeout,
+							   &remaining_timeout)) > 0) {
+		cond_resched();
+		if (signal_pending(current))
+			return -EINTR;
+	}
+
+	if (timeout)
+		return timeout;
+
+	if (remaining_timeout < 0)
+		remaining_timeout = 0;
+
+	return intel_uc_wait_for_idle(&gt->uc, remaining_timeout);
+}
+
+int intel_gt_init(struct intel_gt *gt)
+{
+	int err;
+
+	err = i915_inject_probe_error(gt->i915, -ENODEV);
+	if (err)
+		return err;
+
+	intel_gt_init_workarounds(gt);
+
+	/*
+	 * This is just a security blanket to placate dragons.
+	 * On some systems, we very sporadically observe that the first TLBs
+	 * used by the CS may be stale, despite us poking the TLB reset. If
+	 * we hold the forcewake during initialisation these problems
+	 * just magically go away.
+	 */
+	intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);
+
+	err = intel_gt_init_scratch(gt,
+				    GRAPHICS_VER(gt->i915) == 2 ? SZ_256K : SZ_4K);
+	if (err)
+		goto out_fw;
+
+	intel_gt_pm_init(gt);
+
+	gt->vm = kernel_vm(gt);
+	if (!gt->vm) {
+		err = -ENOMEM;
+		goto err_pm;
+	}
+
+	intel_set_mocs_index(gt);
+
+	err = intel_engines_init(gt);
+	if (err)
+		goto err_engines;
+
+	err = intel_uc_init(&gt->uc);
+	if (err)
+		goto err_engines;
+
+	err = intel_gt_resume(gt);
+	if (err)
+		goto err_uc_init;
+
+	err = intel_gt_init_hwconfig(gt);
+	if (err)
+		drm_err(&gt->i915->drm, "Failed to retrieve hwconfig table: %pe\n",
+			ERR_PTR(err));
+
+	err = __engines_record_defaults(gt);
+	if (err)
+		goto err_gt;
+
+	err = __engines_verify_workarounds(gt);
+	if (err)
+		goto err_gt;
+
+	err = i915_inject_probe_error(gt->i915, -EIO);
+	if (err)
+		goto err_gt;
+
+	intel_uc_init_late(&gt->uc);
+
+	intel_migrate_init(&gt->migrate, gt);
+
+	intel_pxp_init(&gt->pxp);
+
+	goto out_fw;
+err_gt:
+	__intel_gt_disable(gt);
+	intel_uc_fini_hw(&gt->uc);
+err_uc_init:
+	intel_uc_fini(&gt->uc);
+err_engines:
+	intel_engines_release(gt);
+	i915_vm_put(fetch_and_zero(&gt->vm));
+err_pm:
+	intel_gt_pm_fini(gt);
+	intel_gt_fini_scratch(gt);
+out_fw:
+	if (err)
+		intel_gt_set_wedged_on_init(gt);
+	intel_uncore_forcewake_put(gt->uncore, FORCEWAKE_ALL);
+	return err;
+}
+
+void intel_gt_driver_remove(struct intel_gt *gt)
+{
+	__intel_gt_disable(gt);
+
+	intel_migrate_fini(&gt->migrate);
+	intel_uc_driver_remove(&gt->uc);
+
+	intel_engines_release(gt);
+
+	intel_gt_flush_buffer_pool(gt);
+}
+
+void intel_gt_driver_unregister(struct intel_gt *gt)
+{
+	intel_wakeref_t wakeref;
+
+	intel_gt_sysfs_unregister(gt);
+	intel_rps_driver_unregister(&gt->rps);
+	intel_gsc_fini(&gt->gsc);
+
+	intel_pxp_fini(&gt->pxp);
+
+	/*
+	 * Upon unregistering the device to prevent any new users, cancel
+	 * all in-flight requests so that we can quickly unbind the active
+	 * resources.
+	 */
+	intel_gt_set_wedged_on_fini(gt);
+
+	/* Scrub all HW state upon release */
+	with_intel_runtime_pm(gt->uncore->rpm, wakeref)
+		__intel_gt_reset(gt, ALL_ENGINES);
+}
+
+void intel_gt_driver_release(struct intel_gt *gt)
+{
+	struct i915_address_space *vm;
+
+	vm = fetch_and_zero(&gt->vm);
+	if (vm) /* FIXME being called twice on error paths :( */
+		i915_vm_put(vm);
+
+	intel_wa_list_free(&gt->wa_list);
+	intel_gt_pm_fini(gt);
+	intel_gt_fini_scratch(gt);
+	intel_gt_fini_buffer_pool(gt);
+	intel_gt_fini_hwconfig(gt);
+}
+
+void intel_gt_driver_late_release_all(struct drm_i915_private *i915)
+{
+	struct intel_gt *gt;
+	unsigned int id;
+
+	/* We need to wait for inflight RCU frees to release their grip */
+	rcu_barrier();
+
+	for_each_gt(gt, i915, id) {
+		intel_uc_driver_late_release(&gt->uc);
+		intel_gt_fini_requests(gt);
+		intel_gt_fini_reset(gt);
+		intel_gt_fini_timelines(gt);
+		mutex_destroy(&gt->tlb.invalidate_lock);
+		intel_engines_free(gt);
+	}
+}
+
+static int intel_gt_tile_setup(struct intel_gt *gt, phys_addr_t phys_addr)
+{
+	int ret;
+
+	if (!gt_is_root(gt)) {
+		struct intel_uncore *uncore;
+		spinlock_t *irq_lock;
+
+		uncore = drmm_kzalloc(&gt->i915->drm, sizeof(*uncore), GFP_KERNEL);
+		if (!uncore)
+			return -ENOMEM;
+
+		irq_lock = drmm_kzalloc(&gt->i915->drm, sizeof(*irq_lock), GFP_KERNEL);
+		if (!irq_lock)
+			return -ENOMEM;
+
+		gt->uncore = uncore;
+		gt->irq_lock = irq_lock;
+
+		intel_gt_common_init_early(gt);
+	}
+
+	intel_uncore_init_early(gt->uncore, gt);
+
+	ret = intel_uncore_setup_mmio(gt->uncore, phys_addr);
+	if (ret)
+		return ret;
+
+	gt->phys_addr = phys_addr;
+
+	return 0;
+}
+
+int intel_gt_probe_all(struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	struct intel_gt *gt = &i915->gt0;
+	const struct intel_gt_definition *gtdef;
+	phys_addr_t phys_addr;
+	unsigned int mmio_bar;
+	unsigned int i;
+	int ret;
+
+	mmio_bar = GRAPHICS_VER(i915) == 2 ? GEN2_GTTMMADR_BAR : GTTMMADR_BAR;
+	phys_addr = pci_resource_start(pdev, mmio_bar);
+
+	/*
+	 * We always have at least one primary GT on any device
+	 * and it has been already initialized early during probe
+	 * in i915_driver_probe()
+	 */
+	gt->i915 = i915;
+	gt->name = "Primary GT";
+	gt->info.engine_mask = RUNTIME_INFO(i915)->platform_engine_mask;
+
+	drm_dbg(&i915->drm, "Setting up %s\n", gt->name);
+	ret = intel_gt_tile_setup(gt, phys_addr);
+	if (ret)
+		return ret;
+
+	i915->gt[0] = gt;
+
+	if (!HAS_EXTRA_GT_LIST(i915))
+		return 0;
+
+	for (i = 1, gtdef = &INTEL_INFO(i915)->extra_gt_list[i - 1];
+	     gtdef->name != NULL;
+	     i++, gtdef = &INTEL_INFO(i915)->extra_gt_list[i - 1]) {
+		gt = drmm_kzalloc(&i915->drm, sizeof(*gt), GFP_KERNEL);
+		if (!gt) {
+			ret = -ENOMEM;
+			goto err;
+		}
+
+		gt->i915 = i915;
+		gt->name = gtdef->name;
+		gt->type = gtdef->type;
+		gt->info.engine_mask = gtdef->engine_mask;
+		gt->info.id = i;
+
+		drm_dbg(&i915->drm, "Setting up %s\n", gt->name);
+		if (GEM_WARN_ON(range_overflows_t(resource_size_t,
+						  gtdef->mapping_base,
+						  SZ_16M,
+						  pci_resource_len(pdev, mmio_bar)))) {
+			ret = -ENODEV;
+			goto err;
+		}
+
+		switch (gtdef->type) {
+		case GT_TILE:
+			ret = intel_gt_tile_setup(gt, phys_addr + gtdef->mapping_base);
+			break;
+
+		case GT_MEDIA:
+			ret = intel_sa_mediagt_setup(gt, phys_addr + gtdef->mapping_base,
+						     gtdef->gsi_offset);
+			break;
+
+		case GT_PRIMARY:
+			/* Primary GT should not appear in extra GT list */
+		default:
+			MISSING_CASE(gtdef->type);
+			ret = -ENODEV;
+		}
+
+		if (ret)
+			goto err;
+
+		i915->gt[i] = gt;
+	}
+
+	return 0;
+
+err:
+	i915_probe_error(i915, "Failed to initialize %s! (%d)\n", gtdef->name, ret);
+	return ret;
+}
+
+int intel_gt_tiles_init(struct drm_i915_private *i915)
+{
+	struct intel_gt *gt;
+	unsigned int id;
+	int ret;
+
+	for_each_gt(gt, i915, id) {
+		ret = intel_gt_probe_lmem(gt);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+void intel_gt_info_print(const struct intel_gt_info *info,
+			 struct drm_printer *p)
+{
+	drm_printf(p, "available engines: %x\n", info->engine_mask);
+
+	intel_sseu_dump(&info->sseu, p);
+}
+
+struct reg_and_bit {
+	i915_reg_t reg;
+	u32 bit;
+};
+
+static struct reg_and_bit
+get_reg_and_bit(const struct intel_engine_cs *engine, const bool gen8,
+		const i915_reg_t *regs, const unsigned int num)
+{
+	const unsigned int class = engine->class;
+	struct reg_and_bit rb = { };
+
+	if (drm_WARN_ON_ONCE(&engine->i915->drm,
+			     class >= num || !regs[class].reg))
+		return rb;
+
+	rb.reg = regs[class];
+	if (gen8 && class == VIDEO_DECODE_CLASS)
+		rb.reg.reg += 4 * engine->instance; /* GEN8_M2TCR */
+	else
+		rb.bit = engine->instance;
+
+	rb.bit = BIT(rb.bit);
+
+	return rb;
+}
+
+static void mmio_invalidate_full(struct intel_gt *gt)
+{
+	static const i915_reg_t gen8_regs[] = {
+		[RENDER_CLASS]			= GEN8_RTCR,
+		[VIDEO_DECODE_CLASS]		= GEN8_M1TCR, /* , GEN8_M2TCR */
+		[VIDEO_ENHANCEMENT_CLASS]	= GEN8_VTCR,
+		[COPY_ENGINE_CLASS]		= GEN8_BTCR,
+	};
+	static const i915_reg_t gen12_regs[] = {
+		[RENDER_CLASS]			= GEN12_GFX_TLB_INV_CR,
+		[VIDEO_DECODE_CLASS]		= GEN12_VD_TLB_INV_CR,
+		[VIDEO_ENHANCEMENT_CLASS]	= GEN12_VE_TLB_INV_CR,
+		[COPY_ENGINE_CLASS]		= GEN12_BLT_TLB_INV_CR,
+		[COMPUTE_CLASS]			= GEN12_COMPCTX_TLB_INV_CR,
+	};
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+	struct intel_engine_cs *engine;
+	intel_engine_mask_t awake, tmp;
+	enum intel_engine_id id;
+	const i915_reg_t *regs;
+	unsigned int num = 0;
+
+	if (GRAPHICS_VER(i915) == 12) {
+		regs = gen12_regs;
+		num = ARRAY_SIZE(gen12_regs);
+	} else if (GRAPHICS_VER(i915) >= 8 && GRAPHICS_VER(i915) <= 11) {
+		regs = gen8_regs;
+		num = ARRAY_SIZE(gen8_regs);
+	} else if (GRAPHICS_VER(i915) < 8) {
+		return;
+	}
+
+	if (drm_WARN_ONCE(&i915->drm, !num,
+			  "Platform does not implement TLB invalidation!"))
+		return;
+
+	intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+
+	spin_lock_irq(&uncore->lock); /* serialise invalidate with GT reset */
+
+	awake = 0;
+	for_each_engine(engine, gt, id) {
+		struct reg_and_bit rb;
+
+		if (!intel_engine_pm_is_awake(engine))
+			continue;
+
+		rb = get_reg_and_bit(engine, regs == gen8_regs, regs, num);
+		if (!i915_mmio_reg_offset(rb.reg))
+			continue;
+
+		if (GRAPHICS_VER(i915) == 12 && (engine->class == VIDEO_DECODE_CLASS ||
+		    engine->class == VIDEO_ENHANCEMENT_CLASS ||
+		    engine->class == COMPUTE_CLASS))
+			rb.bit = _MASKED_BIT_ENABLE(rb.bit);
+
+		intel_uncore_write_fw(uncore, rb.reg, rb.bit);
+		awake |= engine->mask;
+	}
+
+	GT_TRACE(gt, "invalidated engines %08x\n", awake);
+
+	/* Wa_2207587034:tgl,dg1,rkl,adl-s,adl-p */
+	if (awake &&
+	    (IS_TIGERLAKE(i915) ||
+	     IS_DG1(i915) ||
+	     IS_ROCKETLAKE(i915) ||
+	     IS_ALDERLAKE_S(i915) ||
+	     IS_ALDERLAKE_P(i915)))
+		intel_uncore_write_fw(uncore, GEN12_OA_TLB_INV_CR, 1);
+
+	spin_unlock_irq(&uncore->lock);
+
+	for_each_engine_masked(engine, gt, awake, tmp) {
+		struct reg_and_bit rb;
+
+		/*
+		 * HW architecture suggest typical invalidation time at 40us,
+		 * with pessimistic cases up to 100us and a recommendation to
+		 * cap at 1ms. We go a bit higher just in case.
+		 */
+		const unsigned int timeout_us = 100;
+		const unsigned int timeout_ms = 4;
+
+		rb = get_reg_and_bit(engine, regs == gen8_regs, regs, num);
+		if (__intel_wait_for_register_fw(uncore,
+						 rb.reg, rb.bit, 0,
+						 timeout_us, timeout_ms,
+						 NULL))
+			drm_err_ratelimited(&gt->i915->drm,
+					    "%s TLB invalidation did not complete in %ums!\n",
+					    engine->name, timeout_ms);
+	}
+
+	/*
+	 * Use delayed put since a) we mostly expect a flurry of TLB
+	 * invalidations so it is good to avoid paying the forcewake cost and
+	 * b) it works around a bug in Icelake which cannot cope with too rapid
+	 * transitions.
+	 */
+	intel_uncore_forcewake_put_delayed(uncore, FORCEWAKE_ALL);
+}
+
+static bool tlb_seqno_passed(const struct intel_gt *gt, u32 seqno)
+{
+	u32 cur = intel_gt_tlb_seqno(gt);
+
+	/* Only skip if a *full* TLB invalidate barrier has passed */
+	return (s32)(cur - ALIGN(seqno, 2)) > 0;
+}
+
+void intel_gt_invalidate_tlb(struct intel_gt *gt, u32 seqno)
+{
+	intel_wakeref_t wakeref;
+
+	if (I915_SELFTEST_ONLY(gt->awake == -ENODEV))
+		return;
+
+	if (intel_gt_is_wedged(gt))
+		return;
+
+	if (tlb_seqno_passed(gt, seqno))
+		return;
+
+	with_intel_gt_pm_if_awake(gt, wakeref) {
+		mutex_lock(&gt->tlb.invalidate_lock);
+		if (tlb_seqno_passed(gt, seqno))
+			goto unlock;
+
+		mmio_invalidate_full(gt);
+
+		write_seqcount_invalidate(&gt->tlb.seqno);
+unlock:
+		mutex_unlock(&gt->tlb.invalidate_lock);
+	}
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt.h b/drivers/gpu/drm/i915/gt/intel_gt.h
new file mode 100644
index 000000000..2ee582e28
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt.h
@@ -0,0 +1,116 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_GT__
+#define __INTEL_GT__
+
+#include "intel_engine_types.h"
+#include "intel_gt_types.h"
+#include "intel_reset.h"
+
+struct drm_i915_private;
+struct drm_printer;
+
+#define GT_TRACE(gt, fmt, ...) do {					\
+	const struct intel_gt *gt__ __maybe_unused = (gt);		\
+	GEM_TRACE("%s " fmt, dev_name(gt__->i915->drm.dev),		\
+		  ##__VA_ARGS__);					\
+} while (0)
+
+static inline bool gt_is_root(struct intel_gt *gt)
+{
+	return !gt->info.id;
+}
+
+static inline struct intel_gt *uc_to_gt(struct intel_uc *uc)
+{
+	return container_of(uc, struct intel_gt, uc);
+}
+
+static inline struct intel_gt *guc_to_gt(struct intel_guc *guc)
+{
+	return container_of(guc, struct intel_gt, uc.guc);
+}
+
+static inline struct intel_gt *huc_to_gt(struct intel_huc *huc)
+{
+	return container_of(huc, struct intel_gt, uc.huc);
+}
+
+static inline struct intel_gt *gsc_to_gt(struct intel_gsc *gsc)
+{
+	return container_of(gsc, struct intel_gt, gsc);
+}
+
+void intel_gt_common_init_early(struct intel_gt *gt);
+int intel_root_gt_init_early(struct drm_i915_private *i915);
+int intel_gt_assign_ggtt(struct intel_gt *gt);
+int intel_gt_init_mmio(struct intel_gt *gt);
+int __must_check intel_gt_init_hw(struct intel_gt *gt);
+int intel_gt_init(struct intel_gt *gt);
+void intel_gt_driver_register(struct intel_gt *gt);
+
+void intel_gt_driver_unregister(struct intel_gt *gt);
+void intel_gt_driver_remove(struct intel_gt *gt);
+void intel_gt_driver_release(struct intel_gt *gt);
+void intel_gt_driver_late_release_all(struct drm_i915_private *i915);
+
+int intel_gt_wait_for_idle(struct intel_gt *gt, long timeout);
+
+void intel_gt_check_and_clear_faults(struct intel_gt *gt);
+void intel_gt_clear_error_registers(struct intel_gt *gt,
+				    intel_engine_mask_t engine_mask);
+
+void intel_gt_flush_ggtt_writes(struct intel_gt *gt);
+void intel_gt_chipset_flush(struct intel_gt *gt);
+
+static inline u32 intel_gt_scratch_offset(const struct intel_gt *gt,
+					  enum intel_gt_scratch_field field)
+{
+	return i915_ggtt_offset(gt->scratch) + field;
+}
+
+static inline bool intel_gt_has_unrecoverable_error(const struct intel_gt *gt)
+{
+	return test_bit(I915_WEDGED_ON_INIT, &gt->reset.flags) ||
+	       test_bit(I915_WEDGED_ON_FINI, &gt->reset.flags);
+}
+
+static inline bool intel_gt_is_wedged(const struct intel_gt *gt)
+{
+	GEM_BUG_ON(intel_gt_has_unrecoverable_error(gt) &&
+		   !test_bit(I915_WEDGED, &gt->reset.flags));
+
+	return unlikely(test_bit(I915_WEDGED, &gt->reset.flags));
+}
+
+int intel_gt_probe_all(struct drm_i915_private *i915);
+int intel_gt_tiles_init(struct drm_i915_private *i915);
+void intel_gt_release_all(struct drm_i915_private *i915);
+
+#define for_each_gt(gt__, i915__, id__) \
+	for ((id__) = 0; \
+	     (id__) < I915_MAX_GT; \
+	     (id__)++) \
+		for_each_if(((gt__) = (i915__)->gt[(id__)]))
+
+void intel_gt_info_print(const struct intel_gt_info *info,
+			 struct drm_printer *p);
+
+void intel_gt_watchdog_work(struct work_struct *work);
+
+static inline u32 intel_gt_tlb_seqno(const struct intel_gt *gt)
+{
+	return seqprop_sequence(&gt->tlb.seqno);
+}
+
+static inline u32 intel_gt_next_invalidate_tlb_full(const struct intel_gt *gt)
+{
+	return intel_gt_tlb_seqno(gt) | 1;
+}
+
+void intel_gt_invalidate_tlb(struct intel_gt *gt, u32 seqno);
+
+#endif /* __INTEL_GT_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.c b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.c
new file mode 100644
index 000000000..cadfd8578
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.c
@@ -0,0 +1,246 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014-2018 Intel Corporation
+ */
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_object.h"
+
+#include "i915_drv.h"
+#include "intel_engine_pm.h"
+#include "intel_gt_buffer_pool.h"
+
+static struct list_head *
+bucket_for_size(struct intel_gt_buffer_pool *pool, size_t sz)
+{
+	int n;
+
+	/*
+	 * Compute a power-of-two bucket, but throw everything greater than
+	 * 16KiB into the same bucket: i.e. the buckets hold objects of
+	 * (1 page, 2 pages, 4 pages, 8+ pages).
+	 */
+	n = fls(sz >> PAGE_SHIFT) - 1;
+	if (n >= ARRAY_SIZE(pool->cache_list))
+		n = ARRAY_SIZE(pool->cache_list) - 1;
+
+	return &pool->cache_list[n];
+}
+
+static void node_free(struct intel_gt_buffer_pool_node *node)
+{
+	i915_gem_object_put(node->obj);
+	i915_active_fini(&node->active);
+	kfree_rcu(node, rcu);
+}
+
+static bool pool_free_older_than(struct intel_gt_buffer_pool *pool, long keep)
+{
+	struct intel_gt_buffer_pool_node *node, *stale = NULL;
+	bool active = false;
+	int n;
+
+	/* Free buffers that have not been used in the past second */
+	for (n = 0; n < ARRAY_SIZE(pool->cache_list); n++) {
+		struct list_head *list = &pool->cache_list[n];
+
+		if (list_empty(list))
+			continue;
+
+		if (spin_trylock_irq(&pool->lock)) {
+			struct list_head *pos;
+
+			/* Most recent at head; oldest at tail */
+			list_for_each_prev(pos, list) {
+				unsigned long age;
+
+				node = list_entry(pos, typeof(*node), link);
+
+				age = READ_ONCE(node->age);
+				if (!age || jiffies - age < keep)
+					break;
+
+				/* Check we are the first to claim this node */
+				if (!xchg(&node->age, 0))
+					break;
+
+				node->free = stale;
+				stale = node;
+			}
+			if (!list_is_last(pos, list))
+				__list_del_many(pos, list);
+
+			spin_unlock_irq(&pool->lock);
+		}
+
+		active |= !list_empty(list);
+	}
+
+	while ((node = stale)) {
+		stale = stale->free;
+		node_free(node);
+	}
+
+	return active;
+}
+
+static void pool_free_work(struct work_struct *wrk)
+{
+	struct intel_gt_buffer_pool *pool =
+		container_of(wrk, typeof(*pool), work.work);
+
+	if (pool_free_older_than(pool, HZ))
+		schedule_delayed_work(&pool->work,
+				      round_jiffies_up_relative(HZ));
+}
+
+static void pool_retire(struct i915_active *ref)
+{
+	struct intel_gt_buffer_pool_node *node =
+		container_of(ref, typeof(*node), active);
+	struct intel_gt_buffer_pool *pool = node->pool;
+	struct list_head *list = bucket_for_size(pool, node->obj->base.size);
+	unsigned long flags;
+
+	if (node->pinned) {
+		i915_gem_object_unpin_pages(node->obj);
+
+		/* Return this object to the shrinker pool */
+		i915_gem_object_make_purgeable(node->obj);
+		node->pinned = false;
+	}
+
+	GEM_BUG_ON(node->age);
+	spin_lock_irqsave(&pool->lock, flags);
+	list_add_rcu(&node->link, list);
+	WRITE_ONCE(node->age, jiffies ?: 1); /* 0 reserved for active nodes */
+	spin_unlock_irqrestore(&pool->lock, flags);
+
+	schedule_delayed_work(&pool->work,
+			      round_jiffies_up_relative(HZ));
+}
+
+void intel_gt_buffer_pool_mark_used(struct intel_gt_buffer_pool_node *node)
+{
+	assert_object_held(node->obj);
+
+	if (node->pinned)
+		return;
+
+	__i915_gem_object_pin_pages(node->obj);
+	/* Hide this pinned object from the shrinker until retired */
+	i915_gem_object_make_unshrinkable(node->obj);
+	node->pinned = true;
+}
+
+static struct intel_gt_buffer_pool_node *
+node_create(struct intel_gt_buffer_pool *pool, size_t sz,
+	    enum i915_map_type type)
+{
+	struct intel_gt *gt = container_of(pool, struct intel_gt, buffer_pool);
+	struct intel_gt_buffer_pool_node *node;
+	struct drm_i915_gem_object *obj;
+
+	node = kmalloc(sizeof(*node),
+		       GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN);
+	if (!node)
+		return ERR_PTR(-ENOMEM);
+
+	node->age = 0;
+	node->pool = pool;
+	node->pinned = false;
+	i915_active_init(&node->active, NULL, pool_retire, 0);
+
+	obj = i915_gem_object_create_internal(gt->i915, sz);
+	if (IS_ERR(obj)) {
+		i915_active_fini(&node->active);
+		kfree(node);
+		return ERR_CAST(obj);
+	}
+
+	i915_gem_object_set_readonly(obj);
+
+	node->type = type;
+	node->obj = obj;
+	return node;
+}
+
+struct intel_gt_buffer_pool_node *
+intel_gt_get_buffer_pool(struct intel_gt *gt, size_t size,
+			 enum i915_map_type type)
+{
+	struct intel_gt_buffer_pool *pool = &gt->buffer_pool;
+	struct intel_gt_buffer_pool_node *node;
+	struct list_head *list;
+	int ret;
+
+	size = PAGE_ALIGN(size);
+	list = bucket_for_size(pool, size);
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(node, list, link) {
+		unsigned long age;
+
+		if (node->obj->base.size < size)
+			continue;
+
+		if (node->type != type)
+			continue;
+
+		age = READ_ONCE(node->age);
+		if (!age)
+			continue;
+
+		if (cmpxchg(&node->age, age, 0) == age) {
+			spin_lock_irq(&pool->lock);
+			list_del_rcu(&node->link);
+			spin_unlock_irq(&pool->lock);
+			break;
+		}
+	}
+	rcu_read_unlock();
+
+	if (&node->link == list) {
+		node = node_create(pool, size, type);
+		if (IS_ERR(node))
+			return node;
+	}
+
+	ret = i915_active_acquire(&node->active);
+	if (ret) {
+		node_free(node);
+		return ERR_PTR(ret);
+	}
+
+	return node;
+}
+
+void intel_gt_init_buffer_pool(struct intel_gt *gt)
+{
+	struct intel_gt_buffer_pool *pool = &gt->buffer_pool;
+	int n;
+
+	spin_lock_init(&pool->lock);
+	for (n = 0; n < ARRAY_SIZE(pool->cache_list); n++)
+		INIT_LIST_HEAD(&pool->cache_list[n]);
+	INIT_DELAYED_WORK(&pool->work, pool_free_work);
+}
+
+void intel_gt_flush_buffer_pool(struct intel_gt *gt)
+{
+	struct intel_gt_buffer_pool *pool = &gt->buffer_pool;
+
+	do {
+		while (pool_free_older_than(pool, 0))
+			;
+	} while (cancel_delayed_work_sync(&pool->work));
+}
+
+void intel_gt_fini_buffer_pool(struct intel_gt *gt)
+{
+	struct intel_gt_buffer_pool *pool = &gt->buffer_pool;
+	int n;
+
+	for (n = 0; n < ARRAY_SIZE(pool->cache_list); n++)
+		GEM_BUG_ON(!list_empty(&pool->cache_list[n]));
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.h b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.h
new file mode 100644
index 000000000..487b8a552
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool.h
@@ -0,0 +1,43 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2018 Intel Corporation
+ */
+
+#ifndef INTEL_GT_BUFFER_POOL_H
+#define INTEL_GT_BUFFER_POOL_H
+
+#include <linux/types.h>
+
+#include "i915_active.h"
+#include "intel_gt_buffer_pool_types.h"
+
+struct intel_gt;
+struct i915_request;
+
+struct intel_gt_buffer_pool_node *
+intel_gt_get_buffer_pool(struct intel_gt *gt, size_t size,
+			 enum i915_map_type type);
+
+void intel_gt_buffer_pool_mark_used(struct intel_gt_buffer_pool_node *node);
+
+static inline int
+intel_gt_buffer_pool_mark_active(struct intel_gt_buffer_pool_node *node,
+				 struct i915_request *rq)
+{
+	/* did we call mark_used? */
+	GEM_WARN_ON(!node->pinned);
+
+	return i915_active_add_request(&node->active, rq);
+}
+
+static inline void
+intel_gt_buffer_pool_put(struct intel_gt_buffer_pool_node *node)
+{
+	i915_active_release(&node->active);
+}
+
+void intel_gt_init_buffer_pool(struct intel_gt *gt);
+void intel_gt_flush_buffer_pool(struct intel_gt *gt);
+void intel_gt_fini_buffer_pool(struct intel_gt *gt);
+
+#endif /* INTEL_GT_BUFFER_POOL_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool_types.h b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool_types.h
new file mode 100644
index 000000000..df1d75d08
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_buffer_pool_types.h
@@ -0,0 +1,36 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2018 Intel Corporation
+ */
+
+#ifndef INTEL_GT_BUFFER_POOL_TYPES_H
+#define INTEL_GT_BUFFER_POOL_TYPES_H
+
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/workqueue.h>
+
+#include "gem/i915_gem_object_types.h"
+#include "i915_active_types.h"
+
+struct intel_gt_buffer_pool {
+	spinlock_t lock;
+	struct list_head cache_list[4];
+	struct delayed_work work;
+};
+
+struct intel_gt_buffer_pool_node {
+	struct i915_active active;
+	struct drm_i915_gem_object *obj;
+	struct list_head link;
+	union {
+		struct intel_gt_buffer_pool *pool;
+		struct intel_gt_buffer_pool_node *free;
+		struct rcu_head rcu;
+	};
+	unsigned long age;
+	enum i915_map_type type;
+	u32 pinned;
+};
+
+#endif /* INTEL_GT_BUFFER_POOL_TYPES_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.c b/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.c
new file mode 100644
index 000000000..3f656d3db
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.c
@@ -0,0 +1,212 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_gt.h"
+#include "intel_gt_clock_utils.h"
+#include "intel_gt_regs.h"
+
+static u32 read_reference_ts_freq(struct intel_uncore *uncore)
+{
+	u32 ts_override = intel_uncore_read(uncore, GEN9_TIMESTAMP_OVERRIDE);
+	u32 base_freq, frac_freq;
+
+	base_freq = ((ts_override & GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DIVIDER_MASK) >>
+		     GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DIVIDER_SHIFT) + 1;
+	base_freq *= 1000000;
+
+	frac_freq = ((ts_override &
+		      GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DENOMINATOR_MASK) >>
+		     GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DENOMINATOR_SHIFT);
+	frac_freq = 1000000 / (frac_freq + 1);
+
+	return base_freq + frac_freq;
+}
+
+static u32 gen11_get_crystal_clock_freq(struct intel_uncore *uncore,
+					u32 rpm_config_reg)
+{
+	u32 f19_2_mhz = 19200000;
+	u32 f24_mhz = 24000000;
+	u32 f25_mhz = 25000000;
+	u32 f38_4_mhz = 38400000;
+	u32 crystal_clock =
+		(rpm_config_reg & GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >>
+		GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT;
+
+	switch (crystal_clock) {
+	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ:
+		return f24_mhz;
+	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ:
+		return f19_2_mhz;
+	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_38_4_MHZ:
+		return f38_4_mhz;
+	case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_25_MHZ:
+		return f25_mhz;
+	default:
+		MISSING_CASE(crystal_clock);
+		return 0;
+	}
+}
+
+static u32 gen11_read_clock_frequency(struct intel_uncore *uncore)
+{
+	u32 ctc_reg = intel_uncore_read(uncore, CTC_MODE);
+	u32 freq = 0;
+
+	/*
+	 * Note that on gen11+, the clock frequency may be reconfigured.
+	 * We do not, and we assume nobody else does.
+	 *
+	 * First figure out the reference frequency. There are 2 ways
+	 * we can compute the frequency, either through the
+	 * TIMESTAMP_OVERRIDE register or through RPM_CONFIG. CTC_MODE
+	 * tells us which one we should use.
+	 */
+	if ((ctc_reg & CTC_SOURCE_PARAMETER_MASK) == CTC_SOURCE_DIVIDE_LOGIC) {
+		freq = read_reference_ts_freq(uncore);
+	} else {
+		u32 c0 = intel_uncore_read(uncore, RPM_CONFIG0);
+
+		freq = gen11_get_crystal_clock_freq(uncore, c0);
+
+		/*
+		 * Now figure out how the command stream's timestamp
+		 * register increments from this frequency (it might
+		 * increment only every few clock cycle).
+		 */
+		freq >>= 3 - ((c0 & GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_MASK) >>
+			      GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_SHIFT);
+	}
+
+	return freq;
+}
+
+static u32 gen9_read_clock_frequency(struct intel_uncore *uncore)
+{
+	u32 ctc_reg = intel_uncore_read(uncore, CTC_MODE);
+	u32 freq = 0;
+
+	if ((ctc_reg & CTC_SOURCE_PARAMETER_MASK) == CTC_SOURCE_DIVIDE_LOGIC) {
+		freq = read_reference_ts_freq(uncore);
+	} else {
+		freq = IS_GEN9_LP(uncore->i915) ? 19200000 : 24000000;
+
+		/*
+		 * Now figure out how the command stream's timestamp
+		 * register increments from this frequency (it might
+		 * increment only every few clock cycle).
+		 */
+		freq >>= 3 - ((ctc_reg & CTC_SHIFT_PARAMETER_MASK) >>
+			      CTC_SHIFT_PARAMETER_SHIFT);
+	}
+
+	return freq;
+}
+
+static u32 gen5_read_clock_frequency(struct intel_uncore *uncore)
+{
+	/*
+	 * PRMs say:
+	 *
+	 *     "The PCU TSC counts 10ns increments; this timestamp
+	 *      reflects bits 38:3 of the TSC (i.e. 80ns granularity,
+	 *      rolling over every 1.5 hours).
+	 */
+	return 12500000;
+}
+
+static u32 gen2_read_clock_frequency(struct intel_uncore *uncore)
+{
+	/*
+	 * PRMs say:
+	 *
+	 *     "The value in this register increments once every 16
+	 *      hclks." (through the “Clocking Configuration”
+	 *      (“CLKCFG”) MCHBAR register)
+	 */
+	return RUNTIME_INFO(uncore->i915)->rawclk_freq * 1000 / 16;
+}
+
+static u32 read_clock_frequency(struct intel_uncore *uncore)
+{
+	if (GRAPHICS_VER(uncore->i915) >= 11)
+		return gen11_read_clock_frequency(uncore);
+	else if (GRAPHICS_VER(uncore->i915) >= 9)
+		return gen9_read_clock_frequency(uncore);
+	else if (GRAPHICS_VER(uncore->i915) >= 5)
+		return gen5_read_clock_frequency(uncore);
+	else
+		return gen2_read_clock_frequency(uncore);
+}
+
+void intel_gt_init_clock_frequency(struct intel_gt *gt)
+{
+	gt->clock_frequency = read_clock_frequency(gt->uncore);
+
+	/* Icelake appears to use another fixed frequency for CTX_TIMESTAMP */
+	if (GRAPHICS_VER(gt->i915) == 11)
+		gt->clock_period_ns = NSEC_PER_SEC / 13750000;
+	else if (gt->clock_frequency)
+		gt->clock_period_ns = intel_gt_clock_interval_to_ns(gt, 1);
+
+	GT_TRACE(gt,
+		 "Using clock frequency: %dkHz, period: %dns, wrap: %lldms\n",
+		 gt->clock_frequency / 1000,
+		 gt->clock_period_ns,
+		 div_u64(mul_u32_u32(gt->clock_period_ns, S32_MAX),
+			 USEC_PER_SEC));
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)
+void intel_gt_check_clock_frequency(const struct intel_gt *gt)
+{
+	if (gt->clock_frequency != read_clock_frequency(gt->uncore)) {
+		dev_err(gt->i915->drm.dev,
+			"GT clock frequency changed, was %uHz, now %uHz!\n",
+			gt->clock_frequency,
+			read_clock_frequency(gt->uncore));
+	}
+}
+#endif
+
+static u64 div_u64_roundup(u64 nom, u32 den)
+{
+	return div_u64(nom + den - 1, den);
+}
+
+u64 intel_gt_clock_interval_to_ns(const struct intel_gt *gt, u64 count)
+{
+	return div_u64_roundup(count * NSEC_PER_SEC, gt->clock_frequency);
+}
+
+u64 intel_gt_pm_interval_to_ns(const struct intel_gt *gt, u64 count)
+{
+	return intel_gt_clock_interval_to_ns(gt, 16 * count);
+}
+
+u64 intel_gt_ns_to_clock_interval(const struct intel_gt *gt, u64 ns)
+{
+	return div_u64_roundup(gt->clock_frequency * ns, NSEC_PER_SEC);
+}
+
+u64 intel_gt_ns_to_pm_interval(const struct intel_gt *gt, u64 ns)
+{
+	u64 val;
+
+	/*
+	 * Make these a multiple of magic 25 to avoid SNB (eg. Dell XPS
+	 * 8300) freezing up around GPU hangs. Looks as if even
+	 * scheduling/timer interrupts start misbehaving if the RPS
+	 * EI/thresholds are "bad", leading to a very sluggish or even
+	 * frozen machine.
+	 */
+	val = div_u64_roundup(intel_gt_ns_to_clock_interval(gt, ns), 16);
+	if (GRAPHICS_VER(gt->i915) == 6)
+		val = div_u64_roundup(val, 25) * 25;
+
+	return val;
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.h b/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.h
new file mode 100644
index 000000000..8b03e97a8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_clock_utils.h
@@ -0,0 +1,27 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __INTEL_GT_CLOCK_UTILS_H__
+#define __INTEL_GT_CLOCK_UTILS_H__
+
+#include <linux/types.h>
+
+struct intel_gt;
+
+void intel_gt_init_clock_frequency(struct intel_gt *gt);
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)
+void intel_gt_check_clock_frequency(const struct intel_gt *gt);
+#else
+static inline void intel_gt_check_clock_frequency(const struct intel_gt *gt) {}
+#endif
+
+u64 intel_gt_clock_interval_to_ns(const struct intel_gt *gt, u64 count);
+u64 intel_gt_pm_interval_to_ns(const struct intel_gt *gt, u64 count);
+
+u64 intel_gt_ns_to_clock_interval(const struct intel_gt *gt, u64 ns);
+u64 intel_gt_ns_to_pm_interval(const struct intel_gt *gt, u64 ns);
+
+#endif /* __INTEL_GT_CLOCK_UTILS_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_debugfs.c b/drivers/gpu/drm/i915/gt/intel_gt_debugfs.c
new file mode 100644
index 000000000..dd53641f3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_debugfs.c
@@ -0,0 +1,119 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/debugfs.h>
+
+#include "i915_drv.h"
+#include "intel_gt.h"
+#include "intel_gt_debugfs.h"
+#include "intel_gt_engines_debugfs.h"
+#include "intel_gt_mcr.h"
+#include "intel_gt_pm_debugfs.h"
+#include "intel_sseu_debugfs.h"
+#include "pxp/intel_pxp_debugfs.h"
+#include "uc/intel_uc_debugfs.h"
+
+int intel_gt_debugfs_reset_show(struct intel_gt *gt, u64 *val)
+{
+	int ret = intel_gt_terminally_wedged(gt);
+
+	switch (ret) {
+	case -EIO:
+		*val = 1;
+		return 0;
+	case 0:
+		*val = 0;
+		return 0;
+	default:
+		return ret;
+	}
+}
+
+void intel_gt_debugfs_reset_store(struct intel_gt *gt, u64 val)
+{
+	/* Flush any previous reset before applying for a new one */
+	wait_event(gt->reset.queue,
+		   !test_bit(I915_RESET_BACKOFF, &gt->reset.flags));
+
+	intel_gt_handle_error(gt, val, I915_ERROR_CAPTURE,
+			      "Manually reset engine mask to %llx", val);
+}
+
+/*
+ * keep the interface clean where the first parameter
+ * is a 'struct intel_gt *' instead of 'void *'
+ */
+static int __intel_gt_debugfs_reset_show(void *data, u64 *val)
+{
+	return intel_gt_debugfs_reset_show(data, val);
+}
+
+static int __intel_gt_debugfs_reset_store(void *data, u64 val)
+{
+	intel_gt_debugfs_reset_store(data, val);
+
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(reset_fops, __intel_gt_debugfs_reset_show,
+			__intel_gt_debugfs_reset_store, "%llu\n");
+
+static int steering_show(struct seq_file *m, void *data)
+{
+	struct drm_printer p = drm_seq_file_printer(m);
+	struct intel_gt *gt = m->private;
+
+	intel_gt_mcr_report_steering(&p, gt, true);
+
+	return 0;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(steering);
+
+static void gt_debugfs_register(struct intel_gt *gt, struct dentry *root)
+{
+	static const struct intel_gt_debugfs_file files[] = {
+		{ "reset", &reset_fops, NULL },
+		{ "steering", &steering_fops },
+	};
+
+	intel_gt_debugfs_register_files(root, files, ARRAY_SIZE(files), gt);
+}
+
+void intel_gt_debugfs_register(struct intel_gt *gt)
+{
+	struct dentry *root;
+
+	if (!gt->i915->drm.primary->debugfs_root)
+		return;
+
+	root = debugfs_create_dir("gt", gt->i915->drm.primary->debugfs_root);
+	if (IS_ERR(root))
+		return;
+
+	gt_debugfs_register(gt, root);
+
+	intel_gt_engines_debugfs_register(gt, root);
+	intel_gt_pm_debugfs_register(gt, root);
+	intel_sseu_debugfs_register(gt, root);
+
+	intel_uc_debugfs_register(&gt->uc, root);
+	intel_pxp_debugfs_register(&gt->pxp, root);
+}
+
+void intel_gt_debugfs_register_files(struct dentry *root,
+				     const struct intel_gt_debugfs_file *files,
+				     unsigned long count, void *data)
+{
+	while (count--) {
+		umode_t mode = files->fops->write ? 0644 : 0444;
+
+		if (!files->eval || files->eval(data))
+			debugfs_create_file(files->name,
+					    mode, root, data,
+					    files->fops);
+
+		files++;
+	}
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_debugfs.h b/drivers/gpu/drm/i915/gt/intel_gt_debugfs.h
new file mode 100644
index 000000000..e4110eebf
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_debugfs.h
@@ -0,0 +1,53 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_GT_DEBUGFS_H
+#define INTEL_GT_DEBUGFS_H
+
+#include <linux/file.h>
+
+struct intel_gt;
+
+#define __GT_DEBUGFS_ATTRIBUTE_FOPS(__name)				\
+static const struct file_operations __name ## _fops = {			\
+	.owner = THIS_MODULE,						\
+	.open = __name ## _open,					\
+	.read = seq_read,						\
+	.llseek = seq_lseek,						\
+	.release = single_release,					\
+}
+
+#define DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(__name)			\
+static int __name ## _open(struct inode *inode, struct file *file)	\
+{									\
+	return single_open(file, __name ## _show, inode->i_private);	\
+}									\
+__GT_DEBUGFS_ATTRIBUTE_FOPS(__name)
+
+#define DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE_WITH_SIZE(__name, __size_vf)		\
+static int __name ## _open(struct inode *inode, struct file *file)		\
+{										\
+	return single_open_size(file, __name ## _show, inode->i_private,	\
+			    __size_vf(inode->i_private));			\
+}										\
+__GT_DEBUGFS_ATTRIBUTE_FOPS(__name)
+
+void intel_gt_debugfs_register(struct intel_gt *gt);
+
+struct intel_gt_debugfs_file {
+	const char *name;
+	const struct file_operations *fops;
+	bool (*eval)(void *data);
+};
+
+void intel_gt_debugfs_register_files(struct dentry *root,
+				     const struct intel_gt_debugfs_file *files,
+				     unsigned long count, void *data);
+
+/* functions that need to be accessed by the upper level non-gt interfaces */
+int intel_gt_debugfs_reset_show(struct intel_gt *gt, u64 *val);
+void intel_gt_debugfs_reset_store(struct intel_gt *gt, u64 val);
+
+#endif /* INTEL_GT_DEBUGFS_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_engines_debugfs.c b/drivers/gpu/drm/i915/gt/intel_gt_engines_debugfs.c
new file mode 100644
index 000000000..8f9b874fd
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_engines_debugfs.c
@@ -0,0 +1,36 @@
+// SPDX-License-Identifier: MIT
+
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <drm/drm_print.h>
+
+#include "i915_drv.h" /* for_each_engine! */
+#include "intel_engine.h"
+#include "intel_gt_debugfs.h"
+#include "intel_gt_engines_debugfs.h"
+
+static int engines_show(struct seq_file *m, void *data)
+{
+	struct intel_gt *gt = m->private;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct drm_printer p;
+
+	p = drm_seq_file_printer(m);
+	for_each_engine(engine, gt, id)
+		intel_engine_dump(engine, &p, "%s\n", engine->name);
+
+	return 0;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(engines);
+
+void intel_gt_engines_debugfs_register(struct intel_gt *gt, struct dentry *root)
+{
+	static const struct intel_gt_debugfs_file files[] = {
+		{ "engines", &engines_fops },
+	};
+
+	intel_gt_debugfs_register_files(root, files, ARRAY_SIZE(files), gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_engines_debugfs.h b/drivers/gpu/drm/i915/gt/intel_gt_engines_debugfs.h
new file mode 100644
index 000000000..dda113452
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_engines_debugfs.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_GT_ENGINES_DEBUGFS_H
+#define INTEL_GT_ENGINES_DEBUGFS_H
+
+struct intel_gt;
+struct dentry;
+
+void intel_gt_engines_debugfs_register(struct intel_gt *gt, struct dentry *root);
+
+#endif /* INTEL_GT_ENGINES_DEBUGFS_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_irq.c b/drivers/gpu/drm/i915/gt/intel_gt_irq.c
new file mode 100644
index 000000000..f26882fdc
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_irq.c
@@ -0,0 +1,519 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/sched/clock.h>
+
+#include "i915_drv.h"
+#include "i915_irq.h"
+#include "intel_breadcrumbs.h"
+#include "intel_gt.h"
+#include "intel_gt_irq.h"
+#include "intel_gt_regs.h"
+#include "intel_uncore.h"
+#include "intel_rps.h"
+#include "pxp/intel_pxp_irq.h"
+
+static void guc_irq_handler(struct intel_guc *guc, u16 iir)
+{
+	if (iir & GUC_INTR_GUC2HOST)
+		intel_guc_to_host_event_handler(guc);
+}
+
+static u32
+gen11_gt_engine_identity(struct intel_gt *gt,
+			 const unsigned int bank, const unsigned int bit)
+{
+	void __iomem * const regs = gt->uncore->regs;
+	u32 timeout_ts;
+	u32 ident;
+
+	lockdep_assert_held(gt->irq_lock);
+
+	raw_reg_write(regs, GEN11_IIR_REG_SELECTOR(bank), BIT(bit));
+
+	/*
+	 * NB: Specs do not specify how long to spin wait,
+	 * so we do ~100us as an educated guess.
+	 */
+	timeout_ts = (local_clock() >> 10) + 100;
+	do {
+		ident = raw_reg_read(regs, GEN11_INTR_IDENTITY_REG(bank));
+	} while (!(ident & GEN11_INTR_DATA_VALID) &&
+		 !time_after32(local_clock() >> 10, timeout_ts));
+
+	if (unlikely(!(ident & GEN11_INTR_DATA_VALID))) {
+		DRM_ERROR("INTR_IDENTITY_REG%u:%u 0x%08x not valid!\n",
+			  bank, bit, ident);
+		return 0;
+	}
+
+	raw_reg_write(regs, GEN11_INTR_IDENTITY_REG(bank),
+		      GEN11_INTR_DATA_VALID);
+
+	return ident;
+}
+
+static void
+gen11_other_irq_handler(struct intel_gt *gt, const u8 instance,
+			const u16 iir)
+{
+	struct intel_gt *media_gt = gt->i915->media_gt;
+
+	if (instance == OTHER_GUC_INSTANCE)
+		return guc_irq_handler(&gt->uc.guc, iir);
+	if (instance == OTHER_MEDIA_GUC_INSTANCE && media_gt)
+		return guc_irq_handler(&media_gt->uc.guc, iir);
+
+	if (instance == OTHER_GTPM_INSTANCE)
+		return gen11_rps_irq_handler(&gt->rps, iir);
+	if (instance == OTHER_MEDIA_GTPM_INSTANCE && media_gt)
+		return gen11_rps_irq_handler(&media_gt->rps, iir);
+
+	if (instance == OTHER_KCR_INSTANCE)
+		return intel_pxp_irq_handler(&gt->pxp, iir);
+
+	if (instance == OTHER_GSC_INSTANCE)
+		return intel_gsc_irq_handler(gt, iir);
+
+	WARN_ONCE(1, "unhandled other interrupt instance=0x%x, iir=0x%x\n",
+		  instance, iir);
+}
+
+static void
+gen11_engine_irq_handler(struct intel_gt *gt, const u8 class,
+			 const u8 instance, const u16 iir)
+{
+	struct intel_engine_cs *engine;
+
+	/*
+	 * Platforms with standalone media have their media engines in another
+	 * GT.
+	 */
+	if (MEDIA_VER(gt->i915) >= 13 &&
+	    (class == VIDEO_DECODE_CLASS || class == VIDEO_ENHANCEMENT_CLASS)) {
+		if (!gt->i915->media_gt)
+			goto err;
+
+		gt = gt->i915->media_gt;
+	}
+
+	if (instance <= MAX_ENGINE_INSTANCE)
+		engine = gt->engine_class[class][instance];
+	else
+		engine = NULL;
+
+	if (likely(engine))
+		return intel_engine_cs_irq(engine, iir);
+
+err:
+	WARN_ONCE(1, "unhandled engine interrupt class=0x%x, instance=0x%x\n",
+		  class, instance);
+}
+
+static void
+gen11_gt_identity_handler(struct intel_gt *gt, const u32 identity)
+{
+	const u8 class = GEN11_INTR_ENGINE_CLASS(identity);
+	const u8 instance = GEN11_INTR_ENGINE_INSTANCE(identity);
+	const u16 intr = GEN11_INTR_ENGINE_INTR(identity);
+
+	if (unlikely(!intr))
+		return;
+
+	if (class <= COPY_ENGINE_CLASS || class == COMPUTE_CLASS)
+		return gen11_engine_irq_handler(gt, class, instance, intr);
+
+	if (class == OTHER_CLASS)
+		return gen11_other_irq_handler(gt, instance, intr);
+
+	WARN_ONCE(1, "unknown interrupt class=0x%x, instance=0x%x, intr=0x%x\n",
+		  class, instance, intr);
+}
+
+static void
+gen11_gt_bank_handler(struct intel_gt *gt, const unsigned int bank)
+{
+	void __iomem * const regs = gt->uncore->regs;
+	unsigned long intr_dw;
+	unsigned int bit;
+
+	lockdep_assert_held(gt->irq_lock);
+
+	intr_dw = raw_reg_read(regs, GEN11_GT_INTR_DW(bank));
+
+	for_each_set_bit(bit, &intr_dw, 32) {
+		const u32 ident = gen11_gt_engine_identity(gt, bank, bit);
+
+		gen11_gt_identity_handler(gt, ident);
+	}
+
+	/* Clear must be after shared has been served for engine */
+	raw_reg_write(regs, GEN11_GT_INTR_DW(bank), intr_dw);
+}
+
+void gen11_gt_irq_handler(struct intel_gt *gt, const u32 master_ctl)
+{
+	unsigned int bank;
+
+	spin_lock(gt->irq_lock);
+
+	for (bank = 0; bank < 2; bank++) {
+		if (master_ctl & GEN11_GT_DW_IRQ(bank))
+			gen11_gt_bank_handler(gt, bank);
+	}
+
+	spin_unlock(gt->irq_lock);
+}
+
+bool gen11_gt_reset_one_iir(struct intel_gt *gt,
+			    const unsigned int bank, const unsigned int bit)
+{
+	void __iomem * const regs = gt->uncore->regs;
+	u32 dw;
+
+	lockdep_assert_held(gt->irq_lock);
+
+	dw = raw_reg_read(regs, GEN11_GT_INTR_DW(bank));
+	if (dw & BIT(bit)) {
+		/*
+		 * According to the BSpec, DW_IIR bits cannot be cleared without
+		 * first servicing the Selector & Shared IIR registers.
+		 */
+		gen11_gt_engine_identity(gt, bank, bit);
+
+		/*
+		 * We locked GT INT DW by reading it. If we want to (try
+		 * to) recover from this successfully, we need to clear
+		 * our bit, otherwise we are locking the register for
+		 * everybody.
+		 */
+		raw_reg_write(regs, GEN11_GT_INTR_DW(bank), BIT(bit));
+
+		return true;
+	}
+
+	return false;
+}
+
+void gen11_gt_irq_reset(struct intel_gt *gt)
+{
+	struct intel_uncore *uncore = gt->uncore;
+
+	/* Disable RCS, BCS, VCS and VECS class engines. */
+	intel_uncore_write(uncore, GEN11_RENDER_COPY_INTR_ENABLE, 0);
+	intel_uncore_write(uncore, GEN11_VCS_VECS_INTR_ENABLE,	  0);
+	if (CCS_MASK(gt))
+		intel_uncore_write(uncore, GEN12_CCS_RSVD_INTR_ENABLE, 0);
+	if (HAS_HECI_GSC(gt->i915))
+		intel_uncore_write(uncore, GEN11_GUNIT_CSME_INTR_ENABLE, 0);
+
+	/* Restore masks irqs on RCS, BCS, VCS and VECS engines. */
+	intel_uncore_write(uncore, GEN11_RCS0_RSVD_INTR_MASK,	~0);
+	intel_uncore_write(uncore, GEN11_BCS_RSVD_INTR_MASK,	~0);
+	if (HAS_ENGINE(gt, BCS1) || HAS_ENGINE(gt, BCS2))
+		intel_uncore_write(uncore, XEHPC_BCS1_BCS2_INTR_MASK, ~0);
+	if (HAS_ENGINE(gt, BCS3) || HAS_ENGINE(gt, BCS4))
+		intel_uncore_write(uncore, XEHPC_BCS3_BCS4_INTR_MASK, ~0);
+	if (HAS_ENGINE(gt, BCS5) || HAS_ENGINE(gt, BCS6))
+		intel_uncore_write(uncore, XEHPC_BCS5_BCS6_INTR_MASK, ~0);
+	if (HAS_ENGINE(gt, BCS7) || HAS_ENGINE(gt, BCS8))
+		intel_uncore_write(uncore, XEHPC_BCS7_BCS8_INTR_MASK, ~0);
+	intel_uncore_write(uncore, GEN11_VCS0_VCS1_INTR_MASK,	~0);
+	intel_uncore_write(uncore, GEN11_VCS2_VCS3_INTR_MASK,	~0);
+	if (HAS_ENGINE(gt, VCS4) || HAS_ENGINE(gt, VCS5))
+		intel_uncore_write(uncore, GEN12_VCS4_VCS5_INTR_MASK,   ~0);
+	if (HAS_ENGINE(gt, VCS6) || HAS_ENGINE(gt, VCS7))
+		intel_uncore_write(uncore, GEN12_VCS6_VCS7_INTR_MASK,   ~0);
+	intel_uncore_write(uncore, GEN11_VECS0_VECS1_INTR_MASK,	~0);
+	if (HAS_ENGINE(gt, VECS2) || HAS_ENGINE(gt, VECS3))
+		intel_uncore_write(uncore, GEN12_VECS2_VECS3_INTR_MASK, ~0);
+	if (HAS_ENGINE(gt, CCS0) || HAS_ENGINE(gt, CCS1))
+		intel_uncore_write(uncore, GEN12_CCS0_CCS1_INTR_MASK, ~0);
+	if (HAS_ENGINE(gt, CCS2) || HAS_ENGINE(gt, CCS3))
+		intel_uncore_write(uncore, GEN12_CCS2_CCS3_INTR_MASK, ~0);
+	if (HAS_HECI_GSC(gt->i915))
+		intel_uncore_write(uncore, GEN11_GUNIT_CSME_INTR_MASK, ~0);
+
+	intel_uncore_write(uncore, GEN11_GPM_WGBOXPERF_INTR_ENABLE, 0);
+	intel_uncore_write(uncore, GEN11_GPM_WGBOXPERF_INTR_MASK,  ~0);
+	intel_uncore_write(uncore, GEN11_GUC_SG_INTR_ENABLE, 0);
+	intel_uncore_write(uncore, GEN11_GUC_SG_INTR_MASK,  ~0);
+
+	intel_uncore_write(uncore, GEN11_CRYPTO_RSVD_INTR_ENABLE, 0);
+	intel_uncore_write(uncore, GEN11_CRYPTO_RSVD_INTR_MASK,  ~0);
+}
+
+void gen11_gt_irq_postinstall(struct intel_gt *gt)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	u32 irqs = GT_RENDER_USER_INTERRUPT;
+	const u32 gsc_mask = GSC_IRQ_INTF(0) | GSC_IRQ_INTF(1);
+	u32 dmask;
+	u32 smask;
+
+	if (!intel_uc_wants_guc_submission(&gt->uc))
+		irqs |= GT_CS_MASTER_ERROR_INTERRUPT |
+			GT_CONTEXT_SWITCH_INTERRUPT |
+			GT_WAIT_SEMAPHORE_INTERRUPT;
+
+	dmask = irqs << 16 | irqs;
+	smask = irqs << 16;
+
+	BUILD_BUG_ON(irqs & 0xffff0000);
+
+	/* Enable RCS, BCS, VCS and VECS class interrupts. */
+	intel_uncore_write(uncore, GEN11_RENDER_COPY_INTR_ENABLE, dmask);
+	intel_uncore_write(uncore, GEN11_VCS_VECS_INTR_ENABLE, dmask);
+	if (CCS_MASK(gt))
+		intel_uncore_write(uncore, GEN12_CCS_RSVD_INTR_ENABLE, smask);
+	if (HAS_HECI_GSC(gt->i915))
+		intel_uncore_write(uncore, GEN11_GUNIT_CSME_INTR_ENABLE,
+				   gsc_mask);
+
+	/* Unmask irqs on RCS, BCS, VCS and VECS engines. */
+	intel_uncore_write(uncore, GEN11_RCS0_RSVD_INTR_MASK, ~smask);
+	intel_uncore_write(uncore, GEN11_BCS_RSVD_INTR_MASK, ~smask);
+	if (HAS_ENGINE(gt, BCS1) || HAS_ENGINE(gt, BCS2))
+		intel_uncore_write(uncore, XEHPC_BCS1_BCS2_INTR_MASK, ~dmask);
+	if (HAS_ENGINE(gt, BCS3) || HAS_ENGINE(gt, BCS4))
+		intel_uncore_write(uncore, XEHPC_BCS3_BCS4_INTR_MASK, ~dmask);
+	if (HAS_ENGINE(gt, BCS5) || HAS_ENGINE(gt, BCS6))
+		intel_uncore_write(uncore, XEHPC_BCS5_BCS6_INTR_MASK, ~dmask);
+	if (HAS_ENGINE(gt, BCS7) || HAS_ENGINE(gt, BCS8))
+		intel_uncore_write(uncore, XEHPC_BCS7_BCS8_INTR_MASK, ~dmask);
+	intel_uncore_write(uncore, GEN11_VCS0_VCS1_INTR_MASK, ~dmask);
+	intel_uncore_write(uncore, GEN11_VCS2_VCS3_INTR_MASK, ~dmask);
+	if (HAS_ENGINE(gt, VCS4) || HAS_ENGINE(gt, VCS5))
+		intel_uncore_write(uncore, GEN12_VCS4_VCS5_INTR_MASK, ~dmask);
+	if (HAS_ENGINE(gt, VCS6) || HAS_ENGINE(gt, VCS7))
+		intel_uncore_write(uncore, GEN12_VCS6_VCS7_INTR_MASK, ~dmask);
+	intel_uncore_write(uncore, GEN11_VECS0_VECS1_INTR_MASK, ~dmask);
+	if (HAS_ENGINE(gt, VECS2) || HAS_ENGINE(gt, VECS3))
+		intel_uncore_write(uncore, GEN12_VECS2_VECS3_INTR_MASK, ~dmask);
+	if (HAS_ENGINE(gt, CCS0) || HAS_ENGINE(gt, CCS1))
+		intel_uncore_write(uncore, GEN12_CCS0_CCS1_INTR_MASK, ~dmask);
+	if (HAS_ENGINE(gt, CCS2) || HAS_ENGINE(gt, CCS3))
+		intel_uncore_write(uncore, GEN12_CCS2_CCS3_INTR_MASK, ~dmask);
+	if (HAS_HECI_GSC(gt->i915))
+		intel_uncore_write(uncore, GEN11_GUNIT_CSME_INTR_MASK, ~gsc_mask);
+
+	/*
+	 * RPS interrupts will get enabled/disabled on demand when RPS itself
+	 * is enabled/disabled.
+	 */
+	gt->pm_ier = 0x0;
+	gt->pm_imr = ~gt->pm_ier;
+	intel_uncore_write(uncore, GEN11_GPM_WGBOXPERF_INTR_ENABLE, 0);
+	intel_uncore_write(uncore, GEN11_GPM_WGBOXPERF_INTR_MASK,  ~0);
+
+	/* Same thing for GuC interrupts */
+	intel_uncore_write(uncore, GEN11_GUC_SG_INTR_ENABLE, 0);
+	intel_uncore_write(uncore, GEN11_GUC_SG_INTR_MASK,  ~0);
+}
+
+void gen5_gt_irq_handler(struct intel_gt *gt, u32 gt_iir)
+{
+	if (gt_iir & GT_RENDER_USER_INTERRUPT)
+		intel_engine_cs_irq(gt->engine_class[RENDER_CLASS][0],
+				    gt_iir);
+
+	if (gt_iir & ILK_BSD_USER_INTERRUPT)
+		intel_engine_cs_irq(gt->engine_class[VIDEO_DECODE_CLASS][0],
+				    gt_iir);
+}
+
+static void gen7_parity_error_irq_handler(struct intel_gt *gt, u32 iir)
+{
+	if (!HAS_L3_DPF(gt->i915))
+		return;
+
+	spin_lock(gt->irq_lock);
+	gen5_gt_disable_irq(gt, GT_PARITY_ERROR(gt->i915));
+	spin_unlock(gt->irq_lock);
+
+	if (iir & GT_RENDER_L3_PARITY_ERROR_INTERRUPT_S1)
+		gt->i915->l3_parity.which_slice |= 1 << 1;
+
+	if (iir & GT_RENDER_L3_PARITY_ERROR_INTERRUPT)
+		gt->i915->l3_parity.which_slice |= 1 << 0;
+
+	schedule_work(&gt->i915->l3_parity.error_work);
+}
+
+void gen6_gt_irq_handler(struct intel_gt *gt, u32 gt_iir)
+{
+	if (gt_iir & GT_RENDER_USER_INTERRUPT)
+		intel_engine_cs_irq(gt->engine_class[RENDER_CLASS][0],
+				    gt_iir);
+
+	if (gt_iir & GT_BSD_USER_INTERRUPT)
+		intel_engine_cs_irq(gt->engine_class[VIDEO_DECODE_CLASS][0],
+				    gt_iir >> 12);
+
+	if (gt_iir & GT_BLT_USER_INTERRUPT)
+		intel_engine_cs_irq(gt->engine_class[COPY_ENGINE_CLASS][0],
+				    gt_iir >> 22);
+
+	if (gt_iir & (GT_BLT_CS_ERROR_INTERRUPT |
+		      GT_BSD_CS_ERROR_INTERRUPT |
+		      GT_CS_MASTER_ERROR_INTERRUPT))
+		DRM_DEBUG("Command parser error, gt_iir 0x%08x\n", gt_iir);
+
+	if (gt_iir & GT_PARITY_ERROR(gt->i915))
+		gen7_parity_error_irq_handler(gt, gt_iir);
+}
+
+void gen8_gt_irq_handler(struct intel_gt *gt, u32 master_ctl)
+{
+	void __iomem * const regs = gt->uncore->regs;
+	u32 iir;
+
+	if (master_ctl & (GEN8_GT_RCS_IRQ | GEN8_GT_BCS_IRQ)) {
+		iir = raw_reg_read(regs, GEN8_GT_IIR(0));
+		if (likely(iir)) {
+			intel_engine_cs_irq(gt->engine_class[RENDER_CLASS][0],
+					    iir >> GEN8_RCS_IRQ_SHIFT);
+			intel_engine_cs_irq(gt->engine_class[COPY_ENGINE_CLASS][0],
+					    iir >> GEN8_BCS_IRQ_SHIFT);
+			raw_reg_write(regs, GEN8_GT_IIR(0), iir);
+		}
+	}
+
+	if (master_ctl & (GEN8_GT_VCS0_IRQ | GEN8_GT_VCS1_IRQ)) {
+		iir = raw_reg_read(regs, GEN8_GT_IIR(1));
+		if (likely(iir)) {
+			intel_engine_cs_irq(gt->engine_class[VIDEO_DECODE_CLASS][0],
+					    iir >> GEN8_VCS0_IRQ_SHIFT);
+			intel_engine_cs_irq(gt->engine_class[VIDEO_DECODE_CLASS][1],
+					    iir >> GEN8_VCS1_IRQ_SHIFT);
+			raw_reg_write(regs, GEN8_GT_IIR(1), iir);
+		}
+	}
+
+	if (master_ctl & GEN8_GT_VECS_IRQ) {
+		iir = raw_reg_read(regs, GEN8_GT_IIR(3));
+		if (likely(iir)) {
+			intel_engine_cs_irq(gt->engine_class[VIDEO_ENHANCEMENT_CLASS][0],
+					    iir >> GEN8_VECS_IRQ_SHIFT);
+			raw_reg_write(regs, GEN8_GT_IIR(3), iir);
+		}
+	}
+
+	if (master_ctl & (GEN8_GT_PM_IRQ | GEN8_GT_GUC_IRQ)) {
+		iir = raw_reg_read(regs, GEN8_GT_IIR(2));
+		if (likely(iir)) {
+			gen6_rps_irq_handler(&gt->rps, iir);
+			guc_irq_handler(&gt->uc.guc, iir >> 16);
+			raw_reg_write(regs, GEN8_GT_IIR(2), iir);
+		}
+	}
+}
+
+void gen8_gt_irq_reset(struct intel_gt *gt)
+{
+	struct intel_uncore *uncore = gt->uncore;
+
+	GEN8_IRQ_RESET_NDX(uncore, GT, 0);
+	GEN8_IRQ_RESET_NDX(uncore, GT, 1);
+	GEN8_IRQ_RESET_NDX(uncore, GT, 2);
+	GEN8_IRQ_RESET_NDX(uncore, GT, 3);
+}
+
+void gen8_gt_irq_postinstall(struct intel_gt *gt)
+{
+	/* These are interrupts we'll toggle with the ring mask register */
+	const u32 irqs =
+		GT_CS_MASTER_ERROR_INTERRUPT |
+		GT_RENDER_USER_INTERRUPT |
+		GT_CONTEXT_SWITCH_INTERRUPT |
+		GT_WAIT_SEMAPHORE_INTERRUPT;
+	const u32 gt_interrupts[] = {
+		irqs << GEN8_RCS_IRQ_SHIFT | irqs << GEN8_BCS_IRQ_SHIFT,
+		irqs << GEN8_VCS0_IRQ_SHIFT | irqs << GEN8_VCS1_IRQ_SHIFT,
+		0,
+		irqs << GEN8_VECS_IRQ_SHIFT,
+	};
+	struct intel_uncore *uncore = gt->uncore;
+
+	gt->pm_ier = 0x0;
+	gt->pm_imr = ~gt->pm_ier;
+	GEN8_IRQ_INIT_NDX(uncore, GT, 0, ~gt_interrupts[0], gt_interrupts[0]);
+	GEN8_IRQ_INIT_NDX(uncore, GT, 1, ~gt_interrupts[1], gt_interrupts[1]);
+	/*
+	 * RPS interrupts will get enabled/disabled on demand when RPS itself
+	 * is enabled/disabled. Same wil be the case for GuC interrupts.
+	 */
+	GEN8_IRQ_INIT_NDX(uncore, GT, 2, gt->pm_imr, gt->pm_ier);
+	GEN8_IRQ_INIT_NDX(uncore, GT, 3, ~gt_interrupts[3], gt_interrupts[3]);
+}
+
+static void gen5_gt_update_irq(struct intel_gt *gt,
+			       u32 interrupt_mask,
+			       u32 enabled_irq_mask)
+{
+	lockdep_assert_held(gt->irq_lock);
+
+	GEM_BUG_ON(enabled_irq_mask & ~interrupt_mask);
+
+	gt->gt_imr &= ~interrupt_mask;
+	gt->gt_imr |= (~enabled_irq_mask & interrupt_mask);
+	intel_uncore_write(gt->uncore, GTIMR, gt->gt_imr);
+}
+
+void gen5_gt_enable_irq(struct intel_gt *gt, u32 mask)
+{
+	gen5_gt_update_irq(gt, mask, mask);
+	intel_uncore_posting_read_fw(gt->uncore, GTIMR);
+}
+
+void gen5_gt_disable_irq(struct intel_gt *gt, u32 mask)
+{
+	gen5_gt_update_irq(gt, mask, 0);
+}
+
+void gen5_gt_irq_reset(struct intel_gt *gt)
+{
+	struct intel_uncore *uncore = gt->uncore;
+
+	GEN3_IRQ_RESET(uncore, GT);
+	if (GRAPHICS_VER(gt->i915) >= 6)
+		GEN3_IRQ_RESET(uncore, GEN6_PM);
+}
+
+void gen5_gt_irq_postinstall(struct intel_gt *gt)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	u32 pm_irqs = 0;
+	u32 gt_irqs = 0;
+
+	gt->gt_imr = ~0;
+	if (HAS_L3_DPF(gt->i915)) {
+		/* L3 parity interrupt is always unmasked. */
+		gt->gt_imr = ~GT_PARITY_ERROR(gt->i915);
+		gt_irqs |= GT_PARITY_ERROR(gt->i915);
+	}
+
+	gt_irqs |= GT_RENDER_USER_INTERRUPT;
+	if (GRAPHICS_VER(gt->i915) == 5)
+		gt_irqs |= ILK_BSD_USER_INTERRUPT;
+	else
+		gt_irqs |= GT_BLT_USER_INTERRUPT | GT_BSD_USER_INTERRUPT;
+
+	GEN3_IRQ_INIT(uncore, GT, gt->gt_imr, gt_irqs);
+
+	if (GRAPHICS_VER(gt->i915) >= 6) {
+		/*
+		 * RPS interrupts will get enabled/disabled on demand when RPS
+		 * itself is enabled/disabled.
+		 */
+		if (HAS_ENGINE(gt, VECS0)) {
+			pm_irqs |= PM_VEBOX_USER_INTERRUPT;
+			gt->pm_ier |= PM_VEBOX_USER_INTERRUPT;
+		}
+
+		gt->pm_imr = 0xffffffff;
+		GEN3_IRQ_INIT(uncore, GEN6_PM, gt->pm_imr, pm_irqs);
+	}
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_irq.h b/drivers/gpu/drm/i915/gt/intel_gt_irq.h
new file mode 100644
index 000000000..41cad3866
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_irq.h
@@ -0,0 +1,65 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_GT_IRQ_H
+#define INTEL_GT_IRQ_H
+
+#include <linux/types.h>
+
+#include "intel_engine_types.h"
+
+struct intel_gt;
+
+#define GEN8_GT_IRQS (GEN8_GT_RCS_IRQ | \
+		      GEN8_GT_BCS_IRQ | \
+		      GEN8_GT_VCS0_IRQ | \
+		      GEN8_GT_VCS1_IRQ | \
+		      GEN8_GT_VECS_IRQ | \
+		      GEN8_GT_PM_IRQ | \
+		      GEN8_GT_GUC_IRQ)
+
+void gen11_gt_irq_reset(struct intel_gt *gt);
+void gen11_gt_irq_postinstall(struct intel_gt *gt);
+void gen11_gt_irq_handler(struct intel_gt *gt, const u32 master_ctl);
+
+bool gen11_gt_reset_one_iir(struct intel_gt *gt,
+			    const unsigned int bank,
+			    const unsigned int bit);
+
+void gen5_gt_irq_handler(struct intel_gt *gt, u32 gt_iir);
+
+void gen5_gt_irq_postinstall(struct intel_gt *gt);
+void gen5_gt_irq_reset(struct intel_gt *gt);
+void gen5_gt_disable_irq(struct intel_gt *gt, u32 mask);
+void gen5_gt_enable_irq(struct intel_gt *gt, u32 mask);
+
+void gen6_gt_irq_handler(struct intel_gt *gt, u32 gt_iir);
+
+void gen8_gt_irq_handler(struct intel_gt *gt, u32 master_ctl);
+void gen8_gt_irq_reset(struct intel_gt *gt);
+void gen8_gt_irq_postinstall(struct intel_gt *gt);
+
+static inline void intel_engine_cs_irq(struct intel_engine_cs *engine, u16 iir)
+{
+	if (iir)
+		engine->irq_handler(engine, iir);
+}
+
+static inline void
+intel_engine_set_irq_handler(struct intel_engine_cs *engine,
+			     void (*fn)(struct intel_engine_cs *engine,
+					u16 iir))
+{
+	/*
+	 * As the interrupt is live as allocate and setup the engines,
+	 * err on the side of caution and apply barriers to updating
+	 * the irq handler callback. This assures that when we do use
+	 * the engine, we will receive interrupts only to ourselves,
+	 * and not lose any.
+	 */
+	smp_store_mb(engine->irq_handler, fn);
+}
+
+#endif /* INTEL_GT_IRQ_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_mcr.c b/drivers/gpu/drm/i915/gt/intel_gt_mcr.c
new file mode 100644
index 000000000..e79405a45
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_mcr.c
@@ -0,0 +1,522 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include "i915_drv.h"
+
+#include "intel_gt_mcr.h"
+#include "intel_gt_regs.h"
+
+/**
+ * DOC: GT Multicast/Replicated (MCR) Register Support
+ *
+ * Some GT registers are designed as "multicast" or "replicated" registers:
+ * multiple instances of the same register share a single MMIO offset.  MCR
+ * registers are generally used when the hardware needs to potentially track
+ * independent values of a register per hardware unit (e.g., per-subslice,
+ * per-L3bank, etc.).  The specific types of replication that exist vary
+ * per-platform.
+ *
+ * MMIO accesses to MCR registers are controlled according to the settings
+ * programmed in the platform's MCR_SELECTOR register(s).  MMIO writes to MCR
+ * registers can be done in either a (i.e., a single write updates all
+ * instances of the register to the same value) or unicast (a write updates only
+ * one specific instance).  Reads of MCR registers always operate in a unicast
+ * manner regardless of how the multicast/unicast bit is set in MCR_SELECTOR.
+ * Selection of a specific MCR instance for unicast operations is referred to
+ * as "steering."
+ *
+ * If MCR register operations are steered toward a hardware unit that is
+ * fused off or currently powered down due to power gating, the MMIO operation
+ * is "terminated" by the hardware.  Terminated read operations will return a
+ * value of zero and terminated unicast write operations will be silently
+ * ignored.
+ */
+
+#define HAS_MSLICE_STEERING(dev_priv)	(INTEL_INFO(dev_priv)->has_mslice_steering)
+
+static const char * const intel_steering_types[] = {
+	"L3BANK",
+	"MSLICE",
+	"LNCF",
+	"INSTANCE 0",
+};
+
+static const struct intel_mmio_range icl_l3bank_steering_table[] = {
+	{ 0x00B100, 0x00B3FF },
+	{},
+};
+
+static const struct intel_mmio_range xehpsdv_mslice_steering_table[] = {
+	{ 0x004000, 0x004AFF },
+	{ 0x00C800, 0x00CFFF },
+	{ 0x00DD00, 0x00DDFF },
+	{ 0x00E900, 0x00FFFF }, /* 0xEA00 - OxEFFF is unused */
+	{},
+};
+
+static const struct intel_mmio_range xehpsdv_lncf_steering_table[] = {
+	{ 0x00B000, 0x00B0FF },
+	{ 0x00D800, 0x00D8FF },
+	{},
+};
+
+static const struct intel_mmio_range dg2_lncf_steering_table[] = {
+	{ 0x00B000, 0x00B0FF },
+	{ 0x00D880, 0x00D8FF },
+	{},
+};
+
+/*
+ * We have several types of MCR registers on PVC where steering to (0,0)
+ * will always provide us with a non-terminated value.  We'll stick them
+ * all in the same table for simplicity.
+ */
+static const struct intel_mmio_range pvc_instance0_steering_table[] = {
+	{ 0x004000, 0x004AFF },		/* HALF-BSLICE */
+	{ 0x008800, 0x00887F },		/* CC */
+	{ 0x008A80, 0x008AFF },		/* TILEPSMI */
+	{ 0x00B000, 0x00B0FF },		/* HALF-BSLICE */
+	{ 0x00B100, 0x00B3FF },		/* L3BANK */
+	{ 0x00C800, 0x00CFFF },		/* HALF-BSLICE */
+	{ 0x00D800, 0x00D8FF },		/* HALF-BSLICE */
+	{ 0x00DD00, 0x00DDFF },		/* BSLICE */
+	{ 0x00E900, 0x00E9FF },		/* HALF-BSLICE */
+	{ 0x00EC00, 0x00EEFF },		/* HALF-BSLICE */
+	{ 0x00F000, 0x00FFFF },		/* HALF-BSLICE */
+	{ 0x024180, 0x0241FF },		/* HALF-BSLICE */
+	{},
+};
+
+void intel_gt_mcr_init(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	/*
+	 * An mslice is unavailable only if both the meml3 for the slice is
+	 * disabled *and* all of the DSS in the slice (quadrant) are disabled.
+	 */
+	if (HAS_MSLICE_STEERING(i915)) {
+		gt->info.mslice_mask =
+			intel_slicemask_from_xehp_dssmask(gt->info.sseu.subslice_mask,
+							  GEN_DSS_PER_MSLICE);
+		gt->info.mslice_mask |=
+			(intel_uncore_read(gt->uncore, GEN10_MIRROR_FUSE3) &
+			 GEN12_MEML3_EN_MASK);
+
+		if (!gt->info.mslice_mask) /* should be impossible! */
+			drm_warn(&i915->drm, "mslice mask all zero!\n");
+	}
+
+	if (IS_PONTEVECCHIO(i915)) {
+		gt->steering_table[INSTANCE0] = pvc_instance0_steering_table;
+	} else if (IS_DG2(i915)) {
+		gt->steering_table[MSLICE] = xehpsdv_mslice_steering_table;
+		gt->steering_table[LNCF] = dg2_lncf_steering_table;
+	} else if (IS_XEHPSDV(i915)) {
+		gt->steering_table[MSLICE] = xehpsdv_mslice_steering_table;
+		gt->steering_table[LNCF] = xehpsdv_lncf_steering_table;
+	} else if (GRAPHICS_VER(i915) >= 11 &&
+		   GRAPHICS_VER_FULL(i915) < IP_VER(12, 50)) {
+		gt->steering_table[L3BANK] = icl_l3bank_steering_table;
+		gt->info.l3bank_mask =
+			~intel_uncore_read(gt->uncore, GEN10_MIRROR_FUSE3) &
+			GEN10_L3BANK_MASK;
+		if (!gt->info.l3bank_mask) /* should be impossible! */
+			drm_warn(&i915->drm, "L3 bank mask is all zero!\n");
+	} else if (GRAPHICS_VER(i915) >= 11) {
+		/*
+		 * We expect all modern platforms to have at least some
+		 * type of steering that needs to be initialized.
+		 */
+		MISSING_CASE(INTEL_INFO(i915)->platform);
+	}
+}
+
+/*
+ * rw_with_mcr_steering_fw - Access a register with specific MCR steering
+ * @uncore: pointer to struct intel_uncore
+ * @reg: register being accessed
+ * @rw_flag: FW_REG_READ for read access or FW_REG_WRITE for write access
+ * @group: group number (documented as "sliceid" on older platforms)
+ * @instance: instance number (documented as "subsliceid" on older platforms)
+ * @value: register value to be written (ignored for read)
+ *
+ * Return: 0 for write access. register value for read access.
+ *
+ * Caller needs to make sure the relevant forcewake wells are up.
+ */
+static u32 rw_with_mcr_steering_fw(struct intel_uncore *uncore,
+				   i915_reg_t reg, u8 rw_flag,
+				   int group, int instance, u32 value)
+{
+	u32 mcr_mask, mcr_ss, mcr, old_mcr, val = 0;
+
+	lockdep_assert_held(&uncore->lock);
+
+	if (GRAPHICS_VER(uncore->i915) >= 11) {
+		mcr_mask = GEN11_MCR_SLICE_MASK | GEN11_MCR_SUBSLICE_MASK;
+		mcr_ss = GEN11_MCR_SLICE(group) | GEN11_MCR_SUBSLICE(instance);
+
+		/*
+		 * Wa_22013088509
+		 *
+		 * The setting of the multicast/unicast bit usually wouldn't
+		 * matter for read operations (which always return the value
+		 * from a single register instance regardless of how that bit
+		 * is set), but some platforms have a workaround requiring us
+		 * to remain in multicast mode for reads.  There's no real
+		 * downside to this, so we'll just go ahead and do so on all
+		 * platforms; we'll only clear the multicast bit from the mask
+		 * when exlicitly doing a write operation.
+		 */
+		if (rw_flag == FW_REG_WRITE)
+			mcr_mask |= GEN11_MCR_MULTICAST;
+	} else {
+		mcr_mask = GEN8_MCR_SLICE_MASK | GEN8_MCR_SUBSLICE_MASK;
+		mcr_ss = GEN8_MCR_SLICE(group) | GEN8_MCR_SUBSLICE(instance);
+	}
+
+	old_mcr = mcr = intel_uncore_read_fw(uncore, GEN8_MCR_SELECTOR);
+
+	mcr &= ~mcr_mask;
+	mcr |= mcr_ss;
+	intel_uncore_write_fw(uncore, GEN8_MCR_SELECTOR, mcr);
+
+	if (rw_flag == FW_REG_READ)
+		val = intel_uncore_read_fw(uncore, reg);
+	else
+		intel_uncore_write_fw(uncore, reg, value);
+
+	mcr &= ~mcr_mask;
+	mcr |= old_mcr & mcr_mask;
+
+	intel_uncore_write_fw(uncore, GEN8_MCR_SELECTOR, mcr);
+
+	return val;
+}
+
+static u32 rw_with_mcr_steering(struct intel_uncore *uncore,
+				i915_reg_t reg, u8 rw_flag,
+				int group, int instance,
+				u32 value)
+{
+	enum forcewake_domains fw_domains;
+	u32 val;
+
+	fw_domains = intel_uncore_forcewake_for_reg(uncore, reg,
+						    rw_flag);
+	fw_domains |= intel_uncore_forcewake_for_reg(uncore,
+						     GEN8_MCR_SELECTOR,
+						     FW_REG_READ | FW_REG_WRITE);
+
+	spin_lock_irq(&uncore->lock);
+	intel_uncore_forcewake_get__locked(uncore, fw_domains);
+
+	val = rw_with_mcr_steering_fw(uncore, reg, rw_flag, group, instance, value);
+
+	intel_uncore_forcewake_put__locked(uncore, fw_domains);
+	spin_unlock_irq(&uncore->lock);
+
+	return val;
+}
+
+/**
+ * intel_gt_mcr_read - read a specific instance of an MCR register
+ * @gt: GT structure
+ * @reg: the MCR register to read
+ * @group: the MCR group
+ * @instance: the MCR instance
+ *
+ * Returns the value read from an MCR register after steering toward a specific
+ * group/instance.
+ */
+u32 intel_gt_mcr_read(struct intel_gt *gt,
+		      i915_reg_t reg,
+		      int group, int instance)
+{
+	return rw_with_mcr_steering(gt->uncore, reg, FW_REG_READ, group, instance, 0);
+}
+
+/**
+ * intel_gt_mcr_unicast_write - write a specific instance of an MCR register
+ * @gt: GT structure
+ * @reg: the MCR register to write
+ * @value: value to write
+ * @group: the MCR group
+ * @instance: the MCR instance
+ *
+ * Write an MCR register in unicast mode after steering toward a specific
+ * group/instance.
+ */
+void intel_gt_mcr_unicast_write(struct intel_gt *gt, i915_reg_t reg, u32 value,
+				int group, int instance)
+{
+	rw_with_mcr_steering(gt->uncore, reg, FW_REG_WRITE, group, instance, value);
+}
+
+/**
+ * intel_gt_mcr_multicast_write - write a value to all instances of an MCR register
+ * @gt: GT structure
+ * @reg: the MCR register to write
+ * @value: value to write
+ *
+ * Write an MCR register in multicast mode to update all instances.
+ */
+void intel_gt_mcr_multicast_write(struct intel_gt *gt,
+				i915_reg_t reg, u32 value)
+{
+	intel_uncore_write(gt->uncore, reg, value);
+}
+
+/**
+ * intel_gt_mcr_multicast_write_fw - write a value to all instances of an MCR register
+ * @gt: GT structure
+ * @reg: the MCR register to write
+ * @value: value to write
+ *
+ * Write an MCR register in multicast mode to update all instances.  This
+ * function assumes the caller is already holding any necessary forcewake
+ * domains; use intel_gt_mcr_multicast_write() in cases where forcewake should
+ * be obtained automatically.
+ */
+void intel_gt_mcr_multicast_write_fw(struct intel_gt *gt, i915_reg_t reg, u32 value)
+{
+	intel_uncore_write_fw(gt->uncore, reg, value);
+}
+
+/*
+ * reg_needs_read_steering - determine whether a register read requires
+ *     explicit steering
+ * @gt: GT structure
+ * @reg: the register to check steering requirements for
+ * @type: type of multicast steering to check
+ *
+ * Determines whether @reg needs explicit steering of a specific type for
+ * reads.
+ *
+ * Returns false if @reg does not belong to a register range of the given
+ * steering type, or if the default (subslice-based) steering IDs are suitable
+ * for @type steering too.
+ */
+static bool reg_needs_read_steering(struct intel_gt *gt,
+				    i915_reg_t reg,
+				    enum intel_steering_type type)
+{
+	const u32 offset = i915_mmio_reg_offset(reg);
+	const struct intel_mmio_range *entry;
+
+	if (likely(!gt->steering_table[type]))
+		return false;
+
+	for (entry = gt->steering_table[type]; entry->end; entry++) {
+		if (offset >= entry->start && offset <= entry->end)
+			return true;
+	}
+
+	return false;
+}
+
+/*
+ * get_nonterminated_steering - determines valid IDs for a class of MCR steering
+ * @gt: GT structure
+ * @type: multicast register type
+ * @group: Group ID returned
+ * @instance: Instance ID returned
+ *
+ * Determines group and instance values that will steer reads of the specified
+ * MCR class to a non-terminated instance.
+ */
+static void get_nonterminated_steering(struct intel_gt *gt,
+				       enum intel_steering_type type,
+				       u8 *group, u8 *instance)
+{
+	switch (type) {
+	case L3BANK:
+		*group = 0;		/* unused */
+		*instance = __ffs(gt->info.l3bank_mask);
+		break;
+	case MSLICE:
+		GEM_WARN_ON(!HAS_MSLICE_STEERING(gt->i915));
+		*group = __ffs(gt->info.mslice_mask);
+		*instance = 0;	/* unused */
+		break;
+	case LNCF:
+		/*
+		 * An LNCF is always present if its mslice is present, so we
+		 * can safely just steer to LNCF 0 in all cases.
+		 */
+		GEM_WARN_ON(!HAS_MSLICE_STEERING(gt->i915));
+		*group = __ffs(gt->info.mslice_mask) << 1;
+		*instance = 0;	/* unused */
+		break;
+	case INSTANCE0:
+		/*
+		 * There are a lot of MCR types for which instance (0, 0)
+		 * will always provide a non-terminated value.
+		 */
+		*group = 0;
+		*instance = 0;
+		break;
+	default:
+		MISSING_CASE(type);
+		*group = 0;
+		*instance = 0;
+	}
+}
+
+/**
+ * intel_gt_mcr_get_nonterminated_steering - find group/instance values that
+ *    will steer a register to a non-terminated instance
+ * @gt: GT structure
+ * @reg: register for which the steering is required
+ * @group: return variable for group steering
+ * @instance: return variable for instance steering
+ *
+ * This function returns a group/instance pair that is guaranteed to work for
+ * read steering of the given register. Note that a value will be returned even
+ * if the register is not replicated and therefore does not actually require
+ * steering.
+ */
+void intel_gt_mcr_get_nonterminated_steering(struct intel_gt *gt,
+					     i915_reg_t reg,
+					     u8 *group, u8 *instance)
+{
+	int type;
+
+	for (type = 0; type < NUM_STEERING_TYPES; type++) {
+		if (reg_needs_read_steering(gt, reg, type)) {
+			get_nonterminated_steering(gt, type, group, instance);
+			return;
+		}
+	}
+
+	*group = gt->default_steering.groupid;
+	*instance = gt->default_steering.instanceid;
+}
+
+/**
+ * intel_gt_mcr_read_any_fw - reads one instance of an MCR register
+ * @gt: GT structure
+ * @reg: register to read
+ *
+ * Reads a GT MCR register.  The read will be steered to a non-terminated
+ * instance (i.e., one that isn't fused off or powered down by power gating).
+ * This function assumes the caller is already holding any necessary forcewake
+ * domains; use intel_gt_mcr_read_any() in cases where forcewake should be
+ * obtained automatically.
+ *
+ * Returns the value from a non-terminated instance of @reg.
+ */
+u32 intel_gt_mcr_read_any_fw(struct intel_gt *gt, i915_reg_t reg)
+{
+	int type;
+	u8 group, instance;
+
+	for (type = 0; type < NUM_STEERING_TYPES; type++) {
+		if (reg_needs_read_steering(gt, reg, type)) {
+			get_nonterminated_steering(gt, type, &group, &instance);
+			return rw_with_mcr_steering_fw(gt->uncore, reg,
+						       FW_REG_READ,
+						       group, instance, 0);
+		}
+	}
+
+	return intel_uncore_read_fw(gt->uncore, reg);
+}
+
+/**
+ * intel_gt_mcr_read_any - reads one instance of an MCR register
+ * @gt: GT structure
+ * @reg: register to read
+ *
+ * Reads a GT MCR register.  The read will be steered to a non-terminated
+ * instance (i.e., one that isn't fused off or powered down by power gating).
+ *
+ * Returns the value from a non-terminated instance of @reg.
+ */
+u32 intel_gt_mcr_read_any(struct intel_gt *gt, i915_reg_t reg)
+{
+	int type;
+	u8 group, instance;
+
+	for (type = 0; type < NUM_STEERING_TYPES; type++) {
+		if (reg_needs_read_steering(gt, reg, type)) {
+			get_nonterminated_steering(gt, type, &group, &instance);
+			return rw_with_mcr_steering(gt->uncore, reg,
+						    FW_REG_READ,
+						    group, instance, 0);
+		}
+	}
+
+	return intel_uncore_read(gt->uncore, reg);
+}
+
+static void report_steering_type(struct drm_printer *p,
+				 struct intel_gt *gt,
+				 enum intel_steering_type type,
+				 bool dump_table)
+{
+	const struct intel_mmio_range *entry;
+	u8 group, instance;
+
+	BUILD_BUG_ON(ARRAY_SIZE(intel_steering_types) != NUM_STEERING_TYPES);
+
+	if (!gt->steering_table[type]) {
+		drm_printf(p, "%s steering: uses default steering\n",
+			   intel_steering_types[type]);
+		return;
+	}
+
+	get_nonterminated_steering(gt, type, &group, &instance);
+	drm_printf(p, "%s steering: group=0x%x, instance=0x%x\n",
+		   intel_steering_types[type], group, instance);
+
+	if (!dump_table)
+		return;
+
+	for (entry = gt->steering_table[type]; entry->end; entry++)
+		drm_printf(p, "\t0x%06x - 0x%06x\n", entry->start, entry->end);
+}
+
+void intel_gt_mcr_report_steering(struct drm_printer *p, struct intel_gt *gt,
+				  bool dump_table)
+{
+	drm_printf(p, "Default steering: group=0x%x, instance=0x%x\n",
+		   gt->default_steering.groupid,
+		   gt->default_steering.instanceid);
+
+	if (IS_PONTEVECCHIO(gt->i915)) {
+		report_steering_type(p, gt, INSTANCE0, dump_table);
+	} else if (HAS_MSLICE_STEERING(gt->i915)) {
+		report_steering_type(p, gt, MSLICE, dump_table);
+		report_steering_type(p, gt, LNCF, dump_table);
+	}
+}
+
+/**
+ * intel_gt_mcr_get_ss_steering - returns the group/instance steering for a SS
+ * @gt: GT structure
+ * @dss: DSS ID to obtain steering for
+ * @group: pointer to storage for steering group ID
+ * @instance: pointer to storage for steering instance ID
+ *
+ * Returns the steering IDs (via the @group and @instance parameters) that
+ * correspond to a specific subslice/DSS ID.
+ */
+void intel_gt_mcr_get_ss_steering(struct intel_gt *gt, unsigned int dss,
+				   unsigned int *group, unsigned int *instance)
+{
+	if (IS_PONTEVECCHIO(gt->i915)) {
+		*group = dss / GEN_DSS_PER_CSLICE;
+		*instance = dss % GEN_DSS_PER_CSLICE;
+	} else if (GRAPHICS_VER_FULL(gt->i915) >= IP_VER(12, 50)) {
+		*group = dss / GEN_DSS_PER_GSLICE;
+		*instance = dss % GEN_DSS_PER_GSLICE;
+	} else {
+		*group = dss / GEN_MAX_SS_PER_HSW_SLICE;
+		*instance = dss % GEN_MAX_SS_PER_HSW_SLICE;
+		return;
+	}
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_mcr.h b/drivers/gpu/drm/i915/gt/intel_gt_mcr.h
new file mode 100644
index 000000000..77a8b11c2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_mcr.h
@@ -0,0 +1,58 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_GT_MCR__
+#define __INTEL_GT_MCR__
+
+#include "intel_gt_types.h"
+
+void intel_gt_mcr_init(struct intel_gt *gt);
+
+u32 intel_gt_mcr_read(struct intel_gt *gt,
+		      i915_reg_t reg,
+		      int group, int instance);
+u32 intel_gt_mcr_read_any_fw(struct intel_gt *gt, i915_reg_t reg);
+u32 intel_gt_mcr_read_any(struct intel_gt *gt, i915_reg_t reg);
+
+void intel_gt_mcr_unicast_write(struct intel_gt *gt,
+				i915_reg_t reg, u32 value,
+				int group, int instance);
+void intel_gt_mcr_multicast_write(struct intel_gt *gt,
+				  i915_reg_t reg, u32 value);
+void intel_gt_mcr_multicast_write_fw(struct intel_gt *gt,
+				     i915_reg_t reg, u32 value);
+
+void intel_gt_mcr_get_nonterminated_steering(struct intel_gt *gt,
+					     i915_reg_t reg,
+					     u8 *group, u8 *instance);
+
+void intel_gt_mcr_report_steering(struct drm_printer *p, struct intel_gt *gt,
+				  bool dump_table);
+
+void intel_gt_mcr_get_ss_steering(struct intel_gt *gt, unsigned int dss,
+				  unsigned int *group, unsigned int *instance);
+
+/*
+ * Helper for for_each_ss_steering loop.  On pre-Xe_HP platforms, subslice
+ * presence is determined by using the group/instance as direct lookups in the
+ * slice/subslice topology.  On Xe_HP and beyond, the steering is unrelated to
+ * the topology, so we lookup the DSS ID directly in "slice 0."
+ */
+#define _HAS_SS(ss_, gt_, group_, instance_) ( \
+	GRAPHICS_VER_FULL(gt_->i915) >= IP_VER(12, 50) ? \
+		intel_sseu_has_subslice(&(gt_)->info.sseu, 0, ss_) : \
+		intel_sseu_has_subslice(&(gt_)->info.sseu, group_, instance_))
+
+/*
+ * Loop over each subslice/DSS and determine the group and instance IDs that
+ * should be used to steer MCR accesses toward this DSS.
+ */
+#define for_each_ss_steering(ss_, gt_, group_, instance_) \
+	for (ss_ = 0, intel_gt_mcr_get_ss_steering(gt_, 0, &group_, &instance_); \
+	     ss_ < I915_MAX_SS_FUSE_BITS; \
+	     ss_++, intel_gt_mcr_get_ss_steering(gt_, ss_, &group_, &instance_)) \
+		for_each_if(_HAS_SS(ss_, gt_, group_, instance_))
+
+#endif /* __INTEL_GT_MCR__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_pm.c b/drivers/gpu/drm/i915/gt/intel_gt_pm.c
new file mode 100644
index 000000000..f553e2173
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_pm.c
@@ -0,0 +1,417 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+#include <linux/suspend.h>
+
+#include "i915_drv.h"
+#include "i915_params.h"
+#include "intel_context.h"
+#include "intel_engine_pm.h"
+#include "intel_gt.h"
+#include "intel_gt_clock_utils.h"
+#include "intel_gt_pm.h"
+#include "intel_gt_requests.h"
+#include "intel_llc.h"
+#include "intel_pm.h"
+#include "intel_rc6.h"
+#include "intel_rps.h"
+#include "intel_wakeref.h"
+#include "pxp/intel_pxp_pm.h"
+
+#define I915_GT_SUSPEND_IDLE_TIMEOUT (HZ / 2)
+
+static void user_forcewake(struct intel_gt *gt, bool suspend)
+{
+	int count = atomic_read(&gt->user_wakeref);
+
+	/* Inside suspend/resume so single threaded, no races to worry about. */
+	if (likely(!count))
+		return;
+
+	intel_gt_pm_get(gt);
+	if (suspend) {
+		GEM_BUG_ON(count > atomic_read(&gt->wakeref.count));
+		atomic_sub(count, &gt->wakeref.count);
+	} else {
+		atomic_add(count, &gt->wakeref.count);
+	}
+	intel_gt_pm_put(gt);
+}
+
+static void runtime_begin(struct intel_gt *gt)
+{
+	local_irq_disable();
+	write_seqcount_begin(&gt->stats.lock);
+	gt->stats.start = ktime_get();
+	gt->stats.active = true;
+	write_seqcount_end(&gt->stats.lock);
+	local_irq_enable();
+}
+
+static void runtime_end(struct intel_gt *gt)
+{
+	local_irq_disable();
+	write_seqcount_begin(&gt->stats.lock);
+	gt->stats.active = false;
+	gt->stats.total =
+		ktime_add(gt->stats.total,
+			  ktime_sub(ktime_get(), gt->stats.start));
+	write_seqcount_end(&gt->stats.lock);
+	local_irq_enable();
+}
+
+static int __gt_unpark(struct intel_wakeref *wf)
+{
+	struct intel_gt *gt = container_of(wf, typeof(*gt), wakeref);
+	struct drm_i915_private *i915 = gt->i915;
+
+	GT_TRACE(gt, "\n");
+
+	/*
+	 * It seems that the DMC likes to transition between the DC states a lot
+	 * when there are no connected displays (no active power domains) during
+	 * command submission.
+	 *
+	 * This activity has negative impact on the performance of the chip with
+	 * huge latencies observed in the interrupt handler and elsewhere.
+	 *
+	 * Work around it by grabbing a GT IRQ power domain whilst there is any
+	 * GT activity, preventing any DC state transitions.
+	 */
+	gt->awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ);
+	GEM_BUG_ON(!gt->awake);
+
+	intel_rc6_unpark(&gt->rc6);
+	intel_rps_unpark(&gt->rps);
+	i915_pmu_gt_unparked(i915);
+	intel_guc_busyness_unpark(gt);
+
+	intel_gt_unpark_requests(gt);
+	runtime_begin(gt);
+
+	return 0;
+}
+
+static int __gt_park(struct intel_wakeref *wf)
+{
+	struct intel_gt *gt = container_of(wf, typeof(*gt), wakeref);
+	intel_wakeref_t wakeref = fetch_and_zero(&gt->awake);
+	struct drm_i915_private *i915 = gt->i915;
+
+	GT_TRACE(gt, "\n");
+
+	runtime_end(gt);
+	intel_gt_park_requests(gt);
+
+	intel_guc_busyness_park(gt);
+	i915_vma_parked(gt);
+	i915_pmu_gt_parked(i915);
+	intel_rps_park(&gt->rps);
+	intel_rc6_park(&gt->rc6);
+
+	/* Everything switched off, flush any residual interrupt just in case */
+	intel_synchronize_irq(i915);
+
+	/* Defer dropping the display power well for 100ms, it's slow! */
+	GEM_BUG_ON(!wakeref);
+	intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
+
+	return 0;
+}
+
+static const struct intel_wakeref_ops wf_ops = {
+	.get = __gt_unpark,
+	.put = __gt_park,
+};
+
+void intel_gt_pm_init_early(struct intel_gt *gt)
+{
+	/*
+	 * We access the runtime_pm structure via gt->i915 here rather than
+	 * gt->uncore as we do elsewhere in the file because gt->uncore is not
+	 * yet initialized for all tiles at this point in the driver startup.
+	 * runtime_pm is per-device rather than per-tile, so this is still the
+	 * correct structure.
+	 */
+	intel_wakeref_init(&gt->wakeref, &gt->i915->runtime_pm, &wf_ops);
+	seqcount_mutex_init(&gt->stats.lock, &gt->wakeref.mutex);
+}
+
+void intel_gt_pm_init(struct intel_gt *gt)
+{
+	/*
+	 * Enabling power-management should be "self-healing". If we cannot
+	 * enable a feature, simply leave it disabled with a notice to the
+	 * user.
+	 */
+	intel_rc6_init(&gt->rc6);
+	intel_rps_init(&gt->rps);
+}
+
+static bool reset_engines(struct intel_gt *gt)
+{
+	if (INTEL_INFO(gt->i915)->gpu_reset_clobbers_display)
+		return false;
+
+	return __intel_gt_reset(gt, ALL_ENGINES) == 0;
+}
+
+static void gt_sanitize(struct intel_gt *gt, bool force)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	intel_wakeref_t wakeref;
+
+	GT_TRACE(gt, "force:%s", str_yes_no(force));
+
+	/* Use a raw wakeref to avoid calling intel_display_power_get early */
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+	intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);
+
+	intel_gt_check_clock_frequency(gt);
+
+	/*
+	 * As we have just resumed the machine and woken the device up from
+	 * deep PCI sleep (presumably D3_cold), assume the HW has been reset
+	 * back to defaults, recovering from whatever wedged state we left it
+	 * in and so worth trying to use the device once more.
+	 */
+	if (intel_gt_is_wedged(gt))
+		intel_gt_unset_wedged(gt);
+
+	/* For GuC mode, ensure submission is disabled before stopping ring */
+	intel_uc_reset_prepare(&gt->uc);
+
+	for_each_engine(engine, gt, id) {
+		if (engine->reset.prepare)
+			engine->reset.prepare(engine);
+
+		if (engine->sanitize)
+			engine->sanitize(engine);
+	}
+
+	if (reset_engines(gt) || force) {
+		for_each_engine(engine, gt, id)
+			__intel_engine_reset(engine, false);
+	}
+
+	intel_uc_reset(&gt->uc, false);
+
+	for_each_engine(engine, gt, id)
+		if (engine->reset.finish)
+			engine->reset.finish(engine);
+
+	intel_rps_sanitize(&gt->rps);
+
+	intel_uncore_forcewake_put(gt->uncore, FORCEWAKE_ALL);
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+}
+
+void intel_gt_pm_fini(struct intel_gt *gt)
+{
+	intel_rc6_fini(&gt->rc6);
+}
+
+int intel_gt_resume(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err;
+
+	err = intel_gt_has_unrecoverable_error(gt);
+	if (err)
+		return err;
+
+	GT_TRACE(gt, "\n");
+
+	/*
+	 * After resume, we may need to poke into the pinned kernel
+	 * contexts to paper over any damage caused by the sudden suspend.
+	 * Only the kernel contexts should remain pinned over suspend,
+	 * allowing us to fixup the user contexts on their first pin.
+	 */
+	gt_sanitize(gt, true);
+
+	intel_gt_pm_get(gt);
+
+	intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);
+	intel_rc6_sanitize(&gt->rc6);
+	if (intel_gt_is_wedged(gt)) {
+		err = -EIO;
+		goto out_fw;
+	}
+
+	/* Only when the HW is re-initialised, can we replay the requests */
+	err = intel_gt_init_hw(gt);
+	if (err) {
+		i915_probe_error(gt->i915,
+				 "Failed to initialize GPU, declaring it wedged!\n");
+		goto err_wedged;
+	}
+
+	intel_uc_reset_finish(&gt->uc);
+
+	intel_rps_enable(&gt->rps);
+	intel_llc_enable(&gt->llc);
+
+	for_each_engine(engine, gt, id) {
+		intel_engine_pm_get(engine);
+
+		engine->serial++; /* kernel context lost */
+		err = intel_engine_resume(engine);
+
+		intel_engine_pm_put(engine);
+		if (err) {
+			drm_err(&gt->i915->drm,
+				"Failed to restart %s (%d)\n",
+				engine->name, err);
+			goto err_wedged;
+		}
+	}
+
+	intel_rc6_enable(&gt->rc6);
+
+	intel_uc_resume(&gt->uc);
+
+	intel_pxp_resume(&gt->pxp);
+
+	user_forcewake(gt, false);
+
+out_fw:
+	intel_uncore_forcewake_put(gt->uncore, FORCEWAKE_ALL);
+	intel_gt_pm_put(gt);
+	return err;
+
+err_wedged:
+	intel_gt_set_wedged(gt);
+	goto out_fw;
+}
+
+static void wait_for_suspend(struct intel_gt *gt)
+{
+	if (!intel_gt_pm_is_awake(gt))
+		return;
+
+	if (intel_gt_wait_for_idle(gt, I915_GT_SUSPEND_IDLE_TIMEOUT) == -ETIME) {
+		/*
+		 * Forcibly cancel outstanding work and leave
+		 * the gpu quiet.
+		 */
+		intel_gt_set_wedged(gt);
+		intel_gt_retire_requests(gt);
+	}
+
+	intel_gt_pm_wait_for_idle(gt);
+}
+
+void intel_gt_suspend_prepare(struct intel_gt *gt)
+{
+	user_forcewake(gt, true);
+	wait_for_suspend(gt);
+
+	intel_pxp_suspend_prepare(&gt->pxp);
+}
+
+static suspend_state_t pm_suspend_target(void)
+{
+#if IS_ENABLED(CONFIG_SUSPEND) && IS_ENABLED(CONFIG_PM_SLEEP)
+	return pm_suspend_target_state;
+#else
+	return PM_SUSPEND_TO_IDLE;
+#endif
+}
+
+void intel_gt_suspend_late(struct intel_gt *gt)
+{
+	intel_wakeref_t wakeref;
+
+	/* We expect to be idle already; but also want to be independent */
+	wait_for_suspend(gt);
+
+	if (is_mock_gt(gt))
+		return;
+
+	GEM_BUG_ON(gt->awake);
+
+	intel_uc_suspend(&gt->uc);
+	intel_pxp_suspend(&gt->pxp);
+
+	/*
+	 * On disabling the device, we want to turn off HW access to memory
+	 * that we no longer own.
+	 *
+	 * However, not all suspend-states disable the device. S0 (s2idle)
+	 * is effectively runtime-suspend, the device is left powered on
+	 * but needs to be put into a low power state. We need to keep
+	 * powermanagement enabled, but we also retain system state and so
+	 * it remains safe to keep on using our allocated memory.
+	 */
+	if (pm_suspend_target() == PM_SUSPEND_TO_IDLE)
+		return;
+
+	with_intel_runtime_pm(gt->uncore->rpm, wakeref) {
+		intel_rps_disable(&gt->rps);
+		intel_rc6_disable(&gt->rc6);
+		intel_llc_disable(&gt->llc);
+	}
+
+	gt_sanitize(gt, false);
+
+	GT_TRACE(gt, "\n");
+}
+
+void intel_gt_runtime_suspend(struct intel_gt *gt)
+{
+	intel_pxp_runtime_suspend(&gt->pxp);
+	intel_uc_runtime_suspend(&gt->uc);
+
+	GT_TRACE(gt, "\n");
+}
+
+int intel_gt_runtime_resume(struct intel_gt *gt)
+{
+	int ret;
+
+	GT_TRACE(gt, "\n");
+	intel_gt_init_swizzling(gt);
+	intel_ggtt_restore_fences(gt->ggtt);
+
+	ret = intel_uc_runtime_resume(&gt->uc);
+	if (ret)
+		return ret;
+
+	intel_pxp_runtime_resume(&gt->pxp);
+
+	return 0;
+}
+
+static ktime_t __intel_gt_get_awake_time(const struct intel_gt *gt)
+{
+	ktime_t total = gt->stats.total;
+
+	if (gt->stats.active)
+		total = ktime_add(total,
+				  ktime_sub(ktime_get(), gt->stats.start));
+
+	return total;
+}
+
+ktime_t intel_gt_get_awake_time(const struct intel_gt *gt)
+{
+	unsigned int seq;
+	ktime_t total;
+
+	do {
+		seq = read_seqcount_begin(&gt->stats.lock);
+		total = __intel_gt_get_awake_time(gt);
+	} while (read_seqcount_retry(&gt->stats.lock, seq));
+
+	return total;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_gt_pm.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_pm.h b/drivers/gpu/drm/i915/gt/intel_gt_pm.h
new file mode 100644
index 000000000..6c9a46452
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_pm.h
@@ -0,0 +1,92 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_GT_PM_H
+#define INTEL_GT_PM_H
+
+#include <linux/types.h>
+
+#include "intel_gt_types.h"
+#include "intel_wakeref.h"
+
+static inline bool intel_gt_pm_is_awake(const struct intel_gt *gt)
+{
+	return intel_wakeref_is_active(&gt->wakeref);
+}
+
+static inline void intel_gt_pm_get(struct intel_gt *gt)
+{
+	intel_wakeref_get(&gt->wakeref);
+}
+
+static inline void __intel_gt_pm_get(struct intel_gt *gt)
+{
+	__intel_wakeref_get(&gt->wakeref);
+}
+
+static inline bool intel_gt_pm_get_if_awake(struct intel_gt *gt)
+{
+	return intel_wakeref_get_if_active(&gt->wakeref);
+}
+
+static inline void intel_gt_pm_might_get(struct intel_gt *gt)
+{
+	intel_wakeref_might_get(&gt->wakeref);
+}
+
+static inline void intel_gt_pm_put(struct intel_gt *gt)
+{
+	intel_wakeref_put(&gt->wakeref);
+}
+
+static inline void intel_gt_pm_put_async(struct intel_gt *gt)
+{
+	intel_wakeref_put_async(&gt->wakeref);
+}
+
+static inline void intel_gt_pm_might_put(struct intel_gt *gt)
+{
+	intel_wakeref_might_put(&gt->wakeref);
+}
+
+#define with_intel_gt_pm(gt, tmp) \
+	for (tmp = 1, intel_gt_pm_get(gt); tmp; \
+	     intel_gt_pm_put(gt), tmp = 0)
+
+/**
+ * with_intel_gt_pm_if_awake - if GT is PM awake, get a reference to prevent
+ *	it to sleep, run some code and then asynchrously put the reference
+ *	away.
+ *
+ * @gt: pointer to the gt
+ * @wf: pointer to a temporary wakeref.
+ */
+#define with_intel_gt_pm_if_awake(gt, wf) \
+	for (wf = intel_gt_pm_get_if_awake(gt); wf; intel_gt_pm_put_async(gt), wf = 0)
+
+static inline int intel_gt_pm_wait_for_idle(struct intel_gt *gt)
+{
+	return intel_wakeref_wait_for_idle(&gt->wakeref);
+}
+
+void intel_gt_pm_init_early(struct intel_gt *gt);
+void intel_gt_pm_init(struct intel_gt *gt);
+void intel_gt_pm_fini(struct intel_gt *gt);
+
+void intel_gt_suspend_prepare(struct intel_gt *gt);
+void intel_gt_suspend_late(struct intel_gt *gt);
+int intel_gt_resume(struct intel_gt *gt);
+
+void intel_gt_runtime_suspend(struct intel_gt *gt);
+int intel_gt_runtime_resume(struct intel_gt *gt);
+
+ktime_t intel_gt_get_awake_time(const struct intel_gt *gt);
+
+static inline bool is_mock_gt(const struct intel_gt *gt)
+{
+	return I915_SELFTEST_ONLY(gt->awake == -ENODEV);
+}
+
+#endif /* INTEL_GT_PM_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_pm_debugfs.c b/drivers/gpu/drm/i915/gt/intel_gt_pm_debugfs.c
new file mode 100644
index 000000000..108b9e76c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_pm_debugfs.c
@@ -0,0 +1,670 @@
+// SPDX-License-Identifier: MIT
+
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/seq_file.h>
+#include <linux/string_helpers.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_gt.h"
+#include "intel_gt_clock_utils.h"
+#include "intel_gt_debugfs.h"
+#include "intel_gt_pm.h"
+#include "intel_gt_pm_debugfs.h"
+#include "intel_gt_regs.h"
+#include "intel_llc.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pcode.h"
+#include "intel_rc6.h"
+#include "intel_rps.h"
+#include "intel_runtime_pm.h"
+#include "intel_uncore.h"
+#include "vlv_sideband.h"
+
+void intel_gt_pm_debugfs_forcewake_user_open(struct intel_gt *gt)
+{
+	atomic_inc(&gt->user_wakeref);
+	intel_gt_pm_get(gt);
+	if (GRAPHICS_VER(gt->i915) >= 6)
+		intel_uncore_forcewake_user_get(gt->uncore);
+}
+
+void intel_gt_pm_debugfs_forcewake_user_release(struct intel_gt *gt)
+{
+	if (GRAPHICS_VER(gt->i915) >= 6)
+		intel_uncore_forcewake_user_put(gt->uncore);
+	intel_gt_pm_put(gt);
+	atomic_dec(&gt->user_wakeref);
+}
+
+static int forcewake_user_open(struct inode *inode, struct file *file)
+{
+	struct intel_gt *gt = inode->i_private;
+
+	intel_gt_pm_debugfs_forcewake_user_open(gt);
+
+	return 0;
+}
+
+static int forcewake_user_release(struct inode *inode, struct file *file)
+{
+	struct intel_gt *gt = inode->i_private;
+
+	intel_gt_pm_debugfs_forcewake_user_release(gt);
+
+	return 0;
+}
+
+static const struct file_operations forcewake_user_fops = {
+	.owner = THIS_MODULE,
+	.open = forcewake_user_open,
+	.release = forcewake_user_release,
+};
+
+static int fw_domains_show(struct seq_file *m, void *data)
+{
+	struct intel_gt *gt = m->private;
+	struct intel_uncore *uncore = gt->uncore;
+	struct intel_uncore_forcewake_domain *fw_domain;
+	unsigned int tmp;
+
+	seq_printf(m, "user.bypass_count = %u\n",
+		   uncore->user_forcewake_count);
+
+	for_each_fw_domain(fw_domain, uncore, tmp)
+		seq_printf(m, "%s.wake_count = %u\n",
+			   intel_uncore_forcewake_domain_to_str(fw_domain->id),
+			   READ_ONCE(fw_domain->wake_count));
+
+	return 0;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(fw_domains);
+
+static void print_rc6_res(struct seq_file *m,
+			  const char *title,
+			  const i915_reg_t reg)
+{
+	struct intel_gt *gt = m->private;
+	intel_wakeref_t wakeref;
+
+	with_intel_runtime_pm(gt->uncore->rpm, wakeref)
+		seq_printf(m, "%s %u (%llu us)\n", title,
+			   intel_uncore_read(gt->uncore, reg),
+			   intel_rc6_residency_us(&gt->rc6, reg));
+}
+
+static int vlv_drpc(struct seq_file *m)
+{
+	struct intel_gt *gt = m->private;
+	struct intel_uncore *uncore = gt->uncore;
+	u32 rcctl1, pw_status, mt_fwake_req;
+
+	mt_fwake_req = intel_uncore_read_fw(uncore, FORCEWAKE_MT);
+	pw_status = intel_uncore_read(uncore, VLV_GTLC_PW_STATUS);
+	rcctl1 = intel_uncore_read(uncore, GEN6_RC_CONTROL);
+
+	seq_printf(m, "RC6 Enabled: %s\n",
+		   str_yes_no(rcctl1 & (GEN7_RC_CTL_TO_MODE |
+					GEN6_RC_CTL_EI_MODE(1))));
+	seq_printf(m, "Multi-threaded Forcewake Request: 0x%x\n", mt_fwake_req);
+	seq_printf(m, "Render Power Well: %s\n",
+		   (pw_status & VLV_GTLC_PW_RENDER_STATUS_MASK) ? "Up" : "Down");
+	seq_printf(m, "Media Power Well: %s\n",
+		   (pw_status & VLV_GTLC_PW_MEDIA_STATUS_MASK) ? "Up" : "Down");
+
+	print_rc6_res(m, "Render RC6 residency since boot:", GEN6_GT_GFX_RC6);
+	print_rc6_res(m, "Media RC6 residency since boot:", VLV_GT_MEDIA_RC6);
+
+	return fw_domains_show(m, NULL);
+}
+
+static int gen6_drpc(struct seq_file *m)
+{
+	struct intel_gt *gt = m->private;
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+	u32 gt_core_status, mt_fwake_req, rcctl1, rc6vids = 0;
+	u32 gen9_powergate_enable = 0, gen9_powergate_status = 0;
+
+	mt_fwake_req = intel_uncore_read_fw(uncore, FORCEWAKE_MT);
+	gt_core_status = intel_uncore_read_fw(uncore, GEN6_GT_CORE_STATUS);
+
+	rcctl1 = intel_uncore_read(uncore, GEN6_RC_CONTROL);
+	if (GRAPHICS_VER(i915) >= 9) {
+		gen9_powergate_enable =
+			intel_uncore_read(uncore, GEN9_PG_ENABLE);
+		gen9_powergate_status =
+			intel_uncore_read(uncore, GEN9_PWRGT_DOMAIN_STATUS);
+	}
+
+	if (GRAPHICS_VER(i915) <= 7)
+		snb_pcode_read(gt->uncore, GEN6_PCODE_READ_RC6VIDS, &rc6vids, NULL);
+
+	seq_printf(m, "RC1e Enabled: %s\n",
+		   str_yes_no(rcctl1 & GEN6_RC_CTL_RC1e_ENABLE));
+	seq_printf(m, "RC6 Enabled: %s\n",
+		   str_yes_no(rcctl1 & GEN6_RC_CTL_RC6_ENABLE));
+	if (GRAPHICS_VER(i915) >= 9) {
+		seq_printf(m, "Render Well Gating Enabled: %s\n",
+			   str_yes_no(gen9_powergate_enable & GEN9_RENDER_PG_ENABLE));
+		seq_printf(m, "Media Well Gating Enabled: %s\n",
+			   str_yes_no(gen9_powergate_enable & GEN9_MEDIA_PG_ENABLE));
+	}
+	seq_printf(m, "Deep RC6 Enabled: %s\n",
+		   str_yes_no(rcctl1 & GEN6_RC_CTL_RC6p_ENABLE));
+	seq_printf(m, "Deepest RC6 Enabled: %s\n",
+		   str_yes_no(rcctl1 & GEN6_RC_CTL_RC6pp_ENABLE));
+	seq_puts(m, "Current RC state: ");
+	switch (gt_core_status & GEN6_RCn_MASK) {
+	case GEN6_RC0:
+		if (gt_core_status & GEN6_CORE_CPD_STATE_MASK)
+			seq_puts(m, "Core Power Down\n");
+		else
+			seq_puts(m, "on\n");
+		break;
+	case GEN6_RC3:
+		seq_puts(m, "RC3\n");
+		break;
+	case GEN6_RC6:
+		seq_puts(m, "RC6\n");
+		break;
+	case GEN6_RC7:
+		seq_puts(m, "RC7\n");
+		break;
+	default:
+		seq_puts(m, "Unknown\n");
+		break;
+	}
+
+	seq_printf(m, "Core Power Down: %s\n",
+		   str_yes_no(gt_core_status & GEN6_CORE_CPD_STATE_MASK));
+	seq_printf(m, "Multi-threaded Forcewake Request: 0x%x\n", mt_fwake_req);
+	if (GRAPHICS_VER(i915) >= 9) {
+		seq_printf(m, "Render Power Well: %s\n",
+			   (gen9_powergate_status &
+			    GEN9_PWRGT_RENDER_STATUS_MASK) ? "Up" : "Down");
+		seq_printf(m, "Media Power Well: %s\n",
+			   (gen9_powergate_status &
+			    GEN9_PWRGT_MEDIA_STATUS_MASK) ? "Up" : "Down");
+	}
+
+	/* Not exactly sure what this is */
+	print_rc6_res(m, "RC6 \"Locked to RPn\" residency since boot:",
+		      GEN6_GT_GFX_RC6_LOCKED);
+	print_rc6_res(m, "RC6 residency since boot:", GEN6_GT_GFX_RC6);
+	print_rc6_res(m, "RC6+ residency since boot:", GEN6_GT_GFX_RC6p);
+	print_rc6_res(m, "RC6++ residency since boot:", GEN6_GT_GFX_RC6pp);
+
+	if (GRAPHICS_VER(i915) <= 7) {
+		seq_printf(m, "RC6   voltage: %dmV\n",
+			   GEN6_DECODE_RC6_VID(((rc6vids >> 0) & 0xff)));
+		seq_printf(m, "RC6+  voltage: %dmV\n",
+			   GEN6_DECODE_RC6_VID(((rc6vids >> 8) & 0xff)));
+		seq_printf(m, "RC6++ voltage: %dmV\n",
+			   GEN6_DECODE_RC6_VID(((rc6vids >> 16) & 0xff)));
+	}
+
+	return fw_domains_show(m, NULL);
+}
+
+static int ilk_drpc(struct seq_file *m)
+{
+	struct intel_gt *gt = m->private;
+	struct intel_uncore *uncore = gt->uncore;
+	u32 rgvmodectl, rstdbyctl;
+	u16 crstandvid;
+
+	rgvmodectl = intel_uncore_read(uncore, MEMMODECTL);
+	rstdbyctl = intel_uncore_read(uncore, RSTDBYCTL);
+	crstandvid = intel_uncore_read16(uncore, CRSTANDVID);
+
+	seq_printf(m, "HD boost: %s\n",
+		   str_yes_no(rgvmodectl & MEMMODE_BOOST_EN));
+	seq_printf(m, "Boost freq: %d\n",
+		   (rgvmodectl & MEMMODE_BOOST_FREQ_MASK) >>
+		   MEMMODE_BOOST_FREQ_SHIFT);
+	seq_printf(m, "HW control enabled: %s\n",
+		   str_yes_no(rgvmodectl & MEMMODE_HWIDLE_EN));
+	seq_printf(m, "SW control enabled: %s\n",
+		   str_yes_no(rgvmodectl & MEMMODE_SWMODE_EN));
+	seq_printf(m, "Gated voltage change: %s\n",
+		   str_yes_no(rgvmodectl & MEMMODE_RCLK_GATE));
+	seq_printf(m, "Starting frequency: P%d\n",
+		   (rgvmodectl & MEMMODE_FSTART_MASK) >> MEMMODE_FSTART_SHIFT);
+	seq_printf(m, "Max P-state: P%d\n",
+		   (rgvmodectl & MEMMODE_FMAX_MASK) >> MEMMODE_FMAX_SHIFT);
+	seq_printf(m, "Min P-state: P%d\n", (rgvmodectl & MEMMODE_FMIN_MASK));
+	seq_printf(m, "RS1 VID: %d\n", (crstandvid & 0x3f));
+	seq_printf(m, "RS2 VID: %d\n", ((crstandvid >> 8) & 0x3f));
+	seq_printf(m, "Render standby enabled: %s\n",
+		   str_yes_no(!(rstdbyctl & RCX_SW_EXIT)));
+	seq_puts(m, "Current RS state: ");
+	switch (rstdbyctl & RSX_STATUS_MASK) {
+	case RSX_STATUS_ON:
+		seq_puts(m, "on\n");
+		break;
+	case RSX_STATUS_RC1:
+		seq_puts(m, "RC1\n");
+		break;
+	case RSX_STATUS_RC1E:
+		seq_puts(m, "RC1E\n");
+		break;
+	case RSX_STATUS_RS1:
+		seq_puts(m, "RS1\n");
+		break;
+	case RSX_STATUS_RS2:
+		seq_puts(m, "RS2 (RC6)\n");
+		break;
+	case RSX_STATUS_RS3:
+		seq_puts(m, "RC3 (RC6+)\n");
+		break;
+	default:
+		seq_puts(m, "unknown\n");
+		break;
+	}
+
+	return 0;
+}
+
+static int drpc_show(struct seq_file *m, void *unused)
+{
+	struct intel_gt *gt = m->private;
+	struct drm_i915_private *i915 = gt->i915;
+	intel_wakeref_t wakeref;
+	int err = -ENODEV;
+
+	with_intel_runtime_pm(gt->uncore->rpm, wakeref) {
+		if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+			err = vlv_drpc(m);
+		else if (GRAPHICS_VER(i915) >= 6)
+			err = gen6_drpc(m);
+		else
+			err = ilk_drpc(m);
+	}
+
+	return err;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(drpc);
+
+void intel_gt_pm_frequency_dump(struct intel_gt *gt, struct drm_printer *p)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+	struct intel_rps *rps = &gt->rps;
+	intel_wakeref_t wakeref;
+
+	wakeref = intel_runtime_pm_get(uncore->rpm);
+
+	if (GRAPHICS_VER(i915) == 5) {
+		u16 rgvswctl = intel_uncore_read16(uncore, MEMSWCTL);
+		u16 rgvstat = intel_uncore_read16(uncore, MEMSTAT_ILK);
+
+		drm_printf(p, "Requested P-state: %d\n", (rgvswctl >> 8) & 0xf);
+		drm_printf(p, "Requested VID: %d\n", rgvswctl & 0x3f);
+		drm_printf(p, "Current VID: %d\n", (rgvstat & MEMSTAT_VID_MASK) >>
+			   MEMSTAT_VID_SHIFT);
+		drm_printf(p, "Current P-state: %d\n",
+			   (rgvstat & MEMSTAT_PSTATE_MASK) >> MEMSTAT_PSTATE_SHIFT);
+	} else if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) {
+		u32 rpmodectl, freq_sts;
+
+		rpmodectl = intel_uncore_read(uncore, GEN6_RP_CONTROL);
+		drm_printf(p, "Video Turbo Mode: %s\n",
+			   str_yes_no(rpmodectl & GEN6_RP_MEDIA_TURBO));
+		drm_printf(p, "HW control enabled: %s\n",
+			   str_yes_no(rpmodectl & GEN6_RP_ENABLE));
+		drm_printf(p, "SW control enabled: %s\n",
+			   str_yes_no((rpmodectl & GEN6_RP_MEDIA_MODE_MASK) == GEN6_RP_MEDIA_SW_MODE));
+
+		vlv_punit_get(i915);
+		freq_sts = vlv_punit_read(i915, PUNIT_REG_GPU_FREQ_STS);
+		vlv_punit_put(i915);
+
+		drm_printf(p, "PUNIT_REG_GPU_FREQ_STS: 0x%08x\n", freq_sts);
+		drm_printf(p, "DDR freq: %d MHz\n", i915->mem_freq);
+
+		drm_printf(p, "actual GPU freq: %d MHz\n",
+			   intel_gpu_freq(rps, (freq_sts >> 8) & 0xff));
+
+		drm_printf(p, "current GPU freq: %d MHz\n",
+			   intel_gpu_freq(rps, rps->cur_freq));
+
+		drm_printf(p, "max GPU freq: %d MHz\n",
+			   intel_gpu_freq(rps, rps->max_freq));
+
+		drm_printf(p, "min GPU freq: %d MHz\n",
+			   intel_gpu_freq(rps, rps->min_freq));
+
+		drm_printf(p, "idle GPU freq: %d MHz\n",
+			   intel_gpu_freq(rps, rps->idle_freq));
+
+		drm_printf(p, "efficient (RPe) frequency: %d MHz\n",
+			   intel_gpu_freq(rps, rps->efficient_freq));
+	} else if (GRAPHICS_VER(i915) >= 6) {
+		u32 rp_state_limits;
+		u32 gt_perf_status;
+		struct intel_rps_freq_caps caps;
+		u32 rpmodectl, rpinclimit, rpdeclimit;
+		u32 rpstat, cagf, reqf;
+		u32 rpcurupei, rpcurup, rpprevup;
+		u32 rpcurdownei, rpcurdown, rpprevdown;
+		u32 rpupei, rpupt, rpdownei, rpdownt;
+		u32 pm_ier, pm_imr, pm_isr, pm_iir, pm_mask;
+
+		rp_state_limits = intel_uncore_read(uncore, GEN6_RP_STATE_LIMITS);
+		gen6_rps_get_freq_caps(rps, &caps);
+		if (IS_GEN9_LP(i915))
+			gt_perf_status = intel_uncore_read(uncore, BXT_GT_PERF_STATUS);
+		else
+			gt_perf_status = intel_uncore_read(uncore, GEN6_GT_PERF_STATUS);
+
+		/* RPSTAT1 is in the GT power well */
+		intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+
+		reqf = intel_uncore_read(uncore, GEN6_RPNSWREQ);
+		if (GRAPHICS_VER(i915) >= 9) {
+			reqf >>= 23;
+		} else {
+			reqf &= ~GEN6_TURBO_DISABLE;
+			if (IS_HASWELL(i915) || IS_BROADWELL(i915))
+				reqf >>= 24;
+			else
+				reqf >>= 25;
+		}
+		reqf = intel_gpu_freq(rps, reqf);
+
+		rpmodectl = intel_uncore_read(uncore, GEN6_RP_CONTROL);
+		rpinclimit = intel_uncore_read(uncore, GEN6_RP_UP_THRESHOLD);
+		rpdeclimit = intel_uncore_read(uncore, GEN6_RP_DOWN_THRESHOLD);
+
+		rpstat = intel_uncore_read(uncore, GEN6_RPSTAT1);
+		rpcurupei = intel_uncore_read(uncore, GEN6_RP_CUR_UP_EI) & GEN6_CURICONT_MASK;
+		rpcurup = intel_uncore_read(uncore, GEN6_RP_CUR_UP) & GEN6_CURBSYTAVG_MASK;
+		rpprevup = intel_uncore_read(uncore, GEN6_RP_PREV_UP) & GEN6_CURBSYTAVG_MASK;
+		rpcurdownei = intel_uncore_read(uncore, GEN6_RP_CUR_DOWN_EI) & GEN6_CURIAVG_MASK;
+		rpcurdown = intel_uncore_read(uncore, GEN6_RP_CUR_DOWN) & GEN6_CURBSYTAVG_MASK;
+		rpprevdown = intel_uncore_read(uncore, GEN6_RP_PREV_DOWN) & GEN6_CURBSYTAVG_MASK;
+
+		rpupei = intel_uncore_read(uncore, GEN6_RP_UP_EI);
+		rpupt = intel_uncore_read(uncore, GEN6_RP_UP_THRESHOLD);
+
+		rpdownei = intel_uncore_read(uncore, GEN6_RP_DOWN_EI);
+		rpdownt = intel_uncore_read(uncore, GEN6_RP_DOWN_THRESHOLD);
+
+		cagf = intel_rps_read_actual_frequency(rps);
+
+		intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
+
+		if (GRAPHICS_VER(i915) >= 11) {
+			pm_ier = intel_uncore_read(uncore, GEN11_GPM_WGBOXPERF_INTR_ENABLE);
+			pm_imr = intel_uncore_read(uncore, GEN11_GPM_WGBOXPERF_INTR_MASK);
+			/*
+			 * The equivalent to the PM ISR & IIR cannot be read
+			 * without affecting the current state of the system
+			 */
+			pm_isr = 0;
+			pm_iir = 0;
+		} else if (GRAPHICS_VER(i915) >= 8) {
+			pm_ier = intel_uncore_read(uncore, GEN8_GT_IER(2));
+			pm_imr = intel_uncore_read(uncore, GEN8_GT_IMR(2));
+			pm_isr = intel_uncore_read(uncore, GEN8_GT_ISR(2));
+			pm_iir = intel_uncore_read(uncore, GEN8_GT_IIR(2));
+		} else {
+			pm_ier = intel_uncore_read(uncore, GEN6_PMIER);
+			pm_imr = intel_uncore_read(uncore, GEN6_PMIMR);
+			pm_isr = intel_uncore_read(uncore, GEN6_PMISR);
+			pm_iir = intel_uncore_read(uncore, GEN6_PMIIR);
+		}
+		pm_mask = intel_uncore_read(uncore, GEN6_PMINTRMSK);
+
+		drm_printf(p, "Video Turbo Mode: %s\n",
+			   str_yes_no(rpmodectl & GEN6_RP_MEDIA_TURBO));
+		drm_printf(p, "HW control enabled: %s\n",
+			   str_yes_no(rpmodectl & GEN6_RP_ENABLE));
+		drm_printf(p, "SW control enabled: %s\n",
+			   str_yes_no((rpmodectl & GEN6_RP_MEDIA_MODE_MASK) == GEN6_RP_MEDIA_SW_MODE));
+
+		drm_printf(p, "PM IER=0x%08x IMR=0x%08x, MASK=0x%08x\n",
+			   pm_ier, pm_imr, pm_mask);
+		if (GRAPHICS_VER(i915) <= 10)
+			drm_printf(p, "PM ISR=0x%08x IIR=0x%08x\n",
+				   pm_isr, pm_iir);
+		drm_printf(p, "pm_intrmsk_mbz: 0x%08x\n",
+			   rps->pm_intrmsk_mbz);
+		drm_printf(p, "GT_PERF_STATUS: 0x%08x\n", gt_perf_status);
+		drm_printf(p, "Render p-state ratio: %d\n",
+			   (gt_perf_status & (GRAPHICS_VER(i915) >= 9 ? 0x1ff00 : 0xff00)) >> 8);
+		drm_printf(p, "Render p-state VID: %d\n",
+			   gt_perf_status & 0xff);
+		drm_printf(p, "Render p-state limit: %d\n",
+			   rp_state_limits & 0xff);
+		drm_printf(p, "RPSTAT1: 0x%08x\n", rpstat);
+		drm_printf(p, "RPMODECTL: 0x%08x\n", rpmodectl);
+		drm_printf(p, "RPINCLIMIT: 0x%08x\n", rpinclimit);
+		drm_printf(p, "RPDECLIMIT: 0x%08x\n", rpdeclimit);
+		drm_printf(p, "RPNSWREQ: %dMHz\n", reqf);
+		drm_printf(p, "CAGF: %dMHz\n", cagf);
+		drm_printf(p, "RP CUR UP EI: %d (%lldns)\n",
+			   rpcurupei,
+			   intel_gt_pm_interval_to_ns(gt, rpcurupei));
+		drm_printf(p, "RP CUR UP: %d (%lldns)\n",
+			   rpcurup, intel_gt_pm_interval_to_ns(gt, rpcurup));
+		drm_printf(p, "RP PREV UP: %d (%lldns)\n",
+			   rpprevup, intel_gt_pm_interval_to_ns(gt, rpprevup));
+		drm_printf(p, "Up threshold: %d%%\n",
+			   rps->power.up_threshold);
+		drm_printf(p, "RP UP EI: %d (%lldns)\n",
+			   rpupei, intel_gt_pm_interval_to_ns(gt, rpupei));
+		drm_printf(p, "RP UP THRESHOLD: %d (%lldns)\n",
+			   rpupt, intel_gt_pm_interval_to_ns(gt, rpupt));
+
+		drm_printf(p, "RP CUR DOWN EI: %d (%lldns)\n",
+			   rpcurdownei,
+			   intel_gt_pm_interval_to_ns(gt, rpcurdownei));
+		drm_printf(p, "RP CUR DOWN: %d (%lldns)\n",
+			   rpcurdown,
+			   intel_gt_pm_interval_to_ns(gt, rpcurdown));
+		drm_printf(p, "RP PREV DOWN: %d (%lldns)\n",
+			   rpprevdown,
+			   intel_gt_pm_interval_to_ns(gt, rpprevdown));
+		drm_printf(p, "Down threshold: %d%%\n",
+			   rps->power.down_threshold);
+		drm_printf(p, "RP DOWN EI: %d (%lldns)\n",
+			   rpdownei, intel_gt_pm_interval_to_ns(gt, rpdownei));
+		drm_printf(p, "RP DOWN THRESHOLD: %d (%lldns)\n",
+			   rpdownt, intel_gt_pm_interval_to_ns(gt, rpdownt));
+
+		drm_printf(p, "Lowest (RPN) frequency: %dMHz\n",
+			   intel_gpu_freq(rps, caps.min_freq));
+		drm_printf(p, "Nominal (RP1) frequency: %dMHz\n",
+			   intel_gpu_freq(rps, caps.rp1_freq));
+		drm_printf(p, "Max non-overclocked (RP0) frequency: %dMHz\n",
+			   intel_gpu_freq(rps, caps.rp0_freq));
+		drm_printf(p, "Max overclocked frequency: %dMHz\n",
+			   intel_gpu_freq(rps, rps->max_freq));
+
+		drm_printf(p, "Current freq: %d MHz\n",
+			   intel_gpu_freq(rps, rps->cur_freq));
+		drm_printf(p, "Actual freq: %d MHz\n", cagf);
+		drm_printf(p, "Idle freq: %d MHz\n",
+			   intel_gpu_freq(rps, rps->idle_freq));
+		drm_printf(p, "Min freq: %d MHz\n",
+			   intel_gpu_freq(rps, rps->min_freq));
+		drm_printf(p, "Boost freq: %d MHz\n",
+			   intel_gpu_freq(rps, rps->boost_freq));
+		drm_printf(p, "Max freq: %d MHz\n",
+			   intel_gpu_freq(rps, rps->max_freq));
+		drm_printf(p,
+			   "efficient (RPe) frequency: %d MHz\n",
+			   intel_gpu_freq(rps, rps->efficient_freq));
+	} else {
+		drm_puts(p, "no P-state info available\n");
+	}
+
+	drm_printf(p, "Current CD clock frequency: %d kHz\n", i915->display.cdclk.hw.cdclk);
+	drm_printf(p, "Max CD clock frequency: %d kHz\n", i915->display.cdclk.max_cdclk_freq);
+	drm_printf(p, "Max pixel clock frequency: %d kHz\n", i915->max_dotclk_freq);
+
+	intel_runtime_pm_put(uncore->rpm, wakeref);
+}
+
+static int frequency_show(struct seq_file *m, void *unused)
+{
+	struct intel_gt *gt = m->private;
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	intel_gt_pm_frequency_dump(gt, &p);
+
+	return 0;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(frequency);
+
+static int llc_show(struct seq_file *m, void *data)
+{
+	struct intel_gt *gt = m->private;
+	struct drm_i915_private *i915 = gt->i915;
+	const bool edram = GRAPHICS_VER(i915) > 8;
+	struct intel_rps *rps = &gt->rps;
+	unsigned int max_gpu_freq, min_gpu_freq;
+	intel_wakeref_t wakeref;
+	int gpu_freq, ia_freq;
+
+	seq_printf(m, "LLC: %s\n", str_yes_no(HAS_LLC(i915)));
+	seq_printf(m, "%s: %uMB\n", edram ? "eDRAM" : "eLLC",
+		   i915->edram_size_mb);
+
+	min_gpu_freq = rps->min_freq;
+	max_gpu_freq = rps->max_freq;
+	if (IS_GEN9_BC(i915) || GRAPHICS_VER(i915) >= 11) {
+		/* Convert GT frequency to 50 HZ units */
+		min_gpu_freq /= GEN9_FREQ_SCALER;
+		max_gpu_freq /= GEN9_FREQ_SCALER;
+	}
+
+	seq_puts(m, "GPU freq (MHz)\tEffective CPU freq (MHz)\tEffective Ring freq (MHz)\n");
+
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+	for (gpu_freq = min_gpu_freq; gpu_freq <= max_gpu_freq; gpu_freq++) {
+		ia_freq = gpu_freq;
+		snb_pcode_read(gt->uncore, GEN6_PCODE_READ_MIN_FREQ_TABLE,
+			       &ia_freq, NULL);
+		seq_printf(m, "%d\t\t%d\t\t\t\t%d\n",
+			   intel_gpu_freq(rps,
+					  (gpu_freq *
+					   (IS_GEN9_BC(i915) ||
+					    GRAPHICS_VER(i915) >= 11 ?
+					    GEN9_FREQ_SCALER : 1))),
+			   ((ia_freq >> 0) & 0xff) * 100,
+			   ((ia_freq >> 8) & 0xff) * 100);
+	}
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+
+	return 0;
+}
+
+static bool llc_eval(void *data)
+{
+	struct intel_gt *gt = data;
+
+	return HAS_LLC(gt->i915);
+}
+
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(llc);
+
+static const char *rps_power_to_str(unsigned int power)
+{
+	static const char * const strings[] = {
+		[LOW_POWER] = "low power",
+		[BETWEEN] = "mixed",
+		[HIGH_POWER] = "high power",
+	};
+
+	if (power >= ARRAY_SIZE(strings) || !strings[power])
+		return "unknown";
+
+	return strings[power];
+}
+
+static int rps_boost_show(struct seq_file *m, void *data)
+{
+	struct intel_gt *gt = m->private;
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_rps *rps = &gt->rps;
+
+	seq_printf(m, "RPS enabled? %s\n",
+		   str_yes_no(intel_rps_is_enabled(rps)));
+	seq_printf(m, "RPS active? %s\n",
+		   str_yes_no(intel_rps_is_active(rps)));
+	seq_printf(m, "GPU busy? %s, %llums\n",
+		   str_yes_no(gt->awake),
+		   ktime_to_ms(intel_gt_get_awake_time(gt)));
+	seq_printf(m, "Boosts outstanding? %d\n",
+		   atomic_read(&rps->num_waiters));
+	seq_printf(m, "Interactive? %d\n", READ_ONCE(rps->power.interactive));
+	seq_printf(m, "Frequency requested %d, actual %d\n",
+		   intel_gpu_freq(rps, rps->cur_freq),
+		   intel_rps_read_actual_frequency(rps));
+	seq_printf(m, "  min hard:%d, soft:%d; max soft:%d, hard:%d\n",
+		   intel_gpu_freq(rps, rps->min_freq),
+		   intel_gpu_freq(rps, rps->min_freq_softlimit),
+		   intel_gpu_freq(rps, rps->max_freq_softlimit),
+		   intel_gpu_freq(rps, rps->max_freq));
+	seq_printf(m, "  idle:%d, efficient:%d, boost:%d\n",
+		   intel_gpu_freq(rps, rps->idle_freq),
+		   intel_gpu_freq(rps, rps->efficient_freq),
+		   intel_gpu_freq(rps, rps->boost_freq));
+
+	seq_printf(m, "Wait boosts: %d\n", READ_ONCE(rps->boosts));
+
+	if (GRAPHICS_VER(i915) >= 6 && intel_rps_is_active(rps)) {
+		struct intel_uncore *uncore = gt->uncore;
+		u32 rpup, rpupei;
+		u32 rpdown, rpdownei;
+
+		intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+		rpup = intel_uncore_read_fw(uncore, GEN6_RP_CUR_UP) & GEN6_RP_EI_MASK;
+		rpupei = intel_uncore_read_fw(uncore, GEN6_RP_CUR_UP_EI) & GEN6_RP_EI_MASK;
+		rpdown = intel_uncore_read_fw(uncore, GEN6_RP_CUR_DOWN) & GEN6_RP_EI_MASK;
+		rpdownei = intel_uncore_read_fw(uncore, GEN6_RP_CUR_DOWN_EI) & GEN6_RP_EI_MASK;
+		intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
+
+		seq_printf(m, "\nRPS Autotuning (current \"%s\" window):\n",
+			   rps_power_to_str(rps->power.mode));
+		seq_printf(m, "  Avg. up: %d%% [above threshold? %d%%]\n",
+			   rpup && rpupei ? 100 * rpup / rpupei : 0,
+			   rps->power.up_threshold);
+		seq_printf(m, "  Avg. down: %d%% [below threshold? %d%%]\n",
+			   rpdown && rpdownei ? 100 * rpdown / rpdownei : 0,
+			   rps->power.down_threshold);
+	} else {
+		seq_puts(m, "\nRPS Autotuning inactive\n");
+	}
+
+	return 0;
+}
+
+static bool rps_eval(void *data)
+{
+	struct intel_gt *gt = data;
+
+	return HAS_RPS(gt->i915);
+}
+
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(rps_boost);
+
+void intel_gt_pm_debugfs_register(struct intel_gt *gt, struct dentry *root)
+{
+	static const struct intel_gt_debugfs_file files[] = {
+		{ "drpc", &drpc_fops, NULL },
+		{ "frequency", &frequency_fops, NULL },
+		{ "forcewake", &fw_domains_fops, NULL },
+		{ "forcewake_user", &forcewake_user_fops, NULL},
+		{ "llc", &llc_fops, llc_eval },
+		{ "rps_boost", &rps_boost_fops, rps_eval },
+	};
+
+	intel_gt_debugfs_register_files(root, files, ARRAY_SIZE(files), gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_pm_debugfs.h b/drivers/gpu/drm/i915/gt/intel_gt_pm_debugfs.h
new file mode 100644
index 000000000..0ace8c2da
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_pm_debugfs.h
@@ -0,0 +1,20 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_GT_PM_DEBUGFS_H
+#define INTEL_GT_PM_DEBUGFS_H
+
+struct intel_gt;
+struct dentry;
+struct drm_printer;
+
+void intel_gt_pm_debugfs_register(struct intel_gt *gt, struct dentry *root);
+void intel_gt_pm_frequency_dump(struct intel_gt *gt, struct drm_printer *m);
+
+/* functions that need to be accessed by the upper level non-gt interfaces */
+void intel_gt_pm_debugfs_forcewake_user_open(struct intel_gt *gt);
+void intel_gt_pm_debugfs_forcewake_user_release(struct intel_gt *gt);
+
+#endif /* INTEL_GT_PM_DEBUGFS_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_pm_irq.c b/drivers/gpu/drm/i915/gt/intel_gt_pm_irq.c
new file mode 100644
index 000000000..52f2a28b2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_pm_irq.c
@@ -0,0 +1,109 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_gt.h"
+#include "intel_gt_irq.h"
+#include "intel_gt_pm_irq.h"
+#include "intel_gt_regs.h"
+
+static void write_pm_imr(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+	u32 mask = gt->pm_imr;
+	i915_reg_t reg;
+
+	if (GRAPHICS_VER(i915) >= 11) {
+		reg = GEN11_GPM_WGBOXPERF_INTR_MASK;
+		mask <<= 16; /* pm is in upper half */
+	} else if (GRAPHICS_VER(i915) >= 8) {
+		reg = GEN8_GT_IMR(2);
+	} else {
+		reg = GEN6_PMIMR;
+	}
+
+	intel_uncore_write(uncore, reg, mask);
+}
+
+static void gen6_gt_pm_update_irq(struct intel_gt *gt,
+				  u32 interrupt_mask,
+				  u32 enabled_irq_mask)
+{
+	u32 new_val;
+
+	WARN_ON(enabled_irq_mask & ~interrupt_mask);
+
+	lockdep_assert_held(gt->irq_lock);
+
+	new_val = gt->pm_imr;
+	new_val &= ~interrupt_mask;
+	new_val |= ~enabled_irq_mask & interrupt_mask;
+
+	if (new_val != gt->pm_imr) {
+		gt->pm_imr = new_val;
+		write_pm_imr(gt);
+	}
+}
+
+void gen6_gt_pm_unmask_irq(struct intel_gt *gt, u32 mask)
+{
+	gen6_gt_pm_update_irq(gt, mask, mask);
+}
+
+void gen6_gt_pm_mask_irq(struct intel_gt *gt, u32 mask)
+{
+	gen6_gt_pm_update_irq(gt, mask, 0);
+}
+
+void gen6_gt_pm_reset_iir(struct intel_gt *gt, u32 reset_mask)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	i915_reg_t reg = GRAPHICS_VER(gt->i915) >= 8 ? GEN8_GT_IIR(2) : GEN6_PMIIR;
+
+	lockdep_assert_held(gt->irq_lock);
+
+	intel_uncore_write(uncore, reg, reset_mask);
+	intel_uncore_write(uncore, reg, reset_mask);
+	intel_uncore_posting_read(uncore, reg);
+}
+
+static void write_pm_ier(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+	u32 mask = gt->pm_ier;
+	i915_reg_t reg;
+
+	if (GRAPHICS_VER(i915) >= 11) {
+		reg = GEN11_GPM_WGBOXPERF_INTR_ENABLE;
+		mask <<= 16; /* pm is in upper half */
+	} else if (GRAPHICS_VER(i915) >= 8) {
+		reg = GEN8_GT_IER(2);
+	} else {
+		reg = GEN6_PMIER;
+	}
+
+	intel_uncore_write(uncore, reg, mask);
+}
+
+void gen6_gt_pm_enable_irq(struct intel_gt *gt, u32 enable_mask)
+{
+	lockdep_assert_held(gt->irq_lock);
+
+	gt->pm_ier |= enable_mask;
+	write_pm_ier(gt);
+	gen6_gt_pm_unmask_irq(gt, enable_mask);
+}
+
+void gen6_gt_pm_disable_irq(struct intel_gt *gt, u32 disable_mask)
+{
+	lockdep_assert_held(gt->irq_lock);
+
+	gt->pm_ier &= ~disable_mask;
+	gen6_gt_pm_mask_irq(gt, disable_mask);
+	write_pm_ier(gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_pm_irq.h b/drivers/gpu/drm/i915/gt/intel_gt_pm_irq.h
new file mode 100644
index 000000000..ff766966d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_pm_irq.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_GT_PM_IRQ_H
+#define INTEL_GT_PM_IRQ_H
+
+#include <linux/types.h>
+
+struct intel_gt;
+
+void gen6_gt_pm_unmask_irq(struct intel_gt *gt, u32 mask);
+void gen6_gt_pm_mask_irq(struct intel_gt *gt, u32 mask);
+
+void gen6_gt_pm_enable_irq(struct intel_gt *gt, u32 enable_mask);
+void gen6_gt_pm_disable_irq(struct intel_gt *gt, u32 disable_mask);
+
+void gen6_gt_pm_reset_iir(struct intel_gt *gt, u32 reset_mask);
+
+#endif /* INTEL_GT_PM_IRQ_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_regs.h b/drivers/gpu/drm/i915/gt/intel_gt_regs.h
new file mode 100644
index 000000000..dd006563c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_regs.h
@@ -0,0 +1,1591 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_GT_REGS__
+#define __INTEL_GT_REGS__
+
+#include "i915_reg_defs.h"
+
+/* RPM unit config (Gen8+) */
+#define RPM_CONFIG0				_MMIO(0xd00)
+#define   GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT	3
+#define   GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK	(1 << GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT)
+#define   GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ	0
+#define   GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ	1
+#define   GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT	3
+#define   GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK	(0x7 << GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT)
+#define   GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ	0
+#define   GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ	1
+#define   GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_38_4_MHZ	2
+#define   GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_25_MHZ	3
+#define   GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_SHIFT	1
+#define   GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_MASK	(0x3 << GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_SHIFT)
+
+#define RPM_CONFIG1				_MMIO(0xd04)
+#define   GEN10_GT_NOA_ENABLE			(1 << 9)
+
+/* RCP unit config (Gen8+) */
+#define RCP_CONFIG				_MMIO(0xd08)
+
+#define RC6_LOCATION				_MMIO(0xd40)
+#define   RC6_CTX_IN_DRAM			(1 << 0)
+#define RC6_CTX_BASE				_MMIO(0xd48)
+#define   RC6_CTX_BASE_MASK			0xFFFFFFF0
+
+#define FORCEWAKE_ACK_MEDIA_VDBOX_GEN11(n)	_MMIO(0xd50 + (n) * 4)
+#define FORCEWAKE_ACK_MEDIA_VEBOX_GEN11(n)	_MMIO(0xd70 + (n) * 4)
+#define FORCEWAKE_ACK_RENDER_GEN9		_MMIO(0xd84)
+#define FORCEWAKE_ACK_MEDIA_GEN9		_MMIO(0xd88)
+
+#define MCFG_MCR_SELECTOR			_MMIO(0xfd0)
+#define SF_MCR_SELECTOR				_MMIO(0xfd8)
+#define GEN8_MCR_SELECTOR			_MMIO(0xfdc)
+#define   GEN8_MCR_SLICE(slice)			(((slice) & 3) << 26)
+#define   GEN8_MCR_SLICE_MASK			GEN8_MCR_SLICE(3)
+#define   GEN8_MCR_SUBSLICE(subslice)		(((subslice) & 3) << 24)
+#define   GEN8_MCR_SUBSLICE_MASK		GEN8_MCR_SUBSLICE(3)
+#define   GEN11_MCR_MULTICAST			REG_BIT(31)
+#define   GEN11_MCR_SLICE(slice)		(((slice) & 0xf) << 27)
+#define   GEN11_MCR_SLICE_MASK			GEN11_MCR_SLICE(0xf)
+#define   GEN11_MCR_SUBSLICE(subslice)		(((subslice) & 0x7) << 24)
+#define   GEN11_MCR_SUBSLICE_MASK		GEN11_MCR_SUBSLICE(0x7)
+
+#define IPEIR_I965				_MMIO(0x2064)
+#define IPEHR_I965				_MMIO(0x2068)
+
+/*
+ * On GEN4, only the render ring INSTDONE exists and has a different
+ * layout than the GEN7+ version.
+ * The GEN2 counterpart of this register is GEN2_INSTDONE.
+ */
+#define INSTPS					_MMIO(0x2070) /* 965+ only */
+#define GEN4_INSTDONE1				_MMIO(0x207c) /* 965+ only, aka INSTDONE_2 on SNB */
+#define ACTHD_I965				_MMIO(0x2074)
+#define HWS_PGA					_MMIO(0x2080)
+#define   HWS_ADDRESS_MASK			0xfffff000
+#define   HWS_START_ADDRESS_SHIFT		4
+
+#define _3D_CHICKEN				_MMIO(0x2084)
+#define   _3D_CHICKEN_HIZ_PLANE_DISABLE_MSAA_4X_SNB	(1 << 10)
+
+#define PWRCTXA					_MMIO(0x2088) /* 965GM+ only */
+#define   PWRCTX_EN				(1 << 0)
+
+#define FF_SLICE_CHICKEN			_MMIO(0x2088)
+#define   FF_SLICE_CHICKEN_CL_PROVOKING_VERTEX_FIX	(1 << 1)
+
+/* GM45+ chicken bits -- debug workaround bits that may be required
+ * for various sorts of correct behavior.  The top 16 bits of each are
+ * the enables for writing to the corresponding low bit.
+ */
+#define _3D_CHICKEN2				_MMIO(0x208c)
+/* Disables pipelining of read flushes past the SF-WIZ interface.
+ * Required on all Ironlake steppings according to the B-Spec, but the
+ * particular danger of not doing so is not specified.
+ */
+#define   _3D_CHICKEN2_WM_READ_PIPELINED	(1 << 14)
+
+#define _3D_CHICKEN3				_MMIO(0x2090)
+#define   _3D_CHICKEN_SF_PROVOKING_VERTEX_FIX	(1 << 12)
+#define   _3D_CHICKEN_SF_DISABLE_OBJEND_CULL	(1 << 10)
+#define   _3D_CHICKEN3_AA_LINE_QUALITY_FIX_ENABLE	(1 << 5)
+#define   _3D_CHICKEN3_SF_DISABLE_FASTCLIP_CULL	(1 << 5)
+#define   _3D_CHICKEN_SDE_LIMIT_FIFO_POLY_DEPTH(x)	((x) << 1) /* gen8+ */
+#define   _3D_CHICKEN3_SF_DISABLE_PIPELINED_ATTR_FETCH	(1 << 1) /* gen6 */
+
+#define GEN2_INSTDONE				_MMIO(0x2090)
+#define NOPID					_MMIO(0x2094)
+#define HWSTAM					_MMIO(0x2098)
+
+#define WAIT_FOR_RC6_EXIT			_MMIO(0x20cc)
+/* HSW only */
+#define   HSW_SELECTIVE_READ_ADDRESSING_SHIFT	2
+#define   HSW_SELECTIVE_READ_ADDRESSING_MASK	(0x3 << HSW_SLECTIVE_READ_ADDRESSING_SHIFT)
+#define   HSW_SELECTIVE_WRITE_ADDRESS_SHIFT	4
+#define   HSW_SELECTIVE_WRITE_ADDRESS_MASK	(0x7 << HSW_SELECTIVE_WRITE_ADDRESS_SHIFT)
+/* HSW+ */
+#define   HSW_WAIT_FOR_RC6_EXIT_ENABLE		(1 << 0)
+#define   HSW_RCS_CONTEXT_ENABLE		(1 << 7)
+#define   HSW_RCS_INHIBIT			(1 << 8)
+/* Gen8 */
+#define   GEN8_SELECTIVE_WRITE_ADDRESS_SHIFT	4
+#define   GEN8_SELECTIVE_WRITE_ADDRESS_MASK	(0x3 << GEN8_SELECTIVE_WRITE_ADDRESS_SHIFT)
+#define   GEN8_SELECTIVE_WRITE_ADDRESS_SHIFT	4
+#define   GEN8_SELECTIVE_WRITE_ADDRESS_MASK	(0x3 << GEN8_SELECTIVE_WRITE_ADDRESS_SHIFT)
+#define   GEN8_SELECTIVE_WRITE_ADDRESSING_ENABLE	(1 << 6)
+#define   GEN8_SELECTIVE_READ_SUBSLICE_SELECT_SHIFT	9
+#define   GEN8_SELECTIVE_READ_SUBSLICE_SELECT_MASK	(0x3 << GEN8_SELECTIVE_READ_SUBSLICE_SELECT_SHIFT)
+#define   GEN8_SELECTIVE_READ_SLICE_SELECT_SHIFT	11
+#define   GEN8_SELECTIVE_READ_SLICE_SELECT_MASK	(0x3 << GEN8_SELECTIVE_READ_SLICE_SELECT_SHIFT)
+#define   GEN8_SELECTIVE_READ_ADDRESSING_ENABLE	(1 << 13)
+
+#define GEN6_GT_MODE				_MMIO(0x20d0)
+#define   GEN6_WIZ_HASHING(hi, lo)		(((hi) << 9) | ((lo) << 7))
+#define   GEN6_WIZ_HASHING_8x8			GEN6_WIZ_HASHING(0, 0)
+#define   GEN6_WIZ_HASHING_8x4			GEN6_WIZ_HASHING(0, 1)
+#define   GEN6_WIZ_HASHING_16x4			GEN6_WIZ_HASHING(1, 0)
+#define   GEN6_WIZ_HASHING_MASK			GEN6_WIZ_HASHING(1, 1)
+#define   GEN6_TD_FOUR_ROW_DISPATCH_DISABLE	(1 << 5)
+
+/* chicken reg for WaConextSwitchWithConcurrentTLBInvalidate */
+#define GEN9_CSFE_CHICKEN1_RCS			_MMIO(0x20d4)
+#define   GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE	(1 << 2)
+#define   GEN11_ENABLE_32_PLANE_MODE		(1 << 7)
+
+#define GEN7_FF_SLICE_CS_CHICKEN1		_MMIO(0x20e0)
+#define   GEN9_FFSC_PERCTX_PREEMPT_CTRL		(1 << 14)
+
+#define FF_SLICE_CS_CHICKEN2			_MMIO(0x20e4)
+#define   GEN9_TSG_BARRIER_ACK_DISABLE		(1 << 8)
+#define   GEN9_POOLED_EU_LOAD_BALANCING_FIX_DISABLE	(1 << 10)
+#define   GEN12_PERF_FIX_BALANCING_CFE_DISABLE	REG_BIT(15)
+
+#define GEN9_CS_DEBUG_MODE1			_MMIO(0x20ec)
+#define   FF_DOP_CLOCK_GATE_DISABLE		REG_BIT(1)
+#define GEN12_CS_DEBUG_MODE1_CCCSUNIT_BE_COMMON	_MMIO(0x20ec)
+#define   GEN12_REPLAY_MODE_GRANULARITY		REG_BIT(0)
+
+/* WaClearTdlStateAckDirtyBits */
+#define GEN8_STATE_ACK				_MMIO(0x20f0)
+#define GEN9_STATE_ACK_SLICE1			_MMIO(0x20f8)
+#define GEN9_STATE_ACK_SLICE2			_MMIO(0x2100)
+#define   GEN9_STATE_ACK_TDL0			(1 << 12)
+#define   GEN9_STATE_ACK_TDL1			(1 << 13)
+#define   GEN9_STATE_ACK_TDL2			(1 << 14)
+#define   GEN9_STATE_ACK_TDL3			(1 << 15)
+#define   GEN9_SUBSLICE_TDL_ACK_BITS	\
+	(GEN9_STATE_ACK_TDL3 | GEN9_STATE_ACK_TDL2 | \
+	 GEN9_STATE_ACK_TDL1 | GEN9_STATE_ACK_TDL0)
+
+#define CACHE_MODE_0				_MMIO(0x2120) /* 915+ only */
+#define   CM0_PIPELINED_RENDER_FLUSH_DISABLE	(1 << 8)
+#define   CM0_IZ_OPT_DISABLE			(1 << 6)
+#define   CM0_ZR_OPT_DISABLE			(1 << 5)
+#define	  CM0_STC_EVICT_DISABLE_LRA_SNB		(1 << 5)
+#define   CM0_DEPTH_EVICT_DISABLE		(1 << 4)
+#define   CM0_COLOR_EVICT_DISABLE		(1 << 3)
+#define   CM0_DEPTH_WRITE_DISABLE		(1 << 1)
+#define   CM0_RC_OP_FLUSH_DISABLE		(1 << 0)
+
+#define GFX_FLSH_CNTL				_MMIO(0x2170) /* 915+ only */
+
+/*
+ * Logical Context regs
+ */
+/*
+ * Notes on SNB/IVB/VLV context size:
+ * - Power context is saved elsewhere (LLC or stolen)
+ * - Ring/execlist context is saved on SNB, not on IVB
+ * - Extended context size already includes render context size
+ * - We always need to follow the extended context size.
+ *   SNB BSpec has comments indicating that we should use the
+ *   render context size instead if execlists are disabled, but
+ *   based on empirical testing that's just nonsense.
+ * - Pipelined/VF state is saved on SNB/IVB respectively
+ * - GT1 size just indicates how much of render context
+ *   doesn't need saving on GT1
+ */
+#define CXT_SIZE				_MMIO(0x21a0)
+#define   GEN6_CXT_POWER_SIZE(cxt_reg)		(((cxt_reg) >> 24) & 0x3f)
+#define   GEN6_CXT_RING_SIZE(cxt_reg)		(((cxt_reg) >> 18) & 0x3f)
+#define   GEN6_CXT_RENDER_SIZE(cxt_reg)		(((cxt_reg) >> 12) & 0x3f)
+#define   GEN6_CXT_EXTENDED_SIZE(cxt_reg)	(((cxt_reg) >> 6) & 0x3f)
+#define   GEN6_CXT_PIPELINE_SIZE(cxt_reg)	(((cxt_reg) >> 0) & 0x3f)
+#define   GEN6_CXT_TOTAL_SIZE(cxt_reg)		(GEN6_CXT_RING_SIZE(cxt_reg) + \
+						GEN6_CXT_EXTENDED_SIZE(cxt_reg) + \
+						GEN6_CXT_PIPELINE_SIZE(cxt_reg))
+#define GEN7_CXT_SIZE				_MMIO(0x21a8)
+#define   GEN7_CXT_POWER_SIZE(ctx_reg)		(((ctx_reg) >> 25) & 0x7f)
+#define   GEN7_CXT_RING_SIZE(ctx_reg)		(((ctx_reg) >> 22) & 0x7)
+#define   GEN7_CXT_RENDER_SIZE(ctx_reg)		(((ctx_reg) >> 16) & 0x3f)
+#define   GEN7_CXT_EXTENDED_SIZE(ctx_reg)	(((ctx_reg) >> 9) & 0x7f)
+#define   GEN7_CXT_GT1_SIZE(ctx_reg)		(((ctx_reg) >> 6) & 0x7)
+#define   GEN7_CXT_VFSTATE_SIZE(ctx_reg)	(((ctx_reg) >> 0) & 0x3f)
+#define   GEN7_CXT_TOTAL_SIZE(ctx_reg)		(GEN7_CXT_EXTENDED_SIZE(ctx_reg) + \
+						 GEN7_CXT_VFSTATE_SIZE(ctx_reg))
+
+#define HSW_MI_PREDICATE_RESULT_2		_MMIO(0x2214)
+
+#define GEN9_CTX_PREEMPT_REG			_MMIO(0x2248)
+#define   GEN12_DISABLE_POSH_BUSY_FF_DOP_CG	REG_BIT(11)
+
+#define GPGPU_THREADS_DISPATCHED		_MMIO(0x2290)
+#define GPGPU_THREADS_DISPATCHED_UDW		_MMIO(0x2290 + 4)
+
+#define GEN9_RCS_FE_FSM2			_MMIO(0x22a4)
+#define GEN6_RCS_PWR_FSM			_MMIO(0x22ac)
+
+#define HS_INVOCATION_COUNT			_MMIO(0x2300)
+#define HS_INVOCATION_COUNT_UDW			_MMIO(0x2300 + 4)
+#define DS_INVOCATION_COUNT			_MMIO(0x2308)
+#define DS_INVOCATION_COUNT_UDW			_MMIO(0x2308 + 4)
+#define IA_VERTICES_COUNT			_MMIO(0x2310)
+#define IA_VERTICES_COUNT_UDW			_MMIO(0x2310 + 4)
+#define IA_PRIMITIVES_COUNT			_MMIO(0x2318)
+#define IA_PRIMITIVES_COUNT_UDW			_MMIO(0x2318 + 4)
+#define VS_INVOCATION_COUNT			_MMIO(0x2320)
+#define VS_INVOCATION_COUNT_UDW			_MMIO(0x2320 + 4)
+#define GS_INVOCATION_COUNT			_MMIO(0x2328)
+#define GS_INVOCATION_COUNT_UDW			_MMIO(0x2328 + 4)
+#define GS_PRIMITIVES_COUNT			_MMIO(0x2330)
+#define GS_PRIMITIVES_COUNT_UDW			_MMIO(0x2330 + 4)
+#define CL_INVOCATION_COUNT			_MMIO(0x2338)
+#define CL_INVOCATION_COUNT_UDW			_MMIO(0x2338 + 4)
+#define CL_PRIMITIVES_COUNT			_MMIO(0x2340)
+#define CL_PRIMITIVES_COUNT_UDW			_MMIO(0x2340 + 4)
+#define PS_INVOCATION_COUNT			_MMIO(0x2348)
+#define PS_INVOCATION_COUNT_UDW			_MMIO(0x2348 + 4)
+#define PS_DEPTH_COUNT				_MMIO(0x2350)
+#define PS_DEPTH_COUNT_UDW			_MMIO(0x2350 + 4)
+#define GEN7_3DPRIM_END_OFFSET			_MMIO(0x2420)
+#define GEN7_3DPRIM_START_VERTEX		_MMIO(0x2430)
+#define GEN7_3DPRIM_VERTEX_COUNT		_MMIO(0x2434)
+#define GEN7_3DPRIM_INSTANCE_COUNT		_MMIO(0x2438)
+#define GEN7_3DPRIM_START_INSTANCE		_MMIO(0x243c)
+#define GEN7_3DPRIM_BASE_VERTEX			_MMIO(0x2440)
+#define GEN7_GPGPU_DISPATCHDIMX			_MMIO(0x2500)
+#define GEN7_GPGPU_DISPATCHDIMY			_MMIO(0x2504)
+#define GEN7_GPGPU_DISPATCHDIMZ			_MMIO(0x2508)
+
+#define GFX_MODE				_MMIO(0x2520)
+
+#define GEN8_CS_CHICKEN1			_MMIO(0x2580)
+#define   GEN9_PREEMPT_3D_OBJECT_LEVEL		(1 << 0)
+#define   GEN9_PREEMPT_GPGPU_LEVEL(hi, lo)	(((hi) << 2) | ((lo) << 1))
+#define   GEN9_PREEMPT_GPGPU_MID_THREAD_LEVEL	GEN9_PREEMPT_GPGPU_LEVEL(0, 0)
+#define   GEN9_PREEMPT_GPGPU_THREAD_GROUP_LEVEL	GEN9_PREEMPT_GPGPU_LEVEL(0, 1)
+#define   GEN9_PREEMPT_GPGPU_COMMAND_LEVEL	GEN9_PREEMPT_GPGPU_LEVEL(1, 0)
+#define   GEN9_PREEMPT_GPGPU_LEVEL_MASK		GEN9_PREEMPT_GPGPU_LEVEL(1, 1)
+
+#define DRAW_WATERMARK				_MMIO(0x26c0)
+#define   VERT_WM_VAL				REG_GENMASK(9, 0)
+
+#define GEN12_GLOBAL_MOCS(i)			_MMIO(0x4000 + (i) * 4) /* Global MOCS regs */
+
+#define RENDER_HWS_PGA_GEN7			_MMIO(0x4080)
+
+#define GEN8_GAMW_ECO_DEV_RW_IA			_MMIO(0x4080)
+#define   GAMW_ECO_ENABLE_64K_IPS_FIELD		0xF
+#define   GAMW_ECO_DEV_CTX_RELOAD_DISABLE	(1 << 7)
+
+#define GAM_ECOCHK				_MMIO(0x4090)
+#define   BDW_DISABLE_HDC_INVALIDATION		(1 << 25)
+#define   ECOCHK_SNB_BIT			(1 << 10)
+#define   ECOCHK_DIS_TLB			(1 << 8)
+#define   HSW_ECOCHK_ARB_PRIO_SOL		(1 << 6)
+#define   ECOCHK_PPGTT_CACHE64B			(0x3 << 3)
+#define   ECOCHK_PPGTT_CACHE4B			(0x0 << 3)
+#define   ECOCHK_PPGTT_GFDT_IVB			(0x1 << 4)
+#define   ECOCHK_PPGTT_LLC_IVB			(0x1 << 3)
+#define   ECOCHK_PPGTT_UC_HSW			(0x1 << 3)
+#define   ECOCHK_PPGTT_WT_HSW			(0x2 << 3)
+#define   ECOCHK_PPGTT_WB_HSW			(0x3 << 3)
+
+#define GEN8_RING_FAULT_REG			_MMIO(0x4094)
+#define _RING_FAULT_REG_RCS			0x4094
+#define _RING_FAULT_REG_VCS			0x4194
+#define _RING_FAULT_REG_BCS			0x4294
+#define _RING_FAULT_REG_VECS			0x4394
+#define RING_FAULT_REG(engine)			_MMIO(_PICK((engine)->class, \
+							    _RING_FAULT_REG_RCS, \
+							    _RING_FAULT_REG_VCS, \
+							    _RING_FAULT_REG_VECS, \
+							    _RING_FAULT_REG_BCS))
+
+#define ERROR_GEN6				_MMIO(0x40a0)
+
+#define DONE_REG				_MMIO(0x40b0)
+#define GEN8_PRIVATE_PAT_LO			_MMIO(0x40e0)
+#define GEN8_PRIVATE_PAT_HI			_MMIO(0x40e0 + 4)
+#define GEN10_PAT_INDEX(index)			_MMIO(0x40e0 + (index) * 4)
+#define BSD_HWS_PGA_GEN7			_MMIO(0x4180)
+
+#define GEN12_CCS_AUX_INV			_MMIO(0x4208)
+#define GEN12_VD0_AUX_INV			_MMIO(0x4218)
+#define GEN12_VE0_AUX_INV			_MMIO(0x4238)
+#define GEN12_BCS0_AUX_INV			_MMIO(0x4248)
+
+#define GEN8_RTCR				_MMIO(0x4260)
+#define GEN8_M1TCR				_MMIO(0x4264)
+#define GEN8_M2TCR				_MMIO(0x4268)
+#define GEN8_BTCR				_MMIO(0x426c)
+#define GEN8_VTCR				_MMIO(0x4270)
+
+#define BLT_HWS_PGA_GEN7			_MMIO(0x4280)
+
+#define GEN12_VD2_AUX_INV			_MMIO(0x4298)
+#define GEN12_CCS0_AUX_INV			_MMIO(0x42c8)
+#define   AUX_INV				REG_BIT(0)
+
+#define VEBOX_HWS_PGA_GEN7			_MMIO(0x4380)
+
+#define GEN12_AUX_ERR_DBG			_MMIO(0x43f4)
+
+#define GEN7_TLB_RD_ADDR			_MMIO(0x4700)
+
+#define GEN12_PAT_INDEX(index)			_MMIO(0x4800 + (index) * 4)
+
+#define XEHP_TILE0_ADDR_RANGE			_MMIO(0x4900)
+#define   XEHP_TILE_LMEM_RANGE_SHIFT		8
+
+#define XEHP_FLAT_CCS_BASE_ADDR			_MMIO(0x4910)
+#define   XEHP_CCS_BASE_SHIFT			8
+
+#define GAMTARBMODE				_MMIO(0x4a08)
+#define   ARB_MODE_BWGTLB_DISABLE		(1 << 9)
+#define   ARB_MODE_SWIZZLE_BDW			(1 << 1)
+
+#define GEN9_GAMT_ECO_REG_RW_IA			_MMIO(0x4ab0)
+#define   GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS	(1 << 18)
+
+#define GAMT_CHKN_BIT_REG			_MMIO(0x4ab8)
+#define   GAMT_CHKN_DISABLE_L3_COH_PIPE		(1 << 31)
+#define   GAMT_CHKN_DISABLE_DYNAMIC_CREDIT_SHARING	(1 << 28)
+#define   GAMT_CHKN_DISABLE_I2M_CYCLE_ON_WR_PORT	(1 << 24)
+
+#define GEN8_FAULT_TLB_DATA0			_MMIO(0x4b10)
+#define GEN8_FAULT_TLB_DATA1			_MMIO(0x4b14)
+
+#define GEN11_GACB_PERF_CTRL			_MMIO(0x4b80)
+#define   GEN11_HASH_CTRL_MASK			(0x3 << 12 | 0xf << 0)
+#define   GEN11_HASH_CTRL_BIT0			(1 << 0)
+#define   GEN11_HASH_CTRL_BIT4			(1 << 12)
+
+/* gamt regs */
+#define GEN8_L3_LRA_1_GPGPU			_MMIO(0x4dd4)
+#define   GEN8_L3_LRA_1_GPGPU_DEFAULT_VALUE_BDW	0x67F1427F /* max/min for LRA1/2 */
+#define   GEN8_L3_LRA_1_GPGPU_DEFAULT_VALUE_CHV	0x5FF101FF /* max/min for LRA1/2 */
+#define   GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_SKL	0x67F1427F /*    "        " */
+#define   GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_BXT	0x5FF101FF /*    "        " */
+
+#define MMCD_MISC_CTRL				_MMIO(0x4ddc) /* skl+ */
+#define   MMCD_PCLA				(1 << 31)
+#define   MMCD_HOTSPOT_EN			(1 << 27)
+
+/* There are the 4 64-bit counter registers, one for each stream output */
+#define GEN7_SO_NUM_PRIMS_WRITTEN(n)		_MMIO(0x5200 + (n) * 8)
+#define GEN7_SO_NUM_PRIMS_WRITTEN_UDW(n)	_MMIO(0x5200 + (n) * 8 + 4)
+
+#define GEN7_SO_PRIM_STORAGE_NEEDED(n)		_MMIO(0x5240 + (n) * 8)
+#define GEN7_SO_PRIM_STORAGE_NEEDED_UDW(n)	_MMIO(0x5240 + (n) * 8 + 4)
+
+#define GEN9_WM_CHICKEN3			_MMIO(0x5588)
+#define   GEN9_FACTOR_IN_CLR_VAL_HIZ		(1 << 9)
+
+#define CHICKEN_RASTER_1			_MMIO(0x6204)
+#define   DIS_SF_ROUND_NEAREST_EVEN		REG_BIT(8)
+
+#define CHICKEN_RASTER_2			_MMIO(0x6208)
+#define   TBIMR_FAST_CLIP			REG_BIT(5)
+
+#define VFLSKPD					_MMIO(0x62a8)
+#define   DIS_OVER_FETCH_CACHE			REG_BIT(1)
+#define   DIS_MULT_MISS_RD_SQUASH		REG_BIT(0)
+
+#define FF_MODE2				_MMIO(0x6604)
+#define   FF_MODE2_GS_TIMER_MASK		REG_GENMASK(31, 24)
+#define   FF_MODE2_GS_TIMER_224			REG_FIELD_PREP(FF_MODE2_GS_TIMER_MASK, 224)
+#define   FF_MODE2_TDS_TIMER_MASK		REG_GENMASK(23, 16)
+#define   FF_MODE2_TDS_TIMER_128		REG_FIELD_PREP(FF_MODE2_TDS_TIMER_MASK, 4)
+
+#define XEHPG_INSTDONE_GEOM_SVG			_MMIO(0x666c)
+
+#define CACHE_MODE_0_GEN7			_MMIO(0x7000) /* IVB+ */
+#define   RC_OP_FLUSH_ENABLE			(1 << 0)
+#define   HIZ_RAW_STALL_OPT_DISABLE		(1 << 2)
+#define CACHE_MODE_1				_MMIO(0x7004) /* IVB+ */
+#define   PIXEL_SUBSPAN_COLLECT_OPT_DISABLE	(1 << 6)
+#define   GEN8_4x4_STC_OPTIMIZATION_DISABLE	(1 << 6)
+#define   GEN9_PARTIAL_RESOLVE_IN_VC_DISABLE	(1 << 1)
+
+#define GEN7_GT_MODE				_MMIO(0x7008)
+#define   GEN9_IZ_HASHING_MASK(slice)		(0x3 << ((slice) * 2))
+#define   GEN9_IZ_HASHING(slice, val)		((val) << ((slice) * 2))
+
+/* GEN7 chicken */
+#define GEN7_COMMON_SLICE_CHICKEN1		_MMIO(0x7010)
+#define   GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC	(1 << 10)
+#define   GEN9_RHWO_OPTIMIZATION_DISABLE	(1 << 14)
+
+#define COMMON_SLICE_CHICKEN2			_MMIO(0x7014)
+#define   GEN9_PBE_COMPRESSED_HASH_SELECTION	(1 << 13)
+#define   GEN9_DISABLE_GATHER_AT_SET_SHADER_COMMON_SLICE	(1 << 12)
+#define   GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION	(1 << 8)
+#define   GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE	(1 << 0)
+
+#define HIZ_CHICKEN				_MMIO(0x7018)
+#define   CHV_HZ_8X8_MODE_IN_1X			REG_BIT(15)
+#define   DG1_HZ_READ_SUPPRESSION_OPTIMIZATION_DISABLE	REG_BIT(14)
+#define   BDW_HIZ_POWER_COMPILER_CLOCK_GATING_DISABLE	REG_BIT(3)
+
+#define GEN8_L3CNTLREG				_MMIO(0x7034)
+#define   GEN8_ERRDETBCTRL			(1 << 9)
+
+#define GEN7_SC_INSTDONE			_MMIO(0x7100)
+#define GEN12_SC_INSTDONE_EXTRA			_MMIO(0x7104)
+#define GEN12_SC_INSTDONE_EXTRA2		_MMIO(0x7108)
+
+/* GEN8 chicken */
+#define HDC_CHICKEN0				_MMIO(0x7300)
+#define   HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE	(1 << 15)
+#define   HDC_FENCE_DEST_SLM_DISABLE		(1 << 14)
+#define   HDC_DONOT_FETCH_MEM_WHEN_MASKED	(1 << 11)
+#define   HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT	(1 << 5)
+#define   HDC_FORCE_NON_COHERENT		(1 << 4)
+#define   HDC_BARRIER_PERFORMANCE_DISABLE	(1 << 10)
+
+#define GEN8_HDC_CHICKEN1			_MMIO(0x7304)
+
+#define GEN11_COMMON_SLICE_CHICKEN3		_MMIO(0x7304)
+#define   DG1_FLOAT_POINT_BLEND_OPT_STRICT_MODE_EN	REG_BIT(12)
+#define   XEHP_DUAL_SIMD8_SEQ_MERGE_DISABLE	REG_BIT(12)
+#define   GEN11_BLEND_EMB_FIX_DISABLE_IN_RCC	REG_BIT(11)
+#define   GEN12_DISABLE_CPS_AWARE_COLOR_PIPE	REG_BIT(9)
+
+/* GEN9 chicken */
+#define SLICE_ECO_CHICKEN0			_MMIO(0x7308)
+#define   PIXEL_MASK_CAMMING_DISABLE		(1 << 14)
+
+#define GEN9_SLICE_COMMON_ECO_CHICKEN0		_MMIO(0x7308)
+#define   DISABLE_PIXEL_MASK_CAMMING		(1 << 14)
+
+#define GEN9_SLICE_COMMON_ECO_CHICKEN1		_MMIO(0x731c)
+#define   GEN11_STATE_CACHE_REDIRECT_TO_CS	(1 << 11)
+
+#define SLICE_COMMON_ECO_CHICKEN1		_MMIO(0x731c)
+#define   MSC_MSAA_REODER_BUF_BYPASS_DISABLE	REG_BIT(14)
+
+#define GEN9_SLICE_PGCTL_ACK(slice)		_MMIO(0x804c + (slice) * 0x4)
+#define GEN10_SLICE_PGCTL_ACK(slice)		_MMIO(0x804c + ((slice) / 3) * 0x34 + \
+						      ((slice) % 3) * 0x4)
+#define   GEN9_PGCTL_SLICE_ACK			(1 << 0)
+#define   GEN9_PGCTL_SS_ACK(subslice)		(1 << (2 + (subslice) * 2))
+#define   GEN10_PGCTL_VALID_SS_MASK(slice)	((slice) == 0 ? 0x7F : 0x1F)
+
+#define GEN9_SS01_EU_PGCTL_ACK(slice)		_MMIO(0x805c + (slice) * 0x8)
+#define GEN10_SS01_EU_PGCTL_ACK(slice)		_MMIO(0x805c + ((slice) / 3) * 0x30 + \
+						      ((slice) % 3) * 0x8)
+#define GEN9_SS23_EU_PGCTL_ACK(slice)		_MMIO(0x8060 + (slice) * 0x8)
+#define GEN10_SS23_EU_PGCTL_ACK(slice)		_MMIO(0x8060 + ((slice) / 3) * 0x30 + \
+						      ((slice) % 3) * 0x8)
+#define   GEN9_PGCTL_SSA_EU08_ACK		(1 << 0)
+#define   GEN9_PGCTL_SSA_EU19_ACK		(1 << 2)
+#define   GEN9_PGCTL_SSA_EU210_ACK		(1 << 4)
+#define   GEN9_PGCTL_SSA_EU311_ACK		(1 << 6)
+#define   GEN9_PGCTL_SSB_EU08_ACK		(1 << 8)
+#define   GEN9_PGCTL_SSB_EU19_ACK		(1 << 10)
+#define   GEN9_PGCTL_SSB_EU210_ACK		(1 << 12)
+#define   GEN9_PGCTL_SSB_EU311_ACK		(1 << 14)
+
+#define VF_PREEMPTION				_MMIO(0x83a4)
+#define   PREEMPTION_VERTEX_COUNT		REG_GENMASK(15, 0)
+
+#define GEN8_RC6_CTX_INFO			_MMIO(0x8504)
+
+#define GEN12_SQCM				_MMIO(0x8724)
+#define   EN_32B_ACCESS				REG_BIT(30)
+
+#define HSW_IDICR				_MMIO(0x9008)
+#define   IDIHASHMSK(x)				(((x) & 0x3f) << 16)
+
+#define GEN6_MBCUNIT_SNPCR			_MMIO(0x900c) /* for LLC config */
+#define   GEN6_MBC_SNPCR_SHIFT			21
+#define   GEN6_MBC_SNPCR_MASK			(3 << 21)
+#define   GEN6_MBC_SNPCR_MAX			(0 << 21)
+#define   GEN6_MBC_SNPCR_MED			(1 << 21)
+#define   GEN6_MBC_SNPCR_LOW			(2 << 21)
+#define   GEN6_MBC_SNPCR_MIN			(3 << 21) /* only 1/16th of the cache is shared */
+
+#define VLV_G3DCTL				_MMIO(0x9024)
+#define VLV_GSCKGCTL				_MMIO(0x9028)
+
+/* WaCatErrorRejectionIssue */
+#define GEN7_SQ_CHICKEN_MBCUNIT_CONFIG		_MMIO(0x9030)
+#define   GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB	(1 << 11)
+
+#define FBC_LLC_READ_CTRL			_MMIO(0x9044)
+#define   FBC_LLC_FULLY_OPEN			REG_BIT(30)
+
+#define GEN6_MBCTL				_MMIO(0x907c)
+#define   GEN6_MBCTL_ENABLE_BOOT_FETCH		(1 << 4)
+#define   GEN6_MBCTL_CTX_FETCH_NEEDED		(1 << 3)
+#define   GEN6_MBCTL_BME_UPDATE_ENABLE		(1 << 2)
+#define   GEN6_MBCTL_MAE_UPDATE_ENABLE		(1 << 1)
+#define   GEN6_MBCTL_BOOT_FETCH_MECH		(1 << 0)
+
+/* Fuse readout registers for GT */
+#define	GEN10_MIRROR_FUSE3			_MMIO(0x9118)
+#define   GEN10_L3BANK_PAIR_COUNT		4
+#define   GEN10_L3BANK_MASK			0x0F
+/* on Xe_HP the same fuses indicates mslices instead of L3 banks */
+#define   GEN12_MAX_MSLICES			4
+#define   GEN12_MEML3_EN_MASK			0x0F
+
+#define HSW_PAVP_FUSE1				_MMIO(0x911c)
+#define   XEHP_SFC_ENABLE_MASK			REG_GENMASK(27, 24)
+#define   HSW_F1_EU_DIS_MASK			REG_GENMASK(17, 16)
+#define   HSW_F1_EU_DIS_10EUS			0
+#define   HSW_F1_EU_DIS_8EUS			1
+#define   HSW_F1_EU_DIS_6EUS			2
+
+#define GEN8_FUSE2				_MMIO(0x9120)
+#define   GEN8_F2_SS_DIS_SHIFT			21
+#define   GEN8_F2_SS_DIS_MASK			(0x7 << GEN8_F2_SS_DIS_SHIFT)
+#define   GEN8_F2_S_ENA_SHIFT			25
+#define   GEN8_F2_S_ENA_MASK			(0x7 << GEN8_F2_S_ENA_SHIFT)
+#define   GEN9_F2_SS_DIS_SHIFT			20
+#define   GEN9_F2_SS_DIS_MASK			(0xf << GEN9_F2_SS_DIS_SHIFT)
+#define   GEN10_F2_S_ENA_SHIFT			22
+#define   GEN10_F2_S_ENA_MASK			(0x3f << GEN10_F2_S_ENA_SHIFT)
+#define   GEN10_F2_SS_DIS_SHIFT			18
+#define   GEN10_F2_SS_DIS_MASK			(0xf << GEN10_F2_SS_DIS_SHIFT)
+
+#define GEN8_EU_DISABLE0			_MMIO(0x9134)
+#define GEN9_EU_DISABLE(slice)			_MMIO(0x9134 + (slice) * 0x4)
+#define GEN11_EU_DISABLE			_MMIO(0x9134)
+#define   GEN8_EU_DIS0_S0_MASK			0xffffff
+#define   GEN8_EU_DIS0_S1_SHIFT			24
+#define   GEN8_EU_DIS0_S1_MASK			(0xff << GEN8_EU_DIS0_S1_SHIFT)
+#define   GEN11_EU_DIS_MASK			0xFF
+#define XEHP_EU_ENABLE				_MMIO(0x9134)
+#define   XEHP_EU_ENA_MASK			0xFF
+
+#define GEN8_EU_DISABLE1			_MMIO(0x9138)
+#define   GEN8_EU_DIS1_S1_MASK			0xffff
+#define   GEN8_EU_DIS1_S2_SHIFT			16
+#define   GEN8_EU_DIS1_S2_MASK			(0xffff << GEN8_EU_DIS1_S2_SHIFT)
+
+#define GEN11_GT_SLICE_ENABLE			_MMIO(0x9138)
+#define   GEN11_GT_S_ENA_MASK			0xFF
+
+#define GEN8_EU_DISABLE2			_MMIO(0x913c)
+#define   GEN8_EU_DIS2_S2_MASK			0xff
+
+#define GEN11_GT_SUBSLICE_DISABLE		_MMIO(0x913c)
+#define GEN12_GT_GEOMETRY_DSS_ENABLE		_MMIO(0x913c)
+
+#define GEN10_EU_DISABLE3			_MMIO(0x9140)
+#define   GEN10_EU_DIS_SS_MASK			0xff
+#define GEN11_GT_VEBOX_VDBOX_DISABLE		_MMIO(0x9140)
+#define   GEN11_GT_VDBOX_DISABLE_MASK		0xff
+#define   GEN11_GT_VEBOX_DISABLE_SHIFT		16
+#define   GEN11_GT_VEBOX_DISABLE_MASK		(0x0f << GEN11_GT_VEBOX_DISABLE_SHIFT)
+
+#define GEN12_GT_COMPUTE_DSS_ENABLE		_MMIO(0x9144)
+#define XEHPC_GT_COMPUTE_DSS_ENABLE_EXT		_MMIO(0x9148)
+
+#define GEN6_UCGCTL1				_MMIO(0x9400)
+#define   GEN6_GAMUNIT_CLOCK_GATE_DISABLE	(1 << 22)
+#define   GEN6_EU_TCUNIT_CLOCK_GATE_DISABLE	(1 << 16)
+#define   GEN6_BLBUNIT_CLOCK_GATE_DISABLE	(1 << 5)
+#define   GEN6_CSUNIT_CLOCK_GATE_DISABLE		(1 << 7)
+
+#define GEN6_UCGCTL2				_MMIO(0x9404)
+#define   GEN6_VFUNIT_CLOCK_GATE_DISABLE	(1 << 31)
+#define   GEN7_VDSUNIT_CLOCK_GATE_DISABLE	(1 << 30)
+#define   GEN7_TDLUNIT_CLOCK_GATE_DISABLE	(1 << 22)
+#define   GEN6_RCZUNIT_CLOCK_GATE_DISABLE	(1 << 13)
+#define   GEN6_RCPBUNIT_CLOCK_GATE_DISABLE	(1 << 12)
+#define   GEN6_RCCUNIT_CLOCK_GATE_DISABLE	(1 << 11)
+
+#define GEN6_UCGCTL3				_MMIO(0x9408)
+#define   GEN6_OACSUNIT_CLOCK_GATE_DISABLE	(1 << 20)
+
+#define GEN7_UCGCTL4				_MMIO(0x940c)
+#define   GEN7_L3BANK2X_CLOCK_GATE_DISABLE	(1 << 25)
+#define   GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE	(1 << 14)
+
+#define GEN6_RCGCTL1				_MMIO(0x9410)
+#define GEN6_RCGCTL2				_MMIO(0x9414)
+
+#define GEN6_GDRST				_MMIO(0x941c)
+#define   GEN6_GRDOM_FULL			(1 << 0)
+#define   GEN6_GRDOM_RENDER			(1 << 1)
+#define   GEN6_GRDOM_MEDIA			(1 << 2)
+#define   GEN6_GRDOM_BLT			(1 << 3)
+#define   GEN6_GRDOM_VECS			(1 << 4)
+#define   GEN9_GRDOM_GUC			(1 << 5)
+#define   GEN8_GRDOM_MEDIA2			(1 << 7)
+/* GEN11 changed all bit defs except for FULL & RENDER */
+#define   GEN11_GRDOM_FULL			GEN6_GRDOM_FULL
+#define   GEN11_GRDOM_RENDER			GEN6_GRDOM_RENDER
+#define   XEHPC_GRDOM_BLT8			REG_BIT(31)
+#define   XEHPC_GRDOM_BLT7			REG_BIT(30)
+#define   XEHPC_GRDOM_BLT6			REG_BIT(29)
+#define   XEHPC_GRDOM_BLT5			REG_BIT(28)
+#define   XEHPC_GRDOM_BLT4			REG_BIT(27)
+#define   XEHPC_GRDOM_BLT3			REG_BIT(26)
+#define   XEHPC_GRDOM_BLT2			REG_BIT(25)
+#define   XEHPC_GRDOM_BLT1			REG_BIT(24)
+#define   GEN11_GRDOM_SFC3			REG_BIT(20)
+#define   GEN11_GRDOM_SFC2			REG_BIT(19)
+#define   GEN11_GRDOM_SFC1			REG_BIT(18)
+#define   GEN11_GRDOM_SFC0			REG_BIT(17)
+#define   GEN11_GRDOM_VECS4			REG_BIT(16)
+#define   GEN11_GRDOM_VECS3			REG_BIT(15)
+#define   GEN11_GRDOM_VECS2			REG_BIT(14)
+#define   GEN11_GRDOM_VECS			REG_BIT(13)
+#define   GEN11_GRDOM_MEDIA8			REG_BIT(12)
+#define   GEN11_GRDOM_MEDIA7			REG_BIT(11)
+#define   GEN11_GRDOM_MEDIA6			REG_BIT(10)
+#define   GEN11_GRDOM_MEDIA5			REG_BIT(9)
+#define   GEN11_GRDOM_MEDIA4			REG_BIT(8)
+#define   GEN11_GRDOM_MEDIA3			REG_BIT(7)
+#define   GEN11_GRDOM_MEDIA2			REG_BIT(6)
+#define   GEN11_GRDOM_MEDIA			REG_BIT(5)
+#define   GEN11_GRDOM_GUC			REG_BIT(3)
+#define   GEN11_GRDOM_BLT			REG_BIT(2)
+#define   GEN11_VCS_SFC_RESET_BIT(instance)	(GEN11_GRDOM_SFC0 << ((instance) >> 1))
+#define   GEN11_VECS_SFC_RESET_BIT(instance)	(GEN11_GRDOM_SFC0 << (instance))
+
+#define GEN6_RSTCTL				_MMIO(0x9420)
+
+#define GEN7_MISCCPCTL				_MMIO(0x9424)
+#define   GEN7_DOP_CLOCK_GATE_ENABLE		(1 << 0)
+#define   GEN12_DOP_CLOCK_GATE_RENDER_ENABLE	REG_BIT(1)
+#define   GEN8_DOP_CLOCK_GATE_CFCLK_ENABLE	(1 << 2)
+#define   GEN8_DOP_CLOCK_GATE_GUC_ENABLE	(1 << 4)
+#define   GEN8_DOP_CLOCK_GATE_MEDIA_ENABLE	(1 << 6)
+
+#define GEN8_UCGCTL6				_MMIO(0x9430)
+#define   GEN8_GAPSUNIT_CLOCK_GATE_DISABLE	(1 << 24)
+#define   GEN8_SDEUNIT_CLOCK_GATE_DISABLE	(1 << 14)
+#define   GEN8_HDCUNIT_CLOCK_GATE_DISABLE_HDCREQ	(1 << 28)
+
+#define UNSLCGCTL9430				_MMIO(0x9430)
+#define   MSQDUNIT_CLKGATE_DIS			REG_BIT(3)
+
+#define UNSLICE_UNIT_LEVEL_CLKGATE		_MMIO(0x9434)
+#define   VFUNIT_CLKGATE_DIS			REG_BIT(20)
+#define   TSGUNIT_CLKGATE_DIS			REG_BIT(17) /* XEHPSDV */
+#define   CG3DDISCFEG_CLKGATE_DIS		REG_BIT(17) /* DG2 */
+#define   GAMEDIA_CLKGATE_DIS			REG_BIT(11)
+#define   HSUNIT_CLKGATE_DIS			REG_BIT(8)
+#define   VSUNIT_CLKGATE_DIS			REG_BIT(3)
+
+#define UNSLCGCTL9440				_MMIO(0x9440)
+#define   GAMTLBOACS_CLKGATE_DIS		REG_BIT(28)
+#define   GAMTLBVDBOX5_CLKGATE_DIS		REG_BIT(27)
+#define   GAMTLBVDBOX6_CLKGATE_DIS		REG_BIT(26)
+#define   GAMTLBVDBOX3_CLKGATE_DIS		REG_BIT(24)
+#define   GAMTLBVDBOX4_CLKGATE_DIS		REG_BIT(23)
+#define   GAMTLBVDBOX7_CLKGATE_DIS		REG_BIT(22)
+#define   GAMTLBVDBOX2_CLKGATE_DIS		REG_BIT(21)
+#define   GAMTLBVDBOX0_CLKGATE_DIS		REG_BIT(17)
+#define   GAMTLBKCR_CLKGATE_DIS			REG_BIT(16)
+#define   GAMTLBGUC_CLKGATE_DIS			REG_BIT(15)
+#define   GAMTLBBLT_CLKGATE_DIS			REG_BIT(14)
+#define   GAMTLBVDBOX1_CLKGATE_DIS		REG_BIT(6)
+
+#define UNSLCGCTL9444				_MMIO(0x9444)
+#define   GAMTLBGFXA0_CLKGATE_DIS		REG_BIT(30)
+#define   GAMTLBGFXA1_CLKGATE_DIS		REG_BIT(29)
+#define   GAMTLBCOMPA0_CLKGATE_DIS		REG_BIT(28)
+#define   GAMTLBCOMPA1_CLKGATE_DIS		REG_BIT(27)
+#define   GAMTLBCOMPB0_CLKGATE_DIS		REG_BIT(26)
+#define   GAMTLBCOMPB1_CLKGATE_DIS		REG_BIT(25)
+#define   GAMTLBCOMPC0_CLKGATE_DIS		REG_BIT(24)
+#define   GAMTLBCOMPC1_CLKGATE_DIS		REG_BIT(23)
+#define   GAMTLBCOMPD0_CLKGATE_DIS		REG_BIT(22)
+#define   GAMTLBCOMPD1_CLKGATE_DIS		REG_BIT(21)
+#define   GAMTLBMERT_CLKGATE_DIS		REG_BIT(20)
+#define   GAMTLBVEBOX3_CLKGATE_DIS		REG_BIT(19)
+#define   GAMTLBVEBOX2_CLKGATE_DIS		REG_BIT(18)
+#define   GAMTLBVEBOX1_CLKGATE_DIS		REG_BIT(17)
+#define   GAMTLBVEBOX0_CLKGATE_DIS		REG_BIT(16)
+#define   LTCDD_CLKGATE_DIS			REG_BIT(10)
+
+#define SLICE_UNIT_LEVEL_CLKGATE		_MMIO(0x94d4)
+#define   SARBUNIT_CLKGATE_DIS			(1 << 5)
+#define   RCCUNIT_CLKGATE_DIS			(1 << 7)
+#define   MSCUNIT_CLKGATE_DIS			(1 << 10)
+#define   NODEDSS_CLKGATE_DIS			REG_BIT(12)
+#define   L3_CLKGATE_DIS			REG_BIT(16)
+#define   L3_CR2X_CLKGATE_DIS			REG_BIT(17)
+
+#define SCCGCTL94DC				_MMIO(0x94dc)
+#define   CG3DDISURB				REG_BIT(14)
+
+#define UNSLICE_UNIT_LEVEL_CLKGATE2		_MMIO(0x94e4)
+#define   VSUNIT_CLKGATE_DIS_TGL		REG_BIT(19)
+#define   PSDUNIT_CLKGATE_DIS			REG_BIT(5)
+
+#define SUBSLICE_UNIT_LEVEL_CLKGATE		_MMIO(0x9524)
+#define   DSS_ROUTER_CLKGATE_DIS		REG_BIT(28)
+#define   GWUNIT_CLKGATE_DIS			REG_BIT(16)
+
+#define SUBSLICE_UNIT_LEVEL_CLKGATE2		_MMIO(0x9528)
+#define   CPSSUNIT_CLKGATE_DIS			REG_BIT(9)
+
+#define SSMCGCTL9530				_MMIO(0x9530)
+#define   RTFUNIT_CLKGATE_DIS			REG_BIT(18)
+
+#define GEN10_DFR_RATIO_EN_AND_CHICKEN		_MMIO(0x9550)
+#define   DFR_DISABLE				(1 << 9)
+
+#define INF_UNIT_LEVEL_CLKGATE			_MMIO(0x9560)
+#define   CGPSF_CLKGATE_DIS			(1 << 3)
+
+#define MICRO_BP0_0				_MMIO(0x9800)
+#define MICRO_BP0_2				_MMIO(0x9804)
+#define MICRO_BP0_1				_MMIO(0x9808)
+#define MICRO_BP1_0				_MMIO(0x980c)
+#define MICRO_BP1_2				_MMIO(0x9810)
+#define MICRO_BP1_1				_MMIO(0x9814)
+#define MICRO_BP2_0				_MMIO(0x9818)
+#define MICRO_BP2_2				_MMIO(0x981c)
+#define MICRO_BP2_1				_MMIO(0x9820)
+#define MICRO_BP3_0				_MMIO(0x9824)
+#define MICRO_BP3_2				_MMIO(0x9828)
+#define MICRO_BP3_1				_MMIO(0x982c)
+#define MICRO_BP_TRIGGER			_MMIO(0x9830)
+#define MICRO_BP3_COUNT_STATUS01		_MMIO(0x9834)
+#define MICRO_BP3_COUNT_STATUS23		_MMIO(0x9838)
+#define MICRO_BP_FIRED_ARMED			_MMIO(0x983c)
+
+#define GEN6_GFXPAUSE				_MMIO(0xa000)
+#define GEN6_RPNSWREQ				_MMIO(0xa008)
+#define   GEN6_TURBO_DISABLE			(1 << 31)
+#define   GEN6_FREQUENCY(x)			((x) << 25)
+#define   HSW_FREQUENCY(x)			((x) << 24)
+#define   GEN9_FREQUENCY(x)			((x) << 23)
+#define   GEN6_OFFSET(x)			((x) << 19)
+#define   GEN6_AGGRESSIVE_TURBO			(0 << 15)
+#define   GEN9_SW_REQ_UNSLICE_RATIO_SHIFT	23
+#define   GEN9_IGNORE_SLICE_RATIO		(0 << 0)
+#define   GEN12_MEDIA_FREQ_RATIO		REG_BIT(13)
+
+#define GEN6_RC_VIDEO_FREQ			_MMIO(0xa00c)
+#define   GEN6_RC_CTL_RC6pp_ENABLE		(1 << 16)
+#define   GEN6_RC_CTL_RC6p_ENABLE		(1 << 17)
+#define   GEN6_RC_CTL_RC6_ENABLE		(1 << 18)
+#define   GEN6_RC_CTL_RC1e_ENABLE		(1 << 20)
+#define   GEN6_RC_CTL_RC7_ENABLE		(1 << 22)
+#define   VLV_RC_CTL_CTX_RST_PARALLEL		(1 << 24)
+#define   GEN7_RC_CTL_TO_MODE			(1 << 28)
+#define   GEN6_RC_CTL_EI_MODE(x)		((x) << 27)
+#define   GEN6_RC_CTL_HW_ENABLE			(1 << 31)
+#define GEN6_RP_DOWN_TIMEOUT			_MMIO(0xa010)
+#define GEN6_RP_INTERRUPT_LIMITS		_MMIO(0xa014)
+#define GEN6_RPSTAT1				_MMIO(0xa01c)
+#define   GEN6_CAGF_SHIFT			8
+#define   HSW_CAGF_SHIFT			7
+#define   GEN9_CAGF_SHIFT			23
+#define   GEN6_CAGF_MASK			(0x7f << GEN6_CAGF_SHIFT)
+#define   HSW_CAGF_MASK				(0x7f << HSW_CAGF_SHIFT)
+#define   GEN9_CAGF_MASK			(0x1ff << GEN9_CAGF_SHIFT)
+#define GEN6_RP_CONTROL				_MMIO(0xa024)
+#define   GEN6_RP_MEDIA_TURBO			(1 << 11)
+#define   GEN6_RP_MEDIA_MODE_MASK		(3 << 9)
+#define   GEN6_RP_MEDIA_HW_TURBO_MODE		(3 << 9)
+#define   GEN6_RP_MEDIA_HW_NORMAL_MODE		(2 << 9)
+#define   GEN6_RP_MEDIA_HW_MODE			(1 << 9)
+#define   GEN6_RP_MEDIA_SW_MODE			(0 << 9)
+#define   GEN6_RP_MEDIA_IS_GFX			(1 << 8)
+#define   GEN6_RP_ENABLE			(1 << 7)
+#define   GEN6_RP_UP_IDLE_MIN			(0x1 << 3)
+#define   GEN6_RP_UP_BUSY_AVG			(0x2 << 3)
+#define   GEN6_RP_UP_BUSY_CONT			(0x4 << 3)
+#define   GEN6_RP_DOWN_IDLE_AVG			(0x2 << 0)
+#define   GEN6_RP_DOWN_IDLE_CONT		(0x1 << 0)
+#define   GEN6_RPSWCTL_SHIFT			9
+#define   GEN9_RPSWCTL_ENABLE			(0x2 << GEN6_RPSWCTL_SHIFT)
+#define   GEN9_RPSWCTL_DISABLE			(0x0 << GEN6_RPSWCTL_SHIFT)
+#define GEN6_RP_UP_THRESHOLD			_MMIO(0xa02c)
+#define GEN6_RP_DOWN_THRESHOLD			_MMIO(0xa030)
+#define GEN6_RP_CUR_UP_EI			_MMIO(0xa050)
+#define   GEN6_RP_EI_MASK			0xffffff
+#define   GEN6_CURICONT_MASK			GEN6_RP_EI_MASK
+#define GEN6_RP_CUR_UP				_MMIO(0xa054)
+#define   GEN6_CURBSYTAVG_MASK			GEN6_RP_EI_MASK
+#define GEN6_RP_PREV_UP				_MMIO(0xa058)
+#define GEN6_RP_CUR_DOWN_EI			_MMIO(0xa05c)
+#define   GEN6_CURIAVG_MASK			GEN6_RP_EI_MASK
+#define GEN6_RP_CUR_DOWN			_MMIO(0xa060)
+#define GEN6_RP_PREV_DOWN			_MMIO(0xa064)
+#define GEN6_RP_UP_EI				_MMIO(0xa068)
+#define GEN6_RP_DOWN_EI				_MMIO(0xa06c)
+#define GEN6_RP_IDLE_HYSTERSIS			_MMIO(0xa070)
+#define GEN6_RPDEUHWTC				_MMIO(0xa080)
+#define GEN6_RPDEUC				_MMIO(0xa084)
+#define GEN6_RPDEUCSW				_MMIO(0xa088)
+#define GEN6_RC_CONTROL				_MMIO(0xa090)
+#define GEN6_RC_STATE				_MMIO(0xa094)
+#define   RC_SW_TARGET_STATE_SHIFT		16
+#define   RC_SW_TARGET_STATE_MASK		(7 << RC_SW_TARGET_STATE_SHIFT)
+#define GEN6_RC1_WAKE_RATE_LIMIT		_MMIO(0xa098)
+#define GEN6_RC6_WAKE_RATE_LIMIT		_MMIO(0xa09c)
+#define GEN6_RC6pp_WAKE_RATE_LIMIT		_MMIO(0xa0a0)
+#define GEN10_MEDIA_WAKE_RATE_LIMIT		_MMIO(0xa0a0)
+#define GEN6_RC_EVALUATION_INTERVAL		_MMIO(0xa0a8)
+#define GEN6_RC_IDLE_HYSTERSIS			_MMIO(0xa0ac)
+#define GEN6_RC_SLEEP				_MMIO(0xa0b0)
+#define GEN6_RCUBMABDTMR			_MMIO(0xa0b0)
+#define GEN6_RC1e_THRESHOLD			_MMIO(0xa0b4)
+#define GEN6_RC6_THRESHOLD			_MMIO(0xa0b8)
+#define GEN6_RC6p_THRESHOLD			_MMIO(0xa0bc)
+#define VLV_RCEDATA				_MMIO(0xa0bc)
+#define GEN6_RC6pp_THRESHOLD			_MMIO(0xa0c0)
+#define GEN9_MEDIA_PG_IDLE_HYSTERESIS		_MMIO(0xa0c4)
+#define GEN9_RENDER_PG_IDLE_HYSTERESIS		_MMIO(0xa0c8)
+
+#define GEN6_PMINTRMSK				_MMIO(0xa168)
+#define   GEN8_PMINTR_DISABLE_REDIRECT_TO_GUC	(1 << 31)
+#define   ARAT_EXPIRED_INTRMSK			(1 << 9)
+
+#define GEN8_MISC_CTRL0				_MMIO(0xa180)
+
+#define ECOBUS					_MMIO(0xa180)
+#define    FORCEWAKE_MT_ENABLE			(1 << 5)
+
+#define FORCEWAKE_MT				_MMIO(0xa188) /* multi-threaded */
+#define FORCEWAKE_GT_GEN9			_MMIO(0xa188)
+#define FORCEWAKE				_MMIO(0xa18c)
+
+#define VLV_SPAREG2H				_MMIO(0xa194)
+
+#define GEN9_PG_ENABLE				_MMIO(0xa210)
+#define   GEN9_RENDER_PG_ENABLE			REG_BIT(0)
+#define   GEN9_MEDIA_PG_ENABLE			REG_BIT(1)
+#define   GEN11_MEDIA_SAMPLER_PG_ENABLE		REG_BIT(2)
+#define   VDN_HCP_POWERGATE_ENABLE(n)		REG_BIT(3 + 2 * (n))
+#define   VDN_MFX_POWERGATE_ENABLE(n)		REG_BIT(4 + 2 * (n))
+
+#define GEN8_PUSHBUS_CONTROL			_MMIO(0xa248)
+#define GEN8_PUSHBUS_ENABLE			_MMIO(0xa250)
+#define GEN8_PUSHBUS_SHIFT			_MMIO(0xa25c)
+
+/* GPM unit config (Gen9+) */
+#define CTC_MODE				_MMIO(0xa26c)
+#define   CTC_SOURCE_PARAMETER_MASK		1
+#define   CTC_SOURCE_CRYSTAL_CLOCK		0
+#define   CTC_SOURCE_DIVIDE_LOGIC		1
+#define   CTC_SHIFT_PARAMETER_SHIFT		1
+#define   CTC_SHIFT_PARAMETER_MASK		(0x3 << CTC_SHIFT_PARAMETER_SHIFT)
+
+/* GPM MSG_IDLE */
+#define MSG_IDLE_CS		_MMIO(0x8000)
+#define MSG_IDLE_VCS0		_MMIO(0x8004)
+#define MSG_IDLE_VCS1		_MMIO(0x8008)
+#define MSG_IDLE_BCS		_MMIO(0x800C)
+#define MSG_IDLE_VECS0		_MMIO(0x8010)
+#define MSG_IDLE_VCS2		_MMIO(0x80C0)
+#define MSG_IDLE_VCS3		_MMIO(0x80C4)
+#define MSG_IDLE_VCS4		_MMIO(0x80C8)
+#define MSG_IDLE_VCS5		_MMIO(0x80CC)
+#define MSG_IDLE_VCS6		_MMIO(0x80D0)
+#define MSG_IDLE_VCS7		_MMIO(0x80D4)
+#define MSG_IDLE_VECS1		_MMIO(0x80D8)
+#define MSG_IDLE_VECS2		_MMIO(0x80DC)
+#define MSG_IDLE_VECS3		_MMIO(0x80E0)
+#define  MSG_IDLE_FW_MASK	REG_GENMASK(13, 9)
+#define  MSG_IDLE_FW_SHIFT	9
+
+#define FORCEWAKE_MEDIA_GEN9			_MMIO(0xa270)
+#define FORCEWAKE_RENDER_GEN9			_MMIO(0xa278)
+
+#define VLV_PWRDWNUPCTL				_MMIO(0xa294)
+
+#define GEN9_PWRGT_DOMAIN_STATUS		_MMIO(0xa2a0)
+#define   GEN9_PWRGT_MEDIA_STATUS_MASK		(1 << 0)
+#define   GEN9_PWRGT_RENDER_STATUS_MASK		(1 << 1)
+
+#define MISC_STATUS0				_MMIO(0xa500)
+#define MISC_STATUS1				_MMIO(0xa504)
+
+#define FORCEWAKE_MEDIA_VDBOX_GEN11(n)		_MMIO(0xa540 + (n) * 4)
+#define FORCEWAKE_MEDIA_VEBOX_GEN11(n)		_MMIO(0xa560 + (n) * 4)
+
+#define CHV_POWER_SS0_SIG1			_MMIO(0xa720)
+#define CHV_POWER_SS0_SIG2			_MMIO(0xa724)
+#define CHV_POWER_SS1_SIG1			_MMIO(0xa728)
+#define   CHV_SS_PG_ENABLE			(1 << 1)
+#define   CHV_EU08_PG_ENABLE			(1 << 9)
+#define   CHV_EU19_PG_ENABLE			(1 << 17)
+#define   CHV_EU210_PG_ENABLE			(1 << 25)
+#define CHV_POWER_SS1_SIG2			_MMIO(0xa72c)
+#define   CHV_EU311_PG_ENABLE			(1 << 1)
+
+#define GEN7_SARCHKMD				_MMIO(0xb000)
+#define   GEN7_DISABLE_DEMAND_PREFETCH		(1 << 31)
+#define   GEN7_DISABLE_SAMPLER_PREFETCH		(1 << 30)
+
+#define GEN8_GARBCNTL				_MMIO(0xb004)
+#define   GEN9_GAPS_TSV_CREDIT_DISABLE		(1 << 7)
+#define   GEN11_ARBITRATION_PRIO_ORDER_MASK	(0x3f << 22)
+#define   GEN11_HASH_CTRL_EXCL_MASK		(0x7f << 0)
+#define   GEN11_HASH_CTRL_EXCL_BIT0		(1 << 0)
+
+#define GEN9_SCRATCH_LNCF1			_MMIO(0xb008)
+#define   GEN9_LNCF_NONIA_COHERENT_ATOMICS_ENABLE	REG_BIT(0)
+
+#define GEN7_L3SQCREG1				_MMIO(0xb010)
+#define   VLV_B0_WA_L3SQCREG1_VALUE		0x00D30000
+
+#define GEN7_L3CNTLREG1				_MMIO(0xb01c)
+#define   GEN7_WA_FOR_GEN7_L3_CONTROL		0x3C47FF8C
+#define   GEN7_L3AGDIS				(1 << 19)
+
+#define XEHPC_LNCFMISCCFGREG0			_MMIO(0xb01c)
+#define   XEHPC_OVRLSCCC			REG_BIT(0)
+
+#define GEN7_L3CNTLREG2				_MMIO(0xb020)
+
+/* MOCS (Memory Object Control State) registers */
+#define GEN9_LNCFCMOCS(i)			_MMIO(0xb020 + (i) * 4)	/* L3 Cache Control */
+#define GEN9_LNCFCMOCS_REG_COUNT		32
+
+#define GEN7_L3CNTLREG3				_MMIO(0xb024)
+
+#define GEN7_L3_CHICKEN_MODE_REGISTER		_MMIO(0xb030)
+#define   GEN7_WA_L3_CHICKEN_MODE		0x20000000
+
+#define GEN7_L3SQCREG4				_MMIO(0xb034)
+#define   L3SQ_URB_READ_CAM_MATCH_DISABLE	(1 << 27)
+
+#define HSW_SCRATCH1				_MMIO(0xb038)
+#define   HSW_SCRATCH1_L3_DATA_ATOMICS_DISABLE	(1 << 27)
+
+#define GEN7_L3LOG(slice, i)			_MMIO(0xb070 + (slice) * 0x200 + (i) * 4)
+#define   GEN7_L3LOG_SIZE			0x80
+
+#define GEN10_SCRATCH_LNCF2			_MMIO(0xb0a0)
+#define   PMFLUSHDONE_LNICRSDROP		(1 << 20)
+#define   PMFLUSH_GAPL3UNBLOCK			(1 << 21)
+#define   PMFLUSHDONE_LNEBLK			(1 << 22)
+
+#define XEHP_L3NODEARBCFG			_MMIO(0xb0b4)
+#define   XEHP_LNESPARE				REG_BIT(19)
+
+#define GEN8_L3SQCREG1				_MMIO(0xb100)
+/*
+ * Note that on CHV the following has an off-by-one error wrt. to BSpec.
+ * Using the formula in BSpec leads to a hang, while the formula here works
+ * fine and matches the formulas for all other platforms. A BSpec change
+ * request has been filed to clarify this.
+ */
+#define   L3_GENERAL_PRIO_CREDITS(x)		(((x) >> 1) << 19)
+#define   L3_HIGH_PRIO_CREDITS(x)		(((x) >> 1) << 14)
+#define   L3_PRIO_CREDITS_MASK			((0x1f << 19) | (0x1f << 14))
+
+#define GEN10_L3_CHICKEN_MODE_REGISTER		_MMIO(0xb114)
+#define   GEN11_I2M_WRITE_DISABLE		(1 << 28)
+
+#define GEN8_L3SQCREG4				_MMIO(0xb118)
+#define   GEN11_LQSC_CLEAN_EVICT_DISABLE	(1 << 6)
+#define   GEN8_LQSC_RO_PERF_DIS			(1 << 27)
+#define   GEN8_LQSC_FLUSH_COHERENT_LINES	(1 << 21)
+#define   GEN8_LQSQ_NONIA_COHERENT_ATOMICS_ENABLE	REG_BIT(22)
+
+#define GEN9_SCRATCH1				_MMIO(0xb11c)
+#define   EVICTION_PERF_FIX_ENABLE		REG_BIT(8)
+
+#define BDW_SCRATCH1				_MMIO(0xb11c)
+#define   GEN9_LBS_SLA_RETRY_TIMER_DECREMENT_ENABLE	(1 << 2)
+
+#define GEN11_SCRATCH2				_MMIO(0xb140)
+#define   GEN11_COHERENT_PARTIAL_WRITE_MERGE_ENABLE	(1 << 19)
+
+#define GEN11_L3SQCREG5				_MMIO(0xb158)
+#define   L3_PWM_TIMER_INIT_VAL_MASK		REG_GENMASK(9, 0)
+
+#define MLTICTXCTL				_MMIO(0xb170)
+#define   TDONRENDER				REG_BIT(2)
+
+#define XEHP_L3SCQREG7				_MMIO(0xb188)
+#define   BLEND_FILL_CACHING_OPT_DIS		REG_BIT(3)
+
+#define XEHPC_L3SCRUB				_MMIO(0xb18c)
+#define   SCRUB_CL_DWNGRADE_SHARED		REG_BIT(12)
+#define   SCRUB_RATE_PER_BANK_MASK		REG_GENMASK(2, 0)
+#define   SCRUB_RATE_4B_PER_CLK			REG_FIELD_PREP(SCRUB_RATE_PER_BANK_MASK, 0x6)
+
+#define L3SQCREG1_CCS0				_MMIO(0xb200)
+#define   FLUSHALLNONCOH			REG_BIT(5)
+
+#define GEN11_GLBLINVL				_MMIO(0xb404)
+#define   GEN11_BANK_HASH_ADDR_EXCL_MASK	(0x7f << 5)
+#define   GEN11_BANK_HASH_ADDR_EXCL_BIT0	(1 << 5)
+
+#define GEN11_LSN_UNSLCVC			_MMIO(0xb43c)
+#define   GEN11_LSN_UNSLCVC_GAFS_HALF_CL2_MAXALLOC	(1 << 9)
+#define   GEN11_LSN_UNSLCVC_GAFS_HALF_SF_MAXALLOC	(1 << 7)
+
+#define GUCPMTIMESTAMP				_MMIO(0xc3e8)
+
+#define __GEN9_RCS0_MOCS0			0xc800
+#define GEN9_GFX_MOCS(i)			_MMIO(__GEN9_RCS0_MOCS0 + (i) * 4)
+#define __GEN9_VCS0_MOCS0			0xc900
+#define GEN9_MFX0_MOCS(i)			_MMIO(__GEN9_VCS0_MOCS0 + (i) * 4)
+#define __GEN9_VCS1_MOCS0			0xca00
+#define GEN9_MFX1_MOCS(i)			_MMIO(__GEN9_VCS1_MOCS0 + (i) * 4)
+#define __GEN9_VECS0_MOCS0			0xcb00
+#define GEN9_VEBOX_MOCS(i)			_MMIO(__GEN9_VECS0_MOCS0 + (i) * 4)
+#define __GEN9_BCS0_MOCS0			0xcc00
+#define GEN9_BLT_MOCS(i)			_MMIO(__GEN9_BCS0_MOCS0 + (i) * 4)
+
+#define GEN12_FAULT_TLB_DATA0			_MMIO(0xceb8)
+#define GEN12_FAULT_TLB_DATA1			_MMIO(0xcebc)
+#define   FAULT_VA_HIGH_BITS			(0xf << 0)
+#define   FAULT_GTT_SEL				(1 << 4)
+
+#define GEN12_RING_FAULT_REG			_MMIO(0xcec4)
+#define   GEN8_RING_FAULT_ENGINE_ID(x)		(((x) >> 12) & 0x7)
+#define   RING_FAULT_GTTSEL_MASK		(1 << 11)
+#define   RING_FAULT_SRCID(x)			(((x) >> 3) & 0xff)
+#define   RING_FAULT_FAULT_TYPE(x)		(((x) >> 1) & 0x3)
+#define   RING_FAULT_VALID			(1 << 0)
+
+#define GEN12_GFX_TLB_INV_CR			_MMIO(0xced8)
+#define GEN12_VD_TLB_INV_CR			_MMIO(0xcedc)
+#define GEN12_VE_TLB_INV_CR			_MMIO(0xcee0)
+#define GEN12_BLT_TLB_INV_CR			_MMIO(0xcee4)
+#define GEN12_COMPCTX_TLB_INV_CR		_MMIO(0xcf04)
+
+#define GEN12_MERT_MOD_CTRL			_MMIO(0xcf28)
+#define RENDER_MOD_CTRL				_MMIO(0xcf2c)
+#define COMP_MOD_CTRL				_MMIO(0xcf30)
+#define VDBX_MOD_CTRL				_MMIO(0xcf34)
+#define VEBX_MOD_CTRL				_MMIO(0xcf38)
+#define   FORCE_MISS_FTLB			REG_BIT(3)
+
+#define GEN12_GAMSTLB_CTRL			_MMIO(0xcf4c)
+#define   CONTROL_BLOCK_CLKGATE_DIS		REG_BIT(12)
+#define   EGRESS_BLOCK_CLKGATE_DIS		REG_BIT(11)
+#define   TAG_BLOCK_CLKGATE_DIS			REG_BIT(7)
+
+#define GEN12_GAMCNTRL_CTRL			_MMIO(0xcf54)
+#define   INVALIDATION_BROADCAST_MODE_DIS	REG_BIT(12)
+#define   GLOBAL_INVALIDATION_MODE		REG_BIT(2)
+
+#define GEN12_GAM_DONE				_MMIO(0xcf68)
+
+#define GEN7_HALF_SLICE_CHICKEN1		_MMIO(0xe100) /* IVB GT1 + VLV */
+#define   GEN7_MAX_PS_THREAD_DEP		(8 << 12)
+#define   GEN7_SINGLE_SUBSCAN_DISPATCH_ENABLE	(1 << 10)
+#define   GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE	(1 << 4)
+#define   GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE	(1 << 3)
+
+#define GEN7_SAMPLER_INSTDONE			_MMIO(0xe160)
+#define GEN7_ROW_INSTDONE			_MMIO(0xe164)
+
+#define HALF_SLICE_CHICKEN2			_MMIO(0xe180)
+#define   GEN8_ST_PO_DISABLE			(1 << 13)
+
+#define HALF_SLICE_CHICKEN3			_MMIO(0xe184)
+#define   HSW_SAMPLE_C_PERFORMANCE		(1 << 9)
+#define   GEN8_CENTROID_PIXEL_OPT_DIS		(1 << 8)
+#define   GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC	(1 << 5)
+#define   GEN8_SAMPLER_POWER_BYPASS_DIS		(1 << 1)
+
+#define GEN9_HALF_SLICE_CHICKEN5		_MMIO(0xe188)
+#define   GEN9_DG_MIRROR_FIX_ENABLE		(1 << 5)
+#define   GEN9_CCS_TLB_PREFETCH_ENABLE		(1 << 3)
+
+#define GEN10_SAMPLER_MODE			_MMIO(0xe18c)
+#define   ENABLE_SMALLPL			REG_BIT(15)
+#define   SC_DISABLE_POWER_OPTIMIZATION_EBB	REG_BIT(9)
+#define   GEN11_SAMPLER_ENABLE_HEADLESS_MSG	REG_BIT(5)
+
+#define GEN9_HALF_SLICE_CHICKEN7		_MMIO(0xe194)
+#define   DG2_DISABLE_ROUND_ENABLE_ALLOW_FOR_SSLA	REG_BIT(15)
+#define   GEN9_SAMPLER_HASH_COMPRESSED_READ_ADDR	REG_BIT(8)
+#define   GEN9_ENABLE_YV12_BUGFIX		REG_BIT(4)
+#define   GEN9_ENABLE_GPGPU_PREEMPTION		REG_BIT(2)
+
+#define GEN10_CACHE_MODE_SS			_MMIO(0xe420)
+#define   ENABLE_EU_COUNT_FOR_TDL_FLUSH		REG_BIT(10)
+#define   DISABLE_ECC				REG_BIT(5)
+#define   FLOAT_BLEND_OPTIMIZATION_ENABLE	REG_BIT(4)
+#define   ENABLE_PREFETCH_INTO_IC		REG_BIT(3)
+
+#define EU_PERF_CNTL0				_MMIO(0xe458)
+#define EU_PERF_CNTL4				_MMIO(0xe45c)
+
+#define GEN9_ROW_CHICKEN4			_MMIO(0xe48c)
+#define   GEN12_DISABLE_GRF_CLEAR		REG_BIT(13)
+#define   XEHP_DIS_BBL_SYSPIPE			REG_BIT(11)
+#define   GEN12_DISABLE_TDL_PUSH		REG_BIT(9)
+#define   GEN11_DIS_PICK_2ND_EU			REG_BIT(7)
+#define   GEN12_DISABLE_HDR_PAST_PAYLOAD_HOLD_FIX	REG_BIT(4)
+#define   THREAD_EX_ARB_MODE			REG_GENMASK(3, 2)
+#define   THREAD_EX_ARB_MODE_RR_AFTER_DEP	REG_FIELD_PREP(THREAD_EX_ARB_MODE, 0x2)
+
+#define HSW_ROW_CHICKEN3			_MMIO(0xe49c)
+#define   HSW_ROW_CHICKEN3_L3_GLOBAL_ATOMICS_DISABLE	(1 << 6)
+
+#define GEN8_ROW_CHICKEN			_MMIO(0xe4f0)
+#define   FLOW_CONTROL_ENABLE			REG_BIT(15)
+#define   UGM_BACKUP_MODE			REG_BIT(13)
+#define   MDQ_ARBITRATION_MODE			REG_BIT(12)
+#define   SYSTOLIC_DOP_CLOCK_GATING_DIS		REG_BIT(10)
+#define   PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE	REG_BIT(8)
+#define   STALL_DOP_GATING_DISABLE		REG_BIT(5)
+#define   THROTTLE_12_5				REG_GENMASK(4, 2)
+#define   DISABLE_EARLY_EOT			REG_BIT(1)
+
+#define GEN7_ROW_CHICKEN2			_MMIO(0xe4f4)
+#define   GEN12_DISABLE_READ_SUPPRESSION	REG_BIT(15)
+#define   GEN12_DISABLE_EARLY_READ		REG_BIT(14)
+#define   GEN12_ENABLE_LARGE_GRF_MODE		REG_BIT(12)
+#define   GEN12_PUSH_CONST_DEREF_HOLD_DIS	REG_BIT(8)
+
+#define RT_CTRL					_MMIO(0xe530)
+#define   DIS_NULL_QUERY			REG_BIT(10)
+#define   STACKID_CTRL				REG_GENMASK(6, 5)
+#define   STACKID_CTRL_512			REG_FIELD_PREP(STACKID_CTRL, 0x2)
+
+#define EU_PERF_CNTL1				_MMIO(0xe558)
+#define EU_PERF_CNTL5				_MMIO(0xe55c)
+
+#define GEN12_HDC_CHICKEN0			_MMIO(0xe5f0)
+#define   LSC_L1_FLUSH_CTL_3D_DATAPORT_FLUSH_EVENTS_MASK	REG_GENMASK(13, 11)
+#define ICL_HDC_MODE				_MMIO(0xe5f4)
+
+#define EU_PERF_CNTL2				_MMIO(0xe658)
+#define EU_PERF_CNTL6				_MMIO(0xe65c)
+#define EU_PERF_CNTL3				_MMIO(0xe758)
+
+#define LSC_CHICKEN_BIT_0			_MMIO(0xe7c8)
+#define   DISABLE_D8_D16_COASLESCE		REG_BIT(30)
+#define   FORCE_1_SUB_MESSAGE_PER_FRAGMENT	REG_BIT(15)
+#define LSC_CHICKEN_BIT_0_UDW			_MMIO(0xe7c8 + 4)
+#define   DIS_CHAIN_2XSIMD8			REG_BIT(55 - 32)
+#define   FORCE_SLM_FENCE_SCOPE_TO_TILE		REG_BIT(42 - 32)
+#define   FORCE_UGM_FENCE_SCOPE_TO_TILE		REG_BIT(41 - 32)
+#define   MAXREQS_PER_BANK			REG_GENMASK(39 - 32, 37 - 32)
+#define   DISABLE_128B_EVICTION_COMMAND_UDW	REG_BIT(36 - 32)
+
+#define SARB_CHICKEN1				_MMIO(0xe90c)
+#define   COMP_CKN_IN				REG_GENMASK(30, 29)
+
+#define GEN7_HALF_SLICE_CHICKEN1_GT2		_MMIO(0xf100)
+
+#define GEN7_ROW_CHICKEN2_GT2			_MMIO(0xf4f4)
+#define   DOP_CLOCK_GATING_DISABLE		(1 << 0)
+#define   PUSH_CONSTANT_DEREF_DISABLE		(1 << 8)
+#define   GEN11_TDL_CLOCK_GATING_FIX_DISABLE	(1 << 1)
+
+#define __GEN11_VCS2_MOCS0			0x10000
+#define GEN11_MFX2_MOCS(i)			_MMIO(__GEN11_VCS2_MOCS0 + (i) * 4)
+
+#define CRSTANDVID				_MMIO(0x11100)
+#define PXVFREQ(fstart)				_MMIO(0x11110 + (fstart) * 4)  /* P[0-15]VIDFREQ (0x1114c) (Ironlake) */
+#define   PXVFREQ_PX_MASK			0x7f000000
+#define   PXVFREQ_PX_SHIFT			24
+#define VIDFREQ_BASE				_MMIO(0x11110)
+#define VIDFREQ1				_MMIO(0x11110) /* VIDFREQ1-4 (0x1111c) (Cantiga) */
+#define VIDFREQ2				_MMIO(0x11114)
+#define VIDFREQ3				_MMIO(0x11118)
+#define VIDFREQ4				_MMIO(0x1111c)
+#define   VIDFREQ_P0_MASK			0x1f000000
+#define   VIDFREQ_P0_SHIFT			24
+#define   VIDFREQ_P0_CSCLK_MASK			0x00f00000
+#define   VIDFREQ_P0_CSCLK_SHIFT		20
+#define   VIDFREQ_P0_CRCLK_MASK			0x000f0000
+#define   VIDFREQ_P0_CRCLK_SHIFT		16
+#define   VIDFREQ_P1_MASK			0x00001f00
+#define   VIDFREQ_P1_SHIFT			8
+#define   VIDFREQ_P1_CSCLK_MASK			0x000000f0
+#define   VIDFREQ_P1_CSCLK_SHIFT		4
+#define   VIDFREQ_P1_CRCLK_MASK			0x0000000f
+#define INTTOEXT_BASE				_MMIO(0x11120) /* INTTOEXT1-8 (0x1113c) */
+#define   INTTOEXT_MAP3_SHIFT			24
+#define   INTTOEXT_MAP3_MASK			(0x1f << INTTOEXT_MAP3_SHIFT)
+#define   INTTOEXT_MAP2_SHIFT			16
+#define   INTTOEXT_MAP2_MASK			(0x1f << INTTOEXT_MAP2_SHIFT)
+#define   INTTOEXT_MAP1_SHIFT			8
+#define   INTTOEXT_MAP1_MASK			(0x1f << INTTOEXT_MAP1_SHIFT)
+#define   INTTOEXT_MAP0_SHIFT			0
+#define   INTTOEXT_MAP0_MASK			(0x1f << INTTOEXT_MAP0_SHIFT)
+#define MEMSWCTL				_MMIO(0x11170) /* Ironlake only */
+#define   MEMCTL_CMD_MASK			0xe000
+#define   MEMCTL_CMD_SHIFT			13
+#define   MEMCTL_CMD_RCLK_OFF			0
+#define   MEMCTL_CMD_RCLK_ON			1
+#define   MEMCTL_CMD_CHFREQ			2
+#define   MEMCTL_CMD_CHVID			3
+#define   MEMCTL_CMD_VMMOFF			4
+#define   MEMCTL_CMD_VMMON			5
+#define   MEMCTL_CMD_STS			(1 << 12) /* write 1 triggers command, clears
+							     when command complete */
+#define   MEMCTL_FREQ_MASK			0x0f00 /* jitter, from 0-15 */
+#define   MEMCTL_FREQ_SHIFT			8
+#define   MEMCTL_SFCAVM				(1 << 7)
+#define   MEMCTL_TGT_VID_MASK			0x007f
+#define MEMIHYST				_MMIO(0x1117c)
+#define MEMINTREN				_MMIO(0x11180) /* 16 bits */
+#define   MEMINT_RSEXIT_EN			(1 << 8)
+#define   MEMINT_CX_SUPR_EN			(1 << 7)
+#define   MEMINT_CONT_BUSY_EN			(1 << 6)
+#define   MEMINT_AVG_BUSY_EN			(1 << 5)
+#define   MEMINT_EVAL_CHG_EN			(1 << 4)
+#define   MEMINT_MON_IDLE_EN			(1 << 3)
+#define   MEMINT_UP_EVAL_EN			(1 << 2)
+#define   MEMINT_DOWN_EVAL_EN			(1 << 1)
+#define   MEMINT_SW_CMD_EN			(1 << 0)
+#define MEMINTRSTR				_MMIO(0x11182) /* 16 bits */
+#define   MEM_RSEXIT_MASK			0xc000
+#define   MEM_RSEXIT_SHIFT			14
+#define   MEM_CONT_BUSY_MASK			0x3000
+#define   MEM_CONT_BUSY_SHIFT			12
+#define   MEM_AVG_BUSY_MASK			0x0c00
+#define   MEM_AVG_BUSY_SHIFT			10
+#define   MEM_EVAL_CHG_MASK			0x0300
+#define   MEM_EVAL_BUSY_SHIFT			8
+#define   MEM_MON_IDLE_MASK			0x00c0
+#define   MEM_MON_IDLE_SHIFT			6
+#define   MEM_UP_EVAL_MASK			0x0030
+#define   MEM_UP_EVAL_SHIFT			4
+#define   MEM_DOWN_EVAL_MASK			0x000c
+#define   MEM_DOWN_EVAL_SHIFT			2
+#define   MEM_SW_CMD_MASK			0x0003
+#define   MEM_INT_STEER_GFX			0
+#define   MEM_INT_STEER_CMR			1
+#define   MEM_INT_STEER_SMI			2
+#define   MEM_INT_STEER_SCI			3
+#define MEMINTRSTS				_MMIO(0x11184)
+#define   MEMINT_RSEXIT				(1 << 7)
+#define   MEMINT_CONT_BUSY			(1 << 6)
+#define   MEMINT_AVG_BUSY			(1 << 5)
+#define   MEMINT_EVAL_CHG			(1 << 4)
+#define   MEMINT_MON_IDLE			(1 << 3)
+#define   MEMINT_UP_EVAL			(1 << 2)
+#define   MEMINT_DOWN_EVAL			(1 << 1)
+#define   MEMINT_SW_CMD				(1 << 0)
+#define MEMMODECTL				_MMIO(0x11190)
+#define   MEMMODE_BOOST_EN			(1 << 31)
+#define   MEMMODE_BOOST_FREQ_MASK		0x0f000000 /* jitter for boost, 0-15 */
+#define   MEMMODE_BOOST_FREQ_SHIFT		24
+#define   MEMMODE_IDLE_MODE_MASK		0x00030000
+#define   MEMMODE_IDLE_MODE_SHIFT		16
+#define   MEMMODE_IDLE_MODE_EVAL		0
+#define   MEMMODE_IDLE_MODE_CONT		1
+#define   MEMMODE_HWIDLE_EN			(1 << 15)
+#define   MEMMODE_SWMODE_EN			(1 << 14)
+#define   MEMMODE_RCLK_GATE			(1 << 13)
+#define   MEMMODE_HW_UPDATE			(1 << 12)
+#define   MEMMODE_FSTART_MASK			0x00000f00 /* starting jitter, 0-15 */
+#define   MEMMODE_FSTART_SHIFT			8
+#define   MEMMODE_FMAX_MASK			0x000000f0 /* max jitter, 0-15 */
+#define   MEMMODE_FMAX_SHIFT			4
+#define   MEMMODE_FMIN_MASK			0x0000000f /* min jitter, 0-15 */
+#define RCBMAXAVG				_MMIO(0x1119c)
+#define MEMSWCTL2				_MMIO(0x1119e) /* Cantiga only */
+#define   SWMEMCMD_RENDER_OFF			(0 << 13)
+#define   SWMEMCMD_RENDER_ON			(1 << 13)
+#define   SWMEMCMD_SWFREQ			(2 << 13)
+#define   SWMEMCMD_TARVID			(3 << 13)
+#define   SWMEMCMD_VRM_OFF			(4 << 13)
+#define   SWMEMCMD_VRM_ON			(5 << 13)
+#define   CMDSTS				(1 << 12)
+#define   SFCAVM				(1 << 11)
+#define   SWFREQ_MASK				0x0380 /* P0-7 */
+#define   SWFREQ_SHIFT				7
+#define   TARVID_MASK				0x001f
+#define MEMSTAT_CTG				_MMIO(0x111a0)
+#define RCBMINAVG				_MMIO(0x111a0)
+#define RCUPEI					_MMIO(0x111b0)
+#define RCDNEI					_MMIO(0x111b4)
+#define RSTDBYCTL				_MMIO(0x111b8)
+#define   RS1EN					(1 << 31)
+#define   RS2EN					(1 << 30)
+#define   RS3EN					(1 << 29)
+#define   D3RS3EN				(1 << 28) /* Display D3 imlies RS3 */
+#define   SWPROMORSX				(1 << 27) /* RSx promotion timers ignored */
+#define   RCWAKERW				(1 << 26) /* Resetwarn from PCH causes wakeup */
+#define   DPRSLPVREN				(1 << 25) /* Fast voltage ramp enable */
+#define   GFXTGHYST				(1 << 24) /* Hysteresis to allow trunk gating */
+#define   RCX_SW_EXIT				(1 << 23) /* Leave RSx and prevent re-entry */
+#define   RSX_STATUS_MASK			(7 << 20)
+#define   RSX_STATUS_ON				(0 << 20)
+#define   RSX_STATUS_RC1			(1 << 20)
+#define   RSX_STATUS_RC1E			(2 << 20)
+#define   RSX_STATUS_RS1			(3 << 20)
+#define   RSX_STATUS_RS2			(4 << 20) /* aka rc6 */
+#define   RSX_STATUS_RSVD			(5 << 20) /* deep rc6 unsupported on ilk */
+#define   RSX_STATUS_RS3			(6 << 20) /* rs3 unsupported on ilk */
+#define   RSX_STATUS_RSVD2			(7 << 20)
+#define   UWRCRSXE				(1 << 19) /* wake counter limit prevents rsx */
+#define   RSCRP					(1 << 18) /* rs requests control on rs1/2 reqs */
+#define   JRSC					(1 << 17) /* rsx coupled to cpu c-state */
+#define   RS2INC0				(1 << 16) /* allow rs2 in cpu c0 */
+#define   RS1CONTSAV_MASK			(3 << 14)
+#define   RS1CONTSAV_NO_RS1			(0 << 14) /* rs1 doesn't save/restore context */
+#define   RS1CONTSAV_RSVD			(1 << 14)
+#define   RS1CONTSAV_SAVE_RS1			(2 << 14) /* rs1 saves context */
+#define   RS1CONTSAV_FULL_RS1			(3 << 14) /* rs1 saves and restores context */
+#define   NORMSLEXLAT_MASK			(3 << 12)
+#define   SLOW_RS123				(0 << 12)
+#define   SLOW_RS23				(1 << 12)
+#define   SLOW_RS3				(2 << 12)
+#define   NORMAL_RS123				(3 << 12)
+#define   RCMODE_TIMEOUT			(1 << 11) /* 0 is eval interval method */
+#define   IMPROMOEN				(1 << 10) /* promo is immediate or delayed until next idle interval (only for timeout method above) */
+#define   RCENTSYNC				(1 << 9) /* rs coupled to cpu c-state (3/6/7) */
+#define   STATELOCK				(1 << 7) /* locked to rs_cstate if 0 */
+#define   RS_CSTATE_MASK			(3 << 4)
+#define   RS_CSTATE_C367_RS1			(0 << 4)
+#define   RS_CSTATE_C36_RS1_C7_RS2		(1 << 4)
+#define   RS_CSTATE_RSVD			(2 << 4)
+#define   RS_CSTATE_C367_RS2			(3 << 4)
+#define   REDSAVES				(1 << 3) /* no context save if was idle during rs0 */
+#define   REDRESTORES				(1 << 2) /* no restore if was idle during rs0 */
+#define VIDCTL					_MMIO(0x111c0)
+#define VIDSTS					_MMIO(0x111c8)
+#define VIDSTART				_MMIO(0x111cc) /* 8 bits */
+#define MEMSTAT_ILK				_MMIO(0x111f8)
+#define   MEMSTAT_VID_MASK			0x7f00
+#define   MEMSTAT_VID_SHIFT			8
+#define   MEMSTAT_PSTATE_MASK			0x00f8
+#define   MEMSTAT_PSTATE_SHIFT			3
+#define   MEMSTAT_MON_ACTV			(1 << 2)
+#define   MEMSTAT_SRC_CTL_MASK			0x0003
+#define   MEMSTAT_SRC_CTL_CORE			0
+#define   MEMSTAT_SRC_CTL_TRB			1
+#define   MEMSTAT_SRC_CTL_THM			2
+#define   MEMSTAT_SRC_CTL_STDBY			3
+#define PMMISC					_MMIO(0x11214)
+#define   MCPPCE_EN				(1 << 0) /* enable PM_MSG from PCH->MPC */
+#define SDEW					_MMIO(0x1124c)
+#define CSIEW0					_MMIO(0x11250)
+#define CSIEW1					_MMIO(0x11254)
+#define CSIEW2					_MMIO(0x11258)
+#define PEW(i)					_MMIO(0x1125c + (i) * 4) /* 5 registers */
+#define DEW(i)					_MMIO(0x11270 + (i) * 4) /* 3 registers */
+#define MCHAFE					_MMIO(0x112c0)
+#define CSIEC					_MMIO(0x112e0)
+#define DMIEC					_MMIO(0x112e4)
+#define DDREC					_MMIO(0x112e8)
+#define PEG0EC					_MMIO(0x112ec)
+#define PEG1EC					_MMIO(0x112f0)
+#define GFXEC					_MMIO(0x112f4)
+#define INTTOEXT_BASE_ILK			_MMIO(0x11300)
+#define RPPREVBSYTUPAVG				_MMIO(0x113b8)
+#define RCPREVBSYTUPAVG				_MMIO(0x113b8)
+#define RCPREVBSYTDNAVG				_MMIO(0x113bc)
+#define RPPREVBSYTDNAVG				_MMIO(0x113bc)
+#define ECR					_MMIO(0x11600)
+#define   ECR_GPFE				(1 << 31)
+#define   ECR_IMONE				(1 << 30)
+#define   ECR_CAP_MASK				0x0000001f /* Event range, 0-31 */
+#define OGW0					_MMIO(0x11608)
+#define OGW1					_MMIO(0x1160c)
+#define EG0					_MMIO(0x11610)
+#define EG1					_MMIO(0x11614)
+#define EG2					_MMIO(0x11618)
+#define EG3					_MMIO(0x1161c)
+#define EG4					_MMIO(0x11620)
+#define EG5					_MMIO(0x11624)
+#define EG6					_MMIO(0x11628)
+#define EG7					_MMIO(0x1162c)
+#define PXW(i)					_MMIO(0x11664 + (i) * 4) /* 4 registers */
+#define PXWL(i)					_MMIO(0x11680 + (i) * 8) /* 8 registers */
+#define LCFUSE02				_MMIO(0x116c0)
+#define   LCFUSE_HIV_MASK			0x000000ff
+
+#define GAC_ECO_BITS				_MMIO(0x14090)
+#define   ECOBITS_SNB_BIT			(1 << 13)
+#define   ECOBITS_PPGTT_CACHE64B		(3 << 8)
+#define   ECOBITS_PPGTT_CACHE4B			(0 << 8)
+
+#define GEN12_RCU_MODE				_MMIO(0x14800)
+#define   GEN12_RCU_MODE_CCS_ENABLE		REG_BIT(0)
+
+#define CHV_FUSE_GT				_MMIO(VLV_DISPLAY_BASE + 0x2168)
+#define   CHV_FGT_DISABLE_SS0			(1 << 10)
+#define   CHV_FGT_DISABLE_SS1			(1 << 11)
+#define   CHV_FGT_EU_DIS_SS0_R0_SHIFT		16
+#define   CHV_FGT_EU_DIS_SS0_R0_MASK		(0xf << CHV_FGT_EU_DIS_SS0_R0_SHIFT)
+#define   CHV_FGT_EU_DIS_SS0_R1_SHIFT		20
+#define   CHV_FGT_EU_DIS_SS0_R1_MASK		(0xf << CHV_FGT_EU_DIS_SS0_R1_SHIFT)
+#define   CHV_FGT_EU_DIS_SS1_R0_SHIFT		24
+#define   CHV_FGT_EU_DIS_SS1_R0_MASK		(0xf << CHV_FGT_EU_DIS_SS1_R0_SHIFT)
+#define   CHV_FGT_EU_DIS_SS1_R1_SHIFT		28
+#define   CHV_FGT_EU_DIS_SS1_R1_MASK		(0xf << CHV_FGT_EU_DIS_SS1_R1_SHIFT)
+
+#define BCS_SWCTRL				_MMIO(0x22200)
+#define   BCS_SRC_Y				REG_BIT(0)
+#define   BCS_DST_Y				REG_BIT(1)
+
+#define GAB_CTL					_MMIO(0x24000)
+#define   GAB_CTL_CONT_AFTER_PAGEFAULT		(1 << 8)
+
+#define GEN6_PMISR				_MMIO(0x44020)
+#define GEN6_PMIMR				_MMIO(0x44024) /* rps_lock */
+#define GEN6_PMIIR				_MMIO(0x44028)
+#define GEN6_PMIER				_MMIO(0x4402c)
+#define   GEN6_PM_MBOX_EVENT			(1 << 25)
+#define   GEN6_PM_THERMAL_EVENT			(1 << 24)
+/*
+ * For Gen11 these are in the upper word of the GPM_WGBOXPERF
+ * registers. Shifting is handled on accessing the imr and ier.
+ */
+#define   GEN6_PM_RP_DOWN_TIMEOUT		(1 << 6)
+#define   GEN6_PM_RP_UP_THRESHOLD		(1 << 5)
+#define   GEN6_PM_RP_DOWN_THRESHOLD		(1 << 4)
+#define   GEN6_PM_RP_UP_EI_EXPIRED		(1 << 2)
+#define   GEN6_PM_RP_DOWN_EI_EXPIRED		(1 << 1)
+#define   GEN6_PM_RPS_EVENTS			(GEN6_PM_RP_UP_EI_EXPIRED   | \
+						 GEN6_PM_RP_UP_THRESHOLD    | \
+						 GEN6_PM_RP_DOWN_EI_EXPIRED | \
+						 GEN6_PM_RP_DOWN_THRESHOLD  | \
+						 GEN6_PM_RP_DOWN_TIMEOUT)
+
+#define GEN7_GT_SCRATCH(i)			_MMIO(0x4f100 + (i) * 4)
+#define   GEN7_GT_SCRATCH_REG_NUM		8
+
+#define GFX_FLSH_CNTL_GEN6			_MMIO(0x101008)
+#define   GFX_FLSH_CNTL_EN			(1 << 0)
+
+#define GTFIFODBG				_MMIO(0x120000)
+#define   GT_FIFO_SBDEDICATE_FREE_ENTRY_CHV	(0x1f << 20)
+#define   GT_FIFO_FREE_ENTRIES_CHV		(0x7f << 13)
+#define   GT_FIFO_SBDROPERR			(1 << 6)
+#define   GT_FIFO_BLOBDROPERR			(1 << 5)
+#define   GT_FIFO_SB_READ_ABORTERR		(1 << 4)
+#define   GT_FIFO_DROPERR			(1 << 3)
+#define   GT_FIFO_OVFERR			(1 << 2)
+#define   GT_FIFO_IAWRERR			(1 << 1)
+#define   GT_FIFO_IARDERR			(1 << 0)
+
+#define GTFIFOCTL				_MMIO(0x120008)
+#define   GT_FIFO_FREE_ENTRIES_MASK		0x7f
+#define   GT_FIFO_NUM_RESERVED_ENTRIES		20
+#define   GT_FIFO_CTL_BLOCK_ALL_POLICY_STALL	(1 << 12)
+#define   GT_FIFO_CTL_RC6_POLICY_STALL		(1 << 11)
+
+#define FORCEWAKE_MT_ACK			_MMIO(0x130040)
+#define FORCEWAKE_ACK_HSW			_MMIO(0x130044)
+#define FORCEWAKE_ACK_GT_GEN9			_MMIO(0x130044)
+#define   FORCEWAKE_KERNEL			BIT(0)
+#define   FORCEWAKE_USER			BIT(1)
+#define   FORCEWAKE_KERNEL_FALLBACK		BIT(15)
+#define FORCEWAKE_ACK				_MMIO(0x130090)
+#define VLV_GTLC_WAKE_CTRL			_MMIO(0x130090)
+#define   VLV_GTLC_RENDER_CTX_EXISTS		(1 << 25)
+#define   VLV_GTLC_MEDIA_CTX_EXISTS		(1 << 24)
+#define   VLV_GTLC_ALLOWWAKEREQ			(1 << 0)
+#define VLV_GTLC_PW_STATUS			_MMIO(0x130094)
+#define   VLV_GTLC_ALLOWWAKEACK			(1 << 0)
+#define   VLV_GTLC_ALLOWWAKEERR			(1 << 1)
+#define   VLV_GTLC_PW_MEDIA_STATUS_MASK		(1 << 5)
+#define   VLV_GTLC_PW_RENDER_STATUS_MASK	(1 << 7)
+#define VLV_GTLC_SURVIVABILITY_REG		_MMIO(0x130098)
+#define   VLV_GFX_CLK_STATUS_BIT		(1 << 3)
+#define   VLV_GFX_CLK_FORCE_ON_BIT		(1 << 2)
+#define FORCEWAKE_VLV				_MMIO(0x1300b0)
+#define FORCEWAKE_ACK_VLV			_MMIO(0x1300b4)
+#define FORCEWAKE_MEDIA_VLV			_MMIO(0x1300b8)
+#define FORCEWAKE_ACK_MEDIA_VLV			_MMIO(0x1300bc)
+
+#define GEN6_GT_THREAD_STATUS_REG		_MMIO(0x13805c)
+#define   GEN6_GT_THREAD_STATUS_CORE_MASK	0x7
+
+#define GEN6_GT_CORE_STATUS			_MMIO(0x138060)
+#define   GEN6_CORE_CPD_STATE_MASK		(7 << 4)
+#define   GEN6_RCn_MASK				7
+#define   GEN6_RC0				0
+#define   GEN6_RC3				2
+#define   GEN6_RC6				3
+#define   GEN6_RC7				4
+
+#define GEN8_GT_SLICE_INFO			_MMIO(0x138064)
+#define   GEN8_LSLICESTAT_MASK			0x7
+
+#define GEN6_GT_GFX_RC6_LOCKED			_MMIO(0x138104)
+#define VLV_COUNTER_CONTROL			_MMIO(0x138104)
+#define   VLV_COUNT_RANGE_HIGH			(1 << 15)
+#define   VLV_MEDIA_RC0_COUNT_EN		(1 << 5)
+#define   VLV_RENDER_RC0_COUNT_EN		(1 << 4)
+#define   VLV_MEDIA_RC6_COUNT_EN		(1 << 1)
+#define   VLV_RENDER_RC6_COUNT_EN		(1 << 0)
+#define GEN6_GT_GFX_RC6				_MMIO(0x138108)
+#define VLV_GT_MEDIA_RC6			_MMIO(0x13810c)
+
+#define GEN6_GT_GFX_RC6p			_MMIO(0x13810c)
+#define GEN6_GT_GFX_RC6pp			_MMIO(0x138110)
+#define VLV_RENDER_C0_COUNT			_MMIO(0x138118)
+#define VLV_MEDIA_C0_COUNT			_MMIO(0x13811c)
+
+#define GEN11_GT_INTR_DW(x)			_MMIO(0x190018 + ((x) * 4))
+#define   GEN11_CSME				(31)
+#define   GEN11_GUNIT				(28)
+#define   GEN11_GUC				(25)
+#define   GEN11_WDPERF				(20)
+#define   GEN11_KCR				(19)
+#define   GEN11_GTPM				(16)
+#define   GEN11_BCS				(15)
+#define   XEHPC_BCS1				(14)
+#define   XEHPC_BCS2				(13)
+#define   XEHPC_BCS3				(12)
+#define   XEHPC_BCS4				(11)
+#define   XEHPC_BCS5				(10)
+#define   XEHPC_BCS6				(9)
+#define   XEHPC_BCS7				(8)
+#define   XEHPC_BCS8				(23)
+#define   GEN12_CCS3				(7)
+#define   GEN12_CCS2				(6)
+#define   GEN12_CCS1				(5)
+#define   GEN12_CCS0				(4)
+#define   GEN11_RCS0				(0)
+#define   GEN11_VECS(x)				(31 - (x))
+#define   GEN11_VCS(x)				(x)
+
+#define GEN11_RENDER_COPY_INTR_ENABLE		_MMIO(0x190030)
+#define GEN11_VCS_VECS_INTR_ENABLE		_MMIO(0x190034)
+#define GEN11_GUC_SG_INTR_ENABLE		_MMIO(0x190038)
+#define   ENGINE1_MASK				REG_GENMASK(31, 16)
+#define   ENGINE0_MASK				REG_GENMASK(15, 0)
+#define GEN11_GPM_WGBOXPERF_INTR_ENABLE		_MMIO(0x19003c)
+#define GEN11_CRYPTO_RSVD_INTR_ENABLE		_MMIO(0x190040)
+#define GEN11_GUNIT_CSME_INTR_ENABLE		_MMIO(0x190044)
+#define GEN12_CCS_RSVD_INTR_ENABLE		_MMIO(0x190048)
+
+#define GEN11_INTR_IDENTITY_REG(x)		_MMIO(0x190060 + ((x) * 4))
+#define   GEN11_INTR_DATA_VALID			(1 << 31)
+#define   GEN11_INTR_ENGINE_CLASS(x)		(((x) & GENMASK(18, 16)) >> 16)
+#define   GEN11_INTR_ENGINE_INSTANCE(x)		(((x) & GENMASK(25, 20)) >> 20)
+#define   GEN11_INTR_ENGINE_INTR(x)		((x) & 0xffff)
+/* irq instances for OTHER_CLASS */
+#define   OTHER_GUC_INSTANCE			0
+#define   OTHER_GTPM_INSTANCE			1
+#define   OTHER_KCR_INSTANCE			4
+#define   OTHER_GSC_INSTANCE			6
+#define   OTHER_MEDIA_GUC_INSTANCE		16
+#define   OTHER_MEDIA_GTPM_INSTANCE		17
+
+#define GEN11_IIR_REG_SELECTOR(x)		_MMIO(0x190070 + ((x) * 4))
+
+#define GEN11_RCS0_RSVD_INTR_MASK		_MMIO(0x190090)
+#define GEN11_BCS_RSVD_INTR_MASK		_MMIO(0x1900a0)
+#define GEN11_VCS0_VCS1_INTR_MASK		_MMIO(0x1900a8)
+#define GEN11_VCS2_VCS3_INTR_MASK		_MMIO(0x1900ac)
+#define GEN12_VCS4_VCS5_INTR_MASK		_MMIO(0x1900b0)
+#define GEN12_VCS6_VCS7_INTR_MASK		_MMIO(0x1900b4)
+#define GEN11_VECS0_VECS1_INTR_MASK		_MMIO(0x1900d0)
+#define GEN12_VECS2_VECS3_INTR_MASK		_MMIO(0x1900d4)
+#define GEN11_GUC_SG_INTR_MASK			_MMIO(0x1900e8)
+#define GEN11_GPM_WGBOXPERF_INTR_MASK		_MMIO(0x1900ec)
+#define GEN11_CRYPTO_RSVD_INTR_MASK		_MMIO(0x1900f0)
+#define GEN11_GUNIT_CSME_INTR_MASK		_MMIO(0x1900f4)
+#define GEN12_CCS0_CCS1_INTR_MASK		_MMIO(0x190100)
+#define GEN12_CCS2_CCS3_INTR_MASK		_MMIO(0x190104)
+#define XEHPC_BCS1_BCS2_INTR_MASK		_MMIO(0x190110)
+#define XEHPC_BCS3_BCS4_INTR_MASK		_MMIO(0x190114)
+#define XEHPC_BCS5_BCS6_INTR_MASK		_MMIO(0x190118)
+#define XEHPC_BCS7_BCS8_INTR_MASK		_MMIO(0x19011c)
+
+#define GEN12_SFC_DONE(n)			_MMIO(0x1cc000 + (n) * 0x1000)
+
+/*
+ * Standalone Media's non-engine GT registers are located at their regular GT
+ * offsets plus 0x380000.  This extra offset is stored inside the intel_uncore
+ * structure so that the existing code can be used for both GTs without
+ * modification.
+ */
+#define MTL_MEDIA_GSI_BASE			0x380000
+
+#endif /* __INTEL_GT_REGS__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_requests.c b/drivers/gpu/drm/i915/gt/intel_gt_requests.c
new file mode 100644
index 000000000..1dfd01668
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_requests.c
@@ -0,0 +1,262 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/workqueue.h>
+
+#include "i915_drv.h" /* for_each_engine() */
+#include "i915_request.h"
+#include "intel_engine_heartbeat.h"
+#include "intel_execlists_submission.h"
+#include "intel_gt.h"
+#include "intel_gt_pm.h"
+#include "intel_gt_requests.h"
+#include "intel_timeline.h"
+
+static bool retire_requests(struct intel_timeline *tl)
+{
+	struct i915_request *rq, *rn;
+
+	list_for_each_entry_safe(rq, rn, &tl->requests, link)
+		if (!i915_request_retire(rq))
+			return false;
+
+	/* And check nothing new was submitted */
+	return !i915_active_fence_isset(&tl->last_request);
+}
+
+static bool engine_active(const struct intel_engine_cs *engine)
+{
+	return !list_empty(&engine->kernel_context->timeline->requests);
+}
+
+static bool flush_submission(struct intel_gt *gt, long timeout)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	bool active = false;
+
+	if (!timeout)
+		return false;
+
+	if (!intel_gt_pm_is_awake(gt))
+		return false;
+
+	for_each_engine(engine, gt, id) {
+		intel_engine_flush_submission(engine);
+
+		/* Flush the background retirement and idle barriers */
+		flush_work(&engine->retire_work);
+		flush_delayed_work(&engine->wakeref.work);
+
+		/* Is the idle barrier still outstanding? */
+		active |= engine_active(engine);
+	}
+
+	return active;
+}
+
+static void engine_retire(struct work_struct *work)
+{
+	struct intel_engine_cs *engine =
+		container_of(work, typeof(*engine), retire_work);
+	struct intel_timeline *tl = xchg(&engine->retire, NULL);
+
+	do {
+		struct intel_timeline *next = xchg(&tl->retire, NULL);
+
+		/*
+		 * Our goal here is to retire _idle_ timelines as soon as
+		 * possible (as they are idle, we do not expect userspace
+		 * to be cleaning up anytime soon).
+		 *
+		 * If the timeline is currently locked, either it is being
+		 * retired elsewhere or about to be!
+		 */
+		if (mutex_trylock(&tl->mutex)) {
+			retire_requests(tl);
+			mutex_unlock(&tl->mutex);
+		}
+		intel_timeline_put(tl);
+
+		GEM_BUG_ON(!next);
+		tl = ptr_mask_bits(next, 1);
+	} while (tl);
+}
+
+static bool add_retire(struct intel_engine_cs *engine,
+		       struct intel_timeline *tl)
+{
+#define STUB ((struct intel_timeline *)1)
+	struct intel_timeline *first;
+
+	/*
+	 * We open-code a llist here to include the additional tag [BIT(0)]
+	 * so that we know when the timeline is already on a
+	 * retirement queue: either this engine or another.
+	 */
+
+	if (cmpxchg(&tl->retire, NULL, STUB)) /* already queued */
+		return false;
+
+	intel_timeline_get(tl);
+	first = READ_ONCE(engine->retire);
+	do
+		tl->retire = ptr_pack_bits(first, 1, 1);
+	while (!try_cmpxchg(&engine->retire, &first, tl));
+
+	return !first;
+}
+
+void intel_engine_add_retire(struct intel_engine_cs *engine,
+			     struct intel_timeline *tl)
+{
+	/* We don't deal well with the engine disappearing beneath us */
+	GEM_BUG_ON(intel_engine_is_virtual(engine));
+
+	if (add_retire(engine, tl))
+		schedule_work(&engine->retire_work);
+}
+
+void intel_engine_init_retire(struct intel_engine_cs *engine)
+{
+	INIT_WORK(&engine->retire_work, engine_retire);
+}
+
+void intel_engine_fini_retire(struct intel_engine_cs *engine)
+{
+	flush_work(&engine->retire_work);
+	GEM_BUG_ON(engine->retire);
+}
+
+long intel_gt_retire_requests_timeout(struct intel_gt *gt, long timeout,
+				      long *remaining_timeout)
+{
+	struct intel_gt_timelines *timelines = &gt->timelines;
+	struct intel_timeline *tl, *tn;
+	unsigned long active_count = 0;
+	LIST_HEAD(free);
+
+	flush_submission(gt, timeout); /* kick the ksoftirqd tasklets */
+	spin_lock(&timelines->lock);
+	list_for_each_entry_safe(tl, tn, &timelines->active_list, link) {
+		if (!mutex_trylock(&tl->mutex)) {
+			active_count++; /* report busy to caller, try again? */
+			continue;
+		}
+
+		intel_timeline_get(tl);
+		GEM_BUG_ON(!atomic_read(&tl->active_count));
+		atomic_inc(&tl->active_count); /* pin the list element */
+		spin_unlock(&timelines->lock);
+
+		if (timeout > 0) {
+			struct dma_fence *fence;
+
+			fence = i915_active_fence_get(&tl->last_request);
+			if (fence) {
+				mutex_unlock(&tl->mutex);
+
+				timeout = dma_fence_wait_timeout(fence,
+								 true,
+								 timeout);
+				dma_fence_put(fence);
+
+				/* Retirement is best effort */
+				if (!mutex_trylock(&tl->mutex)) {
+					active_count++;
+					goto out_active;
+				}
+			}
+		}
+
+		if (!retire_requests(tl))
+			active_count++;
+		mutex_unlock(&tl->mutex);
+
+out_active:	spin_lock(&timelines->lock);
+
+		/* Resume list iteration after reacquiring spinlock */
+		list_safe_reset_next(tl, tn, link);
+		if (atomic_dec_and_test(&tl->active_count))
+			list_del(&tl->link);
+
+		/* Defer the final release to after the spinlock */
+		if (refcount_dec_and_test(&tl->kref.refcount)) {
+			GEM_BUG_ON(atomic_read(&tl->active_count));
+			list_add(&tl->link, &free);
+		}
+	}
+	spin_unlock(&timelines->lock);
+
+	list_for_each_entry_safe(tl, tn, &free, link)
+		__intel_timeline_free(&tl->kref);
+
+	if (flush_submission(gt, timeout)) /* Wait, there's more! */
+		active_count++;
+
+	if (remaining_timeout)
+		*remaining_timeout = timeout;
+
+	return active_count ? timeout ?: -ETIME : 0;
+}
+
+static void retire_work_handler(struct work_struct *work)
+{
+	struct intel_gt *gt =
+		container_of(work, typeof(*gt), requests.retire_work.work);
+
+	schedule_delayed_work(&gt->requests.retire_work,
+			      round_jiffies_up_relative(HZ));
+	intel_gt_retire_requests(gt);
+}
+
+void intel_gt_init_requests(struct intel_gt *gt)
+{
+	INIT_DELAYED_WORK(&gt->requests.retire_work, retire_work_handler);
+}
+
+void intel_gt_park_requests(struct intel_gt *gt)
+{
+	cancel_delayed_work(&gt->requests.retire_work);
+}
+
+void intel_gt_unpark_requests(struct intel_gt *gt)
+{
+	schedule_delayed_work(&gt->requests.retire_work,
+			      round_jiffies_up_relative(HZ));
+}
+
+void intel_gt_fini_requests(struct intel_gt *gt)
+{
+	/* Wait until the work is marked as finished before unloading! */
+	cancel_delayed_work_sync(&gt->requests.retire_work);
+
+	flush_work(&gt->watchdog.work);
+}
+
+void intel_gt_watchdog_work(struct work_struct *work)
+{
+	struct intel_gt *gt =
+		container_of(work, typeof(*gt), watchdog.work);
+	struct i915_request *rq, *rn;
+	struct llist_node *first;
+
+	first = llist_del_all(&gt->watchdog.list);
+	if (!first)
+		return;
+
+	llist_for_each_entry_safe(rq, rn, first, watchdog.link) {
+		if (!i915_request_completed(rq)) {
+			struct dma_fence *f = &rq->fence;
+
+			pr_notice("Fence expiration time out i915-%s:%s:%llx!\n",
+				  f->ops->get_driver_name(f),
+				  f->ops->get_timeline_name(f),
+				  f->seqno);
+			i915_request_cancel(rq, -EINTR);
+		}
+		i915_request_put(rq);
+	}
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_requests.h b/drivers/gpu/drm/i915/gt/intel_gt_requests.h
new file mode 100644
index 000000000..d2969f68d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_requests.h
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_GT_REQUESTS_H
+#define INTEL_GT_REQUESTS_H
+
+#include <linux/stddef.h>
+
+struct intel_engine_cs;
+struct intel_gt;
+struct intel_timeline;
+
+long intel_gt_retire_requests_timeout(struct intel_gt *gt, long timeout,
+				      long *remaining_timeout);
+static inline void intel_gt_retire_requests(struct intel_gt *gt)
+{
+	intel_gt_retire_requests_timeout(gt, 0, NULL);
+}
+
+void intel_engine_init_retire(struct intel_engine_cs *engine);
+void intel_engine_add_retire(struct intel_engine_cs *engine,
+			     struct intel_timeline *tl);
+void intel_engine_fini_retire(struct intel_engine_cs *engine);
+
+void intel_gt_init_requests(struct intel_gt *gt);
+void intel_gt_park_requests(struct intel_gt *gt);
+void intel_gt_unpark_requests(struct intel_gt *gt);
+void intel_gt_fini_requests(struct intel_gt *gt);
+
+#endif /* INTEL_GT_REQUESTS_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_sysfs.c b/drivers/gpu/drm/i915/gt/intel_gt_sysfs.c
new file mode 100644
index 000000000..9486dd3be
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_sysfs.c
@@ -0,0 +1,116 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include <drm/drm_device.h>
+#include <linux/device.h>
+#include <linux/kobject.h>
+#include <linux/printk.h>
+#include <linux/sysfs.h>
+
+#include "i915_drv.h"
+#include "i915_sysfs.h"
+#include "intel_gt.h"
+#include "intel_gt_sysfs.h"
+#include "intel_gt_sysfs_pm.h"
+#include "intel_gt_types.h"
+#include "intel_rc6.h"
+
+bool is_object_gt(struct kobject *kobj)
+{
+	return !strncmp(kobj->name, "gt", 2);
+}
+
+struct intel_gt *intel_gt_sysfs_get_drvdata(struct kobject *kobj,
+					    const char *name)
+{
+	/*
+	 * We are interested at knowing from where the interface
+	 * has been called, whether it's called from gt/ or from
+	 * the parent directory.
+	 * From the interface position it depends also the value of
+	 * the private data.
+	 * If the interface is called from gt/ then private data is
+	 * of the "struct intel_gt *" type, otherwise it's * a
+	 * "struct drm_i915_private *" type.
+	 */
+	if (!is_object_gt(kobj)) {
+		struct device *dev = kobj_to_dev(kobj);
+		struct drm_i915_private *i915 = kdev_minor_to_i915(dev);
+
+		return to_gt(i915);
+	}
+
+	return kobj_to_gt(kobj);
+}
+
+static struct kobject *gt_get_parent_obj(struct intel_gt *gt)
+{
+	return &gt->i915->drm.primary->kdev->kobj;
+}
+
+static ssize_t id_show(struct kobject *kobj,
+		       struct kobj_attribute *attr,
+		       char *buf)
+{
+	struct intel_gt *gt = intel_gt_sysfs_get_drvdata(kobj, attr->attr.name);
+
+	return sysfs_emit(buf, "%u\n", gt->info.id);
+}
+static struct kobj_attribute attr_id = __ATTR_RO(id);
+
+static struct attribute *id_attrs[] = {
+	&attr_id.attr,
+	NULL,
+};
+ATTRIBUTE_GROUPS(id);
+
+/* A kobject needs a release() method even if it does nothing */
+static void kobj_gt_release(struct kobject *kobj)
+{
+}
+
+static struct kobj_type kobj_gt_type = {
+	.release = kobj_gt_release,
+	.sysfs_ops = &kobj_sysfs_ops,
+	.default_groups = id_groups,
+};
+
+void intel_gt_sysfs_register(struct intel_gt *gt)
+{
+	/*
+	 * We need to make things right with the
+	 * ABI compatibility. The files were originally
+	 * generated under the parent directory.
+	 *
+	 * We generate the files only for gt 0
+	 * to avoid duplicates.
+	 */
+	if (gt_is_root(gt))
+		intel_gt_sysfs_pm_init(gt, gt_get_parent_obj(gt));
+
+	/* init and xfer ownership to sysfs tree */
+	if (kobject_init_and_add(&gt->sysfs_gt, &kobj_gt_type,
+				 gt->i915->sysfs_gt, "gt%d", gt->info.id))
+		goto exit_fail;
+
+	gt->sysfs_defaults = kobject_create_and_add(".defaults", &gt->sysfs_gt);
+	if (!gt->sysfs_defaults)
+		goto exit_fail;
+
+	intel_gt_sysfs_pm_init(gt, &gt->sysfs_gt);
+
+	return;
+
+exit_fail:
+	kobject_put(&gt->sysfs_gt);
+	drm_warn(&gt->i915->drm,
+		 "failed to initialize gt%d sysfs root\n", gt->info.id);
+}
+
+void intel_gt_sysfs_unregister(struct intel_gt *gt)
+{
+	kobject_put(gt->sysfs_defaults);
+	kobject_put(&gt->sysfs_gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_sysfs.h b/drivers/gpu/drm/i915/gt/intel_gt_sysfs.h
new file mode 100644
index 000000000..c3a123fae
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_sysfs.h
@@ -0,0 +1,36 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __SYSFS_GT_H__
+#define __SYSFS_GT_H__
+
+#include <linux/ctype.h>
+#include <linux/kobject.h>
+
+#include "i915_gem.h" /* GEM_BUG_ON() */
+#include "intel_gt_types.h"
+
+struct intel_gt;
+
+bool is_object_gt(struct kobject *kobj);
+
+struct drm_i915_private *kobj_to_i915(struct kobject *kobj);
+
+struct kobject *
+intel_gt_create_kobj(struct intel_gt *gt,
+		     struct kobject *dir,
+		     const char *name);
+
+static inline struct intel_gt *kobj_to_gt(struct kobject *kobj)
+{
+	return container_of(kobj, struct intel_gt, sysfs_gt);
+}
+
+void intel_gt_sysfs_register(struct intel_gt *gt);
+void intel_gt_sysfs_unregister(struct intel_gt *gt);
+struct intel_gt *intel_gt_sysfs_get_drvdata(struct kobject *kobj,
+					    const char *name);
+
+#endif /* SYSFS_GT_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_sysfs_pm.c b/drivers/gpu/drm/i915/gt/intel_gt_sysfs_pm.c
new file mode 100644
index 000000000..b108f0a8a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_sysfs_pm.c
@@ -0,0 +1,783 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include <drm/drm_device.h>
+#include <linux/sysfs.h>
+#include <linux/printk.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "i915_sysfs.h"
+#include "intel_gt.h"
+#include "intel_gt_regs.h"
+#include "intel_gt_sysfs.h"
+#include "intel_gt_sysfs_pm.h"
+#include "intel_pcode.h"
+#include "intel_rc6.h"
+#include "intel_rps.h"
+
+enum intel_gt_sysfs_op {
+	INTEL_GT_SYSFS_MIN = 0,
+	INTEL_GT_SYSFS_MAX,
+};
+
+static int
+sysfs_gt_attribute_w_func(struct kobject *kobj, struct attribute *attr,
+			  int (func)(struct intel_gt *gt, u32 val), u32 val)
+{
+	struct intel_gt *gt;
+	int ret;
+
+	if (!is_object_gt(kobj)) {
+		int i;
+		struct device *dev = kobj_to_dev(kobj);
+		struct drm_i915_private *i915 = kdev_minor_to_i915(dev);
+
+		for_each_gt(gt, i915, i) {
+			ret = func(gt, val);
+			if (ret)
+				break;
+		}
+	} else {
+		gt = intel_gt_sysfs_get_drvdata(kobj, attr->name);
+		ret = func(gt, val);
+	}
+
+	return ret;
+}
+
+static u32
+sysfs_gt_attribute_r_func(struct kobject *kobj, struct attribute *attr,
+			  u32 (func)(struct intel_gt *gt),
+			  enum intel_gt_sysfs_op op)
+{
+	struct intel_gt *gt;
+	u32 ret;
+
+	ret = (op == INTEL_GT_SYSFS_MAX) ? 0 : (u32) -1;
+
+	if (!is_object_gt(kobj)) {
+		int i;
+		struct device *dev = kobj_to_dev(kobj);
+		struct drm_i915_private *i915 = kdev_minor_to_i915(dev);
+
+		for_each_gt(gt, i915, i) {
+			u32 val = func(gt);
+
+			switch (op) {
+			case INTEL_GT_SYSFS_MIN:
+				if (val < ret)
+					ret = val;
+				break;
+
+			case INTEL_GT_SYSFS_MAX:
+				if (val > ret)
+					ret = val;
+				break;
+			}
+		}
+	} else {
+		gt = intel_gt_sysfs_get_drvdata(kobj, attr->name);
+		ret = func(gt);
+	}
+
+	return ret;
+}
+
+/* RC6 interfaces will show the minimum RC6 residency value */
+#define sysfs_gt_attribute_r_min_func(d, a, f) \
+		sysfs_gt_attribute_r_func(d, a, f, INTEL_GT_SYSFS_MIN)
+
+/* Frequency interfaces will show the maximum frequency value */
+#define sysfs_gt_attribute_r_max_func(d, a, f) \
+		sysfs_gt_attribute_r_func(d, a, f, INTEL_GT_SYSFS_MAX)
+
+#define INTEL_GT_SYSFS_SHOW(_name, _attr_type)							\
+	static ssize_t _name##_show_common(struct kobject *kobj,				\
+					   struct attribute *attr, char *buff)			\
+	{											\
+		u32 val = sysfs_gt_attribute_r_##_attr_type##_func(kobj, attr,			\
+								   __##_name##_show);		\
+												\
+		return sysfs_emit(buff, "%u\n", val);						\
+	}											\
+	static ssize_t _name##_show(struct kobject *kobj,					\
+				    struct kobj_attribute *attr, char *buff)			\
+	{											\
+		return _name ##_show_common(kobj, &attr->attr, buff);				\
+	}											\
+	static ssize_t _name##_dev_show(struct device *dev,					\
+					struct device_attribute *attr, char *buff)		\
+	{											\
+		return _name##_show_common(&dev->kobj, &attr->attr, buff);			\
+	}
+
+#define INTEL_GT_SYSFS_STORE(_name, _func)						\
+	static ssize_t _name##_store_common(struct kobject *kobj,			\
+					    struct attribute *attr,			\
+					    const char *buff, size_t count)		\
+	{										\
+		int ret;								\
+		u32 val;								\
+											\
+		ret = kstrtou32(buff, 0, &val);						\
+		if (ret)								\
+			return ret;							\
+											\
+		ret = sysfs_gt_attribute_w_func(kobj, attr, _func, val);		\
+											\
+		return ret ?: count;							\
+	}										\
+	static ssize_t _name##_store(struct kobject *kobj,				\
+				     struct kobj_attribute *attr, const char *buff,	\
+				     size_t count)					\
+	{										\
+		return _name##_store_common(kobj, &attr->attr, buff, count);		\
+	}										\
+	static ssize_t _name##_dev_store(struct device *dev,				\
+					 struct device_attribute *attr,			\
+					 const char *buff, size_t count)		\
+	{										\
+		return _name##_store_common(&dev->kobj, &attr->attr, buff, count);	\
+	}
+
+#define INTEL_GT_SYSFS_SHOW_MAX(_name) INTEL_GT_SYSFS_SHOW(_name, max)
+#define INTEL_GT_SYSFS_SHOW_MIN(_name) INTEL_GT_SYSFS_SHOW(_name, min)
+
+#define INTEL_GT_ATTR_RW(_name) \
+	static struct kobj_attribute attr_##_name = __ATTR_RW(_name)
+
+#define INTEL_GT_ATTR_RO(_name) \
+	static struct kobj_attribute attr_##_name = __ATTR_RO(_name)
+
+#define INTEL_GT_DUAL_ATTR_RW(_name) \
+	static struct device_attribute dev_attr_##_name = __ATTR(_name, 0644,		\
+								 _name##_dev_show,	\
+								 _name##_dev_store);	\
+	INTEL_GT_ATTR_RW(_name)
+
+#define INTEL_GT_DUAL_ATTR_RO(_name) \
+	static struct device_attribute dev_attr_##_name = __ATTR(_name, 0444,		\
+								 _name##_dev_show,	\
+								 NULL);			\
+	INTEL_GT_ATTR_RO(_name)
+
+#ifdef CONFIG_PM
+static u32 get_residency(struct intel_gt *gt, i915_reg_t reg)
+{
+	intel_wakeref_t wakeref;
+	u64 res = 0;
+
+	with_intel_runtime_pm(gt->uncore->rpm, wakeref)
+		res = intel_rc6_residency_us(&gt->rc6, reg);
+
+	return DIV_ROUND_CLOSEST_ULL(res, 1000);
+}
+
+static u8 get_rc6_mask(struct intel_gt *gt)
+{
+	u8 mask = 0;
+
+	if (HAS_RC6(gt->i915))
+		mask |= BIT(0);
+	if (HAS_RC6p(gt->i915))
+		mask |= BIT(1);
+	if (HAS_RC6pp(gt->i915))
+		mask |= BIT(2);
+
+	return mask;
+}
+
+static ssize_t rc6_enable_show(struct kobject *kobj,
+			       struct kobj_attribute *attr,
+			       char *buff)
+{
+	struct intel_gt *gt = intel_gt_sysfs_get_drvdata(kobj, attr->attr.name);
+
+	return sysfs_emit(buff, "%x\n", get_rc6_mask(gt));
+}
+
+static ssize_t rc6_enable_dev_show(struct device *dev,
+				   struct device_attribute *attr,
+				   char *buff)
+{
+	struct intel_gt *gt = intel_gt_sysfs_get_drvdata(&dev->kobj, attr->attr.name);
+
+	return sysfs_emit(buff, "%x\n", get_rc6_mask(gt));
+}
+
+static u32 __rc6_residency_ms_show(struct intel_gt *gt)
+{
+	return get_residency(gt, GEN6_GT_GFX_RC6);
+}
+
+static u32 __rc6p_residency_ms_show(struct intel_gt *gt)
+{
+	return get_residency(gt, GEN6_GT_GFX_RC6p);
+}
+
+static u32 __rc6pp_residency_ms_show(struct intel_gt *gt)
+{
+	return get_residency(gt, GEN6_GT_GFX_RC6pp);
+}
+
+static u32 __media_rc6_residency_ms_show(struct intel_gt *gt)
+{
+	return get_residency(gt, VLV_GT_MEDIA_RC6);
+}
+
+INTEL_GT_SYSFS_SHOW_MIN(rc6_residency_ms);
+INTEL_GT_SYSFS_SHOW_MIN(rc6p_residency_ms);
+INTEL_GT_SYSFS_SHOW_MIN(rc6pp_residency_ms);
+INTEL_GT_SYSFS_SHOW_MIN(media_rc6_residency_ms);
+
+INTEL_GT_DUAL_ATTR_RO(rc6_enable);
+INTEL_GT_DUAL_ATTR_RO(rc6_residency_ms);
+INTEL_GT_DUAL_ATTR_RO(rc6p_residency_ms);
+INTEL_GT_DUAL_ATTR_RO(rc6pp_residency_ms);
+INTEL_GT_DUAL_ATTR_RO(media_rc6_residency_ms);
+
+static struct attribute *rc6_attrs[] = {
+	&attr_rc6_enable.attr,
+	&attr_rc6_residency_ms.attr,
+	NULL
+};
+
+static struct attribute *rc6p_attrs[] = {
+	&attr_rc6p_residency_ms.attr,
+	&attr_rc6pp_residency_ms.attr,
+	NULL
+};
+
+static struct attribute *media_rc6_attrs[] = {
+	&attr_media_rc6_residency_ms.attr,
+	NULL
+};
+
+static struct attribute *rc6_dev_attrs[] = {
+	&dev_attr_rc6_enable.attr,
+	&dev_attr_rc6_residency_ms.attr,
+	NULL
+};
+
+static struct attribute *rc6p_dev_attrs[] = {
+	&dev_attr_rc6p_residency_ms.attr,
+	&dev_attr_rc6pp_residency_ms.attr,
+	NULL
+};
+
+static struct attribute *media_rc6_dev_attrs[] = {
+	&dev_attr_media_rc6_residency_ms.attr,
+	NULL
+};
+
+static const struct attribute_group rc6_attr_group[] = {
+	{ .attrs = rc6_attrs, },
+	{ .name = power_group_name, .attrs = rc6_dev_attrs, },
+};
+
+static const struct attribute_group rc6p_attr_group[] = {
+	{ .attrs = rc6p_attrs, },
+	{ .name = power_group_name, .attrs = rc6p_dev_attrs, },
+};
+
+static const struct attribute_group media_rc6_attr_group[] = {
+	{ .attrs = media_rc6_attrs, },
+	{ .name = power_group_name, .attrs = media_rc6_dev_attrs, },
+};
+
+static int __intel_gt_sysfs_create_group(struct kobject *kobj,
+					 const struct attribute_group *grp)
+{
+	return is_object_gt(kobj) ?
+	       sysfs_create_group(kobj, &grp[0]) :
+	       sysfs_merge_group(kobj, &grp[1]);
+}
+
+static void intel_sysfs_rc6_init(struct intel_gt *gt, struct kobject *kobj)
+{
+	int ret;
+
+	if (!HAS_RC6(gt->i915))
+		return;
+
+	ret = __intel_gt_sysfs_create_group(kobj, rc6_attr_group);
+	if (ret)
+		drm_warn(&gt->i915->drm,
+			 "failed to create gt%u RC6 sysfs files (%pe)\n",
+			 gt->info.id, ERR_PTR(ret));
+
+	/*
+	 * cannot use the is_visible() attribute because
+	 * the upper object inherits from the parent group.
+	 */
+	if (HAS_RC6p(gt->i915)) {
+		ret = __intel_gt_sysfs_create_group(kobj, rc6p_attr_group);
+		if (ret)
+			drm_warn(&gt->i915->drm,
+				 "failed to create gt%u RC6p sysfs files (%pe)\n",
+				 gt->info.id, ERR_PTR(ret));
+	}
+
+	if (IS_VALLEYVIEW(gt->i915) || IS_CHERRYVIEW(gt->i915)) {
+		ret = __intel_gt_sysfs_create_group(kobj, media_rc6_attr_group);
+		if (ret)
+			drm_warn(&gt->i915->drm,
+				 "failed to create media %u RC6 sysfs files (%pe)\n",
+				 gt->info.id, ERR_PTR(ret));
+	}
+}
+#else
+static void intel_sysfs_rc6_init(struct intel_gt *gt, struct kobject *kobj)
+{
+}
+#endif /* CONFIG_PM */
+
+static u32 __act_freq_mhz_show(struct intel_gt *gt)
+{
+	return intel_rps_read_actual_frequency(&gt->rps);
+}
+
+static u32 __cur_freq_mhz_show(struct intel_gt *gt)
+{
+	return intel_rps_get_requested_frequency(&gt->rps);
+}
+
+static u32 __boost_freq_mhz_show(struct intel_gt *gt)
+{
+	return intel_rps_get_boost_frequency(&gt->rps);
+}
+
+static int __boost_freq_mhz_store(struct intel_gt *gt, u32 val)
+{
+	return intel_rps_set_boost_frequency(&gt->rps, val);
+}
+
+static u32 __RP0_freq_mhz_show(struct intel_gt *gt)
+{
+	return intel_rps_get_rp0_frequency(&gt->rps);
+}
+
+static u32 __RPn_freq_mhz_show(struct intel_gt *gt)
+{
+	return intel_rps_get_rpn_frequency(&gt->rps);
+}
+
+static u32 __RP1_freq_mhz_show(struct intel_gt *gt)
+{
+	return intel_rps_get_rp1_frequency(&gt->rps);
+}
+
+static u32 __max_freq_mhz_show(struct intel_gt *gt)
+{
+	return intel_rps_get_max_frequency(&gt->rps);
+}
+
+static int __set_max_freq(struct intel_gt *gt, u32 val)
+{
+	return intel_rps_set_max_frequency(&gt->rps, val);
+}
+
+static u32 __min_freq_mhz_show(struct intel_gt *gt)
+{
+	return intel_rps_get_min_frequency(&gt->rps);
+}
+
+static int __set_min_freq(struct intel_gt *gt, u32 val)
+{
+	return intel_rps_set_min_frequency(&gt->rps, val);
+}
+
+static u32 __vlv_rpe_freq_mhz_show(struct intel_gt *gt)
+{
+	struct intel_rps *rps = &gt->rps;
+
+	return intel_gpu_freq(rps, rps->efficient_freq);
+}
+
+INTEL_GT_SYSFS_SHOW_MAX(act_freq_mhz);
+INTEL_GT_SYSFS_SHOW_MAX(boost_freq_mhz);
+INTEL_GT_SYSFS_SHOW_MAX(cur_freq_mhz);
+INTEL_GT_SYSFS_SHOW_MAX(RP0_freq_mhz);
+INTEL_GT_SYSFS_SHOW_MAX(RP1_freq_mhz);
+INTEL_GT_SYSFS_SHOW_MAX(RPn_freq_mhz);
+INTEL_GT_SYSFS_SHOW_MAX(max_freq_mhz);
+INTEL_GT_SYSFS_SHOW_MIN(min_freq_mhz);
+INTEL_GT_SYSFS_SHOW_MAX(vlv_rpe_freq_mhz);
+INTEL_GT_SYSFS_STORE(boost_freq_mhz, __boost_freq_mhz_store);
+INTEL_GT_SYSFS_STORE(max_freq_mhz, __set_max_freq);
+INTEL_GT_SYSFS_STORE(min_freq_mhz, __set_min_freq);
+
+#define INTEL_GT_RPS_SYSFS_ATTR(_name, _mode, _show, _store, _show_dev, _store_dev)		\
+	static struct device_attribute dev_attr_gt_##_name = __ATTR(gt_##_name, _mode,		\
+								    _show_dev, _store_dev);	\
+	static struct kobj_attribute attr_rps_##_name = __ATTR(rps_##_name, _mode,		\
+							       _show, _store)
+
+#define INTEL_GT_RPS_SYSFS_ATTR_RO(_name)						\
+		INTEL_GT_RPS_SYSFS_ATTR(_name, 0444, _name##_show, NULL,		\
+					_name##_dev_show, NULL)
+#define INTEL_GT_RPS_SYSFS_ATTR_RW(_name)						\
+		INTEL_GT_RPS_SYSFS_ATTR(_name, 0644, _name##_show, _name##_store,	\
+					_name##_dev_show, _name##_dev_store)
+
+/* The below macros generate static structures */
+INTEL_GT_RPS_SYSFS_ATTR_RO(act_freq_mhz);
+INTEL_GT_RPS_SYSFS_ATTR_RO(cur_freq_mhz);
+INTEL_GT_RPS_SYSFS_ATTR_RW(boost_freq_mhz);
+INTEL_GT_RPS_SYSFS_ATTR_RO(RP0_freq_mhz);
+INTEL_GT_RPS_SYSFS_ATTR_RO(RP1_freq_mhz);
+INTEL_GT_RPS_SYSFS_ATTR_RO(RPn_freq_mhz);
+INTEL_GT_RPS_SYSFS_ATTR_RW(max_freq_mhz);
+INTEL_GT_RPS_SYSFS_ATTR_RW(min_freq_mhz);
+INTEL_GT_RPS_SYSFS_ATTR_RO(vlv_rpe_freq_mhz);
+
+#define GEN6_ATTR(p, s) { \
+		&p##attr_##s##_act_freq_mhz.attr, \
+		&p##attr_##s##_cur_freq_mhz.attr, \
+		&p##attr_##s##_boost_freq_mhz.attr, \
+		&p##attr_##s##_max_freq_mhz.attr, \
+		&p##attr_##s##_min_freq_mhz.attr, \
+		&p##attr_##s##_RP0_freq_mhz.attr, \
+		&p##attr_##s##_RP1_freq_mhz.attr, \
+		&p##attr_##s##_RPn_freq_mhz.attr, \
+		NULL, \
+	}
+
+#define GEN6_RPS_ATTR GEN6_ATTR(, rps)
+#define GEN6_GT_ATTR  GEN6_ATTR(dev_, gt)
+
+static const struct attribute * const gen6_rps_attrs[] = GEN6_RPS_ATTR;
+static const struct attribute * const gen6_gt_attrs[]  = GEN6_GT_ATTR;
+
+static ssize_t punit_req_freq_mhz_show(struct kobject *kobj,
+				       struct kobj_attribute *attr,
+				       char *buff)
+{
+	struct intel_gt *gt = intel_gt_sysfs_get_drvdata(kobj, attr->attr.name);
+	u32 preq = intel_rps_read_punit_req_frequency(&gt->rps);
+
+	return sysfs_emit(buff, "%u\n", preq);
+}
+
+struct intel_gt_bool_throttle_attr {
+	struct attribute attr;
+	ssize_t (*show)(struct kobject *kobj, struct kobj_attribute *attr,
+			char *buf);
+	i915_reg_t reg32;
+	u32 mask;
+};
+
+static ssize_t throttle_reason_bool_show(struct kobject *kobj,
+					 struct kobj_attribute *attr,
+					 char *buff)
+{
+	struct intel_gt *gt = intel_gt_sysfs_get_drvdata(kobj, attr->attr.name);
+	struct intel_gt_bool_throttle_attr *t_attr =
+				(struct intel_gt_bool_throttle_attr *) attr;
+	bool val = rps_read_mask_mmio(&gt->rps, t_attr->reg32, t_attr->mask);
+
+	return sysfs_emit(buff, "%u\n", val);
+}
+
+#define INTEL_GT_RPS_BOOL_ATTR_RO(sysfs_func__, mask__) \
+struct intel_gt_bool_throttle_attr attr_##sysfs_func__ = { \
+	.attr = { .name = __stringify(sysfs_func__), .mode = 0444 }, \
+	.show = throttle_reason_bool_show, \
+	.reg32 = GT0_PERF_LIMIT_REASONS, \
+	.mask = mask__, \
+}
+
+INTEL_GT_ATTR_RO(punit_req_freq_mhz);
+static INTEL_GT_RPS_BOOL_ATTR_RO(throttle_reason_status, GT0_PERF_LIMIT_REASONS_MASK);
+static INTEL_GT_RPS_BOOL_ATTR_RO(throttle_reason_pl1, POWER_LIMIT_1_MASK);
+static INTEL_GT_RPS_BOOL_ATTR_RO(throttle_reason_pl2, POWER_LIMIT_2_MASK);
+static INTEL_GT_RPS_BOOL_ATTR_RO(throttle_reason_pl4, POWER_LIMIT_4_MASK);
+static INTEL_GT_RPS_BOOL_ATTR_RO(throttle_reason_thermal, THERMAL_LIMIT_MASK);
+static INTEL_GT_RPS_BOOL_ATTR_RO(throttle_reason_prochot, PROCHOT_MASK);
+static INTEL_GT_RPS_BOOL_ATTR_RO(throttle_reason_ratl, RATL_MASK);
+static INTEL_GT_RPS_BOOL_ATTR_RO(throttle_reason_vr_thermalert, VR_THERMALERT_MASK);
+static INTEL_GT_RPS_BOOL_ATTR_RO(throttle_reason_vr_tdc, VR_TDC_MASK);
+
+static const struct attribute *throttle_reason_attrs[] = {
+	&attr_throttle_reason_status.attr,
+	&attr_throttle_reason_pl1.attr,
+	&attr_throttle_reason_pl2.attr,
+	&attr_throttle_reason_pl4.attr,
+	&attr_throttle_reason_thermal.attr,
+	&attr_throttle_reason_prochot.attr,
+	&attr_throttle_reason_ratl.attr,
+	&attr_throttle_reason_vr_thermalert.attr,
+	&attr_throttle_reason_vr_tdc.attr,
+	NULL
+};
+
+/*
+ * Scaling for multipliers (aka frequency factors).
+ * The format of the value in the register is u8.8.
+ *
+ * The presentation to userspace is inspired by the perf event framework.
+ * See:
+ *   Documentation/ABI/testing/sysfs-bus-event_source-devices-events
+ * for description of:
+ *   /sys/bus/event_source/devices/<pmu>/events/<event>.scale
+ *
+ * Summary: Expose two sysfs files for each multiplier.
+ *
+ * 1. File <attr> contains a raw hardware value.
+ * 2. File <attr>.scale contains the multiplicative scale factor to be
+ *    used by userspace to compute the actual value.
+ *
+ * So userspace knows that to get the frequency_factor it multiplies the
+ * provided value by the specified scale factor and vice-versa.
+ *
+ * That way there is no precision loss in the kernel interface and API
+ * is future proof should one day the hardware register change to u16.u16,
+ * on some platform. (Or any other fixed point representation.)
+ *
+ * Example:
+ * File <attr> contains the value 2.5, represented as u8.8 0x0280, which
+ * is comprised of:
+ * - an integer part of 2
+ * - a fractional part of 0x80 (representing 0x80 / 2^8 == 0x80 / 256).
+ * File <attr>.scale contains a string representation of floating point
+ * value 0.00390625 (which is (1 / 256)).
+ * Userspace computes the actual value:
+ *   0x0280 * 0.00390625 -> 2.5
+ * or converts an actual value to the value to be written into <attr>:
+ *   2.5 / 0.00390625 -> 0x0280
+ */
+
+#define U8_8_VAL_MASK           0xffff
+#define U8_8_SCALE_TO_VALUE     "0.00390625"
+
+static ssize_t freq_factor_scale_show(struct kobject *kobj,
+				      struct kobj_attribute *attr,
+				      char *buff)
+{
+	return sysfs_emit(buff, "%s\n", U8_8_SCALE_TO_VALUE);
+}
+
+static u32 media_ratio_mode_to_factor(u32 mode)
+{
+	/* 0 -> 0, 1 -> 256, 2 -> 128 */
+	return !mode ? mode : 256 / mode;
+}
+
+static ssize_t media_freq_factor_show(struct kobject *kobj,
+				      struct kobj_attribute *attr,
+				      char *buff)
+{
+	struct intel_gt *gt = intel_gt_sysfs_get_drvdata(kobj, attr->attr.name);
+	struct intel_guc_slpc *slpc = &gt->uc.guc.slpc;
+	intel_wakeref_t wakeref;
+	u32 mode;
+
+	/*
+	 * Retrieve media_ratio_mode from GEN6_RPNSWREQ bit 13 set by
+	 * GuC. GEN6_RPNSWREQ:13 value 0 represents 1:2 and 1 represents 1:1
+	 */
+	if (IS_XEHPSDV(gt->i915) &&
+	    slpc->media_ratio_mode == SLPC_MEDIA_RATIO_MODE_DYNAMIC_CONTROL) {
+		/*
+		 * For XEHPSDV dynamic mode GEN6_RPNSWREQ:13 does not contain
+		 * the media_ratio_mode, just return the cached media ratio
+		 */
+		mode = slpc->media_ratio_mode;
+	} else {
+		with_intel_runtime_pm(gt->uncore->rpm, wakeref)
+			mode = intel_uncore_read(gt->uncore, GEN6_RPNSWREQ);
+		mode = REG_FIELD_GET(GEN12_MEDIA_FREQ_RATIO, mode) ?
+			SLPC_MEDIA_RATIO_MODE_FIXED_ONE_TO_ONE :
+			SLPC_MEDIA_RATIO_MODE_FIXED_ONE_TO_TWO;
+	}
+
+	return sysfs_emit(buff, "%u\n", media_ratio_mode_to_factor(mode));
+}
+
+static ssize_t media_freq_factor_store(struct kobject *kobj,
+				       struct kobj_attribute *attr,
+				       const char *buff, size_t count)
+{
+	struct intel_gt *gt = intel_gt_sysfs_get_drvdata(kobj, attr->attr.name);
+	struct intel_guc_slpc *slpc = &gt->uc.guc.slpc;
+	u32 factor, mode;
+	int err;
+
+	err = kstrtou32(buff, 0, &factor);
+	if (err)
+		return err;
+
+	for (mode = SLPC_MEDIA_RATIO_MODE_DYNAMIC_CONTROL;
+	     mode <= SLPC_MEDIA_RATIO_MODE_FIXED_ONE_TO_TWO; mode++)
+		if (factor == media_ratio_mode_to_factor(mode))
+			break;
+
+	if (mode > SLPC_MEDIA_RATIO_MODE_FIXED_ONE_TO_TWO)
+		return -EINVAL;
+
+	err = intel_guc_slpc_set_media_ratio_mode(slpc, mode);
+	if (!err) {
+		slpc->media_ratio_mode = mode;
+		DRM_DEBUG("Set slpc->media_ratio_mode to %d", mode);
+	}
+	return err ?: count;
+}
+
+static ssize_t media_RP0_freq_mhz_show(struct kobject *kobj,
+				       struct kobj_attribute *attr,
+				       char *buff)
+{
+	struct intel_gt *gt = intel_gt_sysfs_get_drvdata(kobj, attr->attr.name);
+	u32 val;
+	int err;
+
+	err = snb_pcode_read_p(gt->uncore, XEHP_PCODE_FREQUENCY_CONFIG,
+			       PCODE_MBOX_FC_SC_READ_FUSED_P0,
+			       PCODE_MBOX_DOMAIN_MEDIAFF, &val);
+
+	if (err)
+		return err;
+
+	/* Fused media RP0 read from pcode is in units of 50 MHz */
+	val *= GT_FREQUENCY_MULTIPLIER;
+
+	return sysfs_emit(buff, "%u\n", val);
+}
+
+static ssize_t media_RPn_freq_mhz_show(struct kobject *kobj,
+				       struct kobj_attribute *attr,
+				       char *buff)
+{
+	struct intel_gt *gt = intel_gt_sysfs_get_drvdata(kobj, attr->attr.name);
+	u32 val;
+	int err;
+
+	err = snb_pcode_read_p(gt->uncore, XEHP_PCODE_FREQUENCY_CONFIG,
+			       PCODE_MBOX_FC_SC_READ_FUSED_PN,
+			       PCODE_MBOX_DOMAIN_MEDIAFF, &val);
+
+	if (err)
+		return err;
+
+	/* Fused media RPn read from pcode is in units of 50 MHz */
+	val *= GT_FREQUENCY_MULTIPLIER;
+
+	return sysfs_emit(buff, "%u\n", val);
+}
+
+INTEL_GT_ATTR_RW(media_freq_factor);
+static struct kobj_attribute attr_media_freq_factor_scale =
+	__ATTR(media_freq_factor.scale, 0444, freq_factor_scale_show, NULL);
+INTEL_GT_ATTR_RO(media_RP0_freq_mhz);
+INTEL_GT_ATTR_RO(media_RPn_freq_mhz);
+
+static const struct attribute *media_perf_power_attrs[] = {
+	&attr_media_freq_factor.attr,
+	&attr_media_freq_factor_scale.attr,
+	&attr_media_RP0_freq_mhz.attr,
+	&attr_media_RPn_freq_mhz.attr,
+	NULL
+};
+
+static ssize_t
+default_min_freq_mhz_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_gt *gt = kobj_to_gt(kobj->parent);
+
+	return sysfs_emit(buf, "%u\n", gt->defaults.min_freq);
+}
+
+static struct kobj_attribute default_min_freq_mhz =
+__ATTR(rps_min_freq_mhz, 0444, default_min_freq_mhz_show, NULL);
+
+static ssize_t
+default_max_freq_mhz_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_gt *gt = kobj_to_gt(kobj->parent);
+
+	return sysfs_emit(buf, "%u\n", gt->defaults.max_freq);
+}
+
+static struct kobj_attribute default_max_freq_mhz =
+__ATTR(rps_max_freq_mhz, 0444, default_max_freq_mhz_show, NULL);
+
+static const struct attribute * const rps_defaults_attrs[] = {
+	&default_min_freq_mhz.attr,
+	&default_max_freq_mhz.attr,
+	NULL
+};
+
+static int intel_sysfs_rps_init(struct intel_gt *gt, struct kobject *kobj)
+{
+	const struct attribute * const *attrs;
+	struct attribute *vlv_attr;
+	int ret;
+
+	if (GRAPHICS_VER(gt->i915) < 6)
+		return 0;
+
+	if (is_object_gt(kobj)) {
+		attrs = gen6_rps_attrs;
+		vlv_attr = &attr_rps_vlv_rpe_freq_mhz.attr;
+	} else {
+		attrs = gen6_gt_attrs;
+		vlv_attr = &dev_attr_gt_vlv_rpe_freq_mhz.attr;
+	}
+
+	ret = sysfs_create_files(kobj, attrs);
+	if (ret)
+		return ret;
+
+	if (IS_VALLEYVIEW(gt->i915) || IS_CHERRYVIEW(gt->i915))
+		ret = sysfs_create_file(kobj, vlv_attr);
+
+	return ret;
+}
+
+void intel_gt_sysfs_pm_init(struct intel_gt *gt, struct kobject *kobj)
+{
+	int ret;
+
+	intel_sysfs_rc6_init(gt, kobj);
+
+	ret = intel_sysfs_rps_init(gt, kobj);
+	if (ret)
+		drm_warn(&gt->i915->drm,
+			 "failed to create gt%u RPS sysfs files (%pe)",
+			 gt->info.id, ERR_PTR(ret));
+
+	/* end of the legacy interfaces */
+	if (!is_object_gt(kobj))
+		return;
+
+	ret = sysfs_create_file(kobj, &attr_punit_req_freq_mhz.attr);
+	if (ret)
+		drm_warn(&gt->i915->drm,
+			 "failed to create gt%u punit_req_freq_mhz sysfs (%pe)",
+			 gt->info.id, ERR_PTR(ret));
+
+	if (GRAPHICS_VER(gt->i915) >= 11) {
+		ret = sysfs_create_files(kobj, throttle_reason_attrs);
+		if (ret)
+			drm_warn(&gt->i915->drm,
+				 "failed to create gt%u throttle sysfs files (%pe)",
+				 gt->info.id, ERR_PTR(ret));
+	}
+
+	if (HAS_MEDIA_RATIO_MODE(gt->i915) && intel_uc_uses_guc_slpc(&gt->uc)) {
+		ret = sysfs_create_files(kobj, media_perf_power_attrs);
+		if (ret)
+			drm_warn(&gt->i915->drm,
+				 "failed to create gt%u media_perf_power_attrs sysfs (%pe)\n",
+				 gt->info.id, ERR_PTR(ret));
+	}
+
+	ret = sysfs_create_files(gt->sysfs_defaults, rps_defaults_attrs);
+	if (ret)
+		drm_warn(&gt->i915->drm,
+			 "failed to add gt%u rps defaults (%pe)\n",
+			 gt->info.id, ERR_PTR(ret));
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_sysfs_pm.h b/drivers/gpu/drm/i915/gt/intel_gt_sysfs_pm.h
new file mode 100644
index 000000000..f567105a4
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_sysfs_pm.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __SYSFS_GT_PM_H__
+#define __SYSFS_GT_PM_H__
+
+#include <linux/kobject.h>
+
+#include "intel_gt_types.h"
+
+void intel_gt_sysfs_pm_init(struct intel_gt *gt, struct kobject *kobj);
+
+#endif /* SYSFS_RC6_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_types.h b/drivers/gpu/drm/i915/gt/intel_gt_types.h
new file mode 100644
index 000000000..184ee9b11
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gt_types.h
@@ -0,0 +1,299 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_GT_TYPES__
+#define __INTEL_GT_TYPES__
+
+#include <linux/ktime.h>
+#include <linux/list.h>
+#include <linux/llist.h>
+#include <linux/mutex.h>
+#include <linux/notifier.h>
+#include <linux/seqlock.h>
+#include <linux/spinlock.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+
+#include "uc/intel_uc.h"
+#include "intel_gsc.h"
+
+#include "i915_vma.h"
+#include "intel_engine_types.h"
+#include "intel_gt_buffer_pool_types.h"
+#include "intel_hwconfig.h"
+#include "intel_llc_types.h"
+#include "intel_reset_types.h"
+#include "intel_rc6_types.h"
+#include "intel_rps_types.h"
+#include "intel_migrate_types.h"
+#include "intel_wakeref.h"
+#include "pxp/intel_pxp_types.h"
+
+struct drm_i915_private;
+struct i915_ggtt;
+struct intel_engine_cs;
+struct intel_uncore;
+
+struct intel_mmio_range {
+	u32 start;
+	u32 end;
+};
+
+/*
+ * The hardware has multiple kinds of multicast register ranges that need
+ * special register steering (and future platforms are expected to add
+ * additional types).
+ *
+ * During driver startup, we initialize the steering control register to
+ * direct reads to a slice/subslice that are valid for the 'subslice' class
+ * of multicast registers.  If another type of steering does not have any
+ * overlap in valid steering targets with 'subslice' style registers, we will
+ * need to explicitly re-steer reads of registers of the other type.
+ *
+ * Only the replication types that may need additional non-default steering
+ * are listed here.
+ */
+enum intel_steering_type {
+	L3BANK,
+	MSLICE,
+	LNCF,
+
+	/*
+	 * On some platforms there are multiple types of MCR registers that
+	 * will always return a non-terminated value at instance (0, 0).  We'll
+	 * lump those all into a single category to keep things simple.
+	 */
+	INSTANCE0,
+
+	NUM_STEERING_TYPES
+};
+
+enum intel_submission_method {
+	INTEL_SUBMISSION_RING,
+	INTEL_SUBMISSION_ELSP,
+	INTEL_SUBMISSION_GUC,
+};
+
+struct gt_defaults {
+	u32 min_freq;
+	u32 max_freq;
+};
+
+enum intel_gt_type {
+	GT_PRIMARY,
+	GT_TILE,
+	GT_MEDIA,
+};
+
+struct intel_gt {
+	struct drm_i915_private *i915;
+	const char *name;
+	enum intel_gt_type type;
+
+	struct intel_uncore *uncore;
+	struct i915_ggtt *ggtt;
+
+	struct intel_uc uc;
+	struct intel_gsc gsc;
+
+	struct {
+		/* Serialize global tlb invalidations */
+		struct mutex invalidate_lock;
+
+		/*
+		 * Batch TLB invalidations
+		 *
+		 * After unbinding the PTE, we need to ensure the TLB
+		 * are invalidated prior to releasing the physical pages.
+		 * But we only need one such invalidation for all unbinds,
+		 * so we track how many TLB invalidations have been
+		 * performed since unbind the PTE and only emit an extra
+		 * invalidate if no full barrier has been passed.
+		 */
+		seqcount_mutex_t seqno;
+	} tlb;
+
+	struct i915_wa_list wa_list;
+
+	struct intel_gt_timelines {
+		spinlock_t lock; /* protects active_list */
+		struct list_head active_list;
+	} timelines;
+
+	struct intel_gt_requests {
+		/**
+		 * We leave the user IRQ off as much as possible,
+		 * but this means that requests will finish and never
+		 * be retired once the system goes idle. Set a timer to
+		 * fire periodically while the ring is running. When it
+		 * fires, go retire requests.
+		 */
+		struct delayed_work retire_work;
+	} requests;
+
+	struct {
+		struct llist_head list;
+		struct work_struct work;
+	} watchdog;
+
+	struct intel_wakeref wakeref;
+	atomic_t user_wakeref;
+
+	struct list_head closed_vma;
+	spinlock_t closed_lock; /* guards the list of closed_vma */
+
+	ktime_t last_init_time;
+	struct intel_reset reset;
+
+	/**
+	 * Is the GPU currently considered idle, or busy executing
+	 * userspace requests? Whilst idle, we allow runtime power
+	 * management to power down the hardware and display clocks.
+	 * In order to reduce the effect on performance, there
+	 * is a slight delay before we do so.
+	 */
+	intel_wakeref_t awake;
+
+	u32 clock_frequency;
+	u32 clock_period_ns;
+
+	struct intel_llc llc;
+	struct intel_rc6 rc6;
+	struct intel_rps rps;
+
+	spinlock_t *irq_lock;
+	u32 gt_imr;
+	u32 pm_ier;
+	u32 pm_imr;
+
+	u32 pm_guc_events;
+
+	struct {
+		bool active;
+
+		/**
+		 * @lock: Lock protecting the below fields.
+		 */
+		seqcount_mutex_t lock;
+
+		/**
+		 * @total: Total time this engine was busy.
+		 *
+		 * Accumulated time not counting the most recent block in cases
+		 * where engine is currently busy (active > 0).
+		 */
+		ktime_t total;
+
+		/**
+		 * @start: Timestamp of the last idle to active transition.
+		 *
+		 * Idle is defined as active == 0, active is active > 0.
+		 */
+		ktime_t start;
+	} stats;
+
+	struct intel_engine_cs *engine[I915_NUM_ENGINES];
+	struct intel_engine_cs *engine_class[MAX_ENGINE_CLASS + 1]
+					    [MAX_ENGINE_INSTANCE + 1];
+	enum intel_submission_method submission_method;
+
+	/*
+	 * Default address space (either GGTT or ppGTT depending on arch).
+	 *
+	 * Reserved for exclusive use by the kernel.
+	 */
+	struct i915_address_space *vm;
+
+	/*
+	 * A pool of objects to use as shadow copies of client batch buffers
+	 * when the command parser is enabled. Prevents the client from
+	 * modifying the batch contents after software parsing.
+	 *
+	 * Buffers older than 1s are periodically reaped from the pool,
+	 * or may be reclaimed by the shrinker before then.
+	 */
+	struct intel_gt_buffer_pool buffer_pool;
+
+	struct i915_vma *scratch;
+
+	struct intel_migrate migrate;
+
+	const struct intel_mmio_range *steering_table[NUM_STEERING_TYPES];
+
+	struct {
+		u8 groupid;
+		u8 instanceid;
+	} default_steering;
+
+	/*
+	 * Base of per-tile GTTMMADR where we can derive the MMIO and the GGTT.
+	 */
+	phys_addr_t phys_addr;
+
+	struct intel_gt_info {
+		unsigned int id;
+
+		intel_engine_mask_t engine_mask;
+
+		u32 l3bank_mask;
+
+		u8 num_engines;
+
+		/* General presence of SFC units */
+		u8 sfc_mask;
+
+		/* Media engine access to SFC per instance */
+		u8 vdbox_sfc_access;
+
+		/* Slice/subslice/EU info */
+		struct sseu_dev_info sseu;
+
+		unsigned long mslice_mask;
+
+		/** @hwconfig: hardware configuration data */
+		struct intel_hwconfig hwconfig;
+	} info;
+
+	struct {
+		u8 uc_index;
+		u8 wb_index; /* Only used on HAS_L3_CCS_READ() platforms */
+	} mocs;
+
+	struct intel_pxp pxp;
+
+	/* gt/gtN sysfs */
+	struct kobject sysfs_gt;
+
+	/* sysfs defaults per gt */
+	struct gt_defaults defaults;
+	struct kobject *sysfs_defaults;
+};
+
+struct intel_gt_definition {
+	enum intel_gt_type type;
+	char *name;
+	u32 mapping_base;
+	u32 gsi_offset;
+	intel_engine_mask_t engine_mask;
+};
+
+enum intel_gt_scratch_field {
+	/* 8 bytes */
+	INTEL_GT_SCRATCH_FIELD_DEFAULT = 0,
+
+	/* 8 bytes */
+	INTEL_GT_SCRATCH_FIELD_RENDER_FLUSH = 128,
+
+	/* 8 bytes */
+	INTEL_GT_SCRATCH_FIELD_COHERENTL3_WA = 256,
+
+	/* 6 * 8 bytes */
+	INTEL_GT_SCRATCH_FIELD_PERF_CS_GPR = 2048,
+
+	/* 4 bytes */
+	INTEL_GT_SCRATCH_FIELD_PERF_PREDICATE_RESULT_1 = 2096,
+};
+
+#endif /* __INTEL_GT_TYPES_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_gtt.c b/drivers/gpu/drm/i915/gt/intel_gtt.c
new file mode 100644
index 000000000..f4879f437
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gtt.c
@@ -0,0 +1,647 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/slab.h> /* fault-inject.h is not standalone! */
+
+#include <linux/fault-inject.h>
+#include <linux/sched/mm.h>
+
+#include <drm/drm_cache.h>
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_lmem.h"
+#include "i915_trace.h"
+#include "i915_utils.h"
+#include "intel_gt.h"
+#include "intel_gt_regs.h"
+#include "intel_gtt.h"
+
+
+static bool intel_ggtt_update_needs_vtd_wa(struct drm_i915_private *i915)
+{
+	return IS_BROXTON(i915) && i915_vtd_active(i915);
+}
+
+bool intel_vm_no_concurrent_access_wa(struct drm_i915_private *i915)
+{
+	return IS_CHERRYVIEW(i915) || intel_ggtt_update_needs_vtd_wa(i915);
+}
+
+struct drm_i915_gem_object *alloc_pt_lmem(struct i915_address_space *vm, int sz)
+{
+	struct drm_i915_gem_object *obj;
+
+	/*
+	 * To avoid severe over-allocation when dealing with min_page_size
+	 * restrictions, we override that behaviour here by allowing an object
+	 * size and page layout which can be smaller. In practice this should be
+	 * totally fine, since GTT paging structures are not typically inserted
+	 * into the GTT.
+	 *
+	 * Note that we also hit this path for the scratch page, and for this
+	 * case it might need to be 64K, but that should work fine here since we
+	 * used the passed in size for the page size, which should ensure it
+	 * also has the same alignment.
+	 */
+	obj = __i915_gem_object_create_lmem_with_ps(vm->i915, sz, sz,
+						    vm->lmem_pt_obj_flags);
+	/*
+	 * Ensure all paging structures for this vm share the same dma-resv
+	 * object underneath, with the idea that one object_lock() will lock
+	 * them all at once.
+	 */
+	if (!IS_ERR(obj)) {
+		obj->base.resv = i915_vm_resv_get(vm);
+		obj->shares_resv_from = vm;
+	}
+
+	return obj;
+}
+
+struct drm_i915_gem_object *alloc_pt_dma(struct i915_address_space *vm, int sz)
+{
+	struct drm_i915_gem_object *obj;
+
+	if (I915_SELFTEST_ONLY(should_fail(&vm->fault_attr, 1)))
+		i915_gem_shrink_all(vm->i915);
+
+	obj = i915_gem_object_create_internal(vm->i915, sz);
+	/*
+	 * Ensure all paging structures for this vm share the same dma-resv
+	 * object underneath, with the idea that one object_lock() will lock
+	 * them all at once.
+	 */
+	if (!IS_ERR(obj)) {
+		obj->base.resv = i915_vm_resv_get(vm);
+		obj->shares_resv_from = vm;
+	}
+
+	return obj;
+}
+
+int map_pt_dma(struct i915_address_space *vm, struct drm_i915_gem_object *obj)
+{
+	enum i915_map_type type;
+	void *vaddr;
+
+	type = i915_coherent_map_type(vm->i915, obj, true);
+	vaddr = i915_gem_object_pin_map_unlocked(obj, type);
+	if (IS_ERR(vaddr))
+		return PTR_ERR(vaddr);
+
+	i915_gem_object_make_unshrinkable(obj);
+	return 0;
+}
+
+int map_pt_dma_locked(struct i915_address_space *vm, struct drm_i915_gem_object *obj)
+{
+	enum i915_map_type type;
+	void *vaddr;
+
+	type = i915_coherent_map_type(vm->i915, obj, true);
+	vaddr = i915_gem_object_pin_map(obj, type);
+	if (IS_ERR(vaddr))
+		return PTR_ERR(vaddr);
+
+	i915_gem_object_make_unshrinkable(obj);
+	return 0;
+}
+
+static void clear_vm_list(struct list_head *list)
+{
+	struct i915_vma *vma, *vn;
+
+	list_for_each_entry_safe(vma, vn, list, vm_link) {
+		struct drm_i915_gem_object *obj = vma->obj;
+
+		if (!i915_gem_object_get_rcu(obj)) {
+			/*
+			 * Object is dying, but has not yet cleared its
+			 * vma list.
+			 * Unbind the dying vma to ensure our list
+			 * is completely drained. We leave the destruction to
+			 * the object destructor to avoid the vma
+			 * disappearing under it.
+			 */
+			atomic_and(~I915_VMA_PIN_MASK, &vma->flags);
+			WARN_ON(__i915_vma_unbind(vma));
+
+			/* Remove from the unbound list */
+			list_del_init(&vma->vm_link);
+
+			/*
+			 * Delay the vm and vm mutex freeing until the
+			 * object is done with destruction.
+			 */
+			i915_vm_resv_get(vma->vm);
+			vma->vm_ddestroy = true;
+		} else {
+			i915_vma_destroy_locked(vma);
+			i915_gem_object_put(obj);
+		}
+
+	}
+}
+
+static void __i915_vm_close(struct i915_address_space *vm)
+{
+	mutex_lock(&vm->mutex);
+
+	clear_vm_list(&vm->bound_list);
+	clear_vm_list(&vm->unbound_list);
+
+	/* Check for must-fix unanticipated side-effects */
+	GEM_BUG_ON(!list_empty(&vm->bound_list));
+	GEM_BUG_ON(!list_empty(&vm->unbound_list));
+
+	mutex_unlock(&vm->mutex);
+}
+
+/* lock the vm into the current ww, if we lock one, we lock all */
+int i915_vm_lock_objects(struct i915_address_space *vm,
+			 struct i915_gem_ww_ctx *ww)
+{
+	if (vm->scratch[0]->base.resv == &vm->_resv) {
+		return i915_gem_object_lock(vm->scratch[0], ww);
+	} else {
+		struct i915_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
+
+		/* We borrowed the scratch page from ggtt, take the top level object */
+		return i915_gem_object_lock(ppgtt->pd->pt.base, ww);
+	}
+}
+
+void i915_address_space_fini(struct i915_address_space *vm)
+{
+	drm_mm_takedown(&vm->mm);
+}
+
+/**
+ * i915_vm_resv_release - Final struct i915_address_space destructor
+ * @kref: Pointer to the &i915_address_space.resv_ref member.
+ *
+ * This function is called when the last lock sharer no longer shares the
+ * &i915_address_space._resv lock, and also if we raced when
+ * destroying a vma by the vma destruction
+ */
+void i915_vm_resv_release(struct kref *kref)
+{
+	struct i915_address_space *vm =
+		container_of(kref, typeof(*vm), resv_ref);
+
+	dma_resv_fini(&vm->_resv);
+	mutex_destroy(&vm->mutex);
+
+	kfree(vm);
+}
+
+static void __i915_vm_release(struct work_struct *work)
+{
+	struct i915_address_space *vm =
+		container_of(work, struct i915_address_space, release_work);
+
+	__i915_vm_close(vm);
+
+	/* Synchronize async unbinds. */
+	i915_vma_resource_bind_dep_sync_all(vm);
+
+	vm->cleanup(vm);
+	i915_address_space_fini(vm);
+
+	i915_vm_resv_put(vm);
+}
+
+void i915_vm_release(struct kref *kref)
+{
+	struct i915_address_space *vm =
+		container_of(kref, struct i915_address_space, ref);
+
+	GEM_BUG_ON(i915_is_ggtt(vm));
+	trace_i915_ppgtt_release(vm);
+
+	queue_work(vm->i915->wq, &vm->release_work);
+}
+
+void i915_address_space_init(struct i915_address_space *vm, int subclass)
+{
+	kref_init(&vm->ref);
+
+	/*
+	 * Special case for GGTT that has already done an early
+	 * kref_init here.
+	 */
+	if (!kref_read(&vm->resv_ref))
+		kref_init(&vm->resv_ref);
+
+	vm->pending_unbind = RB_ROOT_CACHED;
+	INIT_WORK(&vm->release_work, __i915_vm_release);
+
+	/*
+	 * The vm->mutex must be reclaim safe (for use in the shrinker).
+	 * Do a dummy acquire now under fs_reclaim so that any allocation
+	 * attempt holding the lock is immediately reported by lockdep.
+	 */
+	mutex_init(&vm->mutex);
+	lockdep_set_subclass(&vm->mutex, subclass);
+
+	if (!intel_vm_no_concurrent_access_wa(vm->i915)) {
+		i915_gem_shrinker_taints_mutex(vm->i915, &vm->mutex);
+	} else {
+		/*
+		 * CHV + BXT VTD workaround use stop_machine(),
+		 * which is allowed to allocate memory. This means &vm->mutex
+		 * is the outer lock, and in theory we can allocate memory inside
+		 * it through stop_machine().
+		 *
+		 * Add the annotation for this, we use trylock in shrinker.
+		 */
+		mutex_acquire(&vm->mutex.dep_map, 0, 0, _THIS_IP_);
+		might_alloc(GFP_KERNEL);
+		mutex_release(&vm->mutex.dep_map, _THIS_IP_);
+	}
+	dma_resv_init(&vm->_resv);
+
+	GEM_BUG_ON(!vm->total);
+	drm_mm_init(&vm->mm, 0, vm->total);
+
+	memset64(vm->min_alignment, I915_GTT_MIN_ALIGNMENT,
+		 ARRAY_SIZE(vm->min_alignment));
+
+	if (HAS_64K_PAGES(vm->i915) && NEEDS_COMPACT_PT(vm->i915) &&
+	    subclass == VM_CLASS_PPGTT) {
+		vm->min_alignment[INTEL_MEMORY_LOCAL] = I915_GTT_PAGE_SIZE_2M;
+		vm->min_alignment[INTEL_MEMORY_STOLEN_LOCAL] = I915_GTT_PAGE_SIZE_2M;
+	} else if (HAS_64K_PAGES(vm->i915)) {
+		vm->min_alignment[INTEL_MEMORY_LOCAL] = I915_GTT_PAGE_SIZE_64K;
+		vm->min_alignment[INTEL_MEMORY_STOLEN_LOCAL] = I915_GTT_PAGE_SIZE_64K;
+	}
+
+	vm->mm.head_node.color = I915_COLOR_UNEVICTABLE;
+
+	INIT_LIST_HEAD(&vm->bound_list);
+	INIT_LIST_HEAD(&vm->unbound_list);
+}
+
+void *__px_vaddr(struct drm_i915_gem_object *p)
+{
+	enum i915_map_type type;
+
+	GEM_BUG_ON(!i915_gem_object_has_pages(p));
+	return page_unpack_bits(p->mm.mapping, &type);
+}
+
+dma_addr_t __px_dma(struct drm_i915_gem_object *p)
+{
+	GEM_BUG_ON(!i915_gem_object_has_pages(p));
+	return sg_dma_address(p->mm.pages->sgl);
+}
+
+struct page *__px_page(struct drm_i915_gem_object *p)
+{
+	GEM_BUG_ON(!i915_gem_object_has_pages(p));
+	return sg_page(p->mm.pages->sgl);
+}
+
+void
+fill_page_dma(struct drm_i915_gem_object *p, const u64 val, unsigned int count)
+{
+	void *vaddr = __px_vaddr(p);
+
+	memset64(vaddr, val, count);
+	drm_clflush_virt_range(vaddr, PAGE_SIZE);
+}
+
+static void poison_scratch_page(struct drm_i915_gem_object *scratch)
+{
+	void *vaddr = __px_vaddr(scratch);
+	u8 val;
+
+	val = 0;
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		val = POISON_FREE;
+
+	memset(vaddr, val, scratch->base.size);
+	drm_clflush_virt_range(vaddr, scratch->base.size);
+}
+
+int setup_scratch_page(struct i915_address_space *vm)
+{
+	unsigned long size;
+
+	/*
+	 * In order to utilize 64K pages for an object with a size < 2M, we will
+	 * need to support a 64K scratch page, given that every 16th entry for a
+	 * page-table operating in 64K mode must point to a properly aligned 64K
+	 * region, including any PTEs which happen to point to scratch.
+	 *
+	 * This is only relevant for the 48b PPGTT where we support
+	 * huge-gtt-pages, see also i915_vma_insert(). However, as we share the
+	 * scratch (read-only) between all vm, we create one 64k scratch page
+	 * for all.
+	 */
+	size = I915_GTT_PAGE_SIZE_4K;
+	if (i915_vm_is_4lvl(vm) &&
+	    HAS_PAGE_SIZES(vm->i915, I915_GTT_PAGE_SIZE_64K))
+		size = I915_GTT_PAGE_SIZE_64K;
+
+	do {
+		struct drm_i915_gem_object *obj;
+
+		obj = vm->alloc_scratch_dma(vm, size);
+		if (IS_ERR(obj))
+			goto skip;
+
+		if (map_pt_dma(vm, obj))
+			goto skip_obj;
+
+		/* We need a single contiguous page for our scratch */
+		if (obj->mm.page_sizes.sg < size)
+			goto skip_obj;
+
+		/* And it needs to be correspondingly aligned */
+		if (__px_dma(obj) & (size - 1))
+			goto skip_obj;
+
+		/*
+		 * Use a non-zero scratch page for debugging.
+		 *
+		 * We want a value that should be reasonably obvious
+		 * to spot in the error state, while also causing a GPU hang
+		 * if executed. We prefer using a clear page in production, so
+		 * should it ever be accidentally used, the effect should be
+		 * fairly benign.
+		 */
+		poison_scratch_page(obj);
+
+		vm->scratch[0] = obj;
+		vm->scratch_order = get_order(size);
+		return 0;
+
+skip_obj:
+		i915_gem_object_put(obj);
+skip:
+		if (size == I915_GTT_PAGE_SIZE_4K)
+			return -ENOMEM;
+
+		/*
+		 * If we need 64K minimum GTT pages for device local-memory,
+		 * like on XEHPSDV, then we need to fail the allocation here,
+		 * otherwise we can't safely support the insertion of
+		 * local-memory pages for this vm, since the HW expects the
+		 * correct physical alignment and size when the page-table is
+		 * operating in 64K GTT mode, which includes any scratch PTEs,
+		 * since userspace can still touch them.
+		 */
+		if (HAS_64K_PAGES(vm->i915))
+			return -ENOMEM;
+
+		size = I915_GTT_PAGE_SIZE_4K;
+	} while (1);
+}
+
+void free_scratch(struct i915_address_space *vm)
+{
+	int i;
+
+	if (!vm->scratch[0])
+		return;
+
+	for (i = 0; i <= vm->top; i++)
+		i915_gem_object_put(vm->scratch[i]);
+}
+
+void gtt_write_workarounds(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+
+	/*
+	 * This function is for gtt related workarounds. This function is
+	 * called on driver load and after a GPU reset, so you can place
+	 * workarounds here even if they get overwritten by GPU reset.
+	 */
+	/* WaIncreaseDefaultTLBEntries:chv,bdw,skl,bxt,kbl,glk,cfl,cnl,icl */
+	if (IS_BROADWELL(i915))
+		intel_uncore_write(uncore,
+				   GEN8_L3_LRA_1_GPGPU,
+				   GEN8_L3_LRA_1_GPGPU_DEFAULT_VALUE_BDW);
+	else if (IS_CHERRYVIEW(i915))
+		intel_uncore_write(uncore,
+				   GEN8_L3_LRA_1_GPGPU,
+				   GEN8_L3_LRA_1_GPGPU_DEFAULT_VALUE_CHV);
+	else if (IS_GEN9_LP(i915))
+		intel_uncore_write(uncore,
+				   GEN8_L3_LRA_1_GPGPU,
+				   GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_BXT);
+	else if (GRAPHICS_VER(i915) >= 9 && GRAPHICS_VER(i915) <= 11)
+		intel_uncore_write(uncore,
+				   GEN8_L3_LRA_1_GPGPU,
+				   GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_SKL);
+
+	/*
+	 * To support 64K PTEs we need to first enable the use of the
+	 * Intermediate-Page-Size(IPS) bit of the PDE field via some magical
+	 * mmio, otherwise the page-walker will simply ignore the IPS bit. This
+	 * shouldn't be needed after GEN10.
+	 *
+	 * 64K pages were first introduced from BDW+, although technically they
+	 * only *work* from gen9+. For pre-BDW we instead have the option for
+	 * 32K pages, but we don't currently have any support for it in our
+	 * driver.
+	 */
+	if (HAS_PAGE_SIZES(i915, I915_GTT_PAGE_SIZE_64K) &&
+	    GRAPHICS_VER(i915) <= 10)
+		intel_uncore_rmw(uncore,
+				 GEN8_GAMW_ECO_DEV_RW_IA,
+				 0,
+				 GAMW_ECO_ENABLE_64K_IPS_FIELD);
+
+	if (IS_GRAPHICS_VER(i915, 8, 11)) {
+		bool can_use_gtt_cache = true;
+
+		/*
+		 * According to the BSpec if we use 2M/1G pages then we also
+		 * need to disable the GTT cache. At least on BDW we can see
+		 * visual corruption when using 2M pages, and not disabling the
+		 * GTT cache.
+		 */
+		if (HAS_PAGE_SIZES(i915, I915_GTT_PAGE_SIZE_2M))
+			can_use_gtt_cache = false;
+
+		/* WaGttCachingOffByDefault */
+		intel_uncore_write(uncore,
+				   HSW_GTT_CACHE_EN,
+				   can_use_gtt_cache ? GTT_CACHE_EN_ALL : 0);
+		drm_WARN_ON_ONCE(&i915->drm, can_use_gtt_cache &&
+				 intel_uncore_read(uncore,
+						   HSW_GTT_CACHE_EN) == 0);
+	}
+}
+
+static void tgl_setup_private_ppat(struct intel_uncore *uncore)
+{
+	/* TGL doesn't support LLC or AGE settings */
+	intel_uncore_write(uncore, GEN12_PAT_INDEX(0), GEN8_PPAT_WB);
+	intel_uncore_write(uncore, GEN12_PAT_INDEX(1), GEN8_PPAT_WC);
+	intel_uncore_write(uncore, GEN12_PAT_INDEX(2), GEN8_PPAT_WT);
+	intel_uncore_write(uncore, GEN12_PAT_INDEX(3), GEN8_PPAT_UC);
+	intel_uncore_write(uncore, GEN12_PAT_INDEX(4), GEN8_PPAT_WB);
+	intel_uncore_write(uncore, GEN12_PAT_INDEX(5), GEN8_PPAT_WB);
+	intel_uncore_write(uncore, GEN12_PAT_INDEX(6), GEN8_PPAT_WB);
+	intel_uncore_write(uncore, GEN12_PAT_INDEX(7), GEN8_PPAT_WB);
+}
+
+static void icl_setup_private_ppat(struct intel_uncore *uncore)
+{
+	intel_uncore_write(uncore,
+			   GEN10_PAT_INDEX(0),
+			   GEN8_PPAT_WB | GEN8_PPAT_LLC);
+	intel_uncore_write(uncore,
+			   GEN10_PAT_INDEX(1),
+			   GEN8_PPAT_WC | GEN8_PPAT_LLCELLC);
+	intel_uncore_write(uncore,
+			   GEN10_PAT_INDEX(2),
+			   GEN8_PPAT_WB | GEN8_PPAT_ELLC_OVERRIDE);
+	intel_uncore_write(uncore,
+			   GEN10_PAT_INDEX(3),
+			   GEN8_PPAT_UC);
+	intel_uncore_write(uncore,
+			   GEN10_PAT_INDEX(4),
+			   GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(0));
+	intel_uncore_write(uncore,
+			   GEN10_PAT_INDEX(5),
+			   GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(1));
+	intel_uncore_write(uncore,
+			   GEN10_PAT_INDEX(6),
+			   GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(2));
+	intel_uncore_write(uncore,
+			   GEN10_PAT_INDEX(7),
+			   GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(3));
+}
+
+/*
+ * The GGTT and PPGTT need a private PPAT setup in order to handle cacheability
+ * bits. When using advanced contexts each context stores its own PAT, but
+ * writing this data shouldn't be harmful even in those cases.
+ */
+static void bdw_setup_private_ppat(struct intel_uncore *uncore)
+{
+	struct drm_i915_private *i915 = uncore->i915;
+	u64 pat;
+
+	pat = GEN8_PPAT(0, GEN8_PPAT_WB | GEN8_PPAT_LLC) |	/* for normal objects, no eLLC */
+	      GEN8_PPAT(1, GEN8_PPAT_WC | GEN8_PPAT_LLCELLC) |	/* for something pointing to ptes? */
+	      GEN8_PPAT(3, GEN8_PPAT_UC) |			/* Uncached objects, mostly for scanout */
+	      GEN8_PPAT(4, GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(0)) |
+	      GEN8_PPAT(5, GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(1)) |
+	      GEN8_PPAT(6, GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(2)) |
+	      GEN8_PPAT(7, GEN8_PPAT_WB | GEN8_PPAT_LLCELLC | GEN8_PPAT_AGE(3));
+
+	/* for scanout with eLLC */
+	if (GRAPHICS_VER(i915) >= 9)
+		pat |= GEN8_PPAT(2, GEN8_PPAT_WB | GEN8_PPAT_ELLC_OVERRIDE);
+	else
+		pat |= GEN8_PPAT(2, GEN8_PPAT_WT | GEN8_PPAT_LLCELLC);
+
+	intel_uncore_write(uncore, GEN8_PRIVATE_PAT_LO, lower_32_bits(pat));
+	intel_uncore_write(uncore, GEN8_PRIVATE_PAT_HI, upper_32_bits(pat));
+}
+
+static void chv_setup_private_ppat(struct intel_uncore *uncore)
+{
+	u64 pat;
+
+	/*
+	 * Map WB on BDW to snooped on CHV.
+	 *
+	 * Only the snoop bit has meaning for CHV, the rest is
+	 * ignored.
+	 *
+	 * The hardware will never snoop for certain types of accesses:
+	 * - CPU GTT (GMADR->GGTT->no snoop->memory)
+	 * - PPGTT page tables
+	 * - some other special cycles
+	 *
+	 * As with BDW, we also need to consider the following for GT accesses:
+	 * "For GGTT, there is NO pat_sel[2:0] from the entry,
+	 * so RTL will always use the value corresponding to
+	 * pat_sel = 000".
+	 * Which means we must set the snoop bit in PAT entry 0
+	 * in order to keep the global status page working.
+	 */
+
+	pat = GEN8_PPAT(0, CHV_PPAT_SNOOP) |
+	      GEN8_PPAT(1, 0) |
+	      GEN8_PPAT(2, 0) |
+	      GEN8_PPAT(3, 0) |
+	      GEN8_PPAT(4, CHV_PPAT_SNOOP) |
+	      GEN8_PPAT(5, CHV_PPAT_SNOOP) |
+	      GEN8_PPAT(6, CHV_PPAT_SNOOP) |
+	      GEN8_PPAT(7, CHV_PPAT_SNOOP);
+
+	intel_uncore_write(uncore, GEN8_PRIVATE_PAT_LO, lower_32_bits(pat));
+	intel_uncore_write(uncore, GEN8_PRIVATE_PAT_HI, upper_32_bits(pat));
+}
+
+void setup_private_pat(struct intel_uncore *uncore)
+{
+	struct drm_i915_private *i915 = uncore->i915;
+
+	GEM_BUG_ON(GRAPHICS_VER(i915) < 8);
+
+	if (GRAPHICS_VER(i915) >= 12)
+		tgl_setup_private_ppat(uncore);
+	else if (GRAPHICS_VER(i915) >= 11)
+		icl_setup_private_ppat(uncore);
+	else if (IS_CHERRYVIEW(i915) || IS_GEN9_LP(i915))
+		chv_setup_private_ppat(uncore);
+	else
+		bdw_setup_private_ppat(uncore);
+}
+
+struct i915_vma *
+__vm_create_scratch_for_read(struct i915_address_space *vm, unsigned long size)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+
+	obj = i915_gem_object_create_internal(vm->i915, PAGE_ALIGN(size));
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	i915_gem_object_set_cache_coherency(obj, I915_CACHE_LLC);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return vma;
+	}
+
+	return vma;
+}
+
+struct i915_vma *
+__vm_create_scratch_for_read_pinned(struct i915_address_space *vm, unsigned long size)
+{
+	struct i915_vma *vma;
+	int err;
+
+	vma = __vm_create_scratch_for_read(vm, size);
+	if (IS_ERR(vma))
+		return vma;
+
+	err = i915_vma_pin(vma, 0, 0,
+			   i915_vma_is_ggtt(vma) ? PIN_GLOBAL : PIN_USER);
+	if (err) {
+		i915_vma_put(vma);
+		return ERR_PTR(err);
+	}
+
+	return vma;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_gtt.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_gtt.h b/drivers/gpu/drm/i915/gt/intel_gtt.h
new file mode 100644
index 000000000..c0ca53cba
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_gtt.h
@@ -0,0 +1,696 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ *
+ * Please try to maintain the following order within this file unless it makes
+ * sense to do otherwise. From top to bottom:
+ * 1. typedefs
+ * 2. #defines, and macros
+ * 3. structure definitions
+ * 4. function prototypes
+ *
+ * Within each section, please try to order by generation in ascending order,
+ * from top to bottom (ie. gen6 on the top, gen8 on the bottom).
+ */
+
+#ifndef __INTEL_GTT_H__
+#define __INTEL_GTT_H__
+
+#include <linux/io-mapping.h>
+#include <linux/kref.h>
+#include <linux/mm.h>
+#include <linux/pagevec.h>
+#include <linux/scatterlist.h>
+#include <linux/workqueue.h>
+
+#include <drm/drm_mm.h>
+
+#include "gt/intel_reset.h"
+#include "i915_selftest.h"
+#include "i915_vma_resource.h"
+#include "i915_vma_types.h"
+#include "i915_params.h"
+#include "intel_memory_region.h"
+
+#define I915_GFP_ALLOW_FAIL (GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN)
+
+#if IS_ENABLED(CONFIG_DRM_I915_TRACE_GTT)
+#define DBG(...) trace_printk(__VA_ARGS__)
+#else
+#define DBG(...)
+#endif
+
+#define NALLOC 3 /* 1 normal, 1 for concurrent threads, 1 for preallocation */
+
+#define I915_GTT_PAGE_SIZE_4K	BIT_ULL(12)
+#define I915_GTT_PAGE_SIZE_64K	BIT_ULL(16)
+#define I915_GTT_PAGE_SIZE_2M	BIT_ULL(21)
+
+#define I915_GTT_PAGE_SIZE I915_GTT_PAGE_SIZE_4K
+#define I915_GTT_MAX_PAGE_SIZE I915_GTT_PAGE_SIZE_2M
+
+#define I915_GTT_PAGE_MASK -I915_GTT_PAGE_SIZE
+
+#define I915_GTT_MIN_ALIGNMENT I915_GTT_PAGE_SIZE
+
+#define I915_FENCE_REG_NONE -1
+#define I915_MAX_NUM_FENCES 32
+/* 32 fences + sign bit for FENCE_REG_NONE */
+#define I915_MAX_NUM_FENCE_BITS 6
+
+typedef u32 gen6_pte_t;
+typedef u64 gen8_pte_t;
+
+#define ggtt_total_entries(ggtt) ((ggtt)->vm.total >> PAGE_SHIFT)
+
+#define I915_PTES(pte_len)		((unsigned int)(PAGE_SIZE / (pte_len)))
+#define I915_PTE_MASK(pte_len)		(I915_PTES(pte_len) - 1)
+#define I915_PDES			512
+#define I915_PDE_MASK			(I915_PDES - 1)
+
+/* gen6-hsw has bit 11-4 for physical addr bit 39-32 */
+#define GEN6_GTT_ADDR_ENCODE(addr)	((addr) | (((addr) >> 28) & 0xff0))
+#define GEN6_PTE_ADDR_ENCODE(addr)	GEN6_GTT_ADDR_ENCODE(addr)
+#define GEN6_PDE_ADDR_ENCODE(addr)	GEN6_GTT_ADDR_ENCODE(addr)
+#define GEN6_PTE_CACHE_LLC		(2 << 1)
+#define GEN6_PTE_UNCACHED		(1 << 1)
+#define GEN6_PTE_VALID			REG_BIT(0)
+
+#define GEN6_PTES			I915_PTES(sizeof(gen6_pte_t))
+#define GEN6_PD_SIZE		        (I915_PDES * PAGE_SIZE)
+#define GEN6_PD_ALIGN			(PAGE_SIZE * 16)
+#define GEN6_PDE_SHIFT			22
+#define GEN6_PDE_VALID			REG_BIT(0)
+#define NUM_PTE(pde_shift)     (1 << (pde_shift - PAGE_SHIFT))
+
+#define GEN7_PTE_CACHE_L3_LLC		(3 << 1)
+
+#define BYT_PTE_SNOOPED_BY_CPU_CACHES	REG_BIT(2)
+#define BYT_PTE_WRITEABLE		REG_BIT(1)
+
+#define GEN12_PPGTT_PTE_LM	BIT_ULL(11)
+
+#define GEN12_GGTT_PTE_LM	BIT_ULL(1)
+
+#define GEN12_PDE_64K BIT(6)
+
+/*
+ * Cacheability Control is a 4-bit value. The low three bits are stored in bits
+ * 3:1 of the PTE, while the fourth bit is stored in bit 11 of the PTE.
+ */
+#define HSW_CACHEABILITY_CONTROL(bits)	((((bits) & 0x7) << 1) | \
+					 (((bits) & 0x8) << (11 - 3)))
+#define HSW_WB_LLC_AGE3			HSW_CACHEABILITY_CONTROL(0x2)
+#define HSW_WB_LLC_AGE0			HSW_CACHEABILITY_CONTROL(0x3)
+#define HSW_WB_ELLC_LLC_AGE3		HSW_CACHEABILITY_CONTROL(0x8)
+#define HSW_WB_ELLC_LLC_AGE0		HSW_CACHEABILITY_CONTROL(0xb)
+#define HSW_WT_ELLC_LLC_AGE3		HSW_CACHEABILITY_CONTROL(0x7)
+#define HSW_WT_ELLC_LLC_AGE0		HSW_CACHEABILITY_CONTROL(0x6)
+#define HSW_PTE_UNCACHED		(0)
+#define HSW_GTT_ADDR_ENCODE(addr)	((addr) | (((addr) >> 28) & 0x7f0))
+#define HSW_PTE_ADDR_ENCODE(addr)	HSW_GTT_ADDR_ENCODE(addr)
+
+/*
+ * GEN8 32b style address is defined as a 3 level page table:
+ * 31:30 | 29:21 | 20:12 |  11:0
+ * PDPE  |  PDE  |  PTE  | offset
+ * The difference as compared to normal x86 3 level page table is the PDPEs are
+ * programmed via register.
+ *
+ * GEN8 48b style address is defined as a 4 level page table:
+ * 47:39 | 38:30 | 29:21 | 20:12 |  11:0
+ * PML4E | PDPE  |  PDE  |  PTE  | offset
+ */
+#define GEN8_3LVL_PDPES			4
+
+#define PPAT_UNCACHED			(_PAGE_PWT | _PAGE_PCD)
+#define PPAT_CACHED_PDE			0 /* WB LLC */
+#define PPAT_CACHED			_PAGE_PAT /* WB LLCeLLC */
+#define PPAT_DISPLAY_ELLC		_PAGE_PCD /* WT eLLC */
+
+#define CHV_PPAT_SNOOP			REG_BIT(6)
+#define GEN8_PPAT_AGE(x)		((x)<<4)
+#define GEN8_PPAT_LLCeLLC		(3<<2)
+#define GEN8_PPAT_LLCELLC		(2<<2)
+#define GEN8_PPAT_LLC			(1<<2)
+#define GEN8_PPAT_WB			(3<<0)
+#define GEN8_PPAT_WT			(2<<0)
+#define GEN8_PPAT_WC			(1<<0)
+#define GEN8_PPAT_UC			(0<<0)
+#define GEN8_PPAT_ELLC_OVERRIDE		(0<<2)
+#define GEN8_PPAT(i, x)			((u64)(x) << ((i) * 8))
+
+#define GEN8_PAGE_PRESENT		BIT_ULL(0)
+#define GEN8_PAGE_RW			BIT_ULL(1)
+
+#define GEN8_PDE_IPS_64K BIT(11)
+#define GEN8_PDE_PS_2M   BIT(7)
+
+enum i915_cache_level;
+
+struct drm_i915_gem_object;
+struct i915_fence_reg;
+struct i915_vma;
+struct intel_gt;
+
+#define for_each_sgt_daddr(__dp, __iter, __sgt) \
+	__for_each_sgt_daddr(__dp, __iter, __sgt, I915_GTT_PAGE_SIZE)
+
+struct i915_page_table {
+	struct drm_i915_gem_object *base;
+	union {
+		atomic_t used;
+		struct i915_page_table *stash;
+	};
+	bool is_compact;
+};
+
+struct i915_page_directory {
+	struct i915_page_table pt;
+	spinlock_t lock;
+	void **entry;
+};
+
+#define __px_choose_expr(x, type, expr, other) \
+	__builtin_choose_expr( \
+	__builtin_types_compatible_p(typeof(x), type) || \
+	__builtin_types_compatible_p(typeof(x), const type), \
+	({ type __x = (type)(x); expr; }), \
+	other)
+
+#define px_base(px) \
+	__px_choose_expr(px, struct drm_i915_gem_object *, __x, \
+	__px_choose_expr(px, struct i915_page_table *, __x->base, \
+	__px_choose_expr(px, struct i915_page_directory *, __x->pt.base, \
+	(void)0)))
+
+struct page *__px_page(struct drm_i915_gem_object *p);
+dma_addr_t __px_dma(struct drm_i915_gem_object *p);
+#define px_dma(px) (__px_dma(px_base(px)))
+
+void *__px_vaddr(struct drm_i915_gem_object *p);
+#define px_vaddr(px) (__px_vaddr(px_base(px)))
+
+#define px_pt(px) \
+	__px_choose_expr(px, struct i915_page_table *, __x, \
+	__px_choose_expr(px, struct i915_page_directory *, &__x->pt, \
+	(void)0))
+#define px_used(px) (&px_pt(px)->used)
+
+struct i915_vm_pt_stash {
+	/* preallocated chains of page tables/directories */
+	struct i915_page_table *pt[2];
+	/*
+	 * Optionally override the alignment/size of the physical page that
+	 * contains each PT. If not set defaults back to the usual
+	 * I915_GTT_PAGE_SIZE_4K. This does not influence the other paging
+	 * structures. MUST be a power-of-two. ONLY applicable on discrete
+	 * platforms.
+	 */
+	int pt_sz;
+};
+
+struct i915_vma_ops {
+	/* Map an object into an address space with the given cache flags. */
+	void (*bind_vma)(struct i915_address_space *vm,
+			 struct i915_vm_pt_stash *stash,
+			 struct i915_vma_resource *vma_res,
+			 enum i915_cache_level cache_level,
+			 u32 flags);
+	/*
+	 * Unmap an object from an address space. This usually consists of
+	 * setting the valid PTE entries to a reserved scratch page.
+	 */
+	void (*unbind_vma)(struct i915_address_space *vm,
+			   struct i915_vma_resource *vma_res);
+
+};
+
+struct i915_address_space {
+	struct kref ref;
+	struct work_struct release_work;
+
+	struct drm_mm mm;
+	struct intel_gt *gt;
+	struct drm_i915_private *i915;
+	struct device *dma;
+	u64 total;		/* size addr space maps (ex. 2GB for ggtt) */
+	u64 reserved;		/* size addr space reserved */
+	u64 min_alignment[INTEL_MEMORY_STOLEN_LOCAL + 1];
+
+	unsigned int bind_async_flags;
+
+	struct mutex mutex; /* protects vma and our lists */
+
+	struct kref resv_ref; /* kref to keep the reservation lock alive. */
+	struct dma_resv _resv; /* reservation lock for all pd objects, and buffer pool */
+#define VM_CLASS_GGTT 0
+#define VM_CLASS_PPGTT 1
+#define VM_CLASS_DPT 2
+
+	struct drm_i915_gem_object *scratch[4];
+	/**
+	 * List of vma currently bound.
+	 */
+	struct list_head bound_list;
+
+	/**
+	 * List of vmas not yet bound or evicted.
+	 */
+	struct list_head unbound_list;
+
+	/* Global GTT */
+	bool is_ggtt:1;
+
+	/* Display page table */
+	bool is_dpt:1;
+
+	/* Some systems support read-only mappings for GGTT and/or PPGTT */
+	bool has_read_only:1;
+
+	/* Skip pte rewrite on unbind for suspend. Protected by @mutex */
+	bool skip_pte_rewrite:1;
+
+	u8 top;
+	u8 pd_shift;
+	u8 scratch_order;
+
+	/* Flags used when creating page-table objects for this vm */
+	unsigned long lmem_pt_obj_flags;
+
+	/* Interval tree for pending unbind vma resources */
+	struct rb_root_cached pending_unbind;
+
+	struct drm_i915_gem_object *
+		(*alloc_pt_dma)(struct i915_address_space *vm, int sz);
+	struct drm_i915_gem_object *
+		(*alloc_scratch_dma)(struct i915_address_space *vm, int sz);
+
+	u64 (*pte_encode)(dma_addr_t addr,
+			  enum i915_cache_level level,
+			  u32 flags); /* Create a valid PTE */
+#define PTE_READ_ONLY	BIT(0)
+#define PTE_LM		BIT(1)
+
+	void (*allocate_va_range)(struct i915_address_space *vm,
+				  struct i915_vm_pt_stash *stash,
+				  u64 start, u64 length);
+	void (*clear_range)(struct i915_address_space *vm,
+			    u64 start, u64 length);
+	void (*insert_page)(struct i915_address_space *vm,
+			    dma_addr_t addr,
+			    u64 offset,
+			    enum i915_cache_level cache_level,
+			    u32 flags);
+	void (*insert_entries)(struct i915_address_space *vm,
+			       struct i915_vma_resource *vma_res,
+			       enum i915_cache_level cache_level,
+			       u32 flags);
+	void (*raw_insert_page)(struct i915_address_space *vm,
+				dma_addr_t addr,
+				u64 offset,
+				enum i915_cache_level cache_level,
+				u32 flags);
+	void (*raw_insert_entries)(struct i915_address_space *vm,
+				   struct i915_vma_resource *vma_res,
+				   enum i915_cache_level cache_level,
+				   u32 flags);
+	void (*cleanup)(struct i915_address_space *vm);
+
+	void (*foreach)(struct i915_address_space *vm,
+			u64 start, u64 length,
+			void (*fn)(struct i915_address_space *vm,
+				   struct i915_page_table *pt,
+				   void *data),
+			void *data);
+
+	struct i915_vma_ops vma_ops;
+
+	I915_SELFTEST_DECLARE(struct fault_attr fault_attr);
+	I915_SELFTEST_DECLARE(bool scrub_64K);
+};
+
+/*
+ * The Graphics Translation Table is the way in which GEN hardware translates a
+ * Graphics Virtual Address into a Physical Address. In addition to the normal
+ * collateral associated with any va->pa translations GEN hardware also has a
+ * portion of the GTT which can be mapped by the CPU and remain both coherent
+ * and correct (in cases like swizzling). That region is referred to as GMADR in
+ * the spec.
+ */
+struct i915_ggtt {
+	struct i915_address_space vm;
+
+	struct io_mapping iomap;	/* Mapping to our CPU mappable region */
+	struct resource gmadr;          /* GMADR resource */
+	resource_size_t mappable_end;	/* End offset that we can CPU map */
+
+	/** "Graphics Stolen Memory" holds the global PTEs */
+	void __iomem *gsm;
+	void (*invalidate)(struct i915_ggtt *ggtt);
+
+	/** PPGTT used for aliasing the PPGTT with the GTT */
+	struct i915_ppgtt *alias;
+
+	bool do_idle_maps;
+
+	/**
+	 * @pte_lost: Are ptes lost on resume?
+	 *
+	 * Whether the system was recently restored from hibernate and
+	 * thus may have lost pte content.
+	 */
+	bool pte_lost;
+
+	/**
+	 * @probed_pte: Probed pte value on suspend. Re-checked on resume.
+	 */
+	u64 probed_pte;
+
+	int mtrr;
+
+	/** Bit 6 swizzling required for X tiling */
+	u32 bit_6_swizzle_x;
+	/** Bit 6 swizzling required for Y tiling */
+	u32 bit_6_swizzle_y;
+
+	u32 pin_bias;
+
+	unsigned int num_fences;
+	struct i915_fence_reg *fence_regs;
+	struct list_head fence_list;
+
+	/**
+	 * List of all objects in gtt_space, currently mmaped by userspace.
+	 * All objects within this list must also be on bound_list.
+	 */
+	struct list_head userfault_list;
+
+	struct mutex error_mutex;
+	struct drm_mm_node error_capture;
+	struct drm_mm_node uc_fw;
+};
+
+struct i915_ppgtt {
+	struct i915_address_space vm;
+
+	struct i915_page_directory *pd;
+};
+
+#define i915_is_ggtt(vm) ((vm)->is_ggtt)
+#define i915_is_dpt(vm) ((vm)->is_dpt)
+#define i915_is_ggtt_or_dpt(vm) (i915_is_ggtt(vm) || i915_is_dpt(vm))
+
+bool intel_vm_no_concurrent_access_wa(struct drm_i915_private *i915);
+
+int __must_check
+i915_vm_lock_objects(struct i915_address_space *vm, struct i915_gem_ww_ctx *ww);
+
+static inline bool
+i915_vm_is_4lvl(const struct i915_address_space *vm)
+{
+	return (vm->total - 1) >> 32;
+}
+
+static inline bool
+i915_vm_has_scratch_64K(struct i915_address_space *vm)
+{
+	return vm->scratch_order == get_order(I915_GTT_PAGE_SIZE_64K);
+}
+
+static inline u64 i915_vm_min_alignment(struct i915_address_space *vm,
+					enum intel_memory_type type)
+{
+	/* avoid INTEL_MEMORY_MOCK overflow */
+	if ((int)type >= ARRAY_SIZE(vm->min_alignment))
+		type = INTEL_MEMORY_SYSTEM;
+
+	return vm->min_alignment[type];
+}
+
+static inline u64 i915_vm_obj_min_alignment(struct i915_address_space *vm,
+					    struct drm_i915_gem_object  *obj)
+{
+	struct intel_memory_region *mr = READ_ONCE(obj->mm.region);
+	enum intel_memory_type type = mr ? mr->type : INTEL_MEMORY_SYSTEM;
+
+	return i915_vm_min_alignment(vm, type);
+}
+
+static inline bool
+i915_vm_has_cache_coloring(struct i915_address_space *vm)
+{
+	return i915_is_ggtt(vm) && vm->mm.color_adjust;
+}
+
+static inline struct i915_ggtt *
+i915_vm_to_ggtt(struct i915_address_space *vm)
+{
+	BUILD_BUG_ON(offsetof(struct i915_ggtt, vm));
+	GEM_BUG_ON(!i915_is_ggtt(vm));
+	return container_of(vm, struct i915_ggtt, vm);
+}
+
+static inline struct i915_ppgtt *
+i915_vm_to_ppgtt(struct i915_address_space *vm)
+{
+	BUILD_BUG_ON(offsetof(struct i915_ppgtt, vm));
+	GEM_BUG_ON(i915_is_ggtt_or_dpt(vm));
+	return container_of(vm, struct i915_ppgtt, vm);
+}
+
+static inline struct i915_address_space *
+i915_vm_get(struct i915_address_space *vm)
+{
+	kref_get(&vm->ref);
+	return vm;
+}
+
+static inline struct i915_address_space *
+i915_vm_tryget(struct i915_address_space *vm)
+{
+	return kref_get_unless_zero(&vm->ref) ? vm : NULL;
+}
+
+static inline void assert_vm_alive(struct i915_address_space *vm)
+{
+	GEM_BUG_ON(!kref_read(&vm->ref));
+}
+
+/**
+ * i915_vm_resv_get - Obtain a reference on the vm's reservation lock
+ * @vm: The vm whose reservation lock we want to share.
+ *
+ * Return: A pointer to the vm's reservation lock.
+ */
+static inline struct dma_resv *i915_vm_resv_get(struct i915_address_space *vm)
+{
+	kref_get(&vm->resv_ref);
+	return &vm->_resv;
+}
+
+void i915_vm_release(struct kref *kref);
+
+void i915_vm_resv_release(struct kref *kref);
+
+static inline void i915_vm_put(struct i915_address_space *vm)
+{
+	kref_put(&vm->ref, i915_vm_release);
+}
+
+/**
+ * i915_vm_resv_put - Release a reference on the vm's reservation lock
+ * @resv: Pointer to a reservation lock obtained from i915_vm_resv_get()
+ */
+static inline void i915_vm_resv_put(struct i915_address_space *vm)
+{
+	kref_put(&vm->resv_ref, i915_vm_resv_release);
+}
+
+void i915_address_space_init(struct i915_address_space *vm, int subclass);
+void i915_address_space_fini(struct i915_address_space *vm);
+
+static inline u32 i915_pte_index(u64 address, unsigned int pde_shift)
+{
+	const u32 mask = NUM_PTE(pde_shift) - 1;
+
+	return (address >> PAGE_SHIFT) & mask;
+}
+
+/*
+ * Helper to counts the number of PTEs within the given length. This count
+ * does not cross a page table boundary, so the max value would be
+ * GEN6_PTES for GEN6, and GEN8_PTES for GEN8.
+ */
+static inline u32 i915_pte_count(u64 addr, u64 length, unsigned int pde_shift)
+{
+	const u64 mask = ~((1ULL << pde_shift) - 1);
+	u64 end;
+
+	GEM_BUG_ON(length == 0);
+	GEM_BUG_ON(offset_in_page(addr | length));
+
+	end = addr + length;
+
+	if ((addr & mask) != (end & mask))
+		return NUM_PTE(pde_shift) - i915_pte_index(addr, pde_shift);
+
+	return i915_pte_index(end, pde_shift) - i915_pte_index(addr, pde_shift);
+}
+
+static inline u32 i915_pde_index(u64 addr, u32 shift)
+{
+	return (addr >> shift) & I915_PDE_MASK;
+}
+
+static inline struct i915_page_table *
+i915_pt_entry(const struct i915_page_directory * const pd,
+	      const unsigned short n)
+{
+	return pd->entry[n];
+}
+
+static inline struct i915_page_directory *
+i915_pd_entry(const struct i915_page_directory * const pdp,
+	      const unsigned short n)
+{
+	return pdp->entry[n];
+}
+
+static inline dma_addr_t
+i915_page_dir_dma_addr(const struct i915_ppgtt *ppgtt, const unsigned int n)
+{
+	struct i915_page_table *pt = ppgtt->pd->entry[n];
+
+	return __px_dma(pt ? px_base(pt) : ppgtt->vm.scratch[ppgtt->vm.top]);
+}
+
+void ppgtt_init(struct i915_ppgtt *ppgtt, struct intel_gt *gt,
+		unsigned long lmem_pt_obj_flags);
+void intel_ggtt_bind_vma(struct i915_address_space *vm,
+			 struct i915_vm_pt_stash *stash,
+			 struct i915_vma_resource *vma_res,
+			 enum i915_cache_level cache_level,
+			 u32 flags);
+void intel_ggtt_unbind_vma(struct i915_address_space *vm,
+			   struct i915_vma_resource *vma_res);
+
+int i915_ggtt_probe_hw(struct drm_i915_private *i915);
+int i915_ggtt_init_hw(struct drm_i915_private *i915);
+int i915_ggtt_enable_hw(struct drm_i915_private *i915);
+void i915_ggtt_enable_guc(struct i915_ggtt *ggtt);
+void i915_ggtt_disable_guc(struct i915_ggtt *ggtt);
+int i915_init_ggtt(struct drm_i915_private *i915);
+void i915_ggtt_driver_release(struct drm_i915_private *i915);
+void i915_ggtt_driver_late_release(struct drm_i915_private *i915);
+
+static inline bool i915_ggtt_has_aperture(const struct i915_ggtt *ggtt)
+{
+	return ggtt->mappable_end > 0;
+}
+
+int i915_ppgtt_init_hw(struct intel_gt *gt);
+
+struct i915_ppgtt *i915_ppgtt_create(struct intel_gt *gt,
+				     unsigned long lmem_pt_obj_flags);
+
+void i915_ggtt_suspend_vm(struct i915_address_space *vm);
+bool i915_ggtt_resume_vm(struct i915_address_space *vm);
+void i915_ggtt_suspend(struct i915_ggtt *gtt);
+void i915_ggtt_resume(struct i915_ggtt *ggtt);
+
+/**
+ * i915_ggtt_mark_pte_lost - Mark ggtt ptes as lost or clear such a marking
+ * @i915 The device private.
+ * @val whether the ptes should be marked as lost.
+ *
+ * In some cases pte content is retained across suspend, but typically lost
+ * across hibernate. Typically they should be marked as lost on
+ * hibernation restore and such marking cleared on suspend.
+ */
+void i915_ggtt_mark_pte_lost(struct drm_i915_private *i915, bool val);
+
+void
+fill_page_dma(struct drm_i915_gem_object *p, const u64 val, unsigned int count);
+
+#define fill_px(px, v) fill_page_dma(px_base(px), (v), PAGE_SIZE / sizeof(u64))
+#define fill32_px(px, v) do {						\
+	u64 v__ = lower_32_bits(v);					\
+	fill_px((px), v__ << 32 | v__);					\
+} while (0)
+
+int setup_scratch_page(struct i915_address_space *vm);
+void free_scratch(struct i915_address_space *vm);
+
+struct drm_i915_gem_object *alloc_pt_dma(struct i915_address_space *vm, int sz);
+struct drm_i915_gem_object *alloc_pt_lmem(struct i915_address_space *vm, int sz);
+struct i915_page_table *alloc_pt(struct i915_address_space *vm, int sz);
+struct i915_page_directory *alloc_pd(struct i915_address_space *vm);
+struct i915_page_directory *__alloc_pd(int npde);
+
+int map_pt_dma(struct i915_address_space *vm, struct drm_i915_gem_object *obj);
+int map_pt_dma_locked(struct i915_address_space *vm, struct drm_i915_gem_object *obj);
+
+void free_px(struct i915_address_space *vm,
+	     struct i915_page_table *pt, int lvl);
+#define free_pt(vm, px) free_px(vm, px, 0)
+#define free_pd(vm, px) free_px(vm, px_pt(px), 1)
+
+void
+__set_pd_entry(struct i915_page_directory * const pd,
+	       const unsigned short idx,
+	       struct i915_page_table *pt,
+	       u64 (*encode)(const dma_addr_t, const enum i915_cache_level));
+
+#define set_pd_entry(pd, idx, to) \
+	__set_pd_entry((pd), (idx), px_pt(to), gen8_pde_encode)
+
+void
+clear_pd_entry(struct i915_page_directory * const pd,
+	       const unsigned short idx,
+	       const struct drm_i915_gem_object * const scratch);
+
+bool
+release_pd_entry(struct i915_page_directory * const pd,
+		 const unsigned short idx,
+		 struct i915_page_table * const pt,
+		 const struct drm_i915_gem_object * const scratch);
+void gen6_ggtt_invalidate(struct i915_ggtt *ggtt);
+
+void ppgtt_bind_vma(struct i915_address_space *vm,
+		    struct i915_vm_pt_stash *stash,
+		    struct i915_vma_resource *vma_res,
+		    enum i915_cache_level cache_level,
+		    u32 flags);
+void ppgtt_unbind_vma(struct i915_address_space *vm,
+		      struct i915_vma_resource *vma_res);
+
+void gtt_write_workarounds(struct intel_gt *gt);
+
+void setup_private_pat(struct intel_uncore *uncore);
+
+int i915_vm_alloc_pt_stash(struct i915_address_space *vm,
+			   struct i915_vm_pt_stash *stash,
+			   u64 size);
+int i915_vm_map_pt_stash(struct i915_address_space *vm,
+			 struct i915_vm_pt_stash *stash);
+void i915_vm_free_pt_stash(struct i915_address_space *vm,
+			   struct i915_vm_pt_stash *stash);
+
+struct i915_vma *
+__vm_create_scratch_for_read(struct i915_address_space *vm, unsigned long size);
+
+struct i915_vma *
+__vm_create_scratch_for_read_pinned(struct i915_address_space *vm, unsigned long size);
+
+static inline struct sgt_dma {
+	struct scatterlist *sg;
+	dma_addr_t dma, max;
+} sgt_dma(struct i915_vma_resource *vma_res) {
+	struct scatterlist *sg = vma_res->bi.pages->sgl;
+	dma_addr_t addr = sg_dma_address(sg);
+
+	return (struct sgt_dma){ sg, addr, addr + sg_dma_len(sg) };
+}
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_hwconfig.h b/drivers/gpu/drm/i915/gt/intel_hwconfig.h
new file mode 100644
index 000000000..322290780
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_hwconfig.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef _INTEL_HWCONFIG_H_
+#define _INTEL_HWCONFIG_H_
+
+#include <linux/types.h>
+
+struct intel_gt;
+
+struct intel_hwconfig {
+	u32 size;
+	void *ptr;
+};
+
+int intel_gt_init_hwconfig(struct intel_gt *gt);
+void intel_gt_fini_hwconfig(struct intel_gt *gt);
+
+#endif /* _INTEL_HWCONFIG_H_ */
diff --git a/drivers/gpu/drm/i915/gt/intel_llc.c b/drivers/gpu/drm/i915/gt/intel_llc.c
new file mode 100644
index 000000000..1d19c073b
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_llc.c
@@ -0,0 +1,162 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <asm/tsc.h>
+#include <linux/cpufreq.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_gt.h"
+#include "intel_llc.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pcode.h"
+#include "intel_rps.h"
+
+struct ia_constants {
+	unsigned int min_gpu_freq;
+	unsigned int max_gpu_freq;
+
+	unsigned int min_ring_freq;
+	unsigned int max_ia_freq;
+};
+
+static struct intel_gt *llc_to_gt(struct intel_llc *llc)
+{
+	return container_of(llc, struct intel_gt, llc);
+}
+
+static unsigned int cpu_max_MHz(void)
+{
+	struct cpufreq_policy *policy;
+	unsigned int max_khz;
+
+	policy = cpufreq_cpu_get(0);
+	if (policy) {
+		max_khz = policy->cpuinfo.max_freq;
+		cpufreq_cpu_put(policy);
+	} else {
+		/*
+		 * Default to measured freq if none found, PCU will ensure we
+		 * don't go over
+		 */
+		max_khz = tsc_khz;
+	}
+
+	return max_khz / 1000;
+}
+
+static bool get_ia_constants(struct intel_llc *llc,
+			     struct ia_constants *consts)
+{
+	struct drm_i915_private *i915 = llc_to_gt(llc)->i915;
+	struct intel_rps *rps = &llc_to_gt(llc)->rps;
+
+	if (!HAS_LLC(i915) || IS_DGFX(i915))
+		return false;
+
+	consts->max_ia_freq = cpu_max_MHz();
+
+	consts->min_ring_freq =
+		intel_uncore_read(llc_to_gt(llc)->uncore, DCLK) & 0xf;
+	/* convert DDR frequency from units of 266.6MHz to bandwidth */
+	consts->min_ring_freq = mult_frac(consts->min_ring_freq, 8, 3);
+
+	consts->min_gpu_freq = intel_rps_get_min_raw_freq(rps);
+	consts->max_gpu_freq = intel_rps_get_max_raw_freq(rps);
+
+	return true;
+}
+
+static void calc_ia_freq(struct intel_llc *llc,
+			 unsigned int gpu_freq,
+			 const struct ia_constants *consts,
+			 unsigned int *out_ia_freq,
+			 unsigned int *out_ring_freq)
+{
+	struct drm_i915_private *i915 = llc_to_gt(llc)->i915;
+	const int diff = consts->max_gpu_freq - gpu_freq;
+	unsigned int ia_freq = 0, ring_freq = 0;
+
+	if (GRAPHICS_VER(i915) >= 9) {
+		/*
+		 * ring_freq = 2 * GT. ring_freq is in 100MHz units
+		 * No floor required for ring frequency on SKL.
+		 */
+		ring_freq = gpu_freq;
+	} else if (GRAPHICS_VER(i915) >= 8) {
+		/* max(2 * GT, DDR). NB: GT is 50MHz units */
+		ring_freq = max(consts->min_ring_freq, gpu_freq);
+	} else if (IS_HASWELL(i915)) {
+		ring_freq = mult_frac(gpu_freq, 5, 4);
+		ring_freq = max(consts->min_ring_freq, ring_freq);
+		/* leave ia_freq as the default, chosen by cpufreq */
+	} else {
+		const int min_freq = 15;
+		const int scale = 180;
+
+		/*
+		 * On older processors, there is no separate ring
+		 * clock domain, so in order to boost the bandwidth
+		 * of the ring, we need to upclock the CPU (ia_freq).
+		 *
+		 * For GPU frequencies less than 750MHz,
+		 * just use the lowest ring freq.
+		 */
+		if (gpu_freq < min_freq)
+			ia_freq = 800;
+		else
+			ia_freq = consts->max_ia_freq - diff * scale / 2;
+		ia_freq = DIV_ROUND_CLOSEST(ia_freq, 100);
+	}
+
+	*out_ia_freq = ia_freq;
+	*out_ring_freq = ring_freq;
+}
+
+static void gen6_update_ring_freq(struct intel_llc *llc)
+{
+	struct ia_constants consts;
+	unsigned int gpu_freq;
+
+	if (!get_ia_constants(llc, &consts))
+		return;
+
+	/*
+	 * Although this is unlikely on any platform during initialization,
+	 * let's ensure we don't get accidentally into infinite loop
+	 */
+	if (consts.max_gpu_freq <= consts.min_gpu_freq)
+		return;
+	/*
+	 * For each potential GPU frequency, load a ring frequency we'd like
+	 * to use for memory access.  We do this by specifying the IA frequency
+	 * the PCU should use as a reference to determine the ring frequency.
+	 */
+	for (gpu_freq = consts.max_gpu_freq;
+	     gpu_freq >= consts.min_gpu_freq;
+	     gpu_freq--) {
+		unsigned int ia_freq, ring_freq;
+
+		calc_ia_freq(llc, gpu_freq, &consts, &ia_freq, &ring_freq);
+		snb_pcode_write(llc_to_gt(llc)->uncore, GEN6_PCODE_WRITE_MIN_FREQ_TABLE,
+				ia_freq << GEN6_PCODE_FREQ_IA_RATIO_SHIFT |
+				ring_freq << GEN6_PCODE_FREQ_RING_RATIO_SHIFT |
+				gpu_freq);
+	}
+}
+
+void intel_llc_enable(struct intel_llc *llc)
+{
+	gen6_update_ring_freq(llc);
+}
+
+void intel_llc_disable(struct intel_llc *llc)
+{
+	/* Currently there is no HW configuration to be done to disable. */
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_llc.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_llc.h b/drivers/gpu/drm/i915/gt/intel_llc.h
new file mode 100644
index 000000000..0e2e3871c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_llc.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_LLC_H
+#define INTEL_LLC_H
+
+struct intel_llc;
+
+void intel_llc_enable(struct intel_llc *llc);
+void intel_llc_disable(struct intel_llc *llc);
+
+#endif /* INTEL_LLC_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_llc_types.h b/drivers/gpu/drm/i915/gt/intel_llc_types.h
new file mode 100644
index 000000000..ca5bdf166
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_llc_types.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_LLC_TYPES_H
+#define INTEL_LLC_TYPES_H
+
+struct intel_llc {
+};
+
+#endif /* INTEL_LLC_TYPES_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c
new file mode 100644
index 000000000..7eb01ff17
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
@@ -0,0 +1,1848 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#include "gem/i915_gem_lmem.h"
+
+#include "gen8_engine_cs.h"
+#include "i915_drv.h"
+#include "i915_perf.h"
+#include "i915_reg.h"
+#include "intel_context.h"
+#include "intel_engine.h"
+#include "intel_engine_regs.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt.h"
+#include "intel_gt_regs.h"
+#include "intel_lrc.h"
+#include "intel_lrc_reg.h"
+#include "intel_ring.h"
+#include "shmem_utils.h"
+
+static void set_offsets(u32 *regs,
+			const u8 *data,
+			const struct intel_engine_cs *engine,
+			bool close)
+#define NOP(x) (BIT(7) | (x))
+#define LRI(count, flags) ((flags) << 6 | (count) | BUILD_BUG_ON_ZERO(count >= BIT(6)))
+#define POSTED BIT(0)
+#define REG(x) (((x) >> 2) | BUILD_BUG_ON_ZERO(x >= 0x200))
+#define REG16(x) \
+	(((x) >> 9) | BIT(7) | BUILD_BUG_ON_ZERO(x >= 0x10000)), \
+	(((x) >> 2) & 0x7f)
+#define END 0
+{
+	const u32 base = engine->mmio_base;
+
+	while (*data) {
+		u8 count, flags;
+
+		if (*data & BIT(7)) { /* skip */
+			count = *data++ & ~BIT(7);
+			regs += count;
+			continue;
+		}
+
+		count = *data & 0x3f;
+		flags = *data >> 6;
+		data++;
+
+		*regs = MI_LOAD_REGISTER_IMM(count);
+		if (flags & POSTED)
+			*regs |= MI_LRI_FORCE_POSTED;
+		if (GRAPHICS_VER(engine->i915) >= 11)
+			*regs |= MI_LRI_LRM_CS_MMIO;
+		regs++;
+
+		GEM_BUG_ON(!count);
+		do {
+			u32 offset = 0;
+			u8 v;
+
+			do {
+				v = *data++;
+				offset <<= 7;
+				offset |= v & ~BIT(7);
+			} while (v & BIT(7));
+
+			regs[0] = base + (offset << 2);
+			regs += 2;
+		} while (--count);
+	}
+
+	if (close) {
+		/* Close the batch; used mainly by live_lrc_layout() */
+		*regs = MI_BATCH_BUFFER_END;
+		if (GRAPHICS_VER(engine->i915) >= 11)
+			*regs |= BIT(0);
+	}
+}
+
+static const u8 gen8_xcs_offsets[] = {
+	NOP(1),
+	LRI(11, 0),
+	REG16(0x244),
+	REG(0x034),
+	REG(0x030),
+	REG(0x038),
+	REG(0x03c),
+	REG(0x168),
+	REG(0x140),
+	REG(0x110),
+	REG(0x11c),
+	REG(0x114),
+	REG(0x118),
+
+	NOP(9),
+	LRI(9, 0),
+	REG16(0x3a8),
+	REG16(0x28c),
+	REG16(0x288),
+	REG16(0x284),
+	REG16(0x280),
+	REG16(0x27c),
+	REG16(0x278),
+	REG16(0x274),
+	REG16(0x270),
+
+	NOP(13),
+	LRI(2, 0),
+	REG16(0x200),
+	REG(0x028),
+
+	END
+};
+
+static const u8 gen9_xcs_offsets[] = {
+	NOP(1),
+	LRI(14, POSTED),
+	REG16(0x244),
+	REG(0x034),
+	REG(0x030),
+	REG(0x038),
+	REG(0x03c),
+	REG(0x168),
+	REG(0x140),
+	REG(0x110),
+	REG(0x11c),
+	REG(0x114),
+	REG(0x118),
+	REG(0x1c0),
+	REG(0x1c4),
+	REG(0x1c8),
+
+	NOP(3),
+	LRI(9, POSTED),
+	REG16(0x3a8),
+	REG16(0x28c),
+	REG16(0x288),
+	REG16(0x284),
+	REG16(0x280),
+	REG16(0x27c),
+	REG16(0x278),
+	REG16(0x274),
+	REG16(0x270),
+
+	NOP(13),
+	LRI(1, POSTED),
+	REG16(0x200),
+
+	NOP(13),
+	LRI(44, POSTED),
+	REG(0x028),
+	REG(0x09c),
+	REG(0x0c0),
+	REG(0x178),
+	REG(0x17c),
+	REG16(0x358),
+	REG(0x170),
+	REG(0x150),
+	REG(0x154),
+	REG(0x158),
+	REG16(0x41c),
+	REG16(0x600),
+	REG16(0x604),
+	REG16(0x608),
+	REG16(0x60c),
+	REG16(0x610),
+	REG16(0x614),
+	REG16(0x618),
+	REG16(0x61c),
+	REG16(0x620),
+	REG16(0x624),
+	REG16(0x628),
+	REG16(0x62c),
+	REG16(0x630),
+	REG16(0x634),
+	REG16(0x638),
+	REG16(0x63c),
+	REG16(0x640),
+	REG16(0x644),
+	REG16(0x648),
+	REG16(0x64c),
+	REG16(0x650),
+	REG16(0x654),
+	REG16(0x658),
+	REG16(0x65c),
+	REG16(0x660),
+	REG16(0x664),
+	REG16(0x668),
+	REG16(0x66c),
+	REG16(0x670),
+	REG16(0x674),
+	REG16(0x678),
+	REG16(0x67c),
+	REG(0x068),
+
+	END
+};
+
+static const u8 gen12_xcs_offsets[] = {
+	NOP(1),
+	LRI(13, POSTED),
+	REG16(0x244),
+	REG(0x034),
+	REG(0x030),
+	REG(0x038),
+	REG(0x03c),
+	REG(0x168),
+	REG(0x140),
+	REG(0x110),
+	REG(0x1c0),
+	REG(0x1c4),
+	REG(0x1c8),
+	REG(0x180),
+	REG16(0x2b4),
+
+	NOP(5),
+	LRI(9, POSTED),
+	REG16(0x3a8),
+	REG16(0x28c),
+	REG16(0x288),
+	REG16(0x284),
+	REG16(0x280),
+	REG16(0x27c),
+	REG16(0x278),
+	REG16(0x274),
+	REG16(0x270),
+
+	END
+};
+
+static const u8 dg2_xcs_offsets[] = {
+	NOP(1),
+	LRI(15, POSTED),
+	REG16(0x244),
+	REG(0x034),
+	REG(0x030),
+	REG(0x038),
+	REG(0x03c),
+	REG(0x168),
+	REG(0x140),
+	REG(0x110),
+	REG(0x1c0),
+	REG(0x1c4),
+	REG(0x1c8),
+	REG(0x180),
+	REG16(0x2b4),
+	REG(0x120),
+	REG(0x124),
+
+	NOP(1),
+	LRI(9, POSTED),
+	REG16(0x3a8),
+	REG16(0x28c),
+	REG16(0x288),
+	REG16(0x284),
+	REG16(0x280),
+	REG16(0x27c),
+	REG16(0x278),
+	REG16(0x274),
+	REG16(0x270),
+
+	END
+};
+
+static const u8 gen8_rcs_offsets[] = {
+	NOP(1),
+	LRI(14, POSTED),
+	REG16(0x244),
+	REG(0x034),
+	REG(0x030),
+	REG(0x038),
+	REG(0x03c),
+	REG(0x168),
+	REG(0x140),
+	REG(0x110),
+	REG(0x11c),
+	REG(0x114),
+	REG(0x118),
+	REG(0x1c0),
+	REG(0x1c4),
+	REG(0x1c8),
+
+	NOP(3),
+	LRI(9, POSTED),
+	REG16(0x3a8),
+	REG16(0x28c),
+	REG16(0x288),
+	REG16(0x284),
+	REG16(0x280),
+	REG16(0x27c),
+	REG16(0x278),
+	REG16(0x274),
+	REG16(0x270),
+
+	NOP(13),
+	LRI(1, 0),
+	REG(0x0c8),
+
+	END
+};
+
+static const u8 gen9_rcs_offsets[] = {
+	NOP(1),
+	LRI(14, POSTED),
+	REG16(0x244),
+	REG(0x34),
+	REG(0x30),
+	REG(0x38),
+	REG(0x3c),
+	REG(0x168),
+	REG(0x140),
+	REG(0x110),
+	REG(0x11c),
+	REG(0x114),
+	REG(0x118),
+	REG(0x1c0),
+	REG(0x1c4),
+	REG(0x1c8),
+
+	NOP(3),
+	LRI(9, POSTED),
+	REG16(0x3a8),
+	REG16(0x28c),
+	REG16(0x288),
+	REG16(0x284),
+	REG16(0x280),
+	REG16(0x27c),
+	REG16(0x278),
+	REG16(0x274),
+	REG16(0x270),
+
+	NOP(13),
+	LRI(1, 0),
+	REG(0xc8),
+
+	NOP(13),
+	LRI(44, POSTED),
+	REG(0x28),
+	REG(0x9c),
+	REG(0xc0),
+	REG(0x178),
+	REG(0x17c),
+	REG16(0x358),
+	REG(0x170),
+	REG(0x150),
+	REG(0x154),
+	REG(0x158),
+	REG16(0x41c),
+	REG16(0x600),
+	REG16(0x604),
+	REG16(0x608),
+	REG16(0x60c),
+	REG16(0x610),
+	REG16(0x614),
+	REG16(0x618),
+	REG16(0x61c),
+	REG16(0x620),
+	REG16(0x624),
+	REG16(0x628),
+	REG16(0x62c),
+	REG16(0x630),
+	REG16(0x634),
+	REG16(0x638),
+	REG16(0x63c),
+	REG16(0x640),
+	REG16(0x644),
+	REG16(0x648),
+	REG16(0x64c),
+	REG16(0x650),
+	REG16(0x654),
+	REG16(0x658),
+	REG16(0x65c),
+	REG16(0x660),
+	REG16(0x664),
+	REG16(0x668),
+	REG16(0x66c),
+	REG16(0x670),
+	REG16(0x674),
+	REG16(0x678),
+	REG16(0x67c),
+	REG(0x68),
+
+	END
+};
+
+static const u8 gen11_rcs_offsets[] = {
+	NOP(1),
+	LRI(15, POSTED),
+	REG16(0x244),
+	REG(0x034),
+	REG(0x030),
+	REG(0x038),
+	REG(0x03c),
+	REG(0x168),
+	REG(0x140),
+	REG(0x110),
+	REG(0x11c),
+	REG(0x114),
+	REG(0x118),
+	REG(0x1c0),
+	REG(0x1c4),
+	REG(0x1c8),
+	REG(0x180),
+
+	NOP(1),
+	LRI(9, POSTED),
+	REG16(0x3a8),
+	REG16(0x28c),
+	REG16(0x288),
+	REG16(0x284),
+	REG16(0x280),
+	REG16(0x27c),
+	REG16(0x278),
+	REG16(0x274),
+	REG16(0x270),
+
+	LRI(1, POSTED),
+	REG(0x1b0),
+
+	NOP(10),
+	LRI(1, 0),
+	REG(0x0c8),
+
+	END
+};
+
+static const u8 gen12_rcs_offsets[] = {
+	NOP(1),
+	LRI(13, POSTED),
+	REG16(0x244),
+	REG(0x034),
+	REG(0x030),
+	REG(0x038),
+	REG(0x03c),
+	REG(0x168),
+	REG(0x140),
+	REG(0x110),
+	REG(0x1c0),
+	REG(0x1c4),
+	REG(0x1c8),
+	REG(0x180),
+	REG16(0x2b4),
+
+	NOP(5),
+	LRI(9, POSTED),
+	REG16(0x3a8),
+	REG16(0x28c),
+	REG16(0x288),
+	REG16(0x284),
+	REG16(0x280),
+	REG16(0x27c),
+	REG16(0x278),
+	REG16(0x274),
+	REG16(0x270),
+
+	LRI(3, POSTED),
+	REG(0x1b0),
+	REG16(0x5a8),
+	REG16(0x5ac),
+
+	NOP(6),
+	LRI(1, 0),
+	REG(0x0c8),
+	NOP(3 + 9 + 1),
+
+	LRI(51, POSTED),
+	REG16(0x588),
+	REG16(0x588),
+	REG16(0x588),
+	REG16(0x588),
+	REG16(0x588),
+	REG16(0x588),
+	REG(0x028),
+	REG(0x09c),
+	REG(0x0c0),
+	REG(0x178),
+	REG(0x17c),
+	REG16(0x358),
+	REG(0x170),
+	REG(0x150),
+	REG(0x154),
+	REG(0x158),
+	REG16(0x41c),
+	REG16(0x600),
+	REG16(0x604),
+	REG16(0x608),
+	REG16(0x60c),
+	REG16(0x610),
+	REG16(0x614),
+	REG16(0x618),
+	REG16(0x61c),
+	REG16(0x620),
+	REG16(0x624),
+	REG16(0x628),
+	REG16(0x62c),
+	REG16(0x630),
+	REG16(0x634),
+	REG16(0x638),
+	REG16(0x63c),
+	REG16(0x640),
+	REG16(0x644),
+	REG16(0x648),
+	REG16(0x64c),
+	REG16(0x650),
+	REG16(0x654),
+	REG16(0x658),
+	REG16(0x65c),
+	REG16(0x660),
+	REG16(0x664),
+	REG16(0x668),
+	REG16(0x66c),
+	REG16(0x670),
+	REG16(0x674),
+	REG16(0x678),
+	REG16(0x67c),
+	REG(0x068),
+	REG(0x084),
+	NOP(1),
+
+	END
+};
+
+static const u8 xehp_rcs_offsets[] = {
+	NOP(1),
+	LRI(13, POSTED),
+	REG16(0x244),
+	REG(0x034),
+	REG(0x030),
+	REG(0x038),
+	REG(0x03c),
+	REG(0x168),
+	REG(0x140),
+	REG(0x110),
+	REG(0x1c0),
+	REG(0x1c4),
+	REG(0x1c8),
+	REG(0x180),
+	REG16(0x2b4),
+
+	NOP(5),
+	LRI(9, POSTED),
+	REG16(0x3a8),
+	REG16(0x28c),
+	REG16(0x288),
+	REG16(0x284),
+	REG16(0x280),
+	REG16(0x27c),
+	REG16(0x278),
+	REG16(0x274),
+	REG16(0x270),
+
+	LRI(3, POSTED),
+	REG(0x1b0),
+	REG16(0x5a8),
+	REG16(0x5ac),
+
+	NOP(6),
+	LRI(1, 0),
+	REG(0x0c8),
+
+	END
+};
+
+static const u8 dg2_rcs_offsets[] = {
+	NOP(1),
+	LRI(15, POSTED),
+	REG16(0x244),
+	REG(0x034),
+	REG(0x030),
+	REG(0x038),
+	REG(0x03c),
+	REG(0x168),
+	REG(0x140),
+	REG(0x110),
+	REG(0x1c0),
+	REG(0x1c4),
+	REG(0x1c8),
+	REG(0x180),
+	REG16(0x2b4),
+	REG(0x120),
+	REG(0x124),
+
+	NOP(1),
+	LRI(9, POSTED),
+	REG16(0x3a8),
+	REG16(0x28c),
+	REG16(0x288),
+	REG16(0x284),
+	REG16(0x280),
+	REG16(0x27c),
+	REG16(0x278),
+	REG16(0x274),
+	REG16(0x270),
+
+	LRI(3, POSTED),
+	REG(0x1b0),
+	REG16(0x5a8),
+	REG16(0x5ac),
+
+	NOP(6),
+	LRI(1, 0),
+	REG(0x0c8),
+
+	END
+};
+
+#undef END
+#undef REG16
+#undef REG
+#undef LRI
+#undef NOP
+
+static const u8 *reg_offsets(const struct intel_engine_cs *engine)
+{
+	/*
+	 * The gen12+ lists only have the registers we program in the basic
+	 * default state. We rely on the context image using relative
+	 * addressing to automatic fixup the register state between the
+	 * physical engines for virtual engine.
+	 */
+	GEM_BUG_ON(GRAPHICS_VER(engine->i915) >= 12 &&
+		   !intel_engine_has_relative_mmio(engine));
+
+	if (engine->flags & I915_ENGINE_HAS_RCS_REG_STATE) {
+		if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 55))
+			return dg2_rcs_offsets;
+		else if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 50))
+			return xehp_rcs_offsets;
+		else if (GRAPHICS_VER(engine->i915) >= 12)
+			return gen12_rcs_offsets;
+		else if (GRAPHICS_VER(engine->i915) >= 11)
+			return gen11_rcs_offsets;
+		else if (GRAPHICS_VER(engine->i915) >= 9)
+			return gen9_rcs_offsets;
+		else
+			return gen8_rcs_offsets;
+	} else {
+		if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 55))
+			return dg2_xcs_offsets;
+		else if (GRAPHICS_VER(engine->i915) >= 12)
+			return gen12_xcs_offsets;
+		else if (GRAPHICS_VER(engine->i915) >= 9)
+			return gen9_xcs_offsets;
+		else
+			return gen8_xcs_offsets;
+	}
+}
+
+static int lrc_ring_mi_mode(const struct intel_engine_cs *engine)
+{
+	if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 50))
+		return 0x70;
+	else if (GRAPHICS_VER(engine->i915) >= 12)
+		return 0x60;
+	else if (GRAPHICS_VER(engine->i915) >= 9)
+		return 0x54;
+	else if (engine->class == RENDER_CLASS)
+		return 0x58;
+	else
+		return -1;
+}
+
+static int lrc_ring_bb_offset(const struct intel_engine_cs *engine)
+{
+	if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 50))
+		return 0x80;
+	else if (GRAPHICS_VER(engine->i915) >= 12)
+		return 0x70;
+	else if (GRAPHICS_VER(engine->i915) >= 9)
+		return 0x64;
+	else if (GRAPHICS_VER(engine->i915) >= 8 &&
+		 engine->class == RENDER_CLASS)
+		return 0xc4;
+	else
+		return -1;
+}
+
+static int lrc_ring_gpr0(const struct intel_engine_cs *engine)
+{
+	if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 50))
+		return 0x84;
+	else if (GRAPHICS_VER(engine->i915) >= 12)
+		return 0x74;
+	else if (GRAPHICS_VER(engine->i915) >= 9)
+		return 0x68;
+	else if (engine->class == RENDER_CLASS)
+		return 0xd8;
+	else
+		return -1;
+}
+
+static int lrc_ring_wa_bb_per_ctx(const struct intel_engine_cs *engine)
+{
+	if (GRAPHICS_VER(engine->i915) >= 12)
+		return 0x12;
+	else if (GRAPHICS_VER(engine->i915) >= 9 || engine->class == RENDER_CLASS)
+		return 0x18;
+	else
+		return -1;
+}
+
+static int lrc_ring_indirect_ptr(const struct intel_engine_cs *engine)
+{
+	int x;
+
+	x = lrc_ring_wa_bb_per_ctx(engine);
+	if (x < 0)
+		return x;
+
+	return x + 2;
+}
+
+static int lrc_ring_indirect_offset(const struct intel_engine_cs *engine)
+{
+	int x;
+
+	x = lrc_ring_indirect_ptr(engine);
+	if (x < 0)
+		return x;
+
+	return x + 2;
+}
+
+static int lrc_ring_cmd_buf_cctl(const struct intel_engine_cs *engine)
+{
+
+	if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 50))
+		/*
+		 * Note that the CSFE context has a dummy slot for CMD_BUF_CCTL
+		 * simply to match the RCS context image layout.
+		 */
+		return 0xc6;
+	else if (engine->class != RENDER_CLASS)
+		return -1;
+	else if (GRAPHICS_VER(engine->i915) >= 12)
+		return 0xb6;
+	else if (GRAPHICS_VER(engine->i915) >= 11)
+		return 0xaa;
+	else
+		return -1;
+}
+
+static u32
+lrc_ring_indirect_offset_default(const struct intel_engine_cs *engine)
+{
+	switch (GRAPHICS_VER(engine->i915)) {
+	default:
+		MISSING_CASE(GRAPHICS_VER(engine->i915));
+		fallthrough;
+	case 12:
+		return GEN12_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
+	case 11:
+		return GEN11_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
+	case 9:
+		return GEN9_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
+	case 8:
+		return GEN8_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT;
+	}
+}
+
+static void
+lrc_setup_indirect_ctx(u32 *regs,
+		       const struct intel_engine_cs *engine,
+		       u32 ctx_bb_ggtt_addr,
+		       u32 size)
+{
+	GEM_BUG_ON(!size);
+	GEM_BUG_ON(!IS_ALIGNED(size, CACHELINE_BYTES));
+	GEM_BUG_ON(lrc_ring_indirect_ptr(engine) == -1);
+	regs[lrc_ring_indirect_ptr(engine) + 1] =
+		ctx_bb_ggtt_addr | (size / CACHELINE_BYTES);
+
+	GEM_BUG_ON(lrc_ring_indirect_offset(engine) == -1);
+	regs[lrc_ring_indirect_offset(engine) + 1] =
+		lrc_ring_indirect_offset_default(engine) << 6;
+}
+
+static void init_common_regs(u32 * const regs,
+			     const struct intel_context *ce,
+			     const struct intel_engine_cs *engine,
+			     bool inhibit)
+{
+	u32 ctl;
+	int loc;
+
+	ctl = _MASKED_BIT_ENABLE(CTX_CTRL_INHIBIT_SYN_CTX_SWITCH);
+	ctl |= _MASKED_BIT_DISABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT);
+	if (inhibit)
+		ctl |= CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT;
+	if (GRAPHICS_VER(engine->i915) < 11)
+		ctl |= _MASKED_BIT_DISABLE(CTX_CTRL_ENGINE_CTX_SAVE_INHIBIT |
+					   CTX_CTRL_RS_CTX_ENABLE);
+	regs[CTX_CONTEXT_CONTROL] = ctl;
+
+	regs[CTX_TIMESTAMP] = ce->stats.runtime.last;
+
+	loc = lrc_ring_bb_offset(engine);
+	if (loc != -1)
+		regs[loc + 1] = 0;
+}
+
+static void init_wa_bb_regs(u32 * const regs,
+			    const struct intel_engine_cs *engine)
+{
+	const struct i915_ctx_workarounds * const wa_ctx = &engine->wa_ctx;
+
+	if (wa_ctx->per_ctx.size) {
+		const u32 ggtt_offset = i915_ggtt_offset(wa_ctx->vma);
+
+		GEM_BUG_ON(lrc_ring_wa_bb_per_ctx(engine) == -1);
+		regs[lrc_ring_wa_bb_per_ctx(engine) + 1] =
+			(ggtt_offset + wa_ctx->per_ctx.offset) | 0x01;
+	}
+
+	if (wa_ctx->indirect_ctx.size) {
+		lrc_setup_indirect_ctx(regs, engine,
+				       i915_ggtt_offset(wa_ctx->vma) +
+				       wa_ctx->indirect_ctx.offset,
+				       wa_ctx->indirect_ctx.size);
+	}
+}
+
+static void init_ppgtt_regs(u32 *regs, const struct i915_ppgtt *ppgtt)
+{
+	if (i915_vm_is_4lvl(&ppgtt->vm)) {
+		/* 64b PPGTT (48bit canonical)
+		 * PDP0_DESCRIPTOR contains the base address to PML4 and
+		 * other PDP Descriptors are ignored.
+		 */
+		ASSIGN_CTX_PML4(ppgtt, regs);
+	} else {
+		ASSIGN_CTX_PDP(ppgtt, regs, 3);
+		ASSIGN_CTX_PDP(ppgtt, regs, 2);
+		ASSIGN_CTX_PDP(ppgtt, regs, 1);
+		ASSIGN_CTX_PDP(ppgtt, regs, 0);
+	}
+}
+
+static struct i915_ppgtt *vm_alias(struct i915_address_space *vm)
+{
+	if (i915_is_ggtt(vm))
+		return i915_vm_to_ggtt(vm)->alias;
+	else
+		return i915_vm_to_ppgtt(vm);
+}
+
+static void __reset_stop_ring(u32 *regs, const struct intel_engine_cs *engine)
+{
+	int x;
+
+	x = lrc_ring_mi_mode(engine);
+	if (x != -1) {
+		regs[x + 1] &= ~STOP_RING;
+		regs[x + 1] |= STOP_RING << 16;
+	}
+}
+
+static void __lrc_init_regs(u32 *regs,
+			    const struct intel_context *ce,
+			    const struct intel_engine_cs *engine,
+			    bool inhibit)
+{
+	/*
+	 * A context is actually a big batch buffer with several
+	 * MI_LOAD_REGISTER_IMM commands followed by (reg, value) pairs. The
+	 * values we are setting here are only for the first context restore:
+	 * on a subsequent save, the GPU will recreate this batchbuffer with new
+	 * values (including all the missing MI_LOAD_REGISTER_IMM commands that
+	 * we are not initializing here).
+	 *
+	 * Must keep consistent with virtual_update_register_offsets().
+	 */
+
+	if (inhibit)
+		memset(regs, 0, PAGE_SIZE);
+
+	set_offsets(regs, reg_offsets(engine), engine, inhibit);
+
+	init_common_regs(regs, ce, engine, inhibit);
+	init_ppgtt_regs(regs, vm_alias(ce->vm));
+
+	init_wa_bb_regs(regs, engine);
+
+	__reset_stop_ring(regs, engine);
+}
+
+void lrc_init_regs(const struct intel_context *ce,
+		   const struct intel_engine_cs *engine,
+		   bool inhibit)
+{
+	__lrc_init_regs(ce->lrc_reg_state, ce, engine, inhibit);
+}
+
+void lrc_reset_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine)
+{
+	__reset_stop_ring(ce->lrc_reg_state, engine);
+}
+
+static void
+set_redzone(void *vaddr, const struct intel_engine_cs *engine)
+{
+	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		return;
+
+	vaddr += engine->context_size;
+
+	memset(vaddr, CONTEXT_REDZONE, I915_GTT_PAGE_SIZE);
+}
+
+static void
+check_redzone(const void *vaddr, const struct intel_engine_cs *engine)
+{
+	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		return;
+
+	vaddr += engine->context_size;
+
+	if (memchr_inv(vaddr, CONTEXT_REDZONE, I915_GTT_PAGE_SIZE))
+		drm_err_once(&engine->i915->drm,
+			     "%s context redzone overwritten!\n",
+			     engine->name);
+}
+
+static u32 context_wa_bb_offset(const struct intel_context *ce)
+{
+	return PAGE_SIZE * ce->wa_bb_page;
+}
+
+static u32 *context_indirect_bb(const struct intel_context *ce)
+{
+	void *ptr;
+
+	GEM_BUG_ON(!ce->wa_bb_page);
+
+	ptr = ce->lrc_reg_state;
+	ptr -= LRC_STATE_OFFSET; /* back to start of context image */
+	ptr += context_wa_bb_offset(ce);
+
+	return ptr;
+}
+
+void lrc_init_state(struct intel_context *ce,
+		    struct intel_engine_cs *engine,
+		    void *state)
+{
+	bool inhibit = true;
+
+	set_redzone(state, engine);
+
+	if (engine->default_state) {
+		shmem_read(engine->default_state, 0,
+			   state, engine->context_size);
+		__set_bit(CONTEXT_VALID_BIT, &ce->flags);
+		inhibit = false;
+	}
+
+	/* Clear the ppHWSP (inc. per-context counters) */
+	memset(state, 0, PAGE_SIZE);
+
+	/* Clear the indirect wa and storage */
+	if (ce->wa_bb_page)
+		memset(state + context_wa_bb_offset(ce), 0, PAGE_SIZE);
+
+	/*
+	 * The second page of the context object contains some registers which
+	 * must be set up prior to the first execution.
+	 */
+	__lrc_init_regs(state + LRC_STATE_OFFSET, ce, engine, inhibit);
+}
+
+u32 lrc_indirect_bb(const struct intel_context *ce)
+{
+	return i915_ggtt_offset(ce->state) + context_wa_bb_offset(ce);
+}
+
+static u32 *setup_predicate_disable_wa(const struct intel_context *ce, u32 *cs)
+{
+	/* If predication is active, this will be noop'ed */
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT | (4 - 2);
+	*cs++ = lrc_indirect_bb(ce) + DG2_PREDICATE_RESULT_WA;
+	*cs++ = 0;
+	*cs++ = 0; /* No predication */
+
+	/* predicated end, only terminates if SET_PREDICATE_RESULT:0 is clear */
+	*cs++ = MI_BATCH_BUFFER_END | BIT(15);
+	*cs++ = MI_SET_PREDICATE | MI_SET_PREDICATE_DISABLE;
+
+	/* Instructions are no longer predicated (disabled), we can proceed */
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT | (4 - 2);
+	*cs++ = lrc_indirect_bb(ce) + DG2_PREDICATE_RESULT_WA;
+	*cs++ = 0;
+	*cs++ = 1; /* enable predication before the next BB */
+
+	*cs++ = MI_BATCH_BUFFER_END;
+	GEM_BUG_ON(offset_in_page(cs) > DG2_PREDICATE_RESULT_WA);
+
+	return cs;
+}
+
+static struct i915_vma *
+__lrc_alloc_state(struct intel_context *ce, struct intel_engine_cs *engine)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u32 context_size;
+
+	context_size = round_up(engine->context_size, I915_GTT_PAGE_SIZE);
+
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		context_size += I915_GTT_PAGE_SIZE; /* for redzone */
+
+	if (GRAPHICS_VER(engine->i915) == 12) {
+		ce->wa_bb_page = context_size / PAGE_SIZE;
+		context_size += PAGE_SIZE;
+	}
+
+	if (intel_context_is_parent(ce) && intel_engine_uses_guc(engine)) {
+		ce->parallel.guc.parent_page = context_size / PAGE_SIZE;
+		context_size += PARENT_SCRATCH_SIZE;
+	}
+
+	obj = i915_gem_object_create_lmem(engine->i915, context_size,
+					  I915_BO_ALLOC_PM_VOLATILE);
+	if (IS_ERR(obj))
+		obj = i915_gem_object_create_shmem(engine->i915, context_size);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, &engine->gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return vma;
+	}
+
+	return vma;
+}
+
+static struct intel_timeline *
+pinned_timeline(struct intel_context *ce, struct intel_engine_cs *engine)
+{
+	struct intel_timeline *tl = fetch_and_zero(&ce->timeline);
+
+	return intel_timeline_create_from_engine(engine, page_unmask_bits(tl));
+}
+
+int lrc_alloc(struct intel_context *ce, struct intel_engine_cs *engine)
+{
+	struct intel_ring *ring;
+	struct i915_vma *vma;
+	int err;
+
+	GEM_BUG_ON(ce->state);
+
+	vma = __lrc_alloc_state(ce, engine);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	ring = intel_engine_create_ring(engine, ce->ring_size);
+	if (IS_ERR(ring)) {
+		err = PTR_ERR(ring);
+		goto err_vma;
+	}
+
+	if (!page_mask_bits(ce->timeline)) {
+		struct intel_timeline *tl;
+
+		/*
+		 * Use the static global HWSP for the kernel context, and
+		 * a dynamically allocated cacheline for everyone else.
+		 */
+		if (unlikely(ce->timeline))
+			tl = pinned_timeline(ce, engine);
+		else
+			tl = intel_timeline_create(engine->gt);
+		if (IS_ERR(tl)) {
+			err = PTR_ERR(tl);
+			goto err_ring;
+		}
+
+		ce->timeline = tl;
+	}
+
+	ce->ring = ring;
+	ce->state = vma;
+
+	return 0;
+
+err_ring:
+	intel_ring_put(ring);
+err_vma:
+	i915_vma_put(vma);
+	return err;
+}
+
+void lrc_reset(struct intel_context *ce)
+{
+	GEM_BUG_ON(!intel_context_is_pinned(ce));
+
+	intel_ring_reset(ce->ring, ce->ring->emit);
+
+	/* Scrub away the garbage */
+	lrc_init_regs(ce, ce->engine, true);
+	ce->lrc.lrca = lrc_update_regs(ce, ce->engine, ce->ring->tail);
+}
+
+int
+lrc_pre_pin(struct intel_context *ce,
+	    struct intel_engine_cs *engine,
+	    struct i915_gem_ww_ctx *ww,
+	    void **vaddr)
+{
+	GEM_BUG_ON(!ce->state);
+	GEM_BUG_ON(!i915_vma_is_pinned(ce->state));
+
+	*vaddr = i915_gem_object_pin_map(ce->state->obj,
+					 i915_coherent_map_type(ce->engine->i915,
+								ce->state->obj,
+								false) |
+					 I915_MAP_OVERRIDE);
+
+	return PTR_ERR_OR_ZERO(*vaddr);
+}
+
+int
+lrc_pin(struct intel_context *ce,
+	struct intel_engine_cs *engine,
+	void *vaddr)
+{
+	ce->lrc_reg_state = vaddr + LRC_STATE_OFFSET;
+
+	if (!__test_and_set_bit(CONTEXT_INIT_BIT, &ce->flags))
+		lrc_init_state(ce, engine, vaddr);
+
+	ce->lrc.lrca = lrc_update_regs(ce, engine, ce->ring->tail);
+	return 0;
+}
+
+void lrc_unpin(struct intel_context *ce)
+{
+	if (unlikely(ce->parallel.last_rq)) {
+		i915_request_put(ce->parallel.last_rq);
+		ce->parallel.last_rq = NULL;
+	}
+	check_redzone((void *)ce->lrc_reg_state - LRC_STATE_OFFSET,
+		      ce->engine);
+}
+
+void lrc_post_unpin(struct intel_context *ce)
+{
+	i915_gem_object_unpin_map(ce->state->obj);
+}
+
+void lrc_fini(struct intel_context *ce)
+{
+	if (!ce->state)
+		return;
+
+	intel_ring_put(fetch_and_zero(&ce->ring));
+	i915_vma_put(fetch_and_zero(&ce->state));
+}
+
+void lrc_destroy(struct kref *kref)
+{
+	struct intel_context *ce = container_of(kref, typeof(*ce), ref);
+
+	GEM_BUG_ON(!i915_active_is_idle(&ce->active));
+	GEM_BUG_ON(intel_context_is_pinned(ce));
+
+	lrc_fini(ce);
+
+	intel_context_fini(ce);
+	intel_context_free(ce);
+}
+
+static u32 *
+gen12_emit_timestamp_wa(const struct intel_context *ce, u32 *cs)
+{
+	*cs++ = MI_LOAD_REGISTER_MEM_GEN8 |
+		MI_SRM_LRM_GLOBAL_GTT |
+		MI_LRI_LRM_CS_MMIO;
+	*cs++ = i915_mmio_reg_offset(GEN8_RING_CS_GPR(0, 0));
+	*cs++ = i915_ggtt_offset(ce->state) + LRC_STATE_OFFSET +
+		CTX_TIMESTAMP * sizeof(u32);
+	*cs++ = 0;
+
+	*cs++ = MI_LOAD_REGISTER_REG |
+		MI_LRR_SOURCE_CS_MMIO |
+		MI_LRI_LRM_CS_MMIO;
+	*cs++ = i915_mmio_reg_offset(GEN8_RING_CS_GPR(0, 0));
+	*cs++ = i915_mmio_reg_offset(RING_CTX_TIMESTAMP(0));
+
+	*cs++ = MI_LOAD_REGISTER_REG |
+		MI_LRR_SOURCE_CS_MMIO |
+		MI_LRI_LRM_CS_MMIO;
+	*cs++ = i915_mmio_reg_offset(GEN8_RING_CS_GPR(0, 0));
+	*cs++ = i915_mmio_reg_offset(RING_CTX_TIMESTAMP(0));
+
+	return cs;
+}
+
+static u32 *
+gen12_emit_restore_scratch(const struct intel_context *ce, u32 *cs)
+{
+	GEM_BUG_ON(lrc_ring_gpr0(ce->engine) == -1);
+
+	*cs++ = MI_LOAD_REGISTER_MEM_GEN8 |
+		MI_SRM_LRM_GLOBAL_GTT |
+		MI_LRI_LRM_CS_MMIO;
+	*cs++ = i915_mmio_reg_offset(GEN8_RING_CS_GPR(0, 0));
+	*cs++ = i915_ggtt_offset(ce->state) + LRC_STATE_OFFSET +
+		(lrc_ring_gpr0(ce->engine) + 1) * sizeof(u32);
+	*cs++ = 0;
+
+	return cs;
+}
+
+static u32 *
+gen12_emit_cmd_buf_wa(const struct intel_context *ce, u32 *cs)
+{
+	GEM_BUG_ON(lrc_ring_cmd_buf_cctl(ce->engine) == -1);
+
+	*cs++ = MI_LOAD_REGISTER_MEM_GEN8 |
+		MI_SRM_LRM_GLOBAL_GTT |
+		MI_LRI_LRM_CS_MMIO;
+	*cs++ = i915_mmio_reg_offset(GEN8_RING_CS_GPR(0, 0));
+	*cs++ = i915_ggtt_offset(ce->state) + LRC_STATE_OFFSET +
+		(lrc_ring_cmd_buf_cctl(ce->engine) + 1) * sizeof(u32);
+	*cs++ = 0;
+
+	*cs++ = MI_LOAD_REGISTER_REG |
+		MI_LRR_SOURCE_CS_MMIO |
+		MI_LRI_LRM_CS_MMIO;
+	*cs++ = i915_mmio_reg_offset(GEN8_RING_CS_GPR(0, 0));
+	*cs++ = i915_mmio_reg_offset(RING_CMD_BUF_CCTL(0));
+
+	return cs;
+}
+
+/*
+ * On DG2 during context restore of a preempted context in GPGPU mode,
+ * RCS restore hang is detected. This is extremely timing dependent.
+ * To address this below sw wabb is implemented for DG2 A steppings.
+ */
+static u32 *
+dg2_emit_rcs_hang_wabb(const struct intel_context *ce, u32 *cs)
+{
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(GEN12_STATE_ACK_DEBUG);
+	*cs++ = 0x21;
+
+	*cs++ = MI_LOAD_REGISTER_REG;
+	*cs++ = i915_mmio_reg_offset(RING_NOPID(ce->engine->mmio_base));
+	*cs++ = i915_mmio_reg_offset(GEN12_CULLBIT1);
+
+	*cs++ = MI_LOAD_REGISTER_REG;
+	*cs++ = i915_mmio_reg_offset(RING_NOPID(ce->engine->mmio_base));
+	*cs++ = i915_mmio_reg_offset(GEN12_CULLBIT2);
+
+	return cs;
+}
+
+/*
+ * The bspec's tuning guide asks us to program a vertical watermark value of
+ * 0x3FF.  However this register is not saved/restored properly by the
+ * hardware, so we're required to apply the desired value via INDIRECT_CTX
+ * batch buffer to ensure the value takes effect properly.  All other bits
+ * in this register should remain at 0 (the hardware default).
+ */
+static u32 *
+dg2_emit_draw_watermark_setting(u32 *cs)
+{
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(DRAW_WATERMARK);
+	*cs++ = REG_FIELD_PREP(VERT_WM_VAL, 0x3FF);
+
+	return cs;
+}
+
+static u32 *
+gen12_emit_indirect_ctx_rcs(const struct intel_context *ce, u32 *cs)
+{
+	cs = gen12_emit_timestamp_wa(ce, cs);
+	cs = gen12_emit_cmd_buf_wa(ce, cs);
+	cs = gen12_emit_restore_scratch(ce, cs);
+
+	/* Wa_22011450934:dg2 */
+	if (IS_DG2_GRAPHICS_STEP(ce->engine->i915, G10, STEP_A0, STEP_B0) ||
+	    IS_DG2_GRAPHICS_STEP(ce->engine->i915, G11, STEP_A0, STEP_B0))
+		cs = dg2_emit_rcs_hang_wabb(ce, cs);
+
+	/* Wa_16013000631:dg2 */
+	if (IS_DG2_GRAPHICS_STEP(ce->engine->i915, G10, STEP_B0, STEP_C0) ||
+	    IS_DG2_G11(ce->engine->i915))
+		cs = gen8_emit_pipe_control(cs, PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE, 0);
+
+	cs = gen12_emit_aux_table_inv(ce->engine, cs);
+
+	/* Wa_16014892111 */
+	if (IS_DG2(ce->engine->i915))
+		cs = dg2_emit_draw_watermark_setting(cs);
+
+	return cs;
+}
+
+static u32 *
+gen12_emit_indirect_ctx_xcs(const struct intel_context *ce, u32 *cs)
+{
+	cs = gen12_emit_timestamp_wa(ce, cs);
+	cs = gen12_emit_restore_scratch(ce, cs);
+
+	/* Wa_16013000631:dg2 */
+	if (IS_DG2_GRAPHICS_STEP(ce->engine->i915, G10, STEP_B0, STEP_C0) ||
+	    IS_DG2_G11(ce->engine->i915))
+		if (ce->engine->class == COMPUTE_CLASS)
+			cs = gen8_emit_pipe_control(cs,
+						    PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE,
+						    0);
+
+	return gen12_emit_aux_table_inv(ce->engine, cs);
+}
+
+static void
+setup_indirect_ctx_bb(const struct intel_context *ce,
+		      const struct intel_engine_cs *engine,
+		      u32 *(*emit)(const struct intel_context *, u32 *))
+{
+	u32 * const start = context_indirect_bb(ce);
+	u32 *cs;
+
+	cs = emit(ce, start);
+	GEM_BUG_ON(cs - start > I915_GTT_PAGE_SIZE / sizeof(*cs));
+	while ((unsigned long)cs % CACHELINE_BYTES)
+		*cs++ = MI_NOOP;
+
+	GEM_BUG_ON(cs - start > DG2_PREDICATE_RESULT_BB / sizeof(*start));
+	setup_predicate_disable_wa(ce, start + DG2_PREDICATE_RESULT_BB / sizeof(*start));
+
+	lrc_setup_indirect_ctx(ce->lrc_reg_state, engine,
+			       lrc_indirect_bb(ce),
+			       (cs - start) * sizeof(*cs));
+}
+
+/*
+ * The context descriptor encodes various attributes of a context,
+ * including its GTT address and some flags. Because it's fairly
+ * expensive to calculate, we'll just do it once and cache the result,
+ * which remains valid until the context is unpinned.
+ *
+ * This is what a descriptor looks like, from LSB to MSB::
+ *
+ *      bits  0-11:    flags, GEN8_CTX_* (cached in ctx->desc_template)
+ *      bits 12-31:    LRCA, GTT address of (the HWSP of) this context
+ *      bits 32-52:    ctx ID, a globally unique tag (highest bit used by GuC)
+ *      bits 53-54:    mbz, reserved for use by hardware
+ *      bits 55-63:    group ID, currently unused and set to 0
+ *
+ * Starting from Gen11, the upper dword of the descriptor has a new format:
+ *
+ *      bits 32-36:    reserved
+ *      bits 37-47:    SW context ID
+ *      bits 48:53:    engine instance
+ *      bit 54:        mbz, reserved for use by hardware
+ *      bits 55-60:    SW counter
+ *      bits 61-63:    engine class
+ *
+ * On Xe_HP, the upper dword of the descriptor has a new format:
+ *
+ *      bits 32-37:    virtual function number
+ *      bit 38:        mbz, reserved for use by hardware
+ *      bits 39-54:    SW context ID
+ *      bits 55-57:    reserved
+ *      bits 58-63:    SW counter
+ *
+ * engine info, SW context ID and SW counter need to form a unique number
+ * (Context ID) per lrc.
+ */
+static u32 lrc_descriptor(const struct intel_context *ce)
+{
+	u32 desc;
+
+	desc = INTEL_LEGACY_32B_CONTEXT;
+	if (i915_vm_is_4lvl(ce->vm))
+		desc = INTEL_LEGACY_64B_CONTEXT;
+	desc <<= GEN8_CTX_ADDRESSING_MODE_SHIFT;
+
+	desc |= GEN8_CTX_VALID | GEN8_CTX_PRIVILEGE;
+	if (GRAPHICS_VER(ce->vm->i915) == 8)
+		desc |= GEN8_CTX_L3LLC_COHERENT;
+
+	return i915_ggtt_offset(ce->state) | desc;
+}
+
+u32 lrc_update_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine,
+		    u32 head)
+{
+	struct intel_ring *ring = ce->ring;
+	u32 *regs = ce->lrc_reg_state;
+
+	GEM_BUG_ON(!intel_ring_offset_valid(ring, head));
+	GEM_BUG_ON(!intel_ring_offset_valid(ring, ring->tail));
+
+	regs[CTX_RING_START] = i915_ggtt_offset(ring->vma);
+	regs[CTX_RING_HEAD] = head;
+	regs[CTX_RING_TAIL] = ring->tail;
+	regs[CTX_RING_CTL] = RING_CTL_SIZE(ring->size) | RING_VALID;
+
+	/* RPCS */
+	if (engine->class == RENDER_CLASS) {
+		regs[CTX_R_PWR_CLK_STATE] =
+			intel_sseu_make_rpcs(engine->gt, &ce->sseu);
+
+		i915_oa_init_reg_state(ce, engine);
+	}
+
+	if (ce->wa_bb_page) {
+		u32 *(*fn)(const struct intel_context *ce, u32 *cs);
+
+		fn = gen12_emit_indirect_ctx_xcs;
+		if (ce->engine->class == RENDER_CLASS)
+			fn = gen12_emit_indirect_ctx_rcs;
+
+		/* Mutually exclusive wrt to global indirect bb */
+		GEM_BUG_ON(engine->wa_ctx.indirect_ctx.size);
+		setup_indirect_ctx_bb(ce, engine, fn);
+	}
+
+	return lrc_descriptor(ce) | CTX_DESC_FORCE_RESTORE;
+}
+
+void lrc_update_offsets(struct intel_context *ce,
+			struct intel_engine_cs *engine)
+{
+	set_offsets(ce->lrc_reg_state, reg_offsets(engine), engine, false);
+}
+
+void lrc_check_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine,
+		    const char *when)
+{
+	const struct intel_ring *ring = ce->ring;
+	u32 *regs = ce->lrc_reg_state;
+	bool valid = true;
+	int x;
+
+	if (regs[CTX_RING_START] != i915_ggtt_offset(ring->vma)) {
+		pr_err("%s: context submitted with incorrect RING_START [%08x], expected %08x\n",
+		       engine->name,
+		       regs[CTX_RING_START],
+		       i915_ggtt_offset(ring->vma));
+		regs[CTX_RING_START] = i915_ggtt_offset(ring->vma);
+		valid = false;
+	}
+
+	if ((regs[CTX_RING_CTL] & ~(RING_WAIT | RING_WAIT_SEMAPHORE)) !=
+	    (RING_CTL_SIZE(ring->size) | RING_VALID)) {
+		pr_err("%s: context submitted with incorrect RING_CTL [%08x], expected %08x\n",
+		       engine->name,
+		       regs[CTX_RING_CTL],
+		       (u32)(RING_CTL_SIZE(ring->size) | RING_VALID));
+		regs[CTX_RING_CTL] = RING_CTL_SIZE(ring->size) | RING_VALID;
+		valid = false;
+	}
+
+	x = lrc_ring_mi_mode(engine);
+	if (x != -1 && regs[x + 1] & (regs[x + 1] >> 16) & STOP_RING) {
+		pr_err("%s: context submitted with STOP_RING [%08x] in RING_MI_MODE\n",
+		       engine->name, regs[x + 1]);
+		regs[x + 1] &= ~STOP_RING;
+		regs[x + 1] |= STOP_RING << 16;
+		valid = false;
+	}
+
+	WARN_ONCE(!valid, "Invalid lrc state found %s submission\n", when);
+}
+
+/*
+ * In this WA we need to set GEN8_L3SQCREG4[21:21] and reset it after
+ * PIPE_CONTROL instruction. This is required for the flush to happen correctly
+ * but there is a slight complication as this is applied in WA batch where the
+ * values are only initialized once so we cannot take register value at the
+ * beginning and reuse it further; hence we save its value to memory, upload a
+ * constant value with bit21 set and then we restore it back with the saved value.
+ * To simplify the WA, a constant value is formed by using the default value
+ * of this register. This shouldn't be a problem because we are only modifying
+ * it for a short period and this batch in non-premptible. We can ofcourse
+ * use additional instructions that read the actual value of the register
+ * at that time and set our bit of interest but it makes the WA complicated.
+ *
+ * This WA is also required for Gen9 so extracting as a function avoids
+ * code duplication.
+ */
+static u32 *
+gen8_emit_flush_coherentl3_wa(struct intel_engine_cs *engine, u32 *batch)
+{
+	/* NB no one else is allowed to scribble over scratch + 256! */
+	*batch++ = MI_STORE_REGISTER_MEM_GEN8 | MI_SRM_LRM_GLOBAL_GTT;
+	*batch++ = i915_mmio_reg_offset(GEN8_L3SQCREG4);
+	*batch++ = intel_gt_scratch_offset(engine->gt,
+					   INTEL_GT_SCRATCH_FIELD_COHERENTL3_WA);
+	*batch++ = 0;
+
+	*batch++ = MI_LOAD_REGISTER_IMM(1);
+	*batch++ = i915_mmio_reg_offset(GEN8_L3SQCREG4);
+	*batch++ = 0x40400000 | GEN8_LQSC_FLUSH_COHERENT_LINES;
+
+	batch = gen8_emit_pipe_control(batch,
+				       PIPE_CONTROL_CS_STALL |
+				       PIPE_CONTROL_DC_FLUSH_ENABLE,
+				       0);
+
+	*batch++ = MI_LOAD_REGISTER_MEM_GEN8 | MI_SRM_LRM_GLOBAL_GTT;
+	*batch++ = i915_mmio_reg_offset(GEN8_L3SQCREG4);
+	*batch++ = intel_gt_scratch_offset(engine->gt,
+					   INTEL_GT_SCRATCH_FIELD_COHERENTL3_WA);
+	*batch++ = 0;
+
+	return batch;
+}
+
+/*
+ * Typically we only have one indirect_ctx and per_ctx batch buffer which are
+ * initialized at the beginning and shared across all contexts but this field
+ * helps us to have multiple batches at different offsets and select them based
+ * on a criteria. At the moment this batch always start at the beginning of the page
+ * and at this point we don't have multiple wa_ctx batch buffers.
+ *
+ * The number of WA applied are not known at the beginning; we use this field
+ * to return the no of DWORDS written.
+ *
+ * It is to be noted that this batch does not contain MI_BATCH_BUFFER_END
+ * so it adds NOOPs as padding to make it cacheline aligned.
+ * MI_BATCH_BUFFER_END will be added to perctx batch and both of them together
+ * makes a complete batch buffer.
+ */
+static u32 *gen8_init_indirectctx_bb(struct intel_engine_cs *engine, u32 *batch)
+{
+	/* WaDisableCtxRestoreArbitration:bdw,chv */
+	*batch++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
+	/* WaFlushCoherentL3CacheLinesAtContextSwitch:bdw */
+	if (IS_BROADWELL(engine->i915))
+		batch = gen8_emit_flush_coherentl3_wa(engine, batch);
+
+	/* WaClearSlmSpaceAtContextSwitch:bdw,chv */
+	/* Actual scratch location is at 128 bytes offset */
+	batch = gen8_emit_pipe_control(batch,
+				       PIPE_CONTROL_FLUSH_L3 |
+				       PIPE_CONTROL_STORE_DATA_INDEX |
+				       PIPE_CONTROL_CS_STALL |
+				       PIPE_CONTROL_QW_WRITE,
+				       LRC_PPHWSP_SCRATCH_ADDR);
+
+	*batch++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+
+	/* Pad to end of cacheline */
+	while ((unsigned long)batch % CACHELINE_BYTES)
+		*batch++ = MI_NOOP;
+
+	/*
+	 * MI_BATCH_BUFFER_END is not required in Indirect ctx BB because
+	 * execution depends on the length specified in terms of cache lines
+	 * in the register CTX_RCS_INDIRECT_CTX
+	 */
+
+	return batch;
+}
+
+struct lri {
+	i915_reg_t reg;
+	u32 value;
+};
+
+static u32 *emit_lri(u32 *batch, const struct lri *lri, unsigned int count)
+{
+	GEM_BUG_ON(!count || count > 63);
+
+	*batch++ = MI_LOAD_REGISTER_IMM(count);
+	do {
+		*batch++ = i915_mmio_reg_offset(lri->reg);
+		*batch++ = lri->value;
+	} while (lri++, --count);
+	*batch++ = MI_NOOP;
+
+	return batch;
+}
+
+static u32 *gen9_init_indirectctx_bb(struct intel_engine_cs *engine, u32 *batch)
+{
+	static const struct lri lri[] = {
+		/* WaDisableGatherAtSetShaderCommonSlice:skl,bxt,kbl,glk */
+		{
+			COMMON_SLICE_CHICKEN2,
+			__MASKED_FIELD(GEN9_DISABLE_GATHER_AT_SET_SHADER_COMMON_SLICE,
+				       0),
+		},
+
+		/* BSpec: 11391 */
+		{
+			FF_SLICE_CHICKEN,
+			__MASKED_FIELD(FF_SLICE_CHICKEN_CL_PROVOKING_VERTEX_FIX,
+				       FF_SLICE_CHICKEN_CL_PROVOKING_VERTEX_FIX),
+		},
+
+		/* BSpec: 11299 */
+		{
+			_3D_CHICKEN3,
+			__MASKED_FIELD(_3D_CHICKEN_SF_PROVOKING_VERTEX_FIX,
+				       _3D_CHICKEN_SF_PROVOKING_VERTEX_FIX),
+		}
+	};
+
+	*batch++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
+	/* WaFlushCoherentL3CacheLinesAtContextSwitch:skl,bxt,glk */
+	batch = gen8_emit_flush_coherentl3_wa(engine, batch);
+
+	/* WaClearSlmSpaceAtContextSwitch:skl,bxt,kbl,glk,cfl */
+	batch = gen8_emit_pipe_control(batch,
+				       PIPE_CONTROL_FLUSH_L3 |
+				       PIPE_CONTROL_STORE_DATA_INDEX |
+				       PIPE_CONTROL_CS_STALL |
+				       PIPE_CONTROL_QW_WRITE,
+				       LRC_PPHWSP_SCRATCH_ADDR);
+
+	batch = emit_lri(batch, lri, ARRAY_SIZE(lri));
+
+	/* WaMediaPoolStateCmdInWABB:bxt,glk */
+	if (HAS_POOLED_EU(engine->i915)) {
+		/*
+		 * EU pool configuration is setup along with golden context
+		 * during context initialization. This value depends on
+		 * device type (2x6 or 3x6) and needs to be updated based
+		 * on which subslice is disabled especially for 2x6
+		 * devices, however it is safe to load default
+		 * configuration of 3x6 device instead of masking off
+		 * corresponding bits because HW ignores bits of a disabled
+		 * subslice and drops down to appropriate config. Please
+		 * see render_state_setup() in i915_gem_render_state.c for
+		 * possible configurations, to avoid duplication they are
+		 * not shown here again.
+		 */
+		*batch++ = GEN9_MEDIA_POOL_STATE;
+		*batch++ = GEN9_MEDIA_POOL_ENABLE;
+		*batch++ = 0x00777000;
+		*batch++ = 0;
+		*batch++ = 0;
+		*batch++ = 0;
+	}
+
+	*batch++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+
+	/* Pad to end of cacheline */
+	while ((unsigned long)batch % CACHELINE_BYTES)
+		*batch++ = MI_NOOP;
+
+	return batch;
+}
+
+#define CTX_WA_BB_SIZE (PAGE_SIZE)
+
+static int lrc_create_wa_ctx(struct intel_engine_cs *engine)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err;
+
+	obj = i915_gem_object_create_shmem(engine->i915, CTX_WA_BB_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = i915_vma_instance(obj, &engine->gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err;
+	}
+
+	engine->wa_ctx.vma = vma;
+	return 0;
+
+err:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+void lrc_fini_wa_ctx(struct intel_engine_cs *engine)
+{
+	i915_vma_unpin_and_release(&engine->wa_ctx.vma, 0);
+}
+
+typedef u32 *(*wa_bb_func_t)(struct intel_engine_cs *engine, u32 *batch);
+
+void lrc_init_wa_ctx(struct intel_engine_cs *engine)
+{
+	struct i915_ctx_workarounds *wa_ctx = &engine->wa_ctx;
+	struct i915_wa_ctx_bb *wa_bb[] = {
+		&wa_ctx->indirect_ctx, &wa_ctx->per_ctx
+	};
+	wa_bb_func_t wa_bb_fn[ARRAY_SIZE(wa_bb)];
+	struct i915_gem_ww_ctx ww;
+	void *batch, *batch_ptr;
+	unsigned int i;
+	int err;
+
+	if (!(engine->flags & I915_ENGINE_HAS_RCS_REG_STATE))
+		return;
+
+	switch (GRAPHICS_VER(engine->i915)) {
+	case 12:
+	case 11:
+		return;
+	case 9:
+		wa_bb_fn[0] = gen9_init_indirectctx_bb;
+		wa_bb_fn[1] = NULL;
+		break;
+	case 8:
+		wa_bb_fn[0] = gen8_init_indirectctx_bb;
+		wa_bb_fn[1] = NULL;
+		break;
+	default:
+		MISSING_CASE(GRAPHICS_VER(engine->i915));
+		return;
+	}
+
+	err = lrc_create_wa_ctx(engine);
+	if (err) {
+		/*
+		 * We continue even if we fail to initialize WA batch
+		 * because we only expect rare glitches but nothing
+		 * critical to prevent us from using GPU
+		 */
+		drm_err(&engine->i915->drm,
+			"Ignoring context switch w/a allocation error:%d\n",
+			err);
+		return;
+	}
+
+	if (!engine->wa_ctx.vma)
+		return;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	err = i915_gem_object_lock(wa_ctx->vma->obj, &ww);
+	if (!err)
+		err = i915_ggtt_pin(wa_ctx->vma, &ww, 0, PIN_HIGH);
+	if (err)
+		goto err;
+
+	batch = i915_gem_object_pin_map(wa_ctx->vma->obj, I915_MAP_WB);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		goto err_unpin;
+	}
+
+	/*
+	 * Emit the two workaround batch buffers, recording the offset from the
+	 * start of the workaround batch buffer object for each and their
+	 * respective sizes.
+	 */
+	batch_ptr = batch;
+	for (i = 0; i < ARRAY_SIZE(wa_bb_fn); i++) {
+		wa_bb[i]->offset = batch_ptr - batch;
+		if (GEM_DEBUG_WARN_ON(!IS_ALIGNED(wa_bb[i]->offset,
+						  CACHELINE_BYTES))) {
+			err = -EINVAL;
+			break;
+		}
+		if (wa_bb_fn[i])
+			batch_ptr = wa_bb_fn[i](engine, batch_ptr);
+		wa_bb[i]->size = batch_ptr - (batch + wa_bb[i]->offset);
+	}
+	GEM_BUG_ON(batch_ptr - batch > CTX_WA_BB_SIZE);
+
+	__i915_gem_object_flush_map(wa_ctx->vma->obj, 0, batch_ptr - batch);
+	__i915_gem_object_release_map(wa_ctx->vma->obj);
+
+	/* Verify that we can handle failure to setup the wa_ctx */
+	if (!err)
+		err = i915_inject_probe_error(engine->i915, -ENODEV);
+
+err_unpin:
+	if (err)
+		i915_vma_unpin(wa_ctx->vma);
+err:
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+
+	if (err) {
+		i915_vma_put(engine->wa_ctx.vma);
+
+		/* Clear all flags to prevent further use */
+		memset(wa_ctx, 0, sizeof(*wa_ctx));
+	}
+}
+
+static void st_runtime_underflow(struct intel_context_stats *stats, s32 dt)
+{
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+	stats->runtime.num_underflow++;
+	stats->runtime.max_underflow =
+		max_t(u32, stats->runtime.max_underflow, -dt);
+#endif
+}
+
+static u32 lrc_get_runtime(const struct intel_context *ce)
+{
+	/*
+	 * We can use either ppHWSP[16] which is recorded before the context
+	 * switch (and so excludes the cost of context switches) or use the
+	 * value from the context image itself, which is saved/restored earlier
+	 * and so includes the cost of the save.
+	 */
+	return READ_ONCE(ce->lrc_reg_state[CTX_TIMESTAMP]);
+}
+
+void lrc_update_runtime(struct intel_context *ce)
+{
+	struct intel_context_stats *stats = &ce->stats;
+	u32 old;
+	s32 dt;
+
+	old = stats->runtime.last;
+	stats->runtime.last = lrc_get_runtime(ce);
+	dt = stats->runtime.last - old;
+	if (!dt)
+		return;
+
+	if (unlikely(dt < 0)) {
+		CE_TRACE(ce, "runtime underflow: last=%u, new=%u, delta=%d\n",
+			 old, stats->runtime.last, dt);
+		st_runtime_underflow(stats, dt);
+		return;
+	}
+
+	ewma_runtime_add(&stats->runtime.avg, dt);
+	stats->runtime.total += dt;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_lrc.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.h b/drivers/gpu/drm/i915/gt/intel_lrc.h
new file mode 100644
index 000000000..a390f0813
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_lrc.h
@@ -0,0 +1,143 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#ifndef __INTEL_LRC_H__
+#define __INTEL_LRC_H__
+
+#include "i915_priolist_types.h"
+
+#include <linux/bitfield.h>
+#include <linux/types.h>
+
+#include "intel_context.h"
+
+struct drm_i915_gem_object;
+struct i915_gem_ww_ctx;
+struct intel_engine_cs;
+struct intel_ring;
+struct kref;
+
+/* At the start of the context image is its per-process HWS page */
+#define LRC_PPHWSP_PN	(0)
+#define LRC_PPHWSP_SZ	(1)
+/* After the PPHWSP we have the logical state for the context */
+#define LRC_STATE_PN	(LRC_PPHWSP_PN + LRC_PPHWSP_SZ)
+#define LRC_STATE_OFFSET (LRC_STATE_PN * PAGE_SIZE)
+
+/* Space within PPHWSP reserved to be used as scratch */
+#define LRC_PPHWSP_SCRATCH		0x34
+#define LRC_PPHWSP_SCRATCH_ADDR		(LRC_PPHWSP_SCRATCH * sizeof(u32))
+
+void lrc_init_wa_ctx(struct intel_engine_cs *engine);
+void lrc_fini_wa_ctx(struct intel_engine_cs *engine);
+
+int lrc_alloc(struct intel_context *ce,
+	      struct intel_engine_cs *engine);
+void lrc_reset(struct intel_context *ce);
+void lrc_fini(struct intel_context *ce);
+void lrc_destroy(struct kref *kref);
+
+int
+lrc_pre_pin(struct intel_context *ce,
+	    struct intel_engine_cs *engine,
+	    struct i915_gem_ww_ctx *ww,
+	    void **vaddr);
+int
+lrc_pin(struct intel_context *ce,
+	struct intel_engine_cs *engine,
+	void *vaddr);
+void lrc_unpin(struct intel_context *ce);
+void lrc_post_unpin(struct intel_context *ce);
+
+void lrc_init_state(struct intel_context *ce,
+		    struct intel_engine_cs *engine,
+		    void *state);
+
+void lrc_init_regs(const struct intel_context *ce,
+		   const struct intel_engine_cs *engine,
+		   bool clear);
+void lrc_reset_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine);
+
+u32 lrc_update_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine,
+		    u32 head);
+void lrc_update_offsets(struct intel_context *ce,
+			struct intel_engine_cs *engine);
+
+void lrc_check_regs(const struct intel_context *ce,
+		    const struct intel_engine_cs *engine,
+		    const char *when);
+
+void lrc_update_runtime(struct intel_context *ce);
+
+enum {
+	INTEL_ADVANCED_CONTEXT = 0,
+	INTEL_LEGACY_32B_CONTEXT,
+	INTEL_ADVANCED_AD_CONTEXT,
+	INTEL_LEGACY_64B_CONTEXT
+};
+
+enum {
+	FAULT_AND_HANG = 0,
+	FAULT_AND_HALT, /* Debug only */
+	FAULT_AND_STREAM,
+	FAULT_AND_CONTINUE /* Unsupported */
+};
+
+#define CTX_GTT_ADDRESS_MASK			GENMASK(31, 12)
+#define GEN8_CTX_VALID				(1 << 0)
+#define GEN8_CTX_FORCE_PD_RESTORE		(1 << 1)
+#define GEN8_CTX_FORCE_RESTORE			(1 << 2)
+#define GEN8_CTX_L3LLC_COHERENT			(1 << 5)
+#define GEN8_CTX_PRIVILEGE			(1 << 8)
+#define GEN8_CTX_ADDRESSING_MODE_SHIFT		3
+#define GEN12_CTX_PRIORITY_MASK			GENMASK(10, 9)
+#define GEN12_CTX_PRIORITY_HIGH			FIELD_PREP(GEN12_CTX_PRIORITY_MASK, 2)
+#define GEN12_CTX_PRIORITY_NORMAL		FIELD_PREP(GEN12_CTX_PRIORITY_MASK, 1)
+#define GEN12_CTX_PRIORITY_LOW			FIELD_PREP(GEN12_CTX_PRIORITY_MASK, 0)
+#define GEN8_CTX_ID_SHIFT			32
+#define GEN8_CTX_ID_WIDTH			21
+#define GEN11_SW_CTX_ID_SHIFT			37
+#define GEN11_SW_CTX_ID_WIDTH			11
+#define GEN11_ENGINE_CLASS_SHIFT		61
+#define GEN11_ENGINE_CLASS_WIDTH		3
+#define GEN11_ENGINE_INSTANCE_SHIFT		48
+#define GEN11_ENGINE_INSTANCE_WIDTH		6
+#define XEHP_SW_CTX_ID_SHIFT			39
+#define XEHP_SW_CTX_ID_WIDTH			16
+#define XEHP_SW_COUNTER_SHIFT			58
+#define XEHP_SW_COUNTER_WIDTH			6
+
+static inline void lrc_runtime_start(struct intel_context *ce)
+{
+	struct intel_context_stats *stats = &ce->stats;
+
+	if (intel_context_is_barrier(ce))
+		return;
+
+	if (stats->active)
+		return;
+
+	WRITE_ONCE(stats->active, intel_context_clock());
+}
+
+static inline void lrc_runtime_stop(struct intel_context *ce)
+{
+	struct intel_context_stats *stats = &ce->stats;
+
+	if (!stats->active)
+		return;
+
+	lrc_update_runtime(ce);
+	WRITE_ONCE(stats->active, 0);
+}
+
+#define DG2_PREDICATE_RESULT_WA (PAGE_SIZE - sizeof(u64))
+#define DG2_PREDICATE_RESULT_BB (2048)
+
+u32 lrc_indirect_bb(const struct intel_context *ce);
+
+#endif /* __INTEL_LRC_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc_reg.h b/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
new file mode 100644
index 000000000..304000c7e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_lrc_reg.h
@@ -0,0 +1,81 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2018 Intel Corporation
+ */
+
+#ifndef _INTEL_LRC_REG_H_
+#define _INTEL_LRC_REG_H_
+
+#include <linux/types.h>
+
+#define CTX_DESC_FORCE_RESTORE BIT_ULL(2)
+
+/* GEN8 to GEN12 Reg State Context */
+#define CTX_CONTEXT_CONTROL		(0x02 + 1)
+#define CTX_RING_HEAD			(0x04 + 1)
+#define CTX_RING_TAIL			(0x06 + 1)
+#define CTX_RING_START			(0x08 + 1)
+#define CTX_RING_CTL			(0x0a + 1)
+#define CTX_BB_STATE			(0x10 + 1)
+#define CTX_TIMESTAMP			(0x22 + 1)
+#define CTX_PDP3_UDW			(0x24 + 1)
+#define CTX_PDP3_LDW			(0x26 + 1)
+#define CTX_PDP2_UDW			(0x28 + 1)
+#define CTX_PDP2_LDW			(0x2a + 1)
+#define CTX_PDP1_UDW			(0x2c + 1)
+#define CTX_PDP1_LDW			(0x2e + 1)
+#define CTX_PDP0_UDW			(0x30 + 1)
+#define CTX_PDP0_LDW			(0x32 + 1)
+#define CTX_R_PWR_CLK_STATE		(0x42 + 1)
+
+#define GEN9_CTX_RING_MI_MODE		0x54
+
+#define ASSIGN_CTX_PDP(ppgtt, reg_state, n) do { \
+	u32 *reg_state__ = (reg_state); \
+	const u64 addr__ = i915_page_dir_dma_addr((ppgtt), (n)); \
+	(reg_state__)[CTX_PDP ## n ## _UDW] = upper_32_bits(addr__); \
+	(reg_state__)[CTX_PDP ## n ## _LDW] = lower_32_bits(addr__); \
+} while (0)
+
+#define ASSIGN_CTX_PML4(ppgtt, reg_state) do { \
+	u32 *reg_state__ = (reg_state); \
+	const u64 addr__ = px_dma((ppgtt)->pd); \
+	(reg_state__)[CTX_PDP0_UDW] = upper_32_bits(addr__); \
+	(reg_state__)[CTX_PDP0_LDW] = lower_32_bits(addr__); \
+} while (0)
+
+#define GEN8_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT	0x17
+#define GEN9_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT	0x26
+#define GEN10_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT	0x19
+#define GEN11_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT	0x1A
+#define GEN12_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT	0xD
+
+#define GEN8_EXECLISTS_STATUS_BUF 0x370
+#define GEN11_EXECLISTS_STATUS_BUF2 0x3c0
+
+/*
+ * The docs specify that the write pointer wraps around after 5h, "After status
+ * is written out to the last available status QW at offset 5h, this pointer
+ * wraps to 0."
+ *
+ * Therefore, one must infer than even though there are 3 bits available, 6 and
+ * 7 appear to be * reserved.
+ */
+#define GEN8_CSB_ENTRIES 6
+#define GEN8_CSB_PTR_MASK 0x7
+#define GEN8_CSB_READ_PTR_MASK	(GEN8_CSB_PTR_MASK << 8)
+#define GEN8_CSB_WRITE_PTR_MASK	(GEN8_CSB_PTR_MASK << 0)
+
+#define GEN11_CSB_ENTRIES 12
+#define GEN11_CSB_PTR_MASK 0xf
+#define GEN11_CSB_READ_PTR_MASK		(GEN11_CSB_PTR_MASK << 8)
+#define GEN11_CSB_WRITE_PTR_MASK	(GEN11_CSB_PTR_MASK << 0)
+
+#define MAX_CONTEXT_HW_ID	(1 << 21) /* exclusive */
+#define GEN11_MAX_CONTEXT_HW_ID	(1 << 11) /* exclusive */
+/* in Gen12 ID 0x7FF is reserved to indicate idle */
+#define GEN12_MAX_CONTEXT_HW_ID	(GEN11_MAX_CONTEXT_HW_ID - 1)
+/* in Xe_HP ID 0xFFFF is reserved to indicate "invalid context" */
+#define XEHP_MAX_CONTEXT_HW_ID	0xFFFF
+
+#endif /* _INTEL_LRC_REG_H_ */
diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.c b/drivers/gpu/drm/i915/gt/intel_migrate.c
new file mode 100644
index 000000000..ee072c7d6
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_migrate.c
@@ -0,0 +1,1135 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "intel_context.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt.h"
+#include "intel_gtt.h"
+#include "intel_migrate.h"
+#include "intel_ring.h"
+
+struct insert_pte_data {
+	u64 offset;
+};
+
+#define CHUNK_SZ SZ_8M /* ~1ms at 8GiB/s preemption delay */
+
+#define GET_CCS_BYTES(i915, size)	(HAS_FLAT_CCS(i915) ? \
+					 DIV_ROUND_UP(size, NUM_BYTES_PER_CCS_BYTE) : 0)
+static bool engine_supports_migration(struct intel_engine_cs *engine)
+{
+	if (!engine)
+		return false;
+
+	/*
+	 * We need the ability to prevent aribtration (MI_ARB_ON_OFF),
+	 * the ability to write PTE using inline data (MI_STORE_DATA)
+	 * and of course the ability to do the block transfer (blits).
+	 */
+	GEM_BUG_ON(engine->class != COPY_ENGINE_CLASS);
+
+	return true;
+}
+
+static void xehpsdv_toggle_pdes(struct i915_address_space *vm,
+				struct i915_page_table *pt,
+				void *data)
+{
+	struct insert_pte_data *d = data;
+
+	/*
+	 * Insert a dummy PTE into every PT that will map to LMEM to ensure
+	 * we have a correctly setup PDE structure for later use.
+	 */
+	vm->insert_page(vm, 0, d->offset, I915_CACHE_NONE, PTE_LM);
+	GEM_BUG_ON(!pt->is_compact);
+	d->offset += SZ_2M;
+}
+
+static void xehpsdv_insert_pte(struct i915_address_space *vm,
+			       struct i915_page_table *pt,
+			       void *data)
+{
+	struct insert_pte_data *d = data;
+
+	/*
+	 * We are playing tricks here, since the actual pt, from the hw
+	 * pov, is only 256bytes with 32 entries, or 4096bytes with 512
+	 * entries, but we are still guaranteed that the physical
+	 * alignment is 64K underneath for the pt, and we are careful
+	 * not to access the space in the void.
+	 */
+	vm->insert_page(vm, px_dma(pt), d->offset, I915_CACHE_NONE, PTE_LM);
+	d->offset += SZ_64K;
+}
+
+static void insert_pte(struct i915_address_space *vm,
+		       struct i915_page_table *pt,
+		       void *data)
+{
+	struct insert_pte_data *d = data;
+
+	vm->insert_page(vm, px_dma(pt), d->offset, I915_CACHE_NONE,
+			i915_gem_object_is_lmem(pt->base) ? PTE_LM : 0);
+	d->offset += PAGE_SIZE;
+}
+
+static struct i915_address_space *migrate_vm(struct intel_gt *gt)
+{
+	struct i915_vm_pt_stash stash = {};
+	struct i915_ppgtt *vm;
+	int err;
+	int i;
+
+	/*
+	 * We construct a very special VM for use by all migration contexts,
+	 * it is kept pinned so that it can be used at any time. As we need
+	 * to pre-allocate the page directories for the migration VM, this
+	 * limits us to only using a small number of prepared vma.
+	 *
+	 * To be able to pipeline and reschedule migration operations while
+	 * avoiding unnecessary contention on the vm itself, the PTE updates
+	 * are inline with the blits. All the blits use the same fixed
+	 * addresses, with the backing store redirection being updated on the
+	 * fly. Only 2 implicit vma are used for all migration operations.
+	 *
+	 * We lay the ppGTT out as:
+	 *
+	 *	[0, CHUNK_SZ) -> first object
+	 *	[CHUNK_SZ, 2 * CHUNK_SZ) -> second object
+	 *	[2 * CHUNK_SZ, 2 * CHUNK_SZ + 2 * CHUNK_SZ >> 9] -> PTE
+	 *
+	 * By exposing the dma addresses of the page directories themselves
+	 * within the ppGTT, we are then able to rewrite the PTE prior to use.
+	 * But the PTE update and subsequent migration operation must be atomic,
+	 * i.e. within the same non-preemptible window so that we do not switch
+	 * to another migration context that overwrites the PTE.
+	 *
+	 * This changes quite a bit on platforms with HAS_64K_PAGES support,
+	 * where we instead have three windows, each CHUNK_SIZE in size. The
+	 * first is reserved for mapping system-memory, and that just uses the
+	 * 512 entry layout using 4K GTT pages. The other two windows just map
+	 * lmem pages and must use the new compact 32 entry layout using 64K GTT
+	 * pages, which ensures we can address any lmem object that the user
+	 * throws at us. We then also use the xehpsdv_toggle_pdes as a way of
+	 * just toggling the PDE bit(GEN12_PDE_64K) for us, to enable the
+	 * compact layout for each of these page-tables, that fall within the
+	 * [CHUNK_SIZE, 3 * CHUNK_SIZE) range.
+	 *
+	 * We lay the ppGTT out as:
+	 *
+	 * [0, CHUNK_SZ) -> first window/object, maps smem
+	 * [CHUNK_SZ, 2 * CHUNK_SZ) -> second window/object, maps lmem src
+	 * [2 * CHUNK_SZ, 3 * CHUNK_SZ) -> third window/object, maps lmem dst
+	 *
+	 * For the PTE window it's also quite different, since each PTE must
+	 * point to some 64K page, one for each PT(since it's in lmem), and yet
+	 * each is only <= 4096bytes, but since the unused space within that PTE
+	 * range is never touched, this should be fine.
+	 *
+	 * So basically each PT now needs 64K of virtual memory, instead of 4K,
+	 * which looks like:
+	 *
+	 * [3 * CHUNK_SZ, 3 * CHUNK_SZ + ((3 * CHUNK_SZ / SZ_2M) * SZ_64K)] -> PTE
+	 */
+
+	vm = i915_ppgtt_create(gt, I915_BO_ALLOC_PM_EARLY);
+	if (IS_ERR(vm))
+		return ERR_CAST(vm);
+
+	if (!vm->vm.allocate_va_range || !vm->vm.foreach) {
+		err = -ENODEV;
+		goto err_vm;
+	}
+
+	if (HAS_64K_PAGES(gt->i915))
+		stash.pt_sz = I915_GTT_PAGE_SIZE_64K;
+
+	/*
+	 * Each engine instance is assigned its own chunk in the VM, so
+	 * that we can run multiple instances concurrently
+	 */
+	for (i = 0; i < ARRAY_SIZE(gt->engine_class[COPY_ENGINE_CLASS]); i++) {
+		struct intel_engine_cs *engine;
+		u64 base = (u64)i << 32;
+		struct insert_pte_data d = {};
+		struct i915_gem_ww_ctx ww;
+		u64 sz;
+
+		engine = gt->engine_class[COPY_ENGINE_CLASS][i];
+		if (!engine_supports_migration(engine))
+			continue;
+
+		/*
+		 * We copy in 8MiB chunks. Each PDE covers 2MiB, so we need
+		 * 4x2 page directories for source/destination.
+		 */
+		if (HAS_64K_PAGES(gt->i915))
+			sz = 3 * CHUNK_SZ;
+		else
+			sz = 2 * CHUNK_SZ;
+		d.offset = base + sz;
+
+		/*
+		 * We need another page directory setup so that we can write
+		 * the 8x512 PTE in each chunk.
+		 */
+		if (HAS_64K_PAGES(gt->i915))
+			sz += (sz / SZ_2M) * SZ_64K;
+		else
+			sz += (sz >> 12) * sizeof(u64);
+
+		err = i915_vm_alloc_pt_stash(&vm->vm, &stash, sz);
+		if (err)
+			goto err_vm;
+
+		for_i915_gem_ww(&ww, err, true) {
+			err = i915_vm_lock_objects(&vm->vm, &ww);
+			if (err)
+				continue;
+			err = i915_vm_map_pt_stash(&vm->vm, &stash);
+			if (err)
+				continue;
+
+			vm->vm.allocate_va_range(&vm->vm, &stash, base, sz);
+		}
+		i915_vm_free_pt_stash(&vm->vm, &stash);
+		if (err)
+			goto err_vm;
+
+		/* Now allow the GPU to rewrite the PTE via its own ppGTT */
+		if (HAS_64K_PAGES(gt->i915)) {
+			vm->vm.foreach(&vm->vm, base, d.offset - base,
+				       xehpsdv_insert_pte, &d);
+			d.offset = base + CHUNK_SZ;
+			vm->vm.foreach(&vm->vm,
+				       d.offset,
+				       2 * CHUNK_SZ,
+				       xehpsdv_toggle_pdes, &d);
+		} else {
+			vm->vm.foreach(&vm->vm, base, d.offset - base,
+				       insert_pte, &d);
+		}
+	}
+
+	return &vm->vm;
+
+err_vm:
+	i915_vm_put(&vm->vm);
+	return ERR_PTR(err);
+}
+
+static struct intel_engine_cs *first_copy_engine(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(gt->engine_class[COPY_ENGINE_CLASS]); i++) {
+		engine = gt->engine_class[COPY_ENGINE_CLASS][i];
+		if (engine_supports_migration(engine))
+			return engine;
+	}
+
+	return NULL;
+}
+
+static struct intel_context *pinned_context(struct intel_gt *gt)
+{
+	static struct lock_class_key key;
+	struct intel_engine_cs *engine;
+	struct i915_address_space *vm;
+	struct intel_context *ce;
+
+	engine = first_copy_engine(gt);
+	if (!engine)
+		return ERR_PTR(-ENODEV);
+
+	vm = migrate_vm(gt);
+	if (IS_ERR(vm))
+		return ERR_CAST(vm);
+
+	ce = intel_engine_create_pinned_context(engine, vm, SZ_512K,
+						I915_GEM_HWS_MIGRATE,
+						&key, "migrate");
+	i915_vm_put(vm);
+	return ce;
+}
+
+int intel_migrate_init(struct intel_migrate *m, struct intel_gt *gt)
+{
+	struct intel_context *ce;
+
+	memset(m, 0, sizeof(*m));
+
+	ce = pinned_context(gt);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	m->context = ce;
+	return 0;
+}
+
+static int random_index(unsigned int max)
+{
+	return upper_32_bits(mul_u32_u32(get_random_u32(), max));
+}
+
+static struct intel_context *__migrate_engines(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engines[MAX_ENGINE_INSTANCE];
+	struct intel_engine_cs *engine;
+	unsigned int count, i;
+
+	count = 0;
+	for (i = 0; i < ARRAY_SIZE(gt->engine_class[COPY_ENGINE_CLASS]); i++) {
+		engine = gt->engine_class[COPY_ENGINE_CLASS][i];
+		if (engine_supports_migration(engine))
+			engines[count++] = engine;
+	}
+
+	return intel_context_create(engines[random_index(count)]);
+}
+
+struct intel_context *intel_migrate_create_context(struct intel_migrate *m)
+{
+	struct intel_context *ce;
+
+	/*
+	 * We randomly distribute contexts across the engines upon constrction,
+	 * as they all share the same pinned vm, and so in order to allow
+	 * multiple blits to run in parallel, we must construct each blit
+	 * to use a different range of the vm for its GTT. This has to be
+	 * known at construction, so we can not use the late greedy load
+	 * balancing of the virtual-engine.
+	 */
+	ce = __migrate_engines(m->context->engine->gt);
+	if (IS_ERR(ce))
+		return ce;
+
+	ce->ring = NULL;
+	ce->ring_size = SZ_256K;
+
+	i915_vm_put(ce->vm);
+	ce->vm = i915_vm_get(m->context->vm);
+
+	return ce;
+}
+
+static inline struct sgt_dma sg_sgt(struct scatterlist *sg)
+{
+	dma_addr_t addr = sg_dma_address(sg);
+
+	return (struct sgt_dma){ sg, addr, addr + sg_dma_len(sg) };
+}
+
+static int emit_no_arbitration(struct i915_request *rq)
+{
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/* Explicitly disable preemption for this request. */
+	*cs++ = MI_ARB_ON_OFF;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int max_pte_pkt_size(struct i915_request *rq, int pkt)
+{
+	struct intel_ring *ring = rq->ring;
+
+	pkt = min_t(int, pkt, (ring->space - rq->reserved_space) / sizeof(u32) + 5);
+	pkt = min_t(int, pkt, (ring->size - ring->emit) / sizeof(u32) + 5);
+
+	return pkt;
+}
+
+static int emit_pte(struct i915_request *rq,
+		    struct sgt_dma *it,
+		    enum i915_cache_level cache_level,
+		    bool is_lmem,
+		    u64 offset,
+		    int length)
+{
+	bool has_64K_pages = HAS_64K_PAGES(rq->engine->i915);
+	const u64 encode = rq->context->vm->pte_encode(0, cache_level,
+						       is_lmem ? PTE_LM : 0);
+	struct intel_ring *ring = rq->ring;
+	int pkt, dword_length;
+	u32 total = 0;
+	u32 page_size;
+	u32 *hdr, *cs;
+
+	GEM_BUG_ON(GRAPHICS_VER(rq->engine->i915) < 8);
+
+	page_size = I915_GTT_PAGE_SIZE;
+	dword_length = 0x400;
+
+	/* Compute the page directory offset for the target address range */
+	if (has_64K_pages) {
+		GEM_BUG_ON(!IS_ALIGNED(offset, SZ_2M));
+
+		offset /= SZ_2M;
+		offset *= SZ_64K;
+		offset += 3 * CHUNK_SZ;
+
+		if (is_lmem) {
+			page_size = I915_GTT_PAGE_SIZE_64K;
+			dword_length = 0x40;
+		}
+	} else {
+		offset >>= 12;
+		offset *= sizeof(u64);
+		offset += 2 * CHUNK_SZ;
+	}
+
+	offset += (u64)rq->engine->instance << 32;
+
+	cs = intel_ring_begin(rq, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/* Pack as many PTE updates as possible into a single MI command */
+	pkt = max_pte_pkt_size(rq, dword_length);
+
+	hdr = cs;
+	*cs++ = MI_STORE_DATA_IMM | REG_BIT(21); /* as qword elements */
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+
+	do {
+		if (cs - hdr >= pkt) {
+			int dword_rem;
+
+			*hdr += cs - hdr - 2;
+			*cs++ = MI_NOOP;
+
+			ring->emit = (void *)cs - ring->vaddr;
+			intel_ring_advance(rq, cs);
+			intel_ring_update_space(ring);
+
+			cs = intel_ring_begin(rq, 6);
+			if (IS_ERR(cs))
+				return PTR_ERR(cs);
+
+			dword_rem = dword_length;
+			if (has_64K_pages) {
+				if (IS_ALIGNED(total, SZ_2M)) {
+					offset = round_up(offset, SZ_64K);
+				} else {
+					dword_rem = SZ_2M - (total & (SZ_2M - 1));
+					dword_rem /= page_size;
+					dword_rem *= 2;
+				}
+			}
+
+			pkt = max_pte_pkt_size(rq, dword_rem);
+
+			hdr = cs;
+			*cs++ = MI_STORE_DATA_IMM | REG_BIT(21);
+			*cs++ = lower_32_bits(offset);
+			*cs++ = upper_32_bits(offset);
+		}
+
+		GEM_BUG_ON(!IS_ALIGNED(it->dma, page_size));
+
+		*cs++ = lower_32_bits(encode | it->dma);
+		*cs++ = upper_32_bits(encode | it->dma);
+
+		offset += 8;
+		total += page_size;
+
+		it->dma += page_size;
+		if (it->dma >= it->max) {
+			it->sg = __sg_next(it->sg);
+			if (!it->sg || sg_dma_len(it->sg) == 0)
+				break;
+
+			it->dma = sg_dma_address(it->sg);
+			it->max = it->dma + sg_dma_len(it->sg);
+		}
+	} while (total < length);
+
+	*hdr += cs - hdr - 2;
+	*cs++ = MI_NOOP;
+
+	ring->emit = (void *)cs - ring->vaddr;
+	intel_ring_advance(rq, cs);
+	intel_ring_update_space(ring);
+
+	return total;
+}
+
+static bool wa_1209644611_applies(int ver, u32 size)
+{
+	u32 height = size >> PAGE_SHIFT;
+
+	if (ver != 11)
+		return false;
+
+	return height % 4 == 3 && height <= 8;
+}
+
+/**
+ * DOC: Flat-CCS - Memory compression for Local memory
+ *
+ * On Xe-HP and later devices, we use dedicated compression control state (CCS)
+ * stored in local memory for each surface, to support the 3D and media
+ * compression formats.
+ *
+ * The memory required for the CCS of the entire local memory is 1/256 of the
+ * local memory size. So before the kernel boot, the required memory is reserved
+ * for the CCS data and a secure register will be programmed with the CCS base
+ * address.
+ *
+ * Flat CCS data needs to be cleared when a lmem object is allocated.
+ * And CCS data can be copied in and out of CCS region through
+ * XY_CTRL_SURF_COPY_BLT. CPU can't access the CCS data directly.
+ *
+ * I915 supports Flat-CCS on lmem only objects. When an objects has smem in
+ * its preference list, on memory pressure, i915 needs to migrate the lmem
+ * content into smem. If the lmem object is Flat-CCS compressed by userspace,
+ * then i915 needs to decompress it. But I915 lack the required information
+ * for such decompression. Hence I915 supports Flat-CCS only on lmem only objects.
+ *
+ * When we exhaust the lmem, Flat-CCS capable objects' lmem backing memory can
+ * be temporarily evicted to smem, along with the auxiliary CCS state, where
+ * it can be potentially swapped-out at a later point, if required.
+ * If userspace later touches the evicted pages, then we always move
+ * the backing memory back to lmem, which includes restoring the saved CCS state,
+ * and potentially performing any required swap-in.
+ *
+ * For the migration of the lmem objects with smem in placement list, such as
+ * {lmem, smem}, objects are treated as non Flat-CCS capable objects.
+ */
+
+static inline u32 *i915_flush_dw(u32 *cmd, u32 flags)
+{
+	*cmd++ = MI_FLUSH_DW | flags;
+	*cmd++ = 0;
+	*cmd++ = 0;
+
+	return cmd;
+}
+
+static int emit_copy_ccs(struct i915_request *rq,
+			 u32 dst_offset, u8 dst_access,
+			 u32 src_offset, u8 src_access, int size)
+{
+	struct drm_i915_private *i915 = rq->engine->i915;
+	int mocs = rq->engine->gt->mocs.uc_index << 1;
+	u32 num_ccs_blks;
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 12);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	num_ccs_blks = DIV_ROUND_UP(GET_CCS_BYTES(i915, size),
+				    NUM_CCS_BYTES_PER_BLOCK);
+	GEM_BUG_ON(num_ccs_blks > NUM_CCS_BLKS_PER_XFER);
+	cs = i915_flush_dw(cs, MI_FLUSH_DW_LLC | MI_FLUSH_DW_CCS);
+
+	/*
+	 * The XY_CTRL_SURF_COPY_BLT instruction is used to copy the CCS
+	 * data in and out of the CCS region.
+	 *
+	 * We can copy at most 1024 blocks of 256 bytes using one
+	 * XY_CTRL_SURF_COPY_BLT instruction.
+	 *
+	 * In case we need to copy more than 1024 blocks, we need to add
+	 * another instruction to the same batch buffer.
+	 *
+	 * 1024 blocks of 256 bytes of CCS represent a total 256KB of CCS.
+	 *
+	 * 256 KB of CCS represents 256 * 256 KB = 64 MB of LMEM.
+	 */
+	*cs++ = XY_CTRL_SURF_COPY_BLT |
+		src_access << SRC_ACCESS_TYPE_SHIFT |
+		dst_access << DST_ACCESS_TYPE_SHIFT |
+		((num_ccs_blks - 1) & CCS_SIZE_MASK) << CCS_SIZE_SHIFT;
+	*cs++ = src_offset;
+	*cs++ = rq->engine->instance |
+		FIELD_PREP(XY_CTRL_SURF_MOCS_MASK, mocs);
+	*cs++ = dst_offset;
+	*cs++ = rq->engine->instance |
+		FIELD_PREP(XY_CTRL_SURF_MOCS_MASK, mocs);
+
+	cs = i915_flush_dw(cs, MI_FLUSH_DW_LLC | MI_FLUSH_DW_CCS);
+	*cs++ = MI_NOOP;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int emit_copy(struct i915_request *rq,
+		     u32 dst_offset, u32 src_offset, int size)
+{
+	const int ver = GRAPHICS_VER(rq->engine->i915);
+	u32 instance = rq->engine->instance;
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, ver >= 8 ? 10 : 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	if (ver >= 9 && !wa_1209644611_applies(ver, size)) {
+		*cs++ = GEN9_XY_FAST_COPY_BLT_CMD | (10 - 2);
+		*cs++ = BLT_DEPTH_32 | PAGE_SIZE;
+		*cs++ = 0;
+		*cs++ = size >> PAGE_SHIFT << 16 | PAGE_SIZE / 4;
+		*cs++ = dst_offset;
+		*cs++ = instance;
+		*cs++ = 0;
+		*cs++ = PAGE_SIZE;
+		*cs++ = src_offset;
+		*cs++ = instance;
+	} else if (ver >= 8) {
+		*cs++ = XY_SRC_COPY_BLT_CMD | BLT_WRITE_RGBA | (10 - 2);
+		*cs++ = BLT_DEPTH_32 | BLT_ROP_SRC_COPY | PAGE_SIZE;
+		*cs++ = 0;
+		*cs++ = size >> PAGE_SHIFT << 16 | PAGE_SIZE / 4;
+		*cs++ = dst_offset;
+		*cs++ = instance;
+		*cs++ = 0;
+		*cs++ = PAGE_SIZE;
+		*cs++ = src_offset;
+		*cs++ = instance;
+	} else {
+		GEM_BUG_ON(instance);
+		*cs++ = SRC_COPY_BLT_CMD | BLT_WRITE_RGBA | (6 - 2);
+		*cs++ = BLT_DEPTH_32 | BLT_ROP_SRC_COPY | PAGE_SIZE;
+		*cs++ = size >> PAGE_SHIFT << 16 | PAGE_SIZE;
+		*cs++ = dst_offset;
+		*cs++ = PAGE_SIZE;
+		*cs++ = src_offset;
+	}
+
+	intel_ring_advance(rq, cs);
+	return 0;
+}
+
+static u64 scatter_list_length(struct scatterlist *sg)
+{
+	u64 len = 0;
+
+	while (sg && sg_dma_len(sg)) {
+		len += sg_dma_len(sg);
+		sg = sg_next(sg);
+	}
+
+	return len;
+}
+
+static int
+calculate_chunk_sz(struct drm_i915_private *i915, bool src_is_lmem,
+		   u64 bytes_to_cpy, u64 ccs_bytes_to_cpy)
+{
+	if (ccs_bytes_to_cpy && !src_is_lmem)
+		/*
+		 * When CHUNK_SZ is passed all the pages upto CHUNK_SZ
+		 * will be taken for the blt. in Flat-ccs supported
+		 * platform Smem obj will have more pages than required
+		 * for main meory hence limit it to the required size
+		 * for main memory
+		 */
+		return min_t(u64, bytes_to_cpy, CHUNK_SZ);
+	else
+		return CHUNK_SZ;
+}
+
+static void get_ccs_sg_sgt(struct sgt_dma *it, u64 bytes_to_cpy)
+{
+	u64 len;
+
+	do {
+		GEM_BUG_ON(!it->sg || !sg_dma_len(it->sg));
+		len = it->max - it->dma;
+		if (len > bytes_to_cpy) {
+			it->dma += bytes_to_cpy;
+			break;
+		}
+
+		bytes_to_cpy -= len;
+
+		it->sg = __sg_next(it->sg);
+		it->dma = sg_dma_address(it->sg);
+		it->max = it->dma + sg_dma_len(it->sg);
+	} while (bytes_to_cpy);
+}
+
+int
+intel_context_migrate_copy(struct intel_context *ce,
+			   const struct i915_deps *deps,
+			   struct scatterlist *src,
+			   enum i915_cache_level src_cache_level,
+			   bool src_is_lmem,
+			   struct scatterlist *dst,
+			   enum i915_cache_level dst_cache_level,
+			   bool dst_is_lmem,
+			   struct i915_request **out)
+{
+	struct sgt_dma it_src = sg_sgt(src), it_dst = sg_sgt(dst), it_ccs;
+	struct drm_i915_private *i915 = ce->engine->i915;
+	u64 ccs_bytes_to_cpy = 0, bytes_to_cpy;
+	enum i915_cache_level ccs_cache_level;
+	u32 src_offset, dst_offset;
+	u8 src_access, dst_access;
+	struct i915_request *rq;
+	u64 src_sz, dst_sz;
+	bool ccs_is_src, overwrite_ccs;
+	int err;
+
+	GEM_BUG_ON(ce->vm != ce->engine->gt->migrate.context->vm);
+	GEM_BUG_ON(IS_DGFX(ce->engine->i915) && (!src_is_lmem && !dst_is_lmem));
+	*out = NULL;
+
+	GEM_BUG_ON(ce->ring->size < SZ_64K);
+
+	src_sz = scatter_list_length(src);
+	bytes_to_cpy = src_sz;
+
+	if (HAS_FLAT_CCS(i915) && src_is_lmem ^ dst_is_lmem) {
+		src_access = !src_is_lmem && dst_is_lmem;
+		dst_access = !src_access;
+
+		dst_sz = scatter_list_length(dst);
+		if (src_is_lmem) {
+			it_ccs = it_dst;
+			ccs_cache_level = dst_cache_level;
+			ccs_is_src = false;
+		} else if (dst_is_lmem) {
+			bytes_to_cpy = dst_sz;
+			it_ccs = it_src;
+			ccs_cache_level = src_cache_level;
+			ccs_is_src = true;
+		}
+
+		/*
+		 * When there is a eviction of ccs needed smem will have the
+		 * extra pages for the ccs data
+		 *
+		 * TO-DO: Want to move the size mismatch check to a WARN_ON,
+		 * but still we have some requests of smem->lmem with same size.
+		 * Need to fix it.
+		 */
+		ccs_bytes_to_cpy = src_sz != dst_sz ? GET_CCS_BYTES(i915, bytes_to_cpy) : 0;
+		if (ccs_bytes_to_cpy)
+			get_ccs_sg_sgt(&it_ccs, bytes_to_cpy);
+	}
+
+	overwrite_ccs = HAS_FLAT_CCS(i915) && !ccs_bytes_to_cpy && dst_is_lmem;
+
+	src_offset = 0;
+	dst_offset = CHUNK_SZ;
+	if (HAS_64K_PAGES(ce->engine->i915)) {
+		src_offset = 0;
+		dst_offset = 0;
+		if (src_is_lmem)
+			src_offset = CHUNK_SZ;
+		if (dst_is_lmem)
+			dst_offset = 2 * CHUNK_SZ;
+	}
+
+	do {
+		int len;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_ce;
+		}
+
+		if (deps) {
+			err = i915_request_await_deps(rq, deps);
+			if (err)
+				goto out_rq;
+
+			if (rq->engine->emit_init_breadcrumb) {
+				err = rq->engine->emit_init_breadcrumb(rq);
+				if (err)
+					goto out_rq;
+			}
+
+			deps = NULL;
+		}
+
+		/* The PTE updates + copy must not be interrupted. */
+		err = emit_no_arbitration(rq);
+		if (err)
+			goto out_rq;
+
+		src_sz = calculate_chunk_sz(i915, src_is_lmem,
+					    bytes_to_cpy, ccs_bytes_to_cpy);
+
+		len = emit_pte(rq, &it_src, src_cache_level, src_is_lmem,
+			       src_offset, src_sz);
+		if (!len) {
+			err = -EINVAL;
+			goto out_rq;
+		}
+		if (len < 0) {
+			err = len;
+			goto out_rq;
+		}
+
+		err = emit_pte(rq, &it_dst, dst_cache_level, dst_is_lmem,
+			       dst_offset, len);
+		if (err < 0)
+			goto out_rq;
+		if (err < len) {
+			err = -EINVAL;
+			goto out_rq;
+		}
+
+		err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+		if (err)
+			goto out_rq;
+
+		err = emit_copy(rq, dst_offset,	src_offset, len);
+		if (err)
+			goto out_rq;
+
+		bytes_to_cpy -= len;
+
+		if (ccs_bytes_to_cpy) {
+			int ccs_sz;
+
+			err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+			if (err)
+				goto out_rq;
+
+			ccs_sz = GET_CCS_BYTES(i915, len);
+			err = emit_pte(rq, &it_ccs, ccs_cache_level, false,
+				       ccs_is_src ? src_offset : dst_offset,
+				       ccs_sz);
+			if (err < 0)
+				goto out_rq;
+			if (err < ccs_sz) {
+				err = -EINVAL;
+				goto out_rq;
+			}
+
+			err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+			if (err)
+				goto out_rq;
+
+			err = emit_copy_ccs(rq, dst_offset, dst_access,
+					    src_offset, src_access, len);
+			if (err)
+				goto out_rq;
+
+			err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+			if (err)
+				goto out_rq;
+			ccs_bytes_to_cpy -= ccs_sz;
+		} else if (overwrite_ccs) {
+			err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+			if (err)
+				goto out_rq;
+
+			/*
+			 * While we can't always restore/manage the CCS state,
+			 * we still need to ensure we don't leak the CCS state
+			 * from the previous user, so make sure we overwrite it
+			 * with something.
+			 */
+			err = emit_copy_ccs(rq, dst_offset, INDIRECT_ACCESS,
+					    dst_offset, DIRECT_ACCESS, len);
+			if (err)
+				goto out_rq;
+
+			err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+			if (err)
+				goto out_rq;
+		}
+
+		/* Arbitration is re-enabled between requests. */
+out_rq:
+		if (*out)
+			i915_request_put(*out);
+		*out = i915_request_get(rq);
+		i915_request_add(rq);
+
+		if (err)
+			break;
+
+		if (!bytes_to_cpy && !ccs_bytes_to_cpy) {
+			if (src_is_lmem)
+				WARN_ON(it_src.sg && sg_dma_len(it_src.sg));
+			else
+				WARN_ON(it_dst.sg && sg_dma_len(it_dst.sg));
+			break;
+		}
+
+		if (WARN_ON(!it_src.sg || !sg_dma_len(it_src.sg) ||
+			    !it_dst.sg || !sg_dma_len(it_dst.sg) ||
+			    (ccs_bytes_to_cpy && (!it_ccs.sg ||
+						  !sg_dma_len(it_ccs.sg))))) {
+			err = -EINVAL;
+			break;
+		}
+
+		cond_resched();
+	} while (1);
+
+out_ce:
+	return err;
+}
+
+static int emit_clear(struct i915_request *rq, u32 offset, int size,
+		      u32 value, bool is_lmem)
+{
+	struct drm_i915_private *i915 = rq->engine->i915;
+	int mocs = rq->engine->gt->mocs.uc_index << 1;
+	const int ver = GRAPHICS_VER(i915);
+	int ring_sz;
+	u32 *cs;
+
+	GEM_BUG_ON(size >> PAGE_SHIFT > S16_MAX);
+
+	if (HAS_FLAT_CCS(i915) && ver >= 12)
+		ring_sz = XY_FAST_COLOR_BLT_DW;
+	else if (ver >= 8)
+		ring_sz = 8;
+	else
+		ring_sz = 6;
+
+	cs = intel_ring_begin(rq, ring_sz);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	if (HAS_FLAT_CCS(i915) && ver >= 12) {
+		*cs++ = XY_FAST_COLOR_BLT_CMD | XY_FAST_COLOR_BLT_DEPTH_32 |
+			(XY_FAST_COLOR_BLT_DW - 2);
+		*cs++ = FIELD_PREP(XY_FAST_COLOR_BLT_MOCS_MASK, mocs) |
+			(PAGE_SIZE - 1);
+		*cs++ = 0;
+		*cs++ = size >> PAGE_SHIFT << 16 | PAGE_SIZE / 4;
+		*cs++ = offset;
+		*cs++ = rq->engine->instance;
+		*cs++ = !is_lmem << XY_FAST_COLOR_BLT_MEM_TYPE_SHIFT;
+		/* BG7 */
+		*cs++ = value;
+		*cs++ = 0;
+		*cs++ = 0;
+		*cs++ = 0;
+		/* BG11 */
+		*cs++ = 0;
+		*cs++ = 0;
+		/* BG13 */
+		*cs++ = 0;
+		*cs++ = 0;
+		*cs++ = 0;
+	} else if (ver >= 8) {
+		*cs++ = XY_COLOR_BLT_CMD | BLT_WRITE_RGBA | (7 - 2);
+		*cs++ = BLT_DEPTH_32 | BLT_ROP_COLOR_COPY | PAGE_SIZE;
+		*cs++ = 0;
+		*cs++ = size >> PAGE_SHIFT << 16 | PAGE_SIZE / 4;
+		*cs++ = offset;
+		*cs++ = rq->engine->instance;
+		*cs++ = value;
+		*cs++ = MI_NOOP;
+	} else {
+		*cs++ = XY_COLOR_BLT_CMD | BLT_WRITE_RGBA | (6 - 2);
+		*cs++ = BLT_DEPTH_32 | BLT_ROP_COLOR_COPY | PAGE_SIZE;
+		*cs++ = 0;
+		*cs++ = size >> PAGE_SHIFT << 16 | PAGE_SIZE / 4;
+		*cs++ = offset;
+		*cs++ = value;
+	}
+
+	intel_ring_advance(rq, cs);
+	return 0;
+}
+
+int
+intel_context_migrate_clear(struct intel_context *ce,
+			    const struct i915_deps *deps,
+			    struct scatterlist *sg,
+			    enum i915_cache_level cache_level,
+			    bool is_lmem,
+			    u32 value,
+			    struct i915_request **out)
+{
+	struct drm_i915_private *i915 = ce->engine->i915;
+	struct sgt_dma it = sg_sgt(sg);
+	struct i915_request *rq;
+	u32 offset;
+	int err;
+
+	GEM_BUG_ON(ce->vm != ce->engine->gt->migrate.context->vm);
+	*out = NULL;
+
+	GEM_BUG_ON(ce->ring->size < SZ_64K);
+
+	offset = 0;
+	if (HAS_64K_PAGES(i915) && is_lmem)
+		offset = CHUNK_SZ;
+
+	do {
+		int len;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_ce;
+		}
+
+		if (deps) {
+			err = i915_request_await_deps(rq, deps);
+			if (err)
+				goto out_rq;
+
+			if (rq->engine->emit_init_breadcrumb) {
+				err = rq->engine->emit_init_breadcrumb(rq);
+				if (err)
+					goto out_rq;
+			}
+
+			deps = NULL;
+		}
+
+		/* The PTE updates + clear must not be interrupted. */
+		err = emit_no_arbitration(rq);
+		if (err)
+			goto out_rq;
+
+		len = emit_pte(rq, &it, cache_level, is_lmem, offset, CHUNK_SZ);
+		if (len <= 0) {
+			err = len;
+			goto out_rq;
+		}
+
+		err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+		if (err)
+			goto out_rq;
+
+		err = emit_clear(rq, offset, len, value, is_lmem);
+		if (err)
+			goto out_rq;
+
+		if (HAS_FLAT_CCS(i915) && is_lmem && !value) {
+			/*
+			 * copy the content of memory into corresponding
+			 * ccs surface
+			 */
+			err = emit_copy_ccs(rq, offset, INDIRECT_ACCESS, offset,
+					    DIRECT_ACCESS, len);
+			if (err)
+				goto out_rq;
+		}
+
+		err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+
+		/* Arbitration is re-enabled between requests. */
+out_rq:
+		if (*out)
+			i915_request_put(*out);
+		*out = i915_request_get(rq);
+		i915_request_add(rq);
+		if (err || !it.sg || !sg_dma_len(it.sg))
+			break;
+
+		cond_resched();
+	} while (1);
+
+out_ce:
+	return err;
+}
+
+int intel_migrate_copy(struct intel_migrate *m,
+		       struct i915_gem_ww_ctx *ww,
+		       const struct i915_deps *deps,
+		       struct scatterlist *src,
+		       enum i915_cache_level src_cache_level,
+		       bool src_is_lmem,
+		       struct scatterlist *dst,
+		       enum i915_cache_level dst_cache_level,
+		       bool dst_is_lmem,
+		       struct i915_request **out)
+{
+	struct intel_context *ce;
+	int err;
+
+	*out = NULL;
+	if (!m->context)
+		return -ENODEV;
+
+	ce = intel_migrate_create_context(m);
+	if (IS_ERR(ce))
+		ce = intel_context_get(m->context);
+	GEM_BUG_ON(IS_ERR(ce));
+
+	err = intel_context_pin_ww(ce, ww);
+	if (err)
+		goto out;
+
+	err = intel_context_migrate_copy(ce, deps,
+					 src, src_cache_level, src_is_lmem,
+					 dst, dst_cache_level, dst_is_lmem,
+					 out);
+
+	intel_context_unpin(ce);
+out:
+	intel_context_put(ce);
+	return err;
+}
+
+int
+intel_migrate_clear(struct intel_migrate *m,
+		    struct i915_gem_ww_ctx *ww,
+		    const struct i915_deps *deps,
+		    struct scatterlist *sg,
+		    enum i915_cache_level cache_level,
+		    bool is_lmem,
+		    u32 value,
+		    struct i915_request **out)
+{
+	struct intel_context *ce;
+	int err;
+
+	*out = NULL;
+	if (!m->context)
+		return -ENODEV;
+
+	ce = intel_migrate_create_context(m);
+	if (IS_ERR(ce))
+		ce = intel_context_get(m->context);
+	GEM_BUG_ON(IS_ERR(ce));
+
+	err = intel_context_pin_ww(ce, ww);
+	if (err)
+		goto out;
+
+	err = intel_context_migrate_clear(ce, deps, sg, cache_level,
+					  is_lmem, value, out);
+
+	intel_context_unpin(ce);
+out:
+	intel_context_put(ce);
+	return err;
+}
+
+void intel_migrate_fini(struct intel_migrate *m)
+{
+	struct intel_context *ce;
+
+	ce = fetch_and_zero(&m->context);
+	if (!ce)
+		return;
+
+	intel_engine_destroy_pinned_context(ce);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_migrate.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.h b/drivers/gpu/drm/i915/gt/intel_migrate.h
new file mode 100644
index 000000000..ccc677ec4
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_migrate.h
@@ -0,0 +1,66 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __INTEL_MIGRATE__
+#define __INTEL_MIGRATE__
+
+#include <linux/types.h>
+
+#include "intel_migrate_types.h"
+
+struct dma_fence;
+struct i915_deps;
+struct i915_request;
+struct i915_gem_ww_ctx;
+struct intel_gt;
+struct scatterlist;
+enum i915_cache_level;
+
+int intel_migrate_init(struct intel_migrate *m, struct intel_gt *gt);
+
+struct intel_context *intel_migrate_create_context(struct intel_migrate *m);
+
+int intel_migrate_copy(struct intel_migrate *m,
+		       struct i915_gem_ww_ctx *ww,
+		       const struct i915_deps *deps,
+		       struct scatterlist *src,
+		       enum i915_cache_level src_cache_level,
+		       bool src_is_lmem,
+		       struct scatterlist *dst,
+		       enum i915_cache_level dst_cache_level,
+		       bool dst_is_lmem,
+		       struct i915_request **out);
+
+int intel_context_migrate_copy(struct intel_context *ce,
+			       const struct i915_deps *deps,
+			       struct scatterlist *src,
+			       enum i915_cache_level src_cache_level,
+			       bool src_is_lmem,
+			       struct scatterlist *dst,
+			       enum i915_cache_level dst_cache_level,
+			       bool dst_is_lmem,
+			       struct i915_request **out);
+
+int
+intel_migrate_clear(struct intel_migrate *m,
+		    struct i915_gem_ww_ctx *ww,
+		    const struct i915_deps *deps,
+		    struct scatterlist *sg,
+		    enum i915_cache_level cache_level,
+		    bool is_lmem,
+		    u32 value,
+		    struct i915_request **out);
+int
+intel_context_migrate_clear(struct intel_context *ce,
+			    const struct i915_deps *deps,
+			    struct scatterlist *sg,
+			    enum i915_cache_level cache_level,
+			    bool is_lmem,
+			    u32 value,
+			    struct i915_request **out);
+
+void intel_migrate_fini(struct intel_migrate *m);
+
+#endif /* __INTEL_MIGRATE__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_migrate_types.h b/drivers/gpu/drm/i915/gt/intel_migrate_types.h
new file mode 100644
index 000000000..d98230597
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_migrate_types.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __INTEL_MIGRATE_TYPES__
+#define __INTEL_MIGRATE_TYPES__
+
+struct intel_context;
+
+struct intel_migrate {
+	struct intel_context *context;
+};
+
+#endif /* __INTEL_MIGRATE_TYPES__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_mocs.c b/drivers/gpu/drm/i915/gt/intel_mocs.c
new file mode 100644
index 000000000..152244d7f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_mocs.c
@@ -0,0 +1,680 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2015 Intel Corporation
+ */
+
+#include "i915_drv.h"
+
+#include "intel_engine.h"
+#include "intel_gt.h"
+#include "intel_gt_regs.h"
+#include "intel_mocs.h"
+#include "intel_ring.h"
+
+/* structures required */
+struct drm_i915_mocs_entry {
+	u32 control_value;
+	u16 l3cc_value;
+	u16 used;
+};
+
+struct drm_i915_mocs_table {
+	unsigned int size;
+	unsigned int n_entries;
+	const struct drm_i915_mocs_entry *table;
+	u8 uc_index;
+	u8 wb_index; /* Only used on HAS_L3_CCS_READ() platforms */
+	u8 unused_entries_index;
+};
+
+/* Defines for the tables (XXX_MOCS_0 - XXX_MOCS_63) */
+#define _LE_CACHEABILITY(value)	((value) << 0)
+#define _LE_TGT_CACHE(value)	((value) << 2)
+#define LE_LRUM(value)		((value) << 4)
+#define LE_AOM(value)		((value) << 6)
+#define LE_RSC(value)		((value) << 7)
+#define LE_SCC(value)		((value) << 8)
+#define LE_PFM(value)		((value) << 11)
+#define LE_SCF(value)		((value) << 14)
+#define LE_COS(value)		((value) << 15)
+#define LE_SSE(value)		((value) << 17)
+
+/* Defines for the tables (LNCFMOCS0 - LNCFMOCS31) - two entries per word */
+#define L3_ESC(value)		((value) << 0)
+#define L3_SCC(value)		((value) << 1)
+#define _L3_CACHEABILITY(value)	((value) << 4)
+#define L3_GLBGO(value)		((value) << 6)
+#define L3_LKUP(value)		((value) << 7)
+
+/* Helper defines */
+#define GEN9_NUM_MOCS_ENTRIES	64  /* 63-64 are reserved, but configured. */
+#define PVC_NUM_MOCS_ENTRIES	3
+
+/* (e)LLC caching options */
+/*
+ * Note: LE_0_PAGETABLE works only up to Gen11; for newer gens it means
+ * the same as LE_UC
+ */
+#define LE_0_PAGETABLE		_LE_CACHEABILITY(0)
+#define LE_1_UC			_LE_CACHEABILITY(1)
+#define LE_2_WT			_LE_CACHEABILITY(2)
+#define LE_3_WB			_LE_CACHEABILITY(3)
+
+/* Target cache */
+#define LE_TC_0_PAGETABLE	_LE_TGT_CACHE(0)
+#define LE_TC_1_LLC		_LE_TGT_CACHE(1)
+#define LE_TC_2_LLC_ELLC	_LE_TGT_CACHE(2)
+#define LE_TC_3_LLC_ELLC_ALT	_LE_TGT_CACHE(3)
+
+/* L3 caching options */
+#define L3_0_DIRECT		_L3_CACHEABILITY(0)
+#define L3_1_UC			_L3_CACHEABILITY(1)
+#define L3_2_RESERVED		_L3_CACHEABILITY(2)
+#define L3_3_WB			_L3_CACHEABILITY(3)
+
+#define MOCS_ENTRY(__idx, __control_value, __l3cc_value) \
+	[__idx] = { \
+		.control_value = __control_value, \
+		.l3cc_value = __l3cc_value, \
+		.used = 1, \
+	}
+
+/*
+ * MOCS tables
+ *
+ * These are the MOCS tables that are programmed across all the rings.
+ * The control value is programmed to all the rings that support the
+ * MOCS registers. While the l3cc_values are only programmed to the
+ * LNCFCMOCS0 - LNCFCMOCS32 registers.
+ *
+ * These tables are intended to be kept reasonably consistent across
+ * HW platforms, and for ICL+, be identical across OSes. To achieve
+ * that, for Icelake and above, list of entries is published as part
+ * of bspec.
+ *
+ * Entries not part of the following tables are undefined as far as
+ * userspace is concerned and shouldn't be relied upon.  For Gen < 12
+ * they will be initialized to PTE. Gen >= 12 don't have a setting for
+ * PTE and those platforms except TGL/RKL will be initialized L3 WB to
+ * catch accidental use of reserved and unused mocs indexes.
+ *
+ * The last few entries are reserved by the hardware. For ICL+ they
+ * should be initialized according to bspec and never used, for older
+ * platforms they should never be written to.
+ *
+ * NOTE1: These tables are part of bspec and defined as part of hardware
+ *       interface for ICL+. For older platforms, they are part of kernel
+ *       ABI. It is expected that, for specific hardware platform, existing
+ *       entries will remain constant and the table will only be updated by
+ *       adding new entries, filling unused positions.
+ *
+ * NOTE2: For GEN >= 12 except TGL and RKL, reserved and unspecified MOCS
+ *       indices have been set to L3 WB. These reserved entries should never
+ *       be used, they may be changed to low performant variants with better
+ *       coherency in the future if more entries are needed.
+ *       For TGL/RKL, all the unspecified MOCS indexes are mapped to L3 UC.
+ */
+#define GEN9_MOCS_ENTRIES \
+	MOCS_ENTRY(I915_MOCS_UNCACHED, \
+		   LE_1_UC | LE_TC_2_LLC_ELLC, \
+		   L3_1_UC), \
+	MOCS_ENTRY(I915_MOCS_PTE, \
+		   LE_0_PAGETABLE | LE_TC_0_PAGETABLE | LE_LRUM(3), \
+		   L3_3_WB)
+
+static const struct drm_i915_mocs_entry skl_mocs_table[] = {
+	GEN9_MOCS_ENTRIES,
+	MOCS_ENTRY(I915_MOCS_CACHED,
+		   LE_3_WB | LE_TC_2_LLC_ELLC | LE_LRUM(3),
+		   L3_3_WB),
+
+	/*
+	 * mocs:63
+	 * - used by the L3 for all of its evictions.
+	 *   Thus it is expected to allow LLC cacheability to enable coherent
+	 *   flows to be maintained.
+	 * - used to force L3 uncachable cycles.
+	 *   Thus it is expected to make the surface L3 uncacheable.
+	 */
+	MOCS_ENTRY(63,
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3),
+		   L3_1_UC)
+};
+
+/* NOTE: the LE_TGT_CACHE is not used on Broxton */
+static const struct drm_i915_mocs_entry broxton_mocs_table[] = {
+	GEN9_MOCS_ENTRIES,
+	MOCS_ENTRY(I915_MOCS_CACHED,
+		   LE_1_UC | LE_TC_2_LLC_ELLC | LE_LRUM(3),
+		   L3_3_WB)
+};
+
+#define GEN11_MOCS_ENTRIES \
+	/* Entries 0 and 1 are defined per-platform */ \
+	/* Base - L3 + LLC */ \
+	MOCS_ENTRY(2, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3), \
+		   L3_3_WB), \
+	/* Base - Uncached */ \
+	MOCS_ENTRY(3, \
+		   LE_1_UC | LE_TC_1_LLC, \
+		   L3_1_UC), \
+	/* Base - L3 */ \
+	MOCS_ENTRY(4, \
+		   LE_1_UC | LE_TC_1_LLC, \
+		   L3_3_WB), \
+	/* Base - LLC */ \
+	MOCS_ENTRY(5, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3), \
+		   L3_1_UC), \
+	/* Age 0 - LLC */ \
+	MOCS_ENTRY(6, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(1), \
+		   L3_1_UC), \
+	/* Age 0 - L3 + LLC */ \
+	MOCS_ENTRY(7, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(1), \
+		   L3_3_WB), \
+	/* Age: Don't Chg. - LLC */ \
+	MOCS_ENTRY(8, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(2), \
+		   L3_1_UC), \
+	/* Age: Don't Chg. - L3 + LLC */ \
+	MOCS_ENTRY(9, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(2), \
+		   L3_3_WB), \
+	/* No AOM - LLC */ \
+	MOCS_ENTRY(10, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_AOM(1), \
+		   L3_1_UC), \
+	/* No AOM - L3 + LLC */ \
+	MOCS_ENTRY(11, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_AOM(1), \
+		   L3_3_WB), \
+	/* No AOM; Age 0 - LLC */ \
+	MOCS_ENTRY(12, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(1) | LE_AOM(1), \
+		   L3_1_UC), \
+	/* No AOM; Age 0 - L3 + LLC */ \
+	MOCS_ENTRY(13, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(1) | LE_AOM(1), \
+		   L3_3_WB), \
+	/* No AOM; Age:DC - LLC */ \
+	MOCS_ENTRY(14, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(2) | LE_AOM(1), \
+		   L3_1_UC), \
+	/* No AOM; Age:DC - L3 + LLC */ \
+	MOCS_ENTRY(15, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(2) | LE_AOM(1), \
+		   L3_3_WB), \
+	/* Bypass LLC - Uncached (EHL+) */ \
+	MOCS_ENTRY(16, \
+		   LE_1_UC | LE_TC_1_LLC | LE_SCF(1), \
+		   L3_1_UC), \
+	/* Bypass LLC - L3 (Read-Only) (EHL+) */ \
+	MOCS_ENTRY(17, \
+		   LE_1_UC | LE_TC_1_LLC | LE_SCF(1), \
+		   L3_3_WB), \
+	/* Self-Snoop - L3 + LLC */ \
+	MOCS_ENTRY(18, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_SSE(3), \
+		   L3_3_WB), \
+	/* Skip Caching - L3 + LLC(12.5%) */ \
+	MOCS_ENTRY(19, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_SCC(7), \
+		   L3_3_WB), \
+	/* Skip Caching - L3 + LLC(25%) */ \
+	MOCS_ENTRY(20, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_SCC(3), \
+		   L3_3_WB), \
+	/* Skip Caching - L3 + LLC(50%) */ \
+	MOCS_ENTRY(21, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_SCC(1), \
+		   L3_3_WB), \
+	/* Skip Caching - L3 + LLC(75%) */ \
+	MOCS_ENTRY(22, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_RSC(1) | LE_SCC(3), \
+		   L3_3_WB), \
+	/* Skip Caching - L3 + LLC(87.5%) */ \
+	MOCS_ENTRY(23, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_RSC(1) | LE_SCC(7), \
+		   L3_3_WB), \
+	/* HW Reserved - SW program but never use */ \
+	MOCS_ENTRY(62, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3), \
+		   L3_1_UC), \
+	/* HW Reserved - SW program but never use */ \
+	MOCS_ENTRY(63, \
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3), \
+		   L3_1_UC)
+
+static const struct drm_i915_mocs_entry tgl_mocs_table[] = {
+	/*
+	 * NOTE:
+	 * Reserved and unspecified MOCS indices have been set to (L3 + LCC).
+	 * These reserved entries should never be used, they may be changed
+	 * to low performant variants with better coherency in the future if
+	 * more entries are needed. We are programming index I915_MOCS_PTE(1)
+	 * only, __init_mocs_table() take care to program unused index with
+	 * this entry.
+	 */
+	MOCS_ENTRY(I915_MOCS_PTE,
+		   LE_0_PAGETABLE | LE_TC_0_PAGETABLE,
+		   L3_1_UC),
+	GEN11_MOCS_ENTRIES,
+
+	/* Implicitly enable L1 - HDC:L1 + L3 + LLC */
+	MOCS_ENTRY(48,
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3),
+		   L3_3_WB),
+	/* Implicitly enable L1 - HDC:L1 + L3 */
+	MOCS_ENTRY(49,
+		   LE_1_UC | LE_TC_1_LLC,
+		   L3_3_WB),
+	/* Implicitly enable L1 - HDC:L1 + LLC */
+	MOCS_ENTRY(50,
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3),
+		   L3_1_UC),
+	/* Implicitly enable L1 - HDC:L1 */
+	MOCS_ENTRY(51,
+		   LE_1_UC | LE_TC_1_LLC,
+		   L3_1_UC),
+	/* HW Special Case (CCS) */
+	MOCS_ENTRY(60,
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3),
+		   L3_1_UC),
+	/* HW Special Case (Displayable) */
+	MOCS_ENTRY(61,
+		   LE_1_UC | LE_TC_1_LLC,
+		   L3_3_WB),
+};
+
+static const struct drm_i915_mocs_entry icl_mocs_table[] = {
+	/* Base - Uncached (Deprecated) */
+	MOCS_ENTRY(I915_MOCS_UNCACHED,
+		   LE_1_UC | LE_TC_1_LLC,
+		   L3_1_UC),
+	/* Base - L3 + LeCC:PAT (Deprecated) */
+	MOCS_ENTRY(I915_MOCS_PTE,
+		   LE_0_PAGETABLE | LE_TC_0_PAGETABLE,
+		   L3_3_WB),
+
+	GEN11_MOCS_ENTRIES
+};
+
+static const struct drm_i915_mocs_entry dg1_mocs_table[] = {
+
+	/* UC */
+	MOCS_ENTRY(1, 0, L3_1_UC),
+	/* WB - L3 */
+	MOCS_ENTRY(5, 0, L3_3_WB),
+	/* WB - L3 50% */
+	MOCS_ENTRY(6, 0, L3_ESC(1) | L3_SCC(1) | L3_3_WB),
+	/* WB - L3 25% */
+	MOCS_ENTRY(7, 0, L3_ESC(1) | L3_SCC(3) | L3_3_WB),
+	/* WB - L3 12.5% */
+	MOCS_ENTRY(8, 0, L3_ESC(1) | L3_SCC(7) | L3_3_WB),
+
+	/* HDC:L1 + L3 */
+	MOCS_ENTRY(48, 0, L3_3_WB),
+	/* HDC:L1 */
+	MOCS_ENTRY(49, 0, L3_1_UC),
+
+	/* HW Reserved */
+	MOCS_ENTRY(60, 0, L3_1_UC),
+	MOCS_ENTRY(61, 0, L3_1_UC),
+	MOCS_ENTRY(62, 0, L3_1_UC),
+	MOCS_ENTRY(63, 0, L3_1_UC),
+};
+
+static const struct drm_i915_mocs_entry gen12_mocs_table[] = {
+	GEN11_MOCS_ENTRIES,
+	/* Implicitly enable L1 - HDC:L1 + L3 + LLC */
+	MOCS_ENTRY(48,
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3),
+		   L3_3_WB),
+	/* Implicitly enable L1 - HDC:L1 + L3 */
+	MOCS_ENTRY(49,
+		   LE_1_UC | LE_TC_1_LLC,
+		   L3_3_WB),
+	/* Implicitly enable L1 - HDC:L1 + LLC */
+	MOCS_ENTRY(50,
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3),
+		   L3_1_UC),
+	/* Implicitly enable L1 - HDC:L1 */
+	MOCS_ENTRY(51,
+		   LE_1_UC | LE_TC_1_LLC,
+		   L3_1_UC),
+	/* HW Special Case (CCS) */
+	MOCS_ENTRY(60,
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3),
+		   L3_1_UC),
+	/* HW Special Case (Displayable) */
+	MOCS_ENTRY(61,
+		   LE_1_UC | LE_TC_1_LLC,
+		   L3_3_WB),
+};
+
+static const struct drm_i915_mocs_entry xehpsdv_mocs_table[] = {
+	/* wa_1608975824 */
+	MOCS_ENTRY(0, 0, L3_3_WB | L3_LKUP(1)),
+
+	/* UC - Coherent; GO:L3 */
+	MOCS_ENTRY(1, 0, L3_1_UC | L3_LKUP(1)),
+	/* UC - Coherent; GO:Memory */
+	MOCS_ENTRY(2, 0, L3_1_UC | L3_GLBGO(1) | L3_LKUP(1)),
+	/* UC - Non-Coherent; GO:Memory */
+	MOCS_ENTRY(3, 0, L3_1_UC | L3_GLBGO(1)),
+	/* UC - Non-Coherent; GO:L3 */
+	MOCS_ENTRY(4, 0, L3_1_UC),
+
+	/* WB */
+	MOCS_ENTRY(5, 0, L3_3_WB | L3_LKUP(1)),
+
+	/* HW Reserved - SW program but never use. */
+	MOCS_ENTRY(48, 0, L3_3_WB | L3_LKUP(1)),
+	MOCS_ENTRY(49, 0, L3_1_UC | L3_LKUP(1)),
+	MOCS_ENTRY(60, 0, L3_1_UC),
+	MOCS_ENTRY(61, 0, L3_1_UC),
+	MOCS_ENTRY(62, 0, L3_1_UC),
+	MOCS_ENTRY(63, 0, L3_1_UC),
+};
+
+static const struct drm_i915_mocs_entry dg2_mocs_table[] = {
+	/* UC - Coherent; GO:L3 */
+	MOCS_ENTRY(0, 0, L3_1_UC | L3_LKUP(1)),
+	/* UC - Coherent; GO:Memory */
+	MOCS_ENTRY(1, 0, L3_1_UC | L3_GLBGO(1) | L3_LKUP(1)),
+	/* UC - Non-Coherent; GO:Memory */
+	MOCS_ENTRY(2, 0, L3_1_UC | L3_GLBGO(1)),
+
+	/* WB - LC */
+	MOCS_ENTRY(3, 0, L3_3_WB | L3_LKUP(1)),
+};
+
+static const struct drm_i915_mocs_entry dg2_mocs_table_g10_ax[] = {
+	/* Wa_14011441408: Set Go to Memory for MOCS#0 */
+	MOCS_ENTRY(0, 0, L3_1_UC | L3_GLBGO(1) | L3_LKUP(1)),
+	/* UC - Coherent; GO:Memory */
+	MOCS_ENTRY(1, 0, L3_1_UC | L3_GLBGO(1) | L3_LKUP(1)),
+	/* UC - Non-Coherent; GO:Memory */
+	MOCS_ENTRY(2, 0, L3_1_UC | L3_GLBGO(1)),
+
+	/* WB - LC */
+	MOCS_ENTRY(3, 0, L3_3_WB | L3_LKUP(1)),
+};
+
+static const struct drm_i915_mocs_entry pvc_mocs_table[] = {
+	/* Error */
+	MOCS_ENTRY(0, 0, L3_3_WB),
+
+	/* UC */
+	MOCS_ENTRY(1, 0, L3_1_UC),
+
+	/* WB */
+	MOCS_ENTRY(2, 0, L3_3_WB),
+};
+
+enum {
+	HAS_GLOBAL_MOCS = BIT(0),
+	HAS_ENGINE_MOCS = BIT(1),
+	HAS_RENDER_L3CC = BIT(2),
+};
+
+static bool has_l3cc(const struct drm_i915_private *i915)
+{
+	return true;
+}
+
+static bool has_global_mocs(const struct drm_i915_private *i915)
+{
+	return HAS_GLOBAL_MOCS_REGISTERS(i915);
+}
+
+static bool has_mocs(const struct drm_i915_private *i915)
+{
+	return !IS_DGFX(i915);
+}
+
+static unsigned int get_mocs_settings(const struct drm_i915_private *i915,
+				      struct drm_i915_mocs_table *table)
+{
+	unsigned int flags;
+
+	memset(table, 0, sizeof(struct drm_i915_mocs_table));
+
+	table->unused_entries_index = I915_MOCS_PTE;
+	if (IS_PONTEVECCHIO(i915)) {
+		table->size = ARRAY_SIZE(pvc_mocs_table);
+		table->table = pvc_mocs_table;
+		table->n_entries = PVC_NUM_MOCS_ENTRIES;
+		table->uc_index = 1;
+		table->wb_index = 2;
+		table->unused_entries_index = 2;
+	} else if (IS_DG2(i915)) {
+		if (IS_DG2_GRAPHICS_STEP(i915, G10, STEP_A0, STEP_B0)) {
+			table->size = ARRAY_SIZE(dg2_mocs_table_g10_ax);
+			table->table = dg2_mocs_table_g10_ax;
+		} else {
+			table->size = ARRAY_SIZE(dg2_mocs_table);
+			table->table = dg2_mocs_table;
+		}
+		table->uc_index = 1;
+		table->n_entries = GEN9_NUM_MOCS_ENTRIES;
+		table->unused_entries_index = 3;
+	} else if (IS_XEHPSDV(i915)) {
+		table->size = ARRAY_SIZE(xehpsdv_mocs_table);
+		table->table = xehpsdv_mocs_table;
+		table->uc_index = 2;
+		table->n_entries = GEN9_NUM_MOCS_ENTRIES;
+		table->unused_entries_index = 5;
+	} else if (IS_DG1(i915)) {
+		table->size = ARRAY_SIZE(dg1_mocs_table);
+		table->table = dg1_mocs_table;
+		table->uc_index = 1;
+		table->n_entries = GEN9_NUM_MOCS_ENTRIES;
+		table->uc_index = 1;
+		table->unused_entries_index = 5;
+	} else if (IS_TIGERLAKE(i915) || IS_ROCKETLAKE(i915)) {
+		/* For TGL/RKL, Can't be changed now for ABI reasons */
+		table->size  = ARRAY_SIZE(tgl_mocs_table);
+		table->table = tgl_mocs_table;
+		table->n_entries = GEN9_NUM_MOCS_ENTRIES;
+		table->uc_index = 3;
+	} else if (GRAPHICS_VER(i915) >= 12) {
+		table->size  = ARRAY_SIZE(gen12_mocs_table);
+		table->table = gen12_mocs_table;
+		table->n_entries = GEN9_NUM_MOCS_ENTRIES;
+		table->uc_index = 3;
+		table->unused_entries_index = 2;
+	} else if (GRAPHICS_VER(i915) == 11) {
+		table->size  = ARRAY_SIZE(icl_mocs_table);
+		table->table = icl_mocs_table;
+		table->n_entries = GEN9_NUM_MOCS_ENTRIES;
+	} else if (IS_GEN9_BC(i915)) {
+		table->size  = ARRAY_SIZE(skl_mocs_table);
+		table->n_entries = GEN9_NUM_MOCS_ENTRIES;
+		table->table = skl_mocs_table;
+	} else if (IS_GEN9_LP(i915)) {
+		table->size  = ARRAY_SIZE(broxton_mocs_table);
+		table->n_entries = GEN9_NUM_MOCS_ENTRIES;
+		table->table = broxton_mocs_table;
+	} else {
+		drm_WARN_ONCE(&i915->drm, GRAPHICS_VER(i915) >= 9,
+			      "Platform that should have a MOCS table does not.\n");
+		return 0;
+	}
+
+	if (GEM_DEBUG_WARN_ON(table->size > table->n_entries))
+		return 0;
+
+	/* WaDisableSkipCaching:skl,bxt,kbl,glk */
+	if (GRAPHICS_VER(i915) == 9) {
+		int i;
+
+		for (i = 0; i < table->size; i++)
+			if (GEM_DEBUG_WARN_ON(table->table[i].l3cc_value &
+					      (L3_ESC(1) | L3_SCC(0x7))))
+				return 0;
+	}
+
+	flags = 0;
+	if (has_mocs(i915)) {
+		if (has_global_mocs(i915))
+			flags |= HAS_GLOBAL_MOCS;
+		else
+			flags |= HAS_ENGINE_MOCS;
+	}
+	if (has_l3cc(i915))
+		flags |= HAS_RENDER_L3CC;
+
+	return flags;
+}
+
+/*
+ * Get control_value from MOCS entry taking into account when it's not used
+ * then if unused_entries_index is non-zero then its value will be returned
+ * otherwise I915_MOCS_PTE's value is returned in this case.
+ */
+static u32 get_entry_control(const struct drm_i915_mocs_table *table,
+			     unsigned int index)
+{
+	if (index < table->size && table->table[index].used)
+		return table->table[index].control_value;
+	return table->table[table->unused_entries_index].control_value;
+}
+
+#define for_each_mocs(mocs, t, i) \
+	for (i = 0; \
+	     i < (t)->n_entries ? (mocs = get_entry_control((t), i)), 1 : 0;\
+	     i++)
+
+static void __init_mocs_table(struct intel_uncore *uncore,
+			      const struct drm_i915_mocs_table *table,
+			      u32 addr)
+{
+	unsigned int i;
+	u32 mocs;
+
+	drm_WARN_ONCE(&uncore->i915->drm, !table->unused_entries_index,
+		      "Unused entries index should have been defined\n");
+	for_each_mocs(mocs, table, i)
+		intel_uncore_write_fw(uncore, _MMIO(addr + i * 4), mocs);
+}
+
+static u32 mocs_offset(const struct intel_engine_cs *engine)
+{
+	static const u32 offset[] = {
+		[RCS0]  =  __GEN9_RCS0_MOCS0,
+		[VCS0]  =  __GEN9_VCS0_MOCS0,
+		[VCS1]  =  __GEN9_VCS1_MOCS0,
+		[VECS0] =  __GEN9_VECS0_MOCS0,
+		[BCS0]  =  __GEN9_BCS0_MOCS0,
+		[VCS2]  = __GEN11_VCS2_MOCS0,
+	};
+
+	GEM_BUG_ON(engine->id >= ARRAY_SIZE(offset));
+	return offset[engine->id];
+}
+
+static void init_mocs_table(struct intel_engine_cs *engine,
+			    const struct drm_i915_mocs_table *table)
+{
+	__init_mocs_table(engine->uncore, table, mocs_offset(engine));
+}
+
+/*
+ * Get l3cc_value from MOCS entry taking into account when it's not used
+ * then if unused_entries_index is not zero then its value will be returned
+ * otherwise I915_MOCS_PTE's value is returned in this case.
+ */
+static u16 get_entry_l3cc(const struct drm_i915_mocs_table *table,
+			  unsigned int index)
+{
+	if (index < table->size && table->table[index].used)
+		return table->table[index].l3cc_value;
+	return table->table[table->unused_entries_index].l3cc_value;
+}
+
+static u32 l3cc_combine(u16 low, u16 high)
+{
+	return low | (u32)high << 16;
+}
+
+#define for_each_l3cc(l3cc, t, i) \
+	for (i = 0; \
+	     i < ((t)->n_entries + 1) / 2 ? \
+	     (l3cc = l3cc_combine(get_entry_l3cc((t), 2 * i), \
+				  get_entry_l3cc((t), 2 * i + 1))), 1 : \
+	     0; \
+	     i++)
+
+static void init_l3cc_table(struct intel_uncore *uncore,
+			    const struct drm_i915_mocs_table *table)
+{
+	unsigned int i;
+	u32 l3cc;
+
+	for_each_l3cc(l3cc, table, i)
+		intel_uncore_write_fw(uncore, GEN9_LNCFCMOCS(i), l3cc);
+}
+
+void intel_mocs_init_engine(struct intel_engine_cs *engine)
+{
+	struct drm_i915_mocs_table table;
+	unsigned int flags;
+
+	/* Called under a blanket forcewake */
+	assert_forcewakes_active(engine->uncore, FORCEWAKE_ALL);
+
+	flags = get_mocs_settings(engine->i915, &table);
+	if (!flags)
+		return;
+
+	/* Platforms with global MOCS do not need per-engine initialization. */
+	if (flags & HAS_ENGINE_MOCS)
+		init_mocs_table(engine, &table);
+
+	if (flags & HAS_RENDER_L3CC && engine->class == RENDER_CLASS)
+		init_l3cc_table(engine->uncore, &table);
+}
+
+static u32 global_mocs_offset(void)
+{
+	return i915_mmio_reg_offset(GEN12_GLOBAL_MOCS(0));
+}
+
+void intel_set_mocs_index(struct intel_gt *gt)
+{
+	struct drm_i915_mocs_table table;
+
+	get_mocs_settings(gt->i915, &table);
+	gt->mocs.uc_index = table.uc_index;
+	if (HAS_L3_CCS_READ(gt->i915))
+		gt->mocs.wb_index = table.wb_index;
+}
+
+void intel_mocs_init(struct intel_gt *gt)
+{
+	struct drm_i915_mocs_table table;
+	unsigned int flags;
+
+	/*
+	 * LLC and eDRAM control values are not applicable to dgfx
+	 */
+	flags = get_mocs_settings(gt->i915, &table);
+	if (flags & HAS_GLOBAL_MOCS)
+		__init_mocs_table(gt->uncore, &table, global_mocs_offset());
+
+	/*
+	 * Initialize the L3CC table as part of mocs initalization to make
+	 * sure the LNCFCMOCSx registers are programmed for the subsequent
+	 * memory transactions including guc transactions
+	 */
+	if (flags & HAS_RENDER_L3CC)
+		init_l3cc_table(gt->uncore, &table);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_mocs.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_mocs.h b/drivers/gpu/drm/i915/gt/intel_mocs.h
new file mode 100644
index 000000000..76db82721
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_mocs.h
@@ -0,0 +1,41 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2015 Intel Corporation
+ */
+
+#ifndef INTEL_MOCS_H
+#define INTEL_MOCS_H
+
+/**
+ * DOC: Memory Objects Control State (MOCS)
+ *
+ * Motivation:
+ * In previous Gens the MOCS settings was a value that was set by user land as
+ * part of the batch. In Gen9 this has changed to be a single table (per ring)
+ * that all batches now reference by index instead of programming the MOCS
+ * directly.
+ *
+ * The one wrinkle in this is that only PART of the MOCS tables are included
+ * in context (The GFX_MOCS_0 - GFX_MOCS_64 and the LNCFCMOCS0 - LNCFCMOCS32
+ * registers). The rest are not (the settings for the other rings).
+ *
+ * This table needs to be set at system start-up because the way the table
+ * interacts with the contexts and the GmmLib interface.
+ *
+ *
+ * Implementation:
+ *
+ * The tables (one per supported platform) are defined in intel_mocs.c
+ * and are programmed in the first batch after the context is loaded
+ * (with the hardware workarounds). This will then let the usual
+ * context handling keep the MOCS in step.
+ */
+
+struct intel_engine_cs;
+struct intel_gt;
+
+void intel_mocs_init(struct intel_gt *gt);
+void intel_mocs_init_engine(struct intel_engine_cs *engine);
+void intel_set_mocs_index(struct intel_gt *gt);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_ppgtt.c b/drivers/gpu/drm/i915/gt/intel_ppgtt.c
new file mode 100644
index 000000000..7ecfa672f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_ppgtt.c
@@ -0,0 +1,323 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/slab.h>
+
+#include "gem/i915_gem_lmem.h"
+
+#include "i915_trace.h"
+#include "intel_gtt.h"
+#include "gen6_ppgtt.h"
+#include "gen8_ppgtt.h"
+
+struct i915_page_table *alloc_pt(struct i915_address_space *vm, int sz)
+{
+	struct i915_page_table *pt;
+
+	pt = kmalloc(sizeof(*pt), I915_GFP_ALLOW_FAIL);
+	if (unlikely(!pt))
+		return ERR_PTR(-ENOMEM);
+
+	pt->base = vm->alloc_pt_dma(vm, sz);
+	if (IS_ERR(pt->base)) {
+		kfree(pt);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	pt->is_compact = false;
+	atomic_set(&pt->used, 0);
+	return pt;
+}
+
+struct i915_page_directory *__alloc_pd(int count)
+{
+	struct i915_page_directory *pd;
+
+	pd = kzalloc(sizeof(*pd), I915_GFP_ALLOW_FAIL);
+	if (unlikely(!pd))
+		return NULL;
+
+	pd->entry = kcalloc(count, sizeof(*pd->entry), I915_GFP_ALLOW_FAIL);
+	if (unlikely(!pd->entry)) {
+		kfree(pd);
+		return NULL;
+	}
+
+	spin_lock_init(&pd->lock);
+	return pd;
+}
+
+struct i915_page_directory *alloc_pd(struct i915_address_space *vm)
+{
+	struct i915_page_directory *pd;
+
+	pd = __alloc_pd(I915_PDES);
+	if (unlikely(!pd))
+		return ERR_PTR(-ENOMEM);
+
+	pd->pt.base = vm->alloc_pt_dma(vm, I915_GTT_PAGE_SIZE_4K);
+	if (IS_ERR(pd->pt.base)) {
+		kfree(pd->entry);
+		kfree(pd);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	return pd;
+}
+
+void free_px(struct i915_address_space *vm, struct i915_page_table *pt, int lvl)
+{
+	BUILD_BUG_ON(offsetof(struct i915_page_directory, pt));
+
+	if (lvl) {
+		struct i915_page_directory *pd =
+			container_of(pt, typeof(*pd), pt);
+		kfree(pd->entry);
+	}
+
+	if (pt->base)
+		i915_gem_object_put(pt->base);
+
+	kfree(pt);
+}
+
+static void
+write_dma_entry(struct drm_i915_gem_object * const pdma,
+		const unsigned short idx,
+		const u64 encoded_entry)
+{
+	u64 * const vaddr = __px_vaddr(pdma);
+
+	vaddr[idx] = encoded_entry;
+	drm_clflush_virt_range(&vaddr[idx], sizeof(u64));
+}
+
+void
+__set_pd_entry(struct i915_page_directory * const pd,
+	       const unsigned short idx,
+	       struct i915_page_table * const to,
+	       u64 (*encode)(const dma_addr_t, const enum i915_cache_level))
+{
+	/* Each thread pre-pins the pd, and we may have a thread per pde. */
+	GEM_BUG_ON(atomic_read(px_used(pd)) > NALLOC * I915_PDES);
+
+	atomic_inc(px_used(pd));
+	pd->entry[idx] = to;
+	write_dma_entry(px_base(pd), idx, encode(px_dma(to), I915_CACHE_LLC));
+}
+
+void
+clear_pd_entry(struct i915_page_directory * const pd,
+	       const unsigned short idx,
+	       const struct drm_i915_gem_object * const scratch)
+{
+	GEM_BUG_ON(atomic_read(px_used(pd)) == 0);
+
+	write_dma_entry(px_base(pd), idx, scratch->encode);
+	pd->entry[idx] = NULL;
+	atomic_dec(px_used(pd));
+}
+
+bool
+release_pd_entry(struct i915_page_directory * const pd,
+		 const unsigned short idx,
+		 struct i915_page_table * const pt,
+		 const struct drm_i915_gem_object * const scratch)
+{
+	bool free = false;
+
+	if (atomic_add_unless(&pt->used, -1, 1))
+		return false;
+
+	spin_lock(&pd->lock);
+	if (atomic_dec_and_test(&pt->used)) {
+		clear_pd_entry(pd, idx, scratch);
+		free = true;
+	}
+	spin_unlock(&pd->lock);
+
+	return free;
+}
+
+int i915_ppgtt_init_hw(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	gtt_write_workarounds(gt);
+
+	if (GRAPHICS_VER(i915) == 6)
+		gen6_ppgtt_enable(gt);
+	else if (GRAPHICS_VER(i915) == 7)
+		gen7_ppgtt_enable(gt);
+
+	return 0;
+}
+
+static struct i915_ppgtt *
+__ppgtt_create(struct intel_gt *gt, unsigned long lmem_pt_obj_flags)
+{
+	if (GRAPHICS_VER(gt->i915) < 8)
+		return gen6_ppgtt_create(gt);
+	else
+		return gen8_ppgtt_create(gt, lmem_pt_obj_flags);
+}
+
+struct i915_ppgtt *i915_ppgtt_create(struct intel_gt *gt,
+				     unsigned long lmem_pt_obj_flags)
+{
+	struct i915_ppgtt *ppgtt;
+
+	ppgtt = __ppgtt_create(gt, lmem_pt_obj_flags);
+	if (IS_ERR(ppgtt))
+		return ppgtt;
+
+	trace_i915_ppgtt_create(&ppgtt->vm);
+
+	return ppgtt;
+}
+
+void ppgtt_bind_vma(struct i915_address_space *vm,
+		    struct i915_vm_pt_stash *stash,
+		    struct i915_vma_resource *vma_res,
+		    enum i915_cache_level cache_level,
+		    u32 flags)
+{
+	u32 pte_flags;
+
+	if (!vma_res->allocated) {
+		vm->allocate_va_range(vm, stash, vma_res->start,
+				      vma_res->vma_size);
+		vma_res->allocated = true;
+	}
+
+	/* Applicable to VLV, and gen8+ */
+	pte_flags = 0;
+	if (vma_res->bi.readonly)
+		pte_flags |= PTE_READ_ONLY;
+	if (vma_res->bi.lmem)
+		pte_flags |= PTE_LM;
+
+	vm->insert_entries(vm, vma_res, cache_level, pte_flags);
+	wmb();
+}
+
+void ppgtt_unbind_vma(struct i915_address_space *vm,
+		      struct i915_vma_resource *vma_res)
+{
+	if (!vma_res->allocated)
+		return;
+
+	vm->clear_range(vm, vma_res->start, vma_res->vma_size);
+	if (vma_res->tlb)
+		vma_invalidate_tlb(vm, vma_res->tlb);
+}
+
+static unsigned long pd_count(u64 size, int shift)
+{
+	/* Beware later misalignment */
+	return (size + 2 * (BIT_ULL(shift) - 1)) >> shift;
+}
+
+int i915_vm_alloc_pt_stash(struct i915_address_space *vm,
+			   struct i915_vm_pt_stash *stash,
+			   u64 size)
+{
+	unsigned long count;
+	int shift, n, pt_sz;
+
+	shift = vm->pd_shift;
+	if (!shift)
+		return 0;
+
+	pt_sz = stash->pt_sz;
+	if (!pt_sz)
+		pt_sz = I915_GTT_PAGE_SIZE_4K;
+	else
+		GEM_BUG_ON(!IS_DGFX(vm->i915));
+
+	GEM_BUG_ON(!is_power_of_2(pt_sz));
+
+	count = pd_count(size, shift);
+	while (count--) {
+		struct i915_page_table *pt;
+
+		pt = alloc_pt(vm, pt_sz);
+		if (IS_ERR(pt)) {
+			i915_vm_free_pt_stash(vm, stash);
+			return PTR_ERR(pt);
+		}
+
+		pt->stash = stash->pt[0];
+		stash->pt[0] = pt;
+	}
+
+	for (n = 1; n < vm->top; n++) {
+		shift += ilog2(I915_PDES); /* Each PD holds 512 entries */
+		count = pd_count(size, shift);
+		while (count--) {
+			struct i915_page_directory *pd;
+
+			pd = alloc_pd(vm);
+			if (IS_ERR(pd)) {
+				i915_vm_free_pt_stash(vm, stash);
+				return PTR_ERR(pd);
+			}
+
+			pd->pt.stash = stash->pt[1];
+			stash->pt[1] = &pd->pt;
+		}
+	}
+
+	return 0;
+}
+
+int i915_vm_map_pt_stash(struct i915_address_space *vm,
+			 struct i915_vm_pt_stash *stash)
+{
+	struct i915_page_table *pt;
+	int n, err;
+
+	for (n = 0; n < ARRAY_SIZE(stash->pt); n++) {
+		for (pt = stash->pt[n]; pt; pt = pt->stash) {
+			err = map_pt_dma_locked(vm, pt->base);
+			if (err)
+				return err;
+		}
+	}
+
+	return 0;
+}
+
+void i915_vm_free_pt_stash(struct i915_address_space *vm,
+			   struct i915_vm_pt_stash *stash)
+{
+	struct i915_page_table *pt;
+	int n;
+
+	for (n = 0; n < ARRAY_SIZE(stash->pt); n++) {
+		while ((pt = stash->pt[n])) {
+			stash->pt[n] = pt->stash;
+			free_px(vm, pt, n);
+		}
+	}
+}
+
+void ppgtt_init(struct i915_ppgtt *ppgtt, struct intel_gt *gt,
+		unsigned long lmem_pt_obj_flags)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	ppgtt->vm.gt = gt;
+	ppgtt->vm.i915 = i915;
+	ppgtt->vm.dma = i915->drm.dev;
+	ppgtt->vm.total = BIT_ULL(RUNTIME_INFO(i915)->ppgtt_size);
+	ppgtt->vm.lmem_pt_obj_flags = lmem_pt_obj_flags;
+
+	dma_resv_init(&ppgtt->vm._resv);
+	i915_address_space_init(&ppgtt->vm, VM_CLASS_PPGTT);
+
+	ppgtt->vm.vma_ops.bind_vma    = ppgtt_bind_vma;
+	ppgtt->vm.vma_ops.unbind_vma  = ppgtt_unbind_vma;
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_rc6.c b/drivers/gpu/drm/i915/gt/intel_rc6.c
new file mode 100644
index 000000000..f8d0523f4
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_rc6.c
@@ -0,0 +1,821 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/pm_runtime.h>
+#include <linux/string_helpers.h>
+
+#include "gem/i915_gem_region.h"
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "i915_vgpu.h"
+#include "intel_engine_regs.h"
+#include "intel_gt.h"
+#include "intel_gt_pm.h"
+#include "intel_gt_regs.h"
+#include "intel_pcode.h"
+#include "intel_rc6.h"
+
+/**
+ * DOC: RC6
+ *
+ * RC6 is a special power stage which allows the GPU to enter an very
+ * low-voltage mode when idle, using down to 0V while at this stage.  This
+ * stage is entered automatically when the GPU is idle when RC6 support is
+ * enabled, and as soon as new workload arises GPU wakes up automatically as
+ * well.
+ *
+ * There are different RC6 modes available in Intel GPU, which differentiate
+ * among each other with the latency required to enter and leave RC6 and
+ * voltage consumed by the GPU in different states.
+ *
+ * The combination of the following flags define which states GPU is allowed
+ * to enter, while RC6 is the normal RC6 state, RC6p is the deep RC6, and
+ * RC6pp is deepest RC6. Their support by hardware varies according to the
+ * GPU, BIOS, chipset and platform. RC6 is usually the safest one and the one
+ * which brings the most power savings; deeper states save more power, but
+ * require higher latency to switch to and wake up.
+ */
+
+static struct intel_gt *rc6_to_gt(struct intel_rc6 *rc6)
+{
+	return container_of(rc6, struct intel_gt, rc6);
+}
+
+static struct intel_uncore *rc6_to_uncore(struct intel_rc6 *rc)
+{
+	return rc6_to_gt(rc)->uncore;
+}
+
+static struct drm_i915_private *rc6_to_i915(struct intel_rc6 *rc)
+{
+	return rc6_to_gt(rc)->i915;
+}
+
+static void set(struct intel_uncore *uncore, i915_reg_t reg, u32 val)
+{
+	intel_uncore_write_fw(uncore, reg, val);
+}
+
+static void gen11_rc6_enable(struct intel_rc6 *rc6)
+{
+	struct intel_gt *gt = rc6_to_gt(rc6);
+	struct intel_uncore *uncore = gt->uncore;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	u32 pg_enable;
+	int i;
+
+	/*
+	 * With GuCRC, these parameters are set by GuC
+	 */
+	if (!intel_uc_uses_guc_rc(&gt->uc)) {
+		/* 2b: Program RC6 thresholds.*/
+		set(uncore, GEN6_RC6_WAKE_RATE_LIMIT, 54 << 16 | 85);
+		set(uncore, GEN10_MEDIA_WAKE_RATE_LIMIT, 150);
+
+		set(uncore, GEN6_RC_EVALUATION_INTERVAL, 125000); /* 12500 * 1280ns */
+		set(uncore, GEN6_RC_IDLE_HYSTERSIS, 25); /* 25 * 1280ns */
+		for_each_engine(engine, rc6_to_gt(rc6), id)
+			set(uncore, RING_MAX_IDLE(engine->mmio_base), 10);
+
+		set(uncore, GUC_MAX_IDLE_COUNT, 0xA);
+
+		set(uncore, GEN6_RC_SLEEP, 0);
+
+		set(uncore, GEN6_RC6_THRESHOLD, 50000); /* 50/125ms per EI */
+	}
+
+	/*
+	 * 2c: Program Coarse Power Gating Policies.
+	 *
+	 * Bspec's guidance is to use 25us (really 25 * 1280ns) here. What we
+	 * use instead is a more conservative estimate for the maximum time
+	 * it takes us to service a CS interrupt and submit a new ELSP - that
+	 * is the time which the GPU is idle waiting for the CPU to select the
+	 * next request to execute. If the idle hysteresis is less than that
+	 * interrupt service latency, the hardware will automatically gate
+	 * the power well and we will then incur the wake up cost on top of
+	 * the service latency. A similar guide from plane_state is that we
+	 * do not want the enable hysteresis to less than the wakeup latency.
+	 *
+	 * igt/gem_exec_nop/sequential provides a rough estimate for the
+	 * service latency, and puts it under 10us for Icelake, similar to
+	 * Broadwell+, To be conservative, we want to factor in a context
+	 * switch on top (due to ksoftirqd).
+	 */
+	set(uncore, GEN9_MEDIA_PG_IDLE_HYSTERESIS, 60);
+	set(uncore, GEN9_RENDER_PG_IDLE_HYSTERESIS, 60);
+
+	/* 3a: Enable RC6
+	 *
+	 * With GuCRC, we do not enable bit 31 of RC_CTL,
+	 * thus allowing GuC to control RC6 entry/exit fully instead.
+	 * We will not set the HW ENABLE and EI bits
+	 */
+	if (!intel_guc_rc_enable(&gt->uc.guc))
+		rc6->ctl_enable = GEN6_RC_CTL_RC6_ENABLE;
+	else
+		rc6->ctl_enable =
+			GEN6_RC_CTL_HW_ENABLE |
+			GEN6_RC_CTL_RC6_ENABLE |
+			GEN6_RC_CTL_EI_MODE(1);
+
+	/* Wa_16011777198 - Render powergating must remain disabled */
+	if (IS_DG2_GRAPHICS_STEP(gt->i915, G10, STEP_A0, STEP_C0) ||
+	    IS_DG2_GRAPHICS_STEP(gt->i915, G11, STEP_A0, STEP_B0))
+		pg_enable =
+			GEN9_MEDIA_PG_ENABLE |
+			GEN11_MEDIA_SAMPLER_PG_ENABLE;
+	else
+		pg_enable =
+			GEN9_RENDER_PG_ENABLE |
+			GEN9_MEDIA_PG_ENABLE |
+			GEN11_MEDIA_SAMPLER_PG_ENABLE;
+
+	if (GRAPHICS_VER(gt->i915) >= 12) {
+		for (i = 0; i < I915_MAX_VCS; i++)
+			if (HAS_ENGINE(gt, _VCS(i)))
+				pg_enable |= (VDN_HCP_POWERGATE_ENABLE(i) |
+					      VDN_MFX_POWERGATE_ENABLE(i));
+	}
+
+	set(uncore, GEN9_PG_ENABLE, pg_enable);
+}
+
+static void gen9_rc6_enable(struct intel_rc6 *rc6)
+{
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/* 2b: Program RC6 thresholds.*/
+	if (GRAPHICS_VER(rc6_to_i915(rc6)) >= 11) {
+		set(uncore, GEN6_RC6_WAKE_RATE_LIMIT, 54 << 16 | 85);
+		set(uncore, GEN10_MEDIA_WAKE_RATE_LIMIT, 150);
+	} else if (IS_SKYLAKE(rc6_to_i915(rc6))) {
+		/*
+		 * WaRsDoubleRc6WrlWithCoarsePowerGating:skl Doubling WRL only
+		 * when CPG is enabled
+		 */
+		set(uncore, GEN6_RC6_WAKE_RATE_LIMIT, 108 << 16);
+	} else {
+		set(uncore, GEN6_RC6_WAKE_RATE_LIMIT, 54 << 16);
+	}
+
+	set(uncore, GEN6_RC_EVALUATION_INTERVAL, 125000); /* 12500 * 1280ns */
+	set(uncore, GEN6_RC_IDLE_HYSTERSIS, 25); /* 25 * 1280ns */
+	for_each_engine(engine, rc6_to_gt(rc6), id)
+		set(uncore, RING_MAX_IDLE(engine->mmio_base), 10);
+
+	set(uncore, GUC_MAX_IDLE_COUNT, 0xA);
+
+	set(uncore, GEN6_RC_SLEEP, 0);
+
+	/*
+	 * 2c: Program Coarse Power Gating Policies.
+	 *
+	 * Bspec's guidance is to use 25us (really 25 * 1280ns) here. What we
+	 * use instead is a more conservative estimate for the maximum time
+	 * it takes us to service a CS interrupt and submit a new ELSP - that
+	 * is the time which the GPU is idle waiting for the CPU to select the
+	 * next request to execute. If the idle hysteresis is less than that
+	 * interrupt service latency, the hardware will automatically gate
+	 * the power well and we will then incur the wake up cost on top of
+	 * the service latency. A similar guide from plane_state is that we
+	 * do not want the enable hysteresis to less than the wakeup latency.
+	 *
+	 * igt/gem_exec_nop/sequential provides a rough estimate for the
+	 * service latency, and puts it around 10us for Broadwell (and other
+	 * big core) and around 40us for Broxton (and other low power cores).
+	 * [Note that for legacy ringbuffer submission, this is less than 1us!]
+	 * However, the wakeup latency on Broxton is closer to 100us. To be
+	 * conservative, we have to factor in a context switch on top (due
+	 * to ksoftirqd).
+	 */
+	set(uncore, GEN9_MEDIA_PG_IDLE_HYSTERESIS, 250);
+	set(uncore, GEN9_RENDER_PG_IDLE_HYSTERESIS, 250);
+
+	/* 3a: Enable RC6 */
+	set(uncore, GEN6_RC6_THRESHOLD, 37500); /* 37.5/125ms per EI */
+
+	rc6->ctl_enable =
+		GEN6_RC_CTL_HW_ENABLE |
+		GEN6_RC_CTL_RC6_ENABLE |
+		GEN6_RC_CTL_EI_MODE(1);
+
+	/*
+	 * WaRsDisableCoarsePowerGating:skl,cnl
+	 *   - Render/Media PG need to be disabled with RC6.
+	 */
+	if (!NEEDS_WaRsDisableCoarsePowerGating(rc6_to_i915(rc6)))
+		set(uncore, GEN9_PG_ENABLE,
+		    GEN9_RENDER_PG_ENABLE | GEN9_MEDIA_PG_ENABLE);
+}
+
+static void gen8_rc6_enable(struct intel_rc6 *rc6)
+{
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/* 2b: Program RC6 thresholds.*/
+	set(uncore, GEN6_RC6_WAKE_RATE_LIMIT, 40 << 16);
+	set(uncore, GEN6_RC_EVALUATION_INTERVAL, 125000); /* 12500 * 1280ns */
+	set(uncore, GEN6_RC_IDLE_HYSTERSIS, 25); /* 25 * 1280ns */
+	for_each_engine(engine, rc6_to_gt(rc6), id)
+		set(uncore, RING_MAX_IDLE(engine->mmio_base), 10);
+	set(uncore, GEN6_RC_SLEEP, 0);
+	set(uncore, GEN6_RC6_THRESHOLD, 625); /* 800us/1.28 for TO */
+
+	/* 3: Enable RC6 */
+	rc6->ctl_enable =
+	    GEN6_RC_CTL_HW_ENABLE |
+	    GEN7_RC_CTL_TO_MODE |
+	    GEN6_RC_CTL_RC6_ENABLE;
+}
+
+static void gen6_rc6_enable(struct intel_rc6 *rc6)
+{
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	struct drm_i915_private *i915 = rc6_to_i915(rc6);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	u32 rc6vids, rc6_mask;
+	int ret;
+
+	set(uncore, GEN6_RC1_WAKE_RATE_LIMIT, 1000 << 16);
+	set(uncore, GEN6_RC6_WAKE_RATE_LIMIT, 40 << 16 | 30);
+	set(uncore, GEN6_RC6pp_WAKE_RATE_LIMIT, 30);
+	set(uncore, GEN6_RC_EVALUATION_INTERVAL, 125000);
+	set(uncore, GEN6_RC_IDLE_HYSTERSIS, 25);
+
+	for_each_engine(engine, rc6_to_gt(rc6), id)
+		set(uncore, RING_MAX_IDLE(engine->mmio_base), 10);
+
+	set(uncore, GEN6_RC_SLEEP, 0);
+	set(uncore, GEN6_RC1e_THRESHOLD, 1000);
+	set(uncore, GEN6_RC6_THRESHOLD, 50000);
+	set(uncore, GEN6_RC6p_THRESHOLD, 150000);
+	set(uncore, GEN6_RC6pp_THRESHOLD, 64000); /* unused */
+
+	/* We don't use those on Haswell */
+	rc6_mask = GEN6_RC_CTL_RC6_ENABLE;
+	if (HAS_RC6p(i915))
+		rc6_mask |= GEN6_RC_CTL_RC6p_ENABLE;
+	if (HAS_RC6pp(i915))
+		rc6_mask |= GEN6_RC_CTL_RC6pp_ENABLE;
+	rc6->ctl_enable =
+	    rc6_mask |
+	    GEN6_RC_CTL_EI_MODE(1) |
+	    GEN6_RC_CTL_HW_ENABLE;
+
+	rc6vids = 0;
+	ret = snb_pcode_read(rc6_to_gt(rc6)->uncore, GEN6_PCODE_READ_RC6VIDS, &rc6vids, NULL);
+	if (GRAPHICS_VER(i915) == 6 && ret) {
+		drm_dbg(&i915->drm, "Couldn't check for BIOS workaround\n");
+	} else if (GRAPHICS_VER(i915) == 6 &&
+		   (GEN6_DECODE_RC6_VID(rc6vids & 0xff) < 450)) {
+		drm_dbg(&i915->drm,
+			"You should update your BIOS. Correcting minimum rc6 voltage (%dmV->%dmV)\n",
+			GEN6_DECODE_RC6_VID(rc6vids & 0xff), 450);
+		rc6vids &= 0xffff00;
+		rc6vids |= GEN6_ENCODE_RC6_VID(450);
+		ret = snb_pcode_write(rc6_to_gt(rc6)->uncore, GEN6_PCODE_WRITE_RC6VIDS, rc6vids);
+		if (ret)
+			drm_err(&i915->drm,
+				"Couldn't fix incorrect rc6 voltage\n");
+	}
+}
+
+/* Check that the pcbr address is not empty. */
+static int chv_rc6_init(struct intel_rc6 *rc6)
+{
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	struct drm_i915_private *i915 = rc6_to_i915(rc6);
+	resource_size_t pctx_paddr, paddr;
+	resource_size_t pctx_size = 32 * SZ_1K;
+	u32 pcbr;
+
+	pcbr = intel_uncore_read(uncore, VLV_PCBR);
+	if ((pcbr >> VLV_PCBR_ADDR_SHIFT) == 0) {
+		drm_dbg(&i915->drm, "BIOS didn't set up PCBR, fixing up\n");
+		paddr = i915->dsm.end + 1 - pctx_size;
+		GEM_BUG_ON(paddr > U32_MAX);
+
+		pctx_paddr = (paddr & ~4095);
+		intel_uncore_write(uncore, VLV_PCBR, pctx_paddr);
+	}
+
+	return 0;
+}
+
+static int vlv_rc6_init(struct intel_rc6 *rc6)
+{
+	struct drm_i915_private *i915 = rc6_to_i915(rc6);
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	struct drm_i915_gem_object *pctx;
+	resource_size_t pctx_paddr;
+	resource_size_t pctx_size = 24 * SZ_1K;
+	u32 pcbr;
+
+	pcbr = intel_uncore_read(uncore, VLV_PCBR);
+	if (pcbr) {
+		/* BIOS set it up already, grab the pre-alloc'd space */
+		resource_size_t pcbr_offset;
+
+		pcbr_offset = (pcbr & ~4095) - i915->dsm.start;
+		pctx = i915_gem_object_create_region_at(i915->mm.stolen_region,
+							pcbr_offset,
+							pctx_size,
+							0);
+		if (IS_ERR(pctx))
+			return PTR_ERR(pctx);
+
+		goto out;
+	}
+
+	drm_dbg(&i915->drm, "BIOS didn't set up PCBR, fixing up\n");
+
+	/*
+	 * From the Gunit register HAS:
+	 * The Gfx driver is expected to program this register and ensure
+	 * proper allocation within Gfx stolen memory.  For example, this
+	 * register should be programmed such than the PCBR range does not
+	 * overlap with other ranges, such as the frame buffer, protected
+	 * memory, or any other relevant ranges.
+	 */
+	pctx = i915_gem_object_create_stolen(i915, pctx_size);
+	if (IS_ERR(pctx)) {
+		drm_dbg(&i915->drm,
+			"not enough stolen space for PCTX, disabling\n");
+		return PTR_ERR(pctx);
+	}
+
+	GEM_BUG_ON(range_overflows_end_t(u64,
+					 i915->dsm.start,
+					 pctx->stolen->start,
+					 U32_MAX));
+	pctx_paddr = i915->dsm.start + pctx->stolen->start;
+	intel_uncore_write(uncore, VLV_PCBR, pctx_paddr);
+
+out:
+	rc6->pctx = pctx;
+	return 0;
+}
+
+static void chv_rc6_enable(struct intel_rc6 *rc6)
+{
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/* 2a: Program RC6 thresholds.*/
+	set(uncore, GEN6_RC6_WAKE_RATE_LIMIT, 40 << 16);
+	set(uncore, GEN6_RC_EVALUATION_INTERVAL, 125000); /* 12500 * 1280ns */
+	set(uncore, GEN6_RC_IDLE_HYSTERSIS, 25); /* 25 * 1280ns */
+
+	for_each_engine(engine, rc6_to_gt(rc6), id)
+		set(uncore, RING_MAX_IDLE(engine->mmio_base), 10);
+	set(uncore, GEN6_RC_SLEEP, 0);
+
+	/* TO threshold set to 500 us (0x186 * 1.28 us) */
+	set(uncore, GEN6_RC6_THRESHOLD, 0x186);
+
+	/* Allows RC6 residency counter to work */
+	set(uncore, VLV_COUNTER_CONTROL,
+	    _MASKED_BIT_ENABLE(VLV_COUNT_RANGE_HIGH |
+			       VLV_MEDIA_RC6_COUNT_EN |
+			       VLV_RENDER_RC6_COUNT_EN));
+
+	/* 3: Enable RC6 */
+	rc6->ctl_enable = GEN7_RC_CTL_TO_MODE;
+}
+
+static void vlv_rc6_enable(struct intel_rc6 *rc6)
+{
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	set(uncore, GEN6_RC6_WAKE_RATE_LIMIT, 0x00280000);
+	set(uncore, GEN6_RC_EVALUATION_INTERVAL, 125000);
+	set(uncore, GEN6_RC_IDLE_HYSTERSIS, 25);
+
+	for_each_engine(engine, rc6_to_gt(rc6), id)
+		set(uncore, RING_MAX_IDLE(engine->mmio_base), 10);
+
+	set(uncore, GEN6_RC6_THRESHOLD, 0x557);
+
+	/* Allows RC6 residency counter to work */
+	set(uncore, VLV_COUNTER_CONTROL,
+	    _MASKED_BIT_ENABLE(VLV_COUNT_RANGE_HIGH |
+			       VLV_MEDIA_RC0_COUNT_EN |
+			       VLV_RENDER_RC0_COUNT_EN |
+			       VLV_MEDIA_RC6_COUNT_EN |
+			       VLV_RENDER_RC6_COUNT_EN));
+
+	rc6->ctl_enable =
+	    GEN7_RC_CTL_TO_MODE | VLV_RC_CTL_CTX_RST_PARALLEL;
+}
+
+static bool bxt_check_bios_rc6_setup(struct intel_rc6 *rc6)
+{
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	struct drm_i915_private *i915 = rc6_to_i915(rc6);
+	u32 rc6_ctx_base, rc_ctl, rc_sw_target;
+	bool enable_rc6 = true;
+
+	rc_ctl = intel_uncore_read(uncore, GEN6_RC_CONTROL);
+	rc_sw_target = intel_uncore_read(uncore, GEN6_RC_STATE);
+	rc_sw_target &= RC_SW_TARGET_STATE_MASK;
+	rc_sw_target >>= RC_SW_TARGET_STATE_SHIFT;
+	drm_dbg(&i915->drm, "BIOS enabled RC states: "
+			 "HW_CTRL %s HW_RC6 %s SW_TARGET_STATE %x\n",
+			 str_on_off(rc_ctl & GEN6_RC_CTL_HW_ENABLE),
+			 str_on_off(rc_ctl & GEN6_RC_CTL_RC6_ENABLE),
+			 rc_sw_target);
+
+	if (!(intel_uncore_read(uncore, RC6_LOCATION) & RC6_CTX_IN_DRAM)) {
+		drm_dbg(&i915->drm, "RC6 Base location not set properly.\n");
+		enable_rc6 = false;
+	}
+
+	/*
+	 * The exact context size is not known for BXT, so assume a page size
+	 * for this check.
+	 */
+	rc6_ctx_base =
+		intel_uncore_read(uncore, RC6_CTX_BASE) & RC6_CTX_BASE_MASK;
+	if (!(rc6_ctx_base >= i915->dsm_reserved.start &&
+	      rc6_ctx_base + PAGE_SIZE < i915->dsm_reserved.end)) {
+		drm_dbg(&i915->drm, "RC6 Base address not as expected.\n");
+		enable_rc6 = false;
+	}
+
+	if (!((intel_uncore_read(uncore, PWRCTX_MAXCNT(RENDER_RING_BASE)) & IDLE_TIME_MASK) > 1 &&
+	      (intel_uncore_read(uncore, PWRCTX_MAXCNT(GEN6_BSD_RING_BASE)) & IDLE_TIME_MASK) > 1 &&
+	      (intel_uncore_read(uncore, PWRCTX_MAXCNT(BLT_RING_BASE)) & IDLE_TIME_MASK) > 1 &&
+	      (intel_uncore_read(uncore, PWRCTX_MAXCNT(VEBOX_RING_BASE)) & IDLE_TIME_MASK) > 1)) {
+		drm_dbg(&i915->drm,
+			"Engine Idle wait time not set properly.\n");
+		enable_rc6 = false;
+	}
+
+	if (!intel_uncore_read(uncore, GEN8_PUSHBUS_CONTROL) ||
+	    !intel_uncore_read(uncore, GEN8_PUSHBUS_ENABLE) ||
+	    !intel_uncore_read(uncore, GEN8_PUSHBUS_SHIFT)) {
+		drm_dbg(&i915->drm, "Pushbus not setup properly.\n");
+		enable_rc6 = false;
+	}
+
+	if (!intel_uncore_read(uncore, GEN6_GFXPAUSE)) {
+		drm_dbg(&i915->drm, "GFX pause not setup properly.\n");
+		enable_rc6 = false;
+	}
+
+	if (!intel_uncore_read(uncore, GEN8_MISC_CTRL0)) {
+		drm_dbg(&i915->drm, "GPM control not setup properly.\n");
+		enable_rc6 = false;
+	}
+
+	return enable_rc6;
+}
+
+static bool rc6_supported(struct intel_rc6 *rc6)
+{
+	struct drm_i915_private *i915 = rc6_to_i915(rc6);
+
+	if (!HAS_RC6(i915))
+		return false;
+
+	if (intel_vgpu_active(i915))
+		return false;
+
+	if (is_mock_gt(rc6_to_gt(rc6)))
+		return false;
+
+	if (IS_GEN9_LP(i915) && !bxt_check_bios_rc6_setup(rc6)) {
+		drm_notice(&i915->drm,
+			   "RC6 and powersaving disabled by BIOS\n");
+		return false;
+	}
+
+	return true;
+}
+
+static void rpm_get(struct intel_rc6 *rc6)
+{
+	GEM_BUG_ON(rc6->wakeref);
+	pm_runtime_get_sync(rc6_to_i915(rc6)->drm.dev);
+	rc6->wakeref = true;
+}
+
+static void rpm_put(struct intel_rc6 *rc6)
+{
+	GEM_BUG_ON(!rc6->wakeref);
+	pm_runtime_put(rc6_to_i915(rc6)->drm.dev);
+	rc6->wakeref = false;
+}
+
+static bool pctx_corrupted(struct intel_rc6 *rc6)
+{
+	struct drm_i915_private *i915 = rc6_to_i915(rc6);
+
+	if (!NEEDS_RC6_CTX_CORRUPTION_WA(i915))
+		return false;
+
+	if (intel_uncore_read(rc6_to_uncore(rc6), GEN8_RC6_CTX_INFO))
+		return false;
+
+	drm_notice(&i915->drm,
+		   "RC6 context corruption, disabling runtime power management\n");
+	return true;
+}
+
+static void __intel_rc6_disable(struct intel_rc6 *rc6)
+{
+	struct drm_i915_private *i915 = rc6_to_i915(rc6);
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	struct intel_gt *gt = rc6_to_gt(rc6);
+
+	/* Take control of RC6 back from GuC */
+	intel_guc_rc_disable(&gt->uc.guc);
+
+	intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+	if (GRAPHICS_VER(i915) >= 9)
+		set(uncore, GEN9_PG_ENABLE, 0);
+	set(uncore, GEN6_RC_CONTROL, 0);
+	set(uncore, GEN6_RC_STATE, 0);
+	intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
+}
+
+void intel_rc6_init(struct intel_rc6 *rc6)
+{
+	struct drm_i915_private *i915 = rc6_to_i915(rc6);
+	int err;
+
+	/* Disable runtime-pm until we can save the GPU state with rc6 pctx */
+	rpm_get(rc6);
+
+	if (!rc6_supported(rc6))
+		return;
+
+	if (IS_CHERRYVIEW(i915))
+		err = chv_rc6_init(rc6);
+	else if (IS_VALLEYVIEW(i915))
+		err = vlv_rc6_init(rc6);
+	else
+		err = 0;
+
+	/* Sanitize rc6, ensure it is disabled before we are ready. */
+	__intel_rc6_disable(rc6);
+
+	rc6->supported = err == 0;
+}
+
+void intel_rc6_sanitize(struct intel_rc6 *rc6)
+{
+	memset(rc6->prev_hw_residency, 0, sizeof(rc6->prev_hw_residency));
+
+	if (rc6->enabled) { /* unbalanced suspend/resume */
+		rpm_get(rc6);
+		rc6->enabled = false;
+	}
+
+	if (rc6->supported)
+		__intel_rc6_disable(rc6);
+}
+
+void intel_rc6_enable(struct intel_rc6 *rc6)
+{
+	struct drm_i915_private *i915 = rc6_to_i915(rc6);
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+
+	if (!rc6->supported)
+		return;
+
+	GEM_BUG_ON(rc6->enabled);
+
+	intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+
+	if (IS_CHERRYVIEW(i915))
+		chv_rc6_enable(rc6);
+	else if (IS_VALLEYVIEW(i915))
+		vlv_rc6_enable(rc6);
+	else if (GRAPHICS_VER(i915) >= 11)
+		gen11_rc6_enable(rc6);
+	else if (GRAPHICS_VER(i915) >= 9)
+		gen9_rc6_enable(rc6);
+	else if (IS_BROADWELL(i915))
+		gen8_rc6_enable(rc6);
+	else if (GRAPHICS_VER(i915) >= 6)
+		gen6_rc6_enable(rc6);
+
+	rc6->manual = rc6->ctl_enable & GEN6_RC_CTL_RC6_ENABLE;
+	if (NEEDS_RC6_CTX_CORRUPTION_WA(i915))
+		rc6->ctl_enable = 0;
+
+	intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
+
+	if (unlikely(pctx_corrupted(rc6)))
+		return;
+
+	/* rc6 is ready, runtime-pm is go! */
+	rpm_put(rc6);
+	rc6->enabled = true;
+}
+
+void intel_rc6_unpark(struct intel_rc6 *rc6)
+{
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+
+	if (!rc6->enabled)
+		return;
+
+	/* Restore HW timers for automatic RC6 entry while busy */
+	set(uncore, GEN6_RC_CONTROL, rc6->ctl_enable);
+}
+
+void intel_rc6_park(struct intel_rc6 *rc6)
+{
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	unsigned int target;
+
+	if (!rc6->enabled)
+		return;
+
+	if (unlikely(pctx_corrupted(rc6))) {
+		intel_rc6_disable(rc6);
+		return;
+	}
+
+	if (!rc6->manual)
+		return;
+
+	/* Turn off the HW timers and go directly to rc6 */
+	set(uncore, GEN6_RC_CONTROL, GEN6_RC_CTL_RC6_ENABLE);
+
+	if (HAS_RC6pp(rc6_to_i915(rc6)))
+		target = 0x6; /* deepest rc6 */
+	else if (HAS_RC6p(rc6_to_i915(rc6)))
+		target = 0x5; /* deep rc6 */
+	else
+		target = 0x4; /* normal rc6 */
+	set(uncore, GEN6_RC_STATE, target << RC_SW_TARGET_STATE_SHIFT);
+}
+
+void intel_rc6_disable(struct intel_rc6 *rc6)
+{
+	if (!rc6->enabled)
+		return;
+
+	rpm_get(rc6);
+	rc6->enabled = false;
+
+	__intel_rc6_disable(rc6);
+}
+
+void intel_rc6_fini(struct intel_rc6 *rc6)
+{
+	struct drm_i915_gem_object *pctx;
+
+	intel_rc6_disable(rc6);
+
+	pctx = fetch_and_zero(&rc6->pctx);
+	if (pctx)
+		i915_gem_object_put(pctx);
+
+	if (rc6->wakeref)
+		rpm_put(rc6);
+}
+
+static u64 vlv_residency_raw(struct intel_uncore *uncore, const i915_reg_t reg)
+{
+	u32 lower, upper, tmp;
+	int loop = 2;
+
+	/*
+	 * The register accessed do not need forcewake. We borrow
+	 * uncore lock to prevent concurrent access to range reg.
+	 */
+	lockdep_assert_held(&uncore->lock);
+
+	/*
+	 * vlv and chv residency counters are 40 bits in width.
+	 * With a control bit, we can choose between upper or lower
+	 * 32bit window into this counter.
+	 *
+	 * Although we always use the counter in high-range mode elsewhere,
+	 * userspace may attempt to read the value before rc6 is initialised,
+	 * before we have set the default VLV_COUNTER_CONTROL value. So always
+	 * set the high bit to be safe.
+	 */
+	set(uncore, VLV_COUNTER_CONTROL,
+	    _MASKED_BIT_ENABLE(VLV_COUNT_RANGE_HIGH));
+	upper = intel_uncore_read_fw(uncore, reg);
+	do {
+		tmp = upper;
+
+		set(uncore, VLV_COUNTER_CONTROL,
+		    _MASKED_BIT_DISABLE(VLV_COUNT_RANGE_HIGH));
+		lower = intel_uncore_read_fw(uncore, reg);
+
+		set(uncore, VLV_COUNTER_CONTROL,
+		    _MASKED_BIT_ENABLE(VLV_COUNT_RANGE_HIGH));
+		upper = intel_uncore_read_fw(uncore, reg);
+	} while (upper != tmp && --loop);
+
+	/*
+	 * Everywhere else we always use VLV_COUNTER_CONTROL with the
+	 * VLV_COUNT_RANGE_HIGH bit set - so it is safe to leave it set
+	 * now.
+	 */
+
+	return lower | (u64)upper << 8;
+}
+
+u64 intel_rc6_residency_ns(struct intel_rc6 *rc6, const i915_reg_t reg)
+{
+	struct drm_i915_private *i915 = rc6_to_i915(rc6);
+	struct intel_uncore *uncore = rc6_to_uncore(rc6);
+	u64 time_hw, prev_hw, overflow_hw;
+	unsigned int fw_domains;
+	unsigned long flags;
+	unsigned int i;
+	u32 mul, div;
+
+	if (!rc6->supported)
+		return 0;
+
+	/*
+	 * Store previous hw counter values for counter wrap-around handling.
+	 *
+	 * There are only four interesting registers and they live next to each
+	 * other so we can use the relative address, compared to the smallest
+	 * one as the index into driver storage.
+	 */
+	i = (i915_mmio_reg_offset(reg) -
+	     i915_mmio_reg_offset(GEN6_GT_GFX_RC6_LOCKED)) / sizeof(u32);
+	if (drm_WARN_ON_ONCE(&i915->drm, i >= ARRAY_SIZE(rc6->cur_residency)))
+		return 0;
+
+	fw_domains = intel_uncore_forcewake_for_reg(uncore, reg, FW_REG_READ);
+
+	spin_lock_irqsave(&uncore->lock, flags);
+	intel_uncore_forcewake_get__locked(uncore, fw_domains);
+
+	/* On VLV and CHV, residency time is in CZ units rather than 1.28us */
+	if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) {
+		mul = 1000000;
+		div = i915->czclk_freq;
+		overflow_hw = BIT_ULL(40);
+		time_hw = vlv_residency_raw(uncore, reg);
+	} else {
+		/* 833.33ns units on Gen9LP, 1.28us elsewhere. */
+		if (IS_GEN9_LP(i915)) {
+			mul = 10000;
+			div = 12;
+		} else {
+			mul = 1280;
+			div = 1;
+		}
+
+		overflow_hw = BIT_ULL(32);
+		time_hw = intel_uncore_read_fw(uncore, reg);
+	}
+
+	/*
+	 * Counter wrap handling.
+	 *
+	 * But relying on a sufficient frequency of queries otherwise counters
+	 * can still wrap.
+	 */
+	prev_hw = rc6->prev_hw_residency[i];
+	rc6->prev_hw_residency[i] = time_hw;
+
+	/* RC6 delta from last sample. */
+	if (time_hw >= prev_hw)
+		time_hw -= prev_hw;
+	else
+		time_hw += overflow_hw - prev_hw;
+
+	/* Add delta to RC6 extended raw driver copy. */
+	time_hw += rc6->cur_residency[i];
+	rc6->cur_residency[i] = time_hw;
+
+	intel_uncore_forcewake_put__locked(uncore, fw_domains);
+	spin_unlock_irqrestore(&uncore->lock, flags);
+
+	return mul_u64_u32_div(time_hw, mul, div);
+}
+
+u64 intel_rc6_residency_us(struct intel_rc6 *rc6, i915_reg_t reg)
+{
+	return DIV_ROUND_UP_ULL(intel_rc6_residency_ns(rc6, reg), 1000);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_rc6.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_rc6.h b/drivers/gpu/drm/i915/gt/intel_rc6.h
new file mode 100644
index 000000000..b6fea71af
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_rc6.h
@@ -0,0 +1,27 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_RC6_H
+#define INTEL_RC6_H
+
+#include "i915_reg_defs.h"
+
+struct intel_engine_cs;
+struct intel_rc6;
+
+void intel_rc6_init(struct intel_rc6 *rc6);
+void intel_rc6_fini(struct intel_rc6 *rc6);
+
+void intel_rc6_unpark(struct intel_rc6 *rc6);
+void intel_rc6_park(struct intel_rc6 *rc6);
+
+void intel_rc6_sanitize(struct intel_rc6 *rc6);
+void intel_rc6_enable(struct intel_rc6 *rc6);
+void intel_rc6_disable(struct intel_rc6 *rc6);
+
+u64 intel_rc6_residency_ns(struct intel_rc6 *rc6, i915_reg_t reg);
+u64 intel_rc6_residency_us(struct intel_rc6 *rc6, i915_reg_t reg);
+
+#endif /* INTEL_RC6_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_rc6_types.h b/drivers/gpu/drm/i915/gt/intel_rc6_types.h
new file mode 100644
index 000000000..e747492b2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_rc6_types.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_RC6_TYPES_H
+#define INTEL_RC6_TYPES_H
+
+#include <linux/spinlock.h>
+#include <linux/types.h>
+
+#include "intel_engine_types.h"
+
+struct drm_i915_gem_object;
+
+struct intel_rc6 {
+	u64 prev_hw_residency[4];
+	u64 cur_residency[4];
+
+	u32 ctl_enable;
+
+	struct drm_i915_gem_object *pctx;
+
+	bool supported : 1;
+	bool enabled : 1;
+	bool manual : 1;
+	bool wakeref : 1;
+};
+
+#endif /* INTEL_RC6_TYPES_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_region_lmem.c b/drivers/gpu/drm/i915/gt/intel_region_lmem.c
new file mode 100644
index 000000000..f3ad93db0
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_region_lmem.c
@@ -0,0 +1,287 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_pci.h"
+#include "i915_reg.h"
+#include "intel_memory_region.h"
+#include "intel_pci_config.h"
+#include "intel_region_lmem.h"
+#include "intel_region_ttm.h"
+#include "gem/i915_gem_lmem.h"
+#include "gem/i915_gem_region.h"
+#include "gem/i915_gem_ttm.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_mcr.h"
+#include "gt/intel_gt_regs.h"
+
+#ifdef CONFIG_64BIT
+static void _release_bars(struct pci_dev *pdev)
+{
+	int resno;
+
+	for (resno = PCI_STD_RESOURCES; resno < PCI_STD_RESOURCE_END; resno++) {
+		if (pci_resource_len(pdev, resno))
+			pci_release_resource(pdev, resno);
+	}
+}
+
+static void
+_resize_bar(struct drm_i915_private *i915, int resno, resource_size_t size)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	int bar_size = pci_rebar_bytes_to_size(size);
+	int ret;
+
+	_release_bars(pdev);
+
+	ret = pci_resize_resource(pdev, resno, bar_size);
+	if (ret) {
+		drm_info(&i915->drm, "Failed to resize BAR%d to %dM (%pe)\n",
+			 resno, 1 << bar_size, ERR_PTR(ret));
+		return;
+	}
+
+	drm_info(&i915->drm, "BAR%d resized to %dM\n", resno, 1 << bar_size);
+}
+
+static void i915_resize_lmem_bar(struct drm_i915_private *i915, resource_size_t lmem_size)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	struct pci_bus *root = pdev->bus;
+	struct resource *root_res;
+	resource_size_t rebar_size;
+	resource_size_t current_size;
+	u32 pci_cmd;
+	int i;
+
+	current_size = roundup_pow_of_two(pci_resource_len(pdev, GEN12_LMEM_BAR));
+
+	if (i915->params.lmem_bar_size) {
+		u32 bar_sizes;
+
+		rebar_size = i915->params.lmem_bar_size *
+			(resource_size_t)SZ_1M;
+		bar_sizes = pci_rebar_get_possible_sizes(pdev, GEN12_LMEM_BAR);
+
+		if (rebar_size == current_size)
+			return;
+
+		if (!(bar_sizes & BIT(pci_rebar_bytes_to_size(rebar_size))) ||
+		    rebar_size >= roundup_pow_of_two(lmem_size)) {
+			rebar_size = lmem_size;
+
+			drm_info(&i915->drm,
+				 "Given bar size is not within supported size, setting it to default: %llu\n",
+				 (u64)lmem_size >> 20);
+		}
+	} else {
+		rebar_size = current_size;
+
+		if (rebar_size != roundup_pow_of_two(lmem_size))
+			rebar_size = lmem_size;
+		else
+			return;
+	}
+
+	/* Find out if root bus contains 64bit memory addressing */
+	while (root->parent)
+		root = root->parent;
+
+	pci_bus_for_each_resource(root, root_res, i) {
+		if (root_res && root_res->flags & (IORESOURCE_MEM | IORESOURCE_MEM_64) &&
+		    root_res->start > 0x100000000ull)
+			break;
+	}
+
+	/* pci_resize_resource will fail anyways */
+	if (!root_res) {
+		drm_info(&i915->drm, "Can't resize LMEM BAR - platform support is missing\n");
+		return;
+	}
+
+	/* First disable PCI memory decoding references */
+	pci_read_config_dword(pdev, PCI_COMMAND, &pci_cmd);
+	pci_write_config_dword(pdev, PCI_COMMAND,
+			       pci_cmd & ~PCI_COMMAND_MEMORY);
+
+	_resize_bar(i915, GEN12_LMEM_BAR, rebar_size);
+
+	pci_assign_unassigned_bus_resources(pdev->bus);
+	pci_write_config_dword(pdev, PCI_COMMAND, pci_cmd);
+}
+#else
+static void i915_resize_lmem_bar(struct drm_i915_private *i915, resource_size_t lmem_size) {}
+#endif
+
+static int
+region_lmem_release(struct intel_memory_region *mem)
+{
+	int ret;
+
+	ret = intel_region_ttm_fini(mem);
+	io_mapping_fini(&mem->iomap);
+
+	return ret;
+}
+
+static int
+region_lmem_init(struct intel_memory_region *mem)
+{
+	int ret;
+
+	if (!io_mapping_init_wc(&mem->iomap,
+				mem->io_start,
+				mem->io_size))
+		return -EIO;
+
+	ret = intel_region_ttm_init(mem);
+	if (ret)
+		goto out_no_buddy;
+
+	return 0;
+
+out_no_buddy:
+	io_mapping_fini(&mem->iomap);
+
+	return ret;
+}
+
+static const struct intel_memory_region_ops intel_region_lmem_ops = {
+	.init = region_lmem_init,
+	.release = region_lmem_release,
+	.init_object = __i915_gem_ttm_object_init,
+};
+
+static bool get_legacy_lowmem_region(struct intel_uncore *uncore,
+				     u64 *start, u32 *size)
+{
+	if (!IS_DG1_GRAPHICS_STEP(uncore->i915, STEP_A0, STEP_C0))
+		return false;
+
+	*start = 0;
+	*size = SZ_1M;
+
+	drm_dbg(&uncore->i915->drm, "LMEM: reserved legacy low-memory [0x%llx-0x%llx]\n",
+		*start, *start + *size);
+
+	return true;
+}
+
+static int reserve_lowmem_region(struct intel_uncore *uncore,
+				 struct intel_memory_region *mem)
+{
+	u64 reserve_start;
+	u32 reserve_size;
+	int ret;
+
+	if (!get_legacy_lowmem_region(uncore, &reserve_start, &reserve_size))
+		return 0;
+
+	ret = intel_memory_region_reserve(mem, reserve_start, reserve_size);
+	if (ret)
+		drm_err(&uncore->i915->drm, "LMEM: reserving low memory region failed\n");
+
+	return ret;
+}
+
+static struct intel_memory_region *setup_lmem(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	struct intel_memory_region *mem;
+	resource_size_t min_page_size;
+	resource_size_t io_start;
+	resource_size_t io_size;
+	resource_size_t lmem_size;
+	int err;
+
+	if (!IS_DGFX(i915))
+		return ERR_PTR(-ENODEV);
+
+	if (!i915_pci_resource_valid(pdev, GEN12_LMEM_BAR))
+		return ERR_PTR(-ENXIO);
+
+	if (HAS_FLAT_CCS(i915)) {
+		resource_size_t lmem_range;
+		u64 tile_stolen, flat_ccs_base;
+
+		lmem_range = intel_gt_mcr_read_any(&i915->gt0, XEHP_TILE0_ADDR_RANGE) & 0xFFFF;
+		lmem_size = lmem_range >> XEHP_TILE_LMEM_RANGE_SHIFT;
+		lmem_size *= SZ_1G;
+
+		flat_ccs_base = intel_gt_mcr_read_any(gt, XEHP_FLAT_CCS_BASE_ADDR);
+		flat_ccs_base = (flat_ccs_base >> XEHP_CCS_BASE_SHIFT) * SZ_64K;
+
+		if (GEM_WARN_ON(lmem_size < flat_ccs_base))
+			return ERR_PTR(-EIO);
+
+		tile_stolen = lmem_size - flat_ccs_base;
+
+		/* If the FLAT_CCS_BASE_ADDR register is not populated, flag an error */
+		if (tile_stolen == lmem_size)
+			drm_err(&i915->drm,
+				"CCS_BASE_ADDR register did not have expected value\n");
+
+		lmem_size -= tile_stolen;
+	} else {
+		/* Stolen starts from GSMBASE without CCS */
+		lmem_size = intel_uncore_read64(&i915->uncore, GEN12_GSMBASE);
+	}
+
+	i915_resize_lmem_bar(i915, lmem_size);
+
+	if (i915->params.lmem_size > 0) {
+		lmem_size = min_t(resource_size_t, lmem_size,
+				  mul_u32_u32(i915->params.lmem_size, SZ_1M));
+	}
+
+	io_start = pci_resource_start(pdev, GEN12_LMEM_BAR);
+	io_size = min(pci_resource_len(pdev, GEN12_LMEM_BAR), lmem_size);
+	if (!io_size)
+		return ERR_PTR(-EIO);
+
+	min_page_size = HAS_64K_PAGES(i915) ? I915_GTT_PAGE_SIZE_64K :
+						I915_GTT_PAGE_SIZE_4K;
+	mem = intel_memory_region_create(i915,
+					 0,
+					 lmem_size,
+					 min_page_size,
+					 io_start,
+					 io_size,
+					 INTEL_MEMORY_LOCAL,
+					 0,
+					 &intel_region_lmem_ops);
+	if (IS_ERR(mem))
+		return mem;
+
+	err = reserve_lowmem_region(uncore, mem);
+	if (err)
+		goto err_region_put;
+
+	drm_dbg(&i915->drm, "Local memory: %pR\n", &mem->region);
+	drm_dbg(&i915->drm, "Local memory IO start: %pa\n",
+		&mem->io_start);
+	drm_info(&i915->drm, "Local memory IO size: %pa\n",
+		 &mem->io_size);
+	drm_info(&i915->drm, "Local memory available: %pa\n",
+		 &lmem_size);
+
+	if (io_size < lmem_size)
+		drm_info(&i915->drm, "Using a reduced BAR size of %lluMiB. Consider enabling 'Resizable BAR' or similar, if available in the BIOS.\n",
+			 (u64)io_size >> 20);
+
+	return mem;
+
+err_region_put:
+	intel_memory_region_destroy(mem);
+	return ERR_PTR(err);
+}
+
+struct intel_memory_region *intel_gt_setup_lmem(struct intel_gt *gt)
+{
+	return setup_lmem(gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_region_lmem.h b/drivers/gpu/drm/i915/gt/intel_region_lmem.h
new file mode 100644
index 000000000..1438576b5
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_region_lmem.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_REGION_LMEM_H
+#define __INTEL_REGION_LMEM_H
+
+struct intel_gt;
+
+struct intel_memory_region *intel_gt_setup_lmem(struct intel_gt *gt);
+
+#endif /* !__INTEL_REGION_LMEM_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_renderstate.c b/drivers/gpu/drm/i915/gt/intel_renderstate.c
new file mode 100644
index 000000000..5121e6dc2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_renderstate.c
@@ -0,0 +1,254 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#include "gem/i915_gem_internal.h"
+
+#include "i915_drv.h"
+#include "intel_renderstate.h"
+#include "intel_context.h"
+#include "intel_gpu_commands.h"
+#include "intel_ring.h"
+
+static const struct intel_renderstate_rodata *
+render_state_get_rodata(const struct intel_engine_cs *engine)
+{
+	if (engine->class != RENDER_CLASS)
+		return NULL;
+
+	switch (GRAPHICS_VER(engine->i915)) {
+	case 6:
+		return &gen6_null_state;
+	case 7:
+		return &gen7_null_state;
+	case 8:
+		return &gen8_null_state;
+	case 9:
+		return &gen9_null_state;
+	}
+
+	return NULL;
+}
+
+/*
+ * Macro to add commands to auxiliary batch.
+ * This macro only checks for page overflow before inserting the commands,
+ * this is sufficient as the null state generator makes the final batch
+ * with two passes to build command and state separately. At this point
+ * the size of both are known and it compacts them by relocating the state
+ * right after the commands taking care of alignment so we should sufficient
+ * space below them for adding new commands.
+ */
+#define OUT_BATCH(batch, i, val)				\
+	do {							\
+		if ((i) >= PAGE_SIZE / sizeof(u32))		\
+			goto out;				\
+		(batch)[(i)++] = (val);				\
+	} while (0)
+
+static int render_state_setup(struct intel_renderstate *so,
+			      struct drm_i915_private *i915)
+{
+	const struct intel_renderstate_rodata *rodata = so->rodata;
+	unsigned int i = 0, reloc_index = 0;
+	int ret = -EINVAL;
+	u32 *d;
+
+	d = i915_gem_object_pin_map(so->vma->obj, I915_MAP_WB);
+	if (IS_ERR(d))
+		return PTR_ERR(d);
+
+	while (i < rodata->batch_items) {
+		u32 s = rodata->batch[i];
+
+		if (i * 4  == rodata->reloc[reloc_index]) {
+			u64 r = s + so->vma->node.start;
+
+			s = lower_32_bits(r);
+			if (HAS_64BIT_RELOC(i915)) {
+				if (i + 1 >= rodata->batch_items ||
+				    rodata->batch[i + 1] != 0)
+					goto out;
+
+				d[i++] = s;
+				s = upper_32_bits(r);
+			}
+
+			reloc_index++;
+		}
+
+		d[i++] = s;
+	}
+
+	if (rodata->reloc[reloc_index] != -1) {
+		drm_err(&i915->drm, "only %d relocs resolved\n", reloc_index);
+		goto out;
+	}
+
+	so->batch_offset = i915_ggtt_offset(so->vma);
+	so->batch_size = rodata->batch_items * sizeof(u32);
+
+	while (i % CACHELINE_DWORDS)
+		OUT_BATCH(d, i, MI_NOOP);
+
+	so->aux_offset = i * sizeof(u32);
+
+	if (HAS_POOLED_EU(i915)) {
+		/*
+		 * We always program 3x6 pool config but depending upon which
+		 * subslice is disabled HW drops down to appropriate config
+		 * shown below.
+		 *
+		 * In the below table 2x6 config always refers to
+		 * fused-down version, native 2x6 is not available and can
+		 * be ignored
+		 *
+		 * SNo  subslices config                eu pool configuration
+		 * -----------------------------------------------------------
+		 * 1    3 subslices enabled (3x6)  -    0x00777000  (9+9)
+		 * 2    ss0 disabled (2x6)         -    0x00777000  (3+9)
+		 * 3    ss1 disabled (2x6)         -    0x00770000  (6+6)
+		 * 4    ss2 disabled (2x6)         -    0x00007000  (9+3)
+		 */
+		u32 eu_pool_config = 0x00777000;
+
+		OUT_BATCH(d, i, GEN9_MEDIA_POOL_STATE);
+		OUT_BATCH(d, i, GEN9_MEDIA_POOL_ENABLE);
+		OUT_BATCH(d, i, eu_pool_config);
+		OUT_BATCH(d, i, 0);
+		OUT_BATCH(d, i, 0);
+		OUT_BATCH(d, i, 0);
+	}
+
+	OUT_BATCH(d, i, MI_BATCH_BUFFER_END);
+	so->aux_size = i * sizeof(u32) - so->aux_offset;
+	so->aux_offset += so->batch_offset;
+	/*
+	 * Since we are sending length, we need to strictly conform to
+	 * all requirements. For Gen2 this must be a multiple of 8.
+	 */
+	so->aux_size = ALIGN(so->aux_size, 8);
+
+	ret = 0;
+out:
+	__i915_gem_object_flush_map(so->vma->obj, 0, i * sizeof(u32));
+	__i915_gem_object_release_map(so->vma->obj);
+	return ret;
+}
+
+#undef OUT_BATCH
+
+int intel_renderstate_init(struct intel_renderstate *so,
+			   struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct drm_i915_gem_object *obj = NULL;
+	int err;
+
+	memset(so, 0, sizeof(*so));
+
+	so->rodata = render_state_get_rodata(engine);
+	if (so->rodata) {
+		if (so->rodata->batch_items * 4 > PAGE_SIZE)
+			return -EINVAL;
+
+		obj = i915_gem_object_create_internal(engine->i915, PAGE_SIZE);
+		if (IS_ERR(obj))
+			return PTR_ERR(obj);
+
+		so->vma = i915_vma_instance(obj, &engine->gt->ggtt->vm, NULL);
+		if (IS_ERR(so->vma)) {
+			err = PTR_ERR(so->vma);
+			goto err_obj;
+		}
+	}
+
+	i915_gem_ww_ctx_init(&so->ww, true);
+retry:
+	err = intel_context_pin_ww(ce, &so->ww);
+	if (err)
+		goto err_fini;
+
+	/* return early if there's nothing to setup */
+	if (!err && !so->rodata)
+		return 0;
+
+	err = i915_gem_object_lock(so->vma->obj, &so->ww);
+	if (err)
+		goto err_context;
+
+	err = i915_vma_pin_ww(so->vma, &so->ww, 0, 0, PIN_GLOBAL | PIN_HIGH);
+	if (err)
+		goto err_context;
+
+	err = render_state_setup(so, engine->i915);
+	if (err)
+		goto err_unpin;
+
+	return 0;
+
+err_unpin:
+	i915_vma_unpin(so->vma);
+err_context:
+	intel_context_unpin(ce);
+err_fini:
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&so->ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&so->ww);
+err_obj:
+	if (obj)
+		i915_gem_object_put(obj);
+	so->vma = NULL;
+	return err;
+}
+
+int intel_renderstate_emit(struct intel_renderstate *so,
+			   struct i915_request *rq)
+{
+	struct intel_engine_cs *engine = rq->engine;
+	int err;
+
+	if (!so->vma)
+		return 0;
+
+	err = i915_request_await_object(rq, so->vma->obj, false);
+	if (err == 0)
+		err = i915_vma_move_to_active(so->vma, rq, 0);
+	if (err)
+		return err;
+
+	err = engine->emit_bb_start(rq,
+				    so->batch_offset, so->batch_size,
+				    I915_DISPATCH_SECURE);
+	if (err)
+		return err;
+
+	if (so->aux_size > 8) {
+		err = engine->emit_bb_start(rq,
+					    so->aux_offset, so->aux_size,
+					    I915_DISPATCH_SECURE);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+void intel_renderstate_fini(struct intel_renderstate *so,
+			    struct intel_context *ce)
+{
+	if (so->vma) {
+		i915_vma_unpin(so->vma);
+		i915_vma_close(so->vma);
+	}
+
+	intel_context_unpin(ce);
+	i915_gem_ww_ctx_fini(&so->ww);
+
+	if (so->vma)
+		i915_gem_object_put(so->vma->obj);
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_renderstate.h b/drivers/gpu/drm/i915/gt/intel_renderstate.h
new file mode 100644
index 000000000..4da4c5234
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_renderstate.h
@@ -0,0 +1,52 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#ifndef _INTEL_RENDERSTATE_H_
+#define _INTEL_RENDERSTATE_H_
+
+#include <linux/types.h>
+#include "i915_gem.h"
+#include "i915_gem_ww.h"
+
+struct i915_request;
+struct intel_context;
+struct i915_vma;
+
+struct intel_renderstate_rodata {
+	const u32 *reloc;
+	const u32 *batch;
+	const u32 batch_items;
+};
+
+#define RO_RENDERSTATE(_g)						\
+	const struct intel_renderstate_rodata gen ## _g ## _null_state = { \
+		.reloc = gen ## _g ## _null_state_relocs,		\
+		.batch = gen ## _g ## _null_state_batch,		\
+		.batch_items = sizeof(gen ## _g ## _null_state_batch)/4, \
+	}
+
+extern const struct intel_renderstate_rodata gen6_null_state;
+extern const struct intel_renderstate_rodata gen7_null_state;
+extern const struct intel_renderstate_rodata gen8_null_state;
+extern const struct intel_renderstate_rodata gen9_null_state;
+
+struct intel_renderstate {
+	struct i915_gem_ww_ctx ww;
+	const struct intel_renderstate_rodata *rodata;
+	struct i915_vma *vma;
+	u32 batch_offset;
+	u32 batch_size;
+	u32 aux_offset;
+	u32 aux_size;
+};
+
+int intel_renderstate_init(struct intel_renderstate *so,
+			   struct intel_context *ce);
+int intel_renderstate_emit(struct intel_renderstate *so,
+			   struct i915_request *rq);
+void intel_renderstate_fini(struct intel_renderstate *so,
+			    struct intel_context *ce);
+
+#endif /* _INTEL_RENDERSTATE_H_ */
diff --git a/drivers/gpu/drm/i915/gt/intel_reset.c b/drivers/gpu/drm/i915/gt/intel_reset.c
new file mode 100644
index 000000000..10b930eaa
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_reset.c
@@ -0,0 +1,1557 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2008-2018 Intel Corporation
+ */
+
+#include <linux/sched/mm.h>
+#include <linux/stop_machine.h>
+#include <linux/string_helpers.h>
+
+#include "display/intel_display.h"
+#include "display/intel_overlay.h"
+
+#include "gem/i915_gem_context.h"
+
+#include "gt/intel_gt_regs.h"
+
+#include "i915_drv.h"
+#include "i915_file_private.h"
+#include "i915_gpu_error.h"
+#include "i915_irq.h"
+#include "intel_breadcrumbs.h"
+#include "intel_engine_pm.h"
+#include "intel_engine_regs.h"
+#include "intel_gt.h"
+#include "intel_gt_pm.h"
+#include "intel_gt_requests.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pci_config.h"
+#include "intel_reset.h"
+
+#include "uc/intel_guc.h"
+
+#define RESET_MAX_RETRIES 3
+
+/* XXX How to handle concurrent GGTT updates using tiling registers? */
+#define RESET_UNDER_STOP_MACHINE 0
+
+static void rmw_set_fw(struct intel_uncore *uncore, i915_reg_t reg, u32 set)
+{
+	intel_uncore_rmw_fw(uncore, reg, 0, set);
+}
+
+static void rmw_clear_fw(struct intel_uncore *uncore, i915_reg_t reg, u32 clr)
+{
+	intel_uncore_rmw_fw(uncore, reg, clr, 0);
+}
+
+static void client_mark_guilty(struct i915_gem_context *ctx, bool banned)
+{
+	struct drm_i915_file_private *file_priv = ctx->file_priv;
+	unsigned long prev_hang;
+	unsigned int score;
+
+	if (IS_ERR_OR_NULL(file_priv))
+		return;
+
+	score = 0;
+	if (banned)
+		score = I915_CLIENT_SCORE_CONTEXT_BAN;
+
+	prev_hang = xchg(&file_priv->hang_timestamp, jiffies);
+	if (time_before(jiffies, prev_hang + I915_CLIENT_FAST_HANG_JIFFIES))
+		score += I915_CLIENT_SCORE_HANG_FAST;
+
+	if (score) {
+		atomic_add(score, &file_priv->ban_score);
+
+		drm_dbg(&ctx->i915->drm,
+			"client %s: gained %u ban score, now %u\n",
+			ctx->name, score,
+			atomic_read(&file_priv->ban_score));
+	}
+}
+
+static bool mark_guilty(struct i915_request *rq)
+{
+	struct i915_gem_context *ctx;
+	unsigned long prev_hang;
+	bool banned;
+	int i;
+
+	if (intel_context_is_closed(rq->context))
+		return true;
+
+	rcu_read_lock();
+	ctx = rcu_dereference(rq->context->gem_context);
+	if (ctx && !kref_get_unless_zero(&ctx->ref))
+		ctx = NULL;
+	rcu_read_unlock();
+	if (!ctx)
+		return intel_context_is_banned(rq->context);
+
+	atomic_inc(&ctx->guilty_count);
+
+	/* Cool contexts are too cool to be banned! (Used for reset testing.) */
+	if (!i915_gem_context_is_bannable(ctx)) {
+		banned = false;
+		goto out;
+	}
+
+	drm_notice(&ctx->i915->drm,
+		   "%s context reset due to GPU hang\n",
+		   ctx->name);
+
+	/* Record the timestamp for the last N hangs */
+	prev_hang = ctx->hang_timestamp[0];
+	for (i = 0; i < ARRAY_SIZE(ctx->hang_timestamp) - 1; i++)
+		ctx->hang_timestamp[i] = ctx->hang_timestamp[i + 1];
+	ctx->hang_timestamp[i] = jiffies;
+
+	/* If we have hung N+1 times in rapid succession, we ban the context! */
+	banned = !i915_gem_context_is_recoverable(ctx);
+	if (time_before(jiffies, prev_hang + CONTEXT_FAST_HANG_JIFFIES))
+		banned = true;
+	if (banned)
+		drm_dbg(&ctx->i915->drm, "context %s: guilty %d, banned\n",
+			ctx->name, atomic_read(&ctx->guilty_count));
+
+	client_mark_guilty(ctx, banned);
+
+out:
+	i915_gem_context_put(ctx);
+	return banned;
+}
+
+static void mark_innocent(struct i915_request *rq)
+{
+	struct i915_gem_context *ctx;
+
+	rcu_read_lock();
+	ctx = rcu_dereference(rq->context->gem_context);
+	if (ctx)
+		atomic_inc(&ctx->active_count);
+	rcu_read_unlock();
+}
+
+void __i915_request_reset(struct i915_request *rq, bool guilty)
+{
+	bool banned = false;
+
+	RQ_TRACE(rq, "guilty? %s\n", str_yes_no(guilty));
+	GEM_BUG_ON(__i915_request_is_complete(rq));
+
+	rcu_read_lock(); /* protect the GEM context */
+	if (guilty) {
+		i915_request_set_error_once(rq, -EIO);
+		__i915_request_skip(rq);
+		banned = mark_guilty(rq);
+	} else {
+		i915_request_set_error_once(rq, -EAGAIN);
+		mark_innocent(rq);
+	}
+	rcu_read_unlock();
+
+	if (banned)
+		intel_context_ban(rq->context, rq);
+}
+
+static bool i915_in_reset(struct pci_dev *pdev)
+{
+	u8 gdrst;
+
+	pci_read_config_byte(pdev, I915_GDRST, &gdrst);
+	return gdrst & GRDOM_RESET_STATUS;
+}
+
+static int i915_do_reset(struct intel_gt *gt,
+			 intel_engine_mask_t engine_mask,
+			 unsigned int retry)
+{
+	struct pci_dev *pdev = to_pci_dev(gt->i915->drm.dev);
+	int err;
+
+	/* Assert reset for at least 20 usec, and wait for acknowledgement. */
+	pci_write_config_byte(pdev, I915_GDRST, GRDOM_RESET_ENABLE);
+	udelay(50);
+	err = wait_for_atomic(i915_in_reset(pdev), 50);
+
+	/* Clear the reset request. */
+	pci_write_config_byte(pdev, I915_GDRST, 0);
+	udelay(50);
+	if (!err)
+		err = wait_for_atomic(!i915_in_reset(pdev), 50);
+
+	return err;
+}
+
+static bool g4x_reset_complete(struct pci_dev *pdev)
+{
+	u8 gdrst;
+
+	pci_read_config_byte(pdev, I915_GDRST, &gdrst);
+	return (gdrst & GRDOM_RESET_ENABLE) == 0;
+}
+
+static int g33_do_reset(struct intel_gt *gt,
+			intel_engine_mask_t engine_mask,
+			unsigned int retry)
+{
+	struct pci_dev *pdev = to_pci_dev(gt->i915->drm.dev);
+
+	pci_write_config_byte(pdev, I915_GDRST, GRDOM_RESET_ENABLE);
+	return wait_for_atomic(g4x_reset_complete(pdev), 50);
+}
+
+static int g4x_do_reset(struct intel_gt *gt,
+			intel_engine_mask_t engine_mask,
+			unsigned int retry)
+{
+	struct pci_dev *pdev = to_pci_dev(gt->i915->drm.dev);
+	struct intel_uncore *uncore = gt->uncore;
+	int ret;
+
+	/* WaVcpClkGateDisableForMediaReset:ctg,elk */
+	rmw_set_fw(uncore, VDECCLK_GATE_D, VCP_UNIT_CLOCK_GATE_DISABLE);
+	intel_uncore_posting_read_fw(uncore, VDECCLK_GATE_D);
+
+	pci_write_config_byte(pdev, I915_GDRST,
+			      GRDOM_MEDIA | GRDOM_RESET_ENABLE);
+	ret =  wait_for_atomic(g4x_reset_complete(pdev), 50);
+	if (ret) {
+		GT_TRACE(gt, "Wait for media reset failed\n");
+		goto out;
+	}
+
+	pci_write_config_byte(pdev, I915_GDRST,
+			      GRDOM_RENDER | GRDOM_RESET_ENABLE);
+	ret =  wait_for_atomic(g4x_reset_complete(pdev), 50);
+	if (ret) {
+		GT_TRACE(gt, "Wait for render reset failed\n");
+		goto out;
+	}
+
+out:
+	pci_write_config_byte(pdev, I915_GDRST, 0);
+
+	rmw_clear_fw(uncore, VDECCLK_GATE_D, VCP_UNIT_CLOCK_GATE_DISABLE);
+	intel_uncore_posting_read_fw(uncore, VDECCLK_GATE_D);
+
+	return ret;
+}
+
+static int ilk_do_reset(struct intel_gt *gt, intel_engine_mask_t engine_mask,
+			unsigned int retry)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	int ret;
+
+	intel_uncore_write_fw(uncore, ILK_GDSR,
+			      ILK_GRDOM_RENDER | ILK_GRDOM_RESET_ENABLE);
+	ret = __intel_wait_for_register_fw(uncore, ILK_GDSR,
+					   ILK_GRDOM_RESET_ENABLE, 0,
+					   5000, 0,
+					   NULL);
+	if (ret) {
+		GT_TRACE(gt, "Wait for render reset failed\n");
+		goto out;
+	}
+
+	intel_uncore_write_fw(uncore, ILK_GDSR,
+			      ILK_GRDOM_MEDIA | ILK_GRDOM_RESET_ENABLE);
+	ret = __intel_wait_for_register_fw(uncore, ILK_GDSR,
+					   ILK_GRDOM_RESET_ENABLE, 0,
+					   5000, 0,
+					   NULL);
+	if (ret) {
+		GT_TRACE(gt, "Wait for media reset failed\n");
+		goto out;
+	}
+
+out:
+	intel_uncore_write_fw(uncore, ILK_GDSR, 0);
+	intel_uncore_posting_read_fw(uncore, ILK_GDSR);
+	return ret;
+}
+
+/* Reset the hardware domains (GENX_GRDOM_*) specified by mask */
+static int gen6_hw_domain_reset(struct intel_gt *gt, u32 hw_domain_mask)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	int loops = 2;
+	int err;
+
+	/*
+	 * GEN6_GDRST is not in the gt power well, no need to check
+	 * for fifo space for the write or forcewake the chip for
+	 * the read
+	 */
+	do {
+		intel_uncore_write_fw(uncore, GEN6_GDRST, hw_domain_mask);
+
+		/*
+		 * Wait for the device to ack the reset requests.
+		 *
+		 * On some platforms, e.g. Jasperlake, we see that the
+		 * engine register state is not cleared until shortly after
+		 * GDRST reports completion, causing a failure as we try
+		 * to immediately resume while the internal state is still
+		 * in flux. If we immediately repeat the reset, the second
+		 * reset appears to serialise with the first, and since
+		 * it is a no-op, the registers should retain their reset
+		 * value. However, there is still a concern that upon
+		 * leaving the second reset, the internal engine state
+		 * is still in flux and not ready for resuming.
+		 */
+		err = __intel_wait_for_register_fw(uncore, GEN6_GDRST,
+						   hw_domain_mask, 0,
+						   2000, 0,
+						   NULL);
+	} while (err == 0 && --loops);
+	if (err)
+		GT_TRACE(gt,
+			 "Wait for 0x%08x engines reset failed\n",
+			 hw_domain_mask);
+
+	/*
+	 * As we have observed that the engine state is still volatile
+	 * after GDRST is acked, impose a small delay to let everything settle.
+	 */
+	udelay(50);
+
+	return err;
+}
+
+static int __gen6_reset_engines(struct intel_gt *gt,
+				intel_engine_mask_t engine_mask,
+				unsigned int retry)
+{
+	struct intel_engine_cs *engine;
+	u32 hw_mask;
+
+	if (engine_mask == ALL_ENGINES) {
+		hw_mask = GEN6_GRDOM_FULL;
+	} else {
+		intel_engine_mask_t tmp;
+
+		hw_mask = 0;
+		for_each_engine_masked(engine, gt, engine_mask, tmp) {
+			hw_mask |= engine->reset_domain;
+		}
+	}
+
+	return gen6_hw_domain_reset(gt, hw_mask);
+}
+
+static int gen6_reset_engines(struct intel_gt *gt,
+			      intel_engine_mask_t engine_mask,
+			      unsigned int retry)
+{
+	unsigned long flags;
+	int ret;
+
+	spin_lock_irqsave(&gt->uncore->lock, flags);
+	ret = __gen6_reset_engines(gt, engine_mask, retry);
+	spin_unlock_irqrestore(&gt->uncore->lock, flags);
+
+	return ret;
+}
+
+static struct intel_engine_cs *find_sfc_paired_vecs_engine(struct intel_engine_cs *engine)
+{
+	int vecs_id;
+
+	GEM_BUG_ON(engine->class != VIDEO_DECODE_CLASS);
+
+	vecs_id = _VECS((engine->instance) / 2);
+
+	return engine->gt->engine[vecs_id];
+}
+
+struct sfc_lock_data {
+	i915_reg_t lock_reg;
+	i915_reg_t ack_reg;
+	i915_reg_t usage_reg;
+	u32 lock_bit;
+	u32 ack_bit;
+	u32 usage_bit;
+	u32 reset_bit;
+};
+
+static void get_sfc_forced_lock_data(struct intel_engine_cs *engine,
+				     struct sfc_lock_data *sfc_lock)
+{
+	switch (engine->class) {
+	default:
+		MISSING_CASE(engine->class);
+		fallthrough;
+	case VIDEO_DECODE_CLASS:
+		sfc_lock->lock_reg = GEN11_VCS_SFC_FORCED_LOCK(engine->mmio_base);
+		sfc_lock->lock_bit = GEN11_VCS_SFC_FORCED_LOCK_BIT;
+
+		sfc_lock->ack_reg = GEN11_VCS_SFC_LOCK_STATUS(engine->mmio_base);
+		sfc_lock->ack_bit  = GEN11_VCS_SFC_LOCK_ACK_BIT;
+
+		sfc_lock->usage_reg = GEN11_VCS_SFC_LOCK_STATUS(engine->mmio_base);
+		sfc_lock->usage_bit = GEN11_VCS_SFC_USAGE_BIT;
+		sfc_lock->reset_bit = GEN11_VCS_SFC_RESET_BIT(engine->instance);
+
+		break;
+	case VIDEO_ENHANCEMENT_CLASS:
+		sfc_lock->lock_reg = GEN11_VECS_SFC_FORCED_LOCK(engine->mmio_base);
+		sfc_lock->lock_bit = GEN11_VECS_SFC_FORCED_LOCK_BIT;
+
+		sfc_lock->ack_reg = GEN11_VECS_SFC_LOCK_ACK(engine->mmio_base);
+		sfc_lock->ack_bit  = GEN11_VECS_SFC_LOCK_ACK_BIT;
+
+		sfc_lock->usage_reg = GEN11_VECS_SFC_USAGE(engine->mmio_base);
+		sfc_lock->usage_bit = GEN11_VECS_SFC_USAGE_BIT;
+		sfc_lock->reset_bit = GEN11_VECS_SFC_RESET_BIT(engine->instance);
+
+		break;
+	}
+}
+
+static int gen11_lock_sfc(struct intel_engine_cs *engine,
+			  u32 *reset_mask,
+			  u32 *unlock_mask)
+{
+	struct intel_uncore *uncore = engine->uncore;
+	u8 vdbox_sfc_access = engine->gt->info.vdbox_sfc_access;
+	struct sfc_lock_data sfc_lock;
+	bool lock_obtained, lock_to_other = false;
+	int ret;
+
+	switch (engine->class) {
+	case VIDEO_DECODE_CLASS:
+		if ((BIT(engine->instance) & vdbox_sfc_access) == 0)
+			return 0;
+
+		fallthrough;
+	case VIDEO_ENHANCEMENT_CLASS:
+		get_sfc_forced_lock_data(engine, &sfc_lock);
+
+		break;
+	default:
+		return 0;
+	}
+
+	if (!(intel_uncore_read_fw(uncore, sfc_lock.usage_reg) & sfc_lock.usage_bit)) {
+		struct intel_engine_cs *paired_vecs;
+
+		if (engine->class != VIDEO_DECODE_CLASS ||
+		    GRAPHICS_VER(engine->i915) != 12)
+			return 0;
+
+		/*
+		 * Wa_14010733141
+		 *
+		 * If the VCS-MFX isn't using the SFC, we also need to check
+		 * whether VCS-HCP is using it.  If so, we need to issue a *VE*
+		 * forced lock on the VE engine that shares the same SFC.
+		 */
+		if (!(intel_uncore_read_fw(uncore,
+					   GEN12_HCP_SFC_LOCK_STATUS(engine->mmio_base)) &
+		      GEN12_HCP_SFC_USAGE_BIT))
+			return 0;
+
+		paired_vecs = find_sfc_paired_vecs_engine(engine);
+		get_sfc_forced_lock_data(paired_vecs, &sfc_lock);
+		lock_to_other = true;
+		*unlock_mask |= paired_vecs->mask;
+	} else {
+		*unlock_mask |= engine->mask;
+	}
+
+	/*
+	 * If the engine is using an SFC, tell the engine that a software reset
+	 * is going to happen. The engine will then try to force lock the SFC.
+	 * If SFC ends up being locked to the engine we want to reset, we have
+	 * to reset it as well (we will unlock it once the reset sequence is
+	 * completed).
+	 */
+	rmw_set_fw(uncore, sfc_lock.lock_reg, sfc_lock.lock_bit);
+
+	ret = __intel_wait_for_register_fw(uncore,
+					   sfc_lock.ack_reg,
+					   sfc_lock.ack_bit,
+					   sfc_lock.ack_bit,
+					   1000, 0, NULL);
+
+	/*
+	 * Was the SFC released while we were trying to lock it?
+	 *
+	 * We should reset both the engine and the SFC if:
+	 *  - We were locking the SFC to this engine and the lock succeeded
+	 *       OR
+	 *  - We were locking the SFC to a different engine (Wa_14010733141)
+	 *    but the SFC was released before the lock was obtained.
+	 *
+	 * Otherwise we need only reset the engine by itself and we can
+	 * leave the SFC alone.
+	 */
+	lock_obtained = (intel_uncore_read_fw(uncore, sfc_lock.usage_reg) &
+			sfc_lock.usage_bit) != 0;
+	if (lock_obtained == lock_to_other)
+		return 0;
+
+	if (ret) {
+		ENGINE_TRACE(engine, "Wait for SFC forced lock ack failed\n");
+		return ret;
+	}
+
+	*reset_mask |= sfc_lock.reset_bit;
+	return 0;
+}
+
+static void gen11_unlock_sfc(struct intel_engine_cs *engine)
+{
+	struct intel_uncore *uncore = engine->uncore;
+	u8 vdbox_sfc_access = engine->gt->info.vdbox_sfc_access;
+	struct sfc_lock_data sfc_lock = {};
+
+	if (engine->class != VIDEO_DECODE_CLASS &&
+	    engine->class != VIDEO_ENHANCEMENT_CLASS)
+		return;
+
+	if (engine->class == VIDEO_DECODE_CLASS &&
+	    (BIT(engine->instance) & vdbox_sfc_access) == 0)
+		return;
+
+	get_sfc_forced_lock_data(engine, &sfc_lock);
+
+	rmw_clear_fw(uncore, sfc_lock.lock_reg, sfc_lock.lock_bit);
+}
+
+static int __gen11_reset_engines(struct intel_gt *gt,
+				 intel_engine_mask_t engine_mask,
+				 unsigned int retry)
+{
+	struct intel_engine_cs *engine;
+	intel_engine_mask_t tmp;
+	u32 reset_mask, unlock_mask = 0;
+	int ret;
+
+	if (engine_mask == ALL_ENGINES) {
+		reset_mask = GEN11_GRDOM_FULL;
+	} else {
+		reset_mask = 0;
+		for_each_engine_masked(engine, gt, engine_mask, tmp) {
+			reset_mask |= engine->reset_domain;
+			ret = gen11_lock_sfc(engine, &reset_mask, &unlock_mask);
+			if (ret)
+				goto sfc_unlock;
+		}
+	}
+
+	ret = gen6_hw_domain_reset(gt, reset_mask);
+
+sfc_unlock:
+	/*
+	 * We unlock the SFC based on the lock status and not the result of
+	 * gen11_lock_sfc to make sure that we clean properly if something
+	 * wrong happened during the lock (e.g. lock acquired after timeout
+	 * expiration).
+	 *
+	 * Due to Wa_14010733141, we may have locked an SFC to an engine that
+	 * wasn't being reset.  So instead of calling gen11_unlock_sfc()
+	 * on engine_mask, we instead call it on the mask of engines that our
+	 * gen11_lock_sfc() calls told us actually had locks attempted.
+	 */
+	for_each_engine_masked(engine, gt, unlock_mask, tmp)
+		gen11_unlock_sfc(engine);
+
+	return ret;
+}
+
+static int gen8_engine_reset_prepare(struct intel_engine_cs *engine)
+{
+	struct intel_uncore *uncore = engine->uncore;
+	const i915_reg_t reg = RING_RESET_CTL(engine->mmio_base);
+	u32 request, mask, ack;
+	int ret;
+
+	if (I915_SELFTEST_ONLY(should_fail(&engine->reset_timeout, 1)))
+		return -ETIMEDOUT;
+
+	ack = intel_uncore_read_fw(uncore, reg);
+	if (ack & RESET_CTL_CAT_ERROR) {
+		/*
+		 * For catastrophic errors, ready-for-reset sequence
+		 * needs to be bypassed: HAS#396813
+		 */
+		request = RESET_CTL_CAT_ERROR;
+		mask = RESET_CTL_CAT_ERROR;
+
+		/* Catastrophic errors need to be cleared by HW */
+		ack = 0;
+	} else if (!(ack & RESET_CTL_READY_TO_RESET)) {
+		request = RESET_CTL_REQUEST_RESET;
+		mask = RESET_CTL_READY_TO_RESET;
+		ack = RESET_CTL_READY_TO_RESET;
+	} else {
+		return 0;
+	}
+
+	intel_uncore_write_fw(uncore, reg, _MASKED_BIT_ENABLE(request));
+	ret = __intel_wait_for_register_fw(uncore, reg, mask, ack,
+					   700, 0, NULL);
+	if (ret)
+		drm_err(&engine->i915->drm,
+			"%s reset request timed out: {request: %08x, RESET_CTL: %08x}\n",
+			engine->name, request,
+			intel_uncore_read_fw(uncore, reg));
+
+	return ret;
+}
+
+static void gen8_engine_reset_cancel(struct intel_engine_cs *engine)
+{
+	intel_uncore_write_fw(engine->uncore,
+			      RING_RESET_CTL(engine->mmio_base),
+			      _MASKED_BIT_DISABLE(RESET_CTL_REQUEST_RESET));
+}
+
+static int gen8_reset_engines(struct intel_gt *gt,
+			      intel_engine_mask_t engine_mask,
+			      unsigned int retry)
+{
+	struct intel_engine_cs *engine;
+	const bool reset_non_ready = retry >= 1;
+	intel_engine_mask_t tmp;
+	unsigned long flags;
+	int ret;
+
+	spin_lock_irqsave(&gt->uncore->lock, flags);
+
+	for_each_engine_masked(engine, gt, engine_mask, tmp) {
+		ret = gen8_engine_reset_prepare(engine);
+		if (ret && !reset_non_ready)
+			goto skip_reset;
+
+		/*
+		 * If this is not the first failed attempt to prepare,
+		 * we decide to proceed anyway.
+		 *
+		 * By doing so we risk context corruption and with
+		 * some gens (kbl), possible system hang if reset
+		 * happens during active bb execution.
+		 *
+		 * We rather take context corruption instead of
+		 * failed reset with a wedged driver/gpu. And
+		 * active bb execution case should be covered by
+		 * stop_engines() we have before the reset.
+		 */
+	}
+
+	/*
+	 * Wa_22011100796:dg2, whenever Full soft reset is required,
+	 * reset all individual engines firstly, and then do a full soft reset.
+	 *
+	 * This is best effort, so ignore any error from the initial reset.
+	 */
+	if (IS_DG2(gt->i915) && engine_mask == ALL_ENGINES)
+		__gen11_reset_engines(gt, gt->info.engine_mask, 0);
+
+	if (GRAPHICS_VER(gt->i915) >= 11)
+		ret = __gen11_reset_engines(gt, engine_mask, retry);
+	else
+		ret = __gen6_reset_engines(gt, engine_mask, retry);
+
+skip_reset:
+	for_each_engine_masked(engine, gt, engine_mask, tmp)
+		gen8_engine_reset_cancel(engine);
+
+	spin_unlock_irqrestore(&gt->uncore->lock, flags);
+
+	return ret;
+}
+
+static int mock_reset(struct intel_gt *gt,
+		      intel_engine_mask_t mask,
+		      unsigned int retry)
+{
+	return 0;
+}
+
+typedef int (*reset_func)(struct intel_gt *,
+			  intel_engine_mask_t engine_mask,
+			  unsigned int retry);
+
+static reset_func intel_get_gpu_reset(const struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	if (is_mock_gt(gt))
+		return mock_reset;
+	else if (GRAPHICS_VER(i915) >= 8)
+		return gen8_reset_engines;
+	else if (GRAPHICS_VER(i915) >= 6)
+		return gen6_reset_engines;
+	else if (GRAPHICS_VER(i915) >= 5)
+		return ilk_do_reset;
+	else if (IS_G4X(i915))
+		return g4x_do_reset;
+	else if (IS_G33(i915) || IS_PINEVIEW(i915))
+		return g33_do_reset;
+	else if (GRAPHICS_VER(i915) >= 3)
+		return i915_do_reset;
+	else
+		return NULL;
+}
+
+int __intel_gt_reset(struct intel_gt *gt, intel_engine_mask_t engine_mask)
+{
+	const int retries = engine_mask == ALL_ENGINES ? RESET_MAX_RETRIES : 1;
+	reset_func reset;
+	int ret = -ETIMEDOUT;
+	int retry;
+
+	reset = intel_get_gpu_reset(gt);
+	if (!reset)
+		return -ENODEV;
+
+	/*
+	 * If the power well sleeps during the reset, the reset
+	 * request may be dropped and never completes (causing -EIO).
+	 */
+	intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);
+	for (retry = 0; ret == -ETIMEDOUT && retry < retries; retry++) {
+		GT_TRACE(gt, "engine_mask=%x\n", engine_mask);
+		preempt_disable();
+		ret = reset(gt, engine_mask, retry);
+		preempt_enable();
+	}
+	intel_uncore_forcewake_put(gt->uncore, FORCEWAKE_ALL);
+
+	return ret;
+}
+
+bool intel_has_gpu_reset(const struct intel_gt *gt)
+{
+	if (!gt->i915->params.reset)
+		return NULL;
+
+	return intel_get_gpu_reset(gt);
+}
+
+bool intel_has_reset_engine(const struct intel_gt *gt)
+{
+	if (gt->i915->params.reset < 2)
+		return false;
+
+	return INTEL_INFO(gt->i915)->has_reset_engine;
+}
+
+int intel_reset_guc(struct intel_gt *gt)
+{
+	u32 guc_domain =
+		GRAPHICS_VER(gt->i915) >= 11 ? GEN11_GRDOM_GUC : GEN9_GRDOM_GUC;
+	int ret;
+
+	GEM_BUG_ON(!HAS_GT_UC(gt->i915));
+
+	intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);
+	ret = gen6_hw_domain_reset(gt, guc_domain);
+	intel_uncore_forcewake_put(gt->uncore, FORCEWAKE_ALL);
+
+	return ret;
+}
+
+/*
+ * Ensure irq handler finishes, and not run again.
+ * Also return the active request so that we only search for it once.
+ */
+static void reset_prepare_engine(struct intel_engine_cs *engine)
+{
+	/*
+	 * During the reset sequence, we must prevent the engine from
+	 * entering RC6. As the context state is undefined until we restart
+	 * the engine, if it does enter RC6 during the reset, the state
+	 * written to the powercontext is undefined and so we may lose
+	 * GPU state upon resume, i.e. fail to restart after a reset.
+	 */
+	intel_uncore_forcewake_get(engine->uncore, FORCEWAKE_ALL);
+	if (engine->reset.prepare)
+		engine->reset.prepare(engine);
+}
+
+static void revoke_mmaps(struct intel_gt *gt)
+{
+	int i;
+
+	for (i = 0; i < gt->ggtt->num_fences; i++) {
+		struct drm_vma_offset_node *node;
+		struct i915_vma *vma;
+		u64 vma_offset;
+
+		vma = READ_ONCE(gt->ggtt->fence_regs[i].vma);
+		if (!vma)
+			continue;
+
+		if (!i915_vma_has_userfault(vma))
+			continue;
+
+		GEM_BUG_ON(vma->fence != &gt->ggtt->fence_regs[i]);
+
+		if (!vma->mmo)
+			continue;
+
+		node = &vma->mmo->vma_node;
+		vma_offset = vma->gtt_view.partial.offset << PAGE_SHIFT;
+
+		unmap_mapping_range(gt->i915->drm.anon_inode->i_mapping,
+				    drm_vma_node_offset_addr(node) + vma_offset,
+				    vma->size,
+				    1);
+	}
+}
+
+static intel_engine_mask_t reset_prepare(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	intel_engine_mask_t awake = 0;
+	enum intel_engine_id id;
+
+	/* For GuC mode, ensure submission is disabled before stopping ring */
+	intel_uc_reset_prepare(&gt->uc);
+
+	for_each_engine(engine, gt, id) {
+		if (intel_engine_pm_get_if_awake(engine))
+			awake |= engine->mask;
+		reset_prepare_engine(engine);
+	}
+
+	return awake;
+}
+
+static void gt_revoke(struct intel_gt *gt)
+{
+	revoke_mmaps(gt);
+}
+
+static int gt_reset(struct intel_gt *gt, intel_engine_mask_t stalled_mask)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err;
+
+	/*
+	 * Everything depends on having the GTT running, so we need to start
+	 * there.
+	 */
+	err = i915_ggtt_enable_hw(gt->i915);
+	if (err)
+		return err;
+
+	local_bh_disable();
+	for_each_engine(engine, gt, id)
+		__intel_engine_reset(engine, stalled_mask & engine->mask);
+	local_bh_enable();
+
+	intel_uc_reset(&gt->uc, ALL_ENGINES);
+
+	intel_ggtt_restore_fences(gt->ggtt);
+
+	return err;
+}
+
+static void reset_finish_engine(struct intel_engine_cs *engine)
+{
+	if (engine->reset.finish)
+		engine->reset.finish(engine);
+	intel_uncore_forcewake_put(engine->uncore, FORCEWAKE_ALL);
+
+	intel_engine_signal_breadcrumbs(engine);
+}
+
+static void reset_finish(struct intel_gt *gt, intel_engine_mask_t awake)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id) {
+		reset_finish_engine(engine);
+		if (awake & engine->mask)
+			intel_engine_pm_put(engine);
+	}
+
+	intel_uc_reset_finish(&gt->uc);
+}
+
+static void nop_submit_request(struct i915_request *request)
+{
+	RQ_TRACE(request, "-EIO\n");
+
+	request = i915_request_mark_eio(request);
+	if (request) {
+		i915_request_submit(request);
+		intel_engine_signal_breadcrumbs(request->engine);
+
+		i915_request_put(request);
+	}
+}
+
+static void __intel_gt_set_wedged(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	intel_engine_mask_t awake;
+	enum intel_engine_id id;
+
+	if (test_bit(I915_WEDGED, &gt->reset.flags))
+		return;
+
+	GT_TRACE(gt, "start\n");
+
+	/*
+	 * First, stop submission to hw, but do not yet complete requests by
+	 * rolling the global seqno forward (since this would complete requests
+	 * for which we haven't set the fence error to EIO yet).
+	 */
+	awake = reset_prepare(gt);
+
+	/* Even if the GPU reset fails, it should still stop the engines */
+	if (!INTEL_INFO(gt->i915)->gpu_reset_clobbers_display)
+		__intel_gt_reset(gt, ALL_ENGINES);
+
+	for_each_engine(engine, gt, id)
+		engine->submit_request = nop_submit_request;
+
+	/*
+	 * Make sure no request can slip through without getting completed by
+	 * either this call here to intel_engine_write_global_seqno, or the one
+	 * in nop_submit_request.
+	 */
+	synchronize_rcu_expedited();
+	set_bit(I915_WEDGED, &gt->reset.flags);
+
+	/* Mark all executing requests as skipped */
+	local_bh_disable();
+	for_each_engine(engine, gt, id)
+		if (engine->reset.cancel)
+			engine->reset.cancel(engine);
+	intel_uc_cancel_requests(&gt->uc);
+	local_bh_enable();
+
+	reset_finish(gt, awake);
+
+	GT_TRACE(gt, "end\n");
+}
+
+void intel_gt_set_wedged(struct intel_gt *gt)
+{
+	intel_wakeref_t wakeref;
+
+	if (test_bit(I915_WEDGED, &gt->reset.flags))
+		return;
+
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+	mutex_lock(&gt->reset.mutex);
+
+	if (GEM_SHOW_DEBUG()) {
+		struct drm_printer p = drm_debug_printer(__func__);
+		struct intel_engine_cs *engine;
+		enum intel_engine_id id;
+
+		drm_printf(&p, "called from %pS\n", (void *)_RET_IP_);
+		for_each_engine(engine, gt, id) {
+			if (intel_engine_is_idle(engine))
+				continue;
+
+			intel_engine_dump(engine, &p, "%s\n", engine->name);
+		}
+	}
+
+	__intel_gt_set_wedged(gt);
+
+	mutex_unlock(&gt->reset.mutex);
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+}
+
+static bool __intel_gt_unset_wedged(struct intel_gt *gt)
+{
+	struct intel_gt_timelines *timelines = &gt->timelines;
+	struct intel_timeline *tl;
+	bool ok;
+
+	if (!test_bit(I915_WEDGED, &gt->reset.flags))
+		return true;
+
+	/* Never fully initialised, recovery impossible */
+	if (intel_gt_has_unrecoverable_error(gt))
+		return false;
+
+	GT_TRACE(gt, "start\n");
+
+	/*
+	 * Before unwedging, make sure that all pending operations
+	 * are flushed and errored out - we may have requests waiting upon
+	 * third party fences. We marked all inflight requests as EIO, and
+	 * every execbuf since returned EIO, for consistency we want all
+	 * the currently pending requests to also be marked as EIO, which
+	 * is done inside our nop_submit_request - and so we must wait.
+	 *
+	 * No more can be submitted until we reset the wedged bit.
+	 */
+	spin_lock(&timelines->lock);
+	list_for_each_entry(tl, &timelines->active_list, link) {
+		struct dma_fence *fence;
+
+		fence = i915_active_fence_get(&tl->last_request);
+		if (!fence)
+			continue;
+
+		spin_unlock(&timelines->lock);
+
+		/*
+		 * All internal dependencies (i915_requests) will have
+		 * been flushed by the set-wedge, but we may be stuck waiting
+		 * for external fences. These should all be capped to 10s
+		 * (I915_FENCE_TIMEOUT) so this wait should not be unbounded
+		 * in the worst case.
+		 */
+		dma_fence_default_wait(fence, false, MAX_SCHEDULE_TIMEOUT);
+		dma_fence_put(fence);
+
+		/* Restart iteration after droping lock */
+		spin_lock(&timelines->lock);
+		tl = list_entry(&timelines->active_list, typeof(*tl), link);
+	}
+	spin_unlock(&timelines->lock);
+
+	/* We must reset pending GPU events before restoring our submission */
+	ok = !HAS_EXECLISTS(gt->i915); /* XXX better agnosticism desired */
+	if (!INTEL_INFO(gt->i915)->gpu_reset_clobbers_display)
+		ok = __intel_gt_reset(gt, ALL_ENGINES) == 0;
+	if (!ok) {
+		/*
+		 * Warn CI about the unrecoverable wedged condition.
+		 * Time for a reboot.
+		 */
+		add_taint_for_CI(gt->i915, TAINT_WARN);
+		return false;
+	}
+
+	/*
+	 * Undo nop_submit_request. We prevent all new i915 requests from
+	 * being queued (by disallowing execbuf whilst wedged) so having
+	 * waited for all active requests above, we know the system is idle
+	 * and do not have to worry about a thread being inside
+	 * engine->submit_request() as we swap over. So unlike installing
+	 * the nop_submit_request on reset, we can do this from normal
+	 * context and do not require stop_machine().
+	 */
+	intel_engines_reset_default_submission(gt);
+
+	GT_TRACE(gt, "end\n");
+
+	smp_mb__before_atomic(); /* complete takeover before enabling execbuf */
+	clear_bit(I915_WEDGED, &gt->reset.flags);
+
+	return true;
+}
+
+bool intel_gt_unset_wedged(struct intel_gt *gt)
+{
+	bool result;
+
+	mutex_lock(&gt->reset.mutex);
+	result = __intel_gt_unset_wedged(gt);
+	mutex_unlock(&gt->reset.mutex);
+
+	return result;
+}
+
+static int do_reset(struct intel_gt *gt, intel_engine_mask_t stalled_mask)
+{
+	int err, i;
+
+	err = __intel_gt_reset(gt, ALL_ENGINES);
+	for (i = 0; err && i < RESET_MAX_RETRIES; i++) {
+		msleep(10 * (i + 1));
+		err = __intel_gt_reset(gt, ALL_ENGINES);
+	}
+	if (err)
+		return err;
+
+	return gt_reset(gt, stalled_mask);
+}
+
+static int resume(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int ret;
+
+	for_each_engine(engine, gt, id) {
+		ret = intel_engine_resume(engine);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+/**
+ * intel_gt_reset - reset chip after a hang
+ * @gt: #intel_gt to reset
+ * @stalled_mask: mask of the stalled engines with the guilty requests
+ * @reason: user error message for why we are resetting
+ *
+ * Reset the chip.  Useful if a hang is detected. Marks the device as wedged
+ * on failure.
+ *
+ * Procedure is fairly simple:
+ *   - reset the chip using the reset reg
+ *   - re-init context state
+ *   - re-init hardware status page
+ *   - re-init ring buffer
+ *   - re-init interrupt state
+ *   - re-init display
+ */
+void intel_gt_reset(struct intel_gt *gt,
+		    intel_engine_mask_t stalled_mask,
+		    const char *reason)
+{
+	intel_engine_mask_t awake;
+	int ret;
+
+	GT_TRACE(gt, "flags=%lx\n", gt->reset.flags);
+
+	might_sleep();
+	GEM_BUG_ON(!test_bit(I915_RESET_BACKOFF, &gt->reset.flags));
+
+	/*
+	 * FIXME: Revoking cpu mmap ptes cannot be done from a dma_fence
+	 * critical section like gpu reset.
+	 */
+	gt_revoke(gt);
+
+	mutex_lock(&gt->reset.mutex);
+
+	/* Clear any previous failed attempts at recovery. Time to try again. */
+	if (!__intel_gt_unset_wedged(gt))
+		goto unlock;
+
+	if (reason)
+		drm_notice(&gt->i915->drm,
+			   "Resetting chip for %s\n", reason);
+	atomic_inc(&gt->i915->gpu_error.reset_count);
+
+	awake = reset_prepare(gt);
+
+	if (!intel_has_gpu_reset(gt)) {
+		if (gt->i915->params.reset)
+			drm_err(&gt->i915->drm, "GPU reset not supported\n");
+		else
+			drm_dbg(&gt->i915->drm, "GPU reset disabled\n");
+		goto error;
+	}
+
+	if (INTEL_INFO(gt->i915)->gpu_reset_clobbers_display)
+		intel_runtime_pm_disable_interrupts(gt->i915);
+
+	if (do_reset(gt, stalled_mask)) {
+		drm_err(&gt->i915->drm, "Failed to reset chip\n");
+		goto taint;
+	}
+
+	if (INTEL_INFO(gt->i915)->gpu_reset_clobbers_display)
+		intel_runtime_pm_enable_interrupts(gt->i915);
+
+	intel_overlay_reset(gt->i915);
+
+	/*
+	 * Next we need to restore the context, but we don't use those
+	 * yet either...
+	 *
+	 * Ring buffer needs to be re-initialized in the KMS case, or if X
+	 * was running at the time of the reset (i.e. we weren't VT
+	 * switched away).
+	 */
+	ret = intel_gt_init_hw(gt);
+	if (ret) {
+		drm_err(&gt->i915->drm,
+			"Failed to initialise HW following reset (%d)\n",
+			ret);
+		goto taint;
+	}
+
+	ret = resume(gt);
+	if (ret)
+		goto taint;
+
+finish:
+	reset_finish(gt, awake);
+unlock:
+	mutex_unlock(&gt->reset.mutex);
+	return;
+
+taint:
+	/*
+	 * History tells us that if we cannot reset the GPU now, we
+	 * never will. This then impacts everything that is run
+	 * subsequently. On failing the reset, we mark the driver
+	 * as wedged, preventing further execution on the GPU.
+	 * We also want to go one step further and add a taint to the
+	 * kernel so that any subsequent faults can be traced back to
+	 * this failure. This is important for CI, where if the
+	 * GPU/driver fails we would like to reboot and restart testing
+	 * rather than continue on into oblivion. For everyone else,
+	 * the system should still plod along, but they have been warned!
+	 */
+	add_taint_for_CI(gt->i915, TAINT_WARN);
+error:
+	__intel_gt_set_wedged(gt);
+	goto finish;
+}
+
+static int intel_gt_reset_engine(struct intel_engine_cs *engine)
+{
+	return __intel_gt_reset(engine->gt, engine->mask);
+}
+
+int __intel_engine_reset_bh(struct intel_engine_cs *engine, const char *msg)
+{
+	struct intel_gt *gt = engine->gt;
+	int ret;
+
+	ENGINE_TRACE(engine, "flags=%lx\n", gt->reset.flags);
+	GEM_BUG_ON(!test_bit(I915_RESET_ENGINE + engine->id, &gt->reset.flags));
+
+	if (intel_engine_uses_guc(engine))
+		return -ENODEV;
+
+	if (!intel_engine_pm_get_if_awake(engine))
+		return 0;
+
+	reset_prepare_engine(engine);
+
+	if (msg)
+		drm_notice(&engine->i915->drm,
+			   "Resetting %s for %s\n", engine->name, msg);
+	atomic_inc(&engine->i915->gpu_error.reset_engine_count[engine->uabi_class]);
+
+	ret = intel_gt_reset_engine(engine);
+	if (ret) {
+		/* If we fail here, we expect to fallback to a global reset */
+		ENGINE_TRACE(engine, "Failed to reset %s, err: %d\n", engine->name, ret);
+		goto out;
+	}
+
+	/*
+	 * The request that caused the hang is stuck on elsp, we know the
+	 * active request and can drop it, adjust head to skip the offending
+	 * request to resume executing remaining requests in the queue.
+	 */
+	__intel_engine_reset(engine, true);
+
+	/*
+	 * The engine and its registers (and workarounds in case of render)
+	 * have been reset to their default values. Follow the init_ring
+	 * process to program RING_MODE, HWSP and re-enable submission.
+	 */
+	ret = intel_engine_resume(engine);
+
+out:
+	intel_engine_cancel_stop_cs(engine);
+	reset_finish_engine(engine);
+	intel_engine_pm_put_async(engine);
+	return ret;
+}
+
+/**
+ * intel_engine_reset - reset GPU engine to recover from a hang
+ * @engine: engine to reset
+ * @msg: reason for GPU reset; or NULL for no drm_notice()
+ *
+ * Reset a specific GPU engine. Useful if a hang is detected.
+ * Returns zero on successful reset or otherwise an error code.
+ *
+ * Procedure is:
+ *  - identifies the request that caused the hang and it is dropped
+ *  - reset engine (which will force the engine to idle)
+ *  - re-init/configure engine
+ */
+int intel_engine_reset(struct intel_engine_cs *engine, const char *msg)
+{
+	int err;
+
+	local_bh_disable();
+	err = __intel_engine_reset_bh(engine, msg);
+	local_bh_enable();
+
+	return err;
+}
+
+static void intel_gt_reset_global(struct intel_gt *gt,
+				  u32 engine_mask,
+				  const char *reason)
+{
+	struct kobject *kobj = &gt->i915->drm.primary->kdev->kobj;
+	char *error_event[] = { I915_ERROR_UEVENT "=1", NULL };
+	char *reset_event[] = { I915_RESET_UEVENT "=1", NULL };
+	char *reset_done_event[] = { I915_ERROR_UEVENT "=0", NULL };
+	struct intel_wedge_me w;
+
+	kobject_uevent_env(kobj, KOBJ_CHANGE, error_event);
+
+	GT_TRACE(gt, "resetting chip, engines=%x\n", engine_mask);
+	kobject_uevent_env(kobj, KOBJ_CHANGE, reset_event);
+
+	/* Use a watchdog to ensure that our reset completes */
+	intel_wedge_on_timeout(&w, gt, 5 * HZ) {
+		intel_display_prepare_reset(gt->i915);
+
+		intel_gt_reset(gt, engine_mask, reason);
+
+		intel_display_finish_reset(gt->i915);
+	}
+
+	if (!test_bit(I915_WEDGED, &gt->reset.flags))
+		kobject_uevent_env(kobj, KOBJ_CHANGE, reset_done_event);
+}
+
+/**
+ * intel_gt_handle_error - handle a gpu error
+ * @gt: the intel_gt
+ * @engine_mask: mask representing engines that are hung
+ * @flags: control flags
+ * @fmt: Error message format string
+ *
+ * Do some basic checking of register state at error time and
+ * dump it to the syslog.  Also call i915_capture_error_state() to make
+ * sure we get a record and make it available in debugfs.  Fire a uevent
+ * so userspace knows something bad happened (should trigger collection
+ * of a ring dump etc.).
+ */
+void intel_gt_handle_error(struct intel_gt *gt,
+			   intel_engine_mask_t engine_mask,
+			   unsigned long flags,
+			   const char *fmt, ...)
+{
+	struct intel_engine_cs *engine;
+	intel_wakeref_t wakeref;
+	intel_engine_mask_t tmp;
+	char error_msg[80];
+	char *msg = NULL;
+
+	if (fmt) {
+		va_list args;
+
+		va_start(args, fmt);
+		vscnprintf(error_msg, sizeof(error_msg), fmt, args);
+		va_end(args);
+
+		msg = error_msg;
+	}
+
+	/*
+	 * In most cases it's guaranteed that we get here with an RPM
+	 * reference held, for example because there is a pending GPU
+	 * request that won't finish until the reset is done. This
+	 * isn't the case at least when we get here by doing a
+	 * simulated reset via debugfs, so get an RPM reference.
+	 */
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+
+	engine_mask &= gt->info.engine_mask;
+
+	if (flags & I915_ERROR_CAPTURE) {
+		i915_capture_error_state(gt, engine_mask, CORE_DUMP_FLAG_NONE);
+		intel_gt_clear_error_registers(gt, engine_mask);
+	}
+
+	/*
+	 * Try engine reset when available. We fall back to full reset if
+	 * single reset fails.
+	 */
+	if (!intel_uc_uses_guc_submission(&gt->uc) &&
+	    intel_has_reset_engine(gt) && !intel_gt_is_wedged(gt)) {
+		local_bh_disable();
+		for_each_engine_masked(engine, gt, engine_mask, tmp) {
+			BUILD_BUG_ON(I915_RESET_MODESET >= I915_RESET_ENGINE);
+			if (test_and_set_bit(I915_RESET_ENGINE + engine->id,
+					     &gt->reset.flags))
+				continue;
+
+			if (__intel_engine_reset_bh(engine, msg) == 0)
+				engine_mask &= ~engine->mask;
+
+			clear_and_wake_up_bit(I915_RESET_ENGINE + engine->id,
+					      &gt->reset.flags);
+		}
+		local_bh_enable();
+	}
+
+	if (!engine_mask)
+		goto out;
+
+	/* Full reset needs the mutex, stop any other user trying to do so. */
+	if (test_and_set_bit(I915_RESET_BACKOFF, &gt->reset.flags)) {
+		wait_event(gt->reset.queue,
+			   !test_bit(I915_RESET_BACKOFF, &gt->reset.flags));
+		goto out; /* piggy-back on the other reset */
+	}
+
+	/* Make sure i915_reset_trylock() sees the I915_RESET_BACKOFF */
+	synchronize_rcu_expedited();
+
+	/*
+	 * Prevent any other reset-engine attempt. We don't do this for GuC
+	 * submission the GuC owns the per-engine reset, not the i915.
+	 */
+	if (!intel_uc_uses_guc_submission(&gt->uc)) {
+		for_each_engine(engine, gt, tmp) {
+			while (test_and_set_bit(I915_RESET_ENGINE + engine->id,
+						&gt->reset.flags))
+				wait_on_bit(&gt->reset.flags,
+					    I915_RESET_ENGINE + engine->id,
+					    TASK_UNINTERRUPTIBLE);
+		}
+	}
+
+	/* Flush everyone using a resource about to be clobbered */
+	synchronize_srcu_expedited(&gt->reset.backoff_srcu);
+
+	intel_gt_reset_global(gt, engine_mask, msg);
+
+	if (!intel_uc_uses_guc_submission(&gt->uc)) {
+		for_each_engine(engine, gt, tmp)
+			clear_bit_unlock(I915_RESET_ENGINE + engine->id,
+					 &gt->reset.flags);
+	}
+	clear_bit_unlock(I915_RESET_BACKOFF, &gt->reset.flags);
+	smp_mb__after_atomic();
+	wake_up_all(&gt->reset.queue);
+
+out:
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+}
+
+int intel_gt_reset_trylock(struct intel_gt *gt, int *srcu)
+{
+	might_lock(&gt->reset.backoff_srcu);
+	might_sleep();
+
+	rcu_read_lock();
+	while (test_bit(I915_RESET_BACKOFF, &gt->reset.flags)) {
+		rcu_read_unlock();
+
+		if (wait_event_interruptible(gt->reset.queue,
+					     !test_bit(I915_RESET_BACKOFF,
+						       &gt->reset.flags)))
+			return -EINTR;
+
+		rcu_read_lock();
+	}
+	*srcu = srcu_read_lock(&gt->reset.backoff_srcu);
+	rcu_read_unlock();
+
+	return 0;
+}
+
+void intel_gt_reset_unlock(struct intel_gt *gt, int tag)
+__releases(&gt->reset.backoff_srcu)
+{
+	srcu_read_unlock(&gt->reset.backoff_srcu, tag);
+}
+
+int intel_gt_terminally_wedged(struct intel_gt *gt)
+{
+	might_sleep();
+
+	if (!intel_gt_is_wedged(gt))
+		return 0;
+
+	if (intel_gt_has_unrecoverable_error(gt))
+		return -EIO;
+
+	/* Reset still in progress? Maybe we will recover? */
+	if (wait_event_interruptible(gt->reset.queue,
+				     !test_bit(I915_RESET_BACKOFF,
+					       &gt->reset.flags)))
+		return -EINTR;
+
+	return intel_gt_is_wedged(gt) ? -EIO : 0;
+}
+
+void intel_gt_set_wedged_on_init(struct intel_gt *gt)
+{
+	BUILD_BUG_ON(I915_RESET_ENGINE + I915_NUM_ENGINES >
+		     I915_WEDGED_ON_INIT);
+	intel_gt_set_wedged(gt);
+	i915_disable_error_state(gt->i915, -ENODEV);
+	set_bit(I915_WEDGED_ON_INIT, &gt->reset.flags);
+
+	/* Wedged on init is non-recoverable */
+	add_taint_for_CI(gt->i915, TAINT_WARN);
+}
+
+void intel_gt_set_wedged_on_fini(struct intel_gt *gt)
+{
+	intel_gt_set_wedged(gt);
+	i915_disable_error_state(gt->i915, -ENODEV);
+	set_bit(I915_WEDGED_ON_FINI, &gt->reset.flags);
+	intel_gt_retire_requests(gt); /* cleanup any wedged requests */
+}
+
+void intel_gt_init_reset(struct intel_gt *gt)
+{
+	init_waitqueue_head(&gt->reset.queue);
+	mutex_init(&gt->reset.mutex);
+	init_srcu_struct(&gt->reset.backoff_srcu);
+
+	/*
+	 * While undesirable to wait inside the shrinker, complain anyway.
+	 *
+	 * If we have to wait during shrinking, we guarantee forward progress
+	 * by forcing the reset. Therefore during the reset we must not
+	 * re-enter the shrinker. By declaring that we take the reset mutex
+	 * within the shrinker, we forbid ourselves from performing any
+	 * fs-reclaim or taking related locks during reset.
+	 */
+	i915_gem_shrinker_taints_mutex(gt->i915, &gt->reset.mutex);
+
+	/* no GPU until we are ready! */
+	__set_bit(I915_WEDGED, &gt->reset.flags);
+}
+
+void intel_gt_fini_reset(struct intel_gt *gt)
+{
+	cleanup_srcu_struct(&gt->reset.backoff_srcu);
+}
+
+static void intel_wedge_me(struct work_struct *work)
+{
+	struct intel_wedge_me *w = container_of(work, typeof(*w), work.work);
+
+	drm_err(&w->gt->i915->drm,
+		"%s timed out, cancelling all in-flight rendering.\n",
+		w->name);
+	intel_gt_set_wedged(w->gt);
+}
+
+void __intel_init_wedge(struct intel_wedge_me *w,
+			struct intel_gt *gt,
+			long timeout,
+			const char *name)
+{
+	w->gt = gt;
+	w->name = name;
+
+	INIT_DELAYED_WORK_ONSTACK(&w->work, intel_wedge_me);
+	schedule_delayed_work(&w->work, timeout);
+}
+
+void __intel_fini_wedge(struct intel_wedge_me *w)
+{
+	cancel_delayed_work_sync(&w->work);
+	destroy_delayed_work_on_stack(&w->work);
+	w->gt = NULL;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_reset.c"
+#include "selftest_hangcheck.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_reset.h b/drivers/gpu/drm/i915/gt/intel_reset.h
new file mode 100644
index 000000000..adc734e67
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_reset.h
@@ -0,0 +1,80 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2008-2018 Intel Corporation
+ */
+
+#ifndef I915_RESET_H
+#define I915_RESET_H
+
+#include <linux/compiler.h>
+#include <linux/types.h>
+#include <linux/srcu.h>
+
+#include "intel_engine_types.h"
+#include "intel_reset_types.h"
+
+struct i915_request;
+struct intel_engine_cs;
+struct intel_gt;
+struct intel_guc;
+
+void intel_gt_init_reset(struct intel_gt *gt);
+void intel_gt_fini_reset(struct intel_gt *gt);
+
+__printf(4, 5)
+void intel_gt_handle_error(struct intel_gt *gt,
+			   intel_engine_mask_t engine_mask,
+			   unsigned long flags,
+			   const char *fmt, ...);
+#define I915_ERROR_CAPTURE BIT(0)
+
+void intel_gt_reset(struct intel_gt *gt,
+		    intel_engine_mask_t stalled_mask,
+		    const char *reason);
+int intel_engine_reset(struct intel_engine_cs *engine,
+		       const char *reason);
+int __intel_engine_reset_bh(struct intel_engine_cs *engine,
+			    const char *reason);
+
+void __i915_request_reset(struct i915_request *rq, bool guilty);
+
+int __must_check intel_gt_reset_trylock(struct intel_gt *gt, int *srcu);
+void intel_gt_reset_unlock(struct intel_gt *gt, int tag);
+
+void intel_gt_set_wedged(struct intel_gt *gt);
+bool intel_gt_unset_wedged(struct intel_gt *gt);
+int intel_gt_terminally_wedged(struct intel_gt *gt);
+
+/*
+ * There's no unset_wedged_on_init paired with this one.
+ * Once we're wedged on init, there's no going back.
+ * Same thing for unset_wedged_on_fini.
+ */
+void intel_gt_set_wedged_on_init(struct intel_gt *gt);
+void intel_gt_set_wedged_on_fini(struct intel_gt *gt);
+
+int __intel_gt_reset(struct intel_gt *gt, intel_engine_mask_t engine_mask);
+
+int intel_reset_guc(struct intel_gt *gt);
+
+struct intel_wedge_me {
+	struct delayed_work work;
+	struct intel_gt *gt;
+	const char *name;
+};
+
+void __intel_init_wedge(struct intel_wedge_me *w,
+			struct intel_gt *gt,
+			long timeout,
+			const char *name);
+void __intel_fini_wedge(struct intel_wedge_me *w);
+
+#define intel_wedge_on_timeout(W, GT, TIMEOUT)				\
+	for (__intel_init_wedge((W), (GT), (TIMEOUT), __func__);	\
+	     (W)->gt;							\
+	     __intel_fini_wedge((W)))
+
+bool intel_has_gpu_reset(const struct intel_gt *gt);
+bool intel_has_reset_engine(const struct intel_gt *gt);
+
+#endif /* I915_RESET_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_reset_types.h b/drivers/gpu/drm/i915/gt/intel_reset_types.h
new file mode 100644
index 000000000..9312b29f5
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_reset_types.h
@@ -0,0 +1,61 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_RESET_TYPES_H_
+#define __INTEL_RESET_TYPES_H_
+
+#include <linux/mutex.h>
+#include <linux/wait.h>
+#include <linux/srcu.h>
+
+struct intel_reset {
+	/**
+	 * flags: Control various stages of the GPU reset
+	 *
+	 * #I915_RESET_BACKOFF - When we start a global reset, we need to
+	 * serialise with any other users attempting to do the same, and
+	 * any global resources that may be clobber by the reset (such as
+	 * FENCE registers).
+	 *
+	 * #I915_RESET_ENGINE[num_engines] - Since the driver doesn't need to
+	 * acquire the struct_mutex to reset an engine, we need an explicit
+	 * flag to prevent two concurrent reset attempts in the same engine.
+	 * As the number of engines continues to grow, allocate the flags from
+	 * the most significant bits.
+	 *
+	 * #I915_WEDGED - If reset fails and we can no longer use the GPU,
+	 * we set the #I915_WEDGED bit. Prior to command submission, e.g.
+	 * i915_request_alloc(), this bit is checked and the sequence
+	 * aborted (with -EIO reported to userspace) if set.
+	 *
+	 * #I915_WEDGED_ON_INIT - If we fail to initialize the GPU we can no
+	 * longer use the GPU - similar to #I915_WEDGED bit. The difference in
+	 * the way we're handling "forced" unwedged (e.g. through debugfs),
+	 * which is not allowed in case we failed to initialize.
+	 *
+	 * #I915_WEDGED_ON_FINI - Similar to #I915_WEDGED_ON_INIT, except we
+	 * use it to mark that the GPU is no longer available (and prevent
+	 * users from using it).
+	 */
+	unsigned long flags;
+#define I915_RESET_BACKOFF	0
+#define I915_RESET_MODESET	1
+#define I915_RESET_ENGINE	2
+#define I915_WEDGED_ON_INIT	(BITS_PER_LONG - 3)
+#define I915_WEDGED_ON_FINI	(BITS_PER_LONG - 2)
+#define I915_WEDGED		(BITS_PER_LONG - 1)
+
+	struct mutex mutex; /* serialises wedging/unwedging */
+
+	/**
+	 * Waitqueue to signal when the reset has completed. Used by clients
+	 * that wait for dev_priv->mm.wedged to settle.
+	 */
+	wait_queue_head_t queue;
+
+	struct srcu_struct backoff_srcu;
+};
+
+#endif /* _INTEL_RESET_TYPES_H_ */
diff --git a/drivers/gpu/drm/i915/gt/intel_ring.c b/drivers/gpu/drm/i915/gt/intel_ring.c
new file mode 100644
index 000000000..fb99143be
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_ring.c
@@ -0,0 +1,336 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_lmem.h"
+#include "gem/i915_gem_object.h"
+
+#include "i915_drv.h"
+#include "i915_vma.h"
+#include "intel_engine.h"
+#include "intel_engine_regs.h"
+#include "intel_gpu_commands.h"
+#include "intel_ring.h"
+#include "intel_timeline.h"
+
+unsigned int intel_ring_update_space(struct intel_ring *ring)
+{
+	unsigned int space;
+
+	space = __intel_ring_space(ring->head, ring->emit, ring->size);
+
+	ring->space = space;
+	return space;
+}
+
+void __intel_ring_pin(struct intel_ring *ring)
+{
+	GEM_BUG_ON(!atomic_read(&ring->pin_count));
+	atomic_inc(&ring->pin_count);
+}
+
+int intel_ring_pin(struct intel_ring *ring, struct i915_gem_ww_ctx *ww)
+{
+	struct i915_vma *vma = ring->vma;
+	unsigned int flags;
+	void *addr;
+	int ret;
+
+	if (atomic_fetch_inc(&ring->pin_count))
+		return 0;
+
+	/* Ring wraparound at offset 0 sometimes hangs. No idea why. */
+	flags = PIN_OFFSET_BIAS | i915_ggtt_pin_bias(vma);
+
+	if (i915_gem_object_is_stolen(vma->obj))
+		flags |= PIN_MAPPABLE;
+	else
+		flags |= PIN_HIGH;
+
+	ret = i915_ggtt_pin(vma, ww, 0, flags);
+	if (unlikely(ret))
+		goto err_unpin;
+
+	if (i915_vma_is_map_and_fenceable(vma) && !HAS_LLC(vma->vm->i915)) {
+		addr = (void __force *)i915_vma_pin_iomap(vma);
+	} else {
+		int type = i915_coherent_map_type(vma->vm->i915, vma->obj, false);
+
+		addr = i915_gem_object_pin_map(vma->obj, type);
+	}
+
+	if (IS_ERR(addr)) {
+		ret = PTR_ERR(addr);
+		goto err_ring;
+	}
+
+	i915_vma_make_unshrinkable(vma);
+
+	/* Discard any unused bytes beyond that submitted to hw. */
+	intel_ring_reset(ring, ring->emit);
+
+	ring->vaddr = addr;
+	return 0;
+
+err_ring:
+	i915_vma_unpin(vma);
+err_unpin:
+	atomic_dec(&ring->pin_count);
+	return ret;
+}
+
+void intel_ring_reset(struct intel_ring *ring, u32 tail)
+{
+	tail = intel_ring_wrap(ring, tail);
+	ring->tail = tail;
+	ring->head = tail;
+	ring->emit = tail;
+	intel_ring_update_space(ring);
+}
+
+void intel_ring_unpin(struct intel_ring *ring)
+{
+	struct i915_vma *vma = ring->vma;
+
+	if (!atomic_dec_and_test(&ring->pin_count))
+		return;
+
+	i915_vma_unset_ggtt_write(vma);
+	if (i915_vma_is_map_and_fenceable(vma) && !HAS_LLC(vma->vm->i915))
+		i915_vma_unpin_iomap(vma);
+	else
+		i915_gem_object_unpin_map(vma->obj);
+
+	i915_vma_make_purgeable(vma);
+	i915_vma_unpin(vma);
+}
+
+static struct i915_vma *create_ring_vma(struct i915_ggtt *ggtt, int size)
+{
+	struct i915_address_space *vm = &ggtt->vm;
+	struct drm_i915_private *i915 = vm->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+
+	obj = i915_gem_object_create_lmem(i915, size, I915_BO_ALLOC_VOLATILE |
+					  I915_BO_ALLOC_PM_VOLATILE);
+	if (IS_ERR(obj) && i915_ggtt_has_aperture(ggtt) && !HAS_LLC(i915))
+		obj = i915_gem_object_create_stolen(i915, size);
+	if (IS_ERR(obj))
+		obj = i915_gem_object_create_internal(i915, size);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	/*
+	 * Mark ring buffers as read-only from GPU side (so no stray overwrites)
+	 * if supported by the platform's GGTT.
+	 */
+	if (vm->has_read_only)
+		i915_gem_object_set_readonly(obj);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma))
+		goto err;
+
+	return vma;
+
+err:
+	i915_gem_object_put(obj);
+	return vma;
+}
+
+struct intel_ring *
+intel_engine_create_ring(struct intel_engine_cs *engine, int size)
+{
+	struct drm_i915_private *i915 = engine->i915;
+	struct intel_ring *ring;
+	struct i915_vma *vma;
+
+	GEM_BUG_ON(!is_power_of_2(size));
+	GEM_BUG_ON(RING_CTL_SIZE(size) & ~RING_NR_PAGES);
+
+	ring = kzalloc(sizeof(*ring), GFP_KERNEL);
+	if (!ring)
+		return ERR_PTR(-ENOMEM);
+
+	kref_init(&ring->ref);
+	ring->size = size;
+	ring->wrap = BITS_PER_TYPE(ring->size) - ilog2(size);
+
+	/*
+	 * Workaround an erratum on the i830 which causes a hang if
+	 * the TAIL pointer points to within the last 2 cachelines
+	 * of the buffer.
+	 */
+	ring->effective_size = size;
+	if (IS_I830(i915) || IS_I845G(i915))
+		ring->effective_size -= 2 * CACHELINE_BYTES;
+
+	intel_ring_update_space(ring);
+
+	vma = create_ring_vma(engine->gt->ggtt, size);
+	if (IS_ERR(vma)) {
+		kfree(ring);
+		return ERR_CAST(vma);
+	}
+	ring->vma = vma;
+
+	return ring;
+}
+
+void intel_ring_free(struct kref *ref)
+{
+	struct intel_ring *ring = container_of(ref, typeof(*ring), ref);
+
+	i915_vma_put(ring->vma);
+	kfree(ring);
+}
+
+static noinline int
+wait_for_space(struct intel_ring *ring,
+	       struct intel_timeline *tl,
+	       unsigned int bytes)
+{
+	struct i915_request *target;
+	long timeout;
+
+	if (intel_ring_update_space(ring) >= bytes)
+		return 0;
+
+	GEM_BUG_ON(list_empty(&tl->requests));
+	list_for_each_entry(target, &tl->requests, link) {
+		if (target->ring != ring)
+			continue;
+
+		/* Would completion of this request free enough space? */
+		if (bytes <= __intel_ring_space(target->postfix,
+						ring->emit, ring->size))
+			break;
+	}
+
+	if (GEM_WARN_ON(&target->link == &tl->requests))
+		return -ENOSPC;
+
+	timeout = i915_request_wait(target,
+				    I915_WAIT_INTERRUPTIBLE,
+				    MAX_SCHEDULE_TIMEOUT);
+	if (timeout < 0)
+		return timeout;
+
+	i915_request_retire_upto(target);
+
+	intel_ring_update_space(ring);
+	GEM_BUG_ON(ring->space < bytes);
+	return 0;
+}
+
+u32 *intel_ring_begin(struct i915_request *rq, unsigned int num_dwords)
+{
+	struct intel_ring *ring = rq->ring;
+	const unsigned int remain_usable = ring->effective_size - ring->emit;
+	const unsigned int bytes = num_dwords * sizeof(u32);
+	unsigned int need_wrap = 0;
+	unsigned int total_bytes;
+	u32 *cs;
+
+	/* Packets must be qword aligned. */
+	GEM_BUG_ON(num_dwords & 1);
+
+	total_bytes = bytes + rq->reserved_space;
+	GEM_BUG_ON(total_bytes > ring->effective_size);
+
+	if (unlikely(total_bytes > remain_usable)) {
+		const int remain_actual = ring->size - ring->emit;
+
+		if (bytes > remain_usable) {
+			/*
+			 * Not enough space for the basic request. So need to
+			 * flush out the remainder and then wait for
+			 * base + reserved.
+			 */
+			total_bytes += remain_actual;
+			need_wrap = remain_actual | 1;
+		} else  {
+			/*
+			 * The base request will fit but the reserved space
+			 * falls off the end. So we don't need an immediate
+			 * wrap and only need to effectively wait for the
+			 * reserved size from the start of ringbuffer.
+			 */
+			total_bytes = rq->reserved_space + remain_actual;
+		}
+	}
+
+	if (unlikely(total_bytes > ring->space)) {
+		int ret;
+
+		/*
+		 * Space is reserved in the ringbuffer for finalising the
+		 * request, as that cannot be allowed to fail. During request
+		 * finalisation, reserved_space is set to 0 to stop the
+		 * overallocation and the assumption is that then we never need
+		 * to wait (which has the risk of failing with EINTR).
+		 *
+		 * See also i915_request_alloc() and i915_request_add().
+		 */
+		GEM_BUG_ON(!rq->reserved_space);
+
+		ret = wait_for_space(ring,
+				     i915_request_timeline(rq),
+				     total_bytes);
+		if (unlikely(ret))
+			return ERR_PTR(ret);
+	}
+
+	if (unlikely(need_wrap)) {
+		need_wrap &= ~1;
+		GEM_BUG_ON(need_wrap > ring->space);
+		GEM_BUG_ON(ring->emit + need_wrap > ring->size);
+		GEM_BUG_ON(!IS_ALIGNED(need_wrap, sizeof(u64)));
+
+		/* Fill the tail with MI_NOOP */
+		memset64(ring->vaddr + ring->emit, 0, need_wrap / sizeof(u64));
+		ring->space -= need_wrap;
+		ring->emit = 0;
+	}
+
+	GEM_BUG_ON(ring->emit > ring->size - bytes);
+	GEM_BUG_ON(ring->space < bytes);
+	cs = ring->vaddr + ring->emit;
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		memset32(cs, POISON_INUSE, bytes / sizeof(*cs));
+	ring->emit += bytes;
+	ring->space -= bytes;
+
+	return cs;
+}
+
+/* Align the ring tail to a cacheline boundary */
+int intel_ring_cacheline_align(struct i915_request *rq)
+{
+	int num_dwords;
+	void *cs;
+
+	num_dwords = (rq->ring->emit & (CACHELINE_BYTES - 1)) / sizeof(u32);
+	if (num_dwords == 0)
+		return 0;
+
+	num_dwords = CACHELINE_DWORDS - num_dwords;
+	GEM_BUG_ON(num_dwords & 1);
+
+	cs = intel_ring_begin(rq, num_dwords);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	memset64(cs, (u64)MI_NOOP << 32 | MI_NOOP, num_dwords / 2);
+	intel_ring_advance(rq, cs + num_dwords);
+
+	GEM_BUG_ON(rq->ring->emit & (CACHELINE_BYTES - 1));
+	return 0;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_ring.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_ring.h b/drivers/gpu/drm/i915/gt/intel_ring.h
new file mode 100644
index 000000000..1b32dadfb
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_ring.h
@@ -0,0 +1,142 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_RING_H
+#define INTEL_RING_H
+
+#include "i915_gem.h" /* GEM_BUG_ON */
+#include "i915_request.h"
+#include "intel_ring_types.h"
+
+struct intel_engine_cs;
+
+struct intel_ring *
+intel_engine_create_ring(struct intel_engine_cs *engine, int size);
+
+u32 *intel_ring_begin(struct i915_request *rq, unsigned int num_dwords);
+int intel_ring_cacheline_align(struct i915_request *rq);
+
+unsigned int intel_ring_update_space(struct intel_ring *ring);
+
+void __intel_ring_pin(struct intel_ring *ring);
+int intel_ring_pin(struct intel_ring *ring, struct i915_gem_ww_ctx *ww);
+void intel_ring_unpin(struct intel_ring *ring);
+void intel_ring_reset(struct intel_ring *ring, u32 tail);
+
+void intel_ring_free(struct kref *ref);
+
+static inline struct intel_ring *intel_ring_get(struct intel_ring *ring)
+{
+	kref_get(&ring->ref);
+	return ring;
+}
+
+static inline void intel_ring_put(struct intel_ring *ring)
+{
+	kref_put(&ring->ref, intel_ring_free);
+}
+
+static inline void intel_ring_advance(struct i915_request *rq, u32 *cs)
+{
+	/* Dummy function.
+	 *
+	 * This serves as a placeholder in the code so that the reader
+	 * can compare against the preceding intel_ring_begin() and
+	 * check that the number of dwords emitted matches the space
+	 * reserved for the command packet (i.e. the value passed to
+	 * intel_ring_begin()).
+	 */
+	GEM_BUG_ON((rq->ring->vaddr + rq->ring->emit) != cs);
+	GEM_BUG_ON(!IS_ALIGNED(rq->ring->emit, 8)); /* RING_TAIL qword align */
+}
+
+static inline u32 intel_ring_wrap(const struct intel_ring *ring, u32 pos)
+{
+	return pos & (ring->size - 1);
+}
+
+static inline int intel_ring_direction(const struct intel_ring *ring,
+				       u32 next, u32 prev)
+{
+	typecheck(typeof(ring->size), next);
+	typecheck(typeof(ring->size), prev);
+	return (next - prev) << ring->wrap;
+}
+
+static inline bool
+intel_ring_offset_valid(const struct intel_ring *ring,
+			unsigned int pos)
+{
+	if (pos & -ring->size) /* must be strictly within the ring */
+		return false;
+
+	if (!IS_ALIGNED(pos, 8)) /* must be qword aligned */
+		return false;
+
+	return true;
+}
+
+static inline u32 intel_ring_offset(const struct i915_request *rq, void *addr)
+{
+	/* Don't write ring->size (equivalent to 0) as that hangs some GPUs. */
+	u32 offset = addr - rq->ring->vaddr;
+
+	GEM_BUG_ON(offset > rq->ring->size);
+	return intel_ring_wrap(rq->ring, offset);
+}
+
+static inline void
+assert_ring_tail_valid(const struct intel_ring *ring, unsigned int tail)
+{
+	unsigned int head = READ_ONCE(ring->head);
+
+	GEM_BUG_ON(!intel_ring_offset_valid(ring, tail));
+
+	/*
+	 * "Ring Buffer Use"
+	 *	Gen2 BSpec "1. Programming Environment" / 1.4.4.6
+	 *	Gen3 BSpec "1c Memory Interface Functions" / 2.3.4.5
+	 *	Gen4+ BSpec "1c Memory Interface and Command Stream" / 5.3.4.5
+	 * "If the Ring Buffer Head Pointer and the Tail Pointer are on the
+	 * same cacheline, the Head Pointer must not be greater than the Tail
+	 * Pointer."
+	 *
+	 * We use ring->head as the last known location of the actual RING_HEAD,
+	 * it may have advanced but in the worst case it is equally the same
+	 * as ring->head and so we should never program RING_TAIL to advance
+	 * into the same cacheline as ring->head.
+	 */
+#define cacheline(a) round_down(a, CACHELINE_BYTES)
+	GEM_BUG_ON(cacheline(tail) == cacheline(head) && tail < head);
+#undef cacheline
+}
+
+static inline unsigned int
+intel_ring_set_tail(struct intel_ring *ring, unsigned int tail)
+{
+	/* Whilst writes to the tail are strictly order, there is no
+	 * serialisation between readers and the writers. The tail may be
+	 * read by i915_request_retire() just as it is being updated
+	 * by execlists, as although the breadcrumb is complete, the context
+	 * switch hasn't been seen.
+	 */
+	assert_ring_tail_valid(ring, tail);
+	ring->tail = tail;
+	return tail;
+}
+
+static inline unsigned int
+__intel_ring_space(unsigned int head, unsigned int tail, unsigned int size)
+{
+	/*
+	 * "If the Ring Buffer Head Pointer and the Tail Pointer are on the
+	 * same cacheline, the Head Pointer must not be greater than the Tail
+	 * Pointer."
+	 */
+	GEM_BUG_ON(!is_power_of_2(size));
+	return (head - tail - CACHELINE_BYTES) & (size - 1);
+}
+
+#endif /* INTEL_RING_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
new file mode 100644
index 000000000..d5d6f1fad
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
@@ -0,0 +1,1422 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2008-2021 Intel Corporation
+ */
+
+#include <drm/drm_cache.h>
+
+#include "gem/i915_gem_internal.h"
+
+#include "gen2_engine_cs.h"
+#include "gen6_engine_cs.h"
+#include "gen6_ppgtt.h"
+#include "gen7_renderclear.h"
+#include "i915_drv.h"
+#include "i915_mitigations.h"
+#include "intel_breadcrumbs.h"
+#include "intel_context.h"
+#include "intel_engine_regs.h"
+#include "intel_gt.h"
+#include "intel_gt_irq.h"
+#include "intel_gt_regs.h"
+#include "intel_reset.h"
+#include "intel_ring.h"
+#include "shmem_utils.h"
+#include "intel_engine_heartbeat.h"
+#include "intel_engine_pm.h"
+
+/* Rough estimate of the typical request size, performing a flush,
+ * set-context and then emitting the batch.
+ */
+#define LEGACY_REQUEST_SIZE 200
+
+static void set_hwstam(struct intel_engine_cs *engine, u32 mask)
+{
+	/*
+	 * Keep the render interrupt unmasked as this papers over
+	 * lost interrupts following a reset.
+	 */
+	if (engine->class == RENDER_CLASS) {
+		if (GRAPHICS_VER(engine->i915) >= 6)
+			mask &= ~BIT(0);
+		else
+			mask &= ~I915_USER_INTERRUPT;
+	}
+
+	intel_engine_set_hwsp_writemask(engine, mask);
+}
+
+static void set_hws_pga(struct intel_engine_cs *engine, phys_addr_t phys)
+{
+	u32 addr;
+
+	addr = lower_32_bits(phys);
+	if (GRAPHICS_VER(engine->i915) >= 4)
+		addr |= (phys >> 28) & 0xf0;
+
+	intel_uncore_write(engine->uncore, HWS_PGA, addr);
+}
+
+static struct page *status_page(struct intel_engine_cs *engine)
+{
+	struct drm_i915_gem_object *obj = engine->status_page.vma->obj;
+
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+	return sg_page(obj->mm.pages->sgl);
+}
+
+static void ring_setup_phys_status_page(struct intel_engine_cs *engine)
+{
+	set_hws_pga(engine, PFN_PHYS(page_to_pfn(status_page(engine))));
+	set_hwstam(engine, ~0u);
+}
+
+static void set_hwsp(struct intel_engine_cs *engine, u32 offset)
+{
+	i915_reg_t hwsp;
+
+	/*
+	 * The ring status page addresses are no longer next to the rest of
+	 * the ring registers as of gen7.
+	 */
+	if (GRAPHICS_VER(engine->i915) == 7) {
+		switch (engine->id) {
+		/*
+		 * No more rings exist on Gen7. Default case is only to shut up
+		 * gcc switch check warning.
+		 */
+		default:
+			GEM_BUG_ON(engine->id);
+			fallthrough;
+		case RCS0:
+			hwsp = RENDER_HWS_PGA_GEN7;
+			break;
+		case BCS0:
+			hwsp = BLT_HWS_PGA_GEN7;
+			break;
+		case VCS0:
+			hwsp = BSD_HWS_PGA_GEN7;
+			break;
+		case VECS0:
+			hwsp = VEBOX_HWS_PGA_GEN7;
+			break;
+		}
+	} else if (GRAPHICS_VER(engine->i915) == 6) {
+		hwsp = RING_HWS_PGA_GEN6(engine->mmio_base);
+	} else {
+		hwsp = RING_HWS_PGA(engine->mmio_base);
+	}
+
+	intel_uncore_write_fw(engine->uncore, hwsp, offset);
+	intel_uncore_posting_read_fw(engine->uncore, hwsp);
+}
+
+static void flush_cs_tlb(struct intel_engine_cs *engine)
+{
+	if (!IS_GRAPHICS_VER(engine->i915, 6, 7))
+		return;
+
+	/* ring should be idle before issuing a sync flush*/
+	if ((ENGINE_READ(engine, RING_MI_MODE) & MODE_IDLE) == 0)
+		drm_warn(&engine->i915->drm, "%s not idle before sync flush!\n",
+			 engine->name);
+
+	ENGINE_WRITE_FW(engine, RING_INSTPM,
+			_MASKED_BIT_ENABLE(INSTPM_TLB_INVALIDATE |
+					   INSTPM_SYNC_FLUSH));
+	if (__intel_wait_for_register_fw(engine->uncore,
+					 RING_INSTPM(engine->mmio_base),
+					 INSTPM_SYNC_FLUSH, 0,
+					 2000, 0, NULL))
+		ENGINE_TRACE(engine,
+			     "wait for SyncFlush to complete for TLB invalidation timed out\n");
+}
+
+static void ring_setup_status_page(struct intel_engine_cs *engine)
+{
+	set_hwsp(engine, i915_ggtt_offset(engine->status_page.vma));
+	set_hwstam(engine, ~0u);
+
+	flush_cs_tlb(engine);
+}
+
+static struct i915_address_space *vm_alias(struct i915_address_space *vm)
+{
+	if (i915_is_ggtt(vm))
+		vm = &i915_vm_to_ggtt(vm)->alias->vm;
+
+	return vm;
+}
+
+static u32 pp_dir(struct i915_address_space *vm)
+{
+	return to_gen6_ppgtt(i915_vm_to_ppgtt(vm))->pp_dir;
+}
+
+static void set_pp_dir(struct intel_engine_cs *engine)
+{
+	struct i915_address_space *vm = vm_alias(engine->gt->vm);
+
+	if (!vm)
+		return;
+
+	ENGINE_WRITE_FW(engine, RING_PP_DIR_DCLV, PP_DIR_DCLV_2G);
+	ENGINE_WRITE_FW(engine, RING_PP_DIR_BASE, pp_dir(vm));
+
+	if (GRAPHICS_VER(engine->i915) >= 7) {
+		ENGINE_WRITE_FW(engine,
+				RING_MODE_GEN7,
+				_MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE));
+	}
+}
+
+static bool stop_ring(struct intel_engine_cs *engine)
+{
+	/* Empty the ring by skipping to the end */
+	ENGINE_WRITE_FW(engine, RING_HEAD, ENGINE_READ_FW(engine, RING_TAIL));
+	ENGINE_POSTING_READ(engine, RING_HEAD);
+
+	/* The ring must be empty before it is disabled */
+	ENGINE_WRITE_FW(engine, RING_CTL, 0);
+	ENGINE_POSTING_READ(engine, RING_CTL);
+
+	/* Then reset the disabled ring */
+	ENGINE_WRITE_FW(engine, RING_HEAD, 0);
+	ENGINE_WRITE_FW(engine, RING_TAIL, 0);
+
+	return (ENGINE_READ_FW(engine, RING_HEAD) & HEAD_ADDR) == 0;
+}
+
+static int xcs_resume(struct intel_engine_cs *engine)
+{
+	struct intel_ring *ring = engine->legacy.ring;
+
+	ENGINE_TRACE(engine, "ring:{HEAD:%04x, TAIL:%04x}\n",
+		     ring->head, ring->tail);
+
+	/*
+	 * Double check the ring is empty & disabled before we resume. Called
+	 * from atomic context during PCI probe, so _hardirq().
+	 */
+	intel_synchronize_hardirq(engine->i915);
+	if (!stop_ring(engine))
+		goto err;
+
+	if (HWS_NEEDS_PHYSICAL(engine->i915))
+		ring_setup_phys_status_page(engine);
+	else
+		ring_setup_status_page(engine);
+
+	intel_breadcrumbs_reset(engine->breadcrumbs);
+
+	/* Enforce ordering by reading HEAD register back */
+	ENGINE_POSTING_READ(engine, RING_HEAD);
+
+	/*
+	 * Initialize the ring. This must happen _after_ we've cleared the ring
+	 * registers with the above sequence (the readback of the HEAD registers
+	 * also enforces ordering), otherwise the hw might lose the new ring
+	 * register values.
+	 */
+	ENGINE_WRITE_FW(engine, RING_START, i915_ggtt_offset(ring->vma));
+
+	/* Check that the ring offsets point within the ring! */
+	GEM_BUG_ON(!intel_ring_offset_valid(ring, ring->head));
+	GEM_BUG_ON(!intel_ring_offset_valid(ring, ring->tail));
+	intel_ring_update_space(ring);
+
+	set_pp_dir(engine);
+
+	/* First wake the ring up to an empty/idle ring */
+	ENGINE_WRITE_FW(engine, RING_HEAD, ring->head);
+	ENGINE_WRITE_FW(engine, RING_TAIL, ring->head);
+	ENGINE_POSTING_READ(engine, RING_TAIL);
+
+	ENGINE_WRITE_FW(engine, RING_CTL,
+			RING_CTL_SIZE(ring->size) | RING_VALID);
+
+	/* If the head is still not zero, the ring is dead */
+	if (__intel_wait_for_register_fw(engine->uncore,
+					 RING_CTL(engine->mmio_base),
+					 RING_VALID, RING_VALID,
+					 5000, 0, NULL))
+		goto err;
+
+	if (GRAPHICS_VER(engine->i915) > 2)
+		ENGINE_WRITE_FW(engine,
+				RING_MI_MODE, _MASKED_BIT_DISABLE(STOP_RING));
+
+	/* Now awake, let it get started */
+	if (ring->tail != ring->head) {
+		ENGINE_WRITE_FW(engine, RING_TAIL, ring->tail);
+		ENGINE_POSTING_READ(engine, RING_TAIL);
+	}
+
+	/* Papering over lost _interrupts_ immediately following the restart */
+	intel_engine_signal_breadcrumbs(engine);
+	return 0;
+
+err:
+	drm_err(&engine->i915->drm,
+		"%s initialization failed; "
+		"ctl %08x (valid? %d) head %08x [%08x] tail %08x [%08x] start %08x [expected %08x]\n",
+		engine->name,
+		ENGINE_READ(engine, RING_CTL),
+		ENGINE_READ(engine, RING_CTL) & RING_VALID,
+		ENGINE_READ(engine, RING_HEAD), ring->head,
+		ENGINE_READ(engine, RING_TAIL), ring->tail,
+		ENGINE_READ(engine, RING_START),
+		i915_ggtt_offset(ring->vma));
+	return -EIO;
+}
+
+static void sanitize_hwsp(struct intel_engine_cs *engine)
+{
+	struct intel_timeline *tl;
+
+	list_for_each_entry(tl, &engine->status_page.timelines, engine_link)
+		intel_timeline_reset_seqno(tl);
+}
+
+static void xcs_sanitize(struct intel_engine_cs *engine)
+{
+	/*
+	 * Poison residual state on resume, in case the suspend didn't!
+	 *
+	 * We have to assume that across suspend/resume (or other loss
+	 * of control) that the contents of our pinned buffers has been
+	 * lost, replaced by garbage. Since this doesn't always happen,
+	 * let's poison such state so that we more quickly spot when
+	 * we falsely assume it has been preserved.
+	 */
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		memset(engine->status_page.addr, POISON_INUSE, PAGE_SIZE);
+
+	/*
+	 * The kernel_context HWSP is stored in the status_page. As above,
+	 * that may be lost on resume/initialisation, and so we need to
+	 * reset the value in the HWSP.
+	 */
+	sanitize_hwsp(engine);
+
+	/* And scrub the dirty cachelines for the HWSP */
+	drm_clflush_virt_range(engine->status_page.addr, PAGE_SIZE);
+
+	intel_engine_reset_pinned_contexts(engine);
+}
+
+static void reset_prepare(struct intel_engine_cs *engine)
+{
+	/*
+	 * We stop engines, otherwise we might get failed reset and a
+	 * dead gpu (on elk). Also as modern gpu as kbl can suffer
+	 * from system hang if batchbuffer is progressing when
+	 * the reset is issued, regardless of READY_TO_RESET ack.
+	 * Thus assume it is best to stop engines on all gens
+	 * where we have a gpu reset.
+	 *
+	 * WaKBLVECSSemaphoreWaitPoll:kbl (on ALL_ENGINES)
+	 *
+	 * WaMediaResetMainRingCleanup:ctg,elk (presumably)
+	 * WaClearRingBufHeadRegAtInit:ctg,elk
+	 *
+	 * FIXME: Wa for more modern gens needs to be validated
+	 */
+	ENGINE_TRACE(engine, "\n");
+	intel_engine_stop_cs(engine);
+
+	if (!stop_ring(engine)) {
+		/* G45 ring initialization often fails to reset head to zero */
+		ENGINE_TRACE(engine,
+			     "HEAD not reset to zero, "
+			     "{ CTL:%08x, HEAD:%08x, TAIL:%08x, START:%08x }\n",
+			     ENGINE_READ_FW(engine, RING_CTL),
+			     ENGINE_READ_FW(engine, RING_HEAD),
+			     ENGINE_READ_FW(engine, RING_TAIL),
+			     ENGINE_READ_FW(engine, RING_START));
+		if (!stop_ring(engine)) {
+			drm_err(&engine->i915->drm,
+				"failed to set %s head to zero "
+				"ctl %08x head %08x tail %08x start %08x\n",
+				engine->name,
+				ENGINE_READ_FW(engine, RING_CTL),
+				ENGINE_READ_FW(engine, RING_HEAD),
+				ENGINE_READ_FW(engine, RING_TAIL),
+				ENGINE_READ_FW(engine, RING_START));
+		}
+	}
+}
+
+static void reset_rewind(struct intel_engine_cs *engine, bool stalled)
+{
+	struct i915_request *pos, *rq;
+	unsigned long flags;
+	u32 head;
+
+	rq = NULL;
+	spin_lock_irqsave(&engine->sched_engine->lock, flags);
+	rcu_read_lock();
+	list_for_each_entry(pos, &engine->sched_engine->requests, sched.link) {
+		if (!__i915_request_is_complete(pos)) {
+			rq = pos;
+			break;
+		}
+	}
+	rcu_read_unlock();
+
+	/*
+	 * The guilty request will get skipped on a hung engine.
+	 *
+	 * Users of client default contexts do not rely on logical
+	 * state preserved between batches so it is safe to execute
+	 * queued requests following the hang. Non default contexts
+	 * rely on preserved state, so skipping a batch loses the
+	 * evolution of the state and it needs to be considered corrupted.
+	 * Executing more queued batches on top of corrupted state is
+	 * risky. But we take the risk by trying to advance through
+	 * the queued requests in order to make the client behaviour
+	 * more predictable around resets, by not throwing away random
+	 * amount of batches it has prepared for execution. Sophisticated
+	 * clients can use gem_reset_stats_ioctl and dma fence status
+	 * (exported via sync_file info ioctl on explicit fences) to observe
+	 * when it loses the context state and should rebuild accordingly.
+	 *
+	 * The context ban, and ultimately the client ban, mechanism are safety
+	 * valves if client submission ends up resulting in nothing more than
+	 * subsequent hangs.
+	 */
+
+	if (rq) {
+		/*
+		 * Try to restore the logical GPU state to match the
+		 * continuation of the request queue. If we skip the
+		 * context/PD restore, then the next request may try to execute
+		 * assuming that its context is valid and loaded on the GPU and
+		 * so may try to access invalid memory, prompting repeated GPU
+		 * hangs.
+		 *
+		 * If the request was guilty, we still restore the logical
+		 * state in case the next request requires it (e.g. the
+		 * aliasing ppgtt), but skip over the hung batch.
+		 *
+		 * If the request was innocent, we try to replay the request
+		 * with the restored context.
+		 */
+		__i915_request_reset(rq, stalled);
+
+		GEM_BUG_ON(rq->ring != engine->legacy.ring);
+		head = rq->head;
+	} else {
+		head = engine->legacy.ring->tail;
+	}
+	engine->legacy.ring->head = intel_ring_wrap(engine->legacy.ring, head);
+
+	spin_unlock_irqrestore(&engine->sched_engine->lock, flags);
+}
+
+static void reset_finish(struct intel_engine_cs *engine)
+{
+}
+
+static void reset_cancel(struct intel_engine_cs *engine)
+{
+	struct i915_request *request;
+	unsigned long flags;
+
+	spin_lock_irqsave(&engine->sched_engine->lock, flags);
+
+	/* Mark all submitted requests as skipped. */
+	list_for_each_entry(request, &engine->sched_engine->requests, sched.link)
+		i915_request_put(i915_request_mark_eio(request));
+	intel_engine_signal_breadcrumbs(engine);
+
+	/* Remaining _unready_ requests will be nop'ed when submitted */
+
+	spin_unlock_irqrestore(&engine->sched_engine->lock, flags);
+}
+
+static void i9xx_submit_request(struct i915_request *request)
+{
+	i915_request_submit(request);
+	wmb(); /* paranoid flush writes out of the WCB before mmio */
+
+	ENGINE_WRITE(request->engine, RING_TAIL,
+		     intel_ring_set_tail(request->ring, request->tail));
+}
+
+static void __ring_context_fini(struct intel_context *ce)
+{
+	i915_vma_put(ce->state);
+}
+
+static void ring_context_destroy(struct kref *ref)
+{
+	struct intel_context *ce = container_of(ref, typeof(*ce), ref);
+
+	GEM_BUG_ON(intel_context_is_pinned(ce));
+
+	if (ce->state)
+		__ring_context_fini(ce);
+
+	intel_context_fini(ce);
+	intel_context_free(ce);
+}
+
+static int ring_context_init_default_state(struct intel_context *ce,
+					   struct i915_gem_ww_ctx *ww)
+{
+	struct drm_i915_gem_object *obj = ce->state->obj;
+	void *vaddr;
+
+	vaddr = i915_gem_object_pin_map(obj, I915_MAP_WB);
+	if (IS_ERR(vaddr))
+		return PTR_ERR(vaddr);
+
+	shmem_read(ce->engine->default_state, 0,
+		   vaddr, ce->engine->context_size);
+
+	i915_gem_object_flush_map(obj);
+	__i915_gem_object_release_map(obj);
+
+	__set_bit(CONTEXT_VALID_BIT, &ce->flags);
+	return 0;
+}
+
+static int ring_context_pre_pin(struct intel_context *ce,
+				struct i915_gem_ww_ctx *ww,
+				void **unused)
+{
+	struct i915_address_space *vm;
+	int err = 0;
+
+	if (ce->engine->default_state &&
+	    !test_bit(CONTEXT_VALID_BIT, &ce->flags)) {
+		err = ring_context_init_default_state(ce, ww);
+		if (err)
+			return err;
+	}
+
+	vm = vm_alias(ce->vm);
+	if (vm)
+		err = gen6_ppgtt_pin(i915_vm_to_ppgtt((vm)), ww);
+
+	return err;
+}
+
+static void __context_unpin_ppgtt(struct intel_context *ce)
+{
+	struct i915_address_space *vm;
+
+	vm = vm_alias(ce->vm);
+	if (vm)
+		gen6_ppgtt_unpin(i915_vm_to_ppgtt(vm));
+}
+
+static void ring_context_unpin(struct intel_context *ce)
+{
+}
+
+static void ring_context_post_unpin(struct intel_context *ce)
+{
+	__context_unpin_ppgtt(ce);
+}
+
+static struct i915_vma *
+alloc_context_vma(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err;
+
+	obj = i915_gem_object_create_shmem(i915, engine->context_size);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	/*
+	 * Try to make the context utilize L3 as well as LLC.
+	 *
+	 * On VLV we don't have L3 controls in the PTEs so we
+	 * shouldn't touch the cache level, especially as that
+	 * would make the object snooped which might have a
+	 * negative performance impact.
+	 *
+	 * Snooping is required on non-llc platforms in execlist
+	 * mode, but since all GGTT accesses use PAT entry 0 we
+	 * get snooping anyway regardless of cache_level.
+	 *
+	 * This is only applicable for Ivy Bridge devices since
+	 * later platforms don't have L3 control bits in the PTE.
+	 */
+	if (IS_IVYBRIDGE(i915))
+		i915_gem_object_set_cache_coherency(obj, I915_CACHE_L3_LLC);
+
+	vma = i915_vma_instance(obj, &engine->gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	return vma;
+
+err_obj:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static int ring_context_alloc(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = ce->engine;
+
+	/* One ringbuffer to rule them all */
+	GEM_BUG_ON(!engine->legacy.ring);
+	ce->ring = engine->legacy.ring;
+	ce->timeline = intel_timeline_get(engine->legacy.timeline);
+
+	GEM_BUG_ON(ce->state);
+	if (engine->context_size) {
+		struct i915_vma *vma;
+
+		vma = alloc_context_vma(engine);
+		if (IS_ERR(vma))
+			return PTR_ERR(vma);
+
+		ce->state = vma;
+	}
+
+	return 0;
+}
+
+static int ring_context_pin(struct intel_context *ce, void *unused)
+{
+	return 0;
+}
+
+static void ring_context_reset(struct intel_context *ce)
+{
+	intel_ring_reset(ce->ring, ce->ring->emit);
+	clear_bit(CONTEXT_VALID_BIT, &ce->flags);
+}
+
+static void ring_context_revoke(struct intel_context *ce,
+				struct i915_request *rq,
+				unsigned int preempt_timeout_ms)
+{
+	struct intel_engine_cs *engine;
+
+	if (!rq || !i915_request_is_active(rq))
+		return;
+
+	engine = rq->engine;
+	lockdep_assert_held(&engine->sched_engine->lock);
+	list_for_each_entry_continue(rq, &engine->sched_engine->requests,
+				     sched.link)
+		if (rq->context == ce) {
+			i915_request_set_error_once(rq, -EIO);
+			__i915_request_skip(rq);
+		}
+}
+
+static void ring_context_cancel_request(struct intel_context *ce,
+					struct i915_request *rq)
+{
+	struct intel_engine_cs *engine = NULL;
+
+	i915_request_active_engine(rq, &engine);
+
+	if (engine && intel_engine_pulse(engine))
+		intel_gt_handle_error(engine->gt, engine->mask, 0,
+				      "request cancellation by %s",
+				      current->comm);
+}
+
+static const struct intel_context_ops ring_context_ops = {
+	.alloc = ring_context_alloc,
+
+	.cancel_request = ring_context_cancel_request,
+
+	.revoke = ring_context_revoke,
+
+	.pre_pin = ring_context_pre_pin,
+	.pin = ring_context_pin,
+	.unpin = ring_context_unpin,
+	.post_unpin = ring_context_post_unpin,
+
+	.enter = intel_context_enter_engine,
+	.exit = intel_context_exit_engine,
+
+	.reset = ring_context_reset,
+	.destroy = ring_context_destroy,
+};
+
+static int load_pd_dir(struct i915_request *rq,
+		       struct i915_address_space *vm,
+		       u32 valid)
+{
+	const struct intel_engine_cs * const engine = rq->engine;
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 12);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(RING_PP_DIR_DCLV(engine->mmio_base));
+	*cs++ = valid;
+
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(RING_PP_DIR_BASE(engine->mmio_base));
+	*cs++ = pp_dir(vm);
+
+	/* Stall until the page table load is complete? */
+	*cs++ = MI_STORE_REGISTER_MEM | MI_SRM_LRM_GLOBAL_GTT;
+	*cs++ = i915_mmio_reg_offset(RING_PP_DIR_BASE(engine->mmio_base));
+	*cs++ = intel_gt_scratch_offset(engine->gt,
+					INTEL_GT_SCRATCH_FIELD_DEFAULT);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(RING_INSTPM(engine->mmio_base));
+	*cs++ = _MASKED_BIT_ENABLE(INSTPM_TLB_INVALIDATE);
+
+	intel_ring_advance(rq, cs);
+
+	return rq->engine->emit_flush(rq, EMIT_FLUSH);
+}
+
+static int mi_set_context(struct i915_request *rq,
+			  struct intel_context *ce,
+			  u32 flags)
+{
+	struct intel_engine_cs *engine = rq->engine;
+	struct drm_i915_private *i915 = engine->i915;
+	enum intel_engine_id id;
+	const int num_engines =
+		IS_HASWELL(i915) ? engine->gt->info.num_engines - 1 : 0;
+	bool force_restore = false;
+	int len;
+	u32 *cs;
+
+	len = 4;
+	if (GRAPHICS_VER(i915) == 7)
+		len += 2 + (num_engines ? 4 * num_engines + 6 : 0);
+	else if (GRAPHICS_VER(i915) == 5)
+		len += 2;
+	if (flags & MI_FORCE_RESTORE) {
+		GEM_BUG_ON(flags & MI_RESTORE_INHIBIT);
+		flags &= ~MI_FORCE_RESTORE;
+		force_restore = true;
+		len += 2;
+	}
+
+	cs = intel_ring_begin(rq, len);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/* WaProgramMiArbOnOffAroundMiSetContext:ivb,vlv,hsw,bdw,chv */
+	if (GRAPHICS_VER(i915) == 7) {
+		*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+		if (num_engines) {
+			struct intel_engine_cs *signaller;
+
+			*cs++ = MI_LOAD_REGISTER_IMM(num_engines);
+			for_each_engine(signaller, engine->gt, id) {
+				if (signaller == engine)
+					continue;
+
+				*cs++ = i915_mmio_reg_offset(
+					   RING_PSMI_CTL(signaller->mmio_base));
+				*cs++ = _MASKED_BIT_ENABLE(
+						GEN6_PSMI_SLEEP_MSG_DISABLE);
+			}
+		}
+	} else if (GRAPHICS_VER(i915) == 5) {
+		/*
+		 * This w/a is only listed for pre-production ilk a/b steppings,
+		 * but is also mentioned for programming the powerctx. To be
+		 * safe, just apply the workaround; we do not use SyncFlush so
+		 * this should never take effect and so be a no-op!
+		 */
+		*cs++ = MI_SUSPEND_FLUSH | MI_SUSPEND_FLUSH_EN;
+	}
+
+	if (force_restore) {
+		/*
+		 * The HW doesn't handle being told to restore the current
+		 * context very well. Quite often it likes goes to go off and
+		 * sulk, especially when it is meant to be reloading PP_DIR.
+		 * A very simple fix to force the reload is to simply switch
+		 * away from the current context and back again.
+		 *
+		 * Note that the kernel_context will contain random state
+		 * following the INHIBIT_RESTORE. We accept this since we
+		 * never use the kernel_context state; it is merely a
+		 * placeholder we use to flush other contexts.
+		 */
+		*cs++ = MI_SET_CONTEXT;
+		*cs++ = i915_ggtt_offset(engine->kernel_context->state) |
+			MI_MM_SPACE_GTT |
+			MI_RESTORE_INHIBIT;
+	}
+
+	*cs++ = MI_NOOP;
+	*cs++ = MI_SET_CONTEXT;
+	*cs++ = i915_ggtt_offset(ce->state) | flags;
+	/*
+	 * w/a: MI_SET_CONTEXT must always be followed by MI_NOOP
+	 * WaMiSetContext_Hang:snb,ivb,vlv
+	 */
+	*cs++ = MI_NOOP;
+
+	if (GRAPHICS_VER(i915) == 7) {
+		if (num_engines) {
+			struct intel_engine_cs *signaller;
+			i915_reg_t last_reg = INVALID_MMIO_REG; /* keep gcc quiet */
+
+			*cs++ = MI_LOAD_REGISTER_IMM(num_engines);
+			for_each_engine(signaller, engine->gt, id) {
+				if (signaller == engine)
+					continue;
+
+				last_reg = RING_PSMI_CTL(signaller->mmio_base);
+				*cs++ = i915_mmio_reg_offset(last_reg);
+				*cs++ = _MASKED_BIT_DISABLE(
+						GEN6_PSMI_SLEEP_MSG_DISABLE);
+			}
+
+			/* Insert a delay before the next switch! */
+			*cs++ = MI_STORE_REGISTER_MEM | MI_SRM_LRM_GLOBAL_GTT;
+			*cs++ = i915_mmio_reg_offset(last_reg);
+			*cs++ = intel_gt_scratch_offset(engine->gt,
+							INTEL_GT_SCRATCH_FIELD_DEFAULT);
+			*cs++ = MI_NOOP;
+		}
+		*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	} else if (GRAPHICS_VER(i915) == 5) {
+		*cs++ = MI_SUSPEND_FLUSH;
+	}
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int remap_l3_slice(struct i915_request *rq, int slice)
+{
+#define L3LOG_DW (GEN7_L3LOG_SIZE / sizeof(u32))
+	u32 *cs, *remap_info = rq->engine->i915->l3_parity.remap_info[slice];
+	int i;
+
+	if (!remap_info)
+		return 0;
+
+	cs = intel_ring_begin(rq, L3LOG_DW * 2 + 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/*
+	 * Note: We do not worry about the concurrent register cacheline hang
+	 * here because no other code should access these registers other than
+	 * at initialization time.
+	 */
+	*cs++ = MI_LOAD_REGISTER_IMM(L3LOG_DW);
+	for (i = 0; i < L3LOG_DW; i++) {
+		*cs++ = i915_mmio_reg_offset(GEN7_L3LOG(slice, i));
+		*cs++ = remap_info[i];
+	}
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	return 0;
+#undef L3LOG_DW
+}
+
+static int remap_l3(struct i915_request *rq)
+{
+	struct i915_gem_context *ctx = i915_request_gem_context(rq);
+	int i, err;
+
+	if (!ctx || !ctx->remap_slice)
+		return 0;
+
+	for (i = 0; i < MAX_L3_SLICES; i++) {
+		if (!(ctx->remap_slice & BIT(i)))
+			continue;
+
+		err = remap_l3_slice(rq, i);
+		if (err)
+			return err;
+	}
+
+	ctx->remap_slice = 0;
+	return 0;
+}
+
+static int switch_mm(struct i915_request *rq, struct i915_address_space *vm)
+{
+	int ret;
+
+	if (!vm)
+		return 0;
+
+	ret = rq->engine->emit_flush(rq, EMIT_FLUSH);
+	if (ret)
+		return ret;
+
+	/*
+	 * Not only do we need a full barrier (post-sync write) after
+	 * invalidating the TLBs, but we need to wait a little bit
+	 * longer. Whether this is merely delaying us, or the
+	 * subsequent flush is a key part of serialising with the
+	 * post-sync op, this extra pass appears vital before a
+	 * mm switch!
+	 */
+	ret = load_pd_dir(rq, vm, PP_DIR_DCLV_2G);
+	if (ret)
+		return ret;
+
+	return rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+}
+
+static int clear_residuals(struct i915_request *rq)
+{
+	struct intel_engine_cs *engine = rq->engine;
+	int ret;
+
+	ret = switch_mm(rq, vm_alias(engine->kernel_context->vm));
+	if (ret)
+		return ret;
+
+	if (engine->kernel_context->state) {
+		ret = mi_set_context(rq,
+				     engine->kernel_context,
+				     MI_MM_SPACE_GTT | MI_RESTORE_INHIBIT);
+		if (ret)
+			return ret;
+	}
+
+	ret = engine->emit_bb_start(rq,
+				    engine->wa_ctx.vma->node.start, 0,
+				    0);
+	if (ret)
+		return ret;
+
+	ret = engine->emit_flush(rq, EMIT_FLUSH);
+	if (ret)
+		return ret;
+
+	/* Always invalidate before the next switch_mm() */
+	return engine->emit_flush(rq, EMIT_INVALIDATE);
+}
+
+static int switch_context(struct i915_request *rq)
+{
+	struct intel_engine_cs *engine = rq->engine;
+	struct intel_context *ce = rq->context;
+	void **residuals = NULL;
+	int ret;
+
+	GEM_BUG_ON(HAS_EXECLISTS(engine->i915));
+
+	if (engine->wa_ctx.vma && ce != engine->kernel_context) {
+		if (engine->wa_ctx.vma->private != ce &&
+		    i915_mitigate_clear_residuals()) {
+			ret = clear_residuals(rq);
+			if (ret)
+				return ret;
+
+			residuals = &engine->wa_ctx.vma->private;
+		}
+	}
+
+	ret = switch_mm(rq, vm_alias(ce->vm));
+	if (ret)
+		return ret;
+
+	if (ce->state) {
+		u32 flags;
+
+		GEM_BUG_ON(engine->id != RCS0);
+
+		/* For resource streamer on HSW+ and power context elsewhere */
+		BUILD_BUG_ON(HSW_MI_RS_SAVE_STATE_EN != MI_SAVE_EXT_STATE_EN);
+		BUILD_BUG_ON(HSW_MI_RS_RESTORE_STATE_EN != MI_RESTORE_EXT_STATE_EN);
+
+		flags = MI_SAVE_EXT_STATE_EN | MI_MM_SPACE_GTT;
+		if (test_bit(CONTEXT_VALID_BIT, &ce->flags))
+			flags |= MI_RESTORE_EXT_STATE_EN;
+		else
+			flags |= MI_RESTORE_INHIBIT;
+
+		ret = mi_set_context(rq, ce, flags);
+		if (ret)
+			return ret;
+	}
+
+	ret = remap_l3(rq);
+	if (ret)
+		return ret;
+
+	/*
+	 * Now past the point of no return, this request _will_ be emitted.
+	 *
+	 * Or at least this preamble will be emitted, the request may be
+	 * interrupted prior to submitting the user payload. If so, we
+	 * still submit the "empty" request in order to preserve global
+	 * state tracking such as this, our tracking of the current
+	 * dirty context.
+	 */
+	if (residuals) {
+		intel_context_put(*residuals);
+		*residuals = intel_context_get(ce);
+	}
+
+	return 0;
+}
+
+static int ring_request_alloc(struct i915_request *request)
+{
+	int ret;
+
+	GEM_BUG_ON(!intel_context_is_pinned(request->context));
+	GEM_BUG_ON(i915_request_timeline(request)->has_initial_breadcrumb);
+
+	/*
+	 * Flush enough space to reduce the likelihood of waiting after
+	 * we start building the request - in which case we will just
+	 * have to repeat work.
+	 */
+	request->reserved_space += LEGACY_REQUEST_SIZE;
+
+	/* Unconditionally invalidate GPU caches and TLBs. */
+	ret = request->engine->emit_flush(request, EMIT_INVALIDATE);
+	if (ret)
+		return ret;
+
+	ret = switch_context(request);
+	if (ret)
+		return ret;
+
+	request->reserved_space -= LEGACY_REQUEST_SIZE;
+	return 0;
+}
+
+static void gen6_bsd_submit_request(struct i915_request *request)
+{
+	struct intel_uncore *uncore = request->engine->uncore;
+
+	intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+
+       /* Every tail move must follow the sequence below */
+
+	/* Disable notification that the ring is IDLE. The GT
+	 * will then assume that it is busy and bring it out of rc6.
+	 */
+	intel_uncore_write_fw(uncore, RING_PSMI_CTL(GEN6_BSD_RING_BASE),
+			      _MASKED_BIT_ENABLE(GEN6_PSMI_SLEEP_MSG_DISABLE));
+
+	/* Clear the context id. Here be magic! */
+	intel_uncore_write64_fw(uncore, GEN6_BSD_RNCID, 0x0);
+
+	/* Wait for the ring not to be idle, i.e. for it to wake up. */
+	if (__intel_wait_for_register_fw(uncore,
+					 RING_PSMI_CTL(GEN6_BSD_RING_BASE),
+					 GEN6_BSD_SLEEP_INDICATOR,
+					 0,
+					 1000, 0, NULL))
+		drm_err(&uncore->i915->drm,
+			"timed out waiting for the BSD ring to wake up\n");
+
+	/* Now that the ring is fully powered up, update the tail */
+	i9xx_submit_request(request);
+
+	/* Let the ring send IDLE messages to the GT again,
+	 * and so let it sleep to conserve power when idle.
+	 */
+	intel_uncore_write_fw(uncore, RING_PSMI_CTL(GEN6_BSD_RING_BASE),
+			      _MASKED_BIT_DISABLE(GEN6_PSMI_SLEEP_MSG_DISABLE));
+
+	intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
+}
+
+static void i9xx_set_default_submission(struct intel_engine_cs *engine)
+{
+	engine->submit_request = i9xx_submit_request;
+}
+
+static void gen6_bsd_set_default_submission(struct intel_engine_cs *engine)
+{
+	engine->submit_request = gen6_bsd_submit_request;
+}
+
+static void ring_release(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+
+	drm_WARN_ON(&dev_priv->drm, GRAPHICS_VER(dev_priv) > 2 &&
+		    (ENGINE_READ(engine, RING_MI_MODE) & MODE_IDLE) == 0);
+
+	intel_engine_cleanup_common(engine);
+
+	if (engine->wa_ctx.vma) {
+		intel_context_put(engine->wa_ctx.vma->private);
+		i915_vma_unpin_and_release(&engine->wa_ctx.vma, 0);
+	}
+
+	intel_ring_unpin(engine->legacy.ring);
+	intel_ring_put(engine->legacy.ring);
+
+	intel_timeline_unpin(engine->legacy.timeline);
+	intel_timeline_put(engine->legacy.timeline);
+}
+
+static void irq_handler(struct intel_engine_cs *engine, u16 iir)
+{
+	intel_engine_signal_breadcrumbs(engine);
+}
+
+static void setup_irq(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	intel_engine_set_irq_handler(engine, irq_handler);
+
+	if (GRAPHICS_VER(i915) >= 6) {
+		engine->irq_enable = gen6_irq_enable;
+		engine->irq_disable = gen6_irq_disable;
+	} else if (GRAPHICS_VER(i915) >= 5) {
+		engine->irq_enable = gen5_irq_enable;
+		engine->irq_disable = gen5_irq_disable;
+	} else if (GRAPHICS_VER(i915) >= 3) {
+		engine->irq_enable = gen3_irq_enable;
+		engine->irq_disable = gen3_irq_disable;
+	} else {
+		engine->irq_enable = gen2_irq_enable;
+		engine->irq_disable = gen2_irq_disable;
+	}
+}
+
+static void add_to_engine(struct i915_request *rq)
+{
+	lockdep_assert_held(&rq->engine->sched_engine->lock);
+	list_move_tail(&rq->sched.link, &rq->engine->sched_engine->requests);
+}
+
+static void remove_from_engine(struct i915_request *rq)
+{
+	spin_lock_irq(&rq->engine->sched_engine->lock);
+	list_del_init(&rq->sched.link);
+
+	/* Prevent further __await_execution() registering a cb, then flush */
+	set_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags);
+
+	spin_unlock_irq(&rq->engine->sched_engine->lock);
+
+	i915_request_notify_execute_cb_imm(rq);
+}
+
+static void setup_common(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	/* gen8+ are only supported with execlists */
+	GEM_BUG_ON(GRAPHICS_VER(i915) >= 8);
+
+	setup_irq(engine);
+
+	engine->resume = xcs_resume;
+	engine->sanitize = xcs_sanitize;
+
+	engine->reset.prepare = reset_prepare;
+	engine->reset.rewind = reset_rewind;
+	engine->reset.cancel = reset_cancel;
+	engine->reset.finish = reset_finish;
+
+	engine->add_active_request = add_to_engine;
+	engine->remove_active_request = remove_from_engine;
+
+	engine->cops = &ring_context_ops;
+	engine->request_alloc = ring_request_alloc;
+
+	/*
+	 * Using a global execution timeline; the previous final breadcrumb is
+	 * equivalent to our next initial bread so we can elide
+	 * engine->emit_init_breadcrumb().
+	 */
+	engine->emit_fini_breadcrumb = gen3_emit_breadcrumb;
+	if (GRAPHICS_VER(i915) == 5)
+		engine->emit_fini_breadcrumb = gen5_emit_breadcrumb;
+
+	engine->set_default_submission = i9xx_set_default_submission;
+
+	if (GRAPHICS_VER(i915) >= 6)
+		engine->emit_bb_start = gen6_emit_bb_start;
+	else if (GRAPHICS_VER(i915) >= 4)
+		engine->emit_bb_start = gen4_emit_bb_start;
+	else if (IS_I830(i915) || IS_I845G(i915))
+		engine->emit_bb_start = i830_emit_bb_start;
+	else
+		engine->emit_bb_start = gen3_emit_bb_start;
+}
+
+static void setup_rcs(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	if (HAS_L3_DPF(i915))
+		engine->irq_keep_mask = GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
+
+	engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT;
+
+	if (GRAPHICS_VER(i915) >= 7) {
+		engine->emit_flush = gen7_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen7_emit_breadcrumb_rcs;
+	} else if (GRAPHICS_VER(i915) == 6) {
+		engine->emit_flush = gen6_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen6_emit_breadcrumb_rcs;
+	} else if (GRAPHICS_VER(i915) == 5) {
+		engine->emit_flush = gen4_emit_flush_rcs;
+	} else {
+		if (GRAPHICS_VER(i915) < 4)
+			engine->emit_flush = gen2_emit_flush;
+		else
+			engine->emit_flush = gen4_emit_flush_rcs;
+		engine->irq_enable_mask = I915_USER_INTERRUPT;
+	}
+
+	if (IS_HASWELL(i915))
+		engine->emit_bb_start = hsw_emit_bb_start;
+}
+
+static void setup_vcs(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	if (GRAPHICS_VER(i915) >= 6) {
+		/* gen6 bsd needs a special wa for tail updates */
+		if (GRAPHICS_VER(i915) == 6)
+			engine->set_default_submission = gen6_bsd_set_default_submission;
+		engine->emit_flush = gen6_emit_flush_vcs;
+		engine->irq_enable_mask = GT_BSD_USER_INTERRUPT;
+
+		if (GRAPHICS_VER(i915) == 6)
+			engine->emit_fini_breadcrumb = gen6_emit_breadcrumb_xcs;
+		else
+			engine->emit_fini_breadcrumb = gen7_emit_breadcrumb_xcs;
+	} else {
+		engine->emit_flush = gen4_emit_flush_vcs;
+		if (GRAPHICS_VER(i915) == 5)
+			engine->irq_enable_mask = ILK_BSD_USER_INTERRUPT;
+		else
+			engine->irq_enable_mask = I915_BSD_USER_INTERRUPT;
+	}
+}
+
+static void setup_bcs(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	engine->emit_flush = gen6_emit_flush_xcs;
+	engine->irq_enable_mask = GT_BLT_USER_INTERRUPT;
+
+	if (GRAPHICS_VER(i915) == 6)
+		engine->emit_fini_breadcrumb = gen6_emit_breadcrumb_xcs;
+	else
+		engine->emit_fini_breadcrumb = gen7_emit_breadcrumb_xcs;
+}
+
+static void setup_vecs(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	GEM_BUG_ON(GRAPHICS_VER(i915) < 7);
+
+	engine->emit_flush = gen6_emit_flush_xcs;
+	engine->irq_enable_mask = PM_VEBOX_USER_INTERRUPT;
+	engine->irq_enable = hsw_irq_enable_vecs;
+	engine->irq_disable = hsw_irq_disable_vecs;
+
+	engine->emit_fini_breadcrumb = gen7_emit_breadcrumb_xcs;
+}
+
+static int gen7_ctx_switch_bb_setup(struct intel_engine_cs * const engine,
+				    struct i915_vma * const vma)
+{
+	return gen7_setup_clear_gpr_bb(engine, vma);
+}
+
+static int gen7_ctx_switch_bb_init(struct intel_engine_cs *engine,
+				   struct i915_gem_ww_ctx *ww,
+				   struct i915_vma *vma)
+{
+	int err;
+
+	err = i915_vma_pin_ww(vma, ww, 0, 0, PIN_USER | PIN_HIGH);
+	if (err)
+		return err;
+
+	err = i915_vma_sync(vma);
+	if (err)
+		goto err_unpin;
+
+	err = gen7_ctx_switch_bb_setup(engine, vma);
+	if (err)
+		goto err_unpin;
+
+	engine->wa_ctx.vma = vma;
+	return 0;
+
+err_unpin:
+	i915_vma_unpin(vma);
+	return err;
+}
+
+static struct i915_vma *gen7_ctx_vma(struct intel_engine_cs *engine)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int size, err;
+
+	if (GRAPHICS_VER(engine->i915) != 7 || engine->class != RENDER_CLASS)
+		return NULL;
+
+	err = gen7_ctx_switch_bb_setup(engine, NULL /* probe size */);
+	if (err < 0)
+		return ERR_PTR(err);
+	if (!err)
+		return NULL;
+
+	size = ALIGN(err, PAGE_SIZE);
+
+	obj = i915_gem_object_create_internal(engine->i915, size);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, engine->gt->vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return ERR_CAST(vma);
+	}
+
+	vma->private = intel_context_create(engine); /* dummy residuals */
+	if (IS_ERR(vma->private)) {
+		err = PTR_ERR(vma->private);
+		vma->private = NULL;
+		i915_gem_object_put(obj);
+		return ERR_PTR(err);
+	}
+
+	return vma;
+}
+
+int intel_ring_submission_setup(struct intel_engine_cs *engine)
+{
+	struct i915_gem_ww_ctx ww;
+	struct intel_timeline *timeline;
+	struct intel_ring *ring;
+	struct i915_vma *gen7_wa_vma;
+	int err;
+
+	setup_common(engine);
+
+	switch (engine->class) {
+	case RENDER_CLASS:
+		setup_rcs(engine);
+		break;
+	case VIDEO_DECODE_CLASS:
+		setup_vcs(engine);
+		break;
+	case COPY_ENGINE_CLASS:
+		setup_bcs(engine);
+		break;
+	case VIDEO_ENHANCEMENT_CLASS:
+		setup_vecs(engine);
+		break;
+	default:
+		MISSING_CASE(engine->class);
+		return -ENODEV;
+	}
+
+	timeline = intel_timeline_create_from_engine(engine,
+						     I915_GEM_HWS_SEQNO_ADDR);
+	if (IS_ERR(timeline)) {
+		err = PTR_ERR(timeline);
+		goto err;
+	}
+	GEM_BUG_ON(timeline->has_initial_breadcrumb);
+
+	ring = intel_engine_create_ring(engine, SZ_16K);
+	if (IS_ERR(ring)) {
+		err = PTR_ERR(ring);
+		goto err_timeline;
+	}
+
+	GEM_BUG_ON(engine->legacy.ring);
+	engine->legacy.ring = ring;
+	engine->legacy.timeline = timeline;
+
+	gen7_wa_vma = gen7_ctx_vma(engine);
+	if (IS_ERR(gen7_wa_vma)) {
+		err = PTR_ERR(gen7_wa_vma);
+		goto err_ring;
+	}
+
+	i915_gem_ww_ctx_init(&ww, false);
+
+retry:
+	err = i915_gem_object_lock(timeline->hwsp_ggtt->obj, &ww);
+	if (!err && gen7_wa_vma)
+		err = i915_gem_object_lock(gen7_wa_vma->obj, &ww);
+	if (!err)
+		err = i915_gem_object_lock(engine->legacy.ring->vma->obj, &ww);
+	if (!err)
+		err = intel_timeline_pin(timeline, &ww);
+	if (!err) {
+		err = intel_ring_pin(ring, &ww);
+		if (err)
+			intel_timeline_unpin(timeline);
+	}
+	if (err)
+		goto out;
+
+	GEM_BUG_ON(timeline->hwsp_ggtt != engine->status_page.vma);
+
+	if (gen7_wa_vma) {
+		err = gen7_ctx_switch_bb_init(engine, &ww, gen7_wa_vma);
+		if (err) {
+			intel_ring_unpin(ring);
+			intel_timeline_unpin(timeline);
+		}
+	}
+
+out:
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	if (err)
+		goto err_gen7_put;
+
+	/* Finally, take ownership and responsibility for cleanup! */
+	engine->release = ring_release;
+
+	return 0;
+
+err_gen7_put:
+	if (gen7_wa_vma) {
+		intel_context_put(gen7_wa_vma->private);
+		i915_gem_object_put(gen7_wa_vma->obj);
+	}
+err_ring:
+	intel_ring_put(ring);
+err_timeline:
+	intel_timeline_put(timeline);
+err:
+	intel_engine_cleanup_common(engine);
+	return err;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_ring_submission.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_ring_types.h b/drivers/gpu/drm/i915/gt/intel_ring_types.h
new file mode 100644
index 000000000..49ccb76dd
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_ring_types.h
@@ -0,0 +1,51 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_RING_TYPES_H
+#define INTEL_RING_TYPES_H
+
+#include <linux/atomic.h>
+#include <linux/kref.h>
+#include <linux/types.h>
+
+/*
+ * Early gen2 devices have a cacheline of just 32 bytes, using 64 is overkill,
+ * but keeps the logic simple. Indeed, the whole purpose of this macro is just
+ * to give some inclination as to some of the magic values used in the various
+ * workarounds!
+ */
+#define CACHELINE_BYTES 64
+#define CACHELINE_DWORDS (CACHELINE_BYTES / sizeof(u32))
+
+struct i915_vma;
+
+struct intel_ring {
+	struct kref ref;
+	struct i915_vma *vma;
+	void *vaddr;
+
+	/*
+	 * As we have two types of rings, one global to the engine used
+	 * by ringbuffer submission and those that are exclusive to a
+	 * context used by execlists, we have to play safe and allow
+	 * atomic updates to the pin_count. However, the actual pinning
+	 * of the context is either done during initialisation for
+	 * ringbuffer submission or serialised as part of the context
+	 * pinning for execlists, and so we do not need a mutex ourselves
+	 * to serialise intel_ring_pin/intel_ring_unpin.
+	 */
+	atomic_t pin_count;
+
+	u32 head; /* updated during retire, loosely tracks RING_HEAD */
+	u32 tail; /* updated on submission, used for RING_TAIL */
+	u32 emit; /* updated during request construction */
+
+	u32 space;
+	u32 size;
+	u32 wrap;
+	u32 effective_size;
+};
+
+#endif /* INTEL_RING_TYPES_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_rps.c b/drivers/gpu/drm/i915/gt/intel_rps.c
new file mode 100644
index 000000000..6b86250c3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_rps.c
@@ -0,0 +1,2591 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+
+#include <drm/i915_drm.h>
+
+#include "i915_drv.h"
+#include "i915_irq.h"
+#include "intel_breadcrumbs.h"
+#include "intel_gt.h"
+#include "intel_gt_clock_utils.h"
+#include "intel_gt_irq.h"
+#include "intel_gt_pm_irq.h"
+#include "intel_gt_regs.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pcode.h"
+#include "intel_rps.h"
+#include "vlv_sideband.h"
+#include "../../../platform/x86/intel_ips.h"
+
+#define BUSY_MAX_EI	20u /* ms */
+
+/*
+ * Lock protecting IPS related data structures
+ */
+static DEFINE_SPINLOCK(mchdev_lock);
+
+static struct intel_gt *rps_to_gt(struct intel_rps *rps)
+{
+	return container_of(rps, struct intel_gt, rps);
+}
+
+static struct drm_i915_private *rps_to_i915(struct intel_rps *rps)
+{
+	return rps_to_gt(rps)->i915;
+}
+
+static struct intel_uncore *rps_to_uncore(struct intel_rps *rps)
+{
+	return rps_to_gt(rps)->uncore;
+}
+
+static struct intel_guc_slpc *rps_to_slpc(struct intel_rps *rps)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+
+	return &gt->uc.guc.slpc;
+}
+
+static bool rps_uses_slpc(struct intel_rps *rps)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+
+	return intel_uc_uses_guc_slpc(&gt->uc);
+}
+
+static u32 rps_pm_sanitize_mask(struct intel_rps *rps, u32 mask)
+{
+	return mask & ~rps->pm_intrmsk_mbz;
+}
+
+static void set(struct intel_uncore *uncore, i915_reg_t reg, u32 val)
+{
+	intel_uncore_write_fw(uncore, reg, val);
+}
+
+static void rps_timer(struct timer_list *t)
+{
+	struct intel_rps *rps = from_timer(rps, t, timer);
+	struct intel_engine_cs *engine;
+	ktime_t dt, last, timestamp;
+	enum intel_engine_id id;
+	s64 max_busy[3] = {};
+
+	timestamp = 0;
+	for_each_engine(engine, rps_to_gt(rps), id) {
+		s64 busy;
+		int i;
+
+		dt = intel_engine_get_busy_time(engine, &timestamp);
+		last = engine->stats.rps;
+		engine->stats.rps = dt;
+
+		busy = ktime_to_ns(ktime_sub(dt, last));
+		for (i = 0; i < ARRAY_SIZE(max_busy); i++) {
+			if (busy > max_busy[i])
+				swap(busy, max_busy[i]);
+		}
+	}
+	last = rps->pm_timestamp;
+	rps->pm_timestamp = timestamp;
+
+	if (intel_rps_is_active(rps)) {
+		s64 busy;
+		int i;
+
+		dt = ktime_sub(timestamp, last);
+
+		/*
+		 * Our goal is to evaluate each engine independently, so we run
+		 * at the lowest clocks required to sustain the heaviest
+		 * workload. However, a task may be split into sequential
+		 * dependent operations across a set of engines, such that
+		 * the independent contributions do not account for high load,
+		 * but overall the task is GPU bound. For example, consider
+		 * video decode on vcs followed by colour post-processing
+		 * on vecs, followed by general post-processing on rcs.
+		 * Since multi-engines being active does imply a single
+		 * continuous workload across all engines, we hedge our
+		 * bets by only contributing a factor of the distributed
+		 * load into our busyness calculation.
+		 */
+		busy = max_busy[0];
+		for (i = 1; i < ARRAY_SIZE(max_busy); i++) {
+			if (!max_busy[i])
+				break;
+
+			busy += div_u64(max_busy[i], 1 << i);
+		}
+		GT_TRACE(rps_to_gt(rps),
+			 "busy:%lld [%d%%], max:[%lld, %lld, %lld], interval:%d\n",
+			 busy, (int)div64_u64(100 * busy, dt),
+			 max_busy[0], max_busy[1], max_busy[2],
+			 rps->pm_interval);
+
+		if (100 * busy > rps->power.up_threshold * dt &&
+		    rps->cur_freq < rps->max_freq_softlimit) {
+			rps->pm_iir |= GEN6_PM_RP_UP_THRESHOLD;
+			rps->pm_interval = 1;
+			schedule_work(&rps->work);
+		} else if (100 * busy < rps->power.down_threshold * dt &&
+			   rps->cur_freq > rps->min_freq_softlimit) {
+			rps->pm_iir |= GEN6_PM_RP_DOWN_THRESHOLD;
+			rps->pm_interval = 1;
+			schedule_work(&rps->work);
+		} else {
+			rps->last_adj = 0;
+		}
+
+		mod_timer(&rps->timer,
+			  jiffies + msecs_to_jiffies(rps->pm_interval));
+		rps->pm_interval = min(rps->pm_interval * 2, BUSY_MAX_EI);
+	}
+}
+
+static void rps_start_timer(struct intel_rps *rps)
+{
+	rps->pm_timestamp = ktime_sub(ktime_get(), rps->pm_timestamp);
+	rps->pm_interval = 1;
+	mod_timer(&rps->timer, jiffies + 1);
+}
+
+static void rps_stop_timer(struct intel_rps *rps)
+{
+	del_timer_sync(&rps->timer);
+	rps->pm_timestamp = ktime_sub(ktime_get(), rps->pm_timestamp);
+	cancel_work_sync(&rps->work);
+}
+
+static u32 rps_pm_mask(struct intel_rps *rps, u8 val)
+{
+	u32 mask = 0;
+
+	/* We use UP_EI_EXPIRED interrupts for both up/down in manual mode */
+	if (val > rps->min_freq_softlimit)
+		mask |= (GEN6_PM_RP_UP_EI_EXPIRED |
+			 GEN6_PM_RP_DOWN_THRESHOLD |
+			 GEN6_PM_RP_DOWN_TIMEOUT);
+
+	if (val < rps->max_freq_softlimit)
+		mask |= GEN6_PM_RP_UP_EI_EXPIRED | GEN6_PM_RP_UP_THRESHOLD;
+
+	mask &= rps->pm_events;
+
+	return rps_pm_sanitize_mask(rps, ~mask);
+}
+
+static void rps_reset_ei(struct intel_rps *rps)
+{
+	memset(&rps->ei, 0, sizeof(rps->ei));
+}
+
+static void rps_enable_interrupts(struct intel_rps *rps)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+
+	GEM_BUG_ON(rps_uses_slpc(rps));
+
+	GT_TRACE(gt, "interrupts:on rps->pm_events: %x, rps_pm_mask:%x\n",
+		 rps->pm_events, rps_pm_mask(rps, rps->last_freq));
+
+	rps_reset_ei(rps);
+
+	spin_lock_irq(gt->irq_lock);
+	gen6_gt_pm_enable_irq(gt, rps->pm_events);
+	spin_unlock_irq(gt->irq_lock);
+
+	intel_uncore_write(gt->uncore,
+			   GEN6_PMINTRMSK, rps_pm_mask(rps, rps->last_freq));
+}
+
+static void gen6_rps_reset_interrupts(struct intel_rps *rps)
+{
+	gen6_gt_pm_reset_iir(rps_to_gt(rps), GEN6_PM_RPS_EVENTS);
+}
+
+static void gen11_rps_reset_interrupts(struct intel_rps *rps)
+{
+	while (gen11_gt_reset_one_iir(rps_to_gt(rps), 0, GEN11_GTPM))
+		;
+}
+
+static void rps_reset_interrupts(struct intel_rps *rps)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+
+	spin_lock_irq(gt->irq_lock);
+	if (GRAPHICS_VER(gt->i915) >= 11)
+		gen11_rps_reset_interrupts(rps);
+	else
+		gen6_rps_reset_interrupts(rps);
+
+	rps->pm_iir = 0;
+	spin_unlock_irq(gt->irq_lock);
+}
+
+static void rps_disable_interrupts(struct intel_rps *rps)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+
+	intel_uncore_write(gt->uncore,
+			   GEN6_PMINTRMSK, rps_pm_sanitize_mask(rps, ~0u));
+
+	spin_lock_irq(gt->irq_lock);
+	gen6_gt_pm_disable_irq(gt, GEN6_PM_RPS_EVENTS);
+	spin_unlock_irq(gt->irq_lock);
+
+	intel_synchronize_irq(gt->i915);
+
+	/*
+	 * Now that we will not be generating any more work, flush any
+	 * outstanding tasks. As we are called on the RPS idle path,
+	 * we will reset the GPU to minimum frequencies, so the current
+	 * state of the worker can be discarded.
+	 */
+	cancel_work_sync(&rps->work);
+
+	rps_reset_interrupts(rps);
+	GT_TRACE(gt, "interrupts:off\n");
+}
+
+static const struct cparams {
+	u16 i;
+	u16 t;
+	u16 m;
+	u16 c;
+} cparams[] = {
+	{ 1, 1333, 301, 28664 },
+	{ 1, 1066, 294, 24460 },
+	{ 1, 800, 294, 25192 },
+	{ 0, 1333, 276, 27605 },
+	{ 0, 1066, 276, 27605 },
+	{ 0, 800, 231, 23784 },
+};
+
+static void gen5_rps_init(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	u8 fmax, fmin, fstart;
+	u32 rgvmodectl;
+	int c_m, i;
+
+	if (i915->fsb_freq <= 3200)
+		c_m = 0;
+	else if (i915->fsb_freq <= 4800)
+		c_m = 1;
+	else
+		c_m = 2;
+
+	for (i = 0; i < ARRAY_SIZE(cparams); i++) {
+		if (cparams[i].i == c_m && cparams[i].t == i915->mem_freq) {
+			rps->ips.m = cparams[i].m;
+			rps->ips.c = cparams[i].c;
+			break;
+		}
+	}
+
+	rgvmodectl = intel_uncore_read(uncore, MEMMODECTL);
+
+	/* Set up min, max, and cur for interrupt handling */
+	fmax = (rgvmodectl & MEMMODE_FMAX_MASK) >> MEMMODE_FMAX_SHIFT;
+	fmin = (rgvmodectl & MEMMODE_FMIN_MASK);
+	fstart = (rgvmodectl & MEMMODE_FSTART_MASK) >>
+		MEMMODE_FSTART_SHIFT;
+	drm_dbg(&i915->drm, "fmax: %d, fmin: %d, fstart: %d\n",
+		fmax, fmin, fstart);
+
+	rps->min_freq = fmax;
+	rps->efficient_freq = fstart;
+	rps->max_freq = fmin;
+}
+
+static unsigned long
+__ips_chipset_val(struct intel_ips *ips)
+{
+	struct intel_uncore *uncore =
+		rps_to_uncore(container_of(ips, struct intel_rps, ips));
+	unsigned long now = jiffies_to_msecs(jiffies), dt;
+	unsigned long result;
+	u64 total, delta;
+
+	lockdep_assert_held(&mchdev_lock);
+
+	/*
+	 * Prevent division-by-zero if we are asking too fast.
+	 * Also, we don't get interesting results if we are polling
+	 * faster than once in 10ms, so just return the saved value
+	 * in such cases.
+	 */
+	dt = now - ips->last_time1;
+	if (dt <= 10)
+		return ips->chipset_power;
+
+	/* FIXME: handle per-counter overflow */
+	total = intel_uncore_read(uncore, DMIEC);
+	total += intel_uncore_read(uncore, DDREC);
+	total += intel_uncore_read(uncore, CSIEC);
+
+	delta = total - ips->last_count1;
+
+	result = div_u64(div_u64(ips->m * delta, dt) + ips->c, 10);
+
+	ips->last_count1 = total;
+	ips->last_time1 = now;
+
+	ips->chipset_power = result;
+
+	return result;
+}
+
+static unsigned long ips_mch_val(struct intel_uncore *uncore)
+{
+	unsigned int m, x, b;
+	u32 tsfs;
+
+	tsfs = intel_uncore_read(uncore, TSFS);
+	x = intel_uncore_read8(uncore, TR1);
+
+	b = tsfs & TSFS_INTR_MASK;
+	m = (tsfs & TSFS_SLOPE_MASK) >> TSFS_SLOPE_SHIFT;
+
+	return m * x / 127 - b;
+}
+
+static int _pxvid_to_vd(u8 pxvid)
+{
+	if (pxvid == 0)
+		return 0;
+
+	if (pxvid >= 8 && pxvid < 31)
+		pxvid = 31;
+
+	return (pxvid + 2) * 125;
+}
+
+static u32 pvid_to_extvid(struct drm_i915_private *i915, u8 pxvid)
+{
+	const int vd = _pxvid_to_vd(pxvid);
+
+	if (INTEL_INFO(i915)->is_mobile)
+		return max(vd - 1125, 0);
+
+	return vd;
+}
+
+static void __gen5_ips_update(struct intel_ips *ips)
+{
+	struct intel_uncore *uncore =
+		rps_to_uncore(container_of(ips, struct intel_rps, ips));
+	u64 now, delta, dt;
+	u32 count;
+
+	lockdep_assert_held(&mchdev_lock);
+
+	now = ktime_get_raw_ns();
+	dt = now - ips->last_time2;
+	do_div(dt, NSEC_PER_MSEC);
+
+	/* Don't divide by 0 */
+	if (dt <= 10)
+		return;
+
+	count = intel_uncore_read(uncore, GFXEC);
+	delta = count - ips->last_count2;
+
+	ips->last_count2 = count;
+	ips->last_time2 = now;
+
+	/* More magic constants... */
+	ips->gfx_power = div_u64(delta * 1181, dt * 10);
+}
+
+static void gen5_rps_update(struct intel_rps *rps)
+{
+	spin_lock_irq(&mchdev_lock);
+	__gen5_ips_update(&rps->ips);
+	spin_unlock_irq(&mchdev_lock);
+}
+
+static unsigned int gen5_invert_freq(struct intel_rps *rps,
+				     unsigned int val)
+{
+	/* Invert the frequency bin into an ips delay */
+	val = rps->max_freq - val;
+	val = rps->min_freq + val;
+
+	return val;
+}
+
+static int __gen5_rps_set(struct intel_rps *rps, u8 val)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	u16 rgvswctl;
+
+	lockdep_assert_held(&mchdev_lock);
+
+	rgvswctl = intel_uncore_read16(uncore, MEMSWCTL);
+	if (rgvswctl & MEMCTL_CMD_STS) {
+		DRM_DEBUG("gpu busy, RCS change rejected\n");
+		return -EBUSY; /* still busy with another command */
+	}
+
+	/* Invert the frequency bin into an ips delay */
+	val = gen5_invert_freq(rps, val);
+
+	rgvswctl =
+		(MEMCTL_CMD_CHFREQ << MEMCTL_CMD_SHIFT) |
+		(val << MEMCTL_FREQ_SHIFT) |
+		MEMCTL_SFCAVM;
+	intel_uncore_write16(uncore, MEMSWCTL, rgvswctl);
+	intel_uncore_posting_read16(uncore, MEMSWCTL);
+
+	rgvswctl |= MEMCTL_CMD_STS;
+	intel_uncore_write16(uncore, MEMSWCTL, rgvswctl);
+
+	return 0;
+}
+
+static int gen5_rps_set(struct intel_rps *rps, u8 val)
+{
+	int err;
+
+	spin_lock_irq(&mchdev_lock);
+	err = __gen5_rps_set(rps, val);
+	spin_unlock_irq(&mchdev_lock);
+
+	return err;
+}
+
+static unsigned long intel_pxfreq(u32 vidfreq)
+{
+	int div = (vidfreq & 0x3f0000) >> 16;
+	int post = (vidfreq & 0x3000) >> 12;
+	int pre = (vidfreq & 0x7);
+
+	if (!pre)
+		return 0;
+
+	return div * 133333 / (pre << post);
+}
+
+static unsigned int init_emon(struct intel_uncore *uncore)
+{
+	u8 pxw[16];
+	int i;
+
+	/* Disable to program */
+	intel_uncore_write(uncore, ECR, 0);
+	intel_uncore_posting_read(uncore, ECR);
+
+	/* Program energy weights for various events */
+	intel_uncore_write(uncore, SDEW, 0x15040d00);
+	intel_uncore_write(uncore, CSIEW0, 0x007f0000);
+	intel_uncore_write(uncore, CSIEW1, 0x1e220004);
+	intel_uncore_write(uncore, CSIEW2, 0x04000004);
+
+	for (i = 0; i < 5; i++)
+		intel_uncore_write(uncore, PEW(i), 0);
+	for (i = 0; i < 3; i++)
+		intel_uncore_write(uncore, DEW(i), 0);
+
+	/* Program P-state weights to account for frequency power adjustment */
+	for (i = 0; i < 16; i++) {
+		u32 pxvidfreq = intel_uncore_read(uncore, PXVFREQ(i));
+		unsigned int freq = intel_pxfreq(pxvidfreq);
+		unsigned int vid =
+			(pxvidfreq & PXVFREQ_PX_MASK) >> PXVFREQ_PX_SHIFT;
+		unsigned int val;
+
+		val = vid * vid * freq / 1000 * 255;
+		val /= 127 * 127 * 900;
+
+		pxw[i] = val;
+	}
+	/* Render standby states get 0 weight */
+	pxw[14] = 0;
+	pxw[15] = 0;
+
+	for (i = 0; i < 4; i++) {
+		intel_uncore_write(uncore, PXW(i),
+				   pxw[i * 4 + 0] << 24 |
+				   pxw[i * 4 + 1] << 16 |
+				   pxw[i * 4 + 2] <<  8 |
+				   pxw[i * 4 + 3] <<  0);
+	}
+
+	/* Adjust magic regs to magic values (more experimental results) */
+	intel_uncore_write(uncore, OGW0, 0);
+	intel_uncore_write(uncore, OGW1, 0);
+	intel_uncore_write(uncore, EG0, 0x00007f00);
+	intel_uncore_write(uncore, EG1, 0x0000000e);
+	intel_uncore_write(uncore, EG2, 0x000e0000);
+	intel_uncore_write(uncore, EG3, 0x68000300);
+	intel_uncore_write(uncore, EG4, 0x42000000);
+	intel_uncore_write(uncore, EG5, 0x00140031);
+	intel_uncore_write(uncore, EG6, 0);
+	intel_uncore_write(uncore, EG7, 0);
+
+	for (i = 0; i < 8; i++)
+		intel_uncore_write(uncore, PXWL(i), 0);
+
+	/* Enable PMON + select events */
+	intel_uncore_write(uncore, ECR, 0x80000019);
+
+	return intel_uncore_read(uncore, LCFUSE02) & LCFUSE_HIV_MASK;
+}
+
+static bool gen5_rps_enable(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	u8 fstart, vstart;
+	u32 rgvmodectl;
+
+	spin_lock_irq(&mchdev_lock);
+
+	rgvmodectl = intel_uncore_read(uncore, MEMMODECTL);
+
+	/* Enable temp reporting */
+	intel_uncore_write16(uncore, PMMISC,
+			     intel_uncore_read16(uncore, PMMISC) | MCPPCE_EN);
+	intel_uncore_write16(uncore, TSC1,
+			     intel_uncore_read16(uncore, TSC1) | TSE);
+
+	/* 100ms RC evaluation intervals */
+	intel_uncore_write(uncore, RCUPEI, 100000);
+	intel_uncore_write(uncore, RCDNEI, 100000);
+
+	/* Set max/min thresholds to 90ms and 80ms respectively */
+	intel_uncore_write(uncore, RCBMAXAVG, 90000);
+	intel_uncore_write(uncore, RCBMINAVG, 80000);
+
+	intel_uncore_write(uncore, MEMIHYST, 1);
+
+	/* Set up min, max, and cur for interrupt handling */
+	fstart = (rgvmodectl & MEMMODE_FSTART_MASK) >>
+		MEMMODE_FSTART_SHIFT;
+
+	vstart = (intel_uncore_read(uncore, PXVFREQ(fstart)) &
+		  PXVFREQ_PX_MASK) >> PXVFREQ_PX_SHIFT;
+
+	intel_uncore_write(uncore,
+			   MEMINTREN,
+			   MEMINT_CX_SUPR_EN | MEMINT_EVAL_CHG_EN);
+
+	intel_uncore_write(uncore, VIDSTART, vstart);
+	intel_uncore_posting_read(uncore, VIDSTART);
+
+	rgvmodectl |= MEMMODE_SWMODE_EN;
+	intel_uncore_write(uncore, MEMMODECTL, rgvmodectl);
+
+	if (wait_for_atomic((intel_uncore_read(uncore, MEMSWCTL) &
+			     MEMCTL_CMD_STS) == 0, 10))
+		drm_err(&uncore->i915->drm,
+			"stuck trying to change perf mode\n");
+	mdelay(1);
+
+	__gen5_rps_set(rps, rps->cur_freq);
+
+	rps->ips.last_count1 = intel_uncore_read(uncore, DMIEC);
+	rps->ips.last_count1 += intel_uncore_read(uncore, DDREC);
+	rps->ips.last_count1 += intel_uncore_read(uncore, CSIEC);
+	rps->ips.last_time1 = jiffies_to_msecs(jiffies);
+
+	rps->ips.last_count2 = intel_uncore_read(uncore, GFXEC);
+	rps->ips.last_time2 = ktime_get_raw_ns();
+
+	spin_lock(&i915->irq_lock);
+	ilk_enable_display_irq(i915, DE_PCU_EVENT);
+	spin_unlock(&i915->irq_lock);
+
+	spin_unlock_irq(&mchdev_lock);
+
+	rps->ips.corr = init_emon(uncore);
+
+	return true;
+}
+
+static void gen5_rps_disable(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	u16 rgvswctl;
+
+	spin_lock_irq(&mchdev_lock);
+
+	spin_lock(&i915->irq_lock);
+	ilk_disable_display_irq(i915, DE_PCU_EVENT);
+	spin_unlock(&i915->irq_lock);
+
+	rgvswctl = intel_uncore_read16(uncore, MEMSWCTL);
+
+	/* Ack interrupts, disable EFC interrupt */
+	intel_uncore_write(uncore, MEMINTREN,
+			   intel_uncore_read(uncore, MEMINTREN) &
+			   ~MEMINT_EVAL_CHG_EN);
+	intel_uncore_write(uncore, MEMINTRSTS, MEMINT_EVAL_CHG);
+
+	/* Go back to the starting frequency */
+	__gen5_rps_set(rps, rps->idle_freq);
+	mdelay(1);
+	rgvswctl |= MEMCTL_CMD_STS;
+	intel_uncore_write(uncore, MEMSWCTL, rgvswctl);
+	mdelay(1);
+
+	spin_unlock_irq(&mchdev_lock);
+}
+
+static u32 rps_limits(struct intel_rps *rps, u8 val)
+{
+	u32 limits;
+
+	/*
+	 * Only set the down limit when we've reached the lowest level to avoid
+	 * getting more interrupts, otherwise leave this clear. This prevents a
+	 * race in the hw when coming out of rc6: There's a tiny window where
+	 * the hw runs at the minimal clock before selecting the desired
+	 * frequency, if the down threshold expires in that window we will not
+	 * receive a down interrupt.
+	 */
+	if (GRAPHICS_VER(rps_to_i915(rps)) >= 9) {
+		limits = rps->max_freq_softlimit << 23;
+		if (val <= rps->min_freq_softlimit)
+			limits |= rps->min_freq_softlimit << 14;
+	} else {
+		limits = rps->max_freq_softlimit << 24;
+		if (val <= rps->min_freq_softlimit)
+			limits |= rps->min_freq_softlimit << 16;
+	}
+
+	return limits;
+}
+
+static void rps_set_power(struct intel_rps *rps, int new_power)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+	struct intel_uncore *uncore = gt->uncore;
+	u32 threshold_up = 0, threshold_down = 0; /* in % */
+	u32 ei_up = 0, ei_down = 0;
+
+	lockdep_assert_held(&rps->power.mutex);
+
+	if (new_power == rps->power.mode)
+		return;
+
+	threshold_up = 95;
+	threshold_down = 85;
+
+	/* Note the units here are not exactly 1us, but 1280ns. */
+	switch (new_power) {
+	case LOW_POWER:
+		ei_up = 16000;
+		ei_down = 32000;
+		break;
+
+	case BETWEEN:
+		ei_up = 13000;
+		ei_down = 32000;
+		break;
+
+	case HIGH_POWER:
+		ei_up = 10000;
+		ei_down = 32000;
+		break;
+	}
+
+	/* When byt can survive without system hang with dynamic
+	 * sw freq adjustments, this restriction can be lifted.
+	 */
+	if (IS_VALLEYVIEW(gt->i915))
+		goto skip_hw_write;
+
+	GT_TRACE(gt,
+		 "changing power mode [%d], up %d%% @ %dus, down %d%% @ %dus\n",
+		 new_power, threshold_up, ei_up, threshold_down, ei_down);
+
+	set(uncore, GEN6_RP_UP_EI,
+	    intel_gt_ns_to_pm_interval(gt, ei_up * 1000));
+	set(uncore, GEN6_RP_UP_THRESHOLD,
+	    intel_gt_ns_to_pm_interval(gt, ei_up * threshold_up * 10));
+
+	set(uncore, GEN6_RP_DOWN_EI,
+	    intel_gt_ns_to_pm_interval(gt, ei_down * 1000));
+	set(uncore, GEN6_RP_DOWN_THRESHOLD,
+	    intel_gt_ns_to_pm_interval(gt, ei_down * threshold_down * 10));
+
+	set(uncore, GEN6_RP_CONTROL,
+	    (GRAPHICS_VER(gt->i915) > 9 ? 0 : GEN6_RP_MEDIA_TURBO) |
+	    GEN6_RP_MEDIA_HW_NORMAL_MODE |
+	    GEN6_RP_MEDIA_IS_GFX |
+	    GEN6_RP_ENABLE |
+	    GEN6_RP_UP_BUSY_AVG |
+	    GEN6_RP_DOWN_IDLE_AVG);
+
+skip_hw_write:
+	rps->power.mode = new_power;
+	rps->power.up_threshold = threshold_up;
+	rps->power.down_threshold = threshold_down;
+}
+
+static void gen6_rps_set_thresholds(struct intel_rps *rps, u8 val)
+{
+	int new_power;
+
+	new_power = rps->power.mode;
+	switch (rps->power.mode) {
+	case LOW_POWER:
+		if (val > rps->efficient_freq + 1 &&
+		    val > rps->cur_freq)
+			new_power = BETWEEN;
+		break;
+
+	case BETWEEN:
+		if (val <= rps->efficient_freq &&
+		    val < rps->cur_freq)
+			new_power = LOW_POWER;
+		else if (val >= rps->rp0_freq &&
+			 val > rps->cur_freq)
+			new_power = HIGH_POWER;
+		break;
+
+	case HIGH_POWER:
+		if (val < (rps->rp1_freq + rps->rp0_freq) >> 1 &&
+		    val < rps->cur_freq)
+			new_power = BETWEEN;
+		break;
+	}
+	/* Max/min bins are special */
+	if (val <= rps->min_freq_softlimit)
+		new_power = LOW_POWER;
+	if (val >= rps->max_freq_softlimit)
+		new_power = HIGH_POWER;
+
+	mutex_lock(&rps->power.mutex);
+	if (rps->power.interactive)
+		new_power = HIGH_POWER;
+	rps_set_power(rps, new_power);
+	mutex_unlock(&rps->power.mutex);
+}
+
+void intel_rps_mark_interactive(struct intel_rps *rps, bool interactive)
+{
+	GT_TRACE(rps_to_gt(rps), "mark interactive: %s\n",
+		 str_yes_no(interactive));
+
+	mutex_lock(&rps->power.mutex);
+	if (interactive) {
+		if (!rps->power.interactive++ && intel_rps_is_active(rps))
+			rps_set_power(rps, HIGH_POWER);
+	} else {
+		GEM_BUG_ON(!rps->power.interactive);
+		rps->power.interactive--;
+	}
+	mutex_unlock(&rps->power.mutex);
+}
+
+static int gen6_rps_set(struct intel_rps *rps, u8 val)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 swreq;
+
+	GEM_BUG_ON(rps_uses_slpc(rps));
+
+	if (GRAPHICS_VER(i915) >= 9)
+		swreq = GEN9_FREQUENCY(val);
+	else if (IS_HASWELL(i915) || IS_BROADWELL(i915))
+		swreq = HSW_FREQUENCY(val);
+	else
+		swreq = (GEN6_FREQUENCY(val) |
+			 GEN6_OFFSET(0) |
+			 GEN6_AGGRESSIVE_TURBO);
+	set(uncore, GEN6_RPNSWREQ, swreq);
+
+	GT_TRACE(rps_to_gt(rps), "set val:%x, freq:%d, swreq:%x\n",
+		 val, intel_gpu_freq(rps, val), swreq);
+
+	return 0;
+}
+
+static int vlv_rps_set(struct intel_rps *rps, u8 val)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	int err;
+
+	vlv_punit_get(i915);
+	err = vlv_punit_write(i915, PUNIT_REG_GPU_FREQ_REQ, val);
+	vlv_punit_put(i915);
+
+	GT_TRACE(rps_to_gt(rps), "set val:%x, freq:%d\n",
+		 val, intel_gpu_freq(rps, val));
+
+	return err;
+}
+
+static int rps_set(struct intel_rps *rps, u8 val, bool update)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	int err;
+
+	if (val == rps->last_freq)
+		return 0;
+
+	if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+		err = vlv_rps_set(rps, val);
+	else if (GRAPHICS_VER(i915) >= 6)
+		err = gen6_rps_set(rps, val);
+	else
+		err = gen5_rps_set(rps, val);
+	if (err)
+		return err;
+
+	if (update && GRAPHICS_VER(i915) >= 6)
+		gen6_rps_set_thresholds(rps, val);
+	rps->last_freq = val;
+
+	return 0;
+}
+
+void intel_rps_unpark(struct intel_rps *rps)
+{
+	if (!intel_rps_is_enabled(rps))
+		return;
+
+	GT_TRACE(rps_to_gt(rps), "unpark:%x\n", rps->cur_freq);
+
+	/*
+	 * Use the user's desired frequency as a guide, but for better
+	 * performance, jump directly to RPe as our starting frequency.
+	 */
+	mutex_lock(&rps->lock);
+
+	intel_rps_set_active(rps);
+	intel_rps_set(rps,
+		      clamp(rps->cur_freq,
+			    rps->min_freq_softlimit,
+			    rps->max_freq_softlimit));
+
+	mutex_unlock(&rps->lock);
+
+	rps->pm_iir = 0;
+	if (intel_rps_has_interrupts(rps))
+		rps_enable_interrupts(rps);
+	if (intel_rps_uses_timer(rps))
+		rps_start_timer(rps);
+
+	if (GRAPHICS_VER(rps_to_i915(rps)) == 5)
+		gen5_rps_update(rps);
+}
+
+void intel_rps_park(struct intel_rps *rps)
+{
+	int adj;
+
+	if (!intel_rps_is_enabled(rps))
+		return;
+
+	if (!intel_rps_clear_active(rps))
+		return;
+
+	if (intel_rps_uses_timer(rps))
+		rps_stop_timer(rps);
+	if (intel_rps_has_interrupts(rps))
+		rps_disable_interrupts(rps);
+
+	if (rps->last_freq <= rps->idle_freq)
+		return;
+
+	/*
+	 * The punit delays the write of the frequency and voltage until it
+	 * determines the GPU is awake. During normal usage we don't want to
+	 * waste power changing the frequency if the GPU is sleeping (rc6).
+	 * However, the GPU and driver is now idle and we do not want to delay
+	 * switching to minimum voltage (reducing power whilst idle) as we do
+	 * not expect to be woken in the near future and so must flush the
+	 * change by waking the device.
+	 *
+	 * We choose to take the media powerwell (either would do to trick the
+	 * punit into committing the voltage change) as that takes a lot less
+	 * power than the render powerwell.
+	 */
+	intel_uncore_forcewake_get(rps_to_uncore(rps), FORCEWAKE_MEDIA);
+	rps_set(rps, rps->idle_freq, false);
+	intel_uncore_forcewake_put(rps_to_uncore(rps), FORCEWAKE_MEDIA);
+
+	/*
+	 * Since we will try and restart from the previously requested
+	 * frequency on unparking, treat this idle point as a downclock
+	 * interrupt and reduce the frequency for resume. If we park/unpark
+	 * more frequently than the rps worker can run, we will not respond
+	 * to any EI and never see a change in frequency.
+	 *
+	 * (Note we accommodate Cherryview's limitation of only using an
+	 * even bin by applying it to all.)
+	 */
+	adj = rps->last_adj;
+	if (adj < 0)
+		adj *= 2;
+	else /* CHV needs even encode values */
+		adj = -2;
+	rps->last_adj = adj;
+	rps->cur_freq = max_t(int, rps->cur_freq + adj, rps->min_freq);
+	if (rps->cur_freq < rps->efficient_freq) {
+		rps->cur_freq = rps->efficient_freq;
+		rps->last_adj = 0;
+	}
+
+	GT_TRACE(rps_to_gt(rps), "park:%x\n", rps->cur_freq);
+}
+
+u32 intel_rps_get_boost_frequency(struct intel_rps *rps)
+{
+	struct intel_guc_slpc *slpc;
+
+	if (rps_uses_slpc(rps)) {
+		slpc = rps_to_slpc(rps);
+
+		return slpc->boost_freq;
+	} else {
+		return intel_gpu_freq(rps, rps->boost_freq);
+	}
+}
+
+static int rps_set_boost_freq(struct intel_rps *rps, u32 val)
+{
+	bool boost = false;
+
+	/* Validate against (static) hardware limits */
+	val = intel_freq_opcode(rps, val);
+	if (val < rps->min_freq || val > rps->max_freq)
+		return -EINVAL;
+
+	mutex_lock(&rps->lock);
+	if (val != rps->boost_freq) {
+		rps->boost_freq = val;
+		boost = atomic_read(&rps->num_waiters);
+	}
+	mutex_unlock(&rps->lock);
+	if (boost)
+		schedule_work(&rps->work);
+
+	return 0;
+}
+
+int intel_rps_set_boost_frequency(struct intel_rps *rps, u32 freq)
+{
+	struct intel_guc_slpc *slpc;
+
+	if (rps_uses_slpc(rps)) {
+		slpc = rps_to_slpc(rps);
+
+		return intel_guc_slpc_set_boost_freq(slpc, freq);
+	} else {
+		return rps_set_boost_freq(rps, freq);
+	}
+}
+
+void intel_rps_dec_waiters(struct intel_rps *rps)
+{
+	struct intel_guc_slpc *slpc;
+
+	if (rps_uses_slpc(rps)) {
+		slpc = rps_to_slpc(rps);
+
+		intel_guc_slpc_dec_waiters(slpc);
+	} else {
+		atomic_dec(&rps->num_waiters);
+	}
+}
+
+void intel_rps_boost(struct i915_request *rq)
+{
+	struct intel_guc_slpc *slpc;
+
+	if (i915_request_signaled(rq) || i915_request_has_waitboost(rq))
+		return;
+
+	/* Serializes with i915_request_retire() */
+	if (!test_and_set_bit(I915_FENCE_FLAG_BOOST, &rq->fence.flags)) {
+		struct intel_rps *rps = &READ_ONCE(rq->engine)->gt->rps;
+
+		if (rps_uses_slpc(rps)) {
+			slpc = rps_to_slpc(rps);
+
+			/* Return if old value is non zero */
+			if (!atomic_fetch_inc(&slpc->num_waiters))
+				schedule_work(&slpc->boost_work);
+
+			return;
+		}
+
+		if (atomic_fetch_inc(&rps->num_waiters))
+			return;
+
+		if (!intel_rps_is_active(rps))
+			return;
+
+		GT_TRACE(rps_to_gt(rps), "boost fence:%llx:%llx\n",
+			 rq->fence.context, rq->fence.seqno);
+
+		if (READ_ONCE(rps->cur_freq) < rps->boost_freq)
+			schedule_work(&rps->work);
+
+		WRITE_ONCE(rps->boosts, rps->boosts + 1); /* debug only */
+	}
+}
+
+int intel_rps_set(struct intel_rps *rps, u8 val)
+{
+	int err;
+
+	lockdep_assert_held(&rps->lock);
+	GEM_BUG_ON(val > rps->max_freq);
+	GEM_BUG_ON(val < rps->min_freq);
+
+	if (intel_rps_is_active(rps)) {
+		err = rps_set(rps, val, true);
+		if (err)
+			return err;
+
+		/*
+		 * Make sure we continue to get interrupts
+		 * until we hit the minimum or maximum frequencies.
+		 */
+		if (intel_rps_has_interrupts(rps)) {
+			struct intel_uncore *uncore = rps_to_uncore(rps);
+
+			set(uncore,
+			    GEN6_RP_INTERRUPT_LIMITS, rps_limits(rps, val));
+
+			set(uncore, GEN6_PMINTRMSK, rps_pm_mask(rps, val));
+		}
+	}
+
+	rps->cur_freq = val;
+	return 0;
+}
+
+static u32 intel_rps_read_state_cap(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+
+	if (IS_PONTEVECCHIO(i915))
+		return intel_uncore_read(uncore, PVC_RP_STATE_CAP);
+	else if (IS_XEHPSDV(i915))
+		return intel_uncore_read(uncore, XEHPSDV_RP_STATE_CAP);
+	else if (IS_GEN9_LP(i915))
+		return intel_uncore_read(uncore, BXT_RP_STATE_CAP);
+	else
+		return intel_uncore_read(uncore, GEN6_RP_STATE_CAP);
+}
+
+/**
+ * gen6_rps_get_freq_caps - Get freq caps exposed by HW
+ * @rps: the intel_rps structure
+ * @caps: returned freq caps
+ *
+ * Returned "caps" frequencies should be converted to MHz using
+ * intel_gpu_freq()
+ */
+void gen6_rps_get_freq_caps(struct intel_rps *rps, struct intel_rps_freq_caps *caps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 rp_state_cap;
+
+	rp_state_cap = intel_rps_read_state_cap(rps);
+
+	/* static values from HW: RP0 > RP1 > RPn (min_freq) */
+	if (IS_GEN9_LP(i915)) {
+		caps->rp0_freq = (rp_state_cap >> 16) & 0xff;
+		caps->rp1_freq = (rp_state_cap >>  8) & 0xff;
+		caps->min_freq = (rp_state_cap >>  0) & 0xff;
+	} else {
+		caps->rp0_freq = (rp_state_cap >>  0) & 0xff;
+		if (GRAPHICS_VER(i915) >= 10)
+			caps->rp1_freq = REG_FIELD_GET(RPE_MASK,
+						       intel_uncore_read(to_gt(i915)->uncore,
+						       GEN10_FREQ_INFO_REC));
+		else
+			caps->rp1_freq = (rp_state_cap >>  8) & 0xff;
+		caps->min_freq = (rp_state_cap >> 16) & 0xff;
+	}
+
+	if (IS_GEN9_BC(i915) || GRAPHICS_VER(i915) >= 11) {
+		/*
+		 * In this case rp_state_cap register reports frequencies in
+		 * units of 50 MHz. Convert these to the actual "hw unit", i.e.
+		 * units of 16.67 MHz
+		 */
+		caps->rp0_freq *= GEN9_FREQ_SCALER;
+		caps->rp1_freq *= GEN9_FREQ_SCALER;
+		caps->min_freq *= GEN9_FREQ_SCALER;
+	}
+}
+
+static void gen6_rps_init(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	struct intel_rps_freq_caps caps;
+
+	gen6_rps_get_freq_caps(rps, &caps);
+	rps->rp0_freq = caps.rp0_freq;
+	rps->rp1_freq = caps.rp1_freq;
+	rps->min_freq = caps.min_freq;
+
+	/* hw_max = RP0 until we check for overclocking */
+	rps->max_freq = rps->rp0_freq;
+
+	rps->efficient_freq = rps->rp1_freq;
+	if (IS_HASWELL(i915) || IS_BROADWELL(i915) ||
+	    IS_GEN9_BC(i915) || GRAPHICS_VER(i915) >= 11) {
+		u32 ddcc_status = 0;
+		u32 mult = 1;
+
+		if (IS_GEN9_BC(i915) || GRAPHICS_VER(i915) >= 11)
+			mult = GEN9_FREQ_SCALER;
+		if (snb_pcode_read(rps_to_gt(rps)->uncore,
+				   HSW_PCODE_DYNAMIC_DUTY_CYCLE_CONTROL,
+				   &ddcc_status, NULL) == 0)
+			rps->efficient_freq =
+				clamp_t(u32,
+					((ddcc_status >> 8) & 0xff) * mult,
+					rps->min_freq,
+					rps->max_freq);
+	}
+}
+
+static bool rps_reset(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+
+	/* force a reset */
+	rps->power.mode = -1;
+	rps->last_freq = -1;
+
+	if (rps_set(rps, rps->min_freq, true)) {
+		drm_err(&i915->drm, "Failed to reset RPS to initial values\n");
+		return false;
+	}
+
+	rps->cur_freq = rps->min_freq;
+	return true;
+}
+
+/* See the Gen9_GT_PM_Programming_Guide doc for the below */
+static bool gen9_rps_enable(struct intel_rps *rps)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+	struct intel_uncore *uncore = gt->uncore;
+
+	/* Program defaults and thresholds for RPS */
+	if (GRAPHICS_VER(gt->i915) == 9)
+		intel_uncore_write_fw(uncore, GEN6_RC_VIDEO_FREQ,
+				      GEN9_FREQUENCY(rps->rp1_freq));
+
+	intel_uncore_write_fw(uncore, GEN6_RP_IDLE_HYSTERSIS, 0xa);
+
+	rps->pm_events = GEN6_PM_RP_UP_THRESHOLD | GEN6_PM_RP_DOWN_THRESHOLD;
+
+	return rps_reset(rps);
+}
+
+static bool gen8_rps_enable(struct intel_rps *rps)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+
+	intel_uncore_write_fw(uncore, GEN6_RC_VIDEO_FREQ,
+			      HSW_FREQUENCY(rps->rp1_freq));
+
+	intel_uncore_write_fw(uncore, GEN6_RP_IDLE_HYSTERSIS, 10);
+
+	rps->pm_events = GEN6_PM_RP_UP_THRESHOLD | GEN6_PM_RP_DOWN_THRESHOLD;
+
+	return rps_reset(rps);
+}
+
+static bool gen6_rps_enable(struct intel_rps *rps)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+
+	/* Power down if completely idle for over 50ms */
+	intel_uncore_write_fw(uncore, GEN6_RP_DOWN_TIMEOUT, 50000);
+	intel_uncore_write_fw(uncore, GEN6_RP_IDLE_HYSTERSIS, 10);
+
+	rps->pm_events = (GEN6_PM_RP_UP_THRESHOLD |
+			  GEN6_PM_RP_DOWN_THRESHOLD |
+			  GEN6_PM_RP_DOWN_TIMEOUT);
+
+	return rps_reset(rps);
+}
+
+static int chv_rps_max_freq(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	struct intel_gt *gt = rps_to_gt(rps);
+	u32 val;
+
+	val = vlv_punit_read(i915, FB_GFX_FMAX_AT_VMAX_FUSE);
+
+	switch (gt->info.sseu.eu_total) {
+	case 8:
+		/* (2 * 4) config */
+		val >>= FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT;
+		break;
+	case 12:
+		/* (2 * 6) config */
+		val >>= FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT;
+		break;
+	case 16:
+		/* (2 * 8) config */
+	default:
+		/* Setting (2 * 8) Min RP0 for any other combination */
+		val >>= FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT;
+		break;
+	}
+
+	return val & FB_GFX_FREQ_FUSE_MASK;
+}
+
+static int chv_rps_rpe_freq(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val;
+
+	val = vlv_punit_read(i915, PUNIT_GPU_DUTYCYCLE_REG);
+	val >>= PUNIT_GPU_DUTYCYCLE_RPE_FREQ_SHIFT;
+
+	return val & PUNIT_GPU_DUTYCYCLE_RPE_FREQ_MASK;
+}
+
+static int chv_rps_guar_freq(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val;
+
+	val = vlv_punit_read(i915, FB_GFX_FMAX_AT_VMAX_FUSE);
+
+	return val & FB_GFX_FREQ_FUSE_MASK;
+}
+
+static u32 chv_rps_min_freq(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val;
+
+	val = vlv_punit_read(i915, FB_GFX_FMIN_AT_VMIN_FUSE);
+	val >>= FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT;
+
+	return val & FB_GFX_FREQ_FUSE_MASK;
+}
+
+static bool chv_rps_enable(struct intel_rps *rps)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val;
+
+	/* 1: Program defaults and thresholds for RPS*/
+	intel_uncore_write_fw(uncore, GEN6_RP_DOWN_TIMEOUT, 1000000);
+	intel_uncore_write_fw(uncore, GEN6_RP_UP_THRESHOLD, 59400);
+	intel_uncore_write_fw(uncore, GEN6_RP_DOWN_THRESHOLD, 245000);
+	intel_uncore_write_fw(uncore, GEN6_RP_UP_EI, 66000);
+	intel_uncore_write_fw(uncore, GEN6_RP_DOWN_EI, 350000);
+
+	intel_uncore_write_fw(uncore, GEN6_RP_IDLE_HYSTERSIS, 10);
+
+	/* 2: Enable RPS */
+	intel_uncore_write_fw(uncore, GEN6_RP_CONTROL,
+			      GEN6_RP_MEDIA_HW_NORMAL_MODE |
+			      GEN6_RP_MEDIA_IS_GFX |
+			      GEN6_RP_ENABLE |
+			      GEN6_RP_UP_BUSY_AVG |
+			      GEN6_RP_DOWN_IDLE_AVG);
+
+	rps->pm_events = (GEN6_PM_RP_UP_THRESHOLD |
+			  GEN6_PM_RP_DOWN_THRESHOLD |
+			  GEN6_PM_RP_DOWN_TIMEOUT);
+
+	/* Setting Fixed Bias */
+	vlv_punit_get(i915);
+
+	val = VLV_OVERRIDE_EN | VLV_SOC_TDP_EN | CHV_BIAS_CPU_50_SOC_50;
+	vlv_punit_write(i915, VLV_TURBO_SOC_OVERRIDE, val);
+
+	val = vlv_punit_read(i915, PUNIT_REG_GPU_FREQ_STS);
+
+	vlv_punit_put(i915);
+
+	/* RPS code assumes GPLL is used */
+	drm_WARN_ONCE(&i915->drm, (val & GPLLENABLE) == 0,
+		      "GPLL not enabled\n");
+
+	drm_dbg(&i915->drm, "GPLL enabled? %s\n",
+		str_yes_no(val & GPLLENABLE));
+	drm_dbg(&i915->drm, "GPU status: 0x%08x\n", val);
+
+	return rps_reset(rps);
+}
+
+static int vlv_rps_guar_freq(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val, rp1;
+
+	val = vlv_nc_read(i915, IOSF_NC_FB_GFX_FREQ_FUSE);
+
+	rp1 = val & FB_GFX_FGUARANTEED_FREQ_FUSE_MASK;
+	rp1 >>= FB_GFX_FGUARANTEED_FREQ_FUSE_SHIFT;
+
+	return rp1;
+}
+
+static int vlv_rps_max_freq(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val, rp0;
+
+	val = vlv_nc_read(i915, IOSF_NC_FB_GFX_FREQ_FUSE);
+
+	rp0 = (val & FB_GFX_MAX_FREQ_FUSE_MASK) >> FB_GFX_MAX_FREQ_FUSE_SHIFT;
+	/* Clamp to max */
+	rp0 = min_t(u32, rp0, 0xea);
+
+	return rp0;
+}
+
+static int vlv_rps_rpe_freq(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val, rpe;
+
+	val = vlv_nc_read(i915, IOSF_NC_FB_GFX_FMAX_FUSE_LO);
+	rpe = (val & FB_FMAX_VMIN_FREQ_LO_MASK) >> FB_FMAX_VMIN_FREQ_LO_SHIFT;
+	val = vlv_nc_read(i915, IOSF_NC_FB_GFX_FMAX_FUSE_HI);
+	rpe |= (val & FB_FMAX_VMIN_FREQ_HI_MASK) << 5;
+
+	return rpe;
+}
+
+static int vlv_rps_min_freq(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val;
+
+	val = vlv_punit_read(i915, PUNIT_REG_GPU_LFM) & 0xff;
+	/*
+	 * According to the BYT Punit GPU turbo HAS 1.1.6.3 the minimum value
+	 * for the minimum frequency in GPLL mode is 0xc1. Contrary to this on
+	 * a BYT-M B0 the above register contains 0xbf. Moreover when setting
+	 * a frequency Punit will not allow values below 0xc0. Clamp it 0xc0
+	 * to make sure it matches what Punit accepts.
+	 */
+	return max_t(u32, val, 0xc0);
+}
+
+static bool vlv_rps_enable(struct intel_rps *rps)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val;
+
+	intel_uncore_write_fw(uncore, GEN6_RP_DOWN_TIMEOUT, 1000000);
+	intel_uncore_write_fw(uncore, GEN6_RP_UP_THRESHOLD, 59400);
+	intel_uncore_write_fw(uncore, GEN6_RP_DOWN_THRESHOLD, 245000);
+	intel_uncore_write_fw(uncore, GEN6_RP_UP_EI, 66000);
+	intel_uncore_write_fw(uncore, GEN6_RP_DOWN_EI, 350000);
+
+	intel_uncore_write_fw(uncore, GEN6_RP_IDLE_HYSTERSIS, 10);
+
+	intel_uncore_write_fw(uncore, GEN6_RP_CONTROL,
+			      GEN6_RP_MEDIA_TURBO |
+			      GEN6_RP_MEDIA_HW_NORMAL_MODE |
+			      GEN6_RP_MEDIA_IS_GFX |
+			      GEN6_RP_ENABLE |
+			      GEN6_RP_UP_BUSY_AVG |
+			      GEN6_RP_DOWN_IDLE_CONT);
+
+	/* WaGsvRC0ResidencyMethod:vlv */
+	rps->pm_events = GEN6_PM_RP_UP_EI_EXPIRED;
+
+	vlv_punit_get(i915);
+
+	/* Setting Fixed Bias */
+	val = VLV_OVERRIDE_EN | VLV_SOC_TDP_EN | VLV_BIAS_CPU_125_SOC_875;
+	vlv_punit_write(i915, VLV_TURBO_SOC_OVERRIDE, val);
+
+	val = vlv_punit_read(i915, PUNIT_REG_GPU_FREQ_STS);
+
+	vlv_punit_put(i915);
+
+	/* RPS code assumes GPLL is used */
+	drm_WARN_ONCE(&i915->drm, (val & GPLLENABLE) == 0,
+		      "GPLL not enabled\n");
+
+	drm_dbg(&i915->drm, "GPLL enabled? %s\n",
+		str_yes_no(val & GPLLENABLE));
+	drm_dbg(&i915->drm, "GPU status: 0x%08x\n", val);
+
+	return rps_reset(rps);
+}
+
+static unsigned long __ips_gfx_val(struct intel_ips *ips)
+{
+	struct intel_rps *rps = container_of(ips, typeof(*rps), ips);
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	unsigned int t, state1, state2;
+	u32 pxvid, ext_v;
+	u64 corr, corr2;
+
+	lockdep_assert_held(&mchdev_lock);
+
+	pxvid = intel_uncore_read(uncore, PXVFREQ(rps->cur_freq));
+	pxvid = (pxvid >> 24) & 0x7f;
+	ext_v = pvid_to_extvid(rps_to_i915(rps), pxvid);
+
+	state1 = ext_v;
+
+	/* Revel in the empirically derived constants */
+
+	/* Correction factor in 1/100000 units */
+	t = ips_mch_val(uncore);
+	if (t > 80)
+		corr = t * 2349 + 135940;
+	else if (t >= 50)
+		corr = t * 964 + 29317;
+	else /* < 50 */
+		corr = t * 301 + 1004;
+
+	corr = div_u64(corr * 150142 * state1, 10000) - 78642;
+	corr2 = div_u64(corr, 100000) * ips->corr;
+
+	state2 = div_u64(corr2 * state1, 10000);
+	state2 /= 100; /* convert to mW */
+
+	__gen5_ips_update(ips);
+
+	return ips->gfx_power + state2;
+}
+
+static bool has_busy_stats(struct intel_rps *rps)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, rps_to_gt(rps), id) {
+		if (!intel_engine_supports_stats(engine))
+			return false;
+	}
+
+	return true;
+}
+
+void intel_rps_enable(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	bool enabled = false;
+
+	if (!HAS_RPS(i915))
+		return;
+
+	if (rps_uses_slpc(rps))
+		return;
+
+	intel_gt_check_clock_frequency(rps_to_gt(rps));
+
+	intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+	if (rps->max_freq <= rps->min_freq)
+		/* leave disabled, no room for dynamic reclocking */;
+	else if (IS_CHERRYVIEW(i915))
+		enabled = chv_rps_enable(rps);
+	else if (IS_VALLEYVIEW(i915))
+		enabled = vlv_rps_enable(rps);
+	else if (GRAPHICS_VER(i915) >= 9)
+		enabled = gen9_rps_enable(rps);
+	else if (GRAPHICS_VER(i915) >= 8)
+		enabled = gen8_rps_enable(rps);
+	else if (GRAPHICS_VER(i915) >= 6)
+		enabled = gen6_rps_enable(rps);
+	else if (IS_IRONLAKE_M(i915))
+		enabled = gen5_rps_enable(rps);
+	else
+		MISSING_CASE(GRAPHICS_VER(i915));
+	intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
+	if (!enabled)
+		return;
+
+	GT_TRACE(rps_to_gt(rps),
+		 "min:%x, max:%x, freq:[%d, %d]\n",
+		 rps->min_freq, rps->max_freq,
+		 intel_gpu_freq(rps, rps->min_freq),
+		 intel_gpu_freq(rps, rps->max_freq));
+
+	GEM_BUG_ON(rps->max_freq < rps->min_freq);
+	GEM_BUG_ON(rps->idle_freq > rps->max_freq);
+
+	GEM_BUG_ON(rps->efficient_freq < rps->min_freq);
+	GEM_BUG_ON(rps->efficient_freq > rps->max_freq);
+
+	if (has_busy_stats(rps))
+		intel_rps_set_timer(rps);
+	else if (GRAPHICS_VER(i915) >= 6 && GRAPHICS_VER(i915) <= 11)
+		intel_rps_set_interrupts(rps);
+	else
+		/* Ironlake currently uses intel_ips.ko */ {}
+
+	intel_rps_set_enabled(rps);
+}
+
+static void gen6_rps_disable(struct intel_rps *rps)
+{
+	set(rps_to_uncore(rps), GEN6_RP_CONTROL, 0);
+}
+
+void intel_rps_disable(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+
+	if (!intel_rps_is_enabled(rps))
+		return;
+
+	intel_rps_clear_enabled(rps);
+	intel_rps_clear_interrupts(rps);
+	intel_rps_clear_timer(rps);
+
+	if (GRAPHICS_VER(i915) >= 6)
+		gen6_rps_disable(rps);
+	else if (IS_IRONLAKE_M(i915))
+		gen5_rps_disable(rps);
+}
+
+static int byt_gpu_freq(struct intel_rps *rps, int val)
+{
+	/*
+	 * N = val - 0xb7
+	 * Slow = Fast = GPLL ref * N
+	 */
+	return DIV_ROUND_CLOSEST(rps->gpll_ref_freq * (val - 0xb7), 1000);
+}
+
+static int byt_freq_opcode(struct intel_rps *rps, int val)
+{
+	return DIV_ROUND_CLOSEST(1000 * val, rps->gpll_ref_freq) + 0xb7;
+}
+
+static int chv_gpu_freq(struct intel_rps *rps, int val)
+{
+	/*
+	 * N = val / 2
+	 * CU (slow) = CU2x (fast) / 2 = GPLL ref * N / 2
+	 */
+	return DIV_ROUND_CLOSEST(rps->gpll_ref_freq * val, 2 * 2 * 1000);
+}
+
+static int chv_freq_opcode(struct intel_rps *rps, int val)
+{
+	/* CHV needs even values */
+	return DIV_ROUND_CLOSEST(2 * 1000 * val, rps->gpll_ref_freq) * 2;
+}
+
+int intel_gpu_freq(struct intel_rps *rps, int val)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+
+	if (GRAPHICS_VER(i915) >= 9)
+		return DIV_ROUND_CLOSEST(val * GT_FREQUENCY_MULTIPLIER,
+					 GEN9_FREQ_SCALER);
+	else if (IS_CHERRYVIEW(i915))
+		return chv_gpu_freq(rps, val);
+	else if (IS_VALLEYVIEW(i915))
+		return byt_gpu_freq(rps, val);
+	else if (GRAPHICS_VER(i915) >= 6)
+		return val * GT_FREQUENCY_MULTIPLIER;
+	else
+		return val;
+}
+
+int intel_freq_opcode(struct intel_rps *rps, int val)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+
+	if (GRAPHICS_VER(i915) >= 9)
+		return DIV_ROUND_CLOSEST(val * GEN9_FREQ_SCALER,
+					 GT_FREQUENCY_MULTIPLIER);
+	else if (IS_CHERRYVIEW(i915))
+		return chv_freq_opcode(rps, val);
+	else if (IS_VALLEYVIEW(i915))
+		return byt_freq_opcode(rps, val);
+	else if (GRAPHICS_VER(i915) >= 6)
+		return DIV_ROUND_CLOSEST(val, GT_FREQUENCY_MULTIPLIER);
+	else
+		return val;
+}
+
+static void vlv_init_gpll_ref_freq(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+
+	rps->gpll_ref_freq =
+		vlv_get_cck_clock(i915, "GPLL ref",
+				  CCK_GPLL_CLOCK_CONTROL,
+				  i915->czclk_freq);
+
+	drm_dbg(&i915->drm, "GPLL reference freq: %d kHz\n",
+		rps->gpll_ref_freq);
+}
+
+static void vlv_rps_init(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val;
+
+	vlv_iosf_sb_get(i915,
+			BIT(VLV_IOSF_SB_PUNIT) |
+			BIT(VLV_IOSF_SB_NC) |
+			BIT(VLV_IOSF_SB_CCK));
+
+	vlv_init_gpll_ref_freq(rps);
+
+	val = vlv_punit_read(i915, PUNIT_REG_GPU_FREQ_STS);
+	switch ((val >> 6) & 3) {
+	case 0:
+	case 1:
+		i915->mem_freq = 800;
+		break;
+	case 2:
+		i915->mem_freq = 1066;
+		break;
+	case 3:
+		i915->mem_freq = 1333;
+		break;
+	}
+	drm_dbg(&i915->drm, "DDR speed: %d MHz\n", i915->mem_freq);
+
+	rps->max_freq = vlv_rps_max_freq(rps);
+	rps->rp0_freq = rps->max_freq;
+	drm_dbg(&i915->drm, "max GPU freq: %d MHz (%u)\n",
+		intel_gpu_freq(rps, rps->max_freq), rps->max_freq);
+
+	rps->efficient_freq = vlv_rps_rpe_freq(rps);
+	drm_dbg(&i915->drm, "RPe GPU freq: %d MHz (%u)\n",
+		intel_gpu_freq(rps, rps->efficient_freq), rps->efficient_freq);
+
+	rps->rp1_freq = vlv_rps_guar_freq(rps);
+	drm_dbg(&i915->drm, "RP1(Guar Freq) GPU freq: %d MHz (%u)\n",
+		intel_gpu_freq(rps, rps->rp1_freq), rps->rp1_freq);
+
+	rps->min_freq = vlv_rps_min_freq(rps);
+	drm_dbg(&i915->drm, "min GPU freq: %d MHz (%u)\n",
+		intel_gpu_freq(rps, rps->min_freq), rps->min_freq);
+
+	vlv_iosf_sb_put(i915,
+			BIT(VLV_IOSF_SB_PUNIT) |
+			BIT(VLV_IOSF_SB_NC) |
+			BIT(VLV_IOSF_SB_CCK));
+}
+
+static void chv_rps_init(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 val;
+
+	vlv_iosf_sb_get(i915,
+			BIT(VLV_IOSF_SB_PUNIT) |
+			BIT(VLV_IOSF_SB_NC) |
+			BIT(VLV_IOSF_SB_CCK));
+
+	vlv_init_gpll_ref_freq(rps);
+
+	val = vlv_cck_read(i915, CCK_FUSE_REG);
+
+	switch ((val >> 2) & 0x7) {
+	case 3:
+		i915->mem_freq = 2000;
+		break;
+	default:
+		i915->mem_freq = 1600;
+		break;
+	}
+	drm_dbg(&i915->drm, "DDR speed: %d MHz\n", i915->mem_freq);
+
+	rps->max_freq = chv_rps_max_freq(rps);
+	rps->rp0_freq = rps->max_freq;
+	drm_dbg(&i915->drm, "max GPU freq: %d MHz (%u)\n",
+		intel_gpu_freq(rps, rps->max_freq), rps->max_freq);
+
+	rps->efficient_freq = chv_rps_rpe_freq(rps);
+	drm_dbg(&i915->drm, "RPe GPU freq: %d MHz (%u)\n",
+		intel_gpu_freq(rps, rps->efficient_freq), rps->efficient_freq);
+
+	rps->rp1_freq = chv_rps_guar_freq(rps);
+	drm_dbg(&i915->drm, "RP1(Guar) GPU freq: %d MHz (%u)\n",
+		intel_gpu_freq(rps, rps->rp1_freq), rps->rp1_freq);
+
+	rps->min_freq = chv_rps_min_freq(rps);
+	drm_dbg(&i915->drm, "min GPU freq: %d MHz (%u)\n",
+		intel_gpu_freq(rps, rps->min_freq), rps->min_freq);
+
+	vlv_iosf_sb_put(i915,
+			BIT(VLV_IOSF_SB_PUNIT) |
+			BIT(VLV_IOSF_SB_NC) |
+			BIT(VLV_IOSF_SB_CCK));
+
+	drm_WARN_ONCE(&i915->drm, (rps->max_freq | rps->efficient_freq |
+				   rps->rp1_freq | rps->min_freq) & 1,
+		      "Odd GPU freq values\n");
+}
+
+static void vlv_c0_read(struct intel_uncore *uncore, struct intel_rps_ei *ei)
+{
+	ei->ktime = ktime_get_raw();
+	ei->render_c0 = intel_uncore_read(uncore, VLV_RENDER_C0_COUNT);
+	ei->media_c0 = intel_uncore_read(uncore, VLV_MEDIA_C0_COUNT);
+}
+
+static u32 vlv_wa_c0_ei(struct intel_rps *rps, u32 pm_iir)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	const struct intel_rps_ei *prev = &rps->ei;
+	struct intel_rps_ei now;
+	u32 events = 0;
+
+	if ((pm_iir & GEN6_PM_RP_UP_EI_EXPIRED) == 0)
+		return 0;
+
+	vlv_c0_read(uncore, &now);
+
+	if (prev->ktime) {
+		u64 time, c0;
+		u32 render, media;
+
+		time = ktime_us_delta(now.ktime, prev->ktime);
+
+		time *= rps_to_i915(rps)->czclk_freq;
+
+		/* Workload can be split between render + media,
+		 * e.g. SwapBuffers being blitted in X after being rendered in
+		 * mesa. To account for this we need to combine both engines
+		 * into our activity counter.
+		 */
+		render = now.render_c0 - prev->render_c0;
+		media = now.media_c0 - prev->media_c0;
+		c0 = max(render, media);
+		c0 *= 1000 * 100 << 8; /* to usecs and scale to threshold% */
+
+		if (c0 > time * rps->power.up_threshold)
+			events = GEN6_PM_RP_UP_THRESHOLD;
+		else if (c0 < time * rps->power.down_threshold)
+			events = GEN6_PM_RP_DOWN_THRESHOLD;
+	}
+
+	rps->ei = now;
+	return events;
+}
+
+static void rps_work(struct work_struct *work)
+{
+	struct intel_rps *rps = container_of(work, typeof(*rps), work);
+	struct intel_gt *gt = rps_to_gt(rps);
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	bool client_boost = false;
+	int new_freq, adj, min, max;
+	u32 pm_iir = 0;
+
+	spin_lock_irq(gt->irq_lock);
+	pm_iir = fetch_and_zero(&rps->pm_iir) & rps->pm_events;
+	client_boost = atomic_read(&rps->num_waiters);
+	spin_unlock_irq(gt->irq_lock);
+
+	/* Make sure we didn't queue anything we're not going to process. */
+	if (!pm_iir && !client_boost)
+		goto out;
+
+	mutex_lock(&rps->lock);
+	if (!intel_rps_is_active(rps)) {
+		mutex_unlock(&rps->lock);
+		return;
+	}
+
+	pm_iir |= vlv_wa_c0_ei(rps, pm_iir);
+
+	adj = rps->last_adj;
+	new_freq = rps->cur_freq;
+	min = rps->min_freq_softlimit;
+	max = rps->max_freq_softlimit;
+	if (client_boost)
+		max = rps->max_freq;
+
+	GT_TRACE(gt,
+		 "pm_iir:%x, client_boost:%s, last:%d, cur:%x, min:%x, max:%x\n",
+		 pm_iir, str_yes_no(client_boost),
+		 adj, new_freq, min, max);
+
+	if (client_boost && new_freq < rps->boost_freq) {
+		new_freq = rps->boost_freq;
+		adj = 0;
+	} else if (pm_iir & GEN6_PM_RP_UP_THRESHOLD) {
+		if (adj > 0)
+			adj *= 2;
+		else /* CHV needs even encode values */
+			adj = IS_CHERRYVIEW(gt->i915) ? 2 : 1;
+
+		if (new_freq >= rps->max_freq_softlimit)
+			adj = 0;
+	} else if (client_boost) {
+		adj = 0;
+	} else if (pm_iir & GEN6_PM_RP_DOWN_TIMEOUT) {
+		if (rps->cur_freq > rps->efficient_freq)
+			new_freq = rps->efficient_freq;
+		else if (rps->cur_freq > rps->min_freq_softlimit)
+			new_freq = rps->min_freq_softlimit;
+		adj = 0;
+	} else if (pm_iir & GEN6_PM_RP_DOWN_THRESHOLD) {
+		if (adj < 0)
+			adj *= 2;
+		else /* CHV needs even encode values */
+			adj = IS_CHERRYVIEW(gt->i915) ? -2 : -1;
+
+		if (new_freq <= rps->min_freq_softlimit)
+			adj = 0;
+	} else { /* unknown event */
+		adj = 0;
+	}
+
+	/*
+	 * sysfs frequency limits may have snuck in while
+	 * servicing the interrupt
+	 */
+	new_freq += adj;
+	new_freq = clamp_t(int, new_freq, min, max);
+
+	if (intel_rps_set(rps, new_freq)) {
+		drm_dbg(&i915->drm, "Failed to set new GPU frequency\n");
+		adj = 0;
+	}
+	rps->last_adj = adj;
+
+	mutex_unlock(&rps->lock);
+
+out:
+	spin_lock_irq(gt->irq_lock);
+	gen6_gt_pm_unmask_irq(gt, rps->pm_events);
+	spin_unlock_irq(gt->irq_lock);
+}
+
+void gen11_rps_irq_handler(struct intel_rps *rps, u32 pm_iir)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+	const u32 events = rps->pm_events & pm_iir;
+
+	lockdep_assert_held(gt->irq_lock);
+
+	if (unlikely(!events))
+		return;
+
+	GT_TRACE(gt, "irq events:%x\n", events);
+
+	gen6_gt_pm_mask_irq(gt, events);
+
+	rps->pm_iir |= events;
+	schedule_work(&rps->work);
+}
+
+void gen6_rps_irq_handler(struct intel_rps *rps, u32 pm_iir)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+	u32 events;
+
+	events = pm_iir & rps->pm_events;
+	if (events) {
+		spin_lock(gt->irq_lock);
+
+		GT_TRACE(gt, "irq events:%x\n", events);
+
+		gen6_gt_pm_mask_irq(gt, events);
+		rps->pm_iir |= events;
+
+		schedule_work(&rps->work);
+		spin_unlock(gt->irq_lock);
+	}
+
+	if (GRAPHICS_VER(gt->i915) >= 8)
+		return;
+
+	if (pm_iir & PM_VEBOX_USER_INTERRUPT)
+		intel_engine_cs_irq(gt->engine[VECS0], pm_iir >> 10);
+
+	if (pm_iir & PM_VEBOX_CS_ERROR_INTERRUPT)
+		DRM_DEBUG("Command parser error, pm_iir 0x%08x\n", pm_iir);
+}
+
+void gen5_rps_irq_handler(struct intel_rps *rps)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	u32 busy_up, busy_down, max_avg, min_avg;
+	u8 new_freq;
+
+	spin_lock(&mchdev_lock);
+
+	intel_uncore_write16(uncore,
+			     MEMINTRSTS,
+			     intel_uncore_read(uncore, MEMINTRSTS));
+
+	intel_uncore_write16(uncore, MEMINTRSTS, MEMINT_EVAL_CHG);
+	busy_up = intel_uncore_read(uncore, RCPREVBSYTUPAVG);
+	busy_down = intel_uncore_read(uncore, RCPREVBSYTDNAVG);
+	max_avg = intel_uncore_read(uncore, RCBMAXAVG);
+	min_avg = intel_uncore_read(uncore, RCBMINAVG);
+
+	/* Handle RCS change request from hw */
+	new_freq = rps->cur_freq;
+	if (busy_up > max_avg)
+		new_freq++;
+	else if (busy_down < min_avg)
+		new_freq--;
+	new_freq = clamp(new_freq,
+			 rps->min_freq_softlimit,
+			 rps->max_freq_softlimit);
+
+	if (new_freq != rps->cur_freq && !__gen5_rps_set(rps, new_freq))
+		rps->cur_freq = new_freq;
+
+	spin_unlock(&mchdev_lock);
+}
+
+void intel_rps_init_early(struct intel_rps *rps)
+{
+	mutex_init(&rps->lock);
+	mutex_init(&rps->power.mutex);
+
+	INIT_WORK(&rps->work, rps_work);
+	timer_setup(&rps->timer, rps_timer, 0);
+
+	atomic_set(&rps->num_waiters, 0);
+}
+
+void intel_rps_init(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+
+	if (rps_uses_slpc(rps))
+		return;
+
+	if (IS_CHERRYVIEW(i915))
+		chv_rps_init(rps);
+	else if (IS_VALLEYVIEW(i915))
+		vlv_rps_init(rps);
+	else if (GRAPHICS_VER(i915) >= 6)
+		gen6_rps_init(rps);
+	else if (IS_IRONLAKE_M(i915))
+		gen5_rps_init(rps);
+
+	/* Derive initial user preferences/limits from the hardware limits */
+	rps->max_freq_softlimit = rps->max_freq;
+	rps_to_gt(rps)->defaults.max_freq = rps->max_freq_softlimit;
+	rps->min_freq_softlimit = rps->min_freq;
+	rps_to_gt(rps)->defaults.min_freq = rps->min_freq_softlimit;
+
+	/* After setting max-softlimit, find the overclock max freq */
+	if (GRAPHICS_VER(i915) == 6 || IS_IVYBRIDGE(i915) || IS_HASWELL(i915)) {
+		u32 params = 0;
+
+		snb_pcode_read(rps_to_gt(rps)->uncore, GEN6_READ_OC_PARAMS, &params, NULL);
+		if (params & BIT(31)) { /* OC supported */
+			drm_dbg(&i915->drm,
+				"Overclocking supported, max: %dMHz, overclock: %dMHz\n",
+				(rps->max_freq & 0xff) * 50,
+				(params & 0xff) * 50);
+			rps->max_freq = params & 0xff;
+		}
+	}
+
+	/* Finally allow us to boost to max by default */
+	rps->boost_freq = rps->max_freq;
+	rps->idle_freq = rps->min_freq;
+
+	/* Start in the middle, from here we will autotune based on workload */
+	rps->cur_freq = rps->efficient_freq;
+
+	rps->pm_intrmsk_mbz = 0;
+
+	/*
+	 * SNB,IVB,HSW can while VLV,CHV may hard hang on looping batchbuffer
+	 * if GEN6_PM_UP_EI_EXPIRED is masked.
+	 *
+	 * TODO: verify if this can be reproduced on VLV,CHV.
+	 */
+	if (GRAPHICS_VER(i915) <= 7)
+		rps->pm_intrmsk_mbz |= GEN6_PM_RP_UP_EI_EXPIRED;
+
+	if (GRAPHICS_VER(i915) >= 8 && GRAPHICS_VER(i915) < 11)
+		rps->pm_intrmsk_mbz |= GEN8_PMINTR_DISABLE_REDIRECT_TO_GUC;
+
+	/* GuC needs ARAT expired interrupt unmasked */
+	if (intel_uc_uses_guc_submission(&rps_to_gt(rps)->uc))
+		rps->pm_intrmsk_mbz |= ARAT_EXPIRED_INTRMSK;
+}
+
+void intel_rps_sanitize(struct intel_rps *rps)
+{
+	if (rps_uses_slpc(rps))
+		return;
+
+	if (GRAPHICS_VER(rps_to_i915(rps)) >= 6)
+		rps_disable_interrupts(rps);
+}
+
+u32 intel_rps_get_cagf(struct intel_rps *rps, u32 rpstat)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	u32 cagf;
+
+	if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+		cagf = (rpstat >> 8) & 0xff;
+	else if (GRAPHICS_VER(i915) >= 9)
+		cagf = (rpstat & GEN9_CAGF_MASK) >> GEN9_CAGF_SHIFT;
+	else if (IS_HASWELL(i915) || IS_BROADWELL(i915))
+		cagf = (rpstat & HSW_CAGF_MASK) >> HSW_CAGF_SHIFT;
+	else if (GRAPHICS_VER(i915) >= 6)
+		cagf = (rpstat & GEN6_CAGF_MASK) >> GEN6_CAGF_SHIFT;
+	else
+		cagf = gen5_invert_freq(rps, (rpstat & MEMSTAT_PSTATE_MASK) >>
+					MEMSTAT_PSTATE_SHIFT);
+
+	return cagf;
+}
+
+static u32 read_cagf(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	u32 freq;
+
+	if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) {
+		vlv_punit_get(i915);
+		freq = vlv_punit_read(i915, PUNIT_REG_GPU_FREQ_STS);
+		vlv_punit_put(i915);
+	} else if (GRAPHICS_VER(i915) >= 6) {
+		freq = intel_uncore_read(uncore, GEN6_RPSTAT1);
+	} else {
+		freq = intel_uncore_read(uncore, MEMSTAT_ILK);
+	}
+
+	return intel_rps_get_cagf(rps, freq);
+}
+
+u32 intel_rps_read_actual_frequency(struct intel_rps *rps)
+{
+	struct intel_runtime_pm *rpm = rps_to_uncore(rps)->rpm;
+	intel_wakeref_t wakeref;
+	u32 freq = 0;
+
+	with_intel_runtime_pm_if_in_use(rpm, wakeref)
+		freq = intel_gpu_freq(rps, read_cagf(rps));
+
+	return freq;
+}
+
+u32 intel_rps_read_punit_req(struct intel_rps *rps)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	struct intel_runtime_pm *rpm = rps_to_uncore(rps)->rpm;
+	intel_wakeref_t wakeref;
+	u32 freq = 0;
+
+	with_intel_runtime_pm_if_in_use(rpm, wakeref)
+		freq = intel_uncore_read(uncore, GEN6_RPNSWREQ);
+
+	return freq;
+}
+
+static u32 intel_rps_get_req(u32 pureq)
+{
+	u32 req = pureq >> GEN9_SW_REQ_UNSLICE_RATIO_SHIFT;
+
+	return req;
+}
+
+u32 intel_rps_read_punit_req_frequency(struct intel_rps *rps)
+{
+	u32 freq = intel_rps_get_req(intel_rps_read_punit_req(rps));
+
+	return intel_gpu_freq(rps, freq);
+}
+
+u32 intel_rps_get_requested_frequency(struct intel_rps *rps)
+{
+	if (rps_uses_slpc(rps))
+		return intel_rps_read_punit_req_frequency(rps);
+	else
+		return intel_gpu_freq(rps, rps->cur_freq);
+}
+
+u32 intel_rps_get_max_frequency(struct intel_rps *rps)
+{
+	struct intel_guc_slpc *slpc = rps_to_slpc(rps);
+
+	if (rps_uses_slpc(rps))
+		return slpc->max_freq_softlimit;
+	else
+		return intel_gpu_freq(rps, rps->max_freq_softlimit);
+}
+
+/**
+ * intel_rps_get_max_raw_freq - returns the max frequency in some raw format.
+ * @rps: the intel_rps structure
+ *
+ * Returns the max frequency in a raw format. In newer platforms raw is in
+ * units of 50 MHz.
+ */
+u32 intel_rps_get_max_raw_freq(struct intel_rps *rps)
+{
+	struct intel_guc_slpc *slpc = rps_to_slpc(rps);
+	u32 freq;
+
+	if (rps_uses_slpc(rps)) {
+		return DIV_ROUND_CLOSEST(slpc->rp0_freq,
+					 GT_FREQUENCY_MULTIPLIER);
+	} else {
+		freq = rps->max_freq;
+		if (GRAPHICS_VER(rps_to_i915(rps)) >= 9) {
+			/* Convert GT frequency to 50 MHz units */
+			freq /= GEN9_FREQ_SCALER;
+		}
+		return freq;
+	}
+}
+
+u32 intel_rps_get_rp0_frequency(struct intel_rps *rps)
+{
+	struct intel_guc_slpc *slpc = rps_to_slpc(rps);
+
+	if (rps_uses_slpc(rps))
+		return slpc->rp0_freq;
+	else
+		return intel_gpu_freq(rps, rps->rp0_freq);
+}
+
+u32 intel_rps_get_rp1_frequency(struct intel_rps *rps)
+{
+	struct intel_guc_slpc *slpc = rps_to_slpc(rps);
+
+	if (rps_uses_slpc(rps))
+		return slpc->rp1_freq;
+	else
+		return intel_gpu_freq(rps, rps->rp1_freq);
+}
+
+u32 intel_rps_get_rpn_frequency(struct intel_rps *rps)
+{
+	struct intel_guc_slpc *slpc = rps_to_slpc(rps);
+
+	if (rps_uses_slpc(rps))
+		return slpc->min_freq;
+	else
+		return intel_gpu_freq(rps, rps->min_freq);
+}
+
+static int set_max_freq(struct intel_rps *rps, u32 val)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	int ret = 0;
+
+	mutex_lock(&rps->lock);
+
+	val = intel_freq_opcode(rps, val);
+	if (val < rps->min_freq ||
+	    val > rps->max_freq ||
+	    val < rps->min_freq_softlimit) {
+		ret = -EINVAL;
+		goto unlock;
+	}
+
+	if (val > rps->rp0_freq)
+		drm_dbg(&i915->drm, "User requested overclocking to %d\n",
+			intel_gpu_freq(rps, val));
+
+	rps->max_freq_softlimit = val;
+
+	val = clamp_t(int, rps->cur_freq,
+		      rps->min_freq_softlimit,
+		      rps->max_freq_softlimit);
+
+	/*
+	 * We still need *_set_rps to process the new max_delay and
+	 * update the interrupt limits and PMINTRMSK even though
+	 * frequency request may be unchanged.
+	 */
+	intel_rps_set(rps, val);
+
+unlock:
+	mutex_unlock(&rps->lock);
+
+	return ret;
+}
+
+int intel_rps_set_max_frequency(struct intel_rps *rps, u32 val)
+{
+	struct intel_guc_slpc *slpc = rps_to_slpc(rps);
+
+	if (rps_uses_slpc(rps))
+		return intel_guc_slpc_set_max_freq(slpc, val);
+	else
+		return set_max_freq(rps, val);
+}
+
+u32 intel_rps_get_min_frequency(struct intel_rps *rps)
+{
+	struct intel_guc_slpc *slpc = rps_to_slpc(rps);
+
+	if (rps_uses_slpc(rps))
+		return slpc->min_freq_softlimit;
+	else
+		return intel_gpu_freq(rps, rps->min_freq_softlimit);
+}
+
+/**
+ * intel_rps_get_min_raw_freq - returns the min frequency in some raw format.
+ * @rps: the intel_rps structure
+ *
+ * Returns the min frequency in a raw format. In newer platforms raw is in
+ * units of 50 MHz.
+ */
+u32 intel_rps_get_min_raw_freq(struct intel_rps *rps)
+{
+	struct intel_guc_slpc *slpc = rps_to_slpc(rps);
+	u32 freq;
+
+	if (rps_uses_slpc(rps)) {
+		return DIV_ROUND_CLOSEST(slpc->min_freq,
+					 GT_FREQUENCY_MULTIPLIER);
+	} else {
+		freq = rps->min_freq;
+		if (GRAPHICS_VER(rps_to_i915(rps)) >= 9) {
+			/* Convert GT frequency to 50 MHz units */
+			freq /= GEN9_FREQ_SCALER;
+		}
+		return freq;
+	}
+}
+
+static int set_min_freq(struct intel_rps *rps, u32 val)
+{
+	int ret = 0;
+
+	mutex_lock(&rps->lock);
+
+	val = intel_freq_opcode(rps, val);
+	if (val < rps->min_freq ||
+	    val > rps->max_freq ||
+	    val > rps->max_freq_softlimit) {
+		ret = -EINVAL;
+		goto unlock;
+	}
+
+	rps->min_freq_softlimit = val;
+
+	val = clamp_t(int, rps->cur_freq,
+		      rps->min_freq_softlimit,
+		      rps->max_freq_softlimit);
+
+	/*
+	 * We still need *_set_rps to process the new min_delay and
+	 * update the interrupt limits and PMINTRMSK even though
+	 * frequency request may be unchanged.
+	 */
+	intel_rps_set(rps, val);
+
+unlock:
+	mutex_unlock(&rps->lock);
+
+	return ret;
+}
+
+int intel_rps_set_min_frequency(struct intel_rps *rps, u32 val)
+{
+	struct intel_guc_slpc *slpc = rps_to_slpc(rps);
+
+	if (rps_uses_slpc(rps))
+		return intel_guc_slpc_set_min_freq(slpc, val);
+	else
+		return set_min_freq(rps, val);
+}
+
+static void intel_rps_set_manual(struct intel_rps *rps, bool enable)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+	u32 state = enable ? GEN9_RPSWCTL_ENABLE : GEN9_RPSWCTL_DISABLE;
+
+	/* Allow punit to process software requests */
+	intel_uncore_write(uncore, GEN6_RP_CONTROL, state);
+}
+
+void intel_rps_raise_unslice(struct intel_rps *rps)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+
+	mutex_lock(&rps->lock);
+
+	if (rps_uses_slpc(rps)) {
+		/* RP limits have not been initialized yet for SLPC path */
+		struct intel_rps_freq_caps caps;
+
+		gen6_rps_get_freq_caps(rps, &caps);
+
+		intel_rps_set_manual(rps, true);
+		intel_uncore_write(uncore, GEN6_RPNSWREQ,
+				   ((caps.rp0_freq <<
+				   GEN9_SW_REQ_UNSLICE_RATIO_SHIFT) |
+				   GEN9_IGNORE_SLICE_RATIO));
+		intel_rps_set_manual(rps, false);
+	} else {
+		intel_rps_set(rps, rps->rp0_freq);
+	}
+
+	mutex_unlock(&rps->lock);
+}
+
+void intel_rps_lower_unslice(struct intel_rps *rps)
+{
+	struct intel_uncore *uncore = rps_to_uncore(rps);
+
+	mutex_lock(&rps->lock);
+
+	if (rps_uses_slpc(rps)) {
+		/* RP limits have not been initialized yet for SLPC path */
+		struct intel_rps_freq_caps caps;
+
+		gen6_rps_get_freq_caps(rps, &caps);
+
+		intel_rps_set_manual(rps, true);
+		intel_uncore_write(uncore, GEN6_RPNSWREQ,
+				   ((caps.min_freq <<
+				   GEN9_SW_REQ_UNSLICE_RATIO_SHIFT) |
+				   GEN9_IGNORE_SLICE_RATIO));
+		intel_rps_set_manual(rps, false);
+	} else {
+		intel_rps_set(rps, rps->min_freq);
+	}
+
+	mutex_unlock(&rps->lock);
+}
+
+static u32 rps_read_mmio(struct intel_rps *rps, i915_reg_t reg32)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+	intel_wakeref_t wakeref;
+	u32 val;
+
+	with_intel_runtime_pm(gt->uncore->rpm, wakeref)
+		val = intel_uncore_read(gt->uncore, reg32);
+
+	return val;
+}
+
+bool rps_read_mask_mmio(struct intel_rps *rps,
+			i915_reg_t reg32, u32 mask)
+{
+	return rps_read_mmio(rps, reg32) & mask;
+}
+
+/* External interface for intel_ips.ko */
+
+static struct drm_i915_private __rcu *ips_mchdev;
+
+/**
+ * Tells the intel_ips driver that the i915 driver is now loaded, if
+ * IPS got loaded first.
+ *
+ * This awkward dance is so that neither module has to depend on the
+ * other in order for IPS to do the appropriate communication of
+ * GPU turbo limits to i915.
+ */
+static void
+ips_ping_for_i915_load(void)
+{
+	void (*link)(void);
+
+	link = symbol_get(ips_link_to_i915_driver);
+	if (link) {
+		link();
+		symbol_put(ips_link_to_i915_driver);
+	}
+}
+
+void intel_rps_driver_register(struct intel_rps *rps)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+
+	/*
+	 * We only register the i915 ips part with intel-ips once everything is
+	 * set up, to avoid intel-ips sneaking in and reading bogus values.
+	 */
+	if (GRAPHICS_VER(gt->i915) == 5) {
+		GEM_BUG_ON(ips_mchdev);
+		rcu_assign_pointer(ips_mchdev, gt->i915);
+		ips_ping_for_i915_load();
+	}
+}
+
+void intel_rps_driver_unregister(struct intel_rps *rps)
+{
+	if (rcu_access_pointer(ips_mchdev) == rps_to_i915(rps))
+		rcu_assign_pointer(ips_mchdev, NULL);
+}
+
+static struct drm_i915_private *mchdev_get(void)
+{
+	struct drm_i915_private *i915;
+
+	rcu_read_lock();
+	i915 = rcu_dereference(ips_mchdev);
+	if (i915 && !kref_get_unless_zero(&i915->drm.ref))
+		i915 = NULL;
+	rcu_read_unlock();
+
+	return i915;
+}
+
+/**
+ * i915_read_mch_val - return value for IPS use
+ *
+ * Calculate and return a value for the IPS driver to use when deciding whether
+ * we have thermal and power headroom to increase CPU or GPU power budget.
+ */
+unsigned long i915_read_mch_val(void)
+{
+	struct drm_i915_private *i915;
+	unsigned long chipset_val = 0;
+	unsigned long graphics_val = 0;
+	intel_wakeref_t wakeref;
+
+	i915 = mchdev_get();
+	if (!i915)
+		return 0;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		struct intel_ips *ips = &to_gt(i915)->rps.ips;
+
+		spin_lock_irq(&mchdev_lock);
+		chipset_val = __ips_chipset_val(ips);
+		graphics_val = __ips_gfx_val(ips);
+		spin_unlock_irq(&mchdev_lock);
+	}
+
+	drm_dev_put(&i915->drm);
+	return chipset_val + graphics_val;
+}
+EXPORT_SYMBOL_GPL(i915_read_mch_val);
+
+/**
+ * i915_gpu_raise - raise GPU frequency limit
+ *
+ * Raise the limit; IPS indicates we have thermal headroom.
+ */
+bool i915_gpu_raise(void)
+{
+	struct drm_i915_private *i915;
+	struct intel_rps *rps;
+
+	i915 = mchdev_get();
+	if (!i915)
+		return false;
+
+	rps = &to_gt(i915)->rps;
+
+	spin_lock_irq(&mchdev_lock);
+	if (rps->max_freq_softlimit < rps->max_freq)
+		rps->max_freq_softlimit++;
+	spin_unlock_irq(&mchdev_lock);
+
+	drm_dev_put(&i915->drm);
+	return true;
+}
+EXPORT_SYMBOL_GPL(i915_gpu_raise);
+
+/**
+ * i915_gpu_lower - lower GPU frequency limit
+ *
+ * IPS indicates we're close to a thermal limit, so throttle back the GPU
+ * frequency maximum.
+ */
+bool i915_gpu_lower(void)
+{
+	struct drm_i915_private *i915;
+	struct intel_rps *rps;
+
+	i915 = mchdev_get();
+	if (!i915)
+		return false;
+
+	rps = &to_gt(i915)->rps;
+
+	spin_lock_irq(&mchdev_lock);
+	if (rps->max_freq_softlimit > rps->min_freq)
+		rps->max_freq_softlimit--;
+	spin_unlock_irq(&mchdev_lock);
+
+	drm_dev_put(&i915->drm);
+	return true;
+}
+EXPORT_SYMBOL_GPL(i915_gpu_lower);
+
+/**
+ * i915_gpu_busy - indicate GPU business to IPS
+ *
+ * Tell the IPS driver whether or not the GPU is busy.
+ */
+bool i915_gpu_busy(void)
+{
+	struct drm_i915_private *i915;
+	bool ret;
+
+	i915 = mchdev_get();
+	if (!i915)
+		return false;
+
+	ret = to_gt(i915)->awake;
+
+	drm_dev_put(&i915->drm);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(i915_gpu_busy);
+
+/**
+ * i915_gpu_turbo_disable - disable graphics turbo
+ *
+ * Disable graphics turbo by resetting the max frequency and setting the
+ * current frequency to the default.
+ */
+bool i915_gpu_turbo_disable(void)
+{
+	struct drm_i915_private *i915;
+	struct intel_rps *rps;
+	bool ret;
+
+	i915 = mchdev_get();
+	if (!i915)
+		return false;
+
+	rps = &to_gt(i915)->rps;
+
+	spin_lock_irq(&mchdev_lock);
+	rps->max_freq_softlimit = rps->min_freq;
+	ret = !__gen5_rps_set(&to_gt(i915)->rps, rps->min_freq);
+	spin_unlock_irq(&mchdev_lock);
+
+	drm_dev_put(&i915->drm);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(i915_gpu_turbo_disable);
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_rps.c"
+#include "selftest_slpc.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_rps.h b/drivers/gpu/drm/i915/gt/intel_rps.h
new file mode 100644
index 000000000..4509dfdc5
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_rps.h
@@ -0,0 +1,121 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_RPS_H
+#define INTEL_RPS_H
+
+#include "intel_rps_types.h"
+#include "i915_reg_defs.h"
+
+struct i915_request;
+
+void intel_rps_init_early(struct intel_rps *rps);
+void intel_rps_init(struct intel_rps *rps);
+void intel_rps_sanitize(struct intel_rps *rps);
+
+void intel_rps_driver_register(struct intel_rps *rps);
+void intel_rps_driver_unregister(struct intel_rps *rps);
+
+void intel_rps_enable(struct intel_rps *rps);
+void intel_rps_disable(struct intel_rps *rps);
+
+void intel_rps_park(struct intel_rps *rps);
+void intel_rps_unpark(struct intel_rps *rps);
+void intel_rps_boost(struct i915_request *rq);
+void intel_rps_dec_waiters(struct intel_rps *rps);
+u32 intel_rps_get_boost_frequency(struct intel_rps *rps);
+int intel_rps_set_boost_frequency(struct intel_rps *rps, u32 freq);
+
+int intel_rps_set(struct intel_rps *rps, u8 val);
+void intel_rps_mark_interactive(struct intel_rps *rps, bool interactive);
+
+int intel_gpu_freq(struct intel_rps *rps, int val);
+int intel_freq_opcode(struct intel_rps *rps, int val);
+u32 intel_rps_get_cagf(struct intel_rps *rps, u32 rpstat1);
+u32 intel_rps_read_actual_frequency(struct intel_rps *rps);
+u32 intel_rps_get_requested_frequency(struct intel_rps *rps);
+u32 intel_rps_get_min_frequency(struct intel_rps *rps);
+u32 intel_rps_get_min_raw_freq(struct intel_rps *rps);
+int intel_rps_set_min_frequency(struct intel_rps *rps, u32 val);
+u32 intel_rps_get_max_frequency(struct intel_rps *rps);
+u32 intel_rps_get_max_raw_freq(struct intel_rps *rps);
+int intel_rps_set_max_frequency(struct intel_rps *rps, u32 val);
+u32 intel_rps_get_rp0_frequency(struct intel_rps *rps);
+u32 intel_rps_get_rp1_frequency(struct intel_rps *rps);
+u32 intel_rps_get_rpn_frequency(struct intel_rps *rps);
+u32 intel_rps_read_punit_req(struct intel_rps *rps);
+u32 intel_rps_read_punit_req_frequency(struct intel_rps *rps);
+void gen6_rps_get_freq_caps(struct intel_rps *rps, struct intel_rps_freq_caps *caps);
+void intel_rps_raise_unslice(struct intel_rps *rps);
+void intel_rps_lower_unslice(struct intel_rps *rps);
+
+u32 intel_rps_read_throttle_reason(struct intel_rps *rps);
+bool rps_read_mask_mmio(struct intel_rps *rps, i915_reg_t reg32, u32 mask);
+
+void gen5_rps_irq_handler(struct intel_rps *rps);
+void gen6_rps_irq_handler(struct intel_rps *rps, u32 pm_iir);
+void gen11_rps_irq_handler(struct intel_rps *rps, u32 pm_iir);
+
+static inline bool intel_rps_is_enabled(const struct intel_rps *rps)
+{
+	return test_bit(INTEL_RPS_ENABLED, &rps->flags);
+}
+
+static inline void intel_rps_set_enabled(struct intel_rps *rps)
+{
+	set_bit(INTEL_RPS_ENABLED, &rps->flags);
+}
+
+static inline void intel_rps_clear_enabled(struct intel_rps *rps)
+{
+	clear_bit(INTEL_RPS_ENABLED, &rps->flags);
+}
+
+static inline bool intel_rps_is_active(const struct intel_rps *rps)
+{
+	return test_bit(INTEL_RPS_ACTIVE, &rps->flags);
+}
+
+static inline void intel_rps_set_active(struct intel_rps *rps)
+{
+	set_bit(INTEL_RPS_ACTIVE, &rps->flags);
+}
+
+static inline bool intel_rps_clear_active(struct intel_rps *rps)
+{
+	return test_and_clear_bit(INTEL_RPS_ACTIVE, &rps->flags);
+}
+
+static inline bool intel_rps_has_interrupts(const struct intel_rps *rps)
+{
+	return test_bit(INTEL_RPS_INTERRUPTS, &rps->flags);
+}
+
+static inline void intel_rps_set_interrupts(struct intel_rps *rps)
+{
+	set_bit(INTEL_RPS_INTERRUPTS, &rps->flags);
+}
+
+static inline void intel_rps_clear_interrupts(struct intel_rps *rps)
+{
+	clear_bit(INTEL_RPS_INTERRUPTS, &rps->flags);
+}
+
+static inline bool intel_rps_uses_timer(const struct intel_rps *rps)
+{
+	return test_bit(INTEL_RPS_TIMER, &rps->flags);
+}
+
+static inline void intel_rps_set_timer(struct intel_rps *rps)
+{
+	set_bit(INTEL_RPS_TIMER, &rps->flags);
+}
+
+static inline void intel_rps_clear_timer(struct intel_rps *rps)
+{
+	clear_bit(INTEL_RPS_TIMER, &rps->flags);
+}
+
+#endif /* INTEL_RPS_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_rps_types.h b/drivers/gpu/drm/i915/gt/intel_rps_types.h
new file mode 100644
index 000000000..9173ec75f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_rps_types.h
@@ -0,0 +1,117 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_RPS_TYPES_H
+#define INTEL_RPS_TYPES_H
+
+#include <linux/atomic.h>
+#include <linux/ktime.h>
+#include <linux/mutex.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+
+struct intel_ips {
+	u64 last_count1;
+	unsigned long last_time1;
+	unsigned long chipset_power;
+	u64 last_count2;
+	u64 last_time2;
+	unsigned long gfx_power;
+	u8 corr;
+
+	int c, m;
+};
+
+struct intel_rps_ei {
+	ktime_t ktime;
+	u32 render_c0;
+	u32 media_c0;
+};
+
+enum {
+	INTEL_RPS_ENABLED = 0,
+	INTEL_RPS_ACTIVE,
+	INTEL_RPS_INTERRUPTS,
+	INTEL_RPS_TIMER,
+};
+
+/**
+ * struct intel_rps_freq_caps - rps freq capabilities
+ * @rp0_freq: non-overclocked max frequency
+ * @rp1_freq: "less than" RP0 power/freqency
+ * @min_freq: aka RPn, minimum frequency
+ *
+ * Freq caps exposed by HW, values are in "hw units" and intel_gpu_freq()
+ * should be used to convert to MHz
+ */
+struct intel_rps_freq_caps {
+	u8 rp0_freq;
+	u8 rp1_freq;
+	u8 min_freq;
+};
+
+struct intel_rps {
+	struct mutex lock; /* protects enabling and the worker */
+
+	/*
+	 * work, interrupts_enabled and pm_iir are protected by
+	 * dev_priv->irq_lock
+	 */
+	struct timer_list timer;
+	struct work_struct work;
+	unsigned long flags;
+
+	ktime_t pm_timestamp;
+	u32 pm_interval;
+	u32 pm_iir;
+
+	/* PM interrupt bits that should never be masked */
+	u32 pm_intrmsk_mbz;
+	u32 pm_events;
+
+	/* Frequencies are stored in potentially platform dependent multiples.
+	 * In other words, *_freq needs to be multiplied by X to be interesting.
+	 * Soft limits are those which are used for the dynamic reclocking done
+	 * by the driver (raise frequencies under heavy loads, and lower for
+	 * lighter loads). Hard limits are those imposed by the hardware.
+	 *
+	 * A distinction is made for overclocking, which is never enabled by
+	 * default, and is considered to be above the hard limit if it's
+	 * possible at all.
+	 */
+	u8 cur_freq;		/* Current frequency (cached, may not == HW) */
+	u8 last_freq;		/* Last SWREQ frequency */
+	u8 min_freq_softlimit;	/* Minimum frequency permitted by the driver */
+	u8 max_freq_softlimit;	/* Max frequency permitted by the driver */
+	u8 max_freq;		/* Maximum frequency, RP0 if not overclocking */
+	u8 min_freq;		/* AKA RPn. Minimum frequency */
+	u8 boost_freq;		/* Frequency to request when wait boosting */
+	u8 idle_freq;		/* Frequency to request when we are idle */
+	u8 efficient_freq;	/* AKA RPe. Pre-determined balanced frequency */
+	u8 rp1_freq;		/* "less than" RP0 power/freqency */
+	u8 rp0_freq;		/* Non-overclocked max frequency. */
+	u16 gpll_ref_freq;	/* vlv/chv GPLL reference frequency */
+
+	int last_adj;
+
+	struct {
+		struct mutex mutex;
+
+		enum { LOW_POWER, BETWEEN, HIGH_POWER } mode;
+		unsigned int interactive;
+
+		u8 up_threshold; /* Current %busy required to uplock */
+		u8 down_threshold; /* Current %busy required to downclock */
+	} power;
+
+	atomic_t num_waiters;
+	unsigned int boosts;
+
+	/* manual wa residency calculations */
+	struct intel_rps_ei ei;
+	struct intel_ips ips;
+};
+
+#endif /* INTEL_RPS_TYPES_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_sa_media.c b/drivers/gpu/drm/i915/gt/intel_sa_media.c
new file mode 100644
index 000000000..e8f3d18c1
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_sa_media.c
@@ -0,0 +1,47 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <drm/drm_managed.h>
+
+#include "i915_drv.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_sa_media.h"
+
+int intel_sa_mediagt_setup(struct intel_gt *gt, phys_addr_t phys_addr,
+			   u32 gsi_offset)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore;
+
+	uncore = drmm_kzalloc(&i915->drm, sizeof(*uncore), GFP_KERNEL);
+	if (!uncore)
+		return -ENOMEM;
+
+	uncore->gsi_offset = gsi_offset;
+
+	gt->irq_lock = to_gt(i915)->irq_lock;
+	intel_gt_common_init_early(gt);
+	intel_uncore_init_early(uncore, gt);
+
+	/*
+	 * Standalone media shares the general MMIO space with the primary
+	 * GT.  We'll re-use the primary GT's mapping.
+	 */
+	uncore->regs = i915->uncore.regs;
+	if (drm_WARN_ON(&i915->drm, uncore->regs == NULL))
+		return -EIO;
+
+	gt->uncore = uncore;
+	gt->phys_addr = phys_addr;
+
+	/*
+	 * For current platforms we can assume there's only a single
+	 * media GT and cache it for quick lookup.
+	 */
+	drm_WARN_ON(&i915->drm, i915->media_gt);
+	i915->media_gt = gt;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_sa_media.h b/drivers/gpu/drm/i915/gt/intel_sa_media.h
new file mode 100644
index 000000000..3afb310de
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_sa_media.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+#ifndef __INTEL_SA_MEDIA__
+#define __INTEL_SA_MEDIA__
+
+#include <linux/types.h>
+
+struct intel_gt;
+
+int intel_sa_mediagt_setup(struct intel_gt *gt, phys_addr_t phys_addr,
+			   u32 gsi_offset);
+
+#endif /* __INTEL_SA_MEDIA_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_sseu.c b/drivers/gpu/drm/i915/gt/intel_sseu.c
new file mode 100644
index 000000000..66f21c735
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_sseu.c
@@ -0,0 +1,900 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+
+#include "i915_drv.h"
+#include "intel_engine_regs.h"
+#include "intel_gt_regs.h"
+#include "intel_sseu.h"
+
+void intel_sseu_set_info(struct sseu_dev_info *sseu, u8 max_slices,
+			 u8 max_subslices, u8 max_eus_per_subslice)
+{
+	sseu->max_slices = max_slices;
+	sseu->max_subslices = max_subslices;
+	sseu->max_eus_per_subslice = max_eus_per_subslice;
+}
+
+unsigned int
+intel_sseu_subslice_total(const struct sseu_dev_info *sseu)
+{
+	unsigned int i, total = 0;
+
+	if (sseu->has_xehp_dss)
+		return bitmap_weight(sseu->subslice_mask.xehp,
+				     XEHP_BITMAP_BITS(sseu->subslice_mask));
+
+	for (i = 0; i < ARRAY_SIZE(sseu->subslice_mask.hsw); i++)
+		total += hweight8(sseu->subslice_mask.hsw[i]);
+
+	return total;
+}
+
+unsigned int
+intel_sseu_get_hsw_subslices(const struct sseu_dev_info *sseu, u8 slice)
+{
+	WARN_ON(sseu->has_xehp_dss);
+	if (WARN_ON(slice >= sseu->max_slices))
+		return 0;
+
+	return sseu->subslice_mask.hsw[slice];
+}
+
+static u16 sseu_get_eus(const struct sseu_dev_info *sseu, int slice,
+			int subslice)
+{
+	if (sseu->has_xehp_dss) {
+		WARN_ON(slice > 0);
+		return sseu->eu_mask.xehp[subslice];
+	} else {
+		return sseu->eu_mask.hsw[slice][subslice];
+	}
+}
+
+static void sseu_set_eus(struct sseu_dev_info *sseu, int slice, int subslice,
+			 u16 eu_mask)
+{
+	GEM_WARN_ON(eu_mask && __fls(eu_mask) >= sseu->max_eus_per_subslice);
+	if (sseu->has_xehp_dss) {
+		GEM_WARN_ON(slice > 0);
+		sseu->eu_mask.xehp[subslice] = eu_mask;
+	} else {
+		sseu->eu_mask.hsw[slice][subslice] = eu_mask;
+	}
+}
+
+static u16 compute_eu_total(const struct sseu_dev_info *sseu)
+{
+	int s, ss, total = 0;
+
+	for (s = 0; s < sseu->max_slices; s++)
+		for (ss = 0; ss < sseu->max_subslices; ss++)
+			if (sseu->has_xehp_dss)
+				total += hweight16(sseu->eu_mask.xehp[ss]);
+			else
+				total += hweight16(sseu->eu_mask.hsw[s][ss]);
+
+	return total;
+}
+
+/**
+ * intel_sseu_copy_eumask_to_user - Copy EU mask into a userspace buffer
+ * @to: Pointer to userspace buffer to copy to
+ * @sseu: SSEU structure containing EU mask to copy
+ *
+ * Copies the EU mask to a userspace buffer in the format expected by
+ * the query ioctl's topology queries.
+ *
+ * Returns the result of the copy_to_user() operation.
+ */
+int intel_sseu_copy_eumask_to_user(void __user *to,
+				   const struct sseu_dev_info *sseu)
+{
+	u8 eu_mask[GEN_SS_MASK_SIZE * GEN_MAX_EU_STRIDE] = {};
+	int eu_stride = GEN_SSEU_STRIDE(sseu->max_eus_per_subslice);
+	int len = sseu->max_slices * sseu->max_subslices * eu_stride;
+	int s, ss, i;
+
+	for (s = 0; s < sseu->max_slices; s++) {
+		for (ss = 0; ss < sseu->max_subslices; ss++) {
+			int uapi_offset =
+				s * sseu->max_subslices * eu_stride +
+				ss * eu_stride;
+			u16 mask = sseu_get_eus(sseu, s, ss);
+
+			for (i = 0; i < eu_stride; i++)
+				eu_mask[uapi_offset + i] =
+					(mask >> (BITS_PER_BYTE * i)) & 0xff;
+		}
+	}
+
+	return copy_to_user(to, eu_mask, len);
+}
+
+/**
+ * intel_sseu_copy_ssmask_to_user - Copy subslice mask into a userspace buffer
+ * @to: Pointer to userspace buffer to copy to
+ * @sseu: SSEU structure containing subslice mask to copy
+ *
+ * Copies the subslice mask to a userspace buffer in the format expected by
+ * the query ioctl's topology queries.
+ *
+ * Returns the result of the copy_to_user() operation.
+ */
+int intel_sseu_copy_ssmask_to_user(void __user *to,
+				   const struct sseu_dev_info *sseu)
+{
+	u8 ss_mask[GEN_SS_MASK_SIZE] = {};
+	int ss_stride = GEN_SSEU_STRIDE(sseu->max_subslices);
+	int len = sseu->max_slices * ss_stride;
+	int s, ss, i;
+
+	for (s = 0; s < sseu->max_slices; s++) {
+		for (ss = 0; ss < sseu->max_subslices; ss++) {
+			i = s * ss_stride * BITS_PER_BYTE + ss;
+
+			if (!intel_sseu_has_subslice(sseu, s, ss))
+				continue;
+
+			ss_mask[i / BITS_PER_BYTE] |= BIT(i % BITS_PER_BYTE);
+		}
+	}
+
+	return copy_to_user(to, ss_mask, len);
+}
+
+static void gen11_compute_sseu_info(struct sseu_dev_info *sseu,
+				    u32 ss_en, u16 eu_en)
+{
+	u32 valid_ss_mask = GENMASK(sseu->max_subslices - 1, 0);
+	int ss;
+
+	sseu->slice_mask |= BIT(0);
+	sseu->subslice_mask.hsw[0] = ss_en & valid_ss_mask;
+
+	for (ss = 0; ss < sseu->max_subslices; ss++)
+		if (intel_sseu_has_subslice(sseu, 0, ss))
+			sseu_set_eus(sseu, 0, ss, eu_en);
+
+	sseu->eu_per_subslice = hweight16(eu_en);
+	sseu->eu_total = compute_eu_total(sseu);
+}
+
+static void xehp_compute_sseu_info(struct sseu_dev_info *sseu,
+				   u16 eu_en)
+{
+	int ss;
+
+	sseu->slice_mask |= BIT(0);
+
+	bitmap_or(sseu->subslice_mask.xehp,
+		  sseu->compute_subslice_mask.xehp,
+		  sseu->geometry_subslice_mask.xehp,
+		  XEHP_BITMAP_BITS(sseu->subslice_mask));
+
+	for (ss = 0; ss < sseu->max_subslices; ss++)
+		if (intel_sseu_has_subslice(sseu, 0, ss))
+			sseu_set_eus(sseu, 0, ss, eu_en);
+
+	sseu->eu_per_subslice = hweight16(eu_en);
+	sseu->eu_total = compute_eu_total(sseu);
+}
+
+static void
+xehp_load_dss_mask(struct intel_uncore *uncore,
+		   intel_sseu_ss_mask_t *ssmask,
+		   int numregs,
+		   ...)
+{
+	va_list argp;
+	u32 fuse_val[I915_MAX_SS_FUSE_REGS] = {};
+	int i;
+
+	if (WARN_ON(numregs > I915_MAX_SS_FUSE_REGS))
+		numregs = I915_MAX_SS_FUSE_REGS;
+
+	va_start(argp, numregs);
+	for (i = 0; i < numregs; i++)
+		fuse_val[i] = intel_uncore_read(uncore, va_arg(argp, i915_reg_t));
+	va_end(argp);
+
+	bitmap_from_arr32(ssmask->xehp, fuse_val, numregs * 32);
+}
+
+static void xehp_sseu_info_init(struct intel_gt *gt)
+{
+	struct sseu_dev_info *sseu = &gt->info.sseu;
+	struct intel_uncore *uncore = gt->uncore;
+	u16 eu_en = 0;
+	u8 eu_en_fuse;
+	int num_compute_regs, num_geometry_regs;
+	int eu;
+
+	if (IS_PONTEVECCHIO(gt->i915)) {
+		num_geometry_regs = 0;
+		num_compute_regs = 2;
+	} else {
+		num_geometry_regs = 1;
+		num_compute_regs = 1;
+	}
+
+	/*
+	 * The concept of slice has been removed in Xe_HP.  To be compatible
+	 * with prior generations, assume a single slice across the entire
+	 * device. Then calculate out the DSS for each workload type within
+	 * that software slice.
+	 */
+	intel_sseu_set_info(sseu, 1,
+			    32 * max(num_geometry_regs, num_compute_regs),
+			    HAS_ONE_EU_PER_FUSE_BIT(gt->i915) ? 8 : 16);
+	sseu->has_xehp_dss = 1;
+
+	xehp_load_dss_mask(uncore, &sseu->geometry_subslice_mask,
+			   num_geometry_regs,
+			   GEN12_GT_GEOMETRY_DSS_ENABLE);
+	xehp_load_dss_mask(uncore, &sseu->compute_subslice_mask,
+			   num_compute_regs,
+			   GEN12_GT_COMPUTE_DSS_ENABLE,
+			   XEHPC_GT_COMPUTE_DSS_ENABLE_EXT);
+
+	eu_en_fuse = intel_uncore_read(uncore, XEHP_EU_ENABLE) & XEHP_EU_ENA_MASK;
+
+	if (HAS_ONE_EU_PER_FUSE_BIT(gt->i915))
+		eu_en = eu_en_fuse;
+	else
+		for (eu = 0; eu < sseu->max_eus_per_subslice / 2; eu++)
+			if (eu_en_fuse & BIT(eu))
+				eu_en |= BIT(eu * 2) | BIT(eu * 2 + 1);
+
+	xehp_compute_sseu_info(sseu, eu_en);
+}
+
+static void gen12_sseu_info_init(struct intel_gt *gt)
+{
+	struct sseu_dev_info *sseu = &gt->info.sseu;
+	struct intel_uncore *uncore = gt->uncore;
+	u32 g_dss_en;
+	u16 eu_en = 0;
+	u8 eu_en_fuse;
+	u8 s_en;
+	int eu;
+
+	/*
+	 * Gen12 has Dual-Subslices, which behave similarly to 2 gen11 SS.
+	 * Instead of splitting these, provide userspace with an array
+	 * of DSS to more closely represent the hardware resource.
+	 */
+	intel_sseu_set_info(sseu, 1, 6, 16);
+
+	/*
+	 * Although gen12 architecture supported multiple slices, TGL, RKL,
+	 * DG1, and ADL only had a single slice.
+	 */
+	s_en = intel_uncore_read(uncore, GEN11_GT_SLICE_ENABLE) &
+		GEN11_GT_S_ENA_MASK;
+	drm_WARN_ON(&gt->i915->drm, s_en != 0x1);
+
+	g_dss_en = intel_uncore_read(uncore, GEN12_GT_GEOMETRY_DSS_ENABLE);
+
+	/* one bit per pair of EUs */
+	eu_en_fuse = ~(intel_uncore_read(uncore, GEN11_EU_DISABLE) &
+		       GEN11_EU_DIS_MASK);
+
+	for (eu = 0; eu < sseu->max_eus_per_subslice / 2; eu++)
+		if (eu_en_fuse & BIT(eu))
+			eu_en |= BIT(eu * 2) | BIT(eu * 2 + 1);
+
+	gen11_compute_sseu_info(sseu, g_dss_en, eu_en);
+
+	/* TGL only supports slice-level power gating */
+	sseu->has_slice_pg = 1;
+}
+
+static void gen11_sseu_info_init(struct intel_gt *gt)
+{
+	struct sseu_dev_info *sseu = &gt->info.sseu;
+	struct intel_uncore *uncore = gt->uncore;
+	u32 ss_en;
+	u8 eu_en;
+	u8 s_en;
+
+	if (IS_JSL_EHL(gt->i915))
+		intel_sseu_set_info(sseu, 1, 4, 8);
+	else
+		intel_sseu_set_info(sseu, 1, 8, 8);
+
+	/*
+	 * Although gen11 architecture supported multiple slices, ICL and
+	 * EHL/JSL only had a single slice in practice.
+	 */
+	s_en = intel_uncore_read(uncore, GEN11_GT_SLICE_ENABLE) &
+		GEN11_GT_S_ENA_MASK;
+	drm_WARN_ON(&gt->i915->drm, s_en != 0x1);
+
+	ss_en = ~intel_uncore_read(uncore, GEN11_GT_SUBSLICE_DISABLE);
+
+	eu_en = ~(intel_uncore_read(uncore, GEN11_EU_DISABLE) &
+		  GEN11_EU_DIS_MASK);
+
+	gen11_compute_sseu_info(sseu, ss_en, eu_en);
+
+	/* ICL has no power gating restrictions. */
+	sseu->has_slice_pg = 1;
+	sseu->has_subslice_pg = 1;
+	sseu->has_eu_pg = 1;
+}
+
+static void cherryview_sseu_info_init(struct intel_gt *gt)
+{
+	struct sseu_dev_info *sseu = &gt->info.sseu;
+	u32 fuse;
+
+	fuse = intel_uncore_read(gt->uncore, CHV_FUSE_GT);
+
+	sseu->slice_mask = BIT(0);
+	intel_sseu_set_info(sseu, 1, 2, 8);
+
+	if (!(fuse & CHV_FGT_DISABLE_SS0)) {
+		u8 disabled_mask =
+			((fuse & CHV_FGT_EU_DIS_SS0_R0_MASK) >>
+			 CHV_FGT_EU_DIS_SS0_R0_SHIFT) |
+			(((fuse & CHV_FGT_EU_DIS_SS0_R1_MASK) >>
+			  CHV_FGT_EU_DIS_SS0_R1_SHIFT) << 4);
+
+		sseu->subslice_mask.hsw[0] |= BIT(0);
+		sseu_set_eus(sseu, 0, 0, ~disabled_mask & 0xFF);
+	}
+
+	if (!(fuse & CHV_FGT_DISABLE_SS1)) {
+		u8 disabled_mask =
+			((fuse & CHV_FGT_EU_DIS_SS1_R0_MASK) >>
+			 CHV_FGT_EU_DIS_SS1_R0_SHIFT) |
+			(((fuse & CHV_FGT_EU_DIS_SS1_R1_MASK) >>
+			  CHV_FGT_EU_DIS_SS1_R1_SHIFT) << 4);
+
+		sseu->subslice_mask.hsw[0] |= BIT(1);
+		sseu_set_eus(sseu, 0, 1, ~disabled_mask & 0xFF);
+	}
+
+	sseu->eu_total = compute_eu_total(sseu);
+
+	/*
+	 * CHV expected to always have a uniform distribution of EU
+	 * across subslices.
+	 */
+	sseu->eu_per_subslice = intel_sseu_subslice_total(sseu) ?
+		sseu->eu_total /
+		intel_sseu_subslice_total(sseu) :
+		0;
+	/*
+	 * CHV supports subslice power gating on devices with more than
+	 * one subslice, and supports EU power gating on devices with
+	 * more than one EU pair per subslice.
+	 */
+	sseu->has_slice_pg = 0;
+	sseu->has_subslice_pg = intel_sseu_subslice_total(sseu) > 1;
+	sseu->has_eu_pg = (sseu->eu_per_subslice > 2);
+}
+
+static void gen9_sseu_info_init(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct sseu_dev_info *sseu = &gt->info.sseu;
+	struct intel_uncore *uncore = gt->uncore;
+	u32 fuse2, eu_disable, subslice_mask;
+	const u8 eu_mask = 0xff;
+	int s, ss;
+
+	fuse2 = intel_uncore_read(uncore, GEN8_FUSE2);
+	sseu->slice_mask = (fuse2 & GEN8_F2_S_ENA_MASK) >> GEN8_F2_S_ENA_SHIFT;
+
+	/* BXT has a single slice and at most 3 subslices. */
+	intel_sseu_set_info(sseu, IS_GEN9_LP(i915) ? 1 : 3,
+			    IS_GEN9_LP(i915) ? 3 : 4, 8);
+
+	/*
+	 * The subslice disable field is global, i.e. it applies
+	 * to each of the enabled slices.
+	 */
+	subslice_mask = (1 << sseu->max_subslices) - 1;
+	subslice_mask &= ~((fuse2 & GEN9_F2_SS_DIS_MASK) >>
+			   GEN9_F2_SS_DIS_SHIFT);
+
+	/*
+	 * Iterate through enabled slices and subslices to
+	 * count the total enabled EU.
+	 */
+	for (s = 0; s < sseu->max_slices; s++) {
+		if (!(sseu->slice_mask & BIT(s)))
+			/* skip disabled slice */
+			continue;
+
+		sseu->subslice_mask.hsw[s] = subslice_mask;
+
+		eu_disable = intel_uncore_read(uncore, GEN9_EU_DISABLE(s));
+		for (ss = 0; ss < sseu->max_subslices; ss++) {
+			int eu_per_ss;
+			u8 eu_disabled_mask;
+
+			if (!intel_sseu_has_subslice(sseu, s, ss))
+				/* skip disabled subslice */
+				continue;
+
+			eu_disabled_mask = (eu_disable >> (ss * 8)) & eu_mask;
+
+			sseu_set_eus(sseu, s, ss, ~eu_disabled_mask & eu_mask);
+
+			eu_per_ss = sseu->max_eus_per_subslice -
+				hweight8(eu_disabled_mask);
+
+			/*
+			 * Record which subslice(s) has(have) 7 EUs. we
+			 * can tune the hash used to spread work among
+			 * subslices if they are unbalanced.
+			 */
+			if (eu_per_ss == 7)
+				sseu->subslice_7eu[s] |= BIT(ss);
+		}
+	}
+
+	sseu->eu_total = compute_eu_total(sseu);
+
+	/*
+	 * SKL is expected to always have a uniform distribution
+	 * of EU across subslices with the exception that any one
+	 * EU in any one subslice may be fused off for die
+	 * recovery. BXT is expected to be perfectly uniform in EU
+	 * distribution.
+	 */
+	sseu->eu_per_subslice =
+		intel_sseu_subslice_total(sseu) ?
+		DIV_ROUND_UP(sseu->eu_total, intel_sseu_subslice_total(sseu)) :
+		0;
+
+	/*
+	 * SKL+ supports slice power gating on devices with more than
+	 * one slice, and supports EU power gating on devices with
+	 * more than one EU pair per subslice. BXT+ supports subslice
+	 * power gating on devices with more than one subslice, and
+	 * supports EU power gating on devices with more than one EU
+	 * pair per subslice.
+	 */
+	sseu->has_slice_pg =
+		!IS_GEN9_LP(i915) && hweight8(sseu->slice_mask) > 1;
+	sseu->has_subslice_pg =
+		IS_GEN9_LP(i915) && intel_sseu_subslice_total(sseu) > 1;
+	sseu->has_eu_pg = sseu->eu_per_subslice > 2;
+
+	if (IS_GEN9_LP(i915)) {
+#define IS_SS_DISABLED(ss)	(!(sseu->subslice_mask.hsw[0] & BIT(ss)))
+		RUNTIME_INFO(i915)->has_pooled_eu = hweight8(sseu->subslice_mask.hsw[0]) == 3;
+
+		sseu->min_eu_in_pool = 0;
+		if (HAS_POOLED_EU(i915)) {
+			if (IS_SS_DISABLED(2) || IS_SS_DISABLED(0))
+				sseu->min_eu_in_pool = 3;
+			else if (IS_SS_DISABLED(1))
+				sseu->min_eu_in_pool = 6;
+			else
+				sseu->min_eu_in_pool = 9;
+		}
+#undef IS_SS_DISABLED
+	}
+}
+
+static void bdw_sseu_info_init(struct intel_gt *gt)
+{
+	struct sseu_dev_info *sseu = &gt->info.sseu;
+	struct intel_uncore *uncore = gt->uncore;
+	int s, ss;
+	u32 fuse2, subslice_mask, eu_disable[3]; /* s_max */
+	u32 eu_disable0, eu_disable1, eu_disable2;
+
+	fuse2 = intel_uncore_read(uncore, GEN8_FUSE2);
+	sseu->slice_mask = (fuse2 & GEN8_F2_S_ENA_MASK) >> GEN8_F2_S_ENA_SHIFT;
+	intel_sseu_set_info(sseu, 3, 3, 8);
+
+	/*
+	 * The subslice disable field is global, i.e. it applies
+	 * to each of the enabled slices.
+	 */
+	subslice_mask = GENMASK(sseu->max_subslices - 1, 0);
+	subslice_mask &= ~((fuse2 & GEN8_F2_SS_DIS_MASK) >>
+			   GEN8_F2_SS_DIS_SHIFT);
+	eu_disable0 = intel_uncore_read(uncore, GEN8_EU_DISABLE0);
+	eu_disable1 = intel_uncore_read(uncore, GEN8_EU_DISABLE1);
+	eu_disable2 = intel_uncore_read(uncore, GEN8_EU_DISABLE2);
+	eu_disable[0] = eu_disable0 & GEN8_EU_DIS0_S0_MASK;
+	eu_disable[1] = (eu_disable0 >> GEN8_EU_DIS0_S1_SHIFT) |
+		((eu_disable1 & GEN8_EU_DIS1_S1_MASK) <<
+		 (32 - GEN8_EU_DIS0_S1_SHIFT));
+	eu_disable[2] = (eu_disable1 >> GEN8_EU_DIS1_S2_SHIFT) |
+		((eu_disable2 & GEN8_EU_DIS2_S2_MASK) <<
+		 (32 - GEN8_EU_DIS1_S2_SHIFT));
+
+	/*
+	 * Iterate through enabled slices and subslices to
+	 * count the total enabled EU.
+	 */
+	for (s = 0; s < sseu->max_slices; s++) {
+		if (!(sseu->slice_mask & BIT(s)))
+			/* skip disabled slice */
+			continue;
+
+		sseu->subslice_mask.hsw[s] = subslice_mask;
+
+		for (ss = 0; ss < sseu->max_subslices; ss++) {
+			u8 eu_disabled_mask;
+			u32 n_disabled;
+
+			if (!intel_sseu_has_subslice(sseu, s, ss))
+				/* skip disabled subslice */
+				continue;
+
+			eu_disabled_mask =
+				eu_disable[s] >> (ss * sseu->max_eus_per_subslice);
+
+			sseu_set_eus(sseu, s, ss, ~eu_disabled_mask & 0xFF);
+
+			n_disabled = hweight8(eu_disabled_mask);
+
+			/*
+			 * Record which subslices have 7 EUs.
+			 */
+			if (sseu->max_eus_per_subslice - n_disabled == 7)
+				sseu->subslice_7eu[s] |= 1 << ss;
+		}
+	}
+
+	sseu->eu_total = compute_eu_total(sseu);
+
+	/*
+	 * BDW is expected to always have a uniform distribution of EU across
+	 * subslices with the exception that any one EU in any one subslice may
+	 * be fused off for die recovery.
+	 */
+	sseu->eu_per_subslice =
+		intel_sseu_subslice_total(sseu) ?
+		DIV_ROUND_UP(sseu->eu_total, intel_sseu_subslice_total(sseu)) :
+		0;
+
+	/*
+	 * BDW supports slice power gating on devices with more than
+	 * one slice.
+	 */
+	sseu->has_slice_pg = hweight8(sseu->slice_mask) > 1;
+	sseu->has_subslice_pg = 0;
+	sseu->has_eu_pg = 0;
+}
+
+static void hsw_sseu_info_init(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct sseu_dev_info *sseu = &gt->info.sseu;
+	u32 fuse1;
+	u8 subslice_mask = 0;
+	int s, ss;
+
+	/*
+	 * There isn't a register to tell us how many slices/subslices. We
+	 * work off the PCI-ids here.
+	 */
+	switch (INTEL_INFO(i915)->gt) {
+	default:
+		MISSING_CASE(INTEL_INFO(i915)->gt);
+		fallthrough;
+	case 1:
+		sseu->slice_mask = BIT(0);
+		subslice_mask = BIT(0);
+		break;
+	case 2:
+		sseu->slice_mask = BIT(0);
+		subslice_mask = BIT(0) | BIT(1);
+		break;
+	case 3:
+		sseu->slice_mask = BIT(0) | BIT(1);
+		subslice_mask = BIT(0) | BIT(1);
+		break;
+	}
+
+	fuse1 = intel_uncore_read(gt->uncore, HSW_PAVP_FUSE1);
+	switch (REG_FIELD_GET(HSW_F1_EU_DIS_MASK, fuse1)) {
+	default:
+		MISSING_CASE(REG_FIELD_GET(HSW_F1_EU_DIS_MASK, fuse1));
+		fallthrough;
+	case HSW_F1_EU_DIS_10EUS:
+		sseu->eu_per_subslice = 10;
+		break;
+	case HSW_F1_EU_DIS_8EUS:
+		sseu->eu_per_subslice = 8;
+		break;
+	case HSW_F1_EU_DIS_6EUS:
+		sseu->eu_per_subslice = 6;
+		break;
+	}
+
+	intel_sseu_set_info(sseu, hweight8(sseu->slice_mask),
+			    hweight8(subslice_mask),
+			    sseu->eu_per_subslice);
+
+	for (s = 0; s < sseu->max_slices; s++) {
+		sseu->subslice_mask.hsw[s] = subslice_mask;
+
+		for (ss = 0; ss < sseu->max_subslices; ss++) {
+			sseu_set_eus(sseu, s, ss,
+				     (1UL << sseu->eu_per_subslice) - 1);
+		}
+	}
+
+	sseu->eu_total = compute_eu_total(sseu);
+
+	/* No powergating for you. */
+	sseu->has_slice_pg = 0;
+	sseu->has_subslice_pg = 0;
+	sseu->has_eu_pg = 0;
+}
+
+void intel_sseu_info_init(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50))
+		xehp_sseu_info_init(gt);
+	else if (GRAPHICS_VER(i915) >= 12)
+		gen12_sseu_info_init(gt);
+	else if (GRAPHICS_VER(i915) >= 11)
+		gen11_sseu_info_init(gt);
+	else if (GRAPHICS_VER(i915) >= 9)
+		gen9_sseu_info_init(gt);
+	else if (IS_BROADWELL(i915))
+		bdw_sseu_info_init(gt);
+	else if (IS_CHERRYVIEW(i915))
+		cherryview_sseu_info_init(gt);
+	else if (IS_HASWELL(i915))
+		hsw_sseu_info_init(gt);
+}
+
+u32 intel_sseu_make_rpcs(struct intel_gt *gt,
+			 const struct intel_sseu *req_sseu)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	const struct sseu_dev_info *sseu = &gt->info.sseu;
+	bool subslice_pg = sseu->has_subslice_pg;
+	u8 slices, subslices;
+	u32 rpcs = 0;
+
+	/*
+	 * No explicit RPCS request is needed to ensure full
+	 * slice/subslice/EU enablement prior to Gen9.
+	 */
+	if (GRAPHICS_VER(i915) < 9)
+		return 0;
+
+	/*
+	 * If i915/perf is active, we want a stable powergating configuration
+	 * on the system. Use the configuration pinned by i915/perf.
+	 */
+	if (i915->perf.exclusive_stream)
+		req_sseu = &i915->perf.sseu;
+
+	slices = hweight8(req_sseu->slice_mask);
+	subslices = hweight8(req_sseu->subslice_mask);
+
+	/*
+	 * Since the SScount bitfield in GEN8_R_PWR_CLK_STATE is only three bits
+	 * wide and Icelake has up to eight subslices, specfial programming is
+	 * needed in order to correctly enable all subslices.
+	 *
+	 * According to documentation software must consider the configuration
+	 * as 2x4x8 and hardware will translate this to 1x8x8.
+	 *
+	 * Furthemore, even though SScount is three bits, maximum documented
+	 * value for it is four. From this some rules/restrictions follow:
+	 *
+	 * 1.
+	 * If enabled subslice count is greater than four, two whole slices must
+	 * be enabled instead.
+	 *
+	 * 2.
+	 * When more than one slice is enabled, hardware ignores the subslice
+	 * count altogether.
+	 *
+	 * From these restrictions it follows that it is not possible to enable
+	 * a count of subslices between the SScount maximum of four restriction,
+	 * and the maximum available number on a particular SKU. Either all
+	 * subslices are enabled, or a count between one and four on the first
+	 * slice.
+	 */
+	if (GRAPHICS_VER(i915) == 11 &&
+	    slices == 1 &&
+	    subslices > min_t(u8, 4, hweight8(sseu->subslice_mask.hsw[0]) / 2)) {
+		GEM_BUG_ON(subslices & 1);
+
+		subslice_pg = false;
+		slices *= 2;
+	}
+
+	/*
+	 * Starting in Gen9, render power gating can leave
+	 * slice/subslice/EU in a partially enabled state. We
+	 * must make an explicit request through RPCS for full
+	 * enablement.
+	 */
+	if (sseu->has_slice_pg) {
+		u32 mask, val = slices;
+
+		if (GRAPHICS_VER(i915) >= 11) {
+			mask = GEN11_RPCS_S_CNT_MASK;
+			val <<= GEN11_RPCS_S_CNT_SHIFT;
+		} else {
+			mask = GEN8_RPCS_S_CNT_MASK;
+			val <<= GEN8_RPCS_S_CNT_SHIFT;
+		}
+
+		GEM_BUG_ON(val & ~mask);
+		val &= mask;
+
+		rpcs |= GEN8_RPCS_ENABLE | GEN8_RPCS_S_CNT_ENABLE | val;
+	}
+
+	if (subslice_pg) {
+		u32 val = subslices;
+
+		val <<= GEN8_RPCS_SS_CNT_SHIFT;
+
+		GEM_BUG_ON(val & ~GEN8_RPCS_SS_CNT_MASK);
+		val &= GEN8_RPCS_SS_CNT_MASK;
+
+		rpcs |= GEN8_RPCS_ENABLE | GEN8_RPCS_SS_CNT_ENABLE | val;
+	}
+
+	if (sseu->has_eu_pg) {
+		u32 val;
+
+		val = req_sseu->min_eus_per_subslice << GEN8_RPCS_EU_MIN_SHIFT;
+		GEM_BUG_ON(val & ~GEN8_RPCS_EU_MIN_MASK);
+		val &= GEN8_RPCS_EU_MIN_MASK;
+
+		rpcs |= val;
+
+		val = req_sseu->max_eus_per_subslice << GEN8_RPCS_EU_MAX_SHIFT;
+		GEM_BUG_ON(val & ~GEN8_RPCS_EU_MAX_MASK);
+		val &= GEN8_RPCS_EU_MAX_MASK;
+
+		rpcs |= val;
+
+		rpcs |= GEN8_RPCS_ENABLE;
+	}
+
+	return rpcs;
+}
+
+void intel_sseu_dump(const struct sseu_dev_info *sseu, struct drm_printer *p)
+{
+	int s;
+
+	if (sseu->has_xehp_dss) {
+		drm_printf(p, "subslice total: %u\n",
+			   intel_sseu_subslice_total(sseu));
+		drm_printf(p, "geometry dss mask=%*pb\n",
+			   XEHP_BITMAP_BITS(sseu->geometry_subslice_mask),
+			   sseu->geometry_subslice_mask.xehp);
+		drm_printf(p, "compute dss mask=%*pb\n",
+			   XEHP_BITMAP_BITS(sseu->compute_subslice_mask),
+			   sseu->compute_subslice_mask.xehp);
+	} else {
+		drm_printf(p, "slice total: %u, mask=%04x\n",
+			   hweight8(sseu->slice_mask), sseu->slice_mask);
+		drm_printf(p, "subslice total: %u\n",
+			   intel_sseu_subslice_total(sseu));
+
+		for (s = 0; s < sseu->max_slices; s++) {
+			u8 ss_mask = sseu->subslice_mask.hsw[s];
+
+			drm_printf(p, "slice%d: %u subslices, mask=%08x\n",
+				   s, hweight8(ss_mask), ss_mask);
+		}
+	}
+
+	drm_printf(p, "EU total: %u\n", sseu->eu_total);
+	drm_printf(p, "EU per subslice: %u\n", sseu->eu_per_subslice);
+	drm_printf(p, "has slice power gating: %s\n",
+		   str_yes_no(sseu->has_slice_pg));
+	drm_printf(p, "has subslice power gating: %s\n",
+		   str_yes_no(sseu->has_subslice_pg));
+	drm_printf(p, "has EU power gating: %s\n",
+		   str_yes_no(sseu->has_eu_pg));
+}
+
+static void sseu_print_hsw_topology(const struct sseu_dev_info *sseu,
+				    struct drm_printer *p)
+{
+	int s, ss;
+
+	for (s = 0; s < sseu->max_slices; s++) {
+		u8 ss_mask = sseu->subslice_mask.hsw[s];
+
+		drm_printf(p, "slice%d: %u subslice(s) (0x%08x):\n",
+			   s, hweight8(ss_mask), ss_mask);
+
+		for (ss = 0; ss < sseu->max_subslices; ss++) {
+			u16 enabled_eus = sseu_get_eus(sseu, s, ss);
+
+			drm_printf(p, "\tsubslice%d: %u EUs (0x%hx)\n",
+				   ss, hweight16(enabled_eus), enabled_eus);
+		}
+	}
+}
+
+static void sseu_print_xehp_topology(const struct sseu_dev_info *sseu,
+				     struct drm_printer *p)
+{
+	int dss;
+
+	for (dss = 0; dss < sseu->max_subslices; dss++) {
+		u16 enabled_eus = sseu_get_eus(sseu, 0, dss);
+
+		drm_printf(p, "DSS_%02d: G:%3s C:%3s, %2u EUs (0x%04hx)\n", dss,
+			   str_yes_no(test_bit(dss, sseu->geometry_subslice_mask.xehp)),
+			   str_yes_no(test_bit(dss, sseu->compute_subslice_mask.xehp)),
+			   hweight16(enabled_eus), enabled_eus);
+	}
+}
+
+void intel_sseu_print_topology(struct drm_i915_private *i915,
+			       const struct sseu_dev_info *sseu,
+			       struct drm_printer *p)
+{
+	if (sseu->max_slices == 0) {
+		drm_printf(p, "Unavailable\n");
+	} else if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50)) {
+		sseu_print_xehp_topology(sseu, p);
+	} else {
+		sseu_print_hsw_topology(sseu, p);
+	}
+}
+
+void intel_sseu_print_ss_info(const char *type,
+			      const struct sseu_dev_info *sseu,
+			      struct seq_file *m)
+{
+	int s;
+
+	if (sseu->has_xehp_dss) {
+		seq_printf(m, "  %s Geometry DSS: %u\n", type,
+			   bitmap_weight(sseu->geometry_subslice_mask.xehp,
+					 XEHP_BITMAP_BITS(sseu->geometry_subslice_mask)));
+		seq_printf(m, "  %s Compute DSS: %u\n", type,
+			   bitmap_weight(sseu->compute_subslice_mask.xehp,
+					 XEHP_BITMAP_BITS(sseu->compute_subslice_mask)));
+	} else {
+		for (s = 0; s < fls(sseu->slice_mask); s++)
+			seq_printf(m, "  %s Slice%i subslices: %u\n", type,
+				   s, hweight8(sseu->subslice_mask.hsw[s]));
+	}
+}
+
+u16 intel_slicemask_from_xehp_dssmask(intel_sseu_ss_mask_t dss_mask,
+				      int dss_per_slice)
+{
+	intel_sseu_ss_mask_t per_slice_mask = {};
+	unsigned long slice_mask = 0;
+	int i;
+
+	WARN_ON(DIV_ROUND_UP(XEHP_BITMAP_BITS(dss_mask), dss_per_slice) >
+		8 * sizeof(slice_mask));
+
+	bitmap_fill(per_slice_mask.xehp, dss_per_slice);
+	for (i = 0; !bitmap_empty(dss_mask.xehp, XEHP_BITMAP_BITS(dss_mask)); i++) {
+		if (bitmap_intersects(dss_mask.xehp, per_slice_mask.xehp, dss_per_slice))
+			slice_mask |= BIT(i);
+
+		bitmap_shift_right(dss_mask.xehp, dss_mask.xehp, dss_per_slice,
+				   XEHP_BITMAP_BITS(dss_mask));
+	}
+
+	return slice_mask;
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_sseu.h b/drivers/gpu/drm/i915/gt/intel_sseu.h
new file mode 100644
index 000000000..d7e8c374f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_sseu.h
@@ -0,0 +1,182 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_SSEU_H__
+#define __INTEL_SSEU_H__
+
+#include <linux/types.h>
+#include <linux/kernel.h>
+
+#include "i915_gem.h"
+
+struct drm_i915_private;
+struct intel_gt;
+struct drm_printer;
+
+/*
+ * Maximum number of slices on older platforms.  Slices no longer exist
+ * starting on Xe_HP ("gslices," "cslices," etc. are a different concept and
+ * are not expressed through fusing).
+ */
+#define GEN_MAX_HSW_SLICES		3
+
+/*
+ * Maximum number of subslices that can exist within a HSW-style slice.  This
+ * is only relevant to pre-Xe_HP platforms (Xe_HP and beyond use the
+ * I915_MAX_SS_FUSE_BITS value below).
+ */
+#define GEN_MAX_SS_PER_HSW_SLICE	8
+
+/*
+ * Maximum number of 32-bit registers used by hardware to express the
+ * enabled/disabled subslices.
+ */
+#define I915_MAX_SS_FUSE_REGS	2
+#define I915_MAX_SS_FUSE_BITS	(I915_MAX_SS_FUSE_REGS * 32)
+
+/* Maximum number of EUs that can exist within a subslice or DSS. */
+#define GEN_MAX_EUS_PER_SS		16
+
+#define SSEU_MAX(a, b)			((a) > (b) ? (a) : (b))
+
+/* The maximum number of bits needed to express each subslice/DSS independently */
+#define GEN_SS_MASK_SIZE		SSEU_MAX(I915_MAX_SS_FUSE_BITS, \
+						 GEN_MAX_HSW_SLICES * GEN_MAX_SS_PER_HSW_SLICE)
+
+#define GEN_SSEU_STRIDE(max_entries)	DIV_ROUND_UP(max_entries, BITS_PER_BYTE)
+#define GEN_MAX_SUBSLICE_STRIDE		GEN_SSEU_STRIDE(GEN_SS_MASK_SIZE)
+#define GEN_MAX_EU_STRIDE		GEN_SSEU_STRIDE(GEN_MAX_EUS_PER_SS)
+
+#define GEN_DSS_PER_GSLICE	4
+#define GEN_DSS_PER_CSLICE	8
+#define GEN_DSS_PER_MSLICE	8
+
+#define GEN_MAX_GSLICES		(I915_MAX_SS_FUSE_BITS / GEN_DSS_PER_GSLICE)
+#define GEN_MAX_CSLICES		(I915_MAX_SS_FUSE_BITS / GEN_DSS_PER_CSLICE)
+
+typedef union {
+	u8 hsw[GEN_MAX_HSW_SLICES];
+
+	/* Bitmap compatible with linux/bitmap.h; may exceed size of u64 */
+	unsigned long xehp[BITS_TO_LONGS(I915_MAX_SS_FUSE_BITS)];
+} intel_sseu_ss_mask_t;
+
+#define XEHP_BITMAP_BITS(mask)	((int)BITS_PER_TYPE(typeof(mask.xehp)))
+
+struct sseu_dev_info {
+	u8 slice_mask;
+	intel_sseu_ss_mask_t subslice_mask;
+	intel_sseu_ss_mask_t geometry_subslice_mask;
+	intel_sseu_ss_mask_t compute_subslice_mask;
+	union {
+		u16 hsw[GEN_MAX_HSW_SLICES][GEN_MAX_SS_PER_HSW_SLICE];
+		u16 xehp[I915_MAX_SS_FUSE_BITS];
+	} eu_mask;
+
+	u16 eu_total;
+	u8 eu_per_subslice;
+	u8 min_eu_in_pool;
+	/* For each slice, which subslice(s) has(have) 7 EUs (bitfield)? */
+	u8 subslice_7eu[3];
+	u8 has_slice_pg:1;
+	u8 has_subslice_pg:1;
+	u8 has_eu_pg:1;
+	/*
+	 * For Xe_HP and beyond, the hardware no longer has traditional slices
+	 * so we just report the entire DSS pool under a fake "slice 0."
+	 */
+	u8 has_xehp_dss:1;
+
+	/* Topology fields */
+	u8 max_slices;
+	u8 max_subslices;
+	u8 max_eus_per_subslice;
+};
+
+/*
+ * Powergating configuration for a particular (context,engine).
+ */
+struct intel_sseu {
+	u8 slice_mask;
+	u8 subslice_mask;
+	u8 min_eus_per_subslice;
+	u8 max_eus_per_subslice;
+};
+
+static inline struct intel_sseu
+intel_sseu_from_device_info(const struct sseu_dev_info *sseu)
+{
+	struct intel_sseu value = {
+		.slice_mask = sseu->slice_mask,
+		.subslice_mask = sseu->subslice_mask.hsw[0],
+		.min_eus_per_subslice = sseu->max_eus_per_subslice,
+		.max_eus_per_subslice = sseu->max_eus_per_subslice,
+	};
+
+	return value;
+}
+
+static inline bool
+intel_sseu_has_subslice(const struct sseu_dev_info *sseu, int slice,
+			int subslice)
+{
+	if (slice >= sseu->max_slices ||
+	    subslice >= sseu->max_subslices)
+		return false;
+
+	if (sseu->has_xehp_dss)
+		return test_bit(subslice, sseu->subslice_mask.xehp);
+	else
+		return sseu->subslice_mask.hsw[slice] & BIT(subslice);
+}
+
+/*
+ * Used to obtain the index of the first DSS.  Can start searching from the
+ * beginning of a specific dss group (e.g., gslice, cslice, etc.) if
+ * groupsize and groupnum are non-zero.
+ */
+static inline unsigned int
+intel_sseu_find_first_xehp_dss(const struct sseu_dev_info *sseu, int groupsize,
+			       int groupnum)
+{
+	return find_next_bit(sseu->subslice_mask.xehp,
+			     XEHP_BITMAP_BITS(sseu->subslice_mask),
+			     groupnum * groupsize);
+}
+
+void intel_sseu_set_info(struct sseu_dev_info *sseu, u8 max_slices,
+			 u8 max_subslices, u8 max_eus_per_subslice);
+
+unsigned int
+intel_sseu_subslice_total(const struct sseu_dev_info *sseu);
+
+unsigned int
+intel_sseu_get_hsw_subslices(const struct sseu_dev_info *sseu, u8 slice);
+
+intel_sseu_ss_mask_t
+intel_sseu_get_compute_subslices(const struct sseu_dev_info *sseu);
+
+void intel_sseu_info_init(struct intel_gt *gt);
+
+u32 intel_sseu_make_rpcs(struct intel_gt *gt,
+			 const struct intel_sseu *req_sseu);
+
+void intel_sseu_dump(const struct sseu_dev_info *sseu, struct drm_printer *p);
+void intel_sseu_print_topology(struct drm_i915_private *i915,
+			       const struct sseu_dev_info *sseu,
+			       struct drm_printer *p);
+
+u16 intel_slicemask_from_xehp_dssmask(intel_sseu_ss_mask_t dss_mask, int dss_per_slice);
+
+int intel_sseu_copy_eumask_to_user(void __user *to,
+				   const struct sseu_dev_info *sseu);
+int intel_sseu_copy_ssmask_to_user(void __user *to,
+				   const struct sseu_dev_info *sseu);
+
+void intel_sseu_print_ss_info(const char *type,
+			      const struct sseu_dev_info *sseu,
+			      struct seq_file *m);
+
+#endif /* __INTEL_SSEU_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_sseu_debugfs.c b/drivers/gpu/drm/i915/gt/intel_sseu_debugfs.c
new file mode 100644
index 000000000..c2ee5e182
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_sseu_debugfs.c
@@ -0,0 +1,299 @@
+// SPDX-License-Identifier: MIT
+
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/bitmap.h>
+#include <linux/string_helpers.h>
+
+#include "i915_drv.h"
+#include "intel_gt_debugfs.h"
+#include "intel_gt_regs.h"
+#include "intel_sseu_debugfs.h"
+
+static void cherryview_sseu_device_status(struct intel_gt *gt,
+					  struct sseu_dev_info *sseu)
+{
+#define SS_MAX 2
+	struct intel_uncore *uncore = gt->uncore;
+	const int ss_max = SS_MAX;
+	u32 sig1[SS_MAX], sig2[SS_MAX];
+	int ss;
+
+	sig1[0] = intel_uncore_read(uncore, CHV_POWER_SS0_SIG1);
+	sig1[1] = intel_uncore_read(uncore, CHV_POWER_SS1_SIG1);
+	sig2[0] = intel_uncore_read(uncore, CHV_POWER_SS0_SIG2);
+	sig2[1] = intel_uncore_read(uncore, CHV_POWER_SS1_SIG2);
+
+	for (ss = 0; ss < ss_max; ss++) {
+		unsigned int eu_cnt;
+
+		if (sig1[ss] & CHV_SS_PG_ENABLE)
+			/* skip disabled subslice */
+			continue;
+
+		sseu->slice_mask = BIT(0);
+		sseu->subslice_mask.hsw[0] |= BIT(ss);
+		eu_cnt = ((sig1[ss] & CHV_EU08_PG_ENABLE) ? 0 : 2) +
+			 ((sig1[ss] & CHV_EU19_PG_ENABLE) ? 0 : 2) +
+			 ((sig1[ss] & CHV_EU210_PG_ENABLE) ? 0 : 2) +
+			 ((sig2[ss] & CHV_EU311_PG_ENABLE) ? 0 : 2);
+		sseu->eu_total += eu_cnt;
+		sseu->eu_per_subslice = max_t(unsigned int,
+					      sseu->eu_per_subslice, eu_cnt);
+	}
+#undef SS_MAX
+}
+
+static void gen11_sseu_device_status(struct intel_gt *gt,
+				     struct sseu_dev_info *sseu)
+{
+#define SS_MAX 8
+	struct intel_uncore *uncore = gt->uncore;
+	const struct intel_gt_info *info = &gt->info;
+	u32 s_reg[SS_MAX], eu_reg[2 * SS_MAX], eu_mask[2];
+	int s, ss;
+
+	for (s = 0; s < info->sseu.max_slices; s++) {
+		/*
+		 * FIXME: Valid SS Mask respects the spec and read
+		 * only valid bits for those registers, excluding reserved
+		 * although this seems wrong because it would leave many
+		 * subslices without ACK.
+		 */
+		s_reg[s] = intel_uncore_read(uncore, GEN10_SLICE_PGCTL_ACK(s)) &
+			GEN10_PGCTL_VALID_SS_MASK(s);
+		eu_reg[2 * s] = intel_uncore_read(uncore,
+						  GEN10_SS01_EU_PGCTL_ACK(s));
+		eu_reg[2 * s + 1] = intel_uncore_read(uncore,
+						      GEN10_SS23_EU_PGCTL_ACK(s));
+	}
+
+	eu_mask[0] = GEN9_PGCTL_SSA_EU08_ACK |
+		     GEN9_PGCTL_SSA_EU19_ACK |
+		     GEN9_PGCTL_SSA_EU210_ACK |
+		     GEN9_PGCTL_SSA_EU311_ACK;
+	eu_mask[1] = GEN9_PGCTL_SSB_EU08_ACK |
+		     GEN9_PGCTL_SSB_EU19_ACK |
+		     GEN9_PGCTL_SSB_EU210_ACK |
+		     GEN9_PGCTL_SSB_EU311_ACK;
+
+	for (s = 0; s < info->sseu.max_slices; s++) {
+		if ((s_reg[s] & GEN9_PGCTL_SLICE_ACK) == 0)
+			/* skip disabled slice */
+			continue;
+
+		sseu->slice_mask |= BIT(s);
+		sseu->subslice_mask.hsw[s] = info->sseu.subslice_mask.hsw[s];
+
+		for (ss = 0; ss < info->sseu.max_subslices; ss++) {
+			unsigned int eu_cnt;
+
+			if (info->sseu.has_subslice_pg &&
+			    !(s_reg[s] & (GEN9_PGCTL_SS_ACK(ss))))
+				/* skip disabled subslice */
+				continue;
+
+			eu_cnt = 2 * hweight32(eu_reg[2 * s + ss / 2] &
+					       eu_mask[ss % 2]);
+			sseu->eu_total += eu_cnt;
+			sseu->eu_per_subslice = max_t(unsigned int,
+						      sseu->eu_per_subslice,
+						      eu_cnt);
+		}
+	}
+#undef SS_MAX
+}
+
+static void gen9_sseu_device_status(struct intel_gt *gt,
+				    struct sseu_dev_info *sseu)
+{
+#define SS_MAX 3
+	struct intel_uncore *uncore = gt->uncore;
+	const struct intel_gt_info *info = &gt->info;
+	u32 s_reg[SS_MAX], eu_reg[2 * SS_MAX], eu_mask[2];
+	int s, ss;
+
+	for (s = 0; s < info->sseu.max_slices; s++) {
+		s_reg[s] = intel_uncore_read(uncore, GEN9_SLICE_PGCTL_ACK(s));
+		eu_reg[2 * s] =
+			intel_uncore_read(uncore, GEN9_SS01_EU_PGCTL_ACK(s));
+		eu_reg[2 * s + 1] =
+			intel_uncore_read(uncore, GEN9_SS23_EU_PGCTL_ACK(s));
+	}
+
+	eu_mask[0] = GEN9_PGCTL_SSA_EU08_ACK |
+		     GEN9_PGCTL_SSA_EU19_ACK |
+		     GEN9_PGCTL_SSA_EU210_ACK |
+		     GEN9_PGCTL_SSA_EU311_ACK;
+	eu_mask[1] = GEN9_PGCTL_SSB_EU08_ACK |
+		     GEN9_PGCTL_SSB_EU19_ACK |
+		     GEN9_PGCTL_SSB_EU210_ACK |
+		     GEN9_PGCTL_SSB_EU311_ACK;
+
+	for (s = 0; s < info->sseu.max_slices; s++) {
+		if ((s_reg[s] & GEN9_PGCTL_SLICE_ACK) == 0)
+			/* skip disabled slice */
+			continue;
+
+		sseu->slice_mask |= BIT(s);
+
+		if (IS_GEN9_BC(gt->i915))
+			sseu->subslice_mask.hsw[s] = info->sseu.subslice_mask.hsw[s];
+
+		for (ss = 0; ss < info->sseu.max_subslices; ss++) {
+			unsigned int eu_cnt;
+
+			if (IS_GEN9_LP(gt->i915)) {
+				if (!(s_reg[s] & (GEN9_PGCTL_SS_ACK(ss))))
+					/* skip disabled subslice */
+					continue;
+
+				sseu->subslice_mask.hsw[s] |= BIT(ss);
+			}
+
+			eu_cnt = eu_reg[2 * s + ss / 2] & eu_mask[ss % 2];
+			eu_cnt = 2 * hweight32(eu_cnt);
+
+			sseu->eu_total += eu_cnt;
+			sseu->eu_per_subslice = max_t(unsigned int,
+						      sseu->eu_per_subslice,
+						      eu_cnt);
+		}
+	}
+#undef SS_MAX
+}
+
+static void bdw_sseu_device_status(struct intel_gt *gt,
+				   struct sseu_dev_info *sseu)
+{
+	const struct intel_gt_info *info = &gt->info;
+	u32 slice_info = intel_uncore_read(gt->uncore, GEN8_GT_SLICE_INFO);
+	int s;
+
+	sseu->slice_mask = slice_info & GEN8_LSLICESTAT_MASK;
+
+	if (sseu->slice_mask) {
+		sseu->eu_per_subslice = info->sseu.eu_per_subslice;
+		for (s = 0; s < fls(sseu->slice_mask); s++)
+			sseu->subslice_mask.hsw[s] = info->sseu.subslice_mask.hsw[s];
+		sseu->eu_total = sseu->eu_per_subslice *
+				 intel_sseu_subslice_total(sseu);
+
+		/* subtract fused off EU(s) from enabled slice(s) */
+		for (s = 0; s < fls(sseu->slice_mask); s++) {
+			u8 subslice_7eu = info->sseu.subslice_7eu[s];
+
+			sseu->eu_total -= hweight8(subslice_7eu);
+		}
+	}
+}
+
+static void i915_print_sseu_info(struct seq_file *m,
+				 bool is_available_info,
+				 bool has_pooled_eu,
+				 const struct sseu_dev_info *sseu)
+{
+	const char *type = is_available_info ? "Available" : "Enabled";
+
+	seq_printf(m, "  %s Slice Mask: %04x\n", type,
+		   sseu->slice_mask);
+	seq_printf(m, "  %s Slice Total: %u\n", type,
+		   hweight8(sseu->slice_mask));
+	seq_printf(m, "  %s Subslice Total: %u\n", type,
+		   intel_sseu_subslice_total(sseu));
+	intel_sseu_print_ss_info(type, sseu, m);
+	seq_printf(m, "  %s EU Total: %u\n", type,
+		   sseu->eu_total);
+	seq_printf(m, "  %s EU Per Subslice: %u\n", type,
+		   sseu->eu_per_subslice);
+
+	if (!is_available_info)
+		return;
+
+	seq_printf(m, "  Has Pooled EU: %s\n", str_yes_no(has_pooled_eu));
+	if (has_pooled_eu)
+		seq_printf(m, "  Min EU in pool: %u\n", sseu->min_eu_in_pool);
+
+	seq_printf(m, "  Has Slice Power Gating: %s\n",
+		   str_yes_no(sseu->has_slice_pg));
+	seq_printf(m, "  Has Subslice Power Gating: %s\n",
+		   str_yes_no(sseu->has_subslice_pg));
+	seq_printf(m, "  Has EU Power Gating: %s\n",
+		   str_yes_no(sseu->has_eu_pg));
+}
+
+/*
+ * this is called from top-level debugfs as well, so we can't get the gt from
+ * the seq_file.
+ */
+int intel_sseu_status(struct seq_file *m, struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	const struct intel_gt_info *info = &gt->info;
+	struct sseu_dev_info *sseu;
+	intel_wakeref_t wakeref;
+
+	if (GRAPHICS_VER(i915) < 8)
+		return -ENODEV;
+
+	seq_puts(m, "SSEU Device Info\n");
+	i915_print_sseu_info(m, true, HAS_POOLED_EU(i915), &info->sseu);
+
+	seq_puts(m, "SSEU Device Status\n");
+
+	sseu = kzalloc(sizeof(*sseu), GFP_KERNEL);
+	if (!sseu)
+		return -ENOMEM;
+
+	intel_sseu_set_info(sseu, info->sseu.max_slices,
+			    info->sseu.max_subslices,
+			    info->sseu.max_eus_per_subslice);
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		if (IS_CHERRYVIEW(i915))
+			cherryview_sseu_device_status(gt, sseu);
+		else if (IS_BROADWELL(i915))
+			bdw_sseu_device_status(gt, sseu);
+		else if (GRAPHICS_VER(i915) == 9)
+			gen9_sseu_device_status(gt, sseu);
+		else if (GRAPHICS_VER(i915) >= 11)
+			gen11_sseu_device_status(gt, sseu);
+	}
+
+	i915_print_sseu_info(m, false, HAS_POOLED_EU(i915), sseu);
+
+	kfree(sseu);
+
+	return 0;
+}
+
+static int sseu_status_show(struct seq_file *m, void *unused)
+{
+	struct intel_gt *gt = m->private;
+
+	return intel_sseu_status(m, gt);
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(sseu_status);
+
+static int sseu_topology_show(struct seq_file *m, void *unused)
+{
+	struct intel_gt *gt = m->private;
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	intel_sseu_print_topology(gt->i915, &gt->info.sseu, &p);
+
+	return 0;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(sseu_topology);
+
+void intel_sseu_debugfs_register(struct intel_gt *gt, struct dentry *root)
+{
+	static const struct intel_gt_debugfs_file files[] = {
+		{ "sseu_status", &sseu_status_fops, NULL },
+		{ "sseu_topology", &sseu_topology_fops, NULL },
+	};
+
+	intel_gt_debugfs_register_files(root, files, ARRAY_SIZE(files), gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/intel_sseu_debugfs.h b/drivers/gpu/drm/i915/gt/intel_sseu_debugfs.h
new file mode 100644
index 000000000..73f001589
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_sseu_debugfs.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: MIT */
+
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef INTEL_SSEU_DEBUGFS_H
+#define INTEL_SSEU_DEBUGFS_H
+
+struct intel_gt;
+struct dentry;
+struct seq_file;
+
+int intel_sseu_status(struct seq_file *m, struct intel_gt *gt);
+void intel_sseu_debugfs_register(struct intel_gt *gt, struct dentry *root);
+
+#endif /* INTEL_SSEU_DEBUGFS_H */
diff --git a/drivers/gpu/drm/i915/gt/intel_timeline.c b/drivers/gpu/drm/i915/gt/intel_timeline.c
new file mode 100644
index 000000000..b9640212d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_timeline.c
@@ -0,0 +1,493 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2016-2018 Intel Corporation
+ */
+
+#include <drm/drm_cache.h>
+
+#include "gem/i915_gem_internal.h"
+
+#include "i915_active.h"
+#include "i915_drv.h"
+#include "i915_syncmap.h"
+#include "intel_gt.h"
+#include "intel_ring.h"
+#include "intel_timeline.h"
+
+#define TIMELINE_SEQNO_BYTES 8
+
+static struct i915_vma *hwsp_alloc(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	i915_gem_object_set_cache_coherency(obj, I915_CACHE_LLC);
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma))
+		i915_gem_object_put(obj);
+
+	return vma;
+}
+
+static void __timeline_retire(struct i915_active *active)
+{
+	struct intel_timeline *tl =
+		container_of(active, typeof(*tl), active);
+
+	i915_vma_unpin(tl->hwsp_ggtt);
+	intel_timeline_put(tl);
+}
+
+static int __timeline_active(struct i915_active *active)
+{
+	struct intel_timeline *tl =
+		container_of(active, typeof(*tl), active);
+
+	__i915_vma_pin(tl->hwsp_ggtt);
+	intel_timeline_get(tl);
+	return 0;
+}
+
+I915_SELFTEST_EXPORT int
+intel_timeline_pin_map(struct intel_timeline *timeline)
+{
+	struct drm_i915_gem_object *obj = timeline->hwsp_ggtt->obj;
+	u32 ofs = offset_in_page(timeline->hwsp_offset);
+	void *vaddr;
+
+	vaddr = i915_gem_object_pin_map(obj, I915_MAP_WB);
+	if (IS_ERR(vaddr))
+		return PTR_ERR(vaddr);
+
+	timeline->hwsp_map = vaddr;
+	timeline->hwsp_seqno = memset(vaddr + ofs, 0, TIMELINE_SEQNO_BYTES);
+	drm_clflush_virt_range(vaddr + ofs, TIMELINE_SEQNO_BYTES);
+
+	return 0;
+}
+
+static int intel_timeline_init(struct intel_timeline *timeline,
+			       struct intel_gt *gt,
+			       struct i915_vma *hwsp,
+			       unsigned int offset)
+{
+	kref_init(&timeline->kref);
+	atomic_set(&timeline->pin_count, 0);
+
+	timeline->gt = gt;
+
+	if (hwsp) {
+		timeline->hwsp_offset = offset;
+		timeline->hwsp_ggtt = i915_vma_get(hwsp);
+	} else {
+		timeline->has_initial_breadcrumb = true;
+		hwsp = hwsp_alloc(gt);
+		if (IS_ERR(hwsp))
+			return PTR_ERR(hwsp);
+		timeline->hwsp_ggtt = hwsp;
+	}
+
+	timeline->hwsp_map = NULL;
+	timeline->hwsp_seqno = (void *)(long)timeline->hwsp_offset;
+
+	GEM_BUG_ON(timeline->hwsp_offset >= hwsp->size);
+
+	timeline->fence_context = dma_fence_context_alloc(1);
+
+	mutex_init(&timeline->mutex);
+
+	INIT_ACTIVE_FENCE(&timeline->last_request);
+	INIT_LIST_HEAD(&timeline->requests);
+
+	i915_syncmap_init(&timeline->sync);
+	i915_active_init(&timeline->active, __timeline_active,
+			 __timeline_retire, 0);
+
+	return 0;
+}
+
+void intel_gt_init_timelines(struct intel_gt *gt)
+{
+	struct intel_gt_timelines *timelines = &gt->timelines;
+
+	spin_lock_init(&timelines->lock);
+	INIT_LIST_HEAD(&timelines->active_list);
+}
+
+static void intel_timeline_fini(struct rcu_head *rcu)
+{
+	struct intel_timeline *timeline =
+		container_of(rcu, struct intel_timeline, rcu);
+
+	if (timeline->hwsp_map)
+		i915_gem_object_unpin_map(timeline->hwsp_ggtt->obj);
+
+	i915_vma_put(timeline->hwsp_ggtt);
+	i915_active_fini(&timeline->active);
+
+	/*
+	 * A small race exists between intel_gt_retire_requests_timeout and
+	 * intel_timeline_exit which could result in the syncmap not getting
+	 * free'd. Rather than work to hard to seal this race, simply cleanup
+	 * the syncmap on fini.
+	 */
+	i915_syncmap_free(&timeline->sync);
+
+	kfree(timeline);
+}
+
+struct intel_timeline *
+__intel_timeline_create(struct intel_gt *gt,
+			struct i915_vma *global_hwsp,
+			unsigned int offset)
+{
+	struct intel_timeline *timeline;
+	int err;
+
+	timeline = kzalloc(sizeof(*timeline), GFP_KERNEL);
+	if (!timeline)
+		return ERR_PTR(-ENOMEM);
+
+	err = intel_timeline_init(timeline, gt, global_hwsp, offset);
+	if (err) {
+		kfree(timeline);
+		return ERR_PTR(err);
+	}
+
+	return timeline;
+}
+
+struct intel_timeline *
+intel_timeline_create_from_engine(struct intel_engine_cs *engine,
+				  unsigned int offset)
+{
+	struct i915_vma *hwsp = engine->status_page.vma;
+	struct intel_timeline *tl;
+
+	tl = __intel_timeline_create(engine->gt, hwsp, offset);
+	if (IS_ERR(tl))
+		return tl;
+
+	/* Borrow a nearby lock; we only create these timelines during init */
+	mutex_lock(&hwsp->vm->mutex);
+	list_add_tail(&tl->engine_link, &engine->status_page.timelines);
+	mutex_unlock(&hwsp->vm->mutex);
+
+	return tl;
+}
+
+void __intel_timeline_pin(struct intel_timeline *tl)
+{
+	GEM_BUG_ON(!atomic_read(&tl->pin_count));
+	atomic_inc(&tl->pin_count);
+}
+
+int intel_timeline_pin(struct intel_timeline *tl, struct i915_gem_ww_ctx *ww)
+{
+	int err;
+
+	if (atomic_add_unless(&tl->pin_count, 1, 0))
+		return 0;
+
+	if (!tl->hwsp_map) {
+		err = intel_timeline_pin_map(tl);
+		if (err)
+			return err;
+	}
+
+	err = i915_ggtt_pin(tl->hwsp_ggtt, ww, 0, PIN_HIGH);
+	if (err)
+		return err;
+
+	tl->hwsp_offset =
+		i915_ggtt_offset(tl->hwsp_ggtt) +
+		offset_in_page(tl->hwsp_offset);
+	GT_TRACE(tl->gt, "timeline:%llx using HWSP offset:%x\n",
+		 tl->fence_context, tl->hwsp_offset);
+
+	i915_active_acquire(&tl->active);
+	if (atomic_fetch_inc(&tl->pin_count)) {
+		i915_active_release(&tl->active);
+		__i915_vma_unpin(tl->hwsp_ggtt);
+	}
+
+	return 0;
+}
+
+void intel_timeline_reset_seqno(const struct intel_timeline *tl)
+{
+	u32 *hwsp_seqno = (u32 *)tl->hwsp_seqno;
+	/* Must be pinned to be writable, and no requests in flight. */
+	GEM_BUG_ON(!atomic_read(&tl->pin_count));
+
+	memset(hwsp_seqno + 1, 0, TIMELINE_SEQNO_BYTES - sizeof(*hwsp_seqno));
+	WRITE_ONCE(*hwsp_seqno, tl->seqno);
+	drm_clflush_virt_range(hwsp_seqno, TIMELINE_SEQNO_BYTES);
+}
+
+void intel_timeline_enter(struct intel_timeline *tl)
+{
+	struct intel_gt_timelines *timelines = &tl->gt->timelines;
+
+	/*
+	 * Pretend we are serialised by the timeline->mutex.
+	 *
+	 * While generally true, there are a few exceptions to the rule
+	 * for the engine->kernel_context being used to manage power
+	 * transitions. As the engine_park may be called from under any
+	 * timeline, it uses the power mutex as a global serialisation
+	 * lock to prevent any other request entering its timeline.
+	 *
+	 * The rule is generally tl->mutex, otherwise engine->wakeref.mutex.
+	 *
+	 * However, intel_gt_retire_request() does not know which engine
+	 * it is retiring along and so cannot partake in the engine-pm
+	 * barrier, and there we use the tl->active_count as a means to
+	 * pin the timeline in the active_list while the locks are dropped.
+	 * Ergo, as that is outside of the engine-pm barrier, we need to
+	 * use atomic to manipulate tl->active_count.
+	 */
+	lockdep_assert_held(&tl->mutex);
+
+	if (atomic_add_unless(&tl->active_count, 1, 0))
+		return;
+
+	spin_lock(&timelines->lock);
+	if (!atomic_fetch_inc(&tl->active_count)) {
+		/*
+		 * The HWSP is volatile, and may have been lost while inactive,
+		 * e.g. across suspend/resume. Be paranoid, and ensure that
+		 * the HWSP value matches our seqno so we don't proclaim
+		 * the next request as already complete.
+		 */
+		intel_timeline_reset_seqno(tl);
+		list_add_tail(&tl->link, &timelines->active_list);
+	}
+	spin_unlock(&timelines->lock);
+}
+
+void intel_timeline_exit(struct intel_timeline *tl)
+{
+	struct intel_gt_timelines *timelines = &tl->gt->timelines;
+
+	/* See intel_timeline_enter() */
+	lockdep_assert_held(&tl->mutex);
+
+	GEM_BUG_ON(!atomic_read(&tl->active_count));
+	if (atomic_add_unless(&tl->active_count, -1, 1))
+		return;
+
+	spin_lock(&timelines->lock);
+	if (atomic_dec_and_test(&tl->active_count))
+		list_del(&tl->link);
+	spin_unlock(&timelines->lock);
+
+	/*
+	 * Since this timeline is idle, all bariers upon which we were waiting
+	 * must also be complete and so we can discard the last used barriers
+	 * without loss of information.
+	 */
+	i915_syncmap_free(&tl->sync);
+}
+
+static u32 timeline_advance(struct intel_timeline *tl)
+{
+	GEM_BUG_ON(!atomic_read(&tl->pin_count));
+	GEM_BUG_ON(tl->seqno & tl->has_initial_breadcrumb);
+
+	return tl->seqno += 1 + tl->has_initial_breadcrumb;
+}
+
+static noinline int
+__intel_timeline_get_seqno(struct intel_timeline *tl,
+			   u32 *seqno)
+{
+	u32 next_ofs = offset_in_page(tl->hwsp_offset + TIMELINE_SEQNO_BYTES);
+
+	/* w/a: bit 5 needs to be zero for MI_FLUSH_DW address. */
+	if (TIMELINE_SEQNO_BYTES <= BIT(5) && (next_ofs & BIT(5)))
+		next_ofs = offset_in_page(next_ofs + BIT(5));
+
+	tl->hwsp_offset = i915_ggtt_offset(tl->hwsp_ggtt) + next_ofs;
+	tl->hwsp_seqno = tl->hwsp_map + next_ofs;
+	intel_timeline_reset_seqno(tl);
+
+	*seqno = timeline_advance(tl);
+	GEM_BUG_ON(i915_seqno_passed(*tl->hwsp_seqno, *seqno));
+	return 0;
+}
+
+int intel_timeline_get_seqno(struct intel_timeline *tl,
+			     struct i915_request *rq,
+			     u32 *seqno)
+{
+	*seqno = timeline_advance(tl);
+
+	/* Replace the HWSP on wraparound for HW semaphores */
+	if (unlikely(!*seqno && tl->has_initial_breadcrumb))
+		return __intel_timeline_get_seqno(tl, seqno);
+
+	return 0;
+}
+
+int intel_timeline_read_hwsp(struct i915_request *from,
+			     struct i915_request *to,
+			     u32 *hwsp)
+{
+	struct intel_timeline *tl;
+	int err;
+
+	rcu_read_lock();
+	tl = rcu_dereference(from->timeline);
+	if (i915_request_signaled(from) ||
+	    !i915_active_acquire_if_busy(&tl->active))
+		tl = NULL;
+
+	if (tl) {
+		/* hwsp_offset may wraparound, so use from->hwsp_seqno */
+		*hwsp = i915_ggtt_offset(tl->hwsp_ggtt) +
+			offset_in_page(from->hwsp_seqno);
+	}
+
+	/* ensure we wait on the right request, if not, we completed */
+	if (tl && __i915_request_is_complete(from)) {
+		i915_active_release(&tl->active);
+		tl = NULL;
+	}
+	rcu_read_unlock();
+
+	if (!tl)
+		return 1;
+
+	/* Can't do semaphore waits on kernel context */
+	if (!tl->has_initial_breadcrumb) {
+		err = -EINVAL;
+		goto out;
+	}
+
+	err = i915_active_add_request(&tl->active, to);
+
+out:
+	i915_active_release(&tl->active);
+	return err;
+}
+
+void intel_timeline_unpin(struct intel_timeline *tl)
+{
+	GEM_BUG_ON(!atomic_read(&tl->pin_count));
+	if (!atomic_dec_and_test(&tl->pin_count))
+		return;
+
+	i915_active_release(&tl->active);
+	__i915_vma_unpin(tl->hwsp_ggtt);
+}
+
+void __intel_timeline_free(struct kref *kref)
+{
+	struct intel_timeline *timeline =
+		container_of(kref, typeof(*timeline), kref);
+
+	GEM_BUG_ON(atomic_read(&timeline->pin_count));
+	GEM_BUG_ON(!list_empty(&timeline->requests));
+	GEM_BUG_ON(timeline->retire);
+
+	call_rcu(&timeline->rcu, intel_timeline_fini);
+}
+
+void intel_gt_fini_timelines(struct intel_gt *gt)
+{
+	struct intel_gt_timelines *timelines = &gt->timelines;
+
+	GEM_BUG_ON(!list_empty(&timelines->active_list));
+}
+
+void intel_gt_show_timelines(struct intel_gt *gt,
+			     struct drm_printer *m,
+			     void (*show_request)(struct drm_printer *m,
+						  const struct i915_request *rq,
+						  const char *prefix,
+						  int indent))
+{
+	struct intel_gt_timelines *timelines = &gt->timelines;
+	struct intel_timeline *tl, *tn;
+	LIST_HEAD(free);
+
+	spin_lock(&timelines->lock);
+	list_for_each_entry_safe(tl, tn, &timelines->active_list, link) {
+		unsigned long count, ready, inflight;
+		struct i915_request *rq, *rn;
+		struct dma_fence *fence;
+
+		if (!mutex_trylock(&tl->mutex)) {
+			drm_printf(m, "Timeline %llx: busy; skipping\n",
+				   tl->fence_context);
+			continue;
+		}
+
+		intel_timeline_get(tl);
+		GEM_BUG_ON(!atomic_read(&tl->active_count));
+		atomic_inc(&tl->active_count); /* pin the list element */
+		spin_unlock(&timelines->lock);
+
+		count = 0;
+		ready = 0;
+		inflight = 0;
+		list_for_each_entry_safe(rq, rn, &tl->requests, link) {
+			if (i915_request_completed(rq))
+				continue;
+
+			count++;
+			if (i915_request_is_ready(rq))
+				ready++;
+			if (i915_request_is_active(rq))
+				inflight++;
+		}
+
+		drm_printf(m, "Timeline %llx: { ", tl->fence_context);
+		drm_printf(m, "count: %lu, ready: %lu, inflight: %lu",
+			   count, ready, inflight);
+		drm_printf(m, ", seqno: { current: %d, last: %d }",
+			   *tl->hwsp_seqno, tl->seqno);
+		fence = i915_active_fence_get(&tl->last_request);
+		if (fence) {
+			drm_printf(m, ", engine: %s",
+				   to_request(fence)->engine->name);
+			dma_fence_put(fence);
+		}
+		drm_printf(m, " }\n");
+
+		if (show_request) {
+			list_for_each_entry_safe(rq, rn, &tl->requests, link)
+				show_request(m, rq, "", 2);
+		}
+
+		mutex_unlock(&tl->mutex);
+		spin_lock(&timelines->lock);
+
+		/* Resume list iteration after reacquiring spinlock */
+		list_safe_reset_next(tl, tn, link);
+		if (atomic_dec_and_test(&tl->active_count))
+			list_del(&tl->link);
+
+		/* Defer the final release to after the spinlock */
+		if (refcount_dec_and_test(&tl->kref.refcount)) {
+			GEM_BUG_ON(atomic_read(&tl->active_count));
+			list_add(&tl->link, &free);
+		}
+	}
+	spin_unlock(&timelines->lock);
+
+	list_for_each_entry_safe(tl, tn, &free, link)
+		__intel_timeline_free(&tl->kref);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "gt/selftests/mock_timeline.c"
+#include "gt/selftest_timeline.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_timeline.h b/drivers/gpu/drm/i915/gt/intel_timeline.h
new file mode 100644
index 000000000..57308c4d6
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_timeline.h
@@ -0,0 +1,103 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef I915_TIMELINE_H
+#define I915_TIMELINE_H
+
+#include <linux/lockdep.h>
+
+#include "i915_active.h"
+#include "i915_syncmap.h"
+#include "intel_timeline_types.h"
+
+struct drm_printer;
+
+struct intel_timeline *
+__intel_timeline_create(struct intel_gt *gt,
+			struct i915_vma *global_hwsp,
+			unsigned int offset);
+
+static inline struct intel_timeline *
+intel_timeline_create(struct intel_gt *gt)
+{
+	return __intel_timeline_create(gt, NULL, 0);
+}
+
+struct intel_timeline *
+intel_timeline_create_from_engine(struct intel_engine_cs *engine,
+				  unsigned int offset);
+
+static inline struct intel_timeline *
+intel_timeline_get(struct intel_timeline *timeline)
+{
+	kref_get(&timeline->kref);
+	return timeline;
+}
+
+void __intel_timeline_free(struct kref *kref);
+static inline void intel_timeline_put(struct intel_timeline *timeline)
+{
+	kref_put(&timeline->kref, __intel_timeline_free);
+}
+
+static inline int __intel_timeline_sync_set(struct intel_timeline *tl,
+					    u64 context, u32 seqno)
+{
+	return i915_syncmap_set(&tl->sync, context, seqno);
+}
+
+static inline int intel_timeline_sync_set(struct intel_timeline *tl,
+					  const struct dma_fence *fence)
+{
+	return __intel_timeline_sync_set(tl, fence->context, fence->seqno);
+}
+
+static inline bool __intel_timeline_sync_is_later(struct intel_timeline *tl,
+						  u64 context, u32 seqno)
+{
+	return i915_syncmap_is_later(&tl->sync, context, seqno);
+}
+
+static inline bool intel_timeline_sync_is_later(struct intel_timeline *tl,
+						const struct dma_fence *fence)
+{
+	return __intel_timeline_sync_is_later(tl, fence->context, fence->seqno);
+}
+
+void __intel_timeline_pin(struct intel_timeline *tl);
+int intel_timeline_pin(struct intel_timeline *tl, struct i915_gem_ww_ctx *ww);
+void intel_timeline_enter(struct intel_timeline *tl);
+int intel_timeline_get_seqno(struct intel_timeline *tl,
+			     struct i915_request *rq,
+			     u32 *seqno);
+void intel_timeline_exit(struct intel_timeline *tl);
+void intel_timeline_unpin(struct intel_timeline *tl);
+
+void intel_timeline_reset_seqno(const struct intel_timeline *tl);
+
+int intel_timeline_read_hwsp(struct i915_request *from,
+			     struct i915_request *until,
+			     u32 *hwsp_offset);
+
+void intel_gt_init_timelines(struct intel_gt *gt);
+void intel_gt_fini_timelines(struct intel_gt *gt);
+
+void intel_gt_show_timelines(struct intel_gt *gt,
+			     struct drm_printer *m,
+			     void (*show_request)(struct drm_printer *m,
+						  const struct i915_request *rq,
+						  const char *prefix,
+						  int indent));
+
+static inline bool
+intel_timeline_is_last(const struct intel_timeline *tl,
+		       const struct i915_request *rq)
+{
+	return list_is_last_rcu(&rq->link, &tl->requests);
+}
+
+I915_SELFTEST_DECLARE(int intel_timeline_pin_map(struct intel_timeline *tl));
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_timeline_types.h b/drivers/gpu/drm/i915/gt/intel_timeline_types.h
new file mode 100644
index 000000000..74e67dbf8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_timeline_types.h
@@ -0,0 +1,92 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __I915_TIMELINE_TYPES_H__
+#define __I915_TIMELINE_TYPES_H__
+
+#include <linux/list.h>
+#include <linux/kref.h>
+#include <linux/mutex.h>
+#include <linux/rcupdate.h>
+#include <linux/types.h>
+
+#include "i915_active_types.h"
+
+struct i915_vma;
+struct i915_syncmap;
+struct intel_gt;
+
+struct intel_timeline {
+	u64 fence_context;
+	u32 seqno;
+
+	struct mutex mutex; /* protects the flow of requests */
+
+	/*
+	 * pin_count and active_count track essentially the same thing:
+	 * How many requests are in flight or may be under construction.
+	 *
+	 * We need two distinct counters so that we can assign different
+	 * lifetimes to the events for different use-cases. For example,
+	 * we want to permanently keep the timeline pinned for the kernel
+	 * context so that we can issue requests at any time without having
+	 * to acquire space in the GGTT. However, we want to keep tracking
+	 * the activity (to be able to detect when we become idle) along that
+	 * permanently pinned timeline and so end up requiring two counters.
+	 *
+	 * Note that the active_count is protected by the intel_timeline.mutex,
+	 * but the pin_count is protected by a combination of serialisation
+	 * from the intel_context caller plus internal atomicity.
+	 */
+	atomic_t pin_count;
+	atomic_t active_count;
+
+	void *hwsp_map;
+	const u32 *hwsp_seqno;
+	struct i915_vma *hwsp_ggtt;
+	u32 hwsp_offset;
+
+	bool has_initial_breadcrumb;
+
+	/**
+	 * List of breadcrumbs associated with GPU requests currently
+	 * outstanding.
+	 */
+	struct list_head requests;
+
+	/*
+	 * Contains an RCU guarded pointer to the last request. No reference is
+	 * held to the request, users must carefully acquire a reference to
+	 * the request using i915_active_fence_get(), or manage the RCU
+	 * protection themselves (cf the i915_active_fence API).
+	 */
+	struct i915_active_fence last_request;
+
+	struct i915_active active;
+
+	/** A chain of completed timelines ready for early retirement. */
+	struct intel_timeline *retire;
+
+	/**
+	 * We track the most recent seqno that we wait on in every context so
+	 * that we only have to emit a new await and dependency on a more
+	 * recent sync point. As the contexts may be executed out-of-order, we
+	 * have to track each individually and can not rely on an absolute
+	 * global_seqno. When we know that all tracked fences are completed
+	 * (i.e. when the driver is idle), we know that the syncmap is
+	 * redundant and we can discard it without loss of generality.
+	 */
+	struct i915_syncmap *sync;
+
+	struct list_head link;
+	struct intel_gt *gt;
+
+	struct list_head engine_link;
+
+	struct kref kref;
+	struct rcu_head rcu;
+};
+
+#endif /* __I915_TIMELINE_TYPES_H__ */
diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c
new file mode 100644
index 000000000..86621dff8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c
@@ -0,0 +1,3031 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014-2018 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "intel_context.h"
+#include "intel_engine_pm.h"
+#include "intel_engine_regs.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt.h"
+#include "intel_gt_mcr.h"
+#include "intel_gt_regs.h"
+#include "intel_ring.h"
+#include "intel_workarounds.h"
+
+/**
+ * DOC: Hardware workarounds
+ *
+ * This file is intended as a central place to implement most [1]_ of the
+ * required workarounds for hardware to work as originally intended. They fall
+ * in five basic categories depending on how/when they are applied:
+ *
+ * - Workarounds that touch registers that are saved/restored to/from the HW
+ *   context image. The list is emitted (via Load Register Immediate commands)
+ *   everytime a new context is created.
+ * - GT workarounds. The list of these WAs is applied whenever these registers
+ *   revert to default values (on GPU reset, suspend/resume [2]_, etc..).
+ * - Display workarounds. The list is applied during display clock-gating
+ *   initialization.
+ * - Workarounds that whitelist a privileged register, so that UMDs can manage
+ *   them directly. This is just a special case of a MMMIO workaround (as we
+ *   write the list of these to/be-whitelisted registers to some special HW
+ *   registers).
+ * - Workaround batchbuffers, that get executed automatically by the hardware
+ *   on every HW context restore.
+ *
+ * .. [1] Please notice that there are other WAs that, due to their nature,
+ *    cannot be applied from a central place. Those are peppered around the rest
+ *    of the code, as needed.
+ *
+ * .. [2] Technically, some registers are powercontext saved & restored, so they
+ *    survive a suspend/resume. In practice, writing them again is not too
+ *    costly and simplifies things. We can revisit this in the future.
+ *
+ * Layout
+ * ~~~~~~
+ *
+ * Keep things in this file ordered by WA type, as per the above (context, GT,
+ * display, register whitelist, batchbuffer). Then, inside each type, keep the
+ * following order:
+ *
+ * - Infrastructure functions and macros
+ * - WAs per platform in standard gen/chrono order
+ * - Public functions to init or apply the given workaround type.
+ */
+
+static void wa_init_start(struct i915_wa_list *wal, const char *name, const char *engine_name)
+{
+	wal->name = name;
+	wal->engine_name = engine_name;
+}
+
+#define WA_LIST_CHUNK (1 << 4)
+
+static void wa_init_finish(struct i915_wa_list *wal)
+{
+	/* Trim unused entries. */
+	if (!IS_ALIGNED(wal->count, WA_LIST_CHUNK)) {
+		struct i915_wa *list = kmemdup(wal->list,
+					       wal->count * sizeof(*list),
+					       GFP_KERNEL);
+
+		if (list) {
+			kfree(wal->list);
+			wal->list = list;
+		}
+	}
+
+	if (!wal->count)
+		return;
+
+	DRM_DEBUG_DRIVER("Initialized %u %s workarounds on %s\n",
+			 wal->wa_count, wal->name, wal->engine_name);
+}
+
+static void _wa_add(struct i915_wa_list *wal, const struct i915_wa *wa)
+{
+	unsigned int addr = i915_mmio_reg_offset(wa->reg);
+	unsigned int start = 0, end = wal->count;
+	const unsigned int grow = WA_LIST_CHUNK;
+	struct i915_wa *wa_;
+
+	GEM_BUG_ON(!is_power_of_2(grow));
+
+	if (IS_ALIGNED(wal->count, grow)) { /* Either uninitialized or full. */
+		struct i915_wa *list;
+
+		list = kmalloc_array(ALIGN(wal->count + 1, grow), sizeof(*wa),
+				     GFP_KERNEL);
+		if (!list) {
+			DRM_ERROR("No space for workaround init!\n");
+			return;
+		}
+
+		if (wal->list) {
+			memcpy(list, wal->list, sizeof(*wa) * wal->count);
+			kfree(wal->list);
+		}
+
+		wal->list = list;
+	}
+
+	while (start < end) {
+		unsigned int mid = start + (end - start) / 2;
+
+		if (i915_mmio_reg_offset(wal->list[mid].reg) < addr) {
+			start = mid + 1;
+		} else if (i915_mmio_reg_offset(wal->list[mid].reg) > addr) {
+			end = mid;
+		} else {
+			wa_ = &wal->list[mid];
+
+			if ((wa->clr | wa_->clr) && !(wa->clr & ~wa_->clr)) {
+				DRM_ERROR("Discarding overwritten w/a for reg %04x (clear: %08x, set: %08x)\n",
+					  i915_mmio_reg_offset(wa_->reg),
+					  wa_->clr, wa_->set);
+
+				wa_->set &= ~wa->clr;
+			}
+
+			wal->wa_count++;
+			wa_->set |= wa->set;
+			wa_->clr |= wa->clr;
+			wa_->read |= wa->read;
+			return;
+		}
+	}
+
+	wal->wa_count++;
+	wa_ = &wal->list[wal->count++];
+	*wa_ = *wa;
+
+	while (wa_-- > wal->list) {
+		GEM_BUG_ON(i915_mmio_reg_offset(wa_[0].reg) ==
+			   i915_mmio_reg_offset(wa_[1].reg));
+		if (i915_mmio_reg_offset(wa_[1].reg) >
+		    i915_mmio_reg_offset(wa_[0].reg))
+			break;
+
+		swap(wa_[1], wa_[0]);
+	}
+}
+
+static void wa_add(struct i915_wa_list *wal, i915_reg_t reg,
+		   u32 clear, u32 set, u32 read_mask, bool masked_reg)
+{
+	struct i915_wa wa = {
+		.reg  = reg,
+		.clr  = clear,
+		.set  = set,
+		.read = read_mask,
+		.masked_reg = masked_reg,
+	};
+
+	_wa_add(wal, &wa);
+}
+
+static void
+wa_write_clr_set(struct i915_wa_list *wal, i915_reg_t reg, u32 clear, u32 set)
+{
+	wa_add(wal, reg, clear, set, clear, false);
+}
+
+static void
+wa_write(struct i915_wa_list *wal, i915_reg_t reg, u32 set)
+{
+	wa_write_clr_set(wal, reg, ~0, set);
+}
+
+static void
+wa_write_or(struct i915_wa_list *wal, i915_reg_t reg, u32 set)
+{
+	wa_write_clr_set(wal, reg, set, set);
+}
+
+static void
+wa_write_clr(struct i915_wa_list *wal, i915_reg_t reg, u32 clr)
+{
+	wa_write_clr_set(wal, reg, clr, 0);
+}
+
+/*
+ * WA operations on "masked register". A masked register has the upper 16 bits
+ * documented as "masked" in b-spec. Its purpose is to allow writing to just a
+ * portion of the register without a rmw: you simply write in the upper 16 bits
+ * the mask of bits you are going to modify.
+ *
+ * The wa_masked_* family of functions already does the necessary operations to
+ * calculate the mask based on the parameters passed, so user only has to
+ * provide the lower 16 bits of that register.
+ */
+
+static void
+wa_masked_en(struct i915_wa_list *wal, i915_reg_t reg, u32 val)
+{
+	wa_add(wal, reg, 0, _MASKED_BIT_ENABLE(val), val, true);
+}
+
+static void
+wa_masked_dis(struct i915_wa_list *wal, i915_reg_t reg, u32 val)
+{
+	wa_add(wal, reg, 0, _MASKED_BIT_DISABLE(val), val, true);
+}
+
+static void
+wa_masked_field_set(struct i915_wa_list *wal, i915_reg_t reg,
+		    u32 mask, u32 val)
+{
+	wa_add(wal, reg, 0, _MASKED_FIELD(mask, val), mask, true);
+}
+
+static void gen6_ctx_workarounds_init(struct intel_engine_cs *engine,
+				      struct i915_wa_list *wal)
+{
+	wa_masked_en(wal, INSTPM, INSTPM_FORCE_ORDERING);
+}
+
+static void gen7_ctx_workarounds_init(struct intel_engine_cs *engine,
+				      struct i915_wa_list *wal)
+{
+	wa_masked_en(wal, INSTPM, INSTPM_FORCE_ORDERING);
+}
+
+static void gen8_ctx_workarounds_init(struct intel_engine_cs *engine,
+				      struct i915_wa_list *wal)
+{
+	wa_masked_en(wal, INSTPM, INSTPM_FORCE_ORDERING);
+
+	/* WaDisableAsyncFlipPerfMode:bdw,chv */
+	wa_masked_en(wal, RING_MI_MODE(RENDER_RING_BASE), ASYNC_FLIP_PERF_DISABLE);
+
+	/* WaDisablePartialInstShootdown:bdw,chv */
+	wa_masked_en(wal, GEN8_ROW_CHICKEN,
+		     PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
+
+	/* Use Force Non-Coherent whenever executing a 3D context. This is a
+	 * workaround for a possible hang in the unlikely event a TLB
+	 * invalidation occurs during a PSD flush.
+	 */
+	/* WaForceEnableNonCoherent:bdw,chv */
+	/* WaHdcDisableFetchWhenMasked:bdw,chv */
+	wa_masked_en(wal, HDC_CHICKEN0,
+		     HDC_DONOT_FETCH_MEM_WHEN_MASKED |
+		     HDC_FORCE_NON_COHERENT);
+
+	/* From the Haswell PRM, Command Reference: Registers, CACHE_MODE_0:
+	 * "The Hierarchical Z RAW Stall Optimization allows non-overlapping
+	 *  polygons in the same 8x4 pixel/sample area to be processed without
+	 *  stalling waiting for the earlier ones to write to Hierarchical Z
+	 *  buffer."
+	 *
+	 * This optimization is off by default for BDW and CHV; turn it on.
+	 */
+	wa_masked_dis(wal, CACHE_MODE_0_GEN7, HIZ_RAW_STALL_OPT_DISABLE);
+
+	/* Wa4x4STCOptimizationDisable:bdw,chv */
+	wa_masked_en(wal, CACHE_MODE_1, GEN8_4x4_STC_OPTIMIZATION_DISABLE);
+
+	/*
+	 * BSpec recommends 8x4 when MSAA is used,
+	 * however in practice 16x4 seems fastest.
+	 *
+	 * Note that PS/WM thread counts depend on the WIZ hashing
+	 * disable bit, which we don't touch here, but it's good
+	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
+	 */
+	wa_masked_field_set(wal, GEN7_GT_MODE,
+			    GEN6_WIZ_HASHING_MASK,
+			    GEN6_WIZ_HASHING_16x4);
+}
+
+static void bdw_ctx_workarounds_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	gen8_ctx_workarounds_init(engine, wal);
+
+	/* WaDisableThreadStallDopClockGating:bdw (pre-production) */
+	wa_masked_en(wal, GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE);
+
+	/* WaDisableDopClockGating:bdw
+	 *
+	 * Also see the related UCGTCL1 write in bdw_init_clock_gating()
+	 * to disable EUTC clock gating.
+	 */
+	wa_masked_en(wal, GEN7_ROW_CHICKEN2,
+		     DOP_CLOCK_GATING_DISABLE);
+
+	wa_masked_en(wal, HALF_SLICE_CHICKEN3,
+		     GEN8_SAMPLER_POWER_BYPASS_DIS);
+
+	wa_masked_en(wal, HDC_CHICKEN0,
+		     /* WaForceContextSaveRestoreNonCoherent:bdw */
+		     HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
+		     /* WaDisableFenceDestinationToSLM:bdw (pre-prod) */
+		     (IS_BDW_GT3(i915) ? HDC_FENCE_DEST_SLM_DISABLE : 0));
+}
+
+static void chv_ctx_workarounds_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	gen8_ctx_workarounds_init(engine, wal);
+
+	/* WaDisableThreadStallDopClockGating:chv */
+	wa_masked_en(wal, GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE);
+
+	/* Improve HiZ throughput on CHV. */
+	wa_masked_en(wal, HIZ_CHICKEN, CHV_HZ_8X8_MODE_IN_1X);
+}
+
+static void gen9_ctx_workarounds_init(struct intel_engine_cs *engine,
+				      struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	if (HAS_LLC(i915)) {
+		/* WaCompressedResourceSamplerPbeMediaNewHashMode:skl,kbl
+		 *
+		 * Must match Display Engine. See
+		 * WaCompressedResourceDisplayNewHashMode.
+		 */
+		wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+			     GEN9_PBE_COMPRESSED_HASH_SELECTION);
+		wa_masked_en(wal, GEN9_HALF_SLICE_CHICKEN7,
+			     GEN9_SAMPLER_HASH_COMPRESSED_READ_ADDR);
+	}
+
+	/* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl,glk,cfl */
+	/* WaDisablePartialInstShootdown:skl,bxt,kbl,glk,cfl */
+	wa_masked_en(wal, GEN8_ROW_CHICKEN,
+		     FLOW_CONTROL_ENABLE |
+		     PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
+
+	/* WaEnableYV12BugFixInHalfSliceChicken7:skl,bxt,kbl,glk,cfl */
+	/* WaEnableSamplerGPGPUPreemptionSupport:skl,bxt,kbl,cfl */
+	wa_masked_en(wal, GEN9_HALF_SLICE_CHICKEN7,
+		     GEN9_ENABLE_YV12_BUGFIX |
+		     GEN9_ENABLE_GPGPU_PREEMPTION);
+
+	/* Wa4x4STCOptimizationDisable:skl,bxt,kbl,glk,cfl */
+	/* WaDisablePartialResolveInVc:skl,bxt,kbl,cfl */
+	wa_masked_en(wal, CACHE_MODE_1,
+		     GEN8_4x4_STC_OPTIMIZATION_DISABLE |
+		     GEN9_PARTIAL_RESOLVE_IN_VC_DISABLE);
+
+	/* WaCcsTlbPrefetchDisable:skl,bxt,kbl,glk,cfl */
+	wa_masked_dis(wal, GEN9_HALF_SLICE_CHICKEN5,
+		      GEN9_CCS_TLB_PREFETCH_ENABLE);
+
+	/* WaForceContextSaveRestoreNonCoherent:skl,bxt,kbl,cfl */
+	wa_masked_en(wal, HDC_CHICKEN0,
+		     HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
+		     HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE);
+
+	/* WaForceEnableNonCoherent and WaDisableHDCInvalidation are
+	 * both tied to WaForceContextSaveRestoreNonCoherent
+	 * in some hsds for skl. We keep the tie for all gen9. The
+	 * documentation is a bit hazy and so we want to get common behaviour,
+	 * even though there is no clear evidence we would need both on kbl/bxt.
+	 * This area has been source of system hangs so we play it safe
+	 * and mimic the skl regardless of what bspec says.
+	 *
+	 * Use Force Non-Coherent whenever executing a 3D context. This
+	 * is a workaround for a possible hang in the unlikely event
+	 * a TLB invalidation occurs during a PSD flush.
+	 */
+
+	/* WaForceEnableNonCoherent:skl,bxt,kbl,cfl */
+	wa_masked_en(wal, HDC_CHICKEN0,
+		     HDC_FORCE_NON_COHERENT);
+
+	/* WaDisableSamplerPowerBypassForSOPingPong:skl,bxt,kbl,cfl */
+	if (IS_SKYLAKE(i915) ||
+	    IS_KABYLAKE(i915) ||
+	    IS_COFFEELAKE(i915) ||
+	    IS_COMETLAKE(i915))
+		wa_masked_en(wal, HALF_SLICE_CHICKEN3,
+			     GEN8_SAMPLER_POWER_BYPASS_DIS);
+
+	/* WaDisableSTUnitPowerOptimization:skl,bxt,kbl,glk,cfl */
+	wa_masked_en(wal, HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE);
+
+	/*
+	 * Supporting preemption with fine-granularity requires changes in the
+	 * batch buffer programming. Since we can't break old userspace, we
+	 * need to set our default preemption level to safe value. Userspace is
+	 * still able to use more fine-grained preemption levels, since in
+	 * WaEnablePreemptionGranularityControlByUMD we're whitelisting the
+	 * per-ctx register. As such, WaDisable{3D,GPGPU}MidCmdPreemption are
+	 * not real HW workarounds, but merely a way to start using preemption
+	 * while maintaining old contract with userspace.
+	 */
+
+	/* WaDisable3DMidCmdPreemption:skl,bxt,glk,cfl,[cnl] */
+	wa_masked_dis(wal, GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL);
+
+	/* WaDisableGPGPUMidCmdPreemption:skl,bxt,blk,cfl,[cnl] */
+	wa_masked_field_set(wal, GEN8_CS_CHICKEN1,
+			    GEN9_PREEMPT_GPGPU_LEVEL_MASK,
+			    GEN9_PREEMPT_GPGPU_COMMAND_LEVEL);
+
+	/* WaClearHIZ_WM_CHICKEN3:bxt,glk */
+	if (IS_GEN9_LP(i915))
+		wa_masked_en(wal, GEN9_WM_CHICKEN3, GEN9_FACTOR_IN_CLR_VAL_HIZ);
+}
+
+static void skl_tune_iz_hashing(struct intel_engine_cs *engine,
+				struct i915_wa_list *wal)
+{
+	struct intel_gt *gt = engine->gt;
+	u8 vals[3] = { 0, 0, 0 };
+	unsigned int i;
+
+	for (i = 0; i < 3; i++) {
+		u8 ss;
+
+		/*
+		 * Only consider slices where one, and only one, subslice has 7
+		 * EUs
+		 */
+		if (!is_power_of_2(gt->info.sseu.subslice_7eu[i]))
+			continue;
+
+		/*
+		 * subslice_7eu[i] != 0 (because of the check above) and
+		 * ss_max == 4 (maximum number of subslices possible per slice)
+		 *
+		 * ->    0 <= ss <= 3;
+		 */
+		ss = ffs(gt->info.sseu.subslice_7eu[i]) - 1;
+		vals[i] = 3 - ss;
+	}
+
+	if (vals[0] == 0 && vals[1] == 0 && vals[2] == 0)
+		return;
+
+	/* Tune IZ hashing. See intel_device_info_runtime_init() */
+	wa_masked_field_set(wal, GEN7_GT_MODE,
+			    GEN9_IZ_HASHING_MASK(2) |
+			    GEN9_IZ_HASHING_MASK(1) |
+			    GEN9_IZ_HASHING_MASK(0),
+			    GEN9_IZ_HASHING(2, vals[2]) |
+			    GEN9_IZ_HASHING(1, vals[1]) |
+			    GEN9_IZ_HASHING(0, vals[0]));
+}
+
+static void skl_ctx_workarounds_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	gen9_ctx_workarounds_init(engine, wal);
+	skl_tune_iz_hashing(engine, wal);
+}
+
+static void bxt_ctx_workarounds_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	gen9_ctx_workarounds_init(engine, wal);
+
+	/* WaDisableThreadStallDopClockGating:bxt */
+	wa_masked_en(wal, GEN8_ROW_CHICKEN,
+		     STALL_DOP_GATING_DISABLE);
+
+	/* WaToEnableHwFixForPushConstHWBug:bxt */
+	wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+		     GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+}
+
+static void kbl_ctx_workarounds_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	gen9_ctx_workarounds_init(engine, wal);
+
+	/* WaToEnableHwFixForPushConstHWBug:kbl */
+	if (IS_KBL_GRAPHICS_STEP(i915, STEP_C0, STEP_FOREVER))
+		wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+			     GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+
+	/* WaDisableSbeCacheDispatchPortSharing:kbl */
+	wa_masked_en(wal, GEN7_HALF_SLICE_CHICKEN1,
+		     GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
+}
+
+static void glk_ctx_workarounds_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	gen9_ctx_workarounds_init(engine, wal);
+
+	/* WaToEnableHwFixForPushConstHWBug:glk */
+	wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+		     GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+}
+
+static void cfl_ctx_workarounds_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	gen9_ctx_workarounds_init(engine, wal);
+
+	/* WaToEnableHwFixForPushConstHWBug:cfl */
+	wa_masked_en(wal, COMMON_SLICE_CHICKEN2,
+		     GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+
+	/* WaDisableSbeCacheDispatchPortSharing:cfl */
+	wa_masked_en(wal, GEN7_HALF_SLICE_CHICKEN1,
+		     GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
+}
+
+static void icl_ctx_workarounds_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	/* Wa_1406697149 (WaDisableBankHangMode:icl) */
+	wa_write(wal,
+		 GEN8_L3CNTLREG,
+		 intel_uncore_read(engine->uncore, GEN8_L3CNTLREG) |
+		 GEN8_ERRDETBCTRL);
+
+	/* WaForceEnableNonCoherent:icl
+	 * This is not the same workaround as in early Gen9 platforms, where
+	 * lacking this could cause system hangs, but coherency performance
+	 * overhead is high and only a few compute workloads really need it
+	 * (the register is whitelisted in hardware now, so UMDs can opt in
+	 * for coherency if they have a good reason).
+	 */
+	wa_masked_en(wal, ICL_HDC_MODE, HDC_FORCE_NON_COHERENT);
+
+	/* WaEnableFloatBlendOptimization:icl */
+	wa_add(wal, GEN10_CACHE_MODE_SS, 0,
+	       _MASKED_BIT_ENABLE(FLOAT_BLEND_OPTIMIZATION_ENABLE),
+	       0 /* write-only, so skip validation */,
+	       true);
+
+	/* WaDisableGPGPUMidThreadPreemption:icl */
+	wa_masked_field_set(wal, GEN8_CS_CHICKEN1,
+			    GEN9_PREEMPT_GPGPU_LEVEL_MASK,
+			    GEN9_PREEMPT_GPGPU_THREAD_GROUP_LEVEL);
+
+	/* allow headerless messages for preemptible GPGPU context */
+	wa_masked_en(wal, GEN10_SAMPLER_MODE,
+		     GEN11_SAMPLER_ENABLE_HEADLESS_MSG);
+
+	/* Wa_1604278689:icl,ehl */
+	wa_write(wal, IVB_FBC_RT_BASE, 0xFFFFFFFF & ~ILK_FBC_RT_VALID);
+	wa_write_clr_set(wal, IVB_FBC_RT_BASE_UPPER,
+			 0, /* write-only register; skip validation */
+			 0xFFFFFFFF);
+
+	/* Wa_1406306137:icl,ehl */
+	wa_masked_en(wal, GEN9_ROW_CHICKEN4, GEN11_DIS_PICK_2ND_EU);
+}
+
+/*
+ * These settings aren't actually workarounds, but general tuning settings that
+ * need to be programmed on dg2 platform.
+ */
+static void dg2_ctx_gt_tuning_init(struct intel_engine_cs *engine,
+				   struct i915_wa_list *wal)
+{
+	wa_masked_en(wal, CHICKEN_RASTER_2, TBIMR_FAST_CLIP);
+	wa_write_clr_set(wal, GEN11_L3SQCREG5, L3_PWM_TIMER_INIT_VAL_MASK,
+			 REG_FIELD_PREP(L3_PWM_TIMER_INIT_VAL_MASK, 0x7f));
+	wa_add(wal,
+	       FF_MODE2,
+	       FF_MODE2_TDS_TIMER_MASK,
+	       FF_MODE2_TDS_TIMER_128,
+	       0, false);
+}
+
+/*
+ * These settings aren't actually workarounds, but general tuning settings that
+ * need to be programmed on several platforms.
+ */
+static void gen12_ctx_gt_tuning_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	/*
+	 * Although some platforms refer to it as Wa_1604555607, we need to
+	 * program it even on those that don't explicitly list that
+	 * workaround.
+	 *
+	 * Note that the programming of this register is further modified
+	 * according to the FF_MODE2 guidance given by Wa_1608008084:gen12.
+	 * Wa_1608008084 tells us the FF_MODE2 register will return the wrong
+	 * value when read. The default value for this register is zero for all
+	 * fields and there are no bit masks. So instead of doing a RMW we
+	 * should just write TDS timer value. For the same reason read
+	 * verification is ignored.
+	 */
+	wa_add(wal,
+	       FF_MODE2,
+	       FF_MODE2_TDS_TIMER_MASK,
+	       FF_MODE2_TDS_TIMER_128,
+	       0, false);
+}
+
+static void gen12_ctx_workarounds_init(struct intel_engine_cs *engine,
+				       struct i915_wa_list *wal)
+{
+	gen12_ctx_gt_tuning_init(engine, wal);
+
+	/*
+	 * Wa_1409142259:tgl,dg1,adl-p
+	 * Wa_1409347922:tgl,dg1,adl-p
+	 * Wa_1409252684:tgl,dg1,adl-p
+	 * Wa_1409217633:tgl,dg1,adl-p
+	 * Wa_1409207793:tgl,dg1,adl-p
+	 * Wa_1409178076:tgl,dg1,adl-p
+	 * Wa_1408979724:tgl,dg1,adl-p
+	 * Wa_14010443199:tgl,rkl,dg1,adl-p
+	 * Wa_14010698770:tgl,rkl,dg1,adl-s,adl-p
+	 * Wa_1409342910:tgl,rkl,dg1,adl-s,adl-p
+	 */
+	wa_masked_en(wal, GEN11_COMMON_SLICE_CHICKEN3,
+		     GEN12_DISABLE_CPS_AWARE_COLOR_PIPE);
+
+	/* WaDisableGPGPUMidThreadPreemption:gen12 */
+	wa_masked_field_set(wal, GEN8_CS_CHICKEN1,
+			    GEN9_PREEMPT_GPGPU_LEVEL_MASK,
+			    GEN9_PREEMPT_GPGPU_THREAD_GROUP_LEVEL);
+
+	/*
+	 * Wa_16011163337
+	 *
+	 * Like in gen12_ctx_gt_tuning_init(), read verification is ignored due
+	 * to Wa_1608008084.
+	 */
+	wa_add(wal,
+	       FF_MODE2,
+	       FF_MODE2_GS_TIMER_MASK,
+	       FF_MODE2_GS_TIMER_224,
+	       0, false);
+}
+
+static void dg1_ctx_workarounds_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	gen12_ctx_workarounds_init(engine, wal);
+
+	/* Wa_1409044764 */
+	wa_masked_dis(wal, GEN11_COMMON_SLICE_CHICKEN3,
+		      DG1_FLOAT_POINT_BLEND_OPT_STRICT_MODE_EN);
+
+	/* Wa_22010493298 */
+	wa_masked_en(wal, HIZ_CHICKEN,
+		     DG1_HZ_READ_SUPPRESSION_OPTIMIZATION_DISABLE);
+}
+
+static void dg2_ctx_workarounds_init(struct intel_engine_cs *engine,
+				     struct i915_wa_list *wal)
+{
+	dg2_ctx_gt_tuning_init(engine, wal);
+
+	/* Wa_16011186671:dg2_g11 */
+	if (IS_DG2_GRAPHICS_STEP(engine->i915, G11, STEP_A0, STEP_B0)) {
+		wa_masked_dis(wal, VFLSKPD, DIS_MULT_MISS_RD_SQUASH);
+		wa_masked_en(wal, VFLSKPD, DIS_OVER_FETCH_CACHE);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(engine->i915, G10, STEP_A0, STEP_B0)) {
+		/* Wa_14010469329:dg2_g10 */
+		wa_masked_en(wal, GEN11_COMMON_SLICE_CHICKEN3,
+			     XEHP_DUAL_SIMD8_SEQ_MERGE_DISABLE);
+
+		/*
+		 * Wa_22010465075:dg2_g10
+		 * Wa_22010613112:dg2_g10
+		 * Wa_14010698770:dg2_g10
+		 */
+		wa_masked_en(wal, GEN11_COMMON_SLICE_CHICKEN3,
+			     GEN12_DISABLE_CPS_AWARE_COLOR_PIPE);
+	}
+
+	/* Wa_16013271637:dg2 */
+	wa_masked_en(wal, SLICE_COMMON_ECO_CHICKEN1,
+		     MSC_MSAA_REODER_BUF_BYPASS_DISABLE);
+
+	/* Wa_14014947963:dg2 */
+	if (IS_DG2_GRAPHICS_STEP(engine->i915, G10, STEP_B0, STEP_FOREVER) ||
+		IS_DG2_G11(engine->i915) || IS_DG2_G12(engine->i915))
+		wa_masked_field_set(wal, VF_PREEMPTION, PREEMPTION_VERTEX_COUNT, 0x4000);
+
+	/* Wa_15010599737:dg2 */
+	wa_masked_en(wal, CHICKEN_RASTER_1, DIS_SF_ROUND_NEAREST_EVEN);
+}
+
+static void fakewa_disable_nestedbb_mode(struct intel_engine_cs *engine,
+					 struct i915_wa_list *wal)
+{
+	/*
+	 * This is a "fake" workaround defined by software to ensure we
+	 * maintain reliable, backward-compatible behavior for userspace with
+	 * regards to how nested MI_BATCH_BUFFER_START commands are handled.
+	 *
+	 * The per-context setting of MI_MODE[12] determines whether the bits
+	 * of a nested MI_BATCH_BUFFER_START instruction should be interpreted
+	 * in the traditional manner or whether they should instead use a new
+	 * tgl+ meaning that breaks backward compatibility, but allows nesting
+	 * into 3rd-level batchbuffers.  When this new capability was first
+	 * added in TGL, it remained off by default unless a context
+	 * intentionally opted in to the new behavior.  However Xe_HPG now
+	 * flips this on by default and requires that we explicitly opt out if
+	 * we don't want the new behavior.
+	 *
+	 * From a SW perspective, we want to maintain the backward-compatible
+	 * behavior for userspace, so we'll apply a fake workaround to set it
+	 * back to the legacy behavior on platforms where the hardware default
+	 * is to break compatibility.  At the moment there is no Linux
+	 * userspace that utilizes third-level batchbuffers, so this will avoid
+	 * userspace from needing to make any changes.  using the legacy
+	 * meaning is the correct thing to do.  If/when we have userspace
+	 * consumers that want to utilize third-level batch nesting, we can
+	 * provide a context parameter to allow them to opt-in.
+	 */
+	wa_masked_dis(wal, RING_MI_MODE(engine->mmio_base), TGL_NESTED_BB_EN);
+}
+
+static void gen12_ctx_gt_mocs_init(struct intel_engine_cs *engine,
+				   struct i915_wa_list *wal)
+{
+	u8 mocs;
+
+	/*
+	 * Some blitter commands do not have a field for MOCS, those
+	 * commands will use MOCS index pointed by BLIT_CCTL.
+	 * BLIT_CCTL registers are needed to be programmed to un-cached.
+	 */
+	if (engine->class == COPY_ENGINE_CLASS) {
+		mocs = engine->gt->mocs.uc_index;
+		wa_write_clr_set(wal,
+				 BLIT_CCTL(engine->mmio_base),
+				 BLIT_CCTL_MASK,
+				 BLIT_CCTL_MOCS(mocs, mocs));
+	}
+}
+
+/*
+ * gen12_ctx_gt_fake_wa_init() aren't programmingan official workaround
+ * defined by the hardware team, but it programming general context registers.
+ * Adding those context register programming in context workaround
+ * allow us to use the wa framework for proper application and validation.
+ */
+static void
+gen12_ctx_gt_fake_wa_init(struct intel_engine_cs *engine,
+			  struct i915_wa_list *wal)
+{
+	if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 55))
+		fakewa_disable_nestedbb_mode(engine, wal);
+
+	gen12_ctx_gt_mocs_init(engine, wal);
+}
+
+static void
+__intel_engine_init_ctx_wa(struct intel_engine_cs *engine,
+			   struct i915_wa_list *wal,
+			   const char *name)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	wa_init_start(wal, name, engine->name);
+
+	/* Applies to all engines */
+	/*
+	 * Fake workarounds are not the actual workaround but
+	 * programming of context registers using workaround framework.
+	 */
+	if (GRAPHICS_VER(i915) >= 12)
+		gen12_ctx_gt_fake_wa_init(engine, wal);
+
+	if (engine->class != RENDER_CLASS)
+		goto done;
+
+	if (IS_PONTEVECCHIO(i915))
+		; /* noop; none at this time */
+	else if (IS_DG2(i915))
+		dg2_ctx_workarounds_init(engine, wal);
+	else if (IS_XEHPSDV(i915))
+		; /* noop; none at this time */
+	else if (IS_DG1(i915))
+		dg1_ctx_workarounds_init(engine, wal);
+	else if (GRAPHICS_VER(i915) == 12)
+		gen12_ctx_workarounds_init(engine, wal);
+	else if (GRAPHICS_VER(i915) == 11)
+		icl_ctx_workarounds_init(engine, wal);
+	else if (IS_COFFEELAKE(i915) || IS_COMETLAKE(i915))
+		cfl_ctx_workarounds_init(engine, wal);
+	else if (IS_GEMINILAKE(i915))
+		glk_ctx_workarounds_init(engine, wal);
+	else if (IS_KABYLAKE(i915))
+		kbl_ctx_workarounds_init(engine, wal);
+	else if (IS_BROXTON(i915))
+		bxt_ctx_workarounds_init(engine, wal);
+	else if (IS_SKYLAKE(i915))
+		skl_ctx_workarounds_init(engine, wal);
+	else if (IS_CHERRYVIEW(i915))
+		chv_ctx_workarounds_init(engine, wal);
+	else if (IS_BROADWELL(i915))
+		bdw_ctx_workarounds_init(engine, wal);
+	else if (GRAPHICS_VER(i915) == 7)
+		gen7_ctx_workarounds_init(engine, wal);
+	else if (GRAPHICS_VER(i915) == 6)
+		gen6_ctx_workarounds_init(engine, wal);
+	else if (GRAPHICS_VER(i915) < 8)
+		;
+	else
+		MISSING_CASE(GRAPHICS_VER(i915));
+
+done:
+	wa_init_finish(wal);
+}
+
+void intel_engine_init_ctx_wa(struct intel_engine_cs *engine)
+{
+	__intel_engine_init_ctx_wa(engine, &engine->ctx_wa_list, "context");
+}
+
+int intel_engine_emit_ctx_wa(struct i915_request *rq)
+{
+	struct i915_wa_list *wal = &rq->engine->ctx_wa_list;
+	struct i915_wa *wa;
+	unsigned int i;
+	u32 *cs;
+	int ret;
+
+	if (wal->count == 0)
+		return 0;
+
+	ret = rq->engine->emit_flush(rq, EMIT_BARRIER);
+	if (ret)
+		return ret;
+
+	cs = intel_ring_begin(rq, (wal->count * 2 + 2));
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(wal->count);
+	for (i = 0, wa = wal->list; i < wal->count; i++, wa++) {
+		*cs++ = i915_mmio_reg_offset(wa->reg);
+		*cs++ = wa->set;
+	}
+	*cs++ = MI_NOOP;
+
+	intel_ring_advance(rq, cs);
+
+	ret = rq->engine->emit_flush(rq, EMIT_BARRIER);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static void
+gen4_gt_workarounds_init(struct intel_gt *gt,
+			 struct i915_wa_list *wal)
+{
+	/* WaDisable_RenderCache_OperationalFlush:gen4,ilk */
+	wa_masked_dis(wal, CACHE_MODE_0, RC_OP_FLUSH_ENABLE);
+}
+
+static void
+g4x_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	gen4_gt_workarounds_init(gt, wal);
+
+	/* WaDisableRenderCachePipelinedFlush:g4x,ilk */
+	wa_masked_en(wal, CACHE_MODE_0, CM0_PIPELINED_RENDER_FLUSH_DISABLE);
+}
+
+static void
+ilk_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	g4x_gt_workarounds_init(gt, wal);
+
+	wa_masked_en(wal, _3D_CHICKEN2, _3D_CHICKEN2_WM_READ_PIPELINED);
+}
+
+static void
+snb_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+}
+
+static void
+ivb_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	/* Apply the WaDisableRHWOOptimizationForRenderHang:ivb workaround. */
+	wa_masked_dis(wal,
+		      GEN7_COMMON_SLICE_CHICKEN1,
+		      GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC);
+
+	/* WaApplyL3ControlAndL3ChickenMode:ivb */
+	wa_write(wal, GEN7_L3CNTLREG1, GEN7_WA_FOR_GEN7_L3_CONTROL);
+	wa_write(wal, GEN7_L3_CHICKEN_MODE_REGISTER, GEN7_WA_L3_CHICKEN_MODE);
+
+	/* WaForceL3Serialization:ivb */
+	wa_write_clr(wal, GEN7_L3SQCREG4, L3SQ_URB_READ_CAM_MATCH_DISABLE);
+}
+
+static void
+vlv_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	/* WaForceL3Serialization:vlv */
+	wa_write_clr(wal, GEN7_L3SQCREG4, L3SQ_URB_READ_CAM_MATCH_DISABLE);
+
+	/*
+	 * WaIncreaseL3CreditsForVLVB0:vlv
+	 * This is the hardware default actually.
+	 */
+	wa_write(wal, GEN7_L3SQCREG1, VLV_B0_WA_L3SQCREG1_VALUE);
+}
+
+static void
+hsw_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	/* L3 caching of data atomics doesn't work -- disable it. */
+	wa_write(wal, HSW_SCRATCH1, HSW_SCRATCH1_L3_DATA_ATOMICS_DISABLE);
+
+	wa_add(wal,
+	       HSW_ROW_CHICKEN3, 0,
+	       _MASKED_BIT_ENABLE(HSW_ROW_CHICKEN3_L3_GLOBAL_ATOMICS_DISABLE),
+	       0 /* XXX does this reg exist? */, true);
+
+	/* WaVSRefCountFullforceMissDisable:hsw */
+	wa_write_clr(wal, GEN7_FF_THREAD_MODE, GEN7_FF_VS_REF_CNT_FFME);
+}
+
+static void
+gen9_wa_init_mcr(struct drm_i915_private *i915, struct i915_wa_list *wal)
+{
+	const struct sseu_dev_info *sseu = &to_gt(i915)->info.sseu;
+	unsigned int slice, subslice;
+	u32 mcr, mcr_mask;
+
+	GEM_BUG_ON(GRAPHICS_VER(i915) != 9);
+
+	/*
+	 * WaProgramMgsrForCorrectSliceSpecificMmioReads:gen9,glk,kbl,cml
+	 * Before any MMIO read into slice/subslice specific registers, MCR
+	 * packet control register needs to be programmed to point to any
+	 * enabled s/ss pair. Otherwise, incorrect values will be returned.
+	 * This means each subsequent MMIO read will be forwarded to an
+	 * specific s/ss combination, but this is OK since these registers
+	 * are consistent across s/ss in almost all cases. In the rare
+	 * occasions, such as INSTDONE, where this value is dependent
+	 * on s/ss combo, the read should be done with read_subslice_reg.
+	 */
+	slice = ffs(sseu->slice_mask) - 1;
+	GEM_BUG_ON(slice >= ARRAY_SIZE(sseu->subslice_mask.hsw));
+	subslice = ffs(intel_sseu_get_hsw_subslices(sseu, slice));
+	GEM_BUG_ON(!subslice);
+	subslice--;
+
+	/*
+	 * We use GEN8_MCR..() macros to calculate the |mcr| value for
+	 * Gen9 to address WaProgramMgsrForCorrectSliceSpecificMmioReads
+	 */
+	mcr = GEN8_MCR_SLICE(slice) | GEN8_MCR_SUBSLICE(subslice);
+	mcr_mask = GEN8_MCR_SLICE_MASK | GEN8_MCR_SUBSLICE_MASK;
+
+	drm_dbg(&i915->drm, "MCR slice:%d/subslice:%d = %x\n", slice, subslice, mcr);
+
+	wa_write_clr_set(wal, GEN8_MCR_SELECTOR, mcr_mask, mcr);
+}
+
+static void
+gen9_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	/* WaProgramMgsrForCorrectSliceSpecificMmioReads:glk,kbl,cml,gen9 */
+	gen9_wa_init_mcr(i915, wal);
+
+	/* WaDisableKillLogic:bxt,skl,kbl */
+	if (!IS_COFFEELAKE(i915) && !IS_COMETLAKE(i915))
+		wa_write_or(wal,
+			    GAM_ECOCHK,
+			    ECOCHK_DIS_TLB);
+
+	if (HAS_LLC(i915)) {
+		/* WaCompressedResourceSamplerPbeMediaNewHashMode:skl,kbl
+		 *
+		 * Must match Display Engine. See
+		 * WaCompressedResourceDisplayNewHashMode.
+		 */
+		wa_write_or(wal,
+			    MMCD_MISC_CTRL,
+			    MMCD_PCLA | MMCD_HOTSPOT_EN);
+	}
+
+	/* WaDisableHDCInvalidation:skl,bxt,kbl,cfl */
+	wa_write_or(wal,
+		    GAM_ECOCHK,
+		    BDW_DISABLE_HDC_INVALIDATION);
+}
+
+static void
+skl_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	gen9_gt_workarounds_init(gt, wal);
+
+	/* WaDisableGafsUnitClkGating:skl */
+	wa_write_or(wal,
+		    GEN7_UCGCTL4,
+		    GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE);
+
+	/* WaInPlaceDecompressionHang:skl */
+	if (IS_SKL_GRAPHICS_STEP(gt->i915, STEP_A0, STEP_H0))
+		wa_write_or(wal,
+			    GEN9_GAMT_ECO_REG_RW_IA,
+			    GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
+}
+
+static void
+kbl_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	gen9_gt_workarounds_init(gt, wal);
+
+	/* WaDisableDynamicCreditSharing:kbl */
+	if (IS_KBL_GRAPHICS_STEP(gt->i915, 0, STEP_C0))
+		wa_write_or(wal,
+			    GAMT_CHKN_BIT_REG,
+			    GAMT_CHKN_DISABLE_DYNAMIC_CREDIT_SHARING);
+
+	/* WaDisableGafsUnitClkGating:kbl */
+	wa_write_or(wal,
+		    GEN7_UCGCTL4,
+		    GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE);
+
+	/* WaInPlaceDecompressionHang:kbl */
+	wa_write_or(wal,
+		    GEN9_GAMT_ECO_REG_RW_IA,
+		    GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
+}
+
+static void
+glk_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	gen9_gt_workarounds_init(gt, wal);
+}
+
+static void
+cfl_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	gen9_gt_workarounds_init(gt, wal);
+
+	/* WaDisableGafsUnitClkGating:cfl */
+	wa_write_or(wal,
+		    GEN7_UCGCTL4,
+		    GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE);
+
+	/* WaInPlaceDecompressionHang:cfl */
+	wa_write_or(wal,
+		    GEN9_GAMT_ECO_REG_RW_IA,
+		    GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
+}
+
+static void __set_mcr_steering(struct i915_wa_list *wal,
+			       i915_reg_t steering_reg,
+			       unsigned int slice, unsigned int subslice)
+{
+	u32 mcr, mcr_mask;
+
+	mcr = GEN11_MCR_SLICE(slice) | GEN11_MCR_SUBSLICE(subslice);
+	mcr_mask = GEN11_MCR_SLICE_MASK | GEN11_MCR_SUBSLICE_MASK;
+
+	wa_write_clr_set(wal, steering_reg, mcr_mask, mcr);
+}
+
+static void __add_mcr_wa(struct intel_gt *gt, struct i915_wa_list *wal,
+			 unsigned int slice, unsigned int subslice)
+{
+	struct drm_printer p = drm_debug_printer("MCR Steering:");
+
+	__set_mcr_steering(wal, GEN8_MCR_SELECTOR, slice, subslice);
+
+	gt->default_steering.groupid = slice;
+	gt->default_steering.instanceid = subslice;
+
+	if (drm_debug_enabled(DRM_UT_DRIVER))
+		intel_gt_mcr_report_steering(&p, gt, false);
+}
+
+static void
+icl_wa_init_mcr(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	const struct sseu_dev_info *sseu = &gt->info.sseu;
+	unsigned int subslice;
+
+	GEM_BUG_ON(GRAPHICS_VER(gt->i915) < 11);
+	GEM_BUG_ON(hweight8(sseu->slice_mask) > 1);
+
+	/*
+	 * Although a platform may have subslices, we need to always steer
+	 * reads to the lowest instance that isn't fused off.  When Render
+	 * Power Gating is enabled, grabbing forcewake will only power up a
+	 * single subslice (the "minconfig") if there isn't a real workload
+	 * that needs to be run; this means that if we steer register reads to
+	 * one of the higher subslices, we run the risk of reading back 0's or
+	 * random garbage.
+	 */
+	subslice = __ffs(intel_sseu_get_hsw_subslices(sseu, 0));
+
+	/*
+	 * If the subslice we picked above also steers us to a valid L3 bank,
+	 * then we can just rely on the default steering and won't need to
+	 * worry about explicitly re-steering L3BANK reads later.
+	 */
+	if (gt->info.l3bank_mask & BIT(subslice))
+		gt->steering_table[L3BANK] = NULL;
+
+	__add_mcr_wa(gt, wal, 0, subslice);
+}
+
+static void
+xehp_init_mcr(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	const struct sseu_dev_info *sseu = &gt->info.sseu;
+	unsigned long slice, subslice = 0, slice_mask = 0;
+	u32 lncf_mask = 0;
+	int i;
+
+	/*
+	 * On Xe_HP the steering increases in complexity. There are now several
+	 * more units that require steering and we're not guaranteed to be able
+	 * to find a common setting for all of them. These are:
+	 * - GSLICE (fusable)
+	 * - DSS (sub-unit within gslice; fusable)
+	 * - L3 Bank (fusable)
+	 * - MSLICE (fusable)
+	 * - LNCF (sub-unit within mslice; always present if mslice is present)
+	 *
+	 * We'll do our default/implicit steering based on GSLICE (in the
+	 * sliceid field) and DSS (in the subsliceid field).  If we can
+	 * find overlap between the valid MSLICE and/or LNCF values with
+	 * a suitable GSLICE, then we can just re-use the default value and
+	 * skip and explicit steering at runtime.
+	 *
+	 * We only need to look for overlap between GSLICE/MSLICE/LNCF to find
+	 * a valid sliceid value.  DSS steering is the only type of steering
+	 * that utilizes the 'subsliceid' bits.
+	 *
+	 * Also note that, even though the steering domain is called "GSlice"
+	 * and it is encoded in the register using the gslice format, the spec
+	 * says that the combined (geometry | compute) fuse should be used to
+	 * select the steering.
+	 */
+
+	/* Find the potential gslice candidates */
+	slice_mask = intel_slicemask_from_xehp_dssmask(sseu->subslice_mask,
+						       GEN_DSS_PER_GSLICE);
+
+	/*
+	 * Find the potential LNCF candidates.  Either LNCF within a valid
+	 * mslice is fine.
+	 */
+	for_each_set_bit(i, &gt->info.mslice_mask, GEN12_MAX_MSLICES)
+		lncf_mask |= (0x3 << (i * 2));
+
+	/*
+	 * Are there any sliceid values that work for both GSLICE and LNCF
+	 * steering?
+	 */
+	if (slice_mask & lncf_mask) {
+		slice_mask &= lncf_mask;
+		gt->steering_table[LNCF] = NULL;
+	}
+
+	/* How about sliceid values that also work for MSLICE steering? */
+	if (slice_mask & gt->info.mslice_mask) {
+		slice_mask &= gt->info.mslice_mask;
+		gt->steering_table[MSLICE] = NULL;
+	}
+
+	slice = __ffs(slice_mask);
+	subslice = intel_sseu_find_first_xehp_dss(sseu, GEN_DSS_PER_GSLICE, slice) %
+		GEN_DSS_PER_GSLICE;
+
+	__add_mcr_wa(gt, wal, slice, subslice);
+
+	/*
+	 * SQIDI ranges are special because they use different steering
+	 * registers than everything else we work with.  On XeHP SDV and
+	 * DG2-G10, any value in the steering registers will work fine since
+	 * all instances are present, but DG2-G11 only has SQIDI instances at
+	 * ID's 2 and 3, so we need to steer to one of those.  For simplicity
+	 * we'll just steer to a hardcoded "2" since that value will work
+	 * everywhere.
+	 */
+	__set_mcr_steering(wal, MCFG_MCR_SELECTOR, 0, 2);
+	__set_mcr_steering(wal, SF_MCR_SELECTOR, 0, 2);
+}
+
+static void
+pvc_init_mcr(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	unsigned int dss;
+
+	/*
+	 * Setup implicit steering for COMPUTE and DSS ranges to the first
+	 * non-fused-off DSS.  All other types of MCR registers will be
+	 * explicitly steered.
+	 */
+	dss = intel_sseu_find_first_xehp_dss(&gt->info.sseu, 0, 0);
+	__add_mcr_wa(gt, wal, dss / GEN_DSS_PER_CSLICE, dss % GEN_DSS_PER_CSLICE);
+}
+
+static void
+icl_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	icl_wa_init_mcr(gt, wal);
+
+	/* WaModifyGamTlbPartitioning:icl */
+	wa_write_clr_set(wal,
+			 GEN11_GACB_PERF_CTRL,
+			 GEN11_HASH_CTRL_MASK,
+			 GEN11_HASH_CTRL_BIT0 | GEN11_HASH_CTRL_BIT4);
+
+	/* Wa_1405766107:icl
+	 * Formerly known as WaCL2SFHalfMaxAlloc
+	 */
+	wa_write_or(wal,
+		    GEN11_LSN_UNSLCVC,
+		    GEN11_LSN_UNSLCVC_GAFS_HALF_SF_MAXALLOC |
+		    GEN11_LSN_UNSLCVC_GAFS_HALF_CL2_MAXALLOC);
+
+	/* Wa_220166154:icl
+	 * Formerly known as WaDisCtxReload
+	 */
+	wa_write_or(wal,
+		    GEN8_GAMW_ECO_DEV_RW_IA,
+		    GAMW_ECO_DEV_CTX_RELOAD_DISABLE);
+
+	/* Wa_1406463099:icl
+	 * Formerly known as WaGamTlbPendError
+	 */
+	wa_write_or(wal,
+		    GAMT_CHKN_BIT_REG,
+		    GAMT_CHKN_DISABLE_L3_COH_PIPE);
+
+	/*
+	 * Wa_1408615072:icl,ehl  (vsunit)
+	 * Wa_1407596294:icl,ehl  (hsunit)
+	 */
+	wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE,
+		    VSUNIT_CLKGATE_DIS | HSUNIT_CLKGATE_DIS);
+
+	/* Wa_1407352427:icl,ehl */
+	wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE2,
+		    PSDUNIT_CLKGATE_DIS);
+
+	/* Wa_1406680159:icl,ehl */
+	wa_write_or(wal,
+		    SUBSLICE_UNIT_LEVEL_CLKGATE,
+		    GWUNIT_CLKGATE_DIS);
+
+	/* Wa_1607087056:icl,ehl,jsl */
+	if (IS_ICELAKE(i915) ||
+	    IS_JSL_EHL_GRAPHICS_STEP(i915, STEP_A0, STEP_B0))
+		wa_write_or(wal,
+			    SLICE_UNIT_LEVEL_CLKGATE,
+			    L3_CLKGATE_DIS | L3_CR2X_CLKGATE_DIS);
+
+	/*
+	 * This is not a documented workaround, but rather an optimization
+	 * to reduce sampler power.
+	 */
+	wa_write_clr(wal, GEN10_DFR_RATIO_EN_AND_CHICKEN, DFR_DISABLE);
+}
+
+/*
+ * Though there are per-engine instances of these registers,
+ * they retain their value through engine resets and should
+ * only be provided on the GT workaround list rather than
+ * the engine-specific workaround list.
+ */
+static void
+wa_14011060649(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	struct intel_engine_cs *engine;
+	int id;
+
+	for_each_engine(engine, gt, id) {
+		if (engine->class != VIDEO_DECODE_CLASS ||
+		    (engine->instance % 2))
+			continue;
+
+		wa_write_or(wal, VDBOX_CGCTL3F10(engine->mmio_base),
+			    IECPUNIT_CLKGATE_DIS);
+	}
+}
+
+static void
+gen12_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	icl_wa_init_mcr(gt, wal);
+
+	/* Wa_14011060649:tgl,rkl,dg1,adl-s,adl-p */
+	wa_14011060649(gt, wal);
+
+	/* Wa_14011059788:tgl,rkl,adl-s,dg1,adl-p */
+	wa_write_or(wal, GEN10_DFR_RATIO_EN_AND_CHICKEN, DFR_DISABLE);
+}
+
+static void
+tgl_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	gen12_gt_workarounds_init(gt, wal);
+
+	/* Wa_1409420604:tgl */
+	if (IS_TGL_UY_GRAPHICS_STEP(i915, STEP_A0, STEP_B0))
+		wa_write_or(wal,
+			    SUBSLICE_UNIT_LEVEL_CLKGATE2,
+			    CPSSUNIT_CLKGATE_DIS);
+
+	/* Wa_1607087056:tgl also know as BUG:1409180338 */
+	if (IS_TGL_UY_GRAPHICS_STEP(i915, STEP_A0, STEP_B0))
+		wa_write_or(wal,
+			    SLICE_UNIT_LEVEL_CLKGATE,
+			    L3_CLKGATE_DIS | L3_CR2X_CLKGATE_DIS);
+
+	/* Wa_1408615072:tgl[a0] */
+	if (IS_TGL_UY_GRAPHICS_STEP(i915, STEP_A0, STEP_B0))
+		wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE2,
+			    VSUNIT_CLKGATE_DIS_TGL);
+}
+
+static void
+dg1_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	gen12_gt_workarounds_init(gt, wal);
+
+	/* Wa_1607087056:dg1 */
+	if (IS_DG1_GRAPHICS_STEP(i915, STEP_A0, STEP_B0))
+		wa_write_or(wal,
+			    SLICE_UNIT_LEVEL_CLKGATE,
+			    L3_CLKGATE_DIS | L3_CR2X_CLKGATE_DIS);
+
+	/* Wa_1409420604:dg1 */
+	if (IS_DG1(i915))
+		wa_write_or(wal,
+			    SUBSLICE_UNIT_LEVEL_CLKGATE2,
+			    CPSSUNIT_CLKGATE_DIS);
+
+	/* Wa_1408615072:dg1 */
+	/* Empirical testing shows this register is unaffected by engine reset. */
+	if (IS_DG1(i915))
+		wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE2,
+			    VSUNIT_CLKGATE_DIS_TGL);
+}
+
+static void
+xehpsdv_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	xehp_init_mcr(gt, wal);
+
+	/* Wa_1409757795:xehpsdv */
+	wa_write_or(wal, SCCGCTL94DC, CG3DDISURB);
+
+	/* Wa_16011155590:xehpsdv */
+	if (IS_XEHPSDV_GRAPHICS_STEP(i915, STEP_A0, STEP_B0))
+		wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE,
+			    TSGUNIT_CLKGATE_DIS);
+
+	/* Wa_14011780169:xehpsdv */
+	if (IS_XEHPSDV_GRAPHICS_STEP(i915, STEP_B0, STEP_FOREVER)) {
+		wa_write_or(wal, UNSLCGCTL9440, GAMTLBOACS_CLKGATE_DIS |
+			    GAMTLBVDBOX7_CLKGATE_DIS |
+			    GAMTLBVDBOX6_CLKGATE_DIS |
+			    GAMTLBVDBOX5_CLKGATE_DIS |
+			    GAMTLBVDBOX4_CLKGATE_DIS |
+			    GAMTLBVDBOX3_CLKGATE_DIS |
+			    GAMTLBVDBOX2_CLKGATE_DIS |
+			    GAMTLBVDBOX1_CLKGATE_DIS |
+			    GAMTLBVDBOX0_CLKGATE_DIS |
+			    GAMTLBKCR_CLKGATE_DIS |
+			    GAMTLBGUC_CLKGATE_DIS |
+			    GAMTLBBLT_CLKGATE_DIS);
+		wa_write_or(wal, UNSLCGCTL9444, GAMTLBGFXA0_CLKGATE_DIS |
+			    GAMTLBGFXA1_CLKGATE_DIS |
+			    GAMTLBCOMPA0_CLKGATE_DIS |
+			    GAMTLBCOMPA1_CLKGATE_DIS |
+			    GAMTLBCOMPB0_CLKGATE_DIS |
+			    GAMTLBCOMPB1_CLKGATE_DIS |
+			    GAMTLBCOMPC0_CLKGATE_DIS |
+			    GAMTLBCOMPC1_CLKGATE_DIS |
+			    GAMTLBCOMPD0_CLKGATE_DIS |
+			    GAMTLBCOMPD1_CLKGATE_DIS |
+			    GAMTLBMERT_CLKGATE_DIS   |
+			    GAMTLBVEBOX3_CLKGATE_DIS |
+			    GAMTLBVEBOX2_CLKGATE_DIS |
+			    GAMTLBVEBOX1_CLKGATE_DIS |
+			    GAMTLBVEBOX0_CLKGATE_DIS);
+	}
+
+	/* Wa_16012725990:xehpsdv */
+	if (IS_XEHPSDV_GRAPHICS_STEP(i915, STEP_A1, STEP_FOREVER))
+		wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE, VFUNIT_CLKGATE_DIS);
+
+	/* Wa_14011060649:xehpsdv */
+	wa_14011060649(gt, wal);
+}
+
+static void
+dg2_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	struct intel_engine_cs *engine;
+	int id;
+
+	xehp_init_mcr(gt, wal);
+
+	/* Wa_14011060649:dg2 */
+	wa_14011060649(gt, wal);
+
+	/*
+	 * Although there are per-engine instances of these registers,
+	 * they technically exist outside the engine itself and are not
+	 * impacted by engine resets.  Furthermore, they're part of the
+	 * GuC blacklist so trying to treat them as engine workarounds
+	 * will result in GuC initialization failure and a wedged GPU.
+	 */
+	for_each_engine(engine, gt, id) {
+		if (engine->class != VIDEO_DECODE_CLASS)
+			continue;
+
+		/* Wa_16010515920:dg2_g10 */
+		if (IS_DG2_GRAPHICS_STEP(gt->i915, G10, STEP_A0, STEP_B0))
+			wa_write_or(wal, VDBOX_CGCTL3F18(engine->mmio_base),
+				    ALNUNIT_CLKGATE_DIS);
+	}
+
+	if (IS_DG2_G10(gt->i915)) {
+		/* Wa_22010523718:dg2 */
+		wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE,
+			    CG3DDISCFEG_CLKGATE_DIS);
+
+		/* Wa_14011006942:dg2 */
+		wa_write_or(wal, SUBSLICE_UNIT_LEVEL_CLKGATE,
+			    DSS_ROUTER_CLKGATE_DIS);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(gt->i915, G10, STEP_A0, STEP_B0)) {
+		/* Wa_14010948348:dg2_g10 */
+		wa_write_or(wal, UNSLCGCTL9430, MSQDUNIT_CLKGATE_DIS);
+
+		/* Wa_14011037102:dg2_g10 */
+		wa_write_or(wal, UNSLCGCTL9444, LTCDD_CLKGATE_DIS);
+
+		/* Wa_14011371254:dg2_g10 */
+		wa_write_or(wal, SLICE_UNIT_LEVEL_CLKGATE, NODEDSS_CLKGATE_DIS);
+
+		/* Wa_14011431319:dg2_g10 */
+		wa_write_or(wal, UNSLCGCTL9440, GAMTLBOACS_CLKGATE_DIS |
+			    GAMTLBVDBOX7_CLKGATE_DIS |
+			    GAMTLBVDBOX6_CLKGATE_DIS |
+			    GAMTLBVDBOX5_CLKGATE_DIS |
+			    GAMTLBVDBOX4_CLKGATE_DIS |
+			    GAMTLBVDBOX3_CLKGATE_DIS |
+			    GAMTLBVDBOX2_CLKGATE_DIS |
+			    GAMTLBVDBOX1_CLKGATE_DIS |
+			    GAMTLBVDBOX0_CLKGATE_DIS |
+			    GAMTLBKCR_CLKGATE_DIS |
+			    GAMTLBGUC_CLKGATE_DIS |
+			    GAMTLBBLT_CLKGATE_DIS);
+		wa_write_or(wal, UNSLCGCTL9444, GAMTLBGFXA0_CLKGATE_DIS |
+			    GAMTLBGFXA1_CLKGATE_DIS |
+			    GAMTLBCOMPA0_CLKGATE_DIS |
+			    GAMTLBCOMPA1_CLKGATE_DIS |
+			    GAMTLBCOMPB0_CLKGATE_DIS |
+			    GAMTLBCOMPB1_CLKGATE_DIS |
+			    GAMTLBCOMPC0_CLKGATE_DIS |
+			    GAMTLBCOMPC1_CLKGATE_DIS |
+			    GAMTLBCOMPD0_CLKGATE_DIS |
+			    GAMTLBCOMPD1_CLKGATE_DIS |
+			    GAMTLBMERT_CLKGATE_DIS   |
+			    GAMTLBVEBOX3_CLKGATE_DIS |
+			    GAMTLBVEBOX2_CLKGATE_DIS |
+			    GAMTLBVEBOX1_CLKGATE_DIS |
+			    GAMTLBVEBOX0_CLKGATE_DIS);
+
+		/* Wa_14010569222:dg2_g10 */
+		wa_write_or(wal, UNSLICE_UNIT_LEVEL_CLKGATE,
+			    GAMEDIA_CLKGATE_DIS);
+
+		/* Wa_14011028019:dg2_g10 */
+		wa_write_or(wal, SSMCGCTL9530, RTFUNIT_CLKGATE_DIS);
+	}
+
+	/* Wa_14014830051:dg2 */
+	wa_write_clr(wal, SARB_CHICKEN1, COMP_CKN_IN);
+
+	/*
+	 * The following are not actually "workarounds" but rather
+	 * recommended tuning settings documented in the bspec's
+	 * performance guide section.
+	 */
+	wa_write_or(wal, GEN12_SQCM, EN_32B_ACCESS);
+
+	/* Wa_14015795083 */
+	wa_write_clr(wal, GEN7_MISCCPCTL, GEN12_DOP_CLOCK_GATE_RENDER_ENABLE);
+}
+
+static void
+pvc_gt_workarounds_init(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	pvc_init_mcr(gt, wal);
+
+	/* Wa_14015795083 */
+	wa_write_clr(wal, GEN7_MISCCPCTL, GEN12_DOP_CLOCK_GATE_RENDER_ENABLE);
+}
+
+static void
+gt_init_workarounds(struct intel_gt *gt, struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = gt->i915;
+
+	if (IS_PONTEVECCHIO(i915))
+		pvc_gt_workarounds_init(gt, wal);
+	else if (IS_DG2(i915))
+		dg2_gt_workarounds_init(gt, wal);
+	else if (IS_XEHPSDV(i915))
+		xehpsdv_gt_workarounds_init(gt, wal);
+	else if (IS_DG1(i915))
+		dg1_gt_workarounds_init(gt, wal);
+	else if (IS_TIGERLAKE(i915))
+		tgl_gt_workarounds_init(gt, wal);
+	else if (GRAPHICS_VER(i915) == 12)
+		gen12_gt_workarounds_init(gt, wal);
+	else if (GRAPHICS_VER(i915) == 11)
+		icl_gt_workarounds_init(gt, wal);
+	else if (IS_COFFEELAKE(i915) || IS_COMETLAKE(i915))
+		cfl_gt_workarounds_init(gt, wal);
+	else if (IS_GEMINILAKE(i915))
+		glk_gt_workarounds_init(gt, wal);
+	else if (IS_KABYLAKE(i915))
+		kbl_gt_workarounds_init(gt, wal);
+	else if (IS_BROXTON(i915))
+		gen9_gt_workarounds_init(gt, wal);
+	else if (IS_SKYLAKE(i915))
+		skl_gt_workarounds_init(gt, wal);
+	else if (IS_HASWELL(i915))
+		hsw_gt_workarounds_init(gt, wal);
+	else if (IS_VALLEYVIEW(i915))
+		vlv_gt_workarounds_init(gt, wal);
+	else if (IS_IVYBRIDGE(i915))
+		ivb_gt_workarounds_init(gt, wal);
+	else if (GRAPHICS_VER(i915) == 6)
+		snb_gt_workarounds_init(gt, wal);
+	else if (GRAPHICS_VER(i915) == 5)
+		ilk_gt_workarounds_init(gt, wal);
+	else if (IS_G4X(i915))
+		g4x_gt_workarounds_init(gt, wal);
+	else if (GRAPHICS_VER(i915) == 4)
+		gen4_gt_workarounds_init(gt, wal);
+	else if (GRAPHICS_VER(i915) <= 8)
+		;
+	else
+		MISSING_CASE(GRAPHICS_VER(i915));
+}
+
+void intel_gt_init_workarounds(struct intel_gt *gt)
+{
+	struct i915_wa_list *wal = &gt->wa_list;
+
+	wa_init_start(wal, "GT", "global");
+	gt_init_workarounds(gt, wal);
+	wa_init_finish(wal);
+}
+
+static enum forcewake_domains
+wal_get_fw_for_rmw(struct intel_uncore *uncore, const struct i915_wa_list *wal)
+{
+	enum forcewake_domains fw = 0;
+	struct i915_wa *wa;
+	unsigned int i;
+
+	for (i = 0, wa = wal->list; i < wal->count; i++, wa++)
+		fw |= intel_uncore_forcewake_for_reg(uncore,
+						     wa->reg,
+						     FW_REG_READ |
+						     FW_REG_WRITE);
+
+	return fw;
+}
+
+static bool
+wa_verify(const struct i915_wa *wa, u32 cur, const char *name, const char *from)
+{
+	if ((cur ^ wa->set) & wa->read) {
+		DRM_ERROR("%s workaround lost on %s! (reg[%x]=0x%x, relevant bits were 0x%x vs expected 0x%x)\n",
+			  name, from, i915_mmio_reg_offset(wa->reg),
+			  cur, cur & wa->read, wa->set & wa->read);
+
+		return false;
+	}
+
+	return true;
+}
+
+static void
+wa_list_apply(struct intel_gt *gt, const struct i915_wa_list *wal)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	enum forcewake_domains fw;
+	unsigned long flags;
+	struct i915_wa *wa;
+	unsigned int i;
+
+	if (!wal->count)
+		return;
+
+	fw = wal_get_fw_for_rmw(uncore, wal);
+
+	spin_lock_irqsave(&uncore->lock, flags);
+	intel_uncore_forcewake_get__locked(uncore, fw);
+
+	for (i = 0, wa = wal->list; i < wal->count; i++, wa++) {
+		u32 val, old = 0;
+
+		/* open-coded rmw due to steering */
+		old = wa->clr ? intel_gt_mcr_read_any_fw(gt, wa->reg) : 0;
+		val = (old & ~wa->clr) | wa->set;
+		if (val != old || !wa->clr)
+			intel_uncore_write_fw(uncore, wa->reg, val);
+
+		if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+			wa_verify(wa, intel_gt_mcr_read_any_fw(gt, wa->reg),
+				  wal->name, "application");
+	}
+
+	intel_uncore_forcewake_put__locked(uncore, fw);
+	spin_unlock_irqrestore(&uncore->lock, flags);
+}
+
+void intel_gt_apply_workarounds(struct intel_gt *gt)
+{
+	wa_list_apply(gt, &gt->wa_list);
+}
+
+static bool wa_list_verify(struct intel_gt *gt,
+			   const struct i915_wa_list *wal,
+			   const char *from)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	struct i915_wa *wa;
+	enum forcewake_domains fw;
+	unsigned long flags;
+	unsigned int i;
+	bool ok = true;
+
+	fw = wal_get_fw_for_rmw(uncore, wal);
+
+	spin_lock_irqsave(&uncore->lock, flags);
+	intel_uncore_forcewake_get__locked(uncore, fw);
+
+	for (i = 0, wa = wal->list; i < wal->count; i++, wa++)
+		ok &= wa_verify(wa,
+				intel_gt_mcr_read_any_fw(gt, wa->reg),
+				wal->name, from);
+
+	intel_uncore_forcewake_put__locked(uncore, fw);
+	spin_unlock_irqrestore(&uncore->lock, flags);
+
+	return ok;
+}
+
+bool intel_gt_verify_workarounds(struct intel_gt *gt, const char *from)
+{
+	return wa_list_verify(gt, &gt->wa_list, from);
+}
+
+__maybe_unused
+static bool is_nonpriv_flags_valid(u32 flags)
+{
+	/* Check only valid flag bits are set */
+	if (flags & ~RING_FORCE_TO_NONPRIV_MASK_VALID)
+		return false;
+
+	/* NB: Only 3 out of 4 enum values are valid for access field */
+	if ((flags & RING_FORCE_TO_NONPRIV_ACCESS_MASK) ==
+	    RING_FORCE_TO_NONPRIV_ACCESS_INVALID)
+		return false;
+
+	return true;
+}
+
+static void
+whitelist_reg_ext(struct i915_wa_list *wal, i915_reg_t reg, u32 flags)
+{
+	struct i915_wa wa = {
+		.reg = reg
+	};
+
+	if (GEM_DEBUG_WARN_ON(wal->count >= RING_MAX_NONPRIV_SLOTS))
+		return;
+
+	if (GEM_DEBUG_WARN_ON(!is_nonpriv_flags_valid(flags)))
+		return;
+
+	wa.reg.reg |= flags;
+	_wa_add(wal, &wa);
+}
+
+static void
+whitelist_reg(struct i915_wa_list *wal, i915_reg_t reg)
+{
+	whitelist_reg_ext(wal, reg, RING_FORCE_TO_NONPRIV_ACCESS_RW);
+}
+
+static void gen9_whitelist_build(struct i915_wa_list *w)
+{
+	/* WaVFEStateAfterPipeControlwithMediaStateClear:skl,bxt,glk,cfl */
+	whitelist_reg(w, GEN9_CTX_PREEMPT_REG);
+
+	/* WaEnablePreemptionGranularityControlByUMD:skl,bxt,kbl,cfl,[cnl] */
+	whitelist_reg(w, GEN8_CS_CHICKEN1);
+
+	/* WaAllowUMDToModifyHDCChicken1:skl,bxt,kbl,glk,cfl */
+	whitelist_reg(w, GEN8_HDC_CHICKEN1);
+
+	/* WaSendPushConstantsFromMMIO:skl,bxt */
+	whitelist_reg(w, COMMON_SLICE_CHICKEN2);
+}
+
+static void skl_whitelist_build(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *w = &engine->whitelist;
+
+	if (engine->class != RENDER_CLASS)
+		return;
+
+	gen9_whitelist_build(w);
+
+	/* WaDisableLSQCROPERFforOCL:skl */
+	whitelist_reg(w, GEN8_L3SQCREG4);
+}
+
+static void bxt_whitelist_build(struct intel_engine_cs *engine)
+{
+	if (engine->class != RENDER_CLASS)
+		return;
+
+	gen9_whitelist_build(&engine->whitelist);
+}
+
+static void kbl_whitelist_build(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *w = &engine->whitelist;
+
+	if (engine->class != RENDER_CLASS)
+		return;
+
+	gen9_whitelist_build(w);
+
+	/* WaDisableLSQCROPERFforOCL:kbl */
+	whitelist_reg(w, GEN8_L3SQCREG4);
+}
+
+static void glk_whitelist_build(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *w = &engine->whitelist;
+
+	if (engine->class != RENDER_CLASS)
+		return;
+
+	gen9_whitelist_build(w);
+
+	/* WA #0862: Userspace has to set "Barrier Mode" to avoid hangs. */
+	whitelist_reg(w, GEN9_SLICE_COMMON_ECO_CHICKEN1);
+}
+
+static void cfl_whitelist_build(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *w = &engine->whitelist;
+
+	if (engine->class != RENDER_CLASS)
+		return;
+
+	gen9_whitelist_build(w);
+
+	/*
+	 * WaAllowPMDepthAndInvocationCountAccessFromUMD:cfl,whl,cml,aml
+	 *
+	 * This covers 4 register which are next to one another :
+	 *   - PS_INVOCATION_COUNT
+	 *   - PS_INVOCATION_COUNT_UDW
+	 *   - PS_DEPTH_COUNT
+	 *   - PS_DEPTH_COUNT_UDW
+	 */
+	whitelist_reg_ext(w, PS_INVOCATION_COUNT,
+			  RING_FORCE_TO_NONPRIV_ACCESS_RD |
+			  RING_FORCE_TO_NONPRIV_RANGE_4);
+}
+
+static void allow_read_ctx_timestamp(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *w = &engine->whitelist;
+
+	if (engine->class != RENDER_CLASS)
+		whitelist_reg_ext(w,
+				  RING_CTX_TIMESTAMP(engine->mmio_base),
+				  RING_FORCE_TO_NONPRIV_ACCESS_RD);
+}
+
+static void cml_whitelist_build(struct intel_engine_cs *engine)
+{
+	allow_read_ctx_timestamp(engine);
+
+	cfl_whitelist_build(engine);
+}
+
+static void icl_whitelist_build(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *w = &engine->whitelist;
+
+	allow_read_ctx_timestamp(engine);
+
+	switch (engine->class) {
+	case RENDER_CLASS:
+		/* WaAllowUMDToModifyHalfSliceChicken7:icl */
+		whitelist_reg(w, GEN9_HALF_SLICE_CHICKEN7);
+
+		/* WaAllowUMDToModifySamplerMode:icl */
+		whitelist_reg(w, GEN10_SAMPLER_MODE);
+
+		/* WaEnableStateCacheRedirectToCS:icl */
+		whitelist_reg(w, GEN9_SLICE_COMMON_ECO_CHICKEN1);
+
+		/*
+		 * WaAllowPMDepthAndInvocationCountAccessFromUMD:icl
+		 *
+		 * This covers 4 register which are next to one another :
+		 *   - PS_INVOCATION_COUNT
+		 *   - PS_INVOCATION_COUNT_UDW
+		 *   - PS_DEPTH_COUNT
+		 *   - PS_DEPTH_COUNT_UDW
+		 */
+		whitelist_reg_ext(w, PS_INVOCATION_COUNT,
+				  RING_FORCE_TO_NONPRIV_ACCESS_RD |
+				  RING_FORCE_TO_NONPRIV_RANGE_4);
+		break;
+
+	case VIDEO_DECODE_CLASS:
+		/* hucStatusRegOffset */
+		whitelist_reg_ext(w, _MMIO(0x2000 + engine->mmio_base),
+				  RING_FORCE_TO_NONPRIV_ACCESS_RD);
+		/* hucUKernelHdrInfoRegOffset */
+		whitelist_reg_ext(w, _MMIO(0x2014 + engine->mmio_base),
+				  RING_FORCE_TO_NONPRIV_ACCESS_RD);
+		/* hucStatus2RegOffset */
+		whitelist_reg_ext(w, _MMIO(0x23B0 + engine->mmio_base),
+				  RING_FORCE_TO_NONPRIV_ACCESS_RD);
+		break;
+
+	default:
+		break;
+	}
+}
+
+static void tgl_whitelist_build(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *w = &engine->whitelist;
+
+	allow_read_ctx_timestamp(engine);
+
+	switch (engine->class) {
+	case RENDER_CLASS:
+		/*
+		 * WaAllowPMDepthAndInvocationCountAccessFromUMD:tgl
+		 * Wa_1408556865:tgl
+		 *
+		 * This covers 4 registers which are next to one another :
+		 *   - PS_INVOCATION_COUNT
+		 *   - PS_INVOCATION_COUNT_UDW
+		 *   - PS_DEPTH_COUNT
+		 *   - PS_DEPTH_COUNT_UDW
+		 */
+		whitelist_reg_ext(w, PS_INVOCATION_COUNT,
+				  RING_FORCE_TO_NONPRIV_ACCESS_RD |
+				  RING_FORCE_TO_NONPRIV_RANGE_4);
+
+		/*
+		 * Wa_1808121037:tgl
+		 * Wa_14012131227:dg1
+		 * Wa_1508744258:tgl,rkl,dg1,adl-s,adl-p
+		 */
+		whitelist_reg(w, GEN7_COMMON_SLICE_CHICKEN1);
+
+		/* Wa_1806527549:tgl */
+		whitelist_reg(w, HIZ_CHICKEN);
+		break;
+	default:
+		break;
+	}
+}
+
+static void dg1_whitelist_build(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *w = &engine->whitelist;
+
+	tgl_whitelist_build(engine);
+
+	/* GEN:BUG:1409280441:dg1 */
+	if (IS_DG1_GRAPHICS_STEP(engine->i915, STEP_A0, STEP_B0) &&
+	    (engine->class == RENDER_CLASS ||
+	     engine->class == COPY_ENGINE_CLASS))
+		whitelist_reg_ext(w, RING_ID(engine->mmio_base),
+				  RING_FORCE_TO_NONPRIV_ACCESS_RD);
+}
+
+static void xehpsdv_whitelist_build(struct intel_engine_cs *engine)
+{
+	allow_read_ctx_timestamp(engine);
+}
+
+static void dg2_whitelist_build(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *w = &engine->whitelist;
+
+	allow_read_ctx_timestamp(engine);
+
+	switch (engine->class) {
+	case RENDER_CLASS:
+		/*
+		 * Wa_1507100340:dg2_g10
+		 *
+		 * This covers 4 registers which are next to one another :
+		 *   - PS_INVOCATION_COUNT
+		 *   - PS_INVOCATION_COUNT_UDW
+		 *   - PS_DEPTH_COUNT
+		 *   - PS_DEPTH_COUNT_UDW
+		 */
+		if (IS_DG2_GRAPHICS_STEP(engine->i915, G10, STEP_A0, STEP_B0))
+			whitelist_reg_ext(w, PS_INVOCATION_COUNT,
+					  RING_FORCE_TO_NONPRIV_ACCESS_RD |
+					  RING_FORCE_TO_NONPRIV_RANGE_4);
+
+		break;
+	case COMPUTE_CLASS:
+		/* Wa_16011157294:dg2_g10 */
+		if (IS_DG2_GRAPHICS_STEP(engine->i915, G10, STEP_A0, STEP_B0))
+			whitelist_reg(w, GEN9_CTX_PREEMPT_REG);
+		break;
+	default:
+		break;
+	}
+}
+
+static void blacklist_trtt(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *w = &engine->whitelist;
+
+	/*
+	 * Prevent read/write access to [0x4400, 0x4600) which covers
+	 * the TRTT range across all engines. Note that normally userspace
+	 * cannot access the other engines' trtt control, but for simplicity
+	 * we cover the entire range on each engine.
+	 */
+	whitelist_reg_ext(w, _MMIO(0x4400),
+			  RING_FORCE_TO_NONPRIV_DENY |
+			  RING_FORCE_TO_NONPRIV_RANGE_64);
+	whitelist_reg_ext(w, _MMIO(0x4500),
+			  RING_FORCE_TO_NONPRIV_DENY |
+			  RING_FORCE_TO_NONPRIV_RANGE_64);
+}
+
+static void pvc_whitelist_build(struct intel_engine_cs *engine)
+{
+	allow_read_ctx_timestamp(engine);
+
+	/* Wa_16014440446:pvc */
+	blacklist_trtt(engine);
+}
+
+void intel_engine_init_whitelist(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+	struct i915_wa_list *w = &engine->whitelist;
+
+	wa_init_start(w, "whitelist", engine->name);
+
+	if (IS_PONTEVECCHIO(i915))
+		pvc_whitelist_build(engine);
+	else if (IS_DG2(i915))
+		dg2_whitelist_build(engine);
+	else if (IS_XEHPSDV(i915))
+		xehpsdv_whitelist_build(engine);
+	else if (IS_DG1(i915))
+		dg1_whitelist_build(engine);
+	else if (GRAPHICS_VER(i915) == 12)
+		tgl_whitelist_build(engine);
+	else if (GRAPHICS_VER(i915) == 11)
+		icl_whitelist_build(engine);
+	else if (IS_COMETLAKE(i915))
+		cml_whitelist_build(engine);
+	else if (IS_COFFEELAKE(i915))
+		cfl_whitelist_build(engine);
+	else if (IS_GEMINILAKE(i915))
+		glk_whitelist_build(engine);
+	else if (IS_KABYLAKE(i915))
+		kbl_whitelist_build(engine);
+	else if (IS_BROXTON(i915))
+		bxt_whitelist_build(engine);
+	else if (IS_SKYLAKE(i915))
+		skl_whitelist_build(engine);
+	else if (GRAPHICS_VER(i915) <= 8)
+		;
+	else
+		MISSING_CASE(GRAPHICS_VER(i915));
+
+	wa_init_finish(w);
+}
+
+void intel_engine_apply_whitelist(struct intel_engine_cs *engine)
+{
+	const struct i915_wa_list *wal = &engine->whitelist;
+	struct intel_uncore *uncore = engine->uncore;
+	const u32 base = engine->mmio_base;
+	struct i915_wa *wa;
+	unsigned int i;
+
+	if (!wal->count)
+		return;
+
+	for (i = 0, wa = wal->list; i < wal->count; i++, wa++)
+		intel_uncore_write(uncore,
+				   RING_FORCE_TO_NONPRIV(base, i),
+				   i915_mmio_reg_offset(wa->reg));
+
+	/* And clear the rest just in case of garbage */
+	for (; i < RING_MAX_NONPRIV_SLOTS; i++)
+		intel_uncore_write(uncore,
+				   RING_FORCE_TO_NONPRIV(base, i),
+				   i915_mmio_reg_offset(RING_NOPID(base)));
+}
+
+/*
+ * engine_fake_wa_init(), a place holder to program the registers
+ * which are not part of an official workaround defined by the
+ * hardware team.
+ * Adding programming of those register inside workaround will
+ * allow utilizing wa framework to proper application and verification.
+ */
+static void
+engine_fake_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
+{
+	u8 mocs_w, mocs_r;
+
+	/*
+	 * RING_CMD_CCTL specifies the default MOCS entry that will be used
+	 * by the command streamer when executing commands that don't have
+	 * a way to explicitly specify a MOCS setting.  The default should
+	 * usually reference whichever MOCS entry corresponds to uncached
+	 * behavior, although use of a WB cached entry is recommended by the
+	 * spec in certain circumstances on specific platforms.
+	 */
+	if (GRAPHICS_VER(engine->i915) >= 12) {
+		mocs_r = engine->gt->mocs.uc_index;
+		mocs_w = engine->gt->mocs.uc_index;
+
+		if (HAS_L3_CCS_READ(engine->i915) &&
+		    engine->class == COMPUTE_CLASS) {
+			mocs_r = engine->gt->mocs.wb_index;
+
+			/*
+			 * Even on the few platforms where MOCS 0 is a
+			 * legitimate table entry, it's never the correct
+			 * setting to use here; we can assume the MOCS init
+			 * just forgot to initialize wb_index.
+			 */
+			drm_WARN_ON(&engine->i915->drm, mocs_r == 0);
+		}
+
+		wa_masked_field_set(wal,
+				    RING_CMD_CCTL(engine->mmio_base),
+				    CMD_CCTL_MOCS_MASK,
+				    CMD_CCTL_MOCS_OVERRIDE(mocs_w, mocs_r));
+	}
+}
+
+static bool needs_wa_1308578152(struct intel_engine_cs *engine)
+{
+	return intel_sseu_find_first_xehp_dss(&engine->gt->info.sseu, 0, 0) >=
+		GEN_DSS_PER_GSLICE;
+}
+
+static void
+rcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	if (IS_DG2(i915)) {
+		/* Wa_1509235366:dg2 */
+		wa_write_or(wal, GEN12_GAMCNTRL_CTRL, INVALIDATION_BROADCAST_MODE_DIS |
+			    GLOBAL_INVALIDATION_MODE);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(i915, G11, STEP_A0, STEP_B0)) {
+		/* Wa_14013392000:dg2_g11 */
+		wa_masked_en(wal, GEN7_ROW_CHICKEN2, GEN12_ENABLE_LARGE_GRF_MODE);
+
+		/* Wa_16011620976:dg2_g11 */
+		wa_write_or(wal, LSC_CHICKEN_BIT_0_UDW, DIS_CHAIN_2XSIMD8);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(i915, G10, STEP_B0, STEP_FOREVER) ||
+	    IS_DG2_G11(i915) || IS_DG2_G12(i915)) {
+		/* Wa_1509727124:dg2 */
+		wa_masked_en(wal, GEN10_SAMPLER_MODE,
+			     SC_DISABLE_POWER_OPTIMIZATION_EBB);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(i915, G10, STEP_A0, STEP_B0) ||
+	    IS_DG2_GRAPHICS_STEP(i915, G11, STEP_A0, STEP_B0)) {
+		/* Wa_14012419201:dg2 */
+		wa_masked_en(wal, GEN9_ROW_CHICKEN4,
+			     GEN12_DISABLE_HDR_PAST_PAYLOAD_HOLD_FIX);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(i915, G10, STEP_B0, STEP_C0) ||
+	    IS_DG2_G11(i915)) {
+		/*
+		 * Wa_22012826095:dg2
+		 * Wa_22013059131:dg2
+		 */
+		wa_write_clr_set(wal, LSC_CHICKEN_BIT_0_UDW,
+				 MAXREQS_PER_BANK,
+				 REG_FIELD_PREP(MAXREQS_PER_BANK, 2));
+
+		/* Wa_22013059131:dg2 */
+		wa_write_or(wal, LSC_CHICKEN_BIT_0,
+			    FORCE_1_SUB_MESSAGE_PER_FRAGMENT);
+	}
+
+	/* Wa_1308578152:dg2_g10 when first gslice is fused off */
+	if (IS_DG2_GRAPHICS_STEP(i915, G10, STEP_B0, STEP_C0) &&
+	    needs_wa_1308578152(engine)) {
+		wa_masked_dis(wal, GEN12_CS_DEBUG_MODE1_CCCSUNIT_BE_COMMON,
+			      GEN12_REPLAY_MODE_GRANULARITY);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(i915, G10, STEP_B0, STEP_FOREVER) ||
+	    IS_DG2_G11(i915) || IS_DG2_G12(i915)) {
+		/* Wa_22013037850:dg2 */
+		wa_write_or(wal, LSC_CHICKEN_BIT_0_UDW,
+			    DISABLE_128B_EVICTION_COMMAND_UDW);
+
+		/* Wa_22012856258:dg2 */
+		wa_masked_en(wal, GEN7_ROW_CHICKEN2,
+			     GEN12_DISABLE_READ_SUPPRESSION);
+
+		/*
+		 * Wa_22010960976:dg2
+		 * Wa_14013347512:dg2
+		 */
+		wa_masked_dis(wal, GEN12_HDC_CHICKEN0,
+			      LSC_L1_FLUSH_CTL_3D_DATAPORT_FLUSH_EVENTS_MASK);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(i915, G10, STEP_A0, STEP_B0)) {
+		/*
+		 * Wa_1608949956:dg2_g10
+		 * Wa_14010198302:dg2_g10
+		 */
+		wa_masked_en(wal, GEN8_ROW_CHICKEN,
+			     MDQ_ARBITRATION_MODE | UGM_BACKUP_MODE);
+
+		/*
+		 * Wa_14010918519:dg2_g10
+		 *
+		 * LSC_CHICKEN_BIT_0 always reads back as 0 is this stepping,
+		 * so ignoring verification.
+		 */
+		wa_add(wal, LSC_CHICKEN_BIT_0_UDW, 0,
+		       FORCE_SLM_FENCE_SCOPE_TO_TILE | FORCE_UGM_FENCE_SCOPE_TO_TILE,
+		       0, false);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(i915, G10, STEP_A0, STEP_B0)) {
+		/* Wa_22010430635:dg2 */
+		wa_masked_en(wal,
+			     GEN9_ROW_CHICKEN4,
+			     GEN12_DISABLE_GRF_CLEAR);
+
+		/* Wa_14010648519:dg2 */
+		wa_write_or(wal, XEHP_L3NODEARBCFG, XEHP_LNESPARE);
+	}
+
+	/* Wa_14013202645:dg2 */
+	if (IS_DG2_GRAPHICS_STEP(i915, G10, STEP_B0, STEP_C0) ||
+	    IS_DG2_GRAPHICS_STEP(i915, G11, STEP_A0, STEP_B0))
+		wa_write_or(wal, RT_CTRL, DIS_NULL_QUERY);
+
+	/* Wa_22012532006:dg2 */
+	if (IS_DG2_GRAPHICS_STEP(engine->i915, G10, STEP_A0, STEP_C0) ||
+	    IS_DG2_GRAPHICS_STEP(engine->i915, G11, STEP_A0, STEP_B0))
+		wa_masked_en(wal, GEN9_HALF_SLICE_CHICKEN7,
+			     DG2_DISABLE_ROUND_ENABLE_ALLOW_FOR_SSLA);
+
+	if (IS_DG2_GRAPHICS_STEP(engine->i915, G10, STEP_A0, STEP_B0)) {
+		/* Wa_14010680813:dg2_g10 */
+		wa_write_or(wal, GEN12_GAMSTLB_CTRL, CONTROL_BLOCK_CLKGATE_DIS |
+			    EGRESS_BLOCK_CLKGATE_DIS | TAG_BLOCK_CLKGATE_DIS);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(engine->i915, G10, STEP_A0, STEP_B0) ||
+	    IS_DG2_GRAPHICS_STEP(engine->i915, G11, STEP_A0, STEP_B0)) {
+		/* Wa_14012362059:dg2 */
+		wa_write_or(wal, GEN12_MERT_MOD_CTRL, FORCE_MISS_FTLB);
+	}
+
+	if (IS_DG2_GRAPHICS_STEP(i915, G11, STEP_B0, STEP_FOREVER) ||
+	    IS_DG2_G10(i915)) {
+		/* Wa_22014600077:dg2 */
+		wa_add(wal, GEN10_CACHE_MODE_SS, 0,
+		       _MASKED_BIT_ENABLE(ENABLE_EU_COUNT_FOR_TDL_FLUSH),
+		       0 /* Wa_14012342262 :write-only reg, so skip
+			    verification */,
+		       true);
+	}
+
+	if (IS_DG1_GRAPHICS_STEP(i915, STEP_A0, STEP_B0) ||
+	    IS_TGL_UY_GRAPHICS_STEP(i915, STEP_A0, STEP_B0)) {
+		/*
+		 * Wa_1607138336:tgl[a0],dg1[a0]
+		 * Wa_1607063988:tgl[a0],dg1[a0]
+		 */
+		wa_write_or(wal,
+			    GEN9_CTX_PREEMPT_REG,
+			    GEN12_DISABLE_POSH_BUSY_FF_DOP_CG);
+	}
+
+	if (IS_TGL_UY_GRAPHICS_STEP(i915, STEP_A0, STEP_B0)) {
+		/*
+		 * Wa_1606679103:tgl
+		 * (see also Wa_1606682166:icl)
+		 */
+		wa_write_or(wal,
+			    GEN7_SARCHKMD,
+			    GEN7_DISABLE_SAMPLER_PREFETCH);
+	}
+
+	if (IS_ALDERLAKE_P(i915) || IS_ALDERLAKE_S(i915) || IS_DG1(i915) ||
+	    IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915)) {
+		/* Wa_1606931601:tgl,rkl,dg1,adl-s,adl-p */
+		wa_masked_en(wal, GEN7_ROW_CHICKEN2, GEN12_DISABLE_EARLY_READ);
+
+		/*
+		 * Wa_1407928979:tgl A*
+		 * Wa_18011464164:tgl[B0+],dg1[B0+]
+		 * Wa_22010931296:tgl[B0+],dg1[B0+]
+		 * Wa_14010919138:rkl,dg1,adl-s,adl-p
+		 */
+		wa_write_or(wal, GEN7_FF_THREAD_MODE,
+			    GEN12_FF_TESSELATION_DOP_GATE_DISABLE);
+	}
+
+	if (IS_ALDERLAKE_P(i915) || IS_DG2(i915) || IS_ALDERLAKE_S(i915) ||
+	    IS_DG1(i915) || IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915)) {
+		/*
+		 * Wa_1606700617:tgl,dg1,adl-p
+		 * Wa_22010271021:tgl,rkl,dg1,adl-s,adl-p
+		 * Wa_14010826681:tgl,dg1,rkl,adl-p
+		 * Wa_18019627453:dg2
+		 */
+		wa_masked_en(wal,
+			     GEN9_CS_DEBUG_MODE1,
+			     FF_DOP_CLOCK_GATE_DISABLE);
+	}
+
+	if (IS_ALDERLAKE_P(i915) || IS_ALDERLAKE_S(i915) ||
+	    IS_DG1_GRAPHICS_STEP(i915, STEP_A0, STEP_B0) ||
+	    IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915)) {
+		/* Wa_1409804808:tgl,rkl,dg1[a0],adl-s,adl-p */
+		wa_masked_en(wal, GEN7_ROW_CHICKEN2,
+			     GEN12_PUSH_CONST_DEREF_HOLD_DIS);
+
+		/*
+		 * Wa_1409085225:tgl
+		 * Wa_14010229206:tgl,rkl,dg1[a0],adl-s,adl-p
+		 */
+		wa_masked_en(wal, GEN9_ROW_CHICKEN4, GEN12_DISABLE_TDL_PUSH);
+	}
+
+	if (IS_DG1_GRAPHICS_STEP(i915, STEP_A0, STEP_B0) ||
+	    IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915) || IS_ALDERLAKE_P(i915)) {
+		/*
+		 * Wa_1607030317:tgl
+		 * Wa_1607186500:tgl
+		 * Wa_1607297627:tgl,rkl,dg1[a0],adlp
+		 *
+		 * On TGL and RKL there are multiple entries for this WA in the
+		 * BSpec; some indicate this is an A0-only WA, others indicate
+		 * it applies to all steppings so we trust the "all steppings."
+		 * For DG1 this only applies to A0.
+		 */
+		wa_masked_en(wal,
+			     RING_PSMI_CTL(RENDER_RING_BASE),
+			     GEN12_WAIT_FOR_EVENT_POWER_DOWN_DISABLE |
+			     GEN8_RC_SEMA_IDLE_MSG_DISABLE);
+	}
+
+	if (IS_DG1(i915) || IS_ROCKETLAKE(i915) || IS_TIGERLAKE(i915) ||
+	    IS_ALDERLAKE_S(i915) || IS_ALDERLAKE_P(i915)) {
+		/* Wa_1406941453:tgl,rkl,dg1,adl-s,adl-p */
+		wa_masked_en(wal,
+			     GEN10_SAMPLER_MODE,
+			     ENABLE_SMALLPL);
+	}
+
+	if (GRAPHICS_VER(i915) == 11) {
+		/* This is not an Wa. Enable for better image quality */
+		wa_masked_en(wal,
+			     _3D_CHICKEN3,
+			     _3D_CHICKEN3_AA_LINE_QUALITY_FIX_ENABLE);
+
+		/*
+		 * Wa_1405543622:icl
+		 * Formerly known as WaGAPZPriorityScheme
+		 */
+		wa_write_or(wal,
+			    GEN8_GARBCNTL,
+			    GEN11_ARBITRATION_PRIO_ORDER_MASK);
+
+		/*
+		 * Wa_1604223664:icl
+		 * Formerly known as WaL3BankAddressHashing
+		 */
+		wa_write_clr_set(wal,
+				 GEN8_GARBCNTL,
+				 GEN11_HASH_CTRL_EXCL_MASK,
+				 GEN11_HASH_CTRL_EXCL_BIT0);
+		wa_write_clr_set(wal,
+				 GEN11_GLBLINVL,
+				 GEN11_BANK_HASH_ADDR_EXCL_MASK,
+				 GEN11_BANK_HASH_ADDR_EXCL_BIT0);
+
+		/*
+		 * Wa_1405733216:icl
+		 * Formerly known as WaDisableCleanEvicts
+		 */
+		wa_write_or(wal,
+			    GEN8_L3SQCREG4,
+			    GEN11_LQSC_CLEAN_EVICT_DISABLE);
+
+		/* Wa_1606682166:icl */
+		wa_write_or(wal,
+			    GEN7_SARCHKMD,
+			    GEN7_DISABLE_SAMPLER_PREFETCH);
+
+		/* Wa_1409178092:icl */
+		wa_write_clr_set(wal,
+				 GEN11_SCRATCH2,
+				 GEN11_COHERENT_PARTIAL_WRITE_MERGE_ENABLE,
+				 0);
+
+		/* WaEnable32PlaneMode:icl */
+		wa_masked_en(wal, GEN9_CSFE_CHICKEN1_RCS,
+			     GEN11_ENABLE_32_PLANE_MODE);
+
+		/*
+		 * Wa_1408767742:icl[a2..forever],ehl[all]
+		 * Wa_1605460711:icl[a0..c0]
+		 */
+		wa_write_or(wal,
+			    GEN7_FF_THREAD_MODE,
+			    GEN12_FF_TESSELATION_DOP_GATE_DISABLE);
+
+		/* Wa_22010271021 */
+		wa_masked_en(wal,
+			     GEN9_CS_DEBUG_MODE1,
+			     FF_DOP_CLOCK_GATE_DISABLE);
+	}
+
+	if (IS_GRAPHICS_VER(i915, 9, 12)) {
+		/* FtrPerCtxtPreemptionGranularityControl:skl,bxt,kbl,cfl,cnl,icl,tgl */
+		wa_masked_en(wal,
+			     GEN7_FF_SLICE_CS_CHICKEN1,
+			     GEN9_FFSC_PERCTX_PREEMPT_CTRL);
+	}
+
+	if (IS_SKYLAKE(i915) ||
+	    IS_KABYLAKE(i915) ||
+	    IS_COFFEELAKE(i915) ||
+	    IS_COMETLAKE(i915)) {
+		/* WaEnableGapsTsvCreditFix:skl,kbl,cfl */
+		wa_write_or(wal,
+			    GEN8_GARBCNTL,
+			    GEN9_GAPS_TSV_CREDIT_DISABLE);
+	}
+
+	if (IS_BROXTON(i915)) {
+		/* WaDisablePooledEuLoadBalancingFix:bxt */
+		wa_masked_en(wal,
+			     FF_SLICE_CS_CHICKEN2,
+			     GEN9_POOLED_EU_LOAD_BALANCING_FIX_DISABLE);
+	}
+
+	if (GRAPHICS_VER(i915) == 9) {
+		/* WaContextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl,glk,cfl */
+		wa_masked_en(wal,
+			     GEN9_CSFE_CHICKEN1_RCS,
+			     GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE);
+
+		/* WaEnableLbsSlaRetryTimerDecrement:skl,bxt,kbl,glk,cfl */
+		wa_write_or(wal,
+			    BDW_SCRATCH1,
+			    GEN9_LBS_SLA_RETRY_TIMER_DECREMENT_ENABLE);
+
+		/* WaProgramL3SqcReg1DefaultForPerf:bxt,glk */
+		if (IS_GEN9_LP(i915))
+			wa_write_clr_set(wal,
+					 GEN8_L3SQCREG1,
+					 L3_PRIO_CREDITS_MASK,
+					 L3_GENERAL_PRIO_CREDITS(62) |
+					 L3_HIGH_PRIO_CREDITS(2));
+
+		/* WaOCLCoherentLineFlush:skl,bxt,kbl,cfl */
+		wa_write_or(wal,
+			    GEN8_L3SQCREG4,
+			    GEN8_LQSC_FLUSH_COHERENT_LINES);
+
+		/* Disable atomics in L3 to prevent unrecoverable hangs */
+		wa_write_clr_set(wal, GEN9_SCRATCH_LNCF1,
+				 GEN9_LNCF_NONIA_COHERENT_ATOMICS_ENABLE, 0);
+		wa_write_clr_set(wal, GEN8_L3SQCREG4,
+				 GEN8_LQSQ_NONIA_COHERENT_ATOMICS_ENABLE, 0);
+		wa_write_clr_set(wal, GEN9_SCRATCH1,
+				 EVICTION_PERF_FIX_ENABLE, 0);
+	}
+
+	if (IS_HASWELL(i915)) {
+		/* WaSampleCChickenBitEnable:hsw */
+		wa_masked_en(wal,
+			     HALF_SLICE_CHICKEN3, HSW_SAMPLE_C_PERFORMANCE);
+
+		wa_masked_dis(wal,
+			      CACHE_MODE_0_GEN7,
+			      /* enable HiZ Raw Stall Optimization */
+			      HIZ_RAW_STALL_OPT_DISABLE);
+	}
+
+	if (IS_VALLEYVIEW(i915)) {
+		/* WaDisableEarlyCull:vlv */
+		wa_masked_en(wal,
+			     _3D_CHICKEN3,
+			     _3D_CHICKEN_SF_DISABLE_OBJEND_CULL);
+
+		/*
+		 * WaVSThreadDispatchOverride:ivb,vlv
+		 *
+		 * This actually overrides the dispatch
+		 * mode for all thread types.
+		 */
+		wa_write_clr_set(wal,
+				 GEN7_FF_THREAD_MODE,
+				 GEN7_FF_SCHED_MASK,
+				 GEN7_FF_TS_SCHED_HW |
+				 GEN7_FF_VS_SCHED_HW |
+				 GEN7_FF_DS_SCHED_HW);
+
+		/* WaPsdDispatchEnable:vlv */
+		/* WaDisablePSDDualDispatchEnable:vlv */
+		wa_masked_en(wal,
+			     GEN7_HALF_SLICE_CHICKEN1,
+			     GEN7_MAX_PS_THREAD_DEP |
+			     GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE);
+	}
+
+	if (IS_IVYBRIDGE(i915)) {
+		/* WaDisableEarlyCull:ivb */
+		wa_masked_en(wal,
+			     _3D_CHICKEN3,
+			     _3D_CHICKEN_SF_DISABLE_OBJEND_CULL);
+
+		if (0) { /* causes HiZ corruption on ivb:gt1 */
+			/* enable HiZ Raw Stall Optimization */
+			wa_masked_dis(wal,
+				      CACHE_MODE_0_GEN7,
+				      HIZ_RAW_STALL_OPT_DISABLE);
+		}
+
+		/*
+		 * WaVSThreadDispatchOverride:ivb,vlv
+		 *
+		 * This actually overrides the dispatch
+		 * mode for all thread types.
+		 */
+		wa_write_clr_set(wal,
+				 GEN7_FF_THREAD_MODE,
+				 GEN7_FF_SCHED_MASK,
+				 GEN7_FF_TS_SCHED_HW |
+				 GEN7_FF_VS_SCHED_HW |
+				 GEN7_FF_DS_SCHED_HW);
+
+		/* WaDisablePSDDualDispatchEnable:ivb */
+		if (IS_IVB_GT1(i915))
+			wa_masked_en(wal,
+				     GEN7_HALF_SLICE_CHICKEN1,
+				     GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE);
+	}
+
+	if (GRAPHICS_VER(i915) == 7) {
+		/* WaBCSVCSTlbInvalidationMode:ivb,vlv,hsw */
+		wa_masked_en(wal,
+			     RING_MODE_GEN7(RENDER_RING_BASE),
+			     GFX_TLB_INVALIDATE_EXPLICIT | GFX_REPLAY_MODE);
+
+		/* WaDisable_RenderCache_OperationalFlush:ivb,vlv,hsw */
+		wa_masked_dis(wal, CACHE_MODE_0_GEN7, RC_OP_FLUSH_ENABLE);
+
+		/*
+		 * BSpec says this must be set, even though
+		 * WaDisable4x2SubspanOptimization:ivb,hsw
+		 * WaDisable4x2SubspanOptimization isn't listed for VLV.
+		 */
+		wa_masked_en(wal,
+			     CACHE_MODE_1,
+			     PIXEL_SUBSPAN_COLLECT_OPT_DISABLE);
+
+		/*
+		 * BSpec recommends 8x4 when MSAA is used,
+		 * however in practice 16x4 seems fastest.
+		 *
+		 * Note that PS/WM thread counts depend on the WIZ hashing
+		 * disable bit, which we don't touch here, but it's good
+		 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
+		 */
+		wa_masked_field_set(wal,
+				    GEN7_GT_MODE,
+				    GEN6_WIZ_HASHING_MASK,
+				    GEN6_WIZ_HASHING_16x4);
+	}
+
+	if (IS_GRAPHICS_VER(i915, 6, 7))
+		/*
+		 * We need to disable the AsyncFlip performance optimisations in
+		 * order to use MI_WAIT_FOR_EVENT within the CS. It should
+		 * already be programmed to '1' on all products.
+		 *
+		 * WaDisableAsyncFlipPerfMode:snb,ivb,hsw,vlv
+		 */
+		wa_masked_en(wal,
+			     RING_MI_MODE(RENDER_RING_BASE),
+			     ASYNC_FLIP_PERF_DISABLE);
+
+	if (GRAPHICS_VER(i915) == 6) {
+		/*
+		 * Required for the hardware to program scanline values for
+		 * waiting
+		 * WaEnableFlushTlbInvalidationMode:snb
+		 */
+		wa_masked_en(wal,
+			     GFX_MODE,
+			     GFX_TLB_INVALIDATE_EXPLICIT);
+
+		/* WaDisableHiZPlanesWhenMSAAEnabled:snb */
+		wa_masked_en(wal,
+			     _3D_CHICKEN,
+			     _3D_CHICKEN_HIZ_PLANE_DISABLE_MSAA_4X_SNB);
+
+		wa_masked_en(wal,
+			     _3D_CHICKEN3,
+			     /* WaStripsFansDisableFastClipPerformanceFix:snb */
+			     _3D_CHICKEN3_SF_DISABLE_FASTCLIP_CULL |
+			     /*
+			      * Bspec says:
+			      * "This bit must be set if 3DSTATE_CLIP clip mode is set
+			      * to normal and 3DSTATE_SF number of SF output attributes
+			      * is more than 16."
+			      */
+			     _3D_CHICKEN3_SF_DISABLE_PIPELINED_ATTR_FETCH);
+
+		/*
+		 * BSpec recommends 8x4 when MSAA is used,
+		 * however in practice 16x4 seems fastest.
+		 *
+		 * Note that PS/WM thread counts depend on the WIZ hashing
+		 * disable bit, which we don't touch here, but it's good
+		 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
+		 */
+		wa_masked_field_set(wal,
+				    GEN6_GT_MODE,
+				    GEN6_WIZ_HASHING_MASK,
+				    GEN6_WIZ_HASHING_16x4);
+
+		/* WaDisable_RenderCache_OperationalFlush:snb */
+		wa_masked_dis(wal, CACHE_MODE_0, RC_OP_FLUSH_ENABLE);
+
+		/*
+		 * From the Sandybridge PRM, volume 1 part 3, page 24:
+		 * "If this bit is set, STCunit will have LRA as replacement
+		 *  policy. [...] This bit must be reset. LRA replacement
+		 *  policy is not supported."
+		 */
+		wa_masked_dis(wal,
+			      CACHE_MODE_0,
+			      CM0_STC_EVICT_DISABLE_LRA_SNB);
+	}
+
+	if (IS_GRAPHICS_VER(i915, 4, 6))
+		/* WaTimedSingleVertexDispatch:cl,bw,ctg,elk,ilk,snb */
+		wa_add(wal, RING_MI_MODE(RENDER_RING_BASE),
+		       0, _MASKED_BIT_ENABLE(VS_TIMER_DISPATCH),
+		       /* XXX bit doesn't stick on Broadwater */
+		       IS_I965G(i915) ? 0 : VS_TIMER_DISPATCH, true);
+
+	if (GRAPHICS_VER(i915) == 4)
+		/*
+		 * Disable CONSTANT_BUFFER before it is loaded from the context
+		 * image. For as it is loaded, it is executed and the stored
+		 * address may no longer be valid, leading to a GPU hang.
+		 *
+		 * This imposes the requirement that userspace reload their
+		 * CONSTANT_BUFFER on every batch, fortunately a requirement
+		 * they are already accustomed to from before contexts were
+		 * enabled.
+		 */
+		wa_add(wal, ECOSKPD(RENDER_RING_BASE),
+		       0, _MASKED_BIT_ENABLE(ECO_CONSTANT_BUFFER_SR_DISABLE),
+		       0 /* XXX bit doesn't stick on Broadwater */,
+		       true);
+}
+
+static void
+xcs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	/* WaKBLVECSSemaphoreWaitPoll:kbl */
+	if (IS_KBL_GRAPHICS_STEP(i915, STEP_A0, STEP_F0)) {
+		wa_write(wal,
+			 RING_SEMA_WAIT_POLL(engine->mmio_base),
+			 1);
+	}
+}
+
+static void
+ccs_engine_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
+{
+	if (IS_PVC_CT_STEP(engine->i915, STEP_A0, STEP_C0)) {
+		/* Wa_14014999345:pvc */
+		wa_masked_en(wal, GEN10_CACHE_MODE_SS, DISABLE_ECC);
+	}
+}
+
+/*
+ * The bspec performance guide has recommended MMIO tuning settings.  These
+ * aren't truly "workarounds" but we want to program them with the same
+ * workaround infrastructure to ensure that they're automatically added to
+ * the GuC save/restore lists, re-applied at the right times, and checked for
+ * any conflicting programming requested by real workarounds.
+ *
+ * Programming settings should be added here only if their registers are not
+ * part of an engine's register state context.  If a register is part of a
+ * context, then any tuning settings should be programmed in an appropriate
+ * function invoked by __intel_engine_init_ctx_wa().
+ */
+static void
+add_render_compute_tuning_settings(struct drm_i915_private *i915,
+				   struct i915_wa_list *wal)
+{
+	if (IS_PONTEVECCHIO(i915)) {
+		wa_write(wal, XEHPC_L3SCRUB,
+			 SCRUB_CL_DWNGRADE_SHARED | SCRUB_RATE_4B_PER_CLK);
+	}
+
+	if (IS_DG2(i915)) {
+		wa_write_or(wal, XEHP_L3SCQREG7, BLEND_FILL_CACHING_OPT_DIS);
+		wa_write_clr_set(wal, RT_CTRL, STACKID_CTRL, STACKID_CTRL_512);
+
+		/*
+		 * This is also listed as Wa_22012654132 for certain DG2
+		 * steppings, but the tuning setting programming is a superset
+		 * since it applies to all DG2 variants and steppings.
+		 *
+		 * Note that register 0xE420 is write-only and cannot be read
+		 * back for verification on DG2 (due to Wa_14012342262), so
+		 * we need to explicitly skip the readback.
+		 */
+		wa_add(wal, GEN10_CACHE_MODE_SS, 0,
+		       _MASKED_BIT_ENABLE(ENABLE_PREFETCH_INTO_IC),
+		       0 /* write-only, so skip validation */,
+		       true);
+	}
+
+	/*
+	 * This tuning setting proves beneficial only on ATS-M designs; the
+	 * default "age based" setting is optimal on regular DG2 and other
+	 * platforms.
+	 */
+	if (INTEL_INFO(i915)->tuning_thread_rr_after_dep)
+		wa_masked_field_set(wal, GEN9_ROW_CHICKEN4, THREAD_EX_ARB_MODE,
+				    THREAD_EX_ARB_MODE_RR_AFTER_DEP);
+}
+
+/*
+ * The workarounds in this function apply to shared registers in
+ * the general render reset domain that aren't tied to a
+ * specific engine.  Since all render+compute engines get reset
+ * together, and the contents of these registers are lost during
+ * the shared render domain reset, we'll define such workarounds
+ * here and then add them to just a single RCS or CCS engine's
+ * workaround list (whichever engine has the XXXX flag).
+ */
+static void
+general_render_compute_wa_init(struct intel_engine_cs *engine, struct i915_wa_list *wal)
+{
+	struct drm_i915_private *i915 = engine->i915;
+
+	add_render_compute_tuning_settings(i915, wal);
+
+	if (IS_PONTEVECCHIO(i915)) {
+		/* Wa_16016694945 */
+		wa_masked_en(wal, XEHPC_LNCFMISCCFGREG0, XEHPC_OVRLSCCC);
+	}
+
+	if (IS_XEHPSDV(i915)) {
+		/* Wa_1409954639 */
+		wa_masked_en(wal,
+			     GEN8_ROW_CHICKEN,
+			     SYSTOLIC_DOP_CLOCK_GATING_DIS);
+
+		/* Wa_1607196519 */
+		wa_masked_en(wal,
+			     GEN9_ROW_CHICKEN4,
+			     GEN12_DISABLE_GRF_CLEAR);
+
+		/* Wa_14010670810:xehpsdv */
+		wa_write_or(wal, XEHP_L3NODEARBCFG, XEHP_LNESPARE);
+
+		/* Wa_14010449647:xehpsdv */
+		wa_masked_en(wal, GEN7_HALF_SLICE_CHICKEN1,
+			     GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE);
+
+		/* Wa_18011725039:xehpsdv */
+		if (IS_XEHPSDV_GRAPHICS_STEP(i915, STEP_A1, STEP_B0)) {
+			wa_masked_dis(wal, MLTICTXCTL, TDONRENDER);
+			wa_write_or(wal, L3SQCREG1_CCS0, FLUSHALLNONCOH);
+		}
+
+		/* Wa_14012362059:xehpsdv */
+		wa_write_or(wal, GEN12_MERT_MOD_CTRL, FORCE_MISS_FTLB);
+
+		/* Wa_14014368820:xehpsdv */
+		wa_write_or(wal, GEN12_GAMCNTRL_CTRL, INVALIDATION_BROADCAST_MODE_DIS |
+				GLOBAL_INVALIDATION_MODE);
+	}
+
+	if (IS_DG2(i915) || IS_PONTEVECCHIO(i915)) {
+		/* Wa_14015227452:dg2,pvc */
+		wa_masked_en(wal, GEN9_ROW_CHICKEN4, XEHP_DIS_BBL_SYSPIPE);
+
+		/* Wa_22014226127:dg2,pvc */
+		wa_write_or(wal, LSC_CHICKEN_BIT_0, DISABLE_D8_D16_COASLESCE);
+
+		/* Wa_16015675438:dg2,pvc */
+		wa_masked_en(wal, FF_SLICE_CS_CHICKEN2, GEN12_PERF_FIX_BALANCING_CFE_DISABLE);
+
+		/* Wa_18018781329:dg2,pvc */
+		wa_write_or(wal, RENDER_MOD_CTRL, FORCE_MISS_FTLB);
+		wa_write_or(wal, COMP_MOD_CTRL, FORCE_MISS_FTLB);
+		wa_write_or(wal, VDBX_MOD_CTRL, FORCE_MISS_FTLB);
+		wa_write_or(wal, VEBX_MOD_CTRL, FORCE_MISS_FTLB);
+	}
+}
+
+static void
+engine_init_workarounds(struct intel_engine_cs *engine, struct i915_wa_list *wal)
+{
+	if (I915_SELFTEST_ONLY(GRAPHICS_VER(engine->i915) < 4))
+		return;
+
+	engine_fake_wa_init(engine, wal);
+
+	/*
+	 * These are common workarounds that just need to applied
+	 * to a single RCS/CCS engine's workaround list since
+	 * they're reset as part of the general render domain reset.
+	 */
+	if (engine->flags & I915_ENGINE_FIRST_RENDER_COMPUTE)
+		general_render_compute_wa_init(engine, wal);
+
+	if (engine->class == COMPUTE_CLASS)
+		ccs_engine_wa_init(engine, wal);
+	else if (engine->class == RENDER_CLASS)
+		rcs_engine_wa_init(engine, wal);
+	else
+		xcs_engine_wa_init(engine, wal);
+}
+
+void intel_engine_init_workarounds(struct intel_engine_cs *engine)
+{
+	struct i915_wa_list *wal = &engine->wa_list;
+
+	if (GRAPHICS_VER(engine->i915) < 4)
+		return;
+
+	wa_init_start(wal, "engine", engine->name);
+	engine_init_workarounds(engine, wal);
+	wa_init_finish(wal);
+}
+
+void intel_engine_apply_workarounds(struct intel_engine_cs *engine)
+{
+	wa_list_apply(engine->gt, &engine->wa_list);
+}
+
+static const struct i915_range mcr_ranges_gen8[] = {
+	{ .start = 0x5500, .end = 0x55ff },
+	{ .start = 0x7000, .end = 0x7fff },
+	{ .start = 0x9400, .end = 0x97ff },
+	{ .start = 0xb000, .end = 0xb3ff },
+	{ .start = 0xe000, .end = 0xe7ff },
+	{},
+};
+
+static const struct i915_range mcr_ranges_gen12[] = {
+	{ .start =  0x8150, .end =  0x815f },
+	{ .start =  0x9520, .end =  0x955f },
+	{ .start =  0xb100, .end =  0xb3ff },
+	{ .start =  0xde80, .end =  0xe8ff },
+	{ .start = 0x24a00, .end = 0x24a7f },
+	{},
+};
+
+static const struct i915_range mcr_ranges_xehp[] = {
+	{ .start =  0x4000, .end =  0x4aff },
+	{ .start =  0x5200, .end =  0x52ff },
+	{ .start =  0x5400, .end =  0x7fff },
+	{ .start =  0x8140, .end =  0x815f },
+	{ .start =  0x8c80, .end =  0x8dff },
+	{ .start =  0x94d0, .end =  0x955f },
+	{ .start =  0x9680, .end =  0x96ff },
+	{ .start =  0xb000, .end =  0xb3ff },
+	{ .start =  0xc800, .end =  0xcfff },
+	{ .start =  0xd800, .end =  0xd8ff },
+	{ .start =  0xdc00, .end =  0xffff },
+	{ .start = 0x17000, .end = 0x17fff },
+	{ .start = 0x24a00, .end = 0x24a7f },
+	{},
+};
+
+static bool mcr_range(struct drm_i915_private *i915, u32 offset)
+{
+	const struct i915_range *mcr_ranges;
+	int i;
+
+	if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50))
+		mcr_ranges = mcr_ranges_xehp;
+	else if (GRAPHICS_VER(i915) >= 12)
+		mcr_ranges = mcr_ranges_gen12;
+	else if (GRAPHICS_VER(i915) >= 8)
+		mcr_ranges = mcr_ranges_gen8;
+	else
+		return false;
+
+	/*
+	 * Registers in these ranges are affected by the MCR selector
+	 * which only controls CPU initiated MMIO. Routing does not
+	 * work for CS access so we cannot verify them on this path.
+	 */
+	for (i = 0; mcr_ranges[i].start; i++)
+		if (offset >= mcr_ranges[i].start &&
+		    offset <= mcr_ranges[i].end)
+			return true;
+
+	return false;
+}
+
+static int
+wa_list_srm(struct i915_request *rq,
+	    const struct i915_wa_list *wal,
+	    struct i915_vma *vma)
+{
+	struct drm_i915_private *i915 = rq->engine->i915;
+	unsigned int i, count = 0;
+	const struct i915_wa *wa;
+	u32 srm, *cs;
+
+	srm = MI_STORE_REGISTER_MEM | MI_SRM_LRM_GLOBAL_GTT;
+	if (GRAPHICS_VER(i915) >= 8)
+		srm++;
+
+	for (i = 0, wa = wal->list; i < wal->count; i++, wa++) {
+		if (!mcr_range(i915, i915_mmio_reg_offset(wa->reg)))
+			count++;
+	}
+
+	cs = intel_ring_begin(rq, 4 * count);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	for (i = 0, wa = wal->list; i < wal->count; i++, wa++) {
+		u32 offset = i915_mmio_reg_offset(wa->reg);
+
+		if (mcr_range(i915, offset))
+			continue;
+
+		*cs++ = srm;
+		*cs++ = offset;
+		*cs++ = i915_ggtt_offset(vma) + sizeof(u32) * i;
+		*cs++ = 0;
+	}
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int engine_wa_list_verify(struct intel_context *ce,
+				 const struct i915_wa_list * const wal,
+				 const char *from)
+{
+	const struct i915_wa *wa;
+	struct i915_request *rq;
+	struct i915_vma *vma;
+	struct i915_gem_ww_ctx ww;
+	unsigned int i;
+	u32 *results;
+	int err;
+
+	if (!wal->count)
+		return 0;
+
+	vma = __vm_create_scratch_for_read(&ce->engine->gt->ggtt->vm,
+					   wal->count * sizeof(u32));
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	intel_engine_pm_get(ce->engine);
+	i915_gem_ww_ctx_init(&ww, false);
+retry:
+	err = i915_gem_object_lock(vma->obj, &ww);
+	if (err == 0)
+		err = intel_context_pin_ww(ce, &ww);
+	if (err)
+		goto err_pm;
+
+	err = i915_vma_pin_ww(vma, &ww, 0, 0,
+			   i915_vma_is_ggtt(vma) ? PIN_GLOBAL : PIN_USER);
+	if (err)
+		goto err_unpin;
+
+	rq = i915_request_create(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_vma;
+	}
+
+	err = i915_request_await_object(rq, vma->obj, true);
+	if (err == 0)
+		err = i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE);
+	if (err == 0)
+		err = wa_list_srm(rq, wal, vma);
+
+	i915_request_get(rq);
+	if (err)
+		i915_request_set_error_once(rq, err);
+	i915_request_add(rq);
+
+	if (err)
+		goto err_rq;
+
+	if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+		err = -ETIME;
+		goto err_rq;
+	}
+
+	results = i915_gem_object_pin_map(vma->obj, I915_MAP_WB);
+	if (IS_ERR(results)) {
+		err = PTR_ERR(results);
+		goto err_rq;
+	}
+
+	err = 0;
+	for (i = 0, wa = wal->list; i < wal->count; i++, wa++) {
+		if (mcr_range(rq->engine->i915, i915_mmio_reg_offset(wa->reg)))
+			continue;
+
+		if (!wa_verify(wa, results[i], wal->name, from))
+			err = -ENXIO;
+	}
+
+	i915_gem_object_unpin_map(vma->obj);
+
+err_rq:
+	i915_request_put(rq);
+err_vma:
+	i915_vma_unpin(vma);
+err_unpin:
+	intel_context_unpin(ce);
+err_pm:
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	intel_engine_pm_put(ce->engine);
+	i915_vma_put(vma);
+	return err;
+}
+
+int intel_engine_verify_workarounds(struct intel_engine_cs *engine,
+				    const char *from)
+{
+	return engine_wa_list_verify(engine->kernel_context,
+				     &engine->wa_list,
+				     from);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_workarounds.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.h b/drivers/gpu/drm/i915/gt/intel_workarounds.h
new file mode 100644
index 000000000..9beaab77c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_workarounds.h
@@ -0,0 +1,39 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2018 Intel Corporation
+ */
+
+#ifndef _INTEL_WORKAROUNDS_H_
+#define _INTEL_WORKAROUNDS_H_
+
+#include <linux/slab.h>
+
+#include "intel_workarounds_types.h"
+
+struct drm_i915_private;
+struct i915_request;
+struct intel_engine_cs;
+struct intel_gt;
+
+static inline void intel_wa_list_free(struct i915_wa_list *wal)
+{
+	kfree(wal->list);
+	memset(wal, 0, sizeof(*wal));
+}
+
+void intel_engine_init_ctx_wa(struct intel_engine_cs *engine);
+int intel_engine_emit_ctx_wa(struct i915_request *rq);
+
+void intel_gt_init_workarounds(struct intel_gt *gt);
+void intel_gt_apply_workarounds(struct intel_gt *gt);
+bool intel_gt_verify_workarounds(struct intel_gt *gt, const char *from);
+
+void intel_engine_init_whitelist(struct intel_engine_cs *engine);
+void intel_engine_apply_whitelist(struct intel_engine_cs *engine);
+
+void intel_engine_init_workarounds(struct intel_engine_cs *engine);
+void intel_engine_apply_workarounds(struct intel_engine_cs *engine);
+int intel_engine_verify_workarounds(struct intel_engine_cs *engine,
+				    const char *from);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds_types.h b/drivers/gpu/drm/i915/gt/intel_workarounds_types.h
new file mode 100644
index 000000000..8a4b6de4e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/intel_workarounds_types.h
@@ -0,0 +1,29 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2018 Intel Corporation
+ */
+
+#ifndef __INTEL_WORKAROUNDS_TYPES_H__
+#define __INTEL_WORKAROUNDS_TYPES_H__
+
+#include <linux/types.h>
+
+#include "i915_reg_defs.h"
+
+struct i915_wa {
+	i915_reg_t	reg;
+	u32		clr;
+	u32		set;
+	u32		read;
+	bool		masked_reg;
+};
+
+struct i915_wa_list {
+	const char	*name;
+	const char	*engine_name;
+	struct i915_wa	*list;
+	unsigned int	count;
+	unsigned int	wa_count;
+};
+
+#endif /* __INTEL_WORKAROUNDS_TYPES_H__ */
diff --git a/drivers/gpu/drm/i915/gt/ivb_clear_kernel.c b/drivers/gpu/drm/i915/gt/ivb_clear_kernel.c
new file mode 100644
index 000000000..610ca7687
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/ivb_clear_kernel.c
@@ -0,0 +1,61 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ *
+ * Generated by: IGT Gpu Tools on Fri 21 Feb 2020 05:29:32 AM UTC
+ */
+
+static const u32 ivb_clear_kernel[] = {
+	0x00000001, 0x26020128, 0x00000024, 0x00000000,
+	0x00000040, 0x20280c21, 0x00000028, 0x00000001,
+	0x01000010, 0x20000c20, 0x0000002c, 0x00000000,
+	0x00010220, 0x34001c00, 0x00001400, 0x0000002c,
+	0x00600001, 0x20600061, 0x00000000, 0x00000000,
+	0x00000008, 0x20601c85, 0x00000e00, 0x0000000c,
+	0x00000005, 0x20601ca5, 0x00000060, 0x00000001,
+	0x00000008, 0x20641c85, 0x00000e00, 0x0000000d,
+	0x00000005, 0x20641ca5, 0x00000064, 0x00000003,
+	0x00000041, 0x207424a5, 0x00000064, 0x00000034,
+	0x00000040, 0x206014a5, 0x00000060, 0x00000074,
+	0x00000008, 0x20681c85, 0x00000e00, 0x00000008,
+	0x00000005, 0x20681ca5, 0x00000068, 0x0000000f,
+	0x00000041, 0x20701ca5, 0x00000060, 0x00000010,
+	0x00000040, 0x206814a5, 0x00000068, 0x00000070,
+	0x00600001, 0x20a00061, 0x00000000, 0x00000000,
+	0x00000005, 0x206c1c85, 0x00000e00, 0x00000007,
+	0x00000041, 0x206c1ca5, 0x0000006c, 0x00000004,
+	0x00600001, 0x20800021, 0x008d0000, 0x00000000,
+	0x00000001, 0x20800021, 0x0000006c, 0x00000000,
+	0x00000001, 0x20840021, 0x00000068, 0x00000000,
+	0x00000001, 0x20880061, 0x00000000, 0x00000003,
+	0x00000005, 0x208c0d21, 0x00000086, 0xffffffff,
+	0x05600032, 0x20a00fa1, 0x008d0080, 0x02190001,
+	0x00000040, 0x20a01ca5, 0x000000a0, 0x00000001,
+	0x05600032, 0x20a00fa1, 0x008d0080, 0x040a8001,
+	0x02000040, 0x20281c21, 0x00000028, 0xffffffff,
+	0x00010220, 0x34001c00, 0x00001400, 0xfffffffc,
+	0x00000001, 0x26020128, 0x00000024, 0x00000000,
+	0x00000001, 0x220010e4, 0x00000000, 0x00000000,
+	0x00000001, 0x220831ec, 0x00000000, 0x007f007f,
+	0x00600001, 0x20400021, 0x008d0000, 0x00000000,
+	0x00600001, 0x2fe00021, 0x008d0000, 0x00000000,
+	0x00200001, 0x20400121, 0x00450020, 0x00000000,
+	0x00000001, 0x20480061, 0x00000000, 0x000f000f,
+	0x00000005, 0x204c0d21, 0x00000046, 0xffffffef,
+	0x00800001, 0x20600061, 0x00000000, 0x00000000,
+	0x00800001, 0x20800061, 0x00000000, 0x00000000,
+	0x00800001, 0x20a00061, 0x00000000, 0x00000000,
+	0x00800001, 0x20c00061, 0x00000000, 0x00000000,
+	0x00800001, 0x20e00061, 0x00000000, 0x00000000,
+	0x00800001, 0x21000061, 0x00000000, 0x00000000,
+	0x00800001, 0x21200061, 0x00000000, 0x00000000,
+	0x00800001, 0x21400061, 0x00000000, 0x00000000,
+	0x05600032, 0x20000fa0, 0x008d0040, 0x120a8000,
+	0x00000040, 0x20402d21, 0x00000020, 0x00100010,
+	0x05600032, 0x20000fa0, 0x008d0040, 0x120a8000,
+	0x02000040, 0x22083d8c, 0x00000208, 0xffffffff,
+	0x00800001, 0xa0000109, 0x00000602, 0x00000000,
+	0x00000040, 0x22001c84, 0x00000200, 0x00000020,
+	0x00010220, 0x34001c00, 0x00001400, 0xfffffff8,
+	0x07600032, 0x20000fa0, 0x008d0fe0, 0x82000010,
+};
diff --git a/drivers/gpu/drm/i915/gt/mock_engine.c b/drivers/gpu/drm/i915/gt/mock_engine.c
new file mode 100644
index 000000000..c0637bf79
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/mock_engine.c
@@ -0,0 +1,445 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include "gem/i915_gem_context.h"
+#include "gt/intel_ring.h"
+
+#include "i915_drv.h"
+#include "intel_context.h"
+#include "intel_engine_pm.h"
+
+#include "mock_engine.h"
+#include "selftests/mock_request.h"
+
+static int mock_timeline_pin(struct intel_timeline *tl)
+{
+	int err;
+
+	if (WARN_ON(!i915_gem_object_trylock(tl->hwsp_ggtt->obj, NULL)))
+		return -EBUSY;
+
+	err = intel_timeline_pin_map(tl);
+	i915_gem_object_unlock(tl->hwsp_ggtt->obj);
+	if (err)
+		return err;
+
+	atomic_inc(&tl->pin_count);
+	return 0;
+}
+
+static void mock_timeline_unpin(struct intel_timeline *tl)
+{
+	GEM_BUG_ON(!atomic_read(&tl->pin_count));
+	atomic_dec(&tl->pin_count);
+}
+
+static struct i915_vma *create_ring_vma(struct i915_ggtt *ggtt, int size)
+{
+	struct i915_address_space *vm = &ggtt->vm;
+	struct drm_i915_private *i915 = vm->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+
+	obj = i915_gem_object_create_internal(i915, size);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma))
+		goto err;
+
+	return vma;
+
+err:
+	i915_gem_object_put(obj);
+	return vma;
+}
+
+static struct intel_ring *mock_ring(struct intel_engine_cs *engine)
+{
+	const unsigned long sz = PAGE_SIZE;
+	struct intel_ring *ring;
+
+	ring = kzalloc(sizeof(*ring) + sz, GFP_KERNEL);
+	if (!ring)
+		return NULL;
+
+	kref_init(&ring->ref);
+	ring->size = sz;
+	ring->effective_size = sz;
+	ring->vaddr = (void *)(ring + 1);
+	atomic_set(&ring->pin_count, 1);
+
+	ring->vma = create_ring_vma(engine->gt->ggtt, PAGE_SIZE);
+	if (IS_ERR(ring->vma)) {
+		kfree(ring);
+		return NULL;
+	}
+
+	intel_ring_update_space(ring);
+
+	return ring;
+}
+
+static void mock_ring_free(struct intel_ring *ring)
+{
+	i915_vma_put(ring->vma);
+
+	kfree(ring);
+}
+
+static struct i915_request *first_request(struct mock_engine *engine)
+{
+	return list_first_entry_or_null(&engine->hw_queue,
+					struct i915_request,
+					mock.link);
+}
+
+static void advance(struct i915_request *request)
+{
+	list_del_init(&request->mock.link);
+	i915_request_mark_complete(request);
+	GEM_BUG_ON(!i915_request_completed(request));
+
+	intel_engine_signal_breadcrumbs(request->engine);
+}
+
+static void hw_delay_complete(struct timer_list *t)
+{
+	struct mock_engine *engine = from_timer(engine, t, hw_delay);
+	struct i915_request *request;
+	unsigned long flags;
+
+	spin_lock_irqsave(&engine->hw_lock, flags);
+
+	/* Timer fired, first request is complete */
+	request = first_request(engine);
+	if (request)
+		advance(request);
+
+	/*
+	 * Also immediately signal any subsequent 0-delay requests, but
+	 * requeue the timer for the next delayed request.
+	 */
+	while ((request = first_request(engine))) {
+		if (request->mock.delay) {
+			mod_timer(&engine->hw_delay,
+				  jiffies + request->mock.delay);
+			break;
+		}
+
+		advance(request);
+	}
+
+	spin_unlock_irqrestore(&engine->hw_lock, flags);
+}
+
+static void mock_context_unpin(struct intel_context *ce)
+{
+}
+
+static void mock_context_post_unpin(struct intel_context *ce)
+{
+	i915_vma_unpin(ce->ring->vma);
+}
+
+static void mock_context_destroy(struct kref *ref)
+{
+	struct intel_context *ce = container_of(ref, typeof(*ce), ref);
+
+	GEM_BUG_ON(intel_context_is_pinned(ce));
+
+	if (test_bit(CONTEXT_ALLOC_BIT, &ce->flags)) {
+		mock_ring_free(ce->ring);
+		mock_timeline_unpin(ce->timeline);
+	}
+
+	intel_context_fini(ce);
+	intel_context_free(ce);
+}
+
+static int mock_context_alloc(struct intel_context *ce)
+{
+	int err;
+
+	ce->ring = mock_ring(ce->engine);
+	if (!ce->ring)
+		return -ENOMEM;
+
+	ce->timeline = intel_timeline_create(ce->engine->gt);
+	if (IS_ERR(ce->timeline)) {
+		kfree(ce->engine);
+		return PTR_ERR(ce->timeline);
+	}
+
+	err = mock_timeline_pin(ce->timeline);
+	if (err) {
+		intel_timeline_put(ce->timeline);
+		ce->timeline = NULL;
+		return err;
+	}
+
+	return 0;
+}
+
+static int mock_context_pre_pin(struct intel_context *ce,
+				struct i915_gem_ww_ctx *ww, void **unused)
+{
+	return i915_vma_pin_ww(ce->ring->vma, ww, 0, 0, PIN_GLOBAL | PIN_HIGH);
+}
+
+static int mock_context_pin(struct intel_context *ce, void *unused)
+{
+	return 0;
+}
+
+static void mock_context_reset(struct intel_context *ce)
+{
+}
+
+static const struct intel_context_ops mock_context_ops = {
+	.alloc = mock_context_alloc,
+
+	.pre_pin = mock_context_pre_pin,
+	.pin = mock_context_pin,
+	.unpin = mock_context_unpin,
+	.post_unpin = mock_context_post_unpin,
+
+	.enter = intel_context_enter_engine,
+	.exit = intel_context_exit_engine,
+
+	.reset = mock_context_reset,
+	.destroy = mock_context_destroy,
+};
+
+static int mock_request_alloc(struct i915_request *request)
+{
+	INIT_LIST_HEAD(&request->mock.link);
+	request->mock.delay = 0;
+
+	return 0;
+}
+
+static int mock_emit_flush(struct i915_request *request,
+			   unsigned int flags)
+{
+	return 0;
+}
+
+static u32 *mock_emit_breadcrumb(struct i915_request *request, u32 *cs)
+{
+	return cs;
+}
+
+static void mock_submit_request(struct i915_request *request)
+{
+	struct mock_engine *engine =
+		container_of(request->engine, typeof(*engine), base);
+	unsigned long flags;
+
+	i915_request_submit(request);
+
+	spin_lock_irqsave(&engine->hw_lock, flags);
+	list_add_tail(&request->mock.link, &engine->hw_queue);
+	if (list_is_first(&request->mock.link, &engine->hw_queue)) {
+		if (request->mock.delay)
+			mod_timer(&engine->hw_delay,
+				  jiffies + request->mock.delay);
+		else
+			advance(request);
+	}
+	spin_unlock_irqrestore(&engine->hw_lock, flags);
+}
+
+static void mock_add_to_engine(struct i915_request *rq)
+{
+	lockdep_assert_held(&rq->engine->sched_engine->lock);
+	list_move_tail(&rq->sched.link, &rq->engine->sched_engine->requests);
+}
+
+static void mock_remove_from_engine(struct i915_request *rq)
+{
+	struct intel_engine_cs *engine, *locked;
+
+	/*
+	 * Virtual engines complicate acquiring the engine timeline lock,
+	 * as their rq->engine pointer is not stable until under that
+	 * engine lock. The simple ploy we use is to take the lock then
+	 * check that the rq still belongs to the newly locked engine.
+	 */
+
+	locked = READ_ONCE(rq->engine);
+	spin_lock_irq(&locked->sched_engine->lock);
+	while (unlikely(locked != (engine = READ_ONCE(rq->engine)))) {
+		spin_unlock(&locked->sched_engine->lock);
+		spin_lock(&engine->sched_engine->lock);
+		locked = engine;
+	}
+	list_del_init(&rq->sched.link);
+	spin_unlock_irq(&locked->sched_engine->lock);
+}
+
+static void mock_reset_prepare(struct intel_engine_cs *engine)
+{
+}
+
+static void mock_reset_rewind(struct intel_engine_cs *engine, bool stalled)
+{
+	GEM_BUG_ON(stalled);
+}
+
+static void mock_reset_cancel(struct intel_engine_cs *engine)
+{
+	struct mock_engine *mock =
+		container_of(engine, typeof(*mock), base);
+	struct i915_request *rq;
+	unsigned long flags;
+
+	del_timer_sync(&mock->hw_delay);
+
+	spin_lock_irqsave(&engine->sched_engine->lock, flags);
+
+	/* Mark all submitted requests as skipped. */
+	list_for_each_entry(rq, &engine->sched_engine->requests, sched.link)
+		i915_request_put(i915_request_mark_eio(rq));
+	intel_engine_signal_breadcrumbs(engine);
+
+	/* Cancel and submit all pending requests. */
+	list_for_each_entry(rq, &mock->hw_queue, mock.link) {
+		if (i915_request_mark_eio(rq)) {
+			__i915_request_submit(rq);
+			i915_request_put(rq);
+		}
+	}
+	INIT_LIST_HEAD(&mock->hw_queue);
+
+	spin_unlock_irqrestore(&engine->sched_engine->lock, flags);
+}
+
+static void mock_reset_finish(struct intel_engine_cs *engine)
+{
+}
+
+static void mock_engine_release(struct intel_engine_cs *engine)
+{
+	struct mock_engine *mock =
+		container_of(engine, typeof(*mock), base);
+
+	GEM_BUG_ON(timer_pending(&mock->hw_delay));
+
+	i915_sched_engine_put(engine->sched_engine);
+	intel_breadcrumbs_put(engine->breadcrumbs);
+
+	intel_context_unpin(engine->kernel_context);
+	intel_context_put(engine->kernel_context);
+
+	intel_engine_fini_retire(engine);
+}
+
+struct intel_engine_cs *mock_engine(struct drm_i915_private *i915,
+				    const char *name,
+				    int id)
+{
+	struct mock_engine *engine;
+
+	GEM_BUG_ON(id >= I915_NUM_ENGINES);
+	GEM_BUG_ON(!to_gt(i915)->uncore);
+
+	engine = kzalloc(sizeof(*engine) + PAGE_SIZE, GFP_KERNEL);
+	if (!engine)
+		return NULL;
+
+	/* minimal engine setup for requests */
+	engine->base.i915 = i915;
+	engine->base.gt = to_gt(i915);
+	engine->base.uncore = to_gt(i915)->uncore;
+	snprintf(engine->base.name, sizeof(engine->base.name), "%s", name);
+	engine->base.id = id;
+	engine->base.mask = BIT(id);
+	engine->base.legacy_idx = INVALID_ENGINE;
+	engine->base.instance = id;
+	engine->base.status_page.addr = (void *)(engine + 1);
+
+	engine->base.cops = &mock_context_ops;
+	engine->base.request_alloc = mock_request_alloc;
+	engine->base.emit_flush = mock_emit_flush;
+	engine->base.emit_fini_breadcrumb = mock_emit_breadcrumb;
+	engine->base.submit_request = mock_submit_request;
+	engine->base.add_active_request = mock_add_to_engine;
+	engine->base.remove_active_request = mock_remove_from_engine;
+
+	engine->base.reset.prepare = mock_reset_prepare;
+	engine->base.reset.rewind = mock_reset_rewind;
+	engine->base.reset.cancel = mock_reset_cancel;
+	engine->base.reset.finish = mock_reset_finish;
+
+	engine->base.release = mock_engine_release;
+
+	to_gt(i915)->engine[id] = &engine->base;
+	to_gt(i915)->engine_class[0][id] = &engine->base;
+
+	/* fake hw queue */
+	spin_lock_init(&engine->hw_lock);
+	timer_setup(&engine->hw_delay, hw_delay_complete, 0);
+	INIT_LIST_HEAD(&engine->hw_queue);
+
+	intel_engine_add_user(&engine->base);
+
+	return &engine->base;
+}
+
+int mock_engine_init(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+
+	INIT_LIST_HEAD(&engine->pinned_contexts_list);
+
+	engine->sched_engine = i915_sched_engine_create(ENGINE_MOCK);
+	if (!engine->sched_engine)
+		return -ENOMEM;
+	engine->sched_engine->private_data = engine;
+
+	intel_engine_init_execlists(engine);
+	intel_engine_init__pm(engine);
+	intel_engine_init_retire(engine);
+
+	engine->breadcrumbs = intel_breadcrumbs_create(NULL);
+	if (!engine->breadcrumbs)
+		goto err_schedule;
+
+	ce = create_kernel_context(engine);
+	if (IS_ERR(ce))
+		goto err_breadcrumbs;
+
+	/* We insist the kernel context is using the status_page */
+	engine->status_page.vma = ce->timeline->hwsp_ggtt;
+
+	engine->kernel_context = ce;
+	return 0;
+
+err_breadcrumbs:
+	intel_breadcrumbs_put(engine->breadcrumbs);
+err_schedule:
+	i915_sched_engine_put(engine->sched_engine);
+	return -ENOMEM;
+}
+
+void mock_engine_flush(struct intel_engine_cs *engine)
+{
+	struct mock_engine *mock =
+		container_of(engine, typeof(*mock), base);
+	struct i915_request *request, *rn;
+
+	del_timer_sync(&mock->hw_delay);
+
+	spin_lock_irq(&mock->hw_lock);
+	list_for_each_entry_safe(request, rn, &mock->hw_queue, mock.link)
+		advance(request);
+	spin_unlock_irq(&mock->hw_lock);
+}
+
+void mock_engine_reset(struct intel_engine_cs *engine)
+{
+}
diff --git a/drivers/gpu/drm/i915/gt/mock_engine.h b/drivers/gpu/drm/i915/gt/mock_engine.h
new file mode 100644
index 000000000..cc5ab6e1f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/mock_engine.h
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef __MOCK_ENGINE_H__
+#define __MOCK_ENGINE_H__
+
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/timer.h>
+
+#include "gt/intel_engine.h"
+
+struct mock_engine {
+	struct intel_engine_cs base;
+
+	spinlock_t hw_lock;
+	struct list_head hw_queue;
+	struct timer_list hw_delay;
+};
+
+struct intel_engine_cs *mock_engine(struct drm_i915_private *i915,
+				    const char *name,
+				    int id);
+int mock_engine_init(struct intel_engine_cs *engine);
+
+void mock_engine_flush(struct intel_engine_cs *engine);
+void mock_engine_reset(struct intel_engine_cs *engine);
+void mock_engine_free(struct intel_engine_cs *engine);
+
+#endif /* !__MOCK_ENGINE_H__ */
diff --git a/drivers/gpu/drm/i915/gt/selftest_context.c b/drivers/gpu/drm/i915/gt/selftest_context.c
new file mode 100644
index 000000000..76fbae358
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_context.c
@@ -0,0 +1,451 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "i915_selftest.h"
+#include "intel_engine_heartbeat.h"
+#include "intel_engine_pm.h"
+#include "intel_gt.h"
+
+#include "gem/selftests/mock_context.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/mock_drm.h"
+
+static int request_sync(struct i915_request *rq)
+{
+	struct intel_timeline *tl = i915_request_timeline(rq);
+	long timeout;
+	int err = 0;
+
+	intel_timeline_get(tl);
+	i915_request_get(rq);
+
+	/* Opencode i915_request_add() so we can keep the timeline locked. */
+	__i915_request_commit(rq);
+	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+	__i915_request_queue_bh(rq);
+
+	timeout = i915_request_wait(rq, 0, HZ / 10);
+	if (timeout < 0)
+		err = timeout;
+	else
+		i915_request_retire_upto(rq);
+
+	lockdep_unpin_lock(&tl->mutex, rq->cookie);
+	mutex_unlock(&tl->mutex);
+
+	i915_request_put(rq);
+	intel_timeline_put(tl);
+
+	return err;
+}
+
+static int context_sync(struct intel_context *ce)
+{
+	struct intel_timeline *tl = ce->timeline;
+	int err = 0;
+
+	mutex_lock(&tl->mutex);
+	do {
+		struct i915_request *rq;
+		long timeout;
+
+		if (list_empty(&tl->requests))
+			break;
+
+		rq = list_last_entry(&tl->requests, typeof(*rq), link);
+		i915_request_get(rq);
+
+		timeout = i915_request_wait(rq, 0, HZ / 10);
+		if (timeout < 0)
+			err = timeout;
+		else
+			i915_request_retire_upto(rq);
+
+		i915_request_put(rq);
+	} while (!err);
+	mutex_unlock(&tl->mutex);
+
+	/* Wait for all barriers to complete (remote CPU) before we check */
+	i915_active_unlock_wait(&ce->active);
+	return err;
+}
+
+static int __live_context_size(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+	void *vaddr;
+	int err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	err = intel_context_pin(ce);
+	if (err)
+		goto err;
+
+	vaddr = i915_gem_object_pin_map_unlocked(ce->state->obj,
+						 i915_coherent_map_type(engine->i915,
+									ce->state->obj, false));
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		intel_context_unpin(ce);
+		goto err;
+	}
+
+	/*
+	 * Note that execlists also applies a redzone which it checks on
+	 * context unpin when debugging. We are using the same location
+	 * and same poison value so that our checks overlap. Despite the
+	 * redundancy, we want to keep this little selftest so that we
+	 * get coverage of any and all submission backends, and we can
+	 * always extend this test to ensure we trick the HW into a
+	 * compromising position wrt to the various sections that need
+	 * to be written into the context state.
+	 *
+	 * TLDR; this overlaps with the execlists redzone.
+	 */
+	vaddr += engine->context_size - I915_GTT_PAGE_SIZE;
+	memset(vaddr, POISON_INUSE, I915_GTT_PAGE_SIZE);
+
+	rq = intel_context_create_request(ce);
+	intel_context_unpin(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_unpin;
+	}
+
+	err = request_sync(rq);
+	if (err)
+		goto err_unpin;
+
+	/* Force the context switch */
+	rq = intel_engine_create_kernel_request(engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_unpin;
+	}
+	err = request_sync(rq);
+	if (err)
+		goto err_unpin;
+
+	if (memchr_inv(vaddr, POISON_INUSE, I915_GTT_PAGE_SIZE)) {
+		pr_err("%s context overwrote trailing red-zone!", engine->name);
+		err = -EINVAL;
+	}
+
+err_unpin:
+	i915_gem_object_unpin_map(ce->state->obj);
+err:
+	intel_context_put(ce);
+	return err;
+}
+
+static int live_context_size(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Check that our context sizes are correct by seeing if the
+	 * HW tries to write past the end of one.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		struct file *saved;
+
+		if (!engine->context_size)
+			continue;
+
+		intel_engine_pm_get(engine);
+
+		/*
+		 * Hide the old default state -- we lie about the context size
+		 * and get confused when the default state is smaller than
+		 * expected. For our do nothing request, inheriting the
+		 * active state is sufficient, we are only checking that we
+		 * don't use more than we planned.
+		 */
+		saved = fetch_and_zero(&engine->default_state);
+
+		/* Overlaps with the execlists redzone */
+		engine->context_size += I915_GTT_PAGE_SIZE;
+
+		err = __live_context_size(engine);
+
+		engine->context_size -= I915_GTT_PAGE_SIZE;
+
+		engine->default_state = saved;
+
+		intel_engine_pm_put(engine);
+
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static int __live_active_context(struct intel_engine_cs *engine)
+{
+	unsigned long saved_heartbeat;
+	struct intel_context *ce;
+	int pass;
+	int err;
+
+	/*
+	 * We keep active contexts alive until after a subsequent context
+	 * switch as the final write from the context-save will be after
+	 * we retire the final request. We track when we unpin the context,
+	 * under the presumption that the final pin is from the last request,
+	 * and instead of immediately unpinning the context, we add a task
+	 * to unpin the context from the next idle-barrier.
+	 *
+	 * This test makes sure that the context is kept alive until a
+	 * subsequent idle-barrier (emitted when the engine wakeref hits 0
+	 * with no more outstanding requests).
+	 *
+	 * In GuC submission mode we don't use idle barriers and we instead
+	 * get a message from the GuC to signal that it is safe to unpin the
+	 * context from memory.
+	 */
+	if (intel_engine_uses_guc(engine))
+		return 0;
+
+	if (intel_engine_pm_is_awake(engine)) {
+		pr_err("%s is awake before starting %s!\n",
+		       engine->name, __func__);
+		return -EINVAL;
+	}
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	saved_heartbeat = engine->props.heartbeat_interval_ms;
+	engine->props.heartbeat_interval_ms = 0;
+
+	for (pass = 0; pass <= 2; pass++) {
+		struct i915_request *rq;
+
+		intel_engine_pm_get(engine);
+
+		rq = intel_context_create_request(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_engine;
+		}
+
+		err = request_sync(rq);
+		if (err)
+			goto out_engine;
+
+		/* Context will be kept active until after an idle-barrier. */
+		if (i915_active_is_idle(&ce->active)) {
+			pr_err("context is not active; expected idle-barrier (%s pass %d)\n",
+			       engine->name, pass);
+			err = -EINVAL;
+			goto out_engine;
+		}
+
+		if (!intel_engine_pm_is_awake(engine)) {
+			pr_err("%s is asleep before idle-barrier\n",
+			       engine->name);
+			err = -EINVAL;
+			goto out_engine;
+		}
+
+out_engine:
+		intel_engine_pm_put(engine);
+		if (err)
+			goto err;
+	}
+
+	/* Now make sure our idle-barriers are flushed */
+	err = intel_engine_flush_barriers(engine);
+	if (err)
+		goto err;
+
+	/* Wait for the barrier and in the process wait for engine to park */
+	err = context_sync(engine->kernel_context);
+	if (err)
+		goto err;
+
+	if (!i915_active_is_idle(&ce->active)) {
+		pr_err("context is still active!");
+		err = -EINVAL;
+	}
+
+	intel_engine_pm_flush(engine);
+
+	if (intel_engine_pm_is_awake(engine)) {
+		struct drm_printer p = drm_debug_printer(__func__);
+
+		intel_engine_dump(engine, &p,
+				  "%s is still awake:%d after idle-barriers\n",
+				  engine->name,
+				  atomic_read(&engine->wakeref.count));
+		GEM_TRACE_DUMP();
+
+		err = -EINVAL;
+		goto err;
+	}
+
+err:
+	engine->props.heartbeat_interval_ms = saved_heartbeat;
+	intel_context_put(ce);
+	return err;
+}
+
+static int live_active_context(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	for_each_engine(engine, gt, id) {
+		err = __live_active_context(engine);
+		if (err)
+			break;
+
+		err = igt_flush_test(gt->i915);
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static int __remote_sync(struct intel_context *ce, struct intel_context *remote)
+{
+	struct i915_request *rq;
+	int err;
+
+	err = intel_context_pin(remote);
+	if (err)
+		return err;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto unpin;
+	}
+
+	err = intel_context_prepare_remote_request(remote, rq);
+	if (err) {
+		i915_request_add(rq);
+		goto unpin;
+	}
+
+	err = request_sync(rq);
+
+unpin:
+	intel_context_unpin(remote);
+	return err;
+}
+
+static int __live_remote_context(struct intel_engine_cs *engine)
+{
+	struct intel_context *local, *remote;
+	unsigned long saved_heartbeat;
+	int pass;
+	int err;
+
+	/*
+	 * Check that our idle barriers do not interfere with normal
+	 * activity tracking. In particular, check that operating
+	 * on the context image remotely (intel_context_prepare_remote_request),
+	 * which inserts foreign fences into intel_context.active, does not
+	 * clobber the idle-barrier.
+	 *
+	 * In GuC submission mode we don't use idle barriers.
+	 */
+	if (intel_engine_uses_guc(engine))
+		return 0;
+
+	if (intel_engine_pm_is_awake(engine)) {
+		pr_err("%s is awake before starting %s!\n",
+		       engine->name, __func__);
+		return -EINVAL;
+	}
+
+	remote = intel_context_create(engine);
+	if (IS_ERR(remote))
+		return PTR_ERR(remote);
+
+	local = intel_context_create(engine);
+	if (IS_ERR(local)) {
+		err = PTR_ERR(local);
+		goto err_remote;
+	}
+
+	saved_heartbeat = engine->props.heartbeat_interval_ms;
+	engine->props.heartbeat_interval_ms = 0;
+	intel_engine_pm_get(engine);
+
+	for (pass = 0; pass <= 2; pass++) {
+		err = __remote_sync(local, remote);
+		if (err)
+			break;
+
+		err = __remote_sync(engine->kernel_context, remote);
+		if (err)
+			break;
+
+		if (i915_active_is_idle(&remote->active)) {
+			pr_err("remote context is not active; expected idle-barrier (%s pass %d)\n",
+			       engine->name, pass);
+			err = -EINVAL;
+			break;
+		}
+	}
+
+	intel_engine_pm_put(engine);
+	engine->props.heartbeat_interval_ms = saved_heartbeat;
+
+	intel_context_put(local);
+err_remote:
+	intel_context_put(remote);
+	return err;
+}
+
+static int live_remote_context(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	for_each_engine(engine, gt, id) {
+		err = __live_remote_context(engine);
+		if (err)
+			break;
+
+		err = igt_flush_test(gt->i915);
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+int intel_context_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_context_size),
+		SUBTEST(live_active_context),
+		SUBTEST(live_remote_context),
+	};
+	struct intel_gt *gt = to_gt(i915);
+
+	if (intel_gt_is_wedged(gt))
+		return 0;
+
+	return intel_gt_live_subtests(tests, gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_engine.c b/drivers/gpu/drm/i915/gt/selftest_engine.c
new file mode 100644
index 000000000..57fea9ea1
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_engine.c
@@ -0,0 +1,27 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include "i915_selftest.h"
+#include "selftest_engine.h"
+
+int intel_engine_live_selftests(struct drm_i915_private *i915)
+{
+	static int (* const tests[])(struct intel_gt *) = {
+		live_engine_pm_selftests,
+		NULL,
+	};
+	struct intel_gt *gt = to_gt(i915);
+	typeof(*tests) *fn;
+
+	for (fn = tests; *fn; fn++) {
+		int err;
+
+		err = (*fn)(gt);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_engine.h b/drivers/gpu/drm/i915/gt/selftest_engine.h
new file mode 100644
index 000000000..c6feb3bd2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_engine.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef SELFTEST_ENGINE_H
+#define SELFTEST_ENGINE_H
+
+struct intel_gt;
+
+int live_engine_pm_selftests(struct intel_gt *gt);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/selftest_engine_cs.c b/drivers/gpu/drm/i915/gt/selftest_engine_cs.c
new file mode 100644
index 000000000..1b75f478d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_engine_cs.c
@@ -0,0 +1,423 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/sort.h>
+
+#include "intel_gpu_commands.h"
+#include "intel_gt_pm.h"
+#include "intel_rps.h"
+
+#include "i915_selftest.h"
+#include "selftests/igt_flush_test.h"
+
+#define COUNT 5
+
+static int cmp_u32(const void *A, const void *B)
+{
+	const u32 *a = A, *b = B;
+
+	return *a - *b;
+}
+
+static void perf_begin(struct intel_gt *gt)
+{
+	intel_gt_pm_get(gt);
+
+	/* Boost gpufreq to max [waitboost] and keep it fixed */
+	atomic_inc(&gt->rps.num_waiters);
+	schedule_work(&gt->rps.work);
+	flush_work(&gt->rps.work);
+}
+
+static int perf_end(struct intel_gt *gt)
+{
+	atomic_dec(&gt->rps.num_waiters);
+	intel_gt_pm_put(gt);
+
+	return igt_flush_test(gt->i915);
+}
+
+static int write_timestamp(struct i915_request *rq, int slot)
+{
+	struct intel_timeline *tl =
+		rcu_dereference_protected(rq->timeline,
+					  !i915_request_signaled(rq));
+	u32 cmd;
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	cmd = MI_STORE_REGISTER_MEM | MI_USE_GGTT;
+	if (GRAPHICS_VER(rq->engine->i915) >= 8)
+		cmd++;
+	*cs++ = cmd;
+	*cs++ = i915_mmio_reg_offset(RING_TIMESTAMP(rq->engine->mmio_base));
+	*cs++ = tl->hwsp_offset + slot * sizeof(u32);
+	*cs++ = 0;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static struct i915_vma *create_empty_batch(struct intel_context *ce)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u32 *cs;
+	int err;
+
+	obj = i915_gem_object_create_internal(ce->engine->i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	cs = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_put;
+	}
+
+	cs[0] = MI_BATCH_BUFFER_END;
+
+	i915_gem_object_flush_map(obj);
+
+	vma = i915_vma_instance(obj, ce->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_unpin;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto err_unpin;
+
+	i915_gem_object_unpin_map(obj);
+	return vma;
+
+err_unpin:
+	i915_gem_object_unpin_map(obj);
+err_put:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static u32 trifilter(u32 *a)
+{
+	u64 sum;
+
+	sort(a, COUNT, sizeof(*a), cmp_u32, NULL);
+
+	sum = mul_u32_u32(a[2], 2);
+	sum += a[1];
+	sum += a[3];
+
+	return sum >> 2;
+}
+
+static int perf_mi_bb_start(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	if (GRAPHICS_VER(gt->i915) < 7) /* for per-engine CS_TIMESTAMP */
+		return 0;
+
+	perf_begin(gt);
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce = engine->kernel_context;
+		struct i915_vma *batch;
+		u32 cycles[COUNT];
+		int i;
+
+		intel_engine_pm_get(engine);
+
+		batch = create_empty_batch(ce);
+		if (IS_ERR(batch)) {
+			err = PTR_ERR(batch);
+			intel_engine_pm_put(engine);
+			break;
+		}
+
+		err = i915_vma_sync(batch);
+		if (err) {
+			intel_engine_pm_put(engine);
+			i915_vma_put(batch);
+			break;
+		}
+
+		for (i = 0; i < ARRAY_SIZE(cycles); i++) {
+			struct i915_request *rq;
+
+			rq = i915_request_create(ce);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				break;
+			}
+
+			err = write_timestamp(rq, 2);
+			if (err)
+				goto out;
+
+			err = rq->engine->emit_bb_start(rq,
+							batch->node.start, 8,
+							0);
+			if (err)
+				goto out;
+
+			err = write_timestamp(rq, 3);
+			if (err)
+				goto out;
+
+out:
+			i915_request_get(rq);
+			i915_request_add(rq);
+
+			if (i915_request_wait(rq, 0, HZ / 5) < 0)
+				err = -EIO;
+			i915_request_put(rq);
+			if (err)
+				break;
+
+			cycles[i] = rq->hwsp_seqno[3] - rq->hwsp_seqno[2];
+		}
+		i915_vma_put(batch);
+		intel_engine_pm_put(engine);
+		if (err)
+			break;
+
+		pr_info("%s: MI_BB_START cycles: %u\n",
+			engine->name, trifilter(cycles));
+	}
+	if (perf_end(gt))
+		err = -EIO;
+
+	return err;
+}
+
+static struct i915_vma *create_nop_batch(struct intel_context *ce)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u32 *cs;
+	int err;
+
+	obj = i915_gem_object_create_internal(ce->engine->i915, SZ_64K);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	cs = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_put;
+	}
+
+	memset(cs, 0, SZ_64K);
+	cs[SZ_64K / sizeof(*cs) - 1] = MI_BATCH_BUFFER_END;
+
+	i915_gem_object_flush_map(obj);
+
+	vma = i915_vma_instance(obj, ce->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_unpin;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto err_unpin;
+
+	i915_gem_object_unpin_map(obj);
+	return vma;
+
+err_unpin:
+	i915_gem_object_unpin_map(obj);
+err_put:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static int perf_mi_noop(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	if (GRAPHICS_VER(gt->i915) < 7) /* for per-engine CS_TIMESTAMP */
+		return 0;
+
+	perf_begin(gt);
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce = engine->kernel_context;
+		struct i915_vma *base, *nop;
+		u32 cycles[COUNT];
+		int i;
+
+		intel_engine_pm_get(engine);
+
+		base = create_empty_batch(ce);
+		if (IS_ERR(base)) {
+			err = PTR_ERR(base);
+			intel_engine_pm_put(engine);
+			break;
+		}
+
+		err = i915_vma_sync(base);
+		if (err) {
+			i915_vma_put(base);
+			intel_engine_pm_put(engine);
+			break;
+		}
+
+		nop = create_nop_batch(ce);
+		if (IS_ERR(nop)) {
+			err = PTR_ERR(nop);
+			i915_vma_put(base);
+			intel_engine_pm_put(engine);
+			break;
+		}
+
+		err = i915_vma_sync(nop);
+		if (err) {
+			i915_vma_put(nop);
+			i915_vma_put(base);
+			intel_engine_pm_put(engine);
+			break;
+		}
+
+		for (i = 0; i < ARRAY_SIZE(cycles); i++) {
+			struct i915_request *rq;
+
+			rq = i915_request_create(ce);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				break;
+			}
+
+			err = write_timestamp(rq, 2);
+			if (err)
+				goto out;
+
+			err = rq->engine->emit_bb_start(rq,
+							base->node.start, 8,
+							0);
+			if (err)
+				goto out;
+
+			err = write_timestamp(rq, 3);
+			if (err)
+				goto out;
+
+			err = rq->engine->emit_bb_start(rq,
+							nop->node.start,
+							nop->node.size,
+							0);
+			if (err)
+				goto out;
+
+			err = write_timestamp(rq, 4);
+			if (err)
+				goto out;
+
+out:
+			i915_request_get(rq);
+			i915_request_add(rq);
+
+			if (i915_request_wait(rq, 0, HZ / 5) < 0)
+				err = -EIO;
+			i915_request_put(rq);
+			if (err)
+				break;
+
+			cycles[i] =
+				(rq->hwsp_seqno[4] - rq->hwsp_seqno[3]) -
+				(rq->hwsp_seqno[3] - rq->hwsp_seqno[2]);
+		}
+		i915_vma_put(nop);
+		i915_vma_put(base);
+		intel_engine_pm_put(engine);
+		if (err)
+			break;
+
+		pr_info("%s: 16K MI_NOOP cycles: %u\n",
+			engine->name, trifilter(cycles));
+	}
+	if (perf_end(gt))
+		err = -EIO;
+
+	return err;
+}
+
+int intel_engine_cs_perf_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(perf_mi_bb_start),
+		SUBTEST(perf_mi_noop),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
+
+static int intel_mmio_bases_check(void *arg)
+{
+	int i, j;
+
+	for (i = 0; i < ARRAY_SIZE(intel_engines); i++) {
+		const struct engine_info *info = &intel_engines[i];
+		u8 prev = U8_MAX;
+
+		for (j = 0; j < MAX_MMIO_BASES; j++) {
+			u8 ver = info->mmio_bases[j].graphics_ver;
+			u32 base = info->mmio_bases[j].base;
+
+			if (ver >= prev) {
+				pr_err("%s(%s, class:%d, instance:%d): mmio base for graphics ver %u is before the one for ver %u\n",
+				       __func__,
+				       intel_engine_class_repr(info->class),
+				       info->class, info->instance,
+				       prev, ver);
+				return -EINVAL;
+			}
+
+			if (ver == 0)
+				break;
+
+			if (!base) {
+				pr_err("%s(%s, class:%d, instance:%d): invalid mmio base (%x) for graphics ver %u at entry %u\n",
+				       __func__,
+				       intel_engine_class_repr(info->class),
+				       info->class, info->instance,
+				       base, ver, j);
+				return -EINVAL;
+			}
+
+			prev = ver;
+		}
+
+		pr_debug("%s: min graphics version supported for %s%d is %u\n",
+			 __func__,
+			 intel_engine_class_repr(info->class),
+			 info->instance,
+			 prev);
+	}
+
+	return 0;
+}
+
+int intel_engine_cs_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(intel_mmio_bases_check),
+	};
+
+	return i915_subtests(tests, NULL);
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_engine_heartbeat.c b/drivers/gpu/drm/i915/gt/selftest_engine_heartbeat.c
new file mode 100644
index 000000000..273d440a5
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_engine_heartbeat.c
@@ -0,0 +1,429 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/sort.h>
+
+#include "i915_drv.h"
+
+#include "intel_gt_requests.h"
+#include "i915_selftest.h"
+#include "selftest_engine_heartbeat.h"
+
+static void reset_heartbeat(struct intel_engine_cs *engine)
+{
+	intel_engine_set_heartbeat(engine,
+				   engine->defaults.heartbeat_interval_ms);
+}
+
+static int timeline_sync(struct intel_timeline *tl)
+{
+	struct dma_fence *fence;
+	long timeout;
+
+	fence = i915_active_fence_get(&tl->last_request);
+	if (!fence)
+		return 0;
+
+	timeout = dma_fence_wait_timeout(fence, true, HZ / 2);
+	dma_fence_put(fence);
+	if (timeout < 0)
+		return timeout;
+
+	return 0;
+}
+
+static int engine_sync_barrier(struct intel_engine_cs *engine)
+{
+	return timeline_sync(engine->kernel_context->timeline);
+}
+
+struct pulse {
+	struct i915_active active;
+	struct kref kref;
+};
+
+static int pulse_active(struct i915_active *active)
+{
+	kref_get(&container_of(active, struct pulse, active)->kref);
+	return 0;
+}
+
+static void pulse_free(struct kref *kref)
+{
+	struct pulse *p = container_of(kref, typeof(*p), kref);
+
+	i915_active_fini(&p->active);
+	kfree(p);
+}
+
+static void pulse_put(struct pulse *p)
+{
+	kref_put(&p->kref, pulse_free);
+}
+
+static void pulse_retire(struct i915_active *active)
+{
+	pulse_put(container_of(active, struct pulse, active));
+}
+
+static struct pulse *pulse_create(void)
+{
+	struct pulse *p;
+
+	p = kmalloc(sizeof(*p), GFP_KERNEL);
+	if (!p)
+		return p;
+
+	kref_init(&p->kref);
+	i915_active_init(&p->active, pulse_active, pulse_retire, 0);
+
+	return p;
+}
+
+static void pulse_unlock_wait(struct pulse *p)
+{
+	i915_active_unlock_wait(&p->active);
+}
+
+static int __live_idle_pulse(struct intel_engine_cs *engine,
+			     int (*fn)(struct intel_engine_cs *cs))
+{
+	struct pulse *p;
+	int err;
+
+	GEM_BUG_ON(!intel_engine_pm_is_awake(engine));
+
+	p = pulse_create();
+	if (!p)
+		return -ENOMEM;
+
+	err = i915_active_acquire(&p->active);
+	if (err)
+		goto out;
+
+	err = i915_active_acquire_preallocate_barrier(&p->active, engine);
+	if (err) {
+		i915_active_release(&p->active);
+		goto out;
+	}
+
+	i915_active_acquire_barrier(&p->active);
+	i915_active_release(&p->active);
+
+	GEM_BUG_ON(i915_active_is_idle(&p->active));
+	GEM_BUG_ON(llist_empty(&engine->barrier_tasks));
+
+	err = fn(engine);
+	if (err)
+		goto out;
+
+	GEM_BUG_ON(!llist_empty(&engine->barrier_tasks));
+
+	if (engine_sync_barrier(engine)) {
+		struct drm_printer m = drm_err_printer("pulse");
+
+		pr_err("%s: no heartbeat pulse?\n", engine->name);
+		intel_engine_dump(engine, &m, "%s", engine->name);
+
+		err = -ETIME;
+		goto out;
+	}
+
+	GEM_BUG_ON(READ_ONCE(engine->serial) != engine->wakeref_serial);
+
+	pulse_unlock_wait(p); /* synchronize with the retirement callback */
+
+	if (!i915_active_is_idle(&p->active)) {
+		struct drm_printer m = drm_err_printer("pulse");
+
+		pr_err("%s: heartbeat pulse did not flush idle tasks\n",
+		       engine->name);
+		i915_active_print(&p->active, &m);
+
+		err = -EINVAL;
+		goto out;
+	}
+
+out:
+	pulse_put(p);
+	return err;
+}
+
+static int live_idle_flush(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/* Check that we can flush the idle barriers */
+
+	for_each_engine(engine, gt, id) {
+		st_engine_heartbeat_disable(engine);
+		err = __live_idle_pulse(engine, intel_engine_flush_barriers);
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static int live_idle_pulse(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/* Check that heartbeat pulses flush the idle barriers */
+
+	for_each_engine(engine, gt, id) {
+		st_engine_heartbeat_disable(engine);
+		err = __live_idle_pulse(engine, intel_engine_pulse);
+		st_engine_heartbeat_enable(engine);
+		if (err && err != -ENODEV)
+			break;
+
+		err = 0;
+	}
+
+	return err;
+}
+
+static int cmp_u32(const void *_a, const void *_b)
+{
+	const u32 *a = _a, *b = _b;
+
+	return *a - *b;
+}
+
+static int __live_heartbeat_fast(struct intel_engine_cs *engine)
+{
+	const unsigned int error_threshold = max(20000u, jiffies_to_usecs(6));
+	struct intel_context *ce;
+	struct i915_request *rq;
+	ktime_t t0, t1;
+	u32 times[5];
+	int err;
+	int i;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	intel_engine_pm_get(engine);
+
+	err = intel_engine_set_heartbeat(engine, 1);
+	if (err)
+		goto err_pm;
+
+	for (i = 0; i < ARRAY_SIZE(times); i++) {
+		do {
+			/* Manufacture a tick */
+			intel_engine_park_heartbeat(engine);
+			GEM_BUG_ON(engine->heartbeat.systole);
+			engine->serial++; /*  pretend we are not idle! */
+			intel_engine_unpark_heartbeat(engine);
+
+			flush_delayed_work(&engine->heartbeat.work);
+			if (!delayed_work_pending(&engine->heartbeat.work)) {
+				pr_err("%s: heartbeat %d did not start\n",
+				       engine->name, i);
+				err = -EINVAL;
+				goto err_pm;
+			}
+
+			rcu_read_lock();
+			rq = READ_ONCE(engine->heartbeat.systole);
+			if (rq)
+				rq = i915_request_get_rcu(rq);
+			rcu_read_unlock();
+		} while (!rq);
+
+		t0 = ktime_get();
+		while (rq == READ_ONCE(engine->heartbeat.systole))
+			yield(); /* work is on the local cpu! */
+		t1 = ktime_get();
+
+		i915_request_put(rq);
+		times[i] = ktime_us_delta(t1, t0);
+	}
+
+	sort(times, ARRAY_SIZE(times), sizeof(times[0]), cmp_u32, NULL);
+
+	pr_info("%s: Heartbeat delay: %uus [%u, %u]\n",
+		engine->name,
+		times[ARRAY_SIZE(times) / 2],
+		times[0],
+		times[ARRAY_SIZE(times) - 1]);
+
+	/*
+	 * Ideally, the upper bound on min work delay would be something like
+	 * 2 * 2 (worst), +1 for scheduling, +1 for slack. In practice, we
+	 * are, even with system_wq_highpri, at the mercy of the CPU scheduler
+	 * and may be stuck behind some slow work for many millisecond. Such
+	 * as our very own display workers.
+	 */
+	if (times[ARRAY_SIZE(times) / 2] > error_threshold) {
+		pr_err("%s: Heartbeat delay was %uus, expected less than %dus\n",
+		       engine->name,
+		       times[ARRAY_SIZE(times) / 2],
+		       error_threshold);
+		err = -EINVAL;
+	}
+
+	reset_heartbeat(engine);
+err_pm:
+	intel_engine_pm_put(engine);
+	intel_context_put(ce);
+	return err;
+}
+
+static int live_heartbeat_fast(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/* Check that the heartbeat ticks at the desired rate. */
+	if (!CONFIG_DRM_I915_HEARTBEAT_INTERVAL)
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		err = __live_heartbeat_fast(engine);
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static int __live_heartbeat_off(struct intel_engine_cs *engine)
+{
+	int err;
+
+	intel_engine_pm_get(engine);
+
+	engine->serial++;
+	flush_delayed_work(&engine->heartbeat.work);
+	if (!delayed_work_pending(&engine->heartbeat.work)) {
+		pr_err("%s: heartbeat not running\n",
+		       engine->name);
+		err = -EINVAL;
+		goto err_pm;
+	}
+
+	err = intel_engine_set_heartbeat(engine, 0);
+	if (err)
+		goto err_pm;
+
+	engine->serial++;
+	flush_delayed_work(&engine->heartbeat.work);
+	if (delayed_work_pending(&engine->heartbeat.work)) {
+		pr_err("%s: heartbeat still running\n",
+		       engine->name);
+		err = -EINVAL;
+		goto err_beat;
+	}
+
+	if (READ_ONCE(engine->heartbeat.systole)) {
+		pr_err("%s: heartbeat still allocated\n",
+		       engine->name);
+		err = -EINVAL;
+		goto err_beat;
+	}
+
+err_beat:
+	reset_heartbeat(engine);
+err_pm:
+	intel_engine_pm_put(engine);
+	return err;
+}
+
+static int live_heartbeat_off(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/* Check that we can turn off heartbeat and not interrupt VIP */
+	if (!CONFIG_DRM_I915_HEARTBEAT_INTERVAL)
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		err = __live_heartbeat_off(engine);
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+int intel_heartbeat_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_idle_flush),
+		SUBTEST(live_idle_pulse),
+		SUBTEST(live_heartbeat_fast),
+		SUBTEST(live_heartbeat_off),
+	};
+	int saved_hangcheck;
+	int err;
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	saved_hangcheck = i915->params.enable_hangcheck;
+	i915->params.enable_hangcheck = INT_MAX;
+
+	err = intel_gt_live_subtests(tests, to_gt(i915));
+
+	i915->params.enable_hangcheck = saved_hangcheck;
+	return err;
+}
+
+void st_engine_heartbeat_disable(struct intel_engine_cs *engine)
+{
+	engine->props.heartbeat_interval_ms = 0;
+
+	intel_engine_pm_get(engine);
+	intel_engine_park_heartbeat(engine);
+}
+
+void st_engine_heartbeat_enable(struct intel_engine_cs *engine)
+{
+	intel_engine_pm_put(engine);
+
+	engine->props.heartbeat_interval_ms =
+		engine->defaults.heartbeat_interval_ms;
+}
+
+void st_engine_heartbeat_disable_no_pm(struct intel_engine_cs *engine)
+{
+	engine->props.heartbeat_interval_ms = 0;
+
+	/*
+	 * Park the heartbeat but without holding the PM lock as that
+	 * makes the engines appear not-idle. Note that if/when unpark
+	 * is called due to the PM lock being acquired later the
+	 * heartbeat still won't be enabled because of the above = 0.
+	 */
+	if (intel_engine_pm_get_if_awake(engine)) {
+		intel_engine_park_heartbeat(engine);
+		intel_engine_pm_put(engine);
+	}
+}
+
+void st_engine_heartbeat_enable_no_pm(struct intel_engine_cs *engine)
+{
+	engine->props.heartbeat_interval_ms =
+		engine->defaults.heartbeat_interval_ms;
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_engine_heartbeat.h b/drivers/gpu/drm/i915/gt/selftest_engine_heartbeat.h
new file mode 100644
index 000000000..81da2cd8e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_engine_heartbeat.h
@@ -0,0 +1,16 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef SELFTEST_ENGINE_HEARTBEAT_H
+#define SELFTEST_ENGINE_HEARTBEAT_H
+
+struct intel_engine_cs;
+
+void st_engine_heartbeat_disable(struct intel_engine_cs *engine);
+void st_engine_heartbeat_disable_no_pm(struct intel_engine_cs *engine);
+void st_engine_heartbeat_enable(struct intel_engine_cs *engine);
+void st_engine_heartbeat_enable_no_pm(struct intel_engine_cs *engine);
+
+#endif /* SELFTEST_ENGINE_HEARTBEAT_H */
diff --git a/drivers/gpu/drm/i915/gt/selftest_engine_pm.c b/drivers/gpu/drm/i915/gt/selftest_engine_pm.c
new file mode 100644
index 000000000..0dcb3ed44
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_engine_pm.c
@@ -0,0 +1,423 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/sort.h>
+
+#include "i915_selftest.h"
+#include "intel_engine_regs.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt_clock_utils.h"
+#include "selftest_engine.h"
+#include "selftest_engine_heartbeat.h"
+#include "selftests/igt_atomic.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/igt_spinner.h"
+
+#define COUNT 5
+
+static int cmp_u64(const void *A, const void *B)
+{
+	const u64 *a = A, *b = B;
+
+	return *a - *b;
+}
+
+static u64 trifilter(u64 *a)
+{
+	sort(a, COUNT, sizeof(*a), cmp_u64, NULL);
+	return (a[1] + 2 * a[2] + a[3]) >> 2;
+}
+
+static u32 *emit_wait(u32 *cs, u32 offset, int op, u32 value)
+{
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		op;
+	*cs++ = value;
+	*cs++ = offset;
+	*cs++ = 0;
+
+	return cs;
+}
+
+static u32 *emit_store(u32 *cs, u32 offset, u32 value)
+{
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = offset;
+	*cs++ = 0;
+	*cs++ = value;
+
+	return cs;
+}
+
+static u32 *emit_srm(u32 *cs, i915_reg_t reg, u32 offset)
+{
+	*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+	*cs++ = i915_mmio_reg_offset(reg);
+	*cs++ = offset;
+	*cs++ = 0;
+
+	return cs;
+}
+
+static void write_semaphore(u32 *x, u32 value)
+{
+	WRITE_ONCE(*x, value);
+	wmb();
+}
+
+static int __measure_timestamps(struct intel_context *ce,
+				u64 *dt, u64 *d_ring, u64 *d_ctx)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	u32 *sema = memset32(engine->status_page.addr + 1000, 0, 5);
+	u32 offset = i915_ggtt_offset(engine->status_page.vma);
+	struct i915_request *rq;
+	u32 *cs;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 28);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	/* Signal & wait for start */
+	cs = emit_store(cs, offset + 4008, 1);
+	cs = emit_wait(cs, offset + 4008, MI_SEMAPHORE_SAD_NEQ_SDD, 1);
+
+	cs = emit_srm(cs, RING_TIMESTAMP(engine->mmio_base), offset + 4000);
+	cs = emit_srm(cs, RING_CTX_TIMESTAMP(engine->mmio_base), offset + 4004);
+
+	/* Busy wait */
+	cs = emit_wait(cs, offset + 4008, MI_SEMAPHORE_SAD_EQ_SDD, 1);
+
+	cs = emit_srm(cs, RING_TIMESTAMP(engine->mmio_base), offset + 4016);
+	cs = emit_srm(cs, RING_CTX_TIMESTAMP(engine->mmio_base), offset + 4012);
+
+	intel_ring_advance(rq, cs);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	intel_engine_flush_submission(engine);
+
+	/* Wait for the request to start executing, that then waits for us */
+	while (READ_ONCE(sema[2]) == 0)
+		cpu_relax();
+
+	/* Run the request for a 100us, sampling timestamps before/after */
+	local_irq_disable();
+	write_semaphore(&sema[2], 0);
+	while (READ_ONCE(sema[1]) == 0) /* wait for the gpu to catch up */
+		cpu_relax();
+	*dt = local_clock();
+	udelay(100);
+	*dt = local_clock() - *dt;
+	write_semaphore(&sema[2], 1);
+	local_irq_enable();
+
+	if (i915_request_wait(rq, 0, HZ / 2) < 0) {
+		i915_request_put(rq);
+		return -ETIME;
+	}
+	i915_request_put(rq);
+
+	pr_debug("%s CTX_TIMESTAMP: [%x, %x], RING_TIMESTAMP: [%x, %x]\n",
+		 engine->name, sema[1], sema[3], sema[0], sema[4]);
+
+	*d_ctx = sema[3] - sema[1];
+	*d_ring = sema[4] - sema[0];
+	return 0;
+}
+
+static int __live_engine_timestamps(struct intel_engine_cs *engine)
+{
+	u64 s_ring[COUNT], s_ctx[COUNT], st[COUNT], d_ring, d_ctx, dt;
+	struct intel_context *ce;
+	int i, err = 0;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	for (i = 0; i < COUNT; i++) {
+		err = __measure_timestamps(ce, &st[i], &s_ring[i], &s_ctx[i]);
+		if (err)
+			break;
+	}
+	intel_context_put(ce);
+	if (err)
+		return err;
+
+	dt = trifilter(st);
+	d_ring = trifilter(s_ring);
+	d_ctx = trifilter(s_ctx);
+
+	pr_info("%s elapsed:%lldns, CTX_TIMESTAMP:%lldns, RING_TIMESTAMP:%lldns\n",
+		engine->name, dt,
+		intel_gt_clock_interval_to_ns(engine->gt, d_ctx),
+		intel_gt_clock_interval_to_ns(engine->gt, d_ring));
+
+	d_ring = intel_gt_clock_interval_to_ns(engine->gt, d_ring);
+	if (3 * dt > 4 * d_ring || 4 * dt < 3 * d_ring) {
+		pr_err("%s Mismatch between ring timestamp and walltime!\n",
+		       engine->name);
+		return -EINVAL;
+	}
+
+	d_ring = trifilter(s_ring);
+	d_ctx = trifilter(s_ctx);
+
+	d_ctx *= engine->gt->clock_frequency;
+	if (GRAPHICS_VER(engine->i915) == 11)
+		d_ring *= 12500000; /* Fixed 80ns for GEN11 ctx timestamp? */
+	else
+		d_ring *= engine->gt->clock_frequency;
+
+	if (3 * d_ctx > 4 * d_ring || 4 * d_ctx < 3 * d_ring) {
+		pr_err("%s Mismatch between ring and context timestamps!\n",
+		       engine->name);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int live_engine_timestamps(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * Check that CS_TIMESTAMP / CTX_TIMESTAMP are in sync, i.e. share
+	 * the same CS clock.
+	 */
+
+	if (GRAPHICS_VER(gt->i915) < 8)
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		int err;
+
+		st_engine_heartbeat_disable(engine);
+		err = __live_engine_timestamps(engine);
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int __spin_until_busier(struct intel_engine_cs *engine, ktime_t busyness)
+{
+	ktime_t start, unused, dt;
+
+	if (!intel_engine_uses_guc(engine))
+		return 0;
+
+	/*
+	 * In GuC mode of submission, the busyness stats may get updated after
+	 * the batch starts running. Poll for a change in busyness and timeout
+	 * after 500 us.
+	 */
+	start = ktime_get();
+	while (intel_engine_get_busy_time(engine, &unused) == busyness) {
+		dt = ktime_get() - start;
+		if (dt > 10000000) {
+			pr_err("active wait timed out %lld\n", dt);
+			ENGINE_TRACE(engine, "active wait time out %lld\n", dt);
+			return -ETIME;
+		}
+	}
+
+	return 0;
+}
+
+static int live_engine_busy_stats(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	/*
+	 * Check that if an engine supports busy-stats, they tell the truth.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	GEM_BUG_ON(intel_gt_pm_is_awake(gt));
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq;
+		ktime_t busyness, dummy;
+		ktime_t de, dt;
+		ktime_t t[2];
+
+		if (!intel_engine_supports_stats(engine))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		if (intel_gt_pm_wait_for_idle(gt)) {
+			err = -EBUSY;
+			break;
+		}
+
+		st_engine_heartbeat_disable(engine);
+
+		ENGINE_TRACE(engine, "measuring idle time\n");
+		preempt_disable();
+		de = intel_engine_get_busy_time(engine, &t[0]);
+		udelay(100);
+		de = ktime_sub(intel_engine_get_busy_time(engine, &t[1]), de);
+		preempt_enable();
+		dt = ktime_sub(t[1], t[0]);
+		if (de < 0 || de > 10) {
+			pr_err("%s: reported %lldns [%d%%] busyness while sleeping [for %lldns]\n",
+			       engine->name,
+			       de, (int)div64_u64(100 * de, dt), dt);
+			GEM_TRACE_DUMP();
+			err = -EINVAL;
+			goto end;
+		}
+
+		/* 100% busy */
+		rq = igt_spinner_create_request(&spin,
+						engine->kernel_context,
+						MI_NOOP);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto end;
+		}
+		i915_request_add(rq);
+
+		busyness = intel_engine_get_busy_time(engine, &dummy);
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			intel_gt_set_wedged(engine->gt);
+			err = -ETIME;
+			goto end;
+		}
+
+		err = __spin_until_busier(engine, busyness);
+		if (err) {
+			GEM_TRACE_DUMP();
+			goto end;
+		}
+
+		ENGINE_TRACE(engine, "measuring busy time\n");
+		preempt_disable();
+		de = intel_engine_get_busy_time(engine, &t[0]);
+		mdelay(10);
+		de = ktime_sub(intel_engine_get_busy_time(engine, &t[1]), de);
+		preempt_enable();
+		dt = ktime_sub(t[1], t[0]);
+		if (100 * de < 95 * dt || 95 * de > 100 * dt) {
+			pr_err("%s: reported %lldns [%d%%] busyness while spinning [for %lldns]\n",
+			       engine->name,
+			       de, (int)div64_u64(100 * de, dt), dt);
+			GEM_TRACE_DUMP();
+			err = -EINVAL;
+			goto end;
+		}
+
+end:
+		st_engine_heartbeat_enable(engine);
+		igt_spinner_end(&spin);
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+	return err;
+}
+
+static int live_engine_pm(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * Check we can call intel_engine_pm_put from any context. No
+	 * failures are reported directly, but if we mess up lockdep should
+	 * tell us.
+	 */
+	if (intel_gt_pm_wait_for_idle(gt)) {
+		pr_err("Unable to flush GT pm before test\n");
+		return -EBUSY;
+	}
+
+	GEM_BUG_ON(intel_gt_pm_is_awake(gt));
+	for_each_engine(engine, gt, id) {
+		const typeof(*igt_atomic_phases) *p;
+
+		for (p = igt_atomic_phases; p->name; p++) {
+			/*
+			 * Acquisition is always synchronous, except if we
+			 * know that the engine is already awake, in which
+			 * case we should use intel_engine_pm_get_if_awake()
+			 * to atomically grab the wakeref.
+			 *
+			 * In practice,
+			 *    intel_engine_pm_get();
+			 *    intel_engine_pm_put();
+			 * occurs in one thread, while simultaneously
+			 *    intel_engine_pm_get_if_awake();
+			 *    intel_engine_pm_put();
+			 * occurs from atomic context in another.
+			 */
+			GEM_BUG_ON(intel_engine_pm_is_awake(engine));
+			intel_engine_pm_get(engine);
+
+			p->critical_section_begin();
+			if (!intel_engine_pm_get_if_awake(engine))
+				pr_err("intel_engine_pm_get_if_awake(%s) failed under %s\n",
+				       engine->name, p->name);
+			else
+				intel_engine_pm_put_async(engine);
+			intel_engine_pm_put_async(engine);
+			p->critical_section_end();
+
+			intel_engine_pm_flush(engine);
+
+			if (intel_engine_pm_is_awake(engine)) {
+				pr_err("%s is still awake after flushing pm\n",
+				       engine->name);
+				return -EINVAL;
+			}
+
+			/* gt wakeref is async (deferred to workqueue) */
+			if (intel_gt_pm_wait_for_idle(gt)) {
+				pr_err("GT failed to idle\n");
+				return -EINVAL;
+			}
+		}
+	}
+
+	return 0;
+}
+
+int live_engine_pm_selftests(struct intel_gt *gt)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_engine_timestamps),
+		SUBTEST(live_engine_busy_stats),
+		SUBTEST(live_engine_pm),
+	};
+
+	return intel_gt_live_subtests(tests, gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_execlists.c b/drivers/gpu/drm/i915/gt/selftest_execlists.c
new file mode 100644
index 000000000..b370411d4
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_execlists.c
@@ -0,0 +1,4521 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/prime_numbers.h>
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_engine_heartbeat.h"
+#include "gt/intel_reset.h"
+#include "gt/selftest_engine_heartbeat.h"
+
+#include "i915_selftest.h"
+#include "selftests/i915_random.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/igt_live_test.h"
+#include "selftests/igt_spinner.h"
+#include "selftests/lib_sw_fence.h"
+
+#include "gem/selftests/igt_gem_utils.h"
+#include "gem/selftests/mock_context.h"
+
+#define CS_GPR(engine, n) ((engine)->mmio_base + 0x600 + (n) * 4)
+#define NUM_GPR 16
+#define NUM_GPR_DW (NUM_GPR * 2) /* each GPR is 2 dwords */
+
+static bool is_active(struct i915_request *rq)
+{
+	if (i915_request_is_active(rq))
+		return true;
+
+	if (i915_request_on_hold(rq))
+		return true;
+
+	if (i915_request_has_initial_breadcrumb(rq) && i915_request_started(rq))
+		return true;
+
+	return false;
+}
+
+static int wait_for_submit(struct intel_engine_cs *engine,
+			   struct i915_request *rq,
+			   unsigned long timeout)
+{
+	/* Ignore our own attempts to suppress excess tasklets */
+	tasklet_hi_schedule(&engine->sched_engine->tasklet);
+
+	timeout += jiffies;
+	do {
+		bool done = time_after(jiffies, timeout);
+
+		if (i915_request_completed(rq)) /* that was quick! */
+			return 0;
+
+		/* Wait until the HW has acknowleged the submission (or err) */
+		intel_engine_flush_submission(engine);
+		if (!READ_ONCE(engine->execlists.pending[0]) && is_active(rq))
+			return 0;
+
+		if (done)
+			return -ETIME;
+
+		cond_resched();
+	} while (1);
+}
+
+static int wait_for_reset(struct intel_engine_cs *engine,
+			  struct i915_request *rq,
+			  unsigned long timeout)
+{
+	timeout += jiffies;
+
+	do {
+		cond_resched();
+		intel_engine_flush_submission(engine);
+
+		if (READ_ONCE(engine->execlists.pending[0]))
+			continue;
+
+		if (i915_request_completed(rq))
+			break;
+
+		if (READ_ONCE(rq->fence.error))
+			break;
+	} while (time_before(jiffies, timeout));
+
+	flush_scheduled_work();
+
+	if (rq->fence.error != -EIO) {
+		pr_err("%s: hanging request %llx:%lld not reset\n",
+		       engine->name,
+		       rq->fence.context,
+		       rq->fence.seqno);
+		return -EINVAL;
+	}
+
+	/* Give the request a jiffie to complete after flushing the worker */
+	if (i915_request_wait(rq, 0,
+			      max(0l, (long)(timeout - jiffies)) + 1) < 0) {
+		pr_err("%s: hanging request %llx:%lld did not complete\n",
+		       engine->name,
+		       rq->fence.context,
+		       rq->fence.seqno);
+		return -ETIME;
+	}
+
+	return 0;
+}
+
+static int live_sanitycheck(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	if (!HAS_LOGICAL_RING_CONTEXTS(gt->i915))
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			break;
+		}
+
+		rq = igt_spinner_create_request(&spin, ce, MI_NOOP);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_ctx;
+		}
+
+		i915_request_add(rq);
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			GEM_TRACE("spinner failed to start\n");
+			GEM_TRACE_DUMP();
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto out_ctx;
+		}
+
+		igt_spinner_end(&spin);
+		if (igt_flush_test(gt->i915)) {
+			err = -EIO;
+			goto out_ctx;
+		}
+
+out_ctx:
+		intel_context_put(ce);
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_unlite_restore(struct intel_gt *gt, int prio)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = -ENOMEM;
+
+	/*
+	 * Check that we can correctly context switch between 2 instances
+	 * on the same engine from the same parent context.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return err;
+
+	err = 0;
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce[2] = {};
+		struct i915_request *rq[2];
+		struct igt_live_test t;
+		int n;
+
+		if (prio && !intel_engine_has_preemption(engine))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			break;
+		}
+		st_engine_heartbeat_disable(engine);
+
+		for (n = 0; n < ARRAY_SIZE(ce); n++) {
+			struct intel_context *tmp;
+
+			tmp = intel_context_create(engine);
+			if (IS_ERR(tmp)) {
+				err = PTR_ERR(tmp);
+				goto err_ce;
+			}
+
+			err = intel_context_pin(tmp);
+			if (err) {
+				intel_context_put(tmp);
+				goto err_ce;
+			}
+
+			/*
+			 * Setup the pair of contexts such that if we
+			 * lite-restore using the RING_TAIL from ce[1] it
+			 * will execute garbage from ce[0]->ring.
+			 */
+			memset(tmp->ring->vaddr,
+			       POISON_INUSE, /* IPEHR: 0x5a5a5a5a [hung!] */
+			       tmp->ring->vma->size);
+
+			ce[n] = tmp;
+		}
+		GEM_BUG_ON(!ce[1]->ring->size);
+		intel_ring_reset(ce[1]->ring, ce[1]->ring->size / 2);
+		lrc_update_regs(ce[1], engine, ce[1]->ring->head);
+
+		rq[0] = igt_spinner_create_request(&spin, ce[0], MI_ARB_CHECK);
+		if (IS_ERR(rq[0])) {
+			err = PTR_ERR(rq[0]);
+			goto err_ce;
+		}
+
+		i915_request_get(rq[0]);
+		i915_request_add(rq[0]);
+		GEM_BUG_ON(rq[0]->postfix > ce[1]->ring->emit);
+
+		if (!igt_wait_for_spinner(&spin, rq[0])) {
+			i915_request_put(rq[0]);
+			goto err_ce;
+		}
+
+		rq[1] = i915_request_create(ce[1]);
+		if (IS_ERR(rq[1])) {
+			err = PTR_ERR(rq[1]);
+			i915_request_put(rq[0]);
+			goto err_ce;
+		}
+
+		if (!prio) {
+			/*
+			 * Ensure we do the switch to ce[1] on completion.
+			 *
+			 * rq[0] is already submitted, so this should reduce
+			 * to a no-op (a wait on a request on the same engine
+			 * uses the submit fence, not the completion fence),
+			 * but it will install a dependency on rq[1] for rq[0]
+			 * that will prevent the pair being reordered by
+			 * timeslicing.
+			 */
+			i915_request_await_dma_fence(rq[1], &rq[0]->fence);
+		}
+
+		i915_request_get(rq[1]);
+		i915_request_add(rq[1]);
+		GEM_BUG_ON(rq[1]->postfix <= rq[0]->postfix);
+		i915_request_put(rq[0]);
+
+		if (prio) {
+			struct i915_sched_attr attr = {
+				.priority = prio,
+			};
+
+			/* Alternatively preempt the spinner with ce[1] */
+			engine->sched_engine->schedule(rq[1], &attr);
+		}
+
+		/* And switch back to ce[0] for good measure */
+		rq[0] = i915_request_create(ce[0]);
+		if (IS_ERR(rq[0])) {
+			err = PTR_ERR(rq[0]);
+			i915_request_put(rq[1]);
+			goto err_ce;
+		}
+
+		i915_request_await_dma_fence(rq[0], &rq[1]->fence);
+		i915_request_get(rq[0]);
+		i915_request_add(rq[0]);
+		GEM_BUG_ON(rq[0]->postfix > rq[1]->postfix);
+		i915_request_put(rq[1]);
+		i915_request_put(rq[0]);
+
+err_ce:
+		intel_engine_flush_submission(engine);
+		igt_spinner_end(&spin);
+		for (n = 0; n < ARRAY_SIZE(ce); n++) {
+			if (IS_ERR_OR_NULL(ce[n]))
+				break;
+
+			intel_context_unpin(ce[n]);
+			intel_context_put(ce[n]);
+		}
+
+		st_engine_heartbeat_enable(engine);
+		if (igt_live_test_end(&t))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_unlite_switch(void *arg)
+{
+	return live_unlite_restore(arg, 0);
+}
+
+static int live_unlite_preempt(void *arg)
+{
+	return live_unlite_restore(arg, I915_PRIORITY_MAX);
+}
+
+static int live_unlite_ring(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct igt_spinner spin;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Setup a preemption event that will cause almost the entire ring
+	 * to be unwound, potentially fooling our intel_ring_direction()
+	 * into emitting a forward lite-restore instead of the rollback.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce[2] = {};
+		struct i915_request *rq;
+		struct igt_live_test t;
+		int n;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			break;
+		}
+		st_engine_heartbeat_disable(engine);
+
+		for (n = 0; n < ARRAY_SIZE(ce); n++) {
+			struct intel_context *tmp;
+
+			tmp = intel_context_create(engine);
+			if (IS_ERR(tmp)) {
+				err = PTR_ERR(tmp);
+				goto err_ce;
+			}
+
+			err = intel_context_pin(tmp);
+			if (err) {
+				intel_context_put(tmp);
+				goto err_ce;
+			}
+
+			memset32(tmp->ring->vaddr,
+				 0xdeadbeef, /* trigger a hang if executed */
+				 tmp->ring->vma->size / sizeof(u32));
+
+			ce[n] = tmp;
+		}
+
+		/* Create max prio spinner, followed by N low prio nops */
+		rq = igt_spinner_create_request(&spin, ce[0], MI_ARB_CHECK);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_ce;
+		}
+
+		i915_request_get(rq);
+		rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+		i915_request_add(rq);
+
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			intel_gt_set_wedged(gt);
+			i915_request_put(rq);
+			err = -ETIME;
+			goto err_ce;
+		}
+
+		/* Fill the ring, until we will cause a wrap */
+		n = 0;
+		while (intel_ring_direction(ce[0]->ring,
+					    rq->wa_tail,
+					    ce[0]->ring->tail) <= 0) {
+			struct i915_request *tmp;
+
+			tmp = intel_context_create_request(ce[0]);
+			if (IS_ERR(tmp)) {
+				err = PTR_ERR(tmp);
+				i915_request_put(rq);
+				goto err_ce;
+			}
+
+			i915_request_add(tmp);
+			intel_engine_flush_submission(engine);
+			n++;
+		}
+		intel_engine_flush_submission(engine);
+		pr_debug("%s: Filled ring with %d nop tails {size:%x, tail:%x, emit:%x, rq.tail:%x}\n",
+			 engine->name, n,
+			 ce[0]->ring->size,
+			 ce[0]->ring->tail,
+			 ce[0]->ring->emit,
+			 rq->tail);
+		GEM_BUG_ON(intel_ring_direction(ce[0]->ring,
+						rq->tail,
+						ce[0]->ring->tail) <= 0);
+		i915_request_put(rq);
+
+		/* Create a second ring to preempt the first ring after rq[0] */
+		rq = intel_context_create_request(ce[1]);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_ce;
+		}
+
+		rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		err = wait_for_submit(engine, rq, HZ / 2);
+		i915_request_put(rq);
+		if (err) {
+			pr_err("%s: preemption request was not submitted\n",
+			       engine->name);
+			err = -ETIME;
+		}
+
+		pr_debug("%s: ring[0]:{ tail:%x, emit:%x }, ring[1]:{ tail:%x, emit:%x }\n",
+			 engine->name,
+			 ce[0]->ring->tail, ce[0]->ring->emit,
+			 ce[1]->ring->tail, ce[1]->ring->emit);
+
+err_ce:
+		intel_engine_flush_submission(engine);
+		igt_spinner_end(&spin);
+		for (n = 0; n < ARRAY_SIZE(ce); n++) {
+			if (IS_ERR_OR_NULL(ce[n]))
+				break;
+
+			intel_context_unpin(ce[n]);
+			intel_context_put(ce[n]);
+		}
+		st_engine_heartbeat_enable(engine);
+		if (igt_live_test_end(&t))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_pin_rewind(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * We have to be careful not to trust intel_ring too much, for example
+	 * ring->head is updated upon retire which is out of sync with pinning
+	 * the context. Thus we cannot use ring->head to set CTX_RING_HEAD,
+	 * or else we risk writing an older, stale value.
+	 *
+	 * To simulate this, let's apply a bit of deliberate sabotague.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+		struct intel_ring *ring;
+		struct igt_live_test t;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			break;
+		}
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			break;
+		}
+
+		err = intel_context_pin(ce);
+		if (err) {
+			intel_context_put(ce);
+			break;
+		}
+
+		/* Keep the context awake while we play games */
+		err = i915_active_acquire(&ce->active);
+		if (err) {
+			intel_context_unpin(ce);
+			intel_context_put(ce);
+			break;
+		}
+		ring = ce->ring;
+
+		/* Poison the ring, and offset the next request from HEAD */
+		memset32(ring->vaddr, STACK_MAGIC, ring->size / sizeof(u32));
+		ring->emit = ring->size / 2;
+		ring->tail = ring->emit;
+		GEM_BUG_ON(ring->head);
+
+		intel_context_unpin(ce);
+
+		/* Submit a simple nop request */
+		GEM_BUG_ON(intel_context_is_pinned(ce));
+		rq = intel_context_create_request(ce);
+		i915_active_release(&ce->active); /* e.g. async retire */
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+		GEM_BUG_ON(!rq->head);
+		i915_request_add(rq);
+
+		/* Expect not to hang! */
+		if (igt_live_test_end(&t)) {
+			err = -EIO;
+			break;
+		}
+	}
+
+	return err;
+}
+
+static int engine_lock_reset_tasklet(struct intel_engine_cs *engine)
+{
+	tasklet_disable(&engine->sched_engine->tasklet);
+	local_bh_disable();
+
+	if (test_and_set_bit(I915_RESET_ENGINE + engine->id,
+			     &engine->gt->reset.flags)) {
+		local_bh_enable();
+		tasklet_enable(&engine->sched_engine->tasklet);
+
+		intel_gt_set_wedged(engine->gt);
+		return -EBUSY;
+	}
+
+	return 0;
+}
+
+static void engine_unlock_reset_tasklet(struct intel_engine_cs *engine)
+{
+	clear_and_wake_up_bit(I915_RESET_ENGINE + engine->id,
+			      &engine->gt->reset.flags);
+
+	local_bh_enable();
+	tasklet_enable(&engine->sched_engine->tasklet);
+}
+
+static int live_hold_reset(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	/*
+	 * In order to support offline error capture for fast preempt reset,
+	 * we need to decouple the guilty request and ensure that it and its
+	 * descendents are not executed while the capture is in progress.
+	 */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			break;
+		}
+
+		st_engine_heartbeat_disable(engine);
+
+		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out;
+		}
+		i915_request_add(rq);
+
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			intel_gt_set_wedged(gt);
+			err = -ETIME;
+			goto out;
+		}
+
+		/* We have our request executing, now remove it and reset */
+
+		err = engine_lock_reset_tasklet(engine);
+		if (err)
+			goto out;
+
+		engine->sched_engine->tasklet.callback(&engine->sched_engine->tasklet);
+		GEM_BUG_ON(execlists_active(&engine->execlists) != rq);
+
+		i915_request_get(rq);
+		execlists_hold(engine, rq);
+		GEM_BUG_ON(!i915_request_on_hold(rq));
+
+		__intel_engine_reset_bh(engine, NULL);
+		GEM_BUG_ON(rq->fence.error != -EIO);
+
+		engine_unlock_reset_tasklet(engine);
+
+		/* Check that we do not resubmit the held request */
+		if (!i915_request_wait(rq, 0, HZ / 5)) {
+			pr_err("%s: on hold request completed!\n",
+			       engine->name);
+			i915_request_put(rq);
+			err = -EIO;
+			goto out;
+		}
+		GEM_BUG_ON(!i915_request_on_hold(rq));
+
+		/* But is resubmitted on release */
+		execlists_unhold(engine, rq);
+		if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+			pr_err("%s: held request did not complete!\n",
+			       engine->name);
+			intel_gt_set_wedged(gt);
+			err = -ETIME;
+		}
+		i915_request_put(rq);
+
+out:
+		st_engine_heartbeat_enable(engine);
+		intel_context_put(ce);
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static const char *error_repr(int err)
+{
+	return err ? "bad" : "good";
+}
+
+static int live_error_interrupt(void *arg)
+{
+	static const struct error_phase {
+		enum { GOOD = 0, BAD = -EIO } error[2];
+	} phases[] = {
+		{ { BAD,  GOOD } },
+		{ { BAD,  BAD  } },
+		{ { BAD,  GOOD } },
+		{ { GOOD, GOOD } }, /* sentinel */
+	};
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * We hook up the CS_MASTER_ERROR_INTERRUPT to have forewarning
+	 * of invalid commands in user batches that will cause a GPU hang.
+	 * This is a faster mechanism than using hangcheck/heartbeats, but
+	 * only detects problems the HW knows about -- it will not warn when
+	 * we kill the HW!
+	 *
+	 * To verify our detection and reset, we throw some invalid commands
+	 * at the HW and wait for the interrupt.
+	 */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		const struct error_phase *p;
+		int err = 0;
+
+		st_engine_heartbeat_disable(engine);
+
+		for (p = phases; p->error[0] != GOOD; p++) {
+			struct i915_request *client[ARRAY_SIZE(phases->error)];
+			u32 *cs;
+			int i;
+
+			memset(client, 0, sizeof(*client));
+			for (i = 0; i < ARRAY_SIZE(client); i++) {
+				struct intel_context *ce;
+				struct i915_request *rq;
+
+				ce = intel_context_create(engine);
+				if (IS_ERR(ce)) {
+					err = PTR_ERR(ce);
+					goto out;
+				}
+
+				rq = intel_context_create_request(ce);
+				intel_context_put(ce);
+				if (IS_ERR(rq)) {
+					err = PTR_ERR(rq);
+					goto out;
+				}
+
+				if (rq->engine->emit_init_breadcrumb) {
+					err = rq->engine->emit_init_breadcrumb(rq);
+					if (err) {
+						i915_request_add(rq);
+						goto out;
+					}
+				}
+
+				cs = intel_ring_begin(rq, 2);
+				if (IS_ERR(cs)) {
+					i915_request_add(rq);
+					err = PTR_ERR(cs);
+					goto out;
+				}
+
+				if (p->error[i]) {
+					*cs++ = 0xdeadbeef;
+					*cs++ = 0xdeadbeef;
+				} else {
+					*cs++ = MI_NOOP;
+					*cs++ = MI_NOOP;
+				}
+
+				client[i] = i915_request_get(rq);
+				i915_request_add(rq);
+			}
+
+			err = wait_for_submit(engine, client[0], HZ / 2);
+			if (err) {
+				pr_err("%s: first request did not start within time!\n",
+				       engine->name);
+				err = -ETIME;
+				goto out;
+			}
+
+			for (i = 0; i < ARRAY_SIZE(client); i++) {
+				if (i915_request_wait(client[i], 0, HZ / 5) < 0)
+					pr_debug("%s: %s request incomplete!\n",
+						 engine->name,
+						 error_repr(p->error[i]));
+
+				if (!i915_request_started(client[i])) {
+					pr_err("%s: %s request not started!\n",
+					       engine->name,
+					       error_repr(p->error[i]));
+					err = -ETIME;
+					goto out;
+				}
+
+				/* Kick the tasklet to process the error */
+				intel_engine_flush_submission(engine);
+				if (client[i]->fence.error != p->error[i]) {
+					pr_err("%s: %s request (%s) with wrong error code: %d\n",
+					       engine->name,
+					       error_repr(p->error[i]),
+					       i915_request_completed(client[i]) ? "completed" : "running",
+					       client[i]->fence.error);
+					err = -EINVAL;
+					goto out;
+				}
+			}
+
+out:
+			for (i = 0; i < ARRAY_SIZE(client); i++)
+				if (client[i])
+					i915_request_put(client[i]);
+			if (err) {
+				pr_err("%s: failed at phase[%zd] { %d, %d }\n",
+				       engine->name, p - phases,
+				       p->error[0], p->error[1]);
+				break;
+			}
+		}
+
+		st_engine_heartbeat_enable(engine);
+		if (err) {
+			intel_gt_set_wedged(gt);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+static int
+emit_semaphore_chain(struct i915_request *rq, struct i915_vma *vma, int idx)
+{
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 10);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_NEQ_SDD;
+	*cs++ = 0;
+	*cs++ = i915_ggtt_offset(vma) + 4 * idx;
+	*cs++ = 0;
+
+	if (idx > 0) {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = i915_ggtt_offset(vma) + 4 * (idx - 1);
+		*cs++ = 0;
+		*cs++ = 1;
+	} else {
+		*cs++ = MI_NOOP;
+		*cs++ = MI_NOOP;
+		*cs++ = MI_NOOP;
+		*cs++ = MI_NOOP;
+	}
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
+	intel_ring_advance(rq, cs);
+	return 0;
+}
+
+static struct i915_request *
+semaphore_queue(struct intel_engine_cs *engine, struct i915_vma *vma, int idx)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+	int err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return ERR_CAST(ce);
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		goto out_ce;
+
+	err = 0;
+	if (rq->engine->emit_init_breadcrumb)
+		err = rq->engine->emit_init_breadcrumb(rq);
+	if (err == 0)
+		err = emit_semaphore_chain(rq, vma, idx);
+	if (err == 0)
+		i915_request_get(rq);
+	i915_request_add(rq);
+	if (err)
+		rq = ERR_PTR(err);
+
+out_ce:
+	intel_context_put(ce);
+	return rq;
+}
+
+static int
+release_queue(struct intel_engine_cs *engine,
+	      struct i915_vma *vma,
+	      int idx, int prio)
+{
+	struct i915_sched_attr attr = {
+		.priority = prio,
+	};
+	struct i915_request *rq;
+	u32 *cs;
+
+	rq = intel_engine_create_kernel_request(engine);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = i915_ggtt_offset(vma) + 4 * (idx - 1);
+	*cs++ = 0;
+	*cs++ = 1;
+
+	intel_ring_advance(rq, cs);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	local_bh_disable();
+	engine->sched_engine->schedule(rq, &attr);
+	local_bh_enable(); /* kick tasklet */
+
+	i915_request_put(rq);
+
+	return 0;
+}
+
+static int
+slice_semaphore_queue(struct intel_engine_cs *outer,
+		      struct i915_vma *vma,
+		      int count)
+{
+	struct intel_engine_cs *engine;
+	struct i915_request *head;
+	enum intel_engine_id id;
+	int err, i, n = 0;
+
+	head = semaphore_queue(outer, vma, n++);
+	if (IS_ERR(head))
+		return PTR_ERR(head);
+
+	for_each_engine(engine, outer->gt, id) {
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		for (i = 0; i < count; i++) {
+			struct i915_request *rq;
+
+			rq = semaphore_queue(engine, vma, n++);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				goto out;
+			}
+
+			i915_request_put(rq);
+		}
+	}
+
+	err = release_queue(outer, vma, n, I915_PRIORITY_BARRIER);
+	if (err)
+		goto out;
+
+	if (i915_request_wait(head, 0,
+			      2 * outer->gt->info.num_engines * (count + 2) * (count + 3)) < 0) {
+		pr_err("%s: Failed to slice along semaphore chain of length (%d, %d)!\n",
+		       outer->name, count, n);
+		GEM_TRACE_DUMP();
+		intel_gt_set_wedged(outer->gt);
+		err = -EIO;
+	}
+
+out:
+	i915_request_put(head);
+	return err;
+}
+
+static int live_timeslice_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct drm_i915_gem_object *obj;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct i915_vma *vma;
+	void *vaddr;
+	int err = 0;
+
+	/*
+	 * If a request takes too long, we would like to give other users
+	 * a fair go on the GPU. In particular, users may create batches
+	 * that wait upon external input, where that input may even be
+	 * supplied by another GPU job. To avoid blocking forever, we
+	 * need to preempt the current task and replace it with another
+	 * ready task.
+	 */
+	if (!CONFIG_DRM_I915_TIMESLICE_DURATION)
+		return 0;
+
+	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	vaddr = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto err_obj;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+	if (err)
+		goto err_map;
+
+	err = i915_vma_sync(vma);
+	if (err)
+		goto err_pin;
+
+	for_each_engine(engine, gt, id) {
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		memset(vaddr, 0, PAGE_SIZE);
+
+		st_engine_heartbeat_disable(engine);
+		err = slice_semaphore_queue(engine, vma, 5);
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			goto err_pin;
+
+		if (igt_flush_test(gt->i915)) {
+			err = -EIO;
+			goto err_pin;
+		}
+	}
+
+err_pin:
+	i915_vma_unpin(vma);
+err_map:
+	i915_gem_object_unpin_map(obj);
+err_obj:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static struct i915_request *
+create_rewinder(struct intel_context *ce,
+		struct i915_request *wait,
+		void *slot, int idx)
+{
+	const u32 offset =
+		i915_ggtt_offset(ce->engine->status_page.vma) +
+		offset_in_page(slot);
+	struct i915_request *rq;
+	u32 *cs;
+	int err;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return rq;
+
+	if (wait) {
+		err = i915_request_await_dma_fence(rq, &wait->fence);
+		if (err)
+			goto err;
+	}
+
+	cs = intel_ring_begin(rq, 14);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err;
+	}
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	*cs++ = MI_NOOP;
+
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_GTE_SDD;
+	*cs++ = idx;
+	*cs++ = offset;
+	*cs++ = 0;
+
+	*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+	*cs++ = i915_mmio_reg_offset(RING_TIMESTAMP(rq->engine->mmio_base));
+	*cs++ = offset + idx * sizeof(u32);
+	*cs++ = 0;
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = offset;
+	*cs++ = 0;
+	*cs++ = idx + 1;
+
+	intel_ring_advance(rq, cs);
+
+	err = 0;
+err:
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (err) {
+		i915_request_put(rq);
+		return ERR_PTR(err);
+	}
+
+	return rq;
+}
+
+static int live_timeslice_rewind(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * The usual presumption on timeslice expiration is that we replace
+	 * the active context with another. However, given a chain of
+	 * dependencies we may end up with replacing the context with itself,
+	 * but only a few of those requests, forcing us to rewind the
+	 * RING_TAIL of the original request.
+	 */
+	if (!CONFIG_DRM_I915_TIMESLICE_DURATION)
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		enum { A1, A2, B1 };
+		enum { X = 1, Z, Y };
+		struct i915_request *rq[3] = {};
+		struct intel_context *ce;
+		unsigned long timeslice;
+		int i, err = 0;
+		u32 *slot;
+
+		if (!intel_engine_has_timeslices(engine))
+			continue;
+
+		/*
+		 * A:rq1 -- semaphore wait, timestamp X
+		 * A:rq2 -- write timestamp Y
+		 *
+		 * B:rq1 [await A:rq1] -- write timestamp Z
+		 *
+		 * Force timeslice, release semaphore.
+		 *
+		 * Expect execution/evaluation order XZY
+		 */
+
+		st_engine_heartbeat_disable(engine);
+		timeslice = xchg(&engine->props.timeslice_duration_ms, 1);
+
+		slot = memset32(engine->status_page.addr + 1000, 0, 4);
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto err;
+		}
+
+		rq[A1] = create_rewinder(ce, NULL, slot, X);
+		if (IS_ERR(rq[A1])) {
+			intel_context_put(ce);
+			goto err;
+		}
+
+		rq[A2] = create_rewinder(ce, NULL, slot, Y);
+		intel_context_put(ce);
+		if (IS_ERR(rq[A2]))
+			goto err;
+
+		err = wait_for_submit(engine, rq[A2], HZ / 2);
+		if (err) {
+			pr_err("%s: failed to submit first context\n",
+			       engine->name);
+			goto err;
+		}
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto err;
+		}
+
+		rq[B1] = create_rewinder(ce, rq[A1], slot, Z);
+		intel_context_put(ce);
+		if (IS_ERR(rq[2]))
+			goto err;
+
+		err = wait_for_submit(engine, rq[B1], HZ / 2);
+		if (err) {
+			pr_err("%s: failed to submit second context\n",
+			       engine->name);
+			goto err;
+		}
+
+		/* ELSP[] = { { A:rq1, A:rq2 }, { B:rq1 } } */
+		ENGINE_TRACE(engine, "forcing tasklet for rewind\n");
+		while (i915_request_is_active(rq[A2])) { /* semaphore yield! */
+			/* Wait for the timeslice to kick in */
+			del_timer(&engine->execlists.timer);
+			tasklet_hi_schedule(&engine->sched_engine->tasklet);
+			intel_engine_flush_submission(engine);
+		}
+		/* -> ELSP[] = { { A:rq1 }, { B:rq1 } } */
+		GEM_BUG_ON(!i915_request_is_active(rq[A1]));
+		GEM_BUG_ON(!i915_request_is_active(rq[B1]));
+		GEM_BUG_ON(i915_request_is_active(rq[A2]));
+
+		/* Release the hounds! */
+		slot[0] = 1;
+		wmb(); /* "pairs" with GPU; paranoid kick of internal CPU$ */
+
+		for (i = 1; i <= 3; i++) {
+			unsigned long timeout = jiffies + HZ / 2;
+
+			while (!READ_ONCE(slot[i]) &&
+			       time_before(jiffies, timeout))
+				;
+
+			if (!time_before(jiffies, timeout)) {
+				pr_err("%s: rq[%d] timed out\n",
+				       engine->name, i - 1);
+				err = -ETIME;
+				goto err;
+			}
+
+			pr_debug("%s: slot[%d]:%x\n", engine->name, i, slot[i]);
+		}
+
+		/* XZY: XZ < XY */
+		if (slot[Z] - slot[X] >= slot[Y] - slot[X]) {
+			pr_err("%s: timeslicing did not run context B [%u] before A [%u]!\n",
+			       engine->name,
+			       slot[Z] - slot[X],
+			       slot[Y] - slot[X]);
+			err = -EINVAL;
+		}
+
+err:
+		memset32(&slot[0], -1, 4);
+		wmb();
+
+		engine->props.timeslice_duration_ms = timeslice;
+		st_engine_heartbeat_enable(engine);
+		for (i = 0; i < 3; i++)
+			i915_request_put(rq[i]);
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static struct i915_request *nop_request(struct intel_engine_cs *engine)
+{
+	struct i915_request *rq;
+
+	rq = intel_engine_create_kernel_request(engine);
+	if (IS_ERR(rq))
+		return rq;
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	return rq;
+}
+
+static long slice_timeout(struct intel_engine_cs *engine)
+{
+	long timeout;
+
+	/* Enough time for a timeslice to kick in, and kick out */
+	timeout = 2 * msecs_to_jiffies_timeout(timeslice(engine));
+
+	/* Enough time for the nop request to complete */
+	timeout += HZ / 5;
+
+	return timeout + 1;
+}
+
+static int live_timeslice_queue(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct drm_i915_gem_object *obj;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct i915_vma *vma;
+	void *vaddr;
+	int err = 0;
+
+	/*
+	 * Make sure that even if ELSP[0] and ELSP[1] are filled with
+	 * timeslicing between them disabled, we *do* enable timeslicing
+	 * if the queue demands it. (Normally, we do not submit if
+	 * ELSP[1] is already occupied, so must rely on timeslicing to
+	 * eject ELSP[0] in favour of the queue.)
+	 */
+	if (!CONFIG_DRM_I915_TIMESLICE_DURATION)
+		return 0;
+
+	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	vaddr = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto err_obj;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+	if (err)
+		goto err_map;
+
+	err = i915_vma_sync(vma);
+	if (err)
+		goto err_pin;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_sched_attr attr = { .priority = I915_PRIORITY_MAX };
+		struct i915_request *rq, *nop;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		st_engine_heartbeat_disable(engine);
+		memset(vaddr, 0, PAGE_SIZE);
+
+		/* ELSP[0]: semaphore wait */
+		rq = semaphore_queue(engine, vma, 0);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_heartbeat;
+		}
+		engine->sched_engine->schedule(rq, &attr);
+		err = wait_for_submit(engine, rq, HZ / 2);
+		if (err) {
+			pr_err("%s: Timed out trying to submit semaphores\n",
+			       engine->name);
+			goto err_rq;
+		}
+
+		/* ELSP[1]: nop request */
+		nop = nop_request(engine);
+		if (IS_ERR(nop)) {
+			err = PTR_ERR(nop);
+			goto err_rq;
+		}
+		err = wait_for_submit(engine, nop, HZ / 2);
+		i915_request_put(nop);
+		if (err) {
+			pr_err("%s: Timed out trying to submit nop\n",
+			       engine->name);
+			goto err_rq;
+		}
+
+		GEM_BUG_ON(i915_request_completed(rq));
+		GEM_BUG_ON(execlists_active(&engine->execlists) != rq);
+
+		/* Queue: semaphore signal, matching priority as semaphore */
+		err = release_queue(engine, vma, 1, effective_prio(rq));
+		if (err)
+			goto err_rq;
+
+		/* Wait until we ack the release_queue and start timeslicing */
+		do {
+			cond_resched();
+			intel_engine_flush_submission(engine);
+		} while (READ_ONCE(engine->execlists.pending[0]));
+
+		/* Timeslice every jiffy, so within 2 we should signal */
+		if (i915_request_wait(rq, 0, slice_timeout(engine)) < 0) {
+			struct drm_printer p =
+				drm_info_printer(gt->i915->drm.dev);
+
+			pr_err("%s: Failed to timeslice into queue\n",
+			       engine->name);
+			intel_engine_dump(engine, &p,
+					  "%s\n", engine->name);
+
+			memset(vaddr, 0xff, PAGE_SIZE);
+			err = -EIO;
+		}
+err_rq:
+		i915_request_put(rq);
+err_heartbeat:
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			break;
+	}
+
+err_pin:
+	i915_vma_unpin(vma);
+err_map:
+	i915_gem_object_unpin_map(obj);
+err_obj:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static int live_timeslice_nopreempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	/*
+	 * We should not timeslice into a request that is marked with
+	 * I915_REQUEST_NOPREEMPT.
+	 */
+	if (!CONFIG_DRM_I915_TIMESLICE_DURATION)
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+		unsigned long timeslice;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			break;
+		}
+
+		st_engine_heartbeat_disable(engine);
+		timeslice = xchg(&engine->props.timeslice_duration_ms, 1);
+
+		/* Create an unpreemptible spinner */
+
+		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_heartbeat;
+		}
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			i915_request_put(rq);
+			err = -ETIME;
+			goto out_spin;
+		}
+
+		set_bit(I915_FENCE_FLAG_NOPREEMPT, &rq->fence.flags);
+		i915_request_put(rq);
+
+		/* Followed by a maximum priority barrier (heartbeat) */
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out_spin;
+		}
+
+		rq = intel_context_create_request(ce);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_spin;
+		}
+
+		rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		/*
+		 * Wait until the barrier is in ELSP, and we know timeslicing
+		 * will have been activated.
+		 */
+		if (wait_for_submit(engine, rq, HZ / 2)) {
+			i915_request_put(rq);
+			err = -ETIME;
+			goto out_spin;
+		}
+
+		/*
+		 * Since the ELSP[0] request is unpreemptible, it should not
+		 * allow the maximum priority barrier through. Wait long
+		 * enough to see if it is timesliced in by mistake.
+		 */
+		if (i915_request_wait(rq, 0, slice_timeout(engine)) >= 0) {
+			pr_err("%s: I915_PRIORITY_BARRIER request completed, bypassing no-preempt request\n",
+			       engine->name);
+			err = -EINVAL;
+		}
+		i915_request_put(rq);
+
+out_spin:
+		igt_spinner_end(&spin);
+out_heartbeat:
+		xchg(&engine->props.timeslice_duration_ms, timeslice);
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			break;
+
+		if (igt_flush_test(gt->i915)) {
+			err = -EIO;
+			break;
+		}
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_busywait_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gem_context *ctx_hi, *ctx_lo;
+	struct intel_engine_cs *engine;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	enum intel_engine_id id;
+	u32 *map;
+	int err;
+
+	/*
+	 * Verify that even without HAS_LOGICAL_RING_PREEMPTION, we can
+	 * preempt the busywaits used to synchronise between rings.
+	 */
+
+	ctx_hi = kernel_context(gt->i915, NULL);
+	if (IS_ERR(ctx_hi))
+		return PTR_ERR(ctx_hi);
+
+	ctx_hi->sched.priority = I915_CONTEXT_MAX_USER_PRIORITY;
+
+	ctx_lo = kernel_context(gt->i915, NULL);
+	if (IS_ERR(ctx_lo)) {
+		err = PTR_ERR(ctx_lo);
+		goto err_ctx_hi;
+	}
+
+	ctx_lo->sched.priority = I915_CONTEXT_MIN_USER_PRIORITY;
+
+	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto err_ctx_lo;
+	}
+
+	map = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(map)) {
+		err = PTR_ERR(map);
+		goto err_obj;
+	}
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_map;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+	if (err)
+		goto err_map;
+
+	err = i915_vma_sync(vma);
+	if (err)
+		goto err_vma;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *lo, *hi;
+		struct igt_live_test t;
+		u32 *cs;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			goto err_vma;
+		}
+
+		/*
+		 * We create two requests. The low priority request
+		 * busywaits on a semaphore (inside the ringbuffer where
+		 * is should be preemptible) and the high priority requests
+		 * uses a MI_STORE_DWORD_IMM to update the semaphore value
+		 * allowing the first request to complete. If preemption
+		 * fails, we hang instead.
+		 */
+
+		lo = igt_request_alloc(ctx_lo, engine);
+		if (IS_ERR(lo)) {
+			err = PTR_ERR(lo);
+			goto err_vma;
+		}
+
+		cs = intel_ring_begin(lo, 8);
+		if (IS_ERR(cs)) {
+			err = PTR_ERR(cs);
+			i915_request_add(lo);
+			goto err_vma;
+		}
+
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = i915_ggtt_offset(vma);
+		*cs++ = 0;
+		*cs++ = 1;
+
+		/* XXX Do we need a flush + invalidate here? */
+
+		*cs++ = MI_SEMAPHORE_WAIT |
+			MI_SEMAPHORE_GLOBAL_GTT |
+			MI_SEMAPHORE_POLL |
+			MI_SEMAPHORE_SAD_EQ_SDD;
+		*cs++ = 0;
+		*cs++ = i915_ggtt_offset(vma);
+		*cs++ = 0;
+
+		intel_ring_advance(lo, cs);
+
+		i915_request_get(lo);
+		i915_request_add(lo);
+
+		if (wait_for(READ_ONCE(*map), 10)) {
+			i915_request_put(lo);
+			err = -ETIMEDOUT;
+			goto err_vma;
+		}
+
+		/* Low priority request should be busywaiting now */
+		if (i915_request_wait(lo, 0, 1) != -ETIME) {
+			i915_request_put(lo);
+			pr_err("%s: Busywaiting request did not!\n",
+			       engine->name);
+			err = -EIO;
+			goto err_vma;
+		}
+
+		hi = igt_request_alloc(ctx_hi, engine);
+		if (IS_ERR(hi)) {
+			err = PTR_ERR(hi);
+			i915_request_put(lo);
+			goto err_vma;
+		}
+
+		cs = intel_ring_begin(hi, 4);
+		if (IS_ERR(cs)) {
+			err = PTR_ERR(cs);
+			i915_request_add(hi);
+			i915_request_put(lo);
+			goto err_vma;
+		}
+
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = i915_ggtt_offset(vma);
+		*cs++ = 0;
+		*cs++ = 0;
+
+		intel_ring_advance(hi, cs);
+		i915_request_add(hi);
+
+		if (i915_request_wait(lo, 0, HZ / 5) < 0) {
+			struct drm_printer p = drm_info_printer(gt->i915->drm.dev);
+
+			pr_err("%s: Failed to preempt semaphore busywait!\n",
+			       engine->name);
+
+			intel_engine_dump(engine, &p, "%s\n", engine->name);
+			GEM_TRACE_DUMP();
+
+			i915_request_put(lo);
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto err_vma;
+		}
+		GEM_BUG_ON(READ_ONCE(*map));
+		i915_request_put(lo);
+
+		if (igt_live_test_end(&t)) {
+			err = -EIO;
+			goto err_vma;
+		}
+	}
+
+	err = 0;
+err_vma:
+	i915_vma_unpin(vma);
+err_map:
+	i915_gem_object_unpin_map(obj);
+err_obj:
+	i915_gem_object_put(obj);
+err_ctx_lo:
+	kernel_context_close(ctx_lo);
+err_ctx_hi:
+	kernel_context_close(ctx_hi);
+	return err;
+}
+
+static struct i915_request *
+spinner_create_request(struct igt_spinner *spin,
+		       struct i915_gem_context *ctx,
+		       struct intel_engine_cs *engine,
+		       u32 arb)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+
+	ce = i915_gem_context_get_engine(ctx, engine->legacy_idx);
+	if (IS_ERR(ce))
+		return ERR_CAST(ce);
+
+	rq = igt_spinner_create_request(spin, ce, arb);
+	intel_context_put(ce);
+	return rq;
+}
+
+static int live_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gem_context *ctx_hi, *ctx_lo;
+	struct igt_spinner spin_hi, spin_lo;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	ctx_hi = kernel_context(gt->i915, NULL);
+	if (!ctx_hi)
+		return -ENOMEM;
+	ctx_hi->sched.priority = I915_CONTEXT_MAX_USER_PRIORITY;
+
+	ctx_lo = kernel_context(gt->i915, NULL);
+	if (!ctx_lo)
+		goto err_ctx_hi;
+	ctx_lo->sched.priority = I915_CONTEXT_MIN_USER_PRIORITY;
+
+	if (igt_spinner_init(&spin_hi, gt))
+		goto err_ctx_lo;
+
+	if (igt_spinner_init(&spin_lo, gt))
+		goto err_spin_hi;
+
+	for_each_engine(engine, gt, id) {
+		struct igt_live_test t;
+		struct i915_request *rq;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			goto err_spin_lo;
+		}
+
+		rq = spinner_create_request(&spin_lo, ctx_lo, engine,
+					    MI_ARB_CHECK);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_spin_lo;
+		}
+
+		i915_request_add(rq);
+		if (!igt_wait_for_spinner(&spin_lo, rq)) {
+			GEM_TRACE("lo spinner failed to start\n");
+			GEM_TRACE_DUMP();
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto err_spin_lo;
+		}
+
+		rq = spinner_create_request(&spin_hi, ctx_hi, engine,
+					    MI_ARB_CHECK);
+		if (IS_ERR(rq)) {
+			igt_spinner_end(&spin_lo);
+			err = PTR_ERR(rq);
+			goto err_spin_lo;
+		}
+
+		i915_request_add(rq);
+		if (!igt_wait_for_spinner(&spin_hi, rq)) {
+			GEM_TRACE("hi spinner failed to start\n");
+			GEM_TRACE_DUMP();
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto err_spin_lo;
+		}
+
+		igt_spinner_end(&spin_hi);
+		igt_spinner_end(&spin_lo);
+
+		if (igt_live_test_end(&t)) {
+			err = -EIO;
+			goto err_spin_lo;
+		}
+	}
+
+	err = 0;
+err_spin_lo:
+	igt_spinner_fini(&spin_lo);
+err_spin_hi:
+	igt_spinner_fini(&spin_hi);
+err_ctx_lo:
+	kernel_context_close(ctx_lo);
+err_ctx_hi:
+	kernel_context_close(ctx_hi);
+	return err;
+}
+
+static int live_late_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gem_context *ctx_hi, *ctx_lo;
+	struct igt_spinner spin_hi, spin_lo;
+	struct intel_engine_cs *engine;
+	struct i915_sched_attr attr = {};
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	ctx_hi = kernel_context(gt->i915, NULL);
+	if (!ctx_hi)
+		return -ENOMEM;
+
+	ctx_lo = kernel_context(gt->i915, NULL);
+	if (!ctx_lo)
+		goto err_ctx_hi;
+
+	if (igt_spinner_init(&spin_hi, gt))
+		goto err_ctx_lo;
+
+	if (igt_spinner_init(&spin_lo, gt))
+		goto err_spin_hi;
+
+	/* Make sure ctx_lo stays before ctx_hi until we trigger preemption. */
+	ctx_lo->sched.priority = 1;
+
+	for_each_engine(engine, gt, id) {
+		struct igt_live_test t;
+		struct i915_request *rq;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			goto err_spin_lo;
+		}
+
+		rq = spinner_create_request(&spin_lo, ctx_lo, engine,
+					    MI_ARB_CHECK);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_spin_lo;
+		}
+
+		i915_request_add(rq);
+		if (!igt_wait_for_spinner(&spin_lo, rq)) {
+			pr_err("First context failed to start\n");
+			goto err_wedged;
+		}
+
+		rq = spinner_create_request(&spin_hi, ctx_hi, engine,
+					    MI_NOOP);
+		if (IS_ERR(rq)) {
+			igt_spinner_end(&spin_lo);
+			err = PTR_ERR(rq);
+			goto err_spin_lo;
+		}
+
+		i915_request_add(rq);
+		if (igt_wait_for_spinner(&spin_hi, rq)) {
+			pr_err("Second context overtook first?\n");
+			goto err_wedged;
+		}
+
+		attr.priority = I915_PRIORITY_MAX;
+		engine->sched_engine->schedule(rq, &attr);
+
+		if (!igt_wait_for_spinner(&spin_hi, rq)) {
+			pr_err("High priority context failed to preempt the low priority context\n");
+			GEM_TRACE_DUMP();
+			goto err_wedged;
+		}
+
+		igt_spinner_end(&spin_hi);
+		igt_spinner_end(&spin_lo);
+
+		if (igt_live_test_end(&t)) {
+			err = -EIO;
+			goto err_spin_lo;
+		}
+	}
+
+	err = 0;
+err_spin_lo:
+	igt_spinner_fini(&spin_lo);
+err_spin_hi:
+	igt_spinner_fini(&spin_hi);
+err_ctx_lo:
+	kernel_context_close(ctx_lo);
+err_ctx_hi:
+	kernel_context_close(ctx_hi);
+	return err;
+
+err_wedged:
+	igt_spinner_end(&spin_hi);
+	igt_spinner_end(&spin_lo);
+	intel_gt_set_wedged(gt);
+	err = -EIO;
+	goto err_spin_lo;
+}
+
+struct preempt_client {
+	struct igt_spinner spin;
+	struct i915_gem_context *ctx;
+};
+
+static int preempt_client_init(struct intel_gt *gt, struct preempt_client *c)
+{
+	c->ctx = kernel_context(gt->i915, NULL);
+	if (!c->ctx)
+		return -ENOMEM;
+
+	if (igt_spinner_init(&c->spin, gt))
+		goto err_ctx;
+
+	return 0;
+
+err_ctx:
+	kernel_context_close(c->ctx);
+	return -ENOMEM;
+}
+
+static void preempt_client_fini(struct preempt_client *c)
+{
+	igt_spinner_fini(&c->spin);
+	kernel_context_close(c->ctx);
+}
+
+static int live_nopreempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct preempt_client a, b;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	/*
+	 * Verify that we can disable preemption for an individual request
+	 * that may be being observed and not want to be interrupted.
+	 */
+
+	if (preempt_client_init(gt, &a))
+		return -ENOMEM;
+	if (preempt_client_init(gt, &b))
+		goto err_client_a;
+	b.ctx->sched.priority = I915_PRIORITY_MAX;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq_a, *rq_b;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		engine->execlists.preempt_hang.count = 0;
+
+		rq_a = spinner_create_request(&a.spin,
+					      a.ctx, engine,
+					      MI_ARB_CHECK);
+		if (IS_ERR(rq_a)) {
+			err = PTR_ERR(rq_a);
+			goto err_client_b;
+		}
+
+		/* Low priority client, but unpreemptable! */
+		__set_bit(I915_FENCE_FLAG_NOPREEMPT, &rq_a->fence.flags);
+
+		i915_request_add(rq_a);
+		if (!igt_wait_for_spinner(&a.spin, rq_a)) {
+			pr_err("First client failed to start\n");
+			goto err_wedged;
+		}
+
+		rq_b = spinner_create_request(&b.spin,
+					      b.ctx, engine,
+					      MI_ARB_CHECK);
+		if (IS_ERR(rq_b)) {
+			err = PTR_ERR(rq_b);
+			goto err_client_b;
+		}
+
+		i915_request_add(rq_b);
+
+		/* B is much more important than A! (But A is unpreemptable.) */
+		GEM_BUG_ON(rq_prio(rq_b) <= rq_prio(rq_a));
+
+		/* Wait long enough for preemption and timeslicing */
+		if (igt_wait_for_spinner(&b.spin, rq_b)) {
+			pr_err("Second client started too early!\n");
+			goto err_wedged;
+		}
+
+		igt_spinner_end(&a.spin);
+
+		if (!igt_wait_for_spinner(&b.spin, rq_b)) {
+			pr_err("Second client failed to start\n");
+			goto err_wedged;
+		}
+
+		igt_spinner_end(&b.spin);
+
+		if (engine->execlists.preempt_hang.count) {
+			pr_err("Preemption recorded x%d; should have been suppressed!\n",
+			       engine->execlists.preempt_hang.count);
+			err = -EINVAL;
+			goto err_wedged;
+		}
+
+		if (igt_flush_test(gt->i915))
+			goto err_wedged;
+	}
+
+	err = 0;
+err_client_b:
+	preempt_client_fini(&b);
+err_client_a:
+	preempt_client_fini(&a);
+	return err;
+
+err_wedged:
+	igt_spinner_end(&b.spin);
+	igt_spinner_end(&a.spin);
+	intel_gt_set_wedged(gt);
+	err = -EIO;
+	goto err_client_b;
+}
+
+struct live_preempt_cancel {
+	struct intel_engine_cs *engine;
+	struct preempt_client a, b;
+};
+
+static int __cancel_active0(struct live_preempt_cancel *arg)
+{
+	struct i915_request *rq;
+	struct igt_live_test t;
+	int err;
+
+	/* Preempt cancel of ELSP0 */
+	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
+	if (igt_live_test_begin(&t, arg->engine->i915,
+				__func__, arg->engine->name))
+		return -EIO;
+
+	rq = spinner_create_request(&arg->a.spin,
+				    arg->a.ctx, arg->engine,
+				    MI_ARB_CHECK);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	clear_bit(CONTEXT_BANNED, &rq->context->flags);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
+		err = -EIO;
+		goto out;
+	}
+
+	intel_context_ban(rq->context, rq);
+	err = intel_engine_pulse(arg->engine);
+	if (err)
+		goto out;
+
+	err = wait_for_reset(arg->engine, rq, HZ / 2);
+	if (err) {
+		pr_err("Cancelled inflight0 request did not reset\n");
+		goto out;
+	}
+
+out:
+	i915_request_put(rq);
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	return err;
+}
+
+static int __cancel_active1(struct live_preempt_cancel *arg)
+{
+	struct i915_request *rq[2] = {};
+	struct igt_live_test t;
+	int err;
+
+	/* Preempt cancel of ELSP1 */
+	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
+	if (igt_live_test_begin(&t, arg->engine->i915,
+				__func__, arg->engine->name))
+		return -EIO;
+
+	rq[0] = spinner_create_request(&arg->a.spin,
+				       arg->a.ctx, arg->engine,
+				       MI_NOOP); /* no preemption */
+	if (IS_ERR(rq[0]))
+		return PTR_ERR(rq[0]);
+
+	clear_bit(CONTEXT_BANNED, &rq[0]->context->flags);
+	i915_request_get(rq[0]);
+	i915_request_add(rq[0]);
+	if (!igt_wait_for_spinner(&arg->a.spin, rq[0])) {
+		err = -EIO;
+		goto out;
+	}
+
+	rq[1] = spinner_create_request(&arg->b.spin,
+				       arg->b.ctx, arg->engine,
+				       MI_ARB_CHECK);
+	if (IS_ERR(rq[1])) {
+		err = PTR_ERR(rq[1]);
+		goto out;
+	}
+
+	clear_bit(CONTEXT_BANNED, &rq[1]->context->flags);
+	i915_request_get(rq[1]);
+	err = i915_request_await_dma_fence(rq[1], &rq[0]->fence);
+	i915_request_add(rq[1]);
+	if (err)
+		goto out;
+
+	intel_context_ban(rq[1]->context, rq[1]);
+	err = intel_engine_pulse(arg->engine);
+	if (err)
+		goto out;
+
+	igt_spinner_end(&arg->a.spin);
+	err = wait_for_reset(arg->engine, rq[1], HZ / 2);
+	if (err)
+		goto out;
+
+	if (rq[0]->fence.error != 0) {
+		pr_err("Normal inflight0 request did not complete\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	if (rq[1]->fence.error != -EIO) {
+		pr_err("Cancelled inflight1 request did not report -EIO\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+out:
+	i915_request_put(rq[1]);
+	i915_request_put(rq[0]);
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	return err;
+}
+
+static int __cancel_queued(struct live_preempt_cancel *arg)
+{
+	struct i915_request *rq[3] = {};
+	struct igt_live_test t;
+	int err;
+
+	/* Full ELSP and one in the wings */
+	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
+	if (igt_live_test_begin(&t, arg->engine->i915,
+				__func__, arg->engine->name))
+		return -EIO;
+
+	rq[0] = spinner_create_request(&arg->a.spin,
+				       arg->a.ctx, arg->engine,
+				       MI_ARB_CHECK);
+	if (IS_ERR(rq[0]))
+		return PTR_ERR(rq[0]);
+
+	clear_bit(CONTEXT_BANNED, &rq[0]->context->flags);
+	i915_request_get(rq[0]);
+	i915_request_add(rq[0]);
+	if (!igt_wait_for_spinner(&arg->a.spin, rq[0])) {
+		err = -EIO;
+		goto out;
+	}
+
+	rq[1] = igt_request_alloc(arg->b.ctx, arg->engine);
+	if (IS_ERR(rq[1])) {
+		err = PTR_ERR(rq[1]);
+		goto out;
+	}
+
+	clear_bit(CONTEXT_BANNED, &rq[1]->context->flags);
+	i915_request_get(rq[1]);
+	err = i915_request_await_dma_fence(rq[1], &rq[0]->fence);
+	i915_request_add(rq[1]);
+	if (err)
+		goto out;
+
+	rq[2] = spinner_create_request(&arg->b.spin,
+				       arg->a.ctx, arg->engine,
+				       MI_ARB_CHECK);
+	if (IS_ERR(rq[2])) {
+		err = PTR_ERR(rq[2]);
+		goto out;
+	}
+
+	i915_request_get(rq[2]);
+	err = i915_request_await_dma_fence(rq[2], &rq[1]->fence);
+	i915_request_add(rq[2]);
+	if (err)
+		goto out;
+
+	intel_context_ban(rq[2]->context, rq[2]);
+	err = intel_engine_pulse(arg->engine);
+	if (err)
+		goto out;
+
+	err = wait_for_reset(arg->engine, rq[2], HZ / 2);
+	if (err)
+		goto out;
+
+	if (rq[0]->fence.error != -EIO) {
+		pr_err("Cancelled inflight0 request did not report -EIO\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	/*
+	 * The behavior between having semaphores and not is different. With
+	 * semaphores the subsequent request is on the hardware and not cancelled
+	 * while without the request is held in the driver and cancelled.
+	 */
+	if (intel_engine_has_semaphores(rq[1]->engine) &&
+	    rq[1]->fence.error != 0) {
+		pr_err("Normal inflight1 request did not complete\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	if (rq[2]->fence.error != -EIO) {
+		pr_err("Cancelled queued request did not report -EIO\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+out:
+	i915_request_put(rq[2]);
+	i915_request_put(rq[1]);
+	i915_request_put(rq[0]);
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	return err;
+}
+
+static int __cancel_hostile(struct live_preempt_cancel *arg)
+{
+	struct i915_request *rq;
+	int err;
+
+	/* Preempt cancel non-preemptible spinner in ELSP0 */
+	if (!CONFIG_DRM_I915_PREEMPT_TIMEOUT)
+		return 0;
+
+	if (!intel_has_reset_engine(arg->engine->gt))
+		return 0;
+
+	GEM_TRACE("%s(%s)\n", __func__, arg->engine->name);
+	rq = spinner_create_request(&arg->a.spin,
+				    arg->a.ctx, arg->engine,
+				    MI_NOOP); /* preemption disabled */
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	clear_bit(CONTEXT_BANNED, &rq->context->flags);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
+		err = -EIO;
+		goto out;
+	}
+
+	intel_context_ban(rq->context, rq);
+	err = intel_engine_pulse(arg->engine); /* force reset */
+	if (err)
+		goto out;
+
+	err = wait_for_reset(arg->engine, rq, HZ / 2);
+	if (err) {
+		pr_err("Cancelled inflight0 request did not reset\n");
+		goto out;
+	}
+
+out:
+	i915_request_put(rq);
+	if (igt_flush_test(arg->engine->i915))
+		err = -EIO;
+	return err;
+}
+
+static void force_reset_timeout(struct intel_engine_cs *engine)
+{
+	engine->reset_timeout.probability = 999;
+	atomic_set(&engine->reset_timeout.times, -1);
+}
+
+static void cancel_reset_timeout(struct intel_engine_cs *engine)
+{
+	memset(&engine->reset_timeout, 0, sizeof(engine->reset_timeout));
+}
+
+static int __cancel_fail(struct live_preempt_cancel *arg)
+{
+	struct intel_engine_cs *engine = arg->engine;
+	struct i915_request *rq;
+	int err;
+
+	if (!CONFIG_DRM_I915_PREEMPT_TIMEOUT)
+		return 0;
+
+	if (!intel_has_reset_engine(engine->gt))
+		return 0;
+
+	GEM_TRACE("%s(%s)\n", __func__, engine->name);
+	rq = spinner_create_request(&arg->a.spin,
+				    arg->a.ctx, engine,
+				    MI_NOOP); /* preemption disabled */
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	clear_bit(CONTEXT_BANNED, &rq->context->flags);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
+		err = -EIO;
+		goto out;
+	}
+
+	intel_context_set_banned(rq->context);
+
+	err = intel_engine_pulse(engine);
+	if (err)
+		goto out;
+
+	force_reset_timeout(engine);
+
+	/* force preempt reset [failure] */
+	while (!engine->execlists.pending[0])
+		intel_engine_flush_submission(engine);
+	del_timer_sync(&engine->execlists.preempt);
+	intel_engine_flush_submission(engine);
+
+	cancel_reset_timeout(engine);
+
+	/* after failure, require heartbeats to reset device */
+	intel_engine_set_heartbeat(engine, 1);
+	err = wait_for_reset(engine, rq, HZ / 2);
+	intel_engine_set_heartbeat(engine,
+				   engine->defaults.heartbeat_interval_ms);
+	if (err) {
+		pr_err("Cancelled inflight0 request did not reset\n");
+		goto out;
+	}
+
+out:
+	i915_request_put(rq);
+	if (igt_flush_test(engine->i915))
+		err = -EIO;
+	return err;
+}
+
+static int live_preempt_cancel(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct live_preempt_cancel data;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	/*
+	 * To cancel an inflight context, we need to first remove it from the
+	 * GPU. That sounds like preemption! Plus a little bit of bookkeeping.
+	 */
+
+	if (preempt_client_init(gt, &data.a))
+		return -ENOMEM;
+	if (preempt_client_init(gt, &data.b))
+		goto err_client_a;
+
+	for_each_engine(data.engine, gt, id) {
+		if (!intel_engine_has_preemption(data.engine))
+			continue;
+
+		err = __cancel_active0(&data);
+		if (err)
+			goto err_wedged;
+
+		err = __cancel_active1(&data);
+		if (err)
+			goto err_wedged;
+
+		err = __cancel_queued(&data);
+		if (err)
+			goto err_wedged;
+
+		err = __cancel_hostile(&data);
+		if (err)
+			goto err_wedged;
+
+		err = __cancel_fail(&data);
+		if (err)
+			goto err_wedged;
+	}
+
+	err = 0;
+err_client_b:
+	preempt_client_fini(&data.b);
+err_client_a:
+	preempt_client_fini(&data.a);
+	return err;
+
+err_wedged:
+	GEM_TRACE_DUMP();
+	igt_spinner_end(&data.b.spin);
+	igt_spinner_end(&data.a.spin);
+	intel_gt_set_wedged(gt);
+	goto err_client_b;
+}
+
+static int live_suppress_self_preempt(void *arg)
+{
+	struct i915_sched_attr attr = { .priority = I915_PRIORITY_MAX };
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct preempt_client a, b;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	/*
+	 * Verify that if a preemption request does not cause a change in
+	 * the current execution order, the preempt-to-idle injection is
+	 * skipped and that we do not accidentally apply it after the CS
+	 * completion event.
+	 */
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0; /* presume black blox */
+
+	if (intel_vgpu_active(gt->i915))
+		return 0; /* GVT forces single port & request submission */
+
+	if (preempt_client_init(gt, &a))
+		return -ENOMEM;
+	if (preempt_client_init(gt, &b))
+		goto err_client_a;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq_a, *rq_b;
+		int depth;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (igt_flush_test(gt->i915))
+			goto err_wedged;
+
+		st_engine_heartbeat_disable(engine);
+		engine->execlists.preempt_hang.count = 0;
+
+		rq_a = spinner_create_request(&a.spin,
+					      a.ctx, engine,
+					      MI_NOOP);
+		if (IS_ERR(rq_a)) {
+			err = PTR_ERR(rq_a);
+			st_engine_heartbeat_enable(engine);
+			goto err_client_b;
+		}
+
+		i915_request_add(rq_a);
+		if (!igt_wait_for_spinner(&a.spin, rq_a)) {
+			pr_err("First client failed to start\n");
+			st_engine_heartbeat_enable(engine);
+			goto err_wedged;
+		}
+
+		/* Keep postponing the timer to avoid premature slicing */
+		mod_timer(&engine->execlists.timer, jiffies + HZ);
+		for (depth = 0; depth < 8; depth++) {
+			rq_b = spinner_create_request(&b.spin,
+						      b.ctx, engine,
+						      MI_NOOP);
+			if (IS_ERR(rq_b)) {
+				err = PTR_ERR(rq_b);
+				st_engine_heartbeat_enable(engine);
+				goto err_client_b;
+			}
+			i915_request_add(rq_b);
+
+			GEM_BUG_ON(i915_request_completed(rq_a));
+			engine->sched_engine->schedule(rq_a, &attr);
+			igt_spinner_end(&a.spin);
+
+			if (!igt_wait_for_spinner(&b.spin, rq_b)) {
+				pr_err("Second client failed to start\n");
+				st_engine_heartbeat_enable(engine);
+				goto err_wedged;
+			}
+
+			swap(a, b);
+			rq_a = rq_b;
+		}
+		igt_spinner_end(&a.spin);
+
+		if (engine->execlists.preempt_hang.count) {
+			pr_err("Preemption on %s recorded x%d, depth %d; should have been suppressed!\n",
+			       engine->name,
+			       engine->execlists.preempt_hang.count,
+			       depth);
+			st_engine_heartbeat_enable(engine);
+			err = -EINVAL;
+			goto err_client_b;
+		}
+
+		st_engine_heartbeat_enable(engine);
+		if (igt_flush_test(gt->i915))
+			goto err_wedged;
+	}
+
+	err = 0;
+err_client_b:
+	preempt_client_fini(&b);
+err_client_a:
+	preempt_client_fini(&a);
+	return err;
+
+err_wedged:
+	igt_spinner_end(&b.spin);
+	igt_spinner_end(&a.spin);
+	intel_gt_set_wedged(gt);
+	err = -EIO;
+	goto err_client_b;
+}
+
+static int live_chain_preempt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct preempt_client hi, lo;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	/*
+	 * Build a chain AB...BA between two contexts (A, B) and request
+	 * preemption of the last request. It should then complete before
+	 * the previously submitted spinner in B.
+	 */
+
+	if (preempt_client_init(gt, &hi))
+		return -ENOMEM;
+
+	if (preempt_client_init(gt, &lo))
+		goto err_client_hi;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_sched_attr attr = { .priority = I915_PRIORITY_MAX };
+		struct igt_live_test t;
+		struct i915_request *rq;
+		int ring_size, count, i;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		rq = spinner_create_request(&lo.spin,
+					    lo.ctx, engine,
+					    MI_ARB_CHECK);
+		if (IS_ERR(rq))
+			goto err_wedged;
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		ring_size = rq->wa_tail - rq->head;
+		if (ring_size < 0)
+			ring_size += rq->ring->size;
+		ring_size = rq->ring->size / ring_size;
+		pr_debug("%s(%s): Using maximum of %d requests\n",
+			 __func__, engine->name, ring_size);
+
+		igt_spinner_end(&lo.spin);
+		if (i915_request_wait(rq, 0, HZ / 2) < 0) {
+			pr_err("Timed out waiting to flush %s\n", engine->name);
+			i915_request_put(rq);
+			goto err_wedged;
+		}
+		i915_request_put(rq);
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			goto err_wedged;
+		}
+
+		for_each_prime_number_from(count, 1, ring_size) {
+			rq = spinner_create_request(&hi.spin,
+						    hi.ctx, engine,
+						    MI_ARB_CHECK);
+			if (IS_ERR(rq))
+				goto err_wedged;
+			i915_request_add(rq);
+			if (!igt_wait_for_spinner(&hi.spin, rq))
+				goto err_wedged;
+
+			rq = spinner_create_request(&lo.spin,
+						    lo.ctx, engine,
+						    MI_ARB_CHECK);
+			if (IS_ERR(rq))
+				goto err_wedged;
+			i915_request_add(rq);
+
+			for (i = 0; i < count; i++) {
+				rq = igt_request_alloc(lo.ctx, engine);
+				if (IS_ERR(rq))
+					goto err_wedged;
+				i915_request_add(rq);
+			}
+
+			rq = igt_request_alloc(hi.ctx, engine);
+			if (IS_ERR(rq))
+				goto err_wedged;
+
+			i915_request_get(rq);
+			i915_request_add(rq);
+			engine->sched_engine->schedule(rq, &attr);
+
+			igt_spinner_end(&hi.spin);
+			if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+				struct drm_printer p =
+					drm_info_printer(gt->i915->drm.dev);
+
+				pr_err("Failed to preempt over chain of %d\n",
+				       count);
+				intel_engine_dump(engine, &p,
+						  "%s\n", engine->name);
+				i915_request_put(rq);
+				goto err_wedged;
+			}
+			igt_spinner_end(&lo.spin);
+			i915_request_put(rq);
+
+			rq = igt_request_alloc(lo.ctx, engine);
+			if (IS_ERR(rq))
+				goto err_wedged;
+
+			i915_request_get(rq);
+			i915_request_add(rq);
+
+			if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+				struct drm_printer p =
+					drm_info_printer(gt->i915->drm.dev);
+
+				pr_err("Failed to flush low priority chain of %d requests\n",
+				       count);
+				intel_engine_dump(engine, &p,
+						  "%s\n", engine->name);
+
+				i915_request_put(rq);
+				goto err_wedged;
+			}
+			i915_request_put(rq);
+		}
+
+		if (igt_live_test_end(&t)) {
+			err = -EIO;
+			goto err_wedged;
+		}
+	}
+
+	err = 0;
+err_client_lo:
+	preempt_client_fini(&lo);
+err_client_hi:
+	preempt_client_fini(&hi);
+	return err;
+
+err_wedged:
+	igt_spinner_end(&hi.spin);
+	igt_spinner_end(&lo.spin);
+	intel_gt_set_wedged(gt);
+	err = -EIO;
+	goto err_client_lo;
+}
+
+static int create_gang(struct intel_engine_cs *engine,
+		       struct i915_request **prev)
+{
+	struct drm_i915_gem_object *obj;
+	struct intel_context *ce;
+	struct i915_request *rq;
+	struct i915_vma *vma;
+	u32 *cs;
+	int err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	obj = i915_gem_object_create_internal(engine->i915, 4096);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto err_ce;
+	}
+
+	vma = i915_vma_instance(obj, ce->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto err_obj;
+
+	cs = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_obj;
+	}
+
+	/* Semaphore target: spin until zero */
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_EQ_SDD;
+	*cs++ = 0;
+	*cs++ = lower_32_bits(vma->node.start);
+	*cs++ = upper_32_bits(vma->node.start);
+
+	if (*prev) {
+		u64 offset = (*prev)->batch->node.start;
+
+		/* Terminate the spinner in the next lower priority batch. */
+		*cs++ = MI_STORE_DWORD_IMM_GEN4;
+		*cs++ = lower_32_bits(offset);
+		*cs++ = upper_32_bits(offset);
+		*cs++ = 0;
+	}
+
+	*cs++ = MI_BATCH_BUFFER_END;
+	i915_gem_object_flush_map(obj);
+	i915_gem_object_unpin_map(obj);
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_obj;
+	}
+
+	rq->batch = i915_vma_get(vma);
+	i915_request_get(rq);
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, false);
+	if (!err)
+		err = i915_vma_move_to_active(vma, rq, 0);
+	if (!err)
+		err = rq->engine->emit_bb_start(rq,
+						vma->node.start,
+						PAGE_SIZE, 0);
+	i915_vma_unlock(vma);
+	i915_request_add(rq);
+	if (err)
+		goto err_rq;
+
+	i915_gem_object_put(obj);
+	intel_context_put(ce);
+
+	rq->mock.link.next = &(*prev)->mock.link;
+	*prev = rq;
+	return 0;
+
+err_rq:
+	i915_vma_put(rq->batch);
+	i915_request_put(rq);
+err_obj:
+	i915_gem_object_put(obj);
+err_ce:
+	intel_context_put(ce);
+	return err;
+}
+
+static int __live_preempt_ring(struct intel_engine_cs *engine,
+			       struct igt_spinner *spin,
+			       int queue_sz, int ring_sz)
+{
+	struct intel_context *ce[2] = {};
+	struct i915_request *rq;
+	struct igt_live_test t;
+	int err = 0;
+	int n;
+
+	if (igt_live_test_begin(&t, engine->i915, __func__, engine->name))
+		return -EIO;
+
+	for (n = 0; n < ARRAY_SIZE(ce); n++) {
+		struct intel_context *tmp;
+
+		tmp = intel_context_create(engine);
+		if (IS_ERR(tmp)) {
+			err = PTR_ERR(tmp);
+			goto err_ce;
+		}
+
+		tmp->ring_size = ring_sz;
+
+		err = intel_context_pin(tmp);
+		if (err) {
+			intel_context_put(tmp);
+			goto err_ce;
+		}
+
+		memset32(tmp->ring->vaddr,
+			 0xdeadbeef, /* trigger a hang if executed */
+			 tmp->ring->vma->size / sizeof(u32));
+
+		ce[n] = tmp;
+	}
+
+	rq = igt_spinner_create_request(spin, ce[0], MI_ARB_CHECK);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_ce;
+	}
+
+	i915_request_get(rq);
+	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+	i915_request_add(rq);
+
+	if (!igt_wait_for_spinner(spin, rq)) {
+		intel_gt_set_wedged(engine->gt);
+		i915_request_put(rq);
+		err = -ETIME;
+		goto err_ce;
+	}
+
+	/* Fill the ring, until we will cause a wrap */
+	n = 0;
+	while (ce[0]->ring->tail - rq->wa_tail <= queue_sz) {
+		struct i915_request *tmp;
+
+		tmp = intel_context_create_request(ce[0]);
+		if (IS_ERR(tmp)) {
+			err = PTR_ERR(tmp);
+			i915_request_put(rq);
+			goto err_ce;
+		}
+
+		i915_request_add(tmp);
+		intel_engine_flush_submission(engine);
+		n++;
+	}
+	intel_engine_flush_submission(engine);
+	pr_debug("%s: Filled %d with %d nop tails {size:%x, tail:%x, emit:%x, rq.tail:%x}\n",
+		 engine->name, queue_sz, n,
+		 ce[0]->ring->size,
+		 ce[0]->ring->tail,
+		 ce[0]->ring->emit,
+		 rq->tail);
+	i915_request_put(rq);
+
+	/* Create a second request to preempt the first ring */
+	rq = intel_context_create_request(ce[1]);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_ce;
+	}
+
+	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	err = wait_for_submit(engine, rq, HZ / 2);
+	i915_request_put(rq);
+	if (err) {
+		pr_err("%s: preemption request was not submitted\n",
+		       engine->name);
+		err = -ETIME;
+	}
+
+	pr_debug("%s: ring[0]:{ tail:%x, emit:%x }, ring[1]:{ tail:%x, emit:%x }\n",
+		 engine->name,
+		 ce[0]->ring->tail, ce[0]->ring->emit,
+		 ce[1]->ring->tail, ce[1]->ring->emit);
+
+err_ce:
+	intel_engine_flush_submission(engine);
+	igt_spinner_end(spin);
+	for (n = 0; n < ARRAY_SIZE(ce); n++) {
+		if (IS_ERR_OR_NULL(ce[n]))
+			break;
+
+		intel_context_unpin(ce[n]);
+		intel_context_put(ce[n]);
+	}
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	return err;
+}
+
+static int live_preempt_ring(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct igt_spinner spin;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Check that we rollback large chunks of a ring in order to do a
+	 * preemption event. Similar to live_unlite_ring, but looking at
+	 * ring size rather than the impact of intel_ring_direction().
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		int n;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		st_engine_heartbeat_disable(engine);
+
+		for (n = 0; n <= 3; n++) {
+			err = __live_preempt_ring(engine, &spin,
+						  n * SZ_4K / 4, SZ_4K);
+			if (err)
+				break;
+		}
+
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_preempt_gang(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * Build as long a chain of preempters as we can, with each
+	 * request higher priority than the last. Once we are ready, we release
+	 * the last batch which then precolates down the chain, each releasing
+	 * the next oldest in turn. The intent is to simply push as hard as we
+	 * can with the number of preemptions, trying to exceed narrow HW
+	 * limits. At a minimum, we insist that we can sort all the user
+	 * high priority levels into execution order.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq = NULL;
+		struct igt_live_test t;
+		IGT_TIMEOUT(end_time);
+		int prio = 0;
+		int err = 0;
+		u32 *cs;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name))
+			return -EIO;
+
+		do {
+			struct i915_sched_attr attr = { .priority = prio++ };
+
+			err = create_gang(engine, &rq);
+			if (err)
+				break;
+
+			/* Submit each spinner at increasing priority */
+			engine->sched_engine->schedule(rq, &attr);
+		} while (prio <= I915_PRIORITY_MAX &&
+			 !__igt_timeout(end_time, NULL));
+		pr_debug("%s: Preempt chain of %d requests\n",
+			 engine->name, prio);
+
+		/*
+		 * Such that the last spinner is the highest priority and
+		 * should execute first. When that spinner completes,
+		 * it will terminate the next lowest spinner until there
+		 * are no more spinners and the gang is complete.
+		 */
+		cs = i915_gem_object_pin_map_unlocked(rq->batch->obj, I915_MAP_WC);
+		if (!IS_ERR(cs)) {
+			*cs = 0;
+			i915_gem_object_unpin_map(rq->batch->obj);
+		} else {
+			err = PTR_ERR(cs);
+			intel_gt_set_wedged(gt);
+		}
+
+		while (rq) { /* wait for each rq from highest to lowest prio */
+			struct i915_request *n = list_next_entry(rq, mock.link);
+
+			if (err == 0 && i915_request_wait(rq, 0, HZ / 5) < 0) {
+				struct drm_printer p =
+					drm_info_printer(engine->i915->drm.dev);
+
+				pr_err("Failed to flush chain of %d requests, at %d\n",
+				       prio, rq_prio(rq));
+				intel_engine_dump(engine, &p,
+						  "%s\n", engine->name);
+
+				err = -ETIME;
+			}
+
+			i915_vma_put(rq->batch);
+			i915_request_put(rq);
+			rq = n;
+		}
+
+		if (igt_live_test_end(&t))
+			err = -EIO;
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static struct i915_vma *
+create_gpr_user(struct intel_engine_cs *engine,
+		struct i915_vma *result,
+		unsigned int offset)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u32 *cs;
+	int err;
+	int i;
+
+	obj = i915_gem_object_create_internal(engine->i915, 4096);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, result->vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return vma;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err) {
+		i915_vma_put(vma);
+		return ERR_PTR(err);
+	}
+
+	cs = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(cs)) {
+		i915_vma_put(vma);
+		return ERR_CAST(cs);
+	}
+
+	/* All GPR are clear for new contexts. We use GPR(0) as a constant */
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = CS_GPR(engine, 0);
+	*cs++ = 1;
+
+	for (i = 1; i < NUM_GPR; i++) {
+		u64 addr;
+
+		/*
+		 * Perform: GPR[i]++
+		 *
+		 * As we read and write into the context saved GPR[i], if
+		 * we restart this batch buffer from an earlier point, we
+		 * will repeat the increment and store a value > 1.
+		 */
+		*cs++ = MI_MATH(4);
+		*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCA, MI_MATH_REG(i));
+		*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCB, MI_MATH_REG(0));
+		*cs++ = MI_MATH_ADD;
+		*cs++ = MI_MATH_STORE(MI_MATH_REG(i), MI_MATH_REG_ACCU);
+
+		addr = result->node.start + offset + i * sizeof(*cs);
+		*cs++ = MI_STORE_REGISTER_MEM_GEN8;
+		*cs++ = CS_GPR(engine, 2 * i);
+		*cs++ = lower_32_bits(addr);
+		*cs++ = upper_32_bits(addr);
+
+		*cs++ = MI_SEMAPHORE_WAIT |
+			MI_SEMAPHORE_POLL |
+			MI_SEMAPHORE_SAD_GTE_SDD;
+		*cs++ = i;
+		*cs++ = lower_32_bits(result->node.start);
+		*cs++ = upper_32_bits(result->node.start);
+	}
+
+	*cs++ = MI_BATCH_BUFFER_END;
+	i915_gem_object_flush_map(obj);
+	i915_gem_object_unpin_map(obj);
+
+	return vma;
+}
+
+static struct i915_vma *create_global(struct intel_gt *gt, size_t sz)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err;
+
+	obj = i915_gem_object_create_internal(gt->i915, sz);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return vma;
+	}
+
+	err = i915_ggtt_pin(vma, NULL, 0, 0);
+	if (err) {
+		i915_vma_put(vma);
+		return ERR_PTR(err);
+	}
+
+	return vma;
+}
+
+static struct i915_request *
+create_gpr_client(struct intel_engine_cs *engine,
+		  struct i915_vma *global,
+		  unsigned int offset)
+{
+	struct i915_vma *batch, *vma;
+	struct intel_context *ce;
+	struct i915_request *rq;
+	int err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return ERR_CAST(ce);
+
+	vma = i915_vma_instance(global->obj, ce->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto out_ce;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto out_ce;
+
+	batch = create_gpr_user(engine, vma, offset);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		goto out_vma;
+	}
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_batch;
+	}
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, false);
+	if (!err)
+		err = i915_vma_move_to_active(vma, rq, 0);
+	i915_vma_unlock(vma);
+
+	i915_vma_lock(batch);
+	if (!err)
+		err = i915_request_await_object(rq, batch->obj, false);
+	if (!err)
+		err = i915_vma_move_to_active(batch, rq, 0);
+	if (!err)
+		err = rq->engine->emit_bb_start(rq,
+						batch->node.start,
+						PAGE_SIZE, 0);
+	i915_vma_unlock(batch);
+	i915_vma_unpin(batch);
+
+	if (!err)
+		i915_request_get(rq);
+	i915_request_add(rq);
+
+out_batch:
+	i915_vma_put(batch);
+out_vma:
+	i915_vma_unpin(vma);
+out_ce:
+	intel_context_put(ce);
+	return err ? ERR_PTR(err) : rq;
+}
+
+static int preempt_user(struct intel_engine_cs *engine,
+			struct i915_vma *global,
+			int id)
+{
+	struct i915_sched_attr attr = {
+		.priority = I915_PRIORITY_MAX
+	};
+	struct i915_request *rq;
+	int err = 0;
+	u32 *cs;
+
+	rq = intel_engine_create_kernel_request(engine);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = i915_ggtt_offset(global);
+	*cs++ = 0;
+	*cs++ = id;
+
+	intel_ring_advance(rq, cs);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	engine->sched_engine->schedule(rq, &attr);
+
+	if (i915_request_wait(rq, 0, HZ / 2) < 0)
+		err = -ETIME;
+	i915_request_put(rq);
+
+	return err;
+}
+
+static int live_preempt_user(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct i915_vma *global;
+	enum intel_engine_id id;
+	u32 *result;
+	int err = 0;
+
+	/*
+	 * In our other tests, we look at preemption in carefully
+	 * controlled conditions in the ringbuffer. Since most of the
+	 * time is spent in user batches, most of our preemptions naturally
+	 * occur there. We want to verify that when we preempt inside a batch
+	 * we continue on from the current instruction and do not roll back
+	 * to the start, or another earlier arbitration point.
+	 *
+	 * To verify this, we create a batch which is a mixture of
+	 * MI_MATH (gpr++) MI_SRM (gpr) and preemption points. Then with
+	 * a few preempting contexts thrown into the mix, we look for any
+	 * repeated instructions (which show up as incorrect values).
+	 */
+
+	global = create_global(gt, 4096);
+	if (IS_ERR(global))
+		return PTR_ERR(global);
+
+	result = i915_gem_object_pin_map_unlocked(global->obj, I915_MAP_WC);
+	if (IS_ERR(result)) {
+		i915_vma_unpin_and_release(&global, 0);
+		return PTR_ERR(result);
+	}
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *client[3] = {};
+		struct igt_live_test t;
+		int i;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		if (GRAPHICS_VER(gt->i915) == 8 && engine->class != RENDER_CLASS)
+			continue; /* we need per-context GPR */
+
+		if (igt_live_test_begin(&t, gt->i915, __func__, engine->name)) {
+			err = -EIO;
+			break;
+		}
+
+		memset(result, 0, 4096);
+
+		for (i = 0; i < ARRAY_SIZE(client); i++) {
+			struct i915_request *rq;
+
+			rq = create_gpr_client(engine, global,
+					       NUM_GPR * i * sizeof(u32));
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				goto end_test;
+			}
+
+			client[i] = rq;
+		}
+
+		/* Continuously preempt the set of 3 running contexts */
+		for (i = 1; i <= NUM_GPR; i++) {
+			err = preempt_user(engine, global, i);
+			if (err)
+				goto end_test;
+		}
+
+		if (READ_ONCE(result[0]) != NUM_GPR) {
+			pr_err("%s: Failed to release semaphore\n",
+			       engine->name);
+			err = -EIO;
+			goto end_test;
+		}
+
+		for (i = 0; i < ARRAY_SIZE(client); i++) {
+			int gpr;
+
+			if (i915_request_wait(client[i], 0, HZ / 2) < 0) {
+				err = -ETIME;
+				goto end_test;
+			}
+
+			for (gpr = 1; gpr < NUM_GPR; gpr++) {
+				if (result[NUM_GPR * i + gpr] != 1) {
+					pr_err("%s: Invalid result, client %d, gpr %d, result: %d\n",
+					       engine->name,
+					       i, gpr, result[NUM_GPR * i + gpr]);
+					err = -EINVAL;
+					goto end_test;
+				}
+			}
+		}
+
+end_test:
+		for (i = 0; i < ARRAY_SIZE(client); i++) {
+			if (!client[i])
+				break;
+
+			i915_request_put(client[i]);
+		}
+
+		/* Flush the semaphores on error */
+		smp_store_mb(result[0], -1);
+		if (igt_live_test_end(&t))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	i915_vma_unpin_and_release(&global, I915_VMA_RELEASE_MAP);
+	return err;
+}
+
+static int live_preempt_timeout(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gem_context *ctx_hi, *ctx_lo;
+	struct igt_spinner spin_lo;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = -ENOMEM;
+
+	/*
+	 * Check that we force preemption to occur by cancelling the previous
+	 * context if it refuses to yield the GPU.
+	 */
+	if (!CONFIG_DRM_I915_PREEMPT_TIMEOUT)
+		return 0;
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	ctx_hi = kernel_context(gt->i915, NULL);
+	if (!ctx_hi)
+		return -ENOMEM;
+	ctx_hi->sched.priority = I915_CONTEXT_MAX_USER_PRIORITY;
+
+	ctx_lo = kernel_context(gt->i915, NULL);
+	if (!ctx_lo)
+		goto err_ctx_hi;
+	ctx_lo->sched.priority = I915_CONTEXT_MIN_USER_PRIORITY;
+
+	if (igt_spinner_init(&spin_lo, gt))
+		goto err_ctx_lo;
+
+	for_each_engine(engine, gt, id) {
+		unsigned long saved_timeout;
+		struct i915_request *rq;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		rq = spinner_create_request(&spin_lo, ctx_lo, engine,
+					    MI_NOOP); /* preemption disabled */
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_spin_lo;
+		}
+
+		i915_request_add(rq);
+		if (!igt_wait_for_spinner(&spin_lo, rq)) {
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto err_spin_lo;
+		}
+
+		rq = igt_request_alloc(ctx_hi, engine);
+		if (IS_ERR(rq)) {
+			igt_spinner_end(&spin_lo);
+			err = PTR_ERR(rq);
+			goto err_spin_lo;
+		}
+
+		/* Flush the previous CS ack before changing timeouts */
+		while (READ_ONCE(engine->execlists.pending[0]))
+			cpu_relax();
+
+		saved_timeout = engine->props.preempt_timeout_ms;
+		engine->props.preempt_timeout_ms = 1; /* in ms, -> 1 jiffie */
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		intel_engine_flush_submission(engine);
+		engine->props.preempt_timeout_ms = saved_timeout;
+
+		if (i915_request_wait(rq, 0, HZ / 10) < 0) {
+			intel_gt_set_wedged(gt);
+			i915_request_put(rq);
+			err = -ETIME;
+			goto err_spin_lo;
+		}
+
+		igt_spinner_end(&spin_lo);
+		i915_request_put(rq);
+	}
+
+	err = 0;
+err_spin_lo:
+	igt_spinner_fini(&spin_lo);
+err_ctx_lo:
+	kernel_context_close(ctx_lo);
+err_ctx_hi:
+	kernel_context_close(ctx_hi);
+	return err;
+}
+
+static int random_range(struct rnd_state *rnd, int min, int max)
+{
+	return i915_prandom_u32_max_state(max - min, rnd) + min;
+}
+
+static int random_priority(struct rnd_state *rnd)
+{
+	return random_range(rnd, I915_PRIORITY_MIN, I915_PRIORITY_MAX);
+}
+
+struct preempt_smoke {
+	struct intel_gt *gt;
+	struct kthread_work work;
+	struct i915_gem_context **contexts;
+	struct intel_engine_cs *engine;
+	struct drm_i915_gem_object *batch;
+	unsigned int ncontext;
+	struct rnd_state prng;
+	unsigned long count;
+	int result;
+};
+
+static struct i915_gem_context *smoke_context(struct preempt_smoke *smoke)
+{
+	return smoke->contexts[i915_prandom_u32_max_state(smoke->ncontext,
+							  &smoke->prng)];
+}
+
+static int smoke_submit(struct preempt_smoke *smoke,
+			struct i915_gem_context *ctx, int prio,
+			struct drm_i915_gem_object *batch)
+{
+	struct i915_request *rq;
+	struct i915_vma *vma = NULL;
+	int err = 0;
+
+	if (batch) {
+		struct i915_address_space *vm;
+
+		vm = i915_gem_context_get_eb_vm(ctx);
+		vma = i915_vma_instance(batch, vm, NULL);
+		i915_vm_put(vm);
+		if (IS_ERR(vma))
+			return PTR_ERR(vma);
+
+		err = i915_vma_pin(vma, 0, 0, PIN_USER);
+		if (err)
+			return err;
+	}
+
+	ctx->sched.priority = prio;
+
+	rq = igt_request_alloc(ctx, smoke->engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto unpin;
+	}
+
+	if (vma) {
+		i915_vma_lock(vma);
+		err = i915_request_await_object(rq, vma->obj, false);
+		if (!err)
+			err = i915_vma_move_to_active(vma, rq, 0);
+		if (!err)
+			err = rq->engine->emit_bb_start(rq,
+							vma->node.start,
+							PAGE_SIZE, 0);
+		i915_vma_unlock(vma);
+	}
+
+	i915_request_add(rq);
+
+unpin:
+	if (vma)
+		i915_vma_unpin(vma);
+
+	return err;
+}
+
+static void smoke_crescendo_work(struct kthread_work *work)
+{
+	struct preempt_smoke *smoke = container_of(work, typeof(*smoke), work);
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+
+	count = 0;
+	do {
+		struct i915_gem_context *ctx = smoke_context(smoke);
+
+		smoke->result = smoke_submit(smoke, ctx,
+					     count % I915_PRIORITY_MAX,
+					     smoke->batch);
+
+		count++;
+	} while (!smoke->result && count < smoke->ncontext &&
+		 !__igt_timeout(end_time, NULL));
+
+	smoke->count = count;
+}
+
+static int smoke_crescendo(struct preempt_smoke *smoke, unsigned int flags)
+#define BATCH BIT(0)
+{
+	struct kthread_worker *worker[I915_NUM_ENGINES] = {};
+	struct preempt_smoke *arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	unsigned long count;
+	int err = 0;
+
+	arg = kmalloc_array(I915_NUM_ENGINES, sizeof(*arg), GFP_KERNEL);
+	if (!arg)
+		return -ENOMEM;
+
+	memset(arg, 0, I915_NUM_ENGINES * sizeof(*arg));
+
+	for_each_engine(engine, smoke->gt, id) {
+		arg[id] = *smoke;
+		arg[id].engine = engine;
+		if (!(flags & BATCH))
+			arg[id].batch = NULL;
+		arg[id].count = 0;
+
+		worker[id] = kthread_create_worker(0, "igt/smoke:%d", id);
+		if (IS_ERR(worker[id])) {
+			err = PTR_ERR(worker[id]);
+			break;
+		}
+
+		kthread_init_work(&arg[id].work, smoke_crescendo_work);
+		kthread_queue_work(worker[id], &arg[id].work);
+	}
+
+	count = 0;
+	for_each_engine(engine, smoke->gt, id) {
+		if (IS_ERR_OR_NULL(worker[id]))
+			continue;
+
+		kthread_flush_work(&arg[id].work);
+		if (arg[id].result && !err)
+			err = arg[id].result;
+
+		count += arg[id].count;
+
+		kthread_destroy_worker(worker[id]);
+	}
+
+	pr_info("Submitted %lu crescendo:%x requests across %d engines and %d contexts\n",
+		count, flags, smoke->gt->info.num_engines, smoke->ncontext);
+
+	kfree(arg);
+	return 0;
+}
+
+static int smoke_random(struct preempt_smoke *smoke, unsigned int flags)
+{
+	enum intel_engine_id id;
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+
+	count = 0;
+	do {
+		for_each_engine(smoke->engine, smoke->gt, id) {
+			struct i915_gem_context *ctx = smoke_context(smoke);
+			int err;
+
+			err = smoke_submit(smoke,
+					   ctx, random_priority(&smoke->prng),
+					   flags & BATCH ? smoke->batch : NULL);
+			if (err)
+				return err;
+
+			count++;
+		}
+	} while (count < smoke->ncontext && !__igt_timeout(end_time, NULL));
+
+	pr_info("Submitted %lu random:%x requests across %d engines and %d contexts\n",
+		count, flags, smoke->gt->info.num_engines, smoke->ncontext);
+	return 0;
+}
+
+static int live_preempt_smoke(void *arg)
+{
+	struct preempt_smoke smoke = {
+		.gt = arg,
+		.prng = I915_RND_STATE_INITIALIZER(i915_selftest.random_seed),
+		.ncontext = 256,
+	};
+	const unsigned int phase[] = { 0, BATCH };
+	struct igt_live_test t;
+	int err = -ENOMEM;
+	u32 *cs;
+	int n;
+
+	smoke.contexts = kmalloc_array(smoke.ncontext,
+				       sizeof(*smoke.contexts),
+				       GFP_KERNEL);
+	if (!smoke.contexts)
+		return -ENOMEM;
+
+	smoke.batch =
+		i915_gem_object_create_internal(smoke.gt->i915, PAGE_SIZE);
+	if (IS_ERR(smoke.batch)) {
+		err = PTR_ERR(smoke.batch);
+		goto err_free;
+	}
+
+	cs = i915_gem_object_pin_map_unlocked(smoke.batch, I915_MAP_WB);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_batch;
+	}
+	for (n = 0; n < PAGE_SIZE / sizeof(*cs) - 1; n++)
+		cs[n] = MI_ARB_CHECK;
+	cs[n] = MI_BATCH_BUFFER_END;
+	i915_gem_object_flush_map(smoke.batch);
+	i915_gem_object_unpin_map(smoke.batch);
+
+	if (igt_live_test_begin(&t, smoke.gt->i915, __func__, "all")) {
+		err = -EIO;
+		goto err_batch;
+	}
+
+	for (n = 0; n < smoke.ncontext; n++) {
+		smoke.contexts[n] = kernel_context(smoke.gt->i915, NULL);
+		if (!smoke.contexts[n])
+			goto err_ctx;
+	}
+
+	for (n = 0; n < ARRAY_SIZE(phase); n++) {
+		err = smoke_crescendo(&smoke, phase[n]);
+		if (err)
+			goto err_ctx;
+
+		err = smoke_random(&smoke, phase[n]);
+		if (err)
+			goto err_ctx;
+	}
+
+err_ctx:
+	if (igt_live_test_end(&t))
+		err = -EIO;
+
+	for (n = 0; n < smoke.ncontext; n++) {
+		if (!smoke.contexts[n])
+			break;
+		kernel_context_close(smoke.contexts[n]);
+	}
+
+err_batch:
+	i915_gem_object_put(smoke.batch);
+err_free:
+	kfree(smoke.contexts);
+
+	return err;
+}
+
+static int nop_virtual_engine(struct intel_gt *gt,
+			      struct intel_engine_cs **siblings,
+			      unsigned int nsibling,
+			      unsigned int nctx,
+			      unsigned int flags)
+#define CHAIN BIT(0)
+{
+	IGT_TIMEOUT(end_time);
+	struct i915_request *request[16] = {};
+	struct intel_context *ve[16];
+	unsigned long n, prime, nc;
+	struct igt_live_test t;
+	ktime_t times[2] = {};
+	int err;
+
+	GEM_BUG_ON(!nctx || nctx > ARRAY_SIZE(ve));
+
+	for (n = 0; n < nctx; n++) {
+		ve[n] = intel_engine_create_virtual(siblings, nsibling, 0);
+		if (IS_ERR(ve[n])) {
+			err = PTR_ERR(ve[n]);
+			nctx = n;
+			goto out;
+		}
+
+		err = intel_context_pin(ve[n]);
+		if (err) {
+			intel_context_put(ve[n]);
+			nctx = n;
+			goto out;
+		}
+	}
+
+	err = igt_live_test_begin(&t, gt->i915, __func__, ve[0]->engine->name);
+	if (err)
+		goto out;
+
+	for_each_prime_number_from(prime, 1, 8192) {
+		times[1] = ktime_get_raw();
+
+		if (flags & CHAIN) {
+			for (nc = 0; nc < nctx; nc++) {
+				for (n = 0; n < prime; n++) {
+					struct i915_request *rq;
+
+					rq = i915_request_create(ve[nc]);
+					if (IS_ERR(rq)) {
+						err = PTR_ERR(rq);
+						goto out;
+					}
+
+					if (request[nc])
+						i915_request_put(request[nc]);
+					request[nc] = i915_request_get(rq);
+					i915_request_add(rq);
+				}
+			}
+		} else {
+			for (n = 0; n < prime; n++) {
+				for (nc = 0; nc < nctx; nc++) {
+					struct i915_request *rq;
+
+					rq = i915_request_create(ve[nc]);
+					if (IS_ERR(rq)) {
+						err = PTR_ERR(rq);
+						goto out;
+					}
+
+					if (request[nc])
+						i915_request_put(request[nc]);
+					request[nc] = i915_request_get(rq);
+					i915_request_add(rq);
+				}
+			}
+		}
+
+		for (nc = 0; nc < nctx; nc++) {
+			if (i915_request_wait(request[nc], 0, HZ / 10) < 0) {
+				pr_err("%s(%s): wait for %llx:%lld timed out\n",
+				       __func__, ve[0]->engine->name,
+				       request[nc]->fence.context,
+				       request[nc]->fence.seqno);
+
+				GEM_TRACE("%s(%s) failed at request %llx:%lld\n",
+					  __func__, ve[0]->engine->name,
+					  request[nc]->fence.context,
+					  request[nc]->fence.seqno);
+				GEM_TRACE_DUMP();
+				intel_gt_set_wedged(gt);
+				break;
+			}
+		}
+
+		times[1] = ktime_sub(ktime_get_raw(), times[1]);
+		if (prime == 1)
+			times[0] = times[1];
+
+		for (nc = 0; nc < nctx; nc++) {
+			i915_request_put(request[nc]);
+			request[nc] = NULL;
+		}
+
+		if (__igt_timeout(end_time, NULL))
+			break;
+	}
+
+	err = igt_live_test_end(&t);
+	if (err)
+		goto out;
+
+	pr_info("Requestx%d latencies on %s: 1 = %lluns, %lu = %lluns\n",
+		nctx, ve[0]->engine->name, ktime_to_ns(times[0]),
+		prime, div64_u64(ktime_to_ns(times[1]), prime));
+
+out:
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	for (nc = 0; nc < nctx; nc++) {
+		i915_request_put(request[nc]);
+		intel_context_unpin(ve[nc]);
+		intel_context_put(ve[nc]);
+	}
+	return err;
+}
+
+static unsigned int
+__select_siblings(struct intel_gt *gt,
+		  unsigned int class,
+		  struct intel_engine_cs **siblings,
+		  bool (*filter)(const struct intel_engine_cs *))
+{
+	unsigned int n = 0;
+	unsigned int inst;
+
+	for (inst = 0; inst <= MAX_ENGINE_INSTANCE; inst++) {
+		if (!gt->engine_class[class][inst])
+			continue;
+
+		if (filter && !filter(gt->engine_class[class][inst]))
+			continue;
+
+		siblings[n++] = gt->engine_class[class][inst];
+	}
+
+	return n;
+}
+
+static unsigned int
+select_siblings(struct intel_gt *gt,
+		unsigned int class,
+		struct intel_engine_cs **siblings)
+{
+	return __select_siblings(gt, class, siblings, NULL);
+}
+
+static int live_virtual_engine(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	unsigned int class;
+	int err;
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		err = nop_virtual_engine(gt, &engine, 1, 1, 0);
+		if (err) {
+			pr_err("Failed to wrap engine %s: err=%d\n",
+			       engine->name, err);
+			return err;
+		}
+	}
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		int nsibling, n;
+
+		nsibling = select_siblings(gt, class, siblings);
+		if (nsibling < 2)
+			continue;
+
+		for (n = 1; n <= nsibling + 1; n++) {
+			err = nop_virtual_engine(gt, siblings, nsibling,
+						 n, 0);
+			if (err)
+				return err;
+		}
+
+		err = nop_virtual_engine(gt, siblings, nsibling, n, CHAIN);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int mask_virtual_engine(struct intel_gt *gt,
+			       struct intel_engine_cs **siblings,
+			       unsigned int nsibling)
+{
+	struct i915_request *request[MAX_ENGINE_INSTANCE + 1];
+	struct intel_context *ve;
+	struct igt_live_test t;
+	unsigned int n;
+	int err;
+
+	/*
+	 * Check that by setting the execution mask on a request, we can
+	 * restrict it to our desired engine within the virtual engine.
+	 */
+
+	ve = intel_engine_create_virtual(siblings, nsibling, 0);
+	if (IS_ERR(ve)) {
+		err = PTR_ERR(ve);
+		goto out_close;
+	}
+
+	err = intel_context_pin(ve);
+	if (err)
+		goto out_put;
+
+	err = igt_live_test_begin(&t, gt->i915, __func__, ve->engine->name);
+	if (err)
+		goto out_unpin;
+
+	for (n = 0; n < nsibling; n++) {
+		request[n] = i915_request_create(ve);
+		if (IS_ERR(request[n])) {
+			err = PTR_ERR(request[n]);
+			nsibling = n;
+			goto out;
+		}
+
+		/* Reverse order as it's more likely to be unnatural */
+		request[n]->execution_mask = siblings[nsibling - n - 1]->mask;
+
+		i915_request_get(request[n]);
+		i915_request_add(request[n]);
+	}
+
+	for (n = 0; n < nsibling; n++) {
+		if (i915_request_wait(request[n], 0, HZ / 10) < 0) {
+			pr_err("%s(%s): wait for %llx:%lld timed out\n",
+			       __func__, ve->engine->name,
+			       request[n]->fence.context,
+			       request[n]->fence.seqno);
+
+			GEM_TRACE("%s(%s) failed at request %llx:%lld\n",
+				  __func__, ve->engine->name,
+				  request[n]->fence.context,
+				  request[n]->fence.seqno);
+			GEM_TRACE_DUMP();
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+			goto out;
+		}
+
+		if (request[n]->engine != siblings[nsibling - n - 1]) {
+			pr_err("Executed on wrong sibling '%s', expected '%s'\n",
+			       request[n]->engine->name,
+			       siblings[nsibling - n - 1]->name);
+			err = -EINVAL;
+			goto out;
+		}
+	}
+
+	err = igt_live_test_end(&t);
+out:
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	for (n = 0; n < nsibling; n++)
+		i915_request_put(request[n]);
+
+out_unpin:
+	intel_context_unpin(ve);
+out_put:
+	intel_context_put(ve);
+out_close:
+	return err;
+}
+
+static int live_virtual_mask(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	unsigned int class;
+	int err;
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		unsigned int nsibling;
+
+		nsibling = select_siblings(gt, class, siblings);
+		if (nsibling < 2)
+			continue;
+
+		err = mask_virtual_engine(gt, siblings, nsibling);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int slicein_virtual_engine(struct intel_gt *gt,
+				  struct intel_engine_cs **siblings,
+				  unsigned int nsibling)
+{
+	const long timeout = slice_timeout(siblings[0]);
+	struct intel_context *ce;
+	struct i915_request *rq;
+	struct igt_spinner spin;
+	unsigned int n;
+	int err = 0;
+
+	/*
+	 * Virtual requests must take part in timeslicing on the target engines.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	for (n = 0; n < nsibling; n++) {
+		ce = intel_context_create(siblings[n]);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out;
+		}
+
+		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out;
+		}
+
+		i915_request_add(rq);
+	}
+
+	ce = intel_engine_create_virtual(siblings, nsibling, 0);
+	if (IS_ERR(ce)) {
+		err = PTR_ERR(ce);
+		goto out;
+	}
+
+	rq = intel_context_create_request(ce);
+	intel_context_put(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out;
+	}
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (i915_request_wait(rq, 0, timeout) < 0) {
+		GEM_TRACE_ERR("%s(%s) failed to slice in virtual request\n",
+			      __func__, rq->engine->name);
+		GEM_TRACE_DUMP();
+		intel_gt_set_wedged(gt);
+		err = -EIO;
+	}
+	i915_request_put(rq);
+
+out:
+	igt_spinner_end(&spin);
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int sliceout_virtual_engine(struct intel_gt *gt,
+				   struct intel_engine_cs **siblings,
+				   unsigned int nsibling)
+{
+	const long timeout = slice_timeout(siblings[0]);
+	struct intel_context *ce;
+	struct i915_request *rq;
+	struct igt_spinner spin;
+	unsigned int n;
+	int err = 0;
+
+	/*
+	 * Virtual requests must allow others a fair timeslice.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	/* XXX We do not handle oversubscription and fairness with normal rq */
+	for (n = 0; n < nsibling; n++) {
+		ce = intel_engine_create_virtual(siblings, nsibling, 0);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out;
+		}
+
+		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out;
+		}
+
+		i915_request_add(rq);
+	}
+
+	for (n = 0; !err && n < nsibling; n++) {
+		ce = intel_context_create(siblings[n]);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out;
+		}
+
+		rq = intel_context_create_request(ce);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out;
+		}
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+		if (i915_request_wait(rq, 0, timeout) < 0) {
+			GEM_TRACE_ERR("%s(%s) failed to slice out virtual request\n",
+				      __func__, siblings[n]->name);
+			GEM_TRACE_DUMP();
+			intel_gt_set_wedged(gt);
+			err = -EIO;
+		}
+		i915_request_put(rq);
+	}
+
+out:
+	igt_spinner_end(&spin);
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_virtual_slice(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	unsigned int class;
+	int err;
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		unsigned int nsibling;
+
+		nsibling = __select_siblings(gt, class, siblings,
+					     intel_engine_has_timeslices);
+		if (nsibling < 2)
+			continue;
+
+		err = slicein_virtual_engine(gt, siblings, nsibling);
+		if (err)
+			return err;
+
+		err = sliceout_virtual_engine(gt, siblings, nsibling);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int preserved_virtual_engine(struct intel_gt *gt,
+				    struct intel_engine_cs **siblings,
+				    unsigned int nsibling)
+{
+	struct i915_request *last = NULL;
+	struct intel_context *ve;
+	struct i915_vma *scratch;
+	struct igt_live_test t;
+	unsigned int n;
+	int err = 0;
+	u32 *cs;
+
+	scratch =
+		__vm_create_scratch_for_read_pinned(&siblings[0]->gt->ggtt->vm,
+						    PAGE_SIZE);
+	if (IS_ERR(scratch))
+		return PTR_ERR(scratch);
+
+	err = i915_vma_sync(scratch);
+	if (err)
+		goto out_scratch;
+
+	ve = intel_engine_create_virtual(siblings, nsibling, 0);
+	if (IS_ERR(ve)) {
+		err = PTR_ERR(ve);
+		goto out_scratch;
+	}
+
+	err = intel_context_pin(ve);
+	if (err)
+		goto out_put;
+
+	err = igt_live_test_begin(&t, gt->i915, __func__, ve->engine->name);
+	if (err)
+		goto out_unpin;
+
+	for (n = 0; n < NUM_GPR_DW; n++) {
+		struct intel_engine_cs *engine = siblings[n % nsibling];
+		struct i915_request *rq;
+
+		rq = i915_request_create(ve);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_end;
+		}
+
+		i915_request_put(last);
+		last = i915_request_get(rq);
+
+		cs = intel_ring_begin(rq, 8);
+		if (IS_ERR(cs)) {
+			i915_request_add(rq);
+			err = PTR_ERR(cs);
+			goto out_end;
+		}
+
+		*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+		*cs++ = CS_GPR(engine, n);
+		*cs++ = i915_ggtt_offset(scratch) + n * sizeof(u32);
+		*cs++ = 0;
+
+		*cs++ = MI_LOAD_REGISTER_IMM(1);
+		*cs++ = CS_GPR(engine, (n + 1) % NUM_GPR_DW);
+		*cs++ = n + 1;
+
+		*cs++ = MI_NOOP;
+		intel_ring_advance(rq, cs);
+
+		/* Restrict this request to run on a particular engine */
+		rq->execution_mask = engine->mask;
+		i915_request_add(rq);
+	}
+
+	if (i915_request_wait(last, 0, HZ / 5) < 0) {
+		err = -ETIME;
+		goto out_end;
+	}
+
+	cs = i915_gem_object_pin_map_unlocked(scratch->obj, I915_MAP_WB);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto out_end;
+	}
+
+	for (n = 0; n < NUM_GPR_DW; n++) {
+		if (cs[n] != n) {
+			pr_err("Incorrect value[%d] found for GPR[%d]\n",
+			       cs[n], n);
+			err = -EINVAL;
+			break;
+		}
+	}
+
+	i915_gem_object_unpin_map(scratch->obj);
+
+out_end:
+	if (igt_live_test_end(&t))
+		err = -EIO;
+	i915_request_put(last);
+out_unpin:
+	intel_context_unpin(ve);
+out_put:
+	intel_context_put(ve);
+out_scratch:
+	i915_vma_unpin_and_release(&scratch, 0);
+	return err;
+}
+
+static int live_virtual_preserved(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	unsigned int class;
+
+	/*
+	 * Check that the context image retains non-privileged (user) registers
+	 * from one engine to the next. For this we check that the CS_GPR
+	 * are preserved.
+	 */
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	/* As we use CS_GPR we cannot run before they existed on all engines. */
+	if (GRAPHICS_VER(gt->i915) < 9)
+		return 0;
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		int nsibling, err;
+
+		nsibling = select_siblings(gt, class, siblings);
+		if (nsibling < 2)
+			continue;
+
+		err = preserved_virtual_engine(gt, siblings, nsibling);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int reset_virtual_engine(struct intel_gt *gt,
+				struct intel_engine_cs **siblings,
+				unsigned int nsibling)
+{
+	struct intel_engine_cs *engine;
+	struct intel_context *ve;
+	struct igt_spinner spin;
+	struct i915_request *rq;
+	unsigned int n;
+	int err = 0;
+
+	/*
+	 * In order to support offline error capture for fast preempt reset,
+	 * we need to decouple the guilty request and ensure that it and its
+	 * descendents are not executed while the capture is in progress.
+	 */
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	ve = intel_engine_create_virtual(siblings, nsibling, 0);
+	if (IS_ERR(ve)) {
+		err = PTR_ERR(ve);
+		goto out_spin;
+	}
+
+	for (n = 0; n < nsibling; n++)
+		st_engine_heartbeat_disable(siblings[n]);
+
+	rq = igt_spinner_create_request(&spin, ve, MI_ARB_CHECK);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_heartbeat;
+	}
+	i915_request_add(rq);
+
+	if (!igt_wait_for_spinner(&spin, rq)) {
+		intel_gt_set_wedged(gt);
+		err = -ETIME;
+		goto out_heartbeat;
+	}
+
+	engine = rq->engine;
+	GEM_BUG_ON(engine == ve->engine);
+
+	/* Take ownership of the reset and tasklet */
+	err = engine_lock_reset_tasklet(engine);
+	if (err)
+		goto out_heartbeat;
+
+	engine->sched_engine->tasklet.callback(&engine->sched_engine->tasklet);
+	GEM_BUG_ON(execlists_active(&engine->execlists) != rq);
+
+	/* Fake a preemption event; failed of course */
+	spin_lock_irq(&engine->sched_engine->lock);
+	__unwind_incomplete_requests(engine);
+	spin_unlock_irq(&engine->sched_engine->lock);
+	GEM_BUG_ON(rq->engine != engine);
+
+	/* Reset the engine while keeping our active request on hold */
+	execlists_hold(engine, rq);
+	GEM_BUG_ON(!i915_request_on_hold(rq));
+
+	__intel_engine_reset_bh(engine, NULL);
+	GEM_BUG_ON(rq->fence.error != -EIO);
+
+	/* Release our grasp on the engine, letting CS flow again */
+	engine_unlock_reset_tasklet(engine);
+
+	/* Check that we do not resubmit the held request */
+	i915_request_get(rq);
+	if (!i915_request_wait(rq, 0, HZ / 5)) {
+		pr_err("%s: on hold request completed!\n",
+		       engine->name);
+		intel_gt_set_wedged(gt);
+		err = -EIO;
+		goto out_rq;
+	}
+	GEM_BUG_ON(!i915_request_on_hold(rq));
+
+	/* But is resubmitted on release */
+	execlists_unhold(engine, rq);
+	if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+		pr_err("%s: held request did not complete!\n",
+		       engine->name);
+		intel_gt_set_wedged(gt);
+		err = -ETIME;
+	}
+
+out_rq:
+	i915_request_put(rq);
+out_heartbeat:
+	for (n = 0; n < nsibling; n++)
+		st_engine_heartbeat_enable(siblings[n]);
+
+	intel_context_put(ve);
+out_spin:
+	igt_spinner_fini(&spin);
+	return err;
+}
+
+static int live_virtual_reset(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	unsigned int class;
+
+	/*
+	 * Check that we handle a reset event within a virtual engine.
+	 * Only the physical engine is reset, but we have to check the flow
+	 * of the virtual requests around the reset, and make sure it is not
+	 * forgotten.
+	 */
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	for (class = 0; class <= MAX_ENGINE_CLASS; class++) {
+		int nsibling, err;
+
+		nsibling = select_siblings(gt, class, siblings);
+		if (nsibling < 2)
+			continue;
+
+		err = reset_virtual_engine(gt, siblings, nsibling);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int intel_execlists_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_sanitycheck),
+		SUBTEST(live_unlite_switch),
+		SUBTEST(live_unlite_preempt),
+		SUBTEST(live_unlite_ring),
+		SUBTEST(live_pin_rewind),
+		SUBTEST(live_hold_reset),
+		SUBTEST(live_error_interrupt),
+		SUBTEST(live_timeslice_preempt),
+		SUBTEST(live_timeslice_rewind),
+		SUBTEST(live_timeslice_queue),
+		SUBTEST(live_timeslice_nopreempt),
+		SUBTEST(live_busywait_preempt),
+		SUBTEST(live_preempt),
+		SUBTEST(live_late_preempt),
+		SUBTEST(live_nopreempt),
+		SUBTEST(live_preempt_cancel),
+		SUBTEST(live_suppress_self_preempt),
+		SUBTEST(live_chain_preempt),
+		SUBTEST(live_preempt_ring),
+		SUBTEST(live_preempt_gang),
+		SUBTEST(live_preempt_timeout),
+		SUBTEST(live_preempt_user),
+		SUBTEST(live_preempt_smoke),
+		SUBTEST(live_virtual_engine),
+		SUBTEST(live_virtual_mask),
+		SUBTEST(live_virtual_preserved),
+		SUBTEST(live_virtual_slice),
+		SUBTEST(live_virtual_reset),
+	};
+
+	if (to_gt(i915)->submission_method != INTEL_SUBMISSION_ELSP)
+		return 0;
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_gt_pm.c b/drivers/gpu/drm/i915/gt/selftest_gt_pm.c
new file mode 100644
index 000000000..be94f863b
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_gt_pm.c
@@ -0,0 +1,218 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/sort.h>
+
+#include "intel_engine_regs.h"
+#include "intel_gt_clock_utils.h"
+
+#include "selftest_llc.h"
+#include "selftest_rc6.h"
+#include "selftest_rps.h"
+
+static int cmp_u64(const void *A, const void *B)
+{
+	const u64 *a = A, *b = B;
+
+	if (a < b)
+		return -1;
+	else if (a > b)
+		return 1;
+	else
+		return 0;
+}
+
+static int cmp_u32(const void *A, const void *B)
+{
+	const u32 *a = A, *b = B;
+
+	if (a < b)
+		return -1;
+	else if (a > b)
+		return 1;
+	else
+		return 0;
+}
+
+static void measure_clocks(struct intel_engine_cs *engine,
+			   u32 *out_cycles, ktime_t *out_dt)
+{
+	ktime_t dt[5];
+	u32 cycles[5];
+	int i;
+
+	for (i = 0; i < 5; i++) {
+		local_irq_disable();
+		cycles[i] = -ENGINE_READ_FW(engine, RING_TIMESTAMP);
+		dt[i] = ktime_get();
+
+		udelay(1000);
+
+		dt[i] = ktime_sub(ktime_get(), dt[i]);
+		cycles[i] += ENGINE_READ_FW(engine, RING_TIMESTAMP);
+		local_irq_enable();
+	}
+
+	/* Use the median of both cycle/dt; close enough */
+	sort(cycles, 5, sizeof(*cycles), cmp_u32, NULL);
+	*out_cycles = (cycles[1] + 2 * cycles[2] + cycles[3]) / 4;
+
+	sort(dt, 5, sizeof(*dt), cmp_u64, NULL);
+	*out_dt = div_u64(dt[1] + 2 * dt[2] + dt[3], 4);
+}
+
+static int live_gt_clocks(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	if (!gt->clock_frequency) { /* unknown */
+		pr_info("CS_TIMESTAMP frequency unknown\n");
+		return 0;
+	}
+
+	if (GRAPHICS_VER(gt->i915) < 4) /* Any CS_TIMESTAMP? */
+		return 0;
+
+	if (GRAPHICS_VER(gt->i915) == 5)
+		/*
+		 * XXX CS_TIMESTAMP low dword is dysfunctional?
+		 *
+		 * Ville's experiments indicate the high dword still works,
+		 * but at a correspondingly reduced frequency.
+		 */
+		return 0;
+
+	if (GRAPHICS_VER(gt->i915) == 4)
+		/*
+		 * XXX CS_TIMESTAMP appears gibberish
+		 *
+		 * Ville's experiments indicate that it mostly appears 'stuck'
+		 * in that we see the register report the same cycle count
+		 * for a couple of reads.
+		 */
+		return 0;
+
+	intel_gt_pm_get(gt);
+	intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);
+
+	for_each_engine(engine, gt, id) {
+		u32 cycles;
+		u32 expected;
+		u64 time;
+		u64 dt;
+
+		if (GRAPHICS_VER(engine->i915) < 7 && engine->id != RCS0)
+			continue;
+
+		measure_clocks(engine, &cycles, &dt);
+
+		time = intel_gt_clock_interval_to_ns(engine->gt, cycles);
+		expected = intel_gt_ns_to_clock_interval(engine->gt, dt);
+
+		pr_info("%s: TIMESTAMP %d cycles [%lldns] in %lldns [%d cycles], using CS clock frequency of %uKHz\n",
+			engine->name, cycles, time, dt, expected,
+			engine->gt->clock_frequency / 1000);
+
+		if (9 * time < 8 * dt || 8 * time > 9 * dt) {
+			pr_err("%s: CS ticks did not match walltime!\n",
+			       engine->name);
+			err = -EINVAL;
+			break;
+		}
+
+		if (9 * expected < 8 * cycles || 8 * expected > 9 * cycles) {
+			pr_err("%s: walltime did not match CS ticks!\n",
+			       engine->name);
+			err = -EINVAL;
+			break;
+		}
+	}
+
+	intel_uncore_forcewake_put(gt->uncore, FORCEWAKE_ALL);
+	intel_gt_pm_put(gt);
+
+	return err;
+}
+
+static int live_gt_resume(void *arg)
+{
+	struct intel_gt *gt = arg;
+	IGT_TIMEOUT(end_time);
+	int err;
+
+	/* Do several suspend/resume cycles to check we don't explode! */
+	do {
+		intel_gt_suspend_prepare(gt);
+		intel_gt_suspend_late(gt);
+
+		if (gt->rc6.enabled) {
+			pr_err("rc6 still enabled after suspend!\n");
+			intel_gt_set_wedged_on_init(gt);
+			err = -EINVAL;
+			break;
+		}
+
+		err = intel_gt_resume(gt);
+		if (err)
+			break;
+
+		if (gt->rc6.supported && !gt->rc6.enabled) {
+			pr_err("rc6 not enabled upon resume!\n");
+			intel_gt_set_wedged_on_init(gt);
+			err = -EINVAL;
+			break;
+		}
+
+		err = st_llc_verify(&gt->llc);
+		if (err) {
+			pr_err("llc state not restored upon resume!\n");
+			intel_gt_set_wedged_on_init(gt);
+			break;
+		}
+	} while (!__igt_timeout(end_time, NULL));
+
+	return err;
+}
+
+int intel_gt_pm_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_gt_clocks),
+		SUBTEST(live_rc6_manual),
+		SUBTEST(live_rps_clock_interval),
+		SUBTEST(live_rps_control),
+		SUBTEST(live_rps_frequency_cs),
+		SUBTEST(live_rps_frequency_srm),
+		SUBTEST(live_rps_power),
+		SUBTEST(live_rps_interrupt),
+		SUBTEST(live_rps_dynamic),
+		SUBTEST(live_gt_resume),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
+
+int intel_gt_pm_late_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		/*
+		 * These tests may leave the system in an undesirable state.
+		 * They are intended to be run last in CI and the system
+		 * rebooted afterwards.
+		 */
+		SUBTEST(live_rc6_ctx_wa),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_hangcheck.c b/drivers/gpu/drm/i915/gt/selftest_hangcheck.c
new file mode 100644
index 000000000..71263058a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_hangcheck.c
@@ -0,0 +1,2060 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include <linux/kthread.h>
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_internal.h"
+
+#include "i915_gem_evict.h"
+#include "intel_gt.h"
+#include "intel_engine_heartbeat.h"
+#include "intel_engine_pm.h"
+#include "selftest_engine_heartbeat.h"
+
+#include "i915_selftest.h"
+#include "selftests/i915_random.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/igt_reset.h"
+#include "selftests/igt_atomic.h"
+#include "selftests/igt_spinner.h"
+#include "selftests/intel_scheduler_helpers.h"
+
+#include "selftests/mock_drm.h"
+
+#include "gem/selftests/mock_context.h"
+#include "gem/selftests/igt_gem_utils.h"
+
+#define IGT_IDLE_TIMEOUT 50 /* ms; time to wait after flushing between tests */
+
+struct hang {
+	struct intel_gt *gt;
+	struct drm_i915_gem_object *hws;
+	struct drm_i915_gem_object *obj;
+	struct i915_gem_context *ctx;
+	u32 *seqno;
+	u32 *batch;
+};
+
+static int hang_init(struct hang *h, struct intel_gt *gt)
+{
+	void *vaddr;
+	int err;
+
+	memset(h, 0, sizeof(*h));
+	h->gt = gt;
+
+	h->ctx = kernel_context(gt->i915, NULL);
+	if (IS_ERR(h->ctx))
+		return PTR_ERR(h->ctx);
+
+	GEM_BUG_ON(i915_gem_context_is_bannable(h->ctx));
+
+	h->hws = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(h->hws)) {
+		err = PTR_ERR(h->hws);
+		goto err_ctx;
+	}
+
+	h->obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(h->obj)) {
+		err = PTR_ERR(h->obj);
+		goto err_hws;
+	}
+
+	i915_gem_object_set_cache_coherency(h->hws, I915_CACHE_LLC);
+	vaddr = i915_gem_object_pin_map_unlocked(h->hws, I915_MAP_WB);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto err_obj;
+	}
+	h->seqno = memset(vaddr, 0xff, PAGE_SIZE);
+
+	vaddr = i915_gem_object_pin_map_unlocked(h->obj,
+						 i915_coherent_map_type(gt->i915, h->obj, false));
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto err_unpin_hws;
+	}
+	h->batch = vaddr;
+
+	return 0;
+
+err_unpin_hws:
+	i915_gem_object_unpin_map(h->hws);
+err_obj:
+	i915_gem_object_put(h->obj);
+err_hws:
+	i915_gem_object_put(h->hws);
+err_ctx:
+	kernel_context_close(h->ctx);
+	return err;
+}
+
+static u64 hws_address(const struct i915_vma *hws,
+		       const struct i915_request *rq)
+{
+	return hws->node.start + offset_in_page(sizeof(u32)*rq->fence.context);
+}
+
+static int move_to_active(struct i915_vma *vma,
+			  struct i915_request *rq,
+			  unsigned int flags)
+{
+	int err;
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj,
+					flags & EXEC_OBJECT_WRITE);
+	if (err == 0)
+		err = i915_vma_move_to_active(vma, rq, flags);
+	i915_vma_unlock(vma);
+
+	return err;
+}
+
+static struct i915_request *
+hang_create_request(struct hang *h, struct intel_engine_cs *engine)
+{
+	struct intel_gt *gt = h->gt;
+	struct i915_address_space *vm = i915_gem_context_get_eb_vm(h->ctx);
+	struct drm_i915_gem_object *obj;
+	struct i915_request *rq = NULL;
+	struct i915_vma *hws, *vma;
+	unsigned int flags;
+	void *vaddr;
+	u32 *batch;
+	int err;
+
+	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		i915_vm_put(vm);
+		return ERR_CAST(obj);
+	}
+
+	vaddr = i915_gem_object_pin_map_unlocked(obj, i915_coherent_map_type(gt->i915, obj, false));
+	if (IS_ERR(vaddr)) {
+		i915_gem_object_put(obj);
+		i915_vm_put(vm);
+		return ERR_CAST(vaddr);
+	}
+
+	i915_gem_object_unpin_map(h->obj);
+	i915_gem_object_put(h->obj);
+
+	h->obj = obj;
+	h->batch = vaddr;
+
+	vma = i915_vma_instance(h->obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_vm_put(vm);
+		return ERR_CAST(vma);
+	}
+
+	hws = i915_vma_instance(h->hws, vm, NULL);
+	if (IS_ERR(hws)) {
+		i915_vm_put(vm);
+		return ERR_CAST(hws);
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err) {
+		i915_vm_put(vm);
+		return ERR_PTR(err);
+	}
+
+	err = i915_vma_pin(hws, 0, 0, PIN_USER);
+	if (err)
+		goto unpin_vma;
+
+	rq = igt_request_alloc(h->ctx, engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto unpin_hws;
+	}
+
+	err = move_to_active(vma, rq, 0);
+	if (err)
+		goto cancel_rq;
+
+	err = move_to_active(hws, rq, 0);
+	if (err)
+		goto cancel_rq;
+
+	batch = h->batch;
+	if (GRAPHICS_VER(gt->i915) >= 8) {
+		*batch++ = MI_STORE_DWORD_IMM_GEN4;
+		*batch++ = lower_32_bits(hws_address(hws, rq));
+		*batch++ = upper_32_bits(hws_address(hws, rq));
+		*batch++ = rq->fence.seqno;
+		*batch++ = MI_NOOP;
+
+		memset(batch, 0, 1024);
+		batch += 1024 / sizeof(*batch);
+
+		*batch++ = MI_NOOP;
+		*batch++ = MI_BATCH_BUFFER_START | 1 << 8 | 1;
+		*batch++ = lower_32_bits(vma->node.start);
+		*batch++ = upper_32_bits(vma->node.start);
+	} else if (GRAPHICS_VER(gt->i915) >= 6) {
+		*batch++ = MI_STORE_DWORD_IMM_GEN4;
+		*batch++ = 0;
+		*batch++ = lower_32_bits(hws_address(hws, rq));
+		*batch++ = rq->fence.seqno;
+		*batch++ = MI_NOOP;
+
+		memset(batch, 0, 1024);
+		batch += 1024 / sizeof(*batch);
+
+		*batch++ = MI_NOOP;
+		*batch++ = MI_BATCH_BUFFER_START | 1 << 8;
+		*batch++ = lower_32_bits(vma->node.start);
+	} else if (GRAPHICS_VER(gt->i915) >= 4) {
+		*batch++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*batch++ = 0;
+		*batch++ = lower_32_bits(hws_address(hws, rq));
+		*batch++ = rq->fence.seqno;
+		*batch++ = MI_NOOP;
+
+		memset(batch, 0, 1024);
+		batch += 1024 / sizeof(*batch);
+
+		*batch++ = MI_NOOP;
+		*batch++ = MI_BATCH_BUFFER_START | 2 << 6;
+		*batch++ = lower_32_bits(vma->node.start);
+	} else {
+		*batch++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL;
+		*batch++ = lower_32_bits(hws_address(hws, rq));
+		*batch++ = rq->fence.seqno;
+		*batch++ = MI_NOOP;
+
+		memset(batch, 0, 1024);
+		batch += 1024 / sizeof(*batch);
+
+		*batch++ = MI_NOOP;
+		*batch++ = MI_BATCH_BUFFER_START | 2 << 6;
+		*batch++ = lower_32_bits(vma->node.start);
+	}
+	*batch++ = MI_BATCH_BUFFER_END; /* not reached */
+	intel_gt_chipset_flush(engine->gt);
+
+	if (rq->engine->emit_init_breadcrumb) {
+		err = rq->engine->emit_init_breadcrumb(rq);
+		if (err)
+			goto cancel_rq;
+	}
+
+	flags = 0;
+	if (GRAPHICS_VER(gt->i915) <= 5)
+		flags |= I915_DISPATCH_SECURE;
+
+	err = rq->engine->emit_bb_start(rq, vma->node.start, PAGE_SIZE, flags);
+
+cancel_rq:
+	if (err) {
+		i915_request_set_error_once(rq, err);
+		i915_request_add(rq);
+	}
+unpin_hws:
+	i915_vma_unpin(hws);
+unpin_vma:
+	i915_vma_unpin(vma);
+	i915_vm_put(vm);
+	return err ? ERR_PTR(err) : rq;
+}
+
+static u32 hws_seqno(const struct hang *h, const struct i915_request *rq)
+{
+	return READ_ONCE(h->seqno[rq->fence.context % (PAGE_SIZE/sizeof(u32))]);
+}
+
+static void hang_fini(struct hang *h)
+{
+	*h->batch = MI_BATCH_BUFFER_END;
+	intel_gt_chipset_flush(h->gt);
+
+	i915_gem_object_unpin_map(h->obj);
+	i915_gem_object_put(h->obj);
+
+	i915_gem_object_unpin_map(h->hws);
+	i915_gem_object_put(h->hws);
+
+	kernel_context_close(h->ctx);
+
+	igt_flush_test(h->gt->i915);
+}
+
+static bool wait_until_running(struct hang *h, struct i915_request *rq)
+{
+	return !(wait_for_us(i915_seqno_passed(hws_seqno(h, rq),
+					       rq->fence.seqno),
+			     10) &&
+		 wait_for(i915_seqno_passed(hws_seqno(h, rq),
+					    rq->fence.seqno),
+			  1000));
+}
+
+static int igt_hang_sanitycheck(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_request *rq;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct hang h;
+	int err;
+
+	/* Basic check that we can execute our hanging batch */
+
+	err = hang_init(&h, gt);
+	if (err)
+		return err;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_wedge_me w;
+		long timeout;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		rq = hang_create_request(&h, engine);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			pr_err("Failed to create request for %s, err=%d\n",
+			       engine->name, err);
+			goto fini;
+		}
+
+		i915_request_get(rq);
+
+		*h.batch = MI_BATCH_BUFFER_END;
+		intel_gt_chipset_flush(engine->gt);
+
+		i915_request_add(rq);
+
+		timeout = 0;
+		intel_wedge_on_timeout(&w, gt, HZ / 10 /* 100ms */)
+			timeout = i915_request_wait(rq, 0,
+						    MAX_SCHEDULE_TIMEOUT);
+		if (intel_gt_is_wedged(gt))
+			timeout = -EIO;
+
+		i915_request_put(rq);
+
+		if (timeout < 0) {
+			err = timeout;
+			pr_err("Wait for request failed on %s, err=%d\n",
+			       engine->name, err);
+			goto fini;
+		}
+	}
+
+fini:
+	hang_fini(&h);
+	return err;
+}
+
+static bool wait_for_idle(struct intel_engine_cs *engine)
+{
+	return wait_for(intel_engine_is_idle(engine), IGT_IDLE_TIMEOUT) == 0;
+}
+
+static int igt_reset_nop(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gpu_error *global = &gt->i915->gpu_error;
+	struct intel_engine_cs *engine;
+	unsigned int reset_count, count;
+	enum intel_engine_id id;
+	IGT_TIMEOUT(end_time);
+	int err = 0;
+
+	/* Check that we can reset during non-user portions of requests */
+
+	reset_count = i915_reset_count(global);
+	count = 0;
+	do {
+		for_each_engine(engine, gt, id) {
+			struct intel_context *ce;
+			int i;
+
+			ce = intel_context_create(engine);
+			if (IS_ERR(ce)) {
+				err = PTR_ERR(ce);
+				pr_err("[%s] Create context failed: %d!\n", engine->name, err);
+				break;
+			}
+
+			for (i = 0; i < 16; i++) {
+				struct i915_request *rq;
+
+				rq = intel_context_create_request(ce);
+				if (IS_ERR(rq)) {
+					err = PTR_ERR(rq);
+					pr_err("[%s] Create request failed: %d!\n",
+					       engine->name, err);
+					break;
+				}
+
+				i915_request_add(rq);
+			}
+
+			intel_context_put(ce);
+		}
+
+		igt_global_reset_lock(gt);
+		intel_gt_reset(gt, ALL_ENGINES, NULL);
+		igt_global_reset_unlock(gt);
+
+		if (intel_gt_is_wedged(gt)) {
+			pr_err("[%s] GT is wedged!\n", engine->name);
+			err = -EIO;
+			break;
+		}
+
+		if (i915_reset_count(global) != reset_count + ++count) {
+			pr_err("[%s] Reset not recorded: %d vs %d + %d!\n",
+			       engine->name, i915_reset_count(global), reset_count, count);
+			err = -EINVAL;
+			break;
+		}
+
+		err = igt_flush_test(gt->i915);
+		if (err) {
+			pr_err("[%s] Flush failed: %d!\n", engine->name, err);
+			break;
+		}
+	} while (time_before(jiffies, end_time));
+	pr_info("%s: %d resets\n", __func__, count);
+
+	if (igt_flush_test(gt->i915)) {
+		pr_err("Post flush failed: %d!\n", err);
+		err = -EIO;
+	}
+
+	return err;
+}
+
+static int igt_reset_nop_engine(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gpu_error *global = &gt->i915->gpu_error;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/* Check that we can engine-reset during non-user portions */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		unsigned int reset_count, reset_engine_count, count;
+		struct intel_context *ce;
+		IGT_TIMEOUT(end_time);
+		int err;
+
+		if (intel_engine_uses_guc(engine)) {
+			/* Engine level resets are triggered by GuC when a hang
+			 * is detected. They can't be triggered by the KMD any
+			 * more. Thus a nop batch cannot be used as a reset test
+			 */
+			continue;
+		}
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			pr_err("[%s] Create context failed: %pe!\n", engine->name, ce);
+			return PTR_ERR(ce);
+		}
+
+		reset_count = i915_reset_count(global);
+		reset_engine_count = i915_reset_engine_count(global, engine);
+		count = 0;
+
+		st_engine_heartbeat_disable(engine);
+		GEM_BUG_ON(test_and_set_bit(I915_RESET_ENGINE + id,
+					    &gt->reset.flags));
+		do {
+			int i;
+
+			if (!wait_for_idle(engine)) {
+				pr_err("%s failed to idle before reset\n",
+				       engine->name);
+				err = -EIO;
+				break;
+			}
+
+			for (i = 0; i < 16; i++) {
+				struct i915_request *rq;
+
+				rq = intel_context_create_request(ce);
+				if (IS_ERR(rq)) {
+					struct drm_printer p =
+						drm_info_printer(gt->i915->drm.dev);
+					intel_engine_dump(engine, &p,
+							  "%s(%s): failed to submit request\n",
+							  __func__,
+							  engine->name);
+
+					GEM_TRACE("%s(%s): failed to submit request\n",
+						  __func__,
+						  engine->name);
+					GEM_TRACE_DUMP();
+
+					intel_gt_set_wedged(gt);
+
+					err = PTR_ERR(rq);
+					break;
+				}
+
+				i915_request_add(rq);
+			}
+			err = intel_engine_reset(engine, NULL);
+			if (err) {
+				pr_err("intel_engine_reset(%s) failed, err:%d\n",
+				       engine->name, err);
+				break;
+			}
+
+			if (i915_reset_count(global) != reset_count) {
+				pr_err("Full GPU reset recorded! (engine reset expected)\n");
+				err = -EINVAL;
+				break;
+			}
+
+			if (i915_reset_engine_count(global, engine) !=
+			    reset_engine_count + ++count) {
+				pr_err("%s engine reset not recorded!\n",
+				       engine->name);
+				err = -EINVAL;
+				break;
+			}
+		} while (time_before(jiffies, end_time));
+		clear_and_wake_up_bit(I915_RESET_ENGINE + id, &gt->reset.flags);
+		st_engine_heartbeat_enable(engine);
+
+		pr_info("%s(%s): %d resets\n", __func__, engine->name, count);
+
+		intel_context_put(ce);
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static void force_reset_timeout(struct intel_engine_cs *engine)
+{
+	engine->reset_timeout.probability = 999;
+	atomic_set(&engine->reset_timeout.times, -1);
+}
+
+static void cancel_reset_timeout(struct intel_engine_cs *engine)
+{
+	memset(&engine->reset_timeout, 0, sizeof(engine->reset_timeout));
+}
+
+static int igt_reset_fail_engine(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/* Check that we can recover from engine-reset failues */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		unsigned int count;
+		struct intel_context *ce;
+		IGT_TIMEOUT(end_time);
+		int err;
+
+		/* Can't manually break the reset if i915 doesn't perform it */
+		if (intel_engine_uses_guc(engine))
+			continue;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			pr_err("[%s] Create context failed: %pe!\n", engine->name, ce);
+			return PTR_ERR(ce);
+		}
+
+		st_engine_heartbeat_disable(engine);
+		GEM_BUG_ON(test_and_set_bit(I915_RESET_ENGINE + id,
+					    &gt->reset.flags));
+
+		force_reset_timeout(engine);
+		err = intel_engine_reset(engine, NULL);
+		cancel_reset_timeout(engine);
+		if (err == 0) /* timeouts only generated on gen8+ */
+			goto skip;
+
+		count = 0;
+		do {
+			struct i915_request *last = NULL;
+			int i;
+
+			if (!wait_for_idle(engine)) {
+				pr_err("%s failed to idle before reset\n",
+				       engine->name);
+				err = -EIO;
+				break;
+			}
+
+			for (i = 0; i < count % 15; i++) {
+				struct i915_request *rq;
+
+				rq = intel_context_create_request(ce);
+				if (IS_ERR(rq)) {
+					struct drm_printer p =
+						drm_info_printer(gt->i915->drm.dev);
+					intel_engine_dump(engine, &p,
+							  "%s(%s): failed to submit request\n",
+							  __func__,
+							  engine->name);
+
+					GEM_TRACE("%s(%s): failed to submit request\n",
+						  __func__,
+						  engine->name);
+					GEM_TRACE_DUMP();
+
+					intel_gt_set_wedged(gt);
+					if (last)
+						i915_request_put(last);
+
+					err = PTR_ERR(rq);
+					goto out;
+				}
+
+				if (last)
+					i915_request_put(last);
+				last = i915_request_get(rq);
+				i915_request_add(rq);
+			}
+
+			if (count & 1) {
+				err = intel_engine_reset(engine, NULL);
+				if (err) {
+					GEM_TRACE_ERR("intel_engine_reset(%s) failed, err:%d\n",
+						      engine->name, err);
+					GEM_TRACE_DUMP();
+					i915_request_put(last);
+					break;
+				}
+			} else {
+				force_reset_timeout(engine);
+				err = intel_engine_reset(engine, NULL);
+				cancel_reset_timeout(engine);
+				if (err != -ETIMEDOUT) {
+					pr_err("intel_engine_reset(%s) did not fail, err:%d\n",
+					       engine->name, err);
+					i915_request_put(last);
+					break;
+				}
+			}
+
+			err = 0;
+			if (last) {
+				if (i915_request_wait(last, 0, HZ / 2) < 0) {
+					struct drm_printer p =
+						drm_info_printer(gt->i915->drm.dev);
+
+					intel_engine_dump(engine, &p,
+							  "%s(%s): failed to complete request\n",
+							  __func__,
+							  engine->name);
+
+					GEM_TRACE("%s(%s): failed to complete request\n",
+						  __func__,
+						  engine->name);
+					GEM_TRACE_DUMP();
+
+					err = -EIO;
+				}
+				i915_request_put(last);
+			}
+			count++;
+		} while (err == 0 && time_before(jiffies, end_time));
+out:
+		pr_info("%s(%s): %d resets\n", __func__, engine->name, count);
+skip:
+		clear_and_wake_up_bit(I915_RESET_ENGINE + id, &gt->reset.flags);
+		st_engine_heartbeat_enable(engine);
+		intel_context_put(ce);
+
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int __igt_reset_engine(struct intel_gt *gt, bool active)
+{
+	struct i915_gpu_error *global = &gt->i915->gpu_error;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct hang h;
+	int err = 0;
+
+	/* Check that we can issue an engine reset on an idle engine (no-op) */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	if (active) {
+		err = hang_init(&h, gt);
+		if (err)
+			return err;
+	}
+
+	for_each_engine(engine, gt, id) {
+		unsigned int reset_count, reset_engine_count;
+		unsigned long count;
+		bool using_guc = intel_engine_uses_guc(engine);
+		IGT_TIMEOUT(end_time);
+
+		if (using_guc && !active)
+			continue;
+
+		if (active && !intel_engine_can_store_dword(engine))
+			continue;
+
+		if (!wait_for_idle(engine)) {
+			pr_err("%s failed to idle before reset\n",
+			       engine->name);
+			err = -EIO;
+			break;
+		}
+
+		reset_count = i915_reset_count(global);
+		reset_engine_count = i915_reset_engine_count(global, engine);
+
+		st_engine_heartbeat_disable(engine);
+		GEM_BUG_ON(test_and_set_bit(I915_RESET_ENGINE + id,
+					    &gt->reset.flags));
+		count = 0;
+		do {
+			struct i915_request *rq = NULL;
+			struct intel_selftest_saved_policy saved;
+			int err2;
+
+			err = intel_selftest_modify_policy(engine, &saved,
+							   SELFTEST_SCHEDULER_MODIFY_FAST_RESET);
+			if (err) {
+				pr_err("[%s] Modify policy failed: %d!\n", engine->name, err);
+				break;
+			}
+
+			if (active) {
+				rq = hang_create_request(&h, engine);
+				if (IS_ERR(rq)) {
+					err = PTR_ERR(rq);
+					pr_err("[%s] Create hang request failed: %d!\n",
+					       engine->name, err);
+					goto restore;
+				}
+
+				i915_request_get(rq);
+				i915_request_add(rq);
+
+				if (!wait_until_running(&h, rq)) {
+					struct drm_printer p = drm_info_printer(gt->i915->drm.dev);
+
+					pr_err("%s: Failed to start request %llx, at %x\n",
+					       __func__, rq->fence.seqno, hws_seqno(&h, rq));
+					intel_engine_dump(engine, &p,
+							  "%s\n", engine->name);
+
+					i915_request_put(rq);
+					err = -EIO;
+					goto restore;
+				}
+			}
+
+			if (!using_guc) {
+				err = intel_engine_reset(engine, NULL);
+				if (err) {
+					pr_err("intel_engine_reset(%s) failed, err:%d\n",
+					       engine->name, err);
+					goto skip;
+				}
+			}
+
+			if (rq) {
+				/* Ensure the reset happens and kills the engine */
+				err = intel_selftest_wait_for_rq(rq);
+				if (err)
+					pr_err("[%s] Wait for request %lld:%lld [0x%04X] failed: %d!\n",
+					       engine->name, rq->fence.context,
+					       rq->fence.seqno, rq->context->guc_id.id, err);
+			}
+
+skip:
+			if (rq)
+				i915_request_put(rq);
+
+			if (i915_reset_count(global) != reset_count) {
+				pr_err("Full GPU reset recorded! (engine reset expected)\n");
+				err = -EINVAL;
+				goto restore;
+			}
+
+			/* GuC based resets are not logged per engine */
+			if (!using_guc) {
+				if (i915_reset_engine_count(global, engine) !=
+				    ++reset_engine_count) {
+					pr_err("%s engine reset not recorded!\n",
+					       engine->name);
+					err = -EINVAL;
+					goto restore;
+				}
+			}
+
+			count++;
+
+restore:
+			err2 = intel_selftest_restore_policy(engine, &saved);
+			if (err2)
+				pr_err("[%s] Restore policy failed: %d!\n", engine->name, err);
+			if (err == 0)
+				err = err2;
+			if (err)
+				break;
+		} while (time_before(jiffies, end_time));
+		clear_and_wake_up_bit(I915_RESET_ENGINE + id, &gt->reset.flags);
+		st_engine_heartbeat_enable(engine);
+		pr_info("%s: Completed %lu %s resets\n",
+			engine->name, count, active ? "active" : "idle");
+
+		if (err)
+			break;
+
+		err = igt_flush_test(gt->i915);
+		if (err) {
+			pr_err("[%s] Flush failed: %d!\n", engine->name, err);
+			break;
+		}
+	}
+
+	if (intel_gt_is_wedged(gt)) {
+		pr_err("GT is wedged!\n");
+		err = -EIO;
+	}
+
+	if (active)
+		hang_fini(&h);
+
+	return err;
+}
+
+static int igt_reset_idle_engine(void *arg)
+{
+	return __igt_reset_engine(arg, false);
+}
+
+static int igt_reset_active_engine(void *arg)
+{
+	return __igt_reset_engine(arg, true);
+}
+
+struct active_engine {
+	struct kthread_worker *worker;
+	struct kthread_work work;
+	struct intel_engine_cs *engine;
+	unsigned long resets;
+	unsigned int flags;
+	bool stop;
+	int result;
+};
+
+#define TEST_ACTIVE	BIT(0)
+#define TEST_OTHERS	BIT(1)
+#define TEST_SELF	BIT(2)
+#define TEST_PRIORITY	BIT(3)
+
+static int active_request_put(struct i915_request *rq)
+{
+	int err = 0;
+
+	if (!rq)
+		return 0;
+
+	if (i915_request_wait(rq, 0, 10 * HZ) < 0) {
+		GEM_TRACE("%s timed out waiting for completion of fence %llx:%lld\n",
+			  rq->engine->name,
+			  rq->fence.context,
+			  rq->fence.seqno);
+		GEM_TRACE_DUMP();
+
+		intel_gt_set_wedged(rq->engine->gt);
+		err = -EIO;
+	}
+
+	i915_request_put(rq);
+
+	return err;
+}
+
+static void active_engine(struct kthread_work *work)
+{
+	I915_RND_STATE(prng);
+	struct active_engine *arg = container_of(work, typeof(*arg), work);
+	struct intel_engine_cs *engine = arg->engine;
+	struct i915_request *rq[8] = {};
+	struct intel_context *ce[ARRAY_SIZE(rq)];
+	unsigned long count;
+	int err = 0;
+
+	for (count = 0; count < ARRAY_SIZE(ce); count++) {
+		ce[count] = intel_context_create(engine);
+		if (IS_ERR(ce[count])) {
+			arg->result = PTR_ERR(ce[count]);
+			pr_err("[%s] Create context #%ld failed: %d!\n",
+			       engine->name, count, arg->result);
+			while (--count)
+				intel_context_put(ce[count]);
+			return;
+		}
+	}
+
+	count = 0;
+	while (!READ_ONCE(arg->stop)) {
+		unsigned int idx = count++ & (ARRAY_SIZE(rq) - 1);
+		struct i915_request *old = rq[idx];
+		struct i915_request *new;
+
+		new = intel_context_create_request(ce[idx]);
+		if (IS_ERR(new)) {
+			err = PTR_ERR(new);
+			pr_err("[%s] Create request #%d failed: %d!\n", engine->name, idx, err);
+			break;
+		}
+
+		rq[idx] = i915_request_get(new);
+		i915_request_add(new);
+
+		if (engine->sched_engine->schedule && arg->flags & TEST_PRIORITY) {
+			struct i915_sched_attr attr = {
+				.priority =
+					i915_prandom_u32_max_state(512, &prng),
+			};
+			engine->sched_engine->schedule(rq[idx], &attr);
+		}
+
+		err = active_request_put(old);
+		if (err) {
+			pr_err("[%s] Request put failed: %d!\n", engine->name, err);
+			break;
+		}
+
+		cond_resched();
+	}
+
+	for (count = 0; count < ARRAY_SIZE(rq); count++) {
+		int err__ = active_request_put(rq[count]);
+
+		if (err)
+			pr_err("[%s] Request put #%ld failed: %d!\n", engine->name, count, err);
+
+		/* Keep the first error */
+		if (!err)
+			err = err__;
+
+		intel_context_put(ce[count]);
+	}
+
+	arg->result = err;
+}
+
+static int __igt_reset_engines(struct intel_gt *gt,
+			       const char *test_name,
+			       unsigned int flags)
+{
+	struct i915_gpu_error *global = &gt->i915->gpu_error;
+	struct intel_engine_cs *engine, *other;
+	struct active_engine *threads;
+	enum intel_engine_id id, tmp;
+	struct hang h;
+	int err = 0;
+
+	/* Check that issuing a reset on one engine does not interfere
+	 * with any other engine.
+	 */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	if (flags & TEST_ACTIVE) {
+		err = hang_init(&h, gt);
+		if (err)
+			return err;
+
+		if (flags & TEST_PRIORITY)
+			h.ctx->sched.priority = 1024;
+	}
+
+	threads = kmalloc_array(I915_NUM_ENGINES, sizeof(*threads), GFP_KERNEL);
+	if (!threads)
+		return -ENOMEM;
+
+	for_each_engine(engine, gt, id) {
+		unsigned long device = i915_reset_count(global);
+		unsigned long count = 0, reported;
+		bool using_guc = intel_engine_uses_guc(engine);
+		IGT_TIMEOUT(end_time);
+
+		if (flags & TEST_ACTIVE) {
+			if (!intel_engine_can_store_dword(engine))
+				continue;
+		} else if (using_guc)
+			continue;
+
+		if (!wait_for_idle(engine)) {
+			pr_err("i915_reset_engine(%s:%s): failed to idle before reset\n",
+			       engine->name, test_name);
+			err = -EIO;
+			break;
+		}
+
+		memset(threads, 0, sizeof(*threads) * I915_NUM_ENGINES);
+		for_each_engine(other, gt, tmp) {
+			struct kthread_worker *worker;
+
+			threads[tmp].resets =
+				i915_reset_engine_count(global, other);
+
+			if (other == engine && !(flags & TEST_SELF))
+				continue;
+
+			if (other != engine && !(flags & TEST_OTHERS))
+				continue;
+
+			threads[tmp].engine = other;
+			threads[tmp].flags = flags;
+
+			worker = kthread_create_worker(0, "igt/%s",
+						       other->name);
+			if (IS_ERR(worker)) {
+				err = PTR_ERR(worker);
+				pr_err("[%s] Worker create failed: %d!\n",
+				       engine->name, err);
+				goto unwind;
+			}
+
+			threads[tmp].worker = worker;
+
+			kthread_init_work(&threads[tmp].work, active_engine);
+			kthread_queue_work(threads[tmp].worker,
+					   &threads[tmp].work);
+		}
+
+		st_engine_heartbeat_disable_no_pm(engine);
+		GEM_BUG_ON(test_and_set_bit(I915_RESET_ENGINE + id,
+					    &gt->reset.flags));
+		do {
+			struct i915_request *rq = NULL;
+			struct intel_selftest_saved_policy saved;
+			int err2;
+
+			err = intel_selftest_modify_policy(engine, &saved,
+							   SELFTEST_SCHEDULER_MODIFY_FAST_RESET);
+			if (err) {
+				pr_err("[%s] Modify policy failed: %d!\n", engine->name, err);
+				break;
+			}
+
+			if (flags & TEST_ACTIVE) {
+				rq = hang_create_request(&h, engine);
+				if (IS_ERR(rq)) {
+					err = PTR_ERR(rq);
+					pr_err("[%s] Create hang request failed: %d!\n",
+					       engine->name, err);
+					goto restore;
+				}
+
+				i915_request_get(rq);
+				i915_request_add(rq);
+
+				if (!wait_until_running(&h, rq)) {
+					struct drm_printer p = drm_info_printer(gt->i915->drm.dev);
+
+					pr_err("%s: Failed to start request %llx, at %x\n",
+					       __func__, rq->fence.seqno, hws_seqno(&h, rq));
+					intel_engine_dump(engine, &p,
+							  "%s\n", engine->name);
+
+					i915_request_put(rq);
+					err = -EIO;
+					goto restore;
+				}
+			} else {
+				intel_engine_pm_get(engine);
+			}
+
+			if (!using_guc) {
+				err = intel_engine_reset(engine, NULL);
+				if (err) {
+					pr_err("i915_reset_engine(%s:%s): failed, err=%d\n",
+					       engine->name, test_name, err);
+					goto restore;
+				}
+			}
+
+			if (rq) {
+				/* Ensure the reset happens and kills the engine */
+				err = intel_selftest_wait_for_rq(rq);
+				if (err)
+					pr_err("[%s] Wait for request %lld:%lld [0x%04X] failed: %d!\n",
+					       engine->name, rq->fence.context,
+					       rq->fence.seqno, rq->context->guc_id.id, err);
+			}
+
+			count++;
+
+			if (rq) {
+				if (rq->fence.error != -EIO) {
+					pr_err("i915_reset_engine(%s:%s): failed to reset request %lld:%lld [0x%04X]\n",
+					       engine->name, test_name,
+					       rq->fence.context,
+					       rq->fence.seqno, rq->context->guc_id.id);
+					i915_request_put(rq);
+
+					GEM_TRACE_DUMP();
+					intel_gt_set_wedged(gt);
+					err = -EIO;
+					goto restore;
+				}
+
+				if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+					struct drm_printer p =
+						drm_info_printer(gt->i915->drm.dev);
+
+					pr_err("i915_reset_engine(%s:%s):"
+					       " failed to complete request %llx:%lld after reset\n",
+					       engine->name, test_name,
+					       rq->fence.context,
+					       rq->fence.seqno);
+					intel_engine_dump(engine, &p,
+							  "%s\n", engine->name);
+					i915_request_put(rq);
+
+					GEM_TRACE_DUMP();
+					intel_gt_set_wedged(gt);
+					err = -EIO;
+					goto restore;
+				}
+
+				i915_request_put(rq);
+			}
+
+			if (!(flags & TEST_ACTIVE))
+				intel_engine_pm_put(engine);
+
+			if (!(flags & TEST_SELF) && !wait_for_idle(engine)) {
+				struct drm_printer p =
+					drm_info_printer(gt->i915->drm.dev);
+
+				pr_err("i915_reset_engine(%s:%s):"
+				       " failed to idle after reset\n",
+				       engine->name, test_name);
+				intel_engine_dump(engine, &p,
+						  "%s\n", engine->name);
+
+				err = -EIO;
+				goto restore;
+			}
+
+restore:
+			err2 = intel_selftest_restore_policy(engine, &saved);
+			if (err2)
+				pr_err("[%s] Restore policy failed: %d!\n", engine->name, err2);
+			if (err == 0)
+				err = err2;
+			if (err)
+				break;
+		} while (time_before(jiffies, end_time));
+		clear_and_wake_up_bit(I915_RESET_ENGINE + id, &gt->reset.flags);
+		st_engine_heartbeat_enable_no_pm(engine);
+
+		pr_info("i915_reset_engine(%s:%s): %lu resets\n",
+			engine->name, test_name, count);
+
+		/* GuC based resets are not logged per engine */
+		if (!using_guc) {
+			reported = i915_reset_engine_count(global, engine);
+			reported -= threads[engine->id].resets;
+			if (reported != count) {
+				pr_err("i915_reset_engine(%s:%s): reset %lu times, but reported %lu\n",
+				       engine->name, test_name, count, reported);
+				if (!err)
+					err = -EINVAL;
+			}
+		}
+
+unwind:
+		for_each_engine(other, gt, tmp) {
+			int ret;
+
+			if (!threads[tmp].worker)
+				continue;
+
+			WRITE_ONCE(threads[tmp].stop, true);
+			kthread_flush_work(&threads[tmp].work);
+			ret = READ_ONCE(threads[tmp].result);
+			if (ret) {
+				pr_err("kthread for other engine %s failed, err=%d\n",
+				       other->name, ret);
+				if (!err)
+					err = ret;
+			}
+
+			kthread_destroy_worker(threads[tmp].worker);
+
+			/* GuC based resets are not logged per engine */
+			if (!using_guc) {
+				if (other->uabi_class != engine->uabi_class &&
+				    threads[tmp].resets !=
+				    i915_reset_engine_count(global, other)) {
+					pr_err("Innocent engine %s was reset (count=%ld)\n",
+					       other->name,
+					       i915_reset_engine_count(global, other) -
+					       threads[tmp].resets);
+					if (!err)
+						err = -EINVAL;
+				}
+			}
+		}
+
+		if (device != i915_reset_count(global)) {
+			pr_err("Global reset (count=%ld)!\n",
+			       i915_reset_count(global) - device);
+			if (!err)
+				err = -EINVAL;
+		}
+
+		if (err)
+			break;
+
+		err = igt_flush_test(gt->i915);
+		if (err) {
+			pr_err("[%s] Flush failed: %d!\n", engine->name, err);
+			break;
+		}
+	}
+	kfree(threads);
+
+	if (intel_gt_is_wedged(gt))
+		err = -EIO;
+
+	if (flags & TEST_ACTIVE)
+		hang_fini(&h);
+
+	return err;
+}
+
+static int igt_reset_engines(void *arg)
+{
+	static const struct {
+		const char *name;
+		unsigned int flags;
+	} phases[] = {
+		{ "idle", 0 },
+		{ "active", TEST_ACTIVE },
+		{ "others-idle", TEST_OTHERS },
+		{ "others-active", TEST_OTHERS | TEST_ACTIVE },
+		{
+			"others-priority",
+			TEST_OTHERS | TEST_ACTIVE | TEST_PRIORITY
+		},
+		{
+			"self-priority",
+			TEST_ACTIVE | TEST_PRIORITY | TEST_SELF,
+		},
+		{ }
+	};
+	struct intel_gt *gt = arg;
+	typeof(*phases) *p;
+	int err;
+
+	for (p = phases; p->name; p++) {
+		if (p->flags & TEST_PRIORITY) {
+			if (!(gt->i915->caps.scheduler & I915_SCHEDULER_CAP_PRIORITY))
+				continue;
+		}
+
+		err = __igt_reset_engines(arg, p->name, p->flags);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static u32 fake_hangcheck(struct intel_gt *gt, intel_engine_mask_t mask)
+{
+	u32 count = i915_reset_count(&gt->i915->gpu_error);
+
+	intel_gt_reset(gt, mask, NULL);
+
+	return count;
+}
+
+static int igt_reset_wait(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gpu_error *global = &gt->i915->gpu_error;
+	struct intel_engine_cs *engine;
+	struct i915_request *rq;
+	unsigned int reset_count;
+	struct hang h;
+	long timeout;
+	int err;
+
+	engine = intel_selftest_find_any_engine(gt);
+
+	if (!engine || !intel_engine_can_store_dword(engine))
+		return 0;
+
+	/* Check that we detect a stuck waiter and issue a reset */
+
+	igt_global_reset_lock(gt);
+
+	err = hang_init(&h, gt);
+	if (err) {
+		pr_err("[%s] Hang init failed: %d!\n", engine->name, err);
+		goto unlock;
+	}
+
+	rq = hang_create_request(&h, engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		pr_err("[%s] Create hang request failed: %d!\n", engine->name, err);
+		goto fini;
+	}
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	if (!wait_until_running(&h, rq)) {
+		struct drm_printer p = drm_info_printer(gt->i915->drm.dev);
+
+		pr_err("%s: Failed to start request %llx, at %x\n",
+		       __func__, rq->fence.seqno, hws_seqno(&h, rq));
+		intel_engine_dump(rq->engine, &p, "%s\n", rq->engine->name);
+
+		intel_gt_set_wedged(gt);
+
+		err = -EIO;
+		goto out_rq;
+	}
+
+	reset_count = fake_hangcheck(gt, ALL_ENGINES);
+
+	timeout = i915_request_wait(rq, 0, 10);
+	if (timeout < 0) {
+		pr_err("i915_request_wait failed on a stuck request: err=%ld\n",
+		       timeout);
+		err = timeout;
+		goto out_rq;
+	}
+
+	if (i915_reset_count(global) == reset_count) {
+		pr_err("No GPU reset recorded!\n");
+		err = -EINVAL;
+		goto out_rq;
+	}
+
+out_rq:
+	i915_request_put(rq);
+fini:
+	hang_fini(&h);
+unlock:
+	igt_global_reset_unlock(gt);
+
+	if (intel_gt_is_wedged(gt))
+		return -EIO;
+
+	return err;
+}
+
+struct evict_vma {
+	struct completion completion;
+	struct i915_vma *vma;
+};
+
+static int evict_vma(void *data)
+{
+	struct evict_vma *arg = data;
+	struct i915_address_space *vm = arg->vma->vm;
+	struct drm_mm_node evict = arg->vma->node;
+	int err;
+
+	complete(&arg->completion);
+
+	mutex_lock(&vm->mutex);
+	err = i915_gem_evict_for_node(vm, NULL, &evict, 0);
+	mutex_unlock(&vm->mutex);
+
+	return err;
+}
+
+static int evict_fence(void *data)
+{
+	struct evict_vma *arg = data;
+	int err;
+
+	complete(&arg->completion);
+
+	/* Mark the fence register as dirty to force the mmio update. */
+	err = i915_gem_object_set_tiling(arg->vma->obj, I915_TILING_Y, 512);
+	if (err) {
+		pr_err("Invalid Y-tiling settings; err:%d\n", err);
+		return err;
+	}
+
+	err = i915_vma_pin(arg->vma, 0, 0, PIN_GLOBAL | PIN_MAPPABLE);
+	if (err) {
+		pr_err("Unable to pin vma for Y-tiled fence; err:%d\n", err);
+		return err;
+	}
+
+	err = i915_vma_pin_fence(arg->vma);
+	i915_vma_unpin(arg->vma);
+	if (err) {
+		pr_err("Unable to pin Y-tiled fence; err:%d\n", err);
+		return err;
+	}
+
+	i915_vma_unpin_fence(arg->vma);
+
+	return 0;
+}
+
+static int __igt_reset_evict_vma(struct intel_gt *gt,
+				 struct i915_address_space *vm,
+				 int (*fn)(void *),
+				 unsigned int flags)
+{
+	struct intel_engine_cs *engine;
+	struct drm_i915_gem_object *obj;
+	struct task_struct *tsk = NULL;
+	struct i915_request *rq;
+	struct evict_vma arg;
+	struct hang h;
+	unsigned int pin_flags;
+	int err;
+
+	if (!gt->ggtt->num_fences && flags & EXEC_OBJECT_NEEDS_FENCE)
+		return 0;
+
+	engine = intel_selftest_find_any_engine(gt);
+
+	if (!engine || !intel_engine_can_store_dword(engine))
+		return 0;
+
+	/* Check that we can recover an unbind stuck on a hanging request */
+
+	err = hang_init(&h, gt);
+	if (err) {
+		pr_err("[%s] Hang init failed: %d!\n", engine->name, err);
+		return err;
+	}
+
+	obj = i915_gem_object_create_internal(gt->i915, SZ_1M);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		pr_err("[%s] Create object failed: %d!\n", engine->name, err);
+		goto fini;
+	}
+
+	if (flags & EXEC_OBJECT_NEEDS_FENCE) {
+		err = i915_gem_object_set_tiling(obj, I915_TILING_X, 512);
+		if (err) {
+			pr_err("Invalid X-tiling settings; err:%d\n", err);
+			goto out_obj;
+		}
+	}
+
+	arg.vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(arg.vma)) {
+		err = PTR_ERR(arg.vma);
+		pr_err("[%s] VMA instance failed: %d!\n", engine->name, err);
+		goto out_obj;
+	}
+
+	rq = hang_create_request(&h, engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		pr_err("[%s] Create hang request failed: %d!\n", engine->name, err);
+		goto out_obj;
+	}
+
+	pin_flags = i915_vma_is_ggtt(arg.vma) ? PIN_GLOBAL : PIN_USER;
+
+	if (flags & EXEC_OBJECT_NEEDS_FENCE)
+		pin_flags |= PIN_MAPPABLE;
+
+	err = i915_vma_pin(arg.vma, 0, 0, pin_flags);
+	if (err) {
+		i915_request_add(rq);
+		pr_err("[%s] VMA pin failed: %d!\n", engine->name, err);
+		goto out_obj;
+	}
+
+	if (flags & EXEC_OBJECT_NEEDS_FENCE) {
+		err = i915_vma_pin_fence(arg.vma);
+		if (err) {
+			pr_err("Unable to pin X-tiled fence; err:%d\n", err);
+			i915_vma_unpin(arg.vma);
+			i915_request_add(rq);
+			goto out_obj;
+		}
+	}
+
+	i915_vma_lock(arg.vma);
+	err = i915_request_await_object(rq, arg.vma->obj,
+					flags & EXEC_OBJECT_WRITE);
+	if (err == 0) {
+		err = i915_vma_move_to_active(arg.vma, rq, flags);
+		if (err)
+			pr_err("[%s] Move to active failed: %d!\n", engine->name, err);
+	} else {
+		pr_err("[%s] Request await failed: %d!\n", engine->name, err);
+	}
+
+	i915_vma_unlock(arg.vma);
+
+	if (flags & EXEC_OBJECT_NEEDS_FENCE)
+		i915_vma_unpin_fence(arg.vma);
+	i915_vma_unpin(arg.vma);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (err)
+		goto out_rq;
+
+	if (!wait_until_running(&h, rq)) {
+		struct drm_printer p = drm_info_printer(gt->i915->drm.dev);
+
+		pr_err("%s: Failed to start request %llx, at %x\n",
+		       __func__, rq->fence.seqno, hws_seqno(&h, rq));
+		intel_engine_dump(rq->engine, &p, "%s\n", rq->engine->name);
+
+		intel_gt_set_wedged(gt);
+		goto out_reset;
+	}
+
+	init_completion(&arg.completion);
+
+	tsk = kthread_run(fn, &arg, "igt/evict_vma");
+	if (IS_ERR(tsk)) {
+		err = PTR_ERR(tsk);
+		pr_err("[%s] Thread spawn failed: %d!\n", engine->name, err);
+		tsk = NULL;
+		goto out_reset;
+	}
+	get_task_struct(tsk);
+
+	wait_for_completion(&arg.completion);
+
+	if (wait_for(!list_empty(&rq->fence.cb_list), 10)) {
+		struct drm_printer p = drm_info_printer(gt->i915->drm.dev);
+
+		pr_err("igt/evict_vma kthread did not wait\n");
+		intel_engine_dump(rq->engine, &p, "%s\n", rq->engine->name);
+
+		intel_gt_set_wedged(gt);
+		goto out_reset;
+	}
+
+out_reset:
+	igt_global_reset_lock(gt);
+	fake_hangcheck(gt, rq->engine->mask);
+	igt_global_reset_unlock(gt);
+
+	if (tsk) {
+		struct intel_wedge_me w;
+
+		/* The reset, even indirectly, should take less than 10ms. */
+		intel_wedge_on_timeout(&w, gt, HZ / 10 /* 100ms */)
+			err = kthread_stop(tsk);
+
+		put_task_struct(tsk);
+	}
+
+out_rq:
+	i915_request_put(rq);
+out_obj:
+	i915_gem_object_put(obj);
+fini:
+	hang_fini(&h);
+	if (intel_gt_is_wedged(gt))
+		return -EIO;
+
+	return err;
+}
+
+static int igt_reset_evict_ggtt(void *arg)
+{
+	struct intel_gt *gt = arg;
+
+	return __igt_reset_evict_vma(gt, &gt->ggtt->vm,
+				     evict_vma, EXEC_OBJECT_WRITE);
+}
+
+static int igt_reset_evict_ppgtt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_ppgtt *ppgtt;
+	int err;
+
+	/* aliasing == global gtt locking, covered above */
+	if (INTEL_PPGTT(gt->i915) < INTEL_PPGTT_FULL)
+		return 0;
+
+	ppgtt = i915_ppgtt_create(gt, 0);
+	if (IS_ERR(ppgtt))
+		return PTR_ERR(ppgtt);
+
+	err = __igt_reset_evict_vma(gt, &ppgtt->vm,
+				    evict_vma, EXEC_OBJECT_WRITE);
+	i915_vm_put(&ppgtt->vm);
+
+	return err;
+}
+
+static int igt_reset_evict_fence(void *arg)
+{
+	struct intel_gt *gt = arg;
+
+	return __igt_reset_evict_vma(gt, &gt->ggtt->vm,
+				     evict_fence, EXEC_OBJECT_NEEDS_FENCE);
+}
+
+static int wait_for_others(struct intel_gt *gt,
+			   struct intel_engine_cs *exclude)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id) {
+		if (engine == exclude)
+			continue;
+
+		if (!wait_for_idle(engine))
+			return -EIO;
+	}
+
+	return 0;
+}
+
+static int igt_reset_queue(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gpu_error *global = &gt->i915->gpu_error;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct hang h;
+	int err;
+
+	/* Check that we replay pending requests following a hang */
+
+	igt_global_reset_lock(gt);
+
+	err = hang_init(&h, gt);
+	if (err)
+		goto unlock;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_selftest_saved_policy saved;
+		struct i915_request *prev;
+		IGT_TIMEOUT(end_time);
+		unsigned int count;
+		bool using_guc = intel_engine_uses_guc(engine);
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		if (using_guc) {
+			err = intel_selftest_modify_policy(engine, &saved,
+							   SELFTEST_SCHEDULER_MODIFY_NO_HANGCHECK);
+			if (err) {
+				pr_err("[%s] Modify policy failed: %d!\n", engine->name, err);
+				goto fini;
+			}
+		}
+
+		prev = hang_create_request(&h, engine);
+		if (IS_ERR(prev)) {
+			err = PTR_ERR(prev);
+			pr_err("[%s] Create 'prev' hang request failed: %d!\n", engine->name, err);
+			goto restore;
+		}
+
+		i915_request_get(prev);
+		i915_request_add(prev);
+
+		count = 0;
+		do {
+			struct i915_request *rq;
+			unsigned int reset_count;
+
+			rq = hang_create_request(&h, engine);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				pr_err("[%s] Create hang request failed: %d!\n", engine->name, err);
+				goto restore;
+			}
+
+			i915_request_get(rq);
+			i915_request_add(rq);
+
+			/*
+			 * XXX We don't handle resetting the kernel context
+			 * very well. If we trigger a device reset twice in
+			 * quick succession while the kernel context is
+			 * executing, we may end up skipping the breadcrumb.
+			 * This is really only a problem for the selftest as
+			 * normally there is a large interlude between resets
+			 * (hangcheck), or we focus on resetting just one
+			 * engine and so avoid repeatedly resetting innocents.
+			 */
+			err = wait_for_others(gt, engine);
+			if (err) {
+				pr_err("%s(%s): Failed to idle other inactive engines after device reset\n",
+				       __func__, engine->name);
+				i915_request_put(rq);
+				i915_request_put(prev);
+
+				GEM_TRACE_DUMP();
+				intel_gt_set_wedged(gt);
+				goto restore;
+			}
+
+			if (!wait_until_running(&h, prev)) {
+				struct drm_printer p = drm_info_printer(gt->i915->drm.dev);
+
+				pr_err("%s(%s): Failed to start request %llx, at %x\n",
+				       __func__, engine->name,
+				       prev->fence.seqno, hws_seqno(&h, prev));
+				intel_engine_dump(engine, &p,
+						  "%s\n", engine->name);
+
+				i915_request_put(rq);
+				i915_request_put(prev);
+
+				intel_gt_set_wedged(gt);
+
+				err = -EIO;
+				goto restore;
+			}
+
+			reset_count = fake_hangcheck(gt, BIT(id));
+
+			if (prev->fence.error != -EIO) {
+				pr_err("GPU reset not recorded on hanging request [fence.error=%d]!\n",
+				       prev->fence.error);
+				i915_request_put(rq);
+				i915_request_put(prev);
+				err = -EINVAL;
+				goto restore;
+			}
+
+			if (rq->fence.error) {
+				pr_err("Fence error status not zero [%d] after unrelated reset\n",
+				       rq->fence.error);
+				i915_request_put(rq);
+				i915_request_put(prev);
+				err = -EINVAL;
+				goto restore;
+			}
+
+			if (i915_reset_count(global) == reset_count) {
+				pr_err("No GPU reset recorded!\n");
+				i915_request_put(rq);
+				i915_request_put(prev);
+				err = -EINVAL;
+				goto restore;
+			}
+
+			i915_request_put(prev);
+			prev = rq;
+			count++;
+		} while (time_before(jiffies, end_time));
+		pr_info("%s: Completed %d queued resets\n",
+			engine->name, count);
+
+		*h.batch = MI_BATCH_BUFFER_END;
+		intel_gt_chipset_flush(engine->gt);
+
+		i915_request_put(prev);
+
+restore:
+		if (using_guc) {
+			int err2 = intel_selftest_restore_policy(engine, &saved);
+
+			if (err2)
+				pr_err("%s:%d> [%s] Restore policy failed: %d!\n",
+				       __func__, __LINE__, engine->name, err2);
+			if (err == 0)
+				err = err2;
+		}
+		if (err)
+			goto fini;
+
+		err = igt_flush_test(gt->i915);
+		if (err) {
+			pr_err("[%s] Flush failed: %d!\n", engine->name, err);
+			break;
+		}
+	}
+
+fini:
+	hang_fini(&h);
+unlock:
+	igt_global_reset_unlock(gt);
+
+	if (intel_gt_is_wedged(gt))
+		return -EIO;
+
+	return err;
+}
+
+static int igt_handle_error(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_gpu_error *global = &gt->i915->gpu_error;
+	struct intel_engine_cs *engine;
+	struct hang h;
+	struct i915_request *rq;
+	struct i915_gpu_coredump *error;
+	int err;
+
+	engine = intel_selftest_find_any_engine(gt);
+
+	/* Check that we can issue a global GPU and engine reset */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	if (!engine || !intel_engine_can_store_dword(engine))
+		return 0;
+
+	err = hang_init(&h, gt);
+	if (err) {
+		pr_err("[%s] Hang init failed: %d!\n", engine->name, err);
+		return err;
+	}
+
+	rq = hang_create_request(&h, engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		pr_err("[%s] Create hang request failed: %d!\n", engine->name, err);
+		goto err_fini;
+	}
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	if (!wait_until_running(&h, rq)) {
+		struct drm_printer p = drm_info_printer(gt->i915->drm.dev);
+
+		pr_err("%s: Failed to start request %llx, at %x\n",
+		       __func__, rq->fence.seqno, hws_seqno(&h, rq));
+		intel_engine_dump(rq->engine, &p, "%s\n", rq->engine->name);
+
+		intel_gt_set_wedged(gt);
+
+		err = -EIO;
+		goto err_request;
+	}
+
+	/* Temporarily disable error capture */
+	error = xchg(&global->first_error, (void *)-1);
+
+	intel_gt_handle_error(gt, engine->mask, 0, NULL);
+
+	xchg(&global->first_error, error);
+
+	if (rq->fence.error != -EIO) {
+		pr_err("Guilty request not identified!\n");
+		err = -EINVAL;
+		goto err_request;
+	}
+
+err_request:
+	i915_request_put(rq);
+err_fini:
+	hang_fini(&h);
+	return err;
+}
+
+static int __igt_atomic_reset_engine(struct intel_engine_cs *engine,
+				     const struct igt_atomic_section *p,
+				     const char *mode)
+{
+	struct tasklet_struct * const t = &engine->sched_engine->tasklet;
+	int err;
+
+	GEM_TRACE("i915_reset_engine(%s:%s) under %s\n",
+		  engine->name, mode, p->name);
+
+	if (t->func)
+		tasklet_disable(t);
+	if (strcmp(p->name, "softirq"))
+		local_bh_disable();
+	p->critical_section_begin();
+
+	err = __intel_engine_reset_bh(engine, NULL);
+
+	p->critical_section_end();
+	if (strcmp(p->name, "softirq"))
+		local_bh_enable();
+	if (t->func) {
+		tasklet_enable(t);
+		tasklet_hi_schedule(t);
+	}
+
+	if (err)
+		pr_err("i915_reset_engine(%s:%s) failed under %s\n",
+		       engine->name, mode, p->name);
+
+	return err;
+}
+
+static int igt_atomic_reset_engine(struct intel_engine_cs *engine,
+				   const struct igt_atomic_section *p)
+{
+	struct i915_request *rq;
+	struct hang h;
+	int err;
+
+	err = __igt_atomic_reset_engine(engine, p, "idle");
+	if (err)
+		return err;
+
+	err = hang_init(&h, engine->gt);
+	if (err) {
+		pr_err("[%s] Hang init failed: %d!\n", engine->name, err);
+		return err;
+	}
+
+	rq = hang_create_request(&h, engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		pr_err("[%s] Create hang request failed: %d!\n", engine->name, err);
+		goto out;
+	}
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	if (wait_until_running(&h, rq)) {
+		err = __igt_atomic_reset_engine(engine, p, "active");
+	} else {
+		pr_err("%s(%s): Failed to start request %llx, at %x\n",
+		       __func__, engine->name,
+		       rq->fence.seqno, hws_seqno(&h, rq));
+		intel_gt_set_wedged(engine->gt);
+		err = -EIO;
+	}
+
+	if (err == 0) {
+		struct intel_wedge_me w;
+
+		intel_wedge_on_timeout(&w, engine->gt, HZ / 20 /* 50ms */)
+			i915_request_wait(rq, 0, MAX_SCHEDULE_TIMEOUT);
+		if (intel_gt_is_wedged(engine->gt))
+			err = -EIO;
+	}
+
+	i915_request_put(rq);
+out:
+	hang_fini(&h);
+	return err;
+}
+
+static int igt_reset_engines_atomic(void *arg)
+{
+	struct intel_gt *gt = arg;
+	const typeof(*igt_atomic_phases) *p;
+	int err = 0;
+
+	/* Check that the engines resets are usable from atomic context */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	igt_global_reset_lock(gt);
+
+	/* Flush any requests before we get started and check basics */
+	if (!igt_force_reset(gt))
+		goto unlock;
+
+	for (p = igt_atomic_phases; p->name; p++) {
+		struct intel_engine_cs *engine;
+		enum intel_engine_id id;
+
+		for_each_engine(engine, gt, id) {
+			err = igt_atomic_reset_engine(engine, p);
+			if (err)
+				goto out;
+		}
+	}
+
+out:
+	/* As we poke around the guts, do a full reset before continuing. */
+	igt_force_reset(gt);
+unlock:
+	igt_global_reset_unlock(gt);
+
+	return err;
+}
+
+int intel_hangcheck_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_hang_sanitycheck),
+		SUBTEST(igt_reset_nop),
+		SUBTEST(igt_reset_nop_engine),
+		SUBTEST(igt_reset_idle_engine),
+		SUBTEST(igt_reset_active_engine),
+		SUBTEST(igt_reset_fail_engine),
+		SUBTEST(igt_reset_engines),
+		SUBTEST(igt_reset_engines_atomic),
+		SUBTEST(igt_reset_queue),
+		SUBTEST(igt_reset_wait),
+		SUBTEST(igt_reset_evict_ggtt),
+		SUBTEST(igt_reset_evict_ppgtt),
+		SUBTEST(igt_reset_evict_fence),
+		SUBTEST(igt_handle_error),
+	};
+	struct intel_gt *gt = to_gt(i915);
+	intel_wakeref_t wakeref;
+	int err;
+
+	if (!intel_has_gpu_reset(gt))
+		return 0;
+
+	if (intel_gt_is_wedged(gt))
+		return -EIO; /* we're long past hope of a successful reset */
+
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+
+	err = intel_gt_live_subtests(tests, gt);
+
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_llc.c b/drivers/gpu/drm/i915/gt/selftest_llc.c
new file mode 100644
index 000000000..cfd736d88
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_llc.c
@@ -0,0 +1,71 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "intel_pm.h" /* intel_gpu_freq() */
+#include "selftest_llc.h"
+#include "intel_rps.h"
+
+static int gen6_verify_ring_freq(struct intel_llc *llc)
+{
+	struct drm_i915_private *i915 = llc_to_gt(llc)->i915;
+	struct ia_constants consts;
+	intel_wakeref_t wakeref;
+	unsigned int gpu_freq;
+	int err = 0;
+
+	wakeref = intel_runtime_pm_get(llc_to_gt(llc)->uncore->rpm);
+
+	if (!get_ia_constants(llc, &consts))
+		goto out_rpm;
+
+	for (gpu_freq = consts.min_gpu_freq;
+	     gpu_freq <= consts.max_gpu_freq;
+	     gpu_freq++) {
+		struct intel_rps *rps = &llc_to_gt(llc)->rps;
+
+		unsigned int ia_freq, ring_freq, found;
+		u32 val;
+
+		calc_ia_freq(llc, gpu_freq, &consts, &ia_freq, &ring_freq);
+
+		val = gpu_freq;
+		if (snb_pcode_read(llc_to_gt(llc)->uncore, GEN6_PCODE_READ_MIN_FREQ_TABLE,
+				   &val, NULL)) {
+			pr_err("Failed to read freq table[%d], range [%d, %d]\n",
+			       gpu_freq, consts.min_gpu_freq, consts.max_gpu_freq);
+			err = -ENXIO;
+			break;
+		}
+
+		found = (val >> 0) & 0xff;
+		if (found != ia_freq) {
+			pr_err("Min freq table(%d/[%d, %d]):%dMHz did not match expected CPU freq, found %d, expected %d\n",
+			       gpu_freq, consts.min_gpu_freq, consts.max_gpu_freq,
+			       intel_gpu_freq(rps, gpu_freq * (GRAPHICS_VER(i915) >= 9 ? GEN9_FREQ_SCALER : 1)),
+			       found, ia_freq);
+			err = -EINVAL;
+			break;
+		}
+
+		found = (val >> 8) & 0xff;
+		if (found != ring_freq) {
+			pr_err("Min freq table(%d/[%d, %d]):%dMHz did not match expected ring freq, found %d, expected %d\n",
+			       gpu_freq, consts.min_gpu_freq, consts.max_gpu_freq,
+			       intel_gpu_freq(rps, gpu_freq * (GRAPHICS_VER(i915) >= 9 ? GEN9_FREQ_SCALER : 1)),
+			       found, ring_freq);
+			err = -EINVAL;
+			break;
+		}
+	}
+
+out_rpm:
+	intel_runtime_pm_put(llc_to_gt(llc)->uncore->rpm, wakeref);
+	return err;
+}
+
+int st_llc_verify(struct intel_llc *llc)
+{
+	return gen6_verify_ring_freq(llc);
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_llc.h b/drivers/gpu/drm/i915/gt/selftest_llc.h
new file mode 100644
index 000000000..88ee94800
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_llc.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef SELFTEST_LLC_H
+#define SELFTEST_LLC_H
+
+struct intel_llc;
+
+int st_llc_verify(struct intel_llc *llc);
+
+#endif /* SELFTEST_LLC_H */
diff --git a/drivers/gpu/drm/i915/gt/selftest_lrc.c b/drivers/gpu/drm/i915/gt/selftest_lrc.c
new file mode 100644
index 000000000..82d3f8058
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_lrc.c
@@ -0,0 +1,1977 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/prime_numbers.h>
+
+#include "gem/i915_gem_internal.h"
+
+#include "i915_selftest.h"
+#include "intel_engine_heartbeat.h"
+#include "intel_engine_pm.h"
+#include "intel_reset.h"
+#include "intel_ring.h"
+#include "selftest_engine_heartbeat.h"
+#include "selftests/i915_random.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/igt_live_test.h"
+#include "selftests/igt_spinner.h"
+#include "selftests/lib_sw_fence.h"
+#include "shmem_utils.h"
+
+#include "gem/selftests/igt_gem_utils.h"
+#include "gem/selftests/mock_context.h"
+
+#define CS_GPR(engine, n) ((engine)->mmio_base + 0x600 + (n) * 4)
+#define NUM_GPR 16
+#define NUM_GPR_DW (NUM_GPR * 2) /* each GPR is 2 dwords */
+
+#define LRI_HEADER MI_INSTR(0x22, 0)
+#define LRI_LENGTH_MASK GENMASK(7, 0)
+
+static struct i915_vma *create_scratch(struct intel_gt *gt)
+{
+	return __vm_create_scratch_for_read_pinned(&gt->ggtt->vm, PAGE_SIZE);
+}
+
+static bool is_active(struct i915_request *rq)
+{
+	if (i915_request_is_active(rq))
+		return true;
+
+	if (i915_request_on_hold(rq))
+		return true;
+
+	if (i915_request_has_initial_breadcrumb(rq) && i915_request_started(rq))
+		return true;
+
+	return false;
+}
+
+static int wait_for_submit(struct intel_engine_cs *engine,
+			   struct i915_request *rq,
+			   unsigned long timeout)
+{
+	/* Ignore our own attempts to suppress excess tasklets */
+	tasklet_hi_schedule(&engine->sched_engine->tasklet);
+
+	timeout += jiffies;
+	do {
+		bool done = time_after(jiffies, timeout);
+
+		if (i915_request_completed(rq)) /* that was quick! */
+			return 0;
+
+		/* Wait until the HW has acknowleged the submission (or err) */
+		intel_engine_flush_submission(engine);
+		if (!READ_ONCE(engine->execlists.pending[0]) && is_active(rq))
+			return 0;
+
+		if (done)
+			return -ETIME;
+
+		cond_resched();
+	} while (1);
+}
+
+static int emit_semaphore_signal(struct intel_context *ce, void *slot)
+{
+	const u32 offset =
+		i915_ggtt_offset(ce->engine->status_page.vma) +
+		offset_in_page(slot);
+	struct i915_request *rq;
+	u32 *cs;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = offset;
+	*cs++ = 0;
+	*cs++ = 1;
+
+	intel_ring_advance(rq, cs);
+
+	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+	i915_request_add(rq);
+	return 0;
+}
+
+static int context_flush(struct intel_context *ce, long timeout)
+{
+	struct i915_request *rq;
+	struct dma_fence *fence;
+	int err = 0;
+
+	rq = intel_engine_create_kernel_request(ce->engine);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	fence = i915_active_fence_get(&ce->timeline->last_request);
+	if (fence) {
+		i915_request_await_dma_fence(rq, fence);
+		dma_fence_put(fence);
+	}
+
+	rq = i915_request_get(rq);
+	i915_request_add(rq);
+	if (i915_request_wait(rq, 0, timeout) < 0)
+		err = -ETIME;
+	i915_request_put(rq);
+
+	rmb(); /* We know the request is written, make sure all state is too! */
+	return err;
+}
+
+static int get_lri_mask(struct intel_engine_cs *engine, u32 lri)
+{
+	if ((lri & MI_LRI_LRM_CS_MMIO) == 0)
+		return ~0u;
+
+	if (GRAPHICS_VER(engine->i915) < 12)
+		return 0xfff;
+
+	switch (engine->class) {
+	default:
+	case RENDER_CLASS:
+	case COMPUTE_CLASS:
+		return 0x07ff;
+	case COPY_ENGINE_CLASS:
+		return 0x0fff;
+	case VIDEO_DECODE_CLASS:
+	case VIDEO_ENHANCEMENT_CLASS:
+		return 0x3fff;
+	}
+}
+
+static int live_lrc_layout(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	u32 *lrc;
+	int err;
+
+	/*
+	 * Check the registers offsets we use to create the initial reg state
+	 * match the layout saved by HW.
+	 */
+
+	lrc = (u32 *)__get_free_page(GFP_KERNEL); /* requires page alignment */
+	if (!lrc)
+		return -ENOMEM;
+	GEM_BUG_ON(offset_in_page(lrc));
+
+	err = 0;
+	for_each_engine(engine, gt, id) {
+		u32 *hw;
+		int dw;
+
+		if (!engine->default_state)
+			continue;
+
+		hw = shmem_pin_map(engine->default_state);
+		if (!hw) {
+			err = -ENOMEM;
+			break;
+		}
+		hw += LRC_STATE_OFFSET / sizeof(*hw);
+
+		__lrc_init_regs(memset(lrc, POISON_INUSE, PAGE_SIZE),
+				engine->kernel_context, engine, true);
+
+		dw = 0;
+		do {
+			u32 lri = READ_ONCE(hw[dw]);
+			u32 lri_mask;
+
+			if (lri == 0) {
+				dw++;
+				continue;
+			}
+
+			if (lrc[dw] == 0) {
+				pr_debug("%s: skipped instruction %x at dword %d\n",
+					 engine->name, lri, dw);
+				dw++;
+				continue;
+			}
+
+			if ((lri & GENMASK(31, 23)) != LRI_HEADER) {
+				pr_err("%s: Expected LRI command at dword %d, found %08x\n",
+				       engine->name, dw, lri);
+				err = -EINVAL;
+				break;
+			}
+
+			if (lrc[dw] != lri) {
+				pr_err("%s: LRI command mismatch at dword %d, expected %08x found %08x\n",
+				       engine->name, dw, lri, lrc[dw]);
+				err = -EINVAL;
+				break;
+			}
+
+			/*
+			 * When bit 19 of MI_LOAD_REGISTER_IMM instruction
+			 * opcode is set on Gen12+ devices, HW does not
+			 * care about certain register address offsets, and
+			 * instead check the following for valid address
+			 * ranges on specific engines:
+			 * RCS && CCS: BITS(0 - 10)
+			 * BCS: BITS(0 - 11)
+			 * VECS && VCS: BITS(0 - 13)
+			 */
+			lri_mask = get_lri_mask(engine, lri);
+
+			lri &= 0x7f;
+			lri++;
+			dw++;
+
+			while (lri) {
+				u32 offset = READ_ONCE(hw[dw]);
+
+				if ((offset ^ lrc[dw]) & lri_mask) {
+					pr_err("%s: Different registers found at dword %d, expected %x, found %x\n",
+					       engine->name, dw, offset, lrc[dw]);
+					err = -EINVAL;
+					break;
+				}
+
+				/*
+				 * Skip over the actual register value as we
+				 * expect that to differ.
+				 */
+				dw += 2;
+				lri -= 2;
+			}
+		} while (!err && (lrc[dw] & ~BIT(0)) != MI_BATCH_BUFFER_END);
+
+		if (err) {
+			pr_info("%s: HW register image:\n", engine->name);
+			igt_hexdump(hw, PAGE_SIZE);
+
+			pr_info("%s: SW register image:\n", engine->name);
+			igt_hexdump(lrc, PAGE_SIZE);
+		}
+
+		shmem_unpin_map(engine->default_state, hw);
+		if (err)
+			break;
+	}
+
+	free_page((unsigned long)lrc);
+	return err;
+}
+
+static int find_offset(const u32 *lri, u32 offset)
+{
+	int i;
+
+	for (i = 0; i < PAGE_SIZE / sizeof(u32); i++)
+		if (lri[i] == offset)
+			return i;
+
+	return -1;
+}
+
+static int live_lrc_fixed(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Check the assumed register offsets match the actual locations in
+	 * the context image.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		const struct {
+			u32 reg;
+			u32 offset;
+			const char *name;
+		} tbl[] = {
+			{
+				i915_mmio_reg_offset(RING_START(engine->mmio_base)),
+				CTX_RING_START - 1,
+				"RING_START"
+			},
+			{
+				i915_mmio_reg_offset(RING_CTL(engine->mmio_base)),
+				CTX_RING_CTL - 1,
+				"RING_CTL"
+			},
+			{
+				i915_mmio_reg_offset(RING_HEAD(engine->mmio_base)),
+				CTX_RING_HEAD - 1,
+				"RING_HEAD"
+			},
+			{
+				i915_mmio_reg_offset(RING_TAIL(engine->mmio_base)),
+				CTX_RING_TAIL - 1,
+				"RING_TAIL"
+			},
+			{
+				i915_mmio_reg_offset(RING_MI_MODE(engine->mmio_base)),
+				lrc_ring_mi_mode(engine),
+				"RING_MI_MODE"
+			},
+			{
+				i915_mmio_reg_offset(RING_BBSTATE(engine->mmio_base)),
+				CTX_BB_STATE - 1,
+				"BB_STATE"
+			},
+			{
+				i915_mmio_reg_offset(RING_BB_PER_CTX_PTR(engine->mmio_base)),
+				lrc_ring_wa_bb_per_ctx(engine),
+				"RING_BB_PER_CTX_PTR"
+			},
+			{
+				i915_mmio_reg_offset(RING_INDIRECT_CTX(engine->mmio_base)),
+				lrc_ring_indirect_ptr(engine),
+				"RING_INDIRECT_CTX_PTR"
+			},
+			{
+				i915_mmio_reg_offset(RING_INDIRECT_CTX_OFFSET(engine->mmio_base)),
+				lrc_ring_indirect_offset(engine),
+				"RING_INDIRECT_CTX_OFFSET"
+			},
+			{
+				i915_mmio_reg_offset(RING_CTX_TIMESTAMP(engine->mmio_base)),
+				CTX_TIMESTAMP - 1,
+				"RING_CTX_TIMESTAMP"
+			},
+			{
+				i915_mmio_reg_offset(GEN8_RING_CS_GPR(engine->mmio_base, 0)),
+				lrc_ring_gpr0(engine),
+				"RING_CS_GPR0"
+			},
+			{
+				i915_mmio_reg_offset(RING_CMD_BUF_CCTL(engine->mmio_base)),
+				lrc_ring_cmd_buf_cctl(engine),
+				"RING_CMD_BUF_CCTL"
+			},
+			{
+				i915_mmio_reg_offset(RING_BB_OFFSET(engine->mmio_base)),
+				lrc_ring_bb_offset(engine),
+				"RING_BB_OFFSET"
+			},
+			{ },
+		}, *t;
+		u32 *hw;
+
+		if (!engine->default_state)
+			continue;
+
+		hw = shmem_pin_map(engine->default_state);
+		if (!hw) {
+			err = -ENOMEM;
+			break;
+		}
+		hw += LRC_STATE_OFFSET / sizeof(*hw);
+
+		for (t = tbl; t->name; t++) {
+			int dw = find_offset(hw, t->reg);
+
+			if (dw != t->offset) {
+				pr_err("%s: Offset for %s [0x%x] mismatch, found %x, expected %x\n",
+				       engine->name,
+				       t->name,
+				       t->reg,
+				       dw,
+				       t->offset);
+				err = -EINVAL;
+			}
+		}
+
+		shmem_unpin_map(engine->default_state, hw);
+	}
+
+	return err;
+}
+
+static int __live_lrc_state(struct intel_engine_cs *engine,
+			    struct i915_vma *scratch)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+	struct i915_gem_ww_ctx ww;
+	enum {
+		RING_START_IDX = 0,
+		RING_TAIL_IDX,
+		MAX_IDX
+	};
+	u32 expected[MAX_IDX];
+	u32 *cs;
+	int err;
+	int n;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	i915_gem_ww_ctx_init(&ww, false);
+retry:
+	err = i915_gem_object_lock(scratch->obj, &ww);
+	if (!err)
+		err = intel_context_pin_ww(ce, &ww);
+	if (err)
+		goto err_put;
+
+	rq = i915_request_create(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_unpin;
+	}
+
+	cs = intel_ring_begin(rq, 4 * MAX_IDX);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		i915_request_add(rq);
+		goto err_unpin;
+	}
+
+	*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+	*cs++ = i915_mmio_reg_offset(RING_START(engine->mmio_base));
+	*cs++ = i915_ggtt_offset(scratch) + RING_START_IDX * sizeof(u32);
+	*cs++ = 0;
+
+	expected[RING_START_IDX] = i915_ggtt_offset(ce->ring->vma);
+
+	*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+	*cs++ = i915_mmio_reg_offset(RING_TAIL(engine->mmio_base));
+	*cs++ = i915_ggtt_offset(scratch) + RING_TAIL_IDX * sizeof(u32);
+	*cs++ = 0;
+
+	err = i915_request_await_object(rq, scratch->obj, true);
+	if (!err)
+		err = i915_vma_move_to_active(scratch, rq, EXEC_OBJECT_WRITE);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (err)
+		goto err_rq;
+
+	intel_engine_flush_submission(engine);
+	expected[RING_TAIL_IDX] = ce->ring->tail;
+
+	if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+		err = -ETIME;
+		goto err_rq;
+	}
+
+	cs = i915_gem_object_pin_map(scratch->obj, I915_MAP_WB);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_rq;
+	}
+
+	for (n = 0; n < MAX_IDX; n++) {
+		if (cs[n] != expected[n]) {
+			pr_err("%s: Stored register[%d] value[0x%x] did not match expected[0x%x]\n",
+			       engine->name, n, cs[n], expected[n]);
+			err = -EINVAL;
+			break;
+		}
+	}
+
+	i915_gem_object_unpin_map(scratch->obj);
+
+err_rq:
+	i915_request_put(rq);
+err_unpin:
+	intel_context_unpin(ce);
+err_put:
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	intel_context_put(ce);
+	return err;
+}
+
+static int live_lrc_state(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct i915_vma *scratch;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Check the live register state matches what we expect for this
+	 * intel_context.
+	 */
+
+	scratch = create_scratch(gt);
+	if (IS_ERR(scratch))
+		return PTR_ERR(scratch);
+
+	for_each_engine(engine, gt, id) {
+		err = __live_lrc_state(engine, scratch);
+		if (err)
+			break;
+	}
+
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	i915_vma_unpin_and_release(&scratch, 0);
+	return err;
+}
+
+static int gpr_make_dirty(struct intel_context *ce)
+{
+	struct i915_request *rq;
+	u32 *cs;
+	int n;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 2 * NUM_GPR_DW + 2);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	*cs++ = MI_LOAD_REGISTER_IMM(NUM_GPR_DW);
+	for (n = 0; n < NUM_GPR_DW; n++) {
+		*cs++ = CS_GPR(ce->engine, n);
+		*cs++ = STACK_MAGIC;
+	}
+	*cs++ = MI_NOOP;
+
+	intel_ring_advance(rq, cs);
+
+	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+	i915_request_add(rq);
+
+	return 0;
+}
+
+static struct i915_request *
+__gpr_read(struct intel_context *ce, struct i915_vma *scratch, u32 *slot)
+{
+	const u32 offset =
+		i915_ggtt_offset(ce->engine->status_page.vma) +
+		offset_in_page(slot);
+	struct i915_request *rq;
+	u32 *cs;
+	int err;
+	int n;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return rq;
+
+	cs = intel_ring_begin(rq, 6 + 4 * NUM_GPR_DW);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return ERR_CAST(cs);
+	}
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	*cs++ = MI_NOOP;
+
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_NEQ_SDD;
+	*cs++ = 0;
+	*cs++ = offset;
+	*cs++ = 0;
+
+	for (n = 0; n < NUM_GPR_DW; n++) {
+		*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+		*cs++ = CS_GPR(ce->engine, n);
+		*cs++ = i915_ggtt_offset(scratch) + n * sizeof(u32);
+		*cs++ = 0;
+	}
+
+	i915_vma_lock(scratch);
+	err = i915_request_await_object(rq, scratch->obj, true);
+	if (!err)
+		err = i915_vma_move_to_active(scratch, rq, EXEC_OBJECT_WRITE);
+	i915_vma_unlock(scratch);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (err) {
+		i915_request_put(rq);
+		rq = ERR_PTR(err);
+	}
+
+	return rq;
+}
+
+static int __live_lrc_gpr(struct intel_engine_cs *engine,
+			  struct i915_vma *scratch,
+			  bool preempt)
+{
+	u32 *slot = memset32(engine->status_page.addr + 1000, 0, 4);
+	struct intel_context *ce;
+	struct i915_request *rq;
+	u32 *cs;
+	int err;
+	int n;
+
+	if (GRAPHICS_VER(engine->i915) < 9 && engine->class != RENDER_CLASS)
+		return 0; /* GPR only on rcs0 for gen8 */
+
+	err = gpr_make_dirty(engine->kernel_context);
+	if (err)
+		return err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	rq = __gpr_read(ce, scratch, slot);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_put;
+	}
+
+	err = wait_for_submit(engine, rq, HZ / 2);
+	if (err)
+		goto err_rq;
+
+	if (preempt) {
+		err = gpr_make_dirty(engine->kernel_context);
+		if (err)
+			goto err_rq;
+
+		err = emit_semaphore_signal(engine->kernel_context, slot);
+		if (err)
+			goto err_rq;
+
+		err = wait_for_submit(engine, rq, HZ / 2);
+		if (err)
+			goto err_rq;
+	} else {
+		slot[0] = 1;
+		wmb();
+	}
+
+	if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+		err = -ETIME;
+		goto err_rq;
+	}
+
+	cs = i915_gem_object_pin_map_unlocked(scratch->obj, I915_MAP_WB);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_rq;
+	}
+
+	for (n = 0; n < NUM_GPR_DW; n++) {
+		if (cs[n]) {
+			pr_err("%s: GPR[%d].%s was not zero, found 0x%08x!\n",
+			       engine->name,
+			       n / 2, n & 1 ? "udw" : "ldw",
+			       cs[n]);
+			err = -EINVAL;
+			break;
+		}
+	}
+
+	i915_gem_object_unpin_map(scratch->obj);
+
+err_rq:
+	memset32(&slot[0], -1, 4);
+	wmb();
+	i915_request_put(rq);
+err_put:
+	intel_context_put(ce);
+	return err;
+}
+
+static int live_lrc_gpr(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct i915_vma *scratch;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Check that GPR registers are cleared in new contexts as we need
+	 * to avoid leaking any information from previous contexts.
+	 */
+
+	scratch = create_scratch(gt);
+	if (IS_ERR(scratch))
+		return PTR_ERR(scratch);
+
+	for_each_engine(engine, gt, id) {
+		st_engine_heartbeat_disable(engine);
+
+		err = __live_lrc_gpr(engine, scratch, false);
+		if (err)
+			goto err;
+
+		err = __live_lrc_gpr(engine, scratch, true);
+		if (err)
+			goto err;
+
+err:
+		st_engine_heartbeat_enable(engine);
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	i915_vma_unpin_and_release(&scratch, 0);
+	return err;
+}
+
+static struct i915_request *
+create_timestamp(struct intel_context *ce, void *slot, int idx)
+{
+	const u32 offset =
+		i915_ggtt_offset(ce->engine->status_page.vma) +
+		offset_in_page(slot);
+	struct i915_request *rq;
+	u32 *cs;
+	int err;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return rq;
+
+	cs = intel_ring_begin(rq, 10);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err;
+	}
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	*cs++ = MI_NOOP;
+
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_NEQ_SDD;
+	*cs++ = 0;
+	*cs++ = offset;
+	*cs++ = 0;
+
+	*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+	*cs++ = i915_mmio_reg_offset(RING_CTX_TIMESTAMP(rq->engine->mmio_base));
+	*cs++ = offset + idx * sizeof(u32);
+	*cs++ = 0;
+
+	intel_ring_advance(rq, cs);
+
+	err = 0;
+err:
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (err) {
+		i915_request_put(rq);
+		return ERR_PTR(err);
+	}
+
+	return rq;
+}
+
+struct lrc_timestamp {
+	struct intel_engine_cs *engine;
+	struct intel_context *ce[2];
+	u32 poison;
+};
+
+static bool timestamp_advanced(u32 start, u32 end)
+{
+	return (s32)(end - start) > 0;
+}
+
+static int __lrc_timestamp(const struct lrc_timestamp *arg, bool preempt)
+{
+	u32 *slot = memset32(arg->engine->status_page.addr + 1000, 0, 4);
+	struct i915_request *rq;
+	u32 timestamp;
+	int err = 0;
+
+	arg->ce[0]->lrc_reg_state[CTX_TIMESTAMP] = arg->poison;
+	rq = create_timestamp(arg->ce[0], slot, 1);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	err = wait_for_submit(rq->engine, rq, HZ / 2);
+	if (err)
+		goto err;
+
+	if (preempt) {
+		arg->ce[1]->lrc_reg_state[CTX_TIMESTAMP] = 0xdeadbeef;
+		err = emit_semaphore_signal(arg->ce[1], slot);
+		if (err)
+			goto err;
+	} else {
+		slot[0] = 1;
+		wmb();
+	}
+
+	/* And wait for switch to kernel (to save our context to memory) */
+	err = context_flush(arg->ce[0], HZ / 2);
+	if (err)
+		goto err;
+
+	if (!timestamp_advanced(arg->poison, slot[1])) {
+		pr_err("%s(%s): invalid timestamp on restore, context:%x, request:%x\n",
+		       arg->engine->name, preempt ? "preempt" : "simple",
+		       arg->poison, slot[1]);
+		err = -EINVAL;
+	}
+
+	timestamp = READ_ONCE(arg->ce[0]->lrc_reg_state[CTX_TIMESTAMP]);
+	if (!timestamp_advanced(slot[1], timestamp)) {
+		pr_err("%s(%s): invalid timestamp on save, request:%x, context:%x\n",
+		       arg->engine->name, preempt ? "preempt" : "simple",
+		       slot[1], timestamp);
+		err = -EINVAL;
+	}
+
+err:
+	memset32(slot, -1, 4);
+	i915_request_put(rq);
+	return err;
+}
+
+static int live_lrc_timestamp(void *arg)
+{
+	struct lrc_timestamp data = {};
+	struct intel_gt *gt = arg;
+	enum intel_engine_id id;
+	const u32 poison[] = {
+		0,
+		S32_MAX,
+		(u32)S32_MAX + 1,
+		U32_MAX,
+	};
+
+	/*
+	 * We want to verify that the timestamp is saved and restore across
+	 * context switches and is monotonic.
+	 *
+	 * So we do this with a little bit of LRC poisoning to check various
+	 * boundary conditions, and see what happens if we preempt the context
+	 * with a second request (carrying more poison into the timestamp).
+	 */
+
+	for_each_engine(data.engine, gt, id) {
+		int i, err = 0;
+
+		st_engine_heartbeat_disable(data.engine);
+
+		for (i = 0; i < ARRAY_SIZE(data.ce); i++) {
+			struct intel_context *tmp;
+
+			tmp = intel_context_create(data.engine);
+			if (IS_ERR(tmp)) {
+				err = PTR_ERR(tmp);
+				goto err;
+			}
+
+			err = intel_context_pin(tmp);
+			if (err) {
+				intel_context_put(tmp);
+				goto err;
+			}
+
+			data.ce[i] = tmp;
+		}
+
+		for (i = 0; i < ARRAY_SIZE(poison); i++) {
+			data.poison = poison[i];
+
+			err = __lrc_timestamp(&data, false);
+			if (err)
+				break;
+
+			err = __lrc_timestamp(&data, true);
+			if (err)
+				break;
+		}
+
+err:
+		st_engine_heartbeat_enable(data.engine);
+		for (i = 0; i < ARRAY_SIZE(data.ce); i++) {
+			if (!data.ce[i])
+				break;
+
+			intel_context_unpin(data.ce[i]);
+			intel_context_put(data.ce[i]);
+		}
+
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static struct i915_vma *
+create_user_vma(struct i915_address_space *vm, unsigned long size)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err;
+
+	obj = i915_gem_object_create_internal(vm->i915, size);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return vma;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err) {
+		i915_gem_object_put(obj);
+		return ERR_PTR(err);
+	}
+
+	return vma;
+}
+
+static u32 safe_poison(u32 offset, u32 poison)
+{
+	/*
+	 * Do not enable predication as it will nop all subsequent commands,
+	 * not only disabling the tests (by preventing all the other SRM) but
+	 * also preventing the arbitration events at the end of the request.
+	 */
+	if (offset == i915_mmio_reg_offset(RING_PREDICATE_RESULT(0)))
+		poison &= ~REG_BIT(0);
+
+	return poison;
+}
+
+static struct i915_vma *
+store_context(struct intel_context *ce, struct i915_vma *scratch)
+{
+	struct i915_vma *batch;
+	u32 dw, x, *cs, *hw;
+	u32 *defaults;
+
+	batch = create_user_vma(ce->vm, SZ_64K);
+	if (IS_ERR(batch))
+		return batch;
+
+	cs = i915_gem_object_pin_map_unlocked(batch->obj, I915_MAP_WC);
+	if (IS_ERR(cs)) {
+		i915_vma_put(batch);
+		return ERR_CAST(cs);
+	}
+
+	defaults = shmem_pin_map(ce->engine->default_state);
+	if (!defaults) {
+		i915_gem_object_unpin_map(batch->obj);
+		i915_vma_put(batch);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	x = 0;
+	dw = 0;
+	hw = defaults;
+	hw += LRC_STATE_OFFSET / sizeof(*hw);
+	do {
+		u32 len = hw[dw] & LRI_LENGTH_MASK;
+
+		/*
+		 * Keep it simple, skip parsing complex commands
+		 *
+		 * At present, there are no more MI_LOAD_REGISTER_IMM
+		 * commands after the first 3D state command. Rather
+		 * than include a table (see i915_cmd_parser.c) of all
+		 * the possible commands and their instruction lengths
+		 * (or mask for variable length instructions), assume
+		 * we have gathered the complete list of registers and
+		 * bail out.
+		 */
+		if ((hw[dw] >> INSTR_CLIENT_SHIFT) != INSTR_MI_CLIENT)
+			break;
+
+		if (hw[dw] == 0) {
+			dw++;
+			continue;
+		}
+
+		if ((hw[dw] & GENMASK(31, 23)) != LRI_HEADER) {
+			/* Assume all other MI commands match LRI length mask */
+			dw += len + 2;
+			continue;
+		}
+
+		if (!len) {
+			pr_err("%s: invalid LRI found in context image\n",
+			       ce->engine->name);
+			igt_hexdump(defaults, PAGE_SIZE);
+			break;
+		}
+
+		dw++;
+		len = (len + 1) / 2;
+		while (len--) {
+			*cs++ = MI_STORE_REGISTER_MEM_GEN8;
+			*cs++ = hw[dw];
+			*cs++ = lower_32_bits(scratch->node.start + x);
+			*cs++ = upper_32_bits(scratch->node.start + x);
+
+			dw += 2;
+			x += 4;
+		}
+	} while (dw < PAGE_SIZE / sizeof(u32) &&
+		 (hw[dw] & ~BIT(0)) != MI_BATCH_BUFFER_END);
+
+	*cs++ = MI_BATCH_BUFFER_END;
+
+	shmem_unpin_map(ce->engine->default_state, defaults);
+
+	i915_gem_object_flush_map(batch->obj);
+	i915_gem_object_unpin_map(batch->obj);
+
+	return batch;
+}
+
+static int move_to_active(struct i915_request *rq,
+			  struct i915_vma *vma,
+			  unsigned int flags)
+{
+	int err;
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, flags);
+	if (!err)
+		err = i915_vma_move_to_active(vma, rq, flags);
+	i915_vma_unlock(vma);
+
+	return err;
+}
+
+static struct i915_request *
+record_registers(struct intel_context *ce,
+		 struct i915_vma *before,
+		 struct i915_vma *after,
+		 u32 *sema)
+{
+	struct i915_vma *b_before, *b_after;
+	struct i915_request *rq;
+	u32 *cs;
+	int err;
+
+	b_before = store_context(ce, before);
+	if (IS_ERR(b_before))
+		return ERR_CAST(b_before);
+
+	b_after = store_context(ce, after);
+	if (IS_ERR(b_after)) {
+		rq = ERR_CAST(b_after);
+		goto err_before;
+	}
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		goto err_after;
+
+	err = move_to_active(rq, before, EXEC_OBJECT_WRITE);
+	if (err)
+		goto err_rq;
+
+	err = move_to_active(rq, b_before, 0);
+	if (err)
+		goto err_rq;
+
+	err = move_to_active(rq, after, EXEC_OBJECT_WRITE);
+	if (err)
+		goto err_rq;
+
+	err = move_to_active(rq, b_after, 0);
+	if (err)
+		goto err_rq;
+
+	cs = intel_ring_begin(rq, 14);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_rq;
+	}
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 | BIT(8);
+	*cs++ = lower_32_bits(b_before->node.start);
+	*cs++ = upper_32_bits(b_before->node.start);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		MI_SEMAPHORE_SAD_NEQ_SDD;
+	*cs++ = 0;
+	*cs++ = i915_ggtt_offset(ce->engine->status_page.vma) +
+		offset_in_page(sema);
+	*cs++ = 0;
+	*cs++ = MI_NOOP;
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 | BIT(8);
+	*cs++ = lower_32_bits(b_after->node.start);
+	*cs++ = upper_32_bits(b_after->node.start);
+
+	intel_ring_advance(rq, cs);
+
+	WRITE_ONCE(*sema, 0);
+	i915_request_get(rq);
+	i915_request_add(rq);
+err_after:
+	i915_vma_put(b_after);
+err_before:
+	i915_vma_put(b_before);
+	return rq;
+
+err_rq:
+	i915_request_add(rq);
+	rq = ERR_PTR(err);
+	goto err_after;
+}
+
+static struct i915_vma *load_context(struct intel_context *ce, u32 poison)
+{
+	struct i915_vma *batch;
+	u32 dw, *cs, *hw;
+	u32 *defaults;
+
+	batch = create_user_vma(ce->vm, SZ_64K);
+	if (IS_ERR(batch))
+		return batch;
+
+	cs = i915_gem_object_pin_map_unlocked(batch->obj, I915_MAP_WC);
+	if (IS_ERR(cs)) {
+		i915_vma_put(batch);
+		return ERR_CAST(cs);
+	}
+
+	defaults = shmem_pin_map(ce->engine->default_state);
+	if (!defaults) {
+		i915_gem_object_unpin_map(batch->obj);
+		i915_vma_put(batch);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	dw = 0;
+	hw = defaults;
+	hw += LRC_STATE_OFFSET / sizeof(*hw);
+	do {
+		u32 len = hw[dw] & LRI_LENGTH_MASK;
+
+		/* For simplicity, break parsing at the first complex command */
+		if ((hw[dw] >> INSTR_CLIENT_SHIFT) != INSTR_MI_CLIENT)
+			break;
+
+		if (hw[dw] == 0) {
+			dw++;
+			continue;
+		}
+
+		if ((hw[dw] & GENMASK(31, 23)) != LRI_HEADER) {
+			dw += len + 2;
+			continue;
+		}
+
+		if (!len) {
+			pr_err("%s: invalid LRI found in context image\n",
+			       ce->engine->name);
+			igt_hexdump(defaults, PAGE_SIZE);
+			break;
+		}
+
+		dw++;
+		len = (len + 1) / 2;
+		*cs++ = MI_LOAD_REGISTER_IMM(len);
+		while (len--) {
+			*cs++ = hw[dw];
+			*cs++ = safe_poison(hw[dw] & get_lri_mask(ce->engine,
+								  MI_LRI_LRM_CS_MMIO),
+					    poison);
+			dw += 2;
+		}
+	} while (dw < PAGE_SIZE / sizeof(u32) &&
+		 (hw[dw] & ~BIT(0)) != MI_BATCH_BUFFER_END);
+
+	*cs++ = MI_BATCH_BUFFER_END;
+
+	shmem_unpin_map(ce->engine->default_state, defaults);
+
+	i915_gem_object_flush_map(batch->obj);
+	i915_gem_object_unpin_map(batch->obj);
+
+	return batch;
+}
+
+static int poison_registers(struct intel_context *ce, u32 poison, u32 *sema)
+{
+	struct i915_request *rq;
+	struct i915_vma *batch;
+	u32 *cs;
+	int err;
+
+	batch = load_context(ce, poison);
+	if (IS_ERR(batch))
+		return PTR_ERR(batch);
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_batch;
+	}
+
+	err = move_to_active(rq, batch, 0);
+	if (err)
+		goto err_rq;
+
+	cs = intel_ring_begin(rq, 8);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_rq;
+	}
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 | BIT(8);
+	*cs++ = lower_32_bits(batch->node.start);
+	*cs++ = upper_32_bits(batch->node.start);
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = i915_ggtt_offset(ce->engine->status_page.vma) +
+		offset_in_page(sema);
+	*cs++ = 0;
+	*cs++ = 1;
+
+	intel_ring_advance(rq, cs);
+
+	rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+err_rq:
+	i915_request_add(rq);
+err_batch:
+	i915_vma_put(batch);
+	return err;
+}
+
+static bool is_moving(u32 a, u32 b)
+{
+	return a != b;
+}
+
+static int compare_isolation(struct intel_engine_cs *engine,
+			     struct i915_vma *ref[2],
+			     struct i915_vma *result[2],
+			     struct intel_context *ce,
+			     u32 poison)
+{
+	u32 x, dw, *hw, *lrc;
+	u32 *A[2], *B[2];
+	u32 *defaults;
+	int err = 0;
+
+	A[0] = i915_gem_object_pin_map_unlocked(ref[0]->obj, I915_MAP_WC);
+	if (IS_ERR(A[0]))
+		return PTR_ERR(A[0]);
+
+	A[1] = i915_gem_object_pin_map_unlocked(ref[1]->obj, I915_MAP_WC);
+	if (IS_ERR(A[1])) {
+		err = PTR_ERR(A[1]);
+		goto err_A0;
+	}
+
+	B[0] = i915_gem_object_pin_map_unlocked(result[0]->obj, I915_MAP_WC);
+	if (IS_ERR(B[0])) {
+		err = PTR_ERR(B[0]);
+		goto err_A1;
+	}
+
+	B[1] = i915_gem_object_pin_map_unlocked(result[1]->obj, I915_MAP_WC);
+	if (IS_ERR(B[1])) {
+		err = PTR_ERR(B[1]);
+		goto err_B0;
+	}
+
+	lrc = i915_gem_object_pin_map_unlocked(ce->state->obj,
+					       i915_coherent_map_type(engine->i915,
+								      ce->state->obj,
+								      false));
+	if (IS_ERR(lrc)) {
+		err = PTR_ERR(lrc);
+		goto err_B1;
+	}
+	lrc += LRC_STATE_OFFSET / sizeof(*hw);
+
+	defaults = shmem_pin_map(ce->engine->default_state);
+	if (!defaults) {
+		err = -ENOMEM;
+		goto err_lrc;
+	}
+
+	x = 0;
+	dw = 0;
+	hw = defaults;
+	hw += LRC_STATE_OFFSET / sizeof(*hw);
+	do {
+		u32 len = hw[dw] & LRI_LENGTH_MASK;
+
+		/* For simplicity, break parsing at the first complex command */
+		if ((hw[dw] >> INSTR_CLIENT_SHIFT) != INSTR_MI_CLIENT)
+			break;
+
+		if (hw[dw] == 0) {
+			dw++;
+			continue;
+		}
+
+		if ((hw[dw] & GENMASK(31, 23)) != LRI_HEADER) {
+			dw += len + 2;
+			continue;
+		}
+
+		if (!len) {
+			pr_err("%s: invalid LRI found in context image\n",
+			       engine->name);
+			igt_hexdump(defaults, PAGE_SIZE);
+			break;
+		}
+
+		dw++;
+		len = (len + 1) / 2;
+		while (len--) {
+			if (!is_moving(A[0][x], A[1][x]) &&
+			    (A[0][x] != B[0][x] || A[1][x] != B[1][x])) {
+				switch (hw[dw] & 4095) {
+				case 0x30: /* RING_HEAD */
+				case 0x34: /* RING_TAIL */
+					break;
+
+				default:
+					pr_err("%s[%d]: Mismatch for register %4x, default %08x, reference %08x, result (%08x, %08x), poison %08x, context %08x\n",
+					       engine->name, dw,
+					       hw[dw], hw[dw + 1],
+					       A[0][x], B[0][x], B[1][x],
+					       poison, lrc[dw + 1]);
+					err = -EINVAL;
+				}
+			}
+			dw += 2;
+			x++;
+		}
+	} while (dw < PAGE_SIZE / sizeof(u32) &&
+		 (hw[dw] & ~BIT(0)) != MI_BATCH_BUFFER_END);
+
+	shmem_unpin_map(ce->engine->default_state, defaults);
+err_lrc:
+	i915_gem_object_unpin_map(ce->state->obj);
+err_B1:
+	i915_gem_object_unpin_map(result[1]->obj);
+err_B0:
+	i915_gem_object_unpin_map(result[0]->obj);
+err_A1:
+	i915_gem_object_unpin_map(ref[1]->obj);
+err_A0:
+	i915_gem_object_unpin_map(ref[0]->obj);
+	return err;
+}
+
+static struct i915_vma *
+create_result_vma(struct i915_address_space *vm, unsigned long sz)
+{
+	struct i915_vma *vma;
+	void *ptr;
+
+	vma = create_user_vma(vm, sz);
+	if (IS_ERR(vma))
+		return vma;
+
+	/* Set the results to a known value distinct from the poison */
+	ptr = i915_gem_object_pin_map_unlocked(vma->obj, I915_MAP_WC);
+	if (IS_ERR(ptr)) {
+		i915_vma_put(vma);
+		return ERR_CAST(ptr);
+	}
+
+	memset(ptr, POISON_INUSE, vma->size);
+	i915_gem_object_flush_map(vma->obj);
+	i915_gem_object_unpin_map(vma->obj);
+
+	return vma;
+}
+
+static int __lrc_isolation(struct intel_engine_cs *engine, u32 poison)
+{
+	u32 *sema = memset32(engine->status_page.addr + 1000, 0, 1);
+	struct i915_vma *ref[2], *result[2];
+	struct intel_context *A, *B;
+	struct i915_request *rq;
+	int err;
+
+	A = intel_context_create(engine);
+	if (IS_ERR(A))
+		return PTR_ERR(A);
+
+	B = intel_context_create(engine);
+	if (IS_ERR(B)) {
+		err = PTR_ERR(B);
+		goto err_A;
+	}
+
+	ref[0] = create_result_vma(A->vm, SZ_64K);
+	if (IS_ERR(ref[0])) {
+		err = PTR_ERR(ref[0]);
+		goto err_B;
+	}
+
+	ref[1] = create_result_vma(A->vm, SZ_64K);
+	if (IS_ERR(ref[1])) {
+		err = PTR_ERR(ref[1]);
+		goto err_ref0;
+	}
+
+	rq = record_registers(A, ref[0], ref[1], sema);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_ref1;
+	}
+
+	WRITE_ONCE(*sema, 1);
+	wmb();
+
+	if (i915_request_wait(rq, 0, HZ / 2) < 0) {
+		i915_request_put(rq);
+		err = -ETIME;
+		goto err_ref1;
+	}
+	i915_request_put(rq);
+
+	result[0] = create_result_vma(A->vm, SZ_64K);
+	if (IS_ERR(result[0])) {
+		err = PTR_ERR(result[0]);
+		goto err_ref1;
+	}
+
+	result[1] = create_result_vma(A->vm, SZ_64K);
+	if (IS_ERR(result[1])) {
+		err = PTR_ERR(result[1]);
+		goto err_result0;
+	}
+
+	rq = record_registers(A, result[0], result[1], sema);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_result1;
+	}
+
+	err = poison_registers(B, poison, sema);
+	if (err == 0 && i915_request_wait(rq, 0, HZ / 2) < 0) {
+		pr_err("%s(%s): wait for results timed out\n",
+		       __func__, engine->name);
+		err = -ETIME;
+	}
+
+	/* Always cancel the semaphore wait, just in case the GPU gets stuck */
+	WRITE_ONCE(*sema, -1);
+	i915_request_put(rq);
+	if (err)
+		goto err_result1;
+
+	err = compare_isolation(engine, ref, result, A, poison);
+
+err_result1:
+	i915_vma_put(result[1]);
+err_result0:
+	i915_vma_put(result[0]);
+err_ref1:
+	i915_vma_put(ref[1]);
+err_ref0:
+	i915_vma_put(ref[0]);
+err_B:
+	intel_context_put(B);
+err_A:
+	intel_context_put(A);
+	return err;
+}
+
+static bool skip_isolation(const struct intel_engine_cs *engine)
+{
+	if (engine->class == COPY_ENGINE_CLASS && GRAPHICS_VER(engine->i915) == 9)
+		return true;
+
+	if (engine->class == RENDER_CLASS && GRAPHICS_VER(engine->i915) == 11)
+		return true;
+
+	return false;
+}
+
+static int live_lrc_isolation(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	const u32 poison[] = {
+		STACK_MAGIC,
+		0x3a3a3a3a,
+		0x5c5c5c5c,
+		0xffffffff,
+		0xffff0000,
+	};
+	int err = 0;
+
+	/*
+	 * Our goal is try and verify that per-context state cannot be
+	 * tampered with by another non-privileged client.
+	 *
+	 * We take the list of context registers from the LRI in the default
+	 * context image and attempt to modify that list from a remote context.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		int i;
+
+		/* Just don't even ask */
+		if (!IS_ENABLED(CONFIG_DRM_I915_SELFTEST_BROKEN) &&
+		    skip_isolation(engine))
+			continue;
+
+		intel_engine_pm_get(engine);
+		for (i = 0; i < ARRAY_SIZE(poison); i++) {
+			int result;
+
+			result = __lrc_isolation(engine, poison[i]);
+			if (result && !err)
+				err = result;
+
+			result = __lrc_isolation(engine, ~poison[i]);
+			if (result && !err)
+				err = result;
+		}
+		intel_engine_pm_put(engine);
+		if (igt_flush_test(gt->i915)) {
+			err = -EIO;
+			break;
+		}
+	}
+
+	return err;
+}
+
+static int indirect_ctx_submit_req(struct intel_context *ce)
+{
+	struct i915_request *rq;
+	int err = 0;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	if (i915_request_wait(rq, 0, HZ / 5) < 0)
+		err = -ETIME;
+
+	i915_request_put(rq);
+
+	return err;
+}
+
+#define CTX_BB_CANARY_OFFSET (3 * 1024)
+#define CTX_BB_CANARY_INDEX  (CTX_BB_CANARY_OFFSET / sizeof(u32))
+
+static u32 *
+emit_indirect_ctx_bb_canary(const struct intel_context *ce, u32 *cs)
+{
+	*cs++ = MI_STORE_REGISTER_MEM_GEN8 |
+		MI_SRM_LRM_GLOBAL_GTT |
+		MI_LRI_LRM_CS_MMIO;
+	*cs++ = i915_mmio_reg_offset(RING_START(0));
+	*cs++ = i915_ggtt_offset(ce->state) +
+		context_wa_bb_offset(ce) +
+		CTX_BB_CANARY_OFFSET;
+	*cs++ = 0;
+
+	return cs;
+}
+
+static void
+indirect_ctx_bb_setup(struct intel_context *ce)
+{
+	u32 *cs = context_indirect_bb(ce);
+
+	cs[CTX_BB_CANARY_INDEX] = 0xdeadf00d;
+
+	setup_indirect_ctx_bb(ce, ce->engine, emit_indirect_ctx_bb_canary);
+}
+
+static bool check_ring_start(struct intel_context *ce)
+{
+	const u32 * const ctx_bb = (void *)(ce->lrc_reg_state) -
+		LRC_STATE_OFFSET + context_wa_bb_offset(ce);
+
+	if (ctx_bb[CTX_BB_CANARY_INDEX] == ce->lrc_reg_state[CTX_RING_START])
+		return true;
+
+	pr_err("ring start mismatch: canary 0x%08x vs state 0x%08x\n",
+	       ctx_bb[CTX_BB_CANARY_INDEX],
+	       ce->lrc_reg_state[CTX_RING_START]);
+
+	return false;
+}
+
+static int indirect_ctx_bb_check(struct intel_context *ce)
+{
+	int err;
+
+	err = indirect_ctx_submit_req(ce);
+	if (err)
+		return err;
+
+	if (!check_ring_start(ce))
+		return -EINVAL;
+
+	return 0;
+}
+
+static int __live_lrc_indirect_ctx_bb(struct intel_engine_cs *engine)
+{
+	struct intel_context *a, *b;
+	int err;
+
+	a = intel_context_create(engine);
+	if (IS_ERR(a))
+		return PTR_ERR(a);
+	err = intel_context_pin(a);
+	if (err)
+		goto put_a;
+
+	b = intel_context_create(engine);
+	if (IS_ERR(b)) {
+		err = PTR_ERR(b);
+		goto unpin_a;
+	}
+	err = intel_context_pin(b);
+	if (err)
+		goto put_b;
+
+	/* We use the already reserved extra page in context state */
+	if (!a->wa_bb_page) {
+		GEM_BUG_ON(b->wa_bb_page);
+		GEM_BUG_ON(GRAPHICS_VER(engine->i915) == 12);
+		goto unpin_b;
+	}
+
+	/*
+	 * In order to test that our per context bb is truly per context,
+	 * and executes at the intended spot on context restoring process,
+	 * make the batch store the ring start value to memory.
+	 * As ring start is restored apriori of starting the indirect ctx bb and
+	 * as it will be different for each context, it fits to this purpose.
+	 */
+	indirect_ctx_bb_setup(a);
+	indirect_ctx_bb_setup(b);
+
+	err = indirect_ctx_bb_check(a);
+	if (err)
+		goto unpin_b;
+
+	err = indirect_ctx_bb_check(b);
+
+unpin_b:
+	intel_context_unpin(b);
+put_b:
+	intel_context_put(b);
+unpin_a:
+	intel_context_unpin(a);
+put_a:
+	intel_context_put(a);
+
+	return err;
+}
+
+static int live_lrc_indirect_ctx_bb(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	for_each_engine(engine, gt, id) {
+		intel_engine_pm_get(engine);
+		err = __live_lrc_indirect_ctx_bb(engine);
+		intel_engine_pm_put(engine);
+
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static void garbage_reset(struct intel_engine_cs *engine,
+			  struct i915_request *rq)
+{
+	const unsigned int bit = I915_RESET_ENGINE + engine->id;
+	unsigned long *lock = &engine->gt->reset.flags;
+
+	local_bh_disable();
+	if (!test_and_set_bit(bit, lock)) {
+		tasklet_disable(&engine->sched_engine->tasklet);
+
+		if (!rq->fence.error)
+			__intel_engine_reset_bh(engine, NULL);
+
+		tasklet_enable(&engine->sched_engine->tasklet);
+		clear_and_wake_up_bit(bit, lock);
+	}
+	local_bh_enable();
+}
+
+static struct i915_request *garbage(struct intel_context *ce,
+				    struct rnd_state *prng)
+{
+	struct i915_request *rq;
+	int err;
+
+	err = intel_context_pin(ce);
+	if (err)
+		return ERR_PTR(err);
+
+	prandom_bytes_state(prng,
+			    ce->lrc_reg_state,
+			    ce->engine->context_size -
+			    LRC_STATE_OFFSET);
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_unpin;
+	}
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+	return rq;
+
+err_unpin:
+	intel_context_unpin(ce);
+	return ERR_PTR(err);
+}
+
+static int __lrc_garbage(struct intel_engine_cs *engine, struct rnd_state *prng)
+{
+	struct intel_context *ce;
+	struct i915_request *hang;
+	int err = 0;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	hang = garbage(ce, prng);
+	if (IS_ERR(hang)) {
+		err = PTR_ERR(hang);
+		goto err_ce;
+	}
+
+	if (wait_for_submit(engine, hang, HZ / 2)) {
+		i915_request_put(hang);
+		err = -ETIME;
+		goto err_ce;
+	}
+
+	intel_context_set_banned(ce);
+	garbage_reset(engine, hang);
+
+	intel_engine_flush_submission(engine);
+	if (!hang->fence.error) {
+		i915_request_put(hang);
+		pr_err("%s: corrupted context was not reset\n",
+		       engine->name);
+		err = -EINVAL;
+		goto err_ce;
+	}
+
+	if (i915_request_wait(hang, 0, HZ / 2) < 0) {
+		pr_err("%s: corrupted context did not recover\n",
+		       engine->name);
+		i915_request_put(hang);
+		err = -EIO;
+		goto err_ce;
+	}
+	i915_request_put(hang);
+
+err_ce:
+	intel_context_put(ce);
+	return err;
+}
+
+static int live_lrc_garbage(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * Verify that we can recover if one context state is completely
+	 * corrupted.
+	 */
+
+	if (!IS_ENABLED(CONFIG_DRM_I915_SELFTEST_BROKEN))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		I915_RND_STATE(prng);
+		int err = 0, i;
+
+		if (!intel_has_reset_engine(engine->gt))
+			continue;
+
+		intel_engine_pm_get(engine);
+		for (i = 0; i < 3; i++) {
+			err = __lrc_garbage(engine, &prng);
+			if (err)
+				break;
+		}
+		intel_engine_pm_put(engine);
+
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int __live_pphwsp_runtime(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+	struct i915_request *rq;
+	IGT_TIMEOUT(end_time);
+	int err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	ce->stats.runtime.num_underflow = 0;
+	ce->stats.runtime.max_underflow = 0;
+
+	do {
+		unsigned int loop = 1024;
+
+		while (loop) {
+			rq = intel_context_create_request(ce);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				goto err_rq;
+			}
+
+			if (--loop == 0)
+				i915_request_get(rq);
+
+			i915_request_add(rq);
+		}
+
+		if (__igt_timeout(end_time, NULL))
+			break;
+
+		i915_request_put(rq);
+	} while (1);
+
+	err = i915_request_wait(rq, 0, HZ / 5);
+	if (err < 0) {
+		pr_err("%s: request not completed!\n", engine->name);
+		goto err_wait;
+	}
+
+	igt_flush_test(engine->i915);
+
+	pr_info("%s: pphwsp runtime %lluns, average %lluns\n",
+		engine->name,
+		intel_context_get_total_runtime_ns(ce),
+		intel_context_get_avg_runtime_ns(ce));
+
+	err = 0;
+	if (ce->stats.runtime.num_underflow) {
+		pr_err("%s: pphwsp underflow %u time(s), max %u cycles!\n",
+		       engine->name,
+		       ce->stats.runtime.num_underflow,
+		       ce->stats.runtime.max_underflow);
+		GEM_TRACE_DUMP();
+		err = -EOVERFLOW;
+	}
+
+err_wait:
+	i915_request_put(rq);
+err_rq:
+	intel_context_put(ce);
+	return err;
+}
+
+static int live_pphwsp_runtime(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Check that cumulative context runtime as stored in the pphwsp[16]
+	 * is monotonic.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		err = __live_pphwsp_runtime(engine);
+		if (err)
+			break;
+	}
+
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	return err;
+}
+
+int intel_lrc_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_lrc_layout),
+		SUBTEST(live_lrc_fixed),
+		SUBTEST(live_lrc_state),
+		SUBTEST(live_lrc_gpr),
+		SUBTEST(live_lrc_isolation),
+		SUBTEST(live_lrc_timestamp),
+		SUBTEST(live_lrc_garbage),
+		SUBTEST(live_pphwsp_runtime),
+		SUBTEST(live_lrc_indirect_ctx_bb),
+	};
+
+	if (!HAS_LOGICAL_RING_CONTEXTS(i915))
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_migrate.c b/drivers/gpu/drm/i915/gt/selftest_migrate.c
new file mode 100644
index 000000000..2b0c87999
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_migrate.c
@@ -0,0 +1,871 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/sort.h>
+
+#include "gem/i915_gem_internal.h"
+
+#include "selftests/i915_random.h"
+
+static const unsigned int sizes[] = {
+	SZ_4K,
+	SZ_64K,
+	SZ_2M,
+	CHUNK_SZ - SZ_4K,
+	CHUNK_SZ,
+	CHUNK_SZ + SZ_4K,
+	SZ_64M,
+};
+
+static struct drm_i915_gem_object *
+create_lmem_or_internal(struct drm_i915_private *i915, size_t size)
+{
+	struct drm_i915_gem_object *obj;
+
+	obj = i915_gem_object_create_lmem(i915, size, 0);
+	if (!IS_ERR(obj))
+		return obj;
+
+	return i915_gem_object_create_internal(i915, size);
+}
+
+static int copy(struct intel_migrate *migrate,
+		int (*fn)(struct intel_migrate *migrate,
+			  struct i915_gem_ww_ctx *ww,
+			  struct drm_i915_gem_object *src,
+			  struct drm_i915_gem_object *dst,
+			  struct i915_request **out),
+		u32 sz, struct rnd_state *prng)
+{
+	struct drm_i915_private *i915 = migrate->context->engine->i915;
+	struct drm_i915_gem_object *src, *dst;
+	struct i915_request *rq;
+	struct i915_gem_ww_ctx ww;
+	u32 *vaddr;
+	int err = 0;
+	int i;
+
+	src = create_lmem_or_internal(i915, sz);
+	if (IS_ERR(src))
+		return 0;
+
+	sz = src->base.size;
+	dst = i915_gem_object_create_internal(i915, sz);
+	if (IS_ERR(dst))
+		goto err_free_src;
+
+	for_i915_gem_ww(&ww, err, true) {
+		err = i915_gem_object_lock(src, &ww);
+		if (err)
+			continue;
+
+		err = i915_gem_object_lock(dst, &ww);
+		if (err)
+			continue;
+
+		vaddr = i915_gem_object_pin_map(src, I915_MAP_WC);
+		if (IS_ERR(vaddr)) {
+			err = PTR_ERR(vaddr);
+			continue;
+		}
+
+		for (i = 0; i < sz / sizeof(u32); i++)
+			vaddr[i] = i;
+		i915_gem_object_flush_map(src);
+
+		vaddr = i915_gem_object_pin_map(dst, I915_MAP_WC);
+		if (IS_ERR(vaddr)) {
+			err = PTR_ERR(vaddr);
+			goto unpin_src;
+		}
+
+		for (i = 0; i < sz / sizeof(u32); i++)
+			vaddr[i] = ~i;
+		i915_gem_object_flush_map(dst);
+
+		err = fn(migrate, &ww, src, dst, &rq);
+		if (!err)
+			continue;
+
+		if (err != -EDEADLK && err != -EINTR && err != -ERESTARTSYS)
+			pr_err("%ps failed, size: %u\n", fn, sz);
+		if (rq) {
+			i915_request_wait(rq, 0, HZ);
+			i915_request_put(rq);
+		}
+		i915_gem_object_unpin_map(dst);
+unpin_src:
+		i915_gem_object_unpin_map(src);
+	}
+	if (err)
+		goto err_out;
+
+	if (rq) {
+		if (i915_request_wait(rq, 0, HZ) < 0) {
+			pr_err("%ps timed out, size: %u\n", fn, sz);
+			err = -ETIME;
+		}
+		i915_request_put(rq);
+	}
+
+	for (i = 0; !err && i < sz / PAGE_SIZE; i++) {
+		int x = i * 1024 + i915_prandom_u32_max_state(1024, prng);
+
+		if (vaddr[x] != x) {
+			pr_err("%ps failed, size: %u, offset: %zu\n",
+			       fn, sz, x * sizeof(u32));
+			igt_hexdump(vaddr + i * 1024, 4096);
+			err = -EINVAL;
+		}
+	}
+
+	i915_gem_object_unpin_map(dst);
+	i915_gem_object_unpin_map(src);
+
+err_out:
+	i915_gem_object_put(dst);
+err_free_src:
+	i915_gem_object_put(src);
+
+	return err;
+}
+
+static int intel_context_copy_ccs(struct intel_context *ce,
+				  const struct i915_deps *deps,
+				  struct scatterlist *sg,
+				  enum i915_cache_level cache_level,
+				  bool write_to_ccs,
+				  struct i915_request **out)
+{
+	u8 src_access = write_to_ccs ? DIRECT_ACCESS : INDIRECT_ACCESS;
+	u8 dst_access = write_to_ccs ? INDIRECT_ACCESS : DIRECT_ACCESS;
+	struct sgt_dma it = sg_sgt(sg);
+	struct i915_request *rq;
+	u32 offset;
+	int err;
+
+	GEM_BUG_ON(ce->vm != ce->engine->gt->migrate.context->vm);
+	*out = NULL;
+
+	GEM_BUG_ON(ce->ring->size < SZ_64K);
+
+	offset = 0;
+	if (HAS_64K_PAGES(ce->engine->i915))
+		offset = CHUNK_SZ;
+
+	do {
+		int len;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_ce;
+		}
+
+		if (deps) {
+			err = i915_request_await_deps(rq, deps);
+			if (err)
+				goto out_rq;
+
+			if (rq->engine->emit_init_breadcrumb) {
+				err = rq->engine->emit_init_breadcrumb(rq);
+				if (err)
+					goto out_rq;
+			}
+
+			deps = NULL;
+		}
+
+		/* The PTE updates + clear must not be interrupted. */
+		err = emit_no_arbitration(rq);
+		if (err)
+			goto out_rq;
+
+		len = emit_pte(rq, &it, cache_level, true, offset, CHUNK_SZ);
+		if (len <= 0) {
+			err = len;
+			goto out_rq;
+		}
+
+		err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+		if (err)
+			goto out_rq;
+
+		err = emit_copy_ccs(rq, offset, dst_access,
+				    offset, src_access, len);
+		if (err)
+			goto out_rq;
+
+		err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+
+		/* Arbitration is re-enabled between requests. */
+out_rq:
+		if (*out)
+			i915_request_put(*out);
+		*out = i915_request_get(rq);
+		i915_request_add(rq);
+		if (err || !it.sg || !sg_dma_len(it.sg))
+			break;
+
+		cond_resched();
+	} while (1);
+
+out_ce:
+	return err;
+}
+
+static int
+intel_migrate_ccs_copy(struct intel_migrate *m,
+		       struct i915_gem_ww_ctx *ww,
+		       const struct i915_deps *deps,
+		       struct scatterlist *sg,
+		       enum i915_cache_level cache_level,
+		       bool write_to_ccs,
+		       struct i915_request **out)
+{
+	struct intel_context *ce;
+	int err;
+
+	*out = NULL;
+	if (!m->context)
+		return -ENODEV;
+
+	ce = intel_migrate_create_context(m);
+	if (IS_ERR(ce))
+		ce = intel_context_get(m->context);
+	GEM_BUG_ON(IS_ERR(ce));
+
+	err = intel_context_pin_ww(ce, ww);
+	if (err)
+		goto out;
+
+	err = intel_context_copy_ccs(ce, deps, sg, cache_level,
+				     write_to_ccs, out);
+
+	intel_context_unpin(ce);
+out:
+	intel_context_put(ce);
+	return err;
+}
+
+static int clear(struct intel_migrate *migrate,
+		 int (*fn)(struct intel_migrate *migrate,
+			   struct i915_gem_ww_ctx *ww,
+			   struct drm_i915_gem_object *obj,
+			   u32 value,
+			   struct i915_request **out),
+		 u32 sz, struct rnd_state *prng)
+{
+	struct drm_i915_private *i915 = migrate->context->engine->i915;
+	struct drm_i915_gem_object *obj;
+	struct i915_request *rq;
+	struct i915_gem_ww_ctx ww;
+	u32 *vaddr, val = 0;
+	bool ccs_cap = false;
+	int err = 0;
+	int i;
+
+	obj = create_lmem_or_internal(i915, sz);
+	if (IS_ERR(obj))
+		return 0;
+
+	/* Consider the rounded up memory too */
+	sz = obj->base.size;
+
+	if (HAS_FLAT_CCS(i915) && i915_gem_object_is_lmem(obj))
+		ccs_cap = true;
+
+	for_i915_gem_ww(&ww, err, true) {
+		int ccs_bytes, ccs_bytes_per_chunk;
+
+		err = i915_gem_object_lock(obj, &ww);
+		if (err)
+			continue;
+
+		vaddr = i915_gem_object_pin_map(obj, I915_MAP_WC);
+		if (IS_ERR(vaddr)) {
+			err = PTR_ERR(vaddr);
+			continue;
+		}
+
+		for (i = 0; i < sz / sizeof(u32); i++)
+			vaddr[i] = ~i;
+		i915_gem_object_flush_map(obj);
+
+		if (ccs_cap && !val) {
+			/* Write the obj data into ccs surface */
+			err = intel_migrate_ccs_copy(migrate, &ww, NULL,
+						     obj->mm.pages->sgl,
+						     obj->cache_level,
+						     true, &rq);
+			if (rq && !err) {
+				if (i915_request_wait(rq, 0, HZ) < 0) {
+					pr_err("%ps timed out, size: %u\n",
+					       fn, sz);
+					err = -ETIME;
+				}
+				i915_request_put(rq);
+				rq = NULL;
+			}
+			if (err)
+				continue;
+		}
+
+		err = fn(migrate, &ww, obj, val, &rq);
+		if (rq && !err) {
+			if (i915_request_wait(rq, 0, HZ) < 0) {
+				pr_err("%ps timed out, size: %u\n", fn, sz);
+				err = -ETIME;
+			}
+			i915_request_put(rq);
+			rq = NULL;
+		}
+		if (err)
+			continue;
+
+		i915_gem_object_flush_map(obj);
+
+		/* Verify the set/clear of the obj mem */
+		for (i = 0; !err && i < sz / PAGE_SIZE; i++) {
+			int x = i * 1024 +
+				i915_prandom_u32_max_state(1024, prng);
+
+			if (vaddr[x] != val) {
+				pr_err("%ps failed, (%u != %u), offset: %zu\n",
+				       fn, vaddr[x], val,  x * sizeof(u32));
+				igt_hexdump(vaddr + i * 1024, 4096);
+				err = -EINVAL;
+			}
+		}
+		if (err)
+			continue;
+
+		if (ccs_cap && !val) {
+			for (i = 0; i < sz / sizeof(u32); i++)
+				vaddr[i] = ~i;
+			i915_gem_object_flush_map(obj);
+
+			err = intel_migrate_ccs_copy(migrate, &ww, NULL,
+						     obj->mm.pages->sgl,
+						     obj->cache_level,
+						     false, &rq);
+			if (rq && !err) {
+				if (i915_request_wait(rq, 0, HZ) < 0) {
+					pr_err("%ps timed out, size: %u\n",
+					       fn, sz);
+					err = -ETIME;
+				}
+				i915_request_put(rq);
+				rq = NULL;
+			}
+			if (err)
+				continue;
+
+			ccs_bytes = GET_CCS_BYTES(i915, sz);
+			ccs_bytes_per_chunk = GET_CCS_BYTES(i915, CHUNK_SZ);
+			i915_gem_object_flush_map(obj);
+
+			for (i = 0; !err && i < DIV_ROUND_UP(ccs_bytes, PAGE_SIZE); i++) {
+				int offset = ((i * PAGE_SIZE)  /
+					ccs_bytes_per_chunk) * CHUNK_SZ / sizeof(u32);
+				int ccs_bytes_left = (ccs_bytes - i * PAGE_SIZE) / sizeof(u32);
+				int x = i915_prandom_u32_max_state(min_t(int, 1024,
+									 ccs_bytes_left), prng);
+
+				if (vaddr[offset + x]) {
+					pr_err("%ps ccs clearing failed, offset: %ld/%d\n",
+					       fn, i * PAGE_SIZE + x * sizeof(u32), ccs_bytes);
+					igt_hexdump(vaddr + offset,
+						    min_t(int, 4096,
+							  ccs_bytes_left * sizeof(u32)));
+					err = -EINVAL;
+				}
+			}
+
+			if (err)
+				continue;
+		}
+		i915_gem_object_unpin_map(obj);
+	}
+
+	if (err) {
+		if (err != -EDEADLK && err != -EINTR && err != -ERESTARTSYS)
+			pr_err("%ps failed, size: %u\n", fn, sz);
+		if (rq && err != -EINVAL) {
+			i915_request_wait(rq, 0, HZ);
+			i915_request_put(rq);
+		}
+
+		i915_gem_object_unpin_map(obj);
+	}
+
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static int __migrate_copy(struct intel_migrate *migrate,
+			  struct i915_gem_ww_ctx *ww,
+			  struct drm_i915_gem_object *src,
+			  struct drm_i915_gem_object *dst,
+			  struct i915_request **out)
+{
+	return intel_migrate_copy(migrate, ww, NULL,
+				  src->mm.pages->sgl, src->cache_level,
+				  i915_gem_object_is_lmem(src),
+				  dst->mm.pages->sgl, dst->cache_level,
+				  i915_gem_object_is_lmem(dst),
+				  out);
+}
+
+static int __global_copy(struct intel_migrate *migrate,
+			 struct i915_gem_ww_ctx *ww,
+			 struct drm_i915_gem_object *src,
+			 struct drm_i915_gem_object *dst,
+			 struct i915_request **out)
+{
+	return intel_context_migrate_copy(migrate->context, NULL,
+					  src->mm.pages->sgl, src->cache_level,
+					  i915_gem_object_is_lmem(src),
+					  dst->mm.pages->sgl, dst->cache_level,
+					  i915_gem_object_is_lmem(dst),
+					  out);
+}
+
+static int
+migrate_copy(struct intel_migrate *migrate, u32 sz, struct rnd_state *prng)
+{
+	return copy(migrate, __migrate_copy, sz, prng);
+}
+
+static int
+global_copy(struct intel_migrate *migrate, u32 sz, struct rnd_state *prng)
+{
+	return copy(migrate, __global_copy, sz, prng);
+}
+
+static int __migrate_clear(struct intel_migrate *migrate,
+			   struct i915_gem_ww_ctx *ww,
+			   struct drm_i915_gem_object *obj,
+			   u32 value,
+			   struct i915_request **out)
+{
+	return intel_migrate_clear(migrate, ww, NULL,
+				   obj->mm.pages->sgl,
+				   obj->cache_level,
+				   i915_gem_object_is_lmem(obj),
+				   value, out);
+}
+
+static int __global_clear(struct intel_migrate *migrate,
+			  struct i915_gem_ww_ctx *ww,
+			  struct drm_i915_gem_object *obj,
+			  u32 value,
+			  struct i915_request **out)
+{
+	return intel_context_migrate_clear(migrate->context, NULL,
+					   obj->mm.pages->sgl,
+					   obj->cache_level,
+					   i915_gem_object_is_lmem(obj),
+					   value, out);
+}
+
+static int
+migrate_clear(struct intel_migrate *migrate, u32 sz, struct rnd_state *prng)
+{
+	return clear(migrate, __migrate_clear, sz, prng);
+}
+
+static int
+global_clear(struct intel_migrate *migrate, u32 sz, struct rnd_state *prng)
+{
+	return clear(migrate, __global_clear, sz, prng);
+}
+
+static int live_migrate_copy(void *arg)
+{
+	struct intel_migrate *migrate = arg;
+	struct drm_i915_private *i915 = migrate->context->engine->i915;
+	I915_RND_STATE(prng);
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(sizes); i++) {
+		int err;
+
+		err = migrate_copy(migrate, sizes[i], &prng);
+		if (err == 0)
+			err = global_copy(migrate, sizes[i], &prng);
+		i915_gem_drain_freed_objects(i915);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int live_migrate_clear(void *arg)
+{
+	struct intel_migrate *migrate = arg;
+	struct drm_i915_private *i915 = migrate->context->engine->i915;
+	I915_RND_STATE(prng);
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(sizes); i++) {
+		int err;
+
+		err = migrate_clear(migrate, sizes[i], &prng);
+		if (err == 0)
+			err = global_clear(migrate, sizes[i], &prng);
+
+		i915_gem_drain_freed_objects(i915);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+struct threaded_migrate {
+	struct intel_migrate *migrate;
+	struct task_struct *tsk;
+	struct rnd_state prng;
+};
+
+static int threaded_migrate(struct intel_migrate *migrate,
+			    int (*fn)(void *arg),
+			    unsigned int flags)
+{
+	const unsigned int n_cpus = num_online_cpus() + 1;
+	struct threaded_migrate *thread;
+	I915_RND_STATE(prng);
+	unsigned int i;
+	int err = 0;
+
+	thread = kcalloc(n_cpus, sizeof(*thread), GFP_KERNEL);
+	if (!thread)
+		return 0;
+
+	for (i = 0; i < n_cpus; ++i) {
+		struct task_struct *tsk;
+
+		thread[i].migrate = migrate;
+		thread[i].prng =
+			I915_RND_STATE_INITIALIZER(prandom_u32_state(&prng));
+
+		tsk = kthread_run(fn, &thread[i], "igt-%d", i);
+		if (IS_ERR(tsk)) {
+			err = PTR_ERR(tsk);
+			break;
+		}
+
+		get_task_struct(tsk);
+		thread[i].tsk = tsk;
+	}
+
+	msleep(10); /* start all threads before we kthread_stop() */
+
+	for (i = 0; i < n_cpus; ++i) {
+		struct task_struct *tsk = thread[i].tsk;
+		int status;
+
+		if (IS_ERR_OR_NULL(tsk))
+			continue;
+
+		status = kthread_stop(tsk);
+		if (status && !err)
+			err = status;
+
+		put_task_struct(tsk);
+	}
+
+	kfree(thread);
+	return err;
+}
+
+static int __thread_migrate_copy(void *arg)
+{
+	struct threaded_migrate *tm = arg;
+
+	return migrate_copy(tm->migrate, 2 * CHUNK_SZ, &tm->prng);
+}
+
+static int thread_migrate_copy(void *arg)
+{
+	return threaded_migrate(arg, __thread_migrate_copy, 0);
+}
+
+static int __thread_global_copy(void *arg)
+{
+	struct threaded_migrate *tm = arg;
+
+	return global_copy(tm->migrate, 2 * CHUNK_SZ, &tm->prng);
+}
+
+static int thread_global_copy(void *arg)
+{
+	return threaded_migrate(arg, __thread_global_copy, 0);
+}
+
+static int __thread_migrate_clear(void *arg)
+{
+	struct threaded_migrate *tm = arg;
+
+	return migrate_clear(tm->migrate, 2 * CHUNK_SZ, &tm->prng);
+}
+
+static int __thread_global_clear(void *arg)
+{
+	struct threaded_migrate *tm = arg;
+
+	return global_clear(tm->migrate, 2 * CHUNK_SZ, &tm->prng);
+}
+
+static int thread_migrate_clear(void *arg)
+{
+	return threaded_migrate(arg, __thread_migrate_clear, 0);
+}
+
+static int thread_global_clear(void *arg)
+{
+	return threaded_migrate(arg, __thread_global_clear, 0);
+}
+
+int intel_migrate_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_migrate_copy),
+		SUBTEST(live_migrate_clear),
+		SUBTEST(thread_migrate_copy),
+		SUBTEST(thread_migrate_clear),
+		SUBTEST(thread_global_copy),
+		SUBTEST(thread_global_clear),
+	};
+	struct intel_gt *gt = to_gt(i915);
+
+	if (!gt->migrate.context)
+		return 0;
+
+	return i915_subtests(tests, &gt->migrate);
+}
+
+static struct drm_i915_gem_object *
+create_init_lmem_internal(struct intel_gt *gt, size_t sz, bool try_lmem)
+{
+	struct drm_i915_gem_object *obj = NULL;
+	int err;
+
+	if (try_lmem)
+		obj = i915_gem_object_create_lmem(gt->i915, sz, 0);
+
+	if (IS_ERR_OR_NULL(obj)) {
+		obj = i915_gem_object_create_internal(gt->i915, sz);
+		if (IS_ERR(obj))
+			return obj;
+	}
+
+	i915_gem_object_trylock(obj, NULL);
+	err = i915_gem_object_pin_pages(obj);
+	if (err) {
+		i915_gem_object_unlock(obj);
+		i915_gem_object_put(obj);
+		return ERR_PTR(err);
+	}
+
+	return obj;
+}
+
+static int wrap_ktime_compare(const void *A, const void *B)
+{
+	const ktime_t *a = A, *b = B;
+
+	return ktime_compare(*a, *b);
+}
+
+static int __perf_clear_blt(struct intel_context *ce,
+			    struct scatterlist *sg,
+			    enum i915_cache_level cache_level,
+			    bool is_lmem,
+			    size_t sz)
+{
+	ktime_t t[5];
+	int pass;
+	int err = 0;
+
+	for (pass = 0; pass < ARRAY_SIZE(t); pass++) {
+		struct i915_request *rq;
+		ktime_t t0, t1;
+
+		t0 = ktime_get();
+
+		err = intel_context_migrate_clear(ce, NULL, sg, cache_level,
+						  is_lmem, 0, &rq);
+		if (rq) {
+			if (i915_request_wait(rq, 0, MAX_SCHEDULE_TIMEOUT) < 0)
+				err = -EIO;
+			i915_request_put(rq);
+		}
+		if (err)
+			break;
+
+		t1 = ktime_get();
+		t[pass] = ktime_sub(t1, t0);
+	}
+	if (err)
+		return err;
+
+	sort(t, ARRAY_SIZE(t), sizeof(*t), wrap_ktime_compare, NULL);
+	pr_info("%s: %zd KiB fill: %lld MiB/s\n",
+		ce->engine->name, sz >> 10,
+		div64_u64(mul_u32_u32(4 * sz,
+				      1000 * 1000 * 1000),
+			  t[1] + 2 * t[2] + t[3]) >> 20);
+	return 0;
+}
+
+static int perf_clear_blt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	static const unsigned long sizes[] = {
+		SZ_4K,
+		SZ_64K,
+		SZ_2M,
+		SZ_64M
+	};
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(sizes); i++) {
+		struct drm_i915_gem_object *dst;
+		int err;
+
+		dst = create_init_lmem_internal(gt, sizes[i], true);
+		if (IS_ERR(dst))
+			return PTR_ERR(dst);
+
+		err = __perf_clear_blt(gt->migrate.context,
+				       dst->mm.pages->sgl,
+				       I915_CACHE_NONE,
+				       i915_gem_object_is_lmem(dst),
+				       sizes[i]);
+
+		i915_gem_object_unlock(dst);
+		i915_gem_object_put(dst);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int __perf_copy_blt(struct intel_context *ce,
+			   struct scatterlist *src,
+			   enum i915_cache_level src_cache_level,
+			   bool src_is_lmem,
+			   struct scatterlist *dst,
+			   enum i915_cache_level dst_cache_level,
+			   bool dst_is_lmem,
+			   size_t sz)
+{
+	ktime_t t[5];
+	int pass;
+	int err = 0;
+
+	for (pass = 0; pass < ARRAY_SIZE(t); pass++) {
+		struct i915_request *rq;
+		ktime_t t0, t1;
+
+		t0 = ktime_get();
+
+		err = intel_context_migrate_copy(ce, NULL,
+						 src, src_cache_level,
+						 src_is_lmem,
+						 dst, dst_cache_level,
+						 dst_is_lmem,
+						 &rq);
+		if (rq) {
+			if (i915_request_wait(rq, 0, MAX_SCHEDULE_TIMEOUT) < 0)
+				err = -EIO;
+			i915_request_put(rq);
+		}
+		if (err)
+			break;
+
+		t1 = ktime_get();
+		t[pass] = ktime_sub(t1, t0);
+	}
+	if (err)
+		return err;
+
+	sort(t, ARRAY_SIZE(t), sizeof(*t), wrap_ktime_compare, NULL);
+	pr_info("%s: %zd KiB copy: %lld MiB/s\n",
+		ce->engine->name, sz >> 10,
+		div64_u64(mul_u32_u32(4 * sz,
+				      1000 * 1000 * 1000),
+			  t[1] + 2 * t[2] + t[3]) >> 20);
+	return 0;
+}
+
+static int perf_copy_blt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	static const unsigned long sizes[] = {
+		SZ_4K,
+		SZ_64K,
+		SZ_2M,
+		SZ_64M
+	};
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(sizes); i++) {
+		struct drm_i915_gem_object *src, *dst;
+		size_t sz;
+		int err;
+
+		src = create_init_lmem_internal(gt, sizes[i], true);
+		if (IS_ERR(src))
+			return PTR_ERR(src);
+
+		sz = src->base.size;
+		dst = create_init_lmem_internal(gt, sz, false);
+		if (IS_ERR(dst)) {
+			err = PTR_ERR(dst);
+			goto err_src;
+		}
+
+		err = __perf_copy_blt(gt->migrate.context,
+				      src->mm.pages->sgl,
+				      I915_CACHE_NONE,
+				      i915_gem_object_is_lmem(src),
+				      dst->mm.pages->sgl,
+				      I915_CACHE_NONE,
+				      i915_gem_object_is_lmem(dst),
+				      sz);
+
+		i915_gem_object_unlock(dst);
+		i915_gem_object_put(dst);
+err_src:
+		i915_gem_object_unlock(src);
+		i915_gem_object_put(src);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int intel_migrate_perf_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(perf_clear_blt),
+		SUBTEST(perf_copy_blt),
+	};
+	struct intel_gt *gt = to_gt(i915);
+
+	if (intel_gt_is_wedged(gt))
+		return 0;
+
+	if (!gt->migrate.context)
+		return 0;
+
+	return intel_gt_live_subtests(tests, gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_mocs.c b/drivers/gpu/drm/i915/gt/selftest_mocs.c
new file mode 100644
index 000000000..c1d861333
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_mocs.c
@@ -0,0 +1,455 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_gpu_commands.h"
+#include "i915_selftest.h"
+
+#include "gem/selftests/mock_context.h"
+#include "selftests/igt_reset.h"
+#include "selftests/igt_spinner.h"
+#include "selftests/intel_scheduler_helpers.h"
+
+struct live_mocs {
+	struct drm_i915_mocs_table table;
+	struct drm_i915_mocs_table *mocs;
+	struct drm_i915_mocs_table *l3cc;
+	struct i915_vma *scratch;
+	void *vaddr;
+};
+
+static struct intel_context *mocs_context_create(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return ce;
+
+	/* We build large requests to read the registers from the ring */
+	ce->ring_size = SZ_16K;
+
+	return ce;
+}
+
+static int request_add_sync(struct i915_request *rq, int err)
+{
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (i915_request_wait(rq, 0, HZ / 5) < 0)
+		err = -ETIME;
+	i915_request_put(rq);
+
+	return err;
+}
+
+static int request_add_spin(struct i915_request *rq, struct igt_spinner *spin)
+{
+	int err = 0;
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (spin && !igt_wait_for_spinner(spin, rq))
+		err = -ETIME;
+	i915_request_put(rq);
+
+	return err;
+}
+
+static int live_mocs_init(struct live_mocs *arg, struct intel_gt *gt)
+{
+	unsigned int flags;
+	int err;
+
+	memset(arg, 0, sizeof(*arg));
+
+	flags = get_mocs_settings(gt->i915, &arg->table);
+	if (!flags)
+		return -EINVAL;
+
+	if (flags & HAS_RENDER_L3CC)
+		arg->l3cc = &arg->table;
+
+	if (flags & (HAS_GLOBAL_MOCS | HAS_ENGINE_MOCS))
+		arg->mocs = &arg->table;
+
+	arg->scratch =
+		__vm_create_scratch_for_read_pinned(&gt->ggtt->vm, PAGE_SIZE);
+	if (IS_ERR(arg->scratch))
+		return PTR_ERR(arg->scratch);
+
+	arg->vaddr = i915_gem_object_pin_map_unlocked(arg->scratch->obj, I915_MAP_WB);
+	if (IS_ERR(arg->vaddr)) {
+		err = PTR_ERR(arg->vaddr);
+		goto err_scratch;
+	}
+
+	return 0;
+
+err_scratch:
+	i915_vma_unpin_and_release(&arg->scratch, 0);
+	return err;
+}
+
+static void live_mocs_fini(struct live_mocs *arg)
+{
+	i915_vma_unpin_and_release(&arg->scratch, I915_VMA_RELEASE_MAP);
+}
+
+static int read_regs(struct i915_request *rq,
+		     u32 addr, unsigned int count,
+		     u32 *offset)
+{
+	unsigned int i;
+	u32 *cs;
+
+	GEM_BUG_ON(!IS_ALIGNED(*offset, sizeof(u32)));
+
+	cs = intel_ring_begin(rq, 4 * count);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	for (i = 0; i < count; i++) {
+		*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+		*cs++ = addr;
+		*cs++ = *offset;
+		*cs++ = 0;
+
+		addr += sizeof(u32);
+		*offset += sizeof(u32);
+	}
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int read_mocs_table(struct i915_request *rq,
+			   const struct drm_i915_mocs_table *table,
+			   u32 *offset)
+{
+	u32 addr;
+
+	if (!table)
+		return 0;
+
+	if (HAS_GLOBAL_MOCS_REGISTERS(rq->engine->i915))
+		addr = global_mocs_offset();
+	else
+		addr = mocs_offset(rq->engine);
+
+	return read_regs(rq, addr, table->n_entries, offset);
+}
+
+static int read_l3cc_table(struct i915_request *rq,
+			   const struct drm_i915_mocs_table *table,
+			   u32 *offset)
+{
+	u32 addr = i915_mmio_reg_offset(GEN9_LNCFCMOCS(0));
+
+	if (!table)
+		return 0;
+
+	return read_regs(rq, addr, (table->n_entries + 1) / 2, offset);
+}
+
+static int check_mocs_table(struct intel_engine_cs *engine,
+			    const struct drm_i915_mocs_table *table,
+			    u32 **vaddr)
+{
+	unsigned int i;
+	u32 expect;
+
+	if (!table)
+		return 0;
+
+	for_each_mocs(expect, table, i) {
+		if (**vaddr != expect) {
+			pr_err("%s: Invalid MOCS[%d] entry, found %08x, expected %08x\n",
+			       engine->name, i, **vaddr, expect);
+			return -EINVAL;
+		}
+		++*vaddr;
+	}
+
+	return 0;
+}
+
+static bool mcr_range(struct drm_i915_private *i915, u32 offset)
+{
+	/*
+	 * Registers in this range are affected by the MCR selector
+	 * which only controls CPU initiated MMIO. Routing does not
+	 * work for CS access so we cannot verify them on this path.
+	 */
+	return GRAPHICS_VER(i915) >= 8 && offset >= 0xb000 && offset <= 0xb4ff;
+}
+
+static int check_l3cc_table(struct intel_engine_cs *engine,
+			    const struct drm_i915_mocs_table *table,
+			    u32 **vaddr)
+{
+	/* Can we read the MCR range 0xb00 directly? See intel_workarounds! */
+	u32 reg = i915_mmio_reg_offset(GEN9_LNCFCMOCS(0));
+	unsigned int i;
+	u32 expect;
+
+	if (!table)
+		return 0;
+
+	for_each_l3cc(expect, table, i) {
+		if (!mcr_range(engine->i915, reg) && **vaddr != expect) {
+			pr_err("%s: Invalid L3CC[%d] entry, found %08x, expected %08x\n",
+			       engine->name, i, **vaddr, expect);
+			return -EINVAL;
+		}
+		++*vaddr;
+		reg += 4;
+	}
+
+	return 0;
+}
+
+static int check_mocs_engine(struct live_mocs *arg,
+			     struct intel_context *ce)
+{
+	struct i915_vma *vma = arg->scratch;
+	struct i915_request *rq;
+	u32 offset;
+	u32 *vaddr;
+	int err;
+
+	memset32(arg->vaddr, STACK_MAGIC, PAGE_SIZE / sizeof(u32));
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, true);
+	if (!err)
+		err = i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE);
+	i915_vma_unlock(vma);
+
+	/* Read the mocs tables back using SRM */
+	offset = i915_ggtt_offset(vma);
+	if (!err)
+		err = read_mocs_table(rq, arg->mocs, &offset);
+	if (!err && ce->engine->class == RENDER_CLASS)
+		err = read_l3cc_table(rq, arg->l3cc, &offset);
+	offset -= i915_ggtt_offset(vma);
+	GEM_BUG_ON(offset > PAGE_SIZE);
+
+	err = request_add_sync(rq, err);
+	if (err)
+		return err;
+
+	/* Compare the results against the expected tables */
+	vaddr = arg->vaddr;
+	if (!err)
+		err = check_mocs_table(ce->engine, arg->mocs, &vaddr);
+	if (!err && ce->engine->class == RENDER_CLASS)
+		err = check_l3cc_table(ce->engine, arg->l3cc, &vaddr);
+	if (err)
+		return err;
+
+	GEM_BUG_ON(arg->vaddr + offset != vaddr);
+	return 0;
+}
+
+static int live_mocs_kernel(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct live_mocs mocs;
+	int err;
+
+	/* Basic check the system is configured with the expected mocs table */
+
+	err = live_mocs_init(&mocs, gt);
+	if (err)
+		return err;
+
+	for_each_engine(engine, gt, id) {
+		intel_engine_pm_get(engine);
+		err = check_mocs_engine(&mocs, engine->kernel_context);
+		intel_engine_pm_put(engine);
+		if (err)
+			break;
+	}
+
+	live_mocs_fini(&mocs);
+	return err;
+}
+
+static int live_mocs_clean(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct live_mocs mocs;
+	int err;
+
+	/* Every new context should see the same mocs table */
+
+	err = live_mocs_init(&mocs, gt);
+	if (err)
+		return err;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+
+		ce = mocs_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			break;
+		}
+
+		err = check_mocs_engine(&mocs, ce);
+		intel_context_put(ce);
+		if (err)
+			break;
+	}
+
+	live_mocs_fini(&mocs);
+	return err;
+}
+
+static int active_engine_reset(struct intel_context *ce,
+			       const char *reason,
+			       bool using_guc)
+{
+	struct igt_spinner spin;
+	struct i915_request *rq;
+	int err;
+
+	err = igt_spinner_init(&spin, ce->engine->gt);
+	if (err)
+		return err;
+
+	rq = igt_spinner_create_request(&spin, ce, MI_NOOP);
+	if (IS_ERR(rq)) {
+		igt_spinner_fini(&spin);
+		return PTR_ERR(rq);
+	}
+
+	err = request_add_spin(rq, &spin);
+	if (err == 0 && !using_guc)
+		err = intel_engine_reset(ce->engine, reason);
+
+	/* Ensure the reset happens and kills the engine */
+	if (err == 0)
+		err = intel_selftest_wait_for_rq(rq);
+
+	igt_spinner_end(&spin);
+	igt_spinner_fini(&spin);
+
+	return err;
+}
+
+static int __live_mocs_reset(struct live_mocs *mocs,
+			     struct intel_context *ce, bool using_guc)
+{
+	struct intel_gt *gt = ce->engine->gt;
+	int err;
+
+	if (intel_has_reset_engine(gt)) {
+		if (!using_guc) {
+			err = intel_engine_reset(ce->engine, "mocs");
+			if (err)
+				return err;
+
+			err = check_mocs_engine(mocs, ce);
+			if (err)
+				return err;
+		}
+
+		err = active_engine_reset(ce, "mocs", using_guc);
+		if (err)
+			return err;
+
+		err = check_mocs_engine(mocs, ce);
+		if (err)
+			return err;
+	}
+
+	if (intel_has_gpu_reset(gt)) {
+		intel_gt_reset(gt, ce->engine->mask, "mocs");
+
+		err = check_mocs_engine(mocs, ce);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int live_mocs_reset(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct live_mocs mocs;
+	int err = 0;
+
+	/* Check the mocs setup is retained over per-engine and global resets */
+
+	err = live_mocs_init(&mocs, gt);
+	if (err)
+		return err;
+
+	igt_global_reset_lock(gt);
+	for_each_engine(engine, gt, id) {
+		bool using_guc = intel_engine_uses_guc(engine);
+		struct intel_selftest_saved_policy saved;
+		struct intel_context *ce;
+		int err2;
+
+		err = intel_selftest_modify_policy(engine, &saved,
+						   SELFTEST_SCHEDULER_MODIFY_FAST_RESET);
+		if (err)
+			break;
+
+		ce = mocs_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto restore;
+		}
+
+		intel_engine_pm_get(engine);
+
+		err = __live_mocs_reset(&mocs, ce, using_guc);
+
+		intel_engine_pm_put(engine);
+		intel_context_put(ce);
+
+restore:
+		err2 = intel_selftest_restore_policy(engine, &saved);
+		if (err == 0)
+			err = err2;
+		if (err)
+			break;
+	}
+	igt_global_reset_unlock(gt);
+
+	live_mocs_fini(&mocs);
+	return err;
+}
+
+int intel_mocs_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_mocs_kernel),
+		SUBTEST(live_mocs_clean),
+		SUBTEST(live_mocs_reset),
+	};
+	struct drm_i915_mocs_table table;
+
+	if (!get_mocs_settings(i915, &table))
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_rc6.c b/drivers/gpu/drm/i915/gt/selftest_rc6.c
new file mode 100644
index 000000000..8c70b7e12
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_rc6.c
@@ -0,0 +1,254 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "intel_context.h"
+#include "intel_engine_pm.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt_requests.h"
+#include "intel_ring.h"
+#include "selftest_rc6.h"
+
+#include "selftests/i915_random.h"
+#include "selftests/librapl.h"
+
+static u64 rc6_residency(struct intel_rc6 *rc6)
+{
+	u64 result;
+
+	/* XXX VLV_GT_MEDIA_RC6? */
+
+	result = intel_rc6_residency_ns(rc6, GEN6_GT_GFX_RC6);
+	if (HAS_RC6p(rc6_to_i915(rc6)))
+		result += intel_rc6_residency_ns(rc6, GEN6_GT_GFX_RC6p);
+	if (HAS_RC6pp(rc6_to_i915(rc6)))
+		result += intel_rc6_residency_ns(rc6, GEN6_GT_GFX_RC6pp);
+
+	return result;
+}
+
+int live_rc6_manual(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_rc6 *rc6 = &gt->rc6;
+	u64 rc0_power, rc6_power;
+	intel_wakeref_t wakeref;
+	bool has_power;
+	ktime_t dt;
+	u64 res[2];
+	int err = 0;
+
+	/*
+	 * Our claim is that we can "encourage" the GPU to enter rc6 at will.
+	 * Let's try it!
+	 */
+
+	if (!rc6->enabled)
+		return 0;
+
+	/* bsw/byt use a PCU and decouple RC6 from our manual control */
+	if (IS_VALLEYVIEW(gt->i915) || IS_CHERRYVIEW(gt->i915))
+		return 0;
+
+	has_power = librapl_supported(gt->i915);
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+
+	/* Force RC6 off for starters */
+	__intel_rc6_disable(rc6);
+	msleep(1); /* wakeup is not immediate, takes about 100us on icl */
+
+	res[0] = rc6_residency(rc6);
+
+	dt = ktime_get();
+	rc0_power = librapl_energy_uJ();
+	msleep(250);
+	rc0_power = librapl_energy_uJ() - rc0_power;
+	dt = ktime_sub(ktime_get(), dt);
+	res[1] = rc6_residency(rc6);
+	if ((res[1] - res[0]) >> 10) {
+		pr_err("RC6 residency increased by %lldus while disabled for 250ms!\n",
+		       (res[1] - res[0]) >> 10);
+		err = -EINVAL;
+		goto out_unlock;
+	}
+
+	if (has_power) {
+		rc0_power = div64_u64(NSEC_PER_SEC * rc0_power,
+				      ktime_to_ns(dt));
+		if (!rc0_power) {
+			pr_err("No power measured while in RC0\n");
+			err = -EINVAL;
+			goto out_unlock;
+		}
+	}
+
+	/* Manually enter RC6 */
+	intel_rc6_park(rc6);
+
+	res[0] = rc6_residency(rc6);
+	intel_uncore_forcewake_flush(rc6_to_uncore(rc6), FORCEWAKE_ALL);
+	dt = ktime_get();
+	rc6_power = librapl_energy_uJ();
+	msleep(100);
+	rc6_power = librapl_energy_uJ() - rc6_power;
+	dt = ktime_sub(ktime_get(), dt);
+	res[1] = rc6_residency(rc6);
+	if (res[1] == res[0]) {
+		pr_err("Did not enter RC6! RC6_STATE=%08x, RC6_CONTROL=%08x, residency=%lld\n",
+		       intel_uncore_read_fw(gt->uncore, GEN6_RC_STATE),
+		       intel_uncore_read_fw(gt->uncore, GEN6_RC_CONTROL),
+		       res[0]);
+		err = -EINVAL;
+	}
+
+	if (has_power) {
+		rc6_power = div64_u64(NSEC_PER_SEC * rc6_power,
+				      ktime_to_ns(dt));
+		pr_info("GPU consumed %llduW in RC0 and %llduW in RC6\n",
+			rc0_power, rc6_power);
+		if (2 * rc6_power > rc0_power) {
+			pr_err("GPU leaked energy while in RC6!\n");
+			err = -EINVAL;
+			goto out_unlock;
+		}
+	}
+
+	/* Restore what should have been the original state! */
+	intel_rc6_unpark(rc6);
+
+out_unlock:
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+	return err;
+}
+
+static const u32 *__live_rc6_ctx(struct intel_context *ce)
+{
+	struct i915_request *rq;
+	const u32 *result;
+	u32 cmd;
+	u32 *cs;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return ERR_CAST(rq);
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return cs;
+	}
+
+	cmd = MI_STORE_REGISTER_MEM | MI_USE_GGTT;
+	if (GRAPHICS_VER(rq->engine->i915) >= 8)
+		cmd++;
+
+	*cs++ = cmd;
+	*cs++ = i915_mmio_reg_offset(GEN8_RC6_CTX_INFO);
+	*cs++ = ce->timeline->hwsp_offset + 8;
+	*cs++ = 0;
+	intel_ring_advance(rq, cs);
+
+	result = rq->hwsp_seqno + 2;
+	i915_request_add(rq);
+
+	return result;
+}
+
+static struct intel_engine_cs **
+randomised_engines(struct intel_gt *gt,
+		   struct rnd_state *prng,
+		   unsigned int *count)
+{
+	struct intel_engine_cs *engine, **engines;
+	enum intel_engine_id id;
+	int n;
+
+	n = 0;
+	for_each_engine(engine, gt, id)
+		n++;
+	if (!n)
+		return NULL;
+
+	engines = kmalloc_array(n, sizeof(*engines), GFP_KERNEL);
+	if (!engines)
+		return NULL;
+
+	n = 0;
+	for_each_engine(engine, gt, id)
+		engines[n++] = engine;
+
+	i915_prandom_shuffle(engines, sizeof(*engines), n, prng);
+
+	*count = n;
+	return engines;
+}
+
+int live_rc6_ctx_wa(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs **engines;
+	unsigned int n, count;
+	I915_RND_STATE(prng);
+	int err = 0;
+
+	/* A read of CTX_INFO upsets rc6. Poke the bear! */
+	if (GRAPHICS_VER(gt->i915) < 8)
+		return 0;
+
+	engines = randomised_engines(gt, &prng, &count);
+	if (!engines)
+		return 0;
+
+	for (n = 0; n < count; n++) {
+		struct intel_engine_cs *engine = engines[n];
+		int pass;
+
+		for (pass = 0; pass < 2; pass++) {
+			struct i915_gpu_error *error = &gt->i915->gpu_error;
+			struct intel_context *ce;
+			unsigned int resets =
+				i915_reset_engine_count(error, engine);
+			const u32 *res;
+
+			/* Use a sacrifical context */
+			ce = intel_context_create(engine);
+			if (IS_ERR(ce)) {
+				err = PTR_ERR(ce);
+				goto out;
+			}
+
+			intel_engine_pm_get(engine);
+			res = __live_rc6_ctx(ce);
+			intel_engine_pm_put(engine);
+			intel_context_put(ce);
+			if (IS_ERR(res)) {
+				err = PTR_ERR(res);
+				goto out;
+			}
+
+			if (intel_gt_wait_for_idle(gt, HZ / 5) == -ETIME) {
+				intel_gt_set_wedged(gt);
+				err = -ETIME;
+				goto out;
+			}
+
+			intel_gt_pm_wait_for_idle(gt);
+			pr_debug("%s: CTX_INFO=%0x\n",
+				 engine->name, READ_ONCE(*res));
+
+			if (resets !=
+			    i915_reset_engine_count(error, engine)) {
+				pr_err("%s: GPU reset required\n",
+				       engine->name);
+				add_taint_for_CI(gt->i915, TAINT_WARN);
+				err = -EIO;
+				goto out;
+			}
+		}
+	}
+
+out:
+	kfree(engines);
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_rc6.h b/drivers/gpu/drm/i915/gt/selftest_rc6.h
new file mode 100644
index 000000000..daf092790
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_rc6.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef SELFTEST_RC6_H
+#define SELFTEST_RC6_H
+
+int live_rc6_ctx_wa(void *arg);
+int live_rc6_manual(void *arg);
+
+#endif /* SELFTEST_RC6_H */
diff --git a/drivers/gpu/drm/i915/gt/selftest_reset.c b/drivers/gpu/drm/i915/gt/selftest_reset.c
new file mode 100644
index 000000000..37c38bdd5
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_reset.c
@@ -0,0 +1,388 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/crc32.h>
+
+#include "gem/i915_gem_stolen.h"
+
+#include "i915_memcpy.h"
+#include "i915_selftest.h"
+#include "intel_gpu_commands.h"
+#include "selftests/igt_reset.h"
+#include "selftests/igt_atomic.h"
+#include "selftests/igt_spinner.h"
+
+static int
+__igt_reset_stolen(struct intel_gt *gt,
+		   intel_engine_mask_t mask,
+		   const char *msg)
+{
+	struct i915_ggtt *ggtt = gt->ggtt;
+	const struct resource *dsm = &gt->i915->dsm;
+	resource_size_t num_pages, page;
+	struct intel_engine_cs *engine;
+	intel_wakeref_t wakeref;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	long max, count;
+	void *tmp;
+	u32 *crc;
+	int err;
+
+	if (!drm_mm_node_allocated(&ggtt->error_capture))
+		return 0;
+
+	num_pages = resource_size(dsm) >> PAGE_SHIFT;
+	if (!num_pages)
+		return 0;
+
+	crc = kmalloc_array(num_pages, sizeof(u32), GFP_KERNEL);
+	if (!crc)
+		return -ENOMEM;
+
+	tmp = kmalloc(PAGE_SIZE, GFP_KERNEL);
+	if (!tmp) {
+		err = -ENOMEM;
+		goto err_crc;
+	}
+
+	igt_global_reset_lock(gt);
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+
+	err = igt_spinner_init(&spin, gt);
+	if (err)
+		goto err_lock;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		struct i915_request *rq;
+
+		if (!(mask & engine->mask))
+			continue;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto err_spin;
+		}
+		rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+		intel_context_put(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_spin;
+		}
+		i915_request_add(rq);
+	}
+
+	for (page = 0; page < num_pages; page++) {
+		dma_addr_t dma = (dma_addr_t)dsm->start + (page << PAGE_SHIFT);
+		void __iomem *s;
+		void *in;
+
+		ggtt->vm.insert_page(&ggtt->vm, dma,
+				     ggtt->error_capture.start,
+				     I915_CACHE_NONE, 0);
+		mb();
+
+		s = io_mapping_map_wc(&ggtt->iomap,
+				      ggtt->error_capture.start,
+				      PAGE_SIZE);
+
+		if (!__drm_mm_interval_first(&gt->i915->mm.stolen,
+					     page << PAGE_SHIFT,
+					     ((page + 1) << PAGE_SHIFT) - 1))
+			memset_io(s, STACK_MAGIC, PAGE_SIZE);
+
+		in = (void __force *)s;
+		if (i915_memcpy_from_wc(tmp, in, PAGE_SIZE))
+			in = tmp;
+		crc[page] = crc32_le(0, in, PAGE_SIZE);
+
+		io_mapping_unmap(s);
+	}
+	mb();
+	ggtt->vm.clear_range(&ggtt->vm, ggtt->error_capture.start, PAGE_SIZE);
+
+	if (mask == ALL_ENGINES) {
+		intel_gt_reset(gt, mask, NULL);
+	} else {
+		for_each_engine(engine, gt, id) {
+			if (mask & engine->mask)
+				intel_engine_reset(engine, NULL);
+		}
+	}
+
+	max = -1;
+	count = 0;
+	for (page = 0; page < num_pages; page++) {
+		dma_addr_t dma = (dma_addr_t)dsm->start + (page << PAGE_SHIFT);
+		void __iomem *s;
+		void *in;
+		u32 x;
+
+		ggtt->vm.insert_page(&ggtt->vm, dma,
+				     ggtt->error_capture.start,
+				     I915_CACHE_NONE, 0);
+		mb();
+
+		s = io_mapping_map_wc(&ggtt->iomap,
+				      ggtt->error_capture.start,
+				      PAGE_SIZE);
+
+		in = (void __force *)s;
+		if (i915_memcpy_from_wc(tmp, in, PAGE_SIZE))
+			in = tmp;
+		x = crc32_le(0, in, PAGE_SIZE);
+
+		if (x != crc[page] &&
+		    !__drm_mm_interval_first(&gt->i915->mm.stolen,
+					     page << PAGE_SHIFT,
+					     ((page + 1) << PAGE_SHIFT) - 1)) {
+			pr_debug("unused stolen page %pa modified by GPU reset\n",
+				 &page);
+			if (count++ == 0)
+				igt_hexdump(in, PAGE_SIZE);
+			max = page;
+		}
+
+		io_mapping_unmap(s);
+	}
+	mb();
+	ggtt->vm.clear_range(&ggtt->vm, ggtt->error_capture.start, PAGE_SIZE);
+
+	if (count > 0) {
+		pr_info("%s reset clobbered %ld pages of stolen, last clobber at page %ld\n",
+			msg, count, max);
+	}
+	if (max >= I915_GEM_STOLEN_BIAS >> PAGE_SHIFT) {
+		pr_err("%s reset clobbered unreserved area [above %x] of stolen; may cause severe faults\n",
+		       msg, I915_GEM_STOLEN_BIAS);
+		err = -EINVAL;
+	}
+
+err_spin:
+	igt_spinner_fini(&spin);
+
+err_lock:
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+	igt_global_reset_unlock(gt);
+
+	kfree(tmp);
+err_crc:
+	kfree(crc);
+	return err;
+}
+
+static int igt_reset_device_stolen(void *arg)
+{
+	return __igt_reset_stolen(arg, ALL_ENGINES, "device");
+}
+
+static int igt_reset_engines_stolen(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err;
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		err = __igt_reset_stolen(gt, engine->mask, engine->name);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int igt_global_reset(void *arg)
+{
+	struct intel_gt *gt = arg;
+	unsigned int reset_count;
+	intel_wakeref_t wakeref;
+	int err = 0;
+
+	/* Check that we can issue a global GPU reset */
+
+	igt_global_reset_lock(gt);
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+
+	reset_count = i915_reset_count(&gt->i915->gpu_error);
+
+	intel_gt_reset(gt, ALL_ENGINES, NULL);
+
+	if (i915_reset_count(&gt->i915->gpu_error) == reset_count) {
+		pr_err("No GPU reset recorded!\n");
+		err = -EINVAL;
+	}
+
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+	igt_global_reset_unlock(gt);
+
+	if (intel_gt_is_wedged(gt))
+		err = -EIO;
+
+	return err;
+}
+
+static int igt_wedged_reset(void *arg)
+{
+	struct intel_gt *gt = arg;
+	intel_wakeref_t wakeref;
+
+	/* Check that we can recover a wedged device with a GPU reset */
+
+	igt_global_reset_lock(gt);
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+
+	intel_gt_set_wedged(gt);
+
+	GEM_BUG_ON(!intel_gt_is_wedged(gt));
+	intel_gt_reset(gt, ALL_ENGINES, NULL);
+
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+	igt_global_reset_unlock(gt);
+
+	return intel_gt_is_wedged(gt) ? -EIO : 0;
+}
+
+static int igt_atomic_reset(void *arg)
+{
+	struct intel_gt *gt = arg;
+	const typeof(*igt_atomic_phases) *p;
+	int err = 0;
+
+	/* Check that the resets are usable from atomic context */
+
+	intel_gt_pm_get(gt);
+	igt_global_reset_lock(gt);
+
+	/* Flush any requests before we get started and check basics */
+	if (!igt_force_reset(gt))
+		goto unlock;
+
+	for (p = igt_atomic_phases; p->name; p++) {
+		intel_engine_mask_t awake;
+
+		GEM_TRACE("__intel_gt_reset under %s\n", p->name);
+
+		awake = reset_prepare(gt);
+		p->critical_section_begin();
+
+		err = __intel_gt_reset(gt, ALL_ENGINES);
+
+		p->critical_section_end();
+		reset_finish(gt, awake);
+
+		if (err) {
+			pr_err("__intel_gt_reset failed under %s\n", p->name);
+			break;
+		}
+	}
+
+	/* As we poke around the guts, do a full reset before continuing. */
+	igt_force_reset(gt);
+
+unlock:
+	igt_global_reset_unlock(gt);
+	intel_gt_pm_put(gt);
+
+	return err;
+}
+
+static int igt_atomic_engine_reset(void *arg)
+{
+	struct intel_gt *gt = arg;
+	const typeof(*igt_atomic_phases) *p;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/* Check that the resets are usable from atomic context */
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	if (intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	intel_gt_pm_get(gt);
+	igt_global_reset_lock(gt);
+
+	/* Flush any requests before we get started and check basics */
+	if (!igt_force_reset(gt))
+		goto out_unlock;
+
+	for_each_engine(engine, gt, id) {
+		struct tasklet_struct *t = &engine->sched_engine->tasklet;
+
+		if (t->func)
+			tasklet_disable(t);
+		intel_engine_pm_get(engine);
+
+		for (p = igt_atomic_phases; p->name; p++) {
+			GEM_TRACE("intel_engine_reset(%s) under %s\n",
+				  engine->name, p->name);
+			if (strcmp(p->name, "softirq"))
+				local_bh_disable();
+
+			p->critical_section_begin();
+			err = __intel_engine_reset_bh(engine, NULL);
+			p->critical_section_end();
+
+			if (strcmp(p->name, "softirq"))
+				local_bh_enable();
+
+			if (err) {
+				pr_err("intel_engine_reset(%s) failed under %s\n",
+				       engine->name, p->name);
+				break;
+			}
+		}
+
+		intel_engine_pm_put(engine);
+		if (t->func) {
+			tasklet_enable(t);
+			tasklet_hi_schedule(t);
+		}
+		if (err)
+			break;
+	}
+
+	/* As we poke around the guts, do a full reset before continuing. */
+	igt_force_reset(gt);
+
+out_unlock:
+	igt_global_reset_unlock(gt);
+	intel_gt_pm_put(gt);
+
+	return err;
+}
+
+int intel_reset_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_global_reset), /* attempt to recover GPU first */
+		SUBTEST(igt_reset_device_stolen),
+		SUBTEST(igt_reset_engines_stolen),
+		SUBTEST(igt_wedged_reset),
+		SUBTEST(igt_atomic_reset),
+		SUBTEST(igt_atomic_engine_reset),
+	};
+	struct intel_gt *gt = to_gt(i915);
+
+	if (!intel_has_gpu_reset(gt))
+		return 0;
+
+	if (intel_gt_is_wedged(gt))
+		return -EIO; /* we're long past hope of a successful reset */
+
+	return intel_gt_live_subtests(tests, gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_ring.c b/drivers/gpu/drm/i915/gt/selftest_ring.c
new file mode 100644
index 000000000..2a8c534dc
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_ring.c
@@ -0,0 +1,110 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+static struct intel_ring *mock_ring(unsigned long sz)
+{
+	struct intel_ring *ring;
+
+	ring = kzalloc(sizeof(*ring) + sz, GFP_KERNEL);
+	if (!ring)
+		return NULL;
+
+	kref_init(&ring->ref);
+	ring->size = sz;
+	ring->wrap = BITS_PER_TYPE(ring->size) - ilog2(sz);
+	ring->effective_size = sz;
+	ring->vaddr = (void *)(ring + 1);
+	atomic_set(&ring->pin_count, 1);
+
+	intel_ring_update_space(ring);
+
+	return ring;
+}
+
+static void mock_ring_free(struct intel_ring *ring)
+{
+	kfree(ring);
+}
+
+static int check_ring_direction(struct intel_ring *ring,
+				u32 next, u32 prev,
+				int expected)
+{
+	int result;
+
+	result = intel_ring_direction(ring, next, prev);
+	if (result < 0)
+		result = -1;
+	else if (result > 0)
+		result = 1;
+
+	if (result != expected) {
+		pr_err("intel_ring_direction(%u, %u):%d != %d\n",
+		       next, prev, result, expected);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int check_ring_step(struct intel_ring *ring, u32 x, u32 step)
+{
+	u32 prev = x, next = intel_ring_wrap(ring, x + step);
+	int err = 0;
+
+	err |= check_ring_direction(ring, next, next,  0);
+	err |= check_ring_direction(ring, prev, prev,  0);
+	err |= check_ring_direction(ring, next, prev,  1);
+	err |= check_ring_direction(ring, prev, next, -1);
+
+	return err;
+}
+
+static int check_ring_offset(struct intel_ring *ring, u32 x, u32 step)
+{
+	int err = 0;
+
+	err |= check_ring_step(ring, x, step);
+	err |= check_ring_step(ring, intel_ring_wrap(ring, x + 1), step);
+	err |= check_ring_step(ring, intel_ring_wrap(ring, x - 1), step);
+
+	return err;
+}
+
+static int igt_ring_direction(void *dummy)
+{
+	struct intel_ring *ring;
+	unsigned int half = 2048;
+	int step, err = 0;
+
+	ring = mock_ring(2 * half);
+	if (!ring)
+		return -ENOMEM;
+
+	GEM_BUG_ON(ring->size != 2 * half);
+
+	/* Precision of wrap detection is limited to ring->size / 2 */
+	for (step = 1; step < half; step <<= 1) {
+		err |= check_ring_offset(ring, 0, step);
+		err |= check_ring_offset(ring, half, step);
+	}
+	err |= check_ring_step(ring, 0, half - 64);
+
+	/* And check unwrapped handling for good measure */
+	err |= check_ring_offset(ring, 0, 2 * half + 64);
+	err |= check_ring_offset(ring, 3 * half, 1);
+
+	mock_ring_free(ring);
+	return err;
+}
+
+int intel_ring_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_ring_direction),
+	};
+
+	return i915_subtests(tests, NULL);
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_ring_submission.c b/drivers/gpu/drm/i915/gt/selftest_ring_submission.c
new file mode 100644
index 000000000..70f9ac1ec
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_ring_submission.c
@@ -0,0 +1,298 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "intel_engine_pm.h"
+#include "selftests/igt_flush_test.h"
+
+static struct i915_vma *create_wally(struct intel_engine_cs *engine)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u32 *cs;
+	int err;
+
+	obj = i915_gem_object_create_internal(engine->i915, 4096);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, engine->gt->vm, NULL);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return vma;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER | PIN_HIGH);
+	if (err) {
+		i915_gem_object_put(obj);
+		return ERR_PTR(err);
+	}
+
+	err = i915_vma_sync(vma);
+	if (err) {
+		i915_gem_object_put(obj);
+		return ERR_PTR(err);
+	}
+
+	cs = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(cs)) {
+		i915_gem_object_put(obj);
+		return ERR_CAST(cs);
+	}
+
+	if (GRAPHICS_VER(engine->i915) >= 6) {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4;
+		*cs++ = 0;
+	} else if (GRAPHICS_VER(engine->i915) >= 4) {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = 0;
+	} else {
+		*cs++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL;
+	}
+	*cs++ = vma->node.start + 4000;
+	*cs++ = STACK_MAGIC;
+
+	*cs++ = MI_BATCH_BUFFER_END;
+
+	i915_gem_object_flush_map(obj);
+	i915_gem_object_unpin_map(obj);
+
+	vma->private = intel_context_create(engine); /* dummy residuals */
+	if (IS_ERR(vma->private)) {
+		vma = ERR_CAST(vma->private);
+		i915_gem_object_put(obj);
+	}
+
+	return vma;
+}
+
+static int context_sync(struct intel_context *ce)
+{
+	struct i915_request *rq;
+	int err = 0;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	if (i915_request_wait(rq, 0, HZ / 5) < 0)
+		err = -ETIME;
+	i915_request_put(rq);
+
+	return err;
+}
+
+static int new_context_sync(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+	int err;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	err = context_sync(ce);
+	intel_context_put(ce);
+
+	return err;
+}
+
+static int mixed_contexts_sync(struct intel_engine_cs *engine, u32 *result)
+{
+	int pass;
+	int err;
+
+	for (pass = 0; pass < 2; pass++) {
+		WRITE_ONCE(*result, 0);
+		err = context_sync(engine->kernel_context);
+		if (err || READ_ONCE(*result)) {
+			if (!err) {
+				pr_err("pass[%d] wa_bb emitted for the kernel context\n",
+				       pass);
+				err = -EINVAL;
+			}
+			return err;
+		}
+
+		WRITE_ONCE(*result, 0);
+		err = new_context_sync(engine);
+		if (READ_ONCE(*result) != STACK_MAGIC) {
+			if (!err) {
+				pr_err("pass[%d] wa_bb *NOT* emitted after the kernel context\n",
+				       pass);
+				err = -EINVAL;
+			}
+			return err;
+		}
+
+		WRITE_ONCE(*result, 0);
+		err = new_context_sync(engine);
+		if (READ_ONCE(*result) != STACK_MAGIC) {
+			if (!err) {
+				pr_err("pass[%d] wa_bb *NOT* emitted for the user context switch\n",
+				       pass);
+				err = -EINVAL;
+			}
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+static int double_context_sync_00(struct intel_engine_cs *engine, u32 *result)
+{
+	struct intel_context *ce;
+	int err, i;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	for (i = 0; i < 2; i++) {
+		WRITE_ONCE(*result, 0);
+		err = context_sync(ce);
+		if (err)
+			break;
+	}
+	intel_context_put(ce);
+	if (err)
+		return err;
+
+	if (READ_ONCE(*result)) {
+		pr_err("wa_bb emitted between the same user context\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int kernel_context_sync_00(struct intel_engine_cs *engine, u32 *result)
+{
+	struct intel_context *ce;
+	int err, i;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	for (i = 0; i < 2; i++) {
+		WRITE_ONCE(*result, 0);
+		err = context_sync(ce);
+		if (err)
+			break;
+
+		err = context_sync(engine->kernel_context);
+		if (err)
+			break;
+	}
+	intel_context_put(ce);
+	if (err)
+		return err;
+
+	if (READ_ONCE(*result)) {
+		pr_err("wa_bb emitted between the same user context [with intervening kernel]\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int __live_ctx_switch_wa(struct intel_engine_cs *engine)
+{
+	struct i915_vma *bb;
+	u32 *result;
+	int err;
+
+	bb = create_wally(engine);
+	if (IS_ERR(bb))
+		return PTR_ERR(bb);
+
+	result = i915_gem_object_pin_map_unlocked(bb->obj, I915_MAP_WC);
+	if (IS_ERR(result)) {
+		intel_context_put(bb->private);
+		i915_vma_unpin_and_release(&bb, 0);
+		return PTR_ERR(result);
+	}
+	result += 1000;
+
+	engine->wa_ctx.vma = bb;
+
+	err = mixed_contexts_sync(engine, result);
+	if (err)
+		goto out;
+
+	err = double_context_sync_00(engine, result);
+	if (err)
+		goto out;
+
+	err = kernel_context_sync_00(engine, result);
+	if (err)
+		goto out;
+
+out:
+	intel_context_put(engine->wa_ctx.vma->private);
+	i915_vma_unpin_and_release(&engine->wa_ctx.vma, I915_VMA_RELEASE_MAP);
+	return err;
+}
+
+static int live_ctx_switch_wa(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * Exercise the inter-context wa batch.
+	 *
+	 * Between each user context we run a wa batch, and since it may
+	 * have implications for user visible state, we have to check that
+	 * we do actually execute it.
+	 *
+	 * The trick we use is to replace the normal wa batch with a custom
+	 * one that writes to a marker within it, and we can then look for
+	 * that marker to confirm if the batch was run when we expect it,
+	 * and equally important it was wasn't run when we don't!
+	 */
+
+	for_each_engine(engine, gt, id) {
+		struct i915_vma *saved_wa;
+		int err;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		if (IS_GRAPHICS_VER(gt->i915, 4, 5))
+			continue; /* MI_STORE_DWORD is privileged! */
+
+		saved_wa = fetch_and_zero(&engine->wa_ctx.vma);
+
+		intel_engine_pm_get(engine);
+		err = __live_ctx_switch_wa(engine);
+		intel_engine_pm_put(engine);
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+
+		engine->wa_ctx.vma = saved_wa;
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int intel_ring_submission_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_ctx_switch_wa),
+	};
+
+	if (to_gt(i915)->submission_method > INTEL_SUBMISSION_RING)
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_rps.c b/drivers/gpu/drm/i915/gt/selftest_rps.c
new file mode 100644
index 000000000..cfb4708dd
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_rps.c
@@ -0,0 +1,1319 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/pm_qos.h>
+#include <linux/sort.h>
+
+#include "gem/i915_gem_internal.h"
+
+#include "intel_engine_heartbeat.h"
+#include "intel_engine_pm.h"
+#include "intel_engine_regs.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt_clock_utils.h"
+#include "intel_gt_pm.h"
+#include "intel_rc6.h"
+#include "selftest_engine_heartbeat.h"
+#include "selftest_rps.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/igt_spinner.h"
+#include "selftests/librapl.h"
+
+/* Try to isolate the impact of cstates from determing frequency response */
+#define CPU_LATENCY 0 /* -1 to disable pm_qos, 0 to disable cstates */
+
+static void dummy_rps_work(struct work_struct *wrk)
+{
+}
+
+static int cmp_u64(const void *A, const void *B)
+{
+	const u64 *a = A, *b = B;
+
+	if (*a < *b)
+		return -1;
+	else if (*a > *b)
+		return 1;
+	else
+		return 0;
+}
+
+static int cmp_u32(const void *A, const void *B)
+{
+	const u32 *a = A, *b = B;
+
+	if (*a < *b)
+		return -1;
+	else if (*a > *b)
+		return 1;
+	else
+		return 0;
+}
+
+static struct i915_vma *
+create_spin_counter(struct intel_engine_cs *engine,
+		    struct i915_address_space *vm,
+		    bool srm,
+		    u32 **cancel,
+		    u32 **counter)
+{
+	enum {
+		COUNT,
+		INC,
+		__NGPR__,
+	};
+#define CS_GPR(x) GEN8_RING_CS_GPR(engine->mmio_base, x)
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	unsigned long end;
+	u32 *base, *cs;
+	int loop, i;
+	int err;
+
+	obj = i915_gem_object_create_internal(vm->i915, 64 << 10);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	end = obj->base.size / sizeof(u32) - 1;
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_put;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto err_unlock;
+
+	i915_vma_lock(vma);
+
+	base = i915_gem_object_pin_map(obj, I915_MAP_WC);
+	if (IS_ERR(base)) {
+		err = PTR_ERR(base);
+		goto err_unpin;
+	}
+	cs = base;
+
+	*cs++ = MI_LOAD_REGISTER_IMM(__NGPR__ * 2);
+	for (i = 0; i < __NGPR__; i++) {
+		*cs++ = i915_mmio_reg_offset(CS_GPR(i));
+		*cs++ = 0;
+		*cs++ = i915_mmio_reg_offset(CS_GPR(i)) + 4;
+		*cs++ = 0;
+	}
+
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(CS_GPR(INC));
+	*cs++ = 1;
+
+	loop = cs - base;
+
+	/* Unroll the loop to avoid MI_BB_START stalls impacting measurements */
+	for (i = 0; i < 1024; i++) {
+		*cs++ = MI_MATH(4);
+		*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCA, MI_MATH_REG(COUNT));
+		*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCB, MI_MATH_REG(INC));
+		*cs++ = MI_MATH_ADD;
+		*cs++ = MI_MATH_STORE(MI_MATH_REG(COUNT), MI_MATH_REG_ACCU);
+
+		if (srm) {
+			*cs++ = MI_STORE_REGISTER_MEM_GEN8;
+			*cs++ = i915_mmio_reg_offset(CS_GPR(COUNT));
+			*cs++ = lower_32_bits(vma->node.start + end * sizeof(*cs));
+			*cs++ = upper_32_bits(vma->node.start + end * sizeof(*cs));
+		}
+	}
+
+	*cs++ = MI_BATCH_BUFFER_START_GEN8;
+	*cs++ = lower_32_bits(vma->node.start + loop * sizeof(*cs));
+	*cs++ = upper_32_bits(vma->node.start + loop * sizeof(*cs));
+	GEM_BUG_ON(cs - base > end);
+
+	i915_gem_object_flush_map(obj);
+
+	*cancel = base + loop;
+	*counter = srm ? memset32(base + end, 0, 1) : NULL;
+	return vma;
+
+err_unpin:
+	i915_vma_unpin(vma);
+err_unlock:
+	i915_vma_unlock(vma);
+err_put:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static u8 wait_for_freq(struct intel_rps *rps, u8 freq, int timeout_ms)
+{
+	u8 history[64], i;
+	unsigned long end;
+	int sleep;
+
+	i = 0;
+	memset(history, freq, sizeof(history));
+	sleep = 20;
+
+	/* The PCU does not change instantly, but drifts towards the goal? */
+	end = jiffies + msecs_to_jiffies(timeout_ms);
+	do {
+		u8 act;
+
+		act = read_cagf(rps);
+		if (time_after(jiffies, end))
+			return act;
+
+		/* Target acquired */
+		if (act == freq)
+			return act;
+
+		/* Any change within the last N samples? */
+		if (!memchr_inv(history, act, sizeof(history)))
+			return act;
+
+		history[i] = act;
+		i = (i + 1) % ARRAY_SIZE(history);
+
+		usleep_range(sleep, 2 * sleep);
+		sleep *= 2;
+		if (sleep > timeout_ms * 20)
+			sleep = timeout_ms * 20;
+	} while (1);
+}
+
+static u8 rps_set_check(struct intel_rps *rps, u8 freq)
+{
+	mutex_lock(&rps->lock);
+	GEM_BUG_ON(!intel_rps_is_active(rps));
+	if (wait_for(!intel_rps_set(rps, freq), 50)) {
+		mutex_unlock(&rps->lock);
+		return 0;
+	}
+	GEM_BUG_ON(rps->last_freq != freq);
+	mutex_unlock(&rps->lock);
+
+	return wait_for_freq(rps, freq, 50);
+}
+
+static void show_pstate_limits(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+
+	if (IS_BROXTON(i915)) {
+		pr_info("P_STATE_CAP[%x]: 0x%08x\n",
+			i915_mmio_reg_offset(BXT_RP_STATE_CAP),
+			intel_uncore_read(rps_to_uncore(rps),
+					  BXT_RP_STATE_CAP));
+	} else if (GRAPHICS_VER(i915) == 9) {
+		pr_info("P_STATE_LIMITS[%x]: 0x%08x\n",
+			i915_mmio_reg_offset(GEN9_RP_STATE_LIMITS),
+			intel_uncore_read(rps_to_uncore(rps),
+					  GEN9_RP_STATE_LIMITS));
+	}
+}
+
+int live_rps_clock_interval(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_rps *rps = &gt->rps;
+	void (*saved_work)(struct work_struct *wrk);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	if (!intel_rps_is_enabled(rps) || GRAPHICS_VER(gt->i915) < 6)
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	intel_gt_pm_wait_for_idle(gt);
+	saved_work = rps->work.func;
+	rps->work.func = dummy_rps_work;
+
+	intel_gt_pm_get(gt);
+	intel_rps_disable(&gt->rps);
+
+	intel_gt_check_clock_frequency(gt);
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq;
+		u32 cycles;
+		u64 dt;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		st_engine_heartbeat_disable(engine);
+
+		rq = igt_spinner_create_request(&spin,
+						engine->kernel_context,
+						MI_NOOP);
+		if (IS_ERR(rq)) {
+			st_engine_heartbeat_enable(engine);
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_add(rq);
+
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			pr_err("%s: RPS spinner did not start\n",
+			       engine->name);
+			igt_spinner_end(&spin);
+			st_engine_heartbeat_enable(engine);
+			intel_gt_set_wedged(engine->gt);
+			err = -EIO;
+			break;
+		}
+
+		intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);
+
+		intel_uncore_write_fw(gt->uncore, GEN6_RP_CUR_UP_EI, 0);
+
+		/* Set the evaluation interval to infinity! */
+		intel_uncore_write_fw(gt->uncore,
+				      GEN6_RP_UP_EI, 0xffffffff);
+		intel_uncore_write_fw(gt->uncore,
+				      GEN6_RP_UP_THRESHOLD, 0xffffffff);
+
+		intel_uncore_write_fw(gt->uncore, GEN6_RP_CONTROL,
+				      GEN6_RP_ENABLE | GEN6_RP_UP_BUSY_AVG);
+
+		if (wait_for(intel_uncore_read_fw(gt->uncore,
+						  GEN6_RP_CUR_UP_EI),
+			     10)) {
+			/* Just skip the test; assume lack of HW support */
+			pr_notice("%s: rps evaluation interval not ticking\n",
+				  engine->name);
+			err = -ENODEV;
+		} else {
+			ktime_t dt_[5];
+			u32 cycles_[5];
+			int i;
+
+			for (i = 0; i < 5; i++) {
+				preempt_disable();
+
+				dt_[i] = ktime_get();
+				cycles_[i] = -intel_uncore_read_fw(gt->uncore, GEN6_RP_CUR_UP_EI);
+
+				udelay(1000);
+
+				dt_[i] = ktime_sub(ktime_get(), dt_[i]);
+				cycles_[i] += intel_uncore_read_fw(gt->uncore, GEN6_RP_CUR_UP_EI);
+
+				preempt_enable();
+			}
+
+			/* Use the median of both cycle/dt; close enough */
+			sort(cycles_, 5, sizeof(*cycles_), cmp_u32, NULL);
+			cycles = (cycles_[1] + 2 * cycles_[2] + cycles_[3]) / 4;
+			sort(dt_, 5, sizeof(*dt_), cmp_u64, NULL);
+			dt = div_u64(dt_[1] + 2 * dt_[2] + dt_[3], 4);
+		}
+
+		intel_uncore_write_fw(gt->uncore, GEN6_RP_CONTROL, 0);
+		intel_uncore_forcewake_put(gt->uncore, FORCEWAKE_ALL);
+
+		igt_spinner_end(&spin);
+		st_engine_heartbeat_enable(engine);
+
+		if (err == 0) {
+			u64 time = intel_gt_pm_interval_to_ns(gt, cycles);
+			u32 expected =
+				intel_gt_ns_to_pm_interval(gt, dt);
+
+			pr_info("%s: rps counted %d C0 cycles [%lldns] in %lldns [%d cycles], using GT clock frequency of %uKHz\n",
+				engine->name, cycles, time, dt, expected,
+				gt->clock_frequency / 1000);
+
+			if (10 * time < 8 * dt ||
+			    8 * time > 10 * dt) {
+				pr_err("%s: rps clock time does not match walltime!\n",
+				       engine->name);
+				err = -EINVAL;
+			}
+
+			if (10 * expected < 8 * cycles ||
+			    8 * expected > 10 * cycles) {
+				pr_err("%s: walltime does not match rps clock ticks!\n",
+				       engine->name);
+				err = -EINVAL;
+			}
+		}
+
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+
+		break; /* once is enough */
+	}
+
+	intel_rps_enable(&gt->rps);
+	intel_gt_pm_put(gt);
+
+	igt_spinner_fini(&spin);
+
+	intel_gt_pm_wait_for_idle(gt);
+	rps->work.func = saved_work;
+
+	if (err == -ENODEV) /* skipped, don't report a fail */
+		err = 0;
+
+	return err;
+}
+
+int live_rps_control(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_rps *rps = &gt->rps;
+	void (*saved_work)(struct work_struct *wrk);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	/*
+	 * Check that the actual frequency matches our requested frequency,
+	 * to verify our control mechanism. We have to be careful that the
+	 * PCU may throttle the GPU in which case the actual frequency used
+	 * will be lowered than requested.
+	 */
+
+	if (!intel_rps_is_enabled(rps))
+		return 0;
+
+	if (IS_CHERRYVIEW(gt->i915)) /* XXX fragile PCU */
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	intel_gt_pm_wait_for_idle(gt);
+	saved_work = rps->work.func;
+	rps->work.func = dummy_rps_work;
+
+	intel_gt_pm_get(gt);
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq;
+		ktime_t min_dt, max_dt;
+		int f, limit;
+		int min, max;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		st_engine_heartbeat_disable(engine);
+
+		rq = igt_spinner_create_request(&spin,
+						engine->kernel_context,
+						MI_NOOP);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_add(rq);
+
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			pr_err("%s: RPS spinner did not start\n",
+			       engine->name);
+			igt_spinner_end(&spin);
+			st_engine_heartbeat_enable(engine);
+			intel_gt_set_wedged(engine->gt);
+			err = -EIO;
+			break;
+		}
+
+		if (rps_set_check(rps, rps->min_freq) != rps->min_freq) {
+			pr_err("%s: could not set minimum frequency [%x], only %x!\n",
+			       engine->name, rps->min_freq, read_cagf(rps));
+			igt_spinner_end(&spin);
+			st_engine_heartbeat_enable(engine);
+			show_pstate_limits(rps);
+			err = -EINVAL;
+			break;
+		}
+
+		for (f = rps->min_freq + 1; f < rps->max_freq; f++) {
+			if (rps_set_check(rps, f) < f)
+				break;
+		}
+
+		limit = rps_set_check(rps, f);
+
+		if (rps_set_check(rps, rps->min_freq) != rps->min_freq) {
+			pr_err("%s: could not restore minimum frequency [%x], only %x!\n",
+			       engine->name, rps->min_freq, read_cagf(rps));
+			igt_spinner_end(&spin);
+			st_engine_heartbeat_enable(engine);
+			show_pstate_limits(rps);
+			err = -EINVAL;
+			break;
+		}
+
+		max_dt = ktime_get();
+		max = rps_set_check(rps, limit);
+		max_dt = ktime_sub(ktime_get(), max_dt);
+
+		min_dt = ktime_get();
+		min = rps_set_check(rps, rps->min_freq);
+		min_dt = ktime_sub(ktime_get(), min_dt);
+
+		igt_spinner_end(&spin);
+		st_engine_heartbeat_enable(engine);
+
+		pr_info("%s: range:[%x:%uMHz, %x:%uMHz] limit:[%x:%uMHz], %x:%x response %lluns:%lluns\n",
+			engine->name,
+			rps->min_freq, intel_gpu_freq(rps, rps->min_freq),
+			rps->max_freq, intel_gpu_freq(rps, rps->max_freq),
+			limit, intel_gpu_freq(rps, limit),
+			min, max, ktime_to_ns(min_dt), ktime_to_ns(max_dt));
+
+		if (limit == rps->min_freq) {
+			pr_err("%s: GPU throttled to minimum!\n",
+			       engine->name);
+			show_pstate_limits(rps);
+			err = -ENODEV;
+			break;
+		}
+
+		if (igt_flush_test(gt->i915)) {
+			err = -EIO;
+			break;
+		}
+	}
+	intel_gt_pm_put(gt);
+
+	igt_spinner_fini(&spin);
+
+	intel_gt_pm_wait_for_idle(gt);
+	rps->work.func = saved_work;
+
+	return err;
+}
+
+static void show_pcu_config(struct intel_rps *rps)
+{
+	struct drm_i915_private *i915 = rps_to_i915(rps);
+	unsigned int max_gpu_freq, min_gpu_freq;
+	intel_wakeref_t wakeref;
+	int gpu_freq;
+
+	if (!HAS_LLC(i915))
+		return;
+
+	min_gpu_freq = rps->min_freq;
+	max_gpu_freq = rps->max_freq;
+	if (GRAPHICS_VER(i915) >= 9) {
+		/* Convert GT frequency to 50 HZ units */
+		min_gpu_freq /= GEN9_FREQ_SCALER;
+		max_gpu_freq /= GEN9_FREQ_SCALER;
+	}
+
+	wakeref = intel_runtime_pm_get(rps_to_uncore(rps)->rpm);
+
+	pr_info("%5s  %5s  %5s\n", "GPU", "eCPU", "eRing");
+	for (gpu_freq = min_gpu_freq; gpu_freq <= max_gpu_freq; gpu_freq++) {
+		int ia_freq = gpu_freq;
+
+		snb_pcode_read(rps_to_gt(rps)->uncore, GEN6_PCODE_READ_MIN_FREQ_TABLE,
+			       &ia_freq, NULL);
+
+		pr_info("%5d  %5d  %5d\n",
+			gpu_freq * 50,
+			((ia_freq >> 0) & 0xff) * 100,
+			((ia_freq >> 8) & 0xff) * 100);
+	}
+
+	intel_runtime_pm_put(rps_to_uncore(rps)->rpm, wakeref);
+}
+
+static u64 __measure_frequency(u32 *cntr, int duration_ms)
+{
+	u64 dc, dt;
+
+	dt = ktime_get();
+	dc = READ_ONCE(*cntr);
+	usleep_range(1000 * duration_ms, 2000 * duration_ms);
+	dc = READ_ONCE(*cntr) - dc;
+	dt = ktime_get() - dt;
+
+	return div64_u64(1000 * 1000 * dc, dt);
+}
+
+static u64 measure_frequency_at(struct intel_rps *rps, u32 *cntr, int *freq)
+{
+	u64 x[5];
+	int i;
+
+	*freq = rps_set_check(rps, *freq);
+	for (i = 0; i < 5; i++)
+		x[i] = __measure_frequency(cntr, 2);
+	*freq = (*freq + read_cagf(rps)) / 2;
+
+	/* A simple triangle filter for better result stability */
+	sort(x, 5, sizeof(*x), cmp_u64, NULL);
+	return div_u64(x[1] + 2 * x[2] + x[3], 4);
+}
+
+static u64 __measure_cs_frequency(struct intel_engine_cs *engine,
+				  int duration_ms)
+{
+	u64 dc, dt;
+
+	dt = ktime_get();
+	dc = intel_uncore_read_fw(engine->uncore, CS_GPR(0));
+	usleep_range(1000 * duration_ms, 2000 * duration_ms);
+	dc = intel_uncore_read_fw(engine->uncore, CS_GPR(0)) - dc;
+	dt = ktime_get() - dt;
+
+	return div64_u64(1000 * 1000 * dc, dt);
+}
+
+static u64 measure_cs_frequency_at(struct intel_rps *rps,
+				   struct intel_engine_cs *engine,
+				   int *freq)
+{
+	u64 x[5];
+	int i;
+
+	*freq = rps_set_check(rps, *freq);
+	for (i = 0; i < 5; i++)
+		x[i] = __measure_cs_frequency(engine, 2);
+	*freq = (*freq + read_cagf(rps)) / 2;
+
+	/* A simple triangle filter for better result stability */
+	sort(x, 5, sizeof(*x), cmp_u64, NULL);
+	return div_u64(x[1] + 2 * x[2] + x[3], 4);
+}
+
+static bool scaled_within(u64 x, u64 y, u32 f_n, u32 f_d)
+{
+	return f_d * x > f_n * y && f_n * x < f_d * y;
+}
+
+int live_rps_frequency_cs(void *arg)
+{
+	void (*saved_work)(struct work_struct *wrk);
+	struct intel_gt *gt = arg;
+	struct intel_rps *rps = &gt->rps;
+	struct intel_engine_cs *engine;
+	struct pm_qos_request qos;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * The premise is that the GPU does change frequency at our behest.
+	 * Let's check there is a correspondence between the requested
+	 * frequency, the actual frequency, and the observed clock rate.
+	 */
+
+	if (!intel_rps_is_enabled(rps))
+		return 0;
+
+	if (GRAPHICS_VER(gt->i915) < 8) /* for CS simplicity */
+		return 0;
+
+	if (CPU_LATENCY >= 0)
+		cpu_latency_qos_add_request(&qos, CPU_LATENCY);
+
+	intel_gt_pm_wait_for_idle(gt);
+	saved_work = rps->work.func;
+	rps->work.func = dummy_rps_work;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq;
+		struct i915_vma *vma;
+		u32 *cancel, *cntr;
+		struct {
+			u64 count;
+			int freq;
+		} min, max;
+
+		st_engine_heartbeat_disable(engine);
+
+		vma = create_spin_counter(engine,
+					  engine->kernel_context->vm, false,
+					  &cancel, &cntr);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			st_engine_heartbeat_enable(engine);
+			break;
+		}
+
+		rq = intel_engine_create_kernel_request(engine);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_vma;
+		}
+
+		err = i915_request_await_object(rq, vma->obj, false);
+		if (!err)
+			err = i915_vma_move_to_active(vma, rq, 0);
+		if (!err)
+			err = rq->engine->emit_bb_start(rq,
+							vma->node.start,
+							PAGE_SIZE, 0);
+		i915_request_add(rq);
+		if (err)
+			goto err_vma;
+
+		if (wait_for(intel_uncore_read(engine->uncore, CS_GPR(0)),
+			     10)) {
+			pr_err("%s: timed loop did not start\n",
+			       engine->name);
+			goto err_vma;
+		}
+
+		min.freq = rps->min_freq;
+		min.count = measure_cs_frequency_at(rps, engine, &min.freq);
+
+		max.freq = rps->max_freq;
+		max.count = measure_cs_frequency_at(rps, engine, &max.freq);
+
+		pr_info("%s: min:%lluKHz @ %uMHz, max:%lluKHz @ %uMHz [%d%%]\n",
+			engine->name,
+			min.count, intel_gpu_freq(rps, min.freq),
+			max.count, intel_gpu_freq(rps, max.freq),
+			(int)DIV64_U64_ROUND_CLOSEST(100 * min.freq * max.count,
+						     max.freq * min.count));
+
+		if (!scaled_within(max.freq * min.count,
+				   min.freq * max.count,
+				   2, 3)) {
+			int f;
+
+			pr_err("%s: CS did not scale with frequency! scaled min:%llu, max:%llu\n",
+			       engine->name,
+			       max.freq * min.count,
+			       min.freq * max.count);
+			show_pcu_config(rps);
+
+			for (f = min.freq + 1; f <= rps->max_freq; f++) {
+				int act = f;
+				u64 count;
+
+				count = measure_cs_frequency_at(rps, engine, &act);
+				if (act < f)
+					break;
+
+				pr_info("%s: %x:%uMHz: %lluKHz [%d%%]\n",
+					engine->name,
+					act, intel_gpu_freq(rps, act), count,
+					(int)DIV64_U64_ROUND_CLOSEST(100 * min.freq * count,
+								     act * min.count));
+
+				f = act; /* may skip ahead [pcu granularity] */
+			}
+
+			err = -EINTR; /* ignore error, continue on with test */
+		}
+
+err_vma:
+		*cancel = MI_BATCH_BUFFER_END;
+		i915_gem_object_flush_map(vma->obj);
+		i915_gem_object_unpin_map(vma->obj);
+		i915_vma_unpin(vma);
+		i915_vma_unlock(vma);
+		i915_vma_put(vma);
+
+		st_engine_heartbeat_enable(engine);
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	intel_gt_pm_wait_for_idle(gt);
+	rps->work.func = saved_work;
+
+	if (CPU_LATENCY >= 0)
+		cpu_latency_qos_remove_request(&qos);
+
+	return err;
+}
+
+int live_rps_frequency_srm(void *arg)
+{
+	void (*saved_work)(struct work_struct *wrk);
+	struct intel_gt *gt = arg;
+	struct intel_rps *rps = &gt->rps;
+	struct intel_engine_cs *engine;
+	struct pm_qos_request qos;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * The premise is that the GPU does change frequency at our behest.
+	 * Let's check there is a correspondence between the requested
+	 * frequency, the actual frequency, and the observed clock rate.
+	 */
+
+	if (!intel_rps_is_enabled(rps))
+		return 0;
+
+	if (GRAPHICS_VER(gt->i915) < 8) /* for CS simplicity */
+		return 0;
+
+	if (CPU_LATENCY >= 0)
+		cpu_latency_qos_add_request(&qos, CPU_LATENCY);
+
+	intel_gt_pm_wait_for_idle(gt);
+	saved_work = rps->work.func;
+	rps->work.func = dummy_rps_work;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq;
+		struct i915_vma *vma;
+		u32 *cancel, *cntr;
+		struct {
+			u64 count;
+			int freq;
+		} min, max;
+
+		st_engine_heartbeat_disable(engine);
+
+		vma = create_spin_counter(engine,
+					  engine->kernel_context->vm, true,
+					  &cancel, &cntr);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			st_engine_heartbeat_enable(engine);
+			break;
+		}
+
+		rq = intel_engine_create_kernel_request(engine);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_vma;
+		}
+
+		err = i915_request_await_object(rq, vma->obj, false);
+		if (!err)
+			err = i915_vma_move_to_active(vma, rq, 0);
+		if (!err)
+			err = rq->engine->emit_bb_start(rq,
+							vma->node.start,
+							PAGE_SIZE, 0);
+		i915_request_add(rq);
+		if (err)
+			goto err_vma;
+
+		if (wait_for(READ_ONCE(*cntr), 10)) {
+			pr_err("%s: timed loop did not start\n",
+			       engine->name);
+			goto err_vma;
+		}
+
+		min.freq = rps->min_freq;
+		min.count = measure_frequency_at(rps, cntr, &min.freq);
+
+		max.freq = rps->max_freq;
+		max.count = measure_frequency_at(rps, cntr, &max.freq);
+
+		pr_info("%s: min:%lluKHz @ %uMHz, max:%lluKHz @ %uMHz [%d%%]\n",
+			engine->name,
+			min.count, intel_gpu_freq(rps, min.freq),
+			max.count, intel_gpu_freq(rps, max.freq),
+			(int)DIV64_U64_ROUND_CLOSEST(100 * min.freq * max.count,
+						     max.freq * min.count));
+
+		if (!scaled_within(max.freq * min.count,
+				   min.freq * max.count,
+				   1, 2)) {
+			int f;
+
+			pr_err("%s: CS did not scale with frequency! scaled min:%llu, max:%llu\n",
+			       engine->name,
+			       max.freq * min.count,
+			       min.freq * max.count);
+			show_pcu_config(rps);
+
+			for (f = min.freq + 1; f <= rps->max_freq; f++) {
+				int act = f;
+				u64 count;
+
+				count = measure_frequency_at(rps, cntr, &act);
+				if (act < f)
+					break;
+
+				pr_info("%s: %x:%uMHz: %lluKHz [%d%%]\n",
+					engine->name,
+					act, intel_gpu_freq(rps, act), count,
+					(int)DIV64_U64_ROUND_CLOSEST(100 * min.freq * count,
+								     act * min.count));
+
+				f = act; /* may skip ahead [pcu granularity] */
+			}
+
+			err = -EINTR; /* ignore error, continue on with test */
+		}
+
+err_vma:
+		*cancel = MI_BATCH_BUFFER_END;
+		i915_gem_object_flush_map(vma->obj);
+		i915_gem_object_unpin_map(vma->obj);
+		i915_vma_unpin(vma);
+		i915_vma_unlock(vma);
+		i915_vma_put(vma);
+
+		st_engine_heartbeat_enable(engine);
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	intel_gt_pm_wait_for_idle(gt);
+	rps->work.func = saved_work;
+
+	if (CPU_LATENCY >= 0)
+		cpu_latency_qos_remove_request(&qos);
+
+	return err;
+}
+
+static void sleep_for_ei(struct intel_rps *rps, int timeout_us)
+{
+	/* Flush any previous EI */
+	usleep_range(timeout_us, 2 * timeout_us);
+
+	/* Reset the interrupt status */
+	rps_disable_interrupts(rps);
+	GEM_BUG_ON(rps->pm_iir);
+	rps_enable_interrupts(rps);
+
+	/* And then wait for the timeout, for real this time */
+	usleep_range(2 * timeout_us, 3 * timeout_us);
+}
+
+static int __rps_up_interrupt(struct intel_rps *rps,
+			      struct intel_engine_cs *engine,
+			      struct igt_spinner *spin)
+{
+	struct intel_uncore *uncore = engine->uncore;
+	struct i915_request *rq;
+	u32 timeout;
+
+	if (!intel_engine_can_store_dword(engine))
+		return 0;
+
+	rps_set_check(rps, rps->min_freq);
+
+	rq = igt_spinner_create_request(spin, engine->kernel_context, MI_NOOP);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	if (!igt_wait_for_spinner(spin, rq)) {
+		pr_err("%s: RPS spinner did not start\n",
+		       engine->name);
+		i915_request_put(rq);
+		intel_gt_set_wedged(engine->gt);
+		return -EIO;
+	}
+
+	if (!intel_rps_is_active(rps)) {
+		pr_err("%s: RPS not enabled on starting spinner\n",
+		       engine->name);
+		igt_spinner_end(spin);
+		i915_request_put(rq);
+		return -EINVAL;
+	}
+
+	if (!(rps->pm_events & GEN6_PM_RP_UP_THRESHOLD)) {
+		pr_err("%s: RPS did not register UP interrupt\n",
+		       engine->name);
+		i915_request_put(rq);
+		return -EINVAL;
+	}
+
+	if (rps->last_freq != rps->min_freq) {
+		pr_err("%s: RPS did not program min frequency\n",
+		       engine->name);
+		i915_request_put(rq);
+		return -EINVAL;
+	}
+
+	timeout = intel_uncore_read(uncore, GEN6_RP_UP_EI);
+	timeout = intel_gt_pm_interval_to_ns(engine->gt, timeout);
+	timeout = DIV_ROUND_UP(timeout, 1000);
+
+	sleep_for_ei(rps, timeout);
+	GEM_BUG_ON(i915_request_completed(rq));
+
+	igt_spinner_end(spin);
+	i915_request_put(rq);
+
+	if (rps->cur_freq != rps->min_freq) {
+		pr_err("%s: Frequency unexpectedly changed [up], now %d!\n",
+		       engine->name, intel_rps_read_actual_frequency(rps));
+		return -EINVAL;
+	}
+
+	if (!(rps->pm_iir & GEN6_PM_RP_UP_THRESHOLD)) {
+		pr_err("%s: UP interrupt not recorded for spinner, pm_iir:%x, prev_up:%x, up_threshold:%x, up_ei:%x\n",
+		       engine->name, rps->pm_iir,
+		       intel_uncore_read(uncore, GEN6_RP_PREV_UP),
+		       intel_uncore_read(uncore, GEN6_RP_UP_THRESHOLD),
+		       intel_uncore_read(uncore, GEN6_RP_UP_EI));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int __rps_down_interrupt(struct intel_rps *rps,
+				struct intel_engine_cs *engine)
+{
+	struct intel_uncore *uncore = engine->uncore;
+	u32 timeout;
+
+	rps_set_check(rps, rps->max_freq);
+
+	if (!(rps->pm_events & GEN6_PM_RP_DOWN_THRESHOLD)) {
+		pr_err("%s: RPS did not register DOWN interrupt\n",
+		       engine->name);
+		return -EINVAL;
+	}
+
+	if (rps->last_freq != rps->max_freq) {
+		pr_err("%s: RPS did not program max frequency\n",
+		       engine->name);
+		return -EINVAL;
+	}
+
+	timeout = intel_uncore_read(uncore, GEN6_RP_DOWN_EI);
+	timeout = intel_gt_pm_interval_to_ns(engine->gt, timeout);
+	timeout = DIV_ROUND_UP(timeout, 1000);
+
+	sleep_for_ei(rps, timeout);
+
+	if (rps->cur_freq != rps->max_freq) {
+		pr_err("%s: Frequency unexpectedly changed [down], now %d!\n",
+		       engine->name,
+		       intel_rps_read_actual_frequency(rps));
+		return -EINVAL;
+	}
+
+	if (!(rps->pm_iir & (GEN6_PM_RP_DOWN_THRESHOLD | GEN6_PM_RP_DOWN_TIMEOUT))) {
+		pr_err("%s: DOWN interrupt not recorded for idle, pm_iir:%x, prev_down:%x, down_threshold:%x, down_ei:%x [prev_up:%x, up_threshold:%x, up_ei:%x]\n",
+		       engine->name, rps->pm_iir,
+		       intel_uncore_read(uncore, GEN6_RP_PREV_DOWN),
+		       intel_uncore_read(uncore, GEN6_RP_DOWN_THRESHOLD),
+		       intel_uncore_read(uncore, GEN6_RP_DOWN_EI),
+		       intel_uncore_read(uncore, GEN6_RP_PREV_UP),
+		       intel_uncore_read(uncore, GEN6_RP_UP_THRESHOLD),
+		       intel_uncore_read(uncore, GEN6_RP_UP_EI));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int live_rps_interrupt(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_rps *rps = &gt->rps;
+	void (*saved_work)(struct work_struct *wrk);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	u32 pm_events;
+	int err = 0;
+
+	/*
+	 * First, let's check whether or not we are receiving interrupts.
+	 */
+
+	if (!intel_rps_has_interrupts(rps) || GRAPHICS_VER(gt->i915) < 6)
+		return 0;
+
+	intel_gt_pm_get(gt);
+	pm_events = rps->pm_events;
+	intel_gt_pm_put(gt);
+	if (!pm_events) {
+		pr_err("No RPS PM events registered, but RPS is enabled?\n");
+		return -ENODEV;
+	}
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	intel_gt_pm_wait_for_idle(gt);
+	saved_work = rps->work.func;
+	rps->work.func = dummy_rps_work;
+
+	for_each_engine(engine, gt, id) {
+		/* Keep the engine busy with a spinner; expect an UP! */
+		if (pm_events & GEN6_PM_RP_UP_THRESHOLD) {
+			intel_gt_pm_wait_for_idle(engine->gt);
+			GEM_BUG_ON(intel_rps_is_active(rps));
+
+			st_engine_heartbeat_disable(engine);
+
+			err = __rps_up_interrupt(rps, engine, &spin);
+
+			st_engine_heartbeat_enable(engine);
+			if (err)
+				goto out;
+
+			intel_gt_pm_wait_for_idle(engine->gt);
+		}
+
+		/* Keep the engine awake but idle and check for DOWN */
+		if (pm_events & GEN6_PM_RP_DOWN_THRESHOLD) {
+			st_engine_heartbeat_disable(engine);
+			intel_rc6_disable(&gt->rc6);
+
+			err = __rps_down_interrupt(rps, engine);
+
+			intel_rc6_enable(&gt->rc6);
+			st_engine_heartbeat_enable(engine);
+			if (err)
+				goto out;
+		}
+	}
+
+out:
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	igt_spinner_fini(&spin);
+
+	intel_gt_pm_wait_for_idle(gt);
+	rps->work.func = saved_work;
+
+	return err;
+}
+
+static u64 __measure_power(int duration_ms)
+{
+	u64 dE, dt;
+
+	dt = ktime_get();
+	dE = librapl_energy_uJ();
+	usleep_range(1000 * duration_ms, 2000 * duration_ms);
+	dE = librapl_energy_uJ() - dE;
+	dt = ktime_get() - dt;
+
+	return div64_u64(1000 * 1000 * dE, dt);
+}
+
+static u64 measure_power_at(struct intel_rps *rps, int *freq)
+{
+	u64 x[5];
+	int i;
+
+	*freq = rps_set_check(rps, *freq);
+	for (i = 0; i < 5; i++)
+		x[i] = __measure_power(5);
+	*freq = (*freq + read_cagf(rps)) / 2;
+
+	/* A simple triangle filter for better result stability */
+	sort(x, 5, sizeof(*x), cmp_u64, NULL);
+	return div_u64(x[1] + 2 * x[2] + x[3], 4);
+}
+
+int live_rps_power(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_rps *rps = &gt->rps;
+	void (*saved_work)(struct work_struct *wrk);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	/*
+	 * Our fundamental assumption is that running at lower frequency
+	 * actually saves power. Let's see if our RAPL measurement support
+	 * that theory.
+	 */
+
+	if (!intel_rps_is_enabled(rps) || GRAPHICS_VER(gt->i915) < 6)
+		return 0;
+
+	if (!librapl_supported(gt->i915))
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	intel_gt_pm_wait_for_idle(gt);
+	saved_work = rps->work.func;
+	rps->work.func = dummy_rps_work;
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq;
+		struct {
+			u64 power;
+			int freq;
+		} min, max;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		st_engine_heartbeat_disable(engine);
+
+		rq = igt_spinner_create_request(&spin,
+						engine->kernel_context,
+						MI_NOOP);
+		if (IS_ERR(rq)) {
+			st_engine_heartbeat_enable(engine);
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_add(rq);
+
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			pr_err("%s: RPS spinner did not start\n",
+			       engine->name);
+			igt_spinner_end(&spin);
+			st_engine_heartbeat_enable(engine);
+			intel_gt_set_wedged(engine->gt);
+			err = -EIO;
+			break;
+		}
+
+		max.freq = rps->max_freq;
+		max.power = measure_power_at(rps, &max.freq);
+
+		min.freq = rps->min_freq;
+		min.power = measure_power_at(rps, &min.freq);
+
+		igt_spinner_end(&spin);
+		st_engine_heartbeat_enable(engine);
+
+		pr_info("%s: min:%llumW @ %uMHz, max:%llumW @ %uMHz\n",
+			engine->name,
+			min.power, intel_gpu_freq(rps, min.freq),
+			max.power, intel_gpu_freq(rps, max.freq));
+
+		if (10 * min.freq >= 9 * max.freq) {
+			pr_notice("Could not control frequency, ran at [%d:%uMHz, %d:%uMhz]\n",
+				  min.freq, intel_gpu_freq(rps, min.freq),
+				  max.freq, intel_gpu_freq(rps, max.freq));
+			continue;
+		}
+
+		if (11 * min.power > 10 * max.power) {
+			pr_err("%s: did not conserve power when setting lower frequency!\n",
+			       engine->name);
+			err = -EINVAL;
+			break;
+		}
+
+		if (igt_flush_test(gt->i915)) {
+			err = -EIO;
+			break;
+		}
+	}
+
+	igt_spinner_fini(&spin);
+
+	intel_gt_pm_wait_for_idle(gt);
+	rps->work.func = saved_work;
+
+	return err;
+}
+
+int live_rps_dynamic(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_rps *rps = &gt->rps;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	int err = 0;
+
+	/*
+	 * We've looked at the bascs, and have established that we
+	 * can change the clock frequency and that the HW will generate
+	 * interrupts based on load. Now we check how we integrate those
+	 * moving parts into dynamic reclocking based on load.
+	 */
+
+	if (!intel_rps_is_enabled(rps) || GRAPHICS_VER(gt->i915) < 6)
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	if (intel_rps_has_interrupts(rps))
+		pr_info("RPS has interrupt support\n");
+	if (intel_rps_uses_timer(rps))
+		pr_info("RPS has timer support\n");
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq;
+		struct {
+			ktime_t dt;
+			u8 freq;
+		} min, max;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		intel_gt_pm_wait_for_idle(gt);
+		GEM_BUG_ON(intel_rps_is_active(rps));
+		rps->cur_freq = rps->min_freq;
+
+		intel_engine_pm_get(engine);
+		intel_rc6_disable(&gt->rc6);
+		GEM_BUG_ON(rps->last_freq != rps->min_freq);
+
+		rq = igt_spinner_create_request(&spin,
+						engine->kernel_context,
+						MI_NOOP);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err;
+		}
+
+		i915_request_add(rq);
+
+		max.dt = ktime_get();
+		max.freq = wait_for_freq(rps, rps->max_freq, 500);
+		max.dt = ktime_sub(ktime_get(), max.dt);
+
+		igt_spinner_end(&spin);
+
+		min.dt = ktime_get();
+		min.freq = wait_for_freq(rps, rps->min_freq, 2000);
+		min.dt = ktime_sub(ktime_get(), min.dt);
+
+		pr_info("%s: dynamically reclocked to %u:%uMHz while busy in %lluns, and %u:%uMHz while idle in %lluns\n",
+			engine->name,
+			max.freq, intel_gpu_freq(rps, max.freq),
+			ktime_to_ns(max.dt),
+			min.freq, intel_gpu_freq(rps, min.freq),
+			ktime_to_ns(min.dt));
+		if (min.freq >= max.freq) {
+			pr_err("%s: dynamic reclocking of spinner failed\n!",
+			       engine->name);
+			err = -EINVAL;
+		}
+
+err:
+		intel_rc6_enable(&gt->rc6);
+		intel_engine_pm_put(engine);
+
+		if (igt_flush_test(gt->i915))
+			err = -EIO;
+		if (err)
+			break;
+	}
+
+	igt_spinner_fini(&spin);
+
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_rps.h b/drivers/gpu/drm/i915/gt/selftest_rps.h
new file mode 100644
index 000000000..6e82a631c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_rps.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef SELFTEST_RPS_H
+#define SELFTEST_RPS_H
+
+int live_rps_control(void *arg);
+int live_rps_clock_interval(void *arg);
+int live_rps_frequency_cs(void *arg);
+int live_rps_frequency_srm(void *arg);
+int live_rps_power(void *arg);
+int live_rps_interrupt(void *arg);
+int live_rps_dynamic(void *arg);
+
+#endif /* SELFTEST_RPS_H */
diff --git a/drivers/gpu/drm/i915/gt/selftest_slpc.c b/drivers/gpu/drm/i915/gt/selftest_slpc.c
new file mode 100644
index 000000000..f8a1d27df
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_slpc.c
@@ -0,0 +1,307 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#define NUM_STEPS 5
+#define H2G_DELAY 50000
+#define delay_for_h2g() usleep_range(H2G_DELAY, H2G_DELAY + 10000)
+#define FREQUENCY_REQ_UNIT	DIV_ROUND_CLOSEST(GT_FREQUENCY_MULTIPLIER, \
+						  GEN9_FREQ_SCALER)
+enum test_type {
+	VARY_MIN,
+	VARY_MAX,
+	MAX_GRANTED
+};
+
+static int slpc_set_min_freq(struct intel_guc_slpc *slpc, u32 freq)
+{
+	int ret;
+
+	ret = intel_guc_slpc_set_min_freq(slpc, freq);
+	if (ret)
+		pr_err("Could not set min frequency to [%u]\n", freq);
+	else /* Delay to ensure h2g completes */
+		delay_for_h2g();
+
+	return ret;
+}
+
+static int slpc_set_max_freq(struct intel_guc_slpc *slpc, u32 freq)
+{
+	int ret;
+
+	ret = intel_guc_slpc_set_max_freq(slpc, freq);
+	if (ret)
+		pr_err("Could not set maximum frequency [%u]\n",
+		       freq);
+	else /* Delay to ensure h2g completes */
+		delay_for_h2g();
+
+	return ret;
+}
+
+static int vary_max_freq(struct intel_guc_slpc *slpc, struct intel_rps *rps,
+			 u32 *max_act_freq)
+{
+	u32 step, max_freq, req_freq;
+	u32 act_freq;
+	int err = 0;
+
+	/* Go from max to min in 5 steps */
+	step = (slpc->rp0_freq - slpc->min_freq) / NUM_STEPS;
+	*max_act_freq = slpc->min_freq;
+	for (max_freq = slpc->rp0_freq; max_freq > slpc->min_freq;
+				max_freq -= step) {
+		err = slpc_set_max_freq(slpc, max_freq);
+		if (err)
+			break;
+
+		req_freq = intel_rps_read_punit_req_frequency(rps);
+
+		/* GuC requests freq in multiples of 50/3 MHz */
+		if (req_freq > (max_freq + FREQUENCY_REQ_UNIT)) {
+			pr_err("SWReq is %d, should be at most %d\n", req_freq,
+			       max_freq + FREQUENCY_REQ_UNIT);
+			err = -EINVAL;
+		}
+
+		act_freq =  intel_rps_read_actual_frequency(rps);
+		if (act_freq > *max_act_freq)
+			*max_act_freq = act_freq;
+
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static int vary_min_freq(struct intel_guc_slpc *slpc, struct intel_rps *rps,
+			 u32 *max_act_freq)
+{
+	u32 step, min_freq, req_freq;
+	u32 act_freq;
+	int err = 0;
+
+	/* Go from min to max in 5 steps */
+	step = (slpc->rp0_freq - slpc->min_freq) / NUM_STEPS;
+	*max_act_freq = slpc->min_freq;
+	for (min_freq = slpc->min_freq; min_freq < slpc->rp0_freq;
+				min_freq += step) {
+		err = slpc_set_min_freq(slpc, min_freq);
+		if (err)
+			break;
+
+		req_freq = intel_rps_read_punit_req_frequency(rps);
+
+		/* GuC requests freq in multiples of 50/3 MHz */
+		if (req_freq < (min_freq - FREQUENCY_REQ_UNIT)) {
+			pr_err("SWReq is %d, should be at least %d\n", req_freq,
+			       min_freq - FREQUENCY_REQ_UNIT);
+			err = -EINVAL;
+		}
+
+		act_freq =  intel_rps_read_actual_frequency(rps);
+		if (act_freq > *max_act_freq)
+			*max_act_freq = act_freq;
+
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static int max_granted_freq(struct intel_guc_slpc *slpc, struct intel_rps *rps, u32 *max_act_freq)
+{
+	struct intel_gt *gt = rps_to_gt(rps);
+	u32 perf_limit_reasons;
+	int err = 0;
+
+	err = slpc_set_min_freq(slpc, slpc->rp0_freq);
+	if (err)
+		return err;
+
+	*max_act_freq =  intel_rps_read_actual_frequency(rps);
+	if (*max_act_freq != slpc->rp0_freq) {
+		/* Check if there was some throttling by pcode */
+		perf_limit_reasons = intel_uncore_read(gt->uncore, GT0_PERF_LIMIT_REASONS);
+
+		/* If not, this is an error */
+		if (!(perf_limit_reasons & GT0_PERF_LIMIT_REASONS_MASK)) {
+			pr_err("Pcode did not grant max freq\n");
+			err = -EINVAL;
+		} else {
+			pr_info("Pcode throttled frequency 0x%x\n", perf_limit_reasons);
+		}
+	}
+
+	return err;
+}
+
+static int run_test(struct intel_gt *gt, int test_type)
+{
+	struct intel_guc_slpc *slpc = &gt->uc.guc.slpc;
+	struct intel_rps *rps = &gt->rps;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct igt_spinner spin;
+	u32 slpc_min_freq, slpc_max_freq;
+	int err = 0;
+
+	if (!intel_uc_uses_guc_slpc(&gt->uc))
+		return 0;
+
+	if (igt_spinner_init(&spin, gt))
+		return -ENOMEM;
+
+	if (intel_guc_slpc_get_max_freq(slpc, &slpc_max_freq)) {
+		pr_err("Could not get SLPC max freq\n");
+		return -EIO;
+	}
+
+	if (intel_guc_slpc_get_min_freq(slpc, &slpc_min_freq)) {
+		pr_err("Could not get SLPC min freq\n");
+		return -EIO;
+	}
+
+	/*
+	 * FIXME: With efficient frequency enabled, GuC can request
+	 * frequencies higher than the SLPC max. While this is fixed
+	 * in GuC, we level set these tests with RPn as min.
+	 */
+	err = slpc_set_min_freq(slpc, slpc->min_freq);
+	if (err)
+		return err;
+
+	if (slpc->min_freq == slpc->rp0_freq) {
+		pr_err("Min/Max are fused to the same value\n");
+		return -EINVAL;
+	}
+
+	intel_gt_pm_wait_for_idle(gt);
+	intel_gt_pm_get(gt);
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq;
+		u32 max_act_freq;
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		st_engine_heartbeat_disable(engine);
+
+		rq = igt_spinner_create_request(&spin,
+						engine->kernel_context,
+						MI_NOOP);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			st_engine_heartbeat_enable(engine);
+			break;
+		}
+
+		i915_request_add(rq);
+
+		if (!igt_wait_for_spinner(&spin, rq)) {
+			pr_err("%s: Spinner did not start\n",
+			       engine->name);
+			igt_spinner_end(&spin);
+			st_engine_heartbeat_enable(engine);
+			intel_gt_set_wedged(engine->gt);
+			err = -EIO;
+			break;
+		}
+
+		switch (test_type) {
+		case VARY_MIN:
+			err = vary_min_freq(slpc, rps, &max_act_freq);
+			break;
+
+		case VARY_MAX:
+			err = vary_max_freq(slpc, rps, &max_act_freq);
+			break;
+
+		case MAX_GRANTED:
+			/* Media engines have a different RP0 */
+			if (engine->class == VIDEO_DECODE_CLASS ||
+			    engine->class == VIDEO_ENHANCEMENT_CLASS) {
+				igt_spinner_end(&spin);
+				st_engine_heartbeat_enable(engine);
+				err = 0;
+				continue;
+			}
+
+			err = max_granted_freq(slpc, rps, &max_act_freq);
+			break;
+		}
+
+		pr_info("Max actual frequency for %s was %d\n",
+			engine->name, max_act_freq);
+
+		/* Actual frequency should rise above min */
+		if (max_act_freq <= slpc_min_freq) {
+			pr_err("Actual freq did not rise above min\n");
+			pr_err("Perf Limit Reasons: 0x%x\n",
+			       intel_uncore_read(gt->uncore, GT0_PERF_LIMIT_REASONS));
+			err = -EINVAL;
+		}
+
+		igt_spinner_end(&spin);
+		st_engine_heartbeat_enable(engine);
+
+		if (err)
+			break;
+	}
+
+	/* Restore min/max frequencies */
+	slpc_set_max_freq(slpc, slpc_max_freq);
+	slpc_set_min_freq(slpc, slpc_min_freq);
+
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	intel_gt_pm_put(gt);
+	igt_spinner_fini(&spin);
+	intel_gt_pm_wait_for_idle(gt);
+
+	return err;
+}
+
+static int live_slpc_vary_min(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_gt *gt = to_gt(i915);
+
+	return run_test(gt, VARY_MIN);
+}
+
+static int live_slpc_vary_max(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_gt *gt = to_gt(i915);
+
+	return run_test(gt, VARY_MAX);
+}
+
+/* check if pcode can grant RP0 */
+static int live_slpc_max_granted(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_gt *gt = to_gt(i915);
+
+	return run_test(gt, MAX_GRANTED);
+}
+
+int intel_slpc_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_slpc_vary_max),
+		SUBTEST(live_slpc_vary_min),
+		SUBTEST(live_slpc_max_granted),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_timeline.c b/drivers/gpu/drm/i915/gt/selftest_timeline.c
new file mode 100644
index 000000000..522d01905
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_timeline.c
@@ -0,0 +1,1425 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2017-2018 Intel Corporation
+ */
+
+#include <linux/prime_numbers.h>
+#include <linux/string_helpers.h>
+
+#include "intel_context.h"
+#include "intel_engine_heartbeat.h"
+#include "intel_engine_pm.h"
+#include "intel_engine_regs.h"
+#include "intel_gpu_commands.h"
+#include "intel_gt.h"
+#include "intel_gt_requests.h"
+#include "intel_ring.h"
+#include "selftest_engine_heartbeat.h"
+
+#include "../selftests/i915_random.h"
+#include "../i915_selftest.h"
+
+#include "selftests/igt_flush_test.h"
+#include "selftests/lib_sw_fence.h"
+#include "selftests/mock_gem_device.h"
+#include "selftests/mock_timeline.h"
+
+static struct page *hwsp_page(struct intel_timeline *tl)
+{
+	struct drm_i915_gem_object *obj = tl->hwsp_ggtt->obj;
+
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+	return sg_page(obj->mm.pages->sgl);
+}
+
+static unsigned long hwsp_cacheline(struct intel_timeline *tl)
+{
+	unsigned long address = (unsigned long)page_address(hwsp_page(tl));
+
+	return (address + offset_in_page(tl->hwsp_offset)) / TIMELINE_SEQNO_BYTES;
+}
+
+static int selftest_tl_pin(struct intel_timeline *tl)
+{
+	struct i915_gem_ww_ctx ww;
+	int err;
+
+	i915_gem_ww_ctx_init(&ww, false);
+retry:
+	err = i915_gem_object_lock(tl->hwsp_ggtt->obj, &ww);
+	if (!err)
+		err = intel_timeline_pin(tl, &ww);
+
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	return err;
+}
+
+/* Only half of seqno's are usable, see __intel_timeline_get_seqno() */
+#define CACHELINES_PER_PAGE (PAGE_SIZE / TIMELINE_SEQNO_BYTES / 2)
+
+struct mock_hwsp_freelist {
+	struct intel_gt *gt;
+	struct radix_tree_root cachelines;
+	struct intel_timeline **history;
+	unsigned long count, max;
+	struct rnd_state prng;
+};
+
+enum {
+	SHUFFLE = BIT(0),
+};
+
+static void __mock_hwsp_record(struct mock_hwsp_freelist *state,
+			       unsigned int idx,
+			       struct intel_timeline *tl)
+{
+	tl = xchg(&state->history[idx], tl);
+	if (tl) {
+		radix_tree_delete(&state->cachelines, hwsp_cacheline(tl));
+		intel_timeline_unpin(tl);
+		intel_timeline_put(tl);
+	}
+}
+
+static int __mock_hwsp_timeline(struct mock_hwsp_freelist *state,
+				unsigned int count,
+				unsigned int flags)
+{
+	struct intel_timeline *tl;
+	unsigned int idx;
+
+	while (count--) {
+		unsigned long cacheline;
+		int err;
+
+		tl = intel_timeline_create(state->gt);
+		if (IS_ERR(tl))
+			return PTR_ERR(tl);
+
+		err = selftest_tl_pin(tl);
+		if (err) {
+			intel_timeline_put(tl);
+			return err;
+		}
+
+		cacheline = hwsp_cacheline(tl);
+		err = radix_tree_insert(&state->cachelines, cacheline, tl);
+		if (err) {
+			if (err == -EEXIST) {
+				pr_err("HWSP cacheline %lu already used; duplicate allocation!\n",
+				       cacheline);
+			}
+			intel_timeline_unpin(tl);
+			intel_timeline_put(tl);
+			return err;
+		}
+
+		idx = state->count++ % state->max;
+		__mock_hwsp_record(state, idx, tl);
+	}
+
+	if (flags & SHUFFLE)
+		i915_prandom_shuffle(state->history,
+				     sizeof(*state->history),
+				     min(state->count, state->max),
+				     &state->prng);
+
+	count = i915_prandom_u32_max_state(min(state->count, state->max),
+					   &state->prng);
+	while (count--) {
+		idx = --state->count % state->max;
+		__mock_hwsp_record(state, idx, NULL);
+	}
+
+	return 0;
+}
+
+static int mock_hwsp_freelist(void *arg)
+{
+	struct mock_hwsp_freelist state;
+	struct drm_i915_private *i915;
+	const struct {
+		const char *name;
+		unsigned int flags;
+	} phases[] = {
+		{ "linear", 0 },
+		{ "shuffled", SHUFFLE },
+		{ },
+	}, *p;
+	unsigned int na;
+	int err = 0;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	INIT_RADIX_TREE(&state.cachelines, GFP_KERNEL);
+	state.prng = I915_RND_STATE_INITIALIZER(i915_selftest.random_seed);
+
+	state.gt = to_gt(i915);
+
+	/*
+	 * Create a bunch of timelines and check that their HWSP do not overlap.
+	 * Free some, and try again.
+	 */
+
+	state.max = PAGE_SIZE / sizeof(*state.history);
+	state.count = 0;
+	state.history = kcalloc(state.max, sizeof(*state.history), GFP_KERNEL);
+	if (!state.history) {
+		err = -ENOMEM;
+		goto err_put;
+	}
+
+	for (p = phases; p->name; p++) {
+		pr_debug("%s(%s)\n", __func__, p->name);
+		for_each_prime_number_from(na, 1, 2 * CACHELINES_PER_PAGE) {
+			err = __mock_hwsp_timeline(&state, na, p->flags);
+			if (err)
+				goto out;
+		}
+	}
+
+out:
+	for (na = 0; na < state.max; na++)
+		__mock_hwsp_record(&state, na, NULL);
+	kfree(state.history);
+err_put:
+	mock_destroy_device(i915);
+	return err;
+}
+
+struct __igt_sync {
+	const char *name;
+	u32 seqno;
+	bool expected;
+	bool set;
+};
+
+static int __igt_sync(struct intel_timeline *tl,
+		      u64 ctx,
+		      const struct __igt_sync *p,
+		      const char *name)
+{
+	int ret;
+
+	if (__intel_timeline_sync_is_later(tl, ctx, p->seqno) != p->expected) {
+		pr_err("%s: %s(ctx=%llu, seqno=%u) expected passed %s but failed\n",
+		       name, p->name, ctx, p->seqno, str_yes_no(p->expected));
+		return -EINVAL;
+	}
+
+	if (p->set) {
+		ret = __intel_timeline_sync_set(tl, ctx, p->seqno);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int igt_sync(void *arg)
+{
+	const struct __igt_sync pass[] = {
+		{ "unset", 0, false, false },
+		{ "new", 0, false, true },
+		{ "0a", 0, true, true },
+		{ "1a", 1, false, true },
+		{ "1b", 1, true, true },
+		{ "0b", 0, true, false },
+		{ "2a", 2, false, true },
+		{ "4", 4, false, true },
+		{ "INT_MAX", INT_MAX, false, true },
+		{ "INT_MAX-1", INT_MAX-1, true, false },
+		{ "INT_MAX+1", (u32)INT_MAX+1, false, true },
+		{ "INT_MAX", INT_MAX, true, false },
+		{ "UINT_MAX", UINT_MAX, false, true },
+		{ "wrap", 0, false, true },
+		{ "unwrap", UINT_MAX, true, false },
+		{},
+	}, *p;
+	struct intel_timeline tl;
+	int order, offset;
+	int ret = -ENODEV;
+
+	mock_timeline_init(&tl, 0);
+	for (p = pass; p->name; p++) {
+		for (order = 1; order < 64; order++) {
+			for (offset = -1; offset <= (order > 1); offset++) {
+				u64 ctx = BIT_ULL(order) + offset;
+
+				ret = __igt_sync(&tl, ctx, p, "1");
+				if (ret)
+					goto out;
+			}
+		}
+	}
+	mock_timeline_fini(&tl);
+
+	mock_timeline_init(&tl, 0);
+	for (order = 1; order < 64; order++) {
+		for (offset = -1; offset <= (order > 1); offset++) {
+			u64 ctx = BIT_ULL(order) + offset;
+
+			for (p = pass; p->name; p++) {
+				ret = __igt_sync(&tl, ctx, p, "2");
+				if (ret)
+					goto out;
+			}
+		}
+	}
+
+out:
+	mock_timeline_fini(&tl);
+	return ret;
+}
+
+static unsigned int random_engine(struct rnd_state *rnd)
+{
+	return i915_prandom_u32_max_state(I915_NUM_ENGINES, rnd);
+}
+
+static int bench_sync(void *arg)
+{
+	struct rnd_state prng;
+	struct intel_timeline tl;
+	unsigned long end_time, count;
+	u64 prng32_1M;
+	ktime_t kt;
+	int order, last_order;
+
+	mock_timeline_init(&tl, 0);
+
+	/* Lookups from cache are very fast and so the random number generation
+	 * and the loop itself becomes a significant factor in the per-iteration
+	 * timings. We try to compensate the results by measuring the overhead
+	 * of the prng and subtract it from the reported results.
+	 */
+	prandom_seed_state(&prng, i915_selftest.random_seed);
+	count = 0;
+	kt = ktime_get();
+	end_time = jiffies + HZ/10;
+	do {
+		u32 x;
+
+		/* Make sure the compiler doesn't optimise away the prng call */
+		WRITE_ONCE(x, prandom_u32_state(&prng));
+
+		count++;
+	} while (!time_after(jiffies, end_time));
+	kt = ktime_sub(ktime_get(), kt);
+	pr_debug("%s: %lu random evaluations, %lluns/prng\n",
+		 __func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
+	prng32_1M = div64_ul(ktime_to_ns(kt) << 20, count);
+
+	/* Benchmark (only) setting random context ids */
+	prandom_seed_state(&prng, i915_selftest.random_seed);
+	count = 0;
+	kt = ktime_get();
+	end_time = jiffies + HZ/10;
+	do {
+		u64 id = i915_prandom_u64_state(&prng);
+
+		__intel_timeline_sync_set(&tl, id, 0);
+		count++;
+	} while (!time_after(jiffies, end_time));
+	kt = ktime_sub(ktime_get(), kt);
+	kt = ktime_sub_ns(kt, (count * prng32_1M * 2) >> 20);
+	pr_info("%s: %lu random insertions, %lluns/insert\n",
+		__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
+
+	/* Benchmark looking up the exact same context ids as we just set */
+	prandom_seed_state(&prng, i915_selftest.random_seed);
+	end_time = count;
+	kt = ktime_get();
+	while (end_time--) {
+		u64 id = i915_prandom_u64_state(&prng);
+
+		if (!__intel_timeline_sync_is_later(&tl, id, 0)) {
+			mock_timeline_fini(&tl);
+			pr_err("Lookup of %llu failed\n", id);
+			return -EINVAL;
+		}
+	}
+	kt = ktime_sub(ktime_get(), kt);
+	kt = ktime_sub_ns(kt, (count * prng32_1M * 2) >> 20);
+	pr_info("%s: %lu random lookups, %lluns/lookup\n",
+		__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
+
+	mock_timeline_fini(&tl);
+	cond_resched();
+
+	mock_timeline_init(&tl, 0);
+
+	/* Benchmark setting the first N (in order) contexts */
+	count = 0;
+	kt = ktime_get();
+	end_time = jiffies + HZ/10;
+	do {
+		__intel_timeline_sync_set(&tl, count++, 0);
+	} while (!time_after(jiffies, end_time));
+	kt = ktime_sub(ktime_get(), kt);
+	pr_info("%s: %lu in-order insertions, %lluns/insert\n",
+		__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
+
+	/* Benchmark looking up the exact same context ids as we just set */
+	end_time = count;
+	kt = ktime_get();
+	while (end_time--) {
+		if (!__intel_timeline_sync_is_later(&tl, end_time, 0)) {
+			pr_err("Lookup of %lu failed\n", end_time);
+			mock_timeline_fini(&tl);
+			return -EINVAL;
+		}
+	}
+	kt = ktime_sub(ktime_get(), kt);
+	pr_info("%s: %lu in-order lookups, %lluns/lookup\n",
+		__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
+
+	mock_timeline_fini(&tl);
+	cond_resched();
+
+	mock_timeline_init(&tl, 0);
+
+	/* Benchmark searching for a random context id and maybe changing it */
+	prandom_seed_state(&prng, i915_selftest.random_seed);
+	count = 0;
+	kt = ktime_get();
+	end_time = jiffies + HZ/10;
+	do {
+		u32 id = random_engine(&prng);
+		u32 seqno = prandom_u32_state(&prng);
+
+		if (!__intel_timeline_sync_is_later(&tl, id, seqno))
+			__intel_timeline_sync_set(&tl, id, seqno);
+
+		count++;
+	} while (!time_after(jiffies, end_time));
+	kt = ktime_sub(ktime_get(), kt);
+	kt = ktime_sub_ns(kt, (count * prng32_1M * 2) >> 20);
+	pr_info("%s: %lu repeated insert/lookups, %lluns/op\n",
+		__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
+	mock_timeline_fini(&tl);
+	cond_resched();
+
+	/* Benchmark searching for a known context id and changing the seqno */
+	for (last_order = 1, order = 1; order < 32;
+	     ({ int tmp = last_order; last_order = order; order += tmp; })) {
+		unsigned int mask = BIT(order) - 1;
+
+		mock_timeline_init(&tl, 0);
+
+		count = 0;
+		kt = ktime_get();
+		end_time = jiffies + HZ/10;
+		do {
+			/* Without assuming too many details of the underlying
+			 * implementation, try to identify its phase-changes
+			 * (if any)!
+			 */
+			u64 id = (u64)(count & mask) << order;
+
+			__intel_timeline_sync_is_later(&tl, id, 0);
+			__intel_timeline_sync_set(&tl, id, 0);
+
+			count++;
+		} while (!time_after(jiffies, end_time));
+		kt = ktime_sub(ktime_get(), kt);
+		pr_info("%s: %lu cyclic/%d insert/lookups, %lluns/op\n",
+			__func__, count, order,
+			(long long)div64_ul(ktime_to_ns(kt), count));
+		mock_timeline_fini(&tl);
+		cond_resched();
+	}
+
+	return 0;
+}
+
+int intel_timeline_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(mock_hwsp_freelist),
+		SUBTEST(igt_sync),
+		SUBTEST(bench_sync),
+	};
+
+	return i915_subtests(tests, NULL);
+}
+
+static int emit_ggtt_store_dw(struct i915_request *rq, u32 addr, u32 value)
+{
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	if (GRAPHICS_VER(rq->engine->i915) >= 8) {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = addr;
+		*cs++ = 0;
+		*cs++ = value;
+	} else if (GRAPHICS_VER(rq->engine->i915) >= 4) {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = 0;
+		*cs++ = addr;
+		*cs++ = value;
+	} else {
+		*cs++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL;
+		*cs++ = addr;
+		*cs++ = value;
+		*cs++ = MI_NOOP;
+	}
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static struct i915_request *
+checked_tl_write(struct intel_timeline *tl, struct intel_engine_cs *engine, u32 value)
+{
+	struct i915_request *rq;
+	int err;
+
+	err = selftest_tl_pin(tl);
+	if (err) {
+		rq = ERR_PTR(err);
+		goto out;
+	}
+
+	if (READ_ONCE(*tl->hwsp_seqno) != tl->seqno) {
+		pr_err("Timeline created with incorrect breadcrumb, found %x, expected %x\n",
+		       *tl->hwsp_seqno, tl->seqno);
+		intel_timeline_unpin(tl);
+		return ERR_PTR(-EINVAL);
+	}
+
+	rq = intel_engine_create_kernel_request(engine);
+	if (IS_ERR(rq))
+		goto out_unpin;
+
+	i915_request_get(rq);
+
+	err = emit_ggtt_store_dw(rq, tl->hwsp_offset, value);
+	i915_request_add(rq);
+	if (err) {
+		i915_request_put(rq);
+		rq = ERR_PTR(err);
+	}
+
+out_unpin:
+	intel_timeline_unpin(tl);
+out:
+	if (IS_ERR(rq))
+		pr_err("Failed to write to timeline!\n");
+	return rq;
+}
+
+static int live_hwsp_engine(void *arg)
+{
+#define NUM_TIMELINES 4096
+	struct intel_gt *gt = arg;
+	struct intel_timeline **timelines;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	unsigned long count, n;
+	int err = 0;
+
+	/*
+	 * Create a bunch of timelines and check we can write
+	 * independently to each of their breadcrumb slots.
+	 */
+
+	timelines = kvmalloc_array(NUM_TIMELINES * I915_NUM_ENGINES,
+				   sizeof(*timelines),
+				   GFP_KERNEL);
+	if (!timelines)
+		return -ENOMEM;
+
+	count = 0;
+	for_each_engine(engine, gt, id) {
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		intel_engine_pm_get(engine);
+
+		for (n = 0; n < NUM_TIMELINES; n++) {
+			struct intel_timeline *tl;
+			struct i915_request *rq;
+
+			tl = intel_timeline_create(gt);
+			if (IS_ERR(tl)) {
+				err = PTR_ERR(tl);
+				break;
+			}
+
+			rq = checked_tl_write(tl, engine, count);
+			if (IS_ERR(rq)) {
+				intel_timeline_put(tl);
+				err = PTR_ERR(rq);
+				break;
+			}
+
+			timelines[count++] = tl;
+			i915_request_put(rq);
+		}
+
+		intel_engine_pm_put(engine);
+		if (err)
+			break;
+	}
+
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	for (n = 0; n < count; n++) {
+		struct intel_timeline *tl = timelines[n];
+
+		if (!err && READ_ONCE(*tl->hwsp_seqno) != n) {
+			GEM_TRACE_ERR("Invalid seqno:%lu stored in timeline %llu @ %x, found 0x%x\n",
+				      n, tl->fence_context, tl->hwsp_offset, *tl->hwsp_seqno);
+			GEM_TRACE_DUMP();
+			err = -EINVAL;
+		}
+		intel_timeline_put(tl);
+	}
+
+	kvfree(timelines);
+	return err;
+#undef NUM_TIMELINES
+}
+
+static int live_hwsp_alternate(void *arg)
+{
+#define NUM_TIMELINES 4096
+	struct intel_gt *gt = arg;
+	struct intel_timeline **timelines;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	unsigned long count, n;
+	int err = 0;
+
+	/*
+	 * Create a bunch of timelines and check we can write
+	 * independently to each of their breadcrumb slots with adjacent
+	 * engines.
+	 */
+
+	timelines = kvmalloc_array(NUM_TIMELINES * I915_NUM_ENGINES,
+				   sizeof(*timelines),
+				   GFP_KERNEL);
+	if (!timelines)
+		return -ENOMEM;
+
+	count = 0;
+	for (n = 0; n < NUM_TIMELINES; n++) {
+		for_each_engine(engine, gt, id) {
+			struct intel_timeline *tl;
+			struct i915_request *rq;
+
+			if (!intel_engine_can_store_dword(engine))
+				continue;
+
+			tl = intel_timeline_create(gt);
+			if (IS_ERR(tl)) {
+				err = PTR_ERR(tl);
+				goto out;
+			}
+
+			intel_engine_pm_get(engine);
+			rq = checked_tl_write(tl, engine, count);
+			intel_engine_pm_put(engine);
+			if (IS_ERR(rq)) {
+				intel_timeline_put(tl);
+				err = PTR_ERR(rq);
+				goto out;
+			}
+
+			timelines[count++] = tl;
+			i915_request_put(rq);
+		}
+	}
+
+out:
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	for (n = 0; n < count; n++) {
+		struct intel_timeline *tl = timelines[n];
+
+		if (!err && READ_ONCE(*tl->hwsp_seqno) != n) {
+			GEM_TRACE_ERR("Invalid seqno:%lu stored in timeline %llu @ %x, found 0x%x\n",
+				      n, tl->fence_context, tl->hwsp_offset, *tl->hwsp_seqno);
+			GEM_TRACE_DUMP();
+			err = -EINVAL;
+		}
+		intel_timeline_put(tl);
+	}
+
+	kvfree(timelines);
+	return err;
+#undef NUM_TIMELINES
+}
+
+static int live_hwsp_wrap(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	struct intel_timeline *tl;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Across a seqno wrap, we need to keep the old cacheline alive for
+	 * foreign GPU references.
+	 */
+
+	tl = intel_timeline_create(gt);
+	if (IS_ERR(tl))
+		return PTR_ERR(tl);
+
+	if (!tl->has_initial_breadcrumb)
+		goto out_free;
+
+	err = selftest_tl_pin(tl);
+	if (err)
+		goto out_free;
+
+	for_each_engine(engine, gt, id) {
+		const u32 *hwsp_seqno[2];
+		struct i915_request *rq;
+		u32 seqno[2];
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		rq = intel_engine_create_kernel_request(engine);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out;
+		}
+
+		tl->seqno = -4u;
+
+		mutex_lock_nested(&tl->mutex, SINGLE_DEPTH_NESTING);
+		err = intel_timeline_get_seqno(tl, rq, &seqno[0]);
+		mutex_unlock(&tl->mutex);
+		if (err) {
+			i915_request_add(rq);
+			goto out;
+		}
+		pr_debug("seqno[0]:%08x, hwsp_offset:%08x\n",
+			 seqno[0], tl->hwsp_offset);
+
+		err = emit_ggtt_store_dw(rq, tl->hwsp_offset, seqno[0]);
+		if (err) {
+			i915_request_add(rq);
+			goto out;
+		}
+		hwsp_seqno[0] = tl->hwsp_seqno;
+
+		mutex_lock_nested(&tl->mutex, SINGLE_DEPTH_NESTING);
+		err = intel_timeline_get_seqno(tl, rq, &seqno[1]);
+		mutex_unlock(&tl->mutex);
+		if (err) {
+			i915_request_add(rq);
+			goto out;
+		}
+		pr_debug("seqno[1]:%08x, hwsp_offset:%08x\n",
+			 seqno[1], tl->hwsp_offset);
+
+		err = emit_ggtt_store_dw(rq, tl->hwsp_offset, seqno[1]);
+		if (err) {
+			i915_request_add(rq);
+			goto out;
+		}
+		hwsp_seqno[1] = tl->hwsp_seqno;
+
+		/* With wrap should come a new hwsp */
+		GEM_BUG_ON(seqno[1] >= seqno[0]);
+		GEM_BUG_ON(hwsp_seqno[0] == hwsp_seqno[1]);
+
+		i915_request_add(rq);
+
+		if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+			pr_err("Wait for timeline writes timed out!\n");
+			err = -EIO;
+			goto out;
+		}
+
+		if (READ_ONCE(*hwsp_seqno[0]) != seqno[0] ||
+		    READ_ONCE(*hwsp_seqno[1]) != seqno[1]) {
+			pr_err("Bad timeline values: found (%x, %x), expected (%x, %x)\n",
+			       *hwsp_seqno[0], *hwsp_seqno[1],
+			       seqno[0], seqno[1]);
+			err = -EINVAL;
+			goto out;
+		}
+
+		intel_gt_retire_requests(gt); /* recycle HWSP */
+	}
+
+out:
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	intel_timeline_unpin(tl);
+out_free:
+	intel_timeline_put(tl);
+	return err;
+}
+
+static int emit_read_hwsp(struct i915_request *rq,
+			  u32 seqno, u32 hwsp,
+			  u32 *addr)
+{
+	const u32 gpr = i915_mmio_reg_offset(GEN8_RING_CS_GPR(rq->engine->mmio_base, 0));
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 12);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = *addr;
+	*cs++ = 0;
+	*cs++ = seqno;
+	*addr += 4;
+
+	*cs++ = MI_LOAD_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+	*cs++ = gpr;
+	*cs++ = hwsp;
+	*cs++ = 0;
+
+	*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+	*cs++ = gpr;
+	*cs++ = *addr;
+	*cs++ = 0;
+	*addr += 4;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+struct hwsp_watcher {
+	struct i915_vma *vma;
+	struct i915_request *rq;
+	u32 addr;
+	u32 *map;
+};
+
+static bool cmp_lt(u32 a, u32 b)
+{
+	return a < b;
+}
+
+static bool cmp_gte(u32 a, u32 b)
+{
+	return a >= b;
+}
+
+static int setup_watcher(struct hwsp_watcher *w, struct intel_gt *gt)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+
+	obj = i915_gem_object_create_internal(gt->i915, SZ_2M);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	w->map = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+	if (IS_ERR(w->map)) {
+		i915_gem_object_put(obj);
+		return PTR_ERR(w->map);
+	}
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, 0);
+	if (IS_ERR(vma)) {
+		i915_gem_object_put(obj);
+		return PTR_ERR(vma);
+	}
+
+	w->vma = vma;
+	w->addr = i915_ggtt_offset(vma);
+	return 0;
+}
+
+static void switch_tl_lock(struct i915_request *from, struct i915_request *to)
+{
+	/* some light mutex juggling required; think co-routines */
+
+	if (from) {
+		lockdep_unpin_lock(&from->context->timeline->mutex, from->cookie);
+		mutex_unlock(&from->context->timeline->mutex);
+	}
+
+	if (to) {
+		mutex_lock(&to->context->timeline->mutex);
+		to->cookie = lockdep_pin_lock(&to->context->timeline->mutex);
+	}
+}
+
+static int create_watcher(struct hwsp_watcher *w,
+			  struct intel_engine_cs *engine,
+			  int ringsz)
+{
+	struct intel_context *ce;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	ce->ring_size = ringsz;
+	w->rq = intel_context_create_request(ce);
+	intel_context_put(ce);
+	if (IS_ERR(w->rq))
+		return PTR_ERR(w->rq);
+
+	w->addr = i915_ggtt_offset(w->vma);
+
+	switch_tl_lock(w->rq, NULL);
+
+	return 0;
+}
+
+static int check_watcher(struct hwsp_watcher *w, const char *name,
+			 bool (*op)(u32 hwsp, u32 seqno))
+{
+	struct i915_request *rq = fetch_and_zero(&w->rq);
+	u32 offset, end;
+	int err;
+
+	GEM_BUG_ON(w->addr - i915_ggtt_offset(w->vma) > w->vma->size);
+
+	i915_request_get(rq);
+	switch_tl_lock(NULL, rq);
+	i915_request_add(rq);
+
+	if (i915_request_wait(rq, 0, HZ) < 0) {
+		err = -ETIME;
+		goto out;
+	}
+
+	err = 0;
+	offset = 0;
+	end = (w->addr - i915_ggtt_offset(w->vma)) / sizeof(*w->map);
+	while (offset < end) {
+		if (!op(w->map[offset + 1], w->map[offset])) {
+			pr_err("Watcher '%s' found HWSP value %x for seqno %x\n",
+			       name, w->map[offset + 1], w->map[offset]);
+			err = -EINVAL;
+		}
+
+		offset += 2;
+	}
+
+out:
+	i915_request_put(rq);
+	return err;
+}
+
+static void cleanup_watcher(struct hwsp_watcher *w)
+{
+	if (w->rq) {
+		switch_tl_lock(NULL, w->rq);
+
+		i915_request_add(w->rq);
+	}
+
+	i915_vma_unpin_and_release(&w->vma, I915_VMA_RELEASE_MAP);
+}
+
+static bool retire_requests(struct intel_timeline *tl)
+{
+	struct i915_request *rq, *rn;
+
+	mutex_lock(&tl->mutex);
+	list_for_each_entry_safe(rq, rn, &tl->requests, link)
+		if (!i915_request_retire(rq))
+			break;
+	mutex_unlock(&tl->mutex);
+
+	return !i915_active_fence_isset(&tl->last_request);
+}
+
+static struct i915_request *wrap_timeline(struct i915_request *rq)
+{
+	struct intel_context *ce = rq->context;
+	struct intel_timeline *tl = ce->timeline;
+	u32 seqno = rq->fence.seqno;
+
+	while (tl->seqno >= seqno) { /* Cause a wrap */
+		i915_request_put(rq);
+		rq = intel_context_create_request(ce);
+		if (IS_ERR(rq))
+			return rq;
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+	}
+
+	i915_request_put(rq);
+	rq = i915_request_create(ce);
+	if (IS_ERR(rq))
+		return rq;
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	return rq;
+}
+
+static int live_hwsp_read(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct hwsp_watcher watcher[2] = {};
+	struct intel_engine_cs *engine;
+	struct intel_timeline *tl;
+	enum intel_engine_id id;
+	int err = 0;
+	int i;
+
+	/*
+	 * If we take a reference to the HWSP for reading on the GPU, that
+	 * read may be arbitrarily delayed (either by foreign fence or
+	 * priority saturation) and a wrap can happen within 30 minutes.
+	 * When the GPU read is finally submitted it should be correct,
+	 * even across multiple wraps.
+	 */
+
+	if (GRAPHICS_VER(gt->i915) < 8) /* CS convenience [SRM/LRM] */
+		return 0;
+
+	tl = intel_timeline_create(gt);
+	if (IS_ERR(tl))
+		return PTR_ERR(tl);
+
+	if (!tl->has_initial_breadcrumb)
+		goto out_free;
+
+	for (i = 0; i < ARRAY_SIZE(watcher); i++) {
+		err = setup_watcher(&watcher[i], gt);
+		if (err)
+			goto out;
+	}
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		unsigned long count = 0;
+		IGT_TIMEOUT(end_time);
+
+		/* Create a request we can use for remote reading of the HWSP */
+		err = create_watcher(&watcher[1], engine, SZ_512K);
+		if (err)
+			goto out;
+
+		do {
+			struct i915_sw_fence *submit;
+			struct i915_request *rq;
+			u32 hwsp, dummy;
+
+			submit = heap_fence_create(GFP_KERNEL);
+			if (!submit) {
+				err = -ENOMEM;
+				goto out;
+			}
+
+			err = create_watcher(&watcher[0], engine, SZ_4K);
+			if (err)
+				goto out;
+
+			ce = intel_context_create(engine);
+			if (IS_ERR(ce)) {
+				err = PTR_ERR(ce);
+				goto out;
+			}
+
+			ce->timeline = intel_timeline_get(tl);
+
+			/* Ensure timeline is mapped, done during first pin */
+			err = intel_context_pin(ce);
+			if (err) {
+				intel_context_put(ce);
+				goto out;
+			}
+
+			/*
+			 * Start at a new wrap, and set seqno right before another wrap,
+			 * saving 30 minutes of nops
+			 */
+			tl->seqno = -12u + 2 * (count & 3);
+			__intel_timeline_get_seqno(tl, &dummy);
+
+			rq = i915_request_create(ce);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				intel_context_unpin(ce);
+				intel_context_put(ce);
+				goto out;
+			}
+
+			err = i915_sw_fence_await_dma_fence(&rq->submit,
+							    &watcher[0].rq->fence, 0,
+							    GFP_KERNEL);
+			if (err < 0) {
+				i915_request_add(rq);
+				intel_context_unpin(ce);
+				intel_context_put(ce);
+				goto out;
+			}
+
+			switch_tl_lock(rq, watcher[0].rq);
+			err = intel_timeline_read_hwsp(rq, watcher[0].rq, &hwsp);
+			if (err == 0)
+				err = emit_read_hwsp(watcher[0].rq, /* before */
+						     rq->fence.seqno, hwsp,
+						     &watcher[0].addr);
+			switch_tl_lock(watcher[0].rq, rq);
+			if (err) {
+				i915_request_add(rq);
+				intel_context_unpin(ce);
+				intel_context_put(ce);
+				goto out;
+			}
+
+			switch_tl_lock(rq, watcher[1].rq);
+			err = intel_timeline_read_hwsp(rq, watcher[1].rq, &hwsp);
+			if (err == 0)
+				err = emit_read_hwsp(watcher[1].rq, /* after */
+						     rq->fence.seqno, hwsp,
+						     &watcher[1].addr);
+			switch_tl_lock(watcher[1].rq, rq);
+			if (err) {
+				i915_request_add(rq);
+				intel_context_unpin(ce);
+				intel_context_put(ce);
+				goto out;
+			}
+
+			i915_request_get(rq);
+			i915_request_add(rq);
+
+			rq = wrap_timeline(rq);
+			intel_context_unpin(ce);
+			intel_context_put(ce);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				goto out;
+			}
+
+			err = i915_sw_fence_await_dma_fence(&watcher[1].rq->submit,
+							    &rq->fence, 0,
+							    GFP_KERNEL);
+			if (err < 0) {
+				i915_request_put(rq);
+				goto out;
+			}
+
+			err = check_watcher(&watcher[0], "before", cmp_lt);
+			i915_sw_fence_commit(submit);
+			heap_fence_put(submit);
+			if (err) {
+				i915_request_put(rq);
+				goto out;
+			}
+			count++;
+
+			/* Flush the timeline before manually wrapping again */
+			if (i915_request_wait(rq,
+					      I915_WAIT_INTERRUPTIBLE,
+					      HZ) < 0) {
+				err = -ETIME;
+				i915_request_put(rq);
+				goto out;
+			}
+			retire_requests(tl);
+			i915_request_put(rq);
+
+			/* Single requests are limited to half a ring at most */
+			if (8 * watcher[1].rq->ring->emit >
+			    3 * watcher[1].rq->ring->size)
+				break;
+
+		} while (!__igt_timeout(end_time, NULL) &&
+			 count < (PAGE_SIZE / TIMELINE_SEQNO_BYTES - 1) / 2);
+
+		pr_info("%s: simulated %lu wraps\n", engine->name, count);
+		err = check_watcher(&watcher[1], "after", cmp_gte);
+		if (err)
+			goto out;
+	}
+
+out:
+	for (i = 0; i < ARRAY_SIZE(watcher); i++)
+		cleanup_watcher(&watcher[i]);
+
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+out_free:
+	intel_timeline_put(tl);
+	return err;
+}
+
+static int live_hwsp_rollover_kernel(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Run the host for long enough, and even the kernel context will
+	 * see a seqno rollover.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce = engine->kernel_context;
+		struct intel_timeline *tl = ce->timeline;
+		struct i915_request *rq[3] = {};
+		int i;
+
+		st_engine_heartbeat_disable(engine);
+		if (intel_gt_wait_for_idle(gt, HZ / 2)) {
+			err = -EIO;
+			goto out;
+		}
+
+		GEM_BUG_ON(i915_active_fence_isset(&tl->last_request));
+		tl->seqno = -2u;
+		WRITE_ONCE(*(u32 *)tl->hwsp_seqno, tl->seqno);
+
+		for (i = 0; i < ARRAY_SIZE(rq); i++) {
+			struct i915_request *this;
+
+			this = i915_request_create(ce);
+			if (IS_ERR(this)) {
+				err = PTR_ERR(this);
+				goto out;
+			}
+
+			pr_debug("%s: create fence.seqnp:%d\n",
+				 engine->name,
+				 lower_32_bits(this->fence.seqno));
+
+			GEM_BUG_ON(rcu_access_pointer(this->timeline) != tl);
+
+			rq[i] = i915_request_get(this);
+			i915_request_add(this);
+		}
+
+		/* We expected a wrap! */
+		GEM_BUG_ON(rq[2]->fence.seqno > rq[0]->fence.seqno);
+
+		if (i915_request_wait(rq[2], 0, HZ / 5) < 0) {
+			pr_err("Wait for timeline wrap timed out!\n");
+			err = -EIO;
+			goto out;
+		}
+
+		for (i = 0; i < ARRAY_SIZE(rq); i++) {
+			if (!i915_request_completed(rq[i])) {
+				pr_err("Pre-wrap request not completed!\n");
+				err = -EINVAL;
+				goto out;
+			}
+		}
+
+out:
+		for (i = 0; i < ARRAY_SIZE(rq); i++)
+			i915_request_put(rq[i]);
+		st_engine_heartbeat_enable(engine);
+		if (err)
+			break;
+	}
+
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	return err;
+}
+
+static int live_hwsp_rollover_user(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/*
+	 * Simulate a long running user context, and force the seqno wrap
+	 * on the user's timeline.
+	 */
+
+	for_each_engine(engine, gt, id) {
+		struct i915_request *rq[3] = {};
+		struct intel_timeline *tl;
+		struct intel_context *ce;
+		int i;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce))
+			return PTR_ERR(ce);
+
+		err = intel_context_alloc_state(ce);
+		if (err)
+			goto out;
+
+		tl = ce->timeline;
+		if (!tl->has_initial_breadcrumb)
+			goto out;
+
+		err = intel_context_pin(ce);
+		if (err)
+			goto out;
+
+		tl->seqno = -4u;
+		WRITE_ONCE(*(u32 *)tl->hwsp_seqno, tl->seqno);
+
+		for (i = 0; i < ARRAY_SIZE(rq); i++) {
+			struct i915_request *this;
+
+			this = intel_context_create_request(ce);
+			if (IS_ERR(this)) {
+				err = PTR_ERR(this);
+				goto out_unpin;
+			}
+
+			pr_debug("%s: create fence.seqnp:%d\n",
+				 engine->name,
+				 lower_32_bits(this->fence.seqno));
+
+			GEM_BUG_ON(rcu_access_pointer(this->timeline) != tl);
+
+			rq[i] = i915_request_get(this);
+			i915_request_add(this);
+		}
+
+		/* We expected a wrap! */
+		GEM_BUG_ON(rq[2]->fence.seqno > rq[0]->fence.seqno);
+
+		if (i915_request_wait(rq[2], 0, HZ / 5) < 0) {
+			pr_err("Wait for timeline wrap timed out!\n");
+			err = -EIO;
+			goto out_unpin;
+		}
+
+		for (i = 0; i < ARRAY_SIZE(rq); i++) {
+			if (!i915_request_completed(rq[i])) {
+				pr_err("Pre-wrap request not completed!\n");
+				err = -EINVAL;
+				goto out_unpin;
+			}
+		}
+out_unpin:
+		intel_context_unpin(ce);
+out:
+		for (i = 0; i < ARRAY_SIZE(rq); i++)
+			i915_request_put(rq[i]);
+		intel_context_put(ce);
+		if (err)
+			break;
+	}
+
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	return err;
+}
+
+static int live_hwsp_recycle(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	unsigned long count;
+	int err = 0;
+
+	/*
+	 * Check seqno writes into one timeline at a time. We expect to
+	 * recycle the breadcrumb slot between iterations and neither
+	 * want to confuse ourselves or the GPU.
+	 */
+
+	count = 0;
+	for_each_engine(engine, gt, id) {
+		IGT_TIMEOUT(end_time);
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		intel_engine_pm_get(engine);
+
+		do {
+			struct intel_timeline *tl;
+			struct i915_request *rq;
+
+			tl = intel_timeline_create(gt);
+			if (IS_ERR(tl)) {
+				err = PTR_ERR(tl);
+				break;
+			}
+
+			rq = checked_tl_write(tl, engine, count);
+			if (IS_ERR(rq)) {
+				intel_timeline_put(tl);
+				err = PTR_ERR(rq);
+				break;
+			}
+
+			if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+				pr_err("Wait for timeline writes timed out!\n");
+				i915_request_put(rq);
+				intel_timeline_put(tl);
+				err = -EIO;
+				break;
+			}
+
+			if (READ_ONCE(*tl->hwsp_seqno) != count) {
+				GEM_TRACE_ERR("Invalid seqno:%lu stored in timeline %llu @ %x found 0x%x\n",
+					      count, tl->fence_context,
+					      tl->hwsp_offset, *tl->hwsp_seqno);
+				GEM_TRACE_DUMP();
+				err = -EINVAL;
+			}
+
+			i915_request_put(rq);
+			intel_timeline_put(tl);
+			count++;
+
+			if (err)
+				break;
+		} while (!__igt_timeout(end_time, NULL));
+
+		intel_engine_pm_put(engine);
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+int intel_timeline_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_hwsp_recycle),
+		SUBTEST(live_hwsp_engine),
+		SUBTEST(live_hwsp_alternate),
+		SUBTEST(live_hwsp_wrap),
+		SUBTEST(live_hwsp_read),
+		SUBTEST(live_hwsp_rollover_kernel),
+		SUBTEST(live_hwsp_rollover_user),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
diff --git a/drivers/gpu/drm/i915/gt/selftest_workarounds.c b/drivers/gpu/drm/i915/gt/selftest_workarounds.c
new file mode 100644
index 000000000..67a9aab80
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_workarounds.c
@@ -0,0 +1,1395 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_engine_user.h"
+#include "gt/intel_gt.h"
+#include "i915_selftest.h"
+#include "intel_reset.h"
+
+#include "selftests/igt_flush_test.h"
+#include "selftests/igt_reset.h"
+#include "selftests/igt_spinner.h"
+#include "selftests/intel_scheduler_helpers.h"
+#include "selftests/mock_drm.h"
+
+#include "gem/selftests/igt_gem_utils.h"
+#include "gem/selftests/mock_context.h"
+
+static const struct wo_register {
+	enum intel_platform platform;
+	u32 reg;
+} wo_registers[] = {
+	{ INTEL_GEMINILAKE, 0x731c }
+};
+
+struct wa_lists {
+	struct i915_wa_list gt_wa_list;
+	struct {
+		struct i915_wa_list wa_list;
+		struct i915_wa_list ctx_wa_list;
+	} engine[I915_NUM_ENGINES];
+};
+
+static int request_add_sync(struct i915_request *rq, int err)
+{
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (i915_request_wait(rq, 0, HZ / 5) < 0)
+		err = -EIO;
+	i915_request_put(rq);
+
+	return err;
+}
+
+static int request_add_spin(struct i915_request *rq, struct igt_spinner *spin)
+{
+	int err = 0;
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (spin && !igt_wait_for_spinner(spin, rq))
+		err = -ETIMEDOUT;
+	i915_request_put(rq);
+
+	return err;
+}
+
+static void
+reference_lists_init(struct intel_gt *gt, struct wa_lists *lists)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	memset(lists, 0, sizeof(*lists));
+
+	wa_init_start(&lists->gt_wa_list, "GT_REF", "global");
+	gt_init_workarounds(gt, &lists->gt_wa_list);
+	wa_init_finish(&lists->gt_wa_list);
+
+	for_each_engine(engine, gt, id) {
+		struct i915_wa_list *wal = &lists->engine[id].wa_list;
+
+		wa_init_start(wal, "REF", engine->name);
+		engine_init_workarounds(engine, wal);
+		wa_init_finish(wal);
+
+		__intel_engine_init_ctx_wa(engine,
+					   &lists->engine[id].ctx_wa_list,
+					   "CTX_REF");
+	}
+}
+
+static void
+reference_lists_fini(struct intel_gt *gt, struct wa_lists *lists)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id)
+		intel_wa_list_free(&lists->engine[id].wa_list);
+
+	intel_wa_list_free(&lists->gt_wa_list);
+}
+
+static struct drm_i915_gem_object *
+read_nonprivs(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	const u32 base = engine->mmio_base;
+	struct drm_i915_gem_object *result;
+	struct i915_request *rq;
+	struct i915_vma *vma;
+	u32 srm, *cs;
+	int err;
+	int i;
+
+	result = i915_gem_object_create_internal(engine->i915, PAGE_SIZE);
+	if (IS_ERR(result))
+		return result;
+
+	i915_gem_object_set_cache_coherency(result, I915_CACHE_LLC);
+
+	cs = i915_gem_object_pin_map_unlocked(result, I915_MAP_WB);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_obj;
+	}
+	memset(cs, 0xc5, PAGE_SIZE);
+	i915_gem_object_flush_map(result);
+	i915_gem_object_unpin_map(result);
+
+	vma = i915_vma_instance(result, &engine->gt->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+	if (err)
+		goto err_obj;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_pin;
+	}
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj, true);
+	if (err == 0)
+		err = i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE);
+	i915_vma_unlock(vma);
+	if (err)
+		goto err_req;
+
+	srm = MI_STORE_REGISTER_MEM | MI_SRM_LRM_GLOBAL_GTT;
+	if (GRAPHICS_VER(engine->i915) >= 8)
+		srm++;
+
+	cs = intel_ring_begin(rq, 4 * RING_MAX_NONPRIV_SLOTS);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_req;
+	}
+
+	for (i = 0; i < RING_MAX_NONPRIV_SLOTS; i++) {
+		*cs++ = srm;
+		*cs++ = i915_mmio_reg_offset(RING_FORCE_TO_NONPRIV(base, i));
+		*cs++ = i915_ggtt_offset(vma) + sizeof(u32) * i;
+		*cs++ = 0;
+	}
+	intel_ring_advance(rq, cs);
+
+	i915_request_add(rq);
+	i915_vma_unpin(vma);
+
+	return result;
+
+err_req:
+	i915_request_add(rq);
+err_pin:
+	i915_vma_unpin(vma);
+err_obj:
+	i915_gem_object_put(result);
+	return ERR_PTR(err);
+}
+
+static u32
+get_whitelist_reg(const struct intel_engine_cs *engine, unsigned int i)
+{
+	i915_reg_t reg = i < engine->whitelist.count ?
+			 engine->whitelist.list[i].reg :
+			 RING_NOPID(engine->mmio_base);
+
+	return i915_mmio_reg_offset(reg);
+}
+
+static void
+print_results(const struct intel_engine_cs *engine, const u32 *results)
+{
+	unsigned int i;
+
+	for (i = 0; i < RING_MAX_NONPRIV_SLOTS; i++) {
+		u32 expected = get_whitelist_reg(engine, i);
+		u32 actual = results[i];
+
+		pr_info("RING_NONPRIV[%d]: expected 0x%08x, found 0x%08x\n",
+			i, expected, actual);
+	}
+}
+
+static int check_whitelist(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct drm_i915_gem_object *results;
+	struct intel_wedge_me wedge;
+	u32 *vaddr;
+	int err;
+	int i;
+
+	results = read_nonprivs(ce);
+	if (IS_ERR(results))
+		return PTR_ERR(results);
+
+	err = 0;
+	i915_gem_object_lock(results, NULL);
+	intel_wedge_on_timeout(&wedge, engine->gt, HZ / 5) /* safety net! */
+		err = i915_gem_object_set_to_cpu_domain(results, false);
+
+	if (intel_gt_is_wedged(engine->gt))
+		err = -EIO;
+	if (err)
+		goto out_put;
+
+	vaddr = i915_gem_object_pin_map(results, I915_MAP_WB);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto out_put;
+	}
+
+	for (i = 0; i < RING_MAX_NONPRIV_SLOTS; i++) {
+		u32 expected = get_whitelist_reg(engine, i);
+		u32 actual = vaddr[i];
+
+		if (expected != actual) {
+			print_results(engine, vaddr);
+			pr_err("Invalid RING_NONPRIV[%d], expected 0x%08x, found 0x%08x\n",
+			       i, expected, actual);
+
+			err = -EINVAL;
+			break;
+		}
+	}
+
+	i915_gem_object_unpin_map(results);
+out_put:
+	i915_gem_object_unlock(results);
+	i915_gem_object_put(results);
+	return err;
+}
+
+static int do_device_reset(struct intel_engine_cs *engine)
+{
+	intel_gt_reset(engine->gt, engine->mask, "live_workarounds");
+	return 0;
+}
+
+static int do_engine_reset(struct intel_engine_cs *engine)
+{
+	return intel_engine_reset(engine, "live_workarounds");
+}
+
+static int do_guc_reset(struct intel_engine_cs *engine)
+{
+	/* Currently a no-op as the reset is handled by GuC */
+	return 0;
+}
+
+static int
+switch_to_scratch_context(struct intel_engine_cs *engine,
+			  struct igt_spinner *spin,
+			  struct i915_request **rq)
+{
+	struct intel_context *ce;
+	int err = 0;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	*rq = igt_spinner_create_request(spin, ce, MI_NOOP);
+	intel_context_put(ce);
+
+	if (IS_ERR(*rq)) {
+		spin = NULL;
+		err = PTR_ERR(*rq);
+		goto err;
+	}
+
+	err = request_add_spin(*rq, spin);
+err:
+	if (err && spin)
+		igt_spinner_end(spin);
+
+	return err;
+}
+
+static int check_whitelist_across_reset(struct intel_engine_cs *engine,
+					int (*reset)(struct intel_engine_cs *),
+					const char *name)
+{
+	struct intel_context *ce, *tmp;
+	struct igt_spinner spin;
+	struct i915_request *rq;
+	intel_wakeref_t wakeref;
+	int err;
+
+	pr_info("Checking %d whitelisted registers on %s (RING_NONPRIV) [%s]\n",
+		engine->whitelist.count, engine->name, name);
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	err = igt_spinner_init(&spin, engine->gt);
+	if (err)
+		goto out_ctx;
+
+	err = check_whitelist(ce);
+	if (err) {
+		pr_err("Invalid whitelist *before* %s reset!\n", name);
+		goto out_spin;
+	}
+
+	err = switch_to_scratch_context(engine, &spin, &rq);
+	if (err)
+		goto out_spin;
+
+	/* Ensure the spinner hasn't aborted */
+	if (i915_request_completed(rq)) {
+		pr_err("%s spinner failed to start\n", name);
+		err = -ETIMEDOUT;
+		goto out_spin;
+	}
+
+	with_intel_runtime_pm(engine->uncore->rpm, wakeref)
+		err = reset(engine);
+
+	/* Ensure the reset happens and kills the engine */
+	if (err == 0)
+		err = intel_selftest_wait_for_rq(rq);
+
+	igt_spinner_end(&spin);
+
+	if (err) {
+		pr_err("%s reset failed\n", name);
+		goto out_spin;
+	}
+
+	err = check_whitelist(ce);
+	if (err) {
+		pr_err("Whitelist not preserved in context across %s reset!\n",
+		       name);
+		goto out_spin;
+	}
+
+	tmp = intel_context_create(engine);
+	if (IS_ERR(tmp)) {
+		err = PTR_ERR(tmp);
+		goto out_spin;
+	}
+	intel_context_put(ce);
+	ce = tmp;
+
+	err = check_whitelist(ce);
+	if (err) {
+		pr_err("Invalid whitelist *after* %s reset in fresh context!\n",
+		       name);
+		goto out_spin;
+	}
+
+out_spin:
+	igt_spinner_fini(&spin);
+out_ctx:
+	intel_context_put(ce);
+	return err;
+}
+
+static struct i915_vma *create_batch(struct i915_address_space *vm)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err;
+
+	obj = i915_gem_object_create_internal(vm->i915, 16 * PAGE_SIZE);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto err_obj;
+
+	return vma;
+
+err_obj:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static u32 reg_write(u32 old, u32 new, u32 rsvd)
+{
+	if (rsvd == 0x0000ffff) {
+		old &= ~(new >> 16);
+		old |= new & (new >> 16);
+	} else {
+		old &= ~rsvd;
+		old |= new & rsvd;
+	}
+
+	return old;
+}
+
+static bool wo_register(struct intel_engine_cs *engine, u32 reg)
+{
+	enum intel_platform platform = INTEL_INFO(engine->i915)->platform;
+	int i;
+
+	if ((reg & RING_FORCE_TO_NONPRIV_ACCESS_MASK) ==
+	     RING_FORCE_TO_NONPRIV_ACCESS_WR)
+		return true;
+
+	for (i = 0; i < ARRAY_SIZE(wo_registers); i++) {
+		if (wo_registers[i].platform == platform &&
+		    wo_registers[i].reg == reg)
+			return true;
+	}
+
+	return false;
+}
+
+static bool timestamp(const struct intel_engine_cs *engine, u32 reg)
+{
+	reg = (reg - engine->mmio_base) & ~RING_FORCE_TO_NONPRIV_ACCESS_MASK;
+	switch (reg) {
+	case 0x358:
+	case 0x35c:
+	case 0x3a8:
+		return true;
+
+	default:
+		return false;
+	}
+}
+
+static bool ro_register(u32 reg)
+{
+	if ((reg & RING_FORCE_TO_NONPRIV_ACCESS_MASK) ==
+	     RING_FORCE_TO_NONPRIV_ACCESS_RD)
+		return true;
+
+	return false;
+}
+
+static int whitelist_writable_count(struct intel_engine_cs *engine)
+{
+	int count = engine->whitelist.count;
+	int i;
+
+	for (i = 0; i < engine->whitelist.count; i++) {
+		u32 reg = i915_mmio_reg_offset(engine->whitelist.list[i].reg);
+
+		if (ro_register(reg))
+			count--;
+	}
+
+	return count;
+}
+
+static int check_dirty_whitelist(struct intel_context *ce)
+{
+	const u32 values[] = {
+		0x00000000,
+		0x01010101,
+		0x10100101,
+		0x03030303,
+		0x30300303,
+		0x05050505,
+		0x50500505,
+		0x0f0f0f0f,
+		0xf00ff00f,
+		0x10101010,
+		0xf0f01010,
+		0x30303030,
+		0xa0a03030,
+		0x50505050,
+		0xc0c05050,
+		0xf0f0f0f0,
+		0x11111111,
+		0x33333333,
+		0x55555555,
+		0x0000ffff,
+		0x00ff00ff,
+		0xff0000ff,
+		0xffff00ff,
+		0xffffffff,
+	};
+	struct intel_engine_cs *engine = ce->engine;
+	struct i915_vma *scratch;
+	struct i915_vma *batch;
+	int err = 0, i, v, sz;
+	u32 *cs, *results;
+
+	sz = (2 * ARRAY_SIZE(values) + 1) * sizeof(u32);
+	scratch = __vm_create_scratch_for_read_pinned(ce->vm, sz);
+	if (IS_ERR(scratch))
+		return PTR_ERR(scratch);
+
+	batch = create_batch(ce->vm);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		goto out_scratch;
+	}
+
+	for (i = 0; i < engine->whitelist.count; i++) {
+		u32 reg = i915_mmio_reg_offset(engine->whitelist.list[i].reg);
+		struct i915_gem_ww_ctx ww;
+		u64 addr = scratch->node.start;
+		struct i915_request *rq;
+		u32 srm, lrm, rsvd;
+		u32 expect;
+		int idx;
+		bool ro_reg;
+
+		if (wo_register(engine, reg))
+			continue;
+
+		if (timestamp(engine, reg))
+			continue; /* timestamps are expected to autoincrement */
+
+		ro_reg = ro_register(reg);
+
+		i915_gem_ww_ctx_init(&ww, false);
+retry:
+		cs = NULL;
+		err = i915_gem_object_lock(scratch->obj, &ww);
+		if (!err)
+			err = i915_gem_object_lock(batch->obj, &ww);
+		if (!err)
+			err = intel_context_pin_ww(ce, &ww);
+		if (err)
+			goto out;
+
+		cs = i915_gem_object_pin_map(batch->obj, I915_MAP_WC);
+		if (IS_ERR(cs)) {
+			err = PTR_ERR(cs);
+			goto out_ctx;
+		}
+
+		results = i915_gem_object_pin_map(scratch->obj, I915_MAP_WB);
+		if (IS_ERR(results)) {
+			err = PTR_ERR(results);
+			goto out_unmap_batch;
+		}
+
+		/* Clear non priv flags */
+		reg &= RING_FORCE_TO_NONPRIV_ADDRESS_MASK;
+
+		srm = MI_STORE_REGISTER_MEM;
+		lrm = MI_LOAD_REGISTER_MEM;
+		if (GRAPHICS_VER(engine->i915) >= 8)
+			lrm++, srm++;
+
+		pr_debug("%s: Writing garbage to %x\n",
+			 engine->name, reg);
+
+		/* SRM original */
+		*cs++ = srm;
+		*cs++ = reg;
+		*cs++ = lower_32_bits(addr);
+		*cs++ = upper_32_bits(addr);
+
+		idx = 1;
+		for (v = 0; v < ARRAY_SIZE(values); v++) {
+			/* LRI garbage */
+			*cs++ = MI_LOAD_REGISTER_IMM(1);
+			*cs++ = reg;
+			*cs++ = values[v];
+
+			/* SRM result */
+			*cs++ = srm;
+			*cs++ = reg;
+			*cs++ = lower_32_bits(addr + sizeof(u32) * idx);
+			*cs++ = upper_32_bits(addr + sizeof(u32) * idx);
+			idx++;
+		}
+		for (v = 0; v < ARRAY_SIZE(values); v++) {
+			/* LRI garbage */
+			*cs++ = MI_LOAD_REGISTER_IMM(1);
+			*cs++ = reg;
+			*cs++ = ~values[v];
+
+			/* SRM result */
+			*cs++ = srm;
+			*cs++ = reg;
+			*cs++ = lower_32_bits(addr + sizeof(u32) * idx);
+			*cs++ = upper_32_bits(addr + sizeof(u32) * idx);
+			idx++;
+		}
+		GEM_BUG_ON(idx * sizeof(u32) > scratch->size);
+
+		/* LRM original -- don't leave garbage in the context! */
+		*cs++ = lrm;
+		*cs++ = reg;
+		*cs++ = lower_32_bits(addr);
+		*cs++ = upper_32_bits(addr);
+
+		*cs++ = MI_BATCH_BUFFER_END;
+
+		i915_gem_object_flush_map(batch->obj);
+		i915_gem_object_unpin_map(batch->obj);
+		intel_gt_chipset_flush(engine->gt);
+		cs = NULL;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out_unmap_scratch;
+		}
+
+		if (engine->emit_init_breadcrumb) { /* Be nice if we hang */
+			err = engine->emit_init_breadcrumb(rq);
+			if (err)
+				goto err_request;
+		}
+
+		err = i915_request_await_object(rq, batch->obj, false);
+		if (err == 0)
+			err = i915_vma_move_to_active(batch, rq, 0);
+		if (err)
+			goto err_request;
+
+		err = i915_request_await_object(rq, scratch->obj, true);
+		if (err == 0)
+			err = i915_vma_move_to_active(scratch, rq,
+						      EXEC_OBJECT_WRITE);
+		if (err)
+			goto err_request;
+
+		err = engine->emit_bb_start(rq,
+					    batch->node.start, PAGE_SIZE,
+					    0);
+		if (err)
+			goto err_request;
+
+err_request:
+		err = request_add_sync(rq, err);
+		if (err) {
+			pr_err("%s: Futzing %x timedout; cancelling test\n",
+			       engine->name, reg);
+			intel_gt_set_wedged(engine->gt);
+			goto out_unmap_scratch;
+		}
+
+		GEM_BUG_ON(values[ARRAY_SIZE(values) - 1] != 0xffffffff);
+		if (!ro_reg) {
+			/* detect write masking */
+			rsvd = results[ARRAY_SIZE(values)];
+			if (!rsvd) {
+				pr_err("%s: Unable to write to whitelisted register %x\n",
+				       engine->name, reg);
+				err = -EINVAL;
+				goto out_unmap_scratch;
+			}
+		} else {
+			rsvd = 0;
+		}
+
+		expect = results[0];
+		idx = 1;
+		for (v = 0; v < ARRAY_SIZE(values); v++) {
+			if (ro_reg)
+				expect = results[0];
+			else
+				expect = reg_write(expect, values[v], rsvd);
+
+			if (results[idx] != expect)
+				err++;
+			idx++;
+		}
+		for (v = 0; v < ARRAY_SIZE(values); v++) {
+			if (ro_reg)
+				expect = results[0];
+			else
+				expect = reg_write(expect, ~values[v], rsvd);
+
+			if (results[idx] != expect)
+				err++;
+			idx++;
+		}
+		if (err) {
+			pr_err("%s: %d mismatch between values written to whitelisted register [%x], and values read back!\n",
+			       engine->name, err, reg);
+
+			if (ro_reg)
+				pr_info("%s: Whitelisted read-only register: %x, original value %08x\n",
+					engine->name, reg, results[0]);
+			else
+				pr_info("%s: Whitelisted register: %x, original value %08x, rsvd %08x\n",
+					engine->name, reg, results[0], rsvd);
+
+			expect = results[0];
+			idx = 1;
+			for (v = 0; v < ARRAY_SIZE(values); v++) {
+				u32 w = values[v];
+
+				if (ro_reg)
+					expect = results[0];
+				else
+					expect = reg_write(expect, w, rsvd);
+				pr_info("Wrote %08x, read %08x, expect %08x\n",
+					w, results[idx], expect);
+				idx++;
+			}
+			for (v = 0; v < ARRAY_SIZE(values); v++) {
+				u32 w = ~values[v];
+
+				if (ro_reg)
+					expect = results[0];
+				else
+					expect = reg_write(expect, w, rsvd);
+				pr_info("Wrote %08x, read %08x, expect %08x\n",
+					w, results[idx], expect);
+				idx++;
+			}
+
+			err = -EINVAL;
+		}
+out_unmap_scratch:
+		i915_gem_object_unpin_map(scratch->obj);
+out_unmap_batch:
+		if (cs)
+			i915_gem_object_unpin_map(batch->obj);
+out_ctx:
+		intel_context_unpin(ce);
+out:
+		if (err == -EDEADLK) {
+			err = i915_gem_ww_ctx_backoff(&ww);
+			if (!err)
+				goto retry;
+		}
+		i915_gem_ww_ctx_fini(&ww);
+		if (err)
+			break;
+	}
+
+	if (igt_flush_test(engine->i915))
+		err = -EIO;
+
+	i915_vma_unpin_and_release(&batch, 0);
+out_scratch:
+	i915_vma_unpin_and_release(&scratch, 0);
+	return err;
+}
+
+static int live_dirty_whitelist(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/* Can the user write to the whitelisted registers? */
+
+	if (GRAPHICS_VER(gt->i915) < 7) /* minimum requirement for LRI, SRM, LRM */
+		return 0;
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+		int err;
+
+		if (engine->whitelist.count == 0)
+			continue;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce))
+			return PTR_ERR(ce);
+
+		err = check_dirty_whitelist(ce);
+		intel_context_put(ce);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int live_reset_whitelist(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err = 0;
+
+	/* If we reset the gpu, we should not lose the RING_NONPRIV */
+	igt_global_reset_lock(gt);
+
+	for_each_engine(engine, gt, id) {
+		if (engine->whitelist.count == 0)
+			continue;
+
+		if (intel_has_reset_engine(gt)) {
+			if (intel_engine_uses_guc(engine)) {
+				struct intel_selftest_saved_policy saved;
+				int err2;
+
+				err = intel_selftest_modify_policy(engine, &saved,
+								   SELFTEST_SCHEDULER_MODIFY_FAST_RESET);
+				if (err)
+					goto out;
+
+				err = check_whitelist_across_reset(engine,
+								   do_guc_reset,
+								   "guc");
+
+				err2 = intel_selftest_restore_policy(engine, &saved);
+				if (err == 0)
+					err = err2;
+			} else {
+				err = check_whitelist_across_reset(engine,
+								   do_engine_reset,
+								   "engine");
+			}
+
+			if (err)
+				goto out;
+		}
+
+		if (intel_has_gpu_reset(gt)) {
+			err = check_whitelist_across_reset(engine,
+							   do_device_reset,
+							   "device");
+			if (err)
+				goto out;
+		}
+	}
+
+out:
+	igt_global_reset_unlock(gt);
+	return err;
+}
+
+static int read_whitelisted_registers(struct intel_context *ce,
+				      struct i915_vma *results)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct i915_request *rq;
+	int i, err = 0;
+	u32 srm, *cs;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	i915_vma_lock(results);
+	err = i915_request_await_object(rq, results->obj, true);
+	if (err == 0)
+		err = i915_vma_move_to_active(results, rq, EXEC_OBJECT_WRITE);
+	i915_vma_unlock(results);
+	if (err)
+		goto err_req;
+
+	srm = MI_STORE_REGISTER_MEM;
+	if (GRAPHICS_VER(engine->i915) >= 8)
+		srm++;
+
+	cs = intel_ring_begin(rq, 4 * engine->whitelist.count);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_req;
+	}
+
+	for (i = 0; i < engine->whitelist.count; i++) {
+		u64 offset = results->node.start + sizeof(u32) * i;
+		u32 reg = i915_mmio_reg_offset(engine->whitelist.list[i].reg);
+
+		/* Clear non priv flags */
+		reg &= RING_FORCE_TO_NONPRIV_ADDRESS_MASK;
+
+		*cs++ = srm;
+		*cs++ = reg;
+		*cs++ = lower_32_bits(offset);
+		*cs++ = upper_32_bits(offset);
+	}
+	intel_ring_advance(rq, cs);
+
+err_req:
+	return request_add_sync(rq, err);
+}
+
+static int scrub_whitelisted_registers(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct i915_request *rq;
+	struct i915_vma *batch;
+	int i, err = 0;
+	u32 *cs;
+
+	batch = create_batch(ce->vm);
+	if (IS_ERR(batch))
+		return PTR_ERR(batch);
+
+	cs = i915_gem_object_pin_map_unlocked(batch->obj, I915_MAP_WC);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto err_batch;
+	}
+
+	*cs++ = MI_LOAD_REGISTER_IMM(whitelist_writable_count(engine));
+	for (i = 0; i < engine->whitelist.count; i++) {
+		u32 reg = i915_mmio_reg_offset(engine->whitelist.list[i].reg);
+
+		if (ro_register(reg))
+			continue;
+
+		/* Clear non priv flags */
+		reg &= RING_FORCE_TO_NONPRIV_ADDRESS_MASK;
+
+		*cs++ = reg;
+		*cs++ = 0xffffffff;
+	}
+	*cs++ = MI_BATCH_BUFFER_END;
+
+	i915_gem_object_flush_map(batch->obj);
+	intel_gt_chipset_flush(engine->gt);
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_unpin;
+	}
+
+	if (engine->emit_init_breadcrumb) { /* Be nice if we hang */
+		err = engine->emit_init_breadcrumb(rq);
+		if (err)
+			goto err_request;
+	}
+
+	i915_vma_lock(batch);
+	err = i915_request_await_object(rq, batch->obj, false);
+	if (err == 0)
+		err = i915_vma_move_to_active(batch, rq, 0);
+	i915_vma_unlock(batch);
+	if (err)
+		goto err_request;
+
+	/* Perform the writes from an unprivileged "user" batch */
+	err = engine->emit_bb_start(rq, batch->node.start, 0, 0);
+
+err_request:
+	err = request_add_sync(rq, err);
+
+err_unpin:
+	i915_gem_object_unpin_map(batch->obj);
+err_batch:
+	i915_vma_unpin_and_release(&batch, 0);
+	return err;
+}
+
+struct regmask {
+	i915_reg_t reg;
+	u8 graphics_ver;
+};
+
+static bool find_reg(struct drm_i915_private *i915,
+		     i915_reg_t reg,
+		     const struct regmask *tbl,
+		     unsigned long count)
+{
+	u32 offset = i915_mmio_reg_offset(reg);
+
+	while (count--) {
+		if (GRAPHICS_VER(i915) == tbl->graphics_ver &&
+		    i915_mmio_reg_offset(tbl->reg) == offset)
+			return true;
+		tbl++;
+	}
+
+	return false;
+}
+
+static bool pardon_reg(struct drm_i915_private *i915, i915_reg_t reg)
+{
+	/* Alas, we must pardon some whitelists. Mistakes already made */
+	static const struct regmask pardon[] = {
+		{ GEN9_CTX_PREEMPT_REG, 9 },
+		{ GEN8_L3SQCREG4, 9 },
+	};
+
+	return find_reg(i915, reg, pardon, ARRAY_SIZE(pardon));
+}
+
+static bool result_eq(struct intel_engine_cs *engine,
+		      u32 a, u32 b, i915_reg_t reg)
+{
+	if (a != b && !pardon_reg(engine->i915, reg)) {
+		pr_err("Whitelisted register 0x%4x not context saved: A=%08x, B=%08x\n",
+		       i915_mmio_reg_offset(reg), a, b);
+		return false;
+	}
+
+	return true;
+}
+
+static bool writeonly_reg(struct drm_i915_private *i915, i915_reg_t reg)
+{
+	/* Some registers do not seem to behave and our writes unreadable */
+	static const struct regmask wo[] = {
+		{ GEN9_SLICE_COMMON_ECO_CHICKEN1, 9 },
+	};
+
+	return find_reg(i915, reg, wo, ARRAY_SIZE(wo));
+}
+
+static bool result_neq(struct intel_engine_cs *engine,
+		       u32 a, u32 b, i915_reg_t reg)
+{
+	if (a == b && !writeonly_reg(engine->i915, reg)) {
+		pr_err("Whitelist register 0x%4x:%08x was unwritable\n",
+		       i915_mmio_reg_offset(reg), a);
+		return false;
+	}
+
+	return true;
+}
+
+static int
+check_whitelisted_registers(struct intel_engine_cs *engine,
+			    struct i915_vma *A,
+			    struct i915_vma *B,
+			    bool (*fn)(struct intel_engine_cs *engine,
+				       u32 a, u32 b,
+				       i915_reg_t reg))
+{
+	u32 *a, *b;
+	int i, err;
+
+	a = i915_gem_object_pin_map_unlocked(A->obj, I915_MAP_WB);
+	if (IS_ERR(a))
+		return PTR_ERR(a);
+
+	b = i915_gem_object_pin_map_unlocked(B->obj, I915_MAP_WB);
+	if (IS_ERR(b)) {
+		err = PTR_ERR(b);
+		goto err_a;
+	}
+
+	err = 0;
+	for (i = 0; i < engine->whitelist.count; i++) {
+		const struct i915_wa *wa = &engine->whitelist.list[i];
+
+		if (i915_mmio_reg_offset(wa->reg) &
+		    RING_FORCE_TO_NONPRIV_ACCESS_RD)
+			continue;
+
+		if (!fn(engine, a[i], b[i], wa->reg))
+			err = -EINVAL;
+	}
+
+	i915_gem_object_unpin_map(B->obj);
+err_a:
+	i915_gem_object_unpin_map(A->obj);
+	return err;
+}
+
+static int live_isolated_whitelist(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct {
+		struct i915_vma *scratch[2];
+	} client[2] = {};
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int i, err = 0;
+
+	/*
+	 * Check that a write into a whitelist register works, but
+	 * invisible to a second context.
+	 */
+
+	if (!intel_engines_has_context_isolation(gt->i915))
+		return 0;
+
+	for (i = 0; i < ARRAY_SIZE(client); i++) {
+		client[i].scratch[0] =
+			__vm_create_scratch_for_read_pinned(gt->vm, 4096);
+		if (IS_ERR(client[i].scratch[0])) {
+			err = PTR_ERR(client[i].scratch[0]);
+			goto err;
+		}
+
+		client[i].scratch[1] =
+			__vm_create_scratch_for_read_pinned(gt->vm, 4096);
+		if (IS_ERR(client[i].scratch[1])) {
+			err = PTR_ERR(client[i].scratch[1]);
+			i915_vma_unpin_and_release(&client[i].scratch[0], 0);
+			goto err;
+		}
+	}
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce[2];
+
+		if (!engine->kernel_context->vm)
+			continue;
+
+		if (!whitelist_writable_count(engine))
+			continue;
+
+		ce[0] = intel_context_create(engine);
+		if (IS_ERR(ce[0])) {
+			err = PTR_ERR(ce[0]);
+			break;
+		}
+		ce[1] = intel_context_create(engine);
+		if (IS_ERR(ce[1])) {
+			err = PTR_ERR(ce[1]);
+			intel_context_put(ce[0]);
+			break;
+		}
+
+		/* Read default values */
+		err = read_whitelisted_registers(ce[0], client[0].scratch[0]);
+		if (err)
+			goto err_ce;
+
+		/* Try to overwrite registers (should only affect ctx0) */
+		err = scrub_whitelisted_registers(ce[0]);
+		if (err)
+			goto err_ce;
+
+		/* Read values from ctx1, we expect these to be defaults */
+		err = read_whitelisted_registers(ce[1], client[1].scratch[0]);
+		if (err)
+			goto err_ce;
+
+		/* Verify that both reads return the same default values */
+		err = check_whitelisted_registers(engine,
+						  client[0].scratch[0],
+						  client[1].scratch[0],
+						  result_eq);
+		if (err)
+			goto err_ce;
+
+		/* Read back the updated values in ctx0 */
+		err = read_whitelisted_registers(ce[0], client[0].scratch[1]);
+		if (err)
+			goto err_ce;
+
+		/* User should be granted privilege to overwhite regs */
+		err = check_whitelisted_registers(engine,
+						  client[0].scratch[0],
+						  client[0].scratch[1],
+						  result_neq);
+err_ce:
+		intel_context_put(ce[1]);
+		intel_context_put(ce[0]);
+		if (err)
+			break;
+	}
+
+err:
+	for (i = 0; i < ARRAY_SIZE(client); i++) {
+		i915_vma_unpin_and_release(&client[i].scratch[1], 0);
+		i915_vma_unpin_and_release(&client[i].scratch[0], 0);
+	}
+
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	return err;
+}
+
+static bool
+verify_wa_lists(struct intel_gt *gt, struct wa_lists *lists,
+		const char *str)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	bool ok = true;
+
+	ok &= wa_list_verify(gt, &lists->gt_wa_list, str);
+
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce))
+			return false;
+
+		ok &= engine_wa_list_verify(ce,
+					    &lists->engine[id].wa_list,
+					    str) == 0;
+
+		ok &= engine_wa_list_verify(ce,
+					    &lists->engine[id].ctx_wa_list,
+					    str) == 0;
+
+		intel_context_put(ce);
+	}
+
+	return ok;
+}
+
+static int
+live_gpu_reset_workarounds(void *arg)
+{
+	struct intel_gt *gt = arg;
+	intel_wakeref_t wakeref;
+	struct wa_lists *lists;
+	bool ok;
+
+	if (!intel_has_gpu_reset(gt))
+		return 0;
+
+	lists = kzalloc(sizeof(*lists), GFP_KERNEL);
+	if (!lists)
+		return -ENOMEM;
+
+	pr_info("Verifying after GPU reset...\n");
+
+	igt_global_reset_lock(gt);
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+
+	reference_lists_init(gt, lists);
+
+	ok = verify_wa_lists(gt, lists, "before reset");
+	if (!ok)
+		goto out;
+
+	intel_gt_reset(gt, ALL_ENGINES, "live_workarounds");
+
+	ok = verify_wa_lists(gt, lists, "after reset");
+
+out:
+	reference_lists_fini(gt, lists);
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+	igt_global_reset_unlock(gt);
+	kfree(lists);
+
+	return ok ? 0 : -ESRCH;
+}
+
+static int
+live_engine_reset_workarounds(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct intel_context *ce;
+	struct igt_spinner spin;
+	struct i915_request *rq;
+	intel_wakeref_t wakeref;
+	struct wa_lists *lists;
+	int ret = 0;
+
+	if (!intel_has_reset_engine(gt))
+		return 0;
+
+	lists = kzalloc(sizeof(*lists), GFP_KERNEL);
+	if (!lists)
+		return -ENOMEM;
+
+	igt_global_reset_lock(gt);
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+
+	reference_lists_init(gt, lists);
+
+	for_each_engine(engine, gt, id) {
+		struct intel_selftest_saved_policy saved;
+		bool using_guc = intel_engine_uses_guc(engine);
+		bool ok;
+		int ret2;
+
+		pr_info("Verifying after %s reset...\n", engine->name);
+		ret = intel_selftest_modify_policy(engine, &saved,
+						   SELFTEST_SCHEDULER_MODIFY_FAST_RESET);
+		if (ret)
+			break;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			ret = PTR_ERR(ce);
+			goto restore;
+		}
+
+		if (!using_guc) {
+			ok = verify_wa_lists(gt, lists, "before reset");
+			if (!ok) {
+				ret = -ESRCH;
+				goto err;
+			}
+
+			ret = intel_engine_reset(engine, "live_workarounds:idle");
+			if (ret) {
+				pr_err("%s: Reset failed while idle\n", engine->name);
+				goto err;
+			}
+
+			ok = verify_wa_lists(gt, lists, "after idle reset");
+			if (!ok) {
+				ret = -ESRCH;
+				goto err;
+			}
+		}
+
+		ret = igt_spinner_init(&spin, engine->gt);
+		if (ret)
+			goto err;
+
+		rq = igt_spinner_create_request(&spin, ce, MI_NOOP);
+		if (IS_ERR(rq)) {
+			ret = PTR_ERR(rq);
+			igt_spinner_fini(&spin);
+			goto err;
+		}
+
+		ret = request_add_spin(rq, &spin);
+		if (ret) {
+			pr_err("%s: Spinner failed to start\n", engine->name);
+			igt_spinner_fini(&spin);
+			goto err;
+		}
+
+		/* Ensure the spinner hasn't aborted */
+		if (i915_request_completed(rq)) {
+			ret = -ETIMEDOUT;
+			goto skip;
+		}
+
+		if (!using_guc) {
+			ret = intel_engine_reset(engine, "live_workarounds:active");
+			if (ret) {
+				pr_err("%s: Reset failed on an active spinner\n",
+				       engine->name);
+				igt_spinner_fini(&spin);
+				goto err;
+			}
+		}
+
+		/* Ensure the reset happens and kills the engine */
+		if (ret == 0)
+			ret = intel_selftest_wait_for_rq(rq);
+
+skip:
+		igt_spinner_end(&spin);
+		igt_spinner_fini(&spin);
+
+		ok = verify_wa_lists(gt, lists, "after busy reset");
+		if (!ok)
+			ret = -ESRCH;
+
+err:
+		intel_context_put(ce);
+
+restore:
+		ret2 = intel_selftest_restore_policy(engine, &saved);
+		if (ret == 0)
+			ret = ret2;
+		if (ret)
+			break;
+	}
+
+	reference_lists_fini(gt, lists);
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+	igt_global_reset_unlock(gt);
+	kfree(lists);
+
+	igt_flush_test(gt->i915);
+
+	return ret;
+}
+
+int intel_workarounds_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_dirty_whitelist),
+		SUBTEST(live_reset_whitelist),
+		SUBTEST(live_isolated_whitelist),
+		SUBTEST(live_gpu_reset_workarounds),
+		SUBTEST(live_engine_reset_workarounds),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
diff --git a/drivers/gpu/drm/i915/gt/selftests/mock_timeline.c b/drivers/gpu/drm/i915/gt/selftests/mock_timeline.c
new file mode 100644
index 000000000..aeb1d1f61
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftests/mock_timeline.c
@@ -0,0 +1,29 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2017-2018 Intel Corporation
+ */
+
+#include "../intel_timeline.h"
+
+#include "mock_timeline.h"
+
+void mock_timeline_init(struct intel_timeline *timeline, u64 context)
+{
+	timeline->gt = NULL;
+	timeline->fence_context = context;
+
+	mutex_init(&timeline->mutex);
+
+	INIT_ACTIVE_FENCE(&timeline->last_request);
+	INIT_LIST_HEAD(&timeline->requests);
+
+	i915_syncmap_init(&timeline->sync);
+
+	INIT_LIST_HEAD(&timeline->link);
+}
+
+void mock_timeline_fini(struct intel_timeline *timeline)
+{
+	i915_syncmap_free(&timeline->sync);
+}
diff --git a/drivers/gpu/drm/i915/gt/selftests/mock_timeline.h b/drivers/gpu/drm/i915/gt/selftests/mock_timeline.h
new file mode 100644
index 000000000..d2bcc3df6
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftests/mock_timeline.h
@@ -0,0 +1,17 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2017-2018 Intel Corporation
+ */
+
+#ifndef __MOCK_TIMELINE__
+#define __MOCK_TIMELINE__
+
+#include <linux/types.h>
+
+struct intel_timeline;
+
+void mock_timeline_init(struct intel_timeline *timeline, u64 context);
+void mock_timeline_fini(struct intel_timeline *timeline);
+
+#endif /* !__MOCK_TIMELINE__ */
diff --git a/drivers/gpu/drm/i915/gt/shaders/README b/drivers/gpu/drm/i915/gt/shaders/README
new file mode 100644
index 000000000..e7e96d707
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/shaders/README
@@ -0,0 +1,46 @@
+ASM sources for auto generated shaders
+======================================
+
+The i915/gt/hsw_clear_kernel.c and i915/gt/ivb_clear_kernel.c files contain
+pre-compiled batch chunks that will clear any residual render cache during
+context switch.
+
+They are generated from their respective platform ASM files present on
+i915/gt/shaders/clear_kernel directory.
+
+The generated .c files should never be modified directly. Instead, any modification
+needs to be done on the on their respective ASM files and build instructions below
+needes to be followed.
+
+Building
+========
+
+Environment
+-----------
+
+IGT GPU tool scripts and the Mesa's i965 instruction assembler tool are used
+on building.
+
+Please make sure your Mesa tool is compiled with "-Dtools=intel" and
+"-Ddri-drivers=i965", and run this script from IGT source root directory"
+
+The instructions bellow assume:
+    *  IGT gpu tools source code is located on your home directory (~) as ~/igt
+    *  Mesa source code is located on your home directory (~) as ~/mesa
+       and built under the ~/mesa/build directory
+    *  Linux kernel source code is under your home directory (~) as ~/linux
+
+Instructions
+------------
+
+~ $ cp ~/linux/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm \
+       ~/igt/lib/i915/shaders/clear_kernel/ivb.asm
+~ $ cd ~/igt
+igt $ ./scripts/generate_clear_kernel.sh -g ivb \
+      -m ~/mesa/build/src/intel/tools/i965_asm
+
+~ $ cp ~/linux/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm \
+    ~/igt/lib/i915/shaders/clear_kernel/hsw.asm
+~ $ cd ~/igt
+igt $ ./scripts/generate_clear_kernel.sh -g hsw \
+      -m ~/mesa/build/src/intel/tools/i965_asm
\ No newline at end of file
diff --git a/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm
new file mode 100644
index 000000000..5fdf384bb
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm
@@ -0,0 +1,119 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+/*
+ * Kernel for PAVP buffer clear.
+ *
+ *	1. Clear all 64 GRF registers assigned to the kernel with designated value;
+ *	2. Write 32x16 block of all "0" to render target buffer which indirectly clears
+ *	   512 bytes of Render Cache.
+ */
+
+/* Store designated "clear GRF" value */
+mov(1)          f0.1<1>UW       g1.2<0,1,0>UW                   { align1 1N };
+
+/**
+ * Curbe Format
+ *
+ * DW 1.0 - Block Offset to write Render Cache
+ * DW 1.1 [15:0] - Clear Word
+ * DW 1.2 - Delay iterations
+ * DW 1.3 - Enable Instrumentation (only for debug)
+ * DW 1.4 - Rsvd (intended for context ID)
+ * DW 1.5 - [31:16]:SliceCount, [15:0]:SubSlicePerSliceCount
+ * DW 1.6 - Rsvd MBZ (intended for Enable Wait on Total Thread Count)
+ * DW 1.7 - Rsvd MBZ (inteded for Total Thread Count)
+ *
+ * Binding Table
+ *
+ * BTI 0: 2D Surface to help clear L3 (Render/Data Cache)
+ * BTI 1: Wait/Instrumentation Buffer
+ *  Size : (SliceCount * SubSliceCount  * 16 EUs/SubSlice) rows * (16 threads/EU) cols (Format R32_UINT)
+ *         Expected to be initialized to 0 by driver/another kernel
+ *  Layout:
+ *          RowN: Histogram for EU-N: (SliceID*SubSlicePerSliceCount + SSID)*16 + EUID [assume max 16 EUs / SS]
+ *          Col-k[DW-k]: Threads Executed on ThreadID-k for EU-N
+ */
+add(1)          g1.2<1>UD       g1.2<0,1,0>UD   0x00000001UD    { align1 1N }; /* Loop count to delay kernel: Init to (g1.2 + 1) */
+cmp.z.f0.0(1)   null<1>UD       g1.3<0,1,0>UD   0x00000000UD    { align1 1N };
+(+f0.0) jmpi(1) 352D                                            { align1 WE_all 1N };
+
+/**
+ * State Register has info on where this thread is running
+ *	IVB: sr0.0 :: [15:13]: MBZ, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID
+ *	HSW: sr0.0 :: 15: MBZ, [14:13]: SliceID, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID
+ */
+mov(8)          g3<1>UD         0x00000000UD                    { align1 1Q };
+shr(1)          g3<1>D          sr0<0,1,0>D     12D             { align1 1N };
+and(1)          g3<1>D          g3<0,1,0>D      1D              { align1 1N }; /* g3 has HSID */
+shr(1)          g3.1<1>D        sr0<0,1,0>D     13D             { align1 1N };
+and(1)          g3.1<1>D        g3.1<0,1,0>D    3D              { align1 1N }; /* g3.1 has sliceID */
+mul(1)          g3.5<1>D        g3.1<0,1,0>D    g1.10<0,1,0>UW  { align1 1N };
+add(1)          g3<1>D          g3<0,1,0>D      g3.5<0,1,0>D    { align1 1N }; /* g3 = sliceID * SubSlicePerSliceCount + HSID */
+shr(1)          g3.2<1>D        sr0<0,1,0>D     8D              { align1 1N };
+and(1)          g3.2<1>D        g3.2<0,1,0>D    15D             { align1 1N }; /* g3.2 = EUID */
+mul(1)          g3.4<1>D        g3<0,1,0>D      16D             { align1 1N };
+add(1)          g3.2<1>D        g3.2<0,1,0>D    g3.4<0,1,0>D    { align1 1N }; /* g3.2 now points to EU row number (Y-pixel = V address )  in instrumentation surf */
+
+mov(8)          g5<1>UD         0x00000000UD                    { align1 1Q };
+and(1)          g3.3<1>D        sr0<0,1,0>D     7D              { align1 1N };
+mul(1)          g3.3<1>D        g3.3<0,1,0>D    4D              { align1 1N };
+
+mov(8)          g4<1>UD         g0<8,8,1>UD                     { align1 1Q }; /* Initialize message header with g0 */
+mov(1)          g4<1>UD         g3.3<0,1,0>UD                   { align1 1N }; /* Block offset */
+mov(1)          g4.1<1>UD       g3.2<0,1,0>UD                   { align1 1N }; /* Block offset */
+mov(1)          g4.2<1>UD       0x00000003UD                    { align1 1N }; /* Block size (1 row x 4 bytes) */
+and(1)          g4.3<1>UD       g4.3<0,1,0>UW   0xffffffffUD    { align1 1N };
+
+/* Media block read to fetch current value at specified location in instrumentation buffer */
+sendc(8)        g5<1>UD         g4<8,8,1>F      0x02190001
+
+                            render MsgDesc: media block read MsgCtrl = 0x0 Surface = 1 mlen 1 rlen 1 { align1 1Q };
+add(1)          g5<1>D          g5<0,1,0>D      1D              { align1 1N };
+
+/* Media block write for updated value at specified location in instrumentation buffer */
+sendc(8)        g5<1>UD         g4<8,8,1>F      0x040a8001
+                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 1 mlen 2 rlen 0 { align1 1Q };
+
+/* Delay thread for specified parameter */
+add.nz.f0.0(1)  g1.2<1>UD       g1.2<0,1,0>UD   -1D             { align1 1N };
+(+f0.0) jmpi(1) -32D                                            { align1 WE_all 1N };
+
+/* Store designated "clear GRF" value */
+mov(1)          f0.1<1>UW       g1.2<0,1,0>UW                   { align1 1N };
+
+/* Initialize looping parameters */
+mov(1)          a0<1>D          0D                              { align1 1N }; /* Initialize a0.0:w=0 */
+mov(1)          a0.4<1>W        127W                            { align1 1N }; /* Loop count. Each loop contains 16 GRF's */
+
+/* Write 32x16 all "0" block */
+mov(8)          g2<1>UD         g0<8,8,1>UD                     { align1 1Q };
+mov(8)          g127<1>UD       g0<8,8,1>UD                     { align1 1Q };
+mov(2)          g2<1>UD         g1<2,2,1>UW                     { align1 1N };
+mov(1)          g2.2<1>UD       0x000f000fUD                    { align1 1N }; /* Block size (16x16) */
+and(1)          g2.3<1>UD       g2.3<0,1,0>UW   0xffffffefUD    { align1 1N };
+mov(16)         g3<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g4<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g5<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g6<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g7<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g8<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g9<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g10<1>UD        0x00000000UD                    { align1 1H };
+sendc(8)        null<1>UD       g2<8,8,1>F      0x120a8000
+                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q };
+add(1)          g2<1>UD         g1<0,1,0>UW     0x0010UW        { align1 1N };
+sendc(8)        null<1>UD       g2<8,8,1>F      0x120a8000
+                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q };
+
+/* Now, clear all GRF registers */
+add.nz.f0.0(1)  a0.4<1>W        a0.4<0,1,0>W    -1W             { align1 1N };
+mov(16)         g[a0]<1>UW      f0.1<0,1,0>UW                   { align1 1H };
+add(1)          a0<1>D          a0<0,1,0>D      32D             { align1 1N };
+(+f0.0) jmpi(1) -64D                                            { align1 WE_all 1N };
+
+/* Terminante the thread */
+sendc(8)        null<1>UD       g127<8,8,1>F    0x82000010
+                            thread_spawner MsgDesc: mlen 1 rlen 0           { align1 1Q EOT };
diff --git a/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm
new file mode 100644
index 000000000..97c7ac9e3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm
@@ -0,0 +1,117 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+/*
+ * Kernel for PAVP buffer clear.
+ *
+ *	1. Clear all 64 GRF registers assigned to the kernel with designated value;
+ *	2. Write 32x16 block of all "0" to render target buffer which indirectly clears
+ *	   512 bytes of Render Cache.
+ */
+
+/* Store designated "clear GRF" value */
+mov(1)          f0.1<1>UW       g1.2<0,1,0>UW                   { align1 1N };
+
+/**
+ * Curbe Format
+ *
+ * DW 1.0 - Block Offset to write Render Cache
+ * DW 1.1 [15:0] - Clear Word
+ * DW 1.2 - Delay iterations
+ * DW 1.3 - Enable Instrumentation (only for debug)
+ * DW 1.4 - Rsvd (intended for context ID)
+ * DW 1.5 - [31:16]:SliceCount, [15:0]:SubSlicePerSliceCount
+ * DW 1.6 - Rsvd MBZ (intended for Enable Wait on Total Thread Count)
+ * DW 1.7 - Rsvd MBZ (inteded for Total Thread Count)
+ *
+ * Binding Table
+ *
+ * BTI 0: 2D Surface to help clear L3 (Render/Data Cache)
+ * BTI 1: Wait/Instrumentation Buffer
+ *  Size : (SliceCount * SubSliceCount  * 16 EUs/SubSlice) rows * (16 threads/EU) cols (Format R32_UINT)
+ *         Expected to be initialized to 0 by driver/another kernel
+ *  Layout :
+ *           RowN: Histogram for EU-N: (SliceID*SubSlicePerSliceCount + SSID)*16 + EUID [assume max 16 EUs / SS]
+ *           Col-k[DW-k]: Threads Executed on ThreadID-k for EU-N
+ */
+add(1)          g1.2<1>UD       g1.2<0,1,0>UD   0x00000001UD    { align1 1N }; /* Loop count to delay kernel: Init to (g1.2 + 1) */
+cmp.z.f0.0(1)   null<1>UD       g1.3<0,1,0>UD   0x00000000UD    { align1 1N };
+(+f0.0) jmpi(1) 44D                                             { align1 WE_all 1N };
+
+/**
+ * State Register has info on where this thread is running
+ *	IVB: sr0.0 :: [15:13]: MBZ, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID
+ *	HSW: sr0.0 :: 15: MBZ, [14:13]: SliceID, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID
+ */
+mov(8)          g3<1>UD         0x00000000UD                    { align1 1Q };
+shr(1)          g3<1>D          sr0<0,1,0>D     12D             { align1 1N };
+and(1)          g3<1>D          g3<0,1,0>D      1D              { align1 1N }; /* g3 has HSID */
+shr(1)          g3.1<1>D        sr0<0,1,0>D     13D             { align1 1N };
+and(1)          g3.1<1>D        g3.1<0,1,0>D    3D              { align1 1N }; /* g3.1 has sliceID */
+mul(1)          g3.5<1>D        g3.1<0,1,0>D    g1.10<0,1,0>UW  { align1 1N };
+add(1)          g3<1>D          g3<0,1,0>D      g3.5<0,1,0>D    { align1 1N }; /* g3 = sliceID * SubSlicePerSliceCount + HSID */
+shr(1)          g3.2<1>D        sr0<0,1,0>D     8D              { align1 1N };
+and(1)          g3.2<1>D        g3.2<0,1,0>D    15D             { align1 1N }; /* g3.2 = EUID */
+mul(1)          g3.4<1>D        g3<0,1,0>D      16D             { align1 1N };
+add(1)          g3.2<1>D        g3.2<0,1,0>D    g3.4<0,1,0>D    { align1 1N }; /* g3.2 now points to EU row number (Y-pixel = V address )  in instrumentation surf */
+
+mov(8)          g5<1>UD         0x00000000UD                    { align1 1Q };
+and(1)          g3.3<1>D        sr0<0,1,0>D     7D              { align1 1N };
+mul(1)          g3.3<1>D        g3.3<0,1,0>D    4D              { align1 1N };
+
+mov(8)          g4<1>UD         g0<8,8,1>UD                     { align1 1Q }; /* Initialize message header with g0 */
+mov(1)          g4<1>UD         g3.3<0,1,0>UD                   { align1 1N }; /* Block offset */
+mov(1)          g4.1<1>UD       g3.2<0,1,0>UD                   { align1 1N }; /* Block offset */
+mov(1)          g4.2<1>UD       0x00000003UD                    { align1 1N }; /* Block size (1 row x 4 bytes) */
+and(1)          g4.3<1>UD       g4.3<0,1,0>UW   0xffffffffUD    { align1 1N };
+
+/* Media block read to fetch current value at specified location in instrumentation buffer */
+sendc(8)        g5<1>UD         g4<8,8,1>F      0x02190001
+                            render MsgDesc: media block read MsgCtrl = 0x0 Surface = 1 mlen 1 rlen 1 { align1 1Q };
+add(1)          g5<1>D          g5<0,1,0>D      1D              { align1 1N };
+
+/* Media block write for updated value at specified location in instrumentation buffer */
+sendc(8)        g5<1>UD         g4<8,8,1>F      0x040a8001
+                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 1 mlen 2 rlen 0 { align1 1Q };
+/* Delay thread for specified parameter */
+add.nz.f0.0(1)  g1.2<1>UD       g1.2<0,1,0>UD   -1D             { align1 1N };
+(+f0.0) jmpi(1) -4D                                             { align1 WE_all 1N };
+
+/* Store designated "clear GRF" value */
+mov(1)          f0.1<1>UW       g1.2<0,1,0>UW                   { align1 1N };
+
+/* Initialize looping parameters */
+mov(1)          a0<1>D          0D                              { align1 1N }; /* Initialize a0.0:w=0 */
+mov(1)          a0.4<1>W        127W                            { align1 1N }; /* Loop count. Each loop contains 16 GRF's */
+
+/* Write 32x16 all "0" block */
+mov(8)          g2<1>UD         g0<8,8,1>UD                     { align1 1Q };
+mov(8)          g127<1>UD       g0<8,8,1>UD                     { align1 1Q };
+mov(2)          g2<1>UD         g1<2,2,1>UW                     { align1 1N };
+mov(1)          g2.2<1>UD       0x000f000fUD                    { align1 1N }; /* Block size (16x16) */
+and(1)          g2.3<1>UD       g2.3<0,1,0>UW   0xffffffefUD    { align1 1N };
+mov(16)         g3<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g4<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g5<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g6<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g7<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g8<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g9<1>UD         0x00000000UD                    { align1 1H };
+mov(16)         g10<1>UD        0x00000000UD                    { align1 1H };
+sendc(8)        null<1>UD       g2<8,8,1>F      0x120a8000
+                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q };
+add(1)          g2<1>UD         g1<0,1,0>UW     0x0010UW        { align1 1N };
+sendc(8)        null<1>UD       g2<8,8,1>F      0x120a8000
+                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q };
+
+/* Now, clear all GRF registers */
+add.nz.f0.0(1)  a0.4<1>W        a0.4<0,1,0>W    -1W             { align1 1N };
+mov(16)         g[a0]<1>UW      f0.1<0,1,0>UW                   { align1 1H };
+add(1)          a0<1>D          a0<0,1,0>D      32D             { align1 1N };
+(+f0.0) jmpi(1) -8D                                             { align1 WE_all 1N };
+
+/* Terminante the thread */
+sendc(8)        null<1>UD       g127<8,8,1>F    0x82000010
+                            thread_spawner MsgDesc: mlen 1 rlen 0           { align1 1Q EOT };
diff --git a/drivers/gpu/drm/i915/gt/shmem_utils.c b/drivers/gpu/drm/i915/gt/shmem_utils.c
new file mode 100644
index 000000000..402f085f3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/shmem_utils.c
@@ -0,0 +1,170 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/iosys-map.h>
+#include <linux/mm.h>
+#include <linux/pagemap.h>
+#include <linux/shmem_fs.h>
+
+#include "gem/i915_gem_object.h"
+#include "gem/i915_gem_lmem.h"
+#include "shmem_utils.h"
+
+struct file *shmem_create_from_data(const char *name, void *data, size_t len)
+{
+	struct file *file;
+	int err;
+
+	file = shmem_file_setup(name, PAGE_ALIGN(len), VM_NORESERVE);
+	if (IS_ERR(file))
+		return file;
+
+	err = shmem_write(file, 0, data, len);
+	if (err) {
+		fput(file);
+		return ERR_PTR(err);
+	}
+
+	return file;
+}
+
+struct file *shmem_create_from_object(struct drm_i915_gem_object *obj)
+{
+	struct file *file;
+	void *ptr;
+
+	if (i915_gem_object_is_shmem(obj)) {
+		file = obj->base.filp;
+		atomic_long_inc(&file->f_count);
+		return file;
+	}
+
+	ptr = i915_gem_object_pin_map_unlocked(obj, i915_gem_object_is_lmem(obj) ?
+						I915_MAP_WC : I915_MAP_WB);
+	if (IS_ERR(ptr))
+		return ERR_CAST(ptr);
+
+	file = shmem_create_from_data("", ptr, obj->base.size);
+	i915_gem_object_unpin_map(obj);
+
+	return file;
+}
+
+void *shmem_pin_map(struct file *file)
+{
+	struct page **pages;
+	size_t n_pages, i;
+	void *vaddr;
+
+	n_pages = file->f_mapping->host->i_size >> PAGE_SHIFT;
+	pages = kvmalloc_array(n_pages, sizeof(*pages), GFP_KERNEL);
+	if (!pages)
+		return NULL;
+
+	for (i = 0; i < n_pages; i++) {
+		pages[i] = shmem_read_mapping_page_gfp(file->f_mapping, i,
+						       GFP_KERNEL);
+		if (IS_ERR(pages[i]))
+			goto err_page;
+	}
+
+	vaddr = vmap(pages, n_pages, VM_MAP_PUT_PAGES, PAGE_KERNEL);
+	if (!vaddr)
+		goto err_page;
+	mapping_set_unevictable(file->f_mapping);
+	return vaddr;
+err_page:
+	while (i--)
+		put_page(pages[i]);
+	kvfree(pages);
+	return NULL;
+}
+
+void shmem_unpin_map(struct file *file, void *ptr)
+{
+	mapping_clear_unevictable(file->f_mapping);
+	vfree(ptr);
+}
+
+static int __shmem_rw(struct file *file, loff_t off,
+		      void *ptr, size_t len,
+		      bool write)
+{
+	unsigned long pfn;
+
+	for (pfn = off >> PAGE_SHIFT; len; pfn++) {
+		unsigned int this =
+			min_t(size_t, PAGE_SIZE - offset_in_page(off), len);
+		struct page *page;
+		void *vaddr;
+
+		page = shmem_read_mapping_page_gfp(file->f_mapping, pfn,
+						   GFP_KERNEL);
+		if (IS_ERR(page))
+			return PTR_ERR(page);
+
+		vaddr = kmap(page);
+		if (write) {
+			memcpy(vaddr + offset_in_page(off), ptr, this);
+			set_page_dirty(page);
+		} else {
+			memcpy(ptr, vaddr + offset_in_page(off), this);
+		}
+		mark_page_accessed(page);
+		kunmap(page);
+		put_page(page);
+
+		len -= this;
+		ptr += this;
+		off = 0;
+	}
+
+	return 0;
+}
+
+int shmem_read_to_iosys_map(struct file *file, loff_t off,
+			    struct iosys_map *map, size_t map_off, size_t len)
+{
+	unsigned long pfn;
+
+	for (pfn = off >> PAGE_SHIFT; len; pfn++) {
+		unsigned int this =
+			min_t(size_t, PAGE_SIZE - offset_in_page(off), len);
+		struct page *page;
+		void *vaddr;
+
+		page = shmem_read_mapping_page_gfp(file->f_mapping, pfn,
+						   GFP_KERNEL);
+		if (IS_ERR(page))
+			return PTR_ERR(page);
+
+		vaddr = kmap(page);
+		iosys_map_memcpy_to(map, map_off, vaddr + offset_in_page(off),
+				    this);
+		mark_page_accessed(page);
+		kunmap(page);
+		put_page(page);
+
+		len -= this;
+		map_off += this;
+		off = 0;
+	}
+
+	return 0;
+}
+
+int shmem_read(struct file *file, loff_t off, void *dst, size_t len)
+{
+	return __shmem_rw(file, off, dst, len, false);
+}
+
+int shmem_write(struct file *file, loff_t off, void *src, size_t len)
+{
+	return __shmem_rw(file, off, src, len, true);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "st_shmem_utils.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/shmem_utils.h b/drivers/gpu/drm/i915/gt/shmem_utils.h
new file mode 100644
index 000000000..b2b04d88c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/shmem_utils.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef SHMEM_UTILS_H
+#define SHMEM_UTILS_H
+
+#include <linux/types.h>
+
+struct iosys_map;
+struct drm_i915_gem_object;
+struct file;
+
+struct file *shmem_create_from_data(const char *name, void *data, size_t len);
+struct file *shmem_create_from_object(struct drm_i915_gem_object *obj);
+
+void *shmem_pin_map(struct file *file);
+void shmem_unpin_map(struct file *file, void *ptr);
+
+int shmem_read_to_iosys_map(struct file *file, loff_t off,
+			    struct iosys_map *map, size_t map_off, size_t len);
+int shmem_read(struct file *file, loff_t off, void *dst, size_t len);
+int shmem_write(struct file *file, loff_t off, void *src, size_t len);
+
+#endif /* SHMEM_UTILS_H */
diff --git a/drivers/gpu/drm/i915/gt/st_shmem_utils.c b/drivers/gpu/drm/i915/gt/st_shmem_utils.c
new file mode 100644
index 000000000..b279fe88b
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/st_shmem_utils.c
@@ -0,0 +1,63 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+/* Just a quick and causal check of the shmem_utils API */
+
+static int igt_shmem_basic(void *ignored)
+{
+	u32 datum = 0xdeadbeef, result;
+	struct file *file;
+	u32 *map;
+	int err;
+
+	file = shmem_create_from_data("mock", &datum, sizeof(datum));
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	result = 0;
+	err = shmem_read(file, 0, &result, sizeof(result));
+	if (err)
+		goto out_file;
+
+	if (result != datum) {
+		pr_err("Incorrect read back from shmemfs: %x != %x\n",
+		       result, datum);
+		err = -EINVAL;
+		goto out_file;
+	}
+
+	result = 0xc0ffee;
+	err = shmem_write(file, 0, &result, sizeof(result));
+	if (err)
+		goto out_file;
+
+	map = shmem_pin_map(file);
+	if (!map) {
+		err = -ENOMEM;
+		goto out_file;
+	}
+
+	if (*map != result) {
+		pr_err("Incorrect read back via mmap of last write: %x != %x\n",
+		       *map, result);
+		err = -EINVAL;
+		goto out_map;
+	}
+
+out_map:
+	shmem_unpin_map(file, map);
+out_file:
+	fput(file);
+	return err;
+}
+
+int shmem_utils_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_shmem_basic),
+	};
+
+	return i915_subtests(tests, NULL);
+}
diff --git a/drivers/gpu/drm/i915/gt/sysfs_engines.c b/drivers/gpu/drm/i915/gt/sysfs_engines.c
new file mode 100644
index 000000000..f2d9858d8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/sysfs_engines.c
@@ -0,0 +1,540 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/kobject.h>
+#include <linux/sysfs.h>
+
+#include "i915_drv.h"
+#include "intel_engine.h"
+#include "intel_engine_heartbeat.h"
+#include "sysfs_engines.h"
+
+struct kobj_engine {
+	struct kobject base;
+	struct intel_engine_cs *engine;
+};
+
+static struct intel_engine_cs *kobj_to_engine(struct kobject *kobj)
+{
+	return container_of(kobj, struct kobj_engine, base)->engine;
+}
+
+static ssize_t
+name_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	return sprintf(buf, "%s\n", kobj_to_engine(kobj)->name);
+}
+
+static struct kobj_attribute name_attr =
+__ATTR(name, 0444, name_show, NULL);
+
+static ssize_t
+class_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	return sprintf(buf, "%d\n", kobj_to_engine(kobj)->uabi_class);
+}
+
+static struct kobj_attribute class_attr =
+__ATTR(class, 0444, class_show, NULL);
+
+static ssize_t
+inst_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	return sprintf(buf, "%d\n", kobj_to_engine(kobj)->uabi_instance);
+}
+
+static struct kobj_attribute inst_attr =
+__ATTR(instance, 0444, inst_show, NULL);
+
+static ssize_t
+mmio_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	return sprintf(buf, "0x%x\n", kobj_to_engine(kobj)->mmio_base);
+}
+
+static struct kobj_attribute mmio_attr =
+__ATTR(mmio_base, 0444, mmio_show, NULL);
+
+static const char * const vcs_caps[] = {
+	[ilog2(I915_VIDEO_CLASS_CAPABILITY_HEVC)] = "hevc",
+	[ilog2(I915_VIDEO_AND_ENHANCE_CLASS_CAPABILITY_SFC)] = "sfc",
+};
+
+static const char * const vecs_caps[] = {
+	[ilog2(I915_VIDEO_AND_ENHANCE_CLASS_CAPABILITY_SFC)] = "sfc",
+};
+
+static ssize_t repr_trim(char *buf, ssize_t len)
+{
+	/* Trim off the trailing space and replace with a newline */
+	if (len > PAGE_SIZE)
+		len = PAGE_SIZE;
+	if (len > 0)
+		buf[len - 1] = '\n';
+
+	return len;
+}
+
+static ssize_t
+__caps_show(struct intel_engine_cs *engine,
+	    unsigned long caps, char *buf, bool show_unknown)
+{
+	const char * const *repr;
+	int count, n;
+	ssize_t len;
+
+	switch (engine->class) {
+	case VIDEO_DECODE_CLASS:
+		repr = vcs_caps;
+		count = ARRAY_SIZE(vcs_caps);
+		break;
+
+	case VIDEO_ENHANCEMENT_CLASS:
+		repr = vecs_caps;
+		count = ARRAY_SIZE(vecs_caps);
+		break;
+
+	default:
+		repr = NULL;
+		count = 0;
+		break;
+	}
+	GEM_BUG_ON(count > BITS_PER_LONG);
+
+	len = 0;
+	for_each_set_bit(n, &caps, show_unknown ? BITS_PER_LONG : count) {
+		if (n >= count || !repr[n]) {
+			if (GEM_WARN_ON(show_unknown))
+				len += snprintf(buf + len, PAGE_SIZE - len,
+						"[%x] ", n);
+		} else {
+			len += snprintf(buf + len, PAGE_SIZE - len,
+					"%s ", repr[n]);
+		}
+		if (GEM_WARN_ON(len >= PAGE_SIZE))
+			break;
+	}
+	return repr_trim(buf, len);
+}
+
+static ssize_t
+caps_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return __caps_show(engine, engine->uabi_capabilities, buf, true);
+}
+
+static struct kobj_attribute caps_attr =
+__ATTR(capabilities, 0444, caps_show, NULL);
+
+static ssize_t
+all_caps_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	return __caps_show(kobj_to_engine(kobj), -1, buf, false);
+}
+
+static struct kobj_attribute all_caps_attr =
+__ATTR(known_capabilities, 0444, all_caps_show, NULL);
+
+static ssize_t
+max_spin_store(struct kobject *kobj, struct kobj_attribute *attr,
+	       const char *buf, size_t count)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+	unsigned long long duration, clamped;
+	int err;
+
+	/*
+	 * When waiting for a request, if is it currently being executed
+	 * on the GPU, we busywait for a short while before sleeping. The
+	 * premise is that most requests are short, and if it is already
+	 * executing then there is a good chance that it will complete
+	 * before we can setup the interrupt handler and go to sleep.
+	 * We try to offset the cost of going to sleep, by first spinning
+	 * on the request -- if it completed in less time than it would take
+	 * to go sleep, process the interrupt and return back to the client,
+	 * then we have saved the client some latency, albeit at the cost
+	 * of spinning on an expensive CPU core.
+	 *
+	 * While we try to avoid waiting at all for a request that is unlikely
+	 * to complete, deciding how long it is worth spinning is for is an
+	 * arbitrary decision: trading off power vs latency.
+	 */
+
+	err = kstrtoull(buf, 0, &duration);
+	if (err)
+		return err;
+
+	clamped = intel_clamp_max_busywait_duration_ns(engine, duration);
+	if (duration != clamped)
+		return -EINVAL;
+
+	WRITE_ONCE(engine->props.max_busywait_duration_ns, duration);
+
+	return count;
+}
+
+static ssize_t
+max_spin_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return sprintf(buf, "%lu\n", engine->props.max_busywait_duration_ns);
+}
+
+static struct kobj_attribute max_spin_attr =
+__ATTR(max_busywait_duration_ns, 0644, max_spin_show, max_spin_store);
+
+static ssize_t
+max_spin_default(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return sprintf(buf, "%lu\n", engine->defaults.max_busywait_duration_ns);
+}
+
+static struct kobj_attribute max_spin_def =
+__ATTR(max_busywait_duration_ns, 0444, max_spin_default, NULL);
+
+static ssize_t
+timeslice_store(struct kobject *kobj, struct kobj_attribute *attr,
+		const char *buf, size_t count)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+	unsigned long long duration, clamped;
+	int err;
+
+	/*
+	 * Execlists uses a scheduling quantum (a timeslice) to alternate
+	 * execution between ready-to-run contexts of equal priority. This
+	 * ensures that all users (though only if they of equal importance)
+	 * have the opportunity to run and prevents livelocks where contexts
+	 * may have implicit ordering due to userspace semaphores.
+	 */
+
+	err = kstrtoull(buf, 0, &duration);
+	if (err)
+		return err;
+
+	clamped = intel_clamp_timeslice_duration_ms(engine, duration);
+	if (duration != clamped)
+		return -EINVAL;
+
+	WRITE_ONCE(engine->props.timeslice_duration_ms, duration);
+
+	if (execlists_active(&engine->execlists))
+		set_timer_ms(&engine->execlists.timer, duration);
+
+	return count;
+}
+
+static ssize_t
+timeslice_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return sprintf(buf, "%lu\n", engine->props.timeslice_duration_ms);
+}
+
+static struct kobj_attribute timeslice_duration_attr =
+__ATTR(timeslice_duration_ms, 0644, timeslice_show, timeslice_store);
+
+static ssize_t
+timeslice_default(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return sprintf(buf, "%lu\n", engine->defaults.timeslice_duration_ms);
+}
+
+static struct kobj_attribute timeslice_duration_def =
+__ATTR(timeslice_duration_ms, 0444, timeslice_default, NULL);
+
+static ssize_t
+stop_store(struct kobject *kobj, struct kobj_attribute *attr,
+	   const char *buf, size_t count)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+	unsigned long long duration, clamped;
+	int err;
+
+	/*
+	 * When we allow ourselves to sleep before a GPU reset after disabling
+	 * submission, even for a few milliseconds, gives an innocent context
+	 * the opportunity to clear the GPU before the reset occurs. However,
+	 * how long to sleep depends on the typical non-preemptible duration
+	 * (a similar problem to determining the ideal preempt-reset timeout
+	 * or even the heartbeat interval).
+	 */
+
+	err = kstrtoull(buf, 0, &duration);
+	if (err)
+		return err;
+
+	clamped = intel_clamp_stop_timeout_ms(engine, duration);
+	if (duration != clamped)
+		return -EINVAL;
+
+	WRITE_ONCE(engine->props.stop_timeout_ms, duration);
+	return count;
+}
+
+static ssize_t
+stop_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return sprintf(buf, "%lu\n", engine->props.stop_timeout_ms);
+}
+
+static struct kobj_attribute stop_timeout_attr =
+__ATTR(stop_timeout_ms, 0644, stop_show, stop_store);
+
+static ssize_t
+stop_default(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return sprintf(buf, "%lu\n", engine->defaults.stop_timeout_ms);
+}
+
+static struct kobj_attribute stop_timeout_def =
+__ATTR(stop_timeout_ms, 0444, stop_default, NULL);
+
+static ssize_t
+preempt_timeout_store(struct kobject *kobj, struct kobj_attribute *attr,
+		      const char *buf, size_t count)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+	unsigned long long timeout, clamped;
+	int err;
+
+	/*
+	 * After initialising a preemption request, we give the current
+	 * resident a small amount of time to vacate the GPU. The preemption
+	 * request is for a higher priority context and should be immediate to
+	 * maintain high quality of service (and avoid priority inversion).
+	 * However, the preemption granularity of the GPU can be quite coarse
+	 * and so we need a compromise.
+	 */
+
+	err = kstrtoull(buf, 0, &timeout);
+	if (err)
+		return err;
+
+	clamped = intel_clamp_preempt_timeout_ms(engine, timeout);
+	if (timeout != clamped)
+		return -EINVAL;
+
+	WRITE_ONCE(engine->props.preempt_timeout_ms, timeout);
+
+	if (READ_ONCE(engine->execlists.pending[0]))
+		set_timer_ms(&engine->execlists.preempt, timeout);
+
+	return count;
+}
+
+static ssize_t
+preempt_timeout_show(struct kobject *kobj, struct kobj_attribute *attr,
+		     char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return sprintf(buf, "%lu\n", engine->props.preempt_timeout_ms);
+}
+
+static struct kobj_attribute preempt_timeout_attr =
+__ATTR(preempt_timeout_ms, 0644, preempt_timeout_show, preempt_timeout_store);
+
+static ssize_t
+preempt_timeout_default(struct kobject *kobj, struct kobj_attribute *attr,
+			char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return sprintf(buf, "%lu\n", engine->defaults.preempt_timeout_ms);
+}
+
+static struct kobj_attribute preempt_timeout_def =
+__ATTR(preempt_timeout_ms, 0444, preempt_timeout_default, NULL);
+
+static ssize_t
+heartbeat_store(struct kobject *kobj, struct kobj_attribute *attr,
+		const char *buf, size_t count)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+	unsigned long long delay, clamped;
+	int err;
+
+	/*
+	 * We monitor the health of the system via periodic heartbeat pulses.
+	 * The pulses also provide the opportunity to perform garbage
+	 * collection.  However, we interpret an incomplete pulse (a missed
+	 * heartbeat) as an indication that the system is no longer responsive,
+	 * i.e. hung, and perform an engine or full GPU reset. Given that the
+	 * preemption granularity can be very coarse on a system, the optimal
+	 * value for any workload is unknowable!
+	 */
+
+	err = kstrtoull(buf, 0, &delay);
+	if (err)
+		return err;
+
+	clamped = intel_clamp_heartbeat_interval_ms(engine, delay);
+	if (delay != clamped)
+		return -EINVAL;
+
+	err = intel_engine_set_heartbeat(engine, delay);
+	if (err)
+		return err;
+
+	return count;
+}
+
+static ssize_t
+heartbeat_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return sprintf(buf, "%lu\n", engine->props.heartbeat_interval_ms);
+}
+
+static struct kobj_attribute heartbeat_interval_attr =
+__ATTR(heartbeat_interval_ms, 0644, heartbeat_show, heartbeat_store);
+
+static ssize_t
+heartbeat_default(struct kobject *kobj, struct kobj_attribute *attr, char *buf)
+{
+	struct intel_engine_cs *engine = kobj_to_engine(kobj);
+
+	return sprintf(buf, "%lu\n", engine->defaults.heartbeat_interval_ms);
+}
+
+static struct kobj_attribute heartbeat_interval_def =
+__ATTR(heartbeat_interval_ms, 0444, heartbeat_default, NULL);
+
+static void kobj_engine_release(struct kobject *kobj)
+{
+	kfree(kobj);
+}
+
+static struct kobj_type kobj_engine_type = {
+	.release = kobj_engine_release,
+	.sysfs_ops = &kobj_sysfs_ops
+};
+
+static struct kobject *
+kobj_engine(struct kobject *dir, struct intel_engine_cs *engine)
+{
+	struct kobj_engine *ke;
+
+	ke = kzalloc(sizeof(*ke), GFP_KERNEL);
+	if (!ke)
+		return NULL;
+
+	kobject_init(&ke->base, &kobj_engine_type);
+	ke->engine = engine;
+
+	if (kobject_add(&ke->base, dir, "%s", engine->name)) {
+		kobject_put(&ke->base);
+		return NULL;
+	}
+
+	/* xfer ownership to sysfs tree */
+	return &ke->base;
+}
+
+static void add_defaults(struct kobj_engine *parent)
+{
+	static const struct attribute *files[] = {
+		&max_spin_def.attr,
+		&stop_timeout_def.attr,
+#if CONFIG_DRM_I915_HEARTBEAT_INTERVAL
+		&heartbeat_interval_def.attr,
+#endif
+		NULL
+	};
+	struct kobj_engine *ke;
+
+	ke = kzalloc(sizeof(*ke), GFP_KERNEL);
+	if (!ke)
+		return;
+
+	kobject_init(&ke->base, &kobj_engine_type);
+	ke->engine = parent->engine;
+
+	if (kobject_add(&ke->base, &parent->base, "%s", ".defaults")) {
+		kobject_put(&ke->base);
+		return;
+	}
+
+	if (sysfs_create_files(&ke->base, files))
+		return;
+
+	if (intel_engine_has_timeslices(ke->engine) &&
+	    sysfs_create_file(&ke->base, &timeslice_duration_def.attr))
+		return;
+
+	if (intel_engine_has_preempt_reset(ke->engine) &&
+	    sysfs_create_file(&ke->base, &preempt_timeout_def.attr))
+		return;
+}
+
+void intel_engines_add_sysfs(struct drm_i915_private *i915)
+{
+	static const struct attribute *files[] = {
+		&name_attr.attr,
+		&class_attr.attr,
+		&inst_attr.attr,
+		&mmio_attr.attr,
+		&caps_attr.attr,
+		&all_caps_attr.attr,
+		&max_spin_attr.attr,
+		&stop_timeout_attr.attr,
+#if CONFIG_DRM_I915_HEARTBEAT_INTERVAL
+		&heartbeat_interval_attr.attr,
+#endif
+		NULL
+	};
+
+	struct device *kdev = i915->drm.primary->kdev;
+	struct intel_engine_cs *engine;
+	struct kobject *dir;
+
+	dir = kobject_create_and_add("engine", &kdev->kobj);
+	if (!dir)
+		return;
+
+	for_each_uabi_engine(engine, i915) {
+		struct kobject *kobj;
+
+		kobj = kobj_engine(dir, engine);
+		if (!kobj)
+			goto err_engine;
+
+		if (sysfs_create_files(kobj, files))
+			goto err_object;
+
+		if (intel_engine_has_timeslices(engine) &&
+		    sysfs_create_file(kobj, &timeslice_duration_attr.attr))
+			goto err_engine;
+
+		if (intel_engine_has_preempt_reset(engine) &&
+		    sysfs_create_file(kobj, &preempt_timeout_attr.attr))
+			goto err_engine;
+
+		add_defaults(container_of(kobj, struct kobj_engine, base));
+
+		if (0) {
+err_object:
+			kobject_put(kobj);
+err_engine:
+			dev_err(kdev, "Failed to add sysfs engine '%s'\n",
+				engine->name);
+			break;
+		}
+	}
+}
diff --git a/drivers/gpu/drm/i915/gt/sysfs_engines.h b/drivers/gpu/drm/i915/gt/sysfs_engines.h
new file mode 100644
index 000000000..9546fffe0
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/sysfs_engines.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_ENGINE_SYSFS_H
+#define INTEL_ENGINE_SYSFS_H
+
+struct drm_i915_private;
+
+void intel_engines_add_sysfs(struct drm_i915_private *i915);
+
+#endif /* INTEL_ENGINE_SYSFS_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h b/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h
new file mode 100644
index 000000000..29ef8afc8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h
@@ -0,0 +1,183 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2021 Intel Corporation
+ */
+
+#ifndef _ABI_GUC_ACTIONS_ABI_H
+#define _ABI_GUC_ACTIONS_ABI_H
+
+/**
+ * DOC: HOST2GUC_SELF_CFG
+ *
+ * This message is used by Host KMD to setup of the `GuC Self Config KLVs`_.
+ *
+ * This message must be sent as `MMIO HXG Message`_.
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN = GUC_HXG_ORIGIN_HOST_                                |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_REQUEST_                                 |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 27:16 | DATA0 = MBZ                                                  |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  15:0 | ACTION = _`GUC_ACTION_HOST2GUC_SELF_CFG` = 0x0508            |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 | 31:16 | **KLV_KEY** - KLV key, see `GuC Self Config KLVs`_           |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  15:0 | **KLV_LEN** - KLV length                                     |
+ *  |   |       |                                                              |
+ *  |   |       |   - 32 bit KLV = 1                                           |
+ *  |   |       |   - 64 bit KLV = 2                                           |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 2 |  31:0 | **VALUE32** - Bits 31-0 of the KLV value                     |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 3 |  31:0 | **VALUE64** - Bits 63-32 of the KLV value (**KLV_LEN** = 2)  |
+ *  +---+-------+--------------------------------------------------------------+
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN = GUC_HXG_ORIGIN_GUC_                                 |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_RESPONSE_SUCCESS_                        |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  27:0 | DATA0 = **NUM** - 1 if KLV was parsed, 0 if not recognized   |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+#define GUC_ACTION_HOST2GUC_SELF_CFG			0x0508
+
+#define HOST2GUC_SELF_CFG_REQUEST_MSG_LEN		(GUC_HXG_REQUEST_MSG_MIN_LEN + 3u)
+#define HOST2GUC_SELF_CFG_REQUEST_MSG_0_MBZ		GUC_HXG_REQUEST_MSG_0_DATA0
+#define HOST2GUC_SELF_CFG_REQUEST_MSG_1_KLV_KEY		(0xffffU << 16)
+#define HOST2GUC_SELF_CFG_REQUEST_MSG_1_KLV_LEN		(0xffff << 0)
+#define HOST2GUC_SELF_CFG_REQUEST_MSG_2_VALUE32		GUC_HXG_REQUEST_MSG_n_DATAn
+#define HOST2GUC_SELF_CFG_REQUEST_MSG_3_VALUE64		GUC_HXG_REQUEST_MSG_n_DATAn
+
+#define HOST2GUC_SELF_CFG_RESPONSE_MSG_LEN		GUC_HXG_RESPONSE_MSG_MIN_LEN
+#define HOST2GUC_SELF_CFG_RESPONSE_MSG_0_NUM		GUC_HXG_RESPONSE_MSG_0_DATA0
+
+/**
+ * DOC: HOST2GUC_CONTROL_CTB
+ *
+ * This H2G action allows Vf Host to enable or disable H2G and G2H `CT Buffer`_.
+ *
+ * This message must be sent as `MMIO HXG Message`_.
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN = GUC_HXG_ORIGIN_HOST_                                |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_REQUEST_                                 |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 27:16 | DATA0 = MBZ                                                  |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  15:0 | ACTION = _`GUC_ACTION_HOST2GUC_CONTROL_CTB` = 0x4509         |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 |  31:0 | **CONTROL** - control `CTB based communication`_             |
+ *  |   |       |                                                              |
+ *  |   |       |   - _`GUC_CTB_CONTROL_DISABLE` = 0                           |
+ *  |   |       |   - _`GUC_CTB_CONTROL_ENABLE` = 1                            |
+ *  +---+-------+--------------------------------------------------------------+
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN = GUC_HXG_ORIGIN_GUC_                                 |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_RESPONSE_SUCCESS_                        |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  27:0 | DATA0 = MBZ                                                  |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+#define GUC_ACTION_HOST2GUC_CONTROL_CTB			0x4509
+
+#define HOST2GUC_CONTROL_CTB_REQUEST_MSG_LEN		(GUC_HXG_REQUEST_MSG_MIN_LEN + 1u)
+#define HOST2GUC_CONTROL_CTB_REQUEST_MSG_0_MBZ		GUC_HXG_REQUEST_MSG_0_DATA0
+#define HOST2GUC_CONTROL_CTB_REQUEST_MSG_1_CONTROL	GUC_HXG_REQUEST_MSG_n_DATAn
+#define   GUC_CTB_CONTROL_DISABLE			0u
+#define   GUC_CTB_CONTROL_ENABLE			1u
+
+#define HOST2GUC_CONTROL_CTB_RESPONSE_MSG_LEN		GUC_HXG_RESPONSE_MSG_MIN_LEN
+#define HOST2GUC_CONTROL_CTB_RESPONSE_MSG_0_MBZ		GUC_HXG_RESPONSE_MSG_0_DATA0
+
+/* legacy definitions */
+
+enum intel_guc_action {
+	INTEL_GUC_ACTION_DEFAULT = 0x0,
+	INTEL_GUC_ACTION_REQUEST_PREEMPTION = 0x2,
+	INTEL_GUC_ACTION_REQUEST_ENGINE_RESET = 0x3,
+	INTEL_GUC_ACTION_ALLOCATE_DOORBELL = 0x10,
+	INTEL_GUC_ACTION_DEALLOCATE_DOORBELL = 0x20,
+	INTEL_GUC_ACTION_LOG_BUFFER_FILE_FLUSH_COMPLETE = 0x30,
+	INTEL_GUC_ACTION_UK_LOG_ENABLE_LOGGING = 0x40,
+	INTEL_GUC_ACTION_FORCE_LOG_BUFFER_FLUSH = 0x302,
+	INTEL_GUC_ACTION_ENTER_S_STATE = 0x501,
+	INTEL_GUC_ACTION_EXIT_S_STATE = 0x502,
+	INTEL_GUC_ACTION_GLOBAL_SCHED_POLICY_CHANGE = 0x506,
+	INTEL_GUC_ACTION_SCHED_CONTEXT = 0x1000,
+	INTEL_GUC_ACTION_SCHED_CONTEXT_MODE_SET = 0x1001,
+	INTEL_GUC_ACTION_SCHED_CONTEXT_MODE_DONE = 0x1002,
+	INTEL_GUC_ACTION_SCHED_ENGINE_MODE_SET = 0x1003,
+	INTEL_GUC_ACTION_SCHED_ENGINE_MODE_DONE = 0x1004,
+	INTEL_GUC_ACTION_V69_SET_CONTEXT_PRIORITY = 0x1005,
+	INTEL_GUC_ACTION_V69_SET_CONTEXT_EXECUTION_QUANTUM = 0x1006,
+	INTEL_GUC_ACTION_V69_SET_CONTEXT_PREEMPTION_TIMEOUT = 0x1007,
+	INTEL_GUC_ACTION_CONTEXT_RESET_NOTIFICATION = 0x1008,
+	INTEL_GUC_ACTION_ENGINE_FAILURE_NOTIFICATION = 0x1009,
+	INTEL_GUC_ACTION_HOST2GUC_UPDATE_CONTEXT_POLICIES = 0x100B,
+	INTEL_GUC_ACTION_SETUP_PC_GUCRC = 0x3004,
+	INTEL_GUC_ACTION_AUTHENTICATE_HUC = 0x4000,
+	INTEL_GUC_ACTION_GET_HWCONFIG = 0x4100,
+	INTEL_GUC_ACTION_REGISTER_CONTEXT = 0x4502,
+	INTEL_GUC_ACTION_DEREGISTER_CONTEXT = 0x4503,
+	INTEL_GUC_ACTION_DEREGISTER_CONTEXT_DONE = 0x4600,
+	INTEL_GUC_ACTION_REGISTER_CONTEXT_MULTI_LRC = 0x4601,
+	INTEL_GUC_ACTION_CLIENT_SOFT_RESET = 0x5507,
+	INTEL_GUC_ACTION_SET_ENG_UTIL_BUFF = 0x550A,
+	INTEL_GUC_ACTION_STATE_CAPTURE_NOTIFICATION = 0x8002,
+	INTEL_GUC_ACTION_NOTIFY_FLUSH_LOG_BUFFER_TO_FILE = 0x8003,
+	INTEL_GUC_ACTION_NOTIFY_CRASH_DUMP_POSTED = 0x8004,
+	INTEL_GUC_ACTION_NOTIFY_EXCEPTION = 0x8005,
+	INTEL_GUC_ACTION_LIMIT
+};
+
+enum intel_guc_rc_options {
+	INTEL_GUCRC_HOST_CONTROL,
+	INTEL_GUCRC_FIRMWARE_CONTROL,
+};
+
+enum intel_guc_preempt_options {
+	INTEL_GUC_PREEMPT_OPTION_DROP_WORK_Q = 0x4,
+	INTEL_GUC_PREEMPT_OPTION_DROP_SUBMIT_Q = 0x8,
+};
+
+enum intel_guc_report_status {
+	INTEL_GUC_REPORT_STATUS_UNKNOWN = 0x0,
+	INTEL_GUC_REPORT_STATUS_ACKED = 0x1,
+	INTEL_GUC_REPORT_STATUS_ERROR = 0x2,
+	INTEL_GUC_REPORT_STATUS_COMPLETE = 0x4,
+};
+
+enum intel_guc_sleep_state_status {
+	INTEL_GUC_SLEEP_STATE_SUCCESS = 0x1,
+	INTEL_GUC_SLEEP_STATE_PREEMPT_TO_IDLE_FAILED = 0x2,
+	INTEL_GUC_SLEEP_STATE_ENGINE_RESET_FAILED = 0x3
+#define INTEL_GUC_SLEEP_STATE_INVALID_MASK 0x80000000
+};
+
+#define GUC_LOG_CONTROL_LOGGING_ENABLED	(1 << 0)
+#define GUC_LOG_CONTROL_VERBOSITY_SHIFT	4
+#define GUC_LOG_CONTROL_VERBOSITY_MASK	(0xF << GUC_LOG_CONTROL_VERBOSITY_SHIFT)
+#define GUC_LOG_CONTROL_DEFAULT_LOGGING	(1 << 8)
+
+enum intel_guc_state_capture_event_status {
+	INTEL_GUC_STATE_CAPTURE_EVENT_STATUS_SUCCESS = 0x0,
+	INTEL_GUC_STATE_CAPTURE_EVENT_STATUS_NOSPACE = 0x1,
+};
+
+#define INTEL_GUC_STATE_CAPTURE_EVENT_STATUS_MASK      0x000000FF
+
+#endif /* _ABI_GUC_ACTIONS_ABI_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_slpc_abi.h b/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_slpc_abi.h
new file mode 100644
index 000000000..4c840a263
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_slpc_abi.h
@@ -0,0 +1,240 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _GUC_ACTIONS_SLPC_ABI_H_
+#define _GUC_ACTIONS_SLPC_ABI_H_
+
+#include <linux/types.h>
+
+/**
+ * DOC: SLPC SHARED DATA STRUCTURE
+ *
+ *  +----+------+--------------------------------------------------------------+
+ *  | CL | Bytes| Description                                                  |
+ *  +====+======+==============================================================+
+ *  | 1  | 0-3  | SHARED DATA SIZE                                             |
+ *  |    +------+--------------------------------------------------------------+
+ *  |    | 4-7  | GLOBAL STATE                                                 |
+ *  |    +------+--------------------------------------------------------------+
+ *  |    | 8-11 | DISPLAY DATA ADDRESS                                         |
+ *  |    +------+--------------------------------------------------------------+
+ *  |    | 12:63| PADDING                                                      |
+ *  +----+------+--------------------------------------------------------------+
+ *  |    | 0:63 | PADDING(PLATFORM INFO)                                       |
+ *  +----+------+--------------------------------------------------------------+
+ *  | 3  | 0-3  | TASK STATE DATA                                              |
+ *  +    +------+--------------------------------------------------------------+
+ *  |    | 4:63 | PADDING                                                      |
+ *  +----+------+--------------------------------------------------------------+
+ *  |4-21|0:1087| OVERRIDE PARAMS AND BIT FIELDS                               |
+ *  +----+------+--------------------------------------------------------------+
+ *  |    |      | PADDING + EXTRA RESERVED PAGE                                |
+ *  +----+------+--------------------------------------------------------------+
+ */
+
+/*
+ * SLPC exposes certain parameters for global configuration by the host.
+ * These are referred to as override parameters, because in most cases
+ * the host will not need to modify the default values used by SLPC.
+ * SLPC remembers the default values which allows the host to easily restore
+ * them by simply unsetting the override. The host can set or unset override
+ * parameters during SLPC (re-)initialization using the SLPC Reset event.
+ * The host can also set or unset override parameters on the fly using the
+ * Parameter Set and Parameter Unset events
+ */
+
+#define SLPC_MAX_OVERRIDE_PARAMETERS		256
+#define SLPC_OVERRIDE_BITFIELD_SIZE \
+		(SLPC_MAX_OVERRIDE_PARAMETERS / 32)
+
+#define SLPC_PAGE_SIZE_BYTES			4096
+#define SLPC_CACHELINE_SIZE_BYTES		64
+#define SLPC_SHARED_DATA_SIZE_BYTE_HEADER	SLPC_CACHELINE_SIZE_BYTES
+#define SLPC_SHARED_DATA_SIZE_BYTE_PLATFORM_INFO	SLPC_CACHELINE_SIZE_BYTES
+#define SLPC_SHARED_DATA_SIZE_BYTE_TASK_STATE	SLPC_CACHELINE_SIZE_BYTES
+#define SLPC_SHARED_DATA_MODE_DEFN_TABLE_SIZE	SLPC_PAGE_SIZE_BYTES
+#define SLPC_SHARED_DATA_SIZE_BYTE_MAX		(2 * SLPC_PAGE_SIZE_BYTES)
+
+/*
+ * Cacheline size aligned (Total size needed for
+ * SLPM_KMD_MAX_OVERRIDE_PARAMETERS=256 is 1088 bytes)
+ */
+#define SLPC_OVERRIDE_PARAMS_TOTAL_BYTES	(((((SLPC_MAX_OVERRIDE_PARAMETERS * 4) \
+						+ ((SLPC_MAX_OVERRIDE_PARAMETERS / 32) * 4)) \
+		+ (SLPC_CACHELINE_SIZE_BYTES - 1)) / SLPC_CACHELINE_SIZE_BYTES) * \
+					SLPC_CACHELINE_SIZE_BYTES)
+
+#define SLPC_SHARED_DATA_SIZE_BYTE_OTHER	(SLPC_SHARED_DATA_SIZE_BYTE_MAX - \
+					(SLPC_SHARED_DATA_SIZE_BYTE_HEADER \
+					+ SLPC_SHARED_DATA_SIZE_BYTE_PLATFORM_INFO \
+					+ SLPC_SHARED_DATA_SIZE_BYTE_TASK_STATE \
+					+ SLPC_OVERRIDE_PARAMS_TOTAL_BYTES \
+					+ SLPC_SHARED_DATA_MODE_DEFN_TABLE_SIZE))
+
+enum slpc_task_enable {
+	SLPC_PARAM_TASK_DEFAULT = 0,
+	SLPC_PARAM_TASK_ENABLED,
+	SLPC_PARAM_TASK_DISABLED,
+	SLPC_PARAM_TASK_UNKNOWN
+};
+
+enum slpc_global_state {
+	SLPC_GLOBAL_STATE_NOT_RUNNING = 0,
+	SLPC_GLOBAL_STATE_INITIALIZING = 1,
+	SLPC_GLOBAL_STATE_RESETTING = 2,
+	SLPC_GLOBAL_STATE_RUNNING = 3,
+	SLPC_GLOBAL_STATE_SHUTTING_DOWN = 4,
+	SLPC_GLOBAL_STATE_ERROR = 5
+};
+
+enum slpc_param_id {
+	SLPC_PARAM_TASK_ENABLE_GTPERF = 0,
+	SLPC_PARAM_TASK_DISABLE_GTPERF = 1,
+	SLPC_PARAM_TASK_ENABLE_BALANCER = 2,
+	SLPC_PARAM_TASK_DISABLE_BALANCER = 3,
+	SLPC_PARAM_TASK_ENABLE_DCC = 4,
+	SLPC_PARAM_TASK_DISABLE_DCC = 5,
+	SLPC_PARAM_GLOBAL_MIN_GT_UNSLICE_FREQ_MHZ = 6,
+	SLPC_PARAM_GLOBAL_MAX_GT_UNSLICE_FREQ_MHZ = 7,
+	SLPC_PARAM_GLOBAL_MIN_GT_SLICE_FREQ_MHZ = 8,
+	SLPC_PARAM_GLOBAL_MAX_GT_SLICE_FREQ_MHZ = 9,
+	SLPC_PARAM_GTPERF_THRESHOLD_MAX_FPS = 10,
+	SLPC_PARAM_GLOBAL_DISABLE_GT_FREQ_MANAGEMENT = 11,
+	SLPC_PARAM_GTPERF_ENABLE_FRAMERATE_STALLING = 12,
+	SLPC_PARAM_GLOBAL_DISABLE_RC6_MODE_CHANGE = 13,
+	SLPC_PARAM_GLOBAL_OC_UNSLICE_FREQ_MHZ = 14,
+	SLPC_PARAM_GLOBAL_OC_SLICE_FREQ_MHZ = 15,
+	SLPC_PARAM_GLOBAL_ENABLE_IA_GT_BALANCING = 16,
+	SLPC_PARAM_GLOBAL_ENABLE_ADAPTIVE_BURST_TURBO = 17,
+	SLPC_PARAM_GLOBAL_ENABLE_EVAL_MODE = 18,
+	SLPC_PARAM_GLOBAL_ENABLE_BALANCER_IN_NON_GAMING_MODE = 19,
+	SLPC_PARAM_GLOBAL_RT_MODE_TURBO_FREQ_DELTA_MHZ = 20,
+	SLPC_PARAM_PWRGATE_RC_MODE = 21,
+	SLPC_PARAM_EDR_MODE_COMPUTE_TIMEOUT_MS = 22,
+	SLPC_PARAM_EDR_QOS_FREQ_MHZ = 23,
+	SLPC_PARAM_MEDIA_FF_RATIO_MODE = 24,
+	SLPC_PARAM_ENABLE_IA_FREQ_LIMITING = 25,
+	SLPC_PARAM_STRATEGIES = 26,
+	SLPC_PARAM_POWER_PROFILE = 27,
+	SLPC_PARAM_IGNORE_EFFICIENT_FREQUENCY = 28,
+	SLPC_MAX_PARAM = 32,
+};
+
+enum slpc_media_ratio_mode {
+	SLPC_MEDIA_RATIO_MODE_DYNAMIC_CONTROL = 0,
+	SLPC_MEDIA_RATIO_MODE_FIXED_ONE_TO_ONE = 1,
+	SLPC_MEDIA_RATIO_MODE_FIXED_ONE_TO_TWO = 2,
+};
+
+enum slpc_event_id {
+	SLPC_EVENT_RESET = 0,
+	SLPC_EVENT_SHUTDOWN = 1,
+	SLPC_EVENT_PLATFORM_INFO_CHANGE = 2,
+	SLPC_EVENT_DISPLAY_MODE_CHANGE = 3,
+	SLPC_EVENT_FLIP_COMPLETE = 4,
+	SLPC_EVENT_QUERY_TASK_STATE = 5,
+	SLPC_EVENT_PARAMETER_SET = 6,
+	SLPC_EVENT_PARAMETER_UNSET = 7,
+};
+
+struct slpc_task_state_data {
+	union {
+		u32 task_status_padding;
+		struct {
+			u32 status;
+#define SLPC_GTPERF_TASK_ENABLED	REG_BIT(0)
+#define SLPC_DCC_TASK_ENABLED		REG_BIT(11)
+#define SLPC_IN_DCC			REG_BIT(12)
+#define SLPC_BALANCER_ENABLED		REG_BIT(15)
+#define SLPC_IBC_TASK_ENABLED		REG_BIT(16)
+#define SLPC_BALANCER_IA_LMT_ENABLED	REG_BIT(17)
+#define SLPC_BALANCER_IA_LMT_ACTIVE	REG_BIT(18)
+		};
+	};
+	union {
+		u32 freq_padding;
+		struct {
+#define SLPC_MAX_UNSLICE_FREQ_MASK	REG_GENMASK(7, 0)
+#define SLPC_MIN_UNSLICE_FREQ_MASK	REG_GENMASK(15, 8)
+#define SLPC_MAX_SLICE_FREQ_MASK	REG_GENMASK(23, 16)
+#define SLPC_MIN_SLICE_FREQ_MASK	REG_GENMASK(31, 24)
+			u32 freq;
+		};
+	};
+} __packed;
+
+struct slpc_shared_data_header {
+	/* Total size in bytes of this shared buffer. */
+	u32 size;
+	u32 global_state;
+	u32 display_data_addr;
+} __packed;
+
+struct slpc_override_params {
+	u32 bits[SLPC_OVERRIDE_BITFIELD_SIZE];
+	u32 values[SLPC_MAX_OVERRIDE_PARAMETERS];
+} __packed;
+
+struct slpc_shared_data {
+	struct slpc_shared_data_header header;
+	u8 shared_data_header_pad[SLPC_SHARED_DATA_SIZE_BYTE_HEADER -
+				sizeof(struct slpc_shared_data_header)];
+
+	u8 platform_info_pad[SLPC_SHARED_DATA_SIZE_BYTE_PLATFORM_INFO];
+
+	struct slpc_task_state_data task_state_data;
+	u8 task_state_data_pad[SLPC_SHARED_DATA_SIZE_BYTE_TASK_STATE -
+				sizeof(struct slpc_task_state_data)];
+
+	struct slpc_override_params override_params;
+	u8 override_params_pad[SLPC_OVERRIDE_PARAMS_TOTAL_BYTES -
+				sizeof(struct slpc_override_params)];
+
+	u8 shared_data_pad[SLPC_SHARED_DATA_SIZE_BYTE_OTHER];
+
+	/* PAGE 2 (4096 bytes), mode based parameter will be removed soon */
+	u8 reserved_mode_definition[4096];
+} __packed;
+
+/**
+ * DOC: SLPC H2G MESSAGE FORMAT
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN = GUC_HXG_ORIGIN_HOST_                                |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_REQUEST_                                 |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 27:16 | DATA0 = MBZ                                                  |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  15:0 | ACTION = _`GUC_ACTION_HOST2GUC_PC_SLPM_REQUEST` = 0x3003     |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 |  31:8 | **EVENT_ID**                                                 |
+ *  +   +-------+--------------------------------------------------------------+
+ *  |   |   7:0 | **EVENT_ARGC** - number of data arguments                    |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 2 |  31:0 | **EVENT_DATA1**                                              |
+ *  +---+-------+--------------------------------------------------------------+
+ *  |...|  31:0 | ...                                                          |
+ *  +---+-------+--------------------------------------------------------------+
+ *  |2+n|  31:0 | **EVENT_DATAn**                                              |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_ACTION_HOST2GUC_PC_SLPC_REQUEST		0x3003
+
+#define HOST2GUC_PC_SLPC_REQUEST_MSG_MIN_LEN \
+				(GUC_HXG_REQUEST_MSG_MIN_LEN + 1u)
+#define HOST2GUC_PC_SLPC_EVENT_MAX_INPUT_ARGS		9
+#define HOST2GUC_PC_SLPC_REQUEST_MSG_MAX_LEN \
+		(HOST2GUC_PC_SLPC_REQUEST_REQUEST_MSG_MIN_LEN + \
+			HOST2GUC_PC_SLPC_EVENT_MAX_INPUT_ARGS)
+#define HOST2GUC_PC_SLPC_REQUEST_MSG_0_MBZ		GUC_HXG_REQUEST_MSG_0_DATA0
+#define HOST2GUC_PC_SLPC_REQUEST_MSG_1_EVENT_ID		(0xff << 8)
+#define HOST2GUC_PC_SLPC_REQUEST_MSG_1_EVENT_ARGC	(0xff << 0)
+#define HOST2GUC_PC_SLPC_REQUEST_MSG_N_EVENT_DATA_N	GUC_HXG_REQUEST_MSG_n_DATAn
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/abi/guc_communication_ctb_abi.h b/drivers/gpu/drm/i915/gt/uc/abi/guc_communication_ctb_abi.h
new file mode 100644
index 000000000..28b8387f9
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/abi/guc_communication_ctb_abi.h
@@ -0,0 +1,191 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2021 Intel Corporation
+ */
+
+#ifndef _ABI_GUC_COMMUNICATION_CTB_ABI_H
+#define _ABI_GUC_COMMUNICATION_CTB_ABI_H
+
+#include <linux/types.h>
+#include <linux/build_bug.h>
+
+#include "guc_messages_abi.h"
+
+/**
+ * DOC: CT Buffer
+ *
+ * Circular buffer used to send `CTB Message`_
+ */
+
+/**
+ * DOC: CTB Descriptor
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |  31:0 | **HEAD** - offset (in dwords) to the last dword that was     |
+ *  |   |       | read from the `CT Buffer`_.                                  |
+ *  |   |       | It can only be updated by the receiver.                      |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 |  31:0 | **TAIL** - offset (in dwords) to the last dword that was     |
+ *  |   |       | written to the `CT Buffer`_.                                 |
+ *  |   |       | It can only be updated by the sender.                        |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 2 |  31:0 | **STATUS** - status of the CTB                               |
+ *  |   |       |                                                              |
+ *  |   |       |   - _`GUC_CTB_STATUS_NO_ERROR` = 0 (normal operation)        |
+ *  |   |       |   - _`GUC_CTB_STATUS_OVERFLOW` = 1 (head/tail too large)     |
+ *  |   |       |   - _`GUC_CTB_STATUS_UNDERFLOW` = 2 (truncated message)      |
+ *  |   |       |   - _`GUC_CTB_STATUS_MISMATCH` = 4 (head/tail modified)      |
+ *  |   |       |   - _`GUC_CTB_STATUS_UNUSED` = 8 (CTB is not in use)         |
+ *  +---+-------+--------------------------------------------------------------+
+ *  |...|       | RESERVED = MBZ                                               |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 15|  31:0 | RESERVED = MBZ                                               |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+struct guc_ct_buffer_desc {
+	u32 head;
+	u32 tail;
+	u32 status;
+#define GUC_CTB_STATUS_NO_ERROR				0
+#define GUC_CTB_STATUS_OVERFLOW				BIT(0)
+#define GUC_CTB_STATUS_UNDERFLOW			BIT(1)
+#define GUC_CTB_STATUS_MISMATCH				BIT(2)
+#define GUC_CTB_STATUS_UNUSED				BIT(3)
+	u32 reserved[13];
+} __packed;
+static_assert(sizeof(struct guc_ct_buffer_desc) == 64);
+
+/**
+ * DOC: CTB Message
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 | 31:16 | **FENCE** - message identifier                               |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 15:12 | **FORMAT** - format of the CTB message                       |
+ *  |   |       |  - _`GUC_CTB_FORMAT_HXG` = 0 - see `CTB HXG Message`_        |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  11:8 | **RESERVED**                                                 |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |   7:0 | **NUM_DWORDS** - length of the CTB message (w/o header)      |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 |  31:0 | optional (depends on FORMAT)                                 |
+ *  +---+-------+                                                              |
+ *  |...|       |                                                              |
+ *  +---+-------+                                                              |
+ *  | n |  31:0 |                                                              |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_CTB_HDR_LEN				1u
+#define GUC_CTB_MSG_MIN_LEN			GUC_CTB_HDR_LEN
+#define GUC_CTB_MSG_MAX_LEN			256u
+#define GUC_CTB_MSG_0_FENCE			(0xffffU << 16)
+#define GUC_CTB_MSG_0_FORMAT			(0xf << 12)
+#define   GUC_CTB_FORMAT_HXG			0u
+#define GUC_CTB_MSG_0_RESERVED			(0xf << 8)
+#define GUC_CTB_MSG_0_NUM_DWORDS		(0xff << 0)
+
+/**
+ * DOC: CTB HXG Message
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 | 31:16 | FENCE                                                        |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 15:12 | FORMAT = GUC_CTB_FORMAT_HXG_                                 |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  11:8 | RESERVED = MBZ                                               |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |   7:0 | NUM_DWORDS = length (in dwords) of the embedded HXG message  |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 |  31:0 |                                                              |
+ *  +---+-------+                                                              |
+ *  |...|       | [Embedded `HXG Message`_]                                    |
+ *  +---+-------+                                                              |
+ *  | n |  31:0 |                                                              |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_CTB_HXG_MSG_MIN_LEN		(GUC_CTB_MSG_MIN_LEN + GUC_HXG_MSG_MIN_LEN)
+#define GUC_CTB_HXG_MSG_MAX_LEN		GUC_CTB_MSG_MAX_LEN
+
+/**
+ * DOC: CTB based communication
+ *
+ * The CTB (command transport buffer) communication between Host and GuC
+ * is based on u32 data stream written to the shared buffer. One buffer can
+ * be used to transmit data only in one direction (one-directional channel).
+ *
+ * Current status of the each buffer is stored in the buffer descriptor.
+ * Buffer descriptor holds tail and head fields that represents active data
+ * stream. The tail field is updated by the data producer (sender), and head
+ * field is updated by the data consumer (receiver)::
+ *
+ *      +------------+
+ *      | DESCRIPTOR |          +=================+============+========+
+ *      +============+          |                 | MESSAGE(s) |        |
+ *      | address    |--------->+=================+============+========+
+ *      +------------+
+ *      | head       |          ^-----head--------^
+ *      +------------+
+ *      | tail       |          ^---------tail-----------------^
+ *      +------------+
+ *      | size       |          ^---------------size--------------------^
+ *      +------------+
+ *
+ * Each message in data stream starts with the single u32 treated as a header,
+ * followed by optional set of u32 data that makes message specific payload::
+ *
+ *      +------------+---------+---------+---------+
+ *      |         MESSAGE                          |
+ *      +------------+---------+---------+---------+
+ *      |   msg[0]   |   [1]   |   ...   |  [n-1]  |
+ *      +------------+---------+---------+---------+
+ *      |   MESSAGE  |       MESSAGE PAYLOAD       |
+ *      +   HEADER   +---------+---------+---------+
+ *      |            |    0    |   ...   |    n    |
+ *      +======+=====+=========+=========+=========+
+ *      | 31:16| code|         |         |         |
+ *      +------+-----+         |         |         |
+ *      |  15:5|flags|         |         |         |
+ *      +------+-----+         |         |         |
+ *      |   4:0|  len|         |         |         |
+ *      +------+-----+---------+---------+---------+
+ *
+ *                   ^-------------len-------------^
+ *
+ * The message header consists of:
+ *
+ * - **len**, indicates length of the message payload (in u32)
+ * - **code**, indicates message code
+ * - **flags**, holds various bits to control message handling
+ */
+
+/*
+ * Definition of the command transport message header (DW0)
+ *
+ * bit[4..0]	message len (in dwords)
+ * bit[7..5]	reserved
+ * bit[8]	response (G2H only)
+ * bit[8]	write fence to desc (H2G only)
+ * bit[9]	write status to H2G buff (H2G only)
+ * bit[10]	send status back via G2H (H2G only)
+ * bit[15..11]	reserved
+ * bit[31..16]	action code
+ */
+#define GUC_CT_MSG_LEN_SHIFT			0
+#define GUC_CT_MSG_LEN_MASK			0x1F
+#define GUC_CT_MSG_IS_RESPONSE			(1 << 8)
+#define GUC_CT_MSG_WRITE_FENCE_TO_DESC		(1 << 8)
+#define GUC_CT_MSG_WRITE_STATUS_TO_BUFF		(1 << 9)
+#define GUC_CT_MSG_SEND_STATUS			(1 << 10)
+#define GUC_CT_MSG_ACTION_SHIFT			16
+#define GUC_CT_MSG_ACTION_MASK			0xFFFF
+
+#endif /* _ABI_GUC_COMMUNICATION_CTB_ABI_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/abi/guc_communication_mmio_abi.h b/drivers/gpu/drm/i915/gt/uc/abi/guc_communication_mmio_abi.h
new file mode 100644
index 000000000..9baa3cb07
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/abi/guc_communication_mmio_abi.h
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2021 Intel Corporation
+ */
+
+#ifndef _ABI_GUC_COMMUNICATION_MMIO_ABI_H
+#define _ABI_GUC_COMMUNICATION_MMIO_ABI_H
+
+/**
+ * DOC: GuC MMIO based communication
+ *
+ * The MMIO based communication between Host and GuC relies on special
+ * hardware registers which format could be defined by the software
+ * (so called scratch registers).
+ *
+ * Each MMIO based message, both Host to GuC (H2G) and GuC to Host (G2H)
+ * messages, which maximum length depends on number of available scratch
+ * registers, is directly written into those scratch registers.
+ *
+ * For Gen9+, there are 16 software scratch registers 0xC180-0xC1B8,
+ * but no H2G command takes more than 4 parameters and the GuC firmware
+ * itself uses an 4-element array to store the H2G message.
+ *
+ * For Gen11+, there are additional 4 registers 0x190240-0x19024C, which
+ * are, regardless on lower count, preferred over legacy ones.
+ *
+ * The MMIO based communication is mainly used during driver initialization
+ * phase to setup the `CTB based communication`_ that will be used afterwards.
+ */
+
+#define GUC_MAX_MMIO_MSG_LEN		4
+
+/**
+ * DOC: MMIO HXG Message
+ *
+ * Format of the MMIO messages follows definitions of `HXG Message`_.
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |  31:0 |                                                              |
+ *  +---+-------+                                                              |
+ *  |...|       | [Embedded `HXG Message`_]                                    |
+ *  +---+-------+                                                              |
+ *  | n |  31:0 |                                                              |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#endif /* _ABI_GUC_COMMUNICATION_MMIO_ABI_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/abi/guc_errors_abi.h b/drivers/gpu/drm/i915/gt/uc/abi/guc_errors_abi.h
new file mode 100644
index 000000000..8085fb181
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/abi/guc_errors_abi.h
@@ -0,0 +1,41 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2021 Intel Corporation
+ */
+
+#ifndef _ABI_GUC_ERRORS_ABI_H
+#define _ABI_GUC_ERRORS_ABI_H
+
+enum intel_guc_response_status {
+	INTEL_GUC_RESPONSE_STATUS_SUCCESS = 0x0,
+	INTEL_GUC_RESPONSE_NOT_SUPPORTED = 0x20,
+	INTEL_GUC_RESPONSE_NO_ATTRIBUTE_TABLE = 0x201,
+	INTEL_GUC_RESPONSE_NO_DECRYPTION_KEY = 0x202,
+	INTEL_GUC_RESPONSE_DECRYPTION_FAILED = 0x204,
+	INTEL_GUC_RESPONSE_STATUS_GENERIC_FAIL = 0xF000,
+};
+
+enum intel_guc_load_status {
+	INTEL_GUC_LOAD_STATUS_DEFAULT                          = 0x00,
+	INTEL_GUC_LOAD_STATUS_START                            = 0x01,
+	INTEL_GUC_LOAD_STATUS_ERROR_DEVID_BUILD_MISMATCH       = 0x02,
+	INTEL_GUC_LOAD_STATUS_GUC_PREPROD_BUILD_MISMATCH       = 0x03,
+	INTEL_GUC_LOAD_STATUS_ERROR_DEVID_INVALID_GUCTYPE      = 0x04,
+	INTEL_GUC_LOAD_STATUS_GDT_DONE                         = 0x10,
+	INTEL_GUC_LOAD_STATUS_IDT_DONE                         = 0x20,
+	INTEL_GUC_LOAD_STATUS_LAPIC_DONE                       = 0x30,
+	INTEL_GUC_LOAD_STATUS_GUCINT_DONE                      = 0x40,
+	INTEL_GUC_LOAD_STATUS_DPC_READY                        = 0x50,
+	INTEL_GUC_LOAD_STATUS_DPC_ERROR                        = 0x60,
+	INTEL_GUC_LOAD_STATUS_EXCEPTION                        = 0x70,
+	INTEL_GUC_LOAD_STATUS_INIT_DATA_INVALID                = 0x71,
+	INTEL_GUC_LOAD_STATUS_PXP_TEARDOWN_CTRL_ENABLED        = 0x72,
+	INTEL_GUC_LOAD_STATUS_INVALID_INIT_DATA_RANGE_START,
+	INTEL_GUC_LOAD_STATUS_MPU_DATA_INVALID                 = 0x73,
+	INTEL_GUC_LOAD_STATUS_INIT_MMIO_SAVE_RESTORE_INVALID   = 0x74,
+	INTEL_GUC_LOAD_STATUS_INVALID_INIT_DATA_RANGE_END,
+
+	INTEL_GUC_LOAD_STATUS_READY                            = 0xF0,
+};
+
+#endif /* _ABI_GUC_ERRORS_ABI_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/abi/guc_klvs_abi.h b/drivers/gpu/drm/i915/gt/uc/abi/guc_klvs_abi.h
new file mode 100644
index 000000000..4a59478c3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/abi/guc_klvs_abi.h
@@ -0,0 +1,97 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _ABI_GUC_KLVS_ABI_H
+#define _ABI_GUC_KLVS_ABI_H
+
+#include <linux/types.h>
+
+/**
+ * DOC: GuC KLV
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 | 31:16 | **KEY** - KLV key identifier                                 |
+ *  |   |       |   - `GuC Self Config KLVs`_                                  |
+ *  |   |       |                                                              |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  15:0 | **LEN** - length of VALUE (in 32bit dwords)                  |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 |  31:0 | **VALUE** - actual value of the KLV (format depends on KEY)  |
+ *  +---+-------+                                                              |
+ *  |...|       |                                                              |
+ *  +---+-------+                                                              |
+ *  | n |  31:0 |                                                              |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_KLV_LEN_MIN				1u
+#define GUC_KLV_0_KEY				(0xffff << 16)
+#define GUC_KLV_0_LEN				(0xffff << 0)
+#define GUC_KLV_n_VALUE				(0xffffffff << 0)
+
+/**
+ * DOC: GuC Self Config KLVs
+ *
+ * `GuC KLV`_ keys available for use with HOST2GUC_SELF_CFG_.
+ *
+ * _`GUC_KLV_SELF_CFG_H2G_CTB_ADDR` : 0x0902
+ *      Refers to 64 bit Global Gfx address of H2G `CT Buffer`_.
+ *      Should be above WOPCM address but below APIC base address for native mode.
+ *
+ * _`GUC_KLV_SELF_CFG_H2G_CTB_DESCRIPTOR_ADDR` : 0x0903
+ *      Refers to 64 bit Global Gfx address of H2G `CTB Descriptor`_.
+ *      Should be above WOPCM address but below APIC base address for native mode.
+ *
+ * _`GUC_KLV_SELF_CFG_H2G_CTB_SIZE` : 0x0904
+ *      Refers to size of H2G `CT Buffer`_ in bytes.
+ *      Should be a multiple of 4K.
+ *
+ * _`GUC_KLV_SELF_CFG_G2H_CTB_ADDR` : 0x0905
+ *      Refers to 64 bit Global Gfx address of G2H `CT Buffer`_.
+ *      Should be above WOPCM address but below APIC base address for native mode.
+ *
+ * _`GUC_KLV_SELF_CFG_G2H_CTB_DESCRIPTOR_ADDR` : 0x0906
+ *      Refers to 64 bit Global Gfx address of G2H `CTB Descriptor`_.
+ *      Should be above WOPCM address but below APIC base address for native mode.
+ *
+ * _`GUC_KLV_SELF_CFG_G2H_CTB_SIZE` : 0x0907
+ *      Refers to size of G2H `CT Buffer`_ in bytes.
+ *      Should be a multiple of 4K.
+ */
+
+#define GUC_KLV_SELF_CFG_H2G_CTB_ADDR_KEY		0x0902
+#define GUC_KLV_SELF_CFG_H2G_CTB_ADDR_LEN		2u
+
+#define GUC_KLV_SELF_CFG_H2G_CTB_DESCRIPTOR_ADDR_KEY	0x0903
+#define GUC_KLV_SELF_CFG_H2G_CTB_DESCRIPTOR_ADDR_LEN	2u
+
+#define GUC_KLV_SELF_CFG_H2G_CTB_SIZE_KEY		0x0904
+#define GUC_KLV_SELF_CFG_H2G_CTB_SIZE_LEN		1u
+
+#define GUC_KLV_SELF_CFG_G2H_CTB_ADDR_KEY		0x0905
+#define GUC_KLV_SELF_CFG_G2H_CTB_ADDR_LEN		2u
+
+#define GUC_KLV_SELF_CFG_G2H_CTB_DESCRIPTOR_ADDR_KEY	0x0906
+#define GUC_KLV_SELF_CFG_G2H_CTB_DESCRIPTOR_ADDR_LEN	2u
+
+#define GUC_KLV_SELF_CFG_G2H_CTB_SIZE_KEY		0x0907
+#define GUC_KLV_SELF_CFG_G2H_CTB_SIZE_LEN		1u
+
+/*
+ * Per context scheduling policy update keys.
+ */
+enum  {
+	GUC_CONTEXT_POLICIES_KLV_ID_EXECUTION_QUANTUM			= 0x2001,
+	GUC_CONTEXT_POLICIES_KLV_ID_PREEMPTION_TIMEOUT			= 0x2002,
+	GUC_CONTEXT_POLICIES_KLV_ID_SCHEDULING_PRIORITY			= 0x2003,
+	GUC_CONTEXT_POLICIES_KLV_ID_PREEMPT_TO_IDLE_ON_QUANTUM_EXPIRY	= 0x2004,
+	GUC_CONTEXT_POLICIES_KLV_ID_SLPM_GT_FREQUENCY			= 0x2005,
+
+	GUC_CONTEXT_POLICIES_KLV_NUM_IDS = 5,
+};
+
+#endif /* _ABI_GUC_KLVS_ABI_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/abi/guc_messages_abi.h b/drivers/gpu/drm/i915/gt/uc/abi/guc_messages_abi.h
new file mode 100644
index 000000000..7d5ba4d97
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/abi/guc_messages_abi.h
@@ -0,0 +1,234 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2021 Intel Corporation
+ */
+
+#ifndef _ABI_GUC_MESSAGES_ABI_H
+#define _ABI_GUC_MESSAGES_ABI_H
+
+/**
+ * DOC: HXG Message
+ *
+ * All messages exchanged with GuC are defined using 32 bit dwords.
+ * First dword is treated as a message header. Remaining dwords are optional.
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  |   |       |                                                              |
+ *  | 0 |    31 | **ORIGIN** - originator of the message                       |
+ *  |   |       |   - _`GUC_HXG_ORIGIN_HOST` = 0                               |
+ *  |   |       |   - _`GUC_HXG_ORIGIN_GUC` = 1                                |
+ *  |   |       |                                                              |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | **TYPE** - message type                                      |
+ *  |   |       |   - _`GUC_HXG_TYPE_REQUEST` = 0                              |
+ *  |   |       |   - _`GUC_HXG_TYPE_EVENT` = 1                                |
+ *  |   |       |   - _`GUC_HXG_TYPE_NO_RESPONSE_BUSY` = 3                     |
+ *  |   |       |   - _`GUC_HXG_TYPE_NO_RESPONSE_RETRY` = 5                    |
+ *  |   |       |   - _`GUC_HXG_TYPE_RESPONSE_FAILURE` = 6                     |
+ *  |   |       |   - _`GUC_HXG_TYPE_RESPONSE_SUCCESS` = 7                     |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  27:0 | **AUX** - auxiliary data (depends on TYPE)                   |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 |  31:0 |                                                              |
+ *  +---+-------+                                                              |
+ *  |...|       | **PAYLOAD** - optional payload (depends on TYPE)             |
+ *  +---+-------+                                                              |
+ *  | n |  31:0 |                                                              |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_HXG_MSG_MIN_LEN			1u
+#define GUC_HXG_MSG_0_ORIGIN			(0x1U << 31)
+#define   GUC_HXG_ORIGIN_HOST			0u
+#define   GUC_HXG_ORIGIN_GUC			1u
+#define GUC_HXG_MSG_0_TYPE			(0x7 << 28)
+#define   GUC_HXG_TYPE_REQUEST			0u
+#define   GUC_HXG_TYPE_EVENT			1u
+#define   GUC_HXG_TYPE_NO_RESPONSE_BUSY		3u
+#define   GUC_HXG_TYPE_NO_RESPONSE_RETRY	5u
+#define   GUC_HXG_TYPE_RESPONSE_FAILURE		6u
+#define   GUC_HXG_TYPE_RESPONSE_SUCCESS		7u
+#define GUC_HXG_MSG_0_AUX			(0xfffffff << 0)
+#define GUC_HXG_MSG_n_PAYLOAD			(0xffffffff << 0)
+
+/**
+ * DOC: HXG Request
+ *
+ * The `HXG Request`_ message should be used to initiate synchronous activity
+ * for which confirmation or return data is expected.
+ *
+ * The recipient of this message shall use `HXG Response`_, `HXG Failure`_
+ * or `HXG Retry`_ message as a definite reply, and may use `HXG Busy`_
+ * message as a intermediate reply.
+ *
+ * Format of @DATA0 and all @DATAn fields depends on the @ACTION code.
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN                                                       |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_REQUEST_                                 |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 27:16 | **DATA0** - request data (depends on ACTION)                 |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  15:0 | **ACTION** - requested action code                           |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 |  31:0 |                                                              |
+ *  +---+-------+                                                              |
+ *  |...|       | **DATAn** - optional data (depends on ACTION)                |
+ *  +---+-------+                                                              |
+ *  | n |  31:0 |                                                              |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_HXG_REQUEST_MSG_MIN_LEN		GUC_HXG_MSG_MIN_LEN
+#define GUC_HXG_REQUEST_MSG_0_DATA0		(0xfff << 16)
+#define GUC_HXG_REQUEST_MSG_0_ACTION		(0xffff << 0)
+#define GUC_HXG_REQUEST_MSG_n_DATAn		GUC_HXG_MSG_n_PAYLOAD
+
+/**
+ * DOC: HXG Event
+ *
+ * The `HXG Event`_ message should be used to initiate asynchronous activity
+ * that does not involves immediate confirmation nor data.
+ *
+ * Format of @DATA0 and all @DATAn fields depends on the @ACTION code.
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN                                                       |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_EVENT_                                   |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 27:16 | **DATA0** - event data (depends on ACTION)                   |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  15:0 | **ACTION** - event action code                               |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 |  31:0 |                                                              |
+ *  +---+-------+                                                              |
+ *  |...|       | **DATAn** - optional event  data (depends on ACTION)         |
+ *  +---+-------+                                                              |
+ *  | n |  31:0 |                                                              |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_HXG_EVENT_MSG_MIN_LEN		GUC_HXG_MSG_MIN_LEN
+#define GUC_HXG_EVENT_MSG_0_DATA0		(0xfff << 16)
+#define GUC_HXG_EVENT_MSG_0_ACTION		(0xffff << 0)
+#define GUC_HXG_EVENT_MSG_n_DATAn		GUC_HXG_MSG_n_PAYLOAD
+
+/**
+ * DOC: HXG Busy
+ *
+ * The `HXG Busy`_ message may be used to acknowledge reception of the `HXG Request`_
+ * message if the recipient expects that it processing will be longer than default
+ * timeout.
+ *
+ * The @COUNTER field may be used as a progress indicator.
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN                                                       |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_NO_RESPONSE_BUSY_                        |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  27:0 | **COUNTER** - progress indicator                             |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_HXG_BUSY_MSG_LEN			GUC_HXG_MSG_MIN_LEN
+#define GUC_HXG_BUSY_MSG_0_COUNTER		GUC_HXG_MSG_0_AUX
+
+/**
+ * DOC: HXG Retry
+ *
+ * The `HXG Retry`_ message should be used by recipient to indicate that the
+ * `HXG Request`_ message was dropped and it should be resent again.
+ *
+ * The @REASON field may be used to provide additional information.
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN                                                       |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_NO_RESPONSE_RETRY_                       |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  27:0 | **REASON** - reason for retry                                |
+ *  |   |       |  - _`GUC_HXG_RETRY_REASON_UNSPECIFIED` = 0                   |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_HXG_RETRY_MSG_LEN			GUC_HXG_MSG_MIN_LEN
+#define GUC_HXG_RETRY_MSG_0_REASON		GUC_HXG_MSG_0_AUX
+#define   GUC_HXG_RETRY_REASON_UNSPECIFIED	0u
+
+/**
+ * DOC: HXG Failure
+ *
+ * The `HXG Failure`_ message shall be used as a reply to the `HXG Request`_
+ * message that could not be processed due to an error.
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN                                                       |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_RESPONSE_FAILURE_                        |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 27:16 | **HINT** - additional error hint                             |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  15:0 | **ERROR** - error/result code                                |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_HXG_FAILURE_MSG_LEN			GUC_HXG_MSG_MIN_LEN
+#define GUC_HXG_FAILURE_MSG_0_HINT		(0xfff << 16)
+#define GUC_HXG_FAILURE_MSG_0_ERROR		(0xffff << 0)
+
+/**
+ * DOC: HXG Response
+ *
+ * The `HXG Response`_ message shall be used as a reply to the `HXG Request`_
+ * message that was successfully processed without an error.
+ *
+ *  +---+-------+--------------------------------------------------------------+
+ *  |   | Bits  | Description                                                  |
+ *  +===+=======+==============================================================+
+ *  | 0 |    31 | ORIGIN                                                       |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   | 30:28 | TYPE = GUC_HXG_TYPE_RESPONSE_SUCCESS_                        |
+ *  |   +-------+--------------------------------------------------------------+
+ *  |   |  27:0 | **DATA0** - data (depends on ACTION from `HXG Request`_)     |
+ *  +---+-------+--------------------------------------------------------------+
+ *  | 1 |  31:0 |                                                              |
+ *  +---+-------+                                                              |
+ *  |...|       | **DATAn** - data (depends on ACTION from `HXG Request`_)     |
+ *  +---+-------+                                                              |
+ *  | n |  31:0 |                                                              |
+ *  +---+-------+--------------------------------------------------------------+
+ */
+
+#define GUC_HXG_RESPONSE_MSG_MIN_LEN		GUC_HXG_MSG_MIN_LEN
+#define GUC_HXG_RESPONSE_MSG_0_DATA0		GUC_HXG_MSG_0_AUX
+#define GUC_HXG_RESPONSE_MSG_n_DATAn		GUC_HXG_MSG_n_PAYLOAD
+
+/* deprecated */
+#define INTEL_GUC_MSG_TYPE_SHIFT	28
+#define INTEL_GUC_MSG_TYPE_MASK		(0xF << INTEL_GUC_MSG_TYPE_SHIFT)
+#define INTEL_GUC_MSG_DATA_SHIFT	16
+#define INTEL_GUC_MSG_DATA_MASK		(0xFFF << INTEL_GUC_MSG_DATA_SHIFT)
+#define INTEL_GUC_MSG_CODE_SHIFT	0
+#define INTEL_GUC_MSG_CODE_MASK		(0xFFFF << INTEL_GUC_MSG_CODE_SHIFT)
+
+enum intel_guc_msg_type {
+	INTEL_GUC_MSG_TYPE_REQUEST = 0x0,
+	INTEL_GUC_MSG_TYPE_RESPONSE = 0xF,
+};
+
+#endif /* _ABI_GUC_MESSAGES_ABI_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/guc_capture_fwif.h b/drivers/gpu/drm/i915/gt/uc/guc_capture_fwif.h
new file mode 100644
index 000000000..3624abfd2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/guc_capture_fwif.h
@@ -0,0 +1,218 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021-2022 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_CAPTURE_FWIF_H
+#define _INTEL_GUC_CAPTURE_FWIF_H
+
+#include <linux/types.h>
+#include "intel_guc_fwif.h"
+
+struct intel_guc;
+struct file;
+
+/**
+ * struct __guc_capture_bufstate
+ *
+ * Book-keeping structure used to track read and write pointers
+ * as we extract error capture data from the GuC-log-buffer's
+ * error-capture region as a stream of dwords.
+ */
+struct __guc_capture_bufstate {
+	u32 size;
+	void *data;
+	u32 rd;
+	u32 wr;
+};
+
+/**
+ * struct __guc_capture_parsed_output - extracted error capture node
+ *
+ * A single unit of extracted error-capture output data grouped together
+ * at an engine-instance level. We keep these nodes in a linked list.
+ * See cachelist and outlist below.
+ */
+struct __guc_capture_parsed_output {
+	/*
+	 * A single set of 3 capture lists: a global-list
+	 * an engine-class-list and an engine-instance list.
+	 * outlist in __guc_capture_parsed_output will keep
+	 * a linked list of these nodes that will eventually
+	 * be detached from outlist and attached into to
+	 * i915_gpu_codedump in response to a context reset
+	 */
+	struct list_head link;
+	bool is_partial;
+	u32 eng_class;
+	u32 eng_inst;
+	u32 guc_id;
+	u32 lrca;
+	struct gcap_reg_list_info {
+		u32 vfid;
+		u32 num_regs;
+		struct guc_mmio_reg *regs;
+	} reginfo[GUC_CAPTURE_LIST_TYPE_MAX];
+#define GCAP_PARSED_REGLIST_INDEX_GLOBAL   BIT(GUC_CAPTURE_LIST_TYPE_GLOBAL)
+#define GCAP_PARSED_REGLIST_INDEX_ENGCLASS BIT(GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS)
+#define GCAP_PARSED_REGLIST_INDEX_ENGINST  BIT(GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE)
+};
+
+/**
+ * struct guc_debug_capture_list_header / struct guc_debug_capture_list
+ *
+ * As part of ADS registration, these header structures (followed by
+ * an array of 'struct guc_mmio_reg' entries) are used to register with
+ * GuC microkernel the list of registers we want it to dump out prior
+ * to a engine reset.
+ */
+struct guc_debug_capture_list_header {
+	u32 info;
+#define GUC_CAPTURELISTHDR_NUMDESCR GENMASK(15, 0)
+} __packed;
+
+struct guc_debug_capture_list {
+	struct guc_debug_capture_list_header header;
+	struct guc_mmio_reg regs[0];
+} __packed;
+
+/**
+ * struct __guc_mmio_reg_descr / struct __guc_mmio_reg_descr_group
+ *
+ * intel_guc_capture module uses these structures to maintain static
+ * tables (per unique platform) that consists of lists of registers
+ * (offsets, names, flags,...) that are used at the ADS regisration
+ * time as well as during runtime processing and reporting of error-
+ * capture states generated by GuC just prior to engine reset events.
+ */
+struct __guc_mmio_reg_descr {
+	i915_reg_t reg;
+	u32 flags;
+	u32 mask;
+	const char *regname;
+};
+
+struct __guc_mmio_reg_descr_group {
+	const struct __guc_mmio_reg_descr *list;
+	u32 num_regs;
+	u32 owner; /* see enum guc_capture_owner */
+	u32 type; /* see enum guc_capture_type */
+	u32 engine; /* as per MAX_ENGINE_CLASS */
+	struct __guc_mmio_reg_descr *extlist; /* only used for steered registers */
+};
+
+/**
+ * struct guc_state_capture_header_t / struct guc_state_capture_t /
+ * guc_state_capture_group_header_t / guc_state_capture_group_t
+ *
+ * Prior to resetting engines that have hung or faulted, GuC microkernel
+ * reports the engine error-state (register values that was read) by
+ * logging them into the shared GuC log buffer using these hierarchy
+ * of structures.
+ */
+struct guc_state_capture_header_t {
+	u32 owner;
+#define CAP_HDR_CAPTURE_VFID GENMASK(7, 0)
+	u32 info;
+#define CAP_HDR_CAPTURE_TYPE GENMASK(3, 0) /* see enum guc_capture_type */
+#define CAP_HDR_ENGINE_CLASS GENMASK(7, 4) /* see GUC_MAX_ENGINE_CLASSES */
+#define CAP_HDR_ENGINE_INSTANCE GENMASK(11, 8)
+	u32 lrca; /* if type-instance, LRCA (address) that hung, else set to ~0 */
+	u32 guc_id; /* if type-instance, context index of hung context, else set to ~0 */
+	u32 num_mmios;
+#define CAP_HDR_NUM_MMIOS GENMASK(9, 0)
+} __packed;
+
+struct guc_state_capture_t {
+	struct guc_state_capture_header_t header;
+	struct guc_mmio_reg mmio_entries[0];
+} __packed;
+
+enum guc_capture_group_types {
+	GUC_STATE_CAPTURE_GROUP_TYPE_FULL,
+	GUC_STATE_CAPTURE_GROUP_TYPE_PARTIAL,
+	GUC_STATE_CAPTURE_GROUP_TYPE_MAX,
+};
+
+struct guc_state_capture_group_header_t {
+	u32 owner;
+#define CAP_GRP_HDR_CAPTURE_VFID GENMASK(7, 0)
+	u32 info;
+#define CAP_GRP_HDR_NUM_CAPTURES GENMASK(7, 0)
+#define CAP_GRP_HDR_CAPTURE_TYPE GENMASK(15, 8) /* guc_capture_group_types */
+} __packed;
+
+/* this is the top level structure where an error-capture dump starts */
+struct guc_state_capture_group_t {
+	struct guc_state_capture_group_header_t grp_header;
+	struct guc_state_capture_t capture_entries[0];
+} __packed;
+
+/**
+ * struct __guc_capture_ads_cache
+ *
+ * A structure to cache register lists that were populated and registered
+ * with GuC at startup during ADS registration. This allows much quicker
+ * GuC resets without re-parsing all the tables for the given gt.
+ */
+struct __guc_capture_ads_cache {
+	bool is_valid;
+	void *ptr;
+	size_t size;
+	int status;
+};
+
+/**
+ * struct intel_guc_state_capture
+ *
+ * Internal context of the intel_guc_capture module.
+ */
+struct intel_guc_state_capture {
+	/**
+	 * @reglists: static table of register lists used for error-capture state.
+	 */
+	const struct __guc_mmio_reg_descr_group *reglists;
+
+	/**
+	 * @extlists: allocated table of steered register lists used for error-capture state.
+	 *
+	 * NOTE: steered registers have multiple instances depending on the HW configuration
+	 * (slices or dual-sub-slices) and thus depends on HW fuses discovered at startup
+	 */
+	struct __guc_mmio_reg_descr_group *extlists;
+
+	/**
+	 * @ads_cache: cached register lists that is ADS format ready
+	 */
+	struct __guc_capture_ads_cache ads_cache[GUC_CAPTURE_LIST_INDEX_MAX]
+						[GUC_CAPTURE_LIST_TYPE_MAX]
+						[GUC_MAX_ENGINE_CLASSES];
+	void *ads_null_cache;
+
+	/**
+	 * @cachelist: Pool of pre-allocated nodes for error capture output
+	 *
+	 * We need this pool of pre-allocated nodes because we cannot
+	 * dynamically allocate new nodes when receiving the G2H notification
+	 * because the event handlers for all G2H event-processing is called
+	 * by the ct processing worker queue and when that queue is being
+	 * processed, there is no absoluate guarantee that we are not in the
+	 * midst of a GT reset operation (which doesn't allow allocations).
+	 */
+	struct list_head cachelist;
+#define PREALLOC_NODES_MAX_COUNT (3 * GUC_MAX_ENGINE_CLASSES * GUC_MAX_INSTANCES_PER_CLASS)
+#define PREALLOC_NODES_DEFAULT_NUMREGS 64
+	int max_mmio_per_node;
+
+	/**
+	 * @outlist: Pool of pre-allocated nodes for error capture output
+	 *
+	 * A linked list of parsed GuC error-capture output data before
+	 * reporting with formatting via i915_gpu_coredump. Each node in this linked list shall
+	 * contain a single engine-capture including global, engine-class and
+	 * engine-instance register dumps as per guc_capture_parsed_output_node
+	 */
+	struct list_head outlist;
+};
+
+#endif /* _INTEL_GUC_CAPTURE_FWIF_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc.c b/drivers/gpu/drm/i915/gt/uc/intel_guc.c
new file mode 100644
index 000000000..bac06e3d6
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc.c
@@ -0,0 +1,915 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#include "gem/i915_gem_lmem.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_irq.h"
+#include "gt/intel_gt_pm_irq.h"
+#include "gt/intel_gt_regs.h"
+#include "intel_guc.h"
+#include "intel_guc_ads.h"
+#include "intel_guc_capture.h"
+#include "intel_guc_slpc.h"
+#include "intel_guc_submission.h"
+#include "i915_drv.h"
+#include "i915_irq.h"
+
+/**
+ * DOC: GuC
+ *
+ * The GuC is a microcontroller inside the GT HW, introduced in gen9. The GuC is
+ * designed to offload some of the functionality usually performed by the host
+ * driver; currently the main operations it can take care of are:
+ *
+ * - Authentication of the HuC, which is required to fully enable HuC usage.
+ * - Low latency graphics context scheduling (a.k.a. GuC submission).
+ * - GT Power management.
+ *
+ * The enable_guc module parameter can be used to select which of those
+ * operations to enable within GuC. Note that not all the operations are
+ * supported on all gen9+ platforms.
+ *
+ * Enabling the GuC is not mandatory and therefore the firmware is only loaded
+ * if at least one of the operations is selected. However, not loading the GuC
+ * might result in the loss of some features that do require the GuC (currently
+ * just the HuC, but more are expected to land in the future).
+ */
+
+void intel_guc_notify(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+
+	/*
+	 * On Gen11+, the value written to the register is passes as a payload
+	 * to the FW. However, the FW currently treats all values the same way
+	 * (H2G interrupt), so we can just write the value that the HW expects
+	 * on older gens.
+	 */
+	intel_uncore_write(gt->uncore, guc->notify_reg, GUC_SEND_TRIGGER);
+}
+
+static inline i915_reg_t guc_send_reg(struct intel_guc *guc, u32 i)
+{
+	GEM_BUG_ON(!guc->send_regs.base);
+	GEM_BUG_ON(!guc->send_regs.count);
+	GEM_BUG_ON(i >= guc->send_regs.count);
+
+	return _MMIO(guc->send_regs.base + 4 * i);
+}
+
+void intel_guc_init_send_regs(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	enum forcewake_domains fw_domains = 0;
+	unsigned int i;
+
+	GEM_BUG_ON(!guc->send_regs.base);
+	GEM_BUG_ON(!guc->send_regs.count);
+
+	for (i = 0; i < guc->send_regs.count; i++) {
+		fw_domains |= intel_uncore_forcewake_for_reg(gt->uncore,
+					guc_send_reg(guc, i),
+					FW_REG_READ | FW_REG_WRITE);
+	}
+	guc->send_regs.fw_domains = fw_domains;
+}
+
+static void gen9_reset_guc_interrupts(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+
+	assert_rpm_wakelock_held(&gt->i915->runtime_pm);
+
+	spin_lock_irq(gt->irq_lock);
+	gen6_gt_pm_reset_iir(gt, gt->pm_guc_events);
+	spin_unlock_irq(gt->irq_lock);
+}
+
+static void gen9_enable_guc_interrupts(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+
+	assert_rpm_wakelock_held(&gt->i915->runtime_pm);
+
+	spin_lock_irq(gt->irq_lock);
+	WARN_ON_ONCE(intel_uncore_read(gt->uncore, GEN8_GT_IIR(2)) &
+		     gt->pm_guc_events);
+	gen6_gt_pm_enable_irq(gt, gt->pm_guc_events);
+	spin_unlock_irq(gt->irq_lock);
+}
+
+static void gen9_disable_guc_interrupts(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+
+	assert_rpm_wakelock_held(&gt->i915->runtime_pm);
+
+	spin_lock_irq(gt->irq_lock);
+
+	gen6_gt_pm_disable_irq(gt, gt->pm_guc_events);
+
+	spin_unlock_irq(gt->irq_lock);
+	intel_synchronize_irq(gt->i915);
+
+	gen9_reset_guc_interrupts(guc);
+}
+
+static void gen11_reset_guc_interrupts(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+
+	spin_lock_irq(gt->irq_lock);
+	gen11_gt_reset_one_iir(gt, 0, GEN11_GUC);
+	spin_unlock_irq(gt->irq_lock);
+}
+
+static void gen11_enable_guc_interrupts(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	u32 events = REG_FIELD_PREP(ENGINE1_MASK, GUC_INTR_GUC2HOST);
+
+	spin_lock_irq(gt->irq_lock);
+	WARN_ON_ONCE(gen11_gt_reset_one_iir(gt, 0, GEN11_GUC));
+	intel_uncore_write(gt->uncore,
+			   GEN11_GUC_SG_INTR_ENABLE, events);
+	intel_uncore_write(gt->uncore,
+			   GEN11_GUC_SG_INTR_MASK, ~events);
+	spin_unlock_irq(gt->irq_lock);
+}
+
+static void gen11_disable_guc_interrupts(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+
+	spin_lock_irq(gt->irq_lock);
+
+	intel_uncore_write(gt->uncore, GEN11_GUC_SG_INTR_MASK, ~0);
+	intel_uncore_write(gt->uncore, GEN11_GUC_SG_INTR_ENABLE, 0);
+
+	spin_unlock_irq(gt->irq_lock);
+	intel_synchronize_irq(gt->i915);
+
+	gen11_reset_guc_interrupts(guc);
+}
+
+void intel_guc_init_early(struct intel_guc *guc)
+{
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+
+	intel_uc_fw_init_early(&guc->fw, INTEL_UC_FW_TYPE_GUC);
+	intel_guc_ct_init_early(&guc->ct);
+	intel_guc_log_init_early(&guc->log);
+	intel_guc_submission_init_early(guc);
+	intel_guc_slpc_init_early(&guc->slpc);
+	intel_guc_rc_init_early(guc);
+
+	mutex_init(&guc->send_mutex);
+	spin_lock_init(&guc->irq_lock);
+	if (GRAPHICS_VER(i915) >= 11) {
+		guc->notify_reg = GEN11_GUC_HOST_INTERRUPT;
+		guc->interrupts.reset = gen11_reset_guc_interrupts;
+		guc->interrupts.enable = gen11_enable_guc_interrupts;
+		guc->interrupts.disable = gen11_disable_guc_interrupts;
+		guc->send_regs.base =
+			i915_mmio_reg_offset(GEN11_SOFT_SCRATCH(0));
+		guc->send_regs.count = GEN11_SOFT_SCRATCH_COUNT;
+
+	} else {
+		guc->notify_reg = GUC_SEND_INTERRUPT;
+		guc->interrupts.reset = gen9_reset_guc_interrupts;
+		guc->interrupts.enable = gen9_enable_guc_interrupts;
+		guc->interrupts.disable = gen9_disable_guc_interrupts;
+		guc->send_regs.base = i915_mmio_reg_offset(SOFT_SCRATCH(0));
+		guc->send_regs.count = GUC_MAX_MMIO_MSG_LEN;
+		BUILD_BUG_ON(GUC_MAX_MMIO_MSG_LEN > SOFT_SCRATCH_COUNT);
+	}
+
+	intel_guc_enable_msg(guc, INTEL_GUC_RECV_MSG_EXCEPTION |
+				  INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED);
+}
+
+void intel_guc_init_late(struct intel_guc *guc)
+{
+	intel_guc_ads_init_late(guc);
+}
+
+static u32 guc_ctl_debug_flags(struct intel_guc *guc)
+{
+	u32 level = intel_guc_log_get_level(&guc->log);
+	u32 flags = 0;
+
+	if (!GUC_LOG_LEVEL_IS_VERBOSE(level))
+		flags |= GUC_LOG_DISABLED;
+	else
+		flags |= GUC_LOG_LEVEL_TO_VERBOSITY(level) <<
+			 GUC_LOG_VERBOSITY_SHIFT;
+
+	return flags;
+}
+
+static u32 guc_ctl_feature_flags(struct intel_guc *guc)
+{
+	u32 flags = 0;
+
+	if (!intel_guc_submission_is_used(guc))
+		flags |= GUC_CTL_DISABLE_SCHEDULER;
+
+	if (intel_guc_slpc_is_used(guc))
+		flags |= GUC_CTL_ENABLE_SLPC;
+
+	return flags;
+}
+
+static u32 guc_ctl_log_params_flags(struct intel_guc *guc)
+{
+	struct intel_guc_log *log = &guc->log;
+	u32 offset, flags;
+
+	GEM_BUG_ON(!log->sizes_initialised);
+
+	offset = intel_guc_ggtt_offset(guc, log->vma) >> PAGE_SHIFT;
+
+	flags = GUC_LOG_VALID |
+		GUC_LOG_NOTIFY_ON_HALF_FULL |
+		log->sizes[GUC_LOG_SECTIONS_DEBUG].flag |
+		log->sizes[GUC_LOG_SECTIONS_CAPTURE].flag |
+		(log->sizes[GUC_LOG_SECTIONS_CRASH].count << GUC_LOG_CRASH_SHIFT) |
+		(log->sizes[GUC_LOG_SECTIONS_DEBUG].count << GUC_LOG_DEBUG_SHIFT) |
+		(log->sizes[GUC_LOG_SECTIONS_CAPTURE].count << GUC_LOG_CAPTURE_SHIFT) |
+		(offset << GUC_LOG_BUF_ADDR_SHIFT);
+
+	return flags;
+}
+
+static u32 guc_ctl_ads_flags(struct intel_guc *guc)
+{
+	u32 ads = intel_guc_ggtt_offset(guc, guc->ads_vma) >> PAGE_SHIFT;
+	u32 flags = ads << GUC_ADS_ADDR_SHIFT;
+
+	return flags;
+}
+
+static u32 guc_ctl_wa_flags(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	u32 flags = 0;
+
+	/* Wa_22012773006:gen11,gen12 < XeHP */
+	if (GRAPHICS_VER(gt->i915) >= 11 &&
+	    GRAPHICS_VER_FULL(gt->i915) < IP_VER(12, 50))
+		flags |= GUC_WA_POLLCS;
+
+	/* Wa_16011759253:dg2_g10:a0 */
+	if (IS_DG2_GRAPHICS_STEP(gt->i915, G10, STEP_A0, STEP_B0))
+		flags |= GUC_WA_GAM_CREDITS;
+
+	/* Wa_14014475959:dg2 */
+	if (IS_DG2(gt->i915))
+		flags |= GUC_WA_HOLD_CCS_SWITCHOUT;
+
+	/*
+	 * Wa_14012197797:dg2_g10:a0,dg2_g11:a0
+	 * Wa_22011391025:dg2_g10,dg2_g11,dg2_g12
+	 *
+	 * The same WA bit is used for both and 22011391025 is applicable to
+	 * all DG2.
+	 */
+	if (IS_DG2(gt->i915))
+		flags |= GUC_WA_DUAL_QUEUE;
+
+	/* Wa_22011802037: graphics version 11/12 */
+	if (IS_GRAPHICS_VER(gt->i915, 11, 12))
+		flags |= GUC_WA_PRE_PARSER;
+
+	/* Wa_16011777198:dg2 */
+	if (IS_DG2_GRAPHICS_STEP(gt->i915, G10, STEP_A0, STEP_C0) ||
+	    IS_DG2_GRAPHICS_STEP(gt->i915, G11, STEP_A0, STEP_B0))
+		flags |= GUC_WA_RCS_RESET_BEFORE_RC6;
+
+	/*
+	 * Wa_22012727170:dg2_g10[a0-c0), dg2_g11[a0..)
+	 * Wa_22012727685:dg2_g11[a0..)
+	 */
+	if (IS_DG2_GRAPHICS_STEP(gt->i915, G10, STEP_A0, STEP_C0) ||
+	    IS_DG2_GRAPHICS_STEP(gt->i915, G11, STEP_A0, STEP_FOREVER))
+		flags |= GUC_WA_CONTEXT_ISOLATION;
+
+	/* Wa_16015675438 */
+	if (!RCS_MASK(gt))
+		flags |= GUC_WA_RCS_REGS_IN_CCS_REGS_LIST;
+
+	return flags;
+}
+
+static u32 guc_ctl_devid(struct intel_guc *guc)
+{
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+
+	return (INTEL_DEVID(i915) << 16) | INTEL_REVID(i915);
+}
+
+/*
+ * Initialise the GuC parameter block before starting the firmware
+ * transfer. These parameters are read by the firmware on startup
+ * and cannot be changed thereafter.
+ */
+static void guc_init_params(struct intel_guc *guc)
+{
+	u32 *params = guc->params;
+	int i;
+
+	BUILD_BUG_ON(sizeof(guc->params) != GUC_CTL_MAX_DWORDS * sizeof(u32));
+
+	params[GUC_CTL_LOG_PARAMS] = guc_ctl_log_params_flags(guc);
+	params[GUC_CTL_FEATURE] = guc_ctl_feature_flags(guc);
+	params[GUC_CTL_DEBUG] = guc_ctl_debug_flags(guc);
+	params[GUC_CTL_ADS] = guc_ctl_ads_flags(guc);
+	params[GUC_CTL_WA] = guc_ctl_wa_flags(guc);
+	params[GUC_CTL_DEVID] = guc_ctl_devid(guc);
+
+	for (i = 0; i < GUC_CTL_MAX_DWORDS; i++)
+		DRM_DEBUG_DRIVER("param[%2d] = %#x\n", i, params[i]);
+}
+
+/*
+ * Initialise the GuC parameter block before starting the firmware
+ * transfer. These parameters are read by the firmware on startup
+ * and cannot be changed thereafter.
+ */
+void intel_guc_write_params(struct intel_guc *guc)
+{
+	struct intel_uncore *uncore = guc_to_gt(guc)->uncore;
+	int i;
+
+	/*
+	 * All SOFT_SCRATCH registers are in FORCEWAKE_GT domain and
+	 * they are power context saved so it's ok to release forcewake
+	 * when we are done here and take it again at xfer time.
+	 */
+	intel_uncore_forcewake_get(uncore, FORCEWAKE_GT);
+
+	intel_uncore_write(uncore, SOFT_SCRATCH(0), 0);
+
+	for (i = 0; i < GUC_CTL_MAX_DWORDS; i++)
+		intel_uncore_write(uncore, SOFT_SCRATCH(1 + i), guc->params[i]);
+
+	intel_uncore_forcewake_put(uncore, FORCEWAKE_GT);
+}
+
+void intel_guc_dump_time_info(struct intel_guc *guc, struct drm_printer *p)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	intel_wakeref_t wakeref;
+	u32 stamp = 0;
+	u64 ktime;
+
+	with_intel_runtime_pm(&gt->i915->runtime_pm, wakeref)
+		stamp = intel_uncore_read(gt->uncore, GUCPMTIMESTAMP);
+	ktime = ktime_get_boottime_ns();
+
+	drm_printf(p, "Kernel timestamp: 0x%08llX [%llu]\n", ktime, ktime);
+	drm_printf(p, "GuC timestamp: 0x%08X [%u]\n", stamp, stamp);
+	drm_printf(p, "CS timestamp frequency: %u Hz, %u ns\n",
+		   gt->clock_frequency, gt->clock_period_ns);
+}
+
+int intel_guc_init(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	int ret;
+
+	ret = intel_uc_fw_init(&guc->fw);
+	if (ret)
+		goto out;
+
+	ret = intel_guc_log_create(&guc->log);
+	if (ret)
+		goto err_fw;
+
+	ret = intel_guc_capture_init(guc);
+	if (ret)
+		goto err_log;
+
+	ret = intel_guc_ads_create(guc);
+	if (ret)
+		goto err_capture;
+
+	GEM_BUG_ON(!guc->ads_vma);
+
+	ret = intel_guc_ct_init(&guc->ct);
+	if (ret)
+		goto err_ads;
+
+	if (intel_guc_submission_is_used(guc)) {
+		/*
+		 * This is stuff we need to have available at fw load time
+		 * if we are planning to enable submission later
+		 */
+		ret = intel_guc_submission_init(guc);
+		if (ret)
+			goto err_ct;
+	}
+
+	if (intel_guc_slpc_is_used(guc)) {
+		ret = intel_guc_slpc_init(&guc->slpc);
+		if (ret)
+			goto err_submission;
+	}
+
+	/* now that everything is perma-pinned, initialize the parameters */
+	guc_init_params(guc);
+
+	/* We need to notify the guc whenever we change the GGTT */
+	i915_ggtt_enable_guc(gt->ggtt);
+
+	intel_uc_fw_change_status(&guc->fw, INTEL_UC_FIRMWARE_LOADABLE);
+
+	return 0;
+
+err_submission:
+	intel_guc_submission_fini(guc);
+err_ct:
+	intel_guc_ct_fini(&guc->ct);
+err_ads:
+	intel_guc_ads_destroy(guc);
+err_capture:
+	intel_guc_capture_destroy(guc);
+err_log:
+	intel_guc_log_destroy(&guc->log);
+err_fw:
+	intel_uc_fw_fini(&guc->fw);
+out:
+	i915_probe_error(gt->i915, "failed with %d\n", ret);
+	return ret;
+}
+
+void intel_guc_fini(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+
+	if (!intel_uc_fw_is_loadable(&guc->fw))
+		return;
+
+	i915_ggtt_disable_guc(gt->ggtt);
+
+	if (intel_guc_slpc_is_used(guc))
+		intel_guc_slpc_fini(&guc->slpc);
+
+	if (intel_guc_submission_is_used(guc))
+		intel_guc_submission_fini(guc);
+
+	intel_guc_ct_fini(&guc->ct);
+
+	intel_guc_ads_destroy(guc);
+	intel_guc_capture_destroy(guc);
+	intel_guc_log_destroy(&guc->log);
+	intel_uc_fw_fini(&guc->fw);
+}
+
+/*
+ * This function implements the MMIO based host to GuC interface.
+ */
+int intel_guc_send_mmio(struct intel_guc *guc, const u32 *request, u32 len,
+			u32 *response_buf, u32 response_buf_size)
+{
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	struct intel_uncore *uncore = guc_to_gt(guc)->uncore;
+	u32 header;
+	int i;
+	int ret;
+
+	GEM_BUG_ON(!len);
+	GEM_BUG_ON(len > guc->send_regs.count);
+
+	GEM_BUG_ON(FIELD_GET(GUC_HXG_MSG_0_ORIGIN, request[0]) != GUC_HXG_ORIGIN_HOST);
+	GEM_BUG_ON(FIELD_GET(GUC_HXG_MSG_0_TYPE, request[0]) != GUC_HXG_TYPE_REQUEST);
+
+	mutex_lock(&guc->send_mutex);
+	intel_uncore_forcewake_get(uncore, guc->send_regs.fw_domains);
+
+retry:
+	for (i = 0; i < len; i++)
+		intel_uncore_write(uncore, guc_send_reg(guc, i), request[i]);
+
+	intel_uncore_posting_read(uncore, guc_send_reg(guc, i - 1));
+
+	intel_guc_notify(guc);
+
+	/*
+	 * No GuC command should ever take longer than 10ms.
+	 * Fast commands should still complete in 10us.
+	 */
+	ret = __intel_wait_for_register_fw(uncore,
+					   guc_send_reg(guc, 0),
+					   GUC_HXG_MSG_0_ORIGIN,
+					   FIELD_PREP(GUC_HXG_MSG_0_ORIGIN,
+						      GUC_HXG_ORIGIN_GUC),
+					   10, 10, &header);
+	if (unlikely(ret)) {
+timeout:
+		drm_err(&i915->drm, "mmio request %#x: no reply %x\n",
+			request[0], header);
+		goto out;
+	}
+
+	if (FIELD_GET(GUC_HXG_MSG_0_TYPE, header) == GUC_HXG_TYPE_NO_RESPONSE_BUSY) {
+#define done ({ header = intel_uncore_read(uncore, guc_send_reg(guc, 0)); \
+		FIELD_GET(GUC_HXG_MSG_0_ORIGIN, header) != GUC_HXG_ORIGIN_GUC || \
+		FIELD_GET(GUC_HXG_MSG_0_TYPE, header) != GUC_HXG_TYPE_NO_RESPONSE_BUSY; })
+
+		ret = wait_for(done, 1000);
+		if (unlikely(ret))
+			goto timeout;
+		if (unlikely(FIELD_GET(GUC_HXG_MSG_0_ORIGIN, header) !=
+				       GUC_HXG_ORIGIN_GUC))
+			goto proto;
+#undef done
+	}
+
+	if (FIELD_GET(GUC_HXG_MSG_0_TYPE, header) == GUC_HXG_TYPE_NO_RESPONSE_RETRY) {
+		u32 reason = FIELD_GET(GUC_HXG_RETRY_MSG_0_REASON, header);
+
+		drm_dbg(&i915->drm, "mmio request %#x: retrying, reason %u\n",
+			request[0], reason);
+		goto retry;
+	}
+
+	if (FIELD_GET(GUC_HXG_MSG_0_TYPE, header) == GUC_HXG_TYPE_RESPONSE_FAILURE) {
+		u32 hint = FIELD_GET(GUC_HXG_FAILURE_MSG_0_HINT, header);
+		u32 error = FIELD_GET(GUC_HXG_FAILURE_MSG_0_ERROR, header);
+
+		drm_err(&i915->drm, "mmio request %#x: failure %x/%u\n",
+			request[0], error, hint);
+		ret = -ENXIO;
+		goto out;
+	}
+
+	if (FIELD_GET(GUC_HXG_MSG_0_TYPE, header) != GUC_HXG_TYPE_RESPONSE_SUCCESS) {
+proto:
+		drm_err(&i915->drm, "mmio request %#x: unexpected reply %#x\n",
+			request[0], header);
+		ret = -EPROTO;
+		goto out;
+	}
+
+	if (response_buf) {
+		int count = min(response_buf_size, guc->send_regs.count);
+
+		GEM_BUG_ON(!count);
+
+		response_buf[0] = header;
+
+		for (i = 1; i < count; i++)
+			response_buf[i] = intel_uncore_read(uncore,
+							    guc_send_reg(guc, i));
+
+		/* Use number of copied dwords as our return value */
+		ret = count;
+	} else {
+		/* Use data from the GuC response as our return value */
+		ret = FIELD_GET(GUC_HXG_RESPONSE_MSG_0_DATA0, header);
+	}
+
+out:
+	intel_uncore_forcewake_put(uncore, guc->send_regs.fw_domains);
+	mutex_unlock(&guc->send_mutex);
+
+	return ret;
+}
+
+int intel_guc_to_host_process_recv_msg(struct intel_guc *guc,
+				       const u32 *payload, u32 len)
+{
+	u32 msg;
+
+	if (unlikely(!len))
+		return -EPROTO;
+
+	/* Make sure to handle only enabled messages */
+	msg = payload[0] & guc->msg_enabled_mask;
+
+	if (msg & INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED)
+		drm_err(&guc_to_gt(guc)->i915->drm, "Received early GuC crash dump notification!\n");
+	if (msg & INTEL_GUC_RECV_MSG_EXCEPTION)
+		drm_err(&guc_to_gt(guc)->i915->drm, "Received early GuC exception notification!\n");
+
+	return 0;
+}
+
+/**
+ * intel_guc_auth_huc() - Send action to GuC to authenticate HuC ucode
+ * @guc: intel_guc structure
+ * @rsa_offset: rsa offset w.r.t ggtt base of huc vma
+ *
+ * Triggers a HuC firmware authentication request to the GuC via intel_guc_send
+ * INTEL_GUC_ACTION_AUTHENTICATE_HUC interface. This function is invoked by
+ * intel_huc_auth().
+ *
+ * Return:	non-zero code on error
+ */
+int intel_guc_auth_huc(struct intel_guc *guc, u32 rsa_offset)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_AUTHENTICATE_HUC,
+		rsa_offset
+	};
+
+	return intel_guc_send(guc, action, ARRAY_SIZE(action));
+}
+
+/**
+ * intel_guc_suspend() - notify GuC entering suspend state
+ * @guc:	the guc
+ */
+int intel_guc_suspend(struct intel_guc *guc)
+{
+	int ret;
+	u32 action[] = {
+		INTEL_GUC_ACTION_CLIENT_SOFT_RESET,
+	};
+
+	if (!intel_guc_is_ready(guc))
+		return 0;
+
+	if (intel_guc_submission_is_used(guc)) {
+		/*
+		 * This H2G MMIO command tears down the GuC in two steps. First it will
+		 * generate a G2H CTB for every active context indicating a reset. In
+		 * practice the i915 shouldn't ever get a G2H as suspend should only be
+		 * called when the GPU is idle. Next, it tears down the CTBs and this
+		 * H2G MMIO command completes.
+		 *
+		 * Don't abort on a failure code from the GuC. Keep going and do the
+		 * clean up in santize() and re-initialisation on resume and hopefully
+		 * the error here won't be problematic.
+		 */
+		ret = intel_guc_send_mmio(guc, action, ARRAY_SIZE(action), NULL, 0);
+		if (ret)
+			DRM_ERROR("GuC suspend: RESET_CLIENT action failed with error %d!\n", ret);
+	}
+
+	/* Signal that the GuC isn't running. */
+	intel_guc_sanitize(guc);
+
+	return 0;
+}
+
+/**
+ * intel_guc_resume() - notify GuC resuming from suspend state
+ * @guc:	the guc
+ */
+int intel_guc_resume(struct intel_guc *guc)
+{
+	/*
+	 * NB: This function can still be called even if GuC submission is
+	 * disabled, e.g. if GuC is enabled for HuC authentication only. Thus,
+	 * if any code is later added here, it must be support doing nothing
+	 * if submission is disabled (as per intel_guc_suspend).
+	 */
+	return 0;
+}
+
+/**
+ * DOC: GuC Memory Management
+ *
+ * GuC can't allocate any memory for its own usage, so all the allocations must
+ * be handled by the host driver. GuC accesses the memory via the GGTT, with the
+ * exception of the top and bottom parts of the 4GB address space, which are
+ * instead re-mapped by the GuC HW to memory location of the FW itself (WOPCM)
+ * or other parts of the HW. The driver must take care not to place objects that
+ * the GuC is going to access in these reserved ranges. The layout of the GuC
+ * address space is shown below:
+ *
+ * ::
+ *
+ *     +===========> +====================+ <== FFFF_FFFF
+ *     ^             |      Reserved      |
+ *     |             +====================+ <== GUC_GGTT_TOP
+ *     |             |                    |
+ *     |             |        DRAM        |
+ *    GuC            |                    |
+ *  Address    +===> +====================+ <== GuC ggtt_pin_bias
+ *   Space     ^     |                    |
+ *     |       |     |                    |
+ *     |      GuC    |        GuC         |
+ *     |     WOPCM   |       WOPCM        |
+ *     |      Size   |                    |
+ *     |       |     |                    |
+ *     v       v     |                    |
+ *     +=======+===> +====================+ <== 0000_0000
+ *
+ * The lower part of GuC Address Space [0, ggtt_pin_bias) is mapped to GuC WOPCM
+ * while upper part of GuC Address Space [ggtt_pin_bias, GUC_GGTT_TOP) is mapped
+ * to DRAM. The value of the GuC ggtt_pin_bias is the GuC WOPCM size.
+ */
+
+/**
+ * intel_guc_allocate_vma() - Allocate a GGTT VMA for GuC usage
+ * @guc:	the guc
+ * @size:	size of area to allocate (both virtual space and memory)
+ *
+ * This is a wrapper to create an object for use with the GuC. In order to
+ * use it inside the GuC, an object needs to be pinned lifetime, so we allocate
+ * both some backing storage and a range inside the Global GTT. We must pin
+ * it in the GGTT somewhere other than than [0, GUC ggtt_pin_bias) because that
+ * range is reserved inside GuC.
+ *
+ * Return:	A i915_vma if successful, otherwise an ERR_PTR.
+ */
+struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u64 flags;
+	int ret;
+
+	if (HAS_LMEM(gt->i915))
+		obj = i915_gem_object_create_lmem(gt->i915, size,
+						  I915_BO_ALLOC_CPU_CLEAR |
+						  I915_BO_ALLOC_CONTIGUOUS |
+						  I915_BO_ALLOC_PM_EARLY);
+	else
+		obj = i915_gem_object_create_shmem(gt->i915, size);
+
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, &gt->ggtt->vm, NULL);
+	if (IS_ERR(vma))
+		goto err;
+
+	flags = PIN_OFFSET_BIAS | i915_ggtt_pin_bias(vma);
+	ret = i915_ggtt_pin(vma, NULL, 0, flags);
+	if (ret) {
+		vma = ERR_PTR(ret);
+		goto err;
+	}
+
+	return i915_vma_make_unshrinkable(vma);
+
+err:
+	i915_gem_object_put(obj);
+	return vma;
+}
+
+/**
+ * intel_guc_allocate_and_map_vma() - Allocate and map VMA for GuC usage
+ * @guc:	the guc
+ * @size:	size of area to allocate (both virtual space and memory)
+ * @out_vma:	return variable for the allocated vma pointer
+ * @out_vaddr:	return variable for the obj mapping
+ *
+ * This wrapper calls intel_guc_allocate_vma() and then maps the allocated
+ * object with I915_MAP_WB.
+ *
+ * Return:	0 if successful, a negative errno code otherwise.
+ */
+int intel_guc_allocate_and_map_vma(struct intel_guc *guc, u32 size,
+				   struct i915_vma **out_vma, void **out_vaddr)
+{
+	struct i915_vma *vma;
+	void *vaddr;
+
+	vma = intel_guc_allocate_vma(guc, size);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	vaddr = i915_gem_object_pin_map_unlocked(vma->obj,
+						 i915_coherent_map_type(guc_to_gt(guc)->i915,
+									vma->obj, true));
+	if (IS_ERR(vaddr)) {
+		i915_vma_unpin_and_release(&vma, 0);
+		return PTR_ERR(vaddr);
+	}
+
+	*out_vma = vma;
+	*out_vaddr = vaddr;
+
+	return 0;
+}
+
+static int __guc_action_self_cfg(struct intel_guc *guc, u16 key, u16 len, u64 value)
+{
+	u32 request[HOST2GUC_SELF_CFG_REQUEST_MSG_LEN] = {
+		FIELD_PREP(GUC_HXG_MSG_0_ORIGIN, GUC_HXG_ORIGIN_HOST) |
+		FIELD_PREP(GUC_HXG_MSG_0_TYPE, GUC_HXG_TYPE_REQUEST) |
+		FIELD_PREP(GUC_HXG_REQUEST_MSG_0_ACTION, GUC_ACTION_HOST2GUC_SELF_CFG),
+		FIELD_PREP(HOST2GUC_SELF_CFG_REQUEST_MSG_1_KLV_KEY, key) |
+		FIELD_PREP(HOST2GUC_SELF_CFG_REQUEST_MSG_1_KLV_LEN, len),
+		FIELD_PREP(HOST2GUC_SELF_CFG_REQUEST_MSG_2_VALUE32, lower_32_bits(value)),
+		FIELD_PREP(HOST2GUC_SELF_CFG_REQUEST_MSG_3_VALUE64, upper_32_bits(value)),
+	};
+	int ret;
+
+	GEM_BUG_ON(len > 2);
+	GEM_BUG_ON(len == 1 && upper_32_bits(value));
+
+	/* Self config must go over MMIO */
+	ret = intel_guc_send_mmio(guc, request, ARRAY_SIZE(request), NULL, 0);
+
+	if (unlikely(ret < 0))
+		return ret;
+	if (unlikely(ret > 1))
+		return -EPROTO;
+	if (unlikely(!ret))
+		return -ENOKEY;
+
+	return 0;
+}
+
+static int __guc_self_cfg(struct intel_guc *guc, u16 key, u16 len, u64 value)
+{
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	int err = __guc_action_self_cfg(guc, key, len, value);
+
+	if (unlikely(err))
+		i915_probe_error(i915, "Unsuccessful self-config (%pe) key %#hx value %#llx\n",
+				 ERR_PTR(err), key, value);
+	return err;
+}
+
+int intel_guc_self_cfg32(struct intel_guc *guc, u16 key, u32 value)
+{
+	return __guc_self_cfg(guc, key, 1, value);
+}
+
+int intel_guc_self_cfg64(struct intel_guc *guc, u16 key, u64 value)
+{
+	return __guc_self_cfg(guc, key, 2, value);
+}
+
+/**
+ * intel_guc_load_status - dump information about GuC load status
+ * @guc: the GuC
+ * @p: the &drm_printer
+ *
+ * Pretty printer for GuC load status.
+ */
+void intel_guc_load_status(struct intel_guc *guc, struct drm_printer *p)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct intel_uncore *uncore = gt->uncore;
+	intel_wakeref_t wakeref;
+
+	if (!intel_guc_is_supported(guc)) {
+		drm_printf(p, "GuC not supported\n");
+		return;
+	}
+
+	if (!intel_guc_is_wanted(guc)) {
+		drm_printf(p, "GuC disabled\n");
+		return;
+	}
+
+	intel_uc_fw_dump(&guc->fw, p);
+
+	with_intel_runtime_pm(uncore->rpm, wakeref) {
+		u32 status = intel_uncore_read(uncore, GUC_STATUS);
+		u32 i;
+
+		drm_printf(p, "\nGuC status 0x%08x:\n", status);
+		drm_printf(p, "\tBootrom status = 0x%x\n",
+			   (status & GS_BOOTROM_MASK) >> GS_BOOTROM_SHIFT);
+		drm_printf(p, "\tuKernel status = 0x%x\n",
+			   (status & GS_UKERNEL_MASK) >> GS_UKERNEL_SHIFT);
+		drm_printf(p, "\tMIA Core status = 0x%x\n",
+			   (status & GS_MIA_MASK) >> GS_MIA_SHIFT);
+		drm_puts(p, "\nScratch registers:\n");
+		for (i = 0; i < 16; i++) {
+			drm_printf(p, "\t%2d: \t0x%x\n",
+				   i, intel_uncore_read(uncore, SOFT_SCRATCH(i)));
+		}
+	}
+}
+
+void intel_guc_write_barrier(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+
+	if (i915_gem_object_is_lmem(guc->ct.vma->obj)) {
+		/*
+		 * Ensure intel_uncore_write_fw can be used rather than
+		 * intel_uncore_write.
+		 */
+		GEM_BUG_ON(guc->send_regs.fw_domains);
+
+		/*
+		 * This register is used by the i915 and GuC for MMIO based
+		 * communication. Once we are in this code CTBs are the only
+		 * method the i915 uses to communicate with the GuC so it is
+		 * safe to write to this register (a value of 0 is NOP for MMIO
+		 * communication). If we ever start mixing CTBs and MMIOs a new
+		 * register will have to be chosen. This function is also used
+		 * to enforce ordering of a work queue item write and an update
+		 * to the process descriptor. When a work queue is being used,
+		 * CTBs are also the only mechanism of communication.
+		 */
+		intel_uncore_write_fw(gt->uncore, GEN11_SOFT_SCRATCH(0), 0);
+	} else {
+		/* wmb() sufficient for a barrier if in smem */
+		wmb();
+	}
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc.h b/drivers/gpu/drm/i915/gt/uc/intel_guc.h
new file mode 100644
index 000000000..804133df1
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc.h
@@ -0,0 +1,469 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_H_
+#define _INTEL_GUC_H_
+
+#include <linux/delay.h>
+#include <linux/iosys-map.h>
+#include <linux/xarray.h>
+
+#include "intel_guc_ct.h"
+#include "intel_guc_fw.h"
+#include "intel_guc_fwif.h"
+#include "intel_guc_log.h"
+#include "intel_guc_reg.h"
+#include "intel_guc_slpc_types.h"
+#include "intel_uc_fw.h"
+#include "intel_uncore.h"
+#include "i915_utils.h"
+#include "i915_vma.h"
+
+struct __guc_ads_blob;
+struct intel_guc_state_capture;
+
+/**
+ * struct intel_guc - Top level structure of GuC.
+ *
+ * It handles firmware loading and manages client pool. intel_guc owns an
+ * i915_sched_engine for submission.
+ */
+struct intel_guc {
+	/** @fw: the GuC firmware */
+	struct intel_uc_fw fw;
+	/** @log: sub-structure containing GuC log related data and objects */
+	struct intel_guc_log log;
+	/** @ct: the command transport communication channel */
+	struct intel_guc_ct ct;
+	/** @slpc: sub-structure containing SLPC related data and objects */
+	struct intel_guc_slpc slpc;
+	/** @capture: the error-state-capture module's data and objects */
+	struct intel_guc_state_capture *capture;
+
+	/** @sched_engine: Global engine used to submit requests to GuC */
+	struct i915_sched_engine *sched_engine;
+	/**
+	 * @stalled_request: if GuC can't process a request for any reason, we
+	 * save it until GuC restarts processing. No other request can be
+	 * submitted until the stalled request is processed.
+	 */
+	struct i915_request *stalled_request;
+	/**
+	 * @submission_stall_reason: reason why submission is stalled
+	 */
+	enum {
+		STALL_NONE,
+		STALL_REGISTER_CONTEXT,
+		STALL_MOVE_LRC_TAIL,
+		STALL_ADD_REQUEST,
+	} submission_stall_reason;
+
+	/* intel_guc_recv interrupt related state */
+	/** @irq_lock: protects GuC irq state */
+	spinlock_t irq_lock;
+	/**
+	 * @msg_enabled_mask: mask of events that are processed when receiving
+	 * an INTEL_GUC_ACTION_DEFAULT G2H message.
+	 */
+	unsigned int msg_enabled_mask;
+
+	/**
+	 * @outstanding_submission_g2h: number of outstanding GuC to Host
+	 * responses related to GuC submission, used to determine if the GT is
+	 * idle
+	 */
+	atomic_t outstanding_submission_g2h;
+
+	/** @interrupts: pointers to GuC interrupt-managing functions. */
+	struct {
+		void (*reset)(struct intel_guc *guc);
+		void (*enable)(struct intel_guc *guc);
+		void (*disable)(struct intel_guc *guc);
+	} interrupts;
+
+	/**
+	 * @submission_state: sub-structure for submission state protected by
+	 * single lock
+	 */
+	struct {
+		/**
+		 * @lock: protects everything in submission_state,
+		 * ce->guc_id.id, and ce->guc_id.ref when transitioning in and
+		 * out of zero
+		 */
+		spinlock_t lock;
+		/**
+		 * @guc_ids: used to allocate new guc_ids, single-lrc
+		 */
+		struct ida guc_ids;
+		/**
+		 * @num_guc_ids: Number of guc_ids, selftest feature to be able
+		 * to reduce this number while testing.
+		 */
+		int num_guc_ids;
+		/**
+		 * @guc_ids_bitmap: used to allocate new guc_ids, multi-lrc
+		 */
+		unsigned long *guc_ids_bitmap;
+		/**
+		 * @guc_id_list: list of intel_context with valid guc_ids but no
+		 * refs
+		 */
+		struct list_head guc_id_list;
+		/**
+		 * @destroyed_contexts: list of contexts waiting to be destroyed
+		 * (deregistered with the GuC)
+		 */
+		struct list_head destroyed_contexts;
+		/**
+		 * @destroyed_worker: worker to deregister contexts, need as we
+		 * need to take a GT PM reference and can't from destroy
+		 * function as it might be in an atomic context (no sleeping)
+		 */
+		struct work_struct destroyed_worker;
+		/**
+		 * @reset_fail_worker: worker to trigger a GT reset after an
+		 * engine reset fails
+		 */
+		struct work_struct reset_fail_worker;
+		/**
+		 * @reset_fail_mask: mask of engines that failed to reset
+		 */
+		intel_engine_mask_t reset_fail_mask;
+	} submission_state;
+
+	/**
+	 * @submission_supported: tracks whether we support GuC submission on
+	 * the current platform
+	 */
+	bool submission_supported;
+	/** @submission_selected: tracks whether the user enabled GuC submission */
+	bool submission_selected;
+	/** @submission_initialized: tracks whether GuC submission has been initialised */
+	bool submission_initialized;
+	/**
+	 * @rc_supported: tracks whether we support GuC rc on the current platform
+	 */
+	bool rc_supported;
+	/** @rc_selected: tracks whether the user enabled GuC rc */
+	bool rc_selected;
+
+	/** @ads_vma: object allocated to hold the GuC ADS */
+	struct i915_vma *ads_vma;
+	/** @ads_map: contents of the GuC ADS */
+	struct iosys_map ads_map;
+	/** @ads_regset_size: size of the save/restore regsets in the ADS */
+	u32 ads_regset_size;
+	/**
+	 * @ads_regset_count: number of save/restore registers in the ADS for
+	 * each engine
+	 */
+	u32 ads_regset_count[I915_NUM_ENGINES];
+	/** @ads_regset: save/restore regsets in the ADS */
+	struct guc_mmio_reg *ads_regset;
+	/** @ads_golden_ctxt_size: size of the golden contexts in the ADS */
+	u32 ads_golden_ctxt_size;
+	/** @ads_capture_size: size of register lists in the ADS used for error capture */
+	u32 ads_capture_size;
+	/** @ads_engine_usage_size: size of engine usage in the ADS */
+	u32 ads_engine_usage_size;
+
+	/** @lrc_desc_pool_v69: object allocated to hold the GuC LRC descriptor pool */
+	struct i915_vma *lrc_desc_pool_v69;
+	/** @lrc_desc_pool_vaddr_v69: contents of the GuC LRC descriptor pool */
+	void *lrc_desc_pool_vaddr_v69;
+
+	/**
+	 * @context_lookup: used to resolve intel_context from guc_id, if a
+	 * context is present in this structure it is registered with the GuC
+	 */
+	struct xarray context_lookup;
+
+	/** @params: Control params for fw initialization */
+	u32 params[GUC_CTL_MAX_DWORDS];
+
+	/** @send_regs: GuC's FW specific registers used for sending MMIO H2G */
+	struct {
+		u32 base;
+		unsigned int count;
+		enum forcewake_domains fw_domains;
+	} send_regs;
+
+	/** @notify_reg: register used to send interrupts to the GuC FW */
+	i915_reg_t notify_reg;
+
+	/**
+	 * @mmio_msg: notification bitmask that the GuC writes in one of its
+	 * registers when the CT channel is disabled, to be processed when the
+	 * channel is back up.
+	 */
+	u32 mmio_msg;
+
+	/** @send_mutex: used to serialize the intel_guc_send actions */
+	struct mutex send_mutex;
+
+	/**
+	 * @timestamp: GT timestamp object that stores a copy of the timestamp
+	 * and adjusts it for overflow using a worker.
+	 */
+	struct {
+		/**
+		 * @lock: Lock protecting the below fields and the engine stats.
+		 */
+		spinlock_t lock;
+
+		/**
+		 * @gt_stamp: 64 bit extended value of the GT timestamp.
+		 */
+		u64 gt_stamp;
+
+		/**
+		 * @ping_delay: Period for polling the GT timestamp for
+		 * overflow.
+		 */
+		unsigned long ping_delay;
+
+		/**
+		 * @work: Periodic work to adjust GT timestamp, engine and
+		 * context usage for overflows.
+		 */
+		struct delayed_work work;
+
+		/**
+		 * @shift: Right shift value for the gpm timestamp
+		 */
+		u32 shift;
+
+		/**
+		 * @last_stat_jiffies: jiffies at last actual stats collection time
+		 * We use this timestamp to ensure we don't oversample the
+		 * stats because runtime power management events can trigger
+		 * stats collection at much higher rates than required.
+		 */
+		unsigned long last_stat_jiffies;
+	} timestamp;
+
+#ifdef CONFIG_DRM_I915_SELFTEST
+	/**
+	 * @number_guc_id_stolen: The number of guc_ids that have been stolen
+	 */
+	int number_guc_id_stolen;
+#endif
+};
+
+static inline struct intel_guc *log_to_guc(struct intel_guc_log *log)
+{
+	return container_of(log, struct intel_guc, log);
+}
+
+static
+inline int intel_guc_send(struct intel_guc *guc, const u32 *action, u32 len)
+{
+	return intel_guc_ct_send(&guc->ct, action, len, NULL, 0, 0);
+}
+
+static
+inline int intel_guc_send_nb(struct intel_guc *guc, const u32 *action, u32 len,
+			     u32 g2h_len_dw)
+{
+	return intel_guc_ct_send(&guc->ct, action, len, NULL, 0,
+				 MAKE_SEND_FLAGS(g2h_len_dw));
+}
+
+static inline int
+intel_guc_send_and_receive(struct intel_guc *guc, const u32 *action, u32 len,
+			   u32 *response_buf, u32 response_buf_size)
+{
+	return intel_guc_ct_send(&guc->ct, action, len,
+				 response_buf, response_buf_size, 0);
+}
+
+static inline int intel_guc_send_busy_loop(struct intel_guc *guc,
+					   const u32 *action,
+					   u32 len,
+					   u32 g2h_len_dw,
+					   bool loop)
+{
+	int err;
+	unsigned int sleep_period_ms = 1;
+	bool not_atomic = !in_atomic() && !irqs_disabled();
+
+	/*
+	 * FIXME: Have caller pass in if we are in an atomic context to avoid
+	 * using in_atomic(). It is likely safe here as we check for irqs
+	 * disabled which basically all the spin locks in the i915 do but
+	 * regardless this should be cleaned up.
+	 */
+
+	/* No sleeping with spin locks, just busy loop */
+	might_sleep_if(loop && not_atomic);
+
+retry:
+	err = intel_guc_send_nb(guc, action, len, g2h_len_dw);
+	if (unlikely(err == -EBUSY && loop)) {
+		if (likely(not_atomic)) {
+			if (msleep_interruptible(sleep_period_ms))
+				return -EINTR;
+			sleep_period_ms = sleep_period_ms << 1;
+		} else {
+			cpu_relax();
+		}
+		goto retry;
+	}
+
+	return err;
+}
+
+static inline void intel_guc_to_host_event_handler(struct intel_guc *guc)
+{
+	intel_guc_ct_event_handler(&guc->ct);
+}
+
+/* GuC addresses above GUC_GGTT_TOP also don't map through the GTT */
+#define GUC_GGTT_TOP	0xFEE00000
+
+/**
+ * intel_guc_ggtt_offset() - Get and validate the GGTT offset of @vma
+ * @guc: intel_guc structure.
+ * @vma: i915 graphics virtual memory area.
+ *
+ * GuC does not allow any gfx GGTT address that falls into range
+ * [0, ggtt.pin_bias), which is reserved for Boot ROM, SRAM and WOPCM.
+ * Currently, in order to exclude [0, ggtt.pin_bias) address space from
+ * GGTT, all gfx objects used by GuC are allocated with intel_guc_allocate_vma()
+ * and pinned with PIN_OFFSET_BIAS along with the value of ggtt.pin_bias.
+ *
+ * Return: GGTT offset of the @vma.
+ */
+static inline u32 intel_guc_ggtt_offset(struct intel_guc *guc,
+					struct i915_vma *vma)
+{
+	u32 offset = i915_ggtt_offset(vma);
+
+	GEM_BUG_ON(offset < i915_ggtt_pin_bias(vma));
+	GEM_BUG_ON(range_overflows_t(u64, offset, vma->size, GUC_GGTT_TOP));
+
+	return offset;
+}
+
+void intel_guc_init_early(struct intel_guc *guc);
+void intel_guc_init_late(struct intel_guc *guc);
+void intel_guc_init_send_regs(struct intel_guc *guc);
+void intel_guc_write_params(struct intel_guc *guc);
+int intel_guc_init(struct intel_guc *guc);
+void intel_guc_fini(struct intel_guc *guc);
+void intel_guc_notify(struct intel_guc *guc);
+int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len,
+			u32 *response_buf, u32 response_buf_size);
+int intel_guc_to_host_process_recv_msg(struct intel_guc *guc,
+				       const u32 *payload, u32 len);
+int intel_guc_auth_huc(struct intel_guc *guc, u32 rsa_offset);
+int intel_guc_suspend(struct intel_guc *guc);
+int intel_guc_resume(struct intel_guc *guc);
+struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size);
+int intel_guc_allocate_and_map_vma(struct intel_guc *guc, u32 size,
+				   struct i915_vma **out_vma, void **out_vaddr);
+int intel_guc_self_cfg32(struct intel_guc *guc, u16 key, u32 value);
+int intel_guc_self_cfg64(struct intel_guc *guc, u16 key, u64 value);
+
+static inline bool intel_guc_is_supported(struct intel_guc *guc)
+{
+	return intel_uc_fw_is_supported(&guc->fw);
+}
+
+static inline bool intel_guc_is_wanted(struct intel_guc *guc)
+{
+	return intel_uc_fw_is_enabled(&guc->fw);
+}
+
+static inline bool intel_guc_is_used(struct intel_guc *guc)
+{
+	GEM_BUG_ON(__intel_uc_fw_status(&guc->fw) == INTEL_UC_FIRMWARE_SELECTED);
+	return intel_uc_fw_is_available(&guc->fw);
+}
+
+static inline bool intel_guc_is_fw_running(struct intel_guc *guc)
+{
+	return intel_uc_fw_is_running(&guc->fw);
+}
+
+static inline bool intel_guc_is_ready(struct intel_guc *guc)
+{
+	return intel_guc_is_fw_running(guc) && intel_guc_ct_enabled(&guc->ct);
+}
+
+static inline void intel_guc_reset_interrupts(struct intel_guc *guc)
+{
+	guc->interrupts.reset(guc);
+}
+
+static inline void intel_guc_enable_interrupts(struct intel_guc *guc)
+{
+	guc->interrupts.enable(guc);
+}
+
+static inline void intel_guc_disable_interrupts(struct intel_guc *guc)
+{
+	guc->interrupts.disable(guc);
+}
+
+static inline int intel_guc_sanitize(struct intel_guc *guc)
+{
+	intel_uc_fw_sanitize(&guc->fw);
+	intel_guc_disable_interrupts(guc);
+	intel_guc_ct_sanitize(&guc->ct);
+	guc->mmio_msg = 0;
+
+	return 0;
+}
+
+static inline void intel_guc_enable_msg(struct intel_guc *guc, u32 mask)
+{
+	spin_lock_irq(&guc->irq_lock);
+	guc->msg_enabled_mask |= mask;
+	spin_unlock_irq(&guc->irq_lock);
+}
+
+static inline void intel_guc_disable_msg(struct intel_guc *guc, u32 mask)
+{
+	spin_lock_irq(&guc->irq_lock);
+	guc->msg_enabled_mask &= ~mask;
+	spin_unlock_irq(&guc->irq_lock);
+}
+
+int intel_guc_wait_for_idle(struct intel_guc *guc, long timeout);
+
+int intel_guc_deregister_done_process_msg(struct intel_guc *guc,
+					  const u32 *msg, u32 len);
+int intel_guc_sched_done_process_msg(struct intel_guc *guc,
+				     const u32 *msg, u32 len);
+int intel_guc_context_reset_process_msg(struct intel_guc *guc,
+					const u32 *msg, u32 len);
+int intel_guc_engine_failure_process_msg(struct intel_guc *guc,
+					 const u32 *msg, u32 len);
+int intel_guc_error_capture_process_msg(struct intel_guc *guc,
+					const u32 *msg, u32 len);
+
+struct intel_engine_cs *
+intel_guc_lookup_engine(struct intel_guc *guc, u8 guc_class, u8 instance);
+
+void intel_guc_find_hung_context(struct intel_engine_cs *engine);
+
+int intel_guc_global_policies_update(struct intel_guc *guc);
+
+void intel_guc_context_ban(struct intel_context *ce, struct i915_request *rq);
+
+void intel_guc_submission_reset_prepare(struct intel_guc *guc);
+void intel_guc_submission_reset(struct intel_guc *guc, intel_engine_mask_t stalled);
+void intel_guc_submission_reset_finish(struct intel_guc *guc);
+void intel_guc_submission_cancel_requests(struct intel_guc *guc);
+
+void intel_guc_load_status(struct intel_guc *guc, struct drm_printer *p);
+
+void intel_guc_write_barrier(struct intel_guc *guc);
+
+void intel_guc_dump_time_info(struct intel_guc *guc, struct drm_printer *p);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c
new file mode 100644
index 000000000..74cbe8eaf
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.c
@@ -0,0 +1,905 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#include <linux/bsearch.h>
+
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_mcr.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_lrc.h"
+#include "gt/shmem_utils.h"
+#include "intel_guc_ads.h"
+#include "intel_guc_capture.h"
+#include "intel_guc_fwif.h"
+#include "intel_uc.h"
+#include "i915_drv.h"
+
+/*
+ * The Additional Data Struct (ADS) has pointers for different buffers used by
+ * the GuC. One single gem object contains the ADS struct itself (guc_ads) and
+ * all the extra buffers indirectly linked via the ADS struct's entries.
+ *
+ * Layout of the ADS blob allocated for the GuC:
+ *
+ *      +---------------------------------------+ <== base
+ *      | guc_ads                               |
+ *      +---------------------------------------+
+ *      | guc_policies                          |
+ *      +---------------------------------------+
+ *      | guc_gt_system_info                    |
+ *      +---------------------------------------+
+ *      | guc_engine_usage                      |
+ *      +---------------------------------------+ <== static
+ *      | guc_mmio_reg[countA] (engine 0.0)     |
+ *      | guc_mmio_reg[countB] (engine 0.1)     |
+ *      | guc_mmio_reg[countC] (engine 1.0)     |
+ *      |   ...                                 |
+ *      +---------------------------------------+ <== dynamic
+ *      | padding                               |
+ *      +---------------------------------------+ <== 4K aligned
+ *      | golden contexts                       |
+ *      +---------------------------------------+
+ *      | padding                               |
+ *      +---------------------------------------+ <== 4K aligned
+ *      | capture lists                         |
+ *      +---------------------------------------+
+ *      | padding                               |
+ *      +---------------------------------------+ <== 4K aligned
+ *      | private data                          |
+ *      +---------------------------------------+
+ *      | padding                               |
+ *      +---------------------------------------+ <== 4K aligned
+ */
+struct __guc_ads_blob {
+	struct guc_ads ads;
+	struct guc_policies policies;
+	struct guc_gt_system_info system_info;
+	struct guc_engine_usage engine_usage;
+	/* From here on, location is dynamic! Refer to above diagram. */
+	struct guc_mmio_reg regset[];
+} __packed;
+
+#define ads_blob_read(guc_, field_)					\
+	iosys_map_rd_field(&(guc_)->ads_map, 0, struct __guc_ads_blob, field_)
+
+#define ads_blob_write(guc_, field_, val_)				\
+	iosys_map_wr_field(&(guc_)->ads_map, 0, struct __guc_ads_blob,	\
+			   field_, val_)
+
+#define info_map_write(map_, field_, val_) \
+	iosys_map_wr_field(map_, 0, struct guc_gt_system_info, field_, val_)
+
+#define info_map_read(map_, field_) \
+	iosys_map_rd_field(map_, 0, struct guc_gt_system_info, field_)
+
+static u32 guc_ads_regset_size(struct intel_guc *guc)
+{
+	GEM_BUG_ON(!guc->ads_regset_size);
+	return guc->ads_regset_size;
+}
+
+static u32 guc_ads_golden_ctxt_size(struct intel_guc *guc)
+{
+	return PAGE_ALIGN(guc->ads_golden_ctxt_size);
+}
+
+static u32 guc_ads_capture_size(struct intel_guc *guc)
+{
+	return PAGE_ALIGN(guc->ads_capture_size);
+}
+
+static u32 guc_ads_private_data_size(struct intel_guc *guc)
+{
+	return PAGE_ALIGN(guc->fw.private_data_size);
+}
+
+static u32 guc_ads_regset_offset(struct intel_guc *guc)
+{
+	return offsetof(struct __guc_ads_blob, regset);
+}
+
+static u32 guc_ads_golden_ctxt_offset(struct intel_guc *guc)
+{
+	u32 offset;
+
+	offset = guc_ads_regset_offset(guc) +
+		 guc_ads_regset_size(guc);
+
+	return PAGE_ALIGN(offset);
+}
+
+static u32 guc_ads_capture_offset(struct intel_guc *guc)
+{
+	u32 offset;
+
+	offset = guc_ads_golden_ctxt_offset(guc) +
+		 guc_ads_golden_ctxt_size(guc);
+
+	return PAGE_ALIGN(offset);
+}
+
+static u32 guc_ads_private_data_offset(struct intel_guc *guc)
+{
+	u32 offset;
+
+	offset = guc_ads_capture_offset(guc) +
+		 guc_ads_capture_size(guc);
+
+	return PAGE_ALIGN(offset);
+}
+
+static u32 guc_ads_blob_size(struct intel_guc *guc)
+{
+	return guc_ads_private_data_offset(guc) +
+	       guc_ads_private_data_size(guc);
+}
+
+static void guc_policies_init(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct drm_i915_private *i915 = gt->i915;
+	u32 global_flags = 0;
+
+	ads_blob_write(guc, policies.dpc_promote_time,
+		       GLOBAL_POLICY_DEFAULT_DPC_PROMOTE_TIME_US);
+	ads_blob_write(guc, policies.max_num_work_items,
+		       GLOBAL_POLICY_MAX_NUM_WI);
+
+	if (i915->params.reset < 2)
+		global_flags |= GLOBAL_POLICY_DISABLE_ENGINE_RESET;
+
+	ads_blob_write(guc, policies.global_flags, global_flags);
+	ads_blob_write(guc, policies.is_valid, 1);
+}
+
+void intel_guc_ads_print_policy_info(struct intel_guc *guc,
+				     struct drm_printer *dp)
+{
+	if (unlikely(iosys_map_is_null(&guc->ads_map)))
+		return;
+
+	drm_printf(dp, "Global scheduling policies:\n");
+	drm_printf(dp, "  DPC promote time   = %u\n",
+		   ads_blob_read(guc, policies.dpc_promote_time));
+	drm_printf(dp, "  Max num work items = %u\n",
+		   ads_blob_read(guc, policies.max_num_work_items));
+	drm_printf(dp, "  Flags              = %u\n",
+		   ads_blob_read(guc, policies.global_flags));
+}
+
+static int guc_action_policies_update(struct intel_guc *guc, u32 policy_offset)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_GLOBAL_SCHED_POLICY_CHANGE,
+		policy_offset
+	};
+
+	return intel_guc_send_busy_loop(guc, action, ARRAY_SIZE(action), 0, true);
+}
+
+int intel_guc_global_policies_update(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	u32 scheduler_policies;
+	intel_wakeref_t wakeref;
+	int ret;
+
+	if (iosys_map_is_null(&guc->ads_map))
+		return -EOPNOTSUPP;
+
+	scheduler_policies = ads_blob_read(guc, ads.scheduler_policies);
+	GEM_BUG_ON(!scheduler_policies);
+
+	guc_policies_init(guc);
+
+	if (!intel_guc_is_ready(guc))
+		return 0;
+
+	with_intel_runtime_pm(&gt->i915->runtime_pm, wakeref)
+		ret = guc_action_policies_update(guc, scheduler_policies);
+
+	return ret;
+}
+
+static void guc_mapping_table_init(struct intel_gt *gt,
+				   struct iosys_map *info_map)
+{
+	unsigned int i, j;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/* Table must be set to invalid values for entries not used */
+	for (i = 0; i < GUC_MAX_ENGINE_CLASSES; ++i)
+		for (j = 0; j < GUC_MAX_INSTANCES_PER_CLASS; ++j)
+			info_map_write(info_map, mapping_table[i][j],
+				       GUC_MAX_INSTANCES_PER_CLASS);
+
+	for_each_engine(engine, gt, id) {
+		u8 guc_class = engine_class_to_guc_class(engine->class);
+
+		info_map_write(info_map, mapping_table[guc_class][ilog2(engine->logical_mask)],
+			       engine->instance);
+	}
+}
+
+/*
+ * The save/restore register list must be pre-calculated to a temporary
+ * buffer before it can be copied inside the ADS.
+ */
+struct temp_regset {
+	/*
+	 * ptr to the section of the storage for the engine currently being
+	 * worked on
+	 */
+	struct guc_mmio_reg *registers;
+	/* ptr to the base of the allocated storage for all engines */
+	struct guc_mmio_reg *storage;
+	u32 storage_used;
+	u32 storage_max;
+};
+
+static int guc_mmio_reg_cmp(const void *a, const void *b)
+{
+	const struct guc_mmio_reg *ra = a;
+	const struct guc_mmio_reg *rb = b;
+
+	return (int)ra->offset - (int)rb->offset;
+}
+
+static struct guc_mmio_reg * __must_check
+__mmio_reg_add(struct temp_regset *regset, struct guc_mmio_reg *reg)
+{
+	u32 pos = regset->storage_used;
+	struct guc_mmio_reg *slot;
+
+	if (pos >= regset->storage_max) {
+		size_t size = ALIGN((pos + 1) * sizeof(*slot), PAGE_SIZE);
+		struct guc_mmio_reg *r = krealloc(regset->storage,
+						  size, GFP_KERNEL);
+		if (!r) {
+			WARN_ONCE(1, "Incomplete regset list: can't add register (%d)\n",
+				  -ENOMEM);
+			return ERR_PTR(-ENOMEM);
+		}
+
+		regset->registers = r + (regset->registers - regset->storage);
+		regset->storage = r;
+		regset->storage_max = size / sizeof(*slot);
+	}
+
+	slot = &regset->storage[pos];
+	regset->storage_used++;
+	*slot = *reg;
+
+	return slot;
+}
+
+#define GUC_REGSET_STEERING(group, instance) ( \
+	FIELD_PREP(GUC_REGSET_STEERING_GROUP, (group)) | \
+	FIELD_PREP(GUC_REGSET_STEERING_INSTANCE, (instance)) | \
+	GUC_REGSET_NEEDS_STEERING \
+)
+
+static long __must_check guc_mmio_reg_add(struct intel_gt *gt,
+					  struct temp_regset *regset,
+					  i915_reg_t reg, u32 flags)
+{
+	u32 count = regset->storage_used - (regset->registers - regset->storage);
+	u32 offset = i915_mmio_reg_offset(reg);
+	struct guc_mmio_reg entry = {
+		.offset = offset,
+		.flags = flags,
+	};
+	struct guc_mmio_reg *slot;
+	u8 group, inst;
+
+	/*
+	 * The mmio list is built using separate lists within the driver.
+	 * It's possible that at some point we may attempt to add the same
+	 * register more than once. Do not consider this an error; silently
+	 * move on if the register is already in the list.
+	 */
+	if (bsearch(&entry, regset->registers, count,
+		    sizeof(entry), guc_mmio_reg_cmp))
+		return 0;
+
+	/*
+	 * The GuC doesn't have a default steering, so we need to explicitly
+	 * steer all registers that need steering. However, we do not keep track
+	 * of all the steering ranges, only of those that have a chance of using
+	 * a non-default steering from the i915 pov. Instead of adding such
+	 * tracking, it is easier to just program the default steering for all
+	 * regs that don't need a non-default one.
+	 */
+	intel_gt_mcr_get_nonterminated_steering(gt, reg, &group, &inst);
+	entry.flags |= GUC_REGSET_STEERING(group, inst);
+
+	slot = __mmio_reg_add(regset, &entry);
+	if (IS_ERR(slot))
+		return PTR_ERR(slot);
+
+	while (slot-- > regset->registers) {
+		GEM_BUG_ON(slot[0].offset == slot[1].offset);
+		if (slot[1].offset > slot[0].offset)
+			break;
+
+		swap(slot[1], slot[0]);
+	}
+
+	return 0;
+}
+
+#define GUC_MMIO_REG_ADD(gt, regset, reg, masked) \
+	guc_mmio_reg_add(gt, \
+			 regset, \
+			 (reg), \
+			 (masked) ? GUC_REGSET_MASKED : 0)
+
+static int guc_mmio_regset_init(struct temp_regset *regset,
+				struct intel_engine_cs *engine)
+{
+	struct intel_gt *gt = engine->gt;
+	const u32 base = engine->mmio_base;
+	struct i915_wa_list *wal = &engine->wa_list;
+	struct i915_wa *wa;
+	unsigned int i;
+	int ret = 0;
+
+	/*
+	 * Each engine's registers point to a new start relative to
+	 * storage
+	 */
+	regset->registers = regset->storage + regset->storage_used;
+
+	ret |= GUC_MMIO_REG_ADD(gt, regset, RING_MODE_GEN7(base), true);
+	ret |= GUC_MMIO_REG_ADD(gt, regset, RING_HWS_PGA(base), false);
+	ret |= GUC_MMIO_REG_ADD(gt, regset, RING_IMR(base), false);
+
+	if ((engine->flags & I915_ENGINE_FIRST_RENDER_COMPUTE) &&
+	    CCS_MASK(engine->gt))
+		ret |= GUC_MMIO_REG_ADD(gt, regset, GEN12_RCU_MODE, true);
+
+	for (i = 0, wa = wal->list; i < wal->count; i++, wa++)
+		ret |= GUC_MMIO_REG_ADD(gt, regset, wa->reg, wa->masked_reg);
+
+	/* Be extra paranoid and include all whitelist registers. */
+	for (i = 0; i < RING_MAX_NONPRIV_SLOTS; i++)
+		ret |= GUC_MMIO_REG_ADD(gt, regset,
+					RING_FORCE_TO_NONPRIV(base, i),
+					false);
+
+	/* add in local MOCS registers */
+	for (i = 0; i < GEN9_LNCFCMOCS_REG_COUNT; i++)
+		ret |= GUC_MMIO_REG_ADD(gt, regset, GEN9_LNCFCMOCS(i), false);
+
+	return ret ? -1 : 0;
+}
+
+static long guc_mmio_reg_state_create(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct temp_regset temp_set = {};
+	long total = 0;
+	long ret;
+
+	for_each_engine(engine, gt, id) {
+		u32 used = temp_set.storage_used;
+
+		ret = guc_mmio_regset_init(&temp_set, engine);
+		if (ret < 0)
+			goto fail_regset_init;
+
+		guc->ads_regset_count[id] = temp_set.storage_used - used;
+		total += guc->ads_regset_count[id];
+	}
+
+	guc->ads_regset = temp_set.storage;
+
+	drm_dbg(&guc_to_gt(guc)->i915->drm, "Used %zu KB for temporary ADS regset\n",
+		(temp_set.storage_max * sizeof(struct guc_mmio_reg)) >> 10);
+
+	return total * sizeof(struct guc_mmio_reg);
+
+fail_regset_init:
+	kfree(temp_set.storage);
+	return ret;
+}
+
+static void guc_mmio_reg_state_init(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	u32 addr_ggtt, offset;
+
+	offset = guc_ads_regset_offset(guc);
+	addr_ggtt = intel_guc_ggtt_offset(guc, guc->ads_vma) + offset;
+
+	iosys_map_memcpy_to(&guc->ads_map, offset, guc->ads_regset,
+			    guc->ads_regset_size);
+
+	for_each_engine(engine, gt, id) {
+		u32 count = guc->ads_regset_count[id];
+		u8 guc_class;
+
+		/* Class index is checked in class converter */
+		GEM_BUG_ON(engine->instance >= GUC_MAX_INSTANCES_PER_CLASS);
+
+		guc_class = engine_class_to_guc_class(engine->class);
+
+		if (!count) {
+			ads_blob_write(guc,
+				       ads.reg_state_list[guc_class][engine->instance].address,
+				       0);
+			ads_blob_write(guc,
+				       ads.reg_state_list[guc_class][engine->instance].count,
+				       0);
+			continue;
+		}
+
+		ads_blob_write(guc,
+			       ads.reg_state_list[guc_class][engine->instance].address,
+			       addr_ggtt);
+		ads_blob_write(guc,
+			       ads.reg_state_list[guc_class][engine->instance].count,
+			       count);
+
+		addr_ggtt += count * sizeof(struct guc_mmio_reg);
+	}
+}
+
+static void fill_engine_enable_masks(struct intel_gt *gt,
+				     struct iosys_map *info_map)
+{
+	info_map_write(info_map, engine_enabled_masks[GUC_RENDER_CLASS], RCS_MASK(gt));
+	info_map_write(info_map, engine_enabled_masks[GUC_COMPUTE_CLASS], CCS_MASK(gt));
+	info_map_write(info_map, engine_enabled_masks[GUC_BLITTER_CLASS], BCS_MASK(gt));
+	info_map_write(info_map, engine_enabled_masks[GUC_VIDEO_CLASS], VDBOX_MASK(gt));
+	info_map_write(info_map, engine_enabled_masks[GUC_VIDEOENHANCE_CLASS], VEBOX_MASK(gt));
+}
+
+#define LR_HW_CONTEXT_SIZE (80 * sizeof(u32))
+#define XEHP_LR_HW_CONTEXT_SIZE (96 * sizeof(u32))
+#define LR_HW_CONTEXT_SZ(i915) (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50) ? \
+				    XEHP_LR_HW_CONTEXT_SIZE : \
+				    LR_HW_CONTEXT_SIZE)
+#define LRC_SKIP_SIZE(i915) (LRC_PPHWSP_SZ * PAGE_SIZE + LR_HW_CONTEXT_SZ(i915))
+static int guc_prep_golden_context(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	u32 addr_ggtt, offset;
+	u32 total_size = 0, alloc_size, real_size;
+	u8 engine_class, guc_class;
+	struct guc_gt_system_info local_info;
+	struct iosys_map info_map;
+
+	/*
+	 * Reserve the memory for the golden contexts and point GuC at it but
+	 * leave it empty for now. The context data will be filled in later
+	 * once there is something available to put there.
+	 *
+	 * Note that the HWSP and ring context are not included.
+	 *
+	 * Note also that the storage must be pinned in the GGTT, so that the
+	 * address won't change after GuC has been told where to find it. The
+	 * GuC will also validate that the LRC base + size fall within the
+	 * allowed GGTT range.
+	 */
+	if (!iosys_map_is_null(&guc->ads_map)) {
+		offset = guc_ads_golden_ctxt_offset(guc);
+		addr_ggtt = intel_guc_ggtt_offset(guc, guc->ads_vma) + offset;
+		info_map = IOSYS_MAP_INIT_OFFSET(&guc->ads_map,
+						 offsetof(struct __guc_ads_blob, system_info));
+	} else {
+		memset(&local_info, 0, sizeof(local_info));
+		iosys_map_set_vaddr(&info_map, &local_info);
+		fill_engine_enable_masks(gt, &info_map);
+	}
+
+	for (engine_class = 0; engine_class <= MAX_ENGINE_CLASS; ++engine_class) {
+		if (engine_class == OTHER_CLASS)
+			continue;
+
+		guc_class = engine_class_to_guc_class(engine_class);
+
+		if (!info_map_read(&info_map, engine_enabled_masks[guc_class]))
+			continue;
+
+		real_size = intel_engine_context_size(gt, engine_class);
+		alloc_size = PAGE_ALIGN(real_size);
+		total_size += alloc_size;
+
+		if (iosys_map_is_null(&guc->ads_map))
+			continue;
+
+		/*
+		 * This interface is slightly confusing. We need to pass the
+		 * base address of the full golden context and the size of just
+		 * the engine state, which is the section of the context image
+		 * that starts after the execlists context. This is required to
+		 * allow the GuC to restore just the engine state when a
+		 * watchdog reset occurs.
+		 * We calculate the engine state size by removing the size of
+		 * what comes before it in the context image (which is identical
+		 * on all engines).
+		 */
+		ads_blob_write(guc, ads.eng_state_size[guc_class],
+			       real_size - LRC_SKIP_SIZE(gt->i915));
+		ads_blob_write(guc, ads.golden_context_lrca[guc_class],
+			       addr_ggtt);
+
+		addr_ggtt += alloc_size;
+	}
+
+	/* Make sure current size matches what we calculated previously */
+	if (guc->ads_golden_ctxt_size)
+		GEM_BUG_ON(guc->ads_golden_ctxt_size != total_size);
+
+	return total_size;
+}
+
+static struct intel_engine_cs *find_engine_state(struct intel_gt *gt, u8 engine_class)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id) {
+		if (engine->class != engine_class)
+			continue;
+
+		if (!engine->default_state)
+			continue;
+
+		return engine;
+	}
+
+	return NULL;
+}
+
+static void guc_init_golden_context(struct intel_guc *guc)
+{
+	struct intel_engine_cs *engine;
+	struct intel_gt *gt = guc_to_gt(guc);
+	unsigned long offset;
+	u32 addr_ggtt, total_size = 0, alloc_size, real_size;
+	u8 engine_class, guc_class;
+
+	if (!intel_uc_uses_guc_submission(&gt->uc))
+		return;
+
+	GEM_BUG_ON(iosys_map_is_null(&guc->ads_map));
+
+	/*
+	 * Go back and fill in the golden context data now that it is
+	 * available.
+	 */
+	offset = guc_ads_golden_ctxt_offset(guc);
+	addr_ggtt = intel_guc_ggtt_offset(guc, guc->ads_vma) + offset;
+
+	for (engine_class = 0; engine_class <= MAX_ENGINE_CLASS; ++engine_class) {
+		if (engine_class == OTHER_CLASS)
+			continue;
+
+		guc_class = engine_class_to_guc_class(engine_class);
+		if (!ads_blob_read(guc, system_info.engine_enabled_masks[guc_class]))
+			continue;
+
+		real_size = intel_engine_context_size(gt, engine_class);
+		alloc_size = PAGE_ALIGN(real_size);
+		total_size += alloc_size;
+
+		engine = find_engine_state(gt, engine_class);
+		if (!engine) {
+			drm_err(&gt->i915->drm, "No engine state recorded for class %d!\n",
+				engine_class);
+			ads_blob_write(guc, ads.eng_state_size[guc_class], 0);
+			ads_blob_write(guc, ads.golden_context_lrca[guc_class], 0);
+			continue;
+		}
+
+		GEM_BUG_ON(ads_blob_read(guc, ads.eng_state_size[guc_class]) !=
+			   real_size - LRC_SKIP_SIZE(gt->i915));
+		GEM_BUG_ON(ads_blob_read(guc, ads.golden_context_lrca[guc_class]) != addr_ggtt);
+
+		addr_ggtt += alloc_size;
+
+		shmem_read_to_iosys_map(engine->default_state, 0, &guc->ads_map,
+					offset, real_size);
+		offset += alloc_size;
+	}
+
+	GEM_BUG_ON(guc->ads_golden_ctxt_size != total_size);
+}
+
+static int
+guc_capture_prep_lists(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	u32 ads_ggtt, capture_offset, null_ggtt, total_size = 0;
+	struct guc_gt_system_info local_info;
+	struct iosys_map info_map;
+	bool ads_is_mapped;
+	size_t size = 0;
+	void *ptr;
+	int i, j;
+
+	ads_is_mapped = !iosys_map_is_null(&guc->ads_map);
+	if (ads_is_mapped) {
+		capture_offset = guc_ads_capture_offset(guc);
+		ads_ggtt = intel_guc_ggtt_offset(guc, guc->ads_vma);
+		info_map = IOSYS_MAP_INIT_OFFSET(&guc->ads_map,
+						 offsetof(struct __guc_ads_blob, system_info));
+	} else {
+		memset(&local_info, 0, sizeof(local_info));
+		iosys_map_set_vaddr(&info_map, &local_info);
+		fill_engine_enable_masks(gt, &info_map);
+	}
+
+	/* first, set aside the first page for a capture_list with zero descriptors */
+	total_size = PAGE_SIZE;
+	if (ads_is_mapped) {
+		if (!intel_guc_capture_getnullheader(guc, &ptr, &size))
+			iosys_map_memcpy_to(&guc->ads_map, capture_offset, ptr, size);
+		null_ggtt = ads_ggtt + capture_offset;
+		capture_offset += PAGE_SIZE;
+	}
+
+	for (i = 0; i < GUC_CAPTURE_LIST_INDEX_MAX; i++) {
+		for (j = 0; j < GUC_MAX_ENGINE_CLASSES; j++) {
+
+			/* null list if we dont have said engine or list */
+			if (!info_map_read(&info_map, engine_enabled_masks[j])) {
+				if (ads_is_mapped) {
+					ads_blob_write(guc, ads.capture_class[i][j], null_ggtt);
+					ads_blob_write(guc, ads.capture_instance[i][j], null_ggtt);
+				}
+				continue;
+			}
+			if (intel_guc_capture_getlistsize(guc, i,
+							  GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS,
+							  j, &size)) {
+				if (ads_is_mapped)
+					ads_blob_write(guc, ads.capture_class[i][j], null_ggtt);
+				goto engine_instance_list;
+			}
+			total_size += size;
+			if (ads_is_mapped) {
+				if (total_size > guc->ads_capture_size ||
+				    intel_guc_capture_getlist(guc, i,
+							      GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS,
+							      j, &ptr)) {
+					ads_blob_write(guc, ads.capture_class[i][j], null_ggtt);
+					continue;
+				}
+				ads_blob_write(guc, ads.capture_class[i][j], ads_ggtt +
+					       capture_offset);
+				iosys_map_memcpy_to(&guc->ads_map, capture_offset, ptr, size);
+				capture_offset += size;
+			}
+engine_instance_list:
+			if (intel_guc_capture_getlistsize(guc, i,
+							  GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE,
+							  j, &size)) {
+				if (ads_is_mapped)
+					ads_blob_write(guc, ads.capture_instance[i][j], null_ggtt);
+				continue;
+			}
+			total_size += size;
+			if (ads_is_mapped) {
+				if (total_size > guc->ads_capture_size ||
+				    intel_guc_capture_getlist(guc, i,
+							      GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE,
+							      j, &ptr)) {
+					ads_blob_write(guc, ads.capture_instance[i][j], null_ggtt);
+					continue;
+				}
+				ads_blob_write(guc, ads.capture_instance[i][j], ads_ggtt +
+					       capture_offset);
+				iosys_map_memcpy_to(&guc->ads_map, capture_offset, ptr, size);
+				capture_offset += size;
+			}
+		}
+		if (intel_guc_capture_getlistsize(guc, i, GUC_CAPTURE_LIST_TYPE_GLOBAL, 0, &size)) {
+			if (ads_is_mapped)
+				ads_blob_write(guc, ads.capture_global[i], null_ggtt);
+			continue;
+		}
+		total_size += size;
+		if (ads_is_mapped) {
+			if (total_size > guc->ads_capture_size ||
+			    intel_guc_capture_getlist(guc, i, GUC_CAPTURE_LIST_TYPE_GLOBAL, 0,
+						      &ptr)) {
+				ads_blob_write(guc, ads.capture_global[i], null_ggtt);
+				continue;
+			}
+			ads_blob_write(guc, ads.capture_global[i], ads_ggtt + capture_offset);
+			iosys_map_memcpy_to(&guc->ads_map, capture_offset, ptr, size);
+			capture_offset += size;
+		}
+	}
+
+	if (guc->ads_capture_size && guc->ads_capture_size != PAGE_ALIGN(total_size))
+		drm_warn(&i915->drm, "GuC->ADS->Capture alloc size changed from %d to %d\n",
+			 guc->ads_capture_size, PAGE_ALIGN(total_size));
+
+	return PAGE_ALIGN(total_size);
+}
+
+static void __guc_ads_init(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct drm_i915_private *i915 = gt->i915;
+	struct iosys_map info_map = IOSYS_MAP_INIT_OFFSET(&guc->ads_map,
+			offsetof(struct __guc_ads_blob, system_info));
+	u32 base;
+
+	/* GuC scheduling policies */
+	guc_policies_init(guc);
+
+	/* System info */
+	fill_engine_enable_masks(gt, &info_map);
+
+	ads_blob_write(guc, system_info.generic_gt_sysinfo[GUC_GENERIC_GT_SYSINFO_SLICE_ENABLED],
+		       hweight8(gt->info.sseu.slice_mask));
+	ads_blob_write(guc, system_info.generic_gt_sysinfo[GUC_GENERIC_GT_SYSINFO_VDBOX_SFC_SUPPORT_MASK],
+		       gt->info.vdbox_sfc_access);
+
+	if (GRAPHICS_VER(i915) >= 12 && !IS_DGFX(i915)) {
+		u32 distdbreg = intel_uncore_read(gt->uncore,
+						  GEN12_DIST_DBS_POPULATED);
+		ads_blob_write(guc,
+			       system_info.generic_gt_sysinfo[GUC_GENERIC_GT_SYSINFO_DOORBELL_COUNT_PER_SQIDI],
+			       ((distdbreg >> GEN12_DOORBELLS_PER_SQIDI_SHIFT)
+				& GEN12_DOORBELLS_PER_SQIDI) + 1);
+	}
+
+	/* Golden contexts for re-initialising after a watchdog reset */
+	guc_prep_golden_context(guc);
+
+	guc_mapping_table_init(guc_to_gt(guc), &info_map);
+
+	base = intel_guc_ggtt_offset(guc, guc->ads_vma);
+
+	/* Lists for error capture debug */
+	guc_capture_prep_lists(guc);
+
+	/* ADS */
+	ads_blob_write(guc, ads.scheduler_policies, base +
+		       offsetof(struct __guc_ads_blob, policies));
+	ads_blob_write(guc, ads.gt_system_info, base +
+		       offsetof(struct __guc_ads_blob, system_info));
+
+	/* MMIO save/restore list */
+	guc_mmio_reg_state_init(guc);
+
+	/* Private Data */
+	ads_blob_write(guc, ads.private_data, base +
+		       guc_ads_private_data_offset(guc));
+
+	i915_gem_object_flush_map(guc->ads_vma->obj);
+}
+
+/**
+ * intel_guc_ads_create() - allocates and initializes GuC ADS.
+ * @guc: intel_guc struct
+ *
+ * GuC needs memory block (Additional Data Struct), where it will store
+ * some data. Allocate and initialize such memory block for GuC use.
+ */
+int intel_guc_ads_create(struct intel_guc *guc)
+{
+	void *ads_blob;
+	u32 size;
+	int ret;
+
+	GEM_BUG_ON(guc->ads_vma);
+
+	/*
+	 * Create reg state size dynamically on system memory to be copied to
+	 * the final ads blob on gt init/reset
+	 */
+	ret = guc_mmio_reg_state_create(guc);
+	if (ret < 0)
+		return ret;
+	guc->ads_regset_size = ret;
+
+	/* Likewise the golden contexts: */
+	ret = guc_prep_golden_context(guc);
+	if (ret < 0)
+		return ret;
+	guc->ads_golden_ctxt_size = ret;
+
+	/* Likewise the capture lists: */
+	ret = guc_capture_prep_lists(guc);
+	if (ret < 0)
+		return ret;
+	guc->ads_capture_size = ret;
+
+	/* Now the total size can be determined: */
+	size = guc_ads_blob_size(guc);
+
+	ret = intel_guc_allocate_and_map_vma(guc, size, &guc->ads_vma,
+					     &ads_blob);
+	if (ret)
+		return ret;
+
+	if (i915_gem_object_is_lmem(guc->ads_vma->obj))
+		iosys_map_set_vaddr_iomem(&guc->ads_map, (void __iomem *)ads_blob);
+	else
+		iosys_map_set_vaddr(&guc->ads_map, ads_blob);
+
+	__guc_ads_init(guc);
+
+	return 0;
+}
+
+void intel_guc_ads_init_late(struct intel_guc *guc)
+{
+	/*
+	 * The golden context setup requires the saved engine state from
+	 * __engines_record_defaults(). However, that requires engines to be
+	 * operational which means the ADS must already have been configured.
+	 * Fortunately, the golden context state is not needed until a hang
+	 * occurs, so it can be filled in during this late init phase.
+	 */
+	guc_init_golden_context(guc);
+}
+
+void intel_guc_ads_destroy(struct intel_guc *guc)
+{
+	i915_vma_unpin_and_release(&guc->ads_vma, I915_VMA_RELEASE_MAP);
+	iosys_map_clear(&guc->ads_map);
+	kfree(guc->ads_regset);
+}
+
+static void guc_ads_private_data_reset(struct intel_guc *guc)
+{
+	u32 size;
+
+	size = guc_ads_private_data_size(guc);
+	if (!size)
+		return;
+
+	iosys_map_memset(&guc->ads_map, guc_ads_private_data_offset(guc),
+			 0, size);
+}
+
+/**
+ * intel_guc_ads_reset() - prepares GuC Additional Data Struct for reuse
+ * @guc: intel_guc struct
+ *
+ * GuC stores some data in ADS, which might be stale after a reset.
+ * Reinitialize whole ADS in case any part of it was corrupted during
+ * previous GuC run.
+ */
+void intel_guc_ads_reset(struct intel_guc *guc)
+{
+	if (!guc->ads_vma)
+		return;
+
+	__guc_ads_init(guc);
+
+	guc_ads_private_data_reset(guc);
+}
+
+u32 intel_guc_engine_usage_offset(struct intel_guc *guc)
+{
+	return intel_guc_ggtt_offset(guc, guc->ads_vma) +
+		offsetof(struct __guc_ads_blob, engine_usage);
+}
+
+struct iosys_map intel_guc_engine_usage_record_map(struct intel_engine_cs *engine)
+{
+	struct intel_guc *guc = &engine->gt->uc.guc;
+	u8 guc_class = engine_class_to_guc_class(engine->class);
+	size_t offset = offsetof(struct __guc_ads_blob,
+				 engine_usage.engines[guc_class][ilog2(engine->logical_mask)]);
+
+	return IOSYS_MAP_INIT_OFFSET(&guc->ads_map, offset);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.h
new file mode 100644
index 000000000..1c64f4d6e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ads.h
@@ -0,0 +1,25 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_ADS_H_
+#define _INTEL_GUC_ADS_H_
+
+#include <linux/types.h>
+#include <linux/iosys-map.h>
+
+struct intel_guc;
+struct drm_printer;
+struct intel_engine_cs;
+
+int intel_guc_ads_create(struct intel_guc *guc);
+void intel_guc_ads_destroy(struct intel_guc *guc);
+void intel_guc_ads_init_late(struct intel_guc *guc);
+void intel_guc_ads_reset(struct intel_guc *guc);
+void intel_guc_ads_print_policy_info(struct intel_guc *guc,
+				     struct drm_printer *p);
+struct iosys_map intel_guc_engine_usage_record_map(struct intel_engine_cs *engine);
+u32 intel_guc_engine_usage_offset(struct intel_guc *guc);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c
new file mode 100644
index 000000000..18a8466f8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c
@@ -0,0 +1,1685 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021-2022 Intel Corporation
+ */
+
+#include <linux/types.h>
+
+#include <drm/drm_print.h>
+
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_mcr.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_lrc.h"
+#include "guc_capture_fwif.h"
+#include "intel_guc_capture.h"
+#include "intel_guc_fwif.h"
+#include "i915_drv.h"
+#include "i915_gpu_error.h"
+#include "i915_irq.h"
+#include "i915_memcpy.h"
+#include "i915_reg.h"
+
+/*
+ * Define all device tables of GuC error capture register lists
+ * NOTE: For engine-registers, GuC only needs the register offsets
+ *       from the engine-mmio-base
+ */
+#define COMMON_BASE_GLOBAL \
+	{ FORCEWAKE_MT,             0,      0, "FORCEWAKE" }
+
+#define COMMON_GEN9BASE_GLOBAL \
+	{ ERROR_GEN6,               0,      0, "ERROR_GEN6" }, \
+	{ DONE_REG,                 0,      0, "DONE_REG" }, \
+	{ HSW_GTT_CACHE_EN,         0,      0, "HSW_GTT_CACHE_EN" }
+
+#define GEN9_GLOBAL \
+	{ GEN8_FAULT_TLB_DATA0,     0,      0, "GEN8_FAULT_TLB_DATA0" }, \
+	{ GEN8_FAULT_TLB_DATA1,     0,      0, "GEN8_FAULT_TLB_DATA1" }
+
+#define COMMON_GEN12BASE_GLOBAL \
+	{ GEN12_FAULT_TLB_DATA0,    0,      0, "GEN12_FAULT_TLB_DATA0" }, \
+	{ GEN12_FAULT_TLB_DATA1,    0,      0, "GEN12_FAULT_TLB_DATA1" }, \
+	{ GEN12_AUX_ERR_DBG,        0,      0, "AUX_ERR_DBG" }, \
+	{ GEN12_GAM_DONE,           0,      0, "GAM_DONE" }, \
+	{ GEN12_RING_FAULT_REG,     0,      0, "FAULT_REG" }
+
+#define COMMON_BASE_ENGINE_INSTANCE \
+	{ RING_PSMI_CTL(0),         0,      0, "RC PSMI" }, \
+	{ RING_ESR(0),              0,      0, "ESR" }, \
+	{ RING_DMA_FADD(0),         0,      0, "RING_DMA_FADD_LDW" }, \
+	{ RING_DMA_FADD_UDW(0),     0,      0, "RING_DMA_FADD_UDW" }, \
+	{ RING_IPEIR(0),            0,      0, "IPEIR" }, \
+	{ RING_IPEHR(0),            0,      0, "IPEHR" }, \
+	{ RING_INSTPS(0),           0,      0, "INSTPS" }, \
+	{ RING_BBADDR(0),           0,      0, "RING_BBADDR_LOW32" }, \
+	{ RING_BBADDR_UDW(0),       0,      0, "RING_BBADDR_UP32" }, \
+	{ RING_BBSTATE(0),          0,      0, "BB_STATE" }, \
+	{ CCID(0),                  0,      0, "CCID" }, \
+	{ RING_ACTHD(0),            0,      0, "ACTHD_LDW" }, \
+	{ RING_ACTHD_UDW(0),        0,      0, "ACTHD_UDW" }, \
+	{ RING_INSTPM(0),           0,      0, "INSTPM" }, \
+	{ RING_INSTDONE(0),         0,      0, "INSTDONE" }, \
+	{ RING_NOPID(0),            0,      0, "RING_NOPID" }, \
+	{ RING_START(0),            0,      0, "START" }, \
+	{ RING_HEAD(0),             0,      0, "HEAD" }, \
+	{ RING_TAIL(0),             0,      0, "TAIL" }, \
+	{ RING_CTL(0),              0,      0, "CTL" }, \
+	{ RING_MI_MODE(0),          0,      0, "MODE" }, \
+	{ RING_CONTEXT_CONTROL(0),  0,      0, "RING_CONTEXT_CONTROL" }, \
+	{ RING_HWS_PGA(0),          0,      0, "HWS" }, \
+	{ RING_MODE_GEN7(0),        0,      0, "GFX_MODE" }, \
+	{ GEN8_RING_PDP_LDW(0, 0),  0,      0, "PDP0_LDW" }, \
+	{ GEN8_RING_PDP_UDW(0, 0),  0,      0, "PDP0_UDW" }, \
+	{ GEN8_RING_PDP_LDW(0, 1),  0,      0, "PDP1_LDW" }, \
+	{ GEN8_RING_PDP_UDW(0, 1),  0,      0, "PDP1_UDW" }, \
+	{ GEN8_RING_PDP_LDW(0, 2),  0,      0, "PDP2_LDW" }, \
+	{ GEN8_RING_PDP_UDW(0, 2),  0,      0, "PDP2_UDW" }, \
+	{ GEN8_RING_PDP_LDW(0, 3),  0,      0, "PDP3_LDW" }, \
+	{ GEN8_RING_PDP_UDW(0, 3),  0,      0, "PDP3_UDW" }
+
+#define COMMON_BASE_HAS_EU \
+	{ EIR,                      0,      0, "EIR" }
+
+#define COMMON_BASE_RENDER \
+	{ GEN7_SC_INSTDONE,         0,      0, "GEN7_SC_INSTDONE" }
+
+#define COMMON_GEN12BASE_RENDER \
+	{ GEN12_SC_INSTDONE_EXTRA,  0,      0, "GEN12_SC_INSTDONE_EXTRA" }, \
+	{ GEN12_SC_INSTDONE_EXTRA2, 0,      0, "GEN12_SC_INSTDONE_EXTRA2" }
+
+#define COMMON_GEN12BASE_VEC \
+	{ GEN12_SFC_DONE(0),        0,      0, "SFC_DONE[0]" }, \
+	{ GEN12_SFC_DONE(1),        0,      0, "SFC_DONE[1]" }, \
+	{ GEN12_SFC_DONE(2),        0,      0, "SFC_DONE[2]" }, \
+	{ GEN12_SFC_DONE(3),        0,      0, "SFC_DONE[3]" }
+
+/* XE_LPD - Global */
+static const struct __guc_mmio_reg_descr xe_lpd_global_regs[] = {
+	COMMON_BASE_GLOBAL,
+	COMMON_GEN9BASE_GLOBAL,
+	COMMON_GEN12BASE_GLOBAL,
+};
+
+/* XE_LPD - Render / Compute Per-Class */
+static const struct __guc_mmio_reg_descr xe_lpd_rc_class_regs[] = {
+	COMMON_BASE_HAS_EU,
+	COMMON_BASE_RENDER,
+	COMMON_GEN12BASE_RENDER,
+};
+
+/* GEN9/XE_LPD - Render / Compute Per-Engine-Instance */
+static const struct __guc_mmio_reg_descr xe_lpd_rc_inst_regs[] = {
+	COMMON_BASE_ENGINE_INSTANCE,
+};
+
+/* GEN9/XE_LPD - Media Decode/Encode Per-Engine-Instance */
+static const struct __guc_mmio_reg_descr xe_lpd_vd_inst_regs[] = {
+	COMMON_BASE_ENGINE_INSTANCE,
+};
+
+/* XE_LPD - Video Enhancement Per-Class */
+static const struct __guc_mmio_reg_descr xe_lpd_vec_class_regs[] = {
+	COMMON_GEN12BASE_VEC,
+};
+
+/* GEN9/XE_LPD - Video Enhancement Per-Engine-Instance */
+static const struct __guc_mmio_reg_descr xe_lpd_vec_inst_regs[] = {
+	COMMON_BASE_ENGINE_INSTANCE,
+};
+
+/* GEN9/XE_LPD - Blitter Per-Engine-Instance */
+static const struct __guc_mmio_reg_descr xe_lpd_blt_inst_regs[] = {
+	COMMON_BASE_ENGINE_INSTANCE,
+};
+
+/* GEN9 - Global */
+static const struct __guc_mmio_reg_descr default_global_regs[] = {
+	COMMON_BASE_GLOBAL,
+	COMMON_GEN9BASE_GLOBAL,
+	GEN9_GLOBAL,
+};
+
+static const struct __guc_mmio_reg_descr default_rc_class_regs[] = {
+	COMMON_BASE_HAS_EU,
+	COMMON_BASE_RENDER,
+};
+
+/*
+ * Empty lists:
+ * GEN9/XE_LPD - Blitter Per-Class
+ * GEN9/XE_LPD - Media Decode/Encode Per-Class
+ * GEN9 - VEC Class
+ */
+static const struct __guc_mmio_reg_descr empty_regs_list[] = {
+};
+
+#define TO_GCAP_DEF_OWNER(x) (GUC_CAPTURE_LIST_INDEX_##x)
+#define TO_GCAP_DEF_TYPE(x) (GUC_CAPTURE_LIST_TYPE_##x)
+#define MAKE_REGLIST(regslist, regsowner, regstype, class) \
+	{ \
+		regslist, \
+		ARRAY_SIZE(regslist), \
+		TO_GCAP_DEF_OWNER(regsowner), \
+		TO_GCAP_DEF_TYPE(regstype), \
+		class, \
+		NULL, \
+	}
+
+/* List of lists */
+static const struct __guc_mmio_reg_descr_group default_lists[] = {
+	MAKE_REGLIST(default_global_regs, PF, GLOBAL, 0),
+	MAKE_REGLIST(default_rc_class_regs, PF, ENGINE_CLASS, GUC_RENDER_CLASS),
+	MAKE_REGLIST(xe_lpd_rc_inst_regs, PF, ENGINE_INSTANCE, GUC_RENDER_CLASS),
+	MAKE_REGLIST(empty_regs_list, PF, ENGINE_CLASS, GUC_VIDEO_CLASS),
+	MAKE_REGLIST(xe_lpd_vd_inst_regs, PF, ENGINE_INSTANCE, GUC_VIDEO_CLASS),
+	MAKE_REGLIST(empty_regs_list, PF, ENGINE_CLASS, GUC_VIDEOENHANCE_CLASS),
+	MAKE_REGLIST(xe_lpd_vec_inst_regs, PF, ENGINE_INSTANCE, GUC_VIDEOENHANCE_CLASS),
+	MAKE_REGLIST(empty_regs_list, PF, ENGINE_CLASS, GUC_BLITTER_CLASS),
+	MAKE_REGLIST(xe_lpd_blt_inst_regs, PF, ENGINE_INSTANCE, GUC_BLITTER_CLASS),
+	{}
+};
+
+static const struct __guc_mmio_reg_descr_group xe_lpd_lists[] = {
+	MAKE_REGLIST(xe_lpd_global_regs, PF, GLOBAL, 0),
+	MAKE_REGLIST(xe_lpd_rc_class_regs, PF, ENGINE_CLASS, GUC_RENDER_CLASS),
+	MAKE_REGLIST(xe_lpd_rc_inst_regs, PF, ENGINE_INSTANCE, GUC_RENDER_CLASS),
+	MAKE_REGLIST(empty_regs_list, PF, ENGINE_CLASS, GUC_VIDEO_CLASS),
+	MAKE_REGLIST(xe_lpd_vd_inst_regs, PF, ENGINE_INSTANCE, GUC_VIDEO_CLASS),
+	MAKE_REGLIST(xe_lpd_vec_class_regs, PF, ENGINE_CLASS, GUC_VIDEOENHANCE_CLASS),
+	MAKE_REGLIST(xe_lpd_vec_inst_regs, PF, ENGINE_INSTANCE, GUC_VIDEOENHANCE_CLASS),
+	MAKE_REGLIST(empty_regs_list, PF, ENGINE_CLASS, GUC_BLITTER_CLASS),
+	MAKE_REGLIST(xe_lpd_blt_inst_regs, PF, ENGINE_INSTANCE, GUC_BLITTER_CLASS),
+	{}
+};
+
+static const struct __guc_mmio_reg_descr_group *
+guc_capture_get_one_list(const struct __guc_mmio_reg_descr_group *reglists,
+			 u32 owner, u32 type, u32 id)
+{
+	int i;
+
+	if (!reglists)
+		return NULL;
+
+	for (i = 0; reglists[i].list; ++i) {
+		if (reglists[i].owner == owner && reglists[i].type == type &&
+		    (reglists[i].engine == id || reglists[i].type == GUC_CAPTURE_LIST_TYPE_GLOBAL))
+			return &reglists[i];
+	}
+
+	return NULL;
+}
+
+static struct __guc_mmio_reg_descr_group *
+guc_capture_get_one_ext_list(struct __guc_mmio_reg_descr_group *reglists,
+			     u32 owner, u32 type, u32 id)
+{
+	int i;
+
+	if (!reglists)
+		return NULL;
+
+	for (i = 0; reglists[i].extlist; ++i) {
+		if (reglists[i].owner == owner && reglists[i].type == type &&
+		    (reglists[i].engine == id || reglists[i].type == GUC_CAPTURE_LIST_TYPE_GLOBAL))
+			return &reglists[i];
+	}
+
+	return NULL;
+}
+
+static void guc_capture_free_extlists(struct __guc_mmio_reg_descr_group *reglists)
+{
+	int i = 0;
+
+	if (!reglists)
+		return;
+
+	while (reglists[i].extlist)
+		kfree(reglists[i++].extlist);
+}
+
+struct __ext_steer_reg {
+	const char *name;
+	i915_reg_t reg;
+};
+
+static const struct __ext_steer_reg xe_extregs[] = {
+	{"GEN7_SAMPLER_INSTDONE", GEN7_SAMPLER_INSTDONE},
+	{"GEN7_ROW_INSTDONE", GEN7_ROW_INSTDONE}
+};
+
+static void __fill_ext_reg(struct __guc_mmio_reg_descr *ext,
+			   const struct __ext_steer_reg *extlist,
+			   int slice_id, int subslice_id)
+{
+	ext->reg = extlist->reg;
+	ext->flags = FIELD_PREP(GUC_REGSET_STEERING_GROUP, slice_id);
+	ext->flags |= FIELD_PREP(GUC_REGSET_STEERING_INSTANCE, subslice_id);
+	ext->regname = extlist->name;
+}
+
+static int
+__alloc_ext_regs(struct __guc_mmio_reg_descr_group *newlist,
+		 const struct __guc_mmio_reg_descr_group *rootlist, int num_regs)
+{
+	struct __guc_mmio_reg_descr *list;
+
+	list = kcalloc(num_regs, sizeof(struct __guc_mmio_reg_descr), GFP_KERNEL);
+	if (!list)
+		return -ENOMEM;
+
+	newlist->extlist = list;
+	newlist->num_regs = num_regs;
+	newlist->owner = rootlist->owner;
+	newlist->engine = rootlist->engine;
+	newlist->type = rootlist->type;
+
+	return 0;
+}
+
+static void
+guc_capture_alloc_steered_lists_xe_lpd(struct intel_guc *guc,
+				       const struct __guc_mmio_reg_descr_group *lists)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	int slice, subslice, iter, i, num_steer_regs, num_tot_regs = 0;
+	const struct __guc_mmio_reg_descr_group *list;
+	struct __guc_mmio_reg_descr_group *extlists;
+	struct __guc_mmio_reg_descr *extarray;
+	struct sseu_dev_info *sseu;
+
+	/* In XE_LPD we only have steered registers for the render-class */
+	list = guc_capture_get_one_list(lists, GUC_CAPTURE_LIST_INDEX_PF,
+					GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS, GUC_RENDER_CLASS);
+	/* skip if extlists was previously allocated */
+	if (!list || guc->capture->extlists)
+		return;
+
+	num_steer_regs = ARRAY_SIZE(xe_extregs);
+
+	sseu = &gt->info.sseu;
+	for_each_ss_steering(iter, gt, slice, subslice)
+		num_tot_regs += num_steer_regs;
+
+	if (!num_tot_regs)
+		return;
+
+	/* allocate an extra for an end marker */
+	extlists = kcalloc(2, sizeof(struct __guc_mmio_reg_descr_group), GFP_KERNEL);
+	if (!extlists)
+		return;
+
+	if (__alloc_ext_regs(&extlists[0], list, num_tot_regs)) {
+		kfree(extlists);
+		return;
+	}
+
+	extarray = extlists[0].extlist;
+	for_each_ss_steering(iter, gt, slice, subslice) {
+		for (i = 0; i < num_steer_regs; ++i) {
+			__fill_ext_reg(extarray, &xe_extregs[i], slice, subslice);
+			++extarray;
+		}
+	}
+
+	guc->capture->extlists = extlists;
+}
+
+static const struct __ext_steer_reg xehpg_extregs[] = {
+	{"XEHPG_INSTDONE_GEOM_SVG", XEHPG_INSTDONE_GEOM_SVG}
+};
+
+static bool __has_xehpg_extregs(u32 ipver)
+{
+	return (ipver >= IP_VER(12, 55));
+}
+
+static void
+guc_capture_alloc_steered_lists_xe_hpg(struct intel_guc *guc,
+				       const struct __guc_mmio_reg_descr_group *lists,
+				       u32 ipver)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	struct sseu_dev_info *sseu;
+	int slice, subslice, i, iter, num_steer_regs, num_tot_regs = 0;
+	const struct __guc_mmio_reg_descr_group *list;
+	struct __guc_mmio_reg_descr_group *extlists;
+	struct __guc_mmio_reg_descr *extarray;
+
+	/* In XE_LP / HPG we only have render-class steering registers during error-capture */
+	list = guc_capture_get_one_list(lists, GUC_CAPTURE_LIST_INDEX_PF,
+					GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS, GUC_RENDER_CLASS);
+	/* skip if extlists was previously allocated */
+	if (!list || guc->capture->extlists)
+		return;
+
+	num_steer_regs = ARRAY_SIZE(xe_extregs);
+	if (__has_xehpg_extregs(ipver))
+		num_steer_regs += ARRAY_SIZE(xehpg_extregs);
+
+	sseu = &gt->info.sseu;
+	for_each_ss_steering(iter, gt, slice, subslice)
+		num_tot_regs += num_steer_regs;
+
+	if (!num_tot_regs)
+		return;
+
+	/* allocate an extra for an end marker */
+	extlists = kcalloc(2, sizeof(struct __guc_mmio_reg_descr_group), GFP_KERNEL);
+	if (!extlists)
+		return;
+
+	if (__alloc_ext_regs(&extlists[0], list, num_tot_regs)) {
+		kfree(extlists);
+		return;
+	}
+
+	extarray = extlists[0].extlist;
+	for_each_ss_steering(iter, gt, slice, subslice) {
+		for (i = 0; i < ARRAY_SIZE(xe_extregs); ++i) {
+			__fill_ext_reg(extarray, &xe_extregs[i], slice, subslice);
+			++extarray;
+		}
+		if (__has_xehpg_extregs(ipver)) {
+			for (i = 0; i < ARRAY_SIZE(xehpg_extregs); ++i) {
+				__fill_ext_reg(extarray, &xehpg_extregs[i], slice, subslice);
+				++extarray;
+			}
+		}
+	}
+
+	drm_dbg(&i915->drm, "GuC-capture found %d-ext-regs.\n", num_tot_regs);
+	guc->capture->extlists = extlists;
+}
+
+static const struct __guc_mmio_reg_descr_group *
+guc_capture_get_device_reglist(struct intel_guc *guc)
+{
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+
+	if (GRAPHICS_VER(i915) > 11) {
+		/*
+		 * For certain engine classes, there are slice and subslice
+		 * level registers requiring steering. We allocate and populate
+		 * these at init time based on hw config add it as an extension
+		 * list at the end of the pre-populated render list.
+		 */
+		if (IS_DG2(i915))
+			guc_capture_alloc_steered_lists_xe_hpg(guc, xe_lpd_lists, IP_VER(12, 55));
+		else if (IS_XEHPSDV(i915))
+			guc_capture_alloc_steered_lists_xe_hpg(guc, xe_lpd_lists, IP_VER(12, 50));
+		else
+			guc_capture_alloc_steered_lists_xe_lpd(guc, xe_lpd_lists);
+
+		return xe_lpd_lists;
+	}
+
+	/* if GuC submission is enabled on a non-POR platform, just use a common baseline */
+	return default_lists;
+}
+
+static const char *
+__stringify_type(u32 type)
+{
+	switch (type) {
+	case GUC_CAPTURE_LIST_TYPE_GLOBAL:
+		return "Global";
+	case GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS:
+		return "Class";
+	case GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE:
+		return "Instance";
+	default:
+		break;
+	}
+
+	return "unknown";
+}
+
+static const char *
+__stringify_engclass(u32 class)
+{
+	switch (class) {
+	case GUC_RENDER_CLASS:
+		return "Render";
+	case GUC_VIDEO_CLASS:
+		return "Video";
+	case GUC_VIDEOENHANCE_CLASS:
+		return "VideoEnhance";
+	case GUC_BLITTER_CLASS:
+		return "Blitter";
+	case GUC_COMPUTE_CLASS:
+		return "Compute";
+	default:
+		break;
+	}
+
+	return "unknown";
+}
+
+static int
+guc_capture_list_init(struct intel_guc *guc, u32 owner, u32 type, u32 classid,
+		      struct guc_mmio_reg *ptr, u16 num_entries)
+{
+	u32 i = 0, j = 0;
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	const struct __guc_mmio_reg_descr_group *reglists = guc->capture->reglists;
+	struct __guc_mmio_reg_descr_group *extlists = guc->capture->extlists;
+	const struct __guc_mmio_reg_descr_group *match;
+	struct __guc_mmio_reg_descr_group *matchext;
+
+	if (!reglists)
+		return -ENODEV;
+
+	match = guc_capture_get_one_list(reglists, owner, type, classid);
+	if (!match)
+		return -ENODATA;
+
+	for (i = 0; i < num_entries && i < match->num_regs; ++i) {
+		ptr[i].offset = match->list[i].reg.reg;
+		ptr[i].value = 0xDEADF00D;
+		ptr[i].flags = match->list[i].flags;
+		ptr[i].mask = match->list[i].mask;
+	}
+
+	matchext = guc_capture_get_one_ext_list(extlists, owner, type, classid);
+	if (matchext) {
+		for (i = match->num_regs, j = 0; i < num_entries &&
+		     i < (match->num_regs + matchext->num_regs) &&
+			j < matchext->num_regs; ++i, ++j) {
+			ptr[i].offset = matchext->extlist[j].reg.reg;
+			ptr[i].value = 0xDEADF00D;
+			ptr[i].flags = matchext->extlist[j].flags;
+			ptr[i].mask = matchext->extlist[j].mask;
+		}
+	}
+	if (i < num_entries)
+		drm_dbg(&i915->drm, "GuC-capture: Init reglist short %d out %d.\n",
+			(int)i, (int)num_entries);
+
+	return 0;
+}
+
+static int
+guc_cap_list_num_regs(struct intel_guc_state_capture *gc, u32 owner, u32 type, u32 classid)
+{
+	const struct __guc_mmio_reg_descr_group *match;
+	struct __guc_mmio_reg_descr_group *matchext;
+	int num_regs;
+
+	match = guc_capture_get_one_list(gc->reglists, owner, type, classid);
+	if (!match)
+		return 0;
+
+	num_regs = match->num_regs;
+
+	matchext = guc_capture_get_one_ext_list(gc->extlists, owner, type, classid);
+	if (matchext)
+		num_regs += matchext->num_regs;
+
+	return num_regs;
+}
+
+static int
+guc_capture_getlistsize(struct intel_guc *guc, u32 owner, u32 type, u32 classid,
+			size_t *size, bool is_purpose_est)
+{
+	struct intel_guc_state_capture *gc = guc->capture;
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	struct __guc_capture_ads_cache *cache = &gc->ads_cache[owner][type][classid];
+	int num_regs;
+
+	if (!gc->reglists) {
+		drm_warn(&i915->drm, "GuC-capture: No reglist on this device\n");
+		return -ENODEV;
+	}
+
+	if (cache->is_valid) {
+		*size = cache->size;
+		return cache->status;
+	}
+
+	if (!is_purpose_est && owner == GUC_CAPTURE_LIST_INDEX_PF &&
+	    !guc_capture_get_one_list(gc->reglists, owner, type, classid)) {
+		if (type == GUC_CAPTURE_LIST_TYPE_GLOBAL)
+			drm_warn(&i915->drm, "Missing GuC-Err-Cap reglist Global!\n");
+		else
+			drm_warn(&i915->drm, "Missing GuC-Err-Cap reglist %s(%u):%s(%u)!\n",
+				 __stringify_type(type), type,
+				 __stringify_engclass(classid), classid);
+		return -ENODATA;
+	}
+
+	num_regs = guc_cap_list_num_regs(gc, owner, type, classid);
+	/* intentional empty lists can exist depending on hw config */
+	if (!num_regs)
+		return -ENODATA;
+
+	if (size)
+		*size = PAGE_ALIGN((sizeof(struct guc_debug_capture_list)) +
+				   (num_regs * sizeof(struct guc_mmio_reg)));
+
+	return 0;
+}
+
+int
+intel_guc_capture_getlistsize(struct intel_guc *guc, u32 owner, u32 type, u32 classid,
+			      size_t *size)
+{
+	return guc_capture_getlistsize(guc, owner, type, classid, size, false);
+}
+
+static void guc_capture_create_prealloc_nodes(struct intel_guc *guc);
+
+int
+intel_guc_capture_getlist(struct intel_guc *guc, u32 owner, u32 type, u32 classid,
+			  void **outptr)
+{
+	struct intel_guc_state_capture *gc = guc->capture;
+	struct __guc_capture_ads_cache *cache = &gc->ads_cache[owner][type][classid];
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	struct guc_debug_capture_list *listnode;
+	int ret, num_regs;
+	u8 *caplist, *tmp;
+	size_t size = 0;
+
+	if (!gc->reglists)
+		return -ENODEV;
+
+	if (cache->is_valid) {
+		*outptr = cache->ptr;
+		return cache->status;
+	}
+
+	/*
+	 * ADS population of input registers is a good
+	 * time to pre-allocate cachelist output nodes
+	 */
+	guc_capture_create_prealloc_nodes(guc);
+
+	ret = intel_guc_capture_getlistsize(guc, owner, type, classid, &size);
+	if (ret) {
+		cache->is_valid = true;
+		cache->ptr = NULL;
+		cache->size = 0;
+		cache->status = ret;
+		return ret;
+	}
+
+	caplist = kzalloc(size, GFP_KERNEL);
+	if (!caplist) {
+		drm_dbg(&i915->drm, "GuC-capture: failed to alloc cached caplist");
+		return -ENOMEM;
+	}
+
+	/* populate capture list header */
+	tmp = caplist;
+	num_regs = guc_cap_list_num_regs(guc->capture, owner, type, classid);
+	listnode = (struct guc_debug_capture_list *)tmp;
+	listnode->header.info = FIELD_PREP(GUC_CAPTURELISTHDR_NUMDESCR, (u32)num_regs);
+
+	/* populate list of register descriptor */
+	tmp += sizeof(struct guc_debug_capture_list);
+	guc_capture_list_init(guc, owner, type, classid, (struct guc_mmio_reg *)tmp, num_regs);
+
+	/* cache this list */
+	cache->is_valid = true;
+	cache->ptr = caplist;
+	cache->size = size;
+	cache->status = 0;
+
+	*outptr = caplist;
+
+	return 0;
+}
+
+int
+intel_guc_capture_getnullheader(struct intel_guc *guc,
+				void **outptr, size_t *size)
+{
+	struct intel_guc_state_capture *gc = guc->capture;
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	int tmp = sizeof(u32) * 4;
+	void *null_header;
+
+	if (gc->ads_null_cache) {
+		*outptr = gc->ads_null_cache;
+		*size = tmp;
+		return 0;
+	}
+
+	null_header = kzalloc(tmp, GFP_KERNEL);
+	if (!null_header) {
+		drm_dbg(&i915->drm, "GuC-capture: failed to alloc cached nulllist");
+		return -ENOMEM;
+	}
+
+	gc->ads_null_cache = null_header;
+	*outptr = null_header;
+	*size = tmp;
+
+	return 0;
+}
+
+static int
+guc_capture_output_min_size_est(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int worst_min_size = 0;
+	size_t tmp = 0;
+
+	if (!guc->capture)
+		return -ENODEV;
+
+	/*
+	 * If every single engine-instance suffered a failure in quick succession but
+	 * were all unrelated, then a burst of multiple error-capture events would dump
+	 * registers for every one engine instance, one at a time. In this case, GuC
+	 * would even dump the global-registers repeatedly.
+	 *
+	 * For each engine instance, there would be 1 x guc_state_capture_group_t output
+	 * followed by 3 x guc_state_capture_t lists. The latter is how the register
+	 * dumps are split across different register types (where the '3' are global vs class
+	 * vs instance).
+	 */
+	for_each_engine(engine, gt, id) {
+		worst_min_size += sizeof(struct guc_state_capture_group_header_t) +
+					 (3 * sizeof(struct guc_state_capture_header_t));
+
+		if (!guc_capture_getlistsize(guc, 0, GUC_CAPTURE_LIST_TYPE_GLOBAL, 0, &tmp, true))
+			worst_min_size += tmp;
+
+		if (!guc_capture_getlistsize(guc, 0, GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS,
+					     engine->class, &tmp, true)) {
+			worst_min_size += tmp;
+		}
+		if (!guc_capture_getlistsize(guc, 0, GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE,
+					     engine->class, &tmp, true)) {
+			worst_min_size += tmp;
+		}
+	}
+
+	return worst_min_size;
+}
+
+/*
+ * Add on a 3x multiplier to allow for multiple back-to-back captures occurring
+ * before the i915 can read the data out and process it
+ */
+#define GUC_CAPTURE_OVERBUFFER_MULTIPLIER 3
+
+static void check_guc_capture_size(struct intel_guc *guc)
+{
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	int min_size = guc_capture_output_min_size_est(guc);
+	int spare_size = min_size * GUC_CAPTURE_OVERBUFFER_MULTIPLIER;
+	u32 buffer_size = intel_guc_log_section_size_capture(&guc->log);
+
+	/*
+	 * NOTE: min_size is much smaller than the capture region allocation (DG2: <80K vs 1MB)
+	 * Additionally, its based on space needed to fit all engines getting reset at once
+	 * within the same G2H handler task slot. This is very unlikely. However, if GuC really
+	 * does run out of space for whatever reason, we will see an separate warning message
+	 * when processing the G2H event capture-notification, search for:
+	 * INTEL_GUC_STATE_CAPTURE_EVENT_STATUS_NOSPACE.
+	 */
+	if (min_size < 0)
+		drm_warn(&i915->drm, "Failed to calculate GuC error state capture buffer minimum size: %d!\n",
+			 min_size);
+	else if (min_size > buffer_size)
+		drm_warn(&i915->drm, "GuC error state capture buffer maybe small: %d < %d\n",
+			 buffer_size, min_size);
+	else if (spare_size > buffer_size)
+		drm_dbg(&i915->drm, "GuC error state capture buffer lacks spare size: %d < %d (min = %d)\n",
+			buffer_size, spare_size, min_size);
+}
+
+/*
+ * KMD Init time flows:
+ * --------------------
+ *     --> alloc A: GuC input capture regs lists (registered to GuC via ADS).
+ *                  intel_guc_ads acquires the register lists by calling
+ *                  intel_guc_capture_list_size and intel_guc_capture_list_get 'n' times,
+ *                  where n = 1 for global-reg-list +
+ *                            num_engine_classes for class-reg-list +
+ *                            num_engine_classes for instance-reg-list
+ *                               (since all instances of the same engine-class type
+ *                                have an identical engine-instance register-list).
+ *                  ADS module also calls separately for PF vs VF.
+ *
+ *     --> alloc B: GuC output capture buf (registered via guc_init_params(log_param))
+ *                  Size = #define CAPTURE_BUFFER_SIZE (warns if on too-small)
+ *                  Note2: 'x 3' to hold multiple capture groups
+ *
+ * GUC Runtime notify capture:
+ * --------------------------
+ *     --> G2H STATE_CAPTURE_NOTIFICATION
+ *                   L--> intel_guc_capture_process
+ *                           L--> Loop through B (head..tail) and for each engine instance's
+ *                                err-state-captured register-list we find, we alloc 'C':
+ *      --> alloc C: A capture-output-node structure that includes misc capture info along
+ *                   with 3 register list dumps (global, engine-class and engine-instance)
+ *                   This node is created from a pre-allocated list of blank nodes in
+ *                   guc->capture->cachelist and populated with the error-capture
+ *                   data from GuC and then it's added into guc->capture->outlist linked
+ *                   list. This list is used for matchup and printout by i915_gpu_coredump
+ *                   and err_print_gt, (when user invokes the error capture sysfs).
+ *
+ * GUC --> notify context reset:
+ * -----------------------------
+ *     --> G2H CONTEXT RESET
+ *                   L--> guc_handle_context_reset --> i915_capture_error_state
+ *                          L--> i915_gpu_coredump(..IS_GUC_CAPTURE) --> gt_record_engines
+ *                               --> capture_engine(..IS_GUC_CAPTURE)
+ *                               L--> intel_guc_capture_get_matching_node is where
+ *                                    detach C from internal linked list and add it into
+ *                                    intel_engine_coredump struct (if the context and
+ *                                    engine of the event notification matches a node
+ *                                    in the link list).
+ *
+ * User Sysfs / Debugfs
+ * --------------------
+ *      --> i915_gpu_coredump_copy_to_buffer->
+ *                   L--> err_print_to_sgl --> err_print_gt
+ *                        L--> error_print_guc_captures
+ *                             L--> intel_guc_capture_print_node prints the
+ *                                  register lists values of the attached node
+ *                                  on the error-engine-dump being reported.
+ *                   L--> i915_reset_error_state ... -->__i915_gpu_coredump_free
+ *                        L--> ... cleanup_gt -->
+ *                             L--> intel_guc_capture_free_node returns the
+ *                                  capture-output-node back to the internal
+ *                                  cachelist for reuse.
+ *
+ */
+
+static int guc_capture_buf_cnt(struct __guc_capture_bufstate *buf)
+{
+	if (buf->wr >= buf->rd)
+		return (buf->wr - buf->rd);
+	return (buf->size - buf->rd) + buf->wr;
+}
+
+static int guc_capture_buf_cnt_to_end(struct __guc_capture_bufstate *buf)
+{
+	if (buf->rd > buf->wr)
+		return (buf->size - buf->rd);
+	return (buf->wr - buf->rd);
+}
+
+/*
+ * GuC's error-capture output is a ring buffer populated in a byte-stream fashion:
+ *
+ * The GuC Log buffer region for error-capture is managed like a ring buffer.
+ * The GuC firmware dumps error capture logs into this ring in a byte-stream flow.
+ * Additionally, as per the current and foreseeable future, all packed error-
+ * capture output structures are dword aligned.
+ *
+ * That said, if the GuC firmware is in the midst of writing a structure that is larger
+ * than one dword but the tail end of the err-capture buffer-region has lesser space left,
+ * we would need to extract that structure one dword at a time straddled across the end,
+ * onto the start of the ring.
+ *
+ * Below function, guc_capture_log_remove_dw is a helper for that. All callers of this
+ * function would typically do a straight-up memcpy from the ring contents and will only
+ * call this helper if their structure-extraction is straddling across the end of the
+ * ring. GuC firmware does not add any padding. The reason for the no-padding is to ease
+ * scalability for future expansion of output data types without requiring a redesign
+ * of the flow controls.
+ */
+static int
+guc_capture_log_remove_dw(struct intel_guc *guc, struct __guc_capture_bufstate *buf,
+			  u32 *dw)
+{
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	int tries = 2;
+	int avail = 0;
+	u32 *src_data;
+
+	if (!guc_capture_buf_cnt(buf))
+		return 0;
+
+	while (tries--) {
+		avail = guc_capture_buf_cnt_to_end(buf);
+		if (avail >= sizeof(u32)) {
+			src_data = (u32 *)(buf->data + buf->rd);
+			*dw = *src_data;
+			buf->rd += 4;
+			return 4;
+		}
+		if (avail)
+			drm_dbg(&i915->drm, "GuC-Cap-Logs not dword aligned, skipping.\n");
+		buf->rd = 0;
+	}
+
+	return 0;
+}
+
+static bool
+guc_capture_data_extracted(struct __guc_capture_bufstate *b,
+			   int size, void *dest)
+{
+	if (guc_capture_buf_cnt_to_end(b) >= size) {
+		memcpy(dest, (b->data + b->rd), size);
+		b->rd += size;
+		return true;
+	}
+	return false;
+}
+
+static int
+guc_capture_log_get_group_hdr(struct intel_guc *guc, struct __guc_capture_bufstate *buf,
+			      struct guc_state_capture_group_header_t *ghdr)
+{
+	int read = 0;
+	int fullsize = sizeof(struct guc_state_capture_group_header_t);
+
+	if (fullsize > guc_capture_buf_cnt(buf))
+		return -1;
+
+	if (guc_capture_data_extracted(buf, fullsize, (void *)ghdr))
+		return 0;
+
+	read += guc_capture_log_remove_dw(guc, buf, &ghdr->owner);
+	read += guc_capture_log_remove_dw(guc, buf, &ghdr->info);
+	if (read != fullsize)
+		return -1;
+
+	return 0;
+}
+
+static int
+guc_capture_log_get_data_hdr(struct intel_guc *guc, struct __guc_capture_bufstate *buf,
+			     struct guc_state_capture_header_t *hdr)
+{
+	int read = 0;
+	int fullsize = sizeof(struct guc_state_capture_header_t);
+
+	if (fullsize > guc_capture_buf_cnt(buf))
+		return -1;
+
+	if (guc_capture_data_extracted(buf, fullsize, (void *)hdr))
+		return 0;
+
+	read += guc_capture_log_remove_dw(guc, buf, &hdr->owner);
+	read += guc_capture_log_remove_dw(guc, buf, &hdr->info);
+	read += guc_capture_log_remove_dw(guc, buf, &hdr->lrca);
+	read += guc_capture_log_remove_dw(guc, buf, &hdr->guc_id);
+	read += guc_capture_log_remove_dw(guc, buf, &hdr->num_mmios);
+	if (read != fullsize)
+		return -1;
+
+	return 0;
+}
+
+static int
+guc_capture_log_get_register(struct intel_guc *guc, struct __guc_capture_bufstate *buf,
+			     struct guc_mmio_reg *reg)
+{
+	int read = 0;
+	int fullsize = sizeof(struct guc_mmio_reg);
+
+	if (fullsize > guc_capture_buf_cnt(buf))
+		return -1;
+
+	if (guc_capture_data_extracted(buf, fullsize, (void *)reg))
+		return 0;
+
+	read += guc_capture_log_remove_dw(guc, buf, &reg->offset);
+	read += guc_capture_log_remove_dw(guc, buf, &reg->value);
+	read += guc_capture_log_remove_dw(guc, buf, &reg->flags);
+	read += guc_capture_log_remove_dw(guc, buf, &reg->mask);
+	if (read != fullsize)
+		return -1;
+
+	return 0;
+}
+
+static void
+guc_capture_delete_one_node(struct intel_guc *guc, struct __guc_capture_parsed_output *node)
+{
+	int i;
+
+	for (i = 0; i < GUC_CAPTURE_LIST_TYPE_MAX; ++i)
+		kfree(node->reginfo[i].regs);
+	list_del(&node->link);
+	kfree(node);
+}
+
+static void
+guc_capture_delete_prealloc_nodes(struct intel_guc *guc)
+{
+	struct __guc_capture_parsed_output *n, *ntmp;
+
+	/*
+	 * NOTE: At the end of driver operation, we must assume that we
+	 * have prealloc nodes in both the cachelist as well as outlist
+	 * if unclaimed error capture events occurred prior to shutdown.
+	 */
+	list_for_each_entry_safe(n, ntmp, &guc->capture->outlist, link)
+		guc_capture_delete_one_node(guc, n);
+
+	list_for_each_entry_safe(n, ntmp, &guc->capture->cachelist, link)
+		guc_capture_delete_one_node(guc, n);
+}
+
+static void
+guc_capture_add_node_to_list(struct __guc_capture_parsed_output *node,
+			     struct list_head *list)
+{
+	list_add_tail(&node->link, list);
+}
+
+static void
+guc_capture_add_node_to_outlist(struct intel_guc_state_capture *gc,
+				struct __guc_capture_parsed_output *node)
+{
+	guc_capture_add_node_to_list(node, &gc->outlist);
+}
+
+static void
+guc_capture_add_node_to_cachelist(struct intel_guc_state_capture *gc,
+				  struct __guc_capture_parsed_output *node)
+{
+	guc_capture_add_node_to_list(node, &gc->cachelist);
+}
+
+static void
+guc_capture_init_node(struct intel_guc *guc, struct __guc_capture_parsed_output *node)
+{
+	struct guc_mmio_reg *tmp[GUC_CAPTURE_LIST_TYPE_MAX];
+	int i;
+
+	for (i = 0; i < GUC_CAPTURE_LIST_TYPE_MAX; ++i) {
+		tmp[i] = node->reginfo[i].regs;
+		memset(tmp[i], 0, sizeof(struct guc_mmio_reg) *
+		       guc->capture->max_mmio_per_node);
+	}
+	memset(node, 0, sizeof(*node));
+	for (i = 0; i < GUC_CAPTURE_LIST_TYPE_MAX; ++i)
+		node->reginfo[i].regs = tmp[i];
+
+	INIT_LIST_HEAD(&node->link);
+}
+
+static struct __guc_capture_parsed_output *
+guc_capture_get_prealloc_node(struct intel_guc *guc)
+{
+	struct __guc_capture_parsed_output *found = NULL;
+
+	if (!list_empty(&guc->capture->cachelist)) {
+		struct __guc_capture_parsed_output *n, *ntmp;
+
+		/* get first avail node from the cache list */
+		list_for_each_entry_safe(n, ntmp, &guc->capture->cachelist, link) {
+			found = n;
+			list_del(&n->link);
+			break;
+		}
+	} else {
+		struct __guc_capture_parsed_output *n, *ntmp;
+
+		/* traverse down and steal back the oldest node already allocated */
+		list_for_each_entry_safe(n, ntmp, &guc->capture->outlist, link) {
+			found = n;
+		}
+		if (found)
+			list_del(&found->link);
+	}
+	if (found)
+		guc_capture_init_node(guc, found);
+
+	return found;
+}
+
+static struct __guc_capture_parsed_output *
+guc_capture_alloc_one_node(struct intel_guc *guc)
+{
+	struct __guc_capture_parsed_output *new;
+	int i;
+
+	new = kzalloc(sizeof(*new), GFP_KERNEL);
+	if (!new)
+		return NULL;
+
+	for (i = 0; i < GUC_CAPTURE_LIST_TYPE_MAX; ++i) {
+		new->reginfo[i].regs = kcalloc(guc->capture->max_mmio_per_node,
+					       sizeof(struct guc_mmio_reg), GFP_KERNEL);
+		if (!new->reginfo[i].regs) {
+			while (i)
+				kfree(new->reginfo[--i].regs);
+			kfree(new);
+			return NULL;
+		}
+	}
+	guc_capture_init_node(guc, new);
+
+	return new;
+}
+
+static struct __guc_capture_parsed_output *
+guc_capture_clone_node(struct intel_guc *guc, struct __guc_capture_parsed_output *original,
+		       u32 keep_reglist_mask)
+{
+	struct __guc_capture_parsed_output *new;
+	int i;
+
+	new = guc_capture_get_prealloc_node(guc);
+	if (!new)
+		return NULL;
+	if (!original)
+		return new;
+
+	new->is_partial = original->is_partial;
+
+	/* copy reg-lists that we want to clone */
+	for (i = 0; i < GUC_CAPTURE_LIST_TYPE_MAX; ++i) {
+		if (keep_reglist_mask & BIT(i)) {
+			GEM_BUG_ON(original->reginfo[i].num_regs  >
+				   guc->capture->max_mmio_per_node);
+
+			memcpy(new->reginfo[i].regs, original->reginfo[i].regs,
+			       original->reginfo[i].num_regs * sizeof(struct guc_mmio_reg));
+
+			new->reginfo[i].num_regs = original->reginfo[i].num_regs;
+			new->reginfo[i].vfid  = original->reginfo[i].vfid;
+
+			if (i == GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS) {
+				new->eng_class = original->eng_class;
+			} else if (i == GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE) {
+				new->eng_inst = original->eng_inst;
+				new->guc_id = original->guc_id;
+				new->lrca = original->lrca;
+			}
+		}
+	}
+
+	return new;
+}
+
+static void
+__guc_capture_create_prealloc_nodes(struct intel_guc *guc)
+{
+	struct __guc_capture_parsed_output *node = NULL;
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	int i;
+
+	for (i = 0; i < PREALLOC_NODES_MAX_COUNT; ++i) {
+		node = guc_capture_alloc_one_node(guc);
+		if (!node) {
+			drm_warn(&i915->drm, "GuC Capture pre-alloc-cache failure\n");
+			/* dont free the priors, use what we got and cleanup at shutdown */
+			return;
+		}
+		guc_capture_add_node_to_cachelist(guc->capture, node);
+	}
+}
+
+static int
+guc_get_max_reglist_count(struct intel_guc *guc)
+{
+	int i, j, k, tmp, maxregcount = 0;
+
+	for (i = 0; i < GUC_CAPTURE_LIST_INDEX_MAX; ++i) {
+		for (j = 0; j < GUC_CAPTURE_LIST_TYPE_MAX; ++j) {
+			for (k = 0; k < GUC_MAX_ENGINE_CLASSES; ++k) {
+				if (j == GUC_CAPTURE_LIST_TYPE_GLOBAL && k > 0)
+					continue;
+
+				tmp = guc_cap_list_num_regs(guc->capture, i, j, k);
+				if (tmp > maxregcount)
+					maxregcount = tmp;
+			}
+		}
+	}
+	if (!maxregcount)
+		maxregcount = PREALLOC_NODES_DEFAULT_NUMREGS;
+
+	return maxregcount;
+}
+
+static void
+guc_capture_create_prealloc_nodes(struct intel_guc *guc)
+{
+	/* skip if we've already done the pre-alloc */
+	if (guc->capture->max_mmio_per_node)
+		return;
+
+	guc->capture->max_mmio_per_node = guc_get_max_reglist_count(guc);
+	__guc_capture_create_prealloc_nodes(guc);
+}
+
+static int
+guc_capture_extract_reglists(struct intel_guc *guc, struct __guc_capture_bufstate *buf)
+{
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	struct guc_state_capture_group_header_t ghdr = {0};
+	struct guc_state_capture_header_t hdr = {0};
+	struct __guc_capture_parsed_output *node = NULL;
+	struct guc_mmio_reg *regs = NULL;
+	int i, numlists, numregs, ret = 0;
+	enum guc_capture_type datatype;
+	struct guc_mmio_reg tmp;
+	bool is_partial = false;
+
+	i = guc_capture_buf_cnt(buf);
+	if (!i)
+		return -ENODATA;
+	if (i % sizeof(u32)) {
+		drm_warn(&i915->drm, "GuC Capture new entries unaligned\n");
+		ret = -EIO;
+		goto bailout;
+	}
+
+	/* first get the capture group header */
+	if (guc_capture_log_get_group_hdr(guc, buf, &ghdr)) {
+		ret = -EIO;
+		goto bailout;
+	}
+	/*
+	 * we would typically expect a layout as below where n would be expected to be
+	 * anywhere between 3 to n where n > 3 if we are seeing multiple dependent engine
+	 * instances being reset together.
+	 * ____________________________________________
+	 * | Capture Group                            |
+	 * | ________________________________________ |
+	 * | | Capture Group Header:                | |
+	 * | |  - num_captures = 5                  | |
+	 * | |______________________________________| |
+	 * | ________________________________________ |
+	 * | | Capture1:                            | |
+	 * | |  Hdr: GLOBAL, numregs=a              | |
+	 * | | ____________________________________ | |
+	 * | | | Reglist                          | | |
+	 * | | | - reg1, reg2, ... rega           | | |
+	 * | | |__________________________________| | |
+	 * | |______________________________________| |
+	 * | ________________________________________ |
+	 * | | Capture2:                            | |
+	 * | |  Hdr: CLASS=RENDER/COMPUTE, numregs=b| |
+	 * | | ____________________________________ | |
+	 * | | | Reglist                          | | |
+	 * | | | - reg1, reg2, ... regb           | | |
+	 * | | |__________________________________| | |
+	 * | |______________________________________| |
+	 * | ________________________________________ |
+	 * | | Capture3:                            | |
+	 * | |  Hdr: INSTANCE=RCS, numregs=c        | |
+	 * | | ____________________________________ | |
+	 * | | | Reglist                          | | |
+	 * | | | - reg1, reg2, ... regc           | | |
+	 * | | |__________________________________| | |
+	 * | |______________________________________| |
+	 * | ________________________________________ |
+	 * | | Capture4:                            | |
+	 * | |  Hdr: CLASS=RENDER/COMPUTE, numregs=d| |
+	 * | | ____________________________________ | |
+	 * | | | Reglist                          | | |
+	 * | | | - reg1, reg2, ... regd           | | |
+	 * | | |__________________________________| | |
+	 * | |______________________________________| |
+	 * | ________________________________________ |
+	 * | | Capture5:                            | |
+	 * | |  Hdr: INSTANCE=CCS0, numregs=e       | |
+	 * | | ____________________________________ | |
+	 * | | | Reglist                          | | |
+	 * | | | - reg1, reg2, ... rege           | | |
+	 * | | |__________________________________| | |
+	 * | |______________________________________| |
+	 * |__________________________________________|
+	 */
+	is_partial = FIELD_GET(CAP_GRP_HDR_CAPTURE_TYPE, ghdr.info);
+	numlists = FIELD_GET(CAP_GRP_HDR_NUM_CAPTURES, ghdr.info);
+
+	while (numlists--) {
+		if (guc_capture_log_get_data_hdr(guc, buf, &hdr)) {
+			ret = -EIO;
+			break;
+		}
+
+		datatype = FIELD_GET(CAP_HDR_CAPTURE_TYPE, hdr.info);
+		if (datatype > GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE) {
+			/* unknown capture type - skip over to next capture set */
+			numregs = FIELD_GET(CAP_HDR_NUM_MMIOS, hdr.num_mmios);
+			while (numregs--) {
+				if (guc_capture_log_get_register(guc, buf, &tmp)) {
+					ret = -EIO;
+					break;
+				}
+			}
+			continue;
+		} else if (node) {
+			/*
+			 * Based on the current capture type and what we have so far,
+			 * decide if we should add the current node into the internal
+			 * linked list for match-up when i915_gpu_coredump calls later
+			 * (and alloc a blank node for the next set of reglists)
+			 * or continue with the same node or clone the current node
+			 * but only retain the global or class registers (such as the
+			 * case of dependent engine resets).
+			 */
+			if (datatype == GUC_CAPTURE_LIST_TYPE_GLOBAL) {
+				guc_capture_add_node_to_outlist(guc->capture, node);
+				node = NULL;
+			} else if (datatype == GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS &&
+				   node->reginfo[GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS].num_regs) {
+				/* Add to list, clone node and duplicate global list */
+				guc_capture_add_node_to_outlist(guc->capture, node);
+				node = guc_capture_clone_node(guc, node,
+							      GCAP_PARSED_REGLIST_INDEX_GLOBAL);
+			} else if (datatype == GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE &&
+				   node->reginfo[GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE].num_regs) {
+				/* Add to list, clone node and duplicate global + class lists */
+				guc_capture_add_node_to_outlist(guc->capture, node);
+				node = guc_capture_clone_node(guc, node,
+							      (GCAP_PARSED_REGLIST_INDEX_GLOBAL |
+							      GCAP_PARSED_REGLIST_INDEX_ENGCLASS));
+			}
+		}
+
+		if (!node) {
+			node = guc_capture_get_prealloc_node(guc);
+			if (!node) {
+				ret = -ENOMEM;
+				break;
+			}
+			if (datatype != GUC_CAPTURE_LIST_TYPE_GLOBAL)
+				drm_dbg(&i915->drm, "GuC Capture missing global dump: %08x!\n",
+					datatype);
+		}
+		node->is_partial = is_partial;
+		node->reginfo[datatype].vfid = FIELD_GET(CAP_HDR_CAPTURE_VFID, hdr.owner);
+		switch (datatype) {
+		case GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE:
+			node->eng_class = FIELD_GET(CAP_HDR_ENGINE_CLASS, hdr.info);
+			node->eng_inst = FIELD_GET(CAP_HDR_ENGINE_INSTANCE, hdr.info);
+			node->lrca = hdr.lrca;
+			node->guc_id = hdr.guc_id;
+			break;
+		case GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS:
+			node->eng_class = FIELD_GET(CAP_HDR_ENGINE_CLASS, hdr.info);
+			break;
+		default:
+			break;
+		}
+
+		numregs = FIELD_GET(CAP_HDR_NUM_MMIOS, hdr.num_mmios);
+		if (numregs > guc->capture->max_mmio_per_node) {
+			drm_dbg(&i915->drm, "GuC Capture list extraction clipped by prealloc!\n");
+			numregs = guc->capture->max_mmio_per_node;
+		}
+		node->reginfo[datatype].num_regs = numregs;
+		regs = node->reginfo[datatype].regs;
+		i = 0;
+		while (numregs--) {
+			if (guc_capture_log_get_register(guc, buf, &regs[i++])) {
+				ret = -EIO;
+				break;
+			}
+		}
+	}
+
+bailout:
+	if (node) {
+		/* If we have data, add to linked list for match-up when i915_gpu_coredump calls */
+		for (i = GUC_CAPTURE_LIST_TYPE_GLOBAL; i < GUC_CAPTURE_LIST_TYPE_MAX; ++i) {
+			if (node->reginfo[i].regs) {
+				guc_capture_add_node_to_outlist(guc->capture, node);
+				node = NULL;
+				break;
+			}
+		}
+		if (node) /* else return it back to cache list */
+			guc_capture_add_node_to_cachelist(guc->capture, node);
+	}
+	return ret;
+}
+
+static int __guc_capture_flushlog_complete(struct intel_guc *guc)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_LOG_BUFFER_FILE_FLUSH_COMPLETE,
+		GUC_CAPTURE_LOG_BUFFER
+	};
+
+	return intel_guc_send_nb(guc, action, ARRAY_SIZE(action), 0);
+
+}
+
+static void __guc_capture_process_output(struct intel_guc *guc)
+{
+	unsigned int buffer_size, read_offset, write_offset, full_count;
+	struct intel_uc *uc = container_of(guc, typeof(*uc), guc);
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	struct guc_log_buffer_state log_buf_state_local;
+	struct guc_log_buffer_state *log_buf_state;
+	struct __guc_capture_bufstate buf;
+	void *src_data = NULL;
+	bool new_overflow;
+	int ret;
+
+	log_buf_state = guc->log.buf_addr +
+			(sizeof(struct guc_log_buffer_state) * GUC_CAPTURE_LOG_BUFFER);
+	src_data = guc->log.buf_addr +
+		   intel_guc_get_log_buffer_offset(&guc->log, GUC_CAPTURE_LOG_BUFFER);
+
+	/*
+	 * Make a copy of the state structure, inside GuC log buffer
+	 * (which is uncached mapped), on the stack to avoid reading
+	 * from it multiple times.
+	 */
+	memcpy(&log_buf_state_local, log_buf_state, sizeof(struct guc_log_buffer_state));
+	buffer_size = intel_guc_get_log_buffer_size(&guc->log, GUC_CAPTURE_LOG_BUFFER);
+	read_offset = log_buf_state_local.read_ptr;
+	write_offset = log_buf_state_local.sampled_write_ptr;
+	full_count = log_buf_state_local.buffer_full_cnt;
+
+	/* Bookkeeping stuff */
+	guc->log.stats[GUC_CAPTURE_LOG_BUFFER].flush += log_buf_state_local.flush_to_file;
+	new_overflow = intel_guc_check_log_buf_overflow(&guc->log, GUC_CAPTURE_LOG_BUFFER,
+							full_count);
+
+	/* Now copy the actual logs. */
+	if (unlikely(new_overflow)) {
+		/* copy the whole buffer in case of overflow */
+		read_offset = 0;
+		write_offset = buffer_size;
+	} else if (unlikely((read_offset > buffer_size) ||
+			(write_offset > buffer_size))) {
+		drm_err(&i915->drm, "invalid GuC log capture buffer state!\n");
+		/* copy whole buffer as offsets are unreliable */
+		read_offset = 0;
+		write_offset = buffer_size;
+	}
+
+	buf.size = buffer_size;
+	buf.rd = read_offset;
+	buf.wr = write_offset;
+	buf.data = src_data;
+
+	if (!uc->reset_in_progress) {
+		do {
+			ret = guc_capture_extract_reglists(guc, &buf);
+		} while (ret >= 0);
+	}
+
+	/* Update the state of log buffer err-cap state */
+	log_buf_state->read_ptr = write_offset;
+	log_buf_state->flush_to_file = 0;
+	__guc_capture_flushlog_complete(guc);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+
+static const char *
+guc_capture_reg_to_str(const struct intel_guc *guc, u32 owner, u32 type,
+		       u32 class, u32 id, u32 offset, u32 *is_ext)
+{
+	const struct __guc_mmio_reg_descr_group *reglists = guc->capture->reglists;
+	struct __guc_mmio_reg_descr_group *extlists = guc->capture->extlists;
+	const struct __guc_mmio_reg_descr_group *match;
+	struct __guc_mmio_reg_descr_group *matchext;
+	int j;
+
+	*is_ext = 0;
+	if (!reglists)
+		return NULL;
+
+	match = guc_capture_get_one_list(reglists, owner, type, id);
+	if (!match)
+		return NULL;
+
+	for (j = 0; j < match->num_regs; ++j) {
+		if (offset == match->list[j].reg.reg)
+			return match->list[j].regname;
+	}
+	if (extlists) {
+		matchext = guc_capture_get_one_ext_list(extlists, owner, type, id);
+		if (!matchext)
+			return NULL;
+		for (j = 0; j < matchext->num_regs; ++j) {
+			if (offset == matchext->extlist[j].reg.reg) {
+				*is_ext = 1;
+				return matchext->extlist[j].regname;
+			}
+		}
+	}
+
+	return NULL;
+}
+
+#define GCAP_PRINT_INTEL_ENG_INFO(ebuf, eng) \
+	do { \
+		i915_error_printf(ebuf, "    i915-Eng-Name: %s command stream\n", \
+				  (eng)->name); \
+		i915_error_printf(ebuf, "    i915-Eng-Inst-Class: 0x%02x\n", (eng)->class); \
+		i915_error_printf(ebuf, "    i915-Eng-Inst-Id: 0x%02x\n", (eng)->instance); \
+		i915_error_printf(ebuf, "    i915-Eng-LogicalMask: 0x%08x\n", \
+				  (eng)->logical_mask); \
+	} while (0)
+
+#define GCAP_PRINT_GUC_INST_INFO(ebuf, node) \
+	do { \
+		i915_error_printf(ebuf, "    GuC-Engine-Inst-Id: 0x%08x\n", \
+				  (node)->eng_inst); \
+		i915_error_printf(ebuf, "    GuC-Context-Id: 0x%08x\n", (node)->guc_id); \
+		i915_error_printf(ebuf, "    LRCA: 0x%08x\n", (node)->lrca); \
+	} while (0)
+
+int intel_guc_capture_print_engine_node(struct drm_i915_error_state_buf *ebuf,
+					const struct intel_engine_coredump *ee)
+{
+	const char *grptype[GUC_STATE_CAPTURE_GROUP_TYPE_MAX] = {
+		"full-capture",
+		"partial-capture"
+	};
+	const char *datatype[GUC_CAPTURE_LIST_TYPE_MAX] = {
+		"Global",
+		"Engine-Class",
+		"Engine-Instance"
+	};
+	struct intel_guc_state_capture *cap;
+	struct __guc_capture_parsed_output *node;
+	struct intel_engine_cs *eng;
+	struct guc_mmio_reg *regs;
+	struct intel_guc *guc;
+	const char *str;
+	int numregs, i, j;
+	u32 is_ext;
+
+	if (!ebuf || !ee)
+		return -EINVAL;
+	cap = ee->guc_capture;
+	if (!cap || !ee->engine)
+		return -ENODEV;
+
+	guc = &ee->engine->gt->uc.guc;
+
+	i915_error_printf(ebuf, "global --- GuC Error Capture on %s command stream:\n",
+			  ee->engine->name);
+
+	node = ee->guc_capture_node;
+	if (!node) {
+		i915_error_printf(ebuf, "  No matching ee-node\n");
+		return 0;
+	}
+
+	i915_error_printf(ebuf, "Coverage:  %s\n", grptype[node->is_partial]);
+
+	for (i = GUC_CAPTURE_LIST_TYPE_GLOBAL; i < GUC_CAPTURE_LIST_TYPE_MAX; ++i) {
+		i915_error_printf(ebuf, "  RegListType: %s\n",
+				  datatype[i % GUC_CAPTURE_LIST_TYPE_MAX]);
+		i915_error_printf(ebuf, "    Owner-Id: %d\n", node->reginfo[i].vfid);
+
+		switch (i) {
+		case GUC_CAPTURE_LIST_TYPE_GLOBAL:
+		default:
+			break;
+		case GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS:
+			i915_error_printf(ebuf, "    GuC-Eng-Class: %d\n", node->eng_class);
+			i915_error_printf(ebuf, "    i915-Eng-Class: %d\n",
+					  guc_class_to_engine_class(node->eng_class));
+			break;
+		case GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE:
+			eng = intel_guc_lookup_engine(guc, node->eng_class, node->eng_inst);
+			if (eng)
+				GCAP_PRINT_INTEL_ENG_INFO(ebuf, eng);
+			else
+				i915_error_printf(ebuf, "    i915-Eng-Lookup Fail!\n");
+			GCAP_PRINT_GUC_INST_INFO(ebuf, node);
+			break;
+		}
+
+		numregs = node->reginfo[i].num_regs;
+		i915_error_printf(ebuf, "    NumRegs: %d\n", numregs);
+		j = 0;
+		while (numregs--) {
+			regs = node->reginfo[i].regs;
+			str = guc_capture_reg_to_str(guc, GUC_CAPTURE_LIST_INDEX_PF, i,
+						     node->eng_class, 0, regs[j].offset, &is_ext);
+			if (!str)
+				i915_error_printf(ebuf, "      REG-0x%08x", regs[j].offset);
+			else
+				i915_error_printf(ebuf, "      %s", str);
+			if (is_ext)
+				i915_error_printf(ebuf, "[%ld][%ld]",
+					FIELD_GET(GUC_REGSET_STEERING_GROUP, regs[j].flags),
+					FIELD_GET(GUC_REGSET_STEERING_INSTANCE, regs[j].flags));
+			i915_error_printf(ebuf, ":  0x%08x\n", regs[j].value);
+			++j;
+		}
+	}
+	return 0;
+}
+
+#endif //CONFIG_DRM_I915_CAPTURE_ERROR
+
+static void guc_capture_find_ecode(struct intel_engine_coredump *ee)
+{
+	struct gcap_reg_list_info *reginfo;
+	struct guc_mmio_reg *regs;
+	i915_reg_t reg_ipehr = RING_IPEHR(0);
+	i915_reg_t reg_instdone = RING_INSTDONE(0);
+	int i;
+
+	if (!ee->guc_capture_node)
+		return;
+
+	reginfo = ee->guc_capture_node->reginfo + GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE;
+	regs = reginfo->regs;
+	for (i = 0; i < reginfo->num_regs; i++) {
+		if (regs[i].offset == reg_ipehr.reg)
+			ee->ipehr = regs[i].value;
+		else if (regs[i].offset == reg_instdone.reg)
+			ee->instdone.instdone = regs[i].value;
+	}
+}
+
+void intel_guc_capture_free_node(struct intel_engine_coredump *ee)
+{
+	if (!ee || !ee->guc_capture_node)
+		return;
+
+	guc_capture_add_node_to_cachelist(ee->guc_capture, ee->guc_capture_node);
+	ee->guc_capture = NULL;
+	ee->guc_capture_node = NULL;
+}
+
+void intel_guc_capture_get_matching_node(struct intel_gt *gt,
+					 struct intel_engine_coredump *ee,
+					 struct intel_context *ce)
+{
+	struct __guc_capture_parsed_output *n, *ntmp;
+	struct drm_i915_private *i915;
+	struct intel_guc *guc;
+
+	if (!gt || !ee || !ce)
+		return;
+
+	i915 = gt->i915;
+	guc = &gt->uc.guc;
+	if (!guc->capture)
+		return;
+
+	GEM_BUG_ON(ee->guc_capture_node);
+	/*
+	 * Look for a matching GuC reported error capture node from
+	 * the internal output link-list based on lrca, guc-id and engine
+	 * identification.
+	 */
+	list_for_each_entry_safe(n, ntmp, &guc->capture->outlist, link) {
+		if (n->eng_inst == GUC_ID_TO_ENGINE_INSTANCE(ee->engine->guc_id) &&
+		    n->eng_class == GUC_ID_TO_ENGINE_CLASS(ee->engine->guc_id) &&
+		    n->guc_id && n->guc_id == ce->guc_id.id &&
+		    (n->lrca & CTX_GTT_ADDRESS_MASK) && (n->lrca & CTX_GTT_ADDRESS_MASK) ==
+		    (ce->lrc.lrca & CTX_GTT_ADDRESS_MASK)) {
+			list_del(&n->link);
+			ee->guc_capture_node = n;
+			ee->guc_capture = guc->capture;
+			guc_capture_find_ecode(ee);
+			return;
+		}
+	}
+	drm_dbg(&i915->drm, "GuC capture can't match ee to node\n");
+}
+
+void intel_guc_capture_process(struct intel_guc *guc)
+{
+	if (guc->capture)
+		__guc_capture_process_output(guc);
+}
+
+static void
+guc_capture_free_ads_cache(struct intel_guc_state_capture *gc)
+{
+	int i, j, k;
+	struct __guc_capture_ads_cache *cache;
+
+	for (i = 0; i < GUC_CAPTURE_LIST_INDEX_MAX; ++i) {
+		for (j = 0; j < GUC_CAPTURE_LIST_TYPE_MAX; ++j) {
+			for (k = 0; k < GUC_MAX_ENGINE_CLASSES; ++k) {
+				cache = &gc->ads_cache[i][j][k];
+				if (cache->is_valid)
+					kfree(cache->ptr);
+			}
+		}
+	}
+	kfree(gc->ads_null_cache);
+}
+
+void intel_guc_capture_destroy(struct intel_guc *guc)
+{
+	if (!guc->capture)
+		return;
+
+	guc_capture_free_ads_cache(guc->capture);
+
+	guc_capture_delete_prealloc_nodes(guc);
+
+	guc_capture_free_extlists(guc->capture->extlists);
+	kfree(guc->capture->extlists);
+
+	kfree(guc->capture);
+	guc->capture = NULL;
+}
+
+int intel_guc_capture_init(struct intel_guc *guc)
+{
+	guc->capture = kzalloc(sizeof(*guc->capture), GFP_KERNEL);
+	if (!guc->capture)
+		return -ENOMEM;
+
+	guc->capture->reglists = guc_capture_get_device_reglist(guc);
+
+	INIT_LIST_HEAD(&guc->capture->outlist);
+	INIT_LIST_HEAD(&guc->capture->cachelist);
+
+	check_guc_capture_size(guc);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.h
new file mode 100644
index 000000000..fbd3713c7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.h
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021-2021 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_CAPTURE_H
+#define _INTEL_GUC_CAPTURE_H
+
+#include <linux/types.h>
+
+struct drm_i915_error_state_buf;
+struct guc_gt_system_info;
+struct intel_engine_coredump;
+struct intel_context;
+struct intel_gt;
+struct intel_guc;
+
+void intel_guc_capture_free_node(struct intel_engine_coredump *ee);
+int intel_guc_capture_print_engine_node(struct drm_i915_error_state_buf *m,
+					const struct intel_engine_coredump *ee);
+void intel_guc_capture_get_matching_node(struct intel_gt *gt, struct intel_engine_coredump *ee,
+					 struct intel_context *ce);
+void intel_guc_capture_process(struct intel_guc *guc);
+int intel_guc_capture_getlist(struct intel_guc *guc, u32 owner, u32 type, u32 classid,
+			      void **outptr);
+int intel_guc_capture_getlistsize(struct intel_guc *guc, u32 owner, u32 type, u32 classid,
+				  size_t *size);
+int intel_guc_capture_getnullheader(struct intel_guc *guc, void **outptr, size_t *size);
+void intel_guc_capture_destroy(struct intel_guc *guc);
+int intel_guc_capture_init(struct intel_guc *guc);
+
+#endif /* _INTEL_GUC_CAPTURE_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c
new file mode 100644
index 000000000..2b22065e8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c
@@ -0,0 +1,1250 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2016-2019 Intel Corporation
+ */
+
+#include <linux/circ_buf.h>
+#include <linux/ktime.h>
+#include <linux/time64.h>
+#include <linux/string_helpers.h>
+#include <linux/timekeeping.h>
+
+#include "i915_drv.h"
+#include "intel_guc_ct.h"
+#include "gt/intel_gt.h"
+
+static inline struct intel_guc *ct_to_guc(struct intel_guc_ct *ct)
+{
+	return container_of(ct, struct intel_guc, ct);
+}
+
+static inline struct intel_gt *ct_to_gt(struct intel_guc_ct *ct)
+{
+	return guc_to_gt(ct_to_guc(ct));
+}
+
+static inline struct drm_i915_private *ct_to_i915(struct intel_guc_ct *ct)
+{
+	return ct_to_gt(ct)->i915;
+}
+
+static inline struct drm_device *ct_to_drm(struct intel_guc_ct *ct)
+{
+	return &ct_to_i915(ct)->drm;
+}
+
+#define CT_ERROR(_ct, _fmt, ...) \
+	drm_err(ct_to_drm(_ct), "CT: " _fmt, ##__VA_ARGS__)
+#ifdef CONFIG_DRM_I915_DEBUG_GUC
+#define CT_DEBUG(_ct, _fmt, ...) \
+	drm_dbg(ct_to_drm(_ct), "CT: " _fmt, ##__VA_ARGS__)
+#else
+#define CT_DEBUG(...)	do { } while (0)
+#endif
+#define CT_PROBE_ERROR(_ct, _fmt, ...) \
+	i915_probe_error(ct_to_i915(ct), "CT: " _fmt, ##__VA_ARGS__)
+
+/**
+ * DOC: CTB Blob
+ *
+ * We allocate single blob to hold both CTB descriptors and buffers:
+ *
+ *      +--------+-----------------------------------------------+------+
+ *      | offset | contents                                      | size |
+ *      +========+===============================================+======+
+ *      | 0x0000 | H2G `CTB Descriptor`_ (send)                  |      |
+ *      +--------+-----------------------------------------------+  4K  |
+ *      | 0x0800 | G2H `CTB Descriptor`_ (recv)                  |      |
+ *      +--------+-----------------------------------------------+------+
+ *      | 0x1000 | H2G `CT Buffer`_ (send)                       | n*4K |
+ *      |        |                                               |      |
+ *      +--------+-----------------------------------------------+------+
+ *      | 0x1000 | G2H `CT Buffer`_ (recv)                       | m*4K |
+ *      | + n*4K |                                               |      |
+ *      +--------+-----------------------------------------------+------+
+ *
+ * Size of each `CT Buffer`_ must be multiple of 4K.
+ * We don't expect too many messages in flight at any time, unless we are
+ * using the GuC submission. In that case each request requires a minimum
+ * 2 dwords which gives us a maximum 256 queue'd requests. Hopefully this
+ * enough space to avoid backpressure on the driver. We increase the size
+ * of the receive buffer (relative to the send) to ensure a G2H response
+ * CTB has a landing spot.
+ */
+#define CTB_DESC_SIZE		ALIGN(sizeof(struct guc_ct_buffer_desc), SZ_2K)
+#define CTB_H2G_BUFFER_SIZE	(SZ_4K)
+#define CTB_G2H_BUFFER_SIZE	(4 * CTB_H2G_BUFFER_SIZE)
+#define G2H_ROOM_BUFFER_SIZE	(CTB_G2H_BUFFER_SIZE / 4)
+
+struct ct_request {
+	struct list_head link;
+	u32 fence;
+	u32 status;
+	u32 response_len;
+	u32 *response_buf;
+};
+
+struct ct_incoming_msg {
+	struct list_head link;
+	u32 size;
+	u32 msg[];
+};
+
+enum { CTB_SEND = 0, CTB_RECV = 1 };
+
+enum { CTB_OWNER_HOST = 0 };
+
+static void ct_receive_tasklet_func(struct tasklet_struct *t);
+static void ct_incoming_request_worker_func(struct work_struct *w);
+
+/**
+ * intel_guc_ct_init_early - Initialize CT state without requiring device access
+ * @ct: pointer to CT struct
+ */
+void intel_guc_ct_init_early(struct intel_guc_ct *ct)
+{
+	spin_lock_init(&ct->ctbs.send.lock);
+	spin_lock_init(&ct->ctbs.recv.lock);
+	spin_lock_init(&ct->requests.lock);
+	INIT_LIST_HEAD(&ct->requests.pending);
+	INIT_LIST_HEAD(&ct->requests.incoming);
+	INIT_WORK(&ct->requests.worker, ct_incoming_request_worker_func);
+	tasklet_setup(&ct->receive_tasklet, ct_receive_tasklet_func);
+	init_waitqueue_head(&ct->wq);
+}
+
+static void guc_ct_buffer_desc_init(struct guc_ct_buffer_desc *desc)
+{
+	memset(desc, 0, sizeof(*desc));
+}
+
+static void guc_ct_buffer_reset(struct intel_guc_ct_buffer *ctb)
+{
+	u32 space;
+
+	ctb->broken = false;
+	ctb->tail = 0;
+	ctb->head = 0;
+	space = CIRC_SPACE(ctb->tail, ctb->head, ctb->size) - ctb->resv_space;
+	atomic_set(&ctb->space, space);
+
+	guc_ct_buffer_desc_init(ctb->desc);
+}
+
+static void guc_ct_buffer_init(struct intel_guc_ct_buffer *ctb,
+			       struct guc_ct_buffer_desc *desc,
+			       u32 *cmds, u32 size_in_bytes, u32 resv_space)
+{
+	GEM_BUG_ON(size_in_bytes % 4);
+
+	ctb->desc = desc;
+	ctb->cmds = cmds;
+	ctb->size = size_in_bytes / 4;
+	ctb->resv_space = resv_space / 4;
+
+	guc_ct_buffer_reset(ctb);
+}
+
+static int guc_action_control_ctb(struct intel_guc *guc, u32 control)
+{
+	u32 request[HOST2GUC_CONTROL_CTB_REQUEST_MSG_LEN] = {
+		FIELD_PREP(GUC_HXG_MSG_0_ORIGIN, GUC_HXG_ORIGIN_HOST) |
+		FIELD_PREP(GUC_HXG_MSG_0_TYPE, GUC_HXG_TYPE_REQUEST) |
+		FIELD_PREP(GUC_HXG_REQUEST_MSG_0_ACTION, GUC_ACTION_HOST2GUC_CONTROL_CTB),
+		FIELD_PREP(HOST2GUC_CONTROL_CTB_REQUEST_MSG_1_CONTROL, control),
+	};
+	int ret;
+
+	GEM_BUG_ON(control != GUC_CTB_CONTROL_DISABLE && control != GUC_CTB_CONTROL_ENABLE);
+
+	/* CT control must go over MMIO */
+	ret = intel_guc_send_mmio(guc, request, ARRAY_SIZE(request), NULL, 0);
+
+	return ret > 0 ? -EPROTO : ret;
+}
+
+static int ct_control_enable(struct intel_guc_ct *ct, bool enable)
+{
+	int err;
+
+	err = guc_action_control_ctb(ct_to_guc(ct), enable ?
+				     GUC_CTB_CONTROL_ENABLE : GUC_CTB_CONTROL_DISABLE);
+	if (unlikely(err))
+		CT_PROBE_ERROR(ct, "Failed to control/%s CTB (%pe)\n",
+			       str_enable_disable(enable), ERR_PTR(err));
+
+	return err;
+}
+
+static int ct_register_buffer(struct intel_guc_ct *ct, bool send,
+			      u32 desc_addr, u32 buff_addr, u32 size)
+{
+	int err;
+
+	err = intel_guc_self_cfg64(ct_to_guc(ct), send ?
+				   GUC_KLV_SELF_CFG_H2G_CTB_DESCRIPTOR_ADDR_KEY :
+				   GUC_KLV_SELF_CFG_G2H_CTB_DESCRIPTOR_ADDR_KEY,
+				   desc_addr);
+	if (unlikely(err))
+		goto failed;
+
+	err = intel_guc_self_cfg64(ct_to_guc(ct), send ?
+				   GUC_KLV_SELF_CFG_H2G_CTB_ADDR_KEY :
+				   GUC_KLV_SELF_CFG_G2H_CTB_ADDR_KEY,
+				   buff_addr);
+	if (unlikely(err))
+		goto failed;
+
+	err = intel_guc_self_cfg32(ct_to_guc(ct), send ?
+				   GUC_KLV_SELF_CFG_H2G_CTB_SIZE_KEY :
+				   GUC_KLV_SELF_CFG_G2H_CTB_SIZE_KEY,
+				   size);
+	if (unlikely(err))
+failed:
+		CT_PROBE_ERROR(ct, "Failed to register %s buffer (%pe)\n",
+			       send ? "SEND" : "RECV", ERR_PTR(err));
+
+	return err;
+}
+
+/**
+ * intel_guc_ct_init - Init buffer-based communication
+ * @ct: pointer to CT struct
+ *
+ * Allocate memory required for buffer-based communication.
+ *
+ * Return: 0 on success, a negative errno code on failure.
+ */
+int intel_guc_ct_init(struct intel_guc_ct *ct)
+{
+	struct intel_guc *guc = ct_to_guc(ct);
+	struct guc_ct_buffer_desc *desc;
+	u32 blob_size;
+	u32 cmds_size;
+	u32 resv_space;
+	void *blob;
+	u32 *cmds;
+	int err;
+
+	err = i915_inject_probe_error(guc_to_gt(guc)->i915, -ENXIO);
+	if (err)
+		return err;
+
+	GEM_BUG_ON(ct->vma);
+
+	blob_size = 2 * CTB_DESC_SIZE + CTB_H2G_BUFFER_SIZE + CTB_G2H_BUFFER_SIZE;
+	err = intel_guc_allocate_and_map_vma(guc, blob_size, &ct->vma, &blob);
+	if (unlikely(err)) {
+		CT_PROBE_ERROR(ct, "Failed to allocate %u for CTB data (%pe)\n",
+			       blob_size, ERR_PTR(err));
+		return err;
+	}
+
+	CT_DEBUG(ct, "base=%#x size=%u\n", intel_guc_ggtt_offset(guc, ct->vma), blob_size);
+
+	/* store pointers to desc and cmds for send ctb */
+	desc = blob;
+	cmds = blob + 2 * CTB_DESC_SIZE;
+	cmds_size = CTB_H2G_BUFFER_SIZE;
+	resv_space = 0;
+	CT_DEBUG(ct, "%s desc %#tx cmds %#tx size %u/%u\n", "send",
+		 ptrdiff(desc, blob), ptrdiff(cmds, blob), cmds_size,
+		 resv_space);
+
+	guc_ct_buffer_init(&ct->ctbs.send, desc, cmds, cmds_size, resv_space);
+
+	/* store pointers to desc and cmds for recv ctb */
+	desc = blob + CTB_DESC_SIZE;
+	cmds = blob + 2 * CTB_DESC_SIZE + CTB_H2G_BUFFER_SIZE;
+	cmds_size = CTB_G2H_BUFFER_SIZE;
+	resv_space = G2H_ROOM_BUFFER_SIZE;
+	CT_DEBUG(ct, "%s desc %#tx cmds %#tx size %u/%u\n", "recv",
+		 ptrdiff(desc, blob), ptrdiff(cmds, blob), cmds_size,
+		 resv_space);
+
+	guc_ct_buffer_init(&ct->ctbs.recv, desc, cmds, cmds_size, resv_space);
+
+	return 0;
+}
+
+/**
+ * intel_guc_ct_fini - Fini buffer-based communication
+ * @ct: pointer to CT struct
+ *
+ * Deallocate memory required for buffer-based communication.
+ */
+void intel_guc_ct_fini(struct intel_guc_ct *ct)
+{
+	GEM_BUG_ON(ct->enabled);
+
+	tasklet_kill(&ct->receive_tasklet);
+	i915_vma_unpin_and_release(&ct->vma, I915_VMA_RELEASE_MAP);
+	memset(ct, 0, sizeof(*ct));
+}
+
+/**
+ * intel_guc_ct_enable - Enable buffer based command transport.
+ * @ct: pointer to CT struct
+ *
+ * Return: 0 on success, a negative errno code on failure.
+ */
+int intel_guc_ct_enable(struct intel_guc_ct *ct)
+{
+	struct intel_guc *guc = ct_to_guc(ct);
+	u32 base, desc, cmds, size;
+	void *blob;
+	int err;
+
+	GEM_BUG_ON(ct->enabled);
+
+	/* vma should be already allocated and map'ed */
+	GEM_BUG_ON(!ct->vma);
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(ct->vma->obj));
+	base = intel_guc_ggtt_offset(guc, ct->vma);
+
+	/* blob should start with send descriptor */
+	blob = __px_vaddr(ct->vma->obj);
+	GEM_BUG_ON(blob != ct->ctbs.send.desc);
+
+	/* (re)initialize descriptors */
+	guc_ct_buffer_reset(&ct->ctbs.send);
+	guc_ct_buffer_reset(&ct->ctbs.recv);
+
+	/*
+	 * Register both CT buffers starting with RECV buffer.
+	 * Descriptors are in first half of the blob.
+	 */
+	desc = base + ptrdiff(ct->ctbs.recv.desc, blob);
+	cmds = base + ptrdiff(ct->ctbs.recv.cmds, blob);
+	size = ct->ctbs.recv.size * 4;
+	err = ct_register_buffer(ct, false, desc, cmds, size);
+	if (unlikely(err))
+		goto err_out;
+
+	desc = base + ptrdiff(ct->ctbs.send.desc, blob);
+	cmds = base + ptrdiff(ct->ctbs.send.cmds, blob);
+	size = ct->ctbs.send.size * 4;
+	err = ct_register_buffer(ct, true, desc, cmds, size);
+	if (unlikely(err))
+		goto err_out;
+
+	err = ct_control_enable(ct, true);
+	if (unlikely(err))
+		goto err_out;
+
+	ct->enabled = true;
+	ct->stall_time = KTIME_MAX;
+
+	return 0;
+
+err_out:
+	CT_PROBE_ERROR(ct, "Failed to enable CTB (%pe)\n", ERR_PTR(err));
+	return err;
+}
+
+/**
+ * intel_guc_ct_disable - Disable buffer based command transport.
+ * @ct: pointer to CT struct
+ */
+void intel_guc_ct_disable(struct intel_guc_ct *ct)
+{
+	struct intel_guc *guc = ct_to_guc(ct);
+
+	GEM_BUG_ON(!ct->enabled);
+
+	ct->enabled = false;
+
+	if (intel_guc_is_fw_running(guc)) {
+		ct_control_enable(ct, false);
+	}
+}
+
+static u32 ct_get_next_fence(struct intel_guc_ct *ct)
+{
+	/* For now it's trivial */
+	return ++ct->requests.last_fence;
+}
+
+static int ct_write(struct intel_guc_ct *ct,
+		    const u32 *action,
+		    u32 len /* in dwords */,
+		    u32 fence, u32 flags)
+{
+	struct intel_guc_ct_buffer *ctb = &ct->ctbs.send;
+	struct guc_ct_buffer_desc *desc = ctb->desc;
+	u32 tail = ctb->tail;
+	u32 size = ctb->size;
+	u32 header;
+	u32 hxg;
+	u32 type;
+	u32 *cmds = ctb->cmds;
+	unsigned int i;
+
+	if (unlikely(desc->status))
+		goto corrupted;
+
+	GEM_BUG_ON(tail > size);
+
+#ifdef CONFIG_DRM_I915_DEBUG_GUC
+	if (unlikely(tail != READ_ONCE(desc->tail))) {
+		CT_ERROR(ct, "Tail was modified %u != %u\n",
+			 desc->tail, tail);
+		desc->status |= GUC_CTB_STATUS_MISMATCH;
+		goto corrupted;
+	}
+	if (unlikely(READ_ONCE(desc->head) >= size)) {
+		CT_ERROR(ct, "Invalid head offset %u >= %u)\n",
+			 desc->head, size);
+		desc->status |= GUC_CTB_STATUS_OVERFLOW;
+		goto corrupted;
+	}
+#endif
+
+	/*
+	 * dw0: CT header (including fence)
+	 * dw1: HXG header (including action code)
+	 * dw2+: action data
+	 */
+	header = FIELD_PREP(GUC_CTB_MSG_0_FORMAT, GUC_CTB_FORMAT_HXG) |
+		 FIELD_PREP(GUC_CTB_MSG_0_NUM_DWORDS, len) |
+		 FIELD_PREP(GUC_CTB_MSG_0_FENCE, fence);
+
+	type = (flags & INTEL_GUC_CT_SEND_NB) ? GUC_HXG_TYPE_EVENT :
+		GUC_HXG_TYPE_REQUEST;
+	hxg = FIELD_PREP(GUC_HXG_MSG_0_TYPE, type) |
+		FIELD_PREP(GUC_HXG_EVENT_MSG_0_ACTION |
+			   GUC_HXG_EVENT_MSG_0_DATA0, action[0]);
+
+	CT_DEBUG(ct, "writing (tail %u) %*ph %*ph %*ph\n",
+		 tail, 4, &header, 4, &hxg, 4 * (len - 1), &action[1]);
+
+	cmds[tail] = header;
+	tail = (tail + 1) % size;
+
+	cmds[tail] = hxg;
+	tail = (tail + 1) % size;
+
+	for (i = 1; i < len; i++) {
+		cmds[tail] = action[i];
+		tail = (tail + 1) % size;
+	}
+	GEM_BUG_ON(tail > size);
+
+	/*
+	 * make sure H2G buffer update and LRC tail update (if this triggering a
+	 * submission) are visible before updating the descriptor tail
+	 */
+	intel_guc_write_barrier(ct_to_guc(ct));
+
+	/* update local copies */
+	ctb->tail = tail;
+	GEM_BUG_ON(atomic_read(&ctb->space) < len + GUC_CTB_HDR_LEN);
+	atomic_sub(len + GUC_CTB_HDR_LEN, &ctb->space);
+
+	/* now update descriptor */
+	WRITE_ONCE(desc->tail, tail);
+
+	return 0;
+
+corrupted:
+	CT_ERROR(ct, "Corrupted descriptor head=%u tail=%u status=%#x\n",
+		 desc->head, desc->tail, desc->status);
+	ctb->broken = true;
+	return -EPIPE;
+}
+
+/**
+ * wait_for_ct_request_update - Wait for CT request state update.
+ * @ct:		pointer to CT
+ * @req:	pointer to pending request
+ * @status:	placeholder for status
+ *
+ * For each sent request, GuC shall send back CT response message.
+ * Our message handler will update status of tracked request once
+ * response message with given fence is received. Wait here and
+ * check for valid response status value.
+ *
+ * Return:
+ * *	0 response received (status is valid)
+ * *	-ETIMEDOUT no response within hardcoded timeout
+ */
+static int wait_for_ct_request_update(struct intel_guc_ct *ct, struct ct_request *req, u32 *status)
+{
+	int err;
+	bool ct_enabled;
+
+	/*
+	 * Fast commands should complete in less than 10us, so sample quickly
+	 * up to that length of time, then switch to a slower sleep-wait loop.
+	 * No GuC command should ever take longer than 10ms but many GuC
+	 * commands can be inflight at time, so use a 1s timeout on the slower
+	 * sleep-wait loop.
+	 */
+#define GUC_CTB_RESPONSE_TIMEOUT_SHORT_MS 10
+#define GUC_CTB_RESPONSE_TIMEOUT_LONG_MS 1000
+#define done \
+	(!(ct_enabled = intel_guc_ct_enabled(ct)) || \
+	 FIELD_GET(GUC_HXG_MSG_0_ORIGIN, READ_ONCE(req->status)) == \
+	 GUC_HXG_ORIGIN_GUC)
+	err = wait_for_us(done, GUC_CTB_RESPONSE_TIMEOUT_SHORT_MS);
+	if (err)
+		err = wait_for(done, GUC_CTB_RESPONSE_TIMEOUT_LONG_MS);
+#undef done
+	if (!ct_enabled)
+		err = -ENODEV;
+
+	*status = req->status;
+	return err;
+}
+
+#define GUC_CTB_TIMEOUT_MS	1500
+static inline bool ct_deadlocked(struct intel_guc_ct *ct)
+{
+	long timeout = GUC_CTB_TIMEOUT_MS;
+	bool ret = ktime_ms_delta(ktime_get(), ct->stall_time) > timeout;
+
+	if (unlikely(ret)) {
+		struct guc_ct_buffer_desc *send = ct->ctbs.send.desc;
+		struct guc_ct_buffer_desc *recv = ct->ctbs.send.desc;
+
+		CT_ERROR(ct, "Communication stalled for %lld ms, desc status=%#x,%#x\n",
+			 ktime_ms_delta(ktime_get(), ct->stall_time),
+			 send->status, recv->status);
+		CT_ERROR(ct, "H2G Space: %u (Bytes)\n",
+			 atomic_read(&ct->ctbs.send.space) * 4);
+		CT_ERROR(ct, "Head: %u (Dwords)\n", ct->ctbs.send.desc->head);
+		CT_ERROR(ct, "Tail: %u (Dwords)\n", ct->ctbs.send.desc->tail);
+		CT_ERROR(ct, "G2H Space: %u (Bytes)\n",
+			 atomic_read(&ct->ctbs.recv.space) * 4);
+		CT_ERROR(ct, "Head: %u\n (Dwords)", ct->ctbs.recv.desc->head);
+		CT_ERROR(ct, "Tail: %u\n (Dwords)", ct->ctbs.recv.desc->tail);
+
+		ct->ctbs.send.broken = true;
+	}
+
+	return ret;
+}
+
+static inline bool g2h_has_room(struct intel_guc_ct *ct, u32 g2h_len_dw)
+{
+	struct intel_guc_ct_buffer *ctb = &ct->ctbs.recv;
+
+	/*
+	 * We leave a certain amount of space in the G2H CTB buffer for
+	 * unexpected G2H CTBs (e.g. logging, engine hang, etc...)
+	 */
+	return !g2h_len_dw || atomic_read(&ctb->space) >= g2h_len_dw;
+}
+
+static inline void g2h_reserve_space(struct intel_guc_ct *ct, u32 g2h_len_dw)
+{
+	lockdep_assert_held(&ct->ctbs.send.lock);
+
+	GEM_BUG_ON(!g2h_has_room(ct, g2h_len_dw));
+
+	if (g2h_len_dw)
+		atomic_sub(g2h_len_dw, &ct->ctbs.recv.space);
+}
+
+static inline void g2h_release_space(struct intel_guc_ct *ct, u32 g2h_len_dw)
+{
+	atomic_add(g2h_len_dw, &ct->ctbs.recv.space);
+}
+
+static inline bool h2g_has_room(struct intel_guc_ct *ct, u32 len_dw)
+{
+	struct intel_guc_ct_buffer *ctb = &ct->ctbs.send;
+	struct guc_ct_buffer_desc *desc = ctb->desc;
+	u32 head;
+	u32 space;
+
+	if (atomic_read(&ctb->space) >= len_dw)
+		return true;
+
+	head = READ_ONCE(desc->head);
+	if (unlikely(head > ctb->size)) {
+		CT_ERROR(ct, "Invalid head offset %u >= %u)\n",
+			 head, ctb->size);
+		desc->status |= GUC_CTB_STATUS_OVERFLOW;
+		ctb->broken = true;
+		return false;
+	}
+
+	space = CIRC_SPACE(ctb->tail, head, ctb->size);
+	atomic_set(&ctb->space, space);
+
+	return space >= len_dw;
+}
+
+static int has_room_nb(struct intel_guc_ct *ct, u32 h2g_dw, u32 g2h_dw)
+{
+	bool h2g = h2g_has_room(ct, h2g_dw);
+	bool g2h = g2h_has_room(ct, g2h_dw);
+
+	lockdep_assert_held(&ct->ctbs.send.lock);
+
+	if (unlikely(!h2g || !g2h)) {
+		if (ct->stall_time == KTIME_MAX)
+			ct->stall_time = ktime_get();
+
+		/* Be paranoid and kick G2H tasklet to free credits */
+		if (!g2h)
+			tasklet_hi_schedule(&ct->receive_tasklet);
+
+		if (unlikely(ct_deadlocked(ct)))
+			return -EPIPE;
+		else
+			return -EBUSY;
+	}
+
+	ct->stall_time = KTIME_MAX;
+	return 0;
+}
+
+#define G2H_LEN_DW(f) ({ \
+	typeof(f) f_ = (f); \
+	FIELD_GET(INTEL_GUC_CT_SEND_G2H_DW_MASK, f_) ? \
+	FIELD_GET(INTEL_GUC_CT_SEND_G2H_DW_MASK, f_) + \
+	GUC_CTB_HXG_MSG_MIN_LEN : 0; \
+})
+static int ct_send_nb(struct intel_guc_ct *ct,
+		      const u32 *action,
+		      u32 len,
+		      u32 flags)
+{
+	struct intel_guc_ct_buffer *ctb = &ct->ctbs.send;
+	unsigned long spin_flags;
+	u32 g2h_len_dw = G2H_LEN_DW(flags);
+	u32 fence;
+	int ret;
+
+	spin_lock_irqsave(&ctb->lock, spin_flags);
+
+	ret = has_room_nb(ct, len + GUC_CTB_HDR_LEN, g2h_len_dw);
+	if (unlikely(ret))
+		goto out;
+
+	fence = ct_get_next_fence(ct);
+	ret = ct_write(ct, action, len, fence, flags);
+	if (unlikely(ret))
+		goto out;
+
+	g2h_reserve_space(ct, g2h_len_dw);
+	intel_guc_notify(ct_to_guc(ct));
+
+out:
+	spin_unlock_irqrestore(&ctb->lock, spin_flags);
+
+	return ret;
+}
+
+static int ct_send(struct intel_guc_ct *ct,
+		   const u32 *action,
+		   u32 len,
+		   u32 *response_buf,
+		   u32 response_buf_size,
+		   u32 *status)
+{
+	struct intel_guc_ct_buffer *ctb = &ct->ctbs.send;
+	struct ct_request request;
+	unsigned long flags;
+	unsigned int sleep_period_ms = 1;
+	bool send_again;
+	u32 fence;
+	int err;
+
+	GEM_BUG_ON(!ct->enabled);
+	GEM_BUG_ON(!len);
+	GEM_BUG_ON(len & ~GUC_CT_MSG_LEN_MASK);
+	GEM_BUG_ON(!response_buf && response_buf_size);
+	might_sleep();
+
+resend:
+	send_again = false;
+
+	/*
+	 * We use a lazy spin wait loop here as we believe that if the CT
+	 * buffers are sized correctly the flow control condition should be
+	 * rare. Reserving the maximum size in the G2H credits as we don't know
+	 * how big the response is going to be.
+	 */
+retry:
+	spin_lock_irqsave(&ctb->lock, flags);
+	if (unlikely(!h2g_has_room(ct, len + GUC_CTB_HDR_LEN) ||
+		     !g2h_has_room(ct, GUC_CTB_HXG_MSG_MAX_LEN))) {
+		if (ct->stall_time == KTIME_MAX)
+			ct->stall_time = ktime_get();
+		spin_unlock_irqrestore(&ctb->lock, flags);
+
+		if (unlikely(ct_deadlocked(ct)))
+			return -EPIPE;
+
+		if (msleep_interruptible(sleep_period_ms))
+			return -EINTR;
+		sleep_period_ms = sleep_period_ms << 1;
+
+		goto retry;
+	}
+
+	ct->stall_time = KTIME_MAX;
+
+	fence = ct_get_next_fence(ct);
+	request.fence = fence;
+	request.status = 0;
+	request.response_len = response_buf_size;
+	request.response_buf = response_buf;
+
+	spin_lock(&ct->requests.lock);
+	list_add_tail(&request.link, &ct->requests.pending);
+	spin_unlock(&ct->requests.lock);
+
+	err = ct_write(ct, action, len, fence, 0);
+	g2h_reserve_space(ct, GUC_CTB_HXG_MSG_MAX_LEN);
+
+	spin_unlock_irqrestore(&ctb->lock, flags);
+
+	if (unlikely(err))
+		goto unlink;
+
+	intel_guc_notify(ct_to_guc(ct));
+
+	err = wait_for_ct_request_update(ct, &request, status);
+	g2h_release_space(ct, GUC_CTB_HXG_MSG_MAX_LEN);
+	if (unlikely(err)) {
+		if (err == -ENODEV)
+			/* wait_for_ct_request_update returns -ENODEV on reset/suspend in progress.
+			 * In this case, output is debug rather than error info
+			 */
+			CT_DEBUG(ct, "Request %#x (fence %u) cancelled as CTB is disabled\n",
+				 action[0], request.fence);
+		else
+			CT_ERROR(ct, "No response for request %#x (fence %u)\n",
+				 action[0], request.fence);
+		goto unlink;
+	}
+
+	if (FIELD_GET(GUC_HXG_MSG_0_TYPE, *status) == GUC_HXG_TYPE_NO_RESPONSE_RETRY) {
+		CT_DEBUG(ct, "retrying request %#x (%u)\n", *action,
+			 FIELD_GET(GUC_HXG_RETRY_MSG_0_REASON, *status));
+		send_again = true;
+		goto unlink;
+	}
+
+	if (FIELD_GET(GUC_HXG_MSG_0_TYPE, *status) != GUC_HXG_TYPE_RESPONSE_SUCCESS) {
+		err = -EIO;
+		goto unlink;
+	}
+
+	if (response_buf) {
+		/* There shall be no data in the status */
+		WARN_ON(FIELD_GET(GUC_HXG_RESPONSE_MSG_0_DATA0, request.status));
+		/* Return actual response len */
+		err = request.response_len;
+	} else {
+		/* There shall be no response payload */
+		WARN_ON(request.response_len);
+		/* Return data decoded from the status dword */
+		err = FIELD_GET(GUC_HXG_RESPONSE_MSG_0_DATA0, *status);
+	}
+
+unlink:
+	spin_lock_irqsave(&ct->requests.lock, flags);
+	list_del(&request.link);
+	spin_unlock_irqrestore(&ct->requests.lock, flags);
+
+	if (unlikely(send_again))
+		goto resend;
+
+	return err;
+}
+
+/*
+ * Command Transport (CT) buffer based GuC send function.
+ */
+int intel_guc_ct_send(struct intel_guc_ct *ct, const u32 *action, u32 len,
+		      u32 *response_buf, u32 response_buf_size, u32 flags)
+{
+	u32 status = ~0; /* undefined */
+	int ret;
+
+	if (unlikely(!ct->enabled)) {
+		struct intel_guc *guc = ct_to_guc(ct);
+		struct intel_uc *uc = container_of(guc, struct intel_uc, guc);
+
+		WARN(!uc->reset_in_progress, "Unexpected send: action=%#x\n", *action);
+		return -ENODEV;
+	}
+
+	if (unlikely(ct->ctbs.send.broken))
+		return -EPIPE;
+
+	if (flags & INTEL_GUC_CT_SEND_NB)
+		return ct_send_nb(ct, action, len, flags);
+
+	ret = ct_send(ct, action, len, response_buf, response_buf_size, &status);
+	if (unlikely(ret < 0)) {
+		if (ret != -ENODEV)
+			CT_ERROR(ct, "Sending action %#x failed (%pe) status=%#X\n",
+				 action[0], ERR_PTR(ret), status);
+	} else if (unlikely(ret)) {
+		CT_DEBUG(ct, "send action %#x returned %d (%#x)\n",
+			 action[0], ret, ret);
+	}
+
+	return ret;
+}
+
+static struct ct_incoming_msg *ct_alloc_msg(u32 num_dwords)
+{
+	struct ct_incoming_msg *msg;
+
+	msg = kmalloc(struct_size(msg, msg, num_dwords), GFP_ATOMIC);
+	if (msg)
+		msg->size = num_dwords;
+	return msg;
+}
+
+static void ct_free_msg(struct ct_incoming_msg *msg)
+{
+	kfree(msg);
+}
+
+/*
+ * Return: number available remaining dwords to read (0 if empty)
+ *         or a negative error code on failure
+ */
+static int ct_read(struct intel_guc_ct *ct, struct ct_incoming_msg **msg)
+{
+	struct intel_guc_ct_buffer *ctb = &ct->ctbs.recv;
+	struct guc_ct_buffer_desc *desc = ctb->desc;
+	u32 head = ctb->head;
+	u32 tail = READ_ONCE(desc->tail);
+	u32 size = ctb->size;
+	u32 *cmds = ctb->cmds;
+	s32 available;
+	unsigned int len;
+	unsigned int i;
+	u32 header;
+
+	if (unlikely(ctb->broken))
+		return -EPIPE;
+
+	if (unlikely(desc->status)) {
+		u32 status = desc->status;
+
+		if (status & GUC_CTB_STATUS_UNUSED) {
+			/*
+			 * Potentially valid if a CLIENT_RESET request resulted in
+			 * contexts/engines being reset. But should never happen as
+			 * no contexts should be active when CLIENT_RESET is sent.
+			 */
+			CT_ERROR(ct, "Unexpected G2H after GuC has stopped!\n");
+			status &= ~GUC_CTB_STATUS_UNUSED;
+		}
+
+		if (status)
+			goto corrupted;
+	}
+
+	GEM_BUG_ON(head > size);
+
+#ifdef CONFIG_DRM_I915_DEBUG_GUC
+	if (unlikely(head != READ_ONCE(desc->head))) {
+		CT_ERROR(ct, "Head was modified %u != %u\n",
+			 desc->head, head);
+		desc->status |= GUC_CTB_STATUS_MISMATCH;
+		goto corrupted;
+	}
+#endif
+	if (unlikely(tail >= size)) {
+		CT_ERROR(ct, "Invalid tail offset %u >= %u)\n",
+			 tail, size);
+		desc->status |= GUC_CTB_STATUS_OVERFLOW;
+		goto corrupted;
+	}
+
+	/* tail == head condition indicates empty */
+	available = tail - head;
+	if (unlikely(available == 0)) {
+		*msg = NULL;
+		return 0;
+	}
+
+	/* beware of buffer wrap case */
+	if (unlikely(available < 0))
+		available += size;
+	CT_DEBUG(ct, "available %d (%u:%u:%u)\n", available, head, tail, size);
+	GEM_BUG_ON(available < 0);
+
+	header = cmds[head];
+	head = (head + 1) % size;
+
+	/* message len with header */
+	len = FIELD_GET(GUC_CTB_MSG_0_NUM_DWORDS, header) + GUC_CTB_MSG_MIN_LEN;
+	if (unlikely(len > (u32)available)) {
+		CT_ERROR(ct, "Incomplete message %*ph %*ph %*ph\n",
+			 4, &header,
+			 4 * (head + available - 1 > size ?
+			      size - head : available - 1), &cmds[head],
+			 4 * (head + available - 1 > size ?
+			      available - 1 - size + head : 0), &cmds[0]);
+		desc->status |= GUC_CTB_STATUS_UNDERFLOW;
+		goto corrupted;
+	}
+
+	*msg = ct_alloc_msg(len);
+	if (!*msg) {
+		CT_ERROR(ct, "No memory for message %*ph %*ph %*ph\n",
+			 4, &header,
+			 4 * (head + available - 1 > size ?
+			      size - head : available - 1), &cmds[head],
+			 4 * (head + available - 1 > size ?
+			      available - 1 - size + head : 0), &cmds[0]);
+		return available;
+	}
+
+	(*msg)->msg[0] = header;
+
+	for (i = 1; i < len; i++) {
+		(*msg)->msg[i] = cmds[head];
+		head = (head + 1) % size;
+	}
+	CT_DEBUG(ct, "received %*ph\n", 4 * len, (*msg)->msg);
+
+	/* update local copies */
+	ctb->head = head;
+
+	/* now update descriptor */
+	WRITE_ONCE(desc->head, head);
+
+	return available - len;
+
+corrupted:
+	CT_ERROR(ct, "Corrupted descriptor head=%u tail=%u status=%#x\n",
+		 desc->head, desc->tail, desc->status);
+	ctb->broken = true;
+	return -EPIPE;
+}
+
+static int ct_handle_response(struct intel_guc_ct *ct, struct ct_incoming_msg *response)
+{
+	u32 len = FIELD_GET(GUC_CTB_MSG_0_NUM_DWORDS, response->msg[0]);
+	u32 fence = FIELD_GET(GUC_CTB_MSG_0_FENCE, response->msg[0]);
+	const u32 *hxg = &response->msg[GUC_CTB_MSG_MIN_LEN];
+	const u32 *data = &hxg[GUC_HXG_MSG_MIN_LEN];
+	u32 datalen = len - GUC_HXG_MSG_MIN_LEN;
+	struct ct_request *req;
+	unsigned long flags;
+	bool found = false;
+	int err = 0;
+
+	GEM_BUG_ON(len < GUC_HXG_MSG_MIN_LEN);
+	GEM_BUG_ON(FIELD_GET(GUC_HXG_MSG_0_ORIGIN, hxg[0]) != GUC_HXG_ORIGIN_GUC);
+	GEM_BUG_ON(FIELD_GET(GUC_HXG_MSG_0_TYPE, hxg[0]) != GUC_HXG_TYPE_RESPONSE_SUCCESS &&
+		   FIELD_GET(GUC_HXG_MSG_0_TYPE, hxg[0]) != GUC_HXG_TYPE_NO_RESPONSE_RETRY &&
+		   FIELD_GET(GUC_HXG_MSG_0_TYPE, hxg[0]) != GUC_HXG_TYPE_RESPONSE_FAILURE);
+
+	CT_DEBUG(ct, "response fence %u status %#x\n", fence, hxg[0]);
+
+	spin_lock_irqsave(&ct->requests.lock, flags);
+	list_for_each_entry(req, &ct->requests.pending, link) {
+		if (unlikely(fence != req->fence)) {
+			CT_DEBUG(ct, "request %u awaits response\n",
+				 req->fence);
+			continue;
+		}
+		if (unlikely(datalen > req->response_len)) {
+			CT_ERROR(ct, "Response %u too long (datalen %u > %u)\n",
+				 req->fence, datalen, req->response_len);
+			datalen = min(datalen, req->response_len);
+			err = -EMSGSIZE;
+		}
+		if (datalen)
+			memcpy(req->response_buf, data, 4 * datalen);
+		req->response_len = datalen;
+		WRITE_ONCE(req->status, hxg[0]);
+		found = true;
+		break;
+	}
+	if (!found) {
+		CT_ERROR(ct, "Unsolicited response (fence %u)\n", fence);
+		CT_ERROR(ct, "Could not find fence=%u, last_fence=%u\n", fence,
+			 ct->requests.last_fence);
+		list_for_each_entry(req, &ct->requests.pending, link)
+			CT_ERROR(ct, "request %u awaits response\n",
+				 req->fence);
+		err = -ENOKEY;
+	}
+	spin_unlock_irqrestore(&ct->requests.lock, flags);
+
+	if (unlikely(err))
+		return err;
+
+	ct_free_msg(response);
+	return 0;
+}
+
+static int ct_process_request(struct intel_guc_ct *ct, struct ct_incoming_msg *request)
+{
+	struct intel_guc *guc = ct_to_guc(ct);
+	const u32 *hxg;
+	const u32 *payload;
+	u32 hxg_len, action, len;
+	int ret;
+
+	hxg = &request->msg[GUC_CTB_MSG_MIN_LEN];
+	hxg_len = request->size - GUC_CTB_MSG_MIN_LEN;
+	payload = &hxg[GUC_HXG_MSG_MIN_LEN];
+	action = FIELD_GET(GUC_HXG_EVENT_MSG_0_ACTION, hxg[0]);
+	len = hxg_len - GUC_HXG_MSG_MIN_LEN;
+
+	CT_DEBUG(ct, "request %x %*ph\n", action, 4 * len, payload);
+
+	switch (action) {
+	case INTEL_GUC_ACTION_DEFAULT:
+		ret = intel_guc_to_host_process_recv_msg(guc, payload, len);
+		break;
+	case INTEL_GUC_ACTION_DEREGISTER_CONTEXT_DONE:
+		ret = intel_guc_deregister_done_process_msg(guc, payload,
+							    len);
+		break;
+	case INTEL_GUC_ACTION_SCHED_CONTEXT_MODE_DONE:
+		ret = intel_guc_sched_done_process_msg(guc, payload, len);
+		break;
+	case INTEL_GUC_ACTION_CONTEXT_RESET_NOTIFICATION:
+		ret = intel_guc_context_reset_process_msg(guc, payload, len);
+		break;
+	case INTEL_GUC_ACTION_STATE_CAPTURE_NOTIFICATION:
+		ret = intel_guc_error_capture_process_msg(guc, payload, len);
+		if (unlikely(ret))
+			CT_ERROR(ct, "error capture notification failed %x %*ph\n",
+				 action, 4 * len, payload);
+		break;
+	case INTEL_GUC_ACTION_ENGINE_FAILURE_NOTIFICATION:
+		ret = intel_guc_engine_failure_process_msg(guc, payload, len);
+		break;
+	case INTEL_GUC_ACTION_NOTIFY_FLUSH_LOG_BUFFER_TO_FILE:
+		intel_guc_log_handle_flush_event(&guc->log);
+		ret = 0;
+		break;
+	case INTEL_GUC_ACTION_NOTIFY_CRASH_DUMP_POSTED:
+		CT_ERROR(ct, "Received GuC crash dump notification!\n");
+		ret = 0;
+		break;
+	case INTEL_GUC_ACTION_NOTIFY_EXCEPTION:
+		CT_ERROR(ct, "Received GuC exception notification!\n");
+		ret = 0;
+		break;
+	default:
+		ret = -EOPNOTSUPP;
+		break;
+	}
+
+	if (unlikely(ret)) {
+		CT_ERROR(ct, "Failed to process request %04x (%pe)\n",
+			 action, ERR_PTR(ret));
+		return ret;
+	}
+
+	ct_free_msg(request);
+	return 0;
+}
+
+static bool ct_process_incoming_requests(struct intel_guc_ct *ct)
+{
+	unsigned long flags;
+	struct ct_incoming_msg *request;
+	bool done;
+	int err;
+
+	spin_lock_irqsave(&ct->requests.lock, flags);
+	request = list_first_entry_or_null(&ct->requests.incoming,
+					   struct ct_incoming_msg, link);
+	if (request)
+		list_del(&request->link);
+	done = !!list_empty(&ct->requests.incoming);
+	spin_unlock_irqrestore(&ct->requests.lock, flags);
+
+	if (!request)
+		return true;
+
+	err = ct_process_request(ct, request);
+	if (unlikely(err)) {
+		CT_ERROR(ct, "Failed to process CT message (%pe) %*ph\n",
+			 ERR_PTR(err), 4 * request->size, request->msg);
+		ct_free_msg(request);
+	}
+
+	return done;
+}
+
+static void ct_incoming_request_worker_func(struct work_struct *w)
+{
+	struct intel_guc_ct *ct =
+		container_of(w, struct intel_guc_ct, requests.worker);
+	bool done;
+
+	do {
+		done = ct_process_incoming_requests(ct);
+	} while (!done);
+}
+
+static int ct_handle_event(struct intel_guc_ct *ct, struct ct_incoming_msg *request)
+{
+	const u32 *hxg = &request->msg[GUC_CTB_MSG_MIN_LEN];
+	u32 action = FIELD_GET(GUC_HXG_EVENT_MSG_0_ACTION, hxg[0]);
+	unsigned long flags;
+
+	GEM_BUG_ON(FIELD_GET(GUC_HXG_MSG_0_TYPE, hxg[0]) != GUC_HXG_TYPE_EVENT);
+
+	/*
+	 * Adjusting the space must be done in IRQ or deadlock can occur as the
+	 * CTB processing in the below workqueue can send CTBs which creates a
+	 * circular dependency if the space was returned there.
+	 */
+	switch (action) {
+	case INTEL_GUC_ACTION_SCHED_CONTEXT_MODE_DONE:
+	case INTEL_GUC_ACTION_DEREGISTER_CONTEXT_DONE:
+		g2h_release_space(ct, request->size);
+	}
+
+	spin_lock_irqsave(&ct->requests.lock, flags);
+	list_add_tail(&request->link, &ct->requests.incoming);
+	spin_unlock_irqrestore(&ct->requests.lock, flags);
+
+	queue_work(system_unbound_wq, &ct->requests.worker);
+	return 0;
+}
+
+static int ct_handle_hxg(struct intel_guc_ct *ct, struct ct_incoming_msg *msg)
+{
+	u32 origin, type;
+	u32 *hxg;
+	int err;
+
+	if (unlikely(msg->size < GUC_CTB_HXG_MSG_MIN_LEN))
+		return -EBADMSG;
+
+	hxg = &msg->msg[GUC_CTB_MSG_MIN_LEN];
+
+	origin = FIELD_GET(GUC_HXG_MSG_0_ORIGIN, hxg[0]);
+	if (unlikely(origin != GUC_HXG_ORIGIN_GUC)) {
+		err = -EPROTO;
+		goto failed;
+	}
+
+	type = FIELD_GET(GUC_HXG_MSG_0_TYPE, hxg[0]);
+	switch (type) {
+	case GUC_HXG_TYPE_EVENT:
+		err = ct_handle_event(ct, msg);
+		break;
+	case GUC_HXG_TYPE_RESPONSE_SUCCESS:
+	case GUC_HXG_TYPE_RESPONSE_FAILURE:
+	case GUC_HXG_TYPE_NO_RESPONSE_RETRY:
+		err = ct_handle_response(ct, msg);
+		break;
+	default:
+		err = -EOPNOTSUPP;
+	}
+
+	if (unlikely(err)) {
+failed:
+		CT_ERROR(ct, "Failed to handle HXG message (%pe) %*ph\n",
+			 ERR_PTR(err), 4 * GUC_HXG_MSG_MIN_LEN, hxg);
+	}
+	return err;
+}
+
+static void ct_handle_msg(struct intel_guc_ct *ct, struct ct_incoming_msg *msg)
+{
+	u32 format = FIELD_GET(GUC_CTB_MSG_0_FORMAT, msg->msg[0]);
+	int err;
+
+	if (format == GUC_CTB_FORMAT_HXG)
+		err = ct_handle_hxg(ct, msg);
+	else
+		err = -EOPNOTSUPP;
+
+	if (unlikely(err)) {
+		CT_ERROR(ct, "Failed to process CT message (%pe) %*ph\n",
+			 ERR_PTR(err), 4 * msg->size, msg->msg);
+		ct_free_msg(msg);
+	}
+}
+
+/*
+ * Return: number available remaining dwords to read (0 if empty)
+ *         or a negative error code on failure
+ */
+static int ct_receive(struct intel_guc_ct *ct)
+{
+	struct ct_incoming_msg *msg = NULL;
+	unsigned long flags;
+	int ret;
+
+	spin_lock_irqsave(&ct->ctbs.recv.lock, flags);
+	ret = ct_read(ct, &msg);
+	spin_unlock_irqrestore(&ct->ctbs.recv.lock, flags);
+	if (ret < 0)
+		return ret;
+
+	if (msg)
+		ct_handle_msg(ct, msg);
+
+	return ret;
+}
+
+static void ct_try_receive_message(struct intel_guc_ct *ct)
+{
+	int ret;
+
+	if (GEM_WARN_ON(!ct->enabled))
+		return;
+
+	ret = ct_receive(ct);
+	if (ret > 0)
+		tasklet_hi_schedule(&ct->receive_tasklet);
+}
+
+static void ct_receive_tasklet_func(struct tasklet_struct *t)
+{
+	struct intel_guc_ct *ct = from_tasklet(ct, t, receive_tasklet);
+
+	ct_try_receive_message(ct);
+}
+
+/*
+ * When we're communicating with the GuC over CT, GuC uses events
+ * to notify us about new messages being posted on the RECV buffer.
+ */
+void intel_guc_ct_event_handler(struct intel_guc_ct *ct)
+{
+	if (unlikely(!ct->enabled)) {
+		WARN(1, "Unexpected GuC event received while CT disabled!\n");
+		return;
+	}
+
+	ct_try_receive_message(ct);
+}
+
+void intel_guc_ct_print_info(struct intel_guc_ct *ct,
+			     struct drm_printer *p)
+{
+	drm_printf(p, "CT %s\n", str_enabled_disabled(ct->enabled));
+
+	if (!ct->enabled)
+		return;
+
+	drm_printf(p, "H2G Space: %u\n",
+		   atomic_read(&ct->ctbs.send.space) * 4);
+	drm_printf(p, "Head: %u\n",
+		   ct->ctbs.send.desc->head);
+	drm_printf(p, "Tail: %u\n",
+		   ct->ctbs.send.desc->tail);
+	drm_printf(p, "G2H Space: %u\n",
+		   atomic_read(&ct->ctbs.recv.space) * 4);
+	drm_printf(p, "Head: %u\n",
+		   ct->ctbs.recv.desc->head);
+	drm_printf(p, "Tail: %u\n",
+		   ct->ctbs.recv.desc->tail);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h
new file mode 100644
index 000000000..f709a19c7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h
@@ -0,0 +1,120 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2016-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_CT_H_
+#define _INTEL_GUC_CT_H_
+
+#include <linux/interrupt.h>
+#include <linux/spinlock.h>
+#include <linux/workqueue.h>
+#include <linux/ktime.h>
+#include <linux/wait.h>
+
+#include "intel_guc_fwif.h"
+
+struct i915_vma;
+struct intel_guc;
+struct drm_printer;
+
+/**
+ * DOC: Command Transport (CT).
+ *
+ * Buffer based command transport is a replacement for MMIO based mechanism.
+ * It can be used to perform both host-2-guc and guc-to-host communication.
+ */
+
+/** Represents single command transport buffer.
+ *
+ * A single command transport buffer consists of two parts, the header
+ * record (command transport buffer descriptor) and the actual buffer which
+ * holds the commands.
+ *
+ * @lock: protects access to the commands buffer and buffer descriptor
+ * @desc: pointer to the buffer descriptor
+ * @cmds: pointer to the commands buffer
+ * @size: size of the commands buffer in dwords
+ * @resv_space: reserved space in buffer in dwords
+ * @head: local shadow copy of head in dwords
+ * @tail: local shadow copy of tail in dwords
+ * @space: local shadow copy of space in dwords
+ * @broken: flag to indicate if descriptor data is broken
+ */
+struct intel_guc_ct_buffer {
+	spinlock_t lock;
+	struct guc_ct_buffer_desc *desc;
+	u32 *cmds;
+	u32 size;
+	u32 resv_space;
+	u32 tail;
+	u32 head;
+	atomic_t space;
+	bool broken;
+};
+
+/** Top-level structure for Command Transport related data
+ *
+ * Includes a pair of CT buffers for bi-directional communication and tracking
+ * for the H2G and G2H requests sent and received through the buffers.
+ */
+struct intel_guc_ct {
+	struct i915_vma *vma;
+	bool enabled;
+
+	/* buffers for sending and receiving commands */
+	struct {
+		struct intel_guc_ct_buffer send;
+		struct intel_guc_ct_buffer recv;
+	} ctbs;
+
+	struct tasklet_struct receive_tasklet;
+
+	/** @wq: wait queue for g2h chanenl */
+	wait_queue_head_t wq;
+
+	struct {
+		u16 last_fence; /* last fence used to send request */
+
+		spinlock_t lock; /* protects pending requests list */
+		struct list_head pending; /* requests waiting for response */
+
+		struct list_head incoming; /* incoming requests */
+		struct work_struct worker; /* handler for incoming requests */
+	} requests;
+
+	/** @stall_time: time of first time a CTB submission is stalled */
+	ktime_t stall_time;
+};
+
+void intel_guc_ct_init_early(struct intel_guc_ct *ct);
+int intel_guc_ct_init(struct intel_guc_ct *ct);
+void intel_guc_ct_fini(struct intel_guc_ct *ct);
+int intel_guc_ct_enable(struct intel_guc_ct *ct);
+void intel_guc_ct_disable(struct intel_guc_ct *ct);
+
+static inline void intel_guc_ct_sanitize(struct intel_guc_ct *ct)
+{
+	ct->enabled = false;
+}
+
+static inline bool intel_guc_ct_enabled(struct intel_guc_ct *ct)
+{
+	return ct->enabled;
+}
+
+#define INTEL_GUC_CT_SEND_NB		BIT(31)
+#define INTEL_GUC_CT_SEND_G2H_DW_SHIFT	0
+#define INTEL_GUC_CT_SEND_G2H_DW_MASK	(0xff << INTEL_GUC_CT_SEND_G2H_DW_SHIFT)
+#define MAKE_SEND_FLAGS(len) ({ \
+	typeof(len) len_ = (len); \
+	GEM_BUG_ON(!FIELD_FIT(INTEL_GUC_CT_SEND_G2H_DW_MASK, len_)); \
+	(FIELD_PREP(INTEL_GUC_CT_SEND_G2H_DW_MASK, len_) | INTEL_GUC_CT_SEND_NB); \
+})
+int intel_guc_ct_send(struct intel_guc_ct *ct, const u32 *action, u32 len,
+		      u32 *response_buf, u32 response_buf_size, u32 flags);
+void intel_guc_ct_event_handler(struct intel_guc_ct *ct);
+
+void intel_guc_ct_print_info(struct intel_guc_ct *ct, struct drm_printer *p);
+
+#endif /* _INTEL_GUC_CT_H_ */
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_debugfs.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_debugfs.c
new file mode 100644
index 000000000..25f09a420
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_debugfs.c
@@ -0,0 +1,87 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <drm/drm_print.h>
+
+#include "gt/intel_gt_debugfs.h"
+#include "gt/uc/intel_guc_ads.h"
+#include "gt/uc/intel_guc_ct.h"
+#include "gt/uc/intel_guc_slpc.h"
+#include "gt/uc/intel_guc_submission.h"
+#include "intel_guc.h"
+#include "intel_guc_debugfs.h"
+#include "intel_guc_log_debugfs.h"
+
+static int guc_info_show(struct seq_file *m, void *data)
+{
+	struct intel_guc *guc = m->private;
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	if (!intel_guc_is_supported(guc))
+		return -ENODEV;
+
+	intel_guc_load_status(guc, &p);
+	drm_puts(&p, "\n");
+	intel_guc_log_info(&guc->log, &p);
+
+	if (!intel_guc_submission_is_used(guc))
+		return 0;
+
+	intel_guc_ct_print_info(&guc->ct, &p);
+	intel_guc_submission_print_info(guc, &p);
+	intel_guc_ads_print_policy_info(guc, &p);
+
+	return 0;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(guc_info);
+
+static int guc_registered_contexts_show(struct seq_file *m, void *data)
+{
+	struct intel_guc *guc = m->private;
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	if (!intel_guc_submission_is_used(guc))
+		return -ENODEV;
+
+	intel_guc_submission_print_context_info(guc, &p);
+
+	return 0;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(guc_registered_contexts);
+
+static int guc_slpc_info_show(struct seq_file *m, void *unused)
+{
+	struct intel_guc *guc = m->private;
+	struct intel_guc_slpc *slpc = &guc->slpc;
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	if (!intel_guc_slpc_is_used(guc))
+		return -ENODEV;
+
+	return intel_guc_slpc_print_info(slpc, &p);
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(guc_slpc_info);
+
+static bool intel_eval_slpc_support(void *data)
+{
+	struct intel_guc *guc = (struct intel_guc *)data;
+
+	return intel_guc_slpc_is_used(guc);
+}
+
+void intel_guc_debugfs_register(struct intel_guc *guc, struct dentry *root)
+{
+	static const struct intel_gt_debugfs_file files[] = {
+		{ "guc_info", &guc_info_fops, NULL },
+		{ "guc_registered_contexts", &guc_registered_contexts_fops, NULL },
+		{ "guc_slpc_info", &guc_slpc_info_fops, &intel_eval_slpc_support},
+	};
+
+	if (!intel_guc_is_supported(guc))
+		return;
+
+	intel_gt_debugfs_register_files(root, files, ARRAY_SIZE(files), guc);
+	intel_guc_log_debugfs_register(&guc->log, root);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_debugfs.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_debugfs.h
new file mode 100644
index 000000000..424c26665
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_debugfs.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef DEBUGFS_GUC_H
+#define DEBUGFS_GUC_H
+
+struct intel_guc;
+struct dentry;
+
+void intel_guc_debugfs_register(struct intel_guc *guc, struct dentry *root);
+
+#endif /* DEBUGFS_GUC_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c
new file mode 100644
index 000000000..a0372735c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.c
@@ -0,0 +1,203 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ *
+ * Authors:
+ *    Vinit Azad <vinit.azad@intel.com>
+ *    Ben Widawsky <ben@bwidawsk.net>
+ *    Dave Gordon <david.s.gordon@intel.com>
+ *    Alex Dai <yu.dai@intel.com>
+ */
+
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_regs.h"
+#include "intel_guc_fw.h"
+#include "i915_drv.h"
+
+static void guc_prepare_xfer(struct intel_uncore *uncore)
+{
+	u32 shim_flags = GUC_ENABLE_READ_CACHE_LOGIC |
+			 GUC_ENABLE_READ_CACHE_FOR_SRAM_DATA |
+			 GUC_ENABLE_READ_CACHE_FOR_WOPCM_DATA |
+			 GUC_ENABLE_MIA_CLOCK_GATING;
+
+	if (GRAPHICS_VER_FULL(uncore->i915) < IP_VER(12, 50))
+		shim_flags |= GUC_DISABLE_SRAM_INIT_TO_ZEROES |
+			      GUC_ENABLE_MIA_CACHING;
+
+	/* Must program this register before loading the ucode with DMA */
+	intel_uncore_write(uncore, GUC_SHIM_CONTROL, shim_flags);
+
+	if (IS_GEN9_LP(uncore->i915))
+		intel_uncore_write(uncore, GEN9LP_GT_PM_CONFIG, GT_DOORBELL_ENABLE);
+	else
+		intel_uncore_write(uncore, GEN9_GT_PM_CONFIG, GT_DOORBELL_ENABLE);
+
+	if (GRAPHICS_VER(uncore->i915) == 9) {
+		/* DOP Clock Gating Enable for GuC clocks */
+		intel_uncore_rmw(uncore, GEN7_MISCCPCTL,
+				 0, GEN8_DOP_CLOCK_GATE_GUC_ENABLE);
+
+		/* allows for 5us (in 10ns units) before GT can go to RC6 */
+		intel_uncore_write(uncore, GUC_ARAT_C6DIS, 0x1FF);
+	}
+}
+
+static int guc_xfer_rsa_mmio(struct intel_uc_fw *guc_fw,
+			     struct intel_uncore *uncore)
+{
+	u32 rsa[UOS_RSA_SCRATCH_COUNT];
+	size_t copied;
+	int i;
+
+	copied = intel_uc_fw_copy_rsa(guc_fw, rsa, sizeof(rsa));
+	if (copied < sizeof(rsa))
+		return -ENOMEM;
+
+	for (i = 0; i < UOS_RSA_SCRATCH_COUNT; i++)
+		intel_uncore_write(uncore, UOS_RSA_SCRATCH(i), rsa[i]);
+
+	return 0;
+}
+
+static int guc_xfer_rsa_vma(struct intel_uc_fw *guc_fw,
+			    struct intel_uncore *uncore)
+{
+	struct intel_guc *guc = container_of(guc_fw, struct intel_guc, fw);
+
+	intel_uncore_write(uncore, UOS_RSA_SCRATCH(0),
+			   intel_guc_ggtt_offset(guc, guc_fw->rsa_data));
+
+	return 0;
+}
+
+/* Copy RSA signature from the fw image to HW for verification */
+static int guc_xfer_rsa(struct intel_uc_fw *guc_fw,
+			struct intel_uncore *uncore)
+{
+	if (guc_fw->rsa_data)
+		return guc_xfer_rsa_vma(guc_fw, uncore);
+	else
+		return guc_xfer_rsa_mmio(guc_fw, uncore);
+}
+
+/*
+ * Read the GuC status register (GUC_STATUS) and store it in the
+ * specified location; then return a boolean indicating whether
+ * the value matches either of two values representing completion
+ * of the GuC boot process.
+ *
+ * This is used for polling the GuC status in a wait_for()
+ * loop below.
+ */
+static inline bool guc_ready(struct intel_uncore *uncore, u32 *status)
+{
+	u32 val = intel_uncore_read(uncore, GUC_STATUS);
+	u32 uk_val = REG_FIELD_GET(GS_UKERNEL_MASK, val);
+
+	*status = val;
+	return uk_val == INTEL_GUC_LOAD_STATUS_READY;
+}
+
+static int guc_wait_ucode(struct intel_uncore *uncore)
+{
+	u32 status;
+	int ret;
+
+	/*
+	 * Wait for the GuC to start up.
+	 * NB: Docs recommend not using the interrupt for completion.
+	 * Measurements indicate this should take no more than 20ms
+	 * (assuming the GT clock is at maximum frequency). So, a
+	 * timeout here indicates that the GuC has failed and is unusable.
+	 * (Higher levels of the driver may decide to reset the GuC and
+	 * attempt the ucode load again if this happens.)
+	 *
+	 * FIXME: There is a known (but exceedingly unlikely) race condition
+	 * where the asynchronous frequency management code could reduce
+	 * the GT clock while a GuC reload is in progress (during a full
+	 * GT reset). A fix is in progress but there are complex locking
+	 * issues to be resolved. In the meantime bump the timeout to
+	 * 200ms. Even at slowest clock, this should be sufficient. And
+	 * in the working case, a larger timeout makes no difference.
+	 */
+	ret = wait_for(guc_ready(uncore, &status), 200);
+	if (ret) {
+		struct drm_device *drm = &uncore->i915->drm;
+
+		drm_info(drm, "GuC load failed: status = 0x%08X\n", status);
+		drm_info(drm, "GuC load failed: status: Reset = %d, "
+			"BootROM = 0x%02X, UKernel = 0x%02X, "
+			"MIA = 0x%02X, Auth = 0x%02X\n",
+			REG_FIELD_GET(GS_MIA_IN_RESET, status),
+			REG_FIELD_GET(GS_BOOTROM_MASK, status),
+			REG_FIELD_GET(GS_UKERNEL_MASK, status),
+			REG_FIELD_GET(GS_MIA_MASK, status),
+			REG_FIELD_GET(GS_AUTH_STATUS_MASK, status));
+
+		if ((status & GS_BOOTROM_MASK) == GS_BOOTROM_RSA_FAILED) {
+			drm_info(drm, "GuC firmware signature verification failed\n");
+			ret = -ENOEXEC;
+		}
+
+		if (REG_FIELD_GET(GS_UKERNEL_MASK, status) == INTEL_GUC_LOAD_STATUS_EXCEPTION) {
+			drm_info(drm, "GuC firmware exception. EIP: %#x\n",
+				 intel_uncore_read(uncore, SOFT_SCRATCH(13)));
+			ret = -ENXIO;
+		}
+	}
+
+	return ret;
+}
+
+/**
+ * intel_guc_fw_upload() - load GuC uCode to device
+ * @guc: intel_guc structure
+ *
+ * Called from intel_uc_init_hw() during driver load, resume from sleep and
+ * after a GPU reset.
+ *
+ * The firmware image should have already been fetched into memory, so only
+ * check that fetch succeeded, and then transfer the image to the h/w.
+ *
+ * Return:	non-zero code on error
+ */
+int intel_guc_fw_upload(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct intel_uncore *uncore = gt->uncore;
+	int ret;
+
+	guc_prepare_xfer(uncore);
+
+	/*
+	 * Note that GuC needs the CSS header plus uKernel code to be copied
+	 * by the DMA engine in one operation, whereas the RSA signature is
+	 * loaded separately, either by copying it to the UOS_RSA_SCRATCH
+	 * register (if key size <= 256) or through a ggtt-pinned vma (if key
+	 * size > 256). The RSA size and therefore the way we provide it to the
+	 * HW is fixed for each platform and hard-coded in the bootrom.
+	 */
+	ret = guc_xfer_rsa(&guc->fw, uncore);
+	if (ret)
+		goto out;
+
+	/*
+	 * Current uCode expects the code to be loaded at 8k; locations below
+	 * this are used for the stack.
+	 */
+	ret = intel_uc_fw_upload(&guc->fw, 0x2000, UOS_MOVE);
+	if (ret)
+		goto out;
+
+	ret = guc_wait_ucode(uncore);
+	if (ret)
+		goto out;
+
+	intel_uc_fw_change_status(&guc->fw, INTEL_UC_FIRMWARE_RUNNING);
+	return 0;
+
+out:
+	intel_uc_fw_change_status(&guc->fw, INTEL_UC_FIRMWARE_LOAD_FAIL);
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.h
new file mode 100644
index 000000000..0b4d2a9c9
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_fw.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2017-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_FW_H_
+#define _INTEL_GUC_FW_H_
+
+struct intel_guc;
+
+int intel_guc_fw_upload(struct intel_guc *guc);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h
new file mode 100644
index 000000000..502e7cb5a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h
@@ -0,0 +1,500 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_FWIF_H
+#define _INTEL_GUC_FWIF_H
+
+#include <linux/bits.h>
+#include <linux/compiler.h>
+#include <linux/types.h>
+#include "gt/intel_engine_types.h"
+
+#include "abi/guc_actions_abi.h"
+#include "abi/guc_actions_slpc_abi.h"
+#include "abi/guc_errors_abi.h"
+#include "abi/guc_communication_mmio_abi.h"
+#include "abi/guc_communication_ctb_abi.h"
+#include "abi/guc_klvs_abi.h"
+#include "abi/guc_messages_abi.h"
+
+/* Payload length only i.e. don't include G2H header length */
+#define G2H_LEN_DW_SCHED_CONTEXT_MODE_SET	2
+#define G2H_LEN_DW_DEREGISTER_CONTEXT		1
+
+#define GUC_CONTEXT_DISABLE		0
+#define GUC_CONTEXT_ENABLE		1
+
+#define GUC_CLIENT_PRIORITY_KMD_HIGH	0
+#define GUC_CLIENT_PRIORITY_HIGH	1
+#define GUC_CLIENT_PRIORITY_KMD_NORMAL	2
+#define GUC_CLIENT_PRIORITY_NORMAL	3
+#define GUC_CLIENT_PRIORITY_NUM		4
+
+#define GUC_MAX_CONTEXT_ID		65535
+#define	GUC_INVALID_CONTEXT_ID		GUC_MAX_CONTEXT_ID
+
+#define GUC_RENDER_ENGINE		0
+#define GUC_VIDEO_ENGINE		1
+#define GUC_BLITTER_ENGINE		2
+#define GUC_VIDEOENHANCE_ENGINE		3
+#define GUC_VIDEO_ENGINE2		4
+#define GUC_MAX_ENGINES_NUM		(GUC_VIDEO_ENGINE2 + 1)
+
+#define GUC_RENDER_CLASS		0
+#define GUC_VIDEO_CLASS			1
+#define GUC_VIDEOENHANCE_CLASS		2
+#define GUC_BLITTER_CLASS		3
+#define GUC_COMPUTE_CLASS		4
+#define GUC_LAST_ENGINE_CLASS		GUC_COMPUTE_CLASS
+#define GUC_MAX_ENGINE_CLASSES		16
+#define GUC_MAX_INSTANCES_PER_CLASS	32
+
+#define GUC_DOORBELL_INVALID		256
+
+/*
+ * Work queue item header definitions
+ *
+ * Work queue is circular buffer used to submit complex (multi-lrc) submissions
+ * to the GuC. A work queue item is an entry in the circular buffer.
+ */
+#define WQ_STATUS_ACTIVE		1
+#define WQ_STATUS_SUSPENDED		2
+#define WQ_STATUS_CMD_ERROR		3
+#define WQ_STATUS_ENGINE_ID_NOT_USED	4
+#define WQ_STATUS_SUSPENDED_FROM_RESET	5
+#define WQ_TYPE_BATCH_BUF		0x1
+#define WQ_TYPE_PSEUDO			0x2
+#define WQ_TYPE_INORDER			0x3
+#define WQ_TYPE_NOOP			0x4
+#define WQ_TYPE_MULTI_LRC		0x5
+#define WQ_TYPE_MASK			GENMASK(7, 0)
+#define WQ_LEN_MASK			GENMASK(26, 16)
+
+#define WQ_GUC_ID_MASK			GENMASK(15, 0)
+#define WQ_RING_TAIL_MASK		GENMASK(28, 18)
+
+#define GUC_STAGE_DESC_ATTR_ACTIVE	BIT(0)
+#define GUC_STAGE_DESC_ATTR_PENDING_DB	BIT(1)
+#define GUC_STAGE_DESC_ATTR_KERNEL	BIT(2)
+#define GUC_STAGE_DESC_ATTR_PREEMPT	BIT(3)
+#define GUC_STAGE_DESC_ATTR_RESET	BIT(4)
+#define GUC_STAGE_DESC_ATTR_WQLOCKED	BIT(5)
+#define GUC_STAGE_DESC_ATTR_PCH		BIT(6)
+#define GUC_STAGE_DESC_ATTR_TERMINATED	BIT(7)
+
+#define GUC_CTL_LOG_PARAMS		0
+#define   GUC_LOG_VALID			BIT(0)
+#define   GUC_LOG_NOTIFY_ON_HALF_FULL	BIT(1)
+#define   GUC_LOG_CAPTURE_ALLOC_UNITS	BIT(2)
+#define   GUC_LOG_LOG_ALLOC_UNITS	BIT(3)
+#define   GUC_LOG_CRASH_SHIFT		4
+#define   GUC_LOG_CRASH_MASK		(0x3 << GUC_LOG_CRASH_SHIFT)
+#define   GUC_LOG_DEBUG_SHIFT		6
+#define   GUC_LOG_DEBUG_MASK	        (0xF << GUC_LOG_DEBUG_SHIFT)
+#define   GUC_LOG_CAPTURE_SHIFT		10
+#define   GUC_LOG_CAPTURE_MASK	        (0x3 << GUC_LOG_CAPTURE_SHIFT)
+#define   GUC_LOG_BUF_ADDR_SHIFT	12
+
+#define GUC_CTL_WA			1
+#define   GUC_WA_GAM_CREDITS		BIT(10)
+#define   GUC_WA_DUAL_QUEUE		BIT(11)
+#define   GUC_WA_RCS_RESET_BEFORE_RC6	BIT(13)
+#define   GUC_WA_CONTEXT_ISOLATION	BIT(15)
+#define   GUC_WA_PRE_PARSER		BIT(14)
+#define   GUC_WA_HOLD_CCS_SWITCHOUT	BIT(17)
+#define   GUC_WA_POLLCS			BIT(18)
+#define   GUC_WA_RCS_REGS_IN_CCS_REGS_LIST	BIT(21)
+
+#define GUC_CTL_FEATURE			2
+#define   GUC_CTL_ENABLE_SLPC		BIT(2)
+#define   GUC_CTL_DISABLE_SCHEDULER	BIT(14)
+
+#define GUC_CTL_DEBUG			3
+#define   GUC_LOG_VERBOSITY_SHIFT	0
+#define   GUC_LOG_VERBOSITY_LOW		(0 << GUC_LOG_VERBOSITY_SHIFT)
+#define   GUC_LOG_VERBOSITY_MED		(1 << GUC_LOG_VERBOSITY_SHIFT)
+#define   GUC_LOG_VERBOSITY_HIGH	(2 << GUC_LOG_VERBOSITY_SHIFT)
+#define   GUC_LOG_VERBOSITY_ULTRA	(3 << GUC_LOG_VERBOSITY_SHIFT)
+/* Verbosity range-check limits, without the shift */
+#define	  GUC_LOG_VERBOSITY_MIN		0
+#define	  GUC_LOG_VERBOSITY_MAX		3
+#define	  GUC_LOG_VERBOSITY_MASK	0x0000000f
+#define	  GUC_LOG_DESTINATION_MASK	(3 << 4)
+#define   GUC_LOG_DISABLED		(1 << 6)
+#define   GUC_PROFILE_ENABLED		(1 << 7)
+
+#define GUC_CTL_ADS			4
+#define   GUC_ADS_ADDR_SHIFT		1
+#define   GUC_ADS_ADDR_MASK		(0xFFFFF << GUC_ADS_ADDR_SHIFT)
+
+#define GUC_CTL_DEVID			5
+
+#define GUC_CTL_MAX_DWORDS		(SOFT_SCRATCH_COUNT - 2) /* [1..14] */
+
+/* Generic GT SysInfo data types */
+#define GUC_GENERIC_GT_SYSINFO_SLICE_ENABLED		0
+#define GUC_GENERIC_GT_SYSINFO_VDBOX_SFC_SUPPORT_MASK	1
+#define GUC_GENERIC_GT_SYSINFO_DOORBELL_COUNT_PER_SQIDI	2
+#define GUC_GENERIC_GT_SYSINFO_MAX			16
+
+/*
+ * The class goes in bits [0..2] of the GuC ID, the instance in bits [3..6].
+ * Bit 7 can be used for operations that apply to all engine classes&instances.
+ */
+#define GUC_ENGINE_CLASS_SHIFT		0
+#define GUC_ENGINE_CLASS_MASK		(0x7 << GUC_ENGINE_CLASS_SHIFT)
+#define GUC_ENGINE_INSTANCE_SHIFT	3
+#define GUC_ENGINE_INSTANCE_MASK	(0xf << GUC_ENGINE_INSTANCE_SHIFT)
+#define GUC_ENGINE_ALL_INSTANCES	BIT(7)
+
+#define MAKE_GUC_ID(class, instance) \
+	(((class) << GUC_ENGINE_CLASS_SHIFT) | \
+	 ((instance) << GUC_ENGINE_INSTANCE_SHIFT))
+
+#define GUC_ID_TO_ENGINE_CLASS(guc_id) \
+	(((guc_id) & GUC_ENGINE_CLASS_MASK) >> GUC_ENGINE_CLASS_SHIFT)
+#define GUC_ID_TO_ENGINE_INSTANCE(guc_id) \
+	(((guc_id) & GUC_ENGINE_INSTANCE_MASK) >> GUC_ENGINE_INSTANCE_SHIFT)
+
+#define SLPC_EVENT(id, c) (\
+FIELD_PREP(HOST2GUC_PC_SLPC_REQUEST_MSG_1_EVENT_ID, id) | \
+FIELD_PREP(HOST2GUC_PC_SLPC_REQUEST_MSG_1_EVENT_ARGC, c) \
+)
+
+/* the GuC arrays don't include OTHER_CLASS */
+static u8 engine_class_guc_class_map[] = {
+	[RENDER_CLASS]            = GUC_RENDER_CLASS,
+	[COPY_ENGINE_CLASS]       = GUC_BLITTER_CLASS,
+	[VIDEO_DECODE_CLASS]      = GUC_VIDEO_CLASS,
+	[VIDEO_ENHANCEMENT_CLASS] = GUC_VIDEOENHANCE_CLASS,
+	[COMPUTE_CLASS]           = GUC_COMPUTE_CLASS,
+};
+
+static u8 guc_class_engine_class_map[] = {
+	[GUC_RENDER_CLASS]       = RENDER_CLASS,
+	[GUC_BLITTER_CLASS]      = COPY_ENGINE_CLASS,
+	[GUC_VIDEO_CLASS]        = VIDEO_DECODE_CLASS,
+	[GUC_VIDEOENHANCE_CLASS] = VIDEO_ENHANCEMENT_CLASS,
+	[GUC_COMPUTE_CLASS]      = COMPUTE_CLASS,
+};
+
+static inline u8 engine_class_to_guc_class(u8 class)
+{
+	BUILD_BUG_ON(ARRAY_SIZE(engine_class_guc_class_map) != MAX_ENGINE_CLASS + 1);
+	GEM_BUG_ON(class > MAX_ENGINE_CLASS || class == OTHER_CLASS);
+
+	return engine_class_guc_class_map[class];
+}
+
+static inline u8 guc_class_to_engine_class(u8 guc_class)
+{
+	BUILD_BUG_ON(ARRAY_SIZE(guc_class_engine_class_map) != GUC_LAST_ENGINE_CLASS + 1);
+	GEM_BUG_ON(guc_class > GUC_LAST_ENGINE_CLASS);
+
+	return guc_class_engine_class_map[guc_class];
+}
+
+/* Work item for submitting workloads into work queue of GuC. */
+struct guc_wq_item {
+	u32 header;
+	u32 context_desc;
+	u32 submit_element_info;
+	u32 fence_id;
+} __packed;
+
+struct guc_process_desc_v69 {
+	u32 stage_id;
+	u64 db_base_addr;
+	u32 head;
+	u32 tail;
+	u32 error_offset;
+	u64 wq_base_addr;
+	u32 wq_size_bytes;
+	u32 wq_status;
+	u32 engine_presence;
+	u32 priority;
+	u32 reserved[36];
+} __packed;
+
+struct guc_sched_wq_desc {
+	u32 head;
+	u32 tail;
+	u32 error_offset;
+	u32 wq_status;
+	u32 reserved[28];
+} __packed;
+
+/* Helper for context registration H2G */
+struct guc_ctxt_registration_info {
+	u32 flags;
+	u32 context_idx;
+	u32 engine_class;
+	u32 engine_submit_mask;
+	u32 wq_desc_lo;
+	u32 wq_desc_hi;
+	u32 wq_base_lo;
+	u32 wq_base_hi;
+	u32 wq_size;
+	u32 hwlrca_lo;
+	u32 hwlrca_hi;
+};
+#define CONTEXT_REGISTRATION_FLAG_KMD	BIT(0)
+
+/* Preempt to idle on quantum expiry */
+#define CONTEXT_POLICY_FLAG_PREEMPT_TO_IDLE_V69	BIT(0)
+
+/*
+ * GuC Context registration descriptor.
+ * FIXME: This is only required to exist during context registration.
+ * The current 1:1 between guc_lrc_desc and LRCs for the lifetime of the LRC
+ * is not required.
+ */
+struct guc_lrc_desc_v69 {
+	u32 hw_context_desc;
+	u32 slpm_perf_mode_hint;	/* SPLC v1 only */
+	u32 slpm_freq_hint;
+	u32 engine_submit_mask;		/* In logical space */
+	u8 engine_class;
+	u8 reserved0[3];
+	u32 priority;
+	u32 process_desc;
+	u32 wq_addr;
+	u32 wq_size;
+	u32 context_flags;		/* CONTEXT_REGISTRATION_* */
+	/* Time for one workload to execute. (in micro seconds) */
+	u32 execution_quantum;
+	/* Time to wait for a preemption request to complete before issuing a
+	 * reset. (in micro seconds).
+	 */
+	u32 preemption_timeout;
+	u32 policy_flags;		/* CONTEXT_POLICY_* */
+	u32 reserved1[19];
+} __packed;
+
+/* 32-bit KLV structure as used by policy updates and others */
+struct guc_klv_generic_dw_t {
+	u32 kl;
+	u32 value;
+} __packed;
+
+/* Format of the UPDATE_CONTEXT_POLICIES H2G data packet */
+struct guc_update_context_policy_header {
+	u32 action;
+	u32 ctx_id;
+} __packed;
+
+struct guc_update_context_policy {
+	struct guc_update_context_policy_header header;
+	struct guc_klv_generic_dw_t klv[GUC_CONTEXT_POLICIES_KLV_NUM_IDS];
+} __packed;
+
+#define GUC_POWER_UNSPECIFIED	0
+#define GUC_POWER_D0		1
+#define GUC_POWER_D1		2
+#define GUC_POWER_D2		3
+#define GUC_POWER_D3		4
+
+/* Scheduling policy settings */
+
+#define GLOBAL_POLICY_MAX_NUM_WI 15
+
+/* Don't reset an engine upon preemption failure */
+#define GLOBAL_POLICY_DISABLE_ENGINE_RESET				BIT(0)
+
+#define GLOBAL_POLICY_DEFAULT_DPC_PROMOTE_TIME_US 500000
+
+/*
+ * GuC converts the timeout to clock ticks internally. Different platforms have
+ * different GuC clocks. Thus, the maximum value before overflow is platform
+ * dependent. Current worst case scenario is about 110s. So, the spec says to
+ * limit to 100s to be safe.
+ */
+#define GUC_POLICY_MAX_EXEC_QUANTUM_US		(100 * 1000 * 1000UL)
+#define GUC_POLICY_MAX_PREEMPT_TIMEOUT_US	(100 * 1000 * 1000UL)
+
+static inline u32 guc_policy_max_exec_quantum_ms(void)
+{
+	BUILD_BUG_ON(GUC_POLICY_MAX_EXEC_QUANTUM_US >= UINT_MAX);
+	return GUC_POLICY_MAX_EXEC_QUANTUM_US / 1000;
+}
+
+static inline u32 guc_policy_max_preempt_timeout_ms(void)
+{
+	BUILD_BUG_ON(GUC_POLICY_MAX_PREEMPT_TIMEOUT_US >= UINT_MAX);
+	return GUC_POLICY_MAX_PREEMPT_TIMEOUT_US / 1000;
+}
+
+struct guc_policies {
+	u32 submission_queue_depth[GUC_MAX_ENGINE_CLASSES];
+	/* In micro seconds. How much time to allow before DPC processing is
+	 * called back via interrupt (to prevent DPC queue drain starving).
+	 * Typically 1000s of micro seconds (example only, not granularity). */
+	u32 dpc_promote_time;
+
+	/* Must be set to take these new values. */
+	u32 is_valid;
+
+	/* Max number of WIs to process per call. A large value may keep CS
+	 * idle. */
+	u32 max_num_work_items;
+
+	u32 global_flags;
+	u32 reserved[4];
+} __packed;
+
+/* GuC MMIO reg state struct */
+struct guc_mmio_reg {
+	u32 offset;
+	u32 value;
+	u32 flags;
+#define GUC_REGSET_MASKED		BIT(0)
+#define GUC_REGSET_NEEDS_STEERING	BIT(1)
+#define GUC_REGSET_MASKED_WITH_VALUE	BIT(2)
+#define GUC_REGSET_RESTORE_ONLY		BIT(3)
+#define GUC_REGSET_STEERING_GROUP       GENMASK(15, 12)
+#define GUC_REGSET_STEERING_INSTANCE    GENMASK(23, 20)
+	u32 mask;
+} __packed;
+
+/* GuC register sets */
+struct guc_mmio_reg_set {
+	u32 address;
+	u16 count;
+	u16 reserved;
+} __packed;
+
+/* HW info */
+struct guc_gt_system_info {
+	u8 mapping_table[GUC_MAX_ENGINE_CLASSES][GUC_MAX_INSTANCES_PER_CLASS];
+	u32 engine_enabled_masks[GUC_MAX_ENGINE_CLASSES];
+	u32 generic_gt_sysinfo[GUC_GENERIC_GT_SYSINFO_MAX];
+} __packed;
+
+enum {
+	GUC_CAPTURE_LIST_INDEX_PF = 0,
+	GUC_CAPTURE_LIST_INDEX_VF = 1,
+	GUC_CAPTURE_LIST_INDEX_MAX = 2,
+};
+
+/*Register-types of GuC capture register lists */
+enum guc_capture_type {
+	GUC_CAPTURE_LIST_TYPE_GLOBAL = 0,
+	GUC_CAPTURE_LIST_TYPE_ENGINE_CLASS,
+	GUC_CAPTURE_LIST_TYPE_ENGINE_INSTANCE,
+	GUC_CAPTURE_LIST_TYPE_MAX,
+};
+
+/* GuC Additional Data Struct */
+struct guc_ads {
+	struct guc_mmio_reg_set reg_state_list[GUC_MAX_ENGINE_CLASSES][GUC_MAX_INSTANCES_PER_CLASS];
+	u32 reserved0;
+	u32 scheduler_policies;
+	u32 gt_system_info;
+	u32 reserved1;
+	u32 control_data;
+	u32 golden_context_lrca[GUC_MAX_ENGINE_CLASSES];
+	u32 eng_state_size[GUC_MAX_ENGINE_CLASSES];
+	u32 private_data;
+	u32 reserved2;
+	u32 capture_instance[GUC_CAPTURE_LIST_INDEX_MAX][GUC_MAX_ENGINE_CLASSES];
+	u32 capture_class[GUC_CAPTURE_LIST_INDEX_MAX][GUC_MAX_ENGINE_CLASSES];
+	u32 capture_global[GUC_CAPTURE_LIST_INDEX_MAX];
+	u32 reserved[14];
+} __packed;
+
+/* Engine usage stats */
+struct guc_engine_usage_record {
+	u32 current_context_index;
+	u32 last_switch_in_stamp;
+	u32 reserved0;
+	u32 total_runtime;
+	u32 reserved1[4];
+} __packed;
+
+struct guc_engine_usage {
+	struct guc_engine_usage_record engines[GUC_MAX_ENGINE_CLASSES][GUC_MAX_INSTANCES_PER_CLASS];
+} __packed;
+
+/* GuC logging structures */
+
+enum guc_log_buffer_type {
+	GUC_DEBUG_LOG_BUFFER,
+	GUC_CRASH_DUMP_LOG_BUFFER,
+	GUC_CAPTURE_LOG_BUFFER,
+	GUC_MAX_LOG_BUFFER
+};
+
+/**
+ * struct guc_log_buffer_state - GuC log buffer state
+ *
+ * Below state structure is used for coordination of retrieval of GuC firmware
+ * logs. Separate state is maintained for each log buffer type.
+ * read_ptr points to the location where i915 read last in log buffer and
+ * is read only for GuC firmware. write_ptr is incremented by GuC with number
+ * of bytes written for each log entry and is read only for i915.
+ * When any type of log buffer becomes half full, GuC sends a flush interrupt.
+ * GuC firmware expects that while it is writing to 2nd half of the buffer,
+ * first half would get consumed by Host and then get a flush completed
+ * acknowledgment from Host, so that it does not end up doing any overwrite
+ * causing loss of logs. So when buffer gets half filled & i915 has requested
+ * for interrupt, GuC will set flush_to_file field, set the sampled_write_ptr
+ * to the value of write_ptr and raise the interrupt.
+ * On receiving the interrupt i915 should read the buffer, clear flush_to_file
+ * field and also update read_ptr with the value of sample_write_ptr, before
+ * sending an acknowledgment to GuC. marker & version fields are for internal
+ * usage of GuC and opaque to i915. buffer_full_cnt field is incremented every
+ * time GuC detects the log buffer overflow.
+ */
+struct guc_log_buffer_state {
+	u32 marker[2];
+	u32 read_ptr;
+	u32 write_ptr;
+	u32 size;
+	u32 sampled_write_ptr;
+	u32 wrap_offset;
+	union {
+		struct {
+			u32 flush_to_file:1;
+			u32 buffer_full_cnt:4;
+			u32 reserved:27;
+		};
+		u32 flags;
+	};
+	u32 version;
+} __packed;
+
+struct guc_ctx_report {
+	u32 report_return_status;
+	u32 reserved1[64];
+	u32 affected_count;
+	u32 reserved2[2];
+} __packed;
+
+/* GuC Shared Context Data Struct */
+struct guc_shared_ctx_data {
+	u32 addr_of_last_preempted_data_low;
+	u32 addr_of_last_preempted_data_high;
+	u32 addr_of_last_preempted_data_high_tmp;
+	u32 padding;
+	u32 is_mapped_to_proxy;
+	u32 proxy_ctx_id;
+	u32 engine_reset_ctx_id;
+	u32 media_reset_count;
+	u32 reserved1[8];
+	u32 uk_last_ctx_switch_reason;
+	u32 was_reset;
+	u32 lrca_gpu_addr;
+	u64 execlist_ctx;
+	u32 reserved2[66];
+	struct guc_ctx_report preempt_ctx_report[GUC_MAX_ENGINES_NUM];
+} __packed;
+
+/* This action will be programmed in C1BC - SOFT_SCRATCH_15_REG */
+enum intel_guc_recv_message {
+	INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED = BIT(1),
+	INTEL_GUC_RECV_MSG_EXCEPTION = BIT(30),
+};
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_hwconfig.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_hwconfig.c
new file mode 100644
index 000000000..4781fccc2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_hwconfig.c
@@ -0,0 +1,164 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include "gt/intel_gt.h"
+#include "gt/intel_hwconfig.h"
+#include "i915_drv.h"
+#include "i915_memcpy.h"
+
+/*
+ * GuC has a blob containing hardware configuration information (HWConfig).
+ * This is formatted as a simple and flexible KLV (Key/Length/Value) table.
+ *
+ * For example, a minimal version could be:
+ *   enum device_attr {
+ *     ATTR_SOME_VALUE = 0,
+ *     ATTR_SOME_MASK  = 1,
+ *   };
+ *
+ *   static const u32 hwconfig[] = {
+ *     ATTR_SOME_VALUE,
+ *     1,		// Value Length in DWords
+ *     8,		// Value
+ *
+ *     ATTR_SOME_MASK,
+ *     3,
+ *     0x00FFFFFFFF, 0xFFFFFFFF, 0xFF000000,
+ *   };
+ *
+ * The attribute ids are defined in a hardware spec.
+ */
+
+static int __guc_action_get_hwconfig(struct intel_guc *guc,
+				     u32 ggtt_offset, u32 ggtt_size)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_GET_HWCONFIG,
+		lower_32_bits(ggtt_offset),
+		upper_32_bits(ggtt_offset),
+		ggtt_size,
+	};
+	int ret;
+
+	ret = intel_guc_send_mmio(guc, action, ARRAY_SIZE(action), NULL, 0);
+	if (ret == -ENXIO)
+		return -ENOENT;
+
+	return ret;
+}
+
+static int guc_hwconfig_discover_size(struct intel_guc *guc, struct intel_hwconfig *hwconfig)
+{
+	int ret;
+
+	/*
+	 * Sending a query with zero offset and size will return the
+	 * size of the blob.
+	 */
+	ret = __guc_action_get_hwconfig(guc, 0, 0);
+	if (ret < 0)
+		return ret;
+
+	if (ret == 0)
+		return -EINVAL;
+
+	hwconfig->size = ret;
+	return 0;
+}
+
+static int guc_hwconfig_fill_buffer(struct intel_guc *guc, struct intel_hwconfig *hwconfig)
+{
+	struct i915_vma *vma;
+	u32 ggtt_offset;
+	void *vaddr;
+	int ret;
+
+	GEM_BUG_ON(!hwconfig->size);
+
+	ret = intel_guc_allocate_and_map_vma(guc, hwconfig->size, &vma, &vaddr);
+	if (ret)
+		return ret;
+
+	ggtt_offset = intel_guc_ggtt_offset(guc, vma);
+
+	ret = __guc_action_get_hwconfig(guc, ggtt_offset, hwconfig->size);
+	if (ret >= 0)
+		memcpy(hwconfig->ptr, vaddr, hwconfig->size);
+
+	i915_vma_unpin_and_release(&vma, I915_VMA_RELEASE_MAP);
+
+	return ret;
+}
+
+static bool has_table(struct drm_i915_private *i915)
+{
+	if (IS_ALDERLAKE_P(i915) && !IS_ADLP_N(i915))
+		return true;
+	if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 55))
+		return true;
+
+	return false;
+}
+
+/**
+ * intel_guc_hwconfig_init - Initialize the HWConfig
+ *
+ * Retrieve the HWConfig table from the GuC and save it locally.
+ * It can then be queried on demand by other users later on.
+ */
+static int guc_hwconfig_init(struct intel_gt *gt)
+{
+	struct intel_hwconfig *hwconfig = &gt->info.hwconfig;
+	struct intel_guc *guc = &gt->uc.guc;
+	int ret;
+
+	if (!has_table(gt->i915))
+		return 0;
+
+	ret = guc_hwconfig_discover_size(guc, hwconfig);
+	if (ret)
+		return ret;
+
+	hwconfig->ptr = kmalloc(hwconfig->size, GFP_KERNEL);
+	if (!hwconfig->ptr) {
+		hwconfig->size = 0;
+		return -ENOMEM;
+	}
+
+	ret = guc_hwconfig_fill_buffer(guc, hwconfig);
+	if (ret < 0) {
+		intel_gt_fini_hwconfig(gt);
+		return ret;
+	}
+
+	return 0;
+}
+
+/**
+ * intel_gt_init_hwconfig - Initialize the HWConfig if available
+ *
+ * Retrieve the HWConfig table if available on the current platform.
+ */
+int intel_gt_init_hwconfig(struct intel_gt *gt)
+{
+	if (!intel_uc_uses_guc(&gt->uc))
+		return 0;
+
+	return guc_hwconfig_init(gt);
+}
+
+/**
+ * intel_gt_fini_hwconfig - Finalize the HWConfig
+ *
+ * Free up the memory allocation holding the table.
+ */
+void intel_gt_fini_hwconfig(struct intel_gt *gt)
+{
+	struct intel_hwconfig *hwconfig = &gt->info.hwconfig;
+
+	kfree(hwconfig->ptr);
+	hwconfig->size = 0;
+	hwconfig->ptr = NULL;
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_log.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_log.c
new file mode 100644
index 000000000..68331c538
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_log.c
@@ -0,0 +1,930 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#include <linux/debugfs.h>
+#include <linux/string_helpers.h>
+
+#include "gt/intel_gt.h"
+#include "i915_drv.h"
+#include "i915_irq.h"
+#include "i915_memcpy.h"
+#include "intel_guc_capture.h"
+#include "intel_guc_log.h"
+
+#if defined(CONFIG_DRM_I915_DEBUG_GUC)
+#define GUC_LOG_DEFAULT_CRASH_BUFFER_SIZE	SZ_2M
+#define GUC_LOG_DEFAULT_DEBUG_BUFFER_SIZE	SZ_16M
+#define GUC_LOG_DEFAULT_CAPTURE_BUFFER_SIZE	SZ_1M
+#elif defined(CONFIG_DRM_I915_DEBUG_GEM)
+#define GUC_LOG_DEFAULT_CRASH_BUFFER_SIZE	SZ_1M
+#define GUC_LOG_DEFAULT_DEBUG_BUFFER_SIZE	SZ_2M
+#define GUC_LOG_DEFAULT_CAPTURE_BUFFER_SIZE	SZ_1M
+#else
+#define GUC_LOG_DEFAULT_CRASH_BUFFER_SIZE	SZ_8K
+#define GUC_LOG_DEFAULT_DEBUG_BUFFER_SIZE	SZ_64K
+#define GUC_LOG_DEFAULT_CAPTURE_BUFFER_SIZE	SZ_1M
+#endif
+
+static void guc_log_copy_debuglogs_for_relay(struct intel_guc_log *log);
+
+struct guc_log_section {
+	u32 max;
+	u32 flag;
+	u32 default_val;
+	const char *name;
+};
+
+static void _guc_log_init_sizes(struct intel_guc_log *log)
+{
+	struct intel_guc *guc = log_to_guc(log);
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+	static const struct guc_log_section sections[GUC_LOG_SECTIONS_LIMIT] = {
+		{
+			GUC_LOG_CRASH_MASK >> GUC_LOG_CRASH_SHIFT,
+			GUC_LOG_LOG_ALLOC_UNITS,
+			GUC_LOG_DEFAULT_CRASH_BUFFER_SIZE,
+			"crash dump"
+		},
+		{
+			GUC_LOG_DEBUG_MASK >> GUC_LOG_DEBUG_SHIFT,
+			GUC_LOG_LOG_ALLOC_UNITS,
+			GUC_LOG_DEFAULT_DEBUG_BUFFER_SIZE,
+			"debug",
+		},
+		{
+			GUC_LOG_CAPTURE_MASK >> GUC_LOG_CAPTURE_SHIFT,
+			GUC_LOG_CAPTURE_ALLOC_UNITS,
+			GUC_LOG_DEFAULT_CAPTURE_BUFFER_SIZE,
+			"capture",
+		}
+	};
+	int i;
+
+	for (i = 0; i < GUC_LOG_SECTIONS_LIMIT; i++)
+		log->sizes[i].bytes = sections[i].default_val;
+
+	/* If debug size > 1MB then bump default crash size to keep the same units */
+	if (log->sizes[GUC_LOG_SECTIONS_DEBUG].bytes >= SZ_1M &&
+	    GUC_LOG_DEFAULT_CRASH_BUFFER_SIZE < SZ_1M)
+		log->sizes[GUC_LOG_SECTIONS_CRASH].bytes = SZ_1M;
+
+	/* Prepare the GuC API structure fields: */
+	for (i = 0; i < GUC_LOG_SECTIONS_LIMIT; i++) {
+		/* Convert to correct units */
+		if ((log->sizes[i].bytes % SZ_1M) == 0) {
+			log->sizes[i].units = SZ_1M;
+			log->sizes[i].flag = sections[i].flag;
+		} else {
+			log->sizes[i].units = SZ_4K;
+			log->sizes[i].flag = 0;
+		}
+
+		if (!IS_ALIGNED(log->sizes[i].bytes, log->sizes[i].units))
+			drm_err(&i915->drm, "Mis-aligned GuC log %s size: 0x%X vs 0x%X!",
+				sections[i].name, log->sizes[i].bytes, log->sizes[i].units);
+		log->sizes[i].count = log->sizes[i].bytes / log->sizes[i].units;
+
+		if (!log->sizes[i].count) {
+			drm_err(&i915->drm, "Zero GuC log %s size!", sections[i].name);
+		} else {
+			/* Size is +1 unit */
+			log->sizes[i].count--;
+		}
+
+		/* Clip to field size */
+		if (log->sizes[i].count > sections[i].max) {
+			drm_err(&i915->drm, "GuC log %s size too large: %d vs %d!",
+				sections[i].name, log->sizes[i].count + 1, sections[i].max + 1);
+			log->sizes[i].count = sections[i].max;
+		}
+	}
+
+	if (log->sizes[GUC_LOG_SECTIONS_CRASH].units != log->sizes[GUC_LOG_SECTIONS_DEBUG].units) {
+		drm_err(&i915->drm, "Unit mis-match for GuC log crash and debug sections: %d vs %d!",
+			log->sizes[GUC_LOG_SECTIONS_CRASH].units,
+			log->sizes[GUC_LOG_SECTIONS_DEBUG].units);
+		log->sizes[GUC_LOG_SECTIONS_CRASH].units = log->sizes[GUC_LOG_SECTIONS_DEBUG].units;
+		log->sizes[GUC_LOG_SECTIONS_CRASH].count = 0;
+	}
+
+	log->sizes_initialised = true;
+}
+
+static void guc_log_init_sizes(struct intel_guc_log *log)
+{
+	if (log->sizes_initialised)
+		return;
+
+	_guc_log_init_sizes(log);
+}
+
+static u32 intel_guc_log_section_size_crash(struct intel_guc_log *log)
+{
+	guc_log_init_sizes(log);
+
+	return log->sizes[GUC_LOG_SECTIONS_CRASH].bytes;
+}
+
+static u32 intel_guc_log_section_size_debug(struct intel_guc_log *log)
+{
+	guc_log_init_sizes(log);
+
+	return log->sizes[GUC_LOG_SECTIONS_DEBUG].bytes;
+}
+
+u32 intel_guc_log_section_size_capture(struct intel_guc_log *log)
+{
+	guc_log_init_sizes(log);
+
+	return log->sizes[GUC_LOG_SECTIONS_CAPTURE].bytes;
+}
+
+static u32 intel_guc_log_size(struct intel_guc_log *log)
+{
+	/*
+	 *  GuC Log buffer Layout:
+	 *
+	 *  NB: Ordering must follow "enum guc_log_buffer_type".
+	 *
+	 *  +===============================+ 00B
+	 *  |      Debug state header       |
+	 *  +-------------------------------+ 32B
+	 *  |    Crash dump state header    |
+	 *  +-------------------------------+ 64B
+	 *  |     Capture state header      |
+	 *  +-------------------------------+ 96B
+	 *  |                               |
+	 *  +===============================+ PAGE_SIZE (4KB)
+	 *  |          Debug logs           |
+	 *  +===============================+ + DEBUG_SIZE
+	 *  |        Crash Dump logs        |
+	 *  +===============================+ + CRASH_SIZE
+	 *  |         Capture logs          |
+	 *  +===============================+ + CAPTURE_SIZE
+	 */
+	return PAGE_SIZE +
+		intel_guc_log_section_size_crash(log) +
+		intel_guc_log_section_size_debug(log) +
+		intel_guc_log_section_size_capture(log);
+}
+
+/**
+ * DOC: GuC firmware log
+ *
+ * Firmware log is enabled by setting i915.guc_log_level to the positive level.
+ * Log data is printed out via reading debugfs i915_guc_log_dump. Reading from
+ * i915_guc_load_status will print out firmware loading status and scratch
+ * registers value.
+ */
+
+static int guc_action_flush_log_complete(struct intel_guc *guc)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_LOG_BUFFER_FILE_FLUSH_COMPLETE,
+		GUC_DEBUG_LOG_BUFFER
+	};
+
+	return intel_guc_send_nb(guc, action, ARRAY_SIZE(action), 0);
+}
+
+static int guc_action_flush_log(struct intel_guc *guc)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_FORCE_LOG_BUFFER_FLUSH,
+		0
+	};
+
+	return intel_guc_send(guc, action, ARRAY_SIZE(action));
+}
+
+static int guc_action_control_log(struct intel_guc *guc, bool enable,
+				  bool default_logging, u32 verbosity)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_UK_LOG_ENABLE_LOGGING,
+		(enable ? GUC_LOG_CONTROL_LOGGING_ENABLED : 0) |
+		(verbosity << GUC_LOG_CONTROL_VERBOSITY_SHIFT) |
+		(default_logging ? GUC_LOG_CONTROL_DEFAULT_LOGGING : 0)
+	};
+
+	GEM_BUG_ON(verbosity > GUC_LOG_VERBOSITY_MAX);
+
+	return intel_guc_send(guc, action, ARRAY_SIZE(action));
+}
+
+/*
+ * Sub buffer switch callback. Called whenever relay has to switch to a new
+ * sub buffer, relay stays on the same sub buffer if 0 is returned.
+ */
+static int subbuf_start_callback(struct rchan_buf *buf,
+				 void *subbuf,
+				 void *prev_subbuf,
+				 size_t prev_padding)
+{
+	/*
+	 * Use no-overwrite mode by default, where relay will stop accepting
+	 * new data if there are no empty sub buffers left.
+	 * There is no strict synchronization enforced by relay between Consumer
+	 * and Producer. In overwrite mode, there is a possibility of getting
+	 * inconsistent/garbled data, the producer could be writing on to the
+	 * same sub buffer from which Consumer is reading. This can't be avoided
+	 * unless Consumer is fast enough and can always run in tandem with
+	 * Producer.
+	 */
+	if (relay_buf_full(buf))
+		return 0;
+
+	return 1;
+}
+
+/*
+ * file_create() callback. Creates relay file in debugfs.
+ */
+static struct dentry *create_buf_file_callback(const char *filename,
+					       struct dentry *parent,
+					       umode_t mode,
+					       struct rchan_buf *buf,
+					       int *is_global)
+{
+	struct dentry *buf_file;
+
+	/*
+	 * This to enable the use of a single buffer for the relay channel and
+	 * correspondingly have a single file exposed to User, through which
+	 * it can collect the logs in order without any post-processing.
+	 * Need to set 'is_global' even if parent is NULL for early logging.
+	 */
+	*is_global = 1;
+
+	if (!parent)
+		return NULL;
+
+	buf_file = debugfs_create_file(filename, mode,
+				       parent, buf, &relay_file_operations);
+	if (IS_ERR(buf_file))
+		return NULL;
+
+	return buf_file;
+}
+
+/*
+ * file_remove() default callback. Removes relay file in debugfs.
+ */
+static int remove_buf_file_callback(struct dentry *dentry)
+{
+	debugfs_remove(dentry);
+	return 0;
+}
+
+/* relay channel callbacks */
+static const struct rchan_callbacks relay_callbacks = {
+	.subbuf_start = subbuf_start_callback,
+	.create_buf_file = create_buf_file_callback,
+	.remove_buf_file = remove_buf_file_callback,
+};
+
+static void guc_move_to_next_buf(struct intel_guc_log *log)
+{
+	/*
+	 * Make sure the updates made in the sub buffer are visible when
+	 * Consumer sees the following update to offset inside the sub buffer.
+	 */
+	smp_wmb();
+
+	/* All data has been written, so now move the offset of sub buffer. */
+	relay_reserve(log->relay.channel, log->vma->obj->base.size -
+					  intel_guc_log_section_size_capture(log));
+
+	/* Switch to the next sub buffer */
+	relay_flush(log->relay.channel);
+}
+
+static void *guc_get_write_buffer(struct intel_guc_log *log)
+{
+	/*
+	 * Just get the base address of a new sub buffer and copy data into it
+	 * ourselves. NULL will be returned in no-overwrite mode, if all sub
+	 * buffers are full. Could have used the relay_write() to indirectly
+	 * copy the data, but that would have been bit convoluted, as we need to
+	 * write to only certain locations inside a sub buffer which cannot be
+	 * done without using relay_reserve() along with relay_write(). So its
+	 * better to use relay_reserve() alone.
+	 */
+	return relay_reserve(log->relay.channel, 0);
+}
+
+bool intel_guc_check_log_buf_overflow(struct intel_guc_log *log,
+				      enum guc_log_buffer_type type,
+				      unsigned int full_cnt)
+{
+	unsigned int prev_full_cnt = log->stats[type].sampled_overflow;
+	bool overflow = false;
+
+	if (full_cnt != prev_full_cnt) {
+		overflow = true;
+
+		log->stats[type].overflow = full_cnt;
+		log->stats[type].sampled_overflow += full_cnt - prev_full_cnt;
+
+		if (full_cnt < prev_full_cnt) {
+			/* buffer_full_cnt is a 4 bit counter */
+			log->stats[type].sampled_overflow += 16;
+		}
+
+		dev_notice_ratelimited(guc_to_gt(log_to_guc(log))->i915->drm.dev,
+				       "GuC log buffer overflow\n");
+	}
+
+	return overflow;
+}
+
+unsigned int intel_guc_get_log_buffer_size(struct intel_guc_log *log,
+					   enum guc_log_buffer_type type)
+{
+	switch (type) {
+	case GUC_DEBUG_LOG_BUFFER:
+		return intel_guc_log_section_size_debug(log);
+	case GUC_CRASH_DUMP_LOG_BUFFER:
+		return intel_guc_log_section_size_crash(log);
+	case GUC_CAPTURE_LOG_BUFFER:
+		return intel_guc_log_section_size_capture(log);
+	default:
+		MISSING_CASE(type);
+	}
+
+	return 0;
+}
+
+size_t intel_guc_get_log_buffer_offset(struct intel_guc_log *log,
+				       enum guc_log_buffer_type type)
+{
+	enum guc_log_buffer_type i;
+	size_t offset = PAGE_SIZE;/* for the log_buffer_states */
+
+	for (i = GUC_DEBUG_LOG_BUFFER; i < GUC_MAX_LOG_BUFFER; ++i) {
+		if (i == type)
+			break;
+		offset += intel_guc_get_log_buffer_size(log, i);
+	}
+
+	return offset;
+}
+
+static void _guc_log_copy_debuglogs_for_relay(struct intel_guc_log *log)
+{
+	unsigned int buffer_size, read_offset, write_offset, bytes_to_copy, full_cnt;
+	struct guc_log_buffer_state *log_buf_state, *log_buf_snapshot_state;
+	struct guc_log_buffer_state log_buf_state_local;
+	enum guc_log_buffer_type type;
+	void *src_data, *dst_data;
+	bool new_overflow;
+
+	mutex_lock(&log->relay.lock);
+
+	if (WARN_ON(!intel_guc_log_relay_created(log)))
+		goto out_unlock;
+
+	/* Get the pointer to shared GuC log buffer */
+	src_data = log->buf_addr;
+	log_buf_state = src_data;
+
+	/* Get the pointer to local buffer to store the logs */
+	log_buf_snapshot_state = dst_data = guc_get_write_buffer(log);
+
+	if (unlikely(!log_buf_snapshot_state)) {
+		/*
+		 * Used rate limited to avoid deluge of messages, logs might be
+		 * getting consumed by User at a slow rate.
+		 */
+		DRM_ERROR_RATELIMITED("no sub-buffer to copy general logs\n");
+		log->relay.full_count++;
+
+		goto out_unlock;
+	}
+
+	/* Actual logs are present from the 2nd page */
+	src_data += PAGE_SIZE;
+	dst_data += PAGE_SIZE;
+
+	/* For relay logging, we exclude error state capture */
+	for (type = GUC_DEBUG_LOG_BUFFER; type <= GUC_CRASH_DUMP_LOG_BUFFER; type++) {
+		/*
+		 * Make a copy of the state structure, inside GuC log buffer
+		 * (which is uncached mapped), on the stack to avoid reading
+		 * from it multiple times.
+		 */
+		memcpy(&log_buf_state_local, log_buf_state,
+		       sizeof(struct guc_log_buffer_state));
+		buffer_size = intel_guc_get_log_buffer_size(log, type);
+		read_offset = log_buf_state_local.read_ptr;
+		write_offset = log_buf_state_local.sampled_write_ptr;
+		full_cnt = log_buf_state_local.buffer_full_cnt;
+
+		/* Bookkeeping stuff */
+		log->stats[type].flush += log_buf_state_local.flush_to_file;
+		new_overflow = intel_guc_check_log_buf_overflow(log, type, full_cnt);
+
+		/* Update the state of shared log buffer */
+		log_buf_state->read_ptr = write_offset;
+		log_buf_state->flush_to_file = 0;
+		log_buf_state++;
+
+		/* First copy the state structure in snapshot buffer */
+		memcpy(log_buf_snapshot_state, &log_buf_state_local,
+		       sizeof(struct guc_log_buffer_state));
+
+		/*
+		 * The write pointer could have been updated by GuC firmware,
+		 * after sending the flush interrupt to Host, for consistency
+		 * set write pointer value to same value of sampled_write_ptr
+		 * in the snapshot buffer.
+		 */
+		log_buf_snapshot_state->write_ptr = write_offset;
+		log_buf_snapshot_state++;
+
+		/* Now copy the actual logs. */
+		if (unlikely(new_overflow)) {
+			/* copy the whole buffer in case of overflow */
+			read_offset = 0;
+			write_offset = buffer_size;
+		} else if (unlikely((read_offset > buffer_size) ||
+				    (write_offset > buffer_size))) {
+			DRM_ERROR("invalid log buffer state\n");
+			/* copy whole buffer as offsets are unreliable */
+			read_offset = 0;
+			write_offset = buffer_size;
+		}
+
+		/* Just copy the newly written data */
+		if (read_offset > write_offset) {
+			i915_memcpy_from_wc(dst_data, src_data, write_offset);
+			bytes_to_copy = buffer_size - read_offset;
+		} else {
+			bytes_to_copy = write_offset - read_offset;
+		}
+		i915_memcpy_from_wc(dst_data + read_offset,
+				    src_data + read_offset, bytes_to_copy);
+
+		src_data += buffer_size;
+		dst_data += buffer_size;
+	}
+
+	guc_move_to_next_buf(log);
+
+out_unlock:
+	mutex_unlock(&log->relay.lock);
+}
+
+static void copy_debug_logs_work(struct work_struct *work)
+{
+	struct intel_guc_log *log =
+		container_of(work, struct intel_guc_log, relay.flush_work);
+
+	guc_log_copy_debuglogs_for_relay(log);
+}
+
+static int guc_log_relay_map(struct intel_guc_log *log)
+{
+	lockdep_assert_held(&log->relay.lock);
+
+	if (!log->vma || !log->buf_addr)
+		return -ENODEV;
+
+	/*
+	 * WC vmalloc mapping of log buffer pages was done at
+	 * GuC Log Init time, but lets keep a ref for book-keeping
+	 */
+	i915_gem_object_get(log->vma->obj);
+	log->relay.buf_in_use = true;
+
+	return 0;
+}
+
+static void guc_log_relay_unmap(struct intel_guc_log *log)
+{
+	lockdep_assert_held(&log->relay.lock);
+
+	i915_gem_object_put(log->vma->obj);
+	log->relay.buf_in_use = false;
+}
+
+void intel_guc_log_init_early(struct intel_guc_log *log)
+{
+	mutex_init(&log->relay.lock);
+	INIT_WORK(&log->relay.flush_work, copy_debug_logs_work);
+	log->relay.started = false;
+}
+
+static int guc_log_relay_create(struct intel_guc_log *log)
+{
+	struct intel_guc *guc = log_to_guc(log);
+	struct drm_i915_private *dev_priv = guc_to_gt(guc)->i915;
+	struct rchan *guc_log_relay_chan;
+	size_t n_subbufs, subbuf_size;
+	int ret;
+
+	lockdep_assert_held(&log->relay.lock);
+	GEM_BUG_ON(!log->vma);
+
+	 /*
+	  * Keep the size of sub buffers same as shared log buffer
+	  * but GuC log-events excludes the error-state-capture logs
+	  */
+	subbuf_size = log->vma->size - intel_guc_log_section_size_capture(log);
+
+	/*
+	 * Store up to 8 snapshots, which is large enough to buffer sufficient
+	 * boot time logs and provides enough leeway to User, in terms of
+	 * latency, for consuming the logs from relay. Also doesn't take
+	 * up too much memory.
+	 */
+	n_subbufs = 8;
+
+	guc_log_relay_chan = relay_open("guc_log",
+					dev_priv->drm.primary->debugfs_root,
+					subbuf_size, n_subbufs,
+					&relay_callbacks, dev_priv);
+	if (!guc_log_relay_chan) {
+		DRM_ERROR("Couldn't create relay chan for GuC logging\n");
+
+		ret = -ENOMEM;
+		return ret;
+	}
+
+	GEM_BUG_ON(guc_log_relay_chan->subbuf_size < subbuf_size);
+	log->relay.channel = guc_log_relay_chan;
+
+	return 0;
+}
+
+static void guc_log_relay_destroy(struct intel_guc_log *log)
+{
+	lockdep_assert_held(&log->relay.lock);
+
+	relay_close(log->relay.channel);
+	log->relay.channel = NULL;
+}
+
+static void guc_log_copy_debuglogs_for_relay(struct intel_guc_log *log)
+{
+	struct intel_guc *guc = log_to_guc(log);
+	struct drm_i915_private *dev_priv = guc_to_gt(guc)->i915;
+	intel_wakeref_t wakeref;
+
+	_guc_log_copy_debuglogs_for_relay(log);
+
+	/*
+	 * Generally device is expected to be active only at this
+	 * time, so get/put should be really quick.
+	 */
+	with_intel_runtime_pm(&dev_priv->runtime_pm, wakeref)
+		guc_action_flush_log_complete(guc);
+}
+
+static u32 __get_default_log_level(struct intel_guc_log *log)
+{
+	struct intel_guc *guc = log_to_guc(log);
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+
+	/* A negative value means "use platform/config default" */
+	if (i915->params.guc_log_level < 0) {
+		return (IS_ENABLED(CONFIG_DRM_I915_DEBUG) ||
+			IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)) ?
+			GUC_LOG_LEVEL_MAX : GUC_LOG_LEVEL_NON_VERBOSE;
+	}
+
+	if (i915->params.guc_log_level > GUC_LOG_LEVEL_MAX) {
+		DRM_WARN("Incompatible option detected: %s=%d, %s!\n",
+			 "guc_log_level", i915->params.guc_log_level,
+			 "verbosity too high");
+		return (IS_ENABLED(CONFIG_DRM_I915_DEBUG) ||
+			IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)) ?
+			GUC_LOG_LEVEL_MAX : GUC_LOG_LEVEL_DISABLED;
+	}
+
+	GEM_BUG_ON(i915->params.guc_log_level < GUC_LOG_LEVEL_DISABLED);
+	GEM_BUG_ON(i915->params.guc_log_level > GUC_LOG_LEVEL_MAX);
+	return i915->params.guc_log_level;
+}
+
+int intel_guc_log_create(struct intel_guc_log *log)
+{
+	struct intel_guc *guc = log_to_guc(log);
+	struct i915_vma *vma;
+	void *vaddr;
+	u32 guc_log_size;
+	int ret;
+
+	GEM_BUG_ON(log->vma);
+
+	guc_log_size = intel_guc_log_size(log);
+
+	vma = intel_guc_allocate_vma(guc, guc_log_size);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto err;
+	}
+
+	log->vma = vma;
+	/*
+	 * Create a WC (Uncached for read) vmalloc mapping up front immediate access to
+	 * data from memory during  critical events such as error capture
+	 */
+	vaddr = i915_gem_object_pin_map_unlocked(log->vma->obj, I915_MAP_WC);
+	if (IS_ERR(vaddr)) {
+		ret = PTR_ERR(vaddr);
+		i915_vma_unpin_and_release(&log->vma, 0);
+		goto err;
+	}
+	log->buf_addr = vaddr;
+
+	log->level = __get_default_log_level(log);
+	DRM_DEBUG_DRIVER("guc_log_level=%d (%s, verbose:%s, verbosity:%d)\n",
+			 log->level, str_enabled_disabled(log->level),
+			 str_yes_no(GUC_LOG_LEVEL_IS_VERBOSE(log->level)),
+			 GUC_LOG_LEVEL_TO_VERBOSITY(log->level));
+
+	return 0;
+
+err:
+	DRM_ERROR("Failed to allocate or map GuC log buffer. %d\n", ret);
+	return ret;
+}
+
+void intel_guc_log_destroy(struct intel_guc_log *log)
+{
+	log->buf_addr = NULL;
+	i915_vma_unpin_and_release(&log->vma, I915_VMA_RELEASE_MAP);
+}
+
+int intel_guc_log_set_level(struct intel_guc_log *log, u32 level)
+{
+	struct intel_guc *guc = log_to_guc(log);
+	struct drm_i915_private *dev_priv = guc_to_gt(guc)->i915;
+	intel_wakeref_t wakeref;
+	int ret = 0;
+
+	BUILD_BUG_ON(GUC_LOG_VERBOSITY_MIN != 0);
+	GEM_BUG_ON(!log->vma);
+
+	/*
+	 * GuC is recognizing log levels starting from 0 to max, we're using 0
+	 * as indication that logging should be disabled.
+	 */
+	if (level < GUC_LOG_LEVEL_DISABLED || level > GUC_LOG_LEVEL_MAX)
+		return -EINVAL;
+
+	mutex_lock(&dev_priv->drm.struct_mutex);
+
+	if (log->level == level)
+		goto out_unlock;
+
+	with_intel_runtime_pm(&dev_priv->runtime_pm, wakeref)
+		ret = guc_action_control_log(guc,
+					     GUC_LOG_LEVEL_IS_VERBOSE(level),
+					     GUC_LOG_LEVEL_IS_ENABLED(level),
+					     GUC_LOG_LEVEL_TO_VERBOSITY(level));
+	if (ret) {
+		DRM_DEBUG_DRIVER("guc_log_control action failed %d\n", ret);
+		goto out_unlock;
+	}
+
+	log->level = level;
+
+out_unlock:
+	mutex_unlock(&dev_priv->drm.struct_mutex);
+
+	return ret;
+}
+
+bool intel_guc_log_relay_created(const struct intel_guc_log *log)
+{
+	return log->buf_addr;
+}
+
+int intel_guc_log_relay_open(struct intel_guc_log *log)
+{
+	int ret;
+
+	if (!log->vma)
+		return -ENODEV;
+
+	mutex_lock(&log->relay.lock);
+
+	if (intel_guc_log_relay_created(log)) {
+		ret = -EEXIST;
+		goto out_unlock;
+	}
+
+	/*
+	 * We require SSE 4.1 for fast reads from the GuC log buffer and
+	 * it should be present on the chipsets supporting GuC based
+	 * submissions.
+	 */
+	if (!i915_has_memcpy_from_wc()) {
+		ret = -ENXIO;
+		goto out_unlock;
+	}
+
+	ret = guc_log_relay_create(log);
+	if (ret)
+		goto out_unlock;
+
+	ret = guc_log_relay_map(log);
+	if (ret)
+		goto out_relay;
+
+	mutex_unlock(&log->relay.lock);
+
+	return 0;
+
+out_relay:
+	guc_log_relay_destroy(log);
+out_unlock:
+	mutex_unlock(&log->relay.lock);
+
+	return ret;
+}
+
+int intel_guc_log_relay_start(struct intel_guc_log *log)
+{
+	if (log->relay.started)
+		return -EEXIST;
+
+	/*
+	 * When GuC is logging without us relaying to userspace, we're ignoring
+	 * the flush notification. This means that we need to unconditionally
+	 * flush on relay enabling, since GuC only notifies us once.
+	 */
+	queue_work(system_highpri_wq, &log->relay.flush_work);
+
+	log->relay.started = true;
+
+	return 0;
+}
+
+void intel_guc_log_relay_flush(struct intel_guc_log *log)
+{
+	struct intel_guc *guc = log_to_guc(log);
+	intel_wakeref_t wakeref;
+
+	if (!log->relay.started)
+		return;
+
+	/*
+	 * Before initiating the forceful flush, wait for any pending/ongoing
+	 * flush to complete otherwise forceful flush may not actually happen.
+	 */
+	flush_work(&log->relay.flush_work);
+
+	with_intel_runtime_pm(guc_to_gt(guc)->uncore->rpm, wakeref)
+		guc_action_flush_log(guc);
+
+	/* GuC would have updated log buffer by now, so copy it */
+	guc_log_copy_debuglogs_for_relay(log);
+}
+
+/*
+ * Stops the relay log. Called from intel_guc_log_relay_close(), so no
+ * possibility of race with start/flush since relay_write cannot race
+ * relay_close.
+ */
+static void guc_log_relay_stop(struct intel_guc_log *log)
+{
+	struct intel_guc *guc = log_to_guc(log);
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+
+	if (!log->relay.started)
+		return;
+
+	intel_synchronize_irq(i915);
+
+	flush_work(&log->relay.flush_work);
+
+	log->relay.started = false;
+}
+
+void intel_guc_log_relay_close(struct intel_guc_log *log)
+{
+	guc_log_relay_stop(log);
+
+	mutex_lock(&log->relay.lock);
+	GEM_BUG_ON(!intel_guc_log_relay_created(log));
+	guc_log_relay_unmap(log);
+	guc_log_relay_destroy(log);
+	mutex_unlock(&log->relay.lock);
+}
+
+void intel_guc_log_handle_flush_event(struct intel_guc_log *log)
+{
+	if (log->relay.started)
+		queue_work(system_highpri_wq, &log->relay.flush_work);
+}
+
+static const char *
+stringify_guc_log_type(enum guc_log_buffer_type type)
+{
+	switch (type) {
+	case GUC_DEBUG_LOG_BUFFER:
+		return "DEBUG";
+	case GUC_CRASH_DUMP_LOG_BUFFER:
+		return "CRASH";
+	case GUC_CAPTURE_LOG_BUFFER:
+		return "CAPTURE";
+	default:
+		MISSING_CASE(type);
+	}
+
+	return "";
+}
+
+/**
+ * intel_guc_log_info - dump information about GuC log relay
+ * @log: the GuC log
+ * @p: the &drm_printer
+ *
+ * Pretty printer for GuC log info
+ */
+void intel_guc_log_info(struct intel_guc_log *log, struct drm_printer *p)
+{
+	enum guc_log_buffer_type type;
+
+	if (!intel_guc_log_relay_created(log)) {
+		drm_puts(p, "GuC log relay not created\n");
+		return;
+	}
+
+	drm_puts(p, "GuC logging stats:\n");
+
+	drm_printf(p, "\tRelay full count: %u\n", log->relay.full_count);
+
+	for (type = GUC_DEBUG_LOG_BUFFER; type < GUC_MAX_LOG_BUFFER; type++) {
+		drm_printf(p, "\t%s:\tflush count %10u, overflow count %10u\n",
+			   stringify_guc_log_type(type),
+			   log->stats[type].flush,
+			   log->stats[type].sampled_overflow);
+	}
+}
+
+/**
+ * intel_guc_log_dump - dump the contents of the GuC log
+ * @log: the GuC log
+ * @p: the &drm_printer
+ * @dump_load_err: dump the log saved on GuC load error
+ *
+ * Pretty printer for the GuC log
+ */
+int intel_guc_log_dump(struct intel_guc_log *log, struct drm_printer *p,
+		       bool dump_load_err)
+{
+	struct intel_guc *guc = log_to_guc(log);
+	struct intel_uc *uc = container_of(guc, struct intel_uc, guc);
+	struct drm_i915_gem_object *obj = NULL;
+	void *map;
+	u32 *page;
+	int i, j;
+
+	if (!intel_guc_is_supported(guc))
+		return -ENODEV;
+
+	if (dump_load_err)
+		obj = uc->load_err_log;
+	else if (guc->log.vma)
+		obj = guc->log.vma->obj;
+
+	if (!obj)
+		return 0;
+
+	page = (u32 *)__get_free_page(GFP_KERNEL);
+	if (!page)
+		return -ENOMEM;
+
+	intel_guc_dump_time_info(guc, p);
+
+	map = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(map)) {
+		DRM_DEBUG("Failed to pin object\n");
+		drm_puts(p, "(log data unaccessible)\n");
+		free_page((unsigned long)page);
+		return PTR_ERR(map);
+	}
+
+	for (i = 0; i < obj->base.size; i += PAGE_SIZE) {
+		if (!i915_memcpy_from_wc(page, map + i, PAGE_SIZE))
+			memcpy(page, map + i, PAGE_SIZE);
+
+		for (j = 0; j < PAGE_SIZE / sizeof(u32); j += 4)
+			drm_printf(p, "0x%08x 0x%08x 0x%08x 0x%08x\n",
+				   *(page + j + 0), *(page + j + 1),
+				   *(page + j + 2), *(page + j + 3));
+	}
+
+	drm_puts(p, "\n");
+
+	i915_gem_object_unpin_map(obj);
+	free_page((unsigned long)page);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_log.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_log.h
new file mode 100644
index 000000000..02127703b
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_log.h
@@ -0,0 +1,105 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_LOG_H_
+#define _INTEL_GUC_LOG_H_
+
+#include <linux/mutex.h>
+#include <linux/relay.h>
+#include <linux/workqueue.h>
+
+#include "intel_guc_fwif.h"
+#include "i915_gem.h"
+
+struct intel_guc;
+
+/*
+ * While we're using plain log level in i915, GuC controls are much more...
+ * "elaborate"? We have a couple of bits for verbosity, separate bit for actual
+ * log enabling, and separate bit for default logging - which "conveniently"
+ * ignores the enable bit.
+ */
+#define GUC_LOG_LEVEL_DISABLED		0
+#define GUC_LOG_LEVEL_NON_VERBOSE	1
+#define GUC_LOG_LEVEL_IS_ENABLED(x)	((x) > GUC_LOG_LEVEL_DISABLED)
+#define GUC_LOG_LEVEL_IS_VERBOSE(x)	((x) > GUC_LOG_LEVEL_NON_VERBOSE)
+#define GUC_LOG_LEVEL_TO_VERBOSITY(x) ({		\
+	typeof(x) _x = (x);				\
+	GUC_LOG_LEVEL_IS_VERBOSE(_x) ? _x - 2 : 0;	\
+})
+#define GUC_VERBOSITY_TO_LOG_LEVEL(x)	((x) + 2)
+#define GUC_LOG_LEVEL_MAX GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX)
+
+enum {
+	GUC_LOG_SECTIONS_CRASH,
+	GUC_LOG_SECTIONS_DEBUG,
+	GUC_LOG_SECTIONS_CAPTURE,
+	GUC_LOG_SECTIONS_LIMIT
+};
+
+struct intel_guc_log {
+	u32 level;
+
+	/* Allocation settings */
+	struct {
+		s32 bytes;	/* Size in bytes */
+		s32 units;	/* GuC API units - 1MB or 4KB */
+		s32 count;	/* Number of API units */
+		u32 flag;	/* GuC API units flag */
+	} sizes[GUC_LOG_SECTIONS_LIMIT];
+	bool sizes_initialised;
+
+	/* Combined buffer allocation */
+	struct i915_vma *vma;
+	void *buf_addr;
+
+	/* RelayFS support */
+	struct {
+		bool buf_in_use;
+		bool started;
+		struct work_struct flush_work;
+		struct rchan *channel;
+		struct mutex lock;
+		u32 full_count;
+	} relay;
+
+	/* logging related stats */
+	struct {
+		u32 sampled_overflow;
+		u32 overflow;
+		u32 flush;
+	} stats[GUC_MAX_LOG_BUFFER];
+};
+
+void intel_guc_log_init_early(struct intel_guc_log *log);
+bool intel_guc_check_log_buf_overflow(struct intel_guc_log *log, enum guc_log_buffer_type type,
+				      unsigned int full_cnt);
+unsigned int intel_guc_get_log_buffer_size(struct intel_guc_log *log,
+					   enum guc_log_buffer_type type);
+size_t intel_guc_get_log_buffer_offset(struct intel_guc_log *log, enum guc_log_buffer_type type);
+int intel_guc_log_create(struct intel_guc_log *log);
+void intel_guc_log_destroy(struct intel_guc_log *log);
+
+int intel_guc_log_set_level(struct intel_guc_log *log, u32 level);
+bool intel_guc_log_relay_created(const struct intel_guc_log *log);
+int intel_guc_log_relay_open(struct intel_guc_log *log);
+int intel_guc_log_relay_start(struct intel_guc_log *log);
+void intel_guc_log_relay_flush(struct intel_guc_log *log);
+void intel_guc_log_relay_close(struct intel_guc_log *log);
+
+void intel_guc_log_handle_flush_event(struct intel_guc_log *log);
+
+static inline u32 intel_guc_log_get_level(struct intel_guc_log *log)
+{
+	return log->level;
+}
+
+void intel_guc_log_info(struct intel_guc_log *log, struct drm_printer *p);
+int intel_guc_log_dump(struct intel_guc_log *log, struct drm_printer *p,
+		       bool dump_load_err);
+
+u32 intel_guc_log_section_size_capture(struct intel_guc_log *log);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_log_debugfs.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_log_debugfs.c
new file mode 100644
index 000000000..ddfbe3346
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_log_debugfs.c
@@ -0,0 +1,175 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/fs.h>
+#include <drm/drm_print.h>
+
+#include "gt/intel_gt_debugfs.h"
+#include "intel_guc.h"
+#include "intel_guc_log.h"
+#include "intel_guc_log_debugfs.h"
+#include "intel_uc.h"
+
+static u32 obj_to_guc_log_dump_size(struct drm_i915_gem_object *obj)
+{
+	u32 size;
+
+	if (!obj)
+		return PAGE_SIZE;
+
+	/* "0x%08x 0x%08x 0x%08x 0x%08x\n" => 16 bytes -> 44 chars => x2.75 */
+	size = ((obj->base.size * 11) + 3) / 4;
+
+	/* Add padding for final blank line, any extra header info, etc. */
+	size = PAGE_ALIGN(size + PAGE_SIZE);
+
+	return size;
+}
+
+static u32 guc_log_dump_size(struct intel_guc_log *log)
+{
+	struct intel_guc *guc = log_to_guc(log);
+
+	if (!intel_guc_is_supported(guc))
+		return PAGE_SIZE;
+
+	if (!log->vma)
+		return PAGE_SIZE;
+
+	return obj_to_guc_log_dump_size(log->vma->obj);
+}
+
+static int guc_log_dump_show(struct seq_file *m, void *data)
+{
+	struct drm_printer p = drm_seq_file_printer(m);
+	int ret;
+
+	ret = intel_guc_log_dump(m->private, &p, false);
+
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM) && seq_has_overflowed(m))
+		pr_warn_once("preallocated size:%zx for %s exceeded\n",
+			     m->size, __func__);
+
+	return ret;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE_WITH_SIZE(guc_log_dump, guc_log_dump_size);
+
+static u32 guc_load_err_dump_size(struct intel_guc_log *log)
+{
+	struct intel_guc *guc = log_to_guc(log);
+	struct intel_uc *uc = container_of(guc, struct intel_uc, guc);
+
+	if (!intel_guc_is_supported(guc))
+		return PAGE_SIZE;
+
+	return obj_to_guc_log_dump_size(uc->load_err_log);
+}
+
+static int guc_load_err_log_dump_show(struct seq_file *m, void *data)
+{
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM) && seq_has_overflowed(m))
+		pr_warn_once("preallocated size:%zx for %s exceeded\n",
+			     m->size, __func__);
+
+	return intel_guc_log_dump(m->private, &p, true);
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE_WITH_SIZE(guc_load_err_log_dump, guc_load_err_dump_size);
+
+static int guc_log_level_get(void *data, u64 *val)
+{
+	struct intel_guc_log *log = data;
+
+	if (!log->vma)
+		return -ENODEV;
+
+	*val = intel_guc_log_get_level(log);
+
+	return 0;
+}
+
+static int guc_log_level_set(void *data, u64 val)
+{
+	struct intel_guc_log *log = data;
+
+	if (!log->vma)
+		return -ENODEV;
+
+	return intel_guc_log_set_level(log, val);
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(guc_log_level_fops,
+			guc_log_level_get, guc_log_level_set,
+			"%lld\n");
+
+static int guc_log_relay_open(struct inode *inode, struct file *file)
+{
+	struct intel_guc_log *log = inode->i_private;
+
+	if (!intel_guc_is_ready(log_to_guc(log)))
+		return -ENODEV;
+
+	file->private_data = log;
+
+	return intel_guc_log_relay_open(log);
+}
+
+static ssize_t
+guc_log_relay_write(struct file *filp,
+		    const char __user *ubuf,
+		    size_t cnt,
+		    loff_t *ppos)
+{
+	struct intel_guc_log *log = filp->private_data;
+	int val;
+	int ret;
+
+	ret = kstrtoint_from_user(ubuf, cnt, 0, &val);
+	if (ret < 0)
+		return ret;
+
+	/*
+	 * Enable and start the guc log relay on value of 1.
+	 * Flush log relay for any other value.
+	 */
+	if (val == 1)
+		ret = intel_guc_log_relay_start(log);
+	else
+		intel_guc_log_relay_flush(log);
+
+	return ret ?: cnt;
+}
+
+static int guc_log_relay_release(struct inode *inode, struct file *file)
+{
+	struct intel_guc_log *log = inode->i_private;
+
+	intel_guc_log_relay_close(log);
+	return 0;
+}
+
+static const struct file_operations guc_log_relay_fops = {
+	.owner = THIS_MODULE,
+	.open = guc_log_relay_open,
+	.write = guc_log_relay_write,
+	.release = guc_log_relay_release,
+};
+
+void intel_guc_log_debugfs_register(struct intel_guc_log *log,
+				    struct dentry *root)
+{
+	static const struct intel_gt_debugfs_file files[] = {
+		{ "guc_log_dump", &guc_log_dump_fops, NULL },
+		{ "guc_load_err_log_dump", &guc_load_err_log_dump_fops, NULL },
+		{ "guc_log_level", &guc_log_level_fops, NULL },
+		{ "guc_log_relay", &guc_log_relay_fops, NULL },
+	};
+
+	if (!intel_guc_is_supported(log_to_guc(log)))
+		return;
+
+	intel_gt_debugfs_register_files(root, files, ARRAY_SIZE(files), log);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_log_debugfs.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_log_debugfs.h
new file mode 100644
index 000000000..e8900e3d7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_log_debugfs.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef DEBUGFS_GUC_LOG_H
+#define DEBUGFS_GUC_LOG_H
+
+struct intel_guc_log;
+struct dentry;
+
+void intel_guc_log_debugfs_register(struct intel_guc_log *log,
+				    struct dentry *root);
+
+#endif /* DEBUGFS_GUC_LOG_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_rc.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_rc.c
new file mode 100644
index 000000000..8f8dd0583
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_rc.c
@@ -0,0 +1,81 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+
+#include "intel_guc_rc.h"
+#include "gt/intel_gt.h"
+#include "i915_drv.h"
+
+static bool __guc_rc_supported(struct intel_guc *guc)
+{
+	/* GuC RC is unavailable for pre-Gen12 */
+	return guc->submission_supported &&
+		GRAPHICS_VER(guc_to_gt(guc)->i915) >= 12;
+}
+
+static bool __guc_rc_selected(struct intel_guc *guc)
+{
+	if (!intel_guc_rc_is_supported(guc))
+		return false;
+
+	return guc->submission_selected;
+}
+
+void intel_guc_rc_init_early(struct intel_guc *guc)
+{
+	guc->rc_supported = __guc_rc_supported(guc);
+	guc->rc_selected = __guc_rc_selected(guc);
+}
+
+static int guc_action_control_gucrc(struct intel_guc *guc, bool enable)
+{
+	u32 rc_mode = enable ? INTEL_GUCRC_FIRMWARE_CONTROL :
+				INTEL_GUCRC_HOST_CONTROL;
+	u32 action[] = {
+		INTEL_GUC_ACTION_SETUP_PC_GUCRC,
+		rc_mode
+	};
+	int ret;
+
+	ret = intel_guc_send(guc, action, ARRAY_SIZE(action));
+	ret = ret > 0 ? -EPROTO : ret;
+
+	return ret;
+}
+
+static int __guc_rc_control(struct intel_guc *guc, bool enable)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	int ret;
+
+	if (!intel_uc_uses_guc_rc(&gt->uc))
+		return -EOPNOTSUPP;
+
+	if (!intel_guc_is_ready(guc))
+		return -EINVAL;
+
+	ret = guc_action_control_gucrc(guc, enable);
+	if (ret) {
+		i915_probe_error(guc_to_gt(guc)->i915, "Failed to %s GuC RC (%pe)\n",
+				 str_enable_disable(enable), ERR_PTR(ret));
+		return ret;
+	}
+
+	drm_info(&gt->i915->drm, "GuC RC: %s\n",
+		 str_enabled_disabled(enable));
+
+	return 0;
+}
+
+int intel_guc_rc_enable(struct intel_guc *guc)
+{
+	return __guc_rc_control(guc, true);
+}
+
+int intel_guc_rc_disable(struct intel_guc *guc)
+{
+	return __guc_rc_control(guc, false);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_rc.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_rc.h
new file mode 100644
index 000000000..57e86c337
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_rc.h
@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_RC_H_
+#define _INTEL_GUC_RC_H_
+
+#include "intel_guc_submission.h"
+
+void intel_guc_rc_init_early(struct intel_guc *guc);
+
+static inline bool intel_guc_rc_is_supported(struct intel_guc *guc)
+{
+	return guc->rc_supported;
+}
+
+static inline bool intel_guc_rc_is_wanted(struct intel_guc *guc)
+{
+	return guc->submission_selected && intel_guc_rc_is_supported(guc);
+}
+
+static inline bool intel_guc_rc_is_used(struct intel_guc *guc)
+{
+	return intel_guc_submission_is_used(guc) && intel_guc_rc_is_wanted(guc);
+}
+
+int intel_guc_rc_enable(struct intel_guc *guc);
+int intel_guc_rc_disable(struct intel_guc *guc);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_reg.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_reg.h
new file mode 100644
index 000000000..a7092f711
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_reg.h
@@ -0,0 +1,155 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_REG_H_
+#define _INTEL_GUC_REG_H_
+
+#include <linux/compiler.h>
+#include <linux/types.h>
+
+#include "i915_reg_defs.h"
+
+/* Definitions of GuC H/W registers, bits, etc */
+
+#define GUC_STATUS			_MMIO(0xc000)
+#define   GS_RESET_SHIFT		0
+#define   GS_MIA_IN_RESET		  (0x01 << GS_RESET_SHIFT)
+#define   GS_BOOTROM_SHIFT		1
+#define   GS_BOOTROM_MASK		  (0x7F << GS_BOOTROM_SHIFT)
+#define   GS_BOOTROM_RSA_FAILED		  (0x50 << GS_BOOTROM_SHIFT)
+#define   GS_BOOTROM_JUMP_PASSED	  (0x76 << GS_BOOTROM_SHIFT)
+#define   GS_UKERNEL_SHIFT		8
+#define   GS_UKERNEL_MASK		  (0xFF << GS_UKERNEL_SHIFT)
+#define   GS_MIA_SHIFT			16
+#define   GS_MIA_MASK			  (0x07 << GS_MIA_SHIFT)
+#define   GS_MIA_CORE_STATE		  (0x01 << GS_MIA_SHIFT)
+#define   GS_MIA_HALT_REQUESTED		  (0x02 << GS_MIA_SHIFT)
+#define   GS_MIA_ISR_ENTRY		  (0x04 << GS_MIA_SHIFT)
+#define   GS_AUTH_STATUS_SHIFT		30
+#define   GS_AUTH_STATUS_MASK		  (0x03U << GS_AUTH_STATUS_SHIFT)
+#define   GS_AUTH_STATUS_BAD		  (0x01 << GS_AUTH_STATUS_SHIFT)
+#define   GS_AUTH_STATUS_GOOD		  (0x02 << GS_AUTH_STATUS_SHIFT)
+
+#define SOFT_SCRATCH(n)			_MMIO(0xc180 + (n) * 4)
+#define SOFT_SCRATCH_COUNT		16
+
+#define GEN11_SOFT_SCRATCH(n)		_MMIO(0x190240 + (n) * 4)
+#define GEN11_SOFT_SCRATCH_COUNT	4
+
+#define UOS_RSA_SCRATCH(i)		_MMIO(0xc200 + (i) * 4)
+#define UOS_RSA_SCRATCH_COUNT		64
+
+#define DMA_ADDR_0_LOW			_MMIO(0xc300)
+#define DMA_ADDR_0_HIGH			_MMIO(0xc304)
+#define DMA_ADDR_1_LOW			_MMIO(0xc308)
+#define DMA_ADDR_1_HIGH			_MMIO(0xc30c)
+#define   DMA_ADDRESS_SPACE_WOPCM	  (7 << 16)
+#define   DMA_ADDRESS_SPACE_GTT		  (8 << 16)
+#define DMA_COPY_SIZE			_MMIO(0xc310)
+#define DMA_CTRL			_MMIO(0xc314)
+#define   HUC_UKERNEL			  (1<<9)
+#define   UOS_MOVE			  (1<<4)
+#define   START_DMA			  (1<<0)
+#define DMA_GUC_WOPCM_OFFSET		_MMIO(0xc340)
+#define   GUC_WOPCM_OFFSET_VALID	  (1<<0)
+#define   HUC_LOADING_AGENT_VCR		  (0<<1)
+#define   HUC_LOADING_AGENT_GUC		  (1<<1)
+#define   GUC_WOPCM_OFFSET_SHIFT	14
+#define   GUC_WOPCM_OFFSET_MASK		  (0x3ffff << GUC_WOPCM_OFFSET_SHIFT)
+#define GUC_MAX_IDLE_COUNT		_MMIO(0xC3E4)
+
+#define HUC_STATUS2             _MMIO(0xD3B0)
+#define   HUC_FW_VERIFIED       (1<<7)
+
+#define GEN11_HUC_KERNEL_LOAD_INFO	_MMIO(0xC1DC)
+#define   HUC_LOAD_SUCCESSFUL		  (1 << 0)
+
+#define GUC_WOPCM_SIZE			_MMIO(0xc050)
+#define   GUC_WOPCM_SIZE_LOCKED		  (1<<0)
+#define   GUC_WOPCM_SIZE_SHIFT		12
+#define   GUC_WOPCM_SIZE_MASK		  (0xfffff << GUC_WOPCM_SIZE_SHIFT)
+
+#define GEN8_GT_PM_CONFIG		_MMIO(0x138140)
+#define GEN9LP_GT_PM_CONFIG		_MMIO(0x138140)
+#define GEN9_GT_PM_CONFIG		_MMIO(0x13816c)
+#define   GT_DOORBELL_ENABLE		  (1<<0)
+
+#define GEN8_GTCR			_MMIO(0x4274)
+#define   GEN8_GTCR_INVALIDATE		  (1<<0)
+
+#define GEN12_GUC_TLB_INV_CR		_MMIO(0xcee8)
+#define   GEN12_GUC_TLB_INV_CR_INVALIDATE	(1 << 0)
+
+#define GUC_ARAT_C6DIS			_MMIO(0xA178)
+
+#define GUC_SHIM_CONTROL		_MMIO(0xc064)
+#define   GUC_DISABLE_SRAM_INIT_TO_ZEROES	(1<<0)
+#define   GUC_ENABLE_READ_CACHE_LOGIC		(1<<1)
+#define   GUC_ENABLE_MIA_CACHING		(1<<2)
+#define   GUC_GEN10_MSGCH_ENABLE		(1<<4)
+#define   GUC_ENABLE_READ_CACHE_FOR_SRAM_DATA	(1<<9)
+#define   GUC_ENABLE_READ_CACHE_FOR_WOPCM_DATA	(1<<10)
+#define   GUC_ENABLE_MIA_CLOCK_GATING		(1<<15)
+#define   GUC_GEN10_SHIM_WC_ENABLE		(1<<21)
+
+#define GUC_SHIM_CONTROL2		_MMIO(0xc068)
+#define   GUC_IS_PRIVILEGED		(1<<29)
+#define   GSC_LOADS_HUC			(1<<30)
+
+#define GUC_SEND_INTERRUPT		_MMIO(0xc4c8)
+#define   GUC_SEND_TRIGGER		  (1<<0)
+#define GEN11_GUC_HOST_INTERRUPT	_MMIO(0x1901f0)
+
+#define GEN12_GUC_SEM_INTR_ENABLES	_MMIO(0xc71c)
+#define   GUC_SEM_INTR_ROUTE_TO_GUC	BIT(31)
+#define   GUC_SEM_INTR_ENABLE_ALL	(0xff)
+
+#define GUC_NUM_DOORBELLS		256
+
+/* format of the HW-monitored doorbell cacheline */
+struct guc_doorbell_info {
+	u32 db_status;
+#define GUC_DOORBELL_DISABLED		0
+#define GUC_DOORBELL_ENABLED		1
+
+	u32 cookie;
+	u32 reserved[14];
+} __packed;
+
+#define GEN8_DRBREGL(x)			_MMIO(0x1000 + (x) * 8)
+#define   GEN8_DRB_VALID		  (1<<0)
+#define GEN8_DRBREGU(x)			_MMIO(0x1000 + (x) * 8 + 4)
+
+#define GEN12_DIST_DBS_POPULATED		_MMIO(0xd08)
+#define   GEN12_DOORBELLS_PER_SQIDI_SHIFT	16
+#define   GEN12_DOORBELLS_PER_SQIDI		(0xff)
+#define   GEN12_SQIDIS_DOORBELL_EXIST		(0xffff)
+
+#define DE_GUCRMR			_MMIO(0x44054)
+
+#define GUC_BCS_RCS_IER			_MMIO(0xC550)
+#define GUC_VCS2_VCS1_IER		_MMIO(0xC554)
+#define GUC_WD_VECS_IER			_MMIO(0xC558)
+#define GUC_PM_P24C_IER			_MMIO(0xC55C)
+
+/* GuC Interrupt Vector */
+#define GUC_INTR_GUC2HOST		BIT(15)
+#define GUC_INTR_EXEC_ERROR		BIT(14)
+#define GUC_INTR_DISPLAY_EVENT		BIT(13)
+#define GUC_INTR_SEM_SIG		BIT(12)
+#define GUC_INTR_IOMMU2GUC		BIT(11)
+#define GUC_INTR_DOORBELL_RANG		BIT(10)
+#define GUC_INTR_DMA_DONE		BIT(9)
+#define GUC_INTR_FATAL_ERROR		BIT(8)
+#define GUC_INTR_NOTIF_ERROR		BIT(7)
+#define GUC_INTR_SW_INT_6		BIT(6)
+#define GUC_INTR_SW_INT_5		BIT(5)
+#define GUC_INTR_SW_INT_4		BIT(4)
+#define GUC_INTR_SW_INT_3		BIT(3)
+#define GUC_INTR_SW_INT_2		BIT(2)
+#define GUC_INTR_SW_INT_1		BIT(1)
+#define GUC_INTR_SW_INT_0		BIT(0)
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.c
new file mode 100644
index 000000000..72ba1c758
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.c
@@ -0,0 +1,750 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <drm/drm_cache.h>
+#include <linux/string_helpers.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_guc_slpc.h"
+#include "intel_mchbar_regs.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_rps.h"
+
+static inline struct intel_guc *slpc_to_guc(struct intel_guc_slpc *slpc)
+{
+	return container_of(slpc, struct intel_guc, slpc);
+}
+
+static inline struct intel_gt *slpc_to_gt(struct intel_guc_slpc *slpc)
+{
+	return guc_to_gt(slpc_to_guc(slpc));
+}
+
+static inline struct drm_i915_private *slpc_to_i915(struct intel_guc_slpc *slpc)
+{
+	return slpc_to_gt(slpc)->i915;
+}
+
+static bool __detect_slpc_supported(struct intel_guc *guc)
+{
+	/* GuC SLPC is unavailable for pre-Gen12 */
+	return guc->submission_supported &&
+		GRAPHICS_VER(guc_to_gt(guc)->i915) >= 12;
+}
+
+static bool __guc_slpc_selected(struct intel_guc *guc)
+{
+	if (!intel_guc_slpc_is_supported(guc))
+		return false;
+
+	return guc->submission_selected;
+}
+
+void intel_guc_slpc_init_early(struct intel_guc_slpc *slpc)
+{
+	struct intel_guc *guc = slpc_to_guc(slpc);
+
+	slpc->supported = __detect_slpc_supported(guc);
+	slpc->selected = __guc_slpc_selected(guc);
+}
+
+static void slpc_mem_set_param(struct slpc_shared_data *data,
+			       u32 id, u32 value)
+{
+	GEM_BUG_ON(id >= SLPC_MAX_OVERRIDE_PARAMETERS);
+	/*
+	 * When the flag bit is set, corresponding value will be read
+	 * and applied by SLPC.
+	 */
+	data->override_params.bits[id >> 5] |= (1 << (id % 32));
+	data->override_params.values[id] = value;
+}
+
+static void slpc_mem_set_enabled(struct slpc_shared_data *data,
+				 u8 enable_id, u8 disable_id)
+{
+	/*
+	 * Enabling a param involves setting the enable_id
+	 * to 1 and disable_id to 0.
+	 */
+	slpc_mem_set_param(data, enable_id, 1);
+	slpc_mem_set_param(data, disable_id, 0);
+}
+
+static void slpc_mem_set_disabled(struct slpc_shared_data *data,
+				  u8 enable_id, u8 disable_id)
+{
+	/*
+	 * Disabling a param involves setting the enable_id
+	 * to 0 and disable_id to 1.
+	 */
+	slpc_mem_set_param(data, disable_id, 1);
+	slpc_mem_set_param(data, enable_id, 0);
+}
+
+static u32 slpc_get_state(struct intel_guc_slpc *slpc)
+{
+	struct slpc_shared_data *data;
+
+	GEM_BUG_ON(!slpc->vma);
+
+	drm_clflush_virt_range(slpc->vaddr, sizeof(u32));
+	data = slpc->vaddr;
+
+	return data->header.global_state;
+}
+
+static int guc_action_slpc_set_param_nb(struct intel_guc *guc, u8 id, u32 value)
+{
+	u32 request[] = {
+		GUC_ACTION_HOST2GUC_PC_SLPC_REQUEST,
+		SLPC_EVENT(SLPC_EVENT_PARAMETER_SET, 2),
+		id,
+		value,
+	};
+	int ret;
+
+	ret = intel_guc_send_nb(guc, request, ARRAY_SIZE(request), 0);
+
+	return ret > 0 ? -EPROTO : ret;
+}
+
+static int slpc_set_param_nb(struct intel_guc_slpc *slpc, u8 id, u32 value)
+{
+	struct intel_guc *guc = slpc_to_guc(slpc);
+
+	GEM_BUG_ON(id >= SLPC_MAX_PARAM);
+
+	return guc_action_slpc_set_param_nb(guc, id, value);
+}
+
+static int guc_action_slpc_set_param(struct intel_guc *guc, u8 id, u32 value)
+{
+	u32 request[] = {
+		GUC_ACTION_HOST2GUC_PC_SLPC_REQUEST,
+		SLPC_EVENT(SLPC_EVENT_PARAMETER_SET, 2),
+		id,
+		value,
+	};
+	int ret;
+
+	ret = intel_guc_send(guc, request, ARRAY_SIZE(request));
+
+	return ret > 0 ? -EPROTO : ret;
+}
+
+static bool slpc_is_running(struct intel_guc_slpc *slpc)
+{
+	return slpc_get_state(slpc) == SLPC_GLOBAL_STATE_RUNNING;
+}
+
+static int guc_action_slpc_query(struct intel_guc *guc, u32 offset)
+{
+	u32 request[] = {
+		GUC_ACTION_HOST2GUC_PC_SLPC_REQUEST,
+		SLPC_EVENT(SLPC_EVENT_QUERY_TASK_STATE, 2),
+		offset,
+		0,
+	};
+	int ret;
+
+	ret = intel_guc_send(guc, request, ARRAY_SIZE(request));
+
+	return ret > 0 ? -EPROTO : ret;
+}
+
+static int slpc_query_task_state(struct intel_guc_slpc *slpc)
+{
+	struct intel_guc *guc = slpc_to_guc(slpc);
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	u32 offset = intel_guc_ggtt_offset(guc, slpc->vma);
+	int ret;
+
+	ret = guc_action_slpc_query(guc, offset);
+	if (unlikely(ret))
+		i915_probe_error(i915, "Failed to query task state (%pe)\n",
+				 ERR_PTR(ret));
+
+	drm_clflush_virt_range(slpc->vaddr, SLPC_PAGE_SIZE_BYTES);
+
+	return ret;
+}
+
+static int slpc_set_param(struct intel_guc_slpc *slpc, u8 id, u32 value)
+{
+	struct intel_guc *guc = slpc_to_guc(slpc);
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	int ret;
+
+	GEM_BUG_ON(id >= SLPC_MAX_PARAM);
+
+	ret = guc_action_slpc_set_param(guc, id, value);
+	if (ret)
+		i915_probe_error(i915, "Failed to set param %d to %u (%pe)\n",
+				 id, value, ERR_PTR(ret));
+
+	return ret;
+}
+
+static int slpc_force_min_freq(struct intel_guc_slpc *slpc, u32 freq)
+{
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	struct intel_guc *guc = slpc_to_guc(slpc);
+	intel_wakeref_t wakeref;
+	int ret = 0;
+
+	lockdep_assert_held(&slpc->lock);
+
+	if (!intel_guc_is_ready(guc))
+		return -ENODEV;
+
+	/*
+	 * This function is a little different as compared to
+	 * intel_guc_slpc_set_min_freq(). Softlimit will not be updated
+	 * here since this is used to temporarily change min freq,
+	 * for example, during a waitboost. Caller is responsible for
+	 * checking bounds.
+	 */
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		/* Non-blocking request will avoid stalls */
+		ret = slpc_set_param_nb(slpc,
+					SLPC_PARAM_GLOBAL_MIN_GT_UNSLICE_FREQ_MHZ,
+					freq);
+		if (ret)
+			drm_notice(&i915->drm,
+				   "Failed to send set_param for min freq(%d): (%d)\n",
+				   freq, ret);
+	}
+
+	return ret;
+}
+
+static void slpc_boost_work(struct work_struct *work)
+{
+	struct intel_guc_slpc *slpc = container_of(work, typeof(*slpc), boost_work);
+	int err;
+
+	/*
+	 * Raise min freq to boost. It's possible that
+	 * this is greater than current max. But it will
+	 * certainly be limited by RP0. An error setting
+	 * the min param is not fatal.
+	 */
+	mutex_lock(&slpc->lock);
+	if (atomic_read(&slpc->num_waiters)) {
+		err = slpc_force_min_freq(slpc, slpc->boost_freq);
+		if (!err)
+			slpc->num_boosts++;
+	}
+	mutex_unlock(&slpc->lock);
+}
+
+int intel_guc_slpc_init(struct intel_guc_slpc *slpc)
+{
+	struct intel_guc *guc = slpc_to_guc(slpc);
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	u32 size = PAGE_ALIGN(sizeof(struct slpc_shared_data));
+	int err;
+
+	GEM_BUG_ON(slpc->vma);
+
+	err = intel_guc_allocate_and_map_vma(guc, size, &slpc->vma, (void **)&slpc->vaddr);
+	if (unlikely(err)) {
+		i915_probe_error(i915,
+				 "Failed to allocate SLPC struct (err=%pe)\n",
+				 ERR_PTR(err));
+		return err;
+	}
+
+	slpc->max_freq_softlimit = 0;
+	slpc->min_freq_softlimit = 0;
+
+	slpc->boost_freq = 0;
+	atomic_set(&slpc->num_waiters, 0);
+	slpc->num_boosts = 0;
+	slpc->media_ratio_mode = SLPC_MEDIA_RATIO_MODE_DYNAMIC_CONTROL;
+
+	mutex_init(&slpc->lock);
+	INIT_WORK(&slpc->boost_work, slpc_boost_work);
+
+	return err;
+}
+
+static const char *slpc_global_state_to_string(enum slpc_global_state state)
+{
+	switch (state) {
+	case SLPC_GLOBAL_STATE_NOT_RUNNING:
+		return "not running";
+	case SLPC_GLOBAL_STATE_INITIALIZING:
+		return "initializing";
+	case SLPC_GLOBAL_STATE_RESETTING:
+		return "resetting";
+	case SLPC_GLOBAL_STATE_RUNNING:
+		return "running";
+	case SLPC_GLOBAL_STATE_SHUTTING_DOWN:
+		return "shutting down";
+	case SLPC_GLOBAL_STATE_ERROR:
+		return "error";
+	default:
+		return "unknown";
+	}
+}
+
+static const char *slpc_get_state_string(struct intel_guc_slpc *slpc)
+{
+	return slpc_global_state_to_string(slpc_get_state(slpc));
+}
+
+static int guc_action_slpc_reset(struct intel_guc *guc, u32 offset)
+{
+	u32 request[] = {
+		GUC_ACTION_HOST2GUC_PC_SLPC_REQUEST,
+		SLPC_EVENT(SLPC_EVENT_RESET, 2),
+		offset,
+		0,
+	};
+	int ret;
+
+	ret = intel_guc_send(guc, request, ARRAY_SIZE(request));
+
+	return ret > 0 ? -EPROTO : ret;
+}
+
+static int slpc_reset(struct intel_guc_slpc *slpc)
+{
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	struct intel_guc *guc = slpc_to_guc(slpc);
+	u32 offset = intel_guc_ggtt_offset(guc, slpc->vma);
+	int ret;
+
+	ret = guc_action_slpc_reset(guc, offset);
+
+	if (unlikely(ret < 0)) {
+		i915_probe_error(i915, "SLPC reset action failed (%pe)\n",
+				 ERR_PTR(ret));
+		return ret;
+	}
+
+	if (!ret) {
+		if (wait_for(slpc_is_running(slpc), SLPC_RESET_TIMEOUT_MS)) {
+			i915_probe_error(i915, "SLPC not enabled! State = %s\n",
+					 slpc_get_state_string(slpc));
+			return -EIO;
+		}
+	}
+
+	return 0;
+}
+
+static u32 slpc_decode_min_freq(struct intel_guc_slpc *slpc)
+{
+	struct slpc_shared_data *data = slpc->vaddr;
+
+	GEM_BUG_ON(!slpc->vma);
+
+	return	DIV_ROUND_CLOSEST(REG_FIELD_GET(SLPC_MIN_UNSLICE_FREQ_MASK,
+				  data->task_state_data.freq) *
+				  GT_FREQUENCY_MULTIPLIER, GEN9_FREQ_SCALER);
+}
+
+static u32 slpc_decode_max_freq(struct intel_guc_slpc *slpc)
+{
+	struct slpc_shared_data *data = slpc->vaddr;
+
+	GEM_BUG_ON(!slpc->vma);
+
+	return	DIV_ROUND_CLOSEST(REG_FIELD_GET(SLPC_MAX_UNSLICE_FREQ_MASK,
+				  data->task_state_data.freq) *
+				  GT_FREQUENCY_MULTIPLIER, GEN9_FREQ_SCALER);
+}
+
+static void slpc_shared_data_reset(struct slpc_shared_data *data)
+{
+	memset(data, 0, sizeof(struct slpc_shared_data));
+
+	data->header.size = sizeof(struct slpc_shared_data);
+
+	/* Enable only GTPERF task, disable others */
+	slpc_mem_set_enabled(data, SLPC_PARAM_TASK_ENABLE_GTPERF,
+			     SLPC_PARAM_TASK_DISABLE_GTPERF);
+
+	slpc_mem_set_disabled(data, SLPC_PARAM_TASK_ENABLE_BALANCER,
+			      SLPC_PARAM_TASK_DISABLE_BALANCER);
+
+	slpc_mem_set_disabled(data, SLPC_PARAM_TASK_ENABLE_DCC,
+			      SLPC_PARAM_TASK_DISABLE_DCC);
+}
+
+/**
+ * intel_guc_slpc_set_max_freq() - Set max frequency limit for SLPC.
+ * @slpc: pointer to intel_guc_slpc.
+ * @val: frequency (MHz)
+ *
+ * This function will invoke GuC SLPC action to update the max frequency
+ * limit for unslice.
+ *
+ * Return: 0 on success, non-zero error code on failure.
+ */
+int intel_guc_slpc_set_max_freq(struct intel_guc_slpc *slpc, u32 val)
+{
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	intel_wakeref_t wakeref;
+	int ret;
+
+	if (val < slpc->min_freq ||
+	    val > slpc->rp0_freq ||
+	    val < slpc->min_freq_softlimit)
+		return -EINVAL;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		ret = slpc_set_param(slpc,
+				     SLPC_PARAM_GLOBAL_MAX_GT_UNSLICE_FREQ_MHZ,
+				     val);
+
+		/* Return standardized err code for sysfs calls */
+		if (ret)
+			ret = -EIO;
+	}
+
+	if (!ret)
+		slpc->max_freq_softlimit = val;
+
+	return ret;
+}
+
+/**
+ * intel_guc_slpc_get_max_freq() - Get max frequency limit for SLPC.
+ * @slpc: pointer to intel_guc_slpc.
+ * @val: pointer to val which will hold max frequency (MHz)
+ *
+ * This function will invoke GuC SLPC action to read the max frequency
+ * limit for unslice.
+ *
+ * Return: 0 on success, non-zero error code on failure.
+ */
+int intel_guc_slpc_get_max_freq(struct intel_guc_slpc *slpc, u32 *val)
+{
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	intel_wakeref_t wakeref;
+	int ret = 0;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		/* Force GuC to update task data */
+		ret = slpc_query_task_state(slpc);
+
+		if (!ret)
+			*val = slpc_decode_max_freq(slpc);
+	}
+
+	return ret;
+}
+
+/**
+ * intel_guc_slpc_set_min_freq() - Set min frequency limit for SLPC.
+ * @slpc: pointer to intel_guc_slpc.
+ * @val: frequency (MHz)
+ *
+ * This function will invoke GuC SLPC action to update the min unslice
+ * frequency.
+ *
+ * Return: 0 on success, non-zero error code on failure.
+ */
+int intel_guc_slpc_set_min_freq(struct intel_guc_slpc *slpc, u32 val)
+{
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	intel_wakeref_t wakeref;
+	int ret;
+
+	if (val < slpc->min_freq ||
+	    val > slpc->rp0_freq ||
+	    val > slpc->max_freq_softlimit)
+		return -EINVAL;
+
+	/* Need a lock now since waitboost can be modifying min as well */
+	mutex_lock(&slpc->lock);
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	/* Ignore efficient freq if lower min freq is requested */
+	ret = slpc_set_param(slpc,
+			     SLPC_PARAM_IGNORE_EFFICIENT_FREQUENCY,
+			     val < slpc->rp1_freq);
+	if (ret) {
+		i915_probe_error(i915, "Failed to toggle efficient freq (%pe)\n",
+				 ERR_PTR(ret));
+		goto out;
+	}
+
+	ret = slpc_set_param(slpc,
+			     SLPC_PARAM_GLOBAL_MIN_GT_UNSLICE_FREQ_MHZ,
+			     val);
+
+	if (!ret)
+		slpc->min_freq_softlimit = val;
+
+out:
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+	mutex_unlock(&slpc->lock);
+
+	/* Return standardized err code for sysfs calls */
+	if (ret)
+		ret = -EIO;
+
+	return ret;
+}
+
+/**
+ * intel_guc_slpc_get_min_freq() - Get min frequency limit for SLPC.
+ * @slpc: pointer to intel_guc_slpc.
+ * @val: pointer to val which will hold min frequency (MHz)
+ *
+ * This function will invoke GuC SLPC action to read the min frequency
+ * limit for unslice.
+ *
+ * Return: 0 on success, non-zero error code on failure.
+ */
+int intel_guc_slpc_get_min_freq(struct intel_guc_slpc *slpc, u32 *val)
+{
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	intel_wakeref_t wakeref;
+	int ret = 0;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		/* Force GuC to update task data */
+		ret = slpc_query_task_state(slpc);
+
+		if (!ret)
+			*val = slpc_decode_min_freq(slpc);
+	}
+
+	return ret;
+}
+
+int intel_guc_slpc_set_media_ratio_mode(struct intel_guc_slpc *slpc, u32 val)
+{
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	intel_wakeref_t wakeref;
+	int ret = 0;
+
+	if (!HAS_MEDIA_RATIO_MODE(i915))
+		return -ENODEV;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
+		ret = slpc_set_param(slpc,
+				     SLPC_PARAM_MEDIA_FF_RATIO_MODE,
+				     val);
+	return ret;
+}
+
+void intel_guc_pm_intrmsk_enable(struct intel_gt *gt)
+{
+	u32 pm_intrmsk_mbz = 0;
+
+	/*
+	 * Allow GuC to receive ARAT timer expiry event.
+	 * This interrupt register is setup by RPS code
+	 * when host based Turbo is enabled.
+	 */
+	pm_intrmsk_mbz |= ARAT_EXPIRED_INTRMSK;
+
+	intel_uncore_rmw(gt->uncore,
+			 GEN6_PMINTRMSK, pm_intrmsk_mbz, 0);
+}
+
+static int slpc_set_softlimits(struct intel_guc_slpc *slpc)
+{
+	int ret = 0;
+
+	/*
+	 * Softlimits are initially equivalent to platform limits
+	 * unless they have deviated from defaults, in which case,
+	 * we retain the values and set min/max accordingly.
+	 */
+	if (!slpc->max_freq_softlimit) {
+		slpc->max_freq_softlimit = slpc->rp0_freq;
+		slpc_to_gt(slpc)->defaults.max_freq = slpc->max_freq_softlimit;
+	} else if (slpc->max_freq_softlimit != slpc->rp0_freq) {
+		ret = intel_guc_slpc_set_max_freq(slpc,
+						  slpc->max_freq_softlimit);
+	}
+
+	if (unlikely(ret))
+		return ret;
+
+	if (!slpc->min_freq_softlimit) {
+		ret = intel_guc_slpc_get_min_freq(slpc, &slpc->min_freq_softlimit);
+		if (unlikely(ret))
+			return ret;
+		slpc_to_gt(slpc)->defaults.min_freq = slpc->min_freq_softlimit;
+	} else {
+		return intel_guc_slpc_set_min_freq(slpc,
+						   slpc->min_freq_softlimit);
+	}
+
+	return 0;
+}
+
+static int slpc_use_fused_rp0(struct intel_guc_slpc *slpc)
+{
+	/* Force SLPC to used platform rp0 */
+	return slpc_set_param(slpc,
+			      SLPC_PARAM_GLOBAL_MAX_GT_UNSLICE_FREQ_MHZ,
+			      slpc->rp0_freq);
+}
+
+static void slpc_get_rp_values(struct intel_guc_slpc *slpc)
+{
+	struct intel_rps *rps = &slpc_to_gt(slpc)->rps;
+	struct intel_rps_freq_caps caps;
+
+	gen6_rps_get_freq_caps(rps, &caps);
+	slpc->rp0_freq = intel_gpu_freq(rps, caps.rp0_freq);
+	slpc->rp1_freq = intel_gpu_freq(rps, caps.rp1_freq);
+	slpc->min_freq = intel_gpu_freq(rps, caps.min_freq);
+
+	if (!slpc->boost_freq)
+		slpc->boost_freq = slpc->rp0_freq;
+}
+
+/*
+ * intel_guc_slpc_enable() - Start SLPC
+ * @slpc: pointer to intel_guc_slpc.
+ *
+ * SLPC is enabled by setting up the shared data structure and
+ * sending reset event to GuC SLPC. Initial data is setup in
+ * intel_guc_slpc_init. Here we send the reset event. We do
+ * not currently need a slpc_disable since this is taken care
+ * of automatically when a reset/suspend occurs and the GuC
+ * CTB is destroyed.
+ *
+ * Return: 0 on success, non-zero error code on failure.
+ */
+int intel_guc_slpc_enable(struct intel_guc_slpc *slpc)
+{
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	int ret;
+
+	GEM_BUG_ON(!slpc->vma);
+
+	slpc_shared_data_reset(slpc->vaddr);
+
+	ret = slpc_reset(slpc);
+	if (unlikely(ret < 0)) {
+		i915_probe_error(i915, "SLPC Reset event returned (%pe)\n",
+				 ERR_PTR(ret));
+		return ret;
+	}
+
+	ret = slpc_query_task_state(slpc);
+	if (unlikely(ret < 0))
+		return ret;
+
+	intel_guc_pm_intrmsk_enable(to_gt(i915));
+
+	slpc_get_rp_values(slpc);
+
+	/* Set SLPC max limit to RP0 */
+	ret = slpc_use_fused_rp0(slpc);
+	if (unlikely(ret)) {
+		i915_probe_error(i915, "Failed to set SLPC max to RP0 (%pe)\n",
+				 ERR_PTR(ret));
+		return ret;
+	}
+
+	/* Revert SLPC min/max to softlimits if necessary */
+	ret = slpc_set_softlimits(slpc);
+	if (unlikely(ret)) {
+		i915_probe_error(i915, "Failed to set SLPC softlimits (%pe)\n",
+				 ERR_PTR(ret));
+		return ret;
+	}
+
+	/* Set cached media freq ratio mode */
+	intel_guc_slpc_set_media_ratio_mode(slpc, slpc->media_ratio_mode);
+
+	return 0;
+}
+
+int intel_guc_slpc_set_boost_freq(struct intel_guc_slpc *slpc, u32 val)
+{
+	int ret = 0;
+
+	if (val < slpc->min_freq || val > slpc->rp0_freq)
+		return -EINVAL;
+
+	mutex_lock(&slpc->lock);
+
+	if (slpc->boost_freq != val) {
+		/* Apply only if there are active waiters */
+		if (atomic_read(&slpc->num_waiters)) {
+			ret = slpc_force_min_freq(slpc, val);
+			if (ret) {
+				ret = -EIO;
+				goto done;
+			}
+		}
+
+		slpc->boost_freq = val;
+	}
+
+done:
+	mutex_unlock(&slpc->lock);
+	return ret;
+}
+
+void intel_guc_slpc_dec_waiters(struct intel_guc_slpc *slpc)
+{
+	/*
+	 * Return min back to the softlimit.
+	 * This is called during request retire,
+	 * so we don't need to fail that if the
+	 * set_param fails.
+	 */
+	mutex_lock(&slpc->lock);
+	if (atomic_dec_and_test(&slpc->num_waiters))
+		slpc_force_min_freq(slpc, slpc->min_freq_softlimit);
+	mutex_unlock(&slpc->lock);
+}
+
+int intel_guc_slpc_print_info(struct intel_guc_slpc *slpc, struct drm_printer *p)
+{
+	struct drm_i915_private *i915 = slpc_to_i915(slpc);
+	struct slpc_shared_data *data = slpc->vaddr;
+	struct slpc_task_state_data *slpc_tasks;
+	intel_wakeref_t wakeref;
+	int ret = 0;
+
+	GEM_BUG_ON(!slpc->vma);
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		ret = slpc_query_task_state(slpc);
+
+		if (!ret) {
+			slpc_tasks = &data->task_state_data;
+
+			drm_printf(p, "\tSLPC state: %s\n", slpc_get_state_string(slpc));
+			drm_printf(p, "\tGTPERF task active: %s\n",
+				   str_yes_no(slpc_tasks->status & SLPC_GTPERF_TASK_ENABLED));
+			drm_printf(p, "\tMax freq: %u MHz\n",
+				   slpc_decode_max_freq(slpc));
+			drm_printf(p, "\tMin freq: %u MHz\n",
+				   slpc_decode_min_freq(slpc));
+			drm_printf(p, "\twaitboosts: %u\n",
+				   slpc->num_boosts);
+		}
+	}
+
+	return ret;
+}
+
+void intel_guc_slpc_fini(struct intel_guc_slpc *slpc)
+{
+	if (!slpc->vma)
+		return;
+
+	i915_vma_unpin_and_release(&slpc->vma, I915_VMA_RELEASE_MAP);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.h
new file mode 100644
index 000000000..82a98f78f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc.h
@@ -0,0 +1,46 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_SLPC_H_
+#define _INTEL_GUC_SLPC_H_
+
+#include "intel_guc_submission.h"
+#include "intel_guc_slpc_types.h"
+
+struct intel_gt;
+struct drm_printer;
+
+static inline bool intel_guc_slpc_is_supported(struct intel_guc *guc)
+{
+	return guc->slpc.supported;
+}
+
+static inline bool intel_guc_slpc_is_wanted(struct intel_guc *guc)
+{
+	return guc->slpc.selected;
+}
+
+static inline bool intel_guc_slpc_is_used(struct intel_guc *guc)
+{
+	return intel_guc_submission_is_used(guc) && intel_guc_slpc_is_wanted(guc);
+}
+
+void intel_guc_slpc_init_early(struct intel_guc_slpc *slpc);
+
+int intel_guc_slpc_init(struct intel_guc_slpc *slpc);
+int intel_guc_slpc_enable(struct intel_guc_slpc *slpc);
+void intel_guc_slpc_fini(struct intel_guc_slpc *slpc);
+int intel_guc_slpc_set_max_freq(struct intel_guc_slpc *slpc, u32 val);
+int intel_guc_slpc_set_min_freq(struct intel_guc_slpc *slpc, u32 val);
+int intel_guc_slpc_set_boost_freq(struct intel_guc_slpc *slpc, u32 val);
+int intel_guc_slpc_get_max_freq(struct intel_guc_slpc *slpc, u32 *val);
+int intel_guc_slpc_get_min_freq(struct intel_guc_slpc *slpc, u32 *val);
+int intel_guc_slpc_print_info(struct intel_guc_slpc *slpc, struct drm_printer *p);
+int intel_guc_slpc_set_media_ratio_mode(struct intel_guc_slpc *slpc, u32 val);
+void intel_guc_pm_intrmsk_enable(struct intel_gt *gt);
+void intel_guc_slpc_boost(struct intel_guc_slpc *slpc);
+void intel_guc_slpc_dec_waiters(struct intel_guc_slpc *slpc);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc_types.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc_types.h
new file mode 100644
index 000000000..73d208123
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_slpc_types.h
@@ -0,0 +1,45 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_SLPC_TYPES_H_
+#define _INTEL_GUC_SLPC_TYPES_H_
+
+#include <linux/atomic.h>
+#include <linux/workqueue.h>
+#include <linux/mutex.h>
+#include <linux/types.h>
+
+#define SLPC_RESET_TIMEOUT_MS 5
+
+struct intel_guc_slpc {
+	struct i915_vma *vma;
+	struct slpc_shared_data *vaddr;
+	bool supported;
+	bool selected;
+
+	/* platform frequency limits */
+	u32 min_freq;
+	u32 rp0_freq;
+	u32 rp1_freq;
+	u32 boost_freq;
+
+	/* frequency softlimits */
+	u32 min_freq_softlimit;
+	u32 max_freq_softlimit;
+
+	/* cached media ratio mode */
+	u32 media_ratio_mode;
+
+	/* Protects set/reset of boost freq
+	 * and value of num_waiters
+	 */
+	struct mutex lock;
+
+	struct work_struct boost_work;
+	atomic_t num_waiters;
+	u32 num_boosts;
+};
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
new file mode 100644
index 000000000..fecdc7ea7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
@@ -0,0 +1,5192 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014 Intel Corporation
+ */
+
+#include <linux/circ_buf.h>
+
+#include "gem/i915_gem_context.h"
+#include "gt/gen8_engine_cs.h"
+#include "gt/intel_breadcrumbs.h"
+#include "gt/intel_context.h"
+#include "gt/intel_engine_heartbeat.h"
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_clock_utils.h"
+#include "gt/intel_gt_irq.h"
+#include "gt/intel_gt_pm.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_gt_requests.h"
+#include "gt/intel_lrc.h"
+#include "gt/intel_lrc_reg.h"
+#include "gt/intel_mocs.h"
+#include "gt/intel_ring.h"
+
+#include "intel_guc_ads.h"
+#include "intel_guc_capture.h"
+#include "intel_guc_submission.h"
+
+#include "i915_drv.h"
+#include "i915_trace.h"
+
+/**
+ * DOC: GuC-based command submission
+ *
+ * The Scratch registers:
+ * There are 16 MMIO-based registers start from 0xC180. The kernel driver writes
+ * a value to the action register (SOFT_SCRATCH_0) along with any data. It then
+ * triggers an interrupt on the GuC via another register write (0xC4C8).
+ * Firmware writes a success/fail code back to the action register after
+ * processes the request. The kernel driver polls waiting for this update and
+ * then proceeds.
+ *
+ * Command Transport buffers (CTBs):
+ * Covered in detail in other sections but CTBs (Host to GuC - H2G, GuC to Host
+ * - G2H) are a message interface between the i915 and GuC.
+ *
+ * Context registration:
+ * Before a context can be submitted it must be registered with the GuC via a
+ * H2G. A unique guc_id is associated with each context. The context is either
+ * registered at request creation time (normal operation) or at submission time
+ * (abnormal operation, e.g. after a reset).
+ *
+ * Context submission:
+ * The i915 updates the LRC tail value in memory. The i915 must enable the
+ * scheduling of the context within the GuC for the GuC to actually consider it.
+ * Therefore, the first time a disabled context is submitted we use a schedule
+ * enable H2G, while follow up submissions are done via the context submit H2G,
+ * which informs the GuC that a previously enabled context has new work
+ * available.
+ *
+ * Context unpin:
+ * To unpin a context a H2G is used to disable scheduling. When the
+ * corresponding G2H returns indicating the scheduling disable operation has
+ * completed it is safe to unpin the context. While a disable is in flight it
+ * isn't safe to resubmit the context so a fence is used to stall all future
+ * requests of that context until the G2H is returned.
+ *
+ * Context deregistration:
+ * Before a context can be destroyed or if we steal its guc_id we must
+ * deregister the context with the GuC via H2G. If stealing the guc_id it isn't
+ * safe to submit anything to this guc_id until the deregister completes so a
+ * fence is used to stall all requests associated with this guc_id until the
+ * corresponding G2H returns indicating the guc_id has been deregistered.
+ *
+ * submission_state.guc_ids:
+ * Unique number associated with private GuC context data passed in during
+ * context registration / submission / deregistration. 64k available. Simple ida
+ * is used for allocation.
+ *
+ * Stealing guc_ids:
+ * If no guc_ids are available they can be stolen from another context at
+ * request creation time if that context is unpinned. If a guc_id can't be found
+ * we punt this problem to the user as we believe this is near impossible to hit
+ * during normal use cases.
+ *
+ * Locking:
+ * In the GuC submission code we have 3 basic spin locks which protect
+ * everything. Details about each below.
+ *
+ * sched_engine->lock
+ * This is the submission lock for all contexts that share an i915 schedule
+ * engine (sched_engine), thus only one of the contexts which share a
+ * sched_engine can be submitting at a time. Currently only one sched_engine is
+ * used for all of GuC submission but that could change in the future.
+ *
+ * guc->submission_state.lock
+ * Global lock for GuC submission state. Protects guc_ids and destroyed contexts
+ * list.
+ *
+ * ce->guc_state.lock
+ * Protects everything under ce->guc_state. Ensures that a context is in the
+ * correct state before issuing a H2G. e.g. We don't issue a schedule disable
+ * on a disabled context (bad idea), we don't issue a schedule enable when a
+ * schedule disable is in flight, etc... Also protects list of inflight requests
+ * on the context and the priority management state. Lock is individual to each
+ * context.
+ *
+ * Lock ordering rules:
+ * sched_engine->lock -> ce->guc_state.lock
+ * guc->submission_state.lock -> ce->guc_state.lock
+ *
+ * Reset races:
+ * When a full GT reset is triggered it is assumed that some G2H responses to
+ * H2Gs can be lost as the GuC is also reset. Losing these G2H can prove to be
+ * fatal as we do certain operations upon receiving a G2H (e.g. destroy
+ * contexts, release guc_ids, etc...). When this occurs we can scrub the
+ * context state and cleanup appropriately, however this is quite racey.
+ * To avoid races, the reset code must disable submission before scrubbing for
+ * the missing G2H, while the submission code must check for submission being
+ * disabled and skip sending H2Gs and updating context states when it is. Both
+ * sides must also make sure to hold the relevant locks.
+ */
+
+/* GuC Virtual Engine */
+struct guc_virtual_engine {
+	struct intel_engine_cs base;
+	struct intel_context context;
+};
+
+static struct intel_context *
+guc_create_virtual(struct intel_engine_cs **siblings, unsigned int count,
+		   unsigned long flags);
+
+static struct intel_context *
+guc_create_parallel(struct intel_engine_cs **engines,
+		    unsigned int num_siblings,
+		    unsigned int width);
+
+#define GUC_REQUEST_SIZE 64 /* bytes */
+
+/*
+ * We reserve 1/16 of the guc_ids for multi-lrc as these need to be contiguous
+ * per the GuC submission interface. A different allocation algorithm is used
+ * (bitmap vs. ida) between multi-lrc and single-lrc hence the reason to
+ * partition the guc_id space. We believe the number of multi-lrc contexts in
+ * use should be low and 1/16 should be sufficient. Minimum of 32 guc_ids for
+ * multi-lrc.
+ */
+#define NUMBER_MULTI_LRC_GUC_ID(guc)	\
+	((guc)->submission_state.num_guc_ids / 16)
+
+/*
+ * Below is a set of functions which control the GuC scheduling state which
+ * require a lock.
+ */
+#define SCHED_STATE_WAIT_FOR_DEREGISTER_TO_REGISTER	BIT(0)
+#define SCHED_STATE_DESTROYED				BIT(1)
+#define SCHED_STATE_PENDING_DISABLE			BIT(2)
+#define SCHED_STATE_BANNED				BIT(3)
+#define SCHED_STATE_ENABLED				BIT(4)
+#define SCHED_STATE_PENDING_ENABLE			BIT(5)
+#define SCHED_STATE_REGISTERED				BIT(6)
+#define SCHED_STATE_POLICY_REQUIRED			BIT(7)
+#define SCHED_STATE_BLOCKED_SHIFT			8
+#define SCHED_STATE_BLOCKED		BIT(SCHED_STATE_BLOCKED_SHIFT)
+#define SCHED_STATE_BLOCKED_MASK	(0xfff << SCHED_STATE_BLOCKED_SHIFT)
+
+static inline void init_sched_state(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state &= SCHED_STATE_BLOCKED_MASK;
+}
+
+__maybe_unused
+static bool sched_state_is_init(struct intel_context *ce)
+{
+	/* Kernel contexts can have SCHED_STATE_REGISTERED after suspend. */
+	return !(ce->guc_state.sched_state &
+		 ~(SCHED_STATE_BLOCKED_MASK | SCHED_STATE_REGISTERED));
+}
+
+static inline bool
+context_wait_for_deregister_to_register(struct intel_context *ce)
+{
+	return ce->guc_state.sched_state &
+		SCHED_STATE_WAIT_FOR_DEREGISTER_TO_REGISTER;
+}
+
+static inline void
+set_context_wait_for_deregister_to_register(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state |=
+		SCHED_STATE_WAIT_FOR_DEREGISTER_TO_REGISTER;
+}
+
+static inline void
+clr_context_wait_for_deregister_to_register(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state &=
+		~SCHED_STATE_WAIT_FOR_DEREGISTER_TO_REGISTER;
+}
+
+static inline bool
+context_destroyed(struct intel_context *ce)
+{
+	return ce->guc_state.sched_state & SCHED_STATE_DESTROYED;
+}
+
+static inline void
+set_context_destroyed(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state |= SCHED_STATE_DESTROYED;
+}
+
+static inline bool context_pending_disable(struct intel_context *ce)
+{
+	return ce->guc_state.sched_state & SCHED_STATE_PENDING_DISABLE;
+}
+
+static inline void set_context_pending_disable(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state |= SCHED_STATE_PENDING_DISABLE;
+}
+
+static inline void clr_context_pending_disable(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state &= ~SCHED_STATE_PENDING_DISABLE;
+}
+
+static inline bool context_banned(struct intel_context *ce)
+{
+	return ce->guc_state.sched_state & SCHED_STATE_BANNED;
+}
+
+static inline void set_context_banned(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state |= SCHED_STATE_BANNED;
+}
+
+static inline void clr_context_banned(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state &= ~SCHED_STATE_BANNED;
+}
+
+static inline bool context_enabled(struct intel_context *ce)
+{
+	return ce->guc_state.sched_state & SCHED_STATE_ENABLED;
+}
+
+static inline void set_context_enabled(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state |= SCHED_STATE_ENABLED;
+}
+
+static inline void clr_context_enabled(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state &= ~SCHED_STATE_ENABLED;
+}
+
+static inline bool context_pending_enable(struct intel_context *ce)
+{
+	return ce->guc_state.sched_state & SCHED_STATE_PENDING_ENABLE;
+}
+
+static inline void set_context_pending_enable(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state |= SCHED_STATE_PENDING_ENABLE;
+}
+
+static inline void clr_context_pending_enable(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state &= ~SCHED_STATE_PENDING_ENABLE;
+}
+
+static inline bool context_registered(struct intel_context *ce)
+{
+	return ce->guc_state.sched_state & SCHED_STATE_REGISTERED;
+}
+
+static inline void set_context_registered(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state |= SCHED_STATE_REGISTERED;
+}
+
+static inline void clr_context_registered(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state &= ~SCHED_STATE_REGISTERED;
+}
+
+static inline bool context_policy_required(struct intel_context *ce)
+{
+	return ce->guc_state.sched_state & SCHED_STATE_POLICY_REQUIRED;
+}
+
+static inline void set_context_policy_required(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state |= SCHED_STATE_POLICY_REQUIRED;
+}
+
+static inline void clr_context_policy_required(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	ce->guc_state.sched_state &= ~SCHED_STATE_POLICY_REQUIRED;
+}
+
+static inline u32 context_blocked(struct intel_context *ce)
+{
+	return (ce->guc_state.sched_state & SCHED_STATE_BLOCKED_MASK) >>
+		SCHED_STATE_BLOCKED_SHIFT;
+}
+
+static inline void incr_context_blocked(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+
+	ce->guc_state.sched_state += SCHED_STATE_BLOCKED;
+
+	GEM_BUG_ON(!context_blocked(ce));	/* Overflow check */
+}
+
+static inline void decr_context_blocked(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+
+	GEM_BUG_ON(!context_blocked(ce));	/* Underflow check */
+
+	ce->guc_state.sched_state -= SCHED_STATE_BLOCKED;
+}
+
+static inline bool context_has_committed_requests(struct intel_context *ce)
+{
+	return !!ce->guc_state.number_committed_requests;
+}
+
+static inline void incr_context_committed_requests(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	++ce->guc_state.number_committed_requests;
+	GEM_BUG_ON(ce->guc_state.number_committed_requests < 0);
+}
+
+static inline void decr_context_committed_requests(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	--ce->guc_state.number_committed_requests;
+	GEM_BUG_ON(ce->guc_state.number_committed_requests < 0);
+}
+
+static struct intel_context *
+request_to_scheduling_context(struct i915_request *rq)
+{
+	return intel_context_to_parent(rq->context);
+}
+
+static inline bool context_guc_id_invalid(struct intel_context *ce)
+{
+	return ce->guc_id.id == GUC_INVALID_CONTEXT_ID;
+}
+
+static inline void set_context_guc_id_invalid(struct intel_context *ce)
+{
+	ce->guc_id.id = GUC_INVALID_CONTEXT_ID;
+}
+
+static inline struct intel_guc *ce_to_guc(struct intel_context *ce)
+{
+	return &ce->engine->gt->uc.guc;
+}
+
+static inline struct i915_priolist *to_priolist(struct rb_node *rb)
+{
+	return rb_entry(rb, struct i915_priolist, node);
+}
+
+/*
+ * When using multi-lrc submission a scratch memory area is reserved in the
+ * parent's context state for the process descriptor, work queue, and handshake
+ * between the parent + children contexts to insert safe preemption points
+ * between each of the BBs. Currently the scratch area is sized to a page.
+ *
+ * The layout of this scratch area is below:
+ * 0						guc_process_desc
+ * + sizeof(struct guc_process_desc)		child go
+ * + CACHELINE_BYTES				child join[0]
+ * ...
+ * + CACHELINE_BYTES				child join[n - 1]
+ * ...						unused
+ * PARENT_SCRATCH_SIZE / 2			work queue start
+ * ...						work queue
+ * PARENT_SCRATCH_SIZE - 1			work queue end
+ */
+#define WQ_SIZE			(PARENT_SCRATCH_SIZE / 2)
+#define WQ_OFFSET		(PARENT_SCRATCH_SIZE - WQ_SIZE)
+
+struct sync_semaphore {
+	u32 semaphore;
+	u8 unused[CACHELINE_BYTES - sizeof(u32)];
+};
+
+struct parent_scratch {
+	union guc_descs {
+		struct guc_sched_wq_desc wq_desc;
+		struct guc_process_desc_v69 pdesc;
+	} descs;
+
+	struct sync_semaphore go;
+	struct sync_semaphore join[MAX_ENGINE_INSTANCE + 1];
+
+	u8 unused[WQ_OFFSET - sizeof(union guc_descs) -
+		sizeof(struct sync_semaphore) * (MAX_ENGINE_INSTANCE + 2)];
+
+	u32 wq[WQ_SIZE / sizeof(u32)];
+};
+
+static u32 __get_parent_scratch_offset(struct intel_context *ce)
+{
+	GEM_BUG_ON(!ce->parallel.guc.parent_page);
+
+	return ce->parallel.guc.parent_page * PAGE_SIZE;
+}
+
+static u32 __get_wq_offset(struct intel_context *ce)
+{
+	BUILD_BUG_ON(offsetof(struct parent_scratch, wq) != WQ_OFFSET);
+
+	return __get_parent_scratch_offset(ce) + WQ_OFFSET;
+}
+
+static struct parent_scratch *
+__get_parent_scratch(struct intel_context *ce)
+{
+	BUILD_BUG_ON(sizeof(struct parent_scratch) != PARENT_SCRATCH_SIZE);
+	BUILD_BUG_ON(sizeof(struct sync_semaphore) != CACHELINE_BYTES);
+
+	/*
+	 * Need to subtract LRC_STATE_OFFSET here as the
+	 * parallel.guc.parent_page is the offset into ce->state while
+	 * ce->lrc_reg_reg is ce->state + LRC_STATE_OFFSET.
+	 */
+	return (struct parent_scratch *)
+		(ce->lrc_reg_state +
+		 ((__get_parent_scratch_offset(ce) -
+		   LRC_STATE_OFFSET) / sizeof(u32)));
+}
+
+static struct guc_process_desc_v69 *
+__get_process_desc_v69(struct intel_context *ce)
+{
+	struct parent_scratch *ps = __get_parent_scratch(ce);
+
+	return &ps->descs.pdesc;
+}
+
+static struct guc_sched_wq_desc *
+__get_wq_desc_v70(struct intel_context *ce)
+{
+	struct parent_scratch *ps = __get_parent_scratch(ce);
+
+	return &ps->descs.wq_desc;
+}
+
+static u32 *get_wq_pointer(struct intel_context *ce, u32 wqi_size)
+{
+	/*
+	 * Check for space in work queue. Caching a value of head pointer in
+	 * intel_context structure in order reduce the number accesses to shared
+	 * GPU memory which may be across a PCIe bus.
+	 */
+#define AVAILABLE_SPACE	\
+	CIRC_SPACE(ce->parallel.guc.wqi_tail, ce->parallel.guc.wqi_head, WQ_SIZE)
+	if (wqi_size > AVAILABLE_SPACE) {
+		ce->parallel.guc.wqi_head = READ_ONCE(*ce->parallel.guc.wq_head);
+
+		if (wqi_size > AVAILABLE_SPACE)
+			return NULL;
+	}
+#undef AVAILABLE_SPACE
+
+	return &__get_parent_scratch(ce)->wq[ce->parallel.guc.wqi_tail / sizeof(u32)];
+}
+
+static inline struct intel_context *__get_context(struct intel_guc *guc, u32 id)
+{
+	struct intel_context *ce = xa_load(&guc->context_lookup, id);
+
+	GEM_BUG_ON(id >= GUC_MAX_CONTEXT_ID);
+
+	return ce;
+}
+
+static struct guc_lrc_desc_v69 *__get_lrc_desc_v69(struct intel_guc *guc, u32 index)
+{
+	struct guc_lrc_desc_v69 *base = guc->lrc_desc_pool_vaddr_v69;
+
+	if (!base)
+		return NULL;
+
+	GEM_BUG_ON(index >= GUC_MAX_CONTEXT_ID);
+
+	return &base[index];
+}
+
+static int guc_lrc_desc_pool_create_v69(struct intel_guc *guc)
+{
+	u32 size;
+	int ret;
+
+	size = PAGE_ALIGN(sizeof(struct guc_lrc_desc_v69) *
+			  GUC_MAX_CONTEXT_ID);
+	ret = intel_guc_allocate_and_map_vma(guc, size, &guc->lrc_desc_pool_v69,
+					     (void **)&guc->lrc_desc_pool_vaddr_v69);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static void guc_lrc_desc_pool_destroy_v69(struct intel_guc *guc)
+{
+	if (!guc->lrc_desc_pool_vaddr_v69)
+		return;
+
+	guc->lrc_desc_pool_vaddr_v69 = NULL;
+	i915_vma_unpin_and_release(&guc->lrc_desc_pool_v69, I915_VMA_RELEASE_MAP);
+}
+
+static inline bool guc_submission_initialized(struct intel_guc *guc)
+{
+	return guc->submission_initialized;
+}
+
+static inline void _reset_lrc_desc_v69(struct intel_guc *guc, u32 id)
+{
+	struct guc_lrc_desc_v69 *desc = __get_lrc_desc_v69(guc, id);
+
+	if (desc)
+		memset(desc, 0, sizeof(*desc));
+}
+
+static inline bool ctx_id_mapped(struct intel_guc *guc, u32 id)
+{
+	return __get_context(guc, id);
+}
+
+static inline void set_ctx_id_mapping(struct intel_guc *guc, u32 id,
+				      struct intel_context *ce)
+{
+	unsigned long flags;
+
+	/*
+	 * xarray API doesn't have xa_save_irqsave wrapper, so calling the
+	 * lower level functions directly.
+	 */
+	xa_lock_irqsave(&guc->context_lookup, flags);
+	__xa_store(&guc->context_lookup, id, ce, GFP_ATOMIC);
+	xa_unlock_irqrestore(&guc->context_lookup, flags);
+}
+
+static inline void clr_ctx_id_mapping(struct intel_guc *guc, u32 id)
+{
+	unsigned long flags;
+
+	if (unlikely(!guc_submission_initialized(guc)))
+		return;
+
+	_reset_lrc_desc_v69(guc, id);
+
+	/*
+	 * xarray API doesn't have xa_erase_irqsave wrapper, so calling
+	 * the lower level functions directly.
+	 */
+	xa_lock_irqsave(&guc->context_lookup, flags);
+	__xa_erase(&guc->context_lookup, id);
+	xa_unlock_irqrestore(&guc->context_lookup, flags);
+}
+
+static void decr_outstanding_submission_g2h(struct intel_guc *guc)
+{
+	if (atomic_dec_and_test(&guc->outstanding_submission_g2h))
+		wake_up_all(&guc->ct.wq);
+}
+
+static int guc_submission_send_busy_loop(struct intel_guc *guc,
+					 const u32 *action,
+					 u32 len,
+					 u32 g2h_len_dw,
+					 bool loop)
+{
+	/*
+	 * We always loop when a send requires a reply (i.e. g2h_len_dw > 0),
+	 * so we don't handle the case where we don't get a reply because we
+	 * aborted the send due to the channel being busy.
+	 */
+	GEM_BUG_ON(g2h_len_dw && !loop);
+
+	if (g2h_len_dw)
+		atomic_inc(&guc->outstanding_submission_g2h);
+
+	return intel_guc_send_busy_loop(guc, action, len, g2h_len_dw, loop);
+}
+
+int intel_guc_wait_for_pending_msg(struct intel_guc *guc,
+				   atomic_t *wait_var,
+				   bool interruptible,
+				   long timeout)
+{
+	const int state = interruptible ?
+		TASK_INTERRUPTIBLE : TASK_UNINTERRUPTIBLE;
+	DEFINE_WAIT(wait);
+
+	might_sleep();
+	GEM_BUG_ON(timeout < 0);
+
+	if (!atomic_read(wait_var))
+		return 0;
+
+	if (!timeout)
+		return -ETIME;
+
+	for (;;) {
+		prepare_to_wait(&guc->ct.wq, &wait, state);
+
+		if (!atomic_read(wait_var))
+			break;
+
+		if (signal_pending_state(state, current)) {
+			timeout = -EINTR;
+			break;
+		}
+
+		if (!timeout) {
+			timeout = -ETIME;
+			break;
+		}
+
+		timeout = io_schedule_timeout(timeout);
+	}
+	finish_wait(&guc->ct.wq, &wait);
+
+	return (timeout < 0) ? timeout : 0;
+}
+
+int intel_guc_wait_for_idle(struct intel_guc *guc, long timeout)
+{
+	if (!intel_uc_uses_guc_submission(&guc_to_gt(guc)->uc))
+		return 0;
+
+	return intel_guc_wait_for_pending_msg(guc,
+					      &guc->outstanding_submission_g2h,
+					      true, timeout);
+}
+
+static int guc_context_policy_init_v70(struct intel_context *ce, bool loop);
+static int try_context_registration(struct intel_context *ce, bool loop);
+
+static int __guc_add_request(struct intel_guc *guc, struct i915_request *rq)
+{
+	int err = 0;
+	struct intel_context *ce = request_to_scheduling_context(rq);
+	u32 action[3];
+	int len = 0;
+	u32 g2h_len_dw = 0;
+	bool enabled;
+
+	lockdep_assert_held(&rq->engine->sched_engine->lock);
+
+	/*
+	 * Corner case where requests were sitting in the priority list or a
+	 * request resubmitted after the context was banned.
+	 */
+	if (unlikely(!intel_context_is_schedulable(ce))) {
+		i915_request_put(i915_request_mark_eio(rq));
+		intel_engine_signal_breadcrumbs(ce->engine);
+		return 0;
+	}
+
+	GEM_BUG_ON(!atomic_read(&ce->guc_id.ref));
+	GEM_BUG_ON(context_guc_id_invalid(ce));
+
+	if (context_policy_required(ce)) {
+		err = guc_context_policy_init_v70(ce, false);
+		if (err)
+			return err;
+	}
+
+	spin_lock(&ce->guc_state.lock);
+
+	/*
+	 * The request / context will be run on the hardware when scheduling
+	 * gets enabled in the unblock. For multi-lrc we still submit the
+	 * context to move the LRC tails.
+	 */
+	if (unlikely(context_blocked(ce) && !intel_context_is_parent(ce)))
+		goto out;
+
+	enabled = context_enabled(ce) || context_blocked(ce);
+
+	if (!enabled) {
+		action[len++] = INTEL_GUC_ACTION_SCHED_CONTEXT_MODE_SET;
+		action[len++] = ce->guc_id.id;
+		action[len++] = GUC_CONTEXT_ENABLE;
+		set_context_pending_enable(ce);
+		intel_context_get(ce);
+		g2h_len_dw = G2H_LEN_DW_SCHED_CONTEXT_MODE_SET;
+	} else {
+		action[len++] = INTEL_GUC_ACTION_SCHED_CONTEXT;
+		action[len++] = ce->guc_id.id;
+	}
+
+	err = intel_guc_send_nb(guc, action, len, g2h_len_dw);
+	if (!enabled && !err) {
+		trace_intel_context_sched_enable(ce);
+		atomic_inc(&guc->outstanding_submission_g2h);
+		set_context_enabled(ce);
+
+		/*
+		 * Without multi-lrc KMD does the submission step (moving the
+		 * lrc tail) so enabling scheduling is sufficient to submit the
+		 * context. This isn't the case in multi-lrc submission as the
+		 * GuC needs to move the tails, hence the need for another H2G
+		 * to submit a multi-lrc context after enabling scheduling.
+		 */
+		if (intel_context_is_parent(ce)) {
+			action[0] = INTEL_GUC_ACTION_SCHED_CONTEXT;
+			err = intel_guc_send_nb(guc, action, len - 1, 0);
+		}
+	} else if (!enabled) {
+		clr_context_pending_enable(ce);
+		intel_context_put(ce);
+	}
+	if (likely(!err))
+		trace_i915_request_guc_submit(rq);
+
+out:
+	spin_unlock(&ce->guc_state.lock);
+	return err;
+}
+
+static int guc_add_request(struct intel_guc *guc, struct i915_request *rq)
+{
+	int ret = __guc_add_request(guc, rq);
+
+	if (unlikely(ret == -EBUSY)) {
+		guc->stalled_request = rq;
+		guc->submission_stall_reason = STALL_ADD_REQUEST;
+	}
+
+	return ret;
+}
+
+static inline void guc_set_lrc_tail(struct i915_request *rq)
+{
+	rq->context->lrc_reg_state[CTX_RING_TAIL] =
+		intel_ring_set_tail(rq->ring, rq->tail);
+}
+
+static inline int rq_prio(const struct i915_request *rq)
+{
+	return rq->sched.attr.priority;
+}
+
+static bool is_multi_lrc_rq(struct i915_request *rq)
+{
+	return intel_context_is_parallel(rq->context);
+}
+
+static bool can_merge_rq(struct i915_request *rq,
+			 struct i915_request *last)
+{
+	return request_to_scheduling_context(rq) ==
+		request_to_scheduling_context(last);
+}
+
+static u32 wq_space_until_wrap(struct intel_context *ce)
+{
+	return (WQ_SIZE - ce->parallel.guc.wqi_tail);
+}
+
+static void write_wqi(struct intel_context *ce, u32 wqi_size)
+{
+	BUILD_BUG_ON(!is_power_of_2(WQ_SIZE));
+
+	/*
+	 * Ensure WQI are visible before updating tail
+	 */
+	intel_guc_write_barrier(ce_to_guc(ce));
+
+	ce->parallel.guc.wqi_tail = (ce->parallel.guc.wqi_tail + wqi_size) &
+		(WQ_SIZE - 1);
+	WRITE_ONCE(*ce->parallel.guc.wq_tail, ce->parallel.guc.wqi_tail);
+}
+
+static int guc_wq_noop_append(struct intel_context *ce)
+{
+	u32 *wqi = get_wq_pointer(ce, wq_space_until_wrap(ce));
+	u32 len_dw = wq_space_until_wrap(ce) / sizeof(u32) - 1;
+
+	if (!wqi)
+		return -EBUSY;
+
+	GEM_BUG_ON(!FIELD_FIT(WQ_LEN_MASK, len_dw));
+
+	*wqi = FIELD_PREP(WQ_TYPE_MASK, WQ_TYPE_NOOP) |
+		FIELD_PREP(WQ_LEN_MASK, len_dw);
+	ce->parallel.guc.wqi_tail = 0;
+
+	return 0;
+}
+
+static int __guc_wq_item_append(struct i915_request *rq)
+{
+	struct intel_context *ce = request_to_scheduling_context(rq);
+	struct intel_context *child;
+	unsigned int wqi_size = (ce->parallel.number_children + 4) *
+		sizeof(u32);
+	u32 *wqi;
+	u32 len_dw = (wqi_size / sizeof(u32)) - 1;
+	int ret;
+
+	/* Ensure context is in correct state updating work queue */
+	GEM_BUG_ON(!atomic_read(&ce->guc_id.ref));
+	GEM_BUG_ON(context_guc_id_invalid(ce));
+	GEM_BUG_ON(context_wait_for_deregister_to_register(ce));
+	GEM_BUG_ON(!ctx_id_mapped(ce_to_guc(ce), ce->guc_id.id));
+
+	/* Insert NOOP if this work queue item will wrap the tail pointer. */
+	if (wqi_size > wq_space_until_wrap(ce)) {
+		ret = guc_wq_noop_append(ce);
+		if (ret)
+			return ret;
+	}
+
+	wqi = get_wq_pointer(ce, wqi_size);
+	if (!wqi)
+		return -EBUSY;
+
+	GEM_BUG_ON(!FIELD_FIT(WQ_LEN_MASK, len_dw));
+
+	*wqi++ = FIELD_PREP(WQ_TYPE_MASK, WQ_TYPE_MULTI_LRC) |
+		FIELD_PREP(WQ_LEN_MASK, len_dw);
+	*wqi++ = ce->lrc.lrca;
+	*wqi++ = FIELD_PREP(WQ_GUC_ID_MASK, ce->guc_id.id) |
+	       FIELD_PREP(WQ_RING_TAIL_MASK, ce->ring->tail / sizeof(u64));
+	*wqi++ = 0;	/* fence_id */
+	for_each_child(ce, child)
+		*wqi++ = child->ring->tail / sizeof(u64);
+
+	write_wqi(ce, wqi_size);
+
+	return 0;
+}
+
+static int guc_wq_item_append(struct intel_guc *guc,
+			      struct i915_request *rq)
+{
+	struct intel_context *ce = request_to_scheduling_context(rq);
+	int ret;
+
+	if (unlikely(!intel_context_is_schedulable(ce)))
+		return 0;
+
+	ret = __guc_wq_item_append(rq);
+	if (unlikely(ret == -EBUSY)) {
+		guc->stalled_request = rq;
+		guc->submission_stall_reason = STALL_MOVE_LRC_TAIL;
+	}
+
+	return ret;
+}
+
+static bool multi_lrc_submit(struct i915_request *rq)
+{
+	struct intel_context *ce = request_to_scheduling_context(rq);
+
+	intel_ring_set_tail(rq->ring, rq->tail);
+
+	/*
+	 * We expect the front end (execbuf IOCTL) to set this flag on the last
+	 * request generated from a multi-BB submission. This indicates to the
+	 * backend (GuC interface) that we should submit this context thus
+	 * submitting all the requests generated in parallel.
+	 */
+	return test_bit(I915_FENCE_FLAG_SUBMIT_PARALLEL, &rq->fence.flags) ||
+	       !intel_context_is_schedulable(ce);
+}
+
+static int guc_dequeue_one_context(struct intel_guc *guc)
+{
+	struct i915_sched_engine * const sched_engine = guc->sched_engine;
+	struct i915_request *last = NULL;
+	bool submit = false;
+	struct rb_node *rb;
+	int ret;
+
+	lockdep_assert_held(&sched_engine->lock);
+
+	if (guc->stalled_request) {
+		submit = true;
+		last = guc->stalled_request;
+
+		switch (guc->submission_stall_reason) {
+		case STALL_REGISTER_CONTEXT:
+			goto register_context;
+		case STALL_MOVE_LRC_TAIL:
+			goto move_lrc_tail;
+		case STALL_ADD_REQUEST:
+			goto add_request;
+		default:
+			MISSING_CASE(guc->submission_stall_reason);
+		}
+	}
+
+	while ((rb = rb_first_cached(&sched_engine->queue))) {
+		struct i915_priolist *p = to_priolist(rb);
+		struct i915_request *rq, *rn;
+
+		priolist_for_each_request_consume(rq, rn, p) {
+			if (last && !can_merge_rq(rq, last))
+				goto register_context;
+
+			list_del_init(&rq->sched.link);
+
+			__i915_request_submit(rq);
+
+			trace_i915_request_in(rq, 0);
+			last = rq;
+
+			if (is_multi_lrc_rq(rq)) {
+				/*
+				 * We need to coalesce all multi-lrc requests in
+				 * a relationship into a single H2G. We are
+				 * guaranteed that all of these requests will be
+				 * submitted sequentially.
+				 */
+				if (multi_lrc_submit(rq)) {
+					submit = true;
+					goto register_context;
+				}
+			} else {
+				submit = true;
+			}
+		}
+
+		rb_erase_cached(&p->node, &sched_engine->queue);
+		i915_priolist_free(p);
+	}
+
+register_context:
+	if (submit) {
+		struct intel_context *ce = request_to_scheduling_context(last);
+
+		if (unlikely(!ctx_id_mapped(guc, ce->guc_id.id) &&
+			     intel_context_is_schedulable(ce))) {
+			ret = try_context_registration(ce, false);
+			if (unlikely(ret == -EPIPE)) {
+				goto deadlk;
+			} else if (ret == -EBUSY) {
+				guc->stalled_request = last;
+				guc->submission_stall_reason =
+					STALL_REGISTER_CONTEXT;
+				goto schedule_tasklet;
+			} else if (ret != 0) {
+				GEM_WARN_ON(ret);	/* Unexpected */
+				goto deadlk;
+			}
+		}
+
+move_lrc_tail:
+		if (is_multi_lrc_rq(last)) {
+			ret = guc_wq_item_append(guc, last);
+			if (ret == -EBUSY) {
+				goto schedule_tasklet;
+			} else if (ret != 0) {
+				GEM_WARN_ON(ret);	/* Unexpected */
+				goto deadlk;
+			}
+		} else {
+			guc_set_lrc_tail(last);
+		}
+
+add_request:
+		ret = guc_add_request(guc, last);
+		if (unlikely(ret == -EPIPE)) {
+			goto deadlk;
+		} else if (ret == -EBUSY) {
+			goto schedule_tasklet;
+		} else if (ret != 0) {
+			GEM_WARN_ON(ret);	/* Unexpected */
+			goto deadlk;
+		}
+	}
+
+	guc->stalled_request = NULL;
+	guc->submission_stall_reason = STALL_NONE;
+	return submit;
+
+deadlk:
+	sched_engine->tasklet.callback = NULL;
+	tasklet_disable_nosync(&sched_engine->tasklet);
+	return false;
+
+schedule_tasklet:
+	tasklet_schedule(&sched_engine->tasklet);
+	return false;
+}
+
+static void guc_submission_tasklet(struct tasklet_struct *t)
+{
+	struct i915_sched_engine *sched_engine =
+		from_tasklet(sched_engine, t, tasklet);
+	unsigned long flags;
+	bool loop;
+
+	spin_lock_irqsave(&sched_engine->lock, flags);
+
+	do {
+		loop = guc_dequeue_one_context(sched_engine->private_data);
+	} while (loop);
+
+	i915_sched_engine_reset_on_empty(sched_engine);
+
+	spin_unlock_irqrestore(&sched_engine->lock, flags);
+}
+
+static void cs_irq_handler(struct intel_engine_cs *engine, u16 iir)
+{
+	if (iir & GT_RENDER_USER_INTERRUPT)
+		intel_engine_signal_breadcrumbs(engine);
+}
+
+static void __guc_context_destroy(struct intel_context *ce);
+static void release_guc_id(struct intel_guc *guc, struct intel_context *ce);
+static void guc_signal_context_fence(struct intel_context *ce);
+static void guc_cancel_context_requests(struct intel_context *ce);
+static void guc_blocked_fence_complete(struct intel_context *ce);
+
+static void scrub_guc_desc_for_outstanding_g2h(struct intel_guc *guc)
+{
+	struct intel_context *ce;
+	unsigned long index, flags;
+	bool pending_disable, pending_enable, deregister, destroyed, banned;
+
+	xa_lock_irqsave(&guc->context_lookup, flags);
+	xa_for_each(&guc->context_lookup, index, ce) {
+		/*
+		 * Corner case where the ref count on the object is zero but and
+		 * deregister G2H was lost. In this case we don't touch the ref
+		 * count and finish the destroy of the context.
+		 */
+		bool do_put = kref_get_unless_zero(&ce->ref);
+
+		xa_unlock(&guc->context_lookup);
+
+		spin_lock(&ce->guc_state.lock);
+
+		/*
+		 * Once we are at this point submission_disabled() is guaranteed
+		 * to be visible to all callers who set the below flags (see above
+		 * flush and flushes in reset_prepare). If submission_disabled()
+		 * is set, the caller shouldn't set these flags.
+		 */
+
+		destroyed = context_destroyed(ce);
+		pending_enable = context_pending_enable(ce);
+		pending_disable = context_pending_disable(ce);
+		deregister = context_wait_for_deregister_to_register(ce);
+		banned = context_banned(ce);
+		init_sched_state(ce);
+
+		spin_unlock(&ce->guc_state.lock);
+
+		if (pending_enable || destroyed || deregister) {
+			decr_outstanding_submission_g2h(guc);
+			if (deregister)
+				guc_signal_context_fence(ce);
+			if (destroyed) {
+				intel_gt_pm_put_async(guc_to_gt(guc));
+				release_guc_id(guc, ce);
+				__guc_context_destroy(ce);
+			}
+			if (pending_enable || deregister)
+				intel_context_put(ce);
+		}
+
+		/* Not mutualy exclusive with above if statement. */
+		if (pending_disable) {
+			guc_signal_context_fence(ce);
+			if (banned) {
+				guc_cancel_context_requests(ce);
+				intel_engine_signal_breadcrumbs(ce->engine);
+			}
+			intel_context_sched_disable_unpin(ce);
+			decr_outstanding_submission_g2h(guc);
+
+			spin_lock(&ce->guc_state.lock);
+			guc_blocked_fence_complete(ce);
+			spin_unlock(&ce->guc_state.lock);
+
+			intel_context_put(ce);
+		}
+
+		if (do_put)
+			intel_context_put(ce);
+		xa_lock(&guc->context_lookup);
+	}
+	xa_unlock_irqrestore(&guc->context_lookup, flags);
+}
+
+/*
+ * GuC stores busyness stats for each engine at context in/out boundaries. A
+ * context 'in' logs execution start time, 'out' adds in -> out delta to total.
+ * i915/kmd accesses 'start', 'total' and 'context id' from memory shared with
+ * GuC.
+ *
+ * __i915_pmu_event_read samples engine busyness. When sampling, if context id
+ * is valid (!= ~0) and start is non-zero, the engine is considered to be
+ * active. For an active engine total busyness = total + (now - start), where
+ * 'now' is the time at which the busyness is sampled. For inactive engine,
+ * total busyness = total.
+ *
+ * All times are captured from GUCPMTIMESTAMP reg and are in gt clock domain.
+ *
+ * The start and total values provided by GuC are 32 bits and wrap around in a
+ * few minutes. Since perf pmu provides busyness as 64 bit monotonically
+ * increasing ns values, there is a need for this implementation to account for
+ * overflows and extend the GuC provided values to 64 bits before returning
+ * busyness to the user. In order to do that, a worker runs periodically at
+ * frequency = 1/8th the time it takes for the timestamp to wrap (i.e. once in
+ * 27 seconds for a gt clock frequency of 19.2 MHz).
+ */
+
+#define WRAP_TIME_CLKS U32_MAX
+#define POLL_TIME_CLKS (WRAP_TIME_CLKS >> 3)
+
+static void
+__extend_last_switch(struct intel_guc *guc, u64 *prev_start, u32 new_start)
+{
+	u32 gt_stamp_hi = upper_32_bits(guc->timestamp.gt_stamp);
+	u32 gt_stamp_last = lower_32_bits(guc->timestamp.gt_stamp);
+
+	if (new_start == lower_32_bits(*prev_start))
+		return;
+
+	/*
+	 * When gt is unparked, we update the gt timestamp and start the ping
+	 * worker that updates the gt_stamp every POLL_TIME_CLKS. As long as gt
+	 * is unparked, all switched in contexts will have a start time that is
+	 * within +/- POLL_TIME_CLKS of the most recent gt_stamp.
+	 *
+	 * If neither gt_stamp nor new_start has rolled over, then the
+	 * gt_stamp_hi does not need to be adjusted, however if one of them has
+	 * rolled over, we need to adjust gt_stamp_hi accordingly.
+	 *
+	 * The below conditions address the cases of new_start rollover and
+	 * gt_stamp_last rollover respectively.
+	 */
+	if (new_start < gt_stamp_last &&
+	    (new_start - gt_stamp_last) <= POLL_TIME_CLKS)
+		gt_stamp_hi++;
+
+	if (new_start > gt_stamp_last &&
+	    (gt_stamp_last - new_start) <= POLL_TIME_CLKS && gt_stamp_hi)
+		gt_stamp_hi--;
+
+	*prev_start = ((u64)gt_stamp_hi << 32) | new_start;
+}
+
+#define record_read(map_, field_) \
+	iosys_map_rd_field(map_, 0, struct guc_engine_usage_record, field_)
+
+/*
+ * GuC updates shared memory and KMD reads it. Since this is not synchronized,
+ * we run into a race where the value read is inconsistent. Sometimes the
+ * inconsistency is in reading the upper MSB bytes of the last_in value when
+ * this race occurs. 2 types of cases are seen - upper 8 bits are zero and upper
+ * 24 bits are zero. Since these are non-zero values, it is non-trivial to
+ * determine validity of these values. Instead we read the values multiple times
+ * until they are consistent. In test runs, 3 attempts results in consistent
+ * values. The upper bound is set to 6 attempts and may need to be tuned as per
+ * any new occurences.
+ */
+static void __get_engine_usage_record(struct intel_engine_cs *engine,
+				      u32 *last_in, u32 *id, u32 *total)
+{
+	struct iosys_map rec_map = intel_guc_engine_usage_record_map(engine);
+	int i = 0;
+
+	do {
+		*last_in = record_read(&rec_map, last_switch_in_stamp);
+		*id = record_read(&rec_map, current_context_index);
+		*total = record_read(&rec_map, total_runtime);
+
+		if (record_read(&rec_map, last_switch_in_stamp) == *last_in &&
+		    record_read(&rec_map, current_context_index) == *id &&
+		    record_read(&rec_map, total_runtime) == *total)
+			break;
+	} while (++i < 6);
+}
+
+static void guc_update_engine_gt_clks(struct intel_engine_cs *engine)
+{
+	struct intel_engine_guc_stats *stats = &engine->stats.guc;
+	struct intel_guc *guc = &engine->gt->uc.guc;
+	u32 last_switch, ctx_id, total;
+
+	lockdep_assert_held(&guc->timestamp.lock);
+
+	__get_engine_usage_record(engine, &last_switch, &ctx_id, &total);
+
+	stats->running = ctx_id != ~0U && last_switch;
+	if (stats->running)
+		__extend_last_switch(guc, &stats->start_gt_clk, last_switch);
+
+	/*
+	 * Instead of adjusting the total for overflow, just add the
+	 * difference from previous sample stats->total_gt_clks
+	 */
+	if (total && total != ~0U) {
+		stats->total_gt_clks += (u32)(total - stats->prev_total);
+		stats->prev_total = total;
+	}
+}
+
+static u32 gpm_timestamp_shift(struct intel_gt *gt)
+{
+	intel_wakeref_t wakeref;
+	u32 reg, shift;
+
+	with_intel_runtime_pm(gt->uncore->rpm, wakeref)
+		reg = intel_uncore_read(gt->uncore, RPM_CONFIG0);
+
+	shift = (reg & GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_MASK) >>
+		GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_SHIFT;
+
+	return 3 - shift;
+}
+
+static void guc_update_pm_timestamp(struct intel_guc *guc, ktime_t *now)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	u32 gt_stamp_lo, gt_stamp_hi;
+	u64 gpm_ts;
+
+	lockdep_assert_held(&guc->timestamp.lock);
+
+	gt_stamp_hi = upper_32_bits(guc->timestamp.gt_stamp);
+	gpm_ts = intel_uncore_read64_2x32(gt->uncore, MISC_STATUS0,
+					  MISC_STATUS1) >> guc->timestamp.shift;
+	gt_stamp_lo = lower_32_bits(gpm_ts);
+	*now = ktime_get();
+
+	if (gt_stamp_lo < lower_32_bits(guc->timestamp.gt_stamp))
+		gt_stamp_hi++;
+
+	guc->timestamp.gt_stamp = ((u64)gt_stamp_hi << 32) | gt_stamp_lo;
+}
+
+/*
+ * Unlike the execlist mode of submission total and active times are in terms of
+ * gt clocks. The *now parameter is retained to return the cpu time at which the
+ * busyness was sampled.
+ */
+static ktime_t guc_engine_busyness(struct intel_engine_cs *engine, ktime_t *now)
+{
+	struct intel_engine_guc_stats stats_saved, *stats = &engine->stats.guc;
+	struct i915_gpu_error *gpu_error = &engine->i915->gpu_error;
+	struct intel_gt *gt = engine->gt;
+	struct intel_guc *guc = &gt->uc.guc;
+	u64 total, gt_stamp_saved;
+	unsigned long flags;
+	u32 reset_count;
+	bool in_reset;
+
+	spin_lock_irqsave(&guc->timestamp.lock, flags);
+
+	/*
+	 * If a reset happened, we risk reading partially updated engine
+	 * busyness from GuC, so we just use the driver stored copy of busyness.
+	 * Synchronize with gt reset using reset_count and the
+	 * I915_RESET_BACKOFF flag. Note that reset flow updates the reset_count
+	 * after I915_RESET_BACKOFF flag, so ensure that the reset_count is
+	 * usable by checking the flag afterwards.
+	 */
+	reset_count = i915_reset_count(gpu_error);
+	in_reset = test_bit(I915_RESET_BACKOFF, &gt->reset.flags);
+
+	*now = ktime_get();
+
+	/*
+	 * The active busyness depends on start_gt_clk and gt_stamp.
+	 * gt_stamp is updated by i915 only when gt is awake and the
+	 * start_gt_clk is derived from GuC state. To get a consistent
+	 * view of activity, we query the GuC state only if gt is awake.
+	 */
+	if (!in_reset && intel_gt_pm_get_if_awake(gt)) {
+		stats_saved = *stats;
+		gt_stamp_saved = guc->timestamp.gt_stamp;
+		/*
+		 * Update gt_clks, then gt timestamp to simplify the 'gt_stamp -
+		 * start_gt_clk' calculation below for active engines.
+		 */
+		guc_update_engine_gt_clks(engine);
+		guc_update_pm_timestamp(guc, now);
+		intel_gt_pm_put_async(gt);
+		if (i915_reset_count(gpu_error) != reset_count) {
+			*stats = stats_saved;
+			guc->timestamp.gt_stamp = gt_stamp_saved;
+		}
+	}
+
+	total = intel_gt_clock_interval_to_ns(gt, stats->total_gt_clks);
+	if (stats->running) {
+		u64 clk = guc->timestamp.gt_stamp - stats->start_gt_clk;
+
+		total += intel_gt_clock_interval_to_ns(gt, clk);
+	}
+
+	spin_unlock_irqrestore(&guc->timestamp.lock, flags);
+
+	return ns_to_ktime(total);
+}
+
+static void __reset_guc_busyness_stats(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	unsigned long flags;
+	ktime_t unused;
+
+	cancel_delayed_work_sync(&guc->timestamp.work);
+
+	spin_lock_irqsave(&guc->timestamp.lock, flags);
+
+	guc_update_pm_timestamp(guc, &unused);
+	for_each_engine(engine, gt, id) {
+		guc_update_engine_gt_clks(engine);
+		engine->stats.guc.prev_total = 0;
+	}
+
+	spin_unlock_irqrestore(&guc->timestamp.lock, flags);
+}
+
+static void __update_guc_busyness_stats(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	unsigned long flags;
+	ktime_t unused;
+
+	guc->timestamp.last_stat_jiffies = jiffies;
+
+	spin_lock_irqsave(&guc->timestamp.lock, flags);
+
+	guc_update_pm_timestamp(guc, &unused);
+	for_each_engine(engine, gt, id)
+		guc_update_engine_gt_clks(engine);
+
+	spin_unlock_irqrestore(&guc->timestamp.lock, flags);
+}
+
+static void guc_timestamp_ping(struct work_struct *wrk)
+{
+	struct intel_guc *guc = container_of(wrk, typeof(*guc),
+					     timestamp.work.work);
+	struct intel_uc *uc = container_of(guc, typeof(*uc), guc);
+	struct intel_gt *gt = guc_to_gt(guc);
+	intel_wakeref_t wakeref;
+	int srcu, ret;
+
+	/*
+	 * Synchronize with gt reset to make sure the worker does not
+	 * corrupt the engine/guc stats.
+	 */
+	ret = intel_gt_reset_trylock(gt, &srcu);
+	if (ret)
+		return;
+
+	with_intel_runtime_pm(&gt->i915->runtime_pm, wakeref)
+		__update_guc_busyness_stats(guc);
+
+	intel_gt_reset_unlock(gt, srcu);
+
+	mod_delayed_work(system_highpri_wq, &guc->timestamp.work,
+			 guc->timestamp.ping_delay);
+}
+
+static int guc_action_enable_usage_stats(struct intel_guc *guc)
+{
+	u32 offset = intel_guc_engine_usage_offset(guc);
+	u32 action[] = {
+		INTEL_GUC_ACTION_SET_ENG_UTIL_BUFF,
+		offset,
+		0,
+	};
+
+	return intel_guc_send(guc, action, ARRAY_SIZE(action));
+}
+
+static void guc_init_engine_stats(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	intel_wakeref_t wakeref;
+
+	mod_delayed_work(system_highpri_wq, &guc->timestamp.work,
+			 guc->timestamp.ping_delay);
+
+	with_intel_runtime_pm(&gt->i915->runtime_pm, wakeref) {
+		int ret = guc_action_enable_usage_stats(guc);
+
+		if (ret)
+			drm_err(&gt->i915->drm,
+				"Failed to enable usage stats: %d!\n", ret);
+	}
+}
+
+void intel_guc_busyness_park(struct intel_gt *gt)
+{
+	struct intel_guc *guc = &gt->uc.guc;
+
+	if (!guc_submission_initialized(guc))
+		return;
+
+	/*
+	 * There is a race with suspend flow where the worker runs after suspend
+	 * and causes an unclaimed register access warning. Cancel the worker
+	 * synchronously here.
+	 */
+	cancel_delayed_work_sync(&guc->timestamp.work);
+
+	/*
+	 * Before parking, we should sample engine busyness stats if we need to.
+	 * We can skip it if we are less than half a ping from the last time we
+	 * sampled the busyness stats.
+	 */
+	if (guc->timestamp.last_stat_jiffies &&
+	    !time_after(jiffies, guc->timestamp.last_stat_jiffies +
+			(guc->timestamp.ping_delay / 2)))
+		return;
+
+	__update_guc_busyness_stats(guc);
+}
+
+void intel_guc_busyness_unpark(struct intel_gt *gt)
+{
+	struct intel_guc *guc = &gt->uc.guc;
+	unsigned long flags;
+	ktime_t unused;
+
+	if (!guc_submission_initialized(guc))
+		return;
+
+	spin_lock_irqsave(&guc->timestamp.lock, flags);
+	guc_update_pm_timestamp(guc, &unused);
+	spin_unlock_irqrestore(&guc->timestamp.lock, flags);
+	mod_delayed_work(system_highpri_wq, &guc->timestamp.work,
+			 guc->timestamp.ping_delay);
+}
+
+static inline bool
+submission_disabled(struct intel_guc *guc)
+{
+	struct i915_sched_engine * const sched_engine = guc->sched_engine;
+
+	return unlikely(!sched_engine ||
+			!__tasklet_is_enabled(&sched_engine->tasklet) ||
+			intel_gt_is_wedged(guc_to_gt(guc)));
+}
+
+static void disable_submission(struct intel_guc *guc)
+{
+	struct i915_sched_engine * const sched_engine = guc->sched_engine;
+
+	if (__tasklet_is_enabled(&sched_engine->tasklet)) {
+		GEM_BUG_ON(!guc->ct.enabled);
+		__tasklet_disable_sync_once(&sched_engine->tasklet);
+		sched_engine->tasklet.callback = NULL;
+	}
+}
+
+static void enable_submission(struct intel_guc *guc)
+{
+	struct i915_sched_engine * const sched_engine = guc->sched_engine;
+	unsigned long flags;
+
+	spin_lock_irqsave(&guc->sched_engine->lock, flags);
+	sched_engine->tasklet.callback = guc_submission_tasklet;
+	wmb();	/* Make sure callback visible */
+	if (!__tasklet_is_enabled(&sched_engine->tasklet) &&
+	    __tasklet_enable(&sched_engine->tasklet)) {
+		GEM_BUG_ON(!guc->ct.enabled);
+
+		/* And kick in case we missed a new request submission. */
+		tasklet_hi_schedule(&sched_engine->tasklet);
+	}
+	spin_unlock_irqrestore(&guc->sched_engine->lock, flags);
+}
+
+static void guc_flush_submissions(struct intel_guc *guc)
+{
+	struct i915_sched_engine * const sched_engine = guc->sched_engine;
+	unsigned long flags;
+
+	spin_lock_irqsave(&sched_engine->lock, flags);
+	spin_unlock_irqrestore(&sched_engine->lock, flags);
+}
+
+static void guc_flush_destroyed_contexts(struct intel_guc *guc);
+
+void intel_guc_submission_reset_prepare(struct intel_guc *guc)
+{
+	if (unlikely(!guc_submission_initialized(guc))) {
+		/* Reset called during driver load? GuC not yet initialised! */
+		return;
+	}
+
+	intel_gt_park_heartbeats(guc_to_gt(guc));
+	disable_submission(guc);
+	guc->interrupts.disable(guc);
+	__reset_guc_busyness_stats(guc);
+
+	/* Flush IRQ handler */
+	spin_lock_irq(guc_to_gt(guc)->irq_lock);
+	spin_unlock_irq(guc_to_gt(guc)->irq_lock);
+
+	guc_flush_submissions(guc);
+	guc_flush_destroyed_contexts(guc);
+	flush_work(&guc->ct.requests.worker);
+
+	scrub_guc_desc_for_outstanding_g2h(guc);
+}
+
+static struct intel_engine_cs *
+guc_virtual_get_sibling(struct intel_engine_cs *ve, unsigned int sibling)
+{
+	struct intel_engine_cs *engine;
+	intel_engine_mask_t tmp, mask = ve->mask;
+	unsigned int num_siblings = 0;
+
+	for_each_engine_masked(engine, ve->gt, mask, tmp)
+		if (num_siblings++ == sibling)
+			return engine;
+
+	return NULL;
+}
+
+static inline struct intel_engine_cs *
+__context_to_physical_engine(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = ce->engine;
+
+	if (intel_engine_is_virtual(engine))
+		engine = guc_virtual_get_sibling(engine, 0);
+
+	return engine;
+}
+
+static void guc_reset_state(struct intel_context *ce, u32 head, bool scrub)
+{
+	struct intel_engine_cs *engine = __context_to_physical_engine(ce);
+
+	if (!intel_context_is_schedulable(ce))
+		return;
+
+	GEM_BUG_ON(!intel_context_is_pinned(ce));
+
+	/*
+	 * We want a simple context + ring to execute the breadcrumb update.
+	 * We cannot rely on the context being intact across the GPU hang,
+	 * so clear it and rebuild just what we need for the breadcrumb.
+	 * All pending requests for this context will be zapped, and any
+	 * future request will be after userspace has had the opportunity
+	 * to recreate its own state.
+	 */
+	if (scrub)
+		lrc_init_regs(ce, engine, true);
+
+	/* Rerun the request; its payload has been neutered (if guilty). */
+	lrc_update_regs(ce, engine, head);
+}
+
+static void guc_engine_reset_prepare(struct intel_engine_cs *engine)
+{
+	if (!IS_GRAPHICS_VER(engine->i915, 11, 12))
+		return;
+
+	intel_engine_stop_cs(engine);
+
+	/*
+	 * Wa_22011802037:gen11/gen12: In addition to stopping the cs, we need
+	 * to wait for any pending mi force wakeups
+	 */
+	intel_engine_wait_for_pending_mi_fw(engine);
+}
+
+static void guc_reset_nop(struct intel_engine_cs *engine)
+{
+}
+
+static void guc_rewind_nop(struct intel_engine_cs *engine, bool stalled)
+{
+}
+
+static void
+__unwind_incomplete_requests(struct intel_context *ce)
+{
+	struct i915_request *rq, *rn;
+	struct list_head *pl;
+	int prio = I915_PRIORITY_INVALID;
+	struct i915_sched_engine * const sched_engine =
+		ce->engine->sched_engine;
+	unsigned long flags;
+
+	spin_lock_irqsave(&sched_engine->lock, flags);
+	spin_lock(&ce->guc_state.lock);
+	list_for_each_entry_safe_reverse(rq, rn,
+					 &ce->guc_state.requests,
+					 sched.link) {
+		if (i915_request_completed(rq))
+			continue;
+
+		list_del_init(&rq->sched.link);
+		__i915_request_unsubmit(rq);
+
+		/* Push the request back into the queue for later resubmission. */
+		GEM_BUG_ON(rq_prio(rq) == I915_PRIORITY_INVALID);
+		if (rq_prio(rq) != prio) {
+			prio = rq_prio(rq);
+			pl = i915_sched_lookup_priolist(sched_engine, prio);
+		}
+		GEM_BUG_ON(i915_sched_engine_is_empty(sched_engine));
+
+		list_add(&rq->sched.link, pl);
+		set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+	}
+	spin_unlock(&ce->guc_state.lock);
+	spin_unlock_irqrestore(&sched_engine->lock, flags);
+}
+
+static void __guc_reset_context(struct intel_context *ce, intel_engine_mask_t stalled)
+{
+	bool guilty;
+	struct i915_request *rq;
+	unsigned long flags;
+	u32 head;
+	int i, number_children = ce->parallel.number_children;
+	struct intel_context *parent = ce;
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	intel_context_get(ce);
+
+	/*
+	 * GuC will implicitly mark the context as non-schedulable when it sends
+	 * the reset notification. Make sure our state reflects this change. The
+	 * context will be marked enabled on resubmission.
+	 */
+	spin_lock_irqsave(&ce->guc_state.lock, flags);
+	clr_context_enabled(ce);
+	spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+
+	/*
+	 * For each context in the relationship find the hanging request
+	 * resetting each context / request as needed
+	 */
+	for (i = 0; i < number_children + 1; ++i) {
+		if (!intel_context_is_pinned(ce))
+			goto next_context;
+
+		guilty = false;
+		rq = intel_context_get_active_request(ce);
+		if (!rq) {
+			head = ce->ring->tail;
+			goto out_replay;
+		}
+
+		if (i915_request_started(rq))
+			guilty = stalled & ce->engine->mask;
+
+		GEM_BUG_ON(i915_active_is_idle(&ce->active));
+		head = intel_ring_wrap(ce->ring, rq->head);
+
+		__i915_request_reset(rq, guilty);
+		i915_request_put(rq);
+out_replay:
+		guc_reset_state(ce, head, guilty);
+next_context:
+		if (i != number_children)
+			ce = list_next_entry(ce, parallel.child_link);
+	}
+
+	__unwind_incomplete_requests(parent);
+	intel_context_put(parent);
+}
+
+void intel_guc_submission_reset(struct intel_guc *guc, intel_engine_mask_t stalled)
+{
+	struct intel_context *ce;
+	unsigned long index;
+	unsigned long flags;
+
+	if (unlikely(!guc_submission_initialized(guc))) {
+		/* Reset called during driver load? GuC not yet initialised! */
+		return;
+	}
+
+	xa_lock_irqsave(&guc->context_lookup, flags);
+	xa_for_each(&guc->context_lookup, index, ce) {
+		if (!kref_get_unless_zero(&ce->ref))
+			continue;
+
+		xa_unlock(&guc->context_lookup);
+
+		if (intel_context_is_pinned(ce) &&
+		    !intel_context_is_child(ce))
+			__guc_reset_context(ce, stalled);
+
+		intel_context_put(ce);
+
+		xa_lock(&guc->context_lookup);
+	}
+	xa_unlock_irqrestore(&guc->context_lookup, flags);
+
+	/* GuC is blown away, drop all references to contexts */
+	xa_destroy(&guc->context_lookup);
+}
+
+static void guc_cancel_context_requests(struct intel_context *ce)
+{
+	struct i915_sched_engine *sched_engine = ce_to_guc(ce)->sched_engine;
+	struct i915_request *rq;
+	unsigned long flags;
+
+	/* Mark all executing requests as skipped. */
+	spin_lock_irqsave(&sched_engine->lock, flags);
+	spin_lock(&ce->guc_state.lock);
+	list_for_each_entry(rq, &ce->guc_state.requests, sched.link)
+		i915_request_put(i915_request_mark_eio(rq));
+	spin_unlock(&ce->guc_state.lock);
+	spin_unlock_irqrestore(&sched_engine->lock, flags);
+}
+
+static void
+guc_cancel_sched_engine_requests(struct i915_sched_engine *sched_engine)
+{
+	struct i915_request *rq, *rn;
+	struct rb_node *rb;
+	unsigned long flags;
+
+	/* Can be called during boot if GuC fails to load */
+	if (!sched_engine)
+		return;
+
+	/*
+	 * Before we call engine->cancel_requests(), we should have exclusive
+	 * access to the submission state. This is arranged for us by the
+	 * caller disabling the interrupt generation, the tasklet and other
+	 * threads that may then access the same state, giving us a free hand
+	 * to reset state. However, we still need to let lockdep be aware that
+	 * we know this state may be accessed in hardirq context, so we
+	 * disable the irq around this manipulation and we want to keep
+	 * the spinlock focused on its duties and not accidentally conflate
+	 * coverage to the submission's irq state. (Similarly, although we
+	 * shouldn't need to disable irq around the manipulation of the
+	 * submission's irq state, we also wish to remind ourselves that
+	 * it is irq state.)
+	 */
+	spin_lock_irqsave(&sched_engine->lock, flags);
+
+	/* Flush the queued requests to the timeline list (for retiring). */
+	while ((rb = rb_first_cached(&sched_engine->queue))) {
+		struct i915_priolist *p = to_priolist(rb);
+
+		priolist_for_each_request_consume(rq, rn, p) {
+			list_del_init(&rq->sched.link);
+
+			__i915_request_submit(rq);
+
+			i915_request_put(i915_request_mark_eio(rq));
+		}
+
+		rb_erase_cached(&p->node, &sched_engine->queue);
+		i915_priolist_free(p);
+	}
+
+	/* Remaining _unready_ requests will be nop'ed when submitted */
+
+	sched_engine->queue_priority_hint = INT_MIN;
+	sched_engine->queue = RB_ROOT_CACHED;
+
+	spin_unlock_irqrestore(&sched_engine->lock, flags);
+}
+
+void intel_guc_submission_cancel_requests(struct intel_guc *guc)
+{
+	struct intel_context *ce;
+	unsigned long index;
+	unsigned long flags;
+
+	xa_lock_irqsave(&guc->context_lookup, flags);
+	xa_for_each(&guc->context_lookup, index, ce) {
+		if (!kref_get_unless_zero(&ce->ref))
+			continue;
+
+		xa_unlock(&guc->context_lookup);
+
+		if (intel_context_is_pinned(ce) &&
+		    !intel_context_is_child(ce))
+			guc_cancel_context_requests(ce);
+
+		intel_context_put(ce);
+
+		xa_lock(&guc->context_lookup);
+	}
+	xa_unlock_irqrestore(&guc->context_lookup, flags);
+
+	guc_cancel_sched_engine_requests(guc->sched_engine);
+
+	/* GuC is blown away, drop all references to contexts */
+	xa_destroy(&guc->context_lookup);
+}
+
+void intel_guc_submission_reset_finish(struct intel_guc *guc)
+{
+	/* Reset called during driver load or during wedge? */
+	if (unlikely(!guc_submission_initialized(guc) ||
+		     intel_gt_is_wedged(guc_to_gt(guc)))) {
+		return;
+	}
+
+	/*
+	 * Technically possible for either of these values to be non-zero here,
+	 * but very unlikely + harmless. Regardless let's add a warn so we can
+	 * see in CI if this happens frequently / a precursor to taking down the
+	 * machine.
+	 */
+	GEM_WARN_ON(atomic_read(&guc->outstanding_submission_g2h));
+	atomic_set(&guc->outstanding_submission_g2h, 0);
+
+	intel_guc_global_policies_update(guc);
+	enable_submission(guc);
+	intel_gt_unpark_heartbeats(guc_to_gt(guc));
+}
+
+static void destroyed_worker_func(struct work_struct *w);
+static void reset_fail_worker_func(struct work_struct *w);
+
+/*
+ * Set up the memory resources to be shared with the GuC (via the GGTT)
+ * at firmware loading time.
+ */
+int intel_guc_submission_init(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	int ret;
+
+	if (guc->submission_initialized)
+		return 0;
+
+	if (GET_UC_VER(guc) < MAKE_UC_VER(70, 0, 0)) {
+		ret = guc_lrc_desc_pool_create_v69(guc);
+		if (ret)
+			return ret;
+	}
+
+	guc->submission_state.guc_ids_bitmap =
+		bitmap_zalloc(NUMBER_MULTI_LRC_GUC_ID(guc), GFP_KERNEL);
+	if (!guc->submission_state.guc_ids_bitmap) {
+		ret = -ENOMEM;
+		goto destroy_pool;
+	}
+
+	guc->timestamp.ping_delay = (POLL_TIME_CLKS / gt->clock_frequency + 1) * HZ;
+	guc->timestamp.shift = gpm_timestamp_shift(gt);
+	guc->submission_initialized = true;
+
+	return 0;
+
+destroy_pool:
+	guc_lrc_desc_pool_destroy_v69(guc);
+
+	return ret;
+}
+
+void intel_guc_submission_fini(struct intel_guc *guc)
+{
+	if (!guc->submission_initialized)
+		return;
+
+	guc_flush_destroyed_contexts(guc);
+	guc_lrc_desc_pool_destroy_v69(guc);
+	i915_sched_engine_put(guc->sched_engine);
+	bitmap_free(guc->submission_state.guc_ids_bitmap);
+	guc->submission_initialized = false;
+}
+
+static inline void queue_request(struct i915_sched_engine *sched_engine,
+				 struct i915_request *rq,
+				 int prio)
+{
+	GEM_BUG_ON(!list_empty(&rq->sched.link));
+	list_add_tail(&rq->sched.link,
+		      i915_sched_lookup_priolist(sched_engine, prio));
+	set_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+	tasklet_hi_schedule(&sched_engine->tasklet);
+}
+
+static int guc_bypass_tasklet_submit(struct intel_guc *guc,
+				     struct i915_request *rq)
+{
+	int ret = 0;
+
+	__i915_request_submit(rq);
+
+	trace_i915_request_in(rq, 0);
+
+	if (is_multi_lrc_rq(rq)) {
+		if (multi_lrc_submit(rq)) {
+			ret = guc_wq_item_append(guc, rq);
+			if (!ret)
+				ret = guc_add_request(guc, rq);
+		}
+	} else {
+		guc_set_lrc_tail(rq);
+		ret = guc_add_request(guc, rq);
+	}
+
+	if (unlikely(ret == -EPIPE))
+		disable_submission(guc);
+
+	return ret;
+}
+
+static bool need_tasklet(struct intel_guc *guc, struct i915_request *rq)
+{
+	struct i915_sched_engine *sched_engine = rq->engine->sched_engine;
+	struct intel_context *ce = request_to_scheduling_context(rq);
+
+	return submission_disabled(guc) || guc->stalled_request ||
+		!i915_sched_engine_is_empty(sched_engine) ||
+		!ctx_id_mapped(guc, ce->guc_id.id);
+}
+
+static void guc_submit_request(struct i915_request *rq)
+{
+	struct i915_sched_engine *sched_engine = rq->engine->sched_engine;
+	struct intel_guc *guc = &rq->engine->gt->uc.guc;
+	unsigned long flags;
+
+	/* Will be called from irq-context when using foreign fences. */
+	spin_lock_irqsave(&sched_engine->lock, flags);
+
+	if (need_tasklet(guc, rq))
+		queue_request(sched_engine, rq, rq_prio(rq));
+	else if (guc_bypass_tasklet_submit(guc, rq) == -EBUSY)
+		tasklet_hi_schedule(&sched_engine->tasklet);
+
+	spin_unlock_irqrestore(&sched_engine->lock, flags);
+}
+
+static int new_guc_id(struct intel_guc *guc, struct intel_context *ce)
+{
+	int ret;
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	if (intel_context_is_parent(ce))
+		ret = bitmap_find_free_region(guc->submission_state.guc_ids_bitmap,
+					      NUMBER_MULTI_LRC_GUC_ID(guc),
+					      order_base_2(ce->parallel.number_children
+							   + 1));
+	else
+		ret = ida_simple_get(&guc->submission_state.guc_ids,
+				     NUMBER_MULTI_LRC_GUC_ID(guc),
+				     guc->submission_state.num_guc_ids,
+				     GFP_KERNEL | __GFP_RETRY_MAYFAIL |
+				     __GFP_NOWARN);
+	if (unlikely(ret < 0))
+		return ret;
+
+	ce->guc_id.id = ret;
+	return 0;
+}
+
+static void __release_guc_id(struct intel_guc *guc, struct intel_context *ce)
+{
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	if (!context_guc_id_invalid(ce)) {
+		if (intel_context_is_parent(ce))
+			bitmap_release_region(guc->submission_state.guc_ids_bitmap,
+					      ce->guc_id.id,
+					      order_base_2(ce->parallel.number_children
+							   + 1));
+		else
+			ida_simple_remove(&guc->submission_state.guc_ids,
+					  ce->guc_id.id);
+		clr_ctx_id_mapping(guc, ce->guc_id.id);
+		set_context_guc_id_invalid(ce);
+	}
+	if (!list_empty(&ce->guc_id.link))
+		list_del_init(&ce->guc_id.link);
+}
+
+static void release_guc_id(struct intel_guc *guc, struct intel_context *ce)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&guc->submission_state.lock, flags);
+	__release_guc_id(guc, ce);
+	spin_unlock_irqrestore(&guc->submission_state.lock, flags);
+}
+
+static int steal_guc_id(struct intel_guc *guc, struct intel_context *ce)
+{
+	struct intel_context *cn;
+
+	lockdep_assert_held(&guc->submission_state.lock);
+	GEM_BUG_ON(intel_context_is_child(ce));
+	GEM_BUG_ON(intel_context_is_parent(ce));
+
+	if (!list_empty(&guc->submission_state.guc_id_list)) {
+		cn = list_first_entry(&guc->submission_state.guc_id_list,
+				      struct intel_context,
+				      guc_id.link);
+
+		GEM_BUG_ON(atomic_read(&cn->guc_id.ref));
+		GEM_BUG_ON(context_guc_id_invalid(cn));
+		GEM_BUG_ON(intel_context_is_child(cn));
+		GEM_BUG_ON(intel_context_is_parent(cn));
+
+		list_del_init(&cn->guc_id.link);
+		ce->guc_id.id = cn->guc_id.id;
+
+		spin_lock(&cn->guc_state.lock);
+		clr_context_registered(cn);
+		spin_unlock(&cn->guc_state.lock);
+
+		set_context_guc_id_invalid(cn);
+
+#ifdef CONFIG_DRM_I915_SELFTEST
+		guc->number_guc_id_stolen++;
+#endif
+
+		return 0;
+	} else {
+		return -EAGAIN;
+	}
+}
+
+static int assign_guc_id(struct intel_guc *guc, struct intel_context *ce)
+{
+	int ret;
+
+	lockdep_assert_held(&guc->submission_state.lock);
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	ret = new_guc_id(guc, ce);
+	if (unlikely(ret < 0)) {
+		if (intel_context_is_parent(ce))
+			return -ENOSPC;
+
+		ret = steal_guc_id(guc, ce);
+		if (ret < 0)
+			return ret;
+	}
+
+	if (intel_context_is_parent(ce)) {
+		struct intel_context *child;
+		int i = 1;
+
+		for_each_child(ce, child)
+			child->guc_id.id = ce->guc_id.id + i++;
+	}
+
+	return 0;
+}
+
+#define PIN_GUC_ID_TRIES	4
+static int pin_guc_id(struct intel_guc *guc, struct intel_context *ce)
+{
+	int ret = 0;
+	unsigned long flags, tries = PIN_GUC_ID_TRIES;
+
+	GEM_BUG_ON(atomic_read(&ce->guc_id.ref));
+
+try_again:
+	spin_lock_irqsave(&guc->submission_state.lock, flags);
+
+	might_lock(&ce->guc_state.lock);
+
+	if (context_guc_id_invalid(ce)) {
+		ret = assign_guc_id(guc, ce);
+		if (ret)
+			goto out_unlock;
+		ret = 1;	/* Indidcates newly assigned guc_id */
+	}
+	if (!list_empty(&ce->guc_id.link))
+		list_del_init(&ce->guc_id.link);
+	atomic_inc(&ce->guc_id.ref);
+
+out_unlock:
+	spin_unlock_irqrestore(&guc->submission_state.lock, flags);
+
+	/*
+	 * -EAGAIN indicates no guc_id are available, let's retire any
+	 * outstanding requests to see if that frees up a guc_id. If the first
+	 * retire didn't help, insert a sleep with the timeslice duration before
+	 * attempting to retire more requests. Double the sleep period each
+	 * subsequent pass before finally giving up. The sleep period has max of
+	 * 100ms and minimum of 1ms.
+	 */
+	if (ret == -EAGAIN && --tries) {
+		if (PIN_GUC_ID_TRIES - tries > 1) {
+			unsigned int timeslice_shifted =
+				ce->engine->props.timeslice_duration_ms <<
+				(PIN_GUC_ID_TRIES - tries - 2);
+			unsigned int max = min_t(unsigned int, 100,
+						 timeslice_shifted);
+
+			msleep(max_t(unsigned int, max, 1));
+		}
+		intel_gt_retire_requests(guc_to_gt(guc));
+		goto try_again;
+	}
+
+	return ret;
+}
+
+static void unpin_guc_id(struct intel_guc *guc, struct intel_context *ce)
+{
+	unsigned long flags;
+
+	GEM_BUG_ON(atomic_read(&ce->guc_id.ref) < 0);
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	if (unlikely(context_guc_id_invalid(ce) ||
+		     intel_context_is_parent(ce)))
+		return;
+
+	spin_lock_irqsave(&guc->submission_state.lock, flags);
+	if (!context_guc_id_invalid(ce) && list_empty(&ce->guc_id.link) &&
+	    !atomic_read(&ce->guc_id.ref))
+		list_add_tail(&ce->guc_id.link,
+			      &guc->submission_state.guc_id_list);
+	spin_unlock_irqrestore(&guc->submission_state.lock, flags);
+}
+
+static int __guc_action_register_multi_lrc_v69(struct intel_guc *guc,
+					       struct intel_context *ce,
+					       u32 guc_id,
+					       u32 offset,
+					       bool loop)
+{
+	struct intel_context *child;
+	u32 action[4 + MAX_ENGINE_INSTANCE];
+	int len = 0;
+
+	GEM_BUG_ON(ce->parallel.number_children > MAX_ENGINE_INSTANCE);
+
+	action[len++] = INTEL_GUC_ACTION_REGISTER_CONTEXT_MULTI_LRC;
+	action[len++] = guc_id;
+	action[len++] = ce->parallel.number_children + 1;
+	action[len++] = offset;
+	for_each_child(ce, child) {
+		offset += sizeof(struct guc_lrc_desc_v69);
+		action[len++] = offset;
+	}
+
+	return guc_submission_send_busy_loop(guc, action, len, 0, loop);
+}
+
+static int __guc_action_register_multi_lrc_v70(struct intel_guc *guc,
+					       struct intel_context *ce,
+					       struct guc_ctxt_registration_info *info,
+					       bool loop)
+{
+	struct intel_context *child;
+	u32 action[13 + (MAX_ENGINE_INSTANCE * 2)];
+	int len = 0;
+	u32 next_id;
+
+	GEM_BUG_ON(ce->parallel.number_children > MAX_ENGINE_INSTANCE);
+
+	action[len++] = INTEL_GUC_ACTION_REGISTER_CONTEXT_MULTI_LRC;
+	action[len++] = info->flags;
+	action[len++] = info->context_idx;
+	action[len++] = info->engine_class;
+	action[len++] = info->engine_submit_mask;
+	action[len++] = info->wq_desc_lo;
+	action[len++] = info->wq_desc_hi;
+	action[len++] = info->wq_base_lo;
+	action[len++] = info->wq_base_hi;
+	action[len++] = info->wq_size;
+	action[len++] = ce->parallel.number_children + 1;
+	action[len++] = info->hwlrca_lo;
+	action[len++] = info->hwlrca_hi;
+
+	next_id = info->context_idx + 1;
+	for_each_child(ce, child) {
+		GEM_BUG_ON(next_id++ != child->guc_id.id);
+
+		/*
+		 * NB: GuC interface supports 64 bit LRCA even though i915/HW
+		 * only supports 32 bit currently.
+		 */
+		action[len++] = lower_32_bits(child->lrc.lrca);
+		action[len++] = upper_32_bits(child->lrc.lrca);
+	}
+
+	GEM_BUG_ON(len > ARRAY_SIZE(action));
+
+	return guc_submission_send_busy_loop(guc, action, len, 0, loop);
+}
+
+static int __guc_action_register_context_v69(struct intel_guc *guc,
+					     u32 guc_id,
+					     u32 offset,
+					     bool loop)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_REGISTER_CONTEXT,
+		guc_id,
+		offset,
+	};
+
+	return guc_submission_send_busy_loop(guc, action, ARRAY_SIZE(action),
+					     0, loop);
+}
+
+static int __guc_action_register_context_v70(struct intel_guc *guc,
+					     struct guc_ctxt_registration_info *info,
+					     bool loop)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_REGISTER_CONTEXT,
+		info->flags,
+		info->context_idx,
+		info->engine_class,
+		info->engine_submit_mask,
+		info->wq_desc_lo,
+		info->wq_desc_hi,
+		info->wq_base_lo,
+		info->wq_base_hi,
+		info->wq_size,
+		info->hwlrca_lo,
+		info->hwlrca_hi,
+	};
+
+	return guc_submission_send_busy_loop(guc, action, ARRAY_SIZE(action),
+					     0, loop);
+}
+
+static void prepare_context_registration_info_v69(struct intel_context *ce);
+static void prepare_context_registration_info_v70(struct intel_context *ce,
+						  struct guc_ctxt_registration_info *info);
+
+static int
+register_context_v69(struct intel_guc *guc, struct intel_context *ce, bool loop)
+{
+	u32 offset = intel_guc_ggtt_offset(guc, guc->lrc_desc_pool_v69) +
+		ce->guc_id.id * sizeof(struct guc_lrc_desc_v69);
+
+	prepare_context_registration_info_v69(ce);
+
+	if (intel_context_is_parent(ce))
+		return __guc_action_register_multi_lrc_v69(guc, ce, ce->guc_id.id,
+							   offset, loop);
+	else
+		return __guc_action_register_context_v69(guc, ce->guc_id.id,
+							 offset, loop);
+}
+
+static int
+register_context_v70(struct intel_guc *guc, struct intel_context *ce, bool loop)
+{
+	struct guc_ctxt_registration_info info;
+
+	prepare_context_registration_info_v70(ce, &info);
+
+	if (intel_context_is_parent(ce))
+		return __guc_action_register_multi_lrc_v70(guc, ce, &info, loop);
+	else
+		return __guc_action_register_context_v70(guc, &info, loop);
+}
+
+static int register_context(struct intel_context *ce, bool loop)
+{
+	struct intel_guc *guc = ce_to_guc(ce);
+	int ret;
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+	trace_intel_context_register(ce);
+
+	if (GET_UC_VER(guc) >= MAKE_UC_VER(70, 0, 0))
+		ret = register_context_v70(guc, ce, loop);
+	else
+		ret = register_context_v69(guc, ce, loop);
+
+	if (likely(!ret)) {
+		unsigned long flags;
+
+		spin_lock_irqsave(&ce->guc_state.lock, flags);
+		set_context_registered(ce);
+		spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+
+		if (GET_UC_VER(guc) >= MAKE_UC_VER(70, 0, 0))
+			guc_context_policy_init_v70(ce, loop);
+	}
+
+	return ret;
+}
+
+static int __guc_action_deregister_context(struct intel_guc *guc,
+					   u32 guc_id)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_DEREGISTER_CONTEXT,
+		guc_id,
+	};
+
+	return guc_submission_send_busy_loop(guc, action, ARRAY_SIZE(action),
+					     G2H_LEN_DW_DEREGISTER_CONTEXT,
+					     true);
+}
+
+static int deregister_context(struct intel_context *ce, u32 guc_id)
+{
+	struct intel_guc *guc = ce_to_guc(ce);
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+	trace_intel_context_deregister(ce);
+
+	return __guc_action_deregister_context(guc, guc_id);
+}
+
+static inline void clear_children_join_go_memory(struct intel_context *ce)
+{
+	struct parent_scratch *ps = __get_parent_scratch(ce);
+	int i;
+
+	ps->go.semaphore = 0;
+	for (i = 0; i < ce->parallel.number_children + 1; ++i)
+		ps->join[i].semaphore = 0;
+}
+
+static inline u32 get_children_go_value(struct intel_context *ce)
+{
+	return __get_parent_scratch(ce)->go.semaphore;
+}
+
+static inline u32 get_children_join_value(struct intel_context *ce,
+					  u8 child_index)
+{
+	return __get_parent_scratch(ce)->join[child_index].semaphore;
+}
+
+struct context_policy {
+	u32 count;
+	struct guc_update_context_policy h2g;
+};
+
+static u32 __guc_context_policy_action_size(struct context_policy *policy)
+{
+	size_t bytes = sizeof(policy->h2g.header) +
+		       (sizeof(policy->h2g.klv[0]) * policy->count);
+
+	return bytes / sizeof(u32);
+}
+
+static void __guc_context_policy_start_klv(struct context_policy *policy, u16 guc_id)
+{
+	policy->h2g.header.action = INTEL_GUC_ACTION_HOST2GUC_UPDATE_CONTEXT_POLICIES;
+	policy->h2g.header.ctx_id = guc_id;
+	policy->count = 0;
+}
+
+#define MAKE_CONTEXT_POLICY_ADD(func, id) \
+static void __guc_context_policy_add_##func(struct context_policy *policy, u32 data) \
+{ \
+	GEM_BUG_ON(policy->count >= GUC_CONTEXT_POLICIES_KLV_NUM_IDS); \
+	policy->h2g.klv[policy->count].kl = \
+		FIELD_PREP(GUC_KLV_0_KEY, GUC_CONTEXT_POLICIES_KLV_ID_##id) | \
+		FIELD_PREP(GUC_KLV_0_LEN, 1); \
+	policy->h2g.klv[policy->count].value = data; \
+	policy->count++; \
+}
+
+MAKE_CONTEXT_POLICY_ADD(execution_quantum, EXECUTION_QUANTUM)
+MAKE_CONTEXT_POLICY_ADD(preemption_timeout, PREEMPTION_TIMEOUT)
+MAKE_CONTEXT_POLICY_ADD(priority, SCHEDULING_PRIORITY)
+MAKE_CONTEXT_POLICY_ADD(preempt_to_idle, PREEMPT_TO_IDLE_ON_QUANTUM_EXPIRY)
+
+#undef MAKE_CONTEXT_POLICY_ADD
+
+static int __guc_context_set_context_policies(struct intel_guc *guc,
+					      struct context_policy *policy,
+					      bool loop)
+{
+	return guc_submission_send_busy_loop(guc, (u32 *)&policy->h2g,
+					__guc_context_policy_action_size(policy),
+					0, loop);
+}
+
+static int guc_context_policy_init_v70(struct intel_context *ce, bool loop)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct intel_guc *guc = &engine->gt->uc.guc;
+	struct context_policy policy;
+	u32 execution_quantum;
+	u32 preemption_timeout;
+	unsigned long flags;
+	int ret;
+
+	/* NB: For both of these, zero means disabled. */
+	GEM_BUG_ON(overflows_type(engine->props.timeslice_duration_ms * 1000,
+				  execution_quantum));
+	GEM_BUG_ON(overflows_type(engine->props.preempt_timeout_ms * 1000,
+				  preemption_timeout));
+	execution_quantum = engine->props.timeslice_duration_ms * 1000;
+	preemption_timeout = engine->props.preempt_timeout_ms * 1000;
+
+	__guc_context_policy_start_klv(&policy, ce->guc_id.id);
+
+	__guc_context_policy_add_priority(&policy, ce->guc_state.prio);
+	__guc_context_policy_add_execution_quantum(&policy, execution_quantum);
+	__guc_context_policy_add_preemption_timeout(&policy, preemption_timeout);
+
+	if (engine->flags & I915_ENGINE_WANT_FORCED_PREEMPTION)
+		__guc_context_policy_add_preempt_to_idle(&policy, 1);
+
+	ret = __guc_context_set_context_policies(guc, &policy, loop);
+
+	spin_lock_irqsave(&ce->guc_state.lock, flags);
+	if (ret != 0)
+		set_context_policy_required(ce);
+	else
+		clr_context_policy_required(ce);
+	spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+
+	return ret;
+}
+
+static void guc_context_policy_init_v69(struct intel_engine_cs *engine,
+					struct guc_lrc_desc_v69 *desc)
+{
+	desc->policy_flags = 0;
+
+	if (engine->flags & I915_ENGINE_WANT_FORCED_PREEMPTION)
+		desc->policy_flags |= CONTEXT_POLICY_FLAG_PREEMPT_TO_IDLE_V69;
+
+	/* NB: For both of these, zero means disabled. */
+	GEM_BUG_ON(overflows_type(engine->props.timeslice_duration_ms * 1000,
+				  desc->execution_quantum));
+	GEM_BUG_ON(overflows_type(engine->props.preempt_timeout_ms * 1000,
+				  desc->preemption_timeout));
+	desc->execution_quantum = engine->props.timeslice_duration_ms * 1000;
+	desc->preemption_timeout = engine->props.preempt_timeout_ms * 1000;
+}
+
+static u32 map_guc_prio_to_lrc_desc_prio(u8 prio)
+{
+	/*
+	 * this matches the mapping we do in map_i915_prio_to_guc_prio()
+	 * (e.g. prio < I915_PRIORITY_NORMAL maps to GUC_CLIENT_PRIORITY_NORMAL)
+	 */
+	switch (prio) {
+	default:
+		MISSING_CASE(prio);
+		fallthrough;
+	case GUC_CLIENT_PRIORITY_KMD_NORMAL:
+		return GEN12_CTX_PRIORITY_NORMAL;
+	case GUC_CLIENT_PRIORITY_NORMAL:
+		return GEN12_CTX_PRIORITY_LOW;
+	case GUC_CLIENT_PRIORITY_HIGH:
+	case GUC_CLIENT_PRIORITY_KMD_HIGH:
+		return GEN12_CTX_PRIORITY_HIGH;
+	}
+}
+
+static void prepare_context_registration_info_v69(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct intel_guc *guc = &engine->gt->uc.guc;
+	u32 ctx_id = ce->guc_id.id;
+	struct guc_lrc_desc_v69 *desc;
+	struct intel_context *child;
+
+	GEM_BUG_ON(!engine->mask);
+
+	/*
+	 * Ensure LRC + CT vmas are is same region as write barrier is done
+	 * based on CT vma region.
+	 */
+	GEM_BUG_ON(i915_gem_object_is_lmem(guc->ct.vma->obj) !=
+		   i915_gem_object_is_lmem(ce->ring->vma->obj));
+
+	desc = __get_lrc_desc_v69(guc, ctx_id);
+	desc->engine_class = engine_class_to_guc_class(engine->class);
+	desc->engine_submit_mask = engine->logical_mask;
+	desc->hw_context_desc = ce->lrc.lrca;
+	desc->priority = ce->guc_state.prio;
+	desc->context_flags = CONTEXT_REGISTRATION_FLAG_KMD;
+	guc_context_policy_init_v69(engine, desc);
+
+	/*
+	 * If context is a parent, we need to register a process descriptor
+	 * describing a work queue and register all child contexts.
+	 */
+	if (intel_context_is_parent(ce)) {
+		struct guc_process_desc_v69 *pdesc;
+
+		ce->parallel.guc.wqi_tail = 0;
+		ce->parallel.guc.wqi_head = 0;
+
+		desc->process_desc = i915_ggtt_offset(ce->state) +
+			__get_parent_scratch_offset(ce);
+		desc->wq_addr = i915_ggtt_offset(ce->state) +
+			__get_wq_offset(ce);
+		desc->wq_size = WQ_SIZE;
+
+		pdesc = __get_process_desc_v69(ce);
+		memset(pdesc, 0, sizeof(*(pdesc)));
+		pdesc->stage_id = ce->guc_id.id;
+		pdesc->wq_base_addr = desc->wq_addr;
+		pdesc->wq_size_bytes = desc->wq_size;
+		pdesc->wq_status = WQ_STATUS_ACTIVE;
+
+		ce->parallel.guc.wq_head = &pdesc->head;
+		ce->parallel.guc.wq_tail = &pdesc->tail;
+		ce->parallel.guc.wq_status = &pdesc->wq_status;
+
+		for_each_child(ce, child) {
+			desc = __get_lrc_desc_v69(guc, child->guc_id.id);
+
+			desc->engine_class =
+				engine_class_to_guc_class(engine->class);
+			desc->hw_context_desc = child->lrc.lrca;
+			desc->priority = ce->guc_state.prio;
+			desc->context_flags = CONTEXT_REGISTRATION_FLAG_KMD;
+			guc_context_policy_init_v69(engine, desc);
+		}
+
+		clear_children_join_go_memory(ce);
+	}
+}
+
+static void prepare_context_registration_info_v70(struct intel_context *ce,
+						  struct guc_ctxt_registration_info *info)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct intel_guc *guc = &engine->gt->uc.guc;
+	u32 ctx_id = ce->guc_id.id;
+
+	GEM_BUG_ON(!engine->mask);
+
+	/*
+	 * Ensure LRC + CT vmas are is same region as write barrier is done
+	 * based on CT vma region.
+	 */
+	GEM_BUG_ON(i915_gem_object_is_lmem(guc->ct.vma->obj) !=
+		   i915_gem_object_is_lmem(ce->ring->vma->obj));
+
+	memset(info, 0, sizeof(*info));
+	info->context_idx = ctx_id;
+	info->engine_class = engine_class_to_guc_class(engine->class);
+	info->engine_submit_mask = engine->logical_mask;
+	/*
+	 * NB: GuC interface supports 64 bit LRCA even though i915/HW
+	 * only supports 32 bit currently.
+	 */
+	info->hwlrca_lo = lower_32_bits(ce->lrc.lrca);
+	info->hwlrca_hi = upper_32_bits(ce->lrc.lrca);
+	if (engine->flags & I915_ENGINE_HAS_EU_PRIORITY)
+		info->hwlrca_lo |= map_guc_prio_to_lrc_desc_prio(ce->guc_state.prio);
+	info->flags = CONTEXT_REGISTRATION_FLAG_KMD;
+
+	/*
+	 * If context is a parent, we need to register a process descriptor
+	 * describing a work queue and register all child contexts.
+	 */
+	if (intel_context_is_parent(ce)) {
+		struct guc_sched_wq_desc *wq_desc;
+		u64 wq_desc_offset, wq_base_offset;
+
+		ce->parallel.guc.wqi_tail = 0;
+		ce->parallel.guc.wqi_head = 0;
+
+		wq_desc_offset = i915_ggtt_offset(ce->state) +
+				 __get_parent_scratch_offset(ce);
+		wq_base_offset = i915_ggtt_offset(ce->state) +
+				 __get_wq_offset(ce);
+		info->wq_desc_lo = lower_32_bits(wq_desc_offset);
+		info->wq_desc_hi = upper_32_bits(wq_desc_offset);
+		info->wq_base_lo = lower_32_bits(wq_base_offset);
+		info->wq_base_hi = upper_32_bits(wq_base_offset);
+		info->wq_size = WQ_SIZE;
+
+		wq_desc = __get_wq_desc_v70(ce);
+		memset(wq_desc, 0, sizeof(*wq_desc));
+		wq_desc->wq_status = WQ_STATUS_ACTIVE;
+
+		ce->parallel.guc.wq_head = &wq_desc->head;
+		ce->parallel.guc.wq_tail = &wq_desc->tail;
+		ce->parallel.guc.wq_status = &wq_desc->wq_status;
+
+		clear_children_join_go_memory(ce);
+	}
+}
+
+static int try_context_registration(struct intel_context *ce, bool loop)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct intel_runtime_pm *runtime_pm = engine->uncore->rpm;
+	struct intel_guc *guc = &engine->gt->uc.guc;
+	intel_wakeref_t wakeref;
+	u32 ctx_id = ce->guc_id.id;
+	bool context_registered;
+	int ret = 0;
+
+	GEM_BUG_ON(!sched_state_is_init(ce));
+
+	context_registered = ctx_id_mapped(guc, ctx_id);
+
+	clr_ctx_id_mapping(guc, ctx_id);
+	set_ctx_id_mapping(guc, ctx_id, ce);
+
+	/*
+	 * The context_lookup xarray is used to determine if the hardware
+	 * context is currently registered. There are two cases in which it
+	 * could be registered either the guc_id has been stolen from another
+	 * context or the lrc descriptor address of this context has changed. In
+	 * either case the context needs to be deregistered with the GuC before
+	 * registering this context.
+	 */
+	if (context_registered) {
+		bool disabled;
+		unsigned long flags;
+
+		trace_intel_context_steal_guc_id(ce);
+		GEM_BUG_ON(!loop);
+
+		/* Seal race with Reset */
+		spin_lock_irqsave(&ce->guc_state.lock, flags);
+		disabled = submission_disabled(guc);
+		if (likely(!disabled)) {
+			set_context_wait_for_deregister_to_register(ce);
+			intel_context_get(ce);
+		}
+		spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+		if (unlikely(disabled)) {
+			clr_ctx_id_mapping(guc, ctx_id);
+			return 0;	/* Will get registered later */
+		}
+
+		/*
+		 * If stealing the guc_id, this ce has the same guc_id as the
+		 * context whose guc_id was stolen.
+		 */
+		with_intel_runtime_pm(runtime_pm, wakeref)
+			ret = deregister_context(ce, ce->guc_id.id);
+		if (unlikely(ret == -ENODEV))
+			ret = 0;	/* Will get registered later */
+	} else {
+		with_intel_runtime_pm(runtime_pm, wakeref)
+			ret = register_context(ce, loop);
+		if (unlikely(ret == -EBUSY)) {
+			clr_ctx_id_mapping(guc, ctx_id);
+		} else if (unlikely(ret == -ENODEV)) {
+			clr_ctx_id_mapping(guc, ctx_id);
+			ret = 0;	/* Will get registered later */
+		}
+	}
+
+	return ret;
+}
+
+static int __guc_context_pre_pin(struct intel_context *ce,
+				 struct intel_engine_cs *engine,
+				 struct i915_gem_ww_ctx *ww,
+				 void **vaddr)
+{
+	return lrc_pre_pin(ce, engine, ww, vaddr);
+}
+
+static int __guc_context_pin(struct intel_context *ce,
+			     struct intel_engine_cs *engine,
+			     void *vaddr)
+{
+	if (i915_ggtt_offset(ce->state) !=
+	    (ce->lrc.lrca & CTX_GTT_ADDRESS_MASK))
+		set_bit(CONTEXT_LRCA_DIRTY, &ce->flags);
+
+	/*
+	 * GuC context gets pinned in guc_request_alloc. See that function for
+	 * explaination of why.
+	 */
+
+	return lrc_pin(ce, engine, vaddr);
+}
+
+static int guc_context_pre_pin(struct intel_context *ce,
+			       struct i915_gem_ww_ctx *ww,
+			       void **vaddr)
+{
+	return __guc_context_pre_pin(ce, ce->engine, ww, vaddr);
+}
+
+static int guc_context_pin(struct intel_context *ce, void *vaddr)
+{
+	int ret = __guc_context_pin(ce, ce->engine, vaddr);
+
+	if (likely(!ret && !intel_context_is_barrier(ce)))
+		intel_engine_pm_get(ce->engine);
+
+	return ret;
+}
+
+static void guc_context_unpin(struct intel_context *ce)
+{
+	struct intel_guc *guc = ce_to_guc(ce);
+
+	unpin_guc_id(guc, ce);
+	lrc_unpin(ce);
+
+	if (likely(!intel_context_is_barrier(ce)))
+		intel_engine_pm_put_async(ce->engine);
+}
+
+static void guc_context_post_unpin(struct intel_context *ce)
+{
+	lrc_post_unpin(ce);
+}
+
+static void __guc_context_sched_enable(struct intel_guc *guc,
+				       struct intel_context *ce)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_SCHED_CONTEXT_MODE_SET,
+		ce->guc_id.id,
+		GUC_CONTEXT_ENABLE
+	};
+
+	trace_intel_context_sched_enable(ce);
+
+	guc_submission_send_busy_loop(guc, action, ARRAY_SIZE(action),
+				      G2H_LEN_DW_SCHED_CONTEXT_MODE_SET, true);
+}
+
+static void __guc_context_sched_disable(struct intel_guc *guc,
+					struct intel_context *ce,
+					u16 guc_id)
+{
+	u32 action[] = {
+		INTEL_GUC_ACTION_SCHED_CONTEXT_MODE_SET,
+		guc_id,	/* ce->guc_id.id not stable */
+		GUC_CONTEXT_DISABLE
+	};
+
+	GEM_BUG_ON(guc_id == GUC_INVALID_CONTEXT_ID);
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+	trace_intel_context_sched_disable(ce);
+
+	guc_submission_send_busy_loop(guc, action, ARRAY_SIZE(action),
+				      G2H_LEN_DW_SCHED_CONTEXT_MODE_SET, true);
+}
+
+static void guc_blocked_fence_complete(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+
+	if (!i915_sw_fence_done(&ce->guc_state.blocked))
+		i915_sw_fence_complete(&ce->guc_state.blocked);
+}
+
+static void guc_blocked_fence_reinit(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	GEM_BUG_ON(!i915_sw_fence_done(&ce->guc_state.blocked));
+
+	/*
+	 * This fence is always complete unless a pending schedule disable is
+	 * outstanding. We arm the fence here and complete it when we receive
+	 * the pending schedule disable complete message.
+	 */
+	i915_sw_fence_fini(&ce->guc_state.blocked);
+	i915_sw_fence_reinit(&ce->guc_state.blocked);
+	i915_sw_fence_await(&ce->guc_state.blocked);
+	i915_sw_fence_commit(&ce->guc_state.blocked);
+}
+
+static u16 prep_context_pending_disable(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+
+	set_context_pending_disable(ce);
+	clr_context_enabled(ce);
+	guc_blocked_fence_reinit(ce);
+	intel_context_get(ce);
+
+	return ce->guc_id.id;
+}
+
+static struct i915_sw_fence *guc_context_block(struct intel_context *ce)
+{
+	struct intel_guc *guc = ce_to_guc(ce);
+	unsigned long flags;
+	struct intel_runtime_pm *runtime_pm = ce->engine->uncore->rpm;
+	intel_wakeref_t wakeref;
+	u16 guc_id;
+	bool enabled;
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	spin_lock_irqsave(&ce->guc_state.lock, flags);
+
+	incr_context_blocked(ce);
+
+	enabled = context_enabled(ce);
+	if (unlikely(!enabled || submission_disabled(guc))) {
+		if (enabled)
+			clr_context_enabled(ce);
+		spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+		return &ce->guc_state.blocked;
+	}
+
+	/*
+	 * We add +2 here as the schedule disable complete CTB handler calls
+	 * intel_context_sched_disable_unpin (-2 to pin_count).
+	 */
+	atomic_add(2, &ce->pin_count);
+
+	guc_id = prep_context_pending_disable(ce);
+
+	spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+
+	with_intel_runtime_pm(runtime_pm, wakeref)
+		__guc_context_sched_disable(guc, ce, guc_id);
+
+	return &ce->guc_state.blocked;
+}
+
+#define SCHED_STATE_MULTI_BLOCKED_MASK \
+	(SCHED_STATE_BLOCKED_MASK & ~SCHED_STATE_BLOCKED)
+#define SCHED_STATE_NO_UNBLOCK \
+	(SCHED_STATE_MULTI_BLOCKED_MASK | \
+	 SCHED_STATE_PENDING_DISABLE | \
+	 SCHED_STATE_BANNED)
+
+static bool context_cant_unblock(struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+
+	return (ce->guc_state.sched_state & SCHED_STATE_NO_UNBLOCK) ||
+		context_guc_id_invalid(ce) ||
+		!ctx_id_mapped(ce_to_guc(ce), ce->guc_id.id) ||
+		!intel_context_is_pinned(ce);
+}
+
+static void guc_context_unblock(struct intel_context *ce)
+{
+	struct intel_guc *guc = ce_to_guc(ce);
+	unsigned long flags;
+	struct intel_runtime_pm *runtime_pm = ce->engine->uncore->rpm;
+	intel_wakeref_t wakeref;
+	bool enable;
+
+	GEM_BUG_ON(context_enabled(ce));
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	spin_lock_irqsave(&ce->guc_state.lock, flags);
+
+	if (unlikely(submission_disabled(guc) ||
+		     context_cant_unblock(ce))) {
+		enable = false;
+	} else {
+		enable = true;
+		set_context_pending_enable(ce);
+		set_context_enabled(ce);
+		intel_context_get(ce);
+	}
+
+	decr_context_blocked(ce);
+
+	spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+
+	if (enable) {
+		with_intel_runtime_pm(runtime_pm, wakeref)
+			__guc_context_sched_enable(guc, ce);
+	}
+}
+
+static void guc_context_cancel_request(struct intel_context *ce,
+				       struct i915_request *rq)
+{
+	struct intel_context *block_context =
+		request_to_scheduling_context(rq);
+
+	if (i915_sw_fence_signaled(&rq->submit)) {
+		struct i915_sw_fence *fence;
+
+		intel_context_get(ce);
+		fence = guc_context_block(block_context);
+		i915_sw_fence_wait(fence);
+		if (!i915_request_completed(rq)) {
+			__i915_request_skip(rq);
+			guc_reset_state(ce, intel_ring_wrap(ce->ring, rq->head),
+					true);
+		}
+
+		guc_context_unblock(block_context);
+		intel_context_put(ce);
+	}
+}
+
+static void __guc_context_set_preemption_timeout(struct intel_guc *guc,
+						 u16 guc_id,
+						 u32 preemption_timeout)
+{
+	if (GET_UC_VER(guc) >= MAKE_UC_VER(70, 0, 0)) {
+		struct context_policy policy;
+
+		__guc_context_policy_start_klv(&policy, guc_id);
+		__guc_context_policy_add_preemption_timeout(&policy, preemption_timeout);
+		__guc_context_set_context_policies(guc, &policy, true);
+	} else {
+		u32 action[] = {
+			INTEL_GUC_ACTION_V69_SET_CONTEXT_PREEMPTION_TIMEOUT,
+			guc_id,
+			preemption_timeout
+		};
+
+		intel_guc_send_busy_loop(guc, action, ARRAY_SIZE(action), 0, true);
+	}
+}
+
+static void
+guc_context_revoke(struct intel_context *ce, struct i915_request *rq,
+		   unsigned int preempt_timeout_ms)
+{
+	struct intel_guc *guc = ce_to_guc(ce);
+	struct intel_runtime_pm *runtime_pm =
+		&ce->engine->gt->i915->runtime_pm;
+	intel_wakeref_t wakeref;
+	unsigned long flags;
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	guc_flush_submissions(guc);
+
+	spin_lock_irqsave(&ce->guc_state.lock, flags);
+	set_context_banned(ce);
+
+	if (submission_disabled(guc) ||
+	    (!context_enabled(ce) && !context_pending_disable(ce))) {
+		spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+
+		guc_cancel_context_requests(ce);
+		intel_engine_signal_breadcrumbs(ce->engine);
+	} else if (!context_pending_disable(ce)) {
+		u16 guc_id;
+
+		/*
+		 * We add +2 here as the schedule disable complete CTB handler
+		 * calls intel_context_sched_disable_unpin (-2 to pin_count).
+		 */
+		atomic_add(2, &ce->pin_count);
+
+		guc_id = prep_context_pending_disable(ce);
+		spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+
+		/*
+		 * In addition to disabling scheduling, set the preemption
+		 * timeout to the minimum value (1 us) so the banned context
+		 * gets kicked off the HW ASAP.
+		 */
+		with_intel_runtime_pm(runtime_pm, wakeref) {
+			__guc_context_set_preemption_timeout(guc, guc_id,
+							     preempt_timeout_ms);
+			__guc_context_sched_disable(guc, ce, guc_id);
+		}
+	} else {
+		if (!context_guc_id_invalid(ce))
+			with_intel_runtime_pm(runtime_pm, wakeref)
+				__guc_context_set_preemption_timeout(guc,
+								     ce->guc_id.id,
+								     preempt_timeout_ms);
+		spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+	}
+}
+
+static void guc_context_sched_disable(struct intel_context *ce)
+{
+	struct intel_guc *guc = ce_to_guc(ce);
+	unsigned long flags;
+	struct intel_runtime_pm *runtime_pm = &ce->engine->gt->i915->runtime_pm;
+	intel_wakeref_t wakeref;
+	u16 guc_id;
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	spin_lock_irqsave(&ce->guc_state.lock, flags);
+
+	/*
+	 * We have to check if the context has been disabled by another thread,
+	 * check if submssion has been disabled to seal a race with reset and
+	 * finally check if any more requests have been committed to the
+	 * context ensursing that a request doesn't slip through the
+	 * 'context_pending_disable' fence.
+	 */
+	if (unlikely(!context_enabled(ce) || submission_disabled(guc) ||
+		     context_has_committed_requests(ce))) {
+		clr_context_enabled(ce);
+		spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+		goto unpin;
+	}
+	guc_id = prep_context_pending_disable(ce);
+
+	spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+
+	with_intel_runtime_pm(runtime_pm, wakeref)
+		__guc_context_sched_disable(guc, ce, guc_id);
+
+	return;
+unpin:
+	intel_context_sched_disable_unpin(ce);
+}
+
+static inline void guc_lrc_desc_unpin(struct intel_context *ce)
+{
+	struct intel_guc *guc = ce_to_guc(ce);
+	struct intel_gt *gt = guc_to_gt(guc);
+	unsigned long flags;
+	bool disabled;
+
+	GEM_BUG_ON(!intel_gt_pm_is_awake(gt));
+	GEM_BUG_ON(!ctx_id_mapped(guc, ce->guc_id.id));
+	GEM_BUG_ON(ce != __get_context(guc, ce->guc_id.id));
+	GEM_BUG_ON(context_enabled(ce));
+
+	/* Seal race with Reset */
+	spin_lock_irqsave(&ce->guc_state.lock, flags);
+	disabled = submission_disabled(guc);
+	if (likely(!disabled)) {
+		__intel_gt_pm_get(gt);
+		set_context_destroyed(ce);
+		clr_context_registered(ce);
+	}
+	spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+	if (unlikely(disabled)) {
+		release_guc_id(guc, ce);
+		__guc_context_destroy(ce);
+		return;
+	}
+
+	deregister_context(ce, ce->guc_id.id);
+}
+
+static void __guc_context_destroy(struct intel_context *ce)
+{
+	GEM_BUG_ON(ce->guc_state.prio_count[GUC_CLIENT_PRIORITY_KMD_HIGH] ||
+		   ce->guc_state.prio_count[GUC_CLIENT_PRIORITY_HIGH] ||
+		   ce->guc_state.prio_count[GUC_CLIENT_PRIORITY_KMD_NORMAL] ||
+		   ce->guc_state.prio_count[GUC_CLIENT_PRIORITY_NORMAL]);
+	GEM_BUG_ON(ce->guc_state.number_committed_requests);
+
+	lrc_fini(ce);
+	intel_context_fini(ce);
+
+	if (intel_engine_is_virtual(ce->engine)) {
+		struct guc_virtual_engine *ve =
+			container_of(ce, typeof(*ve), context);
+
+		if (ve->base.breadcrumbs)
+			intel_breadcrumbs_put(ve->base.breadcrumbs);
+
+		kfree(ve);
+	} else {
+		intel_context_free(ce);
+	}
+}
+
+static void guc_flush_destroyed_contexts(struct intel_guc *guc)
+{
+	struct intel_context *ce;
+	unsigned long flags;
+
+	GEM_BUG_ON(!submission_disabled(guc) &&
+		   guc_submission_initialized(guc));
+
+	while (!list_empty(&guc->submission_state.destroyed_contexts)) {
+		spin_lock_irqsave(&guc->submission_state.lock, flags);
+		ce = list_first_entry_or_null(&guc->submission_state.destroyed_contexts,
+					      struct intel_context,
+					      destroyed_link);
+		if (ce)
+			list_del_init(&ce->destroyed_link);
+		spin_unlock_irqrestore(&guc->submission_state.lock, flags);
+
+		if (!ce)
+			break;
+
+		release_guc_id(guc, ce);
+		__guc_context_destroy(ce);
+	}
+}
+
+static void deregister_destroyed_contexts(struct intel_guc *guc)
+{
+	struct intel_context *ce;
+	unsigned long flags;
+
+	while (!list_empty(&guc->submission_state.destroyed_contexts)) {
+		spin_lock_irqsave(&guc->submission_state.lock, flags);
+		ce = list_first_entry_or_null(&guc->submission_state.destroyed_contexts,
+					      struct intel_context,
+					      destroyed_link);
+		if (ce)
+			list_del_init(&ce->destroyed_link);
+		spin_unlock_irqrestore(&guc->submission_state.lock, flags);
+
+		if (!ce)
+			break;
+
+		guc_lrc_desc_unpin(ce);
+	}
+}
+
+static void destroyed_worker_func(struct work_struct *w)
+{
+	struct intel_guc *guc = container_of(w, struct intel_guc,
+					     submission_state.destroyed_worker);
+	struct intel_gt *gt = guc_to_gt(guc);
+	int tmp;
+
+	with_intel_gt_pm(gt, tmp)
+		deregister_destroyed_contexts(guc);
+}
+
+static void guc_context_destroy(struct kref *kref)
+{
+	struct intel_context *ce = container_of(kref, typeof(*ce), ref);
+	struct intel_guc *guc = ce_to_guc(ce);
+	unsigned long flags;
+	bool destroy;
+
+	/*
+	 * If the guc_id is invalid this context has been stolen and we can free
+	 * it immediately. Also can be freed immediately if the context is not
+	 * registered with the GuC or the GuC is in the middle of a reset.
+	 */
+	spin_lock_irqsave(&guc->submission_state.lock, flags);
+	destroy = submission_disabled(guc) || context_guc_id_invalid(ce) ||
+		!ctx_id_mapped(guc, ce->guc_id.id);
+	if (likely(!destroy)) {
+		if (!list_empty(&ce->guc_id.link))
+			list_del_init(&ce->guc_id.link);
+		list_add_tail(&ce->destroyed_link,
+			      &guc->submission_state.destroyed_contexts);
+	} else {
+		__release_guc_id(guc, ce);
+	}
+	spin_unlock_irqrestore(&guc->submission_state.lock, flags);
+	if (unlikely(destroy)) {
+		__guc_context_destroy(ce);
+		return;
+	}
+
+	/*
+	 * We use a worker to issue the H2G to deregister the context as we can
+	 * take the GT PM for the first time which isn't allowed from an atomic
+	 * context.
+	 */
+	queue_work(system_unbound_wq, &guc->submission_state.destroyed_worker);
+}
+
+static int guc_context_alloc(struct intel_context *ce)
+{
+	return lrc_alloc(ce, ce->engine);
+}
+
+static void __guc_context_set_prio(struct intel_guc *guc,
+				   struct intel_context *ce)
+{
+	if (GET_UC_VER(guc) >= MAKE_UC_VER(70, 0, 0)) {
+		struct context_policy policy;
+
+		__guc_context_policy_start_klv(&policy, ce->guc_id.id);
+		__guc_context_policy_add_priority(&policy, ce->guc_state.prio);
+		__guc_context_set_context_policies(guc, &policy, true);
+	} else {
+		u32 action[] = {
+			INTEL_GUC_ACTION_V69_SET_CONTEXT_PRIORITY,
+			ce->guc_id.id,
+			ce->guc_state.prio,
+		};
+
+		guc_submission_send_busy_loop(guc, action, ARRAY_SIZE(action), 0, true);
+	}
+}
+
+static void guc_context_set_prio(struct intel_guc *guc,
+				 struct intel_context *ce,
+				 u8 prio)
+{
+	GEM_BUG_ON(prio < GUC_CLIENT_PRIORITY_KMD_HIGH ||
+		   prio > GUC_CLIENT_PRIORITY_NORMAL);
+	lockdep_assert_held(&ce->guc_state.lock);
+
+	if (ce->guc_state.prio == prio || submission_disabled(guc) ||
+	    !context_registered(ce)) {
+		ce->guc_state.prio = prio;
+		return;
+	}
+
+	ce->guc_state.prio = prio;
+	__guc_context_set_prio(guc, ce);
+
+	trace_intel_context_set_prio(ce);
+}
+
+static inline u8 map_i915_prio_to_guc_prio(int prio)
+{
+	if (prio == I915_PRIORITY_NORMAL)
+		return GUC_CLIENT_PRIORITY_KMD_NORMAL;
+	else if (prio < I915_PRIORITY_NORMAL)
+		return GUC_CLIENT_PRIORITY_NORMAL;
+	else if (prio < I915_PRIORITY_DISPLAY)
+		return GUC_CLIENT_PRIORITY_HIGH;
+	else
+		return GUC_CLIENT_PRIORITY_KMD_HIGH;
+}
+
+static inline void add_context_inflight_prio(struct intel_context *ce,
+					     u8 guc_prio)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	GEM_BUG_ON(guc_prio >= ARRAY_SIZE(ce->guc_state.prio_count));
+
+	++ce->guc_state.prio_count[guc_prio];
+
+	/* Overflow protection */
+	GEM_WARN_ON(!ce->guc_state.prio_count[guc_prio]);
+}
+
+static inline void sub_context_inflight_prio(struct intel_context *ce,
+					     u8 guc_prio)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+	GEM_BUG_ON(guc_prio >= ARRAY_SIZE(ce->guc_state.prio_count));
+
+	/* Underflow protection */
+	GEM_WARN_ON(!ce->guc_state.prio_count[guc_prio]);
+
+	--ce->guc_state.prio_count[guc_prio];
+}
+
+static inline void update_context_prio(struct intel_context *ce)
+{
+	struct intel_guc *guc = &ce->engine->gt->uc.guc;
+	int i;
+
+	BUILD_BUG_ON(GUC_CLIENT_PRIORITY_KMD_HIGH != 0);
+	BUILD_BUG_ON(GUC_CLIENT_PRIORITY_KMD_HIGH > GUC_CLIENT_PRIORITY_NORMAL);
+
+	lockdep_assert_held(&ce->guc_state.lock);
+
+	for (i = 0; i < ARRAY_SIZE(ce->guc_state.prio_count); ++i) {
+		if (ce->guc_state.prio_count[i]) {
+			guc_context_set_prio(guc, ce, i);
+			break;
+		}
+	}
+}
+
+static inline bool new_guc_prio_higher(u8 old_guc_prio, u8 new_guc_prio)
+{
+	/* Lower value is higher priority */
+	return new_guc_prio < old_guc_prio;
+}
+
+static void add_to_context(struct i915_request *rq)
+{
+	struct intel_context *ce = request_to_scheduling_context(rq);
+	u8 new_guc_prio = map_i915_prio_to_guc_prio(rq_prio(rq));
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+	GEM_BUG_ON(rq->guc_prio == GUC_PRIO_FINI);
+
+	spin_lock(&ce->guc_state.lock);
+	list_move_tail(&rq->sched.link, &ce->guc_state.requests);
+
+	if (rq->guc_prio == GUC_PRIO_INIT) {
+		rq->guc_prio = new_guc_prio;
+		add_context_inflight_prio(ce, rq->guc_prio);
+	} else if (new_guc_prio_higher(rq->guc_prio, new_guc_prio)) {
+		sub_context_inflight_prio(ce, rq->guc_prio);
+		rq->guc_prio = new_guc_prio;
+		add_context_inflight_prio(ce, rq->guc_prio);
+	}
+	update_context_prio(ce);
+
+	spin_unlock(&ce->guc_state.lock);
+}
+
+static void guc_prio_fini(struct i915_request *rq, struct intel_context *ce)
+{
+	lockdep_assert_held(&ce->guc_state.lock);
+
+	if (rq->guc_prio != GUC_PRIO_INIT &&
+	    rq->guc_prio != GUC_PRIO_FINI) {
+		sub_context_inflight_prio(ce, rq->guc_prio);
+		update_context_prio(ce);
+	}
+	rq->guc_prio = GUC_PRIO_FINI;
+}
+
+static void remove_from_context(struct i915_request *rq)
+{
+	struct intel_context *ce = request_to_scheduling_context(rq);
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	spin_lock_irq(&ce->guc_state.lock);
+
+	list_del_init(&rq->sched.link);
+	clear_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+
+	/* Prevent further __await_execution() registering a cb, then flush */
+	set_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags);
+
+	guc_prio_fini(rq, ce);
+
+	decr_context_committed_requests(ce);
+
+	spin_unlock_irq(&ce->guc_state.lock);
+
+	atomic_dec(&ce->guc_id.ref);
+	i915_request_notify_execute_cb_imm(rq);
+}
+
+static const struct intel_context_ops guc_context_ops = {
+	.alloc = guc_context_alloc,
+
+	.pre_pin = guc_context_pre_pin,
+	.pin = guc_context_pin,
+	.unpin = guc_context_unpin,
+	.post_unpin = guc_context_post_unpin,
+
+	.revoke = guc_context_revoke,
+
+	.cancel_request = guc_context_cancel_request,
+
+	.enter = intel_context_enter_engine,
+	.exit = intel_context_exit_engine,
+
+	.sched_disable = guc_context_sched_disable,
+
+	.reset = lrc_reset,
+	.destroy = guc_context_destroy,
+
+	.create_virtual = guc_create_virtual,
+	.create_parallel = guc_create_parallel,
+};
+
+static void submit_work_cb(struct irq_work *wrk)
+{
+	struct i915_request *rq = container_of(wrk, typeof(*rq), submit_work);
+
+	might_lock(&rq->engine->sched_engine->lock);
+	i915_sw_fence_complete(&rq->submit);
+}
+
+static void __guc_signal_context_fence(struct intel_context *ce)
+{
+	struct i915_request *rq, *rn;
+
+	lockdep_assert_held(&ce->guc_state.lock);
+
+	if (!list_empty(&ce->guc_state.fences))
+		trace_intel_context_fence_release(ce);
+
+	/*
+	 * Use an IRQ to ensure locking order of sched_engine->lock ->
+	 * ce->guc_state.lock is preserved.
+	 */
+	list_for_each_entry_safe(rq, rn, &ce->guc_state.fences,
+				 guc_fence_link) {
+		list_del(&rq->guc_fence_link);
+		irq_work_queue(&rq->submit_work);
+	}
+
+	INIT_LIST_HEAD(&ce->guc_state.fences);
+}
+
+static void guc_signal_context_fence(struct intel_context *ce)
+{
+	unsigned long flags;
+
+	GEM_BUG_ON(intel_context_is_child(ce));
+
+	spin_lock_irqsave(&ce->guc_state.lock, flags);
+	clr_context_wait_for_deregister_to_register(ce);
+	__guc_signal_context_fence(ce);
+	spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+}
+
+static bool context_needs_register(struct intel_context *ce, bool new_guc_id)
+{
+	return (new_guc_id || test_bit(CONTEXT_LRCA_DIRTY, &ce->flags) ||
+		!ctx_id_mapped(ce_to_guc(ce), ce->guc_id.id)) &&
+		!submission_disabled(ce_to_guc(ce));
+}
+
+static void guc_context_init(struct intel_context *ce)
+{
+	const struct i915_gem_context *ctx;
+	int prio = I915_CONTEXT_DEFAULT_PRIORITY;
+
+	rcu_read_lock();
+	ctx = rcu_dereference(ce->gem_context);
+	if (ctx)
+		prio = ctx->sched.priority;
+	rcu_read_unlock();
+
+	ce->guc_state.prio = map_i915_prio_to_guc_prio(prio);
+	set_bit(CONTEXT_GUC_INIT, &ce->flags);
+}
+
+static int guc_request_alloc(struct i915_request *rq)
+{
+	struct intel_context *ce = request_to_scheduling_context(rq);
+	struct intel_guc *guc = ce_to_guc(ce);
+	unsigned long flags;
+	int ret;
+
+	GEM_BUG_ON(!intel_context_is_pinned(rq->context));
+
+	/*
+	 * Flush enough space to reduce the likelihood of waiting after
+	 * we start building the request - in which case we will just
+	 * have to repeat work.
+	 */
+	rq->reserved_space += GUC_REQUEST_SIZE;
+
+	/*
+	 * Note that after this point, we have committed to using
+	 * this request as it is being used to both track the
+	 * state of engine initialisation and liveness of the
+	 * golden renderstate above. Think twice before you try
+	 * to cancel/unwind this request now.
+	 */
+
+	/* Unconditionally invalidate GPU caches and TLBs. */
+	ret = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+	if (ret)
+		return ret;
+
+	rq->reserved_space -= GUC_REQUEST_SIZE;
+
+	if (unlikely(!test_bit(CONTEXT_GUC_INIT, &ce->flags)))
+		guc_context_init(ce);
+
+	/*
+	 * Call pin_guc_id here rather than in the pinning step as with
+	 * dma_resv, contexts can be repeatedly pinned / unpinned trashing the
+	 * guc_id and creating horrible race conditions. This is especially bad
+	 * when guc_id are being stolen due to over subscription. By the time
+	 * this function is reached, it is guaranteed that the guc_id will be
+	 * persistent until the generated request is retired. Thus, sealing these
+	 * race conditions. It is still safe to fail here if guc_id are
+	 * exhausted and return -EAGAIN to the user indicating that they can try
+	 * again in the future.
+	 *
+	 * There is no need for a lock here as the timeline mutex ensures at
+	 * most one context can be executing this code path at once. The
+	 * guc_id_ref is incremented once for every request in flight and
+	 * decremented on each retire. When it is zero, a lock around the
+	 * increment (in pin_guc_id) is needed to seal a race with unpin_guc_id.
+	 */
+	if (atomic_add_unless(&ce->guc_id.ref, 1, 0))
+		goto out;
+
+	ret = pin_guc_id(guc, ce);	/* returns 1 if new guc_id assigned */
+	if (unlikely(ret < 0))
+		return ret;
+	if (context_needs_register(ce, !!ret)) {
+		ret = try_context_registration(ce, true);
+		if (unlikely(ret)) {	/* unwind */
+			if (ret == -EPIPE) {
+				disable_submission(guc);
+				goto out;	/* GPU will be reset */
+			}
+			atomic_dec(&ce->guc_id.ref);
+			unpin_guc_id(guc, ce);
+			return ret;
+		}
+	}
+
+	clear_bit(CONTEXT_LRCA_DIRTY, &ce->flags);
+
+out:
+	/*
+	 * We block all requests on this context if a G2H is pending for a
+	 * schedule disable or context deregistration as the GuC will fail a
+	 * schedule enable or context registration if either G2H is pending
+	 * respectfully. Once a G2H returns, the fence is released that is
+	 * blocking these requests (see guc_signal_context_fence).
+	 */
+	spin_lock_irqsave(&ce->guc_state.lock, flags);
+	if (context_wait_for_deregister_to_register(ce) ||
+	    context_pending_disable(ce)) {
+		init_irq_work(&rq->submit_work, submit_work_cb);
+		i915_sw_fence_await(&rq->submit);
+
+		list_add_tail(&rq->guc_fence_link, &ce->guc_state.fences);
+	}
+	incr_context_committed_requests(ce);
+	spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+
+	return 0;
+}
+
+static int guc_virtual_context_pre_pin(struct intel_context *ce,
+				       struct i915_gem_ww_ctx *ww,
+				       void **vaddr)
+{
+	struct intel_engine_cs *engine = guc_virtual_get_sibling(ce->engine, 0);
+
+	return __guc_context_pre_pin(ce, engine, ww, vaddr);
+}
+
+static int guc_virtual_context_pin(struct intel_context *ce, void *vaddr)
+{
+	struct intel_engine_cs *engine = guc_virtual_get_sibling(ce->engine, 0);
+	int ret = __guc_context_pin(ce, engine, vaddr);
+	intel_engine_mask_t tmp, mask = ce->engine->mask;
+
+	if (likely(!ret))
+		for_each_engine_masked(engine, ce->engine->gt, mask, tmp)
+			intel_engine_pm_get(engine);
+
+	return ret;
+}
+
+static void guc_virtual_context_unpin(struct intel_context *ce)
+{
+	intel_engine_mask_t tmp, mask = ce->engine->mask;
+	struct intel_engine_cs *engine;
+	struct intel_guc *guc = ce_to_guc(ce);
+
+	GEM_BUG_ON(context_enabled(ce));
+	GEM_BUG_ON(intel_context_is_barrier(ce));
+
+	unpin_guc_id(guc, ce);
+	lrc_unpin(ce);
+
+	for_each_engine_masked(engine, ce->engine->gt, mask, tmp)
+		intel_engine_pm_put_async(engine);
+}
+
+static void guc_virtual_context_enter(struct intel_context *ce)
+{
+	intel_engine_mask_t tmp, mask = ce->engine->mask;
+	struct intel_engine_cs *engine;
+
+	for_each_engine_masked(engine, ce->engine->gt, mask, tmp)
+		intel_engine_pm_get(engine);
+
+	intel_timeline_enter(ce->timeline);
+}
+
+static void guc_virtual_context_exit(struct intel_context *ce)
+{
+	intel_engine_mask_t tmp, mask = ce->engine->mask;
+	struct intel_engine_cs *engine;
+
+	for_each_engine_masked(engine, ce->engine->gt, mask, tmp)
+		intel_engine_pm_put(engine);
+
+	intel_timeline_exit(ce->timeline);
+}
+
+static int guc_virtual_context_alloc(struct intel_context *ce)
+{
+	struct intel_engine_cs *engine = guc_virtual_get_sibling(ce->engine, 0);
+
+	return lrc_alloc(ce, engine);
+}
+
+static const struct intel_context_ops virtual_guc_context_ops = {
+	.alloc = guc_virtual_context_alloc,
+
+	.pre_pin = guc_virtual_context_pre_pin,
+	.pin = guc_virtual_context_pin,
+	.unpin = guc_virtual_context_unpin,
+	.post_unpin = guc_context_post_unpin,
+
+	.revoke = guc_context_revoke,
+
+	.cancel_request = guc_context_cancel_request,
+
+	.enter = guc_virtual_context_enter,
+	.exit = guc_virtual_context_exit,
+
+	.sched_disable = guc_context_sched_disable,
+
+	.destroy = guc_context_destroy,
+
+	.get_sibling = guc_virtual_get_sibling,
+};
+
+static int guc_parent_context_pin(struct intel_context *ce, void *vaddr)
+{
+	struct intel_engine_cs *engine = guc_virtual_get_sibling(ce->engine, 0);
+	struct intel_guc *guc = ce_to_guc(ce);
+	int ret;
+
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+	GEM_BUG_ON(!intel_engine_is_virtual(ce->engine));
+
+	ret = pin_guc_id(guc, ce);
+	if (unlikely(ret < 0))
+		return ret;
+
+	return __guc_context_pin(ce, engine, vaddr);
+}
+
+static int guc_child_context_pin(struct intel_context *ce, void *vaddr)
+{
+	struct intel_engine_cs *engine = guc_virtual_get_sibling(ce->engine, 0);
+
+	GEM_BUG_ON(!intel_context_is_child(ce));
+	GEM_BUG_ON(!intel_engine_is_virtual(ce->engine));
+
+	__intel_context_pin(ce->parallel.parent);
+	return __guc_context_pin(ce, engine, vaddr);
+}
+
+static void guc_parent_context_unpin(struct intel_context *ce)
+{
+	struct intel_guc *guc = ce_to_guc(ce);
+
+	GEM_BUG_ON(context_enabled(ce));
+	GEM_BUG_ON(intel_context_is_barrier(ce));
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+	GEM_BUG_ON(!intel_engine_is_virtual(ce->engine));
+
+	unpin_guc_id(guc, ce);
+	lrc_unpin(ce);
+}
+
+static void guc_child_context_unpin(struct intel_context *ce)
+{
+	GEM_BUG_ON(context_enabled(ce));
+	GEM_BUG_ON(intel_context_is_barrier(ce));
+	GEM_BUG_ON(!intel_context_is_child(ce));
+	GEM_BUG_ON(!intel_engine_is_virtual(ce->engine));
+
+	lrc_unpin(ce);
+}
+
+static void guc_child_context_post_unpin(struct intel_context *ce)
+{
+	GEM_BUG_ON(!intel_context_is_child(ce));
+	GEM_BUG_ON(!intel_context_is_pinned(ce->parallel.parent));
+	GEM_BUG_ON(!intel_engine_is_virtual(ce->engine));
+
+	lrc_post_unpin(ce);
+	intel_context_unpin(ce->parallel.parent);
+}
+
+static void guc_child_context_destroy(struct kref *kref)
+{
+	struct intel_context *ce = container_of(kref, typeof(*ce), ref);
+
+	__guc_context_destroy(ce);
+}
+
+static const struct intel_context_ops virtual_parent_context_ops = {
+	.alloc = guc_virtual_context_alloc,
+
+	.pre_pin = guc_context_pre_pin,
+	.pin = guc_parent_context_pin,
+	.unpin = guc_parent_context_unpin,
+	.post_unpin = guc_context_post_unpin,
+
+	.revoke = guc_context_revoke,
+
+	.cancel_request = guc_context_cancel_request,
+
+	.enter = guc_virtual_context_enter,
+	.exit = guc_virtual_context_exit,
+
+	.sched_disable = guc_context_sched_disable,
+
+	.destroy = guc_context_destroy,
+
+	.get_sibling = guc_virtual_get_sibling,
+};
+
+static const struct intel_context_ops virtual_child_context_ops = {
+	.alloc = guc_virtual_context_alloc,
+
+	.pre_pin = guc_context_pre_pin,
+	.pin = guc_child_context_pin,
+	.unpin = guc_child_context_unpin,
+	.post_unpin = guc_child_context_post_unpin,
+
+	.cancel_request = guc_context_cancel_request,
+
+	.enter = guc_virtual_context_enter,
+	.exit = guc_virtual_context_exit,
+
+	.destroy = guc_child_context_destroy,
+
+	.get_sibling = guc_virtual_get_sibling,
+};
+
+/*
+ * The below override of the breadcrumbs is enabled when the user configures a
+ * context for parallel submission (multi-lrc, parent-child).
+ *
+ * The overridden breadcrumbs implements an algorithm which allows the GuC to
+ * safely preempt all the hw contexts configured for parallel submission
+ * between each BB. The contract between the i915 and GuC is if the parent
+ * context can be preempted, all the children can be preempted, and the GuC will
+ * always try to preempt the parent before the children. A handshake between the
+ * parent / children breadcrumbs ensures the i915 holds up its end of the deal
+ * creating a window to preempt between each set of BBs.
+ */
+static int emit_bb_start_parent_no_preempt_mid_batch(struct i915_request *rq,
+						     u64 offset, u32 len,
+						     const unsigned int flags);
+static int emit_bb_start_child_no_preempt_mid_batch(struct i915_request *rq,
+						    u64 offset, u32 len,
+						    const unsigned int flags);
+static u32 *
+emit_fini_breadcrumb_parent_no_preempt_mid_batch(struct i915_request *rq,
+						 u32 *cs);
+static u32 *
+emit_fini_breadcrumb_child_no_preempt_mid_batch(struct i915_request *rq,
+						u32 *cs);
+
+static struct intel_context *
+guc_create_parallel(struct intel_engine_cs **engines,
+		    unsigned int num_siblings,
+		    unsigned int width)
+{
+	struct intel_engine_cs **siblings = NULL;
+	struct intel_context *parent = NULL, *ce, *err;
+	int i, j;
+
+	siblings = kmalloc_array(num_siblings,
+				 sizeof(*siblings),
+				 GFP_KERNEL);
+	if (!siblings)
+		return ERR_PTR(-ENOMEM);
+
+	for (i = 0; i < width; ++i) {
+		for (j = 0; j < num_siblings; ++j)
+			siblings[j] = engines[i * num_siblings + j];
+
+		ce = intel_engine_create_virtual(siblings, num_siblings,
+						 FORCE_VIRTUAL);
+		if (IS_ERR(ce)) {
+			err = ERR_CAST(ce);
+			goto unwind;
+		}
+
+		if (i == 0) {
+			parent = ce;
+			parent->ops = &virtual_parent_context_ops;
+		} else {
+			ce->ops = &virtual_child_context_ops;
+			intel_context_bind_parent_child(parent, ce);
+		}
+	}
+
+	parent->parallel.fence_context = dma_fence_context_alloc(1);
+
+	parent->engine->emit_bb_start =
+		emit_bb_start_parent_no_preempt_mid_batch;
+	parent->engine->emit_fini_breadcrumb =
+		emit_fini_breadcrumb_parent_no_preempt_mid_batch;
+	parent->engine->emit_fini_breadcrumb_dw =
+		12 + 4 * parent->parallel.number_children;
+	for_each_child(parent, ce) {
+		ce->engine->emit_bb_start =
+			emit_bb_start_child_no_preempt_mid_batch;
+		ce->engine->emit_fini_breadcrumb =
+			emit_fini_breadcrumb_child_no_preempt_mid_batch;
+		ce->engine->emit_fini_breadcrumb_dw = 16;
+	}
+
+	kfree(siblings);
+	return parent;
+
+unwind:
+	if (parent)
+		intel_context_put(parent);
+	kfree(siblings);
+	return err;
+}
+
+static bool
+guc_irq_enable_breadcrumbs(struct intel_breadcrumbs *b)
+{
+	struct intel_engine_cs *sibling;
+	intel_engine_mask_t tmp, mask = b->engine_mask;
+	bool result = false;
+
+	for_each_engine_masked(sibling, b->irq_engine->gt, mask, tmp)
+		result |= intel_engine_irq_enable(sibling);
+
+	return result;
+}
+
+static void
+guc_irq_disable_breadcrumbs(struct intel_breadcrumbs *b)
+{
+	struct intel_engine_cs *sibling;
+	intel_engine_mask_t tmp, mask = b->engine_mask;
+
+	for_each_engine_masked(sibling, b->irq_engine->gt, mask, tmp)
+		intel_engine_irq_disable(sibling);
+}
+
+static void guc_init_breadcrumbs(struct intel_engine_cs *engine)
+{
+	int i;
+
+	/*
+	 * In GuC submission mode we do not know which physical engine a request
+	 * will be scheduled on, this creates a problem because the breadcrumb
+	 * interrupt is per physical engine. To work around this we attach
+	 * requests and direct all breadcrumb interrupts to the first instance
+	 * of an engine per class. In addition all breadcrumb interrupts are
+	 * enabled / disabled across an engine class in unison.
+	 */
+	for (i = 0; i < MAX_ENGINE_INSTANCE; ++i) {
+		struct intel_engine_cs *sibling =
+			engine->gt->engine_class[engine->class][i];
+
+		if (sibling) {
+			if (engine->breadcrumbs != sibling->breadcrumbs) {
+				intel_breadcrumbs_put(engine->breadcrumbs);
+				engine->breadcrumbs =
+					intel_breadcrumbs_get(sibling->breadcrumbs);
+			}
+			break;
+		}
+	}
+
+	if (engine->breadcrumbs) {
+		engine->breadcrumbs->engine_mask |= engine->mask;
+		engine->breadcrumbs->irq_enable = guc_irq_enable_breadcrumbs;
+		engine->breadcrumbs->irq_disable = guc_irq_disable_breadcrumbs;
+	}
+}
+
+static void guc_bump_inflight_request_prio(struct i915_request *rq,
+					   int prio)
+{
+	struct intel_context *ce = request_to_scheduling_context(rq);
+	u8 new_guc_prio = map_i915_prio_to_guc_prio(prio);
+
+	/* Short circuit function */
+	if (prio < I915_PRIORITY_NORMAL ||
+	    rq->guc_prio == GUC_PRIO_FINI ||
+	    (rq->guc_prio != GUC_PRIO_INIT &&
+	     !new_guc_prio_higher(rq->guc_prio, new_guc_prio)))
+		return;
+
+	spin_lock(&ce->guc_state.lock);
+	if (rq->guc_prio != GUC_PRIO_FINI) {
+		if (rq->guc_prio != GUC_PRIO_INIT)
+			sub_context_inflight_prio(ce, rq->guc_prio);
+		rq->guc_prio = new_guc_prio;
+		add_context_inflight_prio(ce, rq->guc_prio);
+		update_context_prio(ce);
+	}
+	spin_unlock(&ce->guc_state.lock);
+}
+
+static void guc_retire_inflight_request_prio(struct i915_request *rq)
+{
+	struct intel_context *ce = request_to_scheduling_context(rq);
+
+	spin_lock(&ce->guc_state.lock);
+	guc_prio_fini(rq, ce);
+	spin_unlock(&ce->guc_state.lock);
+}
+
+static void sanitize_hwsp(struct intel_engine_cs *engine)
+{
+	struct intel_timeline *tl;
+
+	list_for_each_entry(tl, &engine->status_page.timelines, engine_link)
+		intel_timeline_reset_seqno(tl);
+}
+
+static void guc_sanitize(struct intel_engine_cs *engine)
+{
+	/*
+	 * Poison residual state on resume, in case the suspend didn't!
+	 *
+	 * We have to assume that across suspend/resume (or other loss
+	 * of control) that the contents of our pinned buffers has been
+	 * lost, replaced by garbage. Since this doesn't always happen,
+	 * let's poison such state so that we more quickly spot when
+	 * we falsely assume it has been preserved.
+	 */
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		memset(engine->status_page.addr, POISON_INUSE, PAGE_SIZE);
+
+	/*
+	 * The kernel_context HWSP is stored in the status_page. As above,
+	 * that may be lost on resume/initialisation, and so we need to
+	 * reset the value in the HWSP.
+	 */
+	sanitize_hwsp(engine);
+
+	/* And scrub the dirty cachelines for the HWSP */
+	drm_clflush_virt_range(engine->status_page.addr, PAGE_SIZE);
+
+	intel_engine_reset_pinned_contexts(engine);
+}
+
+static void setup_hwsp(struct intel_engine_cs *engine)
+{
+	intel_engine_set_hwsp_writemask(engine, ~0u); /* HWSTAM */
+
+	ENGINE_WRITE_FW(engine,
+			RING_HWS_PGA,
+			i915_ggtt_offset(engine->status_page.vma));
+}
+
+static void start_engine(struct intel_engine_cs *engine)
+{
+	ENGINE_WRITE_FW(engine,
+			RING_MODE_GEN7,
+			_MASKED_BIT_ENABLE(GEN11_GFX_DISABLE_LEGACY_MODE));
+
+	ENGINE_WRITE_FW(engine, RING_MI_MODE, _MASKED_BIT_DISABLE(STOP_RING));
+	ENGINE_POSTING_READ(engine, RING_MI_MODE);
+}
+
+static int guc_resume(struct intel_engine_cs *engine)
+{
+	assert_forcewakes_active(engine->uncore, FORCEWAKE_ALL);
+
+	intel_mocs_init_engine(engine);
+
+	intel_breadcrumbs_reset(engine->breadcrumbs);
+
+	setup_hwsp(engine);
+	start_engine(engine);
+
+	if (engine->flags & I915_ENGINE_FIRST_RENDER_COMPUTE)
+		xehp_enable_ccs_engines(engine);
+
+	return 0;
+}
+
+static bool guc_sched_engine_disabled(struct i915_sched_engine *sched_engine)
+{
+	return !sched_engine->tasklet.callback;
+}
+
+static void guc_set_default_submission(struct intel_engine_cs *engine)
+{
+	engine->submit_request = guc_submit_request;
+}
+
+static inline void guc_kernel_context_pin(struct intel_guc *guc,
+					  struct intel_context *ce)
+{
+	/*
+	 * Note: we purposefully do not check the returns below because
+	 * the registration can only fail if a reset is just starting.
+	 * This is called at the end of reset so presumably another reset
+	 * isn't happening and even it did this code would be run again.
+	 */
+
+	if (context_guc_id_invalid(ce))
+		pin_guc_id(guc, ce);
+
+	try_context_registration(ce, true);
+}
+
+static inline void guc_init_lrc_mapping(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/* make sure all descriptors are clean... */
+	xa_destroy(&guc->context_lookup);
+
+	/*
+	 * A reset might have occurred while we had a pending stalled request,
+	 * so make sure we clean that up.
+	 */
+	guc->stalled_request = NULL;
+	guc->submission_stall_reason = STALL_NONE;
+
+	/*
+	 * Some contexts might have been pinned before we enabled GuC
+	 * submission, so we need to add them to the GuC bookeeping.
+	 * Also, after a reset the of the GuC we want to make sure that the
+	 * information shared with GuC is properly reset. The kernel LRCs are
+	 * not attached to the gem_context, so they need to be added separately.
+	 */
+	for_each_engine(engine, gt, id) {
+		struct intel_context *ce;
+
+		list_for_each_entry(ce, &engine->pinned_contexts_list,
+				    pinned_contexts_link)
+			guc_kernel_context_pin(guc, ce);
+	}
+}
+
+static void guc_release(struct intel_engine_cs *engine)
+{
+	engine->sanitize = NULL; /* no longer in control, nothing to sanitize */
+
+	intel_engine_cleanup_common(engine);
+	lrc_fini_wa_ctx(engine);
+}
+
+static void virtual_guc_bump_serial(struct intel_engine_cs *engine)
+{
+	struct intel_engine_cs *e;
+	intel_engine_mask_t tmp, mask = engine->mask;
+
+	for_each_engine_masked(e, engine->gt, mask, tmp)
+		e->serial++;
+}
+
+static void guc_default_vfuncs(struct intel_engine_cs *engine)
+{
+	/* Default vfuncs which can be overridden by each engine. */
+
+	engine->resume = guc_resume;
+
+	engine->cops = &guc_context_ops;
+	engine->request_alloc = guc_request_alloc;
+	engine->add_active_request = add_to_context;
+	engine->remove_active_request = remove_from_context;
+
+	engine->sched_engine->schedule = i915_schedule;
+
+	engine->reset.prepare = guc_engine_reset_prepare;
+	engine->reset.rewind = guc_rewind_nop;
+	engine->reset.cancel = guc_reset_nop;
+	engine->reset.finish = guc_reset_nop;
+
+	engine->emit_flush = gen8_emit_flush_xcs;
+	engine->emit_init_breadcrumb = gen8_emit_init_breadcrumb;
+	engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb_xcs;
+	if (GRAPHICS_VER(engine->i915) >= 12) {
+		engine->emit_fini_breadcrumb = gen12_emit_fini_breadcrumb_xcs;
+		engine->emit_flush = gen12_emit_flush_xcs;
+	}
+	engine->set_default_submission = guc_set_default_submission;
+	engine->busyness = guc_engine_busyness;
+
+	engine->flags |= I915_ENGINE_SUPPORTS_STATS;
+	engine->flags |= I915_ENGINE_HAS_PREEMPTION;
+	engine->flags |= I915_ENGINE_HAS_TIMESLICES;
+
+	/* Wa_14014475959:dg2 */
+	if (IS_DG2(engine->i915) && engine->class == COMPUTE_CLASS)
+		engine->flags |= I915_ENGINE_USES_WA_HOLD_CCS_SWITCHOUT;
+
+	/*
+	 * TODO: GuC supports timeslicing and semaphores as well, but they're
+	 * handled by the firmware so some minor tweaks are required before
+	 * enabling.
+	 *
+	 * engine->flags |= I915_ENGINE_HAS_SEMAPHORES;
+	 */
+
+	engine->emit_bb_start = gen8_emit_bb_start;
+	if (GRAPHICS_VER_FULL(engine->i915) >= IP_VER(12, 50))
+		engine->emit_bb_start = gen125_emit_bb_start;
+}
+
+static void rcs_submission_override(struct intel_engine_cs *engine)
+{
+	switch (GRAPHICS_VER(engine->i915)) {
+	case 12:
+		engine->emit_flush = gen12_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen12_emit_fini_breadcrumb_rcs;
+		break;
+	case 11:
+		engine->emit_flush = gen11_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen11_emit_fini_breadcrumb_rcs;
+		break;
+	default:
+		engine->emit_flush = gen8_emit_flush_rcs;
+		engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb_rcs;
+		break;
+	}
+}
+
+static inline void guc_default_irqs(struct intel_engine_cs *engine)
+{
+	engine->irq_keep_mask = GT_RENDER_USER_INTERRUPT;
+	intel_engine_set_irq_handler(engine, cs_irq_handler);
+}
+
+static void guc_sched_engine_destroy(struct kref *kref)
+{
+	struct i915_sched_engine *sched_engine =
+		container_of(kref, typeof(*sched_engine), ref);
+	struct intel_guc *guc = sched_engine->private_data;
+
+	guc->sched_engine = NULL;
+	tasklet_kill(&sched_engine->tasklet); /* flush the callback */
+	kfree(sched_engine);
+}
+
+int intel_guc_submission_setup(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *i915 = engine->i915;
+	struct intel_guc *guc = &engine->gt->uc.guc;
+
+	/*
+	 * The setup relies on several assumptions (e.g. irqs always enabled)
+	 * that are only valid on gen11+
+	 */
+	GEM_BUG_ON(GRAPHICS_VER(i915) < 11);
+
+	if (!guc->sched_engine) {
+		guc->sched_engine = i915_sched_engine_create(ENGINE_VIRTUAL);
+		if (!guc->sched_engine)
+			return -ENOMEM;
+
+		guc->sched_engine->schedule = i915_schedule;
+		guc->sched_engine->disabled = guc_sched_engine_disabled;
+		guc->sched_engine->private_data = guc;
+		guc->sched_engine->destroy = guc_sched_engine_destroy;
+		guc->sched_engine->bump_inflight_request_prio =
+			guc_bump_inflight_request_prio;
+		guc->sched_engine->retire_inflight_request_prio =
+			guc_retire_inflight_request_prio;
+		tasklet_setup(&guc->sched_engine->tasklet,
+			      guc_submission_tasklet);
+	}
+	i915_sched_engine_put(engine->sched_engine);
+	engine->sched_engine = i915_sched_engine_get(guc->sched_engine);
+
+	guc_default_vfuncs(engine);
+	guc_default_irqs(engine);
+	guc_init_breadcrumbs(engine);
+
+	if (engine->flags & I915_ENGINE_HAS_RCS_REG_STATE)
+		rcs_submission_override(engine);
+
+	lrc_init_wa_ctx(engine);
+
+	/* Finally, take ownership and responsibility for cleanup! */
+	engine->sanitize = guc_sanitize;
+	engine->release = guc_release;
+
+	return 0;
+}
+
+void intel_guc_submission_enable(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+
+	/* Enable and route to GuC */
+	if (GRAPHICS_VER(gt->i915) >= 12)
+		intel_uncore_write(gt->uncore, GEN12_GUC_SEM_INTR_ENABLES,
+				   GUC_SEM_INTR_ROUTE_TO_GUC |
+				   GUC_SEM_INTR_ENABLE_ALL);
+
+	guc_init_lrc_mapping(guc);
+	guc_init_engine_stats(guc);
+}
+
+void intel_guc_submission_disable(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+
+	/* Note: By the time we're here, GuC may have already been reset */
+
+	/* Disable and route to host */
+	if (GRAPHICS_VER(gt->i915) >= 12)
+		intel_uncore_write(gt->uncore, GEN12_GUC_SEM_INTR_ENABLES, 0x0);
+}
+
+static bool __guc_submission_supported(struct intel_guc *guc)
+{
+	/* GuC submission is unavailable for pre-Gen11 */
+	return intel_guc_is_supported(guc) &&
+	       GRAPHICS_VER(guc_to_gt(guc)->i915) >= 11;
+}
+
+static bool __guc_submission_selected(struct intel_guc *guc)
+{
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+
+	if (!intel_guc_submission_is_supported(guc))
+		return false;
+
+	return i915->params.enable_guc & ENABLE_GUC_SUBMISSION;
+}
+
+void intel_guc_submission_init_early(struct intel_guc *guc)
+{
+	xa_init_flags(&guc->context_lookup, XA_FLAGS_LOCK_IRQ);
+
+	spin_lock_init(&guc->submission_state.lock);
+	INIT_LIST_HEAD(&guc->submission_state.guc_id_list);
+	ida_init(&guc->submission_state.guc_ids);
+	INIT_LIST_HEAD(&guc->submission_state.destroyed_contexts);
+	INIT_WORK(&guc->submission_state.destroyed_worker,
+		  destroyed_worker_func);
+	INIT_WORK(&guc->submission_state.reset_fail_worker,
+		  reset_fail_worker_func);
+
+	spin_lock_init(&guc->timestamp.lock);
+	INIT_DELAYED_WORK(&guc->timestamp.work, guc_timestamp_ping);
+
+	guc->submission_state.num_guc_ids = GUC_MAX_CONTEXT_ID;
+	guc->submission_supported = __guc_submission_supported(guc);
+	guc->submission_selected = __guc_submission_selected(guc);
+}
+
+static inline struct intel_context *
+g2h_context_lookup(struct intel_guc *guc, u32 ctx_id)
+{
+	struct intel_context *ce;
+
+	if (unlikely(ctx_id >= GUC_MAX_CONTEXT_ID)) {
+		drm_err(&guc_to_gt(guc)->i915->drm,
+			"Invalid ctx_id %u\n", ctx_id);
+		return NULL;
+	}
+
+	ce = __get_context(guc, ctx_id);
+	if (unlikely(!ce)) {
+		drm_err(&guc_to_gt(guc)->i915->drm,
+			"Context is NULL, ctx_id %u\n", ctx_id);
+		return NULL;
+	}
+
+	if (unlikely(intel_context_is_child(ce))) {
+		drm_err(&guc_to_gt(guc)->i915->drm,
+			"Context is child, ctx_id %u\n", ctx_id);
+		return NULL;
+	}
+
+	return ce;
+}
+
+int intel_guc_deregister_done_process_msg(struct intel_guc *guc,
+					  const u32 *msg,
+					  u32 len)
+{
+	struct intel_context *ce;
+	u32 ctx_id;
+
+	if (unlikely(len < 1)) {
+		drm_err(&guc_to_gt(guc)->i915->drm, "Invalid length %u\n", len);
+		return -EPROTO;
+	}
+	ctx_id = msg[0];
+
+	ce = g2h_context_lookup(guc, ctx_id);
+	if (unlikely(!ce))
+		return -EPROTO;
+
+	trace_intel_context_deregister_done(ce);
+
+#ifdef CONFIG_DRM_I915_SELFTEST
+	if (unlikely(ce->drop_deregister)) {
+		ce->drop_deregister = false;
+		return 0;
+	}
+#endif
+
+	if (context_wait_for_deregister_to_register(ce)) {
+		struct intel_runtime_pm *runtime_pm =
+			&ce->engine->gt->i915->runtime_pm;
+		intel_wakeref_t wakeref;
+
+		/*
+		 * Previous owner of this guc_id has been deregistered, now safe
+		 * register this context.
+		 */
+		with_intel_runtime_pm(runtime_pm, wakeref)
+			register_context(ce, true);
+		guc_signal_context_fence(ce);
+		intel_context_put(ce);
+	} else if (context_destroyed(ce)) {
+		/* Context has been destroyed */
+		intel_gt_pm_put_async(guc_to_gt(guc));
+		release_guc_id(guc, ce);
+		__guc_context_destroy(ce);
+	}
+
+	decr_outstanding_submission_g2h(guc);
+
+	return 0;
+}
+
+int intel_guc_sched_done_process_msg(struct intel_guc *guc,
+				     const u32 *msg,
+				     u32 len)
+{
+	struct intel_context *ce;
+	unsigned long flags;
+	u32 ctx_id;
+
+	if (unlikely(len < 2)) {
+		drm_err(&guc_to_gt(guc)->i915->drm, "Invalid length %u\n", len);
+		return -EPROTO;
+	}
+	ctx_id = msg[0];
+
+	ce = g2h_context_lookup(guc, ctx_id);
+	if (unlikely(!ce))
+		return -EPROTO;
+
+	if (unlikely(context_destroyed(ce) ||
+		     (!context_pending_enable(ce) &&
+		     !context_pending_disable(ce)))) {
+		drm_err(&guc_to_gt(guc)->i915->drm,
+			"Bad context sched_state 0x%x, ctx_id %u\n",
+			ce->guc_state.sched_state, ctx_id);
+		return -EPROTO;
+	}
+
+	trace_intel_context_sched_done(ce);
+
+	if (context_pending_enable(ce)) {
+#ifdef CONFIG_DRM_I915_SELFTEST
+		if (unlikely(ce->drop_schedule_enable)) {
+			ce->drop_schedule_enable = false;
+			return 0;
+		}
+#endif
+
+		spin_lock_irqsave(&ce->guc_state.lock, flags);
+		clr_context_pending_enable(ce);
+		spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+	} else if (context_pending_disable(ce)) {
+		bool banned;
+
+#ifdef CONFIG_DRM_I915_SELFTEST
+		if (unlikely(ce->drop_schedule_disable)) {
+			ce->drop_schedule_disable = false;
+			return 0;
+		}
+#endif
+
+		/*
+		 * Unpin must be done before __guc_signal_context_fence,
+		 * otherwise a race exists between the requests getting
+		 * submitted + retired before this unpin completes resulting in
+		 * the pin_count going to zero and the context still being
+		 * enabled.
+		 */
+		intel_context_sched_disable_unpin(ce);
+
+		spin_lock_irqsave(&ce->guc_state.lock, flags);
+		banned = context_banned(ce);
+		clr_context_banned(ce);
+		clr_context_pending_disable(ce);
+		__guc_signal_context_fence(ce);
+		guc_blocked_fence_complete(ce);
+		spin_unlock_irqrestore(&ce->guc_state.lock, flags);
+
+		if (banned) {
+			guc_cancel_context_requests(ce);
+			intel_engine_signal_breadcrumbs(ce->engine);
+		}
+	}
+
+	decr_outstanding_submission_g2h(guc);
+	intel_context_put(ce);
+
+	return 0;
+}
+
+static void capture_error_state(struct intel_guc *guc,
+				struct intel_context *ce)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_engine_cs *engine = __context_to_physical_engine(ce);
+	intel_wakeref_t wakeref;
+
+	intel_engine_set_hung_context(engine, ce);
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
+		i915_capture_error_state(gt, engine->mask, CORE_DUMP_FLAG_IS_GUC_CAPTURE);
+	atomic_inc(&i915->gpu_error.reset_engine_count[engine->uabi_class]);
+}
+
+static void guc_context_replay(struct intel_context *ce)
+{
+	struct i915_sched_engine *sched_engine = ce->engine->sched_engine;
+
+	__guc_reset_context(ce, ce->engine->mask);
+	tasklet_hi_schedule(&sched_engine->tasklet);
+}
+
+static void guc_handle_context_reset(struct intel_guc *guc,
+				     struct intel_context *ce)
+{
+	trace_intel_context_reset(ce);
+
+	if (likely(intel_context_is_schedulable(ce))) {
+		capture_error_state(guc, ce);
+		guc_context_replay(ce);
+	} else {
+		drm_info(&guc_to_gt(guc)->i915->drm,
+			 "Ignoring context reset notification of exiting context 0x%04X on %s",
+			 ce->guc_id.id, ce->engine->name);
+	}
+}
+
+int intel_guc_context_reset_process_msg(struct intel_guc *guc,
+					const u32 *msg, u32 len)
+{
+	struct intel_context *ce;
+	unsigned long flags;
+	int ctx_id;
+
+	if (unlikely(len != 1)) {
+		drm_err(&guc_to_gt(guc)->i915->drm, "Invalid length %u", len);
+		return -EPROTO;
+	}
+
+	ctx_id = msg[0];
+
+	/*
+	 * The context lookup uses the xarray but lookups only require an RCU lock
+	 * not the full spinlock. So take the lock explicitly and keep it until the
+	 * context has been reference count locked to ensure it can't be destroyed
+	 * asynchronously until the reset is done.
+	 */
+	xa_lock_irqsave(&guc->context_lookup, flags);
+	ce = g2h_context_lookup(guc, ctx_id);
+	if (ce)
+		intel_context_get(ce);
+	xa_unlock_irqrestore(&guc->context_lookup, flags);
+
+	if (unlikely(!ce))
+		return -EPROTO;
+
+	guc_handle_context_reset(guc, ce);
+	intel_context_put(ce);
+
+	return 0;
+}
+
+int intel_guc_error_capture_process_msg(struct intel_guc *guc,
+					const u32 *msg, u32 len)
+{
+	u32 status;
+
+	if (unlikely(len != 1)) {
+		drm_dbg(&guc_to_gt(guc)->i915->drm, "Invalid length %u", len);
+		return -EPROTO;
+	}
+
+	status = msg[0] & INTEL_GUC_STATE_CAPTURE_EVENT_STATUS_MASK;
+	if (status == INTEL_GUC_STATE_CAPTURE_EVENT_STATUS_NOSPACE)
+		drm_warn(&guc_to_gt(guc)->i915->drm, "G2H-Error capture no space");
+
+	intel_guc_capture_process(guc);
+
+	return 0;
+}
+
+struct intel_engine_cs *
+intel_guc_lookup_engine(struct intel_guc *guc, u8 guc_class, u8 instance)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	u8 engine_class = guc_class_to_engine_class(guc_class);
+
+	/* Class index is checked in class converter */
+	GEM_BUG_ON(instance > MAX_ENGINE_INSTANCE);
+
+	return gt->engine_class[engine_class][instance];
+}
+
+static void reset_fail_worker_func(struct work_struct *w)
+{
+	struct intel_guc *guc = container_of(w, struct intel_guc,
+					     submission_state.reset_fail_worker);
+	struct intel_gt *gt = guc_to_gt(guc);
+	intel_engine_mask_t reset_fail_mask;
+	unsigned long flags;
+
+	spin_lock_irqsave(&guc->submission_state.lock, flags);
+	reset_fail_mask = guc->submission_state.reset_fail_mask;
+	guc->submission_state.reset_fail_mask = 0;
+	spin_unlock_irqrestore(&guc->submission_state.lock, flags);
+
+	if (likely(reset_fail_mask))
+		intel_gt_handle_error(gt, reset_fail_mask,
+				      I915_ERROR_CAPTURE,
+				      "GuC failed to reset engine mask=0x%x\n",
+				      reset_fail_mask);
+}
+
+int intel_guc_engine_failure_process_msg(struct intel_guc *guc,
+					 const u32 *msg, u32 len)
+{
+	struct intel_engine_cs *engine;
+	struct intel_gt *gt = guc_to_gt(guc);
+	u8 guc_class, instance;
+	u32 reason;
+	unsigned long flags;
+
+	if (unlikely(len != 3)) {
+		drm_err(&gt->i915->drm, "Invalid length %u", len);
+		return -EPROTO;
+	}
+
+	guc_class = msg[0];
+	instance = msg[1];
+	reason = msg[2];
+
+	engine = intel_guc_lookup_engine(guc, guc_class, instance);
+	if (unlikely(!engine)) {
+		drm_err(&gt->i915->drm,
+			"Invalid engine %d:%d", guc_class, instance);
+		return -EPROTO;
+	}
+
+	/*
+	 * This is an unexpected failure of a hardware feature. So, log a real
+	 * error message not just the informational that comes with the reset.
+	 */
+	drm_err(&gt->i915->drm, "GuC engine reset request failed on %d:%d (%s) because 0x%08X",
+		guc_class, instance, engine->name, reason);
+
+	spin_lock_irqsave(&guc->submission_state.lock, flags);
+	guc->submission_state.reset_fail_mask |= engine->mask;
+	spin_unlock_irqrestore(&guc->submission_state.lock, flags);
+
+	/*
+	 * A GT reset flushes this worker queue (G2H handler) so we must use
+	 * another worker to trigger a GT reset.
+	 */
+	queue_work(system_unbound_wq, &guc->submission_state.reset_fail_worker);
+
+	return 0;
+}
+
+void intel_guc_find_hung_context(struct intel_engine_cs *engine)
+{
+	struct intel_guc *guc = &engine->gt->uc.guc;
+	struct intel_context *ce;
+	struct i915_request *rq;
+	unsigned long index;
+	unsigned long flags;
+
+	/* Reset called during driver load? GuC not yet initialised! */
+	if (unlikely(!guc_submission_initialized(guc)))
+		return;
+
+	xa_lock_irqsave(&guc->context_lookup, flags);
+	xa_for_each(&guc->context_lookup, index, ce) {
+		bool found;
+
+		if (!kref_get_unless_zero(&ce->ref))
+			continue;
+
+		xa_unlock(&guc->context_lookup);
+
+		if (!intel_context_is_pinned(ce))
+			goto next;
+
+		if (intel_engine_is_virtual(ce->engine)) {
+			if (!(ce->engine->mask & engine->mask))
+				goto next;
+		} else {
+			if (ce->engine != engine)
+				goto next;
+		}
+
+		found = false;
+		spin_lock(&ce->guc_state.lock);
+		list_for_each_entry(rq, &ce->guc_state.requests, sched.link) {
+			if (i915_test_request_state(rq) != I915_REQUEST_ACTIVE)
+				continue;
+
+			found = true;
+			break;
+		}
+		spin_unlock(&ce->guc_state.lock);
+
+		if (found) {
+			intel_engine_set_hung_context(engine, ce);
+
+			/* Can only cope with one hang at a time... */
+			intel_context_put(ce);
+			xa_lock(&guc->context_lookup);
+			goto done;
+		}
+
+next:
+		intel_context_put(ce);
+		xa_lock(&guc->context_lookup);
+	}
+done:
+	xa_unlock_irqrestore(&guc->context_lookup, flags);
+}
+
+void intel_guc_dump_active_requests(struct intel_engine_cs *engine,
+				    struct i915_request *hung_rq,
+				    struct drm_printer *m)
+{
+	struct intel_guc *guc = &engine->gt->uc.guc;
+	struct intel_context *ce;
+	unsigned long index;
+	unsigned long flags;
+
+	/* Reset called during driver load? GuC not yet initialised! */
+	if (unlikely(!guc_submission_initialized(guc)))
+		return;
+
+	xa_lock_irqsave(&guc->context_lookup, flags);
+	xa_for_each(&guc->context_lookup, index, ce) {
+		if (!kref_get_unless_zero(&ce->ref))
+			continue;
+
+		xa_unlock(&guc->context_lookup);
+
+		if (!intel_context_is_pinned(ce))
+			goto next;
+
+		if (intel_engine_is_virtual(ce->engine)) {
+			if (!(ce->engine->mask & engine->mask))
+				goto next;
+		} else {
+			if (ce->engine != engine)
+				goto next;
+		}
+
+		spin_lock(&ce->guc_state.lock);
+		intel_engine_dump_active_requests(&ce->guc_state.requests,
+						  hung_rq, m);
+		spin_unlock(&ce->guc_state.lock);
+
+next:
+		intel_context_put(ce);
+		xa_lock(&guc->context_lookup);
+	}
+	xa_unlock_irqrestore(&guc->context_lookup, flags);
+}
+
+void intel_guc_submission_print_info(struct intel_guc *guc,
+				     struct drm_printer *p)
+{
+	struct i915_sched_engine *sched_engine = guc->sched_engine;
+	struct rb_node *rb;
+	unsigned long flags;
+
+	if (!sched_engine)
+		return;
+
+	drm_printf(p, "GuC Number Outstanding Submission G2H: %u\n",
+		   atomic_read(&guc->outstanding_submission_g2h));
+	drm_printf(p, "GuC tasklet count: %u\n\n",
+		   atomic_read(&sched_engine->tasklet.count));
+
+	spin_lock_irqsave(&sched_engine->lock, flags);
+	drm_printf(p, "Requests in GuC submit tasklet:\n");
+	for (rb = rb_first_cached(&sched_engine->queue); rb; rb = rb_next(rb)) {
+		struct i915_priolist *pl = to_priolist(rb);
+		struct i915_request *rq;
+
+		priolist_for_each_request(rq, pl)
+			drm_printf(p, "guc_id=%u, seqno=%llu\n",
+				   rq->context->guc_id.id,
+				   rq->fence.seqno);
+	}
+	spin_unlock_irqrestore(&sched_engine->lock, flags);
+	drm_printf(p, "\n");
+}
+
+static inline void guc_log_context_priority(struct drm_printer *p,
+					    struct intel_context *ce)
+{
+	int i;
+
+	drm_printf(p, "\t\tPriority: %d\n", ce->guc_state.prio);
+	drm_printf(p, "\t\tNumber Requests (lower index == higher priority)\n");
+	for (i = GUC_CLIENT_PRIORITY_KMD_HIGH;
+	     i < GUC_CLIENT_PRIORITY_NUM; ++i) {
+		drm_printf(p, "\t\tNumber requests in priority band[%d]: %d\n",
+			   i, ce->guc_state.prio_count[i]);
+	}
+	drm_printf(p, "\n");
+}
+
+static inline void guc_log_context(struct drm_printer *p,
+				   struct intel_context *ce)
+{
+	drm_printf(p, "GuC lrc descriptor %u:\n", ce->guc_id.id);
+	drm_printf(p, "\tHW Context Desc: 0x%08x\n", ce->lrc.lrca);
+	drm_printf(p, "\t\tLRC Head: Internal %u, Memory %u\n",
+		   ce->ring->head,
+		   ce->lrc_reg_state[CTX_RING_HEAD]);
+	drm_printf(p, "\t\tLRC Tail: Internal %u, Memory %u\n",
+		   ce->ring->tail,
+		   ce->lrc_reg_state[CTX_RING_TAIL]);
+	drm_printf(p, "\t\tContext Pin Count: %u\n",
+		   atomic_read(&ce->pin_count));
+	drm_printf(p, "\t\tGuC ID Ref Count: %u\n",
+		   atomic_read(&ce->guc_id.ref));
+	drm_printf(p, "\t\tSchedule State: 0x%x\n\n",
+		   ce->guc_state.sched_state);
+}
+
+void intel_guc_submission_print_context_info(struct intel_guc *guc,
+					     struct drm_printer *p)
+{
+	struct intel_context *ce;
+	unsigned long index;
+	unsigned long flags;
+
+	xa_lock_irqsave(&guc->context_lookup, flags);
+	xa_for_each(&guc->context_lookup, index, ce) {
+		GEM_BUG_ON(intel_context_is_child(ce));
+
+		guc_log_context(p, ce);
+		guc_log_context_priority(p, ce);
+
+		if (intel_context_is_parent(ce)) {
+			struct intel_context *child;
+
+			drm_printf(p, "\t\tNumber children: %u\n",
+				   ce->parallel.number_children);
+
+			if (ce->parallel.guc.wq_status) {
+				drm_printf(p, "\t\tWQI Head: %u\n",
+					   READ_ONCE(*ce->parallel.guc.wq_head));
+				drm_printf(p, "\t\tWQI Tail: %u\n",
+					   READ_ONCE(*ce->parallel.guc.wq_tail));
+				drm_printf(p, "\t\tWQI Status: %u\n\n",
+					   READ_ONCE(*ce->parallel.guc.wq_status));
+			}
+
+			if (ce->engine->emit_bb_start ==
+			    emit_bb_start_parent_no_preempt_mid_batch) {
+				u8 i;
+
+				drm_printf(p, "\t\tChildren Go: %u\n\n",
+					   get_children_go_value(ce));
+				for (i = 0; i < ce->parallel.number_children; ++i)
+					drm_printf(p, "\t\tChildren Join: %u\n",
+						   get_children_join_value(ce, i));
+			}
+
+			for_each_child(ce, child)
+				guc_log_context(p, child);
+		}
+	}
+	xa_unlock_irqrestore(&guc->context_lookup, flags);
+}
+
+static inline u32 get_children_go_addr(struct intel_context *ce)
+{
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+
+	return i915_ggtt_offset(ce->state) +
+		__get_parent_scratch_offset(ce) +
+		offsetof(struct parent_scratch, go.semaphore);
+}
+
+static inline u32 get_children_join_addr(struct intel_context *ce,
+					 u8 child_index)
+{
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+
+	return i915_ggtt_offset(ce->state) +
+		__get_parent_scratch_offset(ce) +
+		offsetof(struct parent_scratch, join[child_index].semaphore);
+}
+
+#define PARENT_GO_BB			1
+#define PARENT_GO_FINI_BREADCRUMB	0
+#define CHILD_GO_BB			1
+#define CHILD_GO_FINI_BREADCRUMB	0
+static int emit_bb_start_parent_no_preempt_mid_batch(struct i915_request *rq,
+						     u64 offset, u32 len,
+						     const unsigned int flags)
+{
+	struct intel_context *ce = rq->context;
+	u32 *cs;
+	u8 i;
+
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+
+	cs = intel_ring_begin(rq, 10 + 4 * ce->parallel.number_children);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/* Wait on children */
+	for (i = 0; i < ce->parallel.number_children; ++i) {
+		*cs++ = (MI_SEMAPHORE_WAIT |
+			 MI_SEMAPHORE_GLOBAL_GTT |
+			 MI_SEMAPHORE_POLL |
+			 MI_SEMAPHORE_SAD_EQ_SDD);
+		*cs++ = PARENT_GO_BB;
+		*cs++ = get_children_join_addr(ce, i);
+		*cs++ = 0;
+	}
+
+	/* Turn off preemption */
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+	*cs++ = MI_NOOP;
+
+	/* Tell children go */
+	cs = gen8_emit_ggtt_write(cs,
+				  CHILD_GO_BB,
+				  get_children_go_addr(ce),
+				  0);
+
+	/* Jump to batch */
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 |
+		(flags & I915_DISPATCH_SECURE ? 0 : BIT(8));
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+	*cs++ = MI_NOOP;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int emit_bb_start_child_no_preempt_mid_batch(struct i915_request *rq,
+						    u64 offset, u32 len,
+						    const unsigned int flags)
+{
+	struct intel_context *ce = rq->context;
+	struct intel_context *parent = intel_context_to_parent(ce);
+	u32 *cs;
+
+	GEM_BUG_ON(!intel_context_is_child(ce));
+
+	cs = intel_ring_begin(rq, 12);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/* Signal parent */
+	cs = gen8_emit_ggtt_write(cs,
+				  PARENT_GO_BB,
+				  get_children_join_addr(parent,
+							 ce->parallel.child_index),
+				  0);
+
+	/* Wait on parent for go */
+	*cs++ = (MI_SEMAPHORE_WAIT |
+		 MI_SEMAPHORE_GLOBAL_GTT |
+		 MI_SEMAPHORE_POLL |
+		 MI_SEMAPHORE_SAD_EQ_SDD);
+	*cs++ = CHILD_GO_BB;
+	*cs++ = get_children_go_addr(parent);
+	*cs++ = 0;
+
+	/* Turn off preemption */
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+
+	/* Jump to batch */
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 |
+		(flags & I915_DISPATCH_SECURE ? 0 : BIT(8));
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static u32 *
+__emit_fini_breadcrumb_parent_no_preempt_mid_batch(struct i915_request *rq,
+						   u32 *cs)
+{
+	struct intel_context *ce = rq->context;
+	u8 i;
+
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+
+	/* Wait on children */
+	for (i = 0; i < ce->parallel.number_children; ++i) {
+		*cs++ = (MI_SEMAPHORE_WAIT |
+			 MI_SEMAPHORE_GLOBAL_GTT |
+			 MI_SEMAPHORE_POLL |
+			 MI_SEMAPHORE_SAD_EQ_SDD);
+		*cs++ = PARENT_GO_FINI_BREADCRUMB;
+		*cs++ = get_children_join_addr(ce, i);
+		*cs++ = 0;
+	}
+
+	/* Turn on preemption */
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	*cs++ = MI_NOOP;
+
+	/* Tell children go */
+	cs = gen8_emit_ggtt_write(cs,
+				  CHILD_GO_FINI_BREADCRUMB,
+				  get_children_go_addr(ce),
+				  0);
+
+	return cs;
+}
+
+/*
+ * If this true, a submission of multi-lrc requests had an error and the
+ * requests need to be skipped. The front end (execuf IOCTL) should've called
+ * i915_request_skip which squashes the BB but we still need to emit the fini
+ * breadrcrumbs seqno write. At this point we don't know how many of the
+ * requests in the multi-lrc submission were generated so we can't do the
+ * handshake between the parent and children (e.g. if 4 requests should be
+ * generated but 2nd hit an error only 1 would be seen by the GuC backend).
+ * Simply skip the handshake, but still emit the breadcrumbd seqno, if an error
+ * has occurred on any of the requests in submission / relationship.
+ */
+static inline bool skip_handshake(struct i915_request *rq)
+{
+	return test_bit(I915_FENCE_FLAG_SKIP_PARALLEL, &rq->fence.flags);
+}
+
+#define NON_SKIP_LEN	6
+static u32 *
+emit_fini_breadcrumb_parent_no_preempt_mid_batch(struct i915_request *rq,
+						 u32 *cs)
+{
+	struct intel_context *ce = rq->context;
+	__maybe_unused u32 *before_fini_breadcrumb_user_interrupt_cs;
+	__maybe_unused u32 *start_fini_breadcrumb_cs = cs;
+
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+
+	if (unlikely(skip_handshake(rq))) {
+		/*
+		 * NOP everything in __emit_fini_breadcrumb_parent_no_preempt_mid_batch,
+		 * the NON_SKIP_LEN comes from the length of the emits below.
+		 */
+		memset(cs, 0, sizeof(u32) *
+		       (ce->engine->emit_fini_breadcrumb_dw - NON_SKIP_LEN));
+		cs += ce->engine->emit_fini_breadcrumb_dw - NON_SKIP_LEN;
+	} else {
+		cs = __emit_fini_breadcrumb_parent_no_preempt_mid_batch(rq, cs);
+	}
+
+	/* Emit fini breadcrumb */
+	before_fini_breadcrumb_user_interrupt_cs = cs;
+	cs = gen8_emit_ggtt_write(cs,
+				  rq->fence.seqno,
+				  i915_request_active_timeline(rq)->hwsp_offset,
+				  0);
+
+	/* User interrupt */
+	*cs++ = MI_USER_INTERRUPT;
+	*cs++ = MI_NOOP;
+
+	/* Ensure our math for skip + emit is correct */
+	GEM_BUG_ON(before_fini_breadcrumb_user_interrupt_cs + NON_SKIP_LEN !=
+		   cs);
+	GEM_BUG_ON(start_fini_breadcrumb_cs +
+		   ce->engine->emit_fini_breadcrumb_dw != cs);
+
+	rq->tail = intel_ring_offset(rq, cs);
+
+	return cs;
+}
+
+static u32 *
+__emit_fini_breadcrumb_child_no_preempt_mid_batch(struct i915_request *rq,
+						  u32 *cs)
+{
+	struct intel_context *ce = rq->context;
+	struct intel_context *parent = intel_context_to_parent(ce);
+
+	GEM_BUG_ON(!intel_context_is_child(ce));
+
+	/* Turn on preemption */
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	*cs++ = MI_NOOP;
+
+	/* Signal parent */
+	cs = gen8_emit_ggtt_write(cs,
+				  PARENT_GO_FINI_BREADCRUMB,
+				  get_children_join_addr(parent,
+							 ce->parallel.child_index),
+				  0);
+
+	/* Wait parent on for go */
+	*cs++ = (MI_SEMAPHORE_WAIT |
+		 MI_SEMAPHORE_GLOBAL_GTT |
+		 MI_SEMAPHORE_POLL |
+		 MI_SEMAPHORE_SAD_EQ_SDD);
+	*cs++ = CHILD_GO_FINI_BREADCRUMB;
+	*cs++ = get_children_go_addr(parent);
+	*cs++ = 0;
+
+	return cs;
+}
+
+static u32 *
+emit_fini_breadcrumb_child_no_preempt_mid_batch(struct i915_request *rq,
+						u32 *cs)
+{
+	struct intel_context *ce = rq->context;
+	__maybe_unused u32 *before_fini_breadcrumb_user_interrupt_cs;
+	__maybe_unused u32 *start_fini_breadcrumb_cs = cs;
+
+	GEM_BUG_ON(!intel_context_is_child(ce));
+
+	if (unlikely(skip_handshake(rq))) {
+		/*
+		 * NOP everything in __emit_fini_breadcrumb_child_no_preempt_mid_batch,
+		 * the NON_SKIP_LEN comes from the length of the emits below.
+		 */
+		memset(cs, 0, sizeof(u32) *
+		       (ce->engine->emit_fini_breadcrumb_dw - NON_SKIP_LEN));
+		cs += ce->engine->emit_fini_breadcrumb_dw - NON_SKIP_LEN;
+	} else {
+		cs = __emit_fini_breadcrumb_child_no_preempt_mid_batch(rq, cs);
+	}
+
+	/* Emit fini breadcrumb */
+	before_fini_breadcrumb_user_interrupt_cs = cs;
+	cs = gen8_emit_ggtt_write(cs,
+				  rq->fence.seqno,
+				  i915_request_active_timeline(rq)->hwsp_offset,
+				  0);
+
+	/* User interrupt */
+	*cs++ = MI_USER_INTERRUPT;
+	*cs++ = MI_NOOP;
+
+	/* Ensure our math for skip + emit is correct */
+	GEM_BUG_ON(before_fini_breadcrumb_user_interrupt_cs + NON_SKIP_LEN !=
+		   cs);
+	GEM_BUG_ON(start_fini_breadcrumb_cs +
+		   ce->engine->emit_fini_breadcrumb_dw != cs);
+
+	rq->tail = intel_ring_offset(rq, cs);
+
+	return cs;
+}
+
+#undef NON_SKIP_LEN
+
+static struct intel_context *
+guc_create_virtual(struct intel_engine_cs **siblings, unsigned int count,
+		   unsigned long flags)
+{
+	struct guc_virtual_engine *ve;
+	struct intel_guc *guc;
+	unsigned int n;
+	int err;
+
+	ve = kzalloc(sizeof(*ve), GFP_KERNEL);
+	if (!ve)
+		return ERR_PTR(-ENOMEM);
+
+	guc = &siblings[0]->gt->uc.guc;
+
+	ve->base.i915 = siblings[0]->i915;
+	ve->base.gt = siblings[0]->gt;
+	ve->base.uncore = siblings[0]->uncore;
+	ve->base.id = -1;
+
+	ve->base.uabi_class = I915_ENGINE_CLASS_INVALID;
+	ve->base.instance = I915_ENGINE_CLASS_INVALID_VIRTUAL;
+	ve->base.uabi_instance = I915_ENGINE_CLASS_INVALID_VIRTUAL;
+	ve->base.saturated = ALL_ENGINES;
+
+	snprintf(ve->base.name, sizeof(ve->base.name), "virtual");
+
+	ve->base.sched_engine = i915_sched_engine_get(guc->sched_engine);
+
+	ve->base.cops = &virtual_guc_context_ops;
+	ve->base.request_alloc = guc_request_alloc;
+	ve->base.bump_serial = virtual_guc_bump_serial;
+
+	ve->base.submit_request = guc_submit_request;
+
+	ve->base.flags = I915_ENGINE_IS_VIRTUAL;
+
+	BUILD_BUG_ON(ilog2(VIRTUAL_ENGINES) < I915_NUM_ENGINES);
+	ve->base.mask = VIRTUAL_ENGINES;
+
+	intel_context_init(&ve->context, &ve->base);
+
+	for (n = 0; n < count; n++) {
+		struct intel_engine_cs *sibling = siblings[n];
+
+		GEM_BUG_ON(!is_power_of_2(sibling->mask));
+		if (sibling->mask & ve->base.mask) {
+			DRM_DEBUG("duplicate %s entry in load balancer\n",
+				  sibling->name);
+			err = -EINVAL;
+			goto err_put;
+		}
+
+		ve->base.mask |= sibling->mask;
+		ve->base.logical_mask |= sibling->logical_mask;
+
+		if (n != 0 && ve->base.class != sibling->class) {
+			DRM_DEBUG("invalid mixing of engine class, sibling %d, already %d\n",
+				  sibling->class, ve->base.class);
+			err = -EINVAL;
+			goto err_put;
+		} else if (n == 0) {
+			ve->base.class = sibling->class;
+			ve->base.uabi_class = sibling->uabi_class;
+			snprintf(ve->base.name, sizeof(ve->base.name),
+				 "v%dx%d", ve->base.class, count);
+			ve->base.context_size = sibling->context_size;
+
+			ve->base.add_active_request =
+				sibling->add_active_request;
+			ve->base.remove_active_request =
+				sibling->remove_active_request;
+			ve->base.emit_bb_start = sibling->emit_bb_start;
+			ve->base.emit_flush = sibling->emit_flush;
+			ve->base.emit_init_breadcrumb =
+				sibling->emit_init_breadcrumb;
+			ve->base.emit_fini_breadcrumb =
+				sibling->emit_fini_breadcrumb;
+			ve->base.emit_fini_breadcrumb_dw =
+				sibling->emit_fini_breadcrumb_dw;
+			ve->base.breadcrumbs =
+				intel_breadcrumbs_get(sibling->breadcrumbs);
+
+			ve->base.flags |= sibling->flags;
+
+			ve->base.props.timeslice_duration_ms =
+				sibling->props.timeslice_duration_ms;
+			ve->base.props.preempt_timeout_ms =
+				sibling->props.preempt_timeout_ms;
+		}
+	}
+
+	return &ve->context;
+
+err_put:
+	intel_context_put(&ve->context);
+	return ERR_PTR(err);
+}
+
+bool intel_guc_virtual_engine_has_heartbeat(const struct intel_engine_cs *ve)
+{
+	struct intel_engine_cs *engine;
+	intel_engine_mask_t tmp, mask = ve->mask;
+
+	for_each_engine_masked(engine, ve->gt, mask, tmp)
+		if (READ_ONCE(engine->props.heartbeat_interval_ms))
+			return true;
+
+	return false;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_guc.c"
+#include "selftest_guc_multi_lrc.c"
+#include "selftest_guc_hangcheck.c"
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h
new file mode 100644
index 000000000..5a95a9f0a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.h
@@ -0,0 +1,56 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_GUC_SUBMISSION_H_
+#define _INTEL_GUC_SUBMISSION_H_
+
+#include <linux/types.h>
+
+#include "intel_guc.h"
+
+struct drm_printer;
+struct intel_engine_cs;
+
+void intel_guc_submission_init_early(struct intel_guc *guc);
+int intel_guc_submission_init(struct intel_guc *guc);
+void intel_guc_submission_enable(struct intel_guc *guc);
+void intel_guc_submission_disable(struct intel_guc *guc);
+void intel_guc_submission_fini(struct intel_guc *guc);
+int intel_guc_preempt_work_create(struct intel_guc *guc);
+void intel_guc_preempt_work_destroy(struct intel_guc *guc);
+int intel_guc_submission_setup(struct intel_engine_cs *engine);
+void intel_guc_submission_print_info(struct intel_guc *guc,
+				     struct drm_printer *p);
+void intel_guc_submission_print_context_info(struct intel_guc *guc,
+					     struct drm_printer *p);
+void intel_guc_dump_active_requests(struct intel_engine_cs *engine,
+				    struct i915_request *hung_rq,
+				    struct drm_printer *m);
+void intel_guc_busyness_park(struct intel_gt *gt);
+void intel_guc_busyness_unpark(struct intel_gt *gt);
+
+bool intel_guc_virtual_engine_has_heartbeat(const struct intel_engine_cs *ve);
+
+int intel_guc_wait_for_pending_msg(struct intel_guc *guc,
+				   atomic_t *wait_var,
+				   bool interruptible,
+				   long timeout);
+
+static inline bool intel_guc_submission_is_supported(struct intel_guc *guc)
+{
+	return guc->submission_supported;
+}
+
+static inline bool intel_guc_submission_is_wanted(struct intel_guc *guc)
+{
+	return guc->submission_selected;
+}
+
+static inline bool intel_guc_submission_is_used(struct intel_guc *guc)
+{
+	return intel_guc_is_used(guc) && intel_guc_submission_is_wanted(guc);
+}
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_huc.c b/drivers/gpu/drm/i915/gt/uc/intel_huc.c
new file mode 100644
index 000000000..3bb8838e3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_huc.c
@@ -0,0 +1,266 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2016-2019 Intel Corporation
+ */
+
+#include <linux/types.h>
+
+#include "gt/intel_gt.h"
+#include "intel_guc_reg.h"
+#include "intel_huc.h"
+#include "i915_drv.h"
+
+/**
+ * DOC: HuC
+ *
+ * The HuC is a dedicated microcontroller for usage in media HEVC (High
+ * Efficiency Video Coding) operations. Userspace can directly use the firmware
+ * capabilities by adding HuC specific commands to batch buffers.
+ *
+ * The kernel driver is only responsible for loading the HuC firmware and
+ * triggering its security authentication, which is performed by the GuC on
+ * older platforms and by the GSC on newer ones. For the GuC to correctly
+ * perform the authentication, the HuC binary must be loaded before the GuC one.
+ * Loading the HuC is optional; however, not using the HuC might negatively
+ * impact power usage and/or performance of media workloads, depending on the
+ * use-cases.
+ * HuC must be reloaded on events that cause the WOPCM to lose its contents
+ * (S3/S4, FLR); GuC-authenticated HuC must also be reloaded on GuC/GT reset,
+ * while GSC-managed HuC will survive that.
+ *
+ * See https://github.com/intel/media-driver for the latest details on HuC
+ * functionality.
+ */
+
+/**
+ * DOC: HuC Memory Management
+ *
+ * Similarly to the GuC, the HuC can't do any memory allocations on its own,
+ * with the difference being that the allocations for HuC usage are handled by
+ * the userspace driver instead of the kernel one. The HuC accesses the memory
+ * via the PPGTT belonging to the context loaded on the VCS executing the
+ * HuC-specific commands.
+ */
+
+void intel_huc_init_early(struct intel_huc *huc)
+{
+	struct drm_i915_private *i915 = huc_to_gt(huc)->i915;
+
+	intel_uc_fw_init_early(&huc->fw, INTEL_UC_FW_TYPE_HUC);
+
+	if (GRAPHICS_VER(i915) >= 11) {
+		huc->status.reg = GEN11_HUC_KERNEL_LOAD_INFO;
+		huc->status.mask = HUC_LOAD_SUCCESSFUL;
+		huc->status.value = HUC_LOAD_SUCCESSFUL;
+	} else {
+		huc->status.reg = HUC_STATUS2;
+		huc->status.mask = HUC_FW_VERIFIED;
+		huc->status.value = HUC_FW_VERIFIED;
+	}
+}
+
+#define HUC_LOAD_MODE_STRING(x) (x ? "GSC" : "legacy")
+static int check_huc_loading_mode(struct intel_huc *huc)
+{
+	struct intel_gt *gt = huc_to_gt(huc);
+	bool fw_needs_gsc = intel_huc_is_loaded_by_gsc(huc);
+	bool hw_uses_gsc = false;
+
+	/*
+	 * The fuse for HuC load via GSC is only valid on platforms that have
+	 * GuC deprivilege.
+	 */
+	if (HAS_GUC_DEPRIVILEGE(gt->i915))
+		hw_uses_gsc = intel_uncore_read(gt->uncore, GUC_SHIM_CONTROL2) &
+			      GSC_LOADS_HUC;
+
+	if (fw_needs_gsc != hw_uses_gsc) {
+		drm_err(&gt->i915->drm,
+			"mismatch between HuC FW (%s) and HW (%s) load modes\n",
+			HUC_LOAD_MODE_STRING(fw_needs_gsc),
+			HUC_LOAD_MODE_STRING(hw_uses_gsc));
+		return -ENOEXEC;
+	}
+
+	/* make sure we can access the GSC via the mei driver if we need it */
+	if (!(IS_ENABLED(CONFIG_INTEL_MEI_PXP) && IS_ENABLED(CONFIG_INTEL_MEI_GSC)) &&
+	    fw_needs_gsc) {
+		drm_info(&gt->i915->drm,
+			 "Can't load HuC due to missing MEI modules\n");
+		return -EIO;
+	}
+
+	drm_dbg(&gt->i915->drm, "GSC loads huc=%s\n", str_yes_no(fw_needs_gsc));
+
+	return 0;
+}
+
+int intel_huc_init(struct intel_huc *huc)
+{
+	struct drm_i915_private *i915 = huc_to_gt(huc)->i915;
+	int err;
+
+	err = check_huc_loading_mode(huc);
+	if (err)
+		goto out;
+
+	err = intel_uc_fw_init(&huc->fw);
+	if (err)
+		goto out;
+
+	intel_uc_fw_change_status(&huc->fw, INTEL_UC_FIRMWARE_LOADABLE);
+
+	return 0;
+
+out:
+	drm_info(&i915->drm, "HuC init failed with %d\n", err);
+	return err;
+}
+
+void intel_huc_fini(struct intel_huc *huc)
+{
+	if (!intel_uc_fw_is_loadable(&huc->fw))
+		return;
+
+	intel_uc_fw_fini(&huc->fw);
+}
+
+/**
+ * intel_huc_auth() - Authenticate HuC uCode
+ * @huc: intel_huc structure
+ *
+ * Called after HuC and GuC firmware loading during intel_uc_init_hw().
+ *
+ * This function invokes the GuC action to authenticate the HuC firmware,
+ * passing the offset of the RSA signature to intel_guc_auth_huc(). It then
+ * waits for up to 50ms for firmware verification ACK.
+ */
+int intel_huc_auth(struct intel_huc *huc)
+{
+	struct intel_gt *gt = huc_to_gt(huc);
+	struct intel_guc *guc = &gt->uc.guc;
+	int ret;
+
+	if (!intel_uc_fw_is_loaded(&huc->fw))
+		return -ENOEXEC;
+
+	/* GSC will do the auth */
+	if (intel_huc_is_loaded_by_gsc(huc))
+		return -ENODEV;
+
+	ret = i915_inject_probe_error(gt->i915, -ENXIO);
+	if (ret)
+		goto fail;
+
+	GEM_BUG_ON(intel_uc_fw_is_running(&huc->fw));
+
+	ret = intel_guc_auth_huc(guc, intel_guc_ggtt_offset(guc, huc->fw.rsa_data));
+	if (ret) {
+		DRM_ERROR("HuC: GuC did not ack Auth request %d\n", ret);
+		goto fail;
+	}
+
+	/* Check authentication status, it should be done by now */
+	ret = __intel_wait_for_register(gt->uncore,
+					huc->status.reg,
+					huc->status.mask,
+					huc->status.value,
+					2, 50, NULL);
+	if (ret) {
+		DRM_ERROR("HuC: Firmware not verified %d\n", ret);
+		goto fail;
+	}
+
+	intel_uc_fw_change_status(&huc->fw, INTEL_UC_FIRMWARE_RUNNING);
+	drm_info(&gt->i915->drm, "HuC authenticated\n");
+	return 0;
+
+fail:
+	i915_probe_error(gt->i915, "HuC: Authentication failed %d\n", ret);
+	intel_uc_fw_change_status(&huc->fw, INTEL_UC_FIRMWARE_LOAD_FAIL);
+	return ret;
+}
+
+static bool huc_is_authenticated(struct intel_huc *huc)
+{
+	struct intel_gt *gt = huc_to_gt(huc);
+	intel_wakeref_t wakeref;
+	u32 status = 0;
+
+	with_intel_runtime_pm(gt->uncore->rpm, wakeref)
+		status = intel_uncore_read(gt->uncore, huc->status.reg);
+
+	return (status & huc->status.mask) == huc->status.value;
+}
+
+/**
+ * intel_huc_check_status() - check HuC status
+ * @huc: intel_huc structure
+ *
+ * This function reads status register to verify if HuC
+ * firmware was successfully loaded.
+ *
+ * Returns:
+ *  * -ENODEV if HuC is not present on this platform,
+ *  * -EOPNOTSUPP if HuC firmware is disabled,
+ *  * -ENOPKG if HuC firmware was not installed,
+ *  * -ENOEXEC if HuC firmware is invalid or mismatched,
+ *  * 0 if HuC firmware is not running,
+ *  * 1 if HuC firmware is authenticated and running.
+ */
+int intel_huc_check_status(struct intel_huc *huc)
+{
+	switch (__intel_uc_fw_status(&huc->fw)) {
+	case INTEL_UC_FIRMWARE_NOT_SUPPORTED:
+		return -ENODEV;
+	case INTEL_UC_FIRMWARE_DISABLED:
+		return -EOPNOTSUPP;
+	case INTEL_UC_FIRMWARE_MISSING:
+		return -ENOPKG;
+	case INTEL_UC_FIRMWARE_ERROR:
+		return -ENOEXEC;
+	default:
+		break;
+	}
+
+	return huc_is_authenticated(huc);
+}
+
+void intel_huc_update_auth_status(struct intel_huc *huc)
+{
+	if (!intel_uc_fw_is_loadable(&huc->fw))
+		return;
+
+	if (huc_is_authenticated(huc))
+		intel_uc_fw_change_status(&huc->fw,
+					  INTEL_UC_FIRMWARE_RUNNING);
+}
+
+/**
+ * intel_huc_load_status - dump information about HuC load status
+ * @huc: the HuC
+ * @p: the &drm_printer
+ *
+ * Pretty printer for HuC load status.
+ */
+void intel_huc_load_status(struct intel_huc *huc, struct drm_printer *p)
+{
+	struct intel_gt *gt = huc_to_gt(huc);
+	intel_wakeref_t wakeref;
+
+	if (!intel_huc_is_supported(huc)) {
+		drm_printf(p, "HuC not supported\n");
+		return;
+	}
+
+	if (!intel_huc_is_wanted(huc)) {
+		drm_printf(p, "HuC disabled\n");
+		return;
+	}
+
+	intel_uc_fw_dump(&huc->fw, p);
+
+	with_intel_runtime_pm(gt->uncore->rpm, wakeref)
+		drm_printf(p, "HuC status: 0x%08x\n",
+			   intel_uncore_read(gt->uncore, huc->status.reg));
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_huc.h b/drivers/gpu/drm/i915/gt/uc/intel_huc.h
new file mode 100644
index 000000000..d7e25b6e8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_huc.h
@@ -0,0 +1,61 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_HUC_H_
+#define _INTEL_HUC_H_
+
+#include "i915_reg_defs.h"
+#include "intel_uc_fw.h"
+#include "intel_huc_fw.h"
+
+struct intel_huc {
+	/* Generic uC firmware management */
+	struct intel_uc_fw fw;
+
+	/* HuC-specific additions */
+	struct {
+		i915_reg_t reg;
+		u32 mask;
+		u32 value;
+	} status;
+};
+
+void intel_huc_init_early(struct intel_huc *huc);
+int intel_huc_init(struct intel_huc *huc);
+void intel_huc_fini(struct intel_huc *huc);
+int intel_huc_auth(struct intel_huc *huc);
+int intel_huc_check_status(struct intel_huc *huc);
+void intel_huc_update_auth_status(struct intel_huc *huc);
+
+static inline int intel_huc_sanitize(struct intel_huc *huc)
+{
+	intel_uc_fw_sanitize(&huc->fw);
+	return 0;
+}
+
+static inline bool intel_huc_is_supported(struct intel_huc *huc)
+{
+	return intel_uc_fw_is_supported(&huc->fw);
+}
+
+static inline bool intel_huc_is_wanted(struct intel_huc *huc)
+{
+	return intel_uc_fw_is_enabled(&huc->fw);
+}
+
+static inline bool intel_huc_is_used(struct intel_huc *huc)
+{
+	GEM_BUG_ON(__intel_uc_fw_status(&huc->fw) == INTEL_UC_FIRMWARE_SELECTED);
+	return intel_uc_fw_is_available(&huc->fw);
+}
+
+static inline bool intel_huc_is_loaded_by_gsc(const struct intel_huc *huc)
+{
+	return huc->fw.loaded_via_gsc;
+}
+
+void intel_huc_load_status(struct intel_huc *huc, struct drm_printer *p);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_huc_debugfs.c b/drivers/gpu/drm/i915/gt/uc/intel_huc_debugfs.c
new file mode 100644
index 000000000..15998963b
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_huc_debugfs.c
@@ -0,0 +1,36 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <drm/drm_print.h>
+
+#include "gt/intel_gt_debugfs.h"
+#include "intel_huc.h"
+#include "intel_huc_debugfs.h"
+
+static int huc_info_show(struct seq_file *m, void *data)
+{
+	struct intel_huc *huc = m->private;
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	if (!intel_huc_is_supported(huc))
+		return -ENODEV;
+
+	intel_huc_load_status(huc, &p);
+
+	return 0;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(huc_info);
+
+void intel_huc_debugfs_register(struct intel_huc *huc, struct dentry *root)
+{
+	static const struct intel_gt_debugfs_file files[] = {
+		{ "huc_info", &huc_info_fops, NULL },
+	};
+
+	if (!intel_huc_is_supported(huc))
+		return;
+
+	intel_gt_debugfs_register_files(root, files, ARRAY_SIZE(files), huc);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_huc_debugfs.h b/drivers/gpu/drm/i915/gt/uc/intel_huc_debugfs.h
new file mode 100644
index 000000000..be79e992f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_huc_debugfs.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef DEBUGFS_HUC_H
+#define DEBUGFS_HUC_H
+
+struct intel_huc;
+struct dentry;
+
+void intel_huc_debugfs_register(struct intel_huc *huc, struct dentry *root);
+
+#endif /* DEBUGFS_HUC_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_huc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_huc_fw.c
new file mode 100644
index 000000000..9d6ab1e01
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_huc_fw.c
@@ -0,0 +1,29 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#include "gt/intel_gt.h"
+#include "intel_huc_fw.h"
+#include "i915_drv.h"
+
+/**
+ * intel_huc_fw_upload() - load HuC uCode to device via DMA transfer
+ * @huc: intel_huc structure
+ *
+ * Called from intel_uc_init_hw() during driver load, resume from sleep and
+ * after a GPU reset. Note that HuC must be loaded before GuC.
+ *
+ * The firmware image should have already been fetched into memory, so only
+ * check that fetch succeeded, and then transfer the image to the h/w.
+ *
+ * Return:	non-zero code on error
+ */
+int intel_huc_fw_upload(struct intel_huc *huc)
+{
+	if (intel_huc_is_loaded_by_gsc(huc))
+		return -ENODEV;
+
+	/* HW doesn't look at destination address for HuC, so set it to 0 */
+	return intel_uc_fw_upload(&huc->fw, 0, HUC_UKERNEL);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_huc_fw.h b/drivers/gpu/drm/i915/gt/uc/intel_huc_fw.h
new file mode 100644
index 000000000..12f264ee3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_huc_fw.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_HUC_FW_H_
+#define _INTEL_HUC_FW_H_
+
+struct intel_huc;
+
+int intel_huc_fw_upload(struct intel_huc *huc);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc.c b/drivers/gpu/drm/i915/gt/uc/intel_uc.c
new file mode 100644
index 000000000..dbd048b77
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc.c
@@ -0,0 +1,734 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2016-2019 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+
+#include "gt/intel_gt.h"
+#include "gt/intel_reset.h"
+#include "intel_guc.h"
+#include "intel_guc_ads.h"
+#include "intel_guc_submission.h"
+#include "gt/intel_rps.h"
+#include "intel_uc.h"
+
+#include "i915_drv.h"
+
+static const struct intel_uc_ops uc_ops_off;
+static const struct intel_uc_ops uc_ops_on;
+
+static void uc_expand_default_options(struct intel_uc *uc)
+{
+	struct drm_i915_private *i915 = uc_to_gt(uc)->i915;
+
+	if (i915->params.enable_guc != -1)
+		return;
+
+	/* Don't enable GuC/HuC on pre-Gen12 */
+	if (GRAPHICS_VER(i915) < 12) {
+		i915->params.enable_guc = 0;
+		return;
+	}
+
+	/* Don't enable GuC/HuC on older Gen12 platforms */
+	if (IS_TIGERLAKE(i915) || IS_ROCKETLAKE(i915)) {
+		i915->params.enable_guc = 0;
+		return;
+	}
+
+	/* Intermediate platforms are HuC authentication only */
+	if (IS_ALDERLAKE_S(i915) && !IS_ADLS_RPLS(i915)) {
+		i915->params.enable_guc = ENABLE_GUC_LOAD_HUC;
+		return;
+	}
+
+	/* Default: enable HuC authentication and GuC submission */
+	i915->params.enable_guc = ENABLE_GUC_LOAD_HUC | ENABLE_GUC_SUBMISSION;
+
+	/* XEHPSDV and PVC do not use HuC */
+	if (IS_XEHPSDV(i915) || IS_PONTEVECCHIO(i915))
+		i915->params.enable_guc &= ~ENABLE_GUC_LOAD_HUC;
+}
+
+/* Reset GuC providing us with fresh state for both GuC and HuC.
+ */
+static int __intel_uc_reset_hw(struct intel_uc *uc)
+{
+	struct intel_gt *gt = uc_to_gt(uc);
+	int ret;
+	u32 guc_status;
+
+	ret = i915_inject_probe_error(gt->i915, -ENXIO);
+	if (ret)
+		return ret;
+
+	ret = intel_reset_guc(gt);
+	if (ret) {
+		DRM_ERROR("Failed to reset GuC, ret = %d\n", ret);
+		return ret;
+	}
+
+	guc_status = intel_uncore_read(gt->uncore, GUC_STATUS);
+	WARN(!(guc_status & GS_MIA_IN_RESET),
+	     "GuC status: 0x%x, MIA core expected to be in reset\n",
+	     guc_status);
+
+	return ret;
+}
+
+static void __confirm_options(struct intel_uc *uc)
+{
+	struct drm_i915_private *i915 = uc_to_gt(uc)->i915;
+
+	drm_dbg(&i915->drm,
+		"enable_guc=%d (guc:%s submission:%s huc:%s slpc:%s)\n",
+		i915->params.enable_guc,
+		str_yes_no(intel_uc_wants_guc(uc)),
+		str_yes_no(intel_uc_wants_guc_submission(uc)),
+		str_yes_no(intel_uc_wants_huc(uc)),
+		str_yes_no(intel_uc_wants_guc_slpc(uc)));
+
+	if (i915->params.enable_guc == 0) {
+		GEM_BUG_ON(intel_uc_wants_guc(uc));
+		GEM_BUG_ON(intel_uc_wants_guc_submission(uc));
+		GEM_BUG_ON(intel_uc_wants_huc(uc));
+		GEM_BUG_ON(intel_uc_wants_guc_slpc(uc));
+		return;
+	}
+
+	if (!intel_uc_supports_guc(uc))
+		drm_info(&i915->drm,
+			 "Incompatible option enable_guc=%d - %s\n",
+			 i915->params.enable_guc, "GuC is not supported!");
+
+	if (i915->params.enable_guc & ENABLE_GUC_LOAD_HUC &&
+	    !intel_uc_supports_huc(uc))
+		drm_info(&i915->drm,
+			 "Incompatible option enable_guc=%d - %s\n",
+			 i915->params.enable_guc, "HuC is not supported!");
+
+	if (i915->params.enable_guc & ENABLE_GUC_SUBMISSION &&
+	    !intel_uc_supports_guc_submission(uc))
+		drm_info(&i915->drm,
+			 "Incompatible option enable_guc=%d - %s\n",
+			 i915->params.enable_guc, "GuC submission is N/A");
+
+	if (i915->params.enable_guc & ~ENABLE_GUC_MASK)
+		drm_info(&i915->drm,
+			 "Incompatible option enable_guc=%d - %s\n",
+			 i915->params.enable_guc, "undocumented flag");
+}
+
+void intel_uc_init_early(struct intel_uc *uc)
+{
+	uc_expand_default_options(uc);
+
+	intel_guc_init_early(&uc->guc);
+	intel_huc_init_early(&uc->huc);
+
+	__confirm_options(uc);
+
+	if (intel_uc_wants_guc(uc))
+		uc->ops = &uc_ops_on;
+	else
+		uc->ops = &uc_ops_off;
+}
+
+void intel_uc_init_late(struct intel_uc *uc)
+{
+	intel_guc_init_late(&uc->guc);
+}
+
+void intel_uc_driver_late_release(struct intel_uc *uc)
+{
+}
+
+/**
+ * intel_uc_init_mmio - setup uC MMIO access
+ * @uc: the intel_uc structure
+ *
+ * Setup minimal state necessary for MMIO accesses later in the
+ * initialization sequence.
+ */
+void intel_uc_init_mmio(struct intel_uc *uc)
+{
+	intel_guc_init_send_regs(&uc->guc);
+}
+
+static void __uc_capture_load_err_log(struct intel_uc *uc)
+{
+	struct intel_guc *guc = &uc->guc;
+
+	if (guc->log.vma && !uc->load_err_log)
+		uc->load_err_log = i915_gem_object_get(guc->log.vma->obj);
+}
+
+static void __uc_free_load_err_log(struct intel_uc *uc)
+{
+	struct drm_i915_gem_object *log = fetch_and_zero(&uc->load_err_log);
+
+	if (log)
+		i915_gem_object_put(log);
+}
+
+void intel_uc_driver_remove(struct intel_uc *uc)
+{
+	intel_uc_fini_hw(uc);
+	intel_uc_fini(uc);
+	__uc_free_load_err_log(uc);
+}
+
+/*
+ * Events triggered while CT buffers are disabled are logged in the SCRATCH_15
+ * register using the same bits used in the CT message payload. Since our
+ * communication channel with guc is turned off at this point, we can save the
+ * message and handle it after we turn it back on.
+ */
+static void guc_clear_mmio_msg(struct intel_guc *guc)
+{
+	intel_uncore_write(guc_to_gt(guc)->uncore, SOFT_SCRATCH(15), 0);
+}
+
+static void guc_get_mmio_msg(struct intel_guc *guc)
+{
+	u32 val;
+
+	spin_lock_irq(&guc->irq_lock);
+
+	val = intel_uncore_read(guc_to_gt(guc)->uncore, SOFT_SCRATCH(15));
+	guc->mmio_msg |= val & guc->msg_enabled_mask;
+
+	/*
+	 * clear all events, including the ones we're not currently servicing,
+	 * to make sure we don't try to process a stale message if we enable
+	 * handling of more events later.
+	 */
+	guc_clear_mmio_msg(guc);
+
+	spin_unlock_irq(&guc->irq_lock);
+}
+
+static void guc_handle_mmio_msg(struct intel_guc *guc)
+{
+	/* we need communication to be enabled to reply to GuC */
+	GEM_BUG_ON(!intel_guc_ct_enabled(&guc->ct));
+
+	spin_lock_irq(&guc->irq_lock);
+	if (guc->mmio_msg) {
+		intel_guc_to_host_process_recv_msg(guc, &guc->mmio_msg, 1);
+		guc->mmio_msg = 0;
+	}
+	spin_unlock_irq(&guc->irq_lock);
+}
+
+static int guc_enable_communication(struct intel_guc *guc)
+{
+	struct intel_gt *gt = guc_to_gt(guc);
+	struct drm_i915_private *i915 = gt->i915;
+	int ret;
+
+	GEM_BUG_ON(intel_guc_ct_enabled(&guc->ct));
+
+	ret = i915_inject_probe_error(i915, -ENXIO);
+	if (ret)
+		return ret;
+
+	ret = intel_guc_ct_enable(&guc->ct);
+	if (ret)
+		return ret;
+
+	/* check for mmio messages received before/during the CT enable */
+	guc_get_mmio_msg(guc);
+	guc_handle_mmio_msg(guc);
+
+	intel_guc_enable_interrupts(guc);
+
+	/* check for CT messages received before we enabled interrupts */
+	spin_lock_irq(gt->irq_lock);
+	intel_guc_ct_event_handler(&guc->ct);
+	spin_unlock_irq(gt->irq_lock);
+
+	drm_dbg(&i915->drm, "GuC communication enabled\n");
+
+	return 0;
+}
+
+static void guc_disable_communication(struct intel_guc *guc)
+{
+	struct drm_i915_private *i915 = guc_to_gt(guc)->i915;
+
+	/*
+	 * Events generated during or after CT disable are logged by guc in
+	 * via mmio. Make sure the register is clear before disabling CT since
+	 * all events we cared about have already been processed via CT.
+	 */
+	guc_clear_mmio_msg(guc);
+
+	intel_guc_disable_interrupts(guc);
+
+	intel_guc_ct_disable(&guc->ct);
+
+	/*
+	 * Check for messages received during/after the CT disable. We do not
+	 * expect any messages to have arrived via CT between the interrupt
+	 * disable and the CT disable because GuC should've been idle until we
+	 * triggered the CT disable protocol.
+	 */
+	guc_get_mmio_msg(guc);
+
+	drm_dbg(&i915->drm, "GuC communication disabled\n");
+}
+
+static void __uc_fetch_firmwares(struct intel_uc *uc)
+{
+	int err;
+
+	GEM_BUG_ON(!intel_uc_wants_guc(uc));
+
+	err = intel_uc_fw_fetch(&uc->guc.fw);
+	if (err) {
+		/* Make sure we transition out of transient "SELECTED" state */
+		if (intel_uc_wants_huc(uc)) {
+			drm_dbg(&uc_to_gt(uc)->i915->drm,
+				"Failed to fetch GuC: %d disabling HuC\n", err);
+			intel_uc_fw_change_status(&uc->huc.fw,
+						  INTEL_UC_FIRMWARE_ERROR);
+		}
+
+		return;
+	}
+
+	if (intel_uc_wants_huc(uc))
+		intel_uc_fw_fetch(&uc->huc.fw);
+}
+
+static void __uc_cleanup_firmwares(struct intel_uc *uc)
+{
+	intel_uc_fw_cleanup_fetch(&uc->huc.fw);
+	intel_uc_fw_cleanup_fetch(&uc->guc.fw);
+}
+
+static int __uc_init(struct intel_uc *uc)
+{
+	struct intel_guc *guc = &uc->guc;
+	struct intel_huc *huc = &uc->huc;
+	int ret;
+
+	GEM_BUG_ON(!intel_uc_wants_guc(uc));
+
+	if (!intel_uc_uses_guc(uc))
+		return 0;
+
+	if (i915_inject_probe_failure(uc_to_gt(uc)->i915))
+		return -ENOMEM;
+
+	ret = intel_guc_init(guc);
+	if (ret)
+		return ret;
+
+	if (intel_uc_uses_huc(uc))
+		intel_huc_init(huc);
+
+	return 0;
+}
+
+static void __uc_fini(struct intel_uc *uc)
+{
+	intel_huc_fini(&uc->huc);
+	intel_guc_fini(&uc->guc);
+}
+
+static int __uc_sanitize(struct intel_uc *uc)
+{
+	struct intel_guc *guc = &uc->guc;
+	struct intel_huc *huc = &uc->huc;
+
+	GEM_BUG_ON(!intel_uc_supports_guc(uc));
+
+	intel_huc_sanitize(huc);
+	intel_guc_sanitize(guc);
+
+	return __intel_uc_reset_hw(uc);
+}
+
+/* Initialize and verify the uC regs related to uC positioning in WOPCM */
+static int uc_init_wopcm(struct intel_uc *uc)
+{
+	struct intel_gt *gt = uc_to_gt(uc);
+	struct intel_uncore *uncore = gt->uncore;
+	u32 base = intel_wopcm_guc_base(&gt->i915->wopcm);
+	u32 size = intel_wopcm_guc_size(&gt->i915->wopcm);
+	u32 huc_agent = intel_uc_uses_huc(uc) ? HUC_LOADING_AGENT_GUC : 0;
+	u32 mask;
+	int err;
+
+	if (unlikely(!base || !size)) {
+		i915_probe_error(gt->i915, "Unsuccessful WOPCM partitioning\n");
+		return -E2BIG;
+	}
+
+	GEM_BUG_ON(!intel_uc_supports_guc(uc));
+	GEM_BUG_ON(!(base & GUC_WOPCM_OFFSET_MASK));
+	GEM_BUG_ON(base & ~GUC_WOPCM_OFFSET_MASK);
+	GEM_BUG_ON(!(size & GUC_WOPCM_SIZE_MASK));
+	GEM_BUG_ON(size & ~GUC_WOPCM_SIZE_MASK);
+
+	err = i915_inject_probe_error(gt->i915, -ENXIO);
+	if (err)
+		return err;
+
+	mask = GUC_WOPCM_SIZE_MASK | GUC_WOPCM_SIZE_LOCKED;
+	err = intel_uncore_write_and_verify(uncore, GUC_WOPCM_SIZE, size, mask,
+					    size | GUC_WOPCM_SIZE_LOCKED);
+	if (err)
+		goto err_out;
+
+	mask = GUC_WOPCM_OFFSET_MASK | GUC_WOPCM_OFFSET_VALID | huc_agent;
+	err = intel_uncore_write_and_verify(uncore, DMA_GUC_WOPCM_OFFSET,
+					    base | huc_agent, mask,
+					    base | huc_agent |
+					    GUC_WOPCM_OFFSET_VALID);
+	if (err)
+		goto err_out;
+
+	return 0;
+
+err_out:
+	i915_probe_error(gt->i915, "Failed to init uC WOPCM registers!\n");
+	i915_probe_error(gt->i915, "%s(%#x)=%#x\n", "DMA_GUC_WOPCM_OFFSET",
+			 i915_mmio_reg_offset(DMA_GUC_WOPCM_OFFSET),
+			 intel_uncore_read(uncore, DMA_GUC_WOPCM_OFFSET));
+	i915_probe_error(gt->i915, "%s(%#x)=%#x\n", "GUC_WOPCM_SIZE",
+			 i915_mmio_reg_offset(GUC_WOPCM_SIZE),
+			 intel_uncore_read(uncore, GUC_WOPCM_SIZE));
+
+	return err;
+}
+
+static bool uc_is_wopcm_locked(struct intel_uc *uc)
+{
+	struct intel_gt *gt = uc_to_gt(uc);
+	struct intel_uncore *uncore = gt->uncore;
+
+	return (intel_uncore_read(uncore, GUC_WOPCM_SIZE) & GUC_WOPCM_SIZE_LOCKED) ||
+	       (intel_uncore_read(uncore, DMA_GUC_WOPCM_OFFSET) & GUC_WOPCM_OFFSET_VALID);
+}
+
+static int __uc_check_hw(struct intel_uc *uc)
+{
+	if (!intel_uc_supports_guc(uc))
+		return 0;
+
+	/*
+	 * We can silently continue without GuC only if it was never enabled
+	 * before on this system after reboot, otherwise we risk GPU hangs.
+	 * To check if GuC was loaded before we look at WOPCM registers.
+	 */
+	if (uc_is_wopcm_locked(uc))
+		return -EIO;
+
+	return 0;
+}
+
+static void print_fw_ver(struct intel_uc *uc, struct intel_uc_fw *fw)
+{
+	struct drm_i915_private *i915 = uc_to_gt(uc)->i915;
+
+	drm_info(&i915->drm, "%s firmware %s version %u.%u.%u\n",
+		 intel_uc_fw_type_repr(fw->type), fw->file_selected.path,
+		 fw->file_selected.major_ver,
+		 fw->file_selected.minor_ver,
+		 fw->file_selected.patch_ver);
+}
+
+static int __uc_init_hw(struct intel_uc *uc)
+{
+	struct drm_i915_private *i915 = uc_to_gt(uc)->i915;
+	struct intel_guc *guc = &uc->guc;
+	struct intel_huc *huc = &uc->huc;
+	int ret, attempts;
+
+	GEM_BUG_ON(!intel_uc_supports_guc(uc));
+	GEM_BUG_ON(!intel_uc_wants_guc(uc));
+
+	print_fw_ver(uc, &guc->fw);
+
+	if (intel_uc_uses_huc(uc))
+		print_fw_ver(uc, &huc->fw);
+
+	if (!intel_uc_fw_is_loadable(&guc->fw)) {
+		ret = __uc_check_hw(uc) ||
+		      intel_uc_fw_is_overridden(&guc->fw) ||
+		      intel_uc_wants_guc_submission(uc) ?
+		      intel_uc_fw_status_to_error(guc->fw.status) : 0;
+		goto err_out;
+	}
+
+	ret = uc_init_wopcm(uc);
+	if (ret)
+		goto err_out;
+
+	intel_guc_reset_interrupts(guc);
+
+	/* WaEnableuKernelHeaderValidFix:skl */
+	/* WaEnableGuCBootHashCheckNotSet:skl,bxt,kbl */
+	if (GRAPHICS_VER(i915) == 9)
+		attempts = 3;
+	else
+		attempts = 1;
+
+	intel_rps_raise_unslice(&uc_to_gt(uc)->rps);
+
+	while (attempts--) {
+		/*
+		 * Always reset the GuC just before (re)loading, so
+		 * that the state and timing are fairly predictable
+		 */
+		ret = __uc_sanitize(uc);
+		if (ret)
+			goto err_out;
+
+		intel_huc_fw_upload(huc);
+		intel_guc_ads_reset(guc);
+		intel_guc_write_params(guc);
+		ret = intel_guc_fw_upload(guc);
+		if (ret == 0)
+			break;
+
+		DRM_DEBUG_DRIVER("GuC fw load failed: %d; will reset and "
+				 "retry %d more time(s)\n", ret, attempts);
+	}
+
+	/* Did we succeded or run out of retries? */
+	if (ret)
+		goto err_log_capture;
+
+	ret = guc_enable_communication(guc);
+	if (ret)
+		goto err_log_capture;
+
+	/*
+	 * GSC-loaded HuC is authenticated by the GSC, so we don't need to
+	 * trigger the auth here. However, given that the HuC loaded this way
+	 * survive GT reset, we still need to update our SW bookkeeping to make
+	 * sure it reflects the correct HW status.
+	 */
+	if (intel_huc_is_loaded_by_gsc(huc))
+		intel_huc_update_auth_status(huc);
+	else
+		intel_huc_auth(huc);
+
+	if (intel_uc_uses_guc_submission(uc))
+		intel_guc_submission_enable(guc);
+
+	if (intel_uc_uses_guc_slpc(uc)) {
+		ret = intel_guc_slpc_enable(&guc->slpc);
+		if (ret)
+			goto err_submission;
+	} else {
+		/* Restore GT back to RPn for non-SLPC path */
+		intel_rps_lower_unslice(&uc_to_gt(uc)->rps);
+	}
+
+	drm_info(&i915->drm, "GuC submission %s\n",
+		 str_enabled_disabled(intel_uc_uses_guc_submission(uc)));
+	drm_info(&i915->drm, "GuC SLPC %s\n",
+		 str_enabled_disabled(intel_uc_uses_guc_slpc(uc)));
+
+	return 0;
+
+	/*
+	 * We've failed to load the firmware :(
+	 */
+err_submission:
+	intel_guc_submission_disable(guc);
+err_log_capture:
+	__uc_capture_load_err_log(uc);
+err_out:
+	/* Return GT back to RPn */
+	intel_rps_lower_unslice(&uc_to_gt(uc)->rps);
+
+	__uc_sanitize(uc);
+
+	if (!ret) {
+		drm_notice(&i915->drm, "GuC is uninitialized\n");
+		/* We want to run without GuC submission */
+		return 0;
+	}
+
+	i915_probe_error(i915, "GuC initialization failed %d\n", ret);
+
+	/* We want to keep KMS alive */
+	return -EIO;
+}
+
+static void __uc_fini_hw(struct intel_uc *uc)
+{
+	struct intel_guc *guc = &uc->guc;
+
+	if (!intel_guc_is_fw_running(guc))
+		return;
+
+	if (intel_uc_uses_guc_submission(uc))
+		intel_guc_submission_disable(guc);
+
+	__uc_sanitize(uc);
+}
+
+/**
+ * intel_uc_reset_prepare - Prepare for reset
+ * @uc: the intel_uc structure
+ *
+ * Preparing for full gpu reset.
+ */
+void intel_uc_reset_prepare(struct intel_uc *uc)
+{
+	struct intel_guc *guc = &uc->guc;
+
+	uc->reset_in_progress = true;
+
+	/* Nothing to do if GuC isn't supported */
+	if (!intel_uc_supports_guc(uc))
+		return;
+
+	/* Firmware expected to be running when this function is called */
+	if (!intel_guc_is_ready(guc))
+		goto sanitize;
+
+	if (intel_uc_uses_guc_submission(uc))
+		intel_guc_submission_reset_prepare(guc);
+
+sanitize:
+	__uc_sanitize(uc);
+}
+
+void intel_uc_reset(struct intel_uc *uc, intel_engine_mask_t stalled)
+{
+	struct intel_guc *guc = &uc->guc;
+
+	/* Firmware can not be running when this function is called  */
+	if (intel_uc_uses_guc_submission(uc))
+		intel_guc_submission_reset(guc, stalled);
+}
+
+void intel_uc_reset_finish(struct intel_uc *uc)
+{
+	struct intel_guc *guc = &uc->guc;
+
+	uc->reset_in_progress = false;
+
+	/* Firmware expected to be running when this function is called */
+	if (intel_guc_is_fw_running(guc) && intel_uc_uses_guc_submission(uc))
+		intel_guc_submission_reset_finish(guc);
+}
+
+void intel_uc_cancel_requests(struct intel_uc *uc)
+{
+	struct intel_guc *guc = &uc->guc;
+
+	/* Firmware can not be running when this function is called  */
+	if (intel_uc_uses_guc_submission(uc))
+		intel_guc_submission_cancel_requests(guc);
+}
+
+void intel_uc_runtime_suspend(struct intel_uc *uc)
+{
+	struct intel_guc *guc = &uc->guc;
+
+	if (!intel_guc_is_ready(guc))
+		return;
+
+	/*
+	 * Wait for any outstanding CTB before tearing down communication /w the
+	 * GuC.
+	 */
+#define OUTSTANDING_CTB_TIMEOUT_PERIOD	(HZ / 5)
+	intel_guc_wait_for_pending_msg(guc, &guc->outstanding_submission_g2h,
+				       false, OUTSTANDING_CTB_TIMEOUT_PERIOD);
+	GEM_WARN_ON(atomic_read(&guc->outstanding_submission_g2h));
+
+	guc_disable_communication(guc);
+}
+
+void intel_uc_suspend(struct intel_uc *uc)
+{
+	struct intel_guc *guc = &uc->guc;
+	intel_wakeref_t wakeref;
+	int err;
+
+	if (!intel_guc_is_ready(guc))
+		return;
+
+	with_intel_runtime_pm(&uc_to_gt(uc)->i915->runtime_pm, wakeref) {
+		err = intel_guc_suspend(guc);
+		if (err)
+			DRM_DEBUG_DRIVER("Failed to suspend GuC, err=%d", err);
+	}
+}
+
+static int __uc_resume(struct intel_uc *uc, bool enable_communication)
+{
+	struct intel_guc *guc = &uc->guc;
+	struct intel_gt *gt = guc_to_gt(guc);
+	int err;
+
+	if (!intel_guc_is_fw_running(guc))
+		return 0;
+
+	/* Make sure we enable communication if and only if it's disabled */
+	GEM_BUG_ON(enable_communication == intel_guc_ct_enabled(&guc->ct));
+
+	if (enable_communication)
+		guc_enable_communication(guc);
+
+	/* If we are only resuming GuC communication but not reloading
+	 * GuC, we need to ensure the ARAT timer interrupt is enabled
+	 * again. In case of GuC reload, it is enabled during SLPC enable.
+	 */
+	if (enable_communication && intel_uc_uses_guc_slpc(uc))
+		intel_guc_pm_intrmsk_enable(gt);
+
+	err = intel_guc_resume(guc);
+	if (err) {
+		DRM_DEBUG_DRIVER("Failed to resume GuC, err=%d", err);
+		return err;
+	}
+
+	return 0;
+}
+
+int intel_uc_resume(struct intel_uc *uc)
+{
+	/*
+	 * When coming out of S3/S4 we sanitize and re-init the HW, so
+	 * communication is already re-enabled at this point.
+	 */
+	return __uc_resume(uc, false);
+}
+
+int intel_uc_runtime_resume(struct intel_uc *uc)
+{
+	/*
+	 * During runtime resume we don't sanitize, so we need to re-init
+	 * communication as well.
+	 */
+	return __uc_resume(uc, true);
+}
+
+static const struct intel_uc_ops uc_ops_off = {
+	.init_hw = __uc_check_hw,
+};
+
+static const struct intel_uc_ops uc_ops_on = {
+	.sanitize = __uc_sanitize,
+
+	.init_fw = __uc_fetch_firmwares,
+	.fini_fw = __uc_cleanup_firmwares,
+
+	.init = __uc_init,
+	.fini = __uc_fini,
+
+	.init_hw = __uc_init_hw,
+	.fini_hw = __uc_fini_hw,
+};
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc.h b/drivers/gpu/drm/i915/gt/uc/intel_uc.h
new file mode 100644
index 000000000..a8f38c2c6
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc.h
@@ -0,0 +1,115 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_UC_H_
+#define _INTEL_UC_H_
+
+#include "intel_guc.h"
+#include "intel_guc_rc.h"
+#include "intel_guc_submission.h"
+#include "intel_guc_slpc.h"
+#include "intel_huc.h"
+#include "i915_params.h"
+
+struct intel_uc;
+
+struct intel_uc_ops {
+	int (*sanitize)(struct intel_uc *uc);
+	void (*init_fw)(struct intel_uc *uc);
+	void (*fini_fw)(struct intel_uc *uc);
+	int (*init)(struct intel_uc *uc);
+	void (*fini)(struct intel_uc *uc);
+	int (*init_hw)(struct intel_uc *uc);
+	void (*fini_hw)(struct intel_uc *uc);
+};
+
+struct intel_uc {
+	struct intel_uc_ops const *ops;
+	struct intel_guc guc;
+	struct intel_huc huc;
+
+	/* Snapshot of GuC log from last failed load */
+	struct drm_i915_gem_object *load_err_log;
+
+	bool reset_in_progress;
+};
+
+void intel_uc_init_early(struct intel_uc *uc);
+void intel_uc_init_late(struct intel_uc *uc);
+void intel_uc_driver_late_release(struct intel_uc *uc);
+void intel_uc_driver_remove(struct intel_uc *uc);
+void intel_uc_init_mmio(struct intel_uc *uc);
+void intel_uc_reset_prepare(struct intel_uc *uc);
+void intel_uc_reset(struct intel_uc *uc, intel_engine_mask_t stalled);
+void intel_uc_reset_finish(struct intel_uc *uc);
+void intel_uc_cancel_requests(struct intel_uc *uc);
+void intel_uc_suspend(struct intel_uc *uc);
+void intel_uc_runtime_suspend(struct intel_uc *uc);
+int intel_uc_resume(struct intel_uc *uc);
+int intel_uc_runtime_resume(struct intel_uc *uc);
+
+/*
+ * We need to know as early as possible if we're going to use GuC or not to
+ * take the correct setup paths. Additionally, once we've started loading the
+ * GuC, it is unsafe to keep executing without it because some parts of the HW,
+ * a subset of which is not cleaned on GT reset, will start expecting the GuC FW
+ * to be running.
+ * To solve both these requirements, we commit to using the microcontrollers if
+ * the relevant modparam is set and the blobs are found on the system. At this
+ * stage, the only thing that can stop us from attempting to load the blobs on
+ * the HW and use them is a fundamental issue (e.g. no memory for our
+ * structures); if we hit such a problem during driver load we're broken even
+ * without GuC, so there is no point in trying to fall back.
+ *
+ * Given the above, we can be in one of 4 states, with the last one implying
+ * we're committed to using the microcontroller:
+ * - Not supported: not available in HW and/or firmware not defined.
+ * - Supported: available in HW and firmware defined.
+ * - Wanted: supported + enabled in modparam.
+ * - In use: wanted + firmware found on the system and successfully fetched.
+ */
+
+#define __uc_state_checker(x, func, state, required) \
+static inline bool intel_uc_##state##_##func(struct intel_uc *uc) \
+{ \
+	return intel_##func##_is_##required(&uc->x); \
+}
+
+#define uc_state_checkers(x, func) \
+__uc_state_checker(x, func, supports, supported) \
+__uc_state_checker(x, func, wants, wanted) \
+__uc_state_checker(x, func, uses, used)
+
+uc_state_checkers(guc, guc);
+uc_state_checkers(huc, huc);
+uc_state_checkers(guc, guc_submission);
+uc_state_checkers(guc, guc_slpc);
+uc_state_checkers(guc, guc_rc);
+
+#undef uc_state_checkers
+#undef __uc_state_checker
+
+static inline int intel_uc_wait_for_idle(struct intel_uc *uc, long timeout)
+{
+	return intel_guc_wait_for_idle(&uc->guc, timeout);
+}
+
+#define intel_uc_ops_function(_NAME, _OPS, _TYPE, _RET) \
+static inline _TYPE intel_uc_##_NAME(struct intel_uc *uc) \
+{ \
+	if (uc->ops->_OPS) \
+		return uc->ops->_OPS(uc); \
+	return _RET; \
+}
+intel_uc_ops_function(sanitize, sanitize, int, 0);
+intel_uc_ops_function(fetch_firmwares, init_fw, void, );
+intel_uc_ops_function(cleanup_firmwares, fini_fw, void, );
+intel_uc_ops_function(init, init, int, 0);
+intel_uc_ops_function(fini, fini, void, );
+intel_uc_ops_function(init_hw, init_hw, int, 0);
+intel_uc_ops_function(fini_hw, fini_hw, void, );
+#undef intel_uc_ops_function
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_debugfs.c b/drivers/gpu/drm/i915/gt/uc/intel_uc_debugfs.c
new file mode 100644
index 000000000..284d6fbc2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_debugfs.c
@@ -0,0 +1,61 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/debugfs.h>
+#include <linux/string_helpers.h>
+
+#include <drm/drm_print.h>
+
+#include "gt/intel_gt_debugfs.h"
+#include "intel_guc_debugfs.h"
+#include "intel_huc_debugfs.h"
+#include "intel_uc.h"
+#include "intel_uc_debugfs.h"
+
+static int uc_usage_show(struct seq_file *m, void *data)
+{
+	struct intel_uc *uc = m->private;
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	drm_printf(&p, "[guc] supported:%s wanted:%s used:%s\n",
+		   str_yes_no(intel_uc_supports_guc(uc)),
+		   str_yes_no(intel_uc_wants_guc(uc)),
+		   str_yes_no(intel_uc_uses_guc(uc)));
+	drm_printf(&p, "[huc] supported:%s wanted:%s used:%s\n",
+		   str_yes_no(intel_uc_supports_huc(uc)),
+		   str_yes_no(intel_uc_wants_huc(uc)),
+		   str_yes_no(intel_uc_uses_huc(uc)));
+	drm_printf(&p, "[submission] supported:%s wanted:%s used:%s\n",
+		   str_yes_no(intel_uc_supports_guc_submission(uc)),
+		   str_yes_no(intel_uc_wants_guc_submission(uc)),
+		   str_yes_no(intel_uc_uses_guc_submission(uc)));
+
+	return 0;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(uc_usage);
+
+void intel_uc_debugfs_register(struct intel_uc *uc, struct dentry *gt_root)
+{
+	static const struct intel_gt_debugfs_file files[] = {
+		{ "usage", &uc_usage_fops, NULL },
+	};
+	struct dentry *root;
+
+	if (!gt_root)
+		return;
+
+	/* GuC and HuC go always in pair, no need to check both */
+	if (!intel_uc_supports_guc(uc))
+		return;
+
+	root = debugfs_create_dir("uc", gt_root);
+	if (IS_ERR(root))
+		return;
+
+	intel_gt_debugfs_register_files(root, files, ARRAY_SIZE(files), uc);
+
+	intel_guc_debugfs_register(&uc->guc, root);
+	intel_huc_debugfs_register(&uc->huc, root);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_debugfs.h b/drivers/gpu/drm/i915/gt/uc/intel_uc_debugfs.h
new file mode 100644
index 000000000..010ce250d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_debugfs.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef DEBUGFS_UC_H
+#define DEBUGFS_UC_H
+
+struct intel_uc;
+struct dentry;
+
+void intel_uc_debugfs_register(struct intel_uc *uc, struct dentry *gt_root);
+
+#endif /* DEBUGFS_UC_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
new file mode 100644
index 000000000..b91ad4aed
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
@@ -0,0 +1,1051 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2016-2019 Intel Corporation
+ */
+
+#include <linux/bitfield.h>
+#include <linux/firmware.h>
+#include <linux/highmem.h>
+
+#include <drm/drm_cache.h>
+#include <drm/drm_print.h>
+
+#include "gem/i915_gem_lmem.h"
+#include "intel_uc_fw.h"
+#include "intel_uc_fw_abi.h"
+#include "i915_drv.h"
+#include "i915_reg.h"
+
+static inline struct intel_gt *
+____uc_fw_to_gt(struct intel_uc_fw *uc_fw, enum intel_uc_fw_type type)
+{
+	if (type == INTEL_UC_FW_TYPE_GUC)
+		return container_of(uc_fw, struct intel_gt, uc.guc.fw);
+
+	GEM_BUG_ON(type != INTEL_UC_FW_TYPE_HUC);
+	return container_of(uc_fw, struct intel_gt, uc.huc.fw);
+}
+
+static inline struct intel_gt *__uc_fw_to_gt(struct intel_uc_fw *uc_fw)
+{
+	GEM_BUG_ON(uc_fw->status == INTEL_UC_FIRMWARE_UNINITIALIZED);
+	return ____uc_fw_to_gt(uc_fw, uc_fw->type);
+}
+
+#ifdef CONFIG_DRM_I915_DEBUG_GUC
+void intel_uc_fw_change_status(struct intel_uc_fw *uc_fw,
+			       enum intel_uc_fw_status status)
+{
+	uc_fw->__status =  status;
+	drm_dbg(&__uc_fw_to_gt(uc_fw)->i915->drm,
+		"%s firmware -> %s\n",
+		intel_uc_fw_type_repr(uc_fw->type),
+		status == INTEL_UC_FIRMWARE_SELECTED ?
+		uc_fw->file_selected.path : intel_uc_fw_status_repr(status));
+}
+#endif
+
+/*
+ * List of required GuC and HuC binaries per-platform.
+ * Must be ordered based on platform + revid, from newer to older.
+ *
+ * Note that RKL and ADL-S have the same GuC/HuC device ID's and use the same
+ * firmware as TGL.
+ *
+ * Version numbers:
+ * Originally, the driver required an exact match major/minor/patch furmware
+ * file and only supported that one version for any given platform. However,
+ * the new direction from upstream is to be backwards compatible with all
+ * prior releases and to be as flexible as possible as to what firmware is
+ * loaded.
+ *
+ * For GuC, the major version number signifies a backwards breaking API change.
+ * So, new format GuC firmware files are labelled by their major version only.
+ * For HuC, there is no KMD interaction, hence no version matching requirement.
+ * So, new format HuC firmware files have no version number at all.
+ *
+ * All of which means that the table below must keep all old format files with
+ * full three point version number. But newer files have reduced requirements.
+ * Having said that, the driver still needs to track the minor version number
+ * for GuC at least. As it is useful to report to the user that they are not
+ * running with a recent enough version for all KMD supported features,
+ * security fixes, etc. to be enabled.
+ */
+#define INTEL_GUC_FIRMWARE_DEFS(fw_def, guc_maj, guc_mmp) \
+	fw_def(DG2,          0, guc_maj(dg2,  70, 5)) \
+	fw_def(ALDERLAKE_P,  0, guc_maj(adlp, 70, 5)) \
+	fw_def(ALDERLAKE_P,  0, guc_mmp(adlp, 70, 1, 1)) \
+	fw_def(ALDERLAKE_P,  0, guc_mmp(adlp, 69, 0, 3)) \
+	fw_def(ALDERLAKE_S,  0, guc_maj(tgl,  70, 5)) \
+	fw_def(ALDERLAKE_S,  0, guc_mmp(tgl,  70, 1, 1)) \
+	fw_def(ALDERLAKE_S,  0, guc_mmp(tgl,  69, 0, 3)) \
+	fw_def(DG1,          0, guc_maj(dg1,  70, 5)) \
+	fw_def(ROCKETLAKE,   0, guc_mmp(tgl,  70, 1, 1)) \
+	fw_def(TIGERLAKE,    0, guc_mmp(tgl,  70, 1, 1)) \
+	fw_def(JASPERLAKE,   0, guc_mmp(ehl,  70, 1, 1)) \
+	fw_def(ELKHARTLAKE,  0, guc_mmp(ehl,  70, 1, 1)) \
+	fw_def(ICELAKE,      0, guc_mmp(icl,  70, 1, 1)) \
+	fw_def(COMETLAKE,    5, guc_mmp(cml,  70, 1, 1)) \
+	fw_def(COMETLAKE,    0, guc_mmp(kbl,  70, 1, 1)) \
+	fw_def(COFFEELAKE,   0, guc_mmp(kbl,  70, 1, 1)) \
+	fw_def(GEMINILAKE,   0, guc_mmp(glk,  70, 1, 1)) \
+	fw_def(KABYLAKE,     0, guc_mmp(kbl,  70, 1, 1)) \
+	fw_def(BROXTON,      0, guc_mmp(bxt,  70, 1, 1)) \
+	fw_def(SKYLAKE,      0, guc_mmp(skl,  70, 1, 1))
+
+#define INTEL_HUC_FIRMWARE_DEFS(fw_def, huc_raw, huc_mmp) \
+	fw_def(ALDERLAKE_P,  0, huc_raw(tgl)) \
+	fw_def(ALDERLAKE_P,  0, huc_mmp(tgl,  7, 9, 3)) \
+	fw_def(ALDERLAKE_S,  0, huc_raw(tgl)) \
+	fw_def(ALDERLAKE_S,  0, huc_mmp(tgl,  7, 9, 3)) \
+	fw_def(DG1,          0, huc_raw(dg1)) \
+	fw_def(ROCKETLAKE,   0, huc_mmp(tgl,  7, 9, 3)) \
+	fw_def(TIGERLAKE,    0, huc_mmp(tgl,  7, 9, 3)) \
+	fw_def(JASPERLAKE,   0, huc_mmp(ehl,  9, 0, 0)) \
+	fw_def(ELKHARTLAKE,  0, huc_mmp(ehl,  9, 0, 0)) \
+	fw_def(ICELAKE,      0, huc_mmp(icl,  9, 0, 0)) \
+	fw_def(COMETLAKE,    5, huc_mmp(cml,  4, 0, 0)) \
+	fw_def(COMETLAKE,    0, huc_mmp(kbl,  4, 0, 0)) \
+	fw_def(COFFEELAKE,   0, huc_mmp(kbl,  4, 0, 0)) \
+	fw_def(GEMINILAKE,   0, huc_mmp(glk,  4, 0, 0)) \
+	fw_def(KABYLAKE,     0, huc_mmp(kbl,  4, 0, 0)) \
+	fw_def(BROXTON,      0, huc_mmp(bxt,  2, 0, 0)) \
+	fw_def(SKYLAKE,      0, huc_mmp(skl,  2, 0, 0))
+
+/*
+ * Set of macros for producing a list of filenames from the above table.
+ */
+#define __MAKE_UC_FW_PATH_BLANK(prefix_, name_) \
+	"i915/" \
+	__stringify(prefix_) name_ ".bin"
+
+#define __MAKE_UC_FW_PATH_MAJOR(prefix_, name_, major_) \
+	"i915/" \
+	__stringify(prefix_) name_ \
+	__stringify(major_) ".bin"
+
+#define __MAKE_UC_FW_PATH_MMP(prefix_, name_, major_, minor_, patch_) \
+	"i915/" \
+	__stringify(prefix_) name_ \
+	__stringify(major_) "." \
+	__stringify(minor_) "." \
+	__stringify(patch_) ".bin"
+
+/* Minor for internal driver use, not part of file name */
+#define MAKE_GUC_FW_PATH_MAJOR(prefix_, major_, minor_) \
+	__MAKE_UC_FW_PATH_MAJOR(prefix_, "_guc_", major_)
+
+#define MAKE_GUC_FW_PATH_MMP(prefix_, major_, minor_, patch_) \
+	__MAKE_UC_FW_PATH_MMP(prefix_, "_guc_", major_, minor_, patch_)
+
+#define MAKE_HUC_FW_PATH_BLANK(prefix_) \
+	__MAKE_UC_FW_PATH_BLANK(prefix_, "_huc")
+
+#define MAKE_HUC_FW_PATH_MMP(prefix_, major_, minor_, patch_) \
+	__MAKE_UC_FW_PATH_MMP(prefix_, "_huc_", major_, minor_, patch_)
+
+/*
+ * All blobs need to be declared via MODULE_FIRMWARE().
+ * This first expansion of the table macros is solely to provide
+ * that declaration.
+ */
+#define INTEL_UC_MODULE_FW(platform_, revid_, uc_) \
+	MODULE_FIRMWARE(uc_);
+
+INTEL_GUC_FIRMWARE_DEFS(INTEL_UC_MODULE_FW, MAKE_GUC_FW_PATH_MAJOR, MAKE_GUC_FW_PATH_MMP)
+INTEL_HUC_FIRMWARE_DEFS(INTEL_UC_MODULE_FW, MAKE_HUC_FW_PATH_BLANK, MAKE_HUC_FW_PATH_MMP)
+
+/*
+ * The next expansion of the table macros (in __uc_fw_auto_select below) provides
+ * actual data structures with both the filename and the version information.
+ * These structure arrays are then iterated over to the list of suitable files
+ * for the current platform and to then attempt to load those files, in the order
+ * listed, until one is successfully found.
+ */
+struct __packed uc_fw_blob {
+	const char *path;
+	bool legacy;
+	u8 major;
+	u8 minor;
+	u8 patch;
+};
+
+#define UC_FW_BLOB_BASE(major_, minor_, patch_, path_) \
+	.major = major_, \
+	.minor = minor_, \
+	.patch = patch_, \
+	.path = path_,
+
+#define UC_FW_BLOB_NEW(major_, minor_, patch_, path_) \
+	{ UC_FW_BLOB_BASE(major_, minor_, patch_, path_) \
+	  .legacy = false }
+
+#define UC_FW_BLOB_OLD(major_, minor_, patch_, path_) \
+	{ UC_FW_BLOB_BASE(major_, minor_, patch_, path_) \
+	  .legacy = true }
+
+#define GUC_FW_BLOB(prefix_, major_, minor_) \
+	UC_FW_BLOB_NEW(major_, minor_, 0, \
+		       MAKE_GUC_FW_PATH_MAJOR(prefix_, major_, minor_))
+
+#define GUC_FW_BLOB_MMP(prefix_, major_, minor_, patch_) \
+	UC_FW_BLOB_OLD(major_, minor_, patch_, \
+		       MAKE_GUC_FW_PATH_MMP(prefix_, major_, minor_, patch_))
+
+#define HUC_FW_BLOB(prefix_) \
+	UC_FW_BLOB_NEW(0, 0, 0, MAKE_HUC_FW_PATH_BLANK(prefix_))
+
+#define HUC_FW_BLOB_MMP(prefix_, major_, minor_, patch_) \
+	UC_FW_BLOB_OLD(major_, minor_, patch_, \
+		       MAKE_HUC_FW_PATH_MMP(prefix_, major_, minor_, patch_))
+
+struct __packed uc_fw_platform_requirement {
+	enum intel_platform p;
+	u8 rev; /* first platform rev using this FW */
+	const struct uc_fw_blob blob;
+};
+
+#define MAKE_FW_LIST(platform_, revid_, uc_) \
+{ \
+	.p = INTEL_##platform_, \
+	.rev = revid_, \
+	.blob = uc_, \
+},
+
+struct fw_blobs_by_type {
+	const struct uc_fw_platform_requirement *blobs;
+	u32 count;
+};
+
+static void
+__uc_fw_auto_select(struct drm_i915_private *i915, struct intel_uc_fw *uc_fw)
+{
+	static const struct uc_fw_platform_requirement blobs_guc[] = {
+		INTEL_GUC_FIRMWARE_DEFS(MAKE_FW_LIST, GUC_FW_BLOB, GUC_FW_BLOB_MMP)
+	};
+	static const struct uc_fw_platform_requirement blobs_huc[] = {
+		INTEL_HUC_FIRMWARE_DEFS(MAKE_FW_LIST, HUC_FW_BLOB, HUC_FW_BLOB_MMP)
+	};
+	static const struct fw_blobs_by_type blobs_all[INTEL_UC_FW_NUM_TYPES] = {
+		[INTEL_UC_FW_TYPE_GUC] = { blobs_guc, ARRAY_SIZE(blobs_guc) },
+		[INTEL_UC_FW_TYPE_HUC] = { blobs_huc, ARRAY_SIZE(blobs_huc) },
+	};
+	static bool verified;
+	const struct uc_fw_platform_requirement *fw_blobs;
+	enum intel_platform p = INTEL_INFO(i915)->platform;
+	u32 fw_count;
+	u8 rev = INTEL_REVID(i915);
+	int i;
+	bool found;
+
+	/*
+	 * The only difference between the ADL GuC FWs is the HWConfig support.
+	 * ADL-N does not support HWConfig, so we should use the same binary as
+	 * ADL-S, otherwise the GuC might attempt to fetch a config table that
+	 * does not exist.
+	 */
+	if (IS_ADLP_N(i915))
+		p = INTEL_ALDERLAKE_S;
+
+	GEM_BUG_ON(uc_fw->type >= ARRAY_SIZE(blobs_all));
+	fw_blobs = blobs_all[uc_fw->type].blobs;
+	fw_count = blobs_all[uc_fw->type].count;
+
+	found = false;
+	for (i = 0; i < fw_count && p <= fw_blobs[i].p; i++) {
+		const struct uc_fw_blob *blob = &fw_blobs[i].blob;
+
+		if (p != fw_blobs[i].p)
+			continue;
+
+		if (rev < fw_blobs[i].rev)
+			continue;
+
+		if (uc_fw->file_selected.path) {
+			if (uc_fw->file_selected.path == blob->path)
+				uc_fw->file_selected.path = NULL;
+
+			continue;
+		}
+
+		uc_fw->file_selected.path = blob->path;
+		uc_fw->file_wanted.path = blob->path;
+		uc_fw->file_wanted.major_ver = blob->major;
+		uc_fw->file_wanted.minor_ver = blob->minor;
+		found = true;
+		break;
+	}
+
+	if (!found && uc_fw->file_selected.path) {
+		/* Failed to find a match for the last attempt?! */
+		uc_fw->file_selected.path = NULL;
+	}
+
+	/* make sure the list is ordered as expected */
+	if (IS_ENABLED(CONFIG_DRM_I915_SELFTEST) && !verified) {
+		verified = true;
+
+		for (i = 1; i < fw_count; i++) {
+			/* Next platform is good: */
+			if (fw_blobs[i].p < fw_blobs[i - 1].p)
+				continue;
+
+			/* Next platform revision is good: */
+			if (fw_blobs[i].p == fw_blobs[i - 1].p &&
+			    fw_blobs[i].rev < fw_blobs[i - 1].rev)
+				continue;
+
+			/* Platform/revision must be in order: */
+			if (fw_blobs[i].p != fw_blobs[i - 1].p ||
+			    fw_blobs[i].rev != fw_blobs[i - 1].rev)
+				goto bad;
+
+			/* Next major version is good: */
+			if (fw_blobs[i].blob.major < fw_blobs[i - 1].blob.major)
+				continue;
+
+			/* New must be before legacy: */
+			if (!fw_blobs[i].blob.legacy && fw_blobs[i - 1].blob.legacy)
+				goto bad;
+
+			/* New to legacy also means 0.0 to X.Y (HuC), or X.0 to X.Y (GuC) */
+			if (fw_blobs[i].blob.legacy && !fw_blobs[i - 1].blob.legacy) {
+				if (!fw_blobs[i - 1].blob.major)
+					continue;
+
+				if (fw_blobs[i].blob.major == fw_blobs[i - 1].blob.major)
+					continue;
+			}
+
+			/* Major versions must be in order: */
+			if (fw_blobs[i].blob.major != fw_blobs[i - 1].blob.major)
+				goto bad;
+
+			/* Next minor version is good: */
+			if (fw_blobs[i].blob.minor < fw_blobs[i - 1].blob.minor)
+				continue;
+
+			/* Minor versions must be in order: */
+			if (fw_blobs[i].blob.minor != fw_blobs[i - 1].blob.minor)
+				goto bad;
+
+			/* Patch versions must be in order: */
+			if (fw_blobs[i].blob.patch <= fw_blobs[i - 1].blob.patch)
+				continue;
+
+bad:
+			drm_err(&i915->drm, "Invalid FW blob order: %s r%u %s%d.%d.%d comes before %s r%u %s%d.%d.%d\n",
+				intel_platform_name(fw_blobs[i - 1].p), fw_blobs[i - 1].rev,
+				fw_blobs[i - 1].blob.legacy ? "L" : "v",
+				fw_blobs[i - 1].blob.major,
+				fw_blobs[i - 1].blob.minor,
+				fw_blobs[i - 1].blob.patch,
+				intel_platform_name(fw_blobs[i].p), fw_blobs[i].rev,
+				fw_blobs[i].blob.legacy ? "L" : "v",
+				fw_blobs[i].blob.major,
+				fw_blobs[i].blob.minor,
+				fw_blobs[i].blob.patch);
+
+			uc_fw->file_selected.path = NULL;
+		}
+	}
+}
+
+static const char *__override_guc_firmware_path(struct drm_i915_private *i915)
+{
+	if (i915->params.enable_guc & ENABLE_GUC_MASK)
+		return i915->params.guc_firmware_path;
+	return "";
+}
+
+static const char *__override_huc_firmware_path(struct drm_i915_private *i915)
+{
+	if (i915->params.enable_guc & ENABLE_GUC_LOAD_HUC)
+		return i915->params.huc_firmware_path;
+	return "";
+}
+
+static void __uc_fw_user_override(struct drm_i915_private *i915, struct intel_uc_fw *uc_fw)
+{
+	const char *path = NULL;
+
+	switch (uc_fw->type) {
+	case INTEL_UC_FW_TYPE_GUC:
+		path = __override_guc_firmware_path(i915);
+		break;
+	case INTEL_UC_FW_TYPE_HUC:
+		path = __override_huc_firmware_path(i915);
+		break;
+	}
+
+	if (unlikely(path)) {
+		uc_fw->file_selected.path = path;
+		uc_fw->user_overridden = true;
+	}
+}
+
+/**
+ * intel_uc_fw_init_early - initialize the uC object and select the firmware
+ * @uc_fw: uC firmware
+ * @type: type of uC
+ *
+ * Initialize the state of our uC object and relevant tracking and select the
+ * firmware to fetch and load.
+ */
+void intel_uc_fw_init_early(struct intel_uc_fw *uc_fw,
+			    enum intel_uc_fw_type type)
+{
+	struct drm_i915_private *i915 = ____uc_fw_to_gt(uc_fw, type)->i915;
+
+	/*
+	 * we use FIRMWARE_UNINITIALIZED to detect checks against uc_fw->status
+	 * before we're looked at the HW caps to see if we have uc support
+	 */
+	BUILD_BUG_ON(INTEL_UC_FIRMWARE_UNINITIALIZED);
+	GEM_BUG_ON(uc_fw->status);
+	GEM_BUG_ON(uc_fw->file_selected.path);
+
+	uc_fw->type = type;
+
+	if (HAS_GT_UC(i915)) {
+		__uc_fw_auto_select(i915, uc_fw);
+		__uc_fw_user_override(i915, uc_fw);
+	}
+
+	intel_uc_fw_change_status(uc_fw, uc_fw->file_selected.path ? *uc_fw->file_selected.path ?
+				  INTEL_UC_FIRMWARE_SELECTED :
+				  INTEL_UC_FIRMWARE_DISABLED :
+				  INTEL_UC_FIRMWARE_NOT_SUPPORTED);
+}
+
+static void __force_fw_fetch_failures(struct intel_uc_fw *uc_fw, int e)
+{
+	struct drm_i915_private *i915 = __uc_fw_to_gt(uc_fw)->i915;
+	bool user = e == -EINVAL;
+
+	if (i915_inject_probe_error(i915, e)) {
+		/* non-existing blob */
+		uc_fw->file_selected.path = "<invalid>";
+		uc_fw->user_overridden = user;
+	} else if (i915_inject_probe_error(i915, e)) {
+		/* require next major version */
+		uc_fw->file_wanted.major_ver += 1;
+		uc_fw->file_wanted.minor_ver = 0;
+		uc_fw->user_overridden = user;
+	} else if (i915_inject_probe_error(i915, e)) {
+		/* require next minor version */
+		uc_fw->file_wanted.minor_ver += 1;
+		uc_fw->user_overridden = user;
+	} else if (uc_fw->file_wanted.major_ver &&
+		   i915_inject_probe_error(i915, e)) {
+		/* require prev major version */
+		uc_fw->file_wanted.major_ver -= 1;
+		uc_fw->file_wanted.minor_ver = 0;
+		uc_fw->user_overridden = user;
+	} else if (uc_fw->file_wanted.minor_ver &&
+		   i915_inject_probe_error(i915, e)) {
+		/* require prev minor version - hey, this should work! */
+		uc_fw->file_wanted.minor_ver -= 1;
+		uc_fw->user_overridden = user;
+	} else if (user && i915_inject_probe_error(i915, e)) {
+		/* officially unsupported platform */
+		uc_fw->file_wanted.major_ver = 0;
+		uc_fw->file_wanted.minor_ver = 0;
+		uc_fw->user_overridden = true;
+	}
+}
+
+static int check_gsc_manifest(const struct firmware *fw,
+			      struct intel_uc_fw *uc_fw)
+{
+	u32 *dw = (u32 *)fw->data;
+	u32 version_hi = dw[HUC_GSC_VERSION_HI_DW];
+	u32 version_lo = dw[HUC_GSC_VERSION_LO_DW];
+
+	uc_fw->file_selected.major_ver = FIELD_GET(HUC_GSC_MAJOR_VER_HI_MASK, version_hi);
+	uc_fw->file_selected.minor_ver = FIELD_GET(HUC_GSC_MINOR_VER_HI_MASK, version_hi);
+	uc_fw->file_selected.patch_ver = FIELD_GET(HUC_GSC_PATCH_VER_LO_MASK, version_lo);
+
+	return 0;
+}
+
+static int check_ccs_header(struct drm_i915_private *i915,
+			    const struct firmware *fw,
+			    struct intel_uc_fw *uc_fw)
+{
+	struct uc_css_header *css;
+	size_t size;
+
+	/* Check the size of the blob before examining buffer contents */
+	if (unlikely(fw->size < sizeof(struct uc_css_header))) {
+		drm_warn(&i915->drm, "%s firmware %s: invalid size: %zu < %zu\n",
+			 intel_uc_fw_type_repr(uc_fw->type), uc_fw->file_selected.path,
+			 fw->size, sizeof(struct uc_css_header));
+		return -ENODATA;
+	}
+
+	css = (struct uc_css_header *)fw->data;
+
+	/* Check integrity of size values inside CSS header */
+	size = (css->header_size_dw - css->key_size_dw - css->modulus_size_dw -
+		css->exponent_size_dw) * sizeof(u32);
+	if (unlikely(size != sizeof(struct uc_css_header))) {
+		drm_warn(&i915->drm,
+			 "%s firmware %s: unexpected header size: %zu != %zu\n",
+			 intel_uc_fw_type_repr(uc_fw->type), uc_fw->file_selected.path,
+			 fw->size, sizeof(struct uc_css_header));
+		return -EPROTO;
+	}
+
+	/* uCode size must calculated from other sizes */
+	uc_fw->ucode_size = (css->size_dw - css->header_size_dw) * sizeof(u32);
+
+	/* now RSA */
+	uc_fw->rsa_size = css->key_size_dw * sizeof(u32);
+
+	/* At least, it should have header, uCode and RSA. Size of all three. */
+	size = sizeof(struct uc_css_header) + uc_fw->ucode_size + uc_fw->rsa_size;
+	if (unlikely(fw->size < size)) {
+		drm_warn(&i915->drm, "%s firmware %s: invalid size: %zu < %zu\n",
+			 intel_uc_fw_type_repr(uc_fw->type), uc_fw->file_selected.path,
+			 fw->size, size);
+		return -ENOEXEC;
+	}
+
+	/* Sanity check whether this fw is not larger than whole WOPCM memory */
+	size = __intel_uc_fw_get_upload_size(uc_fw);
+	if (unlikely(size >= i915->wopcm.size)) {
+		drm_warn(&i915->drm, "%s firmware %s: invalid size: %zu > %zu\n",
+			 intel_uc_fw_type_repr(uc_fw->type), uc_fw->file_selected.path,
+			 size, (size_t)i915->wopcm.size);
+		return -E2BIG;
+	}
+
+	/* Get version numbers from the CSS header */
+	uc_fw->file_selected.major_ver = FIELD_GET(CSS_SW_VERSION_UC_MAJOR,
+						   css->sw_version);
+	uc_fw->file_selected.minor_ver = FIELD_GET(CSS_SW_VERSION_UC_MINOR,
+						   css->sw_version);
+	uc_fw->file_selected.patch_ver = FIELD_GET(CSS_SW_VERSION_UC_PATCH,
+						   css->sw_version);
+
+	if (uc_fw->type == INTEL_UC_FW_TYPE_GUC)
+		uc_fw->private_data_size = css->private_data_size;
+
+	return 0;
+}
+
+/**
+ * intel_uc_fw_fetch - fetch uC firmware
+ * @uc_fw: uC firmware
+ *
+ * Fetch uC firmware into GEM obj.
+ *
+ * Return: 0 on success, a negative errno code on failure.
+ */
+int intel_uc_fw_fetch(struct intel_uc_fw *uc_fw)
+{
+	struct drm_i915_private *i915 = __uc_fw_to_gt(uc_fw)->i915;
+	struct intel_uc_fw_file file_ideal;
+	struct device *dev = i915->drm.dev;
+	struct drm_i915_gem_object *obj;
+	const struct firmware *fw = NULL;
+	bool old_ver = false;
+	int err;
+
+	GEM_BUG_ON(!i915->wopcm.size);
+	GEM_BUG_ON(!intel_uc_fw_is_enabled(uc_fw));
+
+	err = i915_inject_probe_error(i915, -ENXIO);
+	if (err)
+		goto fail;
+
+	__force_fw_fetch_failures(uc_fw, -EINVAL);
+	__force_fw_fetch_failures(uc_fw, -ESTALE);
+
+	err = firmware_request_nowarn(&fw, uc_fw->file_selected.path, dev);
+	memcpy(&file_ideal, &uc_fw->file_wanted, sizeof(file_ideal));
+
+	/* Any error is terminal if overriding. Don't bother searching for older versions */
+	if (err && intel_uc_fw_is_overridden(uc_fw))
+		goto fail;
+
+	while (err == -ENOENT) {
+		old_ver = true;
+
+		__uc_fw_auto_select(i915, uc_fw);
+		if (!uc_fw->file_selected.path) {
+			/*
+			 * No more options! But set the path back to something
+			 * valid just in case it gets dereferenced.
+			 */
+			uc_fw->file_selected.path = file_ideal.path;
+
+			/* Also, preserve the version that was really wanted */
+			memcpy(&uc_fw->file_wanted, &file_ideal, sizeof(uc_fw->file_wanted));
+			break;
+		}
+
+		err = firmware_request_nowarn(&fw, uc_fw->file_selected.path, dev);
+	}
+
+	if (err)
+		goto fail;
+
+	if (uc_fw->loaded_via_gsc)
+		err = check_gsc_manifest(fw, uc_fw);
+	else
+		err = check_ccs_header(i915, fw, uc_fw);
+	if (err)
+		goto fail;
+
+	if (uc_fw->file_wanted.major_ver) {
+		/* Check the file's major version was as it claimed */
+		if (uc_fw->file_selected.major_ver != uc_fw->file_wanted.major_ver) {
+			drm_notice(&i915->drm, "%s firmware %s: unexpected version: %u.%u != %u.%u\n",
+				   intel_uc_fw_type_repr(uc_fw->type), uc_fw->file_selected.path,
+				   uc_fw->file_selected.major_ver, uc_fw->file_selected.minor_ver,
+				   uc_fw->file_wanted.major_ver, uc_fw->file_wanted.minor_ver);
+			if (!intel_uc_fw_is_overridden(uc_fw)) {
+				err = -ENOEXEC;
+				goto fail;
+			}
+		} else {
+			if (uc_fw->file_selected.minor_ver < uc_fw->file_wanted.minor_ver)
+				old_ver = true;
+		}
+	}
+
+	if (old_ver) {
+		/* Preserve the version that was really wanted */
+		memcpy(&uc_fw->file_wanted, &file_ideal, sizeof(uc_fw->file_wanted));
+
+		drm_notice(&i915->drm,
+			   "%s firmware %s (%d.%d) is recommended, but only %s (%d.%d) was found\n",
+			   intel_uc_fw_type_repr(uc_fw->type),
+			   uc_fw->file_wanted.path,
+			   uc_fw->file_wanted.major_ver, uc_fw->file_wanted.minor_ver,
+			   uc_fw->file_selected.path,
+			   uc_fw->file_selected.major_ver, uc_fw->file_selected.minor_ver);
+		drm_info(&i915->drm,
+			 "Consider updating your linux-firmware pkg or downloading from %s\n",
+			 INTEL_UC_FIRMWARE_URL);
+	}
+
+	if (HAS_LMEM(i915)) {
+		obj = i915_gem_object_create_lmem_from_data(i915, fw->data, fw->size);
+		if (!IS_ERR(obj))
+			obj->flags |= I915_BO_ALLOC_PM_EARLY;
+	} else {
+		obj = i915_gem_object_create_shmem_from_data(i915, fw->data, fw->size);
+	}
+
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto fail;
+	}
+
+	uc_fw->obj = obj;
+	uc_fw->size = fw->size;
+	intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_AVAILABLE);
+
+	release_firmware(fw);
+	return 0;
+
+fail:
+	intel_uc_fw_change_status(uc_fw, err == -ENOENT ?
+				  INTEL_UC_FIRMWARE_MISSING :
+				  INTEL_UC_FIRMWARE_ERROR);
+
+	i915_probe_error(i915, "%s firmware %s: fetch failed with error %d\n",
+			 intel_uc_fw_type_repr(uc_fw->type), uc_fw->file_selected.path, err);
+	drm_info(&i915->drm, "%s firmware(s) can be downloaded from %s\n",
+		 intel_uc_fw_type_repr(uc_fw->type), INTEL_UC_FIRMWARE_URL);
+
+	release_firmware(fw);		/* OK even if fw is NULL */
+	return err;
+}
+
+static u32 uc_fw_ggtt_offset(struct intel_uc_fw *uc_fw)
+{
+	struct i915_ggtt *ggtt = __uc_fw_to_gt(uc_fw)->ggtt;
+	struct drm_mm_node *node = &ggtt->uc_fw;
+
+	GEM_BUG_ON(!drm_mm_node_allocated(node));
+	GEM_BUG_ON(upper_32_bits(node->start));
+	GEM_BUG_ON(upper_32_bits(node->start + node->size - 1));
+
+	return lower_32_bits(node->start);
+}
+
+static void uc_fw_bind_ggtt(struct intel_uc_fw *uc_fw)
+{
+	struct drm_i915_gem_object *obj = uc_fw->obj;
+	struct i915_ggtt *ggtt = __uc_fw_to_gt(uc_fw)->ggtt;
+	struct i915_vma_resource *dummy = &uc_fw->dummy;
+	u32 pte_flags = 0;
+
+	dummy->start = uc_fw_ggtt_offset(uc_fw);
+	dummy->node_size = obj->base.size;
+	dummy->bi.pages = obj->mm.pages;
+
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+	GEM_BUG_ON(dummy->node_size > ggtt->uc_fw.size);
+
+	/* uc_fw->obj cache domains were not controlled across suspend */
+	if (i915_gem_object_has_struct_page(obj))
+		drm_clflush_sg(dummy->bi.pages);
+
+	if (i915_gem_object_is_lmem(obj))
+		pte_flags |= PTE_LM;
+
+	if (ggtt->vm.raw_insert_entries)
+		ggtt->vm.raw_insert_entries(&ggtt->vm, dummy, I915_CACHE_NONE, pte_flags);
+	else
+		ggtt->vm.insert_entries(&ggtt->vm, dummy, I915_CACHE_NONE, pte_flags);
+}
+
+static void uc_fw_unbind_ggtt(struct intel_uc_fw *uc_fw)
+{
+	struct drm_i915_gem_object *obj = uc_fw->obj;
+	struct i915_ggtt *ggtt = __uc_fw_to_gt(uc_fw)->ggtt;
+	u64 start = uc_fw_ggtt_offset(uc_fw);
+
+	ggtt->vm.clear_range(&ggtt->vm, start, obj->base.size);
+}
+
+static int uc_fw_xfer(struct intel_uc_fw *uc_fw, u32 dst_offset, u32 dma_flags)
+{
+	struct intel_gt *gt = __uc_fw_to_gt(uc_fw);
+	struct intel_uncore *uncore = gt->uncore;
+	u64 offset;
+	int ret;
+
+	ret = i915_inject_probe_error(gt->i915, -ETIMEDOUT);
+	if (ret)
+		return ret;
+
+	intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+
+	/* Set the source address for the uCode */
+	offset = uc_fw_ggtt_offset(uc_fw);
+	GEM_BUG_ON(upper_32_bits(offset) & 0xFFFF0000);
+	intel_uncore_write_fw(uncore, DMA_ADDR_0_LOW, lower_32_bits(offset));
+	intel_uncore_write_fw(uncore, DMA_ADDR_0_HIGH, upper_32_bits(offset));
+
+	/* Set the DMA destination */
+	intel_uncore_write_fw(uncore, DMA_ADDR_1_LOW, dst_offset);
+	intel_uncore_write_fw(uncore, DMA_ADDR_1_HIGH, DMA_ADDRESS_SPACE_WOPCM);
+
+	/*
+	 * Set the transfer size. The header plus uCode will be copied to WOPCM
+	 * via DMA, excluding any other components
+	 */
+	intel_uncore_write_fw(uncore, DMA_COPY_SIZE,
+			      sizeof(struct uc_css_header) + uc_fw->ucode_size);
+
+	/* Start the DMA */
+	intel_uncore_write_fw(uncore, DMA_CTRL,
+			      _MASKED_BIT_ENABLE(dma_flags | START_DMA));
+
+	/* Wait for DMA to finish */
+	ret = intel_wait_for_register_fw(uncore, DMA_CTRL, START_DMA, 0, 100);
+	if (ret)
+		drm_err(&gt->i915->drm, "DMA for %s fw failed, DMA_CTRL=%u\n",
+			intel_uc_fw_type_repr(uc_fw->type),
+			intel_uncore_read_fw(uncore, DMA_CTRL));
+
+	/* Disable the bits once DMA is over */
+	intel_uncore_write_fw(uncore, DMA_CTRL, _MASKED_BIT_DISABLE(dma_flags));
+
+	intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
+
+	return ret;
+}
+
+/**
+ * intel_uc_fw_upload - load uC firmware using custom loader
+ * @uc_fw: uC firmware
+ * @dst_offset: destination offset
+ * @dma_flags: flags for flags for dma ctrl
+ *
+ * Loads uC firmware and updates internal flags.
+ *
+ * Return: 0 on success, non-zero on failure.
+ */
+int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, u32 dst_offset, u32 dma_flags)
+{
+	struct intel_gt *gt = __uc_fw_to_gt(uc_fw);
+	int err;
+
+	/* make sure the status was cleared the last time we reset the uc */
+	GEM_BUG_ON(intel_uc_fw_is_loaded(uc_fw));
+
+	err = i915_inject_probe_error(gt->i915, -ENOEXEC);
+	if (err)
+		return err;
+
+	if (!intel_uc_fw_is_loadable(uc_fw))
+		return -ENOEXEC;
+
+	/* Call custom loader */
+	uc_fw_bind_ggtt(uc_fw);
+	err = uc_fw_xfer(uc_fw, dst_offset, dma_flags);
+	uc_fw_unbind_ggtt(uc_fw);
+	if (err)
+		goto fail;
+
+	intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_TRANSFERRED);
+	return 0;
+
+fail:
+	i915_probe_error(gt->i915, "Failed to load %s firmware %s (%d)\n",
+			 intel_uc_fw_type_repr(uc_fw->type), uc_fw->file_selected.path,
+			 err);
+	intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_LOAD_FAIL);
+	return err;
+}
+
+static inline bool uc_fw_need_rsa_in_memory(struct intel_uc_fw *uc_fw)
+{
+	/*
+	 * The HW reads the GuC RSA from memory if the key size is > 256 bytes,
+	 * while it reads it from the 64 RSA registers if it is smaller.
+	 * The HuC RSA is always read from memory.
+	 */
+	return uc_fw->type == INTEL_UC_FW_TYPE_HUC || uc_fw->rsa_size > 256;
+}
+
+static int uc_fw_rsa_data_create(struct intel_uc_fw *uc_fw)
+{
+	struct intel_gt *gt = __uc_fw_to_gt(uc_fw);
+	struct i915_vma *vma;
+	size_t copied;
+	void *vaddr;
+	int err;
+
+	err = i915_inject_probe_error(gt->i915, -ENXIO);
+	if (err)
+		return err;
+
+	if (!uc_fw_need_rsa_in_memory(uc_fw))
+		return 0;
+
+	/*
+	 * uC firmwares will sit above GUC_GGTT_TOP and will not map through
+	 * GGTT. Unfortunately, this means that the GuC HW cannot perform the uC
+	 * authentication from memory, as the RSA offset now falls within the
+	 * GuC inaccessible range. We resort to perma-pinning an additional vma
+	 * within the accessible range that only contains the RSA signature.
+	 * The GuC HW can use this extra pinning to perform the authentication
+	 * since its GGTT offset will be GuC accessible.
+	 */
+	GEM_BUG_ON(uc_fw->rsa_size > PAGE_SIZE);
+	vma = intel_guc_allocate_vma(&gt->uc.guc, PAGE_SIZE);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	vaddr = i915_gem_object_pin_map_unlocked(vma->obj,
+						 i915_coherent_map_type(gt->i915, vma->obj, true));
+	if (IS_ERR(vaddr)) {
+		i915_vma_unpin_and_release(&vma, 0);
+		err = PTR_ERR(vaddr);
+		goto unpin_out;
+	}
+
+	copied = intel_uc_fw_copy_rsa(uc_fw, vaddr, vma->size);
+	i915_gem_object_unpin_map(vma->obj);
+
+	if (copied < uc_fw->rsa_size) {
+		err = -ENOMEM;
+		goto unpin_out;
+	}
+
+	uc_fw->rsa_data = vma;
+
+	return 0;
+
+unpin_out:
+	i915_vma_unpin_and_release(&vma, 0);
+	return err;
+}
+
+static void uc_fw_rsa_data_destroy(struct intel_uc_fw *uc_fw)
+{
+	i915_vma_unpin_and_release(&uc_fw->rsa_data, 0);
+}
+
+int intel_uc_fw_init(struct intel_uc_fw *uc_fw)
+{
+	int err;
+
+	/* this should happen before the load! */
+	GEM_BUG_ON(intel_uc_fw_is_loaded(uc_fw));
+
+	if (!intel_uc_fw_is_available(uc_fw))
+		return -ENOEXEC;
+
+	err = i915_gem_object_pin_pages_unlocked(uc_fw->obj);
+	if (err) {
+		DRM_DEBUG_DRIVER("%s fw pin-pages err=%d\n",
+				 intel_uc_fw_type_repr(uc_fw->type), err);
+		goto out;
+	}
+
+	err = uc_fw_rsa_data_create(uc_fw);
+	if (err) {
+		DRM_DEBUG_DRIVER("%s fw rsa data creation failed, err=%d\n",
+				 intel_uc_fw_type_repr(uc_fw->type), err);
+		goto out_unpin;
+	}
+
+	return 0;
+
+out_unpin:
+	i915_gem_object_unpin_pages(uc_fw->obj);
+out:
+	intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_INIT_FAIL);
+	return err;
+}
+
+void intel_uc_fw_fini(struct intel_uc_fw *uc_fw)
+{
+	uc_fw_rsa_data_destroy(uc_fw);
+
+	if (i915_gem_object_has_pinned_pages(uc_fw->obj))
+		i915_gem_object_unpin_pages(uc_fw->obj);
+
+	intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_AVAILABLE);
+}
+
+/**
+ * intel_uc_fw_cleanup_fetch - cleanup uC firmware
+ * @uc_fw: uC firmware
+ *
+ * Cleans up uC firmware by releasing the firmware GEM obj.
+ */
+void intel_uc_fw_cleanup_fetch(struct intel_uc_fw *uc_fw)
+{
+	if (!intel_uc_fw_is_available(uc_fw))
+		return;
+
+	i915_gem_object_put(fetch_and_zero(&uc_fw->obj));
+
+	intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_SELECTED);
+}
+
+/**
+ * intel_uc_fw_copy_rsa - copy fw RSA to buffer
+ *
+ * @uc_fw: uC firmware
+ * @dst: dst buffer
+ * @max_len: max number of bytes to copy
+ *
+ * Return: number of copied bytes.
+ */
+size_t intel_uc_fw_copy_rsa(struct intel_uc_fw *uc_fw, void *dst, u32 max_len)
+{
+	struct intel_memory_region *mr = uc_fw->obj->mm.region;
+	u32 size = min_t(u32, uc_fw->rsa_size, max_len);
+	u32 offset = sizeof(struct uc_css_header) + uc_fw->ucode_size;
+	struct sgt_iter iter;
+	size_t count = 0;
+	int idx;
+
+	/* Called during reset handling, must be atomic [no fs_reclaim] */
+	GEM_BUG_ON(!intel_uc_fw_is_available(uc_fw));
+
+	idx = offset >> PAGE_SHIFT;
+	offset = offset_in_page(offset);
+	if (i915_gem_object_has_struct_page(uc_fw->obj)) {
+		struct page *page;
+
+		for_each_sgt_page(page, iter, uc_fw->obj->mm.pages) {
+			u32 len = min_t(u32, size, PAGE_SIZE - offset);
+			void *vaddr;
+
+			if (idx > 0) {
+				idx--;
+				continue;
+			}
+
+			vaddr = kmap_atomic(page);
+			memcpy(dst, vaddr + offset, len);
+			kunmap_atomic(vaddr);
+
+			offset = 0;
+			dst += len;
+			size -= len;
+			count += len;
+			if (!size)
+				break;
+		}
+	} else {
+		dma_addr_t addr;
+
+		for_each_sgt_daddr(addr, iter, uc_fw->obj->mm.pages) {
+			u32 len = min_t(u32, size, PAGE_SIZE - offset);
+			void __iomem *vaddr;
+
+			if (idx > 0) {
+				idx--;
+				continue;
+			}
+
+			vaddr = io_mapping_map_atomic_wc(&mr->iomap,
+							 addr - mr->region.start);
+			memcpy_fromio(dst, vaddr + offset, len);
+			io_mapping_unmap_atomic(vaddr);
+
+			offset = 0;
+			dst += len;
+			size -= len;
+			count += len;
+			if (!size)
+				break;
+		}
+	}
+
+	return count;
+}
+
+/**
+ * intel_uc_fw_dump - dump information about uC firmware
+ * @uc_fw: uC firmware
+ * @p: the &drm_printer
+ *
+ * Pretty printer for uC firmware.
+ */
+void intel_uc_fw_dump(const struct intel_uc_fw *uc_fw, struct drm_printer *p)
+{
+	u32 ver_sel, ver_want;
+
+	drm_printf(p, "%s firmware: %s\n",
+		   intel_uc_fw_type_repr(uc_fw->type), uc_fw->file_selected.path);
+	if (uc_fw->file_selected.path != uc_fw->file_wanted.path)
+		drm_printf(p, "%s firmware wanted: %s\n",
+			   intel_uc_fw_type_repr(uc_fw->type), uc_fw->file_wanted.path);
+	drm_printf(p, "\tstatus: %s\n",
+		   intel_uc_fw_status_repr(uc_fw->status));
+	ver_sel = MAKE_UC_VER(uc_fw->file_selected.major_ver,
+			      uc_fw->file_selected.minor_ver,
+			      uc_fw->file_selected.patch_ver);
+	ver_want = MAKE_UC_VER(uc_fw->file_wanted.major_ver,
+			       uc_fw->file_wanted.minor_ver,
+			       uc_fw->file_wanted.patch_ver);
+	if (ver_sel < ver_want)
+		drm_printf(p, "\tversion: wanted %u.%u.%u, found %u.%u.%u\n",
+			   uc_fw->file_wanted.major_ver,
+			   uc_fw->file_wanted.minor_ver,
+			   uc_fw->file_wanted.patch_ver,
+			   uc_fw->file_selected.major_ver,
+			   uc_fw->file_selected.minor_ver,
+			   uc_fw->file_selected.patch_ver);
+	else
+		drm_printf(p, "\tversion: found %u.%u.%u\n",
+			   uc_fw->file_selected.major_ver,
+			   uc_fw->file_selected.minor_ver,
+			   uc_fw->file_selected.patch_ver);
+	drm_printf(p, "\tuCode: %u bytes\n", uc_fw->ucode_size);
+	drm_printf(p, "\tRSA: %u bytes\n", uc_fw->rsa_size);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h
new file mode 100644
index 000000000..cb586f7df
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.h
@@ -0,0 +1,278 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2014-2019 Intel Corporation
+ */
+
+#ifndef _INTEL_UC_FW_H_
+#define _INTEL_UC_FW_H_
+
+#include <linux/types.h>
+#include "intel_uc_fw_abi.h"
+#include "intel_device_info.h"
+#include "i915_gem.h"
+#include "i915_vma.h"
+
+struct drm_printer;
+struct drm_i915_private;
+struct intel_gt;
+
+/* Home of GuC, HuC and DMC firmwares */
+#define INTEL_UC_FIRMWARE_URL "https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/i915"
+
+/*
+ * +------------+---------------------------------------------------+
+ * |   PHASE    |           FIRMWARE STATUS TRANSITIONS             |
+ * +============+===================================================+
+ * |            |               UNINITIALIZED                       |
+ * +------------+-               /   |   \                         -+
+ * |            |   DISABLED <--/    |    \--> NOT_SUPPORTED        |
+ * | init_early |                    V                              |
+ * |            |                 SELECTED                          |
+ * +------------+-               /   |   \                         -+
+ * |            |    MISSING <--/    |    \--> ERROR                |
+ * |   fetch    |                    V                              |
+ * |            |                 AVAILABLE                         |
+ * +------------+-                   |   \                         -+
+ * |            |                    |    \--> INIT FAIL            |
+ * |   init     |                    V                              |
+ * |            |        /------> LOADABLE <----<-----------\       |
+ * +------------+-       \         /    \        \           \     -+
+ * |            |    LOAD FAIL <--<      \--> TRANSFERRED     \     |
+ * |   upload   |                  \           /   \          /     |
+ * |            |                   \---------/     \--> RUNNING    |
+ * +------------+---------------------------------------------------+
+ */
+
+enum intel_uc_fw_status {
+	INTEL_UC_FIRMWARE_NOT_SUPPORTED = -1, /* no uc HW */
+	INTEL_UC_FIRMWARE_UNINITIALIZED = 0, /* used to catch checks done too early */
+	INTEL_UC_FIRMWARE_DISABLED, /* disabled */
+	INTEL_UC_FIRMWARE_SELECTED, /* selected the blob we want to load */
+	INTEL_UC_FIRMWARE_MISSING, /* blob not found on the system */
+	INTEL_UC_FIRMWARE_ERROR, /* invalid format or version */
+	INTEL_UC_FIRMWARE_AVAILABLE, /* blob found and copied in mem */
+	INTEL_UC_FIRMWARE_INIT_FAIL, /* failed to prepare fw objects for load */
+	INTEL_UC_FIRMWARE_LOADABLE, /* all fw-required objects are ready */
+	INTEL_UC_FIRMWARE_LOAD_FAIL, /* failed to xfer or init/auth the fw */
+	INTEL_UC_FIRMWARE_TRANSFERRED, /* dma xfer done */
+	INTEL_UC_FIRMWARE_RUNNING /* init/auth done */
+};
+
+enum intel_uc_fw_type {
+	INTEL_UC_FW_TYPE_GUC = 0,
+	INTEL_UC_FW_TYPE_HUC
+};
+#define INTEL_UC_FW_NUM_TYPES 2
+
+/*
+ * The firmware build process will generate a version header file with major and
+ * minor version defined. The versions are built into CSS header of firmware.
+ * i915 kernel driver set the minimal firmware version required per platform.
+ */
+struct intel_uc_fw_file {
+	const char *path;
+	u16 major_ver;
+	u16 minor_ver;
+	u16 patch_ver;
+};
+
+/*
+ * This structure encapsulates all the data needed during the process
+ * of fetching, caching, and loading the firmware image into the uC.
+ */
+struct intel_uc_fw {
+	enum intel_uc_fw_type type;
+	union {
+		const enum intel_uc_fw_status status;
+		enum intel_uc_fw_status __status; /* no accidental overwrites */
+	};
+	struct intel_uc_fw_file file_wanted;
+	struct intel_uc_fw_file file_selected;
+	bool user_overridden;
+	size_t size;
+	struct drm_i915_gem_object *obj;
+
+	/**
+	 * @dummy: A vma used in binding the uc fw to ggtt. We can't define this
+	 * vma on the stack as it can lead to a stack overflow, so we define it
+	 * here. Safe to have 1 copy per uc fw because the binding is single
+	 * threaded as it done during driver load (inherently single threaded)
+	 * or during a GT reset (mutex guarantees single threaded).
+	 */
+	struct i915_vma_resource dummy;
+	struct i915_vma *rsa_data;
+
+	u32 rsa_size;
+	u32 ucode_size;
+	u32 private_data_size;
+
+	bool loaded_via_gsc;
+};
+
+#define MAKE_UC_VER(maj, min, pat)	((pat) | ((min) << 8) | ((maj) << 16))
+#define GET_UC_VER(uc)			(MAKE_UC_VER((uc)->fw.file_selected.major_ver, \
+						     (uc)->fw.file_selected.minor_ver, \
+						     (uc)->fw.file_selected.patch_ver))
+
+#ifdef CONFIG_DRM_I915_DEBUG_GUC
+void intel_uc_fw_change_status(struct intel_uc_fw *uc_fw,
+			       enum intel_uc_fw_status status);
+#else
+static inline void intel_uc_fw_change_status(struct intel_uc_fw *uc_fw,
+					     enum intel_uc_fw_status status)
+{
+	uc_fw->__status = status;
+}
+#endif
+
+static inline
+const char *intel_uc_fw_status_repr(enum intel_uc_fw_status status)
+{
+	switch (status) {
+	case INTEL_UC_FIRMWARE_NOT_SUPPORTED:
+		return "N/A";
+	case INTEL_UC_FIRMWARE_UNINITIALIZED:
+		return "UNINITIALIZED";
+	case INTEL_UC_FIRMWARE_DISABLED:
+		return "DISABLED";
+	case INTEL_UC_FIRMWARE_SELECTED:
+		return "SELECTED";
+	case INTEL_UC_FIRMWARE_MISSING:
+		return "MISSING";
+	case INTEL_UC_FIRMWARE_ERROR:
+		return "ERROR";
+	case INTEL_UC_FIRMWARE_AVAILABLE:
+		return "AVAILABLE";
+	case INTEL_UC_FIRMWARE_INIT_FAIL:
+		return "INIT FAIL";
+	case INTEL_UC_FIRMWARE_LOADABLE:
+		return "LOADABLE";
+	case INTEL_UC_FIRMWARE_LOAD_FAIL:
+		return "LOAD FAIL";
+	case INTEL_UC_FIRMWARE_TRANSFERRED:
+		return "TRANSFERRED";
+	case INTEL_UC_FIRMWARE_RUNNING:
+		return "RUNNING";
+	}
+	return "<invalid>";
+}
+
+static inline int intel_uc_fw_status_to_error(enum intel_uc_fw_status status)
+{
+	switch (status) {
+	case INTEL_UC_FIRMWARE_NOT_SUPPORTED:
+		return -ENODEV;
+	case INTEL_UC_FIRMWARE_UNINITIALIZED:
+		return -EACCES;
+	case INTEL_UC_FIRMWARE_DISABLED:
+		return -EPERM;
+	case INTEL_UC_FIRMWARE_MISSING:
+		return -ENOENT;
+	case INTEL_UC_FIRMWARE_ERROR:
+		return -ENOEXEC;
+	case INTEL_UC_FIRMWARE_INIT_FAIL:
+	case INTEL_UC_FIRMWARE_LOAD_FAIL:
+		return -EIO;
+	case INTEL_UC_FIRMWARE_SELECTED:
+		return -ESTALE;
+	case INTEL_UC_FIRMWARE_AVAILABLE:
+	case INTEL_UC_FIRMWARE_LOADABLE:
+	case INTEL_UC_FIRMWARE_TRANSFERRED:
+	case INTEL_UC_FIRMWARE_RUNNING:
+		return 0;
+	}
+	return -EINVAL;
+}
+
+static inline const char *intel_uc_fw_type_repr(enum intel_uc_fw_type type)
+{
+	switch (type) {
+	case INTEL_UC_FW_TYPE_GUC:
+		return "GuC";
+	case INTEL_UC_FW_TYPE_HUC:
+		return "HuC";
+	}
+	return "uC";
+}
+
+static inline enum intel_uc_fw_status
+__intel_uc_fw_status(struct intel_uc_fw *uc_fw)
+{
+	/* shouldn't call this before checking hw/blob availability */
+	GEM_BUG_ON(uc_fw->status == INTEL_UC_FIRMWARE_UNINITIALIZED);
+	return uc_fw->status;
+}
+
+static inline bool intel_uc_fw_is_supported(struct intel_uc_fw *uc_fw)
+{
+	return __intel_uc_fw_status(uc_fw) != INTEL_UC_FIRMWARE_NOT_SUPPORTED;
+}
+
+static inline bool intel_uc_fw_is_enabled(struct intel_uc_fw *uc_fw)
+{
+	return __intel_uc_fw_status(uc_fw) > INTEL_UC_FIRMWARE_DISABLED;
+}
+
+static inline bool intel_uc_fw_is_available(struct intel_uc_fw *uc_fw)
+{
+	return __intel_uc_fw_status(uc_fw) >= INTEL_UC_FIRMWARE_AVAILABLE;
+}
+
+static inline bool intel_uc_fw_is_loadable(struct intel_uc_fw *uc_fw)
+{
+	return __intel_uc_fw_status(uc_fw) >= INTEL_UC_FIRMWARE_LOADABLE;
+}
+
+static inline bool intel_uc_fw_is_loaded(struct intel_uc_fw *uc_fw)
+{
+	return __intel_uc_fw_status(uc_fw) >= INTEL_UC_FIRMWARE_TRANSFERRED;
+}
+
+static inline bool intel_uc_fw_is_running(struct intel_uc_fw *uc_fw)
+{
+	return __intel_uc_fw_status(uc_fw) == INTEL_UC_FIRMWARE_RUNNING;
+}
+
+static inline bool intel_uc_fw_is_overridden(const struct intel_uc_fw *uc_fw)
+{
+	return uc_fw->user_overridden;
+}
+
+static inline void intel_uc_fw_sanitize(struct intel_uc_fw *uc_fw)
+{
+	if (intel_uc_fw_is_loaded(uc_fw))
+		intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_LOADABLE);
+}
+
+static inline u32 __intel_uc_fw_get_upload_size(struct intel_uc_fw *uc_fw)
+{
+	return sizeof(struct uc_css_header) + uc_fw->ucode_size;
+}
+
+/**
+ * intel_uc_fw_get_upload_size() - Get size of firmware needed to be uploaded.
+ * @uc_fw: uC firmware.
+ *
+ * Get the size of the firmware and header that will be uploaded to WOPCM.
+ *
+ * Return: Upload firmware size, or zero on firmware fetch failure.
+ */
+static inline u32 intel_uc_fw_get_upload_size(struct intel_uc_fw *uc_fw)
+{
+	if (!intel_uc_fw_is_available(uc_fw))
+		return 0;
+
+	return __intel_uc_fw_get_upload_size(uc_fw);
+}
+
+void intel_uc_fw_init_early(struct intel_uc_fw *uc_fw,
+			    enum intel_uc_fw_type type);
+int intel_uc_fw_fetch(struct intel_uc_fw *uc_fw);
+void intel_uc_fw_cleanup_fetch(struct intel_uc_fw *uc_fw);
+int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, u32 offset, u32 dma_flags);
+int intel_uc_fw_init(struct intel_uc_fw *uc_fw);
+void intel_uc_fw_fini(struct intel_uc_fw *uc_fw);
+size_t intel_uc_fw_copy_rsa(struct intel_uc_fw *uc_fw, void *dst, u32 max_len);
+void intel_uc_fw_dump(const struct intel_uc_fw *uc_fw, struct drm_printer *p);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw_abi.h b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw_abi.h
new file mode 100644
index 000000000..7a411178b
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw_abi.h
@@ -0,0 +1,92 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef _INTEL_UC_FW_ABI_H
+#define _INTEL_UC_FW_ABI_H
+
+#include <linux/types.h>
+#include <linux/build_bug.h>
+
+/**
+ * DOC: Firmware Layout
+ *
+ * The GuC/HuC firmware layout looks like this::
+ *
+ *      +======================================================================+
+ *      |  Firmware blob                                                       |
+ *      +===============+===============+============+============+============+
+ *      |  CSS header   |     uCode     |  RSA key   |  modulus   |  exponent  |
+ *      +===============+===============+============+============+============+
+ *       <-header size->                 <---header size continued ----------->
+ *       <--- size ----------------------------------------------------------->
+ *                                       <-key size->
+ *                                                    <-mod size->
+ *                                                                 <-exp size->
+ *
+ * The firmware may or may not have modulus key and exponent data. The header,
+ * uCode and RSA signature are must-have components that will be used by driver.
+ * Length of each components, which is all in dwords, can be found in header.
+ * In the case that modulus and exponent are not present in fw, a.k.a truncated
+ * image, the length value still appears in header.
+ *
+ * Driver will do some basic fw size validation based on the following rules:
+ *
+ * 1. Header, uCode and RSA are must-have components.
+ * 2. All firmware components, if they present, are in the sequence illustrated
+ *    in the layout table above.
+ * 3. Length info of each component can be found in header, in dwords.
+ * 4. Modulus and exponent key are not required by driver. They may not appear
+ *    in fw. So driver will load a truncated firmware in this case.
+ *
+ * Starting from DG2, the HuC is loaded by the GSC instead of i915. The GSC
+ * firmware performs all the required integrity checks, we just need to check
+ * the version. Note that the header for GSC-managed blobs is different from the
+ * CSS used for dma-loaded firmwares.
+ */
+
+struct uc_css_header {
+	u32 module_type;
+	/*
+	 * header_size includes all non-uCode bits, including css_header, rsa
+	 * key, modulus key and exponent data.
+	 */
+	u32 header_size_dw;
+	u32 header_version;
+	u32 module_id;
+	u32 module_vendor;
+	u32 date;
+#define CSS_DATE_DAY			(0xFF << 0)
+#define CSS_DATE_MONTH			(0xFF << 8)
+#define CSS_DATE_YEAR			(0xFFFF << 16)
+	u32 size_dw; /* uCode plus header_size_dw */
+	u32 key_size_dw;
+	u32 modulus_size_dw;
+	u32 exponent_size_dw;
+	u32 time;
+#define CSS_TIME_HOUR			(0xFF << 0)
+#define CSS_DATE_MIN			(0xFF << 8)
+#define CSS_DATE_SEC			(0xFFFF << 16)
+	char username[8];
+	char buildnumber[12];
+	u32 sw_version;
+#define CSS_SW_VERSION_UC_MAJOR		(0xFF << 16)
+#define CSS_SW_VERSION_UC_MINOR		(0xFF << 8)
+#define CSS_SW_VERSION_UC_PATCH		(0xFF << 0)
+	u32 reserved0[13];
+	union {
+		u32 private_data_size; /* only applies to GuC */
+		u32 reserved1;
+	};
+	u32 header_info;
+} __packed;
+static_assert(sizeof(struct uc_css_header) == 128);
+
+#define HUC_GSC_VERSION_HI_DW		44
+#define   HUC_GSC_MAJOR_VER_HI_MASK	(0xFF << 0)
+#define   HUC_GSC_MINOR_VER_HI_MASK	(0xFF << 16)
+#define HUC_GSC_VERSION_LO_DW		45
+#define   HUC_GSC_PATCH_VER_LO_MASK	(0xFF << 0)
+
+#endif /* _INTEL_UC_FW_ABI_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/selftest_guc.c b/drivers/gpu/drm/i915/gt/uc/selftest_guc.c
new file mode 100644
index 000000000..e28518fe8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/selftest_guc.c
@@ -0,0 +1,302 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright �� 2021 Intel Corporation
+ */
+
+#include "selftests/igt_spinner.h"
+#include "selftests/intel_scheduler_helpers.h"
+
+static int request_add_spin(struct i915_request *rq, struct igt_spinner *spin)
+{
+	int err = 0;
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (spin && !igt_wait_for_spinner(spin, rq))
+		err = -ETIMEDOUT;
+
+	return err;
+}
+
+static struct i915_request *nop_user_request(struct intel_context *ce,
+					     struct i915_request *from)
+{
+	struct i915_request *rq;
+	int ret;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return rq;
+
+	if (from) {
+		ret = i915_sw_fence_await_dma_fence(&rq->submit,
+						    &from->fence, 0,
+						    I915_FENCE_GFP);
+		if (ret < 0) {
+			i915_request_put(rq);
+			return ERR_PTR(ret);
+		}
+	}
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	return rq;
+}
+
+static int intel_guc_scrub_ctbs(void *arg)
+{
+	struct intel_gt *gt = arg;
+	int ret = 0;
+	int i;
+	struct i915_request *last[3] = {NULL, NULL, NULL}, *rq;
+	intel_wakeref_t wakeref;
+	struct intel_engine_cs *engine;
+	struct intel_context *ce;
+
+	if (!intel_has_gpu_reset(gt))
+		return 0;
+
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+	engine = intel_selftest_find_any_engine(gt);
+
+	/* Submit requests and inject errors forcing G2H to be dropped */
+	for (i = 0; i < 3; ++i) {
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			ret = PTR_ERR(ce);
+			drm_err(&gt->i915->drm, "Failed to create context, %d: %d\n", i, ret);
+			goto err;
+		}
+
+		switch (i) {
+		case 0:
+			ce->drop_schedule_enable = true;
+			break;
+		case 1:
+			ce->drop_schedule_disable = true;
+			break;
+		case 2:
+			ce->drop_deregister = true;
+			break;
+		}
+
+		rq = nop_user_request(ce, NULL);
+		intel_context_put(ce);
+
+		if (IS_ERR(rq)) {
+			ret = PTR_ERR(rq);
+			drm_err(&gt->i915->drm, "Failed to create request, %d: %d\n", i, ret);
+			goto err;
+		}
+
+		last[i] = rq;
+	}
+
+	for (i = 0; i < 3; ++i) {
+		ret = i915_request_wait(last[i], 0, HZ);
+		if (ret < 0) {
+			drm_err(&gt->i915->drm, "Last request failed to complete: %d\n", ret);
+			goto err;
+		}
+		i915_request_put(last[i]);
+		last[i] = NULL;
+	}
+
+	/* Force all H2G / G2H to be submitted / processed */
+	intel_gt_retire_requests(gt);
+	msleep(500);
+
+	/* Scrub missing G2H */
+	intel_gt_handle_error(engine->gt, -1, 0, "selftest reset");
+
+	/* GT will not idle if G2H are lost */
+	ret = intel_gt_wait_for_idle(gt, HZ);
+	if (ret < 0) {
+		drm_err(&gt->i915->drm, "GT failed to idle: %d\n", ret);
+		goto err;
+	}
+
+err:
+	for (i = 0; i < 3; ++i)
+		if (last[i])
+			i915_request_put(last[i]);
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+
+	return ret;
+}
+
+/*
+ * intel_guc_steal_guc_ids - Test to exhaust all guc_ids and then steal one
+ *
+ * This test creates a spinner which is used to block all subsequent submissions
+ * until it completes. Next, a loop creates a context and a NOP request each
+ * iteration until the guc_ids are exhausted (request creation returns -EAGAIN).
+ * The spinner is ended, unblocking all requests created in the loop. At this
+ * point all guc_ids are exhausted but are available to steal. Try to create
+ * another request which should successfully steal a guc_id. Wait on last
+ * request to complete, idle GPU, verify a guc_id was stolen via a counter, and
+ * exit the test. Test also artificially reduces the number of guc_ids so the
+ * test runs in a timely manner.
+ */
+static int intel_guc_steal_guc_ids(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_guc *guc = &gt->uc.guc;
+	int ret, sv, context_index = 0;
+	intel_wakeref_t wakeref;
+	struct intel_engine_cs *engine;
+	struct intel_context **ce;
+	struct igt_spinner spin;
+	struct i915_request *spin_rq = NULL, *rq, *last = NULL;
+	int number_guc_id_stolen = guc->number_guc_id_stolen;
+
+	ce = kcalloc(GUC_MAX_CONTEXT_ID, sizeof(*ce), GFP_KERNEL);
+	if (!ce) {
+		drm_err(&gt->i915->drm, "Context array allocation failed\n");
+		return -ENOMEM;
+	}
+
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+	engine = intel_selftest_find_any_engine(gt);
+	sv = guc->submission_state.num_guc_ids;
+	guc->submission_state.num_guc_ids = 512;
+
+	/* Create spinner to block requests in below loop */
+	ce[context_index] = intel_context_create(engine);
+	if (IS_ERR(ce[context_index])) {
+		ret = PTR_ERR(ce[context_index]);
+		ce[context_index] = NULL;
+		drm_err(&gt->i915->drm, "Failed to create context: %d\n", ret);
+		goto err_wakeref;
+	}
+	ret = igt_spinner_init(&spin, engine->gt);
+	if (ret) {
+		drm_err(&gt->i915->drm, "Failed to create spinner: %d\n", ret);
+		goto err_contexts;
+	}
+	spin_rq = igt_spinner_create_request(&spin, ce[context_index],
+					     MI_ARB_CHECK);
+	if (IS_ERR(spin_rq)) {
+		ret = PTR_ERR(spin_rq);
+		drm_err(&gt->i915->drm, "Failed to create spinner request: %d\n", ret);
+		goto err_contexts;
+	}
+	ret = request_add_spin(spin_rq, &spin);
+	if (ret) {
+		drm_err(&gt->i915->drm, "Failed to add Spinner request: %d\n", ret);
+		goto err_spin_rq;
+	}
+
+	/* Use all guc_ids */
+	while (ret != -EAGAIN) {
+		ce[++context_index] = intel_context_create(engine);
+		if (IS_ERR(ce[context_index])) {
+			ret = PTR_ERR(ce[context_index--]);
+			ce[context_index] = NULL;
+			drm_err(&gt->i915->drm, "Failed to create context: %d\n", ret);
+			goto err_spin_rq;
+		}
+
+		rq = nop_user_request(ce[context_index], spin_rq);
+		if (IS_ERR(rq)) {
+			ret = PTR_ERR(rq);
+			rq = NULL;
+			if (ret != -EAGAIN) {
+				drm_err(&gt->i915->drm, "Failed to create request, %d: %d\n",
+					context_index, ret);
+				goto err_spin_rq;
+			}
+		} else {
+			if (last)
+				i915_request_put(last);
+			last = rq;
+		}
+	}
+
+	/* Release blocked requests */
+	igt_spinner_end(&spin);
+	ret = intel_selftest_wait_for_rq(spin_rq);
+	if (ret) {
+		drm_err(&gt->i915->drm, "Spin request failed to complete: %d\n", ret);
+		i915_request_put(last);
+		goto err_spin_rq;
+	}
+	i915_request_put(spin_rq);
+	igt_spinner_fini(&spin);
+	spin_rq = NULL;
+
+	/* Wait for last request */
+	ret = i915_request_wait(last, 0, HZ * 30);
+	i915_request_put(last);
+	if (ret < 0) {
+		drm_err(&gt->i915->drm, "Last request failed to complete: %d\n", ret);
+		goto err_spin_rq;
+	}
+
+	/* Try to steal guc_id */
+	rq = nop_user_request(ce[context_index], NULL);
+	if (IS_ERR(rq)) {
+		ret = PTR_ERR(rq);
+		drm_err(&gt->i915->drm, "Failed to steal guc_id, %d: %d\n", context_index, ret);
+		goto err_spin_rq;
+	}
+
+	/* Wait for request with stolen guc_id */
+	ret = i915_request_wait(rq, 0, HZ);
+	i915_request_put(rq);
+	if (ret < 0) {
+		drm_err(&gt->i915->drm, "Request with stolen guc_id failed to complete: %d\n", ret);
+		goto err_spin_rq;
+	}
+
+	/* Wait for idle */
+	ret = intel_gt_wait_for_idle(gt, HZ * 30);
+	if (ret < 0) {
+		drm_err(&gt->i915->drm, "GT failed to idle: %d\n", ret);
+		goto err_spin_rq;
+	}
+
+	/* Verify a guc_id was stolen */
+	if (guc->number_guc_id_stolen == number_guc_id_stolen) {
+		drm_err(&gt->i915->drm, "No guc_id was stolen");
+		ret = -EINVAL;
+	} else {
+		ret = 0;
+	}
+
+err_spin_rq:
+	if (spin_rq) {
+		igt_spinner_end(&spin);
+		intel_selftest_wait_for_rq(spin_rq);
+		i915_request_put(spin_rq);
+		igt_spinner_fini(&spin);
+		intel_gt_wait_for_idle(gt, HZ * 30);
+	}
+err_contexts:
+	for (; context_index >= 0 && ce[context_index]; --context_index)
+		intel_context_put(ce[context_index]);
+err_wakeref:
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+	kfree(ce);
+	guc->submission_state.num_guc_ids = sv;
+
+	return ret;
+}
+
+int intel_guc_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(intel_guc_scrub_ctbs),
+		SUBTEST(intel_guc_steal_guc_ids),
+	};
+	struct intel_gt *gt = to_gt(i915);
+
+	if (intel_gt_is_wedged(gt))
+		return 0;
+
+	if (!intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	return intel_gt_live_subtests(tests, gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/selftest_guc_hangcheck.c b/drivers/gpu/drm/i915/gt/uc/selftest_guc_hangcheck.c
new file mode 100644
index 000000000..01f8cd3c3
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/selftest_guc_hangcheck.c
@@ -0,0 +1,159 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include "selftests/igt_spinner.h"
+#include "selftests/igt_reset.h"
+#include "selftests/intel_scheduler_helpers.h"
+#include "gt/intel_engine_heartbeat.h"
+#include "gem/selftests/mock_context.h"
+
+#define BEAT_INTERVAL	100
+
+static struct i915_request *nop_request(struct intel_engine_cs *engine)
+{
+	struct i915_request *rq;
+
+	rq = intel_engine_create_kernel_request(engine);
+	if (IS_ERR(rq))
+		return rq;
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	return rq;
+}
+
+static int intel_hang_guc(void *arg)
+{
+	struct intel_gt *gt = arg;
+	int ret = 0;
+	struct i915_gem_context *ctx;
+	struct intel_context *ce;
+	struct igt_spinner spin;
+	struct i915_request *rq;
+	intel_wakeref_t wakeref;
+	struct i915_gpu_error *global = &gt->i915->gpu_error;
+	struct intel_engine_cs *engine;
+	unsigned int reset_count;
+	u32 guc_status;
+	u32 old_beat;
+
+	ctx = kernel_context(gt->i915, NULL);
+	if (IS_ERR(ctx)) {
+		drm_err(&gt->i915->drm, "Failed get kernel context: %ld\n", PTR_ERR(ctx));
+		return PTR_ERR(ctx);
+	}
+
+	wakeref = intel_runtime_pm_get(gt->uncore->rpm);
+
+	ce = intel_context_create(gt->engine[BCS0]);
+	if (IS_ERR(ce)) {
+		ret = PTR_ERR(ce);
+		drm_err(&gt->i915->drm, "Failed to create spinner request: %d\n", ret);
+		goto err;
+	}
+
+	engine = ce->engine;
+	reset_count = i915_reset_count(global);
+
+	old_beat = engine->props.heartbeat_interval_ms;
+	ret = intel_engine_set_heartbeat(engine, BEAT_INTERVAL);
+	if (ret) {
+		drm_err(&gt->i915->drm, "Failed to boost heatbeat interval: %d\n", ret);
+		goto err;
+	}
+
+	ret = igt_spinner_init(&spin, engine->gt);
+	if (ret) {
+		drm_err(&gt->i915->drm, "Failed to create spinner: %d\n", ret);
+		goto err;
+	}
+
+	rq = igt_spinner_create_request(&spin, ce, MI_NOOP);
+	intel_context_put(ce);
+	if (IS_ERR(rq)) {
+		ret = PTR_ERR(rq);
+		drm_err(&gt->i915->drm, "Failed to create spinner request: %d\n", ret);
+		goto err_spin;
+	}
+
+	ret = request_add_spin(rq, &spin);
+	if (ret) {
+		i915_request_put(rq);
+		drm_err(&gt->i915->drm, "Failed to add Spinner request: %d\n", ret);
+		goto err_spin;
+	}
+
+	ret = intel_reset_guc(gt);
+	if (ret) {
+		i915_request_put(rq);
+		drm_err(&gt->i915->drm, "Failed to reset GuC, ret = %d\n", ret);
+		goto err_spin;
+	}
+
+	guc_status = intel_uncore_read(gt->uncore, GUC_STATUS);
+	if (!(guc_status & GS_MIA_IN_RESET)) {
+		i915_request_put(rq);
+		drm_err(&gt->i915->drm, "GuC failed to reset: status = 0x%08X\n", guc_status);
+		ret = -EIO;
+		goto err_spin;
+	}
+
+	/* Wait for the heartbeat to cause a reset */
+	ret = intel_selftest_wait_for_rq(rq);
+	i915_request_put(rq);
+	if (ret) {
+		drm_err(&gt->i915->drm, "Request failed to complete: %d\n", ret);
+		goto err_spin;
+	}
+
+	if (i915_reset_count(global) == reset_count) {
+		drm_err(&gt->i915->drm, "Failed to record a GPU reset\n");
+		ret = -EINVAL;
+		goto err_spin;
+	}
+
+err_spin:
+	igt_spinner_end(&spin);
+	igt_spinner_fini(&spin);
+	intel_engine_set_heartbeat(engine, old_beat);
+
+	if (ret == 0) {
+		rq = nop_request(engine);
+		if (IS_ERR(rq)) {
+			ret = PTR_ERR(rq);
+			goto err;
+		}
+
+		ret = intel_selftest_wait_for_rq(rq);
+		i915_request_put(rq);
+		if (ret) {
+			drm_err(&gt->i915->drm, "No-op failed to complete: %d\n", ret);
+			goto err;
+		}
+	}
+
+err:
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+	kernel_context_close(ctx);
+
+	return ret;
+}
+
+int intel_guc_hang_check(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(intel_hang_guc),
+	};
+	struct intel_gt *gt = to_gt(i915);
+
+	if (intel_gt_is_wedged(gt))
+		return 0;
+
+	if (!intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	return intel_gt_live_subtests(tests, gt);
+}
diff --git a/drivers/gpu/drm/i915/gt/uc/selftest_guc_multi_lrc.c b/drivers/gpu/drm/i915/gt/uc/selftest_guc_multi_lrc.c
new file mode 100644
index 000000000..d17982c36
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/uc/selftest_guc_multi_lrc.c
@@ -0,0 +1,183 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright �� 2019 Intel Corporation
+ */
+
+#include "selftests/igt_spinner.h"
+#include "selftests/igt_reset.h"
+#include "selftests/intel_scheduler_helpers.h"
+#include "gt/intel_engine_heartbeat.h"
+#include "gem/selftests/mock_context.h"
+
+static void logical_sort(struct intel_engine_cs **engines, int num_engines)
+{
+	struct intel_engine_cs *sorted[MAX_ENGINE_INSTANCE + 1];
+	int i, j;
+
+	for (i = 0; i < num_engines; ++i)
+		for (j = 0; j < MAX_ENGINE_INSTANCE + 1; ++j) {
+			if (engines[j]->logical_mask & BIT(i)) {
+				sorted[i] = engines[j];
+				break;
+			}
+		}
+
+	memcpy(*engines, *sorted,
+	       sizeof(struct intel_engine_cs *) * num_engines);
+}
+
+static struct intel_context *
+multi_lrc_create_parent(struct intel_gt *gt, u8 class,
+			unsigned long flags)
+{
+	struct intel_engine_cs *siblings[MAX_ENGINE_INSTANCE + 1];
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int i = 0;
+
+	for_each_engine(engine, gt, id) {
+		if (engine->class != class)
+			continue;
+
+		siblings[i++] = engine;
+	}
+
+	if (i <= 1)
+		return ERR_PTR(0);
+
+	logical_sort(siblings, i);
+
+	return intel_engine_create_parallel(siblings, 1, i);
+}
+
+static void multi_lrc_context_unpin(struct intel_context *ce)
+{
+	struct intel_context *child;
+
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+
+	for_each_child(ce, child)
+		intel_context_unpin(child);
+	intel_context_unpin(ce);
+}
+
+static void multi_lrc_context_put(struct intel_context *ce)
+{
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+
+	/*
+	 * Only the parent gets the creation ref put in the uAPI, the parent
+	 * itself is responsible for creation ref put on the children.
+	 */
+	intel_context_put(ce);
+}
+
+static struct i915_request *
+multi_lrc_nop_request(struct intel_context *ce)
+{
+	struct intel_context *child;
+	struct i915_request *rq, *child_rq;
+	int i = 0;
+
+	GEM_BUG_ON(!intel_context_is_parent(ce));
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return rq;
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	for_each_child(ce, child) {
+		child_rq = intel_context_create_request(child);
+		if (IS_ERR(child_rq))
+			goto child_error;
+
+		if (++i == ce->parallel.number_children)
+			set_bit(I915_FENCE_FLAG_SUBMIT_PARALLEL,
+				&child_rq->fence.flags);
+		i915_request_add(child_rq);
+	}
+
+	return rq;
+
+child_error:
+	i915_request_put(rq);
+
+	return ERR_PTR(-ENOMEM);
+}
+
+static int __intel_guc_multi_lrc_basic(struct intel_gt *gt, unsigned int class)
+{
+	struct intel_context *parent;
+	struct i915_request *rq;
+	int ret;
+
+	parent = multi_lrc_create_parent(gt, class, 0);
+	if (IS_ERR(parent)) {
+		drm_err(&gt->i915->drm, "Failed creating contexts: %ld", PTR_ERR(parent));
+		return PTR_ERR(parent);
+	} else if (!parent) {
+		drm_dbg(&gt->i915->drm, "Not enough engines in class: %d", class);
+		return 0;
+	}
+
+	rq = multi_lrc_nop_request(parent);
+	if (IS_ERR(rq)) {
+		ret = PTR_ERR(rq);
+		drm_err(&gt->i915->drm, "Failed creating requests: %d", ret);
+		goto out;
+	}
+
+	ret = intel_selftest_wait_for_rq(rq);
+	if (ret)
+		drm_err(&gt->i915->drm, "Failed waiting on request: %d", ret);
+
+	i915_request_put(rq);
+
+	if (ret >= 0) {
+		ret = intel_gt_wait_for_idle(gt, HZ * 5);
+		if (ret < 0)
+			drm_err(&gt->i915->drm, "GT failed to idle: %d\n", ret);
+	}
+
+out:
+	multi_lrc_context_unpin(parent);
+	multi_lrc_context_put(parent);
+	return ret;
+}
+
+static int intel_guc_multi_lrc_basic(void *arg)
+{
+	struct intel_gt *gt = arg;
+	unsigned int class;
+	int ret;
+
+	for (class = 0; class < MAX_ENGINE_CLASS + 1; ++class) {
+		/* We don't support breadcrumb handshake on these classes */
+		if (class == COMPUTE_CLASS || class == RENDER_CLASS)
+			continue;
+
+		ret = __intel_guc_multi_lrc_basic(gt, class);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+int intel_guc_multi_lrc_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(intel_guc_multi_lrc_basic),
+	};
+	struct intel_gt *gt = to_gt(i915);
+
+	if (intel_gt_is_wedged(gt))
+		return 0;
+
+	if (!intel_uc_uses_guc_submission(&gt->uc))
+		return 0;
+
+	return intel_gt_live_subtests(tests, gt);
+}
diff --git a/drivers/gpu/drm/i915/gvt/Makefile b/drivers/gpu/drm/i915/gvt/Makefile
new file mode 100644
index 000000000..1699f6442
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/Makefile
@@ -0,0 +1,25 @@
+# SPDX-License-Identifier: GPL-2.0
+
+kvmgt-$(CONFIG_DRM_I915_GVT) += \
+	gvt/aperture_gm.o \
+	gvt/cfg_space.o \
+	gvt/cmd_parser.o \
+	gvt/debugfs.o \
+	gvt/display.o \
+	gvt/dmabuf.o \
+	gvt/edid.o \
+	gvt/execlist.o \
+	gvt/fb_decoder.o \
+	gvt/firmware.o \
+	gvt/gtt.o \
+	gvt/handlers.o \
+	gvt/interrupt.o \
+	gvt/kvmgt.o \
+	gvt/mmio.o \
+	gvt/mmio_context.o \
+	gvt/opregion.o \
+	gvt/page_track.o \
+	gvt/sched_policy.o \
+	gvt/scheduler.o \
+	gvt/trace_points.o \
+	gvt/vgpu.o
diff --git a/drivers/gpu/drm/i915/gvt/aperture_gm.c b/drivers/gpu/drm/i915/gvt/aperture_gm.c
new file mode 100644
index 000000000..076c779f7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/aperture_gm.c
@@ -0,0 +1,366 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Kevin Tian <kevin.tian@intel.com>
+ *    Dexuan Cui
+ *
+ * Contributors:
+ *    Pei Zhang <pei.zhang@intel.com>
+ *    Min He <min.he@intel.com>
+ *    Niu Bing <bing.niu@intel.com>
+ *    Yulei Zhang <yulei.zhang@intel.com>
+ *    Zhenyu Wang <zhenyuw@linux.intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "gt/intel_ggtt_fencing.h"
+#include "gvt.h"
+
+static int alloc_gm(struct intel_vgpu *vgpu, bool high_gm)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_gt *gt = gvt->gt;
+	unsigned int flags;
+	u64 start, end, size;
+	struct drm_mm_node *node;
+	int ret;
+
+	if (high_gm) {
+		node = &vgpu->gm.high_gm_node;
+		size = vgpu_hidden_sz(vgpu);
+		start = ALIGN(gvt_hidden_gmadr_base(gvt), I915_GTT_PAGE_SIZE);
+		end = ALIGN(gvt_hidden_gmadr_end(gvt), I915_GTT_PAGE_SIZE);
+		flags = PIN_HIGH;
+	} else {
+		node = &vgpu->gm.low_gm_node;
+		size = vgpu_aperture_sz(vgpu);
+		start = ALIGN(gvt_aperture_gmadr_base(gvt), I915_GTT_PAGE_SIZE);
+		end = ALIGN(gvt_aperture_gmadr_end(gvt), I915_GTT_PAGE_SIZE);
+		flags = PIN_MAPPABLE;
+	}
+
+	mutex_lock(&gt->ggtt->vm.mutex);
+	mmio_hw_access_pre(gt);
+	ret = i915_gem_gtt_insert(&gt->ggtt->vm, NULL, node,
+				  size, I915_GTT_PAGE_SIZE,
+				  I915_COLOR_UNEVICTABLE,
+				  start, end, flags);
+	mmio_hw_access_post(gt);
+	mutex_unlock(&gt->ggtt->vm.mutex);
+	if (ret)
+		gvt_err("fail to alloc %s gm space from host\n",
+			high_gm ? "high" : "low");
+
+	return ret;
+}
+
+static int alloc_vgpu_gm(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_gt *gt = gvt->gt;
+	int ret;
+
+	ret = alloc_gm(vgpu, false);
+	if (ret)
+		return ret;
+
+	ret = alloc_gm(vgpu, true);
+	if (ret)
+		goto out_free_aperture;
+
+	gvt_dbg_core("vgpu%d: alloc low GM start %llx size %llx\n", vgpu->id,
+		     vgpu_aperture_offset(vgpu), vgpu_aperture_sz(vgpu));
+
+	gvt_dbg_core("vgpu%d: alloc high GM start %llx size %llx\n", vgpu->id,
+		     vgpu_hidden_offset(vgpu), vgpu_hidden_sz(vgpu));
+
+	return 0;
+out_free_aperture:
+	mutex_lock(&gt->ggtt->vm.mutex);
+	drm_mm_remove_node(&vgpu->gm.low_gm_node);
+	mutex_unlock(&gt->ggtt->vm.mutex);
+	return ret;
+}
+
+static void free_vgpu_gm(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_gt *gt = gvt->gt;
+
+	mutex_lock(&gt->ggtt->vm.mutex);
+	drm_mm_remove_node(&vgpu->gm.low_gm_node);
+	drm_mm_remove_node(&vgpu->gm.high_gm_node);
+	mutex_unlock(&gt->ggtt->vm.mutex);
+}
+
+/**
+ * intel_vgpu_write_fence - write fence registers owned by a vGPU
+ * @vgpu: vGPU instance
+ * @fence: vGPU fence register number
+ * @value: Fence register value to be written
+ *
+ * This function is used to write fence registers owned by a vGPU. The vGPU
+ * fence register number will be translated into HW fence register number.
+ *
+ */
+void intel_vgpu_write_fence(struct intel_vgpu *vgpu,
+		u32 fence, u64 value)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct drm_i915_private *i915 = gvt->gt->i915;
+	struct intel_uncore *uncore = gvt->gt->uncore;
+	struct i915_fence_reg *reg;
+	i915_reg_t fence_reg_lo, fence_reg_hi;
+
+	assert_rpm_wakelock_held(uncore->rpm);
+
+	if (drm_WARN_ON(&i915->drm, fence >= vgpu_fence_sz(vgpu)))
+		return;
+
+	reg = vgpu->fence.regs[fence];
+	if (drm_WARN_ON(&i915->drm, !reg))
+		return;
+
+	fence_reg_lo = FENCE_REG_GEN6_LO(reg->id);
+	fence_reg_hi = FENCE_REG_GEN6_HI(reg->id);
+
+	intel_uncore_write(uncore, fence_reg_lo, 0);
+	intel_uncore_posting_read(uncore, fence_reg_lo);
+
+	intel_uncore_write(uncore, fence_reg_hi, upper_32_bits(value));
+	intel_uncore_write(uncore, fence_reg_lo, lower_32_bits(value));
+	intel_uncore_posting_read(uncore, fence_reg_lo);
+}
+
+static void _clear_vgpu_fence(struct intel_vgpu *vgpu)
+{
+	int i;
+
+	for (i = 0; i < vgpu_fence_sz(vgpu); i++)
+		intel_vgpu_write_fence(vgpu, i, 0);
+}
+
+static void free_vgpu_fence(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_uncore *uncore = gvt->gt->uncore;
+	struct i915_fence_reg *reg;
+	intel_wakeref_t wakeref;
+	u32 i;
+
+	if (drm_WARN_ON(&gvt->gt->i915->drm, !vgpu_fence_sz(vgpu)))
+		return;
+
+	wakeref = intel_runtime_pm_get(uncore->rpm);
+
+	mutex_lock(&gvt->gt->ggtt->vm.mutex);
+	_clear_vgpu_fence(vgpu);
+	for (i = 0; i < vgpu_fence_sz(vgpu); i++) {
+		reg = vgpu->fence.regs[i];
+		i915_unreserve_fence(reg);
+		vgpu->fence.regs[i] = NULL;
+	}
+	mutex_unlock(&gvt->gt->ggtt->vm.mutex);
+
+	intel_runtime_pm_put(uncore->rpm, wakeref);
+}
+
+static int alloc_vgpu_fence(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_uncore *uncore = gvt->gt->uncore;
+	struct i915_fence_reg *reg;
+	intel_wakeref_t wakeref;
+	int i;
+
+	wakeref = intel_runtime_pm_get(uncore->rpm);
+
+	/* Request fences from host */
+	mutex_lock(&gvt->gt->ggtt->vm.mutex);
+
+	for (i = 0; i < vgpu_fence_sz(vgpu); i++) {
+		reg = i915_reserve_fence(gvt->gt->ggtt);
+		if (IS_ERR(reg))
+			goto out_free_fence;
+
+		vgpu->fence.regs[i] = reg;
+	}
+
+	_clear_vgpu_fence(vgpu);
+
+	mutex_unlock(&gvt->gt->ggtt->vm.mutex);
+	intel_runtime_pm_put(uncore->rpm, wakeref);
+	return 0;
+
+out_free_fence:
+	gvt_vgpu_err("Failed to alloc fences\n");
+	/* Return fences to host, if fail */
+	for (i = 0; i < vgpu_fence_sz(vgpu); i++) {
+		reg = vgpu->fence.regs[i];
+		if (!reg)
+			continue;
+		i915_unreserve_fence(reg);
+		vgpu->fence.regs[i] = NULL;
+	}
+	mutex_unlock(&gvt->gt->ggtt->vm.mutex);
+	intel_runtime_pm_put_unchecked(uncore->rpm);
+	return -ENOSPC;
+}
+
+static void free_resource(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+
+	gvt->gm.vgpu_allocated_low_gm_size -= vgpu_aperture_sz(vgpu);
+	gvt->gm.vgpu_allocated_high_gm_size -= vgpu_hidden_sz(vgpu);
+	gvt->fence.vgpu_allocated_fence_num -= vgpu_fence_sz(vgpu);
+}
+
+static int alloc_resource(struct intel_vgpu *vgpu,
+		const struct intel_vgpu_config *conf)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	unsigned long request, avail, max, taken;
+	const char *item;
+
+	if (!conf->low_mm || !conf->high_mm || !conf->fence) {
+		gvt_vgpu_err("Invalid vGPU creation params\n");
+		return -EINVAL;
+	}
+
+	item = "low GM space";
+	max = gvt_aperture_sz(gvt) - HOST_LOW_GM_SIZE;
+	taken = gvt->gm.vgpu_allocated_low_gm_size;
+	avail = max - taken;
+	request = conf->low_mm;
+
+	if (request > avail)
+		goto no_enough_resource;
+
+	vgpu_aperture_sz(vgpu) = ALIGN(request, I915_GTT_PAGE_SIZE);
+
+	item = "high GM space";
+	max = gvt_hidden_sz(gvt) - HOST_HIGH_GM_SIZE;
+	taken = gvt->gm.vgpu_allocated_high_gm_size;
+	avail = max - taken;
+	request = conf->high_mm;
+
+	if (request > avail)
+		goto no_enough_resource;
+
+	vgpu_hidden_sz(vgpu) = ALIGN(request, I915_GTT_PAGE_SIZE);
+
+	item = "fence";
+	max = gvt_fence_sz(gvt) - HOST_FENCE;
+	taken = gvt->fence.vgpu_allocated_fence_num;
+	avail = max - taken;
+	request = conf->fence;
+
+	if (request > avail)
+		goto no_enough_resource;
+
+	vgpu_fence_sz(vgpu) = request;
+
+	gvt->gm.vgpu_allocated_low_gm_size += conf->low_mm;
+	gvt->gm.vgpu_allocated_high_gm_size += conf->high_mm;
+	gvt->fence.vgpu_allocated_fence_num += conf->fence;
+	return 0;
+
+no_enough_resource:
+	gvt_err("fail to allocate resource %s\n", item);
+	gvt_err("request %luMB avail %luMB max %luMB taken %luMB\n",
+		BYTES_TO_MB(request), BYTES_TO_MB(avail),
+		BYTES_TO_MB(max), BYTES_TO_MB(taken));
+	return -ENOSPC;
+}
+
+/**
+ * intel_vgpu_free_resource() - free HW resource owned by a vGPU
+ * @vgpu: a vGPU
+ *
+ * This function is used to free the HW resource owned by a vGPU.
+ *
+ */
+void intel_vgpu_free_resource(struct intel_vgpu *vgpu)
+{
+	free_vgpu_gm(vgpu);
+	free_vgpu_fence(vgpu);
+	free_resource(vgpu);
+}
+
+/**
+ * intel_vgpu_reset_resource - reset resource state owned by a vGPU
+ * @vgpu: a vGPU
+ *
+ * This function is used to reset resource state owned by a vGPU.
+ *
+ */
+void intel_vgpu_reset_resource(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	intel_wakeref_t wakeref;
+
+	with_intel_runtime_pm(gvt->gt->uncore->rpm, wakeref)
+		_clear_vgpu_fence(vgpu);
+}
+
+/**
+ * intel_vgpu_alloc_resource() - allocate HW resource for a vGPU
+ * @vgpu: vGPU
+ * @param: vGPU creation params
+ *
+ * This function is used to allocate HW resource for a vGPU. User specifies
+ * the resource configuration through the creation params.
+ *
+ * Returns:
+ * zero on success, negative error code if failed.
+ *
+ */
+int intel_vgpu_alloc_resource(struct intel_vgpu *vgpu,
+		const struct intel_vgpu_config *conf)
+{
+	int ret;
+
+	ret = alloc_resource(vgpu, conf);
+	if (ret)
+		return ret;
+
+	ret = alloc_vgpu_gm(vgpu);
+	if (ret)
+		goto out_free_resource;
+
+	ret = alloc_vgpu_fence(vgpu);
+	if (ret)
+		goto out_free_vgpu_gm;
+
+	return 0;
+
+out_free_vgpu_gm:
+	free_vgpu_gm(vgpu);
+out_free_resource:
+	free_resource(vgpu);
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/gvt/cfg_space.c b/drivers/gpu/drm/i915/gvt/cfg_space.c
new file mode 100644
index 000000000..eef3bba8a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/cfg_space.c
@@ -0,0 +1,400 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Eddie Dong <eddie.dong@intel.com>
+ *    Jike Song <jike.song@intel.com>
+ *
+ * Contributors:
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *    Min He <min.he@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *
+ */
+
+#include "i915_drv.h"
+#include "gvt.h"
+#include "intel_pci_config.h"
+
+enum {
+	INTEL_GVT_PCI_BAR_GTTMMIO = 0,
+	INTEL_GVT_PCI_BAR_APERTURE,
+	INTEL_GVT_PCI_BAR_PIO,
+	INTEL_GVT_PCI_BAR_MAX,
+};
+
+/* bitmap for writable bits (RW or RW1C bits, but cannot co-exist in one
+ * byte) byte by byte in standard pci configuration space. (not the full
+ * 256 bytes.)
+ */
+static const u8 pci_cfg_space_rw_bmp[PCI_INTERRUPT_LINE + 4] = {
+	[PCI_COMMAND]		= 0xff, 0x07,
+	[PCI_STATUS]		= 0x00, 0xf9, /* the only one RW1C byte */
+	[PCI_CACHE_LINE_SIZE]	= 0xff,
+	[PCI_BASE_ADDRESS_0 ... PCI_CARDBUS_CIS - 1] = 0xff,
+	[PCI_ROM_ADDRESS]	= 0x01, 0xf8, 0xff, 0xff,
+	[PCI_INTERRUPT_LINE]	= 0xff,
+};
+
+/**
+ * vgpu_pci_cfg_mem_write - write virtual cfg space memory
+ * @vgpu: target vgpu
+ * @off: offset
+ * @src: src ptr to write
+ * @bytes: number of bytes
+ *
+ * Use this function to write virtual cfg space memory.
+ * For standard cfg space, only RW bits can be changed,
+ * and we emulates the RW1C behavior of PCI_STATUS register.
+ */
+static void vgpu_pci_cfg_mem_write(struct intel_vgpu *vgpu, unsigned int off,
+				   u8 *src, unsigned int bytes)
+{
+	u8 *cfg_base = vgpu_cfg_space(vgpu);
+	u8 mask, new, old;
+	pci_power_t pwr;
+	int i = 0;
+
+	for (; i < bytes && (off + i < sizeof(pci_cfg_space_rw_bmp)); i++) {
+		mask = pci_cfg_space_rw_bmp[off + i];
+		old = cfg_base[off + i];
+		new = src[i] & mask;
+
+		/**
+		 * The PCI_STATUS high byte has RW1C bits, here
+		 * emulates clear by writing 1 for these bits.
+		 * Writing a 0b to RW1C bits has no effect.
+		 */
+		if (off + i == PCI_STATUS + 1)
+			new = (~new & old) & mask;
+
+		cfg_base[off + i] = (old & ~mask) | new;
+	}
+
+	/* For other configuration space directly copy as it is. */
+	if (i < bytes)
+		memcpy(cfg_base + off + i, src + i, bytes - i);
+
+	if (off == vgpu->cfg_space.pmcsr_off && vgpu->cfg_space.pmcsr_off) {
+		pwr = (pci_power_t __force)(*(u16*)(&vgpu_cfg_space(vgpu)[off])
+			& PCI_PM_CTRL_STATE_MASK);
+		if (pwr == PCI_D3hot)
+			vgpu->d3_entered = true;
+		gvt_dbg_core("vgpu-%d power status changed to %d\n",
+			     vgpu->id, pwr);
+	}
+}
+
+/**
+ * intel_vgpu_emulate_cfg_read - emulate vGPU configuration space read
+ * @vgpu: target vgpu
+ * @offset: offset
+ * @p_data: return data ptr
+ * @bytes: number of bytes to read
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_emulate_cfg_read(struct intel_vgpu *vgpu, unsigned int offset,
+	void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+
+	if (drm_WARN_ON(&i915->drm, bytes > 4))
+		return -EINVAL;
+
+	if (drm_WARN_ON(&i915->drm,
+			offset + bytes > vgpu->gvt->device_info.cfg_space_size))
+		return -EINVAL;
+
+	memcpy(p_data, vgpu_cfg_space(vgpu) + offset, bytes);
+	return 0;
+}
+
+static void map_aperture(struct intel_vgpu *vgpu, bool map)
+{
+	if (map != vgpu->cfg_space.bar[INTEL_GVT_PCI_BAR_APERTURE].tracked)
+		vgpu->cfg_space.bar[INTEL_GVT_PCI_BAR_APERTURE].tracked = map;
+}
+
+static void trap_gttmmio(struct intel_vgpu *vgpu, bool trap)
+{
+	if (trap != vgpu->cfg_space.bar[INTEL_GVT_PCI_BAR_GTTMMIO].tracked)
+		vgpu->cfg_space.bar[INTEL_GVT_PCI_BAR_GTTMMIO].tracked = trap;
+}
+
+static int emulate_pci_command_write(struct intel_vgpu *vgpu,
+	unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u8 old = vgpu_cfg_space(vgpu)[offset];
+	u8 new = *(u8 *)p_data;
+	u8 changed = old ^ new;
+
+	vgpu_pci_cfg_mem_write(vgpu, offset, p_data, bytes);
+	if (!(changed & PCI_COMMAND_MEMORY))
+		return 0;
+
+	if (old & PCI_COMMAND_MEMORY) {
+		trap_gttmmio(vgpu, false);
+		map_aperture(vgpu, false);
+	} else {
+		trap_gttmmio(vgpu, true);
+		map_aperture(vgpu, true);
+	}
+
+	return 0;
+}
+
+static int emulate_pci_rom_bar_write(struct intel_vgpu *vgpu,
+	unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 *pval = (u32 *)(vgpu_cfg_space(vgpu) + offset);
+	u32 new = *(u32 *)(p_data);
+
+	if ((new & PCI_ROM_ADDRESS_MASK) == PCI_ROM_ADDRESS_MASK)
+		/* We don't have rom, return size of 0. */
+		*pval = 0;
+	else
+		vgpu_pci_cfg_mem_write(vgpu, offset, p_data, bytes);
+	return 0;
+}
+
+static void emulate_pci_bar_write(struct intel_vgpu *vgpu, unsigned int offset,
+	void *p_data, unsigned int bytes)
+{
+	u32 new = *(u32 *)(p_data);
+	bool lo = IS_ALIGNED(offset, 8);
+	u64 size;
+	bool mmio_enabled =
+		vgpu_cfg_space(vgpu)[PCI_COMMAND] & PCI_COMMAND_MEMORY;
+	struct intel_vgpu_pci_bar *bars = vgpu->cfg_space.bar;
+
+	/*
+	 * Power-up software can determine how much address
+	 * space the device requires by writing a value of
+	 * all 1's to the register and then reading the value
+	 * back. The device will return 0's in all don't-care
+	 * address bits.
+	 */
+	if (new == 0xffffffff) {
+		switch (offset) {
+		case PCI_BASE_ADDRESS_0:
+		case PCI_BASE_ADDRESS_1:
+			size = ~(bars[INTEL_GVT_PCI_BAR_GTTMMIO].size -1);
+			intel_vgpu_write_pci_bar(vgpu, offset,
+						size >> (lo ? 0 : 32), lo);
+			/*
+			 * Untrap the BAR, since guest hasn't configured a
+			 * valid GPA
+			 */
+			trap_gttmmio(vgpu, false);
+			break;
+		case PCI_BASE_ADDRESS_2:
+		case PCI_BASE_ADDRESS_3:
+			size = ~(bars[INTEL_GVT_PCI_BAR_APERTURE].size -1);
+			intel_vgpu_write_pci_bar(vgpu, offset,
+						size >> (lo ? 0 : 32), lo);
+			map_aperture(vgpu, false);
+			break;
+		default:
+			/* Unimplemented BARs */
+			intel_vgpu_write_pci_bar(vgpu, offset, 0x0, false);
+		}
+	} else {
+		switch (offset) {
+		case PCI_BASE_ADDRESS_0:
+		case PCI_BASE_ADDRESS_1:
+			/*
+			 * Untrap the old BAR first, since guest has
+			 * re-configured the BAR
+			 */
+			trap_gttmmio(vgpu, false);
+			intel_vgpu_write_pci_bar(vgpu, offset, new, lo);
+			trap_gttmmio(vgpu, mmio_enabled);
+			break;
+		case PCI_BASE_ADDRESS_2:
+		case PCI_BASE_ADDRESS_3:
+			map_aperture(vgpu, false);
+			intel_vgpu_write_pci_bar(vgpu, offset, new, lo);
+			map_aperture(vgpu, mmio_enabled);
+			break;
+		default:
+			intel_vgpu_write_pci_bar(vgpu, offset, new, lo);
+		}
+	}
+}
+
+/**
+ * intel_vgpu_emulate_cfg_read - emulate vGPU configuration space write
+ * @vgpu: target vgpu
+ * @offset: offset
+ * @p_data: write data ptr
+ * @bytes: number of bytes to write
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_emulate_cfg_write(struct intel_vgpu *vgpu, unsigned int offset,
+	void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	int ret;
+
+	if (drm_WARN_ON(&i915->drm, bytes > 4))
+		return -EINVAL;
+
+	if (drm_WARN_ON(&i915->drm,
+			offset + bytes > vgpu->gvt->device_info.cfg_space_size))
+		return -EINVAL;
+
+	/* First check if it's PCI_COMMAND */
+	if (IS_ALIGNED(offset, 2) && offset == PCI_COMMAND) {
+		if (drm_WARN_ON(&i915->drm, bytes > 2))
+			return -EINVAL;
+		return emulate_pci_command_write(vgpu, offset, p_data, bytes);
+	}
+
+	switch (rounddown(offset, 4)) {
+	case PCI_ROM_ADDRESS:
+		if (drm_WARN_ON(&i915->drm, !IS_ALIGNED(offset, 4)))
+			return -EINVAL;
+		return emulate_pci_rom_bar_write(vgpu, offset, p_data, bytes);
+
+	case PCI_BASE_ADDRESS_0 ... PCI_BASE_ADDRESS_5:
+		if (drm_WARN_ON(&i915->drm, !IS_ALIGNED(offset, 4)))
+			return -EINVAL;
+		emulate_pci_bar_write(vgpu, offset, p_data, bytes);
+		break;
+	case INTEL_GVT_PCI_SWSCI:
+		if (drm_WARN_ON(&i915->drm, !IS_ALIGNED(offset, 4)))
+			return -EINVAL;
+		ret = intel_vgpu_emulate_opregion_request(vgpu, *(u32 *)p_data);
+		if (ret)
+			return ret;
+		break;
+
+	case INTEL_GVT_PCI_OPREGION:
+		if (drm_WARN_ON(&i915->drm, !IS_ALIGNED(offset, 4)))
+			return -EINVAL;
+		ret = intel_vgpu_opregion_base_write_handler(vgpu,
+						   *(u32 *)p_data);
+		if (ret)
+			return ret;
+
+		vgpu_pci_cfg_mem_write(vgpu, offset, p_data, bytes);
+		break;
+	default:
+		vgpu_pci_cfg_mem_write(vgpu, offset, p_data, bytes);
+		break;
+	}
+	return 0;
+}
+
+/**
+ * intel_vgpu_init_cfg_space - init vGPU configuration space when create vGPU
+ *
+ * @vgpu: a vGPU
+ * @primary: is the vGPU presented as primary
+ *
+ */
+void intel_vgpu_init_cfg_space(struct intel_vgpu *vgpu,
+			       bool primary)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct pci_dev *pdev = to_pci_dev(gvt->gt->i915->drm.dev);
+	const struct intel_gvt_device_info *info = &gvt->device_info;
+	u16 *gmch_ctl;
+	u8 next;
+
+	memcpy(vgpu_cfg_space(vgpu), gvt->firmware.cfg_space,
+	       info->cfg_space_size);
+
+	if (!primary) {
+		vgpu_cfg_space(vgpu)[PCI_CLASS_DEVICE] =
+			INTEL_GVT_PCI_CLASS_VGA_OTHER;
+		vgpu_cfg_space(vgpu)[PCI_CLASS_PROG] =
+			INTEL_GVT_PCI_CLASS_VGA_OTHER;
+	}
+
+	/* Show guest that there isn't any stolen memory.*/
+	gmch_ctl = (u16 *)(vgpu_cfg_space(vgpu) + INTEL_GVT_PCI_GMCH_CONTROL);
+	*gmch_ctl &= ~(BDW_GMCH_GMS_MASK << BDW_GMCH_GMS_SHIFT);
+
+	intel_vgpu_write_pci_bar(vgpu, PCI_BASE_ADDRESS_2,
+				 gvt_aperture_pa_base(gvt), true);
+
+	vgpu_cfg_space(vgpu)[PCI_COMMAND] &= ~(PCI_COMMAND_IO
+					     | PCI_COMMAND_MEMORY
+					     | PCI_COMMAND_MASTER);
+	/*
+	 * Clear the bar upper 32bit and let guest to assign the new value
+	 */
+	memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_1, 0, 4);
+	memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_3, 0, 4);
+	memset(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_4, 0, 8);
+	memset(vgpu_cfg_space(vgpu) + INTEL_GVT_PCI_OPREGION, 0, 4);
+
+	vgpu->cfg_space.bar[INTEL_GVT_PCI_BAR_GTTMMIO].size =
+		pci_resource_len(pdev, GTTMMADR_BAR);
+	vgpu->cfg_space.bar[INTEL_GVT_PCI_BAR_APERTURE].size =
+		pci_resource_len(pdev, GTT_APERTURE_BAR);
+
+	memset(vgpu_cfg_space(vgpu) + PCI_ROM_ADDRESS, 0, 4);
+
+	/* PM Support */
+	vgpu->cfg_space.pmcsr_off = 0;
+	if (vgpu_cfg_space(vgpu)[PCI_STATUS] & PCI_STATUS_CAP_LIST) {
+		next = vgpu_cfg_space(vgpu)[PCI_CAPABILITY_LIST];
+		do {
+			if (vgpu_cfg_space(vgpu)[next + PCI_CAP_LIST_ID] == PCI_CAP_ID_PM) {
+				vgpu->cfg_space.pmcsr_off = next + PCI_PM_CTRL;
+				break;
+			}
+			next = vgpu_cfg_space(vgpu)[next + PCI_CAP_LIST_NEXT];
+		} while (next);
+	}
+}
+
+/**
+ * intel_vgpu_reset_cfg_space - reset vGPU configuration space
+ *
+ * @vgpu: a vGPU
+ *
+ */
+void intel_vgpu_reset_cfg_space(struct intel_vgpu *vgpu)
+{
+	u8 cmd = vgpu_cfg_space(vgpu)[PCI_COMMAND];
+	bool primary = vgpu_cfg_space(vgpu)[PCI_CLASS_DEVICE] !=
+				INTEL_GVT_PCI_CLASS_VGA_OTHER;
+
+	if (cmd & PCI_COMMAND_MEMORY) {
+		trap_gttmmio(vgpu, false);
+		map_aperture(vgpu, false);
+	}
+
+	/**
+	 * Currently we only do such reset when vGPU is not
+	 * owned by any VM, so we simply restore entire cfg
+	 * space to default value.
+	 */
+	intel_vgpu_init_cfg_space(vgpu, primary);
+}
diff --git a/drivers/gpu/drm/i915/gvt/cmd_parser.c b/drivers/gpu/drm/i915/gvt/cmd_parser.c
new file mode 100644
index 000000000..de13f102d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/cmd_parser.c
@@ -0,0 +1,3263 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Ke Yu
+ *    Kevin Tian <kevin.tian@intel.com>
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *
+ * Contributors:
+ *    Min He <min.he@intel.com>
+ *    Ping Gao <ping.a.gao@intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *    Yulei Zhang <yulei.zhang@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#include <linux/slab.h>
+
+#include "i915_drv.h"
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_lrc.h"
+#include "gt/intel_ring.h"
+#include "gt/intel_gt_requests.h"
+#include "gt/shmem_utils.h"
+#include "gvt.h"
+#include "i915_pvinfo.h"
+#include "trace.h"
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_context.h"
+
+#define INVALID_OP    (~0U)
+
+#define OP_LEN_MI           9
+#define OP_LEN_2D           10
+#define OP_LEN_3D_MEDIA     16
+#define OP_LEN_MFX_VC       16
+#define OP_LEN_VEBOX	    16
+
+#define CMD_TYPE(cmd)	(((cmd) >> 29) & 7)
+
+struct sub_op_bits {
+	int hi;
+	int low;
+};
+struct decode_info {
+	const char *name;
+	int op_len;
+	int nr_sub_op;
+	const struct sub_op_bits *sub_op;
+};
+
+#define   MAX_CMD_BUDGET			0x7fffffff
+#define   MI_WAIT_FOR_PLANE_C_FLIP_PENDING      (1<<15)
+#define   MI_WAIT_FOR_PLANE_B_FLIP_PENDING      (1<<9)
+#define   MI_WAIT_FOR_PLANE_A_FLIP_PENDING      (1<<1)
+
+#define   MI_WAIT_FOR_SPRITE_C_FLIP_PENDING      (1<<20)
+#define   MI_WAIT_FOR_SPRITE_B_FLIP_PENDING      (1<<10)
+#define   MI_WAIT_FOR_SPRITE_A_FLIP_PENDING      (1<<2)
+
+/* Render Command Map */
+
+/* MI_* command Opcode (28:23) */
+#define OP_MI_NOOP                          0x0
+#define OP_MI_SET_PREDICATE                 0x1  /* HSW+ */
+#define OP_MI_USER_INTERRUPT                0x2
+#define OP_MI_WAIT_FOR_EVENT                0x3
+#define OP_MI_FLUSH                         0x4
+#define OP_MI_ARB_CHECK                     0x5
+#define OP_MI_RS_CONTROL                    0x6  /* HSW+ */
+#define OP_MI_REPORT_HEAD                   0x7
+#define OP_MI_ARB_ON_OFF                    0x8
+#define OP_MI_URB_ATOMIC_ALLOC              0x9  /* HSW+ */
+#define OP_MI_BATCH_BUFFER_END              0xA
+#define OP_MI_SUSPEND_FLUSH                 0xB
+#define OP_MI_PREDICATE                     0xC  /* IVB+ */
+#define OP_MI_TOPOLOGY_FILTER               0xD  /* IVB+ */
+#define OP_MI_SET_APPID                     0xE  /* IVB+ */
+#define OP_MI_RS_CONTEXT                    0xF  /* HSW+ */
+#define OP_MI_LOAD_SCAN_LINES_INCL          0x12 /* HSW+ */
+#define OP_MI_DISPLAY_FLIP                  0x14
+#define OP_MI_SEMAPHORE_MBOX                0x16
+#define OP_MI_SET_CONTEXT                   0x18
+#define OP_MI_MATH                          0x1A
+#define OP_MI_URB_CLEAR                     0x19
+#define OP_MI_SEMAPHORE_SIGNAL		    0x1B  /* BDW+ */
+#define OP_MI_SEMAPHORE_WAIT		    0x1C  /* BDW+ */
+
+#define OP_MI_STORE_DATA_IMM                0x20
+#define OP_MI_STORE_DATA_INDEX              0x21
+#define OP_MI_LOAD_REGISTER_IMM             0x22
+#define OP_MI_UPDATE_GTT                    0x23
+#define OP_MI_STORE_REGISTER_MEM            0x24
+#define OP_MI_FLUSH_DW                      0x26
+#define OP_MI_CLFLUSH                       0x27
+#define OP_MI_REPORT_PERF_COUNT             0x28
+#define OP_MI_LOAD_REGISTER_MEM             0x29  /* HSW+ */
+#define OP_MI_LOAD_REGISTER_REG             0x2A  /* HSW+ */
+#define OP_MI_RS_STORE_DATA_IMM             0x2B  /* HSW+ */
+#define OP_MI_LOAD_URB_MEM                  0x2C  /* HSW+ */
+#define OP_MI_STORE_URM_MEM                 0x2D  /* HSW+ */
+#define OP_MI_2E			    0x2E  /* BDW+ */
+#define OP_MI_2F			    0x2F  /* BDW+ */
+#define OP_MI_BATCH_BUFFER_START            0x31
+
+/* Bit definition for dword 0 */
+#define _CMDBIT_BB_START_IN_PPGTT	(1UL << 8)
+
+#define OP_MI_CONDITIONAL_BATCH_BUFFER_END  0x36
+
+#define BATCH_BUFFER_ADDR_MASK ((1UL << 32) - (1U << 2))
+#define BATCH_BUFFER_ADDR_HIGH_MASK ((1UL << 16) - (1U))
+#define BATCH_BUFFER_ADR_SPACE_BIT(x)	(((x) >> 8) & 1U)
+#define BATCH_BUFFER_2ND_LEVEL_BIT(x)   ((x) >> 22 & 1U)
+
+/* 2D command: Opcode (28:22) */
+#define OP_2D(x)    ((2<<7) | x)
+
+#define OP_XY_SETUP_BLT                             OP_2D(0x1)
+#define OP_XY_SETUP_CLIP_BLT                        OP_2D(0x3)
+#define OP_XY_SETUP_MONO_PATTERN_SL_BLT             OP_2D(0x11)
+#define OP_XY_PIXEL_BLT                             OP_2D(0x24)
+#define OP_XY_SCANLINES_BLT                         OP_2D(0x25)
+#define OP_XY_TEXT_BLT                              OP_2D(0x26)
+#define OP_XY_TEXT_IMMEDIATE_BLT                    OP_2D(0x31)
+#define OP_XY_COLOR_BLT                             OP_2D(0x50)
+#define OP_XY_PAT_BLT                               OP_2D(0x51)
+#define OP_XY_MONO_PAT_BLT                          OP_2D(0x52)
+#define OP_XY_SRC_COPY_BLT                          OP_2D(0x53)
+#define OP_XY_MONO_SRC_COPY_BLT                     OP_2D(0x54)
+#define OP_XY_FULL_BLT                              OP_2D(0x55)
+#define OP_XY_FULL_MONO_SRC_BLT                     OP_2D(0x56)
+#define OP_XY_FULL_MONO_PATTERN_BLT                 OP_2D(0x57)
+#define OP_XY_FULL_MONO_PATTERN_MONO_SRC_BLT        OP_2D(0x58)
+#define OP_XY_MONO_PAT_FIXED_BLT                    OP_2D(0x59)
+#define OP_XY_MONO_SRC_COPY_IMMEDIATE_BLT           OP_2D(0x71)
+#define OP_XY_PAT_BLT_IMMEDIATE                     OP_2D(0x72)
+#define OP_XY_SRC_COPY_CHROMA_BLT                   OP_2D(0x73)
+#define OP_XY_FULL_IMMEDIATE_PATTERN_BLT            OP_2D(0x74)
+#define OP_XY_FULL_MONO_SRC_IMMEDIATE_PATTERN_BLT   OP_2D(0x75)
+#define OP_XY_PAT_CHROMA_BLT                        OP_2D(0x76)
+#define OP_XY_PAT_CHROMA_BLT_IMMEDIATE              OP_2D(0x77)
+
+/* 3D/Media Command: Pipeline Type(28:27) Opcode(26:24) Sub Opcode(23:16) */
+#define OP_3D_MEDIA(sub_type, opcode, sub_opcode) \
+	((3 << 13) | ((sub_type) << 11) | ((opcode) << 8) | (sub_opcode))
+
+#define OP_STATE_PREFETCH                       OP_3D_MEDIA(0x0, 0x0, 0x03)
+
+#define OP_STATE_BASE_ADDRESS                   OP_3D_MEDIA(0x0, 0x1, 0x01)
+#define OP_STATE_SIP                            OP_3D_MEDIA(0x0, 0x1, 0x02)
+#define OP_3D_MEDIA_0_1_4			OP_3D_MEDIA(0x0, 0x1, 0x04)
+#define OP_SWTESS_BASE_ADDRESS			OP_3D_MEDIA(0x0, 0x1, 0x03)
+
+#define OP_3DSTATE_VF_STATISTICS_GM45           OP_3D_MEDIA(0x1, 0x0, 0x0B)
+
+#define OP_PIPELINE_SELECT                      OP_3D_MEDIA(0x1, 0x1, 0x04)
+
+#define OP_MEDIA_VFE_STATE                      OP_3D_MEDIA(0x2, 0x0, 0x0)
+#define OP_MEDIA_CURBE_LOAD                     OP_3D_MEDIA(0x2, 0x0, 0x1)
+#define OP_MEDIA_INTERFACE_DESCRIPTOR_LOAD      OP_3D_MEDIA(0x2, 0x0, 0x2)
+#define OP_MEDIA_GATEWAY_STATE                  OP_3D_MEDIA(0x2, 0x0, 0x3)
+#define OP_MEDIA_STATE_FLUSH                    OP_3D_MEDIA(0x2, 0x0, 0x4)
+#define OP_MEDIA_POOL_STATE                     OP_3D_MEDIA(0x2, 0x0, 0x5)
+
+#define OP_MEDIA_OBJECT                         OP_3D_MEDIA(0x2, 0x1, 0x0)
+#define OP_MEDIA_OBJECT_PRT                     OP_3D_MEDIA(0x2, 0x1, 0x2)
+#define OP_MEDIA_OBJECT_WALKER                  OP_3D_MEDIA(0x2, 0x1, 0x3)
+#define OP_GPGPU_WALKER                         OP_3D_MEDIA(0x2, 0x1, 0x5)
+
+#define OP_3DSTATE_CLEAR_PARAMS                 OP_3D_MEDIA(0x3, 0x0, 0x04) /* IVB+ */
+#define OP_3DSTATE_DEPTH_BUFFER                 OP_3D_MEDIA(0x3, 0x0, 0x05) /* IVB+ */
+#define OP_3DSTATE_STENCIL_BUFFER               OP_3D_MEDIA(0x3, 0x0, 0x06) /* IVB+ */
+#define OP_3DSTATE_HIER_DEPTH_BUFFER            OP_3D_MEDIA(0x3, 0x0, 0x07) /* IVB+ */
+#define OP_3DSTATE_VERTEX_BUFFERS               OP_3D_MEDIA(0x3, 0x0, 0x08)
+#define OP_3DSTATE_VERTEX_ELEMENTS              OP_3D_MEDIA(0x3, 0x0, 0x09)
+#define OP_3DSTATE_INDEX_BUFFER                 OP_3D_MEDIA(0x3, 0x0, 0x0A)
+#define OP_3DSTATE_VF_STATISTICS                OP_3D_MEDIA(0x3, 0x0, 0x0B)
+#define OP_3DSTATE_VF                           OP_3D_MEDIA(0x3, 0x0, 0x0C)  /* HSW+ */
+#define OP_3DSTATE_CC_STATE_POINTERS            OP_3D_MEDIA(0x3, 0x0, 0x0E)
+#define OP_3DSTATE_SCISSOR_STATE_POINTERS       OP_3D_MEDIA(0x3, 0x0, 0x0F)
+#define OP_3DSTATE_VS                           OP_3D_MEDIA(0x3, 0x0, 0x10)
+#define OP_3DSTATE_GS                           OP_3D_MEDIA(0x3, 0x0, 0x11)
+#define OP_3DSTATE_CLIP                         OP_3D_MEDIA(0x3, 0x0, 0x12)
+#define OP_3DSTATE_SF                           OP_3D_MEDIA(0x3, 0x0, 0x13)
+#define OP_3DSTATE_WM                           OP_3D_MEDIA(0x3, 0x0, 0x14)
+#define OP_3DSTATE_CONSTANT_VS                  OP_3D_MEDIA(0x3, 0x0, 0x15)
+#define OP_3DSTATE_CONSTANT_GS                  OP_3D_MEDIA(0x3, 0x0, 0x16)
+#define OP_3DSTATE_CONSTANT_PS                  OP_3D_MEDIA(0x3, 0x0, 0x17)
+#define OP_3DSTATE_SAMPLE_MASK                  OP_3D_MEDIA(0x3, 0x0, 0x18)
+#define OP_3DSTATE_CONSTANT_HS                  OP_3D_MEDIA(0x3, 0x0, 0x19) /* IVB+ */
+#define OP_3DSTATE_CONSTANT_DS                  OP_3D_MEDIA(0x3, 0x0, 0x1A) /* IVB+ */
+#define OP_3DSTATE_HS                           OP_3D_MEDIA(0x3, 0x0, 0x1B) /* IVB+ */
+#define OP_3DSTATE_TE                           OP_3D_MEDIA(0x3, 0x0, 0x1C) /* IVB+ */
+#define OP_3DSTATE_DS                           OP_3D_MEDIA(0x3, 0x0, 0x1D) /* IVB+ */
+#define OP_3DSTATE_STREAMOUT                    OP_3D_MEDIA(0x3, 0x0, 0x1E) /* IVB+ */
+#define OP_3DSTATE_SBE                          OP_3D_MEDIA(0x3, 0x0, 0x1F) /* IVB+ */
+#define OP_3DSTATE_PS                           OP_3D_MEDIA(0x3, 0x0, 0x20) /* IVB+ */
+#define OP_3DSTATE_VIEWPORT_STATE_POINTERS_SF_CLIP OP_3D_MEDIA(0x3, 0x0, 0x21) /* IVB+ */
+#define OP_3DSTATE_VIEWPORT_STATE_POINTERS_CC   OP_3D_MEDIA(0x3, 0x0, 0x23) /* IVB+ */
+#define OP_3DSTATE_BLEND_STATE_POINTERS         OP_3D_MEDIA(0x3, 0x0, 0x24) /* IVB+ */
+#define OP_3DSTATE_DEPTH_STENCIL_STATE_POINTERS OP_3D_MEDIA(0x3, 0x0, 0x25) /* IVB+ */
+#define OP_3DSTATE_BINDING_TABLE_POINTERS_VS    OP_3D_MEDIA(0x3, 0x0, 0x26) /* IVB+ */
+#define OP_3DSTATE_BINDING_TABLE_POINTERS_HS    OP_3D_MEDIA(0x3, 0x0, 0x27) /* IVB+ */
+#define OP_3DSTATE_BINDING_TABLE_POINTERS_DS    OP_3D_MEDIA(0x3, 0x0, 0x28) /* IVB+ */
+#define OP_3DSTATE_BINDING_TABLE_POINTERS_GS    OP_3D_MEDIA(0x3, 0x0, 0x29) /* IVB+ */
+#define OP_3DSTATE_BINDING_TABLE_POINTERS_PS    OP_3D_MEDIA(0x3, 0x0, 0x2A) /* IVB+ */
+#define OP_3DSTATE_SAMPLER_STATE_POINTERS_VS    OP_3D_MEDIA(0x3, 0x0, 0x2B) /* IVB+ */
+#define OP_3DSTATE_SAMPLER_STATE_POINTERS_HS    OP_3D_MEDIA(0x3, 0x0, 0x2C) /* IVB+ */
+#define OP_3DSTATE_SAMPLER_STATE_POINTERS_DS    OP_3D_MEDIA(0x3, 0x0, 0x2D) /* IVB+ */
+#define OP_3DSTATE_SAMPLER_STATE_POINTERS_GS    OP_3D_MEDIA(0x3, 0x0, 0x2E) /* IVB+ */
+#define OP_3DSTATE_SAMPLER_STATE_POINTERS_PS    OP_3D_MEDIA(0x3, 0x0, 0x2F) /* IVB+ */
+#define OP_3DSTATE_URB_VS                       OP_3D_MEDIA(0x3, 0x0, 0x30) /* IVB+ */
+#define OP_3DSTATE_URB_HS                       OP_3D_MEDIA(0x3, 0x0, 0x31) /* IVB+ */
+#define OP_3DSTATE_URB_DS                       OP_3D_MEDIA(0x3, 0x0, 0x32) /* IVB+ */
+#define OP_3DSTATE_URB_GS                       OP_3D_MEDIA(0x3, 0x0, 0x33) /* IVB+ */
+#define OP_3DSTATE_GATHER_CONSTANT_VS           OP_3D_MEDIA(0x3, 0x0, 0x34) /* HSW+ */
+#define OP_3DSTATE_GATHER_CONSTANT_GS           OP_3D_MEDIA(0x3, 0x0, 0x35) /* HSW+ */
+#define OP_3DSTATE_GATHER_CONSTANT_HS           OP_3D_MEDIA(0x3, 0x0, 0x36) /* HSW+ */
+#define OP_3DSTATE_GATHER_CONSTANT_DS           OP_3D_MEDIA(0x3, 0x0, 0x37) /* HSW+ */
+#define OP_3DSTATE_GATHER_CONSTANT_PS           OP_3D_MEDIA(0x3, 0x0, 0x38) /* HSW+ */
+#define OP_3DSTATE_DX9_CONSTANTF_VS             OP_3D_MEDIA(0x3, 0x0, 0x39) /* HSW+ */
+#define OP_3DSTATE_DX9_CONSTANTF_PS             OP_3D_MEDIA(0x3, 0x0, 0x3A) /* HSW+ */
+#define OP_3DSTATE_DX9_CONSTANTI_VS             OP_3D_MEDIA(0x3, 0x0, 0x3B) /* HSW+ */
+#define OP_3DSTATE_DX9_CONSTANTI_PS             OP_3D_MEDIA(0x3, 0x0, 0x3C) /* HSW+ */
+#define OP_3DSTATE_DX9_CONSTANTB_VS             OP_3D_MEDIA(0x3, 0x0, 0x3D) /* HSW+ */
+#define OP_3DSTATE_DX9_CONSTANTB_PS             OP_3D_MEDIA(0x3, 0x0, 0x3E) /* HSW+ */
+#define OP_3DSTATE_DX9_LOCAL_VALID_VS           OP_3D_MEDIA(0x3, 0x0, 0x3F) /* HSW+ */
+#define OP_3DSTATE_DX9_LOCAL_VALID_PS           OP_3D_MEDIA(0x3, 0x0, 0x40) /* HSW+ */
+#define OP_3DSTATE_DX9_GENERATE_ACTIVE_VS       OP_3D_MEDIA(0x3, 0x0, 0x41) /* HSW+ */
+#define OP_3DSTATE_DX9_GENERATE_ACTIVE_PS       OP_3D_MEDIA(0x3, 0x0, 0x42) /* HSW+ */
+#define OP_3DSTATE_BINDING_TABLE_EDIT_VS        OP_3D_MEDIA(0x3, 0x0, 0x43) /* HSW+ */
+#define OP_3DSTATE_BINDING_TABLE_EDIT_GS        OP_3D_MEDIA(0x3, 0x0, 0x44) /* HSW+ */
+#define OP_3DSTATE_BINDING_TABLE_EDIT_HS        OP_3D_MEDIA(0x3, 0x0, 0x45) /* HSW+ */
+#define OP_3DSTATE_BINDING_TABLE_EDIT_DS        OP_3D_MEDIA(0x3, 0x0, 0x46) /* HSW+ */
+#define OP_3DSTATE_BINDING_TABLE_EDIT_PS        OP_3D_MEDIA(0x3, 0x0, 0x47) /* HSW+ */
+
+#define OP_3DSTATE_VF_INSTANCING 		OP_3D_MEDIA(0x3, 0x0, 0x49) /* BDW+ */
+#define OP_3DSTATE_VF_SGVS  			OP_3D_MEDIA(0x3, 0x0, 0x4A) /* BDW+ */
+#define OP_3DSTATE_VF_TOPOLOGY   		OP_3D_MEDIA(0x3, 0x0, 0x4B) /* BDW+ */
+#define OP_3DSTATE_WM_CHROMAKEY   		OP_3D_MEDIA(0x3, 0x0, 0x4C) /* BDW+ */
+#define OP_3DSTATE_PS_BLEND   			OP_3D_MEDIA(0x3, 0x0, 0x4D) /* BDW+ */
+#define OP_3DSTATE_WM_DEPTH_STENCIL   		OP_3D_MEDIA(0x3, 0x0, 0x4E) /* BDW+ */
+#define OP_3DSTATE_PS_EXTRA   			OP_3D_MEDIA(0x3, 0x0, 0x4F) /* BDW+ */
+#define OP_3DSTATE_RASTER   			OP_3D_MEDIA(0x3, 0x0, 0x50) /* BDW+ */
+#define OP_3DSTATE_SBE_SWIZ   			OP_3D_MEDIA(0x3, 0x0, 0x51) /* BDW+ */
+#define OP_3DSTATE_WM_HZ_OP   			OP_3D_MEDIA(0x3, 0x0, 0x52) /* BDW+ */
+#define OP_3DSTATE_COMPONENT_PACKING		OP_3D_MEDIA(0x3, 0x0, 0x55) /* SKL+ */
+
+#define OP_3DSTATE_DRAWING_RECTANGLE            OP_3D_MEDIA(0x3, 0x1, 0x00)
+#define OP_3DSTATE_SAMPLER_PALETTE_LOAD0        OP_3D_MEDIA(0x3, 0x1, 0x02)
+#define OP_3DSTATE_CHROMA_KEY                   OP_3D_MEDIA(0x3, 0x1, 0x04)
+#define OP_SNB_3DSTATE_DEPTH_BUFFER             OP_3D_MEDIA(0x3, 0x1, 0x05)
+#define OP_3DSTATE_POLY_STIPPLE_OFFSET          OP_3D_MEDIA(0x3, 0x1, 0x06)
+#define OP_3DSTATE_POLY_STIPPLE_PATTERN         OP_3D_MEDIA(0x3, 0x1, 0x07)
+#define OP_3DSTATE_LINE_STIPPLE                 OP_3D_MEDIA(0x3, 0x1, 0x08)
+#define OP_3DSTATE_AA_LINE_PARAMS               OP_3D_MEDIA(0x3, 0x1, 0x0A)
+#define OP_3DSTATE_GS_SVB_INDEX                 OP_3D_MEDIA(0x3, 0x1, 0x0B)
+#define OP_3DSTATE_SAMPLER_PALETTE_LOAD1        OP_3D_MEDIA(0x3, 0x1, 0x0C)
+#define OP_3DSTATE_MULTISAMPLE_BDW		OP_3D_MEDIA(0x3, 0x0, 0x0D)
+#define OP_SNB_3DSTATE_STENCIL_BUFFER           OP_3D_MEDIA(0x3, 0x1, 0x0E)
+#define OP_SNB_3DSTATE_HIER_DEPTH_BUFFER        OP_3D_MEDIA(0x3, 0x1, 0x0F)
+#define OP_SNB_3DSTATE_CLEAR_PARAMS             OP_3D_MEDIA(0x3, 0x1, 0x10)
+#define OP_3DSTATE_MONOFILTER_SIZE              OP_3D_MEDIA(0x3, 0x1, 0x11)
+#define OP_3DSTATE_PUSH_CONSTANT_ALLOC_VS       OP_3D_MEDIA(0x3, 0x1, 0x12) /* IVB+ */
+#define OP_3DSTATE_PUSH_CONSTANT_ALLOC_HS       OP_3D_MEDIA(0x3, 0x1, 0x13) /* IVB+ */
+#define OP_3DSTATE_PUSH_CONSTANT_ALLOC_DS       OP_3D_MEDIA(0x3, 0x1, 0x14) /* IVB+ */
+#define OP_3DSTATE_PUSH_CONSTANT_ALLOC_GS       OP_3D_MEDIA(0x3, 0x1, 0x15) /* IVB+ */
+#define OP_3DSTATE_PUSH_CONSTANT_ALLOC_PS       OP_3D_MEDIA(0x3, 0x1, 0x16) /* IVB+ */
+#define OP_3DSTATE_SO_DECL_LIST                 OP_3D_MEDIA(0x3, 0x1, 0x17)
+#define OP_3DSTATE_SO_BUFFER                    OP_3D_MEDIA(0x3, 0x1, 0x18)
+#define OP_3DSTATE_BINDING_TABLE_POOL_ALLOC     OP_3D_MEDIA(0x3, 0x1, 0x19) /* HSW+ */
+#define OP_3DSTATE_GATHER_POOL_ALLOC            OP_3D_MEDIA(0x3, 0x1, 0x1A) /* HSW+ */
+#define OP_3DSTATE_DX9_CONSTANT_BUFFER_POOL_ALLOC OP_3D_MEDIA(0x3, 0x1, 0x1B) /* HSW+ */
+#define OP_3DSTATE_SAMPLE_PATTERN               OP_3D_MEDIA(0x3, 0x1, 0x1C)
+#define OP_PIPE_CONTROL                         OP_3D_MEDIA(0x3, 0x2, 0x00)
+#define OP_3DPRIMITIVE                          OP_3D_MEDIA(0x3, 0x3, 0x00)
+
+/* VCCP Command Parser */
+
+/*
+ * Below MFX and VBE cmd definition is from vaapi intel driver project (BSD License)
+ * git://anongit.freedesktop.org/vaapi/intel-driver
+ * src/i965_defines.h
+ *
+ */
+
+#define OP_MFX(pipeline, op, sub_opa, sub_opb)     \
+	(3 << 13 | \
+	 (pipeline) << 11 | \
+	 (op) << 8 | \
+	 (sub_opa) << 5 | \
+	 (sub_opb))
+
+#define OP_MFX_PIPE_MODE_SELECT                    OP_MFX(2, 0, 0, 0)  /* ALL */
+#define OP_MFX_SURFACE_STATE                       OP_MFX(2, 0, 0, 1)  /* ALL */
+#define OP_MFX_PIPE_BUF_ADDR_STATE                 OP_MFX(2, 0, 0, 2)  /* ALL */
+#define OP_MFX_IND_OBJ_BASE_ADDR_STATE             OP_MFX(2, 0, 0, 3)  /* ALL */
+#define OP_MFX_BSP_BUF_BASE_ADDR_STATE             OP_MFX(2, 0, 0, 4)  /* ALL */
+#define OP_2_0_0_5                                 OP_MFX(2, 0, 0, 5)  /* ALL */
+#define OP_MFX_STATE_POINTER                       OP_MFX(2, 0, 0, 6)  /* ALL */
+#define OP_MFX_QM_STATE                            OP_MFX(2, 0, 0, 7)  /* IVB+ */
+#define OP_MFX_FQM_STATE                           OP_MFX(2, 0, 0, 8)  /* IVB+ */
+#define OP_MFX_PAK_INSERT_OBJECT                   OP_MFX(2, 0, 2, 8)  /* IVB+ */
+#define OP_MFX_STITCH_OBJECT                       OP_MFX(2, 0, 2, 0xA)  /* IVB+ */
+
+#define OP_MFD_IT_OBJECT                           OP_MFX(2, 0, 1, 9) /* ALL */
+
+#define OP_MFX_WAIT                                OP_MFX(1, 0, 0, 0) /* IVB+ */
+#define OP_MFX_AVC_IMG_STATE                       OP_MFX(2, 1, 0, 0) /* ALL */
+#define OP_MFX_AVC_QM_STATE                        OP_MFX(2, 1, 0, 1) /* ALL */
+#define OP_MFX_AVC_DIRECTMODE_STATE                OP_MFX(2, 1, 0, 2) /* ALL */
+#define OP_MFX_AVC_SLICE_STATE                     OP_MFX(2, 1, 0, 3) /* ALL */
+#define OP_MFX_AVC_REF_IDX_STATE                   OP_MFX(2, 1, 0, 4) /* ALL */
+#define OP_MFX_AVC_WEIGHTOFFSET_STATE              OP_MFX(2, 1, 0, 5) /* ALL */
+#define OP_MFD_AVC_PICID_STATE                     OP_MFX(2, 1, 1, 5) /* HSW+ */
+#define OP_MFD_AVC_DPB_STATE			   OP_MFX(2, 1, 1, 6) /* IVB+ */
+#define OP_MFD_AVC_SLICEADDR                       OP_MFX(2, 1, 1, 7) /* IVB+ */
+#define OP_MFD_AVC_BSD_OBJECT                      OP_MFX(2, 1, 1, 8) /* ALL */
+#define OP_MFC_AVC_PAK_OBJECT                      OP_MFX(2, 1, 2, 9) /* ALL */
+
+#define OP_MFX_VC1_PRED_PIPE_STATE                 OP_MFX(2, 2, 0, 1) /* ALL */
+#define OP_MFX_VC1_DIRECTMODE_STATE                OP_MFX(2, 2, 0, 2) /* ALL */
+#define OP_MFD_VC1_SHORT_PIC_STATE                 OP_MFX(2, 2, 1, 0) /* IVB+ */
+#define OP_MFD_VC1_LONG_PIC_STATE                  OP_MFX(2, 2, 1, 1) /* IVB+ */
+#define OP_MFD_VC1_BSD_OBJECT                      OP_MFX(2, 2, 1, 8) /* ALL */
+
+#define OP_MFX_MPEG2_PIC_STATE                     OP_MFX(2, 3, 0, 0) /* ALL */
+#define OP_MFX_MPEG2_QM_STATE                      OP_MFX(2, 3, 0, 1) /* ALL */
+#define OP_MFD_MPEG2_BSD_OBJECT                    OP_MFX(2, 3, 1, 8) /* ALL */
+#define OP_MFC_MPEG2_SLICEGROUP_STATE              OP_MFX(2, 3, 2, 3) /* ALL */
+#define OP_MFC_MPEG2_PAK_OBJECT                    OP_MFX(2, 3, 2, 9) /* ALL */
+
+#define OP_MFX_2_6_0_0                             OP_MFX(2, 6, 0, 0) /* IVB+ */
+#define OP_MFX_2_6_0_8                             OP_MFX(2, 6, 0, 8) /* IVB+ */
+#define OP_MFX_2_6_0_9                             OP_MFX(2, 6, 0, 9) /* IVB+ */
+
+#define OP_MFX_JPEG_PIC_STATE                      OP_MFX(2, 7, 0, 0)
+#define OP_MFX_JPEG_HUFF_TABLE_STATE               OP_MFX(2, 7, 0, 2)
+#define OP_MFD_JPEG_BSD_OBJECT                     OP_MFX(2, 7, 1, 8)
+
+#define OP_VEB(pipeline, op, sub_opa, sub_opb) \
+	(3 << 13 | \
+	 (pipeline) << 11 | \
+	 (op) << 8 | \
+	 (sub_opa) << 5 | \
+	 (sub_opb))
+
+#define OP_VEB_SURFACE_STATE                       OP_VEB(2, 4, 0, 0)
+#define OP_VEB_STATE                               OP_VEB(2, 4, 0, 2)
+#define OP_VEB_DNDI_IECP_STATE                     OP_VEB(2, 4, 0, 3)
+
+struct parser_exec_state;
+
+typedef int (*parser_cmd_handler)(struct parser_exec_state *s);
+
+#define GVT_CMD_HASH_BITS   7
+
+/* which DWords need address fix */
+#define ADDR_FIX_1(x1)			(1 << (x1))
+#define ADDR_FIX_2(x1, x2)		(ADDR_FIX_1(x1) | ADDR_FIX_1(x2))
+#define ADDR_FIX_3(x1, x2, x3)		(ADDR_FIX_1(x1) | ADDR_FIX_2(x2, x3))
+#define ADDR_FIX_4(x1, x2, x3, x4)	(ADDR_FIX_1(x1) | ADDR_FIX_3(x2, x3, x4))
+#define ADDR_FIX_5(x1, x2, x3, x4, x5)  (ADDR_FIX_1(x1) | ADDR_FIX_4(x2, x3, x4, x5))
+
+#define DWORD_FIELD(dword, end, start) \
+	FIELD_GET(GENMASK(end, start), cmd_val(s, dword))
+
+#define OP_LENGTH_BIAS 2
+#define CMD_LEN(value)  (value + OP_LENGTH_BIAS)
+
+static int gvt_check_valid_cmd_length(int len, int valid_len)
+{
+	if (valid_len != len) {
+		gvt_err("len is not valid:  len=%u  valid_len=%u\n",
+			len, valid_len);
+		return -EFAULT;
+	}
+	return 0;
+}
+
+struct cmd_info {
+	const char *name;
+	u32 opcode;
+
+#define F_LEN_MASK	3U
+#define F_LEN_CONST  1U
+#define F_LEN_VAR    0U
+/* value is const although LEN maybe variable */
+#define F_LEN_VAR_FIXED    (1<<1)
+
+/*
+ * command has its own ip advance logic
+ * e.g. MI_BATCH_START, MI_BATCH_END
+ */
+#define F_IP_ADVANCE_CUSTOM (1<<2)
+	u32 flag;
+
+#define R_RCS	BIT(RCS0)
+#define R_VCS1  BIT(VCS0)
+#define R_VCS2  BIT(VCS1)
+#define R_VCS	(R_VCS1 | R_VCS2)
+#define R_BCS	BIT(BCS0)
+#define R_VECS	BIT(VECS0)
+#define R_ALL (R_RCS | R_VCS | R_BCS | R_VECS)
+	/* rings that support this cmd: BLT/RCS/VCS/VECS */
+	intel_engine_mask_t rings;
+
+	/* devices that support this cmd: SNB/IVB/HSW/... */
+	u16 devices;
+
+	/* which DWords are address that need fix up.
+	 * bit 0 means a 32-bit non address operand in command
+	 * bit 1 means address operand, which could be 32-bit
+	 * or 64-bit depending on different architectures.(
+	 * defined by "gmadr_bytes_in_cmd" in intel_gvt.
+	 * No matter the address length, each address only takes
+	 * one bit in the bitmap.
+	 */
+	u16 addr_bitmap;
+
+	/* flag == F_LEN_CONST : command length
+	 * flag == F_LEN_VAR : length bias bits
+	 * Note: length is in DWord
+	 */
+	u32 len;
+
+	parser_cmd_handler handler;
+
+	/* valid length in DWord */
+	u32 valid_len;
+};
+
+struct cmd_entry {
+	struct hlist_node hlist;
+	const struct cmd_info *info;
+};
+
+enum {
+	RING_BUFFER_INSTRUCTION,
+	BATCH_BUFFER_INSTRUCTION,
+	BATCH_BUFFER_2ND_LEVEL,
+	RING_BUFFER_CTX,
+};
+
+enum {
+	GTT_BUFFER,
+	PPGTT_BUFFER
+};
+
+struct parser_exec_state {
+	struct intel_vgpu *vgpu;
+	const struct intel_engine_cs *engine;
+
+	int buf_type;
+
+	/* batch buffer address type */
+	int buf_addr_type;
+
+	/* graphics memory address of ring buffer start */
+	unsigned long ring_start;
+	unsigned long ring_size;
+	unsigned long ring_head;
+	unsigned long ring_tail;
+
+	/* instruction graphics memory address */
+	unsigned long ip_gma;
+
+	/* mapped va of the instr_gma */
+	void *ip_va;
+	void *rb_va;
+
+	void *ret_bb_va;
+	/* next instruction when return from  batch buffer to ring buffer */
+	unsigned long ret_ip_gma_ring;
+
+	/* next instruction when return from 2nd batch buffer to batch buffer */
+	unsigned long ret_ip_gma_bb;
+
+	/* batch buffer address type (GTT or PPGTT)
+	 * used when ret from 2nd level batch buffer
+	 */
+	int saved_buf_addr_type;
+	bool is_ctx_wa;
+	bool is_init_ctx;
+
+	const struct cmd_info *info;
+
+	struct intel_vgpu_workload *workload;
+};
+
+#define gmadr_dw_number(s)	\
+	(s->vgpu->gvt->device_info.gmadr_bytes_in_cmd >> 2)
+
+static unsigned long bypass_scan_mask = 0;
+
+/* ring ALL, type = 0 */
+static const struct sub_op_bits sub_op_mi[] = {
+	{31, 29},
+	{28, 23},
+};
+
+static const struct decode_info decode_info_mi = {
+	"MI",
+	OP_LEN_MI,
+	ARRAY_SIZE(sub_op_mi),
+	sub_op_mi,
+};
+
+/* ring RCS, command type 2 */
+static const struct sub_op_bits sub_op_2d[] = {
+	{31, 29},
+	{28, 22},
+};
+
+static const struct decode_info decode_info_2d = {
+	"2D",
+	OP_LEN_2D,
+	ARRAY_SIZE(sub_op_2d),
+	sub_op_2d,
+};
+
+/* ring RCS, command type 3 */
+static const struct sub_op_bits sub_op_3d_media[] = {
+	{31, 29},
+	{28, 27},
+	{26, 24},
+	{23, 16},
+};
+
+static const struct decode_info decode_info_3d_media = {
+	"3D_Media",
+	OP_LEN_3D_MEDIA,
+	ARRAY_SIZE(sub_op_3d_media),
+	sub_op_3d_media,
+};
+
+/* ring VCS, command type 3 */
+static const struct sub_op_bits sub_op_mfx_vc[] = {
+	{31, 29},
+	{28, 27},
+	{26, 24},
+	{23, 21},
+	{20, 16},
+};
+
+static const struct decode_info decode_info_mfx_vc = {
+	"MFX_VC",
+	OP_LEN_MFX_VC,
+	ARRAY_SIZE(sub_op_mfx_vc),
+	sub_op_mfx_vc,
+};
+
+/* ring VECS, command type 3 */
+static const struct sub_op_bits sub_op_vebox[] = {
+	{31, 29},
+	{28, 27},
+	{26, 24},
+	{23, 21},
+	{20, 16},
+};
+
+static const struct decode_info decode_info_vebox = {
+	"VEBOX",
+	OP_LEN_VEBOX,
+	ARRAY_SIZE(sub_op_vebox),
+	sub_op_vebox,
+};
+
+static const struct decode_info *ring_decode_info[I915_NUM_ENGINES][8] = {
+	[RCS0] = {
+		&decode_info_mi,
+		NULL,
+		NULL,
+		&decode_info_3d_media,
+		NULL,
+		NULL,
+		NULL,
+		NULL,
+	},
+
+	[VCS0] = {
+		&decode_info_mi,
+		NULL,
+		NULL,
+		&decode_info_mfx_vc,
+		NULL,
+		NULL,
+		NULL,
+		NULL,
+	},
+
+	[BCS0] = {
+		&decode_info_mi,
+		NULL,
+		&decode_info_2d,
+		NULL,
+		NULL,
+		NULL,
+		NULL,
+		NULL,
+	},
+
+	[VECS0] = {
+		&decode_info_mi,
+		NULL,
+		NULL,
+		&decode_info_vebox,
+		NULL,
+		NULL,
+		NULL,
+		NULL,
+	},
+
+	[VCS1] = {
+		&decode_info_mi,
+		NULL,
+		NULL,
+		&decode_info_mfx_vc,
+		NULL,
+		NULL,
+		NULL,
+		NULL,
+	},
+};
+
+static inline u32 get_opcode(u32 cmd, const struct intel_engine_cs *engine)
+{
+	const struct decode_info *d_info;
+
+	d_info = ring_decode_info[engine->id][CMD_TYPE(cmd)];
+	if (d_info == NULL)
+		return INVALID_OP;
+
+	return cmd >> (32 - d_info->op_len);
+}
+
+static inline const struct cmd_info *
+find_cmd_entry(struct intel_gvt *gvt, unsigned int opcode,
+	       const struct intel_engine_cs *engine)
+{
+	struct cmd_entry *e;
+
+	hash_for_each_possible(gvt->cmd_table, e, hlist, opcode) {
+		if (opcode == e->info->opcode &&
+		    e->info->rings & engine->mask)
+			return e->info;
+	}
+	return NULL;
+}
+
+static inline const struct cmd_info *
+get_cmd_info(struct intel_gvt *gvt, u32 cmd,
+	     const struct intel_engine_cs *engine)
+{
+	u32 opcode;
+
+	opcode = get_opcode(cmd, engine);
+	if (opcode == INVALID_OP)
+		return NULL;
+
+	return find_cmd_entry(gvt, opcode, engine);
+}
+
+static inline u32 sub_op_val(u32 cmd, u32 hi, u32 low)
+{
+	return (cmd >> low) & ((1U << (hi - low + 1)) - 1);
+}
+
+static inline void print_opcode(u32 cmd, const struct intel_engine_cs *engine)
+{
+	const struct decode_info *d_info;
+	int i;
+
+	d_info = ring_decode_info[engine->id][CMD_TYPE(cmd)];
+	if (d_info == NULL)
+		return;
+
+	gvt_dbg_cmd("opcode=0x%x %s sub_ops:",
+			cmd >> (32 - d_info->op_len), d_info->name);
+
+	for (i = 0; i < d_info->nr_sub_op; i++)
+		pr_err("0x%x ", sub_op_val(cmd, d_info->sub_op[i].hi,
+					d_info->sub_op[i].low));
+
+	pr_err("\n");
+}
+
+static inline u32 *cmd_ptr(struct parser_exec_state *s, int index)
+{
+	return s->ip_va + (index << 2);
+}
+
+static inline u32 cmd_val(struct parser_exec_state *s, int index)
+{
+	return *cmd_ptr(s, index);
+}
+
+static inline bool is_init_ctx(struct parser_exec_state *s)
+{
+	return (s->buf_type == RING_BUFFER_CTX && s->is_init_ctx);
+}
+
+static void parser_exec_state_dump(struct parser_exec_state *s)
+{
+	int cnt = 0;
+	int i;
+
+	gvt_dbg_cmd("  vgpu%d RING%s: ring_start(%08lx) ring_end(%08lx)"
+		    " ring_head(%08lx) ring_tail(%08lx)\n",
+		    s->vgpu->id, s->engine->name,
+		    s->ring_start, s->ring_start + s->ring_size,
+		    s->ring_head, s->ring_tail);
+
+	gvt_dbg_cmd("  %s %s ip_gma(%08lx) ",
+			s->buf_type == RING_BUFFER_INSTRUCTION ?
+			"RING_BUFFER" : ((s->buf_type == RING_BUFFER_CTX) ?
+				"CTX_BUFFER" : "BATCH_BUFFER"),
+			s->buf_addr_type == GTT_BUFFER ?
+			"GTT" : "PPGTT", s->ip_gma);
+
+	if (s->ip_va == NULL) {
+		gvt_dbg_cmd(" ip_va(NULL)");
+		return;
+	}
+
+	gvt_dbg_cmd("  ip_va=%p: %08x %08x %08x %08x\n",
+			s->ip_va, cmd_val(s, 0), cmd_val(s, 1),
+			cmd_val(s, 2), cmd_val(s, 3));
+
+	print_opcode(cmd_val(s, 0), s->engine);
+
+	s->ip_va = (u32 *)((((u64)s->ip_va) >> 12) << 12);
+
+	while (cnt < 1024) {
+		gvt_dbg_cmd("ip_va=%p: ", s->ip_va);
+		for (i = 0; i < 8; i++)
+			gvt_dbg_cmd("%08x ", cmd_val(s, i));
+		gvt_dbg_cmd("\n");
+
+		s->ip_va += 8 * sizeof(u32);
+		cnt += 8;
+	}
+}
+
+static inline void update_ip_va(struct parser_exec_state *s)
+{
+	unsigned long len = 0;
+
+	if (WARN_ON(s->ring_head == s->ring_tail))
+		return;
+
+	if (s->buf_type == RING_BUFFER_INSTRUCTION ||
+			s->buf_type == RING_BUFFER_CTX) {
+		unsigned long ring_top = s->ring_start + s->ring_size;
+
+		if (s->ring_head > s->ring_tail) {
+			if (s->ip_gma >= s->ring_head && s->ip_gma < ring_top)
+				len = (s->ip_gma - s->ring_head);
+			else if (s->ip_gma >= s->ring_start &&
+					s->ip_gma <= s->ring_tail)
+				len = (ring_top - s->ring_head) +
+					(s->ip_gma - s->ring_start);
+		} else
+			len = (s->ip_gma - s->ring_head);
+
+		s->ip_va = s->rb_va + len;
+	} else {/* shadow batch buffer */
+		s->ip_va = s->ret_bb_va;
+	}
+}
+
+static inline int ip_gma_set(struct parser_exec_state *s,
+		unsigned long ip_gma)
+{
+	WARN_ON(!IS_ALIGNED(ip_gma, 4));
+
+	s->ip_gma = ip_gma;
+	update_ip_va(s);
+	return 0;
+}
+
+static inline int ip_gma_advance(struct parser_exec_state *s,
+		unsigned int dw_len)
+{
+	s->ip_gma += (dw_len << 2);
+
+	if (s->buf_type == RING_BUFFER_INSTRUCTION) {
+		if (s->ip_gma >= s->ring_start + s->ring_size)
+			s->ip_gma -= s->ring_size;
+		update_ip_va(s);
+	} else {
+		s->ip_va += (dw_len << 2);
+	}
+
+	return 0;
+}
+
+static inline int get_cmd_length(const struct cmd_info *info, u32 cmd)
+{
+	if ((info->flag & F_LEN_MASK) == F_LEN_CONST)
+		return info->len;
+	else
+		return (cmd & ((1U << info->len) - 1)) + 2;
+	return 0;
+}
+
+static inline int cmd_length(struct parser_exec_state *s)
+{
+	return get_cmd_length(s->info, cmd_val(s, 0));
+}
+
+/* do not remove this, some platform may need clflush here */
+#define patch_value(s, addr, val) do { \
+	*addr = val; \
+} while (0)
+
+static inline bool is_mocs_mmio(unsigned int offset)
+{
+	return ((offset >= 0xc800) && (offset <= 0xcff8)) ||
+		((offset >= 0xb020) && (offset <= 0xb0a0));
+}
+
+static int is_cmd_update_pdps(unsigned int offset,
+			      struct parser_exec_state *s)
+{
+	u32 base = s->workload->engine->mmio_base;
+	return i915_mmio_reg_equal(_MMIO(offset), GEN8_RING_PDP_UDW(base, 0));
+}
+
+static int cmd_pdp_mmio_update_handler(struct parser_exec_state *s,
+				       unsigned int offset, unsigned int index)
+{
+	struct intel_vgpu *vgpu = s->vgpu;
+	struct intel_vgpu_mm *shadow_mm = s->workload->shadow_mm;
+	struct intel_vgpu_mm *mm;
+	u64 pdps[GEN8_3LVL_PDPES];
+
+	if (shadow_mm->ppgtt_mm.root_entry_type ==
+	    GTT_TYPE_PPGTT_ROOT_L4_ENTRY) {
+		pdps[0] = (u64)cmd_val(s, 2) << 32;
+		pdps[0] |= cmd_val(s, 4);
+
+		mm = intel_vgpu_find_ppgtt_mm(vgpu, pdps);
+		if (!mm) {
+			gvt_vgpu_err("failed to get the 4-level shadow vm\n");
+			return -EINVAL;
+		}
+		intel_vgpu_mm_get(mm);
+		list_add_tail(&mm->ppgtt_mm.link,
+			      &s->workload->lri_shadow_mm);
+		*cmd_ptr(s, 2) = upper_32_bits(mm->ppgtt_mm.shadow_pdps[0]);
+		*cmd_ptr(s, 4) = lower_32_bits(mm->ppgtt_mm.shadow_pdps[0]);
+	} else {
+		/* Currently all guests use PML4 table and now can't
+		 * have a guest with 3-level table but uses LRI for
+		 * PPGTT update. So this is simply un-testable. */
+		GEM_BUG_ON(1);
+		gvt_vgpu_err("invalid shared shadow vm type\n");
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int cmd_reg_handler(struct parser_exec_state *s,
+	unsigned int offset, unsigned int index, char *cmd)
+{
+	struct intel_vgpu *vgpu = s->vgpu;
+	struct intel_gvt *gvt = vgpu->gvt;
+	u32 ctx_sr_ctl;
+	u32 *vreg, vreg_old;
+
+	if (offset + 4 > gvt->device_info.mmio_size) {
+		gvt_vgpu_err("%s access to (%x) outside of MMIO range\n",
+				cmd, offset);
+		return -EFAULT;
+	}
+
+	if (is_init_ctx(s)) {
+		struct intel_gvt_mmio_info *mmio_info;
+
+		intel_gvt_mmio_set_cmd_accessible(gvt, offset);
+		mmio_info = intel_gvt_find_mmio_info(gvt, offset);
+		if (mmio_info && mmio_info->write)
+			intel_gvt_mmio_set_cmd_write_patch(gvt, offset);
+		return 0;
+	}
+
+	if (!intel_gvt_mmio_is_cmd_accessible(gvt, offset)) {
+		gvt_vgpu_err("%s access to non-render register (%x)\n",
+				cmd, offset);
+		return -EBADRQC;
+	}
+
+	if (!strncmp(cmd, "srm", 3) ||
+			!strncmp(cmd, "lrm", 3)) {
+		if (offset == i915_mmio_reg_offset(GEN8_L3SQCREG4) ||
+		    offset == 0x21f0 ||
+		    (IS_BROADWELL(gvt->gt->i915) &&
+		     offset == i915_mmio_reg_offset(INSTPM)))
+			return 0;
+		else {
+			gvt_vgpu_err("%s access to register (%x)\n",
+					cmd, offset);
+			return -EPERM;
+		}
+	}
+
+	if (!strncmp(cmd, "lrr-src", 7) ||
+			!strncmp(cmd, "lrr-dst", 7)) {
+		if (IS_BROADWELL(gvt->gt->i915) && offset == 0x215c)
+			return 0;
+		else {
+			gvt_vgpu_err("not allowed cmd %s reg (%x)\n", cmd, offset);
+			return -EPERM;
+		}
+	}
+
+	if (!strncmp(cmd, "pipe_ctrl", 9)) {
+		/* TODO: add LRI POST logic here */
+		return 0;
+	}
+
+	if (strncmp(cmd, "lri", 3))
+		return -EPERM;
+
+	/* below are all lri handlers */
+	vreg = &vgpu_vreg(s->vgpu, offset);
+
+	if (is_cmd_update_pdps(offset, s) &&
+	    cmd_pdp_mmio_update_handler(s, offset, index))
+		return -EINVAL;
+
+	if (offset == i915_mmio_reg_offset(DERRMR) ||
+		offset == i915_mmio_reg_offset(FORCEWAKE_MT)) {
+		/* Writing to HW VGT_PVINFO_PAGE offset will be discarded */
+		patch_value(s, cmd_ptr(s, index), VGT_PVINFO_PAGE);
+	}
+
+	if (is_mocs_mmio(offset))
+		*vreg = cmd_val(s, index + 1);
+
+	vreg_old = *vreg;
+
+	if (intel_gvt_mmio_is_cmd_write_patch(gvt, offset)) {
+		u32 cmdval_new, cmdval;
+		struct intel_gvt_mmio_info *mmio_info;
+
+		cmdval = cmd_val(s, index + 1);
+
+		mmio_info = intel_gvt_find_mmio_info(gvt, offset);
+		if (!mmio_info) {
+			cmdval_new = cmdval;
+		} else {
+			u64 ro_mask = mmio_info->ro_mask;
+			int ret;
+
+			if (likely(!ro_mask))
+				ret = mmio_info->write(s->vgpu, offset,
+						&cmdval, 4);
+			else {
+				gvt_vgpu_err("try to write RO reg %x\n",
+						offset);
+				ret = -EBADRQC;
+			}
+			if (ret)
+				return ret;
+			cmdval_new = *vreg;
+		}
+		if (cmdval_new != cmdval)
+			patch_value(s, cmd_ptr(s, index+1), cmdval_new);
+	}
+
+	/* only patch cmd. restore vreg value if changed in mmio write handler*/
+	*vreg = vreg_old;
+
+	/* TODO
+	 * In order to let workload with inhibit context to generate
+	 * correct image data into memory, vregs values will be loaded to
+	 * hw via LRIs in the workload with inhibit context. But as
+	 * indirect context is loaded prior to LRIs in workload, we don't
+	 * want reg values specified in indirect context overwritten by
+	 * LRIs in workloads. So, when scanning an indirect context, we
+	 * update reg values in it into vregs, so LRIs in workload with
+	 * inhibit context will restore with correct values
+	 */
+	if (GRAPHICS_VER(s->engine->i915) == 9 &&
+	    intel_gvt_mmio_is_sr_in_ctx(gvt, offset) &&
+	    !strncmp(cmd, "lri", 3)) {
+		intel_gvt_read_gpa(s->vgpu,
+			s->workload->ring_context_gpa + 12, &ctx_sr_ctl, 4);
+		/* check inhibit context */
+		if (ctx_sr_ctl & 1) {
+			u32 data = cmd_val(s, index + 1);
+
+			if (intel_gvt_mmio_has_mode_mask(s->vgpu->gvt, offset))
+				intel_vgpu_mask_mmio_write(vgpu,
+							offset, &data, 4);
+			else
+				vgpu_vreg(vgpu, offset) = data;
+		}
+	}
+
+	return 0;
+}
+
+#define cmd_reg(s, i) \
+	(cmd_val(s, i) & GENMASK(22, 2))
+
+#define cmd_reg_inhibit(s, i) \
+	(cmd_val(s, i) & GENMASK(22, 18))
+
+#define cmd_gma(s, i) \
+	(cmd_val(s, i) & GENMASK(31, 2))
+
+#define cmd_gma_hi(s, i) \
+	(cmd_val(s, i) & GENMASK(15, 0))
+
+static int cmd_handler_lri(struct parser_exec_state *s)
+{
+	int i, ret = 0;
+	int cmd_len = cmd_length(s);
+
+	for (i = 1; i < cmd_len; i += 2) {
+		if (IS_BROADWELL(s->engine->i915) && s->engine->id != RCS0) {
+			if (s->engine->id == BCS0 &&
+			    cmd_reg(s, i) == i915_mmio_reg_offset(DERRMR))
+				ret |= 0;
+			else
+				ret |= cmd_reg_inhibit(s, i) ? -EBADRQC : 0;
+		}
+		if (ret)
+			break;
+		ret |= cmd_reg_handler(s, cmd_reg(s, i), i, "lri");
+		if (ret)
+			break;
+	}
+	return ret;
+}
+
+static int cmd_handler_lrr(struct parser_exec_state *s)
+{
+	int i, ret = 0;
+	int cmd_len = cmd_length(s);
+
+	for (i = 1; i < cmd_len; i += 2) {
+		if (IS_BROADWELL(s->engine->i915))
+			ret |= ((cmd_reg_inhibit(s, i) ||
+				 (cmd_reg_inhibit(s, i + 1)))) ?
+				-EBADRQC : 0;
+		if (ret)
+			break;
+		ret |= cmd_reg_handler(s, cmd_reg(s, i), i, "lrr-src");
+		if (ret)
+			break;
+		ret |= cmd_reg_handler(s, cmd_reg(s, i + 1), i, "lrr-dst");
+		if (ret)
+			break;
+	}
+	return ret;
+}
+
+static inline int cmd_address_audit(struct parser_exec_state *s,
+		unsigned long guest_gma, int op_size, bool index_mode);
+
+static int cmd_handler_lrm(struct parser_exec_state *s)
+{
+	struct intel_gvt *gvt = s->vgpu->gvt;
+	int gmadr_bytes = gvt->device_info.gmadr_bytes_in_cmd;
+	unsigned long gma;
+	int i, ret = 0;
+	int cmd_len = cmd_length(s);
+
+	for (i = 1; i < cmd_len;) {
+		if (IS_BROADWELL(s->engine->i915))
+			ret |= (cmd_reg_inhibit(s, i)) ? -EBADRQC : 0;
+		if (ret)
+			break;
+		ret |= cmd_reg_handler(s, cmd_reg(s, i), i, "lrm");
+		if (ret)
+			break;
+		if (cmd_val(s, 0) & (1 << 22)) {
+			gma = cmd_gma(s, i + 1);
+			if (gmadr_bytes == 8)
+				gma |= (cmd_gma_hi(s, i + 2)) << 32;
+			ret |= cmd_address_audit(s, gma, sizeof(u32), false);
+			if (ret)
+				break;
+		}
+		i += gmadr_dw_number(s) + 1;
+	}
+	return ret;
+}
+
+static int cmd_handler_srm(struct parser_exec_state *s)
+{
+	int gmadr_bytes = s->vgpu->gvt->device_info.gmadr_bytes_in_cmd;
+	unsigned long gma;
+	int i, ret = 0;
+	int cmd_len = cmd_length(s);
+
+	for (i = 1; i < cmd_len;) {
+		ret |= cmd_reg_handler(s, cmd_reg(s, i), i, "srm");
+		if (ret)
+			break;
+		if (cmd_val(s, 0) & (1 << 22)) {
+			gma = cmd_gma(s, i + 1);
+			if (gmadr_bytes == 8)
+				gma |= (cmd_gma_hi(s, i + 2)) << 32;
+			ret |= cmd_address_audit(s, gma, sizeof(u32), false);
+			if (ret)
+				break;
+		}
+		i += gmadr_dw_number(s) + 1;
+	}
+	return ret;
+}
+
+struct cmd_interrupt_event {
+	int pipe_control_notify;
+	int mi_flush_dw;
+	int mi_user_interrupt;
+};
+
+static const struct cmd_interrupt_event cmd_interrupt_events[] = {
+	[RCS0] = {
+		.pipe_control_notify = RCS_PIPE_CONTROL,
+		.mi_flush_dw = INTEL_GVT_EVENT_RESERVED,
+		.mi_user_interrupt = RCS_MI_USER_INTERRUPT,
+	},
+	[BCS0] = {
+		.pipe_control_notify = INTEL_GVT_EVENT_RESERVED,
+		.mi_flush_dw = BCS_MI_FLUSH_DW,
+		.mi_user_interrupt = BCS_MI_USER_INTERRUPT,
+	},
+	[VCS0] = {
+		.pipe_control_notify = INTEL_GVT_EVENT_RESERVED,
+		.mi_flush_dw = VCS_MI_FLUSH_DW,
+		.mi_user_interrupt = VCS_MI_USER_INTERRUPT,
+	},
+	[VCS1] = {
+		.pipe_control_notify = INTEL_GVT_EVENT_RESERVED,
+		.mi_flush_dw = VCS2_MI_FLUSH_DW,
+		.mi_user_interrupt = VCS2_MI_USER_INTERRUPT,
+	},
+	[VECS0] = {
+		.pipe_control_notify = INTEL_GVT_EVENT_RESERVED,
+		.mi_flush_dw = VECS_MI_FLUSH_DW,
+		.mi_user_interrupt = VECS_MI_USER_INTERRUPT,
+	},
+};
+
+static int cmd_handler_pipe_control(struct parser_exec_state *s)
+{
+	int gmadr_bytes = s->vgpu->gvt->device_info.gmadr_bytes_in_cmd;
+	unsigned long gma;
+	bool index_mode = false;
+	unsigned int post_sync;
+	int ret = 0;
+	u32 hws_pga, val;
+
+	post_sync = (cmd_val(s, 1) & PIPE_CONTROL_POST_SYNC_OP_MASK) >> 14;
+
+	/* LRI post sync */
+	if (cmd_val(s, 1) & PIPE_CONTROL_MMIO_WRITE)
+		ret = cmd_reg_handler(s, cmd_reg(s, 2), 1, "pipe_ctrl");
+	/* post sync */
+	else if (post_sync) {
+		if (post_sync == 2)
+			ret = cmd_reg_handler(s, 0x2350, 1, "pipe_ctrl");
+		else if (post_sync == 3)
+			ret = cmd_reg_handler(s, 0x2358, 1, "pipe_ctrl");
+		else if (post_sync == 1) {
+			/* check ggtt*/
+			if ((cmd_val(s, 1) & PIPE_CONTROL_GLOBAL_GTT_IVB)) {
+				gma = cmd_val(s, 2) & GENMASK(31, 3);
+				if (gmadr_bytes == 8)
+					gma |= (cmd_gma_hi(s, 3)) << 32;
+				/* Store Data Index */
+				if (cmd_val(s, 1) & (1 << 21))
+					index_mode = true;
+				ret |= cmd_address_audit(s, gma, sizeof(u64),
+						index_mode);
+				if (ret)
+					return ret;
+				if (index_mode) {
+					hws_pga = s->vgpu->hws_pga[s->engine->id];
+					gma = hws_pga + gma;
+					patch_value(s, cmd_ptr(s, 2), gma);
+					val = cmd_val(s, 1) & (~(1 << 21));
+					patch_value(s, cmd_ptr(s, 1), val);
+				}
+			}
+		}
+	}
+
+	if (ret)
+		return ret;
+
+	if (cmd_val(s, 1) & PIPE_CONTROL_NOTIFY)
+		set_bit(cmd_interrupt_events[s->engine->id].pipe_control_notify,
+			s->workload->pending_events);
+	return 0;
+}
+
+static int cmd_handler_mi_user_interrupt(struct parser_exec_state *s)
+{
+	set_bit(cmd_interrupt_events[s->engine->id].mi_user_interrupt,
+		s->workload->pending_events);
+	patch_value(s, cmd_ptr(s, 0), MI_NOOP);
+	return 0;
+}
+
+static int cmd_advance_default(struct parser_exec_state *s)
+{
+	return ip_gma_advance(s, cmd_length(s));
+}
+
+static int cmd_handler_mi_batch_buffer_end(struct parser_exec_state *s)
+{
+	int ret;
+
+	if (s->buf_type == BATCH_BUFFER_2ND_LEVEL) {
+		s->buf_type = BATCH_BUFFER_INSTRUCTION;
+		ret = ip_gma_set(s, s->ret_ip_gma_bb);
+		s->buf_addr_type = s->saved_buf_addr_type;
+	} else if (s->buf_type == RING_BUFFER_CTX) {
+		ret = ip_gma_set(s, s->ring_tail);
+	} else {
+		s->buf_type = RING_BUFFER_INSTRUCTION;
+		s->buf_addr_type = GTT_BUFFER;
+		if (s->ret_ip_gma_ring >= s->ring_start + s->ring_size)
+			s->ret_ip_gma_ring -= s->ring_size;
+		ret = ip_gma_set(s, s->ret_ip_gma_ring);
+	}
+	return ret;
+}
+
+struct mi_display_flip_command_info {
+	int pipe;
+	int plane;
+	int event;
+	i915_reg_t stride_reg;
+	i915_reg_t ctrl_reg;
+	i915_reg_t surf_reg;
+	u64 stride_val;
+	u64 tile_val;
+	u64 surf_val;
+	bool async_flip;
+};
+
+struct plane_code_mapping {
+	int pipe;
+	int plane;
+	int event;
+};
+
+static int gen8_decode_mi_display_flip(struct parser_exec_state *s,
+		struct mi_display_flip_command_info *info)
+{
+	struct drm_i915_private *dev_priv = s->engine->i915;
+	struct plane_code_mapping gen8_plane_code[] = {
+		[0] = {PIPE_A, PLANE_A, PRIMARY_A_FLIP_DONE},
+		[1] = {PIPE_B, PLANE_A, PRIMARY_B_FLIP_DONE},
+		[2] = {PIPE_A, PLANE_B, SPRITE_A_FLIP_DONE},
+		[3] = {PIPE_B, PLANE_B, SPRITE_B_FLIP_DONE},
+		[4] = {PIPE_C, PLANE_A, PRIMARY_C_FLIP_DONE},
+		[5] = {PIPE_C, PLANE_B, SPRITE_C_FLIP_DONE},
+	};
+	u32 dword0, dword1, dword2;
+	u32 v;
+
+	dword0 = cmd_val(s, 0);
+	dword1 = cmd_val(s, 1);
+	dword2 = cmd_val(s, 2);
+
+	v = (dword0 & GENMASK(21, 19)) >> 19;
+	if (drm_WARN_ON(&dev_priv->drm, v >= ARRAY_SIZE(gen8_plane_code)))
+		return -EBADRQC;
+
+	info->pipe = gen8_plane_code[v].pipe;
+	info->plane = gen8_plane_code[v].plane;
+	info->event = gen8_plane_code[v].event;
+	info->stride_val = (dword1 & GENMASK(15, 6)) >> 6;
+	info->tile_val = (dword1 & 0x1);
+	info->surf_val = (dword2 & GENMASK(31, 12)) >> 12;
+	info->async_flip = ((dword2 & GENMASK(1, 0)) == 0x1);
+
+	if (info->plane == PLANE_A) {
+		info->ctrl_reg = DSPCNTR(info->pipe);
+		info->stride_reg = DSPSTRIDE(info->pipe);
+		info->surf_reg = DSPSURF(info->pipe);
+	} else if (info->plane == PLANE_B) {
+		info->ctrl_reg = SPRCTL(info->pipe);
+		info->stride_reg = SPRSTRIDE(info->pipe);
+		info->surf_reg = SPRSURF(info->pipe);
+	} else {
+		drm_WARN_ON(&dev_priv->drm, 1);
+		return -EBADRQC;
+	}
+	return 0;
+}
+
+static int skl_decode_mi_display_flip(struct parser_exec_state *s,
+		struct mi_display_flip_command_info *info)
+{
+	struct drm_i915_private *dev_priv = s->engine->i915;
+	struct intel_vgpu *vgpu = s->vgpu;
+	u32 dword0 = cmd_val(s, 0);
+	u32 dword1 = cmd_val(s, 1);
+	u32 dword2 = cmd_val(s, 2);
+	u32 plane = (dword0 & GENMASK(12, 8)) >> 8;
+
+	info->plane = PRIMARY_PLANE;
+
+	switch (plane) {
+	case MI_DISPLAY_FLIP_SKL_PLANE_1_A:
+		info->pipe = PIPE_A;
+		info->event = PRIMARY_A_FLIP_DONE;
+		break;
+	case MI_DISPLAY_FLIP_SKL_PLANE_1_B:
+		info->pipe = PIPE_B;
+		info->event = PRIMARY_B_FLIP_DONE;
+		break;
+	case MI_DISPLAY_FLIP_SKL_PLANE_1_C:
+		info->pipe = PIPE_C;
+		info->event = PRIMARY_C_FLIP_DONE;
+		break;
+
+	case MI_DISPLAY_FLIP_SKL_PLANE_2_A:
+		info->pipe = PIPE_A;
+		info->event = SPRITE_A_FLIP_DONE;
+		info->plane = SPRITE_PLANE;
+		break;
+	case MI_DISPLAY_FLIP_SKL_PLANE_2_B:
+		info->pipe = PIPE_B;
+		info->event = SPRITE_B_FLIP_DONE;
+		info->plane = SPRITE_PLANE;
+		break;
+	case MI_DISPLAY_FLIP_SKL_PLANE_2_C:
+		info->pipe = PIPE_C;
+		info->event = SPRITE_C_FLIP_DONE;
+		info->plane = SPRITE_PLANE;
+		break;
+
+	default:
+		gvt_vgpu_err("unknown plane code %d\n", plane);
+		return -EBADRQC;
+	}
+
+	info->stride_val = (dword1 & GENMASK(15, 6)) >> 6;
+	info->tile_val = (dword1 & GENMASK(2, 0));
+	info->surf_val = (dword2 & GENMASK(31, 12)) >> 12;
+	info->async_flip = ((dword2 & GENMASK(1, 0)) == 0x1);
+
+	info->ctrl_reg = DSPCNTR(info->pipe);
+	info->stride_reg = DSPSTRIDE(info->pipe);
+	info->surf_reg = DSPSURF(info->pipe);
+
+	return 0;
+}
+
+static int gen8_check_mi_display_flip(struct parser_exec_state *s,
+		struct mi_display_flip_command_info *info)
+{
+	u32 stride, tile;
+
+	if (!info->async_flip)
+		return 0;
+
+	if (GRAPHICS_VER(s->engine->i915) >= 9) {
+		stride = vgpu_vreg_t(s->vgpu, info->stride_reg) & GENMASK(9, 0);
+		tile = (vgpu_vreg_t(s->vgpu, info->ctrl_reg) &
+				GENMASK(12, 10)) >> 10;
+	} else {
+		stride = (vgpu_vreg_t(s->vgpu, info->stride_reg) &
+				GENMASK(15, 6)) >> 6;
+		tile = (vgpu_vreg_t(s->vgpu, info->ctrl_reg) & (1 << 10)) >> 10;
+	}
+
+	if (stride != info->stride_val)
+		gvt_dbg_cmd("cannot change stride during async flip\n");
+
+	if (tile != info->tile_val)
+		gvt_dbg_cmd("cannot change tile during async flip\n");
+
+	return 0;
+}
+
+static int gen8_update_plane_mmio_from_mi_display_flip(
+		struct parser_exec_state *s,
+		struct mi_display_flip_command_info *info)
+{
+	struct drm_i915_private *dev_priv = s->engine->i915;
+	struct intel_vgpu *vgpu = s->vgpu;
+
+	set_mask_bits(&vgpu_vreg_t(vgpu, info->surf_reg), GENMASK(31, 12),
+		      info->surf_val << 12);
+	if (GRAPHICS_VER(dev_priv) >= 9) {
+		set_mask_bits(&vgpu_vreg_t(vgpu, info->stride_reg), GENMASK(9, 0),
+			      info->stride_val);
+		set_mask_bits(&vgpu_vreg_t(vgpu, info->ctrl_reg), GENMASK(12, 10),
+			      info->tile_val << 10);
+	} else {
+		set_mask_bits(&vgpu_vreg_t(vgpu, info->stride_reg), GENMASK(15, 6),
+			      info->stride_val << 6);
+		set_mask_bits(&vgpu_vreg_t(vgpu, info->ctrl_reg), GENMASK(10, 10),
+			      info->tile_val << 10);
+	}
+
+	if (info->plane == PLANE_PRIMARY)
+		vgpu_vreg_t(vgpu, PIPE_FLIPCOUNT_G4X(info->pipe))++;
+
+	if (info->async_flip)
+		intel_vgpu_trigger_virtual_event(vgpu, info->event);
+	else
+		set_bit(info->event, vgpu->irq.flip_done_event[info->pipe]);
+
+	return 0;
+}
+
+static int decode_mi_display_flip(struct parser_exec_state *s,
+		struct mi_display_flip_command_info *info)
+{
+	if (IS_BROADWELL(s->engine->i915))
+		return gen8_decode_mi_display_flip(s, info);
+	if (GRAPHICS_VER(s->engine->i915) >= 9)
+		return skl_decode_mi_display_flip(s, info);
+
+	return -ENODEV;
+}
+
+static int check_mi_display_flip(struct parser_exec_state *s,
+		struct mi_display_flip_command_info *info)
+{
+	return gen8_check_mi_display_flip(s, info);
+}
+
+static int update_plane_mmio_from_mi_display_flip(
+		struct parser_exec_state *s,
+		struct mi_display_flip_command_info *info)
+{
+	return gen8_update_plane_mmio_from_mi_display_flip(s, info);
+}
+
+static int cmd_handler_mi_display_flip(struct parser_exec_state *s)
+{
+	struct mi_display_flip_command_info info;
+	struct intel_vgpu *vgpu = s->vgpu;
+	int ret;
+	int i;
+	int len = cmd_length(s);
+	u32 valid_len = CMD_LEN(1);
+
+	/* Flip Type == Stereo 3D Flip */
+	if (DWORD_FIELD(2, 1, 0) == 2)
+		valid_len++;
+	ret = gvt_check_valid_cmd_length(cmd_length(s),
+			valid_len);
+	if (ret)
+		return ret;
+
+	ret = decode_mi_display_flip(s, &info);
+	if (ret) {
+		gvt_vgpu_err("fail to decode MI display flip command\n");
+		return ret;
+	}
+
+	ret = check_mi_display_flip(s, &info);
+	if (ret) {
+		gvt_vgpu_err("invalid MI display flip command\n");
+		return ret;
+	}
+
+	ret = update_plane_mmio_from_mi_display_flip(s, &info);
+	if (ret) {
+		gvt_vgpu_err("fail to update plane mmio\n");
+		return ret;
+	}
+
+	for (i = 0; i < len; i++)
+		patch_value(s, cmd_ptr(s, i), MI_NOOP);
+	return 0;
+}
+
+static bool is_wait_for_flip_pending(u32 cmd)
+{
+	return cmd & (MI_WAIT_FOR_PLANE_A_FLIP_PENDING |
+			MI_WAIT_FOR_PLANE_B_FLIP_PENDING |
+			MI_WAIT_FOR_PLANE_C_FLIP_PENDING |
+			MI_WAIT_FOR_SPRITE_A_FLIP_PENDING |
+			MI_WAIT_FOR_SPRITE_B_FLIP_PENDING |
+			MI_WAIT_FOR_SPRITE_C_FLIP_PENDING);
+}
+
+static int cmd_handler_mi_wait_for_event(struct parser_exec_state *s)
+{
+	u32 cmd = cmd_val(s, 0);
+
+	if (!is_wait_for_flip_pending(cmd))
+		return 0;
+
+	patch_value(s, cmd_ptr(s, 0), MI_NOOP);
+	return 0;
+}
+
+static unsigned long get_gma_bb_from_cmd(struct parser_exec_state *s, int index)
+{
+	unsigned long addr;
+	unsigned long gma_high, gma_low;
+	struct intel_vgpu *vgpu = s->vgpu;
+	int gmadr_bytes = vgpu->gvt->device_info.gmadr_bytes_in_cmd;
+
+	if (WARN_ON(gmadr_bytes != 4 && gmadr_bytes != 8)) {
+		gvt_vgpu_err("invalid gma bytes %d\n", gmadr_bytes);
+		return INTEL_GVT_INVALID_ADDR;
+	}
+
+	gma_low = cmd_val(s, index) & BATCH_BUFFER_ADDR_MASK;
+	if (gmadr_bytes == 4) {
+		addr = gma_low;
+	} else {
+		gma_high = cmd_val(s, index + 1) & BATCH_BUFFER_ADDR_HIGH_MASK;
+		addr = (((unsigned long)gma_high) << 32) | gma_low;
+	}
+	return addr;
+}
+
+static inline int cmd_address_audit(struct parser_exec_state *s,
+		unsigned long guest_gma, int op_size, bool index_mode)
+{
+	struct intel_vgpu *vgpu = s->vgpu;
+	u32 max_surface_size = vgpu->gvt->device_info.max_surface_size;
+	int i;
+	int ret;
+
+	if (op_size > max_surface_size) {
+		gvt_vgpu_err("command address audit fail name %s\n",
+			s->info->name);
+		return -EFAULT;
+	}
+
+	if (index_mode)	{
+		if (guest_gma >= I915_GTT_PAGE_SIZE) {
+			ret = -EFAULT;
+			goto err;
+		}
+	} else if (!intel_gvt_ggtt_validate_range(vgpu, guest_gma, op_size)) {
+		ret = -EFAULT;
+		goto err;
+	}
+
+	return 0;
+
+err:
+	gvt_vgpu_err("cmd_parser: Malicious %s detected, addr=0x%lx, len=%d!\n",
+			s->info->name, guest_gma, op_size);
+
+	pr_err("cmd dump: ");
+	for (i = 0; i < cmd_length(s); i++) {
+		if (!(i % 4))
+			pr_err("\n%08x ", cmd_val(s, i));
+		else
+			pr_err("%08x ", cmd_val(s, i));
+	}
+	pr_err("\nvgpu%d: aperture 0x%llx - 0x%llx, hidden 0x%llx - 0x%llx\n",
+			vgpu->id,
+			vgpu_aperture_gmadr_base(vgpu),
+			vgpu_aperture_gmadr_end(vgpu),
+			vgpu_hidden_gmadr_base(vgpu),
+			vgpu_hidden_gmadr_end(vgpu));
+	return ret;
+}
+
+static int cmd_handler_mi_store_data_imm(struct parser_exec_state *s)
+{
+	int gmadr_bytes = s->vgpu->gvt->device_info.gmadr_bytes_in_cmd;
+	int op_size = (cmd_length(s) - 3) * sizeof(u32);
+	int core_id = (cmd_val(s, 2) & (1 << 0)) ? 1 : 0;
+	unsigned long gma, gma_low, gma_high;
+	u32 valid_len = CMD_LEN(2);
+	int ret = 0;
+
+	/* check ppggt */
+	if (!(cmd_val(s, 0) & (1 << 22)))
+		return 0;
+
+	/* check if QWORD */
+	if (DWORD_FIELD(0, 21, 21))
+		valid_len++;
+	ret = gvt_check_valid_cmd_length(cmd_length(s),
+			valid_len);
+	if (ret)
+		return ret;
+
+	gma = cmd_val(s, 2) & GENMASK(31, 2);
+
+	if (gmadr_bytes == 8) {
+		gma_low = cmd_val(s, 1) & GENMASK(31, 2);
+		gma_high = cmd_val(s, 2) & GENMASK(15, 0);
+		gma = (gma_high << 32) | gma_low;
+		core_id = (cmd_val(s, 1) & (1 << 0)) ? 1 : 0;
+	}
+	ret = cmd_address_audit(s, gma + op_size * core_id, op_size, false);
+	return ret;
+}
+
+static inline int unexpected_cmd(struct parser_exec_state *s)
+{
+	struct intel_vgpu *vgpu = s->vgpu;
+
+	gvt_vgpu_err("Unexpected %s in command buffer!\n", s->info->name);
+
+	return -EBADRQC;
+}
+
+static int cmd_handler_mi_semaphore_wait(struct parser_exec_state *s)
+{
+	return unexpected_cmd(s);
+}
+
+static int cmd_handler_mi_report_perf_count(struct parser_exec_state *s)
+{
+	return unexpected_cmd(s);
+}
+
+static int cmd_handler_mi_op_2e(struct parser_exec_state *s)
+{
+	return unexpected_cmd(s);
+}
+
+static int cmd_handler_mi_op_2f(struct parser_exec_state *s)
+{
+	int gmadr_bytes = s->vgpu->gvt->device_info.gmadr_bytes_in_cmd;
+	int op_size = (1 << ((cmd_val(s, 0) & GENMASK(20, 19)) >> 19)) *
+			sizeof(u32);
+	unsigned long gma, gma_high;
+	u32 valid_len = CMD_LEN(1);
+	int ret = 0;
+
+	if (!(cmd_val(s, 0) & (1 << 22)))
+		return ret;
+
+	/* check inline data */
+	if (cmd_val(s, 0) & BIT(18))
+		valid_len = CMD_LEN(9);
+	ret = gvt_check_valid_cmd_length(cmd_length(s),
+			valid_len);
+	if (ret)
+		return ret;
+
+	gma = cmd_val(s, 1) & GENMASK(31, 2);
+	if (gmadr_bytes == 8) {
+		gma_high = cmd_val(s, 2) & GENMASK(15, 0);
+		gma = (gma_high << 32) | gma;
+	}
+	ret = cmd_address_audit(s, gma, op_size, false);
+	return ret;
+}
+
+static int cmd_handler_mi_store_data_index(struct parser_exec_state *s)
+{
+	return unexpected_cmd(s);
+}
+
+static int cmd_handler_mi_clflush(struct parser_exec_state *s)
+{
+	return unexpected_cmd(s);
+}
+
+static int cmd_handler_mi_conditional_batch_buffer_end(
+		struct parser_exec_state *s)
+{
+	return unexpected_cmd(s);
+}
+
+static int cmd_handler_mi_update_gtt(struct parser_exec_state *s)
+{
+	return unexpected_cmd(s);
+}
+
+static int cmd_handler_mi_flush_dw(struct parser_exec_state *s)
+{
+	int gmadr_bytes = s->vgpu->gvt->device_info.gmadr_bytes_in_cmd;
+	unsigned long gma;
+	bool index_mode = false;
+	int ret = 0;
+	u32 hws_pga, val;
+	u32 valid_len = CMD_LEN(2);
+
+	ret = gvt_check_valid_cmd_length(cmd_length(s),
+			valid_len);
+	if (ret) {
+		/* Check again for Qword */
+		ret = gvt_check_valid_cmd_length(cmd_length(s),
+			++valid_len);
+		return ret;
+	}
+
+	/* Check post-sync and ppgtt bit */
+	if (((cmd_val(s, 0) >> 14) & 0x3) && (cmd_val(s, 1) & (1 << 2))) {
+		gma = cmd_val(s, 1) & GENMASK(31, 3);
+		if (gmadr_bytes == 8)
+			gma |= (cmd_val(s, 2) & GENMASK(15, 0)) << 32;
+		/* Store Data Index */
+		if (cmd_val(s, 0) & (1 << 21))
+			index_mode = true;
+		ret = cmd_address_audit(s, gma, sizeof(u64), index_mode);
+		if (ret)
+			return ret;
+		if (index_mode) {
+			hws_pga = s->vgpu->hws_pga[s->engine->id];
+			gma = hws_pga + gma;
+			patch_value(s, cmd_ptr(s, 1), gma);
+			val = cmd_val(s, 0) & (~(1 << 21));
+			patch_value(s, cmd_ptr(s, 0), val);
+		}
+	}
+	/* Check notify bit */
+	if ((cmd_val(s, 0) & (1 << 8)))
+		set_bit(cmd_interrupt_events[s->engine->id].mi_flush_dw,
+			s->workload->pending_events);
+	return ret;
+}
+
+static void addr_type_update_snb(struct parser_exec_state *s)
+{
+	if ((s->buf_type == RING_BUFFER_INSTRUCTION) &&
+			(BATCH_BUFFER_ADR_SPACE_BIT(cmd_val(s, 0)) == 1)) {
+		s->buf_addr_type = PPGTT_BUFFER;
+	}
+}
+
+
+static int copy_gma_to_hva(struct intel_vgpu *vgpu, struct intel_vgpu_mm *mm,
+		unsigned long gma, unsigned long end_gma, void *va)
+{
+	unsigned long copy_len, offset;
+	unsigned long len = 0;
+	unsigned long gpa;
+
+	while (gma != end_gma) {
+		gpa = intel_vgpu_gma_to_gpa(mm, gma);
+		if (gpa == INTEL_GVT_INVALID_ADDR) {
+			gvt_vgpu_err("invalid gma address: %lx\n", gma);
+			return -EFAULT;
+		}
+
+		offset = gma & (I915_GTT_PAGE_SIZE - 1);
+
+		copy_len = (end_gma - gma) >= (I915_GTT_PAGE_SIZE - offset) ?
+			I915_GTT_PAGE_SIZE - offset : end_gma - gma;
+
+		intel_gvt_read_gpa(vgpu, gpa, va + len, copy_len);
+
+		len += copy_len;
+		gma += copy_len;
+	}
+	return len;
+}
+
+
+/*
+ * Check whether a batch buffer needs to be scanned. Currently
+ * the only criteria is based on privilege.
+ */
+static int batch_buffer_needs_scan(struct parser_exec_state *s)
+{
+	/* Decide privilege based on address space */
+	if (cmd_val(s, 0) & BIT(8) &&
+	    !(s->vgpu->scan_nonprivbb & s->engine->mask))
+		return 0;
+
+	return 1;
+}
+
+static const char *repr_addr_type(unsigned int type)
+{
+	return type == PPGTT_BUFFER ? "ppgtt" : "ggtt";
+}
+
+static int find_bb_size(struct parser_exec_state *s,
+			unsigned long *bb_size,
+			unsigned long *bb_end_cmd_offset)
+{
+	unsigned long gma = 0;
+	const struct cmd_info *info;
+	u32 cmd_len = 0;
+	bool bb_end = false;
+	struct intel_vgpu *vgpu = s->vgpu;
+	u32 cmd;
+	struct intel_vgpu_mm *mm = (s->buf_addr_type == GTT_BUFFER) ?
+		s->vgpu->gtt.ggtt_mm : s->workload->shadow_mm;
+
+	*bb_size = 0;
+	*bb_end_cmd_offset = 0;
+
+	/* get the start gm address of the batch buffer */
+	gma = get_gma_bb_from_cmd(s, 1);
+	if (gma == INTEL_GVT_INVALID_ADDR)
+		return -EFAULT;
+
+	cmd = cmd_val(s, 0);
+	info = get_cmd_info(s->vgpu->gvt, cmd, s->engine);
+	if (info == NULL) {
+		gvt_vgpu_err("unknown cmd 0x%x, opcode=0x%x, addr_type=%s, ring %s, workload=%p\n",
+			     cmd, get_opcode(cmd, s->engine),
+			     repr_addr_type(s->buf_addr_type),
+			     s->engine->name, s->workload);
+		return -EBADRQC;
+	}
+	do {
+		if (copy_gma_to_hva(s->vgpu, mm,
+				    gma, gma + 4, &cmd) < 0)
+			return -EFAULT;
+		info = get_cmd_info(s->vgpu->gvt, cmd, s->engine);
+		if (info == NULL) {
+			gvt_vgpu_err("unknown cmd 0x%x, opcode=0x%x, addr_type=%s, ring %s, workload=%p\n",
+				     cmd, get_opcode(cmd, s->engine),
+				     repr_addr_type(s->buf_addr_type),
+				     s->engine->name, s->workload);
+			return -EBADRQC;
+		}
+
+		if (info->opcode == OP_MI_BATCH_BUFFER_END) {
+			bb_end = true;
+		} else if (info->opcode == OP_MI_BATCH_BUFFER_START) {
+			if (BATCH_BUFFER_2ND_LEVEL_BIT(cmd) == 0)
+				/* chained batch buffer */
+				bb_end = true;
+		}
+
+		if (bb_end)
+			*bb_end_cmd_offset = *bb_size;
+
+		cmd_len = get_cmd_length(info, cmd) << 2;
+		*bb_size += cmd_len;
+		gma += cmd_len;
+	} while (!bb_end);
+
+	return 0;
+}
+
+static int audit_bb_end(struct parser_exec_state *s, void *va)
+{
+	struct intel_vgpu *vgpu = s->vgpu;
+	u32 cmd = *(u32 *)va;
+	const struct cmd_info *info;
+
+	info = get_cmd_info(s->vgpu->gvt, cmd, s->engine);
+	if (info == NULL) {
+		gvt_vgpu_err("unknown cmd 0x%x, opcode=0x%x, addr_type=%s, ring %s, workload=%p\n",
+			     cmd, get_opcode(cmd, s->engine),
+			     repr_addr_type(s->buf_addr_type),
+			     s->engine->name, s->workload);
+		return -EBADRQC;
+	}
+
+	if ((info->opcode == OP_MI_BATCH_BUFFER_END) ||
+	    ((info->opcode == OP_MI_BATCH_BUFFER_START) &&
+	     (BATCH_BUFFER_2ND_LEVEL_BIT(cmd) == 0)))
+		return 0;
+
+	return -EBADRQC;
+}
+
+static int perform_bb_shadow(struct parser_exec_state *s)
+{
+	struct intel_vgpu *vgpu = s->vgpu;
+	struct intel_vgpu_shadow_bb *bb;
+	unsigned long gma = 0;
+	unsigned long bb_size;
+	unsigned long bb_end_cmd_offset;
+	int ret = 0;
+	struct intel_vgpu_mm *mm = (s->buf_addr_type == GTT_BUFFER) ?
+		s->vgpu->gtt.ggtt_mm : s->workload->shadow_mm;
+	unsigned long start_offset = 0;
+
+	/* get the start gm address of the batch buffer */
+	gma = get_gma_bb_from_cmd(s, 1);
+	if (gma == INTEL_GVT_INVALID_ADDR)
+		return -EFAULT;
+
+	ret = find_bb_size(s, &bb_size, &bb_end_cmd_offset);
+	if (ret)
+		return ret;
+
+	bb = kzalloc(sizeof(*bb), GFP_KERNEL);
+	if (!bb)
+		return -ENOMEM;
+
+	bb->ppgtt = (s->buf_addr_type == GTT_BUFFER) ? false : true;
+
+	/* the start_offset stores the batch buffer's start gma's
+	 * offset relative to page boundary. so for non-privileged batch
+	 * buffer, the shadowed gem object holds exactly the same page
+	 * layout as original gem object. This is for the convience of
+	 * replacing the whole non-privilged batch buffer page to this
+	 * shadowed one in PPGTT at the same gma address. (this replacing
+	 * action is not implemented yet now, but may be necessary in
+	 * future).
+	 * for prileged batch buffer, we just change start gma address to
+	 * that of shadowed page.
+	 */
+	if (bb->ppgtt)
+		start_offset = gma & ~I915_GTT_PAGE_MASK;
+
+	bb->obj = i915_gem_object_create_shmem(s->engine->i915,
+					       round_up(bb_size + start_offset,
+							PAGE_SIZE));
+	if (IS_ERR(bb->obj)) {
+		ret = PTR_ERR(bb->obj);
+		goto err_free_bb;
+	}
+
+	bb->va = i915_gem_object_pin_map(bb->obj, I915_MAP_WB);
+	if (IS_ERR(bb->va)) {
+		ret = PTR_ERR(bb->va);
+		goto err_free_obj;
+	}
+
+	ret = copy_gma_to_hva(s->vgpu, mm,
+			      gma, gma + bb_size,
+			      bb->va + start_offset);
+	if (ret < 0) {
+		gvt_vgpu_err("fail to copy guest ring buffer\n");
+		ret = -EFAULT;
+		goto err_unmap;
+	}
+
+	ret = audit_bb_end(s, bb->va + start_offset + bb_end_cmd_offset);
+	if (ret)
+		goto err_unmap;
+
+	i915_gem_object_unlock(bb->obj);
+	INIT_LIST_HEAD(&bb->list);
+	list_add(&bb->list, &s->workload->shadow_bb);
+
+	bb->bb_start_cmd_va = s->ip_va;
+
+	if ((s->buf_type == BATCH_BUFFER_INSTRUCTION) && (!s->is_ctx_wa))
+		bb->bb_offset = s->ip_va - s->rb_va;
+	else
+		bb->bb_offset = 0;
+
+	/*
+	 * ip_va saves the virtual address of the shadow batch buffer, while
+	 * ip_gma saves the graphics address of the original batch buffer.
+	 * As the shadow batch buffer is just a copy from the originial one,
+	 * it should be right to use shadow batch buffer'va and original batch
+	 * buffer's gma in pair. After all, we don't want to pin the shadow
+	 * buffer here (too early).
+	 */
+	s->ip_va = bb->va + start_offset;
+	s->ip_gma = gma;
+	return 0;
+err_unmap:
+	i915_gem_object_unpin_map(bb->obj);
+err_free_obj:
+	i915_gem_object_put(bb->obj);
+err_free_bb:
+	kfree(bb);
+	return ret;
+}
+
+static int cmd_handler_mi_batch_buffer_start(struct parser_exec_state *s)
+{
+	bool second_level;
+	int ret = 0;
+	struct intel_vgpu *vgpu = s->vgpu;
+
+	if (s->buf_type == BATCH_BUFFER_2ND_LEVEL) {
+		gvt_vgpu_err("Found MI_BATCH_BUFFER_START in 2nd level BB\n");
+		return -EFAULT;
+	}
+
+	second_level = BATCH_BUFFER_2ND_LEVEL_BIT(cmd_val(s, 0)) == 1;
+	if (second_level && (s->buf_type != BATCH_BUFFER_INSTRUCTION)) {
+		gvt_vgpu_err("Jumping to 2nd level BB from RB is not allowed\n");
+		return -EFAULT;
+	}
+
+	s->saved_buf_addr_type = s->buf_addr_type;
+	addr_type_update_snb(s);
+	if (s->buf_type == RING_BUFFER_INSTRUCTION) {
+		s->ret_ip_gma_ring = s->ip_gma + cmd_length(s) * sizeof(u32);
+		s->buf_type = BATCH_BUFFER_INSTRUCTION;
+	} else if (second_level) {
+		s->buf_type = BATCH_BUFFER_2ND_LEVEL;
+		s->ret_ip_gma_bb = s->ip_gma + cmd_length(s) * sizeof(u32);
+		s->ret_bb_va = s->ip_va + cmd_length(s) * sizeof(u32);
+	}
+
+	if (batch_buffer_needs_scan(s)) {
+		ret = perform_bb_shadow(s);
+		if (ret < 0)
+			gvt_vgpu_err("invalid shadow batch buffer\n");
+	} else {
+		/* emulate a batch buffer end to do return right */
+		ret = cmd_handler_mi_batch_buffer_end(s);
+		if (ret < 0)
+			return ret;
+	}
+	return ret;
+}
+
+static int mi_noop_index;
+
+static const struct cmd_info cmd_info[] = {
+	{"MI_NOOP", OP_MI_NOOP, F_LEN_CONST, R_ALL, D_ALL, 0, 1, NULL},
+
+	{"MI_SET_PREDICATE", OP_MI_SET_PREDICATE, F_LEN_CONST, R_ALL, D_ALL,
+		0, 1, NULL},
+
+	{"MI_USER_INTERRUPT", OP_MI_USER_INTERRUPT, F_LEN_CONST, R_ALL, D_ALL,
+		0, 1, cmd_handler_mi_user_interrupt},
+
+	{"MI_WAIT_FOR_EVENT", OP_MI_WAIT_FOR_EVENT, F_LEN_CONST, R_RCS | R_BCS,
+		D_ALL, 0, 1, cmd_handler_mi_wait_for_event},
+
+	{"MI_FLUSH", OP_MI_FLUSH, F_LEN_CONST, R_ALL, D_ALL, 0, 1, NULL},
+
+	{"MI_ARB_CHECK", OP_MI_ARB_CHECK, F_LEN_CONST, R_ALL, D_ALL, 0, 1,
+		NULL},
+
+	{"MI_RS_CONTROL", OP_MI_RS_CONTROL, F_LEN_CONST, R_RCS, D_ALL, 0, 1,
+		NULL},
+
+	{"MI_REPORT_HEAD", OP_MI_REPORT_HEAD, F_LEN_CONST, R_ALL, D_ALL, 0, 1,
+		NULL},
+
+	{"MI_ARB_ON_OFF", OP_MI_ARB_ON_OFF, F_LEN_CONST, R_ALL, D_ALL, 0, 1,
+		NULL},
+
+	{"MI_URB_ATOMIC_ALLOC", OP_MI_URB_ATOMIC_ALLOC, F_LEN_CONST, R_RCS,
+		D_ALL, 0, 1, NULL},
+
+	{"MI_BATCH_BUFFER_END", OP_MI_BATCH_BUFFER_END,
+		F_IP_ADVANCE_CUSTOM | F_LEN_CONST, R_ALL, D_ALL, 0, 1,
+		cmd_handler_mi_batch_buffer_end},
+
+	{"MI_SUSPEND_FLUSH", OP_MI_SUSPEND_FLUSH, F_LEN_CONST, R_ALL, D_ALL,
+		0, 1, NULL},
+
+	{"MI_PREDICATE", OP_MI_PREDICATE, F_LEN_CONST, R_RCS, D_ALL, 0, 1,
+		NULL},
+
+	{"MI_TOPOLOGY_FILTER", OP_MI_TOPOLOGY_FILTER, F_LEN_CONST, R_ALL,
+		D_ALL, 0, 1, NULL},
+
+	{"MI_SET_APPID", OP_MI_SET_APPID, F_LEN_CONST, R_ALL, D_ALL, 0, 1,
+		NULL},
+
+	{"MI_RS_CONTEXT", OP_MI_RS_CONTEXT, F_LEN_CONST, R_RCS, D_ALL, 0, 1,
+		NULL},
+
+	{"MI_DISPLAY_FLIP", OP_MI_DISPLAY_FLIP, F_LEN_VAR,
+		R_RCS | R_BCS, D_ALL, 0, 8, cmd_handler_mi_display_flip},
+
+	{"MI_SEMAPHORE_MBOX", OP_MI_SEMAPHORE_MBOX, F_LEN_VAR | F_LEN_VAR_FIXED,
+		R_ALL, D_ALL, 0, 8, NULL, CMD_LEN(1)},
+
+	{"MI_MATH", OP_MI_MATH, F_LEN_VAR, R_ALL, D_ALL, 0, 8, NULL},
+
+	{"MI_URB_CLEAR", OP_MI_URB_CLEAR, F_LEN_VAR | F_LEN_VAR_FIXED, R_RCS,
+		D_ALL, 0, 8, NULL, CMD_LEN(0)},
+
+	{"MI_SEMAPHORE_SIGNAL", OP_MI_SEMAPHORE_SIGNAL,
+		F_LEN_VAR | F_LEN_VAR_FIXED, R_ALL, D_BDW_PLUS, 0, 8,
+		NULL, CMD_LEN(0)},
+
+	{"MI_SEMAPHORE_WAIT", OP_MI_SEMAPHORE_WAIT,
+		F_LEN_VAR | F_LEN_VAR_FIXED, R_ALL, D_BDW_PLUS, ADDR_FIX_1(2),
+		8, cmd_handler_mi_semaphore_wait, CMD_LEN(2)},
+
+	{"MI_STORE_DATA_IMM", OP_MI_STORE_DATA_IMM, F_LEN_VAR, R_ALL, D_BDW_PLUS,
+		ADDR_FIX_1(1), 10, cmd_handler_mi_store_data_imm},
+
+	{"MI_STORE_DATA_INDEX", OP_MI_STORE_DATA_INDEX, F_LEN_VAR, R_ALL, D_ALL,
+		0, 8, cmd_handler_mi_store_data_index},
+
+	{"MI_LOAD_REGISTER_IMM", OP_MI_LOAD_REGISTER_IMM, F_LEN_VAR, R_ALL,
+		D_ALL, 0, 8, cmd_handler_lri},
+
+	{"MI_UPDATE_GTT", OP_MI_UPDATE_GTT, F_LEN_VAR, R_ALL, D_BDW_PLUS, 0, 10,
+		cmd_handler_mi_update_gtt},
+
+	{"MI_STORE_REGISTER_MEM", OP_MI_STORE_REGISTER_MEM,
+		F_LEN_VAR | F_LEN_VAR_FIXED, R_ALL, D_ALL, ADDR_FIX_1(2), 8,
+		cmd_handler_srm, CMD_LEN(2)},
+
+	{"MI_FLUSH_DW", OP_MI_FLUSH_DW, F_LEN_VAR, R_ALL, D_ALL, 0, 6,
+		cmd_handler_mi_flush_dw},
+
+	{"MI_CLFLUSH", OP_MI_CLFLUSH, F_LEN_VAR, R_ALL, D_ALL, ADDR_FIX_1(1),
+		10, cmd_handler_mi_clflush},
+
+	{"MI_REPORT_PERF_COUNT", OP_MI_REPORT_PERF_COUNT,
+		F_LEN_VAR | F_LEN_VAR_FIXED, R_ALL, D_ALL, ADDR_FIX_1(1), 6,
+		cmd_handler_mi_report_perf_count, CMD_LEN(2)},
+
+	{"MI_LOAD_REGISTER_MEM", OP_MI_LOAD_REGISTER_MEM,
+		F_LEN_VAR | F_LEN_VAR_FIXED, R_ALL, D_ALL, ADDR_FIX_1(2), 8,
+		cmd_handler_lrm, CMD_LEN(2)},
+
+	{"MI_LOAD_REGISTER_REG", OP_MI_LOAD_REGISTER_REG,
+		F_LEN_VAR | F_LEN_VAR_FIXED, R_ALL, D_ALL, 0, 8,
+		cmd_handler_lrr, CMD_LEN(1)},
+
+	{"MI_RS_STORE_DATA_IMM", OP_MI_RS_STORE_DATA_IMM,
+		F_LEN_VAR | F_LEN_VAR_FIXED, R_RCS, D_ALL, 0,
+		8, NULL, CMD_LEN(2)},
+
+	{"MI_LOAD_URB_MEM", OP_MI_LOAD_URB_MEM, F_LEN_VAR | F_LEN_VAR_FIXED,
+		R_RCS, D_ALL, ADDR_FIX_1(2), 8, NULL, CMD_LEN(2)},
+
+	{"MI_STORE_URM_MEM", OP_MI_STORE_URM_MEM, F_LEN_VAR, R_RCS, D_ALL,
+		ADDR_FIX_1(2), 8, NULL},
+
+	{"MI_OP_2E", OP_MI_2E, F_LEN_VAR | F_LEN_VAR_FIXED, R_ALL, D_BDW_PLUS,
+		ADDR_FIX_2(1, 2), 8, cmd_handler_mi_op_2e, CMD_LEN(3)},
+
+	{"MI_OP_2F", OP_MI_2F, F_LEN_VAR, R_ALL, D_BDW_PLUS, ADDR_FIX_1(1),
+		8, cmd_handler_mi_op_2f},
+
+	{"MI_BATCH_BUFFER_START", OP_MI_BATCH_BUFFER_START,
+		F_IP_ADVANCE_CUSTOM, R_ALL, D_ALL, 0, 8,
+		cmd_handler_mi_batch_buffer_start},
+
+	{"MI_CONDITIONAL_BATCH_BUFFER_END", OP_MI_CONDITIONAL_BATCH_BUFFER_END,
+		F_LEN_VAR | F_LEN_VAR_FIXED, R_ALL, D_ALL, ADDR_FIX_1(2), 8,
+		cmd_handler_mi_conditional_batch_buffer_end, CMD_LEN(2)},
+
+	{"MI_LOAD_SCAN_LINES_INCL", OP_MI_LOAD_SCAN_LINES_INCL, F_LEN_CONST,
+		R_RCS | R_BCS, D_ALL, 0, 2, NULL},
+
+	{"XY_SETUP_BLT", OP_XY_SETUP_BLT, F_LEN_VAR, R_BCS, D_ALL,
+		ADDR_FIX_2(4, 7), 8, NULL},
+
+	{"XY_SETUP_CLIP_BLT", OP_XY_SETUP_CLIP_BLT, F_LEN_VAR, R_BCS, D_ALL,
+		0, 8, NULL},
+
+	{"XY_SETUP_MONO_PATTERN_SL_BLT", OP_XY_SETUP_MONO_PATTERN_SL_BLT,
+		F_LEN_VAR, R_BCS, D_ALL, ADDR_FIX_1(4), 8, NULL},
+
+	{"XY_PIXEL_BLT", OP_XY_PIXEL_BLT, F_LEN_VAR, R_BCS, D_ALL, 0, 8, NULL},
+
+	{"XY_SCANLINES_BLT", OP_XY_SCANLINES_BLT, F_LEN_VAR, R_BCS, D_ALL,
+		0, 8, NULL},
+
+	{"XY_TEXT_BLT", OP_XY_TEXT_BLT, F_LEN_VAR, R_BCS, D_ALL,
+		ADDR_FIX_1(3), 8, NULL},
+
+	{"XY_TEXT_IMMEDIATE_BLT", OP_XY_TEXT_IMMEDIATE_BLT, F_LEN_VAR, R_BCS,
+		D_ALL, 0, 8, NULL},
+
+	{"XY_COLOR_BLT", OP_XY_COLOR_BLT, F_LEN_VAR, R_BCS, D_ALL,
+		ADDR_FIX_1(4), 8, NULL},
+
+	{"XY_PAT_BLT", OP_XY_PAT_BLT, F_LEN_VAR, R_BCS, D_ALL,
+		ADDR_FIX_2(4, 5), 8, NULL},
+
+	{"XY_MONO_PAT_BLT", OP_XY_MONO_PAT_BLT, F_LEN_VAR, R_BCS, D_ALL,
+		ADDR_FIX_1(4), 8, NULL},
+
+	{"XY_SRC_COPY_BLT", OP_XY_SRC_COPY_BLT, F_LEN_VAR, R_BCS, D_ALL,
+		ADDR_FIX_2(4, 7), 8, NULL},
+
+	{"XY_MONO_SRC_COPY_BLT", OP_XY_MONO_SRC_COPY_BLT, F_LEN_VAR, R_BCS,
+		D_ALL, ADDR_FIX_2(4, 5), 8, NULL},
+
+	{"XY_FULL_BLT", OP_XY_FULL_BLT, F_LEN_VAR, R_BCS, D_ALL, 0, 8, NULL},
+
+	{"XY_FULL_MONO_SRC_BLT", OP_XY_FULL_MONO_SRC_BLT, F_LEN_VAR, R_BCS,
+		D_ALL, ADDR_FIX_3(4, 5, 8), 8, NULL},
+
+	{"XY_FULL_MONO_PATTERN_BLT", OP_XY_FULL_MONO_PATTERN_BLT, F_LEN_VAR,
+		R_BCS, D_ALL, ADDR_FIX_2(4, 7), 8, NULL},
+
+	{"XY_FULL_MONO_PATTERN_MONO_SRC_BLT",
+		OP_XY_FULL_MONO_PATTERN_MONO_SRC_BLT,
+		F_LEN_VAR, R_BCS, D_ALL, ADDR_FIX_2(4, 5), 8, NULL},
+
+	{"XY_MONO_PAT_FIXED_BLT", OP_XY_MONO_PAT_FIXED_BLT, F_LEN_VAR, R_BCS,
+		D_ALL, ADDR_FIX_1(4), 8, NULL},
+
+	{"XY_MONO_SRC_COPY_IMMEDIATE_BLT", OP_XY_MONO_SRC_COPY_IMMEDIATE_BLT,
+		F_LEN_VAR, R_BCS, D_ALL, ADDR_FIX_1(4), 8, NULL},
+
+	{"XY_PAT_BLT_IMMEDIATE", OP_XY_PAT_BLT_IMMEDIATE, F_LEN_VAR, R_BCS,
+		D_ALL, ADDR_FIX_1(4), 8, NULL},
+
+	{"XY_SRC_COPY_CHROMA_BLT", OP_XY_SRC_COPY_CHROMA_BLT, F_LEN_VAR, R_BCS,
+		D_ALL, ADDR_FIX_2(4, 7), 8, NULL},
+
+	{"XY_FULL_IMMEDIATE_PATTERN_BLT", OP_XY_FULL_IMMEDIATE_PATTERN_BLT,
+		F_LEN_VAR, R_BCS, D_ALL, ADDR_FIX_2(4, 7), 8, NULL},
+
+	{"XY_FULL_MONO_SRC_IMMEDIATE_PATTERN_BLT",
+		OP_XY_FULL_MONO_SRC_IMMEDIATE_PATTERN_BLT,
+		F_LEN_VAR, R_BCS, D_ALL, ADDR_FIX_2(4, 5), 8, NULL},
+
+	{"XY_PAT_CHROMA_BLT", OP_XY_PAT_CHROMA_BLT, F_LEN_VAR, R_BCS, D_ALL,
+		ADDR_FIX_2(4, 5), 8, NULL},
+
+	{"XY_PAT_CHROMA_BLT_IMMEDIATE", OP_XY_PAT_CHROMA_BLT_IMMEDIATE,
+		F_LEN_VAR, R_BCS, D_ALL, ADDR_FIX_1(4), 8, NULL},
+
+	{"3DSTATE_VIEWPORT_STATE_POINTERS_SF_CLIP",
+		OP_3DSTATE_VIEWPORT_STATE_POINTERS_SF_CLIP,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_VIEWPORT_STATE_POINTERS_CC",
+		OP_3DSTATE_VIEWPORT_STATE_POINTERS_CC,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_BLEND_STATE_POINTERS",
+		OP_3DSTATE_BLEND_STATE_POINTERS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DEPTH_STENCIL_STATE_POINTERS",
+		OP_3DSTATE_DEPTH_STENCIL_STATE_POINTERS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_BINDING_TABLE_POINTERS_VS",
+		OP_3DSTATE_BINDING_TABLE_POINTERS_VS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_BINDING_TABLE_POINTERS_HS",
+		OP_3DSTATE_BINDING_TABLE_POINTERS_HS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_BINDING_TABLE_POINTERS_DS",
+		OP_3DSTATE_BINDING_TABLE_POINTERS_DS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_BINDING_TABLE_POINTERS_GS",
+		OP_3DSTATE_BINDING_TABLE_POINTERS_GS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_BINDING_TABLE_POINTERS_PS",
+		OP_3DSTATE_BINDING_TABLE_POINTERS_PS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SAMPLER_STATE_POINTERS_VS",
+		OP_3DSTATE_SAMPLER_STATE_POINTERS_VS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SAMPLER_STATE_POINTERS_HS",
+		OP_3DSTATE_SAMPLER_STATE_POINTERS_HS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SAMPLER_STATE_POINTERS_DS",
+		OP_3DSTATE_SAMPLER_STATE_POINTERS_DS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SAMPLER_STATE_POINTERS_GS",
+		OP_3DSTATE_SAMPLER_STATE_POINTERS_GS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SAMPLER_STATE_POINTERS_PS",
+		OP_3DSTATE_SAMPLER_STATE_POINTERS_PS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_URB_VS", OP_3DSTATE_URB_VS, F_LEN_VAR, R_RCS, D_ALL,
+		0, 8, NULL},
+
+	{"3DSTATE_URB_HS", OP_3DSTATE_URB_HS, F_LEN_VAR, R_RCS, D_ALL,
+		0, 8, NULL},
+
+	{"3DSTATE_URB_DS", OP_3DSTATE_URB_DS, F_LEN_VAR, R_RCS, D_ALL,
+		0, 8, NULL},
+
+	{"3DSTATE_URB_GS", OP_3DSTATE_URB_GS, F_LEN_VAR, R_RCS, D_ALL,
+		0, 8, NULL},
+
+	{"3DSTATE_GATHER_CONSTANT_VS", OP_3DSTATE_GATHER_CONSTANT_VS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_GATHER_CONSTANT_GS", OP_3DSTATE_GATHER_CONSTANT_GS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_GATHER_CONSTANT_HS", OP_3DSTATE_GATHER_CONSTANT_HS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_GATHER_CONSTANT_DS", OP_3DSTATE_GATHER_CONSTANT_DS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_GATHER_CONSTANT_PS", OP_3DSTATE_GATHER_CONSTANT_PS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DX9_CONSTANTF_VS", OP_3DSTATE_DX9_CONSTANTF_VS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 11, NULL},
+
+	{"3DSTATE_DX9_CONSTANTF_PS", OP_3DSTATE_DX9_CONSTANTF_PS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 11, NULL},
+
+	{"3DSTATE_DX9_CONSTANTI_VS", OP_3DSTATE_DX9_CONSTANTI_VS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DX9_CONSTANTI_PS", OP_3DSTATE_DX9_CONSTANTI_PS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DX9_CONSTANTB_VS", OP_3DSTATE_DX9_CONSTANTB_VS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DX9_CONSTANTB_PS", OP_3DSTATE_DX9_CONSTANTB_PS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DX9_LOCAL_VALID_VS", OP_3DSTATE_DX9_LOCAL_VALID_VS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DX9_LOCAL_VALID_PS", OP_3DSTATE_DX9_LOCAL_VALID_PS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DX9_GENERATE_ACTIVE_VS", OP_3DSTATE_DX9_GENERATE_ACTIVE_VS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DX9_GENERATE_ACTIVE_PS", OP_3DSTATE_DX9_GENERATE_ACTIVE_PS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_BINDING_TABLE_EDIT_VS", OP_3DSTATE_BINDING_TABLE_EDIT_VS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 9, NULL},
+
+	{"3DSTATE_BINDING_TABLE_EDIT_GS", OP_3DSTATE_BINDING_TABLE_EDIT_GS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 9, NULL},
+
+	{"3DSTATE_BINDING_TABLE_EDIT_HS", OP_3DSTATE_BINDING_TABLE_EDIT_HS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 9, NULL},
+
+	{"3DSTATE_BINDING_TABLE_EDIT_DS", OP_3DSTATE_BINDING_TABLE_EDIT_DS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 9, NULL},
+
+	{"3DSTATE_BINDING_TABLE_EDIT_PS", OP_3DSTATE_BINDING_TABLE_EDIT_PS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 9, NULL},
+
+	{"3DSTATE_VF_INSTANCING", OP_3DSTATE_VF_INSTANCING, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, 0, 8, NULL},
+
+	{"3DSTATE_VF_SGVS", OP_3DSTATE_VF_SGVS, F_LEN_VAR, R_RCS, D_BDW_PLUS, 0, 8,
+		NULL},
+
+	{"3DSTATE_VF_TOPOLOGY", OP_3DSTATE_VF_TOPOLOGY, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, 0, 8, NULL},
+
+	{"3DSTATE_WM_CHROMAKEY", OP_3DSTATE_WM_CHROMAKEY, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, 0, 8, NULL},
+
+	{"3DSTATE_PS_BLEND", OP_3DSTATE_PS_BLEND, F_LEN_VAR, R_RCS, D_BDW_PLUS, 0,
+		8, NULL},
+
+	{"3DSTATE_WM_DEPTH_STENCIL", OP_3DSTATE_WM_DEPTH_STENCIL, F_LEN_VAR,
+		R_RCS, D_BDW_PLUS, 0, 8, NULL},
+
+	{"3DSTATE_PS_EXTRA", OP_3DSTATE_PS_EXTRA, F_LEN_VAR, R_RCS, D_BDW_PLUS, 0,
+		8, NULL},
+
+	{"3DSTATE_RASTER", OP_3DSTATE_RASTER, F_LEN_VAR, R_RCS, D_BDW_PLUS, 0, 8,
+		NULL},
+
+	{"3DSTATE_SBE_SWIZ", OP_3DSTATE_SBE_SWIZ, F_LEN_VAR, R_RCS, D_BDW_PLUS, 0, 8,
+		NULL},
+
+	{"3DSTATE_WM_HZ_OP", OP_3DSTATE_WM_HZ_OP, F_LEN_VAR, R_RCS, D_BDW_PLUS, 0, 8,
+		NULL},
+
+	{"3DSTATE_VERTEX_BUFFERS", OP_3DSTATE_VERTEX_BUFFERS, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, 0, 8, NULL},
+
+	{"3DSTATE_VERTEX_ELEMENTS", OP_3DSTATE_VERTEX_ELEMENTS, F_LEN_VAR,
+		R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_INDEX_BUFFER", OP_3DSTATE_INDEX_BUFFER, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, ADDR_FIX_1(2), 8, NULL},
+
+	{"3DSTATE_VF_STATISTICS", OP_3DSTATE_VF_STATISTICS, F_LEN_CONST,
+		R_RCS, D_ALL, 0, 1, NULL},
+
+	{"3DSTATE_VF", OP_3DSTATE_VF, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_CC_STATE_POINTERS", OP_3DSTATE_CC_STATE_POINTERS, F_LEN_VAR,
+		R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SCISSOR_STATE_POINTERS", OP_3DSTATE_SCISSOR_STATE_POINTERS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_GS", OP_3DSTATE_GS, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_CLIP", OP_3DSTATE_CLIP, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_WM", OP_3DSTATE_WM, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_CONSTANT_GS", OP_3DSTATE_CONSTANT_GS, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, 0, 8, NULL},
+
+	{"3DSTATE_CONSTANT_PS", OP_3DSTATE_CONSTANT_PS, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, 0, 8, NULL},
+
+	{"3DSTATE_SAMPLE_MASK", OP_3DSTATE_SAMPLE_MASK, F_LEN_VAR, R_RCS,
+		D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_CONSTANT_HS", OP_3DSTATE_CONSTANT_HS, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, 0, 8, NULL},
+
+	{"3DSTATE_CONSTANT_DS", OP_3DSTATE_CONSTANT_DS, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, 0, 8, NULL},
+
+	{"3DSTATE_HS", OP_3DSTATE_HS, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_TE", OP_3DSTATE_TE, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DS", OP_3DSTATE_DS, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_STREAMOUT", OP_3DSTATE_STREAMOUT, F_LEN_VAR, R_RCS,
+		D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SBE", OP_3DSTATE_SBE, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_PS", OP_3DSTATE_PS, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_DRAWING_RECTANGLE", OP_3DSTATE_DRAWING_RECTANGLE, F_LEN_VAR,
+		R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SAMPLER_PALETTE_LOAD0", OP_3DSTATE_SAMPLER_PALETTE_LOAD0,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_CHROMA_KEY", OP_3DSTATE_CHROMA_KEY, F_LEN_VAR, R_RCS, D_ALL,
+		0, 8, NULL},
+
+	{"3DSTATE_DEPTH_BUFFER", OP_3DSTATE_DEPTH_BUFFER, F_LEN_VAR, R_RCS,
+		D_ALL, ADDR_FIX_1(2), 8, NULL},
+
+	{"3DSTATE_POLY_STIPPLE_OFFSET", OP_3DSTATE_POLY_STIPPLE_OFFSET,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_POLY_STIPPLE_PATTERN", OP_3DSTATE_POLY_STIPPLE_PATTERN,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_LINE_STIPPLE", OP_3DSTATE_LINE_STIPPLE, F_LEN_VAR, R_RCS,
+		D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_AA_LINE_PARAMS", OP_3DSTATE_AA_LINE_PARAMS, F_LEN_VAR, R_RCS,
+		D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_GS_SVB_INDEX", OP_3DSTATE_GS_SVB_INDEX, F_LEN_VAR, R_RCS,
+		D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SAMPLER_PALETTE_LOAD1", OP_3DSTATE_SAMPLER_PALETTE_LOAD1,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_MULTISAMPLE", OP_3DSTATE_MULTISAMPLE_BDW, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, 0, 8, NULL},
+
+	{"3DSTATE_STENCIL_BUFFER", OP_3DSTATE_STENCIL_BUFFER, F_LEN_VAR, R_RCS,
+		D_ALL, ADDR_FIX_1(2), 8, NULL},
+
+	{"3DSTATE_HIER_DEPTH_BUFFER", OP_3DSTATE_HIER_DEPTH_BUFFER, F_LEN_VAR,
+		R_RCS, D_ALL, ADDR_FIX_1(2), 8, NULL},
+
+	{"3DSTATE_CLEAR_PARAMS", OP_3DSTATE_CLEAR_PARAMS, F_LEN_VAR,
+		R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_PUSH_CONSTANT_ALLOC_VS", OP_3DSTATE_PUSH_CONSTANT_ALLOC_VS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_PUSH_CONSTANT_ALLOC_HS", OP_3DSTATE_PUSH_CONSTANT_ALLOC_HS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_PUSH_CONSTANT_ALLOC_DS", OP_3DSTATE_PUSH_CONSTANT_ALLOC_DS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_PUSH_CONSTANT_ALLOC_GS", OP_3DSTATE_PUSH_CONSTANT_ALLOC_GS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_PUSH_CONSTANT_ALLOC_PS", OP_3DSTATE_PUSH_CONSTANT_ALLOC_PS,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_MONOFILTER_SIZE", OP_3DSTATE_MONOFILTER_SIZE, F_LEN_VAR,
+		R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SO_DECL_LIST", OP_3DSTATE_SO_DECL_LIST, F_LEN_VAR, R_RCS,
+		D_ALL, 0, 9, NULL},
+
+	{"3DSTATE_SO_BUFFER", OP_3DSTATE_SO_BUFFER, F_LEN_VAR, R_RCS, D_BDW_PLUS,
+		ADDR_FIX_2(2, 4), 8, NULL},
+
+	{"3DSTATE_BINDING_TABLE_POOL_ALLOC",
+		OP_3DSTATE_BINDING_TABLE_POOL_ALLOC,
+		F_LEN_VAR, R_RCS, D_BDW_PLUS, ADDR_FIX_1(1), 8, NULL},
+
+	{"3DSTATE_GATHER_POOL_ALLOC", OP_3DSTATE_GATHER_POOL_ALLOC,
+		F_LEN_VAR, R_RCS, D_BDW_PLUS, ADDR_FIX_1(1), 8, NULL},
+
+	{"3DSTATE_DX9_CONSTANT_BUFFER_POOL_ALLOC",
+		OP_3DSTATE_DX9_CONSTANT_BUFFER_POOL_ALLOC,
+		F_LEN_VAR, R_RCS, D_BDW_PLUS, ADDR_FIX_1(1), 8, NULL},
+
+	{"3DSTATE_SAMPLE_PATTERN", OP_3DSTATE_SAMPLE_PATTERN, F_LEN_VAR, R_RCS,
+		D_BDW_PLUS, 0, 8, NULL},
+
+	{"PIPE_CONTROL", OP_PIPE_CONTROL, F_LEN_VAR, R_RCS, D_ALL,
+		ADDR_FIX_1(2), 8, cmd_handler_pipe_control},
+
+	{"3DPRIMITIVE", OP_3DPRIMITIVE, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"PIPELINE_SELECT", OP_PIPELINE_SELECT, F_LEN_CONST, R_RCS, D_ALL, 0,
+		1, NULL},
+
+	{"STATE_PREFETCH", OP_STATE_PREFETCH, F_LEN_VAR, R_RCS, D_ALL,
+		ADDR_FIX_1(1), 8, NULL},
+
+	{"STATE_SIP", OP_STATE_SIP, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"STATE_BASE_ADDRESS", OP_STATE_BASE_ADDRESS, F_LEN_VAR, R_RCS, D_BDW_PLUS,
+		ADDR_FIX_5(1, 3, 4, 5, 6), 8, NULL},
+
+	{"OP_3D_MEDIA_0_1_4", OP_3D_MEDIA_0_1_4, F_LEN_VAR, R_RCS, D_ALL,
+		ADDR_FIX_1(1), 8, NULL},
+
+	{"OP_SWTESS_BASE_ADDRESS", OP_SWTESS_BASE_ADDRESS,
+		F_LEN_VAR, R_RCS, D_ALL, ADDR_FIX_2(1, 2), 3, NULL},
+
+	{"3DSTATE_VS", OP_3DSTATE_VS, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_SF", OP_3DSTATE_SF, F_LEN_VAR, R_RCS, D_ALL, 0, 8, NULL},
+
+	{"3DSTATE_CONSTANT_VS", OP_3DSTATE_CONSTANT_VS, F_LEN_VAR, R_RCS, D_BDW_PLUS,
+		0, 8, NULL},
+
+	{"3DSTATE_COMPONENT_PACKING", OP_3DSTATE_COMPONENT_PACKING, F_LEN_VAR, R_RCS,
+		D_SKL_PLUS, 0, 8, NULL},
+
+	{"MEDIA_INTERFACE_DESCRIPTOR_LOAD", OP_MEDIA_INTERFACE_DESCRIPTOR_LOAD,
+		F_LEN_VAR, R_RCS, D_ALL, 0, 16, NULL},
+
+	{"MEDIA_GATEWAY_STATE", OP_MEDIA_GATEWAY_STATE, F_LEN_VAR, R_RCS, D_ALL,
+		0, 16, NULL},
+
+	{"MEDIA_STATE_FLUSH", OP_MEDIA_STATE_FLUSH, F_LEN_VAR, R_RCS, D_ALL,
+		0, 16, NULL},
+
+	{"MEDIA_POOL_STATE", OP_MEDIA_POOL_STATE, F_LEN_VAR, R_RCS, D_ALL,
+		0, 16, NULL},
+
+	{"MEDIA_OBJECT", OP_MEDIA_OBJECT, F_LEN_VAR, R_RCS, D_ALL, 0, 16, NULL},
+
+	{"MEDIA_CURBE_LOAD", OP_MEDIA_CURBE_LOAD, F_LEN_VAR, R_RCS, D_ALL,
+		0, 16, NULL},
+
+	{"MEDIA_OBJECT_PRT", OP_MEDIA_OBJECT_PRT, F_LEN_VAR, R_RCS, D_ALL,
+		0, 16, NULL},
+
+	{"MEDIA_OBJECT_WALKER", OP_MEDIA_OBJECT_WALKER, F_LEN_VAR, R_RCS, D_ALL,
+		0, 16, NULL},
+
+	{"GPGPU_WALKER", OP_GPGPU_WALKER, F_LEN_VAR, R_RCS, D_ALL,
+		0, 8, NULL},
+
+	{"MEDIA_VFE_STATE", OP_MEDIA_VFE_STATE, F_LEN_VAR, R_RCS, D_ALL, 0, 16,
+		NULL},
+
+	{"3DSTATE_VF_STATISTICS_GM45", OP_3DSTATE_VF_STATISTICS_GM45,
+		F_LEN_CONST, R_ALL, D_ALL, 0, 1, NULL},
+
+	{"MFX_PIPE_MODE_SELECT", OP_MFX_PIPE_MODE_SELECT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_SURFACE_STATE", OP_MFX_SURFACE_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_PIPE_BUF_ADDR_STATE", OP_MFX_PIPE_BUF_ADDR_STATE, F_LEN_VAR,
+		R_VCS, D_BDW_PLUS, 0, 12, NULL},
+
+	{"MFX_IND_OBJ_BASE_ADDR_STATE", OP_MFX_IND_OBJ_BASE_ADDR_STATE,
+		F_LEN_VAR, R_VCS, D_BDW_PLUS, 0, 12, NULL},
+
+	{"MFX_BSP_BUF_BASE_ADDR_STATE", OP_MFX_BSP_BUF_BASE_ADDR_STATE,
+		F_LEN_VAR, R_VCS, D_BDW_PLUS, ADDR_FIX_3(1, 3, 5), 12, NULL},
+
+	{"OP_2_0_0_5", OP_2_0_0_5, F_LEN_VAR, R_VCS, D_BDW_PLUS, 0, 12, NULL},
+
+	{"MFX_STATE_POINTER", OP_MFX_STATE_POINTER, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_QM_STATE", OP_MFX_QM_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_FQM_STATE", OP_MFX_FQM_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_PAK_INSERT_OBJECT", OP_MFX_PAK_INSERT_OBJECT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_STITCH_OBJECT", OP_MFX_STITCH_OBJECT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFD_IT_OBJECT", OP_MFD_IT_OBJECT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_WAIT", OP_MFX_WAIT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 6, NULL},
+
+	{"MFX_AVC_IMG_STATE", OP_MFX_AVC_IMG_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_AVC_QM_STATE", OP_MFX_AVC_QM_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_AVC_DIRECTMODE_STATE", OP_MFX_AVC_DIRECTMODE_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_AVC_SLICE_STATE", OP_MFX_AVC_SLICE_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_AVC_REF_IDX_STATE", OP_MFX_AVC_REF_IDX_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_AVC_WEIGHTOFFSET_STATE", OP_MFX_AVC_WEIGHTOFFSET_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFD_AVC_PICID_STATE", OP_MFD_AVC_PICID_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+	{"MFD_AVC_DPB_STATE", OP_MFD_AVC_DPB_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFD_AVC_BSD_OBJECT", OP_MFD_AVC_BSD_OBJECT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFD_AVC_SLICEADDR", OP_MFD_AVC_SLICEADDR, F_LEN_VAR,
+		R_VCS, D_ALL, ADDR_FIX_1(2), 12, NULL},
+
+	{"MFC_AVC_PAK_OBJECT", OP_MFC_AVC_PAK_OBJECT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_VC1_PRED_PIPE_STATE", OP_MFX_VC1_PRED_PIPE_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_VC1_DIRECTMODE_STATE", OP_MFX_VC1_DIRECTMODE_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFD_VC1_SHORT_PIC_STATE", OP_MFD_VC1_SHORT_PIC_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFD_VC1_LONG_PIC_STATE", OP_MFD_VC1_LONG_PIC_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFD_VC1_BSD_OBJECT", OP_MFD_VC1_BSD_OBJECT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFC_MPEG2_SLICEGROUP_STATE", OP_MFC_MPEG2_SLICEGROUP_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFC_MPEG2_PAK_OBJECT", OP_MFC_MPEG2_PAK_OBJECT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_MPEG2_PIC_STATE", OP_MFX_MPEG2_PIC_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_MPEG2_QM_STATE", OP_MFX_MPEG2_QM_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFD_MPEG2_BSD_OBJECT", OP_MFD_MPEG2_BSD_OBJECT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_2_6_0_0", OP_MFX_2_6_0_0, F_LEN_VAR, R_VCS, D_ALL,
+		0, 16, NULL},
+
+	{"MFX_2_6_0_9", OP_MFX_2_6_0_9, F_LEN_VAR, R_VCS, D_ALL, 0, 16, NULL},
+
+	{"MFX_2_6_0_8", OP_MFX_2_6_0_8, F_LEN_VAR, R_VCS, D_ALL, 0, 16, NULL},
+
+	{"MFX_JPEG_PIC_STATE", OP_MFX_JPEG_PIC_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFX_JPEG_HUFF_TABLE_STATE", OP_MFX_JPEG_HUFF_TABLE_STATE, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"MFD_JPEG_BSD_OBJECT", OP_MFD_JPEG_BSD_OBJECT, F_LEN_VAR,
+		R_VCS, D_ALL, 0, 12, NULL},
+
+	{"VEBOX_STATE", OP_VEB_STATE, F_LEN_VAR, R_VECS, D_ALL, 0, 12, NULL},
+
+	{"VEBOX_SURFACE_STATE", OP_VEB_SURFACE_STATE, F_LEN_VAR, R_VECS, D_ALL,
+		0, 12, NULL},
+
+	{"VEB_DI_IECP", OP_VEB_DNDI_IECP_STATE, F_LEN_VAR, R_VECS, D_BDW_PLUS,
+		0, 12, NULL},
+};
+
+static void add_cmd_entry(struct intel_gvt *gvt, struct cmd_entry *e)
+{
+	hash_add(gvt->cmd_table, &e->hlist, e->info->opcode);
+}
+
+/* call the cmd handler, and advance ip */
+static int cmd_parser_exec(struct parser_exec_state *s)
+{
+	struct intel_vgpu *vgpu = s->vgpu;
+	const struct cmd_info *info;
+	u32 cmd;
+	int ret = 0;
+
+	cmd = cmd_val(s, 0);
+
+	/* fastpath for MI_NOOP */
+	if (cmd == MI_NOOP)
+		info = &cmd_info[mi_noop_index];
+	else
+		info = get_cmd_info(s->vgpu->gvt, cmd, s->engine);
+
+	if (info == NULL) {
+		gvt_vgpu_err("unknown cmd 0x%x, opcode=0x%x, addr_type=%s, ring %s, workload=%p\n",
+			     cmd, get_opcode(cmd, s->engine),
+			     repr_addr_type(s->buf_addr_type),
+			     s->engine->name, s->workload);
+		return -EBADRQC;
+	}
+
+	s->info = info;
+
+	trace_gvt_command(vgpu->id, s->engine->id, s->ip_gma, s->ip_va,
+			  cmd_length(s), s->buf_type, s->buf_addr_type,
+			  s->workload, info->name);
+
+	if ((info->flag & F_LEN_MASK) == F_LEN_VAR_FIXED) {
+		ret = gvt_check_valid_cmd_length(cmd_length(s),
+						 info->valid_len);
+		if (ret)
+			return ret;
+	}
+
+	if (info->handler) {
+		ret = info->handler(s);
+		if (ret < 0) {
+			gvt_vgpu_err("%s handler error\n", info->name);
+			return ret;
+		}
+	}
+
+	if (!(info->flag & F_IP_ADVANCE_CUSTOM)) {
+		ret = cmd_advance_default(s);
+		if (ret) {
+			gvt_vgpu_err("%s IP advance error\n", info->name);
+			return ret;
+		}
+	}
+	return 0;
+}
+
+static inline bool gma_out_of_range(unsigned long gma,
+		unsigned long gma_head, unsigned int gma_tail)
+{
+	if (gma_tail >= gma_head)
+		return (gma < gma_head) || (gma > gma_tail);
+	else
+		return (gma > gma_tail) && (gma < gma_head);
+}
+
+/* Keep the consistent return type, e.g EBADRQC for unknown
+ * cmd, EFAULT for invalid address, EPERM for nonpriv. later
+ * works as the input of VM healthy status.
+ */
+static int command_scan(struct parser_exec_state *s,
+		unsigned long rb_head, unsigned long rb_tail,
+		unsigned long rb_start, unsigned long rb_len)
+{
+
+	unsigned long gma_head, gma_tail, gma_bottom;
+	int ret = 0;
+	struct intel_vgpu *vgpu = s->vgpu;
+
+	gma_head = rb_start + rb_head;
+	gma_tail = rb_start + rb_tail;
+	gma_bottom = rb_start +  rb_len;
+
+	while (s->ip_gma != gma_tail) {
+		if (s->buf_type == RING_BUFFER_INSTRUCTION ||
+				s->buf_type == RING_BUFFER_CTX) {
+			if (!(s->ip_gma >= rb_start) ||
+				!(s->ip_gma < gma_bottom)) {
+				gvt_vgpu_err("ip_gma %lx out of ring scope."
+					"(base:0x%lx, bottom: 0x%lx)\n",
+					s->ip_gma, rb_start,
+					gma_bottom);
+				parser_exec_state_dump(s);
+				return -EFAULT;
+			}
+			if (gma_out_of_range(s->ip_gma, gma_head, gma_tail)) {
+				gvt_vgpu_err("ip_gma %lx out of range."
+					"base 0x%lx head 0x%lx tail 0x%lx\n",
+					s->ip_gma, rb_start,
+					rb_head, rb_tail);
+				parser_exec_state_dump(s);
+				break;
+			}
+		}
+		ret = cmd_parser_exec(s);
+		if (ret) {
+			gvt_vgpu_err("cmd parser error\n");
+			parser_exec_state_dump(s);
+			break;
+		}
+	}
+
+	return ret;
+}
+
+static int scan_workload(struct intel_vgpu_workload *workload)
+{
+	unsigned long gma_head, gma_tail, gma_bottom;
+	struct parser_exec_state s;
+	int ret = 0;
+
+	/* ring base is page aligned */
+	if (WARN_ON(!IS_ALIGNED(workload->rb_start, I915_GTT_PAGE_SIZE)))
+		return -EINVAL;
+
+	gma_head = workload->rb_start + workload->rb_head;
+	gma_tail = workload->rb_start + workload->rb_tail;
+	gma_bottom = workload->rb_start +  _RING_CTL_BUF_SIZE(workload->rb_ctl);
+
+	s.buf_type = RING_BUFFER_INSTRUCTION;
+	s.buf_addr_type = GTT_BUFFER;
+	s.vgpu = workload->vgpu;
+	s.engine = workload->engine;
+	s.ring_start = workload->rb_start;
+	s.ring_size = _RING_CTL_BUF_SIZE(workload->rb_ctl);
+	s.ring_head = gma_head;
+	s.ring_tail = gma_tail;
+	s.rb_va = workload->shadow_ring_buffer_va;
+	s.workload = workload;
+	s.is_ctx_wa = false;
+
+	if (bypass_scan_mask & workload->engine->mask || gma_head == gma_tail)
+		return 0;
+
+	ret = ip_gma_set(&s, gma_head);
+	if (ret)
+		goto out;
+
+	ret = command_scan(&s, workload->rb_head, workload->rb_tail,
+		workload->rb_start, _RING_CTL_BUF_SIZE(workload->rb_ctl));
+
+out:
+	return ret;
+}
+
+static int scan_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
+{
+
+	unsigned long gma_head, gma_tail, gma_bottom, ring_size, ring_tail;
+	struct parser_exec_state s;
+	int ret = 0;
+	struct intel_vgpu_workload *workload = container_of(wa_ctx,
+				struct intel_vgpu_workload,
+				wa_ctx);
+
+	/* ring base is page aligned */
+	if (WARN_ON(!IS_ALIGNED(wa_ctx->indirect_ctx.guest_gma,
+					I915_GTT_PAGE_SIZE)))
+		return -EINVAL;
+
+	ring_tail = wa_ctx->indirect_ctx.size + 3 * sizeof(u32);
+	ring_size = round_up(wa_ctx->indirect_ctx.size + CACHELINE_BYTES,
+			PAGE_SIZE);
+	gma_head = wa_ctx->indirect_ctx.guest_gma;
+	gma_tail = wa_ctx->indirect_ctx.guest_gma + ring_tail;
+	gma_bottom = wa_ctx->indirect_ctx.guest_gma + ring_size;
+
+	s.buf_type = RING_BUFFER_INSTRUCTION;
+	s.buf_addr_type = GTT_BUFFER;
+	s.vgpu = workload->vgpu;
+	s.engine = workload->engine;
+	s.ring_start = wa_ctx->indirect_ctx.guest_gma;
+	s.ring_size = ring_size;
+	s.ring_head = gma_head;
+	s.ring_tail = gma_tail;
+	s.rb_va = wa_ctx->indirect_ctx.shadow_va;
+	s.workload = workload;
+	s.is_ctx_wa = true;
+
+	ret = ip_gma_set(&s, gma_head);
+	if (ret)
+		goto out;
+
+	ret = command_scan(&s, 0, ring_tail,
+		wa_ctx->indirect_ctx.guest_gma, ring_size);
+out:
+	return ret;
+}
+
+static int shadow_workload_ring_buffer(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	unsigned long gma_head, gma_tail, gma_top, guest_rb_size;
+	void *shadow_ring_buffer_va;
+	int ret;
+
+	guest_rb_size = _RING_CTL_BUF_SIZE(workload->rb_ctl);
+
+	/* calculate workload ring buffer size */
+	workload->rb_len = (workload->rb_tail + guest_rb_size -
+			workload->rb_head) % guest_rb_size;
+
+	gma_head = workload->rb_start + workload->rb_head;
+	gma_tail = workload->rb_start + workload->rb_tail;
+	gma_top = workload->rb_start + guest_rb_size;
+
+	if (workload->rb_len > s->ring_scan_buffer_size[workload->engine->id]) {
+		void *p;
+
+		/* realloc the new ring buffer if needed */
+		p = krealloc(s->ring_scan_buffer[workload->engine->id],
+			     workload->rb_len, GFP_KERNEL);
+		if (!p) {
+			gvt_vgpu_err("fail to re-alloc ring scan buffer\n");
+			return -ENOMEM;
+		}
+		s->ring_scan_buffer[workload->engine->id] = p;
+		s->ring_scan_buffer_size[workload->engine->id] = workload->rb_len;
+	}
+
+	shadow_ring_buffer_va = s->ring_scan_buffer[workload->engine->id];
+
+	/* get shadow ring buffer va */
+	workload->shadow_ring_buffer_va = shadow_ring_buffer_va;
+
+	/* head > tail --> copy head <-> top */
+	if (gma_head > gma_tail) {
+		ret = copy_gma_to_hva(vgpu, vgpu->gtt.ggtt_mm,
+				      gma_head, gma_top, shadow_ring_buffer_va);
+		if (ret < 0) {
+			gvt_vgpu_err("fail to copy guest ring buffer\n");
+			return ret;
+		}
+		shadow_ring_buffer_va += ret;
+		gma_head = workload->rb_start;
+	}
+
+	/* copy head or start <-> tail */
+	ret = copy_gma_to_hva(vgpu, vgpu->gtt.ggtt_mm, gma_head, gma_tail,
+				shadow_ring_buffer_va);
+	if (ret < 0) {
+		gvt_vgpu_err("fail to copy guest ring buffer\n");
+		return ret;
+	}
+	return 0;
+}
+
+int intel_gvt_scan_and_shadow_ringbuffer(struct intel_vgpu_workload *workload)
+{
+	int ret;
+	struct intel_vgpu *vgpu = workload->vgpu;
+
+	ret = shadow_workload_ring_buffer(workload);
+	if (ret) {
+		gvt_vgpu_err("fail to shadow workload ring_buffer\n");
+		return ret;
+	}
+
+	ret = scan_workload(workload);
+	if (ret) {
+		gvt_vgpu_err("scan workload error\n");
+		return ret;
+	}
+	return 0;
+}
+
+static int shadow_indirect_ctx(struct intel_shadow_wa_ctx *wa_ctx)
+{
+	int ctx_size = wa_ctx->indirect_ctx.size;
+	unsigned long guest_gma = wa_ctx->indirect_ctx.guest_gma;
+	struct intel_vgpu_workload *workload = container_of(wa_ctx,
+					struct intel_vgpu_workload,
+					wa_ctx);
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct drm_i915_gem_object *obj;
+	int ret = 0;
+	void *map;
+
+	obj = i915_gem_object_create_shmem(workload->engine->i915,
+					   roundup(ctx_size + CACHELINE_BYTES,
+						   PAGE_SIZE));
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	/* get the va of the shadow batch buffer */
+	map = i915_gem_object_pin_map(obj, I915_MAP_WB);
+	if (IS_ERR(map)) {
+		gvt_vgpu_err("failed to vmap shadow indirect ctx\n");
+		ret = PTR_ERR(map);
+		goto put_obj;
+	}
+
+	i915_gem_object_lock(obj, NULL);
+	ret = i915_gem_object_set_to_cpu_domain(obj, false);
+	i915_gem_object_unlock(obj);
+	if (ret) {
+		gvt_vgpu_err("failed to set shadow indirect ctx to CPU\n");
+		goto unmap_src;
+	}
+
+	ret = copy_gma_to_hva(workload->vgpu,
+				workload->vgpu->gtt.ggtt_mm,
+				guest_gma, guest_gma + ctx_size,
+				map);
+	if (ret < 0) {
+		gvt_vgpu_err("fail to copy guest indirect ctx\n");
+		goto unmap_src;
+	}
+
+	wa_ctx->indirect_ctx.obj = obj;
+	wa_ctx->indirect_ctx.shadow_va = map;
+	return 0;
+
+unmap_src:
+	i915_gem_object_unpin_map(obj);
+put_obj:
+	i915_gem_object_put(obj);
+	return ret;
+}
+
+static int combine_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
+{
+	u32 per_ctx_start[CACHELINE_DWORDS] = {0};
+	unsigned char *bb_start_sva;
+
+	if (!wa_ctx->per_ctx.valid)
+		return 0;
+
+	per_ctx_start[0] = 0x18800001;
+	per_ctx_start[1] = wa_ctx->per_ctx.guest_gma;
+
+	bb_start_sva = (unsigned char *)wa_ctx->indirect_ctx.shadow_va +
+				wa_ctx->indirect_ctx.size;
+
+	memcpy(bb_start_sva, per_ctx_start, CACHELINE_BYTES);
+
+	return 0;
+}
+
+int intel_gvt_scan_and_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
+{
+	int ret;
+	struct intel_vgpu_workload *workload = container_of(wa_ctx,
+					struct intel_vgpu_workload,
+					wa_ctx);
+	struct intel_vgpu *vgpu = workload->vgpu;
+
+	if (wa_ctx->indirect_ctx.size == 0)
+		return 0;
+
+	ret = shadow_indirect_ctx(wa_ctx);
+	if (ret) {
+		gvt_vgpu_err("fail to shadow indirect ctx\n");
+		return ret;
+	}
+
+	combine_wa_ctx(wa_ctx);
+
+	ret = scan_wa_ctx(wa_ctx);
+	if (ret) {
+		gvt_vgpu_err("scan wa ctx error\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+/* generate dummy contexts by sending empty requests to HW, and let
+ * the HW to fill Engine Contexts. This dummy contexts are used for
+ * initialization purpose (update reg whitelist), so referred to as
+ * init context here
+ */
+void intel_gvt_update_reg_whitelist(struct intel_vgpu *vgpu)
+{
+	const unsigned long start = LRC_STATE_PN * PAGE_SIZE;
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	if (gvt->is_reg_whitelist_updated)
+		return;
+
+	/* scan init ctx to update cmd accessible list */
+	for_each_engine(engine, gvt->gt, id) {
+		struct parser_exec_state s;
+		void *vaddr;
+		int ret;
+
+		if (!engine->default_state)
+			continue;
+
+		vaddr = shmem_pin_map(engine->default_state);
+		if (!vaddr) {
+			gvt_err("failed to map %s->default state\n",
+				engine->name);
+			return;
+		}
+
+		s.buf_type = RING_BUFFER_CTX;
+		s.buf_addr_type = GTT_BUFFER;
+		s.vgpu = vgpu;
+		s.engine = engine;
+		s.ring_start = 0;
+		s.ring_size = engine->context_size - start;
+		s.ring_head = 0;
+		s.ring_tail = s.ring_size;
+		s.rb_va = vaddr + start;
+		s.workload = NULL;
+		s.is_ctx_wa = false;
+		s.is_init_ctx = true;
+
+		/* skipping the first RING_CTX_SIZE(0x50) dwords */
+		ret = ip_gma_set(&s, RING_CTX_SIZE);
+		if (ret == 0) {
+			ret = command_scan(&s, 0, s.ring_size, 0, s.ring_size);
+			if (ret)
+				gvt_err("Scan init ctx error\n");
+		}
+
+		shmem_unpin_map(engine->default_state, vaddr);
+		if (ret)
+			return;
+	}
+
+	gvt->is_reg_whitelist_updated = true;
+}
+
+int intel_gvt_scan_engine_context(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	unsigned long gma_head, gma_tail, gma_start, ctx_size;
+	struct parser_exec_state s;
+	int ring_id = workload->engine->id;
+	struct intel_context *ce = vgpu->submission.shadow[ring_id];
+	int ret;
+
+	GEM_BUG_ON(atomic_read(&ce->pin_count) < 0);
+
+	ctx_size = workload->engine->context_size - PAGE_SIZE;
+
+	/* Only ring contxt is loaded to HW for inhibit context, no need to
+	 * scan engine context
+	 */
+	if (is_inhibit_context(ce))
+		return 0;
+
+	gma_start = i915_ggtt_offset(ce->state) + LRC_STATE_PN*PAGE_SIZE;
+	gma_head = 0;
+	gma_tail = ctx_size;
+
+	s.buf_type = RING_BUFFER_CTX;
+	s.buf_addr_type = GTT_BUFFER;
+	s.vgpu = workload->vgpu;
+	s.engine = workload->engine;
+	s.ring_start = gma_start;
+	s.ring_size = ctx_size;
+	s.ring_head = gma_start + gma_head;
+	s.ring_tail = gma_start + gma_tail;
+	s.rb_va = ce->lrc_reg_state;
+	s.workload = workload;
+	s.is_ctx_wa = false;
+	s.is_init_ctx = false;
+
+	/* don't scan the first RING_CTX_SIZE(0x50) dwords, as it's ring
+	 * context
+	 */
+	ret = ip_gma_set(&s, gma_start + gma_head + RING_CTX_SIZE);
+	if (ret)
+		goto out;
+
+	ret = command_scan(&s, gma_head, gma_tail,
+		gma_start, ctx_size);
+out:
+	if (ret)
+		gvt_vgpu_err("scan shadow ctx error\n");
+
+	return ret;
+}
+
+static int init_cmd_table(struct intel_gvt *gvt)
+{
+	unsigned int gen_type = intel_gvt_get_device_type(gvt);
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(cmd_info); i++) {
+		struct cmd_entry *e;
+
+		if (!(cmd_info[i].devices & gen_type))
+			continue;
+
+		e = kzalloc(sizeof(*e), GFP_KERNEL);
+		if (!e)
+			return -ENOMEM;
+
+		e->info = &cmd_info[i];
+		if (cmd_info[i].opcode == OP_MI_NOOP)
+			mi_noop_index = i;
+
+		INIT_HLIST_NODE(&e->hlist);
+		add_cmd_entry(gvt, e);
+		gvt_dbg_cmd("add %-30s op %04x flag %x devs %02x rings %02x\n",
+			    e->info->name, e->info->opcode, e->info->flag,
+			    e->info->devices, e->info->rings);
+	}
+
+	return 0;
+}
+
+static void clean_cmd_table(struct intel_gvt *gvt)
+{
+	struct hlist_node *tmp;
+	struct cmd_entry *e;
+	int i;
+
+	hash_for_each_safe(gvt->cmd_table, i, tmp, e, hlist)
+		kfree(e);
+
+	hash_init(gvt->cmd_table);
+}
+
+void intel_gvt_clean_cmd_parser(struct intel_gvt *gvt)
+{
+	clean_cmd_table(gvt);
+}
+
+int intel_gvt_init_cmd_parser(struct intel_gvt *gvt)
+{
+	int ret;
+
+	ret = init_cmd_table(gvt);
+	if (ret) {
+		intel_gvt_clean_cmd_parser(gvt);
+		return ret;
+	}
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gvt/cmd_parser.h b/drivers/gpu/drm/i915/gvt/cmd_parser.h
new file mode 100644
index 000000000..416d345e2
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/cmd_parser.h
@@ -0,0 +1,58 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Ke Yu
+ *    Kevin Tian <kevin.tian@intel.com>
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *
+ * Contributors:
+ *    Min He <min.he@intel.com>
+ *    Ping Gao <ping.a.gao@intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *    Yulei Zhang <yulei.zhang@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+#ifndef _GVT_CMD_PARSER_H_
+#define _GVT_CMD_PARSER_H_
+
+#define GVT_CMD_HASH_BITS 7
+
+struct intel_gvt;
+struct intel_shadow_wa_ctx;
+struct intel_vgpu;
+struct intel_vgpu_workload;
+
+void intel_gvt_clean_cmd_parser(struct intel_gvt *gvt);
+
+int intel_gvt_init_cmd_parser(struct intel_gvt *gvt);
+
+int intel_gvt_scan_and_shadow_ringbuffer(struct intel_vgpu_workload *workload);
+
+int intel_gvt_scan_and_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx);
+
+void intel_gvt_update_reg_whitelist(struct intel_vgpu *vgpu);
+
+int intel_gvt_scan_engine_context(struct intel_vgpu_workload *workload);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/debug.h b/drivers/gpu/drm/i915/gvt/debug.h
new file mode 100644
index 000000000..c6027125c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/debug.h
@@ -0,0 +1,65 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __GVT_DEBUG_H__
+#define __GVT_DEBUG_H__
+
+#define gvt_err(fmt, args...) \
+	pr_err("gvt: "fmt, ##args)
+
+#define gvt_vgpu_err(fmt, args...)					\
+do {									\
+	if (IS_ERR_OR_NULL(vgpu))					\
+		pr_err("gvt: "fmt, ##args);			\
+	else								\
+		pr_err("gvt: vgpu %d: "fmt, vgpu->id, ##args);\
+} while (0)
+
+#define gvt_dbg_core(fmt, args...) \
+	pr_debug("gvt: core: "fmt, ##args)
+
+#define gvt_dbg_irq(fmt, args...) \
+	pr_debug("gvt: irq: "fmt, ##args)
+
+#define gvt_dbg_mm(fmt, args...) \
+	pr_debug("gvt: mm: "fmt, ##args)
+
+#define gvt_dbg_mmio(fmt, args...) \
+	pr_debug("gvt: mmio: "fmt, ##args)
+
+#define gvt_dbg_dpy(fmt, args...) \
+	pr_debug("gvt: dpy: "fmt, ##args)
+
+#define gvt_dbg_el(fmt, args...) \
+	pr_debug("gvt: el: "fmt, ##args)
+
+#define gvt_dbg_sched(fmt, args...) \
+	pr_debug("gvt: sched: "fmt, ##args)
+
+#define gvt_dbg_render(fmt, args...) \
+	pr_debug("gvt: render: "fmt, ##args)
+
+#define gvt_dbg_cmd(fmt, args...) \
+	pr_debug("gvt: cmd: "fmt, ##args)
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/debugfs.c b/drivers/gpu/drm/i915/gvt/debugfs.c
new file mode 100644
index 000000000..e08ed0e9f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/debugfs.c
@@ -0,0 +1,213 @@
+/*
+ * Copyright(c) 2011-2017 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#include <linux/debugfs.h>
+#include <linux/list_sort.h>
+#include "i915_drv.h"
+#include "gvt.h"
+
+struct mmio_diff_param {
+	struct intel_vgpu *vgpu;
+	int total;
+	int diff;
+	struct list_head diff_mmio_list;
+};
+
+struct diff_mmio {
+	struct list_head node;
+	u32 offset;
+	u32 preg;
+	u32 vreg;
+};
+
+/* Compare two diff_mmio items. */
+static int mmio_offset_compare(void *priv,
+	const struct list_head *a, const struct list_head *b)
+{
+	struct diff_mmio *ma;
+	struct diff_mmio *mb;
+
+	ma = container_of(a, struct diff_mmio, node);
+	mb = container_of(b, struct diff_mmio, node);
+	if (ma->offset < mb->offset)
+		return -1;
+	else if (ma->offset > mb->offset)
+		return 1;
+	return 0;
+}
+
+static inline int mmio_diff_handler(struct intel_gvt *gvt,
+				    u32 offset, void *data)
+{
+	struct mmio_diff_param *param = data;
+	struct diff_mmio *node;
+	u32 preg, vreg;
+
+	preg = intel_uncore_read_notrace(gvt->gt->uncore, _MMIO(offset));
+	vreg = vgpu_vreg(param->vgpu, offset);
+
+	if (preg != vreg) {
+		node = kmalloc(sizeof(*node), GFP_ATOMIC);
+		if (!node)
+			return -ENOMEM;
+
+		node->offset = offset;
+		node->preg = preg;
+		node->vreg = vreg;
+		list_add(&node->node, &param->diff_mmio_list);
+		param->diff++;
+	}
+	param->total++;
+	return 0;
+}
+
+/* Show the all the different values of tracked mmio. */
+static int vgpu_mmio_diff_show(struct seq_file *s, void *unused)
+{
+	struct intel_vgpu *vgpu = s->private;
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct mmio_diff_param param = {
+		.vgpu = vgpu,
+		.total = 0,
+		.diff = 0,
+	};
+	struct diff_mmio *node, *next;
+
+	INIT_LIST_HEAD(&param.diff_mmio_list);
+
+	mutex_lock(&gvt->lock);
+	spin_lock_bh(&gvt->scheduler.mmio_context_lock);
+
+	mmio_hw_access_pre(gvt->gt);
+	/* Recognize all the diff mmios to list. */
+	intel_gvt_for_each_tracked_mmio(gvt, mmio_diff_handler, &param);
+	mmio_hw_access_post(gvt->gt);
+
+	spin_unlock_bh(&gvt->scheduler.mmio_context_lock);
+	mutex_unlock(&gvt->lock);
+
+	/* In an ascending order by mmio offset. */
+	list_sort(NULL, &param.diff_mmio_list, mmio_offset_compare);
+
+	seq_printf(s, "%-8s %-8s %-8s %-8s\n", "Offset", "HW", "vGPU", "Diff");
+	list_for_each_entry_safe(node, next, &param.diff_mmio_list, node) {
+		u32 diff = node->preg ^ node->vreg;
+
+		seq_printf(s, "%08x %08x %08x %*pbl\n",
+			   node->offset, node->preg, node->vreg,
+			   32, &diff);
+		list_del(&node->node);
+		kfree(node);
+	}
+	seq_printf(s, "Total: %d, Diff: %d\n", param.total, param.diff);
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(vgpu_mmio_diff);
+
+static int
+vgpu_scan_nonprivbb_get(void *data, u64 *val)
+{
+	struct intel_vgpu *vgpu = (struct intel_vgpu *)data;
+
+	*val = vgpu->scan_nonprivbb;
+	return 0;
+}
+
+/*
+ * set/unset bit engine_id of vgpu->scan_nonprivbb to turn on/off scanning
+ * of non-privileged batch buffer. e.g.
+ * if vgpu->scan_nonprivbb=3, then it will scan non-privileged batch buffer
+ * on engine 0 and 1.
+ */
+static int
+vgpu_scan_nonprivbb_set(void *data, u64 val)
+{
+	struct intel_vgpu *vgpu = (struct intel_vgpu *)data;
+
+	vgpu->scan_nonprivbb = val;
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(vgpu_scan_nonprivbb_fops,
+			vgpu_scan_nonprivbb_get, vgpu_scan_nonprivbb_set,
+			"0x%llx\n");
+
+/**
+ * intel_gvt_debugfs_add_vgpu - register debugfs entries for a vGPU
+ * @vgpu: a vGPU
+ */
+void intel_gvt_debugfs_add_vgpu(struct intel_vgpu *vgpu)
+{
+	char name[16] = "";
+
+	snprintf(name, 16, "vgpu%d", vgpu->id);
+	vgpu->debugfs = debugfs_create_dir(name, vgpu->gvt->debugfs_root);
+
+	debugfs_create_bool("active", 0444, vgpu->debugfs, &vgpu->active);
+	debugfs_create_file("mmio_diff", 0444, vgpu->debugfs, vgpu,
+			    &vgpu_mmio_diff_fops);
+	debugfs_create_file("scan_nonprivbb", 0644, vgpu->debugfs, vgpu,
+			    &vgpu_scan_nonprivbb_fops);
+}
+
+/**
+ * intel_gvt_debugfs_remove_vgpu - remove debugfs entries of a vGPU
+ * @vgpu: a vGPU
+ */
+void intel_gvt_debugfs_remove_vgpu(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct drm_minor *minor = gvt->gt->i915->drm.primary;
+
+	if (minor->debugfs_root && gvt->debugfs_root) {
+		debugfs_remove_recursive(vgpu->debugfs);
+		vgpu->debugfs = NULL;
+	}
+}
+
+/**
+ * intel_gvt_debugfs_init - register gvt debugfs root entry
+ * @gvt: GVT device
+ */
+void intel_gvt_debugfs_init(struct intel_gvt *gvt)
+{
+	struct drm_minor *minor = gvt->gt->i915->drm.primary;
+
+	gvt->debugfs_root = debugfs_create_dir("gvt", minor->debugfs_root);
+
+	debugfs_create_ulong("num_tracked_mmio", 0444, gvt->debugfs_root,
+			     &gvt->mmio.num_tracked_mmio);
+}
+
+/**
+ * intel_gvt_debugfs_clean - remove debugfs entries
+ * @gvt: GVT device
+ */
+void intel_gvt_debugfs_clean(struct intel_gvt *gvt)
+{
+	struct drm_minor *minor = gvt->gt->i915->drm.primary;
+
+	if (minor->debugfs_root) {
+		debugfs_remove_recursive(gvt->debugfs_root);
+		gvt->debugfs_root = NULL;
+	}
+}
diff --git a/drivers/gpu/drm/i915/gvt/display.c b/drivers/gpu/drm/i915/gvt/display.c
new file mode 100644
index 000000000..c7722c818
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/display.c
@@ -0,0 +1,808 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Ke Yu
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *
+ * Contributors:
+ *    Terrence Xu <terrence.xu@intel.com>
+ *    Changbin Du <changbin.du@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "gvt.h"
+
+static int get_edp_pipe(struct intel_vgpu *vgpu)
+{
+	u32 data = vgpu_vreg(vgpu, _TRANS_DDI_FUNC_CTL_EDP);
+	int pipe = -1;
+
+	switch (data & TRANS_DDI_EDP_INPUT_MASK) {
+	case TRANS_DDI_EDP_INPUT_A_ON:
+	case TRANS_DDI_EDP_INPUT_A_ONOFF:
+		pipe = PIPE_A;
+		break;
+	case TRANS_DDI_EDP_INPUT_B_ONOFF:
+		pipe = PIPE_B;
+		break;
+	case TRANS_DDI_EDP_INPUT_C_ONOFF:
+		pipe = PIPE_C;
+		break;
+	}
+	return pipe;
+}
+
+static int edp_pipe_is_enabled(struct intel_vgpu *vgpu)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+
+	if (!(vgpu_vreg_t(vgpu, PIPECONF(_PIPE_EDP)) & PIPECONF_ENABLE))
+		return 0;
+
+	if (!(vgpu_vreg(vgpu, _TRANS_DDI_FUNC_CTL_EDP) & TRANS_DDI_FUNC_ENABLE))
+		return 0;
+	return 1;
+}
+
+int pipe_is_enabled(struct intel_vgpu *vgpu, int pipe)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+
+	if (drm_WARN_ON(&dev_priv->drm,
+			pipe < PIPE_A || pipe >= I915_MAX_PIPES))
+		return -EINVAL;
+
+	if (vgpu_vreg_t(vgpu, PIPECONF(pipe)) & PIPECONF_ENABLE)
+		return 1;
+
+	if (edp_pipe_is_enabled(vgpu) &&
+			get_edp_pipe(vgpu) == pipe)
+		return 1;
+	return 0;
+}
+
+static unsigned char virtual_dp_monitor_edid[GVT_EDID_NUM][EDID_SIZE] = {
+	{
+/* EDID with 1024x768 as its resolution */
+		/*Header*/
+		0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00,
+		/* Vendor & Product Identification */
+		0x22, 0xf0, 0x54, 0x29, 0x00, 0x00, 0x00, 0x00, 0x04, 0x17,
+		/* Version & Revision */
+		0x01, 0x04,
+		/* Basic Display Parameters & Features */
+		0xa5, 0x34, 0x20, 0x78, 0x23,
+		/* Color Characteristics */
+		0xfc, 0x81, 0xa4, 0x55, 0x4d, 0x9d, 0x25, 0x12, 0x50, 0x54,
+		/* Established Timings: maximum resolution is 1024x768 */
+		0x21, 0x08, 0x00,
+		/* Standard Timings. All invalid */
+		0x00, 0xc0, 0x00, 0xc0, 0x00, 0x40, 0x00, 0x80, 0x00, 0x00,
+		0x00, 0x40, 0x00, 0x00, 0x00, 0x01,
+		/* 18 Byte Data Blocks 1: invalid */
+		0x00, 0x00, 0x80, 0xa0, 0x70, 0xb0,
+		0x23, 0x40, 0x30, 0x20, 0x36, 0x00, 0x06, 0x44, 0x21, 0x00, 0x00, 0x1a,
+		/* 18 Byte Data Blocks 2: invalid */
+		0x00, 0x00, 0x00, 0xfd, 0x00, 0x18, 0x3c, 0x18, 0x50, 0x11, 0x00, 0x0a,
+		0x20, 0x20, 0x20, 0x20, 0x20, 0x20,
+		/* 18 Byte Data Blocks 3: invalid */
+		0x00, 0x00, 0x00, 0xfc, 0x00, 0x48,
+		0x50, 0x20, 0x5a, 0x52, 0x32, 0x34, 0x34, 0x30, 0x77, 0x0a, 0x20, 0x20,
+		/* 18 Byte Data Blocks 4: invalid */
+		0x00, 0x00, 0x00, 0xff, 0x00, 0x43, 0x4e, 0x34, 0x33, 0x30, 0x34, 0x30,
+		0x44, 0x58, 0x51, 0x0a, 0x20, 0x20,
+		/* Extension Block Count */
+		0x00,
+		/* Checksum */
+		0xef,
+	},
+	{
+/* EDID with 1920x1200 as its resolution */
+		/*Header*/
+		0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00,
+		/* Vendor & Product Identification */
+		0x22, 0xf0, 0x54, 0x29, 0x00, 0x00, 0x00, 0x00, 0x04, 0x17,
+		/* Version & Revision */
+		0x01, 0x04,
+		/* Basic Display Parameters & Features */
+		0xa5, 0x34, 0x20, 0x78, 0x23,
+		/* Color Characteristics */
+		0xfc, 0x81, 0xa4, 0x55, 0x4d, 0x9d, 0x25, 0x12, 0x50, 0x54,
+		/* Established Timings: maximum resolution is 1024x768 */
+		0x21, 0x08, 0x00,
+		/*
+		 * Standard Timings.
+		 * below new resolutions can be supported:
+		 * 1920x1080, 1280x720, 1280x960, 1280x1024,
+		 * 1440x900, 1600x1200, 1680x1050
+		 */
+		0xd1, 0xc0, 0x81, 0xc0, 0x81, 0x40, 0x81, 0x80, 0x95, 0x00,
+		0xa9, 0x40, 0xb3, 0x00, 0x01, 0x01,
+		/* 18 Byte Data Blocks 1: max resolution is 1920x1200 */
+		0x28, 0x3c, 0x80, 0xa0, 0x70, 0xb0,
+		0x23, 0x40, 0x30, 0x20, 0x36, 0x00, 0x06, 0x44, 0x21, 0x00, 0x00, 0x1a,
+		/* 18 Byte Data Blocks 2: invalid */
+		0x00, 0x00, 0x00, 0xfd, 0x00, 0x18, 0x3c, 0x18, 0x50, 0x11, 0x00, 0x0a,
+		0x20, 0x20, 0x20, 0x20, 0x20, 0x20,
+		/* 18 Byte Data Blocks 3: invalid */
+		0x00, 0x00, 0x00, 0xfc, 0x00, 0x48,
+		0x50, 0x20, 0x5a, 0x52, 0x32, 0x34, 0x34, 0x30, 0x77, 0x0a, 0x20, 0x20,
+		/* 18 Byte Data Blocks 4: invalid */
+		0x00, 0x00, 0x00, 0xff, 0x00, 0x43, 0x4e, 0x34, 0x33, 0x30, 0x34, 0x30,
+		0x44, 0x58, 0x51, 0x0a, 0x20, 0x20,
+		/* Extension Block Count */
+		0x00,
+		/* Checksum */
+		0x45,
+	},
+};
+
+#define DPCD_HEADER_SIZE        0xb
+
+/* let the virtual display supports DP1.2 */
+static u8 dpcd_fix_data[DPCD_HEADER_SIZE] = {
+	0x12, 0x014, 0x84, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
+};
+
+static void emulate_monitor_status_change(struct intel_vgpu *vgpu)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+	int pipe;
+
+	if (IS_BROXTON(dev_priv)) {
+		enum transcoder trans;
+		enum port port;
+
+		/* Clear PIPE, DDI, PHY, HPD before setting new */
+		vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) &=
+			~(GEN8_DE_PORT_HOTPLUG(HPD_PORT_A) |
+			  GEN8_DE_PORT_HOTPLUG(HPD_PORT_B) |
+			  GEN8_DE_PORT_HOTPLUG(HPD_PORT_C));
+
+		for_each_pipe(dev_priv, pipe) {
+			vgpu_vreg_t(vgpu, PIPECONF(pipe)) &=
+				~(PIPECONF_ENABLE | PIPECONF_STATE_ENABLE);
+			vgpu_vreg_t(vgpu, DSPCNTR(pipe)) &= ~DISP_ENABLE;
+			vgpu_vreg_t(vgpu, SPRCTL(pipe)) &= ~SPRITE_ENABLE;
+			vgpu_vreg_t(vgpu, CURCNTR(pipe)) &= ~MCURSOR_MODE_MASK;
+			vgpu_vreg_t(vgpu, CURCNTR(pipe)) |= MCURSOR_MODE_DISABLE;
+		}
+
+		for (trans = TRANSCODER_A; trans <= TRANSCODER_EDP; trans++) {
+			vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(trans)) &=
+				~(TRANS_DDI_BPC_MASK | TRANS_DDI_MODE_SELECT_MASK |
+				  TRANS_DDI_PORT_MASK | TRANS_DDI_FUNC_ENABLE);
+		}
+		vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_A)) &=
+			~(TRANS_DDI_BPC_MASK | TRANS_DDI_MODE_SELECT_MASK |
+			  TRANS_DDI_PORT_MASK);
+
+		for (port = PORT_A; port <= PORT_C; port++) {
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(port)) &=
+				~BXT_PHY_LANE_ENABLED;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(port)) |=
+				(BXT_PHY_CMNLANE_POWERDOWN_ACK |
+				 BXT_PHY_LANE_POWERDOWN_ACK);
+
+			vgpu_vreg_t(vgpu, BXT_PORT_PLL_ENABLE(port)) &=
+				~(PORT_PLL_POWER_STATE | PORT_PLL_POWER_ENABLE |
+				  PORT_PLL_REF_SEL | PORT_PLL_LOCK |
+				  PORT_PLL_ENABLE);
+
+			vgpu_vreg_t(vgpu, DDI_BUF_CTL(port)) &=
+				~(DDI_INIT_DISPLAY_DETECTED |
+				  DDI_BUF_CTL_ENABLE);
+			vgpu_vreg_t(vgpu, DDI_BUF_CTL(port)) |= DDI_BUF_IS_IDLE;
+		}
+		vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+			~(PORTA_HOTPLUG_ENABLE | PORTA_HOTPLUG_STATUS_MASK);
+		vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+			~(PORTB_HOTPLUG_ENABLE | PORTB_HOTPLUG_STATUS_MASK);
+		vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+			~(PORTC_HOTPLUG_ENABLE | PORTC_HOTPLUG_STATUS_MASK);
+		/* No hpd_invert set in vgpu vbt, need to clear invert mask */
+		vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &= ~BXT_DDI_HPD_INVERT_MASK;
+		vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) &= ~BXT_DE_PORT_HOTPLUG_MASK;
+
+		vgpu_vreg_t(vgpu, BXT_P_CR_GT_DISP_PWRON) &= ~(BIT(0) | BIT(1));
+		vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY0)) &=
+			~PHY_POWER_GOOD;
+		vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY1)) &=
+			~PHY_POWER_GOOD;
+		vgpu_vreg_t(vgpu, BXT_PHY_CTL_FAMILY(DPIO_PHY0)) &= ~BIT(30);
+		vgpu_vreg_t(vgpu, BXT_PHY_CTL_FAMILY(DPIO_PHY1)) &= ~BIT(30);
+
+		vgpu_vreg_t(vgpu, SFUSE_STRAP) &= ~SFUSE_STRAP_DDIB_DETECTED;
+		vgpu_vreg_t(vgpu, SFUSE_STRAP) &= ~SFUSE_STRAP_DDIC_DETECTED;
+
+		/*
+		 * Only 1 PIPE enabled in current vGPU display and PIPE_A is
+		 *  tied to TRANSCODER_A in HW, so it's safe to assume PIPE_A,
+		 *   TRANSCODER_A can be enabled. PORT_x depends on the input of
+		 *   setup_virtual_dp_monitor.
+		 */
+		vgpu_vreg_t(vgpu, PIPECONF(PIPE_A)) |= PIPECONF_ENABLE;
+		vgpu_vreg_t(vgpu, PIPECONF(PIPE_A)) |= PIPECONF_STATE_ENABLE;
+
+		/*
+		 * Golden M/N are calculated based on:
+		 *   24 bpp, 4 lanes, 154000 pixel clk (from virtual EDID),
+		 *   DP link clk 1620 MHz and non-constant_n.
+		 * TODO: calculate DP link symbol clk and stream clk m/n.
+		 */
+		vgpu_vreg_t(vgpu, PIPE_DATA_M1(TRANSCODER_A)) = TU_SIZE(64);
+		vgpu_vreg_t(vgpu, PIPE_DATA_M1(TRANSCODER_A)) |= 0x5b425e;
+		vgpu_vreg_t(vgpu, PIPE_DATA_N1(TRANSCODER_A)) = 0x800000;
+		vgpu_vreg_t(vgpu, PIPE_LINK_M1(TRANSCODER_A)) = 0x3cd6e;
+		vgpu_vreg_t(vgpu, PIPE_LINK_N1(TRANSCODER_A)) = 0x80000;
+
+		/* Enable per-DDI/PORT vreg */
+		if (intel_vgpu_has_monitor_on_port(vgpu, PORT_A)) {
+			vgpu_vreg_t(vgpu, BXT_P_CR_GT_DISP_PWRON) |= BIT(1);
+			vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY1)) |=
+				PHY_POWER_GOOD;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL_FAMILY(DPIO_PHY1)) |=
+				BIT(30);
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_A)) |=
+				BXT_PHY_LANE_ENABLED;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_A)) &=
+				~(BXT_PHY_CMNLANE_POWERDOWN_ACK |
+				  BXT_PHY_LANE_POWERDOWN_ACK);
+			vgpu_vreg_t(vgpu, BXT_PORT_PLL_ENABLE(PORT_A)) |=
+				(PORT_PLL_POWER_STATE | PORT_PLL_POWER_ENABLE |
+				 PORT_PLL_REF_SEL | PORT_PLL_LOCK |
+				 PORT_PLL_ENABLE);
+			vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_A)) |=
+				(DDI_BUF_CTL_ENABLE | DDI_INIT_DISPLAY_DETECTED);
+			vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_A)) &=
+				~DDI_BUF_IS_IDLE;
+			vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_EDP)) |=
+				(TRANS_DDI_BPC_8 | TRANS_DDI_MODE_SELECT_DP_SST |
+				 TRANS_DDI_FUNC_ENABLE);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTA_HOTPLUG_ENABLE;
+			vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
+				GEN8_DE_PORT_HOTPLUG(HPD_PORT_A);
+		}
+
+		if (intel_vgpu_has_monitor_on_port(vgpu, PORT_B)) {
+			vgpu_vreg_t(vgpu, SFUSE_STRAP) |= SFUSE_STRAP_DDIB_DETECTED;
+			vgpu_vreg_t(vgpu, BXT_P_CR_GT_DISP_PWRON) |= BIT(0);
+			vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY0)) |=
+				PHY_POWER_GOOD;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL_FAMILY(DPIO_PHY0)) |=
+				BIT(30);
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_B)) |=
+				BXT_PHY_LANE_ENABLED;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_B)) &=
+				~(BXT_PHY_CMNLANE_POWERDOWN_ACK |
+				  BXT_PHY_LANE_POWERDOWN_ACK);
+			vgpu_vreg_t(vgpu, BXT_PORT_PLL_ENABLE(PORT_B)) |=
+				(PORT_PLL_POWER_STATE | PORT_PLL_POWER_ENABLE |
+				 PORT_PLL_REF_SEL | PORT_PLL_LOCK |
+				 PORT_PLL_ENABLE);
+			vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_B)) |=
+				DDI_BUF_CTL_ENABLE;
+			vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_B)) &=
+				~DDI_BUF_IS_IDLE;
+			vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_A)) |=
+				(TRANS_DDI_BPC_8 | TRANS_DDI_MODE_SELECT_DP_SST |
+				 (PORT_B << TRANS_DDI_PORT_SHIFT) |
+				 TRANS_DDI_FUNC_ENABLE);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTB_HOTPLUG_ENABLE;
+			vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
+				GEN8_DE_PORT_HOTPLUG(HPD_PORT_B);
+		}
+
+		if (intel_vgpu_has_monitor_on_port(vgpu, PORT_C)) {
+			vgpu_vreg_t(vgpu, SFUSE_STRAP) |= SFUSE_STRAP_DDIC_DETECTED;
+			vgpu_vreg_t(vgpu, BXT_P_CR_GT_DISP_PWRON) |= BIT(0);
+			vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY0)) |=
+				PHY_POWER_GOOD;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL_FAMILY(DPIO_PHY0)) |=
+				BIT(30);
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_C)) |=
+				BXT_PHY_LANE_ENABLED;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_C)) &=
+				~(BXT_PHY_CMNLANE_POWERDOWN_ACK |
+				  BXT_PHY_LANE_POWERDOWN_ACK);
+			vgpu_vreg_t(vgpu, BXT_PORT_PLL_ENABLE(PORT_C)) |=
+				(PORT_PLL_POWER_STATE | PORT_PLL_POWER_ENABLE |
+				 PORT_PLL_REF_SEL | PORT_PLL_LOCK |
+				 PORT_PLL_ENABLE);
+			vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_C)) |=
+				DDI_BUF_CTL_ENABLE;
+			vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_C)) &=
+				~DDI_BUF_IS_IDLE;
+			vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_A)) |=
+				(TRANS_DDI_BPC_8 | TRANS_DDI_MODE_SELECT_DP_SST |
+				 (PORT_B << TRANS_DDI_PORT_SHIFT) |
+				 TRANS_DDI_FUNC_ENABLE);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTC_HOTPLUG_ENABLE;
+			vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
+				GEN8_DE_PORT_HOTPLUG(HPD_PORT_C);
+		}
+
+		return;
+	}
+
+	vgpu_vreg_t(vgpu, SDEISR) &= ~(SDE_PORTB_HOTPLUG_CPT |
+			SDE_PORTC_HOTPLUG_CPT |
+			SDE_PORTD_HOTPLUG_CPT);
+
+	if (IS_SKYLAKE(dev_priv) ||
+	    IS_KABYLAKE(dev_priv) ||
+	    IS_COFFEELAKE(dev_priv) ||
+	    IS_COMETLAKE(dev_priv)) {
+		vgpu_vreg_t(vgpu, SDEISR) &= ~(SDE_PORTA_HOTPLUG_SPT |
+				SDE_PORTE_HOTPLUG_SPT);
+		vgpu_vreg_t(vgpu, SKL_FUSE_STATUS) |=
+				SKL_FUSE_DOWNLOAD_STATUS |
+				SKL_FUSE_PG_DIST_STATUS(SKL_PG0) |
+				SKL_FUSE_PG_DIST_STATUS(SKL_PG1) |
+				SKL_FUSE_PG_DIST_STATUS(SKL_PG2);
+		/*
+		 * Only 1 PIPE enabled in current vGPU display and PIPE_A is
+		 *  tied to TRANSCODER_A in HW, so it's safe to assume PIPE_A,
+		 *   TRANSCODER_A can be enabled. PORT_x depends on the input of
+		 *   setup_virtual_dp_monitor, we can bind DPLL0 to any PORT_x
+		 *   so we fixed to DPLL0 here.
+		 * Setup DPLL0: DP link clk 1620 MHz, non SSC, DP Mode
+		 */
+		vgpu_vreg_t(vgpu, DPLL_CTRL1) =
+			DPLL_CTRL1_OVERRIDE(DPLL_ID_SKL_DPLL0);
+		vgpu_vreg_t(vgpu, DPLL_CTRL1) |=
+			DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1620, DPLL_ID_SKL_DPLL0);
+		vgpu_vreg_t(vgpu, LCPLL1_CTL) =
+			LCPLL_PLL_ENABLE | LCPLL_PLL_LOCK;
+		vgpu_vreg_t(vgpu, DPLL_STATUS) = DPLL_LOCK(DPLL_ID_SKL_DPLL0);
+		/*
+		 * Golden M/N are calculated based on:
+		 *   24 bpp, 4 lanes, 154000 pixel clk (from virtual EDID),
+		 *   DP link clk 1620 MHz and non-constant_n.
+		 * TODO: calculate DP link symbol clk and stream clk m/n.
+		 */
+		vgpu_vreg_t(vgpu, PIPE_DATA_M1(TRANSCODER_A)) = TU_SIZE(64);
+		vgpu_vreg_t(vgpu, PIPE_DATA_M1(TRANSCODER_A)) |= 0x5b425e;
+		vgpu_vreg_t(vgpu, PIPE_DATA_N1(TRANSCODER_A)) = 0x800000;
+		vgpu_vreg_t(vgpu, PIPE_LINK_M1(TRANSCODER_A)) = 0x3cd6e;
+		vgpu_vreg_t(vgpu, PIPE_LINK_N1(TRANSCODER_A)) = 0x80000;
+	}
+
+	if (intel_vgpu_has_monitor_on_port(vgpu, PORT_B)) {
+		vgpu_vreg_t(vgpu, DPLL_CTRL2) &=
+			~DPLL_CTRL2_DDI_CLK_OFF(PORT_B);
+		vgpu_vreg_t(vgpu, DPLL_CTRL2) |=
+			DPLL_CTRL2_DDI_CLK_SEL(DPLL_ID_SKL_DPLL0, PORT_B);
+		vgpu_vreg_t(vgpu, DPLL_CTRL2) |=
+			DPLL_CTRL2_DDI_SEL_OVERRIDE(PORT_B);
+		vgpu_vreg_t(vgpu, SFUSE_STRAP) |= SFUSE_STRAP_DDIB_DETECTED;
+		vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_A)) &=
+			~(TRANS_DDI_BPC_MASK | TRANS_DDI_MODE_SELECT_MASK |
+			TRANS_DDI_PORT_MASK);
+		vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_A)) |=
+			(TRANS_DDI_BPC_8 | TRANS_DDI_MODE_SELECT_DP_SST |
+			(PORT_B << TRANS_DDI_PORT_SHIFT) |
+			TRANS_DDI_FUNC_ENABLE);
+		if (IS_BROADWELL(dev_priv)) {
+			vgpu_vreg_t(vgpu, PORT_CLK_SEL(PORT_B)) &=
+				~PORT_CLK_SEL_MASK;
+			vgpu_vreg_t(vgpu, PORT_CLK_SEL(PORT_B)) |=
+				PORT_CLK_SEL_LCPLL_810;
+		}
+		vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_B)) |= DDI_BUF_CTL_ENABLE;
+		vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_B)) &= ~DDI_BUF_IS_IDLE;
+		vgpu_vreg_t(vgpu, SDEISR) |= SDE_PORTB_HOTPLUG_CPT;
+	}
+
+	if (intel_vgpu_has_monitor_on_port(vgpu, PORT_C)) {
+		vgpu_vreg_t(vgpu, DPLL_CTRL2) &=
+			~DPLL_CTRL2_DDI_CLK_OFF(PORT_C);
+		vgpu_vreg_t(vgpu, DPLL_CTRL2) |=
+			DPLL_CTRL2_DDI_CLK_SEL(DPLL_ID_SKL_DPLL0, PORT_C);
+		vgpu_vreg_t(vgpu, DPLL_CTRL2) |=
+			DPLL_CTRL2_DDI_SEL_OVERRIDE(PORT_C);
+		vgpu_vreg_t(vgpu, SDEISR) |= SDE_PORTC_HOTPLUG_CPT;
+		vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_A)) &=
+			~(TRANS_DDI_BPC_MASK | TRANS_DDI_MODE_SELECT_MASK |
+			TRANS_DDI_PORT_MASK);
+		vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_A)) |=
+			(TRANS_DDI_BPC_8 | TRANS_DDI_MODE_SELECT_DP_SST |
+			(PORT_C << TRANS_DDI_PORT_SHIFT) |
+			TRANS_DDI_FUNC_ENABLE);
+		if (IS_BROADWELL(dev_priv)) {
+			vgpu_vreg_t(vgpu, PORT_CLK_SEL(PORT_C)) &=
+				~PORT_CLK_SEL_MASK;
+			vgpu_vreg_t(vgpu, PORT_CLK_SEL(PORT_C)) |=
+				PORT_CLK_SEL_LCPLL_810;
+		}
+		vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_C)) |= DDI_BUF_CTL_ENABLE;
+		vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_C)) &= ~DDI_BUF_IS_IDLE;
+		vgpu_vreg_t(vgpu, SFUSE_STRAP) |= SFUSE_STRAP_DDIC_DETECTED;
+	}
+
+	if (intel_vgpu_has_monitor_on_port(vgpu, PORT_D)) {
+		vgpu_vreg_t(vgpu, DPLL_CTRL2) &=
+			~DPLL_CTRL2_DDI_CLK_OFF(PORT_D);
+		vgpu_vreg_t(vgpu, DPLL_CTRL2) |=
+			DPLL_CTRL2_DDI_CLK_SEL(DPLL_ID_SKL_DPLL0, PORT_D);
+		vgpu_vreg_t(vgpu, DPLL_CTRL2) |=
+			DPLL_CTRL2_DDI_SEL_OVERRIDE(PORT_D);
+		vgpu_vreg_t(vgpu, SDEISR) |= SDE_PORTD_HOTPLUG_CPT;
+		vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_A)) &=
+			~(TRANS_DDI_BPC_MASK | TRANS_DDI_MODE_SELECT_MASK |
+			TRANS_DDI_PORT_MASK);
+		vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_A)) |=
+			(TRANS_DDI_BPC_8 | TRANS_DDI_MODE_SELECT_DP_SST |
+			(PORT_D << TRANS_DDI_PORT_SHIFT) |
+			TRANS_DDI_FUNC_ENABLE);
+		if (IS_BROADWELL(dev_priv)) {
+			vgpu_vreg_t(vgpu, PORT_CLK_SEL(PORT_D)) &=
+				~PORT_CLK_SEL_MASK;
+			vgpu_vreg_t(vgpu, PORT_CLK_SEL(PORT_D)) |=
+				PORT_CLK_SEL_LCPLL_810;
+		}
+		vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_D)) |= DDI_BUF_CTL_ENABLE;
+		vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_D)) &= ~DDI_BUF_IS_IDLE;
+		vgpu_vreg_t(vgpu, SFUSE_STRAP) |= SFUSE_STRAP_DDID_DETECTED;
+	}
+
+	if ((IS_SKYLAKE(dev_priv) ||
+	     IS_KABYLAKE(dev_priv) ||
+	     IS_COFFEELAKE(dev_priv) ||
+	     IS_COMETLAKE(dev_priv)) &&
+			intel_vgpu_has_monitor_on_port(vgpu, PORT_E)) {
+		vgpu_vreg_t(vgpu, SDEISR) |= SDE_PORTE_HOTPLUG_SPT;
+	}
+
+	if (intel_vgpu_has_monitor_on_port(vgpu, PORT_A)) {
+		if (IS_BROADWELL(dev_priv))
+			vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
+				GEN8_DE_PORT_HOTPLUG(HPD_PORT_A);
+		else
+			vgpu_vreg_t(vgpu, SDEISR) |= SDE_PORTA_HOTPLUG_SPT;
+
+		vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_A)) |= DDI_INIT_DISPLAY_DETECTED;
+	}
+
+	/* Clear host CRT status, so guest couldn't detect this host CRT. */
+	if (IS_BROADWELL(dev_priv))
+		vgpu_vreg_t(vgpu, PCH_ADPA) &= ~ADPA_CRT_HOTPLUG_MONITOR_MASK;
+
+	/* Disable Primary/Sprite/Cursor plane */
+	for_each_pipe(dev_priv, pipe) {
+		vgpu_vreg_t(vgpu, DSPCNTR(pipe)) &= ~DISP_ENABLE;
+		vgpu_vreg_t(vgpu, SPRCTL(pipe)) &= ~SPRITE_ENABLE;
+		vgpu_vreg_t(vgpu, CURCNTR(pipe)) &= ~MCURSOR_MODE_MASK;
+		vgpu_vreg_t(vgpu, CURCNTR(pipe)) |= MCURSOR_MODE_DISABLE;
+	}
+
+	vgpu_vreg_t(vgpu, PIPECONF(PIPE_A)) |= PIPECONF_ENABLE;
+}
+
+static void clean_virtual_dp_monitor(struct intel_vgpu *vgpu, int port_num)
+{
+	struct intel_vgpu_port *port = intel_vgpu_port(vgpu, port_num);
+
+	kfree(port->edid);
+	port->edid = NULL;
+
+	kfree(port->dpcd);
+	port->dpcd = NULL;
+}
+
+static enum hrtimer_restart vblank_timer_fn(struct hrtimer *data)
+{
+	struct intel_vgpu_vblank_timer *vblank_timer;
+	struct intel_vgpu *vgpu;
+
+	vblank_timer = container_of(data, struct intel_vgpu_vblank_timer, timer);
+	vgpu = container_of(vblank_timer, struct intel_vgpu, vblank_timer);
+
+	/* Set vblank emulation request per-vGPU bit */
+	intel_gvt_request_service(vgpu->gvt,
+				  INTEL_GVT_REQUEST_EMULATE_VBLANK + vgpu->id);
+	hrtimer_add_expires_ns(&vblank_timer->timer, vblank_timer->period);
+	return HRTIMER_RESTART;
+}
+
+static int setup_virtual_dp_monitor(struct intel_vgpu *vgpu, int port_num,
+				    int type, unsigned int resolution)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	struct intel_vgpu_port *port = intel_vgpu_port(vgpu, port_num);
+	struct intel_vgpu_vblank_timer *vblank_timer = &vgpu->vblank_timer;
+
+	if (drm_WARN_ON(&i915->drm, resolution >= GVT_EDID_NUM))
+		return -EINVAL;
+
+	port->edid = kzalloc(sizeof(*(port->edid)), GFP_KERNEL);
+	if (!port->edid)
+		return -ENOMEM;
+
+	port->dpcd = kzalloc(sizeof(*(port->dpcd)), GFP_KERNEL);
+	if (!port->dpcd) {
+		kfree(port->edid);
+		return -ENOMEM;
+	}
+
+	memcpy(port->edid->edid_block, virtual_dp_monitor_edid[resolution],
+			EDID_SIZE);
+	port->edid->data_valid = true;
+
+	memcpy(port->dpcd->data, dpcd_fix_data, DPCD_HEADER_SIZE);
+	port->dpcd->data_valid = true;
+	port->dpcd->data[DPCD_SINK_COUNT] = 0x1;
+	port->type = type;
+	port->id = resolution;
+	port->vrefresh_k = GVT_DEFAULT_REFRESH_RATE * MSEC_PER_SEC;
+	vgpu->display.port_num = port_num;
+
+	/* Init hrtimer based on default refresh rate */
+	hrtimer_init(&vblank_timer->timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS);
+	vblank_timer->timer.function = vblank_timer_fn;
+	vblank_timer->vrefresh_k = port->vrefresh_k;
+	vblank_timer->period = DIV64_U64_ROUND_CLOSEST(NSEC_PER_SEC * MSEC_PER_SEC, vblank_timer->vrefresh_k);
+
+	emulate_monitor_status_change(vgpu);
+
+	return 0;
+}
+
+/**
+ * vgpu_update_vblank_emulation - Update per-vGPU vblank_timer
+ * @vgpu: vGPU operated
+ * @turnon: Turn ON/OFF vblank_timer
+ *
+ * This function is used to turn on/off or update the per-vGPU vblank_timer
+ * when PIPECONF is enabled or disabled. vblank_timer period is also updated
+ * if guest changed the refresh rate.
+ *
+ */
+void vgpu_update_vblank_emulation(struct intel_vgpu *vgpu, bool turnon)
+{
+	struct intel_vgpu_vblank_timer *vblank_timer = &vgpu->vblank_timer;
+	struct intel_vgpu_port *port =
+		intel_vgpu_port(vgpu, vgpu->display.port_num);
+
+	if (turnon) {
+		/*
+		 * Skip the re-enable if already active and vrefresh unchanged.
+		 * Otherwise, stop timer if already active and restart with new
+		 *   period.
+		 */
+		if (vblank_timer->vrefresh_k != port->vrefresh_k ||
+		    !hrtimer_active(&vblank_timer->timer)) {
+			/* Stop timer before start with new period if active */
+			if (hrtimer_active(&vblank_timer->timer))
+				hrtimer_cancel(&vblank_timer->timer);
+
+			/* Make sure new refresh rate updated to timer period */
+			vblank_timer->vrefresh_k = port->vrefresh_k;
+			vblank_timer->period = DIV64_U64_ROUND_CLOSEST(NSEC_PER_SEC * MSEC_PER_SEC, vblank_timer->vrefresh_k);
+			hrtimer_start(&vblank_timer->timer,
+				      ktime_add_ns(ktime_get(), vblank_timer->period),
+				      HRTIMER_MODE_ABS);
+		}
+	} else {
+		/* Caller request to stop vblank */
+		hrtimer_cancel(&vblank_timer->timer);
+	}
+}
+
+static void emulate_vblank_on_pipe(struct intel_vgpu *vgpu, int pipe)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+	struct intel_vgpu_irq *irq = &vgpu->irq;
+	int vblank_event[] = {
+		[PIPE_A] = PIPE_A_VBLANK,
+		[PIPE_B] = PIPE_B_VBLANK,
+		[PIPE_C] = PIPE_C_VBLANK,
+	};
+	int event;
+
+	if (pipe < PIPE_A || pipe > PIPE_C)
+		return;
+
+	for_each_set_bit(event, irq->flip_done_event[pipe],
+			INTEL_GVT_EVENT_MAX) {
+		clear_bit(event, irq->flip_done_event[pipe]);
+		if (!pipe_is_enabled(vgpu, pipe))
+			continue;
+
+		intel_vgpu_trigger_virtual_event(vgpu, event);
+	}
+
+	if (pipe_is_enabled(vgpu, pipe)) {
+		vgpu_vreg_t(vgpu, PIPE_FRMCOUNT_G4X(pipe))++;
+		intel_vgpu_trigger_virtual_event(vgpu, vblank_event[pipe]);
+	}
+}
+
+void intel_vgpu_emulate_vblank(struct intel_vgpu *vgpu)
+{
+	int pipe;
+
+	mutex_lock(&vgpu->vgpu_lock);
+	for_each_pipe(vgpu->gvt->gt->i915, pipe)
+		emulate_vblank_on_pipe(vgpu, pipe);
+	mutex_unlock(&vgpu->vgpu_lock);
+}
+
+/**
+ * intel_vgpu_emulate_hotplug - trigger hotplug event for vGPU
+ * @vgpu: a vGPU
+ * @connected: link state
+ *
+ * This function is used to trigger hotplug interrupt for vGPU
+ *
+ */
+void intel_vgpu_emulate_hotplug(struct intel_vgpu *vgpu, bool connected)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+
+	/* TODO: add more platforms support */
+	if (IS_SKYLAKE(i915) ||
+	    IS_KABYLAKE(i915) ||
+	    IS_COFFEELAKE(i915) ||
+	    IS_COMETLAKE(i915)) {
+		if (connected) {
+			vgpu_vreg_t(vgpu, SFUSE_STRAP) |=
+				SFUSE_STRAP_DDID_DETECTED;
+			vgpu_vreg_t(vgpu, SDEISR) |= SDE_PORTD_HOTPLUG_CPT;
+		} else {
+			vgpu_vreg_t(vgpu, SFUSE_STRAP) &=
+				~SFUSE_STRAP_DDID_DETECTED;
+			vgpu_vreg_t(vgpu, SDEISR) &= ~SDE_PORTD_HOTPLUG_CPT;
+		}
+		vgpu_vreg_t(vgpu, SDEIIR) |= SDE_PORTD_HOTPLUG_CPT;
+		vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTD_HOTPLUG_STATUS_MASK;
+		intel_vgpu_trigger_virtual_event(vgpu, DP_D_HOTPLUG);
+	} else if (IS_BROXTON(i915)) {
+		if (intel_vgpu_has_monitor_on_port(vgpu, PORT_A)) {
+			if (connected) {
+				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
+					GEN8_DE_PORT_HOTPLUG(HPD_PORT_A);
+			} else {
+				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) &=
+					~GEN8_DE_PORT_HOTPLUG(HPD_PORT_A);
+			}
+			vgpu_vreg_t(vgpu, GEN8_DE_PORT_IIR) |=
+				GEN8_DE_PORT_HOTPLUG(HPD_PORT_A);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+				~PORTA_HOTPLUG_STATUS_MASK;
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTA_HOTPLUG_LONG_DETECT;
+			intel_vgpu_trigger_virtual_event(vgpu, DP_A_HOTPLUG);
+		}
+		if (intel_vgpu_has_monitor_on_port(vgpu, PORT_B)) {
+			if (connected) {
+				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
+					GEN8_DE_PORT_HOTPLUG(HPD_PORT_B);
+				vgpu_vreg_t(vgpu, SFUSE_STRAP) |=
+					SFUSE_STRAP_DDIB_DETECTED;
+			} else {
+				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) &=
+					~GEN8_DE_PORT_HOTPLUG(HPD_PORT_B);
+				vgpu_vreg_t(vgpu, SFUSE_STRAP) &=
+					~SFUSE_STRAP_DDIB_DETECTED;
+			}
+			vgpu_vreg_t(vgpu, GEN8_DE_PORT_IIR) |=
+				GEN8_DE_PORT_HOTPLUG(HPD_PORT_B);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+				~PORTB_HOTPLUG_STATUS_MASK;
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTB_HOTPLUG_LONG_DETECT;
+			intel_vgpu_trigger_virtual_event(vgpu, DP_B_HOTPLUG);
+		}
+		if (intel_vgpu_has_monitor_on_port(vgpu, PORT_C)) {
+			if (connected) {
+				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) |=
+					GEN8_DE_PORT_HOTPLUG(HPD_PORT_C);
+				vgpu_vreg_t(vgpu, SFUSE_STRAP) |=
+					SFUSE_STRAP_DDIC_DETECTED;
+			} else {
+				vgpu_vreg_t(vgpu, GEN8_DE_PORT_ISR) &=
+					~GEN8_DE_PORT_HOTPLUG(HPD_PORT_C);
+				vgpu_vreg_t(vgpu, SFUSE_STRAP) &=
+					~SFUSE_STRAP_DDIC_DETECTED;
+			}
+			vgpu_vreg_t(vgpu, GEN8_DE_PORT_IIR) |=
+				GEN8_DE_PORT_HOTPLUG(HPD_PORT_C);
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) &=
+				~PORTC_HOTPLUG_STATUS_MASK;
+			vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |=
+				PORTC_HOTPLUG_LONG_DETECT;
+			intel_vgpu_trigger_virtual_event(vgpu, DP_C_HOTPLUG);
+		}
+	}
+}
+
+/**
+ * intel_vgpu_clean_display - clean vGPU virtual display emulation
+ * @vgpu: a vGPU
+ *
+ * This function is used to clean vGPU virtual display emulation stuffs
+ *
+ */
+void intel_vgpu_clean_display(struct intel_vgpu *vgpu)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+
+	if (IS_SKYLAKE(dev_priv) ||
+	    IS_KABYLAKE(dev_priv) ||
+	    IS_COFFEELAKE(dev_priv) ||
+	    IS_COMETLAKE(dev_priv))
+		clean_virtual_dp_monitor(vgpu, PORT_D);
+	else
+		clean_virtual_dp_monitor(vgpu, PORT_B);
+
+	vgpu_update_vblank_emulation(vgpu, false);
+}
+
+/**
+ * intel_vgpu_init_display- initialize vGPU virtual display emulation
+ * @vgpu: a vGPU
+ * @resolution: resolution index for intel_vgpu_edid
+ *
+ * This function is used to initialize vGPU virtual display emulation stuffs
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+int intel_vgpu_init_display(struct intel_vgpu *vgpu, u64 resolution)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+
+	intel_vgpu_init_i2c_edid(vgpu);
+
+	if (IS_SKYLAKE(dev_priv) ||
+	    IS_KABYLAKE(dev_priv) ||
+	    IS_COFFEELAKE(dev_priv) ||
+	    IS_COMETLAKE(dev_priv))
+		return setup_virtual_dp_monitor(vgpu, PORT_D, GVT_DP_D,
+						resolution);
+	else
+		return setup_virtual_dp_monitor(vgpu, PORT_B, GVT_DP_B,
+						resolution);
+}
+
+/**
+ * intel_vgpu_reset_display- reset vGPU virtual display emulation
+ * @vgpu: a vGPU
+ *
+ * This function is used to reset vGPU virtual display emulation stuffs
+ *
+ */
+void intel_vgpu_reset_display(struct intel_vgpu *vgpu)
+{
+	emulate_monitor_status_change(vgpu);
+}
diff --git a/drivers/gpu/drm/i915/gvt/display.h b/drivers/gpu/drm/i915/gvt/display.h
new file mode 100644
index 000000000..f5616f99e
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/display.h
@@ -0,0 +1,224 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Ke Yu
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *
+ * Contributors:
+ *    Terrence Xu <terrence.xu@intel.com>
+ *    Changbin Du <changbin.du@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#ifndef _GVT_DISPLAY_H_
+#define _GVT_DISPLAY_H_
+
+#include <linux/types.h>
+#include <linux/hrtimer.h>
+
+struct intel_gvt;
+struct intel_vgpu;
+
+#define SBI_REG_MAX	20
+#define DPCD_SIZE	0x700
+
+#define intel_vgpu_port(vgpu, port) \
+	(&(vgpu->display.ports[port]))
+
+#define intel_vgpu_has_monitor_on_port(vgpu, port) \
+	(intel_vgpu_port(vgpu, port)->edid && \
+		intel_vgpu_port(vgpu, port)->edid->data_valid)
+
+#define intel_vgpu_port_is_dp(vgpu, port) \
+	((intel_vgpu_port(vgpu, port)->type == GVT_DP_A) || \
+	(intel_vgpu_port(vgpu, port)->type == GVT_DP_B) || \
+	(intel_vgpu_port(vgpu, port)->type == GVT_DP_C) || \
+	(intel_vgpu_port(vgpu, port)->type == GVT_DP_D))
+
+#define INTEL_GVT_MAX_UEVENT_VARS	3
+
+/* DPCD start */
+#define DPCD_SIZE	0x700
+
+/* DPCD */
+#define DP_SET_POWER            0x600
+#define DP_SET_POWER_D0         0x1
+#define AUX_NATIVE_WRITE        0x8
+#define AUX_NATIVE_READ         0x9
+
+#define AUX_NATIVE_REPLY_MASK   (0x3 << 4)
+#define AUX_NATIVE_REPLY_ACK    (0x0 << 4)
+#define AUX_NATIVE_REPLY_NAK    (0x1 << 4)
+#define AUX_NATIVE_REPLY_DEFER  (0x2 << 4)
+
+#define AUX_BURST_SIZE          20
+
+/* DPCD addresses */
+#define DPCD_REV			0x000
+#define DPCD_MAX_LINK_RATE		0x001
+#define DPCD_MAX_LANE_COUNT		0x002
+
+#define DPCD_TRAINING_PATTERN_SET	0x102
+#define	DPCD_SINK_COUNT			0x200
+#define DPCD_LANE0_1_STATUS		0x202
+#define DPCD_LANE2_3_STATUS		0x203
+#define DPCD_LANE_ALIGN_STATUS_UPDATED	0x204
+#define DPCD_SINK_STATUS		0x205
+
+/* link training */
+#define DPCD_TRAINING_PATTERN_SET_MASK	0x03
+#define DPCD_LINK_TRAINING_DISABLED	0x00
+#define DPCD_TRAINING_PATTERN_1		0x01
+#define DPCD_TRAINING_PATTERN_2		0x02
+
+#define DPCD_CP_READY_MASK		(1 << 6)
+
+/* lane status */
+#define DPCD_LANES_CR_DONE		0x11
+#define DPCD_LANES_EQ_DONE		0x22
+#define DPCD_SYMBOL_LOCKED		0x44
+
+#define DPCD_INTERLANE_ALIGN_DONE	0x01
+
+#define DPCD_SINK_IN_SYNC		0x03
+/* DPCD end */
+
+#define SBI_RESPONSE_MASK               0x3
+#define SBI_RESPONSE_SHIFT              0x1
+#define SBI_STAT_MASK                   0x1
+#define SBI_STAT_SHIFT                  0x0
+#define SBI_OPCODE_SHIFT                8
+#define SBI_OPCODE_MASK			(0xff << SBI_OPCODE_SHIFT)
+#define SBI_CMD_IORD                    2
+#define SBI_CMD_IOWR                    3
+#define SBI_CMD_CRRD                    6
+#define SBI_CMD_CRWR                    7
+#define SBI_ADDR_OFFSET_SHIFT           16
+#define SBI_ADDR_OFFSET_MASK            (0xffff << SBI_ADDR_OFFSET_SHIFT)
+
+struct intel_vgpu_sbi_register {
+	unsigned int offset;
+	u32 value;
+};
+
+struct intel_vgpu_sbi {
+	int number;
+	struct intel_vgpu_sbi_register registers[SBI_REG_MAX];
+};
+
+enum intel_gvt_plane_type {
+	PRIMARY_PLANE = 0,
+	CURSOR_PLANE,
+	SPRITE_PLANE,
+	MAX_PLANE
+};
+
+struct intel_vgpu_dpcd_data {
+	bool data_valid;
+	u8 data[DPCD_SIZE];
+};
+
+enum intel_vgpu_port_type {
+	GVT_CRT = 0,
+	GVT_DP_A,
+	GVT_DP_B,
+	GVT_DP_C,
+	GVT_DP_D,
+	GVT_HDMI_B,
+	GVT_HDMI_C,
+	GVT_HDMI_D,
+	GVT_PORT_MAX
+};
+
+enum intel_vgpu_edid {
+	GVT_EDID_1024_768,
+	GVT_EDID_1920_1200,
+	GVT_EDID_NUM,
+};
+
+#define GVT_DEFAULT_REFRESH_RATE 60
+struct intel_vgpu_port {
+	/* per display EDID information */
+	struct intel_vgpu_edid_data *edid;
+	/* per display DPCD information */
+	struct intel_vgpu_dpcd_data *dpcd;
+	int type;
+	enum intel_vgpu_edid id;
+	/* x1000 to get accurate 59.94, 24.976, 29.94, etc. in timing std. */
+	u32 vrefresh_k;
+};
+
+struct intel_vgpu_vblank_timer {
+	struct hrtimer timer;
+	u32 vrefresh_k;
+	u64 period;
+};
+
+static inline char *vgpu_edid_str(enum intel_vgpu_edid id)
+{
+	switch (id) {
+	case GVT_EDID_1024_768:
+		return "1024x768";
+	case GVT_EDID_1920_1200:
+		return "1920x1200";
+	default:
+		return "";
+	}
+}
+
+static inline unsigned int vgpu_edid_xres(enum intel_vgpu_edid id)
+{
+	switch (id) {
+	case GVT_EDID_1024_768:
+		return 1024;
+	case GVT_EDID_1920_1200:
+		return 1920;
+	default:
+		return 0;
+	}
+}
+
+static inline unsigned int vgpu_edid_yres(enum intel_vgpu_edid id)
+{
+	switch (id) {
+	case GVT_EDID_1024_768:
+		return 768;
+	case GVT_EDID_1920_1200:
+		return 1200;
+	default:
+		return 0;
+	}
+}
+
+void intel_vgpu_emulate_vblank(struct intel_vgpu *vgpu);
+void vgpu_update_vblank_emulation(struct intel_vgpu *vgpu, bool turnon);
+
+int intel_vgpu_init_display(struct intel_vgpu *vgpu, u64 resolution);
+void intel_vgpu_reset_display(struct intel_vgpu *vgpu);
+void intel_vgpu_clean_display(struct intel_vgpu *vgpu);
+
+int pipe_is_enabled(struct intel_vgpu *vgpu, int pipe);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/dmabuf.c b/drivers/gpu/drm/i915/gvt/dmabuf.c
new file mode 100644
index 000000000..01e54b45c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/dmabuf.c
@@ -0,0 +1,589 @@
+/*
+ * Copyright 2017 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *
+ * Contributors:
+ *    Xiaoguang Chen
+ *    Tina Zhang <tina.zhang@intel.com>
+ */
+
+#include <linux/dma-buf.h>
+#include <linux/mdev.h>
+
+#include <drm/drm_fourcc.h>
+#include <drm/drm_plane.h>
+
+#include "gem/i915_gem_dmabuf.h"
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "gvt.h"
+
+#define GEN8_DECODE_PTE(pte) (pte & GENMASK_ULL(63, 12))
+
+static int vgpu_gem_get_pages(
+		struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
+	struct intel_vgpu *vgpu;
+	struct sg_table *st;
+	struct scatterlist *sg;
+	int i, j, ret;
+	gen8_pte_t __iomem *gtt_entries;
+	struct intel_vgpu_fb_info *fb_info;
+	u32 page_num;
+
+	fb_info = (struct intel_vgpu_fb_info *)obj->gvt_info;
+	if (drm_WARN_ON(&dev_priv->drm, !fb_info))
+		return -ENODEV;
+
+	vgpu = fb_info->obj->vgpu;
+	if (drm_WARN_ON(&dev_priv->drm, !vgpu))
+		return -ENODEV;
+
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (unlikely(!st))
+		return -ENOMEM;
+
+	page_num = obj->base.size >> PAGE_SHIFT;
+	ret = sg_alloc_table(st, page_num, GFP_KERNEL);
+	if (ret) {
+		kfree(st);
+		return ret;
+	}
+	gtt_entries = (gen8_pte_t __iomem *)to_gt(dev_priv)->ggtt->gsm +
+		(fb_info->start >> PAGE_SHIFT);
+	for_each_sg(st->sgl, sg, page_num, i) {
+		dma_addr_t dma_addr =
+			GEN8_DECODE_PTE(readq(&gtt_entries[i]));
+		if (intel_gvt_dma_pin_guest_page(vgpu, dma_addr)) {
+			ret = -EINVAL;
+			goto out;
+		}
+
+		sg->offset = 0;
+		sg->length = PAGE_SIZE;
+		sg_dma_len(sg) = PAGE_SIZE;
+		sg_dma_address(sg) = dma_addr;
+	}
+
+	__i915_gem_object_set_pages(obj, st, PAGE_SIZE);
+out:
+	if (ret) {
+		dma_addr_t dma_addr;
+
+		for_each_sg(st->sgl, sg, i, j) {
+			dma_addr = sg_dma_address(sg);
+			if (dma_addr)
+				intel_gvt_dma_unmap_guest_page(vgpu, dma_addr);
+		}
+		sg_free_table(st);
+		kfree(st);
+	}
+
+	return ret;
+
+}
+
+static void vgpu_gem_put_pages(struct drm_i915_gem_object *obj,
+		struct sg_table *pages)
+{
+	struct scatterlist *sg;
+
+	if (obj->base.dma_buf) {
+		struct intel_vgpu_fb_info *fb_info = obj->gvt_info;
+		struct intel_vgpu_dmabuf_obj *obj = fb_info->obj;
+		struct intel_vgpu *vgpu = obj->vgpu;
+		int i;
+
+		for_each_sg(pages->sgl, sg, fb_info->size, i)
+			intel_gvt_dma_unmap_guest_page(vgpu,
+					       sg_dma_address(sg));
+	}
+
+	sg_free_table(pages);
+	kfree(pages);
+}
+
+static void dmabuf_gem_object_free(struct kref *kref)
+{
+	struct intel_vgpu_dmabuf_obj *obj =
+		container_of(kref, struct intel_vgpu_dmabuf_obj, kref);
+	struct intel_vgpu *vgpu = obj->vgpu;
+	struct list_head *pos;
+	struct intel_vgpu_dmabuf_obj *dmabuf_obj;
+
+	if (vgpu && vgpu->active && !list_empty(&vgpu->dmabuf_obj_list_head)) {
+		list_for_each(pos, &vgpu->dmabuf_obj_list_head) {
+			dmabuf_obj = list_entry(pos, struct intel_vgpu_dmabuf_obj, list);
+			if (dmabuf_obj == obj) {
+				list_del(pos);
+				idr_remove(&vgpu->object_idr,
+					   dmabuf_obj->dmabuf_id);
+				kfree(dmabuf_obj->info);
+				kfree(dmabuf_obj);
+				break;
+			}
+		}
+	} else {
+		/* Free the orphan dmabuf_objs here */
+		kfree(obj->info);
+		kfree(obj);
+	}
+}
+
+
+static inline void dmabuf_obj_get(struct intel_vgpu_dmabuf_obj *obj)
+{
+	kref_get(&obj->kref);
+}
+
+static inline void dmabuf_obj_put(struct intel_vgpu_dmabuf_obj *obj)
+{
+	kref_put(&obj->kref, dmabuf_gem_object_free);
+}
+
+static void vgpu_gem_release(struct drm_i915_gem_object *gem_obj)
+{
+
+	struct intel_vgpu_fb_info *fb_info = gem_obj->gvt_info;
+	struct intel_vgpu_dmabuf_obj *obj = fb_info->obj;
+	struct intel_vgpu *vgpu = obj->vgpu;
+
+	if (vgpu) {
+		mutex_lock(&vgpu->dmabuf_lock);
+		gem_obj->base.dma_buf = NULL;
+		dmabuf_obj_put(obj);
+		mutex_unlock(&vgpu->dmabuf_lock);
+	} else {
+		/* vgpu is NULL, as it has been removed already */
+		gem_obj->base.dma_buf = NULL;
+		dmabuf_obj_put(obj);
+	}
+}
+
+static const struct drm_i915_gem_object_ops intel_vgpu_gem_ops = {
+	.name = "i915_gem_object_vgpu",
+	.flags = I915_GEM_OBJECT_IS_PROXY,
+	.get_pages = vgpu_gem_get_pages,
+	.put_pages = vgpu_gem_put_pages,
+	.release = vgpu_gem_release,
+};
+
+static struct drm_i915_gem_object *vgpu_create_gem(struct drm_device *dev,
+		struct intel_vgpu_fb_info *info)
+{
+	static struct lock_class_key lock_class;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_i915_gem_object *obj;
+
+	obj = i915_gem_object_alloc();
+	if (obj == NULL)
+		return NULL;
+
+	drm_gem_private_object_init(dev, &obj->base,
+		roundup(info->size, PAGE_SIZE));
+	i915_gem_object_init(obj, &intel_vgpu_gem_ops, &lock_class, 0);
+	i915_gem_object_set_readonly(obj);
+
+	obj->read_domains = I915_GEM_DOMAIN_GTT;
+	obj->write_domain = 0;
+	if (GRAPHICS_VER(dev_priv) >= 9) {
+		unsigned int tiling_mode = 0;
+		unsigned int stride = 0;
+
+		switch (info->drm_format_mod) {
+		case DRM_FORMAT_MOD_LINEAR:
+			tiling_mode = I915_TILING_NONE;
+			break;
+		case I915_FORMAT_MOD_X_TILED:
+			tiling_mode = I915_TILING_X;
+			stride = info->stride;
+			break;
+		case I915_FORMAT_MOD_Y_TILED:
+		case I915_FORMAT_MOD_Yf_TILED:
+			tiling_mode = I915_TILING_Y;
+			stride = info->stride;
+			break;
+		default:
+			gvt_dbg_core("invalid drm_format_mod %llx for tiling\n",
+				     info->drm_format_mod);
+		}
+		obj->tiling_and_stride = tiling_mode | stride;
+	} else {
+		obj->tiling_and_stride = info->drm_format_mod ?
+					I915_TILING_X : 0;
+	}
+
+	return obj;
+}
+
+static bool validate_hotspot(struct intel_vgpu_cursor_plane_format *c)
+{
+	if (c && c->x_hot <= c->width && c->y_hot <= c->height)
+		return true;
+	else
+		return false;
+}
+
+static int vgpu_get_plane_info(struct drm_device *dev,
+		struct intel_vgpu *vgpu,
+		struct intel_vgpu_fb_info *info,
+		int plane_id)
+{
+	struct intel_vgpu_primary_plane_format p;
+	struct intel_vgpu_cursor_plane_format c;
+	int ret, tile_height = 1;
+
+	memset(info, 0, sizeof(*info));
+
+	if (plane_id == DRM_PLANE_TYPE_PRIMARY) {
+		ret = intel_vgpu_decode_primary_plane(vgpu, &p);
+		if (ret)
+			return ret;
+		info->start = p.base;
+		info->start_gpa = p.base_gpa;
+		info->width = p.width;
+		info->height = p.height;
+		info->stride = p.stride;
+		info->drm_format = p.drm_format;
+
+		switch (p.tiled) {
+		case PLANE_CTL_TILED_LINEAR:
+			info->drm_format_mod = DRM_FORMAT_MOD_LINEAR;
+			break;
+		case PLANE_CTL_TILED_X:
+			info->drm_format_mod = I915_FORMAT_MOD_X_TILED;
+			tile_height = 8;
+			break;
+		case PLANE_CTL_TILED_Y:
+			info->drm_format_mod = I915_FORMAT_MOD_Y_TILED;
+			tile_height = 32;
+			break;
+		case PLANE_CTL_TILED_YF:
+			info->drm_format_mod = I915_FORMAT_MOD_Yf_TILED;
+			tile_height = 32;
+			break;
+		default:
+			gvt_vgpu_err("invalid tiling mode: %x\n", p.tiled);
+		}
+	} else if (plane_id == DRM_PLANE_TYPE_CURSOR) {
+		ret = intel_vgpu_decode_cursor_plane(vgpu, &c);
+		if (ret)
+			return ret;
+		info->start = c.base;
+		info->start_gpa = c.base_gpa;
+		info->width = c.width;
+		info->height = c.height;
+		info->stride = c.width * (c.bpp / 8);
+		info->drm_format = c.drm_format;
+		info->drm_format_mod = 0;
+		info->x_pos = c.x_pos;
+		info->y_pos = c.y_pos;
+
+		if (validate_hotspot(&c)) {
+			info->x_hot = c.x_hot;
+			info->y_hot = c.y_hot;
+		} else {
+			info->x_hot = UINT_MAX;
+			info->y_hot = UINT_MAX;
+		}
+	} else {
+		gvt_vgpu_err("invalid plane id:%d\n", plane_id);
+		return -EINVAL;
+	}
+
+	info->size = info->stride * roundup(info->height, tile_height);
+	if (info->size == 0) {
+		gvt_vgpu_err("fb size is zero\n");
+		return -EINVAL;
+	}
+
+	if (info->start & (PAGE_SIZE - 1)) {
+		gvt_vgpu_err("Not aligned fb address:0x%llx\n", info->start);
+		return -EFAULT;
+	}
+
+	if (!intel_gvt_ggtt_validate_range(vgpu, info->start, info->size)) {
+		gvt_vgpu_err("invalid gma addr\n");
+		return -EFAULT;
+	}
+
+	return 0;
+}
+
+static struct intel_vgpu_dmabuf_obj *
+pick_dmabuf_by_info(struct intel_vgpu *vgpu,
+		    struct intel_vgpu_fb_info *latest_info)
+{
+	struct list_head *pos;
+	struct intel_vgpu_fb_info *fb_info;
+	struct intel_vgpu_dmabuf_obj *dmabuf_obj = NULL;
+	struct intel_vgpu_dmabuf_obj *ret = NULL;
+
+	list_for_each(pos, &vgpu->dmabuf_obj_list_head) {
+		dmabuf_obj = list_entry(pos, struct intel_vgpu_dmabuf_obj, list);
+		if (!dmabuf_obj->info)
+			continue;
+
+		fb_info = (struct intel_vgpu_fb_info *)dmabuf_obj->info;
+		if ((fb_info->start == latest_info->start) &&
+		    (fb_info->start_gpa == latest_info->start_gpa) &&
+		    (fb_info->size == latest_info->size) &&
+		    (fb_info->drm_format_mod == latest_info->drm_format_mod) &&
+		    (fb_info->drm_format == latest_info->drm_format) &&
+		    (fb_info->width == latest_info->width) &&
+		    (fb_info->height == latest_info->height)) {
+			ret = dmabuf_obj;
+			break;
+		}
+	}
+
+	return ret;
+}
+
+static struct intel_vgpu_dmabuf_obj *
+pick_dmabuf_by_num(struct intel_vgpu *vgpu, u32 id)
+{
+	struct list_head *pos;
+	struct intel_vgpu_dmabuf_obj *dmabuf_obj = NULL;
+	struct intel_vgpu_dmabuf_obj *ret = NULL;
+
+	list_for_each(pos, &vgpu->dmabuf_obj_list_head) {
+		dmabuf_obj = list_entry(pos, struct intel_vgpu_dmabuf_obj, list);
+		if (dmabuf_obj->dmabuf_id == id) {
+			ret = dmabuf_obj;
+			break;
+		}
+	}
+
+	return ret;
+}
+
+static void update_fb_info(struct vfio_device_gfx_plane_info *gvt_dmabuf,
+		      struct intel_vgpu_fb_info *fb_info)
+{
+	gvt_dmabuf->drm_format = fb_info->drm_format;
+	gvt_dmabuf->drm_format_mod = fb_info->drm_format_mod;
+	gvt_dmabuf->width = fb_info->width;
+	gvt_dmabuf->height = fb_info->height;
+	gvt_dmabuf->stride = fb_info->stride;
+	gvt_dmabuf->size = fb_info->size;
+	gvt_dmabuf->x_pos = fb_info->x_pos;
+	gvt_dmabuf->y_pos = fb_info->y_pos;
+	gvt_dmabuf->x_hot = fb_info->x_hot;
+	gvt_dmabuf->y_hot = fb_info->y_hot;
+}
+
+int intel_vgpu_query_plane(struct intel_vgpu *vgpu, void *args)
+{
+	struct drm_device *dev = &vgpu->gvt->gt->i915->drm;
+	struct vfio_device_gfx_plane_info *gfx_plane_info = args;
+	struct intel_vgpu_dmabuf_obj *dmabuf_obj;
+	struct intel_vgpu_fb_info fb_info;
+	int ret = 0;
+
+	if (gfx_plane_info->flags == (VFIO_GFX_PLANE_TYPE_DMABUF |
+				       VFIO_GFX_PLANE_TYPE_PROBE))
+		return ret;
+	else if ((gfx_plane_info->flags & ~VFIO_GFX_PLANE_TYPE_DMABUF) ||
+			(!gfx_plane_info->flags))
+		return -EINVAL;
+
+	ret = vgpu_get_plane_info(dev, vgpu, &fb_info,
+					gfx_plane_info->drm_plane_type);
+	if (ret != 0)
+		goto out;
+
+	mutex_lock(&vgpu->dmabuf_lock);
+	/* If exists, pick up the exposed dmabuf_obj */
+	dmabuf_obj = pick_dmabuf_by_info(vgpu, &fb_info);
+	if (dmabuf_obj) {
+		update_fb_info(gfx_plane_info, &fb_info);
+		gfx_plane_info->dmabuf_id = dmabuf_obj->dmabuf_id;
+
+		/* This buffer may be released between query_plane ioctl and
+		 * get_dmabuf ioctl. Add the refcount to make sure it won't
+		 * be released between the two ioctls.
+		 */
+		if (!dmabuf_obj->initref) {
+			dmabuf_obj->initref = true;
+			dmabuf_obj_get(dmabuf_obj);
+		}
+		ret = 0;
+		gvt_dbg_dpy("vgpu%d: re-use dmabuf_obj ref %d, id %d\n",
+			    vgpu->id, kref_read(&dmabuf_obj->kref),
+			    gfx_plane_info->dmabuf_id);
+		mutex_unlock(&vgpu->dmabuf_lock);
+		goto out;
+	}
+
+	mutex_unlock(&vgpu->dmabuf_lock);
+
+	/* Need to allocate a new one*/
+	dmabuf_obj = kmalloc(sizeof(struct intel_vgpu_dmabuf_obj), GFP_KERNEL);
+	if (unlikely(!dmabuf_obj)) {
+		gvt_vgpu_err("alloc dmabuf_obj failed\n");
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	dmabuf_obj->info = kmalloc(sizeof(struct intel_vgpu_fb_info),
+				   GFP_KERNEL);
+	if (unlikely(!dmabuf_obj->info)) {
+		gvt_vgpu_err("allocate intel vgpu fb info failed\n");
+		ret = -ENOMEM;
+		goto out_free_dmabuf;
+	}
+	memcpy(dmabuf_obj->info, &fb_info, sizeof(struct intel_vgpu_fb_info));
+
+	((struct intel_vgpu_fb_info *)dmabuf_obj->info)->obj = dmabuf_obj;
+
+	dmabuf_obj->vgpu = vgpu;
+
+	ret = idr_alloc(&vgpu->object_idr, dmabuf_obj, 1, 0, GFP_NOWAIT);
+	if (ret < 0)
+		goto out_free_info;
+	gfx_plane_info->dmabuf_id = ret;
+	dmabuf_obj->dmabuf_id = ret;
+
+	dmabuf_obj->initref = true;
+
+	kref_init(&dmabuf_obj->kref);
+
+	update_fb_info(gfx_plane_info, &fb_info);
+
+	INIT_LIST_HEAD(&dmabuf_obj->list);
+	mutex_lock(&vgpu->dmabuf_lock);
+	list_add_tail(&dmabuf_obj->list, &vgpu->dmabuf_obj_list_head);
+	mutex_unlock(&vgpu->dmabuf_lock);
+
+	gvt_dbg_dpy("vgpu%d: %s new dmabuf_obj ref %d, id %d\n", vgpu->id,
+		    __func__, kref_read(&dmabuf_obj->kref), ret);
+
+	return 0;
+
+out_free_info:
+	kfree(dmabuf_obj->info);
+out_free_dmabuf:
+	kfree(dmabuf_obj);
+out:
+	/* ENODEV means plane isn't ready, which might be a normal case. */
+	return (ret == -ENODEV) ? 0 : ret;
+}
+
+/* To associate an exposed dmabuf with the dmabuf_obj */
+int intel_vgpu_get_dmabuf(struct intel_vgpu *vgpu, unsigned int dmabuf_id)
+{
+	struct drm_device *dev = &vgpu->gvt->gt->i915->drm;
+	struct intel_vgpu_dmabuf_obj *dmabuf_obj;
+	struct drm_i915_gem_object *obj;
+	struct dma_buf *dmabuf;
+	int dmabuf_fd;
+	int ret = 0;
+
+	mutex_lock(&vgpu->dmabuf_lock);
+
+	dmabuf_obj = pick_dmabuf_by_num(vgpu, dmabuf_id);
+	if (dmabuf_obj == NULL) {
+		gvt_vgpu_err("invalid dmabuf id:%d\n", dmabuf_id);
+		ret = -EINVAL;
+		goto out;
+	}
+
+	obj = vgpu_create_gem(dev, dmabuf_obj->info);
+	if (obj == NULL) {
+		gvt_vgpu_err("create gvt gem obj failed\n");
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	obj->gvt_info = dmabuf_obj->info;
+
+	dmabuf = i915_gem_prime_export(&obj->base, DRM_CLOEXEC | DRM_RDWR);
+	if (IS_ERR(dmabuf)) {
+		gvt_vgpu_err("export dma-buf failed\n");
+		ret = PTR_ERR(dmabuf);
+		goto out_free_gem;
+	}
+
+	ret = dma_buf_fd(dmabuf, DRM_CLOEXEC | DRM_RDWR);
+	if (ret < 0) {
+		gvt_vgpu_err("create dma-buf fd failed ret:%d\n", ret);
+		goto out_free_dmabuf;
+	}
+	dmabuf_fd = ret;
+
+	dmabuf_obj_get(dmabuf_obj);
+
+	if (dmabuf_obj->initref) {
+		dmabuf_obj->initref = false;
+		dmabuf_obj_put(dmabuf_obj);
+	}
+
+	mutex_unlock(&vgpu->dmabuf_lock);
+
+	gvt_dbg_dpy("vgpu%d: dmabuf:%d, dmabuf ref %d, fd:%d\n"
+		    "        file count: %ld, GEM ref: %d\n",
+		    vgpu->id, dmabuf_obj->dmabuf_id,
+		    kref_read(&dmabuf_obj->kref),
+		    dmabuf_fd,
+		    file_count(dmabuf->file),
+		    kref_read(&obj->base.refcount));
+
+	i915_gem_object_put(obj);
+
+	return dmabuf_fd;
+
+out_free_dmabuf:
+	dma_buf_put(dmabuf);
+out_free_gem:
+	i915_gem_object_put(obj);
+out:
+	mutex_unlock(&vgpu->dmabuf_lock);
+	return ret;
+}
+
+void intel_vgpu_dmabuf_cleanup(struct intel_vgpu *vgpu)
+{
+	struct list_head *pos, *n;
+	struct intel_vgpu_dmabuf_obj *dmabuf_obj;
+
+	mutex_lock(&vgpu->dmabuf_lock);
+	list_for_each_safe(pos, n, &vgpu->dmabuf_obj_list_head) {
+		dmabuf_obj = list_entry(pos, struct intel_vgpu_dmabuf_obj, list);
+		dmabuf_obj->vgpu = NULL;
+
+		idr_remove(&vgpu->object_idr, dmabuf_obj->dmabuf_id);
+		list_del(pos);
+
+		/* dmabuf_obj might be freed in dmabuf_obj_put */
+		if (dmabuf_obj->initref) {
+			dmabuf_obj->initref = false;
+			dmabuf_obj_put(dmabuf_obj);
+		}
+
+	}
+	mutex_unlock(&vgpu->dmabuf_lock);
+}
diff --git a/drivers/gpu/drm/i915/gvt/dmabuf.h b/drivers/gpu/drm/i915/gvt/dmabuf.h
new file mode 100644
index 000000000..5f8f03fb1
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/dmabuf.h
@@ -0,0 +1,67 @@
+/*
+ * Copyright(c) 2017 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *
+ * Contributors:
+ *    Xiaoguang Chen
+ *    Tina Zhang <tina.zhang@intel.com>
+ */
+
+#ifndef _GVT_DMABUF_H_
+#define _GVT_DMABUF_H_
+#include <linux/vfio.h>
+
+struct intel_vgpu_fb_info {
+	__u64 start;
+	__u64 start_gpa;
+	__u64 drm_format_mod;
+	__u32 drm_format;	/* drm format of plane */
+	__u32 width;	/* width of plane */
+	__u32 height;	/* height of plane */
+	__u32 stride;	/* stride of plane */
+	__u32 size;	/* size of plane in bytes, align on page */
+	__u32 x_pos;	/* horizontal position of cursor plane */
+	__u32 y_pos;	/* vertical position of cursor plane */
+	__u32 x_hot;    /* horizontal position of cursor hotspot */
+	__u32 y_hot;    /* vertical position of cursor hotspot */
+	struct intel_vgpu_dmabuf_obj *obj;
+};
+
+/**
+ * struct intel_vgpu_dmabuf_obj- Intel vGPU device buffer object
+ */
+struct intel_vgpu_dmabuf_obj {
+	struct intel_vgpu *vgpu;
+	struct intel_vgpu_fb_info *info;
+	__u32 dmabuf_id;
+	struct kref kref;
+	bool initref;
+	struct list_head list;
+};
+
+int intel_vgpu_query_plane(struct intel_vgpu *vgpu, void *args);
+int intel_vgpu_get_dmabuf(struct intel_vgpu *vgpu, unsigned int dmabuf_id);
+void intel_vgpu_dmabuf_cleanup(struct intel_vgpu *vgpu);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/edid.c b/drivers/gpu/drm/i915/gvt/edid.c
new file mode 100644
index 000000000..1b509c1a1
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/edid.c
@@ -0,0 +1,585 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Ke Yu
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *
+ * Contributors:
+ *    Terrence Xu <terrence.xu@intel.com>
+ *    Changbin Du <changbin.du@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#include "display/intel_gmbus_regs.h"
+#include "gvt.h"
+#include "i915_drv.h"
+#include "i915_reg.h"
+
+#define GMBUS1_TOTAL_BYTES_SHIFT 16
+#define GMBUS1_TOTAL_BYTES_MASK 0x1ff
+#define gmbus1_total_byte_count(v) (((v) >> \
+	GMBUS1_TOTAL_BYTES_SHIFT) & GMBUS1_TOTAL_BYTES_MASK)
+#define gmbus1_slave_addr(v) (((v) & 0xff) >> 1)
+#define gmbus1_slave_index(v) (((v) >> 8) & 0xff)
+#define gmbus1_bus_cycle(v) (((v) >> 25) & 0x7)
+
+/* GMBUS0 bits definitions */
+#define _GMBUS_PIN_SEL_MASK     (0x7)
+
+static unsigned char edid_get_byte(struct intel_vgpu *vgpu)
+{
+	struct intel_vgpu_i2c_edid *edid = &vgpu->display.i2c_edid;
+	unsigned char chr = 0;
+
+	if (edid->state == I2C_NOT_SPECIFIED || !edid->slave_selected) {
+		gvt_vgpu_err("Driver tries to read EDID without proper sequence!\n");
+		return 0;
+	}
+	if (edid->current_edid_read >= EDID_SIZE) {
+		gvt_vgpu_err("edid_get_byte() exceeds the size of EDID!\n");
+		return 0;
+	}
+
+	if (!edid->edid_available) {
+		gvt_vgpu_err("Reading EDID but EDID is not available!\n");
+		return 0;
+	}
+
+	if (intel_vgpu_has_monitor_on_port(vgpu, edid->port)) {
+		struct intel_vgpu_edid_data *edid_data =
+			intel_vgpu_port(vgpu, edid->port)->edid;
+
+		chr = edid_data->edid_block[edid->current_edid_read];
+		edid->current_edid_read++;
+	} else {
+		gvt_vgpu_err("No EDID available during the reading?\n");
+	}
+	return chr;
+}
+
+static inline int cnp_get_port_from_gmbus0(u32 gmbus0)
+{
+	int port_select = gmbus0 & _GMBUS_PIN_SEL_MASK;
+	int port = -EINVAL;
+
+	if (port_select == GMBUS_PIN_1_BXT)
+		port = PORT_B;
+	else if (port_select == GMBUS_PIN_2_BXT)
+		port = PORT_C;
+	else if (port_select == GMBUS_PIN_3_BXT)
+		port = PORT_D;
+	else if (port_select == GMBUS_PIN_4_CNP)
+		port = PORT_E;
+	return port;
+}
+
+static inline int bxt_get_port_from_gmbus0(u32 gmbus0)
+{
+	int port_select = gmbus0 & _GMBUS_PIN_SEL_MASK;
+	int port = -EINVAL;
+
+	if (port_select == GMBUS_PIN_1_BXT)
+		port = PORT_B;
+	else if (port_select == GMBUS_PIN_2_BXT)
+		port = PORT_C;
+	else if (port_select == GMBUS_PIN_3_BXT)
+		port = PORT_D;
+	return port;
+}
+
+static inline int get_port_from_gmbus0(u32 gmbus0)
+{
+	int port_select = gmbus0 & _GMBUS_PIN_SEL_MASK;
+	int port = -EINVAL;
+
+	if (port_select == GMBUS_PIN_VGADDC)
+		port = PORT_E;
+	else if (port_select == GMBUS_PIN_DPC)
+		port = PORT_C;
+	else if (port_select == GMBUS_PIN_DPB)
+		port = PORT_B;
+	else if (port_select == GMBUS_PIN_DPD)
+		port = PORT_D;
+	return port;
+}
+
+static void reset_gmbus_controller(struct intel_vgpu *vgpu)
+{
+	vgpu_vreg_t(vgpu, PCH_GMBUS2) = GMBUS_HW_RDY;
+	if (!vgpu->display.i2c_edid.edid_available)
+		vgpu_vreg_t(vgpu, PCH_GMBUS2) |= GMBUS_SATOER;
+	vgpu->display.i2c_edid.gmbus.phase = GMBUS_IDLE_PHASE;
+}
+
+/* GMBUS0 */
+static int gmbus0_mmio_write(struct intel_vgpu *vgpu,
+			unsigned int offset, void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	int port, pin_select;
+
+	memcpy(&vgpu_vreg(vgpu, offset), p_data, bytes);
+
+	pin_select = vgpu_vreg(vgpu, offset) & _GMBUS_PIN_SEL_MASK;
+
+	intel_vgpu_init_i2c_edid(vgpu);
+
+	if (pin_select == 0)
+		return 0;
+
+	if (IS_BROXTON(i915))
+		port = bxt_get_port_from_gmbus0(pin_select);
+	else if (IS_COFFEELAKE(i915) || IS_COMETLAKE(i915))
+		port = cnp_get_port_from_gmbus0(pin_select);
+	else
+		port = get_port_from_gmbus0(pin_select);
+	if (drm_WARN_ON(&i915->drm, port < 0))
+		return 0;
+
+	vgpu->display.i2c_edid.state = I2C_GMBUS;
+	vgpu->display.i2c_edid.gmbus.phase = GMBUS_IDLE_PHASE;
+
+	vgpu_vreg_t(vgpu, PCH_GMBUS2) &= ~GMBUS_ACTIVE;
+	vgpu_vreg_t(vgpu, PCH_GMBUS2) |= GMBUS_HW_RDY | GMBUS_HW_WAIT_PHASE;
+
+	if (intel_vgpu_has_monitor_on_port(vgpu, port) &&
+			!intel_vgpu_port_is_dp(vgpu, port)) {
+		vgpu->display.i2c_edid.port = port;
+		vgpu->display.i2c_edid.edid_available = true;
+		vgpu_vreg_t(vgpu, PCH_GMBUS2) &= ~GMBUS_SATOER;
+	} else
+		vgpu_vreg_t(vgpu, PCH_GMBUS2) |= GMBUS_SATOER;
+	return 0;
+}
+
+static int gmbus1_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	struct intel_vgpu_i2c_edid *i2c_edid = &vgpu->display.i2c_edid;
+	u32 slave_addr;
+	u32 wvalue = *(u32 *)p_data;
+
+	if (vgpu_vreg(vgpu, offset) & GMBUS_SW_CLR_INT) {
+		if (!(wvalue & GMBUS_SW_CLR_INT)) {
+			vgpu_vreg(vgpu, offset) &= ~GMBUS_SW_CLR_INT;
+			reset_gmbus_controller(vgpu);
+		}
+		/*
+		 * TODO: "This bit is cleared to zero when an event
+		 * causes the HW_RDY bit transition to occur "
+		 */
+	} else {
+		/*
+		 * per bspec setting this bit can cause:
+		 * 1) INT status bit cleared
+		 * 2) HW_RDY bit asserted
+		 */
+		if (wvalue & GMBUS_SW_CLR_INT) {
+			vgpu_vreg_t(vgpu, PCH_GMBUS2) &= ~GMBUS_INT;
+			vgpu_vreg_t(vgpu, PCH_GMBUS2) |= GMBUS_HW_RDY;
+		}
+
+		/* For virtualization, we suppose that HW is always ready,
+		 * so GMBUS_SW_RDY should always be cleared
+		 */
+		if (wvalue & GMBUS_SW_RDY)
+			wvalue &= ~GMBUS_SW_RDY;
+
+		i2c_edid->gmbus.total_byte_count =
+			gmbus1_total_byte_count(wvalue);
+		slave_addr = gmbus1_slave_addr(wvalue);
+
+		/* vgpu gmbus only support EDID */
+		if (slave_addr == EDID_ADDR) {
+			i2c_edid->slave_selected = true;
+		} else if (slave_addr != 0) {
+			gvt_dbg_dpy(
+				"vgpu%d: unsupported gmbus slave addr(0x%x)\n"
+				"	gmbus operations will be ignored.\n",
+					vgpu->id, slave_addr);
+		}
+
+		if (wvalue & GMBUS_CYCLE_INDEX)
+			i2c_edid->current_edid_read =
+				gmbus1_slave_index(wvalue);
+
+		i2c_edid->gmbus.cycle_type = gmbus1_bus_cycle(wvalue);
+		switch (gmbus1_bus_cycle(wvalue)) {
+		case GMBUS_NOCYCLE:
+			break;
+		case GMBUS_STOP:
+			/* From spec:
+			 * This can only cause a STOP to be generated
+			 * if a GMBUS cycle is generated, the GMBUS is
+			 * currently in a data/wait/idle phase, or it is in a
+			 * WAIT phase
+			 */
+			if (gmbus1_bus_cycle(vgpu_vreg(vgpu, offset))
+				!= GMBUS_NOCYCLE) {
+				intel_vgpu_init_i2c_edid(vgpu);
+				/* After the 'stop' cycle, hw state would become
+				 * 'stop phase' and then 'idle phase' after a
+				 * few milliseconds. In emulation, we just set
+				 * it as 'idle phase' ('stop phase' is not
+				 * visible in gmbus interface)
+				 */
+				i2c_edid->gmbus.phase = GMBUS_IDLE_PHASE;
+				vgpu_vreg_t(vgpu, PCH_GMBUS2) &= ~GMBUS_ACTIVE;
+			}
+			break;
+		case NIDX_NS_W:
+		case IDX_NS_W:
+		case NIDX_STOP:
+		case IDX_STOP:
+			/* From hw spec the GMBUS phase
+			 * transition like this:
+			 * START (-->INDEX) -->DATA
+			 */
+			i2c_edid->gmbus.phase = GMBUS_DATA_PHASE;
+			vgpu_vreg_t(vgpu, PCH_GMBUS2) |= GMBUS_ACTIVE;
+			break;
+		default:
+			gvt_vgpu_err("Unknown/reserved GMBUS cycle detected!\n");
+			break;
+		}
+		/*
+		 * From hw spec the WAIT state will be
+		 * cleared:
+		 * (1) in a new GMBUS cycle
+		 * (2) by generating a stop
+		 */
+		vgpu_vreg(vgpu, offset) = wvalue;
+	}
+	return 0;
+}
+
+static int gmbus3_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+	void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+
+	drm_WARN_ON(&i915->drm, 1);
+	return 0;
+}
+
+static int gmbus3_mmio_read(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	int i;
+	unsigned char byte_data;
+	struct intel_vgpu_i2c_edid *i2c_edid = &vgpu->display.i2c_edid;
+	int byte_left = i2c_edid->gmbus.total_byte_count -
+				i2c_edid->current_edid_read;
+	int byte_count = byte_left;
+	u32 reg_data = 0;
+
+	/* Data can only be recevied if previous settings correct */
+	if (vgpu_vreg_t(vgpu, PCH_GMBUS1) & GMBUS_SLAVE_READ) {
+		if (byte_left <= 0) {
+			memcpy(p_data, &vgpu_vreg(vgpu, offset), bytes);
+			return 0;
+		}
+
+		if (byte_count > 4)
+			byte_count = 4;
+		for (i = 0; i < byte_count; i++) {
+			byte_data = edid_get_byte(vgpu);
+			reg_data |= (byte_data << (i << 3));
+		}
+
+		memcpy(&vgpu_vreg(vgpu, offset), &reg_data, byte_count);
+		memcpy(p_data, &vgpu_vreg(vgpu, offset), bytes);
+
+		if (byte_left <= 4) {
+			switch (i2c_edid->gmbus.cycle_type) {
+			case NIDX_STOP:
+			case IDX_STOP:
+				i2c_edid->gmbus.phase = GMBUS_IDLE_PHASE;
+				break;
+			case NIDX_NS_W:
+			case IDX_NS_W:
+			default:
+				i2c_edid->gmbus.phase = GMBUS_WAIT_PHASE;
+				break;
+			}
+			intel_vgpu_init_i2c_edid(vgpu);
+		}
+		/*
+		 * Read GMBUS3 during send operation,
+		 * return the latest written value
+		 */
+	} else {
+		memcpy(p_data, &vgpu_vreg(vgpu, offset), bytes);
+		gvt_vgpu_err("warning: gmbus3 read with nothing returned\n");
+	}
+	return 0;
+}
+
+static int gmbus2_mmio_read(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 value = vgpu_vreg(vgpu, offset);
+
+	if (!(vgpu_vreg(vgpu, offset) & GMBUS_INUSE))
+		vgpu_vreg(vgpu, offset) |= GMBUS_INUSE;
+	memcpy(p_data, (void *)&value, bytes);
+	return 0;
+}
+
+static int gmbus2_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 wvalue = *(u32 *)p_data;
+
+	if (wvalue & GMBUS_INUSE)
+		vgpu_vreg(vgpu, offset) &= ~GMBUS_INUSE;
+	/* All other bits are read-only */
+	return 0;
+}
+
+/**
+ * intel_gvt_i2c_handle_gmbus_read - emulate gmbus register mmio read
+ * @vgpu: a vGPU
+ * @offset: reg offset
+ * @p_data: data return buffer
+ * @bytes: access data length
+ *
+ * This function is used to emulate gmbus register mmio read
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+int intel_gvt_i2c_handle_gmbus_read(struct intel_vgpu *vgpu,
+	unsigned int offset, void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+
+	if (drm_WARN_ON(&i915->drm, bytes > 8 && (offset & (bytes - 1))))
+		return -EINVAL;
+
+	if (offset == i915_mmio_reg_offset(PCH_GMBUS2))
+		return gmbus2_mmio_read(vgpu, offset, p_data, bytes);
+	else if (offset == i915_mmio_reg_offset(PCH_GMBUS3))
+		return gmbus3_mmio_read(vgpu, offset, p_data, bytes);
+
+	memcpy(p_data, &vgpu_vreg(vgpu, offset), bytes);
+	return 0;
+}
+
+/**
+ * intel_gvt_i2c_handle_gmbus_write - emulate gmbus register mmio write
+ * @vgpu: a vGPU
+ * @offset: reg offset
+ * @p_data: data return buffer
+ * @bytes: access data length
+ *
+ * This function is used to emulate gmbus register mmio write
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+int intel_gvt_i2c_handle_gmbus_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+
+	if (drm_WARN_ON(&i915->drm, bytes > 8 && (offset & (bytes - 1))))
+		return -EINVAL;
+
+	if (offset == i915_mmio_reg_offset(PCH_GMBUS0))
+		return gmbus0_mmio_write(vgpu, offset, p_data, bytes);
+	else if (offset == i915_mmio_reg_offset(PCH_GMBUS1))
+		return gmbus1_mmio_write(vgpu, offset, p_data, bytes);
+	else if (offset == i915_mmio_reg_offset(PCH_GMBUS2))
+		return gmbus2_mmio_write(vgpu, offset, p_data, bytes);
+	else if (offset == i915_mmio_reg_offset(PCH_GMBUS3))
+		return gmbus3_mmio_write(vgpu, offset, p_data, bytes);
+
+	memcpy(&vgpu_vreg(vgpu, offset), p_data, bytes);
+	return 0;
+}
+
+enum {
+	AUX_CH_CTL = 0,
+	AUX_CH_DATA1,
+	AUX_CH_DATA2,
+	AUX_CH_DATA3,
+	AUX_CH_DATA4,
+	AUX_CH_DATA5
+};
+
+static inline int get_aux_ch_reg(unsigned int offset)
+{
+	int reg;
+
+	switch (offset & 0xff) {
+	case 0x10:
+		reg = AUX_CH_CTL;
+		break;
+	case 0x14:
+		reg = AUX_CH_DATA1;
+		break;
+	case 0x18:
+		reg = AUX_CH_DATA2;
+		break;
+	case 0x1c:
+		reg = AUX_CH_DATA3;
+		break;
+	case 0x20:
+		reg = AUX_CH_DATA4;
+		break;
+	case 0x24:
+		reg = AUX_CH_DATA5;
+		break;
+	default:
+		reg = -1;
+		break;
+	}
+	return reg;
+}
+
+#define AUX_CTL_MSG_LENGTH(reg) \
+	((reg & DP_AUX_CH_CTL_MESSAGE_SIZE_MASK) >> \
+		DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT)
+
+/**
+ * intel_gvt_i2c_handle_aux_ch_write - emulate AUX channel register write
+ * @vgpu: a vGPU
+ * @port_idx: port index
+ * @offset: reg offset
+ * @p_data: write ptr
+ *
+ * This function is used to emulate AUX channel register write
+ *
+ */
+void intel_gvt_i2c_handle_aux_ch_write(struct intel_vgpu *vgpu,
+				int port_idx,
+				unsigned int offset,
+				void *p_data)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	struct intel_vgpu_i2c_edid *i2c_edid = &vgpu->display.i2c_edid;
+	int msg_length, ret_msg_size;
+	int msg, addr, ctrl, op;
+	u32 value = *(u32 *)p_data;
+	int aux_data_for_write = 0;
+	int reg = get_aux_ch_reg(offset);
+
+	if (reg != AUX_CH_CTL) {
+		vgpu_vreg(vgpu, offset) = value;
+		return;
+	}
+
+	msg_length = AUX_CTL_MSG_LENGTH(value);
+	// check the msg in DATA register.
+	msg = vgpu_vreg(vgpu, offset + 4);
+	addr = (msg >> 8) & 0xffff;
+	ctrl = (msg >> 24) & 0xff;
+	op = ctrl >> 4;
+	if (!(value & DP_AUX_CH_CTL_SEND_BUSY)) {
+		/* The ctl write to clear some states */
+		return;
+	}
+
+	/* Always set the wanted value for vms. */
+	ret_msg_size = (((op & 0x1) == GVT_AUX_I2C_READ) ? 2 : 1);
+	vgpu_vreg(vgpu, offset) =
+		DP_AUX_CH_CTL_DONE |
+		((ret_msg_size << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) &
+		DP_AUX_CH_CTL_MESSAGE_SIZE_MASK);
+
+	if (msg_length == 3) {
+		if (!(op & GVT_AUX_I2C_MOT)) {
+			/* stop */
+			intel_vgpu_init_i2c_edid(vgpu);
+		} else {
+			/* start or restart */
+			i2c_edid->aux_ch.i2c_over_aux_ch = true;
+			i2c_edid->aux_ch.aux_ch_mot = true;
+			if (addr == 0) {
+				/* reset the address */
+				intel_vgpu_init_i2c_edid(vgpu);
+			} else if (addr == EDID_ADDR) {
+				i2c_edid->state = I2C_AUX_CH;
+				i2c_edid->port = port_idx;
+				i2c_edid->slave_selected = true;
+				if (intel_vgpu_has_monitor_on_port(vgpu,
+					port_idx) &&
+					intel_vgpu_port_is_dp(vgpu, port_idx))
+					i2c_edid->edid_available = true;
+			}
+		}
+	} else if ((op & 0x1) == GVT_AUX_I2C_WRITE) {
+		/* TODO
+		 * We only support EDID reading from I2C_over_AUX. And
+		 * we do not expect the index mode to be used. Right now
+		 * the WRITE operation is ignored. It is good enough to
+		 * support the gfx driver to do EDID access.
+		 */
+	} else {
+		if (drm_WARN_ON(&i915->drm, (op & 0x1) != GVT_AUX_I2C_READ))
+			return;
+		if (drm_WARN_ON(&i915->drm, msg_length != 4))
+			return;
+		if (i2c_edid->edid_available && i2c_edid->slave_selected) {
+			unsigned char val = edid_get_byte(vgpu);
+
+			aux_data_for_write = (val << 16);
+		} else
+			aux_data_for_write = (0xff << 16);
+	}
+	/* write the return value in AUX_CH_DATA reg which includes:
+	 * ACK of I2C_WRITE
+	 * returned byte if it is READ
+	 */
+	aux_data_for_write |= GVT_AUX_I2C_REPLY_ACK << 24;
+	vgpu_vreg(vgpu, offset + 4) = aux_data_for_write;
+}
+
+/**
+ * intel_vgpu_init_i2c_edid - initialize vGPU i2c edid emulation
+ * @vgpu: a vGPU
+ *
+ * This function is used to initialize vGPU i2c edid emulation stuffs
+ *
+ */
+void intel_vgpu_init_i2c_edid(struct intel_vgpu *vgpu)
+{
+	struct intel_vgpu_i2c_edid *edid = &vgpu->display.i2c_edid;
+
+	edid->state = I2C_NOT_SPECIFIED;
+
+	edid->port = -1;
+	edid->slave_selected = false;
+	edid->edid_available = false;
+	edid->current_edid_read = 0;
+
+	memset(&edid->gmbus, 0, sizeof(struct intel_vgpu_i2c_gmbus));
+
+	edid->aux_ch.i2c_over_aux_ch = false;
+	edid->aux_ch.aux_ch_mot = false;
+}
diff --git a/drivers/gpu/drm/i915/gvt/edid.h b/drivers/gpu/drm/i915/gvt/edid.h
new file mode 100644
index 000000000..dfe0cbc6a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/edid.h
@@ -0,0 +1,154 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Ke Yu
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *
+ * Contributors:
+ *    Terrence Xu <terrence.xu@intel.com>
+ *    Changbin Du <changbin.du@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#ifndef _GVT_EDID_H_
+#define _GVT_EDID_H_
+
+#include <linux/types.h>
+
+struct intel_vgpu;
+
+#define EDID_SIZE		128
+#define EDID_ADDR		0x50 /* Linux hvm EDID addr */
+
+#define GVT_AUX_NATIVE_WRITE			0x8
+#define GVT_AUX_NATIVE_READ			0x9
+#define GVT_AUX_I2C_WRITE			0x0
+#define GVT_AUX_I2C_READ			0x1
+#define GVT_AUX_I2C_STATUS			0x2
+#define GVT_AUX_I2C_MOT				0x4
+#define GVT_AUX_I2C_REPLY_ACK			0x0
+
+struct intel_vgpu_edid_data {
+	bool data_valid;
+	unsigned char edid_block[EDID_SIZE];
+};
+
+enum gmbus_cycle_type {
+	GMBUS_NOCYCLE	= 0x0,
+	NIDX_NS_W	= 0x1,
+	IDX_NS_W	= 0x3,
+	GMBUS_STOP	= 0x4,
+	NIDX_STOP	= 0x5,
+	IDX_STOP	= 0x7
+};
+
+/*
+ * States of GMBUS
+ *
+ * GMBUS0-3 could be related to the EDID virtualization. Another two GMBUS
+ * registers, GMBUS4 (interrupt mask) and GMBUS5 (2 byte indes register), are
+ * not considered here. Below describes the usage of GMBUS registers that are
+ * cared by the EDID virtualization
+ *
+ * GMBUS0:
+ *      R/W
+ *      port selection. value of bit0 - bit2 corresponds to the GPIO registers.
+ *
+ * GMBUS1:
+ *      R/W Protect
+ *      Command and Status.
+ *      bit0 is the direction bit: 1 is read; 0 is write.
+ *      bit1 - bit7 is slave 7-bit address.
+ *      bit16 - bit24 total byte count (ignore?)
+ *
+ * GMBUS2:
+ *      Most of bits are read only except bit 15 (IN_USE)
+ *      Status register
+ *      bit0 - bit8 current byte count
+ *      bit 11: hardware ready;
+ *
+ * GMBUS3:
+ *      Read/Write
+ *      Data for transfer
+ */
+
+/* From hw specs, Other phases like START, ADDRESS, INDEX
+ * are invisible to GMBUS MMIO interface. So no definitions
+ * in below enum types
+ */
+enum gvt_gmbus_phase {
+	GMBUS_IDLE_PHASE = 0,
+	GMBUS_DATA_PHASE,
+	GMBUS_WAIT_PHASE,
+	//GMBUS_STOP_PHASE,
+	GMBUS_MAX_PHASE
+};
+
+struct intel_vgpu_i2c_gmbus {
+	unsigned int total_byte_count; /* from GMBUS1 */
+	enum gmbus_cycle_type cycle_type;
+	enum gvt_gmbus_phase phase;
+};
+
+struct intel_vgpu_i2c_aux_ch {
+	bool i2c_over_aux_ch;
+	bool aux_ch_mot;
+};
+
+enum i2c_state {
+	I2C_NOT_SPECIFIED = 0,
+	I2C_GMBUS = 1,
+	I2C_AUX_CH = 2
+};
+
+/* I2C sequences cannot interleave.
+ * GMBUS and AUX_CH sequences cannot interleave.
+ */
+struct intel_vgpu_i2c_edid {
+	enum i2c_state state;
+
+	unsigned int port;
+	bool slave_selected;
+	bool edid_available;
+	unsigned int current_edid_read;
+
+	struct intel_vgpu_i2c_gmbus gmbus;
+	struct intel_vgpu_i2c_aux_ch aux_ch;
+};
+
+void intel_vgpu_init_i2c_edid(struct intel_vgpu *vgpu);
+
+int intel_gvt_i2c_handle_gmbus_read(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes);
+
+int intel_gvt_i2c_handle_gmbus_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes);
+
+void intel_gvt_i2c_handle_aux_ch_write(struct intel_vgpu *vgpu,
+		int port_idx,
+		unsigned int offset,
+		void *p_data);
+
+#endif /*_GVT_EDID_H_*/
diff --git a/drivers/gpu/drm/i915/gvt/execlist.c b/drivers/gpu/drm/i915/gvt/execlist.c
new file mode 100644
index 000000000..274c6ef42
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/execlist.c
@@ -0,0 +1,558 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ * Contributors:
+ *    Min He <min.he@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *    Ping Gao <ping.a.gao@intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *
+ */
+
+#include "i915_drv.h"
+#include "gvt.h"
+
+#define _EL_OFFSET_STATUS       0x234
+#define _EL_OFFSET_STATUS_BUF   0x370
+#define _EL_OFFSET_STATUS_PTR   0x3A0
+
+#define execlist_ring_mmio(e, offset) ((e)->mmio_base + (offset))
+
+#define valid_context(ctx) ((ctx)->valid)
+#define same_context(a, b) (((a)->context_id == (b)->context_id) && \
+		((a)->lrca == (b)->lrca))
+
+static int context_switch_events[] = {
+	[RCS0]  = RCS_AS_CONTEXT_SWITCH,
+	[BCS0]  = BCS_AS_CONTEXT_SWITCH,
+	[VCS0]  = VCS_AS_CONTEXT_SWITCH,
+	[VCS1]  = VCS2_AS_CONTEXT_SWITCH,
+	[VECS0] = VECS_AS_CONTEXT_SWITCH,
+};
+
+static int to_context_switch_event(const struct intel_engine_cs *engine)
+{
+	if (WARN_ON(engine->id >= ARRAY_SIZE(context_switch_events)))
+		return -EINVAL;
+
+	return context_switch_events[engine->id];
+}
+
+static void switch_virtual_execlist_slot(struct intel_vgpu_execlist *execlist)
+{
+	gvt_dbg_el("[before] running slot %d/context %x pending slot %d\n",
+			execlist->running_slot ?
+			execlist->running_slot->index : -1,
+			execlist->running_context ?
+			execlist->running_context->context_id : 0,
+			execlist->pending_slot ?
+			execlist->pending_slot->index : -1);
+
+	execlist->running_slot = execlist->pending_slot;
+	execlist->pending_slot = NULL;
+	execlist->running_context = execlist->running_context ?
+		&execlist->running_slot->ctx[0] : NULL;
+
+	gvt_dbg_el("[after] running slot %d/context %x pending slot %d\n",
+			execlist->running_slot ?
+			execlist->running_slot->index : -1,
+			execlist->running_context ?
+			execlist->running_context->context_id : 0,
+			execlist->pending_slot ?
+			execlist->pending_slot->index : -1);
+}
+
+static void emulate_execlist_status(struct intel_vgpu_execlist *execlist)
+{
+	struct intel_vgpu_execlist_slot *running = execlist->running_slot;
+	struct intel_vgpu_execlist_slot *pending = execlist->pending_slot;
+	struct execlist_ctx_descriptor_format *desc = execlist->running_context;
+	struct intel_vgpu *vgpu = execlist->vgpu;
+	struct execlist_status_format status;
+	u32 status_reg =
+		execlist_ring_mmio(execlist->engine, _EL_OFFSET_STATUS);
+
+	status.ldw = vgpu_vreg(vgpu, status_reg);
+	status.udw = vgpu_vreg(vgpu, status_reg + 4);
+
+	if (running) {
+		status.current_execlist_pointer = !!running->index;
+		status.execlist_write_pointer = !!!running->index;
+		status.execlist_0_active = status.execlist_0_valid =
+			!!!(running->index);
+		status.execlist_1_active = status.execlist_1_valid =
+			!!(running->index);
+	} else {
+		status.context_id = 0;
+		status.execlist_0_active = status.execlist_0_valid = 0;
+		status.execlist_1_active = status.execlist_1_valid = 0;
+	}
+
+	status.context_id = desc ? desc->context_id : 0;
+	status.execlist_queue_full = !!(pending);
+
+	vgpu_vreg(vgpu, status_reg) = status.ldw;
+	vgpu_vreg(vgpu, status_reg + 4) = status.udw;
+
+	gvt_dbg_el("vgpu%d: status reg offset %x ldw %x udw %x\n",
+		vgpu->id, status_reg, status.ldw, status.udw);
+}
+
+static void emulate_csb_update(struct intel_vgpu_execlist *execlist,
+			       struct execlist_context_status_format *status,
+			       bool trigger_interrupt_later)
+{
+	struct intel_vgpu *vgpu = execlist->vgpu;
+	struct execlist_context_status_pointer_format ctx_status_ptr;
+	u32 write_pointer;
+	u32 ctx_status_ptr_reg, ctx_status_buf_reg, offset;
+	unsigned long hwsp_gpa;
+
+	ctx_status_ptr_reg =
+		execlist_ring_mmio(execlist->engine, _EL_OFFSET_STATUS_PTR);
+	ctx_status_buf_reg =
+		execlist_ring_mmio(execlist->engine, _EL_OFFSET_STATUS_BUF);
+
+	ctx_status_ptr.dw = vgpu_vreg(vgpu, ctx_status_ptr_reg);
+
+	write_pointer = ctx_status_ptr.write_ptr;
+
+	if (write_pointer == 0x7)
+		write_pointer = 0;
+	else {
+		++write_pointer;
+		write_pointer %= 0x6;
+	}
+
+	offset = ctx_status_buf_reg + write_pointer * 8;
+
+	vgpu_vreg(vgpu, offset) = status->ldw;
+	vgpu_vreg(vgpu, offset + 4) = status->udw;
+
+	ctx_status_ptr.write_ptr = write_pointer;
+	vgpu_vreg(vgpu, ctx_status_ptr_reg) = ctx_status_ptr.dw;
+
+	/* Update the CSB and CSB write pointer in HWSP */
+	hwsp_gpa = intel_vgpu_gma_to_gpa(vgpu->gtt.ggtt_mm,
+					 vgpu->hws_pga[execlist->engine->id]);
+	if (hwsp_gpa != INTEL_GVT_INVALID_ADDR) {
+		intel_gvt_write_gpa(vgpu,
+			hwsp_gpa + I915_HWS_CSB_BUF0_INDEX * 4 + write_pointer * 8,
+			status, 8);
+		intel_gvt_write_gpa(vgpu,
+			hwsp_gpa + INTEL_HWS_CSB_WRITE_INDEX(execlist->engine->i915) * 4,
+			&write_pointer, 4);
+	}
+
+	gvt_dbg_el("vgpu%d: w pointer %u reg %x csb l %x csb h %x\n",
+		   vgpu->id, write_pointer, offset, status->ldw, status->udw);
+
+	if (trigger_interrupt_later)
+		return;
+
+	intel_vgpu_trigger_virtual_event(vgpu,
+					 to_context_switch_event(execlist->engine));
+}
+
+static int emulate_execlist_ctx_schedule_out(
+		struct intel_vgpu_execlist *execlist,
+		struct execlist_ctx_descriptor_format *ctx)
+{
+	struct intel_vgpu *vgpu = execlist->vgpu;
+	struct intel_vgpu_execlist_slot *running = execlist->running_slot;
+	struct intel_vgpu_execlist_slot *pending = execlist->pending_slot;
+	struct execlist_ctx_descriptor_format *ctx0 = &running->ctx[0];
+	struct execlist_ctx_descriptor_format *ctx1 = &running->ctx[1];
+	struct execlist_context_status_format status;
+
+	memset(&status, 0, sizeof(status));
+
+	gvt_dbg_el("schedule out context id %x\n", ctx->context_id);
+
+	if (WARN_ON(!same_context(ctx, execlist->running_context))) {
+		gvt_vgpu_err("schedule out context is not running context,"
+				"ctx id %x running ctx id %x\n",
+				ctx->context_id,
+				execlist->running_context->context_id);
+		return -EINVAL;
+	}
+
+	/* ctx1 is valid, ctx0/ctx is scheduled-out -> element switch */
+	if (valid_context(ctx1) && same_context(ctx0, ctx)) {
+		gvt_dbg_el("ctx 1 valid, ctx/ctx 0 is scheduled-out\n");
+
+		execlist->running_context = ctx1;
+
+		emulate_execlist_status(execlist);
+
+		status.context_complete = status.element_switch = 1;
+		status.context_id = ctx->context_id;
+
+		emulate_csb_update(execlist, &status, false);
+		/*
+		 * ctx1 is not valid, ctx == ctx0
+		 * ctx1 is valid, ctx1 == ctx
+		 *	--> last element is finished
+		 * emulate:
+		 *	active-to-idle if there is *no* pending execlist
+		 *	context-complete if there *is* pending execlist
+		 */
+	} else if ((!valid_context(ctx1) && same_context(ctx0, ctx))
+			|| (valid_context(ctx1) && same_context(ctx1, ctx))) {
+		gvt_dbg_el("need to switch virtual execlist slot\n");
+
+		switch_virtual_execlist_slot(execlist);
+
+		emulate_execlist_status(execlist);
+
+		status.context_complete = status.active_to_idle = 1;
+		status.context_id = ctx->context_id;
+
+		if (!pending) {
+			emulate_csb_update(execlist, &status, false);
+		} else {
+			emulate_csb_update(execlist, &status, true);
+
+			memset(&status, 0, sizeof(status));
+
+			status.idle_to_active = 1;
+			status.context_id = 0;
+
+			emulate_csb_update(execlist, &status, false);
+		}
+	} else {
+		WARN_ON(1);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static struct intel_vgpu_execlist_slot *get_next_execlist_slot(
+		struct intel_vgpu_execlist *execlist)
+{
+	struct intel_vgpu *vgpu = execlist->vgpu;
+	u32 status_reg =
+		execlist_ring_mmio(execlist->engine, _EL_OFFSET_STATUS);
+	struct execlist_status_format status;
+
+	status.ldw = vgpu_vreg(vgpu, status_reg);
+	status.udw = vgpu_vreg(vgpu, status_reg + 4);
+
+	if (status.execlist_queue_full) {
+		gvt_vgpu_err("virtual execlist slots are full\n");
+		return NULL;
+	}
+
+	return &execlist->slot[status.execlist_write_pointer];
+}
+
+static int emulate_execlist_schedule_in(struct intel_vgpu_execlist *execlist,
+		struct execlist_ctx_descriptor_format ctx[2])
+{
+	struct intel_vgpu_execlist_slot *running = execlist->running_slot;
+	struct intel_vgpu_execlist_slot *slot =
+		get_next_execlist_slot(execlist);
+
+	struct execlist_ctx_descriptor_format *ctx0, *ctx1;
+	struct execlist_context_status_format status;
+	struct intel_vgpu *vgpu = execlist->vgpu;
+
+	gvt_dbg_el("emulate schedule-in\n");
+
+	if (!slot) {
+		gvt_vgpu_err("no available execlist slot\n");
+		return -EINVAL;
+	}
+
+	memset(&status, 0, sizeof(status));
+	memset(slot->ctx, 0, sizeof(slot->ctx));
+
+	slot->ctx[0] = ctx[0];
+	slot->ctx[1] = ctx[1];
+
+	gvt_dbg_el("alloc slot index %d ctx 0 %x ctx 1 %x\n",
+			slot->index, ctx[0].context_id,
+			ctx[1].context_id);
+
+	/*
+	 * no running execlist, make this write bundle as running execlist
+	 * -> idle-to-active
+	 */
+	if (!running) {
+		gvt_dbg_el("no current running execlist\n");
+
+		execlist->running_slot = slot;
+		execlist->pending_slot = NULL;
+		execlist->running_context = &slot->ctx[0];
+
+		gvt_dbg_el("running slot index %d running context %x\n",
+				execlist->running_slot->index,
+				execlist->running_context->context_id);
+
+		emulate_execlist_status(execlist);
+
+		status.idle_to_active = 1;
+		status.context_id = 0;
+
+		emulate_csb_update(execlist, &status, false);
+		return 0;
+	}
+
+	ctx0 = &running->ctx[0];
+	ctx1 = &running->ctx[1];
+
+	gvt_dbg_el("current running slot index %d ctx 0 %x ctx 1 %x\n",
+		running->index, ctx0->context_id, ctx1->context_id);
+
+	/*
+	 * already has an running execlist
+	 *	a. running ctx1 is valid,
+	 *	   ctx0 is finished, and running ctx1 == new execlist ctx[0]
+	 *	b. running ctx1 is not valid,
+	 *	   ctx0 == new execlist ctx[0]
+	 * ----> lite-restore + preempted
+	 */
+	if ((valid_context(ctx1) && same_context(ctx1, &slot->ctx[0]) &&
+		/* condition a */
+		(!same_context(ctx0, execlist->running_context))) ||
+			(!valid_context(ctx1) &&
+			 same_context(ctx0, &slot->ctx[0]))) { /* condition b */
+		gvt_dbg_el("need to switch virtual execlist slot\n");
+
+		execlist->pending_slot = slot;
+		switch_virtual_execlist_slot(execlist);
+
+		emulate_execlist_status(execlist);
+
+		status.lite_restore = status.preempted = 1;
+		status.context_id = ctx[0].context_id;
+
+		emulate_csb_update(execlist, &status, false);
+	} else {
+		gvt_dbg_el("emulate as pending slot\n");
+		/*
+		 * otherwise
+		 * --> emulate pending execlist exist + but no preemption case
+		 */
+		execlist->pending_slot = slot;
+		emulate_execlist_status(execlist);
+	}
+	return 0;
+}
+
+#define get_desc_from_elsp_dwords(ed, i) \
+	((struct execlist_ctx_descriptor_format *)&((ed)->data[i * 2]))
+
+static int prepare_execlist_workload(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct execlist_ctx_descriptor_format ctx[2];
+	int ret;
+
+	if (!workload->emulate_schedule_in)
+		return 0;
+
+	ctx[0] = *get_desc_from_elsp_dwords(&workload->elsp_dwords, 0);
+	ctx[1] = *get_desc_from_elsp_dwords(&workload->elsp_dwords, 1);
+
+	ret = emulate_execlist_schedule_in(&s->execlist[workload->engine->id],
+					   ctx);
+	if (ret) {
+		gvt_vgpu_err("fail to emulate execlist schedule in\n");
+		return ret;
+	}
+	return 0;
+}
+
+static int complete_execlist_workload(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_vgpu_execlist *execlist =
+		&s->execlist[workload->engine->id];
+	struct intel_vgpu_workload *next_workload;
+	struct list_head *next = workload_q_head(vgpu, workload->engine)->next;
+	bool lite_restore = false;
+	int ret = 0;
+
+	gvt_dbg_el("complete workload %p status %d\n",
+		   workload, workload->status);
+
+	if (workload->status || vgpu->resetting_eng & workload->engine->mask)
+		goto out;
+
+	if (!list_empty(workload_q_head(vgpu, workload->engine))) {
+		struct execlist_ctx_descriptor_format *this_desc, *next_desc;
+
+		next_workload = container_of(next,
+				struct intel_vgpu_workload, list);
+		this_desc = &workload->ctx_desc;
+		next_desc = &next_workload->ctx_desc;
+
+		lite_restore = same_context(this_desc, next_desc);
+	}
+
+	if (lite_restore) {
+		gvt_dbg_el("next context == current - no schedule-out\n");
+		goto out;
+	}
+
+	ret = emulate_execlist_ctx_schedule_out(execlist, &workload->ctx_desc);
+out:
+	return ret;
+}
+
+static int submit_context(struct intel_vgpu *vgpu,
+			  const struct intel_engine_cs *engine,
+			  struct execlist_ctx_descriptor_format *desc,
+			  bool emulate_schedule_in)
+{
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_vgpu_workload *workload = NULL;
+
+	workload = intel_vgpu_create_workload(vgpu, engine, desc);
+	if (IS_ERR(workload))
+		return PTR_ERR(workload);
+
+	workload->prepare = prepare_execlist_workload;
+	workload->complete = complete_execlist_workload;
+	workload->emulate_schedule_in = emulate_schedule_in;
+
+	if (emulate_schedule_in)
+		workload->elsp_dwords = s->execlist[engine->id].elsp_dwords;
+
+	gvt_dbg_el("workload %p emulate schedule_in %d\n", workload,
+		   emulate_schedule_in);
+
+	intel_vgpu_queue_workload(workload);
+	return 0;
+}
+
+int intel_vgpu_submit_execlist(struct intel_vgpu *vgpu,
+			       const struct intel_engine_cs *engine)
+{
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_vgpu_execlist *execlist = &s->execlist[engine->id];
+	struct execlist_ctx_descriptor_format *desc[2];
+	int i, ret;
+
+	desc[0] = get_desc_from_elsp_dwords(&execlist->elsp_dwords, 0);
+	desc[1] = get_desc_from_elsp_dwords(&execlist->elsp_dwords, 1);
+
+	if (!desc[0]->valid) {
+		gvt_vgpu_err("invalid elsp submission, desc0 is invalid\n");
+		goto inv_desc;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(desc); i++) {
+		if (!desc[i]->valid)
+			continue;
+		if (!desc[i]->privilege_access) {
+			gvt_vgpu_err("unexpected GGTT elsp submission\n");
+			goto inv_desc;
+		}
+	}
+
+	/* submit workload */
+	for (i = 0; i < ARRAY_SIZE(desc); i++) {
+		if (!desc[i]->valid)
+			continue;
+		ret = submit_context(vgpu, engine, desc[i], i == 0);
+		if (ret) {
+			gvt_vgpu_err("failed to submit desc %d\n", i);
+			return ret;
+		}
+	}
+
+	return 0;
+
+inv_desc:
+	gvt_vgpu_err("descriptors content: desc0 %08x %08x desc1 %08x %08x\n",
+		     desc[0]->udw, desc[0]->ldw, desc[1]->udw, desc[1]->ldw);
+	return -EINVAL;
+}
+
+static void init_vgpu_execlist(struct intel_vgpu *vgpu,
+			       const struct intel_engine_cs *engine)
+{
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_vgpu_execlist *execlist = &s->execlist[engine->id];
+	struct execlist_context_status_pointer_format ctx_status_ptr;
+	u32 ctx_status_ptr_reg;
+
+	memset(execlist, 0, sizeof(*execlist));
+
+	execlist->vgpu = vgpu;
+	execlist->engine = engine;
+	execlist->slot[0].index = 0;
+	execlist->slot[1].index = 1;
+
+	ctx_status_ptr_reg = execlist_ring_mmio(engine, _EL_OFFSET_STATUS_PTR);
+	ctx_status_ptr.dw = vgpu_vreg(vgpu, ctx_status_ptr_reg);
+	ctx_status_ptr.read_ptr = 0;
+	ctx_status_ptr.write_ptr = 0x7;
+	vgpu_vreg(vgpu, ctx_status_ptr_reg) = ctx_status_ptr.dw;
+}
+
+static void clean_execlist(struct intel_vgpu *vgpu,
+			   intel_engine_mask_t engine_mask)
+{
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_engine_cs *engine;
+	intel_engine_mask_t tmp;
+
+	for_each_engine_masked(engine, vgpu->gvt->gt, engine_mask, tmp) {
+		kfree(s->ring_scan_buffer[engine->id]);
+		s->ring_scan_buffer[engine->id] = NULL;
+		s->ring_scan_buffer_size[engine->id] = 0;
+	}
+}
+
+static void reset_execlist(struct intel_vgpu *vgpu,
+			   intel_engine_mask_t engine_mask)
+{
+	struct intel_engine_cs *engine;
+	intel_engine_mask_t tmp;
+
+	for_each_engine_masked(engine, vgpu->gvt->gt, engine_mask, tmp)
+		init_vgpu_execlist(vgpu, engine);
+}
+
+static int init_execlist(struct intel_vgpu *vgpu,
+			 intel_engine_mask_t engine_mask)
+{
+	reset_execlist(vgpu, engine_mask);
+	return 0;
+}
+
+const struct intel_vgpu_submission_ops intel_vgpu_execlist_submission_ops = {
+	.name = "execlist",
+	.init = init_execlist,
+	.reset = reset_execlist,
+	.clean = clean_execlist,
+};
diff --git a/drivers/gpu/drm/i915/gvt/execlist.h b/drivers/gpu/drm/i915/gvt/execlist.h
new file mode 100644
index 000000000..84ad74b37
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/execlist.h
@@ -0,0 +1,185 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ * Contributors:
+ *    Min He <min.he@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *    Ping Gao <ping.a.gao@intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *
+ */
+
+#ifndef _GVT_EXECLIST_H_
+#define _GVT_EXECLIST_H_
+
+#include <linux/types.h>
+
+struct execlist_ctx_descriptor_format {
+	union {
+		u32 ldw;
+		struct {
+			u32 valid                  : 1;
+			u32 force_pd_restore       : 1;
+			u32 force_restore          : 1;
+			u32 addressing_mode        : 2;
+			u32 llc_coherency          : 1;
+			u32 fault_handling         : 2;
+			u32 privilege_access       : 1;
+			u32 reserved               : 3;
+			u32 lrca                   : 20;
+		};
+	};
+	union {
+		u32 udw;
+		u32 context_id;
+	};
+};
+
+struct execlist_status_format {
+	union {
+		u32 ldw;
+		struct {
+			u32 current_execlist_pointer       :1;
+			u32 execlist_write_pointer         :1;
+			u32 execlist_queue_full            :1;
+			u32 execlist_1_valid               :1;
+			u32 execlist_0_valid               :1;
+			u32 last_ctx_switch_reason         :9;
+			u32 current_active_elm_status      :2;
+			u32 arbitration_enable             :1;
+			u32 execlist_1_active              :1;
+			u32 execlist_0_active              :1;
+			u32 reserved                       :13;
+		};
+	};
+	union {
+		u32 udw;
+		u32 context_id;
+	};
+};
+
+struct execlist_context_status_pointer_format {
+	union {
+		u32 dw;
+		struct {
+			u32 write_ptr              :3;
+			u32 reserved               :5;
+			u32 read_ptr               :3;
+			u32 reserved2              :5;
+			u32 mask                   :16;
+		};
+	};
+};
+
+struct execlist_context_status_format {
+	union {
+		u32 ldw;
+		struct {
+			u32 idle_to_active         :1;
+			u32 preempted              :1;
+			u32 element_switch         :1;
+			u32 active_to_idle         :1;
+			u32 context_complete       :1;
+			u32 wait_on_sync_flip      :1;
+			u32 wait_on_vblank         :1;
+			u32 wait_on_semaphore      :1;
+			u32 wait_on_scanline       :1;
+			u32 reserved               :2;
+			u32 semaphore_wait_mode    :1;
+			u32 display_plane          :3;
+			u32 lite_restore           :1;
+			u32 reserved_2             :16;
+		};
+	};
+	union {
+		u32 udw;
+		u32 context_id;
+	};
+};
+
+struct execlist_mmio_pair {
+	u32 addr;
+	u32 val;
+};
+
+/* The first 52 dwords in register state context */
+struct execlist_ring_context {
+	u32 nop1;
+	u32 lri_cmd_1;
+	struct execlist_mmio_pair ctx_ctrl;
+	struct execlist_mmio_pair ring_header;
+	struct execlist_mmio_pair ring_tail;
+	struct execlist_mmio_pair rb_start;
+	struct execlist_mmio_pair rb_ctrl;
+	struct execlist_mmio_pair bb_cur_head_UDW;
+	struct execlist_mmio_pair bb_cur_head_LDW;
+	struct execlist_mmio_pair bb_state;
+	struct execlist_mmio_pair second_bb_addr_UDW;
+	struct execlist_mmio_pair second_bb_addr_LDW;
+	struct execlist_mmio_pair second_bb_state;
+	struct execlist_mmio_pair bb_per_ctx_ptr;
+	struct execlist_mmio_pair rcs_indirect_ctx;
+	struct execlist_mmio_pair rcs_indirect_ctx_offset;
+	u32 nop2;
+	u32 nop3;
+	u32 nop4;
+	u32 lri_cmd_2;
+	struct execlist_mmio_pair ctx_timestamp;
+	/*
+	 * pdps[8]={ pdp3_UDW, pdp3_LDW, pdp2_UDW, pdp2_LDW,
+	 *           pdp1_UDW, pdp1_LDW, pdp0_UDW, pdp0_LDW}
+	 */
+	struct execlist_mmio_pair pdps[8];
+};
+
+struct intel_vgpu_elsp_dwords {
+	u32 data[4];
+	u32 index;
+};
+
+struct intel_vgpu_execlist_slot {
+	struct execlist_ctx_descriptor_format ctx[2];
+	u32 index;
+};
+
+struct intel_vgpu_execlist {
+	struct intel_vgpu_execlist_slot slot[2];
+	struct intel_vgpu_execlist_slot *running_slot;
+	struct intel_vgpu_execlist_slot *pending_slot;
+	struct execlist_ctx_descriptor_format *running_context;
+	struct intel_vgpu *vgpu;
+	struct intel_vgpu_elsp_dwords elsp_dwords;
+	const struct intel_engine_cs *engine;
+};
+
+void intel_vgpu_clean_execlist(struct intel_vgpu *vgpu);
+
+int intel_vgpu_init_execlist(struct intel_vgpu *vgpu);
+
+int intel_vgpu_submit_execlist(struct intel_vgpu *vgpu,
+			       const struct intel_engine_cs *engine);
+
+#endif /*_GVT_EXECLIST_H_*/
diff --git a/drivers/gpu/drm/i915/gvt/fb_decoder.c b/drivers/gpu/drm/i915/gvt/fb_decoder.c
new file mode 100644
index 000000000..835c3fde8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/fb_decoder.c
@@ -0,0 +1,508 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Kevin Tian <kevin.tian@intel.com>
+ *
+ * Contributors:
+ *    Bing Niu <bing.niu@intel.com>
+ *    Xu Han <xu.han@intel.com>
+ *    Ping Gao <ping.a.gao@intel.com>
+ *    Xiaoguang Chen <xiaoguang.chen@intel.com>
+ *    Yang Liu <yang2.liu@intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *
+ */
+
+#include <uapi/drm/drm_fourcc.h>
+#include "i915_drv.h"
+#include "gvt.h"
+#include "i915_pvinfo.h"
+#include "i915_reg.h"
+
+#define PRIMARY_FORMAT_NUM	16
+struct pixel_format {
+	int drm_format;	/* Pixel format in DRM definition */
+	int bpp; /* Bits per pixel, 0 indicates invalid */
+	const char *desc; /* The description */
+};
+
+static const struct pixel_format bdw_pixel_formats[] = {
+	{DRM_FORMAT_C8, 8, "8-bit Indexed"},
+	{DRM_FORMAT_RGB565, 16, "16-bit BGRX (5:6:5 MSB-R:G:B)"},
+	{DRM_FORMAT_XRGB8888, 32, "32-bit BGRX (8:8:8:8 MSB-X:R:G:B)"},
+	{DRM_FORMAT_XBGR2101010, 32, "32-bit RGBX (2:10:10:10 MSB-X:B:G:R)"},
+
+	{DRM_FORMAT_XRGB2101010, 32, "32-bit BGRX (2:10:10:10 MSB-X:R:G:B)"},
+	{DRM_FORMAT_XBGR8888, 32, "32-bit RGBX (8:8:8:8 MSB-X:B:G:R)"},
+
+	/* non-supported format has bpp default to 0 */
+	{0, 0, NULL},
+};
+
+static const struct pixel_format skl_pixel_formats[] = {
+	{DRM_FORMAT_YUYV, 16, "16-bit packed YUYV (8:8:8:8 MSB-V:Y2:U:Y1)"},
+	{DRM_FORMAT_UYVY, 16, "16-bit packed UYVY (8:8:8:8 MSB-Y2:V:Y1:U)"},
+	{DRM_FORMAT_YVYU, 16, "16-bit packed YVYU (8:8:8:8 MSB-U:Y2:V:Y1)"},
+	{DRM_FORMAT_VYUY, 16, "16-bit packed VYUY (8:8:8:8 MSB-Y2:U:Y1:V)"},
+
+	{DRM_FORMAT_C8, 8, "8-bit Indexed"},
+	{DRM_FORMAT_RGB565, 16, "16-bit BGRX (5:6:5 MSB-R:G:B)"},
+	{DRM_FORMAT_ABGR8888, 32, "32-bit RGBA (8:8:8:8 MSB-A:B:G:R)"},
+	{DRM_FORMAT_XBGR8888, 32, "32-bit RGBX (8:8:8:8 MSB-X:B:G:R)"},
+
+	{DRM_FORMAT_ARGB8888, 32, "32-bit BGRA (8:8:8:8 MSB-A:R:G:B)"},
+	{DRM_FORMAT_XRGB8888, 32, "32-bit BGRX (8:8:8:8 MSB-X:R:G:B)"},
+	{DRM_FORMAT_XBGR2101010, 32, "32-bit RGBX (2:10:10:10 MSB-X:B:G:R)"},
+	{DRM_FORMAT_XRGB2101010, 32, "32-bit BGRX (2:10:10:10 MSB-X:R:G:B)"},
+
+	/* non-supported format has bpp default to 0 */
+	{0, 0, NULL},
+};
+
+static int bdw_format_to_drm(int format)
+{
+	int bdw_pixel_formats_index = 6;
+
+	switch (format) {
+	case DISP_FORMAT_8BPP:
+		bdw_pixel_formats_index = 0;
+		break;
+	case DISP_FORMAT_BGRX565:
+		bdw_pixel_formats_index = 1;
+		break;
+	case DISP_FORMAT_BGRX888:
+		bdw_pixel_formats_index = 2;
+		break;
+	case DISP_FORMAT_RGBX101010:
+		bdw_pixel_formats_index = 3;
+		break;
+	case DISP_FORMAT_BGRX101010:
+		bdw_pixel_formats_index = 4;
+		break;
+	case DISP_FORMAT_RGBX888:
+		bdw_pixel_formats_index = 5;
+		break;
+
+	default:
+		break;
+	}
+
+	return bdw_pixel_formats_index;
+}
+
+static int skl_format_to_drm(int format, bool rgb_order, bool alpha,
+	int yuv_order)
+{
+	int skl_pixel_formats_index = 12;
+
+	switch (format) {
+	case PLANE_CTL_FORMAT_INDEXED:
+		skl_pixel_formats_index = 4;
+		break;
+	case PLANE_CTL_FORMAT_RGB_565:
+		skl_pixel_formats_index = 5;
+		break;
+	case PLANE_CTL_FORMAT_XRGB_8888:
+		if (rgb_order)
+			skl_pixel_formats_index = alpha ? 6 : 7;
+		else
+			skl_pixel_formats_index = alpha ? 8 : 9;
+		break;
+	case PLANE_CTL_FORMAT_XRGB_2101010:
+		skl_pixel_formats_index = rgb_order ? 10 : 11;
+		break;
+	case PLANE_CTL_FORMAT_YUV422:
+		skl_pixel_formats_index = yuv_order >> 16;
+		if (skl_pixel_formats_index > 3)
+			return -EINVAL;
+		break;
+
+	default:
+		break;
+	}
+
+	return skl_pixel_formats_index;
+}
+
+static u32 intel_vgpu_get_stride(struct intel_vgpu *vgpu, int pipe,
+	u32 tiled, int stride_mask, int bpp)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+
+	u32 stride_reg = vgpu_vreg_t(vgpu, DSPSTRIDE(pipe)) & stride_mask;
+	u32 stride = stride_reg;
+
+	if (GRAPHICS_VER(dev_priv) >= 9) {
+		switch (tiled) {
+		case PLANE_CTL_TILED_LINEAR:
+			stride = stride_reg * 64;
+			break;
+		case PLANE_CTL_TILED_X:
+			stride = stride_reg * 512;
+			break;
+		case PLANE_CTL_TILED_Y:
+			stride = stride_reg * 128;
+			break;
+		case PLANE_CTL_TILED_YF:
+			if (bpp == 8)
+				stride = stride_reg * 64;
+			else if (bpp == 16 || bpp == 32 || bpp == 64)
+				stride = stride_reg * 128;
+			else
+				gvt_dbg_core("skl: unsupported bpp:%d\n", bpp);
+			break;
+		default:
+			gvt_dbg_core("skl: unsupported tile format:%x\n",
+				tiled);
+		}
+	}
+
+	return stride;
+}
+
+static int get_active_pipe(struct intel_vgpu *vgpu)
+{
+	int i;
+
+	for (i = 0; i < I915_MAX_PIPES; i++)
+		if (pipe_is_enabled(vgpu, i))
+			break;
+
+	return i;
+}
+
+/**
+ * intel_vgpu_decode_primary_plane - Decode primary plane
+ * @vgpu: input vgpu
+ * @plane: primary plane to save decoded info
+ * This function is called for decoding plane
+ *
+ * Returns:
+ * 0 on success, non-zero if failed.
+ */
+int intel_vgpu_decode_primary_plane(struct intel_vgpu *vgpu,
+	struct intel_vgpu_primary_plane_format *plane)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+	u32 val, fmt;
+	int pipe;
+
+	pipe = get_active_pipe(vgpu);
+	if (pipe >= I915_MAX_PIPES)
+		return -ENODEV;
+
+	val = vgpu_vreg_t(vgpu, DSPCNTR(pipe));
+	plane->enabled = !!(val & DISP_ENABLE);
+	if (!plane->enabled)
+		return -ENODEV;
+
+	if (GRAPHICS_VER(dev_priv) >= 9) {
+		plane->tiled = val & PLANE_CTL_TILED_MASK;
+		fmt = skl_format_to_drm(
+			val & PLANE_CTL_FORMAT_MASK_SKL,
+			val & PLANE_CTL_ORDER_RGBX,
+			val & PLANE_CTL_ALPHA_MASK,
+			val & PLANE_CTL_YUV422_ORDER_MASK);
+
+		if (fmt >= ARRAY_SIZE(skl_pixel_formats)) {
+			gvt_vgpu_err("Out-of-bounds pixel format index\n");
+			return -EINVAL;
+		}
+
+		plane->bpp = skl_pixel_formats[fmt].bpp;
+		plane->drm_format = skl_pixel_formats[fmt].drm_format;
+	} else {
+		plane->tiled = val & DISP_TILED;
+		fmt = bdw_format_to_drm(val & DISP_FORMAT_MASK);
+		plane->bpp = bdw_pixel_formats[fmt].bpp;
+		plane->drm_format = bdw_pixel_formats[fmt].drm_format;
+	}
+
+	if (!plane->bpp) {
+		gvt_vgpu_err("Non-supported pixel format (0x%x)\n", fmt);
+		return -EINVAL;
+	}
+
+	plane->hw_format = fmt;
+
+	plane->base = vgpu_vreg_t(vgpu, DSPSURF(pipe)) & I915_GTT_PAGE_MASK;
+	if (!vgpu_gmadr_is_valid(vgpu, plane->base))
+		return  -EINVAL;
+
+	plane->base_gpa = intel_vgpu_gma_to_gpa(vgpu->gtt.ggtt_mm, plane->base);
+	if (plane->base_gpa == INTEL_GVT_INVALID_ADDR) {
+		gvt_vgpu_err("Translate primary plane gma 0x%x to gpa fail\n",
+				plane->base);
+		return  -EINVAL;
+	}
+
+	plane->stride = intel_vgpu_get_stride(vgpu, pipe, plane->tiled,
+		(GRAPHICS_VER(dev_priv) >= 9) ?
+		(_PRI_PLANE_STRIDE_MASK >> 6) :
+		_PRI_PLANE_STRIDE_MASK, plane->bpp);
+
+	plane->width = (vgpu_vreg_t(vgpu, PIPESRC(pipe)) & _PIPE_H_SRCSZ_MASK) >>
+		_PIPE_H_SRCSZ_SHIFT;
+	plane->width += 1;
+	plane->height = (vgpu_vreg_t(vgpu, PIPESRC(pipe)) &
+			_PIPE_V_SRCSZ_MASK) >> _PIPE_V_SRCSZ_SHIFT;
+	plane->height += 1;	/* raw height is one minus the real value */
+
+	val = vgpu_vreg_t(vgpu, DSPTILEOFF(pipe));
+	plane->x_offset = (val & _PRI_PLANE_X_OFF_MASK) >>
+		_PRI_PLANE_X_OFF_SHIFT;
+	plane->y_offset = (val & _PRI_PLANE_Y_OFF_MASK) >>
+		_PRI_PLANE_Y_OFF_SHIFT;
+
+	return 0;
+}
+
+#define CURSOR_FORMAT_NUM	(1 << 6)
+struct cursor_mode_format {
+	int drm_format;	/* Pixel format in DRM definition */
+	u8 bpp; /* Bits per pixel; 0 indicates invalid */
+	u32 width; /* In pixel */
+	u32 height; /* In lines */
+	const char *desc; /* The description */
+};
+
+static const struct cursor_mode_format cursor_pixel_formats[] = {
+	{DRM_FORMAT_ARGB8888, 32, 128, 128, "128x128 32bpp ARGB"},
+	{DRM_FORMAT_ARGB8888, 32, 256, 256, "256x256 32bpp ARGB"},
+	{DRM_FORMAT_ARGB8888, 32, 64, 64, "64x64 32bpp ARGB"},
+	{DRM_FORMAT_ARGB8888, 32, 64, 64, "64x64 32bpp ARGB"},
+
+	/* non-supported format has bpp default to 0 */
+	{0, 0, 0, 0, NULL},
+};
+
+static int cursor_mode_to_drm(int mode)
+{
+	int cursor_pixel_formats_index = 4;
+
+	switch (mode) {
+	case MCURSOR_MODE_128_ARGB_AX:
+		cursor_pixel_formats_index = 0;
+		break;
+	case MCURSOR_MODE_256_ARGB_AX:
+		cursor_pixel_formats_index = 1;
+		break;
+	case MCURSOR_MODE_64_ARGB_AX:
+		cursor_pixel_formats_index = 2;
+		break;
+	case MCURSOR_MODE_64_32B_AX:
+		cursor_pixel_formats_index = 3;
+		break;
+
+	default:
+		break;
+	}
+
+	return cursor_pixel_formats_index;
+}
+
+/**
+ * intel_vgpu_decode_cursor_plane - Decode sprite plane
+ * @vgpu: input vgpu
+ * @plane: cursor plane to save decoded info
+ * This function is called for decoding plane
+ *
+ * Returns:
+ * 0 on success, non-zero if failed.
+ */
+int intel_vgpu_decode_cursor_plane(struct intel_vgpu *vgpu,
+	struct intel_vgpu_cursor_plane_format *plane)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+	u32 val, mode, index;
+	u32 alpha_plane, alpha_force;
+	int pipe;
+
+	pipe = get_active_pipe(vgpu);
+	if (pipe >= I915_MAX_PIPES)
+		return -ENODEV;
+
+	val = vgpu_vreg_t(vgpu, CURCNTR(pipe));
+	mode = val & MCURSOR_MODE_MASK;
+	plane->enabled = (mode != MCURSOR_MODE_DISABLE);
+	if (!plane->enabled)
+		return -ENODEV;
+
+	index = cursor_mode_to_drm(mode);
+
+	if (!cursor_pixel_formats[index].bpp) {
+		gvt_vgpu_err("Non-supported cursor mode (0x%x)\n", mode);
+		return -EINVAL;
+	}
+	plane->mode = mode;
+	plane->bpp = cursor_pixel_formats[index].bpp;
+	plane->drm_format = cursor_pixel_formats[index].drm_format;
+	plane->width = cursor_pixel_formats[index].width;
+	plane->height = cursor_pixel_formats[index].height;
+
+	alpha_plane = (val & _CURSOR_ALPHA_PLANE_MASK) >>
+				_CURSOR_ALPHA_PLANE_SHIFT;
+	alpha_force = (val & _CURSOR_ALPHA_FORCE_MASK) >>
+				_CURSOR_ALPHA_FORCE_SHIFT;
+	if (alpha_plane || alpha_force)
+		gvt_dbg_core("alpha_plane=0x%x, alpha_force=0x%x\n",
+			alpha_plane, alpha_force);
+
+	plane->base = vgpu_vreg_t(vgpu, CURBASE(pipe)) & I915_GTT_PAGE_MASK;
+	if (!vgpu_gmadr_is_valid(vgpu, plane->base))
+		return  -EINVAL;
+
+	plane->base_gpa = intel_vgpu_gma_to_gpa(vgpu->gtt.ggtt_mm, plane->base);
+	if (plane->base_gpa == INTEL_GVT_INVALID_ADDR) {
+		gvt_vgpu_err("Translate cursor plane gma 0x%x to gpa fail\n",
+				plane->base);
+		return  -EINVAL;
+	}
+
+	val = vgpu_vreg_t(vgpu, CURPOS(pipe));
+	plane->x_pos = (val & _CURSOR_POS_X_MASK) >> _CURSOR_POS_X_SHIFT;
+	plane->x_sign = (val & _CURSOR_SIGN_X_MASK) >> _CURSOR_SIGN_X_SHIFT;
+	plane->y_pos = (val & _CURSOR_POS_Y_MASK) >> _CURSOR_POS_Y_SHIFT;
+	plane->y_sign = (val & _CURSOR_SIGN_Y_MASK) >> _CURSOR_SIGN_Y_SHIFT;
+
+	plane->x_hot = vgpu_vreg_t(vgpu, vgtif_reg(cursor_x_hot));
+	plane->y_hot = vgpu_vreg_t(vgpu, vgtif_reg(cursor_y_hot));
+	return 0;
+}
+
+#define SPRITE_FORMAT_NUM	(1 << 3)
+
+static const struct pixel_format sprite_pixel_formats[SPRITE_FORMAT_NUM] = {
+	[0x0] = {DRM_FORMAT_YUV422, 16, "YUV 16-bit 4:2:2 packed"},
+	[0x1] = {DRM_FORMAT_XRGB2101010, 32, "RGB 32-bit 2:10:10:10"},
+	[0x2] = {DRM_FORMAT_XRGB8888, 32, "RGB 32-bit 8:8:8:8"},
+	[0x4] = {DRM_FORMAT_AYUV, 32,
+		"YUV 32-bit 4:4:4 packed (8:8:8:8 MSB-X:Y:U:V)"},
+};
+
+/**
+ * intel_vgpu_decode_sprite_plane - Decode sprite plane
+ * @vgpu: input vgpu
+ * @plane: sprite plane to save decoded info
+ * This function is called for decoding plane
+ *
+ * Returns:
+ * 0 on success, non-zero if failed.
+ */
+int intel_vgpu_decode_sprite_plane(struct intel_vgpu *vgpu,
+	struct intel_vgpu_sprite_plane_format *plane)
+{
+	u32 val, fmt;
+	u32 color_order, yuv_order;
+	int drm_format;
+	int pipe;
+
+	pipe = get_active_pipe(vgpu);
+	if (pipe >= I915_MAX_PIPES)
+		return -ENODEV;
+
+	val = vgpu_vreg_t(vgpu, SPRCTL(pipe));
+	plane->enabled = !!(val & SPRITE_ENABLE);
+	if (!plane->enabled)
+		return -ENODEV;
+
+	plane->tiled = !!(val & SPRITE_TILED);
+	color_order = !!(val & SPRITE_RGB_ORDER_RGBX);
+	yuv_order = (val & SPRITE_YUV_ORDER_MASK) >>
+				_SPRITE_YUV_ORDER_SHIFT;
+
+	fmt = (val & SPRITE_FORMAT_MASK) >> _SPRITE_FMT_SHIFT;
+	if (!sprite_pixel_formats[fmt].bpp) {
+		gvt_vgpu_err("Non-supported pixel format (0x%x)\n", fmt);
+		return -EINVAL;
+	}
+	plane->hw_format = fmt;
+	plane->bpp = sprite_pixel_formats[fmt].bpp;
+	drm_format = sprite_pixel_formats[fmt].drm_format;
+
+	/* Order of RGB values in an RGBxxx buffer may be ordered RGB or
+	 * BGR depending on the state of the color_order field
+	 */
+	if (!color_order) {
+		if (drm_format == DRM_FORMAT_XRGB2101010)
+			drm_format = DRM_FORMAT_XBGR2101010;
+		else if (drm_format == DRM_FORMAT_XRGB8888)
+			drm_format = DRM_FORMAT_XBGR8888;
+	}
+
+	if (drm_format == DRM_FORMAT_YUV422) {
+		switch (yuv_order) {
+		case 0:
+			drm_format = DRM_FORMAT_YUYV;
+			break;
+		case 1:
+			drm_format = DRM_FORMAT_UYVY;
+			break;
+		case 2:
+			drm_format = DRM_FORMAT_YVYU;
+			break;
+		case 3:
+			drm_format = DRM_FORMAT_VYUY;
+			break;
+		default:
+			/* yuv_order has only 2 bits */
+			break;
+		}
+	}
+
+	plane->drm_format = drm_format;
+
+	plane->base = vgpu_vreg_t(vgpu, SPRSURF(pipe)) & I915_GTT_PAGE_MASK;
+	if (!vgpu_gmadr_is_valid(vgpu, plane->base))
+		return  -EINVAL;
+
+	plane->base_gpa = intel_vgpu_gma_to_gpa(vgpu->gtt.ggtt_mm, plane->base);
+	if (plane->base_gpa == INTEL_GVT_INVALID_ADDR) {
+		gvt_vgpu_err("Translate sprite plane gma 0x%x to gpa fail\n",
+				plane->base);
+		return  -EINVAL;
+	}
+
+	plane->stride = vgpu_vreg_t(vgpu, SPRSTRIDE(pipe)) &
+				_SPRITE_STRIDE_MASK;
+
+	val = vgpu_vreg_t(vgpu, SPRSIZE(pipe));
+	plane->height = (val & _SPRITE_SIZE_HEIGHT_MASK) >>
+		_SPRITE_SIZE_HEIGHT_SHIFT;
+	plane->width = (val & _SPRITE_SIZE_WIDTH_MASK) >>
+		_SPRITE_SIZE_WIDTH_SHIFT;
+	plane->height += 1;	/* raw height is one minus the real value */
+	plane->width += 1;	/* raw width is one minus the real value */
+
+	val = vgpu_vreg_t(vgpu, SPRPOS(pipe));
+	plane->x_pos = (val & _SPRITE_POS_X_MASK) >> _SPRITE_POS_X_SHIFT;
+	plane->y_pos = (val & _SPRITE_POS_Y_MASK) >> _SPRITE_POS_Y_SHIFT;
+
+	val = vgpu_vreg_t(vgpu, SPROFFSET(pipe));
+	plane->x_offset = (val & _SPRITE_OFFSET_START_X_MASK) >>
+			   _SPRITE_OFFSET_START_X_SHIFT;
+	plane->y_offset = (val & _SPRITE_OFFSET_START_Y_MASK) >>
+			   _SPRITE_OFFSET_START_Y_SHIFT;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gvt/fb_decoder.h b/drivers/gpu/drm/i915/gvt/fb_decoder.h
new file mode 100644
index 000000000..0daa3931a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/fb_decoder.h
@@ -0,0 +1,173 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Kevin Tian <kevin.tian@intel.com>
+ *
+ * Contributors:
+ *    Bing Niu <bing.niu@intel.com>
+ *    Xu Han <xu.han@intel.com>
+ *    Ping Gao <ping.a.gao@intel.com>
+ *    Xiaoguang Chen <xiaoguang.chen@intel.com>
+ *    Yang Liu <yang2.liu@intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *
+ */
+
+#ifndef _GVT_FB_DECODER_H_
+#define _GVT_FB_DECODER_H_
+
+#include <linux/types.h>
+
+#include "display/intel_display.h"
+
+struct intel_vgpu;
+
+#define _PLANE_CTL_FORMAT_SHIFT		24
+#define _PLANE_CTL_TILED_SHIFT		10
+#define _PIPE_V_SRCSZ_SHIFT		0
+#define _PIPE_V_SRCSZ_MASK		(0xfff << _PIPE_V_SRCSZ_SHIFT)
+#define _PIPE_H_SRCSZ_SHIFT		16
+#define _PIPE_H_SRCSZ_MASK		(0x1fff << _PIPE_H_SRCSZ_SHIFT)
+
+#define _PRI_PLANE_FMT_SHIFT		26
+#define _PRI_PLANE_STRIDE_MASK		(0x3ff << 6)
+#define _PRI_PLANE_X_OFF_SHIFT		0
+#define _PRI_PLANE_X_OFF_MASK		(0x1fff << _PRI_PLANE_X_OFF_SHIFT)
+#define _PRI_PLANE_Y_OFF_SHIFT		16
+#define _PRI_PLANE_Y_OFF_MASK		(0xfff << _PRI_PLANE_Y_OFF_SHIFT)
+
+#define _CURSOR_MODE			0x3f
+#define _CURSOR_ALPHA_FORCE_SHIFT	8
+#define _CURSOR_ALPHA_FORCE_MASK	(0x3 << _CURSOR_ALPHA_FORCE_SHIFT)
+#define _CURSOR_ALPHA_PLANE_SHIFT	10
+#define _CURSOR_ALPHA_PLANE_MASK	(0x3 << _CURSOR_ALPHA_PLANE_SHIFT)
+#define _CURSOR_POS_X_SHIFT		0
+#define _CURSOR_POS_X_MASK		(0x1fff << _CURSOR_POS_X_SHIFT)
+#define _CURSOR_SIGN_X_SHIFT		15
+#define _CURSOR_SIGN_X_MASK		(1 << _CURSOR_SIGN_X_SHIFT)
+#define _CURSOR_POS_Y_SHIFT		16
+#define _CURSOR_POS_Y_MASK		(0xfff << _CURSOR_POS_Y_SHIFT)
+#define _CURSOR_SIGN_Y_SHIFT		31
+#define _CURSOR_SIGN_Y_MASK		(1 << _CURSOR_SIGN_Y_SHIFT)
+
+#define _SPRITE_FMT_SHIFT		25
+#define _SPRITE_COLOR_ORDER_SHIFT	20
+#define _SPRITE_YUV_ORDER_SHIFT		16
+#define _SPRITE_STRIDE_SHIFT		6
+#define _SPRITE_STRIDE_MASK		(0x1ff << _SPRITE_STRIDE_SHIFT)
+#define _SPRITE_SIZE_WIDTH_SHIFT	0
+#define _SPRITE_SIZE_HEIGHT_SHIFT	16
+#define _SPRITE_SIZE_WIDTH_MASK		(0x1fff << _SPRITE_SIZE_WIDTH_SHIFT)
+#define _SPRITE_SIZE_HEIGHT_MASK	(0xfff << _SPRITE_SIZE_HEIGHT_SHIFT)
+#define _SPRITE_POS_X_SHIFT		0
+#define _SPRITE_POS_Y_SHIFT		16
+#define _SPRITE_POS_X_MASK		(0x1fff << _SPRITE_POS_X_SHIFT)
+#define _SPRITE_POS_Y_MASK		(0xfff << _SPRITE_POS_Y_SHIFT)
+#define _SPRITE_OFFSET_START_X_SHIFT	0
+#define _SPRITE_OFFSET_START_Y_SHIFT	16
+#define _SPRITE_OFFSET_START_X_MASK	(0x1fff << _SPRITE_OFFSET_START_X_SHIFT)
+#define _SPRITE_OFFSET_START_Y_MASK	(0xfff << _SPRITE_OFFSET_START_Y_SHIFT)
+
+enum GVT_FB_EVENT {
+	FB_MODE_SET_START = 1,
+	FB_MODE_SET_END,
+	FB_DISPLAY_FLIP,
+};
+
+enum DDI_PORT {
+	DDI_PORT_NONE	= 0,
+	DDI_PORT_B	= 1,
+	DDI_PORT_C	= 2,
+	DDI_PORT_D	= 3,
+	DDI_PORT_E	= 4
+};
+
+/* color space conversion and gamma correction are not included */
+struct intel_vgpu_primary_plane_format {
+	u8	enabled;	/* plane is enabled */
+	u32	tiled;		/* tiling mode: linear, X-tiled, Y tiled, etc */
+	u8	bpp;		/* bits per pixel */
+	u32	hw_format;	/* format field in the PRI_CTL register */
+	u32	drm_format;	/* format in DRM definition */
+	u32	base;		/* framebuffer base in graphics memory */
+	u64     base_gpa;
+	u32	x_offset;	/* in pixels */
+	u32	y_offset;	/* in lines */
+	u32	width;		/* in pixels */
+	u32	height;		/* in lines */
+	u32	stride;		/* in bytes */
+};
+
+struct intel_vgpu_sprite_plane_format {
+	u8	enabled;	/* plane is enabled */
+	u8	tiled;		/* X-tiled */
+	u8	bpp;		/* bits per pixel */
+	u32	hw_format;	/* format field in the SPR_CTL register */
+	u32	drm_format;	/* format in DRM definition */
+	u32	base;		/* sprite base in graphics memory */
+	u64     base_gpa;
+	u32	x_pos;		/* in pixels */
+	u32	y_pos;		/* in lines */
+	u32	x_offset;	/* in pixels */
+	u32	y_offset;	/* in lines */
+	u32	width;		/* in pixels */
+	u32	height;		/* in lines */
+	u32	stride;		/* in bytes */
+};
+
+struct intel_vgpu_cursor_plane_format {
+	u8	enabled;
+	u8	mode;		/* cursor mode select */
+	u8	bpp;		/* bits per pixel */
+	u32	drm_format;	/* format in DRM definition */
+	u32	base;		/* cursor base in graphics memory */
+	u64     base_gpa;
+	u32	x_pos;		/* in pixels */
+	u32	y_pos;		/* in lines */
+	u8	x_sign;		/* X Position Sign */
+	u8	y_sign;		/* Y Position Sign */
+	u32	width;		/* in pixels */
+	u32	height;		/* in lines */
+	u32	x_hot;		/* in pixels */
+	u32	y_hot;		/* in pixels */
+};
+
+struct intel_vgpu_pipe_format {
+	struct intel_vgpu_primary_plane_format	primary;
+	struct intel_vgpu_sprite_plane_format	sprite;
+	struct intel_vgpu_cursor_plane_format	cursor;
+	enum DDI_PORT ddi_port;  /* the DDI port that pipe is connected to */
+};
+
+struct intel_vgpu_fb_format {
+	struct intel_vgpu_pipe_format	pipes[I915_MAX_PIPES];
+};
+
+int intel_vgpu_decode_primary_plane(struct intel_vgpu *vgpu,
+	struct intel_vgpu_primary_plane_format *plane);
+int intel_vgpu_decode_cursor_plane(struct intel_vgpu *vgpu,
+	struct intel_vgpu_cursor_plane_format *plane);
+int intel_vgpu_decode_sprite_plane(struct intel_vgpu *vgpu,
+	struct intel_vgpu_sprite_plane_format *plane);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/firmware.c b/drivers/gpu/drm/i915/gvt/firmware.c
new file mode 100644
index 000000000..54fe44223
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/firmware.c
@@ -0,0 +1,266 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ * Contributors:
+ *    Changbin Du <changbin.du@intel.com>
+ *
+ */
+
+#include <linux/firmware.h>
+#include <linux/crc32.h>
+
+#include "i915_drv.h"
+#include "gvt.h"
+#include "i915_pvinfo.h"
+
+#define FIRMWARE_VERSION (0x0)
+
+struct gvt_firmware_header {
+	u64 magic;
+	u32 crc32;		/* protect the data after this field */
+	u32 version;
+	u64 cfg_space_size;
+	u64 cfg_space_offset;	/* offset in the file */
+	u64 mmio_size;
+	u64 mmio_offset;	/* offset in the file */
+	unsigned char data[1];
+};
+
+#define dev_to_drm_minor(d) dev_get_drvdata((d))
+
+static ssize_t
+gvt_firmware_read(struct file *filp, struct kobject *kobj,
+	     struct bin_attribute *attr, char *buf,
+	     loff_t offset, size_t count)
+{
+	memcpy(buf, attr->private + offset, count);
+	return count;
+}
+
+static struct bin_attribute firmware_attr = {
+	.attr = {.name = "gvt_firmware", .mode = (S_IRUSR)},
+	.read = gvt_firmware_read,
+	.write = NULL,
+	.mmap = NULL,
+};
+
+static int expose_firmware_sysfs(struct intel_gvt *gvt)
+{
+	struct intel_gvt_device_info *info = &gvt->device_info;
+	struct drm_i915_private *i915 = gvt->gt->i915;
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	struct gvt_firmware_header *h;
+	void *firmware;
+	void *p;
+	unsigned long size, crc32_start;
+	int ret;
+
+	size = sizeof(*h) + info->mmio_size + info->cfg_space_size;
+	firmware = vzalloc(size);
+	if (!firmware)
+		return -ENOMEM;
+
+	h = firmware;
+
+	h->magic = VGT_MAGIC;
+	h->version = FIRMWARE_VERSION;
+	h->cfg_space_size = info->cfg_space_size;
+	h->cfg_space_offset = offsetof(struct gvt_firmware_header, data);
+	h->mmio_size = info->mmio_size;
+	h->mmio_offset = h->cfg_space_offset + h->cfg_space_size;
+
+	p = firmware + h->cfg_space_offset;
+
+	memcpy(gvt->firmware.cfg_space, i915->vgpu.initial_cfg_space,
+	       info->cfg_space_size);
+	memcpy(p, gvt->firmware.cfg_space, info->cfg_space_size);
+
+	p = firmware + h->mmio_offset;
+
+	memcpy(gvt->firmware.mmio, i915->vgpu.initial_mmio,
+	       info->mmio_size);
+
+	memcpy(p, gvt->firmware.mmio, info->mmio_size);
+
+	crc32_start = offsetof(struct gvt_firmware_header, crc32) + 4;
+	h->crc32 = crc32_le(0, firmware + crc32_start, size - crc32_start);
+
+	firmware_attr.size = size;
+	firmware_attr.private = firmware;
+
+	ret = device_create_bin_file(&pdev->dev, &firmware_attr);
+	if (ret) {
+		vfree(firmware);
+		return ret;
+	}
+	return 0;
+}
+
+static void clean_firmware_sysfs(struct intel_gvt *gvt)
+{
+	struct pci_dev *pdev = to_pci_dev(gvt->gt->i915->drm.dev);
+
+	device_remove_bin_file(&pdev->dev, &firmware_attr);
+	vfree(firmware_attr.private);
+}
+
+/**
+ * intel_gvt_free_firmware - free GVT firmware
+ * @gvt: intel gvt device
+ *
+ */
+void intel_gvt_free_firmware(struct intel_gvt *gvt)
+{
+	if (!gvt->firmware.firmware_loaded)
+		clean_firmware_sysfs(gvt);
+
+	kfree(gvt->firmware.cfg_space);
+	vfree(gvt->firmware.mmio);
+}
+
+static int verify_firmware(struct intel_gvt *gvt,
+			   const struct firmware *fw)
+{
+	struct intel_gvt_device_info *info = &gvt->device_info;
+	struct pci_dev *pdev = to_pci_dev(gvt->gt->i915->drm.dev);
+	struct gvt_firmware_header *h;
+	unsigned long id, crc32_start;
+	const void *mem;
+	const char *item;
+	u64 file, request;
+
+	h = (struct gvt_firmware_header *)fw->data;
+
+	crc32_start = offsetofend(struct gvt_firmware_header, crc32);
+	mem = fw->data + crc32_start;
+
+#define VERIFY(s, a, b) do { \
+	item = (s); file = (u64)(a); request = (u64)(b); \
+	if ((a) != (b)) \
+		goto invalid_firmware; \
+} while (0)
+
+	VERIFY("magic number", h->magic, VGT_MAGIC);
+	VERIFY("version", h->version, FIRMWARE_VERSION);
+	VERIFY("crc32", h->crc32, crc32_le(0, mem, fw->size - crc32_start));
+	VERIFY("cfg space size", h->cfg_space_size, info->cfg_space_size);
+	VERIFY("mmio size", h->mmio_size, info->mmio_size);
+
+	mem = (fw->data + h->cfg_space_offset);
+
+	id = *(u16 *)(mem + PCI_VENDOR_ID);
+	VERIFY("vender id", id, pdev->vendor);
+
+	id = *(u16 *)(mem + PCI_DEVICE_ID);
+	VERIFY("device id", id, pdev->device);
+
+	id = *(u8 *)(mem + PCI_REVISION_ID);
+	VERIFY("revision id", id, pdev->revision);
+
+#undef VERIFY
+	return 0;
+
+invalid_firmware:
+	gvt_dbg_core("Invalid firmware: %s [file] 0x%llx [request] 0x%llx\n",
+		     item, file, request);
+	return -EINVAL;
+}
+
+#define GVT_FIRMWARE_PATH "i915/gvt"
+
+/**
+ * intel_gvt_load_firmware - load GVT firmware
+ * @gvt: intel gvt device
+ *
+ */
+int intel_gvt_load_firmware(struct intel_gvt *gvt)
+{
+	struct intel_gvt_device_info *info = &gvt->device_info;
+	struct pci_dev *pdev = to_pci_dev(gvt->gt->i915->drm.dev);
+	struct intel_gvt_firmware *firmware = &gvt->firmware;
+	struct gvt_firmware_header *h;
+	const struct firmware *fw;
+	char *path;
+	void *mem;
+	int ret;
+
+	path = kmalloc(PATH_MAX, GFP_KERNEL);
+	if (!path)
+		return -ENOMEM;
+
+	mem = kmalloc(info->cfg_space_size, GFP_KERNEL);
+	if (!mem) {
+		kfree(path);
+		return -ENOMEM;
+	}
+
+	firmware->cfg_space = mem;
+
+	mem = vmalloc(info->mmio_size);
+	if (!mem) {
+		kfree(path);
+		kfree(firmware->cfg_space);
+		return -ENOMEM;
+	}
+
+	firmware->mmio = mem;
+
+	sprintf(path, "%s/vid_0x%04x_did_0x%04x_rid_0x%02x.golden_hw_state",
+		 GVT_FIRMWARE_PATH, pdev->vendor, pdev->device,
+		 pdev->revision);
+
+	gvt_dbg_core("request hw state firmware %s...\n", path);
+
+	ret = request_firmware(&fw, path, gvt->gt->i915->drm.dev);
+	kfree(path);
+
+	if (ret)
+		goto expose_firmware;
+
+	gvt_dbg_core("success.\n");
+
+	ret = verify_firmware(gvt, fw);
+	if (ret)
+		goto out_free_fw;
+
+	gvt_dbg_core("verified.\n");
+
+	h = (struct gvt_firmware_header *)fw->data;
+
+	memcpy(firmware->cfg_space, fw->data + h->cfg_space_offset,
+	       h->cfg_space_size);
+	memcpy(firmware->mmio, fw->data + h->mmio_offset,
+	       h->mmio_size);
+
+	release_firmware(fw);
+	firmware->firmware_loaded = true;
+	return 0;
+
+out_free_fw:
+	release_firmware(fw);
+expose_firmware:
+	expose_firmware_sysfs(gvt);
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gvt/gtt.c b/drivers/gpu/drm/i915/gvt/gtt.c
new file mode 100644
index 000000000..79bf1be68
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/gtt.c
@@ -0,0 +1,2926 @@
+/*
+ * GTT virtualization
+ *
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *    Zhenyu Wang <zhenyuw@linux.intel.com>
+ *    Xiao Zheng <xiao.zheng@intel.com>
+ *
+ * Contributors:
+ *    Min He <min.he@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *
+ */
+
+#include "i915_drv.h"
+#include "gvt.h"
+#include "i915_pvinfo.h"
+#include "trace.h"
+
+#include "gt/intel_gt_regs.h"
+
+#if defined(VERBOSE_DEBUG)
+#define gvt_vdbg_mm(fmt, args...) gvt_dbg_mm(fmt, ##args)
+#else
+#define gvt_vdbg_mm(fmt, args...)
+#endif
+
+static bool enable_out_of_sync = false;
+static int preallocated_oos_pages = 8192;
+
+static bool intel_gvt_is_valid_gfn(struct intel_vgpu *vgpu, unsigned long gfn)
+{
+	struct kvm *kvm = vgpu->vfio_device.kvm;
+	int idx;
+	bool ret;
+
+	if (!vgpu->attached)
+		return false;
+
+	idx = srcu_read_lock(&kvm->srcu);
+	ret = kvm_is_visible_gfn(kvm, gfn);
+	srcu_read_unlock(&kvm->srcu, idx);
+
+	return ret;
+}
+
+/*
+ * validate a gm address and related range size,
+ * translate it to host gm address
+ */
+bool intel_gvt_ggtt_validate_range(struct intel_vgpu *vgpu, u64 addr, u32 size)
+{
+	if (size == 0)
+		return vgpu_gmadr_is_valid(vgpu, addr);
+
+	if (vgpu_gmadr_is_aperture(vgpu, addr) &&
+	    vgpu_gmadr_is_aperture(vgpu, addr + size - 1))
+		return true;
+	else if (vgpu_gmadr_is_hidden(vgpu, addr) &&
+		 vgpu_gmadr_is_hidden(vgpu, addr + size - 1))
+		return true;
+
+	gvt_dbg_mm("Invalid ggtt range at 0x%llx, size: 0x%x\n",
+		     addr, size);
+	return false;
+}
+
+/* translate a guest gmadr to host gmadr */
+int intel_gvt_ggtt_gmadr_g2h(struct intel_vgpu *vgpu, u64 g_addr, u64 *h_addr)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+
+	if (drm_WARN(&i915->drm, !vgpu_gmadr_is_valid(vgpu, g_addr),
+		     "invalid guest gmadr %llx\n", g_addr))
+		return -EACCES;
+
+	if (vgpu_gmadr_is_aperture(vgpu, g_addr))
+		*h_addr = vgpu_aperture_gmadr_base(vgpu)
+			  + (g_addr - vgpu_aperture_offset(vgpu));
+	else
+		*h_addr = vgpu_hidden_gmadr_base(vgpu)
+			  + (g_addr - vgpu_hidden_offset(vgpu));
+	return 0;
+}
+
+/* translate a host gmadr to guest gmadr */
+int intel_gvt_ggtt_gmadr_h2g(struct intel_vgpu *vgpu, u64 h_addr, u64 *g_addr)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+
+	if (drm_WARN(&i915->drm, !gvt_gmadr_is_valid(vgpu->gvt, h_addr),
+		     "invalid host gmadr %llx\n", h_addr))
+		return -EACCES;
+
+	if (gvt_gmadr_is_aperture(vgpu->gvt, h_addr))
+		*g_addr = vgpu_aperture_gmadr_base(vgpu)
+			+ (h_addr - gvt_aperture_gmadr_base(vgpu->gvt));
+	else
+		*g_addr = vgpu_hidden_gmadr_base(vgpu)
+			+ (h_addr - gvt_hidden_gmadr_base(vgpu->gvt));
+	return 0;
+}
+
+int intel_gvt_ggtt_index_g2h(struct intel_vgpu *vgpu, unsigned long g_index,
+			     unsigned long *h_index)
+{
+	u64 h_addr;
+	int ret;
+
+	ret = intel_gvt_ggtt_gmadr_g2h(vgpu, g_index << I915_GTT_PAGE_SHIFT,
+				       &h_addr);
+	if (ret)
+		return ret;
+
+	*h_index = h_addr >> I915_GTT_PAGE_SHIFT;
+	return 0;
+}
+
+int intel_gvt_ggtt_h2g_index(struct intel_vgpu *vgpu, unsigned long h_index,
+			     unsigned long *g_index)
+{
+	u64 g_addr;
+	int ret;
+
+	ret = intel_gvt_ggtt_gmadr_h2g(vgpu, h_index << I915_GTT_PAGE_SHIFT,
+				       &g_addr);
+	if (ret)
+		return ret;
+
+	*g_index = g_addr >> I915_GTT_PAGE_SHIFT;
+	return 0;
+}
+
+#define gtt_type_is_entry(type) \
+	(type > GTT_TYPE_INVALID && type < GTT_TYPE_PPGTT_ENTRY \
+	 && type != GTT_TYPE_PPGTT_PTE_ENTRY \
+	 && type != GTT_TYPE_PPGTT_ROOT_ENTRY)
+
+#define gtt_type_is_pt(type) \
+	(type >= GTT_TYPE_PPGTT_PTE_PT && type < GTT_TYPE_MAX)
+
+#define gtt_type_is_pte_pt(type) \
+	(type == GTT_TYPE_PPGTT_PTE_PT)
+
+#define gtt_type_is_root_pointer(type) \
+	(gtt_type_is_entry(type) && type > GTT_TYPE_PPGTT_ROOT_ENTRY)
+
+#define gtt_init_entry(e, t, p, v) do { \
+	(e)->type = t; \
+	(e)->pdev = p; \
+	memcpy(&(e)->val64, &v, sizeof(v)); \
+} while (0)
+
+/*
+ * Mappings between GTT_TYPE* enumerations.
+ * Following information can be found according to the given type:
+ * - type of next level page table
+ * - type of entry inside this level page table
+ * - type of entry with PSE set
+ *
+ * If the given type doesn't have such a kind of information,
+ * e.g. give a l4 root entry type, then request to get its PSE type,
+ * give a PTE page table type, then request to get its next level page
+ * table type, as we know l4 root entry doesn't have a PSE bit,
+ * and a PTE page table doesn't have a next level page table type,
+ * GTT_TYPE_INVALID will be returned. This is useful when traversing a
+ * page table.
+ */
+
+struct gtt_type_table_entry {
+	int entry_type;
+	int pt_type;
+	int next_pt_type;
+	int pse_entry_type;
+};
+
+#define GTT_TYPE_TABLE_ENTRY(type, e_type, cpt_type, npt_type, pse_type) \
+	[type] = { \
+		.entry_type = e_type, \
+		.pt_type = cpt_type, \
+		.next_pt_type = npt_type, \
+		.pse_entry_type = pse_type, \
+	}
+
+static const struct gtt_type_table_entry gtt_type_table[] = {
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_ROOT_L4_ENTRY,
+			GTT_TYPE_PPGTT_ROOT_L4_ENTRY,
+			GTT_TYPE_INVALID,
+			GTT_TYPE_PPGTT_PML4_PT,
+			GTT_TYPE_INVALID),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PML4_PT,
+			GTT_TYPE_PPGTT_PML4_ENTRY,
+			GTT_TYPE_PPGTT_PML4_PT,
+			GTT_TYPE_PPGTT_PDP_PT,
+			GTT_TYPE_INVALID),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PML4_ENTRY,
+			GTT_TYPE_PPGTT_PML4_ENTRY,
+			GTT_TYPE_PPGTT_PML4_PT,
+			GTT_TYPE_PPGTT_PDP_PT,
+			GTT_TYPE_INVALID),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PDP_PT,
+			GTT_TYPE_PPGTT_PDP_ENTRY,
+			GTT_TYPE_PPGTT_PDP_PT,
+			GTT_TYPE_PPGTT_PDE_PT,
+			GTT_TYPE_PPGTT_PTE_1G_ENTRY),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_ROOT_L3_ENTRY,
+			GTT_TYPE_PPGTT_ROOT_L3_ENTRY,
+			GTT_TYPE_INVALID,
+			GTT_TYPE_PPGTT_PDE_PT,
+			GTT_TYPE_PPGTT_PTE_1G_ENTRY),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PDP_ENTRY,
+			GTT_TYPE_PPGTT_PDP_ENTRY,
+			GTT_TYPE_PPGTT_PDP_PT,
+			GTT_TYPE_PPGTT_PDE_PT,
+			GTT_TYPE_PPGTT_PTE_1G_ENTRY),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PDE_PT,
+			GTT_TYPE_PPGTT_PDE_ENTRY,
+			GTT_TYPE_PPGTT_PDE_PT,
+			GTT_TYPE_PPGTT_PTE_PT,
+			GTT_TYPE_PPGTT_PTE_2M_ENTRY),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PDE_ENTRY,
+			GTT_TYPE_PPGTT_PDE_ENTRY,
+			GTT_TYPE_PPGTT_PDE_PT,
+			GTT_TYPE_PPGTT_PTE_PT,
+			GTT_TYPE_PPGTT_PTE_2M_ENTRY),
+	/* We take IPS bit as 'PSE' for PTE level. */
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PTE_PT,
+			GTT_TYPE_PPGTT_PTE_4K_ENTRY,
+			GTT_TYPE_PPGTT_PTE_PT,
+			GTT_TYPE_INVALID,
+			GTT_TYPE_PPGTT_PTE_64K_ENTRY),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PTE_4K_ENTRY,
+			GTT_TYPE_PPGTT_PTE_4K_ENTRY,
+			GTT_TYPE_PPGTT_PTE_PT,
+			GTT_TYPE_INVALID,
+			GTT_TYPE_PPGTT_PTE_64K_ENTRY),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PTE_64K_ENTRY,
+			GTT_TYPE_PPGTT_PTE_4K_ENTRY,
+			GTT_TYPE_PPGTT_PTE_PT,
+			GTT_TYPE_INVALID,
+			GTT_TYPE_PPGTT_PTE_64K_ENTRY),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PTE_2M_ENTRY,
+			GTT_TYPE_PPGTT_PDE_ENTRY,
+			GTT_TYPE_PPGTT_PDE_PT,
+			GTT_TYPE_INVALID,
+			GTT_TYPE_PPGTT_PTE_2M_ENTRY),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_PPGTT_PTE_1G_ENTRY,
+			GTT_TYPE_PPGTT_PDP_ENTRY,
+			GTT_TYPE_PPGTT_PDP_PT,
+			GTT_TYPE_INVALID,
+			GTT_TYPE_PPGTT_PTE_1G_ENTRY),
+	GTT_TYPE_TABLE_ENTRY(GTT_TYPE_GGTT_PTE,
+			GTT_TYPE_GGTT_PTE,
+			GTT_TYPE_INVALID,
+			GTT_TYPE_INVALID,
+			GTT_TYPE_INVALID),
+};
+
+static inline int get_next_pt_type(int type)
+{
+	return gtt_type_table[type].next_pt_type;
+}
+
+static inline int get_pt_type(int type)
+{
+	return gtt_type_table[type].pt_type;
+}
+
+static inline int get_entry_type(int type)
+{
+	return gtt_type_table[type].entry_type;
+}
+
+static inline int get_pse_type(int type)
+{
+	return gtt_type_table[type].pse_entry_type;
+}
+
+static u64 read_pte64(struct i915_ggtt *ggtt, unsigned long index)
+{
+	void __iomem *addr = (gen8_pte_t __iomem *)ggtt->gsm + index;
+
+	return readq(addr);
+}
+
+static void ggtt_invalidate(struct intel_gt *gt)
+{
+	mmio_hw_access_pre(gt);
+	intel_uncore_write(gt->uncore, GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN);
+	mmio_hw_access_post(gt);
+}
+
+static void write_pte64(struct i915_ggtt *ggtt, unsigned long index, u64 pte)
+{
+	void __iomem *addr = (gen8_pte_t __iomem *)ggtt->gsm + index;
+
+	writeq(pte, addr);
+}
+
+static inline int gtt_get_entry64(void *pt,
+		struct intel_gvt_gtt_entry *e,
+		unsigned long index, bool hypervisor_access, unsigned long gpa,
+		struct intel_vgpu *vgpu)
+{
+	const struct intel_gvt_device_info *info = &vgpu->gvt->device_info;
+	int ret;
+
+	if (WARN_ON(info->gtt_entry_size != 8))
+		return -EINVAL;
+
+	if (hypervisor_access) {
+		ret = intel_gvt_read_gpa(vgpu, gpa +
+				(index << info->gtt_entry_size_shift),
+				&e->val64, 8);
+		if (WARN_ON(ret))
+			return ret;
+	} else if (!pt) {
+		e->val64 = read_pte64(vgpu->gvt->gt->ggtt, index);
+	} else {
+		e->val64 = *((u64 *)pt + index);
+	}
+	return 0;
+}
+
+static inline int gtt_set_entry64(void *pt,
+		struct intel_gvt_gtt_entry *e,
+		unsigned long index, bool hypervisor_access, unsigned long gpa,
+		struct intel_vgpu *vgpu)
+{
+	const struct intel_gvt_device_info *info = &vgpu->gvt->device_info;
+	int ret;
+
+	if (WARN_ON(info->gtt_entry_size != 8))
+		return -EINVAL;
+
+	if (hypervisor_access) {
+		ret = intel_gvt_write_gpa(vgpu, gpa +
+				(index << info->gtt_entry_size_shift),
+				&e->val64, 8);
+		if (WARN_ON(ret))
+			return ret;
+	} else if (!pt) {
+		write_pte64(vgpu->gvt->gt->ggtt, index, e->val64);
+	} else {
+		*((u64 *)pt + index) = e->val64;
+	}
+	return 0;
+}
+
+#define GTT_HAW 46
+
+#define ADDR_1G_MASK	GENMASK_ULL(GTT_HAW - 1, 30)
+#define ADDR_2M_MASK	GENMASK_ULL(GTT_HAW - 1, 21)
+#define ADDR_64K_MASK	GENMASK_ULL(GTT_HAW - 1, 16)
+#define ADDR_4K_MASK	GENMASK_ULL(GTT_HAW - 1, 12)
+
+#define GTT_SPTE_FLAG_MASK GENMASK_ULL(62, 52)
+#define GTT_SPTE_FLAG_64K_SPLITED BIT(52) /* splited 64K gtt entry */
+
+#define GTT_64K_PTE_STRIDE 16
+
+static unsigned long gen8_gtt_get_pfn(struct intel_gvt_gtt_entry *e)
+{
+	unsigned long pfn;
+
+	if (e->type == GTT_TYPE_PPGTT_PTE_1G_ENTRY)
+		pfn = (e->val64 & ADDR_1G_MASK) >> PAGE_SHIFT;
+	else if (e->type == GTT_TYPE_PPGTT_PTE_2M_ENTRY)
+		pfn = (e->val64 & ADDR_2M_MASK) >> PAGE_SHIFT;
+	else if (e->type == GTT_TYPE_PPGTT_PTE_64K_ENTRY)
+		pfn = (e->val64 & ADDR_64K_MASK) >> PAGE_SHIFT;
+	else
+		pfn = (e->val64 & ADDR_4K_MASK) >> PAGE_SHIFT;
+	return pfn;
+}
+
+static void gen8_gtt_set_pfn(struct intel_gvt_gtt_entry *e, unsigned long pfn)
+{
+	if (e->type == GTT_TYPE_PPGTT_PTE_1G_ENTRY) {
+		e->val64 &= ~ADDR_1G_MASK;
+		pfn &= (ADDR_1G_MASK >> PAGE_SHIFT);
+	} else if (e->type == GTT_TYPE_PPGTT_PTE_2M_ENTRY) {
+		e->val64 &= ~ADDR_2M_MASK;
+		pfn &= (ADDR_2M_MASK >> PAGE_SHIFT);
+	} else if (e->type == GTT_TYPE_PPGTT_PTE_64K_ENTRY) {
+		e->val64 &= ~ADDR_64K_MASK;
+		pfn &= (ADDR_64K_MASK >> PAGE_SHIFT);
+	} else {
+		e->val64 &= ~ADDR_4K_MASK;
+		pfn &= (ADDR_4K_MASK >> PAGE_SHIFT);
+	}
+
+	e->val64 |= (pfn << PAGE_SHIFT);
+}
+
+static bool gen8_gtt_test_pse(struct intel_gvt_gtt_entry *e)
+{
+	return !!(e->val64 & _PAGE_PSE);
+}
+
+static void gen8_gtt_clear_pse(struct intel_gvt_gtt_entry *e)
+{
+	if (gen8_gtt_test_pse(e)) {
+		switch (e->type) {
+		case GTT_TYPE_PPGTT_PTE_2M_ENTRY:
+			e->val64 &= ~_PAGE_PSE;
+			e->type = GTT_TYPE_PPGTT_PDE_ENTRY;
+			break;
+		case GTT_TYPE_PPGTT_PTE_1G_ENTRY:
+			e->type = GTT_TYPE_PPGTT_PDP_ENTRY;
+			e->val64 &= ~_PAGE_PSE;
+			break;
+		default:
+			WARN_ON(1);
+		}
+	}
+}
+
+static bool gen8_gtt_test_ips(struct intel_gvt_gtt_entry *e)
+{
+	if (GEM_WARN_ON(e->type != GTT_TYPE_PPGTT_PDE_ENTRY))
+		return false;
+
+	return !!(e->val64 & GEN8_PDE_IPS_64K);
+}
+
+static void gen8_gtt_clear_ips(struct intel_gvt_gtt_entry *e)
+{
+	if (GEM_WARN_ON(e->type != GTT_TYPE_PPGTT_PDE_ENTRY))
+		return;
+
+	e->val64 &= ~GEN8_PDE_IPS_64K;
+}
+
+static bool gen8_gtt_test_present(struct intel_gvt_gtt_entry *e)
+{
+	/*
+	 * i915 writes PDP root pointer registers without present bit,
+	 * it also works, so we need to treat root pointer entry
+	 * specifically.
+	 */
+	if (e->type == GTT_TYPE_PPGTT_ROOT_L3_ENTRY
+			|| e->type == GTT_TYPE_PPGTT_ROOT_L4_ENTRY)
+		return (e->val64 != 0);
+	else
+		return (e->val64 & GEN8_PAGE_PRESENT);
+}
+
+static void gtt_entry_clear_present(struct intel_gvt_gtt_entry *e)
+{
+	e->val64 &= ~GEN8_PAGE_PRESENT;
+}
+
+static void gtt_entry_set_present(struct intel_gvt_gtt_entry *e)
+{
+	e->val64 |= GEN8_PAGE_PRESENT;
+}
+
+static bool gen8_gtt_test_64k_splited(struct intel_gvt_gtt_entry *e)
+{
+	return !!(e->val64 & GTT_SPTE_FLAG_64K_SPLITED);
+}
+
+static void gen8_gtt_set_64k_splited(struct intel_gvt_gtt_entry *e)
+{
+	e->val64 |= GTT_SPTE_FLAG_64K_SPLITED;
+}
+
+static void gen8_gtt_clear_64k_splited(struct intel_gvt_gtt_entry *e)
+{
+	e->val64 &= ~GTT_SPTE_FLAG_64K_SPLITED;
+}
+
+/*
+ * Per-platform GMA routines.
+ */
+static unsigned long gma_to_ggtt_pte_index(unsigned long gma)
+{
+	unsigned long x = (gma >> I915_GTT_PAGE_SHIFT);
+
+	trace_gma_index(__func__, gma, x);
+	return x;
+}
+
+#define DEFINE_PPGTT_GMA_TO_INDEX(prefix, ename, exp) \
+static unsigned long prefix##_gma_to_##ename##_index(unsigned long gma) \
+{ \
+	unsigned long x = (exp); \
+	trace_gma_index(__func__, gma, x); \
+	return x; \
+}
+
+DEFINE_PPGTT_GMA_TO_INDEX(gen8, pte, (gma >> 12 & 0x1ff));
+DEFINE_PPGTT_GMA_TO_INDEX(gen8, pde, (gma >> 21 & 0x1ff));
+DEFINE_PPGTT_GMA_TO_INDEX(gen8, l3_pdp, (gma >> 30 & 0x3));
+DEFINE_PPGTT_GMA_TO_INDEX(gen8, l4_pdp, (gma >> 30 & 0x1ff));
+DEFINE_PPGTT_GMA_TO_INDEX(gen8, pml4, (gma >> 39 & 0x1ff));
+
+static const struct intel_gvt_gtt_pte_ops gen8_gtt_pte_ops = {
+	.get_entry = gtt_get_entry64,
+	.set_entry = gtt_set_entry64,
+	.clear_present = gtt_entry_clear_present,
+	.set_present = gtt_entry_set_present,
+	.test_present = gen8_gtt_test_present,
+	.test_pse = gen8_gtt_test_pse,
+	.clear_pse = gen8_gtt_clear_pse,
+	.clear_ips = gen8_gtt_clear_ips,
+	.test_ips = gen8_gtt_test_ips,
+	.clear_64k_splited = gen8_gtt_clear_64k_splited,
+	.set_64k_splited = gen8_gtt_set_64k_splited,
+	.test_64k_splited = gen8_gtt_test_64k_splited,
+	.get_pfn = gen8_gtt_get_pfn,
+	.set_pfn = gen8_gtt_set_pfn,
+};
+
+static const struct intel_gvt_gtt_gma_ops gen8_gtt_gma_ops = {
+	.gma_to_ggtt_pte_index = gma_to_ggtt_pte_index,
+	.gma_to_pte_index = gen8_gma_to_pte_index,
+	.gma_to_pde_index = gen8_gma_to_pde_index,
+	.gma_to_l3_pdp_index = gen8_gma_to_l3_pdp_index,
+	.gma_to_l4_pdp_index = gen8_gma_to_l4_pdp_index,
+	.gma_to_pml4_index = gen8_gma_to_pml4_index,
+};
+
+/* Update entry type per pse and ips bit. */
+static void update_entry_type_for_real(const struct intel_gvt_gtt_pte_ops *pte_ops,
+	struct intel_gvt_gtt_entry *entry, bool ips)
+{
+	switch (entry->type) {
+	case GTT_TYPE_PPGTT_PDE_ENTRY:
+	case GTT_TYPE_PPGTT_PDP_ENTRY:
+		if (pte_ops->test_pse(entry))
+			entry->type = get_pse_type(entry->type);
+		break;
+	case GTT_TYPE_PPGTT_PTE_4K_ENTRY:
+		if (ips)
+			entry->type = get_pse_type(entry->type);
+		break;
+	default:
+		GEM_BUG_ON(!gtt_type_is_entry(entry->type));
+	}
+
+	GEM_BUG_ON(entry->type == GTT_TYPE_INVALID);
+}
+
+/*
+ * MM helpers.
+ */
+static void _ppgtt_get_root_entry(struct intel_vgpu_mm *mm,
+		struct intel_gvt_gtt_entry *entry, unsigned long index,
+		bool guest)
+{
+	const struct intel_gvt_gtt_pte_ops *pte_ops = mm->vgpu->gvt->gtt.pte_ops;
+
+	GEM_BUG_ON(mm->type != INTEL_GVT_MM_PPGTT);
+
+	entry->type = mm->ppgtt_mm.root_entry_type;
+	pte_ops->get_entry(guest ? mm->ppgtt_mm.guest_pdps :
+			   mm->ppgtt_mm.shadow_pdps,
+			   entry, index, false, 0, mm->vgpu);
+	update_entry_type_for_real(pte_ops, entry, false);
+}
+
+static inline void ppgtt_get_guest_root_entry(struct intel_vgpu_mm *mm,
+		struct intel_gvt_gtt_entry *entry, unsigned long index)
+{
+	_ppgtt_get_root_entry(mm, entry, index, true);
+}
+
+static inline void ppgtt_get_shadow_root_entry(struct intel_vgpu_mm *mm,
+		struct intel_gvt_gtt_entry *entry, unsigned long index)
+{
+	_ppgtt_get_root_entry(mm, entry, index, false);
+}
+
+static void _ppgtt_set_root_entry(struct intel_vgpu_mm *mm,
+		struct intel_gvt_gtt_entry *entry, unsigned long index,
+		bool guest)
+{
+	const struct intel_gvt_gtt_pte_ops *pte_ops = mm->vgpu->gvt->gtt.pte_ops;
+
+	pte_ops->set_entry(guest ? mm->ppgtt_mm.guest_pdps :
+			   mm->ppgtt_mm.shadow_pdps,
+			   entry, index, false, 0, mm->vgpu);
+}
+
+static inline void ppgtt_set_shadow_root_entry(struct intel_vgpu_mm *mm,
+		struct intel_gvt_gtt_entry *entry, unsigned long index)
+{
+	_ppgtt_set_root_entry(mm, entry, index, false);
+}
+
+static void ggtt_get_guest_entry(struct intel_vgpu_mm *mm,
+		struct intel_gvt_gtt_entry *entry, unsigned long index)
+{
+	const struct intel_gvt_gtt_pte_ops *pte_ops = mm->vgpu->gvt->gtt.pte_ops;
+
+	GEM_BUG_ON(mm->type != INTEL_GVT_MM_GGTT);
+
+	entry->type = GTT_TYPE_GGTT_PTE;
+	pte_ops->get_entry(mm->ggtt_mm.virtual_ggtt, entry, index,
+			   false, 0, mm->vgpu);
+}
+
+static void ggtt_set_guest_entry(struct intel_vgpu_mm *mm,
+		struct intel_gvt_gtt_entry *entry, unsigned long index)
+{
+	const struct intel_gvt_gtt_pte_ops *pte_ops = mm->vgpu->gvt->gtt.pte_ops;
+
+	GEM_BUG_ON(mm->type != INTEL_GVT_MM_GGTT);
+
+	pte_ops->set_entry(mm->ggtt_mm.virtual_ggtt, entry, index,
+			   false, 0, mm->vgpu);
+}
+
+static void ggtt_get_host_entry(struct intel_vgpu_mm *mm,
+		struct intel_gvt_gtt_entry *entry, unsigned long index)
+{
+	const struct intel_gvt_gtt_pte_ops *pte_ops = mm->vgpu->gvt->gtt.pte_ops;
+
+	GEM_BUG_ON(mm->type != INTEL_GVT_MM_GGTT);
+
+	pte_ops->get_entry(NULL, entry, index, false, 0, mm->vgpu);
+}
+
+static void ggtt_set_host_entry(struct intel_vgpu_mm *mm,
+		struct intel_gvt_gtt_entry *entry, unsigned long index)
+{
+	const struct intel_gvt_gtt_pte_ops *pte_ops = mm->vgpu->gvt->gtt.pte_ops;
+	unsigned long offset = index;
+
+	GEM_BUG_ON(mm->type != INTEL_GVT_MM_GGTT);
+
+	if (vgpu_gmadr_is_aperture(mm->vgpu, index << I915_GTT_PAGE_SHIFT)) {
+		offset -= (vgpu_aperture_gmadr_base(mm->vgpu) >> PAGE_SHIFT);
+		mm->ggtt_mm.host_ggtt_aperture[offset] = entry->val64;
+	} else if (vgpu_gmadr_is_hidden(mm->vgpu, index << I915_GTT_PAGE_SHIFT)) {
+		offset -= (vgpu_hidden_gmadr_base(mm->vgpu) >> PAGE_SHIFT);
+		mm->ggtt_mm.host_ggtt_hidden[offset] = entry->val64;
+	}
+
+	pte_ops->set_entry(NULL, entry, index, false, 0, mm->vgpu);
+}
+
+/*
+ * PPGTT shadow page table helpers.
+ */
+static inline int ppgtt_spt_get_entry(
+		struct intel_vgpu_ppgtt_spt *spt,
+		void *page_table, int type,
+		struct intel_gvt_gtt_entry *e, unsigned long index,
+		bool guest)
+{
+	struct intel_gvt *gvt = spt->vgpu->gvt;
+	const struct intel_gvt_gtt_pte_ops *ops = gvt->gtt.pte_ops;
+	int ret;
+
+	e->type = get_entry_type(type);
+
+	if (WARN(!gtt_type_is_entry(e->type), "invalid entry type\n"))
+		return -EINVAL;
+
+	ret = ops->get_entry(page_table, e, index, guest,
+			spt->guest_page.gfn << I915_GTT_PAGE_SHIFT,
+			spt->vgpu);
+	if (ret)
+		return ret;
+
+	update_entry_type_for_real(ops, e, guest ?
+				   spt->guest_page.pde_ips : false);
+
+	gvt_vdbg_mm("read ppgtt entry, spt type %d, entry type %d, index %lu, value %llx\n",
+		    type, e->type, index, e->val64);
+	return 0;
+}
+
+static inline int ppgtt_spt_set_entry(
+		struct intel_vgpu_ppgtt_spt *spt,
+		void *page_table, int type,
+		struct intel_gvt_gtt_entry *e, unsigned long index,
+		bool guest)
+{
+	struct intel_gvt *gvt = spt->vgpu->gvt;
+	const struct intel_gvt_gtt_pte_ops *ops = gvt->gtt.pte_ops;
+
+	if (WARN(!gtt_type_is_entry(e->type), "invalid entry type\n"))
+		return -EINVAL;
+
+	gvt_vdbg_mm("set ppgtt entry, spt type %d, entry type %d, index %lu, value %llx\n",
+		    type, e->type, index, e->val64);
+
+	return ops->set_entry(page_table, e, index, guest,
+			spt->guest_page.gfn << I915_GTT_PAGE_SHIFT,
+			spt->vgpu);
+}
+
+#define ppgtt_get_guest_entry(spt, e, index) \
+	ppgtt_spt_get_entry(spt, NULL, \
+		spt->guest_page.type, e, index, true)
+
+#define ppgtt_set_guest_entry(spt, e, index) \
+	ppgtt_spt_set_entry(spt, NULL, \
+		spt->guest_page.type, e, index, true)
+
+#define ppgtt_get_shadow_entry(spt, e, index) \
+	ppgtt_spt_get_entry(spt, spt->shadow_page.vaddr, \
+		spt->shadow_page.type, e, index, false)
+
+#define ppgtt_set_shadow_entry(spt, e, index) \
+	ppgtt_spt_set_entry(spt, spt->shadow_page.vaddr, \
+		spt->shadow_page.type, e, index, false)
+
+static void *alloc_spt(gfp_t gfp_mask)
+{
+	struct intel_vgpu_ppgtt_spt *spt;
+
+	spt = kzalloc(sizeof(*spt), gfp_mask);
+	if (!spt)
+		return NULL;
+
+	spt->shadow_page.page = alloc_page(gfp_mask);
+	if (!spt->shadow_page.page) {
+		kfree(spt);
+		return NULL;
+	}
+	return spt;
+}
+
+static void free_spt(struct intel_vgpu_ppgtt_spt *spt)
+{
+	__free_page(spt->shadow_page.page);
+	kfree(spt);
+}
+
+static int detach_oos_page(struct intel_vgpu *vgpu,
+		struct intel_vgpu_oos_page *oos_page);
+
+static void ppgtt_free_spt(struct intel_vgpu_ppgtt_spt *spt)
+{
+	struct device *kdev = spt->vgpu->gvt->gt->i915->drm.dev;
+
+	trace_spt_free(spt->vgpu->id, spt, spt->guest_page.type);
+
+	dma_unmap_page(kdev, spt->shadow_page.mfn << I915_GTT_PAGE_SHIFT, 4096,
+		       DMA_BIDIRECTIONAL);
+
+	radix_tree_delete(&spt->vgpu->gtt.spt_tree, spt->shadow_page.mfn);
+
+	if (spt->guest_page.gfn) {
+		if (spt->guest_page.oos_page)
+			detach_oos_page(spt->vgpu, spt->guest_page.oos_page);
+
+		intel_vgpu_unregister_page_track(spt->vgpu, spt->guest_page.gfn);
+	}
+
+	list_del_init(&spt->post_shadow_list);
+	free_spt(spt);
+}
+
+static void ppgtt_free_all_spt(struct intel_vgpu *vgpu)
+{
+	struct intel_vgpu_ppgtt_spt *spt, *spn;
+	struct radix_tree_iter iter;
+	LIST_HEAD(all_spt);
+	void __rcu **slot;
+
+	rcu_read_lock();
+	radix_tree_for_each_slot(slot, &vgpu->gtt.spt_tree, &iter, 0) {
+		spt = radix_tree_deref_slot(slot);
+		list_move(&spt->post_shadow_list, &all_spt);
+	}
+	rcu_read_unlock();
+
+	list_for_each_entry_safe(spt, spn, &all_spt, post_shadow_list)
+		ppgtt_free_spt(spt);
+}
+
+static int ppgtt_handle_guest_write_page_table_bytes(
+		struct intel_vgpu_ppgtt_spt *spt,
+		u64 pa, void *p_data, int bytes);
+
+static int ppgtt_write_protection_handler(
+		struct intel_vgpu_page_track *page_track,
+		u64 gpa, void *data, int bytes)
+{
+	struct intel_vgpu_ppgtt_spt *spt = page_track->priv_data;
+
+	int ret;
+
+	if (bytes != 4 && bytes != 8)
+		return -EINVAL;
+
+	ret = ppgtt_handle_guest_write_page_table_bytes(spt, gpa, data, bytes);
+	if (ret)
+		return ret;
+	return ret;
+}
+
+/* Find a spt by guest gfn. */
+static struct intel_vgpu_ppgtt_spt *intel_vgpu_find_spt_by_gfn(
+		struct intel_vgpu *vgpu, unsigned long gfn)
+{
+	struct intel_vgpu_page_track *track;
+
+	track = intel_vgpu_find_page_track(vgpu, gfn);
+	if (track && track->handler == ppgtt_write_protection_handler)
+		return track->priv_data;
+
+	return NULL;
+}
+
+/* Find the spt by shadow page mfn. */
+static inline struct intel_vgpu_ppgtt_spt *intel_vgpu_find_spt_by_mfn(
+		struct intel_vgpu *vgpu, unsigned long mfn)
+{
+	return radix_tree_lookup(&vgpu->gtt.spt_tree, mfn);
+}
+
+static int reclaim_one_ppgtt_mm(struct intel_gvt *gvt);
+
+/* Allocate shadow page table without guest page. */
+static struct intel_vgpu_ppgtt_spt *ppgtt_alloc_spt(
+		struct intel_vgpu *vgpu, enum intel_gvt_gtt_type type)
+{
+	struct device *kdev = vgpu->gvt->gt->i915->drm.dev;
+	struct intel_vgpu_ppgtt_spt *spt = NULL;
+	dma_addr_t daddr;
+	int ret;
+
+retry:
+	spt = alloc_spt(GFP_KERNEL | __GFP_ZERO);
+	if (!spt) {
+		if (reclaim_one_ppgtt_mm(vgpu->gvt))
+			goto retry;
+
+		gvt_vgpu_err("fail to allocate ppgtt shadow page\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	spt->vgpu = vgpu;
+	atomic_set(&spt->refcount, 1);
+	INIT_LIST_HEAD(&spt->post_shadow_list);
+
+	/*
+	 * Init shadow_page.
+	 */
+	spt->shadow_page.type = type;
+	daddr = dma_map_page(kdev, spt->shadow_page.page,
+			     0, 4096, DMA_BIDIRECTIONAL);
+	if (dma_mapping_error(kdev, daddr)) {
+		gvt_vgpu_err("fail to map dma addr\n");
+		ret = -EINVAL;
+		goto err_free_spt;
+	}
+	spt->shadow_page.vaddr = page_address(spt->shadow_page.page);
+	spt->shadow_page.mfn = daddr >> I915_GTT_PAGE_SHIFT;
+
+	ret = radix_tree_insert(&vgpu->gtt.spt_tree, spt->shadow_page.mfn, spt);
+	if (ret)
+		goto err_unmap_dma;
+
+	return spt;
+
+err_unmap_dma:
+	dma_unmap_page(kdev, daddr, PAGE_SIZE, DMA_BIDIRECTIONAL);
+err_free_spt:
+	free_spt(spt);
+	return ERR_PTR(ret);
+}
+
+/* Allocate shadow page table associated with specific gfn. */
+static struct intel_vgpu_ppgtt_spt *ppgtt_alloc_spt_gfn(
+		struct intel_vgpu *vgpu, enum intel_gvt_gtt_type type,
+		unsigned long gfn, bool guest_pde_ips)
+{
+	struct intel_vgpu_ppgtt_spt *spt;
+	int ret;
+
+	spt = ppgtt_alloc_spt(vgpu, type);
+	if (IS_ERR(spt))
+		return spt;
+
+	/*
+	 * Init guest_page.
+	 */
+	ret = intel_vgpu_register_page_track(vgpu, gfn,
+			ppgtt_write_protection_handler, spt);
+	if (ret) {
+		ppgtt_free_spt(spt);
+		return ERR_PTR(ret);
+	}
+
+	spt->guest_page.type = type;
+	spt->guest_page.gfn = gfn;
+	spt->guest_page.pde_ips = guest_pde_ips;
+
+	trace_spt_alloc(vgpu->id, spt, type, spt->shadow_page.mfn, gfn);
+
+	return spt;
+}
+
+#define pt_entry_size_shift(spt) \
+	((spt)->vgpu->gvt->device_info.gtt_entry_size_shift)
+
+#define pt_entries(spt) \
+	(I915_GTT_PAGE_SIZE >> pt_entry_size_shift(spt))
+
+#define for_each_present_guest_entry(spt, e, i) \
+	for (i = 0; i < pt_entries(spt); \
+	     i += spt->guest_page.pde_ips ? GTT_64K_PTE_STRIDE : 1) \
+		if (!ppgtt_get_guest_entry(spt, e, i) && \
+		    spt->vgpu->gvt->gtt.pte_ops->test_present(e))
+
+#define for_each_present_shadow_entry(spt, e, i) \
+	for (i = 0; i < pt_entries(spt); \
+	     i += spt->shadow_page.pde_ips ? GTT_64K_PTE_STRIDE : 1) \
+		if (!ppgtt_get_shadow_entry(spt, e, i) && \
+		    spt->vgpu->gvt->gtt.pte_ops->test_present(e))
+
+#define for_each_shadow_entry(spt, e, i) \
+	for (i = 0; i < pt_entries(spt); \
+	     i += (spt->shadow_page.pde_ips ? GTT_64K_PTE_STRIDE : 1)) \
+		if (!ppgtt_get_shadow_entry(spt, e, i))
+
+static inline void ppgtt_get_spt(struct intel_vgpu_ppgtt_spt *spt)
+{
+	int v = atomic_read(&spt->refcount);
+
+	trace_spt_refcount(spt->vgpu->id, "inc", spt, v, (v + 1));
+	atomic_inc(&spt->refcount);
+}
+
+static inline int ppgtt_put_spt(struct intel_vgpu_ppgtt_spt *spt)
+{
+	int v = atomic_read(&spt->refcount);
+
+	trace_spt_refcount(spt->vgpu->id, "dec", spt, v, (v - 1));
+	return atomic_dec_return(&spt->refcount);
+}
+
+static int ppgtt_invalidate_spt(struct intel_vgpu_ppgtt_spt *spt);
+
+static int ppgtt_invalidate_spt_by_shadow_entry(struct intel_vgpu *vgpu,
+		struct intel_gvt_gtt_entry *e)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	struct intel_vgpu_ppgtt_spt *s;
+	enum intel_gvt_gtt_type cur_pt_type;
+
+	GEM_BUG_ON(!gtt_type_is_pt(get_next_pt_type(e->type)));
+
+	if (e->type != GTT_TYPE_PPGTT_ROOT_L3_ENTRY
+		&& e->type != GTT_TYPE_PPGTT_ROOT_L4_ENTRY) {
+		cur_pt_type = get_next_pt_type(e->type);
+
+		if (!gtt_type_is_pt(cur_pt_type) ||
+				!gtt_type_is_pt(cur_pt_type + 1)) {
+			drm_WARN(&i915->drm, 1,
+				 "Invalid page table type, cur_pt_type is: %d\n",
+				 cur_pt_type);
+			return -EINVAL;
+		}
+
+		cur_pt_type += 1;
+
+		if (ops->get_pfn(e) ==
+			vgpu->gtt.scratch_pt[cur_pt_type].page_mfn)
+			return 0;
+	}
+	s = intel_vgpu_find_spt_by_mfn(vgpu, ops->get_pfn(e));
+	if (!s) {
+		gvt_vgpu_err("fail to find shadow page: mfn: 0x%lx\n",
+				ops->get_pfn(e));
+		return -ENXIO;
+	}
+	return ppgtt_invalidate_spt(s);
+}
+
+static inline void ppgtt_invalidate_pte(struct intel_vgpu_ppgtt_spt *spt,
+		struct intel_gvt_gtt_entry *entry)
+{
+	struct intel_vgpu *vgpu = spt->vgpu;
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	unsigned long pfn;
+	int type;
+
+	pfn = ops->get_pfn(entry);
+	type = spt->shadow_page.type;
+
+	/* Uninitialized spte or unshadowed spte. */
+	if (!pfn || pfn == vgpu->gtt.scratch_pt[type].page_mfn)
+		return;
+
+	intel_gvt_dma_unmap_guest_page(vgpu, pfn << PAGE_SHIFT);
+}
+
+static int ppgtt_invalidate_spt(struct intel_vgpu_ppgtt_spt *spt)
+{
+	struct intel_vgpu *vgpu = spt->vgpu;
+	struct intel_gvt_gtt_entry e;
+	unsigned long index;
+	int ret;
+
+	trace_spt_change(spt->vgpu->id, "die", spt,
+			spt->guest_page.gfn, spt->shadow_page.type);
+
+	if (ppgtt_put_spt(spt) > 0)
+		return 0;
+
+	for_each_present_shadow_entry(spt, &e, index) {
+		switch (e.type) {
+		case GTT_TYPE_PPGTT_PTE_4K_ENTRY:
+			gvt_vdbg_mm("invalidate 4K entry\n");
+			ppgtt_invalidate_pte(spt, &e);
+			break;
+		case GTT_TYPE_PPGTT_PTE_64K_ENTRY:
+			/* We don't setup 64K shadow entry so far. */
+			WARN(1, "suspicious 64K gtt entry\n");
+			continue;
+		case GTT_TYPE_PPGTT_PTE_2M_ENTRY:
+			gvt_vdbg_mm("invalidate 2M entry\n");
+			continue;
+		case GTT_TYPE_PPGTT_PTE_1G_ENTRY:
+			WARN(1, "GVT doesn't support 1GB page\n");
+			continue;
+		case GTT_TYPE_PPGTT_PML4_ENTRY:
+		case GTT_TYPE_PPGTT_PDP_ENTRY:
+		case GTT_TYPE_PPGTT_PDE_ENTRY:
+			gvt_vdbg_mm("invalidate PMUL4/PDP/PDE entry\n");
+			ret = ppgtt_invalidate_spt_by_shadow_entry(
+					spt->vgpu, &e);
+			if (ret)
+				goto fail;
+			break;
+		default:
+			GEM_BUG_ON(1);
+		}
+	}
+
+	trace_spt_change(spt->vgpu->id, "release", spt,
+			 spt->guest_page.gfn, spt->shadow_page.type);
+	ppgtt_free_spt(spt);
+	return 0;
+fail:
+	gvt_vgpu_err("fail: shadow page %p shadow entry 0x%llx type %d\n",
+			spt, e.val64, e.type);
+	return ret;
+}
+
+static bool vgpu_ips_enabled(struct intel_vgpu *vgpu)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+
+	if (GRAPHICS_VER(dev_priv) == 9) {
+		u32 ips = vgpu_vreg_t(vgpu, GEN8_GAMW_ECO_DEV_RW_IA) &
+			GAMW_ECO_ENABLE_64K_IPS_FIELD;
+
+		return ips == GAMW_ECO_ENABLE_64K_IPS_FIELD;
+	} else if (GRAPHICS_VER(dev_priv) >= 11) {
+		/* 64K paging only controlled by IPS bit in PTE now. */
+		return true;
+	} else
+		return false;
+}
+
+static int ppgtt_populate_spt(struct intel_vgpu_ppgtt_spt *spt);
+
+static struct intel_vgpu_ppgtt_spt *ppgtt_populate_spt_by_guest_entry(
+		struct intel_vgpu *vgpu, struct intel_gvt_gtt_entry *we)
+{
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	struct intel_vgpu_ppgtt_spt *spt = NULL;
+	bool ips = false;
+	int ret;
+
+	GEM_BUG_ON(!gtt_type_is_pt(get_next_pt_type(we->type)));
+
+	if (we->type == GTT_TYPE_PPGTT_PDE_ENTRY)
+		ips = vgpu_ips_enabled(vgpu) && ops->test_ips(we);
+
+	spt = intel_vgpu_find_spt_by_gfn(vgpu, ops->get_pfn(we));
+	if (spt) {
+		ppgtt_get_spt(spt);
+
+		if (ips != spt->guest_page.pde_ips) {
+			spt->guest_page.pde_ips = ips;
+
+			gvt_dbg_mm("reshadow PDE since ips changed\n");
+			clear_page(spt->shadow_page.vaddr);
+			ret = ppgtt_populate_spt(spt);
+			if (ret) {
+				ppgtt_put_spt(spt);
+				goto err;
+			}
+		}
+	} else {
+		int type = get_next_pt_type(we->type);
+
+		if (!gtt_type_is_pt(type)) {
+			ret = -EINVAL;
+			goto err;
+		}
+
+		spt = ppgtt_alloc_spt_gfn(vgpu, type, ops->get_pfn(we), ips);
+		if (IS_ERR(spt)) {
+			ret = PTR_ERR(spt);
+			goto err;
+		}
+
+		ret = intel_vgpu_enable_page_track(vgpu, spt->guest_page.gfn);
+		if (ret)
+			goto err_free_spt;
+
+		ret = ppgtt_populate_spt(spt);
+		if (ret)
+			goto err_free_spt;
+
+		trace_spt_change(vgpu->id, "new", spt, spt->guest_page.gfn,
+				 spt->shadow_page.type);
+	}
+	return spt;
+
+err_free_spt:
+	ppgtt_free_spt(spt);
+	spt = NULL;
+err:
+	gvt_vgpu_err("fail: shadow page %p guest entry 0x%llx type %d\n",
+		     spt, we->val64, we->type);
+	return ERR_PTR(ret);
+}
+
+static inline void ppgtt_generate_shadow_entry(struct intel_gvt_gtt_entry *se,
+		struct intel_vgpu_ppgtt_spt *s, struct intel_gvt_gtt_entry *ge)
+{
+	const struct intel_gvt_gtt_pte_ops *ops = s->vgpu->gvt->gtt.pte_ops;
+
+	se->type = ge->type;
+	se->val64 = ge->val64;
+
+	/* Because we always split 64KB pages, so clear IPS in shadow PDE. */
+	if (se->type == GTT_TYPE_PPGTT_PDE_ENTRY)
+		ops->clear_ips(se);
+
+	ops->set_pfn(se, s->shadow_page.mfn);
+}
+
+/*
+ * Check if can do 2M page
+ * @vgpu: target vgpu
+ * @entry: target pfn's gtt entry
+ *
+ * Return 1 if 2MB huge gtt shadowing is possible, 0 if miscondition,
+ * negative if found err.
+ */
+static int is_2MB_gtt_possible(struct intel_vgpu *vgpu,
+	struct intel_gvt_gtt_entry *entry)
+{
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	kvm_pfn_t pfn;
+	int ret;
+
+	if (!HAS_PAGE_SIZES(vgpu->gvt->gt->i915, I915_GTT_PAGE_SIZE_2M))
+		return 0;
+
+	if (!vgpu->attached)
+		return -EINVAL;
+	pfn = gfn_to_pfn(vgpu->vfio_device.kvm, ops->get_pfn(entry));
+	if (is_error_noslot_pfn(pfn))
+		return -EINVAL;
+
+	if (!pfn_valid(pfn))
+		return -EINVAL;
+
+	ret = PageTransHuge(pfn_to_page(pfn));
+	kvm_release_pfn_clean(pfn);
+	return ret;
+}
+
+static int split_2MB_gtt_entry(struct intel_vgpu *vgpu,
+	struct intel_vgpu_ppgtt_spt *spt, unsigned long index,
+	struct intel_gvt_gtt_entry *se)
+{
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	struct intel_vgpu_ppgtt_spt *sub_spt;
+	struct intel_gvt_gtt_entry sub_se;
+	unsigned long start_gfn;
+	dma_addr_t dma_addr;
+	unsigned long sub_index;
+	int ret;
+
+	gvt_dbg_mm("Split 2M gtt entry, index %lu\n", index);
+
+	start_gfn = ops->get_pfn(se);
+
+	sub_spt = ppgtt_alloc_spt(vgpu, GTT_TYPE_PPGTT_PTE_PT);
+	if (IS_ERR(sub_spt))
+		return PTR_ERR(sub_spt);
+
+	for_each_shadow_entry(sub_spt, &sub_se, sub_index) {
+		ret = intel_gvt_dma_map_guest_page(vgpu, start_gfn + sub_index,
+						   PAGE_SIZE, &dma_addr);
+		if (ret)
+			goto err;
+		sub_se.val64 = se->val64;
+
+		/* Copy the PAT field from PDE. */
+		sub_se.val64 &= ~_PAGE_PAT;
+		sub_se.val64 |= (se->val64 & _PAGE_PAT_LARGE) >> 5;
+
+		ops->set_pfn(&sub_se, dma_addr >> PAGE_SHIFT);
+		ppgtt_set_shadow_entry(sub_spt, &sub_se, sub_index);
+	}
+
+	/* Clear dirty field. */
+	se->val64 &= ~_PAGE_DIRTY;
+
+	ops->clear_pse(se);
+	ops->clear_ips(se);
+	ops->set_pfn(se, sub_spt->shadow_page.mfn);
+	ppgtt_set_shadow_entry(spt, se, index);
+	return 0;
+err:
+	/* Cancel the existing addess mappings of DMA addr. */
+	for_each_present_shadow_entry(sub_spt, &sub_se, sub_index) {
+		gvt_vdbg_mm("invalidate 4K entry\n");
+		ppgtt_invalidate_pte(sub_spt, &sub_se);
+	}
+	/* Release the new allocated spt. */
+	trace_spt_change(sub_spt->vgpu->id, "release", sub_spt,
+		sub_spt->guest_page.gfn, sub_spt->shadow_page.type);
+	ppgtt_free_spt(sub_spt);
+	return ret;
+}
+
+static int split_64KB_gtt_entry(struct intel_vgpu *vgpu,
+	struct intel_vgpu_ppgtt_spt *spt, unsigned long index,
+	struct intel_gvt_gtt_entry *se)
+{
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	struct intel_gvt_gtt_entry entry = *se;
+	unsigned long start_gfn;
+	dma_addr_t dma_addr;
+	int i, ret;
+
+	gvt_vdbg_mm("Split 64K gtt entry, index %lu\n", index);
+
+	GEM_BUG_ON(index % GTT_64K_PTE_STRIDE);
+
+	start_gfn = ops->get_pfn(se);
+
+	entry.type = GTT_TYPE_PPGTT_PTE_4K_ENTRY;
+	ops->set_64k_splited(&entry);
+
+	for (i = 0; i < GTT_64K_PTE_STRIDE; i++) {
+		ret = intel_gvt_dma_map_guest_page(vgpu, start_gfn + i,
+						   PAGE_SIZE, &dma_addr);
+		if (ret)
+			return ret;
+
+		ops->set_pfn(&entry, dma_addr >> PAGE_SHIFT);
+		ppgtt_set_shadow_entry(spt, &entry, index + i);
+	}
+	return 0;
+}
+
+static int ppgtt_populate_shadow_entry(struct intel_vgpu *vgpu,
+	struct intel_vgpu_ppgtt_spt *spt, unsigned long index,
+	struct intel_gvt_gtt_entry *ge)
+{
+	const struct intel_gvt_gtt_pte_ops *pte_ops = vgpu->gvt->gtt.pte_ops;
+	struct intel_gvt_gtt_entry se = *ge;
+	unsigned long gfn, page_size = PAGE_SIZE;
+	dma_addr_t dma_addr;
+	int ret;
+
+	if (!pte_ops->test_present(ge))
+		return 0;
+
+	gfn = pte_ops->get_pfn(ge);
+
+	switch (ge->type) {
+	case GTT_TYPE_PPGTT_PTE_4K_ENTRY:
+		gvt_vdbg_mm("shadow 4K gtt entry\n");
+		break;
+	case GTT_TYPE_PPGTT_PTE_64K_ENTRY:
+		gvt_vdbg_mm("shadow 64K gtt entry\n");
+		/*
+		 * The layout of 64K page is special, the page size is
+		 * controlled by uper PDE. To be simple, we always split
+		 * 64K page to smaller 4K pages in shadow PT.
+		 */
+		return split_64KB_gtt_entry(vgpu, spt, index, &se);
+	case GTT_TYPE_PPGTT_PTE_2M_ENTRY:
+		gvt_vdbg_mm("shadow 2M gtt entry\n");
+		ret = is_2MB_gtt_possible(vgpu, ge);
+		if (ret == 0)
+			return split_2MB_gtt_entry(vgpu, spt, index, &se);
+		else if (ret < 0)
+			return ret;
+		page_size = I915_GTT_PAGE_SIZE_2M;
+		break;
+	case GTT_TYPE_PPGTT_PTE_1G_ENTRY:
+		gvt_vgpu_err("GVT doesn't support 1GB entry\n");
+		return -EINVAL;
+	default:
+		GEM_BUG_ON(1);
+	}
+
+	/* direct shadow */
+	ret = intel_gvt_dma_map_guest_page(vgpu, gfn, page_size, &dma_addr);
+	if (ret)
+		return -ENXIO;
+
+	pte_ops->set_pfn(&se, dma_addr >> PAGE_SHIFT);
+	ppgtt_set_shadow_entry(spt, &se, index);
+	return 0;
+}
+
+static int ppgtt_populate_spt(struct intel_vgpu_ppgtt_spt *spt)
+{
+	struct intel_vgpu *vgpu = spt->vgpu;
+	struct intel_gvt *gvt = vgpu->gvt;
+	const struct intel_gvt_gtt_pte_ops *ops = gvt->gtt.pte_ops;
+	struct intel_vgpu_ppgtt_spt *s;
+	struct intel_gvt_gtt_entry se, ge;
+	unsigned long gfn, i;
+	int ret;
+
+	trace_spt_change(spt->vgpu->id, "born", spt,
+			 spt->guest_page.gfn, spt->shadow_page.type);
+
+	for_each_present_guest_entry(spt, &ge, i) {
+		if (gtt_type_is_pt(get_next_pt_type(ge.type))) {
+			s = ppgtt_populate_spt_by_guest_entry(vgpu, &ge);
+			if (IS_ERR(s)) {
+				ret = PTR_ERR(s);
+				goto fail;
+			}
+			ppgtt_get_shadow_entry(spt, &se, i);
+			ppgtt_generate_shadow_entry(&se, s, &ge);
+			ppgtt_set_shadow_entry(spt, &se, i);
+		} else {
+			gfn = ops->get_pfn(&ge);
+			if (!intel_gvt_is_valid_gfn(vgpu, gfn)) {
+				ops->set_pfn(&se, gvt->gtt.scratch_mfn);
+				ppgtt_set_shadow_entry(spt, &se, i);
+				continue;
+			}
+
+			ret = ppgtt_populate_shadow_entry(vgpu, spt, i, &ge);
+			if (ret)
+				goto fail;
+		}
+	}
+	return 0;
+fail:
+	gvt_vgpu_err("fail: shadow page %p guest entry 0x%llx type %d\n",
+			spt, ge.val64, ge.type);
+	return ret;
+}
+
+static int ppgtt_handle_guest_entry_removal(struct intel_vgpu_ppgtt_spt *spt,
+		struct intel_gvt_gtt_entry *se, unsigned long index)
+{
+	struct intel_vgpu *vgpu = spt->vgpu;
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	int ret;
+
+	trace_spt_guest_change(spt->vgpu->id, "remove", spt,
+			       spt->shadow_page.type, se->val64, index);
+
+	gvt_vdbg_mm("destroy old shadow entry, type %d, index %lu, value %llx\n",
+		    se->type, index, se->val64);
+
+	if (!ops->test_present(se))
+		return 0;
+
+	if (ops->get_pfn(se) ==
+	    vgpu->gtt.scratch_pt[spt->shadow_page.type].page_mfn)
+		return 0;
+
+	if (gtt_type_is_pt(get_next_pt_type(se->type))) {
+		struct intel_vgpu_ppgtt_spt *s =
+			intel_vgpu_find_spt_by_mfn(vgpu, ops->get_pfn(se));
+		if (!s) {
+			gvt_vgpu_err("fail to find guest page\n");
+			ret = -ENXIO;
+			goto fail;
+		}
+		ret = ppgtt_invalidate_spt(s);
+		if (ret)
+			goto fail;
+	} else {
+		/* We don't setup 64K shadow entry so far. */
+		WARN(se->type == GTT_TYPE_PPGTT_PTE_64K_ENTRY,
+		     "suspicious 64K entry\n");
+		ppgtt_invalidate_pte(spt, se);
+	}
+
+	return 0;
+fail:
+	gvt_vgpu_err("fail: shadow page %p guest entry 0x%llx type %d\n",
+			spt, se->val64, se->type);
+	return ret;
+}
+
+static int ppgtt_handle_guest_entry_add(struct intel_vgpu_ppgtt_spt *spt,
+		struct intel_gvt_gtt_entry *we, unsigned long index)
+{
+	struct intel_vgpu *vgpu = spt->vgpu;
+	struct intel_gvt_gtt_entry m;
+	struct intel_vgpu_ppgtt_spt *s;
+	int ret;
+
+	trace_spt_guest_change(spt->vgpu->id, "add", spt, spt->shadow_page.type,
+			       we->val64, index);
+
+	gvt_vdbg_mm("add shadow entry: type %d, index %lu, value %llx\n",
+		    we->type, index, we->val64);
+
+	if (gtt_type_is_pt(get_next_pt_type(we->type))) {
+		s = ppgtt_populate_spt_by_guest_entry(vgpu, we);
+		if (IS_ERR(s)) {
+			ret = PTR_ERR(s);
+			goto fail;
+		}
+		ppgtt_get_shadow_entry(spt, &m, index);
+		ppgtt_generate_shadow_entry(&m, s, we);
+		ppgtt_set_shadow_entry(spt, &m, index);
+	} else {
+		ret = ppgtt_populate_shadow_entry(vgpu, spt, index, we);
+		if (ret)
+			goto fail;
+	}
+	return 0;
+fail:
+	gvt_vgpu_err("fail: spt %p guest entry 0x%llx type %d\n",
+		spt, we->val64, we->type);
+	return ret;
+}
+
+static int sync_oos_page(struct intel_vgpu *vgpu,
+		struct intel_vgpu_oos_page *oos_page)
+{
+	const struct intel_gvt_device_info *info = &vgpu->gvt->device_info;
+	struct intel_gvt *gvt = vgpu->gvt;
+	const struct intel_gvt_gtt_pte_ops *ops = gvt->gtt.pte_ops;
+	struct intel_vgpu_ppgtt_spt *spt = oos_page->spt;
+	struct intel_gvt_gtt_entry old, new;
+	int index;
+	int ret;
+
+	trace_oos_change(vgpu->id, "sync", oos_page->id,
+			 spt, spt->guest_page.type);
+
+	old.type = new.type = get_entry_type(spt->guest_page.type);
+	old.val64 = new.val64 = 0;
+
+	for (index = 0; index < (I915_GTT_PAGE_SIZE >>
+				info->gtt_entry_size_shift); index++) {
+		ops->get_entry(oos_page->mem, &old, index, false, 0, vgpu);
+		ops->get_entry(NULL, &new, index, true,
+			       spt->guest_page.gfn << PAGE_SHIFT, vgpu);
+
+		if (old.val64 == new.val64
+			&& !test_and_clear_bit(index, spt->post_shadow_bitmap))
+			continue;
+
+		trace_oos_sync(vgpu->id, oos_page->id,
+				spt, spt->guest_page.type,
+				new.val64, index);
+
+		ret = ppgtt_populate_shadow_entry(vgpu, spt, index, &new);
+		if (ret)
+			return ret;
+
+		ops->set_entry(oos_page->mem, &new, index, false, 0, vgpu);
+	}
+
+	spt->guest_page.write_cnt = 0;
+	list_del_init(&spt->post_shadow_list);
+	return 0;
+}
+
+static int detach_oos_page(struct intel_vgpu *vgpu,
+		struct intel_vgpu_oos_page *oos_page)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_vgpu_ppgtt_spt *spt = oos_page->spt;
+
+	trace_oos_change(vgpu->id, "detach", oos_page->id,
+			 spt, spt->guest_page.type);
+
+	spt->guest_page.write_cnt = 0;
+	spt->guest_page.oos_page = NULL;
+	oos_page->spt = NULL;
+
+	list_del_init(&oos_page->vm_list);
+	list_move_tail(&oos_page->list, &gvt->gtt.oos_page_free_list_head);
+
+	return 0;
+}
+
+static int attach_oos_page(struct intel_vgpu_oos_page *oos_page,
+		struct intel_vgpu_ppgtt_spt *spt)
+{
+	struct intel_gvt *gvt = spt->vgpu->gvt;
+	int ret;
+
+	ret = intel_gvt_read_gpa(spt->vgpu,
+			spt->guest_page.gfn << I915_GTT_PAGE_SHIFT,
+			oos_page->mem, I915_GTT_PAGE_SIZE);
+	if (ret)
+		return ret;
+
+	oos_page->spt = spt;
+	spt->guest_page.oos_page = oos_page;
+
+	list_move_tail(&oos_page->list, &gvt->gtt.oos_page_use_list_head);
+
+	trace_oos_change(spt->vgpu->id, "attach", oos_page->id,
+			 spt, spt->guest_page.type);
+	return 0;
+}
+
+static int ppgtt_set_guest_page_sync(struct intel_vgpu_ppgtt_spt *spt)
+{
+	struct intel_vgpu_oos_page *oos_page = spt->guest_page.oos_page;
+	int ret;
+
+	ret = intel_vgpu_enable_page_track(spt->vgpu, spt->guest_page.gfn);
+	if (ret)
+		return ret;
+
+	trace_oos_change(spt->vgpu->id, "set page sync", oos_page->id,
+			 spt, spt->guest_page.type);
+
+	list_del_init(&oos_page->vm_list);
+	return sync_oos_page(spt->vgpu, oos_page);
+}
+
+static int ppgtt_allocate_oos_page(struct intel_vgpu_ppgtt_spt *spt)
+{
+	struct intel_gvt *gvt = spt->vgpu->gvt;
+	struct intel_gvt_gtt *gtt = &gvt->gtt;
+	struct intel_vgpu_oos_page *oos_page = spt->guest_page.oos_page;
+	int ret;
+
+	WARN(oos_page, "shadow PPGTT page has already has a oos page\n");
+
+	if (list_empty(&gtt->oos_page_free_list_head)) {
+		oos_page = container_of(gtt->oos_page_use_list_head.next,
+			struct intel_vgpu_oos_page, list);
+		ret = ppgtt_set_guest_page_sync(oos_page->spt);
+		if (ret)
+			return ret;
+		ret = detach_oos_page(spt->vgpu, oos_page);
+		if (ret)
+			return ret;
+	} else
+		oos_page = container_of(gtt->oos_page_free_list_head.next,
+			struct intel_vgpu_oos_page, list);
+	return attach_oos_page(oos_page, spt);
+}
+
+static int ppgtt_set_guest_page_oos(struct intel_vgpu_ppgtt_spt *spt)
+{
+	struct intel_vgpu_oos_page *oos_page = spt->guest_page.oos_page;
+
+	if (WARN(!oos_page, "shadow PPGTT page should have a oos page\n"))
+		return -EINVAL;
+
+	trace_oos_change(spt->vgpu->id, "set page out of sync", oos_page->id,
+			 spt, spt->guest_page.type);
+
+	list_add_tail(&oos_page->vm_list, &spt->vgpu->gtt.oos_page_list_head);
+	return intel_vgpu_disable_page_track(spt->vgpu, spt->guest_page.gfn);
+}
+
+/**
+ * intel_vgpu_sync_oos_pages - sync all the out-of-synced shadow for vGPU
+ * @vgpu: a vGPU
+ *
+ * This function is called before submitting a guest workload to host,
+ * to sync all the out-of-synced shadow for vGPU
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_sync_oos_pages(struct intel_vgpu *vgpu)
+{
+	struct list_head *pos, *n;
+	struct intel_vgpu_oos_page *oos_page;
+	int ret;
+
+	if (!enable_out_of_sync)
+		return 0;
+
+	list_for_each_safe(pos, n, &vgpu->gtt.oos_page_list_head) {
+		oos_page = container_of(pos,
+				struct intel_vgpu_oos_page, vm_list);
+		ret = ppgtt_set_guest_page_sync(oos_page->spt);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+/*
+ * The heart of PPGTT shadow page table.
+ */
+static int ppgtt_handle_guest_write_page_table(
+		struct intel_vgpu_ppgtt_spt *spt,
+		struct intel_gvt_gtt_entry *we, unsigned long index)
+{
+	struct intel_vgpu *vgpu = spt->vgpu;
+	int type = spt->shadow_page.type;
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	struct intel_gvt_gtt_entry old_se;
+	int new_present;
+	int i, ret;
+
+	new_present = ops->test_present(we);
+
+	/*
+	 * Adding the new entry first and then removing the old one, that can
+	 * guarantee the ppgtt table is validated during the window between
+	 * adding and removal.
+	 */
+	ppgtt_get_shadow_entry(spt, &old_se, index);
+
+	if (new_present) {
+		ret = ppgtt_handle_guest_entry_add(spt, we, index);
+		if (ret)
+			goto fail;
+	}
+
+	ret = ppgtt_handle_guest_entry_removal(spt, &old_se, index);
+	if (ret)
+		goto fail;
+
+	if (!new_present) {
+		/* For 64KB splited entries, we need clear them all. */
+		if (ops->test_64k_splited(&old_se) &&
+		    !(index % GTT_64K_PTE_STRIDE)) {
+			gvt_vdbg_mm("remove splited 64K shadow entries\n");
+			for (i = 0; i < GTT_64K_PTE_STRIDE; i++) {
+				ops->clear_64k_splited(&old_se);
+				ops->set_pfn(&old_se,
+					vgpu->gtt.scratch_pt[type].page_mfn);
+				ppgtt_set_shadow_entry(spt, &old_se, index + i);
+			}
+		} else if (old_se.type == GTT_TYPE_PPGTT_PTE_2M_ENTRY ||
+			   old_se.type == GTT_TYPE_PPGTT_PTE_1G_ENTRY) {
+			ops->clear_pse(&old_se);
+			ops->set_pfn(&old_se,
+				     vgpu->gtt.scratch_pt[type].page_mfn);
+			ppgtt_set_shadow_entry(spt, &old_se, index);
+		} else {
+			ops->set_pfn(&old_se,
+				     vgpu->gtt.scratch_pt[type].page_mfn);
+			ppgtt_set_shadow_entry(spt, &old_se, index);
+		}
+	}
+
+	return 0;
+fail:
+	gvt_vgpu_err("fail: shadow page %p guest entry 0x%llx type %d.\n",
+			spt, we->val64, we->type);
+	return ret;
+}
+
+
+
+static inline bool can_do_out_of_sync(struct intel_vgpu_ppgtt_spt *spt)
+{
+	return enable_out_of_sync
+		&& gtt_type_is_pte_pt(spt->guest_page.type)
+		&& spt->guest_page.write_cnt >= 2;
+}
+
+static void ppgtt_set_post_shadow(struct intel_vgpu_ppgtt_spt *spt,
+		unsigned long index)
+{
+	set_bit(index, spt->post_shadow_bitmap);
+	if (!list_empty(&spt->post_shadow_list))
+		return;
+
+	list_add_tail(&spt->post_shadow_list,
+			&spt->vgpu->gtt.post_shadow_list_head);
+}
+
+/**
+ * intel_vgpu_flush_post_shadow - flush the post shadow transactions
+ * @vgpu: a vGPU
+ *
+ * This function is called before submitting a guest workload to host,
+ * to flush all the post shadows for a vGPU.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_flush_post_shadow(struct intel_vgpu *vgpu)
+{
+	struct list_head *pos, *n;
+	struct intel_vgpu_ppgtt_spt *spt;
+	struct intel_gvt_gtt_entry ge;
+	unsigned long index;
+	int ret;
+
+	list_for_each_safe(pos, n, &vgpu->gtt.post_shadow_list_head) {
+		spt = container_of(pos, struct intel_vgpu_ppgtt_spt,
+				post_shadow_list);
+
+		for_each_set_bit(index, spt->post_shadow_bitmap,
+				GTT_ENTRY_NUM_IN_ONE_PAGE) {
+			ppgtt_get_guest_entry(spt, &ge, index);
+
+			ret = ppgtt_handle_guest_write_page_table(spt,
+							&ge, index);
+			if (ret)
+				return ret;
+			clear_bit(index, spt->post_shadow_bitmap);
+		}
+		list_del_init(&spt->post_shadow_list);
+	}
+	return 0;
+}
+
+static int ppgtt_handle_guest_write_page_table_bytes(
+		struct intel_vgpu_ppgtt_spt *spt,
+		u64 pa, void *p_data, int bytes)
+{
+	struct intel_vgpu *vgpu = spt->vgpu;
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	const struct intel_gvt_device_info *info = &vgpu->gvt->device_info;
+	struct intel_gvt_gtt_entry we, se;
+	unsigned long index;
+	int ret;
+
+	index = (pa & (PAGE_SIZE - 1)) >> info->gtt_entry_size_shift;
+
+	ppgtt_get_guest_entry(spt, &we, index);
+
+	/*
+	 * For page table which has 64K gtt entry, only PTE#0, PTE#16,
+	 * PTE#32, ... PTE#496 are used. Unused PTEs update should be
+	 * ignored.
+	 */
+	if (we.type == GTT_TYPE_PPGTT_PTE_64K_ENTRY &&
+	    (index % GTT_64K_PTE_STRIDE)) {
+		gvt_vdbg_mm("Ignore write to unused PTE entry, index %lu\n",
+			    index);
+		return 0;
+	}
+
+	if (bytes == info->gtt_entry_size) {
+		ret = ppgtt_handle_guest_write_page_table(spt, &we, index);
+		if (ret)
+			return ret;
+	} else {
+		if (!test_bit(index, spt->post_shadow_bitmap)) {
+			int type = spt->shadow_page.type;
+
+			ppgtt_get_shadow_entry(spt, &se, index);
+			ret = ppgtt_handle_guest_entry_removal(spt, &se, index);
+			if (ret)
+				return ret;
+			ops->set_pfn(&se, vgpu->gtt.scratch_pt[type].page_mfn);
+			ppgtt_set_shadow_entry(spt, &se, index);
+		}
+		ppgtt_set_post_shadow(spt, index);
+	}
+
+	if (!enable_out_of_sync)
+		return 0;
+
+	spt->guest_page.write_cnt++;
+
+	if (spt->guest_page.oos_page)
+		ops->set_entry(spt->guest_page.oos_page->mem, &we, index,
+				false, 0, vgpu);
+
+	if (can_do_out_of_sync(spt)) {
+		if (!spt->guest_page.oos_page)
+			ppgtt_allocate_oos_page(spt);
+
+		ret = ppgtt_set_guest_page_oos(spt);
+		if (ret < 0)
+			return ret;
+	}
+	return 0;
+}
+
+static void invalidate_ppgtt_mm(struct intel_vgpu_mm *mm)
+{
+	struct intel_vgpu *vgpu = mm->vgpu;
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_gvt_gtt *gtt = &gvt->gtt;
+	const struct intel_gvt_gtt_pte_ops *ops = gtt->pte_ops;
+	struct intel_gvt_gtt_entry se;
+	int index;
+
+	if (!mm->ppgtt_mm.shadowed)
+		return;
+
+	for (index = 0; index < ARRAY_SIZE(mm->ppgtt_mm.shadow_pdps); index++) {
+		ppgtt_get_shadow_root_entry(mm, &se, index);
+
+		if (!ops->test_present(&se))
+			continue;
+
+		ppgtt_invalidate_spt_by_shadow_entry(vgpu, &se);
+		se.val64 = 0;
+		ppgtt_set_shadow_root_entry(mm, &se, index);
+
+		trace_spt_guest_change(vgpu->id, "destroy root pointer",
+				       NULL, se.type, se.val64, index);
+	}
+
+	mm->ppgtt_mm.shadowed = false;
+}
+
+
+static int shadow_ppgtt_mm(struct intel_vgpu_mm *mm)
+{
+	struct intel_vgpu *vgpu = mm->vgpu;
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_gvt_gtt *gtt = &gvt->gtt;
+	const struct intel_gvt_gtt_pte_ops *ops = gtt->pte_ops;
+	struct intel_vgpu_ppgtt_spt *spt;
+	struct intel_gvt_gtt_entry ge, se;
+	int index, ret;
+
+	if (mm->ppgtt_mm.shadowed)
+		return 0;
+
+	mm->ppgtt_mm.shadowed = true;
+
+	for (index = 0; index < ARRAY_SIZE(mm->ppgtt_mm.guest_pdps); index++) {
+		ppgtt_get_guest_root_entry(mm, &ge, index);
+
+		if (!ops->test_present(&ge))
+			continue;
+
+		trace_spt_guest_change(vgpu->id, __func__, NULL,
+				       ge.type, ge.val64, index);
+
+		spt = ppgtt_populate_spt_by_guest_entry(vgpu, &ge);
+		if (IS_ERR(spt)) {
+			gvt_vgpu_err("fail to populate guest root pointer\n");
+			ret = PTR_ERR(spt);
+			goto fail;
+		}
+		ppgtt_generate_shadow_entry(&se, spt, &ge);
+		ppgtt_set_shadow_root_entry(mm, &se, index);
+
+		trace_spt_guest_change(vgpu->id, "populate root pointer",
+				       NULL, se.type, se.val64, index);
+	}
+
+	return 0;
+fail:
+	invalidate_ppgtt_mm(mm);
+	return ret;
+}
+
+static struct intel_vgpu_mm *vgpu_alloc_mm(struct intel_vgpu *vgpu)
+{
+	struct intel_vgpu_mm *mm;
+
+	mm = kzalloc(sizeof(*mm), GFP_KERNEL);
+	if (!mm)
+		return NULL;
+
+	mm->vgpu = vgpu;
+	kref_init(&mm->ref);
+	atomic_set(&mm->pincount, 0);
+
+	return mm;
+}
+
+static void vgpu_free_mm(struct intel_vgpu_mm *mm)
+{
+	kfree(mm);
+}
+
+/**
+ * intel_vgpu_create_ppgtt_mm - create a ppgtt mm object for a vGPU
+ * @vgpu: a vGPU
+ * @root_entry_type: ppgtt root entry type
+ * @pdps: guest pdps.
+ *
+ * This function is used to create a ppgtt mm object for a vGPU.
+ *
+ * Returns:
+ * Zero on success, negative error code in pointer if failed.
+ */
+struct intel_vgpu_mm *intel_vgpu_create_ppgtt_mm(struct intel_vgpu *vgpu,
+		enum intel_gvt_gtt_type root_entry_type, u64 pdps[])
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_vgpu_mm *mm;
+	int ret;
+
+	mm = vgpu_alloc_mm(vgpu);
+	if (!mm)
+		return ERR_PTR(-ENOMEM);
+
+	mm->type = INTEL_GVT_MM_PPGTT;
+
+	GEM_BUG_ON(root_entry_type != GTT_TYPE_PPGTT_ROOT_L3_ENTRY &&
+		   root_entry_type != GTT_TYPE_PPGTT_ROOT_L4_ENTRY);
+	mm->ppgtt_mm.root_entry_type = root_entry_type;
+
+	INIT_LIST_HEAD(&mm->ppgtt_mm.list);
+	INIT_LIST_HEAD(&mm->ppgtt_mm.lru_list);
+	INIT_LIST_HEAD(&mm->ppgtt_mm.link);
+
+	if (root_entry_type == GTT_TYPE_PPGTT_ROOT_L4_ENTRY)
+		mm->ppgtt_mm.guest_pdps[0] = pdps[0];
+	else
+		memcpy(mm->ppgtt_mm.guest_pdps, pdps,
+		       sizeof(mm->ppgtt_mm.guest_pdps));
+
+	ret = shadow_ppgtt_mm(mm);
+	if (ret) {
+		gvt_vgpu_err("failed to shadow ppgtt mm\n");
+		vgpu_free_mm(mm);
+		return ERR_PTR(ret);
+	}
+
+	list_add_tail(&mm->ppgtt_mm.list, &vgpu->gtt.ppgtt_mm_list_head);
+
+	mutex_lock(&gvt->gtt.ppgtt_mm_lock);
+	list_add_tail(&mm->ppgtt_mm.lru_list, &gvt->gtt.ppgtt_mm_lru_list_head);
+	mutex_unlock(&gvt->gtt.ppgtt_mm_lock);
+
+	return mm;
+}
+
+static struct intel_vgpu_mm *intel_vgpu_create_ggtt_mm(struct intel_vgpu *vgpu)
+{
+	struct intel_vgpu_mm *mm;
+	unsigned long nr_entries;
+
+	mm = vgpu_alloc_mm(vgpu);
+	if (!mm)
+		return ERR_PTR(-ENOMEM);
+
+	mm->type = INTEL_GVT_MM_GGTT;
+
+	nr_entries = gvt_ggtt_gm_sz(vgpu->gvt) >> I915_GTT_PAGE_SHIFT;
+	mm->ggtt_mm.virtual_ggtt =
+		vzalloc(array_size(nr_entries,
+				   vgpu->gvt->device_info.gtt_entry_size));
+	if (!mm->ggtt_mm.virtual_ggtt) {
+		vgpu_free_mm(mm);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	mm->ggtt_mm.host_ggtt_aperture = vzalloc((vgpu_aperture_sz(vgpu) >> PAGE_SHIFT) * sizeof(u64));
+	if (!mm->ggtt_mm.host_ggtt_aperture) {
+		vfree(mm->ggtt_mm.virtual_ggtt);
+		vgpu_free_mm(mm);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	mm->ggtt_mm.host_ggtt_hidden = vzalloc((vgpu_hidden_sz(vgpu) >> PAGE_SHIFT) * sizeof(u64));
+	if (!mm->ggtt_mm.host_ggtt_hidden) {
+		vfree(mm->ggtt_mm.host_ggtt_aperture);
+		vfree(mm->ggtt_mm.virtual_ggtt);
+		vgpu_free_mm(mm);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	return mm;
+}
+
+/**
+ * _intel_vgpu_mm_release - destroy a mm object
+ * @mm_ref: a kref object
+ *
+ * This function is used to destroy a mm object for vGPU
+ *
+ */
+void _intel_vgpu_mm_release(struct kref *mm_ref)
+{
+	struct intel_vgpu_mm *mm = container_of(mm_ref, typeof(*mm), ref);
+
+	if (GEM_WARN_ON(atomic_read(&mm->pincount)))
+		gvt_err("vgpu mm pin count bug detected\n");
+
+	if (mm->type == INTEL_GVT_MM_PPGTT) {
+		list_del(&mm->ppgtt_mm.list);
+
+		mutex_lock(&mm->vgpu->gvt->gtt.ppgtt_mm_lock);
+		list_del(&mm->ppgtt_mm.lru_list);
+		mutex_unlock(&mm->vgpu->gvt->gtt.ppgtt_mm_lock);
+
+		invalidate_ppgtt_mm(mm);
+	} else {
+		vfree(mm->ggtt_mm.virtual_ggtt);
+		vfree(mm->ggtt_mm.host_ggtt_aperture);
+		vfree(mm->ggtt_mm.host_ggtt_hidden);
+	}
+
+	vgpu_free_mm(mm);
+}
+
+/**
+ * intel_vgpu_unpin_mm - decrease the pin count of a vGPU mm object
+ * @mm: a vGPU mm object
+ *
+ * This function is called when user doesn't want to use a vGPU mm object
+ */
+void intel_vgpu_unpin_mm(struct intel_vgpu_mm *mm)
+{
+	atomic_dec_if_positive(&mm->pincount);
+}
+
+/**
+ * intel_vgpu_pin_mm - increase the pin count of a vGPU mm object
+ * @mm: target vgpu mm
+ *
+ * This function is called when user wants to use a vGPU mm object. If this
+ * mm object hasn't been shadowed yet, the shadow will be populated at this
+ * time.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_pin_mm(struct intel_vgpu_mm *mm)
+{
+	int ret;
+
+	atomic_inc(&mm->pincount);
+
+	if (mm->type == INTEL_GVT_MM_PPGTT) {
+		ret = shadow_ppgtt_mm(mm);
+		if (ret)
+			return ret;
+
+		mutex_lock(&mm->vgpu->gvt->gtt.ppgtt_mm_lock);
+		list_move_tail(&mm->ppgtt_mm.lru_list,
+			       &mm->vgpu->gvt->gtt.ppgtt_mm_lru_list_head);
+		mutex_unlock(&mm->vgpu->gvt->gtt.ppgtt_mm_lock);
+	}
+
+	return 0;
+}
+
+static int reclaim_one_ppgtt_mm(struct intel_gvt *gvt)
+{
+	struct intel_vgpu_mm *mm;
+	struct list_head *pos, *n;
+
+	mutex_lock(&gvt->gtt.ppgtt_mm_lock);
+
+	list_for_each_safe(pos, n, &gvt->gtt.ppgtt_mm_lru_list_head) {
+		mm = container_of(pos, struct intel_vgpu_mm, ppgtt_mm.lru_list);
+
+		if (atomic_read(&mm->pincount))
+			continue;
+
+		list_del_init(&mm->ppgtt_mm.lru_list);
+		mutex_unlock(&gvt->gtt.ppgtt_mm_lock);
+		invalidate_ppgtt_mm(mm);
+		return 1;
+	}
+	mutex_unlock(&gvt->gtt.ppgtt_mm_lock);
+	return 0;
+}
+
+/*
+ * GMA translation APIs.
+ */
+static inline int ppgtt_get_next_level_entry(struct intel_vgpu_mm *mm,
+		struct intel_gvt_gtt_entry *e, unsigned long index, bool guest)
+{
+	struct intel_vgpu *vgpu = mm->vgpu;
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	struct intel_vgpu_ppgtt_spt *s;
+
+	s = intel_vgpu_find_spt_by_mfn(vgpu, ops->get_pfn(e));
+	if (!s)
+		return -ENXIO;
+
+	if (!guest)
+		ppgtt_get_shadow_entry(s, e, index);
+	else
+		ppgtt_get_guest_entry(s, e, index);
+	return 0;
+}
+
+/**
+ * intel_vgpu_gma_to_gpa - translate a gma to GPA
+ * @mm: mm object. could be a PPGTT or GGTT mm object
+ * @gma: graphics memory address in this mm object
+ *
+ * This function is used to translate a graphics memory address in specific
+ * graphics memory space to guest physical address.
+ *
+ * Returns:
+ * Guest physical address on success, INTEL_GVT_INVALID_ADDR if failed.
+ */
+unsigned long intel_vgpu_gma_to_gpa(struct intel_vgpu_mm *mm, unsigned long gma)
+{
+	struct intel_vgpu *vgpu = mm->vgpu;
+	struct intel_gvt *gvt = vgpu->gvt;
+	const struct intel_gvt_gtt_pte_ops *pte_ops = gvt->gtt.pte_ops;
+	const struct intel_gvt_gtt_gma_ops *gma_ops = gvt->gtt.gma_ops;
+	unsigned long gpa = INTEL_GVT_INVALID_ADDR;
+	unsigned long gma_index[4];
+	struct intel_gvt_gtt_entry e;
+	int i, levels = 0;
+	int ret;
+
+	GEM_BUG_ON(mm->type != INTEL_GVT_MM_GGTT &&
+		   mm->type != INTEL_GVT_MM_PPGTT);
+
+	if (mm->type == INTEL_GVT_MM_GGTT) {
+		if (!vgpu_gmadr_is_valid(vgpu, gma))
+			goto err;
+
+		ggtt_get_guest_entry(mm, &e,
+			gma_ops->gma_to_ggtt_pte_index(gma));
+
+		gpa = (pte_ops->get_pfn(&e) << I915_GTT_PAGE_SHIFT)
+			+ (gma & ~I915_GTT_PAGE_MASK);
+
+		trace_gma_translate(vgpu->id, "ggtt", 0, 0, gma, gpa);
+	} else {
+		switch (mm->ppgtt_mm.root_entry_type) {
+		case GTT_TYPE_PPGTT_ROOT_L4_ENTRY:
+			ppgtt_get_shadow_root_entry(mm, &e, 0);
+
+			gma_index[0] = gma_ops->gma_to_pml4_index(gma);
+			gma_index[1] = gma_ops->gma_to_l4_pdp_index(gma);
+			gma_index[2] = gma_ops->gma_to_pde_index(gma);
+			gma_index[3] = gma_ops->gma_to_pte_index(gma);
+			levels = 4;
+			break;
+		case GTT_TYPE_PPGTT_ROOT_L3_ENTRY:
+			ppgtt_get_shadow_root_entry(mm, &e,
+					gma_ops->gma_to_l3_pdp_index(gma));
+
+			gma_index[0] = gma_ops->gma_to_pde_index(gma);
+			gma_index[1] = gma_ops->gma_to_pte_index(gma);
+			levels = 2;
+			break;
+		default:
+			GEM_BUG_ON(1);
+		}
+
+		/* walk the shadow page table and get gpa from guest entry */
+		for (i = 0; i < levels; i++) {
+			ret = ppgtt_get_next_level_entry(mm, &e, gma_index[i],
+				(i == levels - 1));
+			if (ret)
+				goto err;
+
+			if (!pte_ops->test_present(&e)) {
+				gvt_dbg_core("GMA 0x%lx is not present\n", gma);
+				goto err;
+			}
+		}
+
+		gpa = (pte_ops->get_pfn(&e) << I915_GTT_PAGE_SHIFT) +
+					(gma & ~I915_GTT_PAGE_MASK);
+		trace_gma_translate(vgpu->id, "ppgtt", 0,
+				    mm->ppgtt_mm.root_entry_type, gma, gpa);
+	}
+
+	return gpa;
+err:
+	gvt_vgpu_err("invalid mm type: %d gma %lx\n", mm->type, gma);
+	return INTEL_GVT_INVALID_ADDR;
+}
+
+static int emulate_ggtt_mmio_read(struct intel_vgpu *vgpu,
+	unsigned int off, void *p_data, unsigned int bytes)
+{
+	struct intel_vgpu_mm *ggtt_mm = vgpu->gtt.ggtt_mm;
+	const struct intel_gvt_device_info *info = &vgpu->gvt->device_info;
+	unsigned long index = off >> info->gtt_entry_size_shift;
+	unsigned long gma;
+	struct intel_gvt_gtt_entry e;
+
+	if (bytes != 4 && bytes != 8)
+		return -EINVAL;
+
+	gma = index << I915_GTT_PAGE_SHIFT;
+	if (!intel_gvt_ggtt_validate_range(vgpu,
+					   gma, 1 << I915_GTT_PAGE_SHIFT)) {
+		gvt_dbg_mm("read invalid ggtt at 0x%lx\n", gma);
+		memset(p_data, 0, bytes);
+		return 0;
+	}
+
+	ggtt_get_guest_entry(ggtt_mm, &e, index);
+	memcpy(p_data, (void *)&e.val64 + (off & (info->gtt_entry_size - 1)),
+			bytes);
+	return 0;
+}
+
+/**
+ * intel_vgpu_emulate_ggtt_mmio_read - emulate GTT MMIO register read
+ * @vgpu: a vGPU
+ * @off: register offset
+ * @p_data: data will be returned to guest
+ * @bytes: data length
+ *
+ * This function is used to emulate the GTT MMIO register read
+ *
+ * Returns:
+ * Zero on success, error code if failed.
+ */
+int intel_vgpu_emulate_ggtt_mmio_read(struct intel_vgpu *vgpu, unsigned int off,
+	void *p_data, unsigned int bytes)
+{
+	const struct intel_gvt_device_info *info = &vgpu->gvt->device_info;
+	int ret;
+
+	if (bytes != 4 && bytes != 8)
+		return -EINVAL;
+
+	off -= info->gtt_start_offset;
+	ret = emulate_ggtt_mmio_read(vgpu, off, p_data, bytes);
+	return ret;
+}
+
+static void ggtt_invalidate_pte(struct intel_vgpu *vgpu,
+		struct intel_gvt_gtt_entry *entry)
+{
+	const struct intel_gvt_gtt_pte_ops *pte_ops = vgpu->gvt->gtt.pte_ops;
+	unsigned long pfn;
+
+	pfn = pte_ops->get_pfn(entry);
+	if (pfn != vgpu->gvt->gtt.scratch_mfn)
+		intel_gvt_dma_unmap_guest_page(vgpu, pfn << PAGE_SHIFT);
+}
+
+static int emulate_ggtt_mmio_write(struct intel_vgpu *vgpu, unsigned int off,
+	void *p_data, unsigned int bytes)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	const struct intel_gvt_device_info *info = &gvt->device_info;
+	struct intel_vgpu_mm *ggtt_mm = vgpu->gtt.ggtt_mm;
+	const struct intel_gvt_gtt_pte_ops *ops = gvt->gtt.pte_ops;
+	unsigned long g_gtt_index = off >> info->gtt_entry_size_shift;
+	unsigned long gma, gfn;
+	struct intel_gvt_gtt_entry e = {.val64 = 0, .type = GTT_TYPE_GGTT_PTE};
+	struct intel_gvt_gtt_entry m = {.val64 = 0, .type = GTT_TYPE_GGTT_PTE};
+	dma_addr_t dma_addr;
+	int ret;
+	struct intel_gvt_partial_pte *partial_pte, *pos, *n;
+	bool partial_update = false;
+
+	if (bytes != 4 && bytes != 8)
+		return -EINVAL;
+
+	gma = g_gtt_index << I915_GTT_PAGE_SHIFT;
+
+	/* the VM may configure the whole GM space when ballooning is used */
+	if (!vgpu_gmadr_is_valid(vgpu, gma))
+		return 0;
+
+	e.type = GTT_TYPE_GGTT_PTE;
+	memcpy((void *)&e.val64 + (off & (info->gtt_entry_size - 1)), p_data,
+			bytes);
+
+	/* If ggtt entry size is 8 bytes, and it's split into two 4 bytes
+	 * write, save the first 4 bytes in a list and update virtual
+	 * PTE. Only update shadow PTE when the second 4 bytes comes.
+	 */
+	if (bytes < info->gtt_entry_size) {
+		bool found = false;
+
+		list_for_each_entry_safe(pos, n,
+				&ggtt_mm->ggtt_mm.partial_pte_list, list) {
+			if (g_gtt_index == pos->offset >>
+					info->gtt_entry_size_shift) {
+				if (off != pos->offset) {
+					/* the second partial part*/
+					int last_off = pos->offset &
+						(info->gtt_entry_size - 1);
+
+					memcpy((void *)&e.val64 + last_off,
+						(void *)&pos->data + last_off,
+						bytes);
+
+					list_del(&pos->list);
+					kfree(pos);
+					found = true;
+					break;
+				}
+
+				/* update of the first partial part */
+				pos->data = e.val64;
+				ggtt_set_guest_entry(ggtt_mm, &e, g_gtt_index);
+				return 0;
+			}
+		}
+
+		if (!found) {
+			/* the first partial part */
+			partial_pte = kzalloc(sizeof(*partial_pte), GFP_KERNEL);
+			if (!partial_pte)
+				return -ENOMEM;
+			partial_pte->offset = off;
+			partial_pte->data = e.val64;
+			list_add_tail(&partial_pte->list,
+				&ggtt_mm->ggtt_mm.partial_pte_list);
+			partial_update = true;
+		}
+	}
+
+	if (!partial_update && (ops->test_present(&e))) {
+		gfn = ops->get_pfn(&e);
+		m.val64 = e.val64;
+		m.type = e.type;
+
+		/* one PTE update may be issued in multiple writes and the
+		 * first write may not construct a valid gfn
+		 */
+		if (!intel_gvt_is_valid_gfn(vgpu, gfn)) {
+			ops->set_pfn(&m, gvt->gtt.scratch_mfn);
+			goto out;
+		}
+
+		ret = intel_gvt_dma_map_guest_page(vgpu, gfn, PAGE_SIZE,
+						   &dma_addr);
+		if (ret) {
+			gvt_vgpu_err("fail to populate guest ggtt entry\n");
+			/* guest driver may read/write the entry when partial
+			 * update the entry in this situation p2m will fail
+			 * setting the shadow entry to point to a scratch page
+			 */
+			ops->set_pfn(&m, gvt->gtt.scratch_mfn);
+		} else
+			ops->set_pfn(&m, dma_addr >> PAGE_SHIFT);
+	} else {
+		ops->set_pfn(&m, gvt->gtt.scratch_mfn);
+		ops->clear_present(&m);
+	}
+
+out:
+	ggtt_set_guest_entry(ggtt_mm, &e, g_gtt_index);
+
+	ggtt_get_host_entry(ggtt_mm, &e, g_gtt_index);
+	ggtt_invalidate_pte(vgpu, &e);
+
+	ggtt_set_host_entry(ggtt_mm, &m, g_gtt_index);
+	ggtt_invalidate(gvt->gt);
+	return 0;
+}
+
+/*
+ * intel_vgpu_emulate_ggtt_mmio_write - emulate GTT MMIO register write
+ * @vgpu: a vGPU
+ * @off: register offset
+ * @p_data: data from guest write
+ * @bytes: data length
+ *
+ * This function is used to emulate the GTT MMIO register write
+ *
+ * Returns:
+ * Zero on success, error code if failed.
+ */
+int intel_vgpu_emulate_ggtt_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int off, void *p_data, unsigned int bytes)
+{
+	const struct intel_gvt_device_info *info = &vgpu->gvt->device_info;
+	int ret;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_engine_cs *engine;
+	int i;
+
+	if (bytes != 4 && bytes != 8)
+		return -EINVAL;
+
+	off -= info->gtt_start_offset;
+	ret = emulate_ggtt_mmio_write(vgpu, off, p_data, bytes);
+
+	/* if ggtt of last submitted context is written,
+	 * that context is probably got unpinned.
+	 * Set last shadowed ctx to invalid.
+	 */
+	for_each_engine(engine, vgpu->gvt->gt, i) {
+		if (!s->last_ctx[i].valid)
+			continue;
+
+		if (s->last_ctx[i].lrca == (off >> info->gtt_entry_size_shift))
+			s->last_ctx[i].valid = false;
+	}
+	return ret;
+}
+
+static int alloc_scratch_pages(struct intel_vgpu *vgpu,
+		enum intel_gvt_gtt_type type)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	struct intel_vgpu_gtt *gtt = &vgpu->gtt;
+	const struct intel_gvt_gtt_pte_ops *ops = vgpu->gvt->gtt.pte_ops;
+	int page_entry_num = I915_GTT_PAGE_SIZE >>
+				vgpu->gvt->device_info.gtt_entry_size_shift;
+	void *scratch_pt;
+	int i;
+	struct device *dev = vgpu->gvt->gt->i915->drm.dev;
+	dma_addr_t daddr;
+
+	if (drm_WARN_ON(&i915->drm,
+			type < GTT_TYPE_PPGTT_PTE_PT || type >= GTT_TYPE_MAX))
+		return -EINVAL;
+
+	scratch_pt = (void *)get_zeroed_page(GFP_KERNEL);
+	if (!scratch_pt) {
+		gvt_vgpu_err("fail to allocate scratch page\n");
+		return -ENOMEM;
+	}
+
+	daddr = dma_map_page(dev, virt_to_page(scratch_pt), 0, 4096, DMA_BIDIRECTIONAL);
+	if (dma_mapping_error(dev, daddr)) {
+		gvt_vgpu_err("fail to dmamap scratch_pt\n");
+		__free_page(virt_to_page(scratch_pt));
+		return -ENOMEM;
+	}
+	gtt->scratch_pt[type].page_mfn =
+		(unsigned long)(daddr >> I915_GTT_PAGE_SHIFT);
+	gtt->scratch_pt[type].page = virt_to_page(scratch_pt);
+	gvt_dbg_mm("vgpu%d create scratch_pt: type %d mfn=0x%lx\n",
+			vgpu->id, type, gtt->scratch_pt[type].page_mfn);
+
+	/* Build the tree by full filled the scratch pt with the entries which
+	 * point to the next level scratch pt or scratch page. The
+	 * scratch_pt[type] indicate the scratch pt/scratch page used by the
+	 * 'type' pt.
+	 * e.g. scratch_pt[GTT_TYPE_PPGTT_PDE_PT] is used by
+	 * GTT_TYPE_PPGTT_PDE_PT level pt, that means this scratch_pt it self
+	 * is GTT_TYPE_PPGTT_PTE_PT, and full filled by scratch page mfn.
+	 */
+	if (type > GTT_TYPE_PPGTT_PTE_PT) {
+		struct intel_gvt_gtt_entry se;
+
+		memset(&se, 0, sizeof(struct intel_gvt_gtt_entry));
+		se.type = get_entry_type(type - 1);
+		ops->set_pfn(&se, gtt->scratch_pt[type - 1].page_mfn);
+
+		/* The entry parameters like present/writeable/cache type
+		 * set to the same as i915's scratch page tree.
+		 */
+		se.val64 |= GEN8_PAGE_PRESENT | GEN8_PAGE_RW;
+		if (type == GTT_TYPE_PPGTT_PDE_PT)
+			se.val64 |= PPAT_CACHED;
+
+		for (i = 0; i < page_entry_num; i++)
+			ops->set_entry(scratch_pt, &se, i, false, 0, vgpu);
+	}
+
+	return 0;
+}
+
+static int release_scratch_page_tree(struct intel_vgpu *vgpu)
+{
+	int i;
+	struct device *dev = vgpu->gvt->gt->i915->drm.dev;
+	dma_addr_t daddr;
+
+	for (i = GTT_TYPE_PPGTT_PTE_PT; i < GTT_TYPE_MAX; i++) {
+		if (vgpu->gtt.scratch_pt[i].page != NULL) {
+			daddr = (dma_addr_t)(vgpu->gtt.scratch_pt[i].page_mfn <<
+					I915_GTT_PAGE_SHIFT);
+			dma_unmap_page(dev, daddr, 4096, DMA_BIDIRECTIONAL);
+			__free_page(vgpu->gtt.scratch_pt[i].page);
+			vgpu->gtt.scratch_pt[i].page = NULL;
+			vgpu->gtt.scratch_pt[i].page_mfn = 0;
+		}
+	}
+
+	return 0;
+}
+
+static int create_scratch_page_tree(struct intel_vgpu *vgpu)
+{
+	int i, ret;
+
+	for (i = GTT_TYPE_PPGTT_PTE_PT; i < GTT_TYPE_MAX; i++) {
+		ret = alloc_scratch_pages(vgpu, i);
+		if (ret)
+			goto err;
+	}
+
+	return 0;
+
+err:
+	release_scratch_page_tree(vgpu);
+	return ret;
+}
+
+/**
+ * intel_vgpu_init_gtt - initialize per-vGPU graphics memory virulization
+ * @vgpu: a vGPU
+ *
+ * This function is used to initialize per-vGPU graphics memory virtualization
+ * components.
+ *
+ * Returns:
+ * Zero on success, error code if failed.
+ */
+int intel_vgpu_init_gtt(struct intel_vgpu *vgpu)
+{
+	struct intel_vgpu_gtt *gtt = &vgpu->gtt;
+
+	INIT_RADIX_TREE(&gtt->spt_tree, GFP_KERNEL);
+
+	INIT_LIST_HEAD(&gtt->ppgtt_mm_list_head);
+	INIT_LIST_HEAD(&gtt->oos_page_list_head);
+	INIT_LIST_HEAD(&gtt->post_shadow_list_head);
+
+	gtt->ggtt_mm = intel_vgpu_create_ggtt_mm(vgpu);
+	if (IS_ERR(gtt->ggtt_mm)) {
+		gvt_vgpu_err("fail to create mm for ggtt.\n");
+		return PTR_ERR(gtt->ggtt_mm);
+	}
+
+	intel_vgpu_reset_ggtt(vgpu, false);
+
+	INIT_LIST_HEAD(&gtt->ggtt_mm->ggtt_mm.partial_pte_list);
+
+	return create_scratch_page_tree(vgpu);
+}
+
+void intel_vgpu_destroy_all_ppgtt_mm(struct intel_vgpu *vgpu)
+{
+	struct list_head *pos, *n;
+	struct intel_vgpu_mm *mm;
+
+	list_for_each_safe(pos, n, &vgpu->gtt.ppgtt_mm_list_head) {
+		mm = container_of(pos, struct intel_vgpu_mm, ppgtt_mm.list);
+		intel_vgpu_destroy_mm(mm);
+	}
+
+	if (GEM_WARN_ON(!list_empty(&vgpu->gtt.ppgtt_mm_list_head)))
+		gvt_err("vgpu ppgtt mm is not fully destroyed\n");
+
+	if (GEM_WARN_ON(!radix_tree_empty(&vgpu->gtt.spt_tree))) {
+		gvt_err("Why we still has spt not freed?\n");
+		ppgtt_free_all_spt(vgpu);
+	}
+}
+
+static void intel_vgpu_destroy_ggtt_mm(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt_partial_pte *pos, *next;
+
+	list_for_each_entry_safe(pos, next,
+				 &vgpu->gtt.ggtt_mm->ggtt_mm.partial_pte_list,
+				 list) {
+		gvt_dbg_mm("partial PTE update on hold 0x%lx : 0x%llx\n",
+			pos->offset, pos->data);
+		kfree(pos);
+	}
+	intel_vgpu_destroy_mm(vgpu->gtt.ggtt_mm);
+	vgpu->gtt.ggtt_mm = NULL;
+}
+
+/**
+ * intel_vgpu_clean_gtt - clean up per-vGPU graphics memory virulization
+ * @vgpu: a vGPU
+ *
+ * This function is used to clean up per-vGPU graphics memory virtualization
+ * components.
+ *
+ * Returns:
+ * Zero on success, error code if failed.
+ */
+void intel_vgpu_clean_gtt(struct intel_vgpu *vgpu)
+{
+	intel_vgpu_destroy_all_ppgtt_mm(vgpu);
+	intel_vgpu_destroy_ggtt_mm(vgpu);
+	release_scratch_page_tree(vgpu);
+}
+
+static void clean_spt_oos(struct intel_gvt *gvt)
+{
+	struct intel_gvt_gtt *gtt = &gvt->gtt;
+	struct list_head *pos, *n;
+	struct intel_vgpu_oos_page *oos_page;
+
+	WARN(!list_empty(&gtt->oos_page_use_list_head),
+		"someone is still using oos page\n");
+
+	list_for_each_safe(pos, n, &gtt->oos_page_free_list_head) {
+		oos_page = container_of(pos, struct intel_vgpu_oos_page, list);
+		list_del(&oos_page->list);
+		free_page((unsigned long)oos_page->mem);
+		kfree(oos_page);
+	}
+}
+
+static int setup_spt_oos(struct intel_gvt *gvt)
+{
+	struct intel_gvt_gtt *gtt = &gvt->gtt;
+	struct intel_vgpu_oos_page *oos_page;
+	int i;
+	int ret;
+
+	INIT_LIST_HEAD(&gtt->oos_page_free_list_head);
+	INIT_LIST_HEAD(&gtt->oos_page_use_list_head);
+
+	for (i = 0; i < preallocated_oos_pages; i++) {
+		oos_page = kzalloc(sizeof(*oos_page), GFP_KERNEL);
+		if (!oos_page) {
+			ret = -ENOMEM;
+			goto fail;
+		}
+		oos_page->mem = (void *)__get_free_pages(GFP_KERNEL, 0);
+		if (!oos_page->mem) {
+			ret = -ENOMEM;
+			kfree(oos_page);
+			goto fail;
+		}
+
+		INIT_LIST_HEAD(&oos_page->list);
+		INIT_LIST_HEAD(&oos_page->vm_list);
+		oos_page->id = i;
+		list_add_tail(&oos_page->list, &gtt->oos_page_free_list_head);
+	}
+
+	gvt_dbg_mm("%d oos pages preallocated\n", i);
+
+	return 0;
+fail:
+	clean_spt_oos(gvt);
+	return ret;
+}
+
+/**
+ * intel_vgpu_find_ppgtt_mm - find a PPGTT mm object
+ * @vgpu: a vGPU
+ * @pdps: pdp root array
+ *
+ * This function is used to find a PPGTT mm object from mm object pool
+ *
+ * Returns:
+ * pointer to mm object on success, NULL if failed.
+ */
+struct intel_vgpu_mm *intel_vgpu_find_ppgtt_mm(struct intel_vgpu *vgpu,
+		u64 pdps[])
+{
+	struct intel_vgpu_mm *mm;
+	struct list_head *pos;
+
+	list_for_each(pos, &vgpu->gtt.ppgtt_mm_list_head) {
+		mm = container_of(pos, struct intel_vgpu_mm, ppgtt_mm.list);
+
+		switch (mm->ppgtt_mm.root_entry_type) {
+		case GTT_TYPE_PPGTT_ROOT_L4_ENTRY:
+			if (pdps[0] == mm->ppgtt_mm.guest_pdps[0])
+				return mm;
+			break;
+		case GTT_TYPE_PPGTT_ROOT_L3_ENTRY:
+			if (!memcmp(pdps, mm->ppgtt_mm.guest_pdps,
+				    sizeof(mm->ppgtt_mm.guest_pdps)))
+				return mm;
+			break;
+		default:
+			GEM_BUG_ON(1);
+		}
+	}
+	return NULL;
+}
+
+/**
+ * intel_vgpu_get_ppgtt_mm - get or create a PPGTT mm object.
+ * @vgpu: a vGPU
+ * @root_entry_type: ppgtt root entry type
+ * @pdps: guest pdps
+ *
+ * This function is used to find or create a PPGTT mm object from a guest.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+struct intel_vgpu_mm *intel_vgpu_get_ppgtt_mm(struct intel_vgpu *vgpu,
+		enum intel_gvt_gtt_type root_entry_type, u64 pdps[])
+{
+	struct intel_vgpu_mm *mm;
+
+	mm = intel_vgpu_find_ppgtt_mm(vgpu, pdps);
+	if (mm) {
+		intel_vgpu_mm_get(mm);
+	} else {
+		mm = intel_vgpu_create_ppgtt_mm(vgpu, root_entry_type, pdps);
+		if (IS_ERR(mm))
+			gvt_vgpu_err("fail to create mm\n");
+	}
+	return mm;
+}
+
+/**
+ * intel_vgpu_put_ppgtt_mm - find and put a PPGTT mm object.
+ * @vgpu: a vGPU
+ * @pdps: guest pdps
+ *
+ * This function is used to find a PPGTT mm object from a guest and destroy it.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_put_ppgtt_mm(struct intel_vgpu *vgpu, u64 pdps[])
+{
+	struct intel_vgpu_mm *mm;
+
+	mm = intel_vgpu_find_ppgtt_mm(vgpu, pdps);
+	if (!mm) {
+		gvt_vgpu_err("fail to find ppgtt instance.\n");
+		return -EINVAL;
+	}
+	intel_vgpu_mm_put(mm);
+	return 0;
+}
+
+/**
+ * intel_gvt_init_gtt - initialize mm components of a GVT device
+ * @gvt: GVT device
+ *
+ * This function is called at the initialization stage, to initialize
+ * the mm components of a GVT device.
+ *
+ * Returns:
+ * zero on success, negative error code if failed.
+ */
+int intel_gvt_init_gtt(struct intel_gvt *gvt)
+{
+	int ret;
+	void *page;
+	struct device *dev = gvt->gt->i915->drm.dev;
+	dma_addr_t daddr;
+
+	gvt_dbg_core("init gtt\n");
+
+	gvt->gtt.pte_ops = &gen8_gtt_pte_ops;
+	gvt->gtt.gma_ops = &gen8_gtt_gma_ops;
+
+	page = (void *)get_zeroed_page(GFP_KERNEL);
+	if (!page) {
+		gvt_err("fail to allocate scratch ggtt page\n");
+		return -ENOMEM;
+	}
+
+	daddr = dma_map_page(dev, virt_to_page(page), 0,
+			4096, DMA_BIDIRECTIONAL);
+	if (dma_mapping_error(dev, daddr)) {
+		gvt_err("fail to dmamap scratch ggtt page\n");
+		__free_page(virt_to_page(page));
+		return -ENOMEM;
+	}
+
+	gvt->gtt.scratch_page = virt_to_page(page);
+	gvt->gtt.scratch_mfn = (unsigned long)(daddr >> I915_GTT_PAGE_SHIFT);
+
+	if (enable_out_of_sync) {
+		ret = setup_spt_oos(gvt);
+		if (ret) {
+			gvt_err("fail to initialize SPT oos\n");
+			dma_unmap_page(dev, daddr, 4096, DMA_BIDIRECTIONAL);
+			__free_page(gvt->gtt.scratch_page);
+			return ret;
+		}
+	}
+	INIT_LIST_HEAD(&gvt->gtt.ppgtt_mm_lru_list_head);
+	mutex_init(&gvt->gtt.ppgtt_mm_lock);
+	return 0;
+}
+
+/**
+ * intel_gvt_clean_gtt - clean up mm components of a GVT device
+ * @gvt: GVT device
+ *
+ * This function is called at the driver unloading stage, to clean up the
+ * the mm components of a GVT device.
+ *
+ */
+void intel_gvt_clean_gtt(struct intel_gvt *gvt)
+{
+	struct device *dev = gvt->gt->i915->drm.dev;
+	dma_addr_t daddr = (dma_addr_t)(gvt->gtt.scratch_mfn <<
+					I915_GTT_PAGE_SHIFT);
+
+	dma_unmap_page(dev, daddr, 4096, DMA_BIDIRECTIONAL);
+
+	__free_page(gvt->gtt.scratch_page);
+
+	if (enable_out_of_sync)
+		clean_spt_oos(gvt);
+}
+
+/**
+ * intel_vgpu_invalidate_ppgtt - invalidate PPGTT instances
+ * @vgpu: a vGPU
+ *
+ * This function is called when invalidate all PPGTT instances of a vGPU.
+ *
+ */
+void intel_vgpu_invalidate_ppgtt(struct intel_vgpu *vgpu)
+{
+	struct list_head *pos, *n;
+	struct intel_vgpu_mm *mm;
+
+	list_for_each_safe(pos, n, &vgpu->gtt.ppgtt_mm_list_head) {
+		mm = container_of(pos, struct intel_vgpu_mm, ppgtt_mm.list);
+		if (mm->type == INTEL_GVT_MM_PPGTT) {
+			mutex_lock(&vgpu->gvt->gtt.ppgtt_mm_lock);
+			list_del_init(&mm->ppgtt_mm.lru_list);
+			mutex_unlock(&vgpu->gvt->gtt.ppgtt_mm_lock);
+			if (mm->ppgtt_mm.shadowed)
+				invalidate_ppgtt_mm(mm);
+		}
+	}
+}
+
+/**
+ * intel_vgpu_reset_ggtt - reset the GGTT entry
+ * @vgpu: a vGPU
+ * @invalidate_old: invalidate old entries
+ *
+ * This function is called at the vGPU create stage
+ * to reset all the GGTT entries.
+ *
+ */
+void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu, bool invalidate_old)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	const struct intel_gvt_gtt_pte_ops *pte_ops = vgpu->gvt->gtt.pte_ops;
+	struct intel_gvt_gtt_entry entry = {.type = GTT_TYPE_GGTT_PTE};
+	struct intel_gvt_gtt_entry old_entry;
+	u32 index;
+	u32 num_entries;
+
+	pte_ops->set_pfn(&entry, gvt->gtt.scratch_mfn);
+	pte_ops->set_present(&entry);
+
+	index = vgpu_aperture_gmadr_base(vgpu) >> PAGE_SHIFT;
+	num_entries = vgpu_aperture_sz(vgpu) >> PAGE_SHIFT;
+	while (num_entries--) {
+		if (invalidate_old) {
+			ggtt_get_host_entry(vgpu->gtt.ggtt_mm, &old_entry, index);
+			ggtt_invalidate_pte(vgpu, &old_entry);
+		}
+		ggtt_set_host_entry(vgpu->gtt.ggtt_mm, &entry, index++);
+	}
+
+	index = vgpu_hidden_gmadr_base(vgpu) >> PAGE_SHIFT;
+	num_entries = vgpu_hidden_sz(vgpu) >> PAGE_SHIFT;
+	while (num_entries--) {
+		if (invalidate_old) {
+			ggtt_get_host_entry(vgpu->gtt.ggtt_mm, &old_entry, index);
+			ggtt_invalidate_pte(vgpu, &old_entry);
+		}
+		ggtt_set_host_entry(vgpu->gtt.ggtt_mm, &entry, index++);
+	}
+
+	ggtt_invalidate(gvt->gt);
+}
+
+/**
+ * intel_gvt_restore_ggtt - restore all vGPU's ggtt entries
+ * @gvt: intel gvt device
+ *
+ * This function is called at driver resume stage to restore
+ * GGTT entries of every vGPU.
+ *
+ */
+void intel_gvt_restore_ggtt(struct intel_gvt *gvt)
+{
+	struct intel_vgpu *vgpu;
+	struct intel_vgpu_mm *mm;
+	int id;
+	gen8_pte_t pte;
+	u32 idx, num_low, num_hi, offset;
+
+	/* Restore dirty host ggtt for all vGPUs */
+	idr_for_each_entry(&(gvt)->vgpu_idr, vgpu, id) {
+		mm = vgpu->gtt.ggtt_mm;
+
+		num_low = vgpu_aperture_sz(vgpu) >> PAGE_SHIFT;
+		offset = vgpu_aperture_gmadr_base(vgpu) >> PAGE_SHIFT;
+		for (idx = 0; idx < num_low; idx++) {
+			pte = mm->ggtt_mm.host_ggtt_aperture[idx];
+			if (pte & GEN8_PAGE_PRESENT)
+				write_pte64(vgpu->gvt->gt->ggtt, offset + idx, pte);
+		}
+
+		num_hi = vgpu_hidden_sz(vgpu) >> PAGE_SHIFT;
+		offset = vgpu_hidden_gmadr_base(vgpu) >> PAGE_SHIFT;
+		for (idx = 0; idx < num_hi; idx++) {
+			pte = mm->ggtt_mm.host_ggtt_hidden[idx];
+			if (pte & GEN8_PAGE_PRESENT)
+				write_pte64(vgpu->gvt->gt->ggtt, offset + idx, pte);
+		}
+	}
+}
diff --git a/drivers/gpu/drm/i915/gvt/gtt.h b/drivers/gpu/drm/i915/gvt/gtt.h
new file mode 100644
index 000000000..4cb183e06
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/gtt.h
@@ -0,0 +1,296 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *    Zhenyu Wang <zhenyuw@linux.intel.com>
+ *    Xiao Zheng <xiao.zheng@intel.com>
+ *
+ * Contributors:
+ *    Min He <min.he@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *
+ */
+
+#ifndef _GVT_GTT_H_
+#define _GVT_GTT_H_
+
+#include <linux/kernel.h>
+#include <linux/kref.h>
+#include <linux/mutex.h>
+#include <linux/radix-tree.h>
+
+#include "gt/intel_gtt.h"
+
+struct intel_gvt;
+struct intel_vgpu;
+struct intel_vgpu_mm;
+
+#define I915_GTT_PAGE_SHIFT         12
+
+#define INTEL_GVT_INVALID_ADDR (~0UL)
+
+struct intel_gvt_gtt_entry {
+	u64 val64;
+	int type;
+};
+
+struct intel_gvt_gtt_pte_ops {
+	int (*get_entry)(void *pt,
+			 struct intel_gvt_gtt_entry *e,
+			 unsigned long index,
+			 bool hypervisor_access,
+			 unsigned long gpa,
+			 struct intel_vgpu *vgpu);
+	int (*set_entry)(void *pt,
+			 struct intel_gvt_gtt_entry *e,
+			 unsigned long index,
+			 bool hypervisor_access,
+			 unsigned long gpa,
+			 struct intel_vgpu *vgpu);
+	bool (*test_present)(struct intel_gvt_gtt_entry *e);
+	void (*clear_present)(struct intel_gvt_gtt_entry *e);
+	void (*set_present)(struct intel_gvt_gtt_entry *e);
+	bool (*test_pse)(struct intel_gvt_gtt_entry *e);
+	void (*clear_pse)(struct intel_gvt_gtt_entry *e);
+	bool (*test_ips)(struct intel_gvt_gtt_entry *e);
+	void (*clear_ips)(struct intel_gvt_gtt_entry *e);
+	bool (*test_64k_splited)(struct intel_gvt_gtt_entry *e);
+	void (*clear_64k_splited)(struct intel_gvt_gtt_entry *e);
+	void (*set_64k_splited)(struct intel_gvt_gtt_entry *e);
+	void (*set_pfn)(struct intel_gvt_gtt_entry *e, unsigned long pfn);
+	unsigned long (*get_pfn)(struct intel_gvt_gtt_entry *e);
+};
+
+struct intel_gvt_gtt_gma_ops {
+	unsigned long (*gma_to_ggtt_pte_index)(unsigned long gma);
+	unsigned long (*gma_to_pte_index)(unsigned long gma);
+	unsigned long (*gma_to_pde_index)(unsigned long gma);
+	unsigned long (*gma_to_l3_pdp_index)(unsigned long gma);
+	unsigned long (*gma_to_l4_pdp_index)(unsigned long gma);
+	unsigned long (*gma_to_pml4_index)(unsigned long gma);
+};
+
+struct intel_gvt_gtt {
+	const struct intel_gvt_gtt_pte_ops *pte_ops;
+	const struct intel_gvt_gtt_gma_ops *gma_ops;
+	int (*mm_alloc_page_table)(struct intel_vgpu_mm *mm);
+	void (*mm_free_page_table)(struct intel_vgpu_mm *mm);
+	struct list_head oos_page_use_list_head;
+	struct list_head oos_page_free_list_head;
+	struct mutex ppgtt_mm_lock;
+	struct list_head ppgtt_mm_lru_list_head;
+
+	struct page *scratch_page;
+	unsigned long scratch_mfn;
+};
+
+enum intel_gvt_gtt_type {
+	GTT_TYPE_INVALID = 0,
+
+	GTT_TYPE_GGTT_PTE,
+
+	GTT_TYPE_PPGTT_PTE_4K_ENTRY,
+	GTT_TYPE_PPGTT_PTE_64K_ENTRY,
+	GTT_TYPE_PPGTT_PTE_2M_ENTRY,
+	GTT_TYPE_PPGTT_PTE_1G_ENTRY,
+
+	GTT_TYPE_PPGTT_PTE_ENTRY,
+
+	GTT_TYPE_PPGTT_PDE_ENTRY,
+	GTT_TYPE_PPGTT_PDP_ENTRY,
+	GTT_TYPE_PPGTT_PML4_ENTRY,
+
+	GTT_TYPE_PPGTT_ROOT_ENTRY,
+
+	GTT_TYPE_PPGTT_ROOT_L3_ENTRY,
+	GTT_TYPE_PPGTT_ROOT_L4_ENTRY,
+
+	GTT_TYPE_PPGTT_ENTRY,
+
+	GTT_TYPE_PPGTT_PTE_PT,
+	GTT_TYPE_PPGTT_PDE_PT,
+	GTT_TYPE_PPGTT_PDP_PT,
+	GTT_TYPE_PPGTT_PML4_PT,
+
+	GTT_TYPE_MAX,
+};
+
+enum intel_gvt_mm_type {
+	INTEL_GVT_MM_GGTT,
+	INTEL_GVT_MM_PPGTT,
+};
+
+#define GVT_RING_CTX_NR_PDPS	GEN8_3LVL_PDPES
+
+struct intel_gvt_partial_pte {
+	unsigned long offset;
+	u64 data;
+	struct list_head list;
+};
+
+struct intel_vgpu_mm {
+	enum intel_gvt_mm_type type;
+	struct intel_vgpu *vgpu;
+
+	struct kref ref;
+	atomic_t pincount;
+
+	union {
+		struct {
+			enum intel_gvt_gtt_type root_entry_type;
+			/*
+			 * The 4 PDPs in ring context. For 48bit addressing,
+			 * only PDP0 is valid and point to PML4. For 32it
+			 * addressing, all 4 are used as true PDPs.
+			 */
+			u64 guest_pdps[GVT_RING_CTX_NR_PDPS];
+			u64 shadow_pdps[GVT_RING_CTX_NR_PDPS];
+			bool shadowed;
+
+			struct list_head list;
+			struct list_head lru_list;
+			struct list_head link; /* possible LRI shadow mm list */
+		} ppgtt_mm;
+		struct {
+			void *virtual_ggtt;
+			/* Save/restore for PM */
+			u64 *host_ggtt_aperture;
+			u64 *host_ggtt_hidden;
+			struct list_head partial_pte_list;
+		} ggtt_mm;
+	};
+};
+
+struct intel_vgpu_mm *intel_vgpu_create_ppgtt_mm(struct intel_vgpu *vgpu,
+		enum intel_gvt_gtt_type root_entry_type, u64 pdps[]);
+
+static inline void intel_vgpu_mm_get(struct intel_vgpu_mm *mm)
+{
+	kref_get(&mm->ref);
+}
+
+void _intel_vgpu_mm_release(struct kref *mm_ref);
+
+static inline void intel_vgpu_mm_put(struct intel_vgpu_mm *mm)
+{
+	kref_put(&mm->ref, _intel_vgpu_mm_release);
+}
+
+static inline void intel_vgpu_destroy_mm(struct intel_vgpu_mm *mm)
+{
+	intel_vgpu_mm_put(mm);
+}
+
+struct intel_vgpu_guest_page;
+
+struct intel_vgpu_scratch_pt {
+	struct page *page;
+	unsigned long page_mfn;
+};
+
+struct intel_vgpu_gtt {
+	struct intel_vgpu_mm *ggtt_mm;
+	unsigned long active_ppgtt_mm_bitmap;
+	struct list_head ppgtt_mm_list_head;
+	struct radix_tree_root spt_tree;
+	struct list_head oos_page_list_head;
+	struct list_head post_shadow_list_head;
+	struct intel_vgpu_scratch_pt scratch_pt[GTT_TYPE_MAX];
+};
+
+int intel_vgpu_init_gtt(struct intel_vgpu *vgpu);
+void intel_vgpu_clean_gtt(struct intel_vgpu *vgpu);
+void intel_vgpu_reset_ggtt(struct intel_vgpu *vgpu, bool invalidate_old);
+void intel_vgpu_invalidate_ppgtt(struct intel_vgpu *vgpu);
+
+int intel_gvt_init_gtt(struct intel_gvt *gvt);
+void intel_gvt_clean_gtt(struct intel_gvt *gvt);
+
+struct intel_vgpu_mm *intel_gvt_find_ppgtt_mm(struct intel_vgpu *vgpu,
+					      int page_table_level,
+					      void *root_entry);
+
+struct intel_vgpu_oos_page {
+	struct intel_vgpu_ppgtt_spt *spt;
+	struct list_head list;
+	struct list_head vm_list;
+	int id;
+	void *mem;
+};
+
+#define GTT_ENTRY_NUM_IN_ONE_PAGE 512
+
+/* Represent a vgpu shadow page table. */
+struct intel_vgpu_ppgtt_spt {
+	atomic_t refcount;
+	struct intel_vgpu *vgpu;
+
+	struct {
+		enum intel_gvt_gtt_type type;
+		bool pde_ips; /* for 64KB PTEs */
+		void *vaddr;
+		struct page *page;
+		unsigned long mfn;
+	} shadow_page;
+
+	struct {
+		enum intel_gvt_gtt_type type;
+		bool pde_ips; /* for 64KB PTEs */
+		unsigned long gfn;
+		unsigned long write_cnt;
+		struct intel_vgpu_oos_page *oos_page;
+	} guest_page;
+
+	DECLARE_BITMAP(post_shadow_bitmap, GTT_ENTRY_NUM_IN_ONE_PAGE);
+	struct list_head post_shadow_list;
+};
+
+int intel_vgpu_sync_oos_pages(struct intel_vgpu *vgpu);
+
+int intel_vgpu_flush_post_shadow(struct intel_vgpu *vgpu);
+
+int intel_vgpu_pin_mm(struct intel_vgpu_mm *mm);
+
+void intel_vgpu_unpin_mm(struct intel_vgpu_mm *mm);
+
+unsigned long intel_vgpu_gma_to_gpa(struct intel_vgpu_mm *mm,
+		unsigned long gma);
+
+struct intel_vgpu_mm *intel_vgpu_find_ppgtt_mm(struct intel_vgpu *vgpu,
+		u64 pdps[]);
+
+struct intel_vgpu_mm *intel_vgpu_get_ppgtt_mm(struct intel_vgpu *vgpu,
+		enum intel_gvt_gtt_type root_entry_type, u64 pdps[]);
+
+int intel_vgpu_put_ppgtt_mm(struct intel_vgpu *vgpu, u64 pdps[]);
+
+int intel_vgpu_emulate_ggtt_mmio_read(struct intel_vgpu *vgpu,
+	unsigned int off, void *p_data, unsigned int bytes);
+
+int intel_vgpu_emulate_ggtt_mmio_write(struct intel_vgpu *vgpu,
+	unsigned int off, void *p_data, unsigned int bytes);
+
+void intel_vgpu_destroy_all_ppgtt_mm(struct intel_vgpu *vgpu);
+void intel_gvt_restore_ggtt(struct intel_gvt *gvt);
+
+#endif /* _GVT_GTT_H_ */
diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h
new file mode 100644
index 000000000..dbf8d7470
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/gvt.h
@@ -0,0 +1,767 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Kevin Tian <kevin.tian@intel.com>
+ *    Eddie Dong <eddie.dong@intel.com>
+ *
+ * Contributors:
+ *    Niu Bing <bing.niu@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#ifndef _GVT_H_
+#define _GVT_H_
+
+#include <uapi/linux/pci_regs.h>
+#include <linux/kvm_host.h>
+#include <linux/vfio.h>
+#include <linux/mdev.h>
+
+#include "i915_drv.h"
+#include "intel_gvt.h"
+
+#include "debug.h"
+#include "mmio.h"
+#include "reg.h"
+#include "interrupt.h"
+#include "gtt.h"
+#include "display.h"
+#include "edid.h"
+#include "execlist.h"
+#include "scheduler.h"
+#include "sched_policy.h"
+#include "mmio_context.h"
+#include "cmd_parser.h"
+#include "fb_decoder.h"
+#include "dmabuf.h"
+#include "page_track.h"
+
+#define GVT_MAX_VGPU 8
+
+/* Describe per-platform limitations. */
+struct intel_gvt_device_info {
+	u32 max_support_vgpus;
+	u32 cfg_space_size;
+	u32 mmio_size;
+	u32 mmio_bar;
+	unsigned long msi_cap_offset;
+	u32 gtt_start_offset;
+	u32 gtt_entry_size;
+	u32 gtt_entry_size_shift;
+	int gmadr_bytes_in_cmd;
+	u32 max_surface_size;
+};
+
+/* GM resources owned by a vGPU */
+struct intel_vgpu_gm {
+	u64 aperture_sz;
+	u64 hidden_sz;
+	struct drm_mm_node low_gm_node;
+	struct drm_mm_node high_gm_node;
+};
+
+#define INTEL_GVT_MAX_NUM_FENCES 32
+
+/* Fences owned by a vGPU */
+struct intel_vgpu_fence {
+	struct i915_fence_reg *regs[INTEL_GVT_MAX_NUM_FENCES];
+	u32 base;
+	u32 size;
+};
+
+struct intel_vgpu_mmio {
+	void *vreg;
+};
+
+#define INTEL_GVT_MAX_BAR_NUM 4
+
+struct intel_vgpu_pci_bar {
+	u64 size;
+	bool tracked;
+};
+
+struct intel_vgpu_cfg_space {
+	unsigned char virtual_cfg_space[PCI_CFG_SPACE_EXP_SIZE];
+	struct intel_vgpu_pci_bar bar[INTEL_GVT_MAX_BAR_NUM];
+	u32 pmcsr_off;
+};
+
+#define vgpu_cfg_space(vgpu) ((vgpu)->cfg_space.virtual_cfg_space)
+
+struct intel_vgpu_irq {
+	bool irq_warn_once[INTEL_GVT_EVENT_MAX];
+	DECLARE_BITMAP(flip_done_event[I915_MAX_PIPES],
+		       INTEL_GVT_EVENT_MAX);
+};
+
+struct intel_vgpu_opregion {
+	bool mapped;
+	void *va;
+	u32 gfn[INTEL_GVT_OPREGION_PAGES];
+};
+
+#define vgpu_opregion(vgpu) (&(vgpu->opregion))
+
+struct intel_vgpu_display {
+	struct intel_vgpu_i2c_edid i2c_edid;
+	struct intel_vgpu_port ports[I915_MAX_PORTS];
+	struct intel_vgpu_sbi sbi;
+	enum port port_num;
+};
+
+struct vgpu_sched_ctl {
+	int weight;
+};
+
+enum {
+	INTEL_VGPU_EXECLIST_SUBMISSION = 1,
+	INTEL_VGPU_GUC_SUBMISSION,
+};
+
+struct intel_vgpu_submission_ops {
+	const char *name;
+	int (*init)(struct intel_vgpu *vgpu, intel_engine_mask_t engine_mask);
+	void (*clean)(struct intel_vgpu *vgpu, intel_engine_mask_t engine_mask);
+	void (*reset)(struct intel_vgpu *vgpu, intel_engine_mask_t engine_mask);
+};
+
+struct intel_vgpu_submission {
+	struct intel_vgpu_execlist execlist[I915_NUM_ENGINES];
+	struct list_head workload_q_head[I915_NUM_ENGINES];
+	struct intel_context *shadow[I915_NUM_ENGINES];
+	struct kmem_cache *workloads;
+	atomic_t running_workload_num;
+	union {
+		u64 i915_context_pml4;
+		u64 i915_context_pdps[GEN8_3LVL_PDPES];
+	};
+	DECLARE_BITMAP(shadow_ctx_desc_updated, I915_NUM_ENGINES);
+	DECLARE_BITMAP(tlb_handle_pending, I915_NUM_ENGINES);
+	void *ring_scan_buffer[I915_NUM_ENGINES];
+	int ring_scan_buffer_size[I915_NUM_ENGINES];
+	const struct intel_vgpu_submission_ops *ops;
+	int virtual_submission_interface;
+	bool active;
+	struct {
+		u32 lrca;
+		bool valid;
+		u64 ring_context_gpa;
+	} last_ctx[I915_NUM_ENGINES];
+};
+
+#define KVMGT_DEBUGFS_FILENAME		"kvmgt_nr_cache_entries"
+
+struct intel_vgpu {
+	struct vfio_device vfio_device;
+	struct intel_gvt *gvt;
+	struct mutex vgpu_lock;
+	int id;
+	bool active;
+	bool attached;
+	bool pv_notified;
+	bool failsafe;
+	unsigned int resetting_eng;
+
+	/* Both sched_data and sched_ctl can be seen a part of the global gvt
+	 * scheduler structure. So below 2 vgpu data are protected
+	 * by sched_lock, not vgpu_lock.
+	 */
+	void *sched_data;
+	struct vgpu_sched_ctl sched_ctl;
+
+	struct intel_vgpu_fence fence;
+	struct intel_vgpu_gm gm;
+	struct intel_vgpu_cfg_space cfg_space;
+	struct intel_vgpu_mmio mmio;
+	struct intel_vgpu_irq irq;
+	struct intel_vgpu_gtt gtt;
+	struct intel_vgpu_opregion opregion;
+	struct intel_vgpu_display display;
+	struct intel_vgpu_submission submission;
+	struct radix_tree_root page_track_tree;
+	u32 hws_pga[I915_NUM_ENGINES];
+	/* Set on PCI_D3, reset on DMLR, not reflecting the actual PM state */
+	bool d3_entered;
+
+	struct dentry *debugfs;
+
+	struct list_head dmabuf_obj_list_head;
+	struct mutex dmabuf_lock;
+	struct idr object_idr;
+	struct intel_vgpu_vblank_timer vblank_timer;
+
+	u32 scan_nonprivbb;
+
+	struct vfio_region *region;
+	int num_regions;
+	struct eventfd_ctx *intx_trigger;
+	struct eventfd_ctx *msi_trigger;
+
+	/*
+	 * Two caches are used to avoid mapping duplicated pages (eg.
+	 * scratch pages). This help to reduce dma setup overhead.
+	 */
+	struct rb_root gfn_cache;
+	struct rb_root dma_addr_cache;
+	unsigned long nr_cache_entries;
+	struct mutex cache_lock;
+
+	atomic_t released;
+
+	struct kvm_page_track_notifier_node track_node;
+#define NR_BKT (1 << 18)
+	struct hlist_head ptable[NR_BKT];
+#undef NR_BKT
+};
+
+/* validating GM healthy status*/
+#define vgpu_is_vm_unhealthy(ret_val) \
+	(((ret_val) == -EBADRQC) || ((ret_val) == -EFAULT))
+
+struct intel_gvt_gm {
+	unsigned long vgpu_allocated_low_gm_size;
+	unsigned long vgpu_allocated_high_gm_size;
+};
+
+struct intel_gvt_fence {
+	unsigned long vgpu_allocated_fence_num;
+};
+
+/* Special MMIO blocks. */
+struct gvt_mmio_block {
+	unsigned int device;
+	i915_reg_t   offset;
+	unsigned int size;
+	gvt_mmio_func read;
+	gvt_mmio_func write;
+};
+
+#define INTEL_GVT_MMIO_HASH_BITS 11
+
+struct intel_gvt_mmio {
+	u16 *mmio_attribute;
+/* Register contains RO bits */
+#define F_RO		(1 << 0)
+/* Register contains graphics address */
+#define F_GMADR		(1 << 1)
+/* Mode mask registers with high 16 bits as the mask bits */
+#define F_MODE_MASK	(1 << 2)
+/* This reg can be accessed by GPU commands */
+#define F_CMD_ACCESS	(1 << 3)
+/* This reg has been accessed by a VM */
+#define F_ACCESSED	(1 << 4)
+/* This reg requires save & restore during host PM suspend/resume */
+#define F_PM_SAVE	(1 << 5)
+/* This reg could be accessed by unaligned address */
+#define F_UNALIGN	(1 << 6)
+/* This reg is in GVT's mmio save-restor list and in hardware
+ * logical context image
+ */
+#define F_SR_IN_CTX	(1 << 7)
+/* Value of command write of this reg needs to be patched */
+#define F_CMD_WRITE_PATCH	(1 << 8)
+
+	struct gvt_mmio_block *mmio_block;
+	unsigned int num_mmio_block;
+
+	DECLARE_HASHTABLE(mmio_info_table, INTEL_GVT_MMIO_HASH_BITS);
+	unsigned long num_tracked_mmio;
+};
+
+struct intel_gvt_firmware {
+	void *cfg_space;
+	void *mmio;
+	bool firmware_loaded;
+};
+
+struct intel_vgpu_config {
+	unsigned int low_mm;
+	unsigned int high_mm;
+	unsigned int fence;
+
+	/*
+	 * A vGPU with a weight of 8 will get twice as much GPU as a vGPU with
+	 * a weight of 4 on a contended host, different vGPU type has different
+	 * weight set. Legal weights range from 1 to 16.
+	 */
+	unsigned int weight;
+	enum intel_vgpu_edid edid;
+	const char *name;
+};
+
+struct intel_vgpu_type {
+	struct mdev_type type;
+	char name[16];
+	const struct intel_vgpu_config *conf;
+};
+
+struct intel_gvt {
+	/* GVT scope lock, protect GVT itself, and all resource currently
+	 * not yet protected by special locks(vgpu and scheduler lock).
+	 */
+	struct mutex lock;
+	/* scheduler scope lock, protect gvt and vgpu schedule related data */
+	struct mutex sched_lock;
+
+	struct intel_gt *gt;
+	struct idr vgpu_idr;	/* vGPU IDR pool */
+
+	struct intel_gvt_device_info device_info;
+	struct intel_gvt_gm gm;
+	struct intel_gvt_fence fence;
+	struct intel_gvt_mmio mmio;
+	struct intel_gvt_firmware firmware;
+	struct intel_gvt_irq irq;
+	struct intel_gvt_gtt gtt;
+	struct intel_gvt_workload_scheduler scheduler;
+	struct notifier_block shadow_ctx_notifier_block[I915_NUM_ENGINES];
+	DECLARE_HASHTABLE(cmd_table, GVT_CMD_HASH_BITS);
+	struct mdev_parent parent;
+	struct mdev_type **mdev_types;
+	struct intel_vgpu_type *types;
+	unsigned int num_types;
+	struct intel_vgpu *idle_vgpu;
+
+	struct task_struct *service_thread;
+	wait_queue_head_t service_thread_wq;
+
+	/* service_request is always used in bit operation, we should always
+	 * use it with atomic bit ops so that no need to use gvt big lock.
+	 */
+	unsigned long service_request;
+
+	struct {
+		struct engine_mmio *mmio;
+		int ctx_mmio_count[I915_NUM_ENGINES];
+		u32 *tlb_mmio_offset_list;
+		u32 tlb_mmio_offset_list_cnt;
+		u32 *mocs_mmio_offset_list;
+		u32 mocs_mmio_offset_list_cnt;
+	} engine_mmio_list;
+	bool is_reg_whitelist_updated;
+
+	struct dentry *debugfs_root;
+};
+
+static inline struct intel_gvt *to_gvt(struct drm_i915_private *i915)
+{
+	return i915->gvt;
+}
+
+enum {
+	/* Scheduling trigger by timer */
+	INTEL_GVT_REQUEST_SCHED = 0,
+
+	/* Scheduling trigger by event */
+	INTEL_GVT_REQUEST_EVENT_SCHED = 1,
+
+	/* per-vGPU vblank emulation request */
+	INTEL_GVT_REQUEST_EMULATE_VBLANK = 2,
+	INTEL_GVT_REQUEST_EMULATE_VBLANK_MAX = INTEL_GVT_REQUEST_EMULATE_VBLANK
+		+ GVT_MAX_VGPU,
+};
+
+static inline void intel_gvt_request_service(struct intel_gvt *gvt,
+		int service)
+{
+	set_bit(service, (void *)&gvt->service_request);
+	wake_up(&gvt->service_thread_wq);
+}
+
+void intel_gvt_free_firmware(struct intel_gvt *gvt);
+int intel_gvt_load_firmware(struct intel_gvt *gvt);
+
+/* Aperture/GM space definitions for GVT device */
+#define MB_TO_BYTES(mb) ((mb) << 20ULL)
+#define BYTES_TO_MB(b) ((b) >> 20ULL)
+
+#define HOST_LOW_GM_SIZE MB_TO_BYTES(128)
+#define HOST_HIGH_GM_SIZE MB_TO_BYTES(384)
+#define HOST_FENCE 4
+
+#define gvt_to_ggtt(gvt)	((gvt)->gt->ggtt)
+
+/* Aperture/GM space definitions for GVT device */
+#define gvt_aperture_sz(gvt)	  gvt_to_ggtt(gvt)->mappable_end
+#define gvt_aperture_pa_base(gvt) gvt_to_ggtt(gvt)->gmadr.start
+
+#define gvt_ggtt_gm_sz(gvt)	gvt_to_ggtt(gvt)->vm.total
+#define gvt_ggtt_sz(gvt)	(gvt_to_ggtt(gvt)->vm.total >> PAGE_SHIFT << 3)
+#define gvt_hidden_sz(gvt)	(gvt_ggtt_gm_sz(gvt) - gvt_aperture_sz(gvt))
+
+#define gvt_aperture_gmadr_base(gvt) (0)
+#define gvt_aperture_gmadr_end(gvt) (gvt_aperture_gmadr_base(gvt) \
+				     + gvt_aperture_sz(gvt) - 1)
+
+#define gvt_hidden_gmadr_base(gvt) (gvt_aperture_gmadr_base(gvt) \
+				    + gvt_aperture_sz(gvt))
+#define gvt_hidden_gmadr_end(gvt) (gvt_hidden_gmadr_base(gvt) \
+				   + gvt_hidden_sz(gvt) - 1)
+
+#define gvt_fence_sz(gvt) (gvt_to_ggtt(gvt)->num_fences)
+
+/* Aperture/GM space definitions for vGPU */
+#define vgpu_aperture_offset(vgpu)	((vgpu)->gm.low_gm_node.start)
+#define vgpu_hidden_offset(vgpu)	((vgpu)->gm.high_gm_node.start)
+#define vgpu_aperture_sz(vgpu)		((vgpu)->gm.aperture_sz)
+#define vgpu_hidden_sz(vgpu)		((vgpu)->gm.hidden_sz)
+
+#define vgpu_aperture_pa_base(vgpu) \
+	(gvt_aperture_pa_base(vgpu->gvt) + vgpu_aperture_offset(vgpu))
+
+#define vgpu_ggtt_gm_sz(vgpu) ((vgpu)->gm.aperture_sz + (vgpu)->gm.hidden_sz)
+
+#define vgpu_aperture_pa_end(vgpu) \
+	(vgpu_aperture_pa_base(vgpu) + vgpu_aperture_sz(vgpu) - 1)
+
+#define vgpu_aperture_gmadr_base(vgpu) (vgpu_aperture_offset(vgpu))
+#define vgpu_aperture_gmadr_end(vgpu) \
+	(vgpu_aperture_gmadr_base(vgpu) + vgpu_aperture_sz(vgpu) - 1)
+
+#define vgpu_hidden_gmadr_base(vgpu) (vgpu_hidden_offset(vgpu))
+#define vgpu_hidden_gmadr_end(vgpu) \
+	(vgpu_hidden_gmadr_base(vgpu) + vgpu_hidden_sz(vgpu) - 1)
+
+#define vgpu_fence_base(vgpu) (vgpu->fence.base)
+#define vgpu_fence_sz(vgpu) (vgpu->fence.size)
+
+/* ring context size i.e. the first 0x50 dwords*/
+#define RING_CTX_SIZE 320
+
+int intel_vgpu_alloc_resource(struct intel_vgpu *vgpu,
+			      const struct intel_vgpu_config *conf);
+void intel_vgpu_reset_resource(struct intel_vgpu *vgpu);
+void intel_vgpu_free_resource(struct intel_vgpu *vgpu);
+void intel_vgpu_write_fence(struct intel_vgpu *vgpu,
+	u32 fence, u64 value);
+
+/* Macros for easily accessing vGPU virtual/shadow register.
+   Explicitly seperate use for typed MMIO reg or real offset.*/
+#define vgpu_vreg_t(vgpu, reg) \
+	(*(u32 *)(vgpu->mmio.vreg + i915_mmio_reg_offset(reg)))
+#define vgpu_vreg(vgpu, offset) \
+	(*(u32 *)(vgpu->mmio.vreg + (offset)))
+#define vgpu_vreg64_t(vgpu, reg) \
+	(*(u64 *)(vgpu->mmio.vreg + i915_mmio_reg_offset(reg)))
+#define vgpu_vreg64(vgpu, offset) \
+	(*(u64 *)(vgpu->mmio.vreg + (offset)))
+
+#define for_each_active_vgpu(gvt, vgpu, id) \
+	idr_for_each_entry((&(gvt)->vgpu_idr), (vgpu), (id)) \
+		for_each_if(vgpu->active)
+
+static inline void intel_vgpu_write_pci_bar(struct intel_vgpu *vgpu,
+					    u32 offset, u32 val, bool low)
+{
+	u32 *pval;
+
+	/* BAR offset should be 32 bits algiend */
+	offset = rounddown(offset, 4);
+	pval = (u32 *)(vgpu_cfg_space(vgpu) + offset);
+
+	if (low) {
+		/*
+		 * only update bit 31 - bit 4,
+		 * leave the bit 3 - bit 0 unchanged.
+		 */
+		*pval = (val & GENMASK(31, 4)) | (*pval & GENMASK(3, 0));
+	} else {
+		*pval = val;
+	}
+}
+
+int intel_gvt_init_vgpu_types(struct intel_gvt *gvt);
+void intel_gvt_clean_vgpu_types(struct intel_gvt *gvt);
+
+struct intel_vgpu *intel_gvt_create_idle_vgpu(struct intel_gvt *gvt);
+void intel_gvt_destroy_idle_vgpu(struct intel_vgpu *vgpu);
+int intel_gvt_create_vgpu(struct intel_vgpu *vgpu,
+			  const struct intel_vgpu_config *conf);
+void intel_gvt_destroy_vgpu(struct intel_vgpu *vgpu);
+void intel_gvt_release_vgpu(struct intel_vgpu *vgpu);
+void intel_gvt_reset_vgpu_locked(struct intel_vgpu *vgpu, bool dmlr,
+				 intel_engine_mask_t engine_mask);
+void intel_gvt_reset_vgpu(struct intel_vgpu *vgpu);
+void intel_gvt_activate_vgpu(struct intel_vgpu *vgpu);
+void intel_gvt_deactivate_vgpu(struct intel_vgpu *vgpu);
+
+int intel_gvt_set_opregion(struct intel_vgpu *vgpu);
+int intel_gvt_set_edid(struct intel_vgpu *vgpu, int port_num);
+
+/* validating GM functions */
+#define vgpu_gmadr_is_aperture(vgpu, gmadr) \
+	((gmadr >= vgpu_aperture_gmadr_base(vgpu)) && \
+	 (gmadr <= vgpu_aperture_gmadr_end(vgpu)))
+
+#define vgpu_gmadr_is_hidden(vgpu, gmadr) \
+	((gmadr >= vgpu_hidden_gmadr_base(vgpu)) && \
+	 (gmadr <= vgpu_hidden_gmadr_end(vgpu)))
+
+#define vgpu_gmadr_is_valid(vgpu, gmadr) \
+	 ((vgpu_gmadr_is_aperture(vgpu, gmadr) || \
+	  (vgpu_gmadr_is_hidden(vgpu, gmadr))))
+
+#define gvt_gmadr_is_aperture(gvt, gmadr) \
+	 ((gmadr >= gvt_aperture_gmadr_base(gvt)) && \
+	  (gmadr <= gvt_aperture_gmadr_end(gvt)))
+
+#define gvt_gmadr_is_hidden(gvt, gmadr) \
+	  ((gmadr >= gvt_hidden_gmadr_base(gvt)) && \
+	   (gmadr <= gvt_hidden_gmadr_end(gvt)))
+
+#define gvt_gmadr_is_valid(gvt, gmadr) \
+	  (gvt_gmadr_is_aperture(gvt, gmadr) || \
+	    gvt_gmadr_is_hidden(gvt, gmadr))
+
+bool intel_gvt_ggtt_validate_range(struct intel_vgpu *vgpu, u64 addr, u32 size);
+int intel_gvt_ggtt_gmadr_g2h(struct intel_vgpu *vgpu, u64 g_addr, u64 *h_addr);
+int intel_gvt_ggtt_gmadr_h2g(struct intel_vgpu *vgpu, u64 h_addr, u64 *g_addr);
+int intel_gvt_ggtt_index_g2h(struct intel_vgpu *vgpu, unsigned long g_index,
+			     unsigned long *h_index);
+int intel_gvt_ggtt_h2g_index(struct intel_vgpu *vgpu, unsigned long h_index,
+			     unsigned long *g_index);
+
+void intel_vgpu_init_cfg_space(struct intel_vgpu *vgpu,
+		bool primary);
+void intel_vgpu_reset_cfg_space(struct intel_vgpu *vgpu);
+
+int intel_vgpu_emulate_cfg_read(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes);
+
+int intel_vgpu_emulate_cfg_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes);
+
+void intel_vgpu_emulate_hotplug(struct intel_vgpu *vgpu, bool connected);
+
+static inline u64 intel_vgpu_get_bar_gpa(struct intel_vgpu *vgpu, int bar)
+{
+	/* We are 64bit bar. */
+	return (*(u64 *)(vgpu->cfg_space.virtual_cfg_space + bar)) &
+			PCI_BASE_ADDRESS_MEM_MASK;
+}
+
+void intel_vgpu_clean_opregion(struct intel_vgpu *vgpu);
+int intel_vgpu_init_opregion(struct intel_vgpu *vgpu);
+int intel_vgpu_opregion_base_write_handler(struct intel_vgpu *vgpu, u32 gpa);
+
+int intel_vgpu_emulate_opregion_request(struct intel_vgpu *vgpu, u32 swsci);
+void populate_pvinfo_page(struct intel_vgpu *vgpu);
+
+int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload);
+void enter_failsafe_mode(struct intel_vgpu *vgpu, int reason);
+void intel_vgpu_detach_regions(struct intel_vgpu *vgpu);
+
+enum {
+	GVT_FAILSAFE_UNSUPPORTED_GUEST,
+	GVT_FAILSAFE_INSUFFICIENT_RESOURCE,
+	GVT_FAILSAFE_GUEST_ERR,
+};
+
+static inline void mmio_hw_access_pre(struct intel_gt *gt)
+{
+	intel_runtime_pm_get(gt->uncore->rpm);
+}
+
+static inline void mmio_hw_access_post(struct intel_gt *gt)
+{
+	intel_runtime_pm_put_unchecked(gt->uncore->rpm);
+}
+
+/**
+ * intel_gvt_mmio_set_accessed - mark a MMIO has been accessed
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ */
+static inline void intel_gvt_mmio_set_accessed(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	gvt->mmio.mmio_attribute[offset >> 2] |= F_ACCESSED;
+}
+
+/**
+ * intel_gvt_mmio_is_cmd_accessible - if a MMIO could be accessed by command
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ * Returns:
+ * True if an MMIO is able to be accessed by GPU commands
+ */
+static inline bool intel_gvt_mmio_is_cmd_accessible(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	return gvt->mmio.mmio_attribute[offset >> 2] & F_CMD_ACCESS;
+}
+
+/**
+ * intel_gvt_mmio_set_cmd_accessible -
+ *				mark a MMIO could be accessible by command
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ */
+static inline void intel_gvt_mmio_set_cmd_accessible(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	gvt->mmio.mmio_attribute[offset >> 2] |= F_CMD_ACCESS;
+}
+
+/**
+ * intel_gvt_mmio_is_unalign - mark a MMIO could be accessed unaligned
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ */
+static inline bool intel_gvt_mmio_is_unalign(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	return gvt->mmio.mmio_attribute[offset >> 2] & F_UNALIGN;
+}
+
+/**
+ * intel_gvt_mmio_has_mode_mask - if a MMIO has a mode mask
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ * Returns:
+ * True if a MMIO has a mode mask in its higher 16 bits, false if it isn't.
+ *
+ */
+static inline bool intel_gvt_mmio_has_mode_mask(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	return gvt->mmio.mmio_attribute[offset >> 2] & F_MODE_MASK;
+}
+
+/**
+ * intel_gvt_mmio_is_sr_in_ctx -
+ *		check if an MMIO has F_SR_IN_CTX mask
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ * Returns:
+ * True if an MMIO has an F_SR_IN_CTX  mask, false if it isn't.
+ *
+ */
+static inline bool intel_gvt_mmio_is_sr_in_ctx(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	return gvt->mmio.mmio_attribute[offset >> 2] & F_SR_IN_CTX;
+}
+
+/**
+ * intel_gvt_mmio_set_sr_in_ctx -
+ *		mask an MMIO in GVT's mmio save-restore list and also
+ *		in hardware logical context image
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ */
+static inline void intel_gvt_mmio_set_sr_in_ctx(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	gvt->mmio.mmio_attribute[offset >> 2] |= F_SR_IN_CTX;
+}
+
+void intel_gvt_debugfs_add_vgpu(struct intel_vgpu *vgpu);
+/**
+ * intel_gvt_mmio_set_cmd_write_patch -
+ *				mark an MMIO if its cmd write needs to be
+ *				patched
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ */
+static inline void intel_gvt_mmio_set_cmd_write_patch(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	gvt->mmio.mmio_attribute[offset >> 2] |= F_CMD_WRITE_PATCH;
+}
+
+/**
+ * intel_gvt_mmio_is_cmd_write_patch - check if an mmio's cmd access needs to
+ * be patched
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ * Returns:
+ * True if GPU commmand write to an MMIO should be patched
+ */
+static inline bool intel_gvt_mmio_is_cmd_write_patch(
+			struct intel_gvt *gvt, unsigned int offset)
+{
+	return gvt->mmio.mmio_attribute[offset >> 2] & F_CMD_WRITE_PATCH;
+}
+
+/**
+ * intel_gvt_read_gpa - copy data from GPA to host data buffer
+ * @vgpu: a vGPU
+ * @gpa: guest physical address
+ * @buf: host data buffer
+ * @len: data length
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+static inline int intel_gvt_read_gpa(struct intel_vgpu *vgpu, unsigned long gpa,
+		void *buf, unsigned long len)
+{
+	if (!vgpu->attached)
+		return -ESRCH;
+	return vfio_dma_rw(&vgpu->vfio_device, gpa, buf, len, false);
+}
+
+/**
+ * intel_gvt_write_gpa - copy data from host data buffer to GPA
+ * @vgpu: a vGPU
+ * @gpa: guest physical address
+ * @buf: host data buffer
+ * @len: data length
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+static inline int intel_gvt_write_gpa(struct intel_vgpu *vgpu,
+		unsigned long gpa, void *buf, unsigned long len)
+{
+	if (!vgpu->attached)
+		return -ESRCH;
+	return vfio_dma_rw(&vgpu->vfio_device, gpa, buf, len, true);
+}
+
+void intel_gvt_debugfs_remove_vgpu(struct intel_vgpu *vgpu);
+void intel_gvt_debugfs_init(struct intel_gvt *gvt);
+void intel_gvt_debugfs_clean(struct intel_gvt *gvt);
+
+int intel_gvt_page_track_add(struct intel_vgpu *info, u64 gfn);
+int intel_gvt_page_track_remove(struct intel_vgpu *info, u64 gfn);
+int intel_gvt_dma_pin_guest_page(struct intel_vgpu *vgpu, dma_addr_t dma_addr);
+int intel_gvt_dma_map_guest_page(struct intel_vgpu *vgpu, unsigned long gfn,
+		unsigned long size, dma_addr_t *dma_addr);
+void intel_gvt_dma_unmap_guest_page(struct intel_vgpu *vgpu,
+		dma_addr_t dma_addr);
+
+#include "trace.h"
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c
new file mode 100644
index 000000000..daac2050d
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/handlers.c
@@ -0,0 +1,3231 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Kevin Tian <kevin.tian@intel.com>
+ *    Eddie Dong <eddie.dong@intel.com>
+ *    Zhiyuan Lv <zhiyuan.lv@intel.com>
+ *
+ * Contributors:
+ *    Min He <min.he@intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *    Pei Zhang <pei.zhang@intel.com>
+ *    Niu Bing <bing.niu@intel.com>
+ *    Ping Gao <ping.a.gao@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "gvt.h"
+#include "i915_pvinfo.h"
+#include "intel_mchbar_regs.h"
+#include "display/intel_display_types.h"
+#include "display/intel_dmc_regs.h"
+#include "display/intel_fbc.h"
+#include "display/vlv_dsi_pll_regs.h"
+#include "gt/intel_gt_regs.h"
+
+/* XXX FIXME i915 has changed PP_XXX definition */
+#define PCH_PP_STATUS  _MMIO(0xc7200)
+#define PCH_PP_CONTROL _MMIO(0xc7204)
+#define PCH_PP_ON_DELAYS _MMIO(0xc7208)
+#define PCH_PP_OFF_DELAYS _MMIO(0xc720c)
+#define PCH_PP_DIVISOR _MMIO(0xc7210)
+
+unsigned long intel_gvt_get_device_type(struct intel_gvt *gvt)
+{
+	struct drm_i915_private *i915 = gvt->gt->i915;
+
+	if (IS_BROADWELL(i915))
+		return D_BDW;
+	else if (IS_SKYLAKE(i915))
+		return D_SKL;
+	else if (IS_KABYLAKE(i915))
+		return D_KBL;
+	else if (IS_BROXTON(i915))
+		return D_BXT;
+	else if (IS_COFFEELAKE(i915) || IS_COMETLAKE(i915))
+		return D_CFL;
+
+	return 0;
+}
+
+static bool intel_gvt_match_device(struct intel_gvt *gvt,
+		unsigned long device)
+{
+	return intel_gvt_get_device_type(gvt) & device;
+}
+
+static void read_vreg(struct intel_vgpu *vgpu, unsigned int offset,
+	void *p_data, unsigned int bytes)
+{
+	memcpy(p_data, &vgpu_vreg(vgpu, offset), bytes);
+}
+
+static void write_vreg(struct intel_vgpu *vgpu, unsigned int offset,
+	void *p_data, unsigned int bytes)
+{
+	memcpy(&vgpu_vreg(vgpu, offset), p_data, bytes);
+}
+
+struct intel_gvt_mmio_info *intel_gvt_find_mmio_info(struct intel_gvt *gvt,
+						  unsigned int offset)
+{
+	struct intel_gvt_mmio_info *e;
+
+	hash_for_each_possible(gvt->mmio.mmio_info_table, e, node, offset) {
+		if (e->offset == offset)
+			return e;
+	}
+	return NULL;
+}
+
+static int setup_mmio_info(struct intel_gvt *gvt, u32 offset, u32 size,
+			   u16 flags, u32 addr_mask, u32 ro_mask, u32 device,
+			   gvt_mmio_func read, gvt_mmio_func write)
+{
+	struct intel_gvt_mmio_info *p;
+	u32 start, end, i;
+
+	if (!intel_gvt_match_device(gvt, device))
+		return 0;
+
+	if (WARN_ON(!IS_ALIGNED(offset, 4)))
+		return -EINVAL;
+
+	start = offset;
+	end = offset + size;
+
+	for (i = start; i < end; i += 4) {
+		p = intel_gvt_find_mmio_info(gvt, i);
+		if (!p) {
+			WARN(1, "assign a handler to a non-tracked mmio %x\n",
+				i);
+			return -ENODEV;
+		}
+		p->ro_mask = ro_mask;
+		gvt->mmio.mmio_attribute[i / 4] = flags;
+		if (read)
+			p->read = read;
+		if (write)
+			p->write = write;
+	}
+	return 0;
+}
+
+/**
+ * intel_gvt_render_mmio_to_engine - convert a mmio offset into the engine
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ * Returns:
+ * The engine containing the offset within its mmio page.
+ */
+const struct intel_engine_cs *
+intel_gvt_render_mmio_to_engine(struct intel_gvt *gvt, unsigned int offset)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	offset &= ~GENMASK(11, 0);
+	for_each_engine(engine, gvt->gt, id)
+		if (engine->mmio_base == offset)
+			return engine;
+
+	return NULL;
+}
+
+#define offset_to_fence_num(offset) \
+	((offset - i915_mmio_reg_offset(FENCE_REG_GEN6_LO(0))) >> 3)
+
+#define fence_num_to_offset(num) \
+	(num * 8 + i915_mmio_reg_offset(FENCE_REG_GEN6_LO(0)))
+
+
+void enter_failsafe_mode(struct intel_vgpu *vgpu, int reason)
+{
+	switch (reason) {
+	case GVT_FAILSAFE_UNSUPPORTED_GUEST:
+		pr_err("Detected your guest driver doesn't support GVT-g.\n");
+		break;
+	case GVT_FAILSAFE_INSUFFICIENT_RESOURCE:
+		pr_err("Graphics resource is not enough for the guest\n");
+		break;
+	case GVT_FAILSAFE_GUEST_ERR:
+		pr_err("GVT Internal error  for the guest\n");
+		break;
+	default:
+		break;
+	}
+	pr_err("Now vgpu %d will enter failsafe mode.\n", vgpu->id);
+	vgpu->failsafe = true;
+}
+
+static int sanitize_fence_mmio_access(struct intel_vgpu *vgpu,
+		unsigned int fence_num, void *p_data, unsigned int bytes)
+{
+	unsigned int max_fence = vgpu_fence_sz(vgpu);
+
+	if (fence_num >= max_fence) {
+		gvt_vgpu_err("access oob fence reg %d/%d\n",
+			     fence_num, max_fence);
+
+		/* When guest access oob fence regs without access
+		 * pv_info first, we treat guest not supporting GVT,
+		 * and we will let vgpu enter failsafe mode.
+		 */
+		if (!vgpu->pv_notified)
+			enter_failsafe_mode(vgpu,
+					GVT_FAILSAFE_UNSUPPORTED_GUEST);
+
+		memset(p_data, 0, bytes);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static int gamw_echo_dev_rw_ia_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 ips = (*(u32 *)p_data) & GAMW_ECO_ENABLE_64K_IPS_FIELD;
+
+	if (GRAPHICS_VER(vgpu->gvt->gt->i915) <= 10) {
+		if (ips == GAMW_ECO_ENABLE_64K_IPS_FIELD)
+			gvt_dbg_core("vgpu%d: ips enabled\n", vgpu->id);
+		else if (!ips)
+			gvt_dbg_core("vgpu%d: ips disabled\n", vgpu->id);
+		else {
+			/* All engines must be enabled together for vGPU,
+			 * since we don't know which engine the ppgtt will
+			 * bind to when shadowing.
+			 */
+			gvt_vgpu_err("Unsupported IPS setting %x, cannot enable 64K gtt.\n",
+				     ips);
+			return -EINVAL;
+		}
+	}
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	return 0;
+}
+
+static int fence_mmio_read(struct intel_vgpu *vgpu, unsigned int off,
+		void *p_data, unsigned int bytes)
+{
+	int ret;
+
+	ret = sanitize_fence_mmio_access(vgpu, offset_to_fence_num(off),
+			p_data, bytes);
+	if (ret)
+		return ret;
+	read_vreg(vgpu, off, p_data, bytes);
+	return 0;
+}
+
+static int fence_mmio_write(struct intel_vgpu *vgpu, unsigned int off,
+		void *p_data, unsigned int bytes)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	unsigned int fence_num = offset_to_fence_num(off);
+	int ret;
+
+	ret = sanitize_fence_mmio_access(vgpu, fence_num, p_data, bytes);
+	if (ret)
+		return ret;
+	write_vreg(vgpu, off, p_data, bytes);
+
+	mmio_hw_access_pre(gvt->gt);
+	intel_vgpu_write_fence(vgpu, fence_num,
+			vgpu_vreg64(vgpu, fence_num_to_offset(fence_num)));
+	mmio_hw_access_post(gvt->gt);
+	return 0;
+}
+
+#define CALC_MODE_MASK_REG(old, new) \
+	(((new) & GENMASK(31, 16)) \
+	 | ((((old) & GENMASK(15, 0)) & ~((new) >> 16)) \
+	 | ((new) & ((new) >> 16))))
+
+static int mul_force_wake_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 old, new;
+	u32 ack_reg_offset;
+
+	old = vgpu_vreg(vgpu, offset);
+	new = CALC_MODE_MASK_REG(old, *(u32 *)p_data);
+
+	if (GRAPHICS_VER(vgpu->gvt->gt->i915)  >=  9) {
+		switch (offset) {
+		case FORCEWAKE_RENDER_GEN9_REG:
+			ack_reg_offset = FORCEWAKE_ACK_RENDER_GEN9_REG;
+			break;
+		case FORCEWAKE_GT_GEN9_REG:
+			ack_reg_offset = FORCEWAKE_ACK_GT_GEN9_REG;
+			break;
+		case FORCEWAKE_MEDIA_GEN9_REG:
+			ack_reg_offset = FORCEWAKE_ACK_MEDIA_GEN9_REG;
+			break;
+		default:
+			/*should not hit here*/
+			gvt_vgpu_err("invalid forcewake offset 0x%x\n", offset);
+			return -EINVAL;
+		}
+	} else {
+		ack_reg_offset = FORCEWAKE_ACK_HSW_REG;
+	}
+
+	vgpu_vreg(vgpu, offset) = new;
+	vgpu_vreg(vgpu, ack_reg_offset) = (new & GENMASK(15, 0));
+	return 0;
+}
+
+static int gdrst_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+			    void *p_data, unsigned int bytes)
+{
+	intel_engine_mask_t engine_mask = 0;
+	u32 data;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	data = vgpu_vreg(vgpu, offset);
+
+	if (data & GEN6_GRDOM_FULL) {
+		gvt_dbg_mmio("vgpu%d: request full GPU reset\n", vgpu->id);
+		engine_mask = ALL_ENGINES;
+	} else {
+		if (data & GEN6_GRDOM_RENDER) {
+			gvt_dbg_mmio("vgpu%d: request RCS reset\n", vgpu->id);
+			engine_mask |= BIT(RCS0);
+		}
+		if (data & GEN6_GRDOM_MEDIA) {
+			gvt_dbg_mmio("vgpu%d: request VCS reset\n", vgpu->id);
+			engine_mask |= BIT(VCS0);
+		}
+		if (data & GEN6_GRDOM_BLT) {
+			gvt_dbg_mmio("vgpu%d: request BCS Reset\n", vgpu->id);
+			engine_mask |= BIT(BCS0);
+		}
+		if (data & GEN6_GRDOM_VECS) {
+			gvt_dbg_mmio("vgpu%d: request VECS Reset\n", vgpu->id);
+			engine_mask |= BIT(VECS0);
+		}
+		if (data & GEN8_GRDOM_MEDIA2) {
+			gvt_dbg_mmio("vgpu%d: request VCS2 Reset\n", vgpu->id);
+			engine_mask |= BIT(VCS1);
+		}
+		if (data & GEN9_GRDOM_GUC) {
+			gvt_dbg_mmio("vgpu%d: request GUC Reset\n", vgpu->id);
+			vgpu_vreg_t(vgpu, GUC_STATUS) |= GS_MIA_IN_RESET;
+		}
+		engine_mask &= vgpu->gvt->gt->info.engine_mask;
+	}
+
+	/* vgpu_lock already hold by emulate mmio r/w */
+	intel_gvt_reset_vgpu_locked(vgpu, false, engine_mask);
+
+	/* sw will wait for the device to ack the reset request */
+	vgpu_vreg(vgpu, offset) = 0;
+
+	return 0;
+}
+
+static int gmbus_mmio_read(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	return intel_gvt_i2c_handle_gmbus_read(vgpu, offset, p_data, bytes);
+}
+
+static int gmbus_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	return intel_gvt_i2c_handle_gmbus_write(vgpu, offset, p_data, bytes);
+}
+
+static int pch_pp_control_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg(vgpu, offset) & PANEL_POWER_ON) {
+		vgpu_vreg_t(vgpu, PCH_PP_STATUS) |= PP_ON;
+		vgpu_vreg_t(vgpu, PCH_PP_STATUS) |= PP_SEQUENCE_STATE_ON_IDLE;
+		vgpu_vreg_t(vgpu, PCH_PP_STATUS) &= ~PP_SEQUENCE_POWER_DOWN;
+		vgpu_vreg_t(vgpu, PCH_PP_STATUS) &= ~PP_CYCLE_DELAY_ACTIVE;
+
+	} else
+		vgpu_vreg_t(vgpu, PCH_PP_STATUS) &=
+			~(PP_ON | PP_SEQUENCE_POWER_DOWN
+					| PP_CYCLE_DELAY_ACTIVE);
+	return 0;
+}
+
+static int transconf_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg(vgpu, offset) & TRANS_ENABLE)
+		vgpu_vreg(vgpu, offset) |= TRANS_STATE_ENABLE;
+	else
+		vgpu_vreg(vgpu, offset) &= ~TRANS_STATE_ENABLE;
+	return 0;
+}
+
+static int lcpll_ctl_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg(vgpu, offset) & LCPLL_PLL_DISABLE)
+		vgpu_vreg(vgpu, offset) &= ~LCPLL_PLL_LOCK;
+	else
+		vgpu_vreg(vgpu, offset) |= LCPLL_PLL_LOCK;
+
+	if (vgpu_vreg(vgpu, offset) & LCPLL_CD_SOURCE_FCLK)
+		vgpu_vreg(vgpu, offset) |= LCPLL_CD_SOURCE_FCLK_DONE;
+	else
+		vgpu_vreg(vgpu, offset) &= ~LCPLL_CD_SOURCE_FCLK_DONE;
+
+	return 0;
+}
+
+static int dpy_reg_mmio_read(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	switch (offset) {
+	case 0xe651c:
+	case 0xe661c:
+	case 0xe671c:
+	case 0xe681c:
+		vgpu_vreg(vgpu, offset) = 1 << 17;
+		break;
+	case 0xe6c04:
+		vgpu_vreg(vgpu, offset) = 0x3;
+		break;
+	case 0xe6e1c:
+		vgpu_vreg(vgpu, offset) = 0x2f << 16;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	read_vreg(vgpu, offset, p_data, bytes);
+	return 0;
+}
+
+/*
+ * Only PIPE_A is enabled in current vGPU display and PIPE_A is tied to
+ *   TRANSCODER_A in HW. DDI/PORT could be PORT_x depends on
+ *   setup_virtual_dp_monitor().
+ * emulate_monitor_status_change() set up PLL for PORT_x as the initial enabled
+ *   DPLL. Later guest driver may setup a different DPLLx when setting mode.
+ * So the correct sequence to find DP stream clock is:
+ *   Check TRANS_DDI_FUNC_CTL on TRANSCODER_A to get PORT_x.
+ *   Check correct PLLx for PORT_x to get PLL frequency and DP bitrate.
+ * Then Refresh rate then can be calculated based on follow equations:
+ *   Pixel clock = h_total * v_total * refresh_rate
+ *   stream clock = Pixel clock
+ *   ls_clk = DP bitrate
+ *   Link M/N = strm_clk / ls_clk
+ */
+
+static u32 bdw_vgpu_get_dp_bitrate(struct intel_vgpu *vgpu, enum port port)
+{
+	u32 dp_br = 0;
+	u32 ddi_pll_sel = vgpu_vreg_t(vgpu, PORT_CLK_SEL(port));
+
+	switch (ddi_pll_sel) {
+	case PORT_CLK_SEL_LCPLL_2700:
+		dp_br = 270000 * 2;
+		break;
+	case PORT_CLK_SEL_LCPLL_1350:
+		dp_br = 135000 * 2;
+		break;
+	case PORT_CLK_SEL_LCPLL_810:
+		dp_br = 81000 * 2;
+		break;
+	case PORT_CLK_SEL_SPLL:
+	{
+		switch (vgpu_vreg_t(vgpu, SPLL_CTL) & SPLL_FREQ_MASK) {
+		case SPLL_FREQ_810MHz:
+			dp_br = 81000 * 2;
+			break;
+		case SPLL_FREQ_1350MHz:
+			dp_br = 135000 * 2;
+			break;
+		case SPLL_FREQ_2700MHz:
+			dp_br = 270000 * 2;
+			break;
+		default:
+			gvt_dbg_dpy("vgpu-%d PORT_%c can't get freq from SPLL 0x%08x\n",
+				    vgpu->id, port_name(port), vgpu_vreg_t(vgpu, SPLL_CTL));
+			break;
+		}
+		break;
+	}
+	case PORT_CLK_SEL_WRPLL1:
+	case PORT_CLK_SEL_WRPLL2:
+	{
+		u32 wrpll_ctl;
+		int refclk, n, p, r;
+
+		if (ddi_pll_sel == PORT_CLK_SEL_WRPLL1)
+			wrpll_ctl = vgpu_vreg_t(vgpu, WRPLL_CTL(DPLL_ID_WRPLL1));
+		else
+			wrpll_ctl = vgpu_vreg_t(vgpu, WRPLL_CTL(DPLL_ID_WRPLL2));
+
+		switch (wrpll_ctl & WRPLL_REF_MASK) {
+		case WRPLL_REF_PCH_SSC:
+			refclk = vgpu->gvt->gt->i915->display.dpll.ref_clks.ssc;
+			break;
+		case WRPLL_REF_LCPLL:
+			refclk = 2700000;
+			break;
+		default:
+			gvt_dbg_dpy("vgpu-%d PORT_%c WRPLL can't get refclk 0x%08x\n",
+				    vgpu->id, port_name(port), wrpll_ctl);
+			goto out;
+		}
+
+		r = wrpll_ctl & WRPLL_DIVIDER_REF_MASK;
+		p = (wrpll_ctl & WRPLL_DIVIDER_POST_MASK) >> WRPLL_DIVIDER_POST_SHIFT;
+		n = (wrpll_ctl & WRPLL_DIVIDER_FB_MASK) >> WRPLL_DIVIDER_FB_SHIFT;
+
+		dp_br = (refclk * n / 10) / (p * r) * 2;
+		break;
+	}
+	default:
+		gvt_dbg_dpy("vgpu-%d PORT_%c has invalid clock select 0x%08x\n",
+			    vgpu->id, port_name(port), vgpu_vreg_t(vgpu, PORT_CLK_SEL(port)));
+		break;
+	}
+
+out:
+	return dp_br;
+}
+
+static u32 bxt_vgpu_get_dp_bitrate(struct intel_vgpu *vgpu, enum port port)
+{
+	u32 dp_br = 0;
+	int refclk = vgpu->gvt->gt->i915->display.dpll.ref_clks.nssc;
+	enum dpio_phy phy = DPIO_PHY0;
+	enum dpio_channel ch = DPIO_CH0;
+	struct dpll clock = {0};
+	u32 temp;
+
+	/* Port to PHY mapping is fixed, see bxt_ddi_phy_info{} */
+	switch (port) {
+	case PORT_A:
+		phy = DPIO_PHY1;
+		ch = DPIO_CH0;
+		break;
+	case PORT_B:
+		phy = DPIO_PHY0;
+		ch = DPIO_CH0;
+		break;
+	case PORT_C:
+		phy = DPIO_PHY0;
+		ch = DPIO_CH1;
+		break;
+	default:
+		gvt_dbg_dpy("vgpu-%d no PHY for PORT_%c\n", vgpu->id, port_name(port));
+		goto out;
+	}
+
+	temp = vgpu_vreg_t(vgpu, BXT_PORT_PLL_ENABLE(port));
+	if (!(temp & PORT_PLL_ENABLE) || !(temp & PORT_PLL_LOCK)) {
+		gvt_dbg_dpy("vgpu-%d PORT_%c PLL_ENABLE 0x%08x isn't enabled or locked\n",
+			    vgpu->id, port_name(port), temp);
+		goto out;
+	}
+
+	clock.m1 = 2;
+	clock.m2 = REG_FIELD_GET(PORT_PLL_M2_INT_MASK,
+				 vgpu_vreg_t(vgpu, BXT_PORT_PLL(phy, ch, 0))) << 22;
+	if (vgpu_vreg_t(vgpu, BXT_PORT_PLL(phy, ch, 3)) & PORT_PLL_M2_FRAC_ENABLE)
+		clock.m2 |= REG_FIELD_GET(PORT_PLL_M2_FRAC_MASK,
+					  vgpu_vreg_t(vgpu, BXT_PORT_PLL(phy, ch, 2)));
+	clock.n = REG_FIELD_GET(PORT_PLL_N_MASK,
+				vgpu_vreg_t(vgpu, BXT_PORT_PLL(phy, ch, 1)));
+	clock.p1 = REG_FIELD_GET(PORT_PLL_P1_MASK,
+				 vgpu_vreg_t(vgpu, BXT_PORT_PLL_EBB_0(phy, ch)));
+	clock.p2 = REG_FIELD_GET(PORT_PLL_P2_MASK,
+				 vgpu_vreg_t(vgpu, BXT_PORT_PLL_EBB_0(phy, ch)));
+	clock.m = clock.m1 * clock.m2;
+	clock.p = clock.p1 * clock.p2 * 5;
+
+	if (clock.n == 0 || clock.p == 0) {
+		gvt_dbg_dpy("vgpu-%d PORT_%c PLL has invalid divider\n", vgpu->id, port_name(port));
+		goto out;
+	}
+
+	clock.vco = DIV_ROUND_CLOSEST_ULL(mul_u32_u32(refclk, clock.m), clock.n << 22);
+	clock.dot = DIV_ROUND_CLOSEST(clock.vco, clock.p);
+
+	dp_br = clock.dot;
+
+out:
+	return dp_br;
+}
+
+static u32 skl_vgpu_get_dp_bitrate(struct intel_vgpu *vgpu, enum port port)
+{
+	u32 dp_br = 0;
+	enum intel_dpll_id dpll_id = DPLL_ID_SKL_DPLL0;
+
+	/* Find the enabled DPLL for the DDI/PORT */
+	if (!(vgpu_vreg_t(vgpu, DPLL_CTRL2) & DPLL_CTRL2_DDI_CLK_OFF(port)) &&
+	    (vgpu_vreg_t(vgpu, DPLL_CTRL2) & DPLL_CTRL2_DDI_SEL_OVERRIDE(port))) {
+		dpll_id += (vgpu_vreg_t(vgpu, DPLL_CTRL2) &
+			DPLL_CTRL2_DDI_CLK_SEL_MASK(port)) >>
+			DPLL_CTRL2_DDI_CLK_SEL_SHIFT(port);
+	} else {
+		gvt_dbg_dpy("vgpu-%d DPLL for PORT_%c isn't turned on\n",
+			    vgpu->id, port_name(port));
+		return dp_br;
+	}
+
+	/* Find PLL output frequency from correct DPLL, and get bir rate */
+	switch ((vgpu_vreg_t(vgpu, DPLL_CTRL1) &
+		DPLL_CTRL1_LINK_RATE_MASK(dpll_id)) >>
+		DPLL_CTRL1_LINK_RATE_SHIFT(dpll_id)) {
+		case DPLL_CTRL1_LINK_RATE_810:
+			dp_br = 81000 * 2;
+			break;
+		case DPLL_CTRL1_LINK_RATE_1080:
+			dp_br = 108000 * 2;
+			break;
+		case DPLL_CTRL1_LINK_RATE_1350:
+			dp_br = 135000 * 2;
+			break;
+		case DPLL_CTRL1_LINK_RATE_1620:
+			dp_br = 162000 * 2;
+			break;
+		case DPLL_CTRL1_LINK_RATE_2160:
+			dp_br = 216000 * 2;
+			break;
+		case DPLL_CTRL1_LINK_RATE_2700:
+			dp_br = 270000 * 2;
+			break;
+		default:
+			dp_br = 0;
+			gvt_dbg_dpy("vgpu-%d PORT_%c fail to get DPLL-%d freq\n",
+				    vgpu->id, port_name(port), dpll_id);
+	}
+
+	return dp_br;
+}
+
+static void vgpu_update_refresh_rate(struct intel_vgpu *vgpu)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+	enum port port;
+	u32 dp_br, link_m, link_n, htotal, vtotal;
+
+	/* Find DDI/PORT assigned to TRANSCODER_A, expect B or D */
+	port = (vgpu_vreg_t(vgpu, TRANS_DDI_FUNC_CTL(TRANSCODER_A)) &
+		TRANS_DDI_PORT_MASK) >> TRANS_DDI_PORT_SHIFT;
+	if (port != PORT_B && port != PORT_D) {
+		gvt_dbg_dpy("vgpu-%d unsupported PORT_%c\n", vgpu->id, port_name(port));
+		return;
+	}
+
+	/* Calculate DP bitrate from PLL */
+	if (IS_BROADWELL(dev_priv))
+		dp_br = bdw_vgpu_get_dp_bitrate(vgpu, port);
+	else if (IS_BROXTON(dev_priv))
+		dp_br = bxt_vgpu_get_dp_bitrate(vgpu, port);
+	else
+		dp_br = skl_vgpu_get_dp_bitrate(vgpu, port);
+
+	/* Get DP link symbol clock M/N */
+	link_m = vgpu_vreg_t(vgpu, PIPE_LINK_M1(TRANSCODER_A));
+	link_n = vgpu_vreg_t(vgpu, PIPE_LINK_N1(TRANSCODER_A));
+
+	/* Get H/V total from transcoder timing */
+	htotal = (vgpu_vreg_t(vgpu, HTOTAL(TRANSCODER_A)) >> TRANS_HTOTAL_SHIFT);
+	vtotal = (vgpu_vreg_t(vgpu, VTOTAL(TRANSCODER_A)) >> TRANS_VTOTAL_SHIFT);
+
+	if (dp_br && link_n && htotal && vtotal) {
+		u64 pixel_clk = 0;
+		u32 new_rate = 0;
+		u32 *old_rate = &(intel_vgpu_port(vgpu, vgpu->display.port_num)->vrefresh_k);
+
+		/* Calcuate pixel clock by (ls_clk * M / N) */
+		pixel_clk = div_u64(mul_u32_u32(link_m, dp_br), link_n);
+		pixel_clk *= MSEC_PER_SEC;
+
+		/* Calcuate refresh rate by (pixel_clk / (h_total * v_total)) */
+		new_rate = DIV64_U64_ROUND_CLOSEST(mul_u64_u32_shr(pixel_clk, MSEC_PER_SEC, 0), mul_u32_u32(htotal + 1, vtotal + 1));
+
+		if (*old_rate != new_rate)
+			*old_rate = new_rate;
+
+		gvt_dbg_dpy("vgpu-%d PIPE_%c refresh rate updated to %d\n",
+			    vgpu->id, pipe_name(PIPE_A), new_rate);
+	}
+}
+
+static int pipeconf_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 data;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	data = vgpu_vreg(vgpu, offset);
+
+	if (data & PIPECONF_ENABLE) {
+		vgpu_vreg(vgpu, offset) |= PIPECONF_STATE_ENABLE;
+		vgpu_update_refresh_rate(vgpu);
+		vgpu_update_vblank_emulation(vgpu, true);
+	} else {
+		vgpu_vreg(vgpu, offset) &= ~PIPECONF_STATE_ENABLE;
+		vgpu_update_vblank_emulation(vgpu, false);
+	}
+	return 0;
+}
+
+/* sorted in ascending order */
+static i915_reg_t force_nonpriv_white_list[] = {
+	_MMIO(0xd80),
+	GEN9_CS_DEBUG_MODE1, //_MMIO(0x20ec)
+	GEN9_CTX_PREEMPT_REG,//_MMIO(0x2248)
+	CL_PRIMITIVES_COUNT, //_MMIO(0x2340)
+	PS_INVOCATION_COUNT, //_MMIO(0x2348)
+	PS_DEPTH_COUNT, //_MMIO(0x2350)
+	GEN8_CS_CHICKEN1,//_MMIO(0x2580)
+	_MMIO(0x2690),
+	_MMIO(0x2694),
+	_MMIO(0x2698),
+	_MMIO(0x2754),
+	_MMIO(0x28a0),
+	_MMIO(0x4de0),
+	_MMIO(0x4de4),
+	_MMIO(0x4dfc),
+	GEN7_COMMON_SLICE_CHICKEN1,//_MMIO(0x7010)
+	_MMIO(0x7014),
+	HDC_CHICKEN0,//_MMIO(0x7300)
+	GEN8_HDC_CHICKEN1,//_MMIO(0x7304)
+	_MMIO(0x7700),
+	_MMIO(0x7704),
+	_MMIO(0x7708),
+	_MMIO(0x770c),
+	_MMIO(0x83a8),
+	_MMIO(0xb110),
+	GEN8_L3SQCREG4,//_MMIO(0xb118)
+	_MMIO(0xe100),
+	_MMIO(0xe18c),
+	_MMIO(0xe48c),
+	_MMIO(0xe5f4),
+	_MMIO(0x64844),
+};
+
+/* a simple bsearch */
+static inline bool in_whitelist(u32 reg)
+{
+	int left = 0, right = ARRAY_SIZE(force_nonpriv_white_list);
+	i915_reg_t *array = force_nonpriv_white_list;
+
+	while (left < right) {
+		int mid = (left + right)/2;
+
+		if (reg > array[mid].reg)
+			left = mid + 1;
+		else if (reg < array[mid].reg)
+			right = mid;
+		else
+			return true;
+	}
+	return false;
+}
+
+static int force_nonpriv_write(struct intel_vgpu *vgpu,
+	unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 reg_nonpriv = (*(u32 *)p_data) & REG_GENMASK(25, 2);
+	const struct intel_engine_cs *engine =
+		intel_gvt_render_mmio_to_engine(vgpu->gvt, offset);
+
+	if (bytes != 4 || !IS_ALIGNED(offset, bytes) || !engine) {
+		gvt_err("vgpu(%d) Invalid FORCE_NONPRIV offset %x(%dB)\n",
+			vgpu->id, offset, bytes);
+		return -EINVAL;
+	}
+
+	if (!in_whitelist(reg_nonpriv) &&
+	    reg_nonpriv != i915_mmio_reg_offset(RING_NOPID(engine->mmio_base))) {
+		gvt_err("vgpu(%d) Invalid FORCE_NONPRIV write %x at offset %x\n",
+			vgpu->id, reg_nonpriv, offset);
+	} else
+		intel_vgpu_default_mmio_write(vgpu, offset, p_data, bytes);
+
+	return 0;
+}
+
+static int ddi_buf_ctl_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg(vgpu, offset) & DDI_BUF_CTL_ENABLE) {
+		vgpu_vreg(vgpu, offset) &= ~DDI_BUF_IS_IDLE;
+	} else {
+		vgpu_vreg(vgpu, offset) |= DDI_BUF_IS_IDLE;
+		if (offset == i915_mmio_reg_offset(DDI_BUF_CTL(PORT_E)))
+			vgpu_vreg_t(vgpu, DP_TP_STATUS(PORT_E))
+				&= ~DP_TP_STATUS_AUTOTRAIN_DONE;
+	}
+	return 0;
+}
+
+static int fdi_rx_iir_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	vgpu_vreg(vgpu, offset) &= ~*(u32 *)p_data;
+	return 0;
+}
+
+#define FDI_LINK_TRAIN_PATTERN1         0
+#define FDI_LINK_TRAIN_PATTERN2         1
+
+static int fdi_auto_training_started(struct intel_vgpu *vgpu)
+{
+	u32 ddi_buf_ctl = vgpu_vreg_t(vgpu, DDI_BUF_CTL(PORT_E));
+	u32 rx_ctl = vgpu_vreg(vgpu, _FDI_RXA_CTL);
+	u32 tx_ctl = vgpu_vreg_t(vgpu, DP_TP_CTL(PORT_E));
+
+	if ((ddi_buf_ctl & DDI_BUF_CTL_ENABLE) &&
+			(rx_ctl & FDI_RX_ENABLE) &&
+			(rx_ctl & FDI_AUTO_TRAINING) &&
+			(tx_ctl & DP_TP_CTL_ENABLE) &&
+			(tx_ctl & DP_TP_CTL_FDI_AUTOTRAIN))
+		return 1;
+	else
+		return 0;
+}
+
+static int check_fdi_rx_train_status(struct intel_vgpu *vgpu,
+		enum pipe pipe, unsigned int train_pattern)
+{
+	i915_reg_t fdi_rx_imr, fdi_tx_ctl, fdi_rx_ctl;
+	unsigned int fdi_rx_check_bits, fdi_tx_check_bits;
+	unsigned int fdi_rx_train_bits, fdi_tx_train_bits;
+	unsigned int fdi_iir_check_bits;
+
+	fdi_rx_imr = FDI_RX_IMR(pipe);
+	fdi_tx_ctl = FDI_TX_CTL(pipe);
+	fdi_rx_ctl = FDI_RX_CTL(pipe);
+
+	if (train_pattern == FDI_LINK_TRAIN_PATTERN1) {
+		fdi_rx_train_bits = FDI_LINK_TRAIN_PATTERN_1_CPT;
+		fdi_tx_train_bits = FDI_LINK_TRAIN_PATTERN_1;
+		fdi_iir_check_bits = FDI_RX_BIT_LOCK;
+	} else if (train_pattern == FDI_LINK_TRAIN_PATTERN2) {
+		fdi_rx_train_bits = FDI_LINK_TRAIN_PATTERN_2_CPT;
+		fdi_tx_train_bits = FDI_LINK_TRAIN_PATTERN_2;
+		fdi_iir_check_bits = FDI_RX_SYMBOL_LOCK;
+	} else {
+		gvt_vgpu_err("Invalid train pattern %d\n", train_pattern);
+		return -EINVAL;
+	}
+
+	fdi_rx_check_bits = FDI_RX_ENABLE | fdi_rx_train_bits;
+	fdi_tx_check_bits = FDI_TX_ENABLE | fdi_tx_train_bits;
+
+	/* If imr bit has been masked */
+	if (vgpu_vreg_t(vgpu, fdi_rx_imr) & fdi_iir_check_bits)
+		return 0;
+
+	if (((vgpu_vreg_t(vgpu, fdi_tx_ctl) & fdi_tx_check_bits)
+			== fdi_tx_check_bits)
+		&& ((vgpu_vreg_t(vgpu, fdi_rx_ctl) & fdi_rx_check_bits)
+			== fdi_rx_check_bits))
+		return 1;
+	else
+		return 0;
+}
+
+#define INVALID_INDEX (~0U)
+
+static unsigned int calc_index(unsigned int offset, unsigned int start,
+	unsigned int next, unsigned int end, i915_reg_t i915_end)
+{
+	unsigned int range = next - start;
+
+	if (!end)
+		end = i915_mmio_reg_offset(i915_end);
+	if (offset < start || offset > end)
+		return INVALID_INDEX;
+	offset -= start;
+	return offset / range;
+}
+
+#define FDI_RX_CTL_TO_PIPE(offset) \
+	calc_index(offset, _FDI_RXA_CTL, _FDI_RXB_CTL, 0, FDI_RX_CTL(PIPE_C))
+
+#define FDI_TX_CTL_TO_PIPE(offset) \
+	calc_index(offset, _FDI_TXA_CTL, _FDI_TXB_CTL, 0, FDI_TX_CTL(PIPE_C))
+
+#define FDI_RX_IMR_TO_PIPE(offset) \
+	calc_index(offset, _FDI_RXA_IMR, _FDI_RXB_IMR, 0, FDI_RX_IMR(PIPE_C))
+
+static int update_fdi_rx_iir_status(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	i915_reg_t fdi_rx_iir;
+	unsigned int index;
+	int ret;
+
+	if (FDI_RX_CTL_TO_PIPE(offset) != INVALID_INDEX)
+		index = FDI_RX_CTL_TO_PIPE(offset);
+	else if (FDI_TX_CTL_TO_PIPE(offset) != INVALID_INDEX)
+		index = FDI_TX_CTL_TO_PIPE(offset);
+	else if (FDI_RX_IMR_TO_PIPE(offset) != INVALID_INDEX)
+		index = FDI_RX_IMR_TO_PIPE(offset);
+	else {
+		gvt_vgpu_err("Unsupported registers %x\n", offset);
+		return -EINVAL;
+	}
+
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	fdi_rx_iir = FDI_RX_IIR(index);
+
+	ret = check_fdi_rx_train_status(vgpu, index, FDI_LINK_TRAIN_PATTERN1);
+	if (ret < 0)
+		return ret;
+	if (ret)
+		vgpu_vreg_t(vgpu, fdi_rx_iir) |= FDI_RX_BIT_LOCK;
+
+	ret = check_fdi_rx_train_status(vgpu, index, FDI_LINK_TRAIN_PATTERN2);
+	if (ret < 0)
+		return ret;
+	if (ret)
+		vgpu_vreg_t(vgpu, fdi_rx_iir) |= FDI_RX_SYMBOL_LOCK;
+
+	if (offset == _FDI_RXA_CTL)
+		if (fdi_auto_training_started(vgpu))
+			vgpu_vreg_t(vgpu, DP_TP_STATUS(PORT_E)) |=
+				DP_TP_STATUS_AUTOTRAIN_DONE;
+	return 0;
+}
+
+#define DP_TP_CTL_TO_PORT(offset) \
+	calc_index(offset, _DP_TP_CTL_A, _DP_TP_CTL_B, 0, DP_TP_CTL(PORT_E))
+
+static int dp_tp_ctl_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	i915_reg_t status_reg;
+	unsigned int index;
+	u32 data;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	index = DP_TP_CTL_TO_PORT(offset);
+	data = (vgpu_vreg(vgpu, offset) & GENMASK(10, 8)) >> 8;
+	if (data == 0x2) {
+		status_reg = DP_TP_STATUS(index);
+		vgpu_vreg_t(vgpu, status_reg) |= (1 << 25);
+	}
+	return 0;
+}
+
+static int dp_tp_status_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 reg_val;
+	u32 sticky_mask;
+
+	reg_val = *((u32 *)p_data);
+	sticky_mask = GENMASK(27, 26) | (1 << 24);
+
+	vgpu_vreg(vgpu, offset) = (reg_val & ~sticky_mask) |
+		(vgpu_vreg(vgpu, offset) & sticky_mask);
+	vgpu_vreg(vgpu, offset) &= ~(reg_val & sticky_mask);
+	return 0;
+}
+
+static int pch_adpa_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 data;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	data = vgpu_vreg(vgpu, offset);
+
+	if (data & ADPA_CRT_HOTPLUG_FORCE_TRIGGER)
+		vgpu_vreg(vgpu, offset) &= ~ADPA_CRT_HOTPLUG_FORCE_TRIGGER;
+	return 0;
+}
+
+static int south_chicken2_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 data;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	data = vgpu_vreg(vgpu, offset);
+
+	if (data & FDI_MPHY_IOSFSB_RESET_CTL)
+		vgpu_vreg(vgpu, offset) |= FDI_MPHY_IOSFSB_RESET_STATUS;
+	else
+		vgpu_vreg(vgpu, offset) &= ~FDI_MPHY_IOSFSB_RESET_STATUS;
+	return 0;
+}
+
+#define DSPSURF_TO_PIPE(offset) \
+	calc_index(offset, _DSPASURF, _DSPBSURF, 0, DSPSURF(PIPE_C))
+
+static int pri_surf_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+	u32 pipe = DSPSURF_TO_PIPE(offset);
+	int event = SKL_FLIP_EVENT(pipe, PLANE_PRIMARY);
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	vgpu_vreg_t(vgpu, DSPSURFLIVE(pipe)) = vgpu_vreg(vgpu, offset);
+
+	vgpu_vreg_t(vgpu, PIPE_FLIPCOUNT_G4X(pipe))++;
+
+	if (vgpu_vreg_t(vgpu, DSPCNTR(pipe)) & PLANE_CTL_ASYNC_FLIP)
+		intel_vgpu_trigger_virtual_event(vgpu, event);
+	else
+		set_bit(event, vgpu->irq.flip_done_event[pipe]);
+
+	return 0;
+}
+
+#define SPRSURF_TO_PIPE(offset) \
+	calc_index(offset, _SPRA_SURF, _SPRB_SURF, 0, SPRSURF(PIPE_C))
+
+static int spr_surf_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 pipe = SPRSURF_TO_PIPE(offset);
+	int event = SKL_FLIP_EVENT(pipe, PLANE_SPRITE0);
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	vgpu_vreg_t(vgpu, SPRSURFLIVE(pipe)) = vgpu_vreg(vgpu, offset);
+
+	if (vgpu_vreg_t(vgpu, SPRCTL(pipe)) & PLANE_CTL_ASYNC_FLIP)
+		intel_vgpu_trigger_virtual_event(vgpu, event);
+	else
+		set_bit(event, vgpu->irq.flip_done_event[pipe]);
+
+	return 0;
+}
+
+static int reg50080_mmio_write(struct intel_vgpu *vgpu,
+			       unsigned int offset, void *p_data,
+			       unsigned int bytes)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+	enum pipe pipe = REG_50080_TO_PIPE(offset);
+	enum plane_id plane = REG_50080_TO_PLANE(offset);
+	int event = SKL_FLIP_EVENT(pipe, plane);
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	if (plane == PLANE_PRIMARY) {
+		vgpu_vreg_t(vgpu, DSPSURFLIVE(pipe)) = vgpu_vreg(vgpu, offset);
+		vgpu_vreg_t(vgpu, PIPE_FLIPCOUNT_G4X(pipe))++;
+	} else {
+		vgpu_vreg_t(vgpu, SPRSURFLIVE(pipe)) = vgpu_vreg(vgpu, offset);
+	}
+
+	if ((vgpu_vreg(vgpu, offset) & REG50080_FLIP_TYPE_MASK) == REG50080_FLIP_TYPE_ASYNC)
+		intel_vgpu_trigger_virtual_event(vgpu, event);
+	else
+		set_bit(event, vgpu->irq.flip_done_event[pipe]);
+
+	return 0;
+}
+
+static int trigger_aux_channel_interrupt(struct intel_vgpu *vgpu,
+		unsigned int reg)
+{
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+	enum intel_gvt_event_type event;
+
+	if (reg == i915_mmio_reg_offset(DP_AUX_CH_CTL(AUX_CH_A)))
+		event = AUX_CHANNEL_A;
+	else if (reg == _PCH_DPB_AUX_CH_CTL ||
+		 reg == i915_mmio_reg_offset(DP_AUX_CH_CTL(AUX_CH_B)))
+		event = AUX_CHANNEL_B;
+	else if (reg == _PCH_DPC_AUX_CH_CTL ||
+		 reg == i915_mmio_reg_offset(DP_AUX_CH_CTL(AUX_CH_C)))
+		event = AUX_CHANNEL_C;
+	else if (reg == _PCH_DPD_AUX_CH_CTL ||
+		 reg == i915_mmio_reg_offset(DP_AUX_CH_CTL(AUX_CH_D)))
+		event = AUX_CHANNEL_D;
+	else {
+		drm_WARN_ON(&dev_priv->drm, true);
+		return -EINVAL;
+	}
+
+	intel_vgpu_trigger_virtual_event(vgpu, event);
+	return 0;
+}
+
+static int dp_aux_ch_ctl_trans_done(struct intel_vgpu *vgpu, u32 value,
+		unsigned int reg, int len, bool data_valid)
+{
+	/* mark transaction done */
+	value |= DP_AUX_CH_CTL_DONE;
+	value &= ~DP_AUX_CH_CTL_SEND_BUSY;
+	value &= ~DP_AUX_CH_CTL_RECEIVE_ERROR;
+
+	if (data_valid)
+		value &= ~DP_AUX_CH_CTL_TIME_OUT_ERROR;
+	else
+		value |= DP_AUX_CH_CTL_TIME_OUT_ERROR;
+
+	/* message size */
+	value &= ~(0xf << 20);
+	value |= (len << 20);
+	vgpu_vreg(vgpu, reg) = value;
+
+	if (value & DP_AUX_CH_CTL_INTERRUPT)
+		return trigger_aux_channel_interrupt(vgpu, reg);
+	return 0;
+}
+
+static void dp_aux_ch_ctl_link_training(struct intel_vgpu_dpcd_data *dpcd,
+		u8 t)
+{
+	if ((t & DPCD_TRAINING_PATTERN_SET_MASK) == DPCD_TRAINING_PATTERN_1) {
+		/* training pattern 1 for CR */
+		/* set LANE0_CR_DONE, LANE1_CR_DONE */
+		dpcd->data[DPCD_LANE0_1_STATUS] |= DPCD_LANES_CR_DONE;
+		/* set LANE2_CR_DONE, LANE3_CR_DONE */
+		dpcd->data[DPCD_LANE2_3_STATUS] |= DPCD_LANES_CR_DONE;
+	} else if ((t & DPCD_TRAINING_PATTERN_SET_MASK) ==
+			DPCD_TRAINING_PATTERN_2) {
+		/* training pattern 2 for EQ */
+		/* Set CHANNEL_EQ_DONE and  SYMBOL_LOCKED for Lane0_1 */
+		dpcd->data[DPCD_LANE0_1_STATUS] |= DPCD_LANES_EQ_DONE;
+		dpcd->data[DPCD_LANE0_1_STATUS] |= DPCD_SYMBOL_LOCKED;
+		/* Set CHANNEL_EQ_DONE and  SYMBOL_LOCKED for Lane2_3 */
+		dpcd->data[DPCD_LANE2_3_STATUS] |= DPCD_LANES_EQ_DONE;
+		dpcd->data[DPCD_LANE2_3_STATUS] |= DPCD_SYMBOL_LOCKED;
+		/* set INTERLANE_ALIGN_DONE */
+		dpcd->data[DPCD_LANE_ALIGN_STATUS_UPDATED] |=
+			DPCD_INTERLANE_ALIGN_DONE;
+	} else if ((t & DPCD_TRAINING_PATTERN_SET_MASK) ==
+			DPCD_LINK_TRAINING_DISABLED) {
+		/* finish link training */
+		/* set sink status as synchronized */
+		dpcd->data[DPCD_SINK_STATUS] = DPCD_SINK_IN_SYNC;
+	}
+}
+
+#define _REG_HSW_DP_AUX_CH_CTL(dp) \
+	((dp) ? (_PCH_DPB_AUX_CH_CTL + ((dp)-1)*0x100) : 0x64010)
+
+#define _REG_SKL_DP_AUX_CH_CTL(dp) (0x64010 + (dp) * 0x100)
+
+#define OFFSET_TO_DP_AUX_PORT(offset) (((offset) & 0xF00) >> 8)
+
+#define dpy_is_valid_port(port)	\
+		(((port) >= PORT_A) && ((port) < I915_MAX_PORTS))
+
+static int dp_aux_ch_ctl_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	struct intel_vgpu_display *display = &vgpu->display;
+	int msg, addr, ctrl, op, len;
+	int port_index = OFFSET_TO_DP_AUX_PORT(offset);
+	struct intel_vgpu_dpcd_data *dpcd = NULL;
+	struct intel_vgpu_port *port = NULL;
+	u32 data;
+
+	if (!dpy_is_valid_port(port_index)) {
+		gvt_vgpu_err("Unsupported DP port access!\n");
+		return 0;
+	}
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	data = vgpu_vreg(vgpu, offset);
+
+	if ((GRAPHICS_VER(vgpu->gvt->gt->i915) >= 9)
+		&& offset != _REG_SKL_DP_AUX_CH_CTL(port_index)) {
+		/* SKL DPB/C/D aux ctl register changed */
+		return 0;
+	} else if (IS_BROADWELL(vgpu->gvt->gt->i915) &&
+		   offset != _REG_HSW_DP_AUX_CH_CTL(port_index)) {
+		/* write to the data registers */
+		return 0;
+	}
+
+	if (!(data & DP_AUX_CH_CTL_SEND_BUSY)) {
+		/* just want to clear the sticky bits */
+		vgpu_vreg(vgpu, offset) = 0;
+		return 0;
+	}
+
+	port = &display->ports[port_index];
+	dpcd = port->dpcd;
+
+	/* read out message from DATA1 register */
+	msg = vgpu_vreg(vgpu, offset + 4);
+	addr = (msg >> 8) & 0xffff;
+	ctrl = (msg >> 24) & 0xff;
+	len = msg & 0xff;
+	op = ctrl >> 4;
+
+	if (op == GVT_AUX_NATIVE_WRITE) {
+		int t;
+		u8 buf[16];
+
+		if ((addr + len + 1) >= DPCD_SIZE) {
+			/*
+			 * Write request exceeds what we supported,
+			 * DCPD spec: When a Source Device is writing a DPCD
+			 * address not supported by the Sink Device, the Sink
+			 * Device shall reply with AUX NACK and “M” equal to
+			 * zero.
+			 */
+
+			/* NAK the write */
+			vgpu_vreg(vgpu, offset + 4) = AUX_NATIVE_REPLY_NAK;
+			dp_aux_ch_ctl_trans_done(vgpu, data, offset, 2, true);
+			return 0;
+		}
+
+		/*
+		 * Write request format: Headr (command + address + size) occupies
+		 * 4 bytes, followed by (len + 1) bytes of data. See details at
+		 * intel_dp_aux_transfer().
+		 */
+		if ((len + 1 + 4) > AUX_BURST_SIZE) {
+			gvt_vgpu_err("dp_aux_header: len %d is too large\n", len);
+			return -EINVAL;
+		}
+
+		/* unpack data from vreg to buf */
+		for (t = 0; t < 4; t++) {
+			u32 r = vgpu_vreg(vgpu, offset + 8 + t * 4);
+
+			buf[t * 4] = (r >> 24) & 0xff;
+			buf[t * 4 + 1] = (r >> 16) & 0xff;
+			buf[t * 4 + 2] = (r >> 8) & 0xff;
+			buf[t * 4 + 3] = r & 0xff;
+		}
+
+		/* write to virtual DPCD */
+		if (dpcd && dpcd->data_valid) {
+			for (t = 0; t <= len; t++) {
+				int p = addr + t;
+
+				dpcd->data[p] = buf[t];
+				/* check for link training */
+				if (p == DPCD_TRAINING_PATTERN_SET)
+					dp_aux_ch_ctl_link_training(dpcd,
+							buf[t]);
+			}
+		}
+
+		/* ACK the write */
+		vgpu_vreg(vgpu, offset + 4) = 0;
+		dp_aux_ch_ctl_trans_done(vgpu, data, offset, 1,
+				dpcd && dpcd->data_valid);
+		return 0;
+	}
+
+	if (op == GVT_AUX_NATIVE_READ) {
+		int idx, i, ret = 0;
+
+		if ((addr + len + 1) >= DPCD_SIZE) {
+			/*
+			 * read request exceeds what we supported
+			 * DPCD spec: A Sink Device receiving a Native AUX CH
+			 * read request for an unsupported DPCD address must
+			 * reply with an AUX ACK and read data set equal to
+			 * zero instead of replying with AUX NACK.
+			 */
+
+			/* ACK the READ*/
+			vgpu_vreg(vgpu, offset + 4) = 0;
+			vgpu_vreg(vgpu, offset + 8) = 0;
+			vgpu_vreg(vgpu, offset + 12) = 0;
+			vgpu_vreg(vgpu, offset + 16) = 0;
+			vgpu_vreg(vgpu, offset + 20) = 0;
+
+			dp_aux_ch_ctl_trans_done(vgpu, data, offset, len + 2,
+					true);
+			return 0;
+		}
+
+		for (idx = 1; idx <= 5; idx++) {
+			/* clear the data registers */
+			vgpu_vreg(vgpu, offset + 4 * idx) = 0;
+		}
+
+		/*
+		 * Read reply format: ACK (1 byte) plus (len + 1) bytes of data.
+		 */
+		if ((len + 2) > AUX_BURST_SIZE) {
+			gvt_vgpu_err("dp_aux_header: len %d is too large\n", len);
+			return -EINVAL;
+		}
+
+		/* read from virtual DPCD to vreg */
+		/* first 4 bytes: [ACK][addr][addr+1][addr+2] */
+		if (dpcd && dpcd->data_valid) {
+			for (i = 1; i <= (len + 1); i++) {
+				int t;
+
+				t = dpcd->data[addr + i - 1];
+				t <<= (24 - 8 * (i % 4));
+				ret |= t;
+
+				if ((i % 4 == 3) || (i == (len + 1))) {
+					vgpu_vreg(vgpu, offset +
+							(i / 4 + 1) * 4) = ret;
+					ret = 0;
+				}
+			}
+		}
+		dp_aux_ch_ctl_trans_done(vgpu, data, offset, len + 2,
+				dpcd && dpcd->data_valid);
+		return 0;
+	}
+
+	/* i2c transaction starts */
+	intel_gvt_i2c_handle_aux_ch_write(vgpu, port_index, offset, p_data);
+
+	if (data & DP_AUX_CH_CTL_INTERRUPT)
+		trigger_aux_channel_interrupt(vgpu, offset);
+	return 0;
+}
+
+static int mbctl_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	*(u32 *)p_data &= (~GEN6_MBCTL_ENABLE_BOOT_FETCH);
+	write_vreg(vgpu, offset, p_data, bytes);
+	return 0;
+}
+
+static int vga_control_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	bool vga_disable;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	vga_disable = vgpu_vreg(vgpu, offset) & VGA_DISP_DISABLE;
+
+	gvt_dbg_core("vgpu%d: %s VGA mode\n", vgpu->id,
+			vga_disable ? "Disable" : "Enable");
+	return 0;
+}
+
+static u32 read_virtual_sbi_register(struct intel_vgpu *vgpu,
+		unsigned int sbi_offset)
+{
+	struct intel_vgpu_display *display = &vgpu->display;
+	int num = display->sbi.number;
+	int i;
+
+	for (i = 0; i < num; ++i)
+		if (display->sbi.registers[i].offset == sbi_offset)
+			break;
+
+	if (i == num)
+		return 0;
+
+	return display->sbi.registers[i].value;
+}
+
+static void write_virtual_sbi_register(struct intel_vgpu *vgpu,
+		unsigned int offset, u32 value)
+{
+	struct intel_vgpu_display *display = &vgpu->display;
+	int num = display->sbi.number;
+	int i;
+
+	for (i = 0; i < num; ++i) {
+		if (display->sbi.registers[i].offset == offset)
+			break;
+	}
+
+	if (i == num) {
+		if (num == SBI_REG_MAX) {
+			gvt_vgpu_err("SBI caching meets maximum limits\n");
+			return;
+		}
+		display->sbi.number++;
+	}
+
+	display->sbi.registers[i].offset = offset;
+	display->sbi.registers[i].value = value;
+}
+
+static int sbi_data_mmio_read(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	if (((vgpu_vreg_t(vgpu, SBI_CTL_STAT) & SBI_OPCODE_MASK) >>
+				SBI_OPCODE_SHIFT) == SBI_CMD_CRRD) {
+		unsigned int sbi_offset = (vgpu_vreg_t(vgpu, SBI_ADDR) &
+				SBI_ADDR_OFFSET_MASK) >> SBI_ADDR_OFFSET_SHIFT;
+		vgpu_vreg(vgpu, offset) = read_virtual_sbi_register(vgpu,
+				sbi_offset);
+	}
+	read_vreg(vgpu, offset, p_data, bytes);
+	return 0;
+}
+
+static int sbi_ctl_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 data;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	data = vgpu_vreg(vgpu, offset);
+
+	data &= ~(SBI_STAT_MASK << SBI_STAT_SHIFT);
+	data |= SBI_READY;
+
+	data &= ~(SBI_RESPONSE_MASK << SBI_RESPONSE_SHIFT);
+	data |= SBI_RESPONSE_SUCCESS;
+
+	vgpu_vreg(vgpu, offset) = data;
+
+	if (((vgpu_vreg_t(vgpu, SBI_CTL_STAT) & SBI_OPCODE_MASK) >>
+				SBI_OPCODE_SHIFT) == SBI_CMD_CRWR) {
+		unsigned int sbi_offset = (vgpu_vreg_t(vgpu, SBI_ADDR) &
+				SBI_ADDR_OFFSET_MASK) >> SBI_ADDR_OFFSET_SHIFT;
+
+		write_virtual_sbi_register(vgpu, sbi_offset,
+					   vgpu_vreg_t(vgpu, SBI_DATA));
+	}
+	return 0;
+}
+
+#define _vgtif_reg(x) \
+	(VGT_PVINFO_PAGE + offsetof(struct vgt_if, x))
+
+static int pvinfo_mmio_read(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	bool invalid_read = false;
+
+	read_vreg(vgpu, offset, p_data, bytes);
+
+	switch (offset) {
+	case _vgtif_reg(magic) ... _vgtif_reg(vgt_id):
+		if (offset + bytes > _vgtif_reg(vgt_id) + 4)
+			invalid_read = true;
+		break;
+	case _vgtif_reg(avail_rs.mappable_gmadr.base) ...
+		_vgtif_reg(avail_rs.fence_num):
+		if (offset + bytes >
+			_vgtif_reg(avail_rs.fence_num) + 4)
+			invalid_read = true;
+		break;
+	case 0x78010:	/* vgt_caps */
+	case 0x7881c:
+		break;
+	default:
+		invalid_read = true;
+		break;
+	}
+	if (invalid_read)
+		gvt_vgpu_err("invalid pvinfo read: [%x:%x] = %x\n",
+				offset, bytes, *(u32 *)p_data);
+	vgpu->pv_notified = true;
+	return 0;
+}
+
+static int handle_g2v_notification(struct intel_vgpu *vgpu, int notification)
+{
+	enum intel_gvt_gtt_type root_entry_type = GTT_TYPE_PPGTT_ROOT_L4_ENTRY;
+	struct intel_vgpu_mm *mm;
+	u64 *pdps;
+
+	pdps = (u64 *)&vgpu_vreg64_t(vgpu, vgtif_reg(pdp[0]));
+
+	switch (notification) {
+	case VGT_G2V_PPGTT_L3_PAGE_TABLE_CREATE:
+		root_entry_type = GTT_TYPE_PPGTT_ROOT_L3_ENTRY;
+		fallthrough;
+	case VGT_G2V_PPGTT_L4_PAGE_TABLE_CREATE:
+		mm = intel_vgpu_get_ppgtt_mm(vgpu, root_entry_type, pdps);
+		return PTR_ERR_OR_ZERO(mm);
+	case VGT_G2V_PPGTT_L3_PAGE_TABLE_DESTROY:
+	case VGT_G2V_PPGTT_L4_PAGE_TABLE_DESTROY:
+		return intel_vgpu_put_ppgtt_mm(vgpu, pdps);
+	case VGT_G2V_EXECLIST_CONTEXT_CREATE:
+	case VGT_G2V_EXECLIST_CONTEXT_DESTROY:
+	case 1:	/* Remove this in guest driver. */
+		break;
+	default:
+		gvt_vgpu_err("Invalid PV notification %d\n", notification);
+	}
+	return 0;
+}
+
+static int send_display_ready_uevent(struct intel_vgpu *vgpu, int ready)
+{
+	struct kobject *kobj = &vgpu->gvt->gt->i915->drm.primary->kdev->kobj;
+	char *env[3] = {NULL, NULL, NULL};
+	char vmid_str[20];
+	char display_ready_str[20];
+
+	snprintf(display_ready_str, 20, "GVT_DISPLAY_READY=%d", ready);
+	env[0] = display_ready_str;
+
+	snprintf(vmid_str, 20, "VMID=%d", vgpu->id);
+	env[1] = vmid_str;
+
+	return kobject_uevent_env(kobj, KOBJ_ADD, env);
+}
+
+static int pvinfo_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 data = *(u32 *)p_data;
+	bool invalid_write = false;
+
+	switch (offset) {
+	case _vgtif_reg(display_ready):
+		send_display_ready_uevent(vgpu, data ? 1 : 0);
+		break;
+	case _vgtif_reg(g2v_notify):
+		handle_g2v_notification(vgpu, data);
+		break;
+	/* add xhot and yhot to handled list to avoid error log */
+	case _vgtif_reg(cursor_x_hot):
+	case _vgtif_reg(cursor_y_hot):
+	case _vgtif_reg(pdp[0].lo):
+	case _vgtif_reg(pdp[0].hi):
+	case _vgtif_reg(pdp[1].lo):
+	case _vgtif_reg(pdp[1].hi):
+	case _vgtif_reg(pdp[2].lo):
+	case _vgtif_reg(pdp[2].hi):
+	case _vgtif_reg(pdp[3].lo):
+	case _vgtif_reg(pdp[3].hi):
+	case _vgtif_reg(execlist_context_descriptor_lo):
+	case _vgtif_reg(execlist_context_descriptor_hi):
+		break;
+	case _vgtif_reg(rsv5[0])..._vgtif_reg(rsv5[3]):
+		invalid_write = true;
+		enter_failsafe_mode(vgpu, GVT_FAILSAFE_INSUFFICIENT_RESOURCE);
+		break;
+	default:
+		invalid_write = true;
+		gvt_vgpu_err("invalid pvinfo write offset %x bytes %x data %x\n",
+				offset, bytes, data);
+		break;
+	}
+
+	if (!invalid_write)
+		write_vreg(vgpu, offset, p_data, bytes);
+
+	return 0;
+}
+
+static int pf_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	u32 val = *(u32 *)p_data;
+
+	if ((offset == _PS_1A_CTRL || offset == _PS_2A_CTRL ||
+	   offset == _PS_1B_CTRL || offset == _PS_2B_CTRL ||
+	   offset == _PS_1C_CTRL) && (val & PS_PLANE_SEL_MASK) != 0) {
+		drm_WARN_ONCE(&i915->drm, true,
+			      "VM(%d): guest is trying to scaling a plane\n",
+			      vgpu->id);
+		return 0;
+	}
+
+	return intel_vgpu_default_mmio_write(vgpu, offset, p_data, bytes);
+}
+
+static int power_well_ctl_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg(vgpu, offset) &
+	    HSW_PWR_WELL_CTL_REQ(HSW_PW_CTL_IDX_GLOBAL))
+		vgpu_vreg(vgpu, offset) |=
+			HSW_PWR_WELL_CTL_STATE(HSW_PW_CTL_IDX_GLOBAL);
+	else
+		vgpu_vreg(vgpu, offset) &=
+			~HSW_PWR_WELL_CTL_STATE(HSW_PW_CTL_IDX_GLOBAL);
+	return 0;
+}
+
+static int gen9_dbuf_ctl_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg(vgpu, offset) & DBUF_POWER_REQUEST)
+		vgpu_vreg(vgpu, offset) |= DBUF_POWER_STATE;
+	else
+		vgpu_vreg(vgpu, offset) &= ~DBUF_POWER_STATE;
+
+	return 0;
+}
+
+static int fpga_dbg_mmio_write(struct intel_vgpu *vgpu,
+	unsigned int offset, void *p_data, unsigned int bytes)
+{
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg(vgpu, offset) & FPGA_DBG_RM_NOCLAIM)
+		vgpu_vreg(vgpu, offset) &= ~FPGA_DBG_RM_NOCLAIM;
+	return 0;
+}
+
+static int dma_ctrl_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	u32 mode;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	mode = vgpu_vreg(vgpu, offset);
+
+	if (GFX_MODE_BIT_SET_IN_MASK(mode, START_DMA)) {
+		drm_WARN_ONCE(&i915->drm, 1,
+				"VM(%d): iGVT-g doesn't support GuC\n",
+				vgpu->id);
+		return 0;
+	}
+
+	return 0;
+}
+
+static int gen9_trtte_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	u32 trtte = *(u32 *)p_data;
+
+	if ((trtte & 1) && (trtte & (1 << 1)) == 0) {
+		drm_WARN(&i915->drm, 1,
+				"VM(%d): Use physical address for TRTT!\n",
+				vgpu->id);
+		return -EINVAL;
+	}
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	return 0;
+}
+
+static int gen9_trtt_chicken_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	write_vreg(vgpu, offset, p_data, bytes);
+	return 0;
+}
+
+static int dpll_status_read(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 v = 0;
+
+	if (vgpu_vreg(vgpu, 0x46010) & (1 << 31))
+		v |= (1 << 0);
+
+	if (vgpu_vreg(vgpu, 0x46014) & (1 << 31))
+		v |= (1 << 8);
+
+	if (vgpu_vreg(vgpu, 0x46040) & (1 << 31))
+		v |= (1 << 16);
+
+	if (vgpu_vreg(vgpu, 0x46060) & (1 << 31))
+		v |= (1 << 24);
+
+	vgpu_vreg(vgpu, offset) = v;
+
+	return intel_vgpu_default_mmio_read(vgpu, offset, p_data, bytes);
+}
+
+static int mailbox_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 value = *(u32 *)p_data;
+	u32 cmd = value & 0xff;
+	u32 *data0 = &vgpu_vreg_t(vgpu, GEN6_PCODE_DATA);
+
+	switch (cmd) {
+	case GEN9_PCODE_READ_MEM_LATENCY:
+		if (IS_SKYLAKE(vgpu->gvt->gt->i915) ||
+		    IS_KABYLAKE(vgpu->gvt->gt->i915) ||
+		    IS_COFFEELAKE(vgpu->gvt->gt->i915) ||
+		    IS_COMETLAKE(vgpu->gvt->gt->i915)) {
+			/**
+			 * "Read memory latency" command on gen9.
+			 * Below memory latency values are read
+			 * from skylake platform.
+			 */
+			if (!*data0)
+				*data0 = 0x1e1a1100;
+			else
+				*data0 = 0x61514b3d;
+		} else if (IS_BROXTON(vgpu->gvt->gt->i915)) {
+			/**
+			 * "Read memory latency" command on gen9.
+			 * Below memory latency values are read
+			 * from Broxton MRB.
+			 */
+			if (!*data0)
+				*data0 = 0x16080707;
+			else
+				*data0 = 0x16161616;
+		}
+		break;
+	case SKL_PCODE_CDCLK_CONTROL:
+		if (IS_SKYLAKE(vgpu->gvt->gt->i915) ||
+		    IS_KABYLAKE(vgpu->gvt->gt->i915) ||
+		    IS_COFFEELAKE(vgpu->gvt->gt->i915) ||
+		    IS_COMETLAKE(vgpu->gvt->gt->i915))
+			*data0 = SKL_CDCLK_READY_FOR_CHANGE;
+		break;
+	case GEN6_PCODE_READ_RC6VIDS:
+		*data0 |= 0x1;
+		break;
+	}
+
+	gvt_dbg_core("VM(%d) write %x to mailbox, return data0 %x\n",
+		     vgpu->id, value, *data0);
+	/**
+	 * PCODE_READY clear means ready for pcode read/write,
+	 * PCODE_ERROR_MASK clear means no error happened. In GVT-g we
+	 * always emulate as pcode read/write success and ready for access
+	 * anytime, since we don't touch real physical registers here.
+	 */
+	value &= ~(GEN6_PCODE_READY | GEN6_PCODE_ERROR_MASK);
+	return intel_vgpu_default_mmio_write(vgpu, offset, &value, bytes);
+}
+
+static int hws_pga_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 value = *(u32 *)p_data;
+	const struct intel_engine_cs *engine =
+		intel_gvt_render_mmio_to_engine(vgpu->gvt, offset);
+
+	if (value != 0 &&
+	    !intel_gvt_ggtt_validate_range(vgpu, value, I915_GTT_PAGE_SIZE)) {
+		gvt_vgpu_err("write invalid HWSP address, reg:0x%x, value:0x%x\n",
+			      offset, value);
+		return -EINVAL;
+	}
+
+	/*
+	 * Need to emulate all the HWSP register write to ensure host can
+	 * update the VM CSB status correctly. Here listed registers can
+	 * support BDW, SKL or other platforms with same HWSP registers.
+	 */
+	if (unlikely(!engine)) {
+		gvt_vgpu_err("access unknown hardware status page register:0x%x\n",
+			     offset);
+		return -EINVAL;
+	}
+	vgpu->hws_pga[engine->id] = value;
+	gvt_dbg_mmio("VM(%d) write: 0x%x to HWSP: 0x%x\n",
+		     vgpu->id, value, offset);
+
+	return intel_vgpu_default_mmio_write(vgpu, offset, &value, bytes);
+}
+
+static int skl_power_well_ctl_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 v = *(u32 *)p_data;
+
+	if (IS_BROXTON(vgpu->gvt->gt->i915))
+		v &= (1 << 31) | (1 << 29);
+	else
+		v &= (1 << 31) | (1 << 29) | (1 << 9) |
+			(1 << 7) | (1 << 5) | (1 << 3) | (1 << 1);
+	v |= (v >> 1);
+
+	return intel_vgpu_default_mmio_write(vgpu, offset, &v, bytes);
+}
+
+static int skl_lcpll_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 v = *(u32 *)p_data;
+
+	/* other bits are MBZ. */
+	v &= (1 << 31) | (1 << 30);
+	v & (1 << 31) ? (v |= (1 << 30)) : (v &= ~(1 << 30));
+
+	vgpu_vreg(vgpu, offset) = v;
+
+	return 0;
+}
+
+static int bxt_de_pll_enable_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 v = *(u32 *)p_data;
+
+	if (v & BXT_DE_PLL_PLL_ENABLE)
+		v |= BXT_DE_PLL_LOCK;
+
+	vgpu_vreg(vgpu, offset) = v;
+
+	return 0;
+}
+
+static int bxt_port_pll_enable_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 v = *(u32 *)p_data;
+
+	if (v & PORT_PLL_ENABLE)
+		v |= PORT_PLL_LOCK;
+
+	vgpu_vreg(vgpu, offset) = v;
+
+	return 0;
+}
+
+static int bxt_phy_ctl_family_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 v = *(u32 *)p_data;
+	u32 data = v & COMMON_RESET_DIS ? BXT_PHY_LANE_ENABLED : 0;
+
+	switch (offset) {
+	case _PHY_CTL_FAMILY_EDP:
+		vgpu_vreg(vgpu, _BXT_PHY_CTL_DDI_A) = data;
+		break;
+	case _PHY_CTL_FAMILY_DDI:
+		vgpu_vreg(vgpu, _BXT_PHY_CTL_DDI_B) = data;
+		vgpu_vreg(vgpu, _BXT_PHY_CTL_DDI_C) = data;
+		break;
+	}
+
+	vgpu_vreg(vgpu, offset) = v;
+
+	return 0;
+}
+
+static int bxt_port_tx_dw3_read(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 v = vgpu_vreg(vgpu, offset);
+
+	v &= ~UNIQUE_TRANGE_EN_METHOD;
+
+	vgpu_vreg(vgpu, offset) = v;
+
+	return intel_vgpu_default_mmio_read(vgpu, offset, p_data, bytes);
+}
+
+static int bxt_pcs_dw12_grp_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 v = *(u32 *)p_data;
+
+	if (offset == _PORT_PCS_DW12_GRP_A || offset == _PORT_PCS_DW12_GRP_B) {
+		vgpu_vreg(vgpu, offset - 0x600) = v;
+		vgpu_vreg(vgpu, offset - 0x800) = v;
+	} else {
+		vgpu_vreg(vgpu, offset - 0x400) = v;
+		vgpu_vreg(vgpu, offset - 0x600) = v;
+	}
+
+	vgpu_vreg(vgpu, offset) = v;
+
+	return 0;
+}
+
+static int bxt_gt_disp_pwron_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 v = *(u32 *)p_data;
+
+	if (v & BIT(0)) {
+		vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY0)) &=
+			~PHY_RESERVED;
+		vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY0)) |=
+			PHY_POWER_GOOD;
+	}
+
+	if (v & BIT(1)) {
+		vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY1)) &=
+			~PHY_RESERVED;
+		vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY1)) |=
+			PHY_POWER_GOOD;
+	}
+
+
+	vgpu_vreg(vgpu, offset) = v;
+
+	return 0;
+}
+
+static int edp_psr_imr_iir_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	vgpu_vreg(vgpu, offset) = 0;
+	return 0;
+}
+
+/*
+ * FixMe:
+ * If guest fills non-priv batch buffer on ApolloLake/Broxton as Mesa i965 did:
+ * 717e7539124d (i965: Use a WC map and memcpy for the batch instead of pwrite.)
+ * Due to the missing flush of bb filled by VM vCPU, host GPU hangs on executing
+ * these MI_BATCH_BUFFER.
+ * Temporarily workaround this by setting SNOOP bit for PAT3 used by PPGTT
+ * PML4 PTE: PAT(0) PCD(1) PWT(1).
+ * The performance is still expected to be low, will need further improvement.
+ */
+static int bxt_ppat_low_write(struct intel_vgpu *vgpu, unsigned int offset,
+			      void *p_data, unsigned int bytes)
+{
+	u64 pat =
+		GEN8_PPAT(0, CHV_PPAT_SNOOP) |
+		GEN8_PPAT(1, 0) |
+		GEN8_PPAT(2, 0) |
+		GEN8_PPAT(3, CHV_PPAT_SNOOP) |
+		GEN8_PPAT(4, CHV_PPAT_SNOOP) |
+		GEN8_PPAT(5, CHV_PPAT_SNOOP) |
+		GEN8_PPAT(6, CHV_PPAT_SNOOP) |
+		GEN8_PPAT(7, CHV_PPAT_SNOOP);
+
+	vgpu_vreg(vgpu, offset) = lower_32_bits(pat);
+
+	return 0;
+}
+
+static int guc_status_read(struct intel_vgpu *vgpu,
+			   unsigned int offset, void *p_data,
+			   unsigned int bytes)
+{
+	/* keep MIA_IN_RESET before clearing */
+	read_vreg(vgpu, offset, p_data, bytes);
+	vgpu_vreg(vgpu, offset) &= ~GS_MIA_IN_RESET;
+	return 0;
+}
+
+static int mmio_read_from_hw(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	const struct intel_engine_cs *engine =
+		intel_gvt_render_mmio_to_engine(gvt, offset);
+
+	/**
+	 * Read HW reg in following case
+	 * a. the offset isn't a ring mmio
+	 * b. the offset's ring is running on hw.
+	 * c. the offset is ring time stamp mmio
+	 */
+
+	if (!engine ||
+	    vgpu == gvt->scheduler.engine_owner[engine->id] ||
+	    offset == i915_mmio_reg_offset(RING_TIMESTAMP(engine->mmio_base)) ||
+	    offset == i915_mmio_reg_offset(RING_TIMESTAMP_UDW(engine->mmio_base))) {
+		mmio_hw_access_pre(gvt->gt);
+		vgpu_vreg(vgpu, offset) =
+			intel_uncore_read(gvt->gt->uncore, _MMIO(offset));
+		mmio_hw_access_post(gvt->gt);
+	}
+
+	return intel_vgpu_default_mmio_read(vgpu, offset, p_data, bytes);
+}
+
+static int elsp_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	const struct intel_engine_cs *engine = intel_gvt_render_mmio_to_engine(vgpu->gvt, offset);
+	struct intel_vgpu_execlist *execlist;
+	u32 data = *(u32 *)p_data;
+	int ret = 0;
+
+	if (drm_WARN_ON(&i915->drm, !engine))
+		return -EINVAL;
+
+	/*
+	 * Due to d3_entered is used to indicate skipping PPGTT invalidation on
+	 * vGPU reset, it's set on D0->D3 on PCI config write, and cleared after
+	 * vGPU reset if in resuming.
+	 * In S0ix exit, the device power state also transite from D3 to D0 as
+	 * S3 resume, but no vGPU reset (triggered by QEMU devic model). After
+	 * S0ix exit, all engines continue to work. However the d3_entered
+	 * remains set which will break next vGPU reset logic (miss the expected
+	 * PPGTT invalidation).
+	 * Engines can only work in D0. Thus the 1st elsp write gives GVT a
+	 * chance to clear d3_entered.
+	 */
+	if (vgpu->d3_entered)
+		vgpu->d3_entered = false;
+
+	execlist = &vgpu->submission.execlist[engine->id];
+
+	execlist->elsp_dwords.data[3 - execlist->elsp_dwords.index] = data;
+	if (execlist->elsp_dwords.index == 3) {
+		ret = intel_vgpu_submit_execlist(vgpu, engine);
+		if(ret)
+			gvt_vgpu_err("fail submit workload on ring %s\n",
+				     engine->name);
+	}
+
+	++execlist->elsp_dwords.index;
+	execlist->elsp_dwords.index &= 0x3;
+	return ret;
+}
+
+static int ring_mode_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 data = *(u32 *)p_data;
+	const struct intel_engine_cs *engine =
+		intel_gvt_render_mmio_to_engine(vgpu->gvt, offset);
+	bool enable_execlist;
+	int ret;
+
+	(*(u32 *)p_data) &= ~_MASKED_BIT_ENABLE(1);
+	if (IS_COFFEELAKE(vgpu->gvt->gt->i915) ||
+	    IS_COMETLAKE(vgpu->gvt->gt->i915))
+		(*(u32 *)p_data) &= ~_MASKED_BIT_ENABLE(2);
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (IS_MASKED_BITS_ENABLED(data, 1)) {
+		enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST);
+		return 0;
+	}
+
+	if ((IS_COFFEELAKE(vgpu->gvt->gt->i915) ||
+	     IS_COMETLAKE(vgpu->gvt->gt->i915)) &&
+	    IS_MASKED_BITS_ENABLED(data, 2)) {
+		enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST);
+		return 0;
+	}
+
+	/* when PPGTT mode enabled, we will check if guest has called
+	 * pvinfo, if not, we will treat this guest as non-gvtg-aware
+	 * guest, and stop emulating its cfg space, mmio, gtt, etc.
+	 */
+	if ((IS_MASKED_BITS_ENABLED(data, GFX_PPGTT_ENABLE) ||
+	    IS_MASKED_BITS_ENABLED(data, GFX_RUN_LIST_ENABLE)) &&
+	    !vgpu->pv_notified) {
+		enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST);
+		return 0;
+	}
+	if (IS_MASKED_BITS_ENABLED(data, GFX_RUN_LIST_ENABLE) ||
+	    IS_MASKED_BITS_DISABLED(data, GFX_RUN_LIST_ENABLE)) {
+		enable_execlist = !!(data & GFX_RUN_LIST_ENABLE);
+
+		gvt_dbg_core("EXECLIST %s on ring %s\n",
+			     (enable_execlist ? "enabling" : "disabling"),
+			     engine->name);
+
+		if (!enable_execlist)
+			return 0;
+
+		ret = intel_vgpu_select_submission_ops(vgpu,
+						       engine->mask,
+						       INTEL_VGPU_EXECLIST_SUBMISSION);
+		if (ret)
+			return ret;
+
+		intel_vgpu_start_schedule(vgpu);
+	}
+	return 0;
+}
+
+static int gvt_reg_tlb_control_handler(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	unsigned int id = 0;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	vgpu_vreg(vgpu, offset) = 0;
+
+	switch (offset) {
+	case 0x4260:
+		id = RCS0;
+		break;
+	case 0x4264:
+		id = VCS0;
+		break;
+	case 0x4268:
+		id = VCS1;
+		break;
+	case 0x426c:
+		id = BCS0;
+		break;
+	case 0x4270:
+		id = VECS0;
+		break;
+	default:
+		return -EINVAL;
+	}
+	set_bit(id, (void *)vgpu->submission.tlb_handle_pending);
+
+	return 0;
+}
+
+static int ring_reset_ctl_write(struct intel_vgpu *vgpu,
+	unsigned int offset, void *p_data, unsigned int bytes)
+{
+	u32 data;
+
+	write_vreg(vgpu, offset, p_data, bytes);
+	data = vgpu_vreg(vgpu, offset);
+
+	if (IS_MASKED_BITS_ENABLED(data, RESET_CTL_REQUEST_RESET))
+		data |= RESET_CTL_READY_TO_RESET;
+	else if (data & _MASKED_BIT_DISABLE(RESET_CTL_REQUEST_RESET))
+		data &= ~RESET_CTL_READY_TO_RESET;
+
+	vgpu_vreg(vgpu, offset) = data;
+	return 0;
+}
+
+static int csfe_chicken1_mmio_write(struct intel_vgpu *vgpu,
+				    unsigned int offset, void *p_data,
+				    unsigned int bytes)
+{
+	u32 data = *(u32 *)p_data;
+
+	(*(u32 *)p_data) &= ~_MASKED_BIT_ENABLE(0x18);
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (IS_MASKED_BITS_ENABLED(data, 0x10) ||
+	    IS_MASKED_BITS_ENABLED(data, 0x8))
+		enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST);
+
+	return 0;
+}
+
+#define MMIO_F(reg, s, f, am, rm, d, r, w) do { \
+	ret = setup_mmio_info(gvt, i915_mmio_reg_offset(reg), \
+		s, f, am, rm, d, r, w); \
+	if (ret) \
+		return ret; \
+} while (0)
+
+#define MMIO_DH(reg, d, r, w) \
+	MMIO_F(reg, 4, 0, 0, 0, d, r, w)
+
+#define MMIO_DFH(reg, d, f, r, w) \
+	MMIO_F(reg, 4, f, 0, 0, d, r, w)
+
+#define MMIO_GM(reg, d, r, w) \
+	MMIO_F(reg, 4, F_GMADR, 0xFFFFF000, 0, d, r, w)
+
+#define MMIO_GM_RDR(reg, d, r, w) \
+	MMIO_F(reg, 4, F_GMADR | F_CMD_ACCESS, 0xFFFFF000, 0, d, r, w)
+
+#define MMIO_RO(reg, d, f, rm, r, w) \
+	MMIO_F(reg, 4, F_RO | f, 0, rm, d, r, w)
+
+#define MMIO_RING_F(prefix, s, f, am, rm, d, r, w) do { \
+	MMIO_F(prefix(RENDER_RING_BASE), s, f, am, rm, d, r, w); \
+	MMIO_F(prefix(BLT_RING_BASE), s, f, am, rm, d, r, w); \
+	MMIO_F(prefix(GEN6_BSD_RING_BASE), s, f, am, rm, d, r, w); \
+	MMIO_F(prefix(VEBOX_RING_BASE), s, f, am, rm, d, r, w); \
+	if (HAS_ENGINE(gvt->gt, VCS1)) \
+		MMIO_F(prefix(GEN8_BSD2_RING_BASE), s, f, am, rm, d, r, w); \
+} while (0)
+
+#define MMIO_RING_DFH(prefix, d, f, r, w) \
+	MMIO_RING_F(prefix, 4, f, 0, 0, d, r, w)
+
+#define MMIO_RING_GM(prefix, d, r, w) \
+	MMIO_RING_F(prefix, 4, F_GMADR, 0xFFFF0000, 0, d, r, w)
+
+#define MMIO_RING_GM_RDR(prefix, d, r, w) \
+	MMIO_RING_F(prefix, 4, F_GMADR | F_CMD_ACCESS, 0xFFFF0000, 0, d, r, w)
+
+#define MMIO_RING_RO(prefix, d, f, rm, r, w) \
+	MMIO_RING_F(prefix, 4, F_RO | f, 0, rm, d, r, w)
+
+static int init_generic_mmio_info(struct intel_gvt *gvt)
+{
+	struct drm_i915_private *dev_priv = gvt->gt->i915;
+	int ret;
+
+	MMIO_RING_DFH(RING_IMR, D_ALL, 0, NULL,
+		intel_vgpu_reg_imr_handler);
+
+	MMIO_DFH(SDEIMR, D_ALL, 0, NULL, intel_vgpu_reg_imr_handler);
+	MMIO_DFH(SDEIER, D_ALL, 0, NULL, intel_vgpu_reg_ier_handler);
+	MMIO_DFH(SDEIIR, D_ALL, 0, NULL, intel_vgpu_reg_iir_handler);
+
+	MMIO_RING_DFH(RING_HWSTAM, D_ALL, 0, NULL, NULL);
+
+
+	MMIO_DH(GEN8_GAMW_ECO_DEV_RW_IA, D_BDW_PLUS, NULL,
+		gamw_echo_dev_rw_ia_write);
+
+	MMIO_GM_RDR(BSD_HWS_PGA_GEN7, D_ALL, NULL, NULL);
+	MMIO_GM_RDR(BLT_HWS_PGA_GEN7, D_ALL, NULL, NULL);
+	MMIO_GM_RDR(VEBOX_HWS_PGA_GEN7, D_ALL, NULL, NULL);
+
+#define RING_REG(base) _MMIO((base) + 0x28)
+	MMIO_RING_DFH(RING_REG, D_ALL, F_CMD_ACCESS, NULL, NULL);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x134)
+	MMIO_RING_DFH(RING_REG, D_ALL, F_CMD_ACCESS, NULL, NULL);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x6c)
+	MMIO_RING_DFH(RING_REG, D_ALL, 0, mmio_read_from_hw, NULL);
+#undef RING_REG
+	MMIO_DH(GEN7_SC_INSTDONE, D_BDW_PLUS, mmio_read_from_hw, NULL);
+
+	MMIO_GM_RDR(_MMIO(0x2148), D_ALL, NULL, NULL);
+	MMIO_GM_RDR(CCID(RENDER_RING_BASE), D_ALL, NULL, NULL);
+	MMIO_GM_RDR(_MMIO(0x12198), D_ALL, NULL, NULL);
+
+	MMIO_RING_DFH(RING_TAIL, D_ALL, 0, NULL, NULL);
+	MMIO_RING_DFH(RING_HEAD, D_ALL, 0, NULL, NULL);
+	MMIO_RING_DFH(RING_CTL, D_ALL, 0, NULL, NULL);
+	MMIO_RING_DFH(RING_ACTHD, D_ALL, 0, mmio_read_from_hw, NULL);
+	MMIO_RING_GM(RING_START, D_ALL, NULL, NULL);
+
+	/* RING MODE */
+#define RING_REG(base) _MMIO((base) + 0x29c)
+	MMIO_RING_DFH(RING_REG, D_ALL,
+		F_MODE_MASK | F_CMD_ACCESS | F_CMD_WRITE_PATCH, NULL,
+		ring_mode_mmio_write);
+#undef RING_REG
+
+	MMIO_RING_DFH(RING_MI_MODE, D_ALL, F_MODE_MASK | F_CMD_ACCESS,
+		NULL, NULL);
+	MMIO_RING_DFH(RING_INSTPM, D_ALL, F_MODE_MASK | F_CMD_ACCESS,
+			NULL, NULL);
+	MMIO_RING_DFH(RING_TIMESTAMP, D_ALL, F_CMD_ACCESS,
+			mmio_read_from_hw, NULL);
+	MMIO_RING_DFH(RING_TIMESTAMP_UDW, D_ALL, F_CMD_ACCESS,
+			mmio_read_from_hw, NULL);
+
+	MMIO_DFH(GEN7_GT_MODE, D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(CACHE_MODE_0_GEN7, D_ALL, F_MODE_MASK | F_CMD_ACCESS,
+		NULL, NULL);
+	MMIO_DFH(CACHE_MODE_1, D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(CACHE_MODE_0, D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2124), D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DFH(_MMIO(0x20dc), D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_3D_CHICKEN3, D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2088), D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(FF_SLICE_CS_CHICKEN2, D_ALL,
+		 F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2470), D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(GAM_ECOCHK, D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(GEN7_COMMON_SLICE_CHICKEN1, D_ALL, F_MODE_MASK | F_CMD_ACCESS,
+		NULL, NULL);
+	MMIO_DFH(COMMON_SLICE_CHICKEN2, D_ALL, F_MODE_MASK | F_CMD_ACCESS,
+		 NULL, NULL);
+	MMIO_DFH(_MMIO(0x9030), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x20a0), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2420), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2430), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2434), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2438), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x243c), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x7018), D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(HALF_SLICE_CHICKEN3, D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(GEN7_HALF_SLICE_CHICKEN1, D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+
+	/* display */
+	MMIO_DH(PIPECONF(PIPE_A), D_ALL, NULL, pipeconf_mmio_write);
+	MMIO_DH(PIPECONF(PIPE_B), D_ALL, NULL, pipeconf_mmio_write);
+	MMIO_DH(PIPECONF(PIPE_C), D_ALL, NULL, pipeconf_mmio_write);
+	MMIO_DH(PIPECONF(_PIPE_EDP), D_ALL, NULL, pipeconf_mmio_write);
+	MMIO_DH(DSPSURF(PIPE_A), D_ALL, NULL, pri_surf_mmio_write);
+	MMIO_DH(REG_50080(PIPE_A, PLANE_PRIMARY), D_ALL, NULL,
+		reg50080_mmio_write);
+	MMIO_DH(DSPSURF(PIPE_B), D_ALL, NULL, pri_surf_mmio_write);
+	MMIO_DH(REG_50080(PIPE_B, PLANE_PRIMARY), D_ALL, NULL,
+		reg50080_mmio_write);
+	MMIO_DH(DSPSURF(PIPE_C), D_ALL, NULL, pri_surf_mmio_write);
+	MMIO_DH(REG_50080(PIPE_C, PLANE_PRIMARY), D_ALL, NULL,
+		reg50080_mmio_write);
+	MMIO_DH(SPRSURF(PIPE_A), D_ALL, NULL, spr_surf_mmio_write);
+	MMIO_DH(REG_50080(PIPE_A, PLANE_SPRITE0), D_ALL, NULL,
+		reg50080_mmio_write);
+	MMIO_DH(SPRSURF(PIPE_B), D_ALL, NULL, spr_surf_mmio_write);
+	MMIO_DH(REG_50080(PIPE_B, PLANE_SPRITE0), D_ALL, NULL,
+		reg50080_mmio_write);
+	MMIO_DH(SPRSURF(PIPE_C), D_ALL, NULL, spr_surf_mmio_write);
+	MMIO_DH(REG_50080(PIPE_C, PLANE_SPRITE0), D_ALL, NULL,
+		reg50080_mmio_write);
+
+	MMIO_F(PCH_GMBUS0, 4 * 4, 0, 0, 0, D_ALL, gmbus_mmio_read,
+		gmbus_mmio_write);
+	MMIO_F(PCH_GPIO_BASE, 6 * 4, F_UNALIGN, 0, 0, D_ALL, NULL, NULL);
+
+	MMIO_F(_MMIO(_PCH_DPB_AUX_CH_CTL), 6 * 4, 0, 0, 0, D_PRE_SKL, NULL,
+		dp_aux_ch_ctl_mmio_write);
+	MMIO_F(_MMIO(_PCH_DPC_AUX_CH_CTL), 6 * 4, 0, 0, 0, D_PRE_SKL, NULL,
+		dp_aux_ch_ctl_mmio_write);
+	MMIO_F(_MMIO(_PCH_DPD_AUX_CH_CTL), 6 * 4, 0, 0, 0, D_PRE_SKL, NULL,
+		dp_aux_ch_ctl_mmio_write);
+
+	MMIO_DH(PCH_ADPA, D_PRE_SKL, NULL, pch_adpa_mmio_write);
+
+	MMIO_DH(_MMIO(_PCH_TRANSACONF), D_ALL, NULL, transconf_mmio_write);
+	MMIO_DH(_MMIO(_PCH_TRANSBCONF), D_ALL, NULL, transconf_mmio_write);
+
+	MMIO_DH(FDI_RX_IIR(PIPE_A), D_ALL, NULL, fdi_rx_iir_mmio_write);
+	MMIO_DH(FDI_RX_IIR(PIPE_B), D_ALL, NULL, fdi_rx_iir_mmio_write);
+	MMIO_DH(FDI_RX_IIR(PIPE_C), D_ALL, NULL, fdi_rx_iir_mmio_write);
+	MMIO_DH(FDI_RX_IMR(PIPE_A), D_ALL, NULL, update_fdi_rx_iir_status);
+	MMIO_DH(FDI_RX_IMR(PIPE_B), D_ALL, NULL, update_fdi_rx_iir_status);
+	MMIO_DH(FDI_RX_IMR(PIPE_C), D_ALL, NULL, update_fdi_rx_iir_status);
+	MMIO_DH(FDI_RX_CTL(PIPE_A), D_ALL, NULL, update_fdi_rx_iir_status);
+	MMIO_DH(FDI_RX_CTL(PIPE_B), D_ALL, NULL, update_fdi_rx_iir_status);
+	MMIO_DH(FDI_RX_CTL(PIPE_C), D_ALL, NULL, update_fdi_rx_iir_status);
+	MMIO_DH(PCH_PP_CONTROL, D_ALL, NULL, pch_pp_control_mmio_write);
+	MMIO_DH(_MMIO(0xe651c), D_ALL, dpy_reg_mmio_read, NULL);
+	MMIO_DH(_MMIO(0xe661c), D_ALL, dpy_reg_mmio_read, NULL);
+	MMIO_DH(_MMIO(0xe671c), D_ALL, dpy_reg_mmio_read, NULL);
+	MMIO_DH(_MMIO(0xe681c), D_ALL, dpy_reg_mmio_read, NULL);
+	MMIO_DH(_MMIO(0xe6c04), D_ALL, dpy_reg_mmio_read, NULL);
+	MMIO_DH(_MMIO(0xe6e1c), D_ALL, dpy_reg_mmio_read, NULL);
+
+	MMIO_RO(PCH_PORT_HOTPLUG, D_ALL, 0,
+		PORTA_HOTPLUG_STATUS_MASK
+		| PORTB_HOTPLUG_STATUS_MASK
+		| PORTC_HOTPLUG_STATUS_MASK
+		| PORTD_HOTPLUG_STATUS_MASK,
+		NULL, NULL);
+
+	MMIO_DH(LCPLL_CTL, D_ALL, NULL, lcpll_ctl_mmio_write);
+	MMIO_DH(SOUTH_CHICKEN2, D_ALL, NULL, south_chicken2_mmio_write);
+	MMIO_DH(SFUSE_STRAP, D_ALL, NULL, NULL);
+	MMIO_DH(SBI_DATA, D_ALL, sbi_data_mmio_read, NULL);
+	MMIO_DH(SBI_CTL_STAT, D_ALL, NULL, sbi_ctl_mmio_write);
+
+	MMIO_F(_MMIO(_DPA_AUX_CH_CTL), 6 * 4, 0, 0, 0, D_ALL, NULL,
+		dp_aux_ch_ctl_mmio_write);
+
+	MMIO_DH(DDI_BUF_CTL(PORT_A), D_ALL, NULL, ddi_buf_ctl_mmio_write);
+	MMIO_DH(DDI_BUF_CTL(PORT_B), D_ALL, NULL, ddi_buf_ctl_mmio_write);
+	MMIO_DH(DDI_BUF_CTL(PORT_C), D_ALL, NULL, ddi_buf_ctl_mmio_write);
+	MMIO_DH(DDI_BUF_CTL(PORT_D), D_ALL, NULL, ddi_buf_ctl_mmio_write);
+	MMIO_DH(DDI_BUF_CTL(PORT_E), D_ALL, NULL, ddi_buf_ctl_mmio_write);
+
+	MMIO_DH(DP_TP_CTL(PORT_A), D_ALL, NULL, dp_tp_ctl_mmio_write);
+	MMIO_DH(DP_TP_CTL(PORT_B), D_ALL, NULL, dp_tp_ctl_mmio_write);
+	MMIO_DH(DP_TP_CTL(PORT_C), D_ALL, NULL, dp_tp_ctl_mmio_write);
+	MMIO_DH(DP_TP_CTL(PORT_D), D_ALL, NULL, dp_tp_ctl_mmio_write);
+	MMIO_DH(DP_TP_CTL(PORT_E), D_ALL, NULL, dp_tp_ctl_mmio_write);
+
+	MMIO_DH(DP_TP_STATUS(PORT_A), D_ALL, NULL, dp_tp_status_mmio_write);
+	MMIO_DH(DP_TP_STATUS(PORT_B), D_ALL, NULL, dp_tp_status_mmio_write);
+	MMIO_DH(DP_TP_STATUS(PORT_C), D_ALL, NULL, dp_tp_status_mmio_write);
+	MMIO_DH(DP_TP_STATUS(PORT_D), D_ALL, NULL, dp_tp_status_mmio_write);
+	MMIO_DH(DP_TP_STATUS(PORT_E), D_ALL, NULL, NULL);
+
+	MMIO_DH(_MMIO(_TRANS_DDI_FUNC_CTL_A), D_ALL, NULL, NULL);
+	MMIO_DH(_MMIO(_TRANS_DDI_FUNC_CTL_B), D_ALL, NULL, NULL);
+	MMIO_DH(_MMIO(_TRANS_DDI_FUNC_CTL_C), D_ALL, NULL, NULL);
+	MMIO_DH(_MMIO(_TRANS_DDI_FUNC_CTL_EDP), D_ALL, NULL, NULL);
+
+	MMIO_DH(FORCEWAKE, D_ALL, NULL, NULL);
+	MMIO_DFH(GTFIFODBG, D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(GTFIFOCTL, D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DH(FORCEWAKE_MT, D_PRE_SKL, NULL, mul_force_wake_write);
+	MMIO_DH(FORCEWAKE_ACK_HSW, D_BDW, NULL, NULL);
+	MMIO_DH(GEN6_RC_CONTROL, D_ALL, NULL, NULL);
+	MMIO_DH(GEN6_RC_STATE, D_ALL, NULL, NULL);
+	MMIO_DH(HSW_PWR_WELL_CTL1, D_BDW, NULL, power_well_ctl_mmio_write);
+	MMIO_DH(HSW_PWR_WELL_CTL2, D_BDW, NULL, power_well_ctl_mmio_write);
+	MMIO_DH(HSW_PWR_WELL_CTL3, D_BDW, NULL, power_well_ctl_mmio_write);
+	MMIO_DH(HSW_PWR_WELL_CTL4, D_BDW, NULL, power_well_ctl_mmio_write);
+	MMIO_DH(HSW_PWR_WELL_CTL5, D_BDW, NULL, power_well_ctl_mmio_write);
+	MMIO_DH(HSW_PWR_WELL_CTL6, D_BDW, NULL, power_well_ctl_mmio_write);
+
+	MMIO_DH(GEN6_GDRST, D_ALL, NULL, gdrst_mmio_write);
+	MMIO_F(FENCE_REG_GEN6_LO(0), 0x80, 0, 0, 0, D_ALL, fence_mmio_read, fence_mmio_write);
+	MMIO_DH(CPU_VGACNTRL, D_ALL, NULL, vga_control_mmio_write);
+
+	MMIO_DH(GEN7_ERR_INT, D_ALL, NULL, NULL);
+	MMIO_DH(GFX_FLSH_CNTL_GEN6, D_ALL, NULL, NULL);
+
+	MMIO_DH(GEN6_MBCTL, D_ALL, NULL, mbctl_write);
+	MMIO_DFH(GEN7_UCGCTL4, D_ALL, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DH(FPGA_DBG, D_ALL, NULL, fpga_dbg_mmio_write);
+	MMIO_DFH(_MMIO(0x215c), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2178), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x217c), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x12178), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x1217c), D_ALL, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_F(_MMIO(0x2290), 8, F_CMD_ACCESS, 0, 0, D_BDW_PLUS, NULL, NULL);
+	MMIO_F(_MMIO(0x5200), 32, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(_MMIO(0x5240), 32, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(_MMIO(0x5280), 16, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+
+	MMIO_DFH(_MMIO(0x1c17c), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x1c178), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(BCS_SWCTRL, D_ALL, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_F(HS_INVOCATION_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(DS_INVOCATION_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(IA_VERTICES_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(IA_PRIMITIVES_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(VS_INVOCATION_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(GS_INVOCATION_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(GS_PRIMITIVES_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(CL_INVOCATION_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(CL_PRIMITIVES_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(PS_INVOCATION_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_F(PS_DEPTH_COUNT, 8, F_CMD_ACCESS, 0, 0, D_ALL, NULL, NULL);
+	MMIO_DH(_MMIO(0x4260), D_BDW_PLUS, NULL, gvt_reg_tlb_control_handler);
+	MMIO_DH(_MMIO(0x4264), D_BDW_PLUS, NULL, gvt_reg_tlb_control_handler);
+	MMIO_DH(_MMIO(0x4268), D_BDW_PLUS, NULL, gvt_reg_tlb_control_handler);
+	MMIO_DH(_MMIO(0x426c), D_BDW_PLUS, NULL, gvt_reg_tlb_control_handler);
+	MMIO_DH(_MMIO(0x4270), D_BDW_PLUS, NULL, gvt_reg_tlb_control_handler);
+	MMIO_DFH(_MMIO(0x4094), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DFH(ARB_MODE, D_ALL, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_RING_GM(RING_BBADDR, D_ALL, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2220), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x12220), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x22220), D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_RING_DFH(RING_SYNC_1, D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_RING_DFH(RING_SYNC_0, D_ALL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x22178), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x1a178), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x1a17c), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2217c), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DH(EDP_PSR_IMR, D_BDW_PLUS, NULL, edp_psr_imr_iir_write);
+	MMIO_DH(EDP_PSR_IIR, D_BDW_PLUS, NULL, edp_psr_imr_iir_write);
+	MMIO_DH(GUC_STATUS, D_ALL, guc_status_read, NULL);
+
+	return 0;
+}
+
+static int init_bdw_mmio_info(struct intel_gvt *gvt)
+{
+	int ret;
+
+	MMIO_DH(GEN8_GT_IMR(0), D_BDW_PLUS, NULL, intel_vgpu_reg_imr_handler);
+	MMIO_DH(GEN8_GT_IER(0), D_BDW_PLUS, NULL, intel_vgpu_reg_ier_handler);
+	MMIO_DH(GEN8_GT_IIR(0), D_BDW_PLUS, NULL, intel_vgpu_reg_iir_handler);
+
+	MMIO_DH(GEN8_GT_IMR(1), D_BDW_PLUS, NULL, intel_vgpu_reg_imr_handler);
+	MMIO_DH(GEN8_GT_IER(1), D_BDW_PLUS, NULL, intel_vgpu_reg_ier_handler);
+	MMIO_DH(GEN8_GT_IIR(1), D_BDW_PLUS, NULL, intel_vgpu_reg_iir_handler);
+
+	MMIO_DH(GEN8_GT_IMR(2), D_BDW_PLUS, NULL, intel_vgpu_reg_imr_handler);
+	MMIO_DH(GEN8_GT_IER(2), D_BDW_PLUS, NULL, intel_vgpu_reg_ier_handler);
+	MMIO_DH(GEN8_GT_IIR(2), D_BDW_PLUS, NULL, intel_vgpu_reg_iir_handler);
+
+	MMIO_DH(GEN8_GT_IMR(3), D_BDW_PLUS, NULL, intel_vgpu_reg_imr_handler);
+	MMIO_DH(GEN8_GT_IER(3), D_BDW_PLUS, NULL, intel_vgpu_reg_ier_handler);
+	MMIO_DH(GEN8_GT_IIR(3), D_BDW_PLUS, NULL, intel_vgpu_reg_iir_handler);
+
+	MMIO_DH(GEN8_DE_PIPE_IMR(PIPE_A), D_BDW_PLUS, NULL,
+		intel_vgpu_reg_imr_handler);
+	MMIO_DH(GEN8_DE_PIPE_IER(PIPE_A), D_BDW_PLUS, NULL,
+		intel_vgpu_reg_ier_handler);
+	MMIO_DH(GEN8_DE_PIPE_IIR(PIPE_A), D_BDW_PLUS, NULL,
+		intel_vgpu_reg_iir_handler);
+
+	MMIO_DH(GEN8_DE_PIPE_IMR(PIPE_B), D_BDW_PLUS, NULL,
+		intel_vgpu_reg_imr_handler);
+	MMIO_DH(GEN8_DE_PIPE_IER(PIPE_B), D_BDW_PLUS, NULL,
+		intel_vgpu_reg_ier_handler);
+	MMIO_DH(GEN8_DE_PIPE_IIR(PIPE_B), D_BDW_PLUS, NULL,
+		intel_vgpu_reg_iir_handler);
+
+	MMIO_DH(GEN8_DE_PIPE_IMR(PIPE_C), D_BDW_PLUS, NULL,
+		intel_vgpu_reg_imr_handler);
+	MMIO_DH(GEN8_DE_PIPE_IER(PIPE_C), D_BDW_PLUS, NULL,
+		intel_vgpu_reg_ier_handler);
+	MMIO_DH(GEN8_DE_PIPE_IIR(PIPE_C), D_BDW_PLUS, NULL,
+		intel_vgpu_reg_iir_handler);
+
+	MMIO_DH(GEN8_DE_PORT_IMR, D_BDW_PLUS, NULL, intel_vgpu_reg_imr_handler);
+	MMIO_DH(GEN8_DE_PORT_IER, D_BDW_PLUS, NULL, intel_vgpu_reg_ier_handler);
+	MMIO_DH(GEN8_DE_PORT_IIR, D_BDW_PLUS, NULL, intel_vgpu_reg_iir_handler);
+
+	MMIO_DH(GEN8_DE_MISC_IMR, D_BDW_PLUS, NULL, intel_vgpu_reg_imr_handler);
+	MMIO_DH(GEN8_DE_MISC_IER, D_BDW_PLUS, NULL, intel_vgpu_reg_ier_handler);
+	MMIO_DH(GEN8_DE_MISC_IIR, D_BDW_PLUS, NULL, intel_vgpu_reg_iir_handler);
+
+	MMIO_DH(GEN8_PCU_IMR, D_BDW_PLUS, NULL, intel_vgpu_reg_imr_handler);
+	MMIO_DH(GEN8_PCU_IER, D_BDW_PLUS, NULL, intel_vgpu_reg_ier_handler);
+	MMIO_DH(GEN8_PCU_IIR, D_BDW_PLUS, NULL, intel_vgpu_reg_iir_handler);
+
+	MMIO_DH(GEN8_MASTER_IRQ, D_BDW_PLUS, NULL,
+		intel_vgpu_reg_master_irq_handler);
+
+	MMIO_RING_DFH(RING_ACTHD_UDW, D_BDW_PLUS, 0,
+		mmio_read_from_hw, NULL);
+
+#define RING_REG(base) _MMIO((base) + 0xd0)
+	MMIO_RING_F(RING_REG, 4, F_RO, 0,
+		~_MASKED_BIT_ENABLE(RESET_CTL_REQUEST_RESET), D_BDW_PLUS, NULL,
+		ring_reset_ctl_write);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x230)
+	MMIO_RING_DFH(RING_REG, D_BDW_PLUS, 0, NULL, elsp_mmio_write);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x234)
+	MMIO_RING_F(RING_REG, 8, F_RO, 0, ~0, D_BDW_PLUS,
+		NULL, NULL);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x244)
+	MMIO_RING_DFH(RING_REG, D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x370)
+	MMIO_RING_F(RING_REG, 48, F_RO, 0, ~0, D_BDW_PLUS, NULL, NULL);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x3a0)
+	MMIO_RING_DFH(RING_REG, D_BDW_PLUS, F_MODE_MASK, NULL, NULL);
+#undef RING_REG
+
+	MMIO_DH(GEN6_PCODE_MAILBOX, D_BDW_PLUS, NULL, mailbox_write);
+
+#define RING_REG(base) _MMIO((base) + 0x270)
+	MMIO_RING_F(RING_REG, 32, F_CMD_ACCESS, 0, 0, D_BDW_PLUS, NULL, NULL);
+#undef RING_REG
+
+	MMIO_RING_GM(RING_HWS_PGA, D_BDW_PLUS, NULL, hws_pga_write);
+
+	MMIO_DFH(HDC_CHICKEN0, D_BDW_PLUS, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DFH(GEN8_ROW_CHICKEN, D_BDW_PLUS, F_MODE_MASK | F_CMD_ACCESS,
+		NULL, NULL);
+	MMIO_DFH(GEN7_ROW_CHICKEN2, D_BDW_PLUS, F_MODE_MASK | F_CMD_ACCESS,
+		NULL, NULL);
+	MMIO_DFH(GEN8_UCGCTL6, D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DFH(_MMIO(0xb1f0), D_BDW, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xb1c0), D_BDW, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(GEN8_L3SQCREG4, D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xb100), D_BDW, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xb10c), D_BDW, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_F(_MMIO(0x24d0), 48, F_CMD_ACCESS | F_CMD_WRITE_PATCH, 0, 0,
+		D_BDW_PLUS, NULL, force_nonpriv_write);
+
+	MMIO_DFH(_MMIO(0x83a4), D_BDW, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DFH(_MMIO(0x8430), D_BDW, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DFH(_MMIO(0xe194), D_BDW_PLUS, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xe188), D_BDW_PLUS, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(HALF_SLICE_CHICKEN2, D_BDW_PLUS, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x2580), D_BDW_PLUS, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DFH(_MMIO(0x2248), D_BDW, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DFH(_MMIO(0xe220), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xe230), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xe240), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xe260), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xe270), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xe280), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xe2a0), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xe2b0), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xe2c0), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x21f0), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+	return 0;
+}
+
+static int init_skl_mmio_info(struct intel_gvt *gvt)
+{
+	struct drm_i915_private *dev_priv = gvt->gt->i915;
+	int ret;
+
+	MMIO_DH(FORCEWAKE_RENDER_GEN9, D_SKL_PLUS, NULL, mul_force_wake_write);
+	MMIO_DH(FORCEWAKE_ACK_RENDER_GEN9, D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(FORCEWAKE_GT_GEN9, D_SKL_PLUS, NULL, mul_force_wake_write);
+	MMIO_DH(FORCEWAKE_ACK_GT_GEN9, D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(FORCEWAKE_MEDIA_GEN9, D_SKL_PLUS, NULL, mul_force_wake_write);
+	MMIO_DH(FORCEWAKE_ACK_MEDIA_GEN9, D_SKL_PLUS, NULL, NULL);
+
+	MMIO_F(DP_AUX_CH_CTL(AUX_CH_B), 6 * 4, 0, 0, 0, D_SKL_PLUS, NULL,
+						dp_aux_ch_ctl_mmio_write);
+	MMIO_F(DP_AUX_CH_CTL(AUX_CH_C), 6 * 4, 0, 0, 0, D_SKL_PLUS, NULL,
+						dp_aux_ch_ctl_mmio_write);
+	MMIO_F(DP_AUX_CH_CTL(AUX_CH_D), 6 * 4, 0, 0, 0, D_SKL_PLUS, NULL,
+						dp_aux_ch_ctl_mmio_write);
+
+	MMIO_DH(HSW_PWR_WELL_CTL2, D_SKL_PLUS, NULL, skl_power_well_ctl_write);
+
+	MMIO_DH(DBUF_CTL_S(0), D_SKL_PLUS, NULL, gen9_dbuf_ctl_mmio_write);
+
+	MMIO_DFH(GEN9_GAMT_ECO_REG_RW_IA, D_SKL_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(MMCD_MISC_CTRL, D_SKL_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DH(CHICKEN_PAR1_1, D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(LCPLL1_CTL, D_SKL_PLUS, NULL, skl_lcpll_write);
+	MMIO_DH(LCPLL2_CTL, D_SKL_PLUS, NULL, skl_lcpll_write);
+	MMIO_DH(DPLL_STATUS, D_SKL_PLUS, dpll_status_read, NULL);
+
+	MMIO_DH(SKL_PS_WIN_POS(PIPE_A, 0), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_WIN_POS(PIPE_A, 1), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_WIN_POS(PIPE_B, 0), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_WIN_POS(PIPE_B, 1), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_WIN_POS(PIPE_C, 0), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_WIN_POS(PIPE_C, 1), D_SKL_PLUS, NULL, pf_write);
+
+	MMIO_DH(SKL_PS_WIN_SZ(PIPE_A, 0), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_WIN_SZ(PIPE_A, 1), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_WIN_SZ(PIPE_B, 0), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_WIN_SZ(PIPE_B, 1), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_WIN_SZ(PIPE_C, 0), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_WIN_SZ(PIPE_C, 1), D_SKL_PLUS, NULL, pf_write);
+
+	MMIO_DH(SKL_PS_CTRL(PIPE_A, 0), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_CTRL(PIPE_A, 1), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_CTRL(PIPE_B, 0), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_CTRL(PIPE_B, 1), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_CTRL(PIPE_C, 0), D_SKL_PLUS, NULL, pf_write);
+	MMIO_DH(SKL_PS_CTRL(PIPE_C, 1), D_SKL_PLUS, NULL, pf_write);
+
+	MMIO_DH(PLANE_BUF_CFG(PIPE_A, 0), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_BUF_CFG(PIPE_A, 1), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_BUF_CFG(PIPE_A, 2), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_BUF_CFG(PIPE_A, 3), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(PLANE_BUF_CFG(PIPE_B, 0), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_BUF_CFG(PIPE_B, 1), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_BUF_CFG(PIPE_B, 2), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_BUF_CFG(PIPE_B, 3), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(PLANE_BUF_CFG(PIPE_C, 0), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_BUF_CFG(PIPE_C, 1), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_BUF_CFG(PIPE_C, 2), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_BUF_CFG(PIPE_C, 3), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(CUR_BUF_CFG(PIPE_A), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(CUR_BUF_CFG(PIPE_B), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(CUR_BUF_CFG(PIPE_C), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(PLANE_WM_TRANS(PIPE_A, 0), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_WM_TRANS(PIPE_A, 1), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_WM_TRANS(PIPE_A, 2), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(PLANE_WM_TRANS(PIPE_B, 0), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_WM_TRANS(PIPE_B, 1), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_WM_TRANS(PIPE_B, 2), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(PLANE_WM_TRANS(PIPE_C, 0), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_WM_TRANS(PIPE_C, 1), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_WM_TRANS(PIPE_C, 2), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(CUR_WM_TRANS(PIPE_A), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(CUR_WM_TRANS(PIPE_B), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(CUR_WM_TRANS(PIPE_C), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_A, 0), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_A, 1), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_A, 2), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_A, 3), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_B, 0), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_B, 1), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_B, 2), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_B, 3), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_C, 0), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_C, 1), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_C, 2), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(PLANE_NV12_BUF_CFG(PIPE_C, 3), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_A, 1)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_A, 2)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_A, 3)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_A, 4)), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_B, 1)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_B, 2)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_B, 3)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_B, 4)), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_C, 1)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_C, 2)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_C, 3)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C0(PIPE_C, 4)), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_A, 1)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_A, 2)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_A, 3)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_A, 4)), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_B, 1)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_B, 2)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_B, 3)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_B, 4)), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_C, 1)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_C, 2)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_C, 3)), D_SKL_PLUS, NULL, NULL);
+	MMIO_DH(_MMIO(_REG_701C4(PIPE_C, 4)), D_SKL_PLUS, NULL, NULL);
+
+	MMIO_DFH(BDW_SCRATCH1, D_SKL_PLUS, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_F(GEN9_GFX_MOCS(0), 0x7f8, F_CMD_ACCESS, 0, 0, D_SKL_PLUS,
+		NULL, NULL);
+	MMIO_F(GEN7_L3CNTLREG2, 0x80, F_CMD_ACCESS, 0, 0, D_SKL_PLUS,
+		NULL, NULL);
+
+	MMIO_DFH(GEN7_FF_SLICE_CS_CHICKEN1, D_SKL_PLUS,
+		 F_MODE_MASK | F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(GEN9_CS_DEBUG_MODE1, D_SKL_PLUS, F_MODE_MASK | F_CMD_ACCESS,
+		NULL, NULL);
+
+	/* TRTT */
+	MMIO_DFH(TRVATTL3PTRDW(0), D_SKL_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(TRVATTL3PTRDW(1), D_SKL_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(TRVATTL3PTRDW(2), D_SKL_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(TRVATTL3PTRDW(3), D_SKL_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(TRVADR, D_SKL_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(TRTTE, D_SKL_PLUS, F_CMD_ACCESS | F_PM_SAVE,
+		 NULL, gen9_trtte_write);
+	MMIO_DFH(_MMIO(0x4dfc), D_SKL_PLUS, F_PM_SAVE,
+		 NULL, gen9_trtt_chicken_write);
+
+	MMIO_DFH(GEN8_GARBCNTL, D_SKL_PLUS, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DH(DMA_CTRL, D_SKL_PLUS, NULL, dma_ctrl_write);
+
+#define CSFE_CHICKEN1_REG(base) _MMIO((base) + 0xD4)
+	MMIO_RING_DFH(CSFE_CHICKEN1_REG, D_SKL_PLUS, F_MODE_MASK | F_CMD_ACCESS,
+		      NULL, csfe_chicken1_mmio_write);
+#undef CSFE_CHICKEN1_REG
+	MMIO_DFH(GEN8_HDC_CHICKEN1, D_SKL_PLUS, F_MODE_MASK | F_CMD_ACCESS,
+		 NULL, NULL);
+	MMIO_DFH(GEN9_WM_CHICKEN3, D_SKL_PLUS, F_MODE_MASK | F_CMD_ACCESS,
+		 NULL, NULL);
+
+	MMIO_DFH(GAMT_CHKN_BIT_REG, D_KBL | D_CFL, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0xe4cc), D_BDW_PLUS, F_CMD_ACCESS, NULL, NULL);
+
+	return 0;
+}
+
+static int init_bxt_mmio_info(struct intel_gvt *gvt)
+{
+	int ret;
+
+	MMIO_DH(BXT_P_CR_GT_DISP_PWRON, D_BXT, NULL, bxt_gt_disp_pwron_write);
+	MMIO_DH(BXT_PHY_CTL_FAMILY(DPIO_PHY0), D_BXT,
+		NULL, bxt_phy_ctl_family_write);
+	MMIO_DH(BXT_PHY_CTL_FAMILY(DPIO_PHY1), D_BXT,
+		NULL, bxt_phy_ctl_family_write);
+	MMIO_DH(BXT_PORT_PLL_ENABLE(PORT_A), D_BXT,
+		NULL, bxt_port_pll_enable_write);
+	MMIO_DH(BXT_PORT_PLL_ENABLE(PORT_B), D_BXT,
+		NULL, bxt_port_pll_enable_write);
+	MMIO_DH(BXT_PORT_PLL_ENABLE(PORT_C), D_BXT, NULL,
+		bxt_port_pll_enable_write);
+
+	MMIO_DH(BXT_PORT_PCS_DW12_GRP(DPIO_PHY0, DPIO_CH0), D_BXT,
+		NULL, bxt_pcs_dw12_grp_write);
+	MMIO_DH(BXT_PORT_TX_DW3_LN0(DPIO_PHY0, DPIO_CH0), D_BXT,
+		bxt_port_tx_dw3_read, NULL);
+	MMIO_DH(BXT_PORT_PCS_DW12_GRP(DPIO_PHY0, DPIO_CH1), D_BXT,
+		NULL, bxt_pcs_dw12_grp_write);
+	MMIO_DH(BXT_PORT_TX_DW3_LN0(DPIO_PHY0, DPIO_CH1), D_BXT,
+		bxt_port_tx_dw3_read, NULL);
+	MMIO_DH(BXT_PORT_PCS_DW12_GRP(DPIO_PHY1, DPIO_CH0), D_BXT,
+		NULL, bxt_pcs_dw12_grp_write);
+	MMIO_DH(BXT_PORT_TX_DW3_LN0(DPIO_PHY1, DPIO_CH0), D_BXT,
+		bxt_port_tx_dw3_read, NULL);
+	MMIO_DH(BXT_DE_PLL_ENABLE, D_BXT, NULL, bxt_de_pll_enable_write);
+	MMIO_DFH(GEN8_L3SQCREG1, D_BXT, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(GEN8_L3CNTLREG, D_BXT, F_CMD_ACCESS, NULL, NULL);
+	MMIO_DFH(_MMIO(0x20D8), D_BXT, F_CMD_ACCESS, NULL, NULL);
+	MMIO_F(GEN8_RING_CS_GPR(RENDER_RING_BASE, 0), 0x40, F_CMD_ACCESS,
+	       0, 0, D_BXT, NULL, NULL);
+	MMIO_F(GEN8_RING_CS_GPR(GEN6_BSD_RING_BASE, 0), 0x40, F_CMD_ACCESS,
+	       0, 0, D_BXT, NULL, NULL);
+	MMIO_F(GEN8_RING_CS_GPR(BLT_RING_BASE, 0), 0x40, F_CMD_ACCESS,
+	       0, 0, D_BXT, NULL, NULL);
+	MMIO_F(GEN8_RING_CS_GPR(VEBOX_RING_BASE, 0), 0x40, F_CMD_ACCESS,
+	       0, 0, D_BXT, NULL, NULL);
+
+	MMIO_DFH(GEN9_CTX_PREEMPT_REG, D_BXT, F_CMD_ACCESS, NULL, NULL);
+
+	MMIO_DH(GEN8_PRIVATE_PAT_LO, D_BXT, NULL, bxt_ppat_low_write);
+
+	return 0;
+}
+
+static struct gvt_mmio_block *find_mmio_block(struct intel_gvt *gvt,
+					      unsigned int offset)
+{
+	struct gvt_mmio_block *block = gvt->mmio.mmio_block;
+	int num = gvt->mmio.num_mmio_block;
+	int i;
+
+	for (i = 0; i < num; i++, block++) {
+		if (offset >= i915_mmio_reg_offset(block->offset) &&
+		    offset < i915_mmio_reg_offset(block->offset) + block->size)
+			return block;
+	}
+	return NULL;
+}
+
+/**
+ * intel_gvt_clean_mmio_info - clean up MMIO information table for GVT device
+ * @gvt: GVT device
+ *
+ * This function is called at the driver unloading stage, to clean up the MMIO
+ * information table of GVT device
+ *
+ */
+void intel_gvt_clean_mmio_info(struct intel_gvt *gvt)
+{
+	struct hlist_node *tmp;
+	struct intel_gvt_mmio_info *e;
+	int i;
+
+	hash_for_each_safe(gvt->mmio.mmio_info_table, i, tmp, e, node)
+		kfree(e);
+
+	kfree(gvt->mmio.mmio_block);
+	gvt->mmio.mmio_block = NULL;
+	gvt->mmio.num_mmio_block = 0;
+
+	vfree(gvt->mmio.mmio_attribute);
+	gvt->mmio.mmio_attribute = NULL;
+}
+
+static int handle_mmio(struct intel_gvt_mmio_table_iter *iter, u32 offset,
+		       u32 size)
+{
+	struct intel_gvt *gvt = iter->data;
+	struct intel_gvt_mmio_info *info, *p;
+	u32 start, end, i;
+
+	if (WARN_ON(!IS_ALIGNED(offset, 4)))
+		return -EINVAL;
+
+	start = offset;
+	end = offset + size;
+
+	for (i = start; i < end; i += 4) {
+		p = intel_gvt_find_mmio_info(gvt, i);
+		if (p) {
+			WARN(1, "dup mmio definition offset %x\n",
+				info->offset);
+
+			/* We return -EEXIST here to make GVT-g load fail.
+			 * So duplicated MMIO can be found as soon as
+			 * possible.
+			 */
+			return -EEXIST;
+		}
+
+		info = kzalloc(sizeof(*info), GFP_KERNEL);
+		if (!info)
+			return -ENOMEM;
+
+		info->offset = i;
+		info->read = intel_vgpu_default_mmio_read;
+		info->write = intel_vgpu_default_mmio_write;
+		INIT_HLIST_NODE(&info->node);
+		hash_add(gvt->mmio.mmio_info_table, &info->node, info->offset);
+		gvt->mmio.num_tracked_mmio++;
+	}
+	return 0;
+}
+
+static int handle_mmio_block(struct intel_gvt_mmio_table_iter *iter,
+			     u32 offset, u32 size)
+{
+	struct intel_gvt *gvt = iter->data;
+	struct gvt_mmio_block *block = gvt->mmio.mmio_block;
+	void *ret;
+
+	ret = krealloc(block,
+			 (gvt->mmio.num_mmio_block + 1) * sizeof(*block),
+			 GFP_KERNEL);
+	if (!ret)
+		return -ENOMEM;
+
+	gvt->mmio.mmio_block = block = ret;
+
+	block += gvt->mmio.num_mmio_block;
+
+	memset(block, 0, sizeof(*block));
+
+	block->offset = _MMIO(offset);
+	block->size = size;
+
+	gvt->mmio.num_mmio_block++;
+
+	return 0;
+}
+
+static int handle_mmio_cb(struct intel_gvt_mmio_table_iter *iter, u32 offset,
+			  u32 size)
+{
+	if (size < 1024 || offset == i915_mmio_reg_offset(GEN9_GFX_MOCS(0)))
+		return handle_mmio(iter, offset, size);
+	else
+		return handle_mmio_block(iter, offset, size);
+}
+
+static int init_mmio_info(struct intel_gvt *gvt)
+{
+	struct intel_gvt_mmio_table_iter iter = {
+		.i915 = gvt->gt->i915,
+		.data = gvt,
+		.handle_mmio_cb = handle_mmio_cb,
+	};
+
+	return intel_gvt_iterate_mmio_table(&iter);
+}
+
+static int init_mmio_block_handlers(struct intel_gvt *gvt)
+{
+	struct gvt_mmio_block *block;
+
+	block = find_mmio_block(gvt, VGT_PVINFO_PAGE);
+	if (!block) {
+		WARN(1, "fail to assign handlers to mmio block %x\n",
+		     i915_mmio_reg_offset(gvt->mmio.mmio_block->offset));
+		return -ENODEV;
+	}
+
+	block->read = pvinfo_mmio_read;
+	block->write = pvinfo_mmio_write;
+
+	return 0;
+}
+
+/**
+ * intel_gvt_setup_mmio_info - setup MMIO information table for GVT device
+ * @gvt: GVT device
+ *
+ * This function is called at the initialization stage, to setup the MMIO
+ * information table for GVT device
+ *
+ * Returns:
+ * zero on success, negative if failed.
+ */
+int intel_gvt_setup_mmio_info(struct intel_gvt *gvt)
+{
+	struct intel_gvt_device_info *info = &gvt->device_info;
+	struct drm_i915_private *i915 = gvt->gt->i915;
+	int size = info->mmio_size / 4 * sizeof(*gvt->mmio.mmio_attribute);
+	int ret;
+
+	gvt->mmio.mmio_attribute = vzalloc(size);
+	if (!gvt->mmio.mmio_attribute)
+		return -ENOMEM;
+
+	ret = init_mmio_info(gvt);
+	if (ret)
+		goto err;
+
+	ret = init_mmio_block_handlers(gvt);
+	if (ret)
+		goto err;
+
+	ret = init_generic_mmio_info(gvt);
+	if (ret)
+		goto err;
+
+	if (IS_BROADWELL(i915)) {
+		ret = init_bdw_mmio_info(gvt);
+		if (ret)
+			goto err;
+	} else if (IS_SKYLAKE(i915) ||
+		   IS_KABYLAKE(i915) ||
+		   IS_COFFEELAKE(i915) ||
+		   IS_COMETLAKE(i915)) {
+		ret = init_bdw_mmio_info(gvt);
+		if (ret)
+			goto err;
+		ret = init_skl_mmio_info(gvt);
+		if (ret)
+			goto err;
+	} else if (IS_BROXTON(i915)) {
+		ret = init_bdw_mmio_info(gvt);
+		if (ret)
+			goto err;
+		ret = init_skl_mmio_info(gvt);
+		if (ret)
+			goto err;
+		ret = init_bxt_mmio_info(gvt);
+		if (ret)
+			goto err;
+	}
+
+	return 0;
+err:
+	intel_gvt_clean_mmio_info(gvt);
+	return ret;
+}
+
+/**
+ * intel_gvt_for_each_tracked_mmio - iterate each tracked mmio
+ * @gvt: a GVT device
+ * @handler: the handler
+ * @data: private data given to handler
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_gvt_for_each_tracked_mmio(struct intel_gvt *gvt,
+	int (*handler)(struct intel_gvt *gvt, u32 offset, void *data),
+	void *data)
+{
+	struct gvt_mmio_block *block = gvt->mmio.mmio_block;
+	struct intel_gvt_mmio_info *e;
+	int i, j, ret;
+
+	hash_for_each(gvt->mmio.mmio_info_table, i, e, node) {
+		ret = handler(gvt, e->offset, data);
+		if (ret)
+			return ret;
+	}
+
+	for (i = 0; i < gvt->mmio.num_mmio_block; i++, block++) {
+		/* pvinfo data doesn't come from hw mmio */
+		if (i915_mmio_reg_offset(block->offset) == VGT_PVINFO_PAGE)
+			continue;
+
+		for (j = 0; j < block->size; j += 4) {
+			ret = handler(gvt, i915_mmio_reg_offset(block->offset) + j, data);
+			if (ret)
+				return ret;
+		}
+	}
+	return 0;
+}
+
+/**
+ * intel_vgpu_default_mmio_read - default MMIO read handler
+ * @vgpu: a vGPU
+ * @offset: access offset
+ * @p_data: data return buffer
+ * @bytes: access data length
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_default_mmio_read(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	read_vreg(vgpu, offset, p_data, bytes);
+	return 0;
+}
+
+/**
+ * intel_vgpu_default_mmio_write() - default MMIO write handler
+ * @vgpu: a vGPU
+ * @offset: access offset
+ * @p_data: write data buffer
+ * @bytes: access data length
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_default_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	write_vreg(vgpu, offset, p_data, bytes);
+	return 0;
+}
+
+/**
+ * intel_vgpu_mask_mmio_write - write mask register
+ * @vgpu: a vGPU
+ * @offset: access offset
+ * @p_data: write data buffer
+ * @bytes: access data length
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_mask_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+		void *p_data, unsigned int bytes)
+{
+	u32 mask, old_vreg;
+
+	old_vreg = vgpu_vreg(vgpu, offset);
+	write_vreg(vgpu, offset, p_data, bytes);
+	mask = vgpu_vreg(vgpu, offset) >> 16;
+	vgpu_vreg(vgpu, offset) = (old_vreg & ~mask) |
+				(vgpu_vreg(vgpu, offset) & mask);
+
+	return 0;
+}
+
+/**
+ * intel_gvt_in_force_nonpriv_whitelist - if a mmio is in whitelist to be
+ * force-nopriv register
+ *
+ * @gvt: a GVT device
+ * @offset: register offset
+ *
+ * Returns:
+ * True if the register is in force-nonpriv whitelist;
+ * False if outside;
+ */
+bool intel_gvt_in_force_nonpriv_whitelist(struct intel_gvt *gvt,
+					  unsigned int offset)
+{
+	return in_whitelist(offset);
+}
+
+/**
+ * intel_vgpu_mmio_reg_rw - emulate tracked mmio registers
+ * @vgpu: a vGPU
+ * @offset: register offset
+ * @pdata: data buffer
+ * @bytes: data length
+ * @is_read: read or write
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_mmio_reg_rw(struct intel_vgpu *vgpu, unsigned int offset,
+			   void *pdata, unsigned int bytes, bool is_read)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_gvt_mmio_info *mmio_info;
+	struct gvt_mmio_block *mmio_block;
+	gvt_mmio_func func;
+	int ret;
+
+	if (drm_WARN_ON(&i915->drm, bytes > 8))
+		return -EINVAL;
+
+	/*
+	 * Handle special MMIO blocks.
+	 */
+	mmio_block = find_mmio_block(gvt, offset);
+	if (mmio_block) {
+		func = is_read ? mmio_block->read : mmio_block->write;
+		if (func)
+			return func(vgpu, offset, pdata, bytes);
+		goto default_rw;
+	}
+
+	/*
+	 * Normal tracked MMIOs.
+	 */
+	mmio_info = intel_gvt_find_mmio_info(gvt, offset);
+	if (!mmio_info) {
+		gvt_dbg_mmio("untracked MMIO %08x len %d\n", offset, bytes);
+		goto default_rw;
+	}
+
+	if (is_read)
+		return mmio_info->read(vgpu, offset, pdata, bytes);
+	else {
+		u64 ro_mask = mmio_info->ro_mask;
+		u32 old_vreg = 0;
+		u64 data = 0;
+
+		if (intel_gvt_mmio_has_mode_mask(gvt, mmio_info->offset)) {
+			old_vreg = vgpu_vreg(vgpu, offset);
+		}
+
+		if (likely(!ro_mask))
+			ret = mmio_info->write(vgpu, offset, pdata, bytes);
+		else if (!~ro_mask) {
+			gvt_vgpu_err("try to write RO reg %x\n", offset);
+			return 0;
+		} else {
+			/* keep the RO bits in the virtual register */
+			memcpy(&data, pdata, bytes);
+			data &= ~ro_mask;
+			data |= vgpu_vreg(vgpu, offset) & ro_mask;
+			ret = mmio_info->write(vgpu, offset, &data, bytes);
+		}
+
+		/* higher 16bits of mode ctl regs are mask bits for change */
+		if (intel_gvt_mmio_has_mode_mask(gvt, mmio_info->offset)) {
+			u32 mask = vgpu_vreg(vgpu, offset) >> 16;
+
+			vgpu_vreg(vgpu, offset) = (old_vreg & ~mask)
+					| (vgpu_vreg(vgpu, offset) & mask);
+		}
+	}
+
+	return ret;
+
+default_rw:
+	return is_read ?
+		intel_vgpu_default_mmio_read(vgpu, offset, pdata, bytes) :
+		intel_vgpu_default_mmio_write(vgpu, offset, pdata, bytes);
+}
+
+void intel_gvt_restore_fence(struct intel_gvt *gvt)
+{
+	struct intel_vgpu *vgpu;
+	int i, id;
+
+	idr_for_each_entry(&(gvt)->vgpu_idr, vgpu, id) {
+		mmio_hw_access_pre(gvt->gt);
+		for (i = 0; i < vgpu_fence_sz(vgpu); i++)
+			intel_vgpu_write_fence(vgpu, i, vgpu_vreg64(vgpu, fence_num_to_offset(i)));
+		mmio_hw_access_post(gvt->gt);
+	}
+}
+
+static int mmio_pm_restore_handler(struct intel_gvt *gvt, u32 offset, void *data)
+{
+	struct intel_vgpu *vgpu = data;
+	struct drm_i915_private *dev_priv = gvt->gt->i915;
+
+	if (gvt->mmio.mmio_attribute[offset >> 2] & F_PM_SAVE)
+		intel_uncore_write(&dev_priv->uncore, _MMIO(offset), vgpu_vreg(vgpu, offset));
+
+	return 0;
+}
+
+void intel_gvt_restore_mmio(struct intel_gvt *gvt)
+{
+	struct intel_vgpu *vgpu;
+	int id;
+
+	idr_for_each_entry(&(gvt)->vgpu_idr, vgpu, id) {
+		mmio_hw_access_pre(gvt->gt);
+		intel_gvt_for_each_tracked_mmio(gvt, mmio_pm_restore_handler, vgpu);
+		mmio_hw_access_post(gvt->gt);
+	}
+}
diff --git a/drivers/gpu/drm/i915/gvt/interrupt.c b/drivers/gpu/drm/i915/gvt/interrupt.c
new file mode 100644
index 000000000..a6b2021b6
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/interrupt.c
@@ -0,0 +1,717 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Kevin Tian <kevin.tian@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ * Contributors:
+ *    Min he <min.he@intel.com>
+ *
+ */
+
+#include <linux/eventfd.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "gvt.h"
+#include "trace.h"
+
+/* common offset among interrupt control registers */
+#define regbase_to_isr(base)	(base)
+#define regbase_to_imr(base)	(base + 0x4)
+#define regbase_to_iir(base)	(base + 0x8)
+#define regbase_to_ier(base)	(base + 0xC)
+
+#define iir_to_regbase(iir)    (iir - 0x8)
+#define ier_to_regbase(ier)    (ier - 0xC)
+
+#define get_event_virt_handler(irq, e)	(irq->events[e].v_handler)
+#define get_irq_info(irq, e)		(irq->events[e].info)
+
+#define irq_to_gvt(irq) \
+	container_of(irq, struct intel_gvt, irq)
+
+static void update_upstream_irq(struct intel_vgpu *vgpu,
+		struct intel_gvt_irq_info *info);
+
+static const char * const irq_name[INTEL_GVT_EVENT_MAX] = {
+	[RCS_MI_USER_INTERRUPT] = "Render CS MI USER INTERRUPT",
+	[RCS_DEBUG] = "Render EU debug from SVG",
+	[RCS_MMIO_SYNC_FLUSH] = "Render MMIO sync flush status",
+	[RCS_CMD_STREAMER_ERR] = "Render CS error interrupt",
+	[RCS_PIPE_CONTROL] = "Render PIPE CONTROL notify",
+	[RCS_WATCHDOG_EXCEEDED] = "Render CS Watchdog counter exceeded",
+	[RCS_PAGE_DIRECTORY_FAULT] = "Render page directory faults",
+	[RCS_AS_CONTEXT_SWITCH] = "Render AS Context Switch Interrupt",
+
+	[VCS_MI_USER_INTERRUPT] = "Video CS MI USER INTERRUPT",
+	[VCS_MMIO_SYNC_FLUSH] = "Video MMIO sync flush status",
+	[VCS_CMD_STREAMER_ERR] = "Video CS error interrupt",
+	[VCS_MI_FLUSH_DW] = "Video MI FLUSH DW notify",
+	[VCS_WATCHDOG_EXCEEDED] = "Video CS Watchdog counter exceeded",
+	[VCS_PAGE_DIRECTORY_FAULT] = "Video page directory faults",
+	[VCS_AS_CONTEXT_SWITCH] = "Video AS Context Switch Interrupt",
+	[VCS2_MI_USER_INTERRUPT] = "VCS2 Video CS MI USER INTERRUPT",
+	[VCS2_MI_FLUSH_DW] = "VCS2 Video MI FLUSH DW notify",
+	[VCS2_AS_CONTEXT_SWITCH] = "VCS2 Context Switch Interrupt",
+
+	[BCS_MI_USER_INTERRUPT] = "Blitter CS MI USER INTERRUPT",
+	[BCS_MMIO_SYNC_FLUSH] = "Billter MMIO sync flush status",
+	[BCS_CMD_STREAMER_ERR] = "Blitter CS error interrupt",
+	[BCS_MI_FLUSH_DW] = "Blitter MI FLUSH DW notify",
+	[BCS_PAGE_DIRECTORY_FAULT] = "Blitter page directory faults",
+	[BCS_AS_CONTEXT_SWITCH] = "Blitter AS Context Switch Interrupt",
+
+	[VECS_MI_FLUSH_DW] = "Video Enhanced Streamer MI FLUSH DW notify",
+	[VECS_AS_CONTEXT_SWITCH] = "VECS Context Switch Interrupt",
+
+	[PIPE_A_FIFO_UNDERRUN] = "Pipe A FIFO underrun",
+	[PIPE_A_CRC_ERR] = "Pipe A CRC error",
+	[PIPE_A_CRC_DONE] = "Pipe A CRC done",
+	[PIPE_A_VSYNC] = "Pipe A vsync",
+	[PIPE_A_LINE_COMPARE] = "Pipe A line compare",
+	[PIPE_A_ODD_FIELD] = "Pipe A odd field",
+	[PIPE_A_EVEN_FIELD] = "Pipe A even field",
+	[PIPE_A_VBLANK] = "Pipe A vblank",
+	[PIPE_B_FIFO_UNDERRUN] = "Pipe B FIFO underrun",
+	[PIPE_B_CRC_ERR] = "Pipe B CRC error",
+	[PIPE_B_CRC_DONE] = "Pipe B CRC done",
+	[PIPE_B_VSYNC] = "Pipe B vsync",
+	[PIPE_B_LINE_COMPARE] = "Pipe B line compare",
+	[PIPE_B_ODD_FIELD] = "Pipe B odd field",
+	[PIPE_B_EVEN_FIELD] = "Pipe B even field",
+	[PIPE_B_VBLANK] = "Pipe B vblank",
+	[PIPE_C_VBLANK] = "Pipe C vblank",
+	[DPST_PHASE_IN] = "DPST phase in event",
+	[DPST_HISTOGRAM] = "DPST histogram event",
+	[GSE] = "GSE",
+	[DP_A_HOTPLUG] = "DP A Hotplug",
+	[AUX_CHANNEL_A] = "AUX Channel A",
+	[PERF_COUNTER] = "Performance counter",
+	[POISON] = "Poison",
+	[GTT_FAULT] = "GTT fault",
+	[PRIMARY_A_FLIP_DONE] = "Primary Plane A flip done",
+	[PRIMARY_B_FLIP_DONE] = "Primary Plane B flip done",
+	[PRIMARY_C_FLIP_DONE] = "Primary Plane C flip done",
+	[SPRITE_A_FLIP_DONE] = "Sprite Plane A flip done",
+	[SPRITE_B_FLIP_DONE] = "Sprite Plane B flip done",
+	[SPRITE_C_FLIP_DONE] = "Sprite Plane C flip done",
+
+	[PCU_THERMAL] = "PCU Thermal Event",
+	[PCU_PCODE2DRIVER_MAILBOX] = "PCU pcode2driver mailbox event",
+
+	[FDI_RX_INTERRUPTS_TRANSCODER_A] = "FDI RX Interrupts Combined A",
+	[AUDIO_CP_CHANGE_TRANSCODER_A] = "Audio CP Change Transcoder A",
+	[AUDIO_CP_REQUEST_TRANSCODER_A] = "Audio CP Request Transcoder A",
+	[FDI_RX_INTERRUPTS_TRANSCODER_B] = "FDI RX Interrupts Combined B",
+	[AUDIO_CP_CHANGE_TRANSCODER_B] = "Audio CP Change Transcoder B",
+	[AUDIO_CP_REQUEST_TRANSCODER_B] = "Audio CP Request Transcoder B",
+	[FDI_RX_INTERRUPTS_TRANSCODER_C] = "FDI RX Interrupts Combined C",
+	[AUDIO_CP_CHANGE_TRANSCODER_C] = "Audio CP Change Transcoder C",
+	[AUDIO_CP_REQUEST_TRANSCODER_C] = "Audio CP Request Transcoder C",
+	[ERR_AND_DBG] = "South Error and Debug Interrupts Combined",
+	[GMBUS] = "Gmbus",
+	[SDVO_B_HOTPLUG] = "SDVO B hotplug",
+	[CRT_HOTPLUG] = "CRT Hotplug",
+	[DP_B_HOTPLUG] = "DisplayPort/HDMI/DVI B Hotplug",
+	[DP_C_HOTPLUG] = "DisplayPort/HDMI/DVI C Hotplug",
+	[DP_D_HOTPLUG] = "DisplayPort/HDMI/DVI D Hotplug",
+	[AUX_CHANNEL_B] = "AUX Channel B",
+	[AUX_CHANNEL_C] = "AUX Channel C",
+	[AUX_CHANNEL_D] = "AUX Channel D",
+	[AUDIO_POWER_STATE_CHANGE_B] = "Audio Power State change Port B",
+	[AUDIO_POWER_STATE_CHANGE_C] = "Audio Power State change Port C",
+	[AUDIO_POWER_STATE_CHANGE_D] = "Audio Power State change Port D",
+
+	[INTEL_GVT_EVENT_RESERVED] = "RESERVED EVENTS!!!",
+};
+
+static inline struct intel_gvt_irq_info *regbase_to_irq_info(
+		struct intel_gvt *gvt,
+		unsigned int reg)
+{
+	struct intel_gvt_irq *irq = &gvt->irq;
+	int i;
+
+	for_each_set_bit(i, irq->irq_info_bitmap, INTEL_GVT_IRQ_INFO_MAX) {
+		if (i915_mmio_reg_offset(irq->info[i]->reg_base) == reg)
+			return irq->info[i];
+	}
+
+	return NULL;
+}
+
+/**
+ * intel_vgpu_reg_imr_handler - Generic IMR register emulation write handler
+ * @vgpu: a vGPU
+ * @reg: register offset written by guest
+ * @p_data: register data written by guest
+ * @bytes: register data length
+ *
+ * This function is used to emulate the generic IMR register bit change
+ * behavior.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+int intel_vgpu_reg_imr_handler(struct intel_vgpu *vgpu,
+	unsigned int reg, void *p_data, unsigned int bytes)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	const struct intel_gvt_irq_ops *ops = gvt->irq.ops;
+	u32 imr = *(u32 *)p_data;
+
+	trace_write_ir(vgpu->id, "IMR", reg, imr, vgpu_vreg(vgpu, reg),
+		       (vgpu_vreg(vgpu, reg) ^ imr));
+
+	vgpu_vreg(vgpu, reg) = imr;
+
+	ops->check_pending_irq(vgpu);
+
+	return 0;
+}
+
+/**
+ * intel_vgpu_reg_master_irq_handler - master IRQ write emulation handler
+ * @vgpu: a vGPU
+ * @reg: register offset written by guest
+ * @p_data: register data written by guest
+ * @bytes: register data length
+ *
+ * This function is used to emulate the master IRQ register on gen8+.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+int intel_vgpu_reg_master_irq_handler(struct intel_vgpu *vgpu,
+	unsigned int reg, void *p_data, unsigned int bytes)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	const struct intel_gvt_irq_ops *ops = gvt->irq.ops;
+	u32 ier = *(u32 *)p_data;
+	u32 virtual_ier = vgpu_vreg(vgpu, reg);
+
+	trace_write_ir(vgpu->id, "MASTER_IRQ", reg, ier, virtual_ier,
+		       (virtual_ier ^ ier));
+
+	/*
+	 * GEN8_MASTER_IRQ is a special irq register,
+	 * only bit 31 is allowed to be modified
+	 * and treated as an IER bit.
+	 */
+	ier &= GEN8_MASTER_IRQ_CONTROL;
+	virtual_ier &= GEN8_MASTER_IRQ_CONTROL;
+	vgpu_vreg(vgpu, reg) &= ~GEN8_MASTER_IRQ_CONTROL;
+	vgpu_vreg(vgpu, reg) |= ier;
+
+	ops->check_pending_irq(vgpu);
+
+	return 0;
+}
+
+/**
+ * intel_vgpu_reg_ier_handler - Generic IER write emulation handler
+ * @vgpu: a vGPU
+ * @reg: register offset written by guest
+ * @p_data: register data written by guest
+ * @bytes: register data length
+ *
+ * This function is used to emulate the generic IER register behavior.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+int intel_vgpu_reg_ier_handler(struct intel_vgpu *vgpu,
+	unsigned int reg, void *p_data, unsigned int bytes)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct drm_i915_private *i915 = gvt->gt->i915;
+	const struct intel_gvt_irq_ops *ops = gvt->irq.ops;
+	struct intel_gvt_irq_info *info;
+	u32 ier = *(u32 *)p_data;
+
+	trace_write_ir(vgpu->id, "IER", reg, ier, vgpu_vreg(vgpu, reg),
+		       (vgpu_vreg(vgpu, reg) ^ ier));
+
+	vgpu_vreg(vgpu, reg) = ier;
+
+	info = regbase_to_irq_info(gvt, ier_to_regbase(reg));
+	if (drm_WARN_ON(&i915->drm, !info))
+		return -EINVAL;
+
+	if (info->has_upstream_irq)
+		update_upstream_irq(vgpu, info);
+
+	ops->check_pending_irq(vgpu);
+
+	return 0;
+}
+
+/**
+ * intel_vgpu_reg_iir_handler - Generic IIR write emulation handler
+ * @vgpu: a vGPU
+ * @reg: register offset written by guest
+ * @p_data: register data written by guest
+ * @bytes: register data length
+ *
+ * This function is used to emulate the generic IIR register behavior.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+int intel_vgpu_reg_iir_handler(struct intel_vgpu *vgpu, unsigned int reg,
+	void *p_data, unsigned int bytes)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	struct intel_gvt_irq_info *info = regbase_to_irq_info(vgpu->gvt,
+		iir_to_regbase(reg));
+	u32 iir = *(u32 *)p_data;
+
+	trace_write_ir(vgpu->id, "IIR", reg, iir, vgpu_vreg(vgpu, reg),
+		       (vgpu_vreg(vgpu, reg) ^ iir));
+
+	if (drm_WARN_ON(&i915->drm, !info))
+		return -EINVAL;
+
+	vgpu_vreg(vgpu, reg) &= ~iir;
+
+	if (info->has_upstream_irq)
+		update_upstream_irq(vgpu, info);
+	return 0;
+}
+
+static struct intel_gvt_irq_map gen8_irq_map[] = {
+	{ INTEL_GVT_IRQ_INFO_MASTER, 0, INTEL_GVT_IRQ_INFO_GT0, 0xffff },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 1, INTEL_GVT_IRQ_INFO_GT0, 0xffff0000 },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 2, INTEL_GVT_IRQ_INFO_GT1, 0xffff },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 3, INTEL_GVT_IRQ_INFO_GT1, 0xffff0000 },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 4, INTEL_GVT_IRQ_INFO_GT2, 0xffff },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 6, INTEL_GVT_IRQ_INFO_GT3, 0xffff },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 16, INTEL_GVT_IRQ_INFO_DE_PIPE_A, ~0 },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 17, INTEL_GVT_IRQ_INFO_DE_PIPE_B, ~0 },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 18, INTEL_GVT_IRQ_INFO_DE_PIPE_C, ~0 },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 20, INTEL_GVT_IRQ_INFO_DE_PORT, ~0 },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 22, INTEL_GVT_IRQ_INFO_DE_MISC, ~0 },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 23, INTEL_GVT_IRQ_INFO_PCH, ~0 },
+	{ INTEL_GVT_IRQ_INFO_MASTER, 30, INTEL_GVT_IRQ_INFO_PCU, ~0 },
+	{ -1, -1, ~0 },
+};
+
+static void update_upstream_irq(struct intel_vgpu *vgpu,
+		struct intel_gvt_irq_info *info)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	struct intel_gvt_irq *irq = &vgpu->gvt->irq;
+	struct intel_gvt_irq_map *map = irq->irq_map;
+	struct intel_gvt_irq_info *up_irq_info = NULL;
+	u32 set_bits = 0;
+	u32 clear_bits = 0;
+	int bit;
+	u32 val = vgpu_vreg(vgpu,
+			regbase_to_iir(i915_mmio_reg_offset(info->reg_base)))
+		& vgpu_vreg(vgpu,
+			regbase_to_ier(i915_mmio_reg_offset(info->reg_base)));
+
+	if (!info->has_upstream_irq)
+		return;
+
+	for (map = irq->irq_map; map->up_irq_bit != -1; map++) {
+		if (info->group != map->down_irq_group)
+			continue;
+
+		if (!up_irq_info)
+			up_irq_info = irq->info[map->up_irq_group];
+		else
+			drm_WARN_ON(&i915->drm, up_irq_info !=
+				    irq->info[map->up_irq_group]);
+
+		bit = map->up_irq_bit;
+
+		if (val & map->down_irq_bitmask)
+			set_bits |= (1 << bit);
+		else
+			clear_bits |= (1 << bit);
+	}
+
+	if (drm_WARN_ON(&i915->drm, !up_irq_info))
+		return;
+
+	if (up_irq_info->group == INTEL_GVT_IRQ_INFO_MASTER) {
+		u32 isr = i915_mmio_reg_offset(up_irq_info->reg_base);
+
+		vgpu_vreg(vgpu, isr) &= ~clear_bits;
+		vgpu_vreg(vgpu, isr) |= set_bits;
+	} else {
+		u32 iir = regbase_to_iir(
+			i915_mmio_reg_offset(up_irq_info->reg_base));
+		u32 imr = regbase_to_imr(
+			i915_mmio_reg_offset(up_irq_info->reg_base));
+
+		vgpu_vreg(vgpu, iir) |= (set_bits & ~vgpu_vreg(vgpu, imr));
+	}
+
+	if (up_irq_info->has_upstream_irq)
+		update_upstream_irq(vgpu, up_irq_info);
+}
+
+static void init_irq_map(struct intel_gvt_irq *irq)
+{
+	struct intel_gvt_irq_map *map;
+	struct intel_gvt_irq_info *up_info, *down_info;
+	int up_bit;
+
+	for (map = irq->irq_map; map->up_irq_bit != -1; map++) {
+		up_info = irq->info[map->up_irq_group];
+		up_bit = map->up_irq_bit;
+		down_info = irq->info[map->down_irq_group];
+
+		set_bit(up_bit, up_info->downstream_irq_bitmap);
+		down_info->has_upstream_irq = true;
+
+		gvt_dbg_irq("[up] grp %d bit %d -> [down] grp %d bitmask %x\n",
+			up_info->group, up_bit,
+			down_info->group, map->down_irq_bitmask);
+	}
+}
+
+/* =======================vEvent injection===================== */
+
+#define MSI_CAP_CONTROL(offset) (offset + 2)
+#define MSI_CAP_ADDRESS(offset) (offset + 4)
+#define MSI_CAP_DATA(offset) (offset + 8)
+#define MSI_CAP_EN 0x1
+
+static int inject_virtual_interrupt(struct intel_vgpu *vgpu)
+{
+	unsigned long offset = vgpu->gvt->device_info.msi_cap_offset;
+	u16 control, data;
+	u32 addr;
+
+	control = *(u16 *)(vgpu_cfg_space(vgpu) + MSI_CAP_CONTROL(offset));
+	addr = *(u32 *)(vgpu_cfg_space(vgpu) + MSI_CAP_ADDRESS(offset));
+	data = *(u16 *)(vgpu_cfg_space(vgpu) + MSI_CAP_DATA(offset));
+
+	/* Do not generate MSI if MSIEN is disabled */
+	if (!(control & MSI_CAP_EN))
+		return 0;
+
+	if (WARN(control & GENMASK(15, 1), "only support one MSI format\n"))
+		return -EINVAL;
+
+	trace_inject_msi(vgpu->id, addr, data);
+
+	/*
+	 * When guest is powered off, msi_trigger is set to NULL, but vgpu's
+	 * config and mmio register isn't restored to default during guest
+	 * poweroff. If this vgpu is still used in next vm, this vgpu's pipe
+	 * may be enabled, then once this vgpu is active, it will get inject
+	 * vblank interrupt request. But msi_trigger is null until msi is
+	 * enabled by guest. so if msi_trigger is null, success is still
+	 * returned and don't inject interrupt into guest.
+	 */
+	if (!vgpu->attached)
+		return -ESRCH;
+	if (vgpu->msi_trigger && eventfd_signal(vgpu->msi_trigger, 1) != 1)
+		return -EFAULT;
+	return 0;
+}
+
+static void propagate_event(struct intel_gvt_irq *irq,
+	enum intel_gvt_event_type event, struct intel_vgpu *vgpu)
+{
+	struct intel_gvt_irq_info *info;
+	unsigned int reg_base;
+	int bit;
+
+	info = get_irq_info(irq, event);
+	if (WARN_ON(!info))
+		return;
+
+	reg_base = i915_mmio_reg_offset(info->reg_base);
+	bit = irq->events[event].bit;
+
+	if (!test_bit(bit, (void *)&vgpu_vreg(vgpu,
+					regbase_to_imr(reg_base)))) {
+		trace_propagate_event(vgpu->id, irq_name[event], bit);
+		set_bit(bit, (void *)&vgpu_vreg(vgpu,
+					regbase_to_iir(reg_base)));
+	}
+}
+
+/* =======================vEvent Handlers===================== */
+static void handle_default_event_virt(struct intel_gvt_irq *irq,
+	enum intel_gvt_event_type event, struct intel_vgpu *vgpu)
+{
+	if (!vgpu->irq.irq_warn_once[event]) {
+		gvt_dbg_core("vgpu%d: IRQ receive event %d (%s)\n",
+			vgpu->id, event, irq_name[event]);
+		vgpu->irq.irq_warn_once[event] = true;
+	}
+	propagate_event(irq, event, vgpu);
+}
+
+/* =====================GEN specific logic======================= */
+/* GEN8 interrupt routines. */
+
+#define DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(regname, regbase) \
+static struct intel_gvt_irq_info gen8_##regname##_info = { \
+	.name = #regname"-IRQ", \
+	.reg_base = (regbase), \
+	.bit_to_event = {[0 ... INTEL_GVT_IRQ_BITWIDTH-1] = \
+		INTEL_GVT_EVENT_RESERVED}, \
+}
+
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(gt0, GEN8_GT_ISR(0));
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(gt1, GEN8_GT_ISR(1));
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(gt2, GEN8_GT_ISR(2));
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(gt3, GEN8_GT_ISR(3));
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(de_pipe_a, GEN8_DE_PIPE_ISR(PIPE_A));
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(de_pipe_b, GEN8_DE_PIPE_ISR(PIPE_B));
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(de_pipe_c, GEN8_DE_PIPE_ISR(PIPE_C));
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(de_port, GEN8_DE_PORT_ISR);
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(de_misc, GEN8_DE_MISC_ISR);
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(pcu, GEN8_PCU_ISR);
+DEFINE_GVT_GEN8_INTEL_GVT_IRQ_INFO(master, GEN8_MASTER_IRQ);
+
+static struct intel_gvt_irq_info gvt_base_pch_info = {
+	.name = "PCH-IRQ",
+	.reg_base = SDEISR,
+	.bit_to_event = {[0 ... INTEL_GVT_IRQ_BITWIDTH-1] =
+		INTEL_GVT_EVENT_RESERVED},
+};
+
+static void gen8_check_pending_irq(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt_irq *irq = &vgpu->gvt->irq;
+	int i;
+
+	if (!(vgpu_vreg(vgpu, i915_mmio_reg_offset(GEN8_MASTER_IRQ)) &
+				GEN8_MASTER_IRQ_CONTROL))
+		return;
+
+	for_each_set_bit(i, irq->irq_info_bitmap, INTEL_GVT_IRQ_INFO_MAX) {
+		struct intel_gvt_irq_info *info = irq->info[i];
+		u32 reg_base;
+
+		if (!info->has_upstream_irq)
+			continue;
+
+		reg_base = i915_mmio_reg_offset(info->reg_base);
+		if ((vgpu_vreg(vgpu, regbase_to_iir(reg_base))
+				& vgpu_vreg(vgpu, regbase_to_ier(reg_base))))
+			update_upstream_irq(vgpu, info);
+	}
+
+	if (vgpu_vreg(vgpu, i915_mmio_reg_offset(GEN8_MASTER_IRQ))
+			& ~GEN8_MASTER_IRQ_CONTROL)
+		inject_virtual_interrupt(vgpu);
+}
+
+static void gen8_init_irq(
+		struct intel_gvt_irq *irq)
+{
+	struct intel_gvt *gvt = irq_to_gvt(irq);
+
+#define SET_BIT_INFO(s, b, e, i)		\
+	do {					\
+		s->events[e].bit = b;		\
+		s->events[e].info = s->info[i];	\
+		s->info[i]->bit_to_event[b] = e;\
+	} while (0)
+
+#define SET_IRQ_GROUP(s, g, i) \
+	do { \
+		s->info[g] = i; \
+		(i)->group = g; \
+		set_bit(g, s->irq_info_bitmap); \
+	} while (0)
+
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_MASTER, &gen8_master_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_GT0, &gen8_gt0_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_GT1, &gen8_gt1_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_GT2, &gen8_gt2_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_GT3, &gen8_gt3_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_DE_PIPE_A, &gen8_de_pipe_a_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_DE_PIPE_B, &gen8_de_pipe_b_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_DE_PIPE_C, &gen8_de_pipe_c_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_DE_PORT, &gen8_de_port_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_DE_MISC, &gen8_de_misc_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_PCU, &gen8_pcu_info);
+	SET_IRQ_GROUP(irq, INTEL_GVT_IRQ_INFO_PCH, &gvt_base_pch_info);
+
+	/* GEN8 level 2 interrupts. */
+
+	/* GEN8 interrupt GT0 events */
+	SET_BIT_INFO(irq, 0, RCS_MI_USER_INTERRUPT, INTEL_GVT_IRQ_INFO_GT0);
+	SET_BIT_INFO(irq, 4, RCS_PIPE_CONTROL, INTEL_GVT_IRQ_INFO_GT0);
+	SET_BIT_INFO(irq, 8, RCS_AS_CONTEXT_SWITCH, INTEL_GVT_IRQ_INFO_GT0);
+
+	SET_BIT_INFO(irq, 16, BCS_MI_USER_INTERRUPT, INTEL_GVT_IRQ_INFO_GT0);
+	SET_BIT_INFO(irq, 20, BCS_MI_FLUSH_DW, INTEL_GVT_IRQ_INFO_GT0);
+	SET_BIT_INFO(irq, 24, BCS_AS_CONTEXT_SWITCH, INTEL_GVT_IRQ_INFO_GT0);
+
+	/* GEN8 interrupt GT1 events */
+	SET_BIT_INFO(irq, 0, VCS_MI_USER_INTERRUPT, INTEL_GVT_IRQ_INFO_GT1);
+	SET_BIT_INFO(irq, 4, VCS_MI_FLUSH_DW, INTEL_GVT_IRQ_INFO_GT1);
+	SET_BIT_INFO(irq, 8, VCS_AS_CONTEXT_SWITCH, INTEL_GVT_IRQ_INFO_GT1);
+
+	if (HAS_ENGINE(gvt->gt, VCS1)) {
+		SET_BIT_INFO(irq, 16, VCS2_MI_USER_INTERRUPT,
+			INTEL_GVT_IRQ_INFO_GT1);
+		SET_BIT_INFO(irq, 20, VCS2_MI_FLUSH_DW,
+			INTEL_GVT_IRQ_INFO_GT1);
+		SET_BIT_INFO(irq, 24, VCS2_AS_CONTEXT_SWITCH,
+			INTEL_GVT_IRQ_INFO_GT1);
+	}
+
+	/* GEN8 interrupt GT3 events */
+	SET_BIT_INFO(irq, 0, VECS_MI_USER_INTERRUPT, INTEL_GVT_IRQ_INFO_GT3);
+	SET_BIT_INFO(irq, 4, VECS_MI_FLUSH_DW, INTEL_GVT_IRQ_INFO_GT3);
+	SET_BIT_INFO(irq, 8, VECS_AS_CONTEXT_SWITCH, INTEL_GVT_IRQ_INFO_GT3);
+
+	SET_BIT_INFO(irq, 0, PIPE_A_VBLANK, INTEL_GVT_IRQ_INFO_DE_PIPE_A);
+	SET_BIT_INFO(irq, 0, PIPE_B_VBLANK, INTEL_GVT_IRQ_INFO_DE_PIPE_B);
+	SET_BIT_INFO(irq, 0, PIPE_C_VBLANK, INTEL_GVT_IRQ_INFO_DE_PIPE_C);
+
+	/* GEN8 interrupt DE PORT events */
+	SET_BIT_INFO(irq, 0, AUX_CHANNEL_A, INTEL_GVT_IRQ_INFO_DE_PORT);
+	SET_BIT_INFO(irq, 3, DP_A_HOTPLUG, INTEL_GVT_IRQ_INFO_DE_PORT);
+
+	/* GEN8 interrupt DE MISC events */
+	SET_BIT_INFO(irq, 0, GSE, INTEL_GVT_IRQ_INFO_DE_MISC);
+
+	/* PCH events */
+	SET_BIT_INFO(irq, 17, GMBUS, INTEL_GVT_IRQ_INFO_PCH);
+	SET_BIT_INFO(irq, 19, CRT_HOTPLUG, INTEL_GVT_IRQ_INFO_PCH);
+	SET_BIT_INFO(irq, 21, DP_B_HOTPLUG, INTEL_GVT_IRQ_INFO_PCH);
+	SET_BIT_INFO(irq, 22, DP_C_HOTPLUG, INTEL_GVT_IRQ_INFO_PCH);
+	SET_BIT_INFO(irq, 23, DP_D_HOTPLUG, INTEL_GVT_IRQ_INFO_PCH);
+
+	if (IS_BROADWELL(gvt->gt->i915)) {
+		SET_BIT_INFO(irq, 25, AUX_CHANNEL_B, INTEL_GVT_IRQ_INFO_PCH);
+		SET_BIT_INFO(irq, 26, AUX_CHANNEL_C, INTEL_GVT_IRQ_INFO_PCH);
+		SET_BIT_INFO(irq, 27, AUX_CHANNEL_D, INTEL_GVT_IRQ_INFO_PCH);
+
+		SET_BIT_INFO(irq, 4, PRIMARY_A_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_A);
+		SET_BIT_INFO(irq, 5, SPRITE_A_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_A);
+
+		SET_BIT_INFO(irq, 4, PRIMARY_B_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_B);
+		SET_BIT_INFO(irq, 5, SPRITE_B_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_B);
+
+		SET_BIT_INFO(irq, 4, PRIMARY_C_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_C);
+		SET_BIT_INFO(irq, 5, SPRITE_C_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_C);
+	} else if (GRAPHICS_VER(gvt->gt->i915) >= 9) {
+		SET_BIT_INFO(irq, 25, AUX_CHANNEL_B, INTEL_GVT_IRQ_INFO_DE_PORT);
+		SET_BIT_INFO(irq, 26, AUX_CHANNEL_C, INTEL_GVT_IRQ_INFO_DE_PORT);
+		SET_BIT_INFO(irq, 27, AUX_CHANNEL_D, INTEL_GVT_IRQ_INFO_DE_PORT);
+
+		SET_BIT_INFO(irq, 3, PRIMARY_A_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_A);
+		SET_BIT_INFO(irq, 3, PRIMARY_B_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_B);
+		SET_BIT_INFO(irq, 3, PRIMARY_C_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_C);
+
+		SET_BIT_INFO(irq, 4, SPRITE_A_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_A);
+		SET_BIT_INFO(irq, 4, SPRITE_B_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_B);
+		SET_BIT_INFO(irq, 4, SPRITE_C_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_C);
+	}
+
+	/* GEN8 interrupt PCU events */
+	SET_BIT_INFO(irq, 24, PCU_THERMAL, INTEL_GVT_IRQ_INFO_PCU);
+	SET_BIT_INFO(irq, 25, PCU_PCODE2DRIVER_MAILBOX, INTEL_GVT_IRQ_INFO_PCU);
+}
+
+static const struct intel_gvt_irq_ops gen8_irq_ops = {
+	.init_irq = gen8_init_irq,
+	.check_pending_irq = gen8_check_pending_irq,
+};
+
+/**
+ * intel_vgpu_trigger_virtual_event - Trigger a virtual event for a vGPU
+ * @vgpu: a vGPU
+ * @event: interrupt event
+ *
+ * This function is used to trigger a virtual interrupt event for vGPU.
+ * The caller provides the event to be triggered, the framework itself
+ * will emulate the IRQ register bit change.
+ *
+ */
+void intel_vgpu_trigger_virtual_event(struct intel_vgpu *vgpu,
+	enum intel_gvt_event_type event)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_gvt_irq *irq = &gvt->irq;
+	gvt_event_virt_handler_t handler;
+	const struct intel_gvt_irq_ops *ops = gvt->irq.ops;
+
+	handler = get_event_virt_handler(irq, event);
+	drm_WARN_ON(&i915->drm, !handler);
+
+	handler(irq, event, vgpu);
+
+	ops->check_pending_irq(vgpu);
+}
+
+static void init_events(
+	struct intel_gvt_irq *irq)
+{
+	int i;
+
+	for (i = 0; i < INTEL_GVT_EVENT_MAX; i++) {
+		irq->events[i].info = NULL;
+		irq->events[i].v_handler = handle_default_event_virt;
+	}
+}
+
+/**
+ * intel_gvt_init_irq - initialize GVT-g IRQ emulation subsystem
+ * @gvt: a GVT device
+ *
+ * This function is called at driver loading stage, to initialize the GVT-g IRQ
+ * emulation subsystem.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_gvt_init_irq(struct intel_gvt *gvt)
+{
+	struct intel_gvt_irq *irq = &gvt->irq;
+
+	gvt_dbg_core("init irq framework\n");
+
+	irq->ops = &gen8_irq_ops;
+	irq->irq_map = gen8_irq_map;
+
+	/* common event initialization */
+	init_events(irq);
+
+	/* gen specific initialization */
+	irq->ops->init_irq(irq);
+
+	init_irq_map(irq);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gvt/interrupt.h b/drivers/gpu/drm/i915/gvt/interrupt.h
new file mode 100644
index 000000000..b62f04ab4
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/interrupt.h
@@ -0,0 +1,232 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Kevin Tian <kevin.tian@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ * Contributors:
+ *    Min he <min.he@intel.com>
+ *
+ */
+
+#ifndef _GVT_INTERRUPT_H_
+#define _GVT_INTERRUPT_H_
+
+#include <linux/hrtimer.h>
+#include <linux/kernel.h>
+
+#include "i915_reg_defs.h"
+
+enum intel_gvt_event_type {
+	RCS_MI_USER_INTERRUPT = 0,
+	RCS_DEBUG,
+	RCS_MMIO_SYNC_FLUSH,
+	RCS_CMD_STREAMER_ERR,
+	RCS_PIPE_CONTROL,
+	RCS_L3_PARITY_ERR,
+	RCS_WATCHDOG_EXCEEDED,
+	RCS_PAGE_DIRECTORY_FAULT,
+	RCS_AS_CONTEXT_SWITCH,
+	RCS_MONITOR_BUFF_HALF_FULL,
+
+	VCS_MI_USER_INTERRUPT,
+	VCS_MMIO_SYNC_FLUSH,
+	VCS_CMD_STREAMER_ERR,
+	VCS_MI_FLUSH_DW,
+	VCS_WATCHDOG_EXCEEDED,
+	VCS_PAGE_DIRECTORY_FAULT,
+	VCS_AS_CONTEXT_SWITCH,
+
+	VCS2_MI_USER_INTERRUPT,
+	VCS2_MI_FLUSH_DW,
+	VCS2_AS_CONTEXT_SWITCH,
+
+	BCS_MI_USER_INTERRUPT,
+	BCS_MMIO_SYNC_FLUSH,
+	BCS_CMD_STREAMER_ERR,
+	BCS_MI_FLUSH_DW,
+	BCS_PAGE_DIRECTORY_FAULT,
+	BCS_AS_CONTEXT_SWITCH,
+
+	VECS_MI_USER_INTERRUPT,
+	VECS_MI_FLUSH_DW,
+	VECS_AS_CONTEXT_SWITCH,
+
+	PIPE_A_FIFO_UNDERRUN,
+	PIPE_B_FIFO_UNDERRUN,
+	PIPE_A_CRC_ERR,
+	PIPE_B_CRC_ERR,
+	PIPE_A_CRC_DONE,
+	PIPE_B_CRC_DONE,
+	PIPE_A_ODD_FIELD,
+	PIPE_B_ODD_FIELD,
+	PIPE_A_EVEN_FIELD,
+	PIPE_B_EVEN_FIELD,
+	PIPE_A_LINE_COMPARE,
+	PIPE_B_LINE_COMPARE,
+	PIPE_C_LINE_COMPARE,
+	PIPE_A_VBLANK,
+	PIPE_B_VBLANK,
+	PIPE_C_VBLANK,
+	PIPE_A_VSYNC,
+	PIPE_B_VSYNC,
+	PIPE_C_VSYNC,
+	PRIMARY_A_FLIP_DONE,
+	PRIMARY_B_FLIP_DONE,
+	PRIMARY_C_FLIP_DONE,
+	SPRITE_A_FLIP_DONE,
+	SPRITE_B_FLIP_DONE,
+	SPRITE_C_FLIP_DONE,
+
+	PCU_THERMAL,
+	PCU_PCODE2DRIVER_MAILBOX,
+
+	DPST_PHASE_IN,
+	DPST_HISTOGRAM,
+	GSE,
+	DP_A_HOTPLUG,
+	AUX_CHANNEL_A,
+	PERF_COUNTER,
+	POISON,
+	GTT_FAULT,
+	ERROR_INTERRUPT_COMBINED,
+
+	FDI_RX_INTERRUPTS_TRANSCODER_A,
+	AUDIO_CP_CHANGE_TRANSCODER_A,
+	AUDIO_CP_REQUEST_TRANSCODER_A,
+	FDI_RX_INTERRUPTS_TRANSCODER_B,
+	AUDIO_CP_CHANGE_TRANSCODER_B,
+	AUDIO_CP_REQUEST_TRANSCODER_B,
+	FDI_RX_INTERRUPTS_TRANSCODER_C,
+	AUDIO_CP_CHANGE_TRANSCODER_C,
+	AUDIO_CP_REQUEST_TRANSCODER_C,
+	ERR_AND_DBG,
+	GMBUS,
+	SDVO_B_HOTPLUG,
+	CRT_HOTPLUG,
+	DP_B_HOTPLUG,
+	DP_C_HOTPLUG,
+	DP_D_HOTPLUG,
+	AUX_CHANNEL_B,
+	AUX_CHANNEL_C,
+	AUX_CHANNEL_D,
+	AUDIO_POWER_STATE_CHANGE_B,
+	AUDIO_POWER_STATE_CHANGE_C,
+	AUDIO_POWER_STATE_CHANGE_D,
+
+	INTEL_GVT_EVENT_RESERVED,
+	INTEL_GVT_EVENT_MAX,
+};
+
+struct intel_gvt_irq;
+struct intel_gvt;
+struct intel_vgpu;
+
+typedef void (*gvt_event_virt_handler_t)(struct intel_gvt_irq *irq,
+	enum intel_gvt_event_type event, struct intel_vgpu *vgpu);
+
+struct intel_gvt_irq_ops {
+	void (*init_irq)(struct intel_gvt_irq *irq);
+	void (*check_pending_irq)(struct intel_vgpu *vgpu);
+};
+
+/* the list of physical interrupt control register groups */
+enum intel_gvt_irq_type {
+	INTEL_GVT_IRQ_INFO_GT,
+	INTEL_GVT_IRQ_INFO_DPY,
+	INTEL_GVT_IRQ_INFO_PCH,
+	INTEL_GVT_IRQ_INFO_PM,
+
+	INTEL_GVT_IRQ_INFO_MASTER,
+	INTEL_GVT_IRQ_INFO_GT0,
+	INTEL_GVT_IRQ_INFO_GT1,
+	INTEL_GVT_IRQ_INFO_GT2,
+	INTEL_GVT_IRQ_INFO_GT3,
+	INTEL_GVT_IRQ_INFO_DE_PIPE_A,
+	INTEL_GVT_IRQ_INFO_DE_PIPE_B,
+	INTEL_GVT_IRQ_INFO_DE_PIPE_C,
+	INTEL_GVT_IRQ_INFO_DE_PORT,
+	INTEL_GVT_IRQ_INFO_DE_MISC,
+	INTEL_GVT_IRQ_INFO_AUD,
+	INTEL_GVT_IRQ_INFO_PCU,
+
+	INTEL_GVT_IRQ_INFO_MAX,
+};
+
+#define INTEL_GVT_IRQ_BITWIDTH	32
+
+/* device specific interrupt bit definitions */
+struct intel_gvt_irq_info {
+	char *name;
+	i915_reg_t reg_base;
+	enum intel_gvt_event_type bit_to_event[INTEL_GVT_IRQ_BITWIDTH];
+	unsigned long warned;
+	int group;
+	DECLARE_BITMAP(downstream_irq_bitmap, INTEL_GVT_IRQ_BITWIDTH);
+	bool has_upstream_irq;
+};
+
+/* per-event information */
+struct intel_gvt_event_info {
+	int bit;				/* map to register bit */
+	int policy;				/* forwarding policy */
+	struct intel_gvt_irq_info *info;	/* register info */
+	gvt_event_virt_handler_t v_handler;	/* for v_event */
+};
+
+struct intel_gvt_irq_map {
+	int up_irq_group;
+	int up_irq_bit;
+	int down_irq_group;
+	u32 down_irq_bitmask;
+};
+
+/* structure containing device specific IRQ state */
+struct intel_gvt_irq {
+	const struct intel_gvt_irq_ops *ops;
+	struct intel_gvt_irq_info *info[INTEL_GVT_IRQ_INFO_MAX];
+	DECLARE_BITMAP(irq_info_bitmap, INTEL_GVT_IRQ_INFO_MAX);
+	struct intel_gvt_event_info events[INTEL_GVT_EVENT_MAX];
+	DECLARE_BITMAP(pending_events, INTEL_GVT_EVENT_MAX);
+	struct intel_gvt_irq_map *irq_map;
+};
+
+int intel_gvt_init_irq(struct intel_gvt *gvt);
+
+void intel_vgpu_trigger_virtual_event(struct intel_vgpu *vgpu,
+	enum intel_gvt_event_type event);
+
+int intel_vgpu_reg_iir_handler(struct intel_vgpu *vgpu, unsigned int reg,
+	void *p_data, unsigned int bytes);
+int intel_vgpu_reg_ier_handler(struct intel_vgpu *vgpu,
+	unsigned int reg, void *p_data, unsigned int bytes);
+int intel_vgpu_reg_master_irq_handler(struct intel_vgpu *vgpu,
+	unsigned int reg, void *p_data, unsigned int bytes);
+int intel_vgpu_reg_imr_handler(struct intel_vgpu *vgpu,
+	unsigned int reg, void *p_data, unsigned int bytes);
+
+int gvt_ring_id_to_pipe_control_notify_event(int ring_id);
+int gvt_ring_id_to_mi_flush_dw_event(int ring_id);
+int gvt_ring_id_to_mi_user_interrupt_event(int ring_id);
+
+#endif /* _GVT_INTERRUPT_H_ */
diff --git a/drivers/gpu/drm/i915/gvt/kvmgt.c b/drivers/gpu/drm/i915/gvt/kvmgt.c
new file mode 100644
index 000000000..714221f9a
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/kvmgt.c
@@ -0,0 +1,2017 @@
+/*
+ * KVMGT - the implementation of Intel mediated pass-through framework for KVM
+ *
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Kevin Tian <kevin.tian@intel.com>
+ *    Jike Song <jike.song@intel.com>
+ *    Xiaoguang Chen <xiaoguang.chen@intel.com>
+ *    Eddie Dong <eddie.dong@intel.com>
+ *
+ * Contributors:
+ *    Niu Bing <bing.niu@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ */
+
+#include <linux/init.h>
+#include <linux/mm.h>
+#include <linux/kthread.h>
+#include <linux/sched/mm.h>
+#include <linux/types.h>
+#include <linux/list.h>
+#include <linux/rbtree.h>
+#include <linux/spinlock.h>
+#include <linux/eventfd.h>
+#include <linux/mdev.h>
+#include <linux/debugfs.h>
+
+#include <linux/nospec.h>
+
+#include <drm/drm_edid.h>
+
+#include "i915_drv.h"
+#include "intel_gvt.h"
+#include "gvt.h"
+
+MODULE_IMPORT_NS(DMA_BUF);
+MODULE_IMPORT_NS(I915_GVT);
+
+/* helper macros copied from vfio-pci */
+#define VFIO_PCI_OFFSET_SHIFT   40
+#define VFIO_PCI_OFFSET_TO_INDEX(off)   (off >> VFIO_PCI_OFFSET_SHIFT)
+#define VFIO_PCI_INDEX_TO_OFFSET(index) ((u64)(index) << VFIO_PCI_OFFSET_SHIFT)
+#define VFIO_PCI_OFFSET_MASK    (((u64)(1) << VFIO_PCI_OFFSET_SHIFT) - 1)
+
+#define EDID_BLOB_OFFSET (PAGE_SIZE/2)
+
+#define OPREGION_SIGNATURE "IntelGraphicsMem"
+
+struct vfio_region;
+struct intel_vgpu_regops {
+	size_t (*rw)(struct intel_vgpu *vgpu, char *buf,
+			size_t count, loff_t *ppos, bool iswrite);
+	void (*release)(struct intel_vgpu *vgpu,
+			struct vfio_region *region);
+};
+
+struct vfio_region {
+	u32				type;
+	u32				subtype;
+	size_t				size;
+	u32				flags;
+	const struct intel_vgpu_regops	*ops;
+	void				*data;
+};
+
+struct vfio_edid_region {
+	struct vfio_region_gfx_edid vfio_edid_regs;
+	void *edid_blob;
+};
+
+struct kvmgt_pgfn {
+	gfn_t gfn;
+	struct hlist_node hnode;
+};
+
+struct gvt_dma {
+	struct intel_vgpu *vgpu;
+	struct rb_node gfn_node;
+	struct rb_node dma_addr_node;
+	gfn_t gfn;
+	dma_addr_t dma_addr;
+	unsigned long size;
+	struct kref ref;
+};
+
+#define vfio_dev_to_vgpu(vfio_dev) \
+	container_of((vfio_dev), struct intel_vgpu, vfio_device)
+
+static void kvmgt_page_track_write(struct kvm_vcpu *vcpu, gpa_t gpa,
+		const u8 *val, int len,
+		struct kvm_page_track_notifier_node *node);
+static void kvmgt_page_track_flush_slot(struct kvm *kvm,
+		struct kvm_memory_slot *slot,
+		struct kvm_page_track_notifier_node *node);
+
+static ssize_t intel_vgpu_show_description(struct mdev_type *mtype, char *buf)
+{
+	struct intel_vgpu_type *type =
+		container_of(mtype, struct intel_vgpu_type, type);
+
+	return sprintf(buf, "low_gm_size: %dMB\nhigh_gm_size: %dMB\n"
+		       "fence: %d\nresolution: %s\n"
+		       "weight: %d\n",
+		       BYTES_TO_MB(type->conf->low_mm),
+		       BYTES_TO_MB(type->conf->high_mm),
+		       type->conf->fence, vgpu_edid_str(type->conf->edid),
+		       type->conf->weight);
+}
+
+static void gvt_unpin_guest_page(struct intel_vgpu *vgpu, unsigned long gfn,
+		unsigned long size)
+{
+	vfio_unpin_pages(&vgpu->vfio_device, gfn << PAGE_SHIFT,
+			 DIV_ROUND_UP(size, PAGE_SIZE));
+}
+
+/* Pin a normal or compound guest page for dma. */
+static int gvt_pin_guest_page(struct intel_vgpu *vgpu, unsigned long gfn,
+		unsigned long size, struct page **page)
+{
+	int total_pages = DIV_ROUND_UP(size, PAGE_SIZE);
+	struct page *base_page = NULL;
+	int npage;
+	int ret;
+
+	/*
+	 * We pin the pages one-by-one to avoid allocating a big arrary
+	 * on stack to hold pfns.
+	 */
+	for (npage = 0; npage < total_pages; npage++) {
+		dma_addr_t cur_iova = (gfn + npage) << PAGE_SHIFT;
+		struct page *cur_page;
+
+		ret = vfio_pin_pages(&vgpu->vfio_device, cur_iova, 1,
+				     IOMMU_READ | IOMMU_WRITE, &cur_page);
+		if (ret != 1) {
+			gvt_vgpu_err("vfio_pin_pages failed for iova %pad, ret %d\n",
+				     &cur_iova, ret);
+			goto err;
+		}
+
+		if (npage == 0)
+			base_page = cur_page;
+		else if (base_page + npage != cur_page) {
+			gvt_vgpu_err("The pages are not continuous\n");
+			ret = -EINVAL;
+			npage++;
+			goto err;
+		}
+	}
+
+	*page = base_page;
+	return 0;
+err:
+	gvt_unpin_guest_page(vgpu, gfn, npage * PAGE_SIZE);
+	return ret;
+}
+
+static int gvt_dma_map_page(struct intel_vgpu *vgpu, unsigned long gfn,
+		dma_addr_t *dma_addr, unsigned long size)
+{
+	struct device *dev = vgpu->gvt->gt->i915->drm.dev;
+	struct page *page = NULL;
+	int ret;
+
+	ret = gvt_pin_guest_page(vgpu, gfn, size, &page);
+	if (ret)
+		return ret;
+
+	/* Setup DMA mapping. */
+	*dma_addr = dma_map_page(dev, page, 0, size, DMA_BIDIRECTIONAL);
+	if (dma_mapping_error(dev, *dma_addr)) {
+		gvt_vgpu_err("DMA mapping failed for pfn 0x%lx, ret %d\n",
+			     page_to_pfn(page), ret);
+		gvt_unpin_guest_page(vgpu, gfn, size);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void gvt_dma_unmap_page(struct intel_vgpu *vgpu, unsigned long gfn,
+		dma_addr_t dma_addr, unsigned long size)
+{
+	struct device *dev = vgpu->gvt->gt->i915->drm.dev;
+
+	dma_unmap_page(dev, dma_addr, size, DMA_BIDIRECTIONAL);
+	gvt_unpin_guest_page(vgpu, gfn, size);
+}
+
+static struct gvt_dma *__gvt_cache_find_dma_addr(struct intel_vgpu *vgpu,
+		dma_addr_t dma_addr)
+{
+	struct rb_node *node = vgpu->dma_addr_cache.rb_node;
+	struct gvt_dma *itr;
+
+	while (node) {
+		itr = rb_entry(node, struct gvt_dma, dma_addr_node);
+
+		if (dma_addr < itr->dma_addr)
+			node = node->rb_left;
+		else if (dma_addr > itr->dma_addr)
+			node = node->rb_right;
+		else
+			return itr;
+	}
+	return NULL;
+}
+
+static struct gvt_dma *__gvt_cache_find_gfn(struct intel_vgpu *vgpu, gfn_t gfn)
+{
+	struct rb_node *node = vgpu->gfn_cache.rb_node;
+	struct gvt_dma *itr;
+
+	while (node) {
+		itr = rb_entry(node, struct gvt_dma, gfn_node);
+
+		if (gfn < itr->gfn)
+			node = node->rb_left;
+		else if (gfn > itr->gfn)
+			node = node->rb_right;
+		else
+			return itr;
+	}
+	return NULL;
+}
+
+static int __gvt_cache_add(struct intel_vgpu *vgpu, gfn_t gfn,
+		dma_addr_t dma_addr, unsigned long size)
+{
+	struct gvt_dma *new, *itr;
+	struct rb_node **link, *parent = NULL;
+
+	new = kzalloc(sizeof(struct gvt_dma), GFP_KERNEL);
+	if (!new)
+		return -ENOMEM;
+
+	new->vgpu = vgpu;
+	new->gfn = gfn;
+	new->dma_addr = dma_addr;
+	new->size = size;
+	kref_init(&new->ref);
+
+	/* gfn_cache maps gfn to struct gvt_dma. */
+	link = &vgpu->gfn_cache.rb_node;
+	while (*link) {
+		parent = *link;
+		itr = rb_entry(parent, struct gvt_dma, gfn_node);
+
+		if (gfn < itr->gfn)
+			link = &parent->rb_left;
+		else
+			link = &parent->rb_right;
+	}
+	rb_link_node(&new->gfn_node, parent, link);
+	rb_insert_color(&new->gfn_node, &vgpu->gfn_cache);
+
+	/* dma_addr_cache maps dma addr to struct gvt_dma. */
+	parent = NULL;
+	link = &vgpu->dma_addr_cache.rb_node;
+	while (*link) {
+		parent = *link;
+		itr = rb_entry(parent, struct gvt_dma, dma_addr_node);
+
+		if (dma_addr < itr->dma_addr)
+			link = &parent->rb_left;
+		else
+			link = &parent->rb_right;
+	}
+	rb_link_node(&new->dma_addr_node, parent, link);
+	rb_insert_color(&new->dma_addr_node, &vgpu->dma_addr_cache);
+
+	vgpu->nr_cache_entries++;
+	return 0;
+}
+
+static void __gvt_cache_remove_entry(struct intel_vgpu *vgpu,
+				struct gvt_dma *entry)
+{
+	rb_erase(&entry->gfn_node, &vgpu->gfn_cache);
+	rb_erase(&entry->dma_addr_node, &vgpu->dma_addr_cache);
+	kfree(entry);
+	vgpu->nr_cache_entries--;
+}
+
+static void gvt_cache_destroy(struct intel_vgpu *vgpu)
+{
+	struct gvt_dma *dma;
+	struct rb_node *node = NULL;
+
+	for (;;) {
+		mutex_lock(&vgpu->cache_lock);
+		node = rb_first(&vgpu->gfn_cache);
+		if (!node) {
+			mutex_unlock(&vgpu->cache_lock);
+			break;
+		}
+		dma = rb_entry(node, struct gvt_dma, gfn_node);
+		gvt_dma_unmap_page(vgpu, dma->gfn, dma->dma_addr, dma->size);
+		__gvt_cache_remove_entry(vgpu, dma);
+		mutex_unlock(&vgpu->cache_lock);
+	}
+}
+
+static void gvt_cache_init(struct intel_vgpu *vgpu)
+{
+	vgpu->gfn_cache = RB_ROOT;
+	vgpu->dma_addr_cache = RB_ROOT;
+	vgpu->nr_cache_entries = 0;
+	mutex_init(&vgpu->cache_lock);
+}
+
+static void kvmgt_protect_table_init(struct intel_vgpu *info)
+{
+	hash_init(info->ptable);
+}
+
+static void kvmgt_protect_table_destroy(struct intel_vgpu *info)
+{
+	struct kvmgt_pgfn *p;
+	struct hlist_node *tmp;
+	int i;
+
+	hash_for_each_safe(info->ptable, i, tmp, p, hnode) {
+		hash_del(&p->hnode);
+		kfree(p);
+	}
+}
+
+static struct kvmgt_pgfn *
+__kvmgt_protect_table_find(struct intel_vgpu *info, gfn_t gfn)
+{
+	struct kvmgt_pgfn *p, *res = NULL;
+
+	hash_for_each_possible(info->ptable, p, hnode, gfn) {
+		if (gfn == p->gfn) {
+			res = p;
+			break;
+		}
+	}
+
+	return res;
+}
+
+static bool kvmgt_gfn_is_write_protected(struct intel_vgpu *info, gfn_t gfn)
+{
+	struct kvmgt_pgfn *p;
+
+	p = __kvmgt_protect_table_find(info, gfn);
+	return !!p;
+}
+
+static void kvmgt_protect_table_add(struct intel_vgpu *info, gfn_t gfn)
+{
+	struct kvmgt_pgfn *p;
+
+	if (kvmgt_gfn_is_write_protected(info, gfn))
+		return;
+
+	p = kzalloc(sizeof(struct kvmgt_pgfn), GFP_ATOMIC);
+	if (WARN(!p, "gfn: 0x%llx\n", gfn))
+		return;
+
+	p->gfn = gfn;
+	hash_add(info->ptable, &p->hnode, gfn);
+}
+
+static void kvmgt_protect_table_del(struct intel_vgpu *info, gfn_t gfn)
+{
+	struct kvmgt_pgfn *p;
+
+	p = __kvmgt_protect_table_find(info, gfn);
+	if (p) {
+		hash_del(&p->hnode);
+		kfree(p);
+	}
+}
+
+static size_t intel_vgpu_reg_rw_opregion(struct intel_vgpu *vgpu, char *buf,
+		size_t count, loff_t *ppos, bool iswrite)
+{
+	unsigned int i = VFIO_PCI_OFFSET_TO_INDEX(*ppos) -
+			VFIO_PCI_NUM_REGIONS;
+	void *base = vgpu->region[i].data;
+	loff_t pos = *ppos & VFIO_PCI_OFFSET_MASK;
+
+
+	if (pos >= vgpu->region[i].size || iswrite) {
+		gvt_vgpu_err("invalid op or offset for Intel vgpu OpRegion\n");
+		return -EINVAL;
+	}
+	count = min(count, (size_t)(vgpu->region[i].size - pos));
+	memcpy(buf, base + pos, count);
+
+	return count;
+}
+
+static void intel_vgpu_reg_release_opregion(struct intel_vgpu *vgpu,
+		struct vfio_region *region)
+{
+}
+
+static const struct intel_vgpu_regops intel_vgpu_regops_opregion = {
+	.rw = intel_vgpu_reg_rw_opregion,
+	.release = intel_vgpu_reg_release_opregion,
+};
+
+static int handle_edid_regs(struct intel_vgpu *vgpu,
+			struct vfio_edid_region *region, char *buf,
+			size_t count, u16 offset, bool is_write)
+{
+	struct vfio_region_gfx_edid *regs = &region->vfio_edid_regs;
+	unsigned int data;
+
+	if (offset + count > sizeof(*regs))
+		return -EINVAL;
+
+	if (count != 4)
+		return -EINVAL;
+
+	if (is_write) {
+		data = *((unsigned int *)buf);
+		switch (offset) {
+		case offsetof(struct vfio_region_gfx_edid, link_state):
+			if (data == VFIO_DEVICE_GFX_LINK_STATE_UP) {
+				if (!drm_edid_block_valid(
+					(u8 *)region->edid_blob,
+					0,
+					true,
+					NULL)) {
+					gvt_vgpu_err("invalid EDID blob\n");
+					return -EINVAL;
+				}
+				intel_vgpu_emulate_hotplug(vgpu, true);
+			} else if (data == VFIO_DEVICE_GFX_LINK_STATE_DOWN)
+				intel_vgpu_emulate_hotplug(vgpu, false);
+			else {
+				gvt_vgpu_err("invalid EDID link state %d\n",
+					regs->link_state);
+				return -EINVAL;
+			}
+			regs->link_state = data;
+			break;
+		case offsetof(struct vfio_region_gfx_edid, edid_size):
+			if (data > regs->edid_max_size) {
+				gvt_vgpu_err("EDID size is bigger than %d!\n",
+					regs->edid_max_size);
+				return -EINVAL;
+			}
+			regs->edid_size = data;
+			break;
+		default:
+			/* read-only regs */
+			gvt_vgpu_err("write read-only EDID region at offset %d\n",
+				offset);
+			return -EPERM;
+		}
+	} else {
+		memcpy(buf, (char *)regs + offset, count);
+	}
+
+	return count;
+}
+
+static int handle_edid_blob(struct vfio_edid_region *region, char *buf,
+			size_t count, u16 offset, bool is_write)
+{
+	if (offset + count > region->vfio_edid_regs.edid_size)
+		return -EINVAL;
+
+	if (is_write)
+		memcpy(region->edid_blob + offset, buf, count);
+	else
+		memcpy(buf, region->edid_blob + offset, count);
+
+	return count;
+}
+
+static size_t intel_vgpu_reg_rw_edid(struct intel_vgpu *vgpu, char *buf,
+		size_t count, loff_t *ppos, bool iswrite)
+{
+	int ret;
+	unsigned int i = VFIO_PCI_OFFSET_TO_INDEX(*ppos) -
+			VFIO_PCI_NUM_REGIONS;
+	struct vfio_edid_region *region = vgpu->region[i].data;
+	loff_t pos = *ppos & VFIO_PCI_OFFSET_MASK;
+
+	if (pos < region->vfio_edid_regs.edid_offset) {
+		ret = handle_edid_regs(vgpu, region, buf, count, pos, iswrite);
+	} else {
+		pos -= EDID_BLOB_OFFSET;
+		ret = handle_edid_blob(region, buf, count, pos, iswrite);
+	}
+
+	if (ret < 0)
+		gvt_vgpu_err("failed to access EDID region\n");
+
+	return ret;
+}
+
+static void intel_vgpu_reg_release_edid(struct intel_vgpu *vgpu,
+					struct vfio_region *region)
+{
+	kfree(region->data);
+}
+
+static const struct intel_vgpu_regops intel_vgpu_regops_edid = {
+	.rw = intel_vgpu_reg_rw_edid,
+	.release = intel_vgpu_reg_release_edid,
+};
+
+static int intel_vgpu_register_reg(struct intel_vgpu *vgpu,
+		unsigned int type, unsigned int subtype,
+		const struct intel_vgpu_regops *ops,
+		size_t size, u32 flags, void *data)
+{
+	struct vfio_region *region;
+
+	region = krealloc(vgpu->region,
+			(vgpu->num_regions + 1) * sizeof(*region),
+			GFP_KERNEL);
+	if (!region)
+		return -ENOMEM;
+
+	vgpu->region = region;
+	vgpu->region[vgpu->num_regions].type = type;
+	vgpu->region[vgpu->num_regions].subtype = subtype;
+	vgpu->region[vgpu->num_regions].ops = ops;
+	vgpu->region[vgpu->num_regions].size = size;
+	vgpu->region[vgpu->num_regions].flags = flags;
+	vgpu->region[vgpu->num_regions].data = data;
+	vgpu->num_regions++;
+	return 0;
+}
+
+int intel_gvt_set_opregion(struct intel_vgpu *vgpu)
+{
+	void *base;
+	int ret;
+
+	/* Each vgpu has its own opregion, although VFIO would create another
+	 * one later. This one is used to expose opregion to VFIO. And the
+	 * other one created by VFIO later, is used by guest actually.
+	 */
+	base = vgpu_opregion(vgpu)->va;
+	if (!base)
+		return -ENOMEM;
+
+	if (memcmp(base, OPREGION_SIGNATURE, 16)) {
+		memunmap(base);
+		return -EINVAL;
+	}
+
+	ret = intel_vgpu_register_reg(vgpu,
+			PCI_VENDOR_ID_INTEL | VFIO_REGION_TYPE_PCI_VENDOR_TYPE,
+			VFIO_REGION_SUBTYPE_INTEL_IGD_OPREGION,
+			&intel_vgpu_regops_opregion, OPREGION_SIZE,
+			VFIO_REGION_INFO_FLAG_READ, base);
+
+	return ret;
+}
+
+int intel_gvt_set_edid(struct intel_vgpu *vgpu, int port_num)
+{
+	struct intel_vgpu_port *port = intel_vgpu_port(vgpu, port_num);
+	struct vfio_edid_region *base;
+	int ret;
+
+	base = kzalloc(sizeof(*base), GFP_KERNEL);
+	if (!base)
+		return -ENOMEM;
+
+	/* TODO: Add multi-port and EDID extension block support */
+	base->vfio_edid_regs.edid_offset = EDID_BLOB_OFFSET;
+	base->vfio_edid_regs.edid_max_size = EDID_SIZE;
+	base->vfio_edid_regs.edid_size = EDID_SIZE;
+	base->vfio_edid_regs.max_xres = vgpu_edid_xres(port->id);
+	base->vfio_edid_regs.max_yres = vgpu_edid_yres(port->id);
+	base->edid_blob = port->edid->edid_block;
+
+	ret = intel_vgpu_register_reg(vgpu,
+			VFIO_REGION_TYPE_GFX,
+			VFIO_REGION_SUBTYPE_GFX_EDID,
+			&intel_vgpu_regops_edid, EDID_SIZE,
+			VFIO_REGION_INFO_FLAG_READ |
+			VFIO_REGION_INFO_FLAG_WRITE |
+			VFIO_REGION_INFO_FLAG_CAPS, base);
+
+	return ret;
+}
+
+static void intel_vgpu_dma_unmap(struct vfio_device *vfio_dev, u64 iova,
+				 u64 length)
+{
+	struct intel_vgpu *vgpu = vfio_dev_to_vgpu(vfio_dev);
+	struct gvt_dma *entry;
+	u64 iov_pfn = iova >> PAGE_SHIFT;
+	u64 end_iov_pfn = iov_pfn + length / PAGE_SIZE;
+
+	mutex_lock(&vgpu->cache_lock);
+	for (; iov_pfn < end_iov_pfn; iov_pfn++) {
+		entry = __gvt_cache_find_gfn(vgpu, iov_pfn);
+		if (!entry)
+			continue;
+
+		gvt_dma_unmap_page(vgpu, entry->gfn, entry->dma_addr,
+				   entry->size);
+		__gvt_cache_remove_entry(vgpu, entry);
+	}
+	mutex_unlock(&vgpu->cache_lock);
+}
+
+static bool __kvmgt_vgpu_exist(struct intel_vgpu *vgpu)
+{
+	struct intel_vgpu *itr;
+	int id;
+	bool ret = false;
+
+	mutex_lock(&vgpu->gvt->lock);
+	for_each_active_vgpu(vgpu->gvt, itr, id) {
+		if (!itr->attached)
+			continue;
+
+		if (vgpu->vfio_device.kvm == itr->vfio_device.kvm) {
+			ret = true;
+			goto out;
+		}
+	}
+out:
+	mutex_unlock(&vgpu->gvt->lock);
+	return ret;
+}
+
+static int intel_vgpu_open_device(struct vfio_device *vfio_dev)
+{
+	struct intel_vgpu *vgpu = vfio_dev_to_vgpu(vfio_dev);
+
+	if (vgpu->attached)
+		return -EEXIST;
+
+	if (!vgpu->vfio_device.kvm ||
+	    vgpu->vfio_device.kvm->mm != current->mm) {
+		gvt_vgpu_err("KVM is required to use Intel vGPU\n");
+		return -ESRCH;
+	}
+
+	if (__kvmgt_vgpu_exist(vgpu))
+		return -EEXIST;
+
+	vgpu->attached = true;
+
+	kvmgt_protect_table_init(vgpu);
+	gvt_cache_init(vgpu);
+
+	vgpu->track_node.track_write = kvmgt_page_track_write;
+	vgpu->track_node.track_flush_slot = kvmgt_page_track_flush_slot;
+	kvm_get_kvm(vgpu->vfio_device.kvm);
+	kvm_page_track_register_notifier(vgpu->vfio_device.kvm,
+					 &vgpu->track_node);
+
+	debugfs_create_ulong(KVMGT_DEBUGFS_FILENAME, 0444, vgpu->debugfs,
+			     &vgpu->nr_cache_entries);
+
+	intel_gvt_activate_vgpu(vgpu);
+
+	atomic_set(&vgpu->released, 0);
+	return 0;
+}
+
+static void intel_vgpu_release_msi_eventfd_ctx(struct intel_vgpu *vgpu)
+{
+	struct eventfd_ctx *trigger;
+
+	trigger = vgpu->msi_trigger;
+	if (trigger) {
+		eventfd_ctx_put(trigger);
+		vgpu->msi_trigger = NULL;
+	}
+}
+
+static void intel_vgpu_close_device(struct vfio_device *vfio_dev)
+{
+	struct intel_vgpu *vgpu = vfio_dev_to_vgpu(vfio_dev);
+
+	if (!vgpu->attached)
+		return;
+
+	if (atomic_cmpxchg(&vgpu->released, 0, 1))
+		return;
+
+	intel_gvt_release_vgpu(vgpu);
+
+	debugfs_remove(debugfs_lookup(KVMGT_DEBUGFS_FILENAME, vgpu->debugfs));
+
+	kvm_page_track_unregister_notifier(vgpu->vfio_device.kvm,
+					   &vgpu->track_node);
+	kvm_put_kvm(vgpu->vfio_device.kvm);
+
+	kvmgt_protect_table_destroy(vgpu);
+	gvt_cache_destroy(vgpu);
+
+	intel_vgpu_release_msi_eventfd_ctx(vgpu);
+
+	vgpu->attached = false;
+}
+
+static u64 intel_vgpu_get_bar_addr(struct intel_vgpu *vgpu, int bar)
+{
+	u32 start_lo, start_hi;
+	u32 mem_type;
+
+	start_lo = (*(u32 *)(vgpu->cfg_space.virtual_cfg_space + bar)) &
+			PCI_BASE_ADDRESS_MEM_MASK;
+	mem_type = (*(u32 *)(vgpu->cfg_space.virtual_cfg_space + bar)) &
+			PCI_BASE_ADDRESS_MEM_TYPE_MASK;
+
+	switch (mem_type) {
+	case PCI_BASE_ADDRESS_MEM_TYPE_64:
+		start_hi = (*(u32 *)(vgpu->cfg_space.virtual_cfg_space
+						+ bar + 4));
+		break;
+	case PCI_BASE_ADDRESS_MEM_TYPE_32:
+	case PCI_BASE_ADDRESS_MEM_TYPE_1M:
+		/* 1M mem BAR treated as 32-bit BAR */
+	default:
+		/* mem unknown type treated as 32-bit BAR */
+		start_hi = 0;
+		break;
+	}
+
+	return ((u64)start_hi << 32) | start_lo;
+}
+
+static int intel_vgpu_bar_rw(struct intel_vgpu *vgpu, int bar, u64 off,
+			     void *buf, unsigned int count, bool is_write)
+{
+	u64 bar_start = intel_vgpu_get_bar_addr(vgpu, bar);
+	int ret;
+
+	if (is_write)
+		ret = intel_vgpu_emulate_mmio_write(vgpu,
+					bar_start + off, buf, count);
+	else
+		ret = intel_vgpu_emulate_mmio_read(vgpu,
+					bar_start + off, buf, count);
+	return ret;
+}
+
+static inline bool intel_vgpu_in_aperture(struct intel_vgpu *vgpu, u64 off)
+{
+	return off >= vgpu_aperture_offset(vgpu) &&
+	       off < vgpu_aperture_offset(vgpu) + vgpu_aperture_sz(vgpu);
+}
+
+static int intel_vgpu_aperture_rw(struct intel_vgpu *vgpu, u64 off,
+		void *buf, unsigned long count, bool is_write)
+{
+	void __iomem *aperture_va;
+
+	if (!intel_vgpu_in_aperture(vgpu, off) ||
+	    !intel_vgpu_in_aperture(vgpu, off + count)) {
+		gvt_vgpu_err("Invalid aperture offset %llu\n", off);
+		return -EINVAL;
+	}
+
+	aperture_va = io_mapping_map_wc(&vgpu->gvt->gt->ggtt->iomap,
+					ALIGN_DOWN(off, PAGE_SIZE),
+					count + offset_in_page(off));
+	if (!aperture_va)
+		return -EIO;
+
+	if (is_write)
+		memcpy_toio(aperture_va + offset_in_page(off), buf, count);
+	else
+		memcpy_fromio(buf, aperture_va + offset_in_page(off), count);
+
+	io_mapping_unmap(aperture_va);
+
+	return 0;
+}
+
+static ssize_t intel_vgpu_rw(struct intel_vgpu *vgpu, char *buf,
+			size_t count, loff_t *ppos, bool is_write)
+{
+	unsigned int index = VFIO_PCI_OFFSET_TO_INDEX(*ppos);
+	u64 pos = *ppos & VFIO_PCI_OFFSET_MASK;
+	int ret = -EINVAL;
+
+
+	if (index >= VFIO_PCI_NUM_REGIONS + vgpu->num_regions) {
+		gvt_vgpu_err("invalid index: %u\n", index);
+		return -EINVAL;
+	}
+
+	switch (index) {
+	case VFIO_PCI_CONFIG_REGION_INDEX:
+		if (is_write)
+			ret = intel_vgpu_emulate_cfg_write(vgpu, pos,
+						buf, count);
+		else
+			ret = intel_vgpu_emulate_cfg_read(vgpu, pos,
+						buf, count);
+		break;
+	case VFIO_PCI_BAR0_REGION_INDEX:
+		ret = intel_vgpu_bar_rw(vgpu, PCI_BASE_ADDRESS_0, pos,
+					buf, count, is_write);
+		break;
+	case VFIO_PCI_BAR2_REGION_INDEX:
+		ret = intel_vgpu_aperture_rw(vgpu, pos, buf, count, is_write);
+		break;
+	case VFIO_PCI_BAR1_REGION_INDEX:
+	case VFIO_PCI_BAR3_REGION_INDEX:
+	case VFIO_PCI_BAR4_REGION_INDEX:
+	case VFIO_PCI_BAR5_REGION_INDEX:
+	case VFIO_PCI_VGA_REGION_INDEX:
+	case VFIO_PCI_ROM_REGION_INDEX:
+		break;
+	default:
+		if (index >= VFIO_PCI_NUM_REGIONS + vgpu->num_regions)
+			return -EINVAL;
+
+		index -= VFIO_PCI_NUM_REGIONS;
+		return vgpu->region[index].ops->rw(vgpu, buf, count,
+				ppos, is_write);
+	}
+
+	return ret == 0 ? count : ret;
+}
+
+static bool gtt_entry(struct intel_vgpu *vgpu, loff_t *ppos)
+{
+	unsigned int index = VFIO_PCI_OFFSET_TO_INDEX(*ppos);
+	struct intel_gvt *gvt = vgpu->gvt;
+	int offset;
+
+	/* Only allow MMIO GGTT entry access */
+	if (index != PCI_BASE_ADDRESS_0)
+		return false;
+
+	offset = (u64)(*ppos & VFIO_PCI_OFFSET_MASK) -
+		intel_vgpu_get_bar_gpa(vgpu, PCI_BASE_ADDRESS_0);
+
+	return (offset >= gvt->device_info.gtt_start_offset &&
+		offset < gvt->device_info.gtt_start_offset + gvt_ggtt_sz(gvt)) ?
+			true : false;
+}
+
+static ssize_t intel_vgpu_read(struct vfio_device *vfio_dev, char __user *buf,
+			size_t count, loff_t *ppos)
+{
+	struct intel_vgpu *vgpu = vfio_dev_to_vgpu(vfio_dev);
+	unsigned int done = 0;
+	int ret;
+
+	while (count) {
+		size_t filled;
+
+		/* Only support GGTT entry 8 bytes read */
+		if (count >= 8 && !(*ppos % 8) &&
+			gtt_entry(vgpu, ppos)) {
+			u64 val;
+
+			ret = intel_vgpu_rw(vgpu, (char *)&val, sizeof(val),
+					ppos, false);
+			if (ret <= 0)
+				goto read_err;
+
+			if (copy_to_user(buf, &val, sizeof(val)))
+				goto read_err;
+
+			filled = 8;
+		} else if (count >= 4 && !(*ppos % 4)) {
+			u32 val;
+
+			ret = intel_vgpu_rw(vgpu, (char *)&val, sizeof(val),
+					ppos, false);
+			if (ret <= 0)
+				goto read_err;
+
+			if (copy_to_user(buf, &val, sizeof(val)))
+				goto read_err;
+
+			filled = 4;
+		} else if (count >= 2 && !(*ppos % 2)) {
+			u16 val;
+
+			ret = intel_vgpu_rw(vgpu, (char *)&val, sizeof(val),
+					ppos, false);
+			if (ret <= 0)
+				goto read_err;
+
+			if (copy_to_user(buf, &val, sizeof(val)))
+				goto read_err;
+
+			filled = 2;
+		} else {
+			u8 val;
+
+			ret = intel_vgpu_rw(vgpu, &val, sizeof(val), ppos,
+					false);
+			if (ret <= 0)
+				goto read_err;
+
+			if (copy_to_user(buf, &val, sizeof(val)))
+				goto read_err;
+
+			filled = 1;
+		}
+
+		count -= filled;
+		done += filled;
+		*ppos += filled;
+		buf += filled;
+	}
+
+	return done;
+
+read_err:
+	return -EFAULT;
+}
+
+static ssize_t intel_vgpu_write(struct vfio_device *vfio_dev,
+				const char __user *buf,
+				size_t count, loff_t *ppos)
+{
+	struct intel_vgpu *vgpu = vfio_dev_to_vgpu(vfio_dev);
+	unsigned int done = 0;
+	int ret;
+
+	while (count) {
+		size_t filled;
+
+		/* Only support GGTT entry 8 bytes write */
+		if (count >= 8 && !(*ppos % 8) &&
+			gtt_entry(vgpu, ppos)) {
+			u64 val;
+
+			if (copy_from_user(&val, buf, sizeof(val)))
+				goto write_err;
+
+			ret = intel_vgpu_rw(vgpu, (char *)&val, sizeof(val),
+					ppos, true);
+			if (ret <= 0)
+				goto write_err;
+
+			filled = 8;
+		} else if (count >= 4 && !(*ppos % 4)) {
+			u32 val;
+
+			if (copy_from_user(&val, buf, sizeof(val)))
+				goto write_err;
+
+			ret = intel_vgpu_rw(vgpu, (char *)&val, sizeof(val),
+					ppos, true);
+			if (ret <= 0)
+				goto write_err;
+
+			filled = 4;
+		} else if (count >= 2 && !(*ppos % 2)) {
+			u16 val;
+
+			if (copy_from_user(&val, buf, sizeof(val)))
+				goto write_err;
+
+			ret = intel_vgpu_rw(vgpu, (char *)&val,
+					sizeof(val), ppos, true);
+			if (ret <= 0)
+				goto write_err;
+
+			filled = 2;
+		} else {
+			u8 val;
+
+			if (copy_from_user(&val, buf, sizeof(val)))
+				goto write_err;
+
+			ret = intel_vgpu_rw(vgpu, &val, sizeof(val),
+					ppos, true);
+			if (ret <= 0)
+				goto write_err;
+
+			filled = 1;
+		}
+
+		count -= filled;
+		done += filled;
+		*ppos += filled;
+		buf += filled;
+	}
+
+	return done;
+write_err:
+	return -EFAULT;
+}
+
+static int intel_vgpu_mmap(struct vfio_device *vfio_dev,
+		struct vm_area_struct *vma)
+{
+	struct intel_vgpu *vgpu = vfio_dev_to_vgpu(vfio_dev);
+	unsigned int index;
+	u64 virtaddr;
+	unsigned long req_size, pgoff, req_start;
+	pgprot_t pg_prot;
+
+	index = vma->vm_pgoff >> (VFIO_PCI_OFFSET_SHIFT - PAGE_SHIFT);
+	if (index >= VFIO_PCI_ROM_REGION_INDEX)
+		return -EINVAL;
+
+	if (vma->vm_end < vma->vm_start)
+		return -EINVAL;
+	if ((vma->vm_flags & VM_SHARED) == 0)
+		return -EINVAL;
+	if (index != VFIO_PCI_BAR2_REGION_INDEX)
+		return -EINVAL;
+
+	pg_prot = vma->vm_page_prot;
+	virtaddr = vma->vm_start;
+	req_size = vma->vm_end - vma->vm_start;
+	pgoff = vma->vm_pgoff &
+		((1U << (VFIO_PCI_OFFSET_SHIFT - PAGE_SHIFT)) - 1);
+	req_start = pgoff << PAGE_SHIFT;
+
+	if (!intel_vgpu_in_aperture(vgpu, req_start))
+		return -EINVAL;
+	if (req_start + req_size >
+	    vgpu_aperture_offset(vgpu) + vgpu_aperture_sz(vgpu))
+		return -EINVAL;
+
+	pgoff = (gvt_aperture_pa_base(vgpu->gvt) >> PAGE_SHIFT) + pgoff;
+
+	return remap_pfn_range(vma, virtaddr, pgoff, req_size, pg_prot);
+}
+
+static int intel_vgpu_get_irq_count(struct intel_vgpu *vgpu, int type)
+{
+	if (type == VFIO_PCI_INTX_IRQ_INDEX || type == VFIO_PCI_MSI_IRQ_INDEX)
+		return 1;
+
+	return 0;
+}
+
+static int intel_vgpu_set_intx_mask(struct intel_vgpu *vgpu,
+			unsigned int index, unsigned int start,
+			unsigned int count, u32 flags,
+			void *data)
+{
+	return 0;
+}
+
+static int intel_vgpu_set_intx_unmask(struct intel_vgpu *vgpu,
+			unsigned int index, unsigned int start,
+			unsigned int count, u32 flags, void *data)
+{
+	return 0;
+}
+
+static int intel_vgpu_set_intx_trigger(struct intel_vgpu *vgpu,
+		unsigned int index, unsigned int start, unsigned int count,
+		u32 flags, void *data)
+{
+	return 0;
+}
+
+static int intel_vgpu_set_msi_trigger(struct intel_vgpu *vgpu,
+		unsigned int index, unsigned int start, unsigned int count,
+		u32 flags, void *data)
+{
+	struct eventfd_ctx *trigger;
+
+	if (flags & VFIO_IRQ_SET_DATA_EVENTFD) {
+		int fd = *(int *)data;
+
+		trigger = eventfd_ctx_fdget(fd);
+		if (IS_ERR(trigger)) {
+			gvt_vgpu_err("eventfd_ctx_fdget failed\n");
+			return PTR_ERR(trigger);
+		}
+		vgpu->msi_trigger = trigger;
+	} else if ((flags & VFIO_IRQ_SET_DATA_NONE) && !count)
+		intel_vgpu_release_msi_eventfd_ctx(vgpu);
+
+	return 0;
+}
+
+static int intel_vgpu_set_irqs(struct intel_vgpu *vgpu, u32 flags,
+		unsigned int index, unsigned int start, unsigned int count,
+		void *data)
+{
+	int (*func)(struct intel_vgpu *vgpu, unsigned int index,
+			unsigned int start, unsigned int count, u32 flags,
+			void *data) = NULL;
+
+	switch (index) {
+	case VFIO_PCI_INTX_IRQ_INDEX:
+		switch (flags & VFIO_IRQ_SET_ACTION_TYPE_MASK) {
+		case VFIO_IRQ_SET_ACTION_MASK:
+			func = intel_vgpu_set_intx_mask;
+			break;
+		case VFIO_IRQ_SET_ACTION_UNMASK:
+			func = intel_vgpu_set_intx_unmask;
+			break;
+		case VFIO_IRQ_SET_ACTION_TRIGGER:
+			func = intel_vgpu_set_intx_trigger;
+			break;
+		}
+		break;
+	case VFIO_PCI_MSI_IRQ_INDEX:
+		switch (flags & VFIO_IRQ_SET_ACTION_TYPE_MASK) {
+		case VFIO_IRQ_SET_ACTION_MASK:
+		case VFIO_IRQ_SET_ACTION_UNMASK:
+			/* XXX Need masking support exported */
+			break;
+		case VFIO_IRQ_SET_ACTION_TRIGGER:
+			func = intel_vgpu_set_msi_trigger;
+			break;
+		}
+		break;
+	}
+
+	if (!func)
+		return -ENOTTY;
+
+	return func(vgpu, index, start, count, flags, data);
+}
+
+static long intel_vgpu_ioctl(struct vfio_device *vfio_dev, unsigned int cmd,
+			     unsigned long arg)
+{
+	struct intel_vgpu *vgpu = vfio_dev_to_vgpu(vfio_dev);
+	unsigned long minsz;
+
+	gvt_dbg_core("vgpu%d ioctl, cmd: %d\n", vgpu->id, cmd);
+
+	if (cmd == VFIO_DEVICE_GET_INFO) {
+		struct vfio_device_info info;
+
+		minsz = offsetofend(struct vfio_device_info, num_irqs);
+
+		if (copy_from_user(&info, (void __user *)arg, minsz))
+			return -EFAULT;
+
+		if (info.argsz < minsz)
+			return -EINVAL;
+
+		info.flags = VFIO_DEVICE_FLAGS_PCI;
+		info.flags |= VFIO_DEVICE_FLAGS_RESET;
+		info.num_regions = VFIO_PCI_NUM_REGIONS +
+				vgpu->num_regions;
+		info.num_irqs = VFIO_PCI_NUM_IRQS;
+
+		return copy_to_user((void __user *)arg, &info, minsz) ?
+			-EFAULT : 0;
+
+	} else if (cmd == VFIO_DEVICE_GET_REGION_INFO) {
+		struct vfio_region_info info;
+		struct vfio_info_cap caps = { .buf = NULL, .size = 0 };
+		unsigned int i;
+		int ret;
+		struct vfio_region_info_cap_sparse_mmap *sparse = NULL;
+		int nr_areas = 1;
+		int cap_type_id;
+
+		minsz = offsetofend(struct vfio_region_info, offset);
+
+		if (copy_from_user(&info, (void __user *)arg, minsz))
+			return -EFAULT;
+
+		if (info.argsz < minsz)
+			return -EINVAL;
+
+		switch (info.index) {
+		case VFIO_PCI_CONFIG_REGION_INDEX:
+			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
+			info.size = vgpu->gvt->device_info.cfg_space_size;
+			info.flags = VFIO_REGION_INFO_FLAG_READ |
+				     VFIO_REGION_INFO_FLAG_WRITE;
+			break;
+		case VFIO_PCI_BAR0_REGION_INDEX:
+			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
+			info.size = vgpu->cfg_space.bar[info.index].size;
+			if (!info.size) {
+				info.flags = 0;
+				break;
+			}
+
+			info.flags = VFIO_REGION_INFO_FLAG_READ |
+				     VFIO_REGION_INFO_FLAG_WRITE;
+			break;
+		case VFIO_PCI_BAR1_REGION_INDEX:
+			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
+			info.size = 0;
+			info.flags = 0;
+			break;
+		case VFIO_PCI_BAR2_REGION_INDEX:
+			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
+			info.flags = VFIO_REGION_INFO_FLAG_CAPS |
+					VFIO_REGION_INFO_FLAG_MMAP |
+					VFIO_REGION_INFO_FLAG_READ |
+					VFIO_REGION_INFO_FLAG_WRITE;
+			info.size = gvt_aperture_sz(vgpu->gvt);
+
+			sparse = kzalloc(struct_size(sparse, areas, nr_areas),
+					 GFP_KERNEL);
+			if (!sparse)
+				return -ENOMEM;
+
+			sparse->header.id = VFIO_REGION_INFO_CAP_SPARSE_MMAP;
+			sparse->header.version = 1;
+			sparse->nr_areas = nr_areas;
+			cap_type_id = VFIO_REGION_INFO_CAP_SPARSE_MMAP;
+			sparse->areas[0].offset =
+					PAGE_ALIGN(vgpu_aperture_offset(vgpu));
+			sparse->areas[0].size = vgpu_aperture_sz(vgpu);
+			break;
+
+		case VFIO_PCI_BAR3_REGION_INDEX ... VFIO_PCI_BAR5_REGION_INDEX:
+			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
+			info.size = 0;
+			info.flags = 0;
+
+			gvt_dbg_core("get region info bar:%d\n", info.index);
+			break;
+
+		case VFIO_PCI_ROM_REGION_INDEX:
+		case VFIO_PCI_VGA_REGION_INDEX:
+			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
+			info.size = 0;
+			info.flags = 0;
+
+			gvt_dbg_core("get region info index:%d\n", info.index);
+			break;
+		default:
+			{
+				struct vfio_region_info_cap_type cap_type = {
+					.header.id = VFIO_REGION_INFO_CAP_TYPE,
+					.header.version = 1 };
+
+				if (info.index >= VFIO_PCI_NUM_REGIONS +
+						vgpu->num_regions)
+					return -EINVAL;
+				info.index =
+					array_index_nospec(info.index,
+							VFIO_PCI_NUM_REGIONS +
+							vgpu->num_regions);
+
+				i = info.index - VFIO_PCI_NUM_REGIONS;
+
+				info.offset =
+					VFIO_PCI_INDEX_TO_OFFSET(info.index);
+				info.size = vgpu->region[i].size;
+				info.flags = vgpu->region[i].flags;
+
+				cap_type.type = vgpu->region[i].type;
+				cap_type.subtype = vgpu->region[i].subtype;
+
+				ret = vfio_info_add_capability(&caps,
+							&cap_type.header,
+							sizeof(cap_type));
+				if (ret)
+					return ret;
+			}
+		}
+
+		if ((info.flags & VFIO_REGION_INFO_FLAG_CAPS) && sparse) {
+			switch (cap_type_id) {
+			case VFIO_REGION_INFO_CAP_SPARSE_MMAP:
+				ret = vfio_info_add_capability(&caps,
+					&sparse->header,
+					struct_size(sparse, areas,
+						    sparse->nr_areas));
+				if (ret) {
+					kfree(sparse);
+					return ret;
+				}
+				break;
+			default:
+				kfree(sparse);
+				return -EINVAL;
+			}
+		}
+
+		if (caps.size) {
+			info.flags |= VFIO_REGION_INFO_FLAG_CAPS;
+			if (info.argsz < sizeof(info) + caps.size) {
+				info.argsz = sizeof(info) + caps.size;
+				info.cap_offset = 0;
+			} else {
+				vfio_info_cap_shift(&caps, sizeof(info));
+				if (copy_to_user((void __user *)arg +
+						  sizeof(info), caps.buf,
+						  caps.size)) {
+					kfree(caps.buf);
+					kfree(sparse);
+					return -EFAULT;
+				}
+				info.cap_offset = sizeof(info);
+			}
+
+			kfree(caps.buf);
+		}
+
+		kfree(sparse);
+		return copy_to_user((void __user *)arg, &info, minsz) ?
+			-EFAULT : 0;
+	} else if (cmd == VFIO_DEVICE_GET_IRQ_INFO) {
+		struct vfio_irq_info info;
+
+		minsz = offsetofend(struct vfio_irq_info, count);
+
+		if (copy_from_user(&info, (void __user *)arg, minsz))
+			return -EFAULT;
+
+		if (info.argsz < minsz || info.index >= VFIO_PCI_NUM_IRQS)
+			return -EINVAL;
+
+		switch (info.index) {
+		case VFIO_PCI_INTX_IRQ_INDEX:
+		case VFIO_PCI_MSI_IRQ_INDEX:
+			break;
+		default:
+			return -EINVAL;
+		}
+
+		info.flags = VFIO_IRQ_INFO_EVENTFD;
+
+		info.count = intel_vgpu_get_irq_count(vgpu, info.index);
+
+		if (info.index == VFIO_PCI_INTX_IRQ_INDEX)
+			info.flags |= (VFIO_IRQ_INFO_MASKABLE |
+				       VFIO_IRQ_INFO_AUTOMASKED);
+		else
+			info.flags |= VFIO_IRQ_INFO_NORESIZE;
+
+		return copy_to_user((void __user *)arg, &info, minsz) ?
+			-EFAULT : 0;
+	} else if (cmd == VFIO_DEVICE_SET_IRQS) {
+		struct vfio_irq_set hdr;
+		u8 *data = NULL;
+		int ret = 0;
+		size_t data_size = 0;
+
+		minsz = offsetofend(struct vfio_irq_set, count);
+
+		if (copy_from_user(&hdr, (void __user *)arg, minsz))
+			return -EFAULT;
+
+		if (!(hdr.flags & VFIO_IRQ_SET_DATA_NONE)) {
+			int max = intel_vgpu_get_irq_count(vgpu, hdr.index);
+
+			ret = vfio_set_irqs_validate_and_prepare(&hdr, max,
+						VFIO_PCI_NUM_IRQS, &data_size);
+			if (ret) {
+				gvt_vgpu_err("intel:vfio_set_irqs_validate_and_prepare failed\n");
+				return -EINVAL;
+			}
+			if (data_size) {
+				data = memdup_user((void __user *)(arg + minsz),
+						   data_size);
+				if (IS_ERR(data))
+					return PTR_ERR(data);
+			}
+		}
+
+		ret = intel_vgpu_set_irqs(vgpu, hdr.flags, hdr.index,
+					hdr.start, hdr.count, data);
+		kfree(data);
+
+		return ret;
+	} else if (cmd == VFIO_DEVICE_RESET) {
+		intel_gvt_reset_vgpu(vgpu);
+		return 0;
+	} else if (cmd == VFIO_DEVICE_QUERY_GFX_PLANE) {
+		struct vfio_device_gfx_plane_info dmabuf;
+		int ret = 0;
+
+		minsz = offsetofend(struct vfio_device_gfx_plane_info,
+				    dmabuf_id);
+		if (copy_from_user(&dmabuf, (void __user *)arg, minsz))
+			return -EFAULT;
+		if (dmabuf.argsz < minsz)
+			return -EINVAL;
+
+		ret = intel_vgpu_query_plane(vgpu, &dmabuf);
+		if (ret != 0)
+			return ret;
+
+		return copy_to_user((void __user *)arg, &dmabuf, minsz) ?
+								-EFAULT : 0;
+	} else if (cmd == VFIO_DEVICE_GET_GFX_DMABUF) {
+		__u32 dmabuf_id;
+
+		if (get_user(dmabuf_id, (__u32 __user *)arg))
+			return -EFAULT;
+		return intel_vgpu_get_dmabuf(vgpu, dmabuf_id);
+	}
+
+	return -ENOTTY;
+}
+
+static ssize_t
+vgpu_id_show(struct device *dev, struct device_attribute *attr,
+	     char *buf)
+{
+	struct intel_vgpu *vgpu = dev_get_drvdata(dev);
+
+	return sprintf(buf, "%d\n", vgpu->id);
+}
+
+static DEVICE_ATTR_RO(vgpu_id);
+
+static struct attribute *intel_vgpu_attrs[] = {
+	&dev_attr_vgpu_id.attr,
+	NULL
+};
+
+static const struct attribute_group intel_vgpu_group = {
+	.name = "intel_vgpu",
+	.attrs = intel_vgpu_attrs,
+};
+
+static const struct attribute_group *intel_vgpu_groups[] = {
+	&intel_vgpu_group,
+	NULL,
+};
+
+static int intel_vgpu_init_dev(struct vfio_device *vfio_dev)
+{
+	struct mdev_device *mdev = to_mdev_device(vfio_dev->dev);
+	struct intel_vgpu *vgpu = vfio_dev_to_vgpu(vfio_dev);
+	struct intel_vgpu_type *type =
+		container_of(mdev->type, struct intel_vgpu_type, type);
+
+	vgpu->gvt = kdev_to_i915(mdev->type->parent->dev)->gvt;
+	return intel_gvt_create_vgpu(vgpu, type->conf);
+}
+
+static void intel_vgpu_release_dev(struct vfio_device *vfio_dev)
+{
+	struct intel_vgpu *vgpu = vfio_dev_to_vgpu(vfio_dev);
+
+	intel_gvt_destroy_vgpu(vgpu);
+	vfio_free_device(vfio_dev);
+}
+
+static const struct vfio_device_ops intel_vgpu_dev_ops = {
+	.init		= intel_vgpu_init_dev,
+	.release	= intel_vgpu_release_dev,
+	.open_device	= intel_vgpu_open_device,
+	.close_device	= intel_vgpu_close_device,
+	.read		= intel_vgpu_read,
+	.write		= intel_vgpu_write,
+	.mmap		= intel_vgpu_mmap,
+	.ioctl		= intel_vgpu_ioctl,
+	.dma_unmap	= intel_vgpu_dma_unmap,
+};
+
+static int intel_vgpu_probe(struct mdev_device *mdev)
+{
+	struct intel_vgpu *vgpu;
+	int ret;
+
+	vgpu = vfio_alloc_device(intel_vgpu, vfio_device, &mdev->dev,
+				 &intel_vgpu_dev_ops);
+	if (IS_ERR(vgpu)) {
+		gvt_err("failed to create intel vgpu: %ld\n", PTR_ERR(vgpu));
+		return PTR_ERR(vgpu);
+	}
+
+	dev_set_drvdata(&mdev->dev, vgpu);
+	ret = vfio_register_emulated_iommu_dev(&vgpu->vfio_device);
+	if (ret)
+		goto out_put_vdev;
+
+	gvt_dbg_core("intel_vgpu_create succeeded for mdev: %s\n",
+		     dev_name(mdev_dev(mdev)));
+	return 0;
+
+out_put_vdev:
+	vfio_put_device(&vgpu->vfio_device);
+	return ret;
+}
+
+static void intel_vgpu_remove(struct mdev_device *mdev)
+{
+	struct intel_vgpu *vgpu = dev_get_drvdata(&mdev->dev);
+
+	if (WARN_ON_ONCE(vgpu->attached))
+		return;
+
+	vfio_unregister_group_dev(&vgpu->vfio_device);
+	vfio_put_device(&vgpu->vfio_device);
+}
+
+static unsigned int intel_vgpu_get_available(struct mdev_type *mtype)
+{
+	struct intel_vgpu_type *type =
+		container_of(mtype, struct intel_vgpu_type, type);
+	struct intel_gvt *gvt = kdev_to_i915(mtype->parent->dev)->gvt;
+	unsigned int low_gm_avail, high_gm_avail, fence_avail;
+
+	mutex_lock(&gvt->lock);
+	low_gm_avail = gvt_aperture_sz(gvt) - HOST_LOW_GM_SIZE -
+		gvt->gm.vgpu_allocated_low_gm_size;
+	high_gm_avail = gvt_hidden_sz(gvt) - HOST_HIGH_GM_SIZE -
+		gvt->gm.vgpu_allocated_high_gm_size;
+	fence_avail = gvt_fence_sz(gvt) - HOST_FENCE -
+		gvt->fence.vgpu_allocated_fence_num;
+	mutex_unlock(&gvt->lock);
+
+	return min3(low_gm_avail / type->conf->low_mm,
+		    high_gm_avail / type->conf->high_mm,
+		    fence_avail / type->conf->fence);
+}
+
+static struct mdev_driver intel_vgpu_mdev_driver = {
+	.device_api	= VFIO_DEVICE_API_PCI_STRING,
+	.driver = {
+		.name		= "intel_vgpu_mdev",
+		.owner		= THIS_MODULE,
+		.dev_groups	= intel_vgpu_groups,
+	},
+	.probe			= intel_vgpu_probe,
+	.remove			= intel_vgpu_remove,
+	.get_available		= intel_vgpu_get_available,
+	.show_description	= intel_vgpu_show_description,
+};
+
+int intel_gvt_page_track_add(struct intel_vgpu *info, u64 gfn)
+{
+	struct kvm *kvm = info->vfio_device.kvm;
+	struct kvm_memory_slot *slot;
+	int idx;
+
+	if (!info->attached)
+		return -ESRCH;
+
+	idx = srcu_read_lock(&kvm->srcu);
+	slot = gfn_to_memslot(kvm, gfn);
+	if (!slot) {
+		srcu_read_unlock(&kvm->srcu, idx);
+		return -EINVAL;
+	}
+
+	write_lock(&kvm->mmu_lock);
+
+	if (kvmgt_gfn_is_write_protected(info, gfn))
+		goto out;
+
+	kvm_slot_page_track_add_page(kvm, slot, gfn, KVM_PAGE_TRACK_WRITE);
+	kvmgt_protect_table_add(info, gfn);
+
+out:
+	write_unlock(&kvm->mmu_lock);
+	srcu_read_unlock(&kvm->srcu, idx);
+	return 0;
+}
+
+int intel_gvt_page_track_remove(struct intel_vgpu *info, u64 gfn)
+{
+	struct kvm *kvm = info->vfio_device.kvm;
+	struct kvm_memory_slot *slot;
+	int idx;
+
+	if (!info->attached)
+		return 0;
+
+	idx = srcu_read_lock(&kvm->srcu);
+	slot = gfn_to_memslot(kvm, gfn);
+	if (!slot) {
+		srcu_read_unlock(&kvm->srcu, idx);
+		return -EINVAL;
+	}
+
+	write_lock(&kvm->mmu_lock);
+
+	if (!kvmgt_gfn_is_write_protected(info, gfn))
+		goto out;
+
+	kvm_slot_page_track_remove_page(kvm, slot, gfn, KVM_PAGE_TRACK_WRITE);
+	kvmgt_protect_table_del(info, gfn);
+
+out:
+	write_unlock(&kvm->mmu_lock);
+	srcu_read_unlock(&kvm->srcu, idx);
+	return 0;
+}
+
+static void kvmgt_page_track_write(struct kvm_vcpu *vcpu, gpa_t gpa,
+		const u8 *val, int len,
+		struct kvm_page_track_notifier_node *node)
+{
+	struct intel_vgpu *info =
+		container_of(node, struct intel_vgpu, track_node);
+
+	if (kvmgt_gfn_is_write_protected(info, gpa_to_gfn(gpa)))
+		intel_vgpu_page_track_handler(info, gpa,
+						     (void *)val, len);
+}
+
+static void kvmgt_page_track_flush_slot(struct kvm *kvm,
+		struct kvm_memory_slot *slot,
+		struct kvm_page_track_notifier_node *node)
+{
+	int i;
+	gfn_t gfn;
+	struct intel_vgpu *info =
+		container_of(node, struct intel_vgpu, track_node);
+
+	write_lock(&kvm->mmu_lock);
+	for (i = 0; i < slot->npages; i++) {
+		gfn = slot->base_gfn + i;
+		if (kvmgt_gfn_is_write_protected(info, gfn)) {
+			kvm_slot_page_track_remove_page(kvm, slot, gfn,
+						KVM_PAGE_TRACK_WRITE);
+			kvmgt_protect_table_del(info, gfn);
+		}
+	}
+	write_unlock(&kvm->mmu_lock);
+}
+
+void intel_vgpu_detach_regions(struct intel_vgpu *vgpu)
+{
+	int i;
+
+	if (!vgpu->region)
+		return;
+
+	for (i = 0; i < vgpu->num_regions; i++)
+		if (vgpu->region[i].ops->release)
+			vgpu->region[i].ops->release(vgpu,
+					&vgpu->region[i]);
+	vgpu->num_regions = 0;
+	kfree(vgpu->region);
+	vgpu->region = NULL;
+}
+
+int intel_gvt_dma_map_guest_page(struct intel_vgpu *vgpu, unsigned long gfn,
+		unsigned long size, dma_addr_t *dma_addr)
+{
+	struct gvt_dma *entry;
+	int ret;
+
+	if (!vgpu->attached)
+		return -EINVAL;
+
+	mutex_lock(&vgpu->cache_lock);
+
+	entry = __gvt_cache_find_gfn(vgpu, gfn);
+	if (!entry) {
+		ret = gvt_dma_map_page(vgpu, gfn, dma_addr, size);
+		if (ret)
+			goto err_unlock;
+
+		ret = __gvt_cache_add(vgpu, gfn, *dma_addr, size);
+		if (ret)
+			goto err_unmap;
+	} else if (entry->size != size) {
+		/* the same gfn with different size: unmap and re-map */
+		gvt_dma_unmap_page(vgpu, gfn, entry->dma_addr, entry->size);
+		__gvt_cache_remove_entry(vgpu, entry);
+
+		ret = gvt_dma_map_page(vgpu, gfn, dma_addr, size);
+		if (ret)
+			goto err_unlock;
+
+		ret = __gvt_cache_add(vgpu, gfn, *dma_addr, size);
+		if (ret)
+			goto err_unmap;
+	} else {
+		kref_get(&entry->ref);
+		*dma_addr = entry->dma_addr;
+	}
+
+	mutex_unlock(&vgpu->cache_lock);
+	return 0;
+
+err_unmap:
+	gvt_dma_unmap_page(vgpu, gfn, *dma_addr, size);
+err_unlock:
+	mutex_unlock(&vgpu->cache_lock);
+	return ret;
+}
+
+int intel_gvt_dma_pin_guest_page(struct intel_vgpu *vgpu, dma_addr_t dma_addr)
+{
+	struct gvt_dma *entry;
+	int ret = 0;
+
+	if (!vgpu->attached)
+		return -ENODEV;
+
+	mutex_lock(&vgpu->cache_lock);
+	entry = __gvt_cache_find_dma_addr(vgpu, dma_addr);
+	if (entry)
+		kref_get(&entry->ref);
+	else
+		ret = -ENOMEM;
+	mutex_unlock(&vgpu->cache_lock);
+
+	return ret;
+}
+
+static void __gvt_dma_release(struct kref *ref)
+{
+	struct gvt_dma *entry = container_of(ref, typeof(*entry), ref);
+
+	gvt_dma_unmap_page(entry->vgpu, entry->gfn, entry->dma_addr,
+			   entry->size);
+	__gvt_cache_remove_entry(entry->vgpu, entry);
+}
+
+void intel_gvt_dma_unmap_guest_page(struct intel_vgpu *vgpu,
+		dma_addr_t dma_addr)
+{
+	struct gvt_dma *entry;
+
+	if (!vgpu->attached)
+		return;
+
+	mutex_lock(&vgpu->cache_lock);
+	entry = __gvt_cache_find_dma_addr(vgpu, dma_addr);
+	if (entry)
+		kref_put(&entry->ref, __gvt_dma_release);
+	mutex_unlock(&vgpu->cache_lock);
+}
+
+static void init_device_info(struct intel_gvt *gvt)
+{
+	struct intel_gvt_device_info *info = &gvt->device_info;
+	struct pci_dev *pdev = to_pci_dev(gvt->gt->i915->drm.dev);
+
+	info->max_support_vgpus = 8;
+	info->cfg_space_size = PCI_CFG_SPACE_EXP_SIZE;
+	info->mmio_size = 2 * 1024 * 1024;
+	info->mmio_bar = 0;
+	info->gtt_start_offset = 8 * 1024 * 1024;
+	info->gtt_entry_size = 8;
+	info->gtt_entry_size_shift = 3;
+	info->gmadr_bytes_in_cmd = 8;
+	info->max_surface_size = 36 * 1024 * 1024;
+	info->msi_cap_offset = pdev->msi_cap;
+}
+
+static void intel_gvt_test_and_emulate_vblank(struct intel_gvt *gvt)
+{
+	struct intel_vgpu *vgpu;
+	int id;
+
+	mutex_lock(&gvt->lock);
+	idr_for_each_entry((&(gvt)->vgpu_idr), (vgpu), (id)) {
+		if (test_and_clear_bit(INTEL_GVT_REQUEST_EMULATE_VBLANK + id,
+				       (void *)&gvt->service_request)) {
+			if (vgpu->active)
+				intel_vgpu_emulate_vblank(vgpu);
+		}
+	}
+	mutex_unlock(&gvt->lock);
+}
+
+static int gvt_service_thread(void *data)
+{
+	struct intel_gvt *gvt = (struct intel_gvt *)data;
+	int ret;
+
+	gvt_dbg_core("service thread start\n");
+
+	while (!kthread_should_stop()) {
+		ret = wait_event_interruptible(gvt->service_thread_wq,
+				kthread_should_stop() || gvt->service_request);
+
+		if (kthread_should_stop())
+			break;
+
+		if (WARN_ONCE(ret, "service thread is waken up by signal.\n"))
+			continue;
+
+		intel_gvt_test_and_emulate_vblank(gvt);
+
+		if (test_bit(INTEL_GVT_REQUEST_SCHED,
+				(void *)&gvt->service_request) ||
+			test_bit(INTEL_GVT_REQUEST_EVENT_SCHED,
+					(void *)&gvt->service_request)) {
+			intel_gvt_schedule(gvt);
+		}
+	}
+
+	return 0;
+}
+
+static void clean_service_thread(struct intel_gvt *gvt)
+{
+	kthread_stop(gvt->service_thread);
+}
+
+static int init_service_thread(struct intel_gvt *gvt)
+{
+	init_waitqueue_head(&gvt->service_thread_wq);
+
+	gvt->service_thread = kthread_run(gvt_service_thread,
+			gvt, "gvt_service_thread");
+	if (IS_ERR(gvt->service_thread)) {
+		gvt_err("fail to start service thread.\n");
+		return PTR_ERR(gvt->service_thread);
+	}
+	return 0;
+}
+
+/**
+ * intel_gvt_clean_device - clean a GVT device
+ * @i915: i915 private
+ *
+ * This function is called at the driver unloading stage, to free the
+ * resources owned by a GVT device.
+ *
+ */
+static void intel_gvt_clean_device(struct drm_i915_private *i915)
+{
+	struct intel_gvt *gvt = fetch_and_zero(&i915->gvt);
+
+	if (drm_WARN_ON(&i915->drm, !gvt))
+		return;
+
+	mdev_unregister_parent(&gvt->parent);
+	intel_gvt_destroy_idle_vgpu(gvt->idle_vgpu);
+	intel_gvt_clean_vgpu_types(gvt);
+
+	intel_gvt_debugfs_clean(gvt);
+	clean_service_thread(gvt);
+	intel_gvt_clean_cmd_parser(gvt);
+	intel_gvt_clean_sched_policy(gvt);
+	intel_gvt_clean_workload_scheduler(gvt);
+	intel_gvt_clean_gtt(gvt);
+	intel_gvt_free_firmware(gvt);
+	intel_gvt_clean_mmio_info(gvt);
+	idr_destroy(&gvt->vgpu_idr);
+
+	kfree(i915->gvt);
+}
+
+/**
+ * intel_gvt_init_device - initialize a GVT device
+ * @i915: drm i915 private data
+ *
+ * This function is called at the initialization stage, to initialize
+ * necessary GVT components.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+static int intel_gvt_init_device(struct drm_i915_private *i915)
+{
+	struct intel_gvt *gvt;
+	struct intel_vgpu *vgpu;
+	int ret;
+
+	if (drm_WARN_ON(&i915->drm, i915->gvt))
+		return -EEXIST;
+
+	gvt = kzalloc(sizeof(struct intel_gvt), GFP_KERNEL);
+	if (!gvt)
+		return -ENOMEM;
+
+	gvt_dbg_core("init gvt device\n");
+
+	idr_init_base(&gvt->vgpu_idr, 1);
+	spin_lock_init(&gvt->scheduler.mmio_context_lock);
+	mutex_init(&gvt->lock);
+	mutex_init(&gvt->sched_lock);
+	gvt->gt = to_gt(i915);
+	i915->gvt = gvt;
+
+	init_device_info(gvt);
+
+	ret = intel_gvt_setup_mmio_info(gvt);
+	if (ret)
+		goto out_clean_idr;
+
+	intel_gvt_init_engine_mmio_context(gvt);
+
+	ret = intel_gvt_load_firmware(gvt);
+	if (ret)
+		goto out_clean_mmio_info;
+
+	ret = intel_gvt_init_irq(gvt);
+	if (ret)
+		goto out_free_firmware;
+
+	ret = intel_gvt_init_gtt(gvt);
+	if (ret)
+		goto out_free_firmware;
+
+	ret = intel_gvt_init_workload_scheduler(gvt);
+	if (ret)
+		goto out_clean_gtt;
+
+	ret = intel_gvt_init_sched_policy(gvt);
+	if (ret)
+		goto out_clean_workload_scheduler;
+
+	ret = intel_gvt_init_cmd_parser(gvt);
+	if (ret)
+		goto out_clean_sched_policy;
+
+	ret = init_service_thread(gvt);
+	if (ret)
+		goto out_clean_cmd_parser;
+
+	ret = intel_gvt_init_vgpu_types(gvt);
+	if (ret)
+		goto out_clean_thread;
+
+	vgpu = intel_gvt_create_idle_vgpu(gvt);
+	if (IS_ERR(vgpu)) {
+		ret = PTR_ERR(vgpu);
+		gvt_err("failed to create idle vgpu\n");
+		goto out_clean_types;
+	}
+	gvt->idle_vgpu = vgpu;
+
+	intel_gvt_debugfs_init(gvt);
+
+	ret = mdev_register_parent(&gvt->parent, i915->drm.dev,
+				   &intel_vgpu_mdev_driver,
+				   gvt->mdev_types, gvt->num_types);
+	if (ret)
+		goto out_destroy_idle_vgpu;
+
+	gvt_dbg_core("gvt device initialization is done\n");
+	return 0;
+
+out_destroy_idle_vgpu:
+	intel_gvt_destroy_idle_vgpu(gvt->idle_vgpu);
+	intel_gvt_debugfs_clean(gvt);
+out_clean_types:
+	intel_gvt_clean_vgpu_types(gvt);
+out_clean_thread:
+	clean_service_thread(gvt);
+out_clean_cmd_parser:
+	intel_gvt_clean_cmd_parser(gvt);
+out_clean_sched_policy:
+	intel_gvt_clean_sched_policy(gvt);
+out_clean_workload_scheduler:
+	intel_gvt_clean_workload_scheduler(gvt);
+out_clean_gtt:
+	intel_gvt_clean_gtt(gvt);
+out_free_firmware:
+	intel_gvt_free_firmware(gvt);
+out_clean_mmio_info:
+	intel_gvt_clean_mmio_info(gvt);
+out_clean_idr:
+	idr_destroy(&gvt->vgpu_idr);
+	kfree(gvt);
+	i915->gvt = NULL;
+	return ret;
+}
+
+static void intel_gvt_pm_resume(struct drm_i915_private *i915)
+{
+	struct intel_gvt *gvt = i915->gvt;
+
+	intel_gvt_restore_fence(gvt);
+	intel_gvt_restore_mmio(gvt);
+	intel_gvt_restore_ggtt(gvt);
+}
+
+static const struct intel_vgpu_ops intel_gvt_vgpu_ops = {
+	.init_device	= intel_gvt_init_device,
+	.clean_device	= intel_gvt_clean_device,
+	.pm_resume	= intel_gvt_pm_resume,
+};
+
+static int __init kvmgt_init(void)
+{
+	int ret;
+
+	ret = intel_gvt_set_ops(&intel_gvt_vgpu_ops);
+	if (ret)
+		return ret;
+
+	ret = mdev_register_driver(&intel_vgpu_mdev_driver);
+	if (ret)
+		intel_gvt_clear_ops(&intel_gvt_vgpu_ops);
+	return ret;
+}
+
+static void __exit kvmgt_exit(void)
+{
+	mdev_unregister_driver(&intel_vgpu_mdev_driver);
+	intel_gvt_clear_ops(&intel_gvt_vgpu_ops);
+}
+
+module_init(kvmgt_init);
+module_exit(kvmgt_exit);
+
+MODULE_LICENSE("GPL and additional rights");
+MODULE_AUTHOR("Intel Corporation");
diff --git a/drivers/gpu/drm/i915/gvt/mmio.c b/drivers/gpu/drm/i915/gvt/mmio.c
new file mode 100644
index 000000000..9acc00505
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/mmio.c
@@ -0,0 +1,332 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Ke Yu
+ *    Kevin Tian <kevin.tian@intel.com>
+ *    Dexuan Cui
+ *
+ * Contributors:
+ *    Tina Zhang <tina.zhang@intel.com>
+ *    Min He <min.he@intel.com>
+ *    Niu Bing <bing.niu@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "gvt.h"
+
+#include "gt/intel_gt_regs.h"
+
+/**
+ * intel_vgpu_gpa_to_mmio_offset - translate a GPA to MMIO offset
+ * @vgpu: a vGPU
+ * @gpa: guest physical address
+ *
+ * Returns:
+ * Zero on success, negative error code if failed
+ */
+int intel_vgpu_gpa_to_mmio_offset(struct intel_vgpu *vgpu, u64 gpa)
+{
+	u64 gttmmio_gpa = intel_vgpu_get_bar_gpa(vgpu, PCI_BASE_ADDRESS_0);
+	return gpa - gttmmio_gpa;
+}
+
+#define reg_is_mmio(gvt, reg)  \
+	(reg >= 0 && reg < gvt->device_info.mmio_size)
+
+#define reg_is_gtt(gvt, reg)   \
+	(reg >= gvt->device_info.gtt_start_offset \
+	 && reg < gvt->device_info.gtt_start_offset + gvt_ggtt_sz(gvt))
+
+static void failsafe_emulate_mmio_rw(struct intel_vgpu *vgpu, u64 pa,
+		void *p_data, unsigned int bytes, bool read)
+{
+	struct intel_gvt *gvt = NULL;
+	void *pt = NULL;
+	unsigned int offset = 0;
+
+	if (!vgpu || !p_data)
+		return;
+
+	gvt = vgpu->gvt;
+	mutex_lock(&vgpu->vgpu_lock);
+	offset = intel_vgpu_gpa_to_mmio_offset(vgpu, pa);
+	if (reg_is_mmio(gvt, offset)) {
+		if (read)
+			intel_vgpu_default_mmio_read(vgpu, offset, p_data,
+					bytes);
+		else
+			intel_vgpu_default_mmio_write(vgpu, offset, p_data,
+					bytes);
+	} else if (reg_is_gtt(gvt, offset)) {
+		offset -= gvt->device_info.gtt_start_offset;
+		pt = vgpu->gtt.ggtt_mm->ggtt_mm.virtual_ggtt + offset;
+		if (read)
+			memcpy(p_data, pt, bytes);
+		else
+			memcpy(pt, p_data, bytes);
+
+	}
+	mutex_unlock(&vgpu->vgpu_lock);
+}
+
+/**
+ * intel_vgpu_emulate_mmio_read - emulate MMIO read
+ * @vgpu: a vGPU
+ * @pa: guest physical address
+ * @p_data: data return buffer
+ * @bytes: access data length
+ *
+ * Returns:
+ * Zero on success, negative error code if failed
+ */
+int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, u64 pa,
+		void *p_data, unsigned int bytes)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct drm_i915_private *i915 = gvt->gt->i915;
+	unsigned int offset = 0;
+	int ret = -EINVAL;
+
+	if (vgpu->failsafe) {
+		failsafe_emulate_mmio_rw(vgpu, pa, p_data, bytes, true);
+		return 0;
+	}
+	mutex_lock(&vgpu->vgpu_lock);
+
+	offset = intel_vgpu_gpa_to_mmio_offset(vgpu, pa);
+
+	if (drm_WARN_ON(&i915->drm, bytes > 8))
+		goto err;
+
+	if (reg_is_gtt(gvt, offset)) {
+		if (drm_WARN_ON(&i915->drm, !IS_ALIGNED(offset, 4) &&
+				!IS_ALIGNED(offset, 8)))
+			goto err;
+		if (drm_WARN_ON(&i915->drm, bytes != 4 && bytes != 8))
+			goto err;
+		if (drm_WARN_ON(&i915->drm,
+				!reg_is_gtt(gvt, offset + bytes - 1)))
+			goto err;
+
+		ret = intel_vgpu_emulate_ggtt_mmio_read(vgpu, offset,
+				p_data, bytes);
+		if (ret)
+			goto err;
+		goto out;
+	}
+
+	if (drm_WARN_ON_ONCE(&i915->drm, !reg_is_mmio(gvt, offset))) {
+		ret = intel_gvt_read_gpa(vgpu, pa, p_data, bytes);
+		goto out;
+	}
+
+	if (drm_WARN_ON(&i915->drm, !reg_is_mmio(gvt, offset + bytes - 1)))
+		goto err;
+
+	if (!intel_gvt_mmio_is_unalign(gvt, offset)) {
+		if (drm_WARN_ON(&i915->drm, !IS_ALIGNED(offset, bytes)))
+			goto err;
+	}
+
+	ret = intel_vgpu_mmio_reg_rw(vgpu, offset, p_data, bytes, true);
+	if (ret < 0)
+		goto err;
+
+	intel_gvt_mmio_set_accessed(gvt, offset);
+	ret = 0;
+	goto out;
+
+err:
+	gvt_vgpu_err("fail to emulate MMIO read %08x len %d\n",
+			offset, bytes);
+out:
+	mutex_unlock(&vgpu->vgpu_lock);
+	return ret;
+}
+
+/**
+ * intel_vgpu_emulate_mmio_write - emulate MMIO write
+ * @vgpu: a vGPU
+ * @pa: guest physical address
+ * @p_data: write data buffer
+ * @bytes: access data length
+ *
+ * Returns:
+ * Zero on success, negative error code if failed
+ */
+int intel_vgpu_emulate_mmio_write(struct intel_vgpu *vgpu, u64 pa,
+		void *p_data, unsigned int bytes)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct drm_i915_private *i915 = gvt->gt->i915;
+	unsigned int offset = 0;
+	int ret = -EINVAL;
+
+	if (vgpu->failsafe) {
+		failsafe_emulate_mmio_rw(vgpu, pa, p_data, bytes, false);
+		return 0;
+	}
+
+	mutex_lock(&vgpu->vgpu_lock);
+
+	offset = intel_vgpu_gpa_to_mmio_offset(vgpu, pa);
+
+	if (drm_WARN_ON(&i915->drm, bytes > 8))
+		goto err;
+
+	if (reg_is_gtt(gvt, offset)) {
+		if (drm_WARN_ON(&i915->drm, !IS_ALIGNED(offset, 4) &&
+				!IS_ALIGNED(offset, 8)))
+			goto err;
+		if (drm_WARN_ON(&i915->drm, bytes != 4 && bytes != 8))
+			goto err;
+		if (drm_WARN_ON(&i915->drm,
+				!reg_is_gtt(gvt, offset + bytes - 1)))
+			goto err;
+
+		ret = intel_vgpu_emulate_ggtt_mmio_write(vgpu, offset,
+				p_data, bytes);
+		if (ret)
+			goto err;
+		goto out;
+	}
+
+	if (drm_WARN_ON_ONCE(&i915->drm, !reg_is_mmio(gvt, offset))) {
+		ret = intel_gvt_write_gpa(vgpu, pa, p_data, bytes);
+		goto out;
+	}
+
+	ret = intel_vgpu_mmio_reg_rw(vgpu, offset, p_data, bytes, false);
+	if (ret < 0)
+		goto err;
+
+	intel_gvt_mmio_set_accessed(gvt, offset);
+	ret = 0;
+	goto out;
+err:
+	gvt_vgpu_err("fail to emulate MMIO write %08x len %d\n", offset,
+		     bytes);
+out:
+	mutex_unlock(&vgpu->vgpu_lock);
+	return ret;
+}
+
+
+/**
+ * intel_vgpu_reset_mmio - reset virtual MMIO space
+ * @vgpu: a vGPU
+ * @dmlr: whether this is device model level reset
+ */
+void intel_vgpu_reset_mmio(struct intel_vgpu *vgpu, bool dmlr)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	const struct intel_gvt_device_info *info = &gvt->device_info;
+	void  *mmio = gvt->firmware.mmio;
+
+	if (dmlr) {
+		memcpy(vgpu->mmio.vreg, mmio, info->mmio_size);
+
+		vgpu_vreg_t(vgpu, GEN6_GT_THREAD_STATUS_REG) = 0;
+
+		/* set the bit 0:2(Core C-State ) to C0 */
+		vgpu_vreg_t(vgpu, GEN6_GT_CORE_STATUS) = 0;
+
+		/* uc reset hw expect GS_MIA_IN_RESET */
+		vgpu_vreg_t(vgpu, GUC_STATUS) |= GS_MIA_IN_RESET;
+
+		if (IS_BROXTON(vgpu->gvt->gt->i915)) {
+			vgpu_vreg_t(vgpu, BXT_P_CR_GT_DISP_PWRON) &=
+				    ~(BIT(0) | BIT(1));
+			vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY0)) &=
+				    ~PHY_POWER_GOOD;
+			vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY1)) &=
+				    ~PHY_POWER_GOOD;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL_FAMILY(DPIO_PHY0)) &=
+				    ~BIT(30);
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL_FAMILY(DPIO_PHY1)) &=
+				    ~BIT(30);
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_A)) &=
+				    ~BXT_PHY_LANE_ENABLED;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_A)) |=
+				    BXT_PHY_CMNLANE_POWERDOWN_ACK |
+				    BXT_PHY_LANE_POWERDOWN_ACK;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_B)) &=
+				    ~BXT_PHY_LANE_ENABLED;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_B)) |=
+				    BXT_PHY_CMNLANE_POWERDOWN_ACK |
+				    BXT_PHY_LANE_POWERDOWN_ACK;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_C)) &=
+				    ~BXT_PHY_LANE_ENABLED;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_C)) |=
+				    BXT_PHY_CMNLANE_POWERDOWN_ACK |
+				    BXT_PHY_LANE_POWERDOWN_ACK;
+			vgpu_vreg_t(vgpu, SKL_FUSE_STATUS) |=
+				SKL_FUSE_DOWNLOAD_STATUS |
+				SKL_FUSE_PG_DIST_STATUS(SKL_PG0) |
+				SKL_FUSE_PG_DIST_STATUS(SKL_PG1) |
+				SKL_FUSE_PG_DIST_STATUS(SKL_PG2);
+		}
+	} else {
+#define GVT_GEN8_MMIO_RESET_OFFSET		(0x44200)
+		/* only reset the engine related, so starting with 0x44200
+		 * interrupt include DE,display mmio related will not be
+		 * touched
+		 */
+		memcpy(vgpu->mmio.vreg, mmio, GVT_GEN8_MMIO_RESET_OFFSET);
+	}
+
+}
+
+/**
+ * intel_vgpu_init_mmio - init MMIO  space
+ * @vgpu: a vGPU
+ *
+ * Returns:
+ * Zero on success, negative error code if failed
+ */
+int intel_vgpu_init_mmio(struct intel_vgpu *vgpu)
+{
+	const struct intel_gvt_device_info *info = &vgpu->gvt->device_info;
+
+	vgpu->mmio.vreg = vzalloc(info->mmio_size);
+	if (!vgpu->mmio.vreg)
+		return -ENOMEM;
+
+	intel_vgpu_reset_mmio(vgpu, true);
+
+	return 0;
+}
+
+/**
+ * intel_vgpu_clean_mmio - clean MMIO space
+ * @vgpu: a vGPU
+ *
+ */
+void intel_vgpu_clean_mmio(struct intel_vgpu *vgpu)
+{
+	vfree(vgpu->mmio.vreg);
+	vgpu->mmio.vreg = NULL;
+}
diff --git a/drivers/gpu/drm/i915/gvt/mmio.h b/drivers/gpu/drm/i915/gvt/mmio.h
new file mode 100644
index 000000000..bba154e38
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/mmio.h
@@ -0,0 +1,113 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Ke Yu
+ *    Kevin Tian <kevin.tian@intel.com>
+ *    Dexuan Cui
+ *
+ * Contributors:
+ *    Tina Zhang <tina.zhang@intel.com>
+ *    Min He <min.he@intel.com>
+ *    Niu Bing <bing.niu@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#ifndef _GVT_MMIO_H_
+#define _GVT_MMIO_H_
+
+#include <linux/types.h>
+
+struct intel_gvt;
+struct intel_vgpu;
+
+#define D_BDW   (1 << 0)
+#define D_SKL	(1 << 1)
+#define D_KBL	(1 << 2)
+#define D_BXT	(1 << 3)
+#define D_CFL	(1 << 4)
+
+#define D_GEN9PLUS	(D_SKL | D_KBL | D_BXT | D_CFL)
+#define D_GEN8PLUS	(D_BDW | D_SKL | D_KBL | D_BXT | D_CFL)
+
+#define D_SKL_PLUS	(D_SKL | D_KBL | D_BXT | D_CFL)
+#define D_BDW_PLUS	(D_BDW | D_SKL | D_KBL | D_BXT | D_CFL)
+
+#define D_PRE_SKL	(D_BDW)
+#define D_ALL		(D_BDW | D_SKL | D_KBL | D_BXT | D_CFL)
+
+typedef int (*gvt_mmio_func)(struct intel_vgpu *, unsigned int, void *,
+			     unsigned int);
+
+struct intel_gvt_mmio_info {
+	u32 offset;
+	u64 ro_mask;
+	u32 device;
+	gvt_mmio_func read;
+	gvt_mmio_func write;
+	u32 addr_range;
+	struct hlist_node node;
+};
+
+const struct intel_engine_cs *
+intel_gvt_render_mmio_to_engine(struct intel_gvt *gvt, unsigned int reg);
+unsigned long intel_gvt_get_device_type(struct intel_gvt *gvt);
+
+int intel_gvt_setup_mmio_info(struct intel_gvt *gvt);
+void intel_gvt_clean_mmio_info(struct intel_gvt *gvt);
+int intel_gvt_for_each_tracked_mmio(struct intel_gvt *gvt,
+	int (*handler)(struct intel_gvt *gvt, u32 offset, void *data),
+	void *data);
+
+struct intel_gvt_mmio_info *intel_gvt_find_mmio_info(struct intel_gvt *gvt,
+						  unsigned int offset);
+
+int intel_vgpu_init_mmio(struct intel_vgpu *vgpu);
+void intel_vgpu_reset_mmio(struct intel_vgpu *vgpu, bool dmlr);
+void intel_vgpu_clean_mmio(struct intel_vgpu *vgpu);
+
+int intel_vgpu_gpa_to_mmio_offset(struct intel_vgpu *vgpu, u64 gpa);
+
+int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, u64 pa,
+				void *p_data, unsigned int bytes);
+int intel_vgpu_emulate_mmio_write(struct intel_vgpu *vgpu, u64 pa,
+				void *p_data, unsigned int bytes);
+
+int intel_vgpu_default_mmio_read(struct intel_vgpu *vgpu, unsigned int offset,
+				 void *p_data, unsigned int bytes);
+int intel_vgpu_default_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+				  void *p_data, unsigned int bytes);
+
+bool intel_gvt_in_force_nonpriv_whitelist(struct intel_gvt *gvt,
+					  unsigned int offset);
+
+int intel_vgpu_mmio_reg_rw(struct intel_vgpu *vgpu, unsigned int offset,
+			   void *pdata, unsigned int bytes, bool is_read);
+
+int intel_vgpu_mask_mmio_write(struct intel_vgpu *vgpu, unsigned int offset,
+				  void *p_data, unsigned int bytes);
+
+void intel_gvt_restore_fence(struct intel_gvt *gvt);
+void intel_gvt_restore_mmio(struct intel_gvt *gvt);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.c b/drivers/gpu/drm/i915/gvt/mmio_context.c
new file mode 100644
index 000000000..1c6e941c9
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/mmio_context.c
@@ -0,0 +1,606 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Eddie Dong <eddie.dong@intel.com>
+ *    Kevin Tian <kevin.tian@intel.com>
+ *
+ * Contributors:
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *    Changbin Du <changbin.du@intel.com>
+ *    Zhenyu Wang <zhenyuw@linux.intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *
+ */
+
+#include "i915_drv.h"
+#include "gt/intel_context.h"
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_ring.h"
+#include "gvt.h"
+#include "trace.h"
+
+#define GEN9_MOCS_SIZE		64
+
+/* Raw offset is appened to each line for convenience. */
+static struct engine_mmio gen8_engine_mmio_list[] __cacheline_aligned = {
+	{RCS0, RING_MODE_GEN7(RENDER_RING_BASE), 0xffff, false}, /* 0x229c */
+	{RCS0, GEN9_CTX_PREEMPT_REG, 0x0, false}, /* 0x2248 */
+	{RCS0, HWSTAM, 0x0, false}, /* 0x2098 */
+	{RCS0, INSTPM, 0xffff, true}, /* 0x20c0 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 0), 0, false}, /* 0x24d0 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 1), 0, false}, /* 0x24d4 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 2), 0, false}, /* 0x24d8 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 3), 0, false}, /* 0x24dc */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 4), 0, false}, /* 0x24e0 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 5), 0, false}, /* 0x24e4 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 6), 0, false}, /* 0x24e8 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 7), 0, false}, /* 0x24ec */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 8), 0, false}, /* 0x24f0 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 9), 0, false}, /* 0x24f4 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 10), 0, false}, /* 0x24f8 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 11), 0, false}, /* 0x24fc */
+	{RCS0, CACHE_MODE_1, 0xffff, true}, /* 0x7004 */
+	{RCS0, GEN7_GT_MODE, 0xffff, true}, /* 0x7008 */
+	{RCS0, CACHE_MODE_0_GEN7, 0xffff, true}, /* 0x7000 */
+	{RCS0, GEN7_COMMON_SLICE_CHICKEN1, 0xffff, true}, /* 0x7010 */
+	{RCS0, HDC_CHICKEN0, 0xffff, true}, /* 0x7300 */
+	{RCS0, VF_GUARDBAND, 0xffff, true}, /* 0x83a4 */
+
+	{BCS0, RING_GFX_MODE(BLT_RING_BASE), 0xffff, false}, /* 0x2229c */
+	{BCS0, RING_MI_MODE(BLT_RING_BASE), 0xffff, false}, /* 0x2209c */
+	{BCS0, RING_INSTPM(BLT_RING_BASE), 0xffff, false}, /* 0x220c0 */
+	{BCS0, RING_HWSTAM(BLT_RING_BASE), 0x0, false}, /* 0x22098 */
+	{BCS0, RING_EXCC(BLT_RING_BASE), 0xffff, false}, /* 0x22028 */
+	{RCS0, INVALID_MMIO_REG, 0, false } /* Terminated */
+};
+
+static struct engine_mmio gen9_engine_mmio_list[] __cacheline_aligned = {
+	{RCS0, RING_MODE_GEN7(RENDER_RING_BASE), 0xffff, false}, /* 0x229c */
+	{RCS0, GEN9_CTX_PREEMPT_REG, 0x0, false}, /* 0x2248 */
+	{RCS0, HWSTAM, 0x0, false}, /* 0x2098 */
+	{RCS0, INSTPM, 0xffff, true}, /* 0x20c0 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 0), 0, false}, /* 0x24d0 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 1), 0, false}, /* 0x24d4 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 2), 0, false}, /* 0x24d8 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 3), 0, false}, /* 0x24dc */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 4), 0, false}, /* 0x24e0 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 5), 0, false}, /* 0x24e4 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 6), 0, false}, /* 0x24e8 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 7), 0, false}, /* 0x24ec */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 8), 0, false}, /* 0x24f0 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 9), 0, false}, /* 0x24f4 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 10), 0, false}, /* 0x24f8 */
+	{RCS0, RING_FORCE_TO_NONPRIV(RENDER_RING_BASE, 11), 0, false}, /* 0x24fc */
+	{RCS0, CACHE_MODE_1, 0xffff, true}, /* 0x7004 */
+	{RCS0, GEN7_GT_MODE, 0xffff, true}, /* 0x7008 */
+	{RCS0, CACHE_MODE_0_GEN7, 0xffff, true}, /* 0x7000 */
+	{RCS0, GEN7_COMMON_SLICE_CHICKEN1, 0xffff, true}, /* 0x7010 */
+	{RCS0, HDC_CHICKEN0, 0xffff, true}, /* 0x7300 */
+	{RCS0, VF_GUARDBAND, 0xffff, true}, /* 0x83a4 */
+
+	{RCS0, GEN8_PRIVATE_PAT_LO, 0, false}, /* 0x40e0 */
+	{RCS0, GEN8_PRIVATE_PAT_HI, 0, false}, /* 0x40e4 */
+	{RCS0, GEN8_CS_CHICKEN1, 0xffff, true}, /* 0x2580 */
+	{RCS0, COMMON_SLICE_CHICKEN2, 0xffff, true}, /* 0x7014 */
+	{RCS0, GEN9_CS_DEBUG_MODE1, 0xffff, false}, /* 0x20ec */
+	{RCS0, GEN8_L3SQCREG4, 0, false}, /* 0xb118 */
+	{RCS0, GEN9_SCRATCH1, 0, false}, /* 0xb11c */
+	{RCS0, GEN9_SCRATCH_LNCF1, 0, false}, /* 0xb008 */
+	{RCS0, GEN7_HALF_SLICE_CHICKEN1, 0xffff, true}, /* 0xe100 */
+	{RCS0, HALF_SLICE_CHICKEN2, 0xffff, true}, /* 0xe180 */
+	{RCS0, HALF_SLICE_CHICKEN3, 0xffff, true}, /* 0xe184 */
+	{RCS0, GEN9_HALF_SLICE_CHICKEN5, 0xffff, true}, /* 0xe188 */
+	{RCS0, GEN9_HALF_SLICE_CHICKEN7, 0xffff, true}, /* 0xe194 */
+	{RCS0, GEN8_ROW_CHICKEN, 0xffff, true}, /* 0xe4f0 */
+	{RCS0, TRVATTL3PTRDW(0), 0, true}, /* 0x4de0 */
+	{RCS0, TRVATTL3PTRDW(1), 0, true}, /* 0x4de4 */
+	{RCS0, TRNULLDETCT, 0, true}, /* 0x4de8 */
+	{RCS0, TRINVTILEDETCT, 0, true}, /* 0x4dec */
+	{RCS0, TRVADR, 0, true}, /* 0x4df0 */
+	{RCS0, TRTTE, 0, true}, /* 0x4df4 */
+	{RCS0, _MMIO(0x4dfc), 0, true},
+
+	{BCS0, RING_GFX_MODE(BLT_RING_BASE), 0xffff, false}, /* 0x2229c */
+	{BCS0, RING_MI_MODE(BLT_RING_BASE), 0xffff, false}, /* 0x2209c */
+	{BCS0, RING_INSTPM(BLT_RING_BASE), 0xffff, false}, /* 0x220c0 */
+	{BCS0, RING_HWSTAM(BLT_RING_BASE), 0x0, false}, /* 0x22098 */
+	{BCS0, RING_EXCC(BLT_RING_BASE), 0xffff, false}, /* 0x22028 */
+
+	{VCS1, RING_EXCC(GEN8_BSD2_RING_BASE), 0xffff, false}, /* 0x1c028 */
+
+	{VECS0, RING_EXCC(VEBOX_RING_BASE), 0xffff, false}, /* 0x1a028 */
+
+	{RCS0, GEN8_HDC_CHICKEN1, 0xffff, true}, /* 0x7304 */
+	{RCS0, GEN9_CTX_PREEMPT_REG, 0x0, false}, /* 0x2248 */
+	{RCS0, GEN7_UCGCTL4, 0x0, false}, /* 0x940c */
+	{RCS0, GAMT_CHKN_BIT_REG, 0x0, false}, /* 0x4ab8 */
+
+	{RCS0, GEN9_GAMT_ECO_REG_RW_IA, 0x0, false}, /* 0x4ab0 */
+	{RCS0, GEN9_CSFE_CHICKEN1_RCS, 0xffff, false}, /* 0x20d4 */
+	{RCS0, _MMIO(0x20D8), 0xffff, true}, /* 0x20d8 */
+
+	{RCS0, GEN8_GARBCNTL, 0x0, false}, /* 0xb004 */
+	{RCS0, GEN7_FF_THREAD_MODE, 0x0, false}, /* 0x20a0 */
+	{RCS0, FF_SLICE_CS_CHICKEN2, 0xffff, false}, /* 0x20e4 */
+	{RCS0, INVALID_MMIO_REG, 0, false } /* Terminated */
+};
+
+static struct {
+	bool initialized;
+	u32 control_table[I915_NUM_ENGINES][GEN9_MOCS_SIZE];
+	u32 l3cc_table[GEN9_MOCS_SIZE / 2];
+} gen9_render_mocs;
+
+static u32 gen9_mocs_mmio_offset_list[] = {
+	[RCS0]  = 0xc800,
+	[VCS0]  = 0xc900,
+	[VCS1]  = 0xca00,
+	[BCS0]  = 0xcc00,
+	[VECS0] = 0xcb00,
+};
+
+static void load_render_mocs(const struct intel_engine_cs *engine)
+{
+	struct intel_gvt *gvt = engine->i915->gvt;
+	struct intel_uncore *uncore = engine->uncore;
+	u32 cnt = gvt->engine_mmio_list.mocs_mmio_offset_list_cnt;
+	u32 *regs = gvt->engine_mmio_list.mocs_mmio_offset_list;
+	i915_reg_t offset;
+	int ring_id, i;
+
+	/* Platform doesn't have mocs mmios. */
+	if (!regs)
+		return;
+
+	for (ring_id = 0; ring_id < cnt; ring_id++) {
+		if (!HAS_ENGINE(engine->gt, ring_id))
+			continue;
+
+		offset.reg = regs[ring_id];
+		for (i = 0; i < GEN9_MOCS_SIZE; i++) {
+			gen9_render_mocs.control_table[ring_id][i] =
+				intel_uncore_read_fw(uncore, offset);
+			offset.reg += 4;
+		}
+	}
+
+	offset.reg = 0xb020;
+	for (i = 0; i < GEN9_MOCS_SIZE / 2; i++) {
+		gen9_render_mocs.l3cc_table[i] =
+			intel_uncore_read_fw(uncore, offset);
+		offset.reg += 4;
+	}
+	gen9_render_mocs.initialized = true;
+}
+
+static int
+restore_context_mmio_for_inhibit(struct intel_vgpu *vgpu,
+				 struct i915_request *req)
+{
+	u32 *cs;
+	int ret;
+	struct engine_mmio *mmio;
+	struct intel_gvt *gvt = vgpu->gvt;
+	int ring_id = req->engine->id;
+	int count = gvt->engine_mmio_list.ctx_mmio_count[ring_id];
+
+	if (count == 0)
+		return 0;
+
+	ret = req->engine->emit_flush(req, EMIT_BARRIER);
+	if (ret)
+		return ret;
+
+	cs = intel_ring_begin(req, count * 2 + 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(count);
+	for (mmio = gvt->engine_mmio_list.mmio;
+	     i915_mmio_reg_valid(mmio->reg); mmio++) {
+		if (mmio->id != ring_id || !mmio->in_context)
+			continue;
+
+		*cs++ = i915_mmio_reg_offset(mmio->reg);
+		*cs++ = vgpu_vreg_t(vgpu, mmio->reg) | (mmio->mask << 16);
+		gvt_dbg_core("add lri reg pair 0x%x:0x%x in inhibit ctx, vgpu:%d, rind_id:%d\n",
+			      *(cs-2), *(cs-1), vgpu->id, ring_id);
+	}
+
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
+
+	ret = req->engine->emit_flush(req, EMIT_BARRIER);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int
+restore_render_mocs_control_for_inhibit(struct intel_vgpu *vgpu,
+					struct i915_request *req)
+{
+	unsigned int index;
+	u32 *cs;
+
+	cs = intel_ring_begin(req, 2 * GEN9_MOCS_SIZE + 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(GEN9_MOCS_SIZE);
+
+	for (index = 0; index < GEN9_MOCS_SIZE; index++) {
+		*cs++ = i915_mmio_reg_offset(GEN9_GFX_MOCS(index));
+		*cs++ = vgpu_vreg_t(vgpu, GEN9_GFX_MOCS(index));
+		gvt_dbg_core("add lri reg pair 0x%x:0x%x in inhibit ctx, vgpu:%d, rind_id:%d\n",
+			      *(cs-2), *(cs-1), vgpu->id, req->engine->id);
+
+	}
+
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
+
+	return 0;
+}
+
+static int
+restore_render_mocs_l3cc_for_inhibit(struct intel_vgpu *vgpu,
+				     struct i915_request *req)
+{
+	unsigned int index;
+	u32 *cs;
+
+	cs = intel_ring_begin(req, 2 * GEN9_MOCS_SIZE / 2 + 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(GEN9_MOCS_SIZE / 2);
+
+	for (index = 0; index < GEN9_MOCS_SIZE / 2; index++) {
+		*cs++ = i915_mmio_reg_offset(GEN9_LNCFCMOCS(index));
+		*cs++ = vgpu_vreg_t(vgpu, GEN9_LNCFCMOCS(index));
+		gvt_dbg_core("add lri reg pair 0x%x:0x%x in inhibit ctx, vgpu:%d, rind_id:%d\n",
+			      *(cs-2), *(cs-1), vgpu->id, req->engine->id);
+
+	}
+
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
+
+	return 0;
+}
+
+/*
+ * Use lri command to initialize the mmio which is in context state image for
+ * inhibit context, it contains tracked engine mmio, render_mocs and
+ * render_mocs_l3cc.
+ */
+int intel_vgpu_restore_inhibit_context(struct intel_vgpu *vgpu,
+				       struct i915_request *req)
+{
+	int ret;
+	u32 *cs;
+
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
+
+	ret = restore_context_mmio_for_inhibit(vgpu, req);
+	if (ret)
+		goto out;
+
+	/* no MOCS register in context except render engine */
+	if (req->engine->id != RCS0)
+		goto out;
+
+	ret = restore_render_mocs_control_for_inhibit(vgpu, req);
+	if (ret)
+		goto out;
+
+	ret = restore_render_mocs_l3cc_for_inhibit(vgpu, req);
+	if (ret)
+		goto out;
+
+out:
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
+
+	return ret;
+}
+
+static u32 gen8_tlb_mmio_offset_list[] = {
+	[RCS0]  = 0x4260,
+	[VCS0]  = 0x4264,
+	[VCS1]  = 0x4268,
+	[BCS0]  = 0x426c,
+	[VECS0] = 0x4270,
+};
+
+static void handle_tlb_pending_event(struct intel_vgpu *vgpu,
+				     const struct intel_engine_cs *engine)
+{
+	struct intel_uncore *uncore = engine->uncore;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	u32 *regs = vgpu->gvt->engine_mmio_list.tlb_mmio_offset_list;
+	u32 cnt = vgpu->gvt->engine_mmio_list.tlb_mmio_offset_list_cnt;
+	enum forcewake_domains fw;
+	i915_reg_t reg;
+
+	if (!regs)
+		return;
+
+	if (drm_WARN_ON(&engine->i915->drm, engine->id >= cnt))
+		return;
+
+	if (!test_and_clear_bit(engine->id, (void *)s->tlb_handle_pending))
+		return;
+
+	reg = _MMIO(regs[engine->id]);
+
+	/* WaForceWakeRenderDuringMmioTLBInvalidate:skl
+	 * we need to put a forcewake when invalidating RCS TLB caches,
+	 * otherwise device can go to RC6 state and interrupt invalidation
+	 * process
+	 */
+	fw = intel_uncore_forcewake_for_reg(uncore, reg,
+					    FW_REG_READ | FW_REG_WRITE);
+	if (engine->id == RCS0 && GRAPHICS_VER(engine->i915) >= 9)
+		fw |= FORCEWAKE_RENDER;
+
+	intel_uncore_forcewake_get(uncore, fw);
+
+	intel_uncore_write_fw(uncore, reg, 0x1);
+
+	if (wait_for_atomic(intel_uncore_read_fw(uncore, reg) == 0, 50))
+		gvt_vgpu_err("timeout in invalidate ring %s tlb\n",
+			     engine->name);
+	else
+		vgpu_vreg_t(vgpu, reg) = 0;
+
+	intel_uncore_forcewake_put(uncore, fw);
+
+	gvt_dbg_core("invalidate TLB for ring %s\n", engine->name);
+}
+
+static void switch_mocs(struct intel_vgpu *pre, struct intel_vgpu *next,
+			const struct intel_engine_cs *engine)
+{
+	u32 regs[] = {
+		[RCS0]  = 0xc800,
+		[VCS0]  = 0xc900,
+		[VCS1]  = 0xca00,
+		[BCS0]  = 0xcc00,
+		[VECS0] = 0xcb00,
+	};
+	struct intel_uncore *uncore = engine->uncore;
+	i915_reg_t offset, l3_offset;
+	u32 old_v, new_v;
+	int i;
+
+	if (drm_WARN_ON(&engine->i915->drm, engine->id >= ARRAY_SIZE(regs)))
+		return;
+
+	if (engine->id == RCS0 && GRAPHICS_VER(engine->i915) == 9)
+		return;
+
+	if (!pre && !gen9_render_mocs.initialized)
+		load_render_mocs(engine);
+
+	offset.reg = regs[engine->id];
+	for (i = 0; i < GEN9_MOCS_SIZE; i++) {
+		if (pre)
+			old_v = vgpu_vreg_t(pre, offset);
+		else
+			old_v = gen9_render_mocs.control_table[engine->id][i];
+		if (next)
+			new_v = vgpu_vreg_t(next, offset);
+		else
+			new_v = gen9_render_mocs.control_table[engine->id][i];
+
+		if (old_v != new_v)
+			intel_uncore_write_fw(uncore, offset, new_v);
+
+		offset.reg += 4;
+	}
+
+	if (engine->id == RCS0) {
+		l3_offset.reg = 0xb020;
+		for (i = 0; i < GEN9_MOCS_SIZE / 2; i++) {
+			if (pre)
+				old_v = vgpu_vreg_t(pre, l3_offset);
+			else
+				old_v = gen9_render_mocs.l3cc_table[i];
+			if (next)
+				new_v = vgpu_vreg_t(next, l3_offset);
+			else
+				new_v = gen9_render_mocs.l3cc_table[i];
+
+			if (old_v != new_v)
+				intel_uncore_write_fw(uncore, l3_offset, new_v);
+
+			l3_offset.reg += 4;
+		}
+	}
+}
+
+#define CTX_CONTEXT_CONTROL_VAL	0x03
+
+bool is_inhibit_context(struct intel_context *ce)
+{
+	const u32 *reg_state = ce->lrc_reg_state;
+	u32 inhibit_mask =
+		_MASKED_BIT_ENABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT);
+
+	return inhibit_mask ==
+		(reg_state[CTX_CONTEXT_CONTROL_VAL] & inhibit_mask);
+}
+
+/* Switch ring mmio values (context). */
+static void switch_mmio(struct intel_vgpu *pre,
+			struct intel_vgpu *next,
+			const struct intel_engine_cs *engine)
+{
+	struct intel_uncore *uncore = engine->uncore;
+	struct intel_vgpu_submission *s;
+	struct engine_mmio *mmio;
+	u32 old_v, new_v;
+
+	if (GRAPHICS_VER(engine->i915) >= 9)
+		switch_mocs(pre, next, engine);
+
+	for (mmio = engine->i915->gvt->engine_mmio_list.mmio;
+	     i915_mmio_reg_valid(mmio->reg); mmio++) {
+		if (mmio->id != engine->id)
+			continue;
+		/*
+		 * No need to do save or restore of the mmio which is in context
+		 * state image on gen9, it's initialized by lri command and
+		 * save or restore with context together.
+		 */
+		if (GRAPHICS_VER(engine->i915) == 9 && mmio->in_context)
+			continue;
+
+		// save
+		if (pre) {
+			vgpu_vreg_t(pre, mmio->reg) =
+				intel_uncore_read_fw(uncore, mmio->reg);
+			if (mmio->mask)
+				vgpu_vreg_t(pre, mmio->reg) &=
+					~(mmio->mask << 16);
+			old_v = vgpu_vreg_t(pre, mmio->reg);
+		} else {
+			old_v = mmio->value =
+				intel_uncore_read_fw(uncore, mmio->reg);
+		}
+
+		// restore
+		if (next) {
+			s = &next->submission;
+			/*
+			 * No need to restore the mmio which is in context state
+			 * image if it's not inhibit context, it will restore
+			 * itself.
+			 */
+			if (mmio->in_context &&
+			    !is_inhibit_context(s->shadow[engine->id]))
+				continue;
+
+			if (mmio->mask)
+				new_v = vgpu_vreg_t(next, mmio->reg) |
+					(mmio->mask << 16);
+			else
+				new_v = vgpu_vreg_t(next, mmio->reg);
+		} else {
+			if (mmio->in_context)
+				continue;
+			if (mmio->mask)
+				new_v = mmio->value | (mmio->mask << 16);
+			else
+				new_v = mmio->value;
+		}
+
+		intel_uncore_write_fw(uncore, mmio->reg, new_v);
+
+		trace_render_mmio(pre ? pre->id : 0,
+				  next ? next->id : 0,
+				  "switch",
+				  i915_mmio_reg_offset(mmio->reg),
+				  old_v, new_v);
+	}
+
+	if (next)
+		handle_tlb_pending_event(next, engine);
+}
+
+/**
+ * intel_gvt_switch_mmio - switch mmio context of specific engine
+ * @pre: the last vGPU that own the engine
+ * @next: the vGPU to switch to
+ * @engine: the engine
+ *
+ * If pre is null indicates that host own the engine. If next is null
+ * indicates that we are switching to host workload.
+ */
+void intel_gvt_switch_mmio(struct intel_vgpu *pre,
+			   struct intel_vgpu *next,
+			   const struct intel_engine_cs *engine)
+{
+	if (WARN(!pre && !next, "switch ring %s from host to HOST\n",
+		 engine->name))
+		return;
+
+	gvt_dbg_render("switch ring %s from %s to %s\n", engine->name,
+		       pre ? "vGPU" : "host", next ? "vGPU" : "HOST");
+
+	/**
+	 * We are using raw mmio access wrapper to improve the
+	 * performace for batch mmio read/write, so we need
+	 * handle forcewake mannually.
+	 */
+	intel_uncore_forcewake_get(engine->uncore, FORCEWAKE_ALL);
+	switch_mmio(pre, next, engine);
+	intel_uncore_forcewake_put(engine->uncore, FORCEWAKE_ALL);
+}
+
+/**
+ * intel_gvt_init_engine_mmio_context - Initiate the engine mmio list
+ * @gvt: GVT device
+ *
+ */
+void intel_gvt_init_engine_mmio_context(struct intel_gvt *gvt)
+{
+	struct engine_mmio *mmio;
+
+	if (GRAPHICS_VER(gvt->gt->i915) >= 9) {
+		gvt->engine_mmio_list.mmio = gen9_engine_mmio_list;
+		gvt->engine_mmio_list.tlb_mmio_offset_list = gen8_tlb_mmio_offset_list;
+		gvt->engine_mmio_list.tlb_mmio_offset_list_cnt = ARRAY_SIZE(gen8_tlb_mmio_offset_list);
+		gvt->engine_mmio_list.mocs_mmio_offset_list = gen9_mocs_mmio_offset_list;
+		gvt->engine_mmio_list.mocs_mmio_offset_list_cnt = ARRAY_SIZE(gen9_mocs_mmio_offset_list);
+	} else {
+		gvt->engine_mmio_list.mmio = gen8_engine_mmio_list;
+		gvt->engine_mmio_list.tlb_mmio_offset_list = gen8_tlb_mmio_offset_list;
+		gvt->engine_mmio_list.tlb_mmio_offset_list_cnt = ARRAY_SIZE(gen8_tlb_mmio_offset_list);
+	}
+
+	for (mmio = gvt->engine_mmio_list.mmio;
+	     i915_mmio_reg_valid(mmio->reg); mmio++) {
+		if (mmio->in_context) {
+			gvt->engine_mmio_list.ctx_mmio_count[mmio->id]++;
+			intel_gvt_mmio_set_sr_in_ctx(gvt, mmio->reg.reg);
+		}
+	}
+}
diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.h b/drivers/gpu/drm/i915/gvt/mmio_context.h
new file mode 100644
index 000000000..9540813b8
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/mmio_context.h
@@ -0,0 +1,73 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Eddie Dong <eddie.dong@intel.com>
+ *    Kevin Tian <kevin.tian@intel.com>
+ *
+ * Contributors:
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *    Changbin Du <changbin.du@intel.com>
+ *    Zhenyu Wang <zhenyuw@linux.intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *
+ */
+
+#ifndef __GVT_RENDER_H__
+#define __GVT_RENDER_H__
+
+#include <linux/types.h>
+
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_engine_types.h"
+#include "gt/intel_lrc_reg.h"
+
+struct i915_request;
+struct intel_context;
+struct intel_engine_cs;
+struct intel_gvt;
+struct intel_vgpu;
+
+struct engine_mmio {
+	enum intel_engine_id id;
+	i915_reg_t reg;
+	u32 mask;
+	bool in_context;
+	u32 value;
+};
+
+void intel_gvt_switch_mmio(struct intel_vgpu *pre,
+			   struct intel_vgpu *next,
+			   const struct intel_engine_cs *engine);
+
+void intel_gvt_init_engine_mmio_context(struct intel_gvt *gvt);
+
+bool is_inhibit_context(struct intel_context *ce);
+
+int intel_vgpu_restore_inhibit_context(struct intel_vgpu *vgpu,
+				       struct i915_request *req);
+
+#define IS_RESTORE_INHIBIT(a) \
+	IS_MASKED_BITS_ENABLED(a, CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT)
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/opregion.c b/drivers/gpu/drm/i915/gvt/opregion.c
new file mode 100644
index 000000000..d2bed4665
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/opregion.c
@@ -0,0 +1,483 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/acpi.h>
+#include "i915_drv.h"
+#include "gvt.h"
+
+/*
+ * Note: Only for GVT-g virtual VBT generation, other usage must
+ * not do like this.
+ */
+#define _INTEL_BIOS_PRIVATE
+#include "display/intel_vbt_defs.h"
+
+#define OPREGION_SIGNATURE "IntelGraphicsMem"
+#define MBOX_VBT      (1<<3)
+
+/* device handle */
+#define DEVICE_TYPE_CRT    0x01
+#define DEVICE_TYPE_EFP1   0x04
+#define DEVICE_TYPE_EFP2   0x40
+#define DEVICE_TYPE_EFP3   0x20
+#define DEVICE_TYPE_EFP4   0x10
+
+struct opregion_header {
+	u8 signature[16];
+	u32 size;
+	u32 opregion_ver;
+	u8 bios_ver[32];
+	u8 vbios_ver[16];
+	u8 driver_ver[16];
+	u32 mboxes;
+	u32 driver_model;
+	u32 pcon;
+	u8 dver[32];
+	u8 rsvd[124];
+} __packed;
+
+struct bdb_data_header {
+	u8 id;
+	u16 size; /* data size */
+} __packed;
+
+/* For supporting windows guest with opregion, here hardcode the emulated
+ * bdb header version as '186', and the corresponding child_device_config
+ * length should be '33' but not '38'.
+ */
+struct efp_child_device_config {
+	u16 handle;
+	u16 device_type;
+	u16 device_class;
+	u8 i2c_speed;
+	u8 dp_onboard_redriver; /* 158 */
+	u8 dp_ondock_redriver; /* 158 */
+	u8 hdmi_level_shifter_value:4; /* 169 */
+	u8 hdmi_max_data_rate:4; /* 204 */
+	u16 dtd_buf_ptr; /* 161 */
+	u8 edidless_efp:1; /* 161 */
+	u8 compression_enable:1; /* 198 */
+	u8 compression_method:1; /* 198 */
+	u8 ganged_edp:1; /* 202 */
+	u8 skip0:4;
+	u8 compression_structure_index:4; /* 198 */
+	u8 skip1:4;
+	u8 slave_port; /*  202 */
+	u8 skip2;
+	u8 dvo_port;
+	u8 i2c_pin; /* for add-in card */
+	u8 slave_addr; /* for add-in card */
+	u8 ddc_pin;
+	u16 edid_ptr;
+	u8 dvo_config;
+	u8 efp_docked_port:1; /* 158 */
+	u8 lane_reversal:1; /* 184 */
+	u8 onboard_lspcon:1; /* 192 */
+	u8 iboost_enable:1; /* 196 */
+	u8 hpd_invert:1; /* BXT 196 */
+	u8 slip3:3;
+	u8 hdmi_compat:1;
+	u8 dp_compat:1;
+	u8 tmds_compat:1;
+	u8 skip4:5;
+	u8 aux_channel;
+	u8 dongle_detect;
+	u8 pipe_cap:2;
+	u8 sdvo_stall:1; /* 158 */
+	u8 hpd_status:2;
+	u8 integrated_encoder:1;
+	u8 skip5:2;
+	u8 dvo_wiring;
+	u8 mipi_bridge_type; /* 171 */
+	u16 device_class_ext;
+	u8 dvo_function;
+} __packed;
+
+struct vbt {
+	/* header->bdb_offset point to bdb_header offset */
+	struct vbt_header header;
+	struct bdb_header bdb_header;
+
+	struct bdb_data_header general_features_header;
+	struct bdb_general_features general_features;
+
+	struct bdb_data_header general_definitions_header;
+	struct bdb_general_definitions general_definitions;
+
+	struct efp_child_device_config child0;
+	struct efp_child_device_config child1;
+	struct efp_child_device_config child2;
+	struct efp_child_device_config child3;
+
+	struct bdb_data_header driver_features_header;
+	struct bdb_driver_features driver_features;
+};
+
+static void virt_vbt_generation(struct vbt *v)
+{
+	int num_child;
+
+	memset(v, 0, sizeof(struct vbt));
+
+	v->header.signature[0] = '$';
+	v->header.signature[1] = 'V';
+	v->header.signature[2] = 'B';
+	v->header.signature[3] = 'T';
+
+	/* there's features depending on version! */
+	v->header.version = 155;
+	v->header.header_size = sizeof(v->header);
+	v->header.vbt_size = sizeof(struct vbt);
+	v->header.bdb_offset = offsetof(struct vbt, bdb_header);
+
+	strcpy(&v->bdb_header.signature[0], "BIOS_DATA_BLOCK");
+	v->bdb_header.version = 186; /* child_dev_size = 33 */
+	v->bdb_header.header_size = sizeof(v->bdb_header);
+
+	v->bdb_header.bdb_size = sizeof(struct vbt) - sizeof(struct vbt_header);
+
+	/* general features */
+	v->general_features_header.id = BDB_GENERAL_FEATURES;
+	v->general_features_header.size = sizeof(struct bdb_general_features);
+	v->general_features.int_crt_support = 0;
+	v->general_features.int_tv_support = 0;
+
+	/* child device */
+	num_child = 4; /* each port has one child */
+	v->general_definitions.child_dev_size =
+		sizeof(struct efp_child_device_config);
+	v->general_definitions_header.id = BDB_GENERAL_DEFINITIONS;
+	/* size will include child devices */
+	v->general_definitions_header.size =
+		sizeof(struct bdb_general_definitions) +
+			num_child * v->general_definitions.child_dev_size;
+
+	/* portA */
+	v->child0.handle = DEVICE_TYPE_EFP1;
+	v->child0.device_type = DEVICE_TYPE_DP;
+	v->child0.dvo_port = DVO_PORT_DPA;
+	v->child0.aux_channel = DP_AUX_A;
+	v->child0.dp_compat = true;
+	v->child0.integrated_encoder = true;
+
+	/* portB */
+	v->child1.handle = DEVICE_TYPE_EFP2;
+	v->child1.device_type = DEVICE_TYPE_DP;
+	v->child1.dvo_port = DVO_PORT_DPB;
+	v->child1.aux_channel = DP_AUX_B;
+	v->child1.dp_compat = true;
+	v->child1.integrated_encoder = true;
+
+	/* portC */
+	v->child2.handle = DEVICE_TYPE_EFP3;
+	v->child2.device_type = DEVICE_TYPE_DP;
+	v->child2.dvo_port = DVO_PORT_DPC;
+	v->child2.aux_channel = DP_AUX_C;
+	v->child2.dp_compat = true;
+	v->child2.integrated_encoder = true;
+
+	/* portD */
+	v->child3.handle = DEVICE_TYPE_EFP4;
+	v->child3.device_type = DEVICE_TYPE_DP;
+	v->child3.dvo_port = DVO_PORT_DPD;
+	v->child3.aux_channel = DP_AUX_D;
+	v->child3.dp_compat = true;
+	v->child3.integrated_encoder = true;
+
+	/* driver features */
+	v->driver_features_header.id = BDB_DRIVER_FEATURES;
+	v->driver_features_header.size = sizeof(struct bdb_driver_features);
+	v->driver_features.lvds_config = BDB_DRIVER_FEATURE_NO_LVDS;
+}
+
+/**
+ * intel_vgpu_init_opregion - initialize the stuff used to emulate opregion
+ * @vgpu: a vGPU
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_init_opregion(struct intel_vgpu *vgpu)
+{
+	u8 *buf;
+	struct opregion_header *header;
+	struct vbt v;
+	const char opregion_signature[16] = OPREGION_SIGNATURE;
+
+	gvt_dbg_core("init vgpu%d opregion\n", vgpu->id);
+	vgpu_opregion(vgpu)->va = (void *)__get_free_pages(GFP_KERNEL |
+			__GFP_ZERO,
+			get_order(INTEL_GVT_OPREGION_SIZE));
+	if (!vgpu_opregion(vgpu)->va) {
+		gvt_err("fail to get memory for vgpu virt opregion\n");
+		return -ENOMEM;
+	}
+
+	/* emulated opregion with VBT mailbox only */
+	buf = (u8 *)vgpu_opregion(vgpu)->va;
+	header = (struct opregion_header *)buf;
+	memcpy(header->signature, opregion_signature,
+	       sizeof(opregion_signature));
+	header->size = 0x8;
+	header->opregion_ver = 0x02000000;
+	header->mboxes = MBOX_VBT;
+
+	/* for unknown reason, the value in LID field is incorrect
+	 * which block the windows guest, so workaround it by force
+	 * setting it to "OPEN"
+	 */
+	buf[INTEL_GVT_OPREGION_CLID] = 0x3;
+
+	/* emulated vbt from virt vbt generation */
+	virt_vbt_generation(&v);
+	memcpy(buf + INTEL_GVT_OPREGION_VBT_OFFSET, &v, sizeof(struct vbt));
+
+	return 0;
+}
+
+/**
+ * intel_vgpu_opregion_base_write_handler - Opregion base register write handler
+ *
+ * @vgpu: a vGPU
+ * @gpa: guest physical address of opregion
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ */
+int intel_vgpu_opregion_base_write_handler(struct intel_vgpu *vgpu, u32 gpa)
+{
+
+	int i;
+
+	gvt_dbg_core("emulate opregion from kernel\n");
+
+	for (i = 0; i < INTEL_GVT_OPREGION_PAGES; i++)
+		vgpu_opregion(vgpu)->gfn[i] = (gpa >> PAGE_SHIFT) + i;
+	return 0;
+}
+
+/**
+ * intel_vgpu_clean_opregion - clean the stuff used to emulate opregion
+ * @vgpu: a vGPU
+ *
+ */
+void intel_vgpu_clean_opregion(struct intel_vgpu *vgpu)
+{
+	gvt_dbg_core("vgpu%d: clean vgpu opregion\n", vgpu->id);
+
+	if (!vgpu_opregion(vgpu)->va)
+		return;
+
+	/* Guest opregion is released by VFIO */
+	free_pages((unsigned long)vgpu_opregion(vgpu)->va,
+		   get_order(INTEL_GVT_OPREGION_SIZE));
+
+	vgpu_opregion(vgpu)->va = NULL;
+
+}
+
+
+#define GVT_OPREGION_FUNC(scic)					\
+	({							\
+	 u32 __ret;						\
+	 __ret = (scic & OPREGION_SCIC_FUNC_MASK) >>		\
+	 OPREGION_SCIC_FUNC_SHIFT;				\
+	 __ret;							\
+	 })
+
+#define GVT_OPREGION_SUBFUNC(scic)				\
+	({							\
+	 u32 __ret;						\
+	 __ret = (scic & OPREGION_SCIC_SUBFUNC_MASK) >>		\
+	 OPREGION_SCIC_SUBFUNC_SHIFT;				\
+	 __ret;							\
+	 })
+
+static const char *opregion_func_name(u32 func)
+{
+	const char *name = NULL;
+
+	switch (func) {
+	case 0 ... 3:
+	case 5:
+	case 7 ... 15:
+		name = "Reserved";
+		break;
+
+	case 4:
+		name = "Get BIOS Data";
+		break;
+
+	case 6:
+		name = "System BIOS Callbacks";
+		break;
+
+	default:
+		name = "Unknown";
+		break;
+	}
+	return name;
+}
+
+static const char *opregion_subfunc_name(u32 subfunc)
+{
+	const char *name = NULL;
+
+	switch (subfunc) {
+	case 0:
+		name = "Supported Calls";
+		break;
+
+	case 1:
+		name = "Requested Callbacks";
+		break;
+
+	case 2 ... 3:
+	case 8 ... 9:
+		name = "Reserved";
+		break;
+
+	case 5:
+		name = "Boot Display";
+		break;
+
+	case 6:
+		name = "TV-Standard/Video-Connector";
+		break;
+
+	case 7:
+		name = "Internal Graphics";
+		break;
+
+	case 10:
+		name = "Spread Spectrum Clocks";
+		break;
+
+	case 11:
+		name = "Get AKSV";
+		break;
+
+	default:
+		name = "Unknown";
+		break;
+	}
+	return name;
+};
+
+static bool querying_capabilities(u32 scic)
+{
+	u32 func, subfunc;
+
+	func = GVT_OPREGION_FUNC(scic);
+	subfunc = GVT_OPREGION_SUBFUNC(scic);
+
+	if ((func == INTEL_GVT_OPREGION_SCIC_F_GETBIOSDATA &&
+		subfunc == INTEL_GVT_OPREGION_SCIC_SF_SUPPRTEDCALLS)
+		|| (func == INTEL_GVT_OPREGION_SCIC_F_GETBIOSDATA &&
+		 subfunc == INTEL_GVT_OPREGION_SCIC_SF_REQEUSTEDCALLBACKS)
+		|| (func == INTEL_GVT_OPREGION_SCIC_F_GETBIOSCALLBACKS &&
+		 subfunc == INTEL_GVT_OPREGION_SCIC_SF_SUPPRTEDCALLS)) {
+		return true;
+	}
+	return false;
+}
+
+/**
+ * intel_vgpu_emulate_opregion_request - emulating OpRegion request
+ * @vgpu: a vGPU
+ * @swsci: SWSCI request
+ *
+ * Returns:
+ * Zero on success, negative error code if failed
+ */
+int intel_vgpu_emulate_opregion_request(struct intel_vgpu *vgpu, u32 swsci)
+{
+	u32 scic, parm;
+	u32 func, subfunc;
+	u64 scic_pa = 0, parm_pa = 0;
+	int ret;
+
+	scic_pa = (vgpu_opregion(vgpu)->gfn[0] << PAGE_SHIFT) +
+				INTEL_GVT_OPREGION_SCIC;
+	parm_pa = (vgpu_opregion(vgpu)->gfn[0] << PAGE_SHIFT) +
+				INTEL_GVT_OPREGION_PARM;
+	ret = intel_gvt_read_gpa(vgpu, scic_pa, &scic, sizeof(scic));
+	if (ret) {
+		gvt_vgpu_err("guest opregion read error %d, gpa 0x%llx, len %lu\n",
+			ret, scic_pa, sizeof(scic));
+		return ret;
+	}
+
+	ret = intel_gvt_read_gpa(vgpu, parm_pa, &parm, sizeof(parm));
+	if (ret) {
+		gvt_vgpu_err("guest opregion read error %d, gpa 0x%llx, len %lu\n",
+			ret, scic_pa, sizeof(scic));
+		return ret;
+	}
+
+	if (!(swsci & SWSCI_SCI_SELECT)) {
+		gvt_vgpu_err("requesting SMI service\n");
+		return 0;
+	}
+	/* ignore non 0->1 trasitions */
+	if ((vgpu_cfg_space(vgpu)[INTEL_GVT_PCI_SWSCI]
+				& SWSCI_SCI_TRIGGER) ||
+			!(swsci & SWSCI_SCI_TRIGGER)) {
+		return 0;
+	}
+
+	func = GVT_OPREGION_FUNC(scic);
+	subfunc = GVT_OPREGION_SUBFUNC(scic);
+	if (!querying_capabilities(scic)) {
+		gvt_vgpu_err("requesting runtime service: func \"%s\","
+				" subfunc \"%s\"\n",
+				opregion_func_name(func),
+				opregion_subfunc_name(subfunc));
+		/*
+		 * emulate exit status of function call, '0' means
+		 * "failure, generic, unsupported or unknown cause"
+		 */
+		scic &= ~OPREGION_SCIC_EXIT_MASK;
+		goto out;
+	}
+
+	scic = 0;
+	parm = 0;
+
+out:
+	ret = intel_gvt_write_gpa(vgpu, scic_pa, &scic, sizeof(scic));
+	if (ret) {
+		gvt_vgpu_err("guest opregion write error %d, gpa 0x%llx, len %lu\n",
+			ret, scic_pa, sizeof(scic));
+		return ret;
+	}
+
+	ret = intel_gvt_write_gpa(vgpu, parm_pa, &parm, sizeof(parm));
+	if (ret) {
+		gvt_vgpu_err("guest opregion write error %d, gpa 0x%llx, len %lu\n",
+			ret, scic_pa, sizeof(scic));
+		return ret;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/gvt/page_track.c b/drivers/gpu/drm/i915/gvt/page_track.c
new file mode 100644
index 000000000..3375b51c7
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/page_track.c
@@ -0,0 +1,185 @@
+/*
+ * Copyright(c) 2011-2017 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#include "i915_drv.h"
+#include "gvt.h"
+
+/**
+ * intel_vgpu_find_page_track - find page track rcord of guest page
+ * @vgpu: a vGPU
+ * @gfn: the gfn of guest page
+ *
+ * Returns:
+ * A pointer to struct intel_vgpu_page_track if found, else NULL returned.
+ */
+struct intel_vgpu_page_track *intel_vgpu_find_page_track(
+		struct intel_vgpu *vgpu, unsigned long gfn)
+{
+	return radix_tree_lookup(&vgpu->page_track_tree, gfn);
+}
+
+/**
+ * intel_vgpu_register_page_track - register a guest page to be tacked
+ * @vgpu: a vGPU
+ * @gfn: the gfn of guest page
+ * @handler: page track handler
+ * @priv: tracker private
+ *
+ * Returns:
+ * zero on success, negative error code if failed.
+ */
+int intel_vgpu_register_page_track(struct intel_vgpu *vgpu, unsigned long gfn,
+		gvt_page_track_handler_t handler, void *priv)
+{
+	struct intel_vgpu_page_track *track;
+	int ret;
+
+	track = intel_vgpu_find_page_track(vgpu, gfn);
+	if (track)
+		return -EEXIST;
+
+	track = kzalloc(sizeof(*track), GFP_KERNEL);
+	if (!track)
+		return -ENOMEM;
+
+	track->handler = handler;
+	track->priv_data = priv;
+
+	ret = radix_tree_insert(&vgpu->page_track_tree, gfn, track);
+	if (ret) {
+		kfree(track);
+		return ret;
+	}
+
+	return 0;
+}
+
+/**
+ * intel_vgpu_unregister_page_track - unregister the tracked guest page
+ * @vgpu: a vGPU
+ * @gfn: the gfn of guest page
+ *
+ */
+void intel_vgpu_unregister_page_track(struct intel_vgpu *vgpu,
+		unsigned long gfn)
+{
+	struct intel_vgpu_page_track *track;
+
+	track = radix_tree_delete(&vgpu->page_track_tree, gfn);
+	if (track) {
+		if (track->tracked)
+			intel_gvt_page_track_remove(vgpu, gfn);
+		kfree(track);
+	}
+}
+
+/**
+ * intel_vgpu_enable_page_track - set write-protection on guest page
+ * @vgpu: a vGPU
+ * @gfn: the gfn of guest page
+ *
+ * Returns:
+ * zero on success, negative error code if failed.
+ */
+int intel_vgpu_enable_page_track(struct intel_vgpu *vgpu, unsigned long gfn)
+{
+	struct intel_vgpu_page_track *track;
+	int ret;
+
+	track = intel_vgpu_find_page_track(vgpu, gfn);
+	if (!track)
+		return -ENXIO;
+
+	if (track->tracked)
+		return 0;
+
+	ret = intel_gvt_page_track_add(vgpu, gfn);
+	if (ret)
+		return ret;
+	track->tracked = true;
+	return 0;
+}
+
+/**
+ * intel_vgpu_enable_page_track - cancel write-protection on guest page
+ * @vgpu: a vGPU
+ * @gfn: the gfn of guest page
+ *
+ * Returns:
+ * zero on success, negative error code if failed.
+ */
+int intel_vgpu_disable_page_track(struct intel_vgpu *vgpu, unsigned long gfn)
+{
+	struct intel_vgpu_page_track *track;
+	int ret;
+
+	track = intel_vgpu_find_page_track(vgpu, gfn);
+	if (!track)
+		return -ENXIO;
+
+	if (!track->tracked)
+		return 0;
+
+	ret = intel_gvt_page_track_remove(vgpu, gfn);
+	if (ret)
+		return ret;
+	track->tracked = false;
+	return 0;
+}
+
+/**
+ * intel_vgpu_page_track_handler - called when write to write-protected page
+ * @vgpu: a vGPU
+ * @gpa: the gpa of this write
+ * @data: the writed data
+ * @bytes: the length of this write
+ *
+ * Returns:
+ * zero on success, negative error code if failed.
+ */
+int intel_vgpu_page_track_handler(struct intel_vgpu *vgpu, u64 gpa,
+		void *data, unsigned int bytes)
+{
+	struct intel_vgpu_page_track *page_track;
+	int ret = 0;
+
+	mutex_lock(&vgpu->vgpu_lock);
+
+	page_track = intel_vgpu_find_page_track(vgpu, gpa >> PAGE_SHIFT);
+	if (!page_track) {
+		ret = -ENXIO;
+		goto out;
+	}
+
+	if (unlikely(vgpu->failsafe)) {
+		/* Remove write protection to prevent furture traps. */
+		intel_gvt_page_track_remove(vgpu, gpa >> PAGE_SHIFT);
+	} else {
+		ret = page_track->handler(page_track, gpa, data, bytes);
+		if (ret)
+			gvt_err("guest page write error, gpa %llx\n", gpa);
+	}
+
+out:
+	mutex_unlock(&vgpu->vgpu_lock);
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/gvt/page_track.h b/drivers/gpu/drm/i915/gvt/page_track.h
new file mode 100644
index 000000000..f6eb71355
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/page_track.h
@@ -0,0 +1,59 @@
+/*
+ * Copyright(c) 2011-2017 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ */
+
+#ifndef _GVT_PAGE_TRACK_H_
+#define _GVT_PAGE_TRACK_H_
+
+#include <linux/types.h>
+
+struct intel_vgpu;
+struct intel_vgpu_page_track;
+
+typedef int (*gvt_page_track_handler_t)(
+			struct intel_vgpu_page_track *page_track,
+			u64 gpa, void *data, int bytes);
+
+/* Track record for a write-protected guest page. */
+struct intel_vgpu_page_track {
+	gvt_page_track_handler_t handler;
+	bool tracked;
+	void *priv_data;
+};
+
+struct intel_vgpu_page_track *intel_vgpu_find_page_track(
+		struct intel_vgpu *vgpu, unsigned long gfn);
+
+int intel_vgpu_register_page_track(struct intel_vgpu *vgpu,
+		unsigned long gfn, gvt_page_track_handler_t handler,
+		void *priv);
+void intel_vgpu_unregister_page_track(struct intel_vgpu *vgpu,
+		unsigned long gfn);
+
+int intel_vgpu_enable_page_track(struct intel_vgpu *vgpu, unsigned long gfn);
+int intel_vgpu_disable_page_track(struct intel_vgpu *vgpu, unsigned long gfn);
+
+int intel_vgpu_page_track_handler(struct intel_vgpu *vgpu, u64 gpa,
+		void *data, unsigned int bytes);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/reg.h b/drivers/gpu/drm/i915/gvt/reg.h
new file mode 100644
index 000000000..d8216c63c
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/reg.h
@@ -0,0 +1,144 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef _GVT_REG_H
+#define _GVT_REG_H
+
+#define INTEL_GVT_PCI_CLASS_VGA_OTHER   0x80
+
+#define INTEL_GVT_PCI_GMCH_CONTROL	0x50
+#define   BDW_GMCH_GMS_SHIFT		8
+#define   BDW_GMCH_GMS_MASK		0xff
+
+#define INTEL_GVT_PCI_SWSCI		0xe8
+#define   SWSCI_SCI_SELECT		(1 << 15)
+#define   SWSCI_SCI_TRIGGER		1
+
+#define INTEL_GVT_PCI_OPREGION		0xfc
+
+#define INTEL_GVT_OPREGION_CLID		0x1AC
+#define INTEL_GVT_OPREGION_SCIC		0x200
+#define   OPREGION_SCIC_FUNC_MASK	0x1E
+#define   OPREGION_SCIC_FUNC_SHIFT	1
+#define   OPREGION_SCIC_SUBFUNC_MASK	0xFF00
+#define   OPREGION_SCIC_SUBFUNC_SHIFT	8
+#define   OPREGION_SCIC_EXIT_MASK	0xE0
+#define INTEL_GVT_OPREGION_SCIC_F_GETBIOSDATA         4
+#define INTEL_GVT_OPREGION_SCIC_F_GETBIOSCALLBACKS    6
+#define INTEL_GVT_OPREGION_SCIC_SF_SUPPRTEDCALLS      0
+#define INTEL_GVT_OPREGION_SCIC_SF_REQEUSTEDCALLBACKS 1
+#define INTEL_GVT_OPREGION_PARM                   0x204
+
+#define INTEL_GVT_OPREGION_PAGES	2
+#define INTEL_GVT_OPREGION_SIZE		(INTEL_GVT_OPREGION_PAGES * PAGE_SIZE)
+#define INTEL_GVT_OPREGION_VBT_OFFSET	0x400
+#define INTEL_GVT_OPREGION_VBT_SIZE	\
+		(INTEL_GVT_OPREGION_SIZE - INTEL_GVT_OPREGION_VBT_OFFSET)
+
+#define VGT_SPRSTRIDE(pipe)	_PIPE(pipe, _SPRA_STRIDE, _PLANE_STRIDE_2_B)
+
+#define _REG_701C0(pipe, plane) (0x701c0 + pipe * 0x1000 + (plane - 1) * 0x100)
+#define _REG_701C4(pipe, plane) (0x701c4 + pipe * 0x1000 + (plane - 1) * 0x100)
+
+#define SKL_FLIP_EVENT(pipe, plane) (PRIMARY_A_FLIP_DONE + (plane) * 3 + (pipe))
+
+#define REG50080_FLIP_TYPE_MASK	0x3
+#define REG50080_FLIP_TYPE_ASYNC	0x1
+
+#define REG_50080(_pipe, _plane) ({ \
+	typeof(_pipe) (p) = (_pipe); \
+	typeof(_plane) (q) = (_plane); \
+	(((p) == PIPE_A) ? (((q) == PLANE_PRIMARY) ? (_MMIO(0x50080)) : \
+		(_MMIO(0x50090))) : \
+	(((p) == PIPE_B) ? (((q) == PLANE_PRIMARY) ? (_MMIO(0x50088)) : \
+		(_MMIO(0x50098))) : \
+	(((p) == PIPE_C) ? (((q) == PLANE_PRIMARY) ? (_MMIO(0x5008C)) : \
+		(_MMIO(0x5009C))) : \
+		(_MMIO(0x50080))))); })
+
+#define REG_50080_TO_PIPE(_reg) ({ \
+	typeof(_reg) (reg) = (_reg); \
+	(((reg) == 0x50080 || (reg) == 0x50090) ? (PIPE_A) : \
+	(((reg) == 0x50088 || (reg) == 0x50098) ? (PIPE_B) : \
+	(((reg) == 0x5008C || (reg) == 0x5009C) ? (PIPE_C) : \
+	(INVALID_PIPE)))); })
+
+#define REG_50080_TO_PLANE(_reg) ({ \
+	typeof(_reg) (reg) = (_reg); \
+	(((reg) == 0x50080 || (reg) == 0x50088 || (reg) == 0x5008C) ? \
+		(PLANE_PRIMARY) : \
+	(((reg) == 0x50090 || (reg) == 0x50098 || (reg) == 0x5009C) ? \
+		(PLANE_SPRITE0) : (I915_MAX_PLANES))); })
+
+#define GFX_MODE_BIT_SET_IN_MASK(val, bit) \
+		((((bit) & 0xffff0000) == 0) && !!((val) & (((bit) << 16))))
+
+#define IS_MASKED_BITS_ENABLED(_val, _b) \
+		(((_val) & _MASKED_BIT_ENABLE(_b)) == _MASKED_BIT_ENABLE(_b))
+#define IS_MASKED_BITS_DISABLED(_val, _b) \
+		((_val) & _MASKED_BIT_DISABLE(_b))
+
+#define FORCEWAKE_RENDER_GEN9_REG 0xa278
+#define FORCEWAKE_ACK_RENDER_GEN9_REG 0x0D84
+#define FORCEWAKE_GT_GEN9_REG 0xa188
+#define FORCEWAKE_ACK_GT_GEN9_REG 0x130044
+#define FORCEWAKE_MEDIA_GEN9_REG 0xa270
+#define FORCEWAKE_ACK_MEDIA_GEN9_REG 0x0D88
+#define FORCEWAKE_ACK_HSW_REG 0x130044
+
+#define RB_HEAD_WRAP_CNT_MAX	((1 << 11) - 1)
+#define RB_HEAD_WRAP_CNT_OFF	21
+#define RB_HEAD_OFF_MASK	((1U << 21) - (1U << 2))
+#define RB_TAIL_OFF_MASK	((1U << 21) - (1U << 3))
+#define RB_TAIL_SIZE_MASK	((1U << 21) - (1U << 12))
+#define _RING_CTL_BUF_SIZE(ctl) (((ctl) & RB_TAIL_SIZE_MASK) + \
+		I915_GTT_PAGE_SIZE)
+
+#define PCH_GPIO_BASE	_MMIO(0xc5010)
+
+#define PCH_GMBUS0	_MMIO(0xc5100)
+#define PCH_GMBUS1	_MMIO(0xc5104)
+#define PCH_GMBUS2	_MMIO(0xc5108)
+#define PCH_GMBUS3	_MMIO(0xc510c)
+#define PCH_GMBUS4	_MMIO(0xc5110)
+#define PCH_GMBUS5	_MMIO(0xc5120)
+
+#define TRVATTL3PTRDW(i)	_MMIO(0x4de0 + (i) * 4)
+#define TRNULLDETCT		_MMIO(0x4de8)
+#define TRINVTILEDETCT		_MMIO(0x4dec)
+#define TRVADR			_MMIO(0x4df0)
+#define TRTTE			_MMIO(0x4df4)
+#define RING_EXCC(base)		_MMIO((base) + 0x28)
+#define RING_GFX_MODE(base)	_MMIO((base) + 0x29c)
+#define VF_GUARDBAND		_MMIO(0x83a4)
+
+#define BCS_TILE_REGISTER_VAL_OFFSET (0x43*4)
+
+/* XXX FIXME i915 has changed PP_XXX definition */
+#define PCH_PP_STATUS  _MMIO(0xc7200)
+#define PCH_PP_CONTROL _MMIO(0xc7204)
+#define PCH_PP_ON_DELAYS _MMIO(0xc7208)
+#define PCH_PP_OFF_DELAYS _MMIO(0xc720c)
+#define PCH_PP_DIVISOR _MMIO(0xc7210)
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/sched_policy.c b/drivers/gpu/drm/i915/gvt/sched_policy.c
new file mode 100644
index 000000000..c077fb467
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/sched_policy.c
@@ -0,0 +1,480 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Anhua Xu
+ *    Kevin Tian <kevin.tian@intel.com>
+ *
+ * Contributors:
+ *    Min He <min.he@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#include "i915_drv.h"
+#include "gvt.h"
+
+static bool vgpu_has_pending_workload(struct intel_vgpu *vgpu)
+{
+	enum intel_engine_id i;
+	struct intel_engine_cs *engine;
+
+	for_each_engine(engine, vgpu->gvt->gt, i) {
+		if (!list_empty(workload_q_head(vgpu, engine)))
+			return true;
+	}
+
+	return false;
+}
+
+/* We give 2 seconds higher prio for vGPU during start */
+#define GVT_SCHED_VGPU_PRI_TIME  2
+
+struct vgpu_sched_data {
+	struct list_head lru_list;
+	struct intel_vgpu *vgpu;
+	bool active;
+	bool pri_sched;
+	ktime_t pri_time;
+	ktime_t sched_in_time;
+	ktime_t sched_time;
+	ktime_t left_ts;
+	ktime_t allocated_ts;
+
+	struct vgpu_sched_ctl sched_ctl;
+};
+
+struct gvt_sched_data {
+	struct intel_gvt *gvt;
+	struct hrtimer timer;
+	unsigned long period;
+	struct list_head lru_runq_head;
+	ktime_t expire_time;
+};
+
+static void vgpu_update_timeslice(struct intel_vgpu *vgpu, ktime_t cur_time)
+{
+	ktime_t delta_ts;
+	struct vgpu_sched_data *vgpu_data;
+
+	if (!vgpu || vgpu == vgpu->gvt->idle_vgpu)
+		return;
+
+	vgpu_data = vgpu->sched_data;
+	delta_ts = ktime_sub(cur_time, vgpu_data->sched_in_time);
+	vgpu_data->sched_time = ktime_add(vgpu_data->sched_time, delta_ts);
+	vgpu_data->left_ts = ktime_sub(vgpu_data->left_ts, delta_ts);
+	vgpu_data->sched_in_time = cur_time;
+}
+
+#define GVT_TS_BALANCE_PERIOD_MS 100
+#define GVT_TS_BALANCE_STAGE_NUM 10
+
+static void gvt_balance_timeslice(struct gvt_sched_data *sched_data)
+{
+	struct vgpu_sched_data *vgpu_data;
+	struct list_head *pos;
+	static u64 stage_check;
+	int stage = stage_check++ % GVT_TS_BALANCE_STAGE_NUM;
+
+	/* The timeslice accumulation reset at stage 0, which is
+	 * allocated again without adding previous debt.
+	 */
+	if (stage == 0) {
+		int total_weight = 0;
+		ktime_t fair_timeslice;
+
+		list_for_each(pos, &sched_data->lru_runq_head) {
+			vgpu_data = container_of(pos, struct vgpu_sched_data, lru_list);
+			total_weight += vgpu_data->sched_ctl.weight;
+		}
+
+		list_for_each(pos, &sched_data->lru_runq_head) {
+			vgpu_data = container_of(pos, struct vgpu_sched_data, lru_list);
+			fair_timeslice = ktime_divns(ms_to_ktime(GVT_TS_BALANCE_PERIOD_MS),
+						     total_weight) * vgpu_data->sched_ctl.weight;
+
+			vgpu_data->allocated_ts = fair_timeslice;
+			vgpu_data->left_ts = vgpu_data->allocated_ts;
+		}
+	} else {
+		list_for_each(pos, &sched_data->lru_runq_head) {
+			vgpu_data = container_of(pos, struct vgpu_sched_data, lru_list);
+
+			/* timeslice for next 100ms should add the left/debt
+			 * slice of previous stages.
+			 */
+			vgpu_data->left_ts += vgpu_data->allocated_ts;
+		}
+	}
+}
+
+static void try_to_schedule_next_vgpu(struct intel_gvt *gvt)
+{
+	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
+	enum intel_engine_id i;
+	struct intel_engine_cs *engine;
+	struct vgpu_sched_data *vgpu_data;
+	ktime_t cur_time;
+
+	/* no need to schedule if next_vgpu is the same with current_vgpu,
+	 * let scheduler chose next_vgpu again by setting it to NULL.
+	 */
+	if (scheduler->next_vgpu == scheduler->current_vgpu) {
+		scheduler->next_vgpu = NULL;
+		return;
+	}
+
+	/*
+	 * after the flag is set, workload dispatch thread will
+	 * stop dispatching workload for current vgpu
+	 */
+	scheduler->need_reschedule = true;
+
+	/* still have uncompleted workload? */
+	for_each_engine(engine, gvt->gt, i) {
+		if (scheduler->current_workload[engine->id])
+			return;
+	}
+
+	cur_time = ktime_get();
+	vgpu_update_timeslice(scheduler->current_vgpu, cur_time);
+	vgpu_data = scheduler->next_vgpu->sched_data;
+	vgpu_data->sched_in_time = cur_time;
+
+	/* switch current vgpu */
+	scheduler->current_vgpu = scheduler->next_vgpu;
+	scheduler->next_vgpu = NULL;
+
+	scheduler->need_reschedule = false;
+
+	/* wake up workload dispatch thread */
+	for_each_engine(engine, gvt->gt, i)
+		wake_up(&scheduler->waitq[engine->id]);
+}
+
+static struct intel_vgpu *find_busy_vgpu(struct gvt_sched_data *sched_data)
+{
+	struct vgpu_sched_data *vgpu_data;
+	struct intel_vgpu *vgpu = NULL;
+	struct list_head *head = &sched_data->lru_runq_head;
+	struct list_head *pos;
+
+	/* search a vgpu with pending workload */
+	list_for_each(pos, head) {
+
+		vgpu_data = container_of(pos, struct vgpu_sched_data, lru_list);
+		if (!vgpu_has_pending_workload(vgpu_data->vgpu))
+			continue;
+
+		if (vgpu_data->pri_sched) {
+			if (ktime_before(ktime_get(), vgpu_data->pri_time)) {
+				vgpu = vgpu_data->vgpu;
+				break;
+			} else
+				vgpu_data->pri_sched = false;
+		}
+
+		/* Return the vGPU only if it has time slice left */
+		if (vgpu_data->left_ts > 0) {
+			vgpu = vgpu_data->vgpu;
+			break;
+		}
+	}
+
+	return vgpu;
+}
+
+/* in nanosecond */
+#define GVT_DEFAULT_TIME_SLICE 1000000
+
+static void tbs_sched_func(struct gvt_sched_data *sched_data)
+{
+	struct intel_gvt *gvt = sched_data->gvt;
+	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
+	struct vgpu_sched_data *vgpu_data;
+	struct intel_vgpu *vgpu = NULL;
+
+	/* no active vgpu or has already had a target */
+	if (list_empty(&sched_data->lru_runq_head) || scheduler->next_vgpu)
+		goto out;
+
+	vgpu = find_busy_vgpu(sched_data);
+	if (vgpu) {
+		scheduler->next_vgpu = vgpu;
+		vgpu_data = vgpu->sched_data;
+		if (!vgpu_data->pri_sched) {
+			/* Move the last used vGPU to the tail of lru_list */
+			list_del_init(&vgpu_data->lru_list);
+			list_add_tail(&vgpu_data->lru_list,
+				      &sched_data->lru_runq_head);
+		}
+	} else {
+		scheduler->next_vgpu = gvt->idle_vgpu;
+	}
+out:
+	if (scheduler->next_vgpu)
+		try_to_schedule_next_vgpu(gvt);
+}
+
+void intel_gvt_schedule(struct intel_gvt *gvt)
+{
+	struct gvt_sched_data *sched_data = gvt->scheduler.sched_data;
+	ktime_t cur_time;
+
+	mutex_lock(&gvt->sched_lock);
+	cur_time = ktime_get();
+
+	if (test_and_clear_bit(INTEL_GVT_REQUEST_SCHED,
+				(void *)&gvt->service_request)) {
+		if (cur_time >= sched_data->expire_time) {
+			gvt_balance_timeslice(sched_data);
+			sched_data->expire_time = ktime_add_ms(
+				cur_time, GVT_TS_BALANCE_PERIOD_MS);
+		}
+	}
+	clear_bit(INTEL_GVT_REQUEST_EVENT_SCHED, (void *)&gvt->service_request);
+
+	vgpu_update_timeslice(gvt->scheduler.current_vgpu, cur_time);
+	tbs_sched_func(sched_data);
+
+	mutex_unlock(&gvt->sched_lock);
+}
+
+static enum hrtimer_restart tbs_timer_fn(struct hrtimer *timer_data)
+{
+	struct gvt_sched_data *data;
+
+	data = container_of(timer_data, struct gvt_sched_data, timer);
+
+	intel_gvt_request_service(data->gvt, INTEL_GVT_REQUEST_SCHED);
+
+	hrtimer_add_expires_ns(&data->timer, data->period);
+
+	return HRTIMER_RESTART;
+}
+
+static int tbs_sched_init(struct intel_gvt *gvt)
+{
+	struct intel_gvt_workload_scheduler *scheduler =
+		&gvt->scheduler;
+
+	struct gvt_sched_data *data;
+
+	data = kzalloc(sizeof(*data), GFP_KERNEL);
+	if (!data)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&data->lru_runq_head);
+	hrtimer_init(&data->timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS);
+	data->timer.function = tbs_timer_fn;
+	data->period = GVT_DEFAULT_TIME_SLICE;
+	data->gvt = gvt;
+
+	scheduler->sched_data = data;
+
+	return 0;
+}
+
+static void tbs_sched_clean(struct intel_gvt *gvt)
+{
+	struct intel_gvt_workload_scheduler *scheduler =
+		&gvt->scheduler;
+	struct gvt_sched_data *data = scheduler->sched_data;
+
+	hrtimer_cancel(&data->timer);
+
+	kfree(data);
+	scheduler->sched_data = NULL;
+}
+
+static int tbs_sched_init_vgpu(struct intel_vgpu *vgpu)
+{
+	struct vgpu_sched_data *data;
+
+	data = kzalloc(sizeof(*data), GFP_KERNEL);
+	if (!data)
+		return -ENOMEM;
+
+	data->sched_ctl.weight = vgpu->sched_ctl.weight;
+	data->vgpu = vgpu;
+	INIT_LIST_HEAD(&data->lru_list);
+
+	vgpu->sched_data = data;
+
+	return 0;
+}
+
+static void tbs_sched_clean_vgpu(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct gvt_sched_data *sched_data = gvt->scheduler.sched_data;
+
+	kfree(vgpu->sched_data);
+	vgpu->sched_data = NULL;
+
+	/* this vgpu id has been removed */
+	if (idr_is_empty(&gvt->vgpu_idr))
+		hrtimer_cancel(&sched_data->timer);
+}
+
+static void tbs_sched_start_schedule(struct intel_vgpu *vgpu)
+{
+	struct gvt_sched_data *sched_data = vgpu->gvt->scheduler.sched_data;
+	struct vgpu_sched_data *vgpu_data = vgpu->sched_data;
+	ktime_t now;
+
+	if (!list_empty(&vgpu_data->lru_list))
+		return;
+
+	now = ktime_get();
+	vgpu_data->pri_time = ktime_add(now,
+					ktime_set(GVT_SCHED_VGPU_PRI_TIME, 0));
+	vgpu_data->pri_sched = true;
+
+	list_add(&vgpu_data->lru_list, &sched_data->lru_runq_head);
+
+	if (!hrtimer_active(&sched_data->timer))
+		hrtimer_start(&sched_data->timer, ktime_add_ns(ktime_get(),
+			sched_data->period), HRTIMER_MODE_ABS);
+	vgpu_data->active = true;
+}
+
+static void tbs_sched_stop_schedule(struct intel_vgpu *vgpu)
+{
+	struct vgpu_sched_data *vgpu_data = vgpu->sched_data;
+
+	list_del_init(&vgpu_data->lru_list);
+	vgpu_data->active = false;
+}
+
+static const struct intel_gvt_sched_policy_ops tbs_schedule_ops = {
+	.init = tbs_sched_init,
+	.clean = tbs_sched_clean,
+	.init_vgpu = tbs_sched_init_vgpu,
+	.clean_vgpu = tbs_sched_clean_vgpu,
+	.start_schedule = tbs_sched_start_schedule,
+	.stop_schedule = tbs_sched_stop_schedule,
+};
+
+int intel_gvt_init_sched_policy(struct intel_gvt *gvt)
+{
+	int ret;
+
+	mutex_lock(&gvt->sched_lock);
+	gvt->scheduler.sched_ops = &tbs_schedule_ops;
+	ret = gvt->scheduler.sched_ops->init(gvt);
+	mutex_unlock(&gvt->sched_lock);
+
+	return ret;
+}
+
+void intel_gvt_clean_sched_policy(struct intel_gvt *gvt)
+{
+	mutex_lock(&gvt->sched_lock);
+	gvt->scheduler.sched_ops->clean(gvt);
+	mutex_unlock(&gvt->sched_lock);
+}
+
+/* for per-vgpu scheduler policy, there are 2 per-vgpu data:
+ * sched_data, and sched_ctl. We see these 2 data as part of
+ * the global scheduler which are proteced by gvt->sched_lock.
+ * Caller should make their decision if the vgpu_lock should
+ * be hold outside.
+ */
+
+int intel_vgpu_init_sched_policy(struct intel_vgpu *vgpu)
+{
+	int ret;
+
+	mutex_lock(&vgpu->gvt->sched_lock);
+	ret = vgpu->gvt->scheduler.sched_ops->init_vgpu(vgpu);
+	mutex_unlock(&vgpu->gvt->sched_lock);
+
+	return ret;
+}
+
+void intel_vgpu_clean_sched_policy(struct intel_vgpu *vgpu)
+{
+	mutex_lock(&vgpu->gvt->sched_lock);
+	vgpu->gvt->scheduler.sched_ops->clean_vgpu(vgpu);
+	mutex_unlock(&vgpu->gvt->sched_lock);
+}
+
+void intel_vgpu_start_schedule(struct intel_vgpu *vgpu)
+{
+	struct vgpu_sched_data *vgpu_data = vgpu->sched_data;
+
+	mutex_lock(&vgpu->gvt->sched_lock);
+	if (!vgpu_data->active) {
+		gvt_dbg_core("vgpu%d: start schedule\n", vgpu->id);
+		vgpu->gvt->scheduler.sched_ops->start_schedule(vgpu);
+	}
+	mutex_unlock(&vgpu->gvt->sched_lock);
+}
+
+void intel_gvt_kick_schedule(struct intel_gvt *gvt)
+{
+	mutex_lock(&gvt->sched_lock);
+	intel_gvt_request_service(gvt, INTEL_GVT_REQUEST_EVENT_SCHED);
+	mutex_unlock(&gvt->sched_lock);
+}
+
+void intel_vgpu_stop_schedule(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt_workload_scheduler *scheduler =
+		&vgpu->gvt->scheduler;
+	struct vgpu_sched_data *vgpu_data = vgpu->sched_data;
+	struct drm_i915_private *dev_priv = vgpu->gvt->gt->i915;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	if (!vgpu_data->active)
+		return;
+
+	gvt_dbg_core("vgpu%d: stop schedule\n", vgpu->id);
+
+	mutex_lock(&vgpu->gvt->sched_lock);
+	scheduler->sched_ops->stop_schedule(vgpu);
+
+	if (scheduler->next_vgpu == vgpu)
+		scheduler->next_vgpu = NULL;
+
+	if (scheduler->current_vgpu == vgpu) {
+		/* stop workload dispatching */
+		scheduler->need_reschedule = true;
+		scheduler->current_vgpu = NULL;
+	}
+
+	intel_runtime_pm_get(&dev_priv->runtime_pm);
+	spin_lock_bh(&scheduler->mmio_context_lock);
+	for_each_engine(engine, vgpu->gvt->gt, id) {
+		if (scheduler->engine_owner[engine->id] == vgpu) {
+			intel_gvt_switch_mmio(vgpu, NULL, engine);
+			scheduler->engine_owner[engine->id] = NULL;
+		}
+	}
+	spin_unlock_bh(&scheduler->mmio_context_lock);
+	intel_runtime_pm_put_unchecked(&dev_priv->runtime_pm);
+	mutex_unlock(&vgpu->gvt->sched_lock);
+}
diff --git a/drivers/gpu/drm/i915/gvt/sched_policy.h b/drivers/gpu/drm/i915/gvt/sched_policy.h
new file mode 100644
index 000000000..3dacdad5f
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/sched_policy.h
@@ -0,0 +1,65 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Anhua Xu
+ *    Kevin Tian <kevin.tian@intel.com>
+ *
+ * Contributors:
+ *    Min He <min.he@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#ifndef __GVT_SCHED_POLICY__
+#define __GVT_SCHED_POLICY__
+
+struct intel_gvt;
+struct intel_vgpu;
+
+struct intel_gvt_sched_policy_ops {
+	int (*init)(struct intel_gvt *gvt);
+	void (*clean)(struct intel_gvt *gvt);
+	int (*init_vgpu)(struct intel_vgpu *vgpu);
+	void (*clean_vgpu)(struct intel_vgpu *vgpu);
+	void (*start_schedule)(struct intel_vgpu *vgpu);
+	void (*stop_schedule)(struct intel_vgpu *vgpu);
+};
+
+void intel_gvt_schedule(struct intel_gvt *gvt);
+
+int intel_gvt_init_sched_policy(struct intel_gvt *gvt);
+
+void intel_gvt_clean_sched_policy(struct intel_gvt *gvt);
+
+int intel_vgpu_init_sched_policy(struct intel_vgpu *vgpu);
+
+void intel_vgpu_clean_sched_policy(struct intel_vgpu *vgpu);
+
+void intel_vgpu_start_schedule(struct intel_vgpu *vgpu);
+
+void intel_vgpu_stop_schedule(struct intel_vgpu *vgpu);
+
+void intel_gvt_kick_schedule(struct intel_gvt *gvt);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c
new file mode 100644
index 000000000..8342d95f5
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/scheduler.c
@@ -0,0 +1,1786 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ * Contributors:
+ *    Ping Gao <ping.a.gao@intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *    Chanbin Du <changbin.du@intel.com>
+ *    Min He <min.he@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *    Zhenyu Wang <zhenyuw@linux.intel.com>
+ *
+ */
+
+#include <linux/kthread.h>
+
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_context.h"
+#include "gt/intel_execlists_submission.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_lrc.h"
+#include "gt/intel_ring.h"
+
+#include "i915_drv.h"
+#include "i915_gem_gtt.h"
+#include "i915_perf_oa_regs.h"
+#include "gvt.h"
+
+#define RING_CTX_OFF(x) \
+	offsetof(struct execlist_ring_context, x)
+
+static void set_context_pdp_root_pointer(
+		struct execlist_ring_context *ring_context,
+		u32 pdp[8])
+{
+	int i;
+
+	for (i = 0; i < 8; i++)
+		ring_context->pdps[i].val = pdp[7 - i];
+}
+
+static void update_shadow_pdps(struct intel_vgpu_workload *workload)
+{
+	struct execlist_ring_context *shadow_ring_context;
+	struct intel_context *ctx = workload->req->context;
+
+	if (WARN_ON(!workload->shadow_mm))
+		return;
+
+	if (WARN_ON(!atomic_read(&workload->shadow_mm->pincount)))
+		return;
+
+	shadow_ring_context = (struct execlist_ring_context *)ctx->lrc_reg_state;
+	set_context_pdp_root_pointer(shadow_ring_context,
+			(void *)workload->shadow_mm->ppgtt_mm.shadow_pdps);
+}
+
+/*
+ * when populating shadow ctx from guest, we should not overrride oa related
+ * registers, so that they will not be overlapped by guest oa configs. Thus
+ * made it possible to capture oa data from host for both host and guests.
+ */
+static void sr_oa_regs(struct intel_vgpu_workload *workload,
+		u32 *reg_state, bool save)
+{
+	struct drm_i915_private *dev_priv = workload->vgpu->gvt->gt->i915;
+	u32 ctx_oactxctrl = dev_priv->perf.ctx_oactxctrl_offset;
+	u32 ctx_flexeu0 = dev_priv->perf.ctx_flexeu0_offset;
+	int i = 0;
+	u32 flex_mmio[] = {
+		i915_mmio_reg_offset(EU_PERF_CNTL0),
+		i915_mmio_reg_offset(EU_PERF_CNTL1),
+		i915_mmio_reg_offset(EU_PERF_CNTL2),
+		i915_mmio_reg_offset(EU_PERF_CNTL3),
+		i915_mmio_reg_offset(EU_PERF_CNTL4),
+		i915_mmio_reg_offset(EU_PERF_CNTL5),
+		i915_mmio_reg_offset(EU_PERF_CNTL6),
+	};
+
+	if (workload->engine->id != RCS0)
+		return;
+
+	if (save) {
+		workload->oactxctrl = reg_state[ctx_oactxctrl + 1];
+
+		for (i = 0; i < ARRAY_SIZE(workload->flex_mmio); i++) {
+			u32 state_offset = ctx_flexeu0 + i * 2;
+
+			workload->flex_mmio[i] = reg_state[state_offset + 1];
+		}
+	} else {
+		reg_state[ctx_oactxctrl] =
+			i915_mmio_reg_offset(GEN8_OACTXCONTROL);
+		reg_state[ctx_oactxctrl + 1] = workload->oactxctrl;
+
+		for (i = 0; i < ARRAY_SIZE(workload->flex_mmio); i++) {
+			u32 state_offset = ctx_flexeu0 + i * 2;
+			u32 mmio = flex_mmio[i];
+
+			reg_state[state_offset] = mmio;
+			reg_state[state_offset + 1] = workload->flex_mmio[i];
+		}
+	}
+}
+
+static int populate_shadow_context(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_context *ctx = workload->req->context;
+	struct execlist_ring_context *shadow_ring_context;
+	void *dst;
+	void *context_base;
+	unsigned long context_gpa, context_page_num;
+	unsigned long gpa_base; /* first gpa of consecutive GPAs */
+	unsigned long gpa_size; /* size of consecutive GPAs */
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	int i;
+	bool skip = false;
+	int ring_id = workload->engine->id;
+	int ret;
+
+	GEM_BUG_ON(!intel_context_is_pinned(ctx));
+
+	context_base = (void *) ctx->lrc_reg_state -
+				(LRC_STATE_PN << I915_GTT_PAGE_SHIFT);
+
+	shadow_ring_context = (void *) ctx->lrc_reg_state;
+
+	sr_oa_regs(workload, (u32 *)shadow_ring_context, true);
+#define COPY_REG(name) \
+	intel_gvt_read_gpa(vgpu, workload->ring_context_gpa \
+		+ RING_CTX_OFF(name.val), &shadow_ring_context->name.val, 4)
+#define COPY_REG_MASKED(name) {\
+		intel_gvt_read_gpa(vgpu, workload->ring_context_gpa \
+					      + RING_CTX_OFF(name.val),\
+					      &shadow_ring_context->name.val, 4);\
+		shadow_ring_context->name.val |= 0xffff << 16;\
+	}
+
+	COPY_REG_MASKED(ctx_ctrl);
+	COPY_REG(ctx_timestamp);
+
+	if (workload->engine->id == RCS0) {
+		COPY_REG(bb_per_ctx_ptr);
+		COPY_REG(rcs_indirect_ctx);
+		COPY_REG(rcs_indirect_ctx_offset);
+	} else if (workload->engine->id == BCS0)
+		intel_gvt_read_gpa(vgpu,
+				workload->ring_context_gpa +
+				BCS_TILE_REGISTER_VAL_OFFSET,
+				(void *)shadow_ring_context +
+				BCS_TILE_REGISTER_VAL_OFFSET, 4);
+#undef COPY_REG
+#undef COPY_REG_MASKED
+
+	/* don't copy Ring Context (the first 0x50 dwords),
+	 * only copy the Engine Context part from guest
+	 */
+	intel_gvt_read_gpa(vgpu,
+			workload->ring_context_gpa +
+			RING_CTX_SIZE,
+			(void *)shadow_ring_context +
+			RING_CTX_SIZE,
+			I915_GTT_PAGE_SIZE - RING_CTX_SIZE);
+
+	sr_oa_regs(workload, (u32 *)shadow_ring_context, false);
+
+	gvt_dbg_sched("ring %s workload lrca %x, ctx_id %x, ctx gpa %llx",
+			workload->engine->name, workload->ctx_desc.lrca,
+			workload->ctx_desc.context_id,
+			workload->ring_context_gpa);
+
+	/* only need to ensure this context is not pinned/unpinned during the
+	 * period from last submission to this this submission.
+	 * Upon reaching this function, the currently submitted context is not
+	 * supposed to get unpinned. If a misbehaving guest driver ever does
+	 * this, it would corrupt itself.
+	 */
+	if (s->last_ctx[ring_id].valid &&
+			(s->last_ctx[ring_id].lrca ==
+				workload->ctx_desc.lrca) &&
+			(s->last_ctx[ring_id].ring_context_gpa ==
+				workload->ring_context_gpa))
+		skip = true;
+
+	s->last_ctx[ring_id].lrca = workload->ctx_desc.lrca;
+	s->last_ctx[ring_id].ring_context_gpa = workload->ring_context_gpa;
+
+	if (IS_RESTORE_INHIBIT(shadow_ring_context->ctx_ctrl.val) || skip)
+		return 0;
+
+	s->last_ctx[ring_id].valid = false;
+	context_page_num = workload->engine->context_size;
+	context_page_num = context_page_num >> PAGE_SHIFT;
+
+	if (IS_BROADWELL(gvt->gt->i915) && workload->engine->id == RCS0)
+		context_page_num = 19;
+
+	/* find consecutive GPAs from gma until the first inconsecutive GPA.
+	 * read from the continuous GPAs into dst virtual address
+	 */
+	gpa_size = 0;
+	for (i = 2; i < context_page_num; i++) {
+		context_gpa = intel_vgpu_gma_to_gpa(vgpu->gtt.ggtt_mm,
+				(u32)((workload->ctx_desc.lrca + i) <<
+				I915_GTT_PAGE_SHIFT));
+		if (context_gpa == INTEL_GVT_INVALID_ADDR) {
+			gvt_vgpu_err("Invalid guest context descriptor\n");
+			return -EFAULT;
+		}
+
+		if (gpa_size == 0) {
+			gpa_base = context_gpa;
+			dst = context_base + (i << I915_GTT_PAGE_SHIFT);
+		} else if (context_gpa != gpa_base + gpa_size)
+			goto read;
+
+		gpa_size += I915_GTT_PAGE_SIZE;
+
+		if (i == context_page_num - 1)
+			goto read;
+
+		continue;
+
+read:
+		intel_gvt_read_gpa(vgpu, gpa_base, dst, gpa_size);
+		gpa_base = context_gpa;
+		gpa_size = I915_GTT_PAGE_SIZE;
+		dst = context_base + (i << I915_GTT_PAGE_SHIFT);
+	}
+	ret = intel_gvt_scan_engine_context(workload);
+	if (ret) {
+		gvt_vgpu_err("invalid cmd found in guest context pages\n");
+		return ret;
+	}
+	s->last_ctx[ring_id].valid = true;
+	return 0;
+}
+
+static inline bool is_gvt_request(struct i915_request *rq)
+{
+	return intel_context_force_single_submission(rq->context);
+}
+
+static void save_ring_hw_state(struct intel_vgpu *vgpu,
+			       const struct intel_engine_cs *engine)
+{
+	struct intel_uncore *uncore = engine->uncore;
+	i915_reg_t reg;
+
+	reg = RING_INSTDONE(engine->mmio_base);
+	vgpu_vreg(vgpu, i915_mmio_reg_offset(reg)) =
+		intel_uncore_read(uncore, reg);
+
+	reg = RING_ACTHD(engine->mmio_base);
+	vgpu_vreg(vgpu, i915_mmio_reg_offset(reg)) =
+		intel_uncore_read(uncore, reg);
+
+	reg = RING_ACTHD_UDW(engine->mmio_base);
+	vgpu_vreg(vgpu, i915_mmio_reg_offset(reg)) =
+		intel_uncore_read(uncore, reg);
+}
+
+static int shadow_context_status_change(struct notifier_block *nb,
+		unsigned long action, void *data)
+{
+	struct i915_request *rq = data;
+	struct intel_gvt *gvt = container_of(nb, struct intel_gvt,
+				shadow_ctx_notifier_block[rq->engine->id]);
+	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
+	enum intel_engine_id ring_id = rq->engine->id;
+	struct intel_vgpu_workload *workload;
+	unsigned long flags;
+
+	if (!is_gvt_request(rq)) {
+		spin_lock_irqsave(&scheduler->mmio_context_lock, flags);
+		if (action == INTEL_CONTEXT_SCHEDULE_IN &&
+		    scheduler->engine_owner[ring_id]) {
+			/* Switch ring from vGPU to host. */
+			intel_gvt_switch_mmio(scheduler->engine_owner[ring_id],
+					      NULL, rq->engine);
+			scheduler->engine_owner[ring_id] = NULL;
+		}
+		spin_unlock_irqrestore(&scheduler->mmio_context_lock, flags);
+
+		return NOTIFY_OK;
+	}
+
+	workload = scheduler->current_workload[ring_id];
+	if (unlikely(!workload))
+		return NOTIFY_OK;
+
+	switch (action) {
+	case INTEL_CONTEXT_SCHEDULE_IN:
+		spin_lock_irqsave(&scheduler->mmio_context_lock, flags);
+		if (workload->vgpu != scheduler->engine_owner[ring_id]) {
+			/* Switch ring from host to vGPU or vGPU to vGPU. */
+			intel_gvt_switch_mmio(scheduler->engine_owner[ring_id],
+					      workload->vgpu, rq->engine);
+			scheduler->engine_owner[ring_id] = workload->vgpu;
+		} else
+			gvt_dbg_sched("skip ring %d mmio switch for vgpu%d\n",
+				      ring_id, workload->vgpu->id);
+		spin_unlock_irqrestore(&scheduler->mmio_context_lock, flags);
+		atomic_set(&workload->shadow_ctx_active, 1);
+		break;
+	case INTEL_CONTEXT_SCHEDULE_OUT:
+		save_ring_hw_state(workload->vgpu, rq->engine);
+		atomic_set(&workload->shadow_ctx_active, 0);
+		break;
+	case INTEL_CONTEXT_SCHEDULE_PREEMPTED:
+		save_ring_hw_state(workload->vgpu, rq->engine);
+		break;
+	default:
+		WARN_ON(1);
+		return NOTIFY_OK;
+	}
+	wake_up(&workload->shadow_ctx_status_wq);
+	return NOTIFY_OK;
+}
+
+static void
+shadow_context_descriptor_update(struct intel_context *ce,
+				 struct intel_vgpu_workload *workload)
+{
+	u64 desc = ce->lrc.desc;
+
+	/*
+	 * Update bits 0-11 of the context descriptor which includes flags
+	 * like GEN8_CTX_* cached in desc_template
+	 */
+	desc &= ~(0x3ull << GEN8_CTX_ADDRESSING_MODE_SHIFT);
+	desc |= (u64)workload->ctx_desc.addressing_mode <<
+		GEN8_CTX_ADDRESSING_MODE_SHIFT;
+
+	ce->lrc.desc = desc;
+}
+
+static int copy_workload_to_ring_buffer(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct i915_request *req = workload->req;
+	void *shadow_ring_buffer_va;
+	u32 *cs;
+	int err;
+
+	if (GRAPHICS_VER(req->engine->i915) == 9 && is_inhibit_context(req->context))
+		intel_vgpu_restore_inhibit_context(vgpu, req);
+
+	/*
+	 * To track whether a request has started on HW, we can emit a
+	 * breadcrumb at the beginning of the request and check its
+	 * timeline's HWSP to see if the breadcrumb has advanced past the
+	 * start of this request. Actually, the request must have the
+	 * init_breadcrumb if its timeline set has_init_bread_crumb, or the
+	 * scheduler might get a wrong state of it during reset. Since the
+	 * requests from gvt always set the has_init_breadcrumb flag, here
+	 * need to do the emit_init_breadcrumb for all the requests.
+	 */
+	if (req->engine->emit_init_breadcrumb) {
+		err = req->engine->emit_init_breadcrumb(req);
+		if (err) {
+			gvt_vgpu_err("fail to emit init breadcrumb\n");
+			return err;
+		}
+	}
+
+	/* allocate shadow ring buffer */
+	cs = intel_ring_begin(workload->req, workload->rb_len / sizeof(u32));
+	if (IS_ERR(cs)) {
+		gvt_vgpu_err("fail to alloc size =%ld shadow  ring buffer\n",
+			workload->rb_len);
+		return PTR_ERR(cs);
+	}
+
+	shadow_ring_buffer_va = workload->shadow_ring_buffer_va;
+
+	/* get shadow ring buffer va */
+	workload->shadow_ring_buffer_va = cs;
+
+	memcpy(cs, shadow_ring_buffer_va,
+			workload->rb_len);
+
+	cs += workload->rb_len / sizeof(u32);
+	intel_ring_advance(workload->req, cs);
+
+	return 0;
+}
+
+static void release_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
+{
+	if (!wa_ctx->indirect_ctx.obj)
+		return;
+
+	i915_gem_object_lock(wa_ctx->indirect_ctx.obj, NULL);
+	i915_gem_object_unpin_map(wa_ctx->indirect_ctx.obj);
+	i915_gem_object_unlock(wa_ctx->indirect_ctx.obj);
+	i915_gem_object_put(wa_ctx->indirect_ctx.obj);
+
+	wa_ctx->indirect_ctx.obj = NULL;
+	wa_ctx->indirect_ctx.shadow_va = NULL;
+}
+
+static void set_dma_address(struct i915_page_directory *pd, dma_addr_t addr)
+{
+	struct scatterlist *sg = pd->pt.base->mm.pages->sgl;
+
+	/* This is not a good idea */
+	sg->dma_address = addr;
+}
+
+static void set_context_ppgtt_from_shadow(struct intel_vgpu_workload *workload,
+					  struct intel_context *ce)
+{
+	struct intel_vgpu_mm *mm = workload->shadow_mm;
+	struct i915_ppgtt *ppgtt = i915_vm_to_ppgtt(ce->vm);
+	int i = 0;
+
+	if (mm->ppgtt_mm.root_entry_type == GTT_TYPE_PPGTT_ROOT_L4_ENTRY) {
+		set_dma_address(ppgtt->pd, mm->ppgtt_mm.shadow_pdps[0]);
+	} else {
+		for (i = 0; i < GVT_RING_CTX_NR_PDPS; i++) {
+			struct i915_page_directory * const pd =
+				i915_pd_entry(ppgtt->pd, i);
+			/* skip now as current i915 ppgtt alloc won't allocate
+			   top level pdp for non 4-level table, won't impact
+			   shadow ppgtt. */
+			if (!pd)
+				break;
+
+			set_dma_address(pd, mm->ppgtt_mm.shadow_pdps[i]);
+		}
+	}
+}
+
+static int
+intel_gvt_workload_req_alloc(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct i915_request *rq;
+
+	if (workload->req)
+		return 0;
+
+	rq = i915_request_create(s->shadow[workload->engine->id]);
+	if (IS_ERR(rq)) {
+		gvt_vgpu_err("fail to allocate gem request\n");
+		return PTR_ERR(rq);
+	}
+
+	workload->req = i915_request_get(rq);
+	return 0;
+}
+
+/**
+ * intel_gvt_scan_and_shadow_workload - audit the workload by scanning and
+ * shadow it as well, include ringbuffer,wa_ctx and ctx.
+ * @workload: an abstract entity for each execlist submission.
+ *
+ * This function is called before the workload submitting to i915, to make
+ * sure the content of the workload is valid.
+ */
+int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	int ret;
+
+	lockdep_assert_held(&vgpu->vgpu_lock);
+
+	if (workload->shadow)
+		return 0;
+
+	if (!test_and_set_bit(workload->engine->id, s->shadow_ctx_desc_updated))
+		shadow_context_descriptor_update(s->shadow[workload->engine->id],
+						 workload);
+
+	ret = intel_gvt_scan_and_shadow_ringbuffer(workload);
+	if (ret)
+		return ret;
+
+	if (workload->engine->id == RCS0 &&
+	    workload->wa_ctx.indirect_ctx.size) {
+		ret = intel_gvt_scan_and_shadow_wa_ctx(&workload->wa_ctx);
+		if (ret)
+			goto err_shadow;
+	}
+
+	workload->shadow = true;
+	return 0;
+
+err_shadow:
+	release_shadow_wa_ctx(&workload->wa_ctx);
+	return ret;
+}
+
+static void release_shadow_batch_buffer(struct intel_vgpu_workload *workload);
+
+static int prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload)
+{
+	struct intel_gvt *gvt = workload->vgpu->gvt;
+	const int gmadr_bytes = gvt->device_info.gmadr_bytes_in_cmd;
+	struct intel_vgpu_shadow_bb *bb;
+	struct i915_gem_ww_ctx ww;
+	int ret;
+
+	list_for_each_entry(bb, &workload->shadow_bb, list) {
+		/* For privilge batch buffer and not wa_ctx, the bb_start_cmd_va
+		 * is only updated into ring_scan_buffer, not real ring address
+		 * allocated in later copy_workload_to_ring_buffer. pls be noted
+		 * shadow_ring_buffer_va is now pointed to real ring buffer va
+		 * in copy_workload_to_ring_buffer.
+		 */
+
+		if (bb->bb_offset)
+			bb->bb_start_cmd_va = workload->shadow_ring_buffer_va
+				+ bb->bb_offset;
+
+		/*
+		 * For non-priv bb, scan&shadow is only for
+		 * debugging purpose, so the content of shadow bb
+		 * is the same as original bb. Therefore,
+		 * here, rather than switch to shadow bb's gma
+		 * address, we directly use original batch buffer's
+		 * gma address, and send original bb to hardware
+		 * directly
+		 */
+		if (!bb->ppgtt) {
+			i915_gem_ww_ctx_init(&ww, false);
+retry:
+			i915_gem_object_lock(bb->obj, &ww);
+
+			bb->vma = i915_gem_object_ggtt_pin_ww(bb->obj, &ww,
+							      NULL, 0, 0, 0);
+			if (IS_ERR(bb->vma)) {
+				ret = PTR_ERR(bb->vma);
+				if (ret == -EDEADLK) {
+					ret = i915_gem_ww_ctx_backoff(&ww);
+					if (!ret)
+						goto retry;
+				}
+				goto err;
+			}
+
+			/* relocate shadow batch buffer */
+			bb->bb_start_cmd_va[1] = i915_ggtt_offset(bb->vma);
+			if (gmadr_bytes == 8)
+				bb->bb_start_cmd_va[2] = 0;
+
+			ret = i915_vma_move_to_active(bb->vma,
+						      workload->req,
+						      0);
+			if (ret)
+				goto err;
+
+			/* No one is going to touch shadow bb from now on. */
+			i915_gem_object_flush_map(bb->obj);
+			i915_gem_ww_ctx_fini(&ww);
+		}
+	}
+	return 0;
+err:
+	i915_gem_ww_ctx_fini(&ww);
+	release_shadow_batch_buffer(workload);
+	return ret;
+}
+
+static void update_wa_ctx_2_shadow_ctx(struct intel_shadow_wa_ctx *wa_ctx)
+{
+	struct intel_vgpu_workload *workload =
+		container_of(wa_ctx, struct intel_vgpu_workload, wa_ctx);
+	struct i915_request *rq = workload->req;
+	struct execlist_ring_context *shadow_ring_context =
+		(struct execlist_ring_context *)rq->context->lrc_reg_state;
+
+	shadow_ring_context->bb_per_ctx_ptr.val =
+		(shadow_ring_context->bb_per_ctx_ptr.val &
+		(~PER_CTX_ADDR_MASK)) | wa_ctx->per_ctx.shadow_gma;
+	shadow_ring_context->rcs_indirect_ctx.val =
+		(shadow_ring_context->rcs_indirect_ctx.val &
+		(~INDIRECT_CTX_ADDR_MASK)) | wa_ctx->indirect_ctx.shadow_gma;
+}
+
+static int prepare_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
+{
+	struct i915_vma *vma;
+	unsigned char *per_ctx_va =
+		(unsigned char *)wa_ctx->indirect_ctx.shadow_va +
+		wa_ctx->indirect_ctx.size;
+	struct i915_gem_ww_ctx ww;
+	int ret;
+
+	if (wa_ctx->indirect_ctx.size == 0)
+		return 0;
+
+	i915_gem_ww_ctx_init(&ww, false);
+retry:
+	i915_gem_object_lock(wa_ctx->indirect_ctx.obj, &ww);
+
+	vma = i915_gem_object_ggtt_pin_ww(wa_ctx->indirect_ctx.obj, &ww, NULL,
+					  0, CACHELINE_BYTES, 0);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		if (ret == -EDEADLK) {
+			ret = i915_gem_ww_ctx_backoff(&ww);
+			if (!ret)
+				goto retry;
+		}
+		return ret;
+	}
+
+	i915_gem_ww_ctx_fini(&ww);
+
+	/* FIXME: we are not tracking our pinned VMA leaving it
+	 * up to the core to fix up the stray pin_count upon
+	 * free.
+	 */
+
+	wa_ctx->indirect_ctx.shadow_gma = i915_ggtt_offset(vma);
+
+	wa_ctx->per_ctx.shadow_gma = *((unsigned int *)per_ctx_va + 1);
+	memset(per_ctx_va, 0, CACHELINE_BYTES);
+
+	update_wa_ctx_2_shadow_ctx(wa_ctx);
+	return 0;
+}
+
+static void update_vreg_in_ctx(struct intel_vgpu_workload *workload)
+{
+	vgpu_vreg_t(workload->vgpu, RING_START(workload->engine->mmio_base)) =
+		workload->rb_start;
+}
+
+static void release_shadow_batch_buffer(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu_shadow_bb *bb, *pos;
+
+	if (list_empty(&workload->shadow_bb))
+		return;
+
+	bb = list_first_entry(&workload->shadow_bb,
+			struct intel_vgpu_shadow_bb, list);
+
+	list_for_each_entry_safe(bb, pos, &workload->shadow_bb, list) {
+		if (bb->obj) {
+			i915_gem_object_lock(bb->obj, NULL);
+			if (bb->va && !IS_ERR(bb->va))
+				i915_gem_object_unpin_map(bb->obj);
+
+			if (bb->vma && !IS_ERR(bb->vma))
+				i915_vma_unpin(bb->vma);
+
+			i915_gem_object_unlock(bb->obj);
+			i915_gem_object_put(bb->obj);
+		}
+		list_del(&bb->list);
+		kfree(bb);
+	}
+}
+
+static int
+intel_vgpu_shadow_mm_pin(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_vgpu_mm *m;
+	int ret = 0;
+
+	ret = intel_vgpu_pin_mm(workload->shadow_mm);
+	if (ret) {
+		gvt_vgpu_err("fail to vgpu pin mm\n");
+		return ret;
+	}
+
+	if (workload->shadow_mm->type != INTEL_GVT_MM_PPGTT ||
+	    !workload->shadow_mm->ppgtt_mm.shadowed) {
+		intel_vgpu_unpin_mm(workload->shadow_mm);
+		gvt_vgpu_err("workload shadow ppgtt isn't ready\n");
+		return -EINVAL;
+	}
+
+	if (!list_empty(&workload->lri_shadow_mm)) {
+		list_for_each_entry(m, &workload->lri_shadow_mm,
+				    ppgtt_mm.link) {
+			ret = intel_vgpu_pin_mm(m);
+			if (ret) {
+				list_for_each_entry_from_reverse(m,
+								 &workload->lri_shadow_mm,
+								 ppgtt_mm.link)
+					intel_vgpu_unpin_mm(m);
+				gvt_vgpu_err("LRI shadow ppgtt fail to pin\n");
+				break;
+			}
+		}
+	}
+
+	if (ret)
+		intel_vgpu_unpin_mm(workload->shadow_mm);
+
+	return ret;
+}
+
+static void
+intel_vgpu_shadow_mm_unpin(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu_mm *m;
+
+	if (!list_empty(&workload->lri_shadow_mm)) {
+		list_for_each_entry(m, &workload->lri_shadow_mm,
+				    ppgtt_mm.link)
+			intel_vgpu_unpin_mm(m);
+	}
+	intel_vgpu_unpin_mm(workload->shadow_mm);
+}
+
+static int prepare_workload(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	int ret = 0;
+
+	ret = intel_vgpu_shadow_mm_pin(workload);
+	if (ret) {
+		gvt_vgpu_err("fail to pin shadow mm\n");
+		return ret;
+	}
+
+	update_shadow_pdps(workload);
+
+	set_context_ppgtt_from_shadow(workload, s->shadow[workload->engine->id]);
+
+	ret = intel_vgpu_sync_oos_pages(workload->vgpu);
+	if (ret) {
+		gvt_vgpu_err("fail to vgpu sync oos pages\n");
+		goto err_unpin_mm;
+	}
+
+	ret = intel_vgpu_flush_post_shadow(workload->vgpu);
+	if (ret) {
+		gvt_vgpu_err("fail to flush post shadow\n");
+		goto err_unpin_mm;
+	}
+
+	ret = copy_workload_to_ring_buffer(workload);
+	if (ret) {
+		gvt_vgpu_err("fail to generate request\n");
+		goto err_unpin_mm;
+	}
+
+	ret = prepare_shadow_batch_buffer(workload);
+	if (ret) {
+		gvt_vgpu_err("fail to prepare_shadow_batch_buffer\n");
+		goto err_unpin_mm;
+	}
+
+	ret = prepare_shadow_wa_ctx(&workload->wa_ctx);
+	if (ret) {
+		gvt_vgpu_err("fail to prepare_shadow_wa_ctx\n");
+		goto err_shadow_batch;
+	}
+
+	if (workload->prepare) {
+		ret = workload->prepare(workload);
+		if (ret)
+			goto err_shadow_wa_ctx;
+	}
+
+	return 0;
+err_shadow_wa_ctx:
+	release_shadow_wa_ctx(&workload->wa_ctx);
+err_shadow_batch:
+	release_shadow_batch_buffer(workload);
+err_unpin_mm:
+	intel_vgpu_shadow_mm_unpin(workload);
+	return ret;
+}
+
+static int dispatch_workload(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct i915_request *rq;
+	int ret;
+
+	gvt_dbg_sched("ring id %s prepare to dispatch workload %p\n",
+		      workload->engine->name, workload);
+
+	mutex_lock(&vgpu->vgpu_lock);
+
+	ret = intel_gvt_workload_req_alloc(workload);
+	if (ret)
+		goto err_req;
+
+	ret = intel_gvt_scan_and_shadow_workload(workload);
+	if (ret)
+		goto out;
+
+	ret = populate_shadow_context(workload);
+	if (ret) {
+		release_shadow_wa_ctx(&workload->wa_ctx);
+		goto out;
+	}
+
+	ret = prepare_workload(workload);
+out:
+	if (ret) {
+		/* We might still need to add request with
+		 * clean ctx to retire it properly..
+		 */
+		rq = fetch_and_zero(&workload->req);
+		i915_request_put(rq);
+	}
+
+	if (!IS_ERR_OR_NULL(workload->req)) {
+		gvt_dbg_sched("ring id %s submit workload to i915 %p\n",
+			      workload->engine->name, workload->req);
+		i915_request_add(workload->req);
+		workload->dispatched = true;
+	}
+err_req:
+	if (ret)
+		workload->status = ret;
+	mutex_unlock(&vgpu->vgpu_lock);
+	return ret;
+}
+
+static struct intel_vgpu_workload *
+pick_next_workload(struct intel_gvt *gvt, struct intel_engine_cs *engine)
+{
+	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
+	struct intel_vgpu_workload *workload = NULL;
+
+	mutex_lock(&gvt->sched_lock);
+
+	/*
+	 * no current vgpu / will be scheduled out / no workload
+	 * bail out
+	 */
+	if (!scheduler->current_vgpu) {
+		gvt_dbg_sched("ring %s stop - no current vgpu\n", engine->name);
+		goto out;
+	}
+
+	if (scheduler->need_reschedule) {
+		gvt_dbg_sched("ring %s stop - will reschedule\n", engine->name);
+		goto out;
+	}
+
+	if (!scheduler->current_vgpu->active ||
+	    list_empty(workload_q_head(scheduler->current_vgpu, engine)))
+		goto out;
+
+	/*
+	 * still have current workload, maybe the workload disptacher
+	 * fail to submit it for some reason, resubmit it.
+	 */
+	if (scheduler->current_workload[engine->id]) {
+		workload = scheduler->current_workload[engine->id];
+		gvt_dbg_sched("ring %s still have current workload %p\n",
+			      engine->name, workload);
+		goto out;
+	}
+
+	/*
+	 * pick a workload as current workload
+	 * once current workload is set, schedule policy routines
+	 * will wait the current workload is finished when trying to
+	 * schedule out a vgpu.
+	 */
+	scheduler->current_workload[engine->id] =
+		list_first_entry(workload_q_head(scheduler->current_vgpu,
+						 engine),
+				 struct intel_vgpu_workload, list);
+
+	workload = scheduler->current_workload[engine->id];
+
+	gvt_dbg_sched("ring %s pick new workload %p\n", engine->name, workload);
+
+	atomic_inc(&workload->vgpu->submission.running_workload_num);
+out:
+	mutex_unlock(&gvt->sched_lock);
+	return workload;
+}
+
+static void update_guest_pdps(struct intel_vgpu *vgpu,
+			      u64 ring_context_gpa, u32 pdp[8])
+{
+	u64 gpa;
+	int i;
+
+	gpa = ring_context_gpa + RING_CTX_OFF(pdps[0].val);
+
+	for (i = 0; i < 8; i++)
+		intel_gvt_write_gpa(vgpu, gpa + i * 8, &pdp[7 - i], 4);
+}
+
+static __maybe_unused bool
+check_shadow_context_ppgtt(struct execlist_ring_context *c, struct intel_vgpu_mm *m)
+{
+	if (m->ppgtt_mm.root_entry_type == GTT_TYPE_PPGTT_ROOT_L4_ENTRY) {
+		u64 shadow_pdp = c->pdps[7].val | (u64) c->pdps[6].val << 32;
+
+		if (shadow_pdp != m->ppgtt_mm.shadow_pdps[0]) {
+			gvt_dbg_mm("4-level context ppgtt not match LRI command\n");
+			return false;
+		}
+		return true;
+	} else {
+		/* see comment in LRI handler in cmd_parser.c */
+		gvt_dbg_mm("invalid shadow mm type\n");
+		return false;
+	}
+}
+
+static void update_guest_context(struct intel_vgpu_workload *workload)
+{
+	struct i915_request *rq = workload->req;
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct execlist_ring_context *shadow_ring_context;
+	struct intel_context *ctx = workload->req->context;
+	void *context_base;
+	void *src;
+	unsigned long context_gpa, context_page_num;
+	unsigned long gpa_base; /* first gpa of consecutive GPAs */
+	unsigned long gpa_size; /* size of consecutive GPAs*/
+	int i;
+	u32 ring_base;
+	u32 head, tail;
+	u16 wrap_count;
+
+	gvt_dbg_sched("ring id %d workload lrca %x\n", rq->engine->id,
+		      workload->ctx_desc.lrca);
+
+	GEM_BUG_ON(!intel_context_is_pinned(ctx));
+
+	head = workload->rb_head;
+	tail = workload->rb_tail;
+	wrap_count = workload->guest_rb_head >> RB_HEAD_WRAP_CNT_OFF;
+
+	if (tail < head) {
+		if (wrap_count == RB_HEAD_WRAP_CNT_MAX)
+			wrap_count = 0;
+		else
+			wrap_count += 1;
+	}
+
+	head = (wrap_count << RB_HEAD_WRAP_CNT_OFF) | tail;
+
+	ring_base = rq->engine->mmio_base;
+	vgpu_vreg_t(vgpu, RING_TAIL(ring_base)) = tail;
+	vgpu_vreg_t(vgpu, RING_HEAD(ring_base)) = head;
+
+	context_page_num = rq->engine->context_size;
+	context_page_num = context_page_num >> PAGE_SHIFT;
+
+	if (IS_BROADWELL(rq->engine->i915) && rq->engine->id == RCS0)
+		context_page_num = 19;
+
+	context_base = (void *) ctx->lrc_reg_state -
+			(LRC_STATE_PN << I915_GTT_PAGE_SHIFT);
+
+	/* find consecutive GPAs from gma until the first inconsecutive GPA.
+	 * write to the consecutive GPAs from src virtual address
+	 */
+	gpa_size = 0;
+	for (i = 2; i < context_page_num; i++) {
+		context_gpa = intel_vgpu_gma_to_gpa(vgpu->gtt.ggtt_mm,
+				(u32)((workload->ctx_desc.lrca + i) <<
+					I915_GTT_PAGE_SHIFT));
+		if (context_gpa == INTEL_GVT_INVALID_ADDR) {
+			gvt_vgpu_err("invalid guest context descriptor\n");
+			return;
+		}
+
+		if (gpa_size == 0) {
+			gpa_base = context_gpa;
+			src = context_base + (i << I915_GTT_PAGE_SHIFT);
+		} else if (context_gpa != gpa_base + gpa_size)
+			goto write;
+
+		gpa_size += I915_GTT_PAGE_SIZE;
+
+		if (i == context_page_num - 1)
+			goto write;
+
+		continue;
+
+write:
+		intel_gvt_write_gpa(vgpu, gpa_base, src, gpa_size);
+		gpa_base = context_gpa;
+		gpa_size = I915_GTT_PAGE_SIZE;
+		src = context_base + (i << I915_GTT_PAGE_SHIFT);
+	}
+
+	intel_gvt_write_gpa(vgpu, workload->ring_context_gpa +
+		RING_CTX_OFF(ring_header.val), &workload->rb_tail, 4);
+
+	shadow_ring_context = (void *) ctx->lrc_reg_state;
+
+	if (!list_empty(&workload->lri_shadow_mm)) {
+		struct intel_vgpu_mm *m = list_last_entry(&workload->lri_shadow_mm,
+							  struct intel_vgpu_mm,
+							  ppgtt_mm.link);
+		GEM_BUG_ON(!check_shadow_context_ppgtt(shadow_ring_context, m));
+		update_guest_pdps(vgpu, workload->ring_context_gpa,
+				  (void *)m->ppgtt_mm.guest_pdps);
+	}
+
+#define COPY_REG(name) \
+	intel_gvt_write_gpa(vgpu, workload->ring_context_gpa + \
+		RING_CTX_OFF(name.val), &shadow_ring_context->name.val, 4)
+
+	COPY_REG(ctx_ctrl);
+	COPY_REG(ctx_timestamp);
+
+#undef COPY_REG
+
+	intel_gvt_write_gpa(vgpu,
+			workload->ring_context_gpa +
+			sizeof(*shadow_ring_context),
+			(void *)shadow_ring_context +
+			sizeof(*shadow_ring_context),
+			I915_GTT_PAGE_SIZE - sizeof(*shadow_ring_context));
+}
+
+void intel_vgpu_clean_workloads(struct intel_vgpu *vgpu,
+				intel_engine_mask_t engine_mask)
+{
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_engine_cs *engine;
+	struct intel_vgpu_workload *pos, *n;
+	intel_engine_mask_t tmp;
+
+	/* free the unsubmited workloads in the queues. */
+	for_each_engine_masked(engine, vgpu->gvt->gt, engine_mask, tmp) {
+		list_for_each_entry_safe(pos, n,
+			&s->workload_q_head[engine->id], list) {
+			list_del_init(&pos->list);
+			intel_vgpu_destroy_workload(pos);
+		}
+		clear_bit(engine->id, s->shadow_ctx_desc_updated);
+	}
+}
+
+static void complete_current_workload(struct intel_gvt *gvt, int ring_id)
+{
+	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
+	struct intel_vgpu_workload *workload =
+		scheduler->current_workload[ring_id];
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct i915_request *rq = workload->req;
+	int event;
+
+	mutex_lock(&vgpu->vgpu_lock);
+	mutex_lock(&gvt->sched_lock);
+
+	/* For the workload w/ request, needs to wait for the context
+	 * switch to make sure request is completed.
+	 * For the workload w/o request, directly complete the workload.
+	 */
+	if (rq) {
+		wait_event(workload->shadow_ctx_status_wq,
+			   !atomic_read(&workload->shadow_ctx_active));
+
+		/* If this request caused GPU hang, req->fence.error will
+		 * be set to -EIO. Use -EIO to set workload status so
+		 * that when this request caused GPU hang, didn't trigger
+		 * context switch interrupt to guest.
+		 */
+		if (likely(workload->status == -EINPROGRESS)) {
+			if (workload->req->fence.error == -EIO)
+				workload->status = -EIO;
+			else
+				workload->status = 0;
+		}
+
+		if (!workload->status &&
+		    !(vgpu->resetting_eng & BIT(ring_id))) {
+			update_guest_context(workload);
+
+			for_each_set_bit(event, workload->pending_events,
+					 INTEL_GVT_EVENT_MAX)
+				intel_vgpu_trigger_virtual_event(vgpu, event);
+		}
+
+		i915_request_put(fetch_and_zero(&workload->req));
+	}
+
+	gvt_dbg_sched("ring id %d complete workload %p status %d\n",
+			ring_id, workload, workload->status);
+
+	scheduler->current_workload[ring_id] = NULL;
+
+	list_del_init(&workload->list);
+
+	if (workload->status || vgpu->resetting_eng & BIT(ring_id)) {
+		/* if workload->status is not successful means HW GPU
+		 * has occurred GPU hang or something wrong with i915/GVT,
+		 * and GVT won't inject context switch interrupt to guest.
+		 * So this error is a vGPU hang actually to the guest.
+		 * According to this we should emunlate a vGPU hang. If
+		 * there are pending workloads which are already submitted
+		 * from guest, we should clean them up like HW GPU does.
+		 *
+		 * if it is in middle of engine resetting, the pending
+		 * workloads won't be submitted to HW GPU and will be
+		 * cleaned up during the resetting process later, so doing
+		 * the workload clean up here doesn't have any impact.
+		 **/
+		intel_vgpu_clean_workloads(vgpu, BIT(ring_id));
+	}
+
+	workload->complete(workload);
+
+	intel_vgpu_shadow_mm_unpin(workload);
+	intel_vgpu_destroy_workload(workload);
+
+	atomic_dec(&s->running_workload_num);
+	wake_up(&scheduler->workload_complete_wq);
+
+	if (gvt->scheduler.need_reschedule)
+		intel_gvt_request_service(gvt, INTEL_GVT_REQUEST_EVENT_SCHED);
+
+	mutex_unlock(&gvt->sched_lock);
+	mutex_unlock(&vgpu->vgpu_lock);
+}
+
+static int workload_thread(void *arg)
+{
+	struct intel_engine_cs *engine = arg;
+	const bool need_force_wake = GRAPHICS_VER(engine->i915) >= 9;
+	struct intel_gvt *gvt = engine->i915->gvt;
+	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
+	struct intel_vgpu_workload *workload = NULL;
+	struct intel_vgpu *vgpu = NULL;
+	int ret;
+	DEFINE_WAIT_FUNC(wait, woken_wake_function);
+
+	gvt_dbg_core("workload thread for ring %s started\n", engine->name);
+
+	while (!kthread_should_stop()) {
+		intel_wakeref_t wakeref;
+
+		add_wait_queue(&scheduler->waitq[engine->id], &wait);
+		do {
+			workload = pick_next_workload(gvt, engine);
+			if (workload)
+				break;
+			wait_woken(&wait, TASK_INTERRUPTIBLE,
+				   MAX_SCHEDULE_TIMEOUT);
+		} while (!kthread_should_stop());
+		remove_wait_queue(&scheduler->waitq[engine->id], &wait);
+
+		if (!workload)
+			break;
+
+		gvt_dbg_sched("ring %s next workload %p vgpu %d\n",
+			      engine->name, workload,
+			      workload->vgpu->id);
+
+		wakeref = intel_runtime_pm_get(engine->uncore->rpm);
+
+		gvt_dbg_sched("ring %s will dispatch workload %p\n",
+			      engine->name, workload);
+
+		if (need_force_wake)
+			intel_uncore_forcewake_get(engine->uncore,
+						   FORCEWAKE_ALL);
+		/*
+		 * Update the vReg of the vGPU which submitted this
+		 * workload. The vGPU may use these registers for checking
+		 * the context state. The value comes from GPU commands
+		 * in this workload.
+		 */
+		update_vreg_in_ctx(workload);
+
+		ret = dispatch_workload(workload);
+
+		if (ret) {
+			vgpu = workload->vgpu;
+			gvt_vgpu_err("fail to dispatch workload, skip\n");
+			goto complete;
+		}
+
+		gvt_dbg_sched("ring %s wait workload %p\n",
+			      engine->name, workload);
+		i915_request_wait(workload->req, 0, MAX_SCHEDULE_TIMEOUT);
+
+complete:
+		gvt_dbg_sched("will complete workload %p, status: %d\n",
+			      workload, workload->status);
+
+		complete_current_workload(gvt, engine->id);
+
+		if (need_force_wake)
+			intel_uncore_forcewake_put(engine->uncore,
+						   FORCEWAKE_ALL);
+
+		intel_runtime_pm_put(engine->uncore->rpm, wakeref);
+		if (ret && (vgpu_is_vm_unhealthy(ret)))
+			enter_failsafe_mode(vgpu, GVT_FAILSAFE_GUEST_ERR);
+	}
+	return 0;
+}
+
+void intel_gvt_wait_vgpu_idle(struct intel_vgpu *vgpu)
+{
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
+
+	if (atomic_read(&s->running_workload_num)) {
+		gvt_dbg_sched("wait vgpu idle\n");
+
+		wait_event(scheduler->workload_complete_wq,
+				!atomic_read(&s->running_workload_num));
+	}
+}
+
+void intel_gvt_clean_workload_scheduler(struct intel_gvt *gvt)
+{
+	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id i;
+
+	gvt_dbg_core("clean workload scheduler\n");
+
+	for_each_engine(engine, gvt->gt, i) {
+		atomic_notifier_chain_unregister(
+					&engine->context_status_notifier,
+					&gvt->shadow_ctx_notifier_block[i]);
+		kthread_stop(scheduler->thread[i]);
+	}
+}
+
+int intel_gvt_init_workload_scheduler(struct intel_gvt *gvt)
+{
+	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id i;
+	int ret;
+
+	gvt_dbg_core("init workload scheduler\n");
+
+	init_waitqueue_head(&scheduler->workload_complete_wq);
+
+	for_each_engine(engine, gvt->gt, i) {
+		init_waitqueue_head(&scheduler->waitq[i]);
+
+		scheduler->thread[i] = kthread_run(workload_thread, engine,
+						   "gvt:%s", engine->name);
+		if (IS_ERR(scheduler->thread[i])) {
+			gvt_err("fail to create workload thread\n");
+			ret = PTR_ERR(scheduler->thread[i]);
+			goto err;
+		}
+
+		gvt->shadow_ctx_notifier_block[i].notifier_call =
+					shadow_context_status_change;
+		atomic_notifier_chain_register(&engine->context_status_notifier,
+					&gvt->shadow_ctx_notifier_block[i]);
+	}
+
+	return 0;
+
+err:
+	intel_gvt_clean_workload_scheduler(gvt);
+	return ret;
+}
+
+static void
+i915_context_ppgtt_root_restore(struct intel_vgpu_submission *s,
+				struct i915_ppgtt *ppgtt)
+{
+	int i;
+
+	if (i915_vm_is_4lvl(&ppgtt->vm)) {
+		set_dma_address(ppgtt->pd, s->i915_context_pml4);
+	} else {
+		for (i = 0; i < GEN8_3LVL_PDPES; i++) {
+			struct i915_page_directory * const pd =
+				i915_pd_entry(ppgtt->pd, i);
+
+			set_dma_address(pd, s->i915_context_pdps[i]);
+		}
+	}
+}
+
+/**
+ * intel_vgpu_clean_submission - free submission-related resource for vGPU
+ * @vgpu: a vGPU
+ *
+ * This function is called when a vGPU is being destroyed.
+ *
+ */
+void intel_vgpu_clean_submission(struct intel_vgpu *vgpu)
+{
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	intel_vgpu_select_submission_ops(vgpu, ALL_ENGINES, 0);
+
+	i915_context_ppgtt_root_restore(s, i915_vm_to_ppgtt(s->shadow[0]->vm));
+	for_each_engine(engine, vgpu->gvt->gt, id)
+		intel_context_put(s->shadow[id]);
+
+	kmem_cache_destroy(s->workloads);
+}
+
+
+/**
+ * intel_vgpu_reset_submission - reset submission-related resource for vGPU
+ * @vgpu: a vGPU
+ * @engine_mask: engines expected to be reset
+ *
+ * This function is called when a vGPU is being destroyed.
+ *
+ */
+void intel_vgpu_reset_submission(struct intel_vgpu *vgpu,
+				 intel_engine_mask_t engine_mask)
+{
+	struct intel_vgpu_submission *s = &vgpu->submission;
+
+	if (!s->active)
+		return;
+
+	intel_vgpu_clean_workloads(vgpu, engine_mask);
+	s->ops->reset(vgpu, engine_mask);
+}
+
+static void
+i915_context_ppgtt_root_save(struct intel_vgpu_submission *s,
+			     struct i915_ppgtt *ppgtt)
+{
+	int i;
+
+	if (i915_vm_is_4lvl(&ppgtt->vm)) {
+		s->i915_context_pml4 = px_dma(ppgtt->pd);
+	} else {
+		for (i = 0; i < GEN8_3LVL_PDPES; i++) {
+			struct i915_page_directory * const pd =
+				i915_pd_entry(ppgtt->pd, i);
+
+			s->i915_context_pdps[i] = px_dma(pd);
+		}
+	}
+}
+
+/**
+ * intel_vgpu_setup_submission - setup submission-related resource for vGPU
+ * @vgpu: a vGPU
+ *
+ * This function is called when a vGPU is being created.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+int intel_vgpu_setup_submission(struct intel_vgpu *vgpu)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_engine_cs *engine;
+	struct i915_ppgtt *ppgtt;
+	enum intel_engine_id i;
+	int ret;
+
+	ppgtt = i915_ppgtt_create(to_gt(i915), I915_BO_ALLOC_PM_EARLY);
+	if (IS_ERR(ppgtt))
+		return PTR_ERR(ppgtt);
+
+	i915_context_ppgtt_root_save(s, ppgtt);
+
+	for_each_engine(engine, vgpu->gvt->gt, i) {
+		struct intel_context *ce;
+
+		INIT_LIST_HEAD(&s->workload_q_head[i]);
+		s->shadow[i] = ERR_PTR(-EINVAL);
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			ret = PTR_ERR(ce);
+			goto out_shadow_ctx;
+		}
+
+		i915_vm_put(ce->vm);
+		ce->vm = i915_vm_get(&ppgtt->vm);
+		intel_context_set_single_submission(ce);
+
+		/* Max ring buffer size */
+		if (!intel_uc_wants_guc_submission(&engine->gt->uc))
+			ce->ring_size = SZ_2M;
+
+		s->shadow[i] = ce;
+	}
+
+	bitmap_zero(s->shadow_ctx_desc_updated, I915_NUM_ENGINES);
+
+	s->workloads = kmem_cache_create_usercopy("gvt-g_vgpu_workload",
+						  sizeof(struct intel_vgpu_workload), 0,
+						  SLAB_HWCACHE_ALIGN,
+						  offsetof(struct intel_vgpu_workload, rb_tail),
+						  sizeof_field(struct intel_vgpu_workload, rb_tail),
+						  NULL);
+
+	if (!s->workloads) {
+		ret = -ENOMEM;
+		goto out_shadow_ctx;
+	}
+
+	atomic_set(&s->running_workload_num, 0);
+	bitmap_zero(s->tlb_handle_pending, I915_NUM_ENGINES);
+
+	memset(s->last_ctx, 0, sizeof(s->last_ctx));
+
+	i915_vm_put(&ppgtt->vm);
+	return 0;
+
+out_shadow_ctx:
+	i915_context_ppgtt_root_restore(s, ppgtt);
+	for_each_engine(engine, vgpu->gvt->gt, i) {
+		if (IS_ERR(s->shadow[i]))
+			break;
+
+		intel_context_put(s->shadow[i]);
+	}
+	i915_vm_put(&ppgtt->vm);
+	return ret;
+}
+
+/**
+ * intel_vgpu_select_submission_ops - select virtual submission interface
+ * @vgpu: a vGPU
+ * @engine_mask: either ALL_ENGINES or target engine mask
+ * @interface: expected vGPU virtual submission interface
+ *
+ * This function is called when guest configures submission interface.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+int intel_vgpu_select_submission_ops(struct intel_vgpu *vgpu,
+				     intel_engine_mask_t engine_mask,
+				     unsigned int interface)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	const struct intel_vgpu_submission_ops *ops[] = {
+		[INTEL_VGPU_EXECLIST_SUBMISSION] =
+			&intel_vgpu_execlist_submission_ops,
+	};
+	int ret;
+
+	if (drm_WARN_ON(&i915->drm, interface >= ARRAY_SIZE(ops)))
+		return -EINVAL;
+
+	if (drm_WARN_ON(&i915->drm,
+			interface == 0 && engine_mask != ALL_ENGINES))
+		return -EINVAL;
+
+	if (s->active)
+		s->ops->clean(vgpu, engine_mask);
+
+	if (interface == 0) {
+		s->ops = NULL;
+		s->virtual_submission_interface = 0;
+		s->active = false;
+		gvt_dbg_core("vgpu%d: remove submission ops\n", vgpu->id);
+		return 0;
+	}
+
+	ret = ops[interface]->init(vgpu, engine_mask);
+	if (ret)
+		return ret;
+
+	s->ops = ops[interface];
+	s->virtual_submission_interface = interface;
+	s->active = true;
+
+	gvt_dbg_core("vgpu%d: activate ops [ %s ]\n",
+			vgpu->id, s->ops->name);
+
+	return 0;
+}
+
+/**
+ * intel_vgpu_destroy_workload - destroy a vGPU workload
+ * @workload: workload to destroy
+ *
+ * This function is called when destroy a vGPU workload.
+ *
+ */
+void intel_vgpu_destroy_workload(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu_submission *s = &workload->vgpu->submission;
+
+	intel_context_unpin(s->shadow[workload->engine->id]);
+	release_shadow_batch_buffer(workload);
+	release_shadow_wa_ctx(&workload->wa_ctx);
+
+	if (!list_empty(&workload->lri_shadow_mm)) {
+		struct intel_vgpu_mm *m, *mm;
+		list_for_each_entry_safe(m, mm, &workload->lri_shadow_mm,
+					 ppgtt_mm.link) {
+			list_del(&m->ppgtt_mm.link);
+			intel_vgpu_mm_put(m);
+		}
+	}
+
+	GEM_BUG_ON(!list_empty(&workload->lri_shadow_mm));
+	if (workload->shadow_mm)
+		intel_vgpu_mm_put(workload->shadow_mm);
+
+	kmem_cache_free(s->workloads, workload);
+}
+
+static struct intel_vgpu_workload *
+alloc_workload(struct intel_vgpu *vgpu)
+{
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct intel_vgpu_workload *workload;
+
+	workload = kmem_cache_zalloc(s->workloads, GFP_KERNEL);
+	if (!workload)
+		return ERR_PTR(-ENOMEM);
+
+	INIT_LIST_HEAD(&workload->list);
+	INIT_LIST_HEAD(&workload->shadow_bb);
+	INIT_LIST_HEAD(&workload->lri_shadow_mm);
+
+	init_waitqueue_head(&workload->shadow_ctx_status_wq);
+	atomic_set(&workload->shadow_ctx_active, 0);
+
+	workload->status = -EINPROGRESS;
+	workload->vgpu = vgpu;
+
+	return workload;
+}
+
+#define RING_CTX_OFF(x) \
+	offsetof(struct execlist_ring_context, x)
+
+static void read_guest_pdps(struct intel_vgpu *vgpu,
+		u64 ring_context_gpa, u32 pdp[8])
+{
+	u64 gpa;
+	int i;
+
+	gpa = ring_context_gpa + RING_CTX_OFF(pdps[0].val);
+
+	for (i = 0; i < 8; i++)
+		intel_gvt_read_gpa(vgpu,
+				gpa + i * 8, &pdp[7 - i], 4);
+}
+
+static int prepare_mm(struct intel_vgpu_workload *workload)
+{
+	struct execlist_ctx_descriptor_format *desc = &workload->ctx_desc;
+	struct intel_vgpu_mm *mm;
+	struct intel_vgpu *vgpu = workload->vgpu;
+	enum intel_gvt_gtt_type root_entry_type;
+	u64 pdps[GVT_RING_CTX_NR_PDPS];
+
+	switch (desc->addressing_mode) {
+	case 1: /* legacy 32-bit */
+		root_entry_type = GTT_TYPE_PPGTT_ROOT_L3_ENTRY;
+		break;
+	case 3: /* legacy 64-bit */
+		root_entry_type = GTT_TYPE_PPGTT_ROOT_L4_ENTRY;
+		break;
+	default:
+		gvt_vgpu_err("Advanced Context mode(SVM) is not supported!\n");
+		return -EINVAL;
+	}
+
+	read_guest_pdps(workload->vgpu, workload->ring_context_gpa, (void *)pdps);
+
+	mm = intel_vgpu_get_ppgtt_mm(workload->vgpu, root_entry_type, pdps);
+	if (IS_ERR(mm))
+		return PTR_ERR(mm);
+
+	workload->shadow_mm = mm;
+	return 0;
+}
+
+#define same_context(a, b) (((a)->context_id == (b)->context_id) && \
+		((a)->lrca == (b)->lrca))
+
+/**
+ * intel_vgpu_create_workload - create a vGPU workload
+ * @vgpu: a vGPU
+ * @engine: the engine
+ * @desc: a guest context descriptor
+ *
+ * This function is called when creating a vGPU workload.
+ *
+ * Returns:
+ * struct intel_vgpu_workload * on success, negative error code in
+ * pointer if failed.
+ *
+ */
+struct intel_vgpu_workload *
+intel_vgpu_create_workload(struct intel_vgpu *vgpu,
+			   const struct intel_engine_cs *engine,
+			   struct execlist_ctx_descriptor_format *desc)
+{
+	struct intel_vgpu_submission *s = &vgpu->submission;
+	struct list_head *q = workload_q_head(vgpu, engine);
+	struct intel_vgpu_workload *last_workload = NULL;
+	struct intel_vgpu_workload *workload = NULL;
+	u64 ring_context_gpa;
+	u32 head, tail, start, ctl, ctx_ctl, per_ctx, indirect_ctx;
+	u32 guest_head;
+	int ret;
+
+	ring_context_gpa = intel_vgpu_gma_to_gpa(vgpu->gtt.ggtt_mm,
+			(u32)((desc->lrca + 1) << I915_GTT_PAGE_SHIFT));
+	if (ring_context_gpa == INTEL_GVT_INVALID_ADDR) {
+		gvt_vgpu_err("invalid guest context LRCA: %x\n", desc->lrca);
+		return ERR_PTR(-EINVAL);
+	}
+
+	intel_gvt_read_gpa(vgpu, ring_context_gpa +
+			RING_CTX_OFF(ring_header.val), &head, 4);
+
+	intel_gvt_read_gpa(vgpu, ring_context_gpa +
+			RING_CTX_OFF(ring_tail.val), &tail, 4);
+
+	guest_head = head;
+
+	head &= RB_HEAD_OFF_MASK;
+	tail &= RB_TAIL_OFF_MASK;
+
+	list_for_each_entry_reverse(last_workload, q, list) {
+
+		if (same_context(&last_workload->ctx_desc, desc)) {
+			gvt_dbg_el("ring %s cur workload == last\n",
+				   engine->name);
+			gvt_dbg_el("ctx head %x real head %lx\n", head,
+				   last_workload->rb_tail);
+			/*
+			 * cannot use guest context head pointer here,
+			 * as it might not be updated at this time
+			 */
+			head = last_workload->rb_tail;
+			break;
+		}
+	}
+
+	gvt_dbg_el("ring %s begin a new workload\n", engine->name);
+
+	/* record some ring buffer register values for scan and shadow */
+	intel_gvt_read_gpa(vgpu, ring_context_gpa +
+			RING_CTX_OFF(rb_start.val), &start, 4);
+	intel_gvt_read_gpa(vgpu, ring_context_gpa +
+			RING_CTX_OFF(rb_ctrl.val), &ctl, 4);
+	intel_gvt_read_gpa(vgpu, ring_context_gpa +
+			RING_CTX_OFF(ctx_ctrl.val), &ctx_ctl, 4);
+
+	if (!intel_gvt_ggtt_validate_range(vgpu, start,
+				_RING_CTL_BUF_SIZE(ctl))) {
+		gvt_vgpu_err("context contain invalid rb at: 0x%x\n", start);
+		return ERR_PTR(-EINVAL);
+	}
+
+	workload = alloc_workload(vgpu);
+	if (IS_ERR(workload))
+		return workload;
+
+	workload->engine = engine;
+	workload->ctx_desc = *desc;
+	workload->ring_context_gpa = ring_context_gpa;
+	workload->rb_head = head;
+	workload->guest_rb_head = guest_head;
+	workload->rb_tail = tail;
+	workload->rb_start = start;
+	workload->rb_ctl = ctl;
+
+	if (engine->id == RCS0) {
+		intel_gvt_read_gpa(vgpu, ring_context_gpa +
+			RING_CTX_OFF(bb_per_ctx_ptr.val), &per_ctx, 4);
+		intel_gvt_read_gpa(vgpu, ring_context_gpa +
+			RING_CTX_OFF(rcs_indirect_ctx.val), &indirect_ctx, 4);
+
+		workload->wa_ctx.indirect_ctx.guest_gma =
+			indirect_ctx & INDIRECT_CTX_ADDR_MASK;
+		workload->wa_ctx.indirect_ctx.size =
+			(indirect_ctx & INDIRECT_CTX_SIZE_MASK) *
+			CACHELINE_BYTES;
+
+		if (workload->wa_ctx.indirect_ctx.size != 0) {
+			if (!intel_gvt_ggtt_validate_range(vgpu,
+				workload->wa_ctx.indirect_ctx.guest_gma,
+				workload->wa_ctx.indirect_ctx.size)) {
+				gvt_vgpu_err("invalid wa_ctx at: 0x%lx\n",
+				    workload->wa_ctx.indirect_ctx.guest_gma);
+				kmem_cache_free(s->workloads, workload);
+				return ERR_PTR(-EINVAL);
+			}
+		}
+
+		workload->wa_ctx.per_ctx.guest_gma =
+			per_ctx & PER_CTX_ADDR_MASK;
+		workload->wa_ctx.per_ctx.valid = per_ctx & 1;
+		if (workload->wa_ctx.per_ctx.valid) {
+			if (!intel_gvt_ggtt_validate_range(vgpu,
+				workload->wa_ctx.per_ctx.guest_gma,
+				CACHELINE_BYTES)) {
+				gvt_vgpu_err("invalid per_ctx at: 0x%lx\n",
+					workload->wa_ctx.per_ctx.guest_gma);
+				kmem_cache_free(s->workloads, workload);
+				return ERR_PTR(-EINVAL);
+			}
+		}
+	}
+
+	gvt_dbg_el("workload %p ring %s head %x tail %x start %x ctl %x\n",
+		   workload, engine->name, head, tail, start, ctl);
+
+	ret = prepare_mm(workload);
+	if (ret) {
+		kmem_cache_free(s->workloads, workload);
+		return ERR_PTR(ret);
+	}
+
+	/* Only scan and shadow the first workload in the queue
+	 * as there is only one pre-allocated buf-obj for shadow.
+	 */
+	if (list_empty(q)) {
+		intel_wakeref_t wakeref;
+
+		with_intel_runtime_pm(engine->gt->uncore->rpm, wakeref)
+			ret = intel_gvt_scan_and_shadow_workload(workload);
+	}
+
+	if (ret) {
+		if (vgpu_is_vm_unhealthy(ret))
+			enter_failsafe_mode(vgpu, GVT_FAILSAFE_GUEST_ERR);
+		intel_vgpu_destroy_workload(workload);
+		return ERR_PTR(ret);
+	}
+
+	ret = intel_context_pin(s->shadow[engine->id]);
+	if (ret) {
+		intel_vgpu_destroy_workload(workload);
+		return ERR_PTR(ret);
+	}
+
+	return workload;
+}
+
+/**
+ * intel_vgpu_queue_workload - Qeue a vGPU workload
+ * @workload: the workload to queue in
+ */
+void intel_vgpu_queue_workload(struct intel_vgpu_workload *workload)
+{
+	list_add_tail(&workload->list,
+		      workload_q_head(workload->vgpu, workload->engine));
+	intel_gvt_kick_schedule(workload->vgpu->gvt);
+	wake_up(&workload->vgpu->gvt->scheduler.waitq[workload->engine->id]);
+}
diff --git a/drivers/gpu/drm/i915/gvt/scheduler.h b/drivers/gpu/drm/i915/gvt/scheduler.h
new file mode 100644
index 000000000..1f391b3da
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/scheduler.h
@@ -0,0 +1,171 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ * Contributors:
+ *    Ping Gao <ping.a.gao@intel.com>
+ *    Tina Zhang <tina.zhang@intel.com>
+ *    Chanbin Du <changbin.du@intel.com>
+ *    Min He <min.he@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *    Zhenyu Wang <zhenyuw@linux.intel.com>
+ *
+ */
+
+#ifndef _GVT_SCHEDULER_H_
+#define _GVT_SCHEDULER_H_
+
+#include "gt/intel_engine_types.h"
+
+#include "execlist.h"
+#include "interrupt.h"
+
+struct intel_gvt_workload_scheduler {
+	struct intel_vgpu *current_vgpu;
+	struct intel_vgpu *next_vgpu;
+	struct intel_vgpu_workload *current_workload[I915_NUM_ENGINES];
+	bool need_reschedule;
+
+	spinlock_t mmio_context_lock;
+	/* can be null when owner is host */
+	struct intel_vgpu *engine_owner[I915_NUM_ENGINES];
+
+	wait_queue_head_t workload_complete_wq;
+	struct task_struct *thread[I915_NUM_ENGINES];
+	wait_queue_head_t waitq[I915_NUM_ENGINES];
+
+	void *sched_data;
+	const struct intel_gvt_sched_policy_ops *sched_ops;
+};
+
+#define INDIRECT_CTX_ADDR_MASK 0xffffffc0
+#define INDIRECT_CTX_SIZE_MASK 0x3f
+struct shadow_indirect_ctx {
+	struct drm_i915_gem_object *obj;
+	unsigned long guest_gma;
+	unsigned long shadow_gma;
+	void *shadow_va;
+	u32 size;
+};
+
+#define PER_CTX_ADDR_MASK 0xfffff000
+struct shadow_per_ctx {
+	unsigned long guest_gma;
+	unsigned long shadow_gma;
+	unsigned valid;
+};
+
+struct intel_shadow_wa_ctx {
+	struct shadow_indirect_ctx indirect_ctx;
+	struct shadow_per_ctx per_ctx;
+
+};
+
+struct intel_vgpu_workload {
+	struct intel_vgpu *vgpu;
+	const struct intel_engine_cs *engine;
+	struct i915_request *req;
+	/* if this workload has been dispatched to i915? */
+	bool dispatched;
+	bool shadow;      /* if workload has done shadow of guest request */
+	int status;
+
+	struct intel_vgpu_mm *shadow_mm;
+	struct list_head lri_shadow_mm; /* For PPGTT load cmd */
+
+	/* different submission model may need different handler */
+	int (*prepare)(struct intel_vgpu_workload *);
+	int (*complete)(struct intel_vgpu_workload *);
+	struct list_head list;
+
+	DECLARE_BITMAP(pending_events, INTEL_GVT_EVENT_MAX);
+	void *shadow_ring_buffer_va;
+
+	/* execlist context information */
+	struct execlist_ctx_descriptor_format ctx_desc;
+	struct execlist_ring_context *ring_context;
+	unsigned long rb_head, rb_tail, rb_ctl, rb_start, rb_len;
+	unsigned long guest_rb_head;
+	bool restore_inhibit;
+	struct intel_vgpu_elsp_dwords elsp_dwords;
+	bool emulate_schedule_in;
+	atomic_t shadow_ctx_active;
+	wait_queue_head_t shadow_ctx_status_wq;
+	u64 ring_context_gpa;
+
+	/* shadow batch buffer */
+	struct list_head shadow_bb;
+	struct intel_shadow_wa_ctx wa_ctx;
+
+	/* oa registers */
+	u32 oactxctrl;
+	u32 flex_mmio[7];
+};
+
+struct intel_vgpu_shadow_bb {
+	struct list_head list;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	void *va;
+	u32 *bb_start_cmd_va;
+	unsigned long bb_offset;
+	bool ppgtt;
+};
+
+#define workload_q_head(vgpu, e) \
+	(&(vgpu)->submission.workload_q_head[(e)->id])
+
+void intel_vgpu_queue_workload(struct intel_vgpu_workload *workload);
+
+int intel_gvt_init_workload_scheduler(struct intel_gvt *gvt);
+
+void intel_gvt_clean_workload_scheduler(struct intel_gvt *gvt);
+
+void intel_gvt_wait_vgpu_idle(struct intel_vgpu *vgpu);
+
+int intel_vgpu_setup_submission(struct intel_vgpu *vgpu);
+
+void intel_vgpu_reset_submission(struct intel_vgpu *vgpu,
+				 intel_engine_mask_t engine_mask);
+
+void intel_vgpu_clean_submission(struct intel_vgpu *vgpu);
+
+int intel_vgpu_select_submission_ops(struct intel_vgpu *vgpu,
+				     intel_engine_mask_t engine_mask,
+				     unsigned int interface);
+
+extern const struct intel_vgpu_submission_ops
+intel_vgpu_execlist_submission_ops;
+
+struct intel_vgpu_workload *
+intel_vgpu_create_workload(struct intel_vgpu *vgpu,
+			   const struct intel_engine_cs *engine,
+			   struct execlist_ctx_descriptor_format *desc);
+
+void intel_vgpu_destroy_workload(struct intel_vgpu_workload *workload);
+
+void intel_vgpu_clean_workloads(struct intel_vgpu *vgpu,
+				intel_engine_mask_t engine_mask);
+
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/trace.h b/drivers/gpu/drm/i915/gvt/trace.h
new file mode 100644
index 000000000..020f1aa28
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/trace.h
@@ -0,0 +1,383 @@
+/*
+ * Copyright © 2011-2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Jike Song <jike.song@intel.com>
+ *
+ * Contributors:
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#if !defined(_GVT_TRACE_H_) || defined(TRACE_HEADER_MULTI_READ)
+#define _GVT_TRACE_H_
+
+#include <linux/types.h>
+#include <linux/stringify.h>
+#include <linux/tracepoint.h>
+#include <asm/tsc.h>
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM gvt
+
+TRACE_EVENT(spt_alloc,
+	TP_PROTO(int id, void *spt, int type, unsigned long mfn,
+		unsigned long gpt_gfn),
+
+	TP_ARGS(id, spt, type, mfn, gpt_gfn),
+
+	TP_STRUCT__entry(
+		__field(int, id)
+		__field(void *, spt)
+		__field(int, type)
+		__field(unsigned long, mfn)
+		__field(unsigned long, gpt_gfn)
+		),
+
+	TP_fast_assign(
+		__entry->id = id;
+		__entry->spt = spt;
+		__entry->type = type;
+		__entry->mfn = mfn;
+		__entry->gpt_gfn = gpt_gfn;
+	),
+
+	TP_printk("VM%d [alloc] spt %p type %d mfn 0x%lx gfn 0x%lx\n",
+		__entry->id,
+		__entry->spt,
+		__entry->type,
+		__entry->mfn,
+		__entry->gpt_gfn)
+);
+
+TRACE_EVENT(spt_free,
+	TP_PROTO(int id, void *spt, int type),
+
+	TP_ARGS(id, spt, type),
+
+	TP_STRUCT__entry(
+		__field(int, id)
+		__field(void *, spt)
+		__field(int, type)
+		),
+
+	TP_fast_assign(
+		__entry->id = id;
+		__entry->spt = spt;
+		__entry->type = type;
+	),
+
+	TP_printk("VM%u [free] spt %p type %d\n",
+		__entry->id,
+		__entry->spt,
+		__entry->type)
+);
+
+#define MAX_BUF_LEN 256
+
+TRACE_EVENT(gma_index,
+	TP_PROTO(const char *prefix, unsigned long gma,
+		unsigned long index),
+
+	TP_ARGS(prefix, gma, index),
+
+	TP_STRUCT__entry(
+		__array(char, buf, MAX_BUF_LEN)
+	),
+
+	TP_fast_assign(
+		snprintf(__entry->buf, MAX_BUF_LEN,
+			"%s gma 0x%lx index 0x%lx\n", prefix, gma, index);
+	),
+
+	TP_printk("%s", __entry->buf)
+);
+
+TRACE_EVENT(gma_translate,
+	TP_PROTO(int id, char *type, int ring_id, int root_entry_type,
+		unsigned long gma, unsigned long gpa),
+
+	TP_ARGS(id, type, ring_id, root_entry_type, gma, gpa),
+
+	TP_STRUCT__entry(
+		__array(char, buf, MAX_BUF_LEN)
+	),
+
+	TP_fast_assign(
+		snprintf(__entry->buf, MAX_BUF_LEN,
+			"VM%d %s ring %d root_entry_type %d gma 0x%lx -> gpa 0x%lx\n",
+			id, type, ring_id, root_entry_type, gma, gpa);
+	),
+
+	TP_printk("%s", __entry->buf)
+);
+
+TRACE_EVENT(spt_refcount,
+	TP_PROTO(int id, char *action, void *spt, int before, int after),
+
+	TP_ARGS(id, action, spt, before, after),
+
+	TP_STRUCT__entry(
+		__array(char, buf, MAX_BUF_LEN)
+	),
+
+	TP_fast_assign(
+		snprintf(__entry->buf, MAX_BUF_LEN,
+			"VM%d [%s] spt %p before %d -> after %d\n",
+				id, action, spt, before, after);
+	),
+
+	TP_printk("%s", __entry->buf)
+);
+
+TRACE_EVENT(spt_change,
+	TP_PROTO(int id, char *action, void *spt, unsigned long gfn,
+		int type),
+
+	TP_ARGS(id, action, spt, gfn, type),
+
+	TP_STRUCT__entry(
+		__array(char, buf, MAX_BUF_LEN)
+	),
+
+	TP_fast_assign(
+		snprintf(__entry->buf, MAX_BUF_LEN,
+			"VM%d [%s] spt %p gfn 0x%lx type %d\n",
+				id, action, spt, gfn, type);
+	),
+
+	TP_printk("%s", __entry->buf)
+);
+
+TRACE_EVENT(spt_guest_change,
+	TP_PROTO(int id, const char *tag, void *spt, int type, u64 v,
+		unsigned long index),
+
+	TP_ARGS(id, tag, spt, type, v, index),
+
+	TP_STRUCT__entry(
+		__array(char, buf, MAX_BUF_LEN)
+	),
+
+	TP_fast_assign(
+		snprintf(__entry->buf, MAX_BUF_LEN,
+		"VM%d [%s] spt %p type %d entry 0x%llx index 0x%lx\n",
+			id, tag, spt, type, v, index);
+	),
+
+	TP_printk("%s", __entry->buf)
+);
+
+TRACE_EVENT(oos_change,
+	TP_PROTO(int id, const char *tag, int page_id, void *gpt, int type),
+
+	TP_ARGS(id, tag, page_id, gpt, type),
+
+	TP_STRUCT__entry(
+		__array(char, buf, MAX_BUF_LEN)
+	),
+
+	TP_fast_assign(
+		snprintf(__entry->buf, MAX_BUF_LEN,
+		"VM%d [oos %s] page id %d gpt %p type %d\n",
+			id, tag, page_id, gpt, type);
+	),
+
+	TP_printk("%s", __entry->buf)
+);
+
+TRACE_EVENT(oos_sync,
+	TP_PROTO(int id, int page_id, void *gpt, int type, u64 v,
+		unsigned long index),
+
+	TP_ARGS(id, page_id, gpt, type, v, index),
+
+	TP_STRUCT__entry(
+		__array(char, buf, MAX_BUF_LEN)
+	),
+
+	TP_fast_assign(
+	snprintf(__entry->buf, MAX_BUF_LEN,
+	"VM%d [oos sync] page id %d gpt %p type %d entry 0x%llx index 0x%lx\n",
+				id, page_id, gpt, type, v, index);
+	),
+
+	TP_printk("%s", __entry->buf)
+);
+
+#define GVT_CMD_STR_LEN 40
+TRACE_EVENT(gvt_command,
+	TP_PROTO(u8 vgpu_id, u8 ring_id, u32 ip_gma, u32 *cmd_va,
+		u32 cmd_len,  u32 buf_type, u32 buf_addr_type,
+		void *workload, const char *cmd_name),
+
+	TP_ARGS(vgpu_id, ring_id, ip_gma, cmd_va, cmd_len, buf_type,
+		buf_addr_type, workload, cmd_name),
+
+	TP_STRUCT__entry(
+		__field(u8, vgpu_id)
+		__field(u8, ring_id)
+		__field(u32, ip_gma)
+		__field(u32, buf_type)
+		__field(u32, buf_addr_type)
+		__field(u32, cmd_len)
+		__field(void*, workload)
+		__dynamic_array(u32, raw_cmd, cmd_len)
+		__array(char, cmd_name, GVT_CMD_STR_LEN)
+	),
+
+	TP_fast_assign(
+		__entry->vgpu_id = vgpu_id;
+		__entry->ring_id = ring_id;
+		__entry->ip_gma = ip_gma;
+		__entry->buf_type = buf_type;
+		__entry->buf_addr_type = buf_addr_type;
+		__entry->cmd_len = cmd_len;
+		__entry->workload = workload;
+		snprintf(__entry->cmd_name, GVT_CMD_STR_LEN, "%s", cmd_name);
+		memcpy(__get_dynamic_array(raw_cmd), cmd_va, cmd_len * sizeof(*cmd_va));
+	),
+
+
+	TP_printk("vgpu%d ring %d: address_type %u, buf_type %u, ip_gma %08x,cmd (name=%s,len=%u,raw cmd=%s), workload=%p\n",
+		__entry->vgpu_id,
+		__entry->ring_id,
+		__entry->buf_addr_type,
+		__entry->buf_type,
+		__entry->ip_gma,
+		__entry->cmd_name,
+		__entry->cmd_len,
+		__print_array(__get_dynamic_array(raw_cmd),
+			__entry->cmd_len, 4),
+		__entry->workload)
+);
+
+#define GVT_TEMP_STR_LEN 10
+TRACE_EVENT(write_ir,
+	TP_PROTO(int id, char *reg_name, unsigned int reg, unsigned int new_val,
+		 unsigned int old_val, bool changed),
+
+	TP_ARGS(id, reg_name, reg, new_val, old_val, changed),
+
+	TP_STRUCT__entry(
+		__field(int, id)
+		__array(char, buf, GVT_TEMP_STR_LEN)
+		__field(unsigned int, reg)
+		__field(unsigned int, new_val)
+		__field(unsigned int, old_val)
+		__field(bool, changed)
+	),
+
+	TP_fast_assign(
+		__entry->id = id;
+		snprintf(__entry->buf, GVT_TEMP_STR_LEN, "%s", reg_name);
+		__entry->reg = reg;
+		__entry->new_val = new_val;
+		__entry->old_val = old_val;
+		__entry->changed = changed;
+	),
+
+	TP_printk("VM%u write [%s] %x, new %08x, old %08x, changed %08x\n",
+		  __entry->id, __entry->buf, __entry->reg, __entry->new_val,
+		  __entry->old_val, __entry->changed)
+);
+
+TRACE_EVENT(propagate_event,
+	TP_PROTO(int id, const char *irq_name, int bit),
+
+	TP_ARGS(id, irq_name, bit),
+
+	TP_STRUCT__entry(
+		__field(int, id)
+		__array(char, buf, GVT_TEMP_STR_LEN)
+		__field(int, bit)
+	),
+
+	TP_fast_assign(
+		__entry->id = id;
+		snprintf(__entry->buf, GVT_TEMP_STR_LEN, "%s", irq_name);
+		__entry->bit = bit;
+	),
+
+	TP_printk("Set bit (%d) for (%s) for vgpu (%d)\n",
+		  __entry->bit, __entry->buf, __entry->id)
+);
+
+TRACE_EVENT(inject_msi,
+	TP_PROTO(int id, unsigned int address, unsigned int data),
+
+	TP_ARGS(id, address, data),
+
+	TP_STRUCT__entry(
+		__field(int, id)
+		__field(unsigned int, address)
+		__field(unsigned int, data)
+	),
+
+	TP_fast_assign(
+		__entry->id = id;
+		__entry->address = address;
+		__entry->data = data;
+	),
+
+	TP_printk("vgpu%d:inject msi address %x data %x\n",
+		  __entry->id, __entry->address, __entry->data)
+);
+
+TRACE_EVENT(render_mmio,
+	TP_PROTO(int old_id, int new_id, char *action, unsigned int reg,
+		 unsigned int old_val, unsigned int new_val),
+
+	TP_ARGS(old_id, new_id, action, reg, old_val, new_val),
+
+	TP_STRUCT__entry(
+		__field(int, old_id)
+		__field(int, new_id)
+		__array(char, buf, GVT_TEMP_STR_LEN)
+		__field(unsigned int, reg)
+		__field(unsigned int, old_val)
+		__field(unsigned int, new_val)
+	),
+
+	TP_fast_assign(
+		__entry->old_id = old_id;
+		__entry->new_id = new_id;
+		snprintf(__entry->buf, GVT_TEMP_STR_LEN, "%s", action);
+		__entry->reg = reg;
+		__entry->old_val = old_val;
+		__entry->new_val = new_val;
+	),
+
+	TP_printk("VM%u -> VM%u %s reg %x, old %08x new %08x\n",
+		  __entry->old_id, __entry->new_id,
+		  __entry->buf, __entry->reg,
+		  __entry->old_val, __entry->new_val)
+);
+
+#endif /* _GVT_TRACE_H_ */
+
+/* This part must be out of protection */
+#undef TRACE_INCLUDE_PATH
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_PATH ../../drivers/gpu/drm/i915/gvt
+#define TRACE_INCLUDE_FILE trace
+#include <trace/define_trace.h>
diff --git a/drivers/gpu/drm/i915/gvt/trace_points.c b/drivers/gpu/drm/i915/gvt/trace_points.c
new file mode 100644
index 000000000..fe552e877
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/trace_points.c
@@ -0,0 +1,34 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Jike Song <jike.song@intel.com>
+ *
+ * Contributors:
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *
+ */
+
+#ifndef __CHECKER__
+#define CREATE_TRACE_POINTS
+#include "trace.h"
+#endif
diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c
new file mode 100644
index 000000000..56c714740
--- /dev/null
+++ b/drivers/gpu/drm/i915/gvt/vgpu.c
@@ -0,0 +1,512 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Authors:
+ *    Eddie Dong <eddie.dong@intel.com>
+ *    Kevin Tian <kevin.tian@intel.com>
+ *
+ * Contributors:
+ *    Ping Gao <ping.a.gao@intel.com>
+ *    Zhi Wang <zhi.a.wang@intel.com>
+ *    Bing Niu <bing.niu@intel.com>
+ *
+ */
+
+#include "i915_drv.h"
+#include "gvt.h"
+#include "i915_pvinfo.h"
+
+void populate_pvinfo_page(struct intel_vgpu *vgpu)
+{
+	struct drm_i915_private *i915 = vgpu->gvt->gt->i915;
+	/* setup the ballooning information */
+	vgpu_vreg64_t(vgpu, vgtif_reg(magic)) = VGT_MAGIC;
+	vgpu_vreg_t(vgpu, vgtif_reg(version_major)) = 1;
+	vgpu_vreg_t(vgpu, vgtif_reg(version_minor)) = 0;
+	vgpu_vreg_t(vgpu, vgtif_reg(display_ready)) = 0;
+	vgpu_vreg_t(vgpu, vgtif_reg(vgt_id)) = vgpu->id;
+
+	vgpu_vreg_t(vgpu, vgtif_reg(vgt_caps)) = VGT_CAPS_FULL_PPGTT;
+	vgpu_vreg_t(vgpu, vgtif_reg(vgt_caps)) |= VGT_CAPS_HWSP_EMULATION;
+	vgpu_vreg_t(vgpu, vgtif_reg(vgt_caps)) |= VGT_CAPS_HUGE_GTT;
+
+	vgpu_vreg_t(vgpu, vgtif_reg(avail_rs.mappable_gmadr.base)) =
+		vgpu_aperture_gmadr_base(vgpu);
+	vgpu_vreg_t(vgpu, vgtif_reg(avail_rs.mappable_gmadr.size)) =
+		vgpu_aperture_sz(vgpu);
+	vgpu_vreg_t(vgpu, vgtif_reg(avail_rs.nonmappable_gmadr.base)) =
+		vgpu_hidden_gmadr_base(vgpu);
+	vgpu_vreg_t(vgpu, vgtif_reg(avail_rs.nonmappable_gmadr.size)) =
+		vgpu_hidden_sz(vgpu);
+
+	vgpu_vreg_t(vgpu, vgtif_reg(avail_rs.fence_num)) = vgpu_fence_sz(vgpu);
+
+	vgpu_vreg_t(vgpu, vgtif_reg(cursor_x_hot)) = UINT_MAX;
+	vgpu_vreg_t(vgpu, vgtif_reg(cursor_y_hot)) = UINT_MAX;
+
+	gvt_dbg_core("Populate PVINFO PAGE for vGPU %d\n", vgpu->id);
+	gvt_dbg_core("aperture base [GMADR] 0x%llx size 0x%llx\n",
+		vgpu_aperture_gmadr_base(vgpu), vgpu_aperture_sz(vgpu));
+	gvt_dbg_core("hidden base [GMADR] 0x%llx size=0x%llx\n",
+		vgpu_hidden_gmadr_base(vgpu), vgpu_hidden_sz(vgpu));
+	gvt_dbg_core("fence size %d\n", vgpu_fence_sz(vgpu));
+
+	drm_WARN_ON(&i915->drm, sizeof(struct vgt_if) != VGT_PVINFO_SIZE);
+}
+
+/*
+ * vGPU type name is defined as GVTg_Vx_y which contains the physical GPU
+ * generation type (e.g V4 as BDW server, V5 as SKL server).
+ *
+ * Depening on the physical SKU resource, we might see vGPU types like
+ * GVTg_V4_8, GVTg_V4_4, GVTg_V4_2, etc. We can create different types of
+ * vGPU on same physical GPU depending on available resource. Each vGPU
+ * type will have a different number of avail_instance to indicate how
+ * many vGPU instance can be created for this type.
+ */
+#define VGPU_MAX_WEIGHT 16
+#define VGPU_WEIGHT(vgpu_num)	\
+	(VGPU_MAX_WEIGHT / (vgpu_num))
+
+static const struct intel_vgpu_config intel_vgpu_configs[] = {
+	{ MB_TO_BYTES(64), MB_TO_BYTES(384), 4, VGPU_WEIGHT(8), GVT_EDID_1024_768, "8" },
+	{ MB_TO_BYTES(128), MB_TO_BYTES(512), 4, VGPU_WEIGHT(4), GVT_EDID_1920_1200, "4" },
+	{ MB_TO_BYTES(256), MB_TO_BYTES(1024), 4, VGPU_WEIGHT(2), GVT_EDID_1920_1200, "2" },
+	{ MB_TO_BYTES(512), MB_TO_BYTES(2048), 4, VGPU_WEIGHT(1), GVT_EDID_1920_1200, "1" },
+};
+
+/**
+ * intel_gvt_init_vgpu_types - initialize vGPU type list
+ * @gvt : GVT device
+ *
+ * Initialize vGPU type list based on available resource.
+ *
+ */
+int intel_gvt_init_vgpu_types(struct intel_gvt *gvt)
+{
+	unsigned int low_avail = gvt_aperture_sz(gvt) - HOST_LOW_GM_SIZE;
+	unsigned int high_avail = gvt_hidden_sz(gvt) - HOST_HIGH_GM_SIZE;
+	unsigned int num_types = ARRAY_SIZE(intel_vgpu_configs);
+	unsigned int i;
+
+	gvt->types = kcalloc(num_types, sizeof(struct intel_vgpu_type),
+			     GFP_KERNEL);
+	if (!gvt->types)
+		return -ENOMEM;
+
+	gvt->mdev_types = kcalloc(num_types, sizeof(*gvt->mdev_types),
+			     GFP_KERNEL);
+	if (!gvt->mdev_types)
+		goto out_free_types;
+
+	for (i = 0; i < num_types; ++i) {
+		const struct intel_vgpu_config *conf = &intel_vgpu_configs[i];
+
+		if (low_avail / conf->low_mm == 0)
+			break;
+		if (conf->weight < 1 || conf->weight > VGPU_MAX_WEIGHT)
+			goto out_free_mdev_types;
+
+		sprintf(gvt->types[i].name, "GVTg_V%u_%s",
+			GRAPHICS_VER(gvt->gt->i915) == 8 ? 4 : 5, conf->name);
+		gvt->types[i].conf = conf;
+
+		gvt_dbg_core("type[%d]: %s avail %u low %u high %u fence %u weight %u res %s\n",
+			     i, gvt->types[i].name,
+			     min(low_avail / conf->low_mm,
+				 high_avail / conf->high_mm),
+			     conf->low_mm, conf->high_mm, conf->fence,
+			     conf->weight, vgpu_edid_str(conf->edid));
+
+		gvt->mdev_types[i] = &gvt->types[i].type;
+		gvt->mdev_types[i]->sysfs_name = gvt->types[i].name;
+	}
+
+	gvt->num_types = i;
+	return 0;
+
+out_free_mdev_types:
+	kfree(gvt->mdev_types);
+out_free_types:
+	kfree(gvt->types);
+	return -EINVAL;
+}
+
+void intel_gvt_clean_vgpu_types(struct intel_gvt *gvt)
+{
+	kfree(gvt->mdev_types);
+	kfree(gvt->types);
+}
+
+/**
+ * intel_gvt_active_vgpu - activate a virtual GPU
+ * @vgpu: virtual GPU
+ *
+ * This function is called when user wants to activate a virtual GPU.
+ *
+ */
+void intel_gvt_activate_vgpu(struct intel_vgpu *vgpu)
+{
+	mutex_lock(&vgpu->vgpu_lock);
+	vgpu->active = true;
+	mutex_unlock(&vgpu->vgpu_lock);
+}
+
+/**
+ * intel_gvt_deactive_vgpu - deactivate a virtual GPU
+ * @vgpu: virtual GPU
+ *
+ * This function is called when user wants to deactivate a virtual GPU.
+ * The virtual GPU will be stopped.
+ *
+ */
+void intel_gvt_deactivate_vgpu(struct intel_vgpu *vgpu)
+{
+	mutex_lock(&vgpu->vgpu_lock);
+
+	vgpu->active = false;
+
+	if (atomic_read(&vgpu->submission.running_workload_num)) {
+		mutex_unlock(&vgpu->vgpu_lock);
+		intel_gvt_wait_vgpu_idle(vgpu);
+		mutex_lock(&vgpu->vgpu_lock);
+	}
+
+	intel_vgpu_stop_schedule(vgpu);
+
+	mutex_unlock(&vgpu->vgpu_lock);
+}
+
+/**
+ * intel_gvt_release_vgpu - release a virtual GPU
+ * @vgpu: virtual GPU
+ *
+ * This function is called when user wants to release a virtual GPU.
+ * The virtual GPU will be stopped and all runtime information will be
+ * destroyed.
+ *
+ */
+void intel_gvt_release_vgpu(struct intel_vgpu *vgpu)
+{
+	intel_gvt_deactivate_vgpu(vgpu);
+
+	mutex_lock(&vgpu->vgpu_lock);
+	vgpu->d3_entered = false;
+	intel_vgpu_clean_workloads(vgpu, ALL_ENGINES);
+	intel_vgpu_dmabuf_cleanup(vgpu);
+	mutex_unlock(&vgpu->vgpu_lock);
+}
+
+/**
+ * intel_gvt_destroy_vgpu - destroy a virtual GPU
+ * @vgpu: virtual GPU
+ *
+ * This function is called when user wants to destroy a virtual GPU.
+ *
+ */
+void intel_gvt_destroy_vgpu(struct intel_vgpu *vgpu)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct drm_i915_private *i915 = gvt->gt->i915;
+
+	drm_WARN(&i915->drm, vgpu->active, "vGPU is still active!\n");
+
+	/*
+	 * remove idr first so later clean can judge if need to stop
+	 * service if no active vgpu.
+	 */
+	mutex_lock(&gvt->lock);
+	idr_remove(&gvt->vgpu_idr, vgpu->id);
+	mutex_unlock(&gvt->lock);
+
+	mutex_lock(&vgpu->vgpu_lock);
+	intel_gvt_debugfs_remove_vgpu(vgpu);
+	intel_vgpu_clean_sched_policy(vgpu);
+	intel_vgpu_clean_submission(vgpu);
+	intel_vgpu_clean_display(vgpu);
+	intel_vgpu_clean_opregion(vgpu);
+	intel_vgpu_reset_ggtt(vgpu, true);
+	intel_vgpu_clean_gtt(vgpu);
+	intel_vgpu_detach_regions(vgpu);
+	intel_vgpu_free_resource(vgpu);
+	intel_vgpu_clean_mmio(vgpu);
+	intel_vgpu_dmabuf_cleanup(vgpu);
+	mutex_unlock(&vgpu->vgpu_lock);
+}
+
+#define IDLE_VGPU_IDR 0
+
+/**
+ * intel_gvt_create_idle_vgpu - create an idle virtual GPU
+ * @gvt: GVT device
+ *
+ * This function is called when user wants to create an idle virtual GPU.
+ *
+ * Returns:
+ * pointer to intel_vgpu, error pointer if failed.
+ */
+struct intel_vgpu *intel_gvt_create_idle_vgpu(struct intel_gvt *gvt)
+{
+	struct intel_vgpu *vgpu;
+	enum intel_engine_id i;
+	int ret;
+
+	vgpu = vzalloc(sizeof(*vgpu));
+	if (!vgpu)
+		return ERR_PTR(-ENOMEM);
+
+	vgpu->id = IDLE_VGPU_IDR;
+	vgpu->gvt = gvt;
+	mutex_init(&vgpu->vgpu_lock);
+
+	for (i = 0; i < I915_NUM_ENGINES; i++)
+		INIT_LIST_HEAD(&vgpu->submission.workload_q_head[i]);
+
+	ret = intel_vgpu_init_sched_policy(vgpu);
+	if (ret)
+		goto out_free_vgpu;
+
+	vgpu->active = false;
+
+	return vgpu;
+
+out_free_vgpu:
+	vfree(vgpu);
+	return ERR_PTR(ret);
+}
+
+/**
+ * intel_gvt_destroy_vgpu - destroy an idle virtual GPU
+ * @vgpu: virtual GPU
+ *
+ * This function is called when user wants to destroy an idle virtual GPU.
+ *
+ */
+void intel_gvt_destroy_idle_vgpu(struct intel_vgpu *vgpu)
+{
+	mutex_lock(&vgpu->vgpu_lock);
+	intel_vgpu_clean_sched_policy(vgpu);
+	mutex_unlock(&vgpu->vgpu_lock);
+
+	vfree(vgpu);
+}
+
+int intel_gvt_create_vgpu(struct intel_vgpu *vgpu,
+		const struct intel_vgpu_config *conf)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct drm_i915_private *dev_priv = gvt->gt->i915;
+	int ret;
+
+	gvt_dbg_core("low %u MB high %u MB fence %u\n",
+			BYTES_TO_MB(conf->low_mm), BYTES_TO_MB(conf->high_mm),
+			conf->fence);
+
+	mutex_lock(&gvt->lock);
+	ret = idr_alloc(&gvt->vgpu_idr, vgpu, IDLE_VGPU_IDR + 1, GVT_MAX_VGPU,
+		GFP_KERNEL);
+	if (ret < 0)
+		goto out_unlock;;
+
+	vgpu->id = ret;
+	vgpu->sched_ctl.weight = conf->weight;
+	mutex_init(&vgpu->vgpu_lock);
+	mutex_init(&vgpu->dmabuf_lock);
+	INIT_LIST_HEAD(&vgpu->dmabuf_obj_list_head);
+	INIT_RADIX_TREE(&vgpu->page_track_tree, GFP_KERNEL);
+	idr_init_base(&vgpu->object_idr, 1);
+	intel_vgpu_init_cfg_space(vgpu, 1);
+	vgpu->d3_entered = false;
+
+	ret = intel_vgpu_init_mmio(vgpu);
+	if (ret)
+		goto out_clean_idr;
+
+	ret = intel_vgpu_alloc_resource(vgpu, conf);
+	if (ret)
+		goto out_clean_vgpu_mmio;
+
+	populate_pvinfo_page(vgpu);
+
+	ret = intel_vgpu_init_gtt(vgpu);
+	if (ret)
+		goto out_clean_vgpu_resource;
+
+	ret = intel_vgpu_init_opregion(vgpu);
+	if (ret)
+		goto out_clean_gtt;
+
+	ret = intel_vgpu_init_display(vgpu, conf->edid);
+	if (ret)
+		goto out_clean_opregion;
+
+	ret = intel_vgpu_setup_submission(vgpu);
+	if (ret)
+		goto out_clean_display;
+
+	ret = intel_vgpu_init_sched_policy(vgpu);
+	if (ret)
+		goto out_clean_submission;
+
+	intel_gvt_debugfs_add_vgpu(vgpu);
+
+	ret = intel_gvt_set_opregion(vgpu);
+	if (ret)
+		goto out_clean_sched_policy;
+
+	if (IS_BROADWELL(dev_priv) || IS_BROXTON(dev_priv))
+		ret = intel_gvt_set_edid(vgpu, PORT_B);
+	else
+		ret = intel_gvt_set_edid(vgpu, PORT_D);
+	if (ret)
+		goto out_clean_sched_policy;
+
+	intel_gvt_update_reg_whitelist(vgpu);
+	mutex_unlock(&gvt->lock);
+	return 0;
+
+out_clean_sched_policy:
+	intel_vgpu_clean_sched_policy(vgpu);
+out_clean_submission:
+	intel_vgpu_clean_submission(vgpu);
+out_clean_display:
+	intel_vgpu_clean_display(vgpu);
+out_clean_opregion:
+	intel_vgpu_clean_opregion(vgpu);
+out_clean_gtt:
+	intel_vgpu_clean_gtt(vgpu);
+out_clean_vgpu_resource:
+	intel_vgpu_free_resource(vgpu);
+out_clean_vgpu_mmio:
+	intel_vgpu_clean_mmio(vgpu);
+out_clean_idr:
+	idr_remove(&gvt->vgpu_idr, vgpu->id);
+out_unlock:
+	mutex_unlock(&gvt->lock);
+	return ret;
+}
+
+/**
+ * intel_gvt_reset_vgpu_locked - reset a virtual GPU by DMLR or GT reset
+ * @vgpu: virtual GPU
+ * @dmlr: vGPU Device Model Level Reset or GT Reset
+ * @engine_mask: engines to reset for GT reset
+ *
+ * This function is called when user wants to reset a virtual GPU through
+ * device model reset or GT reset. The caller should hold the vgpu lock.
+ *
+ * vGPU Device Model Level Reset (DMLR) simulates the PCI level reset to reset
+ * the whole vGPU to default state as when it is created. This vGPU function
+ * is required both for functionary and security concerns.The ultimate goal
+ * of vGPU FLR is that reuse a vGPU instance by virtual machines. When we
+ * assign a vGPU to a virtual machine we must isse such reset first.
+ *
+ * Full GT Reset and Per-Engine GT Reset are soft reset flow for GPU engines
+ * (Render, Blitter, Video, Video Enhancement). It is defined by GPU Spec.
+ * Unlike the FLR, GT reset only reset particular resource of a vGPU per
+ * the reset request. Guest driver can issue a GT reset by programming the
+ * virtual GDRST register to reset specific virtual GPU engine or all
+ * engines.
+ *
+ * The parameter dev_level is to identify if we will do DMLR or GT reset.
+ * The parameter engine_mask is to specific the engines that need to be
+ * resetted. If value ALL_ENGINES is given for engine_mask, it means
+ * the caller requests a full GT reset that we will reset all virtual
+ * GPU engines. For FLR, engine_mask is ignored.
+ */
+void intel_gvt_reset_vgpu_locked(struct intel_vgpu *vgpu, bool dmlr,
+				 intel_engine_mask_t engine_mask)
+{
+	struct intel_gvt *gvt = vgpu->gvt;
+	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
+	intel_engine_mask_t resetting_eng = dmlr ? ALL_ENGINES : engine_mask;
+
+	gvt_dbg_core("------------------------------------------\n");
+	gvt_dbg_core("resseting vgpu%d, dmlr %d, engine_mask %08x\n",
+		     vgpu->id, dmlr, engine_mask);
+
+	vgpu->resetting_eng = resetting_eng;
+
+	intel_vgpu_stop_schedule(vgpu);
+	/*
+	 * The current_vgpu will set to NULL after stopping the
+	 * scheduler when the reset is triggered by current vgpu.
+	 */
+	if (scheduler->current_vgpu == NULL) {
+		mutex_unlock(&vgpu->vgpu_lock);
+		intel_gvt_wait_vgpu_idle(vgpu);
+		mutex_lock(&vgpu->vgpu_lock);
+	}
+
+	intel_vgpu_reset_submission(vgpu, resetting_eng);
+	/* full GPU reset or device model level reset */
+	if (engine_mask == ALL_ENGINES || dmlr) {
+		intel_vgpu_select_submission_ops(vgpu, ALL_ENGINES, 0);
+		if (engine_mask == ALL_ENGINES)
+			intel_vgpu_invalidate_ppgtt(vgpu);
+		/*fence will not be reset during virtual reset */
+		if (dmlr) {
+			if(!vgpu->d3_entered) {
+				intel_vgpu_invalidate_ppgtt(vgpu);
+				intel_vgpu_destroy_all_ppgtt_mm(vgpu);
+			}
+			intel_vgpu_reset_ggtt(vgpu, true);
+			intel_vgpu_reset_resource(vgpu);
+		}
+
+		intel_vgpu_reset_mmio(vgpu, dmlr);
+		populate_pvinfo_page(vgpu);
+
+		if (dmlr) {
+			intel_vgpu_reset_display(vgpu);
+			intel_vgpu_reset_cfg_space(vgpu);
+			/* only reset the failsafe mode when dmlr reset */
+			vgpu->failsafe = false;
+			/*
+			 * PCI_D0 is set before dmlr, so reset d3_entered here
+			 * after done using.
+			 */
+			if(vgpu->d3_entered)
+				vgpu->d3_entered = false;
+			else
+				vgpu->pv_notified = false;
+		}
+	}
+
+	vgpu->resetting_eng = 0;
+	gvt_dbg_core("reset vgpu%d done\n", vgpu->id);
+	gvt_dbg_core("------------------------------------------\n");
+}
+
+/**
+ * intel_gvt_reset_vgpu - reset a virtual GPU (Function Level)
+ * @vgpu: virtual GPU
+ *
+ * This function is called when user wants to reset a virtual GPU.
+ *
+ */
+void intel_gvt_reset_vgpu(struct intel_vgpu *vgpu)
+{
+	mutex_lock(&vgpu->vgpu_lock);
+	intel_gvt_reset_vgpu_locked(vgpu, true, 0);
+	mutex_unlock(&vgpu->vgpu_lock);
+}
diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
new file mode 100644
index 000000000..5ec293011
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_active.c
@@ -0,0 +1,1208 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/debugobjects.h>
+
+#include "gt/intel_context.h"
+#include "gt/intel_engine_heartbeat.h"
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_ring.h"
+
+#include "i915_drv.h"
+#include "i915_active.h"
+
+/*
+ * Active refs memory management
+ *
+ * To be more economical with memory, we reap all the i915_active trees as
+ * they idle (when we know the active requests are inactive) and allocate the
+ * nodes from a local slab cache to hopefully reduce the fragmentation.
+ */
+static struct kmem_cache *slab_cache;
+
+struct active_node {
+	struct rb_node node;
+	struct i915_active_fence base;
+	struct i915_active *ref;
+	u64 timeline;
+};
+
+#define fetch_node(x) rb_entry(READ_ONCE(x), typeof(struct active_node), node)
+
+static inline struct active_node *
+node_from_active(struct i915_active_fence *active)
+{
+	return container_of(active, struct active_node, base);
+}
+
+#define take_preallocated_barriers(x) llist_del_all(&(x)->preallocated_barriers)
+
+static inline bool is_barrier(const struct i915_active_fence *active)
+{
+	return IS_ERR(rcu_access_pointer(active->fence));
+}
+
+static inline struct llist_node *barrier_to_ll(struct active_node *node)
+{
+	GEM_BUG_ON(!is_barrier(&node->base));
+	return (struct llist_node *)&node->base.cb.node;
+}
+
+static inline struct intel_engine_cs *
+__barrier_to_engine(struct active_node *node)
+{
+	return (struct intel_engine_cs *)READ_ONCE(node->base.cb.node.prev);
+}
+
+static inline struct intel_engine_cs *
+barrier_to_engine(struct active_node *node)
+{
+	GEM_BUG_ON(!is_barrier(&node->base));
+	return __barrier_to_engine(node);
+}
+
+static inline struct active_node *barrier_from_ll(struct llist_node *x)
+{
+	return container_of((struct list_head *)x,
+			    struct active_node, base.cb.node);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM) && IS_ENABLED(CONFIG_DEBUG_OBJECTS)
+
+static void *active_debug_hint(void *addr)
+{
+	struct i915_active *ref = addr;
+
+	return (void *)ref->active ?: (void *)ref->retire ?: (void *)ref;
+}
+
+static const struct debug_obj_descr active_debug_desc = {
+	.name = "i915_active",
+	.debug_hint = active_debug_hint,
+};
+
+static void debug_active_init(struct i915_active *ref)
+{
+	debug_object_init(ref, &active_debug_desc);
+}
+
+static void debug_active_activate(struct i915_active *ref)
+{
+	lockdep_assert_held(&ref->tree_lock);
+	debug_object_activate(ref, &active_debug_desc);
+}
+
+static void debug_active_deactivate(struct i915_active *ref)
+{
+	lockdep_assert_held(&ref->tree_lock);
+	if (!atomic_read(&ref->count)) /* after the last dec */
+		debug_object_deactivate(ref, &active_debug_desc);
+}
+
+static void debug_active_fini(struct i915_active *ref)
+{
+	debug_object_free(ref, &active_debug_desc);
+}
+
+static void debug_active_assert(struct i915_active *ref)
+{
+	debug_object_assert_init(ref, &active_debug_desc);
+}
+
+#else
+
+static inline void debug_active_init(struct i915_active *ref) { }
+static inline void debug_active_activate(struct i915_active *ref) { }
+static inline void debug_active_deactivate(struct i915_active *ref) { }
+static inline void debug_active_fini(struct i915_active *ref) { }
+static inline void debug_active_assert(struct i915_active *ref) { }
+
+#endif
+
+static void
+__active_retire(struct i915_active *ref)
+{
+	struct rb_root root = RB_ROOT;
+	struct active_node *it, *n;
+	unsigned long flags;
+
+	GEM_BUG_ON(i915_active_is_idle(ref));
+
+	/* return the unused nodes to our slabcache -- flushing the allocator */
+	if (!atomic_dec_and_lock_irqsave(&ref->count, &ref->tree_lock, flags))
+		return;
+
+	GEM_BUG_ON(rcu_access_pointer(ref->excl.fence));
+	debug_active_deactivate(ref);
+
+	/* Even if we have not used the cache, we may still have a barrier */
+	if (!ref->cache)
+		ref->cache = fetch_node(ref->tree.rb_node);
+
+	/* Keep the MRU cached node for reuse */
+	if (ref->cache) {
+		/* Discard all other nodes in the tree */
+		rb_erase(&ref->cache->node, &ref->tree);
+		root = ref->tree;
+
+		/* Rebuild the tree with only the cached node */
+		rb_link_node(&ref->cache->node, NULL, &ref->tree.rb_node);
+		rb_insert_color(&ref->cache->node, &ref->tree);
+		GEM_BUG_ON(ref->tree.rb_node != &ref->cache->node);
+
+		/* Make the cached node available for reuse with any timeline */
+		ref->cache->timeline = 0; /* needs cmpxchg(u64) */
+	}
+
+	spin_unlock_irqrestore(&ref->tree_lock, flags);
+
+	/* After the final retire, the entire struct may be freed */
+	if (ref->retire)
+		ref->retire(ref);
+
+	/* ... except if you wait on it, you must manage your own references! */
+	wake_up_var(ref);
+
+	/* Finally free the discarded timeline tree  */
+	rbtree_postorder_for_each_entry_safe(it, n, &root, node) {
+		GEM_BUG_ON(i915_active_fence_isset(&it->base));
+		kmem_cache_free(slab_cache, it);
+	}
+}
+
+static void
+active_work(struct work_struct *wrk)
+{
+	struct i915_active *ref = container_of(wrk, typeof(*ref), work);
+
+	GEM_BUG_ON(!atomic_read(&ref->count));
+	if (atomic_add_unless(&ref->count, -1, 1))
+		return;
+
+	__active_retire(ref);
+}
+
+static void
+active_retire(struct i915_active *ref)
+{
+	GEM_BUG_ON(!atomic_read(&ref->count));
+	if (atomic_add_unless(&ref->count, -1, 1))
+		return;
+
+	if (ref->flags & I915_ACTIVE_RETIRE_SLEEPS) {
+		queue_work(system_unbound_wq, &ref->work);
+		return;
+	}
+
+	__active_retire(ref);
+}
+
+static inline struct dma_fence **
+__active_fence_slot(struct i915_active_fence *active)
+{
+	return (struct dma_fence ** __force)&active->fence;
+}
+
+static inline bool
+active_fence_cb(struct dma_fence *fence, struct dma_fence_cb *cb)
+{
+	struct i915_active_fence *active =
+		container_of(cb, typeof(*active), cb);
+
+	return cmpxchg(__active_fence_slot(active), fence, NULL) == fence;
+}
+
+static void
+node_retire(struct dma_fence *fence, struct dma_fence_cb *cb)
+{
+	if (active_fence_cb(fence, cb))
+		active_retire(container_of(cb, struct active_node, base.cb)->ref);
+}
+
+static void
+excl_retire(struct dma_fence *fence, struct dma_fence_cb *cb)
+{
+	if (active_fence_cb(fence, cb))
+		active_retire(container_of(cb, struct i915_active, excl.cb));
+}
+
+static struct active_node *__active_lookup(struct i915_active *ref, u64 idx)
+{
+	struct active_node *it;
+
+	GEM_BUG_ON(idx == 0); /* 0 is the unordered timeline, rsvd for cache */
+
+	/*
+	 * We track the most recently used timeline to skip a rbtree search
+	 * for the common case, under typical loads we never need the rbtree
+	 * at all. We can reuse the last slot if it is empty, that is
+	 * after the previous activity has been retired, or if it matches the
+	 * current timeline.
+	 */
+	it = READ_ONCE(ref->cache);
+	if (it) {
+		u64 cached = READ_ONCE(it->timeline);
+
+		/* Once claimed, this slot will only belong to this idx */
+		if (cached == idx)
+			return it;
+
+		/*
+		 * An unclaimed cache [.timeline=0] can only be claimed once.
+		 *
+		 * If the value is already non-zero, some other thread has
+		 * claimed the cache and we know that is does not match our
+		 * idx. If, and only if, the timeline is currently zero is it
+		 * worth competing to claim it atomically for ourselves (for
+		 * only the winner of that race will cmpxchg return the old
+		 * value of 0).
+		 */
+		if (!cached && !cmpxchg64(&it->timeline, 0, idx))
+			return it;
+	}
+
+	BUILD_BUG_ON(offsetof(typeof(*it), node));
+
+	/* While active, the tree can only be built; not destroyed */
+	GEM_BUG_ON(i915_active_is_idle(ref));
+
+	it = fetch_node(ref->tree.rb_node);
+	while (it) {
+		if (it->timeline < idx) {
+			it = fetch_node(it->node.rb_right);
+		} else if (it->timeline > idx) {
+			it = fetch_node(it->node.rb_left);
+		} else {
+			WRITE_ONCE(ref->cache, it);
+			break;
+		}
+	}
+
+	/* NB: If the tree rotated beneath us, we may miss our target. */
+	return it;
+}
+
+static struct i915_active_fence *
+active_instance(struct i915_active *ref, u64 idx)
+{
+	struct active_node *node;
+	struct rb_node **p, *parent;
+
+	node = __active_lookup(ref, idx);
+	if (likely(node))
+		return &node->base;
+
+	spin_lock_irq(&ref->tree_lock);
+	GEM_BUG_ON(i915_active_is_idle(ref));
+
+	parent = NULL;
+	p = &ref->tree.rb_node;
+	while (*p) {
+		parent = *p;
+
+		node = rb_entry(parent, struct active_node, node);
+		if (node->timeline == idx)
+			goto out;
+
+		if (node->timeline < idx)
+			p = &parent->rb_right;
+		else
+			p = &parent->rb_left;
+	}
+
+	/*
+	 * XXX: We should preallocate this before i915_active_ref() is ever
+	 *  called, but we cannot call into fs_reclaim() anyway, so use GFP_ATOMIC.
+	 */
+	node = kmem_cache_alloc(slab_cache, GFP_ATOMIC);
+	if (!node)
+		goto out;
+
+	__i915_active_fence_init(&node->base, NULL, node_retire);
+	node->ref = ref;
+	node->timeline = idx;
+
+	rb_link_node(&node->node, parent, p);
+	rb_insert_color(&node->node, &ref->tree);
+
+out:
+	WRITE_ONCE(ref->cache, node);
+	spin_unlock_irq(&ref->tree_lock);
+
+	return &node->base;
+}
+
+void __i915_active_init(struct i915_active *ref,
+			int (*active)(struct i915_active *ref),
+			void (*retire)(struct i915_active *ref),
+			unsigned long flags,
+			struct lock_class_key *mkey,
+			struct lock_class_key *wkey)
+{
+	debug_active_init(ref);
+
+	ref->flags = flags;
+	ref->active = active;
+	ref->retire = retire;
+
+	spin_lock_init(&ref->tree_lock);
+	ref->tree = RB_ROOT;
+	ref->cache = NULL;
+
+	init_llist_head(&ref->preallocated_barriers);
+	atomic_set(&ref->count, 0);
+	__mutex_init(&ref->mutex, "i915_active", mkey);
+	__i915_active_fence_init(&ref->excl, NULL, excl_retire);
+	INIT_WORK(&ref->work, active_work);
+#if IS_ENABLED(CONFIG_LOCKDEP)
+	lockdep_init_map(&ref->work.lockdep_map, "i915_active.work", wkey, 0);
+#endif
+}
+
+static bool ____active_del_barrier(struct i915_active *ref,
+				   struct active_node *node,
+				   struct intel_engine_cs *engine)
+
+{
+	struct llist_node *head = NULL, *tail = NULL;
+	struct llist_node *pos, *next;
+
+	GEM_BUG_ON(node->timeline != engine->kernel_context->timeline->fence_context);
+
+	/*
+	 * Rebuild the llist excluding our node. We may perform this
+	 * outside of the kernel_context timeline mutex and so someone
+	 * else may be manipulating the engine->barrier_tasks, in
+	 * which case either we or they will be upset :)
+	 *
+	 * A second __active_del_barrier() will report failure to claim
+	 * the active_node and the caller will just shrug and know not to
+	 * claim ownership of its node.
+	 *
+	 * A concurrent i915_request_add_active_barriers() will miss adding
+	 * any of the tasks, but we will try again on the next -- and since
+	 * we are actively using the barrier, we know that there will be
+	 * at least another opportunity when we idle.
+	 */
+	llist_for_each_safe(pos, next, llist_del_all(&engine->barrier_tasks)) {
+		if (node == barrier_from_ll(pos)) {
+			node = NULL;
+			continue;
+		}
+
+		pos->next = head;
+		head = pos;
+		if (!tail)
+			tail = pos;
+	}
+	if (head)
+		llist_add_batch(head, tail, &engine->barrier_tasks);
+
+	return !node;
+}
+
+static bool
+__active_del_barrier(struct i915_active *ref, struct active_node *node)
+{
+	return ____active_del_barrier(ref, node, barrier_to_engine(node));
+}
+
+static bool
+replace_barrier(struct i915_active *ref, struct i915_active_fence *active)
+{
+	if (!is_barrier(active)) /* proto-node used by our idle barrier? */
+		return false;
+
+	/*
+	 * This request is on the kernel_context timeline, and so
+	 * we can use it to substitute for the pending idle-barrer
+	 * request that we want to emit on the kernel_context.
+	 */
+	return __active_del_barrier(ref, node_from_active(active));
+}
+
+int i915_active_add_request(struct i915_active *ref, struct i915_request *rq)
+{
+	u64 idx = i915_request_timeline(rq)->fence_context;
+	struct dma_fence *fence = &rq->fence;
+	struct i915_active_fence *active;
+	int err;
+
+	/* Prevent reaping in case we malloc/wait while building the tree */
+	err = i915_active_acquire(ref);
+	if (err)
+		return err;
+
+	do {
+		active = active_instance(ref, idx);
+		if (!active) {
+			err = -ENOMEM;
+			goto out;
+		}
+
+		if (replace_barrier(ref, active)) {
+			RCU_INIT_POINTER(active->fence, NULL);
+			atomic_dec(&ref->count);
+		}
+	} while (unlikely(is_barrier(active)));
+
+	fence = __i915_active_fence_set(active, fence);
+	if (!fence)
+		__i915_active_acquire(ref);
+	else
+		dma_fence_put(fence);
+
+out:
+	i915_active_release(ref);
+	return err;
+}
+
+static struct dma_fence *
+__i915_active_set_fence(struct i915_active *ref,
+			struct i915_active_fence *active,
+			struct dma_fence *fence)
+{
+	struct dma_fence *prev;
+
+	if (replace_barrier(ref, active)) {
+		RCU_INIT_POINTER(active->fence, fence);
+		return NULL;
+	}
+
+	prev = __i915_active_fence_set(active, fence);
+	if (!prev)
+		__i915_active_acquire(ref);
+
+	return prev;
+}
+
+struct dma_fence *
+i915_active_set_exclusive(struct i915_active *ref, struct dma_fence *f)
+{
+	/* We expect the caller to manage the exclusive timeline ordering */
+	return __i915_active_set_fence(ref, &ref->excl, f);
+}
+
+bool i915_active_acquire_if_busy(struct i915_active *ref)
+{
+	debug_active_assert(ref);
+	return atomic_add_unless(&ref->count, 1, 0);
+}
+
+static void __i915_active_activate(struct i915_active *ref)
+{
+	spin_lock_irq(&ref->tree_lock); /* __active_retire() */
+	if (!atomic_fetch_inc(&ref->count))
+		debug_active_activate(ref);
+	spin_unlock_irq(&ref->tree_lock);
+}
+
+int i915_active_acquire(struct i915_active *ref)
+{
+	int err;
+
+	if (i915_active_acquire_if_busy(ref))
+		return 0;
+
+	if (!ref->active) {
+		__i915_active_activate(ref);
+		return 0;
+	}
+
+	err = mutex_lock_interruptible(&ref->mutex);
+	if (err)
+		return err;
+
+	if (likely(!i915_active_acquire_if_busy(ref))) {
+		err = ref->active(ref);
+		if (!err)
+			__i915_active_activate(ref);
+	}
+
+	mutex_unlock(&ref->mutex);
+
+	return err;
+}
+
+int i915_active_acquire_for_context(struct i915_active *ref, u64 idx)
+{
+	struct i915_active_fence *active;
+	int err;
+
+	err = i915_active_acquire(ref);
+	if (err)
+		return err;
+
+	active = active_instance(ref, idx);
+	if (!active) {
+		i915_active_release(ref);
+		return -ENOMEM;
+	}
+
+	return 0; /* return with active ref */
+}
+
+void i915_active_release(struct i915_active *ref)
+{
+	debug_active_assert(ref);
+	active_retire(ref);
+}
+
+static void enable_signaling(struct i915_active_fence *active)
+{
+	struct dma_fence *fence;
+
+	if (unlikely(is_barrier(active)))
+		return;
+
+	fence = i915_active_fence_get(active);
+	if (!fence)
+		return;
+
+	dma_fence_enable_sw_signaling(fence);
+	dma_fence_put(fence);
+}
+
+static int flush_barrier(struct active_node *it)
+{
+	struct intel_engine_cs *engine;
+
+	if (likely(!is_barrier(&it->base)))
+		return 0;
+
+	engine = __barrier_to_engine(it);
+	smp_rmb(); /* serialise with add_active_barriers */
+	if (!is_barrier(&it->base))
+		return 0;
+
+	return intel_engine_flush_barriers(engine);
+}
+
+static int flush_lazy_signals(struct i915_active *ref)
+{
+	struct active_node *it, *n;
+	int err = 0;
+
+	enable_signaling(&ref->excl);
+	rbtree_postorder_for_each_entry_safe(it, n, &ref->tree, node) {
+		err = flush_barrier(it); /* unconnected idle barrier? */
+		if (err)
+			break;
+
+		enable_signaling(&it->base);
+	}
+
+	return err;
+}
+
+int __i915_active_wait(struct i915_active *ref, int state)
+{
+	might_sleep();
+
+	/* Any fence added after the wait begins will not be auto-signaled */
+	if (i915_active_acquire_if_busy(ref)) {
+		int err;
+
+		err = flush_lazy_signals(ref);
+		i915_active_release(ref);
+		if (err)
+			return err;
+
+		if (___wait_var_event(ref, i915_active_is_idle(ref),
+				      state, 0, 0, schedule()))
+			return -EINTR;
+	}
+
+	/*
+	 * After the wait is complete, the caller may free the active.
+	 * We have to flush any concurrent retirement before returning.
+	 */
+	flush_work(&ref->work);
+	return 0;
+}
+
+static int __await_active(struct i915_active_fence *active,
+			  int (*fn)(void *arg, struct dma_fence *fence),
+			  void *arg)
+{
+	struct dma_fence *fence;
+
+	if (is_barrier(active)) /* XXX flush the barrier? */
+		return 0;
+
+	fence = i915_active_fence_get(active);
+	if (fence) {
+		int err;
+
+		err = fn(arg, fence);
+		dma_fence_put(fence);
+		if (err < 0)
+			return err;
+	}
+
+	return 0;
+}
+
+struct wait_barrier {
+	struct wait_queue_entry base;
+	struct i915_active *ref;
+};
+
+static int
+barrier_wake(wait_queue_entry_t *wq, unsigned int mode, int flags, void *key)
+{
+	struct wait_barrier *wb = container_of(wq, typeof(*wb), base);
+
+	if (i915_active_is_idle(wb->ref)) {
+		list_del(&wq->entry);
+		i915_sw_fence_complete(wq->private);
+		kfree(wq);
+	}
+
+	return 0;
+}
+
+static int __await_barrier(struct i915_active *ref, struct i915_sw_fence *fence)
+{
+	struct wait_barrier *wb;
+
+	wb = kmalloc(sizeof(*wb), GFP_KERNEL);
+	if (unlikely(!wb))
+		return -ENOMEM;
+
+	GEM_BUG_ON(i915_active_is_idle(ref));
+	if (!i915_sw_fence_await(fence)) {
+		kfree(wb);
+		return -EINVAL;
+	}
+
+	wb->base.flags = 0;
+	wb->base.func = barrier_wake;
+	wb->base.private = fence;
+	wb->ref = ref;
+
+	add_wait_queue(__var_waitqueue(ref), &wb->base);
+	return 0;
+}
+
+static int await_active(struct i915_active *ref,
+			unsigned int flags,
+			int (*fn)(void *arg, struct dma_fence *fence),
+			void *arg, struct i915_sw_fence *barrier)
+{
+	int err = 0;
+
+	if (!i915_active_acquire_if_busy(ref))
+		return 0;
+
+	if (flags & I915_ACTIVE_AWAIT_EXCL &&
+	    rcu_access_pointer(ref->excl.fence)) {
+		err = __await_active(&ref->excl, fn, arg);
+		if (err)
+			goto out;
+	}
+
+	if (flags & I915_ACTIVE_AWAIT_ACTIVE) {
+		struct active_node *it, *n;
+
+		rbtree_postorder_for_each_entry_safe(it, n, &ref->tree, node) {
+			err = __await_active(&it->base, fn, arg);
+			if (err)
+				goto out;
+		}
+	}
+
+	if (flags & I915_ACTIVE_AWAIT_BARRIER) {
+		err = flush_lazy_signals(ref);
+		if (err)
+			goto out;
+
+		err = __await_barrier(ref, barrier);
+		if (err)
+			goto out;
+	}
+
+out:
+	i915_active_release(ref);
+	return err;
+}
+
+static int rq_await_fence(void *arg, struct dma_fence *fence)
+{
+	return i915_request_await_dma_fence(arg, fence);
+}
+
+int i915_request_await_active(struct i915_request *rq,
+			      struct i915_active *ref,
+			      unsigned int flags)
+{
+	return await_active(ref, flags, rq_await_fence, rq, &rq->submit);
+}
+
+static int sw_await_fence(void *arg, struct dma_fence *fence)
+{
+	return i915_sw_fence_await_dma_fence(arg, fence, 0,
+					     GFP_NOWAIT | __GFP_NOWARN);
+}
+
+int i915_sw_fence_await_active(struct i915_sw_fence *fence,
+			       struct i915_active *ref,
+			       unsigned int flags)
+{
+	return await_active(ref, flags, sw_await_fence, fence, fence);
+}
+
+void i915_active_fini(struct i915_active *ref)
+{
+	debug_active_fini(ref);
+	GEM_BUG_ON(atomic_read(&ref->count));
+	GEM_BUG_ON(work_pending(&ref->work));
+	mutex_destroy(&ref->mutex);
+
+	if (ref->cache)
+		kmem_cache_free(slab_cache, ref->cache);
+}
+
+static inline bool is_idle_barrier(struct active_node *node, u64 idx)
+{
+	return node->timeline == idx && !i915_active_fence_isset(&node->base);
+}
+
+static struct active_node *reuse_idle_barrier(struct i915_active *ref, u64 idx)
+{
+	struct rb_node *prev, *p;
+
+	if (RB_EMPTY_ROOT(&ref->tree))
+		return NULL;
+
+	GEM_BUG_ON(i915_active_is_idle(ref));
+
+	/*
+	 * Try to reuse any existing barrier nodes already allocated for this
+	 * i915_active, due to overlapping active phases there is likely a
+	 * node kept alive (as we reuse before parking). We prefer to reuse
+	 * completely idle barriers (less hassle in manipulating the llists),
+	 * but otherwise any will do.
+	 */
+	if (ref->cache && is_idle_barrier(ref->cache, idx)) {
+		p = &ref->cache->node;
+		goto match;
+	}
+
+	prev = NULL;
+	p = ref->tree.rb_node;
+	while (p) {
+		struct active_node *node =
+			rb_entry(p, struct active_node, node);
+
+		if (is_idle_barrier(node, idx))
+			goto match;
+
+		prev = p;
+		if (node->timeline < idx)
+			p = READ_ONCE(p->rb_right);
+		else
+			p = READ_ONCE(p->rb_left);
+	}
+
+	/*
+	 * No quick match, but we did find the leftmost rb_node for the
+	 * kernel_context. Walk the rb_tree in-order to see if there were
+	 * any idle-barriers on this timeline that we missed, or just use
+	 * the first pending barrier.
+	 */
+	for (p = prev; p; p = rb_next(p)) {
+		struct active_node *node =
+			rb_entry(p, struct active_node, node);
+		struct intel_engine_cs *engine;
+
+		if (node->timeline > idx)
+			break;
+
+		if (node->timeline < idx)
+			continue;
+
+		if (is_idle_barrier(node, idx))
+			goto match;
+
+		/*
+		 * The list of pending barriers is protected by the
+		 * kernel_context timeline, which notably we do not hold
+		 * here. i915_request_add_active_barriers() may consume
+		 * the barrier before we claim it, so we have to check
+		 * for success.
+		 */
+		engine = __barrier_to_engine(node);
+		smp_rmb(); /* serialise with add_active_barriers */
+		if (is_barrier(&node->base) &&
+		    ____active_del_barrier(ref, node, engine))
+			goto match;
+	}
+
+	return NULL;
+
+match:
+	spin_lock_irq(&ref->tree_lock);
+	rb_erase(p, &ref->tree); /* Hide from waits and sibling allocations */
+	if (p == &ref->cache->node)
+		WRITE_ONCE(ref->cache, NULL);
+	spin_unlock_irq(&ref->tree_lock);
+
+	return rb_entry(p, struct active_node, node);
+}
+
+int i915_active_acquire_preallocate_barrier(struct i915_active *ref,
+					    struct intel_engine_cs *engine)
+{
+	intel_engine_mask_t tmp, mask = engine->mask;
+	struct llist_node *first = NULL, *last = NULL;
+	struct intel_gt *gt = engine->gt;
+
+	GEM_BUG_ON(i915_active_is_idle(ref));
+
+	/* Wait until the previous preallocation is completed */
+	while (!llist_empty(&ref->preallocated_barriers))
+		cond_resched();
+
+	/*
+	 * Preallocate a node for each physical engine supporting the target
+	 * engine (remember virtual engines have more than one sibling).
+	 * We can then use the preallocated nodes in
+	 * i915_active_acquire_barrier()
+	 */
+	GEM_BUG_ON(!mask);
+	for_each_engine_masked(engine, gt, mask, tmp) {
+		u64 idx = engine->kernel_context->timeline->fence_context;
+		struct llist_node *prev = first;
+		struct active_node *node;
+
+		rcu_read_lock();
+		node = reuse_idle_barrier(ref, idx);
+		rcu_read_unlock();
+		if (!node) {
+			node = kmem_cache_alloc(slab_cache, GFP_KERNEL);
+			if (!node)
+				goto unwind;
+
+			RCU_INIT_POINTER(node->base.fence, NULL);
+			node->base.cb.func = node_retire;
+			node->timeline = idx;
+			node->ref = ref;
+		}
+
+		if (!i915_active_fence_isset(&node->base)) {
+			/*
+			 * Mark this as being *our* unconnected proto-node.
+			 *
+			 * Since this node is not in any list, and we have
+			 * decoupled it from the rbtree, we can reuse the
+			 * request to indicate this is an idle-barrier node
+			 * and then we can use the rb_node and list pointers
+			 * for our tracking of the pending barrier.
+			 */
+			RCU_INIT_POINTER(node->base.fence, ERR_PTR(-EAGAIN));
+			node->base.cb.node.prev = (void *)engine;
+			__i915_active_acquire(ref);
+		}
+		GEM_BUG_ON(rcu_access_pointer(node->base.fence) != ERR_PTR(-EAGAIN));
+
+		GEM_BUG_ON(barrier_to_engine(node) != engine);
+		first = barrier_to_ll(node);
+		first->next = prev;
+		if (!last)
+			last = first;
+		intel_engine_pm_get(engine);
+	}
+
+	GEM_BUG_ON(!llist_empty(&ref->preallocated_barriers));
+	llist_add_batch(first, last, &ref->preallocated_barriers);
+
+	return 0;
+
+unwind:
+	while (first) {
+		struct active_node *node = barrier_from_ll(first);
+
+		first = first->next;
+
+		atomic_dec(&ref->count);
+		intel_engine_pm_put(barrier_to_engine(node));
+
+		kmem_cache_free(slab_cache, node);
+	}
+	return -ENOMEM;
+}
+
+void i915_active_acquire_barrier(struct i915_active *ref)
+{
+	struct llist_node *pos, *next;
+	unsigned long flags;
+
+	GEM_BUG_ON(i915_active_is_idle(ref));
+
+	/*
+	 * Transfer the list of preallocated barriers into the
+	 * i915_active rbtree, but only as proto-nodes. They will be
+	 * populated by i915_request_add_active_barriers() to point to the
+	 * request that will eventually release them.
+	 */
+	llist_for_each_safe(pos, next, take_preallocated_barriers(ref)) {
+		struct active_node *node = barrier_from_ll(pos);
+		struct intel_engine_cs *engine = barrier_to_engine(node);
+		struct rb_node **p, *parent;
+
+		spin_lock_irqsave_nested(&ref->tree_lock, flags,
+					 SINGLE_DEPTH_NESTING);
+		parent = NULL;
+		p = &ref->tree.rb_node;
+		while (*p) {
+			struct active_node *it;
+
+			parent = *p;
+
+			it = rb_entry(parent, struct active_node, node);
+			if (it->timeline < node->timeline)
+				p = &parent->rb_right;
+			else
+				p = &parent->rb_left;
+		}
+		rb_link_node(&node->node, parent, p);
+		rb_insert_color(&node->node, &ref->tree);
+		spin_unlock_irqrestore(&ref->tree_lock, flags);
+
+		GEM_BUG_ON(!intel_engine_pm_is_awake(engine));
+		llist_add(barrier_to_ll(node), &engine->barrier_tasks);
+		intel_engine_pm_put_delay(engine, 2);
+	}
+}
+
+static struct dma_fence **ll_to_fence_slot(struct llist_node *node)
+{
+	return __active_fence_slot(&barrier_from_ll(node)->base);
+}
+
+void i915_request_add_active_barriers(struct i915_request *rq)
+{
+	struct intel_engine_cs *engine = rq->engine;
+	struct llist_node *node, *next;
+	unsigned long flags;
+
+	GEM_BUG_ON(!intel_context_is_barrier(rq->context));
+	GEM_BUG_ON(intel_engine_is_virtual(engine));
+	GEM_BUG_ON(i915_request_timeline(rq) != engine->kernel_context->timeline);
+
+	node = llist_del_all(&engine->barrier_tasks);
+	if (!node)
+		return;
+	/*
+	 * Attach the list of proto-fences to the in-flight request such
+	 * that the parent i915_active will be released when this request
+	 * is retired.
+	 */
+	spin_lock_irqsave(&rq->lock, flags);
+	llist_for_each_safe(node, next, node) {
+		/* serialise with reuse_idle_barrier */
+		smp_store_mb(*ll_to_fence_slot(node), &rq->fence);
+		list_add_tail((struct list_head *)node, &rq->fence.cb_list);
+	}
+	spin_unlock_irqrestore(&rq->lock, flags);
+}
+
+/*
+ * __i915_active_fence_set: Update the last active fence along its timeline
+ * @active: the active tracker
+ * @fence: the new fence (under construction)
+ *
+ * Records the new @fence as the last active fence along its timeline in
+ * this active tracker, moving the tracking callbacks from the previous
+ * fence onto this one. Gets and returns a reference to the previous fence
+ * (if not already completed), which the caller must put after making sure
+ * that it is executed before the new fence. To ensure that the order of
+ * fences within the timeline of the i915_active_fence is understood, it
+ * should be locked by the caller.
+ */
+struct dma_fence *
+__i915_active_fence_set(struct i915_active_fence *active,
+			struct dma_fence *fence)
+{
+	struct dma_fence *prev;
+	unsigned long flags;
+
+	/*
+	 * In case of fences embedded in i915_requests, their memory is
+	 * SLAB_FAILSAFE_BY_RCU, then it can be reused right after release
+	 * by new requests.  Then, there is a risk of passing back a pointer
+	 * to a new, completely unrelated fence that reuses the same memory
+	 * while tracked under a different active tracker.  Combined with i915
+	 * perf open/close operations that build await dependencies between
+	 * engine kernel context requests and user requests from different
+	 * timelines, this can lead to dependency loops and infinite waits.
+	 *
+	 * As a countermeasure, we try to get a reference to the active->fence
+	 * first, so if we succeed and pass it back to our user then it is not
+	 * released and potentially reused by an unrelated request before the
+	 * user has a chance to set up an await dependency on it.
+	 */
+	prev = i915_active_fence_get(active);
+	if (fence == prev)
+		return fence;
+
+	GEM_BUG_ON(test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags));
+
+	/*
+	 * Consider that we have two threads arriving (A and B), with
+	 * C already resident as the active->fence.
+	 *
+	 * Both A and B have got a reference to C or NULL, depending on the
+	 * timing of the interrupt handler.  Let's assume that if A has got C
+	 * then it has locked C first (before B).
+	 *
+	 * Note the strong ordering of the timeline also provides consistent
+	 * nesting rules for the fence->lock; the inner lock is always the
+	 * older lock.
+	 */
+	spin_lock_irqsave(fence->lock, flags);
+	if (prev)
+		spin_lock_nested(prev->lock, SINGLE_DEPTH_NESTING);
+
+	/*
+	 * A does the cmpxchg first, and so it sees C or NULL, as before, or
+	 * something else, depending on the timing of other threads and/or
+	 * interrupt handler.  If not the same as before then A unlocks C if
+	 * applicable and retries, starting from an attempt to get a new
+	 * active->fence.  Meanwhile, B follows the same path as A.
+	 * Once A succeeds with cmpxch, B fails again, retires, gets A from
+	 * active->fence, locks it as soon as A completes, and possibly
+	 * succeeds with cmpxchg.
+	 */
+	while (cmpxchg(__active_fence_slot(active), prev, fence) != prev) {
+		if (prev) {
+			spin_unlock(prev->lock);
+			dma_fence_put(prev);
+		}
+		spin_unlock_irqrestore(fence->lock, flags);
+
+		prev = i915_active_fence_get(active);
+		GEM_BUG_ON(prev == fence);
+
+		spin_lock_irqsave(fence->lock, flags);
+		if (prev)
+			spin_lock_nested(prev->lock, SINGLE_DEPTH_NESTING);
+	}
+
+	/*
+	 * If prev is NULL then the previous fence must have been signaled
+	 * and we know that we are first on the timeline.  If it is still
+	 * present then, having the lock on that fence already acquired, we
+	 * serialise with the interrupt handler, in the process of removing it
+	 * from any future interrupt callback.  A will then wait on C before
+	 * executing (if present).
+	 *
+	 * As B is second, it sees A as the previous fence and so waits for
+	 * it to complete its transition and takes over the occupancy for
+	 * itself -- remembering that it needs to wait on A before executing.
+	 */
+	if (prev) {
+		__list_del_entry(&active->cb.node);
+		spin_unlock(prev->lock); /* serialise with prev->cb_list */
+	}
+	list_add_tail(&active->cb.node, &fence->cb_list);
+	spin_unlock_irqrestore(fence->lock, flags);
+
+	return prev;
+}
+
+int i915_active_fence_set(struct i915_active_fence *active,
+			  struct i915_request *rq)
+{
+	struct dma_fence *fence;
+	int err = 0;
+
+	/* Must maintain timeline ordering wrt previous active requests */
+	fence = __i915_active_fence_set(active, &rq->fence);
+	if (fence) {
+		err = i915_request_await_dma_fence(rq, fence);
+		dma_fence_put(fence);
+	}
+
+	return err;
+}
+
+void i915_active_noop(struct dma_fence *fence, struct dma_fence_cb *cb)
+{
+	active_fence_cb(fence, cb);
+}
+
+struct auto_active {
+	struct i915_active base;
+	struct kref ref;
+};
+
+struct i915_active *i915_active_get(struct i915_active *ref)
+{
+	struct auto_active *aa = container_of(ref, typeof(*aa), base);
+
+	kref_get(&aa->ref);
+	return &aa->base;
+}
+
+static void auto_release(struct kref *ref)
+{
+	struct auto_active *aa = container_of(ref, typeof(*aa), ref);
+
+	i915_active_fini(&aa->base);
+	kfree(aa);
+}
+
+void i915_active_put(struct i915_active *ref)
+{
+	struct auto_active *aa = container_of(ref, typeof(*aa), base);
+
+	kref_put(&aa->ref, auto_release);
+}
+
+static int auto_active(struct i915_active *ref)
+{
+	i915_active_get(ref);
+	return 0;
+}
+
+static void auto_retire(struct i915_active *ref)
+{
+	i915_active_put(ref);
+}
+
+struct i915_active *i915_active_create(void)
+{
+	struct auto_active *aa;
+
+	aa = kmalloc(sizeof(*aa), GFP_KERNEL);
+	if (!aa)
+		return NULL;
+
+	kref_init(&aa->ref);
+	i915_active_init(&aa->base, auto_active, auto_retire, 0);
+
+	return &aa->base;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/i915_active.c"
+#endif
+
+void i915_active_module_exit(void)
+{
+	kmem_cache_destroy(slab_cache);
+}
+
+int __init i915_active_module_init(void)
+{
+	slab_cache = KMEM_CACHE(active_node, SLAB_HWCACHE_ALIGN);
+	if (!slab_cache)
+		return -ENOMEM;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_active.h b/drivers/gpu/drm/i915/i915_active.h
new file mode 100644
index 000000000..7eb441321
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_active.h
@@ -0,0 +1,238 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef _I915_ACTIVE_H_
+#define _I915_ACTIVE_H_
+
+#include <linux/lockdep.h>
+
+#include "i915_active_types.h"
+#include "i915_request.h"
+
+struct i915_request;
+struct intel_engine_cs;
+struct intel_timeline;
+
+/*
+ * We treat requests as fences. This is not be to confused with our
+ * "fence registers" but pipeline synchronisation objects ala GL_ARB_sync.
+ * We use the fences to synchronize access from the CPU with activity on the
+ * GPU, for example, we should not rewrite an object's PTE whilst the GPU
+ * is reading them. We also track fences at a higher level to provide
+ * implicit synchronisation around GEM objects, e.g. set-domain will wait
+ * for outstanding GPU rendering before marking the object ready for CPU
+ * access, or a pageflip will wait until the GPU is complete before showing
+ * the frame on the scanout.
+ *
+ * In order to use a fence, the object must track the fence it needs to
+ * serialise with. For example, GEM objects want to track both read and
+ * write access so that we can perform concurrent read operations between
+ * the CPU and GPU engines, as well as waiting for all rendering to
+ * complete, or waiting for the last GPU user of a "fence register". The
+ * object then embeds a #i915_active_fence to track the most recent (in
+ * retirement order) request relevant for the desired mode of access.
+ * The #i915_active_fence is updated with i915_active_fence_set() to
+ * track the most recent fence request, typically this is done as part of
+ * i915_vma_move_to_active().
+ *
+ * When the #i915_active_fence completes (is retired), it will
+ * signal its completion to the owner through a callback as well as mark
+ * itself as idle (i915_active_fence.request == NULL). The owner
+ * can then perform any action, such as delayed freeing of an active
+ * resource including itself.
+ */
+
+void i915_active_noop(struct dma_fence *fence, struct dma_fence_cb *cb);
+
+/**
+ * __i915_active_fence_init - prepares the activity tracker for use
+ * @active - the active tracker
+ * @fence - initial fence to track, can be NULL
+ * @func - a callback when then the tracker is retired (becomes idle),
+ *         can be NULL
+ *
+ * i915_active_fence_init() prepares the embedded @active struct for use as
+ * an activity tracker, that is for tracking the last known active fence
+ * associated with it. When the last fence becomes idle, when it is retired
+ * after completion, the optional callback @func is invoked.
+ */
+static inline void
+__i915_active_fence_init(struct i915_active_fence *active,
+			 void *fence,
+			 dma_fence_func_t fn)
+{
+	RCU_INIT_POINTER(active->fence, fence);
+	active->cb.func = fn ?: i915_active_noop;
+}
+
+#define INIT_ACTIVE_FENCE(A) \
+	__i915_active_fence_init((A), NULL, NULL)
+
+struct dma_fence *
+__i915_active_fence_set(struct i915_active_fence *active,
+			struct dma_fence *fence);
+
+/**
+ * i915_active_fence_set - updates the tracker to watch the current fence
+ * @active - the active tracker
+ * @rq - the request to watch
+ *
+ * i915_active_fence_set() watches the given @rq for completion. While
+ * that @rq is busy, the @active reports busy. When that @rq is signaled
+ * (or else retired) the @active tracker is updated to report idle.
+ */
+int __must_check
+i915_active_fence_set(struct i915_active_fence *active,
+		      struct i915_request *rq);
+/**
+ * i915_active_fence_get - return a reference to the active fence
+ * @active - the active tracker
+ *
+ * i915_active_fence_get() returns a reference to the active fence,
+ * or NULL if the active tracker is idle. The reference is obtained under RCU,
+ * so no locking is required by the caller.
+ *
+ * The reference should be freed with dma_fence_put().
+ */
+static inline struct dma_fence *
+i915_active_fence_get(struct i915_active_fence *active)
+{
+	struct dma_fence *fence;
+
+	rcu_read_lock();
+	fence = dma_fence_get_rcu_safe(&active->fence);
+	rcu_read_unlock();
+
+	return fence;
+}
+
+/**
+ * i915_active_fence_isset - report whether the active tracker is assigned
+ * @active - the active tracker
+ *
+ * i915_active_fence_isset() returns true if the active tracker is currently
+ * assigned to a fence. Due to the lazy retiring, that fence may be idle
+ * and this may report stale information.
+ */
+static inline bool
+i915_active_fence_isset(const struct i915_active_fence *active)
+{
+	return rcu_access_pointer(active->fence);
+}
+
+/*
+ * GPU activity tracking
+ *
+ * Each set of commands submitted to the GPU compromises a single request that
+ * signals a fence upon completion. struct i915_request combines the
+ * command submission, scheduling and fence signaling roles. If we want to see
+ * if a particular task is complete, we need to grab the fence (struct
+ * i915_request) for that task and check or wait for it to be signaled. More
+ * often though we want to track the status of a bunch of tasks, for example
+ * to wait for the GPU to finish accessing some memory across a variety of
+ * different command pipelines from different clients. We could choose to
+ * track every single request associated with the task, but knowing that
+ * each request belongs to an ordered timeline (later requests within a
+ * timeline must wait for earlier requests), we need only track the
+ * latest request in each timeline to determine the overall status of the
+ * task.
+ *
+ * struct i915_active provides this tracking across timelines. It builds a
+ * composite shared-fence, and is updated as new work is submitted to the task,
+ * forming a snapshot of the current status. It should be embedded into the
+ * different resources that need to track their associated GPU activity to
+ * provide a callback when that GPU activity has ceased, or otherwise to
+ * provide a serialisation point either for request submission or for CPU
+ * synchronisation.
+ */
+
+void __i915_active_init(struct i915_active *ref,
+			int (*active)(struct i915_active *ref),
+			void (*retire)(struct i915_active *ref),
+			unsigned long flags,
+			struct lock_class_key *mkey,
+			struct lock_class_key *wkey);
+
+/* Specialise each class of i915_active to avoid impossible lockdep cycles. */
+#define i915_active_init(ref, active, retire, flags) do {			\
+	static struct lock_class_key __mkey;					\
+	static struct lock_class_key __wkey;					\
+										\
+	__i915_active_init(ref, active, retire, flags, &__mkey, &__wkey);	\
+} while (0)
+
+int i915_active_add_request(struct i915_active *ref, struct i915_request *rq);
+
+struct dma_fence *
+i915_active_set_exclusive(struct i915_active *ref, struct dma_fence *f);
+
+int __i915_active_wait(struct i915_active *ref, int state);
+static inline int i915_active_wait(struct i915_active *ref)
+{
+	return __i915_active_wait(ref, TASK_INTERRUPTIBLE);
+}
+
+int i915_sw_fence_await_active(struct i915_sw_fence *fence,
+			       struct i915_active *ref,
+			       unsigned int flags);
+int i915_request_await_active(struct i915_request *rq,
+			      struct i915_active *ref,
+			      unsigned int flags);
+#define I915_ACTIVE_AWAIT_EXCL BIT(0)
+#define I915_ACTIVE_AWAIT_ACTIVE BIT(1)
+#define I915_ACTIVE_AWAIT_BARRIER BIT(2)
+
+int i915_active_acquire(struct i915_active *ref);
+int i915_active_acquire_for_context(struct i915_active *ref, u64 idx);
+bool i915_active_acquire_if_busy(struct i915_active *ref);
+
+void i915_active_release(struct i915_active *ref);
+
+static inline void __i915_active_acquire(struct i915_active *ref)
+{
+	GEM_BUG_ON(!atomic_read(&ref->count));
+	atomic_inc(&ref->count);
+}
+
+static inline bool
+i915_active_is_idle(const struct i915_active *ref)
+{
+	return !atomic_read(&ref->count);
+}
+
+void i915_active_fini(struct i915_active *ref);
+
+int i915_active_acquire_preallocate_barrier(struct i915_active *ref,
+					    struct intel_engine_cs *engine);
+void i915_active_acquire_barrier(struct i915_active *ref);
+void i915_request_add_active_barriers(struct i915_request *rq);
+
+void i915_active_print(struct i915_active *ref, struct drm_printer *m);
+void i915_active_unlock_wait(struct i915_active *ref);
+
+struct i915_active *i915_active_create(void);
+struct i915_active *i915_active_get(struct i915_active *ref);
+void i915_active_put(struct i915_active *ref);
+
+static inline int __i915_request_await_exclusive(struct i915_request *rq,
+						 struct i915_active *active)
+{
+	struct dma_fence *fence;
+	int err = 0;
+
+	fence = i915_active_fence_get(&active->excl);
+	if (fence) {
+		err = i915_request_await_dma_fence(rq, fence);
+		dma_fence_put(fence);
+	}
+
+	return err;
+}
+
+void i915_active_module_exit(void);
+int i915_active_module_init(void);
+
+#endif /* _I915_ACTIVE_H_ */
diff --git a/drivers/gpu/drm/i915/i915_active_types.h b/drivers/gpu/drm/i915/i915_active_types.h
new file mode 100644
index 000000000..b02a78ac8
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_active_types.h
@@ -0,0 +1,47 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef _I915_ACTIVE_TYPES_H_
+#define _I915_ACTIVE_TYPES_H_
+
+#include <linux/atomic.h>
+#include <linux/dma-fence.h>
+#include <linux/llist.h>
+#include <linux/mutex.h>
+#include <linux/rbtree.h>
+#include <linux/rcupdate.h>
+#include <linux/workqueue.h>
+
+struct i915_active_fence {
+	struct dma_fence __rcu *fence;
+	struct dma_fence_cb cb;
+};
+
+struct active_node;
+
+struct i915_active {
+	atomic_t count;
+	struct mutex mutex;
+
+	spinlock_t tree_lock;
+	struct active_node *cache;
+	struct rb_root tree;
+
+	/* Preallocated "exclusive" node */
+	struct i915_active_fence excl;
+
+	unsigned long flags;
+#define I915_ACTIVE_RETIRE_SLEEPS BIT(0)
+
+	int (*active)(struct i915_active *ref);
+	void (*retire)(struct i915_active *ref);
+
+	struct work_struct work;
+
+	struct llist_head preallocated_barriers;
+};
+
+#endif /* _I915_ACTIVE_TYPES_H_ */
diff --git a/drivers/gpu/drm/i915/i915_cmd_parser.c b/drivers/gpu/drm/i915/i915_cmd_parser.c
new file mode 100644
index 000000000..f93e6122f
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_cmd_parser.c
@@ -0,0 +1,1624 @@
+/*
+ * Copyright © 2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Brad Volkin <bradley.d.volkin@intel.com>
+ *
+ */
+
+#include <linux/highmem.h>
+
+#include <drm/drm_cache.h>
+
+#include "gt/intel_engine.h"
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt_regs.h"
+
+#include "i915_cmd_parser.h"
+#include "i915_drv.h"
+#include "i915_memcpy.h"
+#include "i915_reg.h"
+
+/**
+ * DOC: batch buffer command parser
+ *
+ * Motivation:
+ * Certain OpenGL features (e.g. transform feedback, performance monitoring)
+ * require userspace code to submit batches containing commands such as
+ * MI_LOAD_REGISTER_IMM to access various registers. Unfortunately, some
+ * generations of the hardware will noop these commands in "unsecure" batches
+ * (which includes all userspace batches submitted via i915) even though the
+ * commands may be safe and represent the intended programming model of the
+ * device.
+ *
+ * The software command parser is similar in operation to the command parsing
+ * done in hardware for unsecure batches. However, the software parser allows
+ * some operations that would be noop'd by hardware, if the parser determines
+ * the operation is safe, and submits the batch as "secure" to prevent hardware
+ * parsing.
+ *
+ * Threats:
+ * At a high level, the hardware (and software) checks attempt to prevent
+ * granting userspace undue privileges. There are three categories of privilege.
+ *
+ * First, commands which are explicitly defined as privileged or which should
+ * only be used by the kernel driver. The parser rejects such commands
+ *
+ * Second, commands which access registers. To support correct/enhanced
+ * userspace functionality, particularly certain OpenGL extensions, the parser
+ * provides a whitelist of registers which userspace may safely access
+ *
+ * Third, commands which access privileged memory (i.e. GGTT, HWS page, etc).
+ * The parser always rejects such commands.
+ *
+ * The majority of the problematic commands fall in the MI_* range, with only a
+ * few specific commands on each engine (e.g. PIPE_CONTROL and MI_FLUSH_DW).
+ *
+ * Implementation:
+ * Each engine maintains tables of commands and registers which the parser
+ * uses in scanning batch buffers submitted to that engine.
+ *
+ * Since the set of commands that the parser must check for is significantly
+ * smaller than the number of commands supported, the parser tables contain only
+ * those commands required by the parser. This generally works because command
+ * opcode ranges have standard command length encodings. So for commands that
+ * the parser does not need to check, it can easily skip them. This is
+ * implemented via a per-engine length decoding vfunc.
+ *
+ * Unfortunately, there are a number of commands that do not follow the standard
+ * length encoding for their opcode range, primarily amongst the MI_* commands.
+ * To handle this, the parser provides a way to define explicit "skip" entries
+ * in the per-engine command tables.
+ *
+ * Other command table entries map fairly directly to high level categories
+ * mentioned above: rejected, register whitelist. The parser implements a number
+ * of checks, including the privileged memory checks, via a general bitmasking
+ * mechanism.
+ */
+
+/*
+ * A command that requires special handling by the command parser.
+ */
+struct drm_i915_cmd_descriptor {
+	/*
+	 * Flags describing how the command parser processes the command.
+	 *
+	 * CMD_DESC_FIXED: The command has a fixed length if this is set,
+	 *                 a length mask if not set
+	 * CMD_DESC_SKIP: The command is allowed but does not follow the
+	 *                standard length encoding for the opcode range in
+	 *                which it falls
+	 * CMD_DESC_REJECT: The command is never allowed
+	 * CMD_DESC_REGISTER: The command should be checked against the
+	 *                    register whitelist for the appropriate ring
+	 */
+	u32 flags;
+#define CMD_DESC_FIXED    (1<<0)
+#define CMD_DESC_SKIP     (1<<1)
+#define CMD_DESC_REJECT   (1<<2)
+#define CMD_DESC_REGISTER (1<<3)
+#define CMD_DESC_BITMASK  (1<<4)
+
+	/*
+	 * The command's unique identification bits and the bitmask to get them.
+	 * This isn't strictly the opcode field as defined in the spec and may
+	 * also include type, subtype, and/or subop fields.
+	 */
+	struct {
+		u32 value;
+		u32 mask;
+	} cmd;
+
+	/*
+	 * The command's length. The command is either fixed length (i.e. does
+	 * not include a length field) or has a length field mask. The flag
+	 * CMD_DESC_FIXED indicates a fixed length. Otherwise, the command has
+	 * a length mask. All command entries in a command table must include
+	 * length information.
+	 */
+	union {
+		u32 fixed;
+		u32 mask;
+	} length;
+
+	/*
+	 * Describes where to find a register address in the command to check
+	 * against the ring's register whitelist. Only valid if flags has the
+	 * CMD_DESC_REGISTER bit set.
+	 *
+	 * A non-zero step value implies that the command may access multiple
+	 * registers in sequence (e.g. LRI), in that case step gives the
+	 * distance in dwords between individual offset fields.
+	 */
+	struct {
+		u32 offset;
+		u32 mask;
+		u32 step;
+	} reg;
+
+#define MAX_CMD_DESC_BITMASKS 3
+	/*
+	 * Describes command checks where a particular dword is masked and
+	 * compared against an expected value. If the command does not match
+	 * the expected value, the parser rejects it. Only valid if flags has
+	 * the CMD_DESC_BITMASK bit set. Only entries where mask is non-zero
+	 * are valid.
+	 *
+	 * If the check specifies a non-zero condition_mask then the parser
+	 * only performs the check when the bits specified by condition_mask
+	 * are non-zero.
+	 */
+	struct {
+		u32 offset;
+		u32 mask;
+		u32 expected;
+		u32 condition_offset;
+		u32 condition_mask;
+	} bits[MAX_CMD_DESC_BITMASKS];
+};
+
+/*
+ * A table of commands requiring special handling by the command parser.
+ *
+ * Each engine has an array of tables. Each table consists of an array of
+ * command descriptors, which must be sorted with command opcodes in
+ * ascending order.
+ */
+struct drm_i915_cmd_table {
+	const struct drm_i915_cmd_descriptor *table;
+	int count;
+};
+
+#define STD_MI_OPCODE_SHIFT  (32 - 9)
+#define STD_3D_OPCODE_SHIFT  (32 - 16)
+#define STD_2D_OPCODE_SHIFT  (32 - 10)
+#define STD_MFX_OPCODE_SHIFT (32 - 16)
+#define MIN_OPCODE_SHIFT 16
+
+#define CMD(op, opm, f, lm, fl, ...)				\
+	{							\
+		.flags = (fl) | ((f) ? CMD_DESC_FIXED : 0),	\
+		.cmd = { (op & ~0u << (opm)), ~0u << (opm) },	\
+		.length = { (lm) },				\
+		__VA_ARGS__					\
+	}
+
+/* Convenience macros to compress the tables */
+#define SMI STD_MI_OPCODE_SHIFT
+#define S3D STD_3D_OPCODE_SHIFT
+#define S2D STD_2D_OPCODE_SHIFT
+#define SMFX STD_MFX_OPCODE_SHIFT
+#define F true
+#define S CMD_DESC_SKIP
+#define R CMD_DESC_REJECT
+#define W CMD_DESC_REGISTER
+#define B CMD_DESC_BITMASK
+
+/*            Command                          Mask   Fixed Len   Action
+	      ---------------------------------------------------------- */
+static const struct drm_i915_cmd_descriptor gen7_common_cmds[] = {
+	CMD(  MI_NOOP,                          SMI,    F,  1,      S  ),
+	CMD(  MI_USER_INTERRUPT,                SMI,    F,  1,      R  ),
+	CMD(  MI_WAIT_FOR_EVENT,                SMI,    F,  1,      R  ),
+	CMD(  MI_ARB_CHECK,                     SMI,    F,  1,      S  ),
+	CMD(  MI_REPORT_HEAD,                   SMI,    F,  1,      S  ),
+	CMD(  MI_SUSPEND_FLUSH,                 SMI,    F,  1,      S  ),
+	CMD(  MI_SEMAPHORE_MBOX,                SMI,   !F,  0xFF,   R  ),
+	CMD(  MI_STORE_DWORD_INDEX,             SMI,   !F,  0xFF,   R  ),
+	CMD(  MI_LOAD_REGISTER_IMM(1),          SMI,   !F,  0xFF,   W,
+	      .reg = { .offset = 1, .mask = 0x007FFFFC, .step = 2 }    ),
+	CMD(  MI_STORE_REGISTER_MEM,            SMI,    F,  3,     W | B,
+	      .reg = { .offset = 1, .mask = 0x007FFFFC },
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_GLOBAL_GTT,
+			.expected = 0,
+	      }},						       ),
+	CMD(  MI_LOAD_REGISTER_MEM,             SMI,    F,  3,     W | B,
+	      .reg = { .offset = 1, .mask = 0x007FFFFC },
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_GLOBAL_GTT,
+			.expected = 0,
+	      }},						       ),
+	/*
+	 * MI_BATCH_BUFFER_START requires some special handling. It's not
+	 * really a 'skip' action but it doesn't seem like it's worth adding
+	 * a new action. See intel_engine_cmd_parser().
+	 */
+	CMD(  MI_BATCH_BUFFER_START,            SMI,   !F,  0xFF,   S  ),
+};
+
+static const struct drm_i915_cmd_descriptor gen7_render_cmds[] = {
+	CMD(  MI_FLUSH,                         SMI,    F,  1,      S  ),
+	CMD(  MI_ARB_ON_OFF,                    SMI,    F,  1,      R  ),
+	CMD(  MI_PREDICATE,                     SMI,    F,  1,      S  ),
+	CMD(  MI_TOPOLOGY_FILTER,               SMI,    F,  1,      S  ),
+	CMD(  MI_SET_APPID,                     SMI,    F,  1,      S  ),
+	CMD(  MI_DISPLAY_FLIP,                  SMI,   !F,  0xFF,   R  ),
+	CMD(  MI_SET_CONTEXT,                   SMI,   !F,  0xFF,   R  ),
+	CMD(  MI_URB_CLEAR,                     SMI,   !F,  0xFF,   S  ),
+	CMD(  MI_STORE_DWORD_IMM,               SMI,   !F,  0x3F,   B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_GLOBAL_GTT,
+			.expected = 0,
+	      }},						       ),
+	CMD(  MI_UPDATE_GTT,                    SMI,   !F,  0xFF,   R  ),
+	CMD(  MI_CLFLUSH,                       SMI,   !F,  0x3FF,  B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_GLOBAL_GTT,
+			.expected = 0,
+	      }},						       ),
+	CMD(  MI_REPORT_PERF_COUNT,             SMI,   !F,  0x3F,   B,
+	      .bits = {{
+			.offset = 1,
+			.mask = MI_REPORT_PERF_COUNT_GGTT,
+			.expected = 0,
+	      }},						       ),
+	CMD(  MI_CONDITIONAL_BATCH_BUFFER_END,  SMI,   !F,  0xFF,   B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_GLOBAL_GTT,
+			.expected = 0,
+	      }},						       ),
+	CMD(  GFX_OP_3DSTATE_VF_STATISTICS,     S3D,    F,  1,      S  ),
+	CMD(  PIPELINE_SELECT,                  S3D,    F,  1,      S  ),
+	CMD(  MEDIA_VFE_STATE,			S3D,   !F,  0xFFFF, B,
+	      .bits = {{
+			.offset = 2,
+			.mask = MEDIA_VFE_STATE_MMIO_ACCESS_MASK,
+			.expected = 0,
+	      }},						       ),
+	CMD(  GPGPU_OBJECT,                     S3D,   !F,  0xFF,   S  ),
+	CMD(  GPGPU_WALKER,                     S3D,   !F,  0xFF,   S  ),
+	CMD(  GFX_OP_3DSTATE_SO_DECL_LIST,      S3D,   !F,  0x1FF,  S  ),
+	CMD(  GFX_OP_PIPE_CONTROL(5),           S3D,   !F,  0xFF,   B,
+	      .bits = {{
+			.offset = 1,
+			.mask = (PIPE_CONTROL_MMIO_WRITE | PIPE_CONTROL_NOTIFY),
+			.expected = 0,
+	      },
+	      {
+			.offset = 1,
+		        .mask = (PIPE_CONTROL_GLOBAL_GTT_IVB |
+				 PIPE_CONTROL_STORE_DATA_INDEX),
+			.expected = 0,
+			.condition_offset = 1,
+			.condition_mask = PIPE_CONTROL_POST_SYNC_OP_MASK,
+	      }},						       ),
+};
+
+static const struct drm_i915_cmd_descriptor hsw_render_cmds[] = {
+	CMD(  MI_SET_PREDICATE,                 SMI,    F,  1,      S  ),
+	CMD(  MI_RS_CONTROL,                    SMI,    F,  1,      S  ),
+	CMD(  MI_URB_ATOMIC_ALLOC,              SMI,    F,  1,      S  ),
+	CMD(  MI_SET_APPID,                     SMI,    F,  1,      S  ),
+	CMD(  MI_RS_CONTEXT,                    SMI,    F,  1,      S  ),
+	CMD(  MI_LOAD_SCAN_LINES_INCL,          SMI,   !F,  0x3F,   R  ),
+	CMD(  MI_LOAD_SCAN_LINES_EXCL,          SMI,   !F,  0x3F,   R  ),
+	CMD(  MI_LOAD_REGISTER_REG,             SMI,   !F,  0xFF,   W,
+	      .reg = { .offset = 1, .mask = 0x007FFFFC, .step = 1 }    ),
+	CMD(  MI_RS_STORE_DATA_IMM,             SMI,   !F,  0xFF,   S  ),
+	CMD(  MI_LOAD_URB_MEM,                  SMI,   !F,  0xFF,   S  ),
+	CMD(  MI_STORE_URB_MEM,                 SMI,   !F,  0xFF,   S  ),
+	CMD(  GFX_OP_3DSTATE_DX9_CONSTANTF_VS,  S3D,   !F,  0x7FF,  S  ),
+	CMD(  GFX_OP_3DSTATE_DX9_CONSTANTF_PS,  S3D,   !F,  0x7FF,  S  ),
+
+	CMD(  GFX_OP_3DSTATE_BINDING_TABLE_EDIT_VS,  S3D,   !F,  0x1FF,  S  ),
+	CMD(  GFX_OP_3DSTATE_BINDING_TABLE_EDIT_GS,  S3D,   !F,  0x1FF,  S  ),
+	CMD(  GFX_OP_3DSTATE_BINDING_TABLE_EDIT_HS,  S3D,   !F,  0x1FF,  S  ),
+	CMD(  GFX_OP_3DSTATE_BINDING_TABLE_EDIT_DS,  S3D,   !F,  0x1FF,  S  ),
+	CMD(  GFX_OP_3DSTATE_BINDING_TABLE_EDIT_PS,  S3D,   !F,  0x1FF,  S  ),
+};
+
+static const struct drm_i915_cmd_descriptor gen7_video_cmds[] = {
+	CMD(  MI_ARB_ON_OFF,                    SMI,    F,  1,      R  ),
+	CMD(  MI_SET_APPID,                     SMI,    F,  1,      S  ),
+	CMD(  MI_STORE_DWORD_IMM,               SMI,   !F,  0xFF,   B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_GLOBAL_GTT,
+			.expected = 0,
+	      }},						       ),
+	CMD(  MI_UPDATE_GTT,                    SMI,   !F,  0x3F,   R  ),
+	CMD(  MI_FLUSH_DW,                      SMI,   !F,  0x3F,   B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_FLUSH_DW_NOTIFY,
+			.expected = 0,
+	      },
+	      {
+			.offset = 1,
+			.mask = MI_FLUSH_DW_USE_GTT,
+			.expected = 0,
+			.condition_offset = 0,
+			.condition_mask = MI_FLUSH_DW_OP_MASK,
+	      },
+	      {
+			.offset = 0,
+			.mask = MI_FLUSH_DW_STORE_INDEX,
+			.expected = 0,
+			.condition_offset = 0,
+			.condition_mask = MI_FLUSH_DW_OP_MASK,
+	      }},						       ),
+	CMD(  MI_CONDITIONAL_BATCH_BUFFER_END,  SMI,   !F,  0xFF,   B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_GLOBAL_GTT,
+			.expected = 0,
+	      }},						       ),
+	/*
+	 * MFX_WAIT doesn't fit the way we handle length for most commands.
+	 * It has a length field but it uses a non-standard length bias.
+	 * It is always 1 dword though, so just treat it as fixed length.
+	 */
+	CMD(  MFX_WAIT,                         SMFX,   F,  1,      S  ),
+};
+
+static const struct drm_i915_cmd_descriptor gen7_vecs_cmds[] = {
+	CMD(  MI_ARB_ON_OFF,                    SMI,    F,  1,      R  ),
+	CMD(  MI_SET_APPID,                     SMI,    F,  1,      S  ),
+	CMD(  MI_STORE_DWORD_IMM,               SMI,   !F,  0xFF,   B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_GLOBAL_GTT,
+			.expected = 0,
+	      }},						       ),
+	CMD(  MI_UPDATE_GTT,                    SMI,   !F,  0x3F,   R  ),
+	CMD(  MI_FLUSH_DW,                      SMI,   !F,  0x3F,   B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_FLUSH_DW_NOTIFY,
+			.expected = 0,
+	      },
+	      {
+			.offset = 1,
+			.mask = MI_FLUSH_DW_USE_GTT,
+			.expected = 0,
+			.condition_offset = 0,
+			.condition_mask = MI_FLUSH_DW_OP_MASK,
+	      },
+	      {
+			.offset = 0,
+			.mask = MI_FLUSH_DW_STORE_INDEX,
+			.expected = 0,
+			.condition_offset = 0,
+			.condition_mask = MI_FLUSH_DW_OP_MASK,
+	      }},						       ),
+	CMD(  MI_CONDITIONAL_BATCH_BUFFER_END,  SMI,   !F,  0xFF,   B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_GLOBAL_GTT,
+			.expected = 0,
+	      }},						       ),
+};
+
+static const struct drm_i915_cmd_descriptor gen7_blt_cmds[] = {
+	CMD(  MI_DISPLAY_FLIP,                  SMI,   !F,  0xFF,   R  ),
+	CMD(  MI_STORE_DWORD_IMM,               SMI,   !F,  0x3FF,  B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_GLOBAL_GTT,
+			.expected = 0,
+	      }},						       ),
+	CMD(  MI_UPDATE_GTT,                    SMI,   !F,  0x3F,   R  ),
+	CMD(  MI_FLUSH_DW,                      SMI,   !F,  0x3F,   B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_FLUSH_DW_NOTIFY,
+			.expected = 0,
+	      },
+	      {
+			.offset = 1,
+			.mask = MI_FLUSH_DW_USE_GTT,
+			.expected = 0,
+			.condition_offset = 0,
+			.condition_mask = MI_FLUSH_DW_OP_MASK,
+	      },
+	      {
+			.offset = 0,
+			.mask = MI_FLUSH_DW_STORE_INDEX,
+			.expected = 0,
+			.condition_offset = 0,
+			.condition_mask = MI_FLUSH_DW_OP_MASK,
+	      }},						       ),
+	CMD(  COLOR_BLT,                        S2D,   !F,  0x3F,   S  ),
+	CMD(  SRC_COPY_BLT,                     S2D,   !F,  0x3F,   S  ),
+};
+
+static const struct drm_i915_cmd_descriptor hsw_blt_cmds[] = {
+	CMD(  MI_LOAD_SCAN_LINES_INCL,          SMI,   !F,  0x3F,   R  ),
+	CMD(  MI_LOAD_SCAN_LINES_EXCL,          SMI,   !F,  0x3F,   R  ),
+};
+
+/*
+ * For Gen9 we can still rely on the h/w to enforce cmd security, and only
+ * need to re-enforce the register access checks. We therefore only need to
+ * teach the cmdparser how to find the end of each command, and identify
+ * register accesses. The table doesn't need to reject any commands, and so
+ * the only commands listed here are:
+ *   1) Those that touch registers
+ *   2) Those that do not have the default 8-bit length
+ *
+ * Note that the default MI length mask chosen for this table is 0xFF, not
+ * the 0x3F used on older devices. This is because the vast majority of MI
+ * cmds on Gen9 use a standard 8-bit Length field.
+ * All the Gen9 blitter instructions are standard 0xFF length mask, and
+ * none allow access to non-general registers, so in fact no BLT cmds are
+ * included in the table at all.
+ *
+ */
+static const struct drm_i915_cmd_descriptor gen9_blt_cmds[] = {
+	CMD(  MI_NOOP,                          SMI,    F,  1,      S  ),
+	CMD(  MI_USER_INTERRUPT,                SMI,    F,  1,      S  ),
+	CMD(  MI_WAIT_FOR_EVENT,                SMI,    F,  1,      S  ),
+	CMD(  MI_FLUSH,                         SMI,    F,  1,      S  ),
+	CMD(  MI_ARB_CHECK,                     SMI,    F,  1,      S  ),
+	CMD(  MI_REPORT_HEAD,                   SMI,    F,  1,      S  ),
+	CMD(  MI_ARB_ON_OFF,                    SMI,    F,  1,      S  ),
+	CMD(  MI_SUSPEND_FLUSH,                 SMI,    F,  1,      S  ),
+	CMD(  MI_LOAD_SCAN_LINES_INCL,          SMI,   !F,  0x3F,   S  ),
+	CMD(  MI_LOAD_SCAN_LINES_EXCL,          SMI,   !F,  0x3F,   S  ),
+	CMD(  MI_STORE_DWORD_IMM,               SMI,   !F,  0x3FF,  S  ),
+	CMD(  MI_LOAD_REGISTER_IMM(1),          SMI,   !F,  0xFF,   W,
+	      .reg = { .offset = 1, .mask = 0x007FFFFC, .step = 2 }    ),
+	CMD(  MI_UPDATE_GTT,                    SMI,   !F,  0x3FF,  S  ),
+	CMD(  MI_STORE_REGISTER_MEM_GEN8,       SMI,    F,  4,      W,
+	      .reg = { .offset = 1, .mask = 0x007FFFFC }               ),
+	CMD(  MI_FLUSH_DW,                      SMI,   !F,  0x3F,   S  ),
+	CMD(  MI_LOAD_REGISTER_MEM_GEN8,        SMI,    F,  4,      W,
+	      .reg = { .offset = 1, .mask = 0x007FFFFC }               ),
+	CMD(  MI_LOAD_REGISTER_REG,             SMI,    !F,  0xFF,  W,
+	      .reg = { .offset = 1, .mask = 0x007FFFFC, .step = 1 }    ),
+
+	/*
+	 * We allow BB_START but apply further checks. We just sanitize the
+	 * basic fields here.
+	 */
+#define MI_BB_START_OPERAND_MASK   GENMASK(SMI-1, 0)
+#define MI_BB_START_OPERAND_EXPECT (MI_BATCH_PPGTT_HSW | 1)
+	CMD(  MI_BATCH_BUFFER_START_GEN8,       SMI,    !F,  0xFF,  B,
+	      .bits = {{
+			.offset = 0,
+			.mask = MI_BB_START_OPERAND_MASK,
+			.expected = MI_BB_START_OPERAND_EXPECT,
+	      }},						       ),
+};
+
+static const struct drm_i915_cmd_descriptor noop_desc =
+	CMD(MI_NOOP, SMI, F, 1, S);
+
+#undef CMD
+#undef SMI
+#undef S3D
+#undef S2D
+#undef SMFX
+#undef F
+#undef S
+#undef R
+#undef W
+#undef B
+
+static const struct drm_i915_cmd_table gen7_render_cmd_table[] = {
+	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
+	{ gen7_render_cmds, ARRAY_SIZE(gen7_render_cmds) },
+};
+
+static const struct drm_i915_cmd_table hsw_render_ring_cmd_table[] = {
+	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
+	{ gen7_render_cmds, ARRAY_SIZE(gen7_render_cmds) },
+	{ hsw_render_cmds, ARRAY_SIZE(hsw_render_cmds) },
+};
+
+static const struct drm_i915_cmd_table gen7_video_cmd_table[] = {
+	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
+	{ gen7_video_cmds, ARRAY_SIZE(gen7_video_cmds) },
+};
+
+static const struct drm_i915_cmd_table hsw_vebox_cmd_table[] = {
+	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
+	{ gen7_vecs_cmds, ARRAY_SIZE(gen7_vecs_cmds) },
+};
+
+static const struct drm_i915_cmd_table gen7_blt_cmd_table[] = {
+	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
+	{ gen7_blt_cmds, ARRAY_SIZE(gen7_blt_cmds) },
+};
+
+static const struct drm_i915_cmd_table hsw_blt_ring_cmd_table[] = {
+	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
+	{ gen7_blt_cmds, ARRAY_SIZE(gen7_blt_cmds) },
+	{ hsw_blt_cmds, ARRAY_SIZE(hsw_blt_cmds) },
+};
+
+static const struct drm_i915_cmd_table gen9_blt_cmd_table[] = {
+	{ gen9_blt_cmds, ARRAY_SIZE(gen9_blt_cmds) },
+};
+
+
+/*
+ * Register whitelists, sorted by increasing register offset.
+ */
+
+/*
+ * An individual whitelist entry granting access to register addr.  If
+ * mask is non-zero the argument of immediate register writes will be
+ * AND-ed with mask, and the command will be rejected if the result
+ * doesn't match value.
+ *
+ * Registers with non-zero mask are only allowed to be written using
+ * LRI.
+ */
+struct drm_i915_reg_descriptor {
+	i915_reg_t addr;
+	u32 mask;
+	u32 value;
+};
+
+/* Convenience macro for adding 32-bit registers. */
+#define REG32(_reg, ...) \
+	{ .addr = (_reg), __VA_ARGS__ }
+
+#define REG32_IDX(_reg, idx) \
+	{ .addr = _reg(idx) }
+
+/*
+ * Convenience macro for adding 64-bit registers.
+ *
+ * Some registers that userspace accesses are 64 bits. The register
+ * access commands only allow 32-bit accesses. Hence, we have to include
+ * entries for both halves of the 64-bit registers.
+ */
+#define REG64(_reg) \
+	{ .addr = _reg }, \
+	{ .addr = _reg ## _UDW }
+
+#define REG64_IDX(_reg, idx) \
+	{ .addr = _reg(idx) }, \
+	{ .addr = _reg ## _UDW(idx) }
+
+#define REG64_BASE_IDX(_reg, base, idx) \
+	{ .addr = _reg(base, idx) }, \
+	{ .addr = _reg ## _UDW(base, idx) }
+
+static const struct drm_i915_reg_descriptor gen7_render_regs[] = {
+	REG64(GPGPU_THREADS_DISPATCHED),
+	REG64(HS_INVOCATION_COUNT),
+	REG64(DS_INVOCATION_COUNT),
+	REG64(IA_VERTICES_COUNT),
+	REG64(IA_PRIMITIVES_COUNT),
+	REG64(VS_INVOCATION_COUNT),
+	REG64(GS_INVOCATION_COUNT),
+	REG64(GS_PRIMITIVES_COUNT),
+	REG64(CL_INVOCATION_COUNT),
+	REG64(CL_PRIMITIVES_COUNT),
+	REG64(PS_INVOCATION_COUNT),
+	REG64(PS_DEPTH_COUNT),
+	REG64_IDX(RING_TIMESTAMP, RENDER_RING_BASE),
+	REG64_IDX(MI_PREDICATE_SRC0, RENDER_RING_BASE),
+	REG64_IDX(MI_PREDICATE_SRC1, RENDER_RING_BASE),
+	REG32(GEN7_3DPRIM_END_OFFSET),
+	REG32(GEN7_3DPRIM_START_VERTEX),
+	REG32(GEN7_3DPRIM_VERTEX_COUNT),
+	REG32(GEN7_3DPRIM_INSTANCE_COUNT),
+	REG32(GEN7_3DPRIM_START_INSTANCE),
+	REG32(GEN7_3DPRIM_BASE_VERTEX),
+	REG32(GEN7_GPGPU_DISPATCHDIMX),
+	REG32(GEN7_GPGPU_DISPATCHDIMY),
+	REG32(GEN7_GPGPU_DISPATCHDIMZ),
+	REG64_IDX(RING_TIMESTAMP, BSD_RING_BASE),
+	REG64_IDX(GEN7_SO_NUM_PRIMS_WRITTEN, 0),
+	REG64_IDX(GEN7_SO_NUM_PRIMS_WRITTEN, 1),
+	REG64_IDX(GEN7_SO_NUM_PRIMS_WRITTEN, 2),
+	REG64_IDX(GEN7_SO_NUM_PRIMS_WRITTEN, 3),
+	REG64_IDX(GEN7_SO_PRIM_STORAGE_NEEDED, 0),
+	REG64_IDX(GEN7_SO_PRIM_STORAGE_NEEDED, 1),
+	REG64_IDX(GEN7_SO_PRIM_STORAGE_NEEDED, 2),
+	REG64_IDX(GEN7_SO_PRIM_STORAGE_NEEDED, 3),
+	REG32(GEN7_SO_WRITE_OFFSET(0)),
+	REG32(GEN7_SO_WRITE_OFFSET(1)),
+	REG32(GEN7_SO_WRITE_OFFSET(2)),
+	REG32(GEN7_SO_WRITE_OFFSET(3)),
+	REG32(GEN7_L3SQCREG1),
+	REG32(GEN7_L3CNTLREG2),
+	REG32(GEN7_L3CNTLREG3),
+	REG64_IDX(RING_TIMESTAMP, BLT_RING_BASE),
+};
+
+static const struct drm_i915_reg_descriptor hsw_render_regs[] = {
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 0),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 1),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 2),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 3),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 4),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 5),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 6),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 7),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 8),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 9),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 10),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 11),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 12),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 13),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 14),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, RENDER_RING_BASE, 15),
+	REG32(HSW_SCRATCH1,
+	      .mask = ~HSW_SCRATCH1_L3_DATA_ATOMICS_DISABLE,
+	      .value = 0),
+	REG32(HSW_ROW_CHICKEN3,
+	      .mask = ~(HSW_ROW_CHICKEN3_L3_GLOBAL_ATOMICS_DISABLE << 16 |
+                        HSW_ROW_CHICKEN3_L3_GLOBAL_ATOMICS_DISABLE),
+	      .value = 0),
+};
+
+static const struct drm_i915_reg_descriptor gen7_blt_regs[] = {
+	REG64_IDX(RING_TIMESTAMP, RENDER_RING_BASE),
+	REG64_IDX(RING_TIMESTAMP, BSD_RING_BASE),
+	REG32(BCS_SWCTRL),
+	REG64_IDX(RING_TIMESTAMP, BLT_RING_BASE),
+};
+
+static const struct drm_i915_reg_descriptor gen9_blt_regs[] = {
+	REG64_IDX(RING_TIMESTAMP, RENDER_RING_BASE),
+	REG64_IDX(RING_TIMESTAMP, BSD_RING_BASE),
+	REG32(BCS_SWCTRL),
+	REG64_IDX(RING_TIMESTAMP, BLT_RING_BASE),
+	REG32_IDX(RING_CTX_TIMESTAMP, BLT_RING_BASE),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 0),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 1),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 2),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 3),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 4),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 5),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 6),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 7),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 8),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 9),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 10),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 11),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 12),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 13),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 14),
+	REG64_BASE_IDX(GEN8_RING_CS_GPR, BLT_RING_BASE, 15),
+};
+
+#undef REG64
+#undef REG32
+
+struct drm_i915_reg_table {
+	const struct drm_i915_reg_descriptor *regs;
+	int num_regs;
+};
+
+static const struct drm_i915_reg_table ivb_render_reg_tables[] = {
+	{ gen7_render_regs, ARRAY_SIZE(gen7_render_regs) },
+};
+
+static const struct drm_i915_reg_table ivb_blt_reg_tables[] = {
+	{ gen7_blt_regs, ARRAY_SIZE(gen7_blt_regs) },
+};
+
+static const struct drm_i915_reg_table hsw_render_reg_tables[] = {
+	{ gen7_render_regs, ARRAY_SIZE(gen7_render_regs) },
+	{ hsw_render_regs, ARRAY_SIZE(hsw_render_regs) },
+};
+
+static const struct drm_i915_reg_table hsw_blt_reg_tables[] = {
+	{ gen7_blt_regs, ARRAY_SIZE(gen7_blt_regs) },
+};
+
+static const struct drm_i915_reg_table gen9_blt_reg_tables[] = {
+	{ gen9_blt_regs, ARRAY_SIZE(gen9_blt_regs) },
+};
+
+static u32 gen7_render_get_cmd_length_mask(u32 cmd_header)
+{
+	u32 client = cmd_header >> INSTR_CLIENT_SHIFT;
+	u32 subclient =
+		(cmd_header & INSTR_SUBCLIENT_MASK) >> INSTR_SUBCLIENT_SHIFT;
+
+	if (client == INSTR_MI_CLIENT)
+		return 0x3F;
+	else if (client == INSTR_RC_CLIENT) {
+		if (subclient == INSTR_MEDIA_SUBCLIENT)
+			return 0xFFFF;
+		else
+			return 0xFF;
+	}
+
+	DRM_DEBUG("CMD: Abnormal rcs cmd length! 0x%08X\n", cmd_header);
+	return 0;
+}
+
+static u32 gen7_bsd_get_cmd_length_mask(u32 cmd_header)
+{
+	u32 client = cmd_header >> INSTR_CLIENT_SHIFT;
+	u32 subclient =
+		(cmd_header & INSTR_SUBCLIENT_MASK) >> INSTR_SUBCLIENT_SHIFT;
+	u32 op = (cmd_header & INSTR_26_TO_24_MASK) >> INSTR_26_TO_24_SHIFT;
+
+	if (client == INSTR_MI_CLIENT)
+		return 0x3F;
+	else if (client == INSTR_RC_CLIENT) {
+		if (subclient == INSTR_MEDIA_SUBCLIENT) {
+			if (op == 6)
+				return 0xFFFF;
+			else
+				return 0xFFF;
+		} else
+			return 0xFF;
+	}
+
+	DRM_DEBUG("CMD: Abnormal bsd cmd length! 0x%08X\n", cmd_header);
+	return 0;
+}
+
+static u32 gen7_blt_get_cmd_length_mask(u32 cmd_header)
+{
+	u32 client = cmd_header >> INSTR_CLIENT_SHIFT;
+
+	if (client == INSTR_MI_CLIENT)
+		return 0x3F;
+	else if (client == INSTR_BC_CLIENT)
+		return 0xFF;
+
+	DRM_DEBUG("CMD: Abnormal blt cmd length! 0x%08X\n", cmd_header);
+	return 0;
+}
+
+static u32 gen9_blt_get_cmd_length_mask(u32 cmd_header)
+{
+	u32 client = cmd_header >> INSTR_CLIENT_SHIFT;
+
+	if (client == INSTR_MI_CLIENT || client == INSTR_BC_CLIENT)
+		return 0xFF;
+
+	DRM_DEBUG("CMD: Abnormal blt cmd length! 0x%08X\n", cmd_header);
+	return 0;
+}
+
+static bool validate_cmds_sorted(const struct intel_engine_cs *engine,
+				 const struct drm_i915_cmd_table *cmd_tables,
+				 int cmd_table_count)
+{
+	int i;
+	bool ret = true;
+
+	if (!cmd_tables || cmd_table_count == 0)
+		return true;
+
+	for (i = 0; i < cmd_table_count; i++) {
+		const struct drm_i915_cmd_table *table = &cmd_tables[i];
+		u32 previous = 0;
+		int j;
+
+		for (j = 0; j < table->count; j++) {
+			const struct drm_i915_cmd_descriptor *desc =
+				&table->table[j];
+			u32 curr = desc->cmd.value & desc->cmd.mask;
+
+			if (curr < previous) {
+				drm_err(&engine->i915->drm,
+					"CMD: %s [%d] command table not sorted: "
+					"table=%d entry=%d cmd=0x%08X prev=0x%08X\n",
+					engine->name, engine->id,
+					i, j, curr, previous);
+				ret = false;
+			}
+
+			previous = curr;
+		}
+	}
+
+	return ret;
+}
+
+static bool check_sorted(const struct intel_engine_cs *engine,
+			 const struct drm_i915_reg_descriptor *reg_table,
+			 int reg_count)
+{
+	int i;
+	u32 previous = 0;
+	bool ret = true;
+
+	for (i = 0; i < reg_count; i++) {
+		u32 curr = i915_mmio_reg_offset(reg_table[i].addr);
+
+		if (curr < previous) {
+			drm_err(&engine->i915->drm,
+				"CMD: %s [%d] register table not sorted: "
+				"entry=%d reg=0x%08X prev=0x%08X\n",
+				engine->name, engine->id,
+				i, curr, previous);
+			ret = false;
+		}
+
+		previous = curr;
+	}
+
+	return ret;
+}
+
+static bool validate_regs_sorted(struct intel_engine_cs *engine)
+{
+	int i;
+	const struct drm_i915_reg_table *table;
+
+	for (i = 0; i < engine->reg_table_count; i++) {
+		table = &engine->reg_tables[i];
+		if (!check_sorted(engine, table->regs, table->num_regs))
+			return false;
+	}
+
+	return true;
+}
+
+struct cmd_node {
+	const struct drm_i915_cmd_descriptor *desc;
+	struct hlist_node node;
+};
+
+/*
+ * Different command ranges have different numbers of bits for the opcode. For
+ * example, MI commands use bits 31:23 while 3D commands use bits 31:16. The
+ * problem is that, for example, MI commands use bits 22:16 for other fields
+ * such as GGTT vs PPGTT bits. If we include those bits in the mask then when
+ * we mask a command from a batch it could hash to the wrong bucket due to
+ * non-opcode bits being set. But if we don't include those bits, some 3D
+ * commands may hash to the same bucket due to not including opcode bits that
+ * make the command unique. For now, we will risk hashing to the same bucket.
+ */
+static inline u32 cmd_header_key(u32 x)
+{
+	switch (x >> INSTR_CLIENT_SHIFT) {
+	default:
+	case INSTR_MI_CLIENT:
+		return x >> STD_MI_OPCODE_SHIFT;
+	case INSTR_RC_CLIENT:
+		return x >> STD_3D_OPCODE_SHIFT;
+	case INSTR_BC_CLIENT:
+		return x >> STD_2D_OPCODE_SHIFT;
+	}
+}
+
+static int init_hash_table(struct intel_engine_cs *engine,
+			   const struct drm_i915_cmd_table *cmd_tables,
+			   int cmd_table_count)
+{
+	int i, j;
+
+	hash_init(engine->cmd_hash);
+
+	for (i = 0; i < cmd_table_count; i++) {
+		const struct drm_i915_cmd_table *table = &cmd_tables[i];
+
+		for (j = 0; j < table->count; j++) {
+			const struct drm_i915_cmd_descriptor *desc =
+				&table->table[j];
+			struct cmd_node *desc_node =
+				kmalloc(sizeof(*desc_node), GFP_KERNEL);
+
+			if (!desc_node)
+				return -ENOMEM;
+
+			desc_node->desc = desc;
+			hash_add(engine->cmd_hash, &desc_node->node,
+				 cmd_header_key(desc->cmd.value));
+		}
+	}
+
+	return 0;
+}
+
+static void fini_hash_table(struct intel_engine_cs *engine)
+{
+	struct hlist_node *tmp;
+	struct cmd_node *desc_node;
+	int i;
+
+	hash_for_each_safe(engine->cmd_hash, i, tmp, desc_node, node) {
+		hash_del(&desc_node->node);
+		kfree(desc_node);
+	}
+}
+
+/**
+ * intel_engine_init_cmd_parser() - set cmd parser related fields for an engine
+ * @engine: the engine to initialize
+ *
+ * Optionally initializes fields related to batch buffer command parsing in the
+ * struct intel_engine_cs based on whether the platform requires software
+ * command parsing.
+ */
+int intel_engine_init_cmd_parser(struct intel_engine_cs *engine)
+{
+	const struct drm_i915_cmd_table *cmd_tables;
+	int cmd_table_count;
+	int ret;
+
+	if (GRAPHICS_VER(engine->i915) != 7 && !(GRAPHICS_VER(engine->i915) == 9 &&
+						 engine->class == COPY_ENGINE_CLASS))
+		return 0;
+
+	switch (engine->class) {
+	case RENDER_CLASS:
+		if (IS_HASWELL(engine->i915)) {
+			cmd_tables = hsw_render_ring_cmd_table;
+			cmd_table_count =
+				ARRAY_SIZE(hsw_render_ring_cmd_table);
+		} else {
+			cmd_tables = gen7_render_cmd_table;
+			cmd_table_count = ARRAY_SIZE(gen7_render_cmd_table);
+		}
+
+		if (IS_HASWELL(engine->i915)) {
+			engine->reg_tables = hsw_render_reg_tables;
+			engine->reg_table_count = ARRAY_SIZE(hsw_render_reg_tables);
+		} else {
+			engine->reg_tables = ivb_render_reg_tables;
+			engine->reg_table_count = ARRAY_SIZE(ivb_render_reg_tables);
+		}
+		engine->get_cmd_length_mask = gen7_render_get_cmd_length_mask;
+		break;
+	case VIDEO_DECODE_CLASS:
+		cmd_tables = gen7_video_cmd_table;
+		cmd_table_count = ARRAY_SIZE(gen7_video_cmd_table);
+		engine->get_cmd_length_mask = gen7_bsd_get_cmd_length_mask;
+		break;
+	case COPY_ENGINE_CLASS:
+		engine->get_cmd_length_mask = gen7_blt_get_cmd_length_mask;
+		if (GRAPHICS_VER(engine->i915) == 9) {
+			cmd_tables = gen9_blt_cmd_table;
+			cmd_table_count = ARRAY_SIZE(gen9_blt_cmd_table);
+			engine->get_cmd_length_mask =
+				gen9_blt_get_cmd_length_mask;
+
+			/* BCS Engine unsafe without parser */
+			engine->flags |= I915_ENGINE_REQUIRES_CMD_PARSER;
+		} else if (IS_HASWELL(engine->i915)) {
+			cmd_tables = hsw_blt_ring_cmd_table;
+			cmd_table_count = ARRAY_SIZE(hsw_blt_ring_cmd_table);
+		} else {
+			cmd_tables = gen7_blt_cmd_table;
+			cmd_table_count = ARRAY_SIZE(gen7_blt_cmd_table);
+		}
+
+		if (GRAPHICS_VER(engine->i915) == 9) {
+			engine->reg_tables = gen9_blt_reg_tables;
+			engine->reg_table_count =
+				ARRAY_SIZE(gen9_blt_reg_tables);
+		} else if (IS_HASWELL(engine->i915)) {
+			engine->reg_tables = hsw_blt_reg_tables;
+			engine->reg_table_count = ARRAY_SIZE(hsw_blt_reg_tables);
+		} else {
+			engine->reg_tables = ivb_blt_reg_tables;
+			engine->reg_table_count = ARRAY_SIZE(ivb_blt_reg_tables);
+		}
+		break;
+	case VIDEO_ENHANCEMENT_CLASS:
+		cmd_tables = hsw_vebox_cmd_table;
+		cmd_table_count = ARRAY_SIZE(hsw_vebox_cmd_table);
+		/* VECS can use the same length_mask function as VCS */
+		engine->get_cmd_length_mask = gen7_bsd_get_cmd_length_mask;
+		break;
+	default:
+		MISSING_CASE(engine->class);
+		goto out;
+	}
+
+	if (!validate_cmds_sorted(engine, cmd_tables, cmd_table_count)) {
+		drm_err(&engine->i915->drm,
+			"%s: command descriptions are not sorted\n",
+			engine->name);
+		goto out;
+	}
+	if (!validate_regs_sorted(engine)) {
+		drm_err(&engine->i915->drm,
+			"%s: registers are not sorted\n", engine->name);
+		goto out;
+	}
+
+	ret = init_hash_table(engine, cmd_tables, cmd_table_count);
+	if (ret) {
+		drm_err(&engine->i915->drm,
+			"%s: initialised failed!\n", engine->name);
+		fini_hash_table(engine);
+		goto out;
+	}
+
+	engine->flags |= I915_ENGINE_USING_CMD_PARSER;
+
+out:
+	if (intel_engine_requires_cmd_parser(engine) &&
+	    !intel_engine_using_cmd_parser(engine))
+		return -EINVAL;
+
+	return 0;
+}
+
+/**
+ * intel_engine_cleanup_cmd_parser() - clean up cmd parser related fields
+ * @engine: the engine to clean up
+ *
+ * Releases any resources related to command parsing that may have been
+ * initialized for the specified engine.
+ */
+void intel_engine_cleanup_cmd_parser(struct intel_engine_cs *engine)
+{
+	if (!intel_engine_using_cmd_parser(engine))
+		return;
+
+	fini_hash_table(engine);
+}
+
+static const struct drm_i915_cmd_descriptor*
+find_cmd_in_table(struct intel_engine_cs *engine,
+		  u32 cmd_header)
+{
+	struct cmd_node *desc_node;
+
+	hash_for_each_possible(engine->cmd_hash, desc_node, node,
+			       cmd_header_key(cmd_header)) {
+		const struct drm_i915_cmd_descriptor *desc = desc_node->desc;
+		if (((cmd_header ^ desc->cmd.value) & desc->cmd.mask) == 0)
+			return desc;
+	}
+
+	return NULL;
+}
+
+/*
+ * Returns a pointer to a descriptor for the command specified by cmd_header.
+ *
+ * The caller must supply space for a default descriptor via the default_desc
+ * parameter. If no descriptor for the specified command exists in the engine's
+ * command parser tables, this function fills in default_desc based on the
+ * engine's default length encoding and returns default_desc.
+ */
+static const struct drm_i915_cmd_descriptor*
+find_cmd(struct intel_engine_cs *engine,
+	 u32 cmd_header,
+	 const struct drm_i915_cmd_descriptor *desc,
+	 struct drm_i915_cmd_descriptor *default_desc)
+{
+	u32 mask;
+
+	if (((cmd_header ^ desc->cmd.value) & desc->cmd.mask) == 0)
+		return desc;
+
+	desc = find_cmd_in_table(engine, cmd_header);
+	if (desc)
+		return desc;
+
+	mask = engine->get_cmd_length_mask(cmd_header);
+	if (!mask)
+		return NULL;
+
+	default_desc->cmd.value = cmd_header;
+	default_desc->cmd.mask = ~0u << MIN_OPCODE_SHIFT;
+	default_desc->length.mask = mask;
+	default_desc->flags = CMD_DESC_SKIP;
+	return default_desc;
+}
+
+static const struct drm_i915_reg_descriptor *
+__find_reg(const struct drm_i915_reg_descriptor *table, int count, u32 addr)
+{
+	int start = 0, end = count;
+	while (start < end) {
+		int mid = start + (end - start) / 2;
+		int ret = addr - i915_mmio_reg_offset(table[mid].addr);
+		if (ret < 0)
+			end = mid;
+		else if (ret > 0)
+			start = mid + 1;
+		else
+			return &table[mid];
+	}
+	return NULL;
+}
+
+static const struct drm_i915_reg_descriptor *
+find_reg(const struct intel_engine_cs *engine, u32 addr)
+{
+	const struct drm_i915_reg_table *table = engine->reg_tables;
+	const struct drm_i915_reg_descriptor *reg = NULL;
+	int count = engine->reg_table_count;
+
+	for (; !reg && (count > 0); ++table, --count)
+		reg = __find_reg(table->regs, table->num_regs, addr);
+
+	return reg;
+}
+
+/* Returns a vmap'd pointer to dst_obj, which the caller must unmap */
+static u32 *copy_batch(struct drm_i915_gem_object *dst_obj,
+		       struct drm_i915_gem_object *src_obj,
+		       unsigned long offset, unsigned long length,
+		       bool *needs_clflush_after)
+{
+	unsigned int src_needs_clflush;
+	unsigned int dst_needs_clflush;
+	void *dst, *src;
+	int ret;
+
+	ret = i915_gem_object_prepare_write(dst_obj, &dst_needs_clflush);
+	if (ret)
+		return ERR_PTR(ret);
+
+	dst = i915_gem_object_pin_map(dst_obj, I915_MAP_WB);
+	i915_gem_object_finish_access(dst_obj);
+	if (IS_ERR(dst))
+		return dst;
+
+	ret = i915_gem_object_prepare_read(src_obj, &src_needs_clflush);
+	if (ret) {
+		i915_gem_object_unpin_map(dst_obj);
+		return ERR_PTR(ret);
+	}
+
+	src = ERR_PTR(-ENODEV);
+	if (src_needs_clflush && i915_has_memcpy_from_wc()) {
+		src = i915_gem_object_pin_map(src_obj, I915_MAP_WC);
+		if (!IS_ERR(src)) {
+			i915_unaligned_memcpy_from_wc(dst,
+						      src + offset,
+						      length);
+			i915_gem_object_unpin_map(src_obj);
+		}
+	}
+	if (IS_ERR(src)) {
+		unsigned long x, n, remain;
+		void *ptr;
+
+		/*
+		 * We can avoid clflushing partial cachelines before the write
+		 * if we only every write full cache-lines. Since we know that
+		 * both the source and destination are in multiples of
+		 * PAGE_SIZE, we can simply round up to the next cacheline.
+		 * We don't care about copying too much here as we only
+		 * validate up to the end of the batch.
+		 */
+		remain = length;
+		if (dst_needs_clflush & CLFLUSH_BEFORE)
+			remain = round_up(remain,
+					  boot_cpu_data.x86_clflush_size);
+
+		ptr = dst;
+		x = offset_in_page(offset);
+		for (n = offset >> PAGE_SHIFT; remain; n++) {
+			int len = min(remain, PAGE_SIZE - x);
+
+			src = kmap_atomic(i915_gem_object_get_page(src_obj, n));
+			if (src_needs_clflush)
+				drm_clflush_virt_range(src + x, len);
+			memcpy(ptr, src + x, len);
+			kunmap_atomic(src);
+
+			ptr += len;
+			remain -= len;
+			x = 0;
+		}
+	}
+
+	i915_gem_object_finish_access(src_obj);
+
+	memset32(dst + length, 0, (dst_obj->base.size - length) / sizeof(u32));
+
+	/* dst_obj is returned with vmap pinned */
+	*needs_clflush_after = dst_needs_clflush & CLFLUSH_AFTER;
+
+	return dst;
+}
+
+static inline bool cmd_desc_is(const struct drm_i915_cmd_descriptor * const desc,
+			       const u32 cmd)
+{
+	return desc->cmd.value == (cmd & desc->cmd.mask);
+}
+
+static bool check_cmd(const struct intel_engine_cs *engine,
+		      const struct drm_i915_cmd_descriptor *desc,
+		      const u32 *cmd, u32 length)
+{
+	if (desc->flags & CMD_DESC_SKIP)
+		return true;
+
+	if (desc->flags & CMD_DESC_REJECT) {
+		DRM_DEBUG("CMD: Rejected command: 0x%08X\n", *cmd);
+		return false;
+	}
+
+	if (desc->flags & CMD_DESC_REGISTER) {
+		/*
+		 * Get the distance between individual register offset
+		 * fields if the command can perform more than one
+		 * access at a time.
+		 */
+		const u32 step = desc->reg.step ? desc->reg.step : length;
+		u32 offset;
+
+		for (offset = desc->reg.offset; offset < length;
+		     offset += step) {
+			const u32 reg_addr = cmd[offset] & desc->reg.mask;
+			const struct drm_i915_reg_descriptor *reg =
+				find_reg(engine, reg_addr);
+
+			if (!reg) {
+				DRM_DEBUG("CMD: Rejected register 0x%08X in command: 0x%08X (%s)\n",
+					  reg_addr, *cmd, engine->name);
+				return false;
+			}
+
+			/*
+			 * Check the value written to the register against the
+			 * allowed mask/value pair given in the whitelist entry.
+			 */
+			if (reg->mask) {
+				if (cmd_desc_is(desc, MI_LOAD_REGISTER_MEM)) {
+					DRM_DEBUG("CMD: Rejected LRM to masked register 0x%08X\n",
+						  reg_addr);
+					return false;
+				}
+
+				if (cmd_desc_is(desc, MI_LOAD_REGISTER_REG)) {
+					DRM_DEBUG("CMD: Rejected LRR to masked register 0x%08X\n",
+						  reg_addr);
+					return false;
+				}
+
+				if (cmd_desc_is(desc, MI_LOAD_REGISTER_IMM(1)) &&
+				    (offset + 2 > length ||
+				     (cmd[offset + 1] & reg->mask) != reg->value)) {
+					DRM_DEBUG("CMD: Rejected LRI to masked register 0x%08X\n",
+						  reg_addr);
+					return false;
+				}
+			}
+		}
+	}
+
+	if (desc->flags & CMD_DESC_BITMASK) {
+		int i;
+
+		for (i = 0; i < MAX_CMD_DESC_BITMASKS; i++) {
+			u32 dword;
+
+			if (desc->bits[i].mask == 0)
+				break;
+
+			if (desc->bits[i].condition_mask != 0) {
+				u32 offset =
+					desc->bits[i].condition_offset;
+				u32 condition = cmd[offset] &
+					desc->bits[i].condition_mask;
+
+				if (condition == 0)
+					continue;
+			}
+
+			if (desc->bits[i].offset >= length) {
+				DRM_DEBUG("CMD: Rejected command 0x%08X, too short to check bitmask (%s)\n",
+					  *cmd, engine->name);
+				return false;
+			}
+
+			dword = cmd[desc->bits[i].offset] &
+				desc->bits[i].mask;
+
+			if (dword != desc->bits[i].expected) {
+				DRM_DEBUG("CMD: Rejected command 0x%08X for bitmask 0x%08X (exp=0x%08X act=0x%08X) (%s)\n",
+					  *cmd,
+					  desc->bits[i].mask,
+					  desc->bits[i].expected,
+					  dword, engine->name);
+				return false;
+			}
+		}
+	}
+
+	return true;
+}
+
+static int check_bbstart(u32 *cmd, u32 offset, u32 length,
+			 u32 batch_length,
+			 u64 batch_addr,
+			 u64 shadow_addr,
+			 const unsigned long *jump_whitelist)
+{
+	u64 jump_offset, jump_target;
+	u32 target_cmd_offset, target_cmd_index;
+
+	/* For igt compatibility on older platforms */
+	if (!jump_whitelist) {
+		DRM_DEBUG("CMD: Rejecting BB_START for ggtt based submission\n");
+		return -EACCES;
+	}
+
+	if (length != 3) {
+		DRM_DEBUG("CMD: Recursive BB_START with bad length(%u)\n",
+			  length);
+		return -EINVAL;
+	}
+
+	jump_target = *(u64 *)(cmd + 1);
+	jump_offset = jump_target - batch_addr;
+
+	/*
+	 * Any underflow of jump_target is guaranteed to be outside the range
+	 * of a u32, so >= test catches both too large and too small
+	 */
+	if (jump_offset >= batch_length) {
+		DRM_DEBUG("CMD: BB_START to 0x%llx jumps out of BB\n",
+			  jump_target);
+		return -EINVAL;
+	}
+
+	/*
+	 * This cannot overflow a u32 because we already checked jump_offset
+	 * is within the BB, and the batch_length is a u32
+	 */
+	target_cmd_offset = lower_32_bits(jump_offset);
+	target_cmd_index = target_cmd_offset / sizeof(u32);
+
+	*(u64 *)(cmd + 1) = shadow_addr + target_cmd_offset;
+
+	if (target_cmd_index == offset)
+		return 0;
+
+	if (IS_ERR(jump_whitelist))
+		return PTR_ERR(jump_whitelist);
+
+	if (!test_bit(target_cmd_index, jump_whitelist)) {
+		DRM_DEBUG("CMD: BB_START to 0x%llx not a previously executed cmd\n",
+			  jump_target);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static unsigned long *alloc_whitelist(u32 batch_length)
+{
+	unsigned long *jmp;
+
+	/*
+	 * We expect batch_length to be less than 256KiB for known users,
+	 * i.e. we need at most an 8KiB bitmap allocation which should be
+	 * reasonably cheap due to kmalloc caches.
+	 */
+
+	/* Prefer to report transient allocation failure rather than hit oom */
+	jmp = bitmap_zalloc(DIV_ROUND_UP(batch_length, sizeof(u32)),
+			    GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN);
+	if (!jmp)
+		return ERR_PTR(-ENOMEM);
+
+	return jmp;
+}
+
+#define LENGTH_BIAS 2
+
+/**
+ * intel_engine_cmd_parser() - parse a batch buffer for privilege violations
+ * @engine: the engine on which the batch is to execute
+ * @batch: the batch buffer in question
+ * @batch_offset: byte offset in the batch at which execution starts
+ * @batch_length: length of the commands in batch_obj
+ * @shadow: validated copy of the batch buffer in question
+ * @trampoline: true if we need to trampoline into privileged execution
+ *
+ * Parses the specified batch buffer looking for privilege violations as
+ * described in the overview.
+ *
+ * Return: non-zero if the parser finds violations or otherwise fails; -EACCES
+ * if the batch appears legal but should use hardware parsing
+ */
+
+int intel_engine_cmd_parser(struct intel_engine_cs *engine,
+			    struct i915_vma *batch,
+			    unsigned long batch_offset,
+			    unsigned long batch_length,
+			    struct i915_vma *shadow,
+			    bool trampoline)
+{
+	u32 *cmd, *batch_end, offset = 0;
+	struct drm_i915_cmd_descriptor default_desc = noop_desc;
+	const struct drm_i915_cmd_descriptor *desc = &default_desc;
+	bool needs_clflush_after = false;
+	unsigned long *jump_whitelist;
+	u64 batch_addr, shadow_addr;
+	int ret = 0;
+
+	GEM_BUG_ON(!IS_ALIGNED(batch_offset, sizeof(*cmd)));
+	GEM_BUG_ON(!IS_ALIGNED(batch_length, sizeof(*cmd)));
+	GEM_BUG_ON(range_overflows_t(u64, batch_offset, batch_length,
+				     batch->size));
+	GEM_BUG_ON(!batch_length);
+
+	cmd = copy_batch(shadow->obj, batch->obj,
+			 batch_offset, batch_length,
+			 &needs_clflush_after);
+	if (IS_ERR(cmd)) {
+		DRM_DEBUG("CMD: Failed to copy batch\n");
+		return PTR_ERR(cmd);
+	}
+
+	jump_whitelist = NULL;
+	if (!trampoline)
+		/* Defer failure until attempted use */
+		jump_whitelist = alloc_whitelist(batch_length);
+
+	shadow_addr = gen8_canonical_addr(shadow->node.start);
+	batch_addr = gen8_canonical_addr(batch->node.start + batch_offset);
+
+	/*
+	 * We use the batch length as size because the shadow object is as
+	 * large or larger and copy_batch() will write MI_NOPs to the extra
+	 * space. Parsing should be faster in some cases this way.
+	 */
+	batch_end = cmd + batch_length / sizeof(*batch_end);
+	do {
+		u32 length;
+
+		if (*cmd == MI_BATCH_BUFFER_END)
+			break;
+
+		desc = find_cmd(engine, *cmd, desc, &default_desc);
+		if (!desc) {
+			DRM_DEBUG("CMD: Unrecognized command: 0x%08X\n", *cmd);
+			ret = -EINVAL;
+			break;
+		}
+
+		if (desc->flags & CMD_DESC_FIXED)
+			length = desc->length.fixed;
+		else
+			length = (*cmd & desc->length.mask) + LENGTH_BIAS;
+
+		if ((batch_end - cmd) < length) {
+			DRM_DEBUG("CMD: Command length exceeds batch length: 0x%08X length=%u batchlen=%td\n",
+				  *cmd,
+				  length,
+				  batch_end - cmd);
+			ret = -EINVAL;
+			break;
+		}
+
+		if (!check_cmd(engine, desc, cmd, length)) {
+			ret = -EACCES;
+			break;
+		}
+
+		if (cmd_desc_is(desc, MI_BATCH_BUFFER_START)) {
+			ret = check_bbstart(cmd, offset, length, batch_length,
+					    batch_addr, shadow_addr,
+					    jump_whitelist);
+			break;
+		}
+
+		if (!IS_ERR_OR_NULL(jump_whitelist))
+			__set_bit(offset, jump_whitelist);
+
+		cmd += length;
+		offset += length;
+		if  (cmd >= batch_end) {
+			DRM_DEBUG("CMD: Got to the end of the buffer w/o a BBE cmd!\n");
+			ret = -EINVAL;
+			break;
+		}
+	} while (1);
+
+	if (trampoline) {
+		/*
+		 * With the trampoline, the shadow is executed twice.
+		 *
+		 *   1 - starting at offset 0, in privileged mode
+		 *   2 - starting at offset batch_len, as non-privileged
+		 *
+		 * Only if the batch is valid and safe to execute, do we
+		 * allow the first privileged execution to proceed. If not,
+		 * we terminate the first batch and use the second batchbuffer
+		 * entry to chain to the original unsafe non-privileged batch,
+		 * leaving it to the HW to validate.
+		 */
+		*batch_end = MI_BATCH_BUFFER_END;
+
+		if (ret) {
+			/* Batch unsafe to execute with privileges, cancel! */
+			cmd = page_mask_bits(shadow->obj->mm.mapping);
+			*cmd = MI_BATCH_BUFFER_END;
+
+			/* If batch is unsafe but valid, jump to the original */
+			if (ret == -EACCES) {
+				unsigned int flags;
+
+				flags = MI_BATCH_NON_SECURE_I965;
+				if (IS_HASWELL(engine->i915))
+					flags = MI_BATCH_NON_SECURE_HSW;
+
+				GEM_BUG_ON(!IS_GRAPHICS_VER(engine->i915, 6, 7));
+				__gen6_emit_bb_start(batch_end,
+						     batch_addr,
+						     flags);
+
+				ret = 0; /* allow execution */
+			}
+		}
+	}
+
+	i915_gem_object_flush_map(shadow->obj);
+
+	if (!IS_ERR_OR_NULL(jump_whitelist))
+		kfree(jump_whitelist);
+	i915_gem_object_unpin_map(shadow->obj);
+	return ret;
+}
+
+/**
+ * i915_cmd_parser_get_version() - get the cmd parser version number
+ * @dev_priv: i915 device private
+ *
+ * The cmd parser maintains a simple increasing integer version number suitable
+ * for passing to userspace clients to determine what operations are permitted.
+ *
+ * Return: the current version number of the cmd parser
+ */
+int i915_cmd_parser_get_version(struct drm_i915_private *dev_priv)
+{
+	struct intel_engine_cs *engine;
+	bool active = false;
+
+	/* If the command parser is not enabled, report 0 - unsupported */
+	for_each_uabi_engine(engine, dev_priv) {
+		if (intel_engine_using_cmd_parser(engine)) {
+			active = true;
+			break;
+		}
+	}
+	if (!active)
+		return 0;
+
+	/*
+	 * Command parser version history
+	 *
+	 * 1. Initial version. Checks batches and reports violations, but leaves
+	 *    hardware parsing enabled (so does not allow new use cases).
+	 * 2. Allow access to the MI_PREDICATE_SRC0 and
+	 *    MI_PREDICATE_SRC1 registers.
+	 * 3. Allow access to the GPGPU_THREADS_DISPATCHED register.
+	 * 4. L3 atomic chicken bits of HSW_SCRATCH1 and HSW_ROW_CHICKEN3.
+	 * 5. GPGPU dispatch compute indirect registers.
+	 * 6. TIMESTAMP register and Haswell CS GPR registers
+	 * 7. Allow MI_LOAD_REGISTER_REG between whitelisted registers.
+	 * 8. Don't report cmd_check() failures as EINVAL errors to userspace;
+	 *    rely on the HW to NOOP disallowed commands as it would without
+	 *    the parser enabled.
+	 * 9. Don't whitelist or handle oacontrol specially, as ownership
+	 *    for oacontrol state is moving to i915-perf.
+	 * 10. Support for Gen9 BCS Parsing
+	 */
+	return 10;
+}
diff --git a/drivers/gpu/drm/i915/i915_cmd_parser.h b/drivers/gpu/drm/i915/i915_cmd_parser.h
new file mode 100644
index 000000000..ba70ac6c9
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_cmd_parser.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_CMD_PARSER_H__
+#define __I915_CMD_PARSER_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct intel_engine_cs;
+struct i915_vma;
+
+int i915_cmd_parser_get_version(struct drm_i915_private *dev_priv);
+int intel_engine_init_cmd_parser(struct intel_engine_cs *engine);
+void intel_engine_cleanup_cmd_parser(struct intel_engine_cs *engine);
+int intel_engine_cmd_parser(struct intel_engine_cs *engine,
+			    struct i915_vma *batch,
+			    unsigned long batch_offset,
+			    unsigned long batch_length,
+			    struct i915_vma *shadow,
+			    bool trampoline);
+#define I915_CMD_PARSER_TRAMPOLINE_SIZE 8
+
+#endif /* __I915_CMD_PARSER_H__ */
diff --git a/drivers/gpu/drm/i915/i915_config.c b/drivers/gpu/drm/i915/i915_config.c
new file mode 100644
index 000000000..afb828dab
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_config.c
@@ -0,0 +1,15 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "i915_drv.h"
+
+unsigned long
+i915_fence_context_timeout(const struct drm_i915_private *i915, u64 context)
+{
+	if (CONFIG_DRM_I915_FENCE_TIMEOUT && context)
+		return msecs_to_jiffies_timeout(CONFIG_DRM_I915_FENCE_TIMEOUT);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
new file mode 100644
index 000000000..ae987e922
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -0,0 +1,800 @@
+/*
+ * Copyright © 2008 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *    Keith Packard <keithp@keithp.com>
+ *
+ */
+
+#include <linux/sched/mm.h>
+#include <linux/sort.h>
+#include <linux/string_helpers.h>
+
+#include <drm/drm_debugfs.h>
+
+#include "gem/i915_gem_context.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_buffer_pool.h"
+#include "gt/intel_gt_clock_utils.h"
+#include "gt/intel_gt_debugfs.h"
+#include "gt/intel_gt_pm.h"
+#include "gt/intel_gt_pm_debugfs.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_gt_requests.h"
+#include "gt/intel_rc6.h"
+#include "gt/intel_reset.h"
+#include "gt/intel_rps.h"
+#include "gt/intel_sseu_debugfs.h"
+
+#include "i915_debugfs.h"
+#include "i915_debugfs_params.h"
+#include "i915_driver.h"
+#include "i915_irq.h"
+#include "i915_scheduler.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pm.h"
+
+static inline struct drm_i915_private *node_to_i915(struct drm_info_node *node)
+{
+	return to_i915(node->minor->dev);
+}
+
+static int i915_capabilities(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *i915 = node_to_i915(m->private);
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	seq_printf(m, "pch: %d\n", INTEL_PCH_TYPE(i915));
+
+	intel_device_info_print(INTEL_INFO(i915), RUNTIME_INFO(i915), &p);
+	i915_print_iommu_status(i915, &p);
+	intel_gt_info_print(&to_gt(i915)->info, &p);
+	intel_driver_caps_print(&i915->caps, &p);
+
+	kernel_param_lock(THIS_MODULE);
+	i915_params_dump(&i915->params, &p);
+	kernel_param_unlock(THIS_MODULE);
+
+	return 0;
+}
+
+static char get_tiling_flag(struct drm_i915_gem_object *obj)
+{
+	switch (i915_gem_object_get_tiling(obj)) {
+	default:
+	case I915_TILING_NONE: return ' ';
+	case I915_TILING_X: return 'X';
+	case I915_TILING_Y: return 'Y';
+	}
+}
+
+static char get_global_flag(struct drm_i915_gem_object *obj)
+{
+	return READ_ONCE(obj->userfault_count) ? 'g' : ' ';
+}
+
+static char get_pin_mapped_flag(struct drm_i915_gem_object *obj)
+{
+	return obj->mm.mapping ? 'M' : ' ';
+}
+
+static const char *
+stringify_page_sizes(unsigned int page_sizes, char *buf, size_t len)
+{
+	size_t x = 0;
+
+	switch (page_sizes) {
+	case 0:
+		return "";
+	case I915_GTT_PAGE_SIZE_4K:
+		return "4K";
+	case I915_GTT_PAGE_SIZE_64K:
+		return "64K";
+	case I915_GTT_PAGE_SIZE_2M:
+		return "2M";
+	default:
+		if (!buf)
+			return "M";
+
+		if (page_sizes & I915_GTT_PAGE_SIZE_2M)
+			x += snprintf(buf + x, len - x, "2M, ");
+		if (page_sizes & I915_GTT_PAGE_SIZE_64K)
+			x += snprintf(buf + x, len - x, "64K, ");
+		if (page_sizes & I915_GTT_PAGE_SIZE_4K)
+			x += snprintf(buf + x, len - x, "4K, ");
+		buf[x-2] = '\0';
+
+		return buf;
+	}
+}
+
+static const char *stringify_vma_type(const struct i915_vma *vma)
+{
+	if (i915_vma_is_ggtt(vma))
+		return "ggtt";
+
+	if (i915_vma_is_dpt(vma))
+		return "dpt";
+
+	return "ppgtt";
+}
+
+static const char *i915_cache_level_str(struct drm_i915_private *i915, int type)
+{
+	switch (type) {
+	case I915_CACHE_NONE: return " uncached";
+	case I915_CACHE_LLC: return HAS_LLC(i915) ? " LLC" : " snooped";
+	case I915_CACHE_L3_LLC: return " L3+LLC";
+	case I915_CACHE_WT: return " WT";
+	default: return "";
+	}
+}
+
+void
+i915_debugfs_describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
+	struct i915_vma *vma;
+	int pin_count = 0;
+
+	seq_printf(m, "%pK: %c%c%c %8zdKiB %02x %02x %s%s%s",
+		   &obj->base,
+		   get_tiling_flag(obj),
+		   get_global_flag(obj),
+		   get_pin_mapped_flag(obj),
+		   obj->base.size / 1024,
+		   obj->read_domains,
+		   obj->write_domain,
+		   i915_cache_level_str(dev_priv, obj->cache_level),
+		   obj->mm.dirty ? " dirty" : "",
+		   obj->mm.madv == I915_MADV_DONTNEED ? " purgeable" : "");
+	if (obj->base.name)
+		seq_printf(m, " (name: %d)", obj->base.name);
+
+	spin_lock(&obj->vma.lock);
+	list_for_each_entry(vma, &obj->vma.list, obj_link) {
+		if (!drm_mm_node_allocated(&vma->node))
+			continue;
+
+		spin_unlock(&obj->vma.lock);
+
+		if (i915_vma_is_pinned(vma))
+			pin_count++;
+
+		seq_printf(m, " (%s offset: %08llx, size: %08llx, pages: %s",
+			   stringify_vma_type(vma),
+			   vma->node.start, vma->node.size,
+			   stringify_page_sizes(vma->resource->page_sizes_gtt,
+						NULL, 0));
+		if (i915_vma_is_ggtt(vma) || i915_vma_is_dpt(vma)) {
+			switch (vma->gtt_view.type) {
+			case I915_GTT_VIEW_NORMAL:
+				seq_puts(m, ", normal");
+				break;
+
+			case I915_GTT_VIEW_PARTIAL:
+				seq_printf(m, ", partial [%08llx+%x]",
+					   vma->gtt_view.partial.offset << PAGE_SHIFT,
+					   vma->gtt_view.partial.size << PAGE_SHIFT);
+				break;
+
+			case I915_GTT_VIEW_ROTATED:
+				seq_printf(m, ", rotated [(%ux%u, src_stride=%u, dst_stride=%u, offset=%u), (%ux%u, src_stride=%u, dst_stride=%u, offset=%u)]",
+					   vma->gtt_view.rotated.plane[0].width,
+					   vma->gtt_view.rotated.plane[0].height,
+					   vma->gtt_view.rotated.plane[0].src_stride,
+					   vma->gtt_view.rotated.plane[0].dst_stride,
+					   vma->gtt_view.rotated.plane[0].offset,
+					   vma->gtt_view.rotated.plane[1].width,
+					   vma->gtt_view.rotated.plane[1].height,
+					   vma->gtt_view.rotated.plane[1].src_stride,
+					   vma->gtt_view.rotated.plane[1].dst_stride,
+					   vma->gtt_view.rotated.plane[1].offset);
+				break;
+
+			case I915_GTT_VIEW_REMAPPED:
+				seq_printf(m, ", remapped [(%ux%u, src_stride=%u, dst_stride=%u, offset=%u), (%ux%u, src_stride=%u, dst_stride=%u, offset=%u)]",
+					   vma->gtt_view.remapped.plane[0].width,
+					   vma->gtt_view.remapped.plane[0].height,
+					   vma->gtt_view.remapped.plane[0].src_stride,
+					   vma->gtt_view.remapped.plane[0].dst_stride,
+					   vma->gtt_view.remapped.plane[0].offset,
+					   vma->gtt_view.remapped.plane[1].width,
+					   vma->gtt_view.remapped.plane[1].height,
+					   vma->gtt_view.remapped.plane[1].src_stride,
+					   vma->gtt_view.remapped.plane[1].dst_stride,
+					   vma->gtt_view.remapped.plane[1].offset);
+				break;
+
+			default:
+				MISSING_CASE(vma->gtt_view.type);
+				break;
+			}
+		}
+		if (vma->fence)
+			seq_printf(m, " , fence: %d", vma->fence->id);
+		seq_puts(m, ")");
+
+		spin_lock(&obj->vma.lock);
+	}
+	spin_unlock(&obj->vma.lock);
+
+	seq_printf(m, " (pinned x %d)", pin_count);
+	if (i915_gem_object_is_stolen(obj))
+		seq_printf(m, " (stolen: %08llx)", obj->stolen->start);
+	if (i915_gem_object_is_framebuffer(obj))
+		seq_printf(m, " (fb)");
+}
+
+static int i915_gem_object_info(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *i915 = node_to_i915(m->private);
+	struct drm_printer p = drm_seq_file_printer(m);
+	struct intel_memory_region *mr;
+	enum intel_region_id id;
+
+	seq_printf(m, "%u shrinkable [%u free] objects, %llu bytes\n",
+		   i915->mm.shrink_count,
+		   atomic_read(&i915->mm.free_count),
+		   i915->mm.shrink_memory);
+	for_each_memory_region(mr, i915, id)
+		intel_memory_region_debug(mr, &p);
+
+	return 0;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+static ssize_t gpu_state_read(struct file *file, char __user *ubuf,
+			      size_t count, loff_t *pos)
+{
+	struct i915_gpu_coredump *error;
+	ssize_t ret;
+	void *buf;
+
+	error = file->private_data;
+	if (!error)
+		return 0;
+
+	/* Bounce buffer required because of kernfs __user API convenience. */
+	buf = kmalloc(count, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	ret = i915_gpu_coredump_copy_to_buffer(error, buf, *pos, count);
+	if (ret <= 0)
+		goto out;
+
+	if (!copy_to_user(ubuf, buf, ret))
+		*pos += ret;
+	else
+		ret = -EFAULT;
+
+out:
+	kfree(buf);
+	return ret;
+}
+
+static int gpu_state_release(struct inode *inode, struct file *file)
+{
+	i915_gpu_coredump_put(file->private_data);
+	return 0;
+}
+
+static int i915_gpu_info_open(struct inode *inode, struct file *file)
+{
+	struct drm_i915_private *i915 = inode->i_private;
+	struct i915_gpu_coredump *gpu;
+	intel_wakeref_t wakeref;
+
+	gpu = NULL;
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
+		gpu = i915_gpu_coredump(to_gt(i915), ALL_ENGINES, CORE_DUMP_FLAG_NONE);
+
+	if (IS_ERR(gpu))
+		return PTR_ERR(gpu);
+
+	file->private_data = gpu;
+	return 0;
+}
+
+static const struct file_operations i915_gpu_info_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_gpu_info_open,
+	.read = gpu_state_read,
+	.llseek = default_llseek,
+	.release = gpu_state_release,
+};
+
+static ssize_t
+i915_error_state_write(struct file *filp,
+		       const char __user *ubuf,
+		       size_t cnt,
+		       loff_t *ppos)
+{
+	struct i915_gpu_coredump *error = filp->private_data;
+
+	if (!error)
+		return 0;
+
+	drm_dbg(&error->i915->drm, "Resetting error state\n");
+	i915_reset_error_state(error->i915);
+
+	return cnt;
+}
+
+static int i915_error_state_open(struct inode *inode, struct file *file)
+{
+	struct i915_gpu_coredump *error;
+
+	error = i915_first_error_state(inode->i_private);
+	if (IS_ERR(error))
+		return PTR_ERR(error);
+
+	file->private_data  = error;
+	return 0;
+}
+
+static const struct file_operations i915_error_state_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_error_state_open,
+	.read = gpu_state_read,
+	.write = i915_error_state_write,
+	.llseek = default_llseek,
+	.release = gpu_state_release,
+};
+#endif
+
+static int i915_frequency_info(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *i915 = node_to_i915(m->private);
+	struct intel_gt *gt = to_gt(i915);
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	intel_gt_pm_frequency_dump(gt, &p);
+
+	return 0;
+}
+
+static const char *swizzle_string(unsigned swizzle)
+{
+	switch (swizzle) {
+	case I915_BIT_6_SWIZZLE_NONE:
+		return "none";
+	case I915_BIT_6_SWIZZLE_9:
+		return "bit9";
+	case I915_BIT_6_SWIZZLE_9_10:
+		return "bit9/bit10";
+	case I915_BIT_6_SWIZZLE_9_11:
+		return "bit9/bit11";
+	case I915_BIT_6_SWIZZLE_9_10_11:
+		return "bit9/bit10/bit11";
+	case I915_BIT_6_SWIZZLE_9_17:
+		return "bit9/bit17";
+	case I915_BIT_6_SWIZZLE_9_10_17:
+		return "bit9/bit10/bit17";
+	case I915_BIT_6_SWIZZLE_UNKNOWN:
+		return "unknown";
+	}
+
+	return "bug";
+}
+
+static int i915_swizzle_info(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	intel_wakeref_t wakeref;
+
+	seq_printf(m, "bit6 swizzle for X-tiling = %s\n",
+		   swizzle_string(to_gt(dev_priv)->ggtt->bit_6_swizzle_x));
+	seq_printf(m, "bit6 swizzle for Y-tiling = %s\n",
+		   swizzle_string(to_gt(dev_priv)->ggtt->bit_6_swizzle_y));
+
+	if (dev_priv->gem_quirks & GEM_QUIRK_PIN_SWIZZLED_PAGES)
+		seq_puts(m, "L-shaped memory detected\n");
+
+	/* On BDW+, swizzling is not used. See detect_bit_6_swizzle() */
+	if (GRAPHICS_VER(dev_priv) >= 8 || IS_VALLEYVIEW(dev_priv))
+		return 0;
+
+	wakeref = intel_runtime_pm_get(&dev_priv->runtime_pm);
+
+	if (IS_GRAPHICS_VER(dev_priv, 3, 4)) {
+		seq_printf(m, "DDC = 0x%08x\n",
+			   intel_uncore_read(uncore, DCC));
+		seq_printf(m, "DDC2 = 0x%08x\n",
+			   intel_uncore_read(uncore, DCC2));
+		seq_printf(m, "C0DRB3 = 0x%04x\n",
+			   intel_uncore_read16(uncore, C0DRB3_BW));
+		seq_printf(m, "C1DRB3 = 0x%04x\n",
+			   intel_uncore_read16(uncore, C1DRB3_BW));
+	} else if (GRAPHICS_VER(dev_priv) >= 6) {
+		seq_printf(m, "MAD_DIMM_C0 = 0x%08x\n",
+			   intel_uncore_read(uncore, MAD_DIMM_C0));
+		seq_printf(m, "MAD_DIMM_C1 = 0x%08x\n",
+			   intel_uncore_read(uncore, MAD_DIMM_C1));
+		seq_printf(m, "MAD_DIMM_C2 = 0x%08x\n",
+			   intel_uncore_read(uncore, MAD_DIMM_C2));
+		seq_printf(m, "TILECTL = 0x%08x\n",
+			   intel_uncore_read(uncore, TILECTL));
+		if (GRAPHICS_VER(dev_priv) >= 8)
+			seq_printf(m, "GAMTARBMODE = 0x%08x\n",
+				   intel_uncore_read(uncore, GAMTARBMODE));
+		else
+			seq_printf(m, "ARB_MODE = 0x%08x\n",
+				   intel_uncore_read(uncore, ARB_MODE));
+		seq_printf(m, "DISP_ARB_CTL = 0x%08x\n",
+			   intel_uncore_read(uncore, DISP_ARB_CTL));
+	}
+
+	intel_runtime_pm_put(&dev_priv->runtime_pm, wakeref);
+
+	return 0;
+}
+
+static int i915_rps_boost_info(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct intel_rps *rps = &to_gt(dev_priv)->rps;
+
+	seq_printf(m, "RPS enabled? %s\n",
+		   str_yes_no(intel_rps_is_enabled(rps)));
+	seq_printf(m, "RPS active? %s\n",
+		   str_yes_no(intel_rps_is_active(rps)));
+	seq_printf(m, "GPU busy? %s\n", str_yes_no(to_gt(dev_priv)->awake));
+	seq_printf(m, "Boosts outstanding? %d\n",
+		   atomic_read(&rps->num_waiters));
+	seq_printf(m, "Interactive? %d\n", READ_ONCE(rps->power.interactive));
+	seq_printf(m, "Frequency requested %d, actual %d\n",
+		   intel_gpu_freq(rps, rps->cur_freq),
+		   intel_rps_read_actual_frequency(rps));
+	seq_printf(m, "  min hard:%d, soft:%d; max soft:%d, hard:%d\n",
+		   intel_gpu_freq(rps, rps->min_freq),
+		   intel_gpu_freq(rps, rps->min_freq_softlimit),
+		   intel_gpu_freq(rps, rps->max_freq_softlimit),
+		   intel_gpu_freq(rps, rps->max_freq));
+	seq_printf(m, "  idle:%d, efficient:%d, boost:%d\n",
+		   intel_gpu_freq(rps, rps->idle_freq),
+		   intel_gpu_freq(rps, rps->efficient_freq),
+		   intel_gpu_freq(rps, rps->boost_freq));
+
+	seq_printf(m, "Wait boosts: %d\n", READ_ONCE(rps->boosts));
+
+	return 0;
+}
+
+static int i915_runtime_pm_status(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *dev_priv = node_to_i915(m->private);
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+
+	if (!HAS_RUNTIME_PM(dev_priv))
+		seq_puts(m, "Runtime power management not supported\n");
+
+	seq_printf(m, "Runtime power status: %s\n",
+		   str_enabled_disabled(!dev_priv->display.power.domains.init_wakeref));
+
+	seq_printf(m, "GPU idle: %s\n", str_yes_no(!to_gt(dev_priv)->awake));
+	seq_printf(m, "IRQs disabled: %s\n",
+		   str_yes_no(!intel_irqs_enabled(dev_priv)));
+#ifdef CONFIG_PM
+	seq_printf(m, "Usage count: %d\n",
+		   atomic_read(&dev_priv->drm.dev->power.usage_count));
+#else
+	seq_printf(m, "Device Power Management (CONFIG_PM) disabled\n");
+#endif
+	seq_printf(m, "PCI device power state: %s [%d]\n",
+		   pci_power_name(pdev->current_state),
+		   pdev->current_state);
+
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)) {
+		struct drm_printer p = drm_seq_file_printer(m);
+
+		print_intel_runtime_pm_wakeref(&dev_priv->runtime_pm, &p);
+	}
+
+	return 0;
+}
+
+static int i915_engine_info(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *i915 = node_to_i915(m->private);
+	struct intel_engine_cs *engine;
+	intel_wakeref_t wakeref;
+	struct drm_printer p;
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	seq_printf(m, "GT awake? %s [%d], %llums\n",
+		   str_yes_no(to_gt(i915)->awake),
+		   atomic_read(&to_gt(i915)->wakeref.count),
+		   ktime_to_ms(intel_gt_get_awake_time(to_gt(i915))));
+	seq_printf(m, "CS timestamp frequency: %u Hz, %d ns\n",
+		   to_gt(i915)->clock_frequency,
+		   to_gt(i915)->clock_period_ns);
+
+	p = drm_seq_file_printer(m);
+	for_each_uabi_engine(engine, i915)
+		intel_engine_dump(engine, &p, "%s\n", engine->name);
+
+	intel_gt_show_timelines(to_gt(i915), &p, i915_request_show_with_schedule);
+
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+
+	return 0;
+}
+
+static int i915_wa_registers(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *i915 = node_to_i915(m->private);
+	struct intel_engine_cs *engine;
+
+	for_each_uabi_engine(engine, i915) {
+		const struct i915_wa_list *wal = &engine->ctx_wa_list;
+		const struct i915_wa *wa;
+		unsigned int count;
+
+		count = wal->count;
+		if (!count)
+			continue;
+
+		seq_printf(m, "%s: Workarounds applied: %u\n",
+			   engine->name, count);
+
+		for (wa = wal->list; count--; wa++)
+			seq_printf(m, "0x%X: 0x%08X, mask: 0x%08X\n",
+				   i915_mmio_reg_offset(wa->reg),
+				   wa->set, wa->clr);
+
+		seq_printf(m, "\n");
+	}
+
+	return 0;
+}
+
+static int i915_wedged_get(void *data, u64 *val)
+{
+	struct drm_i915_private *i915 = data;
+
+	return intel_gt_debugfs_reset_show(to_gt(i915), val);
+}
+
+static int i915_wedged_set(void *data, u64 val)
+{
+	struct drm_i915_private *i915 = data;
+	intel_gt_debugfs_reset_store(to_gt(i915), val);
+
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(i915_wedged_fops,
+			i915_wedged_get, i915_wedged_set,
+			"%llu\n");
+
+static int
+i915_perf_noa_delay_set(void *data, u64 val)
+{
+	struct drm_i915_private *i915 = data;
+
+	/*
+	 * This would lead to infinite waits as we're doing timestamp
+	 * difference on the CS with only 32bits.
+	 */
+	if (intel_gt_ns_to_clock_interval(to_gt(i915), val) > U32_MAX)
+		return -EINVAL;
+
+	atomic64_set(&i915->perf.noa_programming_delay, val);
+	return 0;
+}
+
+static int
+i915_perf_noa_delay_get(void *data, u64 *val)
+{
+	struct drm_i915_private *i915 = data;
+
+	*val = atomic64_read(&i915->perf.noa_programming_delay);
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(i915_perf_noa_delay_fops,
+			i915_perf_noa_delay_get,
+			i915_perf_noa_delay_set,
+			"%llu\n");
+
+#define DROP_UNBOUND	BIT(0)
+#define DROP_BOUND	BIT(1)
+#define DROP_RETIRE	BIT(2)
+#define DROP_ACTIVE	BIT(3)
+#define DROP_FREED	BIT(4)
+#define DROP_SHRINK_ALL	BIT(5)
+#define DROP_IDLE	BIT(6)
+#define DROP_RESET_ACTIVE	BIT(7)
+#define DROP_RESET_SEQNO	BIT(8)
+#define DROP_RCU	BIT(9)
+#define DROP_ALL (DROP_UNBOUND	| \
+		  DROP_BOUND	| \
+		  DROP_RETIRE	| \
+		  DROP_ACTIVE	| \
+		  DROP_FREED	| \
+		  DROP_SHRINK_ALL |\
+		  DROP_IDLE	| \
+		  DROP_RESET_ACTIVE | \
+		  DROP_RESET_SEQNO | \
+		  DROP_RCU)
+static int
+i915_drop_caches_get(void *data, u64 *val)
+{
+	*val = DROP_ALL;
+
+	return 0;
+}
+static int
+gt_drop_caches(struct intel_gt *gt, u64 val)
+{
+	int ret;
+
+	if (val & DROP_RESET_ACTIVE &&
+	    wait_for(intel_engines_are_idle(gt), I915_IDLE_ENGINES_TIMEOUT))
+		intel_gt_set_wedged(gt);
+
+	if (val & DROP_RETIRE)
+		intel_gt_retire_requests(gt);
+
+	if (val & (DROP_IDLE | DROP_ACTIVE)) {
+		ret = intel_gt_wait_for_idle(gt, MAX_SCHEDULE_TIMEOUT);
+		if (ret)
+			return ret;
+	}
+
+	if (val & DROP_IDLE) {
+		ret = intel_gt_pm_wait_for_idle(gt);
+		if (ret)
+			return ret;
+	}
+
+	if (val & DROP_RESET_ACTIVE && intel_gt_terminally_wedged(gt))
+		intel_gt_handle_error(gt, ALL_ENGINES, 0, NULL);
+
+	if (val & DROP_FREED)
+		intel_gt_flush_buffer_pool(gt);
+
+	return 0;
+}
+
+static int
+i915_drop_caches_set(void *data, u64 val)
+{
+	struct drm_i915_private *i915 = data;
+	unsigned int flags;
+	int ret;
+
+	DRM_DEBUG("Dropping caches: 0x%08llx [0x%08llx]\n",
+		  val, val & DROP_ALL);
+
+	ret = gt_drop_caches(to_gt(i915), val);
+	if (ret)
+		return ret;
+
+	fs_reclaim_acquire(GFP_KERNEL);
+	flags = memalloc_noreclaim_save();
+	if (val & DROP_BOUND)
+		i915_gem_shrink(NULL, i915, LONG_MAX, NULL, I915_SHRINK_BOUND);
+
+	if (val & DROP_UNBOUND)
+		i915_gem_shrink(NULL, i915, LONG_MAX, NULL, I915_SHRINK_UNBOUND);
+
+	if (val & DROP_SHRINK_ALL)
+		i915_gem_shrink_all(i915);
+	memalloc_noreclaim_restore(flags);
+	fs_reclaim_release(GFP_KERNEL);
+
+	if (val & DROP_RCU)
+		rcu_barrier();
+
+	if (val & DROP_FREED)
+		i915_gem_drain_freed_objects(i915);
+
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(i915_drop_caches_fops,
+			i915_drop_caches_get, i915_drop_caches_set,
+			"0x%08llx\n");
+
+static int i915_sseu_status(struct seq_file *m, void *unused)
+{
+	struct drm_i915_private *i915 = node_to_i915(m->private);
+	struct intel_gt *gt = to_gt(i915);
+
+	return intel_sseu_status(m, gt);
+}
+
+static int i915_forcewake_open(struct inode *inode, struct file *file)
+{
+	struct drm_i915_private *i915 = inode->i_private;
+	intel_gt_pm_debugfs_forcewake_user_open(to_gt(i915));
+
+	return 0;
+}
+
+static int i915_forcewake_release(struct inode *inode, struct file *file)
+{
+	struct drm_i915_private *i915 = inode->i_private;
+	intel_gt_pm_debugfs_forcewake_user_release(to_gt(i915));
+
+	return 0;
+}
+
+static const struct file_operations i915_forcewake_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_forcewake_open,
+	.release = i915_forcewake_release,
+};
+
+static const struct drm_info_list i915_debugfs_list[] = {
+	{"i915_capabilities", i915_capabilities, 0},
+	{"i915_gem_objects", i915_gem_object_info, 0},
+	{"i915_frequency_info", i915_frequency_info, 0},
+	{"i915_swizzle_info", i915_swizzle_info, 0},
+	{"i915_runtime_pm_status", i915_runtime_pm_status, 0},
+	{"i915_engine_info", i915_engine_info, 0},
+	{"i915_wa_registers", i915_wa_registers, 0},
+	{"i915_sseu_status", i915_sseu_status, 0},
+	{"i915_rps_boost_info", i915_rps_boost_info, 0},
+};
+#define I915_DEBUGFS_ENTRIES ARRAY_SIZE(i915_debugfs_list)
+
+static const struct i915_debugfs_files {
+	const char *name;
+	const struct file_operations *fops;
+} i915_debugfs_files[] = {
+	{"i915_perf_noa_delay", &i915_perf_noa_delay_fops},
+	{"i915_wedged", &i915_wedged_fops},
+	{"i915_gem_drop_caches", &i915_drop_caches_fops},
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+	{"i915_error_state", &i915_error_state_fops},
+	{"i915_gpu_info", &i915_gpu_info_fops},
+#endif
+};
+
+void i915_debugfs_register(struct drm_i915_private *dev_priv)
+{
+	struct drm_minor *minor = dev_priv->drm.primary;
+	int i;
+
+	i915_debugfs_params(dev_priv);
+
+	debugfs_create_file("i915_forcewake_user", S_IRUSR, minor->debugfs_root,
+			    to_i915(minor->dev), &i915_forcewake_fops);
+	for (i = 0; i < ARRAY_SIZE(i915_debugfs_files); i++) {
+		debugfs_create_file(i915_debugfs_files[i].name,
+				    S_IRUGO | S_IWUSR,
+				    minor->debugfs_root,
+				    to_i915(minor->dev),
+				    i915_debugfs_files[i].fops);
+	}
+
+	drm_debugfs_create_files(i915_debugfs_list,
+				 I915_DEBUGFS_ENTRIES,
+				 minor->debugfs_root, minor);
+}
diff --git a/drivers/gpu/drm/i915/i915_debugfs.h b/drivers/gpu/drm/i915/i915_debugfs.h
new file mode 100644
index 000000000..1de2736f1
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_debugfs.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_DEBUGFS_H__
+#define __I915_DEBUGFS_H__
+
+struct drm_connector;
+struct drm_i915_gem_object;
+struct drm_i915_private;
+struct seq_file;
+
+#ifdef CONFIG_DEBUG_FS
+void i915_debugfs_register(struct drm_i915_private *dev_priv);
+void i915_debugfs_describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj);
+#else
+static inline void i915_debugfs_register(struct drm_i915_private *dev_priv) {}
+static inline void i915_debugfs_describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj) {}
+#endif
+
+#endif /* __I915_DEBUGFS_H__ */
diff --git a/drivers/gpu/drm/i915/i915_debugfs_params.c b/drivers/gpu/drm/i915/i915_debugfs_params.c
new file mode 100644
index 000000000..783c8676e
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_debugfs_params.c
@@ -0,0 +1,277 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/kernel.h>
+
+#include "i915_debugfs_params.h"
+#include "gt/intel_gt.h"
+#include "gt/uc/intel_guc.h"
+#include "i915_drv.h"
+#include "i915_params.h"
+
+#define MATCH_DEBUGFS_NODE_NAME(_file, _name) \
+	(strcmp((_file)->f_path.dentry->d_name.name, (_name)) == 0)
+
+#define GET_I915(i915, name, ptr)	\
+	do {	\
+		struct i915_params *params;	\
+		params = container_of(((void *)(ptr)), typeof(*params), name);	\
+		(i915) = container_of(params, typeof(*(i915)), params);	\
+	} while (0)
+
+/* int param */
+static int i915_param_int_show(struct seq_file *m, void *data)
+{
+	int *value = m->private;
+
+	seq_printf(m, "%d\n", *value);
+
+	return 0;
+}
+
+static int i915_param_int_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, i915_param_int_show, inode->i_private);
+}
+
+static int notify_guc(struct drm_i915_private *i915)
+{
+	int ret = 0;
+
+	if (intel_uc_uses_guc_submission(&to_gt(i915)->uc))
+		ret = intel_guc_global_policies_update(&to_gt(i915)->uc.guc);
+
+	return ret;
+}
+
+static ssize_t i915_param_int_write(struct file *file,
+				    const char __user *ubuf, size_t len,
+				    loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	int *value = m->private;
+	int ret;
+
+	ret = kstrtoint_from_user(ubuf, len, 0, value);
+	if (ret) {
+		/* support boolean values too */
+		bool b;
+
+		ret = kstrtobool_from_user(ubuf, len, &b);
+		if (!ret)
+			*value = b;
+	}
+
+	return ret ?: len;
+}
+
+static const struct file_operations i915_param_int_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_param_int_open,
+	.read = seq_read,
+	.write = i915_param_int_write,
+	.llseek = default_llseek,
+	.release = single_release,
+};
+
+static const struct file_operations i915_param_int_fops_ro = {
+	.owner = THIS_MODULE,
+	.open = i915_param_int_open,
+	.read = seq_read,
+	.llseek = default_llseek,
+	.release = single_release,
+};
+
+/* unsigned int param */
+static int i915_param_uint_show(struct seq_file *m, void *data)
+{
+	unsigned int *value = m->private;
+
+	seq_printf(m, "%u\n", *value);
+
+	return 0;
+}
+
+static int i915_param_uint_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, i915_param_uint_show, inode->i_private);
+}
+
+static ssize_t i915_param_uint_write(struct file *file,
+				     const char __user *ubuf, size_t len,
+				     loff_t *offp)
+{
+	struct drm_i915_private *i915;
+	struct seq_file *m = file->private_data;
+	unsigned int *value = m->private;
+	unsigned int old = *value;
+	int ret;
+
+	ret = kstrtouint_from_user(ubuf, len, 0, value);
+	if (ret) {
+		/* support boolean values too */
+		bool b;
+
+		ret = kstrtobool_from_user(ubuf, len, &b);
+		if (!ret)
+			*value = b;
+	}
+
+	if (!ret && MATCH_DEBUGFS_NODE_NAME(file, "reset")) {
+		GET_I915(i915, reset, value);
+
+		ret = notify_guc(i915);
+		if (ret)
+			*value = old;
+	}
+
+	return ret ?: len;
+}
+
+static const struct file_operations i915_param_uint_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_param_uint_open,
+	.read = seq_read,
+	.write = i915_param_uint_write,
+	.llseek = default_llseek,
+	.release = single_release,
+};
+
+static const struct file_operations i915_param_uint_fops_ro = {
+	.owner = THIS_MODULE,
+	.open = i915_param_uint_open,
+	.read = seq_read,
+	.llseek = default_llseek,
+	.release = single_release,
+};
+
+/* char * param */
+static int i915_param_charp_show(struct seq_file *m, void *data)
+{
+	const char **s = m->private;
+
+	seq_printf(m, "%s\n", *s);
+
+	return 0;
+}
+
+static int i915_param_charp_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, i915_param_charp_show, inode->i_private);
+}
+
+static ssize_t i915_param_charp_write(struct file *file,
+				      const char __user *ubuf, size_t len,
+				      loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	char **s = m->private;
+	char *new, *old;
+
+	old = *s;
+	new = strndup_user(ubuf, PAGE_SIZE);
+	if (IS_ERR(new)) {
+		len = PTR_ERR(new);
+		goto out;
+	}
+
+	*s = new;
+
+	kfree(old);
+out:
+	return len;
+}
+
+static const struct file_operations i915_param_charp_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_param_charp_open,
+	.read = seq_read,
+	.write = i915_param_charp_write,
+	.llseek = default_llseek,
+	.release = single_release,
+};
+
+static const struct file_operations i915_param_charp_fops_ro = {
+	.owner = THIS_MODULE,
+	.open = i915_param_charp_open,
+	.read = seq_read,
+	.llseek = default_llseek,
+	.release = single_release,
+};
+
+#define RO(mode) (((mode) & 0222) == 0)
+
+static struct dentry *
+i915_debugfs_create_int(const char *name, umode_t mode,
+			struct dentry *parent, int *value)
+{
+	return debugfs_create_file_unsafe(name, mode, parent, value,
+					  RO(mode) ? &i915_param_int_fops_ro :
+					  &i915_param_int_fops);
+}
+
+static struct dentry *
+i915_debugfs_create_uint(const char *name, umode_t mode,
+			 struct dentry *parent, unsigned int *value)
+{
+	return debugfs_create_file_unsafe(name, mode, parent, value,
+					  RO(mode) ? &i915_param_uint_fops_ro :
+					  &i915_param_uint_fops);
+}
+
+static struct dentry *
+i915_debugfs_create_charp(const char *name, umode_t mode,
+			  struct dentry *parent, char **value)
+{
+	return debugfs_create_file(name, mode, parent, value,
+				   RO(mode) ? &i915_param_charp_fops_ro :
+				   &i915_param_charp_fops);
+}
+
+static __always_inline void
+_i915_param_create_file(struct dentry *parent, const char *name,
+			const char *type, int mode, void *value)
+{
+	if (!mode)
+		return;
+
+	if (!__builtin_strcmp(type, "bool"))
+		debugfs_create_bool(name, mode, parent, value);
+	else if (!__builtin_strcmp(type, "int"))
+		i915_debugfs_create_int(name, mode, parent, value);
+	else if (!__builtin_strcmp(type, "unsigned int"))
+		i915_debugfs_create_uint(name, mode, parent, value);
+	else if (!__builtin_strcmp(type, "unsigned long"))
+		debugfs_create_ulong(name, mode, parent, value);
+	else if (!__builtin_strcmp(type, "char *"))
+		i915_debugfs_create_charp(name, mode, parent, value);
+	else
+		WARN(1, "no debugfs fops defined for param type %s (i915.%s)\n",
+		     type, name);
+}
+
+/* add a subdirectory with files for each i915 param */
+struct dentry *i915_debugfs_params(struct drm_i915_private *i915)
+{
+	struct drm_minor *minor = i915->drm.primary;
+	struct i915_params *params = &i915->params;
+	struct dentry *dir;
+
+	dir = debugfs_create_dir("i915_params", minor->debugfs_root);
+	if (IS_ERR(dir))
+		return dir;
+
+	/*
+	 * Note: We could create files for params needing special handling
+	 * here. Set mode in params to 0 to skip the generic create file, or
+	 * just let the generic create file fail silently with -EEXIST.
+	 */
+
+#define REGISTER(T, x, unused, mode, ...) _i915_param_create_file(dir, #x, #T, mode, &params->x);
+	I915_PARAMS_FOR_EACH(REGISTER);
+#undef REGISTER
+
+	return dir;
+}
diff --git a/drivers/gpu/drm/i915/i915_debugfs_params.h b/drivers/gpu/drm/i915/i915_debugfs_params.h
new file mode 100644
index 000000000..665670765
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_debugfs_params.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_DEBUGFS_PARAMS__
+#define __I915_DEBUGFS_PARAMS__
+
+struct dentry;
+struct drm_i915_private;
+
+struct dentry *i915_debugfs_params(struct drm_i915_private *i915);
+
+#endif /* __I915_DEBUGFS_PARAMS__ */
diff --git a/drivers/gpu/drm/i915/i915_deps.c b/drivers/gpu/drm/i915/i915_deps.c
new file mode 100644
index 000000000..297b8e4e4
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_deps.c
@@ -0,0 +1,237 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/dma-fence.h>
+#include <linux/slab.h>
+
+#include <drm/ttm/ttm_bo_api.h>
+
+#include "i915_deps.h"
+
+/**
+ * DOC: Set of utilities to dynamically collect dependencies into a
+ * structure which is fed into the GT migration code.
+ *
+ * Once we can do async unbinding, this is also needed to coalesce
+ * the migration fence with the unbind fences if these are coalesced
+ * post-migration.
+ *
+ * While collecting the individual dependencies, we store the refcounted
+ * struct dma_fence pointers in a realloc-managed pointer array, since
+ * that can be easily fed into a dma_fence_array. Other options are
+ * available, like for example an xarray for similarity with drm/sched.
+ * Can be changed easily if needed.
+ *
+ * A struct i915_deps need to be initialized using i915_deps_init().
+ * If i915_deps_add_dependency() or i915_deps_add_resv() return an
+ * error code they will internally call i915_deps_fini(), which frees
+ * all internal references and allocations.
+ */
+
+/* Min number of fence pointers in the array when an allocation occurs. */
+#define I915_DEPS_MIN_ALLOC_CHUNK 8U
+
+static void i915_deps_reset_fences(struct i915_deps *deps)
+{
+	if (deps->fences != &deps->single)
+		kfree(deps->fences);
+	deps->num_deps = 0;
+	deps->fences_size = 1;
+	deps->fences = &deps->single;
+}
+
+/**
+ * i915_deps_init - Initialize an i915_deps structure
+ * @deps: Pointer to the i915_deps structure to initialize.
+ * @gfp: The allocation mode for subsequenst allocations.
+ */
+void i915_deps_init(struct i915_deps *deps, gfp_t gfp)
+{
+	deps->fences = NULL;
+	deps->gfp = gfp;
+	i915_deps_reset_fences(deps);
+}
+
+/**
+ * i915_deps_fini - Finalize an i915_deps structure
+ * @deps: Pointer to the i915_deps structure to finalize.
+ *
+ * This function drops all fence references taken, conditionally frees and
+ * then resets the fences array.
+ */
+void i915_deps_fini(struct i915_deps *deps)
+{
+	unsigned int i;
+
+	for (i = 0; i < deps->num_deps; ++i)
+		dma_fence_put(deps->fences[i]);
+
+	if (deps->fences != &deps->single)
+		kfree(deps->fences);
+}
+
+static int i915_deps_grow(struct i915_deps *deps, struct dma_fence *fence,
+			  const struct ttm_operation_ctx *ctx)
+{
+	int ret;
+
+	if (deps->num_deps >= deps->fences_size) {
+		unsigned int new_size = 2 * deps->fences_size;
+		struct dma_fence **new_fences;
+
+		new_size = max(new_size, I915_DEPS_MIN_ALLOC_CHUNK);
+		new_fences = kmalloc_array(new_size, sizeof(*new_fences), deps->gfp);
+		if (!new_fences)
+			goto sync;
+
+		memcpy(new_fences, deps->fences,
+		       deps->fences_size * sizeof(*new_fences));
+		swap(new_fences, deps->fences);
+		if (new_fences != &deps->single)
+			kfree(new_fences);
+		deps->fences_size = new_size;
+	}
+	deps->fences[deps->num_deps++] = dma_fence_get(fence);
+	return 0;
+
+sync:
+	if (ctx->no_wait_gpu && !dma_fence_is_signaled(fence)) {
+		ret = -EBUSY;
+		goto unref;
+	}
+
+	ret = dma_fence_wait(fence, ctx->interruptible);
+	if (ret)
+		goto unref;
+
+	ret = fence->error;
+	if (ret)
+		goto unref;
+
+	return 0;
+
+unref:
+	i915_deps_fini(deps);
+	return ret;
+}
+
+/**
+ * i915_deps_sync - Wait for all the fences in the dependency collection
+ * @deps: Pointer to the i915_deps structure the fences of which to wait for.
+ * @ctx: Pointer to a struct ttm_operation_ctx indicating how the waits
+ * should be performed.
+ *
+ * This function waits for fences in the dependency collection. If it
+ * encounters an error during the wait or a fence error, the wait for
+ * further fences is aborted and the error returned.
+ *
+ * Return: Zero if successful, Negative error code on error.
+ */
+int i915_deps_sync(const struct i915_deps *deps, const struct ttm_operation_ctx *ctx)
+{
+	struct dma_fence **fences = deps->fences;
+	unsigned int i;
+	int ret = 0;
+
+	for (i = 0; i < deps->num_deps; ++i, ++fences) {
+		if (ctx->no_wait_gpu && !dma_fence_is_signaled(*fences)) {
+			ret = -EBUSY;
+			break;
+		}
+
+		ret = dma_fence_wait(*fences, ctx->interruptible);
+		if (!ret)
+			ret = (*fences)->error;
+		if (ret)
+			break;
+	}
+
+	return ret;
+}
+
+/**
+ * i915_deps_add_dependency - Add a fence to the dependency collection
+ * @deps: Pointer to the i915_deps structure a fence is to be added to.
+ * @fence: The fence to add.
+ * @ctx: Pointer to a struct ttm_operation_ctx indicating how waits are to
+ * be performed if waiting.
+ *
+ * Adds a fence to the dependency collection, and takes a reference on it.
+ * If the fence context is not zero and there was a later fence from the
+ * same fence context already added, then the fence is not added to the
+ * dependency collection. If the fence context is not zero and there was
+ * an earlier fence already added, then the fence will replace the older
+ * fence from the same context and the reference on the earlier fence will
+ * be dropped.
+ * If there is a failure to allocate memory to accommodate the new fence to
+ * be added, the new fence will instead be waited for and an error may
+ * be returned; depending on the value of @ctx, or if there was a fence
+ * error. If an error was returned, the dependency collection will be
+ * finalized and all fence reference dropped.
+ *
+ * Return: 0 if success. Negative error code on error.
+ */
+int i915_deps_add_dependency(struct i915_deps *deps,
+			     struct dma_fence *fence,
+			     const struct ttm_operation_ctx *ctx)
+{
+	unsigned int i;
+	int ret;
+
+	if (!fence)
+		return 0;
+
+	if (dma_fence_is_signaled(fence)) {
+		ret = fence->error;
+		if (ret)
+			i915_deps_fini(deps);
+		return ret;
+	}
+
+	for (i = 0; i < deps->num_deps; ++i) {
+		struct dma_fence *entry = deps->fences[i];
+
+		if (!entry->context || entry->context != fence->context)
+			continue;
+
+		if (dma_fence_is_later(fence, entry)) {
+			dma_fence_put(entry);
+			deps->fences[i] = dma_fence_get(fence);
+		}
+
+		return 0;
+	}
+
+	return i915_deps_grow(deps, fence, ctx);
+}
+
+/**
+ * i915_deps_add_resv - Add the fences of a reservation object to a dependency
+ * collection.
+ * @deps: Pointer to the i915_deps structure a fence is to be added to.
+ * @resv: The reservation object, then fences of which to add.
+ * @ctx: Pointer to a struct ttm_operation_ctx indicating how waits are to
+ * be performed if waiting.
+ *
+ * Calls i915_deps_add_depencency() on the indicated fences of @resv.
+ *
+ * Return: Zero on success. Negative error code on error.
+ */
+int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
+		       const struct ttm_operation_ctx *ctx)
+{
+	struct dma_resv_iter iter;
+	struct dma_fence *fence;
+
+	dma_resv_assert_held(resv);
+	dma_resv_for_each_fence(&iter, resv, dma_resv_usage_rw(true), fence) {
+		int ret = i915_deps_add_dependency(deps, fence, ctx);
+
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_deps.h b/drivers/gpu/drm/i915/i915_deps.h
new file mode 100644
index 000000000..d76c0106c
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_deps.h
@@ -0,0 +1,45 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _I915_DEPS_H_
+#define _I915_DEPS_H_
+
+#include <linux/types.h>
+
+struct ttm_operation_ctx;
+struct dma_fence;
+struct dma_resv;
+
+/**
+ * struct i915_deps - Collect dependencies into a single dma-fence
+ * @single: Storage for pointer if the collection is a single fence.
+ * @fences: Allocated array of fence pointers if more than a single fence;
+ * otherwise points to the address of @single.
+ * @num_deps: Current number of dependency fences.
+ * @fences_size: Size of the @fences array in number of pointers.
+ * @gfp: Allocation mode.
+ */
+struct i915_deps {
+	struct dma_fence *single;
+	struct dma_fence **fences;
+	unsigned int num_deps;
+	unsigned int fences_size;
+	gfp_t gfp;
+};
+
+void i915_deps_init(struct i915_deps *deps, gfp_t gfp);
+
+void i915_deps_fini(struct i915_deps *deps);
+
+int i915_deps_add_dependency(struct i915_deps *deps,
+			     struct dma_fence *fence,
+			     const struct ttm_operation_ctx *ctx);
+
+int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
+		       const struct ttm_operation_ctx *ctx);
+
+int i915_deps_sync(const struct i915_deps *deps,
+		   const struct ttm_operation_ctx *ctx);
+#endif
diff --git a/drivers/gpu/drm/i915/i915_driver.c b/drivers/gpu/drm/i915/i915_driver.c
new file mode 100644
index 000000000..be0ebed2a
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_driver.c
@@ -0,0 +1,1940 @@
+/* i915_drv.c -- i830,i845,i855,i865,i915 driver -*- linux-c -*-
+ */
+/*
+ *
+ * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/acpi.h>
+#include <linux/device.h>
+#include <linux/module.h>
+#include <linux/oom.h>
+#include <linux/pci.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+#include <linux/pnp.h>
+#include <linux/slab.h>
+#include <linux/string_helpers.h>
+#include <linux/vga_switcheroo.h>
+#include <linux/vt.h>
+
+#include <drm/drm_aperture.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_ioctl.h>
+#include <drm/drm_managed.h>
+#include <drm/drm_probe_helper.h>
+
+#include "display/intel_acpi.h"
+#include "display/intel_bw.h"
+#include "display/intel_cdclk.h"
+#include "display/intel_display_types.h"
+#include "display/intel_dmc.h"
+#include "display/intel_dp.h"
+#include "display/intel_dpt.h"
+#include "display/intel_fbdev.h"
+#include "display/intel_hotplug.h"
+#include "display/intel_overlay.h"
+#include "display/intel_pch_refclk.h"
+#include "display/intel_pipe_crc.h"
+#include "display/intel_pps.h"
+#include "display/intel_sprite.h"
+#include "display/intel_vga.h"
+#include "display/skl_watermark.h"
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_create.h"
+#include "gem/i915_gem_dmabuf.h"
+#include "gem/i915_gem_ioctls.h"
+#include "gem/i915_gem_mman.h"
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_pm.h"
+#include "gt/intel_rc6.h"
+
+#include "pxp/intel_pxp_pm.h"
+
+#include "i915_file_private.h"
+#include "i915_debugfs.h"
+#include "i915_driver.h"
+#include "i915_drm_client.h"
+#include "i915_drv.h"
+#include "i915_getparam.h"
+#include "i915_ioc32.h"
+#include "i915_ioctl.h"
+#include "i915_irq.h"
+#include "i915_memcpy.h"
+#include "i915_perf.h"
+#include "i915_query.h"
+#include "i915_suspend.h"
+#include "i915_switcheroo.h"
+#include "i915_sysfs.h"
+#include "i915_utils.h"
+#include "i915_vgpu.h"
+#include "intel_dram.h"
+#include "intel_gvt.h"
+#include "intel_memory_region.h"
+#include "intel_pci_config.h"
+#include "intel_pcode.h"
+#include "intel_pm.h"
+#include "intel_region_ttm.h"
+#include "vlv_suspend.h"
+
+/* Intel Rapid Start Technology ACPI device name */
+static const char irst_name[] = "INT3392";
+
+static const struct drm_driver i915_drm_driver;
+
+static void i915_release_bridge_dev(struct drm_device *dev,
+				    void *bridge)
+{
+	pci_dev_put(bridge);
+}
+
+static int i915_get_bridge_dev(struct drm_i915_private *dev_priv)
+{
+	int domain = pci_domain_nr(to_pci_dev(dev_priv->drm.dev)->bus);
+
+	dev_priv->bridge_dev =
+		pci_get_domain_bus_and_slot(domain, 0, PCI_DEVFN(0, 0));
+	if (!dev_priv->bridge_dev) {
+		drm_err(&dev_priv->drm, "bridge device not found\n");
+		return -EIO;
+	}
+
+	return drmm_add_action_or_reset(&dev_priv->drm, i915_release_bridge_dev,
+					dev_priv->bridge_dev);
+}
+
+/* Allocate space for the MCH regs if needed, return nonzero on error */
+static int
+intel_alloc_mchbar_resource(struct drm_i915_private *dev_priv)
+{
+	int reg = GRAPHICS_VER(dev_priv) >= 4 ? MCHBAR_I965 : MCHBAR_I915;
+	u32 temp_lo, temp_hi = 0;
+	u64 mchbar_addr;
+	int ret;
+
+	if (GRAPHICS_VER(dev_priv) >= 4)
+		pci_read_config_dword(dev_priv->bridge_dev, reg + 4, &temp_hi);
+	pci_read_config_dword(dev_priv->bridge_dev, reg, &temp_lo);
+	mchbar_addr = ((u64)temp_hi << 32) | temp_lo;
+
+	/* If ACPI doesn't have it, assume we need to allocate it ourselves */
+#ifdef CONFIG_PNP
+	if (mchbar_addr &&
+	    pnp_range_reserved(mchbar_addr, mchbar_addr + MCHBAR_SIZE))
+		return 0;
+#endif
+
+	/* Get some space for it */
+	dev_priv->mch_res.name = "i915 MCHBAR";
+	dev_priv->mch_res.flags = IORESOURCE_MEM;
+	ret = pci_bus_alloc_resource(dev_priv->bridge_dev->bus,
+				     &dev_priv->mch_res,
+				     MCHBAR_SIZE, MCHBAR_SIZE,
+				     PCIBIOS_MIN_MEM,
+				     0, pcibios_align_resource,
+				     dev_priv->bridge_dev);
+	if (ret) {
+		drm_dbg(&dev_priv->drm, "failed bus alloc: %d\n", ret);
+		dev_priv->mch_res.start = 0;
+		return ret;
+	}
+
+	if (GRAPHICS_VER(dev_priv) >= 4)
+		pci_write_config_dword(dev_priv->bridge_dev, reg + 4,
+				       upper_32_bits(dev_priv->mch_res.start));
+
+	pci_write_config_dword(dev_priv->bridge_dev, reg,
+			       lower_32_bits(dev_priv->mch_res.start));
+	return 0;
+}
+
+/* Setup MCHBAR if possible, return true if we should disable it again */
+static void
+intel_setup_mchbar(struct drm_i915_private *dev_priv)
+{
+	int mchbar_reg = GRAPHICS_VER(dev_priv) >= 4 ? MCHBAR_I965 : MCHBAR_I915;
+	u32 temp;
+	bool enabled;
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		return;
+
+	dev_priv->mchbar_need_disable = false;
+
+	if (IS_I915G(dev_priv) || IS_I915GM(dev_priv)) {
+		pci_read_config_dword(dev_priv->bridge_dev, DEVEN, &temp);
+		enabled = !!(temp & DEVEN_MCHBAR_EN);
+	} else {
+		pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp);
+		enabled = temp & 1;
+	}
+
+	/* If it's already enabled, don't have to do anything */
+	if (enabled)
+		return;
+
+	if (intel_alloc_mchbar_resource(dev_priv))
+		return;
+
+	dev_priv->mchbar_need_disable = true;
+
+	/* Space is allocated or reserved, so enable it. */
+	if (IS_I915G(dev_priv) || IS_I915GM(dev_priv)) {
+		pci_write_config_dword(dev_priv->bridge_dev, DEVEN,
+				       temp | DEVEN_MCHBAR_EN);
+	} else {
+		pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp);
+		pci_write_config_dword(dev_priv->bridge_dev, mchbar_reg, temp | 1);
+	}
+}
+
+static void
+intel_teardown_mchbar(struct drm_i915_private *dev_priv)
+{
+	int mchbar_reg = GRAPHICS_VER(dev_priv) >= 4 ? MCHBAR_I965 : MCHBAR_I915;
+
+	if (dev_priv->mchbar_need_disable) {
+		if (IS_I915G(dev_priv) || IS_I915GM(dev_priv)) {
+			u32 deven_val;
+
+			pci_read_config_dword(dev_priv->bridge_dev, DEVEN,
+					      &deven_val);
+			deven_val &= ~DEVEN_MCHBAR_EN;
+			pci_write_config_dword(dev_priv->bridge_dev, DEVEN,
+					       deven_val);
+		} else {
+			u32 mchbar_val;
+
+			pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg,
+					      &mchbar_val);
+			mchbar_val &= ~1;
+			pci_write_config_dword(dev_priv->bridge_dev, mchbar_reg,
+					       mchbar_val);
+		}
+	}
+
+	if (dev_priv->mch_res.start)
+		release_resource(&dev_priv->mch_res);
+}
+
+static int i915_workqueues_init(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * The i915 workqueue is primarily used for batched retirement of
+	 * requests (and thus managing bo) once the task has been completed
+	 * by the GPU. i915_retire_requests() is called directly when we
+	 * need high-priority retirement, such as waiting for an explicit
+	 * bo.
+	 *
+	 * It is also used for periodic low-priority events, such as
+	 * idle-timers and recording error state.
+	 *
+	 * All tasks on the workqueue are expected to acquire the dev mutex
+	 * so there is no point in running more than one instance of the
+	 * workqueue at any time.  Use an ordered one.
+	 */
+	dev_priv->wq = alloc_ordered_workqueue("i915", 0);
+	if (dev_priv->wq == NULL)
+		goto out_err;
+
+	dev_priv->display.hotplug.dp_wq = alloc_ordered_workqueue("i915-dp", 0);
+	if (dev_priv->display.hotplug.dp_wq == NULL)
+		goto out_free_wq;
+
+	return 0;
+
+out_free_wq:
+	destroy_workqueue(dev_priv->wq);
+out_err:
+	drm_err(&dev_priv->drm, "Failed to allocate workqueues.\n");
+
+	return -ENOMEM;
+}
+
+static void i915_workqueues_cleanup(struct drm_i915_private *dev_priv)
+{
+	destroy_workqueue(dev_priv->display.hotplug.dp_wq);
+	destroy_workqueue(dev_priv->wq);
+}
+
+/*
+ * We don't keep the workarounds for pre-production hardware, so we expect our
+ * driver to fail on these machines in one way or another. A little warning on
+ * dmesg may help both the user and the bug triagers.
+ *
+ * Our policy for removing pre-production workarounds is to keep the
+ * current gen workarounds as a guide to the bring-up of the next gen
+ * (workarounds have a habit of persisting!). Anything older than that
+ * should be removed along with the complications they introduce.
+ */
+static void intel_detect_preproduction_hw(struct drm_i915_private *dev_priv)
+{
+	bool pre = false;
+
+	pre |= IS_HSW_EARLY_SDV(dev_priv);
+	pre |= IS_SKYLAKE(dev_priv) && INTEL_REVID(dev_priv) < 0x6;
+	pre |= IS_BROXTON(dev_priv) && INTEL_REVID(dev_priv) < 0xA;
+	pre |= IS_KABYLAKE(dev_priv) && INTEL_REVID(dev_priv) < 0x1;
+	pre |= IS_GEMINILAKE(dev_priv) && INTEL_REVID(dev_priv) < 0x3;
+	pre |= IS_ICELAKE(dev_priv) && INTEL_REVID(dev_priv) < 0x7;
+
+	if (pre) {
+		drm_err(&dev_priv->drm, "This is a pre-production stepping. "
+			  "It may not be fully functional.\n");
+		add_taint(TAINT_MACHINE_CHECK, LOCKDEP_STILL_OK);
+	}
+}
+
+static void sanitize_gpu(struct drm_i915_private *i915)
+{
+	if (!INTEL_INFO(i915)->gpu_reset_clobbers_display) {
+		struct intel_gt *gt;
+		unsigned int i;
+
+		for_each_gt(gt, i915, i)
+			__intel_gt_reset(gt, ALL_ENGINES);
+	}
+}
+
+/**
+ * i915_driver_early_probe - setup state not requiring device access
+ * @dev_priv: device private
+ *
+ * Initialize everything that is a "SW-only" state, that is state not
+ * requiring accessing the device or exposing the driver via kernel internal
+ * or userspace interfaces. Example steps belonging here: lock initialization,
+ * system memory allocation, setting up device specific attributes and
+ * function hooks not requiring accessing the device.
+ */
+static int i915_driver_early_probe(struct drm_i915_private *dev_priv)
+{
+	int ret = 0;
+
+	if (i915_inject_probe_failure(dev_priv))
+		return -ENODEV;
+
+	intel_device_info_subplatform_init(dev_priv);
+	intel_step_init(dev_priv);
+
+	intel_uncore_mmio_debug_init_early(dev_priv);
+
+	spin_lock_init(&dev_priv->irq_lock);
+	spin_lock_init(&dev_priv->gpu_error.lock);
+	mutex_init(&dev_priv->display.backlight.lock);
+
+	mutex_init(&dev_priv->sb_lock);
+	cpu_latency_qos_add_request(&dev_priv->sb_qos, PM_QOS_DEFAULT_VALUE);
+
+	mutex_init(&dev_priv->display.audio.mutex);
+	mutex_init(&dev_priv->display.wm.wm_mutex);
+	mutex_init(&dev_priv->display.pps.mutex);
+	mutex_init(&dev_priv->display.hdcp.comp_mutex);
+	spin_lock_init(&dev_priv->display.dkl.phy_lock);
+
+	i915_memcpy_init_early(dev_priv);
+	intel_runtime_pm_init_early(&dev_priv->runtime_pm);
+
+	ret = i915_workqueues_init(dev_priv);
+	if (ret < 0)
+		return ret;
+
+	ret = vlv_suspend_init(dev_priv);
+	if (ret < 0)
+		goto err_workqueues;
+
+	ret = intel_region_ttm_device_init(dev_priv);
+	if (ret)
+		goto err_ttm;
+
+	intel_wopcm_init_early(&dev_priv->wopcm);
+
+	ret = intel_root_gt_init_early(dev_priv);
+	if (ret < 0)
+		goto err_rootgt;
+
+	i915_drm_clients_init(&dev_priv->clients, dev_priv);
+
+	i915_gem_init_early(dev_priv);
+
+	/* This must be called before any calls to HAS_PCH_* */
+	intel_detect_pch(dev_priv);
+
+	intel_pm_setup(dev_priv);
+	ret = intel_power_domains_init(dev_priv);
+	if (ret < 0)
+		goto err_gem;
+	intel_irq_init(dev_priv);
+	intel_init_display_hooks(dev_priv);
+	intel_init_clock_gating_hooks(dev_priv);
+
+	intel_detect_preproduction_hw(dev_priv);
+
+	return 0;
+
+err_gem:
+	i915_gem_cleanup_early(dev_priv);
+	intel_gt_driver_late_release_all(dev_priv);
+	i915_drm_clients_fini(&dev_priv->clients);
+err_rootgt:
+	intel_region_ttm_device_fini(dev_priv);
+err_ttm:
+	vlv_suspend_cleanup(dev_priv);
+err_workqueues:
+	i915_workqueues_cleanup(dev_priv);
+	return ret;
+}
+
+/**
+ * i915_driver_late_release - cleanup the setup done in
+ *			       i915_driver_early_probe()
+ * @dev_priv: device private
+ */
+static void i915_driver_late_release(struct drm_i915_private *dev_priv)
+{
+	intel_irq_fini(dev_priv);
+	intel_power_domains_cleanup(dev_priv);
+	i915_gem_cleanup_early(dev_priv);
+	intel_gt_driver_late_release_all(dev_priv);
+	i915_drm_clients_fini(&dev_priv->clients);
+	intel_region_ttm_device_fini(dev_priv);
+	vlv_suspend_cleanup(dev_priv);
+	i915_workqueues_cleanup(dev_priv);
+
+	cpu_latency_qos_remove_request(&dev_priv->sb_qos);
+	mutex_destroy(&dev_priv->sb_lock);
+
+	i915_params_free(&dev_priv->params);
+}
+
+/**
+ * i915_driver_mmio_probe - setup device MMIO
+ * @dev_priv: device private
+ *
+ * Setup minimal device state necessary for MMIO accesses later in the
+ * initialization sequence. The setup here should avoid any other device-wide
+ * side effects or exposing the driver via kernel internal or user space
+ * interfaces.
+ */
+static int i915_driver_mmio_probe(struct drm_i915_private *dev_priv)
+{
+	struct intel_gt *gt;
+	int ret, i;
+
+	if (i915_inject_probe_failure(dev_priv))
+		return -ENODEV;
+
+	ret = i915_get_bridge_dev(dev_priv);
+	if (ret < 0)
+		return ret;
+
+	for_each_gt(gt, dev_priv, i) {
+		ret = intel_uncore_init_mmio(gt->uncore);
+		if (ret)
+			return ret;
+
+		ret = drmm_add_action_or_reset(&dev_priv->drm,
+					       intel_uncore_fini_mmio,
+					       gt->uncore);
+		if (ret)
+			return ret;
+	}
+
+	/* Try to make sure MCHBAR is enabled before poking at it */
+	intel_setup_mchbar(dev_priv);
+	intel_device_info_runtime_init(dev_priv);
+
+	for_each_gt(gt, dev_priv, i) {
+		ret = intel_gt_init_mmio(gt);
+		if (ret)
+			goto err_uncore;
+	}
+
+	/* As early as possible, scrub existing GPU state before clobbering */
+	sanitize_gpu(dev_priv);
+
+	return 0;
+
+err_uncore:
+	intel_teardown_mchbar(dev_priv);
+
+	return ret;
+}
+
+/**
+ * i915_driver_mmio_release - cleanup the setup done in i915_driver_mmio_probe()
+ * @dev_priv: device private
+ */
+static void i915_driver_mmio_release(struct drm_i915_private *dev_priv)
+{
+	intel_teardown_mchbar(dev_priv);
+}
+
+/**
+ * i915_set_dma_info - set all relevant PCI dma info as configured for the
+ * platform
+ * @i915: valid i915 instance
+ *
+ * Set the dma max segment size, device and coherent masks.  The dma mask set
+ * needs to occur before i915_ggtt_probe_hw.
+ *
+ * A couple of platforms have special needs.  Address them as well.
+ *
+ */
+static int i915_set_dma_info(struct drm_i915_private *i915)
+{
+	unsigned int mask_size = INTEL_INFO(i915)->dma_mask_size;
+	int ret;
+
+	GEM_BUG_ON(!mask_size);
+
+	/*
+	 * We don't have a max segment size, so set it to the max so sg's
+	 * debugging layer doesn't complain
+	 */
+	dma_set_max_seg_size(i915->drm.dev, UINT_MAX);
+
+	ret = dma_set_mask(i915->drm.dev, DMA_BIT_MASK(mask_size));
+	if (ret)
+		goto mask_err;
+
+	/* overlay on gen2 is broken and can't address above 1G */
+	if (GRAPHICS_VER(i915) == 2)
+		mask_size = 30;
+
+	/*
+	 * 965GM sometimes incorrectly writes to hardware status page (HWS)
+	 * using 32bit addressing, overwriting memory if HWS is located
+	 * above 4GB.
+	 *
+	 * The documentation also mentions an issue with undefined
+	 * behaviour if any general state is accessed within a page above 4GB,
+	 * which also needs to be handled carefully.
+	 */
+	if (IS_I965G(i915) || IS_I965GM(i915))
+		mask_size = 32;
+
+	ret = dma_set_coherent_mask(i915->drm.dev, DMA_BIT_MASK(mask_size));
+	if (ret)
+		goto mask_err;
+
+	return 0;
+
+mask_err:
+	drm_err(&i915->drm, "Can't set DMA mask/consistent mask (%d)\n", ret);
+	return ret;
+}
+
+static int i915_pcode_init(struct drm_i915_private *i915)
+{
+	struct intel_gt *gt;
+	int id, ret;
+
+	for_each_gt(gt, i915, id) {
+		ret = intel_pcode_init(gt->uncore);
+		if (ret) {
+			drm_err(&gt->i915->drm, "gt%d: intel_pcode_init failed %d\n", id, ret);
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * i915_driver_hw_probe - setup state requiring device access
+ * @dev_priv: device private
+ *
+ * Setup state that requires accessing the device, but doesn't require
+ * exposing the driver via kernel internal or userspace interfaces.
+ */
+static int i915_driver_hw_probe(struct drm_i915_private *dev_priv)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	int ret;
+
+	if (i915_inject_probe_failure(dev_priv))
+		return -ENODEV;
+
+	if (HAS_PPGTT(dev_priv)) {
+		if (intel_vgpu_active(dev_priv) &&
+		    !intel_vgpu_has_full_ppgtt(dev_priv)) {
+			i915_report_error(dev_priv,
+					  "incompatible vGPU found, support for isolated ppGTT required\n");
+			return -ENXIO;
+		}
+	}
+
+	if (HAS_EXECLISTS(dev_priv)) {
+		/*
+		 * Older GVT emulation depends upon intercepting CSB mmio,
+		 * which we no longer use, preferring to use the HWSP cache
+		 * instead.
+		 */
+		if (intel_vgpu_active(dev_priv) &&
+		    !intel_vgpu_has_hwsp_emulation(dev_priv)) {
+			i915_report_error(dev_priv,
+					  "old vGPU host found, support for HWSP emulation required\n");
+			return -ENXIO;
+		}
+	}
+
+	/* needs to be done before ggtt probe */
+	intel_dram_edram_detect(dev_priv);
+
+	ret = i915_set_dma_info(dev_priv);
+	if (ret)
+		return ret;
+
+	i915_perf_init(dev_priv);
+
+	ret = intel_gt_assign_ggtt(to_gt(dev_priv));
+	if (ret)
+		goto err_perf;
+
+	ret = i915_ggtt_probe_hw(dev_priv);
+	if (ret)
+		goto err_perf;
+
+	ret = drm_aperture_remove_conflicting_pci_framebuffers(pdev, dev_priv->drm.driver);
+	if (ret)
+		goto err_ggtt;
+
+	ret = i915_ggtt_init_hw(dev_priv);
+	if (ret)
+		goto err_ggtt;
+
+	ret = intel_memory_regions_hw_probe(dev_priv);
+	if (ret)
+		goto err_ggtt;
+
+	ret = intel_gt_tiles_init(dev_priv);
+	if (ret)
+		goto err_mem_regions;
+
+	ret = i915_ggtt_enable_hw(dev_priv);
+	if (ret) {
+		drm_err(&dev_priv->drm, "failed to enable GGTT\n");
+		goto err_mem_regions;
+	}
+
+	pci_set_master(pdev);
+
+	/* On the 945G/GM, the chipset reports the MSI capability on the
+	 * integrated graphics even though the support isn't actually there
+	 * according to the published specs.  It doesn't appear to function
+	 * correctly in testing on 945G.
+	 * This may be a side effect of MSI having been made available for PEG
+	 * and the registers being closely associated.
+	 *
+	 * According to chipset errata, on the 965GM, MSI interrupts may
+	 * be lost or delayed, and was defeatured. MSI interrupts seem to
+	 * get lost on g4x as well, and interrupt delivery seems to stay
+	 * properly dead afterwards. So we'll just disable them for all
+	 * pre-gen5 chipsets.
+	 *
+	 * dp aux and gmbus irq on gen4 seems to be able to generate legacy
+	 * interrupts even when in MSI mode. This results in spurious
+	 * interrupt warnings if the legacy irq no. is shared with another
+	 * device. The kernel then disables that interrupt source and so
+	 * prevents the other device from working properly.
+	 */
+	if (GRAPHICS_VER(dev_priv) >= 5) {
+		if (pci_enable_msi(pdev) < 0)
+			drm_dbg(&dev_priv->drm, "can't enable MSI");
+	}
+
+	ret = intel_gvt_init(dev_priv);
+	if (ret)
+		goto err_msi;
+
+	intel_opregion_setup(dev_priv);
+
+	ret = i915_pcode_init(dev_priv);
+	if (ret)
+		goto err_msi;
+
+	/*
+	 * Fill the dram structure to get the system dram info. This will be
+	 * used for memory latency calculation.
+	 */
+	intel_dram_detect(dev_priv);
+
+	intel_bw_init_hw(dev_priv);
+
+	return 0;
+
+err_msi:
+	if (pdev->msi_enabled)
+		pci_disable_msi(pdev);
+err_mem_regions:
+	intel_memory_regions_driver_release(dev_priv);
+err_ggtt:
+	i915_ggtt_driver_release(dev_priv);
+	i915_gem_drain_freed_objects(dev_priv);
+	i915_ggtt_driver_late_release(dev_priv);
+err_perf:
+	i915_perf_fini(dev_priv);
+	return ret;
+}
+
+/**
+ * i915_driver_hw_remove - cleanup the setup done in i915_driver_hw_probe()
+ * @dev_priv: device private
+ */
+static void i915_driver_hw_remove(struct drm_i915_private *dev_priv)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+
+	i915_perf_fini(dev_priv);
+
+	if (pdev->msi_enabled)
+		pci_disable_msi(pdev);
+}
+
+/**
+ * i915_driver_register - register the driver with the rest of the system
+ * @dev_priv: device private
+ *
+ * Perform any steps necessary to make the driver available via kernel
+ * internal or userspace interfaces.
+ */
+static void i915_driver_register(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_gt *gt;
+	unsigned int i;
+
+	i915_gem_driver_register(dev_priv);
+	i915_pmu_register(dev_priv);
+
+	intel_vgpu_register(dev_priv);
+
+	/* Reveal our presence to userspace */
+	if (drm_dev_register(dev, 0)) {
+		drm_err(&dev_priv->drm,
+			"Failed to register driver for userspace access!\n");
+		return;
+	}
+
+	i915_debugfs_register(dev_priv);
+	i915_setup_sysfs(dev_priv);
+
+	/* Depends on sysfs having been initialized */
+	i915_perf_register(dev_priv);
+
+	for_each_gt(gt, dev_priv, i)
+		intel_gt_driver_register(gt);
+
+	intel_display_driver_register(dev_priv);
+
+	intel_power_domains_enable(dev_priv);
+	intel_runtime_pm_enable(&dev_priv->runtime_pm);
+
+	intel_register_dsm_handler();
+
+	if (i915_switcheroo_register(dev_priv))
+		drm_err(&dev_priv->drm, "Failed to register vga switcheroo!\n");
+}
+
+/**
+ * i915_driver_unregister - cleanup the registration done in i915_driver_regiser()
+ * @dev_priv: device private
+ */
+static void i915_driver_unregister(struct drm_i915_private *dev_priv)
+{
+	struct intel_gt *gt;
+	unsigned int i;
+
+	i915_switcheroo_unregister(dev_priv);
+
+	intel_unregister_dsm_handler();
+
+	intel_runtime_pm_disable(&dev_priv->runtime_pm);
+	intel_power_domains_disable(dev_priv);
+
+	intel_display_driver_unregister(dev_priv);
+
+	for_each_gt(gt, dev_priv, i)
+		intel_gt_driver_unregister(gt);
+
+	i915_perf_unregister(dev_priv);
+	i915_pmu_unregister(dev_priv);
+
+	i915_teardown_sysfs(dev_priv);
+	drm_dev_unplug(&dev_priv->drm);
+
+	i915_gem_driver_unregister(dev_priv);
+}
+
+void
+i915_print_iommu_status(struct drm_i915_private *i915, struct drm_printer *p)
+{
+	drm_printf(p, "iommu: %s\n",
+		   str_enabled_disabled(i915_vtd_active(i915)));
+}
+
+static void i915_welcome_messages(struct drm_i915_private *dev_priv)
+{
+	if (drm_debug_enabled(DRM_UT_DRIVER)) {
+		struct drm_printer p = drm_debug_printer("i915 device info:");
+		struct intel_gt *gt;
+		unsigned int i;
+
+		drm_printf(&p, "pciid=0x%04x rev=0x%02x platform=%s (subplatform=0x%x) gen=%i\n",
+			   INTEL_DEVID(dev_priv),
+			   INTEL_REVID(dev_priv),
+			   intel_platform_name(INTEL_INFO(dev_priv)->platform),
+			   intel_subplatform(RUNTIME_INFO(dev_priv),
+					     INTEL_INFO(dev_priv)->platform),
+			   GRAPHICS_VER(dev_priv));
+
+		intel_device_info_print(INTEL_INFO(dev_priv),
+					RUNTIME_INFO(dev_priv), &p);
+		i915_print_iommu_status(dev_priv, &p);
+		for_each_gt(gt, dev_priv, i)
+			intel_gt_info_print(&gt->info, &p);
+	}
+
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG))
+		drm_info(&dev_priv->drm, "DRM_I915_DEBUG enabled\n");
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		drm_info(&dev_priv->drm, "DRM_I915_DEBUG_GEM enabled\n");
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM))
+		drm_info(&dev_priv->drm,
+			 "DRM_I915_DEBUG_RUNTIME_PM enabled\n");
+}
+
+static struct drm_i915_private *
+i915_driver_create(struct pci_dev *pdev, const struct pci_device_id *ent)
+{
+	const struct intel_device_info *match_info =
+		(struct intel_device_info *)ent->driver_data;
+	struct intel_device_info *device_info;
+	struct intel_runtime_info *runtime;
+	struct drm_i915_private *i915;
+
+	i915 = devm_drm_dev_alloc(&pdev->dev, &i915_drm_driver,
+				  struct drm_i915_private, drm);
+	if (IS_ERR(i915))
+		return i915;
+
+	pci_set_drvdata(pdev, i915);
+
+	/* Device parameters start as a copy of module parameters. */
+	i915_params_copy(&i915->params, &i915_modparams);
+
+	/* Setup the write-once "constant" device info */
+	device_info = mkwrite_device_info(i915);
+	memcpy(device_info, match_info, sizeof(*device_info));
+
+	/* Initialize initial runtime info from static const data and pdev. */
+	runtime = RUNTIME_INFO(i915);
+	memcpy(runtime, &INTEL_INFO(i915)->__runtime, sizeof(*runtime));
+	runtime->device_id = pdev->device;
+
+	return i915;
+}
+
+/**
+ * i915_driver_probe - setup chip and create an initial config
+ * @pdev: PCI device
+ * @ent: matching PCI ID entry
+ *
+ * The driver probe routine has to do several things:
+ *   - drive output discovery via intel_modeset_init()
+ *   - initialize the memory manager
+ *   - allocate initial config memory
+ *   - setup the DRM framebuffer with the allocated memory
+ */
+int i915_driver_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
+{
+	struct drm_i915_private *i915;
+	int ret;
+
+	i915 = i915_driver_create(pdev, ent);
+	if (IS_ERR(i915))
+		return PTR_ERR(i915);
+
+	/* Disable nuclear pageflip by default on pre-ILK */
+	if (!i915->params.nuclear_pageflip && DISPLAY_VER(i915) < 5)
+		i915->drm.driver_features &= ~DRIVER_ATOMIC;
+
+	ret = pci_enable_device(pdev);
+	if (ret)
+		goto out_fini;
+
+	ret = i915_driver_early_probe(i915);
+	if (ret < 0)
+		goto out_pci_disable;
+
+	disable_rpm_wakeref_asserts(&i915->runtime_pm);
+
+	intel_vgpu_detect(i915);
+
+	ret = intel_gt_probe_all(i915);
+	if (ret < 0)
+		goto out_runtime_pm_put;
+
+	ret = i915_driver_mmio_probe(i915);
+	if (ret < 0)
+		goto out_runtime_pm_put;
+
+	ret = i915_driver_hw_probe(i915);
+	if (ret < 0)
+		goto out_cleanup_mmio;
+
+	ret = intel_modeset_init_noirq(i915);
+	if (ret < 0)
+		goto out_cleanup_hw;
+
+	ret = intel_irq_install(i915);
+	if (ret)
+		goto out_cleanup_modeset;
+
+	ret = intel_modeset_init_nogem(i915);
+	if (ret)
+		goto out_cleanup_irq;
+
+	ret = i915_gem_init(i915);
+	if (ret)
+		goto out_cleanup_modeset2;
+
+	ret = intel_modeset_init(i915);
+	if (ret)
+		goto out_cleanup_gem;
+
+	i915_driver_register(i915);
+
+	enable_rpm_wakeref_asserts(&i915->runtime_pm);
+
+	i915_welcome_messages(i915);
+
+	i915->do_release = true;
+
+	return 0;
+
+out_cleanup_gem:
+	i915_gem_suspend(i915);
+	i915_gem_driver_remove(i915);
+	i915_gem_driver_release(i915);
+out_cleanup_modeset2:
+	/* FIXME clean up the error path */
+	intel_modeset_driver_remove(i915);
+	intel_irq_uninstall(i915);
+	intel_modeset_driver_remove_noirq(i915);
+	goto out_cleanup_modeset;
+out_cleanup_irq:
+	intel_irq_uninstall(i915);
+out_cleanup_modeset:
+	intel_modeset_driver_remove_nogem(i915);
+out_cleanup_hw:
+	i915_driver_hw_remove(i915);
+	intel_memory_regions_driver_release(i915);
+	i915_ggtt_driver_release(i915);
+	i915_gem_drain_freed_objects(i915);
+	i915_ggtt_driver_late_release(i915);
+out_cleanup_mmio:
+	i915_driver_mmio_release(i915);
+out_runtime_pm_put:
+	enable_rpm_wakeref_asserts(&i915->runtime_pm);
+	i915_driver_late_release(i915);
+out_pci_disable:
+	pci_disable_device(pdev);
+out_fini:
+	i915_probe_error(i915, "Device initialization failed (%d)\n", ret);
+	return ret;
+}
+
+void i915_driver_remove(struct drm_i915_private *i915)
+{
+	intel_wakeref_t wakeref;
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	i915_driver_unregister(i915);
+
+	/* Flush any external code that still may be under the RCU lock */
+	synchronize_rcu();
+
+	i915_gem_suspend(i915);
+
+	intel_gvt_driver_remove(i915);
+
+	intel_modeset_driver_remove(i915);
+
+	intel_irq_uninstall(i915);
+
+	intel_modeset_driver_remove_noirq(i915);
+
+	i915_reset_error_state(i915);
+	i915_gem_driver_remove(i915);
+
+	intel_modeset_driver_remove_nogem(i915);
+
+	i915_driver_hw_remove(i915);
+
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+}
+
+static void i915_driver_release(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_runtime_pm *rpm = &dev_priv->runtime_pm;
+	intel_wakeref_t wakeref;
+
+	if (!dev_priv->do_release)
+		return;
+
+	wakeref = intel_runtime_pm_get(rpm);
+
+	i915_gem_driver_release(dev_priv);
+
+	intel_memory_regions_driver_release(dev_priv);
+	i915_ggtt_driver_release(dev_priv);
+	i915_gem_drain_freed_objects(dev_priv);
+	i915_ggtt_driver_late_release(dev_priv);
+
+	i915_driver_mmio_release(dev_priv);
+
+	intel_runtime_pm_put(rpm, wakeref);
+
+	intel_runtime_pm_driver_release(rpm);
+
+	i915_driver_late_release(dev_priv);
+}
+
+static int i915_driver_open(struct drm_device *dev, struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	int ret;
+
+	ret = i915_gem_open(i915, file);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+/**
+ * i915_driver_lastclose - clean up after all DRM clients have exited
+ * @dev: DRM device
+ *
+ * Take care of cleaning up after all DRM clients have exited.  In the
+ * mode setting case, we want to restore the kernel's initial mode (just
+ * in case the last client left us in a bad state).
+ *
+ * Additionally, in the non-mode setting case, we'll tear down the GTT
+ * and DMA structures, since the kernel won't be using them, and clea
+ * up any GEM state.
+ */
+static void i915_driver_lastclose(struct drm_device *dev)
+{
+	intel_fbdev_restore_mode(dev);
+
+	vga_switcheroo_process_delayed_switch();
+}
+
+static void i915_driver_postclose(struct drm_device *dev, struct drm_file *file)
+{
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+
+	i915_gem_context_close(file);
+	i915_drm_client_put(file_priv->client);
+
+	kfree_rcu(file_priv, rcu);
+
+	/* Catch up with all the deferred frees from "this" client */
+	i915_gem_flush_free_objects(to_i915(dev));
+}
+
+static void intel_suspend_encoders(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_encoder *encoder;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	drm_modeset_lock_all(dev);
+	for_each_intel_encoder(dev, encoder)
+		if (encoder->suspend)
+			encoder->suspend(encoder);
+	drm_modeset_unlock_all(dev);
+}
+
+static void intel_shutdown_encoders(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	struct intel_encoder *encoder;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	drm_modeset_lock_all(dev);
+	for_each_intel_encoder(dev, encoder)
+		if (encoder->shutdown)
+			encoder->shutdown(encoder);
+	drm_modeset_unlock_all(dev);
+}
+
+void i915_driver_shutdown(struct drm_i915_private *i915)
+{
+	disable_rpm_wakeref_asserts(&i915->runtime_pm);
+	intel_runtime_pm_disable(&i915->runtime_pm);
+	intel_power_domains_disable(i915);
+
+	if (HAS_DISPLAY(i915)) {
+		drm_kms_helper_poll_disable(&i915->drm);
+
+		drm_atomic_helper_shutdown(&i915->drm);
+	}
+
+	intel_dp_mst_suspend(i915);
+
+	intel_runtime_pm_disable_interrupts(i915);
+	intel_hpd_cancel_work(i915);
+
+	intel_suspend_encoders(i915);
+	intel_shutdown_encoders(i915);
+
+	intel_dmc_ucode_suspend(i915);
+
+	i915_gem_suspend(i915);
+
+	/*
+	 * The only requirement is to reboot with display DC states disabled,
+	 * for now leaving all display power wells in the INIT power domain
+	 * enabled.
+	 *
+	 * TODO:
+	 * - unify the pci_driver::shutdown sequence here with the
+	 *   pci_driver.driver.pm.poweroff,poweroff_late sequence.
+	 * - unify the driver remove and system/runtime suspend sequences with
+	 *   the above unified shutdown/poweroff sequence.
+	 */
+	intel_power_domains_driver_remove(i915);
+	enable_rpm_wakeref_asserts(&i915->runtime_pm);
+
+	intel_runtime_pm_driver_release(&i915->runtime_pm);
+}
+
+static bool suspend_to_idle(struct drm_i915_private *dev_priv)
+{
+#if IS_ENABLED(CONFIG_ACPI_SLEEP)
+	if (acpi_target_system_state() < ACPI_STATE_S3)
+		return true;
+#endif
+	return false;
+}
+
+static int i915_drm_prepare(struct drm_device *dev)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+
+	/*
+	 * NB intel_display_suspend() may issue new requests after we've
+	 * ostensibly marked the GPU as ready-to-sleep here. We need to
+	 * split out that work and pull it forward so that after point,
+	 * the GPU is not woken again.
+	 */
+	return i915_gem_backup_suspend(i915);
+}
+
+static int i915_drm_suspend(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	pci_power_t opregion_target_state;
+
+	disable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	/* We do a lot of poking in a lot of registers, make sure they work
+	 * properly. */
+	intel_power_domains_disable(dev_priv);
+	if (HAS_DISPLAY(dev_priv))
+		drm_kms_helper_poll_disable(dev);
+
+	pci_save_state(pdev);
+
+	intel_display_suspend(dev);
+
+	intel_dp_mst_suspend(dev_priv);
+
+	intel_runtime_pm_disable_interrupts(dev_priv);
+	intel_hpd_cancel_work(dev_priv);
+
+	intel_suspend_encoders(dev_priv);
+
+	intel_suspend_hw(dev_priv);
+
+	/* Must be called before GGTT is suspended. */
+	intel_dpt_suspend(dev_priv);
+	i915_ggtt_suspend(to_gt(dev_priv)->ggtt);
+
+	i915_save_display(dev_priv);
+
+	opregion_target_state = suspend_to_idle(dev_priv) ? PCI_D1 : PCI_D3cold;
+	intel_opregion_suspend(dev_priv, opregion_target_state);
+
+	intel_fbdev_set_suspend(dev, FBINFO_STATE_SUSPENDED, true);
+
+	dev_priv->suspend_count++;
+
+	intel_dmc_ucode_suspend(dev_priv);
+
+	enable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	i915_gem_drain_freed_objects(dev_priv);
+
+	return 0;
+}
+
+static enum i915_drm_suspend_mode
+get_suspend_mode(struct drm_i915_private *dev_priv, bool hibernate)
+{
+	if (hibernate)
+		return I915_DRM_SUSPEND_HIBERNATE;
+
+	if (suspend_to_idle(dev_priv))
+		return I915_DRM_SUSPEND_IDLE;
+
+	return I915_DRM_SUSPEND_MEM;
+}
+
+static int i915_drm_suspend_late(struct drm_device *dev, bool hibernation)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	struct intel_runtime_pm *rpm = &dev_priv->runtime_pm;
+	struct intel_gt *gt;
+	int ret, i;
+
+	disable_rpm_wakeref_asserts(rpm);
+
+	i915_gem_suspend_late(dev_priv);
+
+	for_each_gt(gt, dev_priv, i)
+		intel_uncore_suspend(gt->uncore);
+
+	intel_power_domains_suspend(dev_priv,
+				    get_suspend_mode(dev_priv, hibernation));
+
+	intel_display_power_suspend_late(dev_priv);
+
+	ret = vlv_suspend_complete(dev_priv);
+	if (ret) {
+		drm_err(&dev_priv->drm, "Suspend complete failed: %d\n", ret);
+		intel_power_domains_resume(dev_priv);
+
+		goto out;
+	}
+
+	pci_disable_device(pdev);
+	/*
+	 * During hibernation on some platforms the BIOS may try to access
+	 * the device even though it's already in D3 and hang the machine. So
+	 * leave the device in D0 on those platforms and hope the BIOS will
+	 * power down the device properly. The issue was seen on multiple old
+	 * GENs with different BIOS vendors, so having an explicit blacklist
+	 * is inpractical; apply the workaround on everything pre GEN6. The
+	 * platforms where the issue was seen:
+	 * Lenovo Thinkpad X301, X61s, X60, T60, X41
+	 * Fujitsu FSC S7110
+	 * Acer Aspire 1830T
+	 */
+	if (!(hibernation && GRAPHICS_VER(dev_priv) < 6))
+		pci_set_power_state(pdev, PCI_D3hot);
+
+out:
+	enable_rpm_wakeref_asserts(rpm);
+	if (!dev_priv->uncore.user_forcewake_count)
+		intel_runtime_pm_driver_release(rpm);
+
+	return ret;
+}
+
+int i915_driver_suspend_switcheroo(struct drm_i915_private *i915,
+				   pm_message_t state)
+{
+	int error;
+
+	if (drm_WARN_ON_ONCE(&i915->drm, state.event != PM_EVENT_SUSPEND &&
+			     state.event != PM_EVENT_FREEZE))
+		return -EINVAL;
+
+	if (i915->drm.switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
+
+	error = i915_drm_suspend(&i915->drm);
+	if (error)
+		return error;
+
+	return i915_drm_suspend_late(&i915->drm, false);
+}
+
+static int i915_drm_resume(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	int ret;
+
+	disable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	ret = i915_pcode_init(dev_priv);
+	if (ret)
+		return ret;
+
+	sanitize_gpu(dev_priv);
+
+	ret = i915_ggtt_enable_hw(dev_priv);
+	if (ret)
+		drm_err(&dev_priv->drm, "failed to re-enable GGTT\n");
+
+	i915_ggtt_resume(to_gt(dev_priv)->ggtt);
+	/* Must be called after GGTT is resumed. */
+	intel_dpt_resume(dev_priv);
+
+	intel_dmc_ucode_resume(dev_priv);
+
+	i915_restore_display(dev_priv);
+	intel_pps_unlock_regs_wa(dev_priv);
+
+	intel_init_pch_refclk(dev_priv);
+
+	/*
+	 * Interrupts have to be enabled before any batches are run. If not the
+	 * GPU will hang. i915_gem_init_hw() will initiate batches to
+	 * update/restore the context.
+	 *
+	 * drm_mode_config_reset() needs AUX interrupts.
+	 *
+	 * Modeset enabling in intel_modeset_init_hw() also needs working
+	 * interrupts.
+	 */
+	intel_runtime_pm_enable_interrupts(dev_priv);
+
+	if (HAS_DISPLAY(dev_priv))
+		drm_mode_config_reset(dev);
+
+	i915_gem_resume(dev_priv);
+
+	intel_modeset_init_hw(dev_priv);
+	intel_init_clock_gating(dev_priv);
+	intel_hpd_init(dev_priv);
+
+	/* MST sideband requires HPD interrupts enabled */
+	intel_dp_mst_resume(dev_priv);
+	intel_display_resume(dev);
+
+	intel_hpd_poll_disable(dev_priv);
+	if (HAS_DISPLAY(dev_priv))
+		drm_kms_helper_poll_enable(dev);
+
+	intel_opregion_resume(dev_priv);
+
+	intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING, false);
+
+	intel_power_domains_enable(dev_priv);
+
+	intel_gvt_resume(dev_priv);
+
+	enable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	return 0;
+}
+
+static int i915_drm_resume_early(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	struct intel_gt *gt;
+	int ret, i;
+
+	/*
+	 * We have a resume ordering issue with the snd-hda driver also
+	 * requiring our device to be power up. Due to the lack of a
+	 * parent/child relationship we currently solve this with an early
+	 * resume hook.
+	 *
+	 * FIXME: This should be solved with a special hdmi sink device or
+	 * similar so that power domains can be employed.
+	 */
+
+	/*
+	 * Note that we need to set the power state explicitly, since we
+	 * powered off the device during freeze and the PCI core won't power
+	 * it back up for us during thaw. Powering off the device during
+	 * freeze is not a hard requirement though, and during the
+	 * suspend/resume phases the PCI core makes sure we get here with the
+	 * device powered on. So in case we change our freeze logic and keep
+	 * the device powered we can also remove the following set power state
+	 * call.
+	 */
+	ret = pci_set_power_state(pdev, PCI_D0);
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"failed to set PCI D0 power state (%d)\n", ret);
+		return ret;
+	}
+
+	/*
+	 * Note that pci_enable_device() first enables any parent bridge
+	 * device and only then sets the power state for this device. The
+	 * bridge enabling is a nop though, since bridge devices are resumed
+	 * first. The order of enabling power and enabling the device is
+	 * imposed by the PCI core as described above, so here we preserve the
+	 * same order for the freeze/thaw phases.
+	 *
+	 * TODO: eventually we should remove pci_disable_device() /
+	 * pci_enable_enable_device() from suspend/resume. Due to how they
+	 * depend on the device enable refcount we can't anyway depend on them
+	 * disabling/enabling the device.
+	 */
+	if (pci_enable_device(pdev))
+		return -EIO;
+
+	pci_set_master(pdev);
+
+	disable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	ret = vlv_resume_prepare(dev_priv, false);
+	if (ret)
+		drm_err(&dev_priv->drm,
+			"Resume prepare failed: %d, continuing anyway\n", ret);
+
+	for_each_gt(gt, dev_priv, i) {
+		intel_uncore_resume_early(gt->uncore);
+		intel_gt_check_and_clear_faults(gt);
+	}
+
+	intel_display_power_resume_early(dev_priv);
+
+	intel_power_domains_resume(dev_priv);
+
+	enable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	return ret;
+}
+
+int i915_driver_resume_switcheroo(struct drm_i915_private *i915)
+{
+	int ret;
+
+	if (i915->drm.switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
+
+	ret = i915_drm_resume_early(&i915->drm);
+	if (ret)
+		return ret;
+
+	return i915_drm_resume(&i915->drm);
+}
+
+static int i915_pm_prepare(struct device *kdev)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(kdev);
+
+	if (!i915) {
+		dev_err(kdev, "DRM not initialized, aborting suspend.\n");
+		return -ENODEV;
+	}
+
+	if (i915->drm.switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
+
+	return i915_drm_prepare(&i915->drm);
+}
+
+static int i915_pm_suspend(struct device *kdev)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(kdev);
+
+	if (!i915) {
+		dev_err(kdev, "DRM not initialized, aborting suspend.\n");
+		return -ENODEV;
+	}
+
+	i915_ggtt_mark_pte_lost(i915, false);
+
+	if (i915->drm.switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
+
+	return i915_drm_suspend(&i915->drm);
+}
+
+static int i915_pm_suspend_late(struct device *kdev)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(kdev);
+
+	/*
+	 * We have a suspend ordering issue with the snd-hda driver also
+	 * requiring our device to be power up. Due to the lack of a
+	 * parent/child relationship we currently solve this with an late
+	 * suspend hook.
+	 *
+	 * FIXME: This should be solved with a special hdmi sink device or
+	 * similar so that power domains can be employed.
+	 */
+	if (i915->drm.switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
+
+	return i915_drm_suspend_late(&i915->drm, false);
+}
+
+static int i915_pm_poweroff_late(struct device *kdev)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(kdev);
+
+	if (i915->drm.switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
+
+	return i915_drm_suspend_late(&i915->drm, true);
+}
+
+static int i915_pm_resume_early(struct device *kdev)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(kdev);
+
+	if (i915->drm.switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
+
+	return i915_drm_resume_early(&i915->drm);
+}
+
+static int i915_pm_resume(struct device *kdev)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(kdev);
+
+	if (i915->drm.switch_power_state == DRM_SWITCH_POWER_OFF)
+		return 0;
+
+	/*
+	 * If IRST is enabled, or if we can't detect whether it's enabled,
+	 * then we must assume we lost the GGTT page table entries, since
+	 * they are not retained if IRST decided to enter S4.
+	 */
+	if (!IS_ENABLED(CONFIG_ACPI) || acpi_dev_present(irst_name, NULL, -1))
+		i915_ggtt_mark_pte_lost(i915, true);
+
+	return i915_drm_resume(&i915->drm);
+}
+
+/* freeze: before creating the hibernation_image */
+static int i915_pm_freeze(struct device *kdev)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(kdev);
+	int ret;
+
+	if (i915->drm.switch_power_state != DRM_SWITCH_POWER_OFF) {
+		ret = i915_drm_suspend(&i915->drm);
+		if (ret)
+			return ret;
+	}
+
+	ret = i915_gem_freeze(i915);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int i915_pm_freeze_late(struct device *kdev)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(kdev);
+	int ret;
+
+	if (i915->drm.switch_power_state != DRM_SWITCH_POWER_OFF) {
+		ret = i915_drm_suspend_late(&i915->drm, true);
+		if (ret)
+			return ret;
+	}
+
+	ret = i915_gem_freeze_late(i915);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+/* thaw: called after creating the hibernation image, but before turning off. */
+static int i915_pm_thaw_early(struct device *kdev)
+{
+	return i915_pm_resume_early(kdev);
+}
+
+static int i915_pm_thaw(struct device *kdev)
+{
+	return i915_pm_resume(kdev);
+}
+
+/* restore: called after loading the hibernation image. */
+static int i915_pm_restore_early(struct device *kdev)
+{
+	return i915_pm_resume_early(kdev);
+}
+
+static int i915_pm_restore(struct device *kdev)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(kdev);
+
+	i915_ggtt_mark_pte_lost(i915, true);
+	return i915_pm_resume(kdev);
+}
+
+static int intel_runtime_suspend(struct device *kdev)
+{
+	struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
+	struct intel_runtime_pm *rpm = &dev_priv->runtime_pm;
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	struct pci_dev *root_pdev;
+	struct intel_gt *gt;
+	int ret, i;
+
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !HAS_RUNTIME_PM(dev_priv)))
+		return -ENODEV;
+
+	drm_dbg(&dev_priv->drm, "Suspending device\n");
+
+	disable_rpm_wakeref_asserts(rpm);
+
+	/*
+	 * We are safe here against re-faults, since the fault handler takes
+	 * an RPM reference.
+	 */
+	i915_gem_runtime_suspend(dev_priv);
+
+	for_each_gt(gt, dev_priv, i)
+		intel_gt_runtime_suspend(gt);
+
+	intel_runtime_pm_disable_interrupts(dev_priv);
+
+	for_each_gt(gt, dev_priv, i)
+		intel_uncore_suspend(gt->uncore);
+
+	intel_display_power_suspend(dev_priv);
+
+	ret = vlv_suspend_complete(dev_priv);
+	if (ret) {
+		drm_err(&dev_priv->drm,
+			"Runtime suspend failed, disabling it (%d)\n", ret);
+		intel_uncore_runtime_resume(&dev_priv->uncore);
+
+		intel_runtime_pm_enable_interrupts(dev_priv);
+
+		for_each_gt(gt, dev_priv, i)
+			intel_gt_runtime_resume(gt);
+
+		enable_rpm_wakeref_asserts(rpm);
+
+		return ret;
+	}
+
+	enable_rpm_wakeref_asserts(rpm);
+	intel_runtime_pm_driver_release(rpm);
+
+	if (intel_uncore_arm_unclaimed_mmio_detection(&dev_priv->uncore))
+		drm_err(&dev_priv->drm,
+			"Unclaimed access detected prior to suspending\n");
+
+	/*
+	 * FIXME: Temporary hammer to avoid freezing the machine on our DGFX
+	 * This should be totally removed when we handle the pci states properly
+	 * on runtime PM.
+	 */
+	root_pdev = pcie_find_root_port(pdev);
+	if (root_pdev)
+		pci_d3cold_disable(root_pdev);
+
+	rpm->suspended = true;
+
+	/*
+	 * FIXME: We really should find a document that references the arguments
+	 * used below!
+	 */
+	if (IS_BROADWELL(dev_priv)) {
+		/*
+		 * On Broadwell, if we use PCI_D1 the PCH DDI ports will stop
+		 * being detected, and the call we do at intel_runtime_resume()
+		 * won't be able to restore them. Since PCI_D3hot matches the
+		 * actual specification and appears to be working, use it.
+		 */
+		intel_opregion_notify_adapter(dev_priv, PCI_D3hot);
+	} else {
+		/*
+		 * current versions of firmware which depend on this opregion
+		 * notification have repurposed the D1 definition to mean
+		 * "runtime suspended" vs. what you would normally expect (D3)
+		 * to distinguish it from notifications that might be sent via
+		 * the suspend path.
+		 */
+		intel_opregion_notify_adapter(dev_priv, PCI_D1);
+	}
+
+	assert_forcewakes_inactive(&dev_priv->uncore);
+
+	if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv))
+		intel_hpd_poll_enable(dev_priv);
+
+	drm_dbg(&dev_priv->drm, "Device suspended\n");
+	return 0;
+}
+
+static int intel_runtime_resume(struct device *kdev)
+{
+	struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
+	struct intel_runtime_pm *rpm = &dev_priv->runtime_pm;
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	struct pci_dev *root_pdev;
+	struct intel_gt *gt;
+	int ret, i;
+
+	if (drm_WARN_ON_ONCE(&dev_priv->drm, !HAS_RUNTIME_PM(dev_priv)))
+		return -ENODEV;
+
+	drm_dbg(&dev_priv->drm, "Resuming device\n");
+
+	drm_WARN_ON_ONCE(&dev_priv->drm, atomic_read(&rpm->wakeref_count));
+	disable_rpm_wakeref_asserts(rpm);
+
+	intel_opregion_notify_adapter(dev_priv, PCI_D0);
+	rpm->suspended = false;
+
+	root_pdev = pcie_find_root_port(pdev);
+	if (root_pdev)
+		pci_d3cold_enable(root_pdev);
+
+	if (intel_uncore_unclaimed_mmio(&dev_priv->uncore))
+		drm_dbg(&dev_priv->drm,
+			"Unclaimed access during suspend, bios?\n");
+
+	intel_display_power_resume(dev_priv);
+
+	ret = vlv_resume_prepare(dev_priv, true);
+
+	for_each_gt(gt, dev_priv, i)
+		intel_uncore_runtime_resume(gt->uncore);
+
+	intel_runtime_pm_enable_interrupts(dev_priv);
+
+	/*
+	 * No point of rolling back things in case of an error, as the best
+	 * we can do is to hope that things will still work (and disable RPM).
+	 */
+	for_each_gt(gt, dev_priv, i)
+		intel_gt_runtime_resume(gt);
+
+	/*
+	 * On VLV/CHV display interrupts are part of the display
+	 * power well, so hpd is reinitialized from there. For
+	 * everyone else do it here.
+	 */
+	if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv)) {
+		intel_hpd_init(dev_priv);
+		intel_hpd_poll_disable(dev_priv);
+	}
+
+	skl_watermark_ipc_update(dev_priv);
+
+	enable_rpm_wakeref_asserts(rpm);
+
+	if (ret)
+		drm_err(&dev_priv->drm,
+			"Runtime resume failed, disabling it (%d)\n", ret);
+	else
+		drm_dbg(&dev_priv->drm, "Device resumed\n");
+
+	return ret;
+}
+
+const struct dev_pm_ops i915_pm_ops = {
+	/*
+	 * S0ix (via system suspend) and S3 event handlers [PMSG_SUSPEND,
+	 * PMSG_RESUME]
+	 */
+	.prepare = i915_pm_prepare,
+	.suspend = i915_pm_suspend,
+	.suspend_late = i915_pm_suspend_late,
+	.resume_early = i915_pm_resume_early,
+	.resume = i915_pm_resume,
+
+	/*
+	 * S4 event handlers
+	 * @freeze, @freeze_late    : called (1) before creating the
+	 *                            hibernation image [PMSG_FREEZE] and
+	 *                            (2) after rebooting, before restoring
+	 *                            the image [PMSG_QUIESCE]
+	 * @thaw, @thaw_early       : called (1) after creating the hibernation
+	 *                            image, before writing it [PMSG_THAW]
+	 *                            and (2) after failing to create or
+	 *                            restore the image [PMSG_RECOVER]
+	 * @poweroff, @poweroff_late: called after writing the hibernation
+	 *                            image, before rebooting [PMSG_HIBERNATE]
+	 * @restore, @restore_early : called after rebooting and restoring the
+	 *                            hibernation image [PMSG_RESTORE]
+	 */
+	.freeze = i915_pm_freeze,
+	.freeze_late = i915_pm_freeze_late,
+	.thaw_early = i915_pm_thaw_early,
+	.thaw = i915_pm_thaw,
+	.poweroff = i915_pm_suspend,
+	.poweroff_late = i915_pm_poweroff_late,
+	.restore_early = i915_pm_restore_early,
+	.restore = i915_pm_restore,
+
+	/* S0ix (via runtime suspend) event handlers */
+	.runtime_suspend = intel_runtime_suspend,
+	.runtime_resume = intel_runtime_resume,
+};
+
+static const struct file_operations i915_driver_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_open,
+	.release = drm_release_noglobal,
+	.unlocked_ioctl = drm_ioctl,
+	.mmap = i915_gem_mmap,
+	.poll = drm_poll,
+	.read = drm_read,
+	.compat_ioctl = i915_ioc32_compat_ioctl,
+	.llseek = noop_llseek,
+#ifdef CONFIG_PROC_FS
+	.show_fdinfo = i915_drm_client_fdinfo,
+#endif
+};
+
+static int
+i915_gem_reject_pin_ioctl(struct drm_device *dev, void *data,
+			  struct drm_file *file)
+{
+	return -ENODEV;
+}
+
+static const struct drm_ioctl_desc i915_ioctls[] = {
+	DRM_IOCTL_DEF_DRV(I915_INIT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_FLUSH, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_FLIP, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_BATCHBUFFER, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_IRQ_EMIT, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_IRQ_WAIT, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_GETPARAM, i915_getparam_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_SETPARAM, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_ALLOC, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_FREE, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_INIT_HEAP, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_CMDBUFFER, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_DESTROY_HEAP,  drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_SET_VBLANK_PIPE,  drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GET_VBLANK_PIPE,  drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_VBLANK_SWAP, drm_noop, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_HWS_ADDR, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_INIT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER, drm_invalid_op, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER2_WR, i915_gem_execbuffer2_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_PIN, i915_gem_reject_pin_ioctl, DRM_AUTH|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_UNPIN, i915_gem_reject_pin_ioctl, DRM_AUTH|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_BUSY, i915_gem_busy_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_SET_CACHING, i915_gem_set_caching_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_GET_CACHING, i915_gem_get_caching_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_THROTTLE, i915_gem_throttle_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_ENTERVT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_LEAVEVT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CREATE, i915_gem_create_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CREATE_EXT, i915_gem_create_ext_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_PREAD, i915_gem_pread_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_PWRITE, i915_gem_pwrite_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_MMAP, i915_gem_mmap_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_MMAP_OFFSET, i915_gem_mmap_offset_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_SET_DOMAIN, i915_gem_set_domain_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_SW_FINISH, i915_gem_sw_finish_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_SET_TILING, i915_gem_set_tiling_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_GET_TILING, i915_gem_get_tiling_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_GET_APERTURE, i915_gem_get_aperture_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GET_PIPE_FROM_CRTC_ID, intel_get_pipe_from_crtc_id_ioctl, 0),
+	DRM_IOCTL_DEF_DRV(I915_GEM_MADVISE, i915_gem_madvise_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_OVERLAY_PUT_IMAGE, intel_overlay_put_image_ioctl, DRM_MASTER),
+	DRM_IOCTL_DEF_DRV(I915_OVERLAY_ATTRS, intel_overlay_attrs_ioctl, DRM_MASTER),
+	DRM_IOCTL_DEF_DRV(I915_SET_SPRITE_COLORKEY, intel_sprite_set_colorkey_ioctl, DRM_MASTER),
+	DRM_IOCTL_DEF_DRV(I915_GET_SPRITE_COLORKEY, drm_noop, DRM_MASTER),
+	DRM_IOCTL_DEF_DRV(I915_GEM_WAIT, i915_gem_wait_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_CREATE_EXT, i915_gem_context_create_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_DESTROY, i915_gem_context_destroy_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_REG_READ, i915_reg_read_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GET_RESET_STATS, i915_gem_context_reset_stats_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_USERPTR, i915_gem_userptr_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_GETPARAM, i915_gem_context_getparam_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_SETPARAM, i915_gem_context_setparam_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_PERF_OPEN, i915_perf_open_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_PERF_ADD_CONFIG, i915_perf_add_config_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_PERF_REMOVE_CONFIG, i915_perf_remove_config_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_QUERY, i915_query_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_VM_CREATE, i915_gem_vm_create_ioctl, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_VM_DESTROY, i915_gem_vm_destroy_ioctl, DRM_RENDER_ALLOW),
+};
+
+/*
+ * Interface history:
+ *
+ * 1.1: Original.
+ * 1.2: Add Power Management
+ * 1.3: Add vblank support
+ * 1.4: Fix cmdbuffer path, add heap destroy
+ * 1.5: Add vblank pipe configuration
+ * 1.6: - New ioctl for scheduling buffer swaps on vertical blank
+ *      - Support vertical blank on secondary display pipe
+ */
+#define DRIVER_MAJOR		1
+#define DRIVER_MINOR		6
+#define DRIVER_PATCHLEVEL	0
+
+static const struct drm_driver i915_drm_driver = {
+	/* Don't use MTRRs here; the Xserver or userspace app should
+	 * deal with them for Intel hardware.
+	 */
+	.driver_features =
+	    DRIVER_GEM |
+	    DRIVER_RENDER | DRIVER_MODESET | DRIVER_ATOMIC | DRIVER_SYNCOBJ |
+	    DRIVER_SYNCOBJ_TIMELINE,
+	.release = i915_driver_release,
+	.open = i915_driver_open,
+	.lastclose = i915_driver_lastclose,
+	.postclose = i915_driver_postclose,
+
+	.prime_handle_to_fd = drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle = drm_gem_prime_fd_to_handle,
+	.gem_prime_import = i915_gem_prime_import,
+
+	.dumb_create = i915_gem_dumb_create,
+	.dumb_map_offset = i915_gem_dumb_mmap_offset,
+
+	.ioctls = i915_ioctls,
+	.num_ioctls = ARRAY_SIZE(i915_ioctls),
+	.fops = &i915_driver_fops,
+	.name = DRIVER_NAME,
+	.desc = DRIVER_DESC,
+	.date = DRIVER_DATE,
+	.major = DRIVER_MAJOR,
+	.minor = DRIVER_MINOR,
+	.patchlevel = DRIVER_PATCHLEVEL,
+};
diff --git a/drivers/gpu/drm/i915/i915_driver.h b/drivers/gpu/drm/i915/i915_driver.h
new file mode 100644
index 000000000..44ec543d9
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_driver.h
@@ -0,0 +1,33 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_DRIVER_H__
+#define __I915_DRIVER_H__
+
+#include <linux/pm.h>
+
+struct pci_dev;
+struct pci_device_id;
+struct drm_i915_private;
+struct drm_printer;
+
+#define DRIVER_NAME		"i915"
+#define DRIVER_DESC		"Intel Graphics"
+#define DRIVER_DATE		"20201103"
+#define DRIVER_TIMESTAMP	1604406085
+
+extern const struct dev_pm_ops i915_pm_ops;
+
+int i915_driver_probe(struct pci_dev *pdev, const struct pci_device_id *ent);
+void i915_driver_remove(struct drm_i915_private *i915);
+void i915_driver_shutdown(struct drm_i915_private *i915);
+
+int i915_driver_resume_switcheroo(struct drm_i915_private *i915);
+int i915_driver_suspend_switcheroo(struct drm_i915_private *i915, pm_message_t state);
+
+void
+i915_print_iommu_status(struct drm_i915_private *i915, struct drm_printer *p);
+
+#endif /* __I915_DRIVER_H__ */
diff --git a/drivers/gpu/drm/i915/i915_drm_client.c b/drivers/gpu/drm/i915/i915_drm_client.c
new file mode 100644
index 000000000..b09d1d386
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_drm_client.c
@@ -0,0 +1,160 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+
+#include <uapi/drm/i915_drm.h>
+
+#include <drm/drm_print.h>
+
+#include "gem/i915_gem_context.h"
+#include "i915_drm_client.h"
+#include "i915_file_private.h"
+#include "i915_gem.h"
+#include "i915_utils.h"
+
+void i915_drm_clients_init(struct i915_drm_clients *clients,
+			   struct drm_i915_private *i915)
+{
+	clients->i915 = i915;
+	clients->next_id = 0;
+
+	xa_init_flags(&clients->xarray, XA_FLAGS_ALLOC | XA_FLAGS_LOCK_IRQ);
+}
+
+struct i915_drm_client *i915_drm_client_add(struct i915_drm_clients *clients)
+{
+	struct i915_drm_client *client;
+	struct xarray *xa = &clients->xarray;
+	int ret;
+
+	client = kzalloc(sizeof(*client), GFP_KERNEL);
+	if (!client)
+		return ERR_PTR(-ENOMEM);
+
+	xa_lock_irq(xa);
+	ret = __xa_alloc_cyclic(xa, &client->id, client, xa_limit_32b,
+				&clients->next_id, GFP_KERNEL);
+	xa_unlock_irq(xa);
+	if (ret < 0)
+		goto err;
+
+	kref_init(&client->kref);
+	spin_lock_init(&client->ctx_lock);
+	INIT_LIST_HEAD(&client->ctx_list);
+	client->clients = clients;
+
+	return client;
+
+err:
+	kfree(client);
+
+	return ERR_PTR(ret);
+}
+
+void __i915_drm_client_free(struct kref *kref)
+{
+	struct i915_drm_client *client =
+		container_of(kref, typeof(*client), kref);
+	struct xarray *xa = &client->clients->xarray;
+	unsigned long flags;
+
+	xa_lock_irqsave(xa, flags);
+	__xa_erase(xa, client->id);
+	xa_unlock_irqrestore(xa, flags);
+	kfree(client);
+}
+
+void i915_drm_clients_fini(struct i915_drm_clients *clients)
+{
+	GEM_BUG_ON(!xa_empty(&clients->xarray));
+	xa_destroy(&clients->xarray);
+}
+
+#ifdef CONFIG_PROC_FS
+static const char * const uabi_class_names[] = {
+	[I915_ENGINE_CLASS_RENDER] = "render",
+	[I915_ENGINE_CLASS_COPY] = "copy",
+	[I915_ENGINE_CLASS_VIDEO] = "video",
+	[I915_ENGINE_CLASS_VIDEO_ENHANCE] = "video-enhance",
+	[I915_ENGINE_CLASS_COMPUTE] = "compute",
+};
+
+static u64 busy_add(struct i915_gem_context *ctx, unsigned int class)
+{
+	struct i915_gem_engines_iter it;
+	struct intel_context *ce;
+	u64 total = 0;
+
+	for_each_gem_engine(ce, rcu_dereference(ctx->engines), it) {
+		if (ce->engine->uabi_class != class)
+			continue;
+
+		total += intel_context_get_total_runtime_ns(ce);
+	}
+
+	return total;
+}
+
+static void
+show_client_class(struct seq_file *m,
+		  struct i915_drm_client *client,
+		  unsigned int class)
+{
+	const struct list_head *list = &client->ctx_list;
+	u64 total = atomic64_read(&client->past_runtime[class]);
+	const unsigned int capacity =
+		client->clients->i915->engine_uabi_class_count[class];
+	struct i915_gem_context *ctx;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(ctx, list, client_link)
+		total += busy_add(ctx, class);
+	rcu_read_unlock();
+
+	if (capacity)
+		seq_printf(m, "drm-engine-%s:\t%llu ns\n",
+			   uabi_class_names[class], total);
+
+	if (capacity > 1)
+		seq_printf(m, "drm-engine-capacity-%s:\t%u\n",
+			   uabi_class_names[class],
+			   capacity);
+}
+
+void i915_drm_client_fdinfo(struct seq_file *m, struct file *f)
+{
+	struct drm_file *file = f->private_data;
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct drm_i915_private *i915 = file_priv->dev_priv;
+	struct i915_drm_client *client = file_priv->client;
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	unsigned int i;
+
+	/*
+	 * ******************************************************************
+	 * For text output format description please see drm-usage-stats.rst!
+	 * ******************************************************************
+	 */
+
+	seq_printf(m, "drm-driver:\t%s\n", i915->drm.driver->name);
+	seq_printf(m, "drm-pdev:\t%04x:%02x:%02x.%d\n",
+		   pci_domain_nr(pdev->bus), pdev->bus->number,
+		   PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn));
+	seq_printf(m, "drm-client-id:\t%u\n", client->id);
+
+	/*
+	 * Temporarily skip showing client engine information with GuC submission till
+	 * fetching engine busyness is implemented in the GuC submission backend
+	 */
+	if (GRAPHICS_VER(i915) < 8 || intel_uc_uses_guc_submission(&i915->gt0.uc))
+		return;
+
+	for (i = 0; i < ARRAY_SIZE(uabi_class_names); i++)
+		show_client_class(m, client, i);
+}
+#endif
diff --git a/drivers/gpu/drm/i915/i915_drm_client.h b/drivers/gpu/drm/i915/i915_drm_client.h
new file mode 100644
index 000000000..69496af99
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_drm_client.h
@@ -0,0 +1,68 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __I915_DRM_CLIENT_H__
+#define __I915_DRM_CLIENT_H__
+
+#include <linux/kref.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/xarray.h>
+
+#include <uapi/drm/i915_drm.h>
+
+#define I915_LAST_UABI_ENGINE_CLASS I915_ENGINE_CLASS_COMPUTE
+
+struct drm_i915_private;
+
+struct i915_drm_clients {
+	struct drm_i915_private *i915;
+
+	struct xarray xarray;
+	u32 next_id;
+};
+
+struct i915_drm_client {
+	struct kref kref;
+
+	unsigned int id;
+
+	spinlock_t ctx_lock; /* For add/remove from ctx_list. */
+	struct list_head ctx_list; /* List of contexts belonging to client. */
+
+	struct i915_drm_clients *clients;
+
+	/**
+	 * @past_runtime: Accumulation of pphwsp runtimes from closed contexts.
+	 */
+	atomic64_t past_runtime[I915_LAST_UABI_ENGINE_CLASS + 1];
+};
+
+void i915_drm_clients_init(struct i915_drm_clients *clients,
+			   struct drm_i915_private *i915);
+
+static inline struct i915_drm_client *
+i915_drm_client_get(struct i915_drm_client *client)
+{
+	kref_get(&client->kref);
+	return client;
+}
+
+void __i915_drm_client_free(struct kref *kref);
+
+static inline void i915_drm_client_put(struct i915_drm_client *client)
+{
+	kref_put(&client->kref, __i915_drm_client_free);
+}
+
+struct i915_drm_client *i915_drm_client_add(struct i915_drm_clients *clients);
+
+#ifdef CONFIG_PROC_FS
+void i915_drm_client_fdinfo(struct seq_file *m, struct file *f);
+#endif
+
+void i915_drm_clients_fini(struct i915_drm_clients *clients);
+
+#endif /* !__I915_DRM_CLIENT_H__ */
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
new file mode 100644
index 000000000..bdc81db76
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -0,0 +1,998 @@
+/* i915_drv.h -- Private header for the I915 driver -*- linux-c -*-
+ */
+/*
+ *
+ * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _I915_DRV_H_
+#define _I915_DRV_H_
+
+#include <uapi/drm/i915_drm.h>
+
+#include <linux/pm_qos.h>
+
+#include <drm/ttm/ttm_device.h>
+
+#include "display/intel_display.h"
+#include "display/intel_display_core.h"
+
+#include "gem/i915_gem_context_types.h"
+#include "gem/i915_gem_lmem.h"
+#include "gem/i915_gem_shrinker.h"
+#include "gem/i915_gem_stolen.h"
+
+#include "gt/intel_engine.h"
+#include "gt/intel_gt_types.h"
+#include "gt/intel_region_lmem.h"
+#include "gt/intel_workarounds.h"
+#include "gt/uc/intel_uc.h"
+
+#include "i915_drm_client.h"
+#include "i915_gem.h"
+#include "i915_gpu_error.h"
+#include "i915_params.h"
+#include "i915_perf_types.h"
+#include "i915_scheduler.h"
+#include "i915_utils.h"
+#include "intel_device_info.h"
+#include "intel_memory_region.h"
+#include "intel_pch.h"
+#include "intel_runtime_pm.h"
+#include "intel_step.h"
+#include "intel_uncore.h"
+#include "intel_wopcm.h"
+
+struct drm_i915_clock_gating_funcs;
+struct drm_i915_gem_object;
+struct drm_i915_private;
+struct intel_connector;
+struct intel_dp;
+struct intel_encoder;
+struct intel_limit;
+struct intel_overlay_error_state;
+struct vlv_s0ix_state;
+
+/* Threshold == 5 for long IRQs, 50 for short */
+#define HPD_STORM_DEFAULT_THRESHOLD 50
+
+#define I915_GEM_GPU_DOMAINS \
+	(I915_GEM_DOMAIN_RENDER | \
+	 I915_GEM_DOMAIN_SAMPLER | \
+	 I915_GEM_DOMAIN_COMMAND | \
+	 I915_GEM_DOMAIN_INSTRUCTION | \
+	 I915_GEM_DOMAIN_VERTEX)
+
+#define I915_COLOR_UNEVICTABLE (-1) /* a non-vma sharing the address space */
+
+#define GEM_QUIRK_PIN_SWIZZLED_PAGES	BIT(0)
+
+struct i915_suspend_saved_registers {
+	u32 saveDSPARB;
+	u32 saveSWF0[16];
+	u32 saveSWF1[16];
+	u32 saveSWF3[3];
+	u16 saveGCDGMBUS;
+};
+
+#define MAX_L3_SLICES 2
+struct intel_l3_parity {
+	u32 *remap_info[MAX_L3_SLICES];
+	struct work_struct error_work;
+	int which_slice;
+};
+
+struct i915_gem_mm {
+	/*
+	 * Shortcut for the stolen region. This points to either
+	 * INTEL_REGION_STOLEN_SMEM for integrated platforms, or
+	 * INTEL_REGION_STOLEN_LMEM for discrete, or NULL if the device doesn't
+	 * support stolen.
+	 */
+	struct intel_memory_region *stolen_region;
+	/** Memory allocator for GTT stolen memory */
+	struct drm_mm stolen;
+	/** Protects the usage of the GTT stolen memory allocator. This is
+	 * always the inner lock when overlapping with struct_mutex. */
+	struct mutex stolen_lock;
+
+	/* Protects bound_list/unbound_list and #drm_i915_gem_object.mm.link */
+	spinlock_t obj_lock;
+
+	/**
+	 * List of objects which are purgeable.
+	 */
+	struct list_head purge_list;
+
+	/**
+	 * List of objects which have allocated pages and are shrinkable.
+	 */
+	struct list_head shrink_list;
+
+	/**
+	 * List of objects which are pending destruction.
+	 */
+	struct llist_head free_list;
+	struct work_struct free_work;
+	/**
+	 * Count of objects pending destructions. Used to skip needlessly
+	 * waiting on an RCU barrier if no objects are waiting to be freed.
+	 */
+	atomic_t free_count;
+
+	/**
+	 * tmpfs instance used for shmem backed objects
+	 */
+	struct vfsmount *gemfs;
+
+	struct intel_memory_region *regions[INTEL_REGION_UNKNOWN];
+
+	struct notifier_block oom_notifier;
+	struct notifier_block vmap_notifier;
+	struct shrinker shrinker;
+
+#ifdef CONFIG_MMU_NOTIFIER
+	/**
+	 * notifier_lock for mmu notifiers, memory may not be allocated
+	 * while holding this lock.
+	 */
+	rwlock_t notifier_lock;
+#endif
+
+	/* shrinker accounting, also useful for userland debugging */
+	u64 shrink_memory;
+	u32 shrink_count;
+};
+
+#define I915_IDLE_ENGINES_TIMEOUT (200) /* in ms */
+
+unsigned long i915_fence_context_timeout(const struct drm_i915_private *i915,
+					 u64 context);
+
+static inline unsigned long
+i915_fence_timeout(const struct drm_i915_private *i915)
+{
+	return i915_fence_context_timeout(i915, U64_MAX);
+}
+
+#define HAS_HW_SAGV_WM(i915) (DISPLAY_VER(i915) >= 13 && !IS_DGFX(i915))
+
+struct i915_virtual_gpu {
+	struct mutex lock; /* serialises sending of g2v_notify command pkts */
+	bool active;
+	u32 caps;
+	u32 *initial_mmio;
+	u8 *initial_cfg_space;
+	struct list_head entry;
+};
+
+struct i915_selftest_stash {
+	atomic_t counter;
+	struct ida mock_region_instances;
+};
+
+struct drm_i915_private {
+	struct drm_device drm;
+
+	struct intel_display display;
+
+	/* FIXME: Device release actions should all be moved to drmm_ */
+	bool do_release;
+
+	/* i915 device parameters */
+	struct i915_params params;
+
+	const struct intel_device_info __info; /* Use INTEL_INFO() to access. */
+	struct intel_runtime_info __runtime; /* Use RUNTIME_INFO() to access. */
+	struct intel_driver_caps caps;
+
+	/**
+	 * Data Stolen Memory - aka "i915 stolen memory" gives us the start and
+	 * end of stolen which we can optionally use to create GEM objects
+	 * backed by stolen memory. Note that stolen_usable_size tells us
+	 * exactly how much of this we are actually allowed to use, given that
+	 * some portion of it is in fact reserved for use by hardware functions.
+	 */
+	struct resource dsm;
+	/**
+	 * Reseved portion of Data Stolen Memory
+	 */
+	struct resource dsm_reserved;
+
+	/*
+	 * Stolen memory is segmented in hardware with different portions
+	 * offlimits to certain functions.
+	 *
+	 * The drm_mm is initialised to the total accessible range, as found
+	 * from the PCI config. On Broadwell+, this is further restricted to
+	 * avoid the first page! The upper end of stolen memory is reserved for
+	 * hardware functions and similarly removed from the accessible range.
+	 */
+	resource_size_t stolen_usable_size;	/* Total size minus reserved ranges */
+
+	struct intel_uncore uncore;
+	struct intel_uncore_mmio_debug mmio_debug;
+
+	struct i915_virtual_gpu vgpu;
+
+	struct intel_gvt *gvt;
+
+	struct intel_wopcm wopcm;
+
+	struct pci_dev *bridge_dev;
+
+	struct rb_root uabi_engines;
+	unsigned int engine_uabi_class_count[I915_LAST_UABI_ENGINE_CLASS + 1];
+
+	struct resource mch_res;
+
+	/* protects the irq masks */
+	spinlock_t irq_lock;
+
+	bool display_irqs_enabled;
+
+	/* Sideband mailbox protection */
+	struct mutex sb_lock;
+	struct pm_qos_request sb_qos;
+
+	/** Cached value of IMR to avoid reads in updating the bitfield */
+	union {
+		u32 irq_mask;
+		u32 de_irq_mask[I915_MAX_PIPES];
+	};
+	u32 pipestat_irq_mask[I915_MAX_PIPES];
+
+	bool preserve_bios_swizzle;
+
+	unsigned int fsb_freq, mem_freq, is_ddr3;
+	unsigned int skl_preferred_vco_freq;
+
+	unsigned int max_dotclk_freq;
+	unsigned int hpll_freq;
+	unsigned int czclk_freq;
+
+	/**
+	 * wq - Driver workqueue for GEM.
+	 *
+	 * NOTE: Work items scheduled here are not allowed to grab any modeset
+	 * locks, for otherwise the flushing done in the pageflip code will
+	 * result in deadlocks.
+	 */
+	struct workqueue_struct *wq;
+
+	/* pm private clock gating functions */
+	const struct drm_i915_clock_gating_funcs *clock_gating_funcs;
+
+	/* PCH chipset type */
+	enum intel_pch pch_type;
+	unsigned short pch_id;
+
+	unsigned long gem_quirks;
+
+	struct drm_atomic_state *modeset_restore_state;
+	struct drm_modeset_acquire_ctx reset_ctx;
+
+	struct i915_gem_mm mm;
+
+	/* Kernel Modesetting */
+
+	struct list_head global_obj_list;
+
+	bool mchbar_need_disable;
+
+	struct intel_l3_parity l3_parity;
+
+	/*
+	 * HTI (aka HDPORT) state read during initial hw readout.  Most
+	 * platforms don't have HTI, so this will just stay 0.  Those that do
+	 * will use this later to figure out which PLLs and PHYs are unavailable
+	 * for driver usage.
+	 */
+	u32 hti_state;
+
+	/*
+	 * edram size in MB.
+	 * Cannot be determined by PCIID. You must always read a register.
+	 */
+	u32 edram_size_mb;
+
+	struct i915_gpu_error gpu_error;
+
+	/*
+	 * Shadows for CHV DPLL_MD regs to keep the state
+	 * checker somewhat working in the presence hardware
+	 * crappiness (can't read out DPLL_MD for pipes B & C).
+	 */
+	u32 chv_dpll_md[I915_MAX_PIPES];
+	u32 bxt_phy_grc;
+
+	u32 suspend_count;
+	struct i915_suspend_saved_registers regfile;
+	struct vlv_s0ix_state *vlv_s0ix_state;
+
+	struct dram_info {
+		bool wm_lv_0_adjust_needed;
+		u8 num_channels;
+		bool symmetric_memory;
+		enum intel_dram_type {
+			INTEL_DRAM_UNKNOWN,
+			INTEL_DRAM_DDR3,
+			INTEL_DRAM_DDR4,
+			INTEL_DRAM_LPDDR3,
+			INTEL_DRAM_LPDDR4,
+			INTEL_DRAM_DDR5,
+			INTEL_DRAM_LPDDR5,
+		} type;
+		u8 num_qgv_points;
+		u8 num_psf_gv_points;
+	} dram_info;
+
+	struct intel_runtime_pm runtime_pm;
+
+	struct i915_perf perf;
+
+	/* Abstract the submission mechanism (legacy ringbuffer or execlists) away */
+	struct intel_gt gt0;
+
+	/*
+	 * i915->gt[0] == &i915->gt0
+	 */
+#define I915_MAX_GT 4
+	struct intel_gt *gt[I915_MAX_GT];
+
+	struct kobject *sysfs_gt;
+
+	/* Quick lookup of media GT (current platforms only have one) */
+	struct intel_gt *media_gt;
+
+	struct {
+		struct i915_gem_contexts {
+			spinlock_t lock; /* locks list */
+			struct list_head list;
+		} contexts;
+
+		/*
+		 * We replace the local file with a global mappings as the
+		 * backing storage for the mmap is on the device and not
+		 * on the struct file, and we do not want to prolong the
+		 * lifetime of the local fd. To minimise the number of
+		 * anonymous inodes we create, we use a global singleton to
+		 * share the global mapping.
+		 */
+		struct file *mmap_singleton;
+	} gem;
+
+	u8 pch_ssc_use;
+
+	/* For i915gm/i945gm vblank irq workaround */
+	u8 vblank_enabled;
+
+	bool irq_enabled;
+
+	/*
+	 * DG2: Mask of PHYs that were not calibrated by the firmware
+	 * and should not be used.
+	 */
+	u8 snps_phy_failed_calibration;
+
+	struct i915_pmu pmu;
+
+	struct i915_drm_clients clients;
+
+	/* The TTM device structure. */
+	struct ttm_device bdev;
+
+	I915_SELFTEST_DECLARE(struct i915_selftest_stash selftest;)
+
+	/*
+	 * NOTE: This is the dri1/ums dungeon, don't add stuff here. Your patch
+	 * will be rejected. Instead look for a better place.
+	 */
+};
+
+static inline struct drm_i915_private *to_i915(const struct drm_device *dev)
+{
+	return container_of(dev, struct drm_i915_private, drm);
+}
+
+static inline struct drm_i915_private *kdev_to_i915(struct device *kdev)
+{
+	return dev_get_drvdata(kdev);
+}
+
+static inline struct drm_i915_private *pdev_to_i915(struct pci_dev *pdev)
+{
+	return pci_get_drvdata(pdev);
+}
+
+static inline struct intel_gt *to_gt(struct drm_i915_private *i915)
+{
+	return &i915->gt0;
+}
+
+/* Simple iterator over all initialised engines */
+#define for_each_engine(engine__, dev_priv__, id__) \
+	for ((id__) = 0; \
+	     (id__) < I915_NUM_ENGINES; \
+	     (id__)++) \
+		for_each_if ((engine__) = (dev_priv__)->engine[(id__)])
+
+/* Iterator over subset of engines selected by mask */
+#define for_each_engine_masked(engine__, gt__, mask__, tmp__) \
+	for ((tmp__) = (mask__) & (gt__)->info.engine_mask; \
+	     (tmp__) ? \
+	     ((engine__) = (gt__)->engine[__mask_next_bit(tmp__)]), 1 : \
+	     0;)
+
+#define rb_to_uabi_engine(rb) \
+	rb_entry_safe(rb, struct intel_engine_cs, uabi_node)
+
+#define for_each_uabi_engine(engine__, i915__) \
+	for ((engine__) = rb_to_uabi_engine(rb_first(&(i915__)->uabi_engines));\
+	     (engine__); \
+	     (engine__) = rb_to_uabi_engine(rb_next(&(engine__)->uabi_node)))
+
+#define for_each_uabi_class_engine(engine__, class__, i915__) \
+	for ((engine__) = intel_engine_lookup_user((i915__), (class__), 0); \
+	     (engine__) && (engine__)->uabi_class == (class__); \
+	     (engine__) = rb_to_uabi_engine(rb_next(&(engine__)->uabi_node)))
+
+#define INTEL_INFO(dev_priv)	(&(dev_priv)->__info)
+#define RUNTIME_INFO(dev_priv)	(&(dev_priv)->__runtime)
+#define DRIVER_CAPS(dev_priv)	(&(dev_priv)->caps)
+
+#define INTEL_DEVID(dev_priv)	(RUNTIME_INFO(dev_priv)->device_id)
+
+#define IP_VER(ver, rel)		((ver) << 8 | (rel))
+
+#define GRAPHICS_VER(i915)		(RUNTIME_INFO(i915)->graphics.ip.ver)
+#define GRAPHICS_VER_FULL(i915)		IP_VER(RUNTIME_INFO(i915)->graphics.ip.ver, \
+					       RUNTIME_INFO(i915)->graphics.ip.rel)
+#define IS_GRAPHICS_VER(i915, from, until) \
+	(GRAPHICS_VER(i915) >= (from) && GRAPHICS_VER(i915) <= (until))
+
+#define MEDIA_VER(i915)			(RUNTIME_INFO(i915)->media.ip.ver)
+#define MEDIA_VER_FULL(i915)		IP_VER(RUNTIME_INFO(i915)->media.ip.ver, \
+					       RUNTIME_INFO(i915)->media.ip.rel)
+#define IS_MEDIA_VER(i915, from, until) \
+	(MEDIA_VER(i915) >= (from) && MEDIA_VER(i915) <= (until))
+
+#define DISPLAY_VER(i915)	(RUNTIME_INFO(i915)->display.ip.ver)
+#define IS_DISPLAY_VER(i915, from, until) \
+	(DISPLAY_VER(i915) >= (from) && DISPLAY_VER(i915) <= (until))
+
+#define INTEL_REVID(dev_priv)	(to_pci_dev((dev_priv)->drm.dev)->revision)
+
+#define HAS_DSB(dev_priv)	(INTEL_INFO(dev_priv)->display.has_dsb)
+
+#define INTEL_DISPLAY_STEP(__i915) (RUNTIME_INFO(__i915)->step.display_step)
+#define INTEL_GRAPHICS_STEP(__i915) (RUNTIME_INFO(__i915)->step.graphics_step)
+#define INTEL_MEDIA_STEP(__i915) (RUNTIME_INFO(__i915)->step.media_step)
+#define INTEL_BASEDIE_STEP(__i915) (RUNTIME_INFO(__i915)->step.basedie_step)
+
+#define IS_DISPLAY_STEP(__i915, since, until) \
+	(drm_WARN_ON(&(__i915)->drm, INTEL_DISPLAY_STEP(__i915) == STEP_NONE), \
+	 INTEL_DISPLAY_STEP(__i915) >= (since) && INTEL_DISPLAY_STEP(__i915) < (until))
+
+#define IS_GRAPHICS_STEP(__i915, since, until) \
+	(drm_WARN_ON(&(__i915)->drm, INTEL_GRAPHICS_STEP(__i915) == STEP_NONE), \
+	 INTEL_GRAPHICS_STEP(__i915) >= (since) && INTEL_GRAPHICS_STEP(__i915) < (until))
+
+#define IS_MEDIA_STEP(__i915, since, until) \
+	(drm_WARN_ON(&(__i915)->drm, INTEL_MEDIA_STEP(__i915) == STEP_NONE), \
+	 INTEL_MEDIA_STEP(__i915) >= (since) && INTEL_MEDIA_STEP(__i915) < (until))
+
+#define IS_BASEDIE_STEP(__i915, since, until) \
+	(drm_WARN_ON(&(__i915)->drm, INTEL_BASEDIE_STEP(__i915) == STEP_NONE), \
+	 INTEL_BASEDIE_STEP(__i915) >= (since) && INTEL_BASEDIE_STEP(__i915) < (until))
+
+static __always_inline unsigned int
+__platform_mask_index(const struct intel_runtime_info *info,
+		      enum intel_platform p)
+{
+	const unsigned int pbits =
+		BITS_PER_TYPE(info->platform_mask[0]) - INTEL_SUBPLATFORM_BITS;
+
+	/* Expand the platform_mask array if this fails. */
+	BUILD_BUG_ON(INTEL_MAX_PLATFORMS >
+		     pbits * ARRAY_SIZE(info->platform_mask));
+
+	return p / pbits;
+}
+
+static __always_inline unsigned int
+__platform_mask_bit(const struct intel_runtime_info *info,
+		    enum intel_platform p)
+{
+	const unsigned int pbits =
+		BITS_PER_TYPE(info->platform_mask[0]) - INTEL_SUBPLATFORM_BITS;
+
+	return p % pbits + INTEL_SUBPLATFORM_BITS;
+}
+
+static inline u32
+intel_subplatform(const struct intel_runtime_info *info, enum intel_platform p)
+{
+	const unsigned int pi = __platform_mask_index(info, p);
+
+	return info->platform_mask[pi] & INTEL_SUBPLATFORM_MASK;
+}
+
+static __always_inline bool
+IS_PLATFORM(const struct drm_i915_private *i915, enum intel_platform p)
+{
+	const struct intel_runtime_info *info = RUNTIME_INFO(i915);
+	const unsigned int pi = __platform_mask_index(info, p);
+	const unsigned int pb = __platform_mask_bit(info, p);
+
+	BUILD_BUG_ON(!__builtin_constant_p(p));
+
+	return info->platform_mask[pi] & BIT(pb);
+}
+
+static __always_inline bool
+IS_SUBPLATFORM(const struct drm_i915_private *i915,
+	       enum intel_platform p, unsigned int s)
+{
+	const struct intel_runtime_info *info = RUNTIME_INFO(i915);
+	const unsigned int pi = __platform_mask_index(info, p);
+	const unsigned int pb = __platform_mask_bit(info, p);
+	const unsigned int msb = BITS_PER_TYPE(info->platform_mask[0]) - 1;
+	const u32 mask = info->platform_mask[pi];
+
+	BUILD_BUG_ON(!__builtin_constant_p(p));
+	BUILD_BUG_ON(!__builtin_constant_p(s));
+	BUILD_BUG_ON((s) >= INTEL_SUBPLATFORM_BITS);
+
+	/* Shift and test on the MSB position so sign flag can be used. */
+	return ((mask << (msb - pb)) & (mask << (msb - s))) & BIT(msb);
+}
+
+#define IS_MOBILE(dev_priv)	(INTEL_INFO(dev_priv)->is_mobile)
+#define IS_DGFX(dev_priv)   (INTEL_INFO(dev_priv)->is_dgfx)
+
+#define IS_I830(dev_priv)	IS_PLATFORM(dev_priv, INTEL_I830)
+#define IS_I845G(dev_priv)	IS_PLATFORM(dev_priv, INTEL_I845G)
+#define IS_I85X(dev_priv)	IS_PLATFORM(dev_priv, INTEL_I85X)
+#define IS_I865G(dev_priv)	IS_PLATFORM(dev_priv, INTEL_I865G)
+#define IS_I915G(dev_priv)	IS_PLATFORM(dev_priv, INTEL_I915G)
+#define IS_I915GM(dev_priv)	IS_PLATFORM(dev_priv, INTEL_I915GM)
+#define IS_I945G(dev_priv)	IS_PLATFORM(dev_priv, INTEL_I945G)
+#define IS_I945GM(dev_priv)	IS_PLATFORM(dev_priv, INTEL_I945GM)
+#define IS_I965G(dev_priv)	IS_PLATFORM(dev_priv, INTEL_I965G)
+#define IS_I965GM(dev_priv)	IS_PLATFORM(dev_priv, INTEL_I965GM)
+#define IS_G45(dev_priv)	IS_PLATFORM(dev_priv, INTEL_G45)
+#define IS_GM45(dev_priv)	IS_PLATFORM(dev_priv, INTEL_GM45)
+#define IS_G4X(dev_priv)	(IS_G45(dev_priv) || IS_GM45(dev_priv))
+#define IS_PINEVIEW(dev_priv)	IS_PLATFORM(dev_priv, INTEL_PINEVIEW)
+#define IS_G33(dev_priv)	IS_PLATFORM(dev_priv, INTEL_G33)
+#define IS_IRONLAKE(dev_priv)	IS_PLATFORM(dev_priv, INTEL_IRONLAKE)
+#define IS_IRONLAKE_M(dev_priv) \
+	(IS_PLATFORM(dev_priv, INTEL_IRONLAKE) && IS_MOBILE(dev_priv))
+#define IS_SANDYBRIDGE(dev_priv) IS_PLATFORM(dev_priv, INTEL_SANDYBRIDGE)
+#define IS_IVYBRIDGE(dev_priv)	IS_PLATFORM(dev_priv, INTEL_IVYBRIDGE)
+#define IS_IVB_GT1(dev_priv)	(IS_IVYBRIDGE(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 1)
+#define IS_VALLEYVIEW(dev_priv)	IS_PLATFORM(dev_priv, INTEL_VALLEYVIEW)
+#define IS_CHERRYVIEW(dev_priv)	IS_PLATFORM(dev_priv, INTEL_CHERRYVIEW)
+#define IS_HASWELL(dev_priv)	IS_PLATFORM(dev_priv, INTEL_HASWELL)
+#define IS_BROADWELL(dev_priv)	IS_PLATFORM(dev_priv, INTEL_BROADWELL)
+#define IS_SKYLAKE(dev_priv)	IS_PLATFORM(dev_priv, INTEL_SKYLAKE)
+#define IS_BROXTON(dev_priv)	IS_PLATFORM(dev_priv, INTEL_BROXTON)
+#define IS_KABYLAKE(dev_priv)	IS_PLATFORM(dev_priv, INTEL_KABYLAKE)
+#define IS_GEMINILAKE(dev_priv)	IS_PLATFORM(dev_priv, INTEL_GEMINILAKE)
+#define IS_COFFEELAKE(dev_priv)	IS_PLATFORM(dev_priv, INTEL_COFFEELAKE)
+#define IS_COMETLAKE(dev_priv)	IS_PLATFORM(dev_priv, INTEL_COMETLAKE)
+#define IS_ICELAKE(dev_priv)	IS_PLATFORM(dev_priv, INTEL_ICELAKE)
+#define IS_JSL_EHL(dev_priv)	(IS_PLATFORM(dev_priv, INTEL_JASPERLAKE) || \
+				IS_PLATFORM(dev_priv, INTEL_ELKHARTLAKE))
+#define IS_TIGERLAKE(dev_priv)	IS_PLATFORM(dev_priv, INTEL_TIGERLAKE)
+#define IS_ROCKETLAKE(dev_priv)	IS_PLATFORM(dev_priv, INTEL_ROCKETLAKE)
+#define IS_DG1(dev_priv)        IS_PLATFORM(dev_priv, INTEL_DG1)
+#define IS_ALDERLAKE_S(dev_priv) IS_PLATFORM(dev_priv, INTEL_ALDERLAKE_S)
+#define IS_ALDERLAKE_P(dev_priv) IS_PLATFORM(dev_priv, INTEL_ALDERLAKE_P)
+#define IS_XEHPSDV(dev_priv) IS_PLATFORM(dev_priv, INTEL_XEHPSDV)
+#define IS_DG2(dev_priv)	IS_PLATFORM(dev_priv, INTEL_DG2)
+#define IS_PONTEVECCHIO(dev_priv) IS_PLATFORM(dev_priv, INTEL_PONTEVECCHIO)
+#define IS_METEORLAKE(dev_priv) IS_PLATFORM(dev_priv, INTEL_METEORLAKE)
+
+#define IS_METEORLAKE_M(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_METEORLAKE, INTEL_SUBPLATFORM_M)
+#define IS_METEORLAKE_P(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_METEORLAKE, INTEL_SUBPLATFORM_P)
+#define IS_DG2_G10(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_DG2, INTEL_SUBPLATFORM_G10)
+#define IS_DG2_G11(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_DG2, INTEL_SUBPLATFORM_G11)
+#define IS_DG2_G12(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_DG2, INTEL_SUBPLATFORM_G12)
+#define IS_ADLS_RPLS(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_ALDERLAKE_S, INTEL_SUBPLATFORM_RPL)
+#define IS_ADLP_N(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_ALDERLAKE_P, INTEL_SUBPLATFORM_N)
+#define IS_ADLP_RPLP(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_ALDERLAKE_P, INTEL_SUBPLATFORM_RPL)
+#define IS_HSW_EARLY_SDV(dev_priv) (IS_HASWELL(dev_priv) && \
+				    (INTEL_DEVID(dev_priv) & 0xFF00) == 0x0C00)
+#define IS_BDW_ULT(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_BROADWELL, INTEL_SUBPLATFORM_ULT)
+#define IS_BDW_ULX(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_BROADWELL, INTEL_SUBPLATFORM_ULX)
+#define IS_BDW_GT3(dev_priv)	(IS_BROADWELL(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 3)
+#define IS_HSW_ULT(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_HASWELL, INTEL_SUBPLATFORM_ULT)
+#define IS_HSW_GT3(dev_priv)	(IS_HASWELL(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 3)
+#define IS_HSW_GT1(dev_priv)	(IS_HASWELL(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 1)
+/* ULX machines are also considered ULT. */
+#define IS_HSW_ULX(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_HASWELL, INTEL_SUBPLATFORM_ULX)
+#define IS_SKL_ULT(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_SKYLAKE, INTEL_SUBPLATFORM_ULT)
+#define IS_SKL_ULX(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_SKYLAKE, INTEL_SUBPLATFORM_ULX)
+#define IS_KBL_ULT(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_KABYLAKE, INTEL_SUBPLATFORM_ULT)
+#define IS_KBL_ULX(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_KABYLAKE, INTEL_SUBPLATFORM_ULX)
+#define IS_SKL_GT2(dev_priv)	(IS_SKYLAKE(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 2)
+#define IS_SKL_GT3(dev_priv)	(IS_SKYLAKE(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 3)
+#define IS_SKL_GT4(dev_priv)	(IS_SKYLAKE(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 4)
+#define IS_KBL_GT2(dev_priv)	(IS_KABYLAKE(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 2)
+#define IS_KBL_GT3(dev_priv)	(IS_KABYLAKE(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 3)
+#define IS_CFL_ULT(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_COFFEELAKE, INTEL_SUBPLATFORM_ULT)
+#define IS_CFL_ULX(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_COFFEELAKE, INTEL_SUBPLATFORM_ULX)
+#define IS_CFL_GT2(dev_priv)	(IS_COFFEELAKE(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 2)
+#define IS_CFL_GT3(dev_priv)	(IS_COFFEELAKE(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 3)
+
+#define IS_CML_ULT(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_COMETLAKE, INTEL_SUBPLATFORM_ULT)
+#define IS_CML_ULX(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_COMETLAKE, INTEL_SUBPLATFORM_ULX)
+#define IS_CML_GT2(dev_priv)	(IS_COMETLAKE(dev_priv) && \
+				 INTEL_INFO(dev_priv)->gt == 2)
+
+#define IS_ICL_WITH_PORT_F(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_ICELAKE, INTEL_SUBPLATFORM_PORTF)
+
+#define IS_TGL_UY(dev_priv) \
+	IS_SUBPLATFORM(dev_priv, INTEL_TIGERLAKE, INTEL_SUBPLATFORM_UY)
+
+#define IS_SKL_GRAPHICS_STEP(p, since, until) (IS_SKYLAKE(p) && IS_GRAPHICS_STEP(p, since, until))
+
+#define IS_KBL_GRAPHICS_STEP(dev_priv, since, until) \
+	(IS_KABYLAKE(dev_priv) && IS_GRAPHICS_STEP(dev_priv, since, until))
+#define IS_KBL_DISPLAY_STEP(dev_priv, since, until) \
+	(IS_KABYLAKE(dev_priv) && IS_DISPLAY_STEP(dev_priv, since, until))
+
+#define IS_JSL_EHL_GRAPHICS_STEP(p, since, until) \
+	(IS_JSL_EHL(p) && IS_GRAPHICS_STEP(p, since, until))
+#define IS_JSL_EHL_DISPLAY_STEP(p, since, until) \
+	(IS_JSL_EHL(p) && IS_DISPLAY_STEP(p, since, until))
+
+#define IS_TGL_DISPLAY_STEP(__i915, since, until) \
+	(IS_TIGERLAKE(__i915) && \
+	 IS_DISPLAY_STEP(__i915, since, until))
+
+#define IS_TGL_UY_GRAPHICS_STEP(__i915, since, until) \
+	(IS_TGL_UY(__i915) && \
+	 IS_GRAPHICS_STEP(__i915, since, until))
+
+#define IS_TGL_GRAPHICS_STEP(__i915, since, until) \
+	(IS_TIGERLAKE(__i915) && !IS_TGL_UY(__i915)) && \
+	 IS_GRAPHICS_STEP(__i915, since, until))
+
+#define IS_RKL_DISPLAY_STEP(p, since, until) \
+	(IS_ROCKETLAKE(p) && IS_DISPLAY_STEP(p, since, until))
+
+#define IS_DG1_GRAPHICS_STEP(p, since, until) \
+	(IS_DG1(p) && IS_GRAPHICS_STEP(p, since, until))
+#define IS_DG1_DISPLAY_STEP(p, since, until) \
+	(IS_DG1(p) && IS_DISPLAY_STEP(p, since, until))
+
+#define IS_ADLS_DISPLAY_STEP(__i915, since, until) \
+	(IS_ALDERLAKE_S(__i915) && \
+	 IS_DISPLAY_STEP(__i915, since, until))
+
+#define IS_ADLS_GRAPHICS_STEP(__i915, since, until) \
+	(IS_ALDERLAKE_S(__i915) && \
+	 IS_GRAPHICS_STEP(__i915, since, until))
+
+#define IS_ADLP_DISPLAY_STEP(__i915, since, until) \
+	(IS_ALDERLAKE_P(__i915) && \
+	 IS_DISPLAY_STEP(__i915, since, until))
+
+#define IS_ADLP_GRAPHICS_STEP(__i915, since, until) \
+	(IS_ALDERLAKE_P(__i915) && \
+	 IS_GRAPHICS_STEP(__i915, since, until))
+
+#define IS_XEHPSDV_GRAPHICS_STEP(__i915, since, until) \
+	(IS_XEHPSDV(__i915) && IS_GRAPHICS_STEP(__i915, since, until))
+
+/*
+ * DG2 hardware steppings are a bit unusual.  The hardware design was forked to
+ * create three variants (G10, G11, and G12) which each have distinct
+ * workaround sets.  The G11 and G12 forks of the DG2 design reset the GT
+ * stepping back to "A0" for their first iterations, even though they're more
+ * similar to a G10 B0 stepping and G10 C0 stepping respectively in terms of
+ * functionality and workarounds.  However the display stepping does not reset
+ * in the same manner --- a specific stepping like "B0" has a consistent
+ * meaning regardless of whether it belongs to a G10, G11, or G12 DG2.
+ *
+ * TLDR:  All GT workarounds and stepping-specific logic must be applied in
+ * relation to a specific subplatform (G10/G11/G12), whereas display workarounds
+ * and stepping-specific logic will be applied with a general DG2-wide stepping
+ * number.
+ */
+#define IS_DG2_GRAPHICS_STEP(__i915, variant, since, until) \
+	(IS_SUBPLATFORM(__i915, INTEL_DG2, INTEL_SUBPLATFORM_##variant) && \
+	 IS_GRAPHICS_STEP(__i915, since, until))
+
+#define IS_DG2_DISPLAY_STEP(__i915, since, until) \
+	(IS_DG2(__i915) && \
+	 IS_DISPLAY_STEP(__i915, since, until))
+
+#define IS_PVC_BD_STEP(__i915, since, until) \
+	(IS_PONTEVECCHIO(__i915) && \
+	 IS_BASEDIE_STEP(__i915, since, until))
+
+#define IS_PVC_CT_STEP(__i915, since, until) \
+	(IS_PONTEVECCHIO(__i915) && \
+	 IS_GRAPHICS_STEP(__i915, since, until))
+
+#define IS_LP(dev_priv)		(INTEL_INFO(dev_priv)->is_lp)
+#define IS_GEN9_LP(dev_priv)	(GRAPHICS_VER(dev_priv) == 9 && IS_LP(dev_priv))
+#define IS_GEN9_BC(dev_priv)	(GRAPHICS_VER(dev_priv) == 9 && !IS_LP(dev_priv))
+
+#define __HAS_ENGINE(engine_mask, id) ((engine_mask) & BIT(id))
+#define HAS_ENGINE(gt, id) __HAS_ENGINE((gt)->info.engine_mask, id)
+
+#define ENGINE_INSTANCES_MASK(gt, first, count) ({		\
+	unsigned int first__ = (first);					\
+	unsigned int count__ = (count);					\
+	((gt)->info.engine_mask &						\
+	 GENMASK(first__ + count__ - 1, first__)) >> first__;		\
+})
+#define RCS_MASK(gt) \
+	ENGINE_INSTANCES_MASK(gt, RCS0, I915_MAX_RCS)
+#define BCS_MASK(gt) \
+	ENGINE_INSTANCES_MASK(gt, BCS0, I915_MAX_BCS)
+#define VDBOX_MASK(gt) \
+	ENGINE_INSTANCES_MASK(gt, VCS0, I915_MAX_VCS)
+#define VEBOX_MASK(gt) \
+	ENGINE_INSTANCES_MASK(gt, VECS0, I915_MAX_VECS)
+#define CCS_MASK(gt) \
+	ENGINE_INSTANCES_MASK(gt, CCS0, I915_MAX_CCS)
+
+#define HAS_MEDIA_RATIO_MODE(dev_priv) (INTEL_INFO(dev_priv)->has_media_ratio_mode)
+
+/*
+ * The Gen7 cmdparser copies the scanned buffer to the ggtt for execution
+ * All later gens can run the final buffer from the ppgtt
+ */
+#define CMDPARSER_USES_GGTT(dev_priv) (GRAPHICS_VER(dev_priv) == 7)
+
+#define HAS_LLC(dev_priv)	(INTEL_INFO(dev_priv)->has_llc)
+#define HAS_4TILE(dev_priv)	(INTEL_INFO(dev_priv)->has_4tile)
+#define HAS_SNOOP(dev_priv)	(INTEL_INFO(dev_priv)->has_snoop)
+#define HAS_EDRAM(dev_priv)	((dev_priv)->edram_size_mb)
+#define HAS_SECURE_BATCHES(dev_priv) (GRAPHICS_VER(dev_priv) < 6)
+#define HAS_WT(dev_priv)	HAS_EDRAM(dev_priv)
+
+#define HWS_NEEDS_PHYSICAL(dev_priv)	(INTEL_INFO(dev_priv)->hws_needs_physical)
+
+#define HAS_LOGICAL_RING_CONTEXTS(dev_priv) \
+		(INTEL_INFO(dev_priv)->has_logical_ring_contexts)
+#define HAS_LOGICAL_RING_ELSQ(dev_priv) \
+		(INTEL_INFO(dev_priv)->has_logical_ring_elsq)
+
+#define HAS_EXECLISTS(dev_priv) HAS_LOGICAL_RING_CONTEXTS(dev_priv)
+
+#define INTEL_PPGTT(dev_priv) (RUNTIME_INFO(dev_priv)->ppgtt_type)
+#define HAS_PPGTT(dev_priv) \
+	(INTEL_PPGTT(dev_priv) != INTEL_PPGTT_NONE)
+#define HAS_FULL_PPGTT(dev_priv) \
+	(INTEL_PPGTT(dev_priv) >= INTEL_PPGTT_FULL)
+
+#define HAS_PAGE_SIZES(dev_priv, sizes) ({ \
+	GEM_BUG_ON((sizes) == 0); \
+	((sizes) & ~RUNTIME_INFO(dev_priv)->page_sizes) == 0; \
+})
+
+#define HAS_OVERLAY(dev_priv)		 (INTEL_INFO(dev_priv)->display.has_overlay)
+#define OVERLAY_NEEDS_PHYSICAL(dev_priv) \
+		(INTEL_INFO(dev_priv)->display.overlay_needs_physical)
+
+/* Early gen2 have a totally busted CS tlb and require pinned batches. */
+#define HAS_BROKEN_CS_TLB(dev_priv)	(IS_I830(dev_priv) || IS_I845G(dev_priv))
+
+#define NEEDS_RC6_CTX_CORRUPTION_WA(dev_priv)	\
+	(IS_BROADWELL(dev_priv) || GRAPHICS_VER(dev_priv) == 9)
+
+/* WaRsDisableCoarsePowerGating:skl,cnl */
+#define NEEDS_WaRsDisableCoarsePowerGating(dev_priv)			\
+	(IS_SKL_GT3(dev_priv) || IS_SKL_GT4(dev_priv))
+
+#define HAS_GMBUS_IRQ(dev_priv) (DISPLAY_VER(dev_priv) >= 4)
+#define HAS_GMBUS_BURST_READ(dev_priv) (DISPLAY_VER(dev_priv) >= 11 || \
+					IS_GEMINILAKE(dev_priv) || \
+					IS_KABYLAKE(dev_priv))
+
+/* With the 945 and later, Y tiling got adjusted so that it was 32 128-byte
+ * rows, which changed the alignment requirements and fence programming.
+ */
+#define HAS_128_BYTE_Y_TILING(dev_priv) (GRAPHICS_VER(dev_priv) != 2 && \
+					 !(IS_I915G(dev_priv) || IS_I915GM(dev_priv)))
+#define SUPPORTS_TV(dev_priv)		(INTEL_INFO(dev_priv)->display.supports_tv)
+#define I915_HAS_HOTPLUG(dev_priv)	(INTEL_INFO(dev_priv)->display.has_hotplug)
+
+#define HAS_FW_BLC(dev_priv)	(DISPLAY_VER(dev_priv) > 2)
+#define HAS_FBC(dev_priv)	(RUNTIME_INFO(dev_priv)->fbc_mask != 0)
+#define HAS_CUR_FBC(dev_priv)	(!HAS_GMCH(dev_priv) && DISPLAY_VER(dev_priv) >= 7)
+
+#define HAS_IPS(dev_priv)	(IS_HSW_ULT(dev_priv) || IS_BROADWELL(dev_priv))
+
+#define HAS_DP_MST(dev_priv)	(INTEL_INFO(dev_priv)->display.has_dp_mst)
+#define HAS_DP20(dev_priv)	(IS_DG2(dev_priv) || DISPLAY_VER(dev_priv) >= 14)
+
+#define HAS_DOUBLE_BUFFERED_M_N(dev_priv)	(DISPLAY_VER(dev_priv) >= 9 || IS_BROADWELL(dev_priv))
+
+#define HAS_CDCLK_CRAWL(dev_priv)	 (INTEL_INFO(dev_priv)->display.has_cdclk_crawl)
+#define HAS_DDI(dev_priv)		 (INTEL_INFO(dev_priv)->display.has_ddi)
+#define HAS_FPGA_DBG_UNCLAIMED(dev_priv) (INTEL_INFO(dev_priv)->display.has_fpga_dbg)
+#define HAS_PSR(dev_priv)		 (INTEL_INFO(dev_priv)->display.has_psr)
+#define HAS_PSR_HW_TRACKING(dev_priv) \
+	(INTEL_INFO(dev_priv)->display.has_psr_hw_tracking)
+#define HAS_PSR2_SEL_FETCH(dev_priv)	 (DISPLAY_VER(dev_priv) >= 12)
+#define HAS_TRANSCODER(dev_priv, trans)	 ((RUNTIME_INFO(dev_priv)->cpu_transcoder_mask & BIT(trans)) != 0)
+
+#define HAS_RC6(dev_priv)		 (INTEL_INFO(dev_priv)->has_rc6)
+#define HAS_RC6p(dev_priv)		 (INTEL_INFO(dev_priv)->has_rc6p)
+#define HAS_RC6pp(dev_priv)		 (false) /* HW was never validated */
+
+#define HAS_RPS(dev_priv)	(INTEL_INFO(dev_priv)->has_rps)
+
+#define HAS_DMC(dev_priv)	(RUNTIME_INFO(dev_priv)->has_dmc)
+
+#define HAS_HECI_PXP(dev_priv) \
+	(INTEL_INFO(dev_priv)->has_heci_pxp)
+
+#define HAS_HECI_GSCFI(dev_priv) \
+	(INTEL_INFO(dev_priv)->has_heci_gscfi)
+
+#define HAS_HECI_GSC(dev_priv) (HAS_HECI_PXP(dev_priv) || HAS_HECI_GSCFI(dev_priv))
+
+#define HAS_MSO(i915)		(DISPLAY_VER(i915) >= 12)
+
+#define HAS_RUNTIME_PM(dev_priv) (INTEL_INFO(dev_priv)->has_runtime_pm)
+#define HAS_64BIT_RELOC(dev_priv) (INTEL_INFO(dev_priv)->has_64bit_reloc)
+
+/*
+ * Set this flag, when platform requires 64K GTT page sizes or larger for
+ * device local memory access.
+ */
+#define HAS_64K_PAGES(dev_priv) (INTEL_INFO(dev_priv)->has_64k_pages)
+
+/*
+ * Set this flag when platform doesn't allow both 64k pages and 4k pages in
+ * the same PT. this flag means we need to support compact PT layout for the
+ * ppGTT when using the 64K GTT pages.
+ */
+#define NEEDS_COMPACT_PT(dev_priv) (INTEL_INFO(dev_priv)->needs_compact_pt)
+
+#define HAS_IPC(dev_priv)		 (INTEL_INFO(dev_priv)->display.has_ipc)
+
+#define HAS_REGION(i915, i) (RUNTIME_INFO(i915)->memory_regions & (i))
+#define HAS_LMEM(i915) HAS_REGION(i915, REGION_LMEM)
+
+#define HAS_EXTRA_GT_LIST(dev_priv)   (INTEL_INFO(dev_priv)->extra_gt_list)
+
+/*
+ * Platform has the dedicated compression control state for each lmem surfaces
+ * stored in lmem to support the 3D and media compression formats.
+ */
+#define HAS_FLAT_CCS(dev_priv)   (INTEL_INFO(dev_priv)->has_flat_ccs)
+
+#define HAS_GT_UC(dev_priv)	(INTEL_INFO(dev_priv)->has_gt_uc)
+
+#define HAS_POOLED_EU(dev_priv)	(RUNTIME_INFO(dev_priv)->has_pooled_eu)
+
+#define HAS_GLOBAL_MOCS_REGISTERS(dev_priv)	(INTEL_INFO(dev_priv)->has_global_mocs)
+
+#define HAS_PXP(dev_priv)  ((IS_ENABLED(CONFIG_DRM_I915_PXP) && \
+			    INTEL_INFO(dev_priv)->has_pxp) && \
+			    VDBOX_MASK(to_gt(dev_priv)))
+
+#define HAS_GMCH(dev_priv) (INTEL_INFO(dev_priv)->display.has_gmch)
+
+#define HAS_LSPCON(dev_priv) (IS_DISPLAY_VER(dev_priv, 9, 10))
+
+#define HAS_L3_CCS_READ(i915) (INTEL_INFO(i915)->has_l3_ccs_read)
+
+/* DPF == dynamic parity feature */
+#define HAS_L3_DPF(dev_priv) (INTEL_INFO(dev_priv)->has_l3_dpf)
+#define NUM_L3_SLICES(dev_priv) (IS_HSW_GT3(dev_priv) ? \
+				 2 : HAS_L3_DPF(dev_priv))
+
+#define GT_FREQUENCY_MULTIPLIER 50
+#define GEN9_FREQ_SCALER 3
+
+#define INTEL_NUM_PIPES(dev_priv) (hweight8(RUNTIME_INFO(dev_priv)->pipe_mask))
+
+#define HAS_DISPLAY(dev_priv) (RUNTIME_INFO(dev_priv)->pipe_mask != 0)
+
+#define HAS_VRR(i915)	(DISPLAY_VER(i915) >= 11)
+
+#define HAS_ASYNC_FLIPS(i915)		(DISPLAY_VER(i915) >= 5)
+
+/* Only valid when HAS_DISPLAY() is true */
+#define INTEL_DISPLAY_ENABLED(dev_priv) \
+	(drm_WARN_ON(&(dev_priv)->drm, !HAS_DISPLAY(dev_priv)),		\
+	 !(dev_priv)->params.disable_display &&				\
+	 !intel_opregion_headless_sku(dev_priv))
+
+#define HAS_GUC_DEPRIVILEGE(dev_priv) \
+	(INTEL_INFO(dev_priv)->has_guc_deprivilege)
+
+#define HAS_D12_PLANE_MINIMIZATION(dev_priv) (IS_ROCKETLAKE(dev_priv) || \
+					      IS_ALDERLAKE_S(dev_priv))
+
+#define HAS_MBUS_JOINING(i915) (IS_ALDERLAKE_P(i915) || DISPLAY_VER(i915) >= 14)
+
+#define HAS_3D_PIPELINE(i915)	(INTEL_INFO(i915)->has_3d_pipeline)
+
+#define HAS_ONE_EU_PER_FUSE_BIT(i915)	(INTEL_INFO(i915)->has_one_eu_per_fuse_bit)
+
+/* intel_device_info.c */
+static inline struct intel_device_info *
+mkwrite_device_info(struct drm_i915_private *dev_priv)
+{
+	return (struct intel_device_info *)INTEL_INFO(dev_priv);
+}
+
+static inline enum i915_map_type
+i915_coherent_map_type(struct drm_i915_private *i915,
+		       struct drm_i915_gem_object *obj, bool always_coherent)
+{
+	if (i915_gem_object_is_lmem(obj))
+		return I915_MAP_WC;
+	if (HAS_LLC(i915) || always_coherent)
+		return I915_MAP_WB;
+	else
+		return I915_MAP_WC;
+}
+
+#endif
diff --git a/drivers/gpu/drm/i915/i915_file_private.h b/drivers/gpu/drm/i915/i915_file_private.h
new file mode 100644
index 000000000..f42877869
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_file_private.h
@@ -0,0 +1,111 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_FILE_PRIVATE_H__
+#define __I915_FILE_PRIVATE_H__
+
+#include <linux/mutex.h>
+#include <linux/types.h>
+#include <linux/xarray.h>
+
+struct drm_i915_private;
+struct drm_file;
+struct i915_drm_client;
+
+struct drm_i915_file_private {
+	struct drm_i915_private *dev_priv;
+
+	union {
+		struct drm_file *file;
+		struct rcu_head rcu;
+	};
+
+	/** @proto_context_lock: Guards all struct i915_gem_proto_context
+	 * operations
+	 *
+	 * This not only guards @proto_context_xa, but is always held
+	 * whenever we manipulate any struct i915_gem_proto_context,
+	 * including finalizing it on first actual use of the GEM context.
+	 *
+	 * See i915_gem_proto_context.
+	 */
+	struct mutex proto_context_lock;
+
+	/** @proto_context_xa: xarray of struct i915_gem_proto_context
+	 *
+	 * Historically, the context uAPI allowed for two methods of
+	 * setting context parameters: SET_CONTEXT_PARAM and
+	 * CONTEXT_CREATE_EXT_SETPARAM.  The former is allowed to be called
+	 * at any time while the later happens as part of
+	 * GEM_CONTEXT_CREATE.  Everything settable via one was settable
+	 * via the other.  While some params are fairly simple and setting
+	 * them on a live context is harmless such as the context priority,
+	 * others are far trickier such as the VM or the set of engines.
+	 * In order to swap out the VM, for instance, we have to delay
+	 * until all current in-flight work is complete, swap in the new
+	 * VM, and then continue.  This leads to a plethora of potential
+	 * race conditions we'd really rather avoid.
+	 *
+	 * We have since disallowed setting these more complex parameters
+	 * on active contexts.  This works by delaying the creation of the
+	 * actual context until after the client is done configuring it
+	 * with SET_CONTEXT_PARAM.  From the perspective of the client, it
+	 * has the same u32 context ID the whole time.  From the
+	 * perspective of i915, however, it's a struct i915_gem_proto_context
+	 * right up until the point where we attempt to do something which
+	 * the proto-context can't handle.  Then the struct i915_gem_context
+	 * gets created.
+	 *
+	 * This is accomplished via a little xarray dance.  When
+	 * GEM_CONTEXT_CREATE is called, we create a struct
+	 * i915_gem_proto_context, reserve a slot in @context_xa but leave
+	 * it NULL, and place the proto-context in the corresponding slot
+	 * in @proto_context_xa.  Then, in i915_gem_context_lookup(), we
+	 * first check @context_xa.  If it's there, we return the struct
+	 * i915_gem_context and we're done.  If it's not, we look in
+	 * @proto_context_xa and, if we find it there, we create the actual
+	 * context and kill the proto-context.
+	 *
+	 * In order for this dance to work properly, everything which ever
+	 * touches a struct i915_gem_proto_context is guarded by
+	 * @proto_context_lock, including context creation.  Yes, this
+	 * means context creation now takes a giant global lock but it
+	 * can't really be helped and that should never be on any driver's
+	 * fast-path anyway.
+	 */
+	struct xarray proto_context_xa;
+
+	/** @context_xa: xarray of fully created i915_gem_context
+	 *
+	 * Write access to this xarray is guarded by @proto_context_lock.
+	 * Otherwise, writers may race with finalize_create_context_locked().
+	 *
+	 * See @proto_context_xa.
+	 */
+	struct xarray context_xa;
+	struct xarray vm_xa;
+
+	unsigned int bsd_engine;
+
+/*
+ * Every context ban increments per client ban score. Also
+ * hangs in short succession increments ban score. If ban threshold
+ * is reached, client is considered banned and submitting more work
+ * will fail. This is a stop gap measure to limit the badly behaving
+ * clients access to gpu. Note that unbannable contexts never increment
+ * the client ban score.
+ */
+#define I915_CLIENT_SCORE_HANG_FAST	1
+#define   I915_CLIENT_FAST_HANG_JIFFIES (60 * HZ)
+#define I915_CLIENT_SCORE_CONTEXT_BAN   3
+#define I915_CLIENT_SCORE_BANNED	9
+	/** ban_score: Accumulated score of all ctx bans and fast hangs. */
+	atomic_t ban_score;
+	unsigned long hang_timestamp;
+
+	struct i915_drm_client *client;
+};
+
+#endif /* __I915_FILE_PRIVATE_H__ */
diff --git a/drivers/gpu/drm/i915/i915_fixed.h b/drivers/gpu/drm/i915/i915_fixed.h
new file mode 100644
index 000000000..a327094de
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_fixed.h
@@ -0,0 +1,148 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef _I915_FIXED_H_
+#define _I915_FIXED_H_
+
+#include <linux/bug.h>
+#include <linux/kernel.h>
+#include <linux/math64.h>
+#include <linux/types.h>
+
+typedef struct {
+	u32 val;
+} uint_fixed_16_16_t;
+
+#define FP_16_16_MAX ((uint_fixed_16_16_t){ .val = UINT_MAX })
+
+static inline bool is_fixed16_zero(uint_fixed_16_16_t val)
+{
+	return val.val == 0;
+}
+
+static inline uint_fixed_16_16_t u32_to_fixed16(u32 val)
+{
+	uint_fixed_16_16_t fp = { .val = val << 16 };
+
+	WARN_ON(val > U16_MAX);
+
+	return fp;
+}
+
+static inline u32 fixed16_to_u32_round_up(uint_fixed_16_16_t fp)
+{
+	return DIV_ROUND_UP(fp.val, 1 << 16);
+}
+
+static inline u32 fixed16_to_u32(uint_fixed_16_16_t fp)
+{
+	return fp.val >> 16;
+}
+
+static inline uint_fixed_16_16_t min_fixed16(uint_fixed_16_16_t min1,
+					     uint_fixed_16_16_t min2)
+{
+	uint_fixed_16_16_t min = { .val = min(min1.val, min2.val) };
+
+	return min;
+}
+
+static inline uint_fixed_16_16_t max_fixed16(uint_fixed_16_16_t max1,
+					     uint_fixed_16_16_t max2)
+{
+	uint_fixed_16_16_t max = { .val = max(max1.val, max2.val) };
+
+	return max;
+}
+
+static inline uint_fixed_16_16_t clamp_u64_to_fixed16(u64 val)
+{
+	uint_fixed_16_16_t fp = { .val = (u32)val };
+
+	WARN_ON(val > U32_MAX);
+
+	return fp;
+}
+
+static inline u32 div_round_up_fixed16(uint_fixed_16_16_t val,
+				       uint_fixed_16_16_t d)
+{
+	return DIV_ROUND_UP(val.val, d.val);
+}
+
+static inline u32 mul_round_up_u32_fixed16(u32 val, uint_fixed_16_16_t mul)
+{
+	u64 tmp;
+
+	tmp = mul_u32_u32(val, mul.val);
+	tmp = DIV_ROUND_UP_ULL(tmp, 1 << 16);
+	WARN_ON(tmp > U32_MAX);
+
+	return (u32)tmp;
+}
+
+static inline uint_fixed_16_16_t mul_fixed16(uint_fixed_16_16_t val,
+					     uint_fixed_16_16_t mul)
+{
+	u64 tmp;
+
+	tmp = mul_u32_u32(val.val, mul.val);
+	tmp = tmp >> 16;
+
+	return clamp_u64_to_fixed16(tmp);
+}
+
+static inline uint_fixed_16_16_t div_fixed16(u32 val, u32 d)
+{
+	u64 tmp;
+
+	tmp = (u64)val << 16;
+	tmp = DIV_ROUND_UP_ULL(tmp, d);
+
+	return clamp_u64_to_fixed16(tmp);
+}
+
+static inline u32 div_round_up_u32_fixed16(u32 val, uint_fixed_16_16_t d)
+{
+	u64 tmp;
+
+	tmp = (u64)val << 16;
+	tmp = DIV_ROUND_UP_ULL(tmp, d.val);
+	WARN_ON(tmp > U32_MAX);
+
+	return (u32)tmp;
+}
+
+static inline uint_fixed_16_16_t mul_u32_fixed16(u32 val, uint_fixed_16_16_t mul)
+{
+	u64 tmp;
+
+	tmp = mul_u32_u32(val, mul.val);
+
+	return clamp_u64_to_fixed16(tmp);
+}
+
+static inline uint_fixed_16_16_t add_fixed16(uint_fixed_16_16_t add1,
+					     uint_fixed_16_16_t add2)
+{
+	u64 tmp;
+
+	tmp = (u64)add1.val + add2.val;
+
+	return clamp_u64_to_fixed16(tmp);
+}
+
+static inline uint_fixed_16_16_t add_fixed16_u32(uint_fixed_16_16_t add1,
+						 u32 add2)
+{
+	uint_fixed_16_16_t tmp_add2 = u32_to_fixed16(add2);
+	u64 tmp;
+
+	tmp = (u64)add1.val + tmp_add2.val;
+
+	return clamp_u64_to_fixed16(tmp);
+}
+
+#endif /* _I915_FIXED_H_ */
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
new file mode 100644
index 000000000..38c26668b
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -0,0 +1,1328 @@
+/*
+ * Copyright © 2008-2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *
+ */
+
+#include <linux/dma-fence-array.h>
+#include <linux/kthread.h>
+#include <linux/dma-resv.h>
+#include <linux/shmem_fs.h>
+#include <linux/slab.h>
+#include <linux/stop_machine.h>
+#include <linux/swap.h>
+#include <linux/pci.h>
+#include <linux/dma-buf.h>
+#include <linux/mman.h>
+
+#include <drm/drm_cache.h>
+#include <drm/drm_vma_manager.h>
+
+#include "display/intel_display.h"
+#include "display/intel_frontbuffer.h"
+
+#include "gem/i915_gem_clflush.h"
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_ioctls.h"
+#include "gem/i915_gem_mman.h"
+#include "gem/i915_gem_pm.h"
+#include "gem/i915_gem_region.h"
+#include "gem/i915_gem_userptr.h"
+#include "gt/intel_engine_user.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_pm.h"
+#include "gt/intel_workarounds.h"
+
+#include "i915_drv.h"
+#include "i915_file_private.h"
+#include "i915_trace.h"
+#include "i915_vgpu.h"
+#include "intel_pm.h"
+
+static int
+insert_mappable_node(struct i915_ggtt *ggtt, struct drm_mm_node *node, u32 size)
+{
+	int err;
+
+	err = mutex_lock_interruptible(&ggtt->vm.mutex);
+	if (err)
+		return err;
+
+	memset(node, 0, sizeof(*node));
+	err = drm_mm_insert_node_in_range(&ggtt->vm.mm, node,
+					  size, 0, I915_COLOR_UNEVICTABLE,
+					  0, ggtt->mappable_end,
+					  DRM_MM_INSERT_LOW);
+
+	mutex_unlock(&ggtt->vm.mutex);
+
+	return err;
+}
+
+static void
+remove_mappable_node(struct i915_ggtt *ggtt, struct drm_mm_node *node)
+{
+	mutex_lock(&ggtt->vm.mutex);
+	drm_mm_remove_node(node);
+	mutex_unlock(&ggtt->vm.mutex);
+}
+
+int
+i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
+			    struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	struct drm_i915_gem_get_aperture *args = data;
+	struct i915_vma *vma;
+	u64 pinned;
+
+	if (mutex_lock_interruptible(&ggtt->vm.mutex))
+		return -EINTR;
+
+	pinned = ggtt->vm.reserved;
+	list_for_each_entry(vma, &ggtt->vm.bound_list, vm_link)
+		if (i915_vma_is_pinned(vma))
+			pinned += vma->node.size;
+
+	mutex_unlock(&ggtt->vm.mutex);
+
+	args->aper_size = ggtt->vm.total;
+	args->aper_available_size = args->aper_size - pinned;
+
+	return 0;
+}
+
+int i915_gem_object_unbind(struct drm_i915_gem_object *obj,
+			   unsigned long flags)
+{
+	struct intel_runtime_pm *rpm = &to_i915(obj->base.dev)->runtime_pm;
+	bool vm_trylock = !!(flags & I915_GEM_OBJECT_UNBIND_VM_TRYLOCK);
+	LIST_HEAD(still_in_list);
+	intel_wakeref_t wakeref;
+	struct i915_vma *vma;
+	int ret;
+
+	assert_object_held(obj);
+
+	if (list_empty(&obj->vma.list))
+		return 0;
+
+	/*
+	 * As some machines use ACPI to handle runtime-resume callbacks, and
+	 * ACPI is quite kmalloc happy, we cannot resume beneath the vm->mutex
+	 * as they are required by the shrinker. Ergo, we wake the device up
+	 * first just in case.
+	 */
+	wakeref = intel_runtime_pm_get(rpm);
+
+try_again:
+	ret = 0;
+	spin_lock(&obj->vma.lock);
+	while (!ret && (vma = list_first_entry_or_null(&obj->vma.list,
+						       struct i915_vma,
+						       obj_link))) {
+		list_move_tail(&vma->obj_link, &still_in_list);
+		if (!i915_vma_is_bound(vma, I915_VMA_BIND_MASK))
+			continue;
+
+		if (flags & I915_GEM_OBJECT_UNBIND_TEST) {
+			ret = -EBUSY;
+			break;
+		}
+
+		/*
+		 * Requiring the vm destructor to take the object lock
+		 * before destroying a vma would help us eliminate the
+		 * i915_vm_tryget() here, AND thus also the barrier stuff
+		 * at the end. That's an easy fix, but sleeping locks in
+		 * a kthread should generally be avoided.
+		 */
+		ret = -EAGAIN;
+		if (!i915_vm_tryget(vma->vm))
+			break;
+
+		spin_unlock(&obj->vma.lock);
+
+		/*
+		 * Since i915_vma_parked() takes the object lock
+		 * before vma destruction, it won't race us here,
+		 * and destroy the vma from under us.
+		 */
+
+		ret = -EBUSY;
+		if (flags & I915_GEM_OBJECT_UNBIND_ASYNC) {
+			assert_object_held(vma->obj);
+			ret = i915_vma_unbind_async(vma, vm_trylock);
+		}
+
+		if (ret == -EBUSY && (flags & I915_GEM_OBJECT_UNBIND_ACTIVE ||
+				      !i915_vma_is_active(vma))) {
+			if (vm_trylock) {
+				if (mutex_trylock(&vma->vm->mutex)) {
+					ret = __i915_vma_unbind(vma);
+					mutex_unlock(&vma->vm->mutex);
+				}
+			} else {
+				ret = i915_vma_unbind(vma);
+			}
+		}
+
+		i915_vm_put(vma->vm);
+		spin_lock(&obj->vma.lock);
+	}
+	list_splice_init(&still_in_list, &obj->vma.list);
+	spin_unlock(&obj->vma.lock);
+
+	if (ret == -EAGAIN && flags & I915_GEM_OBJECT_UNBIND_BARRIER) {
+		rcu_barrier(); /* flush the i915_vm_release() */
+		goto try_again;
+	}
+
+	intel_runtime_pm_put(rpm, wakeref);
+
+	return ret;
+}
+
+static int
+shmem_pread(struct page *page, int offset, int len, char __user *user_data,
+	    bool needs_clflush)
+{
+	char *vaddr;
+	int ret;
+
+	vaddr = kmap(page);
+
+	if (needs_clflush)
+		drm_clflush_virt_range(vaddr + offset, len);
+
+	ret = __copy_to_user(user_data, vaddr + offset, len);
+
+	kunmap(page);
+
+	return ret ? -EFAULT : 0;
+}
+
+static int
+i915_gem_shmem_pread(struct drm_i915_gem_object *obj,
+		     struct drm_i915_gem_pread *args)
+{
+	unsigned int needs_clflush;
+	unsigned int idx, offset;
+	char __user *user_data;
+	u64 remain;
+	int ret;
+
+	ret = i915_gem_object_lock_interruptible(obj, NULL);
+	if (ret)
+		return ret;
+
+	ret = i915_gem_object_pin_pages(obj);
+	if (ret)
+		goto err_unlock;
+
+	ret = i915_gem_object_prepare_read(obj, &needs_clflush);
+	if (ret)
+		goto err_unpin;
+
+	i915_gem_object_finish_access(obj);
+	i915_gem_object_unlock(obj);
+
+	remain = args->size;
+	user_data = u64_to_user_ptr(args->data_ptr);
+	offset = offset_in_page(args->offset);
+	for (idx = args->offset >> PAGE_SHIFT; remain; idx++) {
+		struct page *page = i915_gem_object_get_page(obj, idx);
+		unsigned int length = min_t(u64, remain, PAGE_SIZE - offset);
+
+		ret = shmem_pread(page, offset, length, user_data,
+				  needs_clflush);
+		if (ret)
+			break;
+
+		remain -= length;
+		user_data += length;
+		offset = 0;
+	}
+
+	i915_gem_object_unpin_pages(obj);
+	return ret;
+
+err_unpin:
+	i915_gem_object_unpin_pages(obj);
+err_unlock:
+	i915_gem_object_unlock(obj);
+	return ret;
+}
+
+static inline bool
+gtt_user_read(struct io_mapping *mapping,
+	      loff_t base, int offset,
+	      char __user *user_data, int length)
+{
+	void __iomem *vaddr;
+	unsigned long unwritten;
+
+	/* We can use the cpu mem copy function because this is X86. */
+	vaddr = io_mapping_map_atomic_wc(mapping, base);
+	unwritten = __copy_to_user_inatomic(user_data,
+					    (void __force *)vaddr + offset,
+					    length);
+	io_mapping_unmap_atomic(vaddr);
+	if (unwritten) {
+		vaddr = io_mapping_map_wc(mapping, base, PAGE_SIZE);
+		unwritten = copy_to_user(user_data,
+					 (void __force *)vaddr + offset,
+					 length);
+		io_mapping_unmap(vaddr);
+	}
+	return unwritten;
+}
+
+static struct i915_vma *i915_gem_gtt_prepare(struct drm_i915_gem_object *obj,
+					     struct drm_mm_node *node,
+					     bool write)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	struct i915_vma *vma;
+	struct i915_gem_ww_ctx ww;
+	int ret;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	vma = ERR_PTR(-ENODEV);
+	ret = i915_gem_object_lock(obj, &ww);
+	if (ret)
+		goto err_ww;
+
+	ret = i915_gem_object_set_to_gtt_domain(obj, write);
+	if (ret)
+		goto err_ww;
+
+	if (!i915_gem_object_is_tiled(obj))
+		vma = i915_gem_object_ggtt_pin_ww(obj, &ww, NULL, 0, 0,
+						  PIN_MAPPABLE |
+						  PIN_NONBLOCK /* NOWARN */ |
+						  PIN_NOEVICT);
+	if (vma == ERR_PTR(-EDEADLK)) {
+		ret = -EDEADLK;
+		goto err_ww;
+	} else if (!IS_ERR(vma)) {
+		node->start = i915_ggtt_offset(vma);
+		node->flags = 0;
+	} else {
+		ret = insert_mappable_node(ggtt, node, PAGE_SIZE);
+		if (ret)
+			goto err_ww;
+		GEM_BUG_ON(!drm_mm_node_allocated(node));
+		vma = NULL;
+	}
+
+	ret = i915_gem_object_pin_pages(obj);
+	if (ret) {
+		if (drm_mm_node_allocated(node)) {
+			ggtt->vm.clear_range(&ggtt->vm, node->start, node->size);
+			remove_mappable_node(ggtt, node);
+		} else {
+			i915_vma_unpin(vma);
+		}
+	}
+
+err_ww:
+	if (ret == -EDEADLK) {
+		ret = i915_gem_ww_ctx_backoff(&ww);
+		if (!ret)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+
+	return ret ? ERR_PTR(ret) : vma;
+}
+
+static void i915_gem_gtt_cleanup(struct drm_i915_gem_object *obj,
+				 struct drm_mm_node *node,
+				 struct i915_vma *vma)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+
+	i915_gem_object_unpin_pages(obj);
+	if (drm_mm_node_allocated(node)) {
+		ggtt->vm.clear_range(&ggtt->vm, node->start, node->size);
+		remove_mappable_node(ggtt, node);
+	} else {
+		i915_vma_unpin(vma);
+	}
+}
+
+static int
+i915_gem_gtt_pread(struct drm_i915_gem_object *obj,
+		   const struct drm_i915_gem_pread *args)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	intel_wakeref_t wakeref;
+	struct drm_mm_node node;
+	void __user *user_data;
+	struct i915_vma *vma;
+	u64 remain, offset;
+	int ret = 0;
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	vma = i915_gem_gtt_prepare(obj, &node, false);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto out_rpm;
+	}
+
+	user_data = u64_to_user_ptr(args->data_ptr);
+	remain = args->size;
+	offset = args->offset;
+
+	while (remain > 0) {
+		/* Operation in this page
+		 *
+		 * page_base = page offset within aperture
+		 * page_offset = offset within page
+		 * page_length = bytes to copy for this page
+		 */
+		u32 page_base = node.start;
+		unsigned page_offset = offset_in_page(offset);
+		unsigned page_length = PAGE_SIZE - page_offset;
+		page_length = remain < page_length ? remain : page_length;
+		if (drm_mm_node_allocated(&node)) {
+			ggtt->vm.insert_page(&ggtt->vm,
+					     i915_gem_object_get_dma_address(obj, offset >> PAGE_SHIFT),
+					     node.start, I915_CACHE_NONE, 0);
+		} else {
+			page_base += offset & PAGE_MASK;
+		}
+
+		if (gtt_user_read(&ggtt->iomap, page_base, page_offset,
+				  user_data, page_length)) {
+			ret = -EFAULT;
+			break;
+		}
+
+		remain -= page_length;
+		user_data += page_length;
+		offset += page_length;
+	}
+
+	i915_gem_gtt_cleanup(obj, &node, vma);
+out_rpm:
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+	return ret;
+}
+
+/**
+ * Reads data from the object referenced by handle.
+ * @dev: drm device pointer
+ * @data: ioctl data blob
+ * @file: drm file pointer
+ *
+ * On error, the contents of *data are undefined.
+ */
+int
+i915_gem_pread_ioctl(struct drm_device *dev, void *data,
+		     struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_pread *args = data;
+	struct drm_i915_gem_object *obj;
+	int ret;
+
+	/* PREAD is disallowed for all platforms after TGL-LP.  This also
+	 * covers all platforms with local memory.
+	 */
+	if (GRAPHICS_VER(i915) >= 12 && !IS_TIGERLAKE(i915))
+		return -EOPNOTSUPP;
+
+	if (args->size == 0)
+		return 0;
+
+	if (!access_ok(u64_to_user_ptr(args->data_ptr),
+		       args->size))
+		return -EFAULT;
+
+	obj = i915_gem_object_lookup(file, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	/* Bounds check source.  */
+	if (range_overflows_t(u64, args->offset, args->size, obj->base.size)) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	trace_i915_gem_object_pread(obj, args->offset, args->size);
+	ret = -ENODEV;
+	if (obj->ops->pread)
+		ret = obj->ops->pread(obj, args);
+	if (ret != -ENODEV)
+		goto out;
+
+	ret = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE,
+				   MAX_SCHEDULE_TIMEOUT);
+	if (ret)
+		goto out;
+
+	ret = i915_gem_shmem_pread(obj, args);
+	if (ret == -EFAULT || ret == -ENODEV)
+		ret = i915_gem_gtt_pread(obj, args);
+
+out:
+	i915_gem_object_put(obj);
+	return ret;
+}
+
+/* This is the fast write path which cannot handle
+ * page faults in the source data
+ */
+
+static inline bool
+ggtt_write(struct io_mapping *mapping,
+	   loff_t base, int offset,
+	   char __user *user_data, int length)
+{
+	void __iomem *vaddr;
+	unsigned long unwritten;
+
+	/* We can use the cpu mem copy function because this is X86. */
+	vaddr = io_mapping_map_atomic_wc(mapping, base);
+	unwritten = __copy_from_user_inatomic_nocache((void __force *)vaddr + offset,
+						      user_data, length);
+	io_mapping_unmap_atomic(vaddr);
+	if (unwritten) {
+		vaddr = io_mapping_map_wc(mapping, base, PAGE_SIZE);
+		unwritten = copy_from_user((void __force *)vaddr + offset,
+					   user_data, length);
+		io_mapping_unmap(vaddr);
+	}
+
+	return unwritten;
+}
+
+/**
+ * This is the fast pwrite path, where we copy the data directly from the
+ * user into the GTT, uncached.
+ * @obj: i915 GEM object
+ * @args: pwrite arguments structure
+ */
+static int
+i915_gem_gtt_pwrite_fast(struct drm_i915_gem_object *obj,
+			 const struct drm_i915_gem_pwrite *args)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	struct intel_runtime_pm *rpm = &i915->runtime_pm;
+	intel_wakeref_t wakeref;
+	struct drm_mm_node node;
+	struct i915_vma *vma;
+	u64 remain, offset;
+	void __user *user_data;
+	int ret = 0;
+
+	if (i915_gem_object_has_struct_page(obj)) {
+		/*
+		 * Avoid waking the device up if we can fallback, as
+		 * waking/resuming is very slow (worst-case 10-100 ms
+		 * depending on PCI sleeps and our own resume time).
+		 * This easily dwarfs any performance advantage from
+		 * using the cache bypass of indirect GGTT access.
+		 */
+		wakeref = intel_runtime_pm_get_if_in_use(rpm);
+		if (!wakeref)
+			return -EFAULT;
+	} else {
+		/* No backing pages, no fallback, we must force GGTT access */
+		wakeref = intel_runtime_pm_get(rpm);
+	}
+
+	vma = i915_gem_gtt_prepare(obj, &node, true);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto out_rpm;
+	}
+
+	i915_gem_object_invalidate_frontbuffer(obj, ORIGIN_CPU);
+
+	user_data = u64_to_user_ptr(args->data_ptr);
+	offset = args->offset;
+	remain = args->size;
+	while (remain) {
+		/* Operation in this page
+		 *
+		 * page_base = page offset within aperture
+		 * page_offset = offset within page
+		 * page_length = bytes to copy for this page
+		 */
+		u32 page_base = node.start;
+		unsigned int page_offset = offset_in_page(offset);
+		unsigned int page_length = PAGE_SIZE - page_offset;
+		page_length = remain < page_length ? remain : page_length;
+		if (drm_mm_node_allocated(&node)) {
+			/* flush the write before we modify the GGTT */
+			intel_gt_flush_ggtt_writes(ggtt->vm.gt);
+			ggtt->vm.insert_page(&ggtt->vm,
+					     i915_gem_object_get_dma_address(obj, offset >> PAGE_SHIFT),
+					     node.start, I915_CACHE_NONE, 0);
+			wmb(); /* flush modifications to the GGTT (insert_page) */
+		} else {
+			page_base += offset & PAGE_MASK;
+		}
+		/* If we get a fault while copying data, then (presumably) our
+		 * source page isn't available.  Return the error and we'll
+		 * retry in the slow path.
+		 * If the object is non-shmem backed, we retry again with the
+		 * path that handles page fault.
+		 */
+		if (ggtt_write(&ggtt->iomap, page_base, page_offset,
+			       user_data, page_length)) {
+			ret = -EFAULT;
+			break;
+		}
+
+		remain -= page_length;
+		user_data += page_length;
+		offset += page_length;
+	}
+
+	intel_gt_flush_ggtt_writes(ggtt->vm.gt);
+	i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
+
+	i915_gem_gtt_cleanup(obj, &node, vma);
+out_rpm:
+	intel_runtime_pm_put(rpm, wakeref);
+	return ret;
+}
+
+/* Per-page copy function for the shmem pwrite fastpath.
+ * Flushes invalid cachelines before writing to the target if
+ * needs_clflush_before is set and flushes out any written cachelines after
+ * writing if needs_clflush is set.
+ */
+static int
+shmem_pwrite(struct page *page, int offset, int len, char __user *user_data,
+	     bool needs_clflush_before,
+	     bool needs_clflush_after)
+{
+	char *vaddr;
+	int ret;
+
+	vaddr = kmap(page);
+
+	if (needs_clflush_before)
+		drm_clflush_virt_range(vaddr + offset, len);
+
+	ret = __copy_from_user(vaddr + offset, user_data, len);
+	if (!ret && needs_clflush_after)
+		drm_clflush_virt_range(vaddr + offset, len);
+
+	kunmap(page);
+
+	return ret ? -EFAULT : 0;
+}
+
+static int
+i915_gem_shmem_pwrite(struct drm_i915_gem_object *obj,
+		      const struct drm_i915_gem_pwrite *args)
+{
+	unsigned int partial_cacheline_write;
+	unsigned int needs_clflush;
+	unsigned int offset, idx;
+	void __user *user_data;
+	u64 remain;
+	int ret;
+
+	ret = i915_gem_object_lock_interruptible(obj, NULL);
+	if (ret)
+		return ret;
+
+	ret = i915_gem_object_pin_pages(obj);
+	if (ret)
+		goto err_unlock;
+
+	ret = i915_gem_object_prepare_write(obj, &needs_clflush);
+	if (ret)
+		goto err_unpin;
+
+	i915_gem_object_finish_access(obj);
+	i915_gem_object_unlock(obj);
+
+	/* If we don't overwrite a cacheline completely we need to be
+	 * careful to have up-to-date data by first clflushing. Don't
+	 * overcomplicate things and flush the entire patch.
+	 */
+	partial_cacheline_write = 0;
+	if (needs_clflush & CLFLUSH_BEFORE)
+		partial_cacheline_write = boot_cpu_data.x86_clflush_size - 1;
+
+	user_data = u64_to_user_ptr(args->data_ptr);
+	remain = args->size;
+	offset = offset_in_page(args->offset);
+	for (idx = args->offset >> PAGE_SHIFT; remain; idx++) {
+		struct page *page = i915_gem_object_get_page(obj, idx);
+		unsigned int length = min_t(u64, remain, PAGE_SIZE - offset);
+
+		ret = shmem_pwrite(page, offset, length, user_data,
+				   (offset | length) & partial_cacheline_write,
+				   needs_clflush & CLFLUSH_AFTER);
+		if (ret)
+			break;
+
+		remain -= length;
+		user_data += length;
+		offset = 0;
+	}
+
+	i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
+
+	i915_gem_object_unpin_pages(obj);
+	return ret;
+
+err_unpin:
+	i915_gem_object_unpin_pages(obj);
+err_unlock:
+	i915_gem_object_unlock(obj);
+	return ret;
+}
+
+/**
+ * Writes data to the object referenced by handle.
+ * @dev: drm device
+ * @data: ioctl data blob
+ * @file: drm file
+ *
+ * On error, the contents of the buffer that were to be modified are undefined.
+ */
+int
+i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
+		      struct drm_file *file)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_pwrite *args = data;
+	struct drm_i915_gem_object *obj;
+	int ret;
+
+	/* PWRITE is disallowed for all platforms after TGL-LP.  This also
+	 * covers all platforms with local memory.
+	 */
+	if (GRAPHICS_VER(i915) >= 12 && !IS_TIGERLAKE(i915))
+		return -EOPNOTSUPP;
+
+	if (args->size == 0)
+		return 0;
+
+	if (!access_ok(u64_to_user_ptr(args->data_ptr), args->size))
+		return -EFAULT;
+
+	obj = i915_gem_object_lookup(file, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	/* Bounds check destination. */
+	if (range_overflows_t(u64, args->offset, args->size, obj->base.size)) {
+		ret = -EINVAL;
+		goto err;
+	}
+
+	/* Writes not allowed into this read-only object */
+	if (i915_gem_object_is_readonly(obj)) {
+		ret = -EINVAL;
+		goto err;
+	}
+
+	trace_i915_gem_object_pwrite(obj, args->offset, args->size);
+
+	ret = -ENODEV;
+	if (obj->ops->pwrite)
+		ret = obj->ops->pwrite(obj, args);
+	if (ret != -ENODEV)
+		goto err;
+
+	ret = i915_gem_object_wait(obj,
+				   I915_WAIT_INTERRUPTIBLE |
+				   I915_WAIT_ALL,
+				   MAX_SCHEDULE_TIMEOUT);
+	if (ret)
+		goto err;
+
+	ret = -EFAULT;
+	/* We can only do the GTT pwrite on untiled buffers, as otherwise
+	 * it would end up going through the fenced access, and we'll get
+	 * different detiling behavior between reading and writing.
+	 * pread/pwrite currently are reading and writing from the CPU
+	 * perspective, requiring manual detiling by the client.
+	 */
+	if (!i915_gem_object_has_struct_page(obj) ||
+	    i915_gem_cpu_write_needs_clflush(obj))
+		/* Note that the gtt paths might fail with non-page-backed user
+		 * pointers (e.g. gtt mappings when moving data between
+		 * textures). Fallback to the shmem path in that case.
+		 */
+		ret = i915_gem_gtt_pwrite_fast(obj, args);
+
+	if (ret == -EFAULT || ret == -ENOSPC) {
+		if (i915_gem_object_has_struct_page(obj))
+			ret = i915_gem_shmem_pwrite(obj, args);
+	}
+
+err:
+	i915_gem_object_put(obj);
+	return ret;
+}
+
+/**
+ * Called when user space has done writes to this buffer
+ * @dev: drm device
+ * @data: ioctl data blob
+ * @file: drm file
+ */
+int
+i915_gem_sw_finish_ioctl(struct drm_device *dev, void *data,
+			 struct drm_file *file)
+{
+	struct drm_i915_gem_sw_finish *args = data;
+	struct drm_i915_gem_object *obj;
+
+	obj = i915_gem_object_lookup(file, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	/*
+	 * Proxy objects are barred from CPU access, so there is no
+	 * need to ban sw_finish as it is a nop.
+	 */
+
+	/* Pinned buffers may be scanout, so flush the cache */
+	i915_gem_object_flush_if_display(obj);
+	i915_gem_object_put(obj);
+
+	return 0;
+}
+
+void i915_gem_runtime_suspend(struct drm_i915_private *i915)
+{
+	struct drm_i915_gem_object *obj, *on;
+	int i;
+
+	/*
+	 * Only called during RPM suspend. All users of the userfault_list
+	 * must be holding an RPM wakeref to ensure that this can not
+	 * run concurrently with themselves (and use the struct_mutex for
+	 * protection between themselves).
+	 */
+
+	list_for_each_entry_safe(obj, on,
+				 &to_gt(i915)->ggtt->userfault_list, userfault_link)
+		__i915_gem_object_release_mmap_gtt(obj);
+
+	list_for_each_entry_safe(obj, on,
+				 &i915->runtime_pm.lmem_userfault_list, userfault_link)
+		i915_gem_object_runtime_pm_release_mmap_offset(obj);
+
+	/*
+	 * The fence will be lost when the device powers down. If any were
+	 * in use by hardware (i.e. they are pinned), we should not be powering
+	 * down! All other fences will be reacquired by the user upon waking.
+	 */
+	for (i = 0; i < to_gt(i915)->ggtt->num_fences; i++) {
+		struct i915_fence_reg *reg = &to_gt(i915)->ggtt->fence_regs[i];
+
+		/*
+		 * Ideally we want to assert that the fence register is not
+		 * live at this point (i.e. that no piece of code will be
+		 * trying to write through fence + GTT, as that both violates
+		 * our tracking of activity and associated locking/barriers,
+		 * but also is illegal given that the hw is powered down).
+		 *
+		 * Previously we used reg->pin_count as a "liveness" indicator.
+		 * That is not sufficient, and we need a more fine-grained
+		 * tool if we want to have a sanity check here.
+		 */
+
+		if (!reg->vma)
+			continue;
+
+		GEM_BUG_ON(i915_vma_has_userfault(reg->vma));
+		reg->dirty = true;
+	}
+}
+
+static void discard_ggtt_vma(struct i915_vma *vma)
+{
+	struct drm_i915_gem_object *obj = vma->obj;
+
+	spin_lock(&obj->vma.lock);
+	if (!RB_EMPTY_NODE(&vma->obj_node)) {
+		rb_erase(&vma->obj_node, &obj->vma.tree);
+		RB_CLEAR_NODE(&vma->obj_node);
+	}
+	spin_unlock(&obj->vma.lock);
+}
+
+struct i915_vma *
+i915_gem_object_ggtt_pin_ww(struct drm_i915_gem_object *obj,
+			    struct i915_gem_ww_ctx *ww,
+			    const struct i915_gtt_view *view,
+			    u64 size, u64 alignment, u64 flags)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	struct i915_vma *vma;
+	int ret;
+
+	GEM_WARN_ON(!ww);
+
+	if (flags & PIN_MAPPABLE &&
+	    (!view || view->type == I915_GTT_VIEW_NORMAL)) {
+		/*
+		 * If the required space is larger than the available
+		 * aperture, we will not able to find a slot for the
+		 * object and unbinding the object now will be in
+		 * vain. Worse, doing so may cause us to ping-pong
+		 * the object in and out of the Global GTT and
+		 * waste a lot of cycles under the mutex.
+		 */
+		if (obj->base.size > ggtt->mappable_end)
+			return ERR_PTR(-E2BIG);
+
+		/*
+		 * If NONBLOCK is set the caller is optimistically
+		 * trying to cache the full object within the mappable
+		 * aperture, and *must* have a fallback in place for
+		 * situations where we cannot bind the object. We
+		 * can be a little more lax here and use the fallback
+		 * more often to avoid costly migrations of ourselves
+		 * and other objects within the aperture.
+		 *
+		 * Half-the-aperture is used as a simple heuristic.
+		 * More interesting would to do search for a free
+		 * block prior to making the commitment to unbind.
+		 * That caters for the self-harm case, and with a
+		 * little more heuristics (e.g. NOFAULT, NOEVICT)
+		 * we could try to minimise harm to others.
+		 */
+		if (flags & PIN_NONBLOCK &&
+		    obj->base.size > ggtt->mappable_end / 2)
+			return ERR_PTR(-ENOSPC);
+	}
+
+new_vma:
+	vma = i915_vma_instance(obj, &ggtt->vm, view);
+	if (IS_ERR(vma))
+		return vma;
+
+	if (i915_vma_misplaced(vma, size, alignment, flags)) {
+		if (flags & PIN_NONBLOCK) {
+			if (i915_vma_is_pinned(vma) || i915_vma_is_active(vma))
+				return ERR_PTR(-ENOSPC);
+
+			/*
+			 * If this misplaced vma is too big (i.e, at-least
+			 * half the size of aperture) or hasn't been pinned
+			 * mappable before, we ignore the misplacement when
+			 * PIN_NONBLOCK is set in order to avoid the ping-pong
+			 * issue described above. In other words, we try to
+			 * avoid the costly operation of unbinding this vma
+			 * from the GGTT and rebinding it back because there
+			 * may not be enough space for this vma in the aperture.
+			 */
+			if (flags & PIN_MAPPABLE &&
+			    (vma->fence_size > ggtt->mappable_end / 2 ||
+			    !i915_vma_is_map_and_fenceable(vma)))
+				return ERR_PTR(-ENOSPC);
+		}
+
+		if (i915_vma_is_pinned(vma) || i915_vma_is_active(vma)) {
+			discard_ggtt_vma(vma);
+			goto new_vma;
+		}
+
+		ret = i915_vma_unbind(vma);
+		if (ret)
+			return ERR_PTR(ret);
+	}
+
+	ret = i915_vma_pin_ww(vma, ww, size, alignment, flags | PIN_GLOBAL);
+
+	if (ret)
+		return ERR_PTR(ret);
+
+	if (vma->fence && !i915_gem_object_is_tiled(obj)) {
+		mutex_lock(&ggtt->vm.mutex);
+		i915_vma_revoke_fence(vma);
+		mutex_unlock(&ggtt->vm.mutex);
+	}
+
+	ret = i915_vma_wait_for_bind(vma);
+	if (ret) {
+		i915_vma_unpin(vma);
+		return ERR_PTR(ret);
+	}
+
+	return vma;
+}
+
+struct i915_vma * __must_check
+i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj,
+			 const struct i915_gtt_view *view,
+			 u64 size, u64 alignment, u64 flags)
+{
+	struct i915_gem_ww_ctx ww;
+	struct i915_vma *ret;
+	int err;
+
+	for_i915_gem_ww(&ww, err, true) {
+		err = i915_gem_object_lock(obj, &ww);
+		if (err)
+			continue;
+
+		ret = i915_gem_object_ggtt_pin_ww(obj, &ww, view, size,
+						  alignment, flags);
+		if (IS_ERR(ret))
+			err = PTR_ERR(ret);
+	}
+
+	return err ? ERR_PTR(err) : ret;
+}
+
+int
+i915_gem_madvise_ioctl(struct drm_device *dev, void *data,
+		       struct drm_file *file_priv)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct drm_i915_gem_madvise *args = data;
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	switch (args->madv) {
+	case I915_MADV_DONTNEED:
+	case I915_MADV_WILLNEED:
+	    break;
+	default:
+	    return -EINVAL;
+	}
+
+	obj = i915_gem_object_lookup(file_priv, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	err = i915_gem_object_lock_interruptible(obj, NULL);
+	if (err)
+		goto out;
+
+	if (i915_gem_object_has_pages(obj) &&
+	    i915_gem_object_is_tiled(obj) &&
+	    i915->gem_quirks & GEM_QUIRK_PIN_SWIZZLED_PAGES) {
+		if (obj->mm.madv == I915_MADV_WILLNEED) {
+			GEM_BUG_ON(!i915_gem_object_has_tiling_quirk(obj));
+			i915_gem_object_clear_tiling_quirk(obj);
+			i915_gem_object_make_shrinkable(obj);
+		}
+		if (args->madv == I915_MADV_WILLNEED) {
+			GEM_BUG_ON(i915_gem_object_has_tiling_quirk(obj));
+			i915_gem_object_make_unshrinkable(obj);
+			i915_gem_object_set_tiling_quirk(obj);
+		}
+	}
+
+	if (obj->mm.madv != __I915_MADV_PURGED) {
+		obj->mm.madv = args->madv;
+		if (obj->ops->adjust_lru)
+			obj->ops->adjust_lru(obj);
+	}
+
+	if (i915_gem_object_has_pages(obj) ||
+	    i915_gem_object_has_self_managed_shrink_list(obj)) {
+		unsigned long flags;
+
+		spin_lock_irqsave(&i915->mm.obj_lock, flags);
+		if (!list_empty(&obj->mm.link)) {
+			struct list_head *list;
+
+			if (obj->mm.madv != I915_MADV_WILLNEED)
+				list = &i915->mm.purge_list;
+			else
+				list = &i915->mm.shrink_list;
+			list_move_tail(&obj->mm.link, list);
+
+		}
+		spin_unlock_irqrestore(&i915->mm.obj_lock, flags);
+	}
+
+	/* if the object is no longer attached, discard its backing storage */
+	if (obj->mm.madv == I915_MADV_DONTNEED &&
+	    !i915_gem_object_has_pages(obj))
+		i915_gem_object_truncate(obj);
+
+	args->retained = obj->mm.madv != __I915_MADV_PURGED;
+
+	i915_gem_object_unlock(obj);
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+/*
+ * A single pass should suffice to release all the freed objects (along most
+ * call paths), but be a little more paranoid in that freeing the objects does
+ * take a little amount of time, during which the rcu callbacks could have added
+ * new objects into the freed list, and armed the work again.
+ */
+void i915_gem_drain_freed_objects(struct drm_i915_private *i915)
+{
+	while (atomic_read(&i915->mm.free_count)) {
+		flush_work(&i915->mm.free_work);
+		flush_delayed_work(&i915->bdev.wq);
+		rcu_barrier();
+	}
+}
+
+/*
+ * Similar to objects above (see i915_gem_drain_freed-objects), in general we
+ * have workers that are armed by RCU and then rearm themselves in their
+ * callbacks. To be paranoid, we need to drain the workqueue a second time after
+ * waiting for the RCU grace period so that we catch work queued via RCU from
+ * the first pass. As neither drain_workqueue() nor flush_workqueue() report a
+ * result, we make an assumption that we only don't require more than 3 passes
+ * to catch all _recursive_ RCU delayed work.
+ */
+void i915_gem_drain_workqueue(struct drm_i915_private *i915)
+{
+	int i;
+
+	for (i = 0; i < 3; i++) {
+		flush_workqueue(i915->wq);
+		rcu_barrier();
+		i915_gem_drain_freed_objects(i915);
+	}
+
+	drain_workqueue(i915->wq);
+}
+
+int i915_gem_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_gt *gt;
+	unsigned int i;
+	int ret;
+
+	/* We need to fallback to 4K pages if host doesn't support huge gtt. */
+	if (intel_vgpu_active(dev_priv) && !intel_vgpu_has_huge_gtt(dev_priv))
+		RUNTIME_INFO(dev_priv)->page_sizes = I915_GTT_PAGE_SIZE_4K;
+
+	ret = i915_gem_init_userptr(dev_priv);
+	if (ret)
+		return ret;
+
+	intel_uc_fetch_firmwares(&to_gt(dev_priv)->uc);
+	intel_wopcm_init(&dev_priv->wopcm);
+
+	ret = i915_init_ggtt(dev_priv);
+	if (ret) {
+		GEM_BUG_ON(ret == -EIO);
+		goto err_unlock;
+	}
+
+	/*
+	 * Despite its name intel_init_clock_gating applies both display
+	 * clock gating workarounds; GT mmio workarounds and the occasional
+	 * GT power context workaround. Worse, sometimes it includes a context
+	 * register workaround which we need to apply before we record the
+	 * default HW state for all contexts.
+	 *
+	 * FIXME: break up the workarounds and apply them at the right time!
+	 */
+	intel_init_clock_gating(dev_priv);
+
+	for_each_gt(gt, dev_priv, i) {
+		ret = intel_gt_init(gt);
+		if (ret)
+			goto err_unlock;
+	}
+
+	return 0;
+
+	/*
+	 * Unwinding is complicated by that we want to handle -EIO to mean
+	 * disable GPU submission but keep KMS alive. We want to mark the
+	 * HW as irrevisibly wedged, but keep enough state around that the
+	 * driver doesn't explode during runtime.
+	 */
+err_unlock:
+	i915_gem_drain_workqueue(dev_priv);
+
+	if (ret != -EIO) {
+		for_each_gt(gt, dev_priv, i) {
+			intel_gt_driver_remove(gt);
+			intel_gt_driver_release(gt);
+			intel_uc_cleanup_firmwares(&gt->uc);
+		}
+	}
+
+	if (ret == -EIO) {
+		/*
+		 * Allow engines or uC initialisation to fail by marking the GPU
+		 * as wedged. But we only want to do this when the GPU is angry,
+		 * for all other failure, such as an allocation failure, bail.
+		 */
+		for_each_gt(gt, dev_priv, i) {
+			if (!intel_gt_is_wedged(gt)) {
+				i915_probe_error(dev_priv,
+						 "Failed to initialize GPU, declaring it wedged!\n");
+				intel_gt_set_wedged(gt);
+			}
+		}
+
+		/* Minimal basic recovery for KMS */
+		ret = i915_ggtt_enable_hw(dev_priv);
+		i915_ggtt_resume(to_gt(dev_priv)->ggtt);
+		intel_init_clock_gating(dev_priv);
+	}
+
+	i915_gem_drain_freed_objects(dev_priv);
+
+	return ret;
+}
+
+void i915_gem_driver_register(struct drm_i915_private *i915)
+{
+	i915_gem_driver_register__shrinker(i915);
+
+	intel_engines_driver_register(i915);
+}
+
+void i915_gem_driver_unregister(struct drm_i915_private *i915)
+{
+	i915_gem_driver_unregister__shrinker(i915);
+}
+
+void i915_gem_driver_remove(struct drm_i915_private *dev_priv)
+{
+	struct intel_gt *gt;
+	unsigned int i;
+
+	i915_gem_suspend_late(dev_priv);
+	for_each_gt(gt, dev_priv, i)
+		intel_gt_driver_remove(gt);
+	dev_priv->uabi_engines = RB_ROOT;
+
+	/* Flush any outstanding unpin_work. */
+	i915_gem_drain_workqueue(dev_priv);
+
+	i915_gem_drain_freed_objects(dev_priv);
+}
+
+void i915_gem_driver_release(struct drm_i915_private *dev_priv)
+{
+	struct intel_gt *gt;
+	unsigned int i;
+
+	for_each_gt(gt, dev_priv, i) {
+		intel_gt_driver_release(gt);
+		intel_uc_cleanup_firmwares(&gt->uc);
+	}
+
+	/* Flush any outstanding work, including i915_gem_context.release_work. */
+	i915_gem_drain_workqueue(dev_priv);
+
+	drm_WARN_ON(&dev_priv->drm, !list_empty(&dev_priv->gem.contexts.list));
+}
+
+static void i915_gem_init__mm(struct drm_i915_private *i915)
+{
+	spin_lock_init(&i915->mm.obj_lock);
+
+	init_llist_head(&i915->mm.free_list);
+
+	INIT_LIST_HEAD(&i915->mm.purge_list);
+	INIT_LIST_HEAD(&i915->mm.shrink_list);
+
+	i915_gem_init__objects(i915);
+}
+
+void i915_gem_init_early(struct drm_i915_private *dev_priv)
+{
+	i915_gem_init__mm(dev_priv);
+	i915_gem_init__contexts(dev_priv);
+
+	spin_lock_init(&dev_priv->display.fb_tracking.lock);
+}
+
+void i915_gem_cleanup_early(struct drm_i915_private *dev_priv)
+{
+	i915_gem_drain_freed_objects(dev_priv);
+	GEM_BUG_ON(!llist_empty(&dev_priv->mm.free_list));
+	GEM_BUG_ON(atomic_read(&dev_priv->mm.free_count));
+	drm_WARN_ON(&dev_priv->drm, dev_priv->mm.shrink_count);
+}
+
+int i915_gem_open(struct drm_i915_private *i915, struct drm_file *file)
+{
+	struct drm_i915_file_private *file_priv;
+	struct i915_drm_client *client;
+	int ret = -ENOMEM;
+
+	DRM_DEBUG("\n");
+
+	file_priv = kzalloc(sizeof(*file_priv), GFP_KERNEL);
+	if (!file_priv)
+		goto err_alloc;
+
+	client = i915_drm_client_add(&i915->clients);
+	if (IS_ERR(client)) {
+		ret = PTR_ERR(client);
+		goto err_client;
+	}
+
+	file->driver_priv = file_priv;
+	file_priv->dev_priv = i915;
+	file_priv->file = file;
+	file_priv->client = client;
+
+	file_priv->bsd_engine = -1;
+	file_priv->hang_timestamp = jiffies;
+
+	ret = i915_gem_context_open(i915, file);
+	if (ret)
+		goto err_context;
+
+	return 0;
+
+err_context:
+	i915_drm_client_put(client);
+err_client:
+	kfree(file_priv);
+err_alloc:
+	return ret;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_gem_device.c"
+#include "selftests/i915_gem.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_gem.h b/drivers/gpu/drm/i915/i915_gem.h
new file mode 100644
index 000000000..a5cdf6662
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem.h
@@ -0,0 +1,130 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __I915_GEM_H__
+#define __I915_GEM_H__
+
+#include <linux/bug.h>
+#include <linux/types.h>
+
+#include <drm/drm_drv.h>
+
+#include "i915_utils.h"
+
+struct drm_file;
+struct drm_i915_gem_object;
+struct drm_i915_private;
+struct i915_gem_ww_ctx;
+struct i915_gtt_view;
+struct i915_vma;
+
+void i915_gem_init_early(struct drm_i915_private *i915);
+void i915_gem_cleanup_early(struct drm_i915_private *i915);
+
+void i915_gem_drain_freed_objects(struct drm_i915_private *i915);
+void i915_gem_drain_workqueue(struct drm_i915_private *i915);
+
+struct i915_vma * __must_check
+i915_gem_object_ggtt_pin_ww(struct drm_i915_gem_object *obj,
+			    struct i915_gem_ww_ctx *ww,
+			    const struct i915_gtt_view *view,
+			    u64 size, u64 alignment, u64 flags);
+
+struct i915_vma * __must_check
+i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj,
+			 const struct i915_gtt_view *view,
+			 u64 size, u64 alignment, u64 flags);
+
+int i915_gem_object_unbind(struct drm_i915_gem_object *obj,
+			   unsigned long flags);
+#define I915_GEM_OBJECT_UNBIND_ACTIVE BIT(0)
+#define I915_GEM_OBJECT_UNBIND_BARRIER BIT(1)
+#define I915_GEM_OBJECT_UNBIND_TEST BIT(2)
+#define I915_GEM_OBJECT_UNBIND_VM_TRYLOCK BIT(3)
+#define I915_GEM_OBJECT_UNBIND_ASYNC BIT(4)
+
+void i915_gem_runtime_suspend(struct drm_i915_private *i915);
+
+int __must_check i915_gem_init(struct drm_i915_private *i915);
+void i915_gem_driver_register(struct drm_i915_private *i915);
+void i915_gem_driver_unregister(struct drm_i915_private *i915);
+void i915_gem_driver_remove(struct drm_i915_private *i915);
+void i915_gem_driver_release(struct drm_i915_private *i915);
+
+int i915_gem_open(struct drm_i915_private *i915, struct drm_file *file);
+
+/* FIXME: All of the below belong somewhere else. */
+
+#ifdef CONFIG_DRM_I915_DEBUG_GEM
+
+#define GEM_SHOW_DEBUG() drm_debug_enabled(DRM_UT_DRIVER)
+
+#ifdef CONFIG_DRM_I915_DEBUG_GEM_ONCE
+#define __GEM_BUG(cond) BUG()
+#else
+#define __GEM_BUG(cond) \
+	WARN(1, "%s:%d GEM_BUG_ON(%s)\n", __func__, __LINE__, __stringify(cond))
+#endif
+
+#define GEM_BUG_ON(condition) do { if (unlikely((condition))) {	\
+		GEM_TRACE_ERR("%s:%d GEM_BUG_ON(%s)\n", \
+			      __func__, __LINE__, __stringify(condition)); \
+		GEM_TRACE_DUMP(); \
+		__GEM_BUG(condition); \
+		} \
+	} while(0)
+#define GEM_WARN_ON(expr) WARN_ON(expr)
+
+#define GEM_DEBUG_WARN_ON(expr) GEM_WARN_ON(expr)
+
+#else
+
+#define GEM_SHOW_DEBUG() (0)
+
+#define GEM_BUG_ON(expr) BUILD_BUG_ON_INVALID(expr)
+#define GEM_WARN_ON(expr) ({ unlikely(!!(expr)); })
+
+#define GEM_DEBUG_WARN_ON(expr) ({ BUILD_BUG_ON_INVALID(expr); 0; })
+#endif
+
+#if IS_ENABLED(CONFIG_DRM_I915_TRACE_GEM)
+#define GEM_TRACE(...) trace_printk(__VA_ARGS__)
+#define GEM_TRACE_ERR(...) do {						\
+	pr_err(__VA_ARGS__);						\
+	trace_printk(__VA_ARGS__);					\
+} while (0)
+#define GEM_TRACE_DUMP() \
+	do { ftrace_dump(DUMP_ALL); __add_taint_for_CI(TAINT_WARN); } while (0)
+#define GEM_TRACE_DUMP_ON(expr) \
+	do { if (expr) GEM_TRACE_DUMP(); } while (0)
+#else
+#define GEM_TRACE(...) do { } while (0)
+#define GEM_TRACE_ERR(...) do { } while (0)
+#define GEM_TRACE_DUMP() do { } while (0)
+#define GEM_TRACE_DUMP_ON(expr) BUILD_BUG_ON_INVALID(expr)
+#endif
+
+#define I915_GEM_IDLE_TIMEOUT (HZ / 5)
+
+#endif /* __I915_GEM_H__ */
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
new file mode 100644
index 000000000..a4b4d9b7d
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -0,0 +1,520 @@
+/*
+ * Copyright © 2008-2010 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *    Chris Wilson <chris@chris-wilson.co.uuk>
+ *
+ */
+
+#include "gem/i915_gem_context.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_requests.h"
+
+#include "i915_drv.h"
+#include "i915_gem_evict.h"
+#include "i915_trace.h"
+
+I915_SELFTEST_DECLARE(static struct igt_evict_ctl {
+	bool fail_if_busy:1;
+} igt_evict_ctl;)
+
+static bool dying_vma(struct i915_vma *vma)
+{
+	return !kref_read(&vma->obj->base.refcount);
+}
+
+static int ggtt_flush(struct intel_gt *gt)
+{
+	/*
+	 * Not everything in the GGTT is tracked via vma (otherwise we
+	 * could evict as required with minimal stalling) so we are forced
+	 * to idle the GPU and explicitly retire outstanding requests in
+	 * the hopes that we can then remove contexts and the like only
+	 * bound by their active reference.
+	 */
+	return intel_gt_wait_for_idle(gt, MAX_SCHEDULE_TIMEOUT);
+}
+
+static bool grab_vma(struct i915_vma *vma, struct i915_gem_ww_ctx *ww)
+{
+	/*
+	 * We add the extra refcount so the object doesn't drop to zero until
+	 * after ungrab_vma(), this way trylock is always paired with unlock.
+	 */
+	if (i915_gem_object_get_rcu(vma->obj)) {
+		if (!i915_gem_object_trylock(vma->obj, ww)) {
+			i915_gem_object_put(vma->obj);
+			return false;
+		}
+	} else {
+		/* Dead objects don't need pins */
+		atomic_and(~I915_VMA_PIN_MASK, &vma->flags);
+	}
+
+	return true;
+}
+
+static void ungrab_vma(struct i915_vma *vma)
+{
+	if (dying_vma(vma))
+		return;
+
+	i915_gem_object_unlock(vma->obj);
+	i915_gem_object_put(vma->obj);
+}
+
+static bool
+mark_free(struct drm_mm_scan *scan,
+	  struct i915_gem_ww_ctx *ww,
+	  struct i915_vma *vma,
+	  unsigned int flags,
+	  struct list_head *unwind)
+{
+	if (i915_vma_is_pinned(vma))
+		return false;
+
+	if (!grab_vma(vma, ww))
+		return false;
+
+	list_add(&vma->evict_link, unwind);
+	return drm_mm_scan_add_block(scan, &vma->node);
+}
+
+static bool defer_evict(struct i915_vma *vma)
+{
+	if (i915_vma_is_active(vma))
+		return true;
+
+	if (i915_vma_is_scanout(vma))
+		return true;
+
+	return false;
+}
+
+/**
+ * i915_gem_evict_something - Evict vmas to make room for binding a new one
+ * @vm: address space to evict from
+ * @ww: An optional struct i915_gem_ww_ctx.
+ * @min_size: size of the desired free space
+ * @alignment: alignment constraint of the desired free space
+ * @color: color for the desired space
+ * @start: start (inclusive) of the range from which to evict objects
+ * @end: end (exclusive) of the range from which to evict objects
+ * @flags: additional flags to control the eviction algorithm
+ *
+ * This function will try to evict vmas until a free space satisfying the
+ * requirements is found. Callers must check first whether any such hole exists
+ * already before calling this function.
+ *
+ * This function is used by the object/vma binding code.
+ *
+ * Since this function is only used to free up virtual address space it only
+ * ignores pinned vmas, and not object where the backing storage itself is
+ * pinned. Hence obj->pages_pin_count does not protect against eviction.
+ *
+ * To clarify: This is for freeing up virtual address space, not for freeing
+ * memory in e.g. the shrinker.
+ */
+int
+i915_gem_evict_something(struct i915_address_space *vm,
+			 struct i915_gem_ww_ctx *ww,
+			 u64 min_size, u64 alignment,
+			 unsigned long color,
+			 u64 start, u64 end,
+			 unsigned flags)
+{
+	struct drm_mm_scan scan;
+	struct list_head eviction_list;
+	struct i915_vma *vma, *next;
+	struct drm_mm_node *node;
+	enum drm_mm_insert_mode mode;
+	struct i915_vma *active;
+	int ret;
+
+	lockdep_assert_held(&vm->mutex);
+	trace_i915_gem_evict(vm, min_size, alignment, flags);
+
+	/*
+	 * The goal is to evict objects and amalgamate space in rough LRU order.
+	 * Since both active and inactive objects reside on the same list,
+	 * in a mix of creation and last scanned order, as we process the list
+	 * we sort it into inactive/active, which keeps the active portion
+	 * in a rough MRU order.
+	 *
+	 * The retirement sequence is thus:
+	 *   1. Inactive objects (already retired, random order)
+	 *   2. Active objects (will stall on unbinding, oldest scanned first)
+	 */
+	mode = DRM_MM_INSERT_BEST;
+	if (flags & PIN_HIGH)
+		mode = DRM_MM_INSERT_HIGH;
+	if (flags & PIN_MAPPABLE)
+		mode = DRM_MM_INSERT_LOW;
+	drm_mm_scan_init_with_range(&scan, &vm->mm,
+				    min_size, alignment, color,
+				    start, end, mode);
+
+	intel_gt_retire_requests(vm->gt);
+
+search_again:
+	active = NULL;
+	INIT_LIST_HEAD(&eviction_list);
+	list_for_each_entry_safe(vma, next, &vm->bound_list, vm_link) {
+		if (vma == active) { /* now seen this vma twice */
+			if (flags & PIN_NONBLOCK)
+				break;
+
+			active = ERR_PTR(-EAGAIN);
+		}
+
+		/*
+		 * We keep this list in a rough least-recently scanned order
+		 * of active elements (inactive elements are cheap to reap).
+		 * New entries are added to the end, and we move anything we
+		 * scan to the end. The assumption is that the working set
+		 * of applications is either steady state (and thanks to the
+		 * userspace bo cache it almost always is) or volatile and
+		 * frequently replaced after a frame, which are self-evicting!
+		 * Given that assumption, the MRU order of the scan list is
+		 * fairly static, and keeping it in least-recently scan order
+		 * is suitable.
+		 *
+		 * To notice when we complete one full cycle, we record the
+		 * first active element seen, before moving it to the tail.
+		 */
+		if (active != ERR_PTR(-EAGAIN) && defer_evict(vma)) {
+			if (!active)
+				active = vma;
+
+			list_move_tail(&vma->vm_link, &vm->bound_list);
+			continue;
+		}
+
+		if (mark_free(&scan, ww, vma, flags, &eviction_list))
+			goto found;
+	}
+
+	/* Nothing found, clean up and bail out! */
+	list_for_each_entry_safe(vma, next, &eviction_list, evict_link) {
+		ret = drm_mm_scan_remove_block(&scan, &vma->node);
+		BUG_ON(ret);
+		ungrab_vma(vma);
+	}
+
+	/*
+	 * Can we unpin some objects such as idle hw contents,
+	 * or pending flips? But since only the GGTT has global entries
+	 * such as scanouts, rinbuffers and contexts, we can skip the
+	 * purge when inspecting per-process local address spaces.
+	 */
+	if (!i915_is_ggtt(vm) || flags & PIN_NONBLOCK)
+		return -ENOSPC;
+
+	/*
+	 * Not everything in the GGTT is tracked via VMA using
+	 * i915_vma_move_to_active(), otherwise we could evict as required
+	 * with minimal stalling. Instead we are forced to idle the GPU and
+	 * explicitly retire outstanding requests which will then remove
+	 * the pinning for active objects such as contexts and ring,
+	 * enabling us to evict them on the next iteration.
+	 *
+	 * To ensure that all user contexts are evictable, we perform
+	 * a switch to the perma-pinned kernel context. This all also gives
+	 * us a termination condition, when the last retired context is
+	 * the kernel's there is no more we can evict.
+	 */
+	if (I915_SELFTEST_ONLY(igt_evict_ctl.fail_if_busy))
+		return -EBUSY;
+
+	ret = ggtt_flush(vm->gt);
+	if (ret)
+		return ret;
+
+	cond_resched();
+
+	flags |= PIN_NONBLOCK;
+	goto search_again;
+
+found:
+	/* drm_mm doesn't allow any other other operations while
+	 * scanning, therefore store to-be-evicted objects on a
+	 * temporary list and take a reference for all before
+	 * calling unbind (which may remove the active reference
+	 * of any of our objects, thus corrupting the list).
+	 */
+	list_for_each_entry_safe(vma, next, &eviction_list, evict_link) {
+		if (drm_mm_scan_remove_block(&scan, &vma->node)) {
+			__i915_vma_pin(vma);
+		} else {
+			list_del(&vma->evict_link);
+			ungrab_vma(vma);
+		}
+	}
+
+	/* Unbinding will emit any required flushes */
+	ret = 0;
+	list_for_each_entry_safe(vma, next, &eviction_list, evict_link) {
+		__i915_vma_unpin(vma);
+		if (ret == 0)
+			ret = __i915_vma_unbind(vma);
+		ungrab_vma(vma);
+	}
+
+	while (ret == 0 && (node = drm_mm_scan_color_evict(&scan))) {
+		vma = container_of(node, struct i915_vma, node);
+
+		/* If we find any non-objects (!vma), we cannot evict them */
+		if (vma->node.color != I915_COLOR_UNEVICTABLE &&
+		    grab_vma(vma, ww)) {
+			ret = __i915_vma_unbind(vma);
+			ungrab_vma(vma);
+		} else {
+			ret = -ENOSPC;
+		}
+	}
+
+	return ret;
+}
+
+/**
+ * i915_gem_evict_for_node - Evict vmas to make room for binding a new one
+ * @vm: address space to evict from
+ * @ww: An optional struct i915_gem_ww_ctx.
+ * @target: range (and color) to evict for
+ * @flags: additional flags to control the eviction algorithm
+ *
+ * This function will try to evict vmas that overlap the target node.
+ *
+ * To clarify: This is for freeing up virtual address space, not for freeing
+ * memory in e.g. the shrinker.
+ */
+int i915_gem_evict_for_node(struct i915_address_space *vm,
+			    struct i915_gem_ww_ctx *ww,
+			    struct drm_mm_node *target,
+			    unsigned int flags)
+{
+	LIST_HEAD(eviction_list);
+	struct drm_mm_node *node;
+	u64 start = target->start;
+	u64 end = start + target->size;
+	struct i915_vma *vma, *next;
+	int ret = 0;
+
+	lockdep_assert_held(&vm->mutex);
+	GEM_BUG_ON(!IS_ALIGNED(start, I915_GTT_PAGE_SIZE));
+	GEM_BUG_ON(!IS_ALIGNED(end, I915_GTT_PAGE_SIZE));
+
+	trace_i915_gem_evict_node(vm, target, flags);
+
+	/*
+	 * Retire before we search the active list. Although we have
+	 * reasonable accuracy in our retirement lists, we may have
+	 * a stray pin (preventing eviction) that can only be resolved by
+	 * retiring.
+	 */
+	intel_gt_retire_requests(vm->gt);
+
+	if (i915_vm_has_cache_coloring(vm)) {
+		/* Expand search to cover neighbouring guard pages (or lack!) */
+		if (start)
+			start -= I915_GTT_PAGE_SIZE;
+
+		/* Always look at the page afterwards to avoid the end-of-GTT */
+		end += I915_GTT_PAGE_SIZE;
+	}
+	GEM_BUG_ON(start >= end);
+
+	drm_mm_for_each_node_in_range(node, &vm->mm, start, end) {
+		/* If we find any non-objects (!vma), we cannot evict them */
+		if (node->color == I915_COLOR_UNEVICTABLE) {
+			ret = -ENOSPC;
+			break;
+		}
+
+		GEM_BUG_ON(!drm_mm_node_allocated(node));
+		vma = container_of(node, typeof(*vma), node);
+
+		/*
+		 * If we are using coloring to insert guard pages between
+		 * different cache domains within the address space, we have
+		 * to check whether the objects on either side of our range
+		 * abutt and conflict. If they are in conflict, then we evict
+		 * those as well to make room for our guard pages.
+		 */
+		if (i915_vm_has_cache_coloring(vm)) {
+			if (node->start + node->size == target->start) {
+				if (node->color == target->color)
+					continue;
+			}
+			if (node->start == target->start + target->size) {
+				if (node->color == target->color)
+					continue;
+			}
+		}
+
+		if (i915_vma_is_pinned(vma)) {
+			ret = -ENOSPC;
+			break;
+		}
+
+		if (flags & PIN_NONBLOCK && i915_vma_is_active(vma)) {
+			ret = -ENOSPC;
+			break;
+		}
+
+		if (!grab_vma(vma, ww)) {
+			ret = -ENOSPC;
+			break;
+		}
+
+		/*
+		 * Never show fear in the face of dragons!
+		 *
+		 * We cannot directly remove this node from within this
+		 * iterator and as with i915_gem_evict_something() we employ
+		 * the vma pin_count in order to prevent the action of
+		 * unbinding one vma from freeing (by dropping its active
+		 * reference) another in our eviction list.
+		 */
+		__i915_vma_pin(vma);
+		list_add(&vma->evict_link, &eviction_list);
+	}
+
+	list_for_each_entry_safe(vma, next, &eviction_list, evict_link) {
+		__i915_vma_unpin(vma);
+		if (ret == 0)
+			ret = __i915_vma_unbind(vma);
+
+		ungrab_vma(vma);
+	}
+
+	return ret;
+}
+
+/**
+ * i915_gem_evict_vm - Evict all idle vmas from a vm
+ * @vm: Address space to cleanse
+ * @ww: An optional struct i915_gem_ww_ctx. If not NULL, i915_gem_evict_vm
+ * will be able to evict vma's locked by the ww as well.
+ * @busy_bo: Optional pointer to struct drm_i915_gem_object. If not NULL, then
+ * in the event i915_gem_evict_vm() is unable to trylock an object for eviction,
+ * then @busy_bo will point to it. -EBUSY is also returned. The caller must drop
+ * the vm->mutex, before trying again to acquire the contended lock. The caller
+ * also owns a reference to the object.
+ *
+ * This function evicts all vmas from a vm.
+ *
+ * This is used by the execbuf code as a last-ditch effort to defragment the
+ * address space.
+ *
+ * To clarify: This is for freeing up virtual address space, not for freeing
+ * memory in e.g. the shrinker.
+ */
+int i915_gem_evict_vm(struct i915_address_space *vm, struct i915_gem_ww_ctx *ww,
+		      struct drm_i915_gem_object **busy_bo)
+{
+	int ret = 0;
+
+	lockdep_assert_held(&vm->mutex);
+	trace_i915_gem_evict_vm(vm);
+
+	/* Switch back to the default context in order to unpin
+	 * the existing context objects. However, such objects only
+	 * pin themselves inside the global GTT and performing the
+	 * switch otherwise is ineffective.
+	 */
+	if (i915_is_ggtt(vm)) {
+		ret = ggtt_flush(vm->gt);
+		if (ret)
+			return ret;
+	}
+
+	do {
+		struct i915_vma *vma, *vn;
+		LIST_HEAD(eviction_list);
+		LIST_HEAD(locked_eviction_list);
+
+		list_for_each_entry(vma, &vm->bound_list, vm_link) {
+			if (i915_vma_is_pinned(vma))
+				continue;
+
+			/*
+			 * If we already own the lock, trylock fails. In case
+			 * the resv is shared among multiple objects, we still
+			 * need the object ref.
+			 */
+			if (!i915_gem_object_get_rcu(vma->obj) ||
+			    (ww && (dma_resv_locking_ctx(vma->obj->base.resv) == &ww->ctx))) {
+				__i915_vma_pin(vma);
+				list_add(&vma->evict_link, &locked_eviction_list);
+				continue;
+			}
+
+			if (!i915_gem_object_trylock(vma->obj, ww)) {
+				if (busy_bo) {
+					*busy_bo = vma->obj; /* holds ref */
+					ret = -EBUSY;
+					break;
+				}
+				i915_gem_object_put(vma->obj);
+				continue;
+			}
+
+			__i915_vma_pin(vma);
+			list_add(&vma->evict_link, &eviction_list);
+		}
+		if (list_empty(&eviction_list) && list_empty(&locked_eviction_list))
+			break;
+
+		/* Unbind locked objects first, before unlocking the eviction_list */
+		list_for_each_entry_safe(vma, vn, &locked_eviction_list, evict_link) {
+			__i915_vma_unpin(vma);
+
+			if (ret == 0) {
+				ret = __i915_vma_unbind(vma);
+				if (ret != -EINTR) /* "Get me out of here!" */
+					ret = 0;
+			}
+			if (!dying_vma(vma))
+				i915_gem_object_put(vma->obj);
+		}
+
+		list_for_each_entry_safe(vma, vn, &eviction_list, evict_link) {
+			__i915_vma_unpin(vma);
+			if (ret == 0) {
+				ret = __i915_vma_unbind(vma);
+				if (ret != -EINTR) /* "Get me out of here!" */
+					ret = 0;
+			}
+
+			i915_gem_object_unlock(vma->obj);
+			i915_gem_object_put(vma->obj);
+		}
+	} while (ret == 0);
+
+	return ret;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/i915_gem_evict.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.h b/drivers/gpu/drm/i915/i915_gem_evict.h
new file mode 100644
index 000000000..bf0ee0e4f
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_evict.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_GEM_EVICT_H__
+#define __I915_GEM_EVICT_H__
+
+#include <linux/types.h>
+
+struct drm_mm_node;
+struct i915_address_space;
+struct i915_gem_ww_ctx;
+struct drm_i915_gem_object;
+
+int __must_check i915_gem_evict_something(struct i915_address_space *vm,
+					  struct i915_gem_ww_ctx *ww,
+					  u64 min_size, u64 alignment,
+					  unsigned long color,
+					  u64 start, u64 end,
+					  unsigned flags);
+int __must_check i915_gem_evict_for_node(struct i915_address_space *vm,
+					 struct i915_gem_ww_ctx *ww,
+					 struct drm_mm_node *node,
+					 unsigned int flags);
+int i915_gem_evict_vm(struct i915_address_space *vm,
+		      struct i915_gem_ww_ctx *ww,
+		      struct drm_i915_gem_object **busy_bo);
+
+#endif /* __I915_GEM_EVICT_H__ */
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
new file mode 100644
index 000000000..7bd1861dd
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -0,0 +1,297 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2010 Daniel Vetter
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/slab.h> /* fault-inject.h is not standalone! */
+
+#include <linux/fault-inject.h>
+#include <linux/log2.h>
+#include <linux/random.h>
+#include <linux/seq_file.h>
+#include <linux/stop_machine.h>
+
+#include <asm/set_memory.h>
+#include <asm/smp.h>
+
+#include "display/intel_frontbuffer.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_requests.h"
+
+#include "i915_drv.h"
+#include "i915_gem_evict.h"
+#include "i915_scatterlist.h"
+#include "i915_trace.h"
+#include "i915_vgpu.h"
+
+int i915_gem_gtt_prepare_pages(struct drm_i915_gem_object *obj,
+			       struct sg_table *pages)
+{
+	do {
+		if (dma_map_sg_attrs(obj->base.dev->dev,
+				     pages->sgl, pages->nents,
+				     DMA_BIDIRECTIONAL,
+				     DMA_ATTR_SKIP_CPU_SYNC |
+				     DMA_ATTR_NO_KERNEL_MAPPING |
+				     DMA_ATTR_NO_WARN))
+			return 0;
+
+		/*
+		 * If the DMA remap fails, one cause can be that we have
+		 * too many objects pinned in a small remapping table,
+		 * such as swiotlb. Incrementally purge all other objects and
+		 * try again - if there are no more pages to remove from
+		 * the DMA remapper, i915_gem_shrink will return 0.
+		 */
+		GEM_BUG_ON(obj->mm.pages == pages);
+	} while (i915_gem_shrink(NULL, to_i915(obj->base.dev),
+				 obj->base.size >> PAGE_SHIFT, NULL,
+				 I915_SHRINK_BOUND |
+				 I915_SHRINK_UNBOUND));
+
+	return -ENOSPC;
+}
+
+void i915_gem_gtt_finish_pages(struct drm_i915_gem_object *obj,
+			       struct sg_table *pages)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+
+	/* XXX This does not prevent more requests being submitted! */
+	if (unlikely(ggtt->do_idle_maps))
+		/* Wait a bit, in the hope it avoids the hang */
+		usleep_range(100, 250);
+
+	dma_unmap_sg(i915->drm.dev, pages->sgl, pages->nents,
+		     DMA_BIDIRECTIONAL);
+}
+
+/**
+ * i915_gem_gtt_reserve - reserve a node in an address_space (GTT)
+ * @vm: the &struct i915_address_space
+ * @ww: An optional struct i915_gem_ww_ctx.
+ * @node: the &struct drm_mm_node (typically i915_vma.mode)
+ * @size: how much space to allocate inside the GTT,
+ *        must be #I915_GTT_PAGE_SIZE aligned
+ * @offset: where to insert inside the GTT,
+ *          must be #I915_GTT_MIN_ALIGNMENT aligned, and the node
+ *          (@offset + @size) must fit within the address space
+ * @color: color to apply to node, if this node is not from a VMA,
+ *         color must be #I915_COLOR_UNEVICTABLE
+ * @flags: control search and eviction behaviour
+ *
+ * i915_gem_gtt_reserve() tries to insert the @node at the exact @offset inside
+ * the address space (using @size and @color). If the @node does not fit, it
+ * tries to evict any overlapping nodes from the GTT, including any
+ * neighbouring nodes if the colors do not match (to ensure guard pages between
+ * differing domains). See i915_gem_evict_for_node() for the gory details
+ * on the eviction algorithm. #PIN_NONBLOCK may used to prevent waiting on
+ * evicting active overlapping objects, and any overlapping node that is pinned
+ * or marked as unevictable will also result in failure.
+ *
+ * Returns: 0 on success, -ENOSPC if no suitable hole is found, -EINTR if
+ * asked to wait for eviction and interrupted.
+ */
+int i915_gem_gtt_reserve(struct i915_address_space *vm,
+			 struct i915_gem_ww_ctx *ww,
+			 struct drm_mm_node *node,
+			 u64 size, u64 offset, unsigned long color,
+			 unsigned int flags)
+{
+	int err;
+
+	GEM_BUG_ON(!size);
+	GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE));
+	GEM_BUG_ON(!IS_ALIGNED(offset, I915_GTT_MIN_ALIGNMENT));
+	GEM_BUG_ON(range_overflows(offset, size, vm->total));
+	GEM_BUG_ON(vm == &to_gt(vm->i915)->ggtt->alias->vm);
+	GEM_BUG_ON(drm_mm_node_allocated(node));
+
+	node->size = size;
+	node->start = offset;
+	node->color = color;
+
+	err = drm_mm_reserve_node(&vm->mm, node);
+	if (err != -ENOSPC)
+		return err;
+
+	if (flags & PIN_NOEVICT)
+		return -ENOSPC;
+
+	err = i915_gem_evict_for_node(vm, ww, node, flags);
+	if (err == 0)
+		err = drm_mm_reserve_node(&vm->mm, node);
+
+	return err;
+}
+
+static u64 random_offset(u64 start, u64 end, u64 len, u64 align)
+{
+	u64 range, addr;
+
+	GEM_BUG_ON(range_overflows(start, len, end));
+	GEM_BUG_ON(round_up(start, align) > round_down(end - len, align));
+
+	range = round_down(end - len, align) - round_up(start, align);
+	if (range) {
+		if (sizeof(unsigned long) == sizeof(u64)) {
+			addr = get_random_u64();
+		} else {
+			addr = get_random_u32();
+			if (range > U32_MAX) {
+				addr <<= 32;
+				addr |= get_random_u32();
+			}
+		}
+		div64_u64_rem(addr, range, &addr);
+		start += addr;
+	}
+
+	return round_up(start, align);
+}
+
+/**
+ * i915_gem_gtt_insert - insert a node into an address_space (GTT)
+ * @vm: the &struct i915_address_space
+ * @ww: An optional struct i915_gem_ww_ctx.
+ * @node: the &struct drm_mm_node (typically i915_vma.node)
+ * @size: how much space to allocate inside the GTT,
+ *        must be #I915_GTT_PAGE_SIZE aligned
+ * @alignment: required alignment of starting offset, may be 0 but
+ *             if specified, this must be a power-of-two and at least
+ *             #I915_GTT_MIN_ALIGNMENT
+ * @color: color to apply to node
+ * @start: start of any range restriction inside GTT (0 for all),
+ *         must be #I915_GTT_PAGE_SIZE aligned
+ * @end: end of any range restriction inside GTT (U64_MAX for all),
+ *       must be #I915_GTT_PAGE_SIZE aligned if not U64_MAX
+ * @flags: control search and eviction behaviour
+ *
+ * i915_gem_gtt_insert() first searches for an available hole into which
+ * is can insert the node. The hole address is aligned to @alignment and
+ * its @size must then fit entirely within the [@start, @end] bounds. The
+ * nodes on either side of the hole must match @color, or else a guard page
+ * will be inserted between the two nodes (or the node evicted). If no
+ * suitable hole is found, first a victim is randomly selected and tested
+ * for eviction, otherwise then the LRU list of objects within the GTT
+ * is scanned to find the first set of replacement nodes to create the hole.
+ * Those old overlapping nodes are evicted from the GTT (and so must be
+ * rebound before any future use). Any node that is currently pinned cannot
+ * be evicted (see i915_vma_pin()). Similar if the node's VMA is currently
+ * active and #PIN_NONBLOCK is specified, that node is also skipped when
+ * searching for an eviction candidate. See i915_gem_evict_something() for
+ * the gory details on the eviction algorithm.
+ *
+ * Returns: 0 on success, -ENOSPC if no suitable hole is found, -EINTR if
+ * asked to wait for eviction and interrupted.
+ */
+int i915_gem_gtt_insert(struct i915_address_space *vm,
+			struct i915_gem_ww_ctx *ww,
+			struct drm_mm_node *node,
+			u64 size, u64 alignment, unsigned long color,
+			u64 start, u64 end, unsigned int flags)
+{
+	enum drm_mm_insert_mode mode;
+	u64 offset;
+	int err;
+
+	lockdep_assert_held(&vm->mutex);
+
+	GEM_BUG_ON(!size);
+	GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE));
+	GEM_BUG_ON(alignment && !is_power_of_2(alignment));
+	GEM_BUG_ON(alignment && !IS_ALIGNED(alignment, I915_GTT_MIN_ALIGNMENT));
+	GEM_BUG_ON(start >= end);
+	GEM_BUG_ON(start > 0  && !IS_ALIGNED(start, I915_GTT_PAGE_SIZE));
+	GEM_BUG_ON(end < U64_MAX && !IS_ALIGNED(end, I915_GTT_PAGE_SIZE));
+	GEM_BUG_ON(vm == &to_gt(vm->i915)->ggtt->alias->vm);
+	GEM_BUG_ON(drm_mm_node_allocated(node));
+
+	if (unlikely(range_overflows(start, size, end)))
+		return -ENOSPC;
+
+	if (unlikely(round_up(start, alignment) > round_down(end - size, alignment)))
+		return -ENOSPC;
+
+	mode = DRM_MM_INSERT_BEST;
+	if (flags & PIN_HIGH)
+		mode = DRM_MM_INSERT_HIGHEST;
+	if (flags & PIN_MAPPABLE)
+		mode = DRM_MM_INSERT_LOW;
+
+	/* We only allocate in PAGE_SIZE/GTT_PAGE_SIZE (4096) chunks,
+	 * so we know that we always have a minimum alignment of 4096.
+	 * The drm_mm range manager is optimised to return results
+	 * with zero alignment, so where possible use the optimal
+	 * path.
+	 */
+	BUILD_BUG_ON(I915_GTT_MIN_ALIGNMENT > I915_GTT_PAGE_SIZE);
+	if (alignment <= I915_GTT_MIN_ALIGNMENT)
+		alignment = 0;
+
+	err = drm_mm_insert_node_in_range(&vm->mm, node,
+					  size, alignment, color,
+					  start, end, mode);
+	if (err != -ENOSPC)
+		return err;
+
+	if (mode & DRM_MM_INSERT_ONCE) {
+		err = drm_mm_insert_node_in_range(&vm->mm, node,
+						  size, alignment, color,
+						  start, end,
+						  DRM_MM_INSERT_BEST);
+		if (err != -ENOSPC)
+			return err;
+	}
+
+	if (flags & PIN_NOEVICT)
+		return -ENOSPC;
+
+	/*
+	 * No free space, pick a slot at random.
+	 *
+	 * There is a pathological case here using a GTT shared between
+	 * mmap and GPU (i.e. ggtt/aliasing_ppgtt but not full-ppgtt):
+	 *
+	 *    |<-- 256 MiB aperture -->||<-- 1792 MiB unmappable -->|
+	 *         (64k objects)             (448k objects)
+	 *
+	 * Now imagine that the eviction LRU is ordered top-down (just because
+	 * pathology meets real life), and that we need to evict an object to
+	 * make room inside the aperture. The eviction scan then has to walk
+	 * the 448k list before it finds one within range. And now imagine that
+	 * it has to search for a new hole between every byte inside the memcpy,
+	 * for several simultaneous clients.
+	 *
+	 * On a full-ppgtt system, if we have run out of available space, there
+	 * will be lots and lots of objects in the eviction list! Again,
+	 * searching that LRU list may be slow if we are also applying any
+	 * range restrictions (e.g. restriction to low 4GiB) and so, for
+	 * simplicity and similarilty between different GTT, try the single
+	 * random replacement first.
+	 */
+	offset = random_offset(start, end,
+			       size, alignment ?: I915_GTT_MIN_ALIGNMENT);
+	err = i915_gem_gtt_reserve(vm, ww, node, size, offset, color, flags);
+	if (err != -ENOSPC)
+		return err;
+
+	if (flags & PIN_NOSEARCH)
+		return -ENOSPC;
+
+	/* Randomly selected placement is pinned, do a search */
+	err = i915_gem_evict_something(vm, ww, size, alignment, color,
+				       start, end, flags);
+	if (err)
+		return err;
+
+	return drm_mm_insert_node_in_range(&vm->mm, node,
+					   size, alignment, color,
+					   start, end, DRM_MM_INSERT_EVICT);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/i915_gem_gtt.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h
new file mode 100644
index 000000000..8c2f57eb5
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
@@ -0,0 +1,54 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __I915_GEM_GTT_H__
+#define __I915_GEM_GTT_H__
+
+#include <linux/io-mapping.h>
+#include <linux/types.h>
+
+#include <drm/drm_mm.h>
+
+#include "gt/intel_gtt.h"
+#include "i915_scatterlist.h"
+
+struct drm_i915_gem_object;
+struct i915_address_space;
+struct i915_gem_ww_ctx;
+
+int __must_check i915_gem_gtt_prepare_pages(struct drm_i915_gem_object *obj,
+					    struct sg_table *pages);
+void i915_gem_gtt_finish_pages(struct drm_i915_gem_object *obj,
+			       struct sg_table *pages);
+
+int i915_gem_gtt_reserve(struct i915_address_space *vm,
+			 struct i915_gem_ww_ctx *ww,
+			 struct drm_mm_node *node,
+			 u64 size, u64 offset, unsigned long color,
+			 unsigned int flags);
+
+int i915_gem_gtt_insert(struct i915_address_space *vm,
+			struct i915_gem_ww_ctx *ww,
+			struct drm_mm_node *node,
+			u64 size, u64 alignment, unsigned long color,
+			u64 start, u64 end, unsigned int flags);
+
+/* Flags used by pin/bind&friends. */
+#define PIN_NOEVICT		BIT_ULL(0)
+#define PIN_NOSEARCH		BIT_ULL(1)
+#define PIN_NONBLOCK		BIT_ULL(2)
+#define PIN_MAPPABLE		BIT_ULL(3)
+#define PIN_ZONE_4G		BIT_ULL(4)
+#define PIN_HIGH		BIT_ULL(5)
+#define PIN_OFFSET_BIAS		BIT_ULL(6)
+#define PIN_OFFSET_FIXED	BIT_ULL(7)
+#define PIN_VALIDATE		BIT_ULL(8) /* validate placement only, no need to call unpin() */
+
+#define PIN_GLOBAL		BIT_ULL(10) /* I915_VMA_GLOBAL_BIND */
+#define PIN_USER		BIT_ULL(11) /* I915_VMA_LOCAL_BIND */
+
+#define PIN_OFFSET_MASK		I915_GTT_PAGE_MASK
+
+#endif
diff --git a/drivers/gpu/drm/i915/i915_gem_ww.c b/drivers/gpu/drm/i915/i915_gem_ww.c
new file mode 100644
index 000000000..3f6ff1394
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_ww.c
@@ -0,0 +1,63 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+#include <linux/dma-resv.h>
+#include "i915_gem_ww.h"
+#include "gem/i915_gem_object.h"
+
+void i915_gem_ww_ctx_init(struct i915_gem_ww_ctx *ww, bool intr)
+{
+	ww_acquire_init(&ww->ctx, &reservation_ww_class);
+	INIT_LIST_HEAD(&ww->obj_list);
+	ww->intr = intr;
+	ww->contended = NULL;
+}
+
+static void i915_gem_ww_ctx_unlock_all(struct i915_gem_ww_ctx *ww)
+{
+	struct drm_i915_gem_object *obj;
+
+	while ((obj = list_first_entry_or_null(&ww->obj_list, struct drm_i915_gem_object, obj_link))) {
+		list_del(&obj->obj_link);
+		i915_gem_object_unlock(obj);
+		i915_gem_object_put(obj);
+	}
+}
+
+void i915_gem_ww_unlock_single(struct drm_i915_gem_object *obj)
+{
+	list_del(&obj->obj_link);
+	i915_gem_object_unlock(obj);
+	i915_gem_object_put(obj);
+}
+
+void i915_gem_ww_ctx_fini(struct i915_gem_ww_ctx *ww)
+{
+	i915_gem_ww_ctx_unlock_all(ww);
+	WARN_ON(ww->contended);
+	ww_acquire_fini(&ww->ctx);
+}
+
+int __must_check i915_gem_ww_ctx_backoff(struct i915_gem_ww_ctx *ww)
+{
+	int ret = 0;
+
+	if (WARN_ON(!ww->contended))
+		return -EINVAL;
+
+	i915_gem_ww_ctx_unlock_all(ww);
+	if (ww->intr)
+		ret = dma_resv_lock_slow_interruptible(ww->contended->base.resv, &ww->ctx);
+	else
+		dma_resv_lock_slow(ww->contended->base.resv, &ww->ctx);
+
+	if (!ret)
+		list_add_tail(&ww->contended->obj_link, &ww->obj_list);
+	else
+		i915_gem_object_put(ww->contended);
+
+	ww->contended = NULL;
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/i915_gem_ww.h b/drivers/gpu/drm/i915/i915_gem_ww.h
new file mode 100644
index 000000000..86f0fe343
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_ww.h
@@ -0,0 +1,41 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+#ifndef __I915_GEM_WW_H__
+#define __I915_GEM_WW_H__
+
+#include <drm/drm_drv.h>
+
+struct i915_gem_ww_ctx {
+	struct ww_acquire_ctx ctx;
+	struct list_head obj_list;
+	struct drm_i915_gem_object *contended;
+	bool intr;
+};
+
+void i915_gem_ww_ctx_init(struct i915_gem_ww_ctx *ctx, bool intr);
+void i915_gem_ww_ctx_fini(struct i915_gem_ww_ctx *ctx);
+int __must_check i915_gem_ww_ctx_backoff(struct i915_gem_ww_ctx *ctx);
+void i915_gem_ww_unlock_single(struct drm_i915_gem_object *obj);
+
+/* Internal function used by the inlines! Don't use. */
+static inline int __i915_gem_ww_fini(struct i915_gem_ww_ctx *ww, int err)
+{
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(ww);
+		if (!err)
+			err = -EDEADLK;
+	}
+
+	if (err != -EDEADLK)
+		i915_gem_ww_ctx_fini(ww);
+
+	return err;
+}
+
+#define for_i915_gem_ww(_ww, _err, _intr)			  \
+	for (i915_gem_ww_ctx_init(_ww, _intr), (_err) = -EDEADLK; \
+	     (_err) == -EDEADLK;				  \
+	     (_err) = __i915_gem_ww_fini(_ww, _err))
+#endif
diff --git a/drivers/gpu/drm/i915/i915_getparam.c b/drivers/gpu/drm/i915/i915_getparam.c
new file mode 100644
index 000000000..342c8ca64
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_getparam.c
@@ -0,0 +1,187 @@
+/*
+ * SPDX-License-Identifier: MIT
+ */
+
+#include "gem/i915_gem_mman.h"
+#include "gt/intel_engine_user.h"
+
+#include "i915_cmd_parser.h"
+#include "i915_drv.h"
+#include "i915_getparam.h"
+#include "i915_perf.h"
+
+int i915_getparam_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file_priv)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct pci_dev *pdev = to_pci_dev(dev->dev);
+	const struct sseu_dev_info *sseu = &to_gt(i915)->info.sseu;
+	drm_i915_getparam_t *param = data;
+	int value = 0;
+
+	switch (param->param) {
+	case I915_PARAM_IRQ_ACTIVE:
+	case I915_PARAM_ALLOW_BATCHBUFFER:
+	case I915_PARAM_LAST_DISPATCH:
+	case I915_PARAM_HAS_EXEC_CONSTANTS:
+		/* Reject all old ums/dri params. */
+		return -ENODEV;
+	case I915_PARAM_CHIPSET_ID:
+		value = pdev->device;
+		break;
+	case I915_PARAM_REVISION:
+		value = pdev->revision;
+		break;
+	case I915_PARAM_NUM_FENCES_AVAIL:
+		value = to_gt(i915)->ggtt->num_fences;
+		break;
+	case I915_PARAM_HAS_OVERLAY:
+		value = !!i915->display.overlay;
+		break;
+	case I915_PARAM_HAS_BSD:
+		value = !!intel_engine_lookup_user(i915,
+						   I915_ENGINE_CLASS_VIDEO, 0);
+		break;
+	case I915_PARAM_HAS_BLT:
+		value = !!intel_engine_lookup_user(i915,
+						   I915_ENGINE_CLASS_COPY, 0);
+		break;
+	case I915_PARAM_HAS_VEBOX:
+		value = !!intel_engine_lookup_user(i915,
+						   I915_ENGINE_CLASS_VIDEO_ENHANCE, 0);
+		break;
+	case I915_PARAM_HAS_BSD2:
+		value = !!intel_engine_lookup_user(i915,
+						   I915_ENGINE_CLASS_VIDEO, 1);
+		break;
+	case I915_PARAM_HAS_LLC:
+		value = HAS_LLC(i915);
+		break;
+	case I915_PARAM_HAS_WT:
+		value = HAS_WT(i915);
+		break;
+	case I915_PARAM_HAS_ALIASING_PPGTT:
+		value = INTEL_PPGTT(i915);
+		break;
+	case I915_PARAM_HAS_SEMAPHORES:
+		value = !!(i915->caps.scheduler & I915_SCHEDULER_CAP_SEMAPHORES);
+		break;
+	case I915_PARAM_HAS_SECURE_BATCHES:
+		value = HAS_SECURE_BATCHES(i915) && capable(CAP_SYS_ADMIN);
+		break;
+	case I915_PARAM_CMD_PARSER_VERSION:
+		value = i915_cmd_parser_get_version(i915);
+		break;
+	case I915_PARAM_SUBSLICE_TOTAL:
+		value = intel_sseu_subslice_total(sseu);
+		if (!value)
+			return -ENODEV;
+		break;
+	case I915_PARAM_EU_TOTAL:
+		value = sseu->eu_total;
+		if (!value)
+			return -ENODEV;
+		break;
+	case I915_PARAM_HAS_GPU_RESET:
+		value = i915->params.enable_hangcheck &&
+			intel_has_gpu_reset(to_gt(i915));
+		if (value && intel_has_reset_engine(to_gt(i915)))
+			value = 2;
+		break;
+	case I915_PARAM_HAS_RESOURCE_STREAMER:
+		value = 0;
+		break;
+	case I915_PARAM_HAS_POOLED_EU:
+		value = HAS_POOLED_EU(i915);
+		break;
+	case I915_PARAM_MIN_EU_IN_POOL:
+		value = sseu->min_eu_in_pool;
+		break;
+	case I915_PARAM_HUC_STATUS:
+		value = intel_huc_check_status(&to_gt(i915)->uc.huc);
+		if (value < 0)
+			return value;
+		break;
+	case I915_PARAM_MMAP_GTT_VERSION:
+		/* Though we've started our numbering from 1, and so class all
+		 * earlier versions as 0, in effect their value is undefined as
+		 * the ioctl will report EINVAL for the unknown param!
+		 */
+		value = i915_gem_mmap_gtt_version();
+		break;
+	case I915_PARAM_HAS_SCHEDULER:
+		value = i915->caps.scheduler;
+		break;
+
+	case I915_PARAM_MMAP_VERSION:
+		/* Remember to bump this if the version changes! */
+	case I915_PARAM_HAS_GEM:
+	case I915_PARAM_HAS_PAGEFLIPPING:
+	case I915_PARAM_HAS_EXECBUF2: /* depends on GEM */
+	case I915_PARAM_HAS_RELAXED_FENCING:
+	case I915_PARAM_HAS_COHERENT_RINGS:
+	case I915_PARAM_HAS_RELAXED_DELTA:
+	case I915_PARAM_HAS_GEN7_SOL_RESET:
+	case I915_PARAM_HAS_WAIT_TIMEOUT:
+	case I915_PARAM_HAS_PRIME_VMAP_FLUSH:
+	case I915_PARAM_HAS_PINNED_BATCHES:
+	case I915_PARAM_HAS_EXEC_NO_RELOC:
+	case I915_PARAM_HAS_EXEC_HANDLE_LUT:
+	case I915_PARAM_HAS_COHERENT_PHYS_GTT:
+	case I915_PARAM_HAS_EXEC_SOFTPIN:
+	case I915_PARAM_HAS_EXEC_ASYNC:
+	case I915_PARAM_HAS_EXEC_FENCE:
+	case I915_PARAM_HAS_EXEC_CAPTURE:
+	case I915_PARAM_HAS_EXEC_BATCH_FIRST:
+	case I915_PARAM_HAS_EXEC_FENCE_ARRAY:
+	case I915_PARAM_HAS_EXEC_SUBMIT_FENCE:
+	case I915_PARAM_HAS_EXEC_TIMELINE_FENCES:
+	case I915_PARAM_HAS_USERPTR_PROBE:
+		/* For the time being all of these are always true;
+		 * if some supported hardware does not have one of these
+		 * features this value needs to be provided from
+		 * INTEL_INFO(), a feature macro, or similar.
+		 */
+		value = 1;
+		break;
+	case I915_PARAM_HAS_CONTEXT_ISOLATION:
+		value = intel_engines_has_context_isolation(i915);
+		break;
+	case I915_PARAM_SLICE_MASK:
+		/* Not supported from Xe_HP onward; use topology queries */
+		if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50))
+			return -EINVAL;
+
+		value = sseu->slice_mask;
+		if (!value)
+			return -ENODEV;
+		break;
+	case I915_PARAM_SUBSLICE_MASK:
+		/* Not supported from Xe_HP onward; use topology queries */
+		if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50))
+			return -EINVAL;
+
+		/* Only copy bits from the first slice */
+		value = intel_sseu_get_hsw_subslices(sseu, 0);
+		if (!value)
+			return -ENODEV;
+		break;
+	case I915_PARAM_CS_TIMESTAMP_FREQUENCY:
+		value = to_gt(i915)->clock_frequency;
+		break;
+	case I915_PARAM_MMAP_GTT_COHERENT:
+		value = INTEL_INFO(i915)->has_coherent_ggtt;
+		break;
+	case I915_PARAM_PERF_REVISION:
+		value = i915_perf_ioctl_version();
+		break;
+	default:
+		DRM_DEBUG("Unknown parameter %d\n", param->param);
+		return -EINVAL;
+	}
+
+	if (put_user(value, param->value))
+		return -EFAULT;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_getparam.h b/drivers/gpu/drm/i915/i915_getparam.h
new file mode 100644
index 000000000..18e4752e8
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_getparam.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_GETPARAM_H__
+#define __I915_GETPARAM_H__
+
+struct drm_device;
+struct drm_file;
+
+int i915_getparam_ioctl(struct drm_device *dev, void *data,
+			struct drm_file *file_priv);
+
+#endif /* __I915_GETPARAM_H__ */
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
new file mode 100644
index 000000000..847b9e6af
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -0,0 +1,2186 @@
+/*
+ * Copyright (c) 2008 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *    Keith Packard <keithp@keithp.com>
+ *    Mika Kuoppala <mika.kuoppala@intel.com>
+ *
+ */
+
+#include <linux/ascii85.h>
+#include <linux/highmem.h>
+#include <linux/nmi.h>
+#include <linux/pagevec.h>
+#include <linux/scatterlist.h>
+#include <linux/string_helpers.h>
+#include <linux/utsname.h>
+#include <linux/zlib.h>
+
+#include <drm/drm_cache.h>
+#include <drm/drm_print.h>
+
+#include "display/intel_dmc.h"
+#include "display/intel_overlay.h"
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_lmem.h"
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_mcr.h"
+#include "gt/intel_gt_pm.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/uc/intel_guc_capture.h"
+
+#include "i915_driver.h"
+#include "i915_drv.h"
+#include "i915_gpu_error.h"
+#include "i915_memcpy.h"
+#include "i915_scatterlist.h"
+#include "i915_utils.h"
+
+#define ALLOW_FAIL (__GFP_KSWAPD_RECLAIM | __GFP_RETRY_MAYFAIL | __GFP_NOWARN)
+#define ATOMIC_MAYFAIL (GFP_ATOMIC | __GFP_NOWARN)
+
+static void __sg_set_buf(struct scatterlist *sg,
+			 void *addr, unsigned int len, loff_t it)
+{
+	sg->page_link = (unsigned long)virt_to_page(addr);
+	sg->offset = offset_in_page(addr);
+	sg->length = len;
+	sg->dma_address = it;
+}
+
+static bool __i915_error_grow(struct drm_i915_error_state_buf *e, size_t len)
+{
+	if (!len)
+		return false;
+
+	if (e->bytes + len + 1 <= e->size)
+		return true;
+
+	if (e->bytes) {
+		__sg_set_buf(e->cur++, e->buf, e->bytes, e->iter);
+		e->iter += e->bytes;
+		e->buf = NULL;
+		e->bytes = 0;
+	}
+
+	if (e->cur == e->end) {
+		struct scatterlist *sgl;
+
+		sgl = (typeof(sgl))__get_free_page(ALLOW_FAIL);
+		if (!sgl) {
+			e->err = -ENOMEM;
+			return false;
+		}
+
+		if (e->cur) {
+			e->cur->offset = 0;
+			e->cur->length = 0;
+			e->cur->page_link =
+				(unsigned long)sgl | SG_CHAIN;
+		} else {
+			e->sgl = sgl;
+		}
+
+		e->cur = sgl;
+		e->end = sgl + SG_MAX_SINGLE_ALLOC - 1;
+	}
+
+	e->size = ALIGN(len + 1, SZ_64K);
+	e->buf = kmalloc(e->size, ALLOW_FAIL);
+	if (!e->buf) {
+		e->size = PAGE_ALIGN(len + 1);
+		e->buf = kmalloc(e->size, GFP_KERNEL);
+	}
+	if (!e->buf) {
+		e->err = -ENOMEM;
+		return false;
+	}
+
+	return true;
+}
+
+__printf(2, 0)
+static void i915_error_vprintf(struct drm_i915_error_state_buf *e,
+			       const char *fmt, va_list args)
+{
+	va_list ap;
+	int len;
+
+	if (e->err)
+		return;
+
+	va_copy(ap, args);
+	len = vsnprintf(NULL, 0, fmt, ap);
+	va_end(ap);
+	if (len <= 0) {
+		e->err = len;
+		return;
+	}
+
+	if (!__i915_error_grow(e, len))
+		return;
+
+	GEM_BUG_ON(e->bytes >= e->size);
+	len = vscnprintf(e->buf + e->bytes, e->size - e->bytes, fmt, args);
+	if (len < 0) {
+		e->err = len;
+		return;
+	}
+	e->bytes += len;
+}
+
+static void i915_error_puts(struct drm_i915_error_state_buf *e, const char *str)
+{
+	unsigned len;
+
+	if (e->err || !str)
+		return;
+
+	len = strlen(str);
+	if (!__i915_error_grow(e, len))
+		return;
+
+	GEM_BUG_ON(e->bytes + len > e->size);
+	memcpy(e->buf + e->bytes, str, len);
+	e->bytes += len;
+}
+
+#define err_printf(e, ...) i915_error_printf(e, __VA_ARGS__)
+#define err_puts(e, s) i915_error_puts(e, s)
+
+static void __i915_printfn_error(struct drm_printer *p, struct va_format *vaf)
+{
+	i915_error_vprintf(p->arg, vaf->fmt, *vaf->va);
+}
+
+static inline struct drm_printer
+i915_error_printer(struct drm_i915_error_state_buf *e)
+{
+	struct drm_printer p = {
+		.printfn = __i915_printfn_error,
+		.arg = e,
+	};
+	return p;
+}
+
+/* single threaded page allocator with a reserved stash for emergencies */
+static void pool_fini(struct pagevec *pv)
+{
+	pagevec_release(pv);
+}
+
+static int pool_refill(struct pagevec *pv, gfp_t gfp)
+{
+	while (pagevec_space(pv)) {
+		struct page *p;
+
+		p = alloc_page(gfp);
+		if (!p)
+			return -ENOMEM;
+
+		pagevec_add(pv, p);
+	}
+
+	return 0;
+}
+
+static int pool_init(struct pagevec *pv, gfp_t gfp)
+{
+	int err;
+
+	pagevec_init(pv);
+
+	err = pool_refill(pv, gfp);
+	if (err)
+		pool_fini(pv);
+
+	return err;
+}
+
+static void *pool_alloc(struct pagevec *pv, gfp_t gfp)
+{
+	struct page *p;
+
+	p = alloc_page(gfp);
+	if (!p && pagevec_count(pv))
+		p = pv->pages[--pv->nr];
+
+	return p ? page_address(p) : NULL;
+}
+
+static void pool_free(struct pagevec *pv, void *addr)
+{
+	struct page *p = virt_to_page(addr);
+
+	if (pagevec_space(pv))
+		pagevec_add(pv, p);
+	else
+		__free_page(p);
+}
+
+#ifdef CONFIG_DRM_I915_COMPRESS_ERROR
+
+struct i915_vma_compress {
+	struct pagevec pool;
+	struct z_stream_s zstream;
+	void *tmp;
+};
+
+static bool compress_init(struct i915_vma_compress *c)
+{
+	struct z_stream_s *zstream = &c->zstream;
+
+	if (pool_init(&c->pool, ALLOW_FAIL))
+		return false;
+
+	zstream->workspace =
+		kmalloc(zlib_deflate_workspacesize(MAX_WBITS, MAX_MEM_LEVEL),
+			ALLOW_FAIL);
+	if (!zstream->workspace) {
+		pool_fini(&c->pool);
+		return false;
+	}
+
+	c->tmp = NULL;
+	if (i915_has_memcpy_from_wc())
+		c->tmp = pool_alloc(&c->pool, ALLOW_FAIL);
+
+	return true;
+}
+
+static bool compress_start(struct i915_vma_compress *c)
+{
+	struct z_stream_s *zstream = &c->zstream;
+	void *workspace = zstream->workspace;
+
+	memset(zstream, 0, sizeof(*zstream));
+	zstream->workspace = workspace;
+
+	return zlib_deflateInit(zstream, Z_DEFAULT_COMPRESSION) == Z_OK;
+}
+
+static void *compress_next_page(struct i915_vma_compress *c,
+				struct i915_vma_coredump *dst)
+{
+	void *page_addr;
+	struct page *page;
+
+	page_addr = pool_alloc(&c->pool, ALLOW_FAIL);
+	if (!page_addr)
+		return ERR_PTR(-ENOMEM);
+
+	page = virt_to_page(page_addr);
+	list_add_tail(&page->lru, &dst->page_list);
+	return page_addr;
+}
+
+static int compress_page(struct i915_vma_compress *c,
+			 void *src,
+			 struct i915_vma_coredump *dst,
+			 bool wc)
+{
+	struct z_stream_s *zstream = &c->zstream;
+
+	zstream->next_in = src;
+	if (wc && c->tmp && i915_memcpy_from_wc(c->tmp, src, PAGE_SIZE))
+		zstream->next_in = c->tmp;
+	zstream->avail_in = PAGE_SIZE;
+
+	do {
+		if (zstream->avail_out == 0) {
+			zstream->next_out = compress_next_page(c, dst);
+			if (IS_ERR(zstream->next_out))
+				return PTR_ERR(zstream->next_out);
+
+			zstream->avail_out = PAGE_SIZE;
+		}
+
+		if (zlib_deflate(zstream, Z_NO_FLUSH) != Z_OK)
+			return -EIO;
+
+		cond_resched();
+	} while (zstream->avail_in);
+
+	/* Fallback to uncompressed if we increase size? */
+	if (0 && zstream->total_out > zstream->total_in)
+		return -E2BIG;
+
+	return 0;
+}
+
+static int compress_flush(struct i915_vma_compress *c,
+			  struct i915_vma_coredump *dst)
+{
+	struct z_stream_s *zstream = &c->zstream;
+
+	do {
+		switch (zlib_deflate(zstream, Z_FINISH)) {
+		case Z_OK: /* more space requested */
+			zstream->next_out = compress_next_page(c, dst);
+			if (IS_ERR(zstream->next_out))
+				return PTR_ERR(zstream->next_out);
+
+			zstream->avail_out = PAGE_SIZE;
+			break;
+
+		case Z_STREAM_END:
+			goto end;
+
+		default: /* any error */
+			return -EIO;
+		}
+	} while (1);
+
+end:
+	memset(zstream->next_out, 0, zstream->avail_out);
+	dst->unused = zstream->avail_out;
+	return 0;
+}
+
+static void compress_finish(struct i915_vma_compress *c)
+{
+	zlib_deflateEnd(&c->zstream);
+}
+
+static void compress_fini(struct i915_vma_compress *c)
+{
+	kfree(c->zstream.workspace);
+	if (c->tmp)
+		pool_free(&c->pool, c->tmp);
+	pool_fini(&c->pool);
+}
+
+static void err_compression_marker(struct drm_i915_error_state_buf *m)
+{
+	err_puts(m, ":");
+}
+
+#else
+
+struct i915_vma_compress {
+	struct pagevec pool;
+};
+
+static bool compress_init(struct i915_vma_compress *c)
+{
+	return pool_init(&c->pool, ALLOW_FAIL) == 0;
+}
+
+static bool compress_start(struct i915_vma_compress *c)
+{
+	return true;
+}
+
+static int compress_page(struct i915_vma_compress *c,
+			 void *src,
+			 struct i915_vma_coredump *dst,
+			 bool wc)
+{
+	void *ptr;
+
+	ptr = pool_alloc(&c->pool, ALLOW_FAIL);
+	if (!ptr)
+		return -ENOMEM;
+
+	if (!(wc && i915_memcpy_from_wc(ptr, src, PAGE_SIZE)))
+		memcpy(ptr, src, PAGE_SIZE);
+	list_add_tail(&virt_to_page(ptr)->lru, &dst->page_list);
+	cond_resched();
+
+	return 0;
+}
+
+static int compress_flush(struct i915_vma_compress *c,
+			  struct i915_vma_coredump *dst)
+{
+	return 0;
+}
+
+static void compress_finish(struct i915_vma_compress *c)
+{
+}
+
+static void compress_fini(struct i915_vma_compress *c)
+{
+	pool_fini(&c->pool);
+}
+
+static void err_compression_marker(struct drm_i915_error_state_buf *m)
+{
+	err_puts(m, "~");
+}
+
+#endif
+
+static void error_print_instdone(struct drm_i915_error_state_buf *m,
+				 const struct intel_engine_coredump *ee)
+{
+	int slice;
+	int subslice;
+	int iter;
+
+	err_printf(m, "  INSTDONE: 0x%08x\n",
+		   ee->instdone.instdone);
+
+	if (ee->engine->class != RENDER_CLASS || GRAPHICS_VER(m->i915) <= 3)
+		return;
+
+	err_printf(m, "  SC_INSTDONE: 0x%08x\n",
+		   ee->instdone.slice_common);
+
+	if (GRAPHICS_VER(m->i915) <= 6)
+		return;
+
+	for_each_ss_steering(iter, ee->engine->gt, slice, subslice)
+		err_printf(m, "  SAMPLER_INSTDONE[%d][%d]: 0x%08x\n",
+			   slice, subslice,
+			   ee->instdone.sampler[slice][subslice]);
+
+	for_each_ss_steering(iter, ee->engine->gt, slice, subslice)
+		err_printf(m, "  ROW_INSTDONE[%d][%d]: 0x%08x\n",
+			   slice, subslice,
+			   ee->instdone.row[slice][subslice]);
+
+	if (GRAPHICS_VER(m->i915) < 12)
+		return;
+
+	if (GRAPHICS_VER_FULL(m->i915) >= IP_VER(12, 55)) {
+		for_each_ss_steering(iter, ee->engine->gt, slice, subslice)
+			err_printf(m, "  GEOM_SVGUNIT_INSTDONE[%d][%d]: 0x%08x\n",
+				   slice, subslice,
+				   ee->instdone.geom_svg[slice][subslice]);
+	}
+
+	err_printf(m, "  SC_INSTDONE_EXTRA: 0x%08x\n",
+		   ee->instdone.slice_common_extra[0]);
+	err_printf(m, "  SC_INSTDONE_EXTRA2: 0x%08x\n",
+		   ee->instdone.slice_common_extra[1]);
+}
+
+static void error_print_request(struct drm_i915_error_state_buf *m,
+				const char *prefix,
+				const struct i915_request_coredump *erq)
+{
+	if (!erq->seqno)
+		return;
+
+	err_printf(m, "%s pid %d, seqno %8x:%08x%s%s, prio %d, head %08x, tail %08x\n",
+		   prefix, erq->pid, erq->context, erq->seqno,
+		   test_bit(DMA_FENCE_FLAG_SIGNALED_BIT,
+			    &erq->flags) ? "!" : "",
+		   test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT,
+			    &erq->flags) ? "+" : "",
+		   erq->sched_attr.priority,
+		   erq->head, erq->tail);
+}
+
+static void error_print_context(struct drm_i915_error_state_buf *m,
+				const char *header,
+				const struct i915_gem_context_coredump *ctx)
+{
+	err_printf(m, "%s%s[%d] prio %d, guilty %d active %d, runtime total %lluns, avg %lluns\n",
+		   header, ctx->comm, ctx->pid, ctx->sched_attr.priority,
+		   ctx->guilty, ctx->active,
+		   ctx->total_runtime, ctx->avg_runtime);
+}
+
+static struct i915_vma_coredump *
+__find_vma(struct i915_vma_coredump *vma, const char *name)
+{
+	while (vma) {
+		if (strcmp(vma->name, name) == 0)
+			return vma;
+		vma = vma->next;
+	}
+
+	return NULL;
+}
+
+struct i915_vma_coredump *
+intel_gpu_error_find_batch(const struct intel_engine_coredump *ee)
+{
+	return __find_vma(ee->vma, "batch");
+}
+
+static void error_print_engine(struct drm_i915_error_state_buf *m,
+			       const struct intel_engine_coredump *ee)
+{
+	struct i915_vma_coredump *batch;
+	int n;
+
+	err_printf(m, "%s command stream:\n", ee->engine->name);
+	err_printf(m, "  CCID:  0x%08x\n", ee->ccid);
+	err_printf(m, "  START: 0x%08x\n", ee->start);
+	err_printf(m, "  HEAD:  0x%08x [0x%08x]\n", ee->head, ee->rq_head);
+	err_printf(m, "  TAIL:  0x%08x [0x%08x, 0x%08x]\n",
+		   ee->tail, ee->rq_post, ee->rq_tail);
+	err_printf(m, "  CTL:   0x%08x\n", ee->ctl);
+	err_printf(m, "  MODE:  0x%08x\n", ee->mode);
+	err_printf(m, "  HWS:   0x%08x\n", ee->hws);
+	err_printf(m, "  ACTHD: 0x%08x %08x\n",
+		   (u32)(ee->acthd>>32), (u32)ee->acthd);
+	err_printf(m, "  IPEIR: 0x%08x\n", ee->ipeir);
+	err_printf(m, "  IPEHR: 0x%08x\n", ee->ipehr);
+	err_printf(m, "  ESR:   0x%08x\n", ee->esr);
+
+	error_print_instdone(m, ee);
+
+	batch = intel_gpu_error_find_batch(ee);
+	if (batch) {
+		u64 start = batch->gtt_offset;
+		u64 end = start + batch->gtt_size;
+
+		err_printf(m, "  batch: [0x%08x_%08x, 0x%08x_%08x]\n",
+			   upper_32_bits(start), lower_32_bits(start),
+			   upper_32_bits(end), lower_32_bits(end));
+	}
+	if (GRAPHICS_VER(m->i915) >= 4) {
+		err_printf(m, "  BBADDR: 0x%08x_%08x\n",
+			   (u32)(ee->bbaddr>>32), (u32)ee->bbaddr);
+		err_printf(m, "  BB_STATE: 0x%08x\n", ee->bbstate);
+		err_printf(m, "  INSTPS: 0x%08x\n", ee->instps);
+	}
+	err_printf(m, "  INSTPM: 0x%08x\n", ee->instpm);
+	err_printf(m, "  FADDR: 0x%08x %08x\n", upper_32_bits(ee->faddr),
+		   lower_32_bits(ee->faddr));
+	if (GRAPHICS_VER(m->i915) >= 6) {
+		err_printf(m, "  RC PSMI: 0x%08x\n", ee->rc_psmi);
+		err_printf(m, "  FAULT_REG: 0x%08x\n", ee->fault_reg);
+	}
+	if (GRAPHICS_VER(m->i915) >= 11) {
+		err_printf(m, "  NOPID: 0x%08x\n", ee->nopid);
+		err_printf(m, "  EXCC: 0x%08x\n", ee->excc);
+		err_printf(m, "  CMD_CCTL: 0x%08x\n", ee->cmd_cctl);
+		err_printf(m, "  CSCMDOP: 0x%08x\n", ee->cscmdop);
+		err_printf(m, "  CTX_SR_CTL: 0x%08x\n", ee->ctx_sr_ctl);
+		err_printf(m, "  DMA_FADDR_HI: 0x%08x\n", ee->dma_faddr_hi);
+		err_printf(m, "  DMA_FADDR_LO: 0x%08x\n", ee->dma_faddr_lo);
+	}
+	if (HAS_PPGTT(m->i915)) {
+		err_printf(m, "  GFX_MODE: 0x%08x\n", ee->vm_info.gfx_mode);
+
+		if (GRAPHICS_VER(m->i915) >= 8) {
+			int i;
+			for (i = 0; i < 4; i++)
+				err_printf(m, "  PDP%d: 0x%016llx\n",
+					   i, ee->vm_info.pdp[i]);
+		} else {
+			err_printf(m, "  PP_DIR_BASE: 0x%08x\n",
+				   ee->vm_info.pp_dir_base);
+		}
+	}
+
+	for (n = 0; n < ee->num_ports; n++) {
+		err_printf(m, "  ELSP[%d]:", n);
+		error_print_request(m, " ", &ee->execlist[n]);
+	}
+}
+
+void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...)
+{
+	va_list args;
+
+	va_start(args, f);
+	i915_error_vprintf(e, f, args);
+	va_end(args);
+}
+
+void intel_gpu_error_print_vma(struct drm_i915_error_state_buf *m,
+			       const struct intel_engine_cs *engine,
+			       const struct i915_vma_coredump *vma)
+{
+	char out[ASCII85_BUFSZ];
+	struct page *page;
+
+	if (!vma)
+		return;
+
+	err_printf(m, "%s --- %s = 0x%08x %08x\n",
+		   engine ? engine->name : "global", vma->name,
+		   upper_32_bits(vma->gtt_offset),
+		   lower_32_bits(vma->gtt_offset));
+
+	if (vma->gtt_page_sizes > I915_GTT_PAGE_SIZE_4K)
+		err_printf(m, "gtt_page_sizes = 0x%08x\n", vma->gtt_page_sizes);
+
+	err_compression_marker(m);
+	list_for_each_entry(page, &vma->page_list, lru) {
+		int i, len;
+		const u32 *addr = page_address(page);
+
+		len = PAGE_SIZE;
+		if (page == list_last_entry(&vma->page_list, typeof(*page), lru))
+			len -= vma->unused;
+		len = ascii85_encode_len(len);
+
+		for (i = 0; i < len; i++)
+			err_puts(m, ascii85_encode(addr[i], out));
+	}
+	err_puts(m, "\n");
+}
+
+static void err_print_capabilities(struct drm_i915_error_state_buf *m,
+				   struct i915_gpu_coredump *error)
+{
+	struct drm_printer p = i915_error_printer(m);
+
+	intel_device_info_print(&error->device_info, &error->runtime_info, &p);
+	intel_driver_caps_print(&error->driver_caps, &p);
+}
+
+static void err_print_params(struct drm_i915_error_state_buf *m,
+			     const struct i915_params *params)
+{
+	struct drm_printer p = i915_error_printer(m);
+
+	i915_params_dump(params, &p);
+}
+
+static void err_print_pciid(struct drm_i915_error_state_buf *m,
+			    struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+
+	err_printf(m, "PCI ID: 0x%04x\n", pdev->device);
+	err_printf(m, "PCI Revision: 0x%02x\n", pdev->revision);
+	err_printf(m, "PCI Subsystem: %04x:%04x\n",
+		   pdev->subsystem_vendor,
+		   pdev->subsystem_device);
+}
+
+static void err_print_guc_ctb(struct drm_i915_error_state_buf *m,
+			      const char *name,
+			      const struct intel_ctb_coredump *ctb)
+{
+	if (!ctb->size)
+		return;
+
+	err_printf(m, "GuC %s CTB: raw: 0x%08X, 0x%08X/%08X, cached: 0x%08X/%08X, desc = 0x%08X, buf = 0x%08X x 0x%08X\n",
+		   name, ctb->raw_status, ctb->raw_head, ctb->raw_tail,
+		   ctb->head, ctb->tail, ctb->desc_offset, ctb->cmds_offset, ctb->size);
+}
+
+static void err_print_uc(struct drm_i915_error_state_buf *m,
+			 const struct intel_uc_coredump *error_uc)
+{
+	struct drm_printer p = i915_error_printer(m);
+
+	intel_uc_fw_dump(&error_uc->guc_fw, &p);
+	intel_uc_fw_dump(&error_uc->huc_fw, &p);
+	err_printf(m, "GuC timestamp: 0x%08x\n", error_uc->guc.timestamp);
+	intel_gpu_error_print_vma(m, NULL, error_uc->guc.vma_log);
+	err_printf(m, "GuC CTB fence: %d\n", error_uc->guc.last_fence);
+	err_print_guc_ctb(m, "Send", error_uc->guc.ctb + 0);
+	err_print_guc_ctb(m, "Recv", error_uc->guc.ctb + 1);
+	intel_gpu_error_print_vma(m, NULL, error_uc->guc.vma_ctb);
+}
+
+static void err_free_sgl(struct scatterlist *sgl)
+{
+	while (sgl) {
+		struct scatterlist *sg;
+
+		for (sg = sgl; !sg_is_chain(sg); sg++) {
+			kfree(sg_virt(sg));
+			if (sg_is_last(sg))
+				break;
+		}
+
+		sg = sg_is_last(sg) ? NULL : sg_chain_ptr(sg);
+		free_page((unsigned long)sgl);
+		sgl = sg;
+	}
+}
+
+static void err_print_gt_info(struct drm_i915_error_state_buf *m,
+			      struct intel_gt_coredump *gt)
+{
+	struct drm_printer p = i915_error_printer(m);
+
+	intel_gt_info_print(&gt->info, &p);
+	intel_sseu_print_topology(gt->_gt->i915, &gt->info.sseu, &p);
+}
+
+static void err_print_gt_display(struct drm_i915_error_state_buf *m,
+				 struct intel_gt_coredump *gt)
+{
+	err_printf(m, "IER: 0x%08x\n", gt->ier);
+	err_printf(m, "DERRMR: 0x%08x\n", gt->derrmr);
+}
+
+static void err_print_gt_global_nonguc(struct drm_i915_error_state_buf *m,
+				       struct intel_gt_coredump *gt)
+{
+	int i;
+
+	err_printf(m, "GT awake: %s\n", str_yes_no(gt->awake));
+	err_printf(m, "CS timestamp frequency: %u Hz, %d ns\n",
+		   gt->clock_frequency, gt->clock_period_ns);
+	err_printf(m, "EIR: 0x%08x\n", gt->eir);
+	err_printf(m, "PGTBL_ER: 0x%08x\n", gt->pgtbl_er);
+
+	for (i = 0; i < gt->ngtier; i++)
+		err_printf(m, "GTIER[%d]: 0x%08x\n", i, gt->gtier[i]);
+}
+
+static void err_print_gt_global(struct drm_i915_error_state_buf *m,
+				struct intel_gt_coredump *gt)
+{
+	err_printf(m, "FORCEWAKE: 0x%08x\n", gt->forcewake);
+
+	if (IS_GRAPHICS_VER(m->i915, 6, 11)) {
+		err_printf(m, "ERROR: 0x%08x\n", gt->error);
+		err_printf(m, "DONE_REG: 0x%08x\n", gt->done_reg);
+	}
+
+	if (GRAPHICS_VER(m->i915) >= 8)
+		err_printf(m, "FAULT_TLB_DATA: 0x%08x 0x%08x\n",
+			   gt->fault_data1, gt->fault_data0);
+
+	if (GRAPHICS_VER(m->i915) == 7)
+		err_printf(m, "ERR_INT: 0x%08x\n", gt->err_int);
+
+	if (IS_GRAPHICS_VER(m->i915, 8, 11))
+		err_printf(m, "GTT_CACHE_EN: 0x%08x\n", gt->gtt_cache);
+
+	if (GRAPHICS_VER(m->i915) == 12)
+		err_printf(m, "AUX_ERR_DBG: 0x%08x\n", gt->aux_err);
+
+	if (GRAPHICS_VER(m->i915) >= 12) {
+		int i;
+
+		for (i = 0; i < I915_MAX_SFC; i++) {
+			/*
+			 * SFC_DONE resides in the VD forcewake domain, so it
+			 * only exists if the corresponding VCS engine is
+			 * present.
+			 */
+			if ((gt->_gt->info.sfc_mask & BIT(i)) == 0 ||
+			    !HAS_ENGINE(gt->_gt, _VCS(i * 2)))
+				continue;
+
+			err_printf(m, "  SFC_DONE[%d]: 0x%08x\n", i,
+				   gt->sfc_done[i]);
+		}
+
+		err_printf(m, "  GAM_DONE: 0x%08x\n", gt->gam_done);
+	}
+}
+
+static void err_print_gt_fences(struct drm_i915_error_state_buf *m,
+				struct intel_gt_coredump *gt)
+{
+	int i;
+
+	for (i = 0; i < gt->nfence; i++)
+		err_printf(m, "  fence[%d] = %08llx\n", i, gt->fence[i]);
+}
+
+static void err_print_gt_engines(struct drm_i915_error_state_buf *m,
+				 struct intel_gt_coredump *gt)
+{
+	const struct intel_engine_coredump *ee;
+
+	for (ee = gt->engine; ee; ee = ee->next) {
+		const struct i915_vma_coredump *vma;
+
+		if (ee->guc_capture_node)
+			intel_guc_capture_print_engine_node(m, ee);
+		else
+			error_print_engine(m, ee);
+
+		err_printf(m, "  hung: %u\n", ee->hung);
+		err_printf(m, "  engine reset count: %u\n", ee->reset_count);
+		error_print_context(m, "  Active context: ", &ee->context);
+
+		for (vma = ee->vma; vma; vma = vma->next)
+			intel_gpu_error_print_vma(m, ee->engine, vma);
+	}
+
+}
+
+static void __err_print_to_sgl(struct drm_i915_error_state_buf *m,
+			       struct i915_gpu_coredump *error)
+{
+	const struct intel_engine_coredump *ee;
+	struct timespec64 ts;
+
+	if (*error->error_msg)
+		err_printf(m, "%s\n", error->error_msg);
+	err_printf(m, "Kernel: %s %s\n",
+		   init_utsname()->release,
+		   init_utsname()->machine);
+	err_printf(m, "Driver: %s\n", DRIVER_DATE);
+	ts = ktime_to_timespec64(error->time);
+	err_printf(m, "Time: %lld s %ld us\n",
+		   (s64)ts.tv_sec, ts.tv_nsec / NSEC_PER_USEC);
+	ts = ktime_to_timespec64(error->boottime);
+	err_printf(m, "Boottime: %lld s %ld us\n",
+		   (s64)ts.tv_sec, ts.tv_nsec / NSEC_PER_USEC);
+	ts = ktime_to_timespec64(error->uptime);
+	err_printf(m, "Uptime: %lld s %ld us\n",
+		   (s64)ts.tv_sec, ts.tv_nsec / NSEC_PER_USEC);
+	err_printf(m, "Capture: %lu jiffies; %d ms ago\n",
+		   error->capture, jiffies_to_msecs(jiffies - error->capture));
+
+	for (ee = error->gt ? error->gt->engine : NULL; ee; ee = ee->next)
+		err_printf(m, "Active process (on ring %s): %s [%d]\n",
+			   ee->engine->name,
+			   ee->context.comm,
+			   ee->context.pid);
+
+	err_printf(m, "Reset count: %u\n", error->reset_count);
+	err_printf(m, "Suspend count: %u\n", error->suspend_count);
+	err_printf(m, "Platform: %s\n", intel_platform_name(error->device_info.platform));
+	err_printf(m, "Subplatform: 0x%x\n",
+		   intel_subplatform(&error->runtime_info,
+				     error->device_info.platform));
+	err_print_pciid(m, m->i915);
+
+	err_printf(m, "IOMMU enabled?: %d\n", error->iommu);
+
+	intel_dmc_print_error_state(m, m->i915);
+
+	err_printf(m, "RPM wakelock: %s\n", str_yes_no(error->wakelock));
+	err_printf(m, "PM suspended: %s\n", str_yes_no(error->suspended));
+
+	if (error->gt) {
+		bool print_guc_capture = false;
+
+		if (error->gt->uc && error->gt->uc->guc.is_guc_capture)
+			print_guc_capture = true;
+
+		err_print_gt_display(m, error->gt);
+		err_print_gt_global_nonguc(m, error->gt);
+		err_print_gt_fences(m, error->gt);
+
+		/*
+		 * GuC dumped global, eng-class and eng-instance registers together
+		 * as part of engine state dump so we print in err_print_gt_engines
+		 */
+		if (!print_guc_capture)
+			err_print_gt_global(m, error->gt);
+
+		err_print_gt_engines(m, error->gt);
+
+		if (error->gt->uc)
+			err_print_uc(m, error->gt->uc);
+
+		err_print_gt_info(m, error->gt);
+	}
+
+	if (error->overlay)
+		intel_overlay_print_error_state(m, error->overlay);
+
+	err_print_capabilities(m, error);
+	err_print_params(m, &error->params);
+}
+
+static int err_print_to_sgl(struct i915_gpu_coredump *error)
+{
+	struct drm_i915_error_state_buf m;
+
+	if (IS_ERR(error))
+		return PTR_ERR(error);
+
+	if (READ_ONCE(error->sgl))
+		return 0;
+
+	memset(&m, 0, sizeof(m));
+	m.i915 = error->i915;
+
+	__err_print_to_sgl(&m, error);
+
+	if (m.buf) {
+		__sg_set_buf(m.cur++, m.buf, m.bytes, m.iter);
+		m.bytes = 0;
+		m.buf = NULL;
+	}
+	if (m.cur) {
+		GEM_BUG_ON(m.end < m.cur);
+		sg_mark_end(m.cur - 1);
+	}
+	GEM_BUG_ON(m.sgl && !m.cur);
+
+	if (m.err) {
+		err_free_sgl(m.sgl);
+		return m.err;
+	}
+
+	if (cmpxchg(&error->sgl, NULL, m.sgl))
+		err_free_sgl(m.sgl);
+
+	return 0;
+}
+
+ssize_t i915_gpu_coredump_copy_to_buffer(struct i915_gpu_coredump *error,
+					 char *buf, loff_t off, size_t rem)
+{
+	struct scatterlist *sg;
+	size_t count;
+	loff_t pos;
+	int err;
+
+	if (!error || !rem)
+		return 0;
+
+	err = err_print_to_sgl(error);
+	if (err)
+		return err;
+
+	sg = READ_ONCE(error->fit);
+	if (!sg || off < sg->dma_address)
+		sg = error->sgl;
+	if (!sg)
+		return 0;
+
+	pos = sg->dma_address;
+	count = 0;
+	do {
+		size_t len, start;
+
+		if (sg_is_chain(sg)) {
+			sg = sg_chain_ptr(sg);
+			GEM_BUG_ON(sg_is_chain(sg));
+		}
+
+		len = sg->length;
+		if (pos + len <= off) {
+			pos += len;
+			continue;
+		}
+
+		start = sg->offset;
+		if (pos < off) {
+			GEM_BUG_ON(off - pos > len);
+			len -= off - pos;
+			start += off - pos;
+			pos = off;
+		}
+
+		len = min(len, rem);
+		GEM_BUG_ON(!len || len > sg->length);
+
+		memcpy(buf, page_address(sg_page(sg)) + start, len);
+
+		count += len;
+		pos += len;
+
+		buf += len;
+		rem -= len;
+		if (!rem) {
+			WRITE_ONCE(error->fit, sg);
+			break;
+		}
+	} while (!sg_is_last(sg++));
+
+	return count;
+}
+
+static void i915_vma_coredump_free(struct i915_vma_coredump *vma)
+{
+	while (vma) {
+		struct i915_vma_coredump *next = vma->next;
+		struct page *page, *n;
+
+		list_for_each_entry_safe(page, n, &vma->page_list, lru) {
+			list_del_init(&page->lru);
+			__free_page(page);
+		}
+
+		kfree(vma);
+		vma = next;
+	}
+}
+
+static void cleanup_params(struct i915_gpu_coredump *error)
+{
+	i915_params_free(&error->params);
+}
+
+static void cleanup_uc(struct intel_uc_coredump *uc)
+{
+	kfree(uc->guc_fw.file_selected.path);
+	kfree(uc->huc_fw.file_selected.path);
+	kfree(uc->guc_fw.file_wanted.path);
+	kfree(uc->huc_fw.file_wanted.path);
+	i915_vma_coredump_free(uc->guc.vma_log);
+	i915_vma_coredump_free(uc->guc.vma_ctb);
+
+	kfree(uc);
+}
+
+static void cleanup_gt(struct intel_gt_coredump *gt)
+{
+	while (gt->engine) {
+		struct intel_engine_coredump *ee = gt->engine;
+
+		gt->engine = ee->next;
+
+		i915_vma_coredump_free(ee->vma);
+		intel_guc_capture_free_node(ee);
+		kfree(ee);
+	}
+
+	if (gt->uc)
+		cleanup_uc(gt->uc);
+
+	kfree(gt);
+}
+
+void __i915_gpu_coredump_free(struct kref *error_ref)
+{
+	struct i915_gpu_coredump *error =
+		container_of(error_ref, typeof(*error), ref);
+
+	while (error->gt) {
+		struct intel_gt_coredump *gt = error->gt;
+
+		error->gt = gt->next;
+		cleanup_gt(gt);
+	}
+
+	kfree(error->overlay);
+
+	cleanup_params(error);
+
+	err_free_sgl(error->sgl);
+	kfree(error);
+}
+
+static struct i915_vma_coredump *
+i915_vma_coredump_create(const struct intel_gt *gt,
+			 const struct i915_vma_resource *vma_res,
+			 struct i915_vma_compress *compress,
+			 const char *name)
+
+{
+	struct i915_ggtt *ggtt = gt->ggtt;
+	const u64 slot = ggtt->error_capture.start;
+	struct i915_vma_coredump *dst;
+	struct sgt_iter iter;
+	int ret;
+
+	might_sleep();
+
+	if (!vma_res || !vma_res->bi.pages || !compress)
+		return NULL;
+
+	dst = kmalloc(sizeof(*dst), ALLOW_FAIL);
+	if (!dst)
+		return NULL;
+
+	if (!compress_start(compress)) {
+		kfree(dst);
+		return NULL;
+	}
+
+	INIT_LIST_HEAD(&dst->page_list);
+	strcpy(dst->name, name);
+	dst->next = NULL;
+
+	dst->gtt_offset = vma_res->start;
+	dst->gtt_size = vma_res->node_size;
+	dst->gtt_page_sizes = vma_res->page_sizes_gtt;
+	dst->unused = 0;
+
+	ret = -EINVAL;
+	if (drm_mm_node_allocated(&ggtt->error_capture)) {
+		void __iomem *s;
+		dma_addr_t dma;
+
+		for_each_sgt_daddr(dma, iter, vma_res->bi.pages) {
+			mutex_lock(&ggtt->error_mutex);
+			if (ggtt->vm.raw_insert_page)
+				ggtt->vm.raw_insert_page(&ggtt->vm, dma, slot,
+							 I915_CACHE_NONE, 0);
+			else
+				ggtt->vm.insert_page(&ggtt->vm, dma, slot,
+						     I915_CACHE_NONE, 0);
+			mb();
+
+			s = io_mapping_map_wc(&ggtt->iomap, slot, PAGE_SIZE);
+			ret = compress_page(compress,
+					    (void  __force *)s, dst,
+					    true);
+			io_mapping_unmap(s);
+
+			mb();
+			ggtt->vm.clear_range(&ggtt->vm, slot, PAGE_SIZE);
+			mutex_unlock(&ggtt->error_mutex);
+			if (ret)
+				break;
+		}
+	} else if (vma_res->bi.lmem) {
+		struct intel_memory_region *mem = vma_res->mr;
+		dma_addr_t dma;
+
+		for_each_sgt_daddr(dma, iter, vma_res->bi.pages) {
+			dma_addr_t offset = dma - mem->region.start;
+			void __iomem *s;
+
+			if (offset + PAGE_SIZE > mem->io_size) {
+				ret = -EINVAL;
+				break;
+			}
+
+			s = io_mapping_map_wc(&mem->iomap, offset, PAGE_SIZE);
+			ret = compress_page(compress,
+					    (void __force *)s, dst,
+					    true);
+			io_mapping_unmap(s);
+			if (ret)
+				break;
+		}
+	} else {
+		struct page *page;
+
+		for_each_sgt_page(page, iter, vma_res->bi.pages) {
+			void *s;
+
+			drm_clflush_pages(&page, 1);
+
+			s = kmap(page);
+			ret = compress_page(compress, s, dst, false);
+			kunmap(page);
+
+			drm_clflush_pages(&page, 1);
+
+			if (ret)
+				break;
+		}
+	}
+
+	if (ret || compress_flush(compress, dst)) {
+		struct page *page, *n;
+
+		list_for_each_entry_safe_reverse(page, n, &dst->page_list, lru) {
+			list_del_init(&page->lru);
+			pool_free(&compress->pool, page_address(page));
+		}
+
+		kfree(dst);
+		dst = NULL;
+	}
+	compress_finish(compress);
+
+	return dst;
+}
+
+static void gt_record_fences(struct intel_gt_coredump *gt)
+{
+	struct i915_ggtt *ggtt = gt->_gt->ggtt;
+	struct intel_uncore *uncore = gt->_gt->uncore;
+	int i;
+
+	if (GRAPHICS_VER(uncore->i915) >= 6) {
+		for (i = 0; i < ggtt->num_fences; i++)
+			gt->fence[i] =
+				intel_uncore_read64(uncore,
+						    FENCE_REG_GEN6_LO(i));
+	} else if (GRAPHICS_VER(uncore->i915) >= 4) {
+		for (i = 0; i < ggtt->num_fences; i++)
+			gt->fence[i] =
+				intel_uncore_read64(uncore,
+						    FENCE_REG_965_LO(i));
+	} else {
+		for (i = 0; i < ggtt->num_fences; i++)
+			gt->fence[i] =
+				intel_uncore_read(uncore, FENCE_REG(i));
+	}
+	gt->nfence = i;
+}
+
+static void engine_record_registers(struct intel_engine_coredump *ee)
+{
+	const struct intel_engine_cs *engine = ee->engine;
+	struct drm_i915_private *i915 = engine->i915;
+
+	if (GRAPHICS_VER(i915) >= 6) {
+		ee->rc_psmi = ENGINE_READ(engine, RING_PSMI_CTL);
+
+		if (GRAPHICS_VER(i915) >= 12)
+			ee->fault_reg = intel_uncore_read(engine->uncore,
+							  GEN12_RING_FAULT_REG);
+		else if (GRAPHICS_VER(i915) >= 8)
+			ee->fault_reg = intel_uncore_read(engine->uncore,
+							  GEN8_RING_FAULT_REG);
+		else
+			ee->fault_reg = GEN6_RING_FAULT_REG_READ(engine);
+	}
+
+	if (GRAPHICS_VER(i915) >= 4) {
+		ee->esr = ENGINE_READ(engine, RING_ESR);
+		ee->faddr = ENGINE_READ(engine, RING_DMA_FADD);
+		ee->ipeir = ENGINE_READ(engine, RING_IPEIR);
+		ee->ipehr = ENGINE_READ(engine, RING_IPEHR);
+		ee->instps = ENGINE_READ(engine, RING_INSTPS);
+		ee->bbaddr = ENGINE_READ(engine, RING_BBADDR);
+		ee->ccid = ENGINE_READ(engine, CCID);
+		if (GRAPHICS_VER(i915) >= 8) {
+			ee->faddr |= (u64)ENGINE_READ(engine, RING_DMA_FADD_UDW) << 32;
+			ee->bbaddr |= (u64)ENGINE_READ(engine, RING_BBADDR_UDW) << 32;
+		}
+		ee->bbstate = ENGINE_READ(engine, RING_BBSTATE);
+	} else {
+		ee->faddr = ENGINE_READ(engine, DMA_FADD_I8XX);
+		ee->ipeir = ENGINE_READ(engine, IPEIR);
+		ee->ipehr = ENGINE_READ(engine, IPEHR);
+	}
+
+	if (GRAPHICS_VER(i915) >= 11) {
+		ee->cmd_cctl = ENGINE_READ(engine, RING_CMD_CCTL);
+		ee->cscmdop = ENGINE_READ(engine, RING_CSCMDOP);
+		ee->ctx_sr_ctl = ENGINE_READ(engine, RING_CTX_SR_CTL);
+		ee->dma_faddr_hi = ENGINE_READ(engine, RING_DMA_FADD_UDW);
+		ee->dma_faddr_lo = ENGINE_READ(engine, RING_DMA_FADD);
+		ee->nopid = ENGINE_READ(engine, RING_NOPID);
+		ee->excc = ENGINE_READ(engine, RING_EXCC);
+	}
+
+	intel_engine_get_instdone(engine, &ee->instdone);
+
+	ee->instpm = ENGINE_READ(engine, RING_INSTPM);
+	ee->acthd = intel_engine_get_active_head(engine);
+	ee->start = ENGINE_READ(engine, RING_START);
+	ee->head = ENGINE_READ(engine, RING_HEAD);
+	ee->tail = ENGINE_READ(engine, RING_TAIL);
+	ee->ctl = ENGINE_READ(engine, RING_CTL);
+	if (GRAPHICS_VER(i915) > 2)
+		ee->mode = ENGINE_READ(engine, RING_MI_MODE);
+
+	if (!HWS_NEEDS_PHYSICAL(i915)) {
+		i915_reg_t mmio;
+
+		if (GRAPHICS_VER(i915) == 7) {
+			switch (engine->id) {
+			default:
+				MISSING_CASE(engine->id);
+				fallthrough;
+			case RCS0:
+				mmio = RENDER_HWS_PGA_GEN7;
+				break;
+			case BCS0:
+				mmio = BLT_HWS_PGA_GEN7;
+				break;
+			case VCS0:
+				mmio = BSD_HWS_PGA_GEN7;
+				break;
+			case VECS0:
+				mmio = VEBOX_HWS_PGA_GEN7;
+				break;
+			}
+		} else if (GRAPHICS_VER(engine->i915) == 6) {
+			mmio = RING_HWS_PGA_GEN6(engine->mmio_base);
+		} else {
+			/* XXX: gen8 returns to sanity */
+			mmio = RING_HWS_PGA(engine->mmio_base);
+		}
+
+		ee->hws = intel_uncore_read(engine->uncore, mmio);
+	}
+
+	ee->reset_count = i915_reset_engine_count(&i915->gpu_error, engine);
+
+	if (HAS_PPGTT(i915)) {
+		int i;
+
+		ee->vm_info.gfx_mode = ENGINE_READ(engine, RING_MODE_GEN7);
+
+		if (GRAPHICS_VER(i915) == 6) {
+			ee->vm_info.pp_dir_base =
+				ENGINE_READ(engine, RING_PP_DIR_BASE_READ);
+		} else if (GRAPHICS_VER(i915) == 7) {
+			ee->vm_info.pp_dir_base =
+				ENGINE_READ(engine, RING_PP_DIR_BASE);
+		} else if (GRAPHICS_VER(i915) >= 8) {
+			u32 base = engine->mmio_base;
+
+			for (i = 0; i < 4; i++) {
+				ee->vm_info.pdp[i] =
+					intel_uncore_read(engine->uncore,
+							  GEN8_RING_PDP_UDW(base, i));
+				ee->vm_info.pdp[i] <<= 32;
+				ee->vm_info.pdp[i] |=
+					intel_uncore_read(engine->uncore,
+							  GEN8_RING_PDP_LDW(base, i));
+			}
+		}
+	}
+}
+
+static void record_request(const struct i915_request *request,
+			   struct i915_request_coredump *erq)
+{
+	erq->flags = request->fence.flags;
+	erq->context = request->fence.context;
+	erq->seqno = request->fence.seqno;
+	erq->sched_attr = request->sched.attr;
+	erq->head = request->head;
+	erq->tail = request->tail;
+
+	erq->pid = 0;
+	rcu_read_lock();
+	if (!intel_context_is_closed(request->context)) {
+		const struct i915_gem_context *ctx;
+
+		ctx = rcu_dereference(request->context->gem_context);
+		if (ctx)
+			erq->pid = pid_nr(ctx->pid);
+	}
+	rcu_read_unlock();
+}
+
+static void engine_record_execlists(struct intel_engine_coredump *ee)
+{
+	const struct intel_engine_execlists * const el = &ee->engine->execlists;
+	struct i915_request * const *port = el->active;
+	unsigned int n = 0;
+
+	while (*port)
+		record_request(*port++, &ee->execlist[n++]);
+
+	ee->num_ports = n;
+}
+
+static bool record_context(struct i915_gem_context_coredump *e,
+			   const struct i915_request *rq)
+{
+	struct i915_gem_context *ctx;
+	struct task_struct *task;
+	bool simulated;
+
+	rcu_read_lock();
+	ctx = rcu_dereference(rq->context->gem_context);
+	if (ctx && !kref_get_unless_zero(&ctx->ref))
+		ctx = NULL;
+	rcu_read_unlock();
+	if (!ctx)
+		return true;
+
+	rcu_read_lock();
+	task = pid_task(ctx->pid, PIDTYPE_PID);
+	if (task) {
+		strcpy(e->comm, task->comm);
+		e->pid = task->pid;
+	}
+	rcu_read_unlock();
+
+	e->sched_attr = ctx->sched;
+	e->guilty = atomic_read(&ctx->guilty_count);
+	e->active = atomic_read(&ctx->active_count);
+
+	e->total_runtime = intel_context_get_total_runtime_ns(rq->context);
+	e->avg_runtime = intel_context_get_avg_runtime_ns(rq->context);
+
+	simulated = i915_gem_context_no_error_capture(ctx);
+
+	i915_gem_context_put(ctx);
+	return simulated;
+}
+
+struct intel_engine_capture_vma {
+	struct intel_engine_capture_vma *next;
+	struct i915_vma_resource *vma_res;
+	char name[16];
+	bool lockdep_cookie;
+};
+
+static struct intel_engine_capture_vma *
+capture_vma_snapshot(struct intel_engine_capture_vma *next,
+		     struct i915_vma_resource *vma_res,
+		     gfp_t gfp, const char *name)
+{
+	struct intel_engine_capture_vma *c;
+
+	if (!vma_res)
+		return next;
+
+	c = kmalloc(sizeof(*c), gfp);
+	if (!c)
+		return next;
+
+	if (!i915_vma_resource_hold(vma_res, &c->lockdep_cookie)) {
+		kfree(c);
+		return next;
+	}
+
+	strcpy(c->name, name);
+	c->vma_res = i915_vma_resource_get(vma_res);
+
+	c->next = next;
+	return c;
+}
+
+static struct intel_engine_capture_vma *
+capture_vma(struct intel_engine_capture_vma *next,
+	    struct i915_vma *vma,
+	    const char *name,
+	    gfp_t gfp)
+{
+	if (!vma)
+		return next;
+
+	/*
+	 * If the vma isn't pinned, then the vma should be snapshotted
+	 * to a struct i915_vma_snapshot at command submission time.
+	 * Not here.
+	 */
+	if (GEM_WARN_ON(!i915_vma_is_pinned(vma)))
+		return next;
+
+	next = capture_vma_snapshot(next, vma->resource, gfp, name);
+
+	return next;
+}
+
+static struct intel_engine_capture_vma *
+capture_user(struct intel_engine_capture_vma *capture,
+	     const struct i915_request *rq,
+	     gfp_t gfp)
+{
+	struct i915_capture_list *c;
+
+	for (c = rq->capture_list; c; c = c->next)
+		capture = capture_vma_snapshot(capture, c->vma_res, gfp,
+					       "user");
+
+	return capture;
+}
+
+static void add_vma(struct intel_engine_coredump *ee,
+		    struct i915_vma_coredump *vma)
+{
+	if (vma) {
+		vma->next = ee->vma;
+		ee->vma = vma;
+	}
+}
+
+static struct i915_vma_coredump *
+create_vma_coredump(const struct intel_gt *gt, struct i915_vma *vma,
+		    const char *name, struct i915_vma_compress *compress)
+{
+	struct i915_vma_coredump *ret = NULL;
+	struct i915_vma_resource *vma_res;
+	bool lockdep_cookie;
+
+	if (!vma)
+		return NULL;
+
+	vma_res = vma->resource;
+
+	if (i915_vma_resource_hold(vma_res, &lockdep_cookie)) {
+		ret = i915_vma_coredump_create(gt, vma_res, compress, name);
+		i915_vma_resource_unhold(vma_res, lockdep_cookie);
+	}
+
+	return ret;
+}
+
+static void add_vma_coredump(struct intel_engine_coredump *ee,
+			     const struct intel_gt *gt,
+			     struct i915_vma *vma,
+			     const char *name,
+			     struct i915_vma_compress *compress)
+{
+	add_vma(ee, create_vma_coredump(gt, vma, name, compress));
+}
+
+struct intel_engine_coredump *
+intel_engine_coredump_alloc(struct intel_engine_cs *engine, gfp_t gfp, u32 dump_flags)
+{
+	struct intel_engine_coredump *ee;
+
+	ee = kzalloc(sizeof(*ee), gfp);
+	if (!ee)
+		return NULL;
+
+	ee->engine = engine;
+
+	if (!(dump_flags & CORE_DUMP_FLAG_IS_GUC_CAPTURE)) {
+		engine_record_registers(ee);
+		engine_record_execlists(ee);
+	}
+
+	return ee;
+}
+
+struct intel_engine_capture_vma *
+intel_engine_coredump_add_request(struct intel_engine_coredump *ee,
+				  struct i915_request *rq,
+				  gfp_t gfp)
+{
+	struct intel_engine_capture_vma *vma = NULL;
+
+	ee->simulated |= record_context(&ee->context, rq);
+	if (ee->simulated)
+		return NULL;
+
+	/*
+	 * We need to copy these to an anonymous buffer
+	 * as the simplest method to avoid being overwritten
+	 * by userspace.
+	 */
+	vma = capture_vma_snapshot(vma, rq->batch_res, gfp, "batch");
+	vma = capture_user(vma, rq, gfp);
+	vma = capture_vma(vma, rq->ring->vma, "ring", gfp);
+	vma = capture_vma(vma, rq->context->state, "HW context", gfp);
+
+	ee->rq_head = rq->head;
+	ee->rq_post = rq->postfix;
+	ee->rq_tail = rq->tail;
+
+	return vma;
+}
+
+void
+intel_engine_coredump_add_vma(struct intel_engine_coredump *ee,
+			      struct intel_engine_capture_vma *capture,
+			      struct i915_vma_compress *compress)
+{
+	const struct intel_engine_cs *engine = ee->engine;
+
+	while (capture) {
+		struct intel_engine_capture_vma *this = capture;
+		struct i915_vma_resource *vma_res = this->vma_res;
+
+		add_vma(ee,
+			i915_vma_coredump_create(engine->gt, vma_res,
+						 compress, this->name));
+
+		i915_vma_resource_unhold(vma_res, this->lockdep_cookie);
+		i915_vma_resource_put(vma_res);
+
+		capture = this->next;
+		kfree(this);
+	}
+
+	add_vma_coredump(ee, engine->gt, engine->status_page.vma,
+			 "HW Status", compress);
+
+	add_vma_coredump(ee, engine->gt, engine->wa_ctx.vma,
+			 "WA context", compress);
+}
+
+static struct intel_engine_coredump *
+capture_engine(struct intel_engine_cs *engine,
+	       struct i915_vma_compress *compress,
+	       u32 dump_flags)
+{
+	struct intel_engine_capture_vma *capture = NULL;
+	struct intel_engine_coredump *ee;
+	struct intel_context *ce = NULL;
+	struct i915_request *rq = NULL;
+
+	ee = intel_engine_coredump_alloc(engine, ALLOW_FAIL, dump_flags);
+	if (!ee)
+		return NULL;
+
+	intel_engine_get_hung_entity(engine, &ce, &rq);
+	if (!rq || !i915_request_started(rq))
+		goto no_request_capture;
+
+	capture = intel_engine_coredump_add_request(ee, rq, ATOMIC_MAYFAIL);
+	if (!capture)
+		goto no_request_capture;
+	if (dump_flags & CORE_DUMP_FLAG_IS_GUC_CAPTURE)
+		intel_guc_capture_get_matching_node(engine->gt, ee, ce);
+
+	intel_engine_coredump_add_vma(ee, capture, compress);
+	i915_request_put(rq);
+
+	return ee;
+
+no_request_capture:
+	if (rq)
+		i915_request_put(rq);
+	kfree(ee);
+	return NULL;
+}
+
+static void
+gt_record_engines(struct intel_gt_coredump *gt,
+		  intel_engine_mask_t engine_mask,
+		  struct i915_vma_compress *compress,
+		  u32 dump_flags)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt->_gt, id) {
+		struct intel_engine_coredump *ee;
+
+		/* Refill our page pool before entering atomic section */
+		pool_refill(&compress->pool, ALLOW_FAIL);
+
+		ee = capture_engine(engine, compress, dump_flags);
+		if (!ee)
+			continue;
+
+		ee->hung = engine->mask & engine_mask;
+
+		gt->simulated |= ee->simulated;
+		if (ee->simulated) {
+			if (dump_flags & CORE_DUMP_FLAG_IS_GUC_CAPTURE)
+				intel_guc_capture_free_node(ee);
+			kfree(ee);
+			continue;
+		}
+
+		ee->next = gt->engine;
+		gt->engine = ee;
+	}
+}
+
+static void gt_record_guc_ctb(struct intel_ctb_coredump *saved,
+			      const struct intel_guc_ct_buffer *ctb,
+			      const void *blob_ptr, struct intel_guc *guc)
+{
+	if (!ctb || !ctb->desc)
+		return;
+
+	saved->raw_status = ctb->desc->status;
+	saved->raw_head = ctb->desc->head;
+	saved->raw_tail = ctb->desc->tail;
+	saved->head = ctb->head;
+	saved->tail = ctb->tail;
+	saved->size = ctb->size;
+	saved->desc_offset = ((void *)ctb->desc) - blob_ptr;
+	saved->cmds_offset = ((void *)ctb->cmds) - blob_ptr;
+}
+
+static struct intel_uc_coredump *
+gt_record_uc(struct intel_gt_coredump *gt,
+	     struct i915_vma_compress *compress)
+{
+	const struct intel_uc *uc = &gt->_gt->uc;
+	struct intel_uc_coredump *error_uc;
+
+	error_uc = kzalloc(sizeof(*error_uc), ALLOW_FAIL);
+	if (!error_uc)
+		return NULL;
+
+	memcpy(&error_uc->guc_fw, &uc->guc.fw, sizeof(uc->guc.fw));
+	memcpy(&error_uc->huc_fw, &uc->huc.fw, sizeof(uc->huc.fw));
+
+	error_uc->guc_fw.file_selected.path = kstrdup(uc->guc.fw.file_selected.path, ALLOW_FAIL);
+	error_uc->huc_fw.file_selected.path = kstrdup(uc->huc.fw.file_selected.path, ALLOW_FAIL);
+	error_uc->guc_fw.file_wanted.path = kstrdup(uc->guc.fw.file_wanted.path, ALLOW_FAIL);
+	error_uc->huc_fw.file_wanted.path = kstrdup(uc->huc.fw.file_wanted.path, ALLOW_FAIL);
+
+	/*
+	 * Save the GuC log and include a timestamp reference for converting the
+	 * log times to system times (in conjunction with the error->boottime and
+	 * gt->clock_frequency fields saved elsewhere).
+	 */
+	error_uc->guc.timestamp = intel_uncore_read(gt->_gt->uncore, GUCPMTIMESTAMP);
+	error_uc->guc.vma_log = create_vma_coredump(gt->_gt, uc->guc.log.vma,
+						    "GuC log buffer", compress);
+	error_uc->guc.vma_ctb = create_vma_coredump(gt->_gt, uc->guc.ct.vma,
+						    "GuC CT buffer", compress);
+	error_uc->guc.last_fence = uc->guc.ct.requests.last_fence;
+	gt_record_guc_ctb(error_uc->guc.ctb + 0, &uc->guc.ct.ctbs.send,
+			  uc->guc.ct.ctbs.send.desc, (struct intel_guc *)&uc->guc);
+	gt_record_guc_ctb(error_uc->guc.ctb + 1, &uc->guc.ct.ctbs.recv,
+			  uc->guc.ct.ctbs.send.desc, (struct intel_guc *)&uc->guc);
+
+	return error_uc;
+}
+
+/* Capture display registers. */
+static void gt_record_display_regs(struct intel_gt_coredump *gt)
+{
+	struct intel_uncore *uncore = gt->_gt->uncore;
+	struct drm_i915_private *i915 = uncore->i915;
+
+	if (GRAPHICS_VER(i915) >= 6)
+		gt->derrmr = intel_uncore_read(uncore, DERRMR);
+
+	if (GRAPHICS_VER(i915) >= 8)
+		gt->ier = intel_uncore_read(uncore, GEN8_DE_MISC_IER);
+	else if (IS_VALLEYVIEW(i915))
+		gt->ier = intel_uncore_read(uncore, VLV_IER);
+	else if (HAS_PCH_SPLIT(i915))
+		gt->ier = intel_uncore_read(uncore, DEIER);
+	else if (GRAPHICS_VER(i915) == 2)
+		gt->ier = intel_uncore_read16(uncore, GEN2_IER);
+	else
+		gt->ier = intel_uncore_read(uncore, GEN2_IER);
+}
+
+/* Capture all other registers that GuC doesn't capture. */
+static void gt_record_global_nonguc_regs(struct intel_gt_coredump *gt)
+{
+	struct intel_uncore *uncore = gt->_gt->uncore;
+	struct drm_i915_private *i915 = uncore->i915;
+	int i;
+
+	if (IS_VALLEYVIEW(i915)) {
+		gt->gtier[0] = intel_uncore_read(uncore, GTIER);
+		gt->ngtier = 1;
+	} else if (GRAPHICS_VER(i915) >= 11) {
+		gt->gtier[0] =
+			intel_uncore_read(uncore,
+					  GEN11_RENDER_COPY_INTR_ENABLE);
+		gt->gtier[1] =
+			intel_uncore_read(uncore, GEN11_VCS_VECS_INTR_ENABLE);
+		gt->gtier[2] =
+			intel_uncore_read(uncore, GEN11_GUC_SG_INTR_ENABLE);
+		gt->gtier[3] =
+			intel_uncore_read(uncore,
+					  GEN11_GPM_WGBOXPERF_INTR_ENABLE);
+		gt->gtier[4] =
+			intel_uncore_read(uncore,
+					  GEN11_CRYPTO_RSVD_INTR_ENABLE);
+		gt->gtier[5] =
+			intel_uncore_read(uncore,
+					  GEN11_GUNIT_CSME_INTR_ENABLE);
+		gt->ngtier = 6;
+	} else if (GRAPHICS_VER(i915) >= 8) {
+		for (i = 0; i < 4; i++)
+			gt->gtier[i] =
+				intel_uncore_read(uncore, GEN8_GT_IER(i));
+		gt->ngtier = 4;
+	} else if (HAS_PCH_SPLIT(i915)) {
+		gt->gtier[0] = intel_uncore_read(uncore, GTIER);
+		gt->ngtier = 1;
+	}
+
+	gt->eir = intel_uncore_read(uncore, EIR);
+	gt->pgtbl_er = intel_uncore_read(uncore, PGTBL_ER);
+}
+
+/*
+ * Capture all registers that relate to workload submission.
+ * NOTE: In GuC submission, when GuC resets an engine, it can dump these for us
+ */
+static void gt_record_global_regs(struct intel_gt_coredump *gt)
+{
+	struct intel_uncore *uncore = gt->_gt->uncore;
+	struct drm_i915_private *i915 = uncore->i915;
+	int i;
+
+	/*
+	 * General organization
+	 * 1. Registers specific to a single generation
+	 * 2. Registers which belong to multiple generations
+	 * 3. Feature specific registers.
+	 * 4. Everything else
+	 * Please try to follow the order.
+	 */
+
+	/* 1: Registers specific to a single generation */
+	if (IS_VALLEYVIEW(i915))
+		gt->forcewake = intel_uncore_read_fw(uncore, FORCEWAKE_VLV);
+
+	if (GRAPHICS_VER(i915) == 7)
+		gt->err_int = intel_uncore_read(uncore, GEN7_ERR_INT);
+
+	if (GRAPHICS_VER(i915) >= 12) {
+		gt->fault_data0 = intel_uncore_read(uncore,
+						    GEN12_FAULT_TLB_DATA0);
+		gt->fault_data1 = intel_uncore_read(uncore,
+						    GEN12_FAULT_TLB_DATA1);
+	} else if (GRAPHICS_VER(i915) >= 8) {
+		gt->fault_data0 = intel_uncore_read(uncore,
+						    GEN8_FAULT_TLB_DATA0);
+		gt->fault_data1 = intel_uncore_read(uncore,
+						    GEN8_FAULT_TLB_DATA1);
+	}
+
+	if (GRAPHICS_VER(i915) == 6) {
+		gt->forcewake = intel_uncore_read_fw(uncore, FORCEWAKE);
+		gt->gab_ctl = intel_uncore_read(uncore, GAB_CTL);
+		gt->gfx_mode = intel_uncore_read(uncore, GFX_MODE);
+	}
+
+	/* 2: Registers which belong to multiple generations */
+	if (GRAPHICS_VER(i915) >= 7)
+		gt->forcewake = intel_uncore_read_fw(uncore, FORCEWAKE_MT);
+
+	if (GRAPHICS_VER(i915) >= 6) {
+		if (GRAPHICS_VER(i915) < 12) {
+			gt->error = intel_uncore_read(uncore, ERROR_GEN6);
+			gt->done_reg = intel_uncore_read(uncore, DONE_REG);
+		}
+	}
+
+	/* 3: Feature specific registers */
+	if (IS_GRAPHICS_VER(i915, 6, 7)) {
+		gt->gam_ecochk = intel_uncore_read(uncore, GAM_ECOCHK);
+		gt->gac_eco = intel_uncore_read(uncore, GAC_ECO_BITS);
+	}
+
+	if (IS_GRAPHICS_VER(i915, 8, 11))
+		gt->gtt_cache = intel_uncore_read(uncore, HSW_GTT_CACHE_EN);
+
+	if (GRAPHICS_VER(i915) == 12)
+		gt->aux_err = intel_uncore_read(uncore, GEN12_AUX_ERR_DBG);
+
+	if (GRAPHICS_VER(i915) >= 12) {
+		for (i = 0; i < I915_MAX_SFC; i++) {
+			/*
+			 * SFC_DONE resides in the VD forcewake domain, so it
+			 * only exists if the corresponding VCS engine is
+			 * present.
+			 */
+			if ((gt->_gt->info.sfc_mask & BIT(i)) == 0 ||
+			    !HAS_ENGINE(gt->_gt, _VCS(i * 2)))
+				continue;
+
+			gt->sfc_done[i] =
+				intel_uncore_read(uncore, GEN12_SFC_DONE(i));
+		}
+
+		gt->gam_done = intel_uncore_read(uncore, GEN12_GAM_DONE);
+	}
+}
+
+static void gt_record_info(struct intel_gt_coredump *gt)
+{
+	memcpy(&gt->info, &gt->_gt->info, sizeof(struct intel_gt_info));
+	gt->clock_frequency = gt->_gt->clock_frequency;
+	gt->clock_period_ns = gt->_gt->clock_period_ns;
+}
+
+/*
+ * Generate a semi-unique error code. The code is not meant to have meaning, The
+ * code's only purpose is to try to prevent false duplicated bug reports by
+ * grossly estimating a GPU error state.
+ *
+ * TODO Ideally, hashing the batchbuffer would be a very nice way to determine
+ * the hang if we could strip the GTT offset information from it.
+ *
+ * It's only a small step better than a random number in its current form.
+ */
+static u32 generate_ecode(const struct intel_engine_coredump *ee)
+{
+	/*
+	 * IPEHR would be an ideal way to detect errors, as it's the gross
+	 * measure of "the command that hung." However, has some very common
+	 * synchronization commands which almost always appear in the case
+	 * strictly a client bug. Use instdone to differentiate those some.
+	 */
+	return ee ? ee->ipehr ^ ee->instdone.instdone : 0;
+}
+
+static const char *error_msg(struct i915_gpu_coredump *error)
+{
+	struct intel_engine_coredump *first = NULL;
+	unsigned int hung_classes = 0;
+	struct intel_gt_coredump *gt;
+	int len;
+
+	for (gt = error->gt; gt; gt = gt->next) {
+		struct intel_engine_coredump *cs;
+
+		for (cs = gt->engine; cs; cs = cs->next) {
+			if (cs->hung) {
+				hung_classes |= BIT(cs->engine->uabi_class);
+				if (!first)
+					first = cs;
+			}
+		}
+	}
+
+	len = scnprintf(error->error_msg, sizeof(error->error_msg),
+			"GPU HANG: ecode %d:%x:%08x",
+			GRAPHICS_VER(error->i915), hung_classes,
+			generate_ecode(first));
+	if (first && first->context.pid) {
+		/* Just show the first executing process, more is confusing */
+		len += scnprintf(error->error_msg + len,
+				 sizeof(error->error_msg) - len,
+				 ", in %s [%d]",
+				 first->context.comm, first->context.pid);
+	}
+
+	return error->error_msg;
+}
+
+static void capture_gen(struct i915_gpu_coredump *error)
+{
+	struct drm_i915_private *i915 = error->i915;
+
+	error->wakelock = atomic_read(&i915->runtime_pm.wakeref_count);
+	error->suspended = i915->runtime_pm.suspended;
+
+	error->iommu = i915_vtd_active(i915);
+	error->reset_count = i915_reset_count(&i915->gpu_error);
+	error->suspend_count = i915->suspend_count;
+
+	i915_params_copy(&error->params, &i915->params);
+	memcpy(&error->device_info,
+	       INTEL_INFO(i915),
+	       sizeof(error->device_info));
+	memcpy(&error->runtime_info,
+	       RUNTIME_INFO(i915),
+	       sizeof(error->runtime_info));
+	error->driver_caps = i915->caps;
+}
+
+struct i915_gpu_coredump *
+i915_gpu_coredump_alloc(struct drm_i915_private *i915, gfp_t gfp)
+{
+	struct i915_gpu_coredump *error;
+
+	if (!i915->params.error_capture)
+		return NULL;
+
+	error = kzalloc(sizeof(*error), gfp);
+	if (!error)
+		return NULL;
+
+	kref_init(&error->ref);
+	error->i915 = i915;
+
+	error->time = ktime_get_real();
+	error->boottime = ktime_get_boottime();
+	error->uptime = ktime_sub(ktime_get(), to_gt(i915)->last_init_time);
+	error->capture = jiffies;
+
+	capture_gen(error);
+
+	return error;
+}
+
+#define DAY_AS_SECONDS(x) (24 * 60 * 60 * (x))
+
+struct intel_gt_coredump *
+intel_gt_coredump_alloc(struct intel_gt *gt, gfp_t gfp, u32 dump_flags)
+{
+	struct intel_gt_coredump *gc;
+
+	gc = kzalloc(sizeof(*gc), gfp);
+	if (!gc)
+		return NULL;
+
+	gc->_gt = gt;
+	gc->awake = intel_gt_pm_is_awake(gt);
+
+	gt_record_display_regs(gc);
+	gt_record_global_nonguc_regs(gc);
+
+	/*
+	 * GuC dumps global, eng-class and eng-instance registers
+	 * (that can change as part of engine state during execution)
+	 * before an engine is reset due to a hung context.
+	 * GuC captures and reports all three groups of registers
+	 * together as a single set before the engine is reset.
+	 * Thus, if GuC triggered the context reset we retrieve
+	 * the register values as part of gt_record_engines.
+	 */
+	if (!(dump_flags & CORE_DUMP_FLAG_IS_GUC_CAPTURE))
+		gt_record_global_regs(gc);
+
+	gt_record_fences(gc);
+
+	return gc;
+}
+
+struct i915_vma_compress *
+i915_vma_capture_prepare(struct intel_gt_coredump *gt)
+{
+	struct i915_vma_compress *compress;
+
+	compress = kmalloc(sizeof(*compress), ALLOW_FAIL);
+	if (!compress)
+		return NULL;
+
+	if (!compress_init(compress)) {
+		kfree(compress);
+		return NULL;
+	}
+
+	return compress;
+}
+
+void i915_vma_capture_finish(struct intel_gt_coredump *gt,
+			     struct i915_vma_compress *compress)
+{
+	if (!compress)
+		return;
+
+	compress_fini(compress);
+	kfree(compress);
+}
+
+static struct i915_gpu_coredump *
+__i915_gpu_coredump(struct intel_gt *gt, intel_engine_mask_t engine_mask, u32 dump_flags)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct i915_gpu_coredump *error;
+
+	/* Check if GPU capture has been disabled */
+	error = READ_ONCE(i915->gpu_error.first_error);
+	if (IS_ERR(error))
+		return error;
+
+	error = i915_gpu_coredump_alloc(i915, ALLOW_FAIL);
+	if (!error)
+		return ERR_PTR(-ENOMEM);
+
+	error->gt = intel_gt_coredump_alloc(gt, ALLOW_FAIL, dump_flags);
+	if (error->gt) {
+		struct i915_vma_compress *compress;
+
+		compress = i915_vma_capture_prepare(error->gt);
+		if (!compress) {
+			kfree(error->gt);
+			kfree(error);
+			return ERR_PTR(-ENOMEM);
+		}
+
+		if (INTEL_INFO(i915)->has_gt_uc) {
+			error->gt->uc = gt_record_uc(error->gt, compress);
+			if (error->gt->uc) {
+				if (dump_flags & CORE_DUMP_FLAG_IS_GUC_CAPTURE)
+					error->gt->uc->guc.is_guc_capture = true;
+				else
+					GEM_BUG_ON(error->gt->uc->guc.is_guc_capture);
+			}
+		}
+
+		gt_record_info(error->gt);
+		gt_record_engines(error->gt, engine_mask, compress, dump_flags);
+
+
+		i915_vma_capture_finish(error->gt, compress);
+
+		error->simulated |= error->gt->simulated;
+	}
+
+	error->overlay = intel_overlay_capture_error_state(i915);
+
+	return error;
+}
+
+struct i915_gpu_coredump *
+i915_gpu_coredump(struct intel_gt *gt, intel_engine_mask_t engine_mask, u32 dump_flags)
+{
+	static DEFINE_MUTEX(capture_mutex);
+	int ret = mutex_lock_interruptible(&capture_mutex);
+	struct i915_gpu_coredump *dump;
+
+	if (ret)
+		return ERR_PTR(ret);
+
+	dump = __i915_gpu_coredump(gt, engine_mask, dump_flags);
+	mutex_unlock(&capture_mutex);
+
+	return dump;
+}
+
+void i915_error_state_store(struct i915_gpu_coredump *error)
+{
+	struct drm_i915_private *i915;
+	static bool warned;
+
+	if (IS_ERR_OR_NULL(error))
+		return;
+
+	i915 = error->i915;
+	drm_info(&i915->drm, "%s\n", error_msg(error));
+
+	if (error->simulated ||
+	    cmpxchg(&i915->gpu_error.first_error, NULL, error))
+		return;
+
+	i915_gpu_coredump_get(error);
+
+	if (!xchg(&warned, true) &&
+	    ktime_get_real_seconds() - DRIVER_TIMESTAMP < DAY_AS_SECONDS(180)) {
+		pr_info("GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.\n");
+		pr_info("Please file a _new_ bug report at https://gitlab.freedesktop.org/drm/intel/issues/new.\n");
+		pr_info("Please see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details.\n");
+		pr_info("drm/i915 developers can then reassign to the right component if it's not a kernel issue.\n");
+		pr_info("The GPU crash dump is required to analyze GPU hangs, so please always attach it.\n");
+		pr_info("GPU crash dump saved to /sys/class/drm/card%d/error\n",
+			i915->drm.primary->index);
+	}
+}
+
+/**
+ * i915_capture_error_state - capture an error record for later analysis
+ * @gt: intel_gt which originated the hang
+ * @engine_mask: hung engines
+ *
+ *
+ * Should be called when an error is detected (either a hang or an error
+ * interrupt) to capture error state from the time of the error.  Fills
+ * out a structure which becomes available in debugfs for user level tools
+ * to pick up.
+ */
+void i915_capture_error_state(struct intel_gt *gt,
+			      intel_engine_mask_t engine_mask, u32 dump_flags)
+{
+	struct i915_gpu_coredump *error;
+
+	error = i915_gpu_coredump(gt, engine_mask, dump_flags);
+	if (IS_ERR(error)) {
+		cmpxchg(&gt->i915->gpu_error.first_error, NULL, error);
+		return;
+	}
+
+	i915_error_state_store(error);
+	i915_gpu_coredump_put(error);
+}
+
+struct i915_gpu_coredump *
+i915_first_error_state(struct drm_i915_private *i915)
+{
+	struct i915_gpu_coredump *error;
+
+	spin_lock_irq(&i915->gpu_error.lock);
+	error = i915->gpu_error.first_error;
+	if (!IS_ERR_OR_NULL(error))
+		i915_gpu_coredump_get(error);
+	spin_unlock_irq(&i915->gpu_error.lock);
+
+	return error;
+}
+
+void i915_reset_error_state(struct drm_i915_private *i915)
+{
+	struct i915_gpu_coredump *error;
+
+	spin_lock_irq(&i915->gpu_error.lock);
+	error = i915->gpu_error.first_error;
+	if (error != ERR_PTR(-ENODEV)) /* if disabled, always disabled */
+		i915->gpu_error.first_error = NULL;
+	spin_unlock_irq(&i915->gpu_error.lock);
+
+	if (!IS_ERR_OR_NULL(error))
+		i915_gpu_coredump_put(error);
+}
+
+void i915_disable_error_state(struct drm_i915_private *i915, int err)
+{
+	spin_lock_irq(&i915->gpu_error.lock);
+	if (!i915->gpu_error.first_error)
+		i915->gpu_error.first_error = ERR_PTR(err);
+	spin_unlock_irq(&i915->gpu_error.lock);
+}
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.h b/drivers/gpu/drm/i915/i915_gpu_error.h
new file mode 100644
index 000000000..56027ffbc
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gpu_error.h
@@ -0,0 +1,407 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2008-2018 Intel Corporation
+ */
+
+#ifndef _I915_GPU_ERROR_H_
+#define _I915_GPU_ERROR_H_
+
+#include <linux/atomic.h>
+#include <linux/kref.h>
+#include <linux/ktime.h>
+#include <linux/sched.h>
+
+#include <drm/drm_mm.h>
+
+#include "gt/intel_engine.h"
+#include "gt/intel_gt_types.h"
+#include "gt/uc/intel_uc_fw.h"
+
+#include "intel_device_info.h"
+
+#include "i915_gem.h"
+#include "i915_gem_gtt.h"
+#include "i915_params.h"
+#include "i915_scheduler.h"
+
+struct drm_i915_private;
+struct i915_vma_compress;
+struct intel_engine_capture_vma;
+struct intel_overlay_error_state;
+
+struct i915_vma_coredump {
+	struct i915_vma_coredump *next;
+
+	char name[20];
+
+	u64 gtt_offset;
+	u64 gtt_size;
+	u32 gtt_page_sizes;
+
+	int unused;
+	struct list_head page_list;
+};
+
+struct i915_request_coredump {
+	unsigned long flags;
+	pid_t pid;
+	u32 context;
+	u32 seqno;
+	u32 head;
+	u32 tail;
+	struct i915_sched_attr sched_attr;
+};
+
+struct __guc_capture_parsed_output;
+
+struct intel_engine_coredump {
+	const struct intel_engine_cs *engine;
+
+	bool hung;
+	bool simulated;
+	u32 reset_count;
+
+	/* position of active request inside the ring */
+	u32 rq_head, rq_post, rq_tail;
+
+	/* Register state */
+	u32 ccid;
+	u32 start;
+	u32 tail;
+	u32 head;
+	u32 ctl;
+	u32 mode;
+	u32 hws;
+	u32 ipeir;
+	u32 ipehr;
+	u32 esr;
+	u32 bbstate;
+	u32 instpm;
+	u32 instps;
+	u64 bbaddr;
+	u64 acthd;
+	u32 fault_reg;
+	u64 faddr;
+	u32 rc_psmi; /* sleep state */
+	u32 nopid;
+	u32 excc;
+	u32 cmd_cctl;
+	u32 cscmdop;
+	u32 ctx_sr_ctl;
+	u32 dma_faddr_hi;
+	u32 dma_faddr_lo;
+	struct intel_instdone instdone;
+
+	/* GuC matched capture-lists info */
+	struct intel_guc_state_capture *guc_capture;
+	struct __guc_capture_parsed_output *guc_capture_node;
+
+	struct i915_gem_context_coredump {
+		char comm[TASK_COMM_LEN];
+
+		u64 total_runtime;
+		u64 avg_runtime;
+
+		pid_t pid;
+		int active;
+		int guilty;
+		struct i915_sched_attr sched_attr;
+	} context;
+
+	struct i915_vma_coredump *vma;
+
+	struct i915_request_coredump execlist[EXECLIST_MAX_PORTS];
+	unsigned int num_ports;
+
+	struct {
+		u32 gfx_mode;
+		union {
+			u64 pdp[4];
+			u32 pp_dir_base;
+		};
+	} vm_info;
+
+	struct intel_engine_coredump *next;
+};
+
+struct intel_ctb_coredump {
+	u32 raw_head, head;
+	u32 raw_tail, tail;
+	u32 raw_status;
+	u32 desc_offset;
+	u32 cmds_offset;
+	u32 size;
+};
+
+struct intel_gt_coredump {
+	const struct intel_gt *_gt;
+	bool awake;
+	bool simulated;
+
+	struct intel_gt_info info;
+
+	/* Generic register state */
+	u32 eir;
+	u32 pgtbl_er;
+	u32 ier;
+	u32 gtier[6], ngtier;
+	u32 forcewake;
+	u32 error; /* gen6+ */
+	u32 err_int; /* gen7 */
+	u32 fault_data0; /* gen8, gen9 */
+	u32 fault_data1; /* gen8, gen9 */
+	u32 done_reg;
+	u32 gac_eco;
+	u32 gam_ecochk;
+	u32 gab_ctl;
+	u32 gfx_mode;
+	u32 gtt_cache;
+	u32 aux_err; /* gen12 */
+	u32 gam_done; /* gen12 */
+	u32 clock_frequency;
+	u32 clock_period_ns;
+
+	/* Display related */
+	u32 derrmr;
+	u32 sfc_done[I915_MAX_SFC]; /* gen12 */
+
+	u32 nfence;
+	u64 fence[I915_MAX_NUM_FENCES];
+
+	struct intel_engine_coredump *engine;
+
+	struct intel_uc_coredump {
+		struct intel_uc_fw guc_fw;
+		struct intel_uc_fw huc_fw;
+		struct guc_info {
+			struct intel_ctb_coredump ctb[2];
+			struct i915_vma_coredump *vma_ctb;
+			struct i915_vma_coredump *vma_log;
+			u32 timestamp;
+			u16 last_fence;
+			bool is_guc_capture;
+		} guc;
+	} *uc;
+
+	struct intel_gt_coredump *next;
+};
+
+struct i915_gpu_coredump {
+	struct kref ref;
+	ktime_t time;
+	ktime_t boottime;
+	ktime_t uptime;
+	unsigned long capture;
+
+	struct drm_i915_private *i915;
+
+	struct intel_gt_coredump *gt;
+
+	char error_msg[128];
+	bool simulated;
+	bool wakelock;
+	bool suspended;
+	int iommu;
+	u32 reset_count;
+	u32 suspend_count;
+
+	struct intel_device_info device_info;
+	struct intel_runtime_info runtime_info;
+	struct intel_driver_caps driver_caps;
+	struct i915_params params;
+
+	struct intel_overlay_error_state *overlay;
+
+	struct scatterlist *sgl, *fit;
+};
+
+struct i915_gpu_error {
+	/* For reset and error_state handling. */
+	spinlock_t lock;
+	/* Protected by the above dev->gpu_error.lock. */
+	struct i915_gpu_coredump *first_error;
+
+	atomic_t pending_fb_pin;
+
+	/** Number of times the device has been reset (global) */
+	atomic_t reset_count;
+
+	/** Number of times an engine has been reset */
+	atomic_t reset_engine_count[I915_NUM_ENGINES];
+};
+
+struct drm_i915_error_state_buf {
+	struct drm_i915_private *i915;
+	struct scatterlist *sgl, *cur, *end;
+
+	char *buf;
+	size_t bytes;
+	size_t size;
+	loff_t iter;
+
+	int err;
+};
+
+static inline u32 i915_reset_count(struct i915_gpu_error *error)
+{
+	return atomic_read(&error->reset_count);
+}
+
+static inline u32 i915_reset_engine_count(struct i915_gpu_error *error,
+					  const struct intel_engine_cs *engine)
+{
+	return atomic_read(&error->reset_engine_count[engine->uabi_class]);
+}
+
+#define CORE_DUMP_FLAG_NONE           0x0
+#define CORE_DUMP_FLAG_IS_GUC_CAPTURE BIT(0)
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+
+__printf(2, 3)
+void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...);
+void intel_gpu_error_print_vma(struct drm_i915_error_state_buf *m,
+			       const struct intel_engine_cs *engine,
+			       const struct i915_vma_coredump *vma);
+struct i915_vma_coredump *
+intel_gpu_error_find_batch(const struct intel_engine_coredump *ee);
+
+struct i915_gpu_coredump *i915_gpu_coredump(struct intel_gt *gt,
+					    intel_engine_mask_t engine_mask, u32 dump_flags);
+void i915_capture_error_state(struct intel_gt *gt,
+			      intel_engine_mask_t engine_mask, u32 dump_flags);
+
+struct i915_gpu_coredump *
+i915_gpu_coredump_alloc(struct drm_i915_private *i915, gfp_t gfp);
+
+struct intel_gt_coredump *
+intel_gt_coredump_alloc(struct intel_gt *gt, gfp_t gfp, u32 dump_flags);
+
+struct intel_engine_coredump *
+intel_engine_coredump_alloc(struct intel_engine_cs *engine, gfp_t gfp, u32 dump_flags);
+
+struct intel_engine_capture_vma *
+intel_engine_coredump_add_request(struct intel_engine_coredump *ee,
+				  struct i915_request *rq,
+				  gfp_t gfp);
+
+void intel_engine_coredump_add_vma(struct intel_engine_coredump *ee,
+				   struct intel_engine_capture_vma *capture,
+				   struct i915_vma_compress *compress);
+
+struct i915_vma_compress *
+i915_vma_capture_prepare(struct intel_gt_coredump *gt);
+
+void i915_vma_capture_finish(struct intel_gt_coredump *gt,
+			     struct i915_vma_compress *compress);
+
+void i915_error_state_store(struct i915_gpu_coredump *error);
+
+static inline struct i915_gpu_coredump *
+i915_gpu_coredump_get(struct i915_gpu_coredump *gpu)
+{
+	kref_get(&gpu->ref);
+	return gpu;
+}
+
+ssize_t
+i915_gpu_coredump_copy_to_buffer(struct i915_gpu_coredump *error,
+				 char *buf, loff_t offset, size_t count);
+
+void __i915_gpu_coredump_free(struct kref *kref);
+static inline void i915_gpu_coredump_put(struct i915_gpu_coredump *gpu)
+{
+	if (gpu)
+		kref_put(&gpu->ref, __i915_gpu_coredump_free);
+}
+
+struct i915_gpu_coredump *i915_first_error_state(struct drm_i915_private *i915);
+void i915_reset_error_state(struct drm_i915_private *i915);
+void i915_disable_error_state(struct drm_i915_private *i915, int err);
+
+#else
+
+__printf(2, 3)
+static inline void
+i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...)
+{
+}
+
+static inline void
+i915_capture_error_state(struct intel_gt *gt, intel_engine_mask_t engine_mask, u32 dump_flags)
+{
+}
+
+static inline struct i915_gpu_coredump *
+i915_gpu_coredump_alloc(struct drm_i915_private *i915, gfp_t gfp)
+{
+	return NULL;
+}
+
+static inline struct intel_gt_coredump *
+intel_gt_coredump_alloc(struct intel_gt *gt, gfp_t gfp, u32 dump_flags)
+{
+	return NULL;
+}
+
+static inline struct intel_engine_coredump *
+intel_engine_coredump_alloc(struct intel_engine_cs *engine, gfp_t gfp, u32 dump_flags)
+{
+	return NULL;
+}
+
+static inline struct intel_engine_capture_vma *
+intel_engine_coredump_add_request(struct intel_engine_coredump *ee,
+				  struct i915_request *rq,
+				  gfp_t gfp)
+{
+	return NULL;
+}
+
+static inline void
+intel_engine_coredump_add_vma(struct intel_engine_coredump *ee,
+			      struct intel_engine_capture_vma *capture,
+			      struct i915_vma_compress *compress)
+{
+}
+
+static inline struct i915_vma_compress *
+i915_vma_capture_prepare(struct intel_gt_coredump *gt)
+{
+	return NULL;
+}
+
+static inline void
+i915_vma_capture_finish(struct intel_gt_coredump *gt,
+			struct i915_vma_compress *compress)
+{
+}
+
+static inline void
+i915_error_state_store(struct i915_gpu_coredump *error)
+{
+}
+
+static inline void i915_gpu_coredump_put(struct i915_gpu_coredump *gpu)
+{
+}
+
+static inline struct i915_gpu_coredump *
+i915_first_error_state(struct drm_i915_private *i915)
+{
+	return ERR_PTR(-ENODEV);
+}
+
+static inline void i915_reset_error_state(struct drm_i915_private *i915)
+{
+}
+
+static inline void i915_disable_error_state(struct drm_i915_private *i915,
+					    int err)
+{
+}
+
+#endif /* IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) */
+
+#endif /* _I915_GPU_ERROR_H_ */
diff --git a/drivers/gpu/drm/i915/i915_ioc32.c b/drivers/gpu/drm/i915/i915_ioc32.c
new file mode 100644
index 000000000..33348960f
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_ioc32.c
@@ -0,0 +1,94 @@
+/*
+ * 32-bit ioctl compatibility routines for the i915 DRM.
+ *
+ * Copyright (C) Paul Mackerras 2005
+ * Copyright (C) Alan Hourihane 2005
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHOR BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
+ * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Author: Alan Hourihane <alanh@fairlite.demon.co.uk>
+ */
+#include <linux/compat.h>
+
+#include <drm/drm_ioctl.h>
+
+#include "i915_drv.h"
+#include "i915_getparam.h"
+#include "i915_ioc32.h"
+
+struct drm_i915_getparam32 {
+	s32 param;
+	/*
+	 * We screwed up the generic ioctl struct here and used a variable-sized
+	 * pointer. Use u32 in the compat struct to match the 32bit pointer
+	 * userspace expects.
+	 */
+	u32 value;
+};
+
+static int compat_i915_getparam(struct file *file, unsigned int cmd,
+				unsigned long arg)
+{
+	struct drm_i915_getparam32 req32;
+	struct drm_i915_getparam req;
+
+	if (copy_from_user(&req32, (void __user *)arg, sizeof(req32)))
+		return -EFAULT;
+
+	req.param = req32.param;
+	req.value = compat_ptr(req32.value);
+
+	return drm_ioctl_kernel(file, i915_getparam_ioctl, &req,
+				DRM_RENDER_ALLOW);
+}
+
+static drm_ioctl_compat_t *i915_compat_ioctls[] = {
+	[DRM_I915_GETPARAM] = compat_i915_getparam,
+};
+
+/**
+ * i915_ioc32_compat_ioctl - handle the mistakes of the past
+ * @filp: the file pointer
+ * @cmd: the ioctl command (and encoded flags)
+ * @arg: the ioctl argument (from userspace)
+ *
+ * Called whenever a 32-bit process running under a 64-bit kernel
+ * performs an ioctl on /dev/dri/card<n>.
+ */
+long i915_ioc32_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
+{
+	unsigned int nr = DRM_IOCTL_NR(cmd);
+	drm_ioctl_compat_t *fn = NULL;
+	int ret;
+
+	if (nr < DRM_COMMAND_BASE || nr >= DRM_COMMAND_END)
+		return drm_compat_ioctl(filp, cmd, arg);
+
+	if (nr < DRM_COMMAND_BASE + ARRAY_SIZE(i915_compat_ioctls))
+		fn = i915_compat_ioctls[nr - DRM_COMMAND_BASE];
+
+	if (fn != NULL)
+		ret = (*fn) (filp, cmd, arg);
+	else
+		ret = drm_ioctl(filp, cmd, arg);
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/i915_ioc32.h b/drivers/gpu/drm/i915/i915_ioc32.h
new file mode 100644
index 000000000..40dcd55ca
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_ioc32.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __I915_IOC32_H__
+#define __I915_IOC32_H__
+
+#ifdef CONFIG_COMPAT
+struct file;
+long i915_ioc32_compat_ioctl(struct file *filp, unsigned int cmd,
+			     unsigned long arg);
+#else
+#define i915_ioc32_compat_ioctl NULL
+#endif
+
+#endif /* __I915_IOC32_H__ */
diff --git a/drivers/gpu/drm/i915/i915_ioctl.c b/drivers/gpu/drm/i915/i915_ioctl.c
new file mode 100644
index 000000000..06a10ccea
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_ioctl.c
@@ -0,0 +1,94 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#include "gt/intel_engine_regs.h"
+
+#include "i915_drv.h"
+#include "i915_gem.h"
+#include "i915_ioctl.h"
+#include "i915_reg.h"
+#include "intel_runtime_pm.h"
+#include "intel_uncore.h"
+
+/*
+ * This file is for small ioctl functions that are out of place everywhere else,
+ * and not big enough to warrant a file of their own.
+ *
+ * This is not the dumping ground for random ioctls.
+ */
+
+struct reg_whitelist {
+	i915_reg_t offset_ldw;
+	i915_reg_t offset_udw;
+	u8 min_graphics_ver;
+	u8 max_graphics_ver;
+	u8 size;
+};
+
+static const struct reg_whitelist reg_read_whitelist[] = {
+	{
+		.offset_ldw = RING_TIMESTAMP(RENDER_RING_BASE),
+		.offset_udw = RING_TIMESTAMP_UDW(RENDER_RING_BASE),
+		.min_graphics_ver = 4,
+		.max_graphics_ver = 12,
+		.size = 8
+	}
+};
+
+int i915_reg_read_ioctl(struct drm_device *dev,
+			void *data, struct drm_file *unused)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct intel_uncore *uncore = &i915->uncore;
+	struct drm_i915_reg_read *reg = data;
+	struct reg_whitelist const *entry;
+	intel_wakeref_t wakeref;
+	unsigned int flags;
+	int remain;
+	int ret = 0;
+
+	entry = reg_read_whitelist;
+	remain = ARRAY_SIZE(reg_read_whitelist);
+	while (remain) {
+		u32 entry_offset = i915_mmio_reg_offset(entry->offset_ldw);
+
+		GEM_BUG_ON(!is_power_of_2(entry->size));
+		GEM_BUG_ON(entry->size > 8);
+		GEM_BUG_ON(entry_offset & (entry->size - 1));
+
+		if (IS_GRAPHICS_VER(i915, entry->min_graphics_ver, entry->max_graphics_ver) &&
+		    entry_offset == (reg->offset & -entry->size))
+			break;
+		entry++;
+		remain--;
+	}
+
+	if (!remain)
+		return -EINVAL;
+
+	flags = reg->offset & (entry->size - 1);
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		if (entry->size == 8 && flags == I915_REG_READ_8B_WA)
+			reg->val = intel_uncore_read64_2x32(uncore,
+							    entry->offset_ldw,
+							    entry->offset_udw);
+		else if (entry->size == 8 && flags == 0)
+			reg->val = intel_uncore_read64(uncore,
+						       entry->offset_ldw);
+		else if (entry->size == 4 && flags == 0)
+			reg->val = intel_uncore_read(uncore, entry->offset_ldw);
+		else if (entry->size == 2 && flags == 0)
+			reg->val = intel_uncore_read16(uncore,
+						       entry->offset_ldw);
+		else if (entry->size == 1 && flags == 0)
+			reg->val = intel_uncore_read8(uncore,
+						      entry->offset_ldw);
+		else
+			ret = -EINVAL;
+	}
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/i915_ioctl.h b/drivers/gpu/drm/i915/i915_ioctl.h
new file mode 100644
index 000000000..f16ae87b8
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_ioctl.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __I915_IOCTL_H__
+#define __I915_IOCTL_H__
+
+struct drm_device;
+struct drm_file;
+
+int i915_reg_read_ioctl(struct drm_device *dev, void *data, struct drm_file *file);
+
+#endif /* __I915_IOCTL_H__ */
diff --git a/drivers/gpu/drm/i915/i915_iosf_mbi.h b/drivers/gpu/drm/i915/i915_iosf_mbi.h
new file mode 100644
index 000000000..8f81b7603
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_iosf_mbi.h
@@ -0,0 +1,42 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_IOSF_MBI_H__
+#define __I915_IOSF_MBI_H__
+
+#if IS_ENABLED(CONFIG_IOSF_MBI)
+#include <asm/iosf_mbi.h>
+#else
+
+/* Stubs to compile for all non-x86 archs */
+#define MBI_PMIC_BUS_ACCESS_BEGIN       1
+#define MBI_PMIC_BUS_ACCESS_END         2
+
+struct notifier_block;
+
+static inline void iosf_mbi_punit_acquire(void) {}
+static inline void iosf_mbi_punit_release(void) {}
+static inline void iosf_mbi_assert_punit_acquired(void) {}
+
+static inline
+int iosf_mbi_register_pmic_bus_access_notifier(struct notifier_block *nb)
+{
+	return 0;
+}
+
+static inline int
+iosf_mbi_unregister_pmic_bus_access_notifier_unlocked(struct notifier_block *nb)
+{
+	return 0;
+}
+
+static inline
+int iosf_mbi_unregister_pmic_bus_access_notifier(struct notifier_block *nb)
+{
+	return 0;
+}
+#endif
+
+#endif /* __I915_IOSF_MBI_H__ */
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
new file mode 100644
index 000000000..f93ffa662
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -0,0 +1,4648 @@
+/* i915_irq.c -- IRQ support for the I915 -*- linux-c -*-
+ */
+/*
+ * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/circ_buf.h>
+#include <linux/slab.h>
+#include <linux/sysrq.h>
+
+#include <drm/drm_drv.h>
+
+#include "display/icl_dsi_regs.h"
+#include "display/intel_de.h"
+#include "display/intel_display_trace.h"
+#include "display/intel_display_types.h"
+#include "display/intel_fifo_underrun.h"
+#include "display/intel_hotplug.h"
+#include "display/intel_lpe_audio.h"
+#include "display/intel_psr.h"
+
+#include "gt/intel_breadcrumbs.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_irq.h"
+#include "gt/intel_gt_pm_irq.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_rps.h"
+
+#include "i915_driver.h"
+#include "i915_drv.h"
+#include "i915_irq.h"
+#include "intel_pm.h"
+
+/**
+ * DOC: interrupt handling
+ *
+ * These functions provide the basic support for enabling and disabling the
+ * interrupt handling support. There's a lot more functionality in i915_irq.c
+ * and related files, but that will be described in separate chapters.
+ */
+
+/*
+ * Interrupt statistic for PMU. Increments the counter only if the
+ * interrupt originated from the GPU so interrupts from a device which
+ * shares the interrupt line are not accounted.
+ */
+static inline void pmu_irq_stats(struct drm_i915_private *i915,
+				 irqreturn_t res)
+{
+	if (unlikely(res != IRQ_HANDLED))
+		return;
+
+	/*
+	 * A clever compiler translates that into INC. A not so clever one
+	 * should at least prevent store tearing.
+	 */
+	WRITE_ONCE(i915->pmu.irq_count, i915->pmu.irq_count + 1);
+}
+
+typedef bool (*long_pulse_detect_func)(enum hpd_pin pin, u32 val);
+typedef u32 (*hotplug_enables_func)(struct drm_i915_private *i915,
+				    enum hpd_pin pin);
+
+static const u32 hpd_ilk[HPD_NUM_PINS] = {
+	[HPD_PORT_A] = DE_DP_A_HOTPLUG,
+};
+
+static const u32 hpd_ivb[HPD_NUM_PINS] = {
+	[HPD_PORT_A] = DE_DP_A_HOTPLUG_IVB,
+};
+
+static const u32 hpd_bdw[HPD_NUM_PINS] = {
+	[HPD_PORT_A] = GEN8_DE_PORT_HOTPLUG(HPD_PORT_A),
+};
+
+static const u32 hpd_ibx[HPD_NUM_PINS] = {
+	[HPD_CRT] = SDE_CRT_HOTPLUG,
+	[HPD_SDVO_B] = SDE_SDVOB_HOTPLUG,
+	[HPD_PORT_B] = SDE_PORTB_HOTPLUG,
+	[HPD_PORT_C] = SDE_PORTC_HOTPLUG,
+	[HPD_PORT_D] = SDE_PORTD_HOTPLUG,
+};
+
+static const u32 hpd_cpt[HPD_NUM_PINS] = {
+	[HPD_CRT] = SDE_CRT_HOTPLUG_CPT,
+	[HPD_SDVO_B] = SDE_SDVOB_HOTPLUG_CPT,
+	[HPD_PORT_B] = SDE_PORTB_HOTPLUG_CPT,
+	[HPD_PORT_C] = SDE_PORTC_HOTPLUG_CPT,
+	[HPD_PORT_D] = SDE_PORTD_HOTPLUG_CPT,
+};
+
+static const u32 hpd_spt[HPD_NUM_PINS] = {
+	[HPD_PORT_A] = SDE_PORTA_HOTPLUG_SPT,
+	[HPD_PORT_B] = SDE_PORTB_HOTPLUG_CPT,
+	[HPD_PORT_C] = SDE_PORTC_HOTPLUG_CPT,
+	[HPD_PORT_D] = SDE_PORTD_HOTPLUG_CPT,
+	[HPD_PORT_E] = SDE_PORTE_HOTPLUG_SPT,
+};
+
+static const u32 hpd_mask_i915[HPD_NUM_PINS] = {
+	[HPD_CRT] = CRT_HOTPLUG_INT_EN,
+	[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_EN,
+	[HPD_SDVO_C] = SDVOC_HOTPLUG_INT_EN,
+	[HPD_PORT_B] = PORTB_HOTPLUG_INT_EN,
+	[HPD_PORT_C] = PORTC_HOTPLUG_INT_EN,
+	[HPD_PORT_D] = PORTD_HOTPLUG_INT_EN,
+};
+
+static const u32 hpd_status_g4x[HPD_NUM_PINS] = {
+	[HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
+	[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_G4X,
+	[HPD_SDVO_C] = SDVOC_HOTPLUG_INT_STATUS_G4X,
+	[HPD_PORT_B] = PORTB_HOTPLUG_INT_STATUS,
+	[HPD_PORT_C] = PORTC_HOTPLUG_INT_STATUS,
+	[HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS,
+};
+
+static const u32 hpd_status_i915[HPD_NUM_PINS] = {
+	[HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
+	[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_I915,
+	[HPD_SDVO_C] = SDVOC_HOTPLUG_INT_STATUS_I915,
+	[HPD_PORT_B] = PORTB_HOTPLUG_INT_STATUS,
+	[HPD_PORT_C] = PORTC_HOTPLUG_INT_STATUS,
+	[HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS,
+};
+
+static const u32 hpd_bxt[HPD_NUM_PINS] = {
+	[HPD_PORT_A] = GEN8_DE_PORT_HOTPLUG(HPD_PORT_A),
+	[HPD_PORT_B] = GEN8_DE_PORT_HOTPLUG(HPD_PORT_B),
+	[HPD_PORT_C] = GEN8_DE_PORT_HOTPLUG(HPD_PORT_C),
+};
+
+static const u32 hpd_gen11[HPD_NUM_PINS] = {
+	[HPD_PORT_TC1] = GEN11_TC_HOTPLUG(HPD_PORT_TC1) | GEN11_TBT_HOTPLUG(HPD_PORT_TC1),
+	[HPD_PORT_TC2] = GEN11_TC_HOTPLUG(HPD_PORT_TC2) | GEN11_TBT_HOTPLUG(HPD_PORT_TC2),
+	[HPD_PORT_TC3] = GEN11_TC_HOTPLUG(HPD_PORT_TC3) | GEN11_TBT_HOTPLUG(HPD_PORT_TC3),
+	[HPD_PORT_TC4] = GEN11_TC_HOTPLUG(HPD_PORT_TC4) | GEN11_TBT_HOTPLUG(HPD_PORT_TC4),
+	[HPD_PORT_TC5] = GEN11_TC_HOTPLUG(HPD_PORT_TC5) | GEN11_TBT_HOTPLUG(HPD_PORT_TC5),
+	[HPD_PORT_TC6] = GEN11_TC_HOTPLUG(HPD_PORT_TC6) | GEN11_TBT_HOTPLUG(HPD_PORT_TC6),
+};
+
+static const u32 hpd_icp[HPD_NUM_PINS] = {
+	[HPD_PORT_A] = SDE_DDI_HOTPLUG_ICP(HPD_PORT_A),
+	[HPD_PORT_B] = SDE_DDI_HOTPLUG_ICP(HPD_PORT_B),
+	[HPD_PORT_C] = SDE_DDI_HOTPLUG_ICP(HPD_PORT_C),
+	[HPD_PORT_TC1] = SDE_TC_HOTPLUG_ICP(HPD_PORT_TC1),
+	[HPD_PORT_TC2] = SDE_TC_HOTPLUG_ICP(HPD_PORT_TC2),
+	[HPD_PORT_TC3] = SDE_TC_HOTPLUG_ICP(HPD_PORT_TC3),
+	[HPD_PORT_TC4] = SDE_TC_HOTPLUG_ICP(HPD_PORT_TC4),
+	[HPD_PORT_TC5] = SDE_TC_HOTPLUG_ICP(HPD_PORT_TC5),
+	[HPD_PORT_TC6] = SDE_TC_HOTPLUG_ICP(HPD_PORT_TC6),
+};
+
+static const u32 hpd_sde_dg1[HPD_NUM_PINS] = {
+	[HPD_PORT_A] = SDE_DDI_HOTPLUG_ICP(HPD_PORT_A),
+	[HPD_PORT_B] = SDE_DDI_HOTPLUG_ICP(HPD_PORT_B),
+	[HPD_PORT_C] = SDE_DDI_HOTPLUG_ICP(HPD_PORT_C),
+	[HPD_PORT_D] = SDE_DDI_HOTPLUG_ICP(HPD_PORT_D),
+	[HPD_PORT_TC1] = SDE_TC_HOTPLUG_DG2(HPD_PORT_TC1),
+};
+
+static void intel_hpd_init_pins(struct drm_i915_private *dev_priv)
+{
+	struct intel_hotplug *hpd = &dev_priv->display.hotplug;
+
+	if (HAS_GMCH(dev_priv)) {
+		if (IS_G4X(dev_priv) || IS_VALLEYVIEW(dev_priv) ||
+		    IS_CHERRYVIEW(dev_priv))
+			hpd->hpd = hpd_status_g4x;
+		else
+			hpd->hpd = hpd_status_i915;
+		return;
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 11)
+		hpd->hpd = hpd_gen11;
+	else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		hpd->hpd = hpd_bxt;
+	else if (DISPLAY_VER(dev_priv) >= 8)
+		hpd->hpd = hpd_bdw;
+	else if (DISPLAY_VER(dev_priv) >= 7)
+		hpd->hpd = hpd_ivb;
+	else
+		hpd->hpd = hpd_ilk;
+
+	if ((INTEL_PCH_TYPE(dev_priv) < PCH_DG1) &&
+	    (!HAS_PCH_SPLIT(dev_priv) || HAS_PCH_NOP(dev_priv)))
+		return;
+
+	if (INTEL_PCH_TYPE(dev_priv) >= PCH_DG1)
+		hpd->pch_hpd = hpd_sde_dg1;
+	else if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP)
+		hpd->pch_hpd = hpd_icp;
+	else if (HAS_PCH_CNP(dev_priv) || HAS_PCH_SPT(dev_priv))
+		hpd->pch_hpd = hpd_spt;
+	else if (HAS_PCH_LPT(dev_priv) || HAS_PCH_CPT(dev_priv))
+		hpd->pch_hpd = hpd_cpt;
+	else if (HAS_PCH_IBX(dev_priv))
+		hpd->pch_hpd = hpd_ibx;
+	else
+		MISSING_CASE(INTEL_PCH_TYPE(dev_priv));
+}
+
+static void
+intel_handle_vblank(struct drm_i915_private *dev_priv, enum pipe pipe)
+{
+	struct intel_crtc *crtc = intel_crtc_for_pipe(dev_priv, pipe);
+
+	drm_crtc_handle_vblank(&crtc->base);
+}
+
+void gen3_irq_reset(struct intel_uncore *uncore, i915_reg_t imr,
+		    i915_reg_t iir, i915_reg_t ier)
+{
+	intel_uncore_write(uncore, imr, 0xffffffff);
+	intel_uncore_posting_read(uncore, imr);
+
+	intel_uncore_write(uncore, ier, 0);
+
+	/* IIR can theoretically queue up two events. Be paranoid. */
+	intel_uncore_write(uncore, iir, 0xffffffff);
+	intel_uncore_posting_read(uncore, iir);
+	intel_uncore_write(uncore, iir, 0xffffffff);
+	intel_uncore_posting_read(uncore, iir);
+}
+
+void gen2_irq_reset(struct intel_uncore *uncore)
+{
+	intel_uncore_write16(uncore, GEN2_IMR, 0xffff);
+	intel_uncore_posting_read16(uncore, GEN2_IMR);
+
+	intel_uncore_write16(uncore, GEN2_IER, 0);
+
+	/* IIR can theoretically queue up two events. Be paranoid. */
+	intel_uncore_write16(uncore, GEN2_IIR, 0xffff);
+	intel_uncore_posting_read16(uncore, GEN2_IIR);
+	intel_uncore_write16(uncore, GEN2_IIR, 0xffff);
+	intel_uncore_posting_read16(uncore, GEN2_IIR);
+}
+
+/*
+ * We should clear IMR at preinstall/uninstall, and just check at postinstall.
+ */
+static void gen3_assert_iir_is_zero(struct intel_uncore *uncore, i915_reg_t reg)
+{
+	u32 val = intel_uncore_read(uncore, reg);
+
+	if (val == 0)
+		return;
+
+	drm_WARN(&uncore->i915->drm, 1,
+		 "Interrupt register 0x%x is not zero: 0x%08x\n",
+		 i915_mmio_reg_offset(reg), val);
+	intel_uncore_write(uncore, reg, 0xffffffff);
+	intel_uncore_posting_read(uncore, reg);
+	intel_uncore_write(uncore, reg, 0xffffffff);
+	intel_uncore_posting_read(uncore, reg);
+}
+
+static void gen2_assert_iir_is_zero(struct intel_uncore *uncore)
+{
+	u16 val = intel_uncore_read16(uncore, GEN2_IIR);
+
+	if (val == 0)
+		return;
+
+	drm_WARN(&uncore->i915->drm, 1,
+		 "Interrupt register 0x%x is not zero: 0x%08x\n",
+		 i915_mmio_reg_offset(GEN2_IIR), val);
+	intel_uncore_write16(uncore, GEN2_IIR, 0xffff);
+	intel_uncore_posting_read16(uncore, GEN2_IIR);
+	intel_uncore_write16(uncore, GEN2_IIR, 0xffff);
+	intel_uncore_posting_read16(uncore, GEN2_IIR);
+}
+
+void gen3_irq_init(struct intel_uncore *uncore,
+		   i915_reg_t imr, u32 imr_val,
+		   i915_reg_t ier, u32 ier_val,
+		   i915_reg_t iir)
+{
+	gen3_assert_iir_is_zero(uncore, iir);
+
+	intel_uncore_write(uncore, ier, ier_val);
+	intel_uncore_write(uncore, imr, imr_val);
+	intel_uncore_posting_read(uncore, imr);
+}
+
+void gen2_irq_init(struct intel_uncore *uncore,
+		   u32 imr_val, u32 ier_val)
+{
+	gen2_assert_iir_is_zero(uncore);
+
+	intel_uncore_write16(uncore, GEN2_IER, ier_val);
+	intel_uncore_write16(uncore, GEN2_IMR, imr_val);
+	intel_uncore_posting_read16(uncore, GEN2_IMR);
+}
+
+/* For display hotplug interrupt */
+static inline void
+i915_hotplug_interrupt_update_locked(struct drm_i915_private *dev_priv,
+				     u32 mask,
+				     u32 bits)
+{
+	u32 val;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+	drm_WARN_ON(&dev_priv->drm, bits & ~mask);
+
+	val = intel_uncore_read(&dev_priv->uncore, PORT_HOTPLUG_EN);
+	val &= ~mask;
+	val |= bits;
+	intel_uncore_write(&dev_priv->uncore, PORT_HOTPLUG_EN, val);
+}
+
+/**
+ * i915_hotplug_interrupt_update - update hotplug interrupt enable
+ * @dev_priv: driver private
+ * @mask: bits to update
+ * @bits: bits to enable
+ * NOTE: the HPD enable bits are modified both inside and outside
+ * of an interrupt context. To avoid that read-modify-write cycles
+ * interfer, these bits are protected by a spinlock. Since this
+ * function is usually not called from a context where the lock is
+ * held already, this function acquires the lock itself. A non-locking
+ * version is also available.
+ */
+void i915_hotplug_interrupt_update(struct drm_i915_private *dev_priv,
+				   u32 mask,
+				   u32 bits)
+{
+	spin_lock_irq(&dev_priv->irq_lock);
+	i915_hotplug_interrupt_update_locked(dev_priv, mask, bits);
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
+
+/**
+ * ilk_update_display_irq - update DEIMR
+ * @dev_priv: driver private
+ * @interrupt_mask: mask of interrupt bits to update
+ * @enabled_irq_mask: mask of interrupt bits to enable
+ */
+static void ilk_update_display_irq(struct drm_i915_private *dev_priv,
+				   u32 interrupt_mask, u32 enabled_irq_mask)
+{
+	u32 new_val;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+	drm_WARN_ON(&dev_priv->drm, enabled_irq_mask & ~interrupt_mask);
+
+	new_val = dev_priv->irq_mask;
+	new_val &= ~interrupt_mask;
+	new_val |= (~enabled_irq_mask & interrupt_mask);
+
+	if (new_val != dev_priv->irq_mask &&
+	    !drm_WARN_ON(&dev_priv->drm, !intel_irqs_enabled(dev_priv))) {
+		dev_priv->irq_mask = new_val;
+		intel_uncore_write(&dev_priv->uncore, DEIMR, dev_priv->irq_mask);
+		intel_uncore_posting_read(&dev_priv->uncore, DEIMR);
+	}
+}
+
+void ilk_enable_display_irq(struct drm_i915_private *i915, u32 bits)
+{
+	ilk_update_display_irq(i915, bits, bits);
+}
+
+void ilk_disable_display_irq(struct drm_i915_private *i915, u32 bits)
+{
+	ilk_update_display_irq(i915, bits, 0);
+}
+
+/**
+ * bdw_update_port_irq - update DE port interrupt
+ * @dev_priv: driver private
+ * @interrupt_mask: mask of interrupt bits to update
+ * @enabled_irq_mask: mask of interrupt bits to enable
+ */
+static void bdw_update_port_irq(struct drm_i915_private *dev_priv,
+				u32 interrupt_mask,
+				u32 enabled_irq_mask)
+{
+	u32 new_val;
+	u32 old_val;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	drm_WARN_ON(&dev_priv->drm, enabled_irq_mask & ~interrupt_mask);
+
+	if (drm_WARN_ON(&dev_priv->drm, !intel_irqs_enabled(dev_priv)))
+		return;
+
+	old_val = intel_uncore_read(&dev_priv->uncore, GEN8_DE_PORT_IMR);
+
+	new_val = old_val;
+	new_val &= ~interrupt_mask;
+	new_val |= (~enabled_irq_mask & interrupt_mask);
+
+	if (new_val != old_val) {
+		intel_uncore_write(&dev_priv->uncore, GEN8_DE_PORT_IMR, new_val);
+		intel_uncore_posting_read(&dev_priv->uncore, GEN8_DE_PORT_IMR);
+	}
+}
+
+/**
+ * bdw_update_pipe_irq - update DE pipe interrupt
+ * @dev_priv: driver private
+ * @pipe: pipe whose interrupt to update
+ * @interrupt_mask: mask of interrupt bits to update
+ * @enabled_irq_mask: mask of interrupt bits to enable
+ */
+static void bdw_update_pipe_irq(struct drm_i915_private *dev_priv,
+				enum pipe pipe, u32 interrupt_mask,
+				u32 enabled_irq_mask)
+{
+	u32 new_val;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	drm_WARN_ON(&dev_priv->drm, enabled_irq_mask & ~interrupt_mask);
+
+	if (drm_WARN_ON(&dev_priv->drm, !intel_irqs_enabled(dev_priv)))
+		return;
+
+	new_val = dev_priv->de_irq_mask[pipe];
+	new_val &= ~interrupt_mask;
+	new_val |= (~enabled_irq_mask & interrupt_mask);
+
+	if (new_val != dev_priv->de_irq_mask[pipe]) {
+		dev_priv->de_irq_mask[pipe] = new_val;
+		intel_uncore_write(&dev_priv->uncore, GEN8_DE_PIPE_IMR(pipe), dev_priv->de_irq_mask[pipe]);
+		intel_uncore_posting_read(&dev_priv->uncore, GEN8_DE_PIPE_IMR(pipe));
+	}
+}
+
+void bdw_enable_pipe_irq(struct drm_i915_private *i915,
+			 enum pipe pipe, u32 bits)
+{
+	bdw_update_pipe_irq(i915, pipe, bits, bits);
+}
+
+void bdw_disable_pipe_irq(struct drm_i915_private *i915,
+			  enum pipe pipe, u32 bits)
+{
+	bdw_update_pipe_irq(i915, pipe, bits, 0);
+}
+
+/**
+ * ibx_display_interrupt_update - update SDEIMR
+ * @dev_priv: driver private
+ * @interrupt_mask: mask of interrupt bits to update
+ * @enabled_irq_mask: mask of interrupt bits to enable
+ */
+static void ibx_display_interrupt_update(struct drm_i915_private *dev_priv,
+					 u32 interrupt_mask,
+					 u32 enabled_irq_mask)
+{
+	u32 sdeimr = intel_uncore_read(&dev_priv->uncore, SDEIMR);
+	sdeimr &= ~interrupt_mask;
+	sdeimr |= (~enabled_irq_mask & interrupt_mask);
+
+	drm_WARN_ON(&dev_priv->drm, enabled_irq_mask & ~interrupt_mask);
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	if (drm_WARN_ON(&dev_priv->drm, !intel_irqs_enabled(dev_priv)))
+		return;
+
+	intel_uncore_write(&dev_priv->uncore, SDEIMR, sdeimr);
+	intel_uncore_posting_read(&dev_priv->uncore, SDEIMR);
+}
+
+void ibx_enable_display_interrupt(struct drm_i915_private *i915, u32 bits)
+{
+	ibx_display_interrupt_update(i915, bits, bits);
+}
+
+void ibx_disable_display_interrupt(struct drm_i915_private *i915, u32 bits)
+{
+	ibx_display_interrupt_update(i915, bits, 0);
+}
+
+u32 i915_pipestat_enable_mask(struct drm_i915_private *dev_priv,
+			      enum pipe pipe)
+{
+	u32 status_mask = dev_priv->pipestat_irq_mask[pipe];
+	u32 enable_mask = status_mask << 16;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	if (DISPLAY_VER(dev_priv) < 5)
+		goto out;
+
+	/*
+	 * On pipe A we don't support the PSR interrupt yet,
+	 * on pipe B and C the same bit MBZ.
+	 */
+	if (drm_WARN_ON_ONCE(&dev_priv->drm,
+			     status_mask & PIPE_A_PSR_STATUS_VLV))
+		return 0;
+	/*
+	 * On pipe B and C we don't support the PSR interrupt yet, on pipe
+	 * A the same bit is for perf counters which we don't use either.
+	 */
+	if (drm_WARN_ON_ONCE(&dev_priv->drm,
+			     status_mask & PIPE_B_PSR_STATUS_VLV))
+		return 0;
+
+	enable_mask &= ~(PIPE_FIFO_UNDERRUN_STATUS |
+			 SPRITE0_FLIP_DONE_INT_EN_VLV |
+			 SPRITE1_FLIP_DONE_INT_EN_VLV);
+	if (status_mask & SPRITE0_FLIP_DONE_INT_STATUS_VLV)
+		enable_mask |= SPRITE0_FLIP_DONE_INT_EN_VLV;
+	if (status_mask & SPRITE1_FLIP_DONE_INT_STATUS_VLV)
+		enable_mask |= SPRITE1_FLIP_DONE_INT_EN_VLV;
+
+out:
+	drm_WARN_ONCE(&dev_priv->drm,
+		      enable_mask & ~PIPESTAT_INT_ENABLE_MASK ||
+		      status_mask & ~PIPESTAT_INT_STATUS_MASK,
+		      "pipe %c: enable_mask=0x%x, status_mask=0x%x\n",
+		      pipe_name(pipe), enable_mask, status_mask);
+
+	return enable_mask;
+}
+
+void i915_enable_pipestat(struct drm_i915_private *dev_priv,
+			  enum pipe pipe, u32 status_mask)
+{
+	i915_reg_t reg = PIPESTAT(pipe);
+	u32 enable_mask;
+
+	drm_WARN_ONCE(&dev_priv->drm, status_mask & ~PIPESTAT_INT_STATUS_MASK,
+		      "pipe %c: status_mask=0x%x\n",
+		      pipe_name(pipe), status_mask);
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+	drm_WARN_ON(&dev_priv->drm, !intel_irqs_enabled(dev_priv));
+
+	if ((dev_priv->pipestat_irq_mask[pipe] & status_mask) == status_mask)
+		return;
+
+	dev_priv->pipestat_irq_mask[pipe] |= status_mask;
+	enable_mask = i915_pipestat_enable_mask(dev_priv, pipe);
+
+	intel_uncore_write(&dev_priv->uncore, reg, enable_mask | status_mask);
+	intel_uncore_posting_read(&dev_priv->uncore, reg);
+}
+
+void i915_disable_pipestat(struct drm_i915_private *dev_priv,
+			   enum pipe pipe, u32 status_mask)
+{
+	i915_reg_t reg = PIPESTAT(pipe);
+	u32 enable_mask;
+
+	drm_WARN_ONCE(&dev_priv->drm, status_mask & ~PIPESTAT_INT_STATUS_MASK,
+		      "pipe %c: status_mask=0x%x\n",
+		      pipe_name(pipe), status_mask);
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+	drm_WARN_ON(&dev_priv->drm, !intel_irqs_enabled(dev_priv));
+
+	if ((dev_priv->pipestat_irq_mask[pipe] & status_mask) == 0)
+		return;
+
+	dev_priv->pipestat_irq_mask[pipe] &= ~status_mask;
+	enable_mask = i915_pipestat_enable_mask(dev_priv, pipe);
+
+	intel_uncore_write(&dev_priv->uncore, reg, enable_mask | status_mask);
+	intel_uncore_posting_read(&dev_priv->uncore, reg);
+}
+
+static bool i915_has_asle(struct drm_i915_private *dev_priv)
+{
+	if (!dev_priv->display.opregion.asle)
+		return false;
+
+	return IS_PINEVIEW(dev_priv) || IS_MOBILE(dev_priv);
+}
+
+/**
+ * i915_enable_asle_pipestat - enable ASLE pipestat for OpRegion
+ * @dev_priv: i915 device private
+ */
+static void i915_enable_asle_pipestat(struct drm_i915_private *dev_priv)
+{
+	if (!i915_has_asle(dev_priv))
+		return;
+
+	spin_lock_irq(&dev_priv->irq_lock);
+
+	i915_enable_pipestat(dev_priv, PIPE_B, PIPE_LEGACY_BLC_EVENT_STATUS);
+	if (DISPLAY_VER(dev_priv) >= 4)
+		i915_enable_pipestat(dev_priv, PIPE_A,
+				     PIPE_LEGACY_BLC_EVENT_STATUS);
+
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
+
+/*
+ * This timing diagram depicts the video signal in and
+ * around the vertical blanking period.
+ *
+ * Assumptions about the fictitious mode used in this example:
+ *  vblank_start >= 3
+ *  vsync_start = vblank_start + 1
+ *  vsync_end = vblank_start + 2
+ *  vtotal = vblank_start + 3
+ *
+ *           start of vblank:
+ *           latch double buffered registers
+ *           increment frame counter (ctg+)
+ *           generate start of vblank interrupt (gen4+)
+ *           |
+ *           |          frame start:
+ *           |          generate frame start interrupt (aka. vblank interrupt) (gmch)
+ *           |          may be shifted forward 1-3 extra lines via PIPECONF
+ *           |          |
+ *           |          |  start of vsync:
+ *           |          |  generate vsync interrupt
+ *           |          |  |
+ * ___xxxx___    ___xxxx___    ___xxxx___    ___xxxx___    ___xxxx___    ___xxxx
+ *       .   \hs/   .      \hs/          \hs/          \hs/   .      \hs/
+ * ----va---> <-----------------vb--------------------> <--------va-------------
+ *       |          |       <----vs----->                     |
+ * -vbs-----> <---vbs+1---> <---vbs+2---> <-----0-----> <-----1-----> <-----2--- (scanline counter gen2)
+ * -vbs-2---> <---vbs-1---> <---vbs-----> <---vbs+1---> <---vbs+2---> <-----0--- (scanline counter gen3+)
+ * -vbs-2---> <---vbs-2---> <---vbs-1---> <---vbs-----> <---vbs+1---> <---vbs+2- (scanline counter hsw+ hdmi)
+ *       |          |                                         |
+ *       last visible pixel                                   first visible pixel
+ *                  |                                         increment frame counter (gen3/4)
+ *                  pixel counter = vblank_start * htotal     pixel counter = 0 (gen3/4)
+ *
+ * x  = horizontal active
+ * _  = horizontal blanking
+ * hs = horizontal sync
+ * va = vertical active
+ * vb = vertical blanking
+ * vs = vertical sync
+ * vbs = vblank_start (number)
+ *
+ * Summary:
+ * - most events happen at the start of horizontal sync
+ * - frame start happens at the start of horizontal blank, 1-4 lines
+ *   (depending on PIPECONF settings) after the start of vblank
+ * - gen3/4 pixel and frame counter are synchronized with the start
+ *   of horizontal active on the first line of vertical active
+ */
+
+/* Called from drm generic code, passed a 'crtc', which
+ * we use as a pipe index
+ */
+u32 i915_get_vblank_counter(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+	struct drm_vblank_crtc *vblank = &dev_priv->drm.vblank[drm_crtc_index(crtc)];
+	const struct drm_display_mode *mode = &vblank->hwmode;
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	i915_reg_t high_frame, low_frame;
+	u32 high1, high2, low, pixel, vbl_start, hsync_start, htotal;
+	unsigned long irqflags;
+
+	/*
+	 * On i965gm TV output the frame counter only works up to
+	 * the point when we enable the TV encoder. After that the
+	 * frame counter ceases to work and reads zero. We need a
+	 * vblank wait before enabling the TV encoder and so we
+	 * have to enable vblank interrupts while the frame counter
+	 * is still in a working state. However the core vblank code
+	 * does not like us returning non-zero frame counter values
+	 * when we've told it that we don't have a working frame
+	 * counter. Thus we must stop non-zero values leaking out.
+	 */
+	if (!vblank->max_vblank_count)
+		return 0;
+
+	htotal = mode->crtc_htotal;
+	hsync_start = mode->crtc_hsync_start;
+	vbl_start = mode->crtc_vblank_start;
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		vbl_start = DIV_ROUND_UP(vbl_start, 2);
+
+	/* Convert to pixel count */
+	vbl_start *= htotal;
+
+	/* Start of vblank event occurs at start of hsync */
+	vbl_start -= htotal - hsync_start;
+
+	high_frame = PIPEFRAME(pipe);
+	low_frame = PIPEFRAMEPIXEL(pipe);
+
+	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
+
+	/*
+	 * High & low register fields aren't synchronized, so make sure
+	 * we get a low value that's stable across two reads of the high
+	 * register.
+	 */
+	do {
+		high1 = intel_de_read_fw(dev_priv, high_frame) & PIPE_FRAME_HIGH_MASK;
+		low   = intel_de_read_fw(dev_priv, low_frame);
+		high2 = intel_de_read_fw(dev_priv, high_frame) & PIPE_FRAME_HIGH_MASK;
+	} while (high1 != high2);
+
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
+
+	high1 >>= PIPE_FRAME_HIGH_SHIFT;
+	pixel = low & PIPE_PIXEL_MASK;
+	low >>= PIPE_FRAME_LOW_SHIFT;
+
+	/*
+	 * The frame counter increments at beginning of active.
+	 * Cook up a vblank counter by also checking the pixel
+	 * counter against vblank start.
+	 */
+	return (((high1 << 8) | low) + (pixel >= vbl_start)) & 0xffffff;
+}
+
+u32 g4x_get_vblank_counter(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+	struct drm_vblank_crtc *vblank = &dev_priv->drm.vblank[drm_crtc_index(crtc)];
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+
+	if (!vblank->max_vblank_count)
+		return 0;
+
+	return intel_uncore_read(&dev_priv->uncore, PIPE_FRMCOUNT_G4X(pipe));
+}
+
+static u32 intel_crtc_scanlines_since_frame_timestamp(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct drm_vblank_crtc *vblank =
+		&crtc->base.dev->vblank[drm_crtc_index(&crtc->base)];
+	const struct drm_display_mode *mode = &vblank->hwmode;
+	u32 htotal = mode->crtc_htotal;
+	u32 clock = mode->crtc_clock;
+	u32 scan_prev_time, scan_curr_time, scan_post_time;
+
+	/*
+	 * To avoid the race condition where we might cross into the
+	 * next vblank just between the PIPE_FRMTMSTMP and TIMESTAMP_CTR
+	 * reads. We make sure we read PIPE_FRMTMSTMP and TIMESTAMP_CTR
+	 * during the same frame.
+	 */
+	do {
+		/*
+		 * This field provides read back of the display
+		 * pipe frame time stamp. The time stamp value
+		 * is sampled at every start of vertical blank.
+		 */
+		scan_prev_time = intel_de_read_fw(dev_priv,
+						  PIPE_FRMTMSTMP(crtc->pipe));
+
+		/*
+		 * The TIMESTAMP_CTR register has the current
+		 * time stamp value.
+		 */
+		scan_curr_time = intel_de_read_fw(dev_priv, IVB_TIMESTAMP_CTR);
+
+		scan_post_time = intel_de_read_fw(dev_priv,
+						  PIPE_FRMTMSTMP(crtc->pipe));
+	} while (scan_post_time != scan_prev_time);
+
+	return div_u64(mul_u32_u32(scan_curr_time - scan_prev_time,
+				   clock), 1000 * htotal);
+}
+
+/*
+ * On certain encoders on certain platforms, pipe
+ * scanline register will not work to get the scanline,
+ * since the timings are driven from the PORT or issues
+ * with scanline register updates.
+ * This function will use Framestamp and current
+ * timestamp registers to calculate the scanline.
+ */
+static u32 __intel_get_crtc_scanline_from_timestamp(struct intel_crtc *crtc)
+{
+	struct drm_vblank_crtc *vblank =
+		&crtc->base.dev->vblank[drm_crtc_index(&crtc->base)];
+	const struct drm_display_mode *mode = &vblank->hwmode;
+	u32 vblank_start = mode->crtc_vblank_start;
+	u32 vtotal = mode->crtc_vtotal;
+	u32 scanline;
+
+	scanline = intel_crtc_scanlines_since_frame_timestamp(crtc);
+	scanline = min(scanline, vtotal - 1);
+	scanline = (scanline + vblank_start) % vtotal;
+
+	return scanline;
+}
+
+/*
+ * intel_de_read_fw(), only for fast reads of display block, no need for
+ * forcewake etc.
+ */
+static int __intel_get_crtc_scanline(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	const struct drm_display_mode *mode;
+	struct drm_vblank_crtc *vblank;
+	enum pipe pipe = crtc->pipe;
+	int position, vtotal;
+
+	if (!crtc->active)
+		return 0;
+
+	vblank = &crtc->base.dev->vblank[drm_crtc_index(&crtc->base)];
+	mode = &vblank->hwmode;
+
+	if (crtc->mode_flags & I915_MODE_FLAG_GET_SCANLINE_FROM_TIMESTAMP)
+		return __intel_get_crtc_scanline_from_timestamp(crtc);
+
+	vtotal = mode->crtc_vtotal;
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		vtotal /= 2;
+
+	position = intel_de_read_fw(dev_priv, PIPEDSL(pipe)) & PIPEDSL_LINE_MASK;
+
+	/*
+	 * On HSW, the DSL reg (0x70000) appears to return 0 if we
+	 * read it just before the start of vblank.  So try it again
+	 * so we don't accidentally end up spanning a vblank frame
+	 * increment, causing the pipe_update_end() code to squak at us.
+	 *
+	 * The nature of this problem means we can't simply check the ISR
+	 * bit and return the vblank start value; nor can we use the scanline
+	 * debug register in the transcoder as it appears to have the same
+	 * problem.  We may need to extend this to include other platforms,
+	 * but so far testing only shows the problem on HSW.
+	 */
+	if (HAS_DDI(dev_priv) && !position) {
+		int i, temp;
+
+		for (i = 0; i < 100; i++) {
+			udelay(1);
+			temp = intel_de_read_fw(dev_priv, PIPEDSL(pipe)) & PIPEDSL_LINE_MASK;
+			if (temp != position) {
+				position = temp;
+				break;
+			}
+		}
+	}
+
+	/*
+	 * See update_scanline_offset() for the details on the
+	 * scanline_offset adjustment.
+	 */
+	return (position + crtc->scanline_offset) % vtotal;
+}
+
+static bool i915_get_crtc_scanoutpos(struct drm_crtc *_crtc,
+				     bool in_vblank_irq,
+				     int *vpos, int *hpos,
+				     ktime_t *stime, ktime_t *etime,
+				     const struct drm_display_mode *mode)
+{
+	struct drm_device *dev = _crtc->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_crtc *crtc = to_intel_crtc(_crtc);
+	enum pipe pipe = crtc->pipe;
+	int position;
+	int vbl_start, vbl_end, hsync_start, htotal, vtotal;
+	unsigned long irqflags;
+	bool use_scanline_counter = DISPLAY_VER(dev_priv) >= 5 ||
+		IS_G4X(dev_priv) || DISPLAY_VER(dev_priv) == 2 ||
+		crtc->mode_flags & I915_MODE_FLAG_USE_SCANLINE_COUNTER;
+
+	if (drm_WARN_ON(&dev_priv->drm, !mode->crtc_clock)) {
+		drm_dbg(&dev_priv->drm,
+			"trying to get scanoutpos for disabled "
+			"pipe %c\n", pipe_name(pipe));
+		return false;
+	}
+
+	htotal = mode->crtc_htotal;
+	hsync_start = mode->crtc_hsync_start;
+	vtotal = mode->crtc_vtotal;
+	vbl_start = mode->crtc_vblank_start;
+	vbl_end = mode->crtc_vblank_end;
+
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE) {
+		vbl_start = DIV_ROUND_UP(vbl_start, 2);
+		vbl_end /= 2;
+		vtotal /= 2;
+	}
+
+	/*
+	 * Lock uncore.lock, as we will do multiple timing critical raw
+	 * register reads, potentially with preemption disabled, so the
+	 * following code must not block on uncore.lock.
+	 */
+	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
+
+	/* preempt_disable_rt() should go right here in PREEMPT_RT patchset. */
+
+	/* Get optional system timestamp before query. */
+	if (stime)
+		*stime = ktime_get();
+
+	if (crtc->mode_flags & I915_MODE_FLAG_VRR) {
+		int scanlines = intel_crtc_scanlines_since_frame_timestamp(crtc);
+
+		position = __intel_get_crtc_scanline(crtc);
+
+		/*
+		 * Already exiting vblank? If so, shift our position
+		 * so it looks like we're already apporaching the full
+		 * vblank end. This should make the generated timestamp
+		 * more or less match when the active portion will start.
+		 */
+		if (position >= vbl_start && scanlines < position)
+			position = min(crtc->vmax_vblank_start + scanlines, vtotal - 1);
+	} else if (use_scanline_counter) {
+		/* No obvious pixelcount register. Only query vertical
+		 * scanout position from Display scan line register.
+		 */
+		position = __intel_get_crtc_scanline(crtc);
+	} else {
+		/* Have access to pixelcount since start of frame.
+		 * We can split this into vertical and horizontal
+		 * scanout position.
+		 */
+		position = (intel_de_read_fw(dev_priv, PIPEFRAMEPIXEL(pipe)) & PIPE_PIXEL_MASK) >> PIPE_PIXEL_SHIFT;
+
+		/* convert to pixel counts */
+		vbl_start *= htotal;
+		vbl_end *= htotal;
+		vtotal *= htotal;
+
+		/*
+		 * In interlaced modes, the pixel counter counts all pixels,
+		 * so one field will have htotal more pixels. In order to avoid
+		 * the reported position from jumping backwards when the pixel
+		 * counter is beyond the length of the shorter field, just
+		 * clamp the position the length of the shorter field. This
+		 * matches how the scanline counter based position works since
+		 * the scanline counter doesn't count the two half lines.
+		 */
+		if (position >= vtotal)
+			position = vtotal - 1;
+
+		/*
+		 * Start of vblank interrupt is triggered at start of hsync,
+		 * just prior to the first active line of vblank. However we
+		 * consider lines to start at the leading edge of horizontal
+		 * active. So, should we get here before we've crossed into
+		 * the horizontal active of the first line in vblank, we would
+		 * not set the DRM_SCANOUTPOS_INVBL flag. In order to fix that,
+		 * always add htotal-hsync_start to the current pixel position.
+		 */
+		position = (position + htotal - hsync_start) % vtotal;
+	}
+
+	/* Get optional system timestamp after query. */
+	if (etime)
+		*etime = ktime_get();
+
+	/* preempt_enable_rt() should go right here in PREEMPT_RT patchset. */
+
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
+
+	/*
+	 * While in vblank, position will be negative
+	 * counting up towards 0 at vbl_end. And outside
+	 * vblank, position will be positive counting
+	 * up since vbl_end.
+	 */
+	if (position >= vbl_start)
+		position -= vbl_end;
+	else
+		position += vtotal - vbl_end;
+
+	if (use_scanline_counter) {
+		*vpos = position;
+		*hpos = 0;
+	} else {
+		*vpos = position / htotal;
+		*hpos = position - (*vpos * htotal);
+	}
+
+	return true;
+}
+
+bool intel_crtc_get_vblank_timestamp(struct drm_crtc *crtc, int *max_error,
+				     ktime_t *vblank_time, bool in_vblank_irq)
+{
+	return drm_crtc_vblank_helper_get_vblank_timestamp_internal(
+		crtc, max_error, vblank_time, in_vblank_irq,
+		i915_get_crtc_scanoutpos);
+}
+
+int intel_get_crtc_scanline(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	unsigned long irqflags;
+	int position;
+
+	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
+	position = __intel_get_crtc_scanline(crtc);
+	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
+
+	return position;
+}
+
+/**
+ * ivb_parity_work - Workqueue called when a parity error interrupt
+ * occurred.
+ * @work: workqueue struct
+ *
+ * Doesn't actually do anything except notify userspace. As a consequence of
+ * this event, userspace should try to remap the bad rows since statistically
+ * it is likely the same row is more likely to go bad again.
+ */
+static void ivb_parity_work(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, typeof(*dev_priv), l3_parity.error_work);
+	struct intel_gt *gt = to_gt(dev_priv);
+	u32 error_status, row, bank, subbank;
+	char *parity_event[6];
+	u32 misccpctl;
+	u8 slice = 0;
+
+	/* We must turn off DOP level clock gating to access the L3 registers.
+	 * In order to prevent a get/put style interface, acquire struct mutex
+	 * any time we access those registers.
+	 */
+	mutex_lock(&dev_priv->drm.struct_mutex);
+
+	/* If we've screwed up tracking, just let the interrupt fire again */
+	if (drm_WARN_ON(&dev_priv->drm, !dev_priv->l3_parity.which_slice))
+		goto out;
+
+	misccpctl = intel_uncore_read(&dev_priv->uncore, GEN7_MISCCPCTL);
+	intel_uncore_write(&dev_priv->uncore, GEN7_MISCCPCTL, misccpctl & ~GEN7_DOP_CLOCK_GATE_ENABLE);
+	intel_uncore_posting_read(&dev_priv->uncore, GEN7_MISCCPCTL);
+
+	while ((slice = ffs(dev_priv->l3_parity.which_slice)) != 0) {
+		i915_reg_t reg;
+
+		slice--;
+		if (drm_WARN_ON_ONCE(&dev_priv->drm,
+				     slice >= NUM_L3_SLICES(dev_priv)))
+			break;
+
+		dev_priv->l3_parity.which_slice &= ~(1<<slice);
+
+		reg = GEN7_L3CDERRST1(slice);
+
+		error_status = intel_uncore_read(&dev_priv->uncore, reg);
+		row = GEN7_PARITY_ERROR_ROW(error_status);
+		bank = GEN7_PARITY_ERROR_BANK(error_status);
+		subbank = GEN7_PARITY_ERROR_SUBBANK(error_status);
+
+		intel_uncore_write(&dev_priv->uncore, reg, GEN7_PARITY_ERROR_VALID | GEN7_L3CDERRST1_ENABLE);
+		intel_uncore_posting_read(&dev_priv->uncore, reg);
+
+		parity_event[0] = I915_L3_PARITY_UEVENT "=1";
+		parity_event[1] = kasprintf(GFP_KERNEL, "ROW=%d", row);
+		parity_event[2] = kasprintf(GFP_KERNEL, "BANK=%d", bank);
+		parity_event[3] = kasprintf(GFP_KERNEL, "SUBBANK=%d", subbank);
+		parity_event[4] = kasprintf(GFP_KERNEL, "SLICE=%d", slice);
+		parity_event[5] = NULL;
+
+		kobject_uevent_env(&dev_priv->drm.primary->kdev->kobj,
+				   KOBJ_CHANGE, parity_event);
+
+		DRM_DEBUG("Parity error: Slice = %d, Row = %d, Bank = %d, Sub bank = %d.\n",
+			  slice, row, bank, subbank);
+
+		kfree(parity_event[4]);
+		kfree(parity_event[3]);
+		kfree(parity_event[2]);
+		kfree(parity_event[1]);
+	}
+
+	intel_uncore_write(&dev_priv->uncore, GEN7_MISCCPCTL, misccpctl);
+
+out:
+	drm_WARN_ON(&dev_priv->drm, dev_priv->l3_parity.which_slice);
+	spin_lock_irq(gt->irq_lock);
+	gen5_gt_enable_irq(gt, GT_PARITY_ERROR(dev_priv));
+	spin_unlock_irq(gt->irq_lock);
+
+	mutex_unlock(&dev_priv->drm.struct_mutex);
+}
+
+static bool gen11_port_hotplug_long_detect(enum hpd_pin pin, u32 val)
+{
+	switch (pin) {
+	case HPD_PORT_TC1:
+	case HPD_PORT_TC2:
+	case HPD_PORT_TC3:
+	case HPD_PORT_TC4:
+	case HPD_PORT_TC5:
+	case HPD_PORT_TC6:
+		return val & GEN11_HOTPLUG_CTL_LONG_DETECT(pin);
+	default:
+		return false;
+	}
+}
+
+static bool bxt_port_hotplug_long_detect(enum hpd_pin pin, u32 val)
+{
+	switch (pin) {
+	case HPD_PORT_A:
+		return val & PORTA_HOTPLUG_LONG_DETECT;
+	case HPD_PORT_B:
+		return val & PORTB_HOTPLUG_LONG_DETECT;
+	case HPD_PORT_C:
+		return val & PORTC_HOTPLUG_LONG_DETECT;
+	default:
+		return false;
+	}
+}
+
+static bool icp_ddi_port_hotplug_long_detect(enum hpd_pin pin, u32 val)
+{
+	switch (pin) {
+	case HPD_PORT_A:
+	case HPD_PORT_B:
+	case HPD_PORT_C:
+	case HPD_PORT_D:
+		return val & SHOTPLUG_CTL_DDI_HPD_LONG_DETECT(pin);
+	default:
+		return false;
+	}
+}
+
+static bool icp_tc_port_hotplug_long_detect(enum hpd_pin pin, u32 val)
+{
+	switch (pin) {
+	case HPD_PORT_TC1:
+	case HPD_PORT_TC2:
+	case HPD_PORT_TC3:
+	case HPD_PORT_TC4:
+	case HPD_PORT_TC5:
+	case HPD_PORT_TC6:
+		return val & ICP_TC_HPD_LONG_DETECT(pin);
+	default:
+		return false;
+	}
+}
+
+static bool spt_port_hotplug2_long_detect(enum hpd_pin pin, u32 val)
+{
+	switch (pin) {
+	case HPD_PORT_E:
+		return val & PORTE_HOTPLUG_LONG_DETECT;
+	default:
+		return false;
+	}
+}
+
+static bool spt_port_hotplug_long_detect(enum hpd_pin pin, u32 val)
+{
+	switch (pin) {
+	case HPD_PORT_A:
+		return val & PORTA_HOTPLUG_LONG_DETECT;
+	case HPD_PORT_B:
+		return val & PORTB_HOTPLUG_LONG_DETECT;
+	case HPD_PORT_C:
+		return val & PORTC_HOTPLUG_LONG_DETECT;
+	case HPD_PORT_D:
+		return val & PORTD_HOTPLUG_LONG_DETECT;
+	default:
+		return false;
+	}
+}
+
+static bool ilk_port_hotplug_long_detect(enum hpd_pin pin, u32 val)
+{
+	switch (pin) {
+	case HPD_PORT_A:
+		return val & DIGITAL_PORTA_HOTPLUG_LONG_DETECT;
+	default:
+		return false;
+	}
+}
+
+static bool pch_port_hotplug_long_detect(enum hpd_pin pin, u32 val)
+{
+	switch (pin) {
+	case HPD_PORT_B:
+		return val & PORTB_HOTPLUG_LONG_DETECT;
+	case HPD_PORT_C:
+		return val & PORTC_HOTPLUG_LONG_DETECT;
+	case HPD_PORT_D:
+		return val & PORTD_HOTPLUG_LONG_DETECT;
+	default:
+		return false;
+	}
+}
+
+static bool i9xx_port_hotplug_long_detect(enum hpd_pin pin, u32 val)
+{
+	switch (pin) {
+	case HPD_PORT_B:
+		return val & PORTB_HOTPLUG_INT_LONG_PULSE;
+	case HPD_PORT_C:
+		return val & PORTC_HOTPLUG_INT_LONG_PULSE;
+	case HPD_PORT_D:
+		return val & PORTD_HOTPLUG_INT_LONG_PULSE;
+	default:
+		return false;
+	}
+}
+
+/*
+ * Get a bit mask of pins that have triggered, and which ones may be long.
+ * This can be called multiple times with the same masks to accumulate
+ * hotplug detection results from several registers.
+ *
+ * Note that the caller is expected to zero out the masks initially.
+ */
+static void intel_get_hpd_pins(struct drm_i915_private *dev_priv,
+			       u32 *pin_mask, u32 *long_mask,
+			       u32 hotplug_trigger, u32 dig_hotplug_reg,
+			       const u32 hpd[HPD_NUM_PINS],
+			       bool long_pulse_detect(enum hpd_pin pin, u32 val))
+{
+	enum hpd_pin pin;
+
+	BUILD_BUG_ON(BITS_PER_TYPE(*pin_mask) < HPD_NUM_PINS);
+
+	for_each_hpd_pin(pin) {
+		if ((hpd[pin] & hotplug_trigger) == 0)
+			continue;
+
+		*pin_mask |= BIT(pin);
+
+		if (long_pulse_detect(pin, dig_hotplug_reg))
+			*long_mask |= BIT(pin);
+	}
+
+	drm_dbg(&dev_priv->drm,
+		"hotplug event received, stat 0x%08x, dig 0x%08x, pins 0x%08x, long 0x%08x\n",
+		hotplug_trigger, dig_hotplug_reg, *pin_mask, *long_mask);
+
+}
+
+static u32 intel_hpd_enabled_irqs(struct drm_i915_private *dev_priv,
+				  const u32 hpd[HPD_NUM_PINS])
+{
+	struct intel_encoder *encoder;
+	u32 enabled_irqs = 0;
+
+	for_each_intel_encoder(&dev_priv->drm, encoder)
+		if (dev_priv->display.hotplug.stats[encoder->hpd_pin].state == HPD_ENABLED)
+			enabled_irqs |= hpd[encoder->hpd_pin];
+
+	return enabled_irqs;
+}
+
+static u32 intel_hpd_hotplug_irqs(struct drm_i915_private *dev_priv,
+				  const u32 hpd[HPD_NUM_PINS])
+{
+	struct intel_encoder *encoder;
+	u32 hotplug_irqs = 0;
+
+	for_each_intel_encoder(&dev_priv->drm, encoder)
+		hotplug_irqs |= hpd[encoder->hpd_pin];
+
+	return hotplug_irqs;
+}
+
+static u32 intel_hpd_hotplug_enables(struct drm_i915_private *i915,
+				     hotplug_enables_func hotplug_enables)
+{
+	struct intel_encoder *encoder;
+	u32 hotplug = 0;
+
+	for_each_intel_encoder(&i915->drm, encoder)
+		hotplug |= hotplug_enables(i915, encoder->hpd_pin);
+
+	return hotplug;
+}
+
+static void gmbus_irq_handler(struct drm_i915_private *dev_priv)
+{
+	wake_up_all(&dev_priv->display.gmbus.wait_queue);
+}
+
+static void dp_aux_irq_handler(struct drm_i915_private *dev_priv)
+{
+	wake_up_all(&dev_priv->display.gmbus.wait_queue);
+}
+
+#if defined(CONFIG_DEBUG_FS)
+static void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv,
+					 enum pipe pipe,
+					 u32 crc0, u32 crc1,
+					 u32 crc2, u32 crc3,
+					 u32 crc4)
+{
+	struct intel_crtc *crtc = intel_crtc_for_pipe(dev_priv, pipe);
+	struct intel_pipe_crc *pipe_crc = &crtc->pipe_crc;
+	u32 crcs[5] = { crc0, crc1, crc2, crc3, crc4 };
+
+	trace_intel_pipe_crc(crtc, crcs);
+
+	spin_lock(&pipe_crc->lock);
+	/*
+	 * For some not yet identified reason, the first CRC is
+	 * bonkers. So let's just wait for the next vblank and read
+	 * out the buggy result.
+	 *
+	 * On GEN8+ sometimes the second CRC is bonkers as well, so
+	 * don't trust that one either.
+	 */
+	if (pipe_crc->skipped <= 0 ||
+	    (DISPLAY_VER(dev_priv) >= 8 && pipe_crc->skipped == 1)) {
+		pipe_crc->skipped++;
+		spin_unlock(&pipe_crc->lock);
+		return;
+	}
+	spin_unlock(&pipe_crc->lock);
+
+	drm_crtc_add_crc_entry(&crtc->base, true,
+				drm_crtc_accurate_vblank_count(&crtc->base),
+				crcs);
+}
+#else
+static inline void
+display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv,
+			     enum pipe pipe,
+			     u32 crc0, u32 crc1,
+			     u32 crc2, u32 crc3,
+			     u32 crc4) {}
+#endif
+
+static void flip_done_handler(struct drm_i915_private *i915,
+			      enum pipe pipe)
+{
+	struct intel_crtc *crtc = intel_crtc_for_pipe(i915, pipe);
+	struct drm_crtc_state *crtc_state = crtc->base.state;
+	struct drm_pending_vblank_event *e = crtc_state->event;
+	struct drm_device *dev = &i915->drm;
+	unsigned long irqflags;
+
+	spin_lock_irqsave(&dev->event_lock, irqflags);
+
+	crtc_state->event = NULL;
+
+	drm_crtc_send_vblank_event(&crtc->base, e);
+
+	spin_unlock_irqrestore(&dev->event_lock, irqflags);
+}
+
+static void hsw_pipe_crc_irq_handler(struct drm_i915_private *dev_priv,
+				     enum pipe pipe)
+{
+	display_pipe_crc_irq_handler(dev_priv, pipe,
+				     intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_1_IVB(pipe)),
+				     0, 0, 0, 0);
+}
+
+static void ivb_pipe_crc_irq_handler(struct drm_i915_private *dev_priv,
+				     enum pipe pipe)
+{
+	display_pipe_crc_irq_handler(dev_priv, pipe,
+				     intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_1_IVB(pipe)),
+				     intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_2_IVB(pipe)),
+				     intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_3_IVB(pipe)),
+				     intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_4_IVB(pipe)),
+				     intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_5_IVB(pipe)));
+}
+
+static void i9xx_pipe_crc_irq_handler(struct drm_i915_private *dev_priv,
+				      enum pipe pipe)
+{
+	u32 res1, res2;
+
+	if (DISPLAY_VER(dev_priv) >= 3)
+		res1 = intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_RES1_I915(pipe));
+	else
+		res1 = 0;
+
+	if (DISPLAY_VER(dev_priv) >= 5 || IS_G4X(dev_priv))
+		res2 = intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_RES2_G4X(pipe));
+	else
+		res2 = 0;
+
+	display_pipe_crc_irq_handler(dev_priv, pipe,
+				     intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_RED(pipe)),
+				     intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_GREEN(pipe)),
+				     intel_uncore_read(&dev_priv->uncore, PIPE_CRC_RES_BLUE(pipe)),
+				     res1, res2);
+}
+
+static void i9xx_pipestat_irq_reset(struct drm_i915_private *dev_priv)
+{
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe) {
+		intel_uncore_write(&dev_priv->uncore, PIPESTAT(pipe),
+			   PIPESTAT_INT_STATUS_MASK |
+			   PIPE_FIFO_UNDERRUN_STATUS);
+
+		dev_priv->pipestat_irq_mask[pipe] = 0;
+	}
+}
+
+static void i9xx_pipestat_irq_ack(struct drm_i915_private *dev_priv,
+				  u32 iir, u32 pipe_stats[I915_MAX_PIPES])
+{
+	enum pipe pipe;
+
+	spin_lock(&dev_priv->irq_lock);
+
+	if (!dev_priv->display_irqs_enabled) {
+		spin_unlock(&dev_priv->irq_lock);
+		return;
+	}
+
+	for_each_pipe(dev_priv, pipe) {
+		i915_reg_t reg;
+		u32 status_mask, enable_mask, iir_bit = 0;
+
+		/*
+		 * PIPESTAT bits get signalled even when the interrupt is
+		 * disabled with the mask bits, and some of the status bits do
+		 * not generate interrupts at all (like the underrun bit). Hence
+		 * we need to be careful that we only handle what we want to
+		 * handle.
+		 */
+
+		/* fifo underruns are filterered in the underrun handler. */
+		status_mask = PIPE_FIFO_UNDERRUN_STATUS;
+
+		switch (pipe) {
+		default:
+		case PIPE_A:
+			iir_bit = I915_DISPLAY_PIPE_A_EVENT_INTERRUPT;
+			break;
+		case PIPE_B:
+			iir_bit = I915_DISPLAY_PIPE_B_EVENT_INTERRUPT;
+			break;
+		case PIPE_C:
+			iir_bit = I915_DISPLAY_PIPE_C_EVENT_INTERRUPT;
+			break;
+		}
+		if (iir & iir_bit)
+			status_mask |= dev_priv->pipestat_irq_mask[pipe];
+
+		if (!status_mask)
+			continue;
+
+		reg = PIPESTAT(pipe);
+		pipe_stats[pipe] = intel_uncore_read(&dev_priv->uncore, reg) & status_mask;
+		enable_mask = i915_pipestat_enable_mask(dev_priv, pipe);
+
+		/*
+		 * Clear the PIPE*STAT regs before the IIR
+		 *
+		 * Toggle the enable bits to make sure we get an
+		 * edge in the ISR pipe event bit if we don't clear
+		 * all the enabled status bits. Otherwise the edge
+		 * triggered IIR on i965/g4x wouldn't notice that
+		 * an interrupt is still pending.
+		 */
+		if (pipe_stats[pipe]) {
+			intel_uncore_write(&dev_priv->uncore, reg, pipe_stats[pipe]);
+			intel_uncore_write(&dev_priv->uncore, reg, enable_mask);
+		}
+	}
+	spin_unlock(&dev_priv->irq_lock);
+}
+
+static void i8xx_pipestat_irq_handler(struct drm_i915_private *dev_priv,
+				      u16 iir, u32 pipe_stats[I915_MAX_PIPES])
+{
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe) {
+		if (pipe_stats[pipe] & PIPE_VBLANK_INTERRUPT_STATUS)
+			intel_handle_vblank(dev_priv, pipe);
+
+		if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
+			i9xx_pipe_crc_irq_handler(dev_priv, pipe);
+
+		if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
+			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
+	}
+}
+
+static void i915_pipestat_irq_handler(struct drm_i915_private *dev_priv,
+				      u32 iir, u32 pipe_stats[I915_MAX_PIPES])
+{
+	bool blc_event = false;
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe) {
+		if (pipe_stats[pipe] & PIPE_VBLANK_INTERRUPT_STATUS)
+			intel_handle_vblank(dev_priv, pipe);
+
+		if (pipe_stats[pipe] & PIPE_LEGACY_BLC_EVENT_STATUS)
+			blc_event = true;
+
+		if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
+			i9xx_pipe_crc_irq_handler(dev_priv, pipe);
+
+		if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
+			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
+	}
+
+	if (blc_event || (iir & I915_ASLE_INTERRUPT))
+		intel_opregion_asle_intr(dev_priv);
+}
+
+static void i965_pipestat_irq_handler(struct drm_i915_private *dev_priv,
+				      u32 iir, u32 pipe_stats[I915_MAX_PIPES])
+{
+	bool blc_event = false;
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe) {
+		if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS)
+			intel_handle_vblank(dev_priv, pipe);
+
+		if (pipe_stats[pipe] & PIPE_LEGACY_BLC_EVENT_STATUS)
+			blc_event = true;
+
+		if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
+			i9xx_pipe_crc_irq_handler(dev_priv, pipe);
+
+		if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
+			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
+	}
+
+	if (blc_event || (iir & I915_ASLE_INTERRUPT))
+		intel_opregion_asle_intr(dev_priv);
+
+	if (pipe_stats[0] & PIPE_GMBUS_INTERRUPT_STATUS)
+		gmbus_irq_handler(dev_priv);
+}
+
+static void valleyview_pipestat_irq_handler(struct drm_i915_private *dev_priv,
+					    u32 pipe_stats[I915_MAX_PIPES])
+{
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe) {
+		if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS)
+			intel_handle_vblank(dev_priv, pipe);
+
+		if (pipe_stats[pipe] & PLANE_FLIP_DONE_INT_STATUS_VLV)
+			flip_done_handler(dev_priv, pipe);
+
+		if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
+			i9xx_pipe_crc_irq_handler(dev_priv, pipe);
+
+		if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
+			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
+	}
+
+	if (pipe_stats[0] & PIPE_GMBUS_INTERRUPT_STATUS)
+		gmbus_irq_handler(dev_priv);
+}
+
+static u32 i9xx_hpd_irq_ack(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug_status = 0, hotplug_status_mask;
+	int i;
+
+	if (IS_G4X(dev_priv) ||
+	    IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		hotplug_status_mask = HOTPLUG_INT_STATUS_G4X |
+			DP_AUX_CHANNEL_MASK_INT_STATUS_G4X;
+	else
+		hotplug_status_mask = HOTPLUG_INT_STATUS_I915;
+
+	/*
+	 * We absolutely have to clear all the pending interrupt
+	 * bits in PORT_HOTPLUG_STAT. Otherwise the ISR port
+	 * interrupt bit won't have an edge, and the i965/g4x
+	 * edge triggered IIR will not notice that an interrupt
+	 * is still pending. We can't use PORT_HOTPLUG_EN to
+	 * guarantee the edge as the act of toggling the enable
+	 * bits can itself generate a new hotplug interrupt :(
+	 */
+	for (i = 0; i < 10; i++) {
+		u32 tmp = intel_uncore_read(&dev_priv->uncore, PORT_HOTPLUG_STAT) & hotplug_status_mask;
+
+		if (tmp == 0)
+			return hotplug_status;
+
+		hotplug_status |= tmp;
+		intel_uncore_write(&dev_priv->uncore, PORT_HOTPLUG_STAT, hotplug_status);
+	}
+
+	drm_WARN_ONCE(&dev_priv->drm, 1,
+		      "PORT_HOTPLUG_STAT did not clear (0x%08x)\n",
+		      intel_uncore_read(&dev_priv->uncore, PORT_HOTPLUG_STAT));
+
+	return hotplug_status;
+}
+
+static void i9xx_hpd_irq_handler(struct drm_i915_private *dev_priv,
+				 u32 hotplug_status)
+{
+	u32 pin_mask = 0, long_mask = 0;
+	u32 hotplug_trigger;
+
+	if (IS_G4X(dev_priv) ||
+	    IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_G4X;
+	else
+		hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_I915;
+
+	if (hotplug_trigger) {
+		intel_get_hpd_pins(dev_priv, &pin_mask, &long_mask,
+				   hotplug_trigger, hotplug_trigger,
+				   dev_priv->display.hotplug.hpd,
+				   i9xx_port_hotplug_long_detect);
+
+		intel_hpd_irq_handler(dev_priv, pin_mask, long_mask);
+	}
+
+	if ((IS_G4X(dev_priv) ||
+	     IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    hotplug_status & DP_AUX_CHANNEL_MASK_INT_STATUS_G4X)
+		dp_aux_irq_handler(dev_priv);
+}
+
+static irqreturn_t valleyview_irq_handler(int irq, void *arg)
+{
+	struct drm_i915_private *dev_priv = arg;
+	irqreturn_t ret = IRQ_NONE;
+
+	if (!intel_irqs_enabled(dev_priv))
+		return IRQ_NONE;
+
+	/* IRQs are synced during runtime_suspend, we don't require a wakeref */
+	disable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	do {
+		u32 iir, gt_iir, pm_iir;
+		u32 pipe_stats[I915_MAX_PIPES] = {};
+		u32 hotplug_status = 0;
+		u32 ier = 0;
+
+		gt_iir = intel_uncore_read(&dev_priv->uncore, GTIIR);
+		pm_iir = intel_uncore_read(&dev_priv->uncore, GEN6_PMIIR);
+		iir = intel_uncore_read(&dev_priv->uncore, VLV_IIR);
+
+		if (gt_iir == 0 && pm_iir == 0 && iir == 0)
+			break;
+
+		ret = IRQ_HANDLED;
+
+		/*
+		 * Theory on interrupt generation, based on empirical evidence:
+		 *
+		 * x = ((VLV_IIR & VLV_IER) ||
+		 *      (((GT_IIR & GT_IER) || (GEN6_PMIIR & GEN6_PMIER)) &&
+		 *       (VLV_MASTER_IER & MASTER_INTERRUPT_ENABLE)));
+		 *
+		 * A CPU interrupt will only be raised when 'x' has a 0->1 edge.
+		 * Hence we clear MASTER_INTERRUPT_ENABLE and VLV_IER to
+		 * guarantee the CPU interrupt will be raised again even if we
+		 * don't end up clearing all the VLV_IIR, GT_IIR, GEN6_PMIIR
+		 * bits this time around.
+		 */
+		intel_uncore_write(&dev_priv->uncore, VLV_MASTER_IER, 0);
+		ier = intel_uncore_read(&dev_priv->uncore, VLV_IER);
+		intel_uncore_write(&dev_priv->uncore, VLV_IER, 0);
+
+		if (gt_iir)
+			intel_uncore_write(&dev_priv->uncore, GTIIR, gt_iir);
+		if (pm_iir)
+			intel_uncore_write(&dev_priv->uncore, GEN6_PMIIR, pm_iir);
+
+		if (iir & I915_DISPLAY_PORT_INTERRUPT)
+			hotplug_status = i9xx_hpd_irq_ack(dev_priv);
+
+		/* Call regardless, as some status bits might not be
+		 * signalled in iir */
+		i9xx_pipestat_irq_ack(dev_priv, iir, pipe_stats);
+
+		if (iir & (I915_LPE_PIPE_A_INTERRUPT |
+			   I915_LPE_PIPE_B_INTERRUPT))
+			intel_lpe_audio_irq_handler(dev_priv);
+
+		/*
+		 * VLV_IIR is single buffered, and reflects the level
+		 * from PIPESTAT/PORT_HOTPLUG_STAT, hence clear it last.
+		 */
+		if (iir)
+			intel_uncore_write(&dev_priv->uncore, VLV_IIR, iir);
+
+		intel_uncore_write(&dev_priv->uncore, VLV_IER, ier);
+		intel_uncore_write(&dev_priv->uncore, VLV_MASTER_IER, MASTER_INTERRUPT_ENABLE);
+
+		if (gt_iir)
+			gen6_gt_irq_handler(to_gt(dev_priv), gt_iir);
+		if (pm_iir)
+			gen6_rps_irq_handler(&to_gt(dev_priv)->rps, pm_iir);
+
+		if (hotplug_status)
+			i9xx_hpd_irq_handler(dev_priv, hotplug_status);
+
+		valleyview_pipestat_irq_handler(dev_priv, pipe_stats);
+	} while (0);
+
+	pmu_irq_stats(dev_priv, ret);
+
+	enable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	return ret;
+}
+
+static irqreturn_t cherryview_irq_handler(int irq, void *arg)
+{
+	struct drm_i915_private *dev_priv = arg;
+	irqreturn_t ret = IRQ_NONE;
+
+	if (!intel_irqs_enabled(dev_priv))
+		return IRQ_NONE;
+
+	/* IRQs are synced during runtime_suspend, we don't require a wakeref */
+	disable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	do {
+		u32 master_ctl, iir;
+		u32 pipe_stats[I915_MAX_PIPES] = {};
+		u32 hotplug_status = 0;
+		u32 ier = 0;
+
+		master_ctl = intel_uncore_read(&dev_priv->uncore, GEN8_MASTER_IRQ) & ~GEN8_MASTER_IRQ_CONTROL;
+		iir = intel_uncore_read(&dev_priv->uncore, VLV_IIR);
+
+		if (master_ctl == 0 && iir == 0)
+			break;
+
+		ret = IRQ_HANDLED;
+
+		/*
+		 * Theory on interrupt generation, based on empirical evidence:
+		 *
+		 * x = ((VLV_IIR & VLV_IER) ||
+		 *      ((GEN8_MASTER_IRQ & ~GEN8_MASTER_IRQ_CONTROL) &&
+		 *       (GEN8_MASTER_IRQ & GEN8_MASTER_IRQ_CONTROL)));
+		 *
+		 * A CPU interrupt will only be raised when 'x' has a 0->1 edge.
+		 * Hence we clear GEN8_MASTER_IRQ_CONTROL and VLV_IER to
+		 * guarantee the CPU interrupt will be raised again even if we
+		 * don't end up clearing all the VLV_IIR and GEN8_MASTER_IRQ_CONTROL
+		 * bits this time around.
+		 */
+		intel_uncore_write(&dev_priv->uncore, GEN8_MASTER_IRQ, 0);
+		ier = intel_uncore_read(&dev_priv->uncore, VLV_IER);
+		intel_uncore_write(&dev_priv->uncore, VLV_IER, 0);
+
+		gen8_gt_irq_handler(to_gt(dev_priv), master_ctl);
+
+		if (iir & I915_DISPLAY_PORT_INTERRUPT)
+			hotplug_status = i9xx_hpd_irq_ack(dev_priv);
+
+		/* Call regardless, as some status bits might not be
+		 * signalled in iir */
+		i9xx_pipestat_irq_ack(dev_priv, iir, pipe_stats);
+
+		if (iir & (I915_LPE_PIPE_A_INTERRUPT |
+			   I915_LPE_PIPE_B_INTERRUPT |
+			   I915_LPE_PIPE_C_INTERRUPT))
+			intel_lpe_audio_irq_handler(dev_priv);
+
+		/*
+		 * VLV_IIR is single buffered, and reflects the level
+		 * from PIPESTAT/PORT_HOTPLUG_STAT, hence clear it last.
+		 */
+		if (iir)
+			intel_uncore_write(&dev_priv->uncore, VLV_IIR, iir);
+
+		intel_uncore_write(&dev_priv->uncore, VLV_IER, ier);
+		intel_uncore_write(&dev_priv->uncore, GEN8_MASTER_IRQ, GEN8_MASTER_IRQ_CONTROL);
+
+		if (hotplug_status)
+			i9xx_hpd_irq_handler(dev_priv, hotplug_status);
+
+		valleyview_pipestat_irq_handler(dev_priv, pipe_stats);
+	} while (0);
+
+	pmu_irq_stats(dev_priv, ret);
+
+	enable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	return ret;
+}
+
+static void ibx_hpd_irq_handler(struct drm_i915_private *dev_priv,
+				u32 hotplug_trigger)
+{
+	u32 dig_hotplug_reg, pin_mask = 0, long_mask = 0;
+
+	/*
+	 * Somehow the PCH doesn't seem to really ack the interrupt to the CPU
+	 * unless we touch the hotplug register, even if hotplug_trigger is
+	 * zero. Not acking leads to "The master control interrupt lied (SDE)!"
+	 * errors.
+	 */
+	dig_hotplug_reg = intel_uncore_read(&dev_priv->uncore, PCH_PORT_HOTPLUG);
+	if (!hotplug_trigger) {
+		u32 mask = PORTA_HOTPLUG_STATUS_MASK |
+			PORTD_HOTPLUG_STATUS_MASK |
+			PORTC_HOTPLUG_STATUS_MASK |
+			PORTB_HOTPLUG_STATUS_MASK;
+		dig_hotplug_reg &= ~mask;
+	}
+
+	intel_uncore_write(&dev_priv->uncore, PCH_PORT_HOTPLUG, dig_hotplug_reg);
+	if (!hotplug_trigger)
+		return;
+
+	intel_get_hpd_pins(dev_priv, &pin_mask, &long_mask,
+			   hotplug_trigger, dig_hotplug_reg,
+			   dev_priv->display.hotplug.pch_hpd,
+			   pch_port_hotplug_long_detect);
+
+	intel_hpd_irq_handler(dev_priv, pin_mask, long_mask);
+}
+
+static void ibx_irq_handler(struct drm_i915_private *dev_priv, u32 pch_iir)
+{
+	enum pipe pipe;
+	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK;
+
+	ibx_hpd_irq_handler(dev_priv, hotplug_trigger);
+
+	if (pch_iir & SDE_AUDIO_POWER_MASK) {
+		int port = ffs((pch_iir & SDE_AUDIO_POWER_MASK) >>
+			       SDE_AUDIO_POWER_SHIFT);
+		drm_dbg(&dev_priv->drm, "PCH audio power change on port %d\n",
+			port_name(port));
+	}
+
+	if (pch_iir & SDE_AUX_MASK)
+		dp_aux_irq_handler(dev_priv);
+
+	if (pch_iir & SDE_GMBUS)
+		gmbus_irq_handler(dev_priv);
+
+	if (pch_iir & SDE_AUDIO_HDCP_MASK)
+		drm_dbg(&dev_priv->drm, "PCH HDCP audio interrupt\n");
+
+	if (pch_iir & SDE_AUDIO_TRANS_MASK)
+		drm_dbg(&dev_priv->drm, "PCH transcoder audio interrupt\n");
+
+	if (pch_iir & SDE_POISON)
+		drm_err(&dev_priv->drm, "PCH poison interrupt\n");
+
+	if (pch_iir & SDE_FDI_MASK) {
+		for_each_pipe(dev_priv, pipe)
+			drm_dbg(&dev_priv->drm, "  pipe %c FDI IIR: 0x%08x\n",
+				pipe_name(pipe),
+				intel_uncore_read(&dev_priv->uncore, FDI_RX_IIR(pipe)));
+	}
+
+	if (pch_iir & (SDE_TRANSB_CRC_DONE | SDE_TRANSA_CRC_DONE))
+		drm_dbg(&dev_priv->drm, "PCH transcoder CRC done interrupt\n");
+
+	if (pch_iir & (SDE_TRANSB_CRC_ERR | SDE_TRANSA_CRC_ERR))
+		drm_dbg(&dev_priv->drm,
+			"PCH transcoder CRC error interrupt\n");
+
+	if (pch_iir & SDE_TRANSA_FIFO_UNDER)
+		intel_pch_fifo_underrun_irq_handler(dev_priv, PIPE_A);
+
+	if (pch_iir & SDE_TRANSB_FIFO_UNDER)
+		intel_pch_fifo_underrun_irq_handler(dev_priv, PIPE_B);
+}
+
+static void ivb_err_int_handler(struct drm_i915_private *dev_priv)
+{
+	u32 err_int = intel_uncore_read(&dev_priv->uncore, GEN7_ERR_INT);
+	enum pipe pipe;
+
+	if (err_int & ERR_INT_POISON)
+		drm_err(&dev_priv->drm, "Poison interrupt\n");
+
+	for_each_pipe(dev_priv, pipe) {
+		if (err_int & ERR_INT_FIFO_UNDERRUN(pipe))
+			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
+
+		if (err_int & ERR_INT_PIPE_CRC_DONE(pipe)) {
+			if (IS_IVYBRIDGE(dev_priv))
+				ivb_pipe_crc_irq_handler(dev_priv, pipe);
+			else
+				hsw_pipe_crc_irq_handler(dev_priv, pipe);
+		}
+	}
+
+	intel_uncore_write(&dev_priv->uncore, GEN7_ERR_INT, err_int);
+}
+
+static void cpt_serr_int_handler(struct drm_i915_private *dev_priv)
+{
+	u32 serr_int = intel_uncore_read(&dev_priv->uncore, SERR_INT);
+	enum pipe pipe;
+
+	if (serr_int & SERR_INT_POISON)
+		drm_err(&dev_priv->drm, "PCH poison interrupt\n");
+
+	for_each_pipe(dev_priv, pipe)
+		if (serr_int & SERR_INT_TRANS_FIFO_UNDERRUN(pipe))
+			intel_pch_fifo_underrun_irq_handler(dev_priv, pipe);
+
+	intel_uncore_write(&dev_priv->uncore, SERR_INT, serr_int);
+}
+
+static void cpt_irq_handler(struct drm_i915_private *dev_priv, u32 pch_iir)
+{
+	enum pipe pipe;
+	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK_CPT;
+
+	ibx_hpd_irq_handler(dev_priv, hotplug_trigger);
+
+	if (pch_iir & SDE_AUDIO_POWER_MASK_CPT) {
+		int port = ffs((pch_iir & SDE_AUDIO_POWER_MASK_CPT) >>
+			       SDE_AUDIO_POWER_SHIFT_CPT);
+		drm_dbg(&dev_priv->drm, "PCH audio power change on port %c\n",
+			port_name(port));
+	}
+
+	if (pch_iir & SDE_AUX_MASK_CPT)
+		dp_aux_irq_handler(dev_priv);
+
+	if (pch_iir & SDE_GMBUS_CPT)
+		gmbus_irq_handler(dev_priv);
+
+	if (pch_iir & SDE_AUDIO_CP_REQ_CPT)
+		drm_dbg(&dev_priv->drm, "Audio CP request interrupt\n");
+
+	if (pch_iir & SDE_AUDIO_CP_CHG_CPT)
+		drm_dbg(&dev_priv->drm, "Audio CP change interrupt\n");
+
+	if (pch_iir & SDE_FDI_MASK_CPT) {
+		for_each_pipe(dev_priv, pipe)
+			drm_dbg(&dev_priv->drm, "  pipe %c FDI IIR: 0x%08x\n",
+				pipe_name(pipe),
+				intel_uncore_read(&dev_priv->uncore, FDI_RX_IIR(pipe)));
+	}
+
+	if (pch_iir & SDE_ERROR_CPT)
+		cpt_serr_int_handler(dev_priv);
+}
+
+static void icp_irq_handler(struct drm_i915_private *dev_priv, u32 pch_iir)
+{
+	u32 ddi_hotplug_trigger = pch_iir & SDE_DDI_HOTPLUG_MASK_ICP;
+	u32 tc_hotplug_trigger = pch_iir & SDE_TC_HOTPLUG_MASK_ICP;
+	u32 pin_mask = 0, long_mask = 0;
+
+	if (ddi_hotplug_trigger) {
+		u32 dig_hotplug_reg;
+
+		/* Locking due to DSI native GPIO sequences */
+		spin_lock(&dev_priv->irq_lock);
+		dig_hotplug_reg = intel_uncore_read(&dev_priv->uncore, SHOTPLUG_CTL_DDI);
+		intel_uncore_write(&dev_priv->uncore, SHOTPLUG_CTL_DDI, dig_hotplug_reg);
+		spin_unlock(&dev_priv->irq_lock);
+
+		intel_get_hpd_pins(dev_priv, &pin_mask, &long_mask,
+				   ddi_hotplug_trigger, dig_hotplug_reg,
+				   dev_priv->display.hotplug.pch_hpd,
+				   icp_ddi_port_hotplug_long_detect);
+	}
+
+	if (tc_hotplug_trigger) {
+		u32 dig_hotplug_reg;
+
+		dig_hotplug_reg = intel_uncore_read(&dev_priv->uncore, SHOTPLUG_CTL_TC);
+		intel_uncore_write(&dev_priv->uncore, SHOTPLUG_CTL_TC, dig_hotplug_reg);
+
+		intel_get_hpd_pins(dev_priv, &pin_mask, &long_mask,
+				   tc_hotplug_trigger, dig_hotplug_reg,
+				   dev_priv->display.hotplug.pch_hpd,
+				   icp_tc_port_hotplug_long_detect);
+	}
+
+	if (pin_mask)
+		intel_hpd_irq_handler(dev_priv, pin_mask, long_mask);
+
+	if (pch_iir & SDE_GMBUS_ICP)
+		gmbus_irq_handler(dev_priv);
+}
+
+static void spt_irq_handler(struct drm_i915_private *dev_priv, u32 pch_iir)
+{
+	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK_SPT &
+		~SDE_PORTE_HOTPLUG_SPT;
+	u32 hotplug2_trigger = pch_iir & SDE_PORTE_HOTPLUG_SPT;
+	u32 pin_mask = 0, long_mask = 0;
+
+	if (hotplug_trigger) {
+		u32 dig_hotplug_reg;
+
+		dig_hotplug_reg = intel_uncore_read(&dev_priv->uncore, PCH_PORT_HOTPLUG);
+		intel_uncore_write(&dev_priv->uncore, PCH_PORT_HOTPLUG, dig_hotplug_reg);
+
+		intel_get_hpd_pins(dev_priv, &pin_mask, &long_mask,
+				   hotplug_trigger, dig_hotplug_reg,
+				   dev_priv->display.hotplug.pch_hpd,
+				   spt_port_hotplug_long_detect);
+	}
+
+	if (hotplug2_trigger) {
+		u32 dig_hotplug_reg;
+
+		dig_hotplug_reg = intel_uncore_read(&dev_priv->uncore, PCH_PORT_HOTPLUG2);
+		intel_uncore_write(&dev_priv->uncore, PCH_PORT_HOTPLUG2, dig_hotplug_reg);
+
+		intel_get_hpd_pins(dev_priv, &pin_mask, &long_mask,
+				   hotplug2_trigger, dig_hotplug_reg,
+				   dev_priv->display.hotplug.pch_hpd,
+				   spt_port_hotplug2_long_detect);
+	}
+
+	if (pin_mask)
+		intel_hpd_irq_handler(dev_priv, pin_mask, long_mask);
+
+	if (pch_iir & SDE_GMBUS_CPT)
+		gmbus_irq_handler(dev_priv);
+}
+
+static void ilk_hpd_irq_handler(struct drm_i915_private *dev_priv,
+				u32 hotplug_trigger)
+{
+	u32 dig_hotplug_reg, pin_mask = 0, long_mask = 0;
+
+	dig_hotplug_reg = intel_uncore_read(&dev_priv->uncore, DIGITAL_PORT_HOTPLUG_CNTRL);
+	intel_uncore_write(&dev_priv->uncore, DIGITAL_PORT_HOTPLUG_CNTRL, dig_hotplug_reg);
+
+	intel_get_hpd_pins(dev_priv, &pin_mask, &long_mask,
+			   hotplug_trigger, dig_hotplug_reg,
+			   dev_priv->display.hotplug.hpd,
+			   ilk_port_hotplug_long_detect);
+
+	intel_hpd_irq_handler(dev_priv, pin_mask, long_mask);
+}
+
+static void ilk_display_irq_handler(struct drm_i915_private *dev_priv,
+				    u32 de_iir)
+{
+	enum pipe pipe;
+	u32 hotplug_trigger = de_iir & DE_DP_A_HOTPLUG;
+
+	if (hotplug_trigger)
+		ilk_hpd_irq_handler(dev_priv, hotplug_trigger);
+
+	if (de_iir & DE_AUX_CHANNEL_A)
+		dp_aux_irq_handler(dev_priv);
+
+	if (de_iir & DE_GSE)
+		intel_opregion_asle_intr(dev_priv);
+
+	if (de_iir & DE_POISON)
+		drm_err(&dev_priv->drm, "Poison interrupt\n");
+
+	for_each_pipe(dev_priv, pipe) {
+		if (de_iir & DE_PIPE_VBLANK(pipe))
+			intel_handle_vblank(dev_priv, pipe);
+
+		if (de_iir & DE_PLANE_FLIP_DONE(pipe))
+			flip_done_handler(dev_priv, pipe);
+
+		if (de_iir & DE_PIPE_FIFO_UNDERRUN(pipe))
+			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
+
+		if (de_iir & DE_PIPE_CRC_DONE(pipe))
+			i9xx_pipe_crc_irq_handler(dev_priv, pipe);
+	}
+
+	/* check event from PCH */
+	if (de_iir & DE_PCH_EVENT) {
+		u32 pch_iir = intel_uncore_read(&dev_priv->uncore, SDEIIR);
+
+		if (HAS_PCH_CPT(dev_priv))
+			cpt_irq_handler(dev_priv, pch_iir);
+		else
+			ibx_irq_handler(dev_priv, pch_iir);
+
+		/* should clear PCH hotplug event before clear CPU irq */
+		intel_uncore_write(&dev_priv->uncore, SDEIIR, pch_iir);
+	}
+
+	if (DISPLAY_VER(dev_priv) == 5 && de_iir & DE_PCU_EVENT)
+		gen5_rps_irq_handler(&to_gt(dev_priv)->rps);
+}
+
+static void ivb_display_irq_handler(struct drm_i915_private *dev_priv,
+				    u32 de_iir)
+{
+	enum pipe pipe;
+	u32 hotplug_trigger = de_iir & DE_DP_A_HOTPLUG_IVB;
+
+	if (hotplug_trigger)
+		ilk_hpd_irq_handler(dev_priv, hotplug_trigger);
+
+	if (de_iir & DE_ERR_INT_IVB)
+		ivb_err_int_handler(dev_priv);
+
+	if (de_iir & DE_AUX_CHANNEL_A_IVB)
+		dp_aux_irq_handler(dev_priv);
+
+	if (de_iir & DE_GSE_IVB)
+		intel_opregion_asle_intr(dev_priv);
+
+	for_each_pipe(dev_priv, pipe) {
+		if (de_iir & DE_PIPE_VBLANK_IVB(pipe))
+			intel_handle_vblank(dev_priv, pipe);
+
+		if (de_iir & DE_PLANE_FLIP_DONE_IVB(pipe))
+			flip_done_handler(dev_priv, pipe);
+	}
+
+	/* check event from PCH */
+	if (!HAS_PCH_NOP(dev_priv) && (de_iir & DE_PCH_EVENT_IVB)) {
+		u32 pch_iir = intel_uncore_read(&dev_priv->uncore, SDEIIR);
+
+		cpt_irq_handler(dev_priv, pch_iir);
+
+		/* clear PCH hotplug event before clear CPU irq */
+		intel_uncore_write(&dev_priv->uncore, SDEIIR, pch_iir);
+	}
+}
+
+/*
+ * To handle irqs with the minimum potential races with fresh interrupts, we:
+ * 1 - Disable Master Interrupt Control.
+ * 2 - Find the source(s) of the interrupt.
+ * 3 - Clear the Interrupt Identity bits (IIR).
+ * 4 - Process the interrupt(s) that had bits set in the IIRs.
+ * 5 - Re-enable Master Interrupt Control.
+ */
+static irqreturn_t ilk_irq_handler(int irq, void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	void __iomem * const regs = i915->uncore.regs;
+	u32 de_iir, gt_iir, de_ier, sde_ier = 0;
+	irqreturn_t ret = IRQ_NONE;
+
+	if (unlikely(!intel_irqs_enabled(i915)))
+		return IRQ_NONE;
+
+	/* IRQs are synced during runtime_suspend, we don't require a wakeref */
+	disable_rpm_wakeref_asserts(&i915->runtime_pm);
+
+	/* disable master interrupt before clearing iir  */
+	de_ier = raw_reg_read(regs, DEIER);
+	raw_reg_write(regs, DEIER, de_ier & ~DE_MASTER_IRQ_CONTROL);
+
+	/* Disable south interrupts. We'll only write to SDEIIR once, so further
+	 * interrupts will will be stored on its back queue, and then we'll be
+	 * able to process them after we restore SDEIER (as soon as we restore
+	 * it, we'll get an interrupt if SDEIIR still has something to process
+	 * due to its back queue). */
+	if (!HAS_PCH_NOP(i915)) {
+		sde_ier = raw_reg_read(regs, SDEIER);
+		raw_reg_write(regs, SDEIER, 0);
+	}
+
+	/* Find, clear, then process each source of interrupt */
+
+	gt_iir = raw_reg_read(regs, GTIIR);
+	if (gt_iir) {
+		raw_reg_write(regs, GTIIR, gt_iir);
+		if (GRAPHICS_VER(i915) >= 6)
+			gen6_gt_irq_handler(to_gt(i915), gt_iir);
+		else
+			gen5_gt_irq_handler(to_gt(i915), gt_iir);
+		ret = IRQ_HANDLED;
+	}
+
+	de_iir = raw_reg_read(regs, DEIIR);
+	if (de_iir) {
+		raw_reg_write(regs, DEIIR, de_iir);
+		if (DISPLAY_VER(i915) >= 7)
+			ivb_display_irq_handler(i915, de_iir);
+		else
+			ilk_display_irq_handler(i915, de_iir);
+		ret = IRQ_HANDLED;
+	}
+
+	if (GRAPHICS_VER(i915) >= 6) {
+		u32 pm_iir = raw_reg_read(regs, GEN6_PMIIR);
+		if (pm_iir) {
+			raw_reg_write(regs, GEN6_PMIIR, pm_iir);
+			gen6_rps_irq_handler(&to_gt(i915)->rps, pm_iir);
+			ret = IRQ_HANDLED;
+		}
+	}
+
+	raw_reg_write(regs, DEIER, de_ier);
+	if (sde_ier)
+		raw_reg_write(regs, SDEIER, sde_ier);
+
+	pmu_irq_stats(i915, ret);
+
+	/* IRQs are synced during runtime_suspend, we don't require a wakeref */
+	enable_rpm_wakeref_asserts(&i915->runtime_pm);
+
+	return ret;
+}
+
+static void bxt_hpd_irq_handler(struct drm_i915_private *dev_priv,
+				u32 hotplug_trigger)
+{
+	u32 dig_hotplug_reg, pin_mask = 0, long_mask = 0;
+
+	dig_hotplug_reg = intel_uncore_read(&dev_priv->uncore, PCH_PORT_HOTPLUG);
+	intel_uncore_write(&dev_priv->uncore, PCH_PORT_HOTPLUG, dig_hotplug_reg);
+
+	intel_get_hpd_pins(dev_priv, &pin_mask, &long_mask,
+			   hotplug_trigger, dig_hotplug_reg,
+			   dev_priv->display.hotplug.hpd,
+			   bxt_port_hotplug_long_detect);
+
+	intel_hpd_irq_handler(dev_priv, pin_mask, long_mask);
+}
+
+static void gen11_hpd_irq_handler(struct drm_i915_private *dev_priv, u32 iir)
+{
+	u32 pin_mask = 0, long_mask = 0;
+	u32 trigger_tc = iir & GEN11_DE_TC_HOTPLUG_MASK;
+	u32 trigger_tbt = iir & GEN11_DE_TBT_HOTPLUG_MASK;
+
+	if (trigger_tc) {
+		u32 dig_hotplug_reg;
+
+		dig_hotplug_reg = intel_uncore_read(&dev_priv->uncore, GEN11_TC_HOTPLUG_CTL);
+		intel_uncore_write(&dev_priv->uncore, GEN11_TC_HOTPLUG_CTL, dig_hotplug_reg);
+
+		intel_get_hpd_pins(dev_priv, &pin_mask, &long_mask,
+				   trigger_tc, dig_hotplug_reg,
+				   dev_priv->display.hotplug.hpd,
+				   gen11_port_hotplug_long_detect);
+	}
+
+	if (trigger_tbt) {
+		u32 dig_hotplug_reg;
+
+		dig_hotplug_reg = intel_uncore_read(&dev_priv->uncore, GEN11_TBT_HOTPLUG_CTL);
+		intel_uncore_write(&dev_priv->uncore, GEN11_TBT_HOTPLUG_CTL, dig_hotplug_reg);
+
+		intel_get_hpd_pins(dev_priv, &pin_mask, &long_mask,
+				   trigger_tbt, dig_hotplug_reg,
+				   dev_priv->display.hotplug.hpd,
+				   gen11_port_hotplug_long_detect);
+	}
+
+	if (pin_mask)
+		intel_hpd_irq_handler(dev_priv, pin_mask, long_mask);
+	else
+		drm_err(&dev_priv->drm,
+			"Unexpected DE HPD interrupt 0x%08x\n", iir);
+}
+
+static u32 gen8_de_port_aux_mask(struct drm_i915_private *dev_priv)
+{
+	u32 mask;
+
+	if (DISPLAY_VER(dev_priv) >= 13)
+		return TGL_DE_PORT_AUX_DDIA |
+			TGL_DE_PORT_AUX_DDIB |
+			TGL_DE_PORT_AUX_DDIC |
+			XELPD_DE_PORT_AUX_DDID |
+			XELPD_DE_PORT_AUX_DDIE |
+			TGL_DE_PORT_AUX_USBC1 |
+			TGL_DE_PORT_AUX_USBC2 |
+			TGL_DE_PORT_AUX_USBC3 |
+			TGL_DE_PORT_AUX_USBC4;
+	else if (DISPLAY_VER(dev_priv) >= 12)
+		return TGL_DE_PORT_AUX_DDIA |
+			TGL_DE_PORT_AUX_DDIB |
+			TGL_DE_PORT_AUX_DDIC |
+			TGL_DE_PORT_AUX_USBC1 |
+			TGL_DE_PORT_AUX_USBC2 |
+			TGL_DE_PORT_AUX_USBC3 |
+			TGL_DE_PORT_AUX_USBC4 |
+			TGL_DE_PORT_AUX_USBC5 |
+			TGL_DE_PORT_AUX_USBC6;
+
+
+	mask = GEN8_AUX_CHANNEL_A;
+	if (DISPLAY_VER(dev_priv) >= 9)
+		mask |= GEN9_AUX_CHANNEL_B |
+			GEN9_AUX_CHANNEL_C |
+			GEN9_AUX_CHANNEL_D;
+
+	if (DISPLAY_VER(dev_priv) == 11) {
+		mask |= ICL_AUX_CHANNEL_F;
+		mask |= ICL_AUX_CHANNEL_E;
+	}
+
+	return mask;
+}
+
+static u32 gen8_de_pipe_fault_mask(struct drm_i915_private *dev_priv)
+{
+	if (DISPLAY_VER(dev_priv) >= 13 || HAS_D12_PLANE_MINIMIZATION(dev_priv))
+		return RKL_DE_PIPE_IRQ_FAULT_ERRORS;
+	else if (DISPLAY_VER(dev_priv) >= 11)
+		return GEN11_DE_PIPE_IRQ_FAULT_ERRORS;
+	else if (DISPLAY_VER(dev_priv) >= 9)
+		return GEN9_DE_PIPE_IRQ_FAULT_ERRORS;
+	else
+		return GEN8_DE_PIPE_IRQ_FAULT_ERRORS;
+}
+
+static void
+gen8_de_misc_irq_handler(struct drm_i915_private *dev_priv, u32 iir)
+{
+	bool found = false;
+
+	if (iir & GEN8_DE_MISC_GSE) {
+		intel_opregion_asle_intr(dev_priv);
+		found = true;
+	}
+
+	if (iir & GEN8_DE_EDP_PSR) {
+		struct intel_encoder *encoder;
+		u32 psr_iir;
+		i915_reg_t iir_reg;
+
+		for_each_intel_encoder_with_psr(&dev_priv->drm, encoder) {
+			struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
+
+			if (DISPLAY_VER(dev_priv) >= 12)
+				iir_reg = TRANS_PSR_IIR(intel_dp->psr.transcoder);
+			else
+				iir_reg = EDP_PSR_IIR;
+
+			psr_iir = intel_uncore_read(&dev_priv->uncore, iir_reg);
+			intel_uncore_write(&dev_priv->uncore, iir_reg, psr_iir);
+
+			if (psr_iir)
+				found = true;
+
+			intel_psr_irq_handler(intel_dp, psr_iir);
+
+			/* prior GEN12 only have one EDP PSR */
+			if (DISPLAY_VER(dev_priv) < 12)
+				break;
+		}
+	}
+
+	if (!found)
+		drm_err(&dev_priv->drm, "Unexpected DE Misc interrupt\n");
+}
+
+static void gen11_dsi_te_interrupt_handler(struct drm_i915_private *dev_priv,
+					   u32 te_trigger)
+{
+	enum pipe pipe = INVALID_PIPE;
+	enum transcoder dsi_trans;
+	enum port port;
+	u32 val, tmp;
+
+	/*
+	 * Incase of dual link, TE comes from DSI_1
+	 * this is to check if dual link is enabled
+	 */
+	val = intel_uncore_read(&dev_priv->uncore, TRANS_DDI_FUNC_CTL2(TRANSCODER_DSI_0));
+	val &= PORT_SYNC_MODE_ENABLE;
+
+	/*
+	 * if dual link is enabled, then read DSI_0
+	 * transcoder registers
+	 */
+	port = ((te_trigger & DSI1_TE && val) || (te_trigger & DSI0_TE)) ?
+						  PORT_A : PORT_B;
+	dsi_trans = (port == PORT_A) ? TRANSCODER_DSI_0 : TRANSCODER_DSI_1;
+
+	/* Check if DSI configured in command mode */
+	val = intel_uncore_read(&dev_priv->uncore, DSI_TRANS_FUNC_CONF(dsi_trans));
+	val = val & OP_MODE_MASK;
+
+	if (val != CMD_MODE_NO_GATE && val != CMD_MODE_TE_GATE) {
+		drm_err(&dev_priv->drm, "DSI trancoder not configured in command mode\n");
+		return;
+	}
+
+	/* Get PIPE for handling VBLANK event */
+	val = intel_uncore_read(&dev_priv->uncore, TRANS_DDI_FUNC_CTL(dsi_trans));
+	switch (val & TRANS_DDI_EDP_INPUT_MASK) {
+	case TRANS_DDI_EDP_INPUT_A_ON:
+		pipe = PIPE_A;
+		break;
+	case TRANS_DDI_EDP_INPUT_B_ONOFF:
+		pipe = PIPE_B;
+		break;
+	case TRANS_DDI_EDP_INPUT_C_ONOFF:
+		pipe = PIPE_C;
+		break;
+	default:
+		drm_err(&dev_priv->drm, "Invalid PIPE\n");
+		return;
+	}
+
+	intel_handle_vblank(dev_priv, pipe);
+
+	/* clear TE in dsi IIR */
+	port = (te_trigger & DSI1_TE) ? PORT_B : PORT_A;
+	tmp = intel_uncore_read(&dev_priv->uncore, DSI_INTR_IDENT_REG(port));
+	intel_uncore_write(&dev_priv->uncore, DSI_INTR_IDENT_REG(port), tmp);
+}
+
+static u32 gen8_de_pipe_flip_done_mask(struct drm_i915_private *i915)
+{
+	if (DISPLAY_VER(i915) >= 9)
+		return GEN9_PIPE_PLANE1_FLIP_DONE;
+	else
+		return GEN8_PIPE_PRIMARY_FLIP_DONE;
+}
+
+u32 gen8_de_pipe_underrun_mask(struct drm_i915_private *dev_priv)
+{
+	u32 mask = GEN8_PIPE_FIFO_UNDERRUN;
+
+	if (DISPLAY_VER(dev_priv) >= 13)
+		mask |= XELPD_PIPE_SOFT_UNDERRUN |
+			XELPD_PIPE_HARD_UNDERRUN;
+
+	return mask;
+}
+
+static irqreturn_t
+gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl)
+{
+	irqreturn_t ret = IRQ_NONE;
+	u32 iir;
+	enum pipe pipe;
+
+	drm_WARN_ON_ONCE(&dev_priv->drm, !HAS_DISPLAY(dev_priv));
+
+	if (master_ctl & GEN8_DE_MISC_IRQ) {
+		iir = intel_uncore_read(&dev_priv->uncore, GEN8_DE_MISC_IIR);
+		if (iir) {
+			intel_uncore_write(&dev_priv->uncore, GEN8_DE_MISC_IIR, iir);
+			ret = IRQ_HANDLED;
+			gen8_de_misc_irq_handler(dev_priv, iir);
+		} else {
+			drm_err(&dev_priv->drm,
+				"The master control interrupt lied (DE MISC)!\n");
+		}
+	}
+
+	if (DISPLAY_VER(dev_priv) >= 11 && (master_ctl & GEN11_DE_HPD_IRQ)) {
+		iir = intel_uncore_read(&dev_priv->uncore, GEN11_DE_HPD_IIR);
+		if (iir) {
+			intel_uncore_write(&dev_priv->uncore, GEN11_DE_HPD_IIR, iir);
+			ret = IRQ_HANDLED;
+			gen11_hpd_irq_handler(dev_priv, iir);
+		} else {
+			drm_err(&dev_priv->drm,
+				"The master control interrupt lied, (DE HPD)!\n");
+		}
+	}
+
+	if (master_ctl & GEN8_DE_PORT_IRQ) {
+		iir = intel_uncore_read(&dev_priv->uncore, GEN8_DE_PORT_IIR);
+		if (iir) {
+			bool found = false;
+
+			intel_uncore_write(&dev_priv->uncore, GEN8_DE_PORT_IIR, iir);
+			ret = IRQ_HANDLED;
+
+			if (iir & gen8_de_port_aux_mask(dev_priv)) {
+				dp_aux_irq_handler(dev_priv);
+				found = true;
+			}
+
+			if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) {
+				u32 hotplug_trigger = iir & BXT_DE_PORT_HOTPLUG_MASK;
+
+				if (hotplug_trigger) {
+					bxt_hpd_irq_handler(dev_priv, hotplug_trigger);
+					found = true;
+				}
+			} else if (IS_BROADWELL(dev_priv)) {
+				u32 hotplug_trigger = iir & BDW_DE_PORT_HOTPLUG_MASK;
+
+				if (hotplug_trigger) {
+					ilk_hpd_irq_handler(dev_priv, hotplug_trigger);
+					found = true;
+				}
+			}
+
+			if ((IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv)) &&
+			    (iir & BXT_DE_PORT_GMBUS)) {
+				gmbus_irq_handler(dev_priv);
+				found = true;
+			}
+
+			if (DISPLAY_VER(dev_priv) >= 11) {
+				u32 te_trigger = iir & (DSI0_TE | DSI1_TE);
+
+				if (te_trigger) {
+					gen11_dsi_te_interrupt_handler(dev_priv, te_trigger);
+					found = true;
+				}
+			}
+
+			if (!found)
+				drm_err(&dev_priv->drm,
+					"Unexpected DE Port interrupt\n");
+		}
+		else
+			drm_err(&dev_priv->drm,
+				"The master control interrupt lied (DE PORT)!\n");
+	}
+
+	for_each_pipe(dev_priv, pipe) {
+		u32 fault_errors;
+
+		if (!(master_ctl & GEN8_DE_PIPE_IRQ(pipe)))
+			continue;
+
+		iir = intel_uncore_read(&dev_priv->uncore, GEN8_DE_PIPE_IIR(pipe));
+		if (!iir) {
+			drm_err(&dev_priv->drm,
+				"The master control interrupt lied (DE PIPE)!\n");
+			continue;
+		}
+
+		ret = IRQ_HANDLED;
+		intel_uncore_write(&dev_priv->uncore, GEN8_DE_PIPE_IIR(pipe), iir);
+
+		if (iir & GEN8_PIPE_VBLANK)
+			intel_handle_vblank(dev_priv, pipe);
+
+		if (iir & gen8_de_pipe_flip_done_mask(dev_priv))
+			flip_done_handler(dev_priv, pipe);
+
+		if (iir & GEN8_PIPE_CDCLK_CRC_DONE)
+			hsw_pipe_crc_irq_handler(dev_priv, pipe);
+
+		if (iir & gen8_de_pipe_underrun_mask(dev_priv))
+			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
+
+		fault_errors = iir & gen8_de_pipe_fault_mask(dev_priv);
+		if (fault_errors)
+			drm_err(&dev_priv->drm,
+				"Fault errors on pipe %c: 0x%08x\n",
+				pipe_name(pipe),
+				fault_errors);
+	}
+
+	if (HAS_PCH_SPLIT(dev_priv) && !HAS_PCH_NOP(dev_priv) &&
+	    master_ctl & GEN8_DE_PCH_IRQ) {
+		/*
+		 * FIXME(BDW): Assume for now that the new interrupt handling
+		 * scheme also closed the SDE interrupt handling race we've seen
+		 * on older pch-split platforms. But this needs testing.
+		 */
+		iir = intel_uncore_read(&dev_priv->uncore, SDEIIR);
+		if (iir) {
+			intel_uncore_write(&dev_priv->uncore, SDEIIR, iir);
+			ret = IRQ_HANDLED;
+
+			if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP)
+				icp_irq_handler(dev_priv, iir);
+			else if (INTEL_PCH_TYPE(dev_priv) >= PCH_SPT)
+				spt_irq_handler(dev_priv, iir);
+			else
+				cpt_irq_handler(dev_priv, iir);
+		} else {
+			/*
+			 * Like on previous PCH there seems to be something
+			 * fishy going on with forwarding PCH interrupts.
+			 */
+			drm_dbg(&dev_priv->drm,
+				"The master control interrupt lied (SDE)!\n");
+		}
+	}
+
+	return ret;
+}
+
+static inline u32 gen8_master_intr_disable(void __iomem * const regs)
+{
+	raw_reg_write(regs, GEN8_MASTER_IRQ, 0);
+
+	/*
+	 * Now with master disabled, get a sample of level indications
+	 * for this interrupt. Indications will be cleared on related acks.
+	 * New indications can and will light up during processing,
+	 * and will generate new interrupt after enabling master.
+	 */
+	return raw_reg_read(regs, GEN8_MASTER_IRQ);
+}
+
+static inline void gen8_master_intr_enable(void __iomem * const regs)
+{
+	raw_reg_write(regs, GEN8_MASTER_IRQ, GEN8_MASTER_IRQ_CONTROL);
+}
+
+static irqreturn_t gen8_irq_handler(int irq, void *arg)
+{
+	struct drm_i915_private *dev_priv = arg;
+	void __iomem * const regs = dev_priv->uncore.regs;
+	u32 master_ctl;
+
+	if (!intel_irqs_enabled(dev_priv))
+		return IRQ_NONE;
+
+	master_ctl = gen8_master_intr_disable(regs);
+	if (!master_ctl) {
+		gen8_master_intr_enable(regs);
+		return IRQ_NONE;
+	}
+
+	/* Find, queue (onto bottom-halves), then clear each source */
+	gen8_gt_irq_handler(to_gt(dev_priv), master_ctl);
+
+	/* IRQs are synced during runtime_suspend, we don't require a wakeref */
+	if (master_ctl & ~GEN8_GT_IRQS) {
+		disable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+		gen8_de_irq_handler(dev_priv, master_ctl);
+		enable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+	}
+
+	gen8_master_intr_enable(regs);
+
+	pmu_irq_stats(dev_priv, IRQ_HANDLED);
+
+	return IRQ_HANDLED;
+}
+
+static u32
+gen11_gu_misc_irq_ack(struct drm_i915_private *i915, const u32 master_ctl)
+{
+	void __iomem * const regs = i915->uncore.regs;
+	u32 iir;
+
+	if (!(master_ctl & GEN11_GU_MISC_IRQ))
+		return 0;
+
+	iir = raw_reg_read(regs, GEN11_GU_MISC_IIR);
+	if (likely(iir))
+		raw_reg_write(regs, GEN11_GU_MISC_IIR, iir);
+
+	return iir;
+}
+
+static void
+gen11_gu_misc_irq_handler(struct drm_i915_private *i915, const u32 iir)
+{
+	if (iir & GEN11_GU_MISC_GSE)
+		intel_opregion_asle_intr(i915);
+}
+
+static inline u32 gen11_master_intr_disable(void __iomem * const regs)
+{
+	raw_reg_write(regs, GEN11_GFX_MSTR_IRQ, 0);
+
+	/*
+	 * Now with master disabled, get a sample of level indications
+	 * for this interrupt. Indications will be cleared on related acks.
+	 * New indications can and will light up during processing,
+	 * and will generate new interrupt after enabling master.
+	 */
+	return raw_reg_read(regs, GEN11_GFX_MSTR_IRQ);
+}
+
+static inline void gen11_master_intr_enable(void __iomem * const regs)
+{
+	raw_reg_write(regs, GEN11_GFX_MSTR_IRQ, GEN11_MASTER_IRQ);
+}
+
+static void
+gen11_display_irq_handler(struct drm_i915_private *i915)
+{
+	void __iomem * const regs = i915->uncore.regs;
+	const u32 disp_ctl = raw_reg_read(regs, GEN11_DISPLAY_INT_CTL);
+
+	disable_rpm_wakeref_asserts(&i915->runtime_pm);
+	/*
+	 * GEN11_DISPLAY_INT_CTL has same format as GEN8_MASTER_IRQ
+	 * for the display related bits.
+	 */
+	raw_reg_write(regs, GEN11_DISPLAY_INT_CTL, 0x0);
+	gen8_de_irq_handler(i915, disp_ctl);
+	raw_reg_write(regs, GEN11_DISPLAY_INT_CTL,
+		      GEN11_DISPLAY_IRQ_ENABLE);
+
+	enable_rpm_wakeref_asserts(&i915->runtime_pm);
+}
+
+static irqreturn_t gen11_irq_handler(int irq, void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	void __iomem * const regs = i915->uncore.regs;
+	struct intel_gt *gt = to_gt(i915);
+	u32 master_ctl;
+	u32 gu_misc_iir;
+
+	if (!intel_irqs_enabled(i915))
+		return IRQ_NONE;
+
+	master_ctl = gen11_master_intr_disable(regs);
+	if (!master_ctl) {
+		gen11_master_intr_enable(regs);
+		return IRQ_NONE;
+	}
+
+	/* Find, queue (onto bottom-halves), then clear each source */
+	gen11_gt_irq_handler(gt, master_ctl);
+
+	/* IRQs are synced during runtime_suspend, we don't require a wakeref */
+	if (master_ctl & GEN11_DISPLAY_IRQ)
+		gen11_display_irq_handler(i915);
+
+	gu_misc_iir = gen11_gu_misc_irq_ack(i915, master_ctl);
+
+	gen11_master_intr_enable(regs);
+
+	gen11_gu_misc_irq_handler(i915, gu_misc_iir);
+
+	pmu_irq_stats(i915, IRQ_HANDLED);
+
+	return IRQ_HANDLED;
+}
+
+static inline u32 dg1_master_intr_disable(void __iomem * const regs)
+{
+	u32 val;
+
+	/* First disable interrupts */
+	raw_reg_write(regs, DG1_MSTR_TILE_INTR, 0);
+
+	/* Get the indication levels and ack the master unit */
+	val = raw_reg_read(regs, DG1_MSTR_TILE_INTR);
+	if (unlikely(!val))
+		return 0;
+
+	raw_reg_write(regs, DG1_MSTR_TILE_INTR, val);
+
+	return val;
+}
+
+static inline void dg1_master_intr_enable(void __iomem * const regs)
+{
+	raw_reg_write(regs, DG1_MSTR_TILE_INTR, DG1_MSTR_IRQ);
+}
+
+static irqreturn_t dg1_irq_handler(int irq, void *arg)
+{
+	struct drm_i915_private * const i915 = arg;
+	struct intel_gt *gt = to_gt(i915);
+	void __iomem * const regs = gt->uncore->regs;
+	u32 master_tile_ctl, master_ctl;
+	u32 gu_misc_iir;
+
+	if (!intel_irqs_enabled(i915))
+		return IRQ_NONE;
+
+	master_tile_ctl = dg1_master_intr_disable(regs);
+	if (!master_tile_ctl) {
+		dg1_master_intr_enable(regs);
+		return IRQ_NONE;
+	}
+
+	/* FIXME: we only support tile 0 for now. */
+	if (master_tile_ctl & DG1_MSTR_TILE(0)) {
+		master_ctl = raw_reg_read(regs, GEN11_GFX_MSTR_IRQ);
+		raw_reg_write(regs, GEN11_GFX_MSTR_IRQ, master_ctl);
+	} else {
+		DRM_ERROR("Tile not supported: 0x%08x\n", master_tile_ctl);
+		dg1_master_intr_enable(regs);
+		return IRQ_NONE;
+	}
+
+	gen11_gt_irq_handler(gt, master_ctl);
+
+	if (master_ctl & GEN11_DISPLAY_IRQ)
+		gen11_display_irq_handler(i915);
+
+	gu_misc_iir = gen11_gu_misc_irq_ack(i915, master_ctl);
+
+	dg1_master_intr_enable(regs);
+
+	gen11_gu_misc_irq_handler(i915, gu_misc_iir);
+
+	pmu_irq_stats(i915, IRQ_HANDLED);
+
+	return IRQ_HANDLED;
+}
+
+/* Called from drm generic code, passed 'crtc' which
+ * we use as a pipe index
+ */
+int i8xx_enable_vblank(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	unsigned long irqflags;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	i915_enable_pipestat(dev_priv, pipe, PIPE_VBLANK_INTERRUPT_STATUS);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+
+	return 0;
+}
+
+int i915gm_enable_vblank(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+
+	/*
+	 * Vblank interrupts fail to wake the device up from C2+.
+	 * Disabling render clock gating during C-states avoids
+	 * the problem. There is a small power cost so we do this
+	 * only when vblank interrupts are actually enabled.
+	 */
+	if (dev_priv->vblank_enabled++ == 0)
+		intel_uncore_write(&dev_priv->uncore, SCPD0, _MASKED_BIT_ENABLE(CSTATE_RENDER_CLOCK_GATE_DISABLE));
+
+	return i8xx_enable_vblank(crtc);
+}
+
+int i965_enable_vblank(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	unsigned long irqflags;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	i915_enable_pipestat(dev_priv, pipe,
+			     PIPE_START_VBLANK_INTERRUPT_STATUS);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+
+	return 0;
+}
+
+int ilk_enable_vblank(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	unsigned long irqflags;
+	u32 bit = DISPLAY_VER(dev_priv) >= 7 ?
+		DE_PIPE_VBLANK_IVB(pipe) : DE_PIPE_VBLANK(pipe);
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	ilk_enable_display_irq(dev_priv, bit);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+
+	/* Even though there is no DMC, frame counter can get stuck when
+	 * PSR is active as no frames are generated.
+	 */
+	if (HAS_PSR(dev_priv))
+		drm_crtc_vblank_restore(crtc);
+
+	return 0;
+}
+
+static bool gen11_dsi_configure_te(struct intel_crtc *intel_crtc,
+				   bool enable)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_crtc->base.dev);
+	enum port port;
+	u32 tmp;
+
+	if (!(intel_crtc->mode_flags &
+	    (I915_MODE_FLAG_DSI_USE_TE1 | I915_MODE_FLAG_DSI_USE_TE0)))
+		return false;
+
+	/* for dual link cases we consider TE from slave */
+	if (intel_crtc->mode_flags & I915_MODE_FLAG_DSI_USE_TE1)
+		port = PORT_B;
+	else
+		port = PORT_A;
+
+	tmp =  intel_uncore_read(&dev_priv->uncore, DSI_INTR_MASK_REG(port));
+	if (enable)
+		tmp &= ~DSI_TE_EVENT;
+	else
+		tmp |= DSI_TE_EVENT;
+
+	intel_uncore_write(&dev_priv->uncore, DSI_INTR_MASK_REG(port), tmp);
+
+	tmp = intel_uncore_read(&dev_priv->uncore, DSI_INTR_IDENT_REG(port));
+	intel_uncore_write(&dev_priv->uncore, DSI_INTR_IDENT_REG(port), tmp);
+
+	return true;
+}
+
+int bdw_enable_vblank(struct drm_crtc *_crtc)
+{
+	struct intel_crtc *crtc = to_intel_crtc(_crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	unsigned long irqflags;
+
+	if (gen11_dsi_configure_te(crtc, true))
+		return 0;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	bdw_enable_pipe_irq(dev_priv, pipe, GEN8_PIPE_VBLANK);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+
+	/* Even if there is no DMC, frame counter can get stuck when
+	 * PSR is active as no frames are generated, so check only for PSR.
+	 */
+	if (HAS_PSR(dev_priv))
+		drm_crtc_vblank_restore(&crtc->base);
+
+	return 0;
+}
+
+/* Called from drm generic code, passed 'crtc' which
+ * we use as a pipe index
+ */
+void i8xx_disable_vblank(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	unsigned long irqflags;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	i915_disable_pipestat(dev_priv, pipe, PIPE_VBLANK_INTERRUPT_STATUS);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+}
+
+void i915gm_disable_vblank(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+
+	i8xx_disable_vblank(crtc);
+
+	if (--dev_priv->vblank_enabled == 0)
+		intel_uncore_write(&dev_priv->uncore, SCPD0, _MASKED_BIT_DISABLE(CSTATE_RENDER_CLOCK_GATE_DISABLE));
+}
+
+void i965_disable_vblank(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	unsigned long irqflags;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	i915_disable_pipestat(dev_priv, pipe,
+			      PIPE_START_VBLANK_INTERRUPT_STATUS);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+}
+
+void ilk_disable_vblank(struct drm_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	unsigned long irqflags;
+	u32 bit = DISPLAY_VER(dev_priv) >= 7 ?
+		DE_PIPE_VBLANK_IVB(pipe) : DE_PIPE_VBLANK(pipe);
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	ilk_disable_display_irq(dev_priv, bit);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+}
+
+void bdw_disable_vblank(struct drm_crtc *_crtc)
+{
+	struct intel_crtc *crtc = to_intel_crtc(_crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	unsigned long irqflags;
+
+	if (gen11_dsi_configure_te(crtc, false))
+		return;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	bdw_disable_pipe_irq(dev_priv, pipe, GEN8_PIPE_VBLANK);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+}
+
+static void ibx_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	if (HAS_PCH_NOP(dev_priv))
+		return;
+
+	GEN3_IRQ_RESET(uncore, SDE);
+
+	if (HAS_PCH_CPT(dev_priv) || HAS_PCH_LPT(dev_priv))
+		intel_uncore_write(&dev_priv->uncore, SERR_INT, 0xffffffff);
+}
+
+static void vlv_display_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	if (IS_CHERRYVIEW(dev_priv))
+		intel_uncore_write(uncore, DPINVGTT, DPINVGTT_STATUS_MASK_CHV);
+	else
+		intel_uncore_write(uncore, DPINVGTT, DPINVGTT_STATUS_MASK_VLV);
+
+	i915_hotplug_interrupt_update_locked(dev_priv, 0xffffffff, 0);
+	intel_uncore_write(uncore, PORT_HOTPLUG_STAT, intel_uncore_read(&dev_priv->uncore, PORT_HOTPLUG_STAT));
+
+	i9xx_pipestat_irq_reset(dev_priv);
+
+	GEN3_IRQ_RESET(uncore, VLV_);
+	dev_priv->irq_mask = ~0u;
+}
+
+static void vlv_display_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	u32 pipestat_mask;
+	u32 enable_mask;
+	enum pipe pipe;
+
+	pipestat_mask = PIPE_CRC_DONE_INTERRUPT_STATUS;
+
+	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS);
+	for_each_pipe(dev_priv, pipe)
+		i915_enable_pipestat(dev_priv, pipe, pipestat_mask);
+
+	enable_mask = I915_DISPLAY_PORT_INTERRUPT |
+		I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
+		I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
+		I915_LPE_PIPE_A_INTERRUPT |
+		I915_LPE_PIPE_B_INTERRUPT;
+
+	if (IS_CHERRYVIEW(dev_priv))
+		enable_mask |= I915_DISPLAY_PIPE_C_EVENT_INTERRUPT |
+			I915_LPE_PIPE_C_INTERRUPT;
+
+	drm_WARN_ON(&dev_priv->drm, dev_priv->irq_mask != ~0u);
+
+	dev_priv->irq_mask = ~enable_mask;
+
+	GEN3_IRQ_INIT(uncore, VLV_, dev_priv->irq_mask, enable_mask);
+}
+
+/* drm_dma.h hooks
+*/
+static void ilk_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	GEN3_IRQ_RESET(uncore, DE);
+	dev_priv->irq_mask = ~0u;
+
+	if (GRAPHICS_VER(dev_priv) == 7)
+		intel_uncore_write(uncore, GEN7_ERR_INT, 0xffffffff);
+
+	if (IS_HASWELL(dev_priv)) {
+		intel_uncore_write(uncore, EDP_PSR_IMR, 0xffffffff);
+		intel_uncore_write(uncore, EDP_PSR_IIR, 0xffffffff);
+	}
+
+	gen5_gt_irq_reset(to_gt(dev_priv));
+
+	ibx_irq_reset(dev_priv);
+}
+
+static void valleyview_irq_reset(struct drm_i915_private *dev_priv)
+{
+	intel_uncore_write(&dev_priv->uncore, VLV_MASTER_IER, 0);
+	intel_uncore_posting_read(&dev_priv->uncore, VLV_MASTER_IER);
+
+	gen5_gt_irq_reset(to_gt(dev_priv));
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	if (dev_priv->display_irqs_enabled)
+		vlv_display_irq_reset(dev_priv);
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
+
+static void gen8_display_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	enum pipe pipe;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	intel_uncore_write(uncore, EDP_PSR_IMR, 0xffffffff);
+	intel_uncore_write(uncore, EDP_PSR_IIR, 0xffffffff);
+
+	for_each_pipe(dev_priv, pipe)
+		if (intel_display_power_is_enabled(dev_priv,
+						   POWER_DOMAIN_PIPE(pipe)))
+			GEN8_IRQ_RESET_NDX(uncore, DE_PIPE, pipe);
+
+	GEN3_IRQ_RESET(uncore, GEN8_DE_PORT_);
+	GEN3_IRQ_RESET(uncore, GEN8_DE_MISC_);
+}
+
+static void gen8_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	gen8_master_intr_disable(dev_priv->uncore.regs);
+
+	gen8_gt_irq_reset(to_gt(dev_priv));
+	gen8_display_irq_reset(dev_priv);
+	GEN3_IRQ_RESET(uncore, GEN8_PCU_);
+
+	if (HAS_PCH_SPLIT(dev_priv))
+		ibx_irq_reset(dev_priv);
+
+}
+
+static void gen11_display_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	enum pipe pipe;
+	u32 trans_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) |
+		BIT(TRANSCODER_C) | BIT(TRANSCODER_D);
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	intel_uncore_write(uncore, GEN11_DISPLAY_INT_CTL, 0);
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		enum transcoder trans;
+
+		for_each_cpu_transcoder_masked(dev_priv, trans, trans_mask) {
+			enum intel_display_power_domain domain;
+
+			domain = POWER_DOMAIN_TRANSCODER(trans);
+			if (!intel_display_power_is_enabled(dev_priv, domain))
+				continue;
+
+			intel_uncore_write(uncore, TRANS_PSR_IMR(trans), 0xffffffff);
+			intel_uncore_write(uncore, TRANS_PSR_IIR(trans), 0xffffffff);
+		}
+	} else {
+		intel_uncore_write(uncore, EDP_PSR_IMR, 0xffffffff);
+		intel_uncore_write(uncore, EDP_PSR_IIR, 0xffffffff);
+	}
+
+	for_each_pipe(dev_priv, pipe)
+		if (intel_display_power_is_enabled(dev_priv,
+						   POWER_DOMAIN_PIPE(pipe)))
+			GEN8_IRQ_RESET_NDX(uncore, DE_PIPE, pipe);
+
+	GEN3_IRQ_RESET(uncore, GEN8_DE_PORT_);
+	GEN3_IRQ_RESET(uncore, GEN8_DE_MISC_);
+	GEN3_IRQ_RESET(uncore, GEN11_DE_HPD_);
+
+	if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP)
+		GEN3_IRQ_RESET(uncore, SDE);
+}
+
+static void gen11_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_gt *gt = to_gt(dev_priv);
+	struct intel_uncore *uncore = gt->uncore;
+
+	gen11_master_intr_disable(dev_priv->uncore.regs);
+
+	gen11_gt_irq_reset(gt);
+	gen11_display_irq_reset(dev_priv);
+
+	GEN3_IRQ_RESET(uncore, GEN11_GU_MISC_);
+	GEN3_IRQ_RESET(uncore, GEN8_PCU_);
+}
+
+static void dg1_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_gt *gt = to_gt(dev_priv);
+	struct intel_uncore *uncore = gt->uncore;
+
+	dg1_master_intr_disable(dev_priv->uncore.regs);
+
+	gen11_gt_irq_reset(gt);
+	gen11_display_irq_reset(dev_priv);
+
+	GEN3_IRQ_RESET(uncore, GEN11_GU_MISC_);
+	GEN3_IRQ_RESET(uncore, GEN8_PCU_);
+}
+
+void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv,
+				     u8 pipe_mask)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	u32 extra_ier = GEN8_PIPE_VBLANK |
+		gen8_de_pipe_underrun_mask(dev_priv) |
+		gen8_de_pipe_flip_done_mask(dev_priv);
+	enum pipe pipe;
+
+	spin_lock_irq(&dev_priv->irq_lock);
+
+	if (!intel_irqs_enabled(dev_priv)) {
+		spin_unlock_irq(&dev_priv->irq_lock);
+		return;
+	}
+
+	for_each_pipe_masked(dev_priv, pipe, pipe_mask)
+		GEN8_IRQ_INIT_NDX(uncore, DE_PIPE, pipe,
+				  dev_priv->de_irq_mask[pipe],
+				  ~dev_priv->de_irq_mask[pipe] | extra_ier);
+
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
+
+void gen8_irq_power_well_pre_disable(struct drm_i915_private *dev_priv,
+				     u8 pipe_mask)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	enum pipe pipe;
+
+	spin_lock_irq(&dev_priv->irq_lock);
+
+	if (!intel_irqs_enabled(dev_priv)) {
+		spin_unlock_irq(&dev_priv->irq_lock);
+		return;
+	}
+
+	for_each_pipe_masked(dev_priv, pipe, pipe_mask)
+		GEN8_IRQ_RESET_NDX(uncore, DE_PIPE, pipe);
+
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/* make sure we're done processing display irqs */
+	intel_synchronize_irq(dev_priv);
+}
+
+static void cherryview_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	intel_uncore_write(&dev_priv->uncore, GEN8_MASTER_IRQ, 0);
+	intel_uncore_posting_read(&dev_priv->uncore, GEN8_MASTER_IRQ);
+
+	gen8_gt_irq_reset(to_gt(dev_priv));
+
+	GEN3_IRQ_RESET(uncore, GEN8_PCU_);
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	if (dev_priv->display_irqs_enabled)
+		vlv_display_irq_reset(dev_priv);
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
+
+static u32 ibx_hotplug_enables(struct drm_i915_private *i915,
+			       enum hpd_pin pin)
+{
+	switch (pin) {
+	case HPD_PORT_A:
+		/*
+		 * When CPU and PCH are on the same package, port A
+		 * HPD must be enabled in both north and south.
+		 */
+		return HAS_PCH_LPT_LP(i915) ?
+			PORTA_HOTPLUG_ENABLE : 0;
+	case HPD_PORT_B:
+		return PORTB_HOTPLUG_ENABLE |
+			PORTB_PULSE_DURATION_2ms;
+	case HPD_PORT_C:
+		return PORTC_HOTPLUG_ENABLE |
+			PORTC_PULSE_DURATION_2ms;
+	case HPD_PORT_D:
+		return PORTD_HOTPLUG_ENABLE |
+			PORTD_PULSE_DURATION_2ms;
+	default:
+		return 0;
+	}
+}
+
+static void ibx_hpd_detection_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug;
+
+	/*
+	 * Enable digital hotplug on the PCH, and configure the DP short pulse
+	 * duration to 2ms (which is the minimum in the Display Port spec).
+	 * The pulse duration bits are reserved on LPT+.
+	 */
+	hotplug = intel_uncore_read(&dev_priv->uncore, PCH_PORT_HOTPLUG);
+	hotplug &= ~(PORTA_HOTPLUG_ENABLE |
+		     PORTB_HOTPLUG_ENABLE |
+		     PORTC_HOTPLUG_ENABLE |
+		     PORTD_HOTPLUG_ENABLE |
+		     PORTB_PULSE_DURATION_MASK |
+		     PORTC_PULSE_DURATION_MASK |
+		     PORTD_PULSE_DURATION_MASK);
+	hotplug |= intel_hpd_hotplug_enables(dev_priv, ibx_hotplug_enables);
+	intel_uncore_write(&dev_priv->uncore, PCH_PORT_HOTPLUG, hotplug);
+}
+
+static void ibx_hpd_irq_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug_irqs, enabled_irqs;
+
+	enabled_irqs = intel_hpd_enabled_irqs(dev_priv, dev_priv->display.hotplug.pch_hpd);
+	hotplug_irqs = intel_hpd_hotplug_irqs(dev_priv, dev_priv->display.hotplug.pch_hpd);
+
+	ibx_display_interrupt_update(dev_priv, hotplug_irqs, enabled_irqs);
+
+	ibx_hpd_detection_setup(dev_priv);
+}
+
+static u32 icp_ddi_hotplug_enables(struct drm_i915_private *i915,
+				   enum hpd_pin pin)
+{
+	switch (pin) {
+	case HPD_PORT_A:
+	case HPD_PORT_B:
+	case HPD_PORT_C:
+	case HPD_PORT_D:
+		return SHOTPLUG_CTL_DDI_HPD_ENABLE(pin);
+	default:
+		return 0;
+	}
+}
+
+static u32 icp_tc_hotplug_enables(struct drm_i915_private *i915,
+				  enum hpd_pin pin)
+{
+	switch (pin) {
+	case HPD_PORT_TC1:
+	case HPD_PORT_TC2:
+	case HPD_PORT_TC3:
+	case HPD_PORT_TC4:
+	case HPD_PORT_TC5:
+	case HPD_PORT_TC6:
+		return ICP_TC_HPD_ENABLE(pin);
+	default:
+		return 0;
+	}
+}
+
+static void icp_ddi_hpd_detection_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug;
+
+	hotplug = intel_uncore_read(&dev_priv->uncore, SHOTPLUG_CTL_DDI);
+	hotplug &= ~(SHOTPLUG_CTL_DDI_HPD_ENABLE(HPD_PORT_A) |
+		     SHOTPLUG_CTL_DDI_HPD_ENABLE(HPD_PORT_B) |
+		     SHOTPLUG_CTL_DDI_HPD_ENABLE(HPD_PORT_C) |
+		     SHOTPLUG_CTL_DDI_HPD_ENABLE(HPD_PORT_D));
+	hotplug |= intel_hpd_hotplug_enables(dev_priv, icp_ddi_hotplug_enables);
+	intel_uncore_write(&dev_priv->uncore, SHOTPLUG_CTL_DDI, hotplug);
+}
+
+static void icp_tc_hpd_detection_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug;
+
+	hotplug = intel_uncore_read(&dev_priv->uncore, SHOTPLUG_CTL_TC);
+	hotplug &= ~(ICP_TC_HPD_ENABLE(HPD_PORT_TC1) |
+		     ICP_TC_HPD_ENABLE(HPD_PORT_TC2) |
+		     ICP_TC_HPD_ENABLE(HPD_PORT_TC3) |
+		     ICP_TC_HPD_ENABLE(HPD_PORT_TC4) |
+		     ICP_TC_HPD_ENABLE(HPD_PORT_TC5) |
+		     ICP_TC_HPD_ENABLE(HPD_PORT_TC6));
+	hotplug |= intel_hpd_hotplug_enables(dev_priv, icp_tc_hotplug_enables);
+	intel_uncore_write(&dev_priv->uncore, SHOTPLUG_CTL_TC, hotplug);
+}
+
+static void icp_hpd_irq_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug_irqs, enabled_irqs;
+
+	enabled_irqs = intel_hpd_enabled_irqs(dev_priv, dev_priv->display.hotplug.pch_hpd);
+	hotplug_irqs = intel_hpd_hotplug_irqs(dev_priv, dev_priv->display.hotplug.pch_hpd);
+
+	if (INTEL_PCH_TYPE(dev_priv) <= PCH_TGP)
+		intel_uncore_write(&dev_priv->uncore, SHPD_FILTER_CNT, SHPD_FILTER_CNT_500_ADJ);
+
+	ibx_display_interrupt_update(dev_priv, hotplug_irqs, enabled_irqs);
+
+	icp_ddi_hpd_detection_setup(dev_priv);
+	icp_tc_hpd_detection_setup(dev_priv);
+}
+
+static u32 gen11_hotplug_enables(struct drm_i915_private *i915,
+				 enum hpd_pin pin)
+{
+	switch (pin) {
+	case HPD_PORT_TC1:
+	case HPD_PORT_TC2:
+	case HPD_PORT_TC3:
+	case HPD_PORT_TC4:
+	case HPD_PORT_TC5:
+	case HPD_PORT_TC6:
+		return GEN11_HOTPLUG_CTL_ENABLE(pin);
+	default:
+		return 0;
+	}
+}
+
+static void dg1_hpd_irq_setup(struct drm_i915_private *dev_priv)
+{
+	u32 val;
+
+	val = intel_uncore_read(&dev_priv->uncore, SOUTH_CHICKEN1);
+	val |= (INVERT_DDIA_HPD |
+		INVERT_DDIB_HPD |
+		INVERT_DDIC_HPD |
+		INVERT_DDID_HPD);
+	intel_uncore_write(&dev_priv->uncore, SOUTH_CHICKEN1, val);
+
+	icp_hpd_irq_setup(dev_priv);
+}
+
+static void gen11_tc_hpd_detection_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug;
+
+	hotplug = intel_uncore_read(&dev_priv->uncore, GEN11_TC_HOTPLUG_CTL);
+	hotplug &= ~(GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC1) |
+		     GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC2) |
+		     GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC3) |
+		     GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC4) |
+		     GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC5) |
+		     GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC6));
+	hotplug |= intel_hpd_hotplug_enables(dev_priv, gen11_hotplug_enables);
+	intel_uncore_write(&dev_priv->uncore, GEN11_TC_HOTPLUG_CTL, hotplug);
+}
+
+static void gen11_tbt_hpd_detection_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug;
+
+	hotplug = intel_uncore_read(&dev_priv->uncore, GEN11_TBT_HOTPLUG_CTL);
+	hotplug &= ~(GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC1) |
+		     GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC2) |
+		     GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC3) |
+		     GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC4) |
+		     GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC5) |
+		     GEN11_HOTPLUG_CTL_ENABLE(HPD_PORT_TC6));
+	hotplug |= intel_hpd_hotplug_enables(dev_priv, gen11_hotplug_enables);
+	intel_uncore_write(&dev_priv->uncore, GEN11_TBT_HOTPLUG_CTL, hotplug);
+}
+
+static void gen11_hpd_irq_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug_irqs, enabled_irqs;
+	u32 val;
+
+	enabled_irqs = intel_hpd_enabled_irqs(dev_priv, dev_priv->display.hotplug.hpd);
+	hotplug_irqs = intel_hpd_hotplug_irqs(dev_priv, dev_priv->display.hotplug.hpd);
+
+	val = intel_uncore_read(&dev_priv->uncore, GEN11_DE_HPD_IMR);
+	val &= ~hotplug_irqs;
+	val |= ~enabled_irqs & hotplug_irqs;
+	intel_uncore_write(&dev_priv->uncore, GEN11_DE_HPD_IMR, val);
+	intel_uncore_posting_read(&dev_priv->uncore, GEN11_DE_HPD_IMR);
+
+	gen11_tc_hpd_detection_setup(dev_priv);
+	gen11_tbt_hpd_detection_setup(dev_priv);
+
+	if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP)
+		icp_hpd_irq_setup(dev_priv);
+}
+
+static u32 spt_hotplug_enables(struct drm_i915_private *i915,
+			       enum hpd_pin pin)
+{
+	switch (pin) {
+	case HPD_PORT_A:
+		return PORTA_HOTPLUG_ENABLE;
+	case HPD_PORT_B:
+		return PORTB_HOTPLUG_ENABLE;
+	case HPD_PORT_C:
+		return PORTC_HOTPLUG_ENABLE;
+	case HPD_PORT_D:
+		return PORTD_HOTPLUG_ENABLE;
+	default:
+		return 0;
+	}
+}
+
+static u32 spt_hotplug2_enables(struct drm_i915_private *i915,
+				enum hpd_pin pin)
+{
+	switch (pin) {
+	case HPD_PORT_E:
+		return PORTE_HOTPLUG_ENABLE;
+	default:
+		return 0;
+	}
+}
+
+static void spt_hpd_detection_setup(struct drm_i915_private *dev_priv)
+{
+	u32 val, hotplug;
+
+	/* Display WA #1179 WaHardHangonHotPlug: cnp */
+	if (HAS_PCH_CNP(dev_priv)) {
+		val = intel_uncore_read(&dev_priv->uncore, SOUTH_CHICKEN1);
+		val &= ~CHASSIS_CLK_REQ_DURATION_MASK;
+		val |= CHASSIS_CLK_REQ_DURATION(0xf);
+		intel_uncore_write(&dev_priv->uncore, SOUTH_CHICKEN1, val);
+	}
+
+	/* Enable digital hotplug on the PCH */
+	hotplug = intel_uncore_read(&dev_priv->uncore, PCH_PORT_HOTPLUG);
+	hotplug &= ~(PORTA_HOTPLUG_ENABLE |
+		     PORTB_HOTPLUG_ENABLE |
+		     PORTC_HOTPLUG_ENABLE |
+		     PORTD_HOTPLUG_ENABLE);
+	hotplug |= intel_hpd_hotplug_enables(dev_priv, spt_hotplug_enables);
+	intel_uncore_write(&dev_priv->uncore, PCH_PORT_HOTPLUG, hotplug);
+
+	hotplug = intel_uncore_read(&dev_priv->uncore, PCH_PORT_HOTPLUG2);
+	hotplug &= ~PORTE_HOTPLUG_ENABLE;
+	hotplug |= intel_hpd_hotplug_enables(dev_priv, spt_hotplug2_enables);
+	intel_uncore_write(&dev_priv->uncore, PCH_PORT_HOTPLUG2, hotplug);
+}
+
+static void spt_hpd_irq_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug_irqs, enabled_irqs;
+
+	if (INTEL_PCH_TYPE(dev_priv) >= PCH_CNP)
+		intel_uncore_write(&dev_priv->uncore, SHPD_FILTER_CNT, SHPD_FILTER_CNT_500_ADJ);
+
+	enabled_irqs = intel_hpd_enabled_irqs(dev_priv, dev_priv->display.hotplug.pch_hpd);
+	hotplug_irqs = intel_hpd_hotplug_irqs(dev_priv, dev_priv->display.hotplug.pch_hpd);
+
+	ibx_display_interrupt_update(dev_priv, hotplug_irqs, enabled_irqs);
+
+	spt_hpd_detection_setup(dev_priv);
+}
+
+static u32 ilk_hotplug_enables(struct drm_i915_private *i915,
+			       enum hpd_pin pin)
+{
+	switch (pin) {
+	case HPD_PORT_A:
+		return DIGITAL_PORTA_HOTPLUG_ENABLE |
+			DIGITAL_PORTA_PULSE_DURATION_2ms;
+	default:
+		return 0;
+	}
+}
+
+static void ilk_hpd_detection_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug;
+
+	/*
+	 * Enable digital hotplug on the CPU, and configure the DP short pulse
+	 * duration to 2ms (which is the minimum in the Display Port spec)
+	 * The pulse duration bits are reserved on HSW+.
+	 */
+	hotplug = intel_uncore_read(&dev_priv->uncore, DIGITAL_PORT_HOTPLUG_CNTRL);
+	hotplug &= ~(DIGITAL_PORTA_HOTPLUG_ENABLE |
+		     DIGITAL_PORTA_PULSE_DURATION_MASK);
+	hotplug |= intel_hpd_hotplug_enables(dev_priv, ilk_hotplug_enables);
+	intel_uncore_write(&dev_priv->uncore, DIGITAL_PORT_HOTPLUG_CNTRL, hotplug);
+}
+
+static void ilk_hpd_irq_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug_irqs, enabled_irqs;
+
+	enabled_irqs = intel_hpd_enabled_irqs(dev_priv, dev_priv->display.hotplug.hpd);
+	hotplug_irqs = intel_hpd_hotplug_irqs(dev_priv, dev_priv->display.hotplug.hpd);
+
+	if (DISPLAY_VER(dev_priv) >= 8)
+		bdw_update_port_irq(dev_priv, hotplug_irqs, enabled_irqs);
+	else
+		ilk_update_display_irq(dev_priv, hotplug_irqs, enabled_irqs);
+
+	ilk_hpd_detection_setup(dev_priv);
+
+	ibx_hpd_irq_setup(dev_priv);
+}
+
+static u32 bxt_hotplug_enables(struct drm_i915_private *i915,
+			       enum hpd_pin pin)
+{
+	u32 hotplug;
+
+	switch (pin) {
+	case HPD_PORT_A:
+		hotplug = PORTA_HOTPLUG_ENABLE;
+		if (intel_bios_is_port_hpd_inverted(i915, PORT_A))
+			hotplug |= BXT_DDIA_HPD_INVERT;
+		return hotplug;
+	case HPD_PORT_B:
+		hotplug = PORTB_HOTPLUG_ENABLE;
+		if (intel_bios_is_port_hpd_inverted(i915, PORT_B))
+			hotplug |= BXT_DDIB_HPD_INVERT;
+		return hotplug;
+	case HPD_PORT_C:
+		hotplug = PORTC_HOTPLUG_ENABLE;
+		if (intel_bios_is_port_hpd_inverted(i915, PORT_C))
+			hotplug |= BXT_DDIC_HPD_INVERT;
+		return hotplug;
+	default:
+		return 0;
+	}
+}
+
+static void bxt_hpd_detection_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug;
+
+	hotplug = intel_uncore_read(&dev_priv->uncore, PCH_PORT_HOTPLUG);
+	hotplug &= ~(PORTA_HOTPLUG_ENABLE |
+		     PORTB_HOTPLUG_ENABLE |
+		     PORTC_HOTPLUG_ENABLE |
+		     BXT_DDIA_HPD_INVERT |
+		     BXT_DDIB_HPD_INVERT |
+		     BXT_DDIC_HPD_INVERT);
+	hotplug |= intel_hpd_hotplug_enables(dev_priv, bxt_hotplug_enables);
+	intel_uncore_write(&dev_priv->uncore, PCH_PORT_HOTPLUG, hotplug);
+}
+
+static void bxt_hpd_irq_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug_irqs, enabled_irqs;
+
+	enabled_irqs = intel_hpd_enabled_irqs(dev_priv, dev_priv->display.hotplug.hpd);
+	hotplug_irqs = intel_hpd_hotplug_irqs(dev_priv, dev_priv->display.hotplug.hpd);
+
+	bdw_update_port_irq(dev_priv, hotplug_irqs, enabled_irqs);
+
+	bxt_hpd_detection_setup(dev_priv);
+}
+
+/*
+ * SDEIER is also touched by the interrupt handler to work around missed PCH
+ * interrupts. Hence we can't update it after the interrupt handler is enabled -
+ * instead we unconditionally enable all PCH interrupt sources here, but then
+ * only unmask them as needed with SDEIMR.
+ *
+ * Note that we currently do this after installing the interrupt handler,
+ * but before we enable the master interrupt. That should be sufficient
+ * to avoid races with the irq handler, assuming we have MSI. Shared legacy
+ * interrupts could still race.
+ */
+static void ibx_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	u32 mask;
+
+	if (HAS_PCH_NOP(dev_priv))
+		return;
+
+	if (HAS_PCH_IBX(dev_priv))
+		mask = SDE_GMBUS | SDE_AUX_MASK | SDE_POISON;
+	else if (HAS_PCH_CPT(dev_priv) || HAS_PCH_LPT(dev_priv))
+		mask = SDE_GMBUS_CPT | SDE_AUX_MASK_CPT;
+	else
+		mask = SDE_GMBUS_CPT;
+
+	GEN3_IRQ_INIT(uncore, SDE, ~mask, 0xffffffff);
+}
+
+static void ilk_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	u32 display_mask, extra_mask;
+
+	if (GRAPHICS_VER(dev_priv) >= 7) {
+		display_mask = (DE_MASTER_IRQ_CONTROL | DE_GSE_IVB |
+				DE_PCH_EVENT_IVB | DE_AUX_CHANNEL_A_IVB);
+		extra_mask = (DE_PIPEC_VBLANK_IVB | DE_PIPEB_VBLANK_IVB |
+			      DE_PIPEA_VBLANK_IVB | DE_ERR_INT_IVB |
+			      DE_PLANE_FLIP_DONE_IVB(PLANE_C) |
+			      DE_PLANE_FLIP_DONE_IVB(PLANE_B) |
+			      DE_PLANE_FLIP_DONE_IVB(PLANE_A) |
+			      DE_DP_A_HOTPLUG_IVB);
+	} else {
+		display_mask = (DE_MASTER_IRQ_CONTROL | DE_GSE | DE_PCH_EVENT |
+				DE_AUX_CHANNEL_A | DE_PIPEB_CRC_DONE |
+				DE_PIPEA_CRC_DONE | DE_POISON);
+		extra_mask = (DE_PIPEA_VBLANK | DE_PIPEB_VBLANK |
+			      DE_PIPEB_FIFO_UNDERRUN | DE_PIPEA_FIFO_UNDERRUN |
+			      DE_PLANE_FLIP_DONE(PLANE_A) |
+			      DE_PLANE_FLIP_DONE(PLANE_B) |
+			      DE_DP_A_HOTPLUG);
+	}
+
+	if (IS_HASWELL(dev_priv)) {
+		gen3_assert_iir_is_zero(uncore, EDP_PSR_IIR);
+		display_mask |= DE_EDP_PSR_INT_HSW;
+	}
+
+	if (IS_IRONLAKE_M(dev_priv))
+		extra_mask |= DE_PCU_EVENT;
+
+	dev_priv->irq_mask = ~display_mask;
+
+	ibx_irq_postinstall(dev_priv);
+
+	gen5_gt_irq_postinstall(to_gt(dev_priv));
+
+	GEN3_IRQ_INIT(uncore, DE, dev_priv->irq_mask,
+		      display_mask | extra_mask);
+}
+
+void valleyview_enable_display_irqs(struct drm_i915_private *dev_priv)
+{
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	if (dev_priv->display_irqs_enabled)
+		return;
+
+	dev_priv->display_irqs_enabled = true;
+
+	if (intel_irqs_enabled(dev_priv)) {
+		vlv_display_irq_reset(dev_priv);
+		vlv_display_irq_postinstall(dev_priv);
+	}
+}
+
+void valleyview_disable_display_irqs(struct drm_i915_private *dev_priv)
+{
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	if (!dev_priv->display_irqs_enabled)
+		return;
+
+	dev_priv->display_irqs_enabled = false;
+
+	if (intel_irqs_enabled(dev_priv))
+		vlv_display_irq_reset(dev_priv);
+}
+
+
+static void valleyview_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	gen5_gt_irq_postinstall(to_gt(dev_priv));
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	if (dev_priv->display_irqs_enabled)
+		vlv_display_irq_postinstall(dev_priv);
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	intel_uncore_write(&dev_priv->uncore, VLV_MASTER_IER, MASTER_INTERRUPT_ENABLE);
+	intel_uncore_posting_read(&dev_priv->uncore, VLV_MASTER_IER);
+}
+
+static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	u32 de_pipe_masked = gen8_de_pipe_fault_mask(dev_priv) |
+		GEN8_PIPE_CDCLK_CRC_DONE;
+	u32 de_pipe_enables;
+	u32 de_port_masked = gen8_de_port_aux_mask(dev_priv);
+	u32 de_port_enables;
+	u32 de_misc_masked = GEN8_DE_EDP_PSR;
+	u32 trans_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) |
+		BIT(TRANSCODER_C) | BIT(TRANSCODER_D);
+	enum pipe pipe;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	if (DISPLAY_VER(dev_priv) <= 10)
+		de_misc_masked |= GEN8_DE_MISC_GSE;
+
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		de_port_masked |= BXT_DE_PORT_GMBUS;
+
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		enum port port;
+
+		if (intel_bios_is_dsi_present(dev_priv, &port))
+			de_port_masked |= DSI0_TE | DSI1_TE;
+	}
+
+	de_pipe_enables = de_pipe_masked |
+		GEN8_PIPE_VBLANK |
+		gen8_de_pipe_underrun_mask(dev_priv) |
+		gen8_de_pipe_flip_done_mask(dev_priv);
+
+	de_port_enables = de_port_masked;
+	if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+		de_port_enables |= BXT_DE_PORT_HOTPLUG_MASK;
+	else if (IS_BROADWELL(dev_priv))
+		de_port_enables |= BDW_DE_PORT_HOTPLUG_MASK;
+
+	if (DISPLAY_VER(dev_priv) >= 12) {
+		enum transcoder trans;
+
+		for_each_cpu_transcoder_masked(dev_priv, trans, trans_mask) {
+			enum intel_display_power_domain domain;
+
+			domain = POWER_DOMAIN_TRANSCODER(trans);
+			if (!intel_display_power_is_enabled(dev_priv, domain))
+				continue;
+
+			gen3_assert_iir_is_zero(uncore, TRANS_PSR_IIR(trans));
+		}
+	} else {
+		gen3_assert_iir_is_zero(uncore, EDP_PSR_IIR);
+	}
+
+	for_each_pipe(dev_priv, pipe) {
+		dev_priv->de_irq_mask[pipe] = ~de_pipe_masked;
+
+		if (intel_display_power_is_enabled(dev_priv,
+				POWER_DOMAIN_PIPE(pipe)))
+			GEN8_IRQ_INIT_NDX(uncore, DE_PIPE, pipe,
+					  dev_priv->de_irq_mask[pipe],
+					  de_pipe_enables);
+	}
+
+	GEN3_IRQ_INIT(uncore, GEN8_DE_PORT_, ~de_port_masked, de_port_enables);
+	GEN3_IRQ_INIT(uncore, GEN8_DE_MISC_, ~de_misc_masked, de_misc_masked);
+
+	if (DISPLAY_VER(dev_priv) >= 11) {
+		u32 de_hpd_masked = 0;
+		u32 de_hpd_enables = GEN11_DE_TC_HOTPLUG_MASK |
+				     GEN11_DE_TBT_HOTPLUG_MASK;
+
+		GEN3_IRQ_INIT(uncore, GEN11_DE_HPD_, ~de_hpd_masked,
+			      de_hpd_enables);
+	}
+}
+
+static void icp_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	u32 mask = SDE_GMBUS_ICP;
+
+	GEN3_IRQ_INIT(uncore, SDE, ~mask, 0xffffffff);
+}
+
+static void gen8_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP)
+		icp_irq_postinstall(dev_priv);
+	else if (HAS_PCH_SPLIT(dev_priv))
+		ibx_irq_postinstall(dev_priv);
+
+	gen8_gt_irq_postinstall(to_gt(dev_priv));
+	gen8_de_irq_postinstall(dev_priv);
+
+	gen8_master_intr_enable(dev_priv->uncore.regs);
+}
+
+static void gen11_de_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	gen8_de_irq_postinstall(dev_priv);
+
+	intel_uncore_write(&dev_priv->uncore, GEN11_DISPLAY_INT_CTL,
+			   GEN11_DISPLAY_IRQ_ENABLE);
+}
+
+static void gen11_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	struct intel_gt *gt = to_gt(dev_priv);
+	struct intel_uncore *uncore = gt->uncore;
+	u32 gu_misc_masked = GEN11_GU_MISC_GSE;
+
+	if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP)
+		icp_irq_postinstall(dev_priv);
+
+	gen11_gt_irq_postinstall(gt);
+	gen11_de_irq_postinstall(dev_priv);
+
+	GEN3_IRQ_INIT(uncore, GEN11_GU_MISC_, ~gu_misc_masked, gu_misc_masked);
+
+	gen11_master_intr_enable(uncore->regs);
+	intel_uncore_posting_read(&dev_priv->uncore, GEN11_GFX_MSTR_IRQ);
+}
+
+static void dg1_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	struct intel_gt *gt = to_gt(dev_priv);
+	struct intel_uncore *uncore = gt->uncore;
+	u32 gu_misc_masked = GEN11_GU_MISC_GSE;
+
+	gen11_gt_irq_postinstall(gt);
+
+	GEN3_IRQ_INIT(uncore, GEN11_GU_MISC_, ~gu_misc_masked, gu_misc_masked);
+
+	if (HAS_DISPLAY(dev_priv)) {
+		icp_irq_postinstall(dev_priv);
+		gen8_de_irq_postinstall(dev_priv);
+		intel_uncore_write(&dev_priv->uncore, GEN11_DISPLAY_INT_CTL,
+				   GEN11_DISPLAY_IRQ_ENABLE);
+	}
+
+	dg1_master_intr_enable(uncore->regs);
+	intel_uncore_posting_read(uncore, DG1_MSTR_TILE_INTR);
+}
+
+static void cherryview_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	gen8_gt_irq_postinstall(to_gt(dev_priv));
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	if (dev_priv->display_irqs_enabled)
+		vlv_display_irq_postinstall(dev_priv);
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	intel_uncore_write(&dev_priv->uncore, GEN8_MASTER_IRQ, GEN8_MASTER_IRQ_CONTROL);
+	intel_uncore_posting_read(&dev_priv->uncore, GEN8_MASTER_IRQ);
+}
+
+static void i8xx_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	i9xx_pipestat_irq_reset(dev_priv);
+
+	GEN2_IRQ_RESET(uncore);
+	dev_priv->irq_mask = ~0u;
+}
+
+static void i8xx_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	u16 enable_mask;
+
+	intel_uncore_write16(uncore,
+			     EMR,
+			     ~(I915_ERROR_PAGE_TABLE |
+			       I915_ERROR_MEMORY_REFRESH));
+
+	/* Unmask the interrupts that we always want on. */
+	dev_priv->irq_mask =
+		~(I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
+		  I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
+		  I915_MASTER_ERROR_INTERRUPT);
+
+	enable_mask =
+		I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
+		I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
+		I915_MASTER_ERROR_INTERRUPT |
+		I915_USER_INTERRUPT;
+
+	GEN2_IRQ_INIT(uncore, dev_priv->irq_mask, enable_mask);
+
+	/* Interrupt setup is already guaranteed to be single-threaded, this is
+	 * just to make the assert_spin_locked check happy. */
+	spin_lock_irq(&dev_priv->irq_lock);
+	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS);
+	i915_enable_pipestat(dev_priv, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS);
+	spin_unlock_irq(&dev_priv->irq_lock);
+}
+
+static void i8xx_error_irq_ack(struct drm_i915_private *i915,
+			       u16 *eir, u16 *eir_stuck)
+{
+	struct intel_uncore *uncore = &i915->uncore;
+	u16 emr;
+
+	*eir = intel_uncore_read16(uncore, EIR);
+
+	if (*eir)
+		intel_uncore_write16(uncore, EIR, *eir);
+
+	*eir_stuck = intel_uncore_read16(uncore, EIR);
+	if (*eir_stuck == 0)
+		return;
+
+	/*
+	 * Toggle all EMR bits to make sure we get an edge
+	 * in the ISR master error bit if we don't clear
+	 * all the EIR bits. Otherwise the edge triggered
+	 * IIR on i965/g4x wouldn't notice that an interrupt
+	 * is still pending. Also some EIR bits can't be
+	 * cleared except by handling the underlying error
+	 * (or by a GPU reset) so we mask any bit that
+	 * remains set.
+	 */
+	emr = intel_uncore_read16(uncore, EMR);
+	intel_uncore_write16(uncore, EMR, 0xffff);
+	intel_uncore_write16(uncore, EMR, emr | *eir_stuck);
+}
+
+static void i8xx_error_irq_handler(struct drm_i915_private *dev_priv,
+				   u16 eir, u16 eir_stuck)
+{
+	DRM_DEBUG("Master Error: EIR 0x%04x\n", eir);
+
+	if (eir_stuck)
+		drm_dbg(&dev_priv->drm, "EIR stuck: 0x%04x, masked\n",
+			eir_stuck);
+}
+
+static void i9xx_error_irq_ack(struct drm_i915_private *dev_priv,
+			       u32 *eir, u32 *eir_stuck)
+{
+	u32 emr;
+
+	*eir = intel_uncore_read(&dev_priv->uncore, EIR);
+
+	intel_uncore_write(&dev_priv->uncore, EIR, *eir);
+
+	*eir_stuck = intel_uncore_read(&dev_priv->uncore, EIR);
+	if (*eir_stuck == 0)
+		return;
+
+	/*
+	 * Toggle all EMR bits to make sure we get an edge
+	 * in the ISR master error bit if we don't clear
+	 * all the EIR bits. Otherwise the edge triggered
+	 * IIR on i965/g4x wouldn't notice that an interrupt
+	 * is still pending. Also some EIR bits can't be
+	 * cleared except by handling the underlying error
+	 * (or by a GPU reset) so we mask any bit that
+	 * remains set.
+	 */
+	emr = intel_uncore_read(&dev_priv->uncore, EMR);
+	intel_uncore_write(&dev_priv->uncore, EMR, 0xffffffff);
+	intel_uncore_write(&dev_priv->uncore, EMR, emr | *eir_stuck);
+}
+
+static void i9xx_error_irq_handler(struct drm_i915_private *dev_priv,
+				   u32 eir, u32 eir_stuck)
+{
+	DRM_DEBUG("Master Error, EIR 0x%08x\n", eir);
+
+	if (eir_stuck)
+		drm_dbg(&dev_priv->drm, "EIR stuck: 0x%08x, masked\n",
+			eir_stuck);
+}
+
+static irqreturn_t i8xx_irq_handler(int irq, void *arg)
+{
+	struct drm_i915_private *dev_priv = arg;
+	irqreturn_t ret = IRQ_NONE;
+
+	if (!intel_irqs_enabled(dev_priv))
+		return IRQ_NONE;
+
+	/* IRQs are synced during runtime_suspend, we don't require a wakeref */
+	disable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	do {
+		u32 pipe_stats[I915_MAX_PIPES] = {};
+		u16 eir = 0, eir_stuck = 0;
+		u16 iir;
+
+		iir = intel_uncore_read16(&dev_priv->uncore, GEN2_IIR);
+		if (iir == 0)
+			break;
+
+		ret = IRQ_HANDLED;
+
+		/* Call regardless, as some status bits might not be
+		 * signalled in iir */
+		i9xx_pipestat_irq_ack(dev_priv, iir, pipe_stats);
+
+		if (iir & I915_MASTER_ERROR_INTERRUPT)
+			i8xx_error_irq_ack(dev_priv, &eir, &eir_stuck);
+
+		intel_uncore_write16(&dev_priv->uncore, GEN2_IIR, iir);
+
+		if (iir & I915_USER_INTERRUPT)
+			intel_engine_cs_irq(to_gt(dev_priv)->engine[RCS0], iir);
+
+		if (iir & I915_MASTER_ERROR_INTERRUPT)
+			i8xx_error_irq_handler(dev_priv, eir, eir_stuck);
+
+		i8xx_pipestat_irq_handler(dev_priv, iir, pipe_stats);
+	} while (0);
+
+	pmu_irq_stats(dev_priv, ret);
+
+	enable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	return ret;
+}
+
+static void i915_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	if (I915_HAS_HOTPLUG(dev_priv)) {
+		i915_hotplug_interrupt_update(dev_priv, 0xffffffff, 0);
+		intel_uncore_write(&dev_priv->uncore, PORT_HOTPLUG_STAT, intel_uncore_read(&dev_priv->uncore, PORT_HOTPLUG_STAT));
+	}
+
+	i9xx_pipestat_irq_reset(dev_priv);
+
+	GEN3_IRQ_RESET(uncore, GEN2_);
+	dev_priv->irq_mask = ~0u;
+}
+
+static void i915_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	u32 enable_mask;
+
+	intel_uncore_write(&dev_priv->uncore, EMR, ~(I915_ERROR_PAGE_TABLE |
+			  I915_ERROR_MEMORY_REFRESH));
+
+	/* Unmask the interrupts that we always want on. */
+	dev_priv->irq_mask =
+		~(I915_ASLE_INTERRUPT |
+		  I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
+		  I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
+		  I915_MASTER_ERROR_INTERRUPT);
+
+	enable_mask =
+		I915_ASLE_INTERRUPT |
+		I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
+		I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
+		I915_MASTER_ERROR_INTERRUPT |
+		I915_USER_INTERRUPT;
+
+	if (I915_HAS_HOTPLUG(dev_priv)) {
+		/* Enable in IER... */
+		enable_mask |= I915_DISPLAY_PORT_INTERRUPT;
+		/* and unmask in IMR */
+		dev_priv->irq_mask &= ~I915_DISPLAY_PORT_INTERRUPT;
+	}
+
+	GEN3_IRQ_INIT(uncore, GEN2_, dev_priv->irq_mask, enable_mask);
+
+	/* Interrupt setup is already guaranteed to be single-threaded, this is
+	 * just to make the assert_spin_locked check happy. */
+	spin_lock_irq(&dev_priv->irq_lock);
+	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS);
+	i915_enable_pipestat(dev_priv, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS);
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	i915_enable_asle_pipestat(dev_priv);
+}
+
+static irqreturn_t i915_irq_handler(int irq, void *arg)
+{
+	struct drm_i915_private *dev_priv = arg;
+	irqreturn_t ret = IRQ_NONE;
+
+	if (!intel_irqs_enabled(dev_priv))
+		return IRQ_NONE;
+
+	/* IRQs are synced during runtime_suspend, we don't require a wakeref */
+	disable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	do {
+		u32 pipe_stats[I915_MAX_PIPES] = {};
+		u32 eir = 0, eir_stuck = 0;
+		u32 hotplug_status = 0;
+		u32 iir;
+
+		iir = intel_uncore_read(&dev_priv->uncore, GEN2_IIR);
+		if (iir == 0)
+			break;
+
+		ret = IRQ_HANDLED;
+
+		if (I915_HAS_HOTPLUG(dev_priv) &&
+		    iir & I915_DISPLAY_PORT_INTERRUPT)
+			hotplug_status = i9xx_hpd_irq_ack(dev_priv);
+
+		/* Call regardless, as some status bits might not be
+		 * signalled in iir */
+		i9xx_pipestat_irq_ack(dev_priv, iir, pipe_stats);
+
+		if (iir & I915_MASTER_ERROR_INTERRUPT)
+			i9xx_error_irq_ack(dev_priv, &eir, &eir_stuck);
+
+		intel_uncore_write(&dev_priv->uncore, GEN2_IIR, iir);
+
+		if (iir & I915_USER_INTERRUPT)
+			intel_engine_cs_irq(to_gt(dev_priv)->engine[RCS0], iir);
+
+		if (iir & I915_MASTER_ERROR_INTERRUPT)
+			i9xx_error_irq_handler(dev_priv, eir, eir_stuck);
+
+		if (hotplug_status)
+			i9xx_hpd_irq_handler(dev_priv, hotplug_status);
+
+		i915_pipestat_irq_handler(dev_priv, iir, pipe_stats);
+	} while (0);
+
+	pmu_irq_stats(dev_priv, ret);
+
+	enable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	return ret;
+}
+
+static void i965_irq_reset(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	i915_hotplug_interrupt_update(dev_priv, 0xffffffff, 0);
+	intel_uncore_write(&dev_priv->uncore, PORT_HOTPLUG_STAT, intel_uncore_read(&dev_priv->uncore, PORT_HOTPLUG_STAT));
+
+	i9xx_pipestat_irq_reset(dev_priv);
+
+	GEN3_IRQ_RESET(uncore, GEN2_);
+	dev_priv->irq_mask = ~0u;
+}
+
+static void i965_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	u32 enable_mask;
+	u32 error_mask;
+
+	/*
+	 * Enable some error detection, note the instruction error mask
+	 * bit is reserved, so we leave it masked.
+	 */
+	if (IS_G4X(dev_priv)) {
+		error_mask = ~(GM45_ERROR_PAGE_TABLE |
+			       GM45_ERROR_MEM_PRIV |
+			       GM45_ERROR_CP_PRIV |
+			       I915_ERROR_MEMORY_REFRESH);
+	} else {
+		error_mask = ~(I915_ERROR_PAGE_TABLE |
+			       I915_ERROR_MEMORY_REFRESH);
+	}
+	intel_uncore_write(&dev_priv->uncore, EMR, error_mask);
+
+	/* Unmask the interrupts that we always want on. */
+	dev_priv->irq_mask =
+		~(I915_ASLE_INTERRUPT |
+		  I915_DISPLAY_PORT_INTERRUPT |
+		  I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
+		  I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
+		  I915_MASTER_ERROR_INTERRUPT);
+
+	enable_mask =
+		I915_ASLE_INTERRUPT |
+		I915_DISPLAY_PORT_INTERRUPT |
+		I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
+		I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
+		I915_MASTER_ERROR_INTERRUPT |
+		I915_USER_INTERRUPT;
+
+	if (IS_G4X(dev_priv))
+		enable_mask |= I915_BSD_USER_INTERRUPT;
+
+	GEN3_IRQ_INIT(uncore, GEN2_, dev_priv->irq_mask, enable_mask);
+
+	/* Interrupt setup is already guaranteed to be single-threaded, this is
+	 * just to make the assert_spin_locked check happy. */
+	spin_lock_irq(&dev_priv->irq_lock);
+	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS);
+	i915_enable_pipestat(dev_priv, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS);
+	i915_enable_pipestat(dev_priv, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS);
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	i915_enable_asle_pipestat(dev_priv);
+}
+
+static void i915_hpd_irq_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug_en;
+
+	lockdep_assert_held(&dev_priv->irq_lock);
+
+	/* Note HDMI and DP share hotplug bits */
+	/* enable bits are the same for all generations */
+	hotplug_en = intel_hpd_enabled_irqs(dev_priv, hpd_mask_i915);
+	/* Programming the CRT detection parameters tends
+	   to generate a spurious hotplug event about three
+	   seconds later.  So just do it once.
+	*/
+	if (IS_G4X(dev_priv))
+		hotplug_en |= CRT_HOTPLUG_ACTIVATION_PERIOD_64;
+	hotplug_en |= CRT_HOTPLUG_VOLTAGE_COMPARE_50;
+
+	/* Ignore TV since it's buggy */
+	i915_hotplug_interrupt_update_locked(dev_priv,
+					     HOTPLUG_INT_EN_MASK |
+					     CRT_HOTPLUG_VOLTAGE_COMPARE_MASK |
+					     CRT_HOTPLUG_ACTIVATION_PERIOD_64,
+					     hotplug_en);
+}
+
+static irqreturn_t i965_irq_handler(int irq, void *arg)
+{
+	struct drm_i915_private *dev_priv = arg;
+	irqreturn_t ret = IRQ_NONE;
+
+	if (!intel_irqs_enabled(dev_priv))
+		return IRQ_NONE;
+
+	/* IRQs are synced during runtime_suspend, we don't require a wakeref */
+	disable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	do {
+		u32 pipe_stats[I915_MAX_PIPES] = {};
+		u32 eir = 0, eir_stuck = 0;
+		u32 hotplug_status = 0;
+		u32 iir;
+
+		iir = intel_uncore_read(&dev_priv->uncore, GEN2_IIR);
+		if (iir == 0)
+			break;
+
+		ret = IRQ_HANDLED;
+
+		if (iir & I915_DISPLAY_PORT_INTERRUPT)
+			hotplug_status = i9xx_hpd_irq_ack(dev_priv);
+
+		/* Call regardless, as some status bits might not be
+		 * signalled in iir */
+		i9xx_pipestat_irq_ack(dev_priv, iir, pipe_stats);
+
+		if (iir & I915_MASTER_ERROR_INTERRUPT)
+			i9xx_error_irq_ack(dev_priv, &eir, &eir_stuck);
+
+		intel_uncore_write(&dev_priv->uncore, GEN2_IIR, iir);
+
+		if (iir & I915_USER_INTERRUPT)
+			intel_engine_cs_irq(to_gt(dev_priv)->engine[RCS0],
+					    iir);
+
+		if (iir & I915_BSD_USER_INTERRUPT)
+			intel_engine_cs_irq(to_gt(dev_priv)->engine[VCS0],
+					    iir >> 25);
+
+		if (iir & I915_MASTER_ERROR_INTERRUPT)
+			i9xx_error_irq_handler(dev_priv, eir, eir_stuck);
+
+		if (hotplug_status)
+			i9xx_hpd_irq_handler(dev_priv, hotplug_status);
+
+		i965_pipestat_irq_handler(dev_priv, iir, pipe_stats);
+	} while (0);
+
+	pmu_irq_stats(dev_priv, IRQ_HANDLED);
+
+	enable_rpm_wakeref_asserts(&dev_priv->runtime_pm);
+
+	return ret;
+}
+
+struct intel_hotplug_funcs {
+	void (*hpd_irq_setup)(struct drm_i915_private *i915);
+};
+
+#define HPD_FUNCS(platform)					 \
+static const struct intel_hotplug_funcs platform##_hpd_funcs = { \
+	.hpd_irq_setup = platform##_hpd_irq_setup,		 \
+}
+
+HPD_FUNCS(i915);
+HPD_FUNCS(dg1);
+HPD_FUNCS(gen11);
+HPD_FUNCS(bxt);
+HPD_FUNCS(icp);
+HPD_FUNCS(spt);
+HPD_FUNCS(ilk);
+#undef HPD_FUNCS
+
+void intel_hpd_irq_setup(struct drm_i915_private *i915)
+{
+	if (i915->display_irqs_enabled && i915->display.funcs.hotplug)
+		i915->display.funcs.hotplug->hpd_irq_setup(i915);
+}
+
+/**
+ * intel_irq_init - initializes irq support
+ * @dev_priv: i915 device instance
+ *
+ * This function initializes all the irq support including work items, timers
+ * and all the vtables. It does not setup the interrupt itself though.
+ */
+void intel_irq_init(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *dev = &dev_priv->drm;
+	int i;
+
+	INIT_WORK(&dev_priv->l3_parity.error_work, ivb_parity_work);
+	for (i = 0; i < MAX_L3_SLICES; ++i)
+		dev_priv->l3_parity.remap_info[i] = NULL;
+
+	/* pre-gen11 the guc irqs bits are in the upper 16 bits of the pm reg */
+	if (HAS_GT_UC(dev_priv) && GRAPHICS_VER(dev_priv) < 11)
+		to_gt(dev_priv)->pm_guc_events = GUC_INTR_GUC2HOST << 16;
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	intel_hpd_init_pins(dev_priv);
+
+	intel_hpd_init_work(dev_priv);
+
+	dev->vblank_disable_immediate = true;
+
+	/* Most platforms treat the display irq block as an always-on
+	 * power domain. vlv/chv can disable it at runtime and need
+	 * special care to avoid writing any of the display block registers
+	 * outside of the power domain. We defer setting up the display irqs
+	 * in this case to the runtime pm.
+	 */
+	dev_priv->display_irqs_enabled = true;
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		dev_priv->display_irqs_enabled = false;
+
+	dev_priv->display.hotplug.hpd_storm_threshold = HPD_STORM_DEFAULT_THRESHOLD;
+	/* If we have MST support, we want to avoid doing short HPD IRQ storm
+	 * detection, as short HPD storms will occur as a natural part of
+	 * sideband messaging with MST.
+	 * On older platforms however, IRQ storms can occur with both long and
+	 * short pulses, as seen on some G4x systems.
+	 */
+	dev_priv->display.hotplug.hpd_short_storm_enabled = !HAS_DP_MST(dev_priv);
+
+	if (HAS_GMCH(dev_priv)) {
+		if (I915_HAS_HOTPLUG(dev_priv))
+			dev_priv->display.funcs.hotplug = &i915_hpd_funcs;
+	} else {
+		if (HAS_PCH_DG2(dev_priv))
+			dev_priv->display.funcs.hotplug = &icp_hpd_funcs;
+		else if (HAS_PCH_DG1(dev_priv))
+			dev_priv->display.funcs.hotplug = &dg1_hpd_funcs;
+		else if (DISPLAY_VER(dev_priv) >= 11)
+			dev_priv->display.funcs.hotplug = &gen11_hpd_funcs;
+		else if (IS_GEMINILAKE(dev_priv) || IS_BROXTON(dev_priv))
+			dev_priv->display.funcs.hotplug = &bxt_hpd_funcs;
+		else if (INTEL_PCH_TYPE(dev_priv) >= PCH_ICP)
+			dev_priv->display.funcs.hotplug = &icp_hpd_funcs;
+		else if (INTEL_PCH_TYPE(dev_priv) >= PCH_SPT)
+			dev_priv->display.funcs.hotplug = &spt_hpd_funcs;
+		else
+			dev_priv->display.funcs.hotplug = &ilk_hpd_funcs;
+	}
+}
+
+/**
+ * intel_irq_fini - deinitializes IRQ support
+ * @i915: i915 device instance
+ *
+ * This function deinitializes all the IRQ support.
+ */
+void intel_irq_fini(struct drm_i915_private *i915)
+{
+	int i;
+
+	for (i = 0; i < MAX_L3_SLICES; ++i)
+		kfree(i915->l3_parity.remap_info[i]);
+}
+
+static irq_handler_t intel_irq_handler(struct drm_i915_private *dev_priv)
+{
+	if (HAS_GMCH(dev_priv)) {
+		if (IS_CHERRYVIEW(dev_priv))
+			return cherryview_irq_handler;
+		else if (IS_VALLEYVIEW(dev_priv))
+			return valleyview_irq_handler;
+		else if (GRAPHICS_VER(dev_priv) == 4)
+			return i965_irq_handler;
+		else if (GRAPHICS_VER(dev_priv) == 3)
+			return i915_irq_handler;
+		else
+			return i8xx_irq_handler;
+	} else {
+		if (GRAPHICS_VER_FULL(dev_priv) >= IP_VER(12, 10))
+			return dg1_irq_handler;
+		else if (GRAPHICS_VER(dev_priv) >= 11)
+			return gen11_irq_handler;
+		else if (GRAPHICS_VER(dev_priv) >= 8)
+			return gen8_irq_handler;
+		else
+			return ilk_irq_handler;
+	}
+}
+
+static void intel_irq_reset(struct drm_i915_private *dev_priv)
+{
+	if (HAS_GMCH(dev_priv)) {
+		if (IS_CHERRYVIEW(dev_priv))
+			cherryview_irq_reset(dev_priv);
+		else if (IS_VALLEYVIEW(dev_priv))
+			valleyview_irq_reset(dev_priv);
+		else if (GRAPHICS_VER(dev_priv) == 4)
+			i965_irq_reset(dev_priv);
+		else if (GRAPHICS_VER(dev_priv) == 3)
+			i915_irq_reset(dev_priv);
+		else
+			i8xx_irq_reset(dev_priv);
+	} else {
+		if (GRAPHICS_VER_FULL(dev_priv) >= IP_VER(12, 10))
+			dg1_irq_reset(dev_priv);
+		else if (GRAPHICS_VER(dev_priv) >= 11)
+			gen11_irq_reset(dev_priv);
+		else if (GRAPHICS_VER(dev_priv) >= 8)
+			gen8_irq_reset(dev_priv);
+		else
+			ilk_irq_reset(dev_priv);
+	}
+}
+
+static void intel_irq_postinstall(struct drm_i915_private *dev_priv)
+{
+	if (HAS_GMCH(dev_priv)) {
+		if (IS_CHERRYVIEW(dev_priv))
+			cherryview_irq_postinstall(dev_priv);
+		else if (IS_VALLEYVIEW(dev_priv))
+			valleyview_irq_postinstall(dev_priv);
+		else if (GRAPHICS_VER(dev_priv) == 4)
+			i965_irq_postinstall(dev_priv);
+		else if (GRAPHICS_VER(dev_priv) == 3)
+			i915_irq_postinstall(dev_priv);
+		else
+			i8xx_irq_postinstall(dev_priv);
+	} else {
+		if (GRAPHICS_VER_FULL(dev_priv) >= IP_VER(12, 10))
+			dg1_irq_postinstall(dev_priv);
+		else if (GRAPHICS_VER(dev_priv) >= 11)
+			gen11_irq_postinstall(dev_priv);
+		else if (GRAPHICS_VER(dev_priv) >= 8)
+			gen8_irq_postinstall(dev_priv);
+		else
+			ilk_irq_postinstall(dev_priv);
+	}
+}
+
+/**
+ * intel_irq_install - enables the hardware interrupt
+ * @dev_priv: i915 device instance
+ *
+ * This function enables the hardware interrupt handling, but leaves the hotplug
+ * handling still disabled. It is called after intel_irq_init().
+ *
+ * In the driver load and resume code we need working interrupts in a few places
+ * but don't want to deal with the hassle of concurrent probe and hotplug
+ * workers. Hence the split into this two-stage approach.
+ */
+int intel_irq_install(struct drm_i915_private *dev_priv)
+{
+	int irq = to_pci_dev(dev_priv->drm.dev)->irq;
+	int ret;
+
+	/*
+	 * We enable some interrupt sources in our postinstall hooks, so mark
+	 * interrupts as enabled _before_ actually enabling them to avoid
+	 * special cases in our ordering checks.
+	 */
+	dev_priv->runtime_pm.irqs_enabled = true;
+
+	dev_priv->irq_enabled = true;
+
+	intel_irq_reset(dev_priv);
+
+	ret = request_irq(irq, intel_irq_handler(dev_priv),
+			  IRQF_SHARED, DRIVER_NAME, dev_priv);
+	if (ret < 0) {
+		dev_priv->irq_enabled = false;
+		return ret;
+	}
+
+	intel_irq_postinstall(dev_priv);
+
+	return ret;
+}
+
+/**
+ * intel_irq_uninstall - finilizes all irq handling
+ * @dev_priv: i915 device instance
+ *
+ * This stops interrupt and hotplug handling and unregisters and frees all
+ * resources acquired in the init functions.
+ */
+void intel_irq_uninstall(struct drm_i915_private *dev_priv)
+{
+	int irq = to_pci_dev(dev_priv->drm.dev)->irq;
+
+	/*
+	 * FIXME we can get called twice during driver probe
+	 * error handling as well as during driver remove due to
+	 * intel_modeset_driver_remove() calling us out of sequence.
+	 * Would be nice if it didn't do that...
+	 */
+	if (!dev_priv->irq_enabled)
+		return;
+
+	dev_priv->irq_enabled = false;
+
+	intel_irq_reset(dev_priv);
+
+	free_irq(irq, dev_priv);
+
+	intel_hpd_cancel_work(dev_priv);
+	dev_priv->runtime_pm.irqs_enabled = false;
+}
+
+/**
+ * intel_runtime_pm_disable_interrupts - runtime interrupt disabling
+ * @dev_priv: i915 device instance
+ *
+ * This function is used to disable interrupts at runtime, both in the runtime
+ * pm and the system suspend/resume code.
+ */
+void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv)
+{
+	intel_irq_reset(dev_priv);
+	dev_priv->runtime_pm.irqs_enabled = false;
+	intel_synchronize_irq(dev_priv);
+}
+
+/**
+ * intel_runtime_pm_enable_interrupts - runtime interrupt enabling
+ * @dev_priv: i915 device instance
+ *
+ * This function is used to enable interrupts at runtime, both in the runtime
+ * pm and the system suspend/resume code.
+ */
+void intel_runtime_pm_enable_interrupts(struct drm_i915_private *dev_priv)
+{
+	dev_priv->runtime_pm.irqs_enabled = true;
+	intel_irq_reset(dev_priv);
+	intel_irq_postinstall(dev_priv);
+}
+
+bool intel_irqs_enabled(struct drm_i915_private *dev_priv)
+{
+	return dev_priv->runtime_pm.irqs_enabled;
+}
+
+void intel_synchronize_irq(struct drm_i915_private *i915)
+{
+	synchronize_irq(to_pci_dev(i915->drm.dev)->irq);
+}
+
+void intel_synchronize_hardirq(struct drm_i915_private *i915)
+{
+	synchronize_hardirq(to_pci_dev(i915->drm.dev)->irq);
+}
diff --git a/drivers/gpu/drm/i915/i915_irq.h b/drivers/gpu/drm/i915/i915_irq.h
new file mode 100644
index 000000000..82639d9d7
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_irq.h
@@ -0,0 +1,135 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_IRQ_H__
+#define __I915_IRQ_H__
+
+#include <linux/ktime.h>
+#include <linux/types.h>
+
+#include "i915_reg.h"
+
+enum pipe;
+struct drm_crtc;
+struct drm_device;
+struct drm_display_mode;
+struct drm_i915_private;
+struct intel_crtc;
+struct intel_uncore;
+
+void intel_irq_init(struct drm_i915_private *dev_priv);
+void intel_irq_fini(struct drm_i915_private *dev_priv);
+int intel_irq_install(struct drm_i915_private *dev_priv);
+void intel_irq_uninstall(struct drm_i915_private *dev_priv);
+
+u32 i915_pipestat_enable_mask(struct drm_i915_private *dev_priv,
+			      enum pipe pipe);
+void
+i915_enable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
+		     u32 status_mask);
+
+void
+i915_disable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
+		      u32 status_mask);
+
+void valleyview_enable_display_irqs(struct drm_i915_private *dev_priv);
+void valleyview_disable_display_irqs(struct drm_i915_private *dev_priv);
+
+void intel_hpd_irq_setup(struct drm_i915_private *i915);
+void i915_hotplug_interrupt_update(struct drm_i915_private *dev_priv,
+				   u32 mask,
+				   u32 bits);
+
+void ilk_enable_display_irq(struct drm_i915_private *i915, u32 bits);
+void ilk_disable_display_irq(struct drm_i915_private *i915, u32 bits);
+
+void bdw_enable_pipe_irq(struct drm_i915_private *i915, enum pipe pipe, u32 bits);
+void bdw_disable_pipe_irq(struct drm_i915_private *i915, enum pipe pipe, u32 bits);
+
+void ibx_enable_display_interrupt(struct drm_i915_private *i915, u32 bits);
+void ibx_disable_display_interrupt(struct drm_i915_private *i915, u32 bits);
+
+void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, u32 mask);
+void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, u32 mask);
+void gen11_reset_rps_interrupts(struct drm_i915_private *dev_priv);
+void gen6_reset_rps_interrupts(struct drm_i915_private *dev_priv);
+void gen6_enable_rps_interrupts(struct drm_i915_private *dev_priv);
+void gen6_disable_rps_interrupts(struct drm_i915_private *dev_priv);
+void gen6_rps_reset_ei(struct drm_i915_private *dev_priv);
+u32 gen6_sanitize_rps_pm_mask(const struct drm_i915_private *i915, u32 mask);
+
+void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv);
+void intel_runtime_pm_enable_interrupts(struct drm_i915_private *dev_priv);
+bool intel_irqs_enabled(struct drm_i915_private *dev_priv);
+void intel_synchronize_irq(struct drm_i915_private *i915);
+void intel_synchronize_hardirq(struct drm_i915_private *i915);
+
+int intel_get_crtc_scanline(struct intel_crtc *crtc);
+void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv,
+				     u8 pipe_mask);
+void gen8_irq_power_well_pre_disable(struct drm_i915_private *dev_priv,
+				     u8 pipe_mask);
+u32 gen8_de_pipe_underrun_mask(struct drm_i915_private *dev_priv);
+
+bool intel_crtc_get_vblank_timestamp(struct drm_crtc *crtc, int *max_error,
+				     ktime_t *vblank_time, bool in_vblank_irq);
+
+u32 i915_get_vblank_counter(struct drm_crtc *crtc);
+u32 g4x_get_vblank_counter(struct drm_crtc *crtc);
+
+int i8xx_enable_vblank(struct drm_crtc *crtc);
+int i915gm_enable_vblank(struct drm_crtc *crtc);
+int i965_enable_vblank(struct drm_crtc *crtc);
+int ilk_enable_vblank(struct drm_crtc *crtc);
+int bdw_enable_vblank(struct drm_crtc *crtc);
+void i8xx_disable_vblank(struct drm_crtc *crtc);
+void i915gm_disable_vblank(struct drm_crtc *crtc);
+void i965_disable_vblank(struct drm_crtc *crtc);
+void ilk_disable_vblank(struct drm_crtc *crtc);
+void bdw_disable_vblank(struct drm_crtc *crtc);
+
+void gen2_irq_reset(struct intel_uncore *uncore);
+void gen3_irq_reset(struct intel_uncore *uncore, i915_reg_t imr,
+		    i915_reg_t iir, i915_reg_t ier);
+
+void gen2_irq_init(struct intel_uncore *uncore,
+		   u32 imr_val, u32 ier_val);
+void gen3_irq_init(struct intel_uncore *uncore,
+		   i915_reg_t imr, u32 imr_val,
+		   i915_reg_t ier, u32 ier_val,
+		   i915_reg_t iir);
+
+#define GEN8_IRQ_RESET_NDX(uncore, type, which) \
+({ \
+	unsigned int which_ = which; \
+	gen3_irq_reset((uncore), GEN8_##type##_IMR(which_), \
+		       GEN8_##type##_IIR(which_), GEN8_##type##_IER(which_)); \
+})
+
+#define GEN3_IRQ_RESET(uncore, type) \
+	gen3_irq_reset((uncore), type##IMR, type##IIR, type##IER)
+
+#define GEN2_IRQ_RESET(uncore) \
+	gen2_irq_reset(uncore)
+
+#define GEN8_IRQ_INIT_NDX(uncore, type, which, imr_val, ier_val) \
+({ \
+	unsigned int which_ = which; \
+	gen3_irq_init((uncore), \
+		      GEN8_##type##_IMR(which_), imr_val, \
+		      GEN8_##type##_IER(which_), ier_val, \
+		      GEN8_##type##_IIR(which_)); \
+})
+
+#define GEN3_IRQ_INIT(uncore, type, imr_val, ier_val) \
+	gen3_irq_init((uncore), \
+		      type##IMR, imr_val, \
+		      type##IER, ier_val, \
+		      type##IIR)
+
+#define GEN2_IRQ_INIT(uncore, imr_val, ier_val) \
+	gen2_irq_init((uncore), imr_val, ier_val)
+
+#endif /* __I915_IRQ_H__ */
diff --git a/drivers/gpu/drm/i915/i915_memcpy.c b/drivers/gpu/drm/i915/i915_memcpy.c
new file mode 100644
index 000000000..1b021a490
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_memcpy.c
@@ -0,0 +1,168 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/kernel.h>
+#include <asm/fpu/api.h>
+
+#include "i915_memcpy.h"
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG)
+#define CI_BUG_ON(expr) BUG_ON(expr)
+#else
+#define CI_BUG_ON(expr) BUILD_BUG_ON_INVALID(expr)
+#endif
+
+static DEFINE_STATIC_KEY_FALSE(has_movntdqa);
+
+static void __memcpy_ntdqa(void *dst, const void *src, unsigned long len)
+{
+	kernel_fpu_begin();
+
+	while (len >= 4) {
+		asm("movntdqa   (%0), %%xmm0\n"
+		    "movntdqa 16(%0), %%xmm1\n"
+		    "movntdqa 32(%0), %%xmm2\n"
+		    "movntdqa 48(%0), %%xmm3\n"
+		    "movaps %%xmm0,   (%1)\n"
+		    "movaps %%xmm1, 16(%1)\n"
+		    "movaps %%xmm2, 32(%1)\n"
+		    "movaps %%xmm3, 48(%1)\n"
+		    :: "r" (src), "r" (dst) : "memory");
+		src += 64;
+		dst += 64;
+		len -= 4;
+	}
+	while (len--) {
+		asm("movntdqa (%0), %%xmm0\n"
+		    "movaps %%xmm0, (%1)\n"
+		    :: "r" (src), "r" (dst) : "memory");
+		src += 16;
+		dst += 16;
+	}
+
+	kernel_fpu_end();
+}
+
+static void __memcpy_ntdqu(void *dst, const void *src, unsigned long len)
+{
+	kernel_fpu_begin();
+
+	while (len >= 4) {
+		asm("movntdqa   (%0), %%xmm0\n"
+		    "movntdqa 16(%0), %%xmm1\n"
+		    "movntdqa 32(%0), %%xmm2\n"
+		    "movntdqa 48(%0), %%xmm3\n"
+		    "movups %%xmm0,   (%1)\n"
+		    "movups %%xmm1, 16(%1)\n"
+		    "movups %%xmm2, 32(%1)\n"
+		    "movups %%xmm3, 48(%1)\n"
+		    :: "r" (src), "r" (dst) : "memory");
+		src += 64;
+		dst += 64;
+		len -= 4;
+	}
+	while (len--) {
+		asm("movntdqa (%0), %%xmm0\n"
+		    "movups %%xmm0, (%1)\n"
+		    :: "r" (src), "r" (dst) : "memory");
+		src += 16;
+		dst += 16;
+	}
+
+	kernel_fpu_end();
+}
+
+/**
+ * i915_memcpy_from_wc: perform an accelerated *aligned* read from WC
+ * @dst: destination pointer
+ * @src: source pointer
+ * @len: how many bytes to copy
+ *
+ * i915_memcpy_from_wc copies @len bytes from @src to @dst using
+ * non-temporal instructions where available. Note that all arguments
+ * (@src, @dst) must be aligned to 16 bytes and @len must be a multiple
+ * of 16.
+ *
+ * To test whether accelerated reads from WC are supported, use
+ * i915_memcpy_from_wc(NULL, NULL, 0);
+ *
+ * Returns true if the copy was successful, false if the preconditions
+ * are not met.
+ */
+bool i915_memcpy_from_wc(void *dst, const void *src, unsigned long len)
+{
+	if (unlikely(((unsigned long)dst | (unsigned long)src | len) & 15))
+		return false;
+
+	if (static_branch_likely(&has_movntdqa)) {
+		if (likely(len))
+			__memcpy_ntdqa(dst, src, len >> 4);
+		return true;
+	}
+
+	return false;
+}
+
+/**
+ * i915_unaligned_memcpy_from_wc: perform a mostly accelerated read from WC
+ * @dst: destination pointer
+ * @src: source pointer
+ * @len: how many bytes to copy
+ *
+ * Like i915_memcpy_from_wc(), the unaligned variant copies @len bytes from
+ * @src to @dst using * non-temporal instructions where available, but
+ * accepts that its arguments may not be aligned, but are valid for the
+ * potential 16-byte read past the end.
+ */
+void i915_unaligned_memcpy_from_wc(void *dst, const void *src, unsigned long len)
+{
+	unsigned long addr;
+
+	CI_BUG_ON(!i915_has_memcpy_from_wc());
+
+	addr = (unsigned long)src;
+	if (!IS_ALIGNED(addr, 16)) {
+		unsigned long x = min(ALIGN(addr, 16) - addr, len);
+
+		memcpy(dst, src, x);
+
+		len -= x;
+		dst += x;
+		src += x;
+	}
+
+	if (likely(len))
+		__memcpy_ntdqu(dst, src, DIV_ROUND_UP(len, 16));
+}
+
+void i915_memcpy_init_early(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * Some hypervisors (e.g. KVM) don't support VEX-prefix instructions
+	 * emulation. So don't enable movntdqa in hypervisor guest.
+	 */
+	if (static_cpu_has(X86_FEATURE_XMM4_1) &&
+	    !boot_cpu_has(X86_FEATURE_HYPERVISOR))
+		static_branch_enable(&has_movntdqa);
+}
diff --git a/drivers/gpu/drm/i915/i915_memcpy.h b/drivers/gpu/drm/i915/i915_memcpy.h
new file mode 100644
index 000000000..3df063a32
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_memcpy.h
@@ -0,0 +1,34 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_MEMCPY_H__
+#define __I915_MEMCPY_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+
+void i915_memcpy_init_early(struct drm_i915_private *i915);
+
+bool i915_memcpy_from_wc(void *dst, const void *src, unsigned long len);
+void i915_unaligned_memcpy_from_wc(void *dst, const void *src, unsigned long len);
+
+/* The movntdqa instructions used for memcpy-from-wc require 16-byte alignment,
+ * as well as SSE4.1 support. i915_memcpy_from_wc() will report if it cannot
+ * perform the operation. To check beforehand, pass in the parameters to
+ * to i915_can_memcpy_from_wc() - since we only care about the low 4 bits,
+ * you only need to pass in the minor offsets, page-aligned pointers are
+ * always valid.
+ *
+ * For just checking for SSE4.1, in the foreknowledge that the future use
+ * will be correctly aligned, just use i915_has_memcpy_from_wc().
+ */
+#define i915_can_memcpy_from_wc(dst, src, len) \
+	i915_memcpy_from_wc((void *)((unsigned long)(dst) | (unsigned long)(src) | (len)), NULL, 0)
+
+#define i915_has_memcpy_from_wc() \
+	i915_memcpy_from_wc(NULL, NULL, 0)
+
+#endif /* __I915_MEMCPY_H__ */
diff --git a/drivers/gpu/drm/i915/i915_mitigations.c b/drivers/gpu/drm/i915/i915_mitigations.c
new file mode 100644
index 000000000..def7302ef
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_mitigations.c
@@ -0,0 +1,147 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/kernel.h>
+#include <linux/moduleparam.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+
+#include "i915_driver.h"
+#include "i915_drv.h"
+#include "i915_mitigations.h"
+
+static unsigned long mitigations __read_mostly = ~0UL;
+
+enum {
+	CLEAR_RESIDUALS = 0,
+};
+
+static const char * const names[] = {
+	[CLEAR_RESIDUALS] = "residuals",
+};
+
+bool i915_mitigate_clear_residuals(void)
+{
+	return READ_ONCE(mitigations) & BIT(CLEAR_RESIDUALS);
+}
+
+static int mitigations_set(const char *val, const struct kernel_param *kp)
+{
+	unsigned long new = ~0UL;
+	char *str, *sep, *tok;
+	bool first = true;
+	int err = 0;
+
+	BUILD_BUG_ON(ARRAY_SIZE(names) >= BITS_PER_TYPE(mitigations));
+
+	str = kstrdup(val, GFP_KERNEL);
+	if (!str)
+		return -ENOMEM;
+
+	for (sep = str; (tok = strsep(&sep, ","));) {
+		bool enable = true;
+		int i;
+
+		/* Be tolerant of leading/trailing whitespace */
+		tok = strim(tok);
+
+		if (first) {
+			first = false;
+
+			if (!strcmp(tok, "auto"))
+				continue;
+
+			new = 0;
+			if (!strcmp(tok, "off"))
+				continue;
+		}
+
+		if (*tok == '!') {
+			enable = !enable;
+			tok++;
+		}
+
+		if (!strncmp(tok, "no", 2)) {
+			enable = !enable;
+			tok += 2;
+		}
+
+		if (*tok == '\0')
+			continue;
+
+		for (i = 0; i < ARRAY_SIZE(names); i++) {
+			if (!strcmp(tok, names[i])) {
+				if (enable)
+					new |= BIT(i);
+				else
+					new &= ~BIT(i);
+				break;
+			}
+		}
+		if (i == ARRAY_SIZE(names)) {
+			pr_err("Bad \"%s.mitigations=%s\", '%s' is unknown\n",
+			       DRIVER_NAME, val, tok);
+			err = -EINVAL;
+			break;
+		}
+	}
+	kfree(str);
+	if (err)
+		return err;
+
+	WRITE_ONCE(mitigations, new);
+	return 0;
+}
+
+static int mitigations_get(char *buffer, const struct kernel_param *kp)
+{
+	unsigned long local = READ_ONCE(mitigations);
+	int count, i;
+	bool enable;
+
+	if (!local)
+		return scnprintf(buffer, PAGE_SIZE, "%s\n", "off");
+
+	if (local & BIT(BITS_PER_LONG - 1)) {
+		count = scnprintf(buffer, PAGE_SIZE, "%s,", "auto");
+		enable = false;
+	} else {
+		enable = true;
+		count = 0;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(names); i++) {
+		if ((local & BIT(i)) != enable)
+			continue;
+
+		count += scnprintf(buffer + count, PAGE_SIZE - count,
+				   "%s%s,", enable ? "" : "!", names[i]);
+	}
+
+	buffer[count - 1] = '\n';
+	return count;
+}
+
+static const struct kernel_param_ops ops = {
+	.set = mitigations_set,
+	.get = mitigations_get,
+};
+
+module_param_cb_unsafe(mitigations, &ops, NULL, 0600);
+MODULE_PARM_DESC(mitigations,
+"Selectively enable security mitigations for all Intel® GPUs in the system.\n"
+"\n"
+"  auto -- enables all mitigations required for the platform [default]\n"
+"  off  -- disables all mitigations\n"
+"\n"
+"Individual mitigations can be enabled by passing a comma-separated string,\n"
+"e.g. mitigations=residuals to enable only clearing residuals or\n"
+"mitigations=auto,noresiduals to disable only the clear residual mitigation.\n"
+"Either '!' or 'no' may be used to switch from enabling the mitigation to\n"
+"disabling it.\n"
+"\n"
+"Active mitigations for Ivybridge, Baytrail, Haswell:\n"
+"  residuals -- clear all thread-local registers between contexts"
+);
diff --git a/drivers/gpu/drm/i915/i915_mitigations.h b/drivers/gpu/drm/i915/i915_mitigations.h
new file mode 100644
index 000000000..1359d8135
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_mitigations.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_MITIGATIONS_H__
+#define __I915_MITIGATIONS_H__
+
+#include <linux/types.h>
+
+bool i915_mitigate_clear_residuals(void);
+
+#endif /* __I915_MITIGATIONS_H__ */
diff --git a/drivers/gpu/drm/i915/i915_mm.c b/drivers/gpu/drm/i915/i915_mm.c
new file mode 100644
index 000000000..7998bc74a
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_mm.c
@@ -0,0 +1,154 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/mm.h>
+#include <linux/io-mapping.h>
+
+
+#include "i915_drv.h"
+#include "i915_mm.h"
+
+struct remap_pfn {
+	struct mm_struct *mm;
+	unsigned long pfn;
+	pgprot_t prot;
+
+	struct sgt_iter sgt;
+	resource_size_t iobase;
+};
+
+#define use_dma(io) ((io) != -1)
+
+static inline unsigned long sgt_pfn(const struct remap_pfn *r)
+{
+	if (use_dma(r->iobase))
+		return (r->sgt.dma + r->sgt.curr + r->iobase) >> PAGE_SHIFT;
+	else
+		return r->sgt.pfn + (r->sgt.curr >> PAGE_SHIFT);
+}
+
+static int remap_sg(pte_t *pte, unsigned long addr, void *data)
+{
+	struct remap_pfn *r = data;
+
+	if (GEM_WARN_ON(!r->sgt.sgp))
+		return -EINVAL;
+
+	/* Special PTE are not associated with any struct page */
+	set_pte_at(r->mm, addr, pte,
+		   pte_mkspecial(pfn_pte(sgt_pfn(r), r->prot)));
+	r->pfn++; /* track insertions in case we need to unwind later */
+
+	r->sgt.curr += PAGE_SIZE;
+	if (r->sgt.curr >= r->sgt.max)
+		r->sgt = __sgt_iter(__sg_next(r->sgt.sgp), use_dma(r->iobase));
+
+	return 0;
+}
+
+#define EXPECTED_FLAGS (VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP)
+
+#if IS_ENABLED(CONFIG_X86)
+static int remap_pfn(pte_t *pte, unsigned long addr, void *data)
+{
+	struct remap_pfn *r = data;
+
+	/* Special PTE are not associated with any struct page */
+	set_pte_at(r->mm, addr, pte, pte_mkspecial(pfn_pte(r->pfn, r->prot)));
+	r->pfn++;
+
+	return 0;
+}
+
+/**
+ * remap_io_mapping - remap an IO mapping to userspace
+ * @vma: user vma to map to
+ * @addr: target user address to start at
+ * @pfn: physical address of kernel memory
+ * @size: size of map area
+ * @iomap: the source io_mapping
+ *
+ *  Note: this is only safe if the mm semaphore is held when called.
+ */
+int remap_io_mapping(struct vm_area_struct *vma,
+		     unsigned long addr, unsigned long pfn, unsigned long size,
+		     struct io_mapping *iomap)
+{
+	struct remap_pfn r;
+	int err;
+
+	GEM_BUG_ON((vma->vm_flags & EXPECTED_FLAGS) != EXPECTED_FLAGS);
+
+	/* We rely on prevalidation of the io-mapping to skip track_pfn(). */
+	r.mm = vma->vm_mm;
+	r.pfn = pfn;
+	r.prot = __pgprot((pgprot_val(iomap->prot) & _PAGE_CACHE_MASK) |
+			  (pgprot_val(vma->vm_page_prot) & ~_PAGE_CACHE_MASK));
+
+	err = apply_to_page_range(r.mm, addr, size, remap_pfn, &r);
+	if (unlikely(err)) {
+		zap_vma_ptes(vma, addr, (r.pfn - pfn) << PAGE_SHIFT);
+		return err;
+	}
+
+	return 0;
+}
+#endif
+
+/**
+ * remap_io_sg - remap an IO mapping to userspace
+ * @vma: user vma to map to
+ * @addr: target user address to start at
+ * @size: size of map area
+ * @sgl: Start sg entry
+ * @iobase: Use stored dma address offset by this address or pfn if -1
+ *
+ *  Note: this is only safe if the mm semaphore is held when called.
+ */
+int remap_io_sg(struct vm_area_struct *vma,
+		unsigned long addr, unsigned long size,
+		struct scatterlist *sgl, resource_size_t iobase)
+{
+	struct remap_pfn r = {
+		.mm = vma->vm_mm,
+		.prot = vma->vm_page_prot,
+		.sgt = __sgt_iter(sgl, use_dma(iobase)),
+		.iobase = iobase,
+	};
+	int err;
+
+	/* We rely on prevalidation of the io-mapping to skip track_pfn(). */
+	GEM_BUG_ON((vma->vm_flags & EXPECTED_FLAGS) != EXPECTED_FLAGS);
+
+	if (!use_dma(iobase))
+		flush_cache_range(vma, addr, size);
+
+	err = apply_to_page_range(r.mm, addr, size, remap_sg, &r);
+	if (unlikely(err)) {
+		zap_vma_ptes(vma, addr, r.pfn << PAGE_SHIFT);
+		return err;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_mm.h b/drivers/gpu/drm/i915/i915_mm.h
new file mode 100644
index 000000000..04c8974d8
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_mm.h
@@ -0,0 +1,35 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_MM_H__
+#define __I915_MM_H__
+
+#include <linux/bug.h>
+#include <linux/types.h>
+
+struct vm_area_struct;
+struct io_mapping;
+struct scatterlist;
+
+#if IS_ENABLED(CONFIG_X86)
+int remap_io_mapping(struct vm_area_struct *vma,
+		     unsigned long addr, unsigned long pfn, unsigned long size,
+		     struct io_mapping *iomap);
+#else
+static inline
+int remap_io_mapping(struct vm_area_struct *vma,
+		     unsigned long addr, unsigned long pfn, unsigned long size,
+		     struct io_mapping *iomap)
+{
+	WARN_ONCE(1, "Architecture has no drm_cache.c support\n");
+	return 0;
+}
+#endif
+
+int remap_io_sg(struct vm_area_struct *vma,
+		unsigned long addr, unsigned long size,
+		struct scatterlist *sgl, resource_size_t iobase);
+
+#endif /* __I915_MM_H__ */
diff --git a/drivers/gpu/drm/i915/i915_module.c b/drivers/gpu/drm/i915/i915_module.c
new file mode 100644
index 000000000..65acd7bf7
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_module.c
@@ -0,0 +1,125 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <drm/drm_drv.h>
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_object.h"
+#include "i915_active.h"
+#include "i915_driver.h"
+#include "i915_params.h"
+#include "i915_pci.h"
+#include "i915_perf.h"
+#include "i915_request.h"
+#include "i915_scheduler.h"
+#include "i915_selftest.h"
+#include "i915_vma.h"
+#include "i915_vma_resource.h"
+
+static int i915_check_nomodeset(void)
+{
+	bool use_kms = true;
+
+	/*
+	 * Enable KMS by default, unless explicitly overriden by
+	 * either the i915.modeset parameter or by the
+	 * nomodeset boot option.
+	 */
+
+	if (i915_modparams.modeset == 0)
+		use_kms = false;
+
+	if (drm_firmware_drivers_only() && i915_modparams.modeset == -1)
+		use_kms = false;
+
+	if (!use_kms) {
+		/* Silently fail loading to not upset userspace. */
+		DRM_DEBUG_DRIVER("KMS disabled.\n");
+		return 1;
+	}
+
+	return 0;
+}
+
+static const struct {
+   int (*init)(void);
+   void (*exit)(void);
+} init_funcs[] = {
+	{ .init = i915_check_nomodeset },
+	{ .init = i915_active_module_init,
+	  .exit = i915_active_module_exit },
+	{ .init = i915_context_module_init,
+	  .exit = i915_context_module_exit },
+	{ .init = i915_gem_context_module_init,
+	  .exit = i915_gem_context_module_exit },
+	{ .init = i915_objects_module_init,
+	  .exit = i915_objects_module_exit },
+	{ .init = i915_request_module_init,
+	  .exit = i915_request_module_exit },
+	{ .init = i915_scheduler_module_init,
+	  .exit = i915_scheduler_module_exit },
+	{ .init = i915_vma_module_init,
+	  .exit = i915_vma_module_exit },
+	{ .init = i915_vma_resource_module_init,
+	  .exit = i915_vma_resource_module_exit },
+	{ .init = i915_mock_selftests },
+	{ .init = i915_pmu_init,
+	  .exit = i915_pmu_exit },
+	{ .init = i915_pci_register_driver,
+	  .exit = i915_pci_unregister_driver },
+	{ .init = i915_perf_sysctl_register,
+	  .exit = i915_perf_sysctl_unregister },
+};
+static int init_progress;
+
+static int __init i915_init(void)
+{
+	int err, i;
+
+	for (i = 0; i < ARRAY_SIZE(init_funcs); i++) {
+		err = init_funcs[i].init();
+		if (err < 0) {
+			while (i--) {
+				if (init_funcs[i].exit)
+					init_funcs[i].exit();
+			}
+			return err;
+		} else if (err > 0) {
+			/*
+			 * Early-exit success is reserved for things which
+			 * don't have an exit() function because we have no
+			 * idea how far they got or how to partially tear
+			 * them down.
+			 */
+			WARN_ON(init_funcs[i].exit);
+			break;
+		}
+	}
+
+	init_progress = i;
+
+	return 0;
+}
+
+static void __exit i915_exit(void)
+{
+	int i;
+
+	for (i = init_progress - 1; i >= 0; i--) {
+		GEM_BUG_ON(i >= ARRAY_SIZE(init_funcs));
+		if (init_funcs[i].exit)
+			init_funcs[i].exit();
+	}
+}
+
+module_init(i915_init);
+module_exit(i915_exit);
+
+MODULE_AUTHOR("Tungsten Graphics, Inc.");
+MODULE_AUTHOR("Intel Corporation");
+
+MODULE_DESCRIPTION(DRIVER_DESC);
+MODULE_LICENSE("GPL and additional rights");
diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c
new file mode 100644
index 000000000..5b24dd50f
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_params.c
@@ -0,0 +1,285 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/string_helpers.h>
+
+#include <drm/drm_print.h>
+
+#include "i915_params.h"
+#include "i915_drv.h"
+
+DECLARE_DYNDBG_CLASSMAP(drm_debug_classes, DD_CLASS_TYPE_DISJOINT_BITS, 0,
+			"DRM_UT_CORE",
+			"DRM_UT_DRIVER",
+			"DRM_UT_KMS",
+			"DRM_UT_PRIME",
+			"DRM_UT_ATOMIC",
+			"DRM_UT_VBL",
+			"DRM_UT_STATE",
+			"DRM_UT_LEASE",
+			"DRM_UT_DP",
+			"DRM_UT_DRMRES");
+
+#define i915_param_named(name, T, perm, desc) \
+	module_param_named(name, i915_modparams.name, T, perm); \
+	MODULE_PARM_DESC(name, desc)
+#define i915_param_named_unsafe(name, T, perm, desc) \
+	module_param_named_unsafe(name, i915_modparams.name, T, perm); \
+	MODULE_PARM_DESC(name, desc)
+
+struct i915_params i915_modparams __read_mostly = {
+#define MEMBER(T, member, value, ...) .member = (value),
+	I915_PARAMS_FOR_EACH(MEMBER)
+#undef MEMBER
+};
+
+/*
+ * Note: As a rule, keep module parameter sysfs permissions read-only
+ * 0400. Runtime changes are only supported through i915 debugfs.
+ *
+ * For any exceptions requiring write access and runtime changes through module
+ * parameter sysfs, prevent debugfs file creation by setting the parameter's
+ * debugfs mode to 0.
+ */
+
+i915_param_named(modeset, int, 0400,
+	"Use kernel modesetting [KMS] (0=disable, "
+	"1=on, -1=force vga console preference [default])");
+
+i915_param_named_unsafe(enable_dc, int, 0400,
+	"Enable power-saving display C-states. "
+	"(-1=auto [default]; 0=disable; 1=up to DC5; 2=up to DC6; "
+	"3=up to DC5 with DC3CO; 4=up to DC6 with DC3CO)");
+
+i915_param_named_unsafe(enable_fbc, int, 0400,
+	"Enable frame buffer compression for power savings "
+	"(default: -1 (use per-chip default))");
+
+i915_param_named_unsafe(lvds_channel_mode, int, 0400,
+	 "Specify LVDS channel mode "
+	 "(0=probe BIOS [default], 1=single-channel, 2=dual-channel)");
+
+i915_param_named_unsafe(panel_use_ssc, int, 0400,
+	"Use Spread Spectrum Clock with panels [LVDS/eDP] "
+	"(default: auto from VBT)");
+
+i915_param_named_unsafe(vbt_sdvo_panel_type, int, 0400,
+	"Override/Ignore selection of SDVO panel mode in the VBT "
+	"(-2=ignore, -1=auto [default], index in VBT BIOS table)");
+
+i915_param_named_unsafe(reset, uint, 0400,
+	"Attempt GPU resets (0=disabled, 1=full gpu reset, 2=engine reset [default])");
+
+i915_param_named_unsafe(vbt_firmware, charp, 0400,
+	"Load VBT from specified file under /lib/firmware");
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+i915_param_named(error_capture, bool, 0400,
+	"Record the GPU state following a hang. "
+	"This information in /sys/class/drm/card<N>/error is vital for "
+	"triaging and debugging hangs.");
+#endif
+
+i915_param_named_unsafe(enable_hangcheck, bool, 0400,
+	"Periodically check GPU activity for detecting hangs. "
+	"WARNING: Disabling this can cause system wide hangs. "
+	"(default: true)");
+
+i915_param_named_unsafe(enable_psr, int, 0400,
+	"Enable PSR "
+	"(0=disabled, 1=enable up to PSR1, 2=enable up to PSR2) "
+	"Default: -1 (use per-chip default)");
+
+i915_param_named(psr_safest_params, bool, 0400,
+	"Replace PSR VBT parameters by the safest and not optimal ones. This "
+	"is helpful to detect if PSR issues are related to bad values set in "
+	" VBT. (0=use VBT parameters, 1=use safest parameters)");
+
+i915_param_named_unsafe(enable_psr2_sel_fetch, bool, 0400,
+	"Enable PSR2 selective fetch "
+	"(0=disabled, 1=enabled) "
+	"Default: 0");
+
+i915_param_named_unsafe(force_probe, charp, 0400,
+	"Force probe options for specified supported devices. "
+	"See CONFIG_DRM_I915_FORCE_PROBE for details.");
+
+i915_param_named_unsafe(disable_power_well, int, 0400,
+	"Disable display power wells when possible "
+	"(-1=auto [default], 0=power wells always on, 1=power wells disabled when possible)");
+
+i915_param_named_unsafe(enable_ips, int, 0400, "Enable IPS (default: true)");
+
+i915_param_named(fastboot, int, 0400,
+	"Try to skip unnecessary mode sets at boot time "
+	"(0=disabled, 1=enabled) "
+	"Default: -1 (use per-chip default)");
+
+i915_param_named_unsafe(load_detect_test, bool, 0400,
+	"Force-enable the VGA load detect code for testing (default:false). "
+	"For developers only.");
+
+i915_param_named_unsafe(force_reset_modeset_test, bool, 0400,
+	"Force a modeset during gpu reset for testing (default:false). "
+	"For developers only.");
+
+i915_param_named_unsafe(invert_brightness, int, 0400,
+	"Invert backlight brightness "
+	"(-1 force normal, 0 machine defaults, 1 force inversion), please "
+	"report PCI device ID, subsystem vendor and subsystem device ID "
+	"to dri-devel@lists.freedesktop.org, if your machine needs it. "
+	"It will then be included in an upcoming module version.");
+
+i915_param_named(disable_display, bool, 0400,
+	"Disable display (default: false)");
+
+i915_param_named(memtest, bool, 0400,
+	"Perform a read/write test of all device memory on module load (default: off)");
+
+i915_param_named(mmio_debug, int, 0400,
+	"Enable the MMIO debug code for the first N failures (default: off). "
+	"This may negatively affect performance.");
+
+/* Special case writable file */
+i915_param_named(verbose_state_checks, bool, 0600,
+	"Enable verbose logs (ie. WARN_ON()) in case of unexpected hw state conditions.");
+
+i915_param_named_unsafe(nuclear_pageflip, bool, 0400,
+	"Force enable atomic functionality on platforms that don't have full support yet.");
+
+/* WA to get away with the default setting in VBT for early platforms.Will be removed */
+i915_param_named_unsafe(edp_vswing, int, 0400,
+	"Ignore/Override vswing pre-emph table selection from VBT "
+	"(0=use value from vbt [default], 1=low power swing(200mV),"
+	"2=default swing(400mV))");
+
+i915_param_named_unsafe(enable_guc, int, 0400,
+	"Enable GuC load for GuC submission and/or HuC load. "
+	"Required functionality can be selected using bitmask values. "
+	"(-1=auto [default], 0=disable, 1=GuC submission, 2=HuC load)");
+
+i915_param_named(guc_log_level, int, 0400,
+	"GuC firmware logging level. Requires GuC to be loaded. "
+	"(-1=auto [default], 0=disable, 1..4=enable with verbosity min..max)");
+
+i915_param_named_unsafe(guc_firmware_path, charp, 0400,
+	"GuC firmware path to use instead of the default one");
+
+i915_param_named_unsafe(huc_firmware_path, charp, 0400,
+	"HuC firmware path to use instead of the default one");
+
+i915_param_named_unsafe(dmc_firmware_path, charp, 0400,
+	"DMC firmware path to use instead of the default one");
+
+i915_param_named_unsafe(enable_dp_mst, bool, 0400,
+	"Enable multi-stream transport (MST) for new DisplayPort sinks. (default: true)");
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG)
+i915_param_named_unsafe(inject_probe_failure, uint, 0400,
+	"Force an error after a number of failure check points (0:disabled (default), N:force failure at the Nth failure check point)");
+#endif
+
+i915_param_named(enable_dpcd_backlight, int, 0400,
+	"Enable support for DPCD backlight control"
+	"(-1=use per-VBT LFP backlight type setting [default], 0=disabled, 1=enable, 2=force VESA interface, 3=force Intel interface)");
+
+#if IS_ENABLED(CONFIG_DRM_I915_GVT)
+i915_param_named(enable_gvt, bool, 0400,
+	"Enable support for Intel GVT-g graphics virtualization host support(default:false)");
+#endif
+
+#if CONFIG_DRM_I915_REQUEST_TIMEOUT
+i915_param_named_unsafe(request_timeout_ms, uint, 0600,
+			"Default request/fence/batch buffer expiration timeout.");
+#endif
+
+i915_param_named_unsafe(lmem_size, uint, 0400,
+			"Set the lmem size(in MiB) for each region. (default: 0, all memory)");
+i915_param_named_unsafe(lmem_bar_size, uint, 0400,
+			"Set the lmem bar size(in MiB).");
+
+static __always_inline void _print_param(struct drm_printer *p,
+					 const char *name,
+					 const char *type,
+					 const void *x)
+{
+	if (!__builtin_strcmp(type, "bool"))
+		drm_printf(p, "i915.%s=%s\n", name,
+			   str_yes_no(*(const bool *)x));
+	else if (!__builtin_strcmp(type, "int"))
+		drm_printf(p, "i915.%s=%d\n", name, *(const int *)x);
+	else if (!__builtin_strcmp(type, "unsigned int"))
+		drm_printf(p, "i915.%s=%u\n", name, *(const unsigned int *)x);
+	else if (!__builtin_strcmp(type, "unsigned long"))
+		drm_printf(p, "i915.%s=%lu\n", name, *(const unsigned long *)x);
+	else if (!__builtin_strcmp(type, "char *"))
+		drm_printf(p, "i915.%s=%s\n", name, *(const char **)x);
+	else
+		WARN_ONCE(1, "no printer defined for param type %s (i915.%s)\n",
+			  type, name);
+}
+
+/**
+ * i915_params_dump - dump i915 modparams
+ * @params: i915 modparams
+ * @p: the &drm_printer
+ *
+ * Pretty printer for i915 modparams.
+ */
+void i915_params_dump(const struct i915_params *params, struct drm_printer *p)
+{
+#define PRINT(T, x, ...) _print_param(p, #x, #T, &params->x);
+	I915_PARAMS_FOR_EACH(PRINT);
+#undef PRINT
+}
+
+static __always_inline void dup_param(const char *type, void *x)
+{
+	if (!__builtin_strcmp(type, "char *"))
+		*(void **)x = kstrdup(*(void **)x, GFP_ATOMIC);
+}
+
+void i915_params_copy(struct i915_params *dest, const struct i915_params *src)
+{
+	*dest = *src;
+#define DUP(T, x, ...) dup_param(#T, &dest->x);
+	I915_PARAMS_FOR_EACH(DUP);
+#undef DUP
+}
+
+static __always_inline void free_param(const char *type, void *x)
+{
+	if (!__builtin_strcmp(type, "char *")) {
+		kfree(*(void **)x);
+		*(void **)x = NULL;
+	}
+}
+
+/* free the allocated members, *not* the passed in params itself */
+void i915_params_free(struct i915_params *params)
+{
+#define FREE(T, x, ...) free_param(#T, &params->x);
+	I915_PARAMS_FOR_EACH(FREE);
+#undef FREE
+}
diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h
new file mode 100644
index 000000000..2733cb6cf
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_params.h
@@ -0,0 +1,101 @@
+/*
+ * Copyright © 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _I915_PARAMS_H_
+#define _I915_PARAMS_H_
+
+#include <linux/bitops.h>
+#include <linux/cache.h> /* for __read_mostly */
+
+struct drm_printer;
+
+#define ENABLE_GUC_SUBMISSION		BIT(0)
+#define ENABLE_GUC_LOAD_HUC		BIT(1)
+#define ENABLE_GUC_MASK			GENMASK(1, 0)
+
+/*
+ * Invoke param, a function-like macro, for each i915 param, with arguments:
+ *
+ * param(type, name, value, mode)
+ *
+ * type: parameter type, one of {bool, int, unsigned int, unsigned long, char *}
+ * name: name of the parameter
+ * value: initial/default value of the parameter
+ * mode: debugfs file permissions, one of {0400, 0600, 0}, use 0 to not create
+ *       debugfs file
+ */
+#define I915_PARAMS_FOR_EACH(param) \
+	param(char *, vbt_firmware, NULL, 0400) \
+	param(int, modeset, -1, 0400) \
+	param(int, lvds_channel_mode, 0, 0400) \
+	param(int, panel_use_ssc, -1, 0600) \
+	param(int, vbt_sdvo_panel_type, -1, 0400) \
+	param(int, enable_dc, -1, 0400) \
+	param(int, enable_fbc, -1, 0600) \
+	param(int, enable_psr, -1, 0600) \
+	param(bool, psr_safest_params, false, 0400) \
+	param(bool, enable_psr2_sel_fetch, true, 0400) \
+	param(int, disable_power_well, -1, 0400) \
+	param(int, enable_ips, 1, 0600) \
+	param(int, invert_brightness, 0, 0600) \
+	param(int, enable_guc, -1, 0400) \
+	param(int, guc_log_level, -1, 0400) \
+	param(char *, guc_firmware_path, NULL, 0400) \
+	param(char *, huc_firmware_path, NULL, 0400) \
+	param(char *, dmc_firmware_path, NULL, 0400) \
+	param(bool, memtest, false, 0400) \
+	param(int, mmio_debug, -IS_ENABLED(CONFIG_DRM_I915_DEBUG_MMIO), 0600) \
+	param(int, edp_vswing, 0, 0400) \
+	param(unsigned int, reset, 3, 0600) \
+	param(unsigned int, inject_probe_failure, 0, 0) \
+	param(int, fastboot, -1, 0600) \
+	param(int, enable_dpcd_backlight, -1, 0600) \
+	param(char *, force_probe, CONFIG_DRM_I915_FORCE_PROBE, 0400) \
+	param(unsigned int, request_timeout_ms, CONFIG_DRM_I915_REQUEST_TIMEOUT, CONFIG_DRM_I915_REQUEST_TIMEOUT ? 0600 : 0) \
+	param(unsigned int, lmem_size, 0, 0400) \
+	param(unsigned int, lmem_bar_size, 0, 0400) \
+	/* leave bools at the end to not create holes */ \
+	param(bool, enable_hangcheck, true, 0600) \
+	param(bool, load_detect_test, false, 0600) \
+	param(bool, force_reset_modeset_test, false, 0600) \
+	param(bool, error_capture, true, IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) ? 0600 : 0) \
+	param(bool, disable_display, false, 0400) \
+	param(bool, verbose_state_checks, true, 0) \
+	param(bool, nuclear_pageflip, false, 0400) \
+	param(bool, enable_dp_mst, true, 0600) \
+	param(bool, enable_gvt, false, IS_ENABLED(CONFIG_DRM_I915_GVT) ? 0400 : 0)
+
+#define MEMBER(T, member, ...) T member;
+struct i915_params {
+	I915_PARAMS_FOR_EACH(MEMBER);
+};
+#undef MEMBER
+
+extern struct i915_params i915_modparams __read_mostly;
+
+void i915_params_dump(const struct i915_params *params, struct drm_printer *p);
+void i915_params_copy(struct i915_params *dest, const struct i915_params *src);
+void i915_params_free(struct i915_params *params);
+
+#endif
diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c
new file mode 100644
index 000000000..efa80475f
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_pci.c
@@ -0,0 +1,1415 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <drm/drm_color_mgmt.h>
+#include <drm/drm_drv.h>
+#include <drm/i915_pciids.h>
+
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_sa_media.h"
+
+#include "i915_driver.h"
+#include "i915_drv.h"
+#include "i915_pci.h"
+#include "i915_reg.h"
+#include "intel_pci_config.h"
+
+#define PLATFORM(x) .platform = (x)
+#define GEN(x) \
+	.__runtime.graphics.ip.ver = (x), \
+	.__runtime.media.ip.ver = (x), \
+	.__runtime.display.ip.ver = (x)
+
+#define NO_DISPLAY .__runtime.pipe_mask = 0
+
+#define I845_PIPE_OFFSETS \
+	.display.pipe_offsets = { \
+		[TRANSCODER_A] = PIPE_A_OFFSET,	\
+	}, \
+	.display.trans_offsets = { \
+		[TRANSCODER_A] = TRANSCODER_A_OFFSET, \
+	}
+
+#define I9XX_PIPE_OFFSETS \
+	.display.pipe_offsets = { \
+		[TRANSCODER_A] = PIPE_A_OFFSET,	\
+		[TRANSCODER_B] = PIPE_B_OFFSET, \
+	}, \
+	.display.trans_offsets = { \
+		[TRANSCODER_A] = TRANSCODER_A_OFFSET, \
+		[TRANSCODER_B] = TRANSCODER_B_OFFSET, \
+	}
+
+#define IVB_PIPE_OFFSETS \
+	.display.pipe_offsets = { \
+		[TRANSCODER_A] = PIPE_A_OFFSET,	\
+		[TRANSCODER_B] = PIPE_B_OFFSET, \
+		[TRANSCODER_C] = PIPE_C_OFFSET, \
+	}, \
+	.display.trans_offsets = { \
+		[TRANSCODER_A] = TRANSCODER_A_OFFSET, \
+		[TRANSCODER_B] = TRANSCODER_B_OFFSET, \
+		[TRANSCODER_C] = TRANSCODER_C_OFFSET, \
+	}
+
+#define HSW_PIPE_OFFSETS \
+	.display.pipe_offsets = { \
+		[TRANSCODER_A] = PIPE_A_OFFSET,	\
+		[TRANSCODER_B] = PIPE_B_OFFSET, \
+		[TRANSCODER_C] = PIPE_C_OFFSET, \
+		[TRANSCODER_EDP] = PIPE_EDP_OFFSET, \
+	}, \
+	.display.trans_offsets = { \
+		[TRANSCODER_A] = TRANSCODER_A_OFFSET, \
+		[TRANSCODER_B] = TRANSCODER_B_OFFSET, \
+		[TRANSCODER_C] = TRANSCODER_C_OFFSET, \
+		[TRANSCODER_EDP] = TRANSCODER_EDP_OFFSET, \
+	}
+
+#define CHV_PIPE_OFFSETS \
+	.display.pipe_offsets = { \
+		[TRANSCODER_A] = PIPE_A_OFFSET, \
+		[TRANSCODER_B] = PIPE_B_OFFSET, \
+		[TRANSCODER_C] = CHV_PIPE_C_OFFSET, \
+	}, \
+	.display.trans_offsets = { \
+		[TRANSCODER_A] = TRANSCODER_A_OFFSET, \
+		[TRANSCODER_B] = TRANSCODER_B_OFFSET, \
+		[TRANSCODER_C] = CHV_TRANSCODER_C_OFFSET, \
+	}
+
+#define I845_CURSOR_OFFSETS \
+	.display.cursor_offsets = { \
+		[PIPE_A] = CURSOR_A_OFFSET, \
+	}
+
+#define I9XX_CURSOR_OFFSETS \
+	.display.cursor_offsets = { \
+		[PIPE_A] = CURSOR_A_OFFSET, \
+		[PIPE_B] = CURSOR_B_OFFSET, \
+	}
+
+#define CHV_CURSOR_OFFSETS \
+	.display.cursor_offsets = { \
+		[PIPE_A] = CURSOR_A_OFFSET, \
+		[PIPE_B] = CURSOR_B_OFFSET, \
+		[PIPE_C] = CHV_CURSOR_C_OFFSET, \
+	}
+
+#define IVB_CURSOR_OFFSETS \
+	.display.cursor_offsets = { \
+		[PIPE_A] = CURSOR_A_OFFSET, \
+		[PIPE_B] = IVB_CURSOR_B_OFFSET, \
+		[PIPE_C] = IVB_CURSOR_C_OFFSET, \
+	}
+
+#define TGL_CURSOR_OFFSETS \
+	.display.cursor_offsets = { \
+		[PIPE_A] = CURSOR_A_OFFSET, \
+		[PIPE_B] = IVB_CURSOR_B_OFFSET, \
+		[PIPE_C] = IVB_CURSOR_C_OFFSET, \
+		[PIPE_D] = TGL_CURSOR_D_OFFSET, \
+	}
+
+#define I9XX_COLORS \
+	.display.color = { .gamma_lut_size = 256 }
+#define I965_COLORS \
+	.display.color = { .gamma_lut_size = 129, \
+		   .gamma_lut_tests = DRM_COLOR_LUT_NON_DECREASING, \
+	}
+#define ILK_COLORS \
+	.display.color = { .gamma_lut_size = 1024 }
+#define IVB_COLORS \
+	.display.color = { .degamma_lut_size = 1024, .gamma_lut_size = 1024 }
+#define CHV_COLORS \
+	.display.color = { \
+		.degamma_lut_size = 65, .gamma_lut_size = 257, \
+		.degamma_lut_tests = DRM_COLOR_LUT_NON_DECREASING, \
+		.gamma_lut_tests = DRM_COLOR_LUT_NON_DECREASING, \
+	}
+#define GLK_COLORS \
+	.display.color = { \
+		.degamma_lut_size = 33, .gamma_lut_size = 1024, \
+		.degamma_lut_tests = DRM_COLOR_LUT_NON_DECREASING | \
+				     DRM_COLOR_LUT_EQUAL_CHANNELS, \
+	}
+#define ICL_COLORS \
+	.display.color = { \
+		.degamma_lut_size = 33, .gamma_lut_size = 262145, \
+		.degamma_lut_tests = DRM_COLOR_LUT_NON_DECREASING | \
+				     DRM_COLOR_LUT_EQUAL_CHANNELS, \
+		.gamma_lut_tests = DRM_COLOR_LUT_NON_DECREASING, \
+	}
+
+/* Keep in gen based order, and chronological order within a gen */
+
+#define GEN_DEFAULT_PAGE_SIZES \
+	.__runtime.page_sizes = I915_GTT_PAGE_SIZE_4K
+
+#define GEN_DEFAULT_REGIONS \
+	.__runtime.memory_regions = REGION_SMEM | REGION_STOLEN_SMEM
+
+#define I830_FEATURES \
+	GEN(2), \
+	.is_mobile = 1, \
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B), \
+	.display.has_overlay = 1, \
+	.display.cursor_needs_physical = 1, \
+	.display.overlay_needs_physical = 1, \
+	.display.has_gmch = 1, \
+	.gpu_reset_clobbers_display = true, \
+	.has_3d_pipeline = 1, \
+	.hws_needs_physical = 1, \
+	.unfenced_needs_alignment = 1, \
+	.__runtime.platform_engine_mask = BIT(RCS0), \
+	.has_snoop = true, \
+	.has_coherent_ggtt = false, \
+	.dma_mask_size = 32, \
+	I9XX_PIPE_OFFSETS, \
+	I9XX_CURSOR_OFFSETS, \
+	I9XX_COLORS, \
+	GEN_DEFAULT_PAGE_SIZES, \
+	GEN_DEFAULT_REGIONS
+
+#define I845_FEATURES \
+	GEN(2), \
+	.__runtime.pipe_mask = BIT(PIPE_A), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A), \
+	.display.has_overlay = 1, \
+	.display.overlay_needs_physical = 1, \
+	.display.has_gmch = 1, \
+	.has_3d_pipeline = 1, \
+	.gpu_reset_clobbers_display = true, \
+	.hws_needs_physical = 1, \
+	.unfenced_needs_alignment = 1, \
+	.__runtime.platform_engine_mask = BIT(RCS0), \
+	.has_snoop = true, \
+	.has_coherent_ggtt = false, \
+	.dma_mask_size = 32, \
+	I845_PIPE_OFFSETS, \
+	I845_CURSOR_OFFSETS, \
+	I9XX_COLORS, \
+	GEN_DEFAULT_PAGE_SIZES, \
+	GEN_DEFAULT_REGIONS
+
+static const struct intel_device_info i830_info = {
+	I830_FEATURES,
+	PLATFORM(INTEL_I830),
+};
+
+static const struct intel_device_info i845g_info = {
+	I845_FEATURES,
+	PLATFORM(INTEL_I845G),
+};
+
+static const struct intel_device_info i85x_info = {
+	I830_FEATURES,
+	PLATFORM(INTEL_I85X),
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A),
+};
+
+static const struct intel_device_info i865g_info = {
+	I845_FEATURES,
+	PLATFORM(INTEL_I865G),
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A),
+};
+
+#define GEN3_FEATURES \
+	GEN(3), \
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B), \
+	.display.has_gmch = 1, \
+	.gpu_reset_clobbers_display = true, \
+	.__runtime.platform_engine_mask = BIT(RCS0), \
+	.has_3d_pipeline = 1, \
+	.has_snoop = true, \
+	.has_coherent_ggtt = true, \
+	.dma_mask_size = 32, \
+	I9XX_PIPE_OFFSETS, \
+	I9XX_CURSOR_OFFSETS, \
+	I9XX_COLORS, \
+	GEN_DEFAULT_PAGE_SIZES, \
+	GEN_DEFAULT_REGIONS
+
+static const struct intel_device_info i915g_info = {
+	GEN3_FEATURES,
+	PLATFORM(INTEL_I915G),
+	.has_coherent_ggtt = false,
+	.display.cursor_needs_physical = 1,
+	.display.has_overlay = 1,
+	.display.overlay_needs_physical = 1,
+	.hws_needs_physical = 1,
+	.unfenced_needs_alignment = 1,
+};
+
+static const struct intel_device_info i915gm_info = {
+	GEN3_FEATURES,
+	PLATFORM(INTEL_I915GM),
+	.is_mobile = 1,
+	.display.cursor_needs_physical = 1,
+	.display.has_overlay = 1,
+	.display.overlay_needs_physical = 1,
+	.display.supports_tv = 1,
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A),
+	.hws_needs_physical = 1,
+	.unfenced_needs_alignment = 1,
+};
+
+static const struct intel_device_info i945g_info = {
+	GEN3_FEATURES,
+	PLATFORM(INTEL_I945G),
+	.display.has_hotplug = 1,
+	.display.cursor_needs_physical = 1,
+	.display.has_overlay = 1,
+	.display.overlay_needs_physical = 1,
+	.hws_needs_physical = 1,
+	.unfenced_needs_alignment = 1,
+};
+
+static const struct intel_device_info i945gm_info = {
+	GEN3_FEATURES,
+	PLATFORM(INTEL_I945GM),
+	.is_mobile = 1,
+	.display.has_hotplug = 1,
+	.display.cursor_needs_physical = 1,
+	.display.has_overlay = 1,
+	.display.overlay_needs_physical = 1,
+	.display.supports_tv = 1,
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A),
+	.hws_needs_physical = 1,
+	.unfenced_needs_alignment = 1,
+};
+
+static const struct intel_device_info g33_info = {
+	GEN3_FEATURES,
+	PLATFORM(INTEL_G33),
+	.display.has_hotplug = 1,
+	.display.has_overlay = 1,
+	.dma_mask_size = 36,
+};
+
+static const struct intel_device_info pnv_g_info = {
+	GEN3_FEATURES,
+	PLATFORM(INTEL_PINEVIEW),
+	.display.has_hotplug = 1,
+	.display.has_overlay = 1,
+	.dma_mask_size = 36,
+};
+
+static const struct intel_device_info pnv_m_info = {
+	GEN3_FEATURES,
+	PLATFORM(INTEL_PINEVIEW),
+	.is_mobile = 1,
+	.display.has_hotplug = 1,
+	.display.has_overlay = 1,
+	.dma_mask_size = 36,
+};
+
+#define GEN4_FEATURES \
+	GEN(4), \
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B), \
+	.display.has_hotplug = 1, \
+	.display.has_gmch = 1, \
+	.gpu_reset_clobbers_display = true, \
+	.__runtime.platform_engine_mask = BIT(RCS0), \
+	.has_3d_pipeline = 1, \
+	.has_snoop = true, \
+	.has_coherent_ggtt = true, \
+	.dma_mask_size = 36, \
+	I9XX_PIPE_OFFSETS, \
+	I9XX_CURSOR_OFFSETS, \
+	I965_COLORS, \
+	GEN_DEFAULT_PAGE_SIZES, \
+	GEN_DEFAULT_REGIONS
+
+static const struct intel_device_info i965g_info = {
+	GEN4_FEATURES,
+	PLATFORM(INTEL_I965G),
+	.display.has_overlay = 1,
+	.hws_needs_physical = 1,
+	.has_snoop = false,
+};
+
+static const struct intel_device_info i965gm_info = {
+	GEN4_FEATURES,
+	PLATFORM(INTEL_I965GM),
+	.is_mobile = 1,
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A),
+	.display.has_overlay = 1,
+	.display.supports_tv = 1,
+	.hws_needs_physical = 1,
+	.has_snoop = false,
+};
+
+static const struct intel_device_info g45_info = {
+	GEN4_FEATURES,
+	PLATFORM(INTEL_G45),
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(VCS0),
+	.gpu_reset_clobbers_display = false,
+};
+
+static const struct intel_device_info gm45_info = {
+	GEN4_FEATURES,
+	PLATFORM(INTEL_GM45),
+	.is_mobile = 1,
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A),
+	.display.supports_tv = 1,
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(VCS0),
+	.gpu_reset_clobbers_display = false,
+};
+
+#define GEN5_FEATURES \
+	GEN(5), \
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B), \
+	.display.has_hotplug = 1, \
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(VCS0), \
+	.has_3d_pipeline = 1, \
+	.has_snoop = true, \
+	.has_coherent_ggtt = true, \
+	/* ilk does support rc6, but we do not implement [power] contexts */ \
+	.has_rc6 = 0, \
+	.dma_mask_size = 36, \
+	I9XX_PIPE_OFFSETS, \
+	I9XX_CURSOR_OFFSETS, \
+	ILK_COLORS, \
+	GEN_DEFAULT_PAGE_SIZES, \
+	GEN_DEFAULT_REGIONS
+
+static const struct intel_device_info ilk_d_info = {
+	GEN5_FEATURES,
+	PLATFORM(INTEL_IRONLAKE),
+};
+
+static const struct intel_device_info ilk_m_info = {
+	GEN5_FEATURES,
+	PLATFORM(INTEL_IRONLAKE),
+	.is_mobile = 1,
+	.has_rps = true,
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A),
+};
+
+#define GEN6_FEATURES \
+	GEN(6), \
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B), \
+	.display.has_hotplug = 1, \
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A), \
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(VCS0) | BIT(BCS0), \
+	.has_3d_pipeline = 1, \
+	.has_coherent_ggtt = true, \
+	.has_llc = 1, \
+	.has_rc6 = 1, \
+	/* snb does support rc6p, but enabling it causes various issues */ \
+	.has_rc6p = 0, \
+	.has_rps = true, \
+	.dma_mask_size = 40, \
+	.__runtime.ppgtt_type = INTEL_PPGTT_ALIASING, \
+	.__runtime.ppgtt_size = 31, \
+	I9XX_PIPE_OFFSETS, \
+	I9XX_CURSOR_OFFSETS, \
+	ILK_COLORS, \
+	GEN_DEFAULT_PAGE_SIZES, \
+	GEN_DEFAULT_REGIONS
+
+#define SNB_D_PLATFORM \
+	GEN6_FEATURES, \
+	PLATFORM(INTEL_SANDYBRIDGE)
+
+static const struct intel_device_info snb_d_gt1_info = {
+	SNB_D_PLATFORM,
+	.gt = 1,
+};
+
+static const struct intel_device_info snb_d_gt2_info = {
+	SNB_D_PLATFORM,
+	.gt = 2,
+};
+
+#define SNB_M_PLATFORM \
+	GEN6_FEATURES, \
+	PLATFORM(INTEL_SANDYBRIDGE), \
+	.is_mobile = 1
+
+
+static const struct intel_device_info snb_m_gt1_info = {
+	SNB_M_PLATFORM,
+	.gt = 1,
+};
+
+static const struct intel_device_info snb_m_gt2_info = {
+	SNB_M_PLATFORM,
+	.gt = 2,
+};
+
+#define GEN7_FEATURES  \
+	GEN(7), \
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) | BIT(TRANSCODER_C), \
+	.display.has_hotplug = 1, \
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A), \
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(VCS0) | BIT(BCS0), \
+	.has_3d_pipeline = 1, \
+	.has_coherent_ggtt = true, \
+	.has_llc = 1, \
+	.has_rc6 = 1, \
+	.has_rc6p = 1, \
+	.has_reset_engine = true, \
+	.has_rps = true, \
+	.dma_mask_size = 40, \
+	.__runtime.ppgtt_type = INTEL_PPGTT_ALIASING, \
+	.__runtime.ppgtt_size = 31, \
+	IVB_PIPE_OFFSETS, \
+	IVB_CURSOR_OFFSETS, \
+	IVB_COLORS, \
+	GEN_DEFAULT_PAGE_SIZES, \
+	GEN_DEFAULT_REGIONS
+
+#define IVB_D_PLATFORM \
+	GEN7_FEATURES, \
+	PLATFORM(INTEL_IVYBRIDGE), \
+	.has_l3_dpf = 1
+
+static const struct intel_device_info ivb_d_gt1_info = {
+	IVB_D_PLATFORM,
+	.gt = 1,
+};
+
+static const struct intel_device_info ivb_d_gt2_info = {
+	IVB_D_PLATFORM,
+	.gt = 2,
+};
+
+#define IVB_M_PLATFORM \
+	GEN7_FEATURES, \
+	PLATFORM(INTEL_IVYBRIDGE), \
+	.is_mobile = 1, \
+	.has_l3_dpf = 1
+
+static const struct intel_device_info ivb_m_gt1_info = {
+	IVB_M_PLATFORM,
+	.gt = 1,
+};
+
+static const struct intel_device_info ivb_m_gt2_info = {
+	IVB_M_PLATFORM,
+	.gt = 2,
+};
+
+static const struct intel_device_info ivb_q_info = {
+	GEN7_FEATURES,
+	PLATFORM(INTEL_IVYBRIDGE),
+	NO_DISPLAY,
+	.gt = 2,
+	.has_l3_dpf = 1,
+};
+
+static const struct intel_device_info vlv_info = {
+	PLATFORM(INTEL_VALLEYVIEW),
+	GEN(7),
+	.is_lp = 1,
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B),
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B),
+	.has_runtime_pm = 1,
+	.has_rc6 = 1,
+	.has_reset_engine = true,
+	.has_rps = true,
+	.display.has_gmch = 1,
+	.display.has_hotplug = 1,
+	.dma_mask_size = 40,
+	.__runtime.ppgtt_type = INTEL_PPGTT_ALIASING,
+	.__runtime.ppgtt_size = 31,
+	.has_snoop = true,
+	.has_coherent_ggtt = false,
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(VCS0) | BIT(BCS0),
+	.display.mmio_offset = VLV_DISPLAY_BASE,
+	I9XX_PIPE_OFFSETS,
+	I9XX_CURSOR_OFFSETS,
+	I965_COLORS,
+	GEN_DEFAULT_PAGE_SIZES,
+	GEN_DEFAULT_REGIONS,
+};
+
+#define G75_FEATURES  \
+	GEN7_FEATURES, \
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(VCS0) | BIT(BCS0) | BIT(VECS0), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) | \
+		BIT(TRANSCODER_C) | BIT(TRANSCODER_EDP), \
+	.display.has_ddi = 1, \
+	.display.has_fpga_dbg = 1, \
+	.display.has_dp_mst = 1, \
+	.has_rc6p = 0 /* RC6p removed-by HSW */, \
+	HSW_PIPE_OFFSETS, \
+	.has_runtime_pm = 1
+
+#define HSW_PLATFORM \
+	G75_FEATURES, \
+	PLATFORM(INTEL_HASWELL), \
+	.has_l3_dpf = 1
+
+static const struct intel_device_info hsw_gt1_info = {
+	HSW_PLATFORM,
+	.gt = 1,
+};
+
+static const struct intel_device_info hsw_gt2_info = {
+	HSW_PLATFORM,
+	.gt = 2,
+};
+
+static const struct intel_device_info hsw_gt3_info = {
+	HSW_PLATFORM,
+	.gt = 3,
+};
+
+#define GEN8_FEATURES \
+	G75_FEATURES, \
+	GEN(8), \
+	.has_logical_ring_contexts = 1, \
+	.dma_mask_size = 39, \
+	.__runtime.ppgtt_type = INTEL_PPGTT_FULL, \
+	.__runtime.ppgtt_size = 48, \
+	.has_64bit_reloc = 1
+
+#define BDW_PLATFORM \
+	GEN8_FEATURES, \
+	PLATFORM(INTEL_BROADWELL)
+
+static const struct intel_device_info bdw_gt1_info = {
+	BDW_PLATFORM,
+	.gt = 1,
+};
+
+static const struct intel_device_info bdw_gt2_info = {
+	BDW_PLATFORM,
+	.gt = 2,
+};
+
+static const struct intel_device_info bdw_rsvd_info = {
+	BDW_PLATFORM,
+	.gt = 3,
+	/* According to the device ID those devices are GT3, they were
+	 * previously treated as not GT3, keep it like that.
+	 */
+};
+
+static const struct intel_device_info bdw_gt3_info = {
+	BDW_PLATFORM,
+	.gt = 3,
+	.__runtime.platform_engine_mask =
+		BIT(RCS0) | BIT(VCS0) | BIT(BCS0) | BIT(VECS0) | BIT(VCS1),
+};
+
+static const struct intel_device_info chv_info = {
+	PLATFORM(INTEL_CHERRYVIEW),
+	GEN(8),
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C),
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) | BIT(TRANSCODER_C),
+	.display.has_hotplug = 1,
+	.is_lp = 1,
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(VCS0) | BIT(BCS0) | BIT(VECS0),
+	.has_64bit_reloc = 1,
+	.has_runtime_pm = 1,
+	.has_rc6 = 1,
+	.has_rps = true,
+	.has_logical_ring_contexts = 1,
+	.display.has_gmch = 1,
+	.dma_mask_size = 39,
+	.__runtime.ppgtt_type = INTEL_PPGTT_FULL,
+	.__runtime.ppgtt_size = 32,
+	.has_reset_engine = 1,
+	.has_snoop = true,
+	.has_coherent_ggtt = false,
+	.display.mmio_offset = VLV_DISPLAY_BASE,
+	CHV_PIPE_OFFSETS,
+	CHV_CURSOR_OFFSETS,
+	CHV_COLORS,
+	GEN_DEFAULT_PAGE_SIZES,
+	GEN_DEFAULT_REGIONS,
+};
+
+#define GEN9_DEFAULT_PAGE_SIZES \
+	.__runtime.page_sizes = I915_GTT_PAGE_SIZE_4K | \
+		I915_GTT_PAGE_SIZE_64K
+
+#define GEN9_FEATURES \
+	GEN8_FEATURES, \
+	GEN(9), \
+	GEN9_DEFAULT_PAGE_SIZES, \
+	.__runtime.has_dmc = 1, \
+	.has_gt_uc = 1, \
+	.__runtime.has_hdcp = 1, \
+	.display.has_ipc = 1, \
+	.display.has_psr = 1, \
+	.display.has_psr_hw_tracking = 1, \
+	.display.dbuf.size = 896 - 4, /* 4 blocks for bypass path allocation */ \
+	.display.dbuf.slice_mask = BIT(DBUF_S1)
+
+#define SKL_PLATFORM \
+	GEN9_FEATURES, \
+	PLATFORM(INTEL_SKYLAKE)
+
+static const struct intel_device_info skl_gt1_info = {
+	SKL_PLATFORM,
+	.gt = 1,
+};
+
+static const struct intel_device_info skl_gt2_info = {
+	SKL_PLATFORM,
+	.gt = 2,
+};
+
+#define SKL_GT3_PLUS_PLATFORM \
+	SKL_PLATFORM, \
+	.__runtime.platform_engine_mask = \
+		BIT(RCS0) | BIT(VCS0) | BIT(BCS0) | BIT(VECS0) | BIT(VCS1)
+
+
+static const struct intel_device_info skl_gt3_info = {
+	SKL_GT3_PLUS_PLATFORM,
+	.gt = 3,
+};
+
+static const struct intel_device_info skl_gt4_info = {
+	SKL_GT3_PLUS_PLATFORM,
+	.gt = 4,
+};
+
+#define GEN9_LP_FEATURES \
+	GEN(9), \
+	.is_lp = 1, \
+	.display.dbuf.slice_mask = BIT(DBUF_S1), \
+	.display.has_hotplug = 1, \
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(VCS0) | BIT(BCS0) | BIT(VECS0), \
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) | \
+		BIT(TRANSCODER_C) | BIT(TRANSCODER_EDP) | \
+		BIT(TRANSCODER_DSI_A) | BIT(TRANSCODER_DSI_C), \
+	.has_3d_pipeline = 1, \
+	.has_64bit_reloc = 1, \
+	.display.has_ddi = 1, \
+	.display.has_fpga_dbg = 1, \
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A), \
+	.__runtime.has_hdcp = 1, \
+	.display.has_psr = 1, \
+	.display.has_psr_hw_tracking = 1, \
+	.has_runtime_pm = 1, \
+	.__runtime.has_dmc = 1, \
+	.has_rc6 = 1, \
+	.has_rps = true, \
+	.display.has_dp_mst = 1, \
+	.has_logical_ring_contexts = 1, \
+	.has_gt_uc = 1, \
+	.dma_mask_size = 39, \
+	.__runtime.ppgtt_type = INTEL_PPGTT_FULL, \
+	.__runtime.ppgtt_size = 48, \
+	.has_reset_engine = 1, \
+	.has_snoop = true, \
+	.has_coherent_ggtt = false, \
+	.display.has_ipc = 1, \
+	HSW_PIPE_OFFSETS, \
+	IVB_CURSOR_OFFSETS, \
+	IVB_COLORS, \
+	GEN9_DEFAULT_PAGE_SIZES, \
+	GEN_DEFAULT_REGIONS
+
+static const struct intel_device_info bxt_info = {
+	GEN9_LP_FEATURES,
+	PLATFORM(INTEL_BROXTON),
+	.display.dbuf.size = 512 - 4, /* 4 blocks for bypass path allocation */
+};
+
+static const struct intel_device_info glk_info = {
+	GEN9_LP_FEATURES,
+	PLATFORM(INTEL_GEMINILAKE),
+	.__runtime.display.ip.ver = 10,
+	.display.dbuf.size = 1024 - 4, /* 4 blocks for bypass path allocation */
+	GLK_COLORS,
+};
+
+#define KBL_PLATFORM \
+	GEN9_FEATURES, \
+	PLATFORM(INTEL_KABYLAKE)
+
+static const struct intel_device_info kbl_gt1_info = {
+	KBL_PLATFORM,
+	.gt = 1,
+};
+
+static const struct intel_device_info kbl_gt2_info = {
+	KBL_PLATFORM,
+	.gt = 2,
+};
+
+static const struct intel_device_info kbl_gt3_info = {
+	KBL_PLATFORM,
+	.gt = 3,
+	.__runtime.platform_engine_mask =
+		BIT(RCS0) | BIT(VCS0) | BIT(BCS0) | BIT(VECS0) | BIT(VCS1),
+};
+
+#define CFL_PLATFORM \
+	GEN9_FEATURES, \
+	PLATFORM(INTEL_COFFEELAKE)
+
+static const struct intel_device_info cfl_gt1_info = {
+	CFL_PLATFORM,
+	.gt = 1,
+};
+
+static const struct intel_device_info cfl_gt2_info = {
+	CFL_PLATFORM,
+	.gt = 2,
+};
+
+static const struct intel_device_info cfl_gt3_info = {
+	CFL_PLATFORM,
+	.gt = 3,
+	.__runtime.platform_engine_mask =
+		BIT(RCS0) | BIT(VCS0) | BIT(BCS0) | BIT(VECS0) | BIT(VCS1),
+};
+
+#define CML_PLATFORM \
+	GEN9_FEATURES, \
+	PLATFORM(INTEL_COMETLAKE)
+
+static const struct intel_device_info cml_gt1_info = {
+	CML_PLATFORM,
+	.gt = 1,
+};
+
+static const struct intel_device_info cml_gt2_info = {
+	CML_PLATFORM,
+	.gt = 2,
+};
+
+#define GEN11_DEFAULT_PAGE_SIZES \
+	.__runtime.page_sizes = I915_GTT_PAGE_SIZE_4K | \
+		I915_GTT_PAGE_SIZE_64K |		\
+		I915_GTT_PAGE_SIZE_2M
+
+#define GEN11_FEATURES \
+	GEN9_FEATURES, \
+	GEN11_DEFAULT_PAGE_SIZES, \
+	.display.abox_mask = BIT(0), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) | \
+		BIT(TRANSCODER_C) | BIT(TRANSCODER_EDP) | \
+		BIT(TRANSCODER_DSI_0) | BIT(TRANSCODER_DSI_1), \
+	.display.pipe_offsets = { \
+		[TRANSCODER_A] = PIPE_A_OFFSET, \
+		[TRANSCODER_B] = PIPE_B_OFFSET, \
+		[TRANSCODER_C] = PIPE_C_OFFSET, \
+		[TRANSCODER_EDP] = PIPE_EDP_OFFSET, \
+		[TRANSCODER_DSI_0] = PIPE_DSI0_OFFSET, \
+		[TRANSCODER_DSI_1] = PIPE_DSI1_OFFSET, \
+	}, \
+	.display.trans_offsets = { \
+		[TRANSCODER_A] = TRANSCODER_A_OFFSET, \
+		[TRANSCODER_B] = TRANSCODER_B_OFFSET, \
+		[TRANSCODER_C] = TRANSCODER_C_OFFSET, \
+		[TRANSCODER_EDP] = TRANSCODER_EDP_OFFSET, \
+		[TRANSCODER_DSI_0] = TRANSCODER_DSI0_OFFSET, \
+		[TRANSCODER_DSI_1] = TRANSCODER_DSI1_OFFSET, \
+	}, \
+	GEN(11), \
+	ICL_COLORS, \
+	.display.dbuf.size = 2048, \
+	.display.dbuf.slice_mask = BIT(DBUF_S1) | BIT(DBUF_S2), \
+	.__runtime.has_dsc = 1, \
+	.has_coherent_ggtt = false, \
+	.has_logical_ring_elsq = 1
+
+static const struct intel_device_info icl_info = {
+	GEN11_FEATURES,
+	PLATFORM(INTEL_ICELAKE),
+	.__runtime.platform_engine_mask =
+		BIT(RCS0) | BIT(BCS0) | BIT(VECS0) | BIT(VCS0) | BIT(VCS2),
+};
+
+static const struct intel_device_info ehl_info = {
+	GEN11_FEATURES,
+	PLATFORM(INTEL_ELKHARTLAKE),
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(BCS0) | BIT(VCS0) | BIT(VECS0),
+	.__runtime.ppgtt_size = 36,
+};
+
+static const struct intel_device_info jsl_info = {
+	GEN11_FEATURES,
+	PLATFORM(INTEL_JASPERLAKE),
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(BCS0) | BIT(VCS0) | BIT(VECS0),
+	.__runtime.ppgtt_size = 36,
+};
+
+#define GEN12_FEATURES \
+	GEN11_FEATURES, \
+	GEN(12), \
+	.display.abox_mask = GENMASK(2, 1), \
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D), \
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) | \
+		BIT(TRANSCODER_C) | BIT(TRANSCODER_D) | \
+		BIT(TRANSCODER_DSI_0) | BIT(TRANSCODER_DSI_1), \
+	.display.pipe_offsets = { \
+		[TRANSCODER_A] = PIPE_A_OFFSET, \
+		[TRANSCODER_B] = PIPE_B_OFFSET, \
+		[TRANSCODER_C] = PIPE_C_OFFSET, \
+		[TRANSCODER_D] = PIPE_D_OFFSET, \
+		[TRANSCODER_DSI_0] = PIPE_DSI0_OFFSET, \
+		[TRANSCODER_DSI_1] = PIPE_DSI1_OFFSET, \
+	}, \
+	.display.trans_offsets = { \
+		[TRANSCODER_A] = TRANSCODER_A_OFFSET, \
+		[TRANSCODER_B] = TRANSCODER_B_OFFSET, \
+		[TRANSCODER_C] = TRANSCODER_C_OFFSET, \
+		[TRANSCODER_D] = TRANSCODER_D_OFFSET, \
+		[TRANSCODER_DSI_0] = TRANSCODER_DSI0_OFFSET, \
+		[TRANSCODER_DSI_1] = TRANSCODER_DSI1_OFFSET, \
+	}, \
+	TGL_CURSOR_OFFSETS, \
+	.has_global_mocs = 1, \
+	.has_pxp = 1, \
+	.display.has_dsb = 0 /* FIXME: LUT load is broken with DSB */
+
+static const struct intel_device_info tgl_info = {
+	GEN12_FEATURES,
+	PLATFORM(INTEL_TIGERLAKE),
+	.display.has_modular_fia = 1,
+	.__runtime.platform_engine_mask =
+		BIT(RCS0) | BIT(BCS0) | BIT(VECS0) | BIT(VCS0) | BIT(VCS2),
+};
+
+static const struct intel_device_info rkl_info = {
+	GEN12_FEATURES,
+	PLATFORM(INTEL_ROCKETLAKE),
+	.display.abox_mask = BIT(0),
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C),
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) |
+		BIT(TRANSCODER_C),
+	.display.has_hti = 1,
+	.display.has_psr_hw_tracking = 0,
+	.__runtime.platform_engine_mask =
+		BIT(RCS0) | BIT(BCS0) | BIT(VECS0) | BIT(VCS0),
+};
+
+#define DGFX_FEATURES \
+	.__runtime.memory_regions = REGION_SMEM | REGION_LMEM | REGION_STOLEN_LMEM, \
+	.has_llc = 0, \
+	.has_pxp = 0, \
+	.has_snoop = 1, \
+	.is_dgfx = 1, \
+	.has_heci_gscfi = 1
+
+static const struct intel_device_info dg1_info = {
+	GEN12_FEATURES,
+	DGFX_FEATURES,
+	.__runtime.graphics.ip.rel = 10,
+	PLATFORM(INTEL_DG1),
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D),
+	.require_force_probe = 1,
+	.__runtime.platform_engine_mask =
+		BIT(RCS0) | BIT(BCS0) | BIT(VECS0) |
+		BIT(VCS0) | BIT(VCS2),
+	/* Wa_16011227922 */
+	.__runtime.ppgtt_size = 47,
+};
+
+static const struct intel_device_info adl_s_info = {
+	GEN12_FEATURES,
+	PLATFORM(INTEL_ALDERLAKE_S),
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D),
+	.display.has_hti = 1,
+	.display.has_psr_hw_tracking = 0,
+	.__runtime.platform_engine_mask =
+		BIT(RCS0) | BIT(BCS0) | BIT(VECS0) | BIT(VCS0) | BIT(VCS2),
+	.dma_mask_size = 39,
+};
+
+#define XE_LPD_FEATURES \
+	.display.abox_mask = GENMASK(1, 0),					\
+	.display.color = {							\
+		.degamma_lut_size = 128, .gamma_lut_size = 1024,		\
+		.degamma_lut_tests = DRM_COLOR_LUT_NON_DECREASING |		\
+				     DRM_COLOR_LUT_EQUAL_CHANNELS,		\
+	},									\
+	.display.dbuf.size = 4096,						\
+	.display.dbuf.slice_mask = BIT(DBUF_S1) | BIT(DBUF_S2) | BIT(DBUF_S3) |	\
+		BIT(DBUF_S4),							\
+	.display.has_ddi = 1,							\
+	.__runtime.has_dmc = 1,							\
+	.display.has_dp_mst = 1,						\
+	.display.has_dsb = 1,							\
+	.__runtime.has_dsc = 1,							\
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A),					\
+	.display.has_fpga_dbg = 1,						\
+	.__runtime.has_hdcp = 1,						\
+	.display.has_hotplug = 1,						\
+	.display.has_ipc = 1,							\
+	.display.has_psr = 1,							\
+	.__runtime.display.ip.ver = 13,							\
+	.__runtime.pipe_mask = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D),	\
+	.display.pipe_offsets = {						\
+		[TRANSCODER_A] = PIPE_A_OFFSET,					\
+		[TRANSCODER_B] = PIPE_B_OFFSET,					\
+		[TRANSCODER_C] = PIPE_C_OFFSET,					\
+		[TRANSCODER_D] = PIPE_D_OFFSET,					\
+		[TRANSCODER_DSI_0] = PIPE_DSI0_OFFSET,				\
+		[TRANSCODER_DSI_1] = PIPE_DSI1_OFFSET,				\
+	},									\
+	.display.trans_offsets = {						\
+		[TRANSCODER_A] = TRANSCODER_A_OFFSET,				\
+		[TRANSCODER_B] = TRANSCODER_B_OFFSET,				\
+		[TRANSCODER_C] = TRANSCODER_C_OFFSET,				\
+		[TRANSCODER_D] = TRANSCODER_D_OFFSET,				\
+		[TRANSCODER_DSI_0] = TRANSCODER_DSI0_OFFSET,			\
+		[TRANSCODER_DSI_1] = TRANSCODER_DSI1_OFFSET,			\
+	},									\
+	TGL_CURSOR_OFFSETS
+
+static const struct intel_device_info adl_p_info = {
+	GEN12_FEATURES,
+	XE_LPD_FEATURES,
+	PLATFORM(INTEL_ALDERLAKE_P),
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) |
+			       BIT(TRANSCODER_C) | BIT(TRANSCODER_D) |
+			       BIT(TRANSCODER_DSI_0) | BIT(TRANSCODER_DSI_1),
+	.display.has_cdclk_crawl = 1,
+	.display.has_modular_fia = 1,
+	.display.has_psr_hw_tracking = 0,
+	.__runtime.platform_engine_mask =
+		BIT(RCS0) | BIT(BCS0) | BIT(VECS0) | BIT(VCS0) | BIT(VCS2),
+	.__runtime.ppgtt_size = 48,
+	.dma_mask_size = 39,
+};
+
+#undef GEN
+
+#define XE_HP_PAGE_SIZES \
+	.__runtime.page_sizes = I915_GTT_PAGE_SIZE_4K | \
+		I915_GTT_PAGE_SIZE_64K |		\
+		I915_GTT_PAGE_SIZE_2M
+
+#define XE_HP_FEATURES \
+	.__runtime.graphics.ip.ver = 12, \
+	.__runtime.graphics.ip.rel = 50, \
+	XE_HP_PAGE_SIZES, \
+	.dma_mask_size = 46, \
+	.has_3d_pipeline = 1, \
+	.has_64bit_reloc = 1, \
+	.has_flat_ccs = 1, \
+	.has_global_mocs = 1, \
+	.has_gt_uc = 1, \
+	.has_llc = 1, \
+	.has_logical_ring_contexts = 1, \
+	.has_logical_ring_elsq = 1, \
+	.has_mslice_steering = 1, \
+	.has_rc6 = 1, \
+	.has_reset_engine = 1, \
+	.has_rps = 1, \
+	.has_runtime_pm = 1, \
+	.__runtime.ppgtt_size = 48, \
+	.__runtime.ppgtt_type = INTEL_PPGTT_FULL
+
+#define XE_HPM_FEATURES \
+	.__runtime.media.ip.ver = 12, \
+	.__runtime.media.ip.rel = 50
+
+__maybe_unused
+static const struct intel_device_info xehpsdv_info = {
+	XE_HP_FEATURES,
+	XE_HPM_FEATURES,
+	DGFX_FEATURES,
+	PLATFORM(INTEL_XEHPSDV),
+	NO_DISPLAY,
+	.has_64k_pages = 1,
+	.needs_compact_pt = 1,
+	.has_media_ratio_mode = 1,
+	.__runtime.platform_engine_mask =
+		BIT(RCS0) | BIT(BCS0) |
+		BIT(VECS0) | BIT(VECS1) | BIT(VECS2) | BIT(VECS3) |
+		BIT(VCS0) | BIT(VCS1) | BIT(VCS2) | BIT(VCS3) |
+		BIT(VCS4) | BIT(VCS5) | BIT(VCS6) | BIT(VCS7) |
+		BIT(CCS0) | BIT(CCS1) | BIT(CCS2) | BIT(CCS3),
+	.require_force_probe = 1,
+};
+
+#define DG2_FEATURES \
+	XE_HP_FEATURES, \
+	XE_HPM_FEATURES, \
+	DGFX_FEATURES, \
+	.__runtime.graphics.ip.rel = 55, \
+	.__runtime.media.ip.rel = 55, \
+	PLATFORM(INTEL_DG2), \
+	.has_4tile = 1, \
+	.has_64k_pages = 1, \
+	.has_guc_deprivilege = 1, \
+	.has_heci_pxp = 1, \
+	.needs_compact_pt = 1, \
+	.has_media_ratio_mode = 1, \
+	.__runtime.platform_engine_mask = \
+		BIT(RCS0) | BIT(BCS0) | \
+		BIT(VECS0) | BIT(VECS1) | \
+		BIT(VCS0) | BIT(VCS2) | \
+		BIT(CCS0) | BIT(CCS1) | BIT(CCS2) | BIT(CCS3)
+
+static const struct intel_device_info dg2_info = {
+	DG2_FEATURES,
+	XE_LPD_FEATURES,
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) |
+			       BIT(TRANSCODER_C) | BIT(TRANSCODER_D),
+	.require_force_probe = 1,
+};
+
+static const struct intel_device_info ats_m_info = {
+	DG2_FEATURES,
+	NO_DISPLAY,
+	.require_force_probe = 1,
+	.tuning_thread_rr_after_dep = 1,
+};
+
+#define XE_HPC_FEATURES \
+	XE_HP_FEATURES, \
+	.dma_mask_size = 52, \
+	.has_3d_pipeline = 0, \
+	.has_guc_deprivilege = 1, \
+	.has_l3_ccs_read = 1, \
+	.has_mslice_steering = 0, \
+	.has_one_eu_per_fuse_bit = 1
+
+__maybe_unused
+static const struct intel_device_info pvc_info = {
+	XE_HPC_FEATURES,
+	XE_HPM_FEATURES,
+	DGFX_FEATURES,
+	.__runtime.graphics.ip.rel = 60,
+	.__runtime.media.ip.rel = 60,
+	PLATFORM(INTEL_PONTEVECCHIO),
+	NO_DISPLAY,
+	.has_flat_ccs = 0,
+	.__runtime.platform_engine_mask =
+		BIT(BCS0) |
+		BIT(VCS0) |
+		BIT(CCS0) | BIT(CCS1) | BIT(CCS2) | BIT(CCS3),
+	.require_force_probe = 1,
+};
+
+#define XE_LPDP_FEATURES	\
+	XE_LPD_FEATURES,	\
+	.__runtime.display.ip.ver = 14,	\
+	.display.has_cdclk_crawl = 1, \
+	.__runtime.fbc_mask = BIT(INTEL_FBC_A) | BIT(INTEL_FBC_B)
+
+static const struct intel_gt_definition xelpmp_extra_gt[] = {
+	{
+		.type = GT_MEDIA,
+		.name = "Standalone Media GT",
+		.gsi_offset = MTL_MEDIA_GSI_BASE,
+		.engine_mask = BIT(VECS0) | BIT(VCS0) | BIT(VCS2),
+	},
+	{}
+};
+
+static const struct intel_device_info mtl_info = {
+	XE_HP_FEATURES,
+	XE_LPDP_FEATURES,
+	.__runtime.cpu_transcoder_mask = BIT(TRANSCODER_A) | BIT(TRANSCODER_B) |
+			       BIT(TRANSCODER_C) | BIT(TRANSCODER_D),
+	/*
+	 * Real graphics IP version will be obtained from hardware GMD_ID
+	 * register.  Value provided here is just for sanity checking.
+	 */
+	.__runtime.graphics.ip.ver = 12,
+	.__runtime.graphics.ip.rel = 70,
+	.__runtime.media.ip.ver = 13,
+	PLATFORM(INTEL_METEORLAKE),
+	.display.has_modular_fia = 1,
+	.extra_gt_list = xelpmp_extra_gt,
+	.has_flat_ccs = 0,
+	.has_snoop = 1,
+	.__runtime.memory_regions = REGION_SMEM | REGION_STOLEN_LMEM,
+	.__runtime.platform_engine_mask = BIT(RCS0) | BIT(BCS0) | BIT(CCS0),
+	.require_force_probe = 1,
+};
+
+#undef PLATFORM
+
+/*
+ * Make sure any device matches here are from most specific to most
+ * general.  For example, since the Quanta match is based on the subsystem
+ * and subvendor IDs, we need it to come before the more general IVB
+ * PCI ID matches, otherwise we'll use the wrong info struct above.
+ */
+static const struct pci_device_id pciidlist[] = {
+	INTEL_I830_IDS(&i830_info),
+	INTEL_I845G_IDS(&i845g_info),
+	INTEL_I85X_IDS(&i85x_info),
+	INTEL_I865G_IDS(&i865g_info),
+	INTEL_I915G_IDS(&i915g_info),
+	INTEL_I915GM_IDS(&i915gm_info),
+	INTEL_I945G_IDS(&i945g_info),
+	INTEL_I945GM_IDS(&i945gm_info),
+	INTEL_I965G_IDS(&i965g_info),
+	INTEL_G33_IDS(&g33_info),
+	INTEL_I965GM_IDS(&i965gm_info),
+	INTEL_GM45_IDS(&gm45_info),
+	INTEL_G45_IDS(&g45_info),
+	INTEL_PINEVIEW_G_IDS(&pnv_g_info),
+	INTEL_PINEVIEW_M_IDS(&pnv_m_info),
+	INTEL_IRONLAKE_D_IDS(&ilk_d_info),
+	INTEL_IRONLAKE_M_IDS(&ilk_m_info),
+	INTEL_SNB_D_GT1_IDS(&snb_d_gt1_info),
+	INTEL_SNB_D_GT2_IDS(&snb_d_gt2_info),
+	INTEL_SNB_M_GT1_IDS(&snb_m_gt1_info),
+	INTEL_SNB_M_GT2_IDS(&snb_m_gt2_info),
+	INTEL_IVB_Q_IDS(&ivb_q_info), /* must be first IVB */
+	INTEL_IVB_M_GT1_IDS(&ivb_m_gt1_info),
+	INTEL_IVB_M_GT2_IDS(&ivb_m_gt2_info),
+	INTEL_IVB_D_GT1_IDS(&ivb_d_gt1_info),
+	INTEL_IVB_D_GT2_IDS(&ivb_d_gt2_info),
+	INTEL_HSW_GT1_IDS(&hsw_gt1_info),
+	INTEL_HSW_GT2_IDS(&hsw_gt2_info),
+	INTEL_HSW_GT3_IDS(&hsw_gt3_info),
+	INTEL_VLV_IDS(&vlv_info),
+	INTEL_BDW_GT1_IDS(&bdw_gt1_info),
+	INTEL_BDW_GT2_IDS(&bdw_gt2_info),
+	INTEL_BDW_GT3_IDS(&bdw_gt3_info),
+	INTEL_BDW_RSVD_IDS(&bdw_rsvd_info),
+	INTEL_CHV_IDS(&chv_info),
+	INTEL_SKL_GT1_IDS(&skl_gt1_info),
+	INTEL_SKL_GT2_IDS(&skl_gt2_info),
+	INTEL_SKL_GT3_IDS(&skl_gt3_info),
+	INTEL_SKL_GT4_IDS(&skl_gt4_info),
+	INTEL_BXT_IDS(&bxt_info),
+	INTEL_GLK_IDS(&glk_info),
+	INTEL_KBL_GT1_IDS(&kbl_gt1_info),
+	INTEL_KBL_GT2_IDS(&kbl_gt2_info),
+	INTEL_KBL_GT3_IDS(&kbl_gt3_info),
+	INTEL_KBL_GT4_IDS(&kbl_gt3_info),
+	INTEL_AML_KBL_GT2_IDS(&kbl_gt2_info),
+	INTEL_CFL_S_GT1_IDS(&cfl_gt1_info),
+	INTEL_CFL_S_GT2_IDS(&cfl_gt2_info),
+	INTEL_CFL_H_GT1_IDS(&cfl_gt1_info),
+	INTEL_CFL_H_GT2_IDS(&cfl_gt2_info),
+	INTEL_CFL_U_GT2_IDS(&cfl_gt2_info),
+	INTEL_CFL_U_GT3_IDS(&cfl_gt3_info),
+	INTEL_WHL_U_GT1_IDS(&cfl_gt1_info),
+	INTEL_WHL_U_GT2_IDS(&cfl_gt2_info),
+	INTEL_AML_CFL_GT2_IDS(&cfl_gt2_info),
+	INTEL_WHL_U_GT3_IDS(&cfl_gt3_info),
+	INTEL_CML_GT1_IDS(&cml_gt1_info),
+	INTEL_CML_GT2_IDS(&cml_gt2_info),
+	INTEL_CML_U_GT1_IDS(&cml_gt1_info),
+	INTEL_CML_U_GT2_IDS(&cml_gt2_info),
+	INTEL_ICL_11_IDS(&icl_info),
+	INTEL_EHL_IDS(&ehl_info),
+	INTEL_JSL_IDS(&jsl_info),
+	INTEL_TGL_12_IDS(&tgl_info),
+	INTEL_RKL_IDS(&rkl_info),
+	INTEL_ADLS_IDS(&adl_s_info),
+	INTEL_ADLP_IDS(&adl_p_info),
+	INTEL_ADLN_IDS(&adl_p_info),
+	INTEL_DG1_IDS(&dg1_info),
+	INTEL_RPLS_IDS(&adl_s_info),
+	INTEL_RPLP_IDS(&adl_p_info),
+	INTEL_DG2_IDS(&dg2_info),
+	INTEL_ATS_M_IDS(&ats_m_info),
+	INTEL_MTL_IDS(&mtl_info),
+	{0, 0, 0}
+};
+MODULE_DEVICE_TABLE(pci, pciidlist);
+
+static void i915_pci_remove(struct pci_dev *pdev)
+{
+	struct drm_i915_private *i915;
+
+	i915 = pci_get_drvdata(pdev);
+	if (!i915) /* driver load aborted, nothing to cleanup */
+		return;
+
+	i915_driver_remove(i915);
+	pci_set_drvdata(pdev, NULL);
+}
+
+/* is device_id present in comma separated list of ids */
+static bool device_id_in_list(u16 device_id, const char *devices, bool negative)
+{
+	char *s, *p, *tok;
+	bool ret;
+
+	if (!devices || !*devices)
+		return false;
+
+	/* match everything */
+	if (negative && strcmp(devices, "!*") == 0)
+		return true;
+	if (!negative && strcmp(devices, "*") == 0)
+		return true;
+
+	s = kstrdup(devices, GFP_KERNEL);
+	if (!s)
+		return false;
+
+	for (p = s, ret = false; (tok = strsep(&p, ",")) != NULL; ) {
+		u16 val;
+
+		if (negative && tok[0] == '!')
+			tok++;
+		else if ((negative && tok[0] != '!') ||
+			 (!negative && tok[0] == '!'))
+			continue;
+
+		if (kstrtou16(tok, 16, &val) == 0 && val == device_id) {
+			ret = true;
+			break;
+		}
+	}
+
+	kfree(s);
+
+	return ret;
+}
+
+static bool id_forced(u16 device_id)
+{
+	return device_id_in_list(device_id, i915_modparams.force_probe, false);
+}
+
+static bool id_blocked(u16 device_id)
+{
+	return device_id_in_list(device_id, i915_modparams.force_probe, true);
+}
+
+bool i915_pci_resource_valid(struct pci_dev *pdev, int bar)
+{
+	if (!pci_resource_flags(pdev, bar))
+		return false;
+
+	if (pci_resource_flags(pdev, bar) & IORESOURCE_UNSET)
+		return false;
+
+	if (!pci_resource_len(pdev, bar))
+		return false;
+
+	return true;
+}
+
+static bool intel_mmio_bar_valid(struct pci_dev *pdev, struct intel_device_info *intel_info)
+{
+	int gttmmaddr_bar = intel_info->__runtime.graphics.ip.ver == 2 ? GEN2_GTTMMADR_BAR : GTTMMADR_BAR;
+
+	return i915_pci_resource_valid(pdev, gttmmaddr_bar);
+}
+
+static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
+{
+	struct intel_device_info *intel_info =
+		(struct intel_device_info *) ent->driver_data;
+	int err;
+
+	if (intel_info->require_force_probe && !id_forced(pdev->device)) {
+		dev_info(&pdev->dev,
+			 "Your graphics device %04x is not properly supported by i915 in this\n"
+			 "kernel version. To force driver probe anyway, use i915.force_probe=%04x\n"
+			 "module parameter or CONFIG_DRM_I915_FORCE_PROBE=%04x configuration option,\n"
+			 "or (recommended) check for kernel updates.\n",
+			 pdev->device, pdev->device, pdev->device);
+		return -ENODEV;
+	}
+
+	if (id_blocked(pdev->device)) {
+		dev_info(&pdev->dev, "I915 probe blocked for Device ID %04x.\n",
+			 pdev->device);
+		return -ENODEV;
+	}
+
+	if (intel_info->require_force_probe) {
+		dev_info(&pdev->dev, "Force probing unsupported Device ID %04x, tainting kernel\n",
+			 pdev->device);
+		add_taint(TAINT_USER, LOCKDEP_STILL_OK);
+	}
+
+	/* Only bind to function 0 of the device. Early generations
+	 * used function 1 as a placeholder for multi-head. This causes
+	 * us confusion instead, especially on the systems where both
+	 * functions have the same PCI-ID!
+	 */
+	if (PCI_FUNC(pdev->devfn))
+		return -ENODEV;
+
+	if (!intel_mmio_bar_valid(pdev, intel_info))
+		return -ENXIO;
+
+	/* Detect if we need to wait for other drivers early on */
+	if (intel_modeset_probe_defer(pdev))
+		return -EPROBE_DEFER;
+
+	err = i915_driver_probe(pdev, ent);
+	if (err)
+		return err;
+
+	if (i915_inject_probe_failure(pci_get_drvdata(pdev))) {
+		i915_pci_remove(pdev);
+		return -ENODEV;
+	}
+
+	err = i915_live_selftests(pdev);
+	if (err) {
+		i915_pci_remove(pdev);
+		return err > 0 ? -ENOTTY : err;
+	}
+
+	err = i915_perf_selftests(pdev);
+	if (err) {
+		i915_pci_remove(pdev);
+		return err > 0 ? -ENOTTY : err;
+	}
+
+	return 0;
+}
+
+static void i915_pci_shutdown(struct pci_dev *pdev)
+{
+	struct drm_i915_private *i915 = pci_get_drvdata(pdev);
+
+	i915_driver_shutdown(i915);
+}
+
+static struct pci_driver i915_pci_driver = {
+	.name = DRIVER_NAME,
+	.id_table = pciidlist,
+	.probe = i915_pci_probe,
+	.remove = i915_pci_remove,
+	.shutdown = i915_pci_shutdown,
+	.driver.pm = &i915_pm_ops,
+};
+
+int i915_pci_register_driver(void)
+{
+	return pci_register_driver(&i915_pci_driver);
+}
+
+void i915_pci_unregister_driver(void)
+{
+	pci_unregister_driver(&i915_pci_driver);
+}
diff --git a/drivers/gpu/drm/i915/i915_pci.h b/drivers/gpu/drm/i915/i915_pci.h
new file mode 100644
index 000000000..8dfe19f9a
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_pci.h
@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_PCI_H__
+#define __I915_PCI_H__
+
+#include <linux/types.h>
+
+struct pci_dev;
+
+int i915_pci_register_driver(void);
+void i915_pci_unregister_driver(void);
+
+bool i915_pci_resource_valid(struct pci_dev *pdev, int bar);
+
+#endif /* __I915_PCI_H__ */
diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c
new file mode 100644
index 000000000..2d4f09813
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_perf.c
@@ -0,0 +1,4624 @@
+/*
+ * Copyright © 2015-2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *   Robert Bragg <robert@sixbynine.org>
+ */
+
+
+/**
+ * DOC: i915 Perf Overview
+ *
+ * Gen graphics supports a large number of performance counters that can help
+ * driver and application developers understand and optimize their use of the
+ * GPU.
+ *
+ * This i915 perf interface enables userspace to configure and open a file
+ * descriptor representing a stream of GPU metrics which can then be read() as
+ * a stream of sample records.
+ *
+ * The interface is particularly suited to exposing buffered metrics that are
+ * captured by DMA from the GPU, unsynchronized with and unrelated to the CPU.
+ *
+ * Streams representing a single context are accessible to applications with a
+ * corresponding drm file descriptor, such that OpenGL can use the interface
+ * without special privileges. Access to system-wide metrics requires root
+ * privileges by default, unless changed via the dev.i915.perf_event_paranoid
+ * sysctl option.
+ *
+ */
+
+/**
+ * DOC: i915 Perf History and Comparison with Core Perf
+ *
+ * The interface was initially inspired by the core Perf infrastructure but
+ * some notable differences are:
+ *
+ * i915 perf file descriptors represent a "stream" instead of an "event"; where
+ * a perf event primarily corresponds to a single 64bit value, while a stream
+ * might sample sets of tightly-coupled counters, depending on the
+ * configuration.  For example the Gen OA unit isn't designed to support
+ * orthogonal configurations of individual counters; it's configured for a set
+ * of related counters. Samples for an i915 perf stream capturing OA metrics
+ * will include a set of counter values packed in a compact HW specific format.
+ * The OA unit supports a number of different packing formats which can be
+ * selected by the user opening the stream. Perf has support for grouping
+ * events, but each event in the group is configured, validated and
+ * authenticated individually with separate system calls.
+ *
+ * i915 perf stream configurations are provided as an array of u64 (key,value)
+ * pairs, instead of a fixed struct with multiple miscellaneous config members,
+ * interleaved with event-type specific members.
+ *
+ * i915 perf doesn't support exposing metrics via an mmap'd circular buffer.
+ * The supported metrics are being written to memory by the GPU unsynchronized
+ * with the CPU, using HW specific packing formats for counter sets. Sometimes
+ * the constraints on HW configuration require reports to be filtered before it
+ * would be acceptable to expose them to unprivileged applications - to hide
+ * the metrics of other processes/contexts. For these use cases a read() based
+ * interface is a good fit, and provides an opportunity to filter data as it
+ * gets copied from the GPU mapped buffers to userspace buffers.
+ *
+ *
+ * Issues hit with first prototype based on Core Perf
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ *
+ * The first prototype of this driver was based on the core perf
+ * infrastructure, and while we did make that mostly work, with some changes to
+ * perf, we found we were breaking or working around too many assumptions baked
+ * into perf's currently cpu centric design.
+ *
+ * In the end we didn't see a clear benefit to making perf's implementation and
+ * interface more complex by changing design assumptions while we knew we still
+ * wouldn't be able to use any existing perf based userspace tools.
+ *
+ * Also considering the Gen specific nature of the Observability hardware and
+ * how userspace will sometimes need to combine i915 perf OA metrics with
+ * side-band OA data captured via MI_REPORT_PERF_COUNT commands; we're
+ * expecting the interface to be used by a platform specific userspace such as
+ * OpenGL or tools. This is to say; we aren't inherently missing out on having
+ * a standard vendor/architecture agnostic interface by not using perf.
+ *
+ *
+ * For posterity, in case we might re-visit trying to adapt core perf to be
+ * better suited to exposing i915 metrics these were the main pain points we
+ * hit:
+ *
+ * - The perf based OA PMU driver broke some significant design assumptions:
+ *
+ *   Existing perf pmus are used for profiling work on a cpu and we were
+ *   introducing the idea of _IS_DEVICE pmus with different security
+ *   implications, the need to fake cpu-related data (such as user/kernel
+ *   registers) to fit with perf's current design, and adding _DEVICE records
+ *   as a way to forward device-specific status records.
+ *
+ *   The OA unit writes reports of counters into a circular buffer, without
+ *   involvement from the CPU, making our PMU driver the first of a kind.
+ *
+ *   Given the way we were periodically forward data from the GPU-mapped, OA
+ *   buffer to perf's buffer, those bursts of sample writes looked to perf like
+ *   we were sampling too fast and so we had to subvert its throttling checks.
+ *
+ *   Perf supports groups of counters and allows those to be read via
+ *   transactions internally but transactions currently seem designed to be
+ *   explicitly initiated from the cpu (say in response to a userspace read())
+ *   and while we could pull a report out of the OA buffer we can't
+ *   trigger a report from the cpu on demand.
+ *
+ *   Related to being report based; the OA counters are configured in HW as a
+ *   set while perf generally expects counter configurations to be orthogonal.
+ *   Although counters can be associated with a group leader as they are
+ *   opened, there's no clear precedent for being able to provide group-wide
+ *   configuration attributes (for example we want to let userspace choose the
+ *   OA unit report format used to capture all counters in a set, or specify a
+ *   GPU context to filter metrics on). We avoided using perf's grouping
+ *   feature and forwarded OA reports to userspace via perf's 'raw' sample
+ *   field. This suited our userspace well considering how coupled the counters
+ *   are when dealing with normalizing. It would be inconvenient to split
+ *   counters up into separate events, only to require userspace to recombine
+ *   them. For Mesa it's also convenient to be forwarded raw, periodic reports
+ *   for combining with the side-band raw reports it captures using
+ *   MI_REPORT_PERF_COUNT commands.
+ *
+ *   - As a side note on perf's grouping feature; there was also some concern
+ *     that using PERF_FORMAT_GROUP as a way to pack together counter values
+ *     would quite drastically inflate our sample sizes, which would likely
+ *     lower the effective sampling resolutions we could use when the available
+ *     memory bandwidth is limited.
+ *
+ *     With the OA unit's report formats, counters are packed together as 32
+ *     or 40bit values, with the largest report size being 256 bytes.
+ *
+ *     PERF_FORMAT_GROUP values are 64bit, but there doesn't appear to be a
+ *     documented ordering to the values, implying PERF_FORMAT_ID must also be
+ *     used to add a 64bit ID before each value; giving 16 bytes per counter.
+ *
+ *   Related to counter orthogonality; we can't time share the OA unit, while
+ *   event scheduling is a central design idea within perf for allowing
+ *   userspace to open + enable more events than can be configured in HW at any
+ *   one time.  The OA unit is not designed to allow re-configuration while in
+ *   use. We can't reconfigure the OA unit without losing internal OA unit
+ *   state which we can't access explicitly to save and restore. Reconfiguring
+ *   the OA unit is also relatively slow, involving ~100 register writes. From
+ *   userspace Mesa also depends on a stable OA configuration when emitting
+ *   MI_REPORT_PERF_COUNT commands and importantly the OA unit can't be
+ *   disabled while there are outstanding MI_RPC commands lest we hang the
+ *   command streamer.
+ *
+ *   The contents of sample records aren't extensible by device drivers (i.e.
+ *   the sample_type bits). As an example; Sourab Gupta had been looking to
+ *   attach GPU timestamps to our OA samples. We were shoehorning OA reports
+ *   into sample records by using the 'raw' field, but it's tricky to pack more
+ *   than one thing into this field because events/core.c currently only lets a
+ *   pmu give a single raw data pointer plus len which will be copied into the
+ *   ring buffer. To include more than the OA report we'd have to copy the
+ *   report into an intermediate larger buffer. I'd been considering allowing a
+ *   vector of data+len values to be specified for copying the raw data, but
+ *   it felt like a kludge to being using the raw field for this purpose.
+ *
+ * - It felt like our perf based PMU was making some technical compromises
+ *   just for the sake of using perf:
+ *
+ *   perf_event_open() requires events to either relate to a pid or a specific
+ *   cpu core, while our device pmu related to neither.  Events opened with a
+ *   pid will be automatically enabled/disabled according to the scheduling of
+ *   that process - so not appropriate for us. When an event is related to a
+ *   cpu id, perf ensures pmu methods will be invoked via an inter process
+ *   interrupt on that core. To avoid invasive changes our userspace opened OA
+ *   perf events for a specific cpu. This was workable but it meant the
+ *   majority of the OA driver ran in atomic context, including all OA report
+ *   forwarding, which wasn't really necessary in our case and seems to make
+ *   our locking requirements somewhat complex as we handled the interaction
+ *   with the rest of the i915 driver.
+ */
+
+#include <linux/anon_inodes.h>
+#include <linux/sizes.h>
+#include <linux/uuid.h>
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_internal.h"
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_engine_user.h"
+#include "gt/intel_execlists_submission.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_clock_utils.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_lrc.h"
+#include "gt/intel_lrc_reg.h"
+#include "gt/intel_ring.h"
+
+#include "i915_drv.h"
+#include "i915_file_private.h"
+#include "i915_perf.h"
+#include "i915_perf_oa_regs.h"
+
+/* HW requires this to be a power of two, between 128k and 16M, though driver
+ * is currently generally designed assuming the largest 16M size is used such
+ * that the overflow cases are unlikely in normal operation.
+ */
+#define OA_BUFFER_SIZE		SZ_16M
+
+#define OA_TAKEN(tail, head)	((tail - head) & (OA_BUFFER_SIZE - 1))
+
+/**
+ * DOC: OA Tail Pointer Race
+ *
+ * There's a HW race condition between OA unit tail pointer register updates and
+ * writes to memory whereby the tail pointer can sometimes get ahead of what's
+ * been written out to the OA buffer so far (in terms of what's visible to the
+ * CPU).
+ *
+ * Although this can be observed explicitly while copying reports to userspace
+ * by checking for a zeroed report-id field in tail reports, we want to account
+ * for this earlier, as part of the oa_buffer_check_unlocked to avoid lots of
+ * redundant read() attempts.
+ *
+ * We workaround this issue in oa_buffer_check_unlocked() by reading the reports
+ * in the OA buffer, starting from the tail reported by the HW until we find a
+ * report with its first 2 dwords not 0 meaning its previous report is
+ * completely in memory and ready to be read. Those dwords are also set to 0
+ * once read and the whole buffer is cleared upon OA buffer initialization. The
+ * first dword is the reason for this report while the second is the timestamp,
+ * making the chances of having those 2 fields at 0 fairly unlikely. A more
+ * detailed explanation is available in oa_buffer_check_unlocked().
+ *
+ * Most of the implementation details for this workaround are in
+ * oa_buffer_check_unlocked() and _append_oa_reports()
+ *
+ * Note for posterity: previously the driver used to define an effective tail
+ * pointer that lagged the real pointer by a 'tail margin' measured in bytes
+ * derived from %OA_TAIL_MARGIN_NSEC and the configured sampling frequency.
+ * This was flawed considering that the OA unit may also automatically generate
+ * non-periodic reports (such as on context switch) or the OA unit may be
+ * enabled without any periodic sampling.
+ */
+#define OA_TAIL_MARGIN_NSEC	100000ULL
+#define INVALID_TAIL_PTR	0xffffffff
+
+/* The default frequency for checking whether the OA unit has written new
+ * reports to the circular OA buffer...
+ */
+#define DEFAULT_POLL_FREQUENCY_HZ 200
+#define DEFAULT_POLL_PERIOD_NS (NSEC_PER_SEC / DEFAULT_POLL_FREQUENCY_HZ)
+
+/* for sysctl proc_dointvec_minmax of dev.i915.perf_stream_paranoid */
+static u32 i915_perf_stream_paranoid = true;
+
+/* The maximum exponent the hardware accepts is 63 (essentially it selects one
+ * of the 64bit timestamp bits to trigger reports from) but there's currently
+ * no known use case for sampling as infrequently as once per 47 thousand years.
+ *
+ * Since the timestamps included in OA reports are only 32bits it seems
+ * reasonable to limit the OA exponent where it's still possible to account for
+ * overflow in OA report timestamps.
+ */
+#define OA_EXPONENT_MAX 31
+
+#define INVALID_CTX_ID 0xffffffff
+
+/* On Gen8+ automatically triggered OA reports include a 'reason' field... */
+#define OAREPORT_REASON_MASK           0x3f
+#define OAREPORT_REASON_MASK_EXTENDED  0x7f
+#define OAREPORT_REASON_SHIFT          19
+#define OAREPORT_REASON_TIMER          (1<<0)
+#define OAREPORT_REASON_CTX_SWITCH     (1<<3)
+#define OAREPORT_REASON_CLK_RATIO      (1<<5)
+
+
+/* For sysctl proc_dointvec_minmax of i915_oa_max_sample_rate
+ *
+ * The highest sampling frequency we can theoretically program the OA unit
+ * with is always half the timestamp frequency: E.g. 6.25Mhz for Haswell.
+ *
+ * Initialized just before we register the sysctl parameter.
+ */
+static int oa_sample_rate_hard_limit;
+
+/* Theoretically we can program the OA unit to sample every 160ns but don't
+ * allow that by default unless root...
+ *
+ * The default threshold of 100000Hz is based on perf's similar
+ * kernel.perf_event_max_sample_rate sysctl parameter.
+ */
+static u32 i915_oa_max_sample_rate = 100000;
+
+/* XXX: beware if future OA HW adds new report formats that the current
+ * code assumes all reports have a power-of-two size and ~(size - 1) can
+ * be used as a mask to align the OA tail pointer.
+ */
+static const struct i915_oa_format oa_formats[I915_OA_FORMAT_MAX] = {
+	[I915_OA_FORMAT_A13]	    = { 0, 64 },
+	[I915_OA_FORMAT_A29]	    = { 1, 128 },
+	[I915_OA_FORMAT_A13_B8_C8]  = { 2, 128 },
+	/* A29_B8_C8 Disallowed as 192 bytes doesn't factor into buffer size */
+	[I915_OA_FORMAT_B4_C8]	    = { 4, 64 },
+	[I915_OA_FORMAT_A45_B8_C8]  = { 5, 256 },
+	[I915_OA_FORMAT_B4_C8_A16]  = { 6, 128 },
+	[I915_OA_FORMAT_C4_B8]	    = { 7, 64 },
+	[I915_OA_FORMAT_A12]		    = { 0, 64 },
+	[I915_OA_FORMAT_A12_B8_C8]	    = { 2, 128 },
+	[I915_OA_FORMAT_A32u40_A4u32_B8_C8] = { 5, 256 },
+};
+
+#define SAMPLE_OA_REPORT      (1<<0)
+
+/**
+ * struct perf_open_properties - for validated properties given to open a stream
+ * @sample_flags: `DRM_I915_PERF_PROP_SAMPLE_*` properties are tracked as flags
+ * @single_context: Whether a single or all gpu contexts should be monitored
+ * @hold_preemption: Whether the preemption is disabled for the filtered
+ *                   context
+ * @ctx_handle: A gem ctx handle for use with @single_context
+ * @metrics_set: An ID for an OA unit metric set advertised via sysfs
+ * @oa_format: An OA unit HW report format
+ * @oa_periodic: Whether to enable periodic OA unit sampling
+ * @oa_period_exponent: The OA unit sampling period is derived from this
+ * @engine: The engine (typically rcs0) being monitored by the OA unit
+ * @has_sseu: Whether @sseu was specified by userspace
+ * @sseu: internal SSEU configuration computed either from the userspace
+ *        specified configuration in the opening parameters or a default value
+ *        (see get_default_sseu_config())
+ * @poll_oa_period: The period in nanoseconds at which the CPU will check for OA
+ * data availability
+ *
+ * As read_properties_unlocked() enumerates and validates the properties given
+ * to open a stream of metrics the configuration is built up in the structure
+ * which starts out zero initialized.
+ */
+struct perf_open_properties {
+	u32 sample_flags;
+
+	u64 single_context:1;
+	u64 hold_preemption:1;
+	u64 ctx_handle;
+
+	/* OA sampling state */
+	int metrics_set;
+	int oa_format;
+	bool oa_periodic;
+	int oa_period_exponent;
+
+	struct intel_engine_cs *engine;
+
+	bool has_sseu;
+	struct intel_sseu sseu;
+
+	u64 poll_oa_period;
+};
+
+struct i915_oa_config_bo {
+	struct llist_node node;
+
+	struct i915_oa_config *oa_config;
+	struct i915_vma *vma;
+};
+
+static struct ctl_table_header *sysctl_header;
+
+static enum hrtimer_restart oa_poll_check_timer_cb(struct hrtimer *hrtimer);
+
+void i915_oa_config_release(struct kref *ref)
+{
+	struct i915_oa_config *oa_config =
+		container_of(ref, typeof(*oa_config), ref);
+
+	kfree(oa_config->flex_regs);
+	kfree(oa_config->b_counter_regs);
+	kfree(oa_config->mux_regs);
+
+	kfree_rcu(oa_config, rcu);
+}
+
+struct i915_oa_config *
+i915_perf_get_oa_config(struct i915_perf *perf, int metrics_set)
+{
+	struct i915_oa_config *oa_config;
+
+	rcu_read_lock();
+	oa_config = idr_find(&perf->metrics_idr, metrics_set);
+	if (oa_config)
+		oa_config = i915_oa_config_get(oa_config);
+	rcu_read_unlock();
+
+	return oa_config;
+}
+
+static void free_oa_config_bo(struct i915_oa_config_bo *oa_bo)
+{
+	i915_oa_config_put(oa_bo->oa_config);
+	i915_vma_put(oa_bo->vma);
+	kfree(oa_bo);
+}
+
+static u32 gen12_oa_hw_tail_read(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+
+	return intel_uncore_read(uncore, GEN12_OAG_OATAILPTR) &
+	       GEN12_OAG_OATAILPTR_MASK;
+}
+
+static u32 gen8_oa_hw_tail_read(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+
+	return intel_uncore_read(uncore, GEN8_OATAILPTR) & GEN8_OATAILPTR_MASK;
+}
+
+static u32 gen7_oa_hw_tail_read(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	u32 oastatus1 = intel_uncore_read(uncore, GEN7_OASTATUS1);
+
+	return oastatus1 & GEN7_OASTATUS1_TAIL_MASK;
+}
+
+/**
+ * oa_buffer_check_unlocked - check for data and update tail ptr state
+ * @stream: i915 stream instance
+ *
+ * This is either called via fops (for blocking reads in user ctx) or the poll
+ * check hrtimer (atomic ctx) to check the OA buffer tail pointer and check
+ * if there is data available for userspace to read.
+ *
+ * This function is central to providing a workaround for the OA unit tail
+ * pointer having a race with respect to what data is visible to the CPU.
+ * It is responsible for reading tail pointers from the hardware and giving
+ * the pointers time to 'age' before they are made available for reading.
+ * (See description of OA_TAIL_MARGIN_NSEC above for further details.)
+ *
+ * Besides returning true when there is data available to read() this function
+ * also updates the tail, aging_tail and aging_timestamp in the oa_buffer
+ * object.
+ *
+ * Note: It's safe to read OA config state here unlocked, assuming that this is
+ * only called while the stream is enabled, while the global OA configuration
+ * can't be modified.
+ *
+ * Returns: %true if the OA buffer contains data, else %false
+ */
+static bool oa_buffer_check_unlocked(struct i915_perf_stream *stream)
+{
+	u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
+	int report_size = stream->oa_buffer.format_size;
+	unsigned long flags;
+	bool pollin;
+	u32 hw_tail;
+	u64 now;
+
+	/* We have to consider the (unlikely) possibility that read() errors
+	 * could result in an OA buffer reset which might reset the head and
+	 * tail state.
+	 */
+	spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags);
+
+	hw_tail = stream->perf->ops.oa_hw_tail_read(stream);
+
+	/* The tail pointer increases in 64 byte increments,
+	 * not in report_size steps...
+	 */
+	hw_tail &= ~(report_size - 1);
+
+	now = ktime_get_mono_fast_ns();
+
+	if (hw_tail == stream->oa_buffer.aging_tail &&
+	    (now - stream->oa_buffer.aging_timestamp) > OA_TAIL_MARGIN_NSEC) {
+		/* If the HW tail hasn't move since the last check and the HW
+		 * tail has been aging for long enough, declare it the new
+		 * tail.
+		 */
+		stream->oa_buffer.tail = stream->oa_buffer.aging_tail;
+	} else {
+		u32 head, tail, aged_tail;
+
+		/* NB: The head we observe here might effectively be a little
+		 * out of date. If a read() is in progress, the head could be
+		 * anywhere between this head and stream->oa_buffer.tail.
+		 */
+		head = stream->oa_buffer.head - gtt_offset;
+		aged_tail = stream->oa_buffer.tail - gtt_offset;
+
+		hw_tail -= gtt_offset;
+		tail = hw_tail;
+
+		/* Walk the stream backward until we find a report with dword 0
+		 * & 1 not at 0. Since the circular buffer pointers progress by
+		 * increments of 64 bytes and that reports can be up to 256
+		 * bytes long, we can't tell whether a report has fully landed
+		 * in memory before the first 2 dwords of the following report
+		 * have effectively landed.
+		 *
+		 * This is assuming that the writes of the OA unit land in
+		 * memory in the order they were written to.
+		 * If not : (╯°□°）╯︵ ┻━┻
+		 */
+		while (OA_TAKEN(tail, aged_tail) >= report_size) {
+			u32 *report32 = (void *)(stream->oa_buffer.vaddr + tail);
+
+			if (report32[0] != 0 || report32[1] != 0)
+				break;
+
+			tail = (tail - report_size) & (OA_BUFFER_SIZE - 1);
+		}
+
+		if (OA_TAKEN(hw_tail, tail) > report_size &&
+		    __ratelimit(&stream->perf->tail_pointer_race))
+			DRM_NOTE("unlanded report(s) head=0x%x "
+				 "tail=0x%x hw_tail=0x%x\n",
+				 head, tail, hw_tail);
+
+		stream->oa_buffer.tail = gtt_offset + tail;
+		stream->oa_buffer.aging_tail = gtt_offset + hw_tail;
+		stream->oa_buffer.aging_timestamp = now;
+	}
+
+	pollin = OA_TAKEN(stream->oa_buffer.tail - gtt_offset,
+			  stream->oa_buffer.head - gtt_offset) >= report_size;
+
+	spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags);
+
+	return pollin;
+}
+
+/**
+ * append_oa_status - Appends a status record to a userspace read() buffer.
+ * @stream: An i915-perf stream opened for OA metrics
+ * @buf: destination buffer given by userspace
+ * @count: the number of bytes userspace wants to read
+ * @offset: (inout): the current position for writing into @buf
+ * @type: The kind of status to report to userspace
+ *
+ * Writes a status record (such as `DRM_I915_PERF_RECORD_OA_REPORT_LOST`)
+ * into the userspace read() buffer.
+ *
+ * The @buf @offset will only be updated on success.
+ *
+ * Returns: 0 on success, negative error code on failure.
+ */
+static int append_oa_status(struct i915_perf_stream *stream,
+			    char __user *buf,
+			    size_t count,
+			    size_t *offset,
+			    enum drm_i915_perf_record_type type)
+{
+	struct drm_i915_perf_record_header header = { type, 0, sizeof(header) };
+
+	if ((count - *offset) < header.size)
+		return -ENOSPC;
+
+	if (copy_to_user(buf + *offset, &header, sizeof(header)))
+		return -EFAULT;
+
+	(*offset) += header.size;
+
+	return 0;
+}
+
+/**
+ * append_oa_sample - Copies single OA report into userspace read() buffer.
+ * @stream: An i915-perf stream opened for OA metrics
+ * @buf: destination buffer given by userspace
+ * @count: the number of bytes userspace wants to read
+ * @offset: (inout): the current position for writing into @buf
+ * @report: A single OA report to (optionally) include as part of the sample
+ *
+ * The contents of a sample are configured through `DRM_I915_PERF_PROP_SAMPLE_*`
+ * properties when opening a stream, tracked as `stream->sample_flags`. This
+ * function copies the requested components of a single sample to the given
+ * read() @buf.
+ *
+ * The @buf @offset will only be updated on success.
+ *
+ * Returns: 0 on success, negative error code on failure.
+ */
+static int append_oa_sample(struct i915_perf_stream *stream,
+			    char __user *buf,
+			    size_t count,
+			    size_t *offset,
+			    const u8 *report)
+{
+	int report_size = stream->oa_buffer.format_size;
+	struct drm_i915_perf_record_header header;
+
+	header.type = DRM_I915_PERF_RECORD_SAMPLE;
+	header.pad = 0;
+	header.size = stream->sample_size;
+
+	if ((count - *offset) < header.size)
+		return -ENOSPC;
+
+	buf += *offset;
+	if (copy_to_user(buf, &header, sizeof(header)))
+		return -EFAULT;
+	buf += sizeof(header);
+
+	if (copy_to_user(buf, report, report_size))
+		return -EFAULT;
+
+	(*offset) += header.size;
+
+	return 0;
+}
+
+/**
+ * gen8_append_oa_reports - Copies all buffered OA reports into
+ *			    userspace read() buffer.
+ * @stream: An i915-perf stream opened for OA metrics
+ * @buf: destination buffer given by userspace
+ * @count: the number of bytes userspace wants to read
+ * @offset: (inout): the current position for writing into @buf
+ *
+ * Notably any error condition resulting in a short read (-%ENOSPC or
+ * -%EFAULT) will be returned even though one or more records may
+ * have been successfully copied. In this case it's up to the caller
+ * to decide if the error should be squashed before returning to
+ * userspace.
+ *
+ * Note: reports are consumed from the head, and appended to the
+ * tail, so the tail chases the head?... If you think that's mad
+ * and back-to-front you're not alone, but this follows the
+ * Gen PRM naming convention.
+ *
+ * Returns: 0 on success, negative error code on failure.
+ */
+static int gen8_append_oa_reports(struct i915_perf_stream *stream,
+				  char __user *buf,
+				  size_t count,
+				  size_t *offset)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	int report_size = stream->oa_buffer.format_size;
+	u8 *oa_buf_base = stream->oa_buffer.vaddr;
+	u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
+	u32 mask = (OA_BUFFER_SIZE - 1);
+	size_t start_offset = *offset;
+	unsigned long flags;
+	u32 head, tail;
+	u32 taken;
+	int ret = 0;
+
+	if (drm_WARN_ON(&uncore->i915->drm, !stream->enabled))
+		return -EIO;
+
+	spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags);
+
+	head = stream->oa_buffer.head;
+	tail = stream->oa_buffer.tail;
+
+	spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags);
+
+	/*
+	 * NB: oa_buffer.head/tail include the gtt_offset which we don't want
+	 * while indexing relative to oa_buf_base.
+	 */
+	head -= gtt_offset;
+	tail -= gtt_offset;
+
+	/*
+	 * An out of bounds or misaligned head or tail pointer implies a driver
+	 * bug since we validate + align the tail pointers we read from the
+	 * hardware and we are in full control of the head pointer which should
+	 * only be incremented by multiples of the report size (notably also
+	 * all a power of two).
+	 */
+	if (drm_WARN_ONCE(&uncore->i915->drm,
+			  head > OA_BUFFER_SIZE || head % report_size ||
+			  tail > OA_BUFFER_SIZE || tail % report_size,
+			  "Inconsistent OA buffer pointers: head = %u, tail = %u\n",
+			  head, tail))
+		return -EIO;
+
+
+	for (/* none */;
+	     (taken = OA_TAKEN(tail, head));
+	     head = (head + report_size) & mask) {
+		u8 *report = oa_buf_base + head;
+		u32 *report32 = (void *)report;
+		u32 ctx_id;
+		u32 reason;
+
+		/*
+		 * All the report sizes factor neatly into the buffer
+		 * size so we never expect to see a report split
+		 * between the beginning and end of the buffer.
+		 *
+		 * Given the initial alignment check a misalignment
+		 * here would imply a driver bug that would result
+		 * in an overrun.
+		 */
+		if (drm_WARN_ON(&uncore->i915->drm,
+				(OA_BUFFER_SIZE - head) < report_size)) {
+			drm_err(&uncore->i915->drm,
+				"Spurious OA head ptr: non-integral report offset\n");
+			break;
+		}
+
+		/*
+		 * The reason field includes flags identifying what
+		 * triggered this specific report (mostly timer
+		 * triggered or e.g. due to a context switch).
+		 *
+		 * This field is never expected to be zero so we can
+		 * check that the report isn't invalid before copying
+		 * it to userspace...
+		 */
+		reason = ((report32[0] >> OAREPORT_REASON_SHIFT) &
+			  (GRAPHICS_VER(stream->perf->i915) == 12 ?
+			   OAREPORT_REASON_MASK_EXTENDED :
+			   OAREPORT_REASON_MASK));
+
+		ctx_id = report32[2] & stream->specific_ctx_id_mask;
+
+		/*
+		 * Squash whatever is in the CTX_ID field if it's marked as
+		 * invalid to be sure we avoid false-positive, single-context
+		 * filtering below...
+		 *
+		 * Note: that we don't clear the valid_ctx_bit so userspace can
+		 * understand that the ID has been squashed by the kernel.
+		 */
+		if (!(report32[0] & stream->perf->gen8_valid_ctx_bit) &&
+		    GRAPHICS_VER(stream->perf->i915) <= 11)
+			ctx_id = report32[2] = INVALID_CTX_ID;
+
+		/*
+		 * NB: For Gen 8 the OA unit no longer supports clock gating
+		 * off for a specific context and the kernel can't securely
+		 * stop the counters from updating as system-wide / global
+		 * values.
+		 *
+		 * Automatic reports now include a context ID so reports can be
+		 * filtered on the cpu but it's not worth trying to
+		 * automatically subtract/hide counter progress for other
+		 * contexts while filtering since we can't stop userspace
+		 * issuing MI_REPORT_PERF_COUNT commands which would still
+		 * provide a side-band view of the real values.
+		 *
+		 * To allow userspace (such as Mesa/GL_INTEL_performance_query)
+		 * to normalize counters for a single filtered context then it
+		 * needs be forwarded bookend context-switch reports so that it
+		 * can track switches in between MI_REPORT_PERF_COUNT commands
+		 * and can itself subtract/ignore the progress of counters
+		 * associated with other contexts. Note that the hardware
+		 * automatically triggers reports when switching to a new
+		 * context which are tagged with the ID of the newly active
+		 * context. To avoid the complexity (and likely fragility) of
+		 * reading ahead while parsing reports to try and minimize
+		 * forwarding redundant context switch reports (i.e. between
+		 * other, unrelated contexts) we simply elect to forward them
+		 * all.
+		 *
+		 * We don't rely solely on the reason field to identify context
+		 * switches since it's not-uncommon for periodic samples to
+		 * identify a switch before any 'context switch' report.
+		 */
+		if (!stream->perf->exclusive_stream->ctx ||
+		    stream->specific_ctx_id == ctx_id ||
+		    stream->oa_buffer.last_ctx_id == stream->specific_ctx_id ||
+		    reason & OAREPORT_REASON_CTX_SWITCH) {
+
+			/*
+			 * While filtering for a single context we avoid
+			 * leaking the IDs of other contexts.
+			 */
+			if (stream->perf->exclusive_stream->ctx &&
+			    stream->specific_ctx_id != ctx_id) {
+				report32[2] = INVALID_CTX_ID;
+			}
+
+			ret = append_oa_sample(stream, buf, count, offset,
+					       report);
+			if (ret)
+				break;
+
+			stream->oa_buffer.last_ctx_id = ctx_id;
+		}
+
+		/*
+		 * Clear out the first 2 dword as a mean to detect unlanded
+		 * reports.
+		 */
+		report32[0] = 0;
+		report32[1] = 0;
+	}
+
+	if (start_offset != *offset) {
+		i915_reg_t oaheadptr;
+
+		oaheadptr = GRAPHICS_VER(stream->perf->i915) == 12 ?
+			    GEN12_OAG_OAHEADPTR : GEN8_OAHEADPTR;
+
+		spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags);
+
+		/*
+		 * We removed the gtt_offset for the copy loop above, indexing
+		 * relative to oa_buf_base so put back here...
+		 */
+		head += gtt_offset;
+		intel_uncore_write(uncore, oaheadptr,
+				   head & GEN12_OAG_OAHEADPTR_MASK);
+		stream->oa_buffer.head = head;
+
+		spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags);
+	}
+
+	return ret;
+}
+
+/**
+ * gen8_oa_read - copy status records then buffered OA reports
+ * @stream: An i915-perf stream opened for OA metrics
+ * @buf: destination buffer given by userspace
+ * @count: the number of bytes userspace wants to read
+ * @offset: (inout): the current position for writing into @buf
+ *
+ * Checks OA unit status registers and if necessary appends corresponding
+ * status records for userspace (such as for a buffer full condition) and then
+ * initiate appending any buffered OA reports.
+ *
+ * Updates @offset according to the number of bytes successfully copied into
+ * the userspace buffer.
+ *
+ * NB: some data may be successfully copied to the userspace buffer
+ * even if an error is returned, and this is reflected in the
+ * updated @offset.
+ *
+ * Returns: zero on success or a negative error code
+ */
+static int gen8_oa_read(struct i915_perf_stream *stream,
+			char __user *buf,
+			size_t count,
+			size_t *offset)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	u32 oastatus;
+	i915_reg_t oastatus_reg;
+	int ret;
+
+	if (drm_WARN_ON(&uncore->i915->drm, !stream->oa_buffer.vaddr))
+		return -EIO;
+
+	oastatus_reg = GRAPHICS_VER(stream->perf->i915) == 12 ?
+		       GEN12_OAG_OASTATUS : GEN8_OASTATUS;
+
+	oastatus = intel_uncore_read(uncore, oastatus_reg);
+
+	/*
+	 * We treat OABUFFER_OVERFLOW as a significant error:
+	 *
+	 * Although theoretically we could handle this more gracefully
+	 * sometimes, some Gens don't correctly suppress certain
+	 * automatically triggered reports in this condition and so we
+	 * have to assume that old reports are now being trampled
+	 * over.
+	 *
+	 * Considering how we don't currently give userspace control
+	 * over the OA buffer size and always configure a large 16MB
+	 * buffer, then a buffer overflow does anyway likely indicate
+	 * that something has gone quite badly wrong.
+	 */
+	if (oastatus & GEN8_OASTATUS_OABUFFER_OVERFLOW) {
+		ret = append_oa_status(stream, buf, count, offset,
+				       DRM_I915_PERF_RECORD_OA_BUFFER_LOST);
+		if (ret)
+			return ret;
+
+		drm_dbg(&stream->perf->i915->drm,
+			"OA buffer overflow (exponent = %d): force restart\n",
+			stream->period_exponent);
+
+		stream->perf->ops.oa_disable(stream);
+		stream->perf->ops.oa_enable(stream);
+
+		/*
+		 * Note: .oa_enable() is expected to re-init the oabuffer and
+		 * reset GEN8_OASTATUS for us
+		 */
+		oastatus = intel_uncore_read(uncore, oastatus_reg);
+	}
+
+	if (oastatus & GEN8_OASTATUS_REPORT_LOST) {
+		ret = append_oa_status(stream, buf, count, offset,
+				       DRM_I915_PERF_RECORD_OA_REPORT_LOST);
+		if (ret)
+			return ret;
+
+		intel_uncore_rmw(uncore, oastatus_reg,
+				 GEN8_OASTATUS_COUNTER_OVERFLOW |
+				 GEN8_OASTATUS_REPORT_LOST,
+				 IS_GRAPHICS_VER(uncore->i915, 8, 11) ?
+				 (GEN8_OASTATUS_HEAD_POINTER_WRAP |
+				  GEN8_OASTATUS_TAIL_POINTER_WRAP) : 0);
+	}
+
+	return gen8_append_oa_reports(stream, buf, count, offset);
+}
+
+/**
+ * gen7_append_oa_reports - Copies all buffered OA reports into
+ *			    userspace read() buffer.
+ * @stream: An i915-perf stream opened for OA metrics
+ * @buf: destination buffer given by userspace
+ * @count: the number of bytes userspace wants to read
+ * @offset: (inout): the current position for writing into @buf
+ *
+ * Notably any error condition resulting in a short read (-%ENOSPC or
+ * -%EFAULT) will be returned even though one or more records may
+ * have been successfully copied. In this case it's up to the caller
+ * to decide if the error should be squashed before returning to
+ * userspace.
+ *
+ * Note: reports are consumed from the head, and appended to the
+ * tail, so the tail chases the head?... If you think that's mad
+ * and back-to-front you're not alone, but this follows the
+ * Gen PRM naming convention.
+ *
+ * Returns: 0 on success, negative error code on failure.
+ */
+static int gen7_append_oa_reports(struct i915_perf_stream *stream,
+				  char __user *buf,
+				  size_t count,
+				  size_t *offset)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	int report_size = stream->oa_buffer.format_size;
+	u8 *oa_buf_base = stream->oa_buffer.vaddr;
+	u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
+	u32 mask = (OA_BUFFER_SIZE - 1);
+	size_t start_offset = *offset;
+	unsigned long flags;
+	u32 head, tail;
+	u32 taken;
+	int ret = 0;
+
+	if (drm_WARN_ON(&uncore->i915->drm, !stream->enabled))
+		return -EIO;
+
+	spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags);
+
+	head = stream->oa_buffer.head;
+	tail = stream->oa_buffer.tail;
+
+	spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags);
+
+	/* NB: oa_buffer.head/tail include the gtt_offset which we don't want
+	 * while indexing relative to oa_buf_base.
+	 */
+	head -= gtt_offset;
+	tail -= gtt_offset;
+
+	/* An out of bounds or misaligned head or tail pointer implies a driver
+	 * bug since we validate + align the tail pointers we read from the
+	 * hardware and we are in full control of the head pointer which should
+	 * only be incremented by multiples of the report size (notably also
+	 * all a power of two).
+	 */
+	if (drm_WARN_ONCE(&uncore->i915->drm,
+			  head > OA_BUFFER_SIZE || head % report_size ||
+			  tail > OA_BUFFER_SIZE || tail % report_size,
+			  "Inconsistent OA buffer pointers: head = %u, tail = %u\n",
+			  head, tail))
+		return -EIO;
+
+
+	for (/* none */;
+	     (taken = OA_TAKEN(tail, head));
+	     head = (head + report_size) & mask) {
+		u8 *report = oa_buf_base + head;
+		u32 *report32 = (void *)report;
+
+		/* All the report sizes factor neatly into the buffer
+		 * size so we never expect to see a report split
+		 * between the beginning and end of the buffer.
+		 *
+		 * Given the initial alignment check a misalignment
+		 * here would imply a driver bug that would result
+		 * in an overrun.
+		 */
+		if (drm_WARN_ON(&uncore->i915->drm,
+				(OA_BUFFER_SIZE - head) < report_size)) {
+			drm_err(&uncore->i915->drm,
+				"Spurious OA head ptr: non-integral report offset\n");
+			break;
+		}
+
+		/* The report-ID field for periodic samples includes
+		 * some undocumented flags related to what triggered
+		 * the report and is never expected to be zero so we
+		 * can check that the report isn't invalid before
+		 * copying it to userspace...
+		 */
+		if (report32[0] == 0) {
+			if (__ratelimit(&stream->perf->spurious_report_rs))
+				DRM_NOTE("Skipping spurious, invalid OA report\n");
+			continue;
+		}
+
+		ret = append_oa_sample(stream, buf, count, offset, report);
+		if (ret)
+			break;
+
+		/* Clear out the first 2 dwords as a mean to detect unlanded
+		 * reports.
+		 */
+		report32[0] = 0;
+		report32[1] = 0;
+	}
+
+	if (start_offset != *offset) {
+		spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags);
+
+		/* We removed the gtt_offset for the copy loop above, indexing
+		 * relative to oa_buf_base so put back here...
+		 */
+		head += gtt_offset;
+
+		intel_uncore_write(uncore, GEN7_OASTATUS2,
+				   (head & GEN7_OASTATUS2_HEAD_MASK) |
+				   GEN7_OASTATUS2_MEM_SELECT_GGTT);
+		stream->oa_buffer.head = head;
+
+		spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags);
+	}
+
+	return ret;
+}
+
+/**
+ * gen7_oa_read - copy status records then buffered OA reports
+ * @stream: An i915-perf stream opened for OA metrics
+ * @buf: destination buffer given by userspace
+ * @count: the number of bytes userspace wants to read
+ * @offset: (inout): the current position for writing into @buf
+ *
+ * Checks Gen 7 specific OA unit status registers and if necessary appends
+ * corresponding status records for userspace (such as for a buffer full
+ * condition) and then initiate appending any buffered OA reports.
+ *
+ * Updates @offset according to the number of bytes successfully copied into
+ * the userspace buffer.
+ *
+ * Returns: zero on success or a negative error code
+ */
+static int gen7_oa_read(struct i915_perf_stream *stream,
+			char __user *buf,
+			size_t count,
+			size_t *offset)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	u32 oastatus1;
+	int ret;
+
+	if (drm_WARN_ON(&uncore->i915->drm, !stream->oa_buffer.vaddr))
+		return -EIO;
+
+	oastatus1 = intel_uncore_read(uncore, GEN7_OASTATUS1);
+
+	/* XXX: On Haswell we don't have a safe way to clear oastatus1
+	 * bits while the OA unit is enabled (while the tail pointer
+	 * may be updated asynchronously) so we ignore status bits
+	 * that have already been reported to userspace.
+	 */
+	oastatus1 &= ~stream->perf->gen7_latched_oastatus1;
+
+	/* We treat OABUFFER_OVERFLOW as a significant error:
+	 *
+	 * - The status can be interpreted to mean that the buffer is
+	 *   currently full (with a higher precedence than OA_TAKEN()
+	 *   which will start to report a near-empty buffer after an
+	 *   overflow) but it's awkward that we can't clear the status
+	 *   on Haswell, so without a reset we won't be able to catch
+	 *   the state again.
+	 *
+	 * - Since it also implies the HW has started overwriting old
+	 *   reports it may also affect our sanity checks for invalid
+	 *   reports when copying to userspace that assume new reports
+	 *   are being written to cleared memory.
+	 *
+	 * - In the future we may want to introduce a flight recorder
+	 *   mode where the driver will automatically maintain a safe
+	 *   guard band between head/tail, avoiding this overflow
+	 *   condition, but we avoid the added driver complexity for
+	 *   now.
+	 */
+	if (unlikely(oastatus1 & GEN7_OASTATUS1_OABUFFER_OVERFLOW)) {
+		ret = append_oa_status(stream, buf, count, offset,
+				       DRM_I915_PERF_RECORD_OA_BUFFER_LOST);
+		if (ret)
+			return ret;
+
+		drm_dbg(&stream->perf->i915->drm,
+			"OA buffer overflow (exponent = %d): force restart\n",
+			stream->period_exponent);
+
+		stream->perf->ops.oa_disable(stream);
+		stream->perf->ops.oa_enable(stream);
+
+		oastatus1 = intel_uncore_read(uncore, GEN7_OASTATUS1);
+	}
+
+	if (unlikely(oastatus1 & GEN7_OASTATUS1_REPORT_LOST)) {
+		ret = append_oa_status(stream, buf, count, offset,
+				       DRM_I915_PERF_RECORD_OA_REPORT_LOST);
+		if (ret)
+			return ret;
+		stream->perf->gen7_latched_oastatus1 |=
+			GEN7_OASTATUS1_REPORT_LOST;
+	}
+
+	return gen7_append_oa_reports(stream, buf, count, offset);
+}
+
+/**
+ * i915_oa_wait_unlocked - handles blocking IO until OA data available
+ * @stream: An i915-perf stream opened for OA metrics
+ *
+ * Called when userspace tries to read() from a blocking stream FD opened
+ * for OA metrics. It waits until the hrtimer callback finds a non-empty
+ * OA buffer and wakes us.
+ *
+ * Note: it's acceptable to have this return with some false positives
+ * since any subsequent read handling will return -EAGAIN if there isn't
+ * really data ready for userspace yet.
+ *
+ * Returns: zero on success or a negative error code
+ */
+static int i915_oa_wait_unlocked(struct i915_perf_stream *stream)
+{
+	/* We would wait indefinitely if periodic sampling is not enabled */
+	if (!stream->periodic)
+		return -EIO;
+
+	return wait_event_interruptible(stream->poll_wq,
+					oa_buffer_check_unlocked(stream));
+}
+
+/**
+ * i915_oa_poll_wait - call poll_wait() for an OA stream poll()
+ * @stream: An i915-perf stream opened for OA metrics
+ * @file: An i915 perf stream file
+ * @wait: poll() state table
+ *
+ * For handling userspace polling on an i915 perf stream opened for OA metrics,
+ * this starts a poll_wait with the wait queue that our hrtimer callback wakes
+ * when it sees data ready to read in the circular OA buffer.
+ */
+static void i915_oa_poll_wait(struct i915_perf_stream *stream,
+			      struct file *file,
+			      poll_table *wait)
+{
+	poll_wait(file, &stream->poll_wq, wait);
+}
+
+/**
+ * i915_oa_read - just calls through to &i915_oa_ops->read
+ * @stream: An i915-perf stream opened for OA metrics
+ * @buf: destination buffer given by userspace
+ * @count: the number of bytes userspace wants to read
+ * @offset: (inout): the current position for writing into @buf
+ *
+ * Updates @offset according to the number of bytes successfully copied into
+ * the userspace buffer.
+ *
+ * Returns: zero on success or a negative error code
+ */
+static int i915_oa_read(struct i915_perf_stream *stream,
+			char __user *buf,
+			size_t count,
+			size_t *offset)
+{
+	return stream->perf->ops.read(stream, buf, count, offset);
+}
+
+static struct intel_context *oa_pin_context(struct i915_perf_stream *stream)
+{
+	struct i915_gem_engines_iter it;
+	struct i915_gem_context *ctx = stream->ctx;
+	struct intel_context *ce;
+	struct i915_gem_ww_ctx ww;
+	int err = -ENODEV;
+
+	for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it) {
+		if (ce->engine != stream->engine) /* first match! */
+			continue;
+
+		err = 0;
+		break;
+	}
+	i915_gem_context_unlock_engines(ctx);
+
+	if (err)
+		return ERR_PTR(err);
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	/*
+	 * As the ID is the gtt offset of the context's vma we
+	 * pin the vma to ensure the ID remains fixed.
+	 */
+	err = intel_context_pin_ww(ce, &ww);
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+
+	if (err)
+		return ERR_PTR(err);
+
+	stream->pinned_ctx = ce;
+	return stream->pinned_ctx;
+}
+
+/**
+ * oa_get_render_ctx_id - determine and hold ctx hw id
+ * @stream: An i915-perf stream opened for OA metrics
+ *
+ * Determine the render context hw id, and ensure it remains fixed for the
+ * lifetime of the stream. This ensures that we don't have to worry about
+ * updating the context ID in OACONTROL on the fly.
+ *
+ * Returns: zero on success or a negative error code
+ */
+static int oa_get_render_ctx_id(struct i915_perf_stream *stream)
+{
+	struct intel_context *ce;
+
+	ce = oa_pin_context(stream);
+	if (IS_ERR(ce))
+		return PTR_ERR(ce);
+
+	switch (GRAPHICS_VER(ce->engine->i915)) {
+	case 7: {
+		/*
+		 * On Haswell we don't do any post processing of the reports
+		 * and don't need to use the mask.
+		 */
+		stream->specific_ctx_id = i915_ggtt_offset(ce->state);
+		stream->specific_ctx_id_mask = 0;
+		break;
+	}
+
+	case 8:
+	case 9:
+		if (intel_engine_uses_guc(ce->engine)) {
+			/*
+			 * When using GuC, the context descriptor we write in
+			 * i915 is read by GuC and rewritten before it's
+			 * actually written into the hardware. The LRCA is
+			 * what is put into the context id field of the
+			 * context descriptor by GuC. Because it's aligned to
+			 * a page, the lower 12bits are always at 0 and
+			 * dropped by GuC. They won't be part of the context
+			 * ID in the OA reports, so squash those lower bits.
+			 */
+			stream->specific_ctx_id = ce->lrc.lrca >> 12;
+
+			/*
+			 * GuC uses the top bit to signal proxy submission, so
+			 * ignore that bit.
+			 */
+			stream->specific_ctx_id_mask =
+				(1U << (GEN8_CTX_ID_WIDTH - 1)) - 1;
+		} else {
+			stream->specific_ctx_id_mask =
+				(1U << GEN8_CTX_ID_WIDTH) - 1;
+			stream->specific_ctx_id = stream->specific_ctx_id_mask;
+		}
+		break;
+
+	case 11:
+	case 12:
+		if (GRAPHICS_VER_FULL(ce->engine->i915) >= IP_VER(12, 50)) {
+			stream->specific_ctx_id_mask =
+				((1U << XEHP_SW_CTX_ID_WIDTH) - 1) <<
+				(XEHP_SW_CTX_ID_SHIFT - 32);
+			stream->specific_ctx_id =
+				(XEHP_MAX_CONTEXT_HW_ID - 1) <<
+				(XEHP_SW_CTX_ID_SHIFT - 32);
+		} else {
+			stream->specific_ctx_id_mask =
+				((1U << GEN11_SW_CTX_ID_WIDTH) - 1) << (GEN11_SW_CTX_ID_SHIFT - 32);
+			/*
+			 * Pick an unused context id
+			 * 0 - BITS_PER_LONG are used by other contexts
+			 * GEN12_MAX_CONTEXT_HW_ID (0x7ff) is used by idle context
+			 */
+			stream->specific_ctx_id =
+				(GEN12_MAX_CONTEXT_HW_ID - 1) << (GEN11_SW_CTX_ID_SHIFT - 32);
+		}
+		break;
+
+	default:
+		MISSING_CASE(GRAPHICS_VER(ce->engine->i915));
+	}
+
+	ce->tag = stream->specific_ctx_id;
+
+	drm_dbg(&stream->perf->i915->drm,
+		"filtering on ctx_id=0x%x ctx_id_mask=0x%x\n",
+		stream->specific_ctx_id,
+		stream->specific_ctx_id_mask);
+
+	return 0;
+}
+
+/**
+ * oa_put_render_ctx_id - counterpart to oa_get_render_ctx_id releases hold
+ * @stream: An i915-perf stream opened for OA metrics
+ *
+ * In case anything needed doing to ensure the context HW ID would remain valid
+ * for the lifetime of the stream, then that can be undone here.
+ */
+static void oa_put_render_ctx_id(struct i915_perf_stream *stream)
+{
+	struct intel_context *ce;
+
+	ce = fetch_and_zero(&stream->pinned_ctx);
+	if (ce) {
+		ce->tag = 0; /* recomputed on next submission after parking */
+		intel_context_unpin(ce);
+	}
+
+	stream->specific_ctx_id = INVALID_CTX_ID;
+	stream->specific_ctx_id_mask = 0;
+}
+
+static void
+free_oa_buffer(struct i915_perf_stream *stream)
+{
+	i915_vma_unpin_and_release(&stream->oa_buffer.vma,
+				   I915_VMA_RELEASE_MAP);
+
+	stream->oa_buffer.vaddr = NULL;
+}
+
+static void
+free_oa_configs(struct i915_perf_stream *stream)
+{
+	struct i915_oa_config_bo *oa_bo, *tmp;
+
+	i915_oa_config_put(stream->oa_config);
+	llist_for_each_entry_safe(oa_bo, tmp, stream->oa_config_bos.first, node)
+		free_oa_config_bo(oa_bo);
+}
+
+static void
+free_noa_wait(struct i915_perf_stream *stream)
+{
+	i915_vma_unpin_and_release(&stream->noa_wait, 0);
+}
+
+static void i915_oa_stream_destroy(struct i915_perf_stream *stream)
+{
+	struct i915_perf *perf = stream->perf;
+
+	if (WARN_ON(stream != perf->exclusive_stream))
+		return;
+
+	/*
+	 * Unset exclusive_stream first, it will be checked while disabling
+	 * the metric set on gen8+.
+	 *
+	 * See i915_oa_init_reg_state() and lrc_configure_all_contexts()
+	 */
+	WRITE_ONCE(perf->exclusive_stream, NULL);
+	perf->ops.disable_metric_set(stream);
+
+	free_oa_buffer(stream);
+
+	intel_uncore_forcewake_put(stream->uncore, FORCEWAKE_ALL);
+	intel_engine_pm_put(stream->engine);
+
+	if (stream->ctx)
+		oa_put_render_ctx_id(stream);
+
+	free_oa_configs(stream);
+	free_noa_wait(stream);
+
+	if (perf->spurious_report_rs.missed) {
+		DRM_NOTE("%d spurious OA report notices suppressed due to ratelimiting\n",
+			 perf->spurious_report_rs.missed);
+	}
+}
+
+static void gen7_init_oa_buffer(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
+	unsigned long flags;
+
+	spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags);
+
+	/* Pre-DevBDW: OABUFFER must be set with counters off,
+	 * before OASTATUS1, but after OASTATUS2
+	 */
+	intel_uncore_write(uncore, GEN7_OASTATUS2, /* head */
+			   gtt_offset | GEN7_OASTATUS2_MEM_SELECT_GGTT);
+	stream->oa_buffer.head = gtt_offset;
+
+	intel_uncore_write(uncore, GEN7_OABUFFER, gtt_offset);
+
+	intel_uncore_write(uncore, GEN7_OASTATUS1, /* tail */
+			   gtt_offset | OABUFFER_SIZE_16M);
+
+	/* Mark that we need updated tail pointers to read from... */
+	stream->oa_buffer.aging_tail = INVALID_TAIL_PTR;
+	stream->oa_buffer.tail = gtt_offset;
+
+	spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags);
+
+	/* On Haswell we have to track which OASTATUS1 flags we've
+	 * already seen since they can't be cleared while periodic
+	 * sampling is enabled.
+	 */
+	stream->perf->gen7_latched_oastatus1 = 0;
+
+	/* NB: although the OA buffer will initially be allocated
+	 * zeroed via shmfs (and so this memset is redundant when
+	 * first allocating), we may re-init the OA buffer, either
+	 * when re-enabling a stream or in error/reset paths.
+	 *
+	 * The reason we clear the buffer for each re-init is for the
+	 * sanity check in gen7_append_oa_reports() that looks at the
+	 * report-id field to make sure it's non-zero which relies on
+	 * the assumption that new reports are being written to zeroed
+	 * memory...
+	 */
+	memset(stream->oa_buffer.vaddr, 0, OA_BUFFER_SIZE);
+}
+
+static void gen8_init_oa_buffer(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
+	unsigned long flags;
+
+	spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags);
+
+	intel_uncore_write(uncore, GEN8_OASTATUS, 0);
+	intel_uncore_write(uncore, GEN8_OAHEADPTR, gtt_offset);
+	stream->oa_buffer.head = gtt_offset;
+
+	intel_uncore_write(uncore, GEN8_OABUFFER_UDW, 0);
+
+	/*
+	 * PRM says:
+	 *
+	 *  "This MMIO must be set before the OATAILPTR
+	 *  register and after the OAHEADPTR register. This is
+	 *  to enable proper functionality of the overflow
+	 *  bit."
+	 */
+	intel_uncore_write(uncore, GEN8_OABUFFER, gtt_offset |
+		   OABUFFER_SIZE_16M | GEN8_OABUFFER_MEM_SELECT_GGTT);
+	intel_uncore_write(uncore, GEN8_OATAILPTR, gtt_offset & GEN8_OATAILPTR_MASK);
+
+	/* Mark that we need updated tail pointers to read from... */
+	stream->oa_buffer.aging_tail = INVALID_TAIL_PTR;
+	stream->oa_buffer.tail = gtt_offset;
+
+	/*
+	 * Reset state used to recognise context switches, affecting which
+	 * reports we will forward to userspace while filtering for a single
+	 * context.
+	 */
+	stream->oa_buffer.last_ctx_id = INVALID_CTX_ID;
+
+	spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags);
+
+	/*
+	 * NB: although the OA buffer will initially be allocated
+	 * zeroed via shmfs (and so this memset is redundant when
+	 * first allocating), we may re-init the OA buffer, either
+	 * when re-enabling a stream or in error/reset paths.
+	 *
+	 * The reason we clear the buffer for each re-init is for the
+	 * sanity check in gen8_append_oa_reports() that looks at the
+	 * reason field to make sure it's non-zero which relies on
+	 * the assumption that new reports are being written to zeroed
+	 * memory...
+	 */
+	memset(stream->oa_buffer.vaddr, 0, OA_BUFFER_SIZE);
+}
+
+static void gen12_init_oa_buffer(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	u32 gtt_offset = i915_ggtt_offset(stream->oa_buffer.vma);
+	unsigned long flags;
+
+	spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags);
+
+	intel_uncore_write(uncore, GEN12_OAG_OASTATUS, 0);
+	intel_uncore_write(uncore, GEN12_OAG_OAHEADPTR,
+			   gtt_offset & GEN12_OAG_OAHEADPTR_MASK);
+	stream->oa_buffer.head = gtt_offset;
+
+	/*
+	 * PRM says:
+	 *
+	 *  "This MMIO must be set before the OATAILPTR
+	 *  register and after the OAHEADPTR register. This is
+	 *  to enable proper functionality of the overflow
+	 *  bit."
+	 */
+	intel_uncore_write(uncore, GEN12_OAG_OABUFFER, gtt_offset |
+			   OABUFFER_SIZE_16M | GEN8_OABUFFER_MEM_SELECT_GGTT);
+	intel_uncore_write(uncore, GEN12_OAG_OATAILPTR,
+			   gtt_offset & GEN12_OAG_OATAILPTR_MASK);
+
+	/* Mark that we need updated tail pointers to read from... */
+	stream->oa_buffer.aging_tail = INVALID_TAIL_PTR;
+	stream->oa_buffer.tail = gtt_offset;
+
+	/*
+	 * Reset state used to recognise context switches, affecting which
+	 * reports we will forward to userspace while filtering for a single
+	 * context.
+	 */
+	stream->oa_buffer.last_ctx_id = INVALID_CTX_ID;
+
+	spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags);
+
+	/*
+	 * NB: although the OA buffer will initially be allocated
+	 * zeroed via shmfs (and so this memset is redundant when
+	 * first allocating), we may re-init the OA buffer, either
+	 * when re-enabling a stream or in error/reset paths.
+	 *
+	 * The reason we clear the buffer for each re-init is for the
+	 * sanity check in gen8_append_oa_reports() that looks at the
+	 * reason field to make sure it's non-zero which relies on
+	 * the assumption that new reports are being written to zeroed
+	 * memory...
+	 */
+	memset(stream->oa_buffer.vaddr, 0,
+	       stream->oa_buffer.vma->size);
+}
+
+static int alloc_oa_buffer(struct i915_perf_stream *stream)
+{
+	struct drm_i915_private *i915 = stream->perf->i915;
+	struct drm_i915_gem_object *bo;
+	struct i915_vma *vma;
+	int ret;
+
+	if (drm_WARN_ON(&i915->drm, stream->oa_buffer.vma))
+		return -ENODEV;
+
+	BUILD_BUG_ON_NOT_POWER_OF_2(OA_BUFFER_SIZE);
+	BUILD_BUG_ON(OA_BUFFER_SIZE < SZ_128K || OA_BUFFER_SIZE > SZ_16M);
+
+	bo = i915_gem_object_create_shmem(stream->perf->i915, OA_BUFFER_SIZE);
+	if (IS_ERR(bo)) {
+		drm_err(&i915->drm, "Failed to allocate OA buffer\n");
+		return PTR_ERR(bo);
+	}
+
+	i915_gem_object_set_cache_coherency(bo, I915_CACHE_LLC);
+
+	/* PreHSW required 512K alignment, HSW requires 16M */
+	vma = i915_gem_object_ggtt_pin(bo, NULL, 0, SZ_16M, 0);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto err_unref;
+	}
+	stream->oa_buffer.vma = vma;
+
+	stream->oa_buffer.vaddr =
+		i915_gem_object_pin_map_unlocked(bo, I915_MAP_WB);
+	if (IS_ERR(stream->oa_buffer.vaddr)) {
+		ret = PTR_ERR(stream->oa_buffer.vaddr);
+		goto err_unpin;
+	}
+
+	return 0;
+
+err_unpin:
+	__i915_vma_unpin(vma);
+
+err_unref:
+	i915_gem_object_put(bo);
+
+	stream->oa_buffer.vaddr = NULL;
+	stream->oa_buffer.vma = NULL;
+
+	return ret;
+}
+
+static u32 *save_restore_register(struct i915_perf_stream *stream, u32 *cs,
+				  bool save, i915_reg_t reg, u32 offset,
+				  u32 dword_count)
+{
+	u32 cmd;
+	u32 d;
+
+	cmd = save ? MI_STORE_REGISTER_MEM : MI_LOAD_REGISTER_MEM;
+	cmd |= MI_SRM_LRM_GLOBAL_GTT;
+	if (GRAPHICS_VER(stream->perf->i915) >= 8)
+		cmd++;
+
+	for (d = 0; d < dword_count; d++) {
+		*cs++ = cmd;
+		*cs++ = i915_mmio_reg_offset(reg) + 4 * d;
+		*cs++ = intel_gt_scratch_offset(stream->engine->gt,
+						offset) + 4 * d;
+		*cs++ = 0;
+	}
+
+	return cs;
+}
+
+static int alloc_noa_wait(struct i915_perf_stream *stream)
+{
+	struct drm_i915_private *i915 = stream->perf->i915;
+	struct drm_i915_gem_object *bo;
+	struct i915_vma *vma;
+	const u64 delay_ticks = 0xffffffffffffffff -
+		intel_gt_ns_to_clock_interval(to_gt(stream->perf->i915),
+		atomic64_read(&stream->perf->noa_programming_delay));
+	const u32 base = stream->engine->mmio_base;
+#define CS_GPR(x) GEN8_RING_CS_GPR(base, x)
+	u32 *batch, *ts0, *cs, *jump;
+	struct i915_gem_ww_ctx ww;
+	int ret, i;
+	enum {
+		START_TS,
+		NOW_TS,
+		DELTA_TS,
+		JUMP_PREDICATE,
+		DELTA_TARGET,
+		N_CS_GPR
+	};
+
+	bo = i915_gem_object_create_internal(i915, 4096);
+	if (IS_ERR(bo)) {
+		drm_err(&i915->drm,
+			"Failed to allocate NOA wait batchbuffer\n");
+		return PTR_ERR(bo);
+	}
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	ret = i915_gem_object_lock(bo, &ww);
+	if (ret)
+		goto out_ww;
+
+	/*
+	 * We pin in GGTT because we jump into this buffer now because
+	 * multiple OA config BOs will have a jump to this address and it
+	 * needs to be fixed during the lifetime of the i915/perf stream.
+	 */
+	vma = i915_gem_object_ggtt_pin_ww(bo, &ww, NULL, 0, 0, PIN_HIGH);
+	if (IS_ERR(vma)) {
+		ret = PTR_ERR(vma);
+		goto out_ww;
+	}
+
+	batch = cs = i915_gem_object_pin_map(bo, I915_MAP_WB);
+	if (IS_ERR(batch)) {
+		ret = PTR_ERR(batch);
+		goto err_unpin;
+	}
+
+	/* Save registers. */
+	for (i = 0; i < N_CS_GPR; i++)
+		cs = save_restore_register(
+			stream, cs, true /* save */, CS_GPR(i),
+			INTEL_GT_SCRATCH_FIELD_PERF_CS_GPR + 8 * i, 2);
+	cs = save_restore_register(
+		stream, cs, true /* save */, MI_PREDICATE_RESULT_1(RENDER_RING_BASE),
+		INTEL_GT_SCRATCH_FIELD_PERF_PREDICATE_RESULT_1, 1);
+
+	/* First timestamp snapshot location. */
+	ts0 = cs;
+
+	/*
+	 * Initial snapshot of the timestamp register to implement the wait.
+	 * We work with 32b values, so clear out the top 32b bits of the
+	 * register because the ALU works 64bits.
+	 */
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(CS_GPR(START_TS)) + 4;
+	*cs++ = 0;
+	*cs++ = MI_LOAD_REGISTER_REG | (3 - 2);
+	*cs++ = i915_mmio_reg_offset(RING_TIMESTAMP(base));
+	*cs++ = i915_mmio_reg_offset(CS_GPR(START_TS));
+
+	/*
+	 * This is the location we're going to jump back into until the
+	 * required amount of time has passed.
+	 */
+	jump = cs;
+
+	/*
+	 * Take another snapshot of the timestamp register. Take care to clear
+	 * up the top 32bits of CS_GPR(1) as we're using it for other
+	 * operations below.
+	 */
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(CS_GPR(NOW_TS)) + 4;
+	*cs++ = 0;
+	*cs++ = MI_LOAD_REGISTER_REG | (3 - 2);
+	*cs++ = i915_mmio_reg_offset(RING_TIMESTAMP(base));
+	*cs++ = i915_mmio_reg_offset(CS_GPR(NOW_TS));
+
+	/*
+	 * Do a diff between the 2 timestamps and store the result back into
+	 * CS_GPR(1).
+	 */
+	*cs++ = MI_MATH(5);
+	*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCA, MI_MATH_REG(NOW_TS));
+	*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCB, MI_MATH_REG(START_TS));
+	*cs++ = MI_MATH_SUB;
+	*cs++ = MI_MATH_STORE(MI_MATH_REG(DELTA_TS), MI_MATH_REG_ACCU);
+	*cs++ = MI_MATH_STORE(MI_MATH_REG(JUMP_PREDICATE), MI_MATH_REG_CF);
+
+	/*
+	 * Transfer the carry flag (set to 1 if ts1 < ts0, meaning the
+	 * timestamp have rolled over the 32bits) into the predicate register
+	 * to be used for the predicated jump.
+	 */
+	*cs++ = MI_LOAD_REGISTER_REG | (3 - 2);
+	*cs++ = i915_mmio_reg_offset(CS_GPR(JUMP_PREDICATE));
+	*cs++ = i915_mmio_reg_offset(MI_PREDICATE_RESULT_1(RENDER_RING_BASE));
+
+	/* Restart from the beginning if we had timestamps roll over. */
+	*cs++ = (GRAPHICS_VER(i915) < 8 ?
+		 MI_BATCH_BUFFER_START :
+		 MI_BATCH_BUFFER_START_GEN8) |
+		MI_BATCH_PREDICATE;
+	*cs++ = i915_ggtt_offset(vma) + (ts0 - batch) * 4;
+	*cs++ = 0;
+
+	/*
+	 * Now add the diff between to previous timestamps and add it to :
+	 *      (((1 * << 64) - 1) - delay_ns)
+	 *
+	 * When the Carry Flag contains 1 this means the elapsed time is
+	 * longer than the expected delay, and we can exit the wait loop.
+	 */
+	*cs++ = MI_LOAD_REGISTER_IMM(2);
+	*cs++ = i915_mmio_reg_offset(CS_GPR(DELTA_TARGET));
+	*cs++ = lower_32_bits(delay_ticks);
+	*cs++ = i915_mmio_reg_offset(CS_GPR(DELTA_TARGET)) + 4;
+	*cs++ = upper_32_bits(delay_ticks);
+
+	*cs++ = MI_MATH(4);
+	*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCA, MI_MATH_REG(DELTA_TS));
+	*cs++ = MI_MATH_LOAD(MI_MATH_REG_SRCB, MI_MATH_REG(DELTA_TARGET));
+	*cs++ = MI_MATH_ADD;
+	*cs++ = MI_MATH_STOREINV(MI_MATH_REG(JUMP_PREDICATE), MI_MATH_REG_CF);
+
+	*cs++ = MI_ARB_CHECK;
+
+	/*
+	 * Transfer the result into the predicate register to be used for the
+	 * predicated jump.
+	 */
+	*cs++ = MI_LOAD_REGISTER_REG | (3 - 2);
+	*cs++ = i915_mmio_reg_offset(CS_GPR(JUMP_PREDICATE));
+	*cs++ = i915_mmio_reg_offset(MI_PREDICATE_RESULT_1(RENDER_RING_BASE));
+
+	/* Predicate the jump.  */
+	*cs++ = (GRAPHICS_VER(i915) < 8 ?
+		 MI_BATCH_BUFFER_START :
+		 MI_BATCH_BUFFER_START_GEN8) |
+		MI_BATCH_PREDICATE;
+	*cs++ = i915_ggtt_offset(vma) + (jump - batch) * 4;
+	*cs++ = 0;
+
+	/* Restore registers. */
+	for (i = 0; i < N_CS_GPR; i++)
+		cs = save_restore_register(
+			stream, cs, false /* restore */, CS_GPR(i),
+			INTEL_GT_SCRATCH_FIELD_PERF_CS_GPR + 8 * i, 2);
+	cs = save_restore_register(
+		stream, cs, false /* restore */, MI_PREDICATE_RESULT_1(RENDER_RING_BASE),
+		INTEL_GT_SCRATCH_FIELD_PERF_PREDICATE_RESULT_1, 1);
+
+	/* And return to the ring. */
+	*cs++ = MI_BATCH_BUFFER_END;
+
+	GEM_BUG_ON(cs - batch > PAGE_SIZE / sizeof(*batch));
+
+	i915_gem_object_flush_map(bo);
+	__i915_gem_object_release_map(bo);
+
+	stream->noa_wait = vma;
+	goto out_ww;
+
+err_unpin:
+	i915_vma_unpin_and_release(&vma, 0);
+out_ww:
+	if (ret == -EDEADLK) {
+		ret = i915_gem_ww_ctx_backoff(&ww);
+		if (!ret)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	if (ret)
+		i915_gem_object_put(bo);
+	return ret;
+}
+
+static u32 *write_cs_mi_lri(u32 *cs,
+			    const struct i915_oa_reg *reg_data,
+			    u32 n_regs)
+{
+	u32 i;
+
+	for (i = 0; i < n_regs; i++) {
+		if ((i % MI_LOAD_REGISTER_IMM_MAX_REGS) == 0) {
+			u32 n_lri = min_t(u32,
+					  n_regs - i,
+					  MI_LOAD_REGISTER_IMM_MAX_REGS);
+
+			*cs++ = MI_LOAD_REGISTER_IMM(n_lri);
+		}
+		*cs++ = i915_mmio_reg_offset(reg_data[i].addr);
+		*cs++ = reg_data[i].value;
+	}
+
+	return cs;
+}
+
+static int num_lri_dwords(int num_regs)
+{
+	int count = 0;
+
+	if (num_regs > 0) {
+		count += DIV_ROUND_UP(num_regs, MI_LOAD_REGISTER_IMM_MAX_REGS);
+		count += num_regs * 2;
+	}
+
+	return count;
+}
+
+static struct i915_oa_config_bo *
+alloc_oa_config_buffer(struct i915_perf_stream *stream,
+		       struct i915_oa_config *oa_config)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_oa_config_bo *oa_bo;
+	struct i915_gem_ww_ctx ww;
+	size_t config_length = 0;
+	u32 *cs;
+	int err;
+
+	oa_bo = kzalloc(sizeof(*oa_bo), GFP_KERNEL);
+	if (!oa_bo)
+		return ERR_PTR(-ENOMEM);
+
+	config_length += num_lri_dwords(oa_config->mux_regs_len);
+	config_length += num_lri_dwords(oa_config->b_counter_regs_len);
+	config_length += num_lri_dwords(oa_config->flex_regs_len);
+	config_length += 3; /* MI_BATCH_BUFFER_START */
+	config_length = ALIGN(sizeof(u32) * config_length, I915_GTT_PAGE_SIZE);
+
+	obj = i915_gem_object_create_shmem(stream->perf->i915, config_length);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto err_free;
+	}
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	err = i915_gem_object_lock(obj, &ww);
+	if (err)
+		goto out_ww;
+
+	cs = i915_gem_object_pin_map(obj, I915_MAP_WB);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto out_ww;
+	}
+
+	cs = write_cs_mi_lri(cs,
+			     oa_config->mux_regs,
+			     oa_config->mux_regs_len);
+	cs = write_cs_mi_lri(cs,
+			     oa_config->b_counter_regs,
+			     oa_config->b_counter_regs_len);
+	cs = write_cs_mi_lri(cs,
+			     oa_config->flex_regs,
+			     oa_config->flex_regs_len);
+
+	/* Jump into the active wait. */
+	*cs++ = (GRAPHICS_VER(stream->perf->i915) < 8 ?
+		 MI_BATCH_BUFFER_START :
+		 MI_BATCH_BUFFER_START_GEN8);
+	*cs++ = i915_ggtt_offset(stream->noa_wait);
+	*cs++ = 0;
+
+	i915_gem_object_flush_map(obj);
+	__i915_gem_object_release_map(obj);
+
+	oa_bo->vma = i915_vma_instance(obj,
+				       &stream->engine->gt->ggtt->vm,
+				       NULL);
+	if (IS_ERR(oa_bo->vma)) {
+		err = PTR_ERR(oa_bo->vma);
+		goto out_ww;
+	}
+
+	oa_bo->oa_config = i915_oa_config_get(oa_config);
+	llist_add(&oa_bo->node, &stream->oa_config_bos);
+
+out_ww:
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+
+	if (err)
+		i915_gem_object_put(obj);
+err_free:
+	if (err) {
+		kfree(oa_bo);
+		return ERR_PTR(err);
+	}
+	return oa_bo;
+}
+
+static struct i915_vma *
+get_oa_vma(struct i915_perf_stream *stream, struct i915_oa_config *oa_config)
+{
+	struct i915_oa_config_bo *oa_bo;
+
+	/*
+	 * Look for the buffer in the already allocated BOs attached
+	 * to the stream.
+	 */
+	llist_for_each_entry(oa_bo, stream->oa_config_bos.first, node) {
+		if (oa_bo->oa_config == oa_config &&
+		    memcmp(oa_bo->oa_config->uuid,
+			   oa_config->uuid,
+			   sizeof(oa_config->uuid)) == 0)
+			goto out;
+	}
+
+	oa_bo = alloc_oa_config_buffer(stream, oa_config);
+	if (IS_ERR(oa_bo))
+		return ERR_CAST(oa_bo);
+
+out:
+	return i915_vma_get(oa_bo->vma);
+}
+
+static int
+emit_oa_config(struct i915_perf_stream *stream,
+	       struct i915_oa_config *oa_config,
+	       struct intel_context *ce,
+	       struct i915_active *active)
+{
+	struct i915_request *rq;
+	struct i915_vma *vma;
+	struct i915_gem_ww_ctx ww;
+	int err;
+
+	vma = get_oa_vma(stream, oa_config);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	err = i915_gem_object_lock(vma->obj, &ww);
+	if (err)
+		goto err;
+
+	err = i915_vma_pin_ww(vma, &ww, 0, 0, PIN_GLOBAL | PIN_HIGH);
+	if (err)
+		goto err;
+
+	intel_engine_pm_get(ce->engine);
+	rq = i915_request_create(ce);
+	intel_engine_pm_put(ce->engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_vma_unpin;
+	}
+
+	if (!IS_ERR_OR_NULL(active)) {
+		/* After all individual context modifications */
+		err = i915_request_await_active(rq, active,
+						I915_ACTIVE_AWAIT_ACTIVE);
+		if (err)
+			goto err_add_request;
+
+		err = i915_active_add_request(active, rq);
+		if (err)
+			goto err_add_request;
+	}
+
+	err = i915_request_await_object(rq, vma->obj, 0);
+	if (!err)
+		err = i915_vma_move_to_active(vma, rq, 0);
+	if (err)
+		goto err_add_request;
+
+	err = rq->engine->emit_bb_start(rq,
+					vma->node.start, 0,
+					I915_DISPATCH_SECURE);
+	if (err)
+		goto err_add_request;
+
+err_add_request:
+	i915_request_add(rq);
+err_vma_unpin:
+	i915_vma_unpin(vma);
+err:
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+
+	i915_gem_ww_ctx_fini(&ww);
+	i915_vma_put(vma);
+	return err;
+}
+
+static struct intel_context *oa_context(struct i915_perf_stream *stream)
+{
+	return stream->pinned_ctx ?: stream->engine->kernel_context;
+}
+
+static int
+hsw_enable_metric_set(struct i915_perf_stream *stream,
+		      struct i915_active *active)
+{
+	struct intel_uncore *uncore = stream->uncore;
+
+	/*
+	 * PRM:
+	 *
+	 * OA unit is using “crclk” for its functionality. When trunk
+	 * level clock gating takes place, OA clock would be gated,
+	 * unable to count the events from non-render clock domain.
+	 * Render clock gating must be disabled when OA is enabled to
+	 * count the events from non-render domain. Unit level clock
+	 * gating for RCS should also be disabled.
+	 */
+	intel_uncore_rmw(uncore, GEN7_MISCCPCTL,
+			 GEN7_DOP_CLOCK_GATE_ENABLE, 0);
+	intel_uncore_rmw(uncore, GEN6_UCGCTL1,
+			 0, GEN6_CSUNIT_CLOCK_GATE_DISABLE);
+
+	return emit_oa_config(stream,
+			      stream->oa_config, oa_context(stream),
+			      active);
+}
+
+static void hsw_disable_metric_set(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+
+	intel_uncore_rmw(uncore, GEN6_UCGCTL1,
+			 GEN6_CSUNIT_CLOCK_GATE_DISABLE, 0);
+	intel_uncore_rmw(uncore, GEN7_MISCCPCTL,
+			 0, GEN7_DOP_CLOCK_GATE_ENABLE);
+
+	intel_uncore_rmw(uncore, GDT_CHICKEN_BITS, GT_NOA_ENABLE, 0);
+}
+
+static u32 oa_config_flex_reg(const struct i915_oa_config *oa_config,
+			      i915_reg_t reg)
+{
+	u32 mmio = i915_mmio_reg_offset(reg);
+	int i;
+
+	/*
+	 * This arbitrary default will select the 'EU FPU0 Pipeline
+	 * Active' event. In the future it's anticipated that there
+	 * will be an explicit 'No Event' we can select, but not yet...
+	 */
+	if (!oa_config)
+		return 0;
+
+	for (i = 0; i < oa_config->flex_regs_len; i++) {
+		if (i915_mmio_reg_offset(oa_config->flex_regs[i].addr) == mmio)
+			return oa_config->flex_regs[i].value;
+	}
+
+	return 0;
+}
+/*
+ * NB: It must always remain pointer safe to run this even if the OA unit
+ * has been disabled.
+ *
+ * It's fine to put out-of-date values into these per-context registers
+ * in the case that the OA unit has been disabled.
+ */
+static void
+gen8_update_reg_state_unlocked(const struct intel_context *ce,
+			       const struct i915_perf_stream *stream)
+{
+	u32 ctx_oactxctrl = stream->perf->ctx_oactxctrl_offset;
+	u32 ctx_flexeu0 = stream->perf->ctx_flexeu0_offset;
+	/* The MMIO offsets for Flex EU registers aren't contiguous */
+	static const i915_reg_t flex_regs[] = {
+		EU_PERF_CNTL0,
+		EU_PERF_CNTL1,
+		EU_PERF_CNTL2,
+		EU_PERF_CNTL3,
+		EU_PERF_CNTL4,
+		EU_PERF_CNTL5,
+		EU_PERF_CNTL6,
+	};
+	u32 *reg_state = ce->lrc_reg_state;
+	int i;
+
+	reg_state[ctx_oactxctrl + 1] =
+		(stream->period_exponent << GEN8_OA_TIMER_PERIOD_SHIFT) |
+		(stream->periodic ? GEN8_OA_TIMER_ENABLE : 0) |
+		GEN8_OA_COUNTER_RESUME;
+
+	for (i = 0; i < ARRAY_SIZE(flex_regs); i++)
+		reg_state[ctx_flexeu0 + i * 2 + 1] =
+			oa_config_flex_reg(stream->oa_config, flex_regs[i]);
+}
+
+struct flex {
+	i915_reg_t reg;
+	u32 offset;
+	u32 value;
+};
+
+static int
+gen8_store_flex(struct i915_request *rq,
+		struct intel_context *ce,
+		const struct flex *flex, unsigned int count)
+{
+	u32 offset;
+	u32 *cs;
+
+	cs = intel_ring_begin(rq, 4 * count);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	offset = i915_ggtt_offset(ce->state) + LRC_STATE_OFFSET;
+	do {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*cs++ = offset + flex->offset * sizeof(u32);
+		*cs++ = 0;
+		*cs++ = flex->value;
+	} while (flex++, --count);
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int
+gen8_load_flex(struct i915_request *rq,
+	       struct intel_context *ce,
+	       const struct flex *flex, unsigned int count)
+{
+	u32 *cs;
+
+	GEM_BUG_ON(!count || count > 63);
+
+	cs = intel_ring_begin(rq, 2 * count + 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(count);
+	do {
+		*cs++ = i915_mmio_reg_offset(flex->reg);
+		*cs++ = flex->value;
+	} while (flex++, --count);
+	*cs++ = MI_NOOP;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static int gen8_modify_context(struct intel_context *ce,
+			       const struct flex *flex, unsigned int count)
+{
+	struct i915_request *rq;
+	int err;
+
+	rq = intel_engine_create_kernel_request(ce->engine);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	/* Serialise with the remote context */
+	err = intel_context_prepare_remote_request(ce, rq);
+	if (err == 0)
+		err = gen8_store_flex(rq, ce, flex, count);
+
+	i915_request_add(rq);
+	return err;
+}
+
+static int
+gen8_modify_self(struct intel_context *ce,
+		 const struct flex *flex, unsigned int count,
+		 struct i915_active *active)
+{
+	struct i915_request *rq;
+	int err;
+
+	intel_engine_pm_get(ce->engine);
+	rq = i915_request_create(ce);
+	intel_engine_pm_put(ce->engine);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	if (!IS_ERR_OR_NULL(active)) {
+		err = i915_active_add_request(active, rq);
+		if (err)
+			goto err_add_request;
+	}
+
+	err = gen8_load_flex(rq, ce, flex, count);
+	if (err)
+		goto err_add_request;
+
+err_add_request:
+	i915_request_add(rq);
+	return err;
+}
+
+static int gen8_configure_context(struct i915_gem_context *ctx,
+				  struct flex *flex, unsigned int count)
+{
+	struct i915_gem_engines_iter it;
+	struct intel_context *ce;
+	int err = 0;
+
+	for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it) {
+		GEM_BUG_ON(ce == ce->engine->kernel_context);
+
+		if (ce->engine->class != RENDER_CLASS)
+			continue;
+
+		/* Otherwise OA settings will be set upon first use */
+		if (!intel_context_pin_if_active(ce))
+			continue;
+
+		flex->value = intel_sseu_make_rpcs(ce->engine->gt, &ce->sseu);
+		err = gen8_modify_context(ce, flex, count);
+
+		intel_context_unpin(ce);
+		if (err)
+			break;
+	}
+	i915_gem_context_unlock_engines(ctx);
+
+	return err;
+}
+
+static int gen12_configure_oar_context(struct i915_perf_stream *stream,
+				       struct i915_active *active)
+{
+	int err;
+	struct intel_context *ce = stream->pinned_ctx;
+	u32 format = stream->oa_buffer.format;
+	struct flex regs_context[] = {
+		{
+			GEN8_OACTXCONTROL,
+			stream->perf->ctx_oactxctrl_offset + 1,
+			active ? GEN8_OA_COUNTER_RESUME : 0,
+		},
+	};
+	/* Offsets in regs_lri are not used since this configuration is only
+	 * applied using LRI. Initialize the correct offsets for posterity.
+	 */
+#define GEN12_OAR_OACONTROL_OFFSET 0x5B0
+	struct flex regs_lri[] = {
+		{
+			GEN12_OAR_OACONTROL,
+			GEN12_OAR_OACONTROL_OFFSET + 1,
+			(format << GEN12_OAR_OACONTROL_COUNTER_FORMAT_SHIFT) |
+			(active ? GEN12_OAR_OACONTROL_COUNTER_ENABLE : 0)
+		},
+		{
+			RING_CONTEXT_CONTROL(ce->engine->mmio_base),
+			CTX_CONTEXT_CONTROL,
+			_MASKED_FIELD(GEN12_CTX_CTRL_OAR_CONTEXT_ENABLE,
+				      active ?
+				      GEN12_CTX_CTRL_OAR_CONTEXT_ENABLE :
+				      0)
+		},
+	};
+
+	/* Modify the context image of pinned context with regs_context*/
+	err = intel_context_lock_pinned(ce);
+	if (err)
+		return err;
+
+	err = gen8_modify_context(ce, regs_context, ARRAY_SIZE(regs_context));
+	intel_context_unlock_pinned(ce);
+	if (err)
+		return err;
+
+	/* Apply regs_lri using LRI with pinned context */
+	return gen8_modify_self(ce, regs_lri, ARRAY_SIZE(regs_lri), active);
+}
+
+/*
+ * Manages updating the per-context aspects of the OA stream
+ * configuration across all contexts.
+ *
+ * The awkward consideration here is that OACTXCONTROL controls the
+ * exponent for periodic sampling which is primarily used for system
+ * wide profiling where we'd like a consistent sampling period even in
+ * the face of context switches.
+ *
+ * Our approach of updating the register state context (as opposed to
+ * say using a workaround batch buffer) ensures that the hardware
+ * won't automatically reload an out-of-date timer exponent even
+ * transiently before a WA BB could be parsed.
+ *
+ * This function needs to:
+ * - Ensure the currently running context's per-context OA state is
+ *   updated
+ * - Ensure that all existing contexts will have the correct per-context
+ *   OA state if they are scheduled for use.
+ * - Ensure any new contexts will be initialized with the correct
+ *   per-context OA state.
+ *
+ * Note: it's only the RCS/Render context that has any OA state.
+ * Note: the first flex register passed must always be R_PWR_CLK_STATE
+ */
+static int
+oa_configure_all_contexts(struct i915_perf_stream *stream,
+			  struct flex *regs,
+			  size_t num_regs,
+			  struct i915_active *active)
+{
+	struct drm_i915_private *i915 = stream->perf->i915;
+	struct intel_engine_cs *engine;
+	struct i915_gem_context *ctx, *cn;
+	int err;
+
+	lockdep_assert_held(&stream->perf->lock);
+
+	/*
+	 * The OA register config is setup through the context image. This image
+	 * might be written to by the GPU on context switch (in particular on
+	 * lite-restore). This means we can't safely update a context's image,
+	 * if this context is scheduled/submitted to run on the GPU.
+	 *
+	 * We could emit the OA register config through the batch buffer but
+	 * this might leave small interval of time where the OA unit is
+	 * configured at an invalid sampling period.
+	 *
+	 * Note that since we emit all requests from a single ring, there
+	 * is still an implicit global barrier here that may cause a high
+	 * priority context to wait for an otherwise independent low priority
+	 * context. Contexts idle at the time of reconfiguration are not
+	 * trapped behind the barrier.
+	 */
+	spin_lock(&i915->gem.contexts.lock);
+	list_for_each_entry_safe(ctx, cn, &i915->gem.contexts.list, link) {
+		if (!kref_get_unless_zero(&ctx->ref))
+			continue;
+
+		spin_unlock(&i915->gem.contexts.lock);
+
+		err = gen8_configure_context(ctx, regs, num_regs);
+		if (err) {
+			i915_gem_context_put(ctx);
+			return err;
+		}
+
+		spin_lock(&i915->gem.contexts.lock);
+		list_safe_reset_next(ctx, cn, link);
+		i915_gem_context_put(ctx);
+	}
+	spin_unlock(&i915->gem.contexts.lock);
+
+	/*
+	 * After updating all other contexts, we need to modify ourselves.
+	 * If we don't modify the kernel_context, we do not get events while
+	 * idle.
+	 */
+	for_each_uabi_engine(engine, i915) {
+		struct intel_context *ce = engine->kernel_context;
+
+		if (engine->class != RENDER_CLASS)
+			continue;
+
+		regs[0].value = intel_sseu_make_rpcs(engine->gt, &ce->sseu);
+
+		err = gen8_modify_self(ce, regs, num_regs, active);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int
+gen12_configure_all_contexts(struct i915_perf_stream *stream,
+			     const struct i915_oa_config *oa_config,
+			     struct i915_active *active)
+{
+	struct flex regs[] = {
+		{
+			GEN8_R_PWR_CLK_STATE(RENDER_RING_BASE),
+			CTX_R_PWR_CLK_STATE,
+		},
+	};
+
+	return oa_configure_all_contexts(stream,
+					 regs, ARRAY_SIZE(regs),
+					 active);
+}
+
+static int
+lrc_configure_all_contexts(struct i915_perf_stream *stream,
+			   const struct i915_oa_config *oa_config,
+			   struct i915_active *active)
+{
+	/* The MMIO offsets for Flex EU registers aren't contiguous */
+	const u32 ctx_flexeu0 = stream->perf->ctx_flexeu0_offset;
+#define ctx_flexeuN(N) (ctx_flexeu0 + 2 * (N) + 1)
+	struct flex regs[] = {
+		{
+			GEN8_R_PWR_CLK_STATE(RENDER_RING_BASE),
+			CTX_R_PWR_CLK_STATE,
+		},
+		{
+			GEN8_OACTXCONTROL,
+			stream->perf->ctx_oactxctrl_offset + 1,
+		},
+		{ EU_PERF_CNTL0, ctx_flexeuN(0) },
+		{ EU_PERF_CNTL1, ctx_flexeuN(1) },
+		{ EU_PERF_CNTL2, ctx_flexeuN(2) },
+		{ EU_PERF_CNTL3, ctx_flexeuN(3) },
+		{ EU_PERF_CNTL4, ctx_flexeuN(4) },
+		{ EU_PERF_CNTL5, ctx_flexeuN(5) },
+		{ EU_PERF_CNTL6, ctx_flexeuN(6) },
+	};
+#undef ctx_flexeuN
+	int i;
+
+	regs[1].value =
+		(stream->period_exponent << GEN8_OA_TIMER_PERIOD_SHIFT) |
+		(stream->periodic ? GEN8_OA_TIMER_ENABLE : 0) |
+		GEN8_OA_COUNTER_RESUME;
+
+	for (i = 2; i < ARRAY_SIZE(regs); i++)
+		regs[i].value = oa_config_flex_reg(oa_config, regs[i].reg);
+
+	return oa_configure_all_contexts(stream,
+					 regs, ARRAY_SIZE(regs),
+					 active);
+}
+
+static int
+gen8_enable_metric_set(struct i915_perf_stream *stream,
+		       struct i915_active *active)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	struct i915_oa_config *oa_config = stream->oa_config;
+	int ret;
+
+	/*
+	 * We disable slice/unslice clock ratio change reports on SKL since
+	 * they are too noisy. The HW generates a lot of redundant reports
+	 * where the ratio hasn't really changed causing a lot of redundant
+	 * work to processes and increasing the chances we'll hit buffer
+	 * overruns.
+	 *
+	 * Although we don't currently use the 'disable overrun' OABUFFER
+	 * feature it's worth noting that clock ratio reports have to be
+	 * disabled before considering to use that feature since the HW doesn't
+	 * correctly block these reports.
+	 *
+	 * Currently none of the high-level metrics we have depend on knowing
+	 * this ratio to normalize.
+	 *
+	 * Note: This register is not power context saved and restored, but
+	 * that's OK considering that we disable RC6 while the OA unit is
+	 * enabled.
+	 *
+	 * The _INCLUDE_CLK_RATIO bit allows the slice/unslice frequency to
+	 * be read back from automatically triggered reports, as part of the
+	 * RPT_ID field.
+	 */
+	if (IS_GRAPHICS_VER(stream->perf->i915, 9, 11)) {
+		intel_uncore_write(uncore, GEN8_OA_DEBUG,
+				   _MASKED_BIT_ENABLE(GEN9_OA_DEBUG_DISABLE_CLK_RATIO_REPORTS |
+						      GEN9_OA_DEBUG_INCLUDE_CLK_RATIO));
+	}
+
+	/*
+	 * Update all contexts prior writing the mux configurations as we need
+	 * to make sure all slices/subslices are ON before writing to NOA
+	 * registers.
+	 */
+	ret = lrc_configure_all_contexts(stream, oa_config, active);
+	if (ret)
+		return ret;
+
+	return emit_oa_config(stream,
+			      stream->oa_config, oa_context(stream),
+			      active);
+}
+
+static u32 oag_report_ctx_switches(const struct i915_perf_stream *stream)
+{
+	return _MASKED_FIELD(GEN12_OAG_OA_DEBUG_DISABLE_CTX_SWITCH_REPORTS,
+			     (stream->sample_flags & SAMPLE_OA_REPORT) ?
+			     0 : GEN12_OAG_OA_DEBUG_DISABLE_CTX_SWITCH_REPORTS);
+}
+
+static int
+gen12_enable_metric_set(struct i915_perf_stream *stream,
+			struct i915_active *active)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	struct i915_oa_config *oa_config = stream->oa_config;
+	bool periodic = stream->periodic;
+	u32 period_exponent = stream->period_exponent;
+	int ret;
+
+	intel_uncore_write(uncore, GEN12_OAG_OA_DEBUG,
+			   /* Disable clk ratio reports, like previous Gens. */
+			   _MASKED_BIT_ENABLE(GEN12_OAG_OA_DEBUG_DISABLE_CLK_RATIO_REPORTS |
+					      GEN12_OAG_OA_DEBUG_INCLUDE_CLK_RATIO) |
+			   /*
+			    * If the user didn't require OA reports, instruct
+			    * the hardware not to emit ctx switch reports.
+			    */
+			   oag_report_ctx_switches(stream));
+
+	intel_uncore_write(uncore, GEN12_OAG_OAGLBCTXCTRL, periodic ?
+			   (GEN12_OAG_OAGLBCTXCTRL_COUNTER_RESUME |
+			    GEN12_OAG_OAGLBCTXCTRL_TIMER_ENABLE |
+			    (period_exponent << GEN12_OAG_OAGLBCTXCTRL_TIMER_PERIOD_SHIFT))
+			    : 0);
+
+	/*
+	 * Update all contexts prior writing the mux configurations as we need
+	 * to make sure all slices/subslices are ON before writing to NOA
+	 * registers.
+	 */
+	ret = gen12_configure_all_contexts(stream, oa_config, active);
+	if (ret)
+		return ret;
+
+	/*
+	 * For Gen12, performance counters are context
+	 * saved/restored. Only enable it for the context that
+	 * requested this.
+	 */
+	if (stream->ctx) {
+		ret = gen12_configure_oar_context(stream, active);
+		if (ret)
+			return ret;
+	}
+
+	return emit_oa_config(stream,
+			      stream->oa_config, oa_context(stream),
+			      active);
+}
+
+static void gen8_disable_metric_set(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+
+	/* Reset all contexts' slices/subslices configurations. */
+	lrc_configure_all_contexts(stream, NULL, NULL);
+
+	intel_uncore_rmw(uncore, GDT_CHICKEN_BITS, GT_NOA_ENABLE, 0);
+}
+
+static void gen11_disable_metric_set(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+
+	/* Reset all contexts' slices/subslices configurations. */
+	lrc_configure_all_contexts(stream, NULL, NULL);
+
+	/* Make sure we disable noa to save power. */
+	intel_uncore_rmw(uncore, RPM_CONFIG1, GEN10_GT_NOA_ENABLE, 0);
+}
+
+static void gen12_disable_metric_set(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+
+	/* Reset all contexts' slices/subslices configurations. */
+	gen12_configure_all_contexts(stream, NULL, NULL);
+
+	/* disable the context save/restore or OAR counters */
+	if (stream->ctx)
+		gen12_configure_oar_context(stream, NULL);
+
+	/* Make sure we disable noa to save power. */
+	intel_uncore_rmw(uncore, RPM_CONFIG1, GEN10_GT_NOA_ENABLE, 0);
+}
+
+static void gen7_oa_enable(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	struct i915_gem_context *ctx = stream->ctx;
+	u32 ctx_id = stream->specific_ctx_id;
+	bool periodic = stream->periodic;
+	u32 period_exponent = stream->period_exponent;
+	u32 report_format = stream->oa_buffer.format;
+
+	/*
+	 * Reset buf pointers so we don't forward reports from before now.
+	 *
+	 * Think carefully if considering trying to avoid this, since it
+	 * also ensures status flags and the buffer itself are cleared
+	 * in error paths, and we have checks for invalid reports based
+	 * on the assumption that certain fields are written to zeroed
+	 * memory which this helps maintains.
+	 */
+	gen7_init_oa_buffer(stream);
+
+	intel_uncore_write(uncore, GEN7_OACONTROL,
+			   (ctx_id & GEN7_OACONTROL_CTX_MASK) |
+			   (period_exponent <<
+			    GEN7_OACONTROL_TIMER_PERIOD_SHIFT) |
+			   (periodic ? GEN7_OACONTROL_TIMER_ENABLE : 0) |
+			   (report_format << GEN7_OACONTROL_FORMAT_SHIFT) |
+			   (ctx ? GEN7_OACONTROL_PER_CTX_ENABLE : 0) |
+			   GEN7_OACONTROL_ENABLE);
+}
+
+static void gen8_oa_enable(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	u32 report_format = stream->oa_buffer.format;
+
+	/*
+	 * Reset buf pointers so we don't forward reports from before now.
+	 *
+	 * Think carefully if considering trying to avoid this, since it
+	 * also ensures status flags and the buffer itself are cleared
+	 * in error paths, and we have checks for invalid reports based
+	 * on the assumption that certain fields are written to zeroed
+	 * memory which this helps maintains.
+	 */
+	gen8_init_oa_buffer(stream);
+
+	/*
+	 * Note: we don't rely on the hardware to perform single context
+	 * filtering and instead filter on the cpu based on the context-id
+	 * field of reports
+	 */
+	intel_uncore_write(uncore, GEN8_OACONTROL,
+			   (report_format << GEN8_OA_REPORT_FORMAT_SHIFT) |
+			   GEN8_OA_COUNTER_ENABLE);
+}
+
+static void gen12_oa_enable(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+	u32 report_format = stream->oa_buffer.format;
+
+	/*
+	 * If we don't want OA reports from the OA buffer, then we don't even
+	 * need to program the OAG unit.
+	 */
+	if (!(stream->sample_flags & SAMPLE_OA_REPORT))
+		return;
+
+	gen12_init_oa_buffer(stream);
+
+	intel_uncore_write(uncore, GEN12_OAG_OACONTROL,
+			   (report_format << GEN12_OAG_OACONTROL_OA_COUNTER_FORMAT_SHIFT) |
+			   GEN12_OAG_OACONTROL_OA_COUNTER_ENABLE);
+}
+
+/**
+ * i915_oa_stream_enable - handle `I915_PERF_IOCTL_ENABLE` for OA stream
+ * @stream: An i915 perf stream opened for OA metrics
+ *
+ * [Re]enables hardware periodic sampling according to the period configured
+ * when opening the stream. This also starts a hrtimer that will periodically
+ * check for data in the circular OA buffer for notifying userspace (e.g.
+ * during a read() or poll()).
+ */
+static void i915_oa_stream_enable(struct i915_perf_stream *stream)
+{
+	stream->pollin = false;
+
+	stream->perf->ops.oa_enable(stream);
+
+	if (stream->sample_flags & SAMPLE_OA_REPORT)
+		hrtimer_start(&stream->poll_check_timer,
+			      ns_to_ktime(stream->poll_oa_period),
+			      HRTIMER_MODE_REL_PINNED);
+}
+
+static void gen7_oa_disable(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+
+	intel_uncore_write(uncore, GEN7_OACONTROL, 0);
+	if (intel_wait_for_register(uncore,
+				    GEN7_OACONTROL, GEN7_OACONTROL_ENABLE, 0,
+				    50))
+		drm_err(&stream->perf->i915->drm,
+			"wait for OA to be disabled timed out\n");
+}
+
+static void gen8_oa_disable(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+
+	intel_uncore_write(uncore, GEN8_OACONTROL, 0);
+	if (intel_wait_for_register(uncore,
+				    GEN8_OACONTROL, GEN8_OA_COUNTER_ENABLE, 0,
+				    50))
+		drm_err(&stream->perf->i915->drm,
+			"wait for OA to be disabled timed out\n");
+}
+
+static void gen12_oa_disable(struct i915_perf_stream *stream)
+{
+	struct intel_uncore *uncore = stream->uncore;
+
+	intel_uncore_write(uncore, GEN12_OAG_OACONTROL, 0);
+	if (intel_wait_for_register(uncore,
+				    GEN12_OAG_OACONTROL,
+				    GEN12_OAG_OACONTROL_OA_COUNTER_ENABLE, 0,
+				    50))
+		drm_err(&stream->perf->i915->drm,
+			"wait for OA to be disabled timed out\n");
+
+	intel_uncore_write(uncore, GEN12_OA_TLB_INV_CR, 1);
+	if (intel_wait_for_register(uncore,
+				    GEN12_OA_TLB_INV_CR,
+				    1, 0,
+				    50))
+		drm_err(&stream->perf->i915->drm,
+			"wait for OA tlb invalidate timed out\n");
+}
+
+/**
+ * i915_oa_stream_disable - handle `I915_PERF_IOCTL_DISABLE` for OA stream
+ * @stream: An i915 perf stream opened for OA metrics
+ *
+ * Stops the OA unit from periodically writing counter reports into the
+ * circular OA buffer. This also stops the hrtimer that periodically checks for
+ * data in the circular OA buffer, for notifying userspace.
+ */
+static void i915_oa_stream_disable(struct i915_perf_stream *stream)
+{
+	stream->perf->ops.oa_disable(stream);
+
+	if (stream->sample_flags & SAMPLE_OA_REPORT)
+		hrtimer_cancel(&stream->poll_check_timer);
+}
+
+static const struct i915_perf_stream_ops i915_oa_stream_ops = {
+	.destroy = i915_oa_stream_destroy,
+	.enable = i915_oa_stream_enable,
+	.disable = i915_oa_stream_disable,
+	.wait_unlocked = i915_oa_wait_unlocked,
+	.poll_wait = i915_oa_poll_wait,
+	.read = i915_oa_read,
+};
+
+static int i915_perf_stream_enable_sync(struct i915_perf_stream *stream)
+{
+	struct i915_active *active;
+	int err;
+
+	active = i915_active_create();
+	if (!active)
+		return -ENOMEM;
+
+	err = stream->perf->ops.enable_metric_set(stream, active);
+	if (err == 0)
+		__i915_active_wait(active, TASK_UNINTERRUPTIBLE);
+
+	i915_active_put(active);
+	return err;
+}
+
+static void
+get_default_sseu_config(struct intel_sseu *out_sseu,
+			struct intel_engine_cs *engine)
+{
+	const struct sseu_dev_info *devinfo_sseu = &engine->gt->info.sseu;
+
+	*out_sseu = intel_sseu_from_device_info(devinfo_sseu);
+
+	if (GRAPHICS_VER(engine->i915) == 11) {
+		/*
+		 * We only need subslice count so it doesn't matter which ones
+		 * we select - just turn off low bits in the amount of half of
+		 * all available subslices per slice.
+		 */
+		out_sseu->subslice_mask =
+			~(~0 << (hweight8(out_sseu->subslice_mask) / 2));
+		out_sseu->slice_mask = 0x1;
+	}
+}
+
+static int
+get_sseu_config(struct intel_sseu *out_sseu,
+		struct intel_engine_cs *engine,
+		const struct drm_i915_gem_context_param_sseu *drm_sseu)
+{
+	if (drm_sseu->engine.engine_class != engine->uabi_class ||
+	    drm_sseu->engine.engine_instance != engine->uabi_instance)
+		return -EINVAL;
+
+	return i915_gem_user_to_context_sseu(engine->gt, drm_sseu, out_sseu);
+}
+
+/**
+ * i915_oa_stream_init - validate combined props for OA stream and init
+ * @stream: An i915 perf stream
+ * @param: The open parameters passed to `DRM_I915_PERF_OPEN`
+ * @props: The property state that configures stream (individually validated)
+ *
+ * While read_properties_unlocked() validates properties in isolation it
+ * doesn't ensure that the combination necessarily makes sense.
+ *
+ * At this point it has been determined that userspace wants a stream of
+ * OA metrics, but still we need to further validate the combined
+ * properties are OK.
+ *
+ * If the configuration makes sense then we can allocate memory for
+ * a circular OA buffer and apply the requested metric set configuration.
+ *
+ * Returns: zero on success or a negative error code.
+ */
+static int i915_oa_stream_init(struct i915_perf_stream *stream,
+			       struct drm_i915_perf_open_param *param,
+			       struct perf_open_properties *props)
+{
+	struct drm_i915_private *i915 = stream->perf->i915;
+	struct i915_perf *perf = stream->perf;
+	int format_size;
+	int ret;
+
+	if (!props->engine) {
+		drm_dbg(&stream->perf->i915->drm,
+			"OA engine not specified\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * If the sysfs metrics/ directory wasn't registered for some
+	 * reason then don't let userspace try their luck with config
+	 * IDs
+	 */
+	if (!perf->metrics_kobj) {
+		drm_dbg(&stream->perf->i915->drm,
+			"OA metrics weren't advertised via sysfs\n");
+		return -EINVAL;
+	}
+
+	if (!(props->sample_flags & SAMPLE_OA_REPORT) &&
+	    (GRAPHICS_VER(perf->i915) < 12 || !stream->ctx)) {
+		drm_dbg(&stream->perf->i915->drm,
+			"Only OA report sampling supported\n");
+		return -EINVAL;
+	}
+
+	if (!perf->ops.enable_metric_set) {
+		drm_dbg(&stream->perf->i915->drm,
+			"OA unit not supported\n");
+		return -ENODEV;
+	}
+
+	/*
+	 * To avoid the complexity of having to accurately filter
+	 * counter reports and marshal to the appropriate client
+	 * we currently only allow exclusive access
+	 */
+	if (perf->exclusive_stream) {
+		drm_dbg(&stream->perf->i915->drm,
+			"OA unit already in use\n");
+		return -EBUSY;
+	}
+
+	if (!props->oa_format) {
+		drm_dbg(&stream->perf->i915->drm,
+			"OA report format not specified\n");
+		return -EINVAL;
+	}
+
+	stream->engine = props->engine;
+	stream->uncore = stream->engine->gt->uncore;
+
+	stream->sample_size = sizeof(struct drm_i915_perf_record_header);
+
+	format_size = perf->oa_formats[props->oa_format].size;
+
+	stream->sample_flags = props->sample_flags;
+	stream->sample_size += format_size;
+
+	stream->oa_buffer.format_size = format_size;
+	if (drm_WARN_ON(&i915->drm, stream->oa_buffer.format_size == 0))
+		return -EINVAL;
+
+	stream->hold_preemption = props->hold_preemption;
+
+	stream->oa_buffer.format =
+		perf->oa_formats[props->oa_format].format;
+
+	stream->periodic = props->oa_periodic;
+	if (stream->periodic)
+		stream->period_exponent = props->oa_period_exponent;
+
+	if (stream->ctx) {
+		ret = oa_get_render_ctx_id(stream);
+		if (ret) {
+			drm_dbg(&stream->perf->i915->drm,
+				"Invalid context id to filter with\n");
+			return ret;
+		}
+	}
+
+	ret = alloc_noa_wait(stream);
+	if (ret) {
+		drm_dbg(&stream->perf->i915->drm,
+			"Unable to allocate NOA wait batch buffer\n");
+		goto err_noa_wait_alloc;
+	}
+
+	stream->oa_config = i915_perf_get_oa_config(perf, props->metrics_set);
+	if (!stream->oa_config) {
+		drm_dbg(&stream->perf->i915->drm,
+			"Invalid OA config id=%i\n", props->metrics_set);
+		ret = -EINVAL;
+		goto err_config;
+	}
+
+	/* PRM - observability performance counters:
+	 *
+	 *   OACONTROL, performance counter enable, note:
+	 *
+	 *   "When this bit is set, in order to have coherent counts,
+	 *   RC6 power state and trunk clock gating must be disabled.
+	 *   This can be achieved by programming MMIO registers as
+	 *   0xA094=0 and 0xA090[31]=1"
+	 *
+	 *   In our case we are expecting that taking pm + FORCEWAKE
+	 *   references will effectively disable RC6.
+	 */
+	intel_engine_pm_get(stream->engine);
+	intel_uncore_forcewake_get(stream->uncore, FORCEWAKE_ALL);
+
+	ret = alloc_oa_buffer(stream);
+	if (ret)
+		goto err_oa_buf_alloc;
+
+	stream->ops = &i915_oa_stream_ops;
+
+	perf->sseu = props->sseu;
+	WRITE_ONCE(perf->exclusive_stream, stream);
+
+	ret = i915_perf_stream_enable_sync(stream);
+	if (ret) {
+		drm_dbg(&stream->perf->i915->drm,
+			"Unable to enable metric set\n");
+		goto err_enable;
+	}
+
+	drm_dbg(&stream->perf->i915->drm,
+		"opening stream oa config uuid=%s\n",
+		  stream->oa_config->uuid);
+
+	hrtimer_init(&stream->poll_check_timer,
+		     CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+	stream->poll_check_timer.function = oa_poll_check_timer_cb;
+	init_waitqueue_head(&stream->poll_wq);
+	spin_lock_init(&stream->oa_buffer.ptr_lock);
+
+	return 0;
+
+err_enable:
+	WRITE_ONCE(perf->exclusive_stream, NULL);
+	perf->ops.disable_metric_set(stream);
+
+	free_oa_buffer(stream);
+
+err_oa_buf_alloc:
+	free_oa_configs(stream);
+
+	intel_uncore_forcewake_put(stream->uncore, FORCEWAKE_ALL);
+	intel_engine_pm_put(stream->engine);
+
+err_config:
+	free_noa_wait(stream);
+
+err_noa_wait_alloc:
+	if (stream->ctx)
+		oa_put_render_ctx_id(stream);
+
+	return ret;
+}
+
+void i915_oa_init_reg_state(const struct intel_context *ce,
+			    const struct intel_engine_cs *engine)
+{
+	struct i915_perf_stream *stream;
+
+	if (engine->class != RENDER_CLASS)
+		return;
+
+	/* perf.exclusive_stream serialised by lrc_configure_all_contexts() */
+	stream = READ_ONCE(engine->i915->perf.exclusive_stream);
+	if (stream && GRAPHICS_VER(stream->perf->i915) < 12)
+		gen8_update_reg_state_unlocked(ce, stream);
+}
+
+/**
+ * i915_perf_read - handles read() FOP for i915 perf stream FDs
+ * @file: An i915 perf stream file
+ * @buf: destination buffer given by userspace
+ * @count: the number of bytes userspace wants to read
+ * @ppos: (inout) file seek position (unused)
+ *
+ * The entry point for handling a read() on a stream file descriptor from
+ * userspace. Most of the work is left to the i915_perf_read_locked() and
+ * &i915_perf_stream_ops->read but to save having stream implementations (of
+ * which we might have multiple later) we handle blocking read here.
+ *
+ * We can also consistently treat trying to read from a disabled stream
+ * as an IO error so implementations can assume the stream is enabled
+ * while reading.
+ *
+ * Returns: The number of bytes copied or a negative error code on failure.
+ */
+static ssize_t i915_perf_read(struct file *file,
+			      char __user *buf,
+			      size_t count,
+			      loff_t *ppos)
+{
+	struct i915_perf_stream *stream = file->private_data;
+	struct i915_perf *perf = stream->perf;
+	size_t offset = 0;
+	int ret;
+
+	/* To ensure it's handled consistently we simply treat all reads of a
+	 * disabled stream as an error. In particular it might otherwise lead
+	 * to a deadlock for blocking file descriptors...
+	 */
+	if (!stream->enabled || !(stream->sample_flags & SAMPLE_OA_REPORT))
+		return -EIO;
+
+	if (!(file->f_flags & O_NONBLOCK)) {
+		/* There's the small chance of false positives from
+		 * stream->ops->wait_unlocked.
+		 *
+		 * E.g. with single context filtering since we only wait until
+		 * oabuffer has >= 1 report we don't immediately know whether
+		 * any reports really belong to the current context
+		 */
+		do {
+			ret = stream->ops->wait_unlocked(stream);
+			if (ret)
+				return ret;
+
+			mutex_lock(&perf->lock);
+			ret = stream->ops->read(stream, buf, count, &offset);
+			mutex_unlock(&perf->lock);
+		} while (!offset && !ret);
+	} else {
+		mutex_lock(&perf->lock);
+		ret = stream->ops->read(stream, buf, count, &offset);
+		mutex_unlock(&perf->lock);
+	}
+
+	/* We allow the poll checking to sometimes report false positive EPOLLIN
+	 * events where we might actually report EAGAIN on read() if there's
+	 * not really any data available. In this situation though we don't
+	 * want to enter a busy loop between poll() reporting a EPOLLIN event
+	 * and read() returning -EAGAIN. Clearing the oa.pollin state here
+	 * effectively ensures we back off until the next hrtimer callback
+	 * before reporting another EPOLLIN event.
+	 * The exception to this is if ops->read() returned -ENOSPC which means
+	 * that more OA data is available than could fit in the user provided
+	 * buffer. In this case we want the next poll() call to not block.
+	 */
+	if (ret != -ENOSPC)
+		stream->pollin = false;
+
+	/* Possible values for ret are 0, -EFAULT, -ENOSPC, -EIO, ... */
+	return offset ?: (ret ?: -EAGAIN);
+}
+
+static enum hrtimer_restart oa_poll_check_timer_cb(struct hrtimer *hrtimer)
+{
+	struct i915_perf_stream *stream =
+		container_of(hrtimer, typeof(*stream), poll_check_timer);
+
+	if (oa_buffer_check_unlocked(stream)) {
+		stream->pollin = true;
+		wake_up(&stream->poll_wq);
+	}
+
+	hrtimer_forward_now(hrtimer,
+			    ns_to_ktime(stream->poll_oa_period));
+
+	return HRTIMER_RESTART;
+}
+
+/**
+ * i915_perf_poll_locked - poll_wait() with a suitable wait queue for stream
+ * @stream: An i915 perf stream
+ * @file: An i915 perf stream file
+ * @wait: poll() state table
+ *
+ * For handling userspace polling on an i915 perf stream, this calls through to
+ * &i915_perf_stream_ops->poll_wait to call poll_wait() with a wait queue that
+ * will be woken for new stream data.
+ *
+ * Note: The &perf->lock mutex has been taken to serialize
+ * with any non-file-operation driver hooks.
+ *
+ * Returns: any poll events that are ready without sleeping
+ */
+static __poll_t i915_perf_poll_locked(struct i915_perf_stream *stream,
+				      struct file *file,
+				      poll_table *wait)
+{
+	__poll_t events = 0;
+
+	stream->ops->poll_wait(stream, file, wait);
+
+	/* Note: we don't explicitly check whether there's something to read
+	 * here since this path may be very hot depending on what else
+	 * userspace is polling, or on the timeout in use. We rely solely on
+	 * the hrtimer/oa_poll_check_timer_cb to notify us when there are
+	 * samples to read.
+	 */
+	if (stream->pollin)
+		events |= EPOLLIN;
+
+	return events;
+}
+
+/**
+ * i915_perf_poll - call poll_wait() with a suitable wait queue for stream
+ * @file: An i915 perf stream file
+ * @wait: poll() state table
+ *
+ * For handling userspace polling on an i915 perf stream, this ensures
+ * poll_wait() gets called with a wait queue that will be woken for new stream
+ * data.
+ *
+ * Note: Implementation deferred to i915_perf_poll_locked()
+ *
+ * Returns: any poll events that are ready without sleeping
+ */
+static __poll_t i915_perf_poll(struct file *file, poll_table *wait)
+{
+	struct i915_perf_stream *stream = file->private_data;
+	struct i915_perf *perf = stream->perf;
+	__poll_t ret;
+
+	mutex_lock(&perf->lock);
+	ret = i915_perf_poll_locked(stream, file, wait);
+	mutex_unlock(&perf->lock);
+
+	return ret;
+}
+
+/**
+ * i915_perf_enable_locked - handle `I915_PERF_IOCTL_ENABLE` ioctl
+ * @stream: A disabled i915 perf stream
+ *
+ * [Re]enables the associated capture of data for this stream.
+ *
+ * If a stream was previously enabled then there's currently no intention
+ * to provide userspace any guarantee about the preservation of previously
+ * buffered data.
+ */
+static void i915_perf_enable_locked(struct i915_perf_stream *stream)
+{
+	if (stream->enabled)
+		return;
+
+	/* Allow stream->ops->enable() to refer to this */
+	stream->enabled = true;
+
+	if (stream->ops->enable)
+		stream->ops->enable(stream);
+
+	if (stream->hold_preemption)
+		intel_context_set_nopreempt(stream->pinned_ctx);
+}
+
+/**
+ * i915_perf_disable_locked - handle `I915_PERF_IOCTL_DISABLE` ioctl
+ * @stream: An enabled i915 perf stream
+ *
+ * Disables the associated capture of data for this stream.
+ *
+ * The intention is that disabling an re-enabling a stream will ideally be
+ * cheaper than destroying and re-opening a stream with the same configuration,
+ * though there are no formal guarantees about what state or buffered data
+ * must be retained between disabling and re-enabling a stream.
+ *
+ * Note: while a stream is disabled it's considered an error for userspace
+ * to attempt to read from the stream (-EIO).
+ */
+static void i915_perf_disable_locked(struct i915_perf_stream *stream)
+{
+	if (!stream->enabled)
+		return;
+
+	/* Allow stream->ops->disable() to refer to this */
+	stream->enabled = false;
+
+	if (stream->hold_preemption)
+		intel_context_clear_nopreempt(stream->pinned_ctx);
+
+	if (stream->ops->disable)
+		stream->ops->disable(stream);
+}
+
+static long i915_perf_config_locked(struct i915_perf_stream *stream,
+				    unsigned long metrics_set)
+{
+	struct i915_oa_config *config;
+	long ret = stream->oa_config->id;
+
+	config = i915_perf_get_oa_config(stream->perf, metrics_set);
+	if (!config)
+		return -EINVAL;
+
+	if (config != stream->oa_config) {
+		int err;
+
+		/*
+		 * If OA is bound to a specific context, emit the
+		 * reconfiguration inline from that context. The update
+		 * will then be ordered with respect to submission on that
+		 * context.
+		 *
+		 * When set globally, we use a low priority kernel context,
+		 * so it will effectively take effect when idle.
+		 */
+		err = emit_oa_config(stream, config, oa_context(stream), NULL);
+		if (!err)
+			config = xchg(&stream->oa_config, config);
+		else
+			ret = err;
+	}
+
+	i915_oa_config_put(config);
+
+	return ret;
+}
+
+/**
+ * i915_perf_ioctl_locked - support ioctl() usage with i915 perf stream FDs
+ * @stream: An i915 perf stream
+ * @cmd: the ioctl request
+ * @arg: the ioctl data
+ *
+ * Note: The &perf->lock mutex has been taken to serialize
+ * with any non-file-operation driver hooks.
+ *
+ * Returns: zero on success or a negative error code. Returns -EINVAL for
+ * an unknown ioctl request.
+ */
+static long i915_perf_ioctl_locked(struct i915_perf_stream *stream,
+				   unsigned int cmd,
+				   unsigned long arg)
+{
+	switch (cmd) {
+	case I915_PERF_IOCTL_ENABLE:
+		i915_perf_enable_locked(stream);
+		return 0;
+	case I915_PERF_IOCTL_DISABLE:
+		i915_perf_disable_locked(stream);
+		return 0;
+	case I915_PERF_IOCTL_CONFIG:
+		return i915_perf_config_locked(stream, arg);
+	}
+
+	return -EINVAL;
+}
+
+/**
+ * i915_perf_ioctl - support ioctl() usage with i915 perf stream FDs
+ * @file: An i915 perf stream file
+ * @cmd: the ioctl request
+ * @arg: the ioctl data
+ *
+ * Implementation deferred to i915_perf_ioctl_locked().
+ *
+ * Returns: zero on success or a negative error code. Returns -EINVAL for
+ * an unknown ioctl request.
+ */
+static long i915_perf_ioctl(struct file *file,
+			    unsigned int cmd,
+			    unsigned long arg)
+{
+	struct i915_perf_stream *stream = file->private_data;
+	struct i915_perf *perf = stream->perf;
+	long ret;
+
+	mutex_lock(&perf->lock);
+	ret = i915_perf_ioctl_locked(stream, cmd, arg);
+	mutex_unlock(&perf->lock);
+
+	return ret;
+}
+
+/**
+ * i915_perf_destroy_locked - destroy an i915 perf stream
+ * @stream: An i915 perf stream
+ *
+ * Frees all resources associated with the given i915 perf @stream, disabling
+ * any associated data capture in the process.
+ *
+ * Note: The &perf->lock mutex has been taken to serialize
+ * with any non-file-operation driver hooks.
+ */
+static void i915_perf_destroy_locked(struct i915_perf_stream *stream)
+{
+	if (stream->enabled)
+		i915_perf_disable_locked(stream);
+
+	if (stream->ops->destroy)
+		stream->ops->destroy(stream);
+
+	if (stream->ctx)
+		i915_gem_context_put(stream->ctx);
+
+	kfree(stream);
+}
+
+/**
+ * i915_perf_release - handles userspace close() of a stream file
+ * @inode: anonymous inode associated with file
+ * @file: An i915 perf stream file
+ *
+ * Cleans up any resources associated with an open i915 perf stream file.
+ *
+ * NB: close() can't really fail from the userspace point of view.
+ *
+ * Returns: zero on success or a negative error code.
+ */
+static int i915_perf_release(struct inode *inode, struct file *file)
+{
+	struct i915_perf_stream *stream = file->private_data;
+	struct i915_perf *perf = stream->perf;
+
+	mutex_lock(&perf->lock);
+	i915_perf_destroy_locked(stream);
+	mutex_unlock(&perf->lock);
+
+	/* Release the reference the perf stream kept on the driver. */
+	drm_dev_put(&perf->i915->drm);
+
+	return 0;
+}
+
+
+static const struct file_operations fops = {
+	.owner		= THIS_MODULE,
+	.llseek		= no_llseek,
+	.release	= i915_perf_release,
+	.poll		= i915_perf_poll,
+	.read		= i915_perf_read,
+	.unlocked_ioctl	= i915_perf_ioctl,
+	/* Our ioctl have no arguments, so it's safe to use the same function
+	 * to handle 32bits compatibility.
+	 */
+	.compat_ioctl   = i915_perf_ioctl,
+};
+
+
+/**
+ * i915_perf_open_ioctl_locked - DRM ioctl() for userspace to open a stream FD
+ * @perf: i915 perf instance
+ * @param: The open parameters passed to 'DRM_I915_PERF_OPEN`
+ * @props: individually validated u64 property value pairs
+ * @file: drm file
+ *
+ * See i915_perf_ioctl_open() for interface details.
+ *
+ * Implements further stream config validation and stream initialization on
+ * behalf of i915_perf_open_ioctl() with the &perf->lock mutex
+ * taken to serialize with any non-file-operation driver hooks.
+ *
+ * Note: at this point the @props have only been validated in isolation and
+ * it's still necessary to validate that the combination of properties makes
+ * sense.
+ *
+ * In the case where userspace is interested in OA unit metrics then further
+ * config validation and stream initialization details will be handled by
+ * i915_oa_stream_init(). The code here should only validate config state that
+ * will be relevant to all stream types / backends.
+ *
+ * Returns: zero on success or a negative error code.
+ */
+static int
+i915_perf_open_ioctl_locked(struct i915_perf *perf,
+			    struct drm_i915_perf_open_param *param,
+			    struct perf_open_properties *props,
+			    struct drm_file *file)
+{
+	struct i915_gem_context *specific_ctx = NULL;
+	struct i915_perf_stream *stream = NULL;
+	unsigned long f_flags = 0;
+	bool privileged_op = true;
+	int stream_fd;
+	int ret;
+
+	if (props->single_context) {
+		u32 ctx_handle = props->ctx_handle;
+		struct drm_i915_file_private *file_priv = file->driver_priv;
+
+		specific_ctx = i915_gem_context_lookup(file_priv, ctx_handle);
+		if (IS_ERR(specific_ctx)) {
+			drm_dbg(&perf->i915->drm,
+				"Failed to look up context with ID %u for opening perf stream\n",
+				  ctx_handle);
+			ret = PTR_ERR(specific_ctx);
+			goto err;
+		}
+	}
+
+	/*
+	 * On Haswell the OA unit supports clock gating off for a specific
+	 * context and in this mode there's no visibility of metrics for the
+	 * rest of the system, which we consider acceptable for a
+	 * non-privileged client.
+	 *
+	 * For Gen8->11 the OA unit no longer supports clock gating off for a
+	 * specific context and the kernel can't securely stop the counters
+	 * from updating as system-wide / global values. Even though we can
+	 * filter reports based on the included context ID we can't block
+	 * clients from seeing the raw / global counter values via
+	 * MI_REPORT_PERF_COUNT commands and so consider it a privileged op to
+	 * enable the OA unit by default.
+	 *
+	 * For Gen12+ we gain a new OAR unit that only monitors the RCS on a
+	 * per context basis. So we can relax requirements there if the user
+	 * doesn't request global stream access (i.e. query based sampling
+	 * using MI_RECORD_PERF_COUNT.
+	 */
+	if (IS_HASWELL(perf->i915) && specific_ctx)
+		privileged_op = false;
+	else if (GRAPHICS_VER(perf->i915) == 12 && specific_ctx &&
+		 (props->sample_flags & SAMPLE_OA_REPORT) == 0)
+		privileged_op = false;
+
+	if (props->hold_preemption) {
+		if (!props->single_context) {
+			drm_dbg(&perf->i915->drm,
+				"preemption disable with no context\n");
+			ret = -EINVAL;
+			goto err;
+		}
+		privileged_op = true;
+	}
+
+	/*
+	 * Asking for SSEU configuration is a priviliged operation.
+	 */
+	if (props->has_sseu)
+		privileged_op = true;
+	else
+		get_default_sseu_config(&props->sseu, props->engine);
+
+	/* Similar to perf's kernel.perf_paranoid_cpu sysctl option
+	 * we check a dev.i915.perf_stream_paranoid sysctl option
+	 * to determine if it's ok to access system wide OA counters
+	 * without CAP_PERFMON or CAP_SYS_ADMIN privileges.
+	 */
+	if (privileged_op &&
+	    i915_perf_stream_paranoid && !perfmon_capable()) {
+		drm_dbg(&perf->i915->drm,
+			"Insufficient privileges to open i915 perf stream\n");
+		ret = -EACCES;
+		goto err_ctx;
+	}
+
+	stream = kzalloc(sizeof(*stream), GFP_KERNEL);
+	if (!stream) {
+		ret = -ENOMEM;
+		goto err_ctx;
+	}
+
+	stream->perf = perf;
+	stream->ctx = specific_ctx;
+	stream->poll_oa_period = props->poll_oa_period;
+
+	ret = i915_oa_stream_init(stream, param, props);
+	if (ret)
+		goto err_alloc;
+
+	/* we avoid simply assigning stream->sample_flags = props->sample_flags
+	 * to have _stream_init check the combination of sample flags more
+	 * thoroughly, but still this is the expected result at this point.
+	 */
+	if (WARN_ON(stream->sample_flags != props->sample_flags)) {
+		ret = -ENODEV;
+		goto err_flags;
+	}
+
+	if (param->flags & I915_PERF_FLAG_FD_CLOEXEC)
+		f_flags |= O_CLOEXEC;
+	if (param->flags & I915_PERF_FLAG_FD_NONBLOCK)
+		f_flags |= O_NONBLOCK;
+
+	stream_fd = anon_inode_getfd("[i915_perf]", &fops, stream, f_flags);
+	if (stream_fd < 0) {
+		ret = stream_fd;
+		goto err_flags;
+	}
+
+	if (!(param->flags & I915_PERF_FLAG_DISABLED))
+		i915_perf_enable_locked(stream);
+
+	/* Take a reference on the driver that will be kept with stream_fd
+	 * until its release.
+	 */
+	drm_dev_get(&perf->i915->drm);
+
+	return stream_fd;
+
+err_flags:
+	if (stream->ops->destroy)
+		stream->ops->destroy(stream);
+err_alloc:
+	kfree(stream);
+err_ctx:
+	if (specific_ctx)
+		i915_gem_context_put(specific_ctx);
+err:
+	return ret;
+}
+
+static u64 oa_exponent_to_ns(struct i915_perf *perf, int exponent)
+{
+	return intel_gt_clock_interval_to_ns(to_gt(perf->i915),
+					     2ULL << exponent);
+}
+
+static __always_inline bool
+oa_format_valid(struct i915_perf *perf, enum drm_i915_oa_format format)
+{
+	return test_bit(format, perf->format_mask);
+}
+
+static __always_inline void
+oa_format_add(struct i915_perf *perf, enum drm_i915_oa_format format)
+{
+	__set_bit(format, perf->format_mask);
+}
+
+/**
+ * read_properties_unlocked - validate + copy userspace stream open properties
+ * @perf: i915 perf instance
+ * @uprops: The array of u64 key value pairs given by userspace
+ * @n_props: The number of key value pairs expected in @uprops
+ * @props: The stream configuration built up while validating properties
+ *
+ * Note this function only validates properties in isolation it doesn't
+ * validate that the combination of properties makes sense or that all
+ * properties necessary for a particular kind of stream have been set.
+ *
+ * Note that there currently aren't any ordering requirements for properties so
+ * we shouldn't validate or assume anything about ordering here. This doesn't
+ * rule out defining new properties with ordering requirements in the future.
+ */
+static int read_properties_unlocked(struct i915_perf *perf,
+				    u64 __user *uprops,
+				    u32 n_props,
+				    struct perf_open_properties *props)
+{
+	u64 __user *uprop = uprops;
+	u32 i;
+	int ret;
+
+	memset(props, 0, sizeof(struct perf_open_properties));
+	props->poll_oa_period = DEFAULT_POLL_PERIOD_NS;
+
+	if (!n_props) {
+		drm_dbg(&perf->i915->drm,
+			"No i915 perf properties given\n");
+		return -EINVAL;
+	}
+
+	/* At the moment we only support using i915-perf on the RCS. */
+	props->engine = intel_engine_lookup_user(perf->i915,
+						 I915_ENGINE_CLASS_RENDER,
+						 0);
+	if (!props->engine) {
+		drm_dbg(&perf->i915->drm,
+			"No RENDER-capable engines\n");
+		return -EINVAL;
+	}
+
+	/* Considering that ID = 0 is reserved and assuming that we don't
+	 * (currently) expect any configurations to ever specify duplicate
+	 * values for a particular property ID then the last _PROP_MAX value is
+	 * one greater than the maximum number of properties we expect to get
+	 * from userspace.
+	 */
+	if (n_props >= DRM_I915_PERF_PROP_MAX) {
+		drm_dbg(&perf->i915->drm,
+			"More i915 perf properties specified than exist\n");
+		return -EINVAL;
+	}
+
+	for (i = 0; i < n_props; i++) {
+		u64 oa_period, oa_freq_hz;
+		u64 id, value;
+
+		ret = get_user(id, uprop);
+		if (ret)
+			return ret;
+
+		ret = get_user(value, uprop + 1);
+		if (ret)
+			return ret;
+
+		if (id == 0 || id >= DRM_I915_PERF_PROP_MAX) {
+			drm_dbg(&perf->i915->drm,
+				"Unknown i915 perf property ID\n");
+			return -EINVAL;
+		}
+
+		switch ((enum drm_i915_perf_property_id)id) {
+		case DRM_I915_PERF_PROP_CTX_HANDLE:
+			props->single_context = 1;
+			props->ctx_handle = value;
+			break;
+		case DRM_I915_PERF_PROP_SAMPLE_OA:
+			if (value)
+				props->sample_flags |= SAMPLE_OA_REPORT;
+			break;
+		case DRM_I915_PERF_PROP_OA_METRICS_SET:
+			if (value == 0) {
+				drm_dbg(&perf->i915->drm,
+					"Unknown OA metric set ID\n");
+				return -EINVAL;
+			}
+			props->metrics_set = value;
+			break;
+		case DRM_I915_PERF_PROP_OA_FORMAT:
+			if (value == 0 || value >= I915_OA_FORMAT_MAX) {
+				drm_dbg(&perf->i915->drm,
+					"Out-of-range OA report format %llu\n",
+					  value);
+				return -EINVAL;
+			}
+			if (!oa_format_valid(perf, value)) {
+				drm_dbg(&perf->i915->drm,
+					"Unsupported OA report format %llu\n",
+					  value);
+				return -EINVAL;
+			}
+			props->oa_format = value;
+			break;
+		case DRM_I915_PERF_PROP_OA_EXPONENT:
+			if (value > OA_EXPONENT_MAX) {
+				drm_dbg(&perf->i915->drm,
+					"OA timer exponent too high (> %u)\n",
+					 OA_EXPONENT_MAX);
+				return -EINVAL;
+			}
+
+			/* Theoretically we can program the OA unit to sample
+			 * e.g. every 160ns for HSW, 167ns for BDW/SKL or 104ns
+			 * for BXT. We don't allow such high sampling
+			 * frequencies by default unless root.
+			 */
+
+			BUILD_BUG_ON(sizeof(oa_period) != 8);
+			oa_period = oa_exponent_to_ns(perf, value);
+
+			/* This check is primarily to ensure that oa_period <=
+			 * UINT32_MAX (before passing to do_div which only
+			 * accepts a u32 denominator), but we can also skip
+			 * checking anything < 1Hz which implicitly can't be
+			 * limited via an integer oa_max_sample_rate.
+			 */
+			if (oa_period <= NSEC_PER_SEC) {
+				u64 tmp = NSEC_PER_SEC;
+				do_div(tmp, oa_period);
+				oa_freq_hz = tmp;
+			} else
+				oa_freq_hz = 0;
+
+			if (oa_freq_hz > i915_oa_max_sample_rate && !perfmon_capable()) {
+				drm_dbg(&perf->i915->drm,
+					"OA exponent would exceed the max sampling frequency (sysctl dev.i915.oa_max_sample_rate) %uHz without CAP_PERFMON or CAP_SYS_ADMIN privileges\n",
+					  i915_oa_max_sample_rate);
+				return -EACCES;
+			}
+
+			props->oa_periodic = true;
+			props->oa_period_exponent = value;
+			break;
+		case DRM_I915_PERF_PROP_HOLD_PREEMPTION:
+			props->hold_preemption = !!value;
+			break;
+		case DRM_I915_PERF_PROP_GLOBAL_SSEU: {
+			struct drm_i915_gem_context_param_sseu user_sseu;
+
+			if (GRAPHICS_VER_FULL(perf->i915) >= IP_VER(12, 50)) {
+				drm_dbg(&perf->i915->drm,
+					"SSEU config not supported on gfx %x\n",
+					GRAPHICS_VER_FULL(perf->i915));
+				return -ENODEV;
+			}
+
+			if (copy_from_user(&user_sseu,
+					   u64_to_user_ptr(value),
+					   sizeof(user_sseu))) {
+				drm_dbg(&perf->i915->drm,
+					"Unable to copy global sseu parameter\n");
+				return -EFAULT;
+			}
+
+			ret = get_sseu_config(&props->sseu, props->engine, &user_sseu);
+			if (ret) {
+				drm_dbg(&perf->i915->drm,
+					"Invalid SSEU configuration\n");
+				return ret;
+			}
+			props->has_sseu = true;
+			break;
+		}
+		case DRM_I915_PERF_PROP_POLL_OA_PERIOD:
+			if (value < 100000 /* 100us */) {
+				drm_dbg(&perf->i915->drm,
+					"OA availability timer too small (%lluns < 100us)\n",
+					  value);
+				return -EINVAL;
+			}
+			props->poll_oa_period = value;
+			break;
+		case DRM_I915_PERF_PROP_MAX:
+			MISSING_CASE(id);
+			return -EINVAL;
+		}
+
+		uprop += 2;
+	}
+
+	return 0;
+}
+
+/**
+ * i915_perf_open_ioctl - DRM ioctl() for userspace to open a stream FD
+ * @dev: drm device
+ * @data: ioctl data copied from userspace (unvalidated)
+ * @file: drm file
+ *
+ * Validates the stream open parameters given by userspace including flags
+ * and an array of u64 key, value pair properties.
+ *
+ * Very little is assumed up front about the nature of the stream being
+ * opened (for instance we don't assume it's for periodic OA unit metrics). An
+ * i915-perf stream is expected to be a suitable interface for other forms of
+ * buffered data written by the GPU besides periodic OA metrics.
+ *
+ * Note we copy the properties from userspace outside of the i915 perf
+ * mutex to avoid an awkward lockdep with mmap_lock.
+ *
+ * Most of the implementation details are handled by
+ * i915_perf_open_ioctl_locked() after taking the &perf->lock
+ * mutex for serializing with any non-file-operation driver hooks.
+ *
+ * Return: A newly opened i915 Perf stream file descriptor or negative
+ * error code on failure.
+ */
+int i915_perf_open_ioctl(struct drm_device *dev, void *data,
+			 struct drm_file *file)
+{
+	struct i915_perf *perf = &to_i915(dev)->perf;
+	struct drm_i915_perf_open_param *param = data;
+	struct perf_open_properties props;
+	u32 known_open_flags;
+	int ret;
+
+	if (!perf->i915)
+		return -ENOTSUPP;
+
+	known_open_flags = I915_PERF_FLAG_FD_CLOEXEC |
+			   I915_PERF_FLAG_FD_NONBLOCK |
+			   I915_PERF_FLAG_DISABLED;
+	if (param->flags & ~known_open_flags) {
+		drm_dbg(&perf->i915->drm,
+			"Unknown drm_i915_perf_open_param flag\n");
+		return -EINVAL;
+	}
+
+	ret = read_properties_unlocked(perf,
+				       u64_to_user_ptr(param->properties_ptr),
+				       param->num_properties,
+				       &props);
+	if (ret)
+		return ret;
+
+	mutex_lock(&perf->lock);
+	ret = i915_perf_open_ioctl_locked(perf, param, &props, file);
+	mutex_unlock(&perf->lock);
+
+	return ret;
+}
+
+/**
+ * i915_perf_register - exposes i915-perf to userspace
+ * @i915: i915 device instance
+ *
+ * In particular OA metric sets are advertised under a sysfs metrics/
+ * directory allowing userspace to enumerate valid IDs that can be
+ * used to open an i915-perf stream.
+ */
+void i915_perf_register(struct drm_i915_private *i915)
+{
+	struct i915_perf *perf = &i915->perf;
+
+	if (!perf->i915)
+		return;
+
+	/* To be sure we're synchronized with an attempted
+	 * i915_perf_open_ioctl(); considering that we register after
+	 * being exposed to userspace.
+	 */
+	mutex_lock(&perf->lock);
+
+	perf->metrics_kobj =
+		kobject_create_and_add("metrics",
+				       &i915->drm.primary->kdev->kobj);
+
+	mutex_unlock(&perf->lock);
+}
+
+/**
+ * i915_perf_unregister - hide i915-perf from userspace
+ * @i915: i915 device instance
+ *
+ * i915-perf state cleanup is split up into an 'unregister' and
+ * 'deinit' phase where the interface is first hidden from
+ * userspace by i915_perf_unregister() before cleaning up
+ * remaining state in i915_perf_fini().
+ */
+void i915_perf_unregister(struct drm_i915_private *i915)
+{
+	struct i915_perf *perf = &i915->perf;
+
+	if (!perf->metrics_kobj)
+		return;
+
+	kobject_put(perf->metrics_kobj);
+	perf->metrics_kobj = NULL;
+}
+
+static bool gen8_is_valid_flex_addr(struct i915_perf *perf, u32 addr)
+{
+	static const i915_reg_t flex_eu_regs[] = {
+		EU_PERF_CNTL0,
+		EU_PERF_CNTL1,
+		EU_PERF_CNTL2,
+		EU_PERF_CNTL3,
+		EU_PERF_CNTL4,
+		EU_PERF_CNTL5,
+		EU_PERF_CNTL6,
+	};
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(flex_eu_regs); i++) {
+		if (i915_mmio_reg_offset(flex_eu_regs[i]) == addr)
+			return true;
+	}
+	return false;
+}
+
+static bool reg_in_range_table(u32 addr, const struct i915_range *table)
+{
+	while (table->start || table->end) {
+		if (addr >= table->start && addr <= table->end)
+			return true;
+
+		table++;
+	}
+
+	return false;
+}
+
+#define REG_EQUAL(addr, mmio) \
+	((addr) == i915_mmio_reg_offset(mmio))
+
+static const struct i915_range gen7_oa_b_counters[] = {
+	{ .start = 0x2710, .end = 0x272c },	/* OASTARTTRIG[1-8] */
+	{ .start = 0x2740, .end = 0x275c },	/* OAREPORTTRIG[1-8] */
+	{ .start = 0x2770, .end = 0x27ac },	/* OACEC[0-7][0-1] */
+	{}
+};
+
+static const struct i915_range gen12_oa_b_counters[] = {
+	{ .start = 0x2b2c, .end = 0x2b2c },	/* GEN12_OAG_OA_PESS */
+	{ .start = 0xd900, .end = 0xd91c },	/* GEN12_OAG_OASTARTTRIG[1-8] */
+	{ .start = 0xd920, .end = 0xd93c },	/* GEN12_OAG_OAREPORTTRIG1[1-8] */
+	{ .start = 0xd940, .end = 0xd97c },	/* GEN12_OAG_CEC[0-7][0-1] */
+	{ .start = 0xdc00, .end = 0xdc3c },	/* GEN12_OAG_SCEC[0-7][0-1] */
+	{ .start = 0xdc40, .end = 0xdc40 },	/* GEN12_OAG_SPCTR_CNF */
+	{ .start = 0xdc44, .end = 0xdc44 },	/* GEN12_OAA_DBG_REG */
+	{}
+};
+
+static const struct i915_range gen7_oa_mux_regs[] = {
+	{ .start = 0x91b8, .end = 0x91cc },	/* OA_PERFCNT[1-2], OA_PERFMATRIX */
+	{ .start = 0x9800, .end = 0x9888 },	/* MICRO_BP0_0 - NOA_WRITE */
+	{ .start = 0xe180, .end = 0xe180 },	/* HALF_SLICE_CHICKEN2 */
+	{}
+};
+
+static const struct i915_range hsw_oa_mux_regs[] = {
+	{ .start = 0x09e80, .end = 0x09ea4 }, /* HSW_MBVID2_NOA[0-9] */
+	{ .start = 0x09ec0, .end = 0x09ec0 }, /* HSW_MBVID2_MISR0 */
+	{ .start = 0x25100, .end = 0x2ff90 },
+	{}
+};
+
+static const struct i915_range chv_oa_mux_regs[] = {
+	{ .start = 0x182300, .end = 0x1823a4 },
+	{}
+};
+
+static const struct i915_range gen8_oa_mux_regs[] = {
+	{ .start = 0x0d00, .end = 0x0d2c },	/* RPM_CONFIG[0-1], NOA_CONFIG[0-8] */
+	{ .start = 0x20cc, .end = 0x20cc },	/* WAIT_FOR_RC6_EXIT */
+	{}
+};
+
+static const struct i915_range gen11_oa_mux_regs[] = {
+	{ .start = 0x91c8, .end = 0x91dc },	/* OA_PERFCNT[3-4] */
+	{}
+};
+
+static const struct i915_range gen12_oa_mux_regs[] = {
+	{ .start = 0x0d00, .end = 0x0d04 },     /* RPM_CONFIG[0-1] */
+	{ .start = 0x0d0c, .end = 0x0d2c },     /* NOA_CONFIG[0-8] */
+	{ .start = 0x9840, .end = 0x9840 },	/* GDT_CHICKEN_BITS */
+	{ .start = 0x9884, .end = 0x9888 },	/* NOA_WRITE */
+	{ .start = 0x20cc, .end = 0x20cc },	/* WAIT_FOR_RC6_EXIT */
+	{}
+};
+
+static bool gen7_is_valid_b_counter_addr(struct i915_perf *perf, u32 addr)
+{
+	return reg_in_range_table(addr, gen7_oa_b_counters);
+}
+
+static bool gen8_is_valid_mux_addr(struct i915_perf *perf, u32 addr)
+{
+	return reg_in_range_table(addr, gen7_oa_mux_regs) ||
+		reg_in_range_table(addr, gen8_oa_mux_regs);
+}
+
+static bool gen11_is_valid_mux_addr(struct i915_perf *perf, u32 addr)
+{
+	return reg_in_range_table(addr, gen7_oa_mux_regs) ||
+		reg_in_range_table(addr, gen8_oa_mux_regs) ||
+		reg_in_range_table(addr, gen11_oa_mux_regs);
+}
+
+static bool hsw_is_valid_mux_addr(struct i915_perf *perf, u32 addr)
+{
+	return reg_in_range_table(addr, gen7_oa_mux_regs) ||
+		reg_in_range_table(addr, hsw_oa_mux_regs);
+}
+
+static bool chv_is_valid_mux_addr(struct i915_perf *perf, u32 addr)
+{
+	return reg_in_range_table(addr, gen7_oa_mux_regs) ||
+		reg_in_range_table(addr, chv_oa_mux_regs);
+}
+
+static bool gen12_is_valid_b_counter_addr(struct i915_perf *perf, u32 addr)
+{
+	return reg_in_range_table(addr, gen12_oa_b_counters);
+}
+
+static bool gen12_is_valid_mux_addr(struct i915_perf *perf, u32 addr)
+{
+	return reg_in_range_table(addr, gen12_oa_mux_regs);
+}
+
+static u32 mask_reg_value(u32 reg, u32 val)
+{
+	/* HALF_SLICE_CHICKEN2 is programmed with a the
+	 * WaDisableSTUnitPowerOptimization workaround. Make sure the value
+	 * programmed by userspace doesn't change this.
+	 */
+	if (REG_EQUAL(reg, HALF_SLICE_CHICKEN2))
+		val = val & ~_MASKED_BIT_ENABLE(GEN8_ST_PO_DISABLE);
+
+	/* WAIT_FOR_RC6_EXIT has only one bit fullfilling the function
+	 * indicated by its name and a bunch of selection fields used by OA
+	 * configs.
+	 */
+	if (REG_EQUAL(reg, WAIT_FOR_RC6_EXIT))
+		val = val & ~_MASKED_BIT_ENABLE(HSW_WAIT_FOR_RC6_EXIT_ENABLE);
+
+	return val;
+}
+
+static struct i915_oa_reg *alloc_oa_regs(struct i915_perf *perf,
+					 bool (*is_valid)(struct i915_perf *perf, u32 addr),
+					 u32 __user *regs,
+					 u32 n_regs)
+{
+	struct i915_oa_reg *oa_regs;
+	int err;
+	u32 i;
+
+	if (!n_regs)
+		return NULL;
+
+	/* No is_valid function means we're not allowing any register to be programmed. */
+	GEM_BUG_ON(!is_valid);
+	if (!is_valid)
+		return ERR_PTR(-EINVAL);
+
+	oa_regs = kmalloc_array(n_regs, sizeof(*oa_regs), GFP_KERNEL);
+	if (!oa_regs)
+		return ERR_PTR(-ENOMEM);
+
+	for (i = 0; i < n_regs; i++) {
+		u32 addr, value;
+
+		err = get_user(addr, regs);
+		if (err)
+			goto addr_err;
+
+		if (!is_valid(perf, addr)) {
+			drm_dbg(&perf->i915->drm,
+				"Invalid oa_reg address: %X\n", addr);
+			err = -EINVAL;
+			goto addr_err;
+		}
+
+		err = get_user(value, regs + 1);
+		if (err)
+			goto addr_err;
+
+		oa_regs[i].addr = _MMIO(addr);
+		oa_regs[i].value = mask_reg_value(addr, value);
+
+		regs += 2;
+	}
+
+	return oa_regs;
+
+addr_err:
+	kfree(oa_regs);
+	return ERR_PTR(err);
+}
+
+static ssize_t show_dynamic_id(struct kobject *kobj,
+			       struct kobj_attribute *attr,
+			       char *buf)
+{
+	struct i915_oa_config *oa_config =
+		container_of(attr, typeof(*oa_config), sysfs_metric_id);
+
+	return sprintf(buf, "%d\n", oa_config->id);
+}
+
+static int create_dynamic_oa_sysfs_entry(struct i915_perf *perf,
+					 struct i915_oa_config *oa_config)
+{
+	sysfs_attr_init(&oa_config->sysfs_metric_id.attr);
+	oa_config->sysfs_metric_id.attr.name = "id";
+	oa_config->sysfs_metric_id.attr.mode = S_IRUGO;
+	oa_config->sysfs_metric_id.show = show_dynamic_id;
+	oa_config->sysfs_metric_id.store = NULL;
+
+	oa_config->attrs[0] = &oa_config->sysfs_metric_id.attr;
+	oa_config->attrs[1] = NULL;
+
+	oa_config->sysfs_metric.name = oa_config->uuid;
+	oa_config->sysfs_metric.attrs = oa_config->attrs;
+
+	return sysfs_create_group(perf->metrics_kobj,
+				  &oa_config->sysfs_metric);
+}
+
+/**
+ * i915_perf_add_config_ioctl - DRM ioctl() for userspace to add a new OA config
+ * @dev: drm device
+ * @data: ioctl data (pointer to struct drm_i915_perf_oa_config) copied from
+ *        userspace (unvalidated)
+ * @file: drm file
+ *
+ * Validates the submitted OA register to be saved into a new OA config that
+ * can then be used for programming the OA unit and its NOA network.
+ *
+ * Returns: A new allocated config number to be used with the perf open ioctl
+ * or a negative error code on failure.
+ */
+int i915_perf_add_config_ioctl(struct drm_device *dev, void *data,
+			       struct drm_file *file)
+{
+	struct i915_perf *perf = &to_i915(dev)->perf;
+	struct drm_i915_perf_oa_config *args = data;
+	struct i915_oa_config *oa_config, *tmp;
+	struct i915_oa_reg *regs;
+	int err, id;
+
+	if (!perf->i915)
+		return -ENOTSUPP;
+
+	if (!perf->metrics_kobj) {
+		drm_dbg(&perf->i915->drm,
+			"OA metrics weren't advertised via sysfs\n");
+		return -EINVAL;
+	}
+
+	if (i915_perf_stream_paranoid && !perfmon_capable()) {
+		drm_dbg(&perf->i915->drm,
+			"Insufficient privileges to add i915 OA config\n");
+		return -EACCES;
+	}
+
+	if ((!args->mux_regs_ptr || !args->n_mux_regs) &&
+	    (!args->boolean_regs_ptr || !args->n_boolean_regs) &&
+	    (!args->flex_regs_ptr || !args->n_flex_regs)) {
+		drm_dbg(&perf->i915->drm,
+			"No OA registers given\n");
+		return -EINVAL;
+	}
+
+	oa_config = kzalloc(sizeof(*oa_config), GFP_KERNEL);
+	if (!oa_config) {
+		drm_dbg(&perf->i915->drm,
+			"Failed to allocate memory for the OA config\n");
+		return -ENOMEM;
+	}
+
+	oa_config->perf = perf;
+	kref_init(&oa_config->ref);
+
+	if (!uuid_is_valid(args->uuid)) {
+		drm_dbg(&perf->i915->drm,
+			"Invalid uuid format for OA config\n");
+		err = -EINVAL;
+		goto reg_err;
+	}
+
+	/* Last character in oa_config->uuid will be 0 because oa_config is
+	 * kzalloc.
+	 */
+	memcpy(oa_config->uuid, args->uuid, sizeof(args->uuid));
+
+	oa_config->mux_regs_len = args->n_mux_regs;
+	regs = alloc_oa_regs(perf,
+			     perf->ops.is_valid_mux_reg,
+			     u64_to_user_ptr(args->mux_regs_ptr),
+			     args->n_mux_regs);
+
+	if (IS_ERR(regs)) {
+		drm_dbg(&perf->i915->drm,
+			"Failed to create OA config for mux_regs\n");
+		err = PTR_ERR(regs);
+		goto reg_err;
+	}
+	oa_config->mux_regs = regs;
+
+	oa_config->b_counter_regs_len = args->n_boolean_regs;
+	regs = alloc_oa_regs(perf,
+			     perf->ops.is_valid_b_counter_reg,
+			     u64_to_user_ptr(args->boolean_regs_ptr),
+			     args->n_boolean_regs);
+
+	if (IS_ERR(regs)) {
+		drm_dbg(&perf->i915->drm,
+			"Failed to create OA config for b_counter_regs\n");
+		err = PTR_ERR(regs);
+		goto reg_err;
+	}
+	oa_config->b_counter_regs = regs;
+
+	if (GRAPHICS_VER(perf->i915) < 8) {
+		if (args->n_flex_regs != 0) {
+			err = -EINVAL;
+			goto reg_err;
+		}
+	} else {
+		oa_config->flex_regs_len = args->n_flex_regs;
+		regs = alloc_oa_regs(perf,
+				     perf->ops.is_valid_flex_reg,
+				     u64_to_user_ptr(args->flex_regs_ptr),
+				     args->n_flex_regs);
+
+		if (IS_ERR(regs)) {
+			drm_dbg(&perf->i915->drm,
+				"Failed to create OA config for flex_regs\n");
+			err = PTR_ERR(regs);
+			goto reg_err;
+		}
+		oa_config->flex_regs = regs;
+	}
+
+	err = mutex_lock_interruptible(&perf->metrics_lock);
+	if (err)
+		goto reg_err;
+
+	/* We shouldn't have too many configs, so this iteration shouldn't be
+	 * too costly.
+	 */
+	idr_for_each_entry(&perf->metrics_idr, tmp, id) {
+		if (!strcmp(tmp->uuid, oa_config->uuid)) {
+			drm_dbg(&perf->i915->drm,
+				"OA config already exists with this uuid\n");
+			err = -EADDRINUSE;
+			goto sysfs_err;
+		}
+	}
+
+	err = create_dynamic_oa_sysfs_entry(perf, oa_config);
+	if (err) {
+		drm_dbg(&perf->i915->drm,
+			"Failed to create sysfs entry for OA config\n");
+		goto sysfs_err;
+	}
+
+	/* Config id 0 is invalid, id 1 for kernel stored test config. */
+	oa_config->id = idr_alloc(&perf->metrics_idr,
+				  oa_config, 2,
+				  0, GFP_KERNEL);
+	if (oa_config->id < 0) {
+		drm_dbg(&perf->i915->drm,
+			"Failed to create sysfs entry for OA config\n");
+		err = oa_config->id;
+		goto sysfs_err;
+	}
+	id = oa_config->id;
+
+	drm_dbg(&perf->i915->drm,
+		"Added config %s id=%i\n", oa_config->uuid, oa_config->id);
+	mutex_unlock(&perf->metrics_lock);
+
+	return id;
+
+sysfs_err:
+	mutex_unlock(&perf->metrics_lock);
+reg_err:
+	i915_oa_config_put(oa_config);
+	drm_dbg(&perf->i915->drm,
+		"Failed to add new OA config\n");
+	return err;
+}
+
+/**
+ * i915_perf_remove_config_ioctl - DRM ioctl() for userspace to remove an OA config
+ * @dev: drm device
+ * @data: ioctl data (pointer to u64 integer) copied from userspace
+ * @file: drm file
+ *
+ * Configs can be removed while being used, the will stop appearing in sysfs
+ * and their content will be freed when the stream using the config is closed.
+ *
+ * Returns: 0 on success or a negative error code on failure.
+ */
+int i915_perf_remove_config_ioctl(struct drm_device *dev, void *data,
+				  struct drm_file *file)
+{
+	struct i915_perf *perf = &to_i915(dev)->perf;
+	u64 *arg = data;
+	struct i915_oa_config *oa_config;
+	int ret;
+
+	if (!perf->i915)
+		return -ENOTSUPP;
+
+	if (i915_perf_stream_paranoid && !perfmon_capable()) {
+		drm_dbg(&perf->i915->drm,
+			"Insufficient privileges to remove i915 OA config\n");
+		return -EACCES;
+	}
+
+	ret = mutex_lock_interruptible(&perf->metrics_lock);
+	if (ret)
+		return ret;
+
+	oa_config = idr_find(&perf->metrics_idr, *arg);
+	if (!oa_config) {
+		drm_dbg(&perf->i915->drm,
+			"Failed to remove unknown OA config\n");
+		ret = -ENOENT;
+		goto err_unlock;
+	}
+
+	GEM_BUG_ON(*arg != oa_config->id);
+
+	sysfs_remove_group(perf->metrics_kobj, &oa_config->sysfs_metric);
+
+	idr_remove(&perf->metrics_idr, *arg);
+
+	mutex_unlock(&perf->metrics_lock);
+
+	drm_dbg(&perf->i915->drm,
+		"Removed config %s id=%i\n", oa_config->uuid, oa_config->id);
+
+	i915_oa_config_put(oa_config);
+
+	return 0;
+
+err_unlock:
+	mutex_unlock(&perf->metrics_lock);
+	return ret;
+}
+
+static struct ctl_table oa_table[] = {
+	{
+	 .procname = "perf_stream_paranoid",
+	 .data = &i915_perf_stream_paranoid,
+	 .maxlen = sizeof(i915_perf_stream_paranoid),
+	 .mode = 0644,
+	 .proc_handler = proc_dointvec_minmax,
+	 .extra1 = SYSCTL_ZERO,
+	 .extra2 = SYSCTL_ONE,
+	 },
+	{
+	 .procname = "oa_max_sample_rate",
+	 .data = &i915_oa_max_sample_rate,
+	 .maxlen = sizeof(i915_oa_max_sample_rate),
+	 .mode = 0644,
+	 .proc_handler = proc_dointvec_minmax,
+	 .extra1 = SYSCTL_ZERO,
+	 .extra2 = &oa_sample_rate_hard_limit,
+	 },
+	{}
+};
+
+static void oa_init_supported_formats(struct i915_perf *perf)
+{
+	struct drm_i915_private *i915 = perf->i915;
+	enum intel_platform platform = INTEL_INFO(i915)->platform;
+
+	switch (platform) {
+	case INTEL_HASWELL:
+		oa_format_add(perf, I915_OA_FORMAT_A13);
+		oa_format_add(perf, I915_OA_FORMAT_A13);
+		oa_format_add(perf, I915_OA_FORMAT_A29);
+		oa_format_add(perf, I915_OA_FORMAT_A13_B8_C8);
+		oa_format_add(perf, I915_OA_FORMAT_B4_C8);
+		oa_format_add(perf, I915_OA_FORMAT_A45_B8_C8);
+		oa_format_add(perf, I915_OA_FORMAT_B4_C8_A16);
+		oa_format_add(perf, I915_OA_FORMAT_C4_B8);
+		break;
+
+	case INTEL_BROADWELL:
+	case INTEL_CHERRYVIEW:
+	case INTEL_SKYLAKE:
+	case INTEL_BROXTON:
+	case INTEL_KABYLAKE:
+	case INTEL_GEMINILAKE:
+	case INTEL_COFFEELAKE:
+	case INTEL_COMETLAKE:
+	case INTEL_ICELAKE:
+	case INTEL_ELKHARTLAKE:
+	case INTEL_JASPERLAKE:
+	case INTEL_TIGERLAKE:
+	case INTEL_ROCKETLAKE:
+	case INTEL_DG1:
+	case INTEL_ALDERLAKE_S:
+	case INTEL_ALDERLAKE_P:
+		oa_format_add(perf, I915_OA_FORMAT_A12);
+		oa_format_add(perf, I915_OA_FORMAT_A12_B8_C8);
+		oa_format_add(perf, I915_OA_FORMAT_A32u40_A4u32_B8_C8);
+		oa_format_add(perf, I915_OA_FORMAT_C4_B8);
+		break;
+
+	default:
+		MISSING_CASE(platform);
+	}
+}
+
+/**
+ * i915_perf_init - initialize i915-perf state on module bind
+ * @i915: i915 device instance
+ *
+ * Initializes i915-perf state without exposing anything to userspace.
+ *
+ * Note: i915-perf initialization is split into an 'init' and 'register'
+ * phase with the i915_perf_register() exposing state to userspace.
+ */
+void i915_perf_init(struct drm_i915_private *i915)
+{
+	struct i915_perf *perf = &i915->perf;
+
+	/* XXX const struct i915_perf_ops! */
+
+	/* i915_perf is not enabled for DG2 yet */
+	if (IS_DG2(i915))
+		return;
+
+	perf->oa_formats = oa_formats;
+	if (IS_HASWELL(i915)) {
+		perf->ops.is_valid_b_counter_reg = gen7_is_valid_b_counter_addr;
+		perf->ops.is_valid_mux_reg = hsw_is_valid_mux_addr;
+		perf->ops.is_valid_flex_reg = NULL;
+		perf->ops.enable_metric_set = hsw_enable_metric_set;
+		perf->ops.disable_metric_set = hsw_disable_metric_set;
+		perf->ops.oa_enable = gen7_oa_enable;
+		perf->ops.oa_disable = gen7_oa_disable;
+		perf->ops.read = gen7_oa_read;
+		perf->ops.oa_hw_tail_read = gen7_oa_hw_tail_read;
+	} else if (HAS_LOGICAL_RING_CONTEXTS(i915)) {
+		/* Note: that although we could theoretically also support the
+		 * legacy ringbuffer mode on BDW (and earlier iterations of
+		 * this driver, before upstreaming did this) it didn't seem
+		 * worth the complexity to maintain now that BDW+ enable
+		 * execlist mode by default.
+		 */
+		perf->ops.read = gen8_oa_read;
+
+		if (IS_GRAPHICS_VER(i915, 8, 9)) {
+			perf->ops.is_valid_b_counter_reg =
+				gen7_is_valid_b_counter_addr;
+			perf->ops.is_valid_mux_reg =
+				gen8_is_valid_mux_addr;
+			perf->ops.is_valid_flex_reg =
+				gen8_is_valid_flex_addr;
+
+			if (IS_CHERRYVIEW(i915)) {
+				perf->ops.is_valid_mux_reg =
+					chv_is_valid_mux_addr;
+			}
+
+			perf->ops.oa_enable = gen8_oa_enable;
+			perf->ops.oa_disable = gen8_oa_disable;
+			perf->ops.enable_metric_set = gen8_enable_metric_set;
+			perf->ops.disable_metric_set = gen8_disable_metric_set;
+			perf->ops.oa_hw_tail_read = gen8_oa_hw_tail_read;
+
+			if (GRAPHICS_VER(i915) == 8) {
+				perf->ctx_oactxctrl_offset = 0x120;
+				perf->ctx_flexeu0_offset = 0x2ce;
+
+				perf->gen8_valid_ctx_bit = BIT(25);
+			} else {
+				perf->ctx_oactxctrl_offset = 0x128;
+				perf->ctx_flexeu0_offset = 0x3de;
+
+				perf->gen8_valid_ctx_bit = BIT(16);
+			}
+		} else if (GRAPHICS_VER(i915) == 11) {
+			perf->ops.is_valid_b_counter_reg =
+				gen7_is_valid_b_counter_addr;
+			perf->ops.is_valid_mux_reg =
+				gen11_is_valid_mux_addr;
+			perf->ops.is_valid_flex_reg =
+				gen8_is_valid_flex_addr;
+
+			perf->ops.oa_enable = gen8_oa_enable;
+			perf->ops.oa_disable = gen8_oa_disable;
+			perf->ops.enable_metric_set = gen8_enable_metric_set;
+			perf->ops.disable_metric_set = gen11_disable_metric_set;
+			perf->ops.oa_hw_tail_read = gen8_oa_hw_tail_read;
+
+			perf->ctx_oactxctrl_offset = 0x124;
+			perf->ctx_flexeu0_offset = 0x78e;
+
+			perf->gen8_valid_ctx_bit = BIT(16);
+		} else if (GRAPHICS_VER(i915) == 12) {
+			perf->ops.is_valid_b_counter_reg =
+				gen12_is_valid_b_counter_addr;
+			perf->ops.is_valid_mux_reg =
+				gen12_is_valid_mux_addr;
+			perf->ops.is_valid_flex_reg =
+				gen8_is_valid_flex_addr;
+
+			perf->ops.oa_enable = gen12_oa_enable;
+			perf->ops.oa_disable = gen12_oa_disable;
+			perf->ops.enable_metric_set = gen12_enable_metric_set;
+			perf->ops.disable_metric_set = gen12_disable_metric_set;
+			perf->ops.oa_hw_tail_read = gen12_oa_hw_tail_read;
+
+			perf->ctx_flexeu0_offset = 0;
+			perf->ctx_oactxctrl_offset = 0x144;
+		}
+	}
+
+	if (perf->ops.enable_metric_set) {
+		mutex_init(&perf->lock);
+
+		/* Choose a representative limit */
+		oa_sample_rate_hard_limit = to_gt(i915)->clock_frequency / 2;
+
+		mutex_init(&perf->metrics_lock);
+		idr_init_base(&perf->metrics_idr, 1);
+
+		/* We set up some ratelimit state to potentially throttle any
+		 * _NOTES about spurious, invalid OA reports which we don't
+		 * forward to userspace.
+		 *
+		 * We print a _NOTE about any throttling when closing the
+		 * stream instead of waiting until driver _fini which no one
+		 * would ever see.
+		 *
+		 * Using the same limiting factors as printk_ratelimit()
+		 */
+		ratelimit_state_init(&perf->spurious_report_rs, 5 * HZ, 10);
+		/* Since we use a DRM_NOTE for spurious reports it would be
+		 * inconsistent to let __ratelimit() automatically print a
+		 * warning for throttling.
+		 */
+		ratelimit_set_flags(&perf->spurious_report_rs,
+				    RATELIMIT_MSG_ON_RELEASE);
+
+		ratelimit_state_init(&perf->tail_pointer_race,
+				     5 * HZ, 10);
+		ratelimit_set_flags(&perf->tail_pointer_race,
+				    RATELIMIT_MSG_ON_RELEASE);
+
+		atomic64_set(&perf->noa_programming_delay,
+			     500 * 1000 /* 500us */);
+
+		perf->i915 = i915;
+
+		oa_init_supported_formats(perf);
+	}
+}
+
+static int destroy_config(int id, void *p, void *data)
+{
+	i915_oa_config_put(p);
+	return 0;
+}
+
+int i915_perf_sysctl_register(void)
+{
+	sysctl_header = register_sysctl("dev/i915", oa_table);
+	return 0;
+}
+
+void i915_perf_sysctl_unregister(void)
+{
+	unregister_sysctl_table(sysctl_header);
+}
+
+/**
+ * i915_perf_fini - Counter part to i915_perf_init()
+ * @i915: i915 device instance
+ */
+void i915_perf_fini(struct drm_i915_private *i915)
+{
+	struct i915_perf *perf = &i915->perf;
+
+	if (!perf->i915)
+		return;
+
+	idr_for_each(&perf->metrics_idr, destroy_config, perf);
+	idr_destroy(&perf->metrics_idr);
+
+	memset(&perf->ops, 0, sizeof(perf->ops));
+	perf->i915 = NULL;
+}
+
+/**
+ * i915_perf_ioctl_version - Version of the i915-perf subsystem
+ *
+ * This version number is used by userspace to detect available features.
+ */
+int i915_perf_ioctl_version(void)
+{
+	/*
+	 * 1: Initial version
+	 *   I915_PERF_IOCTL_ENABLE
+	 *   I915_PERF_IOCTL_DISABLE
+	 *
+	 * 2: Added runtime modification of OA config.
+	 *   I915_PERF_IOCTL_CONFIG
+	 *
+	 * 3: Add DRM_I915_PERF_PROP_HOLD_PREEMPTION parameter to hold
+	 *    preemption on a particular context so that performance data is
+	 *    accessible from a delta of MI_RPC reports without looking at the
+	 *    OA buffer.
+	 *
+	 * 4: Add DRM_I915_PERF_PROP_ALLOWED_SSEU to limit what contexts can
+	 *    be run for the duration of the performance recording based on
+	 *    their SSEU configuration.
+	 *
+	 * 5: Add DRM_I915_PERF_PROP_POLL_OA_PERIOD parameter that controls the
+	 *    interval for the hrtimer used to check for OA data.
+	 */
+	return 5;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/i915_perf.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_perf.h b/drivers/gpu/drm/i915/i915_perf.h
new file mode 100644
index 000000000..1d1329e5a
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_perf.h
@@ -0,0 +1,60 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_PERF_H__
+#define __I915_PERF_H__
+
+#include <linux/kref.h>
+#include <linux/types.h>
+
+#include "i915_perf_types.h"
+
+struct drm_device;
+struct drm_file;
+struct drm_i915_private;
+struct i915_oa_config;
+struct intel_context;
+struct intel_engine_cs;
+
+void i915_perf_init(struct drm_i915_private *i915);
+void i915_perf_fini(struct drm_i915_private *i915);
+void i915_perf_register(struct drm_i915_private *i915);
+void i915_perf_unregister(struct drm_i915_private *i915);
+int i915_perf_ioctl_version(void);
+int i915_perf_sysctl_register(void);
+void i915_perf_sysctl_unregister(void);
+
+int i915_perf_open_ioctl(struct drm_device *dev, void *data,
+			 struct drm_file *file);
+int i915_perf_add_config_ioctl(struct drm_device *dev, void *data,
+			       struct drm_file *file);
+int i915_perf_remove_config_ioctl(struct drm_device *dev, void *data,
+				  struct drm_file *file);
+
+void i915_oa_init_reg_state(const struct intel_context *ce,
+			    const struct intel_engine_cs *engine);
+
+struct i915_oa_config *
+i915_perf_get_oa_config(struct i915_perf *perf, int metrics_set);
+
+static inline struct i915_oa_config *
+i915_oa_config_get(struct i915_oa_config *oa_config)
+{
+	if (kref_get_unless_zero(&oa_config->ref))
+		return oa_config;
+	else
+		return NULL;
+}
+
+void i915_oa_config_release(struct kref *ref);
+static inline void i915_oa_config_put(struct i915_oa_config *oa_config)
+{
+	if (!oa_config)
+		return;
+
+	kref_put(&oa_config->ref, i915_oa_config_release);
+}
+
+#endif /* __I915_PERF_H__ */
diff --git a/drivers/gpu/drm/i915/i915_perf_oa_regs.h b/drivers/gpu/drm/i915/i915_perf_oa_regs.h
new file mode 100644
index 000000000..f31c9f13a
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_perf_oa_regs.h
@@ -0,0 +1,137 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_PERF_OA_REGS__
+#define __INTEL_PERF_OA_REGS__
+
+#include "i915_reg_defs.h"
+
+#define GEN7_OACONTROL _MMIO(0x2360)
+#define  GEN7_OACONTROL_CTX_MASK	    0xFFFFF000
+#define  GEN7_OACONTROL_TIMER_PERIOD_MASK   0x3F
+#define  GEN7_OACONTROL_TIMER_PERIOD_SHIFT  6
+#define  GEN7_OACONTROL_TIMER_ENABLE	    (1 << 5)
+#define  GEN7_OACONTROL_FORMAT_A13	    (0 << 2)
+#define  GEN7_OACONTROL_FORMAT_A29	    (1 << 2)
+#define  GEN7_OACONTROL_FORMAT_A13_B8_C8    (2 << 2)
+#define  GEN7_OACONTROL_FORMAT_A29_B8_C8    (3 << 2)
+#define  GEN7_OACONTROL_FORMAT_B4_C8	    (4 << 2)
+#define  GEN7_OACONTROL_FORMAT_A45_B8_C8    (5 << 2)
+#define  GEN7_OACONTROL_FORMAT_B4_C8_A16    (6 << 2)
+#define  GEN7_OACONTROL_FORMAT_C4_B8	    (7 << 2)
+#define  GEN7_OACONTROL_FORMAT_SHIFT	    2
+#define  GEN7_OACONTROL_PER_CTX_ENABLE	    (1 << 1)
+#define  GEN7_OACONTROL_ENABLE		    (1 << 0)
+
+#define GEN8_OACTXID _MMIO(0x2364)
+
+#define GEN8_OA_DEBUG _MMIO(0x2B04)
+#define  GEN9_OA_DEBUG_DISABLE_CLK_RATIO_REPORTS    (1 << 5)
+#define  GEN9_OA_DEBUG_INCLUDE_CLK_RATIO	    (1 << 6)
+#define  GEN9_OA_DEBUG_DISABLE_GO_1_0_REPORTS	    (1 << 2)
+#define  GEN9_OA_DEBUG_DISABLE_CTX_SWITCH_REPORTS   (1 << 1)
+
+#define GEN8_OACONTROL _MMIO(0x2B00)
+#define  GEN8_OA_REPORT_FORMAT_A12	    (0 << 2)
+#define  GEN8_OA_REPORT_FORMAT_A12_B8_C8    (2 << 2)
+#define  GEN8_OA_REPORT_FORMAT_A36_B8_C8    (5 << 2)
+#define  GEN8_OA_REPORT_FORMAT_C4_B8	    (7 << 2)
+#define  GEN8_OA_REPORT_FORMAT_SHIFT	    2
+#define  GEN8_OA_SPECIFIC_CONTEXT_ENABLE    (1 << 1)
+#define  GEN8_OA_COUNTER_ENABLE             (1 << 0)
+
+#define GEN8_OACTXCONTROL _MMIO(0x2360)
+#define  GEN8_OA_TIMER_PERIOD_MASK	    0x3F
+#define  GEN8_OA_TIMER_PERIOD_SHIFT	    2
+#define  GEN8_OA_TIMER_ENABLE		    (1 << 1)
+#define  GEN8_OA_COUNTER_RESUME		    (1 << 0)
+
+#define GEN7_OABUFFER _MMIO(0x23B0) /* R/W */
+#define  GEN7_OABUFFER_OVERRUN_DISABLE	    (1 << 3)
+#define  GEN7_OABUFFER_EDGE_TRIGGER	    (1 << 2)
+#define  GEN7_OABUFFER_STOP_RESUME_ENABLE   (1 << 1)
+#define  GEN7_OABUFFER_RESUME		    (1 << 0)
+
+#define GEN8_OABUFFER_UDW _MMIO(0x23b4)
+#define GEN8_OABUFFER _MMIO(0x2b14)
+#define  GEN8_OABUFFER_MEM_SELECT_GGTT      (1 << 0)  /* 0: PPGTT, 1: GGTT */
+
+#define GEN7_OASTATUS1 _MMIO(0x2364)
+#define  GEN7_OASTATUS1_TAIL_MASK	    0xffffffc0
+#define  GEN7_OASTATUS1_COUNTER_OVERFLOW    (1 << 2)
+#define  GEN7_OASTATUS1_OABUFFER_OVERFLOW   (1 << 1)
+#define  GEN7_OASTATUS1_REPORT_LOST	    (1 << 0)
+
+#define GEN7_OASTATUS2 _MMIO(0x2368)
+#define  GEN7_OASTATUS2_HEAD_MASK           0xffffffc0
+#define  GEN7_OASTATUS2_MEM_SELECT_GGTT     (1 << 0) /* 0: PPGTT, 1: GGTT */
+
+#define GEN8_OASTATUS _MMIO(0x2b08)
+#define  GEN8_OASTATUS_TAIL_POINTER_WRAP    (1 << 17)
+#define  GEN8_OASTATUS_HEAD_POINTER_WRAP    (1 << 16)
+#define  GEN8_OASTATUS_OVERRUN_STATUS	    (1 << 3)
+#define  GEN8_OASTATUS_COUNTER_OVERFLOW     (1 << 2)
+#define  GEN8_OASTATUS_OABUFFER_OVERFLOW    (1 << 1)
+#define  GEN8_OASTATUS_REPORT_LOST	    (1 << 0)
+
+#define GEN8_OAHEADPTR _MMIO(0x2B0C)
+#define GEN8_OAHEADPTR_MASK    0xffffffc0
+#define GEN8_OATAILPTR _MMIO(0x2B10)
+#define GEN8_OATAILPTR_MASK    0xffffffc0
+
+#define OABUFFER_SIZE_128K  (0 << 3)
+#define OABUFFER_SIZE_256K  (1 << 3)
+#define OABUFFER_SIZE_512K  (2 << 3)
+#define OABUFFER_SIZE_1M    (3 << 3)
+#define OABUFFER_SIZE_2M    (4 << 3)
+#define OABUFFER_SIZE_4M    (5 << 3)
+#define OABUFFER_SIZE_8M    (6 << 3)
+#define OABUFFER_SIZE_16M   (7 << 3)
+
+#define GEN12_OA_TLB_INV_CR _MMIO(0xceec)
+
+/* Gen12 OAR unit */
+#define GEN12_OAR_OACONTROL _MMIO(0x2960)
+#define  GEN12_OAR_OACONTROL_COUNTER_FORMAT_SHIFT 1
+#define  GEN12_OAR_OACONTROL_COUNTER_ENABLE       (1 << 0)
+
+#define GEN12_OACTXCONTROL _MMIO(0x2360)
+#define GEN12_OAR_OASTATUS _MMIO(0x2968)
+
+/* Gen12 OAG unit */
+#define GEN12_OAG_OAHEADPTR _MMIO(0xdb00)
+#define  GEN12_OAG_OAHEADPTR_MASK 0xffffffc0
+#define GEN12_OAG_OATAILPTR _MMIO(0xdb04)
+#define  GEN12_OAG_OATAILPTR_MASK 0xffffffc0
+
+#define GEN12_OAG_OABUFFER  _MMIO(0xdb08)
+#define  GEN12_OAG_OABUFFER_BUFFER_SIZE_MASK  (0x7)
+#define  GEN12_OAG_OABUFFER_BUFFER_SIZE_SHIFT (3)
+#define  GEN12_OAG_OABUFFER_MEMORY_SELECT     (1 << 0) /* 0: PPGTT, 1: GGTT */
+
+#define GEN12_OAG_OAGLBCTXCTRL _MMIO(0x2b28)
+#define  GEN12_OAG_OAGLBCTXCTRL_TIMER_PERIOD_SHIFT 2
+#define  GEN12_OAG_OAGLBCTXCTRL_TIMER_ENABLE       (1 << 1)
+#define  GEN12_OAG_OAGLBCTXCTRL_COUNTER_RESUME     (1 << 0)
+
+#define GEN12_OAG_OACONTROL _MMIO(0xdaf4)
+#define  GEN12_OAG_OACONTROL_OA_COUNTER_FORMAT_SHIFT 2
+#define  GEN12_OAG_OACONTROL_OA_COUNTER_ENABLE       (1 << 0)
+
+#define GEN12_OAG_OA_DEBUG _MMIO(0xdaf8)
+#define  GEN12_OAG_OA_DEBUG_INCLUDE_CLK_RATIO          (1 << 6)
+#define  GEN12_OAG_OA_DEBUG_DISABLE_CLK_RATIO_REPORTS  (1 << 5)
+#define  GEN12_OAG_OA_DEBUG_DISABLE_GO_1_0_REPORTS     (1 << 2)
+#define  GEN12_OAG_OA_DEBUG_DISABLE_CTX_SWITCH_REPORTS (1 << 1)
+
+#define GEN12_OAG_OASTATUS _MMIO(0xdafc)
+#define  GEN12_OAG_OASTATUS_COUNTER_OVERFLOW (1 << 2)
+#define  GEN12_OAG_OASTATUS_BUFFER_OVERFLOW  (1 << 1)
+#define  GEN12_OAG_OASTATUS_REPORT_LOST      (1 << 0)
+
+#define GDT_CHICKEN_BITS    _MMIO(0x9840)
+#define   GT_NOA_ENABLE	    0x00000080
+
+#endif /* __INTEL_PERF_OA_REGS__ */
diff --git a/drivers/gpu/drm/i915/i915_perf_types.h b/drivers/gpu/drm/i915/i915_perf_types.h
new file mode 100644
index 000000000..05cb9a335
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_perf_types.h
@@ -0,0 +1,455 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef _I915_PERF_TYPES_H_
+#define _I915_PERF_TYPES_H_
+
+#include <linux/atomic.h>
+#include <linux/device.h>
+#include <linux/hrtimer.h>
+#include <linux/llist.h>
+#include <linux/poll.h>
+#include <linux/sysfs.h>
+#include <linux/types.h>
+#include <linux/uuid.h>
+#include <linux/wait.h>
+#include <uapi/drm/i915_drm.h>
+
+#include "gt/intel_sseu.h"
+#include "i915_reg_defs.h"
+#include "intel_wakeref.h"
+
+struct drm_i915_private;
+struct file;
+struct i915_active;
+struct i915_gem_context;
+struct i915_perf;
+struct i915_vma;
+struct intel_context;
+struct intel_engine_cs;
+
+struct i915_oa_format {
+	u32 format;
+	int size;
+};
+
+struct i915_oa_reg {
+	i915_reg_t addr;
+	u32 value;
+};
+
+struct i915_oa_config {
+	struct i915_perf *perf;
+
+	char uuid[UUID_STRING_LEN + 1];
+	int id;
+
+	const struct i915_oa_reg *mux_regs;
+	u32 mux_regs_len;
+	const struct i915_oa_reg *b_counter_regs;
+	u32 b_counter_regs_len;
+	const struct i915_oa_reg *flex_regs;
+	u32 flex_regs_len;
+
+	struct attribute_group sysfs_metric;
+	struct attribute *attrs[2];
+	struct kobj_attribute sysfs_metric_id;
+
+	struct kref ref;
+	struct rcu_head rcu;
+};
+
+struct i915_perf_stream;
+
+/**
+ * struct i915_perf_stream_ops - the OPs to support a specific stream type
+ */
+struct i915_perf_stream_ops {
+	/**
+	 * @enable: Enables the collection of HW samples, either in response to
+	 * `I915_PERF_IOCTL_ENABLE` or implicitly called when stream is opened
+	 * without `I915_PERF_FLAG_DISABLED`.
+	 */
+	void (*enable)(struct i915_perf_stream *stream);
+
+	/**
+	 * @disable: Disables the collection of HW samples, either in response
+	 * to `I915_PERF_IOCTL_DISABLE` or implicitly called before destroying
+	 * the stream.
+	 */
+	void (*disable)(struct i915_perf_stream *stream);
+
+	/**
+	 * @poll_wait: Call poll_wait, passing a wait queue that will be woken
+	 * once there is something ready to read() for the stream
+	 */
+	void (*poll_wait)(struct i915_perf_stream *stream,
+			  struct file *file,
+			  poll_table *wait);
+
+	/**
+	 * @wait_unlocked: For handling a blocking read, wait until there is
+	 * something to ready to read() for the stream. E.g. wait on the same
+	 * wait queue that would be passed to poll_wait().
+	 */
+	int (*wait_unlocked)(struct i915_perf_stream *stream);
+
+	/**
+	 * @read: Copy buffered metrics as records to userspace
+	 * **buf**: the userspace, destination buffer
+	 * **count**: the number of bytes to copy, requested by userspace
+	 * **offset**: zero at the start of the read, updated as the read
+	 * proceeds, it represents how many bytes have been copied so far and
+	 * the buffer offset for copying the next record.
+	 *
+	 * Copy as many buffered i915 perf samples and records for this stream
+	 * to userspace as will fit in the given buffer.
+	 *
+	 * Only write complete records; returning -%ENOSPC if there isn't room
+	 * for a complete record.
+	 *
+	 * Return any error condition that results in a short read such as
+	 * -%ENOSPC or -%EFAULT, even though these may be squashed before
+	 * returning to userspace.
+	 */
+	int (*read)(struct i915_perf_stream *stream,
+		    char __user *buf,
+		    size_t count,
+		    size_t *offset);
+
+	/**
+	 * @destroy: Cleanup any stream specific resources.
+	 *
+	 * The stream will always be disabled before this is called.
+	 */
+	void (*destroy)(struct i915_perf_stream *stream);
+};
+
+/**
+ * struct i915_perf_stream - state for a single open stream FD
+ */
+struct i915_perf_stream {
+	/**
+	 * @perf: i915_perf backpointer
+	 */
+	struct i915_perf *perf;
+
+	/**
+	 * @uncore: mmio access path
+	 */
+	struct intel_uncore *uncore;
+
+	/**
+	 * @engine: Engine associated with this performance stream.
+	 */
+	struct intel_engine_cs *engine;
+
+	/**
+	 * @sample_flags: Flags representing the `DRM_I915_PERF_PROP_SAMPLE_*`
+	 * properties given when opening a stream, representing the contents
+	 * of a single sample as read() by userspace.
+	 */
+	u32 sample_flags;
+
+	/**
+	 * @sample_size: Considering the configured contents of a sample
+	 * combined with the required header size, this is the total size
+	 * of a single sample record.
+	 */
+	int sample_size;
+
+	/**
+	 * @ctx: %NULL if measuring system-wide across all contexts or a
+	 * specific context that is being monitored.
+	 */
+	struct i915_gem_context *ctx;
+
+	/**
+	 * @enabled: Whether the stream is currently enabled, considering
+	 * whether the stream was opened in a disabled state and based
+	 * on `I915_PERF_IOCTL_ENABLE` and `I915_PERF_IOCTL_DISABLE` calls.
+	 */
+	bool enabled;
+
+	/**
+	 * @hold_preemption: Whether preemption is put on hold for command
+	 * submissions done on the @ctx. This is useful for some drivers that
+	 * cannot easily post process the OA buffer context to subtract delta
+	 * of performance counters not associated with @ctx.
+	 */
+	bool hold_preemption;
+
+	/**
+	 * @ops: The callbacks providing the implementation of this specific
+	 * type of configured stream.
+	 */
+	const struct i915_perf_stream_ops *ops;
+
+	/**
+	 * @oa_config: The OA configuration used by the stream.
+	 */
+	struct i915_oa_config *oa_config;
+
+	/**
+	 * @oa_config_bos: A list of struct i915_oa_config_bo allocated lazily
+	 * each time @oa_config changes.
+	 */
+	struct llist_head oa_config_bos;
+
+	/**
+	 * @pinned_ctx: The OA context specific information.
+	 */
+	struct intel_context *pinned_ctx;
+
+	/**
+	 * @specific_ctx_id: The id of the specific context.
+	 */
+	u32 specific_ctx_id;
+
+	/**
+	 * @specific_ctx_id_mask: The mask used to masking specific_ctx_id bits.
+	 */
+	u32 specific_ctx_id_mask;
+
+	/**
+	 * @poll_check_timer: High resolution timer that will periodically
+	 * check for data in the circular OA buffer for notifying userspace
+	 * (e.g. during a read() or poll()).
+	 */
+	struct hrtimer poll_check_timer;
+
+	/**
+	 * @poll_wq: The wait queue that hrtimer callback wakes when it
+	 * sees data ready to read in the circular OA buffer.
+	 */
+	wait_queue_head_t poll_wq;
+
+	/**
+	 * @pollin: Whether there is data available to read.
+	 */
+	bool pollin;
+
+	/**
+	 * @periodic: Whether periodic sampling is currently enabled.
+	 */
+	bool periodic;
+
+	/**
+	 * @period_exponent: The OA unit sampling frequency is derived from this.
+	 */
+	int period_exponent;
+
+	/**
+	 * @oa_buffer: State of the OA buffer.
+	 */
+	struct {
+		struct i915_vma *vma;
+		u8 *vaddr;
+		u32 last_ctx_id;
+		int format;
+		int format_size;
+		int size_exponent;
+
+		/**
+		 * @ptr_lock: Locks reads and writes to all head/tail state
+		 *
+		 * Consider: the head and tail pointer state needs to be read
+		 * consistently from a hrtimer callback (atomic context) and
+		 * read() fop (user context) with tail pointer updates happening
+		 * in atomic context and head updates in user context and the
+		 * (unlikely) possibility of read() errors needing to reset all
+		 * head/tail state.
+		 *
+		 * Note: Contention/performance aren't currently a significant
+		 * concern here considering the relatively low frequency of
+		 * hrtimer callbacks (5ms period) and that reads typically only
+		 * happen in response to a hrtimer event and likely complete
+		 * before the next callback.
+		 *
+		 * Note: This lock is not held *while* reading and copying data
+		 * to userspace so the value of head observed in htrimer
+		 * callbacks won't represent any partial consumption of data.
+		 */
+		spinlock_t ptr_lock;
+
+		/**
+		 * @aging_tail: The last HW tail reported by HW. The data
+		 * might not have made it to memory yet though.
+		 */
+		u32 aging_tail;
+
+		/**
+		 * @aging_timestamp: A monotonic timestamp for when the current aging tail pointer
+		 * was read; used to determine when it is old enough to trust.
+		 */
+		u64 aging_timestamp;
+
+		/**
+		 * @head: Although we can always read back the head pointer register,
+		 * we prefer to avoid trusting the HW state, just to avoid any
+		 * risk that some hardware condition could * somehow bump the
+		 * head pointer unpredictably and cause us to forward the wrong
+		 * OA buffer data to userspace.
+		 */
+		u32 head;
+
+		/**
+		 * @tail: The last verified tail that can be read by userspace.
+		 */
+		u32 tail;
+	} oa_buffer;
+
+	/**
+	 * @noa_wait: A batch buffer doing a wait on the GPU for the NOA logic to be
+	 * reprogrammed.
+	 */
+	struct i915_vma *noa_wait;
+
+	/**
+	 * @poll_oa_period: The period in nanoseconds at which the OA
+	 * buffer should be checked for available data.
+	 */
+	u64 poll_oa_period;
+};
+
+/**
+ * struct i915_oa_ops - Gen specific implementation of an OA unit stream
+ */
+struct i915_oa_ops {
+	/**
+	 * @is_valid_b_counter_reg: Validates register's address for
+	 * programming boolean counters for a particular platform.
+	 */
+	bool (*is_valid_b_counter_reg)(struct i915_perf *perf, u32 addr);
+
+	/**
+	 * @is_valid_mux_reg: Validates register's address for programming mux
+	 * for a particular platform.
+	 */
+	bool (*is_valid_mux_reg)(struct i915_perf *perf, u32 addr);
+
+	/**
+	 * @is_valid_flex_reg: Validates register's address for programming
+	 * flex EU filtering for a particular platform.
+	 */
+	bool (*is_valid_flex_reg)(struct i915_perf *perf, u32 addr);
+
+	/**
+	 * @enable_metric_set: Selects and applies any MUX configuration to set
+	 * up the Boolean and Custom (B/C) counters that are part of the
+	 * counter reports being sampled. May apply system constraints such as
+	 * disabling EU clock gating as required.
+	 */
+	int (*enable_metric_set)(struct i915_perf_stream *stream,
+				 struct i915_active *active);
+
+	/**
+	 * @disable_metric_set: Remove system constraints associated with using
+	 * the OA unit.
+	 */
+	void (*disable_metric_set)(struct i915_perf_stream *stream);
+
+	/**
+	 * @oa_enable: Enable periodic sampling
+	 */
+	void (*oa_enable)(struct i915_perf_stream *stream);
+
+	/**
+	 * @oa_disable: Disable periodic sampling
+	 */
+	void (*oa_disable)(struct i915_perf_stream *stream);
+
+	/**
+	 * @read: Copy data from the circular OA buffer into a given userspace
+	 * buffer.
+	 */
+	int (*read)(struct i915_perf_stream *stream,
+		    char __user *buf,
+		    size_t count,
+		    size_t *offset);
+
+	/**
+	 * @oa_hw_tail_read: read the OA tail pointer register
+	 *
+	 * In particular this enables us to share all the fiddly code for
+	 * handling the OA unit tail pointer race that affects multiple
+	 * generations.
+	 */
+	u32 (*oa_hw_tail_read)(struct i915_perf_stream *stream);
+};
+
+struct i915_perf {
+	struct drm_i915_private *i915;
+
+	struct kobject *metrics_kobj;
+
+	/*
+	 * Lock associated with adding/modifying/removing OA configs
+	 * in perf->metrics_idr.
+	 */
+	struct mutex metrics_lock;
+
+	/*
+	 * List of dynamic configurations (struct i915_oa_config), you
+	 * need to hold perf->metrics_lock to access it.
+	 */
+	struct idr metrics_idr;
+
+	/*
+	 * Lock associated with anything below within this structure
+	 * except exclusive_stream.
+	 */
+	struct mutex lock;
+
+	/*
+	 * The stream currently using the OA unit. If accessed
+	 * outside a syscall associated to its file
+	 * descriptor.
+	 */
+	struct i915_perf_stream *exclusive_stream;
+
+	/**
+	 * @sseu: sseu configuration selected to run while perf is active,
+	 * applies to all contexts.
+	 */
+	struct intel_sseu sseu;
+
+	/**
+	 * For rate limiting any notifications of spurious
+	 * invalid OA reports
+	 */
+	struct ratelimit_state spurious_report_rs;
+
+	/**
+	 * For rate limiting any notifications of tail pointer
+	 * race.
+	 */
+	struct ratelimit_state tail_pointer_race;
+
+	u32 gen7_latched_oastatus1;
+	u32 ctx_oactxctrl_offset;
+	u32 ctx_flexeu0_offset;
+
+	/**
+	 * The RPT_ID/reason field for Gen8+ includes a bit
+	 * to determine if the CTX ID in the report is valid
+	 * but the specific bit differs between Gen 8 and 9
+	 */
+	u32 gen8_valid_ctx_bit;
+
+	struct i915_oa_ops ops;
+	const struct i915_oa_format *oa_formats;
+
+	/**
+	 * Use a format mask to store the supported formats
+	 * for a platform.
+	 */
+#define FORMAT_MASK_SIZE DIV_ROUND_UP(I915_OA_FORMAT_MAX - 1, BITS_PER_LONG)
+	unsigned long format_mask[FORMAT_MASK_SIZE];
+
+	atomic64_t noa_programming_delay;
+};
+
+#endif /* _I915_PERF_TYPES_H_ */
diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c
new file mode 100644
index 000000000..89eec5858
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_pmu.c
@@ -0,0 +1,1252 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2017-2018 Intel Corporation
+ */
+
+#include <linux/pm_runtime.h>
+
+#include "gt/intel_engine.h"
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_engine_user.h"
+#include "gt/intel_gt_pm.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_rc6.h"
+#include "gt/intel_rps.h"
+
+#include "i915_drv.h"
+#include "i915_pmu.h"
+#include "intel_pm.h"
+
+/* Frequency for the sampling timer for events which need it. */
+#define FREQUENCY 200
+#define PERIOD max_t(u64, 10000, NSEC_PER_SEC / FREQUENCY)
+
+#define ENGINE_SAMPLE_MASK \
+	(BIT(I915_SAMPLE_BUSY) | \
+	 BIT(I915_SAMPLE_WAIT) | \
+	 BIT(I915_SAMPLE_SEMA))
+
+static cpumask_t i915_pmu_cpumask;
+static unsigned int i915_pmu_target_cpu = -1;
+
+static u8 engine_config_sample(u64 config)
+{
+	return config & I915_PMU_SAMPLE_MASK;
+}
+
+static u8 engine_event_sample(struct perf_event *event)
+{
+	return engine_config_sample(event->attr.config);
+}
+
+static u8 engine_event_class(struct perf_event *event)
+{
+	return (event->attr.config >> I915_PMU_CLASS_SHIFT) & 0xff;
+}
+
+static u8 engine_event_instance(struct perf_event *event)
+{
+	return (event->attr.config >> I915_PMU_SAMPLE_BITS) & 0xff;
+}
+
+static bool is_engine_config(u64 config)
+{
+	return config < __I915_PMU_OTHER(0);
+}
+
+static unsigned int other_bit(const u64 config)
+{
+	unsigned int val;
+
+	switch (config) {
+	case I915_PMU_ACTUAL_FREQUENCY:
+		val =  __I915_PMU_ACTUAL_FREQUENCY_ENABLED;
+		break;
+	case I915_PMU_REQUESTED_FREQUENCY:
+		val = __I915_PMU_REQUESTED_FREQUENCY_ENABLED;
+		break;
+	case I915_PMU_RC6_RESIDENCY:
+		val = __I915_PMU_RC6_RESIDENCY_ENABLED;
+		break;
+	default:
+		/*
+		 * Events that do not require sampling, or tracking state
+		 * transitions between enabled and disabled can be ignored.
+		 */
+		return -1;
+	}
+
+	return I915_ENGINE_SAMPLE_COUNT + val;
+}
+
+static unsigned int config_bit(const u64 config)
+{
+	if (is_engine_config(config))
+		return engine_config_sample(config);
+	else
+		return other_bit(config);
+}
+
+static u64 config_mask(u64 config)
+{
+	return BIT_ULL(config_bit(config));
+}
+
+static bool is_engine_event(struct perf_event *event)
+{
+	return is_engine_config(event->attr.config);
+}
+
+static unsigned int event_bit(struct perf_event *event)
+{
+	return config_bit(event->attr.config);
+}
+
+static bool pmu_needs_timer(struct i915_pmu *pmu, bool gpu_active)
+{
+	struct drm_i915_private *i915 = container_of(pmu, typeof(*i915), pmu);
+	u32 enable;
+
+	/*
+	 * Only some counters need the sampling timer.
+	 *
+	 * We start with a bitmask of all currently enabled events.
+	 */
+	enable = pmu->enable;
+
+	/*
+	 * Mask out all the ones which do not need the timer, or in
+	 * other words keep all the ones that could need the timer.
+	 */
+	enable &= config_mask(I915_PMU_ACTUAL_FREQUENCY) |
+		  config_mask(I915_PMU_REQUESTED_FREQUENCY) |
+		  ENGINE_SAMPLE_MASK;
+
+	/*
+	 * When the GPU is idle per-engine counters do not need to be
+	 * running so clear those bits out.
+	 */
+	if (!gpu_active)
+		enable &= ~ENGINE_SAMPLE_MASK;
+	/*
+	 * Also there is software busyness tracking available we do not
+	 * need the timer for I915_SAMPLE_BUSY counter.
+	 */
+	else if (i915->caps.scheduler & I915_SCHEDULER_CAP_ENGINE_BUSY_STATS)
+		enable &= ~BIT(I915_SAMPLE_BUSY);
+
+	/*
+	 * If some bits remain it means we need the sampling timer running.
+	 */
+	return enable;
+}
+
+static u64 __get_rc6(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	u64 val;
+
+	val = intel_rc6_residency_ns(&gt->rc6, GEN6_GT_GFX_RC6);
+
+	if (HAS_RC6p(i915))
+		val += intel_rc6_residency_ns(&gt->rc6, GEN6_GT_GFX_RC6p);
+
+	if (HAS_RC6pp(i915))
+		val += intel_rc6_residency_ns(&gt->rc6, GEN6_GT_GFX_RC6pp);
+
+	return val;
+}
+
+static inline s64 ktime_since_raw(const ktime_t kt)
+{
+	return ktime_to_ns(ktime_sub(ktime_get_raw(), kt));
+}
+
+static u64 get_rc6(struct intel_gt *gt)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct i915_pmu *pmu = &i915->pmu;
+	unsigned long flags;
+	bool awake = false;
+	u64 val;
+
+	if (intel_gt_pm_get_if_awake(gt)) {
+		val = __get_rc6(gt);
+		intel_gt_pm_put_async(gt);
+		awake = true;
+	}
+
+	spin_lock_irqsave(&pmu->lock, flags);
+
+	if (awake) {
+		pmu->sample[__I915_SAMPLE_RC6].cur = val;
+	} else {
+		/*
+		 * We think we are runtime suspended.
+		 *
+		 * Report the delta from when the device was suspended to now,
+		 * on top of the last known real value, as the approximated RC6
+		 * counter value.
+		 */
+		val = ktime_since_raw(pmu->sleep_last);
+		val += pmu->sample[__I915_SAMPLE_RC6].cur;
+	}
+
+	if (val < pmu->sample[__I915_SAMPLE_RC6_LAST_REPORTED].cur)
+		val = pmu->sample[__I915_SAMPLE_RC6_LAST_REPORTED].cur;
+	else
+		pmu->sample[__I915_SAMPLE_RC6_LAST_REPORTED].cur = val;
+
+	spin_unlock_irqrestore(&pmu->lock, flags);
+
+	return val;
+}
+
+static void init_rc6(struct i915_pmu *pmu)
+{
+	struct drm_i915_private *i915 = container_of(pmu, typeof(*i915), pmu);
+	intel_wakeref_t wakeref;
+
+	with_intel_runtime_pm(to_gt(i915)->uncore->rpm, wakeref) {
+		pmu->sample[__I915_SAMPLE_RC6].cur = __get_rc6(to_gt(i915));
+		pmu->sample[__I915_SAMPLE_RC6_LAST_REPORTED].cur =
+					pmu->sample[__I915_SAMPLE_RC6].cur;
+		pmu->sleep_last = ktime_get_raw();
+	}
+}
+
+static void park_rc6(struct drm_i915_private *i915)
+{
+	struct i915_pmu *pmu = &i915->pmu;
+
+	pmu->sample[__I915_SAMPLE_RC6].cur = __get_rc6(to_gt(i915));
+	pmu->sleep_last = ktime_get_raw();
+}
+
+static void __i915_pmu_maybe_start_timer(struct i915_pmu *pmu)
+{
+	if (!pmu->timer_enabled && pmu_needs_timer(pmu, true)) {
+		pmu->timer_enabled = true;
+		pmu->timer_last = ktime_get();
+		hrtimer_start_range_ns(&pmu->timer,
+				       ns_to_ktime(PERIOD), 0,
+				       HRTIMER_MODE_REL_PINNED);
+	}
+}
+
+void i915_pmu_gt_parked(struct drm_i915_private *i915)
+{
+	struct i915_pmu *pmu = &i915->pmu;
+
+	if (!pmu->base.event_init)
+		return;
+
+	spin_lock_irq(&pmu->lock);
+
+	park_rc6(i915);
+
+	/*
+	 * Signal sampling timer to stop if only engine events are enabled and
+	 * GPU went idle.
+	 */
+	pmu->timer_enabled = pmu_needs_timer(pmu, false);
+
+	spin_unlock_irq(&pmu->lock);
+}
+
+void i915_pmu_gt_unparked(struct drm_i915_private *i915)
+{
+	struct i915_pmu *pmu = &i915->pmu;
+
+	if (!pmu->base.event_init)
+		return;
+
+	spin_lock_irq(&pmu->lock);
+
+	/*
+	 * Re-enable sampling timer when GPU goes active.
+	 */
+	__i915_pmu_maybe_start_timer(pmu);
+
+	spin_unlock_irq(&pmu->lock);
+}
+
+static void
+add_sample(struct i915_pmu_sample *sample, u32 val)
+{
+	sample->cur += val;
+}
+
+static bool exclusive_mmio_access(const struct drm_i915_private *i915)
+{
+	/*
+	 * We have to avoid concurrent mmio cache line access on gen7 or
+	 * risk a machine hang. For a fun history lesson dig out the old
+	 * userspace intel_gpu_top and run it on Ivybridge or Haswell!
+	 */
+	return GRAPHICS_VER(i915) == 7;
+}
+
+static void engine_sample(struct intel_engine_cs *engine, unsigned int period_ns)
+{
+	struct intel_engine_pmu *pmu = &engine->pmu;
+	bool busy;
+	u32 val;
+
+	val = ENGINE_READ_FW(engine, RING_CTL);
+	if (val == 0) /* powerwell off => engine idle */
+		return;
+
+	if (val & RING_WAIT)
+		add_sample(&pmu->sample[I915_SAMPLE_WAIT], period_ns);
+	if (val & RING_WAIT_SEMAPHORE)
+		add_sample(&pmu->sample[I915_SAMPLE_SEMA], period_ns);
+
+	/* No need to sample when busy stats are supported. */
+	if (intel_engine_supports_stats(engine))
+		return;
+
+	/*
+	 * While waiting on a semaphore or event, MI_MODE reports the
+	 * ring as idle. However, previously using the seqno, and with
+	 * execlists sampling, we account for the ring waiting as the
+	 * engine being busy. Therefore, we record the sample as being
+	 * busy if either waiting or !idle.
+	 */
+	busy = val & (RING_WAIT_SEMAPHORE | RING_WAIT);
+	if (!busy) {
+		val = ENGINE_READ_FW(engine, RING_MI_MODE);
+		busy = !(val & MODE_IDLE);
+	}
+	if (busy)
+		add_sample(&pmu->sample[I915_SAMPLE_BUSY], period_ns);
+}
+
+static void
+engines_sample(struct intel_gt *gt, unsigned int period_ns)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	unsigned long flags;
+
+	if ((i915->pmu.enable & ENGINE_SAMPLE_MASK) == 0)
+		return;
+
+	if (!intel_gt_pm_is_awake(gt))
+		return;
+
+	for_each_engine(engine, gt, id) {
+		if (!intel_engine_pm_get_if_awake(engine))
+			continue;
+
+		if (exclusive_mmio_access(i915)) {
+			spin_lock_irqsave(&engine->uncore->lock, flags);
+			engine_sample(engine, period_ns);
+			spin_unlock_irqrestore(&engine->uncore->lock, flags);
+		} else {
+			engine_sample(engine, period_ns);
+		}
+
+		intel_engine_pm_put_async(engine);
+	}
+}
+
+static void
+add_sample_mult(struct i915_pmu_sample *sample, u32 val, u32 mul)
+{
+	sample->cur += mul_u32_u32(val, mul);
+}
+
+static bool frequency_sampling_enabled(struct i915_pmu *pmu)
+{
+	return pmu->enable &
+	       (config_mask(I915_PMU_ACTUAL_FREQUENCY) |
+		config_mask(I915_PMU_REQUESTED_FREQUENCY));
+}
+
+static void
+frequency_sample(struct intel_gt *gt, unsigned int period_ns)
+{
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_uncore *uncore = gt->uncore;
+	struct i915_pmu *pmu = &i915->pmu;
+	struct intel_rps *rps = &gt->rps;
+
+	if (!frequency_sampling_enabled(pmu))
+		return;
+
+	/* Report 0/0 (actual/requested) frequency while parked. */
+	if (!intel_gt_pm_get_if_awake(gt))
+		return;
+
+	if (pmu->enable & config_mask(I915_PMU_ACTUAL_FREQUENCY)) {
+		u32 val;
+
+		/*
+		 * We take a quick peek here without using forcewake
+		 * so that we don't perturb the system under observation
+		 * (forcewake => !rc6 => increased power use). We expect
+		 * that if the read fails because it is outside of the
+		 * mmio power well, then it will return 0 -- in which
+		 * case we assume the system is running at the intended
+		 * frequency. Fortunately, the read should rarely fail!
+		 */
+		val = intel_uncore_read_fw(uncore, GEN6_RPSTAT1);
+		if (val)
+			val = intel_rps_get_cagf(rps, val);
+		else
+			val = rps->cur_freq;
+
+		add_sample_mult(&pmu->sample[__I915_SAMPLE_FREQ_ACT],
+				intel_gpu_freq(rps, val), period_ns / 1000);
+	}
+
+	if (pmu->enable & config_mask(I915_PMU_REQUESTED_FREQUENCY)) {
+		add_sample_mult(&pmu->sample[__I915_SAMPLE_FREQ_REQ],
+				intel_rps_get_requested_frequency(rps),
+				period_ns / 1000);
+	}
+
+	intel_gt_pm_put_async(gt);
+}
+
+static enum hrtimer_restart i915_sample(struct hrtimer *hrtimer)
+{
+	struct drm_i915_private *i915 =
+		container_of(hrtimer, struct drm_i915_private, pmu.timer);
+	struct i915_pmu *pmu = &i915->pmu;
+	struct intel_gt *gt = to_gt(i915);
+	unsigned int period_ns;
+	ktime_t now;
+
+	if (!READ_ONCE(pmu->timer_enabled))
+		return HRTIMER_NORESTART;
+
+	now = ktime_get();
+	period_ns = ktime_to_ns(ktime_sub(now, pmu->timer_last));
+	pmu->timer_last = now;
+
+	/*
+	 * Strictly speaking the passed in period may not be 100% accurate for
+	 * all internal calculation, since some amount of time can be spent on
+	 * grabbing the forcewake. However the potential error from timer call-
+	 * back delay greatly dominates this so we keep it simple.
+	 */
+	engines_sample(gt, period_ns);
+	frequency_sample(gt, period_ns);
+
+	hrtimer_forward(hrtimer, now, ns_to_ktime(PERIOD));
+
+	return HRTIMER_RESTART;
+}
+
+static void i915_pmu_event_destroy(struct perf_event *event)
+{
+	struct drm_i915_private *i915 =
+		container_of(event->pmu, typeof(*i915), pmu.base);
+
+	drm_WARN_ON(&i915->drm, event->parent);
+
+	drm_dev_put(&i915->drm);
+}
+
+static int
+engine_event_status(struct intel_engine_cs *engine,
+		    enum drm_i915_pmu_engine_sample sample)
+{
+	switch (sample) {
+	case I915_SAMPLE_BUSY:
+	case I915_SAMPLE_WAIT:
+		break;
+	case I915_SAMPLE_SEMA:
+		if (GRAPHICS_VER(engine->i915) < 6)
+			return -ENODEV;
+		break;
+	default:
+		return -ENOENT;
+	}
+
+	return 0;
+}
+
+static int
+config_status(struct drm_i915_private *i915, u64 config)
+{
+	struct intel_gt *gt = to_gt(i915);
+
+	switch (config) {
+	case I915_PMU_ACTUAL_FREQUENCY:
+		if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+			/* Requires a mutex for sampling! */
+			return -ENODEV;
+		fallthrough;
+	case I915_PMU_REQUESTED_FREQUENCY:
+		if (GRAPHICS_VER(i915) < 6)
+			return -ENODEV;
+		break;
+	case I915_PMU_INTERRUPTS:
+		break;
+	case I915_PMU_RC6_RESIDENCY:
+		if (!gt->rc6.supported)
+			return -ENODEV;
+		break;
+	case I915_PMU_SOFTWARE_GT_AWAKE_TIME:
+		break;
+	default:
+		return -ENOENT;
+	}
+
+	return 0;
+}
+
+static int engine_event_init(struct perf_event *event)
+{
+	struct drm_i915_private *i915 =
+		container_of(event->pmu, typeof(*i915), pmu.base);
+	struct intel_engine_cs *engine;
+
+	engine = intel_engine_lookup_user(i915, engine_event_class(event),
+					  engine_event_instance(event));
+	if (!engine)
+		return -ENODEV;
+
+	return engine_event_status(engine, engine_event_sample(event));
+}
+
+static int i915_pmu_event_init(struct perf_event *event)
+{
+	struct drm_i915_private *i915 =
+		container_of(event->pmu, typeof(*i915), pmu.base);
+	struct i915_pmu *pmu = &i915->pmu;
+	int ret;
+
+	if (pmu->closed)
+		return -ENODEV;
+
+	if (event->attr.type != event->pmu->type)
+		return -ENOENT;
+
+	/* unsupported modes and filters */
+	if (event->attr.sample_period) /* no sampling */
+		return -EINVAL;
+
+	if (has_branch_stack(event))
+		return -EOPNOTSUPP;
+
+	if (event->cpu < 0)
+		return -EINVAL;
+
+	/* only allow running on one cpu at a time */
+	if (!cpumask_test_cpu(event->cpu, &i915_pmu_cpumask))
+		return -EINVAL;
+
+	if (is_engine_event(event))
+		ret = engine_event_init(event);
+	else
+		ret = config_status(i915, event->attr.config);
+	if (ret)
+		return ret;
+
+	if (!event->parent) {
+		drm_dev_get(&i915->drm);
+		event->destroy = i915_pmu_event_destroy;
+	}
+
+	return 0;
+}
+
+static u64 __i915_pmu_event_read(struct perf_event *event)
+{
+	struct drm_i915_private *i915 =
+		container_of(event->pmu, typeof(*i915), pmu.base);
+	struct i915_pmu *pmu = &i915->pmu;
+	u64 val = 0;
+
+	if (is_engine_event(event)) {
+		u8 sample = engine_event_sample(event);
+		struct intel_engine_cs *engine;
+
+		engine = intel_engine_lookup_user(i915,
+						  engine_event_class(event),
+						  engine_event_instance(event));
+
+		if (drm_WARN_ON_ONCE(&i915->drm, !engine)) {
+			/* Do nothing */
+		} else if (sample == I915_SAMPLE_BUSY &&
+			   intel_engine_supports_stats(engine)) {
+			ktime_t unused;
+
+			val = ktime_to_ns(intel_engine_get_busy_time(engine,
+								     &unused));
+		} else {
+			val = engine->pmu.sample[sample].cur;
+		}
+	} else {
+		switch (event->attr.config) {
+		case I915_PMU_ACTUAL_FREQUENCY:
+			val =
+			   div_u64(pmu->sample[__I915_SAMPLE_FREQ_ACT].cur,
+				   USEC_PER_SEC /* to MHz */);
+			break;
+		case I915_PMU_REQUESTED_FREQUENCY:
+			val =
+			   div_u64(pmu->sample[__I915_SAMPLE_FREQ_REQ].cur,
+				   USEC_PER_SEC /* to MHz */);
+			break;
+		case I915_PMU_INTERRUPTS:
+			val = READ_ONCE(pmu->irq_count);
+			break;
+		case I915_PMU_RC6_RESIDENCY:
+			val = get_rc6(to_gt(i915));
+			break;
+		case I915_PMU_SOFTWARE_GT_AWAKE_TIME:
+			val = ktime_to_ns(intel_gt_get_awake_time(to_gt(i915)));
+			break;
+		}
+	}
+
+	return val;
+}
+
+static void i915_pmu_event_read(struct perf_event *event)
+{
+	struct drm_i915_private *i915 =
+		container_of(event->pmu, typeof(*i915), pmu.base);
+	struct hw_perf_event *hwc = &event->hw;
+	struct i915_pmu *pmu = &i915->pmu;
+	u64 prev, new;
+
+	if (pmu->closed) {
+		event->hw.state = PERF_HES_STOPPED;
+		return;
+	}
+again:
+	prev = local64_read(&hwc->prev_count);
+	new = __i915_pmu_event_read(event);
+
+	if (local64_cmpxchg(&hwc->prev_count, prev, new) != prev)
+		goto again;
+
+	local64_add(new - prev, &event->count);
+}
+
+static void i915_pmu_enable(struct perf_event *event)
+{
+	struct drm_i915_private *i915 =
+		container_of(event->pmu, typeof(*i915), pmu.base);
+	struct i915_pmu *pmu = &i915->pmu;
+	unsigned long flags;
+	unsigned int bit;
+
+	bit = event_bit(event);
+	if (bit == -1)
+		goto update;
+
+	spin_lock_irqsave(&pmu->lock, flags);
+
+	/*
+	 * Update the bitmask of enabled events and increment
+	 * the event reference counter.
+	 */
+	BUILD_BUG_ON(ARRAY_SIZE(pmu->enable_count) != I915_PMU_MASK_BITS);
+	GEM_BUG_ON(bit >= ARRAY_SIZE(pmu->enable_count));
+	GEM_BUG_ON(pmu->enable_count[bit] == ~0);
+
+	pmu->enable |= BIT_ULL(bit);
+	pmu->enable_count[bit]++;
+
+	/*
+	 * Start the sampling timer if needed and not already enabled.
+	 */
+	__i915_pmu_maybe_start_timer(pmu);
+
+	/*
+	 * For per-engine events the bitmask and reference counting
+	 * is stored per engine.
+	 */
+	if (is_engine_event(event)) {
+		u8 sample = engine_event_sample(event);
+		struct intel_engine_cs *engine;
+
+		engine = intel_engine_lookup_user(i915,
+						  engine_event_class(event),
+						  engine_event_instance(event));
+
+		BUILD_BUG_ON(ARRAY_SIZE(engine->pmu.enable_count) !=
+			     I915_ENGINE_SAMPLE_COUNT);
+		BUILD_BUG_ON(ARRAY_SIZE(engine->pmu.sample) !=
+			     I915_ENGINE_SAMPLE_COUNT);
+		GEM_BUG_ON(sample >= ARRAY_SIZE(engine->pmu.enable_count));
+		GEM_BUG_ON(sample >= ARRAY_SIZE(engine->pmu.sample));
+		GEM_BUG_ON(engine->pmu.enable_count[sample] == ~0);
+
+		engine->pmu.enable |= BIT(sample);
+		engine->pmu.enable_count[sample]++;
+	}
+
+	spin_unlock_irqrestore(&pmu->lock, flags);
+
+update:
+	/*
+	 * Store the current counter value so we can report the correct delta
+	 * for all listeners. Even when the event was already enabled and has
+	 * an existing non-zero value.
+	 */
+	local64_set(&event->hw.prev_count, __i915_pmu_event_read(event));
+}
+
+static void i915_pmu_disable(struct perf_event *event)
+{
+	struct drm_i915_private *i915 =
+		container_of(event->pmu, typeof(*i915), pmu.base);
+	unsigned int bit = event_bit(event);
+	struct i915_pmu *pmu = &i915->pmu;
+	unsigned long flags;
+
+	if (bit == -1)
+		return;
+
+	spin_lock_irqsave(&pmu->lock, flags);
+
+	if (is_engine_event(event)) {
+		u8 sample = engine_event_sample(event);
+		struct intel_engine_cs *engine;
+
+		engine = intel_engine_lookup_user(i915,
+						  engine_event_class(event),
+						  engine_event_instance(event));
+
+		GEM_BUG_ON(sample >= ARRAY_SIZE(engine->pmu.enable_count));
+		GEM_BUG_ON(sample >= ARRAY_SIZE(engine->pmu.sample));
+		GEM_BUG_ON(engine->pmu.enable_count[sample] == 0);
+
+		/*
+		 * Decrement the reference count and clear the enabled
+		 * bitmask when the last listener on an event goes away.
+		 */
+		if (--engine->pmu.enable_count[sample] == 0)
+			engine->pmu.enable &= ~BIT(sample);
+	}
+
+	GEM_BUG_ON(bit >= ARRAY_SIZE(pmu->enable_count));
+	GEM_BUG_ON(pmu->enable_count[bit] == 0);
+	/*
+	 * Decrement the reference count and clear the enabled
+	 * bitmask when the last listener on an event goes away.
+	 */
+	if (--pmu->enable_count[bit] == 0) {
+		pmu->enable &= ~BIT_ULL(bit);
+		pmu->timer_enabled &= pmu_needs_timer(pmu, true);
+	}
+
+	spin_unlock_irqrestore(&pmu->lock, flags);
+}
+
+static void i915_pmu_event_start(struct perf_event *event, int flags)
+{
+	struct drm_i915_private *i915 =
+		container_of(event->pmu, typeof(*i915), pmu.base);
+	struct i915_pmu *pmu = &i915->pmu;
+
+	if (pmu->closed)
+		return;
+
+	i915_pmu_enable(event);
+	event->hw.state = 0;
+}
+
+static void i915_pmu_event_stop(struct perf_event *event, int flags)
+{
+	struct drm_i915_private *i915 =
+		container_of(event->pmu, typeof(*i915), pmu.base);
+	struct i915_pmu *pmu = &i915->pmu;
+
+	if (pmu->closed)
+		goto out;
+
+	if (flags & PERF_EF_UPDATE)
+		i915_pmu_event_read(event);
+	i915_pmu_disable(event);
+
+out:
+	event->hw.state = PERF_HES_STOPPED;
+}
+
+static int i915_pmu_event_add(struct perf_event *event, int flags)
+{
+	struct drm_i915_private *i915 =
+		container_of(event->pmu, typeof(*i915), pmu.base);
+	struct i915_pmu *pmu = &i915->pmu;
+
+	if (pmu->closed)
+		return -ENODEV;
+
+	if (flags & PERF_EF_START)
+		i915_pmu_event_start(event, flags);
+
+	return 0;
+}
+
+static void i915_pmu_event_del(struct perf_event *event, int flags)
+{
+	i915_pmu_event_stop(event, PERF_EF_UPDATE);
+}
+
+static int i915_pmu_event_event_idx(struct perf_event *event)
+{
+	return 0;
+}
+
+struct i915_str_attribute {
+	struct device_attribute attr;
+	const char *str;
+};
+
+static ssize_t i915_pmu_format_show(struct device *dev,
+				    struct device_attribute *attr, char *buf)
+{
+	struct i915_str_attribute *eattr;
+
+	eattr = container_of(attr, struct i915_str_attribute, attr);
+	return sprintf(buf, "%s\n", eattr->str);
+}
+
+#define I915_PMU_FORMAT_ATTR(_name, _config) \
+	(&((struct i915_str_attribute[]) { \
+		{ .attr = __ATTR(_name, 0444, i915_pmu_format_show, NULL), \
+		  .str = _config, } \
+	})[0].attr.attr)
+
+static struct attribute *i915_pmu_format_attrs[] = {
+	I915_PMU_FORMAT_ATTR(i915_eventid, "config:0-20"),
+	NULL,
+};
+
+static const struct attribute_group i915_pmu_format_attr_group = {
+	.name = "format",
+	.attrs = i915_pmu_format_attrs,
+};
+
+struct i915_ext_attribute {
+	struct device_attribute attr;
+	unsigned long val;
+};
+
+static ssize_t i915_pmu_event_show(struct device *dev,
+				   struct device_attribute *attr, char *buf)
+{
+	struct i915_ext_attribute *eattr;
+
+	eattr = container_of(attr, struct i915_ext_attribute, attr);
+	return sprintf(buf, "config=0x%lx\n", eattr->val);
+}
+
+static ssize_t cpumask_show(struct device *dev,
+			    struct device_attribute *attr, char *buf)
+{
+	return cpumap_print_to_pagebuf(true, buf, &i915_pmu_cpumask);
+}
+
+static DEVICE_ATTR_RO(cpumask);
+
+static struct attribute *i915_cpumask_attrs[] = {
+	&dev_attr_cpumask.attr,
+	NULL,
+};
+
+static const struct attribute_group i915_pmu_cpumask_attr_group = {
+	.attrs = i915_cpumask_attrs,
+};
+
+#define __event(__config, __name, __unit) \
+{ \
+	.config = (__config), \
+	.name = (__name), \
+	.unit = (__unit), \
+}
+
+#define __engine_event(__sample, __name) \
+{ \
+	.sample = (__sample), \
+	.name = (__name), \
+}
+
+static struct i915_ext_attribute *
+add_i915_attr(struct i915_ext_attribute *attr, const char *name, u64 config)
+{
+	sysfs_attr_init(&attr->attr.attr);
+	attr->attr.attr.name = name;
+	attr->attr.attr.mode = 0444;
+	attr->attr.show = i915_pmu_event_show;
+	attr->val = config;
+
+	return ++attr;
+}
+
+static struct perf_pmu_events_attr *
+add_pmu_attr(struct perf_pmu_events_attr *attr, const char *name,
+	     const char *str)
+{
+	sysfs_attr_init(&attr->attr.attr);
+	attr->attr.attr.name = name;
+	attr->attr.attr.mode = 0444;
+	attr->attr.show = perf_event_sysfs_show;
+	attr->event_str = str;
+
+	return ++attr;
+}
+
+static struct attribute **
+create_event_attributes(struct i915_pmu *pmu)
+{
+	struct drm_i915_private *i915 = container_of(pmu, typeof(*i915), pmu);
+	static const struct {
+		u64 config;
+		const char *name;
+		const char *unit;
+	} events[] = {
+		__event(I915_PMU_ACTUAL_FREQUENCY, "actual-frequency", "M"),
+		__event(I915_PMU_REQUESTED_FREQUENCY, "requested-frequency", "M"),
+		__event(I915_PMU_INTERRUPTS, "interrupts", NULL),
+		__event(I915_PMU_RC6_RESIDENCY, "rc6-residency", "ns"),
+		__event(I915_PMU_SOFTWARE_GT_AWAKE_TIME, "software-gt-awake-time", "ns"),
+	};
+	static const struct {
+		enum drm_i915_pmu_engine_sample sample;
+		char *name;
+	} engine_events[] = {
+		__engine_event(I915_SAMPLE_BUSY, "busy"),
+		__engine_event(I915_SAMPLE_SEMA, "sema"),
+		__engine_event(I915_SAMPLE_WAIT, "wait"),
+	};
+	unsigned int count = 0;
+	struct perf_pmu_events_attr *pmu_attr = NULL, *pmu_iter;
+	struct i915_ext_attribute *i915_attr = NULL, *i915_iter;
+	struct attribute **attr = NULL, **attr_iter;
+	struct intel_engine_cs *engine;
+	unsigned int i;
+
+	/* Count how many counters we will be exposing. */
+	for (i = 0; i < ARRAY_SIZE(events); i++) {
+		if (!config_status(i915, events[i].config))
+			count++;
+	}
+
+	for_each_uabi_engine(engine, i915) {
+		for (i = 0; i < ARRAY_SIZE(engine_events); i++) {
+			if (!engine_event_status(engine,
+						 engine_events[i].sample))
+				count++;
+		}
+	}
+
+	/* Allocate attribute objects and table. */
+	i915_attr = kcalloc(count, sizeof(*i915_attr), GFP_KERNEL);
+	if (!i915_attr)
+		goto err_alloc;
+
+	pmu_attr = kcalloc(count, sizeof(*pmu_attr), GFP_KERNEL);
+	if (!pmu_attr)
+		goto err_alloc;
+
+	/* Max one pointer of each attribute type plus a termination entry. */
+	attr = kcalloc(count * 2 + 1, sizeof(*attr), GFP_KERNEL);
+	if (!attr)
+		goto err_alloc;
+
+	i915_iter = i915_attr;
+	pmu_iter = pmu_attr;
+	attr_iter = attr;
+
+	/* Initialize supported non-engine counters. */
+	for (i = 0; i < ARRAY_SIZE(events); i++) {
+		char *str;
+
+		if (config_status(i915, events[i].config))
+			continue;
+
+		str = kstrdup(events[i].name, GFP_KERNEL);
+		if (!str)
+			goto err;
+
+		*attr_iter++ = &i915_iter->attr.attr;
+		i915_iter = add_i915_attr(i915_iter, str, events[i].config);
+
+		if (events[i].unit) {
+			str = kasprintf(GFP_KERNEL, "%s.unit", events[i].name);
+			if (!str)
+				goto err;
+
+			*attr_iter++ = &pmu_iter->attr.attr;
+			pmu_iter = add_pmu_attr(pmu_iter, str, events[i].unit);
+		}
+	}
+
+	/* Initialize supported engine counters. */
+	for_each_uabi_engine(engine, i915) {
+		for (i = 0; i < ARRAY_SIZE(engine_events); i++) {
+			char *str;
+
+			if (engine_event_status(engine,
+						engine_events[i].sample))
+				continue;
+
+			str = kasprintf(GFP_KERNEL, "%s-%s",
+					engine->name, engine_events[i].name);
+			if (!str)
+				goto err;
+
+			*attr_iter++ = &i915_iter->attr.attr;
+			i915_iter =
+				add_i915_attr(i915_iter, str,
+					      __I915_PMU_ENGINE(engine->uabi_class,
+								engine->uabi_instance,
+								engine_events[i].sample));
+
+			str = kasprintf(GFP_KERNEL, "%s-%s.unit",
+					engine->name, engine_events[i].name);
+			if (!str)
+				goto err;
+
+			*attr_iter++ = &pmu_iter->attr.attr;
+			pmu_iter = add_pmu_attr(pmu_iter, str, "ns");
+		}
+	}
+
+	pmu->i915_attr = i915_attr;
+	pmu->pmu_attr = pmu_attr;
+
+	return attr;
+
+err:;
+	for (attr_iter = attr; *attr_iter; attr_iter++)
+		kfree((*attr_iter)->name);
+
+err_alloc:
+	kfree(attr);
+	kfree(i915_attr);
+	kfree(pmu_attr);
+
+	return NULL;
+}
+
+static void free_event_attributes(struct i915_pmu *pmu)
+{
+	struct attribute **attr_iter = pmu->events_attr_group.attrs;
+
+	for (; *attr_iter; attr_iter++)
+		kfree((*attr_iter)->name);
+
+	kfree(pmu->events_attr_group.attrs);
+	kfree(pmu->i915_attr);
+	kfree(pmu->pmu_attr);
+
+	pmu->events_attr_group.attrs = NULL;
+	pmu->i915_attr = NULL;
+	pmu->pmu_attr = NULL;
+}
+
+static int i915_pmu_cpu_online(unsigned int cpu, struct hlist_node *node)
+{
+	struct i915_pmu *pmu = hlist_entry_safe(node, typeof(*pmu), cpuhp.node);
+
+	GEM_BUG_ON(!pmu->base.event_init);
+
+	/* Select the first online CPU as a designated reader. */
+	if (cpumask_empty(&i915_pmu_cpumask))
+		cpumask_set_cpu(cpu, &i915_pmu_cpumask);
+
+	return 0;
+}
+
+static int i915_pmu_cpu_offline(unsigned int cpu, struct hlist_node *node)
+{
+	struct i915_pmu *pmu = hlist_entry_safe(node, typeof(*pmu), cpuhp.node);
+	unsigned int target = i915_pmu_target_cpu;
+
+	GEM_BUG_ON(!pmu->base.event_init);
+
+	/*
+	 * Unregistering an instance generates a CPU offline event which we must
+	 * ignore to avoid incorrectly modifying the shared i915_pmu_cpumask.
+	 */
+	if (pmu->closed)
+		return 0;
+
+	if (cpumask_test_and_clear_cpu(cpu, &i915_pmu_cpumask)) {
+		target = cpumask_any_but(topology_sibling_cpumask(cpu), cpu);
+
+		/* Migrate events if there is a valid target */
+		if (target < nr_cpu_ids) {
+			cpumask_set_cpu(target, &i915_pmu_cpumask);
+			i915_pmu_target_cpu = target;
+		}
+	}
+
+	if (target < nr_cpu_ids && target != pmu->cpuhp.cpu) {
+		perf_pmu_migrate_context(&pmu->base, cpu, target);
+		pmu->cpuhp.cpu = target;
+	}
+
+	return 0;
+}
+
+static enum cpuhp_state cpuhp_slot = CPUHP_INVALID;
+
+int i915_pmu_init(void)
+{
+	int ret;
+
+	ret = cpuhp_setup_state_multi(CPUHP_AP_ONLINE_DYN,
+				      "perf/x86/intel/i915:online",
+				      i915_pmu_cpu_online,
+				      i915_pmu_cpu_offline);
+	if (ret < 0)
+		pr_notice("Failed to setup cpuhp state for i915 PMU! (%d)\n",
+			  ret);
+	else
+		cpuhp_slot = ret;
+
+	return 0;
+}
+
+void i915_pmu_exit(void)
+{
+	if (cpuhp_slot != CPUHP_INVALID)
+		cpuhp_remove_multi_state(cpuhp_slot);
+}
+
+static int i915_pmu_register_cpuhp_state(struct i915_pmu *pmu)
+{
+	if (cpuhp_slot == CPUHP_INVALID)
+		return -EINVAL;
+
+	return cpuhp_state_add_instance(cpuhp_slot, &pmu->cpuhp.node);
+}
+
+static void i915_pmu_unregister_cpuhp_state(struct i915_pmu *pmu)
+{
+	cpuhp_state_remove_instance(cpuhp_slot, &pmu->cpuhp.node);
+}
+
+static bool is_igp(struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+
+	/* IGP is 0000:00:02.0 */
+	return pci_domain_nr(pdev->bus) == 0 &&
+	       pdev->bus->number == 0 &&
+	       PCI_SLOT(pdev->devfn) == 2 &&
+	       PCI_FUNC(pdev->devfn) == 0;
+}
+
+void i915_pmu_register(struct drm_i915_private *i915)
+{
+	struct i915_pmu *pmu = &i915->pmu;
+	const struct attribute_group *attr_groups[] = {
+		&i915_pmu_format_attr_group,
+		&pmu->events_attr_group,
+		&i915_pmu_cpumask_attr_group,
+		NULL
+	};
+
+	int ret = -ENOMEM;
+
+	if (GRAPHICS_VER(i915) <= 2) {
+		drm_info(&i915->drm, "PMU not supported for this GPU.");
+		return;
+	}
+
+	spin_lock_init(&pmu->lock);
+	hrtimer_init(&pmu->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+	pmu->timer.function = i915_sample;
+	pmu->cpuhp.cpu = -1;
+	init_rc6(pmu);
+
+	if (!is_igp(i915)) {
+		pmu->name = kasprintf(GFP_KERNEL,
+				      "i915_%s",
+				      dev_name(i915->drm.dev));
+		if (pmu->name) {
+			/* tools/perf reserves colons as special. */
+			strreplace((char *)pmu->name, ':', '_');
+		}
+	} else {
+		pmu->name = "i915";
+	}
+	if (!pmu->name)
+		goto err;
+
+	pmu->events_attr_group.name = "events";
+	pmu->events_attr_group.attrs = create_event_attributes(pmu);
+	if (!pmu->events_attr_group.attrs)
+		goto err_name;
+
+	pmu->base.attr_groups = kmemdup(attr_groups, sizeof(attr_groups),
+					GFP_KERNEL);
+	if (!pmu->base.attr_groups)
+		goto err_attr;
+
+	pmu->base.module	= THIS_MODULE;
+	pmu->base.task_ctx_nr	= perf_invalid_context;
+	pmu->base.event_init	= i915_pmu_event_init;
+	pmu->base.add		= i915_pmu_event_add;
+	pmu->base.del		= i915_pmu_event_del;
+	pmu->base.start		= i915_pmu_event_start;
+	pmu->base.stop		= i915_pmu_event_stop;
+	pmu->base.read		= i915_pmu_event_read;
+	pmu->base.event_idx	= i915_pmu_event_event_idx;
+
+	ret = perf_pmu_register(&pmu->base, pmu->name, -1);
+	if (ret)
+		goto err_groups;
+
+	ret = i915_pmu_register_cpuhp_state(pmu);
+	if (ret)
+		goto err_unreg;
+
+	return;
+
+err_unreg:
+	perf_pmu_unregister(&pmu->base);
+err_groups:
+	kfree(pmu->base.attr_groups);
+err_attr:
+	pmu->base.event_init = NULL;
+	free_event_attributes(pmu);
+err_name:
+	if (!is_igp(i915))
+		kfree(pmu->name);
+err:
+	drm_notice(&i915->drm, "Failed to register PMU!\n");
+}
+
+void i915_pmu_unregister(struct drm_i915_private *i915)
+{
+	struct i915_pmu *pmu = &i915->pmu;
+
+	if (!pmu->base.event_init)
+		return;
+
+	/*
+	 * "Disconnect" the PMU callbacks - since all are atomic synchronize_rcu
+	 * ensures all currently executing ones will have exited before we
+	 * proceed with unregistration.
+	 */
+	pmu->closed = true;
+	synchronize_rcu();
+
+	hrtimer_cancel(&pmu->timer);
+
+	i915_pmu_unregister_cpuhp_state(pmu);
+
+	perf_pmu_unregister(&pmu->base);
+	pmu->base.event_init = NULL;
+	kfree(pmu->base.attr_groups);
+	if (!is_igp(i915))
+		kfree(pmu->name);
+	free_event_attributes(pmu);
+}
diff --git a/drivers/gpu/drm/i915/i915_pmu.h b/drivers/gpu/drm/i915/i915_pmu.h
new file mode 100644
index 000000000..449057648
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_pmu.h
@@ -0,0 +1,165 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2017-2018 Intel Corporation
+ */
+
+#ifndef __I915_PMU_H__
+#define __I915_PMU_H__
+
+#include <linux/hrtimer.h>
+#include <linux/perf_event.h>
+#include <linux/spinlock_types.h>
+#include <uapi/drm/i915_drm.h>
+
+struct drm_i915_private;
+
+/**
+ * Non-engine events that we need to track enabled-disabled transition and
+ * current state.
+ */
+enum i915_pmu_tracked_events {
+	__I915_PMU_ACTUAL_FREQUENCY_ENABLED = 0,
+	__I915_PMU_REQUESTED_FREQUENCY_ENABLED,
+	__I915_PMU_RC6_RESIDENCY_ENABLED,
+	__I915_PMU_TRACKED_EVENT_COUNT, /* count marker */
+};
+
+/**
+ * Slots used from the sampling timer (non-engine events) with some extras for
+ * convenience.
+ */
+enum {
+	__I915_SAMPLE_FREQ_ACT = 0,
+	__I915_SAMPLE_FREQ_REQ,
+	__I915_SAMPLE_RC6,
+	__I915_SAMPLE_RC6_LAST_REPORTED,
+	__I915_NUM_PMU_SAMPLERS
+};
+
+/**
+ * How many different events we track in the global PMU mask.
+ *
+ * It is also used to know to needed number of event reference counters.
+ */
+#define I915_PMU_MASK_BITS \
+	(I915_ENGINE_SAMPLE_COUNT + __I915_PMU_TRACKED_EVENT_COUNT)
+
+#define I915_ENGINE_SAMPLE_COUNT (I915_SAMPLE_SEMA + 1)
+
+struct i915_pmu_sample {
+	u64 cur;
+};
+
+struct i915_pmu {
+	/**
+	 * @cpuhp: Struct used for CPU hotplug handling.
+	 */
+	struct {
+		struct hlist_node node;
+		unsigned int cpu;
+	} cpuhp;
+	/**
+	 * @base: PMU base.
+	 */
+	struct pmu base;
+	/**
+	 * @closed: i915 is unregistering.
+	 */
+	bool closed;
+	/**
+	 * @name: Name as registered with perf core.
+	 */
+	const char *name;
+	/**
+	 * @lock: Lock protecting enable mask and ref count handling.
+	 */
+	spinlock_t lock;
+	/**
+	 * @timer: Timer for internal i915 PMU sampling.
+	 */
+	struct hrtimer timer;
+	/**
+	 * @enable: Bitmask of specific enabled events.
+	 *
+	 * For some events we need to track their state and do some internal
+	 * house keeping.
+	 *
+	 * Each engine event sampler type and event listed in enum
+	 * i915_pmu_tracked_events gets a bit in this field.
+	 *
+	 * Low bits are engine samplers and other events continue from there.
+	 */
+	u32 enable;
+
+	/**
+	 * @timer_last:
+	 *
+	 * Timestmap of the previous timer invocation.
+	 */
+	ktime_t timer_last;
+
+	/**
+	 * @enable_count: Reference counts for the enabled events.
+	 *
+	 * Array indices are mapped in the same way as bits in the @enable field
+	 * and they are used to control sampling on/off when multiple clients
+	 * are using the PMU API.
+	 */
+	unsigned int enable_count[I915_PMU_MASK_BITS];
+	/**
+	 * @timer_enabled: Should the internal sampling timer be running.
+	 */
+	bool timer_enabled;
+	/**
+	 * @sample: Current and previous (raw) counters for sampling events.
+	 *
+	 * These counters are updated from the i915 PMU sampling timer.
+	 *
+	 * Only global counters are held here, while the per-engine ones are in
+	 * struct intel_engine_cs.
+	 */
+	struct i915_pmu_sample sample[__I915_NUM_PMU_SAMPLERS];
+	/**
+	 * @sleep_last: Last time GT parked for RC6 estimation.
+	 */
+	ktime_t sleep_last;
+	/**
+	 * @irq_count: Number of interrupts
+	 *
+	 * Intentionally unsigned long to avoid atomics or heuristics on 32bit.
+	 * 4e9 interrupts are a lot and postprocessing can really deal with an
+	 * occasional wraparound easily. It's 32bit after all.
+	 */
+	unsigned long irq_count;
+	/**
+	 * @events_attr_group: Device events attribute group.
+	 */
+	struct attribute_group events_attr_group;
+	/**
+	 * @i915_attr: Memory block holding device attributes.
+	 */
+	void *i915_attr;
+	/**
+	 * @pmu_attr: Memory block holding device attributes.
+	 */
+	void *pmu_attr;
+};
+
+#ifdef CONFIG_PERF_EVENTS
+int i915_pmu_init(void);
+void i915_pmu_exit(void);
+void i915_pmu_register(struct drm_i915_private *i915);
+void i915_pmu_unregister(struct drm_i915_private *i915);
+void i915_pmu_gt_parked(struct drm_i915_private *i915);
+void i915_pmu_gt_unparked(struct drm_i915_private *i915);
+#else
+static inline int i915_pmu_init(void) { return 0; }
+static inline void i915_pmu_exit(void) {}
+static inline void i915_pmu_register(struct drm_i915_private *i915) {}
+static inline void i915_pmu_unregister(struct drm_i915_private *i915) {}
+static inline void i915_pmu_gt_parked(struct drm_i915_private *i915) {}
+static inline void i915_pmu_gt_unparked(struct drm_i915_private *i915) {}
+#endif
+
+#endif
diff --git a/drivers/gpu/drm/i915/i915_priolist_types.h b/drivers/gpu/drm/i915/i915_priolist_types.h
new file mode 100644
index 000000000..bc2fa84f9
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_priolist_types.h
@@ -0,0 +1,47 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef _I915_PRIOLIST_TYPES_H_
+#define _I915_PRIOLIST_TYPES_H_
+
+#include <linux/list.h>
+#include <linux/rbtree.h>
+
+#include <uapi/drm/i915_drm.h>
+
+enum {
+	I915_PRIORITY_MIN = I915_CONTEXT_MIN_USER_PRIORITY - 1,
+	I915_PRIORITY_NORMAL = I915_CONTEXT_DEFAULT_PRIORITY,
+	I915_PRIORITY_MAX = I915_CONTEXT_MAX_USER_PRIORITY + 1,
+
+	/* A preemptive pulse used to monitor the health of each engine */
+	I915_PRIORITY_HEARTBEAT,
+
+	/* Interactive workload, scheduled for immediate pageflipping */
+	I915_PRIORITY_DISPLAY,
+};
+
+/* Smallest priority value that cannot be bumped. */
+#define I915_PRIORITY_INVALID (INT_MIN)
+
+/*
+ * Requests containing performance queries must not be preempted by
+ * another context. They get scheduled with their default priority and
+ * once they reach the execlist ports we ensure that they stick on the
+ * HW until finished by pretending that they have maximum priority,
+ * i.e. nothing can have higher priority and force us to usurp the
+ * active request.
+ */
+#define I915_PRIORITY_UNPREEMPTABLE INT_MAX
+#define I915_PRIORITY_BARRIER (I915_PRIORITY_UNPREEMPTABLE - 1)
+
+struct i915_priolist {
+	struct list_head requests;
+	struct rb_node node;
+	int priority;
+};
+
+#endif /* _I915_PRIOLIST_TYPES_H_ */
diff --git a/drivers/gpu/drm/i915/i915_pvinfo.h b/drivers/gpu/drm/i915/i915_pvinfo.h
new file mode 100644
index 000000000..683e97ac2
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_pvinfo.h
@@ -0,0 +1,123 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef _I915_PVINFO_H_
+#define _I915_PVINFO_H_
+
+#include <linux/types.h>
+
+/* The MMIO offset of the shared info between guest and host emulator */
+#define VGT_PVINFO_PAGE	0x78000
+#define VGT_PVINFO_SIZE	0x1000
+
+/*
+ * The following structure pages are defined in GEN MMIO space
+ * for virtualization. (One page for now)
+ */
+#define VGT_MAGIC         0x4776544776544776ULL	/* 'vGTvGTvG' */
+#define VGT_VERSION_MAJOR 1
+#define VGT_VERSION_MINOR 0
+
+/*
+ * notifications from guest to vgpu device model
+ */
+enum vgt_g2v_type {
+	VGT_G2V_PPGTT_L3_PAGE_TABLE_CREATE = 2,
+	VGT_G2V_PPGTT_L3_PAGE_TABLE_DESTROY,
+	VGT_G2V_PPGTT_L4_PAGE_TABLE_CREATE,
+	VGT_G2V_PPGTT_L4_PAGE_TABLE_DESTROY,
+	VGT_G2V_EXECLIST_CONTEXT_CREATE,
+	VGT_G2V_EXECLIST_CONTEXT_DESTROY,
+	VGT_G2V_MAX,
+};
+
+/*
+ * VGT capabilities type
+ */
+#define VGT_CAPS_FULL_PPGTT		BIT(2)
+#define VGT_CAPS_HWSP_EMULATION		BIT(3)
+#define VGT_CAPS_HUGE_GTT		BIT(4)
+
+struct vgt_if {
+	u64 magic;		/* VGT_MAGIC */
+	u16 version_major;
+	u16 version_minor;
+	u32 vgt_id;		/* ID of vGT instance */
+	u32 vgt_caps;		/* VGT capabilities */
+	u32 rsv1[11];		/* pad to offset 0x40 */
+	/*
+	 *  Data structure to describe the balooning info of resources.
+	 *  Each VM can only have one portion of continuous area for now.
+	 *  (May support scattered resource in future)
+	 *  (starting from offset 0x40)
+	 */
+	struct {
+		/* Aperture register balooning */
+		struct {
+			u32 base;
+			u32 size;
+		} mappable_gmadr;	/* aperture */
+		/* GMADR register balooning */
+		struct {
+			u32 base;
+			u32 size;
+		} nonmappable_gmadr;	/* non aperture */
+		/* allowed fence registers */
+		u32 fence_num;
+		u32 rsv2[3];
+	} avail_rs;		/* available/assigned resource */
+	u32 rsv3[0x200 - 24];	/* pad to half page */
+	/*
+	 * The bottom half page is for response from Gfx driver to hypervisor.
+	 */
+	u32 rsv4;
+	u32 display_ready;	/* ready for display owner switch */
+
+	u32 rsv5[4];
+
+	u32 g2v_notify;
+	u32 rsv6[5];
+
+	u32 cursor_x_hot;
+	u32 cursor_y_hot;
+
+	struct {
+		u32 lo;
+		u32 hi;
+	} pdp[4];
+
+	u32 execlist_context_descriptor_lo;
+	u32 execlist_context_descriptor_hi;
+
+	u32  rsv7[0x200 - 24];    /* pad to one page */
+} __packed;
+
+#define vgtif_offset(x) (offsetof(struct vgt_if, x))
+
+#define vgtif_reg(x) _MMIO(VGT_PVINFO_PAGE + vgtif_offset(x))
+
+/* vGPU display status to be used by the host side */
+#define VGT_DRV_DISPLAY_NOT_READY 0
+#define VGT_DRV_DISPLAY_READY     1  /* ready for display switch */
+
+#endif /* _I915_PVINFO_H_ */
diff --git a/drivers/gpu/drm/i915/i915_query.c b/drivers/gpu/drm/i915/i915_query.c
new file mode 100644
index 000000000..6ec9c9fb7
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_query.c
@@ -0,0 +1,602 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/nospec.h>
+
+#include "i915_drv.h"
+#include "i915_perf.h"
+#include "i915_query.h"
+#include "gt/intel_engine_user.h"
+#include <uapi/drm/i915_drm.h>
+
+static int copy_query_item(void *query_hdr, size_t query_sz,
+			   u32 total_length,
+			   struct drm_i915_query_item *query_item)
+{
+	if (query_item->length == 0)
+		return total_length;
+
+	if (query_item->length < total_length)
+		return -EINVAL;
+
+	if (copy_from_user(query_hdr, u64_to_user_ptr(query_item->data_ptr),
+			   query_sz))
+		return -EFAULT;
+
+	return 0;
+}
+
+static int fill_topology_info(const struct sseu_dev_info *sseu,
+			      struct drm_i915_query_item *query_item,
+			      intel_sseu_ss_mask_t subslice_mask)
+{
+	struct drm_i915_query_topology_info topo;
+	u32 slice_length, subslice_length, eu_length, total_length;
+	int ss_stride = GEN_SSEU_STRIDE(sseu->max_subslices);
+	int eu_stride = GEN_SSEU_STRIDE(sseu->max_eus_per_subslice);
+	int ret;
+
+	BUILD_BUG_ON(sizeof(u8) != sizeof(sseu->slice_mask));
+
+	if (sseu->max_slices == 0)
+		return -ENODEV;
+
+	slice_length = sizeof(sseu->slice_mask);
+	subslice_length = sseu->max_slices * ss_stride;
+	eu_length = sseu->max_slices * sseu->max_subslices * eu_stride;
+	total_length = sizeof(topo) + slice_length + subslice_length +
+		       eu_length;
+
+	ret = copy_query_item(&topo, sizeof(topo), total_length, query_item);
+
+	if (ret != 0)
+		return ret;
+
+	memset(&topo, 0, sizeof(topo));
+	topo.max_slices = sseu->max_slices;
+	topo.max_subslices = sseu->max_subslices;
+	topo.max_eus_per_subslice = sseu->max_eus_per_subslice;
+
+	topo.subslice_offset = slice_length;
+	topo.subslice_stride = ss_stride;
+	topo.eu_offset = slice_length + subslice_length;
+	topo.eu_stride = eu_stride;
+
+	if (copy_to_user(u64_to_user_ptr(query_item->data_ptr),
+			 &topo, sizeof(topo)))
+		return -EFAULT;
+
+	if (copy_to_user(u64_to_user_ptr(query_item->data_ptr + sizeof(topo)),
+			 &sseu->slice_mask, slice_length))
+		return -EFAULT;
+
+	if (intel_sseu_copy_ssmask_to_user(u64_to_user_ptr(query_item->data_ptr +
+							   sizeof(topo) + slice_length),
+					   sseu))
+		return -EFAULT;
+
+	if (intel_sseu_copy_eumask_to_user(u64_to_user_ptr(query_item->data_ptr +
+							   sizeof(topo) +
+							   slice_length + subslice_length),
+					   sseu))
+		return -EFAULT;
+
+	return total_length;
+}
+
+static int query_topology_info(struct drm_i915_private *dev_priv,
+			       struct drm_i915_query_item *query_item)
+{
+	const struct sseu_dev_info *sseu = &to_gt(dev_priv)->info.sseu;
+
+	if (query_item->flags != 0)
+		return -EINVAL;
+
+	return fill_topology_info(sseu, query_item, sseu->subslice_mask);
+}
+
+static int query_geometry_subslices(struct drm_i915_private *i915,
+				    struct drm_i915_query_item *query_item)
+{
+	const struct sseu_dev_info *sseu;
+	struct intel_engine_cs *engine;
+	struct i915_engine_class_instance classinstance;
+
+	if (GRAPHICS_VER_FULL(i915) < IP_VER(12, 50))
+		return -ENODEV;
+
+	classinstance = *((struct i915_engine_class_instance *)&query_item->flags);
+
+	engine = intel_engine_lookup_user(i915, (u8)classinstance.engine_class,
+					  (u8)classinstance.engine_instance);
+
+	if (!engine)
+		return -EINVAL;
+
+	if (engine->class != RENDER_CLASS)
+		return -EINVAL;
+
+	sseu = &engine->gt->info.sseu;
+
+	return fill_topology_info(sseu, query_item, sseu->geometry_subslice_mask);
+}
+
+static int
+query_engine_info(struct drm_i915_private *i915,
+		  struct drm_i915_query_item *query_item)
+{
+	struct drm_i915_query_engine_info __user *query_ptr =
+				u64_to_user_ptr(query_item->data_ptr);
+	struct drm_i915_engine_info __user *info_ptr;
+	struct drm_i915_query_engine_info query;
+	struct drm_i915_engine_info info = { };
+	unsigned int num_uabi_engines = 0;
+	struct intel_engine_cs *engine;
+	int len, ret;
+
+	if (query_item->flags)
+		return -EINVAL;
+
+	for_each_uabi_engine(engine, i915)
+		num_uabi_engines++;
+
+	len = struct_size(query_ptr, engines, num_uabi_engines);
+
+	ret = copy_query_item(&query, sizeof(query), len, query_item);
+	if (ret != 0)
+		return ret;
+
+	if (query.num_engines || query.rsvd[0] || query.rsvd[1] ||
+	    query.rsvd[2])
+		return -EINVAL;
+
+	info_ptr = &query_ptr->engines[0];
+
+	for_each_uabi_engine(engine, i915) {
+		info.engine.engine_class = engine->uabi_class;
+		info.engine.engine_instance = engine->uabi_instance;
+		info.flags = I915_ENGINE_INFO_HAS_LOGICAL_INSTANCE;
+		info.capabilities = engine->uabi_capabilities;
+		info.logical_instance = ilog2(engine->logical_mask);
+
+		if (copy_to_user(info_ptr, &info, sizeof(info)))
+			return -EFAULT;
+
+		query.num_engines++;
+		info_ptr++;
+	}
+
+	if (copy_to_user(query_ptr, &query, sizeof(query)))
+		return -EFAULT;
+
+	return len;
+}
+
+static int can_copy_perf_config_registers_or_number(u32 user_n_regs,
+						    u64 user_regs_ptr,
+						    u32 kernel_n_regs)
+{
+	/*
+	 * We'll just put the number of registers, and won't copy the
+	 * register.
+	 */
+	if (user_n_regs == 0)
+		return 0;
+
+	if (user_n_regs < kernel_n_regs)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int copy_perf_config_registers_or_number(const struct i915_oa_reg *kernel_regs,
+						u32 kernel_n_regs,
+						u64 user_regs_ptr,
+						u32 *user_n_regs)
+{
+	u32 __user *p = u64_to_user_ptr(user_regs_ptr);
+	u32 r;
+
+	if (*user_n_regs == 0) {
+		*user_n_regs = kernel_n_regs;
+		return 0;
+	}
+
+	*user_n_regs = kernel_n_regs;
+
+	if (!user_write_access_begin(p, 2 * sizeof(u32) * kernel_n_regs))
+		return -EFAULT;
+
+	for (r = 0; r < kernel_n_regs; r++, p += 2) {
+		unsafe_put_user(i915_mmio_reg_offset(kernel_regs[r].addr),
+				p, Efault);
+		unsafe_put_user(kernel_regs[r].value, p + 1, Efault);
+	}
+	user_write_access_end();
+	return 0;
+Efault:
+	user_write_access_end();
+	return -EFAULT;
+}
+
+static int query_perf_config_data(struct drm_i915_private *i915,
+				  struct drm_i915_query_item *query_item,
+				  bool use_uuid)
+{
+	struct drm_i915_query_perf_config __user *user_query_config_ptr =
+		u64_to_user_ptr(query_item->data_ptr);
+	struct drm_i915_perf_oa_config __user *user_config_ptr =
+		u64_to_user_ptr(query_item->data_ptr +
+				sizeof(struct drm_i915_query_perf_config));
+	struct drm_i915_perf_oa_config user_config;
+	struct i915_perf *perf = &i915->perf;
+	struct i915_oa_config *oa_config;
+	char uuid[UUID_STRING_LEN + 1];
+	u64 config_id;
+	u32 flags, total_size;
+	int ret;
+
+	if (!perf->i915)
+		return -ENODEV;
+
+	total_size =
+		sizeof(struct drm_i915_query_perf_config) +
+		sizeof(struct drm_i915_perf_oa_config);
+
+	if (query_item->length == 0)
+		return total_size;
+
+	if (query_item->length < total_size) {
+		DRM_DEBUG("Invalid query config data item size=%u expected=%u\n",
+			  query_item->length, total_size);
+		return -EINVAL;
+	}
+
+	if (get_user(flags, &user_query_config_ptr->flags))
+		return -EFAULT;
+
+	if (flags != 0)
+		return -EINVAL;
+
+	if (use_uuid) {
+		struct i915_oa_config *tmp;
+		int id;
+
+		BUILD_BUG_ON(sizeof(user_query_config_ptr->uuid) >= sizeof(uuid));
+
+		memset(&uuid, 0, sizeof(uuid));
+		if (copy_from_user(uuid, user_query_config_ptr->uuid,
+				     sizeof(user_query_config_ptr->uuid)))
+			return -EFAULT;
+
+		oa_config = NULL;
+		rcu_read_lock();
+		idr_for_each_entry(&perf->metrics_idr, tmp, id) {
+			if (!strcmp(tmp->uuid, uuid)) {
+				oa_config = i915_oa_config_get(tmp);
+				break;
+			}
+		}
+		rcu_read_unlock();
+	} else {
+		if (get_user(config_id, &user_query_config_ptr->config))
+			return -EFAULT;
+
+		oa_config = i915_perf_get_oa_config(perf, config_id);
+	}
+	if (!oa_config)
+		return -ENOENT;
+
+	if (copy_from_user(&user_config, user_config_ptr, sizeof(user_config))) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	ret = can_copy_perf_config_registers_or_number(user_config.n_boolean_regs,
+						       user_config.boolean_regs_ptr,
+						       oa_config->b_counter_regs_len);
+	if (ret)
+		goto out;
+
+	ret = can_copy_perf_config_registers_or_number(user_config.n_flex_regs,
+						       user_config.flex_regs_ptr,
+						       oa_config->flex_regs_len);
+	if (ret)
+		goto out;
+
+	ret = can_copy_perf_config_registers_or_number(user_config.n_mux_regs,
+						       user_config.mux_regs_ptr,
+						       oa_config->mux_regs_len);
+	if (ret)
+		goto out;
+
+	ret = copy_perf_config_registers_or_number(oa_config->b_counter_regs,
+						   oa_config->b_counter_regs_len,
+						   user_config.boolean_regs_ptr,
+						   &user_config.n_boolean_regs);
+	if (ret)
+		goto out;
+
+	ret = copy_perf_config_registers_or_number(oa_config->flex_regs,
+						   oa_config->flex_regs_len,
+						   user_config.flex_regs_ptr,
+						   &user_config.n_flex_regs);
+	if (ret)
+		goto out;
+
+	ret = copy_perf_config_registers_or_number(oa_config->mux_regs,
+						   oa_config->mux_regs_len,
+						   user_config.mux_regs_ptr,
+						   &user_config.n_mux_regs);
+	if (ret)
+		goto out;
+
+	memcpy(user_config.uuid, oa_config->uuid, sizeof(user_config.uuid));
+
+	if (copy_to_user(user_config_ptr, &user_config, sizeof(user_config))) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	ret = total_size;
+
+out:
+	i915_oa_config_put(oa_config);
+	return ret;
+}
+
+static size_t sizeof_perf_config_list(size_t count)
+{
+	return sizeof(struct drm_i915_query_perf_config) + sizeof(u64) * count;
+}
+
+static size_t sizeof_perf_metrics(struct i915_perf *perf)
+{
+	struct i915_oa_config *tmp;
+	size_t i;
+	int id;
+
+	i = 1;
+	rcu_read_lock();
+	idr_for_each_entry(&perf->metrics_idr, tmp, id)
+		i++;
+	rcu_read_unlock();
+
+	return sizeof_perf_config_list(i);
+}
+
+static int query_perf_config_list(struct drm_i915_private *i915,
+				  struct drm_i915_query_item *query_item)
+{
+	struct drm_i915_query_perf_config __user *user_query_config_ptr =
+		u64_to_user_ptr(query_item->data_ptr);
+	struct i915_perf *perf = &i915->perf;
+	u64 *oa_config_ids = NULL;
+	int alloc, n_configs;
+	u32 flags;
+	int ret;
+
+	if (!perf->i915)
+		return -ENODEV;
+
+	if (query_item->length == 0)
+		return sizeof_perf_metrics(perf);
+
+	if (get_user(flags, &user_query_config_ptr->flags))
+		return -EFAULT;
+
+	if (flags != 0)
+		return -EINVAL;
+
+	n_configs = 1;
+	do {
+		struct i915_oa_config *tmp;
+		u64 *ids;
+		int id;
+
+		ids = krealloc(oa_config_ids,
+			       n_configs * sizeof(*oa_config_ids),
+			       GFP_KERNEL);
+		if (!ids)
+			return -ENOMEM;
+
+		alloc = fetch_and_zero(&n_configs);
+
+		ids[n_configs++] = 1ull; /* reserved for test_config */
+		rcu_read_lock();
+		idr_for_each_entry(&perf->metrics_idr, tmp, id) {
+			if (n_configs < alloc)
+				ids[n_configs] = id;
+			n_configs++;
+		}
+		rcu_read_unlock();
+
+		oa_config_ids = ids;
+	} while (n_configs > alloc);
+
+	if (query_item->length < sizeof_perf_config_list(n_configs)) {
+		DRM_DEBUG("Invalid query config list item size=%u expected=%zu\n",
+			  query_item->length,
+			  sizeof_perf_config_list(n_configs));
+		kfree(oa_config_ids);
+		return -EINVAL;
+	}
+
+	if (put_user(n_configs, &user_query_config_ptr->config)) {
+		kfree(oa_config_ids);
+		return -EFAULT;
+	}
+
+	ret = copy_to_user(user_query_config_ptr + 1,
+			   oa_config_ids,
+			   n_configs * sizeof(*oa_config_ids));
+	kfree(oa_config_ids);
+	if (ret)
+		return -EFAULT;
+
+	return sizeof_perf_config_list(n_configs);
+}
+
+static int query_perf_config(struct drm_i915_private *i915,
+			     struct drm_i915_query_item *query_item)
+{
+	switch (query_item->flags) {
+	case DRM_I915_QUERY_PERF_CONFIG_LIST:
+		return query_perf_config_list(i915, query_item);
+	case DRM_I915_QUERY_PERF_CONFIG_DATA_FOR_UUID:
+		return query_perf_config_data(i915, query_item, true);
+	case DRM_I915_QUERY_PERF_CONFIG_DATA_FOR_ID:
+		return query_perf_config_data(i915, query_item, false);
+	default:
+		return -EINVAL;
+	}
+}
+
+static int query_memregion_info(struct drm_i915_private *i915,
+				struct drm_i915_query_item *query_item)
+{
+	struct drm_i915_query_memory_regions __user *query_ptr =
+		u64_to_user_ptr(query_item->data_ptr);
+	struct drm_i915_memory_region_info __user *info_ptr =
+		&query_ptr->regions[0];
+	struct drm_i915_memory_region_info info = { };
+	struct drm_i915_query_memory_regions query;
+	struct intel_memory_region *mr;
+	u32 total_length;
+	int ret, id, i;
+
+	if (query_item->flags != 0)
+		return -EINVAL;
+
+	total_length = sizeof(query);
+	for_each_memory_region(mr, i915, id) {
+		if (mr->private)
+			continue;
+
+		total_length += sizeof(info);
+	}
+
+	ret = copy_query_item(&query, sizeof(query), total_length, query_item);
+	if (ret != 0)
+		return ret;
+
+	if (query.num_regions)
+		return -EINVAL;
+
+	for (i = 0; i < ARRAY_SIZE(query.rsvd); i++) {
+		if (query.rsvd[i])
+			return -EINVAL;
+	}
+
+	for_each_memory_region(mr, i915, id) {
+		if (mr->private)
+			continue;
+
+		info.region.memory_class = mr->type;
+		info.region.memory_instance = mr->instance;
+		info.probed_size = mr->total;
+
+		if (mr->type == INTEL_MEMORY_LOCAL)
+			info.probed_cpu_visible_size = mr->io_size;
+		else
+			info.probed_cpu_visible_size = mr->total;
+
+		if (perfmon_capable()) {
+			intel_memory_region_avail(mr,
+						  &info.unallocated_size,
+						  &info.unallocated_cpu_visible_size);
+		} else {
+			info.unallocated_size = info.probed_size;
+			info.unallocated_cpu_visible_size =
+				info.probed_cpu_visible_size;
+		}
+
+		if (__copy_to_user(info_ptr, &info, sizeof(info)))
+			return -EFAULT;
+
+		query.num_regions++;
+		info_ptr++;
+	}
+
+	if (__copy_to_user(query_ptr, &query, sizeof(query)))
+		return -EFAULT;
+
+	return total_length;
+}
+
+static int query_hwconfig_blob(struct drm_i915_private *i915,
+			       struct drm_i915_query_item *query_item)
+{
+	struct intel_gt *gt = to_gt(i915);
+	struct intel_hwconfig *hwconfig = &gt->info.hwconfig;
+
+	if (!hwconfig->size || !hwconfig->ptr)
+		return -ENODEV;
+
+	if (query_item->length == 0)
+		return hwconfig->size;
+
+	if (query_item->length < hwconfig->size)
+		return -EINVAL;
+
+	if (copy_to_user(u64_to_user_ptr(query_item->data_ptr),
+			 hwconfig->ptr, hwconfig->size))
+		return -EFAULT;
+
+	return hwconfig->size;
+}
+
+static int (* const i915_query_funcs[])(struct drm_i915_private *dev_priv,
+					struct drm_i915_query_item *query_item) = {
+	query_topology_info,
+	query_engine_info,
+	query_perf_config,
+	query_memregion_info,
+	query_hwconfig_blob,
+	query_geometry_subslices,
+};
+
+int i915_query_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_i915_query *args = data;
+	struct drm_i915_query_item __user *user_item_ptr =
+		u64_to_user_ptr(args->items_ptr);
+	u32 i;
+
+	if (args->flags != 0)
+		return -EINVAL;
+
+	for (i = 0; i < args->num_items; i++, user_item_ptr++) {
+		struct drm_i915_query_item item;
+		unsigned long func_idx;
+		int ret;
+
+		if (copy_from_user(&item, user_item_ptr, sizeof(item)))
+			return -EFAULT;
+
+		if (item.query_id == 0)
+			return -EINVAL;
+
+		if (overflows_type(item.query_id - 1, unsigned long))
+			return -EINVAL;
+
+		func_idx = item.query_id - 1;
+
+		ret = -EINVAL;
+		if (func_idx < ARRAY_SIZE(i915_query_funcs)) {
+			func_idx = array_index_nospec(func_idx,
+						      ARRAY_SIZE(i915_query_funcs));
+			ret = i915_query_funcs[func_idx](dev_priv, &item);
+		}
+
+		/* Only write the length back to userspace if they differ. */
+		if (ret != item.length && put_user(ret, &user_item_ptr->length))
+			return -EFAULT;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_query.h b/drivers/gpu/drm/i915/i915_query.h
new file mode 100644
index 000000000..31dcef181
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_query.h
@@ -0,0 +1,15 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef _I915_QUERY_H_
+#define _I915_QUERY_H_
+
+struct drm_device;
+struct drm_file;
+
+int i915_query_ioctl(struct drm_device *dev, void *data, struct drm_file *file);
+
+#endif
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
new file mode 100644
index 000000000..c67667043
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -0,0 +1,8370 @@
+/* Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef _I915_REG_H_
+#define _I915_REG_H_
+
+#include "i915_reg_defs.h"
+
+/**
+ * DOC: The i915 register macro definition style guide
+ *
+ * Follow the style described here for new macros, and while changing existing
+ * macros. Do **not** mass change existing definitions just to update the style.
+ *
+ * File Layout
+ * ~~~~~~~~~~~
+ *
+ * Keep helper macros near the top. For example, _PIPE() and friends.
+ *
+ * Prefix macros that generally should not be used outside of this file with
+ * underscore '_'. For example, _PIPE() and friends, single instances of
+ * registers that are defined solely for the use by function-like macros.
+ *
+ * Avoid using the underscore prefixed macros outside of this file. There are
+ * exceptions, but keep them to a minimum.
+ *
+ * There are two basic types of register definitions: Single registers and
+ * register groups. Register groups are registers which have two or more
+ * instances, for example one per pipe, port, transcoder, etc. Register groups
+ * should be defined using function-like macros.
+ *
+ * For single registers, define the register offset first, followed by register
+ * contents.
+ *
+ * For register groups, define the register instance offsets first, prefixed
+ * with underscore, followed by a function-like macro choosing the right
+ * instance based on the parameter, followed by register contents.
+ *
+ * Define the register contents (i.e. bit and bit field macros) from most
+ * significant to least significant bit. Indent the register content macros
+ * using two extra spaces between ``#define`` and the macro name.
+ *
+ * Define bit fields using ``REG_GENMASK(h, l)``. Define bit field contents
+ * using ``REG_FIELD_PREP(mask, value)``. This will define the values already
+ * shifted in place, so they can be directly OR'd together. For convenience,
+ * function-like macros may be used to define bit fields, but do note that the
+ * macros may be needed to read as well as write the register contents.
+ *
+ * Define bits using ``REG_BIT(N)``. Do **not** add ``_BIT`` suffix to the name.
+ *
+ * Group the register and its contents together without blank lines, separate
+ * from other registers and their contents with one blank line.
+ *
+ * Indent macro values from macro names using TABs. Align values vertically. Use
+ * braces in macro values as needed to avoid unintended precedence after macro
+ * substitution. Use spaces in macro values according to kernel coding
+ * style. Use lower case in hexadecimal values.
+ *
+ * Naming
+ * ~~~~~~
+ *
+ * Try to name registers according to the specs. If the register name changes in
+ * the specs from platform to another, stick to the original name.
+ *
+ * Try to re-use existing register macro definitions. Only add new macros for
+ * new register offsets, or when the register contents have changed enough to
+ * warrant a full redefinition.
+ *
+ * When a register macro changes for a new platform, prefix the new macro using
+ * the platform acronym or generation. For example, ``SKL_`` or ``GEN8_``. The
+ * prefix signifies the start platform/generation using the register.
+ *
+ * When a bit (field) macro changes or gets added for a new platform, while
+ * retaining the existing register macro, add a platform acronym or generation
+ * suffix to the name. For example, ``_SKL`` or ``_GEN8``.
+ *
+ * Examples
+ * ~~~~~~~~
+ *
+ * (Note that the values in the example are indented using spaces instead of
+ * TABs to avoid misalignment in generated documentation. Use TABs in the
+ * definitions.)::
+ *
+ *  #define _FOO_A                      0xf000
+ *  #define _FOO_B                      0xf001
+ *  #define FOO(pipe)                   _MMIO_PIPE(pipe, _FOO_A, _FOO_B)
+ *  #define   FOO_ENABLE                REG_BIT(31)
+ *  #define   FOO_MODE_MASK             REG_GENMASK(19, 16)
+ *  #define   FOO_MODE_BAR              REG_FIELD_PREP(FOO_MODE_MASK, 0)
+ *  #define   FOO_MODE_BAZ              REG_FIELD_PREP(FOO_MODE_MASK, 1)
+ *  #define   FOO_MODE_QUX_SNB          REG_FIELD_PREP(FOO_MODE_MASK, 2)
+ *
+ *  #define BAR                         _MMIO(0xb000)
+ *  #define GEN8_BAR                    _MMIO(0xb888)
+ */
+
+#define DISPLAY_MMIO_BASE(dev_priv)	(INTEL_INFO(dev_priv)->display.mmio_offset)
+
+/*
+ * Given the first two numbers __a and __b of arbitrarily many evenly spaced
+ * numbers, pick the 0-based __index'th value.
+ *
+ * Always prefer this over _PICK() if the numbers are evenly spaced.
+ */
+#define _PICK_EVEN(__index, __a, __b) ((__a) + (__index) * ((__b) - (__a)))
+
+/*
+ * Given the arbitrary numbers in varargs, pick the 0-based __index'th number.
+ *
+ * Always prefer _PICK_EVEN() over this if the numbers are evenly spaced.
+ */
+#define _PICK(__index, ...) (((const u32 []){ __VA_ARGS__ })[__index])
+
+/*
+ * Named helper wrappers around _PICK_EVEN() and _PICK().
+ */
+#define _PIPE(pipe, a, b)		_PICK_EVEN(pipe, a, b)
+#define _PLANE(plane, a, b)		_PICK_EVEN(plane, a, b)
+#define _TRANS(tran, a, b)		_PICK_EVEN(tran, a, b)
+#define _PORT(port, a, b)		_PICK_EVEN(port, a, b)
+#define _PLL(pll, a, b)			_PICK_EVEN(pll, a, b)
+#define _PHY(phy, a, b)			_PICK_EVEN(phy, a, b)
+
+#define _MMIO_PIPE(pipe, a, b)		_MMIO(_PIPE(pipe, a, b))
+#define _MMIO_PLANE(plane, a, b)	_MMIO(_PLANE(plane, a, b))
+#define _MMIO_TRANS(tran, a, b)		_MMIO(_TRANS(tran, a, b))
+#define _MMIO_PORT(port, a, b)		_MMIO(_PORT(port, a, b))
+#define _MMIO_PLL(pll, a, b)		_MMIO(_PLL(pll, a, b))
+#define _MMIO_PHY(phy, a, b)		_MMIO(_PHY(phy, a, b))
+
+#define _PHY3(phy, ...)			_PICK(phy, __VA_ARGS__)
+
+#define _MMIO_PIPE3(pipe, a, b, c)	_MMIO(_PICK(pipe, a, b, c))
+#define _MMIO_PORT3(pipe, a, b, c)	_MMIO(_PICK(pipe, a, b, c))
+#define _MMIO_PHY3(phy, a, b, c)	_MMIO(_PHY3(phy, a, b, c))
+#define _MMIO_PLL3(pll, ...)		_MMIO(_PICK(pll, __VA_ARGS__))
+
+
+/*
+ * Device info offset array based helpers for groups of registers with unevenly
+ * spaced base offsets.
+ */
+#define _MMIO_PIPE2(pipe, reg)		_MMIO(INTEL_INFO(dev_priv)->display.pipe_offsets[(pipe)] - \
+					      INTEL_INFO(dev_priv)->display.pipe_offsets[PIPE_A] + \
+					      DISPLAY_MMIO_BASE(dev_priv) + (reg))
+#define _MMIO_TRANS2(tran, reg)		_MMIO(INTEL_INFO(dev_priv)->display.trans_offsets[(tran)] - \
+					      INTEL_INFO(dev_priv)->display.trans_offsets[TRANSCODER_A] + \
+					      DISPLAY_MMIO_BASE(dev_priv) + (reg))
+#define _MMIO_CURSOR2(pipe, reg)	_MMIO(INTEL_INFO(dev_priv)->display.cursor_offsets[(pipe)] - \
+					      INTEL_INFO(dev_priv)->display.cursor_offsets[PIPE_A] + \
+					      DISPLAY_MMIO_BASE(dev_priv) + (reg))
+
+#define __MASKED_FIELD(mask, value) ((mask) << 16 | (value))
+#define _MASKED_FIELD(mask, value) ({					   \
+	if (__builtin_constant_p(mask))					   \
+		BUILD_BUG_ON_MSG(((mask) & 0xffff0000), "Incorrect mask"); \
+	if (__builtin_constant_p(value))				   \
+		BUILD_BUG_ON_MSG((value) & 0xffff0000, "Incorrect value"); \
+	if (__builtin_constant_p(mask) && __builtin_constant_p(value))	   \
+		BUILD_BUG_ON_MSG((value) & ~(mask),			   \
+				 "Incorrect value for mask");		   \
+	__MASKED_FIELD(mask, value); })
+#define _MASKED_BIT_ENABLE(a)	({ typeof(a) _a = (a); _MASKED_FIELD(_a, _a); })
+#define _MASKED_BIT_DISABLE(a)	(_MASKED_FIELD((a), 0))
+
+#define GU_CNTL				_MMIO(0x101010)
+#define   LMEM_INIT			REG_BIT(7)
+
+#define GEN6_STOLEN_RESERVED		_MMIO(0x1082C0)
+#define GEN6_STOLEN_RESERVED_ADDR_MASK	(0xFFF << 20)
+#define GEN7_STOLEN_RESERVED_ADDR_MASK	(0x3FFF << 18)
+#define GEN6_STOLEN_RESERVED_SIZE_MASK	(3 << 4)
+#define GEN6_STOLEN_RESERVED_1M		(0 << 4)
+#define GEN6_STOLEN_RESERVED_512K	(1 << 4)
+#define GEN6_STOLEN_RESERVED_256K	(2 << 4)
+#define GEN6_STOLEN_RESERVED_128K	(3 << 4)
+#define GEN7_STOLEN_RESERVED_SIZE_MASK	(1 << 5)
+#define GEN7_STOLEN_RESERVED_1M		(0 << 5)
+#define GEN7_STOLEN_RESERVED_256K	(1 << 5)
+#define GEN8_STOLEN_RESERVED_SIZE_MASK	(3 << 7)
+#define GEN8_STOLEN_RESERVED_1M		(0 << 7)
+#define GEN8_STOLEN_RESERVED_2M		(1 << 7)
+#define GEN8_STOLEN_RESERVED_4M		(2 << 7)
+#define GEN8_STOLEN_RESERVED_8M		(3 << 7)
+#define GEN6_STOLEN_RESERVED_ENABLE	(1 << 0)
+#define GEN11_STOLEN_RESERVED_ADDR_MASK	(0xFFFFFFFFFFFULL << 20)
+
+#define _VGA_MSR_WRITE _MMIO(0x3c2)
+
+#define _GEN7_PIPEA_DE_LOAD_SL	0x70068
+#define _GEN7_PIPEB_DE_LOAD_SL	0x71068
+#define GEN7_PIPE_DE_LOAD_SL(pipe) _MMIO_PIPE(pipe, _GEN7_PIPEA_DE_LOAD_SL, _GEN7_PIPEB_DE_LOAD_SL)
+
+/*
+ * Reset registers
+ */
+#define DEBUG_RESET_I830		_MMIO(0x6070)
+#define  DEBUG_RESET_FULL		(1 << 7)
+#define  DEBUG_RESET_RENDER		(1 << 8)
+#define  DEBUG_RESET_DISPLAY		(1 << 9)
+
+/*
+ * IOSF sideband
+ */
+#define VLV_IOSF_DOORBELL_REQ			_MMIO(VLV_DISPLAY_BASE + 0x2100)
+#define   IOSF_DEVFN_SHIFT			24
+#define   IOSF_OPCODE_SHIFT			16
+#define   IOSF_PORT_SHIFT			8
+#define   IOSF_BYTE_ENABLES_SHIFT		4
+#define   IOSF_BAR_SHIFT			1
+#define   IOSF_SB_BUSY				(1 << 0)
+#define   IOSF_PORT_BUNIT			0x03
+#define   IOSF_PORT_PUNIT			0x04
+#define   IOSF_PORT_NC				0x11
+#define   IOSF_PORT_DPIO			0x12
+#define   IOSF_PORT_GPIO_NC			0x13
+#define   IOSF_PORT_CCK				0x14
+#define   IOSF_PORT_DPIO_2			0x1a
+#define   IOSF_PORT_FLISDSI			0x1b
+#define   IOSF_PORT_GPIO_SC			0x48
+#define   IOSF_PORT_GPIO_SUS			0xa8
+#define   IOSF_PORT_CCU				0xa9
+#define   CHV_IOSF_PORT_GPIO_N			0x13
+#define   CHV_IOSF_PORT_GPIO_SE			0x48
+#define   CHV_IOSF_PORT_GPIO_E			0xa8
+#define   CHV_IOSF_PORT_GPIO_SW			0xb2
+#define VLV_IOSF_DATA				_MMIO(VLV_DISPLAY_BASE + 0x2104)
+#define VLV_IOSF_ADDR				_MMIO(VLV_DISPLAY_BASE + 0x2108)
+
+/* DPIO registers */
+#define DPIO_DEVFN			0
+
+#define DPIO_CTL			_MMIO(VLV_DISPLAY_BASE + 0x2110)
+#define  DPIO_MODSEL1			(1 << 3) /* if ref clk b == 27 */
+#define  DPIO_MODSEL0			(1 << 2) /* if ref clk a == 27 */
+#define  DPIO_SFR_BYPASS		(1 << 1)
+#define  DPIO_CMNRST			(1 << 0)
+
+#define DPIO_PHY(pipe)			((pipe) >> 1)
+
+/*
+ * Per pipe/PLL DPIO regs
+ */
+#define _VLV_PLL_DW3_CH0		0x800c
+#define   DPIO_POST_DIV_SHIFT		(28) /* 3 bits */
+#define   DPIO_POST_DIV_DAC		0
+#define   DPIO_POST_DIV_HDMIDP		1 /* DAC 225-400M rate */
+#define   DPIO_POST_DIV_LVDS1		2
+#define   DPIO_POST_DIV_LVDS2		3
+#define   DPIO_K_SHIFT			(24) /* 4 bits */
+#define   DPIO_P1_SHIFT			(21) /* 3 bits */
+#define   DPIO_P2_SHIFT			(16) /* 5 bits */
+#define   DPIO_N_SHIFT			(12) /* 4 bits */
+#define   DPIO_ENABLE_CALIBRATION	(1 << 11)
+#define   DPIO_M1DIV_SHIFT		(8) /* 3 bits */
+#define   DPIO_M2DIV_MASK		0xff
+#define _VLV_PLL_DW3_CH1		0x802c
+#define VLV_PLL_DW3(ch) _PIPE(ch, _VLV_PLL_DW3_CH0, _VLV_PLL_DW3_CH1)
+
+#define _VLV_PLL_DW5_CH0		0x8014
+#define   DPIO_REFSEL_OVERRIDE		27
+#define   DPIO_PLL_MODESEL_SHIFT	24 /* 3 bits */
+#define   DPIO_BIAS_CURRENT_CTL_SHIFT	21 /* 3 bits, always 0x7 */
+#define   DPIO_PLL_REFCLK_SEL_SHIFT	16 /* 2 bits */
+#define   DPIO_PLL_REFCLK_SEL_MASK	3
+#define   DPIO_DRIVER_CTL_SHIFT		12 /* always set to 0x8 */
+#define   DPIO_CLK_BIAS_CTL_SHIFT	8 /* always set to 0x5 */
+#define _VLV_PLL_DW5_CH1		0x8034
+#define VLV_PLL_DW5(ch) _PIPE(ch, _VLV_PLL_DW5_CH0, _VLV_PLL_DW5_CH1)
+
+#define _VLV_PLL_DW7_CH0		0x801c
+#define _VLV_PLL_DW7_CH1		0x803c
+#define VLV_PLL_DW7(ch) _PIPE(ch, _VLV_PLL_DW7_CH0, _VLV_PLL_DW7_CH1)
+
+#define _VLV_PLL_DW8_CH0		0x8040
+#define _VLV_PLL_DW8_CH1		0x8060
+#define VLV_PLL_DW8(ch) _PIPE(ch, _VLV_PLL_DW8_CH0, _VLV_PLL_DW8_CH1)
+
+#define VLV_PLL_DW9_BCAST		0xc044
+#define _VLV_PLL_DW9_CH0		0x8044
+#define _VLV_PLL_DW9_CH1		0x8064
+#define VLV_PLL_DW9(ch) _PIPE(ch, _VLV_PLL_DW9_CH0, _VLV_PLL_DW9_CH1)
+
+#define _VLV_PLL_DW10_CH0		0x8048
+#define _VLV_PLL_DW10_CH1		0x8068
+#define VLV_PLL_DW10(ch) _PIPE(ch, _VLV_PLL_DW10_CH0, _VLV_PLL_DW10_CH1)
+
+#define _VLV_PLL_DW11_CH0		0x804c
+#define _VLV_PLL_DW11_CH1		0x806c
+#define VLV_PLL_DW11(ch) _PIPE(ch, _VLV_PLL_DW11_CH0, _VLV_PLL_DW11_CH1)
+
+/* Spec for ref block start counts at DW10 */
+#define VLV_REF_DW13			0x80ac
+
+#define VLV_CMN_DW0			0x8100
+
+/*
+ * Per DDI channel DPIO regs
+ */
+
+#define _VLV_PCS_DW0_CH0		0x8200
+#define _VLV_PCS_DW0_CH1		0x8400
+#define   DPIO_PCS_TX_LANE2_RESET	(1 << 16)
+#define   DPIO_PCS_TX_LANE1_RESET	(1 << 7)
+#define   DPIO_LEFT_TXFIFO_RST_MASTER2	(1 << 4)
+#define   DPIO_RIGHT_TXFIFO_RST_MASTER2	(1 << 3)
+#define VLV_PCS_DW0(ch) _PORT(ch, _VLV_PCS_DW0_CH0, _VLV_PCS_DW0_CH1)
+
+#define _VLV_PCS01_DW0_CH0		0x200
+#define _VLV_PCS23_DW0_CH0		0x400
+#define _VLV_PCS01_DW0_CH1		0x2600
+#define _VLV_PCS23_DW0_CH1		0x2800
+#define VLV_PCS01_DW0(ch) _PORT(ch, _VLV_PCS01_DW0_CH0, _VLV_PCS01_DW0_CH1)
+#define VLV_PCS23_DW0(ch) _PORT(ch, _VLV_PCS23_DW0_CH0, _VLV_PCS23_DW0_CH1)
+
+#define _VLV_PCS_DW1_CH0		0x8204
+#define _VLV_PCS_DW1_CH1		0x8404
+#define   CHV_PCS_REQ_SOFTRESET_EN	(1 << 23)
+#define   DPIO_PCS_CLK_CRI_RXEB_EIOS_EN	(1 << 22)
+#define   DPIO_PCS_CLK_CRI_RXDIGFILTSG_EN (1 << 21)
+#define   DPIO_PCS_CLK_DATAWIDTH_SHIFT	(6)
+#define   DPIO_PCS_CLK_SOFT_RESET	(1 << 5)
+#define VLV_PCS_DW1(ch) _PORT(ch, _VLV_PCS_DW1_CH0, _VLV_PCS_DW1_CH1)
+
+#define _VLV_PCS01_DW1_CH0		0x204
+#define _VLV_PCS23_DW1_CH0		0x404
+#define _VLV_PCS01_DW1_CH1		0x2604
+#define _VLV_PCS23_DW1_CH1		0x2804
+#define VLV_PCS01_DW1(ch) _PORT(ch, _VLV_PCS01_DW1_CH0, _VLV_PCS01_DW1_CH1)
+#define VLV_PCS23_DW1(ch) _PORT(ch, _VLV_PCS23_DW1_CH0, _VLV_PCS23_DW1_CH1)
+
+#define _VLV_PCS_DW8_CH0		0x8220
+#define _VLV_PCS_DW8_CH1		0x8420
+#define   CHV_PCS_USEDCLKCHANNEL_OVRRIDE	(1 << 20)
+#define   CHV_PCS_USEDCLKCHANNEL		(1 << 21)
+#define VLV_PCS_DW8(ch) _PORT(ch, _VLV_PCS_DW8_CH0, _VLV_PCS_DW8_CH1)
+
+#define _VLV_PCS01_DW8_CH0		0x0220
+#define _VLV_PCS23_DW8_CH0		0x0420
+#define _VLV_PCS01_DW8_CH1		0x2620
+#define _VLV_PCS23_DW8_CH1		0x2820
+#define VLV_PCS01_DW8(port) _PORT(port, _VLV_PCS01_DW8_CH0, _VLV_PCS01_DW8_CH1)
+#define VLV_PCS23_DW8(port) _PORT(port, _VLV_PCS23_DW8_CH0, _VLV_PCS23_DW8_CH1)
+
+#define _VLV_PCS_DW9_CH0		0x8224
+#define _VLV_PCS_DW9_CH1		0x8424
+#define   DPIO_PCS_TX2MARGIN_MASK	(0x7 << 13)
+#define   DPIO_PCS_TX2MARGIN_000	(0 << 13)
+#define   DPIO_PCS_TX2MARGIN_101	(1 << 13)
+#define   DPIO_PCS_TX1MARGIN_MASK	(0x7 << 10)
+#define   DPIO_PCS_TX1MARGIN_000	(0 << 10)
+#define   DPIO_PCS_TX1MARGIN_101	(1 << 10)
+#define	VLV_PCS_DW9(ch) _PORT(ch, _VLV_PCS_DW9_CH0, _VLV_PCS_DW9_CH1)
+
+#define _VLV_PCS01_DW9_CH0		0x224
+#define _VLV_PCS23_DW9_CH0		0x424
+#define _VLV_PCS01_DW9_CH1		0x2624
+#define _VLV_PCS23_DW9_CH1		0x2824
+#define VLV_PCS01_DW9(ch) _PORT(ch, _VLV_PCS01_DW9_CH0, _VLV_PCS01_DW9_CH1)
+#define VLV_PCS23_DW9(ch) _PORT(ch, _VLV_PCS23_DW9_CH0, _VLV_PCS23_DW9_CH1)
+
+#define _CHV_PCS_DW10_CH0		0x8228
+#define _CHV_PCS_DW10_CH1		0x8428
+#define   DPIO_PCS_SWING_CALC_TX0_TX2	(1 << 30)
+#define   DPIO_PCS_SWING_CALC_TX1_TX3	(1 << 31)
+#define   DPIO_PCS_TX2DEEMP_MASK	(0xf << 24)
+#define   DPIO_PCS_TX2DEEMP_9P5		(0 << 24)
+#define   DPIO_PCS_TX2DEEMP_6P0		(2 << 24)
+#define   DPIO_PCS_TX1DEEMP_MASK	(0xf << 16)
+#define   DPIO_PCS_TX1DEEMP_9P5		(0 << 16)
+#define   DPIO_PCS_TX1DEEMP_6P0		(2 << 16)
+#define CHV_PCS_DW10(ch) _PORT(ch, _CHV_PCS_DW10_CH0, _CHV_PCS_DW10_CH1)
+
+#define _VLV_PCS01_DW10_CH0		0x0228
+#define _VLV_PCS23_DW10_CH0		0x0428
+#define _VLV_PCS01_DW10_CH1		0x2628
+#define _VLV_PCS23_DW10_CH1		0x2828
+#define VLV_PCS01_DW10(port) _PORT(port, _VLV_PCS01_DW10_CH0, _VLV_PCS01_DW10_CH1)
+#define VLV_PCS23_DW10(port) _PORT(port, _VLV_PCS23_DW10_CH0, _VLV_PCS23_DW10_CH1)
+
+#define _VLV_PCS_DW11_CH0		0x822c
+#define _VLV_PCS_DW11_CH1		0x842c
+#define   DPIO_TX2_STAGGER_MASK(x)	((x) << 24)
+#define   DPIO_LANEDESKEW_STRAP_OVRD	(1 << 3)
+#define   DPIO_LEFT_TXFIFO_RST_MASTER	(1 << 1)
+#define   DPIO_RIGHT_TXFIFO_RST_MASTER	(1 << 0)
+#define VLV_PCS_DW11(ch) _PORT(ch, _VLV_PCS_DW11_CH0, _VLV_PCS_DW11_CH1)
+
+#define _VLV_PCS01_DW11_CH0		0x022c
+#define _VLV_PCS23_DW11_CH0		0x042c
+#define _VLV_PCS01_DW11_CH1		0x262c
+#define _VLV_PCS23_DW11_CH1		0x282c
+#define VLV_PCS01_DW11(ch) _PORT(ch, _VLV_PCS01_DW11_CH0, _VLV_PCS01_DW11_CH1)
+#define VLV_PCS23_DW11(ch) _PORT(ch, _VLV_PCS23_DW11_CH0, _VLV_PCS23_DW11_CH1)
+
+#define _VLV_PCS01_DW12_CH0		0x0230
+#define _VLV_PCS23_DW12_CH0		0x0430
+#define _VLV_PCS01_DW12_CH1		0x2630
+#define _VLV_PCS23_DW12_CH1		0x2830
+#define VLV_PCS01_DW12(ch) _PORT(ch, _VLV_PCS01_DW12_CH0, _VLV_PCS01_DW12_CH1)
+#define VLV_PCS23_DW12(ch) _PORT(ch, _VLV_PCS23_DW12_CH0, _VLV_PCS23_DW12_CH1)
+
+#define _VLV_PCS_DW12_CH0		0x8230
+#define _VLV_PCS_DW12_CH1		0x8430
+#define   DPIO_TX2_STAGGER_MULT(x)	((x) << 20)
+#define   DPIO_TX1_STAGGER_MULT(x)	((x) << 16)
+#define   DPIO_TX1_STAGGER_MASK(x)	((x) << 8)
+#define   DPIO_LANESTAGGER_STRAP_OVRD	(1 << 6)
+#define   DPIO_LANESTAGGER_STRAP(x)	((x) << 0)
+#define VLV_PCS_DW12(ch) _PORT(ch, _VLV_PCS_DW12_CH0, _VLV_PCS_DW12_CH1)
+
+#define _VLV_PCS_DW14_CH0		0x8238
+#define _VLV_PCS_DW14_CH1		0x8438
+#define	VLV_PCS_DW14(ch) _PORT(ch, _VLV_PCS_DW14_CH0, _VLV_PCS_DW14_CH1)
+
+#define _VLV_PCS_DW23_CH0		0x825c
+#define _VLV_PCS_DW23_CH1		0x845c
+#define VLV_PCS_DW23(ch) _PORT(ch, _VLV_PCS_DW23_CH0, _VLV_PCS_DW23_CH1)
+
+#define _VLV_TX_DW2_CH0			0x8288
+#define _VLV_TX_DW2_CH1			0x8488
+#define   DPIO_SWING_MARGIN000_SHIFT	16
+#define   DPIO_SWING_MARGIN000_MASK	(0xff << DPIO_SWING_MARGIN000_SHIFT)
+#define   DPIO_UNIQ_TRANS_SCALE_SHIFT	8
+#define VLV_TX_DW2(ch) _PORT(ch, _VLV_TX_DW2_CH0, _VLV_TX_DW2_CH1)
+
+#define _VLV_TX_DW3_CH0			0x828c
+#define _VLV_TX_DW3_CH1			0x848c
+/* The following bit for CHV phy */
+#define   DPIO_TX_UNIQ_TRANS_SCALE_EN	(1 << 27)
+#define   DPIO_SWING_MARGIN101_SHIFT	16
+#define   DPIO_SWING_MARGIN101_MASK	(0xff << DPIO_SWING_MARGIN101_SHIFT)
+#define VLV_TX_DW3(ch) _PORT(ch, _VLV_TX_DW3_CH0, _VLV_TX_DW3_CH1)
+
+#define _VLV_TX_DW4_CH0			0x8290
+#define _VLV_TX_DW4_CH1			0x8490
+#define   DPIO_SWING_DEEMPH9P5_SHIFT	24
+#define   DPIO_SWING_DEEMPH9P5_MASK	(0xff << DPIO_SWING_DEEMPH9P5_SHIFT)
+#define   DPIO_SWING_DEEMPH6P0_SHIFT	16
+#define   DPIO_SWING_DEEMPH6P0_MASK	(0xff << DPIO_SWING_DEEMPH6P0_SHIFT)
+#define VLV_TX_DW4(ch) _PORT(ch, _VLV_TX_DW4_CH0, _VLV_TX_DW4_CH1)
+
+#define _VLV_TX3_DW4_CH0		0x690
+#define _VLV_TX3_DW4_CH1		0x2a90
+#define VLV_TX3_DW4(ch) _PORT(ch, _VLV_TX3_DW4_CH0, _VLV_TX3_DW4_CH1)
+
+#define _VLV_TX_DW5_CH0			0x8294
+#define _VLV_TX_DW5_CH1			0x8494
+#define   DPIO_TX_OCALINIT_EN		(1 << 31)
+#define VLV_TX_DW5(ch) _PORT(ch, _VLV_TX_DW5_CH0, _VLV_TX_DW5_CH1)
+
+#define _VLV_TX_DW11_CH0		0x82ac
+#define _VLV_TX_DW11_CH1		0x84ac
+#define VLV_TX_DW11(ch) _PORT(ch, _VLV_TX_DW11_CH0, _VLV_TX_DW11_CH1)
+
+#define _VLV_TX_DW14_CH0		0x82b8
+#define _VLV_TX_DW14_CH1		0x84b8
+#define VLV_TX_DW14(ch) _PORT(ch, _VLV_TX_DW14_CH0, _VLV_TX_DW14_CH1)
+
+/* CHV dpPhy registers */
+#define _CHV_PLL_DW0_CH0		0x8000
+#define _CHV_PLL_DW0_CH1		0x8180
+#define CHV_PLL_DW0(ch) _PIPE(ch, _CHV_PLL_DW0_CH0, _CHV_PLL_DW0_CH1)
+
+#define _CHV_PLL_DW1_CH0		0x8004
+#define _CHV_PLL_DW1_CH1		0x8184
+#define   DPIO_CHV_N_DIV_SHIFT		8
+#define   DPIO_CHV_M1_DIV_BY_2		(0 << 0)
+#define CHV_PLL_DW1(ch) _PIPE(ch, _CHV_PLL_DW1_CH0, _CHV_PLL_DW1_CH1)
+
+#define _CHV_PLL_DW2_CH0		0x8008
+#define _CHV_PLL_DW2_CH1		0x8188
+#define CHV_PLL_DW2(ch) _PIPE(ch, _CHV_PLL_DW2_CH0, _CHV_PLL_DW2_CH1)
+
+#define _CHV_PLL_DW3_CH0		0x800c
+#define _CHV_PLL_DW3_CH1		0x818c
+#define  DPIO_CHV_FRAC_DIV_EN		(1 << 16)
+#define  DPIO_CHV_FIRST_MOD		(0 << 8)
+#define  DPIO_CHV_SECOND_MOD		(1 << 8)
+#define  DPIO_CHV_FEEDFWD_GAIN_SHIFT	0
+#define  DPIO_CHV_FEEDFWD_GAIN_MASK		(0xF << 0)
+#define CHV_PLL_DW3(ch) _PIPE(ch, _CHV_PLL_DW3_CH0, _CHV_PLL_DW3_CH1)
+
+#define _CHV_PLL_DW6_CH0		0x8018
+#define _CHV_PLL_DW6_CH1		0x8198
+#define   DPIO_CHV_GAIN_CTRL_SHIFT	16
+#define	  DPIO_CHV_INT_COEFF_SHIFT	8
+#define   DPIO_CHV_PROP_COEFF_SHIFT	0
+#define CHV_PLL_DW6(ch) _PIPE(ch, _CHV_PLL_DW6_CH0, _CHV_PLL_DW6_CH1)
+
+#define _CHV_PLL_DW8_CH0		0x8020
+#define _CHV_PLL_DW8_CH1		0x81A0
+#define   DPIO_CHV_TDC_TARGET_CNT_SHIFT 0
+#define   DPIO_CHV_TDC_TARGET_CNT_MASK  (0x3FF << 0)
+#define CHV_PLL_DW8(ch) _PIPE(ch, _CHV_PLL_DW8_CH0, _CHV_PLL_DW8_CH1)
+
+#define _CHV_PLL_DW9_CH0		0x8024
+#define _CHV_PLL_DW9_CH1		0x81A4
+#define  DPIO_CHV_INT_LOCK_THRESHOLD_SHIFT		1 /* 3 bits */
+#define  DPIO_CHV_INT_LOCK_THRESHOLD_MASK		(7 << 1)
+#define  DPIO_CHV_INT_LOCK_THRESHOLD_SEL_COARSE	1 /* 1: coarse & 0 : fine  */
+#define CHV_PLL_DW9(ch) _PIPE(ch, _CHV_PLL_DW9_CH0, _CHV_PLL_DW9_CH1)
+
+#define _CHV_CMN_DW0_CH0               0x8100
+#define   DPIO_ALLDL_POWERDOWN_SHIFT_CH0	19
+#define   DPIO_ANYDL_POWERDOWN_SHIFT_CH0	18
+#define   DPIO_ALLDL_POWERDOWN			(1 << 1)
+#define   DPIO_ANYDL_POWERDOWN			(1 << 0)
+
+#define _CHV_CMN_DW5_CH0               0x8114
+#define   CHV_BUFRIGHTENA1_DISABLE	(0 << 20)
+#define   CHV_BUFRIGHTENA1_NORMAL	(1 << 20)
+#define   CHV_BUFRIGHTENA1_FORCE	(3 << 20)
+#define   CHV_BUFRIGHTENA1_MASK		(3 << 20)
+#define   CHV_BUFLEFTENA1_DISABLE	(0 << 22)
+#define   CHV_BUFLEFTENA1_NORMAL	(1 << 22)
+#define   CHV_BUFLEFTENA1_FORCE		(3 << 22)
+#define   CHV_BUFLEFTENA1_MASK		(3 << 22)
+
+#define _CHV_CMN_DW13_CH0		0x8134
+#define _CHV_CMN_DW0_CH1		0x8080
+#define   DPIO_CHV_S1_DIV_SHIFT		21
+#define   DPIO_CHV_P1_DIV_SHIFT		13 /* 3 bits */
+#define   DPIO_CHV_P2_DIV_SHIFT		8  /* 5 bits */
+#define   DPIO_CHV_K_DIV_SHIFT		4
+#define   DPIO_PLL_FREQLOCK		(1 << 1)
+#define   DPIO_PLL_LOCK			(1 << 0)
+#define CHV_CMN_DW13(ch) _PIPE(ch, _CHV_CMN_DW13_CH0, _CHV_CMN_DW0_CH1)
+
+#define _CHV_CMN_DW14_CH0		0x8138
+#define _CHV_CMN_DW1_CH1		0x8084
+#define   DPIO_AFC_RECAL		(1 << 14)
+#define   DPIO_DCLKP_EN			(1 << 13)
+#define   CHV_BUFLEFTENA2_DISABLE	(0 << 17) /* CL2 DW1 only */
+#define   CHV_BUFLEFTENA2_NORMAL	(1 << 17) /* CL2 DW1 only */
+#define   CHV_BUFLEFTENA2_FORCE		(3 << 17) /* CL2 DW1 only */
+#define   CHV_BUFLEFTENA2_MASK		(3 << 17) /* CL2 DW1 only */
+#define   CHV_BUFRIGHTENA2_DISABLE	(0 << 19) /* CL2 DW1 only */
+#define   CHV_BUFRIGHTENA2_NORMAL	(1 << 19) /* CL2 DW1 only */
+#define   CHV_BUFRIGHTENA2_FORCE	(3 << 19) /* CL2 DW1 only */
+#define   CHV_BUFRIGHTENA2_MASK		(3 << 19) /* CL2 DW1 only */
+#define CHV_CMN_DW14(ch) _PIPE(ch, _CHV_CMN_DW14_CH0, _CHV_CMN_DW1_CH1)
+
+#define _CHV_CMN_DW19_CH0		0x814c
+#define _CHV_CMN_DW6_CH1		0x8098
+#define   DPIO_ALLDL_POWERDOWN_SHIFT_CH1	30 /* CL2 DW6 only */
+#define   DPIO_ANYDL_POWERDOWN_SHIFT_CH1	29 /* CL2 DW6 only */
+#define   DPIO_DYNPWRDOWNEN_CH1		(1 << 28) /* CL2 DW6 only */
+#define   CHV_CMN_USEDCLKCHANNEL	(1 << 13)
+
+#define CHV_CMN_DW19(ch) _PIPE(ch, _CHV_CMN_DW19_CH0, _CHV_CMN_DW6_CH1)
+
+#define CHV_CMN_DW28			0x8170
+#define   DPIO_CL1POWERDOWNEN		(1 << 23)
+#define   DPIO_DYNPWRDOWNEN_CH0		(1 << 22)
+#define   DPIO_SUS_CLK_CONFIG_ON		(0 << 0)
+#define   DPIO_SUS_CLK_CONFIG_CLKREQ		(1 << 0)
+#define   DPIO_SUS_CLK_CONFIG_GATE		(2 << 0)
+#define   DPIO_SUS_CLK_CONFIG_GATE_CLKREQ	(3 << 0)
+
+#define CHV_CMN_DW30			0x8178
+#define   DPIO_CL2_LDOFUSE_PWRENB	(1 << 6)
+#define   DPIO_LRC_BYPASS		(1 << 3)
+
+#define _TXLANE(ch, lane, offset) ((ch ? 0x2400 : 0) + \
+					(lane) * 0x200 + (offset))
+
+#define CHV_TX_DW0(ch, lane) _TXLANE(ch, lane, 0x80)
+#define CHV_TX_DW1(ch, lane) _TXLANE(ch, lane, 0x84)
+#define CHV_TX_DW2(ch, lane) _TXLANE(ch, lane, 0x88)
+#define CHV_TX_DW3(ch, lane) _TXLANE(ch, lane, 0x8c)
+#define CHV_TX_DW4(ch, lane) _TXLANE(ch, lane, 0x90)
+#define CHV_TX_DW5(ch, lane) _TXLANE(ch, lane, 0x94)
+#define CHV_TX_DW6(ch, lane) _TXLANE(ch, lane, 0x98)
+#define CHV_TX_DW7(ch, lane) _TXLANE(ch, lane, 0x9c)
+#define CHV_TX_DW8(ch, lane) _TXLANE(ch, lane, 0xa0)
+#define CHV_TX_DW9(ch, lane) _TXLANE(ch, lane, 0xa4)
+#define CHV_TX_DW10(ch, lane) _TXLANE(ch, lane, 0xa8)
+#define CHV_TX_DW11(ch, lane) _TXLANE(ch, lane, 0xac)
+#define   DPIO_FRC_LATENCY_SHFIT	8
+#define CHV_TX_DW14(ch, lane) _TXLANE(ch, lane, 0xb8)
+#define   DPIO_UPAR_SHIFT		30
+
+/* BXT PHY registers */
+#define _BXT_PHY0_BASE			0x6C000
+#define _BXT_PHY1_BASE			0x162000
+#define _BXT_PHY2_BASE			0x163000
+#define BXT_PHY_BASE(phy)		_PHY3((phy), _BXT_PHY0_BASE, \
+						     _BXT_PHY1_BASE, \
+						     _BXT_PHY2_BASE)
+
+#define _BXT_PHY(phy, reg)						\
+	_MMIO(BXT_PHY_BASE(phy) - _BXT_PHY0_BASE + (reg))
+
+#define _BXT_PHY_CH(phy, ch, reg_ch0, reg_ch1)		\
+	(BXT_PHY_BASE(phy) + _PIPE((ch), (reg_ch0) - _BXT_PHY0_BASE,	\
+					 (reg_ch1) - _BXT_PHY0_BASE))
+#define _MMIO_BXT_PHY_CH(phy, ch, reg_ch0, reg_ch1)		\
+	_MMIO(_BXT_PHY_CH(phy, ch, reg_ch0, reg_ch1))
+
+#define BXT_P_CR_GT_DISP_PWRON		_MMIO(0x138090)
+#define  MIPIO_RST_CTRL				(1 << 2)
+
+#define _BXT_PHY_CTL_DDI_A		0x64C00
+#define _BXT_PHY_CTL_DDI_B		0x64C10
+#define _BXT_PHY_CTL_DDI_C		0x64C20
+#define   BXT_PHY_CMNLANE_POWERDOWN_ACK	(1 << 10)
+#define   BXT_PHY_LANE_POWERDOWN_ACK	(1 << 9)
+#define   BXT_PHY_LANE_ENABLED		(1 << 8)
+#define BXT_PHY_CTL(port)		_MMIO_PORT(port, _BXT_PHY_CTL_DDI_A, \
+							 _BXT_PHY_CTL_DDI_B)
+
+#define _PHY_CTL_FAMILY_EDP		0x64C80
+#define _PHY_CTL_FAMILY_DDI		0x64C90
+#define _PHY_CTL_FAMILY_DDI_C		0x64CA0
+#define   COMMON_RESET_DIS		(1 << 31)
+#define BXT_PHY_CTL_FAMILY(phy)		_MMIO_PHY3((phy), _PHY_CTL_FAMILY_DDI, \
+							  _PHY_CTL_FAMILY_EDP, \
+							  _PHY_CTL_FAMILY_DDI_C)
+
+/* BXT PHY PLL registers */
+#define _PORT_PLL_A			0x46074
+#define _PORT_PLL_B			0x46078
+#define _PORT_PLL_C			0x4607c
+#define   PORT_PLL_ENABLE		REG_BIT(31)
+#define   PORT_PLL_LOCK			REG_BIT(30)
+#define   PORT_PLL_REF_SEL		REG_BIT(27)
+#define   PORT_PLL_POWER_ENABLE		REG_BIT(26)
+#define   PORT_PLL_POWER_STATE		REG_BIT(25)
+#define BXT_PORT_PLL_ENABLE(port)	_MMIO_PORT(port, _PORT_PLL_A, _PORT_PLL_B)
+
+#define _PORT_PLL_EBB_0_A		0x162034
+#define _PORT_PLL_EBB_0_B		0x6C034
+#define _PORT_PLL_EBB_0_C		0x6C340
+#define   PORT_PLL_P1_MASK		REG_GENMASK(15, 13)
+#define   PORT_PLL_P1(p1)		REG_FIELD_PREP(PORT_PLL_P1_MASK, (p1))
+#define   PORT_PLL_P2_MASK		REG_GENMASK(12, 8)
+#define   PORT_PLL_P2(p2)		REG_FIELD_PREP(PORT_PLL_P2_MASK, (p2))
+#define BXT_PORT_PLL_EBB_0(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_PLL_EBB_0_B, \
+							 _PORT_PLL_EBB_0_C)
+
+#define _PORT_PLL_EBB_4_A		0x162038
+#define _PORT_PLL_EBB_4_B		0x6C038
+#define _PORT_PLL_EBB_4_C		0x6C344
+#define   PORT_PLL_RECALIBRATE		REG_BIT(14)
+#define   PORT_PLL_10BIT_CLK_ENABLE	REG_BIT(13)
+#define BXT_PORT_PLL_EBB_4(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_PLL_EBB_4_B, \
+							 _PORT_PLL_EBB_4_C)
+
+#define _PORT_PLL_0_A			0x162100
+#define _PORT_PLL_0_B			0x6C100
+#define _PORT_PLL_0_C			0x6C380
+/* PORT_PLL_0_A */
+#define   PORT_PLL_M2_INT_MASK		REG_GENMASK(7, 0)
+#define   PORT_PLL_M2_INT(m2_int)	REG_FIELD_PREP(PORT_PLL_M2_INT_MASK, (m2_int))
+/* PORT_PLL_1_A */
+#define   PORT_PLL_N_MASK		REG_GENMASK(11, 8)
+#define   PORT_PLL_N(n)			REG_FIELD_PREP(PORT_PLL_N_MASK, (n))
+/* PORT_PLL_2_A */
+#define   PORT_PLL_M2_FRAC_MASK		REG_GENMASK(21, 0)
+#define   PORT_PLL_M2_FRAC(m2_frac)	REG_FIELD_PREP(PORT_PLL_M2_FRAC_MASK, (m2_frac))
+/* PORT_PLL_3_A */
+#define   PORT_PLL_M2_FRAC_ENABLE	REG_BIT(16)
+/* PORT_PLL_6_A */
+#define   PORT_PLL_GAIN_CTL_MASK	REG_GENMASK(18, 16)
+#define   PORT_PLL_GAIN_CTL(x)		REG_FIELD_PREP(PORT_PLL_GAIN_CTL_MASK, (x))
+#define   PORT_PLL_INT_COEFF_MASK	REG_GENMASK(12, 8)
+#define   PORT_PLL_INT_COEFF(x)		REG_FIELD_PREP(PORT_PLL_INT_COEFF_MASK, (x))
+#define   PORT_PLL_PROP_COEFF_MASK	REG_GENMASK(3, 0)
+#define   PORT_PLL_PROP_COEFF(x)	REG_FIELD_PREP(PORT_PLL_PROP_COEFF_MASK, (x))
+/* PORT_PLL_8_A */
+#define   PORT_PLL_TARGET_CNT_MASK	REG_GENMASK(9, 0)
+#define   PORT_PLL_TARGET_CNT(x)	REG_FIELD_PREP(PORT_PLL_TARGET_CNT_MASK, (x))
+/* PORT_PLL_9_A */
+#define  PORT_PLL_LOCK_THRESHOLD_MASK	REG_GENMASK(3, 1)
+#define  PORT_PLL_LOCK_THRESHOLD(x)	REG_FIELD_PREP(PORT_PLL_LOCK_THRESHOLD_MASK, (x))
+/* PORT_PLL_10_A */
+#define  PORT_PLL_DCO_AMP_OVR_EN_H	REG_BIT(27)
+#define  PORT_PLL_DCO_AMP_MASK		REG_GENMASK(13, 10)
+#define  PORT_PLL_DCO_AMP(x)		REG_FIELD_PREP(PORT_PLL_DCO_AMP_MASK, (x))
+#define _PORT_PLL_BASE(phy, ch)		_BXT_PHY_CH(phy, ch, \
+						    _PORT_PLL_0_B, \
+						    _PORT_PLL_0_C)
+#define BXT_PORT_PLL(phy, ch, idx)	_MMIO(_PORT_PLL_BASE(phy, ch) + \
+					      (idx) * 4)
+
+/* BXT PHY common lane registers */
+#define _PORT_CL1CM_DW0_A		0x162000
+#define _PORT_CL1CM_DW0_BC		0x6C000
+#define   PHY_POWER_GOOD		(1 << 16)
+#define   PHY_RESERVED			(1 << 7)
+#define BXT_PORT_CL1CM_DW0(phy)		_BXT_PHY((phy), _PORT_CL1CM_DW0_BC)
+
+#define _PORT_CL1CM_DW9_A		0x162024
+#define _PORT_CL1CM_DW9_BC		0x6C024
+#define   IREF0RC_OFFSET_SHIFT		8
+#define   IREF0RC_OFFSET_MASK		(0xFF << IREF0RC_OFFSET_SHIFT)
+#define BXT_PORT_CL1CM_DW9(phy)		_BXT_PHY((phy), _PORT_CL1CM_DW9_BC)
+
+#define _PORT_CL1CM_DW10_A		0x162028
+#define _PORT_CL1CM_DW10_BC		0x6C028
+#define   IREF1RC_OFFSET_SHIFT		8
+#define   IREF1RC_OFFSET_MASK		(0xFF << IREF1RC_OFFSET_SHIFT)
+#define BXT_PORT_CL1CM_DW10(phy)	_BXT_PHY((phy), _PORT_CL1CM_DW10_BC)
+
+#define _PORT_CL1CM_DW28_A		0x162070
+#define _PORT_CL1CM_DW28_BC		0x6C070
+#define   OCL1_POWER_DOWN_EN		(1 << 23)
+#define   DW28_OLDO_DYN_PWR_DOWN_EN	(1 << 22)
+#define   SUS_CLK_CONFIG		0x3
+#define BXT_PORT_CL1CM_DW28(phy)	_BXT_PHY((phy), _PORT_CL1CM_DW28_BC)
+
+#define _PORT_CL1CM_DW30_A		0x162078
+#define _PORT_CL1CM_DW30_BC		0x6C078
+#define   OCL2_LDOFUSE_PWR_DIS		(1 << 6)
+#define BXT_PORT_CL1CM_DW30(phy)	_BXT_PHY((phy), _PORT_CL1CM_DW30_BC)
+
+/* The spec defines this only for BXT PHY0, but lets assume that this
+ * would exist for PHY1 too if it had a second channel.
+ */
+#define _PORT_CL2CM_DW6_A		0x162358
+#define _PORT_CL2CM_DW6_BC		0x6C358
+#define BXT_PORT_CL2CM_DW6(phy)		_BXT_PHY((phy), _PORT_CL2CM_DW6_BC)
+#define   DW6_OLDO_DYN_PWR_DOWN_EN	(1 << 28)
+
+/* BXT PHY Ref registers */
+#define _PORT_REF_DW3_A			0x16218C
+#define _PORT_REF_DW3_BC		0x6C18C
+#define   GRC_DONE			(1 << 22)
+#define BXT_PORT_REF_DW3(phy)		_BXT_PHY((phy), _PORT_REF_DW3_BC)
+
+#define _PORT_REF_DW6_A			0x162198
+#define _PORT_REF_DW6_BC		0x6C198
+#define   GRC_CODE_SHIFT		24
+#define   GRC_CODE_MASK			(0xFF << GRC_CODE_SHIFT)
+#define   GRC_CODE_FAST_SHIFT		16
+#define   GRC_CODE_FAST_MASK		(0xFF << GRC_CODE_FAST_SHIFT)
+#define   GRC_CODE_SLOW_SHIFT		8
+#define   GRC_CODE_SLOW_MASK		(0xFF << GRC_CODE_SLOW_SHIFT)
+#define   GRC_CODE_NOM_MASK		0xFF
+#define BXT_PORT_REF_DW6(phy)		_BXT_PHY((phy), _PORT_REF_DW6_BC)
+
+#define _PORT_REF_DW8_A			0x1621A0
+#define _PORT_REF_DW8_BC		0x6C1A0
+#define   GRC_DIS			(1 << 15)
+#define   GRC_RDY_OVRD			(1 << 1)
+#define BXT_PORT_REF_DW8(phy)		_BXT_PHY((phy), _PORT_REF_DW8_BC)
+
+/* BXT PHY PCS registers */
+#define _PORT_PCS_DW10_LN01_A		0x162428
+#define _PORT_PCS_DW10_LN01_B		0x6C428
+#define _PORT_PCS_DW10_LN01_C		0x6C828
+#define _PORT_PCS_DW10_GRP_A		0x162C28
+#define _PORT_PCS_DW10_GRP_B		0x6CC28
+#define _PORT_PCS_DW10_GRP_C		0x6CE28
+#define BXT_PORT_PCS_DW10_LN01(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_PCS_DW10_LN01_B, \
+							 _PORT_PCS_DW10_LN01_C)
+#define BXT_PORT_PCS_DW10_GRP(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_PCS_DW10_GRP_B, \
+							 _PORT_PCS_DW10_GRP_C)
+
+#define   TX2_SWING_CALC_INIT		(1 << 31)
+#define   TX1_SWING_CALC_INIT		(1 << 30)
+
+#define _PORT_PCS_DW12_LN01_A		0x162430
+#define _PORT_PCS_DW12_LN01_B		0x6C430
+#define _PORT_PCS_DW12_LN01_C		0x6C830
+#define _PORT_PCS_DW12_LN23_A		0x162630
+#define _PORT_PCS_DW12_LN23_B		0x6C630
+#define _PORT_PCS_DW12_LN23_C		0x6CA30
+#define _PORT_PCS_DW12_GRP_A		0x162c30
+#define _PORT_PCS_DW12_GRP_B		0x6CC30
+#define _PORT_PCS_DW12_GRP_C		0x6CE30
+#define   LANESTAGGER_STRAP_OVRD	(1 << 6)
+#define   LANE_STAGGER_MASK		0x1F
+#define BXT_PORT_PCS_DW12_LN01(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_PCS_DW12_LN01_B, \
+							 _PORT_PCS_DW12_LN01_C)
+#define BXT_PORT_PCS_DW12_LN23(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_PCS_DW12_LN23_B, \
+							 _PORT_PCS_DW12_LN23_C)
+#define BXT_PORT_PCS_DW12_GRP(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_PCS_DW12_GRP_B, \
+							 _PORT_PCS_DW12_GRP_C)
+
+/* BXT PHY TX registers */
+#define _BXT_LANE_OFFSET(lane)           (((lane) >> 1) * 0x200 +	\
+					  ((lane) & 1) * 0x80)
+
+#define _PORT_TX_DW2_LN0_A		0x162508
+#define _PORT_TX_DW2_LN0_B		0x6C508
+#define _PORT_TX_DW2_LN0_C		0x6C908
+#define _PORT_TX_DW2_GRP_A		0x162D08
+#define _PORT_TX_DW2_GRP_B		0x6CD08
+#define _PORT_TX_DW2_GRP_C		0x6CF08
+#define BXT_PORT_TX_DW2_LN0(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_TX_DW2_LN0_B, \
+							 _PORT_TX_DW2_LN0_C)
+#define BXT_PORT_TX_DW2_GRP(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_TX_DW2_GRP_B, \
+							 _PORT_TX_DW2_GRP_C)
+#define   MARGIN_000_SHIFT		16
+#define   MARGIN_000			(0xFF << MARGIN_000_SHIFT)
+#define   UNIQ_TRANS_SCALE_SHIFT	8
+#define   UNIQ_TRANS_SCALE		(0xFF << UNIQ_TRANS_SCALE_SHIFT)
+
+#define _PORT_TX_DW3_LN0_A		0x16250C
+#define _PORT_TX_DW3_LN0_B		0x6C50C
+#define _PORT_TX_DW3_LN0_C		0x6C90C
+#define _PORT_TX_DW3_GRP_A		0x162D0C
+#define _PORT_TX_DW3_GRP_B		0x6CD0C
+#define _PORT_TX_DW3_GRP_C		0x6CF0C
+#define BXT_PORT_TX_DW3_LN0(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_TX_DW3_LN0_B, \
+							 _PORT_TX_DW3_LN0_C)
+#define BXT_PORT_TX_DW3_GRP(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_TX_DW3_GRP_B, \
+							 _PORT_TX_DW3_GRP_C)
+#define   SCALE_DCOMP_METHOD		(1 << 26)
+#define   UNIQUE_TRANGE_EN_METHOD	(1 << 27)
+
+#define _PORT_TX_DW4_LN0_A		0x162510
+#define _PORT_TX_DW4_LN0_B		0x6C510
+#define _PORT_TX_DW4_LN0_C		0x6C910
+#define _PORT_TX_DW4_GRP_A		0x162D10
+#define _PORT_TX_DW4_GRP_B		0x6CD10
+#define _PORT_TX_DW4_GRP_C		0x6CF10
+#define BXT_PORT_TX_DW4_LN0(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_TX_DW4_LN0_B, \
+							 _PORT_TX_DW4_LN0_C)
+#define BXT_PORT_TX_DW4_GRP(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_TX_DW4_GRP_B, \
+							 _PORT_TX_DW4_GRP_C)
+#define   DEEMPH_SHIFT			24
+#define   DE_EMPHASIS			(0xFF << DEEMPH_SHIFT)
+
+#define _PORT_TX_DW5_LN0_A		0x162514
+#define _PORT_TX_DW5_LN0_B		0x6C514
+#define _PORT_TX_DW5_LN0_C		0x6C914
+#define _PORT_TX_DW5_GRP_A		0x162D14
+#define _PORT_TX_DW5_GRP_B		0x6CD14
+#define _PORT_TX_DW5_GRP_C		0x6CF14
+#define BXT_PORT_TX_DW5_LN0(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_TX_DW5_LN0_B, \
+							 _PORT_TX_DW5_LN0_C)
+#define BXT_PORT_TX_DW5_GRP(phy, ch)	_MMIO_BXT_PHY_CH(phy, ch, \
+							 _PORT_TX_DW5_GRP_B, \
+							 _PORT_TX_DW5_GRP_C)
+#define   DCC_DELAY_RANGE_1		(1 << 9)
+#define   DCC_DELAY_RANGE_2		(1 << 8)
+
+#define _PORT_TX_DW14_LN0_A		0x162538
+#define _PORT_TX_DW14_LN0_B		0x6C538
+#define _PORT_TX_DW14_LN0_C		0x6C938
+#define   LATENCY_OPTIM_SHIFT		30
+#define   LATENCY_OPTIM			(1 << LATENCY_OPTIM_SHIFT)
+#define BXT_PORT_TX_DW14_LN(phy, ch, lane)				\
+	_MMIO(_BXT_PHY_CH(phy, ch, _PORT_TX_DW14_LN0_B,			\
+				   _PORT_TX_DW14_LN0_C) +		\
+	      _BXT_LANE_OFFSET(lane))
+
+/* UAIMI scratch pad register 1 */
+#define UAIMI_SPR1			_MMIO(0x4F074)
+/* SKL VccIO mask */
+#define SKL_VCCIO_MASK			0x1
+/* SKL balance leg register */
+#define DISPIO_CR_TX_BMU_CR0		_MMIO(0x6C00C)
+/* I_boost values */
+#define BALANCE_LEG_SHIFT(port)		(8 + 3 * (port))
+#define BALANCE_LEG_MASK(port)		(7 << (8 + 3 * (port)))
+/* Balance leg disable bits */
+#define BALANCE_LEG_DISABLE_SHIFT	23
+#define BALANCE_LEG_DISABLE(port)	(1 << (23 + (port)))
+
+/*
+ * Fence registers
+ * [0-7]  @ 0x2000 gen2,gen3
+ * [8-15] @ 0x3000 945,g33,pnv
+ *
+ * [0-15] @ 0x3000 gen4,gen5
+ *
+ * [0-15] @ 0x100000 gen6,vlv,chv
+ * [0-31] @ 0x100000 gen7+
+ */
+#define FENCE_REG(i)			_MMIO(0x2000 + (((i) & 8) << 9) + ((i) & 7) * 4)
+#define   I830_FENCE_START_MASK		0x07f80000
+#define   I830_FENCE_TILING_Y_SHIFT	12
+#define   I830_FENCE_SIZE_BITS(size)	((ffs((size) >> 19) - 1) << 8)
+#define   I830_FENCE_PITCH_SHIFT	4
+#define   I830_FENCE_REG_VALID		(1 << 0)
+#define   I915_FENCE_MAX_PITCH_VAL	4
+#define   I830_FENCE_MAX_PITCH_VAL	6
+#define   I830_FENCE_MAX_SIZE_VAL	(1 << 8)
+
+#define   I915_FENCE_START_MASK		0x0ff00000
+#define   I915_FENCE_SIZE_BITS(size)	((ffs((size) >> 20) - 1) << 8)
+
+#define FENCE_REG_965_LO(i)		_MMIO(0x03000 + (i) * 8)
+#define FENCE_REG_965_HI(i)		_MMIO(0x03000 + (i) * 8 + 4)
+#define   I965_FENCE_PITCH_SHIFT	2
+#define   I965_FENCE_TILING_Y_SHIFT	1
+#define   I965_FENCE_REG_VALID		(1 << 0)
+#define   I965_FENCE_MAX_PITCH_VAL	0x0400
+
+#define FENCE_REG_GEN6_LO(i)		_MMIO(0x100000 + (i) * 8)
+#define FENCE_REG_GEN6_HI(i)		_MMIO(0x100000 + (i) * 8 + 4)
+#define   GEN6_FENCE_PITCH_SHIFT	32
+#define   GEN7_FENCE_MAX_PITCH_VAL	0x0800
+
+
+/* control register for cpu gtt access */
+#define TILECTL				_MMIO(0x101000)
+#define   TILECTL_SWZCTL			(1 << 0)
+#define   TILECTL_TLBPF			(1 << 1)
+#define   TILECTL_TLB_PREFETCH_DIS	(1 << 2)
+#define   TILECTL_BACKSNOOP_DIS		(1 << 3)
+
+/*
+ * Instruction and interrupt control regs
+ */
+#define PGTBL_CTL	_MMIO(0x02020)
+#define   PGTBL_ADDRESS_LO_MASK	0xfffff000 /* bits [31:12] */
+#define   PGTBL_ADDRESS_HI_MASK	0x000000f0 /* bits [35:32] (gen4) */
+#define PGTBL_ER	_MMIO(0x02024)
+#define PRB0_BASE	(0x2030 - 0x30)
+#define PRB1_BASE	(0x2040 - 0x30) /* 830,gen3 */
+#define PRB2_BASE	(0x2050 - 0x30) /* gen3 */
+#define SRB0_BASE	(0x2100 - 0x30) /* gen2 */
+#define SRB1_BASE	(0x2110 - 0x30) /* gen2 */
+#define SRB2_BASE	(0x2120 - 0x30) /* 830 */
+#define SRB3_BASE	(0x2130 - 0x30) /* 830 */
+#define RENDER_RING_BASE	0x02000
+#define BSD_RING_BASE		0x04000
+#define GEN6_BSD_RING_BASE	0x12000
+#define GEN8_BSD2_RING_BASE	0x1c000
+#define GEN11_BSD_RING_BASE	0x1c0000
+#define GEN11_BSD2_RING_BASE	0x1c4000
+#define GEN11_BSD3_RING_BASE	0x1d0000
+#define GEN11_BSD4_RING_BASE	0x1d4000
+#define XEHP_BSD5_RING_BASE	0x1e0000
+#define XEHP_BSD6_RING_BASE	0x1e4000
+#define XEHP_BSD7_RING_BASE	0x1f0000
+#define XEHP_BSD8_RING_BASE	0x1f4000
+#define VEBOX_RING_BASE		0x1a000
+#define GEN11_VEBOX_RING_BASE		0x1c8000
+#define GEN11_VEBOX2_RING_BASE		0x1d8000
+#define XEHP_VEBOX3_RING_BASE		0x1e8000
+#define XEHP_VEBOX4_RING_BASE		0x1f8000
+#define GEN12_COMPUTE0_RING_BASE	0x1a000
+#define GEN12_COMPUTE1_RING_BASE	0x1c000
+#define GEN12_COMPUTE2_RING_BASE	0x1e000
+#define GEN12_COMPUTE3_RING_BASE	0x26000
+#define BLT_RING_BASE		0x22000
+#define XEHPC_BCS1_RING_BASE	0x3e0000
+#define XEHPC_BCS2_RING_BASE	0x3e2000
+#define XEHPC_BCS3_RING_BASE	0x3e4000
+#define XEHPC_BCS4_RING_BASE	0x3e6000
+#define XEHPC_BCS5_RING_BASE	0x3e8000
+#define XEHPC_BCS6_RING_BASE	0x3ea000
+#define XEHPC_BCS7_RING_BASE	0x3ec000
+#define XEHPC_BCS8_RING_BASE	0x3ee000
+#define DG1_GSC_HECI1_BASE	0x00258000
+#define DG1_GSC_HECI2_BASE	0x00259000
+#define DG2_GSC_HECI1_BASE	0x00373000
+#define DG2_GSC_HECI2_BASE	0x00374000
+
+
+
+#define HSW_GTT_CACHE_EN	_MMIO(0x4024)
+#define   GTT_CACHE_EN_ALL	0xF0007FFF
+#define GEN7_WR_WATERMARK	_MMIO(0x4028)
+#define GEN7_GFX_PRIO_CTRL	_MMIO(0x402C)
+#define ARB_MODE		_MMIO(0x4030)
+#define   ARB_MODE_SWIZZLE_SNB	(1 << 4)
+#define   ARB_MODE_SWIZZLE_IVB	(1 << 5)
+#define GEN7_GFX_PEND_TLB0	_MMIO(0x4034)
+#define GEN7_GFX_PEND_TLB1	_MMIO(0x4038)
+/* L3, CVS, ZTLB, RCC, CASC LRA min, max values */
+#define GEN7_LRA_LIMITS(i)	_MMIO(0x403C + (i) * 4)
+#define GEN7_LRA_LIMITS_REG_NUM	13
+#define GEN7_MEDIA_MAX_REQ_COUNT	_MMIO(0x4070)
+#define GEN7_GFX_MAX_REQ_COUNT		_MMIO(0x4074)
+
+#define GEN7_ERR_INT	_MMIO(0x44040)
+#define   ERR_INT_POISON		(1 << 31)
+#define   ERR_INT_MMIO_UNCLAIMED	(1 << 13)
+#define   ERR_INT_PIPE_CRC_DONE_C	(1 << 8)
+#define   ERR_INT_FIFO_UNDERRUN_C	(1 << 6)
+#define   ERR_INT_PIPE_CRC_DONE_B	(1 << 5)
+#define   ERR_INT_FIFO_UNDERRUN_B	(1 << 3)
+#define   ERR_INT_PIPE_CRC_DONE_A	(1 << 2)
+#define   ERR_INT_PIPE_CRC_DONE(pipe)	(1 << (2 + (pipe) * 3))
+#define   ERR_INT_FIFO_UNDERRUN_A	(1 << 0)
+#define   ERR_INT_FIFO_UNDERRUN(pipe)	(1 << ((pipe) * 3))
+
+#define FPGA_DBG		_MMIO(0x42300)
+#define   FPGA_DBG_RM_NOCLAIM	REG_BIT(31)
+
+#define CLAIM_ER		_MMIO(VLV_DISPLAY_BASE + 0x2028)
+#define   CLAIM_ER_CLR		REG_BIT(31)
+#define   CLAIM_ER_OVERFLOW	REG_BIT(16)
+#define   CLAIM_ER_CTR_MASK	REG_GENMASK(15, 0)
+
+#define DERRMR		_MMIO(0x44050)
+/* Note that HBLANK events are reserved on bdw+ */
+#define   DERRMR_PIPEA_SCANLINE		(1 << 0)
+#define   DERRMR_PIPEA_PRI_FLIP_DONE	(1 << 1)
+#define   DERRMR_PIPEA_SPR_FLIP_DONE	(1 << 2)
+#define   DERRMR_PIPEA_VBLANK		(1 << 3)
+#define   DERRMR_PIPEA_HBLANK		(1 << 5)
+#define   DERRMR_PIPEB_SCANLINE		(1 << 8)
+#define   DERRMR_PIPEB_PRI_FLIP_DONE	(1 << 9)
+#define   DERRMR_PIPEB_SPR_FLIP_DONE	(1 << 10)
+#define   DERRMR_PIPEB_VBLANK		(1 << 11)
+#define   DERRMR_PIPEB_HBLANK		(1 << 13)
+/* Note that PIPEC is not a simple translation of PIPEA/PIPEB */
+#define   DERRMR_PIPEC_SCANLINE		(1 << 14)
+#define   DERRMR_PIPEC_PRI_FLIP_DONE	(1 << 15)
+#define   DERRMR_PIPEC_SPR_FLIP_DONE	(1 << 20)
+#define   DERRMR_PIPEC_VBLANK		(1 << 21)
+#define   DERRMR_PIPEC_HBLANK		(1 << 22)
+
+#define VLV_GU_CTL0	_MMIO(VLV_DISPLAY_BASE + 0x2030)
+#define VLV_GU_CTL1	_MMIO(VLV_DISPLAY_BASE + 0x2034)
+#define SCPD0		_MMIO(0x209c) /* 915+ only */
+#define  SCPD_FBC_IGNORE_3D			(1 << 6)
+#define  CSTATE_RENDER_CLOCK_GATE_DISABLE	(1 << 5)
+#define GEN2_IER	_MMIO(0x20a0)
+#define GEN2_IIR	_MMIO(0x20a4)
+#define GEN2_IMR	_MMIO(0x20a8)
+#define GEN2_ISR	_MMIO(0x20ac)
+#define VLV_GUNIT_CLOCK_GATE	_MMIO(VLV_DISPLAY_BASE + 0x2060)
+#define   GINT_DIS		(1 << 22)
+#define   GCFG_DIS		(1 << 8)
+#define VLV_GUNIT_CLOCK_GATE2	_MMIO(VLV_DISPLAY_BASE + 0x2064)
+#define VLV_IIR_RW	_MMIO(VLV_DISPLAY_BASE + 0x2084)
+#define VLV_IER		_MMIO(VLV_DISPLAY_BASE + 0x20a0)
+#define VLV_IIR		_MMIO(VLV_DISPLAY_BASE + 0x20a4)
+#define VLV_IMR		_MMIO(VLV_DISPLAY_BASE + 0x20a8)
+#define VLV_ISR		_MMIO(VLV_DISPLAY_BASE + 0x20ac)
+#define VLV_PCBR	_MMIO(VLV_DISPLAY_BASE + 0x2120)
+#define VLV_PCBR_ADDR_SHIFT	12
+
+#define   DISPLAY_PLANE_FLIP_PENDING(plane) (1 << (11 - (plane))) /* A and B only */
+#define EIR		_MMIO(0x20b0)
+#define EMR		_MMIO(0x20b4)
+#define ESR		_MMIO(0x20b8)
+#define   GM45_ERROR_PAGE_TABLE				(1 << 5)
+#define   GM45_ERROR_MEM_PRIV				(1 << 4)
+#define   I915_ERROR_PAGE_TABLE				(1 << 4)
+#define   GM45_ERROR_CP_PRIV				(1 << 3)
+#define   I915_ERROR_MEMORY_REFRESH			(1 << 1)
+#define   I915_ERROR_INSTRUCTION			(1 << 0)
+#define INSTPM	        _MMIO(0x20c0)
+#define   INSTPM_SELF_EN (1 << 12) /* 915GM only */
+#define   INSTPM_AGPBUSY_INT_EN (1 << 11) /* gen3: when disabled, pending interrupts
+					will not assert AGPBUSY# and will only
+					be delivered when out of C3. */
+#define   INSTPM_FORCE_ORDERING				(1 << 7) /* GEN6+ */
+#define   INSTPM_TLB_INVALIDATE	(1 << 9)
+#define   INSTPM_SYNC_FLUSH	(1 << 5)
+#define MEM_MODE	_MMIO(0x20cc)
+#define   MEM_DISPLAY_B_TRICKLE_FEED_DISABLE (1 << 3) /* 830 only */
+#define   MEM_DISPLAY_A_TRICKLE_FEED_DISABLE (1 << 2) /* 830/845 only */
+#define   MEM_DISPLAY_TRICKLE_FEED_DISABLE (1 << 2) /* 85x only */
+#define FW_BLC		_MMIO(0x20d8)
+#define FW_BLC2		_MMIO(0x20dc)
+#define FW_BLC_SELF	_MMIO(0x20e0) /* 915+ only */
+#define   FW_BLC_SELF_EN_MASK      (1 << 31)
+#define   FW_BLC_SELF_FIFO_MASK    (1 << 16) /* 945 only */
+#define   FW_BLC_SELF_EN           (1 << 15) /* 945 only */
+#define MM_BURST_LENGTH     0x00700000
+#define MM_FIFO_WATERMARK   0x0001F000
+#define LM_BURST_LENGTH     0x00000700
+#define LM_FIFO_WATERMARK   0x0000001F
+#define MI_ARB_STATE	_MMIO(0x20e4) /* 915+ only */
+
+#define _MBUS_ABOX0_CTL			0x45038
+#define _MBUS_ABOX1_CTL			0x45048
+#define _MBUS_ABOX2_CTL			0x4504C
+#define MBUS_ABOX_CTL(x)		_MMIO(_PICK(x, _MBUS_ABOX0_CTL, \
+						    _MBUS_ABOX1_CTL, \
+						    _MBUS_ABOX2_CTL))
+#define MBUS_ABOX_BW_CREDIT_MASK	(3 << 20)
+#define MBUS_ABOX_BW_CREDIT(x)		((x) << 20)
+#define MBUS_ABOX_B_CREDIT_MASK		(0xF << 16)
+#define MBUS_ABOX_B_CREDIT(x)		((x) << 16)
+#define MBUS_ABOX_BT_CREDIT_POOL2_MASK	(0x1F << 8)
+#define MBUS_ABOX_BT_CREDIT_POOL2(x)	((x) << 8)
+#define MBUS_ABOX_BT_CREDIT_POOL1_MASK	(0x1F << 0)
+#define MBUS_ABOX_BT_CREDIT_POOL1(x)	((x) << 0)
+
+#define _PIPEA_MBUS_DBOX_CTL			0x7003C
+#define _PIPEB_MBUS_DBOX_CTL			0x7103C
+#define PIPE_MBUS_DBOX_CTL(pipe)		_MMIO_PIPE(pipe, _PIPEA_MBUS_DBOX_CTL, \
+							   _PIPEB_MBUS_DBOX_CTL)
+#define MBUS_DBOX_B2B_TRANSACTIONS_MAX_MASK	REG_GENMASK(24, 20) /* tgl+ */
+#define MBUS_DBOX_B2B_TRANSACTIONS_MAX(x)	REG_FIELD_PREP(MBUS_DBOX_B2B_TRANSACTIONS_MAX_MASK, x)
+#define MBUS_DBOX_B2B_TRANSACTIONS_DELAY_MASK	REG_GENMASK(19, 17) /* tgl+ */
+#define MBUS_DBOX_B2B_TRANSACTIONS_DELAY(x)	REG_FIELD_PREP(MBUS_DBOX_B2B_TRANSACTIONS_DELAY_MASK, x)
+#define MBUS_DBOX_REGULATE_B2B_TRANSACTIONS_EN	REG_BIT(16) /* tgl+ */
+#define MBUS_DBOX_BW_CREDIT_MASK		REG_GENMASK(15, 14)
+#define MBUS_DBOX_BW_CREDIT(x)			REG_FIELD_PREP(MBUS_DBOX_BW_CREDIT_MASK, x)
+#define MBUS_DBOX_BW_4CREDITS_MTL		REG_FIELD_PREP(MBUS_DBOX_BW_CREDIT_MASK, 0x2)
+#define MBUS_DBOX_BW_8CREDITS_MTL		REG_FIELD_PREP(MBUS_DBOX_BW_CREDIT_MASK, 0x3)
+#define MBUS_DBOX_B_CREDIT_MASK			REG_GENMASK(12, 8)
+#define MBUS_DBOX_B_CREDIT(x)			REG_FIELD_PREP(MBUS_DBOX_B_CREDIT_MASK, x)
+#define MBUS_DBOX_I_CREDIT_MASK			REG_GENMASK(7, 5)
+#define MBUS_DBOX_I_CREDIT(x)			REG_FIELD_PREP(MBUS_DBOX_I_CREDIT_MASK, x)
+#define MBUS_DBOX_A_CREDIT_MASK			REG_GENMASK(3, 0)
+#define MBUS_DBOX_A_CREDIT(x)			REG_FIELD_PREP(MBUS_DBOX_A_CREDIT_MASK, x)
+
+#define MBUS_UBOX_CTL			_MMIO(0x4503C)
+#define MBUS_BBOX_CTL_S1		_MMIO(0x45040)
+#define MBUS_BBOX_CTL_S2		_MMIO(0x45044)
+
+#define MBUS_CTL			_MMIO(0x4438C)
+#define MBUS_JOIN			REG_BIT(31)
+#define MBUS_HASHING_MODE_MASK		REG_BIT(30)
+#define MBUS_HASHING_MODE_2x2		REG_FIELD_PREP(MBUS_HASHING_MODE_MASK, 0)
+#define MBUS_HASHING_MODE_1x4		REG_FIELD_PREP(MBUS_HASHING_MODE_MASK, 1)
+#define MBUS_JOIN_PIPE_SELECT_MASK	REG_GENMASK(28, 26)
+#define MBUS_JOIN_PIPE_SELECT(pipe)	REG_FIELD_PREP(MBUS_JOIN_PIPE_SELECT_MASK, pipe)
+#define MBUS_JOIN_PIPE_SELECT_NONE	MBUS_JOIN_PIPE_SELECT(7)
+
+#define HDPORT_STATE			_MMIO(0x45050)
+#define   HDPORT_DPLL_USED_MASK		REG_GENMASK(15, 12)
+#define   HDPORT_DDI_USED(phy)		REG_BIT(2 * (phy) + 1)
+#define   HDPORT_ENABLED		REG_BIT(0)
+
+/* Make render/texture TLB fetches lower priorty than associated data
+ *   fetches. This is not turned on by default
+ */
+#define   MI_ARB_RENDER_TLB_LOW_PRIORITY	(1 << 15)
+
+/* Isoch request wait on GTT enable (Display A/B/C streams).
+ * Make isoch requests stall on the TLB update. May cause
+ * display underruns (test mode only)
+ */
+#define   MI_ARB_ISOCH_WAIT_GTT			(1 << 14)
+
+/* Block grant count for isoch requests when block count is
+ * set to a finite value.
+ */
+#define   MI_ARB_BLOCK_GRANT_MASK		(3 << 12)
+#define   MI_ARB_BLOCK_GRANT_8			(0 << 12)	/* for 3 display planes */
+#define   MI_ARB_BLOCK_GRANT_4			(1 << 12)	/* for 2 display planes */
+#define   MI_ARB_BLOCK_GRANT_2			(2 << 12)	/* for 1 display plane */
+#define   MI_ARB_BLOCK_GRANT_0			(3 << 12)	/* don't use */
+
+/* Enable render writes to complete in C2/C3/C4 power states.
+ * If this isn't enabled, render writes are prevented in low
+ * power states. That seems bad to me.
+ */
+#define   MI_ARB_C3_LP_WRITE_ENABLE		(1 << 11)
+
+/* This acknowledges an async flip immediately instead
+ * of waiting for 2TLB fetches.
+ */
+#define   MI_ARB_ASYNC_FLIP_ACK_IMMEDIATE	(1 << 10)
+
+/* Enables non-sequential data reads through arbiter
+ */
+#define   MI_ARB_DUAL_DATA_PHASE_DISABLE	(1 << 9)
+
+/* Disable FSB snooping of cacheable write cycles from binner/render
+ * command stream
+ */
+#define   MI_ARB_CACHE_SNOOP_DISABLE		(1 << 8)
+
+/* Arbiter time slice for non-isoch streams */
+#define   MI_ARB_TIME_SLICE_MASK		(7 << 5)
+#define   MI_ARB_TIME_SLICE_1			(0 << 5)
+#define   MI_ARB_TIME_SLICE_2			(1 << 5)
+#define   MI_ARB_TIME_SLICE_4			(2 << 5)
+#define   MI_ARB_TIME_SLICE_6			(3 << 5)
+#define   MI_ARB_TIME_SLICE_8			(4 << 5)
+#define   MI_ARB_TIME_SLICE_10			(5 << 5)
+#define   MI_ARB_TIME_SLICE_14			(6 << 5)
+#define   MI_ARB_TIME_SLICE_16			(7 << 5)
+
+/* Low priority grace period page size */
+#define   MI_ARB_LOW_PRIORITY_GRACE_4KB		(0 << 4)	/* default */
+#define   MI_ARB_LOW_PRIORITY_GRACE_8KB		(1 << 4)
+
+/* Disable display A/B trickle feed */
+#define   MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE	(1 << 2)
+
+/* Set display plane priority */
+#define   MI_ARB_DISPLAY_PRIORITY_A_B		(0 << 0)	/* display A > display B */
+#define   MI_ARB_DISPLAY_PRIORITY_B_A		(1 << 0)	/* display B > display A */
+
+#define MI_STATE	_MMIO(0x20e4) /* gen2 only */
+#define   MI_AGPBUSY_INT_EN			(1 << 1) /* 85x only */
+#define   MI_AGPBUSY_830_MODE			(1 << 0) /* 85x only */
+
+/* On modern GEN architectures interrupt control consists of two sets
+ * of registers. The first set pertains to the ring generating the
+ * interrupt. The second control is for the functional block generating the
+ * interrupt. These are PM, GT, DE, etc.
+ *
+ * Luckily *knocks on wood* all the ring interrupt bits match up with the
+ * GT interrupt bits, so we don't need to duplicate the defines.
+ *
+ * These defines should cover us well from SNB->HSW with minor exceptions
+ * it can also work on ILK.
+ */
+#define GT_BLT_FLUSHDW_NOTIFY_INTERRUPT		(1 << 26)
+#define GT_BLT_CS_ERROR_INTERRUPT		(1 << 25)
+#define GT_BLT_USER_INTERRUPT			(1 << 22)
+#define GT_BSD_CS_ERROR_INTERRUPT		(1 << 15)
+#define GT_BSD_USER_INTERRUPT			(1 << 12)
+#define GT_RENDER_L3_PARITY_ERROR_INTERRUPT_S1	(1 << 11) /* hsw+; rsvd on snb, ivb, vlv */
+#define GT_WAIT_SEMAPHORE_INTERRUPT		REG_BIT(11) /* bdw+ */
+#define GT_CONTEXT_SWITCH_INTERRUPT		(1 <<  8)
+#define GT_RENDER_L3_PARITY_ERROR_INTERRUPT	(1 <<  5) /* !snb */
+#define GT_RENDER_PIPECTL_NOTIFY_INTERRUPT	(1 <<  4)
+#define GT_CS_MASTER_ERROR_INTERRUPT		REG_BIT(3)
+#define GT_RENDER_SYNC_STATUS_INTERRUPT		(1 <<  2)
+#define GT_RENDER_DEBUG_INTERRUPT		(1 <<  1)
+#define GT_RENDER_USER_INTERRUPT		(1 <<  0)
+
+#define PM_VEBOX_CS_ERROR_INTERRUPT		(1 << 12) /* hsw+ */
+#define PM_VEBOX_USER_INTERRUPT			(1 << 10) /* hsw+ */
+
+#define GT_PARITY_ERROR(dev_priv) \
+	(GT_RENDER_L3_PARITY_ERROR_INTERRUPT | \
+	 (IS_HASWELL(dev_priv) ? GT_RENDER_L3_PARITY_ERROR_INTERRUPT_S1 : 0))
+
+/* These are all the "old" interrupts */
+#define ILK_BSD_USER_INTERRUPT				(1 << 5)
+
+#define I915_PM_INTERRUPT				(1 << 31)
+#define I915_ISP_INTERRUPT				(1 << 22)
+#define I915_LPE_PIPE_B_INTERRUPT			(1 << 21)
+#define I915_LPE_PIPE_A_INTERRUPT			(1 << 20)
+#define I915_MIPIC_INTERRUPT				(1 << 19)
+#define I915_MIPIA_INTERRUPT				(1 << 18)
+#define I915_PIPE_CONTROL_NOTIFY_INTERRUPT		(1 << 18)
+#define I915_DISPLAY_PORT_INTERRUPT			(1 << 17)
+#define I915_DISPLAY_PIPE_C_HBLANK_INTERRUPT		(1 << 16)
+#define I915_MASTER_ERROR_INTERRUPT			(1 << 15)
+#define I915_DISPLAY_PIPE_B_HBLANK_INTERRUPT		(1 << 14)
+#define I915_GMCH_THERMAL_SENSOR_EVENT_INTERRUPT	(1 << 14) /* p-state */
+#define I915_DISPLAY_PIPE_A_HBLANK_INTERRUPT		(1 << 13)
+#define I915_HWB_OOM_INTERRUPT				(1 << 13)
+#define I915_LPE_PIPE_C_INTERRUPT			(1 << 12)
+#define I915_SYNC_STATUS_INTERRUPT			(1 << 12)
+#define I915_MISC_INTERRUPT				(1 << 11)
+#define I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT	(1 << 11)
+#define I915_DISPLAY_PIPE_C_VBLANK_INTERRUPT		(1 << 10)
+#define I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT	(1 << 10)
+#define I915_DISPLAY_PIPE_C_EVENT_INTERRUPT		(1 << 9)
+#define I915_OVERLAY_PLANE_FLIP_PENDING_INTERRUPT	(1 << 9)
+#define I915_DISPLAY_PIPE_C_DPBM_INTERRUPT		(1 << 8)
+#define I915_DISPLAY_PLANE_C_FLIP_PENDING_INTERRUPT	(1 << 8)
+#define I915_DISPLAY_PIPE_A_VBLANK_INTERRUPT		(1 << 7)
+#define I915_DISPLAY_PIPE_A_EVENT_INTERRUPT		(1 << 6)
+#define I915_DISPLAY_PIPE_B_VBLANK_INTERRUPT		(1 << 5)
+#define I915_DISPLAY_PIPE_B_EVENT_INTERRUPT		(1 << 4)
+#define I915_DISPLAY_PIPE_A_DPBM_INTERRUPT		(1 << 3)
+#define I915_DISPLAY_PIPE_B_DPBM_INTERRUPT		(1 << 2)
+#define I915_DEBUG_INTERRUPT				(1 << 2)
+#define I915_WINVALID_INTERRUPT				(1 << 1)
+#define I915_USER_INTERRUPT				(1 << 1)
+#define I915_ASLE_INTERRUPT				(1 << 0)
+#define I915_BSD_USER_INTERRUPT				(1 << 25)
+
+#define I915_HDMI_LPE_AUDIO_BASE	(VLV_DISPLAY_BASE + 0x65000)
+#define I915_HDMI_LPE_AUDIO_SIZE	0x1000
+
+/* DisplayPort Audio w/ LPE */
+#define VLV_AUD_CHICKEN_BIT_REG		_MMIO(VLV_DISPLAY_BASE + 0x62F38)
+#define VLV_CHICKEN_BIT_DBG_ENABLE	(1 << 0)
+
+#define _VLV_AUD_PORT_EN_B_DBG		(VLV_DISPLAY_BASE + 0x62F20)
+#define _VLV_AUD_PORT_EN_C_DBG		(VLV_DISPLAY_BASE + 0x62F30)
+#define _VLV_AUD_PORT_EN_D_DBG		(VLV_DISPLAY_BASE + 0x62F34)
+#define VLV_AUD_PORT_EN_DBG(port)	_MMIO_PORT3((port) - PORT_B,	   \
+						    _VLV_AUD_PORT_EN_B_DBG, \
+						    _VLV_AUD_PORT_EN_C_DBG, \
+						    _VLV_AUD_PORT_EN_D_DBG)
+#define VLV_AMP_MUTE		        (1 << 1)
+
+#define GEN6_BSD_RNCID			_MMIO(0x12198)
+
+#define GEN7_FF_THREAD_MODE		_MMIO(0x20a0)
+#define   GEN7_FF_SCHED_MASK		0x0077070
+#define   GEN8_FF_DS_REF_CNT_FFME	(1 << 19)
+#define   GEN12_FF_TESSELATION_DOP_GATE_DISABLE BIT(19)
+#define   GEN7_FF_TS_SCHED_HS1		(0x5 << 16)
+#define   GEN7_FF_TS_SCHED_HS0		(0x3 << 16)
+#define   GEN7_FF_TS_SCHED_LOAD_BALANCE	(0x1 << 16)
+#define   GEN7_FF_TS_SCHED_HW		(0x0 << 16) /* Default */
+#define   GEN7_FF_VS_REF_CNT_FFME	(1 << 15)
+#define   GEN7_FF_VS_SCHED_HS1		(0x5 << 12)
+#define   GEN7_FF_VS_SCHED_HS0		(0x3 << 12)
+#define   GEN7_FF_VS_SCHED_LOAD_BALANCE	(0x1 << 12) /* Default */
+#define   GEN7_FF_VS_SCHED_HW		(0x0 << 12)
+#define   GEN7_FF_DS_SCHED_HS1		(0x5 << 4)
+#define   GEN7_FF_DS_SCHED_HS0		(0x3 << 4)
+#define   GEN7_FF_DS_SCHED_LOAD_BALANCE	(0x1 << 4)  /* Default */
+#define   GEN7_FF_DS_SCHED_HW		(0x0 << 4)
+
+/*
+ * Framebuffer compression (915+ only)
+ */
+
+#define FBC_CFB_BASE		_MMIO(0x3200) /* 4k page aligned */
+#define FBC_LL_BASE		_MMIO(0x3204) /* 4k page aligned */
+#define FBC_CONTROL		_MMIO(0x3208)
+#define   FBC_CTL_EN			REG_BIT(31)
+#define   FBC_CTL_PERIODIC		REG_BIT(30)
+#define   FBC_CTL_INTERVAL_MASK		REG_GENMASK(29, 16)
+#define   FBC_CTL_INTERVAL(x)		REG_FIELD_PREP(FBC_CTL_INTERVAL_MASK, (x))
+#define   FBC_CTL_STOP_ON_MOD		REG_BIT(15)
+#define   FBC_CTL_UNCOMPRESSIBLE	REG_BIT(14) /* i915+ */
+#define   FBC_CTL_C3_IDLE		REG_BIT(13) /* i945gm only */
+#define   FBC_CTL_STRIDE_MASK		REG_GENMASK(12, 5)
+#define   FBC_CTL_STRIDE(x)		REG_FIELD_PREP(FBC_CTL_STRIDE_MASK, (x))
+#define   FBC_CTL_FENCENO_MASK		REG_GENMASK(3, 0)
+#define   FBC_CTL_FENCENO(x)		REG_FIELD_PREP(FBC_CTL_FENCENO_MASK, (x))
+#define FBC_COMMAND		_MMIO(0x320c)
+#define   FBC_CMD_COMPRESS		REG_BIT(0)
+#define FBC_STATUS		_MMIO(0x3210)
+#define   FBC_STAT_COMPRESSING		REG_BIT(31)
+#define   FBC_STAT_COMPRESSED		REG_BIT(30)
+#define   FBC_STAT_MODIFIED		REG_BIT(29)
+#define   FBC_STAT_CURRENT_LINE_MASK	REG_GENMASK(10, 0)
+#define FBC_CONTROL2		_MMIO(0x3214) /* i965gm only */
+#define   FBC_CTL_FENCE_DBL		REG_BIT(4)
+#define   FBC_CTL_IDLE_MASK		REG_GENMASK(3, 2)
+#define   FBC_CTL_IDLE_IMM		REG_FIELD_PREP(FBC_CTL_IDLE_MASK, 0)
+#define   FBC_CTL_IDLE_FULL		REG_FIELD_PREP(FBC_CTL_IDLE_MASK, 1)
+#define   FBC_CTL_IDLE_LINE		REG_FIELD_PREP(FBC_CTL_IDLE_MASK, 2)
+#define   FBC_CTL_IDLE_DEBUG		REG_FIELD_PREP(FBC_CTL_IDLE_MASK, 3)
+#define   FBC_CTL_CPU_FENCE_EN		REG_BIT(1)
+#define   FBC_CTL_PLANE_MASK		REG_GENMASK(1, 0)
+#define   FBC_CTL_PLANE(i9xx_plane)	REG_FIELD_PREP(FBC_CTL_PLANE_MASK, (i9xx_plane))
+#define FBC_FENCE_OFF		_MMIO(0x3218)  /* i965gm only, BSpec typo has 321Bh */
+#define FBC_MOD_NUM		_MMIO(0x3220)  /* i965gm only */
+#define   FBC_MOD_NUM_MASK		REG_GENMASK(31, 1)
+#define   FBC_MOD_NUM_VALID		REG_BIT(0)
+#define FBC_TAG(i)		_MMIO(0x3300 + (i) * 4) /* 49 reisters */
+#define   FBC_TAG_MASK			REG_GENMASK(1, 0) /* 16 tags per register */
+#define   FBC_TAG_MODIFIED		REG_FIELD_PREP(FBC_TAG_MASK, 0)
+#define   FBC_TAG_UNCOMPRESSED		REG_FIELD_PREP(FBC_TAG_MASK, 1)
+#define   FBC_TAG_UNCOMPRESSIBLE	REG_FIELD_PREP(FBC_TAG_MASK, 2)
+#define   FBC_TAG_COMPRESSED		REG_FIELD_PREP(FBC_TAG_MASK, 3)
+
+#define FBC_LL_SIZE		(1536)
+
+/* Framebuffer compression for GM45+ */
+#define DPFC_CB_BASE			_MMIO(0x3200)
+#define ILK_DPFC_CB_BASE(fbc_id)	_MMIO_PIPE((fbc_id), 0x43200, 0x43240)
+#define DPFC_CONTROL			_MMIO(0x3208)
+#define ILK_DPFC_CONTROL(fbc_id)	_MMIO_PIPE((fbc_id), 0x43208, 0x43248)
+#define   DPFC_CTL_EN				REG_BIT(31)
+#define   DPFC_CTL_PLANE_MASK_G4X		REG_BIT(30) /* g4x-snb */
+#define   DPFC_CTL_PLANE_G4X(i9xx_plane)	REG_FIELD_PREP(DPFC_CTL_PLANE_MASK_G4X, (i9xx_plane))
+#define   DPFC_CTL_FENCE_EN_G4X			REG_BIT(29) /* g4x-snb */
+#define   DPFC_CTL_PLANE_MASK_IVB		REG_GENMASK(30, 29) /* ivb only */
+#define   DPFC_CTL_PLANE_IVB(i9xx_plane)	REG_FIELD_PREP(DPFC_CTL_PLANE_MASK_IVB, (i9xx_plane))
+#define   DPFC_CTL_FENCE_EN_IVB			REG_BIT(28) /* ivb+ */
+#define   DPFC_CTL_PERSISTENT_MODE		REG_BIT(25) /* g4x-snb */
+#define   DPFC_CTL_FALSE_COLOR			REG_BIT(10) /* ivb+ */
+#define   DPFC_CTL_SR_EN			REG_BIT(10) /* g4x only */
+#define   DPFC_CTL_SR_EXIT_DIS			REG_BIT(9) /* g4x only */
+#define   DPFC_CTL_LIMIT_MASK			REG_GENMASK(7, 6)
+#define   DPFC_CTL_LIMIT_1X			REG_FIELD_PREP(DPFC_CTL_LIMIT_MASK, 0)
+#define   DPFC_CTL_LIMIT_2X			REG_FIELD_PREP(DPFC_CTL_LIMIT_MASK, 1)
+#define   DPFC_CTL_LIMIT_4X			REG_FIELD_PREP(DPFC_CTL_LIMIT_MASK, 2)
+#define   DPFC_CTL_FENCENO_MASK			REG_GENMASK(3, 0)
+#define   DPFC_CTL_FENCENO(fence)		REG_FIELD_PREP(DPFC_CTL_FENCENO_MASK, (fence))
+#define DPFC_RECOMP_CTL			_MMIO(0x320c)
+#define ILK_DPFC_RECOMP_CTL(fbc_id)	_MMIO_PIPE((fbc_id), 0x4320c, 0x4324c)
+#define   DPFC_RECOMP_STALL_EN			REG_BIT(27)
+#define   DPFC_RECOMP_STALL_WM_MASK		REG_GENMASK(26, 16)
+#define   DPFC_RECOMP_TIMER_COUNT_MASK		REG_GENMASK(5, 0)
+#define DPFC_STATUS			_MMIO(0x3210)
+#define ILK_DPFC_STATUS(fbc_id)		_MMIO_PIPE((fbc_id), 0x43210, 0x43250)
+#define   DPFC_INVAL_SEG_MASK			REG_GENMASK(26, 16)
+#define   DPFC_COMP_SEG_MASK			REG_GENMASK(10, 0)
+#define DPFC_STATUS2			_MMIO(0x3214)
+#define ILK_DPFC_STATUS2(fbc_id)	_MMIO_PIPE((fbc_id), 0x43214, 0x43254)
+#define   DPFC_COMP_SEG_MASK_IVB		REG_GENMASK(11, 0)
+#define DPFC_FENCE_YOFF			_MMIO(0x3218)
+#define ILK_DPFC_FENCE_YOFF(fbc_id)	_MMIO_PIPE((fbc_id), 0x43218, 0x43258)
+#define DPFC_CHICKEN			_MMIO(0x3224)
+#define ILK_DPFC_CHICKEN(fbc_id)	_MMIO_PIPE((fbc_id), 0x43224, 0x43264)
+#define   DPFC_HT_MODIFY			REG_BIT(31) /* pre-ivb */
+#define   DPFC_NUKE_ON_ANY_MODIFICATION		REG_BIT(23) /* bdw+ */
+#define   DPFC_CHICKEN_COMP_DUMMY_PIXEL		REG_BIT(14) /* glk+ */
+#define   DPFC_CHICKEN_FORCE_SLB_INVALIDATION	REG_BIT(13) /* icl+ */
+#define   DPFC_DISABLE_DUMMY0			REG_BIT(8) /* ivb+ */
+
+#define GLK_FBC_STRIDE(fbc_id)	_MMIO_PIPE((fbc_id), 0x43228, 0x43268)
+#define   FBC_STRIDE_OVERRIDE	REG_BIT(15)
+#define   FBC_STRIDE_MASK	REG_GENMASK(14, 0)
+#define   FBC_STRIDE(x)		REG_FIELD_PREP(FBC_STRIDE_MASK, (x))
+
+#define ILK_FBC_RT_BASE		_MMIO(0x2128)
+#define   ILK_FBC_RT_VALID	REG_BIT(0)
+#define   SNB_FBC_FRONT_BUFFER	REG_BIT(1)
+
+#define ILK_DISPLAY_CHICKEN1	_MMIO(0x42000)
+#define   ILK_FBCQ_DIS		(1 << 22)
+#define   ILK_PABSTRETCH_DIS	REG_BIT(21)
+#define   ILK_SABSTRETCH_DIS	REG_BIT(20)
+#define   IVB_PRI_STRETCH_MAX_MASK	REG_GENMASK(21, 20)
+#define   IVB_PRI_STRETCH_MAX_X8	REG_FIELD_PREP(IVB_PRI_STRETCH_MAX_MASK, 0)
+#define   IVB_PRI_STRETCH_MAX_X4	REG_FIELD_PREP(IVB_PRI_STRETCH_MAX_MASK, 1)
+#define   IVB_PRI_STRETCH_MAX_X2	REG_FIELD_PREP(IVB_PRI_STRETCH_MAX_MASK, 2)
+#define   IVB_PRI_STRETCH_MAX_X1	REG_FIELD_PREP(IVB_PRI_STRETCH_MAX_MASK, 3)
+#define   IVB_SPR_STRETCH_MAX_MASK	REG_GENMASK(19, 18)
+#define   IVB_SPR_STRETCH_MAX_X8	REG_FIELD_PREP(IVB_SPR_STRETCH_MAX_MASK, 0)
+#define   IVB_SPR_STRETCH_MAX_X4	REG_FIELD_PREP(IVB_SPR_STRETCH_MAX_MASK, 1)
+#define   IVB_SPR_STRETCH_MAX_X2	REG_FIELD_PREP(IVB_SPR_STRETCH_MAX_MASK, 2)
+#define   IVB_SPR_STRETCH_MAX_X1	REG_FIELD_PREP(IVB_SPR_STRETCH_MAX_MASK, 3)
+
+
+/*
+ * Framebuffer compression for Sandybridge
+ *
+ * The following two registers are of type GTTMMADR
+ */
+#define SNB_DPFC_CTL_SA		_MMIO(0x100100)
+#define   SNB_DPFC_FENCE_EN		REG_BIT(29)
+#define   SNB_DPFC_FENCENO_MASK		REG_GENMASK(4, 0)
+#define   SNB_DPFC_FENCENO(fence)	REG_FIELD_PREP(SNB_DPFC_FENCENO_MASK, (fence))
+#define SNB_DPFC_CPU_FENCE_OFFSET	_MMIO(0x100104)
+
+/* Framebuffer compression for Ivybridge */
+#define IVB_FBC_RT_BASE			_MMIO(0x7020)
+#define IVB_FBC_RT_BASE_UPPER		_MMIO(0x7024)
+
+#define IPS_CTL		_MMIO(0x43408)
+#define   IPS_ENABLE	(1 << 31)
+
+#define MSG_FBC_REND_STATE(fbc_id)	_MMIO_PIPE((fbc_id), 0x50380, 0x50384)
+#define   FBC_REND_NUKE			REG_BIT(2)
+#define   FBC_REND_CACHE_CLEAN		REG_BIT(1)
+
+/*
+ * Clock control & power management
+ */
+#define _DPLL_A (DISPLAY_MMIO_BASE(dev_priv) + 0x6014)
+#define _DPLL_B (DISPLAY_MMIO_BASE(dev_priv) + 0x6018)
+#define _CHV_DPLL_C (DISPLAY_MMIO_BASE(dev_priv) + 0x6030)
+#define DPLL(pipe) _MMIO_PIPE3((pipe), _DPLL_A, _DPLL_B, _CHV_DPLL_C)
+
+#define VGA0	_MMIO(0x6000)
+#define VGA1	_MMIO(0x6004)
+#define VGA_PD	_MMIO(0x6010)
+#define   VGA0_PD_P2_DIV_4	(1 << 7)
+#define   VGA0_PD_P1_DIV_2	(1 << 5)
+#define   VGA0_PD_P1_SHIFT	0
+#define   VGA0_PD_P1_MASK	(0x1f << 0)
+#define   VGA1_PD_P2_DIV_4	(1 << 15)
+#define   VGA1_PD_P1_DIV_2	(1 << 13)
+#define   VGA1_PD_P1_SHIFT	8
+#define   VGA1_PD_P1_MASK	(0x1f << 8)
+#define   DPLL_VCO_ENABLE		(1 << 31)
+#define   DPLL_SDVO_HIGH_SPEED		(1 << 30)
+#define   DPLL_DVO_2X_MODE		(1 << 30)
+#define   DPLL_EXT_BUFFER_ENABLE_VLV	(1 << 30)
+#define   DPLL_SYNCLOCK_ENABLE		(1 << 29)
+#define   DPLL_REF_CLK_ENABLE_VLV	(1 << 29)
+#define   DPLL_VGA_MODE_DIS		(1 << 28)
+#define   DPLLB_MODE_DAC_SERIAL		(1 << 26) /* i915 */
+#define   DPLLB_MODE_LVDS		(2 << 26) /* i915 */
+#define   DPLL_MODE_MASK		(3 << 26)
+#define   DPLL_DAC_SERIAL_P2_CLOCK_DIV_10 (0 << 24) /* i915 */
+#define   DPLL_DAC_SERIAL_P2_CLOCK_DIV_5 (1 << 24) /* i915 */
+#define   DPLLB_LVDS_P2_CLOCK_DIV_14	(0 << 24) /* i915 */
+#define   DPLLB_LVDS_P2_CLOCK_DIV_7	(1 << 24) /* i915 */
+#define   DPLL_P2_CLOCK_DIV_MASK	0x03000000 /* i915 */
+#define   DPLL_FPA01_P1_POST_DIV_MASK	0x00ff0000 /* i915 */
+#define   DPLL_FPA01_P1_POST_DIV_MASK_PINEVIEW	0x00ff8000 /* Pineview */
+#define   DPLL_LOCK_VLV			(1 << 15)
+#define   DPLL_INTEGRATED_CRI_CLK_VLV	(1 << 14)
+#define   DPLL_INTEGRATED_REF_CLK_VLV	(1 << 13)
+#define   DPLL_SSC_REF_CLK_CHV		(1 << 13)
+#define   DPLL_PORTC_READY_MASK		(0xf << 4)
+#define   DPLL_PORTB_READY_MASK		(0xf)
+
+#define   DPLL_FPA01_P1_POST_DIV_MASK_I830	0x001f0000
+
+/* Additional CHV pll/phy registers */
+#define DPIO_PHY_STATUS			_MMIO(VLV_DISPLAY_BASE + 0x6240)
+#define   DPLL_PORTD_READY_MASK		(0xf)
+#define DISPLAY_PHY_CONTROL _MMIO(VLV_DISPLAY_BASE + 0x60100)
+#define   PHY_CH_POWER_DOWN_OVRD_EN(phy, ch)	(1 << (2 * (phy) + (ch) + 27))
+#define   PHY_LDO_DELAY_0NS			0x0
+#define   PHY_LDO_DELAY_200NS			0x1
+#define   PHY_LDO_DELAY_600NS			0x2
+#define   PHY_LDO_SEQ_DELAY(delay, phy)		((delay) << (2 * (phy) + 23))
+#define   PHY_CH_POWER_DOWN_OVRD(mask, phy, ch)	((mask) << (8 * (phy) + 4 * (ch) + 11))
+#define   PHY_CH_SU_PSR				0x1
+#define   PHY_CH_DEEP_PSR			0x7
+#define   PHY_CH_POWER_MODE(mode, phy, ch)	((mode) << (6 * (phy) + 3 * (ch) + 2))
+#define   PHY_COM_LANE_RESET_DEASSERT(phy)	(1 << (phy))
+#define DISPLAY_PHY_STATUS _MMIO(VLV_DISPLAY_BASE + 0x60104)
+#define   PHY_POWERGOOD(phy)	(((phy) == DPIO_PHY0) ? (1 << 31) : (1 << 30))
+#define   PHY_STATUS_CMN_LDO(phy, ch)                   (1 << (6 - (6 * (phy) + 3 * (ch))))
+#define   PHY_STATUS_SPLINE_LDO(phy, ch, spline)        (1 << (8 - (6 * (phy) + 3 * (ch) + (spline))))
+
+/*
+ * The i830 generation, in LVDS mode, defines P1 as the bit number set within
+ * this field (only one bit may be set).
+ */
+#define   DPLL_FPA01_P1_POST_DIV_MASK_I830_LVDS	0x003f0000
+#define   DPLL_FPA01_P1_POST_DIV_SHIFT	16
+#define   DPLL_FPA01_P1_POST_DIV_SHIFT_PINEVIEW 15
+/* i830, required in DVO non-gang */
+#define   PLL_P2_DIVIDE_BY_4		(1 << 23)
+#define   PLL_P1_DIVIDE_BY_TWO		(1 << 21) /* i830 */
+#define   PLL_REF_INPUT_DREFCLK		(0 << 13)
+#define   PLL_REF_INPUT_TVCLKINA	(1 << 13) /* i830 */
+#define   PLL_REF_INPUT_TVCLKINBC	(2 << 13) /* SDVO TVCLKIN */
+#define   PLLB_REF_INPUT_SPREADSPECTRUMIN (3 << 13)
+#define   PLL_REF_INPUT_MASK		(3 << 13)
+#define   PLL_LOAD_PULSE_PHASE_SHIFT		9
+/* Ironlake */
+# define PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT     9
+# define PLL_REF_SDVO_HDMI_MULTIPLIER_MASK      (7 << 9)
+# define PLL_REF_SDVO_HDMI_MULTIPLIER(x)	(((x) - 1) << 9)
+# define DPLL_FPA1_P1_POST_DIV_SHIFT            0
+# define DPLL_FPA1_P1_POST_DIV_MASK             0xff
+
+/*
+ * Parallel to Serial Load Pulse phase selection.
+ * Selects the phase for the 10X DPLL clock for the PCIe
+ * digital display port. The range is 4 to 13; 10 or more
+ * is just a flip delay. The default is 6
+ */
+#define   PLL_LOAD_PULSE_PHASE_MASK		(0xf << PLL_LOAD_PULSE_PHASE_SHIFT)
+#define   DISPLAY_RATE_SELECT_FPA1		(1 << 8)
+/*
+ * SDVO multiplier for 945G/GM. Not used on 965.
+ */
+#define   SDVO_MULTIPLIER_MASK			0x000000ff
+#define   SDVO_MULTIPLIER_SHIFT_HIRES		4
+#define   SDVO_MULTIPLIER_SHIFT_VGA		0
+
+#define _DPLL_A_MD (DISPLAY_MMIO_BASE(dev_priv) + 0x601c)
+#define _DPLL_B_MD (DISPLAY_MMIO_BASE(dev_priv) + 0x6020)
+#define _CHV_DPLL_C_MD (DISPLAY_MMIO_BASE(dev_priv) + 0x603c)
+#define DPLL_MD(pipe) _MMIO_PIPE3((pipe), _DPLL_A_MD, _DPLL_B_MD, _CHV_DPLL_C_MD)
+
+/*
+ * UDI pixel divider, controlling how many pixels are stuffed into a packet.
+ *
+ * Value is pixels minus 1.  Must be set to 1 pixel for SDVO.
+ */
+#define   DPLL_MD_UDI_DIVIDER_MASK		0x3f000000
+#define   DPLL_MD_UDI_DIVIDER_SHIFT		24
+/* UDI pixel divider for VGA, same as DPLL_MD_UDI_DIVIDER_MASK. */
+#define   DPLL_MD_VGA_UDI_DIVIDER_MASK		0x003f0000
+#define   DPLL_MD_VGA_UDI_DIVIDER_SHIFT		16
+/*
+ * SDVO/UDI pixel multiplier.
+ *
+ * SDVO requires that the bus clock rate be between 1 and 2 Ghz, and the bus
+ * clock rate is 10 times the DPLL clock.  At low resolution/refresh rate
+ * modes, the bus rate would be below the limits, so SDVO allows for stuffing
+ * dummy bytes in the datastream at an increased clock rate, with both sides of
+ * the link knowing how many bytes are fill.
+ *
+ * So, for a mode with a dotclock of 65Mhz, we would want to double the clock
+ * rate to 130Mhz to get a bus rate of 1.30Ghz.  The DPLL clock rate would be
+ * set to 130Mhz, and the SDVO multiplier set to 2x in this register and
+ * through an SDVO command.
+ *
+ * This register field has values of multiplication factor minus 1, with
+ * a maximum multiplier of 5 for SDVO.
+ */
+#define   DPLL_MD_UDI_MULTIPLIER_MASK		0x00003f00
+#define   DPLL_MD_UDI_MULTIPLIER_SHIFT		8
+/*
+ * SDVO/UDI pixel multiplier for VGA, same as DPLL_MD_UDI_MULTIPLIER_MASK.
+ * This best be set to the default value (3) or the CRT won't work. No,
+ * I don't entirely understand what this does...
+ */
+#define   DPLL_MD_VGA_UDI_MULTIPLIER_MASK	0x0000003f
+#define   DPLL_MD_VGA_UDI_MULTIPLIER_SHIFT	0
+
+#define RAWCLK_FREQ_VLV		_MMIO(VLV_DISPLAY_BASE + 0x6024)
+
+#define _FPA0	0x6040
+#define _FPA1	0x6044
+#define _FPB0	0x6048
+#define _FPB1	0x604c
+#define FP0(pipe) _MMIO_PIPE(pipe, _FPA0, _FPB0)
+#define FP1(pipe) _MMIO_PIPE(pipe, _FPA1, _FPB1)
+#define   FP_N_DIV_MASK		0x003f0000
+#define   FP_N_PINEVIEW_DIV_MASK	0x00ff0000
+#define   FP_N_DIV_SHIFT		16
+#define   FP_M1_DIV_MASK	0x00003f00
+#define   FP_M1_DIV_SHIFT		 8
+#define   FP_M2_DIV_MASK	0x0000003f
+#define   FP_M2_PINEVIEW_DIV_MASK	0x000000ff
+#define   FP_M2_DIV_SHIFT		 0
+#define DPLL_TEST	_MMIO(0x606c)
+#define   DPLLB_TEST_SDVO_DIV_1		(0 << 22)
+#define   DPLLB_TEST_SDVO_DIV_2		(1 << 22)
+#define   DPLLB_TEST_SDVO_DIV_4		(2 << 22)
+#define   DPLLB_TEST_SDVO_DIV_MASK	(3 << 22)
+#define   DPLLB_TEST_N_BYPASS		(1 << 19)
+#define   DPLLB_TEST_M_BYPASS		(1 << 18)
+#define   DPLLB_INPUT_BUFFER_ENABLE	(1 << 16)
+#define   DPLLA_TEST_N_BYPASS		(1 << 3)
+#define   DPLLA_TEST_M_BYPASS		(1 << 2)
+#define   DPLLA_INPUT_BUFFER_ENABLE	(1 << 0)
+#define D_STATE		_MMIO(0x6104)
+#define  DSTATE_GFX_RESET_I830			(1 << 6)
+#define  DSTATE_PLL_D3_OFF			(1 << 3)
+#define  DSTATE_GFX_CLOCK_GATING		(1 << 1)
+#define  DSTATE_DOT_CLOCK_GATING		(1 << 0)
+#define DSPCLK_GATE_D(__i915)		_MMIO(DISPLAY_MMIO_BASE(__i915) + 0x6200)
+# define DPUNIT_B_CLOCK_GATE_DISABLE		(1 << 30) /* 965 */
+# define VSUNIT_CLOCK_GATE_DISABLE		(1 << 29) /* 965 */
+# define VRHUNIT_CLOCK_GATE_DISABLE		(1 << 28) /* 965 */
+# define VRDUNIT_CLOCK_GATE_DISABLE		(1 << 27) /* 965 */
+# define AUDUNIT_CLOCK_GATE_DISABLE		(1 << 26) /* 965 */
+# define DPUNIT_A_CLOCK_GATE_DISABLE		(1 << 25) /* 965 */
+# define DPCUNIT_CLOCK_GATE_DISABLE		(1 << 24) /* 965 */
+# define PNV_GMBUSUNIT_CLOCK_GATE_DISABLE	(1 << 24) /* pnv */
+# define TVRUNIT_CLOCK_GATE_DISABLE		(1 << 23) /* 915-945 */
+# define TVCUNIT_CLOCK_GATE_DISABLE		(1 << 22) /* 915-945 */
+# define TVFUNIT_CLOCK_GATE_DISABLE		(1 << 21) /* 915-945 */
+# define TVEUNIT_CLOCK_GATE_DISABLE		(1 << 20) /* 915-945 */
+# define DVSUNIT_CLOCK_GATE_DISABLE		(1 << 19) /* 915-945 */
+# define DSSUNIT_CLOCK_GATE_DISABLE		(1 << 18) /* 915-945 */
+# define DDBUNIT_CLOCK_GATE_DISABLE		(1 << 17) /* 915-945 */
+# define DPRUNIT_CLOCK_GATE_DISABLE		(1 << 16) /* 915-945 */
+# define DPFUNIT_CLOCK_GATE_DISABLE		(1 << 15) /* 915-945 */
+# define DPBMUNIT_CLOCK_GATE_DISABLE		(1 << 14) /* 915-945 */
+# define DPLSUNIT_CLOCK_GATE_DISABLE		(1 << 13) /* 915-945 */
+# define DPLUNIT_CLOCK_GATE_DISABLE		(1 << 12) /* 915-945 */
+# define DPOUNIT_CLOCK_GATE_DISABLE		(1 << 11)
+# define DPBUNIT_CLOCK_GATE_DISABLE		(1 << 10)
+# define DCUNIT_CLOCK_GATE_DISABLE		(1 << 9)
+# define DPUNIT_CLOCK_GATE_DISABLE		(1 << 8)
+# define VRUNIT_CLOCK_GATE_DISABLE		(1 << 7) /* 915+: reserved */
+# define OVHUNIT_CLOCK_GATE_DISABLE		(1 << 6) /* 830-865 */
+# define DPIOUNIT_CLOCK_GATE_DISABLE		(1 << 6) /* 915-945 */
+# define OVFUNIT_CLOCK_GATE_DISABLE		(1 << 5)
+# define OVBUNIT_CLOCK_GATE_DISABLE		(1 << 4)
+/*
+ * This bit must be set on the 830 to prevent hangs when turning off the
+ * overlay scaler.
+ */
+# define OVRUNIT_CLOCK_GATE_DISABLE		(1 << 3)
+# define OVCUNIT_CLOCK_GATE_DISABLE		(1 << 2)
+# define OVUUNIT_CLOCK_GATE_DISABLE		(1 << 1)
+# define ZVUNIT_CLOCK_GATE_DISABLE		(1 << 0) /* 830 */
+# define OVLUNIT_CLOCK_GATE_DISABLE		(1 << 0) /* 845,865 */
+
+#define RENCLK_GATE_D1		_MMIO(0x6204)
+# define BLITTER_CLOCK_GATE_DISABLE		(1 << 13) /* 945GM only */
+# define MPEG_CLOCK_GATE_DISABLE		(1 << 12) /* 945GM only */
+# define PC_FE_CLOCK_GATE_DISABLE		(1 << 11)
+# define PC_BE_CLOCK_GATE_DISABLE		(1 << 10)
+# define WINDOWER_CLOCK_GATE_DISABLE		(1 << 9)
+# define INTERPOLATOR_CLOCK_GATE_DISABLE	(1 << 8)
+# define COLOR_CALCULATOR_CLOCK_GATE_DISABLE	(1 << 7)
+# define MOTION_COMP_CLOCK_GATE_DISABLE		(1 << 6)
+# define MAG_CLOCK_GATE_DISABLE			(1 << 5)
+/* This bit must be unset on 855,865 */
+# define MECI_CLOCK_GATE_DISABLE		(1 << 4)
+# define DCMP_CLOCK_GATE_DISABLE		(1 << 3)
+# define MEC_CLOCK_GATE_DISABLE			(1 << 2)
+# define MECO_CLOCK_GATE_DISABLE		(1 << 1)
+/* This bit must be set on 855,865. */
+# define SV_CLOCK_GATE_DISABLE			(1 << 0)
+# define I915_MPEG_CLOCK_GATE_DISABLE		(1 << 16)
+# define I915_VLD_IP_PR_CLOCK_GATE_DISABLE	(1 << 15)
+# define I915_MOTION_COMP_CLOCK_GATE_DISABLE	(1 << 14)
+# define I915_BD_BF_CLOCK_GATE_DISABLE		(1 << 13)
+# define I915_SF_SE_CLOCK_GATE_DISABLE		(1 << 12)
+# define I915_WM_CLOCK_GATE_DISABLE		(1 << 11)
+# define I915_IZ_CLOCK_GATE_DISABLE		(1 << 10)
+# define I915_PI_CLOCK_GATE_DISABLE		(1 << 9)
+# define I915_DI_CLOCK_GATE_DISABLE		(1 << 8)
+# define I915_SH_SV_CLOCK_GATE_DISABLE		(1 << 7)
+# define I915_PL_DG_QC_FT_CLOCK_GATE_DISABLE	(1 << 6)
+# define I915_SC_CLOCK_GATE_DISABLE		(1 << 5)
+# define I915_FL_CLOCK_GATE_DISABLE		(1 << 4)
+# define I915_DM_CLOCK_GATE_DISABLE		(1 << 3)
+# define I915_PS_CLOCK_GATE_DISABLE		(1 << 2)
+# define I915_CC_CLOCK_GATE_DISABLE		(1 << 1)
+# define I915_BY_CLOCK_GATE_DISABLE		(1 << 0)
+
+# define I965_RCZ_CLOCK_GATE_DISABLE		(1 << 30)
+/* This bit must always be set on 965G/965GM */
+# define I965_RCC_CLOCK_GATE_DISABLE		(1 << 29)
+# define I965_RCPB_CLOCK_GATE_DISABLE		(1 << 28)
+# define I965_DAP_CLOCK_GATE_DISABLE		(1 << 27)
+# define I965_ROC_CLOCK_GATE_DISABLE		(1 << 26)
+# define I965_GW_CLOCK_GATE_DISABLE		(1 << 25)
+# define I965_TD_CLOCK_GATE_DISABLE		(1 << 24)
+/* This bit must always be set on 965G */
+# define I965_ISC_CLOCK_GATE_DISABLE		(1 << 23)
+# define I965_IC_CLOCK_GATE_DISABLE		(1 << 22)
+# define I965_EU_CLOCK_GATE_DISABLE		(1 << 21)
+# define I965_IF_CLOCK_GATE_DISABLE		(1 << 20)
+# define I965_TC_CLOCK_GATE_DISABLE		(1 << 19)
+# define I965_SO_CLOCK_GATE_DISABLE		(1 << 17)
+# define I965_FBC_CLOCK_GATE_DISABLE		(1 << 16)
+# define I965_MARI_CLOCK_GATE_DISABLE		(1 << 15)
+# define I965_MASF_CLOCK_GATE_DISABLE		(1 << 14)
+# define I965_MAWB_CLOCK_GATE_DISABLE		(1 << 13)
+# define I965_EM_CLOCK_GATE_DISABLE		(1 << 12)
+# define I965_UC_CLOCK_GATE_DISABLE		(1 << 11)
+# define I965_SI_CLOCK_GATE_DISABLE		(1 << 6)
+# define I965_MT_CLOCK_GATE_DISABLE		(1 << 5)
+# define I965_PL_CLOCK_GATE_DISABLE		(1 << 4)
+# define I965_DG_CLOCK_GATE_DISABLE		(1 << 3)
+# define I965_QC_CLOCK_GATE_DISABLE		(1 << 2)
+# define I965_FT_CLOCK_GATE_DISABLE		(1 << 1)
+# define I965_DM_CLOCK_GATE_DISABLE		(1 << 0)
+
+#define RENCLK_GATE_D2		_MMIO(0x6208)
+#define VF_UNIT_CLOCK_GATE_DISABLE		(1 << 9)
+#define GS_UNIT_CLOCK_GATE_DISABLE		(1 << 7)
+#define CL_UNIT_CLOCK_GATE_DISABLE		(1 << 6)
+
+#define VDECCLK_GATE_D		_MMIO(0x620C)		/* g4x only */
+#define  VCP_UNIT_CLOCK_GATE_DISABLE		(1 << 4)
+
+#define RAMCLK_GATE_D		_MMIO(0x6210)		/* CRL only */
+#define DEUC			_MMIO(0x6214)          /* CRL only */
+
+#define FW_BLC_SELF_VLV		_MMIO(VLV_DISPLAY_BASE + 0x6500)
+#define  FW_CSPWRDWNEN		(1 << 15)
+
+#define MI_ARB_VLV		_MMIO(VLV_DISPLAY_BASE + 0x6504)
+
+#define CZCLK_CDCLK_FREQ_RATIO	_MMIO(VLV_DISPLAY_BASE + 0x6508)
+#define   CDCLK_FREQ_SHIFT	4
+#define   CDCLK_FREQ_MASK	(0x1f << CDCLK_FREQ_SHIFT)
+#define   CZCLK_FREQ_MASK	0xf
+
+#define GCI_CONTROL		_MMIO(VLV_DISPLAY_BASE + 0x650C)
+#define   PFI_CREDIT_63		(9 << 28)		/* chv only */
+#define   PFI_CREDIT_31		(8 << 28)		/* chv only */
+#define   PFI_CREDIT(x)		(((x) - 8) << 28)	/* 8-15 */
+#define   PFI_CREDIT_RESEND	(1 << 27)
+#define   VGA_FAST_MODE_DISABLE	(1 << 14)
+
+#define GMBUSFREQ_VLV		_MMIO(VLV_DISPLAY_BASE + 0x6510)
+
+/*
+ * Palette regs
+ */
+#define _PALETTE_A		0xa000
+#define _PALETTE_B		0xa800
+#define _CHV_PALETTE_C		0xc000
+#define PALETTE_RED_MASK        REG_GENMASK(23, 16)
+#define PALETTE_GREEN_MASK      REG_GENMASK(15, 8)
+#define PALETTE_BLUE_MASK       REG_GENMASK(7, 0)
+#define PALETTE(pipe, i)	_MMIO(DISPLAY_MMIO_BASE(dev_priv) + \
+				      _PICK((pipe), _PALETTE_A,		\
+					    _PALETTE_B, _CHV_PALETTE_C) + \
+				      (i) * 4)
+
+#define PEG_BAND_GAP_DATA	_MMIO(0x14d68)
+
+#define BXT_RP_STATE_CAP        _MMIO(0x138170)
+#define GEN9_RP_STATE_LIMITS	_MMIO(0x138148)
+#define XEHPSDV_RP_STATE_CAP	_MMIO(0x250014)
+#define PVC_RP_STATE_CAP	_MMIO(0x281014)
+
+#define GT0_PERF_LIMIT_REASONS		_MMIO(0x1381a8)
+#define   GT0_PERF_LIMIT_REASONS_MASK	0xde3
+#define   PROCHOT_MASK			REG_BIT(0)
+#define   THERMAL_LIMIT_MASK		REG_BIT(1)
+#define   RATL_MASK			REG_BIT(5)
+#define   VR_THERMALERT_MASK		REG_BIT(6)
+#define   VR_TDC_MASK			REG_BIT(7)
+#define   POWER_LIMIT_4_MASK		REG_BIT(8)
+#define   POWER_LIMIT_1_MASK		REG_BIT(10)
+#define   POWER_LIMIT_2_MASK		REG_BIT(11)
+
+#define CHV_CLK_CTL1			_MMIO(0x101100)
+#define VLV_CLK_CTL2			_MMIO(0x101104)
+#define   CLK_CTL2_CZCOUNT_30NS_SHIFT	28
+
+/*
+ * Overlay regs
+ */
+
+#define OVADD			_MMIO(0x30000)
+#define DOVSTA			_MMIO(0x30008)
+#define OC_BUF			(0x3 << 20)
+#define OGAMC5			_MMIO(0x30010)
+#define OGAMC4			_MMIO(0x30014)
+#define OGAMC3			_MMIO(0x30018)
+#define OGAMC2			_MMIO(0x3001c)
+#define OGAMC1			_MMIO(0x30020)
+#define OGAMC0			_MMIO(0x30024)
+
+/*
+ * GEN9 clock gating regs
+ */
+#define GEN9_CLKGATE_DIS_0		_MMIO(0x46530)
+#define   DARBF_GATING_DIS		(1 << 27)
+#define   PWM2_GATING_DIS		(1 << 14)
+#define   PWM1_GATING_DIS		(1 << 13)
+
+#define GEN9_CLKGATE_DIS_3		_MMIO(0x46538)
+#define   TGL_VRH_GATING_DIS		REG_BIT(31)
+#define   DPT_GATING_DIS		REG_BIT(22)
+
+#define GEN9_CLKGATE_DIS_4		_MMIO(0x4653C)
+#define   BXT_GMBUS_GATING_DIS		(1 << 14)
+
+#define GEN9_CLKGATE_DIS_5		_MMIO(0x46540)
+#define   DPCE_GATING_DIS		REG_BIT(17)
+
+#define _CLKGATE_DIS_PSL_A		0x46520
+#define _CLKGATE_DIS_PSL_B		0x46524
+#define _CLKGATE_DIS_PSL_C		0x46528
+#define   DUPS1_GATING_DIS		(1 << 15)
+#define   DUPS2_GATING_DIS		(1 << 19)
+#define   DUPS3_GATING_DIS		(1 << 23)
+#define   CURSOR_GATING_DIS		REG_BIT(28)
+#define   DPF_GATING_DIS		(1 << 10)
+#define   DPF_RAM_GATING_DIS		(1 << 9)
+#define   DPFR_GATING_DIS		(1 << 8)
+
+#define CLKGATE_DIS_PSL(pipe) \
+	_MMIO_PIPE(pipe, _CLKGATE_DIS_PSL_A, _CLKGATE_DIS_PSL_B)
+
+#define _CLKGATE_DIS_PSL_EXT_A		0x4654C
+#define _CLKGATE_DIS_PSL_EXT_B		0x46550
+#define   PIPEDMC_GATING_DIS		REG_BIT(12)
+
+#define CLKGATE_DIS_PSL_EXT(pipe) \
+	_MMIO_PIPE(pipe, _CLKGATE_DIS_PSL_EXT_A, _CLKGATE_DIS_PSL_EXT_B)
+
+/*
+ * Display engine regs
+ */
+
+/* Pipe A CRC regs */
+#define _PIPE_CRC_CTL_A			0x60050
+#define   PIPE_CRC_ENABLE		REG_BIT(31)
+/* skl+ source selection */
+#define   PIPE_CRC_SOURCE_MASK_SKL	REG_GENMASK(30, 28)
+#define   PIPE_CRC_SOURCE_PLANE_1_SKL	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_SKL, 0)
+#define   PIPE_CRC_SOURCE_PLANE_2_SKL	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_SKL, 2)
+#define   PIPE_CRC_SOURCE_DMUX_SKL	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_SKL, 4)
+#define   PIPE_CRC_SOURCE_PLANE_3_SKL	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_SKL, 6)
+#define   PIPE_CRC_SOURCE_PLANE_4_SKL	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_SKL, 7)
+#define   PIPE_CRC_SOURCE_PLANE_5_SKL	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_SKL, 5)
+#define   PIPE_CRC_SOURCE_PLANE_6_SKL	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_SKL, 3)
+#define   PIPE_CRC_SOURCE_PLANE_7_SKL	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_SKL, 1)
+/* ivb+ source selection */
+#define   PIPE_CRC_SOURCE_MASK_IVB	REG_GENMASK(30, 29)
+#define   PIPE_CRC_SOURCE_PRIMARY_IVB	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_IVB, 0)
+#define   PIPE_CRC_SOURCE_SPRITE_IVB	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_IVB, 1)
+#define   PIPE_CRC_SOURCE_PF_IVB	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_IVB, 2)
+/* ilk+ source selection */
+#define   PIPE_CRC_SOURCE_MASK_ILK	REG_GENMASK(30, 28)
+#define   PIPE_CRC_SOURCE_PRIMARY_ILK	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_ILK, 0)
+#define   PIPE_CRC_SOURCE_SPRITE_ILK	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_ILK, 1)
+#define   PIPE_CRC_SOURCE_PIPE_ILK	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_ILK, 2)
+/* embedded DP port on the north display block */
+#define   PIPE_CRC_SOURCE_PORT_A_ILK	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_ILK, 4)
+#define   PIPE_CRC_SOURCE_FDI_ILK	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_ILK, 5)
+/* vlv source selection */
+#define   PIPE_CRC_SOURCE_MASK_VLV	REG_GENMASK(30, 27)
+#define   PIPE_CRC_SOURCE_PIPE_VLV	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_VLV, 0)
+#define   PIPE_CRC_SOURCE_HDMIB_VLV	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_VLV, 1)
+#define   PIPE_CRC_SOURCE_HDMIC_VLV	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_VLV, 2)
+/* with DP port the pipe source is invalid */
+#define   PIPE_CRC_SOURCE_DP_D_VLV	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_VLV, 3)
+#define   PIPE_CRC_SOURCE_DP_B_VLV	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_VLV, 6)
+#define   PIPE_CRC_SOURCE_DP_C_VLV	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_VLV, 7)
+/* gen3+ source selection */
+#define   PIPE_CRC_SOURCE_MASK_I9XX	REG_GENMASK(30, 28)
+#define   PIPE_CRC_SOURCE_PIPE_I9XX	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_I9XX, 0)
+#define   PIPE_CRC_SOURCE_SDVOB_I9XX	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_I9XX, 1)
+#define   PIPE_CRC_SOURCE_SDVOC_I9XX	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_I9XX, 2)
+/* with DP/TV port the pipe source is invalid */
+#define   PIPE_CRC_SOURCE_DP_D_G4X	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_I9XX, 3)
+#define   PIPE_CRC_SOURCE_TV_PRE	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_I9XX, 4)
+#define   PIPE_CRC_SOURCE_TV_POST	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_I9XX, 5)
+#define   PIPE_CRC_SOURCE_DP_B_G4X	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_I9XX, 6)
+#define   PIPE_CRC_SOURCE_DP_C_G4X	REG_FIELD_PREP(PIPE_CRC_SOURCE_MASK_I9XX, 7)
+/* gen2 doesn't have source selection bits */
+#define   PIPE_CRC_INCLUDE_BORDER_I8XX	REG_BIT(30)
+
+#define _PIPE_CRC_RES_1_A_IVB		0x60064
+#define _PIPE_CRC_RES_2_A_IVB		0x60068
+#define _PIPE_CRC_RES_3_A_IVB		0x6006c
+#define _PIPE_CRC_RES_4_A_IVB		0x60070
+#define _PIPE_CRC_RES_5_A_IVB		0x60074
+
+#define _PIPE_CRC_RES_RED_A		0x60060
+#define _PIPE_CRC_RES_GREEN_A		0x60064
+#define _PIPE_CRC_RES_BLUE_A		0x60068
+#define _PIPE_CRC_RES_RES1_A_I915	0x6006c
+#define _PIPE_CRC_RES_RES2_A_G4X	0x60080
+
+/* Pipe B CRC regs */
+#define _PIPE_CRC_RES_1_B_IVB		0x61064
+#define _PIPE_CRC_RES_2_B_IVB		0x61068
+#define _PIPE_CRC_RES_3_B_IVB		0x6106c
+#define _PIPE_CRC_RES_4_B_IVB		0x61070
+#define _PIPE_CRC_RES_5_B_IVB		0x61074
+
+#define PIPE_CRC_CTL(pipe)		_MMIO_TRANS2(pipe, _PIPE_CRC_CTL_A)
+#define PIPE_CRC_RES_1_IVB(pipe)	_MMIO_TRANS2(pipe, _PIPE_CRC_RES_1_A_IVB)
+#define PIPE_CRC_RES_2_IVB(pipe)	_MMIO_TRANS2(pipe, _PIPE_CRC_RES_2_A_IVB)
+#define PIPE_CRC_RES_3_IVB(pipe)	_MMIO_TRANS2(pipe, _PIPE_CRC_RES_3_A_IVB)
+#define PIPE_CRC_RES_4_IVB(pipe)	_MMIO_TRANS2(pipe, _PIPE_CRC_RES_4_A_IVB)
+#define PIPE_CRC_RES_5_IVB(pipe)	_MMIO_TRANS2(pipe, _PIPE_CRC_RES_5_A_IVB)
+
+#define PIPE_CRC_RES_RED(pipe)		_MMIO_TRANS2(pipe, _PIPE_CRC_RES_RED_A)
+#define PIPE_CRC_RES_GREEN(pipe)	_MMIO_TRANS2(pipe, _PIPE_CRC_RES_GREEN_A)
+#define PIPE_CRC_RES_BLUE(pipe)		_MMIO_TRANS2(pipe, _PIPE_CRC_RES_BLUE_A)
+#define PIPE_CRC_RES_RES1_I915(pipe)	_MMIO_TRANS2(pipe, _PIPE_CRC_RES_RES1_A_I915)
+#define PIPE_CRC_RES_RES2_G4X(pipe)	_MMIO_TRANS2(pipe, _PIPE_CRC_RES_RES2_A_G4X)
+
+/* Pipe A timing regs */
+#define _HTOTAL_A	0x60000
+#define _HBLANK_A	0x60004
+#define _HSYNC_A	0x60008
+#define _VTOTAL_A	0x6000c
+#define _VBLANK_A	0x60010
+#define _VSYNC_A	0x60014
+#define _EXITLINE_A	0x60018
+#define _PIPEASRC	0x6001c
+#define   PIPESRC_WIDTH_MASK	REG_GENMASK(31, 16)
+#define   PIPESRC_WIDTH(w)	REG_FIELD_PREP(PIPESRC_WIDTH_MASK, (w))
+#define   PIPESRC_HEIGHT_MASK	REG_GENMASK(15, 0)
+#define   PIPESRC_HEIGHT(h)	REG_FIELD_PREP(PIPESRC_HEIGHT_MASK, (h))
+#define _BCLRPAT_A	0x60020
+#define _VSYNCSHIFT_A	0x60028
+#define _PIPE_MULT_A	0x6002c
+
+/* Pipe B timing regs */
+#define _HTOTAL_B	0x61000
+#define _HBLANK_B	0x61004
+#define _HSYNC_B	0x61008
+#define _VTOTAL_B	0x6100c
+#define _VBLANK_B	0x61010
+#define _VSYNC_B	0x61014
+#define _PIPEBSRC	0x6101c
+#define _BCLRPAT_B	0x61020
+#define _VSYNCSHIFT_B	0x61028
+#define _PIPE_MULT_B	0x6102c
+
+/* DSI 0 timing regs */
+#define _HTOTAL_DSI0		0x6b000
+#define _HSYNC_DSI0		0x6b008
+#define _VTOTAL_DSI0		0x6b00c
+#define _VSYNC_DSI0		0x6b014
+#define _VSYNCSHIFT_DSI0	0x6b028
+
+/* DSI 1 timing regs */
+#define _HTOTAL_DSI1		0x6b800
+#define _HSYNC_DSI1		0x6b808
+#define _VTOTAL_DSI1		0x6b80c
+#define _VSYNC_DSI1		0x6b814
+#define _VSYNCSHIFT_DSI1	0x6b828
+
+#define TRANSCODER_A_OFFSET 0x60000
+#define TRANSCODER_B_OFFSET 0x61000
+#define TRANSCODER_C_OFFSET 0x62000
+#define CHV_TRANSCODER_C_OFFSET 0x63000
+#define TRANSCODER_D_OFFSET 0x63000
+#define TRANSCODER_EDP_OFFSET 0x6f000
+#define TRANSCODER_DSI0_OFFSET	0x6b000
+#define TRANSCODER_DSI1_OFFSET	0x6b800
+
+#define HTOTAL(trans)		_MMIO_TRANS2(trans, _HTOTAL_A)
+#define HBLANK(trans)		_MMIO_TRANS2(trans, _HBLANK_A)
+#define HSYNC(trans)		_MMIO_TRANS2(trans, _HSYNC_A)
+#define VTOTAL(trans)		_MMIO_TRANS2(trans, _VTOTAL_A)
+#define VBLANK(trans)		_MMIO_TRANS2(trans, _VBLANK_A)
+#define VSYNC(trans)		_MMIO_TRANS2(trans, _VSYNC_A)
+#define BCLRPAT(trans)		_MMIO_TRANS2(trans, _BCLRPAT_A)
+#define VSYNCSHIFT(trans)	_MMIO_TRANS2(trans, _VSYNCSHIFT_A)
+#define PIPESRC(trans)		_MMIO_TRANS2(trans, _PIPEASRC)
+#define PIPE_MULT(trans)	_MMIO_TRANS2(trans, _PIPE_MULT_A)
+
+#define EXITLINE(trans)		_MMIO_TRANS2(trans, _EXITLINE_A)
+#define   EXITLINE_ENABLE	REG_BIT(31)
+#define   EXITLINE_MASK		REG_GENMASK(12, 0)
+#define   EXITLINE_SHIFT	0
+
+/* VRR registers */
+#define _TRANS_VRR_CTL_A		0x60420
+#define _TRANS_VRR_CTL_B		0x61420
+#define _TRANS_VRR_CTL_C		0x62420
+#define _TRANS_VRR_CTL_D		0x63420
+#define TRANS_VRR_CTL(trans)			_MMIO_TRANS2(trans, _TRANS_VRR_CTL_A)
+#define   VRR_CTL_VRR_ENABLE			REG_BIT(31)
+#define   VRR_CTL_IGN_MAX_SHIFT			REG_BIT(30)
+#define   VRR_CTL_FLIP_LINE_EN			REG_BIT(29)
+#define   VRR_CTL_PIPELINE_FULL_MASK		REG_GENMASK(10, 3)
+#define   VRR_CTL_PIPELINE_FULL(x)		REG_FIELD_PREP(VRR_CTL_PIPELINE_FULL_MASK, (x))
+#define   VRR_CTL_PIPELINE_FULL_OVERRIDE	REG_BIT(0)
+#define	  XELPD_VRR_CTL_VRR_GUARDBAND_MASK	REG_GENMASK(15, 0)
+#define	  XELPD_VRR_CTL_VRR_GUARDBAND(x)	REG_FIELD_PREP(XELPD_VRR_CTL_VRR_GUARDBAND_MASK, (x))
+
+#define _TRANS_VRR_VMAX_A		0x60424
+#define _TRANS_VRR_VMAX_B		0x61424
+#define _TRANS_VRR_VMAX_C		0x62424
+#define _TRANS_VRR_VMAX_D		0x63424
+#define TRANS_VRR_VMAX(trans)		_MMIO_TRANS2(trans, _TRANS_VRR_VMAX_A)
+#define   VRR_VMAX_MASK			REG_GENMASK(19, 0)
+
+#define _TRANS_VRR_VMIN_A		0x60434
+#define _TRANS_VRR_VMIN_B		0x61434
+#define _TRANS_VRR_VMIN_C		0x62434
+#define _TRANS_VRR_VMIN_D		0x63434
+#define TRANS_VRR_VMIN(trans)		_MMIO_TRANS2(trans, _TRANS_VRR_VMIN_A)
+#define   VRR_VMIN_MASK			REG_GENMASK(15, 0)
+
+#define _TRANS_VRR_VMAXSHIFT_A		0x60428
+#define _TRANS_VRR_VMAXSHIFT_B		0x61428
+#define _TRANS_VRR_VMAXSHIFT_C		0x62428
+#define _TRANS_VRR_VMAXSHIFT_D		0x63428
+#define TRANS_VRR_VMAXSHIFT(trans)	_MMIO_TRANS2(trans, \
+					_TRANS_VRR_VMAXSHIFT_A)
+#define   VRR_VMAXSHIFT_DEC_MASK	REG_GENMASK(29, 16)
+#define   VRR_VMAXSHIFT_DEC		REG_BIT(16)
+#define   VRR_VMAXSHIFT_INC_MASK	REG_GENMASK(12, 0)
+
+#define _TRANS_VRR_STATUS_A		0x6042C
+#define _TRANS_VRR_STATUS_B		0x6142C
+#define _TRANS_VRR_STATUS_C		0x6242C
+#define _TRANS_VRR_STATUS_D		0x6342C
+#define TRANS_VRR_STATUS(trans)		_MMIO_TRANS2(trans, _TRANS_VRR_STATUS_A)
+#define   VRR_STATUS_VMAX_REACHED	REG_BIT(31)
+#define   VRR_STATUS_NOFLIP_TILL_BNDR	REG_BIT(30)
+#define   VRR_STATUS_FLIP_BEF_BNDR	REG_BIT(29)
+#define   VRR_STATUS_NO_FLIP_FRAME	REG_BIT(28)
+#define   VRR_STATUS_VRR_EN_LIVE	REG_BIT(27)
+#define   VRR_STATUS_FLIPS_SERVICED	REG_BIT(26)
+#define   VRR_STATUS_VBLANK_MASK	REG_GENMASK(22, 20)
+#define   STATUS_FSM_IDLE		REG_FIELD_PREP(VRR_STATUS_VBLANK_MASK, 0)
+#define   STATUS_FSM_WAIT_TILL_FDB	REG_FIELD_PREP(VRR_STATUS_VBLANK_MASK, 1)
+#define   STATUS_FSM_WAIT_TILL_FS	REG_FIELD_PREP(VRR_STATUS_VBLANK_MASK, 2)
+#define   STATUS_FSM_WAIT_TILL_FLIP	REG_FIELD_PREP(VRR_STATUS_VBLANK_MASK, 3)
+#define   STATUS_FSM_PIPELINE_FILL	REG_FIELD_PREP(VRR_STATUS_VBLANK_MASK, 4)
+#define   STATUS_FSM_ACTIVE		REG_FIELD_PREP(VRR_STATUS_VBLANK_MASK, 5)
+#define   STATUS_FSM_LEGACY_VBLANK	REG_FIELD_PREP(VRR_STATUS_VBLANK_MASK, 6)
+
+#define _TRANS_VRR_VTOTAL_PREV_A	0x60480
+#define _TRANS_VRR_VTOTAL_PREV_B	0x61480
+#define _TRANS_VRR_VTOTAL_PREV_C	0x62480
+#define _TRANS_VRR_VTOTAL_PREV_D	0x63480
+#define TRANS_VRR_VTOTAL_PREV(trans)	_MMIO_TRANS2(trans, \
+					_TRANS_VRR_VTOTAL_PREV_A)
+#define   VRR_VTOTAL_FLIP_BEFR_BNDR	REG_BIT(31)
+#define   VRR_VTOTAL_FLIP_AFTER_BNDR	REG_BIT(30)
+#define   VRR_VTOTAL_FLIP_AFTER_DBLBUF	REG_BIT(29)
+#define   VRR_VTOTAL_PREV_FRAME_MASK	REG_GENMASK(19, 0)
+
+#define _TRANS_VRR_FLIPLINE_A		0x60438
+#define _TRANS_VRR_FLIPLINE_B		0x61438
+#define _TRANS_VRR_FLIPLINE_C		0x62438
+#define _TRANS_VRR_FLIPLINE_D		0x63438
+#define TRANS_VRR_FLIPLINE(trans)	_MMIO_TRANS2(trans, \
+					_TRANS_VRR_FLIPLINE_A)
+#define   VRR_FLIPLINE_MASK		REG_GENMASK(19, 0)
+
+#define _TRANS_VRR_STATUS2_A		0x6043C
+#define _TRANS_VRR_STATUS2_B		0x6143C
+#define _TRANS_VRR_STATUS2_C		0x6243C
+#define _TRANS_VRR_STATUS2_D		0x6343C
+#define TRANS_VRR_STATUS2(trans)	_MMIO_TRANS2(trans, _TRANS_VRR_STATUS2_A)
+#define   VRR_STATUS2_VERT_LN_CNT_MASK	REG_GENMASK(19, 0)
+
+#define _TRANS_PUSH_A			0x60A70
+#define _TRANS_PUSH_B			0x61A70
+#define _TRANS_PUSH_C			0x62A70
+#define _TRANS_PUSH_D			0x63A70
+#define TRANS_PUSH(trans)		_MMIO_TRANS2(trans, _TRANS_PUSH_A)
+#define   TRANS_PUSH_EN			REG_BIT(31)
+#define   TRANS_PUSH_SEND		REG_BIT(30)
+
+/*
+ * HSW+ eDP PSR registers
+ *
+ * HSW PSR registers are relative to DDIA(_DDI_BUF_CTL_A + 0x800) with just one
+ * instance of it
+ */
+#define _SRD_CTL_A				0x60800
+#define _SRD_CTL_EDP				0x6f800
+#define EDP_PSR_CTL(tran)			_MMIO_TRANS2(tran, _SRD_CTL_A)
+#define   EDP_PSR_ENABLE			(1 << 31)
+#define   BDW_PSR_SINGLE_FRAME			(1 << 30)
+#define   EDP_PSR_RESTORE_PSR_ACTIVE_CTX_MASK	(1 << 29) /* SW can't modify */
+#define   EDP_PSR_LINK_STANDBY			(1 << 27)
+#define   EDP_PSR_MIN_LINK_ENTRY_TIME_MASK	(3 << 25)
+#define   EDP_PSR_MIN_LINK_ENTRY_TIME_8_LINES	(0 << 25)
+#define   EDP_PSR_MIN_LINK_ENTRY_TIME_4_LINES	(1 << 25)
+#define   EDP_PSR_MIN_LINK_ENTRY_TIME_2_LINES	(2 << 25)
+#define   EDP_PSR_MIN_LINK_ENTRY_TIME_0_LINES	(3 << 25)
+#define   EDP_PSR_MAX_SLEEP_TIME_SHIFT		20
+#define   EDP_PSR_SKIP_AUX_EXIT			(1 << 12)
+#define   EDP_PSR_TP1_TP2_SEL			(0 << 11)
+#define   EDP_PSR_TP1_TP3_SEL			(1 << 11)
+#define   EDP_PSR_CRC_ENABLE			(1 << 10) /* BDW+ */
+#define   EDP_PSR_TP2_TP3_TIME_500us		(0 << 8)
+#define   EDP_PSR_TP2_TP3_TIME_100us		(1 << 8)
+#define   EDP_PSR_TP2_TP3_TIME_2500us		(2 << 8)
+#define   EDP_PSR_TP2_TP3_TIME_0us		(3 << 8)
+#define   EDP_PSR_TP4_TIME_0US			(3 << 6) /* ICL+ */
+#define   EDP_PSR_TP1_TIME_500us		(0 << 4)
+#define   EDP_PSR_TP1_TIME_100us		(1 << 4)
+#define   EDP_PSR_TP1_TIME_2500us		(2 << 4)
+#define   EDP_PSR_TP1_TIME_0us			(3 << 4)
+#define   EDP_PSR_IDLE_FRAME_SHIFT		0
+
+/*
+ * Until TGL, IMR/IIR are fixed at 0x648xx. On TGL+ those registers are relative
+ * to transcoder and bits defined for each one as if using no shift (i.e. as if
+ * it was for TRANSCODER_EDP)
+ */
+#define EDP_PSR_IMR				_MMIO(0x64834)
+#define EDP_PSR_IIR				_MMIO(0x64838)
+#define _PSR_IMR_A				0x60814
+#define _PSR_IIR_A				0x60818
+#define TRANS_PSR_IMR(tran)			_MMIO_TRANS2(tran, _PSR_IMR_A)
+#define TRANS_PSR_IIR(tran)			_MMIO_TRANS2(tran, _PSR_IIR_A)
+#define   _EDP_PSR_TRANS_SHIFT(trans)		((trans) == TRANSCODER_EDP ? \
+						 0 : ((trans) - TRANSCODER_A + 1) * 8)
+#define   TGL_PSR_MASK			REG_GENMASK(2, 0)
+#define   TGL_PSR_ERROR			REG_BIT(2)
+#define   TGL_PSR_POST_EXIT		REG_BIT(1)
+#define   TGL_PSR_PRE_ENTRY		REG_BIT(0)
+#define   EDP_PSR_MASK(trans)		(TGL_PSR_MASK <<		\
+					 _EDP_PSR_TRANS_SHIFT(trans))
+#define   EDP_PSR_ERROR(trans)		(TGL_PSR_ERROR <<		\
+					 _EDP_PSR_TRANS_SHIFT(trans))
+#define   EDP_PSR_POST_EXIT(trans)	(TGL_PSR_POST_EXIT <<		\
+					 _EDP_PSR_TRANS_SHIFT(trans))
+#define   EDP_PSR_PRE_ENTRY(trans)	(TGL_PSR_PRE_ENTRY <<		\
+					 _EDP_PSR_TRANS_SHIFT(trans))
+
+#define _SRD_AUX_DATA_A				0x60814
+#define _SRD_AUX_DATA_EDP			0x6f814
+#define EDP_PSR_AUX_DATA(tran, i)		_MMIO_TRANS2(tran, _SRD_AUX_DATA_A + (i) + 4) /* 5 registers */
+
+#define _SRD_STATUS_A				0x60840
+#define _SRD_STATUS_EDP				0x6f840
+#define EDP_PSR_STATUS(tran)			_MMIO_TRANS2(tran, _SRD_STATUS_A)
+#define   EDP_PSR_STATUS_STATE_MASK		(7 << 29)
+#define   EDP_PSR_STATUS_STATE_SHIFT		29
+#define   EDP_PSR_STATUS_STATE_IDLE		(0 << 29)
+#define   EDP_PSR_STATUS_STATE_SRDONACK		(1 << 29)
+#define   EDP_PSR_STATUS_STATE_SRDENT		(2 << 29)
+#define   EDP_PSR_STATUS_STATE_BUFOFF		(3 << 29)
+#define   EDP_PSR_STATUS_STATE_BUFON		(4 << 29)
+#define   EDP_PSR_STATUS_STATE_AUXACK		(5 << 29)
+#define   EDP_PSR_STATUS_STATE_SRDOFFACK	(6 << 29)
+#define   EDP_PSR_STATUS_LINK_MASK		(3 << 26)
+#define   EDP_PSR_STATUS_LINK_FULL_OFF		(0 << 26)
+#define   EDP_PSR_STATUS_LINK_FULL_ON		(1 << 26)
+#define   EDP_PSR_STATUS_LINK_STANDBY		(2 << 26)
+#define   EDP_PSR_STATUS_MAX_SLEEP_TIMER_SHIFT	20
+#define   EDP_PSR_STATUS_MAX_SLEEP_TIMER_MASK	0x1f
+#define   EDP_PSR_STATUS_COUNT_SHIFT		16
+#define   EDP_PSR_STATUS_COUNT_MASK		0xf
+#define   EDP_PSR_STATUS_AUX_ERROR		(1 << 15)
+#define   EDP_PSR_STATUS_AUX_SENDING		(1 << 12)
+#define   EDP_PSR_STATUS_SENDING_IDLE		(1 << 9)
+#define   EDP_PSR_STATUS_SENDING_TP2_TP3	(1 << 8)
+#define   EDP_PSR_STATUS_SENDING_TP1		(1 << 4)
+#define   EDP_PSR_STATUS_IDLE_MASK		0xf
+
+#define _SRD_PERF_CNT_A			0x60844
+#define _SRD_PERF_CNT_EDP		0x6f844
+#define EDP_PSR_PERF_CNT(tran)		_MMIO_TRANS2(tran, _SRD_PERF_CNT_A)
+#define   EDP_PSR_PERF_CNT_MASK		0xffffff
+
+/* PSR_MASK on SKL+ */
+#define _SRD_DEBUG_A				0x60860
+#define _SRD_DEBUG_EDP				0x6f860
+#define EDP_PSR_DEBUG(tran)			_MMIO_TRANS2(tran, _SRD_DEBUG_A)
+#define   EDP_PSR_DEBUG_MASK_MAX_SLEEP         (1 << 28)
+#define   EDP_PSR_DEBUG_MASK_LPSP              (1 << 27)
+#define   EDP_PSR_DEBUG_MASK_MEMUP             (1 << 26)
+#define   EDP_PSR_DEBUG_MASK_HPD               (1 << 25)
+#define   EDP_PSR_DEBUG_MASK_DISP_REG_WRITE    (1 << 16) /* Reserved in ICL+ */
+#define   EDP_PSR_DEBUG_EXIT_ON_PIXEL_UNDERRUN (1 << 15) /* SKL+ */
+
+#define _PSR2_CTL_A				0x60900
+#define _PSR2_CTL_EDP				0x6f900
+#define EDP_PSR2_CTL(tran)			_MMIO_TRANS2(tran, _PSR2_CTL_A)
+#define   EDP_PSR2_ENABLE			(1 << 31)
+#define   EDP_SU_TRACK_ENABLE			(1 << 30) /* up to adl-p */
+#define   TGL_EDP_PSR2_BLOCK_COUNT_NUM_2	(0 << 28)
+#define   TGL_EDP_PSR2_BLOCK_COUNT_NUM_3	(1 << 28)
+#define   EDP_Y_COORDINATE_ENABLE		REG_BIT(25) /* display 10, 11 and 12 */
+#define   EDP_PSR2_SU_SDP_SCANLINE		REG_BIT(25) /* display 13+ */
+#define   EDP_MAX_SU_DISABLE_TIME(t)		((t) << 20)
+#define   EDP_MAX_SU_DISABLE_TIME_MASK		(0x1f << 20)
+#define   EDP_PSR2_IO_BUFFER_WAKE_MAX_LINES	8
+#define   EDP_PSR2_IO_BUFFER_WAKE(lines)	((EDP_PSR2_IO_BUFFER_WAKE_MAX_LINES - (lines)) << 13)
+#define   EDP_PSR2_IO_BUFFER_WAKE_MASK		(3 << 13)
+#define   TGL_EDP_PSR2_IO_BUFFER_WAKE_MIN_LINES	5
+#define   TGL_EDP_PSR2_IO_BUFFER_WAKE_SHIFT	13
+#define   TGL_EDP_PSR2_IO_BUFFER_WAKE(lines)	(((lines) - TGL_EDP_PSR2_IO_BUFFER_WAKE_MIN_LINES) << TGL_EDP_PSR2_IO_BUFFER_WAKE_SHIFT)
+#define   TGL_EDP_PSR2_IO_BUFFER_WAKE_MASK	(7 << 13)
+#define   EDP_PSR2_FAST_WAKE_MAX_LINES		8
+#define   EDP_PSR2_FAST_WAKE(lines)		((EDP_PSR2_FAST_WAKE_MAX_LINES - (lines)) << 11)
+#define   EDP_PSR2_FAST_WAKE_MASK		(3 << 11)
+#define   TGL_EDP_PSR2_FAST_WAKE_MIN_LINES	5
+#define   TGL_EDP_PSR2_FAST_WAKE_MIN_SHIFT	10
+#define   TGL_EDP_PSR2_FAST_WAKE(lines)		(((lines) - TGL_EDP_PSR2_FAST_WAKE_MIN_LINES) << TGL_EDP_PSR2_FAST_WAKE_MIN_SHIFT)
+#define   TGL_EDP_PSR2_FAST_WAKE_MASK		(7 << 10)
+#define   EDP_PSR2_TP2_TIME_500us		(0 << 8)
+#define   EDP_PSR2_TP2_TIME_100us		(1 << 8)
+#define   EDP_PSR2_TP2_TIME_2500us		(2 << 8)
+#define   EDP_PSR2_TP2_TIME_50us		(3 << 8)
+#define   EDP_PSR2_TP2_TIME_MASK		(3 << 8)
+#define   EDP_PSR2_FRAME_BEFORE_SU_SHIFT	4
+#define   EDP_PSR2_FRAME_BEFORE_SU_MASK		(0xf << 4)
+#define   EDP_PSR2_FRAME_BEFORE_SU(a)		((a) << 4)
+#define   EDP_PSR2_IDLE_FRAME_MASK		0xf
+#define   EDP_PSR2_IDLE_FRAME_SHIFT		0
+
+#define _PSR_EVENT_TRANS_A			0x60848
+#define _PSR_EVENT_TRANS_B			0x61848
+#define _PSR_EVENT_TRANS_C			0x62848
+#define _PSR_EVENT_TRANS_D			0x63848
+#define _PSR_EVENT_TRANS_EDP			0x6f848
+#define PSR_EVENT(tran)				_MMIO_TRANS2(tran, _PSR_EVENT_TRANS_A)
+#define  PSR_EVENT_PSR2_WD_TIMER_EXPIRE		(1 << 17)
+#define  PSR_EVENT_PSR2_DISABLED		(1 << 16)
+#define  PSR_EVENT_SU_DIRTY_FIFO_UNDERRUN	(1 << 15)
+#define  PSR_EVENT_SU_CRC_FIFO_UNDERRUN		(1 << 14)
+#define  PSR_EVENT_GRAPHICS_RESET		(1 << 12)
+#define  PSR_EVENT_PCH_INTERRUPT		(1 << 11)
+#define  PSR_EVENT_MEMORY_UP			(1 << 10)
+#define  PSR_EVENT_FRONT_BUFFER_MODIFY		(1 << 9)
+#define  PSR_EVENT_WD_TIMER_EXPIRE		(1 << 8)
+#define  PSR_EVENT_PIPE_REGISTERS_UPDATE	(1 << 6)
+#define  PSR_EVENT_REGISTER_UPDATE		(1 << 5) /* Reserved in ICL+ */
+#define  PSR_EVENT_HDCP_ENABLE			(1 << 4)
+#define  PSR_EVENT_KVMR_SESSION_ENABLE		(1 << 3)
+#define  PSR_EVENT_VBI_ENABLE			(1 << 2)
+#define  PSR_EVENT_LPSP_MODE_EXIT		(1 << 1)
+#define  PSR_EVENT_PSR_DISABLE			(1 << 0)
+
+#define _PSR2_STATUS_A				0x60940
+#define _PSR2_STATUS_EDP			0x6f940
+#define EDP_PSR2_STATUS(tran)			_MMIO_TRANS2(tran, _PSR2_STATUS_A)
+#define EDP_PSR2_STATUS_STATE_MASK		REG_GENMASK(31, 28)
+#define EDP_PSR2_STATUS_STATE_DEEP_SLEEP	REG_FIELD_PREP(EDP_PSR2_STATUS_STATE_MASK, 0x8)
+
+#define _PSR2_SU_STATUS_A		0x60914
+#define _PSR2_SU_STATUS_EDP		0x6f914
+#define _PSR2_SU_STATUS(tran, index)	_MMIO_TRANS2(tran, _PSR2_SU_STATUS_A + (index) * 4)
+#define PSR2_SU_STATUS(tran, frame)	(_PSR2_SU_STATUS(tran, (frame) / 3))
+#define PSR2_SU_STATUS_SHIFT(frame)	(((frame) % 3) * 10)
+#define PSR2_SU_STATUS_MASK(frame)	(0x3ff << PSR2_SU_STATUS_SHIFT(frame))
+#define PSR2_SU_STATUS_FRAMES		8
+
+#define _PSR2_MAN_TRK_CTL_A					0x60910
+#define _PSR2_MAN_TRK_CTL_EDP					0x6f910
+#define PSR2_MAN_TRK_CTL(tran)					_MMIO_TRANS2(tran, _PSR2_MAN_TRK_CTL_A)
+#define  PSR2_MAN_TRK_CTL_ENABLE				REG_BIT(31)
+#define  PSR2_MAN_TRK_CTL_SU_REGION_START_ADDR_MASK		REG_GENMASK(30, 21)
+#define  PSR2_MAN_TRK_CTL_SU_REGION_START_ADDR(val)		REG_FIELD_PREP(PSR2_MAN_TRK_CTL_SU_REGION_START_ADDR_MASK, val)
+#define  PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR_MASK		REG_GENMASK(20, 11)
+#define  PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR(val)		REG_FIELD_PREP(PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR_MASK, val)
+#define  PSR2_MAN_TRK_CTL_SF_SINGLE_FULL_FRAME			REG_BIT(3)
+#define  PSR2_MAN_TRK_CTL_SF_CONTINUOS_FULL_FRAME		REG_BIT(2)
+#define  PSR2_MAN_TRK_CTL_SF_PARTIAL_FRAME_UPDATE		REG_BIT(1)
+#define  ADLP_PSR2_MAN_TRK_CTL_SU_REGION_START_ADDR_MASK	REG_GENMASK(28, 16)
+#define  ADLP_PSR2_MAN_TRK_CTL_SU_REGION_START_ADDR(val)	REG_FIELD_PREP(ADLP_PSR2_MAN_TRK_CTL_SU_REGION_START_ADDR_MASK, val)
+#define  ADLP_PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR_MASK		REG_GENMASK(12, 0)
+#define  ADLP_PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR(val)		REG_FIELD_PREP(ADLP_PSR2_MAN_TRK_CTL_SU_REGION_END_ADDR_MASK, val)
+#define  ADLP_PSR2_MAN_TRK_CTL_SF_PARTIAL_FRAME_UPDATE		REG_BIT(31)
+#define  ADLP_PSR2_MAN_TRK_CTL_SF_SINGLE_FULL_FRAME		REG_BIT(14)
+#define  ADLP_PSR2_MAN_TRK_CTL_SF_CONTINUOS_FULL_FRAME		REG_BIT(13)
+
+/* Icelake DSC Rate Control Range Parameter Registers */
+#define DSCA_RC_RANGE_PARAMETERS_0		_MMIO(0x6B240)
+#define DSCA_RC_RANGE_PARAMETERS_0_UDW		_MMIO(0x6B240 + 4)
+#define DSCC_RC_RANGE_PARAMETERS_0		_MMIO(0x6BA40)
+#define DSCC_RC_RANGE_PARAMETERS_0_UDW		_MMIO(0x6BA40 + 4)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_0_PB	(0x78208)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_0_UDW_PB	(0x78208 + 4)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_0_PB	(0x78308)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_0_UDW_PB	(0x78308 + 4)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_0_PC	(0x78408)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_0_UDW_PC	(0x78408 + 4)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_0_PC	(0x78508)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_0_UDW_PC	(0x78508 + 4)
+#define ICL_DSC0_RC_RANGE_PARAMETERS_0(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_0_PB, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_0_PC)
+#define ICL_DSC0_RC_RANGE_PARAMETERS_0_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_0_UDW_PB, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_0_UDW_PC)
+#define ICL_DSC1_RC_RANGE_PARAMETERS_0(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_0_PB, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_0_PC)
+#define ICL_DSC1_RC_RANGE_PARAMETERS_0_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_0_UDW_PB, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_0_UDW_PC)
+#define RC_BPG_OFFSET_SHIFT			10
+#define RC_MAX_QP_SHIFT				5
+#define RC_MIN_QP_SHIFT				0
+
+#define DSCA_RC_RANGE_PARAMETERS_1		_MMIO(0x6B248)
+#define DSCA_RC_RANGE_PARAMETERS_1_UDW		_MMIO(0x6B248 + 4)
+#define DSCC_RC_RANGE_PARAMETERS_1		_MMIO(0x6BA48)
+#define DSCC_RC_RANGE_PARAMETERS_1_UDW		_MMIO(0x6BA48 + 4)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_1_PB	(0x78210)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_1_UDW_PB	(0x78210 + 4)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_1_PB	(0x78310)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_1_UDW_PB	(0x78310 + 4)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_1_PC	(0x78410)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_1_UDW_PC	(0x78410 + 4)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_1_PC	(0x78510)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_1_UDW_PC	(0x78510 + 4)
+#define ICL_DSC0_RC_RANGE_PARAMETERS_1(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_1_PB, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_1_PC)
+#define ICL_DSC0_RC_RANGE_PARAMETERS_1_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_1_UDW_PB, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_1_UDW_PC)
+#define ICL_DSC1_RC_RANGE_PARAMETERS_1(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_1_PB, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_1_PC)
+#define ICL_DSC1_RC_RANGE_PARAMETERS_1_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_1_UDW_PB, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_1_UDW_PC)
+
+#define DSCA_RC_RANGE_PARAMETERS_2		_MMIO(0x6B250)
+#define DSCA_RC_RANGE_PARAMETERS_2_UDW		_MMIO(0x6B250 + 4)
+#define DSCC_RC_RANGE_PARAMETERS_2		_MMIO(0x6BA50)
+#define DSCC_RC_RANGE_PARAMETERS_2_UDW		_MMIO(0x6BA50 + 4)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_2_PB	(0x78218)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_2_UDW_PB	(0x78218 + 4)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_2_PB	(0x78318)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_2_UDW_PB	(0x78318 + 4)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_2_PC	(0x78418)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_2_UDW_PC	(0x78418 + 4)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_2_PC	(0x78518)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_2_UDW_PC	(0x78518 + 4)
+#define ICL_DSC0_RC_RANGE_PARAMETERS_2(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_2_PB, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_2_PC)
+#define ICL_DSC0_RC_RANGE_PARAMETERS_2_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_2_UDW_PB, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_2_UDW_PC)
+#define ICL_DSC1_RC_RANGE_PARAMETERS_2(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_2_PB, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_2_PC)
+#define ICL_DSC1_RC_RANGE_PARAMETERS_2_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_2_UDW_PB, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_2_UDW_PC)
+
+#define DSCA_RC_RANGE_PARAMETERS_3		_MMIO(0x6B258)
+#define DSCA_RC_RANGE_PARAMETERS_3_UDW		_MMIO(0x6B258 + 4)
+#define DSCC_RC_RANGE_PARAMETERS_3		_MMIO(0x6BA58)
+#define DSCC_RC_RANGE_PARAMETERS_3_UDW		_MMIO(0x6BA58 + 4)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_3_PB	(0x78220)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_3_UDW_PB	(0x78220 + 4)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_3_PB	(0x78320)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_3_UDW_PB	(0x78320 + 4)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_3_PC	(0x78420)
+#define _ICL_DSC0_RC_RANGE_PARAMETERS_3_UDW_PC	(0x78420 + 4)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_3_PC	(0x78520)
+#define _ICL_DSC1_RC_RANGE_PARAMETERS_3_UDW_PC	(0x78520 + 4)
+#define ICL_DSC0_RC_RANGE_PARAMETERS_3(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_3_PB, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_3_PC)
+#define ICL_DSC0_RC_RANGE_PARAMETERS_3_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_3_UDW_PB, \
+							_ICL_DSC0_RC_RANGE_PARAMETERS_3_UDW_PC)
+#define ICL_DSC1_RC_RANGE_PARAMETERS_3(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_3_PB, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_3_PC)
+#define ICL_DSC1_RC_RANGE_PARAMETERS_3_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_3_UDW_PB, \
+							_ICL_DSC1_RC_RANGE_PARAMETERS_3_UDW_PC)
+
+/* VGA port control */
+#define ADPA			_MMIO(0x61100)
+#define PCH_ADPA                _MMIO(0xe1100)
+#define VLV_ADPA		_MMIO(VLV_DISPLAY_BASE + 0x61100)
+
+#define   ADPA_DAC_ENABLE	(1 << 31)
+#define   ADPA_DAC_DISABLE	0
+#define   ADPA_PIPE_SEL_SHIFT		30
+#define   ADPA_PIPE_SEL_MASK		(1 << 30)
+#define   ADPA_PIPE_SEL(pipe)		((pipe) << 30)
+#define   ADPA_PIPE_SEL_SHIFT_CPT	29
+#define   ADPA_PIPE_SEL_MASK_CPT	(3 << 29)
+#define   ADPA_PIPE_SEL_CPT(pipe)	((pipe) << 29)
+#define   ADPA_CRT_HOTPLUG_MASK  0x03ff0000 /* bit 25-16 */
+#define   ADPA_CRT_HOTPLUG_MONITOR_NONE  (0 << 24)
+#define   ADPA_CRT_HOTPLUG_MONITOR_MASK  (3 << 24)
+#define   ADPA_CRT_HOTPLUG_MONITOR_COLOR (3 << 24)
+#define   ADPA_CRT_HOTPLUG_MONITOR_MONO  (2 << 24)
+#define   ADPA_CRT_HOTPLUG_ENABLE        (1 << 23)
+#define   ADPA_CRT_HOTPLUG_PERIOD_64     (0 << 22)
+#define   ADPA_CRT_HOTPLUG_PERIOD_128    (1 << 22)
+#define   ADPA_CRT_HOTPLUG_WARMUP_5MS    (0 << 21)
+#define   ADPA_CRT_HOTPLUG_WARMUP_10MS   (1 << 21)
+#define   ADPA_CRT_HOTPLUG_SAMPLE_2S     (0 << 20)
+#define   ADPA_CRT_HOTPLUG_SAMPLE_4S     (1 << 20)
+#define   ADPA_CRT_HOTPLUG_VOLTAGE_40    (0 << 18)
+#define   ADPA_CRT_HOTPLUG_VOLTAGE_50    (1 << 18)
+#define   ADPA_CRT_HOTPLUG_VOLTAGE_60    (2 << 18)
+#define   ADPA_CRT_HOTPLUG_VOLTAGE_70    (3 << 18)
+#define   ADPA_CRT_HOTPLUG_VOLREF_325MV  (0 << 17)
+#define   ADPA_CRT_HOTPLUG_VOLREF_475MV  (1 << 17)
+#define   ADPA_CRT_HOTPLUG_FORCE_TRIGGER (1 << 16)
+#define   ADPA_USE_VGA_HVPOLARITY (1 << 15)
+#define   ADPA_SETS_HVPOLARITY	0
+#define   ADPA_VSYNC_CNTL_DISABLE (1 << 10)
+#define   ADPA_VSYNC_CNTL_ENABLE 0
+#define   ADPA_HSYNC_CNTL_DISABLE (1 << 11)
+#define   ADPA_HSYNC_CNTL_ENABLE 0
+#define   ADPA_VSYNC_ACTIVE_HIGH (1 << 4)
+#define   ADPA_VSYNC_ACTIVE_LOW	0
+#define   ADPA_HSYNC_ACTIVE_HIGH (1 << 3)
+#define   ADPA_HSYNC_ACTIVE_LOW	0
+#define   ADPA_DPMS_MASK	(~(3 << 10))
+#define   ADPA_DPMS_ON		(0 << 10)
+#define   ADPA_DPMS_SUSPEND	(1 << 10)
+#define   ADPA_DPMS_STANDBY	(2 << 10)
+#define   ADPA_DPMS_OFF		(3 << 10)
+
+
+/* Hotplug control (945+ only) */
+#define PORT_HOTPLUG_EN		_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x61110)
+#define   PORTB_HOTPLUG_INT_EN			(1 << 29)
+#define   PORTC_HOTPLUG_INT_EN			(1 << 28)
+#define   PORTD_HOTPLUG_INT_EN			(1 << 27)
+#define   SDVOB_HOTPLUG_INT_EN			(1 << 26)
+#define   SDVOC_HOTPLUG_INT_EN			(1 << 25)
+#define   TV_HOTPLUG_INT_EN			(1 << 18)
+#define   CRT_HOTPLUG_INT_EN			(1 << 9)
+#define HOTPLUG_INT_EN_MASK			(PORTB_HOTPLUG_INT_EN | \
+						 PORTC_HOTPLUG_INT_EN | \
+						 PORTD_HOTPLUG_INT_EN | \
+						 SDVOC_HOTPLUG_INT_EN | \
+						 SDVOB_HOTPLUG_INT_EN | \
+						 CRT_HOTPLUG_INT_EN)
+#define   CRT_HOTPLUG_FORCE_DETECT		(1 << 3)
+#define CRT_HOTPLUG_ACTIVATION_PERIOD_32	(0 << 8)
+/* must use period 64 on GM45 according to docs */
+#define CRT_HOTPLUG_ACTIVATION_PERIOD_64	(1 << 8)
+#define CRT_HOTPLUG_DAC_ON_TIME_2M		(0 << 7)
+#define CRT_HOTPLUG_DAC_ON_TIME_4M		(1 << 7)
+#define CRT_HOTPLUG_VOLTAGE_COMPARE_40		(0 << 5)
+#define CRT_HOTPLUG_VOLTAGE_COMPARE_50		(1 << 5)
+#define CRT_HOTPLUG_VOLTAGE_COMPARE_60		(2 << 5)
+#define CRT_HOTPLUG_VOLTAGE_COMPARE_70		(3 << 5)
+#define CRT_HOTPLUG_VOLTAGE_COMPARE_MASK	(3 << 5)
+#define CRT_HOTPLUG_DETECT_DELAY_1G		(0 << 4)
+#define CRT_HOTPLUG_DETECT_DELAY_2G		(1 << 4)
+#define CRT_HOTPLUG_DETECT_VOLTAGE_325MV	(0 << 2)
+#define CRT_HOTPLUG_DETECT_VOLTAGE_475MV	(1 << 2)
+
+#define PORT_HOTPLUG_STAT	_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x61114)
+/*
+ * HDMI/DP bits are g4x+
+ *
+ * WARNING: Bspec for hpd status bits on gen4 seems to be completely confused.
+ * Please check the detailed lore in the commit message for for experimental
+ * evidence.
+ */
+/* Bspec says GM45 should match G4X/VLV/CHV, but reality disagrees */
+#define   PORTD_HOTPLUG_LIVE_STATUS_GM45	(1 << 29)
+#define   PORTC_HOTPLUG_LIVE_STATUS_GM45	(1 << 28)
+#define   PORTB_HOTPLUG_LIVE_STATUS_GM45	(1 << 27)
+/* G4X/VLV/CHV DP/HDMI bits again match Bspec */
+#define   PORTD_HOTPLUG_LIVE_STATUS_G4X		(1 << 27)
+#define   PORTC_HOTPLUG_LIVE_STATUS_G4X		(1 << 28)
+#define   PORTB_HOTPLUG_LIVE_STATUS_G4X		(1 << 29)
+#define   PORTD_HOTPLUG_INT_STATUS		(3 << 21)
+#define   PORTD_HOTPLUG_INT_LONG_PULSE		(2 << 21)
+#define   PORTD_HOTPLUG_INT_SHORT_PULSE		(1 << 21)
+#define   PORTC_HOTPLUG_INT_STATUS		(3 << 19)
+#define   PORTC_HOTPLUG_INT_LONG_PULSE		(2 << 19)
+#define   PORTC_HOTPLUG_INT_SHORT_PULSE		(1 << 19)
+#define   PORTB_HOTPLUG_INT_STATUS		(3 << 17)
+#define   PORTB_HOTPLUG_INT_LONG_PULSE		(2 << 17)
+#define   PORTB_HOTPLUG_INT_SHORT_PLUSE		(1 << 17)
+/* CRT/TV common between gen3+ */
+#define   CRT_HOTPLUG_INT_STATUS		(1 << 11)
+#define   TV_HOTPLUG_INT_STATUS			(1 << 10)
+#define   CRT_HOTPLUG_MONITOR_MASK		(3 << 8)
+#define   CRT_HOTPLUG_MONITOR_COLOR		(3 << 8)
+#define   CRT_HOTPLUG_MONITOR_MONO		(2 << 8)
+#define   CRT_HOTPLUG_MONITOR_NONE		(0 << 8)
+#define   DP_AUX_CHANNEL_D_INT_STATUS_G4X	(1 << 6)
+#define   DP_AUX_CHANNEL_C_INT_STATUS_G4X	(1 << 5)
+#define   DP_AUX_CHANNEL_B_INT_STATUS_G4X	(1 << 4)
+#define   DP_AUX_CHANNEL_MASK_INT_STATUS_G4X	(7 << 4)
+
+/* SDVO is different across gen3/4 */
+#define   SDVOC_HOTPLUG_INT_STATUS_G4X		(1 << 3)
+#define   SDVOB_HOTPLUG_INT_STATUS_G4X		(1 << 2)
+/*
+ * Bspec seems to be seriously misleaded about the SDVO hpd bits on i965g/gm,
+ * since reality corrobates that they're the same as on gen3. But keep these
+ * bits here (and the comment!) to help any other lost wanderers back onto the
+ * right tracks.
+ */
+#define   SDVOC_HOTPLUG_INT_STATUS_I965		(3 << 4)
+#define   SDVOB_HOTPLUG_INT_STATUS_I965		(3 << 2)
+#define   SDVOC_HOTPLUG_INT_STATUS_I915		(1 << 7)
+#define   SDVOB_HOTPLUG_INT_STATUS_I915		(1 << 6)
+#define   HOTPLUG_INT_STATUS_G4X		(CRT_HOTPLUG_INT_STATUS | \
+						 SDVOB_HOTPLUG_INT_STATUS_G4X | \
+						 SDVOC_HOTPLUG_INT_STATUS_G4X | \
+						 PORTB_HOTPLUG_INT_STATUS | \
+						 PORTC_HOTPLUG_INT_STATUS | \
+						 PORTD_HOTPLUG_INT_STATUS)
+
+#define HOTPLUG_INT_STATUS_I915			(CRT_HOTPLUG_INT_STATUS | \
+						 SDVOB_HOTPLUG_INT_STATUS_I915 | \
+						 SDVOC_HOTPLUG_INT_STATUS_I915 | \
+						 PORTB_HOTPLUG_INT_STATUS | \
+						 PORTC_HOTPLUG_INT_STATUS | \
+						 PORTD_HOTPLUG_INT_STATUS)
+
+/* SDVO and HDMI port control.
+ * The same register may be used for SDVO or HDMI */
+#define _GEN3_SDVOB	0x61140
+#define _GEN3_SDVOC	0x61160
+#define GEN3_SDVOB	_MMIO(_GEN3_SDVOB)
+#define GEN3_SDVOC	_MMIO(_GEN3_SDVOC)
+#define GEN4_HDMIB	GEN3_SDVOB
+#define GEN4_HDMIC	GEN3_SDVOC
+#define VLV_HDMIB	_MMIO(VLV_DISPLAY_BASE + 0x61140)
+#define VLV_HDMIC	_MMIO(VLV_DISPLAY_BASE + 0x61160)
+#define CHV_HDMID	_MMIO(VLV_DISPLAY_BASE + 0x6116C)
+#define PCH_SDVOB	_MMIO(0xe1140)
+#define PCH_HDMIB	PCH_SDVOB
+#define PCH_HDMIC	_MMIO(0xe1150)
+#define PCH_HDMID	_MMIO(0xe1160)
+
+#define PORT_DFT_I9XX				_MMIO(0x61150)
+#define   DC_BALANCE_RESET			(1 << 25)
+#define PORT_DFT2_G4X		_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x61154)
+#define   DC_BALANCE_RESET_VLV			(1 << 31)
+#define   PIPE_SCRAMBLE_RESET_MASK		((1 << 14) | (0x3 << 0))
+#define   PIPE_C_SCRAMBLE_RESET			REG_BIT(14) /* chv */
+#define   PIPE_B_SCRAMBLE_RESET			REG_BIT(1)
+#define   PIPE_A_SCRAMBLE_RESET			REG_BIT(0)
+
+/* Gen 3 SDVO bits: */
+#define   SDVO_ENABLE				(1 << 31)
+#define   SDVO_PIPE_SEL_SHIFT			30
+#define   SDVO_PIPE_SEL_MASK			(1 << 30)
+#define   SDVO_PIPE_SEL(pipe)			((pipe) << 30)
+#define   SDVO_STALL_SELECT			(1 << 29)
+#define   SDVO_INTERRUPT_ENABLE			(1 << 26)
+/*
+ * 915G/GM SDVO pixel multiplier.
+ * Programmed value is multiplier - 1, up to 5x.
+ * \sa DPLL_MD_UDI_MULTIPLIER_MASK
+ */
+#define   SDVO_PORT_MULTIPLY_MASK		(7 << 23)
+#define   SDVO_PORT_MULTIPLY_SHIFT		23
+#define   SDVO_PHASE_SELECT_MASK		(15 << 19)
+#define   SDVO_PHASE_SELECT_DEFAULT		(6 << 19)
+#define   SDVO_CLOCK_OUTPUT_INVERT		(1 << 18)
+#define   SDVOC_GANG_MODE			(1 << 16) /* Port C only */
+#define   SDVO_BORDER_ENABLE			(1 << 7) /* SDVO only */
+#define   SDVOB_PCIE_CONCURRENCY		(1 << 3) /* Port B only */
+#define   SDVO_DETECTED				(1 << 2)
+/* Bits to be preserved when writing */
+#define   SDVOB_PRESERVE_MASK ((1 << 17) | (1 << 16) | (1 << 14) | \
+			       SDVO_INTERRUPT_ENABLE)
+#define   SDVOC_PRESERVE_MASK ((1 << 17) | SDVO_INTERRUPT_ENABLE)
+
+/* Gen 4 SDVO/HDMI bits: */
+#define   SDVO_COLOR_FORMAT_8bpc		(0 << 26)
+#define   SDVO_COLOR_FORMAT_MASK		(7 << 26)
+#define   SDVO_ENCODING_SDVO			(0 << 10)
+#define   SDVO_ENCODING_HDMI			(2 << 10)
+#define   HDMI_MODE_SELECT_HDMI			(1 << 9) /* HDMI only */
+#define   HDMI_MODE_SELECT_DVI			(0 << 9) /* HDMI only */
+#define   HDMI_COLOR_RANGE_16_235		(1 << 8) /* HDMI only */
+#define   HDMI_AUDIO_ENABLE			(1 << 6) /* HDMI only */
+/* VSYNC/HSYNC bits new with 965, default is to be set */
+#define   SDVO_VSYNC_ACTIVE_HIGH		(1 << 4)
+#define   SDVO_HSYNC_ACTIVE_HIGH		(1 << 3)
+
+/* Gen 5 (IBX) SDVO/HDMI bits: */
+#define   HDMI_COLOR_FORMAT_12bpc		(3 << 26) /* HDMI only */
+#define   SDVOB_HOTPLUG_ENABLE			(1 << 23) /* SDVO only */
+
+/* Gen 6 (CPT) SDVO/HDMI bits: */
+#define   SDVO_PIPE_SEL_SHIFT_CPT		29
+#define   SDVO_PIPE_SEL_MASK_CPT		(3 << 29)
+#define   SDVO_PIPE_SEL_CPT(pipe)		((pipe) << 29)
+
+/* CHV SDVO/HDMI bits: */
+#define   SDVO_PIPE_SEL_SHIFT_CHV		24
+#define   SDVO_PIPE_SEL_MASK_CHV		(3 << 24)
+#define   SDVO_PIPE_SEL_CHV(pipe)		((pipe) << 24)
+
+
+/* DVO port control */
+#define _DVOA			0x61120
+#define DVOA			_MMIO(_DVOA)
+#define _DVOB			0x61140
+#define DVOB			_MMIO(_DVOB)
+#define _DVOC			0x61160
+#define DVOC			_MMIO(_DVOC)
+#define   DVO_ENABLE			(1 << 31)
+#define   DVO_PIPE_SEL_SHIFT		30
+#define   DVO_PIPE_SEL_MASK		(1 << 30)
+#define   DVO_PIPE_SEL(pipe)		((pipe) << 30)
+#define   DVO_PIPE_STALL_UNUSED		(0 << 28)
+#define   DVO_PIPE_STALL		(1 << 28)
+#define   DVO_PIPE_STALL_TV		(2 << 28)
+#define   DVO_PIPE_STALL_MASK		(3 << 28)
+#define   DVO_USE_VGA_SYNC		(1 << 15)
+#define   DVO_DATA_ORDER_I740		(0 << 14)
+#define   DVO_DATA_ORDER_FP		(1 << 14)
+#define   DVO_VSYNC_DISABLE		(1 << 11)
+#define   DVO_HSYNC_DISABLE		(1 << 10)
+#define   DVO_VSYNC_TRISTATE		(1 << 9)
+#define   DVO_HSYNC_TRISTATE		(1 << 8)
+#define   DVO_BORDER_ENABLE		(1 << 7)
+#define   DVO_DATA_ORDER_GBRG		(1 << 6)
+#define   DVO_DATA_ORDER_RGGB		(0 << 6)
+#define   DVO_DATA_ORDER_GBRG_ERRATA	(0 << 6)
+#define   DVO_DATA_ORDER_RGGB_ERRATA	(1 << 6)
+#define   DVO_VSYNC_ACTIVE_HIGH		(1 << 4)
+#define   DVO_HSYNC_ACTIVE_HIGH		(1 << 3)
+#define   DVO_BLANK_ACTIVE_HIGH		(1 << 2)
+#define   DVO_OUTPUT_CSTATE_PIXELS	(1 << 1)	/* SDG only */
+#define   DVO_OUTPUT_SOURCE_SIZE_PIXELS	(1 << 0)	/* SDG only */
+#define   DVO_PRESERVE_MASK		(0x7 << 24)
+#define DVOA_SRCDIM		_MMIO(0x61124)
+#define DVOB_SRCDIM		_MMIO(0x61144)
+#define DVOC_SRCDIM		_MMIO(0x61164)
+#define   DVO_SRCDIM_HORIZONTAL_SHIFT	12
+#define   DVO_SRCDIM_VERTICAL_SHIFT	0
+
+/* LVDS port control */
+#define LVDS			_MMIO(0x61180)
+/*
+ * Enables the LVDS port.  This bit must be set before DPLLs are enabled, as
+ * the DPLL semantics change when the LVDS is assigned to that pipe.
+ */
+#define   LVDS_PORT_EN			REG_BIT(31)
+/* Selects pipe B for LVDS data.  Must be set on pre-965. */
+#define   LVDS_PIPE_SEL_MASK		REG_BIT(30)
+#define   LVDS_PIPE_SEL(pipe)		REG_FIELD_PREP(LVDS_PIPE_SEL_MASK, (pipe))
+#define   LVDS_PIPE_SEL_MASK_CPT	REG_GENMASK(30, 29)
+#define   LVDS_PIPE_SEL_CPT(pipe)	REG_FIELD_PREP(LVDS_PIPE_SEL_MASK_CPT, (pipe))
+/* LVDS dithering flag on 965/g4x platform */
+#define   LVDS_ENABLE_DITHER		REG_BIT(25)
+/* LVDS sync polarity flags. Set to invert (i.e. negative) */
+#define   LVDS_VSYNC_POLARITY		REG_BIT(21)
+#define   LVDS_HSYNC_POLARITY		REG_BIT(20)
+
+/* Enable border for unscaled (or aspect-scaled) display */
+#define   LVDS_BORDER_ENABLE		REG_BIT(15)
+/*
+ * Enables the A0-A2 data pairs and CLKA, containing 18 bits of color data per
+ * pixel.
+ */
+#define   LVDS_A0A2_CLKA_POWER_MASK	REG_GENMASK(9, 8)
+#define   LVDS_A0A2_CLKA_POWER_DOWN	REG_FIELD_PREP(LVDS_A0A2_CLKA_POWER_MASK, 0)
+#define   LVDS_A0A2_CLKA_POWER_UP	REG_FIELD_PREP(LVDS_A0A2_CLKA_POWER_MASK, 3)
+/*
+ * Controls the A3 data pair, which contains the additional LSBs for 24 bit
+ * mode.  Only enabled if LVDS_A0A2_CLKA_POWER_UP also indicates it should be
+ * on.
+ */
+#define   LVDS_A3_POWER_MASK		REG_GENMASK(7, 6)
+#define   LVDS_A3_POWER_DOWN		REG_FIELD_PREP(LVDS_A3_POWER_MASK, 0)
+#define   LVDS_A3_POWER_UP		REG_FIELD_PREP(LVDS_A3_POWER_MASK, 3)
+/*
+ * Controls the CLKB pair.  This should only be set when LVDS_B0B3_POWER_UP
+ * is set.
+ */
+#define   LVDS_CLKB_POWER_MASK		REG_GENMASK(5, 4)
+#define   LVDS_CLKB_POWER_DOWN		REG_FIELD_PREP(LVDS_CLKB_POWER_MASK, 0)
+#define   LVDS_CLKB_POWER_UP		REG_FIELD_PREP(LVDS_CLKB_POWER_MASK, 3)
+/*
+ * Controls the B0-B3 data pairs.  This must be set to match the DPLL p2
+ * setting for whether we are in dual-channel mode.  The B3 pair will
+ * additionally only be powered up when LVDS_A3_POWER_UP is set.
+ */
+#define   LVDS_B0B3_POWER_MASK		REG_GENMASK(3, 2)
+#define   LVDS_B0B3_POWER_DOWN		REG_FIELD_PREP(LVDS_B0B3_POWER_MASK, 0)
+#define   LVDS_B0B3_POWER_UP		REG_FIELD_PREP(LVDS_B0B3_POWER_MASK, 3)
+
+/* Video Data Island Packet control */
+#define VIDEO_DIP_DATA		_MMIO(0x61178)
+/* Read the description of VIDEO_DIP_DATA (before Haswell) or VIDEO_DIP_ECC
+ * (Haswell and newer) to see which VIDEO_DIP_DATA byte corresponds to each byte
+ * of the infoframe structure specified by CEA-861. */
+#define   VIDEO_DIP_DATA_SIZE	32
+#define   VIDEO_DIP_GMP_DATA_SIZE	36
+#define   VIDEO_DIP_VSC_DATA_SIZE	36
+#define   VIDEO_DIP_PPS_DATA_SIZE	132
+#define VIDEO_DIP_CTL		_MMIO(0x61170)
+/* Pre HSW: */
+#define   VIDEO_DIP_ENABLE		(1 << 31)
+#define   VIDEO_DIP_PORT(port)		((port) << 29)
+#define   VIDEO_DIP_PORT_MASK		(3 << 29)
+#define   VIDEO_DIP_ENABLE_GCP		(1 << 25) /* ilk+ */
+#define   VIDEO_DIP_ENABLE_AVI		(1 << 21)
+#define   VIDEO_DIP_ENABLE_VENDOR	(2 << 21)
+#define   VIDEO_DIP_ENABLE_GAMUT	(4 << 21) /* ilk+ */
+#define   VIDEO_DIP_ENABLE_SPD		(8 << 21)
+#define   VIDEO_DIP_SELECT_AVI		(0 << 19)
+#define   VIDEO_DIP_SELECT_VENDOR	(1 << 19)
+#define   VIDEO_DIP_SELECT_GAMUT	(2 << 19)
+#define   VIDEO_DIP_SELECT_SPD		(3 << 19)
+#define   VIDEO_DIP_SELECT_MASK		(3 << 19)
+#define   VIDEO_DIP_FREQ_ONCE		(0 << 16)
+#define   VIDEO_DIP_FREQ_VSYNC		(1 << 16)
+#define   VIDEO_DIP_FREQ_2VSYNC		(2 << 16)
+#define   VIDEO_DIP_FREQ_MASK		(3 << 16)
+/* HSW and later: */
+#define   VIDEO_DIP_ENABLE_DRM_GLK	(1 << 28)
+#define   PSR_VSC_BIT_7_SET		(1 << 27)
+#define   VSC_SELECT_MASK		(0x3 << 25)
+#define   VSC_SELECT_SHIFT		25
+#define   VSC_DIP_HW_HEA_DATA		(0 << 25)
+#define   VSC_DIP_HW_HEA_SW_DATA	(1 << 25)
+#define   VSC_DIP_HW_DATA_SW_HEA	(2 << 25)
+#define   VSC_DIP_SW_HEA_DATA		(3 << 25)
+#define   VDIP_ENABLE_PPS		(1 << 24)
+#define   VIDEO_DIP_ENABLE_VSC_HSW	(1 << 20)
+#define   VIDEO_DIP_ENABLE_GCP_HSW	(1 << 16)
+#define   VIDEO_DIP_ENABLE_AVI_HSW	(1 << 12)
+#define   VIDEO_DIP_ENABLE_VS_HSW	(1 << 8)
+#define   VIDEO_DIP_ENABLE_GMP_HSW	(1 << 4)
+#define   VIDEO_DIP_ENABLE_SPD_HSW	(1 << 0)
+
+/* Panel power sequencing */
+#define PPS_BASE			0x61200
+#define VLV_PPS_BASE			(VLV_DISPLAY_BASE + PPS_BASE)
+#define PCH_PPS_BASE			0xC7200
+
+#define _MMIO_PPS(pps_idx, reg)		_MMIO(dev_priv->display.pps.mmio_base -	\
+					      PPS_BASE + (reg) +	\
+					      (pps_idx) * 0x100)
+
+#define _PP_STATUS			0x61200
+#define PP_STATUS(pps_idx)		_MMIO_PPS(pps_idx, _PP_STATUS)
+#define   PP_ON				REG_BIT(31)
+/*
+ * Indicates that all dependencies of the panel are on:
+ *
+ * - PLL enabled
+ * - pipe enabled
+ * - LVDS/DVOB/DVOC on
+ */
+#define   PP_READY			REG_BIT(30)
+#define   PP_SEQUENCE_MASK		REG_GENMASK(29, 28)
+#define   PP_SEQUENCE_NONE		REG_FIELD_PREP(PP_SEQUENCE_MASK, 0)
+#define   PP_SEQUENCE_POWER_UP		REG_FIELD_PREP(PP_SEQUENCE_MASK, 1)
+#define   PP_SEQUENCE_POWER_DOWN	REG_FIELD_PREP(PP_SEQUENCE_MASK, 2)
+#define   PP_CYCLE_DELAY_ACTIVE		REG_BIT(27)
+#define   PP_SEQUENCE_STATE_MASK	REG_GENMASK(3, 0)
+#define   PP_SEQUENCE_STATE_OFF_IDLE	REG_FIELD_PREP(PP_SEQUENCE_STATE_MASK, 0x0)
+#define   PP_SEQUENCE_STATE_OFF_S0_1	REG_FIELD_PREP(PP_SEQUENCE_STATE_MASK, 0x1)
+#define   PP_SEQUENCE_STATE_OFF_S0_2	REG_FIELD_PREP(PP_SEQUENCE_STATE_MASK, 0x2)
+#define   PP_SEQUENCE_STATE_OFF_S0_3	REG_FIELD_PREP(PP_SEQUENCE_STATE_MASK, 0x3)
+#define   PP_SEQUENCE_STATE_ON_IDLE	REG_FIELD_PREP(PP_SEQUENCE_STATE_MASK, 0x8)
+#define   PP_SEQUENCE_STATE_ON_S1_1	REG_FIELD_PREP(PP_SEQUENCE_STATE_MASK, 0x9)
+#define   PP_SEQUENCE_STATE_ON_S1_2	REG_FIELD_PREP(PP_SEQUENCE_STATE_MASK, 0xa)
+#define   PP_SEQUENCE_STATE_ON_S1_3	REG_FIELD_PREP(PP_SEQUENCE_STATE_MASK, 0xb)
+#define   PP_SEQUENCE_STATE_RESET	REG_FIELD_PREP(PP_SEQUENCE_STATE_MASK, 0xf)
+
+#define _PP_CONTROL			0x61204
+#define PP_CONTROL(pps_idx)		_MMIO_PPS(pps_idx, _PP_CONTROL)
+#define  PANEL_UNLOCK_MASK		REG_GENMASK(31, 16)
+#define  PANEL_UNLOCK_REGS		REG_FIELD_PREP(PANEL_UNLOCK_MASK, 0xabcd)
+#define  BXT_POWER_CYCLE_DELAY_MASK	REG_GENMASK(8, 4)
+#define  EDP_FORCE_VDD			REG_BIT(3)
+#define  EDP_BLC_ENABLE			REG_BIT(2)
+#define  PANEL_POWER_RESET		REG_BIT(1)
+#define  PANEL_POWER_ON			REG_BIT(0)
+
+#define _PP_ON_DELAYS			0x61208
+#define PP_ON_DELAYS(pps_idx)		_MMIO_PPS(pps_idx, _PP_ON_DELAYS)
+#define  PANEL_PORT_SELECT_MASK		REG_GENMASK(31, 30)
+#define  PANEL_PORT_SELECT_LVDS		REG_FIELD_PREP(PANEL_PORT_SELECT_MASK, 0)
+#define  PANEL_PORT_SELECT_DPA		REG_FIELD_PREP(PANEL_PORT_SELECT_MASK, 1)
+#define  PANEL_PORT_SELECT_DPC		REG_FIELD_PREP(PANEL_PORT_SELECT_MASK, 2)
+#define  PANEL_PORT_SELECT_DPD		REG_FIELD_PREP(PANEL_PORT_SELECT_MASK, 3)
+#define  PANEL_PORT_SELECT_VLV(port)	REG_FIELD_PREP(PANEL_PORT_SELECT_MASK, port)
+#define  PANEL_POWER_UP_DELAY_MASK	REG_GENMASK(28, 16)
+#define  PANEL_LIGHT_ON_DELAY_MASK	REG_GENMASK(12, 0)
+
+#define _PP_OFF_DELAYS			0x6120C
+#define PP_OFF_DELAYS(pps_idx)		_MMIO_PPS(pps_idx, _PP_OFF_DELAYS)
+#define  PANEL_POWER_DOWN_DELAY_MASK	REG_GENMASK(28, 16)
+#define  PANEL_LIGHT_OFF_DELAY_MASK	REG_GENMASK(12, 0)
+
+#define _PP_DIVISOR			0x61210
+#define PP_DIVISOR(pps_idx)		_MMIO_PPS(pps_idx, _PP_DIVISOR)
+#define  PP_REFERENCE_DIVIDER_MASK	REG_GENMASK(31, 8)
+#define  PANEL_POWER_CYCLE_DELAY_MASK	REG_GENMASK(4, 0)
+
+/* Panel fitting */
+#define PFIT_CONTROL	_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x61230)
+#define   PFIT_ENABLE		(1 << 31)
+#define   PFIT_PIPE_MASK	(3 << 29)
+#define   PFIT_PIPE_SHIFT	29
+#define   PFIT_PIPE(pipe)	((pipe) << 29)
+#define   VERT_INTERP_DISABLE	(0 << 10)
+#define   VERT_INTERP_BILINEAR	(1 << 10)
+#define   VERT_INTERP_MASK	(3 << 10)
+#define   VERT_AUTO_SCALE	(1 << 9)
+#define   HORIZ_INTERP_DISABLE	(0 << 6)
+#define   HORIZ_INTERP_BILINEAR	(1 << 6)
+#define   HORIZ_INTERP_MASK	(3 << 6)
+#define   HORIZ_AUTO_SCALE	(1 << 5)
+#define   PANEL_8TO6_DITHER_ENABLE (1 << 3)
+#define   PFIT_FILTER_FUZZY	(0 << 24)
+#define   PFIT_SCALING_AUTO	(0 << 26)
+#define   PFIT_SCALING_PROGRAMMED (1 << 26)
+#define   PFIT_SCALING_PILLAR	(2 << 26)
+#define   PFIT_SCALING_LETTER	(3 << 26)
+#define PFIT_PGM_RATIOS _MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x61234)
+/* Pre-965 */
+#define		PFIT_VERT_SCALE_SHIFT		20
+#define		PFIT_VERT_SCALE_MASK		0xfff00000
+#define		PFIT_HORIZ_SCALE_SHIFT		4
+#define		PFIT_HORIZ_SCALE_MASK		0x0000fff0
+/* 965+ */
+#define		PFIT_VERT_SCALE_SHIFT_965	16
+#define		PFIT_VERT_SCALE_MASK_965	0x1fff0000
+#define		PFIT_HORIZ_SCALE_SHIFT_965	0
+#define		PFIT_HORIZ_SCALE_MASK_965	0x00001fff
+
+#define PFIT_AUTO_RATIOS _MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x61238)
+
+#define PCH_GTC_CTL		_MMIO(0xe7000)
+#define   PCH_GTC_ENABLE	(1 << 31)
+
+/* TV port control */
+#define TV_CTL			_MMIO(0x68000)
+/* Enables the TV encoder */
+# define TV_ENC_ENABLE			(1 << 31)
+/* Sources the TV encoder input from pipe B instead of A. */
+# define TV_ENC_PIPE_SEL_SHIFT		30
+# define TV_ENC_PIPE_SEL_MASK		(1 << 30)
+# define TV_ENC_PIPE_SEL(pipe)		((pipe) << 30)
+/* Outputs composite video (DAC A only) */
+# define TV_ENC_OUTPUT_COMPOSITE	(0 << 28)
+/* Outputs SVideo video (DAC B/C) */
+# define TV_ENC_OUTPUT_SVIDEO		(1 << 28)
+/* Outputs Component video (DAC A/B/C) */
+# define TV_ENC_OUTPUT_COMPONENT	(2 << 28)
+/* Outputs Composite and SVideo (DAC A/B/C) */
+# define TV_ENC_OUTPUT_SVIDEO_COMPOSITE	(3 << 28)
+# define TV_TRILEVEL_SYNC		(1 << 21)
+/* Enables slow sync generation (945GM only) */
+# define TV_SLOW_SYNC			(1 << 20)
+/* Selects 4x oversampling for 480i and 576p */
+# define TV_OVERSAMPLE_4X		(0 << 18)
+/* Selects 2x oversampling for 720p and 1080i */
+# define TV_OVERSAMPLE_2X		(1 << 18)
+/* Selects no oversampling for 1080p */
+# define TV_OVERSAMPLE_NONE		(2 << 18)
+/* Selects 8x oversampling */
+# define TV_OVERSAMPLE_8X		(3 << 18)
+# define TV_OVERSAMPLE_MASK		(3 << 18)
+/* Selects progressive mode rather than interlaced */
+# define TV_PROGRESSIVE			(1 << 17)
+/* Sets the colorburst to PAL mode.  Required for non-M PAL modes. */
+# define TV_PAL_BURST			(1 << 16)
+/* Field for setting delay of Y compared to C */
+# define TV_YC_SKEW_MASK		(7 << 12)
+/* Enables a fix for 480p/576p standard definition modes on the 915GM only */
+# define TV_ENC_SDP_FIX			(1 << 11)
+/*
+ * Enables a fix for the 915GM only.
+ *
+ * Not sure what it does.
+ */
+# define TV_ENC_C0_FIX			(1 << 10)
+/* Bits that must be preserved by software */
+# define TV_CTL_SAVE			((1 << 11) | (3 << 9) | (7 << 6) | 0xf)
+# define TV_FUSE_STATE_MASK		(3 << 4)
+/* Read-only state that reports all features enabled */
+# define TV_FUSE_STATE_ENABLED		(0 << 4)
+/* Read-only state that reports that Macrovision is disabled in hardware*/
+# define TV_FUSE_STATE_NO_MACROVISION	(1 << 4)
+/* Read-only state that reports that TV-out is disabled in hardware. */
+# define TV_FUSE_STATE_DISABLED		(2 << 4)
+/* Normal operation */
+# define TV_TEST_MODE_NORMAL		(0 << 0)
+/* Encoder test pattern 1 - combo pattern */
+# define TV_TEST_MODE_PATTERN_1		(1 << 0)
+/* Encoder test pattern 2 - full screen vertical 75% color bars */
+# define TV_TEST_MODE_PATTERN_2		(2 << 0)
+/* Encoder test pattern 3 - full screen horizontal 75% color bars */
+# define TV_TEST_MODE_PATTERN_3		(3 << 0)
+/* Encoder test pattern 4 - random noise */
+# define TV_TEST_MODE_PATTERN_4		(4 << 0)
+/* Encoder test pattern 5 - linear color ramps */
+# define TV_TEST_MODE_PATTERN_5		(5 << 0)
+/*
+ * This test mode forces the DACs to 50% of full output.
+ *
+ * This is used for load detection in combination with TVDAC_SENSE_MASK
+ */
+# define TV_TEST_MODE_MONITOR_DETECT	(7 << 0)
+# define TV_TEST_MODE_MASK		(7 << 0)
+
+#define TV_DAC			_MMIO(0x68004)
+# define TV_DAC_SAVE		0x00ffff00
+/*
+ * Reports that DAC state change logic has reported change (RO).
+ *
+ * This gets cleared when TV_DAC_STATE_EN is cleared
+*/
+# define TVDAC_STATE_CHG		(1 << 31)
+# define TVDAC_SENSE_MASK		(7 << 28)
+/* Reports that DAC A voltage is above the detect threshold */
+# define TVDAC_A_SENSE			(1 << 30)
+/* Reports that DAC B voltage is above the detect threshold */
+# define TVDAC_B_SENSE			(1 << 29)
+/* Reports that DAC C voltage is above the detect threshold */
+# define TVDAC_C_SENSE			(1 << 28)
+/*
+ * Enables DAC state detection logic, for load-based TV detection.
+ *
+ * The PLL of the chosen pipe (in TV_CTL) must be running, and the encoder set
+ * to off, for load detection to work.
+ */
+# define TVDAC_STATE_CHG_EN		(1 << 27)
+/* Sets the DAC A sense value to high */
+# define TVDAC_A_SENSE_CTL		(1 << 26)
+/* Sets the DAC B sense value to high */
+# define TVDAC_B_SENSE_CTL		(1 << 25)
+/* Sets the DAC C sense value to high */
+# define TVDAC_C_SENSE_CTL		(1 << 24)
+/* Overrides the ENC_ENABLE and DAC voltage levels */
+# define DAC_CTL_OVERRIDE		(1 << 7)
+/* Sets the slew rate.  Must be preserved in software */
+# define ENC_TVDAC_SLEW_FAST		(1 << 6)
+# define DAC_A_1_3_V			(0 << 4)
+# define DAC_A_1_1_V			(1 << 4)
+# define DAC_A_0_7_V			(2 << 4)
+# define DAC_A_MASK			(3 << 4)
+# define DAC_B_1_3_V			(0 << 2)
+# define DAC_B_1_1_V			(1 << 2)
+# define DAC_B_0_7_V			(2 << 2)
+# define DAC_B_MASK			(3 << 2)
+# define DAC_C_1_3_V			(0 << 0)
+# define DAC_C_1_1_V			(1 << 0)
+# define DAC_C_0_7_V			(2 << 0)
+# define DAC_C_MASK			(3 << 0)
+
+/*
+ * CSC coefficients are stored in a floating point format with 9 bits of
+ * mantissa and 2 or 3 bits of exponent.  The exponent is represented as 2**-n,
+ * where 2-bit exponents are unsigned n, and 3-bit exponents are signed n with
+ * -1 (0x3) being the only legal negative value.
+ */
+#define TV_CSC_Y		_MMIO(0x68010)
+# define TV_RY_MASK			0x07ff0000
+# define TV_RY_SHIFT			16
+# define TV_GY_MASK			0x00000fff
+# define TV_GY_SHIFT			0
+
+#define TV_CSC_Y2		_MMIO(0x68014)
+# define TV_BY_MASK			0x07ff0000
+# define TV_BY_SHIFT			16
+/*
+ * Y attenuation for component video.
+ *
+ * Stored in 1.9 fixed point.
+ */
+# define TV_AY_MASK			0x000003ff
+# define TV_AY_SHIFT			0
+
+#define TV_CSC_U		_MMIO(0x68018)
+# define TV_RU_MASK			0x07ff0000
+# define TV_RU_SHIFT			16
+# define TV_GU_MASK			0x000007ff
+# define TV_GU_SHIFT			0
+
+#define TV_CSC_U2		_MMIO(0x6801c)
+# define TV_BU_MASK			0x07ff0000
+# define TV_BU_SHIFT			16
+/*
+ * U attenuation for component video.
+ *
+ * Stored in 1.9 fixed point.
+ */
+# define TV_AU_MASK			0x000003ff
+# define TV_AU_SHIFT			0
+
+#define TV_CSC_V		_MMIO(0x68020)
+# define TV_RV_MASK			0x0fff0000
+# define TV_RV_SHIFT			16
+# define TV_GV_MASK			0x000007ff
+# define TV_GV_SHIFT			0
+
+#define TV_CSC_V2		_MMIO(0x68024)
+# define TV_BV_MASK			0x07ff0000
+# define TV_BV_SHIFT			16
+/*
+ * V attenuation for component video.
+ *
+ * Stored in 1.9 fixed point.
+ */
+# define TV_AV_MASK			0x000007ff
+# define TV_AV_SHIFT			0
+
+#define TV_CLR_KNOBS		_MMIO(0x68028)
+/* 2s-complement brightness adjustment */
+# define TV_BRIGHTNESS_MASK		0xff000000
+# define TV_BRIGHTNESS_SHIFT		24
+/* Contrast adjustment, as a 2.6 unsigned floating point number */
+# define TV_CONTRAST_MASK		0x00ff0000
+# define TV_CONTRAST_SHIFT		16
+/* Saturation adjustment, as a 2.6 unsigned floating point number */
+# define TV_SATURATION_MASK		0x0000ff00
+# define TV_SATURATION_SHIFT		8
+/* Hue adjustment, as an integer phase angle in degrees */
+# define TV_HUE_MASK			0x000000ff
+# define TV_HUE_SHIFT			0
+
+#define TV_CLR_LEVEL		_MMIO(0x6802c)
+/* Controls the DAC level for black */
+# define TV_BLACK_LEVEL_MASK		0x01ff0000
+# define TV_BLACK_LEVEL_SHIFT		16
+/* Controls the DAC level for blanking */
+# define TV_BLANK_LEVEL_MASK		0x000001ff
+# define TV_BLANK_LEVEL_SHIFT		0
+
+#define TV_H_CTL_1		_MMIO(0x68030)
+/* Number of pixels in the hsync. */
+# define TV_HSYNC_END_MASK		0x1fff0000
+# define TV_HSYNC_END_SHIFT		16
+/* Total number of pixels minus one in the line (display and blanking). */
+# define TV_HTOTAL_MASK			0x00001fff
+# define TV_HTOTAL_SHIFT		0
+
+#define TV_H_CTL_2		_MMIO(0x68034)
+/* Enables the colorburst (needed for non-component color) */
+# define TV_BURST_ENA			(1 << 31)
+/* Offset of the colorburst from the start of hsync, in pixels minus one. */
+# define TV_HBURST_START_SHIFT		16
+# define TV_HBURST_START_MASK		0x1fff0000
+/* Length of the colorburst */
+# define TV_HBURST_LEN_SHIFT		0
+# define TV_HBURST_LEN_MASK		0x0001fff
+
+#define TV_H_CTL_3		_MMIO(0x68038)
+/* End of hblank, measured in pixels minus one from start of hsync */
+# define TV_HBLANK_END_SHIFT		16
+# define TV_HBLANK_END_MASK		0x1fff0000
+/* Start of hblank, measured in pixels minus one from start of hsync */
+# define TV_HBLANK_START_SHIFT		0
+# define TV_HBLANK_START_MASK		0x0001fff
+
+#define TV_V_CTL_1		_MMIO(0x6803c)
+/* XXX */
+# define TV_NBR_END_SHIFT		16
+# define TV_NBR_END_MASK		0x07ff0000
+/* XXX */
+# define TV_VI_END_F1_SHIFT		8
+# define TV_VI_END_F1_MASK		0x00003f00
+/* XXX */
+# define TV_VI_END_F2_SHIFT		0
+# define TV_VI_END_F2_MASK		0x0000003f
+
+#define TV_V_CTL_2		_MMIO(0x68040)
+/* Length of vsync, in half lines */
+# define TV_VSYNC_LEN_MASK		0x07ff0000
+# define TV_VSYNC_LEN_SHIFT		16
+/* Offset of the start of vsync in field 1, measured in one less than the
+ * number of half lines.
+ */
+# define TV_VSYNC_START_F1_MASK		0x00007f00
+# define TV_VSYNC_START_F1_SHIFT	8
+/*
+ * Offset of the start of vsync in field 2, measured in one less than the
+ * number of half lines.
+ */
+# define TV_VSYNC_START_F2_MASK		0x0000007f
+# define TV_VSYNC_START_F2_SHIFT	0
+
+#define TV_V_CTL_3		_MMIO(0x68044)
+/* Enables generation of the equalization signal */
+# define TV_EQUAL_ENA			(1 << 31)
+/* Length of vsync, in half lines */
+# define TV_VEQ_LEN_MASK		0x007f0000
+# define TV_VEQ_LEN_SHIFT		16
+/* Offset of the start of equalization in field 1, measured in one less than
+ * the number of half lines.
+ */
+# define TV_VEQ_START_F1_MASK		0x0007f00
+# define TV_VEQ_START_F1_SHIFT		8
+/*
+ * Offset of the start of equalization in field 2, measured in one less than
+ * the number of half lines.
+ */
+# define TV_VEQ_START_F2_MASK		0x000007f
+# define TV_VEQ_START_F2_SHIFT		0
+
+#define TV_V_CTL_4		_MMIO(0x68048)
+/*
+ * Offset to start of vertical colorburst, measured in one less than the
+ * number of lines from vertical start.
+ */
+# define TV_VBURST_START_F1_MASK	0x003f0000
+# define TV_VBURST_START_F1_SHIFT	16
+/*
+ * Offset to the end of vertical colorburst, measured in one less than the
+ * number of lines from the start of NBR.
+ */
+# define TV_VBURST_END_F1_MASK		0x000000ff
+# define TV_VBURST_END_F1_SHIFT		0
+
+#define TV_V_CTL_5		_MMIO(0x6804c)
+/*
+ * Offset to start of vertical colorburst, measured in one less than the
+ * number of lines from vertical start.
+ */
+# define TV_VBURST_START_F2_MASK	0x003f0000
+# define TV_VBURST_START_F2_SHIFT	16
+/*
+ * Offset to the end of vertical colorburst, measured in one less than the
+ * number of lines from the start of NBR.
+ */
+# define TV_VBURST_END_F2_MASK		0x000000ff
+# define TV_VBURST_END_F2_SHIFT		0
+
+#define TV_V_CTL_6		_MMIO(0x68050)
+/*
+ * Offset to start of vertical colorburst, measured in one less than the
+ * number of lines from vertical start.
+ */
+# define TV_VBURST_START_F3_MASK	0x003f0000
+# define TV_VBURST_START_F3_SHIFT	16
+/*
+ * Offset to the end of vertical colorburst, measured in one less than the
+ * number of lines from the start of NBR.
+ */
+# define TV_VBURST_END_F3_MASK		0x000000ff
+# define TV_VBURST_END_F3_SHIFT		0
+
+#define TV_V_CTL_7		_MMIO(0x68054)
+/*
+ * Offset to start of vertical colorburst, measured in one less than the
+ * number of lines from vertical start.
+ */
+# define TV_VBURST_START_F4_MASK	0x003f0000
+# define TV_VBURST_START_F4_SHIFT	16
+/*
+ * Offset to the end of vertical colorburst, measured in one less than the
+ * number of lines from the start of NBR.
+ */
+# define TV_VBURST_END_F4_MASK		0x000000ff
+# define TV_VBURST_END_F4_SHIFT		0
+
+#define TV_SC_CTL_1		_MMIO(0x68060)
+/* Turns on the first subcarrier phase generation DDA */
+# define TV_SC_DDA1_EN			(1 << 31)
+/* Turns on the first subcarrier phase generation DDA */
+# define TV_SC_DDA2_EN			(1 << 30)
+/* Turns on the first subcarrier phase generation DDA */
+# define TV_SC_DDA3_EN			(1 << 29)
+/* Sets the subcarrier DDA to reset frequency every other field */
+# define TV_SC_RESET_EVERY_2		(0 << 24)
+/* Sets the subcarrier DDA to reset frequency every fourth field */
+# define TV_SC_RESET_EVERY_4		(1 << 24)
+/* Sets the subcarrier DDA to reset frequency every eighth field */
+# define TV_SC_RESET_EVERY_8		(2 << 24)
+/* Sets the subcarrier DDA to never reset the frequency */
+# define TV_SC_RESET_NEVER		(3 << 24)
+/* Sets the peak amplitude of the colorburst.*/
+# define TV_BURST_LEVEL_MASK		0x00ff0000
+# define TV_BURST_LEVEL_SHIFT		16
+/* Sets the increment of the first subcarrier phase generation DDA */
+# define TV_SCDDA1_INC_MASK		0x00000fff
+# define TV_SCDDA1_INC_SHIFT		0
+
+#define TV_SC_CTL_2		_MMIO(0x68064)
+/* Sets the rollover for the second subcarrier phase generation DDA */
+# define TV_SCDDA2_SIZE_MASK		0x7fff0000
+# define TV_SCDDA2_SIZE_SHIFT		16
+/* Sets the increent of the second subcarrier phase generation DDA */
+# define TV_SCDDA2_INC_MASK		0x00007fff
+# define TV_SCDDA2_INC_SHIFT		0
+
+#define TV_SC_CTL_3		_MMIO(0x68068)
+/* Sets the rollover for the third subcarrier phase generation DDA */
+# define TV_SCDDA3_SIZE_MASK		0x7fff0000
+# define TV_SCDDA3_SIZE_SHIFT		16
+/* Sets the increent of the third subcarrier phase generation DDA */
+# define TV_SCDDA3_INC_MASK		0x00007fff
+# define TV_SCDDA3_INC_SHIFT		0
+
+#define TV_WIN_POS		_MMIO(0x68070)
+/* X coordinate of the display from the start of horizontal active */
+# define TV_XPOS_MASK			0x1fff0000
+# define TV_XPOS_SHIFT			16
+/* Y coordinate of the display from the start of vertical active (NBR) */
+# define TV_YPOS_MASK			0x00000fff
+# define TV_YPOS_SHIFT			0
+
+#define TV_WIN_SIZE		_MMIO(0x68074)
+/* Horizontal size of the display window, measured in pixels*/
+# define TV_XSIZE_MASK			0x1fff0000
+# define TV_XSIZE_SHIFT			16
+/*
+ * Vertical size of the display window, measured in pixels.
+ *
+ * Must be even for interlaced modes.
+ */
+# define TV_YSIZE_MASK			0x00000fff
+# define TV_YSIZE_SHIFT			0
+
+#define TV_FILTER_CTL_1		_MMIO(0x68080)
+/*
+ * Enables automatic scaling calculation.
+ *
+ * If set, the rest of the registers are ignored, and the calculated values can
+ * be read back from the register.
+ */
+# define TV_AUTO_SCALE			(1 << 31)
+/*
+ * Disables the vertical filter.
+ *
+ * This is required on modes more than 1024 pixels wide */
+# define TV_V_FILTER_BYPASS		(1 << 29)
+/* Enables adaptive vertical filtering */
+# define TV_VADAPT			(1 << 28)
+# define TV_VADAPT_MODE_MASK		(3 << 26)
+/* Selects the least adaptive vertical filtering mode */
+# define TV_VADAPT_MODE_LEAST		(0 << 26)
+/* Selects the moderately adaptive vertical filtering mode */
+# define TV_VADAPT_MODE_MODERATE	(1 << 26)
+/* Selects the most adaptive vertical filtering mode */
+# define TV_VADAPT_MODE_MOST		(3 << 26)
+/*
+ * Sets the horizontal scaling factor.
+ *
+ * This should be the fractional part of the horizontal scaling factor divided
+ * by the oversampling rate.  TV_HSCALE should be less than 1, and set to:
+ *
+ * (src width - 1) / ((oversample * dest width) - 1)
+ */
+# define TV_HSCALE_FRAC_MASK		0x00003fff
+# define TV_HSCALE_FRAC_SHIFT		0
+
+#define TV_FILTER_CTL_2		_MMIO(0x68084)
+/*
+ * Sets the integer part of the 3.15 fixed-point vertical scaling factor.
+ *
+ * TV_VSCALE should be (src height - 1) / ((interlace * dest height) - 1)
+ */
+# define TV_VSCALE_INT_MASK		0x00038000
+# define TV_VSCALE_INT_SHIFT		15
+/*
+ * Sets the fractional part of the 3.15 fixed-point vertical scaling factor.
+ *
+ * \sa TV_VSCALE_INT_MASK
+ */
+# define TV_VSCALE_FRAC_MASK		0x00007fff
+# define TV_VSCALE_FRAC_SHIFT		0
+
+#define TV_FILTER_CTL_3		_MMIO(0x68088)
+/*
+ * Sets the integer part of the 3.15 fixed-point vertical scaling factor.
+ *
+ * TV_VSCALE should be (src height - 1) / (1/4 * (dest height - 1))
+ *
+ * For progressive modes, TV_VSCALE_IP_INT should be set to zeroes.
+ */
+# define TV_VSCALE_IP_INT_MASK		0x00038000
+# define TV_VSCALE_IP_INT_SHIFT		15
+/*
+ * Sets the fractional part of the 3.15 fixed-point vertical scaling factor.
+ *
+ * For progressive modes, TV_VSCALE_IP_INT should be set to zeroes.
+ *
+ * \sa TV_VSCALE_IP_INT_MASK
+ */
+# define TV_VSCALE_IP_FRAC_MASK		0x00007fff
+# define TV_VSCALE_IP_FRAC_SHIFT		0
+
+#define TV_CC_CONTROL		_MMIO(0x68090)
+# define TV_CC_ENABLE			(1 << 31)
+/*
+ * Specifies which field to send the CC data in.
+ *
+ * CC data is usually sent in field 0.
+ */
+# define TV_CC_FID_MASK			(1 << 27)
+# define TV_CC_FID_SHIFT		27
+/* Sets the horizontal position of the CC data.  Usually 135. */
+# define TV_CC_HOFF_MASK		0x03ff0000
+# define TV_CC_HOFF_SHIFT		16
+/* Sets the vertical position of the CC data.  Usually 21 */
+# define TV_CC_LINE_MASK		0x0000003f
+# define TV_CC_LINE_SHIFT		0
+
+#define TV_CC_DATA		_MMIO(0x68094)
+# define TV_CC_RDY			(1 << 31)
+/* Second word of CC data to be transmitted. */
+# define TV_CC_DATA_2_MASK		0x007f0000
+# define TV_CC_DATA_2_SHIFT		16
+/* First word of CC data to be transmitted. */
+# define TV_CC_DATA_1_MASK		0x0000007f
+# define TV_CC_DATA_1_SHIFT		0
+
+#define TV_H_LUMA(i)		_MMIO(0x68100 + (i) * 4) /* 60 registers */
+#define TV_H_CHROMA(i)		_MMIO(0x68200 + (i) * 4) /* 60 registers */
+#define TV_V_LUMA(i)		_MMIO(0x68300 + (i) * 4) /* 43 registers */
+#define TV_V_CHROMA(i)		_MMIO(0x68400 + (i) * 4) /* 43 registers */
+
+/* Display Port */
+#define DP_A			_MMIO(0x64000) /* eDP */
+#define DP_B			_MMIO(0x64100)
+#define DP_C			_MMIO(0x64200)
+#define DP_D			_MMIO(0x64300)
+
+#define VLV_DP_B		_MMIO(VLV_DISPLAY_BASE + 0x64100)
+#define VLV_DP_C		_MMIO(VLV_DISPLAY_BASE + 0x64200)
+#define CHV_DP_D		_MMIO(VLV_DISPLAY_BASE + 0x64300)
+
+#define   DP_PORT_EN			(1 << 31)
+#define   DP_PIPE_SEL_SHIFT		30
+#define   DP_PIPE_SEL_MASK		(1 << 30)
+#define   DP_PIPE_SEL(pipe)		((pipe) << 30)
+#define   DP_PIPE_SEL_SHIFT_IVB		29
+#define   DP_PIPE_SEL_MASK_IVB		(3 << 29)
+#define   DP_PIPE_SEL_IVB(pipe)		((pipe) << 29)
+#define   DP_PIPE_SEL_SHIFT_CHV		16
+#define   DP_PIPE_SEL_MASK_CHV		(3 << 16)
+#define   DP_PIPE_SEL_CHV(pipe)		((pipe) << 16)
+
+/* Link training mode - select a suitable mode for each stage */
+#define   DP_LINK_TRAIN_PAT_1		(0 << 28)
+#define   DP_LINK_TRAIN_PAT_2		(1 << 28)
+#define   DP_LINK_TRAIN_PAT_IDLE	(2 << 28)
+#define   DP_LINK_TRAIN_OFF		(3 << 28)
+#define   DP_LINK_TRAIN_MASK		(3 << 28)
+#define   DP_LINK_TRAIN_SHIFT		28
+
+/* CPT Link training mode */
+#define   DP_LINK_TRAIN_PAT_1_CPT	(0 << 8)
+#define   DP_LINK_TRAIN_PAT_2_CPT	(1 << 8)
+#define   DP_LINK_TRAIN_PAT_IDLE_CPT	(2 << 8)
+#define   DP_LINK_TRAIN_OFF_CPT		(3 << 8)
+#define   DP_LINK_TRAIN_MASK_CPT	(7 << 8)
+#define   DP_LINK_TRAIN_SHIFT_CPT	8
+
+/* Signal voltages. These are mostly controlled by the other end */
+#define   DP_VOLTAGE_0_4		(0 << 25)
+#define   DP_VOLTAGE_0_6		(1 << 25)
+#define   DP_VOLTAGE_0_8		(2 << 25)
+#define   DP_VOLTAGE_1_2		(3 << 25)
+#define   DP_VOLTAGE_MASK		(7 << 25)
+#define   DP_VOLTAGE_SHIFT		25
+
+/* Signal pre-emphasis levels, like voltages, the other end tells us what
+ * they want
+ */
+#define   DP_PRE_EMPHASIS_0		(0 << 22)
+#define   DP_PRE_EMPHASIS_3_5		(1 << 22)
+#define   DP_PRE_EMPHASIS_6		(2 << 22)
+#define   DP_PRE_EMPHASIS_9_5		(3 << 22)
+#define   DP_PRE_EMPHASIS_MASK		(7 << 22)
+#define   DP_PRE_EMPHASIS_SHIFT		22
+
+/* How many wires to use. I guess 3 was too hard */
+#define   DP_PORT_WIDTH(width)		(((width) - 1) << 19)
+#define   DP_PORT_WIDTH_MASK		(7 << 19)
+#define   DP_PORT_WIDTH_SHIFT		19
+
+/* Mystic DPCD version 1.1 special mode */
+#define   DP_ENHANCED_FRAMING		(1 << 18)
+
+/* eDP */
+#define   DP_PLL_FREQ_270MHZ		(0 << 16)
+#define   DP_PLL_FREQ_162MHZ		(1 << 16)
+#define   DP_PLL_FREQ_MASK		(3 << 16)
+
+/* locked once port is enabled */
+#define   DP_PORT_REVERSAL		(1 << 15)
+
+/* eDP */
+#define   DP_PLL_ENABLE			(1 << 14)
+
+/* sends the clock on lane 15 of the PEG for debug */
+#define   DP_CLOCK_OUTPUT_ENABLE	(1 << 13)
+
+#define   DP_SCRAMBLING_DISABLE		(1 << 12)
+#define   DP_SCRAMBLING_DISABLE_IRONLAKE	(1 << 7)
+
+/* limit RGB values to avoid confusing TVs */
+#define   DP_COLOR_RANGE_16_235		(1 << 8)
+
+/* Turn on the audio link */
+#define   DP_AUDIO_OUTPUT_ENABLE	(1 << 6)
+
+/* vs and hs sync polarity */
+#define   DP_SYNC_VS_HIGH		(1 << 4)
+#define   DP_SYNC_HS_HIGH		(1 << 3)
+
+/* A fantasy */
+#define   DP_DETECTED			(1 << 2)
+
+/* The aux channel provides a way to talk to the
+ * signal sink for DDC etc. Max packet size supported
+ * is 20 bytes in each direction, hence the 5 fixed
+ * data registers
+ */
+#define _DPA_AUX_CH_CTL		(DISPLAY_MMIO_BASE(dev_priv) + 0x64010)
+#define _DPA_AUX_CH_DATA1	(DISPLAY_MMIO_BASE(dev_priv) + 0x64014)
+
+#define _DPB_AUX_CH_CTL		(DISPLAY_MMIO_BASE(dev_priv) + 0x64110)
+#define _DPB_AUX_CH_DATA1	(DISPLAY_MMIO_BASE(dev_priv) + 0x64114)
+
+#define DP_AUX_CH_CTL(aux_ch)	_MMIO_PORT(aux_ch, _DPA_AUX_CH_CTL, _DPB_AUX_CH_CTL)
+#define DP_AUX_CH_DATA(aux_ch, i)	_MMIO(_PORT(aux_ch, _DPA_AUX_CH_DATA1, _DPB_AUX_CH_DATA1) + (i) * 4) /* 5 registers */
+
+#define _XELPDP_USBC1_AUX_CH_CTL	0x16F210
+#define _XELPDP_USBC2_AUX_CH_CTL	0x16F410
+#define _XELPDP_USBC3_AUX_CH_CTL	0x16F610
+#define _XELPDP_USBC4_AUX_CH_CTL	0x16F810
+
+#define XELPDP_DP_AUX_CH_CTL(aux_ch)		_MMIO(_PICK(aux_ch, \
+						       _DPA_AUX_CH_CTL, \
+						       _DPB_AUX_CH_CTL, \
+						       0, /* port/aux_ch C is non-existent */ \
+						       _XELPDP_USBC1_AUX_CH_CTL, \
+						       _XELPDP_USBC2_AUX_CH_CTL, \
+						       _XELPDP_USBC3_AUX_CH_CTL, \
+						       _XELPDP_USBC4_AUX_CH_CTL))
+
+#define _XELPDP_USBC1_AUX_CH_DATA1      0x16F214
+#define _XELPDP_USBC2_AUX_CH_DATA1      0x16F414
+#define _XELPDP_USBC3_AUX_CH_DATA1      0x16F614
+#define _XELPDP_USBC4_AUX_CH_DATA1      0x16F814
+
+#define XELPDP_DP_AUX_CH_DATA(aux_ch, i)	_MMIO(_PICK(aux_ch, \
+						       _DPA_AUX_CH_DATA1, \
+						       _DPB_AUX_CH_DATA1, \
+						       0, /* port/aux_ch C is non-existent */ \
+						       _XELPDP_USBC1_AUX_CH_DATA1, \
+						       _XELPDP_USBC2_AUX_CH_DATA1, \
+						       _XELPDP_USBC3_AUX_CH_DATA1, \
+						       _XELPDP_USBC4_AUX_CH_DATA1) + (i) * 4)
+
+#define   DP_AUX_CH_CTL_SEND_BUSY	    (1 << 31)
+#define   DP_AUX_CH_CTL_DONE		    (1 << 30)
+#define   DP_AUX_CH_CTL_INTERRUPT	    (1 << 29)
+#define   DP_AUX_CH_CTL_TIME_OUT_ERROR	    (1 << 28)
+#define   DP_AUX_CH_CTL_TIME_OUT_400us	    (0 << 26)
+#define   DP_AUX_CH_CTL_TIME_OUT_600us	    (1 << 26)
+#define   DP_AUX_CH_CTL_TIME_OUT_800us	    (2 << 26)
+#define   DP_AUX_CH_CTL_TIME_OUT_MAX	    (3 << 26) /* Varies per platform */
+#define   DP_AUX_CH_CTL_TIME_OUT_MASK	    (3 << 26)
+#define   DP_AUX_CH_CTL_RECEIVE_ERROR	    (1 << 25)
+#define   DP_AUX_CH_CTL_MESSAGE_SIZE_MASK    (0x1f << 20)
+#define   DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT   20
+#define   XELPDP_DP_AUX_CH_CTL_POWER_REQUEST REG_BIT(19)
+#define   XELPDP_DP_AUX_CH_CTL_POWER_STATUS  REG_BIT(18)
+#define   DP_AUX_CH_CTL_PRECHARGE_2US_MASK   (0xf << 16)
+#define   DP_AUX_CH_CTL_PRECHARGE_2US_SHIFT  16
+#define   DP_AUX_CH_CTL_AUX_AKSV_SELECT	    (1 << 15)
+#define   DP_AUX_CH_CTL_MANCHESTER_TEST	    (1 << 14)
+#define   DP_AUX_CH_CTL_SYNC_TEST	    (1 << 13)
+#define   DP_AUX_CH_CTL_DEGLITCH_TEST	    (1 << 12)
+#define   DP_AUX_CH_CTL_PRECHARGE_TEST	    (1 << 11)
+#define   DP_AUX_CH_CTL_BIT_CLOCK_2X_MASK    (0x7ff)
+#define   DP_AUX_CH_CTL_BIT_CLOCK_2X_SHIFT   0
+#define   DP_AUX_CH_CTL_PSR_DATA_AUX_REG_SKL	(1 << 14)
+#define   DP_AUX_CH_CTL_FS_DATA_AUX_REG_SKL	(1 << 13)
+#define   DP_AUX_CH_CTL_GTC_DATA_AUX_REG_SKL	(1 << 12)
+#define   DP_AUX_CH_CTL_TBT_IO			(1 << 11)
+#define   DP_AUX_CH_CTL_FW_SYNC_PULSE_SKL_MASK (0x1f << 5)
+#define   DP_AUX_CH_CTL_FW_SYNC_PULSE_SKL(c) (((c) - 1) << 5)
+#define   DP_AUX_CH_CTL_SYNC_PULSE_SKL(c)   ((c) - 1)
+
+/*
+ * Computing GMCH M and N values for the Display Port link
+ *
+ * GMCH M/N = dot clock * bytes per pixel / ls_clk * # of lanes
+ *
+ * ls_clk (we assume) is the DP link clock (1.62 or 2.7 GHz)
+ *
+ * The GMCH value is used internally
+ *
+ * bytes_per_pixel is the number of bytes coming out of the plane,
+ * which is after the LUTs, so we want the bytes for our color format.
+ * For our current usage, this is always 3, one byte for R, G and B.
+ */
+#define _PIPEA_DATA_M_G4X	0x70050
+#define _PIPEB_DATA_M_G4X	0x71050
+
+/* Transfer unit size for display port - 1, default is 0x3f (for TU size 64) */
+#define  TU_SIZE_MASK		REG_GENMASK(30, 25)
+#define  TU_SIZE(x)		REG_FIELD_PREP(TU_SIZE_MASK, (x) - 1) /* default size 64 */
+
+#define  DATA_LINK_M_N_MASK	REG_GENMASK(23, 0)
+#define  DATA_LINK_N_MAX	(0x800000)
+
+#define _PIPEA_DATA_N_G4X	0x70054
+#define _PIPEB_DATA_N_G4X	0x71054
+
+/*
+ * Computing Link M and N values for the Display Port link
+ *
+ * Link M / N = pixel_clock / ls_clk
+ *
+ * (the DP spec calls pixel_clock the 'strm_clk')
+ *
+ * The Link value is transmitted in the Main Stream
+ * Attributes and VB-ID.
+ */
+
+#define _PIPEA_LINK_M_G4X	0x70060
+#define _PIPEB_LINK_M_G4X	0x71060
+#define _PIPEA_LINK_N_G4X	0x70064
+#define _PIPEB_LINK_N_G4X	0x71064
+
+#define PIPE_DATA_M_G4X(pipe) _MMIO_PIPE(pipe, _PIPEA_DATA_M_G4X, _PIPEB_DATA_M_G4X)
+#define PIPE_DATA_N_G4X(pipe) _MMIO_PIPE(pipe, _PIPEA_DATA_N_G4X, _PIPEB_DATA_N_G4X)
+#define PIPE_LINK_M_G4X(pipe) _MMIO_PIPE(pipe, _PIPEA_LINK_M_G4X, _PIPEB_LINK_M_G4X)
+#define PIPE_LINK_N_G4X(pipe) _MMIO_PIPE(pipe, _PIPEA_LINK_N_G4X, _PIPEB_LINK_N_G4X)
+
+/* Display & cursor control */
+
+/* Pipe A */
+#define _PIPEADSL		0x70000
+#define   PIPEDSL_CURR_FIELD	REG_BIT(31) /* ctg+ */
+#define   PIPEDSL_LINE_MASK	REG_GENMASK(19, 0)
+#define _PIPEACONF		0x70008
+#define   PIPECONF_ENABLE			REG_BIT(31)
+#define   PIPECONF_DOUBLE_WIDE			REG_BIT(30) /* pre-i965 */
+#define   PIPECONF_STATE_ENABLE			REG_BIT(30) /* i965+ */
+#define   PIPECONF_DSI_PLL_LOCKED		REG_BIT(29) /* vlv & pipe A only */
+#define   PIPECONF_FRAME_START_DELAY_MASK	REG_GENMASK(28, 27) /* pre-hsw */
+#define   PIPECONF_FRAME_START_DELAY(x)		REG_FIELD_PREP(PIPECONF_FRAME_START_DELAY_MASK, (x)) /* pre-hsw: 0-3 */
+#define   PIPECONF_PIPE_LOCKED			REG_BIT(25)
+#define   PIPECONF_FORCE_BORDER			REG_BIT(25)
+#define   PIPECONF_GAMMA_MODE_MASK_I9XX		REG_BIT(24) /* gmch */
+#define   PIPECONF_GAMMA_MODE_MASK_ILK		REG_GENMASK(25, 24) /* ilk-ivb */
+#define   PIPECONF_GAMMA_MODE_8BIT		REG_FIELD_PREP(PIPECONF_GAMMA_MODE_MASK, 0)
+#define   PIPECONF_GAMMA_MODE_10BIT		REG_FIELD_PREP(PIPECONF_GAMMA_MODE_MASK, 1)
+#define   PIPECONF_GAMMA_MODE_12BIT		REG_FIELD_PREP(PIPECONF_GAMMA_MODE_MASK_ILK, 2) /* ilk-ivb */
+#define   PIPECONF_GAMMA_MODE_SPLIT		REG_FIELD_PREP(PIPECONF_GAMMA_MODE_MASK_ILK, 3) /* ivb */
+#define   PIPECONF_GAMMA_MODE(x)		REG_FIELD_PREP(PIPECONF_GAMMA_MODE_MASK_ILK, (x)) /* pass in GAMMA_MODE_MODE_* */
+#define   PIPECONF_INTERLACE_MASK		REG_GENMASK(23, 21) /* gen3+ */
+#define   PIPECONF_INTERLACE_PROGRESSIVE	REG_FIELD_PREP(PIPECONF_INTERLACE_MASK, 0)
+#define   PIPECONF_INTERLACE_W_SYNC_SHIFT_PANEL	REG_FIELD_PREP(PIPECONF_INTERLACE_MASK, 4) /* gen4 only */
+#define   PIPECONF_INTERLACE_W_SYNC_SHIFT	REG_FIELD_PREP(PIPECONF_INTERLACE_MASK, 5) /* gen4 only */
+#define   PIPECONF_INTERLACE_W_FIELD_INDICATION	REG_FIELD_PREP(PIPECONF_INTERLACE_MASK, 6)
+#define   PIPECONF_INTERLACE_FIELD_0_ONLY	REG_FIELD_PREP(PIPECONF_INTERLACE_MASK, 7) /* gen3 only */
+/*
+ * ilk+: PF/D=progressive fetch/display, IF/D=interlaced fetch/display,
+ * DBL=power saving pixel doubling, PF-ID* requires panel fitter
+ */
+#define   PIPECONF_INTERLACE_MASK_ILK		REG_GENMASK(23, 21) /* ilk+ */
+#define   PIPECONF_INTERLACE_MASK_HSW		REG_GENMASK(22, 21) /* hsw+ */
+#define   PIPECONF_INTERLACE_PF_PD_ILK		REG_FIELD_PREP(PIPECONF_INTERLACE_MASK_ILK, 0)
+#define   PIPECONF_INTERLACE_PF_ID_ILK		REG_FIELD_PREP(PIPECONF_INTERLACE_MASK_ILK, 1)
+#define   PIPECONF_INTERLACE_IF_ID_ILK		REG_FIELD_PREP(PIPECONF_INTERLACE_MASK_ILK, 3)
+#define   PIPECONF_INTERLACE_IF_ID_DBL_ILK	REG_FIELD_PREP(PIPECONF_INTERLACE_MASK_ILK, 4) /* ilk/snb only */
+#define   PIPECONF_INTERLACE_PF_ID_DBL_ILK	REG_FIELD_PREP(PIPECONF_INTERLACE_MASK_ILK, 5) /* ilk/snb only */
+#define   PIPECONF_REFRESH_RATE_ALT_ILK		REG_BIT(20)
+#define   PIPECONF_MSA_TIMING_DELAY_MASK	REG_GENMASK(19, 18) /* ilk/snb/ivb */
+#define   PIPECONF_MSA_TIMING_DELAY(x)		REG_FIELD_PREP(PIPECONF_MSA_TIMING_DELAY_MASK, (x))
+#define   PIPECONF_CXSR_DOWNCLOCK		REG_BIT(16)
+#define   PIPECONF_REFRESH_RATE_ALT_VLV		REG_BIT(14)
+#define   PIPECONF_COLOR_RANGE_SELECT		REG_BIT(13)
+#define   PIPECONF_OUTPUT_COLORSPACE_MASK	REG_GENMASK(12, 11) /* ilk-ivb */
+#define   PIPECONF_OUTPUT_COLORSPACE_RGB	REG_FIELD_PREP(PIPECONF_OUTPUT_COLORSPACE_MASK, 0) /* ilk-ivb */
+#define   PIPECONF_OUTPUT_COLORSPACE_YUV601	REG_FIELD_PREP(PIPECONF_OUTPUT_COLORSPACE_MASK, 1) /* ilk-ivb */
+#define   PIPECONF_OUTPUT_COLORSPACE_YUV709	REG_FIELD_PREP(PIPECONF_OUTPUT_COLORSPACE_MASK, 2) /* ilk-ivb */
+#define   PIPECONF_OUTPUT_COLORSPACE_YUV_HSW	REG_BIT(11) /* hsw only */
+#define   PIPECONF_BPC_MASK			REG_GENMASK(7, 5) /* ctg-ivb */
+#define   PIPECONF_BPC_8			REG_FIELD_PREP(PIPECONF_BPC_MASK, 0)
+#define   PIPECONF_BPC_10			REG_FIELD_PREP(PIPECONF_BPC_MASK, 1)
+#define   PIPECONF_BPC_6			REG_FIELD_PREP(PIPECONF_BPC_MASK, 2)
+#define   PIPECONF_BPC_12			REG_FIELD_PREP(PIPECONF_BPC_MASK, 3)
+#define   PIPECONF_DITHER_EN			REG_BIT(4)
+#define   PIPECONF_DITHER_TYPE_MASK		REG_GENMASK(3, 2)
+#define   PIPECONF_DITHER_TYPE_SP		REG_FIELD_PREP(PIPECONF_DITHER_TYPE_MASK, 0)
+#define   PIPECONF_DITHER_TYPE_ST1		REG_FIELD_PREP(PIPECONF_DITHER_TYPE_MASK, 1)
+#define   PIPECONF_DITHER_TYPE_ST2		REG_FIELD_PREP(PIPECONF_DITHER_TYPE_MASK, 2)
+#define   PIPECONF_DITHER_TYPE_TEMP		REG_FIELD_PREP(PIPECONF_DITHER_TYPE_MASK, 3)
+#define _PIPEASTAT		0x70024
+#define   PIPE_FIFO_UNDERRUN_STATUS		(1UL << 31)
+#define   SPRITE1_FLIP_DONE_INT_EN_VLV		(1UL << 30)
+#define   PIPE_CRC_ERROR_ENABLE			(1UL << 29)
+#define   PIPE_CRC_DONE_ENABLE			(1UL << 28)
+#define   PERF_COUNTER2_INTERRUPT_EN		(1UL << 27)
+#define   PIPE_GMBUS_EVENT_ENABLE		(1UL << 27)
+#define   PLANE_FLIP_DONE_INT_EN_VLV		(1UL << 26)
+#define   PIPE_HOTPLUG_INTERRUPT_ENABLE		(1UL << 26)
+#define   PIPE_VSYNC_INTERRUPT_ENABLE		(1UL << 25)
+#define   PIPE_DISPLAY_LINE_COMPARE_ENABLE	(1UL << 24)
+#define   PIPE_DPST_EVENT_ENABLE		(1UL << 23)
+#define   SPRITE0_FLIP_DONE_INT_EN_VLV		(1UL << 22)
+#define   PIPE_LEGACY_BLC_EVENT_ENABLE		(1UL << 22)
+#define   PIPE_ODD_FIELD_INTERRUPT_ENABLE	(1UL << 21)
+#define   PIPE_EVEN_FIELD_INTERRUPT_ENABLE	(1UL << 20)
+#define   PIPE_B_PSR_INTERRUPT_ENABLE_VLV	(1UL << 19)
+#define   PERF_COUNTER_INTERRUPT_EN		(1UL << 19)
+#define   PIPE_HOTPLUG_TV_INTERRUPT_ENABLE	(1UL << 18) /* pre-965 */
+#define   PIPE_START_VBLANK_INTERRUPT_ENABLE	(1UL << 18) /* 965 or later */
+#define   PIPE_FRAMESTART_INTERRUPT_ENABLE	(1UL << 17)
+#define   PIPE_VBLANK_INTERRUPT_ENABLE		(1UL << 17)
+#define   PIPEA_HBLANK_INT_EN_VLV		(1UL << 16)
+#define   PIPE_OVERLAY_UPDATED_ENABLE		(1UL << 16)
+#define   SPRITE1_FLIP_DONE_INT_STATUS_VLV	(1UL << 15)
+#define   SPRITE0_FLIP_DONE_INT_STATUS_VLV	(1UL << 14)
+#define   PIPE_CRC_ERROR_INTERRUPT_STATUS	(1UL << 13)
+#define   PIPE_CRC_DONE_INTERRUPT_STATUS	(1UL << 12)
+#define   PERF_COUNTER2_INTERRUPT_STATUS	(1UL << 11)
+#define   PIPE_GMBUS_INTERRUPT_STATUS		(1UL << 11)
+#define   PLANE_FLIP_DONE_INT_STATUS_VLV	(1UL << 10)
+#define   PIPE_HOTPLUG_INTERRUPT_STATUS		(1UL << 10)
+#define   PIPE_VSYNC_INTERRUPT_STATUS		(1UL << 9)
+#define   PIPE_DISPLAY_LINE_COMPARE_STATUS	(1UL << 8)
+#define   PIPE_DPST_EVENT_STATUS		(1UL << 7)
+#define   PIPE_A_PSR_STATUS_VLV			(1UL << 6)
+#define   PIPE_LEGACY_BLC_EVENT_STATUS		(1UL << 6)
+#define   PIPE_ODD_FIELD_INTERRUPT_STATUS	(1UL << 5)
+#define   PIPE_EVEN_FIELD_INTERRUPT_STATUS	(1UL << 4)
+#define   PIPE_B_PSR_STATUS_VLV			(1UL << 3)
+#define   PERF_COUNTER_INTERRUPT_STATUS		(1UL << 3)
+#define   PIPE_HOTPLUG_TV_INTERRUPT_STATUS	(1UL << 2) /* pre-965 */
+#define   PIPE_START_VBLANK_INTERRUPT_STATUS	(1UL << 2) /* 965 or later */
+#define   PIPE_FRAMESTART_INTERRUPT_STATUS	(1UL << 1)
+#define   PIPE_VBLANK_INTERRUPT_STATUS		(1UL << 1)
+#define   PIPE_HBLANK_INT_STATUS		(1UL << 0)
+#define   PIPE_OVERLAY_UPDATED_STATUS		(1UL << 0)
+
+#define PIPESTAT_INT_ENABLE_MASK		0x7fff0000
+#define PIPESTAT_INT_STATUS_MASK		0x0000ffff
+
+#define PIPE_A_OFFSET		0x70000
+#define PIPE_B_OFFSET		0x71000
+#define PIPE_C_OFFSET		0x72000
+#define PIPE_D_OFFSET		0x73000
+#define CHV_PIPE_C_OFFSET	0x74000
+/*
+ * There's actually no pipe EDP. Some pipe registers have
+ * simply shifted from the pipe to the transcoder, while
+ * keeping their original offset. Thus we need PIPE_EDP_OFFSET
+ * to access such registers in transcoder EDP.
+ */
+#define PIPE_EDP_OFFSET	0x7f000
+
+/* ICL DSI 0 and 1 */
+#define PIPE_DSI0_OFFSET	0x7b000
+#define PIPE_DSI1_OFFSET	0x7b800
+
+#define PIPECONF(pipe)		_MMIO_PIPE2(pipe, _PIPEACONF)
+#define PIPEDSL(pipe)		_MMIO_PIPE2(pipe, _PIPEADSL)
+#define PIPEFRAME(pipe)		_MMIO_PIPE2(pipe, _PIPEAFRAMEHIGH)
+#define PIPEFRAMEPIXEL(pipe)	_MMIO_PIPE2(pipe, _PIPEAFRAMEPIXEL)
+#define PIPESTAT(pipe)		_MMIO_PIPE2(pipe, _PIPEASTAT)
+
+#define  _PIPEAGCMAX           0x70010
+#define  _PIPEBGCMAX           0x71010
+#define PIPEGCMAX(pipe, i)     _MMIO_PIPE2(pipe, _PIPEAGCMAX + (i) * 4)
+
+#define _PIPE_ARB_CTL_A			0x70028 /* icl+ */
+#define PIPE_ARB_CTL(pipe)		_MMIO_PIPE2(pipe, _PIPE_ARB_CTL_A)
+#define   PIPE_ARB_USE_PROG_SLOTS	REG_BIT(13)
+
+#define _PIPE_MISC_A			0x70030
+#define _PIPE_MISC_B			0x71030
+#define   PIPEMISC_YUV420_ENABLE		REG_BIT(27) /* glk+ */
+#define   PIPEMISC_YUV420_MODE_FULL_BLEND	REG_BIT(26) /* glk+ */
+#define   PIPEMISC_HDR_MODE_PRECISION		REG_BIT(23) /* icl+ */
+#define   PIPEMISC_OUTPUT_COLORSPACE_YUV	REG_BIT(11)
+#define   PIPEMISC_PIXEL_ROUNDING_TRUNC		REG_BIT(8) /* tgl+ */
+/*
+ * For Display < 13, Bits 5-7 of PIPE MISC represent DITHER BPC with
+ * valid values of: 6, 8, 10 BPC.
+ * ADLP+, the bits 5-7 represent PORT OUTPUT BPC with valid values of:
+ * 6, 8, 10, 12 BPC.
+ */
+#define   PIPEMISC_BPC_MASK			REG_GENMASK(7, 5)
+#define   PIPEMISC_BPC_8			REG_FIELD_PREP(PIPEMISC_BPC_MASK, 0)
+#define   PIPEMISC_BPC_10			REG_FIELD_PREP(PIPEMISC_BPC_MASK, 1)
+#define   PIPEMISC_BPC_6			REG_FIELD_PREP(PIPEMISC_BPC_MASK, 2)
+#define   PIPEMISC_BPC_12_ADLP			REG_FIELD_PREP(PIPEMISC_BPC_MASK, 4) /* adlp+ */
+#define   PIPEMISC_DITHER_ENABLE		REG_BIT(4)
+#define   PIPEMISC_DITHER_TYPE_MASK		REG_GENMASK(3, 2)
+#define   PIPEMISC_DITHER_TYPE_SP		REG_FIELD_PREP(PIPEMISC_DITHER_TYPE_MASK, 0)
+#define   PIPEMISC_DITHER_TYPE_ST1		REG_FIELD_PREP(PIPEMISC_DITHER_TYPE_MASK, 1)
+#define   PIPEMISC_DITHER_TYPE_ST2		REG_FIELD_PREP(PIPEMISC_DITHER_TYPE_MASK, 2)
+#define   PIPEMISC_DITHER_TYPE_TEMP		REG_FIELD_PREP(PIPEMISC_DITHER_TYPE_MASK, 3)
+#define PIPEMISC(pipe)			_MMIO_PIPE2(pipe, _PIPE_MISC_A)
+
+#define _PIPE_MISC2_A					0x7002C
+#define _PIPE_MISC2_B					0x7102C
+#define   PIPE_MISC2_BUBBLE_COUNTER_MASK	REG_GENMASK(31, 24)
+#define   PIPE_MISC2_BUBBLE_COUNTER_SCALER_EN	REG_FIELD_PREP(PIPE_MISC2_BUBBLE_COUNTER_MASK, 80)
+#define   PIPE_MISC2_BUBBLE_COUNTER_SCALER_DIS	REG_FIELD_PREP(PIPE_MISC2_BUBBLE_COUNTER_MASK, 20)
+#define PIPE_MISC2(pipe)					_MMIO_PIPE2(pipe, _PIPE_MISC2_A)
+
+/* Skylake+ pipe bottom (background) color */
+#define _SKL_BOTTOM_COLOR_A		0x70034
+#define _SKL_BOTTOM_COLOR_B		0x71034
+#define   SKL_BOTTOM_COLOR_GAMMA_ENABLE		REG_BIT(31)
+#define   SKL_BOTTOM_COLOR_CSC_ENABLE		REG_BIT(30)
+#define SKL_BOTTOM_COLOR(pipe)		_MMIO_PIPE(pipe, _SKL_BOTTOM_COLOR_A, _SKL_BOTTOM_COLOR_B)
+
+#define _ICL_PIPE_A_STATUS			0x70058
+#define ICL_PIPESTATUS(pipe)			_MMIO_PIPE2(pipe, _ICL_PIPE_A_STATUS)
+#define   PIPE_STATUS_UNDERRUN				REG_BIT(31)
+#define   PIPE_STATUS_SOFT_UNDERRUN_XELPD		REG_BIT(28)
+#define   PIPE_STATUS_HARD_UNDERRUN_XELPD		REG_BIT(27)
+#define   PIPE_STATUS_PORT_UNDERRUN_XELPD		REG_BIT(26)
+
+#define VLV_DPFLIPSTAT				_MMIO(VLV_DISPLAY_BASE + 0x70028)
+#define   PIPEB_LINE_COMPARE_INT_EN			REG_BIT(29)
+#define   PIPEB_HLINE_INT_EN			REG_BIT(28)
+#define   PIPEB_VBLANK_INT_EN			REG_BIT(27)
+#define   SPRITED_FLIP_DONE_INT_EN			REG_BIT(26)
+#define   SPRITEC_FLIP_DONE_INT_EN			REG_BIT(25)
+#define   PLANEB_FLIP_DONE_INT_EN			REG_BIT(24)
+#define   PIPE_PSR_INT_EN			REG_BIT(22)
+#define   PIPEA_LINE_COMPARE_INT_EN			REG_BIT(21)
+#define   PIPEA_HLINE_INT_EN			REG_BIT(20)
+#define   PIPEA_VBLANK_INT_EN			REG_BIT(19)
+#define   SPRITEB_FLIP_DONE_INT_EN			REG_BIT(18)
+#define   SPRITEA_FLIP_DONE_INT_EN			REG_BIT(17)
+#define   PLANEA_FLIPDONE_INT_EN			REG_BIT(16)
+#define   PIPEC_LINE_COMPARE_INT_EN			REG_BIT(13)
+#define   PIPEC_HLINE_INT_EN			REG_BIT(12)
+#define   PIPEC_VBLANK_INT_EN			REG_BIT(11)
+#define   SPRITEF_FLIPDONE_INT_EN			REG_BIT(10)
+#define   SPRITEE_FLIPDONE_INT_EN			REG_BIT(9)
+#define   PLANEC_FLIPDONE_INT_EN			REG_BIT(8)
+
+#define DPINVGTT				_MMIO(VLV_DISPLAY_BASE + 0x7002c) /* VLV/CHV only */
+#define   DPINVGTT_EN_MASK_CHV				REG_GENMASK(27, 16)
+#define   DPINVGTT_EN_MASK_VLV				REG_GENMASK(23, 16)
+#define   SPRITEF_INVALID_GTT_INT_EN			REG_BIT(27)
+#define   SPRITEE_INVALID_GTT_INT_EN			REG_BIT(26)
+#define   PLANEC_INVALID_GTT_INT_EN			REG_BIT(25)
+#define   CURSORC_INVALID_GTT_INT_EN			REG_BIT(24)
+#define   CURSORB_INVALID_GTT_INT_EN			REG_BIT(23)
+#define   CURSORA_INVALID_GTT_INT_EN			REG_BIT(22)
+#define   SPRITED_INVALID_GTT_INT_EN			REG_BIT(21)
+#define   SPRITEC_INVALID_GTT_INT_EN			REG_BIT(20)
+#define   PLANEB_INVALID_GTT_INT_EN			REG_BIT(19)
+#define   SPRITEB_INVALID_GTT_INT_EN			REG_BIT(18)
+#define   SPRITEA_INVALID_GTT_INT_EN			REG_BIT(17)
+#define   PLANEA_INVALID_GTT_INT_EN			REG_BIT(16)
+#define   DPINVGTT_STATUS_MASK_CHV			REG_GENMASK(11, 0)
+#define   DPINVGTT_STATUS_MASK_VLV			REG_GENMASK(7, 0)
+#define   SPRITEF_INVALID_GTT_STATUS			REG_BIT(11)
+#define   SPRITEE_INVALID_GTT_STATUS			REG_BIT(10)
+#define   PLANEC_INVALID_GTT_STATUS			REG_BIT(9)
+#define   CURSORC_INVALID_GTT_STATUS			REG_BIT(8)
+#define   CURSORB_INVALID_GTT_STATUS			REG_BIT(7)
+#define   CURSORA_INVALID_GTT_STATUS			REG_BIT(6)
+#define   SPRITED_INVALID_GTT_STATUS			REG_BIT(5)
+#define   SPRITEC_INVALID_GTT_STATUS			REG_BIT(4)
+#define   PLANEB_INVALID_GTT_STATUS			REG_BIT(3)
+#define   SPRITEB_INVALID_GTT_STATUS			REG_BIT(2)
+#define   SPRITEA_INVALID_GTT_STATUS			REG_BIT(1)
+#define   PLANEA_INVALID_GTT_STATUS			REG_BIT(0)
+
+#define DSPARB			_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x70030)
+#define   DSPARB_CSTART_MASK	(0x7f << 7)
+#define   DSPARB_CSTART_SHIFT	7
+#define   DSPARB_BSTART_MASK	(0x7f)
+#define   DSPARB_BSTART_SHIFT	0
+#define   DSPARB_BEND_SHIFT	9 /* on 855 */
+#define   DSPARB_AEND_SHIFT	0
+#define   DSPARB_SPRITEA_SHIFT_VLV	0
+#define   DSPARB_SPRITEA_MASK_VLV	(0xff << 0)
+#define   DSPARB_SPRITEB_SHIFT_VLV	8
+#define   DSPARB_SPRITEB_MASK_VLV	(0xff << 8)
+#define   DSPARB_SPRITEC_SHIFT_VLV	16
+#define   DSPARB_SPRITEC_MASK_VLV	(0xff << 16)
+#define   DSPARB_SPRITED_SHIFT_VLV	24
+#define   DSPARB_SPRITED_MASK_VLV	(0xff << 24)
+#define DSPARB2				_MMIO(VLV_DISPLAY_BASE + 0x70060) /* vlv/chv */
+#define   DSPARB_SPRITEA_HI_SHIFT_VLV	0
+#define   DSPARB_SPRITEA_HI_MASK_VLV	(0x1 << 0)
+#define   DSPARB_SPRITEB_HI_SHIFT_VLV	4
+#define   DSPARB_SPRITEB_HI_MASK_VLV	(0x1 << 4)
+#define   DSPARB_SPRITEC_HI_SHIFT_VLV	8
+#define   DSPARB_SPRITEC_HI_MASK_VLV	(0x1 << 8)
+#define   DSPARB_SPRITED_HI_SHIFT_VLV	12
+#define   DSPARB_SPRITED_HI_MASK_VLV	(0x1 << 12)
+#define   DSPARB_SPRITEE_HI_SHIFT_VLV	16
+#define   DSPARB_SPRITEE_HI_MASK_VLV	(0x1 << 16)
+#define   DSPARB_SPRITEF_HI_SHIFT_VLV	20
+#define   DSPARB_SPRITEF_HI_MASK_VLV	(0x1 << 20)
+#define DSPARB3				_MMIO(VLV_DISPLAY_BASE + 0x7006c) /* chv */
+#define   DSPARB_SPRITEE_SHIFT_VLV	0
+#define   DSPARB_SPRITEE_MASK_VLV	(0xff << 0)
+#define   DSPARB_SPRITEF_SHIFT_VLV	8
+#define   DSPARB_SPRITEF_MASK_VLV	(0xff << 8)
+
+/* pnv/gen4/g4x/vlv/chv */
+#define DSPFW1		_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x70034)
+#define   DSPFW_SR_SHIFT		23
+#define   DSPFW_SR_MASK			(0x1ff << 23)
+#define   DSPFW_CURSORB_SHIFT		16
+#define   DSPFW_CURSORB_MASK		(0x3f << 16)
+#define   DSPFW_PLANEB_SHIFT		8
+#define   DSPFW_PLANEB_MASK		(0x7f << 8)
+#define   DSPFW_PLANEB_MASK_VLV		(0xff << 8) /* vlv/chv */
+#define   DSPFW_PLANEA_SHIFT		0
+#define   DSPFW_PLANEA_MASK		(0x7f << 0)
+#define   DSPFW_PLANEA_MASK_VLV		(0xff << 0) /* vlv/chv */
+#define DSPFW2		_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x70038)
+#define   DSPFW_FBC_SR_EN		(1 << 31)	  /* g4x */
+#define   DSPFW_FBC_SR_SHIFT		28
+#define   DSPFW_FBC_SR_MASK		(0x7 << 28) /* g4x */
+#define   DSPFW_FBC_HPLL_SR_SHIFT	24
+#define   DSPFW_FBC_HPLL_SR_MASK	(0xf << 24) /* g4x */
+#define   DSPFW_SPRITEB_SHIFT		(16)
+#define   DSPFW_SPRITEB_MASK		(0x7f << 16) /* g4x */
+#define   DSPFW_SPRITEB_MASK_VLV	(0xff << 16) /* vlv/chv */
+#define   DSPFW_CURSORA_SHIFT		8
+#define   DSPFW_CURSORA_MASK		(0x3f << 8)
+#define   DSPFW_PLANEC_OLD_SHIFT	0
+#define   DSPFW_PLANEC_OLD_MASK		(0x7f << 0) /* pre-gen4 sprite C */
+#define   DSPFW_SPRITEA_SHIFT		0
+#define   DSPFW_SPRITEA_MASK		(0x7f << 0) /* g4x */
+#define   DSPFW_SPRITEA_MASK_VLV	(0xff << 0) /* vlv/chv */
+#define DSPFW3		_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x7003c)
+#define   DSPFW_HPLL_SR_EN		(1 << 31)
+#define   PINEVIEW_SELF_REFRESH_EN	(1 << 30)
+#define   DSPFW_CURSOR_SR_SHIFT		24
+#define   DSPFW_CURSOR_SR_MASK		(0x3f << 24)
+#define   DSPFW_HPLL_CURSOR_SHIFT	16
+#define   DSPFW_HPLL_CURSOR_MASK	(0x3f << 16)
+#define   DSPFW_HPLL_SR_SHIFT		0
+#define   DSPFW_HPLL_SR_MASK		(0x1ff << 0)
+
+/* vlv/chv */
+#define DSPFW4		_MMIO(VLV_DISPLAY_BASE + 0x70070)
+#define   DSPFW_SPRITEB_WM1_SHIFT	16
+#define   DSPFW_SPRITEB_WM1_MASK	(0xff << 16)
+#define   DSPFW_CURSORA_WM1_SHIFT	8
+#define   DSPFW_CURSORA_WM1_MASK	(0x3f << 8)
+#define   DSPFW_SPRITEA_WM1_SHIFT	0
+#define   DSPFW_SPRITEA_WM1_MASK	(0xff << 0)
+#define DSPFW5		_MMIO(VLV_DISPLAY_BASE + 0x70074)
+#define   DSPFW_PLANEB_WM1_SHIFT	24
+#define   DSPFW_PLANEB_WM1_MASK		(0xff << 24)
+#define   DSPFW_PLANEA_WM1_SHIFT	16
+#define   DSPFW_PLANEA_WM1_MASK		(0xff << 16)
+#define   DSPFW_CURSORB_WM1_SHIFT	8
+#define   DSPFW_CURSORB_WM1_MASK	(0x3f << 8)
+#define   DSPFW_CURSOR_SR_WM1_SHIFT	0
+#define   DSPFW_CURSOR_SR_WM1_MASK	(0x3f << 0)
+#define DSPFW6		_MMIO(VLV_DISPLAY_BASE + 0x70078)
+#define   DSPFW_SR_WM1_SHIFT		0
+#define   DSPFW_SR_WM1_MASK		(0x1ff << 0)
+#define DSPFW7		_MMIO(VLV_DISPLAY_BASE + 0x7007c)
+#define DSPFW7_CHV	_MMIO(VLV_DISPLAY_BASE + 0x700b4) /* wtf #1? */
+#define   DSPFW_SPRITED_WM1_SHIFT	24
+#define   DSPFW_SPRITED_WM1_MASK	(0xff << 24)
+#define   DSPFW_SPRITED_SHIFT		16
+#define   DSPFW_SPRITED_MASK_VLV	(0xff << 16)
+#define   DSPFW_SPRITEC_WM1_SHIFT	8
+#define   DSPFW_SPRITEC_WM1_MASK	(0xff << 8)
+#define   DSPFW_SPRITEC_SHIFT		0
+#define   DSPFW_SPRITEC_MASK_VLV	(0xff << 0)
+#define DSPFW8_CHV	_MMIO(VLV_DISPLAY_BASE + 0x700b8)
+#define   DSPFW_SPRITEF_WM1_SHIFT	24
+#define   DSPFW_SPRITEF_WM1_MASK	(0xff << 24)
+#define   DSPFW_SPRITEF_SHIFT		16
+#define   DSPFW_SPRITEF_MASK_VLV	(0xff << 16)
+#define   DSPFW_SPRITEE_WM1_SHIFT	8
+#define   DSPFW_SPRITEE_WM1_MASK	(0xff << 8)
+#define   DSPFW_SPRITEE_SHIFT		0
+#define   DSPFW_SPRITEE_MASK_VLV	(0xff << 0)
+#define DSPFW9_CHV	_MMIO(VLV_DISPLAY_BASE + 0x7007c) /* wtf #2? */
+#define   DSPFW_PLANEC_WM1_SHIFT	24
+#define   DSPFW_PLANEC_WM1_MASK		(0xff << 24)
+#define   DSPFW_PLANEC_SHIFT		16
+#define   DSPFW_PLANEC_MASK_VLV		(0xff << 16)
+#define   DSPFW_CURSORC_WM1_SHIFT	8
+#define   DSPFW_CURSORC_WM1_MASK	(0x3f << 16)
+#define   DSPFW_CURSORC_SHIFT		0
+#define   DSPFW_CURSORC_MASK		(0x3f << 0)
+
+/* vlv/chv high order bits */
+#define DSPHOWM		_MMIO(VLV_DISPLAY_BASE + 0x70064)
+#define   DSPFW_SR_HI_SHIFT		24
+#define   DSPFW_SR_HI_MASK		(3 << 24) /* 2 bits for chv, 1 for vlv */
+#define   DSPFW_SPRITEF_HI_SHIFT	23
+#define   DSPFW_SPRITEF_HI_MASK		(1 << 23)
+#define   DSPFW_SPRITEE_HI_SHIFT	22
+#define   DSPFW_SPRITEE_HI_MASK		(1 << 22)
+#define   DSPFW_PLANEC_HI_SHIFT		21
+#define   DSPFW_PLANEC_HI_MASK		(1 << 21)
+#define   DSPFW_SPRITED_HI_SHIFT	20
+#define   DSPFW_SPRITED_HI_MASK		(1 << 20)
+#define   DSPFW_SPRITEC_HI_SHIFT	16
+#define   DSPFW_SPRITEC_HI_MASK		(1 << 16)
+#define   DSPFW_PLANEB_HI_SHIFT		12
+#define   DSPFW_PLANEB_HI_MASK		(1 << 12)
+#define   DSPFW_SPRITEB_HI_SHIFT	8
+#define   DSPFW_SPRITEB_HI_MASK		(1 << 8)
+#define   DSPFW_SPRITEA_HI_SHIFT	4
+#define   DSPFW_SPRITEA_HI_MASK		(1 << 4)
+#define   DSPFW_PLANEA_HI_SHIFT		0
+#define   DSPFW_PLANEA_HI_MASK		(1 << 0)
+#define DSPHOWM1	_MMIO(VLV_DISPLAY_BASE + 0x70068)
+#define   DSPFW_SR_WM1_HI_SHIFT		24
+#define   DSPFW_SR_WM1_HI_MASK		(3 << 24) /* 2 bits for chv, 1 for vlv */
+#define   DSPFW_SPRITEF_WM1_HI_SHIFT	23
+#define   DSPFW_SPRITEF_WM1_HI_MASK	(1 << 23)
+#define   DSPFW_SPRITEE_WM1_HI_SHIFT	22
+#define   DSPFW_SPRITEE_WM1_HI_MASK	(1 << 22)
+#define   DSPFW_PLANEC_WM1_HI_SHIFT	21
+#define   DSPFW_PLANEC_WM1_HI_MASK	(1 << 21)
+#define   DSPFW_SPRITED_WM1_HI_SHIFT	20
+#define   DSPFW_SPRITED_WM1_HI_MASK	(1 << 20)
+#define   DSPFW_SPRITEC_WM1_HI_SHIFT	16
+#define   DSPFW_SPRITEC_WM1_HI_MASK	(1 << 16)
+#define   DSPFW_PLANEB_WM1_HI_SHIFT	12
+#define   DSPFW_PLANEB_WM1_HI_MASK	(1 << 12)
+#define   DSPFW_SPRITEB_WM1_HI_SHIFT	8
+#define   DSPFW_SPRITEB_WM1_HI_MASK	(1 << 8)
+#define   DSPFW_SPRITEA_WM1_HI_SHIFT	4
+#define   DSPFW_SPRITEA_WM1_HI_MASK	(1 << 4)
+#define   DSPFW_PLANEA_WM1_HI_SHIFT	0
+#define   DSPFW_PLANEA_WM1_HI_MASK	(1 << 0)
+
+/* drain latency register values*/
+#define VLV_DDL(pipe)			_MMIO(VLV_DISPLAY_BASE + 0x70050 + 4 * (pipe))
+#define DDL_CURSOR_SHIFT		24
+#define DDL_SPRITE_SHIFT(sprite)	(8 + 8 * (sprite))
+#define DDL_PLANE_SHIFT			0
+#define DDL_PRECISION_HIGH		(1 << 7)
+#define DDL_PRECISION_LOW		(0 << 7)
+#define DRAIN_LATENCY_MASK		0x7f
+
+#define CBR1_VLV			_MMIO(VLV_DISPLAY_BASE + 0x70400)
+#define  CBR_PND_DEADLINE_DISABLE	(1 << 31)
+#define  CBR_PWM_CLOCK_MUX_SELECT	(1 << 30)
+
+#define CBR4_VLV			_MMIO(VLV_DISPLAY_BASE + 0x70450)
+#define  CBR_DPLLBMD_PIPE(pipe)		(1 << (7 + (pipe) * 11)) /* pipes B and C */
+
+/* FIFO watermark sizes etc */
+#define G4X_FIFO_LINE_SIZE	64
+#define I915_FIFO_LINE_SIZE	64
+#define I830_FIFO_LINE_SIZE	32
+
+#define VALLEYVIEW_FIFO_SIZE	255
+#define G4X_FIFO_SIZE		127
+#define I965_FIFO_SIZE		512
+#define I945_FIFO_SIZE		127
+#define I915_FIFO_SIZE		95
+#define I855GM_FIFO_SIZE	127 /* In cachelines */
+#define I830_FIFO_SIZE		95
+
+#define VALLEYVIEW_MAX_WM	0xff
+#define G4X_MAX_WM		0x3f
+#define I915_MAX_WM		0x3f
+
+#define PINEVIEW_DISPLAY_FIFO	512 /* in 64byte unit */
+#define PINEVIEW_FIFO_LINE_SIZE	64
+#define PINEVIEW_MAX_WM		0x1ff
+#define PINEVIEW_DFT_WM		0x3f
+#define PINEVIEW_DFT_HPLLOFF_WM	0
+#define PINEVIEW_GUARD_WM		10
+#define PINEVIEW_CURSOR_FIFO		64
+#define PINEVIEW_CURSOR_MAX_WM	0x3f
+#define PINEVIEW_CURSOR_DFT_WM	0
+#define PINEVIEW_CURSOR_GUARD_WM	5
+
+#define VALLEYVIEW_CURSOR_MAX_WM 64
+#define I965_CURSOR_FIFO	64
+#define I965_CURSOR_MAX_WM	32
+#define I965_CURSOR_DFT_WM	8
+
+/* Watermark register definitions for SKL */
+#define _CUR_WM_A_0		0x70140
+#define _CUR_WM_B_0		0x71140
+#define _CUR_WM_SAGV_A		0x70158
+#define _CUR_WM_SAGV_B		0x71158
+#define _CUR_WM_SAGV_TRANS_A	0x7015C
+#define _CUR_WM_SAGV_TRANS_B	0x7115C
+#define _CUR_WM_TRANS_A		0x70168
+#define _CUR_WM_TRANS_B		0x71168
+#define _PLANE_WM_1_A_0		0x70240
+#define _PLANE_WM_1_B_0		0x71240
+#define _PLANE_WM_2_A_0		0x70340
+#define _PLANE_WM_2_B_0		0x71340
+#define _PLANE_WM_SAGV_1_A	0x70258
+#define _PLANE_WM_SAGV_1_B	0x71258
+#define _PLANE_WM_SAGV_2_A	0x70358
+#define _PLANE_WM_SAGV_2_B	0x71358
+#define _PLANE_WM_SAGV_TRANS_1_A	0x7025C
+#define _PLANE_WM_SAGV_TRANS_1_B	0x7125C
+#define _PLANE_WM_SAGV_TRANS_2_A	0x7035C
+#define _PLANE_WM_SAGV_TRANS_2_B	0x7135C
+#define _PLANE_WM_TRANS_1_A	0x70268
+#define _PLANE_WM_TRANS_1_B	0x71268
+#define _PLANE_WM_TRANS_2_A	0x70368
+#define _PLANE_WM_TRANS_2_B	0x71368
+#define   PLANE_WM_EN		(1 << 31)
+#define   PLANE_WM_IGNORE_LINES	(1 << 30)
+#define   PLANE_WM_LINES_MASK	REG_GENMASK(26, 14)
+#define   PLANE_WM_BLOCKS_MASK	REG_GENMASK(11, 0)
+
+#define _CUR_WM_0(pipe) _PIPE(pipe, _CUR_WM_A_0, _CUR_WM_B_0)
+#define CUR_WM(pipe, level) _MMIO(_CUR_WM_0(pipe) + ((4) * (level)))
+#define CUR_WM_SAGV(pipe) _MMIO_PIPE(pipe, _CUR_WM_SAGV_A, _CUR_WM_SAGV_B)
+#define CUR_WM_SAGV_TRANS(pipe) _MMIO_PIPE(pipe, _CUR_WM_SAGV_TRANS_A, _CUR_WM_SAGV_TRANS_B)
+#define CUR_WM_TRANS(pipe) _MMIO_PIPE(pipe, _CUR_WM_TRANS_A, _CUR_WM_TRANS_B)
+#define _PLANE_WM_1(pipe) _PIPE(pipe, _PLANE_WM_1_A_0, _PLANE_WM_1_B_0)
+#define _PLANE_WM_2(pipe) _PIPE(pipe, _PLANE_WM_2_A_0, _PLANE_WM_2_B_0)
+#define _PLANE_WM_BASE(pipe, plane) \
+	_PLANE(plane, _PLANE_WM_1(pipe), _PLANE_WM_2(pipe))
+#define PLANE_WM(pipe, plane, level) \
+	_MMIO(_PLANE_WM_BASE(pipe, plane) + ((4) * (level)))
+#define _PLANE_WM_SAGV_1(pipe) \
+	_PIPE(pipe, _PLANE_WM_SAGV_1_A, _PLANE_WM_SAGV_1_B)
+#define _PLANE_WM_SAGV_2(pipe) \
+	_PIPE(pipe, _PLANE_WM_SAGV_2_A, _PLANE_WM_SAGV_2_B)
+#define PLANE_WM_SAGV(pipe, plane) \
+	_MMIO(_PLANE(plane, _PLANE_WM_SAGV_1(pipe), _PLANE_WM_SAGV_2(pipe)))
+#define _PLANE_WM_SAGV_TRANS_1(pipe) \
+	_PIPE(pipe, _PLANE_WM_SAGV_TRANS_1_A, _PLANE_WM_SAGV_TRANS_1_B)
+#define _PLANE_WM_SAGV_TRANS_2(pipe) \
+	_PIPE(pipe, _PLANE_WM_SAGV_TRANS_2_A, _PLANE_WM_SAGV_TRANS_2_B)
+#define PLANE_WM_SAGV_TRANS(pipe, plane) \
+	_MMIO(_PLANE(plane, _PLANE_WM_SAGV_TRANS_1(pipe), _PLANE_WM_SAGV_TRANS_2(pipe)))
+#define _PLANE_WM_TRANS_1(pipe) \
+	_PIPE(pipe, _PLANE_WM_TRANS_1_A, _PLANE_WM_TRANS_1_B)
+#define _PLANE_WM_TRANS_2(pipe) \
+	_PIPE(pipe, _PLANE_WM_TRANS_2_A, _PLANE_WM_TRANS_2_B)
+#define PLANE_WM_TRANS(pipe, plane) \
+	_MMIO(_PLANE(plane, _PLANE_WM_TRANS_1(pipe), _PLANE_WM_TRANS_2(pipe)))
+
+/* define the Watermark register on Ironlake */
+#define _WM0_PIPEA_ILK		0x45100
+#define _WM0_PIPEB_ILK		0x45104
+#define _WM0_PIPEC_IVB		0x45200
+#define WM0_PIPE_ILK(pipe)	_MMIO_PIPE3((pipe), _WM0_PIPEA_ILK, \
+					    _WM0_PIPEB_ILK, _WM0_PIPEC_IVB)
+#define  WM0_PIPE_PRIMARY_MASK	REG_GENMASK(31, 16)
+#define  WM0_PIPE_SPRITE_MASK	REG_GENMASK(15, 8)
+#define  WM0_PIPE_CURSOR_MASK	REG_GENMASK(7, 0)
+#define  WM0_PIPE_PRIMARY(x)	REG_FIELD_PREP(WM0_PIPE_PRIMARY_MASK, (x))
+#define  WM0_PIPE_SPRITE(x)	REG_FIELD_PREP(WM0_PIPE_SPRITE_MASK, (x))
+#define  WM0_PIPE_CURSOR(x)	REG_FIELD_PREP(WM0_PIPE_CURSOR_MASK, (x))
+#define WM1_LP_ILK		_MMIO(0x45108)
+#define WM2_LP_ILK		_MMIO(0x4510c)
+#define WM3_LP_ILK		_MMIO(0x45110)
+#define  WM_LP_ENABLE		REG_BIT(31)
+#define  WM_LP_LATENCY_MASK	REG_GENMASK(30, 24)
+#define  WM_LP_FBC_MASK_BDW	REG_GENMASK(23, 19)
+#define  WM_LP_FBC_MASK_ILK	REG_GENMASK(23, 20)
+#define  WM_LP_PRIMARY_MASK	REG_GENMASK(18, 8)
+#define  WM_LP_CURSOR_MASK	REG_GENMASK(7, 0)
+#define  WM_LP_LATENCY(x)	REG_FIELD_PREP(WM_LP_LATENCY_MASK, (x))
+#define  WM_LP_FBC_BDW(x)	REG_FIELD_PREP(WM_LP_FBC_MASK_BDW, (x))
+#define  WM_LP_FBC_ILK(x)	REG_FIELD_PREP(WM_LP_FBC_MASK_ILK, (x))
+#define  WM_LP_PRIMARY(x)	REG_FIELD_PREP(WM_LP_PRIMARY_MASK, (x))
+#define  WM_LP_CURSOR(x)	REG_FIELD_PREP(WM_LP_CURSOR_MASK, (x))
+#define WM1S_LP_ILK		_MMIO(0x45120)
+#define WM2S_LP_IVB		_MMIO(0x45124)
+#define WM3S_LP_IVB		_MMIO(0x45128)
+#define  WM_LP_SPRITE_ENABLE	REG_BIT(31) /* ilk/snb WM1S only */
+#define  WM_LP_SPRITE_MASK	REG_GENMASK(10, 0)
+#define  WM_LP_SPRITE(x)	REG_FIELD_PREP(WM_LP_SPRITE_MASK, (x))
+
+/*
+ * The two pipe frame counter registers are not synchronized, so
+ * reading a stable value is somewhat tricky. The following code
+ * should work:
+ *
+ *  do {
+ *    high1 = ((INREG(PIPEAFRAMEHIGH) & PIPE_FRAME_HIGH_MASK) >>
+ *             PIPE_FRAME_HIGH_SHIFT;
+ *    low1 =  ((INREG(PIPEAFRAMEPIXEL) & PIPE_FRAME_LOW_MASK) >>
+ *             PIPE_FRAME_LOW_SHIFT);
+ *    high2 = ((INREG(PIPEAFRAMEHIGH) & PIPE_FRAME_HIGH_MASK) >>
+ *             PIPE_FRAME_HIGH_SHIFT);
+ *  } while (high1 != high2);
+ *  frame = (high1 << 8) | low1;
+ */
+#define _PIPEAFRAMEHIGH          0x70040
+#define   PIPE_FRAME_HIGH_MASK    0x0000ffff
+#define   PIPE_FRAME_HIGH_SHIFT   0
+#define _PIPEAFRAMEPIXEL         0x70044
+#define   PIPE_FRAME_LOW_MASK     0xff000000
+#define   PIPE_FRAME_LOW_SHIFT    24
+#define   PIPE_PIXEL_MASK         0x00ffffff
+#define   PIPE_PIXEL_SHIFT        0
+/* GM45+ just has to be different */
+#define _PIPEA_FRMCOUNT_G4X	0x70040
+#define _PIPEA_FLIPCOUNT_G4X	0x70044
+#define PIPE_FRMCOUNT_G4X(pipe) _MMIO_PIPE2(pipe, _PIPEA_FRMCOUNT_G4X)
+#define PIPE_FLIPCOUNT_G4X(pipe) _MMIO_PIPE2(pipe, _PIPEA_FLIPCOUNT_G4X)
+
+/* Cursor A & B regs */
+#define _CURACNTR		0x70080
+/* Old style CUR*CNTR flags (desktop 8xx) */
+#define   CURSOR_ENABLE			REG_BIT(31)
+#define   CURSOR_PIPE_GAMMA_ENABLE	REG_BIT(30)
+#define   CURSOR_STRIDE_MASK	REG_GENMASK(29, 28)
+#define   CURSOR_STRIDE(stride)	REG_FIELD_PREP(CURSOR_STRIDE_MASK, ffs(stride) - 9) /* 256,512,1k,2k */
+#define   CURSOR_FORMAT_MASK	REG_GENMASK(26, 24)
+#define   CURSOR_FORMAT_2C	REG_FIELD_PREP(CURSOR_FORMAT_MASK, 0)
+#define   CURSOR_FORMAT_3C	REG_FIELD_PREP(CURSOR_FORMAT_MASK, 1)
+#define   CURSOR_FORMAT_4C	REG_FIELD_PREP(CURSOR_FORMAT_MASK, 2)
+#define   CURSOR_FORMAT_ARGB	REG_FIELD_PREP(CURSOR_FORMAT_MASK, 4)
+#define   CURSOR_FORMAT_XRGB	REG_FIELD_PREP(CURSOR_FORMAT_MASK, 5)
+/* New style CUR*CNTR flags */
+#define   MCURSOR_ARB_SLOTS_MASK	REG_GENMASK(30, 28) /* icl+ */
+#define   MCURSOR_ARB_SLOTS(x)		REG_FIELD_PREP(MCURSOR_ARB_SLOTS_MASK, (x)) /* icl+ */
+#define   MCURSOR_PIPE_SEL_MASK		REG_GENMASK(29, 28)
+#define   MCURSOR_PIPE_SEL(pipe)	REG_FIELD_PREP(MCURSOR_PIPE_SEL_MASK, (pipe))
+#define   MCURSOR_PIPE_GAMMA_ENABLE	REG_BIT(26)
+#define   MCURSOR_PIPE_CSC_ENABLE	REG_BIT(24) /* ilk+ */
+#define   MCURSOR_ROTATE_180		REG_BIT(15)
+#define   MCURSOR_TRICKLE_FEED_DISABLE	REG_BIT(14)
+#define   MCURSOR_MODE_MASK		0x27
+#define   MCURSOR_MODE_DISABLE		0x00
+#define   MCURSOR_MODE_128_32B_AX	0x02
+#define   MCURSOR_MODE_256_32B_AX	0x03
+#define   MCURSOR_MODE_64_32B_AX	0x07
+#define   MCURSOR_MODE_128_ARGB_AX	(0x20 | MCURSOR_MODE_128_32B_AX)
+#define   MCURSOR_MODE_256_ARGB_AX	(0x20 | MCURSOR_MODE_256_32B_AX)
+#define   MCURSOR_MODE_64_ARGB_AX	(0x20 | MCURSOR_MODE_64_32B_AX)
+#define _CURABASE		0x70084
+#define _CURAPOS		0x70088
+#define   CURSOR_POS_Y_SIGN		REG_BIT(31)
+#define   CURSOR_POS_Y_MASK		REG_GENMASK(30, 16)
+#define   CURSOR_POS_Y(y)		REG_FIELD_PREP(CURSOR_POS_Y_MASK, (y))
+#define   CURSOR_POS_X_SIGN		REG_BIT(15)
+#define   CURSOR_POS_X_MASK		REG_GENMASK(14, 0)
+#define   CURSOR_POS_X(x)		REG_FIELD_PREP(CURSOR_POS_X_MASK, (x))
+#define _CURASIZE		0x700a0 /* 845/865 */
+#define   CURSOR_HEIGHT_MASK		REG_GENMASK(21, 12)
+#define   CURSOR_HEIGHT(h)		REG_FIELD_PREP(CURSOR_HEIGHT_MASK, (h))
+#define   CURSOR_WIDTH_MASK		REG_GENMASK(9, 0)
+#define   CURSOR_WIDTH(w)		REG_FIELD_PREP(CURSOR_WIDTH_MASK, (w))
+#define _CUR_FBC_CTL_A		0x700a0 /* ivb+ */
+#define   CUR_FBC_EN			REG_BIT(31)
+#define   CUR_FBC_HEIGHT_MASK		REG_GENMASK(7, 0)
+#define   CUR_FBC_HEIGHT(h)		REG_FIELD_PREP(CUR_FBC_HEIGHT_MASK, (h))
+#define _CURASURFLIVE		0x700ac /* g4x+ */
+#define _CURBCNTR		0x700c0
+#define _CURBBASE		0x700c4
+#define _CURBPOS		0x700c8
+
+#define _CURBCNTR_IVB		0x71080
+#define _CURBBASE_IVB		0x71084
+#define _CURBPOS_IVB		0x71088
+
+#define CURCNTR(pipe) _MMIO_CURSOR2(pipe, _CURACNTR)
+#define CURBASE(pipe) _MMIO_CURSOR2(pipe, _CURABASE)
+#define CURPOS(pipe) _MMIO_CURSOR2(pipe, _CURAPOS)
+#define CURSIZE(pipe) _MMIO_CURSOR2(pipe, _CURASIZE)
+#define CUR_FBC_CTL(pipe) _MMIO_CURSOR2(pipe, _CUR_FBC_CTL_A)
+#define CURSURFLIVE(pipe) _MMIO_CURSOR2(pipe, _CURASURFLIVE)
+
+#define CURSOR_A_OFFSET 0x70080
+#define CURSOR_B_OFFSET 0x700c0
+#define CHV_CURSOR_C_OFFSET 0x700e0
+#define IVB_CURSOR_B_OFFSET 0x71080
+#define IVB_CURSOR_C_OFFSET 0x72080
+#define TGL_CURSOR_D_OFFSET 0x73080
+
+/* Display A control */
+#define _DSPAADDR_VLV				0x7017C /* vlv/chv */
+#define _DSPACNTR				0x70180
+#define   DISP_ENABLE			REG_BIT(31)
+#define   DISP_PIPE_GAMMA_ENABLE	REG_BIT(30)
+#define   DISP_FORMAT_MASK		REG_GENMASK(29, 26)
+#define   DISP_FORMAT_8BPP		REG_FIELD_PREP(DISP_FORMAT_MASK, 2)
+#define   DISP_FORMAT_BGRA555		REG_FIELD_PREP(DISP_FORMAT_MASK, 3)
+#define   DISP_FORMAT_BGRX555		REG_FIELD_PREP(DISP_FORMAT_MASK, 4)
+#define   DISP_FORMAT_BGRX565		REG_FIELD_PREP(DISP_FORMAT_MASK, 5)
+#define   DISP_FORMAT_BGRX888		REG_FIELD_PREP(DISP_FORMAT_MASK, 6)
+#define   DISP_FORMAT_BGRA888		REG_FIELD_PREP(DISP_FORMAT_MASK, 7)
+#define   DISP_FORMAT_RGBX101010	REG_FIELD_PREP(DISP_FORMAT_MASK, 8)
+#define   DISP_FORMAT_RGBA101010	REG_FIELD_PREP(DISP_FORMAT_MASK, 9)
+#define   DISP_FORMAT_BGRX101010	REG_FIELD_PREP(DISP_FORMAT_MASK, 10)
+#define   DISP_FORMAT_BGRA101010	REG_FIELD_PREP(DISP_FORMAT_MASK, 11)
+#define   DISP_FORMAT_RGBX161616	REG_FIELD_PREP(DISP_FORMAT_MASK, 12)
+#define   DISP_FORMAT_RGBX888		REG_FIELD_PREP(DISP_FORMAT_MASK, 14)
+#define   DISP_FORMAT_RGBA888		REG_FIELD_PREP(DISP_FORMAT_MASK, 15)
+#define   DISP_STEREO_ENABLE		REG_BIT(25)
+#define   DISP_PIPE_CSC_ENABLE		REG_BIT(24) /* ilk+ */
+#define   DISP_PIPE_SEL_MASK		REG_GENMASK(25, 24)
+#define   DISP_PIPE_SEL(pipe)		REG_FIELD_PREP(DISP_PIPE_SEL_MASK, (pipe))
+#define   DISP_SRC_KEY_ENABLE		REG_BIT(22)
+#define   DISP_LINE_DOUBLE		REG_BIT(20)
+#define   DISP_STEREO_POLARITY_SECOND	REG_BIT(18)
+#define   DISP_ALPHA_PREMULTIPLY	REG_BIT(16) /* CHV pipe B */
+#define   DISP_ROTATE_180		REG_BIT(15)
+#define   DISP_TRICKLE_FEED_DISABLE	REG_BIT(14) /* g4x+ */
+#define   DISP_TILED			REG_BIT(10)
+#define   DISP_ASYNC_FLIP		REG_BIT(9) /* g4x+ */
+#define   DISP_MIRROR			REG_BIT(8) /* CHV pipe B */
+#define _DSPAADDR				0x70184
+#define _DSPASTRIDE				0x70188
+#define _DSPAPOS				0x7018C /* reserved */
+#define   DISP_POS_Y_MASK		REG_GENMASK(31, 16)
+#define   DISP_POS_Y(y)			REG_FIELD_PREP(DISP_POS_Y_MASK, (y))
+#define   DISP_POS_X_MASK		REG_GENMASK(15, 0)
+#define   DISP_POS_X(x)			REG_FIELD_PREP(DISP_POS_X_MASK, (x))
+#define _DSPASIZE				0x70190
+#define   DISP_HEIGHT_MASK		REG_GENMASK(31, 16)
+#define   DISP_HEIGHT(h)		REG_FIELD_PREP(DISP_HEIGHT_MASK, (h))
+#define   DISP_WIDTH_MASK		REG_GENMASK(15, 0)
+#define   DISP_WIDTH(w)			REG_FIELD_PREP(DISP_WIDTH_MASK, (w))
+#define _DSPASURF				0x7019C /* 965+ only */
+#define   DISP_ADDR_MASK		REG_GENMASK(31, 12)
+#define _DSPATILEOFF				0x701A4 /* 965+ only */
+#define   DISP_OFFSET_Y_MASK		REG_GENMASK(31, 16)
+#define   DISP_OFFSET_Y(y)		REG_FIELD_PREP(DISP_OFFSET_Y_MASK, (y))
+#define   DISP_OFFSET_X_MASK		REG_GENMASK(15, 0)
+#define   DISP_OFFSET_X(x)		REG_FIELD_PREP(DISP_OFFSET_X_MASK, (x))
+#define _DSPAOFFSET				0x701A4 /* HSW */
+#define _DSPASURFLIVE				0x701AC
+#define _DSPAGAMC				0x701E0
+
+#define DSPADDR_VLV(plane)	_MMIO_PIPE2(plane, _DSPAADDR_VLV)
+#define DSPCNTR(plane)		_MMIO_PIPE2(plane, _DSPACNTR)
+#define DSPADDR(plane)		_MMIO_PIPE2(plane, _DSPAADDR)
+#define DSPSTRIDE(plane)	_MMIO_PIPE2(plane, _DSPASTRIDE)
+#define DSPPOS(plane)		_MMIO_PIPE2(plane, _DSPAPOS)
+#define DSPSIZE(plane)		_MMIO_PIPE2(plane, _DSPASIZE)
+#define DSPSURF(plane)		_MMIO_PIPE2(plane, _DSPASURF)
+#define DSPTILEOFF(plane)	_MMIO_PIPE2(plane, _DSPATILEOFF)
+#define DSPLINOFF(plane)	DSPADDR(plane)
+#define DSPOFFSET(plane)	_MMIO_PIPE2(plane, _DSPAOFFSET)
+#define DSPSURFLIVE(plane)	_MMIO_PIPE2(plane, _DSPASURFLIVE)
+#define DSPGAMC(plane, i)	_MMIO_PIPE2(plane, _DSPAGAMC + (5 - (i)) * 4) /* plane C only, 6 x u0.8 */
+
+/* CHV pipe B blender and primary plane */
+#define _CHV_BLEND_A		0x60a00
+#define   CHV_BLEND_MASK	REG_GENMASK(31, 30)
+#define   CHV_BLEND_LEGACY	REG_FIELD_PREP(CHV_BLEND_MASK, 0)
+#define   CHV_BLEND_ANDROID	REG_FIELD_PREP(CHV_BLEND_MASK, 1)
+#define   CHV_BLEND_MPO		REG_FIELD_PREP(CHV_BLEND_MASK, 2)
+#define _CHV_CANVAS_A		0x60a04
+#define   CHV_CANVAS_RED_MASK	REG_GENMASK(29, 20)
+#define   CHV_CANVAS_GREEN_MASK	REG_GENMASK(19, 10)
+#define   CHV_CANVAS_BLUE_MASK	REG_GENMASK(9, 0)
+#define _PRIMPOS_A		0x60a08
+#define   PRIM_POS_Y_MASK	REG_GENMASK(31, 16)
+#define   PRIM_POS_Y(y)		REG_FIELD_PREP(PRIM_POS_Y_MASK, (y))
+#define   PRIM_POS_X_MASK	REG_GENMASK(15, 0)
+#define   PRIM_POS_X(x)		REG_FIELD_PREP(PRIM_POS_X_MASK, (x))
+#define _PRIMSIZE_A		0x60a0c
+#define   PRIM_HEIGHT_MASK	REG_GENMASK(31, 16)
+#define   PRIM_HEIGHT(h)	REG_FIELD_PREP(PRIM_HEIGHT_MASK, (h))
+#define   PRIM_WIDTH_MASK	REG_GENMASK(15, 0)
+#define   PRIM_WIDTH(w)		REG_FIELD_PREP(PRIM_WIDTH_MASK, (w))
+#define _PRIMCNSTALPHA_A	0x60a10
+#define   PRIM_CONST_ALPHA_ENABLE	REG_BIT(31)
+#define   PRIM_CONST_ALPHA_MASK		REG_GENMASK(7, 0)
+#define   PRIM_CONST_ALPHA(alpha)	REG_FIELD_PREP(PRIM_CONST_ALPHA_MASK, (alpha))
+
+#define CHV_BLEND(pipe)		_MMIO_TRANS2(pipe, _CHV_BLEND_A)
+#define CHV_CANVAS(pipe)	_MMIO_TRANS2(pipe, _CHV_CANVAS_A)
+#define PRIMPOS(plane)		_MMIO_TRANS2(plane, _PRIMPOS_A)
+#define PRIMSIZE(plane)		_MMIO_TRANS2(plane, _PRIMSIZE_A)
+#define PRIMCNSTALPHA(plane)	_MMIO_TRANS2(plane, _PRIMCNSTALPHA_A)
+
+/* Display/Sprite base address macros */
+#define DISP_BASEADDR_MASK	(0xfffff000)
+#define I915_LO_DISPBASE(val)	((val) & ~DISP_BASEADDR_MASK)
+#define I915_HI_DISPBASE(val)	((val) & DISP_BASEADDR_MASK)
+
+/*
+ * VBIOS flags
+ * gen2:
+ * [00:06] alm,mgm
+ * [10:16] all
+ * [30:32] alm,mgm
+ * gen3+:
+ * [00:0f] all
+ * [10:1f] all
+ * [30:32] all
+ */
+#define SWF0(i)	_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x70410 + (i) * 4)
+#define SWF1(i)	_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x71410 + (i) * 4)
+#define SWF3(i)	_MMIO(DISPLAY_MMIO_BASE(dev_priv) + 0x72414 + (i) * 4)
+#define SWF_ILK(i)	_MMIO(0x4F000 + (i) * 4)
+
+/* Pipe B */
+#define _PIPEBDSL		(DISPLAY_MMIO_BASE(dev_priv) + 0x71000)
+#define _PIPEBCONF		(DISPLAY_MMIO_BASE(dev_priv) + 0x71008)
+#define _PIPEBSTAT		(DISPLAY_MMIO_BASE(dev_priv) + 0x71024)
+#define _PIPEBFRAMEHIGH		0x71040
+#define _PIPEBFRAMEPIXEL	0x71044
+#define _PIPEB_FRMCOUNT_G4X	(DISPLAY_MMIO_BASE(dev_priv) + 0x71040)
+#define _PIPEB_FLIPCOUNT_G4X	(DISPLAY_MMIO_BASE(dev_priv) + 0x71044)
+
+
+/* Display B control */
+#define _DSPBCNTR		(DISPLAY_MMIO_BASE(dev_priv) + 0x71180)
+#define   DISP_ALPHA_TRANS_ENABLE	REG_BIT(15)
+#define   DISP_SPRITE_ABOVE_OVERLAY	REG_BIT(0)
+#define _DSPBADDR		(DISPLAY_MMIO_BASE(dev_priv) + 0x71184)
+#define _DSPBSTRIDE		(DISPLAY_MMIO_BASE(dev_priv) + 0x71188)
+#define _DSPBPOS		(DISPLAY_MMIO_BASE(dev_priv) + 0x7118C)
+#define _DSPBSIZE		(DISPLAY_MMIO_BASE(dev_priv) + 0x71190)
+#define _DSPBSURF		(DISPLAY_MMIO_BASE(dev_priv) + 0x7119C)
+#define _DSPBTILEOFF		(DISPLAY_MMIO_BASE(dev_priv) + 0x711A4)
+#define _DSPBOFFSET		(DISPLAY_MMIO_BASE(dev_priv) + 0x711A4)
+#define _DSPBSURFLIVE		(DISPLAY_MMIO_BASE(dev_priv) + 0x711AC)
+
+/* ICL DSI 0 and 1 */
+#define _PIPEDSI0CONF		0x7b008
+#define _PIPEDSI1CONF		0x7b808
+
+/* Sprite A control */
+#define _DVSACNTR		0x72180
+#define   DVS_ENABLE			REG_BIT(31)
+#define   DVS_PIPE_GAMMA_ENABLE		REG_BIT(30)
+#define   DVS_YUV_RANGE_CORRECTION_DISABLE	REG_BIT(27)
+#define   DVS_FORMAT_MASK		REG_GENMASK(26, 25)
+#define   DVS_FORMAT_YUV422		REG_FIELD_PREP(DVS_FORMAT_MASK, 0)
+#define   DVS_FORMAT_RGBX101010		REG_FIELD_PREP(DVS_FORMAT_MASK, 1)
+#define   DVS_FORMAT_RGBX888		REG_FIELD_PREP(DVS_FORMAT_MASK, 2)
+#define   DVS_FORMAT_RGBX161616		REG_FIELD_PREP(DVS_FORMAT_MASK, 3)
+#define   DVS_PIPE_CSC_ENABLE		REG_BIT(24)
+#define   DVS_SOURCE_KEY		REG_BIT(22)
+#define   DVS_RGB_ORDER_XBGR		REG_BIT(20)
+#define   DVS_YUV_FORMAT_BT709		REG_BIT(18)
+#define   DVS_YUV_ORDER_MASK		REG_GENMASK(17, 16)
+#define   DVS_YUV_ORDER_YUYV		REG_FIELD_PREP(DVS_YUV_ORDER_MASK, 0)
+#define   DVS_YUV_ORDER_UYVY		REG_FIELD_PREP(DVS_YUV_ORDER_MASK, 1)
+#define   DVS_YUV_ORDER_YVYU		REG_FIELD_PREP(DVS_YUV_ORDER_MASK, 2)
+#define   DVS_YUV_ORDER_VYUY		REG_FIELD_PREP(DVS_YUV_ORDER_MASK, 3)
+#define   DVS_ROTATE_180		REG_BIT(15)
+#define   DVS_TRICKLE_FEED_DISABLE	REG_BIT(14)
+#define   DVS_TILED			REG_BIT(10)
+#define   DVS_DEST_KEY			REG_BIT(2)
+#define _DVSALINOFF		0x72184
+#define _DVSASTRIDE		0x72188
+#define _DVSAPOS		0x7218c
+#define   DVS_POS_Y_MASK		REG_GENMASK(31, 16)
+#define   DVS_POS_Y(y)			REG_FIELD_PREP(DVS_POS_Y_MASK, (y))
+#define   DVS_POS_X_MASK		REG_GENMASK(15, 0)
+#define   DVS_POS_X(x)			REG_FIELD_PREP(DVS_POS_X_MASK, (x))
+#define _DVSASIZE		0x72190
+#define   DVS_HEIGHT_MASK		REG_GENMASK(31, 16)
+#define   DVS_HEIGHT(h)			REG_FIELD_PREP(DVS_HEIGHT_MASK, (h))
+#define   DVS_WIDTH_MASK		REG_GENMASK(15, 0)
+#define   DVS_WIDTH(w)			REG_FIELD_PREP(DVS_WIDTH_MASK, (w))
+#define _DVSAKEYVAL		0x72194
+#define _DVSAKEYMSK		0x72198
+#define _DVSASURF		0x7219c
+#define   DVS_ADDR_MASK			REG_GENMASK(31, 12)
+#define _DVSAKEYMAXVAL		0x721a0
+#define _DVSATILEOFF		0x721a4
+#define   DVS_OFFSET_Y_MASK		REG_GENMASK(31, 16)
+#define   DVS_OFFSET_Y(y)		REG_FIELD_PREP(DVS_OFFSET_Y_MASK, (y))
+#define   DVS_OFFSET_X_MASK		REG_GENMASK(15, 0)
+#define   DVS_OFFSET_X(x)		REG_FIELD_PREP(DVS_OFFSET_X_MASK, (x))
+#define _DVSASURFLIVE		0x721ac
+#define _DVSAGAMC_G4X		0x721e0 /* g4x */
+#define _DVSASCALE		0x72204
+#define   DVS_SCALE_ENABLE		REG_BIT(31)
+#define   DVS_FILTER_MASK		REG_GENMASK(30, 29)
+#define   DVS_FILTER_MEDIUM		REG_FIELD_PREP(DVS_FILTER_MASK, 0)
+#define   DVS_FILTER_ENHANCING		REG_FIELD_PREP(DVS_FILTER_MASK, 1)
+#define   DVS_FILTER_SOFTENING		REG_FIELD_PREP(DVS_FILTER_MASK, 2)
+#define   DVS_VERTICAL_OFFSET_HALF	REG_BIT(28) /* must be enabled below */
+#define   DVS_VERTICAL_OFFSET_ENABLE	REG_BIT(27)
+#define   DVS_SRC_WIDTH_MASK		REG_GENMASK(26, 16)
+#define   DVS_SRC_WIDTH(w)		REG_FIELD_PREP(DVS_SRC_WIDTH_MASK, (w))
+#define   DVS_SRC_HEIGHT_MASK		REG_GENMASK(10, 0)
+#define   DVS_SRC_HEIGHT(h)		REG_FIELD_PREP(DVS_SRC_HEIGHT_MASK, (h))
+#define _DVSAGAMC_ILK		0x72300 /* ilk/snb */
+#define _DVSAGAMCMAX_ILK	0x72340 /* ilk/snb */
+
+#define _DVSBCNTR		0x73180
+#define _DVSBLINOFF		0x73184
+#define _DVSBSTRIDE		0x73188
+#define _DVSBPOS		0x7318c
+#define _DVSBSIZE		0x73190
+#define _DVSBKEYVAL		0x73194
+#define _DVSBKEYMSK		0x73198
+#define _DVSBSURF		0x7319c
+#define _DVSBKEYMAXVAL		0x731a0
+#define _DVSBTILEOFF		0x731a4
+#define _DVSBSURFLIVE		0x731ac
+#define _DVSBGAMC_G4X		0x731e0 /* g4x */
+#define _DVSBSCALE		0x73204
+#define _DVSBGAMC_ILK		0x73300 /* ilk/snb */
+#define _DVSBGAMCMAX_ILK	0x73340 /* ilk/snb */
+
+#define DVSCNTR(pipe) _MMIO_PIPE(pipe, _DVSACNTR, _DVSBCNTR)
+#define DVSLINOFF(pipe) _MMIO_PIPE(pipe, _DVSALINOFF, _DVSBLINOFF)
+#define DVSSTRIDE(pipe) _MMIO_PIPE(pipe, _DVSASTRIDE, _DVSBSTRIDE)
+#define DVSPOS(pipe) _MMIO_PIPE(pipe, _DVSAPOS, _DVSBPOS)
+#define DVSSURF(pipe) _MMIO_PIPE(pipe, _DVSASURF, _DVSBSURF)
+#define DVSKEYMAX(pipe) _MMIO_PIPE(pipe, _DVSAKEYMAXVAL, _DVSBKEYMAXVAL)
+#define DVSSIZE(pipe) _MMIO_PIPE(pipe, _DVSASIZE, _DVSBSIZE)
+#define DVSSCALE(pipe) _MMIO_PIPE(pipe, _DVSASCALE, _DVSBSCALE)
+#define DVSTILEOFF(pipe) _MMIO_PIPE(pipe, _DVSATILEOFF, _DVSBTILEOFF)
+#define DVSKEYVAL(pipe) _MMIO_PIPE(pipe, _DVSAKEYVAL, _DVSBKEYVAL)
+#define DVSKEYMSK(pipe) _MMIO_PIPE(pipe, _DVSAKEYMSK, _DVSBKEYMSK)
+#define DVSSURFLIVE(pipe) _MMIO_PIPE(pipe, _DVSASURFLIVE, _DVSBSURFLIVE)
+#define DVSGAMC_G4X(pipe, i) _MMIO(_PIPE(pipe, _DVSAGAMC_G4X, _DVSBGAMC_G4X) + (5 - (i)) * 4) /* 6 x u0.8 */
+#define DVSGAMC_ILK(pipe, i) _MMIO(_PIPE(pipe, _DVSAGAMC_ILK, _DVSBGAMC_ILK) + (i) * 4) /* 16 x u0.10 */
+#define DVSGAMCMAX_ILK(pipe, i) _MMIO(_PIPE(pipe, _DVSAGAMCMAX_ILK, _DVSBGAMCMAX_ILK) + (i) * 4) /* 3 x u1.10 */
+
+#define _SPRA_CTL		0x70280
+#define   SPRITE_ENABLE				REG_BIT(31)
+#define   SPRITE_PIPE_GAMMA_ENABLE		REG_BIT(30)
+#define   SPRITE_YUV_RANGE_CORRECTION_DISABLE	REG_BIT(28)
+#define   SPRITE_FORMAT_MASK			REG_GENMASK(27, 25)
+#define   SPRITE_FORMAT_YUV422			REG_FIELD_PREP(SPRITE_FORMAT_MASK, 0)
+#define   SPRITE_FORMAT_RGBX101010		REG_FIELD_PREP(SPRITE_FORMAT_MASK, 1)
+#define   SPRITE_FORMAT_RGBX888			REG_FIELD_PREP(SPRITE_FORMAT_MASK, 2)
+#define   SPRITE_FORMAT_RGBX161616		REG_FIELD_PREP(SPRITE_FORMAT_MASK, 3)
+#define   SPRITE_FORMAT_YUV444			REG_FIELD_PREP(SPRITE_FORMAT_MASK, 4)
+#define   SPRITE_FORMAT_XR_BGR101010		REG_FIELD_PREP(SPRITE_FORMAT_MASK, 5) /* Extended range */
+#define   SPRITE_PIPE_CSC_ENABLE		REG_BIT(24)
+#define   SPRITE_SOURCE_KEY			REG_BIT(22)
+#define   SPRITE_RGB_ORDER_RGBX			REG_BIT(20) /* only for 888 and 161616 */
+#define   SPRITE_YUV_TO_RGB_CSC_DISABLE		REG_BIT(19)
+#define   SPRITE_YUV_TO_RGB_CSC_FORMAT_BT709	REG_BIT(18) /* 0 is BT601 */
+#define   SPRITE_YUV_ORDER_MASK			REG_GENMASK(17, 16)
+#define   SPRITE_YUV_ORDER_YUYV			REG_FIELD_PREP(SPRITE_YUV_ORDER_MASK, 0)
+#define   SPRITE_YUV_ORDER_UYVY			REG_FIELD_PREP(SPRITE_YUV_ORDER_MASK, 1)
+#define   SPRITE_YUV_ORDER_YVYU			REG_FIELD_PREP(SPRITE_YUV_ORDER_MASK, 2)
+#define   SPRITE_YUV_ORDER_VYUY			REG_FIELD_PREP(SPRITE_YUV_ORDER_MASK, 3)
+#define   SPRITE_ROTATE_180			REG_BIT(15)
+#define   SPRITE_TRICKLE_FEED_DISABLE		REG_BIT(14)
+#define   SPRITE_PLANE_GAMMA_DISABLE		REG_BIT(13)
+#define   SPRITE_TILED				REG_BIT(10)
+#define   SPRITE_DEST_KEY			REG_BIT(2)
+#define _SPRA_LINOFF		0x70284
+#define _SPRA_STRIDE		0x70288
+#define _SPRA_POS		0x7028c
+#define   SPRITE_POS_Y_MASK	REG_GENMASK(31, 16)
+#define   SPRITE_POS_Y(y)	REG_FIELD_PREP(SPRITE_POS_Y_MASK, (y))
+#define   SPRITE_POS_X_MASK	REG_GENMASK(15, 0)
+#define   SPRITE_POS_X(x)	REG_FIELD_PREP(SPRITE_POS_X_MASK, (x))
+#define _SPRA_SIZE		0x70290
+#define   SPRITE_HEIGHT_MASK	REG_GENMASK(31, 16)
+#define   SPRITE_HEIGHT(h)	REG_FIELD_PREP(SPRITE_HEIGHT_MASK, (h))
+#define   SPRITE_WIDTH_MASK	REG_GENMASK(15, 0)
+#define   SPRITE_WIDTH(w)	REG_FIELD_PREP(SPRITE_WIDTH_MASK, (w))
+#define _SPRA_KEYVAL		0x70294
+#define _SPRA_KEYMSK		0x70298
+#define _SPRA_SURF		0x7029c
+#define   SPRITE_ADDR_MASK	REG_GENMASK(31, 12)
+#define _SPRA_KEYMAX		0x702a0
+#define _SPRA_TILEOFF		0x702a4
+#define   SPRITE_OFFSET_Y_MASK	REG_GENMASK(31, 16)
+#define   SPRITE_OFFSET_Y(y)	REG_FIELD_PREP(SPRITE_OFFSET_Y_MASK, (y))
+#define   SPRITE_OFFSET_X_MASK	REG_GENMASK(15, 0)
+#define   SPRITE_OFFSET_X(x)	REG_FIELD_PREP(SPRITE_OFFSET_X_MASK, (x))
+#define _SPRA_OFFSET		0x702a4
+#define _SPRA_SURFLIVE		0x702ac
+#define _SPRA_SCALE		0x70304
+#define   SPRITE_SCALE_ENABLE			REG_BIT(31)
+#define   SPRITE_FILTER_MASK			REG_GENMASK(30, 29)
+#define   SPRITE_FILTER_MEDIUM			REG_FIELD_PREP(SPRITE_FILTER_MASK, 0)
+#define   SPRITE_FILTER_ENHANCING		REG_FIELD_PREP(SPRITE_FILTER_MASK, 1)
+#define   SPRITE_FILTER_SOFTENING		REG_FIELD_PREP(SPRITE_FILTER_MASK, 2)
+#define   SPRITE_VERTICAL_OFFSET_HALF		REG_BIT(28) /* must be enabled below */
+#define   SPRITE_VERTICAL_OFFSET_ENABLE		REG_BIT(27)
+#define   SPRITE_SRC_WIDTH_MASK			REG_GENMASK(26, 16)
+#define   SPRITE_SRC_WIDTH(w)			REG_FIELD_PREP(SPRITE_SRC_WIDTH_MASK, (w))
+#define   SPRITE_SRC_HEIGHT_MASK		REG_GENMASK(10, 0)
+#define   SPRITE_SRC_HEIGHT(h)			REG_FIELD_PREP(SPRITE_SRC_HEIGHT_MASK, (h))
+#define _SPRA_GAMC		0x70400
+#define _SPRA_GAMC16		0x70440
+#define _SPRA_GAMC17		0x7044c
+
+#define _SPRB_CTL		0x71280
+#define _SPRB_LINOFF		0x71284
+#define _SPRB_STRIDE		0x71288
+#define _SPRB_POS		0x7128c
+#define _SPRB_SIZE		0x71290
+#define _SPRB_KEYVAL		0x71294
+#define _SPRB_KEYMSK		0x71298
+#define _SPRB_SURF		0x7129c
+#define _SPRB_KEYMAX		0x712a0
+#define _SPRB_TILEOFF		0x712a4
+#define _SPRB_OFFSET		0x712a4
+#define _SPRB_SURFLIVE		0x712ac
+#define _SPRB_SCALE		0x71304
+#define _SPRB_GAMC		0x71400
+#define _SPRB_GAMC16		0x71440
+#define _SPRB_GAMC17		0x7144c
+
+#define SPRCTL(pipe) _MMIO_PIPE(pipe, _SPRA_CTL, _SPRB_CTL)
+#define SPRLINOFF(pipe) _MMIO_PIPE(pipe, _SPRA_LINOFF, _SPRB_LINOFF)
+#define SPRSTRIDE(pipe) _MMIO_PIPE(pipe, _SPRA_STRIDE, _SPRB_STRIDE)
+#define SPRPOS(pipe) _MMIO_PIPE(pipe, _SPRA_POS, _SPRB_POS)
+#define SPRSIZE(pipe) _MMIO_PIPE(pipe, _SPRA_SIZE, _SPRB_SIZE)
+#define SPRKEYVAL(pipe) _MMIO_PIPE(pipe, _SPRA_KEYVAL, _SPRB_KEYVAL)
+#define SPRKEYMSK(pipe) _MMIO_PIPE(pipe, _SPRA_KEYMSK, _SPRB_KEYMSK)
+#define SPRSURF(pipe) _MMIO_PIPE(pipe, _SPRA_SURF, _SPRB_SURF)
+#define SPRKEYMAX(pipe) _MMIO_PIPE(pipe, _SPRA_KEYMAX, _SPRB_KEYMAX)
+#define SPRTILEOFF(pipe) _MMIO_PIPE(pipe, _SPRA_TILEOFF, _SPRB_TILEOFF)
+#define SPROFFSET(pipe) _MMIO_PIPE(pipe, _SPRA_OFFSET, _SPRB_OFFSET)
+#define SPRSCALE(pipe) _MMIO_PIPE(pipe, _SPRA_SCALE, _SPRB_SCALE)
+#define SPRGAMC(pipe, i) _MMIO(_PIPE(pipe, _SPRA_GAMC, _SPRB_GAMC) + (i) * 4) /* 16 x u0.10 */
+#define SPRGAMC16(pipe, i) _MMIO(_PIPE(pipe, _SPRA_GAMC16, _SPRB_GAMC16) + (i) * 4) /* 3 x u1.10 */
+#define SPRGAMC17(pipe, i) _MMIO(_PIPE(pipe, _SPRA_GAMC17, _SPRB_GAMC17) + (i) * 4) /* 3 x u2.10 */
+#define SPRSURFLIVE(pipe) _MMIO_PIPE(pipe, _SPRA_SURFLIVE, _SPRB_SURFLIVE)
+
+#define _SPACNTR		(VLV_DISPLAY_BASE + 0x72180)
+#define   SP_ENABLE			REG_BIT(31)
+#define   SP_PIPE_GAMMA_ENABLE		REG_BIT(30)
+#define   SP_FORMAT_MASK		REG_GENMASK(29, 26)
+#define   SP_FORMAT_YUV422		REG_FIELD_PREP(SP_FORMAT_MASK, 0)
+#define   SP_FORMAT_8BPP		REG_FIELD_PREP(SP_FORMAT_MASK, 2)
+#define   SP_FORMAT_BGR565		REG_FIELD_PREP(SP_FORMAT_MASK, 5)
+#define   SP_FORMAT_BGRX8888		REG_FIELD_PREP(SP_FORMAT_MASK, 6)
+#define   SP_FORMAT_BGRA8888		REG_FIELD_PREP(SP_FORMAT_MASK, 7)
+#define   SP_FORMAT_RGBX1010102		REG_FIELD_PREP(SP_FORMAT_MASK, 8)
+#define   SP_FORMAT_RGBA1010102		REG_FIELD_PREP(SP_FORMAT_MASK, 9)
+#define   SP_FORMAT_BGRX1010102		REG_FIELD_PREP(SP_FORMAT_MASK, 10) /* CHV pipe B */
+#define   SP_FORMAT_BGRA1010102		REG_FIELD_PREP(SP_FORMAT_MASK, 11) /* CHV pipe B */
+#define   SP_FORMAT_RGBX8888		REG_FIELD_PREP(SP_FORMAT_MASK, 14)
+#define   SP_FORMAT_RGBA8888		REG_FIELD_PREP(SP_FORMAT_MASK, 15)
+#define   SP_ALPHA_PREMULTIPLY		REG_BIT(23) /* CHV pipe B */
+#define   SP_SOURCE_KEY			REG_BIT(22)
+#define   SP_YUV_FORMAT_BT709		REG_BIT(18)
+#define   SP_YUV_ORDER_MASK		REG_GENMASK(17, 16)
+#define   SP_YUV_ORDER_YUYV		REG_FIELD_PREP(SP_YUV_ORDER_MASK, 0)
+#define   SP_YUV_ORDER_UYVY		REG_FIELD_PREP(SP_YUV_ORDER_MASK, 1)
+#define   SP_YUV_ORDER_YVYU		REG_FIELD_PREP(SP_YUV_ORDER_MASK, 2)
+#define   SP_YUV_ORDER_VYUY		REG_FIELD_PREP(SP_YUV_ORDER_MASK, 3)
+#define   SP_ROTATE_180			REG_BIT(15)
+#define   SP_TILED			REG_BIT(10)
+#define   SP_MIRROR			REG_BIT(8) /* CHV pipe B */
+#define _SPALINOFF		(VLV_DISPLAY_BASE + 0x72184)
+#define _SPASTRIDE		(VLV_DISPLAY_BASE + 0x72188)
+#define _SPAPOS			(VLV_DISPLAY_BASE + 0x7218c)
+#define   SP_POS_Y_MASK			REG_GENMASK(31, 16)
+#define   SP_POS_Y(y)			REG_FIELD_PREP(SP_POS_Y_MASK, (y))
+#define   SP_POS_X_MASK			REG_GENMASK(15, 0)
+#define   SP_POS_X(x)			REG_FIELD_PREP(SP_POS_X_MASK, (x))
+#define _SPASIZE		(VLV_DISPLAY_BASE + 0x72190)
+#define   SP_HEIGHT_MASK		REG_GENMASK(31, 16)
+#define   SP_HEIGHT(h)			REG_FIELD_PREP(SP_HEIGHT_MASK, (h))
+#define   SP_WIDTH_MASK			REG_GENMASK(15, 0)
+#define   SP_WIDTH(w)			REG_FIELD_PREP(SP_WIDTH_MASK, (w))
+#define _SPAKEYMINVAL		(VLV_DISPLAY_BASE + 0x72194)
+#define _SPAKEYMSK		(VLV_DISPLAY_BASE + 0x72198)
+#define _SPASURF		(VLV_DISPLAY_BASE + 0x7219c)
+#define   SP_ADDR_MASK			REG_GENMASK(31, 12)
+#define _SPAKEYMAXVAL		(VLV_DISPLAY_BASE + 0x721a0)
+#define _SPATILEOFF		(VLV_DISPLAY_BASE + 0x721a4)
+#define   SP_OFFSET_Y_MASK		REG_GENMASK(31, 16)
+#define   SP_OFFSET_Y(y)		REG_FIELD_PREP(SP_OFFSET_Y_MASK, (y))
+#define   SP_OFFSET_X_MASK		REG_GENMASK(15, 0)
+#define   SP_OFFSET_X(x)		REG_FIELD_PREP(SP_OFFSET_X_MASK, (x))
+#define _SPACONSTALPHA		(VLV_DISPLAY_BASE + 0x721a8)
+#define   SP_CONST_ALPHA_ENABLE		REG_BIT(31)
+#define   SP_CONST_ALPHA_MASK		REG_GENMASK(7, 0)
+#define   SP_CONST_ALPHA(alpha)		REG_FIELD_PREP(SP_CONST_ALPHA_MASK, (alpha))
+#define _SPACLRC0		(VLV_DISPLAY_BASE + 0x721d0)
+#define   SP_CONTRAST_MASK		REG_GENMASK(26, 18)
+#define   SP_CONTRAST(x)		REG_FIELD_PREP(SP_CONTRAST_MASK, (x)) /* u3.6 */
+#define   SP_BRIGHTNESS_MASK		REG_GENMASK(7, 0)
+#define   SP_BRIGHTNESS(x)		REG_FIELD_PREP(SP_BRIGHTNESS_MASK, (x)) /* s8 */
+#define _SPACLRC1		(VLV_DISPLAY_BASE + 0x721d4)
+#define   SP_SH_SIN_MASK		REG_GENMASK(26, 16)
+#define   SP_SH_SIN(x)			REG_FIELD_PREP(SP_SH_SIN_MASK, (x)) /* s4.7 */
+#define   SP_SH_COS_MASK		REG_GENMASK(9, 0)
+#define   SP_SH_COS(x)			REG_FIELD_PREP(SP_SH_COS_MASK, (x)) /* u3.7 */
+#define _SPAGAMC		(VLV_DISPLAY_BASE + 0x721e0)
+
+#define _SPBCNTR		(VLV_DISPLAY_BASE + 0x72280)
+#define _SPBLINOFF		(VLV_DISPLAY_BASE + 0x72284)
+#define _SPBSTRIDE		(VLV_DISPLAY_BASE + 0x72288)
+#define _SPBPOS			(VLV_DISPLAY_BASE + 0x7228c)
+#define _SPBSIZE		(VLV_DISPLAY_BASE + 0x72290)
+#define _SPBKEYMINVAL		(VLV_DISPLAY_BASE + 0x72294)
+#define _SPBKEYMSK		(VLV_DISPLAY_BASE + 0x72298)
+#define _SPBSURF		(VLV_DISPLAY_BASE + 0x7229c)
+#define _SPBKEYMAXVAL		(VLV_DISPLAY_BASE + 0x722a0)
+#define _SPBTILEOFF		(VLV_DISPLAY_BASE + 0x722a4)
+#define _SPBCONSTALPHA		(VLV_DISPLAY_BASE + 0x722a8)
+#define _SPBCLRC0		(VLV_DISPLAY_BASE + 0x722d0)
+#define _SPBCLRC1		(VLV_DISPLAY_BASE + 0x722d4)
+#define _SPBGAMC		(VLV_DISPLAY_BASE + 0x722e0)
+
+#define _VLV_SPR(pipe, plane_id, reg_a, reg_b) \
+	_PIPE((pipe) * 2 + (plane_id) - PLANE_SPRITE0, (reg_a), (reg_b))
+#define _MMIO_VLV_SPR(pipe, plane_id, reg_a, reg_b) \
+	_MMIO(_VLV_SPR((pipe), (plane_id), (reg_a), (reg_b)))
+
+#define SPCNTR(pipe, plane_id)		_MMIO_VLV_SPR((pipe), (plane_id), _SPACNTR, _SPBCNTR)
+#define SPLINOFF(pipe, plane_id)	_MMIO_VLV_SPR((pipe), (plane_id), _SPALINOFF, _SPBLINOFF)
+#define SPSTRIDE(pipe, plane_id)	_MMIO_VLV_SPR((pipe), (plane_id), _SPASTRIDE, _SPBSTRIDE)
+#define SPPOS(pipe, plane_id)		_MMIO_VLV_SPR((pipe), (plane_id), _SPAPOS, _SPBPOS)
+#define SPSIZE(pipe, plane_id)		_MMIO_VLV_SPR((pipe), (plane_id), _SPASIZE, _SPBSIZE)
+#define SPKEYMINVAL(pipe, plane_id)	_MMIO_VLV_SPR((pipe), (plane_id), _SPAKEYMINVAL, _SPBKEYMINVAL)
+#define SPKEYMSK(pipe, plane_id)	_MMIO_VLV_SPR((pipe), (plane_id), _SPAKEYMSK, _SPBKEYMSK)
+#define SPSURF(pipe, plane_id)		_MMIO_VLV_SPR((pipe), (plane_id), _SPASURF, _SPBSURF)
+#define SPKEYMAXVAL(pipe, plane_id)	_MMIO_VLV_SPR((pipe), (plane_id), _SPAKEYMAXVAL, _SPBKEYMAXVAL)
+#define SPTILEOFF(pipe, plane_id)	_MMIO_VLV_SPR((pipe), (plane_id), _SPATILEOFF, _SPBTILEOFF)
+#define SPCONSTALPHA(pipe, plane_id)	_MMIO_VLV_SPR((pipe), (plane_id), _SPACONSTALPHA, _SPBCONSTALPHA)
+#define SPCLRC0(pipe, plane_id)		_MMIO_VLV_SPR((pipe), (plane_id), _SPACLRC0, _SPBCLRC0)
+#define SPCLRC1(pipe, plane_id)		_MMIO_VLV_SPR((pipe), (plane_id), _SPACLRC1, _SPBCLRC1)
+#define SPGAMC(pipe, plane_id, i)	_MMIO(_VLV_SPR((pipe), (plane_id), _SPAGAMC, _SPBGAMC) + (5 - (i)) * 4) /* 6 x u0.10 */
+
+/*
+ * CHV pipe B sprite CSC
+ *
+ * |cr|   |c0 c1 c2|   |cr + cr_ioff|   |cr_ooff|
+ * |yg| = |c3 c4 c5| x |yg + yg_ioff| + |yg_ooff|
+ * |cb|   |c6 c7 c8|   |cb + cr_ioff|   |cb_ooff|
+ */
+#define _MMIO_CHV_SPCSC(plane_id, reg) \
+	_MMIO(VLV_DISPLAY_BASE + ((plane_id) - PLANE_SPRITE0) * 0x1000 + (reg))
+
+#define SPCSCYGOFF(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d900)
+#define SPCSCCBOFF(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d904)
+#define SPCSCCROFF(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d908)
+#define  SPCSC_OOFF_MASK	REG_GENMASK(26, 16)
+#define  SPCSC_OOFF(x)		REG_FIELD_PREP(SPCSC_OOFF_MASK, (x) & 0x7ff) /* s11 */
+#define  SPCSC_IOFF_MASK	REG_GENMASK(10, 0)
+#define  SPCSC_IOFF(x)		REG_FIELD_PREP(SPCSC_IOFF_MASK, (x) & 0x7ff) /* s11 */
+
+#define SPCSCC01(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d90c)
+#define SPCSCC23(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d910)
+#define SPCSCC45(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d914)
+#define SPCSCC67(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d918)
+#define SPCSCC8(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d91c)
+#define  SPCSC_C1_MASK		REG_GENMASK(30, 16)
+#define  SPCSC_C1(x)		REG_FIELD_PREP(SPCSC_C1_MASK, (x) & 0x7fff) /* s3.12 */
+#define  SPCSC_C0_MASK		REG_GENMASK(14, 0)
+#define  SPCSC_C0(x)		REG_FIELD_PREP(SPCSC_C0_MASK, (x) & 0x7fff) /* s3.12 */
+
+#define SPCSCYGICLAMP(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d920)
+#define SPCSCCBICLAMP(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d924)
+#define SPCSCCRICLAMP(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d928)
+#define  SPCSC_IMAX_MASK	REG_GENMASK(26, 16)
+#define  SPCSC_IMAX(x)		REG_FIELD_PREP(SPCSC_IMAX_MASK, (x) & 0x7ff) /* s11 */
+#define  SPCSC_IMIN_MASK	REG_GENMASK(10, 0)
+#define  SPCSC_IMIN(x)		REG_FIELD_PREP(SPCSC_IMIN_MASK, (x) & 0x7ff) /* s11 */
+
+#define SPCSCYGOCLAMP(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d92c)
+#define SPCSCCBOCLAMP(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d930)
+#define SPCSCCROCLAMP(plane_id)	_MMIO_CHV_SPCSC(plane_id, 0x6d934)
+#define  SPCSC_OMAX_MASK	REG_GENMASK(25, 16)
+#define  SPCSC_OMAX(x)		REG_FIELD_PREP(SPCSC_OMAX_MASK, (x)) /* u10 */
+#define  SPCSC_OMIN_MASK	REG_GENMASK(9, 0)
+#define  SPCSC_OMIN(x)		REG_FIELD_PREP(SPCSC_OMIN_MASK, (x)) /* u10 */
+
+/* Skylake plane registers */
+
+#define _PLANE_CTL_1_A				0x70180
+#define _PLANE_CTL_2_A				0x70280
+#define _PLANE_CTL_3_A				0x70380
+#define   PLANE_CTL_ENABLE			REG_BIT(31)
+#define   PLANE_CTL_ARB_SLOTS_MASK		REG_GENMASK(30, 28) /* icl+ */
+#define   PLANE_CTL_ARB_SLOTS(x)		REG_FIELD_PREP(PLANE_CTL_ARB_SLOTS_MASK, (x)) /* icl+ */
+#define   PLANE_CTL_PIPE_GAMMA_ENABLE		REG_BIT(30) /* Pre-GLK */
+#define   PLANE_CTL_YUV_RANGE_CORRECTION_DISABLE	REG_BIT(28)
+/*
+ * ICL+ uses the same PLANE_CTL_FORMAT bits, but the field definition
+ * expanded to include bit 23 as well. However, the shift-24 based values
+ * correctly map to the same formats in ICL, as long as bit 23 is set to 0
+ */
+#define   PLANE_CTL_FORMAT_MASK_SKL		REG_GENMASK(27, 24) /* pre-icl */
+#define   PLANE_CTL_FORMAT_MASK_ICL		REG_GENMASK(27, 23) /* icl+ */
+#define   PLANE_CTL_FORMAT_YUV422		REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 0)
+#define   PLANE_CTL_FORMAT_NV12			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 1)
+#define   PLANE_CTL_FORMAT_XRGB_2101010		REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 2)
+#define   PLANE_CTL_FORMAT_P010			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 3)
+#define   PLANE_CTL_FORMAT_XRGB_8888		REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 4)
+#define   PLANE_CTL_FORMAT_P012			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 5)
+#define   PLANE_CTL_FORMAT_XRGB_16161616F	REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 6)
+#define   PLANE_CTL_FORMAT_P016			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 7)
+#define   PLANE_CTL_FORMAT_XYUV			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 8)
+#define   PLANE_CTL_FORMAT_INDEXED		REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 12)
+#define   PLANE_CTL_FORMAT_RGB_565		REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_SKL, 14)
+#define   PLANE_CTL_FORMAT_Y210			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_ICL, 1)
+#define   PLANE_CTL_FORMAT_Y212			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_ICL, 3)
+#define   PLANE_CTL_FORMAT_Y216			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_ICL, 5)
+#define   PLANE_CTL_FORMAT_Y410			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_ICL, 7)
+#define   PLANE_CTL_FORMAT_Y412			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_ICL, 9)
+#define   PLANE_CTL_FORMAT_Y416			REG_FIELD_PREP(PLANE_CTL_FORMAT_MASK_ICL, 11)
+#define   PLANE_CTL_PIPE_CSC_ENABLE		REG_BIT(23) /* Pre-GLK */
+#define   PLANE_CTL_KEY_ENABLE_MASK		REG_GENMASK(22, 21)
+#define   PLANE_CTL_KEY_ENABLE_SOURCE		REG_FIELD_PREP(PLANE_CTL_KEY_ENABLE_MASK, 1)
+#define   PLANE_CTL_KEY_ENABLE_DESTINATION	REG_FIELD_PREP(PLANE_CTL_KEY_ENABLE_MASK, 2)
+#define   PLANE_CTL_ORDER_RGBX			REG_BIT(20)
+#define   PLANE_CTL_YUV420_Y_PLANE		REG_BIT(19)
+#define   PLANE_CTL_YUV_TO_RGB_CSC_FORMAT_BT709	REG_BIT(18)
+#define   PLANE_CTL_YUV422_ORDER_MASK		REG_GENMASK(17, 16)
+#define   PLANE_CTL_YUV422_ORDER_YUYV		REG_FIELD_PREP(PLANE_CTL_YUV422_ORDER_MASK, 0)
+#define   PLANE_CTL_YUV422_ORDER_UYVY		REG_FIELD_PREP(PLANE_CTL_YUV422_ORDER_MASK, 1)
+#define   PLANE_CTL_YUV422_ORDER_YVYU		REG_FIELD_PREP(PLANE_CTL_YUV422_ORDER_MASK, 2)
+#define   PLANE_CTL_YUV422_ORDER_VYUY		REG_FIELD_PREP(PLANE_CTL_YUV422_ORDER_MASK, 3)
+#define   PLANE_CTL_RENDER_DECOMPRESSION_ENABLE	REG_BIT(15)
+#define   PLANE_CTL_TRICKLE_FEED_DISABLE	REG_BIT(14)
+#define   PLANE_CTL_CLEAR_COLOR_DISABLE		REG_BIT(13) /* TGL+ */
+#define   PLANE_CTL_PLANE_GAMMA_DISABLE		REG_BIT(13) /* Pre-GLK */
+#define   PLANE_CTL_TILED_MASK			REG_GENMASK(12, 10)
+#define   PLANE_CTL_TILED_LINEAR		REG_FIELD_PREP(PLANE_CTL_TILED_MASK, 0)
+#define   PLANE_CTL_TILED_X			REG_FIELD_PREP(PLANE_CTL_TILED_MASK, 1)
+#define   PLANE_CTL_TILED_Y			REG_FIELD_PREP(PLANE_CTL_TILED_MASK, 4)
+#define   PLANE_CTL_TILED_YF			REG_FIELD_PREP(PLANE_CTL_TILED_MASK, 5)
+#define   PLANE_CTL_TILED_4                     REG_FIELD_PREP(PLANE_CTL_TILED_MASK, 5)
+#define   PLANE_CTL_ASYNC_FLIP			REG_BIT(9)
+#define   PLANE_CTL_FLIP_HORIZONTAL		REG_BIT(8)
+#define   PLANE_CTL_MEDIA_DECOMPRESSION_ENABLE	REG_BIT(4) /* TGL+ */
+#define   PLANE_CTL_ALPHA_MASK			REG_GENMASK(5, 4) /* Pre-GLK */
+#define   PLANE_CTL_ALPHA_DISABLE		REG_FIELD_PREP(PLANE_CTL_ALPHA_MASK, 0)
+#define   PLANE_CTL_ALPHA_SW_PREMULTIPLY	REG_FIELD_PREP(PLANE_CTL_ALPHA_MASK, 2)
+#define   PLANE_CTL_ALPHA_HW_PREMULTIPLY	REG_FIELD_PREP(PLANE_CTL_ALPHA_MASK, 3)
+#define   PLANE_CTL_ROTATE_MASK			REG_GENMASK(1, 0)
+#define   PLANE_CTL_ROTATE_0			REG_FIELD_PREP(PLANE_CTL_ROTATE_MASK, 0)
+#define   PLANE_CTL_ROTATE_90			REG_FIELD_PREP(PLANE_CTL_ROTATE_MASK, 1)
+#define   PLANE_CTL_ROTATE_180			REG_FIELD_PREP(PLANE_CTL_ROTATE_MASK, 2)
+#define   PLANE_CTL_ROTATE_270			REG_FIELD_PREP(PLANE_CTL_ROTATE_MASK, 3)
+#define _PLANE_STRIDE_1_A			0x70188
+#define _PLANE_STRIDE_2_A			0x70288
+#define _PLANE_STRIDE_3_A			0x70388
+#define   PLANE_STRIDE__MASK			REG_GENMASK(11, 0)
+#define   PLANE_STRIDE_(stride)			REG_FIELD_PREP(PLANE_STRIDE__MASK, (stride))
+#define _PLANE_POS_1_A				0x7018c
+#define _PLANE_POS_2_A				0x7028c
+#define _PLANE_POS_3_A				0x7038c
+#define   PLANE_POS_Y_MASK			REG_GENMASK(31, 16)
+#define   PLANE_POS_Y(y)			REG_FIELD_PREP(PLANE_POS_Y_MASK, (y))
+#define   PLANE_POS_X_MASK			REG_GENMASK(15, 0)
+#define   PLANE_POS_X(x)			REG_FIELD_PREP(PLANE_POS_X_MASK, (x))
+#define _PLANE_SIZE_1_A				0x70190
+#define _PLANE_SIZE_2_A				0x70290
+#define _PLANE_SIZE_3_A				0x70390
+#define   PLANE_HEIGHT_MASK			REG_GENMASK(31, 16)
+#define   PLANE_HEIGHT(h)			REG_FIELD_PREP(PLANE_HEIGHT_MASK, (h))
+#define   PLANE_WIDTH_MASK			REG_GENMASK(15, 0)
+#define   PLANE_WIDTH(w)			REG_FIELD_PREP(PLANE_WIDTH_MASK, (w))
+#define _PLANE_SURF_1_A				0x7019c
+#define _PLANE_SURF_2_A				0x7029c
+#define _PLANE_SURF_3_A				0x7039c
+#define   PLANE_SURF_ADDR_MASK			REG_GENMASK(31, 12)
+#define   PLANE_SURF_DECRYPT			REG_BIT(2)
+#define _PLANE_OFFSET_1_A			0x701a4
+#define _PLANE_OFFSET_2_A			0x702a4
+#define _PLANE_OFFSET_3_A			0x703a4
+#define   PLANE_OFFSET_Y_MASK			REG_GENMASK(31, 16)
+#define   PLANE_OFFSET_Y(y)			REG_FIELD_PREP(PLANE_OFFSET_Y_MASK, (y))
+#define   PLANE_OFFSET_X_MASK			REG_GENMASK(15, 0)
+#define   PLANE_OFFSET_X(x)			REG_FIELD_PREP(PLANE_OFFSET_X_MASK, (x))
+#define _PLANE_KEYVAL_1_A			0x70194
+#define _PLANE_KEYVAL_2_A			0x70294
+#define _PLANE_KEYMSK_1_A			0x70198
+#define _PLANE_KEYMSK_2_A			0x70298
+#define  PLANE_KEYMSK_ALPHA_ENABLE		(1 << 31)
+#define _PLANE_KEYMAX_1_A			0x701a0
+#define _PLANE_KEYMAX_2_A			0x702a0
+#define  PLANE_KEYMAX_ALPHA(a)			((a) << 24)
+#define _PLANE_CC_VAL_1_A			0x701b4
+#define _PLANE_CC_VAL_2_A			0x702b4
+#define _PLANE_AUX_DIST_1_A			0x701c0
+#define   PLANE_AUX_DISTANCE_MASK		REG_GENMASK(31, 12)
+#define   PLANE_AUX_STRIDE_MASK			REG_GENMASK(11, 0)
+#define   PLANE_AUX_STRIDE(stride)		REG_FIELD_PREP(PLANE_AUX_STRIDE_MASK, (stride))
+#define _PLANE_AUX_DIST_2_A			0x702c0
+#define _PLANE_AUX_OFFSET_1_A			0x701c4
+#define _PLANE_AUX_OFFSET_2_A			0x702c4
+#define _PLANE_CUS_CTL_1_A			0x701c8
+#define _PLANE_CUS_CTL_2_A			0x702c8
+#define   PLANE_CUS_ENABLE			REG_BIT(31)
+#define   PLANE_CUS_Y_PLANE_MASK			REG_BIT(30)
+#define   PLANE_CUS_Y_PLANE_4_RKL		REG_FIELD_PREP(PLANE_CUS_Y_PLANE_MASK, 0)
+#define   PLANE_CUS_Y_PLANE_5_RKL		REG_FIELD_PREP(PLANE_CUS_Y_PLANE_MASK, 1)
+#define   PLANE_CUS_Y_PLANE_6_ICL		REG_FIELD_PREP(PLANE_CUS_Y_PLANE_MASK, 0)
+#define   PLANE_CUS_Y_PLANE_7_ICL		REG_FIELD_PREP(PLANE_CUS_Y_PLANE_MASK, 1)
+#define   PLANE_CUS_HPHASE_SIGN_NEGATIVE		REG_BIT(19)
+#define   PLANE_CUS_HPHASE_MASK			REG_GENMASK(17, 16)
+#define   PLANE_CUS_HPHASE_0			REG_FIELD_PREP(PLANE_CUS_HPHASE_MASK, 0)
+#define   PLANE_CUS_HPHASE_0_25			REG_FIELD_PREP(PLANE_CUS_HPHASE_MASK, 1)
+#define   PLANE_CUS_HPHASE_0_5			REG_FIELD_PREP(PLANE_CUS_HPHASE_MASK, 2)
+#define   PLANE_CUS_VPHASE_SIGN_NEGATIVE		REG_BIT(15)
+#define   PLANE_CUS_VPHASE_MASK			REG_GENMASK(13, 12)
+#define   PLANE_CUS_VPHASE_0			REG_FIELD_PREP(PLANE_CUS_VPHASE_MASK, 0)
+#define   PLANE_CUS_VPHASE_0_25			REG_FIELD_PREP(PLANE_CUS_VPHASE_MASK, 1)
+#define   PLANE_CUS_VPHASE_0_5			REG_FIELD_PREP(PLANE_CUS_VPHASE_MASK, 2)
+#define _PLANE_COLOR_CTL_1_A			0x701CC /* GLK+ */
+#define _PLANE_COLOR_CTL_2_A			0x702CC /* GLK+ */
+#define _PLANE_COLOR_CTL_3_A			0x703CC /* GLK+ */
+#define   PLANE_COLOR_PIPE_GAMMA_ENABLE			REG_BIT(30) /* Pre-ICL */
+#define   PLANE_COLOR_YUV_RANGE_CORRECTION_DISABLE	REG_BIT(28)
+#define   PLANE_COLOR_PIPE_CSC_ENABLE			REG_BIT(23) /* Pre-ICL */
+#define   PLANE_COLOR_PLANE_CSC_ENABLE			REG_BIT(21) /* ICL+ */
+#define   PLANE_COLOR_INPUT_CSC_ENABLE			REG_BIT(20) /* ICL+ */
+#define   PLANE_COLOR_CSC_MODE_MASK			REG_GENMASK(19, 17)
+#define   PLANE_COLOR_CSC_MODE_BYPASS			REG_FIELD_PREP(PLANE_COLOR_CSC_MODE_MASK, 0)
+#define   PLANE_COLOR_CSC_MODE_YUV601_TO_RGB601		REG_FIELD_PREP(PLANE_COLOR_CSC_MODE_MASK, 1)
+#define   PLANE_COLOR_CSC_MODE_YUV709_TO_RGB709		REG_FIELD_PREP(PLANE_COLOR_CSC_MODE_MASK, 2)
+#define   PLANE_COLOR_CSC_MODE_YUV2020_TO_RGB2020	REG_FIELD_PREP(PLANE_COLOR_CSC_MODE_MASK, 3)
+#define   PLANE_COLOR_CSC_MODE_RGB709_TO_RGB2020	REG_FIELD_PREP(PLANE_COLOR_CSC_MODE_MASK, 4)
+#define   PLANE_COLOR_PLANE_GAMMA_DISABLE		REG_BIT(13)
+#define   PLANE_COLOR_ALPHA_MASK			REG_GENMASK(5, 4)
+#define   PLANE_COLOR_ALPHA_DISABLE			REG_FIELD_PREP(PLANE_COLOR_ALPHA_MASK, 0)
+#define   PLANE_COLOR_ALPHA_SW_PREMULTIPLY		REG_FIELD_PREP(PLANE_COLOR_ALPHA_MASK, 2)
+#define   PLANE_COLOR_ALPHA_HW_PREMULTIPLY		REG_FIELD_PREP(PLANE_COLOR_ALPHA_MASK, 3)
+#define _PLANE_BUF_CFG_1_A			0x7027c
+#define _PLANE_BUF_CFG_2_A			0x7037c
+#define _PLANE_NV12_BUF_CFG_1_A		0x70278
+#define _PLANE_NV12_BUF_CFG_2_A		0x70378
+
+#define _PLANE_CC_VAL_1_B		0x711b4
+#define _PLANE_CC_VAL_2_B		0x712b4
+#define _PLANE_CC_VAL_1(pipe, dw)	(_PIPE(pipe, _PLANE_CC_VAL_1_A, _PLANE_CC_VAL_1_B) + (dw) * 4)
+#define _PLANE_CC_VAL_2(pipe, dw)	(_PIPE(pipe, _PLANE_CC_VAL_2_A, _PLANE_CC_VAL_2_B) + (dw) * 4)
+#define PLANE_CC_VAL(pipe, plane, dw) \
+	_MMIO_PLANE((plane), _PLANE_CC_VAL_1((pipe), (dw)), _PLANE_CC_VAL_2((pipe), (dw)))
+
+/* Input CSC Register Definitions */
+#define _PLANE_INPUT_CSC_RY_GY_1_A	0x701E0
+#define _PLANE_INPUT_CSC_RY_GY_2_A	0x702E0
+
+#define _PLANE_INPUT_CSC_RY_GY_1_B	0x711E0
+#define _PLANE_INPUT_CSC_RY_GY_2_B	0x712E0
+
+#define _PLANE_INPUT_CSC_RY_GY_1(pipe)	\
+	_PIPE(pipe, _PLANE_INPUT_CSC_RY_GY_1_A, \
+	     _PLANE_INPUT_CSC_RY_GY_1_B)
+#define _PLANE_INPUT_CSC_RY_GY_2(pipe)	\
+	_PIPE(pipe, _PLANE_INPUT_CSC_RY_GY_2_A, \
+	     _PLANE_INPUT_CSC_RY_GY_2_B)
+
+#define PLANE_INPUT_CSC_COEFF(pipe, plane, index)	\
+	_MMIO_PLANE(plane, _PLANE_INPUT_CSC_RY_GY_1(pipe) +  (index) * 4, \
+		    _PLANE_INPUT_CSC_RY_GY_2(pipe) + (index) * 4)
+
+#define _PLANE_INPUT_CSC_PREOFF_HI_1_A		0x701F8
+#define _PLANE_INPUT_CSC_PREOFF_HI_2_A		0x702F8
+
+#define _PLANE_INPUT_CSC_PREOFF_HI_1_B		0x711F8
+#define _PLANE_INPUT_CSC_PREOFF_HI_2_B		0x712F8
+
+#define _PLANE_INPUT_CSC_PREOFF_HI_1(pipe)	\
+	_PIPE(pipe, _PLANE_INPUT_CSC_PREOFF_HI_1_A, \
+	     _PLANE_INPUT_CSC_PREOFF_HI_1_B)
+#define _PLANE_INPUT_CSC_PREOFF_HI_2(pipe)	\
+	_PIPE(pipe, _PLANE_INPUT_CSC_PREOFF_HI_2_A, \
+	     _PLANE_INPUT_CSC_PREOFF_HI_2_B)
+#define PLANE_INPUT_CSC_PREOFF(pipe, plane, index)	\
+	_MMIO_PLANE(plane, _PLANE_INPUT_CSC_PREOFF_HI_1(pipe) + (index) * 4, \
+		    _PLANE_INPUT_CSC_PREOFF_HI_2(pipe) + (index) * 4)
+
+#define _PLANE_INPUT_CSC_POSTOFF_HI_1_A		0x70204
+#define _PLANE_INPUT_CSC_POSTOFF_HI_2_A		0x70304
+
+#define _PLANE_INPUT_CSC_POSTOFF_HI_1_B		0x71204
+#define _PLANE_INPUT_CSC_POSTOFF_HI_2_B		0x71304
+
+#define _PLANE_INPUT_CSC_POSTOFF_HI_1(pipe)	\
+	_PIPE(pipe, _PLANE_INPUT_CSC_POSTOFF_HI_1_A, \
+	     _PLANE_INPUT_CSC_POSTOFF_HI_1_B)
+#define _PLANE_INPUT_CSC_POSTOFF_HI_2(pipe)	\
+	_PIPE(pipe, _PLANE_INPUT_CSC_POSTOFF_HI_2_A, \
+	     _PLANE_INPUT_CSC_POSTOFF_HI_2_B)
+#define PLANE_INPUT_CSC_POSTOFF(pipe, plane, index)	\
+	_MMIO_PLANE(plane, _PLANE_INPUT_CSC_POSTOFF_HI_1(pipe) + (index) * 4, \
+		    _PLANE_INPUT_CSC_POSTOFF_HI_2(pipe) + (index) * 4)
+
+#define _PLANE_CTL_1_B				0x71180
+#define _PLANE_CTL_2_B				0x71280
+#define _PLANE_CTL_3_B				0x71380
+#define _PLANE_CTL_1(pipe)	_PIPE(pipe, _PLANE_CTL_1_A, _PLANE_CTL_1_B)
+#define _PLANE_CTL_2(pipe)	_PIPE(pipe, _PLANE_CTL_2_A, _PLANE_CTL_2_B)
+#define _PLANE_CTL_3(pipe)	_PIPE(pipe, _PLANE_CTL_3_A, _PLANE_CTL_3_B)
+#define PLANE_CTL(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_CTL_1(pipe), _PLANE_CTL_2(pipe))
+
+#define _PLANE_STRIDE_1_B			0x71188
+#define _PLANE_STRIDE_2_B			0x71288
+#define _PLANE_STRIDE_3_B			0x71388
+#define _PLANE_STRIDE_1(pipe)	\
+	_PIPE(pipe, _PLANE_STRIDE_1_A, _PLANE_STRIDE_1_B)
+#define _PLANE_STRIDE_2(pipe)	\
+	_PIPE(pipe, _PLANE_STRIDE_2_A, _PLANE_STRIDE_2_B)
+#define _PLANE_STRIDE_3(pipe)	\
+	_PIPE(pipe, _PLANE_STRIDE_3_A, _PLANE_STRIDE_3_B)
+#define PLANE_STRIDE(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_STRIDE_1(pipe), _PLANE_STRIDE_2(pipe))
+
+#define _PLANE_POS_1_B				0x7118c
+#define _PLANE_POS_2_B				0x7128c
+#define _PLANE_POS_3_B				0x7138c
+#define _PLANE_POS_1(pipe)	_PIPE(pipe, _PLANE_POS_1_A, _PLANE_POS_1_B)
+#define _PLANE_POS_2(pipe)	_PIPE(pipe, _PLANE_POS_2_A, _PLANE_POS_2_B)
+#define _PLANE_POS_3(pipe)	_PIPE(pipe, _PLANE_POS_3_A, _PLANE_POS_3_B)
+#define PLANE_POS(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_POS_1(pipe), _PLANE_POS_2(pipe))
+
+#define _PLANE_SIZE_1_B				0x71190
+#define _PLANE_SIZE_2_B				0x71290
+#define _PLANE_SIZE_3_B				0x71390
+#define _PLANE_SIZE_1(pipe)	_PIPE(pipe, _PLANE_SIZE_1_A, _PLANE_SIZE_1_B)
+#define _PLANE_SIZE_2(pipe)	_PIPE(pipe, _PLANE_SIZE_2_A, _PLANE_SIZE_2_B)
+#define _PLANE_SIZE_3(pipe)	_PIPE(pipe, _PLANE_SIZE_3_A, _PLANE_SIZE_3_B)
+#define PLANE_SIZE(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_SIZE_1(pipe), _PLANE_SIZE_2(pipe))
+
+#define _PLANE_SURF_1_B				0x7119c
+#define _PLANE_SURF_2_B				0x7129c
+#define _PLANE_SURF_3_B				0x7139c
+#define _PLANE_SURF_1(pipe)	_PIPE(pipe, _PLANE_SURF_1_A, _PLANE_SURF_1_B)
+#define _PLANE_SURF_2(pipe)	_PIPE(pipe, _PLANE_SURF_2_A, _PLANE_SURF_2_B)
+#define _PLANE_SURF_3(pipe)	_PIPE(pipe, _PLANE_SURF_3_A, _PLANE_SURF_3_B)
+#define PLANE_SURF(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_SURF_1(pipe), _PLANE_SURF_2(pipe))
+
+#define _PLANE_OFFSET_1_B			0x711a4
+#define _PLANE_OFFSET_2_B			0x712a4
+#define _PLANE_OFFSET_1(pipe) _PIPE(pipe, _PLANE_OFFSET_1_A, _PLANE_OFFSET_1_B)
+#define _PLANE_OFFSET_2(pipe) _PIPE(pipe, _PLANE_OFFSET_2_A, _PLANE_OFFSET_2_B)
+#define PLANE_OFFSET(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_OFFSET_1(pipe), _PLANE_OFFSET_2(pipe))
+
+#define _PLANE_KEYVAL_1_B			0x71194
+#define _PLANE_KEYVAL_2_B			0x71294
+#define _PLANE_KEYVAL_1(pipe) _PIPE(pipe, _PLANE_KEYVAL_1_A, _PLANE_KEYVAL_1_B)
+#define _PLANE_KEYVAL_2(pipe) _PIPE(pipe, _PLANE_KEYVAL_2_A, _PLANE_KEYVAL_2_B)
+#define PLANE_KEYVAL(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_KEYVAL_1(pipe), _PLANE_KEYVAL_2(pipe))
+
+#define _PLANE_KEYMSK_1_B			0x71198
+#define _PLANE_KEYMSK_2_B			0x71298
+#define _PLANE_KEYMSK_1(pipe) _PIPE(pipe, _PLANE_KEYMSK_1_A, _PLANE_KEYMSK_1_B)
+#define _PLANE_KEYMSK_2(pipe) _PIPE(pipe, _PLANE_KEYMSK_2_A, _PLANE_KEYMSK_2_B)
+#define PLANE_KEYMSK(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_KEYMSK_1(pipe), _PLANE_KEYMSK_2(pipe))
+
+#define _PLANE_KEYMAX_1_B			0x711a0
+#define _PLANE_KEYMAX_2_B			0x712a0
+#define _PLANE_KEYMAX_1(pipe) _PIPE(pipe, _PLANE_KEYMAX_1_A, _PLANE_KEYMAX_1_B)
+#define _PLANE_KEYMAX_2(pipe) _PIPE(pipe, _PLANE_KEYMAX_2_A, _PLANE_KEYMAX_2_B)
+#define PLANE_KEYMAX(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_KEYMAX_1(pipe), _PLANE_KEYMAX_2(pipe))
+
+#define _PLANE_BUF_CFG_1_B			0x7127c
+#define _PLANE_BUF_CFG_2_B			0x7137c
+/* skl+: 10 bits, icl+ 11 bits, adlp+ 12 bits */
+#define   PLANE_BUF_END_MASK		REG_GENMASK(27, 16)
+#define   PLANE_BUF_END(end)		REG_FIELD_PREP(PLANE_BUF_END_MASK, (end))
+#define   PLANE_BUF_START_MASK		REG_GENMASK(11, 0)
+#define   PLANE_BUF_START(start)	REG_FIELD_PREP(PLANE_BUF_START_MASK, (start))
+#define _PLANE_BUF_CFG_1(pipe)	\
+	_PIPE(pipe, _PLANE_BUF_CFG_1_A, _PLANE_BUF_CFG_1_B)
+#define _PLANE_BUF_CFG_2(pipe)	\
+	_PIPE(pipe, _PLANE_BUF_CFG_2_A, _PLANE_BUF_CFG_2_B)
+#define PLANE_BUF_CFG(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_BUF_CFG_1(pipe), _PLANE_BUF_CFG_2(pipe))
+
+#define _PLANE_NV12_BUF_CFG_1_B		0x71278
+#define _PLANE_NV12_BUF_CFG_2_B		0x71378
+#define _PLANE_NV12_BUF_CFG_1(pipe)	\
+	_PIPE(pipe, _PLANE_NV12_BUF_CFG_1_A, _PLANE_NV12_BUF_CFG_1_B)
+#define _PLANE_NV12_BUF_CFG_2(pipe)	\
+	_PIPE(pipe, _PLANE_NV12_BUF_CFG_2_A, _PLANE_NV12_BUF_CFG_2_B)
+#define PLANE_NV12_BUF_CFG(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_NV12_BUF_CFG_1(pipe), _PLANE_NV12_BUF_CFG_2(pipe))
+
+#define _PLANE_AUX_DIST_1_B		0x711c0
+#define _PLANE_AUX_DIST_2_B		0x712c0
+#define _PLANE_AUX_DIST_1(pipe) \
+			_PIPE(pipe, _PLANE_AUX_DIST_1_A, _PLANE_AUX_DIST_1_B)
+#define _PLANE_AUX_DIST_2(pipe) \
+			_PIPE(pipe, _PLANE_AUX_DIST_2_A, _PLANE_AUX_DIST_2_B)
+#define PLANE_AUX_DIST(pipe, plane)     \
+	_MMIO_PLANE(plane, _PLANE_AUX_DIST_1(pipe), _PLANE_AUX_DIST_2(pipe))
+
+#define _PLANE_AUX_OFFSET_1_B		0x711c4
+#define _PLANE_AUX_OFFSET_2_B		0x712c4
+#define _PLANE_AUX_OFFSET_1(pipe)       \
+		_PIPE(pipe, _PLANE_AUX_OFFSET_1_A, _PLANE_AUX_OFFSET_1_B)
+#define _PLANE_AUX_OFFSET_2(pipe)       \
+		_PIPE(pipe, _PLANE_AUX_OFFSET_2_A, _PLANE_AUX_OFFSET_2_B)
+#define PLANE_AUX_OFFSET(pipe, plane)   \
+	_MMIO_PLANE(plane, _PLANE_AUX_OFFSET_1(pipe), _PLANE_AUX_OFFSET_2(pipe))
+
+#define _PLANE_CUS_CTL_1_B		0x711c8
+#define _PLANE_CUS_CTL_2_B		0x712c8
+#define _PLANE_CUS_CTL_1(pipe)       \
+		_PIPE(pipe, _PLANE_CUS_CTL_1_A, _PLANE_CUS_CTL_1_B)
+#define _PLANE_CUS_CTL_2(pipe)       \
+		_PIPE(pipe, _PLANE_CUS_CTL_2_A, _PLANE_CUS_CTL_2_B)
+#define PLANE_CUS_CTL(pipe, plane)   \
+	_MMIO_PLANE(plane, _PLANE_CUS_CTL_1(pipe), _PLANE_CUS_CTL_2(pipe))
+
+#define _PLANE_COLOR_CTL_1_B			0x711CC
+#define _PLANE_COLOR_CTL_2_B			0x712CC
+#define _PLANE_COLOR_CTL_3_B			0x713CC
+#define _PLANE_COLOR_CTL_1(pipe)	\
+	_PIPE(pipe, _PLANE_COLOR_CTL_1_A, _PLANE_COLOR_CTL_1_B)
+#define _PLANE_COLOR_CTL_2(pipe)	\
+	_PIPE(pipe, _PLANE_COLOR_CTL_2_A, _PLANE_COLOR_CTL_2_B)
+#define PLANE_COLOR_CTL(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_COLOR_CTL_1(pipe), _PLANE_COLOR_CTL_2(pipe))
+
+#define _SEL_FETCH_PLANE_BASE_1_A		0x70890
+#define _SEL_FETCH_PLANE_BASE_2_A		0x708B0
+#define _SEL_FETCH_PLANE_BASE_3_A		0x708D0
+#define _SEL_FETCH_PLANE_BASE_4_A		0x708F0
+#define _SEL_FETCH_PLANE_BASE_5_A		0x70920
+#define _SEL_FETCH_PLANE_BASE_6_A		0x70940
+#define _SEL_FETCH_PLANE_BASE_7_A		0x70960
+#define _SEL_FETCH_PLANE_BASE_CUR_A		0x70880
+#define _SEL_FETCH_PLANE_BASE_1_B		0x71890
+
+#define _SEL_FETCH_PLANE_BASE_A(plane) _PICK(plane, \
+					     _SEL_FETCH_PLANE_BASE_1_A, \
+					     _SEL_FETCH_PLANE_BASE_2_A, \
+					     _SEL_FETCH_PLANE_BASE_3_A, \
+					     _SEL_FETCH_PLANE_BASE_4_A, \
+					     _SEL_FETCH_PLANE_BASE_5_A, \
+					     _SEL_FETCH_PLANE_BASE_6_A, \
+					     _SEL_FETCH_PLANE_BASE_7_A, \
+					     _SEL_FETCH_PLANE_BASE_CUR_A)
+#define _SEL_FETCH_PLANE_BASE_1(pipe) _PIPE(pipe, _SEL_FETCH_PLANE_BASE_1_A, _SEL_FETCH_PLANE_BASE_1_B)
+#define _SEL_FETCH_PLANE_BASE(pipe, plane) (_SEL_FETCH_PLANE_BASE_1(pipe) - \
+					    _SEL_FETCH_PLANE_BASE_1_A + \
+					    _SEL_FETCH_PLANE_BASE_A(plane))
+
+#define _SEL_FETCH_PLANE_CTL_1_A		0x70890
+#define PLANE_SEL_FETCH_CTL(pipe, plane) _MMIO(_SEL_FETCH_PLANE_BASE(pipe, plane) + \
+					       _SEL_FETCH_PLANE_CTL_1_A - \
+					       _SEL_FETCH_PLANE_BASE_1_A)
+#define PLANE_SEL_FETCH_CTL_ENABLE		REG_BIT(31)
+
+#define _SEL_FETCH_PLANE_POS_1_A		0x70894
+#define PLANE_SEL_FETCH_POS(pipe, plane) _MMIO(_SEL_FETCH_PLANE_BASE(pipe, plane) + \
+					       _SEL_FETCH_PLANE_POS_1_A - \
+					       _SEL_FETCH_PLANE_BASE_1_A)
+
+#define _SEL_FETCH_PLANE_SIZE_1_A		0x70898
+#define PLANE_SEL_FETCH_SIZE(pipe, plane) _MMIO(_SEL_FETCH_PLANE_BASE(pipe, plane) + \
+						_SEL_FETCH_PLANE_SIZE_1_A - \
+						_SEL_FETCH_PLANE_BASE_1_A)
+
+#define _SEL_FETCH_PLANE_OFFSET_1_A		0x7089C
+#define PLANE_SEL_FETCH_OFFSET(pipe, plane) _MMIO(_SEL_FETCH_PLANE_BASE(pipe, plane) + \
+						  _SEL_FETCH_PLANE_OFFSET_1_A - \
+						  _SEL_FETCH_PLANE_BASE_1_A)
+
+/* SKL new cursor registers */
+#define _CUR_BUF_CFG_A				0x7017c
+#define _CUR_BUF_CFG_B				0x7117c
+#define CUR_BUF_CFG(pipe)	_MMIO_PIPE(pipe, _CUR_BUF_CFG_A, _CUR_BUF_CFG_B)
+
+/* VBIOS regs */
+#define VGACNTRL		_MMIO(0x71400)
+# define VGA_DISP_DISABLE			(1 << 31)
+# define VGA_2X_MODE				(1 << 30)
+# define VGA_PIPE_B_SELECT			(1 << 29)
+
+#define VLV_VGACNTRL		_MMIO(VLV_DISPLAY_BASE + 0x71400)
+
+/* Ironlake */
+
+#define CPU_VGACNTRL	_MMIO(0x41000)
+
+#define DIGITAL_PORT_HOTPLUG_CNTRL	_MMIO(0x44030)
+#define  DIGITAL_PORTA_HOTPLUG_ENABLE		(1 << 4)
+#define  DIGITAL_PORTA_PULSE_DURATION_2ms	(0 << 2) /* pre-HSW */
+#define  DIGITAL_PORTA_PULSE_DURATION_4_5ms	(1 << 2) /* pre-HSW */
+#define  DIGITAL_PORTA_PULSE_DURATION_6ms	(2 << 2) /* pre-HSW */
+#define  DIGITAL_PORTA_PULSE_DURATION_100ms	(3 << 2) /* pre-HSW */
+#define  DIGITAL_PORTA_PULSE_DURATION_MASK	(3 << 2) /* pre-HSW */
+#define  DIGITAL_PORTA_HOTPLUG_STATUS_MASK	(3 << 0)
+#define  DIGITAL_PORTA_HOTPLUG_NO_DETECT	(0 << 0)
+#define  DIGITAL_PORTA_HOTPLUG_SHORT_DETECT	(1 << 0)
+#define  DIGITAL_PORTA_HOTPLUG_LONG_DETECT	(2 << 0)
+
+/* refresh rate hardware control */
+#define RR_HW_CTL       _MMIO(0x45300)
+#define  RR_HW_LOW_POWER_FRAMES_MASK    0xff
+#define  RR_HW_HIGH_POWER_FRAMES_MASK   0xff00
+
+#define FDI_PLL_BIOS_0  _MMIO(0x46000)
+#define  FDI_PLL_FB_CLOCK_MASK  0xff
+#define FDI_PLL_BIOS_1  _MMIO(0x46004)
+#define FDI_PLL_BIOS_2  _MMIO(0x46008)
+#define DISPLAY_PORT_PLL_BIOS_0         _MMIO(0x4600c)
+#define DISPLAY_PORT_PLL_BIOS_1         _MMIO(0x46010)
+#define DISPLAY_PORT_PLL_BIOS_2         _MMIO(0x46014)
+
+#define PCH_3DCGDIS0		_MMIO(0x46020)
+# define MARIUNIT_CLOCK_GATE_DISABLE		(1 << 18)
+# define SVSMUNIT_CLOCK_GATE_DISABLE		(1 << 1)
+
+#define PCH_3DCGDIS1		_MMIO(0x46024)
+# define VFMUNIT_CLOCK_GATE_DISABLE		(1 << 11)
+
+#define FDI_PLL_FREQ_CTL        _MMIO(0x46030)
+#define  FDI_PLL_FREQ_CHANGE_REQUEST    (1 << 24)
+#define  FDI_PLL_FREQ_LOCK_LIMIT_MASK   0xfff00
+#define  FDI_PLL_FREQ_DISABLE_COUNT_LIMIT_MASK  0xff
+
+
+#define _PIPEA_DATA_M1		0x60030
+#define _PIPEA_DATA_N1		0x60034
+#define _PIPEA_DATA_M2		0x60038
+#define _PIPEA_DATA_N2		0x6003c
+#define _PIPEA_LINK_M1		0x60040
+#define _PIPEA_LINK_N1		0x60044
+#define _PIPEA_LINK_M2		0x60048
+#define _PIPEA_LINK_N2		0x6004c
+
+/* PIPEB timing regs are same start from 0x61000 */
+
+#define _PIPEB_DATA_M1		0x61030
+#define _PIPEB_DATA_N1		0x61034
+#define _PIPEB_DATA_M2		0x61038
+#define _PIPEB_DATA_N2		0x6103c
+#define _PIPEB_LINK_M1		0x61040
+#define _PIPEB_LINK_N1		0x61044
+#define _PIPEB_LINK_M2		0x61048
+#define _PIPEB_LINK_N2		0x6104c
+
+#define PIPE_DATA_M1(tran) _MMIO_TRANS2(tran, _PIPEA_DATA_M1)
+#define PIPE_DATA_N1(tran) _MMIO_TRANS2(tran, _PIPEA_DATA_N1)
+#define PIPE_DATA_M2(tran) _MMIO_TRANS2(tran, _PIPEA_DATA_M2)
+#define PIPE_DATA_N2(tran) _MMIO_TRANS2(tran, _PIPEA_DATA_N2)
+#define PIPE_LINK_M1(tran) _MMIO_TRANS2(tran, _PIPEA_LINK_M1)
+#define PIPE_LINK_N1(tran) _MMIO_TRANS2(tran, _PIPEA_LINK_N1)
+#define PIPE_LINK_M2(tran) _MMIO_TRANS2(tran, _PIPEA_LINK_M2)
+#define PIPE_LINK_N2(tran) _MMIO_TRANS2(tran, _PIPEA_LINK_N2)
+
+/* CPU panel fitter */
+/* IVB+ has 3 fitters, 0 is 7x5 capable, the other two only 3x3 */
+#define _PFA_CTL_1               0x68080
+#define _PFB_CTL_1               0x68880
+#define  PF_ENABLE              (1 << 31)
+#define  PF_PIPE_SEL_MASK_IVB	(3 << 29)
+#define  PF_PIPE_SEL_IVB(pipe)	((pipe) << 29)
+#define  PF_FILTER_MASK		(3 << 23)
+#define  PF_FILTER_PROGRAMMED	(0 << 23)
+#define  PF_FILTER_MED_3x3	(1 << 23)
+#define  PF_FILTER_EDGE_ENHANCE	(2 << 23)
+#define  PF_FILTER_EDGE_SOFTEN	(3 << 23)
+#define _PFA_WIN_SZ		0x68074
+#define _PFB_WIN_SZ		0x68874
+#define _PFA_WIN_POS		0x68070
+#define _PFB_WIN_POS		0x68870
+#define _PFA_VSCALE		0x68084
+#define _PFB_VSCALE		0x68884
+#define _PFA_HSCALE		0x68090
+#define _PFB_HSCALE		0x68890
+
+#define PF_CTL(pipe)		_MMIO_PIPE(pipe, _PFA_CTL_1, _PFB_CTL_1)
+#define PF_WIN_SZ(pipe)		_MMIO_PIPE(pipe, _PFA_WIN_SZ, _PFB_WIN_SZ)
+#define PF_WIN_POS(pipe)	_MMIO_PIPE(pipe, _PFA_WIN_POS, _PFB_WIN_POS)
+#define PF_VSCALE(pipe)		_MMIO_PIPE(pipe, _PFA_VSCALE, _PFB_VSCALE)
+#define PF_HSCALE(pipe)		_MMIO_PIPE(pipe, _PFA_HSCALE, _PFB_HSCALE)
+
+#define _PSA_CTL		0x68180
+#define _PSB_CTL		0x68980
+#define PS_ENABLE		(1 << 31)
+#define _PSA_WIN_SZ		0x68174
+#define _PSB_WIN_SZ		0x68974
+#define _PSA_WIN_POS		0x68170
+#define _PSB_WIN_POS		0x68970
+
+#define PS_CTL(pipe)		_MMIO_PIPE(pipe, _PSA_CTL, _PSB_CTL)
+#define PS_WIN_SZ(pipe)		_MMIO_PIPE(pipe, _PSA_WIN_SZ, _PSB_WIN_SZ)
+#define PS_WIN_POS(pipe)	_MMIO_PIPE(pipe, _PSA_WIN_POS, _PSB_WIN_POS)
+
+/*
+ * Skylake scalers
+ */
+#define _PS_1A_CTRL      0x68180
+#define _PS_2A_CTRL      0x68280
+#define _PS_1B_CTRL      0x68980
+#define _PS_2B_CTRL      0x68A80
+#define _PS_1C_CTRL      0x69180
+#define PS_SCALER_EN        (1 << 31)
+#define SKL_PS_SCALER_MODE_MASK (3 << 28)
+#define SKL_PS_SCALER_MODE_DYN  (0 << 28)
+#define SKL_PS_SCALER_MODE_HQ  (1 << 28)
+#define SKL_PS_SCALER_MODE_NV12 (2 << 28)
+#define PS_SCALER_MODE_PLANAR (1 << 29)
+#define PS_SCALER_MODE_NORMAL (0 << 29)
+#define PS_PLANE_SEL_MASK  (7 << 25)
+#define PS_PLANE_SEL(plane) (((plane) + 1) << 25)
+#define PS_FILTER_MASK         (3 << 23)
+#define PS_FILTER_MEDIUM       (0 << 23)
+#define PS_FILTER_PROGRAMMED   (1 << 23)
+#define PS_FILTER_EDGE_ENHANCE (2 << 23)
+#define PS_FILTER_BILINEAR     (3 << 23)
+#define PS_VERT3TAP            (1 << 21)
+#define PS_VERT_INT_INVERT_FIELD1 (0 << 20)
+#define PS_VERT_INT_INVERT_FIELD0 (1 << 20)
+#define PS_PWRUP_PROGRESS         (1 << 17)
+#define PS_V_FILTER_BYPASS        (1 << 8)
+#define PS_VADAPT_EN              (1 << 7)
+#define PS_VADAPT_MODE_MASK        (3 << 5)
+#define PS_VADAPT_MODE_LEAST_ADAPT (0 << 5)
+#define PS_VADAPT_MODE_MOD_ADAPT   (1 << 5)
+#define PS_VADAPT_MODE_MOST_ADAPT  (3 << 5)
+#define PS_PLANE_Y_SEL_MASK  (7 << 5)
+#define PS_PLANE_Y_SEL(plane) (((plane) + 1) << 5)
+#define PS_Y_VERT_FILTER_SELECT(set)   ((set) << 4)
+#define PS_Y_HORZ_FILTER_SELECT(set)   ((set) << 3)
+#define PS_UV_VERT_FILTER_SELECT(set)  ((set) << 2)
+#define PS_UV_HORZ_FILTER_SELECT(set)  ((set) << 1)
+
+#define _PS_PWR_GATE_1A     0x68160
+#define _PS_PWR_GATE_2A     0x68260
+#define _PS_PWR_GATE_1B     0x68960
+#define _PS_PWR_GATE_2B     0x68A60
+#define _PS_PWR_GATE_1C     0x69160
+#define PS_PWR_GATE_DIS_OVERRIDE       (1 << 31)
+#define PS_PWR_GATE_SETTLING_TIME_32   (0 << 3)
+#define PS_PWR_GATE_SETTLING_TIME_64   (1 << 3)
+#define PS_PWR_GATE_SETTLING_TIME_96   (2 << 3)
+#define PS_PWR_GATE_SETTLING_TIME_128  (3 << 3)
+#define PS_PWR_GATE_SLPEN_8             0
+#define PS_PWR_GATE_SLPEN_16            1
+#define PS_PWR_GATE_SLPEN_24            2
+#define PS_PWR_GATE_SLPEN_32            3
+
+#define _PS_WIN_POS_1A      0x68170
+#define _PS_WIN_POS_2A      0x68270
+#define _PS_WIN_POS_1B      0x68970
+#define _PS_WIN_POS_2B      0x68A70
+#define _PS_WIN_POS_1C      0x69170
+
+#define _PS_WIN_SZ_1A       0x68174
+#define _PS_WIN_SZ_2A       0x68274
+#define _PS_WIN_SZ_1B       0x68974
+#define _PS_WIN_SZ_2B       0x68A74
+#define _PS_WIN_SZ_1C       0x69174
+
+#define _PS_VSCALE_1A       0x68184
+#define _PS_VSCALE_2A       0x68284
+#define _PS_VSCALE_1B       0x68984
+#define _PS_VSCALE_2B       0x68A84
+#define _PS_VSCALE_1C       0x69184
+
+#define _PS_HSCALE_1A       0x68190
+#define _PS_HSCALE_2A       0x68290
+#define _PS_HSCALE_1B       0x68990
+#define _PS_HSCALE_2B       0x68A90
+#define _PS_HSCALE_1C       0x69190
+
+#define _PS_VPHASE_1A       0x68188
+#define _PS_VPHASE_2A       0x68288
+#define _PS_VPHASE_1B       0x68988
+#define _PS_VPHASE_2B       0x68A88
+#define _PS_VPHASE_1C       0x69188
+#define  PS_Y_PHASE(x)		((x) << 16)
+#define  PS_UV_RGB_PHASE(x)	((x) << 0)
+#define   PS_PHASE_MASK	(0x7fff << 1) /* u2.13 */
+#define   PS_PHASE_TRIP	(1 << 0)
+
+#define _PS_HPHASE_1A       0x68194
+#define _PS_HPHASE_2A       0x68294
+#define _PS_HPHASE_1B       0x68994
+#define _PS_HPHASE_2B       0x68A94
+#define _PS_HPHASE_1C       0x69194
+
+#define _PS_ECC_STAT_1A     0x681D0
+#define _PS_ECC_STAT_2A     0x682D0
+#define _PS_ECC_STAT_1B     0x689D0
+#define _PS_ECC_STAT_2B     0x68AD0
+#define _PS_ECC_STAT_1C     0x691D0
+
+#define _PS_COEF_SET0_INDEX_1A	   0x68198
+#define _PS_COEF_SET0_INDEX_2A	   0x68298
+#define _PS_COEF_SET0_INDEX_1B	   0x68998
+#define _PS_COEF_SET0_INDEX_2B	   0x68A98
+#define PS_COEE_INDEX_AUTO_INC	   (1 << 10)
+
+#define _PS_COEF_SET0_DATA_1A	   0x6819C
+#define _PS_COEF_SET0_DATA_2A	   0x6829C
+#define _PS_COEF_SET0_DATA_1B	   0x6899C
+#define _PS_COEF_SET0_DATA_2B	   0x68A9C
+
+#define _ID(id, a, b) _PICK_EVEN(id, a, b)
+#define SKL_PS_CTRL(pipe, id) _MMIO_PIPE(pipe,        \
+			_ID(id, _PS_1A_CTRL, _PS_2A_CTRL),       \
+			_ID(id, _PS_1B_CTRL, _PS_2B_CTRL))
+#define SKL_PS_PWR_GATE(pipe, id) _MMIO_PIPE(pipe,    \
+			_ID(id, _PS_PWR_GATE_1A, _PS_PWR_GATE_2A), \
+			_ID(id, _PS_PWR_GATE_1B, _PS_PWR_GATE_2B))
+#define SKL_PS_WIN_POS(pipe, id) _MMIO_PIPE(pipe,     \
+			_ID(id, _PS_WIN_POS_1A, _PS_WIN_POS_2A), \
+			_ID(id, _PS_WIN_POS_1B, _PS_WIN_POS_2B))
+#define SKL_PS_WIN_SZ(pipe, id)  _MMIO_PIPE(pipe,     \
+			_ID(id, _PS_WIN_SZ_1A, _PS_WIN_SZ_2A),   \
+			_ID(id, _PS_WIN_SZ_1B, _PS_WIN_SZ_2B))
+#define SKL_PS_VSCALE(pipe, id)  _MMIO_PIPE(pipe,     \
+			_ID(id, _PS_VSCALE_1A, _PS_VSCALE_2A),   \
+			_ID(id, _PS_VSCALE_1B, _PS_VSCALE_2B))
+#define SKL_PS_HSCALE(pipe, id)  _MMIO_PIPE(pipe,     \
+			_ID(id, _PS_HSCALE_1A, _PS_HSCALE_2A),   \
+			_ID(id, _PS_HSCALE_1B, _PS_HSCALE_2B))
+#define SKL_PS_VPHASE(pipe, id)  _MMIO_PIPE(pipe,     \
+			_ID(id, _PS_VPHASE_1A, _PS_VPHASE_2A),   \
+			_ID(id, _PS_VPHASE_1B, _PS_VPHASE_2B))
+#define SKL_PS_HPHASE(pipe, id)  _MMIO_PIPE(pipe,     \
+			_ID(id, _PS_HPHASE_1A, _PS_HPHASE_2A),   \
+			_ID(id, _PS_HPHASE_1B, _PS_HPHASE_2B))
+#define SKL_PS_ECC_STAT(pipe, id)  _MMIO_PIPE(pipe,     \
+			_ID(id, _PS_ECC_STAT_1A, _PS_ECC_STAT_2A),   \
+			_ID(id, _PS_ECC_STAT_1B, _PS_ECC_STAT_2B))
+#define GLK_PS_COEF_INDEX_SET(pipe, id, set)  _MMIO_PIPE(pipe,    \
+			_ID(id, _PS_COEF_SET0_INDEX_1A, _PS_COEF_SET0_INDEX_2A) + (set) * 8, \
+			_ID(id, _PS_COEF_SET0_INDEX_1B, _PS_COEF_SET0_INDEX_2B) + (set) * 8)
+
+#define GLK_PS_COEF_DATA_SET(pipe, id, set)  _MMIO_PIPE(pipe,     \
+			_ID(id, _PS_COEF_SET0_DATA_1A, _PS_COEF_SET0_DATA_2A) + (set) * 8, \
+			_ID(id, _PS_COEF_SET0_DATA_1B, _PS_COEF_SET0_DATA_2B) + (set) * 8)
+/* legacy palette */
+#define _LGC_PALETTE_A           0x4a000
+#define _LGC_PALETTE_B           0x4a800
+#define LGC_PALETTE_RED_MASK     REG_GENMASK(23, 16)
+#define LGC_PALETTE_GREEN_MASK   REG_GENMASK(15, 8)
+#define LGC_PALETTE_BLUE_MASK    REG_GENMASK(7, 0)
+#define LGC_PALETTE(pipe, i) _MMIO(_PIPE(pipe, _LGC_PALETTE_A, _LGC_PALETTE_B) + (i) * 4)
+
+/* ilk/snb precision palette */
+#define _PREC_PALETTE_A           0x4b000
+#define _PREC_PALETTE_B           0x4c000
+#define   PREC_PALETTE_RED_MASK   REG_GENMASK(29, 20)
+#define   PREC_PALETTE_GREEN_MASK REG_GENMASK(19, 10)
+#define   PREC_PALETTE_BLUE_MASK  REG_GENMASK(9, 0)
+#define PREC_PALETTE(pipe, i) _MMIO(_PIPE(pipe, _PREC_PALETTE_A, _PREC_PALETTE_B) + (i) * 4)
+
+#define  _PREC_PIPEAGCMAX              0x4d000
+#define  _PREC_PIPEBGCMAX              0x4d010
+#define PREC_PIPEGCMAX(pipe, i)        _MMIO(_PIPE(pipe, _PIPEAGCMAX, _PIPEBGCMAX) + (i) * 4)
+
+#define _GAMMA_MODE_A		0x4a480
+#define _GAMMA_MODE_B		0x4ac80
+#define GAMMA_MODE(pipe) _MMIO_PIPE(pipe, _GAMMA_MODE_A, _GAMMA_MODE_B)
+#define  PRE_CSC_GAMMA_ENABLE	(1 << 31)
+#define  POST_CSC_GAMMA_ENABLE	(1 << 30)
+#define  GAMMA_MODE_MODE_MASK	(3 << 0)
+#define  GAMMA_MODE_MODE_8BIT	(0 << 0)
+#define  GAMMA_MODE_MODE_10BIT	(1 << 0)
+#define  GAMMA_MODE_MODE_12BIT	(2 << 0)
+#define  GAMMA_MODE_MODE_SPLIT	(3 << 0) /* ivb-bdw */
+#define  GAMMA_MODE_MODE_12BIT_MULTI_SEGMENTED	(3 << 0) /* icl + */
+
+/* Display Internal Timeout Register */
+#define RM_TIMEOUT		_MMIO(0x42060)
+#define  MMIO_TIMEOUT_US(us)	((us) << 0)
+
+/* interrupts */
+#define DE_MASTER_IRQ_CONTROL   (1 << 31)
+#define DE_SPRITEB_FLIP_DONE    (1 << 29)
+#define DE_SPRITEA_FLIP_DONE    (1 << 28)
+#define DE_PLANEB_FLIP_DONE     (1 << 27)
+#define DE_PLANEA_FLIP_DONE     (1 << 26)
+#define DE_PLANE_FLIP_DONE(plane) (1 << (26 + (plane)))
+#define DE_PCU_EVENT            (1 << 25)
+#define DE_GTT_FAULT            (1 << 24)
+#define DE_POISON               (1 << 23)
+#define DE_PERFORM_COUNTER      (1 << 22)
+#define DE_PCH_EVENT            (1 << 21)
+#define DE_AUX_CHANNEL_A        (1 << 20)
+#define DE_DP_A_HOTPLUG         (1 << 19)
+#define DE_GSE                  (1 << 18)
+#define DE_PIPEB_VBLANK         (1 << 15)
+#define DE_PIPEB_EVEN_FIELD     (1 << 14)
+#define DE_PIPEB_ODD_FIELD      (1 << 13)
+#define DE_PIPEB_LINE_COMPARE   (1 << 12)
+#define DE_PIPEB_VSYNC          (1 << 11)
+#define DE_PIPEB_CRC_DONE	(1 << 10)
+#define DE_PIPEB_FIFO_UNDERRUN  (1 << 8)
+#define DE_PIPEA_VBLANK         (1 << 7)
+#define DE_PIPE_VBLANK(pipe)    (1 << (7 + 8 * (pipe)))
+#define DE_PIPEA_EVEN_FIELD     (1 << 6)
+#define DE_PIPEA_ODD_FIELD      (1 << 5)
+#define DE_PIPEA_LINE_COMPARE   (1 << 4)
+#define DE_PIPEA_VSYNC          (1 << 3)
+#define DE_PIPEA_CRC_DONE	(1 << 2)
+#define DE_PIPE_CRC_DONE(pipe)	(1 << (2 + 8 * (pipe)))
+#define DE_PIPEA_FIFO_UNDERRUN  (1 << 0)
+#define DE_PIPE_FIFO_UNDERRUN(pipe)  (1 << (8 * (pipe)))
+
+/* More Ivybridge lolz */
+#define DE_ERR_INT_IVB			(1 << 30)
+#define DE_GSE_IVB			(1 << 29)
+#define DE_PCH_EVENT_IVB		(1 << 28)
+#define DE_DP_A_HOTPLUG_IVB		(1 << 27)
+#define DE_AUX_CHANNEL_A_IVB		(1 << 26)
+#define DE_EDP_PSR_INT_HSW		(1 << 19)
+#define DE_SPRITEC_FLIP_DONE_IVB	(1 << 14)
+#define DE_PLANEC_FLIP_DONE_IVB		(1 << 13)
+#define DE_PIPEC_VBLANK_IVB		(1 << 10)
+#define DE_SPRITEB_FLIP_DONE_IVB	(1 << 9)
+#define DE_PLANEB_FLIP_DONE_IVB		(1 << 8)
+#define DE_PIPEB_VBLANK_IVB		(1 << 5)
+#define DE_SPRITEA_FLIP_DONE_IVB	(1 << 4)
+#define DE_PLANEA_FLIP_DONE_IVB		(1 << 3)
+#define DE_PLANE_FLIP_DONE_IVB(plane)	(1 << (3 + 5 * (plane)))
+#define DE_PIPEA_VBLANK_IVB		(1 << 0)
+#define DE_PIPE_VBLANK_IVB(pipe)	(1 << ((pipe) * 5))
+
+#define VLV_MASTER_IER			_MMIO(0x4400c) /* Gunit master IER */
+#define   MASTER_INTERRUPT_ENABLE	(1 << 31)
+
+#define DEISR   _MMIO(0x44000)
+#define DEIMR   _MMIO(0x44004)
+#define DEIIR   _MMIO(0x44008)
+#define DEIER   _MMIO(0x4400c)
+
+#define GTISR   _MMIO(0x44010)
+#define GTIMR   _MMIO(0x44014)
+#define GTIIR   _MMIO(0x44018)
+#define GTIER   _MMIO(0x4401c)
+
+#define GEN8_MASTER_IRQ			_MMIO(0x44200)
+#define  GEN8_MASTER_IRQ_CONTROL	(1 << 31)
+#define  GEN8_PCU_IRQ			(1 << 30)
+#define  GEN8_DE_PCH_IRQ		(1 << 23)
+#define  GEN8_DE_MISC_IRQ		(1 << 22)
+#define  GEN8_DE_PORT_IRQ		(1 << 20)
+#define  GEN8_DE_PIPE_C_IRQ		(1 << 18)
+#define  GEN8_DE_PIPE_B_IRQ		(1 << 17)
+#define  GEN8_DE_PIPE_A_IRQ		(1 << 16)
+#define  GEN8_DE_PIPE_IRQ(pipe)		(1 << (16 + (pipe)))
+#define  GEN8_GT_VECS_IRQ		(1 << 6)
+#define  GEN8_GT_GUC_IRQ		(1 << 5)
+#define  GEN8_GT_PM_IRQ			(1 << 4)
+#define  GEN8_GT_VCS1_IRQ		(1 << 3) /* NB: VCS2 in bspec! */
+#define  GEN8_GT_VCS0_IRQ		(1 << 2) /* NB: VCS1 in bpsec! */
+#define  GEN8_GT_BCS_IRQ		(1 << 1)
+#define  GEN8_GT_RCS_IRQ		(1 << 0)
+
+#define XELPD_DISPLAY_ERR_FATAL_MASK	_MMIO(0x4421c)
+
+#define GEN8_GT_ISR(which) _MMIO(0x44300 + (0x10 * (which)))
+#define GEN8_GT_IMR(which) _MMIO(0x44304 + (0x10 * (which)))
+#define GEN8_GT_IIR(which) _MMIO(0x44308 + (0x10 * (which)))
+#define GEN8_GT_IER(which) _MMIO(0x4430c + (0x10 * (which)))
+
+#define GEN8_RCS_IRQ_SHIFT 0
+#define GEN8_BCS_IRQ_SHIFT 16
+#define GEN8_VCS0_IRQ_SHIFT 0  /* NB: VCS1 in bspec! */
+#define GEN8_VCS1_IRQ_SHIFT 16 /* NB: VCS2 in bpsec! */
+#define GEN8_VECS_IRQ_SHIFT 0
+#define GEN8_WD_IRQ_SHIFT 16
+
+#define GEN8_DE_PIPE_ISR(pipe) _MMIO(0x44400 + (0x10 * (pipe)))
+#define GEN8_DE_PIPE_IMR(pipe) _MMIO(0x44404 + (0x10 * (pipe)))
+#define GEN8_DE_PIPE_IIR(pipe) _MMIO(0x44408 + (0x10 * (pipe)))
+#define GEN8_DE_PIPE_IER(pipe) _MMIO(0x4440c + (0x10 * (pipe)))
+#define  GEN8_PIPE_FIFO_UNDERRUN	(1 << 31)
+#define  GEN8_PIPE_CDCLK_CRC_ERROR	(1 << 29)
+#define  GEN8_PIPE_CDCLK_CRC_DONE	(1 << 28)
+#define  XELPD_PIPE_SOFT_UNDERRUN	(1 << 22)
+#define  XELPD_PIPE_HARD_UNDERRUN	(1 << 21)
+#define  GEN8_PIPE_CURSOR_FAULT		(1 << 10)
+#define  GEN8_PIPE_SPRITE_FAULT		(1 << 9)
+#define  GEN8_PIPE_PRIMARY_FAULT	(1 << 8)
+#define  GEN8_PIPE_SPRITE_FLIP_DONE	(1 << 5)
+#define  GEN8_PIPE_PRIMARY_FLIP_DONE	(1 << 4)
+#define  GEN8_PIPE_SCAN_LINE_EVENT	(1 << 2)
+#define  GEN8_PIPE_VSYNC		(1 << 1)
+#define  GEN8_PIPE_VBLANK		(1 << 0)
+#define  GEN9_PIPE_CURSOR_FAULT		(1 << 11)
+#define  GEN11_PIPE_PLANE7_FAULT	(1 << 22)
+#define  GEN11_PIPE_PLANE6_FAULT	(1 << 21)
+#define  GEN11_PIPE_PLANE5_FAULT	(1 << 20)
+#define  GEN9_PIPE_PLANE4_FAULT		(1 << 10)
+#define  GEN9_PIPE_PLANE3_FAULT		(1 << 9)
+#define  GEN9_PIPE_PLANE2_FAULT		(1 << 8)
+#define  GEN9_PIPE_PLANE1_FAULT		(1 << 7)
+#define  GEN9_PIPE_PLANE4_FLIP_DONE	(1 << 6)
+#define  GEN9_PIPE_PLANE3_FLIP_DONE	(1 << 5)
+#define  GEN9_PIPE_PLANE2_FLIP_DONE	(1 << 4)
+#define  GEN9_PIPE_PLANE1_FLIP_DONE	(1 << 3)
+#define  GEN9_PIPE_PLANE_FLIP_DONE(p)	(1 << (3 + (p)))
+#define GEN8_DE_PIPE_IRQ_FAULT_ERRORS \
+	(GEN8_PIPE_CURSOR_FAULT | \
+	 GEN8_PIPE_SPRITE_FAULT | \
+	 GEN8_PIPE_PRIMARY_FAULT)
+#define GEN9_DE_PIPE_IRQ_FAULT_ERRORS \
+	(GEN9_PIPE_CURSOR_FAULT | \
+	 GEN9_PIPE_PLANE4_FAULT | \
+	 GEN9_PIPE_PLANE3_FAULT | \
+	 GEN9_PIPE_PLANE2_FAULT | \
+	 GEN9_PIPE_PLANE1_FAULT)
+#define GEN11_DE_PIPE_IRQ_FAULT_ERRORS \
+	(GEN9_DE_PIPE_IRQ_FAULT_ERRORS | \
+	 GEN11_PIPE_PLANE7_FAULT | \
+	 GEN11_PIPE_PLANE6_FAULT | \
+	 GEN11_PIPE_PLANE5_FAULT)
+#define RKL_DE_PIPE_IRQ_FAULT_ERRORS \
+	(GEN9_DE_PIPE_IRQ_FAULT_ERRORS | \
+	 GEN11_PIPE_PLANE5_FAULT)
+
+#define _HPD_PIN_DDI(hpd_pin)	((hpd_pin) - HPD_PORT_A)
+#define _HPD_PIN_TC(hpd_pin)	((hpd_pin) - HPD_PORT_TC1)
+
+#define GEN8_DE_PORT_ISR _MMIO(0x44440)
+#define GEN8_DE_PORT_IMR _MMIO(0x44444)
+#define GEN8_DE_PORT_IIR _MMIO(0x44448)
+#define GEN8_DE_PORT_IER _MMIO(0x4444c)
+#define  DSI1_NON_TE			(1 << 31)
+#define  DSI0_NON_TE			(1 << 30)
+#define  ICL_AUX_CHANNEL_E		(1 << 29)
+#define  ICL_AUX_CHANNEL_F		(1 << 28)
+#define  GEN9_AUX_CHANNEL_D		(1 << 27)
+#define  GEN9_AUX_CHANNEL_C		(1 << 26)
+#define  GEN9_AUX_CHANNEL_B		(1 << 25)
+#define  DSI1_TE			(1 << 24)
+#define  DSI0_TE			(1 << 23)
+#define  GEN8_DE_PORT_HOTPLUG(hpd_pin)	REG_BIT(3 + _HPD_PIN_DDI(hpd_pin))
+#define  BXT_DE_PORT_HOTPLUG_MASK	(GEN8_DE_PORT_HOTPLUG(HPD_PORT_A) | \
+					 GEN8_DE_PORT_HOTPLUG(HPD_PORT_B) | \
+					 GEN8_DE_PORT_HOTPLUG(HPD_PORT_C))
+#define  BDW_DE_PORT_HOTPLUG_MASK	GEN8_DE_PORT_HOTPLUG(HPD_PORT_A)
+#define  BXT_DE_PORT_GMBUS		(1 << 1)
+#define  GEN8_AUX_CHANNEL_A		(1 << 0)
+#define  TGL_DE_PORT_AUX_USBC6		REG_BIT(13)
+#define  XELPD_DE_PORT_AUX_DDIE		REG_BIT(13)
+#define  TGL_DE_PORT_AUX_USBC5		REG_BIT(12)
+#define  XELPD_DE_PORT_AUX_DDID		REG_BIT(12)
+#define  TGL_DE_PORT_AUX_USBC4		REG_BIT(11)
+#define  TGL_DE_PORT_AUX_USBC3		REG_BIT(10)
+#define  TGL_DE_PORT_AUX_USBC2		REG_BIT(9)
+#define  TGL_DE_PORT_AUX_USBC1		REG_BIT(8)
+#define  TGL_DE_PORT_AUX_DDIC		REG_BIT(2)
+#define  TGL_DE_PORT_AUX_DDIB		REG_BIT(1)
+#define  TGL_DE_PORT_AUX_DDIA		REG_BIT(0)
+
+#define GEN8_DE_MISC_ISR _MMIO(0x44460)
+#define GEN8_DE_MISC_IMR _MMIO(0x44464)
+#define GEN8_DE_MISC_IIR _MMIO(0x44468)
+#define GEN8_DE_MISC_IER _MMIO(0x4446c)
+#define  GEN8_DE_MISC_GSE		(1 << 27)
+#define  GEN8_DE_EDP_PSR		(1 << 19)
+
+#define GEN8_PCU_ISR _MMIO(0x444e0)
+#define GEN8_PCU_IMR _MMIO(0x444e4)
+#define GEN8_PCU_IIR _MMIO(0x444e8)
+#define GEN8_PCU_IER _MMIO(0x444ec)
+
+#define GEN11_GU_MISC_ISR	_MMIO(0x444f0)
+#define GEN11_GU_MISC_IMR	_MMIO(0x444f4)
+#define GEN11_GU_MISC_IIR	_MMIO(0x444f8)
+#define GEN11_GU_MISC_IER	_MMIO(0x444fc)
+#define  GEN11_GU_MISC_GSE	(1 << 27)
+
+#define GEN11_GFX_MSTR_IRQ		_MMIO(0x190010)
+#define  GEN11_MASTER_IRQ		(1 << 31)
+#define  GEN11_PCU_IRQ			(1 << 30)
+#define  GEN11_GU_MISC_IRQ		(1 << 29)
+#define  GEN11_DISPLAY_IRQ		(1 << 16)
+#define  GEN11_GT_DW_IRQ(x)		(1 << (x))
+#define  GEN11_GT_DW1_IRQ		(1 << 1)
+#define  GEN11_GT_DW0_IRQ		(1 << 0)
+
+#define DG1_MSTR_TILE_INTR		_MMIO(0x190008)
+#define   DG1_MSTR_IRQ			REG_BIT(31)
+#define   DG1_MSTR_TILE(t)		REG_BIT(t)
+
+#define GEN11_DISPLAY_INT_CTL		_MMIO(0x44200)
+#define  GEN11_DISPLAY_IRQ_ENABLE	(1 << 31)
+#define  GEN11_AUDIO_CODEC_IRQ		(1 << 24)
+#define  GEN11_DE_PCH_IRQ		(1 << 23)
+#define  GEN11_DE_MISC_IRQ		(1 << 22)
+#define  GEN11_DE_HPD_IRQ		(1 << 21)
+#define  GEN11_DE_PORT_IRQ		(1 << 20)
+#define  GEN11_DE_PIPE_C		(1 << 18)
+#define  GEN11_DE_PIPE_B		(1 << 17)
+#define  GEN11_DE_PIPE_A		(1 << 16)
+
+#define GEN11_DE_HPD_ISR		_MMIO(0x44470)
+#define GEN11_DE_HPD_IMR		_MMIO(0x44474)
+#define GEN11_DE_HPD_IIR		_MMIO(0x44478)
+#define GEN11_DE_HPD_IER		_MMIO(0x4447c)
+#define  GEN11_TC_HOTPLUG(hpd_pin)		REG_BIT(16 + _HPD_PIN_TC(hpd_pin))
+#define  GEN11_DE_TC_HOTPLUG_MASK		(GEN11_TC_HOTPLUG(HPD_PORT_TC6) | \
+						 GEN11_TC_HOTPLUG(HPD_PORT_TC5) | \
+						 GEN11_TC_HOTPLUG(HPD_PORT_TC4) | \
+						 GEN11_TC_HOTPLUG(HPD_PORT_TC3) | \
+						 GEN11_TC_HOTPLUG(HPD_PORT_TC2) | \
+						 GEN11_TC_HOTPLUG(HPD_PORT_TC1))
+#define  GEN11_TBT_HOTPLUG(hpd_pin)		REG_BIT(_HPD_PIN_TC(hpd_pin))
+#define  GEN11_DE_TBT_HOTPLUG_MASK		(GEN11_TBT_HOTPLUG(HPD_PORT_TC6) | \
+						 GEN11_TBT_HOTPLUG(HPD_PORT_TC5) | \
+						 GEN11_TBT_HOTPLUG(HPD_PORT_TC4) | \
+						 GEN11_TBT_HOTPLUG(HPD_PORT_TC3) | \
+						 GEN11_TBT_HOTPLUG(HPD_PORT_TC2) | \
+						 GEN11_TBT_HOTPLUG(HPD_PORT_TC1))
+
+#define GEN11_TBT_HOTPLUG_CTL				_MMIO(0x44030)
+#define GEN11_TC_HOTPLUG_CTL				_MMIO(0x44038)
+#define  GEN11_HOTPLUG_CTL_ENABLE(hpd_pin)		(8 << (_HPD_PIN_TC(hpd_pin) * 4))
+#define  GEN11_HOTPLUG_CTL_LONG_DETECT(hpd_pin)		(2 << (_HPD_PIN_TC(hpd_pin) * 4))
+#define  GEN11_HOTPLUG_CTL_SHORT_DETECT(hpd_pin)	(1 << (_HPD_PIN_TC(hpd_pin) * 4))
+#define  GEN11_HOTPLUG_CTL_NO_DETECT(hpd_pin)		(0 << (_HPD_PIN_TC(hpd_pin) * 4))
+
+#define ILK_DISPLAY_CHICKEN2	_MMIO(0x42004)
+/* Required on all Ironlake and Sandybridge according to the B-Spec. */
+#define  ILK_ELPIN_409_SELECT	(1 << 25)
+#define  ILK_DPARB_GATE	(1 << 22)
+#define  ILK_VSDPFD_FULL	(1 << 21)
+#define FUSE_STRAP			_MMIO(0x42014)
+#define  ILK_INTERNAL_GRAPHICS_DISABLE	(1 << 31)
+#define  ILK_INTERNAL_DISPLAY_DISABLE	(1 << 30)
+#define  ILK_DISPLAY_DEBUG_DISABLE	(1 << 29)
+#define  IVB_PIPE_C_DISABLE		(1 << 28)
+#define  ILK_HDCP_DISABLE		(1 << 25)
+#define  ILK_eDP_A_DISABLE		(1 << 24)
+#define  HSW_CDCLK_LIMIT		(1 << 24)
+#define  ILK_DESKTOP			(1 << 23)
+#define  HSW_CPU_SSC_ENABLE		(1 << 21)
+
+#define FUSE_STRAP3			_MMIO(0x42020)
+#define  HSW_REF_CLK_SELECT		(1 << 1)
+
+#define ILK_DSPCLK_GATE_D			_MMIO(0x42020)
+#define   ILK_VRHUNIT_CLOCK_GATE_DISABLE	(1 << 28)
+#define   ILK_DPFCUNIT_CLOCK_GATE_DISABLE	(1 << 9)
+#define   ILK_DPFCRUNIT_CLOCK_GATE_DISABLE	(1 << 8)
+#define   ILK_DPFDUNIT_CLOCK_GATE_ENABLE	(1 << 7)
+#define   ILK_DPARBUNIT_CLOCK_GATE_ENABLE	(1 << 5)
+
+#define IVB_CHICKEN3	_MMIO(0x4200c)
+# define CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE	(1 << 5)
+# define CHICKEN3_DGMG_DONE_FIX_DISABLE		(1 << 2)
+
+#define CHICKEN_PAR1_1			_MMIO(0x42080)
+#define  IGNORE_KVMR_PIPE_A		REG_BIT(23)
+#define  KBL_ARB_FILL_SPARE_22		REG_BIT(22)
+#define  DIS_RAM_BYPASS_PSR2_MAN_TRACK	(1 << 16)
+#define  SKL_DE_COMPRESSED_HASH_MODE	(1 << 15)
+#define  DPA_MASK_VBLANK_SRD		(1 << 15)
+#define  FORCE_ARB_IDLE_PLANES		(1 << 14)
+#define  SKL_EDP_PSR_FIX_RDWRAP		(1 << 3)
+#define  IGNORE_PSR2_HW_TRACKING	(1 << 1)
+
+#define CHICKEN_PAR2_1		_MMIO(0x42090)
+#define  KVM_CONFIG_CHANGE_NOTIFICATION_SELECT	(1 << 14)
+
+#define CHICKEN_MISC_2		_MMIO(0x42084)
+#define  KBL_ARB_FILL_SPARE_14	REG_BIT(14)
+#define  KBL_ARB_FILL_SPARE_13	REG_BIT(13)
+#define  GLK_CL2_PWR_DOWN	(1 << 12)
+#define  GLK_CL1_PWR_DOWN	(1 << 11)
+#define  GLK_CL0_PWR_DOWN	(1 << 10)
+
+#define CHICKEN_MISC_4		_MMIO(0x4208c)
+#define   CHICKEN_FBC_STRIDE_OVERRIDE	REG_BIT(13)
+#define   CHICKEN_FBC_STRIDE_MASK	REG_GENMASK(12, 0)
+#define   CHICKEN_FBC_STRIDE(x)		REG_FIELD_PREP(CHICKEN_FBC_STRIDE_MASK, (x))
+
+#define _CHICKEN_PIPESL_1_A	0x420b0
+#define _CHICKEN_PIPESL_1_B	0x420b4
+#define  HSW_PRI_STRETCH_MAX_MASK	REG_GENMASK(28, 27)
+#define  HSW_PRI_STRETCH_MAX_X8		REG_FIELD_PREP(HSW_PRI_STRETCH_MAX_MASK, 0)
+#define  HSW_PRI_STRETCH_MAX_X4		REG_FIELD_PREP(HSW_PRI_STRETCH_MAX_MASK, 1)
+#define  HSW_PRI_STRETCH_MAX_X2		REG_FIELD_PREP(HSW_PRI_STRETCH_MAX_MASK, 2)
+#define  HSW_PRI_STRETCH_MAX_X1		REG_FIELD_PREP(HSW_PRI_STRETCH_MAX_MASK, 3)
+#define  HSW_SPR_STRETCH_MAX_MASK	REG_GENMASK(26, 25)
+#define  HSW_SPR_STRETCH_MAX_X8		REG_FIELD_PREP(HSW_SPR_STRETCH_MAX_MASK, 0)
+#define  HSW_SPR_STRETCH_MAX_X4		REG_FIELD_PREP(HSW_SPR_STRETCH_MAX_MASK, 1)
+#define  HSW_SPR_STRETCH_MAX_X2		REG_FIELD_PREP(HSW_SPR_STRETCH_MAX_MASK, 2)
+#define  HSW_SPR_STRETCH_MAX_X1		REG_FIELD_PREP(HSW_SPR_STRETCH_MAX_MASK, 3)
+#define  HSW_FBCQ_DIS			(1 << 22)
+#define  BDW_DPRS_MASK_VBLANK_SRD	(1 << 0)
+#define  SKL_PLANE1_STRETCH_MAX_MASK	REG_GENMASK(1, 0)
+#define  SKL_PLANE1_STRETCH_MAX_X8	REG_FIELD_PREP(SKL_PLANE1_STRETCH_MAX_MASK, 0)
+#define  SKL_PLANE1_STRETCH_MAX_X4	REG_FIELD_PREP(SKL_PLANE1_STRETCH_MAX_MASK, 1)
+#define  SKL_PLANE1_STRETCH_MAX_X2	REG_FIELD_PREP(SKL_PLANE1_STRETCH_MAX_MASK, 2)
+#define  SKL_PLANE1_STRETCH_MAX_X1	REG_FIELD_PREP(SKL_PLANE1_STRETCH_MAX_MASK, 3)
+#define CHICKEN_PIPESL_1(pipe) _MMIO_PIPE(pipe, _CHICKEN_PIPESL_1_A, _CHICKEN_PIPESL_1_B)
+
+#define _CHICKEN_TRANS_A	0x420c0
+#define _CHICKEN_TRANS_B	0x420c4
+#define _CHICKEN_TRANS_C	0x420c8
+#define _CHICKEN_TRANS_EDP	0x420cc
+#define _CHICKEN_TRANS_D	0x420d8
+#define CHICKEN_TRANS(trans)	_MMIO(_PICK((trans), \
+					    [TRANSCODER_EDP] = _CHICKEN_TRANS_EDP, \
+					    [TRANSCODER_A] = _CHICKEN_TRANS_A, \
+					    [TRANSCODER_B] = _CHICKEN_TRANS_B, \
+					    [TRANSCODER_C] = _CHICKEN_TRANS_C, \
+					    [TRANSCODER_D] = _CHICKEN_TRANS_D))
+
+#define _MTL_CHICKEN_TRANS_A	0x604e0
+#define _MTL_CHICKEN_TRANS_B	0x614e0
+#define MTL_CHICKEN_TRANS(trans)	_MMIO_TRANS((trans), \
+						    _MTL_CHICKEN_TRANS_A, \
+						    _MTL_CHICKEN_TRANS_B)
+
+#define  HSW_FRAME_START_DELAY_MASK	REG_GENMASK(28, 27)
+#define  HSW_FRAME_START_DELAY(x)	REG_FIELD_PREP(HSW_FRAME_START_DELAY_MASK, x)
+#define  VSC_DATA_SEL_SOFTWARE_CONTROL	REG_BIT(25) /* GLK */
+#define  FECSTALL_DIS_DPTSTREAM_DPTTG	REG_BIT(23)
+#define  DDI_TRAINING_OVERRIDE_ENABLE	REG_BIT(19)
+#define  ADLP_1_BASED_X_GRANULARITY	REG_BIT(18)
+#define  DDI_TRAINING_OVERRIDE_VALUE	REG_BIT(18)
+#define  DDIE_TRAINING_OVERRIDE_ENABLE	REG_BIT(17) /* CHICKEN_TRANS_A only */
+#define  DDIE_TRAINING_OVERRIDE_VALUE	REG_BIT(16) /* CHICKEN_TRANS_A only */
+#define  PSR2_ADD_VERTICAL_LINE_COUNT	REG_BIT(15)
+#define  PSR2_VSC_ENABLE_PROG_HEADER	REG_BIT(12)
+
+#define DISP_ARB_CTL	_MMIO(0x45000)
+#define  DISP_FBC_MEMORY_WAKE		(1 << 31)
+#define  DISP_TILE_SURFACE_SWIZZLING	(1 << 13)
+#define  DISP_FBC_WM_DIS		(1 << 15)
+#define DISP_ARB_CTL2	_MMIO(0x45004)
+#define  DISP_DATA_PARTITION_5_6	(1 << 6)
+#define  DISP_IPC_ENABLE		(1 << 3)
+
+/*
+ * The below are numbered starting from "S1" on gen11/gen12, but starting
+ * with display 13, the bspec switches to a 0-based numbering scheme
+ * (although the addresses stay the same so new S0 = old S1, new S1 = old S2).
+ * We'll just use the 0-based numbering here for all platforms since it's the
+ * way things will be named by the hardware team going forward, plus it's more
+ * consistent with how most of the rest of our registers are named.
+ */
+#define _DBUF_CTL_S0				0x45008
+#define _DBUF_CTL_S1				0x44FE8
+#define _DBUF_CTL_S2				0x44300
+#define _DBUF_CTL_S3				0x44304
+#define DBUF_CTL_S(slice)			_MMIO(_PICK(slice, \
+							    _DBUF_CTL_S0, \
+							    _DBUF_CTL_S1, \
+							    _DBUF_CTL_S2, \
+							    _DBUF_CTL_S3))
+#define  DBUF_POWER_REQUEST			REG_BIT(31)
+#define  DBUF_POWER_STATE			REG_BIT(30)
+#define  DBUF_TRACKER_STATE_SERVICE_MASK	REG_GENMASK(23, 19)
+#define  DBUF_TRACKER_STATE_SERVICE(x)		REG_FIELD_PREP(DBUF_TRACKER_STATE_SERVICE_MASK, x)
+#define  DBUF_MIN_TRACKER_STATE_SERVICE_MASK	REG_GENMASK(18, 16) /* ADL-P+ */
+#define  DBUF_MIN_TRACKER_STATE_SERVICE(x)		REG_FIELD_PREP(DBUF_MIN_TRACKER_STATE_SERVICE_MASK, x) /* ADL-P+ */
+
+#define GEN7_MSG_CTL	_MMIO(0x45010)
+#define  WAIT_FOR_PCH_RESET_ACK		(1 << 1)
+#define  WAIT_FOR_PCH_FLR_ACK		(1 << 0)
+
+#define _BW_BUDDY0_CTL			0x45130
+#define _BW_BUDDY1_CTL			0x45140
+#define BW_BUDDY_CTL(x)			_MMIO(_PICK_EVEN(x, \
+							 _BW_BUDDY0_CTL, \
+							 _BW_BUDDY1_CTL))
+#define   BW_BUDDY_DISABLE		REG_BIT(31)
+#define   BW_BUDDY_TLB_REQ_TIMER_MASK	REG_GENMASK(21, 16)
+#define   BW_BUDDY_TLB_REQ_TIMER(x)	REG_FIELD_PREP(BW_BUDDY_TLB_REQ_TIMER_MASK, x)
+
+#define _BW_BUDDY0_PAGE_MASK		0x45134
+#define _BW_BUDDY1_PAGE_MASK		0x45144
+#define BW_BUDDY_PAGE_MASK(x)		_MMIO(_PICK_EVEN(x, \
+							 _BW_BUDDY0_PAGE_MASK, \
+							 _BW_BUDDY1_PAGE_MASK))
+
+#define HSW_NDE_RSTWRN_OPT	_MMIO(0x46408)
+#define  MTL_RESET_PICA_HANDSHAKE_EN	REG_BIT(6)
+#define  RESET_PCH_HANDSHAKE_ENABLE	REG_BIT(4)
+
+#define GEN8_CHICKEN_DCPR_1			_MMIO(0x46430)
+#define   SKL_SELECT_ALTERNATE_DC_EXIT		REG_BIT(30)
+#define   LATENCY_REPORTING_REMOVED_PIPE_C	REG_BIT(25)
+#define   LATENCY_REPORTING_REMOVED_PIPE_B	REG_BIT(24)
+#define   LATENCY_REPORTING_REMOVED_PIPE_A	REG_BIT(23)
+#define   ICL_DELAY_PMRSP			REG_BIT(22)
+#define   DISABLE_FLR_SRC			REG_BIT(15)
+#define   MASK_WAKEMEM				REG_BIT(13)
+#define   DDI_CLOCK_REG_ACCESS			REG_BIT(7)
+
+#define GEN11_CHICKEN_DCPR_2			_MMIO(0x46434)
+#define   DCPR_MASK_MAXLATENCY_MEMUP_CLR	REG_BIT(27)
+#define   DCPR_MASK_LPMODE			REG_BIT(26)
+#define   DCPR_SEND_RESP_IMM			REG_BIT(25)
+#define   DCPR_CLEAR_MEMSTAT_DIS		REG_BIT(24)
+
+#define SKL_DFSM			_MMIO(0x51000)
+#define   SKL_DFSM_DISPLAY_PM_DISABLE	(1 << 27)
+#define   SKL_DFSM_DISPLAY_HDCP_DISABLE	(1 << 25)
+#define   SKL_DFSM_CDCLK_LIMIT_MASK	(3 << 23)
+#define   SKL_DFSM_CDCLK_LIMIT_675	(0 << 23)
+#define   SKL_DFSM_CDCLK_LIMIT_540	(1 << 23)
+#define   SKL_DFSM_CDCLK_LIMIT_450	(2 << 23)
+#define   SKL_DFSM_CDCLK_LIMIT_337_5	(3 << 23)
+#define   ICL_DFSM_DMC_DISABLE		(1 << 23)
+#define   SKL_DFSM_PIPE_A_DISABLE	(1 << 30)
+#define   SKL_DFSM_PIPE_B_DISABLE	(1 << 21)
+#define   SKL_DFSM_PIPE_C_DISABLE	(1 << 28)
+#define   TGL_DFSM_PIPE_D_DISABLE	(1 << 22)
+#define   GLK_DFSM_DISPLAY_DSC_DISABLE	(1 << 7)
+
+#define SKL_DSSM				_MMIO(0x51004)
+#define ICL_DSSM_CDCLK_PLL_REFCLK_MASK		(7 << 29)
+#define ICL_DSSM_CDCLK_PLL_REFCLK_24MHz		(0 << 29)
+#define ICL_DSSM_CDCLK_PLL_REFCLK_19_2MHz	(1 << 29)
+#define ICL_DSSM_CDCLK_PLL_REFCLK_38_4MHz	(2 << 29)
+
+/*GEN11 chicken */
+#define _PIPEA_CHICKEN				0x70038
+#define _PIPEB_CHICKEN				0x71038
+#define _PIPEC_CHICKEN				0x72038
+#define PIPE_CHICKEN(pipe)			_MMIO_PIPE(pipe, _PIPEA_CHICKEN,\
+							   _PIPEB_CHICKEN)
+#define   UNDERRUN_RECOVERY_DISABLE_ADLP	REG_BIT(30)
+#define   UNDERRUN_RECOVERY_ENABLE_DG2		REG_BIT(30)
+#define   PIXEL_ROUNDING_TRUNC_FB_PASSTHRU	REG_BIT(15)
+#define   DG2_RENDER_CCSTAG_4_3_EN		REG_BIT(12)
+#define   PER_PIXEL_ALPHA_BYPASS_EN		REG_BIT(7)
+
+/* PCH */
+
+#define PCH_DISPLAY_BASE	0xc0000u
+
+/* south display engine interrupt: IBX */
+#define SDE_AUDIO_POWER_D	(1 << 27)
+#define SDE_AUDIO_POWER_C	(1 << 26)
+#define SDE_AUDIO_POWER_B	(1 << 25)
+#define SDE_AUDIO_POWER_SHIFT	(25)
+#define SDE_AUDIO_POWER_MASK	(7 << SDE_AUDIO_POWER_SHIFT)
+#define SDE_GMBUS		(1 << 24)
+#define SDE_AUDIO_HDCP_TRANSB	(1 << 23)
+#define SDE_AUDIO_HDCP_TRANSA	(1 << 22)
+#define SDE_AUDIO_HDCP_MASK	(3 << 22)
+#define SDE_AUDIO_TRANSB	(1 << 21)
+#define SDE_AUDIO_TRANSA	(1 << 20)
+#define SDE_AUDIO_TRANS_MASK	(3 << 20)
+#define SDE_POISON		(1 << 19)
+/* 18 reserved */
+#define SDE_FDI_RXB		(1 << 17)
+#define SDE_FDI_RXA		(1 << 16)
+#define SDE_FDI_MASK		(3 << 16)
+#define SDE_AUXD		(1 << 15)
+#define SDE_AUXC		(1 << 14)
+#define SDE_AUXB		(1 << 13)
+#define SDE_AUX_MASK		(7 << 13)
+/* 12 reserved */
+#define SDE_CRT_HOTPLUG         (1 << 11)
+#define SDE_PORTD_HOTPLUG       (1 << 10)
+#define SDE_PORTC_HOTPLUG       (1 << 9)
+#define SDE_PORTB_HOTPLUG       (1 << 8)
+#define SDE_SDVOB_HOTPLUG       (1 << 6)
+#define SDE_HOTPLUG_MASK        (SDE_CRT_HOTPLUG | \
+				 SDE_SDVOB_HOTPLUG |	\
+				 SDE_PORTB_HOTPLUG |	\
+				 SDE_PORTC_HOTPLUG |	\
+				 SDE_PORTD_HOTPLUG)
+#define SDE_TRANSB_CRC_DONE	(1 << 5)
+#define SDE_TRANSB_CRC_ERR	(1 << 4)
+#define SDE_TRANSB_FIFO_UNDER	(1 << 3)
+#define SDE_TRANSA_CRC_DONE	(1 << 2)
+#define SDE_TRANSA_CRC_ERR	(1 << 1)
+#define SDE_TRANSA_FIFO_UNDER	(1 << 0)
+#define SDE_TRANS_MASK		(0x3f)
+
+/* south display engine interrupt: CPT - CNP */
+#define SDE_AUDIO_POWER_D_CPT	(1 << 31)
+#define SDE_AUDIO_POWER_C_CPT	(1 << 30)
+#define SDE_AUDIO_POWER_B_CPT	(1 << 29)
+#define SDE_AUDIO_POWER_SHIFT_CPT   29
+#define SDE_AUDIO_POWER_MASK_CPT    (7 << 29)
+#define SDE_AUXD_CPT		(1 << 27)
+#define SDE_AUXC_CPT		(1 << 26)
+#define SDE_AUXB_CPT		(1 << 25)
+#define SDE_AUX_MASK_CPT	(7 << 25)
+#define SDE_PORTE_HOTPLUG_SPT	(1 << 25)
+#define SDE_PORTA_HOTPLUG_SPT	(1 << 24)
+#define SDE_PORTD_HOTPLUG_CPT	(1 << 23)
+#define SDE_PORTC_HOTPLUG_CPT	(1 << 22)
+#define SDE_PORTB_HOTPLUG_CPT	(1 << 21)
+#define SDE_CRT_HOTPLUG_CPT	(1 << 19)
+#define SDE_SDVOB_HOTPLUG_CPT	(1 << 18)
+#define SDE_HOTPLUG_MASK_CPT	(SDE_CRT_HOTPLUG_CPT |		\
+				 SDE_SDVOB_HOTPLUG_CPT |	\
+				 SDE_PORTD_HOTPLUG_CPT |	\
+				 SDE_PORTC_HOTPLUG_CPT |	\
+				 SDE_PORTB_HOTPLUG_CPT)
+#define SDE_HOTPLUG_MASK_SPT	(SDE_PORTE_HOTPLUG_SPT |	\
+				 SDE_PORTD_HOTPLUG_CPT |	\
+				 SDE_PORTC_HOTPLUG_CPT |	\
+				 SDE_PORTB_HOTPLUG_CPT |	\
+				 SDE_PORTA_HOTPLUG_SPT)
+#define SDE_GMBUS_CPT		(1 << 17)
+#define SDE_ERROR_CPT		(1 << 16)
+#define SDE_AUDIO_CP_REQ_C_CPT	(1 << 10)
+#define SDE_AUDIO_CP_CHG_C_CPT	(1 << 9)
+#define SDE_FDI_RXC_CPT		(1 << 8)
+#define SDE_AUDIO_CP_REQ_B_CPT	(1 << 6)
+#define SDE_AUDIO_CP_CHG_B_CPT	(1 << 5)
+#define SDE_FDI_RXB_CPT		(1 << 4)
+#define SDE_AUDIO_CP_REQ_A_CPT	(1 << 2)
+#define SDE_AUDIO_CP_CHG_A_CPT	(1 << 1)
+#define SDE_FDI_RXA_CPT		(1 << 0)
+#define SDE_AUDIO_CP_REQ_CPT	(SDE_AUDIO_CP_REQ_C_CPT | \
+				 SDE_AUDIO_CP_REQ_B_CPT | \
+				 SDE_AUDIO_CP_REQ_A_CPT)
+#define SDE_AUDIO_CP_CHG_CPT	(SDE_AUDIO_CP_CHG_C_CPT | \
+				 SDE_AUDIO_CP_CHG_B_CPT | \
+				 SDE_AUDIO_CP_CHG_A_CPT)
+#define SDE_FDI_MASK_CPT	(SDE_FDI_RXC_CPT | \
+				 SDE_FDI_RXB_CPT | \
+				 SDE_FDI_RXA_CPT)
+
+/* south display engine interrupt: ICP/TGP */
+#define SDE_GMBUS_ICP			(1 << 23)
+#define SDE_TC_HOTPLUG_ICP(hpd_pin)	REG_BIT(24 + _HPD_PIN_TC(hpd_pin))
+#define SDE_TC_HOTPLUG_DG2(hpd_pin)	REG_BIT(25 + _HPD_PIN_TC(hpd_pin)) /* sigh */
+#define SDE_DDI_HOTPLUG_ICP(hpd_pin)	REG_BIT(16 + _HPD_PIN_DDI(hpd_pin))
+#define SDE_DDI_HOTPLUG_MASK_ICP	(SDE_DDI_HOTPLUG_ICP(HPD_PORT_D) | \
+					 SDE_DDI_HOTPLUG_ICP(HPD_PORT_C) | \
+					 SDE_DDI_HOTPLUG_ICP(HPD_PORT_B) | \
+					 SDE_DDI_HOTPLUG_ICP(HPD_PORT_A))
+#define SDE_TC_HOTPLUG_MASK_ICP		(SDE_TC_HOTPLUG_ICP(HPD_PORT_TC6) | \
+					 SDE_TC_HOTPLUG_ICP(HPD_PORT_TC5) | \
+					 SDE_TC_HOTPLUG_ICP(HPD_PORT_TC4) | \
+					 SDE_TC_HOTPLUG_ICP(HPD_PORT_TC3) | \
+					 SDE_TC_HOTPLUG_ICP(HPD_PORT_TC2) | \
+					 SDE_TC_HOTPLUG_ICP(HPD_PORT_TC1))
+
+#define SDEISR  _MMIO(0xc4000)
+#define SDEIMR  _MMIO(0xc4004)
+#define SDEIIR  _MMIO(0xc4008)
+#define SDEIER  _MMIO(0xc400c)
+
+#define SERR_INT			_MMIO(0xc4040)
+#define  SERR_INT_POISON		(1 << 31)
+#define  SERR_INT_TRANS_FIFO_UNDERRUN(pipe)	(1 << ((pipe) * 3))
+
+/* digital port hotplug */
+#define PCH_PORT_HOTPLUG		_MMIO(0xc4030)	/* SHOTPLUG_CTL */
+#define  PORTA_HOTPLUG_ENABLE		(1 << 28) /* LPT:LP+ & BXT */
+#define  BXT_DDIA_HPD_INVERT            (1 << 27)
+#define  PORTA_HOTPLUG_STATUS_MASK	(3 << 24) /* SPT+ & BXT */
+#define  PORTA_HOTPLUG_NO_DETECT	(0 << 24) /* SPT+ & BXT */
+#define  PORTA_HOTPLUG_SHORT_DETECT	(1 << 24) /* SPT+ & BXT */
+#define  PORTA_HOTPLUG_LONG_DETECT	(2 << 24) /* SPT+ & BXT */
+#define  PORTD_HOTPLUG_ENABLE		(1 << 20)
+#define  PORTD_PULSE_DURATION_2ms	(0 << 18) /* pre-LPT */
+#define  PORTD_PULSE_DURATION_4_5ms	(1 << 18) /* pre-LPT */
+#define  PORTD_PULSE_DURATION_6ms	(2 << 18) /* pre-LPT */
+#define  PORTD_PULSE_DURATION_100ms	(3 << 18) /* pre-LPT */
+#define  PORTD_PULSE_DURATION_MASK	(3 << 18) /* pre-LPT */
+#define  PORTD_HOTPLUG_STATUS_MASK	(3 << 16)
+#define  PORTD_HOTPLUG_NO_DETECT	(0 << 16)
+#define  PORTD_HOTPLUG_SHORT_DETECT	(1 << 16)
+#define  PORTD_HOTPLUG_LONG_DETECT	(2 << 16)
+#define  PORTC_HOTPLUG_ENABLE		(1 << 12)
+#define  BXT_DDIC_HPD_INVERT            (1 << 11)
+#define  PORTC_PULSE_DURATION_2ms	(0 << 10) /* pre-LPT */
+#define  PORTC_PULSE_DURATION_4_5ms	(1 << 10) /* pre-LPT */
+#define  PORTC_PULSE_DURATION_6ms	(2 << 10) /* pre-LPT */
+#define  PORTC_PULSE_DURATION_100ms	(3 << 10) /* pre-LPT */
+#define  PORTC_PULSE_DURATION_MASK	(3 << 10) /* pre-LPT */
+#define  PORTC_HOTPLUG_STATUS_MASK	(3 << 8)
+#define  PORTC_HOTPLUG_NO_DETECT	(0 << 8)
+#define  PORTC_HOTPLUG_SHORT_DETECT	(1 << 8)
+#define  PORTC_HOTPLUG_LONG_DETECT	(2 << 8)
+#define  PORTB_HOTPLUG_ENABLE		(1 << 4)
+#define  BXT_DDIB_HPD_INVERT            (1 << 3)
+#define  PORTB_PULSE_DURATION_2ms	(0 << 2) /* pre-LPT */
+#define  PORTB_PULSE_DURATION_4_5ms	(1 << 2) /* pre-LPT */
+#define  PORTB_PULSE_DURATION_6ms	(2 << 2) /* pre-LPT */
+#define  PORTB_PULSE_DURATION_100ms	(3 << 2) /* pre-LPT */
+#define  PORTB_PULSE_DURATION_MASK	(3 << 2) /* pre-LPT */
+#define  PORTB_HOTPLUG_STATUS_MASK	(3 << 0)
+#define  PORTB_HOTPLUG_NO_DETECT	(0 << 0)
+#define  PORTB_HOTPLUG_SHORT_DETECT	(1 << 0)
+#define  PORTB_HOTPLUG_LONG_DETECT	(2 << 0)
+#define  BXT_DDI_HPD_INVERT_MASK	(BXT_DDIA_HPD_INVERT | \
+					BXT_DDIB_HPD_INVERT | \
+					BXT_DDIC_HPD_INVERT)
+
+#define PCH_PORT_HOTPLUG2		_MMIO(0xc403C)	/* SHOTPLUG_CTL2 SPT+ */
+#define  PORTE_HOTPLUG_ENABLE		(1 << 4)
+#define  PORTE_HOTPLUG_STATUS_MASK	(3 << 0)
+#define  PORTE_HOTPLUG_NO_DETECT	(0 << 0)
+#define  PORTE_HOTPLUG_SHORT_DETECT	(1 << 0)
+#define  PORTE_HOTPLUG_LONG_DETECT	(2 << 0)
+
+/* This register is a reuse of PCH_PORT_HOTPLUG register. The
+ * functionality covered in PCH_PORT_HOTPLUG is split into
+ * SHOTPLUG_CTL_DDI and SHOTPLUG_CTL_TC.
+ */
+
+#define SHOTPLUG_CTL_DDI				_MMIO(0xc4030)
+#define   SHOTPLUG_CTL_DDI_HPD_ENABLE(hpd_pin)			(0x8 << (_HPD_PIN_DDI(hpd_pin) * 4))
+#define   SHOTPLUG_CTL_DDI_HPD_OUTPUT_DATA(hpd_pin)		(0x4 << (_HPD_PIN_DDI(hpd_pin) * 4))
+#define   SHOTPLUG_CTL_DDI_HPD_STATUS_MASK(hpd_pin)		(0x3 << (_HPD_PIN_DDI(hpd_pin) * 4))
+#define   SHOTPLUG_CTL_DDI_HPD_NO_DETECT(hpd_pin)		(0x0 << (_HPD_PIN_DDI(hpd_pin) * 4))
+#define   SHOTPLUG_CTL_DDI_HPD_SHORT_DETECT(hpd_pin)		(0x1 << (_HPD_PIN_DDI(hpd_pin) * 4))
+#define   SHOTPLUG_CTL_DDI_HPD_LONG_DETECT(hpd_pin)		(0x2 << (_HPD_PIN_DDI(hpd_pin) * 4))
+#define   SHOTPLUG_CTL_DDI_HPD_SHORT_LONG_DETECT(hpd_pin)	(0x3 << (_HPD_PIN_DDI(hpd_pin) * 4))
+
+#define SHOTPLUG_CTL_TC				_MMIO(0xc4034)
+#define   ICP_TC_HPD_ENABLE(hpd_pin)		(8 << (_HPD_PIN_TC(hpd_pin) * 4))
+#define   ICP_TC_HPD_LONG_DETECT(hpd_pin)	(2 << (_HPD_PIN_TC(hpd_pin) * 4))
+#define   ICP_TC_HPD_SHORT_DETECT(hpd_pin)	(1 << (_HPD_PIN_TC(hpd_pin) * 4))
+
+#define SHPD_FILTER_CNT				_MMIO(0xc4038)
+#define   SHPD_FILTER_CNT_500_ADJ		0x001D9
+
+#define _PCH_DPLL_A              0xc6014
+#define _PCH_DPLL_B              0xc6018
+#define PCH_DPLL(pll) _MMIO((pll) == 0 ? _PCH_DPLL_A : _PCH_DPLL_B)
+
+#define _PCH_FPA0                0xc6040
+#define  FP_CB_TUNE		(0x3 << 22)
+#define _PCH_FPA1                0xc6044
+#define _PCH_FPB0                0xc6048
+#define _PCH_FPB1                0xc604c
+#define PCH_FP0(pll) _MMIO((pll) == 0 ? _PCH_FPA0 : _PCH_FPB0)
+#define PCH_FP1(pll) _MMIO((pll) == 0 ? _PCH_FPA1 : _PCH_FPB1)
+
+#define PCH_DPLL_TEST           _MMIO(0xc606c)
+
+#define PCH_DREF_CONTROL        _MMIO(0xC6200)
+#define  DREF_CONTROL_MASK      0x7fc3
+#define  DREF_CPU_SOURCE_OUTPUT_DISABLE         (0 << 13)
+#define  DREF_CPU_SOURCE_OUTPUT_DOWNSPREAD      (2 << 13)
+#define  DREF_CPU_SOURCE_OUTPUT_NONSPREAD       (3 << 13)
+#define  DREF_CPU_SOURCE_OUTPUT_MASK		(3 << 13)
+#define  DREF_SSC_SOURCE_DISABLE                (0 << 11)
+#define  DREF_SSC_SOURCE_ENABLE                 (2 << 11)
+#define  DREF_SSC_SOURCE_MASK			(3 << 11)
+#define  DREF_NONSPREAD_SOURCE_DISABLE          (0 << 9)
+#define  DREF_NONSPREAD_CK505_ENABLE		(1 << 9)
+#define  DREF_NONSPREAD_SOURCE_ENABLE           (2 << 9)
+#define  DREF_NONSPREAD_SOURCE_MASK		(3 << 9)
+#define  DREF_SUPERSPREAD_SOURCE_DISABLE        (0 << 7)
+#define  DREF_SUPERSPREAD_SOURCE_ENABLE         (2 << 7)
+#define  DREF_SUPERSPREAD_SOURCE_MASK		(3 << 7)
+#define  DREF_SSC4_DOWNSPREAD                   (0 << 6)
+#define  DREF_SSC4_CENTERSPREAD                 (1 << 6)
+#define  DREF_SSC1_DISABLE                      (0 << 1)
+#define  DREF_SSC1_ENABLE                       (1 << 1)
+#define  DREF_SSC4_DISABLE                      (0)
+#define  DREF_SSC4_ENABLE                       (1)
+
+#define PCH_RAWCLK_FREQ         _MMIO(0xc6204)
+#define  FDL_TP1_TIMER_SHIFT    12
+#define  FDL_TP1_TIMER_MASK     (3 << 12)
+#define  FDL_TP2_TIMER_SHIFT    10
+#define  FDL_TP2_TIMER_MASK     (3 << 10)
+#define  RAWCLK_FREQ_MASK       0x3ff
+#define  CNP_RAWCLK_DIV_MASK	(0x3ff << 16)
+#define  CNP_RAWCLK_DIV(div)	((div) << 16)
+#define  CNP_RAWCLK_FRAC_MASK	(0xf << 26)
+#define  CNP_RAWCLK_DEN(den)	((den) << 26)
+#define  ICP_RAWCLK_NUM(num)	((num) << 11)
+
+#define PCH_DPLL_TMR_CFG        _MMIO(0xc6208)
+
+#define PCH_SSC4_PARMS          _MMIO(0xc6210)
+#define PCH_SSC4_AUX_PARMS      _MMIO(0xc6214)
+
+#define PCH_DPLL_SEL		_MMIO(0xc7000)
+#define	 TRANS_DPLLB_SEL(pipe)		(1 << ((pipe) * 4))
+#define	 TRANS_DPLLA_SEL(pipe)		0
+#define  TRANS_DPLL_ENABLE(pipe)	(1 << ((pipe) * 4 + 3))
+
+/* transcoder */
+
+#define _PCH_TRANS_HTOTAL_A		0xe0000
+#define  TRANS_HTOTAL_SHIFT		16
+#define  TRANS_HACTIVE_SHIFT		0
+#define _PCH_TRANS_HBLANK_A		0xe0004
+#define  TRANS_HBLANK_END_SHIFT		16
+#define  TRANS_HBLANK_START_SHIFT	0
+#define _PCH_TRANS_HSYNC_A		0xe0008
+#define  TRANS_HSYNC_END_SHIFT		16
+#define  TRANS_HSYNC_START_SHIFT	0
+#define _PCH_TRANS_VTOTAL_A		0xe000c
+#define  TRANS_VTOTAL_SHIFT		16
+#define  TRANS_VACTIVE_SHIFT		0
+#define _PCH_TRANS_VBLANK_A		0xe0010
+#define  TRANS_VBLANK_END_SHIFT		16
+#define  TRANS_VBLANK_START_SHIFT	0
+#define _PCH_TRANS_VSYNC_A		0xe0014
+#define  TRANS_VSYNC_END_SHIFT		16
+#define  TRANS_VSYNC_START_SHIFT	0
+#define _PCH_TRANS_VSYNCSHIFT_A		0xe0028
+
+#define _PCH_TRANSA_DATA_M1	0xe0030
+#define _PCH_TRANSA_DATA_N1	0xe0034
+#define _PCH_TRANSA_DATA_M2	0xe0038
+#define _PCH_TRANSA_DATA_N2	0xe003c
+#define _PCH_TRANSA_LINK_M1	0xe0040
+#define _PCH_TRANSA_LINK_N1	0xe0044
+#define _PCH_TRANSA_LINK_M2	0xe0048
+#define _PCH_TRANSA_LINK_N2	0xe004c
+
+/* Per-transcoder DIP controls (PCH) */
+#define _VIDEO_DIP_CTL_A         0xe0200
+#define _VIDEO_DIP_DATA_A        0xe0208
+#define _VIDEO_DIP_GCP_A         0xe0210
+#define  GCP_COLOR_INDICATION		(1 << 2)
+#define  GCP_DEFAULT_PHASE_ENABLE	(1 << 1)
+#define  GCP_AV_MUTE			(1 << 0)
+
+#define _VIDEO_DIP_CTL_B         0xe1200
+#define _VIDEO_DIP_DATA_B        0xe1208
+#define _VIDEO_DIP_GCP_B         0xe1210
+
+#define TVIDEO_DIP_CTL(pipe) _MMIO_PIPE(pipe, _VIDEO_DIP_CTL_A, _VIDEO_DIP_CTL_B)
+#define TVIDEO_DIP_DATA(pipe) _MMIO_PIPE(pipe, _VIDEO_DIP_DATA_A, _VIDEO_DIP_DATA_B)
+#define TVIDEO_DIP_GCP(pipe) _MMIO_PIPE(pipe, _VIDEO_DIP_GCP_A, _VIDEO_DIP_GCP_B)
+
+/* Per-transcoder DIP controls (VLV) */
+#define _VLV_VIDEO_DIP_CTL_A		(VLV_DISPLAY_BASE + 0x60200)
+#define _VLV_VIDEO_DIP_DATA_A		(VLV_DISPLAY_BASE + 0x60208)
+#define _VLV_VIDEO_DIP_GDCP_PAYLOAD_A	(VLV_DISPLAY_BASE + 0x60210)
+
+#define _VLV_VIDEO_DIP_CTL_B		(VLV_DISPLAY_BASE + 0x61170)
+#define _VLV_VIDEO_DIP_DATA_B		(VLV_DISPLAY_BASE + 0x61174)
+#define _VLV_VIDEO_DIP_GDCP_PAYLOAD_B	(VLV_DISPLAY_BASE + 0x61178)
+
+#define _CHV_VIDEO_DIP_CTL_C		(VLV_DISPLAY_BASE + 0x611f0)
+#define _CHV_VIDEO_DIP_DATA_C		(VLV_DISPLAY_BASE + 0x611f4)
+#define _CHV_VIDEO_DIP_GDCP_PAYLOAD_C	(VLV_DISPLAY_BASE + 0x611f8)
+
+#define VLV_TVIDEO_DIP_CTL(pipe) \
+	_MMIO_PIPE3((pipe), _VLV_VIDEO_DIP_CTL_A, \
+	       _VLV_VIDEO_DIP_CTL_B, _CHV_VIDEO_DIP_CTL_C)
+#define VLV_TVIDEO_DIP_DATA(pipe) \
+	_MMIO_PIPE3((pipe), _VLV_VIDEO_DIP_DATA_A, \
+	       _VLV_VIDEO_DIP_DATA_B, _CHV_VIDEO_DIP_DATA_C)
+#define VLV_TVIDEO_DIP_GCP(pipe) \
+	_MMIO_PIPE3((pipe), _VLV_VIDEO_DIP_GDCP_PAYLOAD_A, \
+		_VLV_VIDEO_DIP_GDCP_PAYLOAD_B, _CHV_VIDEO_DIP_GDCP_PAYLOAD_C)
+
+/* Haswell DIP controls */
+
+#define _HSW_VIDEO_DIP_CTL_A		0x60200
+#define _HSW_VIDEO_DIP_AVI_DATA_A	0x60220
+#define _HSW_VIDEO_DIP_VS_DATA_A	0x60260
+#define _HSW_VIDEO_DIP_SPD_DATA_A	0x602A0
+#define _HSW_VIDEO_DIP_GMP_DATA_A	0x602E0
+#define _HSW_VIDEO_DIP_VSC_DATA_A	0x60320
+#define _GLK_VIDEO_DIP_DRM_DATA_A	0x60440
+#define _HSW_VIDEO_DIP_AVI_ECC_A	0x60240
+#define _HSW_VIDEO_DIP_VS_ECC_A		0x60280
+#define _HSW_VIDEO_DIP_SPD_ECC_A	0x602C0
+#define _HSW_VIDEO_DIP_GMP_ECC_A	0x60300
+#define _HSW_VIDEO_DIP_VSC_ECC_A	0x60344
+#define _HSW_VIDEO_DIP_GCP_A		0x60210
+
+#define _HSW_VIDEO_DIP_CTL_B		0x61200
+#define _HSW_VIDEO_DIP_AVI_DATA_B	0x61220
+#define _HSW_VIDEO_DIP_VS_DATA_B	0x61260
+#define _HSW_VIDEO_DIP_SPD_DATA_B	0x612A0
+#define _HSW_VIDEO_DIP_GMP_DATA_B	0x612E0
+#define _HSW_VIDEO_DIP_VSC_DATA_B	0x61320
+#define _GLK_VIDEO_DIP_DRM_DATA_B	0x61440
+#define _HSW_VIDEO_DIP_BVI_ECC_B	0x61240
+#define _HSW_VIDEO_DIP_VS_ECC_B		0x61280
+#define _HSW_VIDEO_DIP_SPD_ECC_B	0x612C0
+#define _HSW_VIDEO_DIP_GMP_ECC_B	0x61300
+#define _HSW_VIDEO_DIP_VSC_ECC_B	0x61344
+#define _HSW_VIDEO_DIP_GCP_B		0x61210
+
+/* Icelake PPS_DATA and _ECC DIP Registers.
+ * These are available for transcoders B,C and eDP.
+ * Adding the _A so as to reuse the _MMIO_TRANS2
+ * definition, with which it offsets to the right location.
+ */
+
+#define _ICL_VIDEO_DIP_PPS_DATA_A	0x60350
+#define _ICL_VIDEO_DIP_PPS_DATA_B	0x61350
+#define _ICL_VIDEO_DIP_PPS_ECC_A	0x603D4
+#define _ICL_VIDEO_DIP_PPS_ECC_B	0x613D4
+
+#define HSW_TVIDEO_DIP_CTL(trans)		_MMIO_TRANS2(trans, _HSW_VIDEO_DIP_CTL_A)
+#define HSW_TVIDEO_DIP_GCP(trans)		_MMIO_TRANS2(trans, _HSW_VIDEO_DIP_GCP_A)
+#define HSW_TVIDEO_DIP_AVI_DATA(trans, i)	_MMIO_TRANS2(trans, _HSW_VIDEO_DIP_AVI_DATA_A + (i) * 4)
+#define HSW_TVIDEO_DIP_VS_DATA(trans, i)	_MMIO_TRANS2(trans, _HSW_VIDEO_DIP_VS_DATA_A + (i) * 4)
+#define HSW_TVIDEO_DIP_SPD_DATA(trans, i)	_MMIO_TRANS2(trans, _HSW_VIDEO_DIP_SPD_DATA_A + (i) * 4)
+#define HSW_TVIDEO_DIP_GMP_DATA(trans, i)	_MMIO_TRANS2(trans, _HSW_VIDEO_DIP_GMP_DATA_A + (i) * 4)
+#define HSW_TVIDEO_DIP_VSC_DATA(trans, i)	_MMIO_TRANS2(trans, _HSW_VIDEO_DIP_VSC_DATA_A + (i) * 4)
+#define GLK_TVIDEO_DIP_DRM_DATA(trans, i)	_MMIO_TRANS2(trans, _GLK_VIDEO_DIP_DRM_DATA_A + (i) * 4)
+#define ICL_VIDEO_DIP_PPS_DATA(trans, i)	_MMIO_TRANS2(trans, _ICL_VIDEO_DIP_PPS_DATA_A + (i) * 4)
+#define ICL_VIDEO_DIP_PPS_ECC(trans, i)		_MMIO_TRANS2(trans, _ICL_VIDEO_DIP_PPS_ECC_A + (i) * 4)
+
+#define _HSW_STEREO_3D_CTL_A		0x70020
+#define   S3D_ENABLE			(1 << 31)
+#define _HSW_STEREO_3D_CTL_B		0x71020
+
+#define HSW_STEREO_3D_CTL(trans)	_MMIO_PIPE2(trans, _HSW_STEREO_3D_CTL_A)
+
+#define _PCH_TRANS_HTOTAL_B          0xe1000
+#define _PCH_TRANS_HBLANK_B          0xe1004
+#define _PCH_TRANS_HSYNC_B           0xe1008
+#define _PCH_TRANS_VTOTAL_B          0xe100c
+#define _PCH_TRANS_VBLANK_B          0xe1010
+#define _PCH_TRANS_VSYNC_B           0xe1014
+#define _PCH_TRANS_VSYNCSHIFT_B 0xe1028
+
+#define PCH_TRANS_HTOTAL(pipe)		_MMIO_PIPE(pipe, _PCH_TRANS_HTOTAL_A, _PCH_TRANS_HTOTAL_B)
+#define PCH_TRANS_HBLANK(pipe)		_MMIO_PIPE(pipe, _PCH_TRANS_HBLANK_A, _PCH_TRANS_HBLANK_B)
+#define PCH_TRANS_HSYNC(pipe)		_MMIO_PIPE(pipe, _PCH_TRANS_HSYNC_A, _PCH_TRANS_HSYNC_B)
+#define PCH_TRANS_VTOTAL(pipe)		_MMIO_PIPE(pipe, _PCH_TRANS_VTOTAL_A, _PCH_TRANS_VTOTAL_B)
+#define PCH_TRANS_VBLANK(pipe)		_MMIO_PIPE(pipe, _PCH_TRANS_VBLANK_A, _PCH_TRANS_VBLANK_B)
+#define PCH_TRANS_VSYNC(pipe)		_MMIO_PIPE(pipe, _PCH_TRANS_VSYNC_A, _PCH_TRANS_VSYNC_B)
+#define PCH_TRANS_VSYNCSHIFT(pipe)	_MMIO_PIPE(pipe, _PCH_TRANS_VSYNCSHIFT_A, _PCH_TRANS_VSYNCSHIFT_B)
+
+#define _PCH_TRANSB_DATA_M1	0xe1030
+#define _PCH_TRANSB_DATA_N1	0xe1034
+#define _PCH_TRANSB_DATA_M2	0xe1038
+#define _PCH_TRANSB_DATA_N2	0xe103c
+#define _PCH_TRANSB_LINK_M1	0xe1040
+#define _PCH_TRANSB_LINK_N1	0xe1044
+#define _PCH_TRANSB_LINK_M2	0xe1048
+#define _PCH_TRANSB_LINK_N2	0xe104c
+
+#define PCH_TRANS_DATA_M1(pipe)	_MMIO_PIPE(pipe, _PCH_TRANSA_DATA_M1, _PCH_TRANSB_DATA_M1)
+#define PCH_TRANS_DATA_N1(pipe)	_MMIO_PIPE(pipe, _PCH_TRANSA_DATA_N1, _PCH_TRANSB_DATA_N1)
+#define PCH_TRANS_DATA_M2(pipe)	_MMIO_PIPE(pipe, _PCH_TRANSA_DATA_M2, _PCH_TRANSB_DATA_M2)
+#define PCH_TRANS_DATA_N2(pipe)	_MMIO_PIPE(pipe, _PCH_TRANSA_DATA_N2, _PCH_TRANSB_DATA_N2)
+#define PCH_TRANS_LINK_M1(pipe)	_MMIO_PIPE(pipe, _PCH_TRANSA_LINK_M1, _PCH_TRANSB_LINK_M1)
+#define PCH_TRANS_LINK_N1(pipe)	_MMIO_PIPE(pipe, _PCH_TRANSA_LINK_N1, _PCH_TRANSB_LINK_N1)
+#define PCH_TRANS_LINK_M2(pipe)	_MMIO_PIPE(pipe, _PCH_TRANSA_LINK_M2, _PCH_TRANSB_LINK_M2)
+#define PCH_TRANS_LINK_N2(pipe)	_MMIO_PIPE(pipe, _PCH_TRANSA_LINK_N2, _PCH_TRANSB_LINK_N2)
+
+#define _PCH_TRANSACONF              0xf0008
+#define _PCH_TRANSBCONF              0xf1008
+#define PCH_TRANSCONF(pipe)	_MMIO_PIPE(pipe, _PCH_TRANSACONF, _PCH_TRANSBCONF)
+#define LPT_TRANSCONF		PCH_TRANSCONF(PIPE_A) /* lpt has only one transcoder */
+#define  TRANS_ENABLE			REG_BIT(31)
+#define  TRANS_STATE_ENABLE		REG_BIT(30)
+#define  TRANS_FRAME_START_DELAY_MASK	REG_GENMASK(28, 27) /* ibx */
+#define  TRANS_FRAME_START_DELAY(x)	REG_FIELD_PREP(TRANS_FRAME_START_DELAY_MASK, (x)) /* ibx: 0-3 */
+#define  TRANS_INTERLACE_MASK		REG_GENMASK(23, 21)
+#define  TRANS_INTERLACE_PROGRESSIVE	REG_FIELD_PREP(TRANS_INTERLACE_MASK, 0)
+#define  TRANS_INTERLACE_LEGACY_VSYNC_IBX	REG_FIELD_PREP(TRANS_INTERLACE_MASK, 2) /* ibx */
+#define  TRANS_INTERLACE_INTERLACED	REG_FIELD_PREP(TRANS_INTERLACE_MASK, 3)
+#define  TRANS_BPC_MASK			REG_GENMASK(7, 5) /* ibx */
+#define  TRANS_BPC_8			REG_FIELD_PREP(TRANS_BPC_MASK, 0)
+#define  TRANS_BPC_10			REG_FIELD_PREP(TRANS_BPC_MASK, 1)
+#define  TRANS_BPC_6			REG_FIELD_PREP(TRANS_BPC_MASK, 2)
+#define  TRANS_BPC_12			REG_FIELD_PREP(TRANS_BPC_MASK, 3)
+#define _TRANSA_CHICKEN1	 0xf0060
+#define _TRANSB_CHICKEN1	 0xf1060
+#define TRANS_CHICKEN1(pipe)	_MMIO_PIPE(pipe, _TRANSA_CHICKEN1, _TRANSB_CHICKEN1)
+#define  TRANS_CHICKEN1_HDMIUNIT_GC_DISABLE	(1 << 10)
+#define  TRANS_CHICKEN1_DP0UNIT_GC_DISABLE	(1 << 4)
+#define _TRANSA_CHICKEN2	 0xf0064
+#define _TRANSB_CHICKEN2	 0xf1064
+#define TRANS_CHICKEN2(pipe)	_MMIO_PIPE(pipe, _TRANSA_CHICKEN2, _TRANSB_CHICKEN2)
+#define  TRANS_CHICKEN2_TIMING_OVERRIDE			(1 << 31)
+#define  TRANS_CHICKEN2_FDI_POLARITY_REVERSED		(1 << 29)
+#define  TRANS_CHICKEN2_FRAME_START_DELAY_MASK		(3 << 27)
+#define  TRANS_CHICKEN2_FRAME_START_DELAY(x)		((x) << 27) /* 0-3 */
+#define  TRANS_CHICKEN2_DISABLE_DEEP_COLOR_COUNTER	(1 << 26)
+#define  TRANS_CHICKEN2_DISABLE_DEEP_COLOR_MODESWITCH	(1 << 25)
+
+#define SOUTH_CHICKEN1		_MMIO(0xc2000)
+#define  FDIA_PHASE_SYNC_SHIFT_OVR	19
+#define  FDIA_PHASE_SYNC_SHIFT_EN	18
+#define  INVERT_DDID_HPD			(1 << 18)
+#define  INVERT_DDIC_HPD			(1 << 17)
+#define  INVERT_DDIB_HPD			(1 << 16)
+#define  INVERT_DDIA_HPD			(1 << 15)
+#define  FDI_PHASE_SYNC_OVR(pipe) (1 << (FDIA_PHASE_SYNC_SHIFT_OVR - ((pipe) * 2)))
+#define  FDI_PHASE_SYNC_EN(pipe) (1 << (FDIA_PHASE_SYNC_SHIFT_EN - ((pipe) * 2)))
+#define  FDI_BC_BIFURCATION_SELECT	(1 << 12)
+#define  CHASSIS_CLK_REQ_DURATION_MASK	(0xf << 8)
+#define  CHASSIS_CLK_REQ_DURATION(x)	((x) << 8)
+#define  SBCLK_RUN_REFCLK_DIS		(1 << 7)
+#define  SPT_PWM_GRANULARITY		(1 << 0)
+#define SOUTH_CHICKEN2		_MMIO(0xc2004)
+#define  FDI_MPHY_IOSFSB_RESET_STATUS	(1 << 13)
+#define  FDI_MPHY_IOSFSB_RESET_CTL	(1 << 12)
+#define  LPT_PWM_GRANULARITY		(1 << 5)
+#define  DPLS_EDP_PPS_FIX_DIS		(1 << 0)
+
+#define _FDI_RXA_CHICKEN        0xc200c
+#define _FDI_RXB_CHICKEN        0xc2010
+#define  FDI_RX_PHASE_SYNC_POINTER_OVR	(1 << 1)
+#define  FDI_RX_PHASE_SYNC_POINTER_EN	(1 << 0)
+#define FDI_RX_CHICKEN(pipe)	_MMIO_PIPE(pipe, _FDI_RXA_CHICKEN, _FDI_RXB_CHICKEN)
+
+#define SOUTH_DSPCLK_GATE_D	_MMIO(0xc2020)
+#define  PCH_GMBUSUNIT_CLOCK_GATE_DISABLE (1 << 31)
+#define  PCH_DPLUNIT_CLOCK_GATE_DISABLE (1 << 30)
+#define  PCH_DPLSUNIT_CLOCK_GATE_DISABLE (1 << 29)
+#define  PCH_DPMGUNIT_CLOCK_GATE_DISABLE (1 << 15)
+#define  PCH_CPUNIT_CLOCK_GATE_DISABLE (1 << 14)
+#define  CNP_PWM_CGE_GATING_DISABLE (1 << 13)
+#define  PCH_LP_PARTITION_LEVEL_DISABLE  (1 << 12)
+
+/* CPU: FDI_TX */
+#define _FDI_TXA_CTL            0x60100
+#define _FDI_TXB_CTL            0x61100
+#define FDI_TX_CTL(pipe)	_MMIO_PIPE(pipe, _FDI_TXA_CTL, _FDI_TXB_CTL)
+#define  FDI_TX_DISABLE         (0 << 31)
+#define  FDI_TX_ENABLE          (1 << 31)
+#define  FDI_LINK_TRAIN_PATTERN_1       (0 << 28)
+#define  FDI_LINK_TRAIN_PATTERN_2       (1 << 28)
+#define  FDI_LINK_TRAIN_PATTERN_IDLE    (2 << 28)
+#define  FDI_LINK_TRAIN_NONE            (3 << 28)
+#define  FDI_LINK_TRAIN_VOLTAGE_0_4V    (0 << 25)
+#define  FDI_LINK_TRAIN_VOLTAGE_0_6V    (1 << 25)
+#define  FDI_LINK_TRAIN_VOLTAGE_0_8V    (2 << 25)
+#define  FDI_LINK_TRAIN_VOLTAGE_1_2V    (3 << 25)
+#define  FDI_LINK_TRAIN_PRE_EMPHASIS_NONE (0 << 22)
+#define  FDI_LINK_TRAIN_PRE_EMPHASIS_1_5X (1 << 22)
+#define  FDI_LINK_TRAIN_PRE_EMPHASIS_2X   (2 << 22)
+#define  FDI_LINK_TRAIN_PRE_EMPHASIS_3X   (3 << 22)
+/* ILK always use 400mV 0dB for voltage swing and pre-emphasis level.
+   SNB has different settings. */
+/* SNB A-stepping */
+#define  FDI_LINK_TRAIN_400MV_0DB_SNB_A		(0x38 << 22)
+#define  FDI_LINK_TRAIN_400MV_6DB_SNB_A		(0x02 << 22)
+#define  FDI_LINK_TRAIN_600MV_3_5DB_SNB_A	(0x01 << 22)
+#define  FDI_LINK_TRAIN_800MV_0DB_SNB_A		(0x0 << 22)
+/* SNB B-stepping */
+#define  FDI_LINK_TRAIN_400MV_0DB_SNB_B		(0x0 << 22)
+#define  FDI_LINK_TRAIN_400MV_6DB_SNB_B		(0x3a << 22)
+#define  FDI_LINK_TRAIN_600MV_3_5DB_SNB_B	(0x39 << 22)
+#define  FDI_LINK_TRAIN_800MV_0DB_SNB_B		(0x38 << 22)
+#define  FDI_LINK_TRAIN_VOL_EMP_MASK		(0x3f << 22)
+#define  FDI_DP_PORT_WIDTH_SHIFT		19
+#define  FDI_DP_PORT_WIDTH_MASK			(7 << FDI_DP_PORT_WIDTH_SHIFT)
+#define  FDI_DP_PORT_WIDTH(width)           (((width) - 1) << FDI_DP_PORT_WIDTH_SHIFT)
+#define  FDI_TX_ENHANCE_FRAME_ENABLE    (1 << 18)
+/* Ironlake: hardwired to 1 */
+#define  FDI_TX_PLL_ENABLE              (1 << 14)
+
+/* Ivybridge has different bits for lolz */
+#define  FDI_LINK_TRAIN_PATTERN_1_IVB       (0 << 8)
+#define  FDI_LINK_TRAIN_PATTERN_2_IVB       (1 << 8)
+#define  FDI_LINK_TRAIN_PATTERN_IDLE_IVB    (2 << 8)
+#define  FDI_LINK_TRAIN_NONE_IVB            (3 << 8)
+
+/* both Tx and Rx */
+#define  FDI_COMPOSITE_SYNC		(1 << 11)
+#define  FDI_LINK_TRAIN_AUTO		(1 << 10)
+#define  FDI_SCRAMBLING_ENABLE          (0 << 7)
+#define  FDI_SCRAMBLING_DISABLE         (1 << 7)
+
+/* FDI_RX, FDI_X is hard-wired to Transcoder_X */
+#define _FDI_RXA_CTL             0xf000c
+#define _FDI_RXB_CTL             0xf100c
+#define FDI_RX_CTL(pipe)	_MMIO_PIPE(pipe, _FDI_RXA_CTL, _FDI_RXB_CTL)
+#define  FDI_RX_ENABLE          (1 << 31)
+/* train, dp width same as FDI_TX */
+#define  FDI_FS_ERRC_ENABLE		(1 << 27)
+#define  FDI_FE_ERRC_ENABLE		(1 << 26)
+#define  FDI_RX_POLARITY_REVERSED_LPT	(1 << 16)
+#define  FDI_8BPC                       (0 << 16)
+#define  FDI_10BPC                      (1 << 16)
+#define  FDI_6BPC                       (2 << 16)
+#define  FDI_12BPC                      (3 << 16)
+#define  FDI_RX_LINK_REVERSAL_OVERRIDE  (1 << 15)
+#define  FDI_DMI_LINK_REVERSE_MASK      (1 << 14)
+#define  FDI_RX_PLL_ENABLE              (1 << 13)
+#define  FDI_FS_ERR_CORRECT_ENABLE      (1 << 11)
+#define  FDI_FE_ERR_CORRECT_ENABLE      (1 << 10)
+#define  FDI_FS_ERR_REPORT_ENABLE       (1 << 9)
+#define  FDI_FE_ERR_REPORT_ENABLE       (1 << 8)
+#define  FDI_RX_ENHANCE_FRAME_ENABLE    (1 << 6)
+#define  FDI_PCDCLK	                (1 << 4)
+/* CPT */
+#define  FDI_AUTO_TRAINING			(1 << 10)
+#define  FDI_LINK_TRAIN_PATTERN_1_CPT		(0 << 8)
+#define  FDI_LINK_TRAIN_PATTERN_2_CPT		(1 << 8)
+#define  FDI_LINK_TRAIN_PATTERN_IDLE_CPT	(2 << 8)
+#define  FDI_LINK_TRAIN_NORMAL_CPT		(3 << 8)
+#define  FDI_LINK_TRAIN_PATTERN_MASK_CPT	(3 << 8)
+
+#define _FDI_RXA_MISC			0xf0010
+#define _FDI_RXB_MISC			0xf1010
+#define  FDI_RX_PWRDN_LANE1_MASK	(3 << 26)
+#define  FDI_RX_PWRDN_LANE1_VAL(x)	((x) << 26)
+#define  FDI_RX_PWRDN_LANE0_MASK	(3 << 24)
+#define  FDI_RX_PWRDN_LANE0_VAL(x)	((x) << 24)
+#define  FDI_RX_TP1_TO_TP2_48		(2 << 20)
+#define  FDI_RX_TP1_TO_TP2_64		(3 << 20)
+#define  FDI_RX_FDI_DELAY_90		(0x90 << 0)
+#define FDI_RX_MISC(pipe)	_MMIO_PIPE(pipe, _FDI_RXA_MISC, _FDI_RXB_MISC)
+
+#define _FDI_RXA_TUSIZE1        0xf0030
+#define _FDI_RXA_TUSIZE2        0xf0038
+#define _FDI_RXB_TUSIZE1        0xf1030
+#define _FDI_RXB_TUSIZE2        0xf1038
+#define FDI_RX_TUSIZE1(pipe)	_MMIO_PIPE(pipe, _FDI_RXA_TUSIZE1, _FDI_RXB_TUSIZE1)
+#define FDI_RX_TUSIZE2(pipe)	_MMIO_PIPE(pipe, _FDI_RXA_TUSIZE2, _FDI_RXB_TUSIZE2)
+
+/* FDI_RX interrupt register format */
+#define FDI_RX_INTER_LANE_ALIGN         (1 << 10)
+#define FDI_RX_SYMBOL_LOCK              (1 << 9) /* train 2 */
+#define FDI_RX_BIT_LOCK                 (1 << 8) /* train 1 */
+#define FDI_RX_TRAIN_PATTERN_2_FAIL     (1 << 7)
+#define FDI_RX_FS_CODE_ERR              (1 << 6)
+#define FDI_RX_FE_CODE_ERR              (1 << 5)
+#define FDI_RX_SYMBOL_ERR_RATE_ABOVE    (1 << 4)
+#define FDI_RX_HDCP_LINK_FAIL           (1 << 3)
+#define FDI_RX_PIXEL_FIFO_OVERFLOW      (1 << 2)
+#define FDI_RX_CROSS_CLOCK_OVERFLOW     (1 << 1)
+#define FDI_RX_SYMBOL_QUEUE_OVERFLOW    (1 << 0)
+
+#define _FDI_RXA_IIR            0xf0014
+#define _FDI_RXA_IMR            0xf0018
+#define _FDI_RXB_IIR            0xf1014
+#define _FDI_RXB_IMR            0xf1018
+#define FDI_RX_IIR(pipe)	_MMIO_PIPE(pipe, _FDI_RXA_IIR, _FDI_RXB_IIR)
+#define FDI_RX_IMR(pipe)	_MMIO_PIPE(pipe, _FDI_RXA_IMR, _FDI_RXB_IMR)
+
+#define FDI_PLL_CTL_1           _MMIO(0xfe000)
+#define FDI_PLL_CTL_2           _MMIO(0xfe004)
+
+#define PCH_LVDS	_MMIO(0xe1180)
+#define   LVDS_DETECTED	REG_BIT(1)
+
+#define _PCH_DP_B		0xe4100
+#define PCH_DP_B		_MMIO(_PCH_DP_B)
+#define _PCH_DPB_AUX_CH_CTL	0xe4110
+#define _PCH_DPB_AUX_CH_DATA1	0xe4114
+#define _PCH_DPB_AUX_CH_DATA2	0xe4118
+#define _PCH_DPB_AUX_CH_DATA3	0xe411c
+#define _PCH_DPB_AUX_CH_DATA4	0xe4120
+#define _PCH_DPB_AUX_CH_DATA5	0xe4124
+
+#define _PCH_DP_C		0xe4200
+#define PCH_DP_C		_MMIO(_PCH_DP_C)
+#define _PCH_DPC_AUX_CH_CTL	0xe4210
+#define _PCH_DPC_AUX_CH_DATA1	0xe4214
+#define _PCH_DPC_AUX_CH_DATA2	0xe4218
+#define _PCH_DPC_AUX_CH_DATA3	0xe421c
+#define _PCH_DPC_AUX_CH_DATA4	0xe4220
+#define _PCH_DPC_AUX_CH_DATA5	0xe4224
+
+#define _PCH_DP_D		0xe4300
+#define PCH_DP_D		_MMIO(_PCH_DP_D)
+#define _PCH_DPD_AUX_CH_CTL	0xe4310
+#define _PCH_DPD_AUX_CH_DATA1	0xe4314
+#define _PCH_DPD_AUX_CH_DATA2	0xe4318
+#define _PCH_DPD_AUX_CH_DATA3	0xe431c
+#define _PCH_DPD_AUX_CH_DATA4	0xe4320
+#define _PCH_DPD_AUX_CH_DATA5	0xe4324
+
+#define PCH_DP_AUX_CH_CTL(aux_ch)		_MMIO_PORT((aux_ch) - AUX_CH_B, _PCH_DPB_AUX_CH_CTL, _PCH_DPC_AUX_CH_CTL)
+#define PCH_DP_AUX_CH_DATA(aux_ch, i)	_MMIO(_PORT((aux_ch) - AUX_CH_B, _PCH_DPB_AUX_CH_DATA1, _PCH_DPC_AUX_CH_DATA1) + (i) * 4) /* 5 registers */
+
+/* CPT */
+#define _TRANS_DP_CTL_A		0xe0300
+#define _TRANS_DP_CTL_B		0xe1300
+#define _TRANS_DP_CTL_C		0xe2300
+#define TRANS_DP_CTL(pipe)	_MMIO_PIPE(pipe, _TRANS_DP_CTL_A, _TRANS_DP_CTL_B)
+#define  TRANS_DP_OUTPUT_ENABLE		REG_BIT(31)
+#define  TRANS_DP_PORT_SEL_MASK		REG_GENMASK(30, 29)
+#define  TRANS_DP_PORT_SEL_NONE		REG_FIELD_PREP(TRANS_DP_PORT_SEL_MASK, 3)
+#define  TRANS_DP_PORT_SEL(port)	REG_FIELD_PREP(TRANS_DP_PORT_SEL_MASK, (port) - PORT_B)
+#define  TRANS_DP_AUDIO_ONLY		REG_BIT(26)
+#define  TRANS_DP_ENH_FRAMING		REG_BIT(18)
+#define  TRANS_DP_BPC_MASK		REG_GENMASK(10, 9)
+#define  TRANS_DP_BPC_8			REG_FIELD_PREP(TRANS_DP_BPC_MASK, 0)
+#define  TRANS_DP_BPC_10		REG_FIELD_PREP(TRANS_DP_BPC_MASK, 1)
+#define  TRANS_DP_BPC_6			REG_FIELD_PREP(TRANS_DP_BPC_MASK, 2)
+#define  TRANS_DP_BPC_12		REG_FIELD_PREP(TRANS_DP_BPC_MASK, 3)
+#define  TRANS_DP_VSYNC_ACTIVE_HIGH	REG_BIT(4)
+#define  TRANS_DP_HSYNC_ACTIVE_HIGH	REG_BIT(3)
+
+#define _TRANS_DP2_CTL_A			0x600a0
+#define _TRANS_DP2_CTL_B			0x610a0
+#define _TRANS_DP2_CTL_C			0x620a0
+#define _TRANS_DP2_CTL_D			0x630a0
+#define TRANS_DP2_CTL(trans)			_MMIO_TRANS(trans, _TRANS_DP2_CTL_A, _TRANS_DP2_CTL_B)
+#define  TRANS_DP2_128B132B_CHANNEL_CODING	REG_BIT(31)
+#define  TRANS_DP2_PANEL_REPLAY_ENABLE		REG_BIT(30)
+#define  TRANS_DP2_DEBUG_ENABLE			REG_BIT(23)
+
+#define _TRANS_DP2_VFREQHIGH_A			0x600a4
+#define _TRANS_DP2_VFREQHIGH_B			0x610a4
+#define _TRANS_DP2_VFREQHIGH_C			0x620a4
+#define _TRANS_DP2_VFREQHIGH_D			0x630a4
+#define TRANS_DP2_VFREQHIGH(trans)		_MMIO_TRANS(trans, _TRANS_DP2_VFREQHIGH_A, _TRANS_DP2_VFREQHIGH_B)
+#define  TRANS_DP2_VFREQ_PIXEL_CLOCK_MASK	REG_GENMASK(31, 8)
+#define  TRANS_DP2_VFREQ_PIXEL_CLOCK(clk_hz)	REG_FIELD_PREP(TRANS_DP2_VFREQ_PIXEL_CLOCK_MASK, (clk_hz))
+
+#define _TRANS_DP2_VFREQLOW_A			0x600a8
+#define _TRANS_DP2_VFREQLOW_B			0x610a8
+#define _TRANS_DP2_VFREQLOW_C			0x620a8
+#define _TRANS_DP2_VFREQLOW_D			0x630a8
+#define TRANS_DP2_VFREQLOW(trans)		_MMIO_TRANS(trans, _TRANS_DP2_VFREQLOW_A, _TRANS_DP2_VFREQLOW_B)
+
+/* SNB eDP training params */
+/* SNB A-stepping */
+#define  EDP_LINK_TRAIN_400MV_0DB_SNB_A		(0x38 << 22)
+#define  EDP_LINK_TRAIN_400MV_6DB_SNB_A		(0x02 << 22)
+#define  EDP_LINK_TRAIN_600MV_3_5DB_SNB_A	(0x01 << 22)
+#define  EDP_LINK_TRAIN_800MV_0DB_SNB_A		(0x0 << 22)
+/* SNB B-stepping */
+#define  EDP_LINK_TRAIN_400_600MV_0DB_SNB_B	(0x0 << 22)
+#define  EDP_LINK_TRAIN_400MV_3_5DB_SNB_B	(0x1 << 22)
+#define  EDP_LINK_TRAIN_400_600MV_6DB_SNB_B	(0x3a << 22)
+#define  EDP_LINK_TRAIN_600_800MV_3_5DB_SNB_B	(0x39 << 22)
+#define  EDP_LINK_TRAIN_800_1200MV_0DB_SNB_B	(0x38 << 22)
+#define  EDP_LINK_TRAIN_VOL_EMP_MASK_SNB	(0x3f << 22)
+
+/* IVB */
+#define EDP_LINK_TRAIN_400MV_0DB_IVB		(0x24 << 22)
+#define EDP_LINK_TRAIN_400MV_3_5DB_IVB		(0x2a << 22)
+#define EDP_LINK_TRAIN_400MV_6DB_IVB		(0x2f << 22)
+#define EDP_LINK_TRAIN_600MV_0DB_IVB		(0x30 << 22)
+#define EDP_LINK_TRAIN_600MV_3_5DB_IVB		(0x36 << 22)
+#define EDP_LINK_TRAIN_800MV_0DB_IVB		(0x38 << 22)
+#define EDP_LINK_TRAIN_800MV_3_5DB_IVB		(0x3e << 22)
+
+/* legacy values */
+#define EDP_LINK_TRAIN_500MV_0DB_IVB		(0x00 << 22)
+#define EDP_LINK_TRAIN_1000MV_0DB_IVB		(0x20 << 22)
+#define EDP_LINK_TRAIN_500MV_3_5DB_IVB		(0x02 << 22)
+#define EDP_LINK_TRAIN_1000MV_3_5DB_IVB		(0x22 << 22)
+#define EDP_LINK_TRAIN_1000MV_6DB_IVB		(0x23 << 22)
+
+#define  EDP_LINK_TRAIN_VOL_EMP_MASK_IVB	(0x3f << 22)
+
+#define  VLV_PMWGICZ				_MMIO(0x1300a4)
+
+#define  HSW_EDRAM_CAP				_MMIO(0x120010)
+#define    EDRAM_ENABLED			0x1
+#define    EDRAM_NUM_BANKS(cap)			(((cap) >> 1) & 0xf)
+#define    EDRAM_WAYS_IDX(cap)			(((cap) >> 5) & 0x7)
+#define    EDRAM_SETS_IDX(cap)			(((cap) >> 8) & 0x3)
+
+#define VLV_CHICKEN_3				_MMIO(VLV_DISPLAY_BASE + 0x7040C)
+#define  PIXEL_OVERLAP_CNT_MASK			(3 << 30)
+#define  PIXEL_OVERLAP_CNT_SHIFT		30
+
+#define GEN6_PCODE_MAILBOX			_MMIO(0x138124)
+#define   GEN6_PCODE_READY			(1 << 31)
+#define   GEN6_PCODE_MB_PARAM2			REG_GENMASK(23, 16)
+#define   GEN6_PCODE_MB_PARAM1			REG_GENMASK(15, 8)
+#define   GEN6_PCODE_MB_COMMAND			REG_GENMASK(7, 0)
+#define   GEN6_PCODE_ERROR_MASK			0xFF
+#define     GEN6_PCODE_SUCCESS			0x0
+#define     GEN6_PCODE_ILLEGAL_CMD		0x1
+#define     GEN6_PCODE_MIN_FREQ_TABLE_GT_RATIO_OUT_OF_RANGE 0x2
+#define     GEN6_PCODE_TIMEOUT			0x3
+#define     GEN6_PCODE_UNIMPLEMENTED_CMD	0xFF
+#define     GEN7_PCODE_TIMEOUT			0x2
+#define     GEN7_PCODE_ILLEGAL_DATA		0x3
+#define     GEN11_PCODE_ILLEGAL_SUBCOMMAND	0x4
+#define     GEN11_PCODE_LOCKED			0x6
+#define     GEN11_PCODE_REJECTED		0x11
+#define     GEN7_PCODE_MIN_FREQ_TABLE_GT_RATIO_OUT_OF_RANGE 0x10
+#define   GEN6_PCODE_WRITE_RC6VIDS		0x4
+#define   GEN6_PCODE_READ_RC6VIDS		0x5
+#define     GEN6_ENCODE_RC6_VID(mv)		(((mv) - 245) / 5)
+#define     GEN6_DECODE_RC6_VID(vids)		(((vids) * 5) + 245)
+#define   BDW_PCODE_DISPLAY_FREQ_CHANGE_REQ	0x18
+#define   GEN9_PCODE_READ_MEM_LATENCY		0x6
+#define     GEN9_MEM_LATENCY_LEVEL_3_7_MASK	REG_GENMASK(31, 24)
+#define     GEN9_MEM_LATENCY_LEVEL_2_6_MASK	REG_GENMASK(23, 16)
+#define     GEN9_MEM_LATENCY_LEVEL_1_5_MASK	REG_GENMASK(15, 8)
+#define     GEN9_MEM_LATENCY_LEVEL_0_4_MASK	REG_GENMASK(7, 0)
+#define   SKL_PCODE_LOAD_HDCP_KEYS		0x5
+#define   SKL_PCODE_CDCLK_CONTROL		0x7
+#define     SKL_CDCLK_PREPARE_FOR_CHANGE	0x3
+#define     SKL_CDCLK_READY_FOR_CHANGE		0x1
+#define   GEN6_PCODE_WRITE_MIN_FREQ_TABLE	0x8
+#define   GEN6_PCODE_READ_MIN_FREQ_TABLE	0x9
+#define   GEN6_READ_OC_PARAMS			0xc
+#define   ICL_PCODE_MEM_SUBSYSYSTEM_INFO	0xd
+#define     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO	(0x0 << 8)
+#define     ICL_PCODE_MEM_SS_READ_QGV_POINT_INFO(point)	(((point) << 16) | (0x1 << 8))
+#define     ADL_PCODE_MEM_SS_READ_PSF_GV_INFO	((0) | (0x2 << 8))
+#define   ICL_PCODE_SAGV_DE_MEM_SS_CONFIG	0xe
+#define     ICL_PCODE_REP_QGV_MASK		REG_GENMASK(1, 0)
+#define     ICL_PCODE_REP_QGV_SAFE		REG_FIELD_PREP(ICL_PCODE_REP_QGV_MASK, 0)
+#define     ICL_PCODE_REP_QGV_POLL		REG_FIELD_PREP(ICL_PCODE_REP_QGV_MASK, 1)
+#define     ICL_PCODE_REP_QGV_REJECTED		REG_FIELD_PREP(ICL_PCODE_REP_QGV_MASK, 2)
+#define     ADLS_PCODE_REP_PSF_MASK		REG_GENMASK(3, 2)
+#define     ADLS_PCODE_REP_PSF_SAFE		REG_FIELD_PREP(ADLS_PCODE_REP_PSF_MASK, 0)
+#define     ADLS_PCODE_REP_PSF_POLL		REG_FIELD_PREP(ADLS_PCODE_REP_PSF_MASK, 1)
+#define     ADLS_PCODE_REP_PSF_REJECTED		REG_FIELD_PREP(ADLS_PCODE_REP_PSF_MASK, 2)
+#define     ICL_PCODE_REQ_QGV_PT_MASK		REG_GENMASK(7, 0)
+#define     ICL_PCODE_REQ_QGV_PT(x)		REG_FIELD_PREP(ICL_PCODE_REQ_QGV_PT_MASK, (x))
+#define     ADLS_PCODE_REQ_PSF_PT_MASK		REG_GENMASK(10, 8)
+#define     ADLS_PCODE_REQ_PSF_PT(x)		REG_FIELD_PREP(ADLS_PCODE_REQ_PSF_PT_MASK, (x))
+#define   GEN6_PCODE_READ_D_COMP		0x10
+#define   GEN6_PCODE_WRITE_D_COMP		0x11
+#define   ICL_PCODE_EXIT_TCCOLD			0x12
+#define   HSW_PCODE_DE_WRITE_FREQ_REQ		0x17
+#define   DISPLAY_IPS_CONTROL			0x19
+#define   TGL_PCODE_TCCOLD			0x26
+#define     TGL_PCODE_EXIT_TCCOLD_DATA_L_EXIT_FAILED	REG_BIT(0)
+#define     TGL_PCODE_EXIT_TCCOLD_DATA_L_BLOCK_REQ	0
+#define     TGL_PCODE_EXIT_TCCOLD_DATA_L_UNBLOCK_REQ	REG_BIT(0)
+            /* See also IPS_CTL */
+#define     IPS_PCODE_CONTROL			(1 << 30)
+#define   HSW_PCODE_DYNAMIC_DUTY_CYCLE_CONTROL	0x1A
+#define   GEN9_PCODE_SAGV_CONTROL		0x21
+#define     GEN9_SAGV_DISABLE			0x0
+#define     GEN9_SAGV_IS_DISABLED		0x1
+#define     GEN9_SAGV_ENABLE			0x3
+#define   DG1_PCODE_STATUS			0x7E
+#define     DG1_UNCORE_GET_INIT_STATUS		0x0
+#define     DG1_UNCORE_INIT_STATUS_COMPLETE	0x1
+#define GEN12_PCODE_READ_SAGV_BLOCK_TIME_US	0x23
+#define   XEHP_PCODE_FREQUENCY_CONFIG		0x6e	/* xehpsdv, pvc */
+/* XEHP_PCODE_FREQUENCY_CONFIG sub-commands (param1) */
+#define     PCODE_MBOX_FC_SC_READ_FUSED_P0	0x0
+#define     PCODE_MBOX_FC_SC_READ_FUSED_PN	0x1
+/* PCODE_MBOX_DOMAIN_* - mailbox domain IDs */
+/*   XEHP_PCODE_FREQUENCY_CONFIG param2 */
+#define     PCODE_MBOX_DOMAIN_NONE		0x0
+#define     PCODE_MBOX_DOMAIN_MEDIAFF		0x3
+#define GEN6_PCODE_DATA				_MMIO(0x138128)
+#define   GEN6_PCODE_FREQ_IA_RATIO_SHIFT	8
+#define   GEN6_PCODE_FREQ_RING_RATIO_SHIFT	16
+#define GEN6_PCODE_DATA1			_MMIO(0x13812C)
+
+/* IVYBRIDGE DPF */
+#define GEN7_L3CDERRST1(slice)		_MMIO(0xB008 + (slice) * 0x200) /* L3CD Error Status 1 */
+#define   GEN7_L3CDERRST1_ROW_MASK	(0x7ff << 14)
+#define   GEN7_PARITY_ERROR_VALID	(1 << 13)
+#define   GEN7_L3CDERRST1_BANK_MASK	(3 << 11)
+#define   GEN7_L3CDERRST1_SUBBANK_MASK	(7 << 8)
+#define GEN7_PARITY_ERROR_ROW(reg) \
+		(((reg) & GEN7_L3CDERRST1_ROW_MASK) >> 14)
+#define GEN7_PARITY_ERROR_BANK(reg) \
+		(((reg) & GEN7_L3CDERRST1_BANK_MASK) >> 11)
+#define GEN7_PARITY_ERROR_SUBBANK(reg) \
+		(((reg) & GEN7_L3CDERRST1_SUBBANK_MASK) >> 8)
+#define   GEN7_L3CDERRST1_ENABLE	(1 << 7)
+
+/* These are the 4 32-bit write offset registers for each stream
+ * output buffer.  It determines the offset from the
+ * 3DSTATE_SO_BUFFERs that the next streamed vertex output goes to.
+ */
+#define GEN7_SO_WRITE_OFFSET(n)		_MMIO(0x5280 + (n) * 4)
+
+/*
+ * HSW - ICL power wells
+ *
+ * Platforms have up to 3 power well control register sets, each set
+ * controlling up to 16 power wells via a request/status HW flag tuple:
+ * - main (HSW_PWR_WELL_CTL[1-4])
+ * - AUX  (ICL_PWR_WELL_CTL_AUX[1-4])
+ * - DDI  (ICL_PWR_WELL_CTL_DDI[1-4])
+ * Each control register set consists of up to 4 registers used by different
+ * sources that can request a power well to be enabled:
+ * - BIOS   (HSW_PWR_WELL_CTL1/ICL_PWR_WELL_CTL_AUX1/ICL_PWR_WELL_CTL_DDI1)
+ * - DRIVER (HSW_PWR_WELL_CTL2/ICL_PWR_WELL_CTL_AUX2/ICL_PWR_WELL_CTL_DDI2)
+ * - KVMR   (HSW_PWR_WELL_CTL3)   (only in the main register set)
+ * - DEBUG  (HSW_PWR_WELL_CTL4/ICL_PWR_WELL_CTL_AUX4/ICL_PWR_WELL_CTL_DDI4)
+ */
+#define HSW_PWR_WELL_CTL1			_MMIO(0x45400)
+#define HSW_PWR_WELL_CTL2			_MMIO(0x45404)
+#define HSW_PWR_WELL_CTL3			_MMIO(0x45408)
+#define HSW_PWR_WELL_CTL4			_MMIO(0x4540C)
+#define   HSW_PWR_WELL_CTL_REQ(pw_idx)		(0x2 << ((pw_idx) * 2))
+#define   HSW_PWR_WELL_CTL_STATE(pw_idx)	(0x1 << ((pw_idx) * 2))
+
+/* HSW/BDW power well */
+#define   HSW_PW_CTL_IDX_GLOBAL			15
+
+/* SKL/BXT/GLK power wells */
+#define   SKL_PW_CTL_IDX_PW_2			15
+#define   SKL_PW_CTL_IDX_PW_1			14
+#define   GLK_PW_CTL_IDX_AUX_C			10
+#define   GLK_PW_CTL_IDX_AUX_B			9
+#define   GLK_PW_CTL_IDX_AUX_A			8
+#define   SKL_PW_CTL_IDX_DDI_D			4
+#define   SKL_PW_CTL_IDX_DDI_C			3
+#define   SKL_PW_CTL_IDX_DDI_B			2
+#define   SKL_PW_CTL_IDX_DDI_A_E		1
+#define   GLK_PW_CTL_IDX_DDI_A			1
+#define   SKL_PW_CTL_IDX_MISC_IO		0
+
+/* ICL/TGL - power wells */
+#define   TGL_PW_CTL_IDX_PW_5			4
+#define   ICL_PW_CTL_IDX_PW_4			3
+#define   ICL_PW_CTL_IDX_PW_3			2
+#define   ICL_PW_CTL_IDX_PW_2			1
+#define   ICL_PW_CTL_IDX_PW_1			0
+
+/* XE_LPD - power wells */
+#define   XELPD_PW_CTL_IDX_PW_D			8
+#define   XELPD_PW_CTL_IDX_PW_C			7
+#define   XELPD_PW_CTL_IDX_PW_B			6
+#define   XELPD_PW_CTL_IDX_PW_A			5
+
+#define ICL_PWR_WELL_CTL_AUX1			_MMIO(0x45440)
+#define ICL_PWR_WELL_CTL_AUX2			_MMIO(0x45444)
+#define ICL_PWR_WELL_CTL_AUX4			_MMIO(0x4544C)
+#define   TGL_PW_CTL_IDX_AUX_TBT6		14
+#define   TGL_PW_CTL_IDX_AUX_TBT5		13
+#define   TGL_PW_CTL_IDX_AUX_TBT4		12
+#define   ICL_PW_CTL_IDX_AUX_TBT4		11
+#define   TGL_PW_CTL_IDX_AUX_TBT3		11
+#define   ICL_PW_CTL_IDX_AUX_TBT3		10
+#define   TGL_PW_CTL_IDX_AUX_TBT2		10
+#define   ICL_PW_CTL_IDX_AUX_TBT2		9
+#define   TGL_PW_CTL_IDX_AUX_TBT1		9
+#define   ICL_PW_CTL_IDX_AUX_TBT1		8
+#define   TGL_PW_CTL_IDX_AUX_TC6		8
+#define   XELPD_PW_CTL_IDX_AUX_E			8
+#define   TGL_PW_CTL_IDX_AUX_TC5		7
+#define   XELPD_PW_CTL_IDX_AUX_D			7
+#define   TGL_PW_CTL_IDX_AUX_TC4		6
+#define   ICL_PW_CTL_IDX_AUX_F			5
+#define   TGL_PW_CTL_IDX_AUX_TC3		5
+#define   ICL_PW_CTL_IDX_AUX_E			4
+#define   TGL_PW_CTL_IDX_AUX_TC2		4
+#define   ICL_PW_CTL_IDX_AUX_D			3
+#define   TGL_PW_CTL_IDX_AUX_TC1		3
+#define   ICL_PW_CTL_IDX_AUX_C			2
+#define   ICL_PW_CTL_IDX_AUX_B			1
+#define   ICL_PW_CTL_IDX_AUX_A			0
+
+#define ICL_PWR_WELL_CTL_DDI1			_MMIO(0x45450)
+#define ICL_PWR_WELL_CTL_DDI2			_MMIO(0x45454)
+#define ICL_PWR_WELL_CTL_DDI4			_MMIO(0x4545C)
+#define   XELPD_PW_CTL_IDX_DDI_E			8
+#define   TGL_PW_CTL_IDX_DDI_TC6		8
+#define   XELPD_PW_CTL_IDX_DDI_D			7
+#define   TGL_PW_CTL_IDX_DDI_TC5		7
+#define   TGL_PW_CTL_IDX_DDI_TC4		6
+#define   ICL_PW_CTL_IDX_DDI_F			5
+#define   TGL_PW_CTL_IDX_DDI_TC3		5
+#define   ICL_PW_CTL_IDX_DDI_E			4
+#define   TGL_PW_CTL_IDX_DDI_TC2		4
+#define   ICL_PW_CTL_IDX_DDI_D			3
+#define   TGL_PW_CTL_IDX_DDI_TC1		3
+#define   ICL_PW_CTL_IDX_DDI_C			2
+#define   ICL_PW_CTL_IDX_DDI_B			1
+#define   ICL_PW_CTL_IDX_DDI_A			0
+
+/* HSW - power well misc debug registers */
+#define HSW_PWR_WELL_CTL5			_MMIO(0x45410)
+#define   HSW_PWR_WELL_ENABLE_SINGLE_STEP	(1 << 31)
+#define   HSW_PWR_WELL_PWR_GATE_OVERRIDE	(1 << 20)
+#define   HSW_PWR_WELL_FORCE_ON			(1 << 19)
+#define HSW_PWR_WELL_CTL6			_MMIO(0x45414)
+
+/* SKL Fuse Status */
+enum skl_power_gate {
+	SKL_PG0,
+	SKL_PG1,
+	SKL_PG2,
+	ICL_PG3,
+	ICL_PG4,
+};
+
+#define SKL_FUSE_STATUS				_MMIO(0x42000)
+#define  SKL_FUSE_DOWNLOAD_STATUS		(1 << 31)
+/*
+ * PG0 is HW controlled, so doesn't have a corresponding power well control knob
+ * SKL_DISP_PW1_IDX..SKL_DISP_PW2_IDX -> PG1..PG2
+ */
+#define  SKL_PW_CTL_IDX_TO_PG(pw_idx)		\
+	((pw_idx) - SKL_PW_CTL_IDX_PW_1 + SKL_PG1)
+/*
+ * PG0 is HW controlled, so doesn't have a corresponding power well control knob
+ * ICL_DISP_PW1_IDX..ICL_DISP_PW4_IDX -> PG1..PG4
+ */
+#define  ICL_PW_CTL_IDX_TO_PG(pw_idx)		\
+	((pw_idx) - ICL_PW_CTL_IDX_PW_1 + SKL_PG1)
+#define  SKL_FUSE_PG_DIST_STATUS(pg)		(1 << (27 - (pg)))
+
+#define _ICL_AUX_REG_IDX(pw_idx)	((pw_idx) - ICL_PW_CTL_IDX_AUX_A)
+#define _ICL_AUX_ANAOVRD1_A		0x162398
+#define _ICL_AUX_ANAOVRD1_B		0x6C398
+#define ICL_AUX_ANAOVRD1(pw_idx)	_MMIO(_PICK(_ICL_AUX_REG_IDX(pw_idx), \
+						    _ICL_AUX_ANAOVRD1_A, \
+						    _ICL_AUX_ANAOVRD1_B))
+#define   ICL_AUX_ANAOVRD1_LDO_BYPASS	(1 << 7)
+#define   ICL_AUX_ANAOVRD1_ENABLE	(1 << 0)
+
+/* Per-pipe DDI Function Control */
+#define _TRANS_DDI_FUNC_CTL_A		0x60400
+#define _TRANS_DDI_FUNC_CTL_B		0x61400
+#define _TRANS_DDI_FUNC_CTL_C		0x62400
+#define _TRANS_DDI_FUNC_CTL_D		0x63400
+#define _TRANS_DDI_FUNC_CTL_EDP		0x6F400
+#define _TRANS_DDI_FUNC_CTL_DSI0	0x6b400
+#define _TRANS_DDI_FUNC_CTL_DSI1	0x6bc00
+#define TRANS_DDI_FUNC_CTL(tran) _MMIO_TRANS2(tran, _TRANS_DDI_FUNC_CTL_A)
+
+#define  TRANS_DDI_FUNC_ENABLE		(1 << 31)
+/* Those bits are ignored by pipe EDP since it can only connect to DDI A */
+#define  TRANS_DDI_PORT_SHIFT		28
+#define  TGL_TRANS_DDI_PORT_SHIFT	27
+#define  TRANS_DDI_PORT_MASK		(7 << TRANS_DDI_PORT_SHIFT)
+#define  TGL_TRANS_DDI_PORT_MASK	(0xf << TGL_TRANS_DDI_PORT_SHIFT)
+#define  TRANS_DDI_SELECT_PORT(x)	((x) << TRANS_DDI_PORT_SHIFT)
+#define  TGL_TRANS_DDI_SELECT_PORT(x)	(((x) + 1) << TGL_TRANS_DDI_PORT_SHIFT)
+#define  TRANS_DDI_MODE_SELECT_MASK	(7 << 24)
+#define  TRANS_DDI_MODE_SELECT_HDMI	(0 << 24)
+#define  TRANS_DDI_MODE_SELECT_DVI	(1 << 24)
+#define  TRANS_DDI_MODE_SELECT_DP_SST	(2 << 24)
+#define  TRANS_DDI_MODE_SELECT_DP_MST	(3 << 24)
+#define  TRANS_DDI_MODE_SELECT_FDI_OR_128B132B	(4 << 24)
+#define  TRANS_DDI_BPC_MASK		(7 << 20)
+#define  TRANS_DDI_BPC_8		(0 << 20)
+#define  TRANS_DDI_BPC_10		(1 << 20)
+#define  TRANS_DDI_BPC_6		(2 << 20)
+#define  TRANS_DDI_BPC_12		(3 << 20)
+#define  TRANS_DDI_PORT_SYNC_MASTER_SELECT_MASK	REG_GENMASK(19, 18)
+#define  TRANS_DDI_PORT_SYNC_MASTER_SELECT(x)	REG_FIELD_PREP(TRANS_DDI_PORT_SYNC_MASTER_SELECT_MASK, (x))
+#define  TRANS_DDI_PVSYNC		(1 << 17)
+#define  TRANS_DDI_PHSYNC		(1 << 16)
+#define  TRANS_DDI_PORT_SYNC_ENABLE	REG_BIT(15)
+#define  TRANS_DDI_EDP_INPUT_MASK	(7 << 12)
+#define  TRANS_DDI_EDP_INPUT_A_ON	(0 << 12)
+#define  TRANS_DDI_EDP_INPUT_A_ONOFF	(4 << 12)
+#define  TRANS_DDI_EDP_INPUT_B_ONOFF	(5 << 12)
+#define  TRANS_DDI_EDP_INPUT_C_ONOFF	(6 << 12)
+#define  TRANS_DDI_EDP_INPUT_D_ONOFF	(7 << 12)
+#define  TRANS_DDI_MST_TRANSPORT_SELECT_MASK	REG_GENMASK(11, 10)
+#define  TRANS_DDI_MST_TRANSPORT_SELECT(trans)	\
+	REG_FIELD_PREP(TRANS_DDI_MST_TRANSPORT_SELECT_MASK, trans)
+#define  TRANS_DDI_HDCP_SIGNALLING	(1 << 9)
+#define  TRANS_DDI_DP_VC_PAYLOAD_ALLOC	(1 << 8)
+#define  TRANS_DDI_HDMI_SCRAMBLER_CTS_ENABLE (1 << 7)
+#define  TRANS_DDI_HDMI_SCRAMBLER_RESET_FREQ (1 << 6)
+#define  TRANS_DDI_HDCP_SELECT		REG_BIT(5)
+#define  TRANS_DDI_BFI_ENABLE		(1 << 4)
+#define  TRANS_DDI_HIGH_TMDS_CHAR_RATE	(1 << 4)
+#define  TRANS_DDI_HDMI_SCRAMBLING	(1 << 0)
+#define  TRANS_DDI_HDMI_SCRAMBLING_MASK (TRANS_DDI_HDMI_SCRAMBLER_CTS_ENABLE \
+					| TRANS_DDI_HDMI_SCRAMBLER_RESET_FREQ \
+					| TRANS_DDI_HDMI_SCRAMBLING)
+
+#define _TRANS_DDI_FUNC_CTL2_A		0x60404
+#define _TRANS_DDI_FUNC_CTL2_B		0x61404
+#define _TRANS_DDI_FUNC_CTL2_C		0x62404
+#define _TRANS_DDI_FUNC_CTL2_EDP	0x6f404
+#define _TRANS_DDI_FUNC_CTL2_DSI0	0x6b404
+#define _TRANS_DDI_FUNC_CTL2_DSI1	0x6bc04
+#define TRANS_DDI_FUNC_CTL2(tran)	_MMIO_TRANS2(tran, _TRANS_DDI_FUNC_CTL2_A)
+#define  PORT_SYNC_MODE_ENABLE			REG_BIT(4)
+#define  PORT_SYNC_MODE_MASTER_SELECT_MASK	REG_GENMASK(2, 0)
+#define  PORT_SYNC_MODE_MASTER_SELECT(x)	REG_FIELD_PREP(PORT_SYNC_MODE_MASTER_SELECT_MASK, (x))
+
+#define TRANS_CMTG_CHICKEN		_MMIO(0x6fa90)
+#define  DISABLE_DPT_CLK_GATING		REG_BIT(1)
+
+/* DisplayPort Transport Control */
+#define _DP_TP_CTL_A			0x64040
+#define _DP_TP_CTL_B			0x64140
+#define _TGL_DP_TP_CTL_A		0x60540
+#define DP_TP_CTL(port) _MMIO_PORT(port, _DP_TP_CTL_A, _DP_TP_CTL_B)
+#define TGL_DP_TP_CTL(tran) _MMIO_TRANS2((tran), _TGL_DP_TP_CTL_A)
+#define  DP_TP_CTL_ENABLE			(1 << 31)
+#define  DP_TP_CTL_FEC_ENABLE			(1 << 30)
+#define  DP_TP_CTL_MODE_SST			(0 << 27)
+#define  DP_TP_CTL_MODE_MST			(1 << 27)
+#define  DP_TP_CTL_FORCE_ACT			(1 << 25)
+#define  DP_TP_CTL_ENHANCED_FRAME_ENABLE	(1 << 18)
+#define  DP_TP_CTL_FDI_AUTOTRAIN		(1 << 15)
+#define  DP_TP_CTL_LINK_TRAIN_MASK		(7 << 8)
+#define  DP_TP_CTL_LINK_TRAIN_PAT1		(0 << 8)
+#define  DP_TP_CTL_LINK_TRAIN_PAT2		(1 << 8)
+#define  DP_TP_CTL_LINK_TRAIN_PAT3		(4 << 8)
+#define  DP_TP_CTL_LINK_TRAIN_PAT4		(5 << 8)
+#define  DP_TP_CTL_LINK_TRAIN_IDLE		(2 << 8)
+#define  DP_TP_CTL_LINK_TRAIN_NORMAL		(3 << 8)
+#define  DP_TP_CTL_SCRAMBLE_DISABLE		(1 << 7)
+
+/* DisplayPort Transport Status */
+#define _DP_TP_STATUS_A			0x64044
+#define _DP_TP_STATUS_B			0x64144
+#define _TGL_DP_TP_STATUS_A		0x60544
+#define DP_TP_STATUS(port) _MMIO_PORT(port, _DP_TP_STATUS_A, _DP_TP_STATUS_B)
+#define TGL_DP_TP_STATUS(tran) _MMIO_TRANS2((tran), _TGL_DP_TP_STATUS_A)
+#define  DP_TP_STATUS_FEC_ENABLE_LIVE		(1 << 28)
+#define  DP_TP_STATUS_IDLE_DONE			(1 << 25)
+#define  DP_TP_STATUS_ACT_SENT			(1 << 24)
+#define  DP_TP_STATUS_MODE_STATUS_MST		(1 << 23)
+#define  DP_TP_STATUS_AUTOTRAIN_DONE		(1 << 12)
+#define  DP_TP_STATUS_PAYLOAD_MAPPING_VC2	(3 << 8)
+#define  DP_TP_STATUS_PAYLOAD_MAPPING_VC1	(3 << 4)
+#define  DP_TP_STATUS_PAYLOAD_MAPPING_VC0	(3 << 0)
+
+/* DDI Buffer Control */
+#define _DDI_BUF_CTL_A				0x64000
+#define _DDI_BUF_CTL_B				0x64100
+#define DDI_BUF_CTL(port) _MMIO_PORT(port, _DDI_BUF_CTL_A, _DDI_BUF_CTL_B)
+#define  DDI_BUF_CTL_ENABLE			(1 << 31)
+#define  DDI_BUF_TRANS_SELECT(n)	((n) << 24)
+#define  DDI_BUF_EMP_MASK			(0xf << 24)
+#define  DDI_BUF_PHY_LINK_RATE(r)		((r) << 20)
+#define  DDI_BUF_PORT_REVERSAL			(1 << 16)
+#define  DDI_BUF_IS_IDLE			(1 << 7)
+#define  DDI_BUF_CTL_TC_PHY_OWNERSHIP		REG_BIT(6)
+#define  DDI_A_4_LANES				(1 << 4)
+#define  DDI_PORT_WIDTH(width)			(((width) - 1) << 1)
+#define  DDI_PORT_WIDTH_MASK			(7 << 1)
+#define  DDI_PORT_WIDTH_SHIFT			1
+#define  DDI_INIT_DISPLAY_DETECTED		(1 << 0)
+
+/* DDI Buffer Translations */
+#define _DDI_BUF_TRANS_A		0x64E00
+#define _DDI_BUF_TRANS_B		0x64E60
+#define DDI_BUF_TRANS_LO(port, i)	_MMIO(_PORT(port, _DDI_BUF_TRANS_A, _DDI_BUF_TRANS_B) + (i) * 8)
+#define  DDI_BUF_BALANCE_LEG_ENABLE	(1 << 31)
+#define DDI_BUF_TRANS_HI(port, i)	_MMIO(_PORT(port, _DDI_BUF_TRANS_A, _DDI_BUF_TRANS_B) + (i) * 8 + 4)
+
+/* DDI DP Compliance Control */
+#define _DDI_DP_COMP_CTL_A			0x605F0
+#define _DDI_DP_COMP_CTL_B			0x615F0
+#define DDI_DP_COMP_CTL(pipe)			_MMIO_PIPE(pipe, _DDI_DP_COMP_CTL_A, _DDI_DP_COMP_CTL_B)
+#define   DDI_DP_COMP_CTL_ENABLE		(1 << 31)
+#define   DDI_DP_COMP_CTL_D10_2			(0 << 28)
+#define   DDI_DP_COMP_CTL_SCRAMBLED_0		(1 << 28)
+#define   DDI_DP_COMP_CTL_PRBS7			(2 << 28)
+#define   DDI_DP_COMP_CTL_CUSTOM80		(3 << 28)
+#define   DDI_DP_COMP_CTL_HBR2			(4 << 28)
+#define   DDI_DP_COMP_CTL_SCRAMBLED_1		(5 << 28)
+#define   DDI_DP_COMP_CTL_HBR2_RESET		(0xFC << 0)
+
+/* DDI DP Compliance Pattern */
+#define _DDI_DP_COMP_PAT_A			0x605F4
+#define _DDI_DP_COMP_PAT_B			0x615F4
+#define DDI_DP_COMP_PAT(pipe, i)		_MMIO(_PIPE(pipe, _DDI_DP_COMP_PAT_A, _DDI_DP_COMP_PAT_B) + (i) * 4)
+
+/* Sideband Interface (SBI) is programmed indirectly, via
+ * SBI_ADDR, which contains the register offset; and SBI_DATA,
+ * which contains the payload */
+#define SBI_ADDR			_MMIO(0xC6000)
+#define SBI_DATA			_MMIO(0xC6004)
+#define SBI_CTL_STAT			_MMIO(0xC6008)
+#define  SBI_CTL_DEST_ICLK		(0x0 << 16)
+#define  SBI_CTL_DEST_MPHY		(0x1 << 16)
+#define  SBI_CTL_OP_IORD		(0x2 << 8)
+#define  SBI_CTL_OP_IOWR		(0x3 << 8)
+#define  SBI_CTL_OP_CRRD		(0x6 << 8)
+#define  SBI_CTL_OP_CRWR		(0x7 << 8)
+#define  SBI_RESPONSE_FAIL		(0x1 << 1)
+#define  SBI_RESPONSE_SUCCESS		(0x0 << 1)
+#define  SBI_BUSY			(0x1 << 0)
+#define  SBI_READY			(0x0 << 0)
+
+/* SBI offsets */
+#define  SBI_SSCDIVINTPHASE			0x0200
+#define  SBI_SSCDIVINTPHASE6			0x0600
+#define   SBI_SSCDIVINTPHASE_DIVSEL_SHIFT	1
+#define   SBI_SSCDIVINTPHASE_DIVSEL_MASK	(0x7f << 1)
+#define   SBI_SSCDIVINTPHASE_DIVSEL(x)		((x) << 1)
+#define   SBI_SSCDIVINTPHASE_INCVAL_SHIFT	8
+#define   SBI_SSCDIVINTPHASE_INCVAL_MASK	(0x7f << 8)
+#define   SBI_SSCDIVINTPHASE_INCVAL(x)		((x) << 8)
+#define   SBI_SSCDIVINTPHASE_DIR(x)		((x) << 15)
+#define   SBI_SSCDIVINTPHASE_PROPAGATE		(1 << 0)
+#define  SBI_SSCDITHPHASE			0x0204
+#define  SBI_SSCCTL				0x020c
+#define  SBI_SSCCTL6				0x060C
+#define   SBI_SSCCTL_PATHALT			(1 << 3)
+#define   SBI_SSCCTL_DISABLE			(1 << 0)
+#define  SBI_SSCAUXDIV6				0x0610
+#define   SBI_SSCAUXDIV_FINALDIV2SEL_SHIFT	4
+#define   SBI_SSCAUXDIV_FINALDIV2SEL_MASK	(1 << 4)
+#define   SBI_SSCAUXDIV_FINALDIV2SEL(x)		((x) << 4)
+#define  SBI_DBUFF0				0x2a00
+#define  SBI_GEN0				0x1f00
+#define   SBI_GEN0_CFG_BUFFENABLE_DISABLE	(1 << 0)
+
+/* LPT PIXCLK_GATE */
+#define PIXCLK_GATE			_MMIO(0xC6020)
+#define  PIXCLK_GATE_UNGATE		(1 << 0)
+#define  PIXCLK_GATE_GATE		(0 << 0)
+
+/* SPLL */
+#define SPLL_CTL			_MMIO(0x46020)
+#define  SPLL_PLL_ENABLE		(1 << 31)
+#define  SPLL_REF_BCLK			(0 << 28)
+#define  SPLL_REF_MUXED_SSC		(1 << 28) /* CPU SSC if fused enabled, PCH SSC otherwise */
+#define  SPLL_REF_NON_SSC_HSW		(2 << 28)
+#define  SPLL_REF_PCH_SSC_BDW		(2 << 28)
+#define  SPLL_REF_LCPLL			(3 << 28)
+#define  SPLL_REF_MASK			(3 << 28)
+#define  SPLL_FREQ_810MHz		(0 << 26)
+#define  SPLL_FREQ_1350MHz		(1 << 26)
+#define  SPLL_FREQ_2700MHz		(2 << 26)
+#define  SPLL_FREQ_MASK			(3 << 26)
+
+/* WRPLL */
+#define _WRPLL_CTL1			0x46040
+#define _WRPLL_CTL2			0x46060
+#define WRPLL_CTL(pll)			_MMIO_PIPE(pll, _WRPLL_CTL1, _WRPLL_CTL2)
+#define  WRPLL_PLL_ENABLE		(1 << 31)
+#define  WRPLL_REF_BCLK			(0 << 28)
+#define  WRPLL_REF_PCH_SSC		(1 << 28)
+#define  WRPLL_REF_MUXED_SSC_BDW	(2 << 28) /* CPU SSC if fused enabled, PCH SSC otherwise */
+#define  WRPLL_REF_SPECIAL_HSW		(2 << 28) /* muxed SSC (ULT), non-SSC (non-ULT) */
+#define  WRPLL_REF_LCPLL		(3 << 28)
+#define  WRPLL_REF_MASK			(3 << 28)
+/* WRPLL divider programming */
+#define  WRPLL_DIVIDER_REFERENCE(x)	((x) << 0)
+#define  WRPLL_DIVIDER_REF_MASK		(0xff)
+#define  WRPLL_DIVIDER_POST(x)		((x) << 8)
+#define  WRPLL_DIVIDER_POST_MASK	(0x3f << 8)
+#define  WRPLL_DIVIDER_POST_SHIFT	8
+#define  WRPLL_DIVIDER_FEEDBACK(x)	((x) << 16)
+#define  WRPLL_DIVIDER_FB_SHIFT		16
+#define  WRPLL_DIVIDER_FB_MASK		(0xff << 16)
+
+/* Port clock selection */
+#define _PORT_CLK_SEL_A			0x46100
+#define _PORT_CLK_SEL_B			0x46104
+#define PORT_CLK_SEL(port) _MMIO_PORT(port, _PORT_CLK_SEL_A, _PORT_CLK_SEL_B)
+#define  PORT_CLK_SEL_MASK		REG_GENMASK(31, 29)
+#define  PORT_CLK_SEL_LCPLL_2700	REG_FIELD_PREP(PORT_CLK_SEL_MASK, 0)
+#define  PORT_CLK_SEL_LCPLL_1350	REG_FIELD_PREP(PORT_CLK_SEL_MASK, 1)
+#define  PORT_CLK_SEL_LCPLL_810		REG_FIELD_PREP(PORT_CLK_SEL_MASK, 2)
+#define  PORT_CLK_SEL_SPLL		REG_FIELD_PREP(PORT_CLK_SEL_MASK, 3)
+#define  PORT_CLK_SEL_WRPLL(pll)	REG_FIELD_PREP(PORT_CLK_SEL_MASK, 4 + (pll))
+#define  PORT_CLK_SEL_WRPLL1		REG_FIELD_PREP(PORT_CLK_SEL_MASK, 4)
+#define  PORT_CLK_SEL_WRPLL2		REG_FIELD_PREP(PORT_CLK_SEL_MASK, 5)
+#define  PORT_CLK_SEL_NONE		REG_FIELD_PREP(PORT_CLK_SEL_MASK, 7)
+
+/* On ICL+ this is the same as PORT_CLK_SEL, but all bits change. */
+#define DDI_CLK_SEL(port)		PORT_CLK_SEL(port)
+#define  DDI_CLK_SEL_MASK		REG_GENMASK(31, 28)
+#define  DDI_CLK_SEL_NONE		REG_FIELD_PREP(DDI_CLK_SEL_MASK, 0x0)
+#define  DDI_CLK_SEL_MG			REG_FIELD_PREP(DDI_CLK_SEL_MASK, 0x8)
+#define  DDI_CLK_SEL_TBT_162		REG_FIELD_PREP(DDI_CLK_SEL_MASK, 0xC)
+#define  DDI_CLK_SEL_TBT_270		REG_FIELD_PREP(DDI_CLK_SEL_MASK, 0xD)
+#define  DDI_CLK_SEL_TBT_540		REG_FIELD_PREP(DDI_CLK_SEL_MASK, 0xE)
+#define  DDI_CLK_SEL_TBT_810		REG_FIELD_PREP(DDI_CLK_SEL_MASK, 0xF)
+
+/* Transcoder clock selection */
+#define _TRANS_CLK_SEL_A		0x46140
+#define _TRANS_CLK_SEL_B		0x46144
+#define TRANS_CLK_SEL(tran) _MMIO_TRANS(tran, _TRANS_CLK_SEL_A, _TRANS_CLK_SEL_B)
+/* For each transcoder, we need to select the corresponding port clock */
+#define  TRANS_CLK_SEL_DISABLED		(0x0 << 29)
+#define  TRANS_CLK_SEL_PORT(x)		(((x) + 1) << 29)
+#define  TGL_TRANS_CLK_SEL_DISABLED	(0x0 << 28)
+#define  TGL_TRANS_CLK_SEL_PORT(x)	(((x) + 1) << 28)
+
+
+#define CDCLK_FREQ			_MMIO(0x46200)
+
+#define _TRANSA_MSA_MISC		0x60410
+#define _TRANSB_MSA_MISC		0x61410
+#define _TRANSC_MSA_MISC		0x62410
+#define _TRANS_EDP_MSA_MISC		0x6f410
+#define TRANS_MSA_MISC(tran) _MMIO_TRANS2(tran, _TRANSA_MSA_MISC)
+/* See DP_MSA_MISC_* for the bit definitions */
+
+#define _TRANS_A_SET_CONTEXT_LATENCY		0x6007C
+#define _TRANS_B_SET_CONTEXT_LATENCY		0x6107C
+#define _TRANS_C_SET_CONTEXT_LATENCY		0x6207C
+#define _TRANS_D_SET_CONTEXT_LATENCY		0x6307C
+#define TRANS_SET_CONTEXT_LATENCY(tran)		_MMIO_TRANS2(tran, _TRANS_A_SET_CONTEXT_LATENCY)
+#define  TRANS_SET_CONTEXT_LATENCY_MASK		REG_GENMASK(15, 0)
+#define  TRANS_SET_CONTEXT_LATENCY_VALUE(x)	REG_FIELD_PREP(TRANS_SET_CONTEXT_LATENCY_MASK, (x))
+
+/* LCPLL Control */
+#define LCPLL_CTL			_MMIO(0x130040)
+#define  LCPLL_PLL_DISABLE		(1 << 31)
+#define  LCPLL_PLL_LOCK			(1 << 30)
+#define  LCPLL_REF_NON_SSC		(0 << 28)
+#define  LCPLL_REF_BCLK			(2 << 28)
+#define  LCPLL_REF_PCH_SSC		(3 << 28)
+#define  LCPLL_REF_MASK			(3 << 28)
+#define  LCPLL_CLK_FREQ_MASK		(3 << 26)
+#define  LCPLL_CLK_FREQ_450		(0 << 26)
+#define  LCPLL_CLK_FREQ_54O_BDW		(1 << 26)
+#define  LCPLL_CLK_FREQ_337_5_BDW	(2 << 26)
+#define  LCPLL_CLK_FREQ_675_BDW		(3 << 26)
+#define  LCPLL_CD_CLOCK_DISABLE		(1 << 25)
+#define  LCPLL_ROOT_CD_CLOCK_DISABLE	(1 << 24)
+#define  LCPLL_CD2X_CLOCK_DISABLE	(1 << 23)
+#define  LCPLL_POWER_DOWN_ALLOW		(1 << 22)
+#define  LCPLL_CD_SOURCE_FCLK		(1 << 21)
+#define  LCPLL_CD_SOURCE_FCLK_DONE	(1 << 19)
+
+/*
+ * SKL Clocks
+ */
+
+/* CDCLK_CTL */
+#define CDCLK_CTL			_MMIO(0x46000)
+#define  CDCLK_FREQ_SEL_MASK		REG_GENMASK(27, 26)
+#define  CDCLK_FREQ_450_432		REG_FIELD_PREP(CDCLK_FREQ_SEL_MASK, 0)
+#define  CDCLK_FREQ_540		REG_FIELD_PREP(CDCLK_FREQ_SEL_MASK, 1)
+#define  CDCLK_FREQ_337_308		REG_FIELD_PREP(CDCLK_FREQ_SEL_MASK, 2)
+#define  CDCLK_FREQ_675_617		REG_FIELD_PREP(CDCLK_FREQ_SEL_MASK, 3)
+#define  BXT_CDCLK_CD2X_DIV_SEL_MASK	REG_GENMASK(23, 22)
+#define  BXT_CDCLK_CD2X_DIV_SEL_1	REG_FIELD_PREP(BXT_CDCLK_CD2X_DIV_SEL_MASK, 0)
+#define  BXT_CDCLK_CD2X_DIV_SEL_1_5	REG_FIELD_PREP(BXT_CDCLK_CD2X_DIV_SEL_MASK, 1)
+#define  BXT_CDCLK_CD2X_DIV_SEL_2	REG_FIELD_PREP(BXT_CDCLK_CD2X_DIV_SEL_MASK, 2)
+#define  BXT_CDCLK_CD2X_DIV_SEL_4	REG_FIELD_PREP(BXT_CDCLK_CD2X_DIV_SEL_MASK, 3)
+#define  BXT_CDCLK_CD2X_PIPE(pipe)	((pipe) << 20)
+#define  CDCLK_DIVMUX_CD_OVERRIDE	(1 << 19)
+#define  BXT_CDCLK_CD2X_PIPE_NONE	BXT_CDCLK_CD2X_PIPE(3)
+#define  ICL_CDCLK_CD2X_PIPE(pipe)	(_PICK(pipe, 0, 2, 6) << 19)
+#define  ICL_CDCLK_CD2X_PIPE_NONE	(7 << 19)
+#define  TGL_CDCLK_CD2X_PIPE(pipe)	BXT_CDCLK_CD2X_PIPE(pipe)
+#define  TGL_CDCLK_CD2X_PIPE_NONE	ICL_CDCLK_CD2X_PIPE_NONE
+#define  BXT_CDCLK_SSA_PRECHARGE_ENABLE	(1 << 16)
+#define  CDCLK_FREQ_DECIMAL_MASK	(0x7ff)
+
+/* CDCLK_SQUASH_CTL */
+#define CDCLK_SQUASH_CTL		_MMIO(0x46008)
+#define  CDCLK_SQUASH_ENABLE		REG_BIT(31)
+#define  CDCLK_SQUASH_WINDOW_SIZE_MASK	REG_GENMASK(27, 24)
+#define  CDCLK_SQUASH_WINDOW_SIZE(x)	REG_FIELD_PREP(CDCLK_SQUASH_WINDOW_SIZE_MASK, (x))
+#define  CDCLK_SQUASH_WAVEFORM_MASK	REG_GENMASK(15, 0)
+#define  CDCLK_SQUASH_WAVEFORM(x)	REG_FIELD_PREP(CDCLK_SQUASH_WAVEFORM_MASK, (x))
+
+/* LCPLL_CTL */
+#define LCPLL1_CTL		_MMIO(0x46010)
+#define LCPLL2_CTL		_MMIO(0x46014)
+#define  LCPLL_PLL_ENABLE	(1 << 31)
+
+/* DPLL control1 */
+#define DPLL_CTRL1		_MMIO(0x6C058)
+#define  DPLL_CTRL1_HDMI_MODE(id)		(1 << ((id) * 6 + 5))
+#define  DPLL_CTRL1_SSC(id)			(1 << ((id) * 6 + 4))
+#define  DPLL_CTRL1_LINK_RATE_MASK(id)		(7 << ((id) * 6 + 1))
+#define  DPLL_CTRL1_LINK_RATE_SHIFT(id)		((id) * 6 + 1)
+#define  DPLL_CTRL1_LINK_RATE(linkrate, id)	((linkrate) << ((id) * 6 + 1))
+#define  DPLL_CTRL1_OVERRIDE(id)		(1 << ((id) * 6))
+#define  DPLL_CTRL1_LINK_RATE_2700		0
+#define  DPLL_CTRL1_LINK_RATE_1350		1
+#define  DPLL_CTRL1_LINK_RATE_810		2
+#define  DPLL_CTRL1_LINK_RATE_1620		3
+#define  DPLL_CTRL1_LINK_RATE_1080		4
+#define  DPLL_CTRL1_LINK_RATE_2160		5
+
+/* DPLL control2 */
+#define DPLL_CTRL2				_MMIO(0x6C05C)
+#define  DPLL_CTRL2_DDI_CLK_OFF(port)		(1 << ((port) + 15))
+#define  DPLL_CTRL2_DDI_CLK_SEL_MASK(port)	(3 << ((port) * 3 + 1))
+#define  DPLL_CTRL2_DDI_CLK_SEL_SHIFT(port)    ((port) * 3 + 1)
+#define  DPLL_CTRL2_DDI_CLK_SEL(clk, port)	((clk) << ((port) * 3 + 1))
+#define  DPLL_CTRL2_DDI_SEL_OVERRIDE(port)     (1 << ((port) * 3))
+
+/* DPLL Status */
+#define DPLL_STATUS	_MMIO(0x6C060)
+#define  DPLL_LOCK(id) (1 << ((id) * 8))
+
+/* DPLL cfg */
+#define _DPLL1_CFGCR1	0x6C040
+#define _DPLL2_CFGCR1	0x6C048
+#define _DPLL3_CFGCR1	0x6C050
+#define  DPLL_CFGCR1_FREQ_ENABLE	(1 << 31)
+#define  DPLL_CFGCR1_DCO_FRACTION_MASK	(0x7fff << 9)
+#define  DPLL_CFGCR1_DCO_FRACTION(x)	((x) << 9)
+#define  DPLL_CFGCR1_DCO_INTEGER_MASK	(0x1ff)
+
+#define _DPLL1_CFGCR2	0x6C044
+#define _DPLL2_CFGCR2	0x6C04C
+#define _DPLL3_CFGCR2	0x6C054
+#define  DPLL_CFGCR2_QDIV_RATIO_MASK	(0xff << 8)
+#define  DPLL_CFGCR2_QDIV_RATIO(x)	((x) << 8)
+#define  DPLL_CFGCR2_QDIV_MODE(x)	((x) << 7)
+#define  DPLL_CFGCR2_KDIV_MASK		(3 << 5)
+#define  DPLL_CFGCR2_KDIV(x)		((x) << 5)
+#define  DPLL_CFGCR2_KDIV_5 (0 << 5)
+#define  DPLL_CFGCR2_KDIV_2 (1 << 5)
+#define  DPLL_CFGCR2_KDIV_3 (2 << 5)
+#define  DPLL_CFGCR2_KDIV_1 (3 << 5)
+#define  DPLL_CFGCR2_PDIV_MASK		(7 << 2)
+#define  DPLL_CFGCR2_PDIV(x)		((x) << 2)
+#define  DPLL_CFGCR2_PDIV_1 (0 << 2)
+#define  DPLL_CFGCR2_PDIV_2 (1 << 2)
+#define  DPLL_CFGCR2_PDIV_3 (2 << 2)
+#define  DPLL_CFGCR2_PDIV_7 (4 << 2)
+#define  DPLL_CFGCR2_PDIV_7_INVALID	(5 << 2)
+#define  DPLL_CFGCR2_CENTRAL_FREQ_MASK	(3)
+
+#define DPLL_CFGCR1(id)	_MMIO_PIPE((id) - SKL_DPLL1, _DPLL1_CFGCR1, _DPLL2_CFGCR1)
+#define DPLL_CFGCR2(id)	_MMIO_PIPE((id) - SKL_DPLL1, _DPLL1_CFGCR2, _DPLL2_CFGCR2)
+
+/* ICL Clocks */
+#define ICL_DPCLKA_CFGCR0			_MMIO(0x164280)
+#define  ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy)	(1 << _PICK(phy, 10, 11, 24, 4, 5))
+#define  RKL_DPCLKA_CFGCR0_DDI_CLK_OFF(phy)	REG_BIT((phy) + 10)
+#define  ICL_DPCLKA_CFGCR0_TC_CLK_OFF(tc_port)	(1 << ((tc_port) < TC_PORT_4 ? \
+						       (tc_port) + 12 : \
+						       (tc_port) - TC_PORT_4 + 21))
+#define  ICL_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy)	((phy) * 2)
+#define  ICL_DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(phy)	(3 << ICL_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy))
+#define  ICL_DPCLKA_CFGCR0_DDI_CLK_SEL(pll, phy)	((pll) << ICL_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy))
+#define  RKL_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy)	_PICK(phy, 0, 2, 4, 27)
+#define  RKL_DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(phy) \
+	(3 << RKL_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy))
+#define  RKL_DPCLKA_CFGCR0_DDI_CLK_SEL(pll, phy) \
+	((pll) << RKL_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy))
+
+/*
+ * DG1 Clocks
+ * First registers controls the first A and B, while the second register
+ * controls the phy C and D. The bits on these registers are the
+ * same, but refer to different phys
+ */
+#define _DG1_DPCLKA_CFGCR0				0x164280
+#define _DG1_DPCLKA1_CFGCR0				0x16C280
+#define _DG1_DPCLKA_PHY_IDX(phy)			((phy) % 2)
+#define _DG1_DPCLKA_PLL_IDX(pll)			((pll) % 2)
+#define DG1_DPCLKA_CFGCR0(phy)				_MMIO_PHY((phy) / 2, \
+								  _DG1_DPCLKA_CFGCR0, \
+								  _DG1_DPCLKA1_CFGCR0)
+#define   DG1_DPCLKA_CFGCR0_DDI_CLK_OFF(phy)		REG_BIT(_DG1_DPCLKA_PHY_IDX(phy) + 10)
+#define   DG1_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy)	(_DG1_DPCLKA_PHY_IDX(phy) * 2)
+#define   DG1_DPCLKA_CFGCR0_DDI_CLK_SEL(pll, phy)	(_DG1_DPCLKA_PLL_IDX(pll) << DG1_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy))
+#define   DG1_DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(phy)	(0x3 << DG1_DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(phy))
+
+/* ADLS Clocks */
+#define _ADLS_DPCLKA_CFGCR0			0x164280
+#define _ADLS_DPCLKA_CFGCR1			0x1642BC
+#define ADLS_DPCLKA_CFGCR(phy)			_MMIO_PHY((phy) / 3, \
+							  _ADLS_DPCLKA_CFGCR0, \
+							  _ADLS_DPCLKA_CFGCR1)
+#define  ADLS_DPCLKA_CFGCR_DDI_SHIFT(phy)		(((phy) % 3) * 2)
+/* ADLS DPCLKA_CFGCR0 DDI mask */
+#define  ADLS_DPCLKA_DDII_SEL_MASK			REG_GENMASK(5, 4)
+#define  ADLS_DPCLKA_DDIB_SEL_MASK			REG_GENMASK(3, 2)
+#define  ADLS_DPCLKA_DDIA_SEL_MASK			REG_GENMASK(1, 0)
+/* ADLS DPCLKA_CFGCR1 DDI mask */
+#define  ADLS_DPCLKA_DDIK_SEL_MASK			REG_GENMASK(3, 2)
+#define  ADLS_DPCLKA_DDIJ_SEL_MASK			REG_GENMASK(1, 0)
+#define  ADLS_DPCLKA_CFGCR_DDI_CLK_SEL_MASK(phy)	_PICK((phy), \
+							ADLS_DPCLKA_DDIA_SEL_MASK, \
+							ADLS_DPCLKA_DDIB_SEL_MASK, \
+							ADLS_DPCLKA_DDII_SEL_MASK, \
+							ADLS_DPCLKA_DDIJ_SEL_MASK, \
+							ADLS_DPCLKA_DDIK_SEL_MASK)
+
+/* ICL PLL */
+#define DPLL0_ENABLE		0x46010
+#define DPLL1_ENABLE		0x46014
+#define _ADLS_DPLL2_ENABLE	0x46018
+#define _ADLS_DPLL3_ENABLE	0x46030
+#define  PLL_ENABLE		(1 << 31)
+#define  PLL_LOCK		(1 << 30)
+#define  PLL_POWER_ENABLE	(1 << 27)
+#define  PLL_POWER_STATE	(1 << 26)
+#define ICL_DPLL_ENABLE(pll)	_MMIO_PLL3(pll, DPLL0_ENABLE, DPLL1_ENABLE, \
+					   _ADLS_DPLL2_ENABLE, _ADLS_DPLL3_ENABLE)
+
+#define _DG2_PLL3_ENABLE	0x4601C
+
+#define DG2_PLL_ENABLE(pll) _MMIO_PLL3(pll, DPLL0_ENABLE, DPLL1_ENABLE, \
+				       _ADLS_DPLL2_ENABLE, _DG2_PLL3_ENABLE)
+
+#define TBT_PLL_ENABLE		_MMIO(0x46020)
+
+#define _MG_PLL1_ENABLE		0x46030
+#define _MG_PLL2_ENABLE		0x46034
+#define _MG_PLL3_ENABLE		0x46038
+#define _MG_PLL4_ENABLE		0x4603C
+/* Bits are the same as DPLL0_ENABLE */
+#define MG_PLL_ENABLE(tc_port)	_MMIO_PORT((tc_port), _MG_PLL1_ENABLE, \
+					   _MG_PLL2_ENABLE)
+
+/* DG1 PLL */
+#define DG1_DPLL_ENABLE(pll)    _MMIO_PLL3(pll, DPLL0_ENABLE, DPLL1_ENABLE, \
+					   _MG_PLL1_ENABLE, _MG_PLL2_ENABLE)
+
+/* ADL-P Type C PLL */
+#define PORTTC1_PLL_ENABLE	0x46038
+#define PORTTC2_PLL_ENABLE	0x46040
+
+#define ADLP_PORTTC_PLL_ENABLE(tc_port)		_MMIO_PORT((tc_port), \
+							    PORTTC1_PLL_ENABLE, \
+							    PORTTC2_PLL_ENABLE)
+
+#define _ICL_DPLL0_CFGCR0		0x164000
+#define _ICL_DPLL1_CFGCR0		0x164080
+#define ICL_DPLL_CFGCR0(pll)		_MMIO_PLL(pll, _ICL_DPLL0_CFGCR0, \
+						  _ICL_DPLL1_CFGCR0)
+#define   DPLL_CFGCR0_HDMI_MODE		(1 << 30)
+#define   DPLL_CFGCR0_SSC_ENABLE	(1 << 29)
+#define   DPLL_CFGCR0_SSC_ENABLE_ICL	(1 << 25)
+#define   DPLL_CFGCR0_LINK_RATE_MASK	(0xf << 25)
+#define   DPLL_CFGCR0_LINK_RATE_2700	(0 << 25)
+#define   DPLL_CFGCR0_LINK_RATE_1350	(1 << 25)
+#define   DPLL_CFGCR0_LINK_RATE_810	(2 << 25)
+#define   DPLL_CFGCR0_LINK_RATE_1620	(3 << 25)
+#define   DPLL_CFGCR0_LINK_RATE_1080	(4 << 25)
+#define   DPLL_CFGCR0_LINK_RATE_2160	(5 << 25)
+#define   DPLL_CFGCR0_LINK_RATE_3240	(6 << 25)
+#define   DPLL_CFGCR0_LINK_RATE_4050	(7 << 25)
+#define   DPLL_CFGCR0_DCO_FRACTION_MASK	(0x7fff << 10)
+#define   DPLL_CFGCR0_DCO_FRACTION_SHIFT	(10)
+#define   DPLL_CFGCR0_DCO_FRACTION(x)	((x) << 10)
+#define   DPLL_CFGCR0_DCO_INTEGER_MASK	(0x3ff)
+
+#define _ICL_DPLL0_CFGCR1		0x164004
+#define _ICL_DPLL1_CFGCR1		0x164084
+#define ICL_DPLL_CFGCR1(pll)		_MMIO_PLL(pll, _ICL_DPLL0_CFGCR1, \
+						  _ICL_DPLL1_CFGCR1)
+#define   DPLL_CFGCR1_QDIV_RATIO_MASK	(0xff << 10)
+#define   DPLL_CFGCR1_QDIV_RATIO_SHIFT	(10)
+#define   DPLL_CFGCR1_QDIV_RATIO(x)	((x) << 10)
+#define   DPLL_CFGCR1_QDIV_MODE_SHIFT	(9)
+#define   DPLL_CFGCR1_QDIV_MODE(x)	((x) << 9)
+#define   DPLL_CFGCR1_KDIV_MASK		(7 << 6)
+#define   DPLL_CFGCR1_KDIV_SHIFT		(6)
+#define   DPLL_CFGCR1_KDIV(x)		((x) << 6)
+#define   DPLL_CFGCR1_KDIV_1		(1 << 6)
+#define   DPLL_CFGCR1_KDIV_2		(2 << 6)
+#define   DPLL_CFGCR1_KDIV_3		(4 << 6)
+#define   DPLL_CFGCR1_PDIV_MASK		(0xf << 2)
+#define   DPLL_CFGCR1_PDIV_SHIFT		(2)
+#define   DPLL_CFGCR1_PDIV(x)		((x) << 2)
+#define   DPLL_CFGCR1_PDIV_2		(1 << 2)
+#define   DPLL_CFGCR1_PDIV_3		(2 << 2)
+#define   DPLL_CFGCR1_PDIV_5		(4 << 2)
+#define   DPLL_CFGCR1_PDIV_7		(8 << 2)
+#define   DPLL_CFGCR1_CENTRAL_FREQ	(3 << 0)
+#define   DPLL_CFGCR1_CENTRAL_FREQ_8400	(3 << 0)
+#define   TGL_DPLL_CFGCR1_CFSELOVRD_NORMAL_XTAL	(0 << 0)
+
+#define _TGL_DPLL0_CFGCR0		0x164284
+#define _TGL_DPLL1_CFGCR0		0x16428C
+#define _TGL_TBTPLL_CFGCR0		0x16429C
+#define TGL_DPLL_CFGCR0(pll)		_MMIO_PLL3(pll, _TGL_DPLL0_CFGCR0, \
+						  _TGL_DPLL1_CFGCR0, \
+						  _TGL_TBTPLL_CFGCR0)
+#define RKL_DPLL_CFGCR0(pll)		_MMIO_PLL(pll, _TGL_DPLL0_CFGCR0, \
+						  _TGL_DPLL1_CFGCR0)
+
+#define _TGL_DPLL0_DIV0					0x164B00
+#define _TGL_DPLL1_DIV0					0x164C00
+#define TGL_DPLL0_DIV0(pll)				_MMIO_PLL(pll, _TGL_DPLL0_DIV0, _TGL_DPLL1_DIV0)
+#define   TGL_DPLL0_DIV0_AFC_STARTUP_MASK		REG_GENMASK(27, 25)
+#define   TGL_DPLL0_DIV0_AFC_STARTUP(val)		REG_FIELD_PREP(TGL_DPLL0_DIV0_AFC_STARTUP_MASK, (val))
+
+#define _TGL_DPLL0_CFGCR1		0x164288
+#define _TGL_DPLL1_CFGCR1		0x164290
+#define _TGL_TBTPLL_CFGCR1		0x1642A0
+#define TGL_DPLL_CFGCR1(pll)		_MMIO_PLL3(pll, _TGL_DPLL0_CFGCR1, \
+						   _TGL_DPLL1_CFGCR1, \
+						   _TGL_TBTPLL_CFGCR1)
+#define RKL_DPLL_CFGCR1(pll)		_MMIO_PLL(pll, _TGL_DPLL0_CFGCR1, \
+						  _TGL_DPLL1_CFGCR1)
+
+#define _DG1_DPLL2_CFGCR0		0x16C284
+#define _DG1_DPLL3_CFGCR0		0x16C28C
+#define DG1_DPLL_CFGCR0(pll)		_MMIO_PLL3(pll, _TGL_DPLL0_CFGCR0, \
+						   _TGL_DPLL1_CFGCR0, \
+						   _DG1_DPLL2_CFGCR0, \
+						   _DG1_DPLL3_CFGCR0)
+
+#define _DG1_DPLL2_CFGCR1               0x16C288
+#define _DG1_DPLL3_CFGCR1               0x16C290
+#define DG1_DPLL_CFGCR1(pll)            _MMIO_PLL3(pll, _TGL_DPLL0_CFGCR1, \
+						   _TGL_DPLL1_CFGCR1, \
+						   _DG1_DPLL2_CFGCR1, \
+						   _DG1_DPLL3_CFGCR1)
+
+/* For ADL-S DPLL4_CFGCR0/1 are used to control DPLL2 */
+#define _ADLS_DPLL3_CFGCR0		0x1642C0
+#define _ADLS_DPLL4_CFGCR0		0x164294
+#define ADLS_DPLL_CFGCR0(pll)		_MMIO_PLL3(pll, _TGL_DPLL0_CFGCR0, \
+						   _TGL_DPLL1_CFGCR0, \
+						   _ADLS_DPLL4_CFGCR0, \
+						   _ADLS_DPLL3_CFGCR0)
+
+#define _ADLS_DPLL3_CFGCR1		0x1642C4
+#define _ADLS_DPLL4_CFGCR1		0x164298
+#define ADLS_DPLL_CFGCR1(pll)		_MMIO_PLL3(pll, _TGL_DPLL0_CFGCR1, \
+						   _TGL_DPLL1_CFGCR1, \
+						   _ADLS_DPLL4_CFGCR1, \
+						   _ADLS_DPLL3_CFGCR1)
+
+#define _DKL_PHY1_BASE			0x168000
+#define _DKL_PHY2_BASE			0x169000
+#define _DKL_PHY3_BASE			0x16A000
+#define _DKL_PHY4_BASE			0x16B000
+#define _DKL_PHY5_BASE			0x16C000
+#define _DKL_PHY6_BASE			0x16D000
+
+#define DKL_REG_TC_PORT(__reg) \
+	(TC_PORT_1 + ((__reg).reg - _DKL_PHY1_BASE) / (_DKL_PHY2_BASE - _DKL_PHY1_BASE))
+
+/* DEKEL PHY MMIO Address = Phy base + (internal address & ~index_mask) */
+#define _DKL_PCS_DW5			0x14
+#define DKL_PCS_DW5(tc_port)		_MMIO(_PORT(tc_port, _DKL_PHY1_BASE, \
+						    _DKL_PHY2_BASE) + \
+						    _DKL_PCS_DW5)
+#define   DKL_PCS_DW5_CORE_SOFTRESET	REG_BIT(11)
+
+#define _DKL_PLL_DIV0			0x200
+#define   DKL_PLL_DIV0_AFC_STARTUP_MASK	REG_GENMASK(27, 25)
+#define   DKL_PLL_DIV0_AFC_STARTUP(val)	REG_FIELD_PREP(DKL_PLL_DIV0_AFC_STARTUP_MASK, (val))
+#define   DKL_PLL_DIV0_INTEG_COEFF(x)	((x) << 16)
+#define   DKL_PLL_DIV0_INTEG_COEFF_MASK	(0x1F << 16)
+#define   DKL_PLL_DIV0_PROP_COEFF(x)	((x) << 12)
+#define   DKL_PLL_DIV0_PROP_COEFF_MASK	(0xF << 12)
+#define   DKL_PLL_DIV0_FBPREDIV_SHIFT   (8)
+#define   DKL_PLL_DIV0_FBPREDIV(x)	((x) << DKL_PLL_DIV0_FBPREDIV_SHIFT)
+#define   DKL_PLL_DIV0_FBPREDIV_MASK	(0xF << DKL_PLL_DIV0_FBPREDIV_SHIFT)
+#define   DKL_PLL_DIV0_FBDIV_INT(x)	((x) << 0)
+#define   DKL_PLL_DIV0_FBDIV_INT_MASK	(0xFF << 0)
+#define   DKL_PLL_DIV0_MASK		(DKL_PLL_DIV0_INTEG_COEFF_MASK | \
+					 DKL_PLL_DIV0_PROP_COEFF_MASK | \
+					 DKL_PLL_DIV0_FBPREDIV_MASK | \
+					 DKL_PLL_DIV0_FBDIV_INT_MASK)
+#define DKL_PLL_DIV0(tc_port)		_MMIO(_PORT(tc_port, _DKL_PHY1_BASE, \
+						    _DKL_PHY2_BASE) + \
+						    _DKL_PLL_DIV0)
+
+#define _DKL_PLL_DIV1				0x204
+#define   DKL_PLL_DIV1_IREF_TRIM(x)		((x) << 16)
+#define   DKL_PLL_DIV1_IREF_TRIM_MASK		(0x1F << 16)
+#define   DKL_PLL_DIV1_TDC_TARGET_CNT(x)	((x) << 0)
+#define   DKL_PLL_DIV1_TDC_TARGET_CNT_MASK	(0xFF << 0)
+#define DKL_PLL_DIV1(tc_port)		_MMIO(_PORT(tc_port, _DKL_PHY1_BASE, \
+						    _DKL_PHY2_BASE) + \
+						    _DKL_PLL_DIV1)
+
+#define _DKL_PLL_SSC				0x210
+#define   DKL_PLL_SSC_IREF_NDIV_RATIO(x)	((x) << 29)
+#define   DKL_PLL_SSC_IREF_NDIV_RATIO_MASK	(0x7 << 29)
+#define   DKL_PLL_SSC_STEP_LEN(x)		((x) << 16)
+#define   DKL_PLL_SSC_STEP_LEN_MASK		(0xFF << 16)
+#define   DKL_PLL_SSC_STEP_NUM(x)		((x) << 11)
+#define   DKL_PLL_SSC_STEP_NUM_MASK		(0x7 << 11)
+#define   DKL_PLL_SSC_EN			(1 << 9)
+#define DKL_PLL_SSC(tc_port)		_MMIO(_PORT(tc_port, _DKL_PHY1_BASE, \
+						    _DKL_PHY2_BASE) + \
+						    _DKL_PLL_SSC)
+
+#define _DKL_PLL_BIAS			0x214
+#define   DKL_PLL_BIAS_FRAC_EN_H	(1 << 30)
+#define   DKL_PLL_BIAS_FBDIV_SHIFT	(8)
+#define   DKL_PLL_BIAS_FBDIV_FRAC(x)	((x) << DKL_PLL_BIAS_FBDIV_SHIFT)
+#define   DKL_PLL_BIAS_FBDIV_FRAC_MASK	(0x3FFFFF << DKL_PLL_BIAS_FBDIV_SHIFT)
+#define DKL_PLL_BIAS(tc_port)		_MMIO(_PORT(tc_port, _DKL_PHY1_BASE, \
+						    _DKL_PHY2_BASE) + \
+						    _DKL_PLL_BIAS)
+
+#define _DKL_PLL_TDC_COLDST_BIAS		0x218
+#define   DKL_PLL_TDC_SSC_STEP_SIZE(x)		((x) << 8)
+#define   DKL_PLL_TDC_SSC_STEP_SIZE_MASK	(0xFF << 8)
+#define   DKL_PLL_TDC_FEED_FWD_GAIN(x)		((x) << 0)
+#define   DKL_PLL_TDC_FEED_FWD_GAIN_MASK	(0xFF << 0)
+#define DKL_PLL_TDC_COLDST_BIAS(tc_port) _MMIO(_PORT(tc_port, \
+						     _DKL_PHY1_BASE, \
+						     _DKL_PHY2_BASE) + \
+						     _DKL_PLL_TDC_COLDST_BIAS)
+
+#define _DKL_REFCLKIN_CTL		0x12C
+/* Bits are the same as MG_REFCLKIN_CTL */
+#define DKL_REFCLKIN_CTL(tc_port)	_MMIO(_PORT(tc_port, \
+						    _DKL_PHY1_BASE, \
+						    _DKL_PHY2_BASE) + \
+					      _DKL_REFCLKIN_CTL)
+
+#define _DKL_CLKTOP2_HSCLKCTL		0xD4
+/* Bits are the same as MG_CLKTOP2_HSCLKCTL */
+#define DKL_CLKTOP2_HSCLKCTL(tc_port)	_MMIO(_PORT(tc_port, \
+						    _DKL_PHY1_BASE, \
+						    _DKL_PHY2_BASE) + \
+					      _DKL_CLKTOP2_HSCLKCTL)
+
+#define _DKL_CLKTOP2_CORECLKCTL1		0xD8
+/* Bits are the same as MG_CLKTOP2_CORECLKCTL1 */
+#define DKL_CLKTOP2_CORECLKCTL1(tc_port)	_MMIO(_PORT(tc_port, \
+							    _DKL_PHY1_BASE, \
+							    _DKL_PHY2_BASE) + \
+						      _DKL_CLKTOP2_CORECLKCTL1)
+
+#define _DKL_TX_DPCNTL0				0x2C0
+#define  DKL_TX_PRESHOOT_COEFF(x)			((x) << 13)
+#define  DKL_TX_PRESHOOT_COEFF_MASK			(0x1f << 13)
+#define  DKL_TX_DE_EMPHASIS_COEFF(x)		((x) << 8)
+#define  DKL_TX_DE_EMPAHSIS_COEFF_MASK		(0x1f << 8)
+#define  DKL_TX_VSWING_CONTROL(x)			((x) << 0)
+#define  DKL_TX_VSWING_CONTROL_MASK			(0x7 << 0)
+#define DKL_TX_DPCNTL0(tc_port) _MMIO(_PORT(tc_port, \
+						     _DKL_PHY1_BASE, \
+						     _DKL_PHY2_BASE) + \
+						     _DKL_TX_DPCNTL0)
+
+#define _DKL_TX_DPCNTL1				0x2C4
+/* Bits are the same as DKL_TX_DPCNTRL0 */
+#define DKL_TX_DPCNTL1(tc_port) _MMIO(_PORT(tc_port, \
+						     _DKL_PHY1_BASE, \
+						     _DKL_PHY2_BASE) + \
+						     _DKL_TX_DPCNTL1)
+
+#define _DKL_TX_DPCNTL2					0x2C8
+#define  DKL_TX_DP20BITMODE				REG_BIT(2)
+#define  DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX1_MASK	REG_GENMASK(4, 3)
+#define  DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX1(val)	REG_FIELD_PREP(DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX1_MASK, (val))
+#define  DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX2_MASK	REG_GENMASK(6, 5)
+#define  DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX2(val)	REG_FIELD_PREP(DKL_TX_DPCNTL2_CFG_LOADGENSELECT_TX2_MASK, (val))
+#define DKL_TX_DPCNTL2(tc_port) _MMIO(_PORT(tc_port, \
+						     _DKL_PHY1_BASE, \
+						     _DKL_PHY2_BASE) + \
+						     _DKL_TX_DPCNTL2)
+
+#define _DKL_TX_FW_CALIB				0x2F8
+#define  DKL_TX_CFG_DISABLE_WAIT_INIT			(1 << 7)
+#define DKL_TX_FW_CALIB(tc_port) _MMIO(_PORT(tc_port, \
+						     _DKL_PHY1_BASE, \
+						     _DKL_PHY2_BASE) + \
+						     _DKL_TX_FW_CALIB)
+
+#define _DKL_TX_PMD_LANE_SUS				0xD00
+#define DKL_TX_PMD_LANE_SUS(tc_port) _MMIO(_PORT(tc_port, \
+							  _DKL_PHY1_BASE, \
+							  _DKL_PHY2_BASE) + \
+							  _DKL_TX_PMD_LANE_SUS)
+
+#define _DKL_TX_DW17					0xDC4
+#define DKL_TX_DW17(tc_port) _MMIO(_PORT(tc_port, \
+						     _DKL_PHY1_BASE, \
+						     _DKL_PHY2_BASE) + \
+						     _DKL_TX_DW17)
+
+#define _DKL_TX_DW18					0xDC8
+#define DKL_TX_DW18(tc_port) _MMIO(_PORT(tc_port, \
+						     _DKL_PHY1_BASE, \
+						     _DKL_PHY2_BASE) + \
+						     _DKL_TX_DW18)
+
+#define _DKL_DP_MODE					0xA0
+#define DKL_DP_MODE(tc_port) _MMIO(_PORT(tc_port, \
+						     _DKL_PHY1_BASE, \
+						     _DKL_PHY2_BASE) + \
+						     _DKL_DP_MODE)
+
+#define _DKL_CMN_UC_DW27			0x36C
+#define  DKL_CMN_UC_DW27_UC_HEALTH		(0x1 << 15)
+#define DKL_CMN_UC_DW_27(tc_port)		_MMIO(_PORT(tc_port, \
+							    _DKL_PHY1_BASE, \
+							    _DKL_PHY2_BASE) + \
+							    _DKL_CMN_UC_DW27)
+
+/*
+ * Each Dekel PHY is addressed through a 4KB aperture. Each PHY has more than
+ * 4KB of register space, so a separate index is programmed in HIP_INDEX_REG0
+ * or HIP_INDEX_REG1, based on the port number, to set the upper 2 address
+ * bits that point the 4KB window into the full PHY register space.
+ */
+#define _HIP_INDEX_REG0			0x1010A0
+#define _HIP_INDEX_REG1			0x1010A4
+#define HIP_INDEX_REG(tc_port)		_MMIO((tc_port) < 4 ? _HIP_INDEX_REG0 \
+					      : _HIP_INDEX_REG1)
+#define _HIP_INDEX_SHIFT(tc_port)	(8 * ((tc_port) % 4))
+#define HIP_INDEX_VAL(tc_port, val)	((val) << _HIP_INDEX_SHIFT(tc_port))
+
+/* BXT display engine PLL */
+#define BXT_DE_PLL_CTL			_MMIO(0x6d000)
+#define   BXT_DE_PLL_RATIO(x)		(x)	/* {60,65,100} * 19.2MHz */
+#define   BXT_DE_PLL_RATIO_MASK		0xff
+
+#define BXT_DE_PLL_ENABLE		_MMIO(0x46070)
+#define   BXT_DE_PLL_PLL_ENABLE		(1 << 31)
+#define   BXT_DE_PLL_LOCK		(1 << 30)
+#define   BXT_DE_PLL_FREQ_REQ		(1 << 23)
+#define   BXT_DE_PLL_FREQ_REQ_ACK	(1 << 22)
+#define   ICL_CDCLK_PLL_RATIO(x)	(x)
+#define   ICL_CDCLK_PLL_RATIO_MASK	0xff
+
+/* GEN9 DC */
+#define DC_STATE_EN			_MMIO(0x45504)
+#define  DC_STATE_DISABLE		0
+#define  DC_STATE_EN_DC3CO		REG_BIT(30)
+#define  DC_STATE_DC3CO_STATUS		REG_BIT(29)
+#define  DC_STATE_EN_UPTO_DC5		(1 << 0)
+#define  DC_STATE_EN_DC9		(1 << 3)
+#define  DC_STATE_EN_UPTO_DC6		(2 << 0)
+#define  DC_STATE_EN_UPTO_DC5_DC6_MASK   0x3
+
+#define  DC_STATE_DEBUG                  _MMIO(0x45520)
+#define  DC_STATE_DEBUG_MASK_CORES	(1 << 0)
+#define  DC_STATE_DEBUG_MASK_MEMORY_UP	(1 << 1)
+
+#define D_COMP_BDW			_MMIO(0x138144)
+
+/* Pipe WM_LINETIME - watermark line time */
+#define _WM_LINETIME_A		0x45270
+#define _WM_LINETIME_B		0x45274
+#define WM_LINETIME(pipe) _MMIO_PIPE(pipe, _WM_LINETIME_A, _WM_LINETIME_B)
+#define  HSW_LINETIME_MASK	REG_GENMASK(8, 0)
+#define  HSW_LINETIME(x)	REG_FIELD_PREP(HSW_LINETIME_MASK, (x))
+#define  HSW_IPS_LINETIME_MASK	REG_GENMASK(24, 16)
+#define  HSW_IPS_LINETIME(x)	REG_FIELD_PREP(HSW_IPS_LINETIME_MASK, (x))
+
+/* SFUSE_STRAP */
+#define SFUSE_STRAP			_MMIO(0xc2014)
+#define  SFUSE_STRAP_FUSE_LOCK		(1 << 13)
+#define  SFUSE_STRAP_RAW_FREQUENCY	(1 << 8)
+#define  SFUSE_STRAP_DISPLAY_DISABLED	(1 << 7)
+#define  SFUSE_STRAP_CRT_DISABLED	(1 << 6)
+#define  SFUSE_STRAP_DDIF_DETECTED	(1 << 3)
+#define  SFUSE_STRAP_DDIB_DETECTED	(1 << 2)
+#define  SFUSE_STRAP_DDIC_DETECTED	(1 << 1)
+#define  SFUSE_STRAP_DDID_DETECTED	(1 << 0)
+
+#define WM_MISC				_MMIO(0x45260)
+#define  WM_MISC_DATA_PARTITION_5_6	(1 << 0)
+
+#define WM_DBG				_MMIO(0x45280)
+#define  WM_DBG_DISALLOW_MULTIPLE_LP	(1 << 0)
+#define  WM_DBG_DISALLOW_MAXFIFO	(1 << 1)
+#define  WM_DBG_DISALLOW_SPRITE		(1 << 2)
+
+/* pipe CSC */
+#define _PIPE_A_CSC_COEFF_RY_GY	0x49010
+#define _PIPE_A_CSC_COEFF_BY	0x49014
+#define _PIPE_A_CSC_COEFF_RU_GU	0x49018
+#define _PIPE_A_CSC_COEFF_BU	0x4901c
+#define _PIPE_A_CSC_COEFF_RV_GV	0x49020
+#define _PIPE_A_CSC_COEFF_BV	0x49024
+
+#define _PIPE_A_CSC_MODE	0x49028
+#define  ICL_CSC_ENABLE			(1 << 31) /* icl+ */
+#define  ICL_OUTPUT_CSC_ENABLE		(1 << 30) /* icl+ */
+#define  CSC_BLACK_SCREEN_OFFSET	(1 << 2) /* ilk/snb */
+#define  CSC_POSITION_BEFORE_GAMMA	(1 << 1) /* pre-glk */
+#define  CSC_MODE_YUV_TO_RGB		(1 << 0) /* ilk/snb */
+
+#define _PIPE_A_CSC_PREOFF_HI	0x49030
+#define _PIPE_A_CSC_PREOFF_ME	0x49034
+#define _PIPE_A_CSC_PREOFF_LO	0x49038
+#define _PIPE_A_CSC_POSTOFF_HI	0x49040
+#define _PIPE_A_CSC_POSTOFF_ME	0x49044
+#define _PIPE_A_CSC_POSTOFF_LO	0x49048
+
+#define _PIPE_B_CSC_COEFF_RY_GY	0x49110
+#define _PIPE_B_CSC_COEFF_BY	0x49114
+#define _PIPE_B_CSC_COEFF_RU_GU	0x49118
+#define _PIPE_B_CSC_COEFF_BU	0x4911c
+#define _PIPE_B_CSC_COEFF_RV_GV	0x49120
+#define _PIPE_B_CSC_COEFF_BV	0x49124
+#define _PIPE_B_CSC_MODE	0x49128
+#define _PIPE_B_CSC_PREOFF_HI	0x49130
+#define _PIPE_B_CSC_PREOFF_ME	0x49134
+#define _PIPE_B_CSC_PREOFF_LO	0x49138
+#define _PIPE_B_CSC_POSTOFF_HI	0x49140
+#define _PIPE_B_CSC_POSTOFF_ME	0x49144
+#define _PIPE_B_CSC_POSTOFF_LO	0x49148
+
+#define PIPE_CSC_COEFF_RY_GY(pipe)	_MMIO_PIPE(pipe, _PIPE_A_CSC_COEFF_RY_GY, _PIPE_B_CSC_COEFF_RY_GY)
+#define PIPE_CSC_COEFF_BY(pipe)		_MMIO_PIPE(pipe, _PIPE_A_CSC_COEFF_BY, _PIPE_B_CSC_COEFF_BY)
+#define PIPE_CSC_COEFF_RU_GU(pipe)	_MMIO_PIPE(pipe, _PIPE_A_CSC_COEFF_RU_GU, _PIPE_B_CSC_COEFF_RU_GU)
+#define PIPE_CSC_COEFF_BU(pipe)		_MMIO_PIPE(pipe, _PIPE_A_CSC_COEFF_BU, _PIPE_B_CSC_COEFF_BU)
+#define PIPE_CSC_COEFF_RV_GV(pipe)	_MMIO_PIPE(pipe, _PIPE_A_CSC_COEFF_RV_GV, _PIPE_B_CSC_COEFF_RV_GV)
+#define PIPE_CSC_COEFF_BV(pipe)		_MMIO_PIPE(pipe, _PIPE_A_CSC_COEFF_BV, _PIPE_B_CSC_COEFF_BV)
+#define PIPE_CSC_MODE(pipe)		_MMIO_PIPE(pipe, _PIPE_A_CSC_MODE, _PIPE_B_CSC_MODE)
+#define PIPE_CSC_PREOFF_HI(pipe)	_MMIO_PIPE(pipe, _PIPE_A_CSC_PREOFF_HI, _PIPE_B_CSC_PREOFF_HI)
+#define PIPE_CSC_PREOFF_ME(pipe)	_MMIO_PIPE(pipe, _PIPE_A_CSC_PREOFF_ME, _PIPE_B_CSC_PREOFF_ME)
+#define PIPE_CSC_PREOFF_LO(pipe)	_MMIO_PIPE(pipe, _PIPE_A_CSC_PREOFF_LO, _PIPE_B_CSC_PREOFF_LO)
+#define PIPE_CSC_POSTOFF_HI(pipe)	_MMIO_PIPE(pipe, _PIPE_A_CSC_POSTOFF_HI, _PIPE_B_CSC_POSTOFF_HI)
+#define PIPE_CSC_POSTOFF_ME(pipe)	_MMIO_PIPE(pipe, _PIPE_A_CSC_POSTOFF_ME, _PIPE_B_CSC_POSTOFF_ME)
+#define PIPE_CSC_POSTOFF_LO(pipe)	_MMIO_PIPE(pipe, _PIPE_A_CSC_POSTOFF_LO, _PIPE_B_CSC_POSTOFF_LO)
+
+/* Pipe Output CSC */
+#define _PIPE_A_OUTPUT_CSC_COEFF_RY_GY	0x49050
+#define _PIPE_A_OUTPUT_CSC_COEFF_BY	0x49054
+#define _PIPE_A_OUTPUT_CSC_COEFF_RU_GU	0x49058
+#define _PIPE_A_OUTPUT_CSC_COEFF_BU	0x4905c
+#define _PIPE_A_OUTPUT_CSC_COEFF_RV_GV	0x49060
+#define _PIPE_A_OUTPUT_CSC_COEFF_BV	0x49064
+#define _PIPE_A_OUTPUT_CSC_PREOFF_HI	0x49068
+#define _PIPE_A_OUTPUT_CSC_PREOFF_ME	0x4906c
+#define _PIPE_A_OUTPUT_CSC_PREOFF_LO	0x49070
+#define _PIPE_A_OUTPUT_CSC_POSTOFF_HI	0x49074
+#define _PIPE_A_OUTPUT_CSC_POSTOFF_ME	0x49078
+#define _PIPE_A_OUTPUT_CSC_POSTOFF_LO	0x4907c
+
+#define _PIPE_B_OUTPUT_CSC_COEFF_RY_GY	0x49150
+#define _PIPE_B_OUTPUT_CSC_COEFF_BY	0x49154
+#define _PIPE_B_OUTPUT_CSC_COEFF_RU_GU	0x49158
+#define _PIPE_B_OUTPUT_CSC_COEFF_BU	0x4915c
+#define _PIPE_B_OUTPUT_CSC_COEFF_RV_GV	0x49160
+#define _PIPE_B_OUTPUT_CSC_COEFF_BV	0x49164
+#define _PIPE_B_OUTPUT_CSC_PREOFF_HI	0x49168
+#define _PIPE_B_OUTPUT_CSC_PREOFF_ME	0x4916c
+#define _PIPE_B_OUTPUT_CSC_PREOFF_LO	0x49170
+#define _PIPE_B_OUTPUT_CSC_POSTOFF_HI	0x49174
+#define _PIPE_B_OUTPUT_CSC_POSTOFF_ME	0x49178
+#define _PIPE_B_OUTPUT_CSC_POSTOFF_LO	0x4917c
+
+#define PIPE_CSC_OUTPUT_COEFF_RY_GY(pipe)	_MMIO_PIPE(pipe,\
+							   _PIPE_A_OUTPUT_CSC_COEFF_RY_GY,\
+							   _PIPE_B_OUTPUT_CSC_COEFF_RY_GY)
+#define PIPE_CSC_OUTPUT_COEFF_BY(pipe)		_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_COEFF_BY, \
+							   _PIPE_B_OUTPUT_CSC_COEFF_BY)
+#define PIPE_CSC_OUTPUT_COEFF_RU_GU(pipe)	_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_COEFF_RU_GU, \
+							   _PIPE_B_OUTPUT_CSC_COEFF_RU_GU)
+#define PIPE_CSC_OUTPUT_COEFF_BU(pipe)		_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_COEFF_BU, \
+							   _PIPE_B_OUTPUT_CSC_COEFF_BU)
+#define PIPE_CSC_OUTPUT_COEFF_RV_GV(pipe)	_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_COEFF_RV_GV, \
+							   _PIPE_B_OUTPUT_CSC_COEFF_RV_GV)
+#define PIPE_CSC_OUTPUT_COEFF_BV(pipe)		_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_COEFF_BV, \
+							   _PIPE_B_OUTPUT_CSC_COEFF_BV)
+#define PIPE_CSC_OUTPUT_PREOFF_HI(pipe)		_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_PREOFF_HI, \
+							   _PIPE_B_OUTPUT_CSC_PREOFF_HI)
+#define PIPE_CSC_OUTPUT_PREOFF_ME(pipe)		_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_PREOFF_ME, \
+							   _PIPE_B_OUTPUT_CSC_PREOFF_ME)
+#define PIPE_CSC_OUTPUT_PREOFF_LO(pipe)		_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_PREOFF_LO, \
+							   _PIPE_B_OUTPUT_CSC_PREOFF_LO)
+#define PIPE_CSC_OUTPUT_POSTOFF_HI(pipe)	_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_POSTOFF_HI, \
+							   _PIPE_B_OUTPUT_CSC_POSTOFF_HI)
+#define PIPE_CSC_OUTPUT_POSTOFF_ME(pipe)	_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_POSTOFF_ME, \
+							   _PIPE_B_OUTPUT_CSC_POSTOFF_ME)
+#define PIPE_CSC_OUTPUT_POSTOFF_LO(pipe)	_MMIO_PIPE(pipe, \
+							   _PIPE_A_OUTPUT_CSC_POSTOFF_LO, \
+							   _PIPE_B_OUTPUT_CSC_POSTOFF_LO)
+
+/* pipe degamma/gamma LUTs on IVB+ */
+#define _PAL_PREC_INDEX_A	0x4A400
+#define _PAL_PREC_INDEX_B	0x4AC00
+#define _PAL_PREC_INDEX_C	0x4B400
+#define   PAL_PREC_10_12_BIT		(0 << 31)
+#define   PAL_PREC_SPLIT_MODE		(1 << 31)
+#define   PAL_PREC_AUTO_INCREMENT	(1 << 15)
+#define   PAL_PREC_INDEX_VALUE_MASK	(0x3ff << 0)
+#define   PAL_PREC_INDEX_VALUE(x)	((x) << 0)
+#define _PAL_PREC_DATA_A	0x4A404
+#define _PAL_PREC_DATA_B	0x4AC04
+#define _PAL_PREC_DATA_C	0x4B404
+#define _PAL_PREC_GC_MAX_A	0x4A410
+#define _PAL_PREC_GC_MAX_B	0x4AC10
+#define _PAL_PREC_GC_MAX_C	0x4B410
+#define   PREC_PAL_DATA_RED_MASK	REG_GENMASK(29, 20)
+#define   PREC_PAL_DATA_GREEN_MASK	REG_GENMASK(19, 10)
+#define   PREC_PAL_DATA_BLUE_MASK	REG_GENMASK(9, 0)
+#define _PAL_PREC_EXT_GC_MAX_A	0x4A420
+#define _PAL_PREC_EXT_GC_MAX_B	0x4AC20
+#define _PAL_PREC_EXT_GC_MAX_C	0x4B420
+#define _PAL_PREC_EXT2_GC_MAX_A	0x4A430
+#define _PAL_PREC_EXT2_GC_MAX_B	0x4AC30
+#define _PAL_PREC_EXT2_GC_MAX_C	0x4B430
+
+#define PREC_PAL_INDEX(pipe)		_MMIO_PIPE(pipe, _PAL_PREC_INDEX_A, _PAL_PREC_INDEX_B)
+#define PREC_PAL_DATA(pipe)		_MMIO_PIPE(pipe, _PAL_PREC_DATA_A, _PAL_PREC_DATA_B)
+#define PREC_PAL_GC_MAX(pipe, i)	_MMIO(_PIPE(pipe, _PAL_PREC_GC_MAX_A, _PAL_PREC_GC_MAX_B) + (i) * 4)
+#define PREC_PAL_EXT_GC_MAX(pipe, i)	_MMIO(_PIPE(pipe, _PAL_PREC_EXT_GC_MAX_A, _PAL_PREC_EXT_GC_MAX_B) + (i) * 4)
+#define PREC_PAL_EXT2_GC_MAX(pipe, i)	_MMIO(_PIPE(pipe, _PAL_PREC_EXT2_GC_MAX_A, _PAL_PREC_EXT2_GC_MAX_B) + (i) * 4)
+
+#define _PRE_CSC_GAMC_INDEX_A	0x4A484
+#define _PRE_CSC_GAMC_INDEX_B	0x4AC84
+#define _PRE_CSC_GAMC_INDEX_C	0x4B484
+#define   PRE_CSC_GAMC_AUTO_INCREMENT	(1 << 10)
+#define _PRE_CSC_GAMC_DATA_A	0x4A488
+#define _PRE_CSC_GAMC_DATA_B	0x4AC88
+#define _PRE_CSC_GAMC_DATA_C	0x4B488
+
+#define PRE_CSC_GAMC_INDEX(pipe)	_MMIO_PIPE(pipe, _PRE_CSC_GAMC_INDEX_A, _PRE_CSC_GAMC_INDEX_B)
+#define PRE_CSC_GAMC_DATA(pipe)		_MMIO_PIPE(pipe, _PRE_CSC_GAMC_DATA_A, _PRE_CSC_GAMC_DATA_B)
+
+/* ICL Multi segmented gamma */
+#define _PAL_PREC_MULTI_SEG_INDEX_A	0x4A408
+#define _PAL_PREC_MULTI_SEG_INDEX_B	0x4AC08
+#define  PAL_PREC_MULTI_SEGMENT_AUTO_INCREMENT		REG_BIT(15)
+#define  PAL_PREC_MULTI_SEGMENT_INDEX_VALUE_MASK	REG_GENMASK(4, 0)
+
+#define _PAL_PREC_MULTI_SEG_DATA_A	0x4A40C
+#define _PAL_PREC_MULTI_SEG_DATA_B	0x4AC0C
+#define  PAL_PREC_MULTI_SEG_RED_LDW_MASK   REG_GENMASK(29, 24)
+#define  PAL_PREC_MULTI_SEG_RED_UDW_MASK   REG_GENMASK(29, 20)
+#define  PAL_PREC_MULTI_SEG_GREEN_LDW_MASK REG_GENMASK(19, 14)
+#define  PAL_PREC_MULTI_SEG_GREEN_UDW_MASK REG_GENMASK(19, 10)
+#define  PAL_PREC_MULTI_SEG_BLUE_LDW_MASK  REG_GENMASK(9, 4)
+#define  PAL_PREC_MULTI_SEG_BLUE_UDW_MASK  REG_GENMASK(9, 0)
+
+#define PREC_PAL_MULTI_SEG_INDEX(pipe)	_MMIO_PIPE(pipe, \
+					_PAL_PREC_MULTI_SEG_INDEX_A, \
+					_PAL_PREC_MULTI_SEG_INDEX_B)
+#define PREC_PAL_MULTI_SEG_DATA(pipe)	_MMIO_PIPE(pipe, \
+					_PAL_PREC_MULTI_SEG_DATA_A, \
+					_PAL_PREC_MULTI_SEG_DATA_B)
+
+#define _MMIO_PLANE_GAMC(plane, i, a, b)  _MMIO(_PIPE(plane, a, b) + (i) * 4)
+
+/* Plane CSC Registers */
+#define _PLANE_CSC_RY_GY_1_A	0x70210
+#define _PLANE_CSC_RY_GY_2_A	0x70310
+
+#define _PLANE_CSC_RY_GY_1_B	0x71210
+#define _PLANE_CSC_RY_GY_2_B	0x71310
+
+#define _PLANE_CSC_RY_GY_1(pipe)	_PIPE(pipe, _PLANE_CSC_RY_GY_1_A, \
+					      _PLANE_CSC_RY_GY_1_B)
+#define _PLANE_CSC_RY_GY_2(pipe)	_PIPE(pipe, _PLANE_CSC_RY_GY_2_A, \
+					      _PLANE_CSC_RY_GY_2_B)
+#define PLANE_CSC_COEFF(pipe, plane, index)	_MMIO_PLANE(plane, \
+							    _PLANE_CSC_RY_GY_1(pipe) +  (index) * 4, \
+							    _PLANE_CSC_RY_GY_2(pipe) + (index) * 4)
+
+#define _PLANE_CSC_PREOFF_HI_1_A		0x70228
+#define _PLANE_CSC_PREOFF_HI_2_A		0x70328
+
+#define _PLANE_CSC_PREOFF_HI_1_B		0x71228
+#define _PLANE_CSC_PREOFF_HI_2_B		0x71328
+
+#define _PLANE_CSC_PREOFF_HI_1(pipe)	_PIPE(pipe, _PLANE_CSC_PREOFF_HI_1_A, \
+					      _PLANE_CSC_PREOFF_HI_1_B)
+#define _PLANE_CSC_PREOFF_HI_2(pipe)	_PIPE(pipe, _PLANE_CSC_PREOFF_HI_2_A, \
+					      _PLANE_CSC_PREOFF_HI_2_B)
+#define PLANE_CSC_PREOFF(pipe, plane, index)	_MMIO_PLANE(plane, _PLANE_CSC_PREOFF_HI_1(pipe) + \
+							    (index) * 4, _PLANE_CSC_PREOFF_HI_2(pipe) + \
+							    (index) * 4)
+
+#define _PLANE_CSC_POSTOFF_HI_1_A		0x70234
+#define _PLANE_CSC_POSTOFF_HI_2_A		0x70334
+
+#define _PLANE_CSC_POSTOFF_HI_1_B		0x71234
+#define _PLANE_CSC_POSTOFF_HI_2_B		0x71334
+
+#define _PLANE_CSC_POSTOFF_HI_1(pipe)	_PIPE(pipe, _PLANE_CSC_POSTOFF_HI_1_A, \
+					      _PLANE_CSC_POSTOFF_HI_1_B)
+#define _PLANE_CSC_POSTOFF_HI_2(pipe)	_PIPE(pipe, _PLANE_CSC_POSTOFF_HI_2_A, \
+					      _PLANE_CSC_POSTOFF_HI_2_B)
+#define PLANE_CSC_POSTOFF(pipe, plane, index)	_MMIO_PLANE(plane, _PLANE_CSC_POSTOFF_HI_1(pipe) + \
+							    (index) * 4, _PLANE_CSC_POSTOFF_HI_2(pipe) + \
+							    (index) * 4)
+
+/* pipe CSC & degamma/gamma LUTs on CHV */
+#define _CGM_PIPE_A_CSC_COEFF01	(VLV_DISPLAY_BASE + 0x67900)
+#define _CGM_PIPE_A_CSC_COEFF23	(VLV_DISPLAY_BASE + 0x67904)
+#define _CGM_PIPE_A_CSC_COEFF45	(VLV_DISPLAY_BASE + 0x67908)
+#define _CGM_PIPE_A_CSC_COEFF67	(VLV_DISPLAY_BASE + 0x6790C)
+#define _CGM_PIPE_A_CSC_COEFF8	(VLV_DISPLAY_BASE + 0x67910)
+#define _CGM_PIPE_A_DEGAMMA	(VLV_DISPLAY_BASE + 0x66000)
+#define   CGM_PIPE_DEGAMMA_RED_MASK	REG_GENMASK(13, 0)
+#define   CGM_PIPE_DEGAMMA_GREEN_MASK	REG_GENMASK(29, 16)
+#define   CGM_PIPE_DEGAMMA_BLUE_MASK	REG_GENMASK(13, 0)
+#define _CGM_PIPE_A_GAMMA	(VLV_DISPLAY_BASE + 0x67000)
+#define   CGM_PIPE_GAMMA_RED_MASK	REG_GENMASK(9, 0)
+#define   CGM_PIPE_GAMMA_GREEN_MASK	REG_GENMASK(25, 16)
+#define   CGM_PIPE_GAMMA_BLUE_MASK	REG_GENMASK(9, 0)
+#define _CGM_PIPE_A_MODE	(VLV_DISPLAY_BASE + 0x67A00)
+#define   CGM_PIPE_MODE_GAMMA	(1 << 2)
+#define   CGM_PIPE_MODE_CSC	(1 << 1)
+#define   CGM_PIPE_MODE_DEGAMMA	(1 << 0)
+
+#define _CGM_PIPE_B_CSC_COEFF01	(VLV_DISPLAY_BASE + 0x69900)
+#define _CGM_PIPE_B_CSC_COEFF23	(VLV_DISPLAY_BASE + 0x69904)
+#define _CGM_PIPE_B_CSC_COEFF45	(VLV_DISPLAY_BASE + 0x69908)
+#define _CGM_PIPE_B_CSC_COEFF67	(VLV_DISPLAY_BASE + 0x6990C)
+#define _CGM_PIPE_B_CSC_COEFF8	(VLV_DISPLAY_BASE + 0x69910)
+#define _CGM_PIPE_B_DEGAMMA	(VLV_DISPLAY_BASE + 0x68000)
+#define _CGM_PIPE_B_GAMMA	(VLV_DISPLAY_BASE + 0x69000)
+#define _CGM_PIPE_B_MODE	(VLV_DISPLAY_BASE + 0x69A00)
+
+#define CGM_PIPE_CSC_COEFF01(pipe)	_MMIO_PIPE(pipe, _CGM_PIPE_A_CSC_COEFF01, _CGM_PIPE_B_CSC_COEFF01)
+#define CGM_PIPE_CSC_COEFF23(pipe)	_MMIO_PIPE(pipe, _CGM_PIPE_A_CSC_COEFF23, _CGM_PIPE_B_CSC_COEFF23)
+#define CGM_PIPE_CSC_COEFF45(pipe)	_MMIO_PIPE(pipe, _CGM_PIPE_A_CSC_COEFF45, _CGM_PIPE_B_CSC_COEFF45)
+#define CGM_PIPE_CSC_COEFF67(pipe)	_MMIO_PIPE(pipe, _CGM_PIPE_A_CSC_COEFF67, _CGM_PIPE_B_CSC_COEFF67)
+#define CGM_PIPE_CSC_COEFF8(pipe)	_MMIO_PIPE(pipe, _CGM_PIPE_A_CSC_COEFF8, _CGM_PIPE_B_CSC_COEFF8)
+#define CGM_PIPE_DEGAMMA(pipe, i, w)	_MMIO(_PIPE(pipe, _CGM_PIPE_A_DEGAMMA, _CGM_PIPE_B_DEGAMMA) + (i) * 8 + (w) * 4)
+#define CGM_PIPE_GAMMA(pipe, i, w)	_MMIO(_PIPE(pipe, _CGM_PIPE_A_GAMMA, _CGM_PIPE_B_GAMMA) + (i) * 8 + (w) * 4)
+#define CGM_PIPE_MODE(pipe)		_MMIO_PIPE(pipe, _CGM_PIPE_A_MODE, _CGM_PIPE_B_MODE)
+
+/* Gen4+ Timestamp and Pipe Frame time stamp registers */
+#define GEN4_TIMESTAMP		_MMIO(0x2358)
+#define ILK_TIMESTAMP_HI	_MMIO(0x70070)
+#define IVB_TIMESTAMP_CTR	_MMIO(0x44070)
+
+#define GEN9_TIMESTAMP_OVERRIDE				_MMIO(0x44074)
+#define  GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DIVIDER_SHIFT	0
+#define  GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DIVIDER_MASK	0x3ff
+#define  GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DENOMINATOR_SHIFT	12
+#define  GEN9_TIMESTAMP_OVERRIDE_US_COUNTER_DENOMINATOR_MASK	(0xf << 12)
+
+#define _PIPE_FRMTMSTMP_A		0x70048
+#define PIPE_FRMTMSTMP(pipe)		\
+			_MMIO_PIPE2(pipe, _PIPE_FRMTMSTMP_A)
+
+/* Display Stream Splitter Control */
+#define DSS_CTL1				_MMIO(0x67400)
+#define  SPLITTER_ENABLE			(1 << 31)
+#define  JOINER_ENABLE				(1 << 30)
+#define  DUAL_LINK_MODE_INTERLEAVE		(1 << 24)
+#define  DUAL_LINK_MODE_FRONTBACK		(0 << 24)
+#define  OVERLAP_PIXELS_MASK			(0xf << 16)
+#define  OVERLAP_PIXELS(pixels)			((pixels) << 16)
+#define  LEFT_DL_BUF_TARGET_DEPTH_MASK		(0xfff << 0)
+#define  LEFT_DL_BUF_TARGET_DEPTH(pixels)	((pixels) << 0)
+#define  MAX_DL_BUFFER_TARGET_DEPTH		0x5a0
+
+#define DSS_CTL2				_MMIO(0x67404)
+#define  LEFT_BRANCH_VDSC_ENABLE		(1 << 31)
+#define  RIGHT_BRANCH_VDSC_ENABLE		(1 << 15)
+#define  RIGHT_DL_BUF_TARGET_DEPTH_MASK		(0xfff << 0)
+#define  RIGHT_DL_BUF_TARGET_DEPTH(pixels)	((pixels) << 0)
+
+#define _ICL_PIPE_DSS_CTL1_PB			0x78200
+#define _ICL_PIPE_DSS_CTL1_PC			0x78400
+#define ICL_PIPE_DSS_CTL1(pipe)			_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_PIPE_DSS_CTL1_PB, \
+							   _ICL_PIPE_DSS_CTL1_PC)
+#define  BIG_JOINER_ENABLE			(1 << 29)
+#define  MASTER_BIG_JOINER_ENABLE		(1 << 28)
+#define  VGA_CENTERING_ENABLE			(1 << 27)
+#define  SPLITTER_CONFIGURATION_MASK		REG_GENMASK(26, 25)
+#define  SPLITTER_CONFIGURATION_2_SEGMENT	REG_FIELD_PREP(SPLITTER_CONFIGURATION_MASK, 0)
+#define  SPLITTER_CONFIGURATION_4_SEGMENT	REG_FIELD_PREP(SPLITTER_CONFIGURATION_MASK, 1)
+#define  UNCOMPRESSED_JOINER_MASTER		(1 << 21)
+#define  UNCOMPRESSED_JOINER_SLAVE		(1 << 20)
+
+#define _ICL_PIPE_DSS_CTL2_PB			0x78204
+#define _ICL_PIPE_DSS_CTL2_PC			0x78404
+#define ICL_PIPE_DSS_CTL2(pipe)			_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_PIPE_DSS_CTL2_PB, \
+							   _ICL_PIPE_DSS_CTL2_PC)
+
+#define GEN12_GSMBASE			_MMIO(0x108100)
+#define GEN12_DSMBASE			_MMIO(0x1080C0)
+
+#define XEHP_CLOCK_GATE_DIS		_MMIO(0x101014)
+#define   SGSI_SIDECLK_DIS		REG_BIT(17)
+#define   SGGI_DIS			REG_BIT(15)
+#define   SGR_DIS			REG_BIT(13)
+
+#define _ICL_PHY_MISC_A		0x64C00
+#define _ICL_PHY_MISC_B		0x64C04
+#define _DG2_PHY_MISC_TC1	0x64C14 /* TC1="PHY E" but offset as if "PHY F" */
+#define ICL_PHY_MISC(port)	_MMIO_PORT(port, _ICL_PHY_MISC_A, _ICL_PHY_MISC_B)
+#define DG2_PHY_MISC(port)	((port) == PHY_E ? _MMIO(_DG2_PHY_MISC_TC1) : \
+				 ICL_PHY_MISC(port))
+#define  ICL_PHY_MISC_MUX_DDID			(1 << 28)
+#define  ICL_PHY_MISC_DE_IO_COMP_PWR_DOWN	(1 << 23)
+#define  DG2_PHY_DP_TX_ACK_MASK			REG_GENMASK(23, 20)
+
+/* Icelake Display Stream Compression Registers */
+#define DSCA_PICTURE_PARAMETER_SET_0		_MMIO(0x6B200)
+#define DSCC_PICTURE_PARAMETER_SET_0		_MMIO(0x6BA00)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_0_PB	0x78270
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_0_PB	0x78370
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_0_PC	0x78470
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_0_PC	0x78570
+#define ICL_DSC0_PICTURE_PARAMETER_SET_0(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_0_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_0_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_0(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_0_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_0_PC)
+#define  DSC_ALT_ICH_SEL		(1 << 20)
+#define  DSC_VBR_ENABLE			(1 << 19)
+#define  DSC_422_ENABLE			(1 << 18)
+#define  DSC_COLOR_SPACE_CONVERSION	(1 << 17)
+#define  DSC_BLOCK_PREDICTION		(1 << 16)
+#define  DSC_LINE_BUF_DEPTH_SHIFT	12
+#define  DSC_BPC_SHIFT			8
+#define  DSC_VER_MIN_SHIFT		4
+#define  DSC_VER_MAJ			(0x1 << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_1		_MMIO(0x6B204)
+#define DSCC_PICTURE_PARAMETER_SET_1		_MMIO(0x6BA04)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_1_PB	0x78274
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_1_PB	0x78374
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_1_PC	0x78474
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_1_PC	0x78574
+#define ICL_DSC0_PICTURE_PARAMETER_SET_1(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_1_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_1_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_1(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_1_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_1_PC)
+#define  DSC_BPP(bpp)				((bpp) << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_2		_MMIO(0x6B208)
+#define DSCC_PICTURE_PARAMETER_SET_2		_MMIO(0x6BA08)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_2_PB	0x78278
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_2_PB	0x78378
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_2_PC	0x78478
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_2_PC	0x78578
+#define ICL_DSC0_PICTURE_PARAMETER_SET_2(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_2_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_2_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_2(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+					    _ICL_DSC1_PICTURE_PARAMETER_SET_2_PB, \
+					    _ICL_DSC1_PICTURE_PARAMETER_SET_2_PC)
+#define  DSC_PIC_WIDTH(pic_width)	((pic_width) << 16)
+#define  DSC_PIC_HEIGHT(pic_height)	((pic_height) << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_3		_MMIO(0x6B20C)
+#define DSCC_PICTURE_PARAMETER_SET_3		_MMIO(0x6BA0C)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_3_PB	0x7827C
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_3_PB	0x7837C
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_3_PC	0x7847C
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_3_PC	0x7857C
+#define ICL_DSC0_PICTURE_PARAMETER_SET_3(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_3_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_3_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_3(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_3_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_3_PC)
+#define  DSC_SLICE_WIDTH(slice_width)   ((slice_width) << 16)
+#define  DSC_SLICE_HEIGHT(slice_height) ((slice_height) << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_4		_MMIO(0x6B210)
+#define DSCC_PICTURE_PARAMETER_SET_4		_MMIO(0x6BA10)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_4_PB	0x78280
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_4_PB	0x78380
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_4_PC	0x78480
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_4_PC	0x78580
+#define ICL_DSC0_PICTURE_PARAMETER_SET_4(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_4_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_4_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_4(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_4_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_4_PC)
+#define  DSC_INITIAL_DEC_DELAY(dec_delay)       ((dec_delay) << 16)
+#define  DSC_INITIAL_XMIT_DELAY(xmit_delay)     ((xmit_delay) << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_5		_MMIO(0x6B214)
+#define DSCC_PICTURE_PARAMETER_SET_5		_MMIO(0x6BA14)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_5_PB	0x78284
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_5_PB	0x78384
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_5_PC	0x78484
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_5_PC	0x78584
+#define ICL_DSC0_PICTURE_PARAMETER_SET_5(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_5_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_5_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_5(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_5_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_5_PC)
+#define  DSC_SCALE_DEC_INT(scale_dec)	((scale_dec) << 16)
+#define  DSC_SCALE_INC_INT(scale_inc)		((scale_inc) << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_6		_MMIO(0x6B218)
+#define DSCC_PICTURE_PARAMETER_SET_6		_MMIO(0x6BA18)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_6_PB	0x78288
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_6_PB	0x78388
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_6_PC	0x78488
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_6_PC	0x78588
+#define ICL_DSC0_PICTURE_PARAMETER_SET_6(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_6_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_6_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_6(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_6_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_6_PC)
+#define  DSC_FLATNESS_MAX_QP(max_qp)		((max_qp) << 24)
+#define  DSC_FLATNESS_MIN_QP(min_qp)		((min_qp) << 16)
+#define  DSC_FIRST_LINE_BPG_OFFSET(offset)	((offset) << 8)
+#define  DSC_INITIAL_SCALE_VALUE(value)		((value) << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_7		_MMIO(0x6B21C)
+#define DSCC_PICTURE_PARAMETER_SET_7		_MMIO(0x6BA1C)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_7_PB	0x7828C
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_7_PB	0x7838C
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_7_PC	0x7848C
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_7_PC	0x7858C
+#define ICL_DSC0_PICTURE_PARAMETER_SET_7(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							    _ICL_DSC0_PICTURE_PARAMETER_SET_7_PB, \
+							    _ICL_DSC0_PICTURE_PARAMETER_SET_7_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_7(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							    _ICL_DSC1_PICTURE_PARAMETER_SET_7_PB, \
+							    _ICL_DSC1_PICTURE_PARAMETER_SET_7_PC)
+#define  DSC_NFL_BPG_OFFSET(bpg_offset)		((bpg_offset) << 16)
+#define  DSC_SLICE_BPG_OFFSET(bpg_offset)	((bpg_offset) << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_8		_MMIO(0x6B220)
+#define DSCC_PICTURE_PARAMETER_SET_8		_MMIO(0x6BA20)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_8_PB	0x78290
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_8_PB	0x78390
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_8_PC	0x78490
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_8_PC	0x78590
+#define ICL_DSC0_PICTURE_PARAMETER_SET_8(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_8_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_8_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_8(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_8_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_8_PC)
+#define  DSC_INITIAL_OFFSET(initial_offset)		((initial_offset) << 16)
+#define  DSC_FINAL_OFFSET(final_offset)			((final_offset) << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_9		_MMIO(0x6B224)
+#define DSCC_PICTURE_PARAMETER_SET_9		_MMIO(0x6BA24)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_9_PB	0x78294
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_9_PB	0x78394
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_9_PC	0x78494
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_9_PC	0x78594
+#define ICL_DSC0_PICTURE_PARAMETER_SET_9(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_9_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_9_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_9(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_9_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_9_PC)
+#define  DSC_RC_EDGE_FACTOR(rc_edge_fact)	((rc_edge_fact) << 16)
+#define  DSC_RC_MODEL_SIZE(rc_model_size)	((rc_model_size) << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_10		_MMIO(0x6B228)
+#define DSCC_PICTURE_PARAMETER_SET_10		_MMIO(0x6BA28)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_10_PB	0x78298
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_10_PB	0x78398
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_10_PC	0x78498
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_10_PC	0x78598
+#define ICL_DSC0_PICTURE_PARAMETER_SET_10(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_10_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_10_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_10(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_10_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_10_PC)
+#define  DSC_RC_TARGET_OFF_LOW(rc_tgt_off_low)		((rc_tgt_off_low) << 20)
+#define  DSC_RC_TARGET_OFF_HIGH(rc_tgt_off_high)	((rc_tgt_off_high) << 16)
+#define  DSC_RC_QUANT_INC_LIMIT1(lim)			((lim) << 8)
+#define  DSC_RC_QUANT_INC_LIMIT0(lim)			((lim) << 0)
+
+#define DSCA_PICTURE_PARAMETER_SET_11		_MMIO(0x6B22C)
+#define DSCC_PICTURE_PARAMETER_SET_11		_MMIO(0x6BA2C)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_11_PB	0x7829C
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_11_PB	0x7839C
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_11_PC	0x7849C
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_11_PC	0x7859C
+#define ICL_DSC0_PICTURE_PARAMETER_SET_11(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_11_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_11_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_11(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_11_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_11_PC)
+
+#define DSCA_PICTURE_PARAMETER_SET_12		_MMIO(0x6B260)
+#define DSCC_PICTURE_PARAMETER_SET_12		_MMIO(0x6BA60)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_12_PB	0x782A0
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_12_PB	0x783A0
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_12_PC	0x784A0
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_12_PC	0x785A0
+#define ICL_DSC0_PICTURE_PARAMETER_SET_12(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_12_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_12_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_12(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_12_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_12_PC)
+
+#define DSCA_PICTURE_PARAMETER_SET_13		_MMIO(0x6B264)
+#define DSCC_PICTURE_PARAMETER_SET_13		_MMIO(0x6BA64)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_13_PB	0x782A4
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_13_PB	0x783A4
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_13_PC	0x784A4
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_13_PC	0x785A4
+#define ICL_DSC0_PICTURE_PARAMETER_SET_13(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_13_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_13_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_13(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_13_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_13_PC)
+
+#define DSCA_PICTURE_PARAMETER_SET_14		_MMIO(0x6B268)
+#define DSCC_PICTURE_PARAMETER_SET_14		_MMIO(0x6BA68)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_14_PB	0x782A8
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_14_PB	0x783A8
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_14_PC	0x784A8
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_14_PC	0x785A8
+#define ICL_DSC0_PICTURE_PARAMETER_SET_14(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_14_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_14_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_14(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_14_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_14_PC)
+
+#define DSCA_PICTURE_PARAMETER_SET_15		_MMIO(0x6B26C)
+#define DSCC_PICTURE_PARAMETER_SET_15		_MMIO(0x6BA6C)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_15_PB	0x782AC
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_15_PB	0x783AC
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_15_PC	0x784AC
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_15_PC	0x785AC
+#define ICL_DSC0_PICTURE_PARAMETER_SET_15(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_15_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_15_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_15(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_15_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_15_PC)
+
+#define DSCA_PICTURE_PARAMETER_SET_16		_MMIO(0x6B270)
+#define DSCC_PICTURE_PARAMETER_SET_16		_MMIO(0x6BA70)
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_16_PB	0x782B0
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_16_PB	0x783B0
+#define _ICL_DSC0_PICTURE_PARAMETER_SET_16_PC	0x784B0
+#define _ICL_DSC1_PICTURE_PARAMETER_SET_16_PC	0x785B0
+#define ICL_DSC0_PICTURE_PARAMETER_SET_16(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_16_PB, \
+							   _ICL_DSC0_PICTURE_PARAMETER_SET_16_PC)
+#define ICL_DSC1_PICTURE_PARAMETER_SET_16(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_16_PB, \
+							   _ICL_DSC1_PICTURE_PARAMETER_SET_16_PC)
+#define  DSC_SLICE_ROW_PER_FRAME(slice_row_per_frame)	((slice_row_per_frame) << 20)
+#define  DSC_SLICE_PER_LINE(slice_per_line)		((slice_per_line) << 16)
+#define  DSC_SLICE_CHUNK_SIZE(slice_chunk_size)		((slice_chunk_size) << 0)
+
+/* Icelake Rate Control Buffer Threshold Registers */
+#define DSCA_RC_BUF_THRESH_0			_MMIO(0x6B230)
+#define DSCA_RC_BUF_THRESH_0_UDW		_MMIO(0x6B230 + 4)
+#define DSCC_RC_BUF_THRESH_0			_MMIO(0x6BA30)
+#define DSCC_RC_BUF_THRESH_0_UDW		_MMIO(0x6BA30 + 4)
+#define _ICL_DSC0_RC_BUF_THRESH_0_PB		(0x78254)
+#define _ICL_DSC0_RC_BUF_THRESH_0_UDW_PB	(0x78254 + 4)
+#define _ICL_DSC1_RC_BUF_THRESH_0_PB		(0x78354)
+#define _ICL_DSC1_RC_BUF_THRESH_0_UDW_PB	(0x78354 + 4)
+#define _ICL_DSC0_RC_BUF_THRESH_0_PC		(0x78454)
+#define _ICL_DSC0_RC_BUF_THRESH_0_UDW_PC	(0x78454 + 4)
+#define _ICL_DSC1_RC_BUF_THRESH_0_PC		(0x78554)
+#define _ICL_DSC1_RC_BUF_THRESH_0_UDW_PC	(0x78554 + 4)
+#define ICL_DSC0_RC_BUF_THRESH_0(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+						_ICL_DSC0_RC_BUF_THRESH_0_PB, \
+						_ICL_DSC0_RC_BUF_THRESH_0_PC)
+#define ICL_DSC0_RC_BUF_THRESH_0_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+						_ICL_DSC0_RC_BUF_THRESH_0_UDW_PB, \
+						_ICL_DSC0_RC_BUF_THRESH_0_UDW_PC)
+#define ICL_DSC1_RC_BUF_THRESH_0(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+						_ICL_DSC1_RC_BUF_THRESH_0_PB, \
+						_ICL_DSC1_RC_BUF_THRESH_0_PC)
+#define ICL_DSC1_RC_BUF_THRESH_0_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+						_ICL_DSC1_RC_BUF_THRESH_0_UDW_PB, \
+						_ICL_DSC1_RC_BUF_THRESH_0_UDW_PC)
+
+#define DSCA_RC_BUF_THRESH_1			_MMIO(0x6B238)
+#define DSCA_RC_BUF_THRESH_1_UDW		_MMIO(0x6B238 + 4)
+#define DSCC_RC_BUF_THRESH_1			_MMIO(0x6BA38)
+#define DSCC_RC_BUF_THRESH_1_UDW		_MMIO(0x6BA38 + 4)
+#define _ICL_DSC0_RC_BUF_THRESH_1_PB		(0x7825C)
+#define _ICL_DSC0_RC_BUF_THRESH_1_UDW_PB	(0x7825C + 4)
+#define _ICL_DSC1_RC_BUF_THRESH_1_PB		(0x7835C)
+#define _ICL_DSC1_RC_BUF_THRESH_1_UDW_PB	(0x7835C + 4)
+#define _ICL_DSC0_RC_BUF_THRESH_1_PC		(0x7845C)
+#define _ICL_DSC0_RC_BUF_THRESH_1_UDW_PC	(0x7845C + 4)
+#define _ICL_DSC1_RC_BUF_THRESH_1_PC		(0x7855C)
+#define _ICL_DSC1_RC_BUF_THRESH_1_UDW_PC	(0x7855C + 4)
+#define ICL_DSC0_RC_BUF_THRESH_1(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+						_ICL_DSC0_RC_BUF_THRESH_1_PB, \
+						_ICL_DSC0_RC_BUF_THRESH_1_PC)
+#define ICL_DSC0_RC_BUF_THRESH_1_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+						_ICL_DSC0_RC_BUF_THRESH_1_UDW_PB, \
+						_ICL_DSC0_RC_BUF_THRESH_1_UDW_PC)
+#define ICL_DSC1_RC_BUF_THRESH_1(pipe)		_MMIO_PIPE((pipe) - PIPE_B, \
+						_ICL_DSC1_RC_BUF_THRESH_1_PB, \
+						_ICL_DSC1_RC_BUF_THRESH_1_PC)
+#define ICL_DSC1_RC_BUF_THRESH_1_UDW(pipe)	_MMIO_PIPE((pipe) - PIPE_B, \
+						_ICL_DSC1_RC_BUF_THRESH_1_UDW_PB, \
+						_ICL_DSC1_RC_BUF_THRESH_1_UDW_PC)
+
+#define PORT_TX_DFLEXDPSP(fia)			_MMIO_FIA((fia), 0x008A0)
+#define   MODULAR_FIA_MASK			(1 << 4)
+#define   TC_LIVE_STATE_TBT(idx)		(1 << ((idx) * 8 + 6))
+#define   TC_LIVE_STATE_TC(idx)			(1 << ((idx) * 8 + 5))
+#define   DP_LANE_ASSIGNMENT_SHIFT(idx)		((idx) * 8)
+#define   DP_LANE_ASSIGNMENT_MASK(idx)		(0xf << ((idx) * 8))
+#define   DP_LANE_ASSIGNMENT(idx, x)		((x) << ((idx) * 8))
+
+#define PORT_TX_DFLEXDPPMS(fia)			_MMIO_FIA((fia), 0x00890)
+#define   DP_PHY_MODE_STATUS_COMPLETED(idx)	(1 << (idx))
+
+#define PORT_TX_DFLEXDPCSSS(fia)		_MMIO_FIA((fia), 0x00894)
+#define   DP_PHY_MODE_STATUS_NOT_SAFE(idx)	(1 << (idx))
+
+#define PORT_TX_DFLEXPA1(fia)			_MMIO_FIA((fia), 0x00880)
+#define   DP_PIN_ASSIGNMENT_SHIFT(idx)		((idx) * 4)
+#define   DP_PIN_ASSIGNMENT_MASK(idx)		(0xf << ((idx) * 4))
+#define   DP_PIN_ASSIGNMENT(idx, x)		((x) << ((idx) * 4))
+
+#define _TCSS_DDI_STATUS_1			0x161500
+#define _TCSS_DDI_STATUS_2			0x161504
+#define TCSS_DDI_STATUS(tc)			_MMIO(_PICK_EVEN(tc, \
+								 _TCSS_DDI_STATUS_1, \
+								 _TCSS_DDI_STATUS_2))
+#define  TCSS_DDI_STATUS_READY			REG_BIT(2)
+#define  TCSS_DDI_STATUS_HPD_LIVE_STATUS_TBT	REG_BIT(1)
+#define  TCSS_DDI_STATUS_HPD_LIVE_STATUS_ALT	REG_BIT(0)
+
+#define PRIMARY_SPI_TRIGGER			_MMIO(0x102040)
+#define PRIMARY_SPI_ADDRESS			_MMIO(0x102080)
+#define PRIMARY_SPI_REGIONID			_MMIO(0x102084)
+#define SPI_STATIC_REGIONS			_MMIO(0x102090)
+#define   OPTIONROM_SPI_REGIONID_MASK		REG_GENMASK(7, 0)
+#define OROM_OFFSET				_MMIO(0x1020c0)
+#define   OROM_OFFSET_MASK			REG_GENMASK(20, 16)
+
+/* This register controls the Display State Buffer (DSB) engines. */
+#define _DSBSL_INSTANCE_BASE		0x70B00
+#define DSBSL_INSTANCE(pipe, id)	(_DSBSL_INSTANCE_BASE + \
+					 (pipe) * 0x1000 + (id) * 0x100)
+#define DSB_HEAD(pipe, id)		_MMIO(DSBSL_INSTANCE(pipe, id) + 0x0)
+#define DSB_TAIL(pipe, id)		_MMIO(DSBSL_INSTANCE(pipe, id) + 0x4)
+#define DSB_CTRL(pipe, id)		_MMIO(DSBSL_INSTANCE(pipe, id) + 0x8)
+#define   DSB_ENABLE			(1 << 31)
+#define   DSB_STATUS			(1 << 0)
+
+#define CLKREQ_POLICY			_MMIO(0x101038)
+#define  CLKREQ_POLICY_MEM_UP_OVRD	REG_BIT(1)
+
+#define CLKGATE_DIS_MISC			_MMIO(0x46534)
+#define  CLKGATE_DIS_MISC_DMASC_GATING_DIS	REG_BIT(21)
+
+#define GEN12_CULLBIT1			_MMIO(0x6100)
+#define GEN12_CULLBIT2			_MMIO(0x7030)
+#define GEN12_STATE_ACK_DEBUG		_MMIO(0x20BC)
+
+#define MTL_LATENCY_LP0_LP1		_MMIO(0x45780)
+#define MTL_LATENCY_LP2_LP3		_MMIO(0x45784)
+#define MTL_LATENCY_LP4_LP5		_MMIO(0x45788)
+#define  MTL_LATENCY_LEVEL_EVEN_MASK	REG_GENMASK(12, 0)
+#define  MTL_LATENCY_LEVEL_ODD_MASK	REG_GENMASK(28, 16)
+
+#define MTL_LATENCY_SAGV		_MMIO(0x4578b)
+#define   MTL_LATENCY_QCLK_SAGV		REG_GENMASK(12, 0)
+
+#define MTL_MEM_SS_INFO_GLOBAL			_MMIO(0x45700)
+#define   MTL_N_OF_ENABLED_QGV_POINTS_MASK	REG_GENMASK(11, 8)
+#define   MTL_N_OF_POPULATED_CH_MASK		REG_GENMASK(7, 4)
+#define   MTL_DDR_TYPE_MASK			REG_GENMASK(3, 0)
+
+#define MTL_MEM_SS_INFO_QGV_POINT_LOW(point)	 _MMIO(0x45710 + (point) * 2)
+#define   MTL_TRCD_MASK			REG_GENMASK(31, 24)
+#define   MTL_TRP_MASK			REG_GENMASK(23, 16)
+#define   MTL_DCLK_MASK			REG_GENMASK(15, 0)
+
+#define MTL_MEM_SS_INFO_QGV_POINT_HIGH(point)	 _MMIO(0x45714 + (point) * 2)
+#define   MTL_TRAS_MASK			REG_GENMASK(16, 8)
+#define   MTL_TRDPRE_MASK		REG_GENMASK(7, 0)
+
+#endif /* _I915_REG_H_ */
diff --git a/drivers/gpu/drm/i915/i915_reg_defs.h b/drivers/gpu/drm/i915/i915_reg_defs.h
new file mode 100644
index 000000000..8f486f776
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_reg_defs.h
@@ -0,0 +1,126 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __I915_REG_DEFS__
+#define __I915_REG_DEFS__
+
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+
+/**
+ * REG_BIT() - Prepare a u32 bit value
+ * @__n: 0-based bit number
+ *
+ * Local wrapper for BIT() to force u32, with compile time checks.
+ *
+ * @return: Value with bit @__n set.
+ */
+#define REG_BIT(__n)							\
+	((u32)(BIT(__n) +						\
+	       BUILD_BUG_ON_ZERO(__is_constexpr(__n) &&		\
+				 ((__n) < 0 || (__n) > 31))))
+
+/**
+ * REG_GENMASK() - Prepare a continuous u32 bitmask
+ * @__high: 0-based high bit
+ * @__low: 0-based low bit
+ *
+ * Local wrapper for GENMASK() to force u32, with compile time checks.
+ *
+ * @return: Continuous bitmask from @__high to @__low, inclusive.
+ */
+#define REG_GENMASK(__high, __low)					\
+	((u32)(GENMASK(__high, __low) +					\
+	       BUILD_BUG_ON_ZERO(__is_constexpr(__high) &&	\
+				 __is_constexpr(__low) &&		\
+				 ((__low) < 0 || (__high) > 31 || (__low) > (__high)))))
+
+/**
+ * REG_GENMASK64() - Prepare a continuous u64 bitmask
+ * @__high: 0-based high bit
+ * @__low: 0-based low bit
+ *
+ * Local wrapper for GENMASK_ULL() to force u64, with compile time checks.
+ *
+ * @return: Continuous bitmask from @__high to @__low, inclusive.
+ */
+#define REG_GENMASK64(__high, __low)					\
+	((u64)(GENMASK_ULL(__high, __low) +				\
+	       BUILD_BUG_ON_ZERO(__is_constexpr(__high) &&		\
+				 __is_constexpr(__low) &&		\
+				 ((__low) < 0 || (__high) > 63 || (__low) > (__high)))))
+
+/*
+ * Local integer constant expression version of is_power_of_2().
+ */
+#define IS_POWER_OF_2(__x)		((__x) && (((__x) & ((__x) - 1)) == 0))
+
+/**
+ * REG_FIELD_PREP() - Prepare a u32 bitfield value
+ * @__mask: shifted mask defining the field's length and position
+ * @__val: value to put in the field
+ *
+ * Local copy of FIELD_PREP() to generate an integer constant expression, force
+ * u32 and for consistency with REG_FIELD_GET(), REG_BIT() and REG_GENMASK().
+ *
+ * @return: @__val masked and shifted into the field defined by @__mask.
+ */
+#define REG_FIELD_PREP(__mask, __val)						\
+	((u32)((((typeof(__mask))(__val) << __bf_shf(__mask)) & (__mask)) +	\
+	       BUILD_BUG_ON_ZERO(!__is_constexpr(__mask)) +		\
+	       BUILD_BUG_ON_ZERO((__mask) == 0 || (__mask) > U32_MAX) +		\
+	       BUILD_BUG_ON_ZERO(!IS_POWER_OF_2((__mask) + (1ULL << __bf_shf(__mask)))) + \
+	       BUILD_BUG_ON_ZERO(__builtin_choose_expr(__is_constexpr(__val), (~((__mask) >> __bf_shf(__mask)) & (__val)), 0))))
+
+/**
+ * REG_FIELD_GET() - Extract a u32 bitfield value
+ * @__mask: shifted mask defining the field's length and position
+ * @__val: value to extract the bitfield value from
+ *
+ * Local wrapper for FIELD_GET() to force u32 and for consistency with
+ * REG_FIELD_PREP(), REG_BIT() and REG_GENMASK().
+ *
+ * @return: Masked and shifted value of the field defined by @__mask in @__val.
+ */
+#define REG_FIELD_GET(__mask, __val)	((u32)FIELD_GET(__mask, __val))
+
+/**
+ * REG_FIELD_GET64() - Extract a u64 bitfield value
+ * @__mask: shifted mask defining the field's length and position
+ * @__val: value to extract the bitfield value from
+ *
+ * Local wrapper for FIELD_GET() to force u64 and for consistency with
+ * REG_GENMASK64().
+ *
+ * @return: Masked and shifted value of the field defined by @__mask in @__val.
+ */
+#define REG_FIELD_GET64(__mask, __val)	((u64)FIELD_GET(__mask, __val))
+
+typedef struct {
+	u32 reg;
+} i915_reg_t;
+
+#define _MMIO(r) ((const i915_reg_t){ .reg = (r) })
+
+#define INVALID_MMIO_REG _MMIO(0)
+
+static __always_inline u32 i915_mmio_reg_offset(i915_reg_t reg)
+{
+	return reg.reg;
+}
+
+static inline bool i915_mmio_reg_equal(i915_reg_t a, i915_reg_t b)
+{
+	return i915_mmio_reg_offset(a) == i915_mmio_reg_offset(b);
+}
+
+static inline bool i915_mmio_reg_valid(i915_reg_t reg)
+{
+	return !i915_mmio_reg_equal(reg, INVALID_MMIO_REG);
+}
+
+#define VLV_DISPLAY_BASE		0x180000
+
+#endif /* __I915_REG_DEFS__ */
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
new file mode 100644
index 000000000..7ce126a01
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -0,0 +1,2288 @@
+/*
+ * Copyright © 2008-2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/dma-fence-array.h>
+#include <linux/dma-fence-chain.h>
+#include <linux/irq_work.h>
+#include <linux/prefetch.h>
+#include <linux/sched.h>
+#include <linux/sched/clock.h>
+#include <linux/sched/signal.h>
+#include <linux/sched/mm.h>
+
+#include "gem/i915_gem_context.h"
+#include "gt/intel_breadcrumbs.h"
+#include "gt/intel_context.h"
+#include "gt/intel_engine.h"
+#include "gt/intel_engine_heartbeat.h"
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_reset.h"
+#include "gt/intel_ring.h"
+#include "gt/intel_rps.h"
+
+#include "i915_active.h"
+#include "i915_deps.h"
+#include "i915_driver.h"
+#include "i915_drv.h"
+#include "i915_trace.h"
+#include "intel_pm.h"
+
+struct execute_cb {
+	struct irq_work work;
+	struct i915_sw_fence *fence;
+	struct i915_request *signal;
+};
+
+static struct kmem_cache *slab_requests;
+static struct kmem_cache *slab_execute_cbs;
+
+static const char *i915_fence_get_driver_name(struct dma_fence *fence)
+{
+	return dev_name(to_request(fence)->i915->drm.dev);
+}
+
+static const char *i915_fence_get_timeline_name(struct dma_fence *fence)
+{
+	const struct i915_gem_context *ctx;
+
+	/*
+	 * The timeline struct (as part of the ppgtt underneath a context)
+	 * may be freed when the request is no longer in use by the GPU.
+	 * We could extend the life of a context to beyond that of all
+	 * fences, possibly keeping the hw resource around indefinitely,
+	 * or we just give them a false name. Since
+	 * dma_fence_ops.get_timeline_name is a debug feature, the occasional
+	 * lie seems justifiable.
+	 */
+	if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags))
+		return "signaled";
+
+	ctx = i915_request_gem_context(to_request(fence));
+	if (!ctx)
+		return "[" DRIVER_NAME "]";
+
+	return ctx->name;
+}
+
+static bool i915_fence_signaled(struct dma_fence *fence)
+{
+	return i915_request_completed(to_request(fence));
+}
+
+static bool i915_fence_enable_signaling(struct dma_fence *fence)
+{
+	return i915_request_enable_breadcrumb(to_request(fence));
+}
+
+static signed long i915_fence_wait(struct dma_fence *fence,
+				   bool interruptible,
+				   signed long timeout)
+{
+	return i915_request_wait_timeout(to_request(fence),
+					 interruptible | I915_WAIT_PRIORITY,
+					 timeout);
+}
+
+struct kmem_cache *i915_request_slab_cache(void)
+{
+	return slab_requests;
+}
+
+static void i915_fence_release(struct dma_fence *fence)
+{
+	struct i915_request *rq = to_request(fence);
+
+	GEM_BUG_ON(rq->guc_prio != GUC_PRIO_INIT &&
+		   rq->guc_prio != GUC_PRIO_FINI);
+
+	i915_request_free_capture_list(fetch_and_zero(&rq->capture_list));
+	if (rq->batch_res) {
+		i915_vma_resource_put(rq->batch_res);
+		rq->batch_res = NULL;
+	}
+
+	/*
+	 * The request is put onto a RCU freelist (i.e. the address
+	 * is immediately reused), mark the fences as being freed now.
+	 * Otherwise the debugobjects for the fences are only marked as
+	 * freed when the slab cache itself is freed, and so we would get
+	 * caught trying to reuse dead objects.
+	 */
+	i915_sw_fence_fini(&rq->submit);
+	i915_sw_fence_fini(&rq->semaphore);
+
+	/*
+	 * Keep one request on each engine for reserved use under mempressure.
+	 *
+	 * We do not hold a reference to the engine here and so have to be
+	 * very careful in what rq->engine we poke. The virtual engine is
+	 * referenced via the rq->context and we released that ref during
+	 * i915_request_retire(), ergo we must not dereference a virtual
+	 * engine here. Not that we would want to, as the only consumer of
+	 * the reserved engine->request_pool is the power management parking,
+	 * which must-not-fail, and that is only run on the physical engines.
+	 *
+	 * Since the request must have been executed to be have completed,
+	 * we know that it will have been processed by the HW and will
+	 * not be unsubmitted again, so rq->engine and rq->execution_mask
+	 * at this point is stable. rq->execution_mask will be a single
+	 * bit if the last and _only_ engine it could execution on was a
+	 * physical engine, if it's multiple bits then it started on and
+	 * could still be on a virtual engine. Thus if the mask is not a
+	 * power-of-two we assume that rq->engine may still be a virtual
+	 * engine and so a dangling invalid pointer that we cannot dereference
+	 *
+	 * For example, consider the flow of a bonded request through a virtual
+	 * engine. The request is created with a wide engine mask (all engines
+	 * that we might execute on). On processing the bond, the request mask
+	 * is reduced to one or more engines. If the request is subsequently
+	 * bound to a single engine, it will then be constrained to only
+	 * execute on that engine and never returned to the virtual engine
+	 * after timeslicing away, see __unwind_incomplete_requests(). Thus we
+	 * know that if the rq->execution_mask is a single bit, rq->engine
+	 * can be a physical engine with the exact corresponding mask.
+	 */
+	if (is_power_of_2(rq->execution_mask) &&
+	    !cmpxchg(&rq->engine->request_pool, NULL, rq))
+		return;
+
+	kmem_cache_free(slab_requests, rq);
+}
+
+const struct dma_fence_ops i915_fence_ops = {
+	.get_driver_name = i915_fence_get_driver_name,
+	.get_timeline_name = i915_fence_get_timeline_name,
+	.enable_signaling = i915_fence_enable_signaling,
+	.signaled = i915_fence_signaled,
+	.wait = i915_fence_wait,
+	.release = i915_fence_release,
+};
+
+static void irq_execute_cb(struct irq_work *wrk)
+{
+	struct execute_cb *cb = container_of(wrk, typeof(*cb), work);
+
+	i915_sw_fence_complete(cb->fence);
+	kmem_cache_free(slab_execute_cbs, cb);
+}
+
+static __always_inline void
+__notify_execute_cb(struct i915_request *rq, bool (*fn)(struct irq_work *wrk))
+{
+	struct execute_cb *cb, *cn;
+
+	if (llist_empty(&rq->execute_cb))
+		return;
+
+	llist_for_each_entry_safe(cb, cn,
+				  llist_del_all(&rq->execute_cb),
+				  work.node.llist)
+		fn(&cb->work);
+}
+
+static void __notify_execute_cb_irq(struct i915_request *rq)
+{
+	__notify_execute_cb(rq, irq_work_queue);
+}
+
+static bool irq_work_imm(struct irq_work *wrk)
+{
+	wrk->func(wrk);
+	return false;
+}
+
+void i915_request_notify_execute_cb_imm(struct i915_request *rq)
+{
+	__notify_execute_cb(rq, irq_work_imm);
+}
+
+static void __i915_request_fill(struct i915_request *rq, u8 val)
+{
+	void *vaddr = rq->ring->vaddr;
+	u32 head;
+
+	head = rq->infix;
+	if (rq->postfix < head) {
+		memset(vaddr + head, val, rq->ring->size - head);
+		head = 0;
+	}
+	memset(vaddr + head, val, rq->postfix - head);
+}
+
+/**
+ * i915_request_active_engine
+ * @rq: request to inspect
+ * @active: pointer in which to return the active engine
+ *
+ * Fills the currently active engine to the @active pointer if the request
+ * is active and still not completed.
+ *
+ * Returns true if request was active or false otherwise.
+ */
+bool
+i915_request_active_engine(struct i915_request *rq,
+			   struct intel_engine_cs **active)
+{
+	struct intel_engine_cs *engine, *locked;
+	bool ret = false;
+
+	/*
+	 * Serialise with __i915_request_submit() so that it sees
+	 * is-banned?, or we know the request is already inflight.
+	 *
+	 * Note that rq->engine is unstable, and so we double
+	 * check that we have acquired the lock on the final engine.
+	 */
+	locked = READ_ONCE(rq->engine);
+	spin_lock_irq(&locked->sched_engine->lock);
+	while (unlikely(locked != (engine = READ_ONCE(rq->engine)))) {
+		spin_unlock(&locked->sched_engine->lock);
+		locked = engine;
+		spin_lock(&locked->sched_engine->lock);
+	}
+
+	if (i915_request_is_active(rq)) {
+		if (!__i915_request_is_complete(rq))
+			*active = locked;
+		ret = true;
+	}
+
+	spin_unlock_irq(&locked->sched_engine->lock);
+
+	return ret;
+}
+
+static void __rq_init_watchdog(struct i915_request *rq)
+{
+	rq->watchdog.timer.function = NULL;
+}
+
+static enum hrtimer_restart __rq_watchdog_expired(struct hrtimer *hrtimer)
+{
+	struct i915_request *rq =
+		container_of(hrtimer, struct i915_request, watchdog.timer);
+	struct intel_gt *gt = rq->engine->gt;
+
+	if (!i915_request_completed(rq)) {
+		if (llist_add(&rq->watchdog.link, &gt->watchdog.list))
+			schedule_work(&gt->watchdog.work);
+	} else {
+		i915_request_put(rq);
+	}
+
+	return HRTIMER_NORESTART;
+}
+
+static void __rq_arm_watchdog(struct i915_request *rq)
+{
+	struct i915_request_watchdog *wdg = &rq->watchdog;
+	struct intel_context *ce = rq->context;
+
+	if (!ce->watchdog.timeout_us)
+		return;
+
+	i915_request_get(rq);
+
+	hrtimer_init(&wdg->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+	wdg->timer.function = __rq_watchdog_expired;
+	hrtimer_start_range_ns(&wdg->timer,
+			       ns_to_ktime(ce->watchdog.timeout_us *
+					   NSEC_PER_USEC),
+			       NSEC_PER_MSEC,
+			       HRTIMER_MODE_REL);
+}
+
+static void __rq_cancel_watchdog(struct i915_request *rq)
+{
+	struct i915_request_watchdog *wdg = &rq->watchdog;
+
+	if (wdg->timer.function && hrtimer_try_to_cancel(&wdg->timer) > 0)
+		i915_request_put(rq);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+
+/**
+ * i915_request_free_capture_list - Free a capture list
+ * @capture: Pointer to the first list item or NULL
+ *
+ */
+void i915_request_free_capture_list(struct i915_capture_list *capture)
+{
+	while (capture) {
+		struct i915_capture_list *next = capture->next;
+
+		i915_vma_resource_put(capture->vma_res);
+		kfree(capture);
+		capture = next;
+	}
+}
+
+#define assert_capture_list_is_null(_rq) GEM_BUG_ON((_rq)->capture_list)
+
+#define clear_capture_list(_rq) ((_rq)->capture_list = NULL)
+
+#else
+
+#define i915_request_free_capture_list(_a) do {} while (0)
+
+#define assert_capture_list_is_null(_a) do {} while (0)
+
+#define clear_capture_list(_rq) do {} while (0)
+
+#endif
+
+bool i915_request_retire(struct i915_request *rq)
+{
+	if (!__i915_request_is_complete(rq))
+		return false;
+
+	RQ_TRACE(rq, "\n");
+
+	GEM_BUG_ON(!i915_sw_fence_signaled(&rq->submit));
+	trace_i915_request_retire(rq);
+	i915_request_mark_complete(rq);
+
+	__rq_cancel_watchdog(rq);
+
+	/*
+	 * We know the GPU must have read the request to have
+	 * sent us the seqno + interrupt, so use the position
+	 * of tail of the request to update the last known position
+	 * of the GPU head.
+	 *
+	 * Note this requires that we are always called in request
+	 * completion order.
+	 */
+	GEM_BUG_ON(!list_is_first(&rq->link,
+				  &i915_request_timeline(rq)->requests));
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		/* Poison before we release our space in the ring */
+		__i915_request_fill(rq, POISON_FREE);
+	rq->ring->head = rq->postfix;
+
+	if (!i915_request_signaled(rq)) {
+		spin_lock_irq(&rq->lock);
+		dma_fence_signal_locked(&rq->fence);
+		spin_unlock_irq(&rq->lock);
+	}
+
+	if (test_and_set_bit(I915_FENCE_FLAG_BOOST, &rq->fence.flags))
+		intel_rps_dec_waiters(&rq->engine->gt->rps);
+
+	/*
+	 * We only loosely track inflight requests across preemption,
+	 * and so we may find ourselves attempting to retire a _completed_
+	 * request that we have removed from the HW and put back on a run
+	 * queue.
+	 *
+	 * As we set I915_FENCE_FLAG_ACTIVE on the request, this should be
+	 * after removing the breadcrumb and signaling it, so that we do not
+	 * inadvertently attach the breadcrumb to a completed request.
+	 */
+	rq->engine->remove_active_request(rq);
+	GEM_BUG_ON(!llist_empty(&rq->execute_cb));
+
+	__list_del_entry(&rq->link); /* poison neither prev/next (RCU walks) */
+
+	intel_context_exit(rq->context);
+	intel_context_unpin(rq->context);
+
+	i915_sched_node_fini(&rq->sched);
+	i915_request_put(rq);
+
+	return true;
+}
+
+void i915_request_retire_upto(struct i915_request *rq)
+{
+	struct intel_timeline * const tl = i915_request_timeline(rq);
+	struct i915_request *tmp;
+
+	RQ_TRACE(rq, "\n");
+	GEM_BUG_ON(!__i915_request_is_complete(rq));
+
+	do {
+		tmp = list_first_entry(&tl->requests, typeof(*tmp), link);
+		GEM_BUG_ON(!i915_request_completed(tmp));
+	} while (i915_request_retire(tmp) && tmp != rq);
+}
+
+static struct i915_request * const *
+__engine_active(struct intel_engine_cs *engine)
+{
+	return READ_ONCE(engine->execlists.active);
+}
+
+static bool __request_in_flight(const struct i915_request *signal)
+{
+	struct i915_request * const *port, *rq;
+	bool inflight = false;
+
+	if (!i915_request_is_ready(signal))
+		return false;
+
+	/*
+	 * Even if we have unwound the request, it may still be on
+	 * the GPU (preempt-to-busy). If that request is inside an
+	 * unpreemptible critical section, it will not be removed. Some
+	 * GPU functions may even be stuck waiting for the paired request
+	 * (__await_execution) to be submitted and cannot be preempted
+	 * until the bond is executing.
+	 *
+	 * As we know that there are always preemption points between
+	 * requests, we know that only the currently executing request
+	 * may be still active even though we have cleared the flag.
+	 * However, we can't rely on our tracking of ELSP[0] to know
+	 * which request is currently active and so maybe stuck, as
+	 * the tracking maybe an event behind. Instead assume that
+	 * if the context is still inflight, then it is still active
+	 * even if the active flag has been cleared.
+	 *
+	 * To further complicate matters, if there a pending promotion, the HW
+	 * may either perform a context switch to the second inflight execlists,
+	 * or it may switch to the pending set of execlists. In the case of the
+	 * latter, it may send the ACK and we process the event copying the
+	 * pending[] over top of inflight[], _overwriting_ our *active. Since
+	 * this implies the HW is arbitrating and not struck in *active, we do
+	 * not worry about complete accuracy, but we do require no read/write
+	 * tearing of the pointer [the read of the pointer must be valid, even
+	 * as the array is being overwritten, for which we require the writes
+	 * to avoid tearing.]
+	 *
+	 * Note that the read of *execlists->active may race with the promotion
+	 * of execlists->pending[] to execlists->inflight[], overwritting
+	 * the value at *execlists->active. This is fine. The promotion implies
+	 * that we received an ACK from the HW, and so the context is not
+	 * stuck -- if we do not see ourselves in *active, the inflight status
+	 * is valid. If instead we see ourselves being copied into *active,
+	 * we are inflight and may signal the callback.
+	 */
+	if (!intel_context_inflight(signal->context))
+		return false;
+
+	rcu_read_lock();
+	for (port = __engine_active(signal->engine);
+	     (rq = READ_ONCE(*port)); /* may race with promotion of pending[] */
+	     port++) {
+		if (rq->context == signal->context) {
+			inflight = i915_seqno_passed(rq->fence.seqno,
+						     signal->fence.seqno);
+			break;
+		}
+	}
+	rcu_read_unlock();
+
+	return inflight;
+}
+
+static int
+__await_execution(struct i915_request *rq,
+		  struct i915_request *signal,
+		  gfp_t gfp)
+{
+	struct execute_cb *cb;
+
+	if (i915_request_is_active(signal))
+		return 0;
+
+	cb = kmem_cache_alloc(slab_execute_cbs, gfp);
+	if (!cb)
+		return -ENOMEM;
+
+	cb->fence = &rq->submit;
+	i915_sw_fence_await(cb->fence);
+	init_irq_work(&cb->work, irq_execute_cb);
+
+	/*
+	 * Register the callback first, then see if the signaler is already
+	 * active. This ensures that if we race with the
+	 * __notify_execute_cb from i915_request_submit() and we are not
+	 * included in that list, we get a second bite of the cherry and
+	 * execute it ourselves. After this point, a future
+	 * i915_request_submit() will notify us.
+	 *
+	 * In i915_request_retire() we set the ACTIVE bit on a completed
+	 * request (then flush the execute_cb). So by registering the
+	 * callback first, then checking the ACTIVE bit, we serialise with
+	 * the completed/retired request.
+	 */
+	if (llist_add(&cb->work.node.llist, &signal->execute_cb)) {
+		if (i915_request_is_active(signal) ||
+		    __request_in_flight(signal))
+			i915_request_notify_execute_cb_imm(signal);
+	}
+
+	return 0;
+}
+
+static bool fatal_error(int error)
+{
+	switch (error) {
+	case 0: /* not an error! */
+	case -EAGAIN: /* innocent victim of a GT reset (__i915_request_reset) */
+	case -ETIMEDOUT: /* waiting for Godot (timer_i915_sw_fence_wake) */
+		return false;
+	default:
+		return true;
+	}
+}
+
+void __i915_request_skip(struct i915_request *rq)
+{
+	GEM_BUG_ON(!fatal_error(rq->fence.error));
+
+	if (rq->infix == rq->postfix)
+		return;
+
+	RQ_TRACE(rq, "error: %d\n", rq->fence.error);
+
+	/*
+	 * As this request likely depends on state from the lost
+	 * context, clear out all the user operations leaving the
+	 * breadcrumb at the end (so we get the fence notifications).
+	 */
+	__i915_request_fill(rq, 0);
+	rq->infix = rq->postfix;
+}
+
+bool i915_request_set_error_once(struct i915_request *rq, int error)
+{
+	int old;
+
+	GEM_BUG_ON(!IS_ERR_VALUE((long)error));
+
+	if (i915_request_signaled(rq))
+		return false;
+
+	old = READ_ONCE(rq->fence.error);
+	do {
+		if (fatal_error(old))
+			return false;
+	} while (!try_cmpxchg(&rq->fence.error, &old, error));
+
+	return true;
+}
+
+struct i915_request *i915_request_mark_eio(struct i915_request *rq)
+{
+	if (__i915_request_is_complete(rq))
+		return NULL;
+
+	GEM_BUG_ON(i915_request_signaled(rq));
+
+	/* As soon as the request is completed, it may be retired */
+	rq = i915_request_get(rq);
+
+	i915_request_set_error_once(rq, -EIO);
+	i915_request_mark_complete(rq);
+
+	return rq;
+}
+
+bool __i915_request_submit(struct i915_request *request)
+{
+	struct intel_engine_cs *engine = request->engine;
+	bool result = false;
+
+	RQ_TRACE(request, "\n");
+
+	GEM_BUG_ON(!irqs_disabled());
+	lockdep_assert_held(&engine->sched_engine->lock);
+
+	/*
+	 * With the advent of preempt-to-busy, we frequently encounter
+	 * requests that we have unsubmitted from HW, but left running
+	 * until the next ack and so have completed in the meantime. On
+	 * resubmission of that completed request, we can skip
+	 * updating the payload, and execlists can even skip submitting
+	 * the request.
+	 *
+	 * We must remove the request from the caller's priority queue,
+	 * and the caller must only call us when the request is in their
+	 * priority queue, under the sched_engine->lock. This ensures that the
+	 * request has *not* yet been retired and we can safely move
+	 * the request into the engine->active.list where it will be
+	 * dropped upon retiring. (Otherwise if resubmit a *retired*
+	 * request, this would be a horrible use-after-free.)
+	 */
+	if (__i915_request_is_complete(request)) {
+		list_del_init(&request->sched.link);
+		goto active;
+	}
+
+	if (unlikely(!intel_context_is_schedulable(request->context)))
+		i915_request_set_error_once(request, -EIO);
+
+	if (unlikely(fatal_error(request->fence.error)))
+		__i915_request_skip(request);
+
+	/*
+	 * Are we using semaphores when the gpu is already saturated?
+	 *
+	 * Using semaphores incurs a cost in having the GPU poll a
+	 * memory location, busywaiting for it to change. The continual
+	 * memory reads can have a noticeable impact on the rest of the
+	 * system with the extra bus traffic, stalling the cpu as it too
+	 * tries to access memory across the bus (perf stat -e bus-cycles).
+	 *
+	 * If we installed a semaphore on this request and we only submit
+	 * the request after the signaler completed, that indicates the
+	 * system is overloaded and using semaphores at this time only
+	 * increases the amount of work we are doing. If so, we disable
+	 * further use of semaphores until we are idle again, whence we
+	 * optimistically try again.
+	 */
+	if (request->sched.semaphores &&
+	    i915_sw_fence_signaled(&request->semaphore))
+		engine->saturated |= request->sched.semaphores;
+
+	engine->emit_fini_breadcrumb(request,
+				     request->ring->vaddr + request->postfix);
+
+	trace_i915_request_execute(request);
+	if (engine->bump_serial)
+		engine->bump_serial(engine);
+	else
+		engine->serial++;
+
+	result = true;
+
+	GEM_BUG_ON(test_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags));
+	engine->add_active_request(request);
+active:
+	clear_bit(I915_FENCE_FLAG_PQUEUE, &request->fence.flags);
+	set_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags);
+
+	/*
+	 * XXX Rollback bonded-execution on __i915_request_unsubmit()?
+	 *
+	 * In the future, perhaps when we have an active time-slicing scheduler,
+	 * it will be interesting to unsubmit parallel execution and remove
+	 * busywaits from the GPU until their master is restarted. This is
+	 * quite hairy, we have to carefully rollback the fence and do a
+	 * preempt-to-idle cycle on the target engine, all the while the
+	 * master execute_cb may refire.
+	 */
+	__notify_execute_cb_irq(request);
+
+	/* We may be recursing from the signal callback of another i915 fence */
+	if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &request->fence.flags))
+		i915_request_enable_breadcrumb(request);
+
+	return result;
+}
+
+void i915_request_submit(struct i915_request *request)
+{
+	struct intel_engine_cs *engine = request->engine;
+	unsigned long flags;
+
+	/* Will be called from irq-context when using foreign fences. */
+	spin_lock_irqsave(&engine->sched_engine->lock, flags);
+
+	__i915_request_submit(request);
+
+	spin_unlock_irqrestore(&engine->sched_engine->lock, flags);
+}
+
+void __i915_request_unsubmit(struct i915_request *request)
+{
+	struct intel_engine_cs *engine = request->engine;
+
+	/*
+	 * Only unwind in reverse order, required so that the per-context list
+	 * is kept in seqno/ring order.
+	 */
+	RQ_TRACE(request, "\n");
+
+	GEM_BUG_ON(!irqs_disabled());
+	lockdep_assert_held(&engine->sched_engine->lock);
+
+	/*
+	 * Before we remove this breadcrumb from the signal list, we have
+	 * to ensure that a concurrent dma_fence_enable_signaling() does not
+	 * attach itself. We first mark the request as no longer active and
+	 * make sure that is visible to other cores, and then remove the
+	 * breadcrumb if attached.
+	 */
+	GEM_BUG_ON(!test_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags));
+	clear_bit_unlock(I915_FENCE_FLAG_ACTIVE, &request->fence.flags);
+	if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &request->fence.flags))
+		i915_request_cancel_breadcrumb(request);
+
+	/* We've already spun, don't charge on resubmitting. */
+	if (request->sched.semaphores && __i915_request_has_started(request))
+		request->sched.semaphores = 0;
+
+	/*
+	 * We don't need to wake_up any waiters on request->execute, they
+	 * will get woken by any other event or us re-adding this request
+	 * to the engine timeline (__i915_request_submit()). The waiters
+	 * should be quite adapt at finding that the request now has a new
+	 * global_seqno to the one they went to sleep on.
+	 */
+}
+
+void i915_request_unsubmit(struct i915_request *request)
+{
+	struct intel_engine_cs *engine = request->engine;
+	unsigned long flags;
+
+	/* Will be called from irq-context when using foreign fences. */
+	spin_lock_irqsave(&engine->sched_engine->lock, flags);
+
+	__i915_request_unsubmit(request);
+
+	spin_unlock_irqrestore(&engine->sched_engine->lock, flags);
+}
+
+void i915_request_cancel(struct i915_request *rq, int error)
+{
+	if (!i915_request_set_error_once(rq, error))
+		return;
+
+	set_bit(I915_FENCE_FLAG_SENTINEL, &rq->fence.flags);
+
+	intel_context_cancel_request(rq->context, rq);
+}
+
+static int
+submit_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
+{
+	struct i915_request *request =
+		container_of(fence, typeof(*request), submit);
+
+	switch (state) {
+	case FENCE_COMPLETE:
+		trace_i915_request_submit(request);
+
+		if (unlikely(fence->error))
+			i915_request_set_error_once(request, fence->error);
+		else
+			__rq_arm_watchdog(request);
+
+		/*
+		 * We need to serialize use of the submit_request() callback
+		 * with its hotplugging performed during an emergency
+		 * i915_gem_set_wedged().  We use the RCU mechanism to mark the
+		 * critical section in order to force i915_gem_set_wedged() to
+		 * wait until the submit_request() is completed before
+		 * proceeding.
+		 */
+		rcu_read_lock();
+		request->engine->submit_request(request);
+		rcu_read_unlock();
+		break;
+
+	case FENCE_FREE:
+		i915_request_put(request);
+		break;
+	}
+
+	return NOTIFY_DONE;
+}
+
+static int
+semaphore_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
+{
+	struct i915_request *rq = container_of(fence, typeof(*rq), semaphore);
+
+	switch (state) {
+	case FENCE_COMPLETE:
+		break;
+
+	case FENCE_FREE:
+		i915_request_put(rq);
+		break;
+	}
+
+	return NOTIFY_DONE;
+}
+
+static void retire_requests(struct intel_timeline *tl)
+{
+	struct i915_request *rq, *rn;
+
+	list_for_each_entry_safe(rq, rn, &tl->requests, link)
+		if (!i915_request_retire(rq))
+			break;
+}
+
+static noinline struct i915_request *
+request_alloc_slow(struct intel_timeline *tl,
+		   struct i915_request **rsvd,
+		   gfp_t gfp)
+{
+	struct i915_request *rq;
+
+	/* If we cannot wait, dip into our reserves */
+	if (!gfpflags_allow_blocking(gfp)) {
+		rq = xchg(rsvd, NULL);
+		if (!rq) /* Use the normal failure path for one final WARN */
+			goto out;
+
+		return rq;
+	}
+
+	if (list_empty(&tl->requests))
+		goto out;
+
+	/* Move our oldest request to the slab-cache (if not in use!) */
+	rq = list_first_entry(&tl->requests, typeof(*rq), link);
+	i915_request_retire(rq);
+
+	rq = kmem_cache_alloc(slab_requests,
+			      gfp | __GFP_RETRY_MAYFAIL | __GFP_NOWARN);
+	if (rq)
+		return rq;
+
+	/* Ratelimit ourselves to prevent oom from malicious clients */
+	rq = list_last_entry(&tl->requests, typeof(*rq), link);
+	cond_synchronize_rcu(rq->rcustate);
+
+	/* Retire our old requests in the hope that we free some */
+	retire_requests(tl);
+
+out:
+	return kmem_cache_alloc(slab_requests, gfp);
+}
+
+static void __i915_request_ctor(void *arg)
+{
+	struct i915_request *rq = arg;
+
+	spin_lock_init(&rq->lock);
+	i915_sched_node_init(&rq->sched);
+	i915_sw_fence_init(&rq->submit, submit_notify);
+	i915_sw_fence_init(&rq->semaphore, semaphore_notify);
+
+	clear_capture_list(rq);
+	rq->batch_res = NULL;
+
+	init_llist_head(&rq->execute_cb);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#define clear_batch_ptr(_rq) ((_rq)->batch = NULL)
+#else
+#define clear_batch_ptr(_a) do {} while (0)
+#endif
+
+struct i915_request *
+__i915_request_create(struct intel_context *ce, gfp_t gfp)
+{
+	struct intel_timeline *tl = ce->timeline;
+	struct i915_request *rq;
+	u32 seqno;
+	int ret;
+
+	might_alloc(gfp);
+
+	/* Check that the caller provided an already pinned context */
+	__intel_context_pin(ce);
+
+	/*
+	 * Beware: Dragons be flying overhead.
+	 *
+	 * We use RCU to look up requests in flight. The lookups may
+	 * race with the request being allocated from the slab freelist.
+	 * That is the request we are writing to here, may be in the process
+	 * of being read by __i915_active_request_get_rcu(). As such,
+	 * we have to be very careful when overwriting the contents. During
+	 * the RCU lookup, we change chase the request->engine pointer,
+	 * read the request->global_seqno and increment the reference count.
+	 *
+	 * The reference count is incremented atomically. If it is zero,
+	 * the lookup knows the request is unallocated and complete. Otherwise,
+	 * it is either still in use, or has been reallocated and reset
+	 * with dma_fence_init(). This increment is safe for release as we
+	 * check that the request we have a reference to and matches the active
+	 * request.
+	 *
+	 * Before we increment the refcount, we chase the request->engine
+	 * pointer. We must not call kmem_cache_zalloc() or else we set
+	 * that pointer to NULL and cause a crash during the lookup. If
+	 * we see the request is completed (based on the value of the
+	 * old engine and seqno), the lookup is complete and reports NULL.
+	 * If we decide the request is not completed (new engine or seqno),
+	 * then we grab a reference and double check that it is still the
+	 * active request - which it won't be and restart the lookup.
+	 *
+	 * Do not use kmem_cache_zalloc() here!
+	 */
+	rq = kmem_cache_alloc(slab_requests,
+			      gfp | __GFP_RETRY_MAYFAIL | __GFP_NOWARN);
+	if (unlikely(!rq)) {
+		rq = request_alloc_slow(tl, &ce->engine->request_pool, gfp);
+		if (!rq) {
+			ret = -ENOMEM;
+			goto err_unreserve;
+		}
+	}
+
+	rq->context = ce;
+	rq->engine = ce->engine;
+	rq->ring = ce->ring;
+	rq->execution_mask = ce->engine->mask;
+	rq->i915 = ce->engine->i915;
+
+	ret = intel_timeline_get_seqno(tl, rq, &seqno);
+	if (ret)
+		goto err_free;
+
+	dma_fence_init(&rq->fence, &i915_fence_ops, &rq->lock,
+		       tl->fence_context, seqno);
+
+	RCU_INIT_POINTER(rq->timeline, tl);
+	rq->hwsp_seqno = tl->hwsp_seqno;
+	GEM_BUG_ON(__i915_request_is_complete(rq));
+
+	rq->rcustate = get_state_synchronize_rcu(); /* acts as smp_mb() */
+
+	rq->guc_prio = GUC_PRIO_INIT;
+
+	/* We bump the ref for the fence chain */
+	i915_sw_fence_reinit(&i915_request_get(rq)->submit);
+	i915_sw_fence_reinit(&i915_request_get(rq)->semaphore);
+
+	i915_sched_node_reinit(&rq->sched);
+
+	/* No zalloc, everything must be cleared after use */
+	clear_batch_ptr(rq);
+	__rq_init_watchdog(rq);
+	assert_capture_list_is_null(rq);
+	GEM_BUG_ON(!llist_empty(&rq->execute_cb));
+	GEM_BUG_ON(rq->batch_res);
+
+	/*
+	 * Reserve space in the ring buffer for all the commands required to
+	 * eventually emit this request. This is to guarantee that the
+	 * i915_request_add() call can't fail. Note that the reserve may need
+	 * to be redone if the request is not actually submitted straight
+	 * away, e.g. because a GPU scheduler has deferred it.
+	 *
+	 * Note that due to how we add reserved_space to intel_ring_begin()
+	 * we need to double our request to ensure that if we need to wrap
+	 * around inside i915_request_add() there is sufficient space at
+	 * the beginning of the ring as well.
+	 */
+	rq->reserved_space =
+		2 * rq->engine->emit_fini_breadcrumb_dw * sizeof(u32);
+
+	/*
+	 * Record the position of the start of the request so that
+	 * should we detect the updated seqno part-way through the
+	 * GPU processing the request, we never over-estimate the
+	 * position of the head.
+	 */
+	rq->head = rq->ring->emit;
+
+	ret = rq->engine->request_alloc(rq);
+	if (ret)
+		goto err_unwind;
+
+	rq->infix = rq->ring->emit; /* end of header; start of user payload */
+
+	intel_context_mark_active(ce);
+	list_add_tail_rcu(&rq->link, &tl->requests);
+
+	return rq;
+
+err_unwind:
+	ce->ring->emit = rq->head;
+
+	/* Make sure we didn't add ourselves to external state before freeing */
+	GEM_BUG_ON(!list_empty(&rq->sched.signalers_list));
+	GEM_BUG_ON(!list_empty(&rq->sched.waiters_list));
+
+err_free:
+	kmem_cache_free(slab_requests, rq);
+err_unreserve:
+	intel_context_unpin(ce);
+	return ERR_PTR(ret);
+}
+
+struct i915_request *
+i915_request_create(struct intel_context *ce)
+{
+	struct i915_request *rq;
+	struct intel_timeline *tl;
+
+	tl = intel_context_timeline_lock(ce);
+	if (IS_ERR(tl))
+		return ERR_CAST(tl);
+
+	/* Move our oldest request to the slab-cache (if not in use!) */
+	rq = list_first_entry(&tl->requests, typeof(*rq), link);
+	if (!list_is_last(&rq->link, &tl->requests))
+		i915_request_retire(rq);
+
+	intel_context_enter(ce);
+	rq = __i915_request_create(ce, GFP_KERNEL);
+	intel_context_exit(ce); /* active reference transferred to request */
+	if (IS_ERR(rq))
+		goto err_unlock;
+
+	/* Check that we do not interrupt ourselves with a new request */
+	rq->cookie = lockdep_pin_lock(&tl->mutex);
+
+	return rq;
+
+err_unlock:
+	intel_context_timeline_unlock(tl);
+	return rq;
+}
+
+static int
+i915_request_await_start(struct i915_request *rq, struct i915_request *signal)
+{
+	struct dma_fence *fence;
+	int err;
+
+	if (i915_request_timeline(rq) == rcu_access_pointer(signal->timeline))
+		return 0;
+
+	if (i915_request_started(signal))
+		return 0;
+
+	/*
+	 * The caller holds a reference on @signal, but we do not serialise
+	 * against it being retired and removed from the lists.
+	 *
+	 * We do not hold a reference to the request before @signal, and
+	 * so must be very careful to ensure that it is not _recycled_ as
+	 * we follow the link backwards.
+	 */
+	fence = NULL;
+	rcu_read_lock();
+	do {
+		struct list_head *pos = READ_ONCE(signal->link.prev);
+		struct i915_request *prev;
+
+		/* Confirm signal has not been retired, the link is valid */
+		if (unlikely(__i915_request_has_started(signal)))
+			break;
+
+		/* Is signal the earliest request on its timeline? */
+		if (pos == &rcu_dereference(signal->timeline)->requests)
+			break;
+
+		/*
+		 * Peek at the request before us in the timeline. That
+		 * request will only be valid before it is retired, so
+		 * after acquiring a reference to it, confirm that it is
+		 * still part of the signaler's timeline.
+		 */
+		prev = list_entry(pos, typeof(*prev), link);
+		if (!i915_request_get_rcu(prev))
+			break;
+
+		/* After the strong barrier, confirm prev is still attached */
+		if (unlikely(READ_ONCE(prev->link.next) != &signal->link)) {
+			i915_request_put(prev);
+			break;
+		}
+
+		fence = &prev->fence;
+	} while (0);
+	rcu_read_unlock();
+	if (!fence)
+		return 0;
+
+	err = 0;
+	if (!intel_timeline_sync_is_later(i915_request_timeline(rq), fence))
+		err = i915_sw_fence_await_dma_fence(&rq->submit,
+						    fence, 0,
+						    I915_FENCE_GFP);
+	dma_fence_put(fence);
+
+	return err;
+}
+
+static intel_engine_mask_t
+already_busywaiting(struct i915_request *rq)
+{
+	/*
+	 * Polling a semaphore causes bus traffic, delaying other users of
+	 * both the GPU and CPU. We want to limit the impact on others,
+	 * while taking advantage of early submission to reduce GPU
+	 * latency. Therefore we restrict ourselves to not using more
+	 * than one semaphore from each source, and not using a semaphore
+	 * if we have detected the engine is saturated (i.e. would not be
+	 * submitted early and cause bus traffic reading an already passed
+	 * semaphore).
+	 *
+	 * See the are-we-too-late? check in __i915_request_submit().
+	 */
+	return rq->sched.semaphores | READ_ONCE(rq->engine->saturated);
+}
+
+static int
+__emit_semaphore_wait(struct i915_request *to,
+		      struct i915_request *from,
+		      u32 seqno)
+{
+	const int has_token = GRAPHICS_VER(to->engine->i915) >= 12;
+	u32 hwsp_offset;
+	int len, err;
+	u32 *cs;
+
+	GEM_BUG_ON(GRAPHICS_VER(to->engine->i915) < 8);
+	GEM_BUG_ON(i915_request_has_initial_breadcrumb(to));
+
+	/* We need to pin the signaler's HWSP until we are finished reading. */
+	err = intel_timeline_read_hwsp(from, to, &hwsp_offset);
+	if (err)
+		return err;
+
+	len = 4;
+	if (has_token)
+		len += 2;
+
+	cs = intel_ring_begin(to, len);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	/*
+	 * Using greater-than-or-equal here means we have to worry
+	 * about seqno wraparound. To side step that issue, we swap
+	 * the timeline HWSP upon wrapping, so that everyone listening
+	 * for the old (pre-wrap) values do not see the much smaller
+	 * (post-wrap) values than they were expecting (and so wait
+	 * forever).
+	 */
+	*cs++ = (MI_SEMAPHORE_WAIT |
+		 MI_SEMAPHORE_GLOBAL_GTT |
+		 MI_SEMAPHORE_POLL |
+		 MI_SEMAPHORE_SAD_GTE_SDD) +
+		has_token;
+	*cs++ = seqno;
+	*cs++ = hwsp_offset;
+	*cs++ = 0;
+	if (has_token) {
+		*cs++ = 0;
+		*cs++ = MI_NOOP;
+	}
+
+	intel_ring_advance(to, cs);
+	return 0;
+}
+
+static bool
+can_use_semaphore_wait(struct i915_request *to, struct i915_request *from)
+{
+	return to->engine->gt->ggtt == from->engine->gt->ggtt;
+}
+
+static int
+emit_semaphore_wait(struct i915_request *to,
+		    struct i915_request *from,
+		    gfp_t gfp)
+{
+	const intel_engine_mask_t mask = READ_ONCE(from->engine)->mask;
+	struct i915_sw_fence *wait = &to->submit;
+
+	if (!can_use_semaphore_wait(to, from))
+		goto await_fence;
+
+	if (!intel_context_use_semaphores(to->context))
+		goto await_fence;
+
+	if (i915_request_has_initial_breadcrumb(to))
+		goto await_fence;
+
+	/*
+	 * If this or its dependents are waiting on an external fence
+	 * that may fail catastrophically, then we want to avoid using
+	 * sempahores as they bypass the fence signaling metadata, and we
+	 * lose the fence->error propagation.
+	 */
+	if (from->sched.flags & I915_SCHED_HAS_EXTERNAL_CHAIN)
+		goto await_fence;
+
+	/* Just emit the first semaphore we see as request space is limited. */
+	if (already_busywaiting(to) & mask)
+		goto await_fence;
+
+	if (i915_request_await_start(to, from) < 0)
+		goto await_fence;
+
+	/* Only submit our spinner after the signaler is running! */
+	if (__await_execution(to, from, gfp))
+		goto await_fence;
+
+	if (__emit_semaphore_wait(to, from, from->fence.seqno))
+		goto await_fence;
+
+	to->sched.semaphores |= mask;
+	wait = &to->semaphore;
+
+await_fence:
+	return i915_sw_fence_await_dma_fence(wait,
+					     &from->fence, 0,
+					     I915_FENCE_GFP);
+}
+
+static bool intel_timeline_sync_has_start(struct intel_timeline *tl,
+					  struct dma_fence *fence)
+{
+	return __intel_timeline_sync_is_later(tl,
+					      fence->context,
+					      fence->seqno - 1);
+}
+
+static int intel_timeline_sync_set_start(struct intel_timeline *tl,
+					 const struct dma_fence *fence)
+{
+	return __intel_timeline_sync_set(tl, fence->context, fence->seqno - 1);
+}
+
+static int
+__i915_request_await_execution(struct i915_request *to,
+			       struct i915_request *from)
+{
+	int err;
+
+	GEM_BUG_ON(intel_context_is_barrier(from->context));
+
+	/* Submit both requests at the same time */
+	err = __await_execution(to, from, I915_FENCE_GFP);
+	if (err)
+		return err;
+
+	/* Squash repeated depenendices to the same timelines */
+	if (intel_timeline_sync_has_start(i915_request_timeline(to),
+					  &from->fence))
+		return 0;
+
+	/*
+	 * Wait until the start of this request.
+	 *
+	 * The execution cb fires when we submit the request to HW. But in
+	 * many cases this may be long before the request itself is ready to
+	 * run (consider that we submit 2 requests for the same context, where
+	 * the request of interest is behind an indefinite spinner). So we hook
+	 * up to both to reduce our queues and keep the execution lag minimised
+	 * in the worst case, though we hope that the await_start is elided.
+	 */
+	err = i915_request_await_start(to, from);
+	if (err < 0)
+		return err;
+
+	/*
+	 * Ensure both start together [after all semaphores in signal]
+	 *
+	 * Now that we are queued to the HW at roughly the same time (thanks
+	 * to the execute cb) and are ready to run at roughly the same time
+	 * (thanks to the await start), our signaler may still be indefinitely
+	 * delayed by waiting on a semaphore from a remote engine. If our
+	 * signaler depends on a semaphore, so indirectly do we, and we do not
+	 * want to start our payload until our signaler also starts theirs.
+	 * So we wait.
+	 *
+	 * However, there is also a second condition for which we need to wait
+	 * for the precise start of the signaler. Consider that the signaler
+	 * was submitted in a chain of requests following another context
+	 * (with just an ordinary intra-engine fence dependency between the
+	 * two). In this case the signaler is queued to HW, but not for
+	 * immediate execution, and so we must wait until it reaches the
+	 * active slot.
+	 */
+	if (can_use_semaphore_wait(to, from) &&
+	    intel_engine_has_semaphores(to->engine) &&
+	    !i915_request_has_initial_breadcrumb(to)) {
+		err = __emit_semaphore_wait(to, from, from->fence.seqno - 1);
+		if (err < 0)
+			return err;
+	}
+
+	/* Couple the dependency tree for PI on this exposed to->fence */
+	if (to->engine->sched_engine->schedule) {
+		err = i915_sched_node_add_dependency(&to->sched,
+						     &from->sched,
+						     I915_DEPENDENCY_WEAK);
+		if (err < 0)
+			return err;
+	}
+
+	return intel_timeline_sync_set_start(i915_request_timeline(to),
+					     &from->fence);
+}
+
+static void mark_external(struct i915_request *rq)
+{
+	/*
+	 * The downside of using semaphores is that we lose metadata passing
+	 * along the signaling chain. This is particularly nasty when we
+	 * need to pass along a fatal error such as EFAULT or EDEADLK. For
+	 * fatal errors we want to scrub the request before it is executed,
+	 * which means that we cannot preload the request onto HW and have
+	 * it wait upon a semaphore.
+	 */
+	rq->sched.flags |= I915_SCHED_HAS_EXTERNAL_CHAIN;
+}
+
+static int
+__i915_request_await_external(struct i915_request *rq, struct dma_fence *fence)
+{
+	mark_external(rq);
+	return i915_sw_fence_await_dma_fence(&rq->submit, fence,
+					     i915_fence_context_timeout(rq->engine->i915,
+									fence->context),
+					     I915_FENCE_GFP);
+}
+
+static int
+i915_request_await_external(struct i915_request *rq, struct dma_fence *fence)
+{
+	struct dma_fence *iter;
+	int err = 0;
+
+	if (!to_dma_fence_chain(fence))
+		return __i915_request_await_external(rq, fence);
+
+	dma_fence_chain_for_each(iter, fence) {
+		struct dma_fence_chain *chain = to_dma_fence_chain(iter);
+
+		if (!dma_fence_is_i915(chain->fence)) {
+			err = __i915_request_await_external(rq, iter);
+			break;
+		}
+
+		err = i915_request_await_dma_fence(rq, chain->fence);
+		if (err < 0)
+			break;
+	}
+
+	dma_fence_put(iter);
+	return err;
+}
+
+static inline bool is_parallel_rq(struct i915_request *rq)
+{
+	return intel_context_is_parallel(rq->context);
+}
+
+static inline struct intel_context *request_to_parent(struct i915_request *rq)
+{
+	return intel_context_to_parent(rq->context);
+}
+
+static bool is_same_parallel_context(struct i915_request *to,
+				     struct i915_request *from)
+{
+	if (is_parallel_rq(to))
+		return request_to_parent(to) == request_to_parent(from);
+
+	return false;
+}
+
+int
+i915_request_await_execution(struct i915_request *rq,
+			     struct dma_fence *fence)
+{
+	struct dma_fence **child = &fence;
+	unsigned int nchild = 1;
+	int ret;
+
+	if (dma_fence_is_array(fence)) {
+		struct dma_fence_array *array = to_dma_fence_array(fence);
+
+		/* XXX Error for signal-on-any fence arrays */
+
+		child = array->fences;
+		nchild = array->num_fences;
+		GEM_BUG_ON(!nchild);
+	}
+
+	do {
+		fence = *child++;
+		if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags))
+			continue;
+
+		if (fence->context == rq->fence.context)
+			continue;
+
+		/*
+		 * We don't squash repeated fence dependencies here as we
+		 * want to run our callback in all cases.
+		 */
+
+		if (dma_fence_is_i915(fence)) {
+			if (is_same_parallel_context(rq, to_request(fence)))
+				continue;
+			ret = __i915_request_await_execution(rq,
+							     to_request(fence));
+		} else {
+			ret = i915_request_await_external(rq, fence);
+		}
+		if (ret < 0)
+			return ret;
+	} while (--nchild);
+
+	return 0;
+}
+
+static int
+await_request_submit(struct i915_request *to, struct i915_request *from)
+{
+	/*
+	 * If we are waiting on a virtual engine, then it may be
+	 * constrained to execute on a single engine *prior* to submission.
+	 * When it is submitted, it will be first submitted to the virtual
+	 * engine and then passed to the physical engine. We cannot allow
+	 * the waiter to be submitted immediately to the physical engine
+	 * as it may then bypass the virtual request.
+	 */
+	if (to->engine == READ_ONCE(from->engine))
+		return i915_sw_fence_await_sw_fence_gfp(&to->submit,
+							&from->submit,
+							I915_FENCE_GFP);
+	else
+		return __i915_request_await_execution(to, from);
+}
+
+static int
+i915_request_await_request(struct i915_request *to, struct i915_request *from)
+{
+	int ret;
+
+	GEM_BUG_ON(to == from);
+	GEM_BUG_ON(to->timeline == from->timeline);
+
+	if (i915_request_completed(from)) {
+		i915_sw_fence_set_error_once(&to->submit, from->fence.error);
+		return 0;
+	}
+
+	if (to->engine->sched_engine->schedule) {
+		ret = i915_sched_node_add_dependency(&to->sched,
+						     &from->sched,
+						     I915_DEPENDENCY_EXTERNAL);
+		if (ret < 0)
+			return ret;
+	}
+
+	if (!intel_engine_uses_guc(to->engine) &&
+	    is_power_of_2(to->execution_mask | READ_ONCE(from->execution_mask)))
+		ret = await_request_submit(to, from);
+	else
+		ret = emit_semaphore_wait(to, from, I915_FENCE_GFP);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+int
+i915_request_await_dma_fence(struct i915_request *rq, struct dma_fence *fence)
+{
+	struct dma_fence **child = &fence;
+	unsigned int nchild = 1;
+	int ret;
+
+	/*
+	 * Note that if the fence-array was created in signal-on-any mode,
+	 * we should *not* decompose it into its individual fences. However,
+	 * we don't currently store which mode the fence-array is operating
+	 * in. Fortunately, the only user of signal-on-any is private to
+	 * amdgpu and we should not see any incoming fence-array from
+	 * sync-file being in signal-on-any mode.
+	 */
+	if (dma_fence_is_array(fence)) {
+		struct dma_fence_array *array = to_dma_fence_array(fence);
+
+		child = array->fences;
+		nchild = array->num_fences;
+		GEM_BUG_ON(!nchild);
+	}
+
+	do {
+		fence = *child++;
+		if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags))
+			continue;
+
+		/*
+		 * Requests on the same timeline are explicitly ordered, along
+		 * with their dependencies, by i915_request_add() which ensures
+		 * that requests are submitted in-order through each ring.
+		 */
+		if (fence->context == rq->fence.context)
+			continue;
+
+		/* Squash repeated waits to the same timelines */
+		if (fence->context &&
+		    intel_timeline_sync_is_later(i915_request_timeline(rq),
+						 fence))
+			continue;
+
+		if (dma_fence_is_i915(fence)) {
+			if (is_same_parallel_context(rq, to_request(fence)))
+				continue;
+			ret = i915_request_await_request(rq, to_request(fence));
+		} else {
+			ret = i915_request_await_external(rq, fence);
+		}
+		if (ret < 0)
+			return ret;
+
+		/* Record the latest fence used against each timeline */
+		if (fence->context)
+			intel_timeline_sync_set(i915_request_timeline(rq),
+						fence);
+	} while (--nchild);
+
+	return 0;
+}
+
+/**
+ * i915_request_await_deps - set this request to (async) wait upon a struct
+ * i915_deps dma_fence collection
+ * @rq: request we are wishing to use
+ * @deps: The struct i915_deps containing the dependencies.
+ *
+ * Returns 0 if successful, negative error code on error.
+ */
+int i915_request_await_deps(struct i915_request *rq, const struct i915_deps *deps)
+{
+	int i, err;
+
+	for (i = 0; i < deps->num_deps; ++i) {
+		err = i915_request_await_dma_fence(rq, deps->fences[i]);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+/**
+ * i915_request_await_object - set this request to (async) wait upon a bo
+ * @to: request we are wishing to use
+ * @obj: object which may be in use on another ring.
+ * @write: whether the wait is on behalf of a writer
+ *
+ * This code is meant to abstract object synchronization with the GPU.
+ * Conceptually we serialise writes between engines inside the GPU.
+ * We only allow one engine to write into a buffer at any time, but
+ * multiple readers. To ensure each has a coherent view of memory, we must:
+ *
+ * - If there is an outstanding write request to the object, the new
+ *   request must wait for it to complete (either CPU or in hw, requests
+ *   on the same ring will be naturally ordered).
+ *
+ * - If we are a write request (pending_write_domain is set), the new
+ *   request must wait for outstanding read requests to complete.
+ *
+ * Returns 0 if successful, else propagates up the lower layer error.
+ */
+int
+i915_request_await_object(struct i915_request *to,
+			  struct drm_i915_gem_object *obj,
+			  bool write)
+{
+	struct dma_resv_iter cursor;
+	struct dma_fence *fence;
+	int ret = 0;
+
+	dma_resv_for_each_fence(&cursor, obj->base.resv,
+				dma_resv_usage_rw(write), fence) {
+		ret = i915_request_await_dma_fence(to, fence);
+		if (ret)
+			break;
+	}
+
+	return ret;
+}
+
+static struct i915_request *
+__i915_request_ensure_parallel_ordering(struct i915_request *rq,
+					struct intel_timeline *timeline)
+{
+	struct i915_request *prev;
+
+	GEM_BUG_ON(!is_parallel_rq(rq));
+
+	prev = request_to_parent(rq)->parallel.last_rq;
+	if (prev) {
+		if (!__i915_request_is_complete(prev)) {
+			i915_sw_fence_await_sw_fence(&rq->submit,
+						     &prev->submit,
+						     &rq->submitq);
+
+			if (rq->engine->sched_engine->schedule)
+				__i915_sched_node_add_dependency(&rq->sched,
+								 &prev->sched,
+								 &rq->dep,
+								 0);
+		}
+		i915_request_put(prev);
+	}
+
+	request_to_parent(rq)->parallel.last_rq = i915_request_get(rq);
+
+	/*
+	 * Users have to put a reference potentially got by
+	 * __i915_active_fence_set() to the returned request
+	 * when no longer needed
+	 */
+	return to_request(__i915_active_fence_set(&timeline->last_request,
+						  &rq->fence));
+}
+
+static struct i915_request *
+__i915_request_ensure_ordering(struct i915_request *rq,
+			       struct intel_timeline *timeline)
+{
+	struct i915_request *prev;
+
+	GEM_BUG_ON(is_parallel_rq(rq));
+
+	prev = to_request(__i915_active_fence_set(&timeline->last_request,
+						  &rq->fence));
+
+	if (prev && !__i915_request_is_complete(prev)) {
+		bool uses_guc = intel_engine_uses_guc(rq->engine);
+		bool pow2 = is_power_of_2(READ_ONCE(prev->engine)->mask |
+					  rq->engine->mask);
+		bool same_context = prev->context == rq->context;
+
+		/*
+		 * The requests are supposed to be kept in order. However,
+		 * we need to be wary in case the timeline->last_request
+		 * is used as a barrier for external modification to this
+		 * context.
+		 */
+		GEM_BUG_ON(same_context &&
+			   i915_seqno_passed(prev->fence.seqno,
+					     rq->fence.seqno));
+
+		if ((same_context && uses_guc) || (!uses_guc && pow2))
+			i915_sw_fence_await_sw_fence(&rq->submit,
+						     &prev->submit,
+						     &rq->submitq);
+		else
+			__i915_sw_fence_await_dma_fence(&rq->submit,
+							&prev->fence,
+							&rq->dmaq);
+		if (rq->engine->sched_engine->schedule)
+			__i915_sched_node_add_dependency(&rq->sched,
+							 &prev->sched,
+							 &rq->dep,
+							 0);
+	}
+
+	/*
+	 * Users have to put the reference to prev potentially got
+	 * by __i915_active_fence_set() when no longer needed
+	 */
+	return prev;
+}
+
+static struct i915_request *
+__i915_request_add_to_timeline(struct i915_request *rq)
+{
+	struct intel_timeline *timeline = i915_request_timeline(rq);
+	struct i915_request *prev;
+
+	/*
+	 * Dependency tracking and request ordering along the timeline
+	 * is special cased so that we can eliminate redundant ordering
+	 * operations while building the request (we know that the timeline
+	 * itself is ordered, and here we guarantee it).
+	 *
+	 * As we know we will need to emit tracking along the timeline,
+	 * we embed the hooks into our request struct -- at the cost of
+	 * having to have specialised no-allocation interfaces (which will
+	 * be beneficial elsewhere).
+	 *
+	 * A second benefit to open-coding i915_request_await_request is
+	 * that we can apply a slight variant of the rules specialised
+	 * for timelines that jump between engines (such as virtual engines).
+	 * If we consider the case of virtual engine, we must emit a dma-fence
+	 * to prevent scheduling of the second request until the first is
+	 * complete (to maximise our greedy late load balancing) and this
+	 * precludes optimising to use semaphores serialisation of a single
+	 * timeline across engines.
+	 *
+	 * We do not order parallel submission requests on the timeline as each
+	 * parallel submission context has its own timeline and the ordering
+	 * rules for parallel requests are that they must be submitted in the
+	 * order received from the execbuf IOCTL. So rather than using the
+	 * timeline we store a pointer to last request submitted in the
+	 * relationship in the gem context and insert a submission fence
+	 * between that request and request passed into this function or
+	 * alternatively we use completion fence if gem context has a single
+	 * timeline and this is the first submission of an execbuf IOCTL.
+	 */
+	if (likely(!is_parallel_rq(rq)))
+		prev = __i915_request_ensure_ordering(rq, timeline);
+	else
+		prev = __i915_request_ensure_parallel_ordering(rq, timeline);
+	if (prev)
+		i915_request_put(prev);
+
+	/*
+	 * Make sure that no request gazumped us - if it was allocated after
+	 * our i915_request_alloc() and called __i915_request_add() before
+	 * us, the timeline will hold its seqno which is later than ours.
+	 */
+	GEM_BUG_ON(timeline->seqno != rq->fence.seqno);
+
+	return prev;
+}
+
+/*
+ * NB: This function is not allowed to fail. Doing so would mean the the
+ * request is not being tracked for completion but the work itself is
+ * going to happen on the hardware. This would be a Bad Thing(tm).
+ */
+struct i915_request *__i915_request_commit(struct i915_request *rq)
+{
+	struct intel_engine_cs *engine = rq->engine;
+	struct intel_ring *ring = rq->ring;
+	u32 *cs;
+
+	RQ_TRACE(rq, "\n");
+
+	/*
+	 * To ensure that this call will not fail, space for its emissions
+	 * should already have been reserved in the ring buffer. Let the ring
+	 * know that it is time to use that space up.
+	 */
+	GEM_BUG_ON(rq->reserved_space > ring->space);
+	rq->reserved_space = 0;
+	rq->emitted_jiffies = jiffies;
+
+	/*
+	 * Record the position of the start of the breadcrumb so that
+	 * should we detect the updated seqno part-way through the
+	 * GPU processing the request, we never over-estimate the
+	 * position of the ring's HEAD.
+	 */
+	cs = intel_ring_begin(rq, engine->emit_fini_breadcrumb_dw);
+	GEM_BUG_ON(IS_ERR(cs));
+	rq->postfix = intel_ring_offset(rq, cs);
+
+	return __i915_request_add_to_timeline(rq);
+}
+
+void __i915_request_queue_bh(struct i915_request *rq)
+{
+	i915_sw_fence_commit(&rq->semaphore);
+	i915_sw_fence_commit(&rq->submit);
+}
+
+void __i915_request_queue(struct i915_request *rq,
+			  const struct i915_sched_attr *attr)
+{
+	/*
+	 * Let the backend know a new request has arrived that may need
+	 * to adjust the existing execution schedule due to a high priority
+	 * request - i.e. we may want to preempt the current request in order
+	 * to run a high priority dependency chain *before* we can execute this
+	 * request.
+	 *
+	 * This is called before the request is ready to run so that we can
+	 * decide whether to preempt the entire chain so that it is ready to
+	 * run at the earliest possible convenience.
+	 */
+	if (attr && rq->engine->sched_engine->schedule)
+		rq->engine->sched_engine->schedule(rq, attr);
+
+	local_bh_disable();
+	__i915_request_queue_bh(rq);
+	local_bh_enable(); /* kick tasklets */
+}
+
+void i915_request_add(struct i915_request *rq)
+{
+	struct intel_timeline * const tl = i915_request_timeline(rq);
+	struct i915_sched_attr attr = {};
+	struct i915_gem_context *ctx;
+
+	lockdep_assert_held(&tl->mutex);
+	lockdep_unpin_lock(&tl->mutex, rq->cookie);
+
+	trace_i915_request_add(rq);
+	__i915_request_commit(rq);
+
+	/* XXX placeholder for selftests */
+	rcu_read_lock();
+	ctx = rcu_dereference(rq->context->gem_context);
+	if (ctx)
+		attr = ctx->sched;
+	rcu_read_unlock();
+
+	__i915_request_queue(rq, &attr);
+
+	mutex_unlock(&tl->mutex);
+}
+
+static unsigned long local_clock_ns(unsigned int *cpu)
+{
+	unsigned long t;
+
+	/*
+	 * Cheaply and approximately convert from nanoseconds to microseconds.
+	 * The result and subsequent calculations are also defined in the same
+	 * approximate microseconds units. The principal source of timing
+	 * error here is from the simple truncation.
+	 *
+	 * Note that local_clock() is only defined wrt to the current CPU;
+	 * the comparisons are no longer valid if we switch CPUs. Instead of
+	 * blocking preemption for the entire busywait, we can detect the CPU
+	 * switch and use that as indicator of system load and a reason to
+	 * stop busywaiting, see busywait_stop().
+	 */
+	*cpu = get_cpu();
+	t = local_clock();
+	put_cpu();
+
+	return t;
+}
+
+static bool busywait_stop(unsigned long timeout, unsigned int cpu)
+{
+	unsigned int this_cpu;
+
+	if (time_after(local_clock_ns(&this_cpu), timeout))
+		return true;
+
+	return this_cpu != cpu;
+}
+
+static bool __i915_spin_request(struct i915_request * const rq, int state)
+{
+	unsigned long timeout_ns;
+	unsigned int cpu;
+
+	/*
+	 * Only wait for the request if we know it is likely to complete.
+	 *
+	 * We don't track the timestamps around requests, nor the average
+	 * request length, so we do not have a good indicator that this
+	 * request will complete within the timeout. What we do know is the
+	 * order in which requests are executed by the context and so we can
+	 * tell if the request has been started. If the request is not even
+	 * running yet, it is a fair assumption that it will not complete
+	 * within our relatively short timeout.
+	 */
+	if (!i915_request_is_running(rq))
+		return false;
+
+	/*
+	 * When waiting for high frequency requests, e.g. during synchronous
+	 * rendering split between the CPU and GPU, the finite amount of time
+	 * required to set up the irq and wait upon it limits the response
+	 * rate. By busywaiting on the request completion for a short while we
+	 * can service the high frequency waits as quick as possible. However,
+	 * if it is a slow request, we want to sleep as quickly as possible.
+	 * The tradeoff between waiting and sleeping is roughly the time it
+	 * takes to sleep on a request, on the order of a microsecond.
+	 */
+
+	timeout_ns = READ_ONCE(rq->engine->props.max_busywait_duration_ns);
+	timeout_ns += local_clock_ns(&cpu);
+	do {
+		if (dma_fence_is_signaled(&rq->fence))
+			return true;
+
+		if (signal_pending_state(state, current))
+			break;
+
+		if (busywait_stop(timeout_ns, cpu))
+			break;
+
+		cpu_relax();
+	} while (!need_resched());
+
+	return false;
+}
+
+struct request_wait {
+	struct dma_fence_cb cb;
+	struct task_struct *tsk;
+};
+
+static void request_wait_wake(struct dma_fence *fence, struct dma_fence_cb *cb)
+{
+	struct request_wait *wait = container_of(cb, typeof(*wait), cb);
+
+	wake_up_process(fetch_and_zero(&wait->tsk));
+}
+
+/**
+ * i915_request_wait_timeout - wait until execution of request has finished
+ * @rq: the request to wait upon
+ * @flags: how to wait
+ * @timeout: how long to wait in jiffies
+ *
+ * i915_request_wait_timeout() waits for the request to be completed, for a
+ * maximum of @timeout jiffies (with MAX_SCHEDULE_TIMEOUT implying an
+ * unbounded wait).
+ *
+ * Returns the remaining time (in jiffies) if the request completed, which may
+ * be zero if the request is unfinished after the timeout expires.
+ * If the timeout is 0, it will return 1 if the fence is signaled.
+ *
+ * May return -EINTR is called with I915_WAIT_INTERRUPTIBLE and a signal is
+ * pending before the request completes.
+ *
+ * NOTE: This function has the same wait semantics as dma-fence.
+ */
+long i915_request_wait_timeout(struct i915_request *rq,
+			       unsigned int flags,
+			       long timeout)
+{
+	const int state = flags & I915_WAIT_INTERRUPTIBLE ?
+		TASK_INTERRUPTIBLE : TASK_UNINTERRUPTIBLE;
+	struct request_wait wait;
+
+	might_sleep();
+	GEM_BUG_ON(timeout < 0);
+
+	if (dma_fence_is_signaled(&rq->fence))
+		return timeout ?: 1;
+
+	if (!timeout)
+		return -ETIME;
+
+	trace_i915_request_wait_begin(rq, flags);
+
+	/*
+	 * We must never wait on the GPU while holding a lock as we
+	 * may need to perform a GPU reset. So while we don't need to
+	 * serialise wait/reset with an explicit lock, we do want
+	 * lockdep to detect potential dependency cycles.
+	 */
+	mutex_acquire(&rq->engine->gt->reset.mutex.dep_map, 0, 0, _THIS_IP_);
+
+	/*
+	 * Optimistic spin before touching IRQs.
+	 *
+	 * We may use a rather large value here to offset the penalty of
+	 * switching away from the active task. Frequently, the client will
+	 * wait upon an old swapbuffer to throttle itself to remain within a
+	 * frame of the gpu. If the client is running in lockstep with the gpu,
+	 * then it should not be waiting long at all, and a sleep now will incur
+	 * extra scheduler latency in producing the next frame. To try to
+	 * avoid adding the cost of enabling/disabling the interrupt to the
+	 * short wait, we first spin to see if the request would have completed
+	 * in the time taken to setup the interrupt.
+	 *
+	 * We need upto 5us to enable the irq, and upto 20us to hide the
+	 * scheduler latency of a context switch, ignoring the secondary
+	 * impacts from a context switch such as cache eviction.
+	 *
+	 * The scheme used for low-latency IO is called "hybrid interrupt
+	 * polling". The suggestion there is to sleep until just before you
+	 * expect to be woken by the device interrupt and then poll for its
+	 * completion. That requires having a good predictor for the request
+	 * duration, which we currently lack.
+	 */
+	if (CONFIG_DRM_I915_MAX_REQUEST_BUSYWAIT &&
+	    __i915_spin_request(rq, state))
+		goto out;
+
+	/*
+	 * This client is about to stall waiting for the GPU. In many cases
+	 * this is undesirable and limits the throughput of the system, as
+	 * many clients cannot continue processing user input/output whilst
+	 * blocked. RPS autotuning may take tens of milliseconds to respond
+	 * to the GPU load and thus incurs additional latency for the client.
+	 * We can circumvent that by promoting the GPU frequency to maximum
+	 * before we sleep. This makes the GPU throttle up much more quickly
+	 * (good for benchmarks and user experience, e.g. window animations),
+	 * but at a cost of spending more power processing the workload
+	 * (bad for battery).
+	 */
+	if (flags & I915_WAIT_PRIORITY && !i915_request_started(rq))
+		intel_rps_boost(rq);
+
+	wait.tsk = current;
+	if (dma_fence_add_callback(&rq->fence, &wait.cb, request_wait_wake))
+		goto out;
+
+	/*
+	 * Flush the submission tasklet, but only if it may help this request.
+	 *
+	 * We sometimes experience some latency between the HW interrupts and
+	 * tasklet execution (mostly due to ksoftirqd latency, but it can also
+	 * be due to lazy CS events), so lets run the tasklet manually if there
+	 * is a chance it may submit this request. If the request is not ready
+	 * to run, as it is waiting for other fences to be signaled, flushing
+	 * the tasklet is busy work without any advantage for this client.
+	 *
+	 * If the HW is being lazy, this is the last chance before we go to
+	 * sleep to catch any pending events. We will check periodically in
+	 * the heartbeat to flush the submission tasklets as a last resort
+	 * for unhappy HW.
+	 */
+	if (i915_request_is_ready(rq))
+		__intel_engine_flush_submission(rq->engine, false);
+
+	for (;;) {
+		set_current_state(state);
+
+		if (dma_fence_is_signaled(&rq->fence))
+			break;
+
+		if (signal_pending_state(state, current)) {
+			timeout = -ERESTARTSYS;
+			break;
+		}
+
+		if (!timeout) {
+			timeout = -ETIME;
+			break;
+		}
+
+		timeout = io_schedule_timeout(timeout);
+	}
+	__set_current_state(TASK_RUNNING);
+
+	if (READ_ONCE(wait.tsk))
+		dma_fence_remove_callback(&rq->fence, &wait.cb);
+	GEM_BUG_ON(!list_empty(&wait.cb.node));
+
+out:
+	mutex_release(&rq->engine->gt->reset.mutex.dep_map, _THIS_IP_);
+	trace_i915_request_wait_end(rq);
+	return timeout;
+}
+
+/**
+ * i915_request_wait - wait until execution of request has finished
+ * @rq: the request to wait upon
+ * @flags: how to wait
+ * @timeout: how long to wait in jiffies
+ *
+ * i915_request_wait() waits for the request to be completed, for a
+ * maximum of @timeout jiffies (with MAX_SCHEDULE_TIMEOUT implying an
+ * unbounded wait).
+ *
+ * Returns the remaining time (in jiffies) if the request completed, which may
+ * be zero or -ETIME if the request is unfinished after the timeout expires.
+ * May return -EINTR is called with I915_WAIT_INTERRUPTIBLE and a signal is
+ * pending before the request completes.
+ *
+ * NOTE: This function behaves differently from dma-fence wait semantics for
+ * timeout = 0. It returns 0 on success, and -ETIME if not signaled.
+ */
+long i915_request_wait(struct i915_request *rq,
+		       unsigned int flags,
+		       long timeout)
+{
+	long ret = i915_request_wait_timeout(rq, flags, timeout);
+
+	if (!ret)
+		return -ETIME;
+
+	if (ret > 0 && !timeout)
+		return 0;
+
+	return ret;
+}
+
+static int print_sched_attr(const struct i915_sched_attr *attr,
+			    char *buf, int x, int len)
+{
+	if (attr->priority == I915_PRIORITY_INVALID)
+		return x;
+
+	x += snprintf(buf + x, len - x,
+		      " prio=%d", attr->priority);
+
+	return x;
+}
+
+static char queue_status(const struct i915_request *rq)
+{
+	if (i915_request_is_active(rq))
+		return 'E';
+
+	if (i915_request_is_ready(rq))
+		return intel_engine_is_virtual(rq->engine) ? 'V' : 'R';
+
+	return 'U';
+}
+
+static const char *run_status(const struct i915_request *rq)
+{
+	if (__i915_request_is_complete(rq))
+		return "!";
+
+	if (__i915_request_has_started(rq))
+		return "*";
+
+	if (!i915_sw_fence_signaled(&rq->semaphore))
+		return "&";
+
+	return "";
+}
+
+static const char *fence_status(const struct i915_request *rq)
+{
+	if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags))
+		return "+";
+
+	if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &rq->fence.flags))
+		return "-";
+
+	return "";
+}
+
+void i915_request_show(struct drm_printer *m,
+		       const struct i915_request *rq,
+		       const char *prefix,
+		       int indent)
+{
+	const char *name = rq->fence.ops->get_timeline_name((struct dma_fence *)&rq->fence);
+	char buf[80] = "";
+	int x = 0;
+
+	/*
+	 * The prefix is used to show the queue status, for which we use
+	 * the following flags:
+	 *
+	 *  U [Unready]
+	 *    - initial status upon being submitted by the user
+	 *
+	 *    - the request is not ready for execution as it is waiting
+	 *      for external fences
+	 *
+	 *  R [Ready]
+	 *    - all fences the request was waiting on have been signaled,
+	 *      and the request is now ready for execution and will be
+	 *      in a backend queue
+	 *
+	 *    - a ready request may still need to wait on semaphores
+	 *      [internal fences]
+	 *
+	 *  V [Ready/virtual]
+	 *    - same as ready, but queued over multiple backends
+	 *
+	 *  E [Executing]
+	 *    - the request has been transferred from the backend queue and
+	 *      submitted for execution on HW
+	 *
+	 *    - a completed request may still be regarded as executing, its
+	 *      status may not be updated until it is retired and removed
+	 *      from the lists
+	 */
+
+	x = print_sched_attr(&rq->sched.attr, buf, x, sizeof(buf));
+
+	drm_printf(m, "%s%.*s%c %llx:%lld%s%s %s @ %dms: %s\n",
+		   prefix, indent, "                ",
+		   queue_status(rq),
+		   rq->fence.context, rq->fence.seqno,
+		   run_status(rq),
+		   fence_status(rq),
+		   buf,
+		   jiffies_to_msecs(jiffies - rq->emitted_jiffies),
+		   name);
+}
+
+static bool engine_match_ring(struct intel_engine_cs *engine, struct i915_request *rq)
+{
+	u32 ring = ENGINE_READ(engine, RING_START);
+
+	return ring == i915_ggtt_offset(rq->ring->vma);
+}
+
+static bool match_ring(struct i915_request *rq)
+{
+	struct intel_engine_cs *engine;
+	bool found;
+	int i;
+
+	if (!intel_engine_is_virtual(rq->engine))
+		return engine_match_ring(rq->engine, rq);
+
+	found = false;
+	i = 0;
+	while ((engine = intel_engine_get_sibling(rq->engine, i++))) {
+		found = engine_match_ring(engine, rq);
+		if (found)
+			break;
+	}
+
+	return found;
+}
+
+enum i915_request_state i915_test_request_state(struct i915_request *rq)
+{
+	if (i915_request_completed(rq))
+		return I915_REQUEST_COMPLETE;
+
+	if (!i915_request_started(rq))
+		return I915_REQUEST_PENDING;
+
+	if (match_ring(rq))
+		return I915_REQUEST_ACTIVE;
+
+	return I915_REQUEST_QUEUED;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_request.c"
+#include "selftests/i915_request.c"
+#endif
+
+void i915_request_module_exit(void)
+{
+	kmem_cache_destroy(slab_execute_cbs);
+	kmem_cache_destroy(slab_requests);
+}
+
+int __init i915_request_module_init(void)
+{
+	slab_requests =
+		kmem_cache_create("i915_request",
+				  sizeof(struct i915_request),
+				  __alignof__(struct i915_request),
+				  SLAB_HWCACHE_ALIGN |
+				  SLAB_RECLAIM_ACCOUNT |
+				  SLAB_TYPESAFE_BY_RCU,
+				  __i915_request_ctor);
+	if (!slab_requests)
+		return -ENOMEM;
+
+	slab_execute_cbs = KMEM_CACHE(execute_cb,
+					     SLAB_HWCACHE_ALIGN |
+					     SLAB_RECLAIM_ACCOUNT |
+					     SLAB_TYPESAFE_BY_RCU);
+	if (!slab_execute_cbs)
+		goto err_requests;
+
+	return 0;
+
+err_requests:
+	kmem_cache_destroy(slab_requests);
+	return -ENOMEM;
+}
diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h
new file mode 100644
index 000000000..47041ec68
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_request.h
@@ -0,0 +1,727 @@
+/*
+ * Copyright © 2008-2018 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef I915_REQUEST_H
+#define I915_REQUEST_H
+
+#include <linux/dma-fence.h>
+#include <linux/hrtimer.h>
+#include <linux/irq_work.h>
+#include <linux/llist.h>
+#include <linux/lockdep.h>
+
+#include "gem/i915_gem_context_types.h"
+#include "gt/intel_context_types.h"
+#include "gt/intel_engine_types.h"
+#include "gt/intel_timeline_types.h"
+
+#include "i915_gem.h"
+#include "i915_scheduler.h"
+#include "i915_selftest.h"
+#include "i915_sw_fence.h"
+#include "i915_vma_resource.h"
+
+#include <uapi/drm/i915_drm.h>
+
+struct drm_file;
+struct drm_i915_gem_object;
+struct drm_printer;
+struct i915_deps;
+struct i915_request;
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+struct i915_capture_list {
+	struct i915_vma_resource *vma_res;
+	struct i915_capture_list *next;
+};
+
+void i915_request_free_capture_list(struct i915_capture_list *capture);
+#else
+#define i915_request_free_capture_list(_a) do {} while (0)
+#endif
+
+#define RQ_TRACE(rq, fmt, ...) do {					\
+	const struct i915_request *rq__ = (rq);				\
+	ENGINE_TRACE(rq__->engine, "fence %llx:%lld, current %d " fmt,	\
+		     rq__->fence.context, rq__->fence.seqno,		\
+		     hwsp_seqno(rq__), ##__VA_ARGS__);			\
+} while (0)
+
+enum {
+	/*
+	 * I915_FENCE_FLAG_ACTIVE - this request is currently submitted to HW.
+	 *
+	 * Set by __i915_request_submit() on handing over to HW, and cleared
+	 * by __i915_request_unsubmit() if we preempt this request.
+	 *
+	 * Finally cleared for consistency on retiring the request, when
+	 * we know the HW is no longer running this request.
+	 *
+	 * See i915_request_is_active()
+	 */
+	I915_FENCE_FLAG_ACTIVE = DMA_FENCE_FLAG_USER_BITS,
+
+	/*
+	 * I915_FENCE_FLAG_PQUEUE - this request is ready for execution
+	 *
+	 * Using the scheduler, when a request is ready for execution it is put
+	 * into the priority queue, and removed from that queue when transferred
+	 * to the HW runlists. We want to track its membership within the
+	 * priority queue so that we can easily check before rescheduling.
+	 *
+	 * See i915_request_in_priority_queue()
+	 */
+	I915_FENCE_FLAG_PQUEUE,
+
+	/*
+	 * I915_FENCE_FLAG_HOLD - this request is currently on hold
+	 *
+	 * This request has been suspended, pending an ongoing investigation.
+	 */
+	I915_FENCE_FLAG_HOLD,
+
+	/*
+	 * I915_FENCE_FLAG_INITIAL_BREADCRUMB - this request has the initial
+	 * breadcrumb that marks the end of semaphore waits and start of the
+	 * user payload.
+	 */
+	I915_FENCE_FLAG_INITIAL_BREADCRUMB,
+
+	/*
+	 * I915_FENCE_FLAG_SIGNAL - this request is currently on signal_list
+	 *
+	 * Internal bookkeeping used by the breadcrumb code to track when
+	 * a request is on the various signal_list.
+	 */
+	I915_FENCE_FLAG_SIGNAL,
+
+	/*
+	 * I915_FENCE_FLAG_NOPREEMPT - this request should not be preempted
+	 *
+	 * The execution of some requests should not be interrupted. This is
+	 * a sensitive operation as it makes the request super important,
+	 * blocking other higher priority work. Abuse of this flag will
+	 * lead to quality of service issues.
+	 */
+	I915_FENCE_FLAG_NOPREEMPT,
+
+	/*
+	 * I915_FENCE_FLAG_SENTINEL - this request should be last in the queue
+	 *
+	 * A high priority sentinel request may be submitted to clear the
+	 * submission queue. As it will be the only request in-flight, upon
+	 * execution all other active requests will have been preempted and
+	 * unsubmitted. This preemptive pulse is used to re-evaluate the
+	 * in-flight requests, particularly in cases where an active context
+	 * is banned and those active requests need to be cancelled.
+	 */
+	I915_FENCE_FLAG_SENTINEL,
+
+	/*
+	 * I915_FENCE_FLAG_BOOST - upclock the gpu for this request
+	 *
+	 * Some requests are more important than others! In particular, a
+	 * request that the user is waiting on is typically required for
+	 * interactive latency, for which we want to minimise by upclocking
+	 * the GPU. Here we track such boost requests on a per-request basis.
+	 */
+	I915_FENCE_FLAG_BOOST,
+
+	/*
+	 * I915_FENCE_FLAG_SUBMIT_PARALLEL - request with a context in a
+	 * parent-child relationship (parallel submission, multi-lrc) should
+	 * trigger a submission to the GuC rather than just moving the context
+	 * tail.
+	 */
+	I915_FENCE_FLAG_SUBMIT_PARALLEL,
+
+	/*
+	 * I915_FENCE_FLAG_SKIP_PARALLEL - request with a context in a
+	 * parent-child relationship (parallel submission, multi-lrc) that
+	 * hit an error while generating requests in the execbuf IOCTL.
+	 * Indicates this request should be skipped as another request in
+	 * submission / relationship encoutered an error.
+	 */
+	I915_FENCE_FLAG_SKIP_PARALLEL,
+
+	/*
+	 * I915_FENCE_FLAG_COMPOSITE - Indicates fence is part of a composite
+	 * fence (dma_fence_array) and i915 generated for parallel submission.
+	 */
+	I915_FENCE_FLAG_COMPOSITE,
+};
+
+/**
+ * Request queue structure.
+ *
+ * The request queue allows us to note sequence numbers that have been emitted
+ * and may be associated with active buffers to be retired.
+ *
+ * By keeping this list, we can avoid having to do questionable sequence
+ * number comparisons on buffer last_read|write_seqno. It also allows an
+ * emission time to be associated with the request for tracking how far ahead
+ * of the GPU the submission is.
+ *
+ * When modifying this structure be very aware that we perform a lockless
+ * RCU lookup of it that may race against reallocation of the struct
+ * from the slab freelist. We intentionally do not zero the structure on
+ * allocation so that the lookup can use the dangling pointers (and is
+ * cogniscent that those pointers may be wrong). Instead, everything that
+ * needs to be initialised must be done so explicitly.
+ *
+ * The requests are reference counted.
+ */
+struct i915_request {
+	struct dma_fence fence;
+	spinlock_t lock;
+
+	struct drm_i915_private *i915;
+
+	/**
+	 * Context and ring buffer related to this request
+	 * Contexts are refcounted, so when this request is associated with a
+	 * context, we must increment the context's refcount, to guarantee that
+	 * it persists while any request is linked to it. Requests themselves
+	 * are also refcounted, so the request will only be freed when the last
+	 * reference to it is dismissed, and the code in
+	 * i915_request_free() will then decrement the refcount on the
+	 * context.
+	 */
+	struct intel_engine_cs *engine;
+	struct intel_context *context;
+	struct intel_ring *ring;
+	struct intel_timeline __rcu *timeline;
+
+	struct list_head signal_link;
+	struct llist_node signal_node;
+
+	/*
+	 * The rcu epoch of when this request was allocated. Used to judiciously
+	 * apply backpressure on future allocations to ensure that under
+	 * mempressure there is sufficient RCU ticks for us to reclaim our
+	 * RCU protected slabs.
+	 */
+	unsigned long rcustate;
+
+	/*
+	 * We pin the timeline->mutex while constructing the request to
+	 * ensure that no caller accidentally drops it during construction.
+	 * The timeline->mutex must be held to ensure that only this caller
+	 * can use the ring and manipulate the associated timeline during
+	 * construction.
+	 */
+	struct pin_cookie cookie;
+
+	/*
+	 * Fences for the various phases in the request's lifetime.
+	 *
+	 * The submit fence is used to await upon all of the request's
+	 * dependencies. When it is signaled, the request is ready to run.
+	 * It is used by the driver to then queue the request for execution.
+	 */
+	struct i915_sw_fence submit;
+	union {
+		wait_queue_entry_t submitq;
+		struct i915_sw_dma_fence_cb dmaq;
+		struct i915_request_duration_cb {
+			struct dma_fence_cb cb;
+			ktime_t emitted;
+		} duration;
+	};
+	struct llist_head execute_cb;
+	struct i915_sw_fence semaphore;
+	/**
+	 * @submit_work: complete submit fence from an IRQ if needed for
+	 * locking hierarchy reasons.
+	 */
+	struct irq_work submit_work;
+
+	/*
+	 * A list of everyone we wait upon, and everyone who waits upon us.
+	 * Even though we will not be submitted to the hardware before the
+	 * submit fence is signaled (it waits for all external events as well
+	 * as our own requests), the scheduler still needs to know the
+	 * dependency tree for the lifetime of the request (from execbuf
+	 * to retirement), i.e. bidirectional dependency information for the
+	 * request not tied to individual fences.
+	 */
+	struct i915_sched_node sched;
+	struct i915_dependency dep;
+	intel_engine_mask_t execution_mask;
+
+	/*
+	 * A convenience pointer to the current breadcrumb value stored in
+	 * the HW status page (or our timeline's local equivalent). The full
+	 * path would be rq->hw_context->ring->timeline->hwsp_seqno.
+	 */
+	const u32 *hwsp_seqno;
+
+	/** Position in the ring of the start of the request */
+	u32 head;
+
+	/** Position in the ring of the start of the user packets */
+	u32 infix;
+
+	/**
+	 * Position in the ring of the start of the postfix.
+	 * This is required to calculate the maximum available ring space
+	 * without overwriting the postfix.
+	 */
+	u32 postfix;
+
+	/** Position in the ring of the end of the whole request */
+	u32 tail;
+
+	/** Position in the ring of the end of any workarounds after the tail */
+	u32 wa_tail;
+
+	/** Preallocate space in the ring for the emitting the request */
+	u32 reserved_space;
+
+	/** Batch buffer pointer for selftest internal use. */
+	I915_SELFTEST_DECLARE(struct i915_vma *batch);
+
+	struct i915_vma_resource *batch_res;
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+	/**
+	 * Additional buffers requested by userspace to be captured upon
+	 * a GPU hang. The vma/obj on this list are protected by their
+	 * active reference - all objects on this list must also be
+	 * on the active_list (of their final request).
+	 */
+	struct i915_capture_list *capture_list;
+#endif
+
+	/** Time at which this request was emitted, in jiffies. */
+	unsigned long emitted_jiffies;
+
+	/** timeline->request entry for this request */
+	struct list_head link;
+
+	/** Watchdog support fields. */
+	struct i915_request_watchdog {
+		struct llist_node link;
+		struct hrtimer timer;
+	} watchdog;
+
+	/**
+	 * @guc_fence_link: Requests may need to be stalled when using GuC
+	 * submission waiting for certain GuC operations to complete. If that is
+	 * the case, stalled requests are added to a per context list of stalled
+	 * requests. The below list_head is the link in that list. Protected by
+	 * ce->guc_state.lock.
+	 */
+	struct list_head guc_fence_link;
+
+	/**
+	 * @guc_prio: Priority level while the request is in flight. Differs
+	 * from i915 scheduler priority. See comment above
+	 * I915_SCHEDULER_CAP_STATIC_PRIORITY_MAP for details. Protected by
+	 * ce->guc_active.lock. Two special values (GUC_PRIO_INIT and
+	 * GUC_PRIO_FINI) outside the GuC priority range are used to indicate
+	 * if the priority has not been initialized yet or if no more updates
+	 * are possible because the request has completed.
+	 */
+#define	GUC_PRIO_INIT	0xff
+#define	GUC_PRIO_FINI	0xfe
+	u8 guc_prio;
+
+	I915_SELFTEST_DECLARE(struct {
+		struct list_head link;
+		unsigned long delay;
+	} mock;)
+};
+
+#define I915_FENCE_GFP (GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN)
+
+extern const struct dma_fence_ops i915_fence_ops;
+
+static inline bool dma_fence_is_i915(const struct dma_fence *fence)
+{
+	return fence->ops == &i915_fence_ops;
+}
+
+struct kmem_cache *i915_request_slab_cache(void);
+
+struct i915_request * __must_check
+__i915_request_create(struct intel_context *ce, gfp_t gfp);
+struct i915_request * __must_check
+i915_request_create(struct intel_context *ce);
+
+void __i915_request_skip(struct i915_request *rq);
+bool i915_request_set_error_once(struct i915_request *rq, int error);
+struct i915_request *i915_request_mark_eio(struct i915_request *rq);
+
+struct i915_request *__i915_request_commit(struct i915_request *request);
+void __i915_request_queue(struct i915_request *rq,
+			  const struct i915_sched_attr *attr);
+void __i915_request_queue_bh(struct i915_request *rq);
+
+bool i915_request_retire(struct i915_request *rq);
+void i915_request_retire_upto(struct i915_request *rq);
+
+static inline struct i915_request *
+to_request(struct dma_fence *fence)
+{
+	/* We assume that NULL fence/request are interoperable */
+	BUILD_BUG_ON(offsetof(struct i915_request, fence) != 0);
+	GEM_BUG_ON(fence && !dma_fence_is_i915(fence));
+	return container_of(fence, struct i915_request, fence);
+}
+
+static inline struct i915_request *
+i915_request_get(struct i915_request *rq)
+{
+	return to_request(dma_fence_get(&rq->fence));
+}
+
+static inline struct i915_request *
+i915_request_get_rcu(struct i915_request *rq)
+{
+	return to_request(dma_fence_get_rcu(&rq->fence));
+}
+
+static inline void
+i915_request_put(struct i915_request *rq)
+{
+	dma_fence_put(&rq->fence);
+}
+
+int i915_request_await_object(struct i915_request *to,
+			      struct drm_i915_gem_object *obj,
+			      bool write);
+int i915_request_await_dma_fence(struct i915_request *rq,
+				 struct dma_fence *fence);
+int i915_request_await_deps(struct i915_request *rq, const struct i915_deps *deps);
+int i915_request_await_execution(struct i915_request *rq,
+				 struct dma_fence *fence);
+
+void i915_request_add(struct i915_request *rq);
+
+bool __i915_request_submit(struct i915_request *request);
+void i915_request_submit(struct i915_request *request);
+
+void __i915_request_unsubmit(struct i915_request *request);
+void i915_request_unsubmit(struct i915_request *request);
+
+void i915_request_cancel(struct i915_request *rq, int error);
+
+long i915_request_wait_timeout(struct i915_request *rq,
+			       unsigned int flags,
+			       long timeout)
+	__attribute__((nonnull(1)));
+
+long i915_request_wait(struct i915_request *rq,
+		       unsigned int flags,
+		       long timeout)
+	__attribute__((nonnull(1)));
+#define I915_WAIT_INTERRUPTIBLE	BIT(0)
+#define I915_WAIT_PRIORITY	BIT(1) /* small priority bump for the request */
+#define I915_WAIT_ALL		BIT(2) /* used by i915_gem_object_wait() */
+
+void i915_request_show(struct drm_printer *m,
+		       const struct i915_request *rq,
+		       const char *prefix,
+		       int indent);
+
+static inline bool i915_request_signaled(const struct i915_request *rq)
+{
+	/* The request may live longer than its HWSP, so check flags first! */
+	return test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags);
+}
+
+static inline bool i915_request_is_active(const struct i915_request *rq)
+{
+	return test_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags);
+}
+
+static inline bool i915_request_in_priority_queue(const struct i915_request *rq)
+{
+	return test_bit(I915_FENCE_FLAG_PQUEUE, &rq->fence.flags);
+}
+
+static inline bool
+i915_request_has_initial_breadcrumb(const struct i915_request *rq)
+{
+	return test_bit(I915_FENCE_FLAG_INITIAL_BREADCRUMB, &rq->fence.flags);
+}
+
+/**
+ * Returns true if seq1 is later than seq2.
+ */
+static inline bool i915_seqno_passed(u32 seq1, u32 seq2)
+{
+	return (s32)(seq1 - seq2) >= 0;
+}
+
+static inline u32 __hwsp_seqno(const struct i915_request *rq)
+{
+	const u32 *hwsp = READ_ONCE(rq->hwsp_seqno);
+
+	return READ_ONCE(*hwsp);
+}
+
+/**
+ * hwsp_seqno - the current breadcrumb value in the HW status page
+ * @rq: the request, to chase the relevant HW status page
+ *
+ * The emphasis in naming here is that hwsp_seqno() is not a property of the
+ * request, but an indication of the current HW state (associated with this
+ * request). Its value will change as the GPU executes more requests.
+ *
+ * Returns the current breadcrumb value in the associated HW status page (or
+ * the local timeline's equivalent) for this request. The request itself
+ * has the associated breadcrumb value of rq->fence.seqno, when the HW
+ * status page has that breadcrumb or later, this request is complete.
+ */
+static inline u32 hwsp_seqno(const struct i915_request *rq)
+{
+	u32 seqno;
+
+	rcu_read_lock(); /* the HWSP may be freed at runtime */
+	seqno = __hwsp_seqno(rq);
+	rcu_read_unlock();
+
+	return seqno;
+}
+
+static inline bool __i915_request_has_started(const struct i915_request *rq)
+{
+	return i915_seqno_passed(__hwsp_seqno(rq), rq->fence.seqno - 1);
+}
+
+/**
+ * i915_request_started - check if the request has begun being executed
+ * @rq: the request
+ *
+ * If the timeline is not using initial breadcrumbs, a request is
+ * considered started if the previous request on its timeline (i.e.
+ * context) has been signaled.
+ *
+ * If the timeline is using semaphores, it will also be emitting an
+ * "initial breadcrumb" after the semaphores are complete and just before
+ * it began executing the user payload. A request can therefore be active
+ * on the HW and not yet started as it is still busywaiting on its
+ * dependencies (via HW semaphores).
+ *
+ * If the request has started, its dependencies will have been signaled
+ * (either by fences or by semaphores) and it will have begun processing
+ * the user payload.
+ *
+ * However, even if a request has started, it may have been preempted and
+ * so no longer active, or it may have already completed.
+ *
+ * See also i915_request_is_active().
+ *
+ * Returns true if the request has begun executing the user payload, or
+ * has completed:
+ */
+static inline bool i915_request_started(const struct i915_request *rq)
+{
+	bool result;
+
+	if (i915_request_signaled(rq))
+		return true;
+
+	result = true;
+	rcu_read_lock(); /* the HWSP may be freed at runtime */
+	if (likely(!i915_request_signaled(rq)))
+		/* Remember: started but may have since been preempted! */
+		result = __i915_request_has_started(rq);
+	rcu_read_unlock();
+
+	return result;
+}
+
+/**
+ * i915_request_is_running - check if the request may actually be executing
+ * @rq: the request
+ *
+ * Returns true if the request is currently submitted to hardware, has passed
+ * its start point (i.e. the context is setup and not busywaiting). Note that
+ * it may no longer be running by the time the function returns!
+ */
+static inline bool i915_request_is_running(const struct i915_request *rq)
+{
+	bool result;
+
+	if (!i915_request_is_active(rq))
+		return false;
+
+	rcu_read_lock();
+	result = __i915_request_has_started(rq) && i915_request_is_active(rq);
+	rcu_read_unlock();
+
+	return result;
+}
+
+/**
+ * i915_request_is_ready - check if the request is ready for execution
+ * @rq: the request
+ *
+ * Upon construction, the request is instructed to wait upon various
+ * signals before it is ready to be executed by the HW. That is, we do
+ * not want to start execution and read data before it is written. In practice,
+ * this is controlled with a mixture of interrupts and semaphores. Once
+ * the submit fence is completed, the backend scheduler will place the
+ * request into its queue and from there submit it for execution. So we
+ * can detect when a request is eligible for execution (and is under control
+ * of the scheduler) by querying where it is in any of the scheduler's lists.
+ *
+ * Returns true if the request is ready for execution (it may be inflight),
+ * false otherwise.
+ */
+static inline bool i915_request_is_ready(const struct i915_request *rq)
+{
+	return !list_empty(&rq->sched.link);
+}
+
+static inline bool __i915_request_is_complete(const struct i915_request *rq)
+{
+	return i915_seqno_passed(__hwsp_seqno(rq), rq->fence.seqno);
+}
+
+static inline bool i915_request_completed(const struct i915_request *rq)
+{
+	bool result;
+
+	if (i915_request_signaled(rq))
+		return true;
+
+	result = true;
+	rcu_read_lock(); /* the HWSP may be freed at runtime */
+	if (likely(!i915_request_signaled(rq)))
+		result = __i915_request_is_complete(rq);
+	rcu_read_unlock();
+
+	return result;
+}
+
+static inline void i915_request_mark_complete(struct i915_request *rq)
+{
+	WRITE_ONCE(rq->hwsp_seqno, /* decouple from HWSP */
+		   (u32 *)&rq->fence.seqno);
+}
+
+static inline bool i915_request_has_waitboost(const struct i915_request *rq)
+{
+	return test_bit(I915_FENCE_FLAG_BOOST, &rq->fence.flags);
+}
+
+static inline bool i915_request_has_nopreempt(const struct i915_request *rq)
+{
+	/* Preemption should only be disabled very rarely */
+	return unlikely(test_bit(I915_FENCE_FLAG_NOPREEMPT, &rq->fence.flags));
+}
+
+static inline bool i915_request_has_sentinel(const struct i915_request *rq)
+{
+	return unlikely(test_bit(I915_FENCE_FLAG_SENTINEL, &rq->fence.flags));
+}
+
+static inline bool i915_request_on_hold(const struct i915_request *rq)
+{
+	return unlikely(test_bit(I915_FENCE_FLAG_HOLD, &rq->fence.flags));
+}
+
+static inline void i915_request_set_hold(struct i915_request *rq)
+{
+	set_bit(I915_FENCE_FLAG_HOLD, &rq->fence.flags);
+}
+
+static inline void i915_request_clear_hold(struct i915_request *rq)
+{
+	clear_bit(I915_FENCE_FLAG_HOLD, &rq->fence.flags);
+}
+
+static inline struct intel_timeline *
+i915_request_timeline(const struct i915_request *rq)
+{
+	/* Valid only while the request is being constructed (or retired). */
+	return rcu_dereference_protected(rq->timeline,
+					 lockdep_is_held(&rcu_access_pointer(rq->timeline)->mutex) ||
+					 test_bit(CONTEXT_IS_PARKING, &rq->context->flags));
+}
+
+static inline struct i915_gem_context *
+i915_request_gem_context(const struct i915_request *rq)
+{
+	/* Valid only while the request is being constructed (or retired). */
+	return rcu_dereference_protected(rq->context->gem_context, true);
+}
+
+static inline struct intel_timeline *
+i915_request_active_timeline(const struct i915_request *rq)
+{
+	/*
+	 * When in use during submission, we are protected by a guarantee that
+	 * the context/timeline is pinned and must remain pinned until after
+	 * this submission.
+	 */
+	return rcu_dereference_protected(rq->timeline,
+					 lockdep_is_held(&rq->engine->sched_engine->lock));
+}
+
+static inline u32
+i915_request_active_seqno(const struct i915_request *rq)
+{
+	u32 hwsp_phys_base =
+		page_mask_bits(i915_request_active_timeline(rq)->hwsp_offset);
+	u32 hwsp_relative_offset = offset_in_page(rq->hwsp_seqno);
+
+	/*
+	 * Because of wraparound, we cannot simply take tl->hwsp_offset,
+	 * but instead use the fact that the relative for vaddr is the
+	 * offset as for hwsp_offset. Take the top bits from tl->hwsp_offset
+	 * and combine them with the relative offset in rq->hwsp_seqno.
+	 *
+	 * As rw->hwsp_seqno is rewritten when signaled, this only works
+	 * when the request isn't signaled yet, but at that point you
+	 * no longer need the offset.
+	 */
+
+	return hwsp_phys_base + hwsp_relative_offset;
+}
+
+bool
+i915_request_active_engine(struct i915_request *rq,
+			   struct intel_engine_cs **active);
+
+void i915_request_notify_execute_cb_imm(struct i915_request *rq);
+
+enum i915_request_state {
+	I915_REQUEST_UNKNOWN = 0,
+	I915_REQUEST_COMPLETE,
+	I915_REQUEST_PENDING,
+	I915_REQUEST_QUEUED,
+	I915_REQUEST_ACTIVE,
+};
+
+enum i915_request_state i915_test_request_state(struct i915_request *rq);
+
+void i915_request_module_exit(void);
+int i915_request_module_init(void);
+
+#endif /* I915_REQUEST_H */
diff --git a/drivers/gpu/drm/i915/i915_scatterlist.c b/drivers/gpu/drm/i915/i915_scatterlist.c
new file mode 100644
index 000000000..dcc081874
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_scatterlist.c
@@ -0,0 +1,229 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#include "i915_scatterlist.h"
+#include "i915_ttm_buddy_manager.h"
+
+#include <drm/drm_buddy.h>
+#include <drm/drm_mm.h>
+
+#include <linux/slab.h>
+
+bool i915_sg_trim(struct sg_table *orig_st)
+{
+	struct sg_table new_st;
+	struct scatterlist *sg, *new_sg;
+	unsigned int i;
+
+	if (orig_st->nents == orig_st->orig_nents)
+		return false;
+
+	if (sg_alloc_table(&new_st, orig_st->nents, GFP_KERNEL | __GFP_NOWARN))
+		return false;
+
+	new_sg = new_st.sgl;
+	for_each_sg(orig_st->sgl, sg, orig_st->nents, i) {
+		sg_set_page(new_sg, sg_page(sg), sg->length, 0);
+		sg_dma_address(new_sg) = sg_dma_address(sg);
+		sg_dma_len(new_sg) = sg_dma_len(sg);
+
+		new_sg = sg_next(new_sg);
+	}
+	GEM_BUG_ON(new_sg); /* Should walk exactly nents and hit the end */
+
+	sg_free_table(orig_st);
+
+	*orig_st = new_st;
+	return true;
+}
+
+static void i915_refct_sgt_release(struct kref *ref)
+{
+	struct i915_refct_sgt *rsgt =
+		container_of(ref, typeof(*rsgt), kref);
+
+	sg_free_table(&rsgt->table);
+	kfree(rsgt);
+}
+
+static const struct i915_refct_sgt_ops rsgt_ops = {
+	.release = i915_refct_sgt_release
+};
+
+/**
+ * i915_refct_sgt_init - Initialize a struct i915_refct_sgt with default ops
+ * @rsgt: The struct i915_refct_sgt to initialize.
+ * size: The size of the underlying memory buffer.
+ */
+void i915_refct_sgt_init(struct i915_refct_sgt *rsgt, size_t size)
+{
+	__i915_refct_sgt_init(rsgt, size, &rsgt_ops);
+}
+
+/**
+ * i915_rsgt_from_mm_node - Create a refcounted sg_table from a struct
+ * drm_mm_node
+ * @node: The drm_mm_node.
+ * @region_start: An offset to add to the dma addresses of the sg list.
+ * @page_alignment: Required page alignment for each sg entry. Power of two.
+ *
+ * Create a struct sg_table, initializing it from a struct drm_mm_node,
+ * taking a maximum segment length into account, splitting into segments
+ * if necessary.
+ *
+ * Return: A pointer to a kmalloced struct i915_refct_sgt on success, negative
+ * error code cast to an error pointer on failure.
+ */
+struct i915_refct_sgt *i915_rsgt_from_mm_node(const struct drm_mm_node *node,
+					      u64 region_start,
+					      u32 page_alignment)
+{
+	const u32 max_segment = round_down(UINT_MAX, page_alignment);
+	const u32 segment_pages = max_segment >> PAGE_SHIFT;
+	u64 block_size, offset, prev_end;
+	struct i915_refct_sgt *rsgt;
+	struct sg_table *st;
+	struct scatterlist *sg;
+
+	GEM_BUG_ON(!max_segment);
+
+	rsgt = kmalloc(sizeof(*rsgt), GFP_KERNEL);
+	if (!rsgt)
+		return ERR_PTR(-ENOMEM);
+
+	i915_refct_sgt_init(rsgt, node->size << PAGE_SHIFT);
+	st = &rsgt->table;
+	if (sg_alloc_table(st, DIV_ROUND_UP_ULL(node->size, segment_pages),
+			   GFP_KERNEL)) {
+		i915_refct_sgt_put(rsgt);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	sg = st->sgl;
+	st->nents = 0;
+	prev_end = (resource_size_t)-1;
+	block_size = node->size << PAGE_SHIFT;
+	offset = node->start << PAGE_SHIFT;
+
+	while (block_size) {
+		u64 len;
+
+		if (offset != prev_end || sg->length >= max_segment) {
+			if (st->nents)
+				sg = __sg_next(sg);
+
+			sg_dma_address(sg) = region_start + offset;
+			GEM_BUG_ON(!IS_ALIGNED(sg_dma_address(sg),
+					       page_alignment));
+			sg_dma_len(sg) = 0;
+			sg->length = 0;
+			st->nents++;
+		}
+
+		len = min_t(u64, block_size, max_segment - sg->length);
+		sg->length += len;
+		sg_dma_len(sg) += len;
+
+		offset += len;
+		block_size -= len;
+
+		prev_end = offset;
+	}
+
+	sg_mark_end(sg);
+	i915_sg_trim(st);
+
+	return rsgt;
+}
+
+/**
+ * i915_rsgt_from_buddy_resource - Create a refcounted sg_table from a struct
+ * i915_buddy_block list
+ * @res: The struct i915_ttm_buddy_resource.
+ * @region_start: An offset to add to the dma addresses of the sg list.
+ * @page_alignment: Required page alignment for each sg entry. Power of two.
+ *
+ * Create a struct sg_table, initializing it from struct i915_buddy_block list,
+ * taking a maximum segment length into account, splitting into segments
+ * if necessary.
+ *
+ * Return: A pointer to a kmalloced struct i915_refct_sgts on success, negative
+ * error code cast to an error pointer on failure.
+ */
+struct i915_refct_sgt *i915_rsgt_from_buddy_resource(struct ttm_resource *res,
+						     u64 region_start,
+						     u32 page_alignment)
+{
+	struct i915_ttm_buddy_resource *bman_res = to_ttm_buddy_resource(res);
+	const u64 size = res->num_pages << PAGE_SHIFT;
+	const u32 max_segment = round_down(UINT_MAX, page_alignment);
+	struct drm_buddy *mm = bman_res->mm;
+	struct list_head *blocks = &bman_res->blocks;
+	struct drm_buddy_block *block;
+	struct i915_refct_sgt *rsgt;
+	struct scatterlist *sg;
+	struct sg_table *st;
+	resource_size_t prev_end;
+
+	GEM_BUG_ON(list_empty(blocks));
+	GEM_BUG_ON(!max_segment);
+
+	rsgt = kmalloc(sizeof(*rsgt), GFP_KERNEL);
+	if (!rsgt)
+		return ERR_PTR(-ENOMEM);
+
+	i915_refct_sgt_init(rsgt, size);
+	st = &rsgt->table;
+	if (sg_alloc_table(st, res->num_pages, GFP_KERNEL)) {
+		i915_refct_sgt_put(rsgt);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	sg = st->sgl;
+	st->nents = 0;
+	prev_end = (resource_size_t)-1;
+
+	list_for_each_entry(block, blocks, link) {
+		u64 block_size, offset;
+
+		block_size = min_t(u64, size, drm_buddy_block_size(mm, block));
+		offset = drm_buddy_block_offset(block);
+
+		while (block_size) {
+			u64 len;
+
+			if (offset != prev_end || sg->length >= max_segment) {
+				if (st->nents)
+					sg = __sg_next(sg);
+
+				sg_dma_address(sg) = region_start + offset;
+				GEM_BUG_ON(!IS_ALIGNED(sg_dma_address(sg),
+						       page_alignment));
+				sg_dma_len(sg) = 0;
+				sg->length = 0;
+				st->nents++;
+			}
+
+			len = min_t(u64, block_size, max_segment - sg->length);
+			sg->length += len;
+			sg_dma_len(sg) += len;
+
+			offset += len;
+			block_size -= len;
+
+			prev_end = offset;
+		}
+	}
+
+	sg_mark_end(sg);
+	i915_sg_trim(st);
+
+	return rsgt;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/scatterlist.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_scatterlist.h b/drivers/gpu/drm/i915/i915_scatterlist.h
new file mode 100644
index 000000000..b0a1db44f
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_scatterlist.h
@@ -0,0 +1,231 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2016 Intel Corporation
+ */
+
+#ifndef I915_SCATTERLIST_H
+#define I915_SCATTERLIST_H
+
+#include <linux/pfn.h>
+#include <linux/scatterlist.h>
+#include <linux/dma-mapping.h>
+#include <xen/xen.h>
+
+#include "i915_gem.h"
+
+struct drm_mm_node;
+struct ttm_resource;
+
+/*
+ * Optimised SGL iterator for GEM objects
+ */
+static __always_inline struct sgt_iter {
+	struct scatterlist *sgp;
+	union {
+		unsigned long pfn;
+		dma_addr_t dma;
+	};
+	unsigned int curr;
+	unsigned int max;
+} __sgt_iter(struct scatterlist *sgl, bool dma) {
+	struct sgt_iter s = { .sgp = sgl };
+
+	if (dma && s.sgp && sg_dma_len(s.sgp) == 0) {
+		s.sgp = NULL;
+	} else if (s.sgp) {
+		s.max = s.curr = s.sgp->offset;
+		if (dma) {
+			s.dma = sg_dma_address(s.sgp);
+			s.max += sg_dma_len(s.sgp);
+		} else {
+			s.pfn = page_to_pfn(sg_page(s.sgp));
+			s.max += s.sgp->length;
+		}
+	}
+
+	return s;
+}
+
+static inline int __sg_page_count(const struct scatterlist *sg)
+{
+	return sg->length >> PAGE_SHIFT;
+}
+
+static inline int __sg_dma_page_count(const struct scatterlist *sg)
+{
+	return sg_dma_len(sg) >> PAGE_SHIFT;
+}
+
+static inline struct scatterlist *____sg_next(struct scatterlist *sg)
+{
+	++sg;
+	if (unlikely(sg_is_chain(sg)))
+		sg = sg_chain_ptr(sg);
+	return sg;
+}
+
+/**
+ * __sg_next - return the next scatterlist entry in a list
+ * @sg:		The current sg entry
+ *
+ * Description:
+ *   If the entry is the last, return NULL; otherwise, step to the next
+ *   element in the array (@sg@+1). If that's a chain pointer, follow it;
+ *   otherwise just return the pointer to the current element.
+ **/
+static inline struct scatterlist *__sg_next(struct scatterlist *sg)
+{
+	return sg_is_last(sg) ? NULL : ____sg_next(sg);
+}
+
+/**
+ * __for_each_sgt_daddr - iterate over the device addresses of the given sg_table
+ * @__dp:	Device address (output)
+ * @__iter:	'struct sgt_iter' (iterator state, internal)
+ * @__sgt:	sg_table to iterate over (input)
+ * @__step:	step size
+ */
+#define __for_each_sgt_daddr(__dp, __iter, __sgt, __step)		\
+	for ((__iter) = __sgt_iter((__sgt)->sgl, true);			\
+	     ((__dp) = (__iter).dma + (__iter).curr), (__iter).sgp;	\
+	     (((__iter).curr += (__step)) >= (__iter).max) ?		\
+	     (__iter) = __sgt_iter(__sg_next((__iter).sgp), true), 0 : 0)
+
+/**
+ * for_each_sgt_page - iterate over the pages of the given sg_table
+ * @__pp:	page pointer (output)
+ * @__iter:	'struct sgt_iter' (iterator state, internal)
+ * @__sgt:	sg_table to iterate over (input)
+ */
+#define for_each_sgt_page(__pp, __iter, __sgt)				\
+	for ((__iter) = __sgt_iter((__sgt)->sgl, false);		\
+	     ((__pp) = (__iter).pfn == 0 ? NULL :			\
+	      pfn_to_page((__iter).pfn + ((__iter).curr >> PAGE_SHIFT))); \
+	     (((__iter).curr += PAGE_SIZE) >= (__iter).max) ?		\
+	     (__iter) = __sgt_iter(__sg_next((__iter).sgp), false), 0 : 0)
+
+/**
+ * i915_sg_dma_sizes - Record the dma segment sizes of a scatterlist
+ * @sg: The scatterlist
+ *
+ * Return: An unsigned int with segment sizes logically or'ed together.
+ * A caller can use this information to determine what hardware page table
+ * entry sizes can be used to map the memory represented by the scatterlist.
+ */
+static inline unsigned int i915_sg_dma_sizes(struct scatterlist *sg)
+{
+	unsigned int page_sizes;
+
+	page_sizes = 0;
+	while (sg && sg_dma_len(sg)) {
+		GEM_BUG_ON(sg->offset);
+		GEM_BUG_ON(!IS_ALIGNED(sg_dma_len(sg), PAGE_SIZE));
+		page_sizes |= sg_dma_len(sg);
+		sg = __sg_next(sg);
+	}
+
+	return page_sizes;
+}
+
+static inline unsigned int i915_sg_segment_size(struct device *dev)
+{
+	size_t max = min_t(size_t, UINT_MAX, dma_max_mapping_size(dev));
+
+	/*
+	 * For Xen PV guests pages aren't contiguous in DMA (machine) address
+	 * space.  The DMA API takes care of that both in dma_alloc_* (by
+	 * calling into the hypervisor to make the pages contiguous) and in
+	 * dma_map_* (by bounce buffering).  But i915 abuses ignores the
+	 * coherency aspects of the DMA API and thus can't cope with bounce
+	 * buffering actually happening, so add a hack here to force small
+	 * allocations and mappings when running in PV mode on Xen.
+	 *
+	 * Note this will still break if bounce buffering is required for other
+	 * reasons, like confidential computing hypervisors or PCIe root ports
+	 * with addressing limitations.
+	 */
+	if (xen_pv_domain())
+		max = PAGE_SIZE;
+	return round_down(max, PAGE_SIZE);
+}
+
+bool i915_sg_trim(struct sg_table *orig_st);
+
+/**
+ * struct i915_refct_sgt_ops - Operations structure for struct i915_refct_sgt
+ */
+struct i915_refct_sgt_ops {
+	/**
+	 * release() - Free the memory of the struct i915_refct_sgt
+	 * @ref: struct kref that is embedded in the struct i915_refct_sgt
+	 */
+	void (*release)(struct kref *ref);
+};
+
+/**
+ * struct i915_refct_sgt - A refcounted scatter-gather table
+ * @kref: struct kref for refcounting
+ * @table: struct sg_table holding the scatter-gather table itself. Note that
+ * @table->sgl = NULL can be used to determine whether a scatter-gather table
+ * is present or not.
+ * @size: The size in bytes of the underlying memory buffer
+ * @ops: The operations structure.
+ */
+struct i915_refct_sgt {
+	struct kref kref;
+	struct sg_table table;
+	size_t size;
+	const struct i915_refct_sgt_ops *ops;
+};
+
+/**
+ * i915_refct_sgt_put - Put a refcounted sg-table
+ * @rsgt the struct i915_refct_sgt to put.
+ */
+static inline void i915_refct_sgt_put(struct i915_refct_sgt *rsgt)
+{
+	if (rsgt)
+		kref_put(&rsgt->kref, rsgt->ops->release);
+}
+
+/**
+ * i915_refct_sgt_get - Get a refcounted sg-table
+ * @rsgt the struct i915_refct_sgt to get.
+ */
+static inline struct i915_refct_sgt *
+i915_refct_sgt_get(struct i915_refct_sgt *rsgt)
+{
+	kref_get(&rsgt->kref);
+	return rsgt;
+}
+
+/**
+ * __i915_refct_sgt_init - Initialize a refcounted sg-list with a custom
+ * operations structure
+ * @rsgt The struct i915_refct_sgt to initialize.
+ * @size: Size in bytes of the underlying memory buffer.
+ * @ops: A customized operations structure in case the refcounted sg-list
+ * is embedded into another structure.
+ */
+static inline void __i915_refct_sgt_init(struct i915_refct_sgt *rsgt,
+					 size_t size,
+					 const struct i915_refct_sgt_ops *ops)
+{
+	kref_init(&rsgt->kref);
+	rsgt->table.sgl = NULL;
+	rsgt->size = size;
+	rsgt->ops = ops;
+}
+
+void i915_refct_sgt_init(struct i915_refct_sgt *rsgt, size_t size);
+
+struct i915_refct_sgt *i915_rsgt_from_mm_node(const struct drm_mm_node *node,
+					      u64 region_start,
+					      u32 page_alignment);
+
+struct i915_refct_sgt *i915_rsgt_from_buddy_resource(struct ttm_resource *res,
+						     u64 region_start,
+						     u32 page_alignment);
+
+#endif
diff --git a/drivers/gpu/drm/i915/i915_scheduler.c b/drivers/gpu/drm/i915/i915_scheduler.c
new file mode 100644
index 000000000..762127dd5
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_scheduler.c
@@ -0,0 +1,511 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/mutex.h>
+
+#include "i915_drv.h"
+#include "i915_request.h"
+#include "i915_scheduler.h"
+
+static struct kmem_cache *slab_dependencies;
+static struct kmem_cache *slab_priorities;
+
+static DEFINE_SPINLOCK(schedule_lock);
+
+static const struct i915_request *
+node_to_request(const struct i915_sched_node *node)
+{
+	return container_of(node, const struct i915_request, sched);
+}
+
+static inline bool node_started(const struct i915_sched_node *node)
+{
+	return i915_request_started(node_to_request(node));
+}
+
+static inline bool node_signaled(const struct i915_sched_node *node)
+{
+	return i915_request_completed(node_to_request(node));
+}
+
+static inline struct i915_priolist *to_priolist(struct rb_node *rb)
+{
+	return rb_entry(rb, struct i915_priolist, node);
+}
+
+static void assert_priolists(struct i915_sched_engine * const sched_engine)
+{
+	struct rb_node *rb;
+	long last_prio;
+
+	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		return;
+
+	GEM_BUG_ON(rb_first_cached(&sched_engine->queue) !=
+		   rb_first(&sched_engine->queue.rb_root));
+
+	last_prio = INT_MAX;
+	for (rb = rb_first_cached(&sched_engine->queue); rb; rb = rb_next(rb)) {
+		const struct i915_priolist *p = to_priolist(rb);
+
+		GEM_BUG_ON(p->priority > last_prio);
+		last_prio = p->priority;
+	}
+}
+
+struct list_head *
+i915_sched_lookup_priolist(struct i915_sched_engine *sched_engine, int prio)
+{
+	struct i915_priolist *p;
+	struct rb_node **parent, *rb;
+	bool first = true;
+
+	lockdep_assert_held(&sched_engine->lock);
+	assert_priolists(sched_engine);
+
+	if (unlikely(sched_engine->no_priolist))
+		prio = I915_PRIORITY_NORMAL;
+
+find_priolist:
+	/* most positive priority is scheduled first, equal priorities fifo */
+	rb = NULL;
+	parent = &sched_engine->queue.rb_root.rb_node;
+	while (*parent) {
+		rb = *parent;
+		p = to_priolist(rb);
+		if (prio > p->priority) {
+			parent = &rb->rb_left;
+		} else if (prio < p->priority) {
+			parent = &rb->rb_right;
+			first = false;
+		} else {
+			return &p->requests;
+		}
+	}
+
+	if (prio == I915_PRIORITY_NORMAL) {
+		p = &sched_engine->default_priolist;
+	} else {
+		p = kmem_cache_alloc(slab_priorities, GFP_ATOMIC);
+		/* Convert an allocation failure to a priority bump */
+		if (unlikely(!p)) {
+			prio = I915_PRIORITY_NORMAL; /* recurses just once */
+
+			/* To maintain ordering with all rendering, after an
+			 * allocation failure we have to disable all scheduling.
+			 * Requests will then be executed in fifo, and schedule
+			 * will ensure that dependencies are emitted in fifo.
+			 * There will be still some reordering with existing
+			 * requests, so if userspace lied about their
+			 * dependencies that reordering may be visible.
+			 */
+			sched_engine->no_priolist = true;
+			goto find_priolist;
+		}
+	}
+
+	p->priority = prio;
+	INIT_LIST_HEAD(&p->requests);
+
+	rb_link_node(&p->node, rb, parent);
+	rb_insert_color_cached(&p->node, &sched_engine->queue, first);
+
+	return &p->requests;
+}
+
+void __i915_priolist_free(struct i915_priolist *p)
+{
+	kmem_cache_free(slab_priorities, p);
+}
+
+struct sched_cache {
+	struct list_head *priolist;
+};
+
+static struct i915_sched_engine *
+lock_sched_engine(struct i915_sched_node *node,
+		  struct i915_sched_engine *locked,
+		  struct sched_cache *cache)
+{
+	const struct i915_request *rq = node_to_request(node);
+	struct i915_sched_engine *sched_engine;
+
+	GEM_BUG_ON(!locked);
+
+	/*
+	 * Virtual engines complicate acquiring the engine timeline lock,
+	 * as their rq->engine pointer is not stable until under that
+	 * engine lock. The simple ploy we use is to take the lock then
+	 * check that the rq still belongs to the newly locked engine.
+	 */
+	while (locked != (sched_engine = READ_ONCE(rq->engine)->sched_engine)) {
+		spin_unlock(&locked->lock);
+		memset(cache, 0, sizeof(*cache));
+		spin_lock(&sched_engine->lock);
+		locked = sched_engine;
+	}
+
+	GEM_BUG_ON(locked != sched_engine);
+	return locked;
+}
+
+static void __i915_schedule(struct i915_sched_node *node,
+			    const struct i915_sched_attr *attr)
+{
+	const int prio = max(attr->priority, node->attr.priority);
+	struct i915_sched_engine *sched_engine;
+	struct i915_dependency *dep, *p;
+	struct i915_dependency stack;
+	struct sched_cache cache;
+	LIST_HEAD(dfs);
+
+	/* Needed in order to use the temporary link inside i915_dependency */
+	lockdep_assert_held(&schedule_lock);
+	GEM_BUG_ON(prio == I915_PRIORITY_INVALID);
+
+	if (node_signaled(node))
+		return;
+
+	stack.signaler = node;
+	list_add(&stack.dfs_link, &dfs);
+
+	/*
+	 * Recursively bump all dependent priorities to match the new request.
+	 *
+	 * A naive approach would be to use recursion:
+	 * static void update_priorities(struct i915_sched_node *node, prio) {
+	 *	list_for_each_entry(dep, &node->signalers_list, signal_link)
+	 *		update_priorities(dep->signal, prio)
+	 *	queue_request(node);
+	 * }
+	 * but that may have unlimited recursion depth and so runs a very
+	 * real risk of overunning the kernel stack. Instead, we build
+	 * a flat list of all dependencies starting with the current request.
+	 * As we walk the list of dependencies, we add all of its dependencies
+	 * to the end of the list (this may include an already visited
+	 * request) and continue to walk onwards onto the new dependencies. The
+	 * end result is a topological list of requests in reverse order, the
+	 * last element in the list is the request we must execute first.
+	 */
+	list_for_each_entry(dep, &dfs, dfs_link) {
+		struct i915_sched_node *node = dep->signaler;
+
+		/* If we are already flying, we know we have no signalers */
+		if (node_started(node))
+			continue;
+
+		/*
+		 * Within an engine, there can be no cycle, but we may
+		 * refer to the same dependency chain multiple times
+		 * (redundant dependencies are not eliminated) and across
+		 * engines.
+		 */
+		list_for_each_entry(p, &node->signalers_list, signal_link) {
+			GEM_BUG_ON(p == dep); /* no cycles! */
+
+			if (node_signaled(p->signaler))
+				continue;
+
+			if (prio > READ_ONCE(p->signaler->attr.priority))
+				list_move_tail(&p->dfs_link, &dfs);
+		}
+	}
+
+	/*
+	 * If we didn't need to bump any existing priorities, and we haven't
+	 * yet submitted this request (i.e. there is no potential race with
+	 * execlists_submit_request()), we can set our own priority and skip
+	 * acquiring the engine locks.
+	 */
+	if (node->attr.priority == I915_PRIORITY_INVALID) {
+		GEM_BUG_ON(!list_empty(&node->link));
+		node->attr = *attr;
+
+		if (stack.dfs_link.next == stack.dfs_link.prev)
+			return;
+
+		__list_del_entry(&stack.dfs_link);
+	}
+
+	memset(&cache, 0, sizeof(cache));
+	sched_engine = node_to_request(node)->engine->sched_engine;
+	spin_lock(&sched_engine->lock);
+
+	/* Fifo and depth-first replacement ensure our deps execute before us */
+	sched_engine = lock_sched_engine(node, sched_engine, &cache);
+	list_for_each_entry_safe_reverse(dep, p, &dfs, dfs_link) {
+		struct i915_request *from = container_of(dep->signaler,
+							 struct i915_request,
+							 sched);
+		INIT_LIST_HEAD(&dep->dfs_link);
+
+		node = dep->signaler;
+		sched_engine = lock_sched_engine(node, sched_engine, &cache);
+		lockdep_assert_held(&sched_engine->lock);
+
+		/* Recheck after acquiring the engine->timeline.lock */
+		if (prio <= node->attr.priority || node_signaled(node))
+			continue;
+
+		GEM_BUG_ON(node_to_request(node)->engine->sched_engine !=
+			   sched_engine);
+
+		/* Must be called before changing the nodes priority */
+		if (sched_engine->bump_inflight_request_prio)
+			sched_engine->bump_inflight_request_prio(from, prio);
+
+		WRITE_ONCE(node->attr.priority, prio);
+
+		/*
+		 * Once the request is ready, it will be placed into the
+		 * priority lists and then onto the HW runlist. Before the
+		 * request is ready, it does not contribute to our preemption
+		 * decisions and we can safely ignore it, as it will, and
+		 * any preemption required, be dealt with upon submission.
+		 * See engine->submit_request()
+		 */
+		if (list_empty(&node->link))
+			continue;
+
+		if (i915_request_in_priority_queue(node_to_request(node))) {
+			if (!cache.priolist)
+				cache.priolist =
+					i915_sched_lookup_priolist(sched_engine,
+								   prio);
+			list_move_tail(&node->link, cache.priolist);
+		}
+
+		/* Defer (tasklet) submission until after all of our updates. */
+		if (sched_engine->kick_backend)
+			sched_engine->kick_backend(node_to_request(node), prio);
+	}
+
+	spin_unlock(&sched_engine->lock);
+}
+
+void i915_schedule(struct i915_request *rq, const struct i915_sched_attr *attr)
+{
+	spin_lock_irq(&schedule_lock);
+	__i915_schedule(&rq->sched, attr);
+	spin_unlock_irq(&schedule_lock);
+}
+
+void i915_sched_node_init(struct i915_sched_node *node)
+{
+	INIT_LIST_HEAD(&node->signalers_list);
+	INIT_LIST_HEAD(&node->waiters_list);
+	INIT_LIST_HEAD(&node->link);
+
+	i915_sched_node_reinit(node);
+}
+
+void i915_sched_node_reinit(struct i915_sched_node *node)
+{
+	node->attr.priority = I915_PRIORITY_INVALID;
+	node->semaphores = 0;
+	node->flags = 0;
+
+	GEM_BUG_ON(!list_empty(&node->signalers_list));
+	GEM_BUG_ON(!list_empty(&node->waiters_list));
+	GEM_BUG_ON(!list_empty(&node->link));
+}
+
+static struct i915_dependency *
+i915_dependency_alloc(void)
+{
+	return kmem_cache_alloc(slab_dependencies, GFP_KERNEL);
+}
+
+static void
+i915_dependency_free(struct i915_dependency *dep)
+{
+	kmem_cache_free(slab_dependencies, dep);
+}
+
+bool __i915_sched_node_add_dependency(struct i915_sched_node *node,
+				      struct i915_sched_node *signal,
+				      struct i915_dependency *dep,
+				      unsigned long flags)
+{
+	bool ret = false;
+
+	spin_lock_irq(&schedule_lock);
+
+	if (!node_signaled(signal)) {
+		INIT_LIST_HEAD(&dep->dfs_link);
+		dep->signaler = signal;
+		dep->waiter = node;
+		dep->flags = flags;
+
+		/* All set, now publish. Beware the lockless walkers. */
+		list_add_rcu(&dep->signal_link, &node->signalers_list);
+		list_add_rcu(&dep->wait_link, &signal->waiters_list);
+
+		/* Propagate the chains */
+		node->flags |= signal->flags;
+		ret = true;
+	}
+
+	spin_unlock_irq(&schedule_lock);
+
+	return ret;
+}
+
+int i915_sched_node_add_dependency(struct i915_sched_node *node,
+				   struct i915_sched_node *signal,
+				   unsigned long flags)
+{
+	struct i915_dependency *dep;
+
+	dep = i915_dependency_alloc();
+	if (!dep)
+		return -ENOMEM;
+
+	if (!__i915_sched_node_add_dependency(node, signal, dep,
+					      flags | I915_DEPENDENCY_ALLOC))
+		i915_dependency_free(dep);
+
+	return 0;
+}
+
+void i915_sched_node_fini(struct i915_sched_node *node)
+{
+	struct i915_dependency *dep, *tmp;
+
+	spin_lock_irq(&schedule_lock);
+
+	/*
+	 * Everyone we depended upon (the fences we wait to be signaled)
+	 * should retire before us and remove themselves from our list.
+	 * However, retirement is run independently on each timeline and
+	 * so we may be called out-of-order.
+	 */
+	list_for_each_entry_safe(dep, tmp, &node->signalers_list, signal_link) {
+		GEM_BUG_ON(!list_empty(&dep->dfs_link));
+
+		list_del_rcu(&dep->wait_link);
+		if (dep->flags & I915_DEPENDENCY_ALLOC)
+			i915_dependency_free(dep);
+	}
+	INIT_LIST_HEAD(&node->signalers_list);
+
+	/* Remove ourselves from everyone who depends upon us */
+	list_for_each_entry_safe(dep, tmp, &node->waiters_list, wait_link) {
+		GEM_BUG_ON(dep->signaler != node);
+		GEM_BUG_ON(!list_empty(&dep->dfs_link));
+
+		list_del_rcu(&dep->signal_link);
+		if (dep->flags & I915_DEPENDENCY_ALLOC)
+			i915_dependency_free(dep);
+	}
+	INIT_LIST_HEAD(&node->waiters_list);
+
+	spin_unlock_irq(&schedule_lock);
+}
+
+void i915_request_show_with_schedule(struct drm_printer *m,
+				     const struct i915_request *rq,
+				     const char *prefix,
+				     int indent)
+{
+	struct i915_dependency *dep;
+
+	i915_request_show(m, rq, prefix, indent);
+	if (i915_request_completed(rq))
+		return;
+
+	rcu_read_lock();
+	for_each_signaler(dep, rq) {
+		const struct i915_request *signaler =
+			node_to_request(dep->signaler);
+
+		/* Dependencies along the same timeline are expected. */
+		if (signaler->timeline == rq->timeline)
+			continue;
+
+		if (__i915_request_is_complete(signaler))
+			continue;
+
+		i915_request_show(m, signaler, prefix, indent + 2);
+	}
+	rcu_read_unlock();
+}
+
+static void default_destroy(struct kref *kref)
+{
+	struct i915_sched_engine *sched_engine =
+		container_of(kref, typeof(*sched_engine), ref);
+
+	tasklet_kill(&sched_engine->tasklet); /* flush the callback */
+	kfree(sched_engine);
+}
+
+static bool default_disabled(struct i915_sched_engine *sched_engine)
+{
+	return false;
+}
+
+struct i915_sched_engine *
+i915_sched_engine_create(unsigned int subclass)
+{
+	struct i915_sched_engine *sched_engine;
+
+	sched_engine = kzalloc(sizeof(*sched_engine), GFP_KERNEL);
+	if (!sched_engine)
+		return NULL;
+
+	kref_init(&sched_engine->ref);
+
+	sched_engine->queue = RB_ROOT_CACHED;
+	sched_engine->queue_priority_hint = INT_MIN;
+	sched_engine->destroy = default_destroy;
+	sched_engine->disabled = default_disabled;
+
+	INIT_LIST_HEAD(&sched_engine->requests);
+	INIT_LIST_HEAD(&sched_engine->hold);
+
+	spin_lock_init(&sched_engine->lock);
+	lockdep_set_subclass(&sched_engine->lock, subclass);
+
+	/*
+	 * Due to an interesting quirk in lockdep's internal debug tracking,
+	 * after setting a subclass we must ensure the lock is used. Otherwise,
+	 * nr_unused_locks is incremented once too often.
+	 */
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+	local_irq_disable();
+	lock_map_acquire(&sched_engine->lock.dep_map);
+	lock_map_release(&sched_engine->lock.dep_map);
+	local_irq_enable();
+#endif
+
+	return sched_engine;
+}
+
+void i915_scheduler_module_exit(void)
+{
+	kmem_cache_destroy(slab_dependencies);
+	kmem_cache_destroy(slab_priorities);
+}
+
+int __init i915_scheduler_module_init(void)
+{
+	slab_dependencies = KMEM_CACHE(i915_dependency,
+					      SLAB_HWCACHE_ALIGN |
+					      SLAB_TYPESAFE_BY_RCU);
+	if (!slab_dependencies)
+		return -ENOMEM;
+
+	slab_priorities = KMEM_CACHE(i915_priolist, 0);
+	if (!slab_priorities)
+		goto err_priorities;
+
+	return 0;
+
+err_priorities:
+	kmem_cache_destroy(slab_priorities);
+	return -ENOMEM;
+}
diff --git a/drivers/gpu/drm/i915/i915_scheduler.h b/drivers/gpu/drm/i915/i915_scheduler.h
new file mode 100644
index 000000000..c229c9107
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_scheduler.h
@@ -0,0 +1,109 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef _I915_SCHEDULER_H_
+#define _I915_SCHEDULER_H_
+
+#include <linux/bitops.h>
+#include <linux/list.h>
+#include <linux/kernel.h>
+
+#include "i915_scheduler_types.h"
+#include "i915_tasklet.h"
+
+struct drm_printer;
+
+#define priolist_for_each_request(it, plist) \
+	list_for_each_entry(it, &(plist)->requests, sched.link)
+
+#define priolist_for_each_request_consume(it, n, plist) \
+	list_for_each_entry_safe(it, n, &(plist)->requests, sched.link)
+
+void i915_sched_node_init(struct i915_sched_node *node);
+void i915_sched_node_reinit(struct i915_sched_node *node);
+
+bool __i915_sched_node_add_dependency(struct i915_sched_node *node,
+				      struct i915_sched_node *signal,
+				      struct i915_dependency *dep,
+				      unsigned long flags);
+
+int i915_sched_node_add_dependency(struct i915_sched_node *node,
+				   struct i915_sched_node *signal,
+				   unsigned long flags);
+
+void i915_sched_node_fini(struct i915_sched_node *node);
+
+void i915_schedule(struct i915_request *request,
+		   const struct i915_sched_attr *attr);
+
+struct list_head *
+i915_sched_lookup_priolist(struct i915_sched_engine *sched_engine, int prio);
+
+void __i915_priolist_free(struct i915_priolist *p);
+static inline void i915_priolist_free(struct i915_priolist *p)
+{
+	if (p->priority != I915_PRIORITY_NORMAL)
+		__i915_priolist_free(p);
+}
+
+struct i915_sched_engine *
+i915_sched_engine_create(unsigned int subclass);
+
+static inline struct i915_sched_engine *
+i915_sched_engine_get(struct i915_sched_engine *sched_engine)
+{
+	kref_get(&sched_engine->ref);
+	return sched_engine;
+}
+
+static inline void
+i915_sched_engine_put(struct i915_sched_engine *sched_engine)
+{
+	kref_put(&sched_engine->ref, sched_engine->destroy);
+}
+
+static inline bool
+i915_sched_engine_is_empty(struct i915_sched_engine *sched_engine)
+{
+	return RB_EMPTY_ROOT(&sched_engine->queue.rb_root);
+}
+
+static inline void
+i915_sched_engine_reset_on_empty(struct i915_sched_engine *sched_engine)
+{
+	if (i915_sched_engine_is_empty(sched_engine))
+		sched_engine->no_priolist = false;
+}
+
+static inline void
+i915_sched_engine_active_lock_bh(struct i915_sched_engine *sched_engine)
+{
+	local_bh_disable(); /* prevent local softirq and lock recursion */
+	tasklet_lock(&sched_engine->tasklet);
+}
+
+static inline void
+i915_sched_engine_active_unlock_bh(struct i915_sched_engine *sched_engine)
+{
+	tasklet_unlock(&sched_engine->tasklet);
+	local_bh_enable(); /* restore softirq, and kick ksoftirqd! */
+}
+
+void i915_request_show_with_schedule(struct drm_printer *m,
+				     const struct i915_request *rq,
+				     const char *prefix,
+				     int indent);
+
+static inline bool
+i915_sched_engine_disabled(struct i915_sched_engine *sched_engine)
+{
+	return sched_engine->disabled(sched_engine);
+}
+
+void i915_scheduler_module_exit(void);
+int i915_scheduler_module_init(void);
+
+#endif /* _I915_SCHEDULER_H_ */
diff --git a/drivers/gpu/drm/i915/i915_scheduler_types.h b/drivers/gpu/drm/i915/i915_scheduler_types.h
new file mode 100644
index 000000000..b0a1b58c7
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_scheduler_types.h
@@ -0,0 +1,205 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef _I915_SCHEDULER_TYPES_H_
+#define _I915_SCHEDULER_TYPES_H_
+
+#include <linux/list.h>
+
+#include "gt/intel_engine_types.h"
+#include "i915_priolist_types.h"
+
+struct drm_i915_private;
+struct i915_request;
+struct intel_engine_cs;
+
+struct i915_sched_attr {
+	/**
+	 * @priority: execution and service priority
+	 *
+	 * All clients are equal, but some are more equal than others!
+	 *
+	 * Requests from a context with a greater (more positive) value of
+	 * @priority will be executed before those with a lower @priority
+	 * value, forming a simple QoS.
+	 *
+	 * The &drm_i915_private.kernel_context is assigned the lowest priority.
+	 */
+	int priority;
+};
+
+/*
+ * "People assume that time is a strict progression of cause to effect, but
+ * actually, from a nonlinear, non-subjective viewpoint, it's more like a big
+ * ball of wibbly-wobbly, timey-wimey ... stuff." -The Doctor, 2015
+ *
+ * Requests exist in a complex web of interdependencies. Each request
+ * has to wait for some other request to complete before it is ready to be run
+ * (e.g. we have to wait until the pixels have been rendering into a texture
+ * before we can copy from it). We track the readiness of a request in terms
+ * of fences, but we also need to keep the dependency tree for the lifetime
+ * of the request (beyond the life of an individual fence). We use the tree
+ * at various points to reorder the requests whilst keeping the requests
+ * in order with respect to their various dependencies.
+ *
+ * There is no active component to the "scheduler". As we know the dependency
+ * DAG of each request, we are able to insert it into a sorted queue when it
+ * is ready, and are able to reorder its portion of the graph to accommodate
+ * dynamic priority changes.
+ *
+ * Ok, there is now one active element to the "scheduler" in the backends.
+ * We let a new context run for a small amount of time before re-evaluating
+ * the run order. As we re-evaluate, we maintain the strict ordering of
+ * dependencies, but attempt to rotate the active contexts (the current context
+ * is put to the back of its priority queue, then reshuffling its dependents).
+ * This provides minimal timeslicing and prevents a userspace hog (e.g.
+ * something waiting on a user semaphore [VkEvent]) from denying service to
+ * others.
+ */
+struct i915_sched_node {
+	struct list_head signalers_list; /* those before us, we depend upon */
+	struct list_head waiters_list; /* those after us, they depend upon us */
+	struct list_head link;
+	struct i915_sched_attr attr;
+	unsigned int flags;
+#define I915_SCHED_HAS_EXTERNAL_CHAIN	BIT(0)
+	intel_engine_mask_t semaphores;
+};
+
+struct i915_dependency {
+	struct i915_sched_node *signaler;
+	struct i915_sched_node *waiter;
+	struct list_head signal_link;
+	struct list_head wait_link;
+	struct list_head dfs_link;
+	unsigned long flags;
+#define I915_DEPENDENCY_ALLOC		BIT(0)
+#define I915_DEPENDENCY_EXTERNAL	BIT(1)
+#define I915_DEPENDENCY_WEAK		BIT(2)
+};
+
+#define for_each_waiter(p__, rq__) \
+	list_for_each_entry_lockless(p__, \
+				     &(rq__)->sched.waiters_list, \
+				     wait_link)
+
+#define for_each_signaler(p__, rq__) \
+	list_for_each_entry_rcu(p__, \
+				&(rq__)->sched.signalers_list, \
+				signal_link)
+
+/**
+ * struct i915_sched_engine - scheduler engine
+ *
+ * A schedule engine represents a submission queue with different priority
+ * bands. It contains all the common state (relative to the backend) to queue,
+ * track, and submit a request.
+ *
+ * This object at the moment is quite i915 specific but will transition into a
+ * container for the drm_gpu_scheduler plus a few other variables once the i915
+ * is integrated with the DRM scheduler.
+ */
+struct i915_sched_engine {
+	/**
+	 * @ref: reference count of schedule engine object
+	 */
+	struct kref ref;
+
+	/**
+	 * @lock: protects requests in priority lists, requests, hold and
+	 * tasklet while running
+	 */
+	spinlock_t lock;
+
+	/**
+	 * @requests: list of requests inflight on this schedule engine
+	 */
+	struct list_head requests;
+
+	/**
+	 * @hold: list of ready requests, but on hold
+	 */
+	struct list_head hold;
+
+	/**
+	 * @tasklet: softirq tasklet for submission
+	 */
+	struct tasklet_struct tasklet;
+
+	/**
+	 * @default_priolist: priority list for I915_PRIORITY_NORMAL
+	 */
+	struct i915_priolist default_priolist;
+
+	/**
+	 * @queue_priority_hint: Highest pending priority.
+	 *
+	 * When we add requests into the queue, or adjust the priority of
+	 * executing requests, we compute the maximum priority of those
+	 * pending requests. We can then use this value to determine if
+	 * we need to preempt the executing requests to service the queue.
+	 * However, since the we may have recorded the priority of an inflight
+	 * request we wanted to preempt but since completed, at the time of
+	 * dequeuing the priority hint may no longer may match the highest
+	 * available request priority.
+	 */
+	int queue_priority_hint;
+
+	/**
+	 * @queue: queue of requests, in priority lists
+	 */
+	struct rb_root_cached queue;
+
+	/**
+	 * @no_priolist: priority lists disabled
+	 */
+	bool no_priolist;
+
+	/**
+	 * @private_data: private data of the submission backend
+	 */
+	void *private_data;
+
+	/**
+	 * @destroy: destroy schedule engine / cleanup in backend
+	 */
+	void	(*destroy)(struct kref *kref);
+
+	/**
+	 * @disabled: check if backend has disabled submission
+	 */
+	bool	(*disabled)(struct i915_sched_engine *sched_engine);
+
+	/**
+	 * @kick_backend: kick backend after a request's priority has changed
+	 */
+	void	(*kick_backend)(const struct i915_request *rq,
+				int prio);
+
+	/**
+	 * @bump_inflight_request_prio: update priority of an inflight request
+	 */
+	void	(*bump_inflight_request_prio)(struct i915_request *rq,
+					      int prio);
+
+	/**
+	 * @retire_inflight_request_prio: indicate request is retired to
+	 * priority tracking
+	 */
+	void	(*retire_inflight_request_prio)(struct i915_request *rq);
+
+	/**
+	 * @schedule: adjust priority of request
+	 *
+	 * Call when the priority on a request has changed and it and its
+	 * dependencies may need rescheduling. Note the request itself may
+	 * not be ready to run!
+	 */
+	void	(*schedule)(struct i915_request *request,
+			    const struct i915_sched_attr *attr);
+};
+
+#endif /* _I915_SCHEDULER_TYPES_H_ */
diff --git a/drivers/gpu/drm/i915/i915_selftest.h b/drivers/gpu/drm/i915/i915_selftest.h
new file mode 100644
index 000000000..f54de0499
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_selftest.h
@@ -0,0 +1,140 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __I915_SELFTEST_H__
+#define __I915_SELFTEST_H__
+
+#include <linux/types.h>
+
+struct pci_dev;
+struct drm_i915_private;
+
+struct i915_selftest {
+	unsigned long timeout_jiffies;
+	unsigned int timeout_ms;
+	unsigned int random_seed;
+	char *filter;
+	int mock;
+	int live;
+	int perf;
+};
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include <linux/fault-inject.h>
+
+extern struct i915_selftest i915_selftest;
+
+int i915_mock_selftests(void);
+int i915_live_selftests(struct pci_dev *pdev);
+int i915_perf_selftests(struct pci_dev *pdev);
+
+/* We extract the function declarations from i915_mock_selftests.h and
+ * i915_live_selftests.h Add your unit test declarations there!
+ *
+ * Mock unit tests are run very early upon module load, before the driver
+ * is probed. All hardware interactions, as well as other subsystems, must
+ * be "mocked".
+ *
+ * Live unit tests are run after the driver is loaded - all hardware
+ * interactions are real.
+ */
+#define selftest(name, func) int func(void);
+#include "selftests/i915_mock_selftests.h"
+#undef selftest
+#define selftest(name, func) int func(struct drm_i915_private *i915);
+#include "selftests/i915_live_selftests.h"
+#include "selftests/i915_perf_selftests.h"
+#undef selftest
+
+struct i915_subtest {
+	int (*func)(void *data);
+	const char *name;
+};
+
+int __i915_nop_setup(void *data);
+int __i915_nop_teardown(int err, void *data);
+
+int __i915_live_setup(void *data);
+int __i915_live_teardown(int err, void *data);
+
+int __intel_gt_live_setup(void *data);
+int __intel_gt_live_teardown(int err, void *data);
+
+int __i915_subtests(const char *caller,
+		    int (*setup)(void *data),
+		    int (*teardown)(int err, void *data),
+		    const struct i915_subtest *st,
+		    unsigned int count,
+		    void *data);
+#define i915_subtests(T, data) \
+	__i915_subtests(__func__, \
+			__i915_nop_setup, __i915_nop_teardown, \
+			T, ARRAY_SIZE(T), data)
+#define i915_live_subtests(T, data) ({ \
+	typecheck(struct drm_i915_private *, data); \
+	__i915_subtests(__func__, \
+			__i915_live_setup, __i915_live_teardown, \
+			T, ARRAY_SIZE(T), data); \
+})
+#define intel_gt_live_subtests(T, data) ({ \
+	typecheck(struct intel_gt *, data); \
+	__i915_subtests(__func__, \
+			__intel_gt_live_setup, __intel_gt_live_teardown, \
+			T, ARRAY_SIZE(T), data); \
+})
+
+#define SUBTEST(x) { x, #x }
+
+#define I915_SELFTEST_DECLARE(x) x
+#define I915_SELFTEST_ONLY(x) unlikely(x)
+#define I915_SELFTEST_EXPORT
+
+#else /* !IS_ENABLED(CONFIG_DRM_I915_SELFTEST) */
+
+static inline int i915_mock_selftests(void) { return 0; }
+static inline int i915_live_selftests(struct pci_dev *pdev) { return 0; }
+static inline int i915_perf_selftests(struct pci_dev *pdev) { return 0; }
+
+#define I915_SELFTEST_DECLARE(x)
+#define I915_SELFTEST_ONLY(x) 0
+#define I915_SELFTEST_EXPORT static
+
+#endif
+
+/* Using the i915_selftest_ prefix becomes a little unwieldy with the helpers.
+ * Instead we use the igt_ shorthand, in reference to the intel-gpu-tools
+ * suite of uabi test cases (which includes a test runner for our selftests).
+ */
+
+#define IGT_TIMEOUT(name__) \
+	unsigned long name__ = jiffies + i915_selftest.timeout_jiffies
+
+__printf(2, 3)
+bool __igt_timeout(unsigned long timeout, const char *fmt, ...);
+
+#define igt_timeout(t, fmt, ...) \
+	__igt_timeout((t), KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__)
+
+void igt_hexdump(const void *buf, size_t len);
+
+#endif /* !__I915_SELFTEST_H__ */
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
new file mode 100644
index 000000000..81def10eb
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -0,0 +1,124 @@
+/*
+ *
+ * Copyright 2008 (c) Intel Corporation
+ *   Jesse Barnes <jbarnes@virtuousgeek.org>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "display/intel_de.h"
+#include "display/intel_gmbus.h"
+#include "display/intel_vga.h"
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "i915_suspend.h"
+#include "intel_pci_config.h"
+
+static void intel_save_swf(struct drm_i915_private *dev_priv)
+{
+	int i;
+
+	/* Scratch space */
+	if (GRAPHICS_VER(dev_priv) == 2 && IS_MOBILE(dev_priv)) {
+		for (i = 0; i < 7; i++) {
+			dev_priv->regfile.saveSWF0[i] = intel_de_read(dev_priv, SWF0(i));
+			dev_priv->regfile.saveSWF1[i] = intel_de_read(dev_priv, SWF1(i));
+		}
+		for (i = 0; i < 3; i++)
+			dev_priv->regfile.saveSWF3[i] = intel_de_read(dev_priv, SWF3(i));
+	} else if (GRAPHICS_VER(dev_priv) == 2) {
+		for (i = 0; i < 7; i++)
+			dev_priv->regfile.saveSWF1[i] = intel_de_read(dev_priv, SWF1(i));
+	} else if (HAS_GMCH(dev_priv)) {
+		for (i = 0; i < 16; i++) {
+			dev_priv->regfile.saveSWF0[i] = intel_de_read(dev_priv, SWF0(i));
+			dev_priv->regfile.saveSWF1[i] = intel_de_read(dev_priv, SWF1(i));
+		}
+		for (i = 0; i < 3; i++)
+			dev_priv->regfile.saveSWF3[i] = intel_de_read(dev_priv, SWF3(i));
+	}
+}
+
+static void intel_restore_swf(struct drm_i915_private *dev_priv)
+{
+	int i;
+
+	/* Scratch space */
+	if (GRAPHICS_VER(dev_priv) == 2 && IS_MOBILE(dev_priv)) {
+		for (i = 0; i < 7; i++) {
+			intel_de_write(dev_priv, SWF0(i), dev_priv->regfile.saveSWF0[i]);
+			intel_de_write(dev_priv, SWF1(i), dev_priv->regfile.saveSWF1[i]);
+		}
+		for (i = 0; i < 3; i++)
+			intel_de_write(dev_priv, SWF3(i), dev_priv->regfile.saveSWF3[i]);
+	} else if (GRAPHICS_VER(dev_priv) == 2) {
+		for (i = 0; i < 7; i++)
+			intel_de_write(dev_priv, SWF1(i), dev_priv->regfile.saveSWF1[i]);
+	} else if (HAS_GMCH(dev_priv)) {
+		for (i = 0; i < 16; i++) {
+			intel_de_write(dev_priv, SWF0(i), dev_priv->regfile.saveSWF0[i]);
+			intel_de_write(dev_priv, SWF1(i), dev_priv->regfile.saveSWF1[i]);
+		}
+		for (i = 0; i < 3; i++)
+			intel_de_write(dev_priv, SWF3(i), dev_priv->regfile.saveSWF3[i]);
+	}
+}
+
+void i915_save_display(struct drm_i915_private *dev_priv)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	/* Display arbitration control */
+	if (GRAPHICS_VER(dev_priv) <= 4)
+		dev_priv->regfile.saveDSPARB = intel_de_read(dev_priv, DSPARB);
+
+	if (GRAPHICS_VER(dev_priv) == 4)
+		pci_read_config_word(pdev, GCDGMBUS,
+				     &dev_priv->regfile.saveGCDGMBUS);
+
+	intel_save_swf(dev_priv);
+}
+
+void i915_restore_display(struct drm_i915_private *dev_priv)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+
+	if (!HAS_DISPLAY(dev_priv))
+		return;
+
+	intel_restore_swf(dev_priv);
+
+	if (GRAPHICS_VER(dev_priv) == 4)
+		pci_write_config_word(pdev, GCDGMBUS,
+				      dev_priv->regfile.saveGCDGMBUS);
+
+	/* Display arbitration */
+	if (GRAPHICS_VER(dev_priv) <= 4)
+		intel_de_write(dev_priv, DSPARB, dev_priv->regfile.saveDSPARB);
+
+	intel_vga_redisable(dev_priv);
+
+	intel_gmbus_reset(dev_priv);
+}
diff --git a/drivers/gpu/drm/i915/i915_suspend.h b/drivers/gpu/drm/i915/i915_suspend.h
new file mode 100644
index 000000000..e5a611ee3
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_suspend.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_SUSPEND_H__
+#define __I915_SUSPEND_H__
+
+struct drm_i915_private;
+
+void i915_save_display(struct drm_i915_private *i915);
+void i915_restore_display(struct drm_i915_private *i915);
+
+#endif /* __I915_SUSPEND_H__ */
diff --git a/drivers/gpu/drm/i915/i915_sw_fence.c b/drivers/gpu/drm/i915/i915_sw_fence.c
new file mode 100644
index 000000000..6fc0d1b89
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_sw_fence.c
@@ -0,0 +1,604 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * (C) Copyright 2016 Intel Corporation
+ */
+
+#include <linux/slab.h>
+#include <linux/dma-fence.h>
+#include <linux/irq_work.h>
+#include <linux/dma-resv.h>
+
+#include "i915_sw_fence.h"
+#include "i915_selftest.h"
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG)
+#define I915_SW_FENCE_BUG_ON(expr) BUG_ON(expr)
+#else
+#define I915_SW_FENCE_BUG_ON(expr) BUILD_BUG_ON_INVALID(expr)
+#endif
+
+#ifdef CONFIG_DRM_I915_SW_FENCE_CHECK_DAG
+static DEFINE_SPINLOCK(i915_sw_fence_lock);
+#endif
+
+#define WQ_FLAG_BITS \
+	BITS_PER_TYPE(typeof_member(struct wait_queue_entry, flags))
+
+/* after WQ_FLAG_* for safety */
+#define I915_SW_FENCE_FLAG_FENCE BIT(WQ_FLAG_BITS - 1)
+#define I915_SW_FENCE_FLAG_ALLOC BIT(WQ_FLAG_BITS - 2)
+
+enum {
+	DEBUG_FENCE_IDLE = 0,
+	DEBUG_FENCE_NOTIFY,
+};
+
+static void *i915_sw_fence_debug_hint(void *addr)
+{
+	return (void *)(((struct i915_sw_fence *)addr)->fn);
+}
+
+#ifdef CONFIG_DRM_I915_SW_FENCE_DEBUG_OBJECTS
+
+static const struct debug_obj_descr i915_sw_fence_debug_descr = {
+	.name = "i915_sw_fence",
+	.debug_hint = i915_sw_fence_debug_hint,
+};
+
+static inline void debug_fence_init(struct i915_sw_fence *fence)
+{
+	debug_object_init(fence, &i915_sw_fence_debug_descr);
+}
+
+static inline void debug_fence_init_onstack(struct i915_sw_fence *fence)
+{
+	debug_object_init_on_stack(fence, &i915_sw_fence_debug_descr);
+}
+
+static inline void debug_fence_activate(struct i915_sw_fence *fence)
+{
+	debug_object_activate(fence, &i915_sw_fence_debug_descr);
+}
+
+static inline void debug_fence_set_state(struct i915_sw_fence *fence,
+					 int old, int new)
+{
+	debug_object_active_state(fence, &i915_sw_fence_debug_descr, old, new);
+}
+
+static inline void debug_fence_deactivate(struct i915_sw_fence *fence)
+{
+	debug_object_deactivate(fence, &i915_sw_fence_debug_descr);
+}
+
+static inline void debug_fence_destroy(struct i915_sw_fence *fence)
+{
+	debug_object_destroy(fence, &i915_sw_fence_debug_descr);
+}
+
+static inline void debug_fence_free(struct i915_sw_fence *fence)
+{
+	debug_object_free(fence, &i915_sw_fence_debug_descr);
+	smp_wmb(); /* flush the change in state before reallocation */
+}
+
+static inline void debug_fence_assert(struct i915_sw_fence *fence)
+{
+	debug_object_assert_init(fence, &i915_sw_fence_debug_descr);
+}
+
+#else
+
+static inline void debug_fence_init(struct i915_sw_fence *fence)
+{
+}
+
+static inline void debug_fence_init_onstack(struct i915_sw_fence *fence)
+{
+}
+
+static inline void debug_fence_activate(struct i915_sw_fence *fence)
+{
+}
+
+static inline void debug_fence_set_state(struct i915_sw_fence *fence,
+					 int old, int new)
+{
+}
+
+static inline void debug_fence_deactivate(struct i915_sw_fence *fence)
+{
+}
+
+static inline void debug_fence_destroy(struct i915_sw_fence *fence)
+{
+}
+
+static inline void debug_fence_free(struct i915_sw_fence *fence)
+{
+}
+
+static inline void debug_fence_assert(struct i915_sw_fence *fence)
+{
+}
+
+#endif
+
+static int __i915_sw_fence_notify(struct i915_sw_fence *fence,
+				  enum i915_sw_fence_notify state)
+{
+	return fence->fn(fence, state);
+}
+
+#ifdef CONFIG_DRM_I915_SW_FENCE_DEBUG_OBJECTS
+void i915_sw_fence_fini(struct i915_sw_fence *fence)
+{
+	debug_fence_free(fence);
+}
+#endif
+
+static void __i915_sw_fence_wake_up_all(struct i915_sw_fence *fence,
+					struct list_head *continuation)
+{
+	wait_queue_head_t *x = &fence->wait;
+	wait_queue_entry_t *pos, *next;
+	unsigned long flags;
+
+	debug_fence_deactivate(fence);
+	atomic_set_release(&fence->pending, -1); /* 0 -> -1 [done] */
+
+	/*
+	 * To prevent unbounded recursion as we traverse the graph of
+	 * i915_sw_fences, we move the entry list from this, the next ready
+	 * fence, to the tail of the original fence's entry list
+	 * (and so added to the list to be woken).
+	 */
+
+	spin_lock_irqsave_nested(&x->lock, flags, 1 + !!continuation);
+	if (continuation) {
+		list_for_each_entry_safe(pos, next, &x->head, entry) {
+			if (pos->flags & I915_SW_FENCE_FLAG_FENCE)
+				list_move_tail(&pos->entry, continuation);
+			else
+				pos->func(pos, TASK_NORMAL, 0, continuation);
+		}
+	} else {
+		LIST_HEAD(extra);
+
+		do {
+			list_for_each_entry_safe(pos, next, &x->head, entry) {
+				int wake_flags;
+
+				wake_flags = 0;
+				if (pos->flags & I915_SW_FENCE_FLAG_FENCE)
+					wake_flags = fence->error;
+
+				pos->func(pos, TASK_NORMAL, wake_flags, &extra);
+			}
+
+			if (list_empty(&extra))
+				break;
+
+			list_splice_tail_init(&extra, &x->head);
+		} while (1);
+	}
+	spin_unlock_irqrestore(&x->lock, flags);
+
+	debug_fence_assert(fence);
+}
+
+static void __i915_sw_fence_complete(struct i915_sw_fence *fence,
+				     struct list_head *continuation)
+{
+	debug_fence_assert(fence);
+
+	if (!atomic_dec_and_test(&fence->pending))
+		return;
+
+	debug_fence_set_state(fence, DEBUG_FENCE_IDLE, DEBUG_FENCE_NOTIFY);
+
+	if (__i915_sw_fence_notify(fence, FENCE_COMPLETE) != NOTIFY_DONE)
+		return;
+
+	debug_fence_set_state(fence, DEBUG_FENCE_NOTIFY, DEBUG_FENCE_IDLE);
+
+	__i915_sw_fence_wake_up_all(fence, continuation);
+
+	debug_fence_destroy(fence);
+	__i915_sw_fence_notify(fence, FENCE_FREE);
+}
+
+void i915_sw_fence_complete(struct i915_sw_fence *fence)
+{
+	debug_fence_assert(fence);
+
+	if (WARN_ON(i915_sw_fence_done(fence)))
+		return;
+
+	__i915_sw_fence_complete(fence, NULL);
+}
+
+bool i915_sw_fence_await(struct i915_sw_fence *fence)
+{
+	int pending;
+
+	/*
+	 * It is only safe to add a new await to the fence while it has
+	 * not yet been signaled (i.e. there are still existing signalers).
+	 */
+	pending = atomic_read(&fence->pending);
+	do {
+		if (pending < 1)
+			return false;
+	} while (!atomic_try_cmpxchg(&fence->pending, &pending, pending + 1));
+
+	return true;
+}
+
+void __i915_sw_fence_init(struct i915_sw_fence *fence,
+			  i915_sw_fence_notify_t fn,
+			  const char *name,
+			  struct lock_class_key *key)
+{
+	__init_waitqueue_head(&fence->wait, name, key);
+	fence->fn = fn;
+#ifdef CONFIG_DRM_I915_SW_FENCE_CHECK_DAG
+	fence->flags = 0;
+#endif
+
+	i915_sw_fence_reinit(fence);
+}
+
+void i915_sw_fence_reinit(struct i915_sw_fence *fence)
+{
+	debug_fence_init(fence);
+
+	atomic_set(&fence->pending, 1);
+	fence->error = 0;
+
+	I915_SW_FENCE_BUG_ON(!list_empty(&fence->wait.head));
+}
+
+void i915_sw_fence_commit(struct i915_sw_fence *fence)
+{
+	debug_fence_activate(fence);
+	i915_sw_fence_complete(fence);
+}
+
+static int i915_sw_fence_wake(wait_queue_entry_t *wq, unsigned mode, int flags, void *key)
+{
+	i915_sw_fence_set_error_once(wq->private, flags);
+
+	list_del(&wq->entry);
+	__i915_sw_fence_complete(wq->private, key);
+
+	if (wq->flags & I915_SW_FENCE_FLAG_ALLOC)
+		kfree(wq);
+	return 0;
+}
+
+#ifdef CONFIG_DRM_I915_SW_FENCE_CHECK_DAG
+static bool __i915_sw_fence_check_if_after(struct i915_sw_fence *fence,
+				    const struct i915_sw_fence * const signaler)
+{
+	wait_queue_entry_t *wq;
+
+	if (__test_and_set_bit(I915_SW_FENCE_CHECKED_BIT, &fence->flags))
+		return false;
+
+	if (fence == signaler)
+		return true;
+
+	list_for_each_entry(wq, &fence->wait.head, entry) {
+		if (wq->func != i915_sw_fence_wake)
+			continue;
+
+		if (__i915_sw_fence_check_if_after(wq->private, signaler))
+			return true;
+	}
+
+	return false;
+}
+
+static void __i915_sw_fence_clear_checked_bit(struct i915_sw_fence *fence)
+{
+	wait_queue_entry_t *wq;
+
+	if (!__test_and_clear_bit(I915_SW_FENCE_CHECKED_BIT, &fence->flags))
+		return;
+
+	list_for_each_entry(wq, &fence->wait.head, entry) {
+		if (wq->func != i915_sw_fence_wake)
+			continue;
+
+		__i915_sw_fence_clear_checked_bit(wq->private);
+	}
+}
+
+static bool i915_sw_fence_check_if_after(struct i915_sw_fence *fence,
+				  const struct i915_sw_fence * const signaler)
+{
+	unsigned long flags;
+	bool err;
+
+	spin_lock_irqsave(&i915_sw_fence_lock, flags);
+	err = __i915_sw_fence_check_if_after(fence, signaler);
+	__i915_sw_fence_clear_checked_bit(fence);
+	spin_unlock_irqrestore(&i915_sw_fence_lock, flags);
+
+	return err;
+}
+#else
+static bool i915_sw_fence_check_if_after(struct i915_sw_fence *fence,
+					 const struct i915_sw_fence * const signaler)
+{
+	return false;
+}
+#endif
+
+static int __i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
+					  struct i915_sw_fence *signaler,
+					  wait_queue_entry_t *wq, gfp_t gfp)
+{
+	unsigned int pending;
+	unsigned long flags;
+
+	debug_fence_assert(fence);
+	might_sleep_if(gfpflags_allow_blocking(gfp));
+
+	if (i915_sw_fence_done(signaler)) {
+		i915_sw_fence_set_error_once(fence, signaler->error);
+		return 0;
+	}
+
+	debug_fence_assert(signaler);
+
+	/* The dependency graph must be acyclic. */
+	if (unlikely(i915_sw_fence_check_if_after(fence, signaler)))
+		return -EINVAL;
+
+	pending = I915_SW_FENCE_FLAG_FENCE;
+	if (!wq) {
+		wq = kmalloc(sizeof(*wq), gfp);
+		if (!wq) {
+			if (!gfpflags_allow_blocking(gfp))
+				return -ENOMEM;
+
+			i915_sw_fence_wait(signaler);
+			i915_sw_fence_set_error_once(fence, signaler->error);
+			return 0;
+		}
+
+		pending |= I915_SW_FENCE_FLAG_ALLOC;
+	}
+
+	INIT_LIST_HEAD(&wq->entry);
+	wq->flags = pending;
+	wq->func = i915_sw_fence_wake;
+	wq->private = fence;
+
+	i915_sw_fence_await(fence);
+
+	spin_lock_irqsave(&signaler->wait.lock, flags);
+	if (likely(!i915_sw_fence_done(signaler))) {
+		__add_wait_queue_entry_tail(&signaler->wait, wq);
+		pending = 1;
+	} else {
+		i915_sw_fence_wake(wq, 0, signaler->error, NULL);
+		pending = 0;
+	}
+	spin_unlock_irqrestore(&signaler->wait.lock, flags);
+
+	return pending;
+}
+
+int i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
+				 struct i915_sw_fence *signaler,
+				 wait_queue_entry_t *wq)
+{
+	return __i915_sw_fence_await_sw_fence(fence, signaler, wq, 0);
+}
+
+int i915_sw_fence_await_sw_fence_gfp(struct i915_sw_fence *fence,
+				     struct i915_sw_fence *signaler,
+				     gfp_t gfp)
+{
+	return __i915_sw_fence_await_sw_fence(fence, signaler, NULL, gfp);
+}
+
+struct i915_sw_dma_fence_cb_timer {
+	struct i915_sw_dma_fence_cb base;
+	struct dma_fence *dma;
+	struct timer_list timer;
+	struct irq_work work;
+	struct rcu_head rcu;
+};
+
+static void dma_i915_sw_fence_wake(struct dma_fence *dma,
+				   struct dma_fence_cb *data)
+{
+	struct i915_sw_dma_fence_cb *cb = container_of(data, typeof(*cb), base);
+
+	i915_sw_fence_set_error_once(cb->fence, dma->error);
+	i915_sw_fence_complete(cb->fence);
+	kfree(cb);
+}
+
+static void timer_i915_sw_fence_wake(struct timer_list *t)
+{
+	struct i915_sw_dma_fence_cb_timer *cb = from_timer(cb, t, timer);
+	struct i915_sw_fence *fence;
+
+	fence = xchg(&cb->base.fence, NULL);
+	if (!fence)
+		return;
+
+	pr_notice("Asynchronous wait on fence %s:%s:%llx timed out (hint:%ps)\n",
+		  cb->dma->ops->get_driver_name(cb->dma),
+		  cb->dma->ops->get_timeline_name(cb->dma),
+		  cb->dma->seqno,
+		  i915_sw_fence_debug_hint(fence));
+
+	i915_sw_fence_set_error_once(fence, -ETIMEDOUT);
+	i915_sw_fence_complete(fence);
+}
+
+static void dma_i915_sw_fence_wake_timer(struct dma_fence *dma,
+					 struct dma_fence_cb *data)
+{
+	struct i915_sw_dma_fence_cb_timer *cb =
+		container_of(data, typeof(*cb), base.base);
+	struct i915_sw_fence *fence;
+
+	fence = xchg(&cb->base.fence, NULL);
+	if (fence) {
+		i915_sw_fence_set_error_once(fence, dma->error);
+		i915_sw_fence_complete(fence);
+	}
+
+	irq_work_queue(&cb->work);
+}
+
+static void irq_i915_sw_fence_work(struct irq_work *wrk)
+{
+	struct i915_sw_dma_fence_cb_timer *cb =
+		container_of(wrk, typeof(*cb), work);
+
+	del_timer_sync(&cb->timer);
+	dma_fence_put(cb->dma);
+
+	kfree_rcu(cb, rcu);
+}
+
+int i915_sw_fence_await_dma_fence(struct i915_sw_fence *fence,
+				  struct dma_fence *dma,
+				  unsigned long timeout,
+				  gfp_t gfp)
+{
+	struct i915_sw_dma_fence_cb *cb;
+	dma_fence_func_t func;
+	int ret;
+
+	debug_fence_assert(fence);
+	might_sleep_if(gfpflags_allow_blocking(gfp));
+
+	if (dma_fence_is_signaled(dma)) {
+		i915_sw_fence_set_error_once(fence, dma->error);
+		return 0;
+	}
+
+	cb = kmalloc(timeout ?
+		     sizeof(struct i915_sw_dma_fence_cb_timer) :
+		     sizeof(struct i915_sw_dma_fence_cb),
+		     gfp);
+	if (!cb) {
+		if (!gfpflags_allow_blocking(gfp))
+			return -ENOMEM;
+
+		ret = dma_fence_wait(dma, false);
+		if (ret)
+			return ret;
+
+		i915_sw_fence_set_error_once(fence, dma->error);
+		return 0;
+	}
+
+	cb->fence = fence;
+	i915_sw_fence_await(fence);
+
+	func = dma_i915_sw_fence_wake;
+	if (timeout) {
+		struct i915_sw_dma_fence_cb_timer *timer =
+			container_of(cb, typeof(*timer), base);
+
+		timer->dma = dma_fence_get(dma);
+		init_irq_work(&timer->work, irq_i915_sw_fence_work);
+
+		timer_setup(&timer->timer,
+			    timer_i915_sw_fence_wake, TIMER_IRQSAFE);
+		mod_timer(&timer->timer, round_jiffies_up(jiffies + timeout));
+
+		func = dma_i915_sw_fence_wake_timer;
+	}
+
+	ret = dma_fence_add_callback(dma, &cb->base, func);
+	if (ret == 0) {
+		ret = 1;
+	} else {
+		func(dma, &cb->base);
+		if (ret == -ENOENT) /* fence already signaled */
+			ret = 0;
+	}
+
+	return ret;
+}
+
+static void __dma_i915_sw_fence_wake(struct dma_fence *dma,
+				     struct dma_fence_cb *data)
+{
+	struct i915_sw_dma_fence_cb *cb = container_of(data, typeof(*cb), base);
+
+	i915_sw_fence_set_error_once(cb->fence, dma->error);
+	i915_sw_fence_complete(cb->fence);
+}
+
+int __i915_sw_fence_await_dma_fence(struct i915_sw_fence *fence,
+				    struct dma_fence *dma,
+				    struct i915_sw_dma_fence_cb *cb)
+{
+	int ret;
+
+	debug_fence_assert(fence);
+
+	if (dma_fence_is_signaled(dma)) {
+		i915_sw_fence_set_error_once(fence, dma->error);
+		return 0;
+	}
+
+	cb->fence = fence;
+	i915_sw_fence_await(fence);
+
+	ret = 1;
+	if (dma_fence_add_callback(dma, &cb->base, __dma_i915_sw_fence_wake)) {
+		/* fence already signaled */
+		__dma_i915_sw_fence_wake(dma, &cb->base);
+		ret = 0;
+	}
+
+	return ret;
+}
+
+int i915_sw_fence_await_reservation(struct i915_sw_fence *fence,
+				    struct dma_resv *resv,
+				    const struct dma_fence_ops *exclude,
+				    bool write,
+				    unsigned long timeout,
+				    gfp_t gfp)
+{
+	struct dma_resv_iter cursor;
+	struct dma_fence *f;
+	int ret = 0, pending;
+
+	debug_fence_assert(fence);
+	might_sleep_if(gfpflags_allow_blocking(gfp));
+
+	dma_resv_iter_begin(&cursor, resv, dma_resv_usage_rw(write));
+	dma_resv_for_each_fence_unlocked(&cursor, f) {
+		pending = i915_sw_fence_await_dma_fence(fence, f, timeout,
+							gfp);
+		if (pending < 0) {
+			ret = pending;
+			break;
+		}
+
+		ret |= pending;
+	}
+	dma_resv_iter_end(&cursor);
+	return ret;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/lib_sw_fence.c"
+#include "selftests/i915_sw_fence.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_sw_fence.h b/drivers/gpu/drm/i915/i915_sw_fence.h
new file mode 100644
index 000000000..619fc5a22
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_sw_fence.h
@@ -0,0 +1,124 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * i915_sw_fence.h - library routines for N:M synchronisation points
+ *
+ * Copyright (C) 2016 Intel Corporation
+ */
+
+#ifndef _I915_SW_FENCE_H_
+#define _I915_SW_FENCE_H_
+
+#include <linux/dma-fence.h>
+#include <linux/gfp.h>
+#include <linux/kref.h>
+#include <linux/notifier.h> /* for NOTIFY_DONE */
+#include <linux/wait.h>
+
+struct completion;
+struct dma_resv;
+struct i915_sw_fence;
+
+enum i915_sw_fence_notify {
+	FENCE_COMPLETE,
+	FENCE_FREE
+};
+
+typedef int (*i915_sw_fence_notify_t)(struct i915_sw_fence *,
+				      enum i915_sw_fence_notify state);
+
+struct i915_sw_fence {
+	wait_queue_head_t wait;
+	i915_sw_fence_notify_t fn;
+#ifdef CONFIG_DRM_I915_SW_FENCE_CHECK_DAG
+	unsigned long flags;
+#endif
+	atomic_t pending;
+	int error;
+};
+
+#define I915_SW_FENCE_CHECKED_BIT	0 /* used internally for DAG checking */
+
+void __i915_sw_fence_init(struct i915_sw_fence *fence,
+			  i915_sw_fence_notify_t fn,
+			  const char *name,
+			  struct lock_class_key *key);
+#ifdef CONFIG_LOCKDEP
+#define i915_sw_fence_init(fence, fn)				\
+do {								\
+	static struct lock_class_key __key;			\
+								\
+	BUILD_BUG_ON((fn) == NULL);				\
+	__i915_sw_fence_init((fence), (fn), #fence, &__key);	\
+} while (0)
+#else
+#define i915_sw_fence_init(fence, fn)				\
+do {								\
+	BUILD_BUG_ON((fn) == NULL);				\
+	__i915_sw_fence_init((fence), (fn), NULL, NULL);	\
+} while (0)
+#endif
+
+void i915_sw_fence_reinit(struct i915_sw_fence *fence);
+
+#ifdef CONFIG_DRM_I915_SW_FENCE_DEBUG_OBJECTS
+void i915_sw_fence_fini(struct i915_sw_fence *fence);
+#else
+static inline void i915_sw_fence_fini(struct i915_sw_fence *fence) {}
+#endif
+
+void i915_sw_fence_commit(struct i915_sw_fence *fence);
+
+int i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
+				 struct i915_sw_fence *after,
+				 wait_queue_entry_t *wq);
+int i915_sw_fence_await_sw_fence_gfp(struct i915_sw_fence *fence,
+				     struct i915_sw_fence *after,
+				     gfp_t gfp);
+
+struct i915_sw_dma_fence_cb {
+	struct dma_fence_cb base;
+	struct i915_sw_fence *fence;
+};
+
+int __i915_sw_fence_await_dma_fence(struct i915_sw_fence *fence,
+				    struct dma_fence *dma,
+				    struct i915_sw_dma_fence_cb *cb);
+int i915_sw_fence_await_dma_fence(struct i915_sw_fence *fence,
+				  struct dma_fence *dma,
+				  unsigned long timeout,
+				  gfp_t gfp);
+
+int i915_sw_fence_await_reservation(struct i915_sw_fence *fence,
+				    struct dma_resv *resv,
+				    const struct dma_fence_ops *exclude,
+				    bool write,
+				    unsigned long timeout,
+				    gfp_t gfp);
+
+bool i915_sw_fence_await(struct i915_sw_fence *fence);
+void i915_sw_fence_complete(struct i915_sw_fence *fence);
+
+static inline bool i915_sw_fence_signaled(const struct i915_sw_fence *fence)
+{
+	return atomic_read(&fence->pending) <= 0;
+}
+
+static inline bool i915_sw_fence_done(const struct i915_sw_fence *fence)
+{
+	return atomic_read(&fence->pending) < 0;
+}
+
+static inline void i915_sw_fence_wait(struct i915_sw_fence *fence)
+{
+	wait_event(fence->wait, i915_sw_fence_done(fence));
+}
+
+static inline void
+i915_sw_fence_set_error_once(struct i915_sw_fence *fence, int error)
+{
+	if (unlikely(error))
+		cmpxchg(&fence->error, 0, error);
+}
+
+#endif /* _I915_SW_FENCE_H_ */
diff --git a/drivers/gpu/drm/i915/i915_sw_fence_work.c b/drivers/gpu/drm/i915/i915_sw_fence_work.c
new file mode 100644
index 000000000..d2e56b387
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_sw_fence_work.c
@@ -0,0 +1,99 @@
+// SPDX-License-Identifier: MIT
+
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include "i915_sw_fence_work.h"
+
+static void fence_complete(struct dma_fence_work *f)
+{
+	if (f->ops->release)
+		f->ops->release(f);
+	dma_fence_signal(&f->dma);
+}
+
+static void fence_work(struct work_struct *work)
+{
+	struct dma_fence_work *f = container_of(work, typeof(*f), work);
+
+	f->ops->work(f);
+
+	fence_complete(f);
+	dma_fence_put(&f->dma);
+}
+
+static int
+fence_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
+{
+	struct dma_fence_work *f = container_of(fence, typeof(*f), chain);
+
+	switch (state) {
+	case FENCE_COMPLETE:
+		if (fence->error)
+			dma_fence_set_error(&f->dma, fence->error);
+
+		if (!f->dma.error) {
+			dma_fence_get(&f->dma);
+			if (test_bit(DMA_FENCE_WORK_IMM, &f->dma.flags))
+				fence_work(&f->work);
+			else
+				queue_work(system_unbound_wq, &f->work);
+		} else {
+			fence_complete(f);
+		}
+		break;
+
+	case FENCE_FREE:
+		dma_fence_put(&f->dma);
+		break;
+	}
+
+	return NOTIFY_DONE;
+}
+
+static const char *get_driver_name(struct dma_fence *fence)
+{
+	return "dma-fence";
+}
+
+static const char *get_timeline_name(struct dma_fence *fence)
+{
+	struct dma_fence_work *f = container_of(fence, typeof(*f), dma);
+
+	return f->ops->name ?: "work";
+}
+
+static void fence_release(struct dma_fence *fence)
+{
+	struct dma_fence_work *f = container_of(fence, typeof(*f), dma);
+
+	i915_sw_fence_fini(&f->chain);
+
+	BUILD_BUG_ON(offsetof(typeof(*f), dma));
+	dma_fence_free(&f->dma);
+}
+
+static const struct dma_fence_ops fence_ops = {
+	.get_driver_name = get_driver_name,
+	.get_timeline_name = get_timeline_name,
+	.release = fence_release,
+};
+
+void dma_fence_work_init(struct dma_fence_work *f,
+			 const struct dma_fence_work_ops *ops)
+{
+	f->ops = ops;
+	spin_lock_init(&f->lock);
+	dma_fence_init(&f->dma, &fence_ops, &f->lock, 0, 0);
+	i915_sw_fence_init(&f->chain, fence_notify);
+	INIT_WORK(&f->work, fence_work);
+}
+
+int dma_fence_work_chain(struct dma_fence_work *f, struct dma_fence *signal)
+{
+	if (!signal)
+		return 0;
+
+	return __i915_sw_fence_await_dma_fence(&f->chain, signal, &f->cb);
+}
diff --git a/drivers/gpu/drm/i915/i915_sw_fence_work.h b/drivers/gpu/drm/i915/i915_sw_fence_work.h
new file mode 100644
index 000000000..d56806918
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_sw_fence_work.h
@@ -0,0 +1,67 @@
+/* SPDX-License-Identifier: MIT */
+
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef I915_SW_FENCE_WORK_H
+#define I915_SW_FENCE_WORK_H
+
+#include <linux/dma-fence.h>
+#include <linux/spinlock.h>
+#include <linux/workqueue.h>
+
+#include "i915_sw_fence.h"
+
+struct dma_fence_work;
+
+struct dma_fence_work_ops {
+	const char *name;
+	void (*work)(struct dma_fence_work *f);
+	void (*release)(struct dma_fence_work *f);
+};
+
+struct dma_fence_work {
+	struct dma_fence dma;
+	spinlock_t lock;
+
+	struct i915_sw_fence chain;
+	struct i915_sw_dma_fence_cb cb;
+
+	struct work_struct work;
+	const struct dma_fence_work_ops *ops;
+};
+
+enum {
+	DMA_FENCE_WORK_IMM = DMA_FENCE_FLAG_USER_BITS,
+};
+
+void dma_fence_work_init(struct dma_fence_work *f,
+			 const struct dma_fence_work_ops *ops);
+int dma_fence_work_chain(struct dma_fence_work *f, struct dma_fence *signal);
+
+static inline void dma_fence_work_commit(struct dma_fence_work *f)
+{
+	i915_sw_fence_commit(&f->chain);
+}
+
+/**
+ * dma_fence_work_commit_imm: Commit the fence, and if possible execute locally.
+ * @f: the fenced worker
+ *
+ * Instead of always scheduling a worker to execute the callback (see
+ * dma_fence_work_commit()), we try to execute the callback immediately in
+ * the local context. It is required that the fence be committed before it
+ * is published, and that no other threads try to tamper with the number
+ * of asynchronous waits on the fence (or else the callback will be
+ * executed in the wrong context, i.e. not the callers).
+ */
+static inline void dma_fence_work_commit_imm(struct dma_fence_work *f)
+{
+	if (atomic_read(&f->chain.pending) <= 1)
+		__set_bit(DMA_FENCE_WORK_IMM, &f->dma.flags);
+
+	dma_fence_work_commit(f);
+}
+
+#endif /* I915_SW_FENCE_WORK_H */
diff --git a/drivers/gpu/drm/i915/i915_switcheroo.c b/drivers/gpu/drm/i915/i915_switcheroo.c
new file mode 100644
index 000000000..f45bd6b6c
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_switcheroo.c
@@ -0,0 +1,72 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/vga_switcheroo.h>
+
+#include "i915_driver.h"
+#include "i915_drv.h"
+#include "i915_switcheroo.h"
+
+static void i915_switcheroo_set_state(struct pci_dev *pdev,
+				      enum vga_switcheroo_state state)
+{
+	struct drm_i915_private *i915 = pdev_to_i915(pdev);
+	pm_message_t pmm = { .event = PM_EVENT_SUSPEND };
+
+	if (!i915) {
+		dev_err(&pdev->dev, "DRM not initialized, aborting switch.\n");
+		return;
+	}
+	if (!HAS_DISPLAY(i915)) {
+		dev_err(&pdev->dev, "Device state not initialized, aborting switch.\n");
+		return;
+	}
+
+	if (state == VGA_SWITCHEROO_ON) {
+		drm_info(&i915->drm, "switched on\n");
+		i915->drm.switch_power_state = DRM_SWITCH_POWER_CHANGING;
+		/* i915 resume handler doesn't set to D0 */
+		pci_set_power_state(pdev, PCI_D0);
+		i915_driver_resume_switcheroo(i915);
+		i915->drm.switch_power_state = DRM_SWITCH_POWER_ON;
+	} else {
+		drm_info(&i915->drm, "switched off\n");
+		i915->drm.switch_power_state = DRM_SWITCH_POWER_CHANGING;
+		i915_driver_suspend_switcheroo(i915, pmm);
+		i915->drm.switch_power_state = DRM_SWITCH_POWER_OFF;
+	}
+}
+
+static bool i915_switcheroo_can_switch(struct pci_dev *pdev)
+{
+	struct drm_i915_private *i915 = pdev_to_i915(pdev);
+
+	/*
+	 * FIXME: open_count is protected by drm_global_mutex but that would lead to
+	 * locking inversion with the driver load path. And the access here is
+	 * completely racy anyway. So don't bother with locking for now.
+	 */
+	return i915 && HAS_DISPLAY(i915) && atomic_read(&i915->drm.open_count) == 0;
+}
+
+static const struct vga_switcheroo_client_ops i915_switcheroo_ops = {
+	.set_gpu_state = i915_switcheroo_set_state,
+	.reprobe = NULL,
+	.can_switch = i915_switcheroo_can_switch,
+};
+
+int i915_switcheroo_register(struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+
+	return vga_switcheroo_register_client(pdev, &i915_switcheroo_ops, false);
+}
+
+void i915_switcheroo_unregister(struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+
+	vga_switcheroo_unregister_client(pdev);
+}
diff --git a/drivers/gpu/drm/i915/i915_switcheroo.h b/drivers/gpu/drm/i915/i915_switcheroo.h
new file mode 100644
index 000000000..59b6c1e07
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_switcheroo.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_SWITCHEROO__
+#define __I915_SWITCHEROO__
+
+struct drm_i915_private;
+
+int i915_switcheroo_register(struct drm_i915_private *i915);
+void i915_switcheroo_unregister(struct drm_i915_private *i915);
+
+#endif /* __I915_SWITCHEROO__ */
diff --git a/drivers/gpu/drm/i915/i915_syncmap.c b/drivers/gpu/drm/i915/i915_syncmap.c
new file mode 100644
index 000000000..60404dbb2
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_syncmap.c
@@ -0,0 +1,412 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/slab.h>
+
+#include "i915_syncmap.h"
+
+#include "i915_gem.h" /* GEM_BUG_ON() */
+#include "i915_selftest.h"
+
+#define SHIFT ilog2(KSYNCMAP)
+#define MASK (KSYNCMAP - 1)
+
+/*
+ * struct i915_syncmap is a layer of a radixtree that maps a u64 fence
+ * context id to the last u32 fence seqno waited upon from that context.
+ * Unlike lib/radixtree it uses a parent pointer that allows traversal back to
+ * the root. This allows us to access the whole tree via a single pointer
+ * to the most recently used layer. We expect fence contexts to be dense
+ * and most reuse to be on the same i915_gem_context but on neighbouring
+ * engines (i.e. on adjacent contexts) and reuse the same leaf, a very
+ * effective lookup cache. If the new lookup is not on the same leaf, we
+ * expect it to be on the neighbouring branch.
+ *
+ * A leaf holds an array of u32 seqno, and has height 0. The bitmap field
+ * allows us to store whether a particular seqno is valid (i.e. allows us
+ * to distinguish unset from 0).
+ *
+ * A branch holds an array of layer pointers, and has height > 0, and always
+ * has at least 2 layers (either branches or leaves) below it.
+ *
+ * For example,
+ *	for x in
+ *	  0 1 2 0x10 0x11 0x200 0x201
+ *	  0x500000 0x500001 0x503000 0x503001
+ *	  0xE<<60:
+ *		i915_syncmap_set(&sync, x, lower_32_bits(x));
+ * will build a tree like:
+ *	0xXXXXXXXXXXXXXXXX
+ *	0-> 0x0000000000XXXXXX
+ *	|   0-> 0x0000000000000XXX
+ *	|   |   0-> 0x00000000000000XX
+ *	|   |   |   0-> 0x000000000000000X 0:0, 1:1, 2:2
+ *	|   |   |   1-> 0x000000000000001X 0:10, 1:11
+ *	|   |   2-> 0x000000000000020X 0:200, 1:201
+ *	|   5-> 0x000000000050XXXX
+ *	|       0-> 0x000000000050000X 0:500000, 1:500001
+ *	|       3-> 0x000000000050300X 0:503000, 1:503001
+ *	e-> 0xe00000000000000X e:e
+ */
+
+struct i915_syncmap {
+	u64 prefix;
+	unsigned int height;
+	unsigned int bitmap;
+	struct i915_syncmap *parent;
+	/*
+	 * Following this header is an array of either seqno or child pointers:
+	 * union {
+	 *	u32 seqno[KSYNCMAP];
+	 *	struct i915_syncmap *child[KSYNCMAP];
+	 * };
+	 */
+};
+
+/**
+ * i915_syncmap_init -- initialise the #i915_syncmap
+ * @root: pointer to the #i915_syncmap
+ */
+void i915_syncmap_init(struct i915_syncmap **root)
+{
+	BUILD_BUG_ON_NOT_POWER_OF_2(KSYNCMAP);
+	BUILD_BUG_ON_NOT_POWER_OF_2(SHIFT);
+	BUILD_BUG_ON(KSYNCMAP > BITS_PER_TYPE((*root)->bitmap));
+	*root = NULL;
+}
+
+static inline u32 *__sync_seqno(struct i915_syncmap *p)
+{
+	GEM_BUG_ON(p->height);
+	return (u32 *)(p + 1);
+}
+
+static inline struct i915_syncmap **__sync_child(struct i915_syncmap *p)
+{
+	GEM_BUG_ON(!p->height);
+	return (struct i915_syncmap **)(p + 1);
+}
+
+static inline unsigned int
+__sync_branch_idx(const struct i915_syncmap *p, u64 id)
+{
+	return (id >> p->height) & MASK;
+}
+
+static inline unsigned int
+__sync_leaf_idx(const struct i915_syncmap *p, u64 id)
+{
+	GEM_BUG_ON(p->height);
+	return id & MASK;
+}
+
+static inline u64 __sync_branch_prefix(const struct i915_syncmap *p, u64 id)
+{
+	return id >> p->height >> SHIFT;
+}
+
+static inline u64 __sync_leaf_prefix(const struct i915_syncmap *p, u64 id)
+{
+	GEM_BUG_ON(p->height);
+	return id >> SHIFT;
+}
+
+static inline bool seqno_later(u32 a, u32 b)
+{
+	return (s32)(a - b) >= 0;
+}
+
+/**
+ * i915_syncmap_is_later -- compare against the last know sync point
+ * @root: pointer to the #i915_syncmap
+ * @id: the context id (other timeline) we are synchronising to
+ * @seqno: the sequence number along the other timeline
+ *
+ * If we have already synchronised this @root timeline with another (@id) then
+ * we can omit any repeated or earlier synchronisation requests. If the two
+ * timelines are already coupled, we can also omit the dependency between the
+ * two as that is already known via the timeline.
+ *
+ * Returns true if the two timelines are already synchronised wrt to @seqno,
+ * false if not and the synchronisation must be emitted.
+ */
+bool i915_syncmap_is_later(struct i915_syncmap **root, u64 id, u32 seqno)
+{
+	struct i915_syncmap *p;
+	unsigned int idx;
+
+	p = *root;
+	if (!p)
+		return false;
+
+	if (likely(__sync_leaf_prefix(p, id) == p->prefix))
+		goto found;
+
+	/* First climb the tree back to a parent branch */
+	do {
+		p = p->parent;
+		if (!p)
+			return false;
+
+		if (__sync_branch_prefix(p, id) == p->prefix)
+			break;
+	} while (1);
+
+	/* And then descend again until we find our leaf */
+	do {
+		if (!p->height)
+			break;
+
+		p = __sync_child(p)[__sync_branch_idx(p, id)];
+		if (!p)
+			return false;
+
+		if (__sync_branch_prefix(p, id) != p->prefix)
+			return false;
+	} while (1);
+
+	*root = p;
+found:
+	idx = __sync_leaf_idx(p, id);
+	if (!(p->bitmap & BIT(idx)))
+		return false;
+
+	return seqno_later(__sync_seqno(p)[idx], seqno);
+}
+
+static struct i915_syncmap *
+__sync_alloc_leaf(struct i915_syncmap *parent, u64 id)
+{
+	struct i915_syncmap *p;
+
+	p = kmalloc(sizeof(*p) + KSYNCMAP * sizeof(u32), GFP_KERNEL);
+	if (unlikely(!p))
+		return NULL;
+
+	p->parent = parent;
+	p->height = 0;
+	p->bitmap = 0;
+	p->prefix = __sync_leaf_prefix(p, id);
+	return p;
+}
+
+static inline void __sync_set_seqno(struct i915_syncmap *p, u64 id, u32 seqno)
+{
+	unsigned int idx = __sync_leaf_idx(p, id);
+
+	p->bitmap |= BIT(idx);
+	__sync_seqno(p)[idx] = seqno;
+}
+
+static inline void __sync_set_child(struct i915_syncmap *p,
+				    unsigned int idx,
+				    struct i915_syncmap *child)
+{
+	p->bitmap |= BIT(idx);
+	__sync_child(p)[idx] = child;
+}
+
+static noinline int __sync_set(struct i915_syncmap **root, u64 id, u32 seqno)
+{
+	struct i915_syncmap *p = *root;
+	unsigned int idx;
+
+	if (!p) {
+		p = __sync_alloc_leaf(NULL, id);
+		if (unlikely(!p))
+			return -ENOMEM;
+
+		goto found;
+	}
+
+	/* Caller handled the likely cached case */
+	GEM_BUG_ON(__sync_leaf_prefix(p, id) == p->prefix);
+
+	/* Climb back up the tree until we find a common prefix */
+	do {
+		if (!p->parent)
+			break;
+
+		p = p->parent;
+
+		if (__sync_branch_prefix(p, id) == p->prefix)
+			break;
+	} while (1);
+
+	/*
+	 * No shortcut, we have to descend the tree to find the right layer
+	 * containing this fence.
+	 *
+	 * Each layer in the tree holds 16 (KSYNCMAP) pointers, either fences
+	 * or lower layers. Leaf nodes (height = 0) contain the fences, all
+	 * other nodes (height > 0) are internal layers that point to a lower
+	 * node. Each internal layer has at least 2 descendents.
+	 *
+	 * Starting at the top, we check whether the current prefix matches. If
+	 * it doesn't, we have gone past our target and need to insert a join
+	 * into the tree, and a new leaf node for the target as a descendent
+	 * of the join, as well as the original layer.
+	 *
+	 * The matching prefix means we are still following the right branch
+	 * of the tree. If it has height 0, we have found our leaf and just
+	 * need to replace the fence slot with ourselves. If the height is
+	 * not zero, our slot contains the next layer in the tree (unless
+	 * it is empty, in which case we can add ourselves as a new leaf).
+	 * As descend the tree the prefix grows (and height decreases).
+	 */
+	do {
+		struct i915_syncmap *next;
+
+		if (__sync_branch_prefix(p, id) != p->prefix) {
+			unsigned int above;
+
+			/* Insert a join above the current layer */
+			next = kzalloc(sizeof(*next) + KSYNCMAP * sizeof(next),
+				       GFP_KERNEL);
+			if (unlikely(!next))
+				return -ENOMEM;
+
+			/* Compute the height at which these two diverge */
+			above = fls64(__sync_branch_prefix(p, id) ^ p->prefix);
+			above = round_up(above, SHIFT);
+			next->height = above + p->height;
+			next->prefix = __sync_branch_prefix(next, id);
+
+			/* Insert the join into the parent */
+			if (p->parent) {
+				idx = __sync_branch_idx(p->parent, id);
+				__sync_child(p->parent)[idx] = next;
+				GEM_BUG_ON(!(p->parent->bitmap & BIT(idx)));
+			}
+			next->parent = p->parent;
+
+			/* Compute the idx of the other branch, not our id! */
+			idx = p->prefix >> (above - SHIFT) & MASK;
+			__sync_set_child(next, idx, p);
+			p->parent = next;
+
+			/* Ascend to the join */
+			p = next;
+		} else {
+			if (!p->height)
+				break;
+		}
+
+		/* Descend into the next layer */
+		GEM_BUG_ON(!p->height);
+		idx = __sync_branch_idx(p, id);
+		next = __sync_child(p)[idx];
+		if (!next) {
+			next = __sync_alloc_leaf(p, id);
+			if (unlikely(!next))
+				return -ENOMEM;
+
+			__sync_set_child(p, idx, next);
+			p = next;
+			break;
+		}
+
+		p = next;
+	} while (1);
+
+found:
+	GEM_BUG_ON(p->prefix != __sync_leaf_prefix(p, id));
+	__sync_set_seqno(p, id, seqno);
+	*root = p;
+	return 0;
+}
+
+/**
+ * i915_syncmap_set -- mark the most recent syncpoint between contexts
+ * @root: pointer to the #i915_syncmap
+ * @id: the context id (other timeline) we have synchronised to
+ * @seqno: the sequence number along the other timeline
+ *
+ * When we synchronise this @root timeline with another (@id), we also know
+ * that we have synchronized with all previous seqno along that timeline. If
+ * we then have a request to synchronise with the same seqno or older, we can
+ * omit it, see i915_syncmap_is_later()
+ *
+ * Returns 0 on success, or a negative error code.
+ */
+int i915_syncmap_set(struct i915_syncmap **root, u64 id, u32 seqno)
+{
+	struct i915_syncmap *p = *root;
+
+	/*
+	 * We expect to be called in sequence following is_later(id), which
+	 * should have preloaded the root for us.
+	 */
+	if (likely(p && __sync_leaf_prefix(p, id) == p->prefix)) {
+		__sync_set_seqno(p, id, seqno);
+		return 0;
+	}
+
+	return __sync_set(root, id, seqno);
+}
+
+static void __sync_free(struct i915_syncmap *p)
+{
+	if (p->height) {
+		unsigned int i;
+
+		while ((i = ffs(p->bitmap))) {
+			p->bitmap &= ~0u << i;
+			__sync_free(__sync_child(p)[i - 1]);
+		}
+	}
+
+	kfree(p);
+}
+
+/**
+ * i915_syncmap_free -- free all memory associated with the syncmap
+ * @root: pointer to the #i915_syncmap
+ *
+ * Either when the timeline is to be freed and we no longer need the sync
+ * point tracking, or when the fences are all known to be signaled and the
+ * sync point tracking is redundant, we can free the #i915_syncmap to recover
+ * its allocations.
+ *
+ * Will reinitialise the @root pointer so that the #i915_syncmap is ready for
+ * reuse.
+ */
+void i915_syncmap_free(struct i915_syncmap **root)
+{
+	struct i915_syncmap *p;
+
+	p = *root;
+	if (!p)
+		return;
+
+	while (p->parent)
+		p = p->parent;
+
+	__sync_free(p);
+	*root = NULL;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/i915_syncmap.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_syncmap.h b/drivers/gpu/drm/i915/i915_syncmap.h
new file mode 100644
index 000000000..0653f70be
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_syncmap.h
@@ -0,0 +1,38 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __I915_SYNCMAP_H__
+#define __I915_SYNCMAP_H__
+
+#include <linux/types.h>
+
+struct i915_syncmap;
+#define KSYNCMAP 16 /* radix of the tree, how many slots in each layer */
+
+void i915_syncmap_init(struct i915_syncmap **root);
+int i915_syncmap_set(struct i915_syncmap **root, u64 id, u32 seqno);
+bool i915_syncmap_is_later(struct i915_syncmap **root, u64 id, u32 seqno);
+void i915_syncmap_free(struct i915_syncmap **root);
+
+#endif /* __I915_SYNCMAP_H__ */
diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c
new file mode 100644
index 000000000..1e2750210
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_sysfs.c
@@ -0,0 +1,273 @@
+/*
+ * Copyright © 2012 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Ben Widawsky <ben@bwidawsk.net>
+ *
+ */
+
+#include <linux/device.h>
+#include <linux/module.h>
+#include <linux/stat.h>
+#include <linux/sysfs.h>
+
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_rc6.h"
+#include "gt/intel_rps.h"
+#include "gt/sysfs_engines.h"
+
+#include "i915_drv.h"
+#include "i915_sysfs.h"
+#include "intel_pm.h"
+
+struct drm_i915_private *kdev_minor_to_i915(struct device *kdev)
+{
+	struct drm_minor *minor = dev_get_drvdata(kdev);
+	return to_i915(minor->dev);
+}
+
+static int l3_access_valid(struct drm_i915_private *i915, loff_t offset)
+{
+	if (!HAS_L3_DPF(i915))
+		return -EPERM;
+
+	if (!IS_ALIGNED(offset, sizeof(u32)))
+		return -EINVAL;
+
+	if (offset >= GEN7_L3LOG_SIZE)
+		return -ENXIO;
+
+	return 0;
+}
+
+static ssize_t
+i915_l3_read(struct file *filp, struct kobject *kobj,
+	     struct bin_attribute *attr, char *buf,
+	     loff_t offset, size_t count)
+{
+	struct device *kdev = kobj_to_dev(kobj);
+	struct drm_i915_private *i915 = kdev_minor_to_i915(kdev);
+	int slice = (int)(uintptr_t)attr->private;
+	int ret;
+
+	ret = l3_access_valid(i915, offset);
+	if (ret)
+		return ret;
+
+	count = round_down(count, sizeof(u32));
+	count = min_t(size_t, GEN7_L3LOG_SIZE - offset, count);
+	memset(buf, 0, count);
+
+	spin_lock(&i915->gem.contexts.lock);
+	if (i915->l3_parity.remap_info[slice])
+		memcpy(buf,
+		       i915->l3_parity.remap_info[slice] + offset / sizeof(u32),
+		       count);
+	spin_unlock(&i915->gem.contexts.lock);
+
+	return count;
+}
+
+static ssize_t
+i915_l3_write(struct file *filp, struct kobject *kobj,
+	      struct bin_attribute *attr, char *buf,
+	      loff_t offset, size_t count)
+{
+	struct device *kdev = kobj_to_dev(kobj);
+	struct drm_i915_private *i915 = kdev_minor_to_i915(kdev);
+	int slice = (int)(uintptr_t)attr->private;
+	u32 *remap_info, *freeme = NULL;
+	struct i915_gem_context *ctx;
+	int ret;
+
+	ret = l3_access_valid(i915, offset);
+	if (ret)
+		return ret;
+
+	if (count < sizeof(u32))
+		return -EINVAL;
+
+	remap_info = kzalloc(GEN7_L3LOG_SIZE, GFP_KERNEL);
+	if (!remap_info)
+		return -ENOMEM;
+
+	spin_lock(&i915->gem.contexts.lock);
+
+	if (i915->l3_parity.remap_info[slice]) {
+		freeme = remap_info;
+		remap_info = i915->l3_parity.remap_info[slice];
+	} else {
+		i915->l3_parity.remap_info[slice] = remap_info;
+	}
+
+	count = round_down(count, sizeof(u32));
+	memcpy(remap_info + offset / sizeof(u32), buf, count);
+
+	/* NB: We defer the remapping until we switch to the context */
+	list_for_each_entry(ctx, &i915->gem.contexts.list, link)
+		ctx->remap_slice |= BIT(slice);
+
+	spin_unlock(&i915->gem.contexts.lock);
+	kfree(freeme);
+
+	/*
+	 * TODO: Ideally we really want a GPU reset here to make sure errors
+	 * aren't propagated. Since I cannot find a stable way to reset the GPU
+	 * at this point it is left as a TODO.
+	*/
+
+	return count;
+}
+
+static const struct bin_attribute dpf_attrs = {
+	.attr = {.name = "l3_parity", .mode = (S_IRUSR | S_IWUSR)},
+	.size = GEN7_L3LOG_SIZE,
+	.read = i915_l3_read,
+	.write = i915_l3_write,
+	.mmap = NULL,
+	.private = (void *)0
+};
+
+static const struct bin_attribute dpf_attrs_1 = {
+	.attr = {.name = "l3_parity_slice_1", .mode = (S_IRUSR | S_IWUSR)},
+	.size = GEN7_L3LOG_SIZE,
+	.read = i915_l3_read,
+	.write = i915_l3_write,
+	.mmap = NULL,
+	.private = (void *)1
+};
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+
+static ssize_t error_state_read(struct file *filp, struct kobject *kobj,
+				struct bin_attribute *attr, char *buf,
+				loff_t off, size_t count)
+{
+
+	struct device *kdev = kobj_to_dev(kobj);
+	struct drm_i915_private *i915 = kdev_minor_to_i915(kdev);
+	struct i915_gpu_coredump *gpu;
+	ssize_t ret = 0;
+
+	/*
+	 * FIXME: Concurrent clients triggering resets and reading + clearing
+	 * dumps can cause inconsistent sysfs reads when a user calls in with a
+	 * non-zero offset to complete a prior partial read but the
+	 * gpu_coredump has been cleared or replaced.
+	 */
+
+	gpu = i915_first_error_state(i915);
+	if (IS_ERR(gpu)) {
+		ret = PTR_ERR(gpu);
+	} else if (gpu) {
+		ret = i915_gpu_coredump_copy_to_buffer(gpu, buf, off, count);
+		i915_gpu_coredump_put(gpu);
+	} else {
+		const char *str = "No error state collected\n";
+		size_t len = strlen(str);
+
+		if (off < len) {
+			ret = min_t(size_t, count, len - off);
+			memcpy(buf, str + off, ret);
+		}
+	}
+
+	return ret;
+}
+
+static ssize_t error_state_write(struct file *file, struct kobject *kobj,
+				 struct bin_attribute *attr, char *buf,
+				 loff_t off, size_t count)
+{
+	struct device *kdev = kobj_to_dev(kobj);
+	struct drm_i915_private *dev_priv = kdev_minor_to_i915(kdev);
+
+	drm_dbg(&dev_priv->drm, "Resetting error state\n");
+	i915_reset_error_state(dev_priv);
+
+	return count;
+}
+
+static const struct bin_attribute error_state_attr = {
+	.attr.name = "error",
+	.attr.mode = S_IRUSR | S_IWUSR,
+	.size = 0,
+	.read = error_state_read,
+	.write = error_state_write,
+};
+
+static void i915_setup_error_capture(struct device *kdev)
+{
+	if (sysfs_create_bin_file(&kdev->kobj, &error_state_attr))
+		DRM_ERROR("error_state sysfs setup failed\n");
+}
+
+static void i915_teardown_error_capture(struct device *kdev)
+{
+	sysfs_remove_bin_file(&kdev->kobj, &error_state_attr);
+}
+#else
+static void i915_setup_error_capture(struct device *kdev) {}
+static void i915_teardown_error_capture(struct device *kdev) {}
+#endif
+
+void i915_setup_sysfs(struct drm_i915_private *dev_priv)
+{
+	struct device *kdev = dev_priv->drm.primary->kdev;
+	int ret;
+
+	if (HAS_L3_DPF(dev_priv)) {
+		ret = device_create_bin_file(kdev, &dpf_attrs);
+		if (ret)
+			drm_err(&dev_priv->drm,
+				"l3 parity sysfs setup failed\n");
+
+		if (NUM_L3_SLICES(dev_priv) > 1) {
+			ret = device_create_bin_file(kdev,
+						     &dpf_attrs_1);
+			if (ret)
+				drm_err(&dev_priv->drm,
+					"l3 parity slice 1 setup failed\n");
+		}
+	}
+
+	dev_priv->sysfs_gt = kobject_create_and_add("gt", &kdev->kobj);
+	if (!dev_priv->sysfs_gt)
+		drm_warn(&dev_priv->drm,
+			 "failed to register GT sysfs directory\n");
+
+	i915_setup_error_capture(kdev);
+
+	intel_engines_add_sysfs(dev_priv);
+}
+
+void i915_teardown_sysfs(struct drm_i915_private *dev_priv)
+{
+	struct device *kdev = dev_priv->drm.primary->kdev;
+
+	i915_teardown_error_capture(kdev);
+
+	device_remove_bin_file(kdev,  &dpf_attrs_1);
+	device_remove_bin_file(kdev,  &dpf_attrs);
+
+	kobject_put(dev_priv->sysfs_gt);
+}
diff --git a/drivers/gpu/drm/i915/i915_sysfs.h b/drivers/gpu/drm/i915/i915_sysfs.h
new file mode 100644
index 000000000..243a17741
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_sysfs.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __I915_SYSFS_H__
+#define __I915_SYSFS_H__
+
+struct device;
+struct drm_i915_private;
+
+struct drm_i915_private *kdev_minor_to_i915(struct device *kdev);
+
+void i915_setup_sysfs(struct drm_i915_private *i915);
+void i915_teardown_sysfs(struct drm_i915_private *i915);
+
+#endif /* __I915_SYSFS_H__ */
diff --git a/drivers/gpu/drm/i915/i915_tasklet.h b/drivers/gpu/drm/i915/i915_tasklet.h
new file mode 100644
index 000000000..5d7069bdf
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_tasklet.h
@@ -0,0 +1,43 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __I915_TASKLET_H__
+#define __I915_TASKLET_H__
+
+#include <linux/interrupt.h>
+
+static inline void tasklet_lock(struct tasklet_struct *t)
+{
+	while (!tasklet_trylock(t))
+		cpu_relax();
+}
+
+static inline bool tasklet_is_locked(const struct tasklet_struct *t)
+{
+	return test_bit(TASKLET_STATE_RUN, &t->state);
+}
+
+static inline void __tasklet_disable_sync_once(struct tasklet_struct *t)
+{
+	if (!atomic_fetch_inc(&t->count))
+		tasklet_unlock_spin_wait(t);
+}
+
+static inline bool __tasklet_is_enabled(const struct tasklet_struct *t)
+{
+	return !atomic_read(&t->count);
+}
+
+static inline bool __tasklet_enable(struct tasklet_struct *t)
+{
+	return atomic_dec_and_test(&t->count);
+}
+
+static inline bool __tasklet_is_scheduled(struct tasklet_struct *t)
+{
+	return test_bit(TASKLET_STATE_SCHED, &t->state);
+}
+
+#endif /* __I915_TASKLET_H__ */
diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h
new file mode 100644
index 000000000..37b5c9e9d
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_trace.h
@@ -0,0 +1,770 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM i915
+
+#if !defined(_I915_TRACE_H_) || defined(TRACE_HEADER_MULTI_READ)
+#define _I915_TRACE_H_
+
+#include <linux/stringify.h>
+#include <linux/types.h>
+#include <linux/tracepoint.h>
+
+#include <drm/drm_drv.h>
+
+#include "gt/intel_engine.h"
+
+#include "i915_drv.h"
+#include "i915_irq.h"
+
+/* object tracking */
+
+TRACE_EVENT(i915_gem_object_create,
+	    TP_PROTO(struct drm_i915_gem_object *obj),
+	    TP_ARGS(obj),
+
+	    TP_STRUCT__entry(
+			     __field(struct drm_i915_gem_object *, obj)
+			     __field(u64, size)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->obj = obj;
+			   __entry->size = obj->base.size;
+			   ),
+
+	    TP_printk("obj=%p, size=0x%llx", __entry->obj, __entry->size)
+);
+
+TRACE_EVENT(i915_gem_shrink,
+	    TP_PROTO(struct drm_i915_private *i915, unsigned long target, unsigned flags),
+	    TP_ARGS(i915, target, flags),
+
+	    TP_STRUCT__entry(
+			     __field(int, dev)
+			     __field(unsigned long, target)
+			     __field(unsigned, flags)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->dev = i915->drm.primary->index;
+			   __entry->target = target;
+			   __entry->flags = flags;
+			   ),
+
+	    TP_printk("dev=%d, target=%lu, flags=%x",
+		      __entry->dev, __entry->target, __entry->flags)
+);
+
+TRACE_EVENT(i915_vma_bind,
+	    TP_PROTO(struct i915_vma *vma, unsigned flags),
+	    TP_ARGS(vma, flags),
+
+	    TP_STRUCT__entry(
+			     __field(struct drm_i915_gem_object *, obj)
+			     __field(struct i915_address_space *, vm)
+			     __field(u64, offset)
+			     __field(u64, size)
+			     __field(unsigned, flags)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->obj = vma->obj;
+			   __entry->vm = vma->vm;
+			   __entry->offset = vma->node.start;
+			   __entry->size = vma->node.size;
+			   __entry->flags = flags;
+			   ),
+
+	    TP_printk("obj=%p, offset=0x%016llx size=0x%llx%s vm=%p",
+		      __entry->obj, __entry->offset, __entry->size,
+		      __entry->flags & PIN_MAPPABLE ? ", mappable" : "",
+		      __entry->vm)
+);
+
+TRACE_EVENT(i915_vma_unbind,
+	    TP_PROTO(struct i915_vma *vma),
+	    TP_ARGS(vma),
+
+	    TP_STRUCT__entry(
+			     __field(struct drm_i915_gem_object *, obj)
+			     __field(struct i915_address_space *, vm)
+			     __field(u64, offset)
+			     __field(u64, size)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->obj = vma->obj;
+			   __entry->vm = vma->vm;
+			   __entry->offset = vma->node.start;
+			   __entry->size = vma->node.size;
+			   ),
+
+	    TP_printk("obj=%p, offset=0x%016llx size=0x%llx vm=%p",
+		      __entry->obj, __entry->offset, __entry->size, __entry->vm)
+);
+
+TRACE_EVENT(i915_gem_object_pwrite,
+	    TP_PROTO(struct drm_i915_gem_object *obj, u64 offset, u64 len),
+	    TP_ARGS(obj, offset, len),
+
+	    TP_STRUCT__entry(
+			     __field(struct drm_i915_gem_object *, obj)
+			     __field(u64, offset)
+			     __field(u64, len)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->obj = obj;
+			   __entry->offset = offset;
+			   __entry->len = len;
+			   ),
+
+	    TP_printk("obj=%p, offset=0x%llx, len=0x%llx",
+		      __entry->obj, __entry->offset, __entry->len)
+);
+
+TRACE_EVENT(i915_gem_object_pread,
+	    TP_PROTO(struct drm_i915_gem_object *obj, u64 offset, u64 len),
+	    TP_ARGS(obj, offset, len),
+
+	    TP_STRUCT__entry(
+			     __field(struct drm_i915_gem_object *, obj)
+			     __field(u64, offset)
+			     __field(u64, len)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->obj = obj;
+			   __entry->offset = offset;
+			   __entry->len = len;
+			   ),
+
+	    TP_printk("obj=%p, offset=0x%llx, len=0x%llx",
+		      __entry->obj, __entry->offset, __entry->len)
+);
+
+TRACE_EVENT(i915_gem_object_fault,
+	    TP_PROTO(struct drm_i915_gem_object *obj, u64 index, bool gtt, bool write),
+	    TP_ARGS(obj, index, gtt, write),
+
+	    TP_STRUCT__entry(
+			     __field(struct drm_i915_gem_object *, obj)
+			     __field(u64, index)
+			     __field(bool, gtt)
+			     __field(bool, write)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->obj = obj;
+			   __entry->index = index;
+			   __entry->gtt = gtt;
+			   __entry->write = write;
+			   ),
+
+	    TP_printk("obj=%p, %s index=%llu %s",
+		      __entry->obj,
+		      __entry->gtt ? "GTT" : "CPU",
+		      __entry->index,
+		      __entry->write ? ", writable" : "")
+);
+
+DECLARE_EVENT_CLASS(i915_gem_object,
+	    TP_PROTO(struct drm_i915_gem_object *obj),
+	    TP_ARGS(obj),
+
+	    TP_STRUCT__entry(
+			     __field(struct drm_i915_gem_object *, obj)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->obj = obj;
+			   ),
+
+	    TP_printk("obj=%p", __entry->obj)
+);
+
+DEFINE_EVENT(i915_gem_object, i915_gem_object_clflush,
+	     TP_PROTO(struct drm_i915_gem_object *obj),
+	     TP_ARGS(obj)
+);
+
+DEFINE_EVENT(i915_gem_object, i915_gem_object_destroy,
+	    TP_PROTO(struct drm_i915_gem_object *obj),
+	    TP_ARGS(obj)
+);
+
+TRACE_EVENT(i915_gem_evict,
+	    TP_PROTO(struct i915_address_space *vm, u64 size, u64 align, unsigned int flags),
+	    TP_ARGS(vm, size, align, flags),
+
+	    TP_STRUCT__entry(
+			     __field(u32, dev)
+			     __field(struct i915_address_space *, vm)
+			     __field(u64, size)
+			     __field(u64, align)
+			     __field(unsigned int, flags)
+			    ),
+
+	    TP_fast_assign(
+			   __entry->dev = vm->i915->drm.primary->index;
+			   __entry->vm = vm;
+			   __entry->size = size;
+			   __entry->align = align;
+			   __entry->flags = flags;
+			  ),
+
+	    TP_printk("dev=%d, vm=%p, size=0x%llx, align=0x%llx %s",
+		      __entry->dev, __entry->vm, __entry->size, __entry->align,
+		      __entry->flags & PIN_MAPPABLE ? ", mappable" : "")
+);
+
+TRACE_EVENT(i915_gem_evict_node,
+	    TP_PROTO(struct i915_address_space *vm, struct drm_mm_node *node, unsigned int flags),
+	    TP_ARGS(vm, node, flags),
+
+	    TP_STRUCT__entry(
+			     __field(u32, dev)
+			     __field(struct i915_address_space *, vm)
+			     __field(u64, start)
+			     __field(u64, size)
+			     __field(unsigned long, color)
+			     __field(unsigned int, flags)
+			    ),
+
+	    TP_fast_assign(
+			   __entry->dev = vm->i915->drm.primary->index;
+			   __entry->vm = vm;
+			   __entry->start = node->start;
+			   __entry->size = node->size;
+			   __entry->color = node->color;
+			   __entry->flags = flags;
+			  ),
+
+	    TP_printk("dev=%d, vm=%p, start=0x%llx size=0x%llx, color=0x%lx, flags=%x",
+		      __entry->dev, __entry->vm,
+		      __entry->start, __entry->size,
+		      __entry->color, __entry->flags)
+);
+
+TRACE_EVENT(i915_gem_evict_vm,
+	    TP_PROTO(struct i915_address_space *vm),
+	    TP_ARGS(vm),
+
+	    TP_STRUCT__entry(
+			     __field(u32, dev)
+			     __field(struct i915_address_space *, vm)
+			    ),
+
+	    TP_fast_assign(
+			   __entry->dev = vm->i915->drm.primary->index;
+			   __entry->vm = vm;
+			  ),
+
+	    TP_printk("dev=%d, vm=%p", __entry->dev, __entry->vm)
+);
+
+TRACE_EVENT(i915_request_queue,
+	    TP_PROTO(struct i915_request *rq, u32 flags),
+	    TP_ARGS(rq, flags),
+
+	    TP_STRUCT__entry(
+			     __field(u32, dev)
+			     __field(u64, ctx)
+			     __field(u16, class)
+			     __field(u16, instance)
+			     __field(u32, seqno)
+			     __field(u32, flags)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->dev = rq->engine->i915->drm.primary->index;
+			   __entry->class = rq->engine->uabi_class;
+			   __entry->instance = rq->engine->uabi_instance;
+			   __entry->ctx = rq->fence.context;
+			   __entry->seqno = rq->fence.seqno;
+			   __entry->flags = flags;
+			   ),
+
+	    TP_printk("dev=%u, engine=%u:%u, ctx=%llu, seqno=%u, flags=0x%x",
+		      __entry->dev, __entry->class, __entry->instance,
+		      __entry->ctx, __entry->seqno, __entry->flags)
+);
+
+DECLARE_EVENT_CLASS(i915_request,
+	    TP_PROTO(struct i915_request *rq),
+	    TP_ARGS(rq),
+
+	    TP_STRUCT__entry(
+			     __field(u32, dev)
+			     __field(u64, ctx)
+			     __field(u16, class)
+			     __field(u16, instance)
+			     __field(u32, seqno)
+			     __field(u32, tail)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->dev = rq->engine->i915->drm.primary->index;
+			   __entry->class = rq->engine->uabi_class;
+			   __entry->instance = rq->engine->uabi_instance;
+			   __entry->ctx = rq->fence.context;
+			   __entry->seqno = rq->fence.seqno;
+			   __entry->tail = rq->tail;
+			   ),
+
+	    TP_printk("dev=%u, engine=%u:%u, ctx=%llu, seqno=%u, tail=%u",
+		      __entry->dev, __entry->class, __entry->instance,
+		      __entry->ctx, __entry->seqno, __entry->tail)
+);
+
+DEFINE_EVENT(i915_request, i915_request_add,
+	     TP_PROTO(struct i915_request *rq),
+	     TP_ARGS(rq)
+);
+
+#if defined(CONFIG_DRM_I915_LOW_LEVEL_TRACEPOINTS)
+DEFINE_EVENT(i915_request, i915_request_guc_submit,
+	     TP_PROTO(struct i915_request *rq),
+	     TP_ARGS(rq)
+);
+
+DEFINE_EVENT(i915_request, i915_request_submit,
+	     TP_PROTO(struct i915_request *rq),
+	     TP_ARGS(rq)
+);
+
+DEFINE_EVENT(i915_request, i915_request_execute,
+	     TP_PROTO(struct i915_request *rq),
+	     TP_ARGS(rq)
+);
+
+TRACE_EVENT(i915_request_in,
+	    TP_PROTO(struct i915_request *rq, unsigned int port),
+	    TP_ARGS(rq, port),
+
+	    TP_STRUCT__entry(
+			     __field(u32, dev)
+			     __field(u64, ctx)
+			     __field(u16, class)
+			     __field(u16, instance)
+			     __field(u32, seqno)
+			     __field(u32, port)
+			     __field(s32, prio)
+			    ),
+
+	    TP_fast_assign(
+			   __entry->dev = rq->engine->i915->drm.primary->index;
+			   __entry->class = rq->engine->uabi_class;
+			   __entry->instance = rq->engine->uabi_instance;
+			   __entry->ctx = rq->fence.context;
+			   __entry->seqno = rq->fence.seqno;
+			   __entry->prio = rq->sched.attr.priority;
+			   __entry->port = port;
+			   ),
+
+	    TP_printk("dev=%u, engine=%u:%u, ctx=%llu, seqno=%u, prio=%d, port=%u",
+		      __entry->dev, __entry->class, __entry->instance,
+		      __entry->ctx, __entry->seqno,
+		      __entry->prio, __entry->port)
+);
+
+TRACE_EVENT(i915_request_out,
+	    TP_PROTO(struct i915_request *rq),
+	    TP_ARGS(rq),
+
+	    TP_STRUCT__entry(
+			     __field(u32, dev)
+			     __field(u64, ctx)
+			     __field(u16, class)
+			     __field(u16, instance)
+			     __field(u32, seqno)
+			     __field(u32, completed)
+			    ),
+
+	    TP_fast_assign(
+			   __entry->dev = rq->engine->i915->drm.primary->index;
+			   __entry->class = rq->engine->uabi_class;
+			   __entry->instance = rq->engine->uabi_instance;
+			   __entry->ctx = rq->fence.context;
+			   __entry->seqno = rq->fence.seqno;
+			   __entry->completed = i915_request_completed(rq);
+			   ),
+
+		    TP_printk("dev=%u, engine=%u:%u, ctx=%llu, seqno=%u, completed?=%u",
+			      __entry->dev, __entry->class, __entry->instance,
+			      __entry->ctx, __entry->seqno, __entry->completed)
+);
+
+DECLARE_EVENT_CLASS(intel_context,
+		    TP_PROTO(struct intel_context *ce),
+		    TP_ARGS(ce),
+
+		    TP_STRUCT__entry(
+			     __field(u32, guc_id)
+			     __field(int, pin_count)
+			     __field(u32, sched_state)
+			     __field(u8, guc_prio)
+			     ),
+
+		    TP_fast_assign(
+			   __entry->guc_id = ce->guc_id.id;
+			   __entry->pin_count = atomic_read(&ce->pin_count);
+			   __entry->sched_state = ce->guc_state.sched_state;
+			   __entry->guc_prio = ce->guc_state.prio;
+			   ),
+
+		    TP_printk("guc_id=%d, pin_count=%d sched_state=0x%x, guc_prio=%u",
+			      __entry->guc_id, __entry->pin_count,
+			      __entry->sched_state,
+			      __entry->guc_prio)
+);
+
+DEFINE_EVENT(intel_context, intel_context_set_prio,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_reset,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_ban,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_register,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_deregister,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_deregister_done,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_sched_enable,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_sched_disable,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_sched_done,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_create,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_fence_release,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_free,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_steal_guc_id,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_do_pin,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+DEFINE_EVENT(intel_context, intel_context_do_unpin,
+	     TP_PROTO(struct intel_context *ce),
+	     TP_ARGS(ce)
+);
+
+#else
+#if !defined(TRACE_HEADER_MULTI_READ)
+static inline void
+trace_i915_request_guc_submit(struct i915_request *rq)
+{
+}
+
+static inline void
+trace_i915_request_submit(struct i915_request *rq)
+{
+}
+
+static inline void
+trace_i915_request_execute(struct i915_request *rq)
+{
+}
+
+static inline void
+trace_i915_request_in(struct i915_request *rq, unsigned int port)
+{
+}
+
+static inline void
+trace_i915_request_out(struct i915_request *rq)
+{
+}
+
+static inline void
+trace_intel_context_set_prio(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_reset(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_ban(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_register(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_deregister(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_deregister_done(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_sched_enable(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_sched_disable(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_sched_done(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_create(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_fence_release(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_free(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_steal_guc_id(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_do_pin(struct intel_context *ce)
+{
+}
+
+static inline void
+trace_intel_context_do_unpin(struct intel_context *ce)
+{
+}
+#endif
+#endif
+
+DEFINE_EVENT(i915_request, i915_request_retire,
+	    TP_PROTO(struct i915_request *rq),
+	    TP_ARGS(rq)
+);
+
+TRACE_EVENT(i915_request_wait_begin,
+	    TP_PROTO(struct i915_request *rq, unsigned int flags),
+	    TP_ARGS(rq, flags),
+
+	    TP_STRUCT__entry(
+			     __field(u32, dev)
+			     __field(u64, ctx)
+			     __field(u16, class)
+			     __field(u16, instance)
+			     __field(u32, seqno)
+			     __field(unsigned int, flags)
+			     ),
+
+	    /* NB: the blocking information is racy since mutex_is_locked
+	     * doesn't check that the current thread holds the lock. The only
+	     * other option would be to pass the boolean information of whether
+	     * or not the class was blocking down through the stack which is
+	     * less desirable.
+	     */
+	    TP_fast_assign(
+			   __entry->dev = rq->engine->i915->drm.primary->index;
+			   __entry->class = rq->engine->uabi_class;
+			   __entry->instance = rq->engine->uabi_instance;
+			   __entry->ctx = rq->fence.context;
+			   __entry->seqno = rq->fence.seqno;
+			   __entry->flags = flags;
+			   ),
+
+	    TP_printk("dev=%u, engine=%u:%u, ctx=%llu, seqno=%u, flags=0x%x",
+		      __entry->dev, __entry->class, __entry->instance,
+		      __entry->ctx, __entry->seqno,
+		      __entry->flags)
+);
+
+DEFINE_EVENT(i915_request, i915_request_wait_end,
+	    TP_PROTO(struct i915_request *rq),
+	    TP_ARGS(rq)
+);
+
+TRACE_EVENT_CONDITION(i915_reg_rw,
+	TP_PROTO(bool write, i915_reg_t reg, u64 val, int len, bool trace),
+
+	TP_ARGS(write, reg, val, len, trace),
+
+	TP_CONDITION(trace),
+
+	TP_STRUCT__entry(
+		__field(u64, val)
+		__field(u32, reg)
+		__field(u16, write)
+		__field(u16, len)
+		),
+
+	TP_fast_assign(
+		__entry->val = (u64)val;
+		__entry->reg = i915_mmio_reg_offset(reg);
+		__entry->write = write;
+		__entry->len = len;
+		),
+
+	TP_printk("%s reg=0x%x, len=%d, val=(0x%x, 0x%x)",
+		__entry->write ? "write" : "read",
+		__entry->reg, __entry->len,
+		(u32)(__entry->val & 0xffffffff),
+		(u32)(__entry->val >> 32))
+);
+
+TRACE_EVENT(intel_gpu_freq_change,
+	    TP_PROTO(u32 freq),
+	    TP_ARGS(freq),
+
+	    TP_STRUCT__entry(
+			     __field(u32, freq)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->freq = freq;
+			   ),
+
+	    TP_printk("new_freq=%u", __entry->freq)
+);
+
+/**
+ * DOC: i915_ppgtt_create and i915_ppgtt_release tracepoints
+ *
+ * With full ppgtt enabled each process using drm will allocate at least one
+ * translation table. With these traces it is possible to keep track of the
+ * allocation and of the lifetime of the tables; this can be used during
+ * testing/debug to verify that we are not leaking ppgtts.
+ * These traces identify the ppgtt through the vm pointer, which is also printed
+ * by the i915_vma_bind and i915_vma_unbind tracepoints.
+ */
+DECLARE_EVENT_CLASS(i915_ppgtt,
+	TP_PROTO(struct i915_address_space *vm),
+	TP_ARGS(vm),
+
+	TP_STRUCT__entry(
+			__field(struct i915_address_space *, vm)
+			__field(u32, dev)
+	),
+
+	TP_fast_assign(
+			__entry->vm = vm;
+			__entry->dev = vm->i915->drm.primary->index;
+	),
+
+	TP_printk("dev=%u, vm=%p", __entry->dev, __entry->vm)
+)
+
+DEFINE_EVENT(i915_ppgtt, i915_ppgtt_create,
+	TP_PROTO(struct i915_address_space *vm),
+	TP_ARGS(vm)
+);
+
+DEFINE_EVENT(i915_ppgtt, i915_ppgtt_release,
+	TP_PROTO(struct i915_address_space *vm),
+	TP_ARGS(vm)
+);
+
+/**
+ * DOC: i915_context_create and i915_context_free tracepoints
+ *
+ * These tracepoints are used to track creation and deletion of contexts.
+ * If full ppgtt is enabled, they also print the address of the vm assigned to
+ * the context.
+ */
+DECLARE_EVENT_CLASS(i915_context,
+	TP_PROTO(struct i915_gem_context *ctx),
+	TP_ARGS(ctx),
+
+	TP_STRUCT__entry(
+			__field(u32, dev)
+			__field(struct i915_gem_context *, ctx)
+			__field(struct i915_address_space *, vm)
+	),
+
+	TP_fast_assign(
+			__entry->dev = ctx->i915->drm.primary->index;
+			__entry->ctx = ctx;
+			__entry->vm = ctx->vm;
+	),
+
+	TP_printk("dev=%u, ctx=%p, ctx_vm=%p",
+		  __entry->dev, __entry->ctx, __entry->vm)
+)
+
+DEFINE_EVENT(i915_context, i915_context_create,
+	TP_PROTO(struct i915_gem_context *ctx),
+	TP_ARGS(ctx)
+);
+
+DEFINE_EVENT(i915_context, i915_context_free,
+	TP_PROTO(struct i915_gem_context *ctx),
+	TP_ARGS(ctx)
+);
+
+#endif /* _I915_TRACE_H_ */
+
+/* This part must be outside protection */
+#undef TRACE_INCLUDE_PATH
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_PATH ../../drivers/gpu/drm/i915
+#define TRACE_INCLUDE_FILE i915_trace
+#include <trace/define_trace.h>
diff --git a/drivers/gpu/drm/i915/i915_trace_points.c b/drivers/gpu/drm/i915/i915_trace_points.c
new file mode 100644
index 000000000..463a71779
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_trace_points.c
@@ -0,0 +1,14 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright © 2009 Intel Corporation
+ *
+ * Authors:
+ *    Chris Wilson <chris@chris-wilson.co.uk>
+ */
+
+#include "i915_drv.h"
+
+#ifndef __CHECKER__
+#define CREATE_TRACE_POINTS
+#include "i915_trace.h"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_ttm_buddy_manager.c b/drivers/gpu/drm/i915/i915_ttm_buddy_manager.c
new file mode 100644
index 000000000..e19452f0e
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_ttm_buddy_manager.c
@@ -0,0 +1,464 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/slab.h>
+
+#include <drm/ttm/ttm_bo_driver.h>
+#include <drm/ttm/ttm_placement.h>
+
+#include <drm/drm_buddy.h>
+
+#include "i915_ttm_buddy_manager.h"
+
+#include "i915_gem.h"
+
+struct i915_ttm_buddy_manager {
+	struct ttm_resource_manager manager;
+	struct drm_buddy mm;
+	struct list_head reserved;
+	struct mutex lock;
+	unsigned long visible_size;
+	unsigned long visible_avail;
+	unsigned long visible_reserved;
+	u64 default_page_size;
+};
+
+static struct i915_ttm_buddy_manager *
+to_buddy_manager(struct ttm_resource_manager *man)
+{
+	return container_of(man, struct i915_ttm_buddy_manager, manager);
+}
+
+static int i915_ttm_buddy_man_alloc(struct ttm_resource_manager *man,
+				    struct ttm_buffer_object *bo,
+				    const struct ttm_place *place,
+				    struct ttm_resource **res)
+{
+	struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
+	struct i915_ttm_buddy_resource *bman_res;
+	struct drm_buddy *mm = &bman->mm;
+	unsigned long n_pages, lpfn;
+	u64 min_page_size;
+	u64 size;
+	int err;
+
+	lpfn = place->lpfn;
+	if (!lpfn)
+		lpfn = man->size;
+
+	bman_res = kzalloc(sizeof(*bman_res), GFP_KERNEL);
+	if (!bman_res)
+		return -ENOMEM;
+
+	ttm_resource_init(bo, place, &bman_res->base);
+	INIT_LIST_HEAD(&bman_res->blocks);
+	bman_res->mm = mm;
+
+	if (place->flags & TTM_PL_FLAG_TOPDOWN)
+		bman_res->flags |= DRM_BUDDY_TOPDOWN_ALLOCATION;
+
+	if (place->fpfn || lpfn != man->size)
+		bman_res->flags |= DRM_BUDDY_RANGE_ALLOCATION;
+
+	GEM_BUG_ON(!bman_res->base.num_pages);
+	size = bman_res->base.num_pages << PAGE_SHIFT;
+
+	min_page_size = bman->default_page_size;
+	if (bo->page_alignment)
+		min_page_size = bo->page_alignment << PAGE_SHIFT;
+
+	GEM_BUG_ON(min_page_size < mm->chunk_size);
+	GEM_BUG_ON(!IS_ALIGNED(size, min_page_size));
+
+	if (place->fpfn + bman_res->base.num_pages != place->lpfn &&
+	    place->flags & TTM_PL_FLAG_CONTIGUOUS) {
+		unsigned long pages;
+
+		size = roundup_pow_of_two(size);
+		min_page_size = size;
+
+		pages = size >> ilog2(mm->chunk_size);
+		if (pages > lpfn)
+			lpfn = pages;
+	}
+
+	if (size > lpfn << PAGE_SHIFT) {
+		err = -E2BIG;
+		goto err_free_res;
+	}
+
+	n_pages = size >> ilog2(mm->chunk_size);
+
+	mutex_lock(&bman->lock);
+	if (lpfn <= bman->visible_size && n_pages > bman->visible_avail) {
+		mutex_unlock(&bman->lock);
+		err = -ENOSPC;
+		goto err_free_res;
+	}
+
+	err = drm_buddy_alloc_blocks(mm, (u64)place->fpfn << PAGE_SHIFT,
+				     (u64)lpfn << PAGE_SHIFT,
+				     (u64)n_pages << PAGE_SHIFT,
+				     min_page_size,
+				     &bman_res->blocks,
+				     bman_res->flags);
+	if (unlikely(err))
+		goto err_free_blocks;
+
+	if (place->flags & TTM_PL_FLAG_CONTIGUOUS) {
+		u64 original_size = (u64)bman_res->base.num_pages << PAGE_SHIFT;
+
+		drm_buddy_block_trim(mm,
+				     original_size,
+				     &bman_res->blocks);
+	}
+
+	if (lpfn <= bman->visible_size) {
+		bman_res->used_visible_size = bman_res->base.num_pages;
+	} else {
+		struct drm_buddy_block *block;
+
+		list_for_each_entry(block, &bman_res->blocks, link) {
+			unsigned long start =
+				drm_buddy_block_offset(block) >> PAGE_SHIFT;
+
+			if (start < bman->visible_size) {
+				unsigned long end = start +
+					(drm_buddy_block_size(mm, block) >> PAGE_SHIFT);
+
+				bman_res->used_visible_size +=
+					min(end, bman->visible_size) - start;
+			}
+		}
+	}
+
+	if (bman_res->used_visible_size)
+		bman->visible_avail -= bman_res->used_visible_size;
+
+	mutex_unlock(&bman->lock);
+
+	if (place->lpfn - place->fpfn == n_pages)
+		bman_res->base.start = place->fpfn;
+	else if (lpfn <= bman->visible_size)
+		bman_res->base.start = 0;
+	else
+		bman_res->base.start = bman->visible_size;
+
+	*res = &bman_res->base;
+	return 0;
+
+err_free_blocks:
+	drm_buddy_free_list(mm, &bman_res->blocks);
+	mutex_unlock(&bman->lock);
+err_free_res:
+	ttm_resource_fini(man, &bman_res->base);
+	kfree(bman_res);
+	return err;
+}
+
+static void i915_ttm_buddy_man_free(struct ttm_resource_manager *man,
+				    struct ttm_resource *res)
+{
+	struct i915_ttm_buddy_resource *bman_res = to_ttm_buddy_resource(res);
+	struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
+
+	mutex_lock(&bman->lock);
+	drm_buddy_free_list(&bman->mm, &bman_res->blocks);
+	bman->visible_avail += bman_res->used_visible_size;
+	mutex_unlock(&bman->lock);
+
+	ttm_resource_fini(man, res);
+	kfree(bman_res);
+}
+
+static bool i915_ttm_buddy_man_intersects(struct ttm_resource_manager *man,
+					  struct ttm_resource *res,
+					  const struct ttm_place *place,
+					  size_t size)
+{
+	struct i915_ttm_buddy_resource *bman_res = to_ttm_buddy_resource(res);
+	struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
+	struct drm_buddy *mm = &bman->mm;
+	struct drm_buddy_block *block;
+
+	if (!place->fpfn && !place->lpfn)
+		return true;
+
+	GEM_BUG_ON(!place->lpfn);
+
+	/*
+	 * If we just want something mappable then we can quickly check
+	 * if the current victim resource is using any of the CPU
+	 * visible portion.
+	 */
+	if (!place->fpfn &&
+	    place->lpfn == i915_ttm_buddy_man_visible_size(man))
+		return bman_res->used_visible_size > 0;
+
+	/* Check each drm buddy block individually */
+	list_for_each_entry(block, &bman_res->blocks, link) {
+		unsigned long fpfn =
+			drm_buddy_block_offset(block) >> PAGE_SHIFT;
+		unsigned long lpfn = fpfn +
+			(drm_buddy_block_size(mm, block) >> PAGE_SHIFT);
+
+		if (place->fpfn < lpfn && place->lpfn > fpfn)
+			return true;
+	}
+
+	return false;
+}
+
+static bool i915_ttm_buddy_man_compatible(struct ttm_resource_manager *man,
+					  struct ttm_resource *res,
+					  const struct ttm_place *place,
+					  size_t size)
+{
+	struct i915_ttm_buddy_resource *bman_res = to_ttm_buddy_resource(res);
+	struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
+	struct drm_buddy *mm = &bman->mm;
+	struct drm_buddy_block *block;
+
+	if (!place->fpfn && !place->lpfn)
+		return true;
+
+	GEM_BUG_ON(!place->lpfn);
+
+	if (!place->fpfn &&
+	    place->lpfn == i915_ttm_buddy_man_visible_size(man))
+		return bman_res->used_visible_size == res->num_pages;
+
+	/* Check each drm buddy block individually */
+	list_for_each_entry(block, &bman_res->blocks, link) {
+		unsigned long fpfn =
+			drm_buddy_block_offset(block) >> PAGE_SHIFT;
+		unsigned long lpfn = fpfn +
+			(drm_buddy_block_size(mm, block) >> PAGE_SHIFT);
+
+		if (fpfn < place->fpfn || lpfn > place->lpfn)
+			return false;
+	}
+
+	return true;
+}
+
+static void i915_ttm_buddy_man_debug(struct ttm_resource_manager *man,
+				     struct drm_printer *printer)
+{
+	struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
+	struct drm_buddy_block *block;
+
+	mutex_lock(&bman->lock);
+	drm_printf(printer, "default_page_size: %lluKiB\n",
+		   bman->default_page_size >> 10);
+	drm_printf(printer, "visible_avail: %lluMiB\n",
+		   (u64)bman->visible_avail << PAGE_SHIFT >> 20);
+	drm_printf(printer, "visible_size: %lluMiB\n",
+		   (u64)bman->visible_size << PAGE_SHIFT >> 20);
+	drm_printf(printer, "visible_reserved: %lluMiB\n",
+		   (u64)bman->visible_reserved << PAGE_SHIFT >> 20);
+
+	drm_buddy_print(&bman->mm, printer);
+
+	drm_printf(printer, "reserved:\n");
+	list_for_each_entry(block, &bman->reserved, link)
+		drm_buddy_block_print(&bman->mm, block, printer);
+	mutex_unlock(&bman->lock);
+}
+
+static const struct ttm_resource_manager_func i915_ttm_buddy_manager_func = {
+	.alloc = i915_ttm_buddy_man_alloc,
+	.free = i915_ttm_buddy_man_free,
+	.intersects = i915_ttm_buddy_man_intersects,
+	.compatible = i915_ttm_buddy_man_compatible,
+	.debug = i915_ttm_buddy_man_debug,
+};
+
+/**
+ * i915_ttm_buddy_man_init - Setup buddy allocator based ttm manager
+ * @bdev: The ttm device
+ * @type: Memory type we want to manage
+ * @use_tt: Set use_tt for the manager
+ * @size: The size in bytes to manage
+ * @visible_size: The CPU visible size in bytes to manage
+ * @default_page_size: The default minimum page size in bytes for allocations,
+ * this must be at least as large as @chunk_size, and can be overridden by
+ * setting the BO page_alignment, to be larger or smaller as needed.
+ * @chunk_size: The minimum page size in bytes for our allocations i.e
+ * order-zero
+ *
+ * Note that the starting address is assumed to be zero here, since this
+ * simplifies keeping the property where allocated blocks having natural
+ * power-of-two alignment. So long as the real starting address is some large
+ * power-of-two, or naturally start from zero, then this should be fine.  Also
+ * the &i915_ttm_buddy_man_reserve interface can be used to preserve alignment
+ * if say there is some unusable range from the start of the region. We can
+ * revisit this in the future and make the interface accept an actual starting
+ * offset and let it take care of the rest.
+ *
+ * Note that if the @size is not aligned to the @chunk_size then we perform the
+ * required rounding to get the usable size. The final size in pages can be
+ * taken from &ttm_resource_manager.size.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int i915_ttm_buddy_man_init(struct ttm_device *bdev,
+			    unsigned int type, bool use_tt,
+			    u64 size, u64 visible_size, u64 default_page_size,
+			    u64 chunk_size)
+{
+	struct ttm_resource_manager *man;
+	struct i915_ttm_buddy_manager *bman;
+	int err;
+
+	bman = kzalloc(sizeof(*bman), GFP_KERNEL);
+	if (!bman)
+		return -ENOMEM;
+
+	err = drm_buddy_init(&bman->mm, size, chunk_size);
+	if (err)
+		goto err_free_bman;
+
+	mutex_init(&bman->lock);
+	INIT_LIST_HEAD(&bman->reserved);
+	GEM_BUG_ON(default_page_size < chunk_size);
+	bman->default_page_size = default_page_size;
+	bman->visible_size = visible_size >> PAGE_SHIFT;
+	bman->visible_avail = bman->visible_size;
+
+	man = &bman->manager;
+	man->use_tt = use_tt;
+	man->func = &i915_ttm_buddy_manager_func;
+	ttm_resource_manager_init(man, bdev, bman->mm.size >> PAGE_SHIFT);
+
+	ttm_resource_manager_set_used(man, true);
+	ttm_set_driver_manager(bdev, type, man);
+
+	return 0;
+
+err_free_bman:
+	kfree(bman);
+	return err;
+}
+
+/**
+ * i915_ttm_buddy_man_fini - Destroy the buddy allocator ttm manager
+ * @bdev: The ttm device
+ * @type: Memory type we want to manage
+ *
+ * Note that if we reserved anything with &i915_ttm_buddy_man_reserve, this will
+ * also be freed for us here.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int i915_ttm_buddy_man_fini(struct ttm_device *bdev, unsigned int type)
+{
+	struct ttm_resource_manager *man = ttm_manager_type(bdev, type);
+	struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
+	struct drm_buddy *mm = &bman->mm;
+	int ret;
+
+	ttm_resource_manager_set_used(man, false);
+
+	ret = ttm_resource_manager_evict_all(bdev, man);
+	if (ret)
+		return ret;
+
+	ttm_set_driver_manager(bdev, type, NULL);
+
+	mutex_lock(&bman->lock);
+	drm_buddy_free_list(mm, &bman->reserved);
+	drm_buddy_fini(mm);
+	bman->visible_avail += bman->visible_reserved;
+	WARN_ON_ONCE(bman->visible_avail != bman->visible_size);
+	mutex_unlock(&bman->lock);
+
+	ttm_resource_manager_cleanup(man);
+	kfree(bman);
+
+	return 0;
+}
+
+/**
+ * i915_ttm_buddy_man_reserve - Reserve address range
+ * @man: The buddy allocator ttm manager
+ * @start: The offset in bytes, where the region start is assumed to be zero
+ * @size: The size in bytes
+ *
+ * Note that the starting address for the region is always assumed to be zero.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int i915_ttm_buddy_man_reserve(struct ttm_resource_manager *man,
+			       u64 start, u64 size)
+{
+	struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
+	struct drm_buddy *mm = &bman->mm;
+	unsigned long fpfn = start >> PAGE_SHIFT;
+	unsigned long flags = 0;
+	int ret;
+
+	flags |= DRM_BUDDY_RANGE_ALLOCATION;
+
+	mutex_lock(&bman->lock);
+	ret = drm_buddy_alloc_blocks(mm, start,
+				     start + size,
+				     size, mm->chunk_size,
+				     &bman->reserved,
+				     flags);
+
+	if (fpfn < bman->visible_size) {
+		unsigned long lpfn = fpfn + (size >> PAGE_SHIFT);
+		unsigned long visible = min(lpfn, bman->visible_size) - fpfn;
+
+		bman->visible_reserved += visible;
+		bman->visible_avail -= visible;
+	}
+	mutex_unlock(&bman->lock);
+
+	return ret;
+}
+
+/**
+ * i915_ttm_buddy_man_visible_size - Return the size of the CPU visible portion
+ * in pages.
+ * @man: The buddy allocator ttm manager
+ */
+u64 i915_ttm_buddy_man_visible_size(struct ttm_resource_manager *man)
+{
+	struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
+
+	return bman->visible_size;
+}
+
+/**
+ * i915_ttm_buddy_man_avail - Query the avail tracking for the manager.
+ *
+ * @man: The buddy allocator ttm manager
+ * @avail: The total available memory in pages for the entire manager.
+ * @visible_avail: The total available memory in pages for the CPU visible
+ * portion. Note that this will always give the same value as @avail on
+ * configurations that don't have a small BAR.
+ */
+void i915_ttm_buddy_man_avail(struct ttm_resource_manager *man,
+			      u64 *avail, u64 *visible_avail)
+{
+	struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
+
+	mutex_lock(&bman->lock);
+	*avail = bman->mm.avail >> PAGE_SHIFT;
+	*visible_avail = bman->visible_avail;
+	mutex_unlock(&bman->lock);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+void i915_ttm_buddy_man_force_visible_size(struct ttm_resource_manager *man,
+					   u64 size)
+{
+	struct i915_ttm_buddy_manager *bman = to_buddy_manager(man);
+
+	bman->visible_size = size;
+}
+#endif
diff --git a/drivers/gpu/drm/i915/i915_ttm_buddy_manager.h b/drivers/gpu/drm/i915/i915_ttm_buddy_manager.h
new file mode 100644
index 000000000..d64620712
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_ttm_buddy_manager.h
@@ -0,0 +1,72 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_TTM_BUDDY_MANAGER_H__
+#define __I915_TTM_BUDDY_MANAGER_H__
+
+#include <linux/list.h>
+#include <linux/types.h>
+
+#include <drm/ttm/ttm_resource.h>
+
+struct ttm_device;
+struct ttm_resource_manager;
+struct drm_buddy;
+
+/**
+ * struct i915_ttm_buddy_resource
+ *
+ * @base: struct ttm_resource base class we extend
+ * @blocks: the list of struct i915_buddy_block for this resource/allocation
+ * @flags: DRM_BUDDY_*_ALLOCATION flags
+ * @used_visible_size: How much of this resource, if any, uses the CPU visible
+ * portion, in pages.
+ * @mm: the struct i915_buddy_mm for this resource
+ *
+ * Extends the struct ttm_resource to manage an address space allocation with
+ * one or more struct i915_buddy_block.
+ */
+struct i915_ttm_buddy_resource {
+	struct ttm_resource base;
+	struct list_head blocks;
+	unsigned long flags;
+	unsigned long used_visible_size;
+	struct drm_buddy *mm;
+};
+
+/**
+ * to_ttm_buddy_resource
+ *
+ * @res: the resource to upcast
+ *
+ * Upcast the struct ttm_resource object into a struct i915_ttm_buddy_resource.
+ */
+static inline struct i915_ttm_buddy_resource *
+to_ttm_buddy_resource(struct ttm_resource *res)
+{
+	return container_of(res, struct i915_ttm_buddy_resource, base);
+}
+
+int i915_ttm_buddy_man_init(struct ttm_device *bdev,
+			    unsigned type, bool use_tt,
+			    u64 size, u64 visible_size,
+			    u64 default_page_size, u64 chunk_size);
+int i915_ttm_buddy_man_fini(struct ttm_device *bdev,
+			    unsigned int type);
+
+int i915_ttm_buddy_man_reserve(struct ttm_resource_manager *man,
+			       u64 start, u64 size);
+
+u64 i915_ttm_buddy_man_visible_size(struct ttm_resource_manager *man);
+
+void i915_ttm_buddy_man_avail(struct ttm_resource_manager *man,
+			      u64 *avail, u64 *avail_visible);
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+void i915_ttm_buddy_man_force_visible_size(struct ttm_resource_manager *man,
+					   u64 size);
+#endif
+
+#endif
diff --git a/drivers/gpu/drm/i915/i915_user_extensions.c b/drivers/gpu/drm/i915/i915_user_extensions.c
new file mode 100644
index 000000000..c822d0aaf
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_user_extensions.c
@@ -0,0 +1,61 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/nospec.h>
+#include <linux/sched/signal.h>
+#include <linux/uaccess.h>
+
+#include <uapi/drm/i915_drm.h>
+
+#include "i915_user_extensions.h"
+#include "i915_utils.h"
+
+int i915_user_extensions(struct i915_user_extension __user *ext,
+			 const i915_user_extension_fn *tbl,
+			 unsigned int count,
+			 void *data)
+{
+	unsigned int stackdepth = 512;
+
+	while (ext) {
+		int i, err;
+		u32 name;
+		u64 next;
+
+		if (!stackdepth--) /* recursion vs useful flexibility */
+			return -E2BIG;
+
+		err = check_user_mbz(&ext->flags);
+		if (err)
+			return err;
+
+		for (i = 0; i < ARRAY_SIZE(ext->rsvd); i++) {
+			err = check_user_mbz(&ext->rsvd[i]);
+			if (err)
+				return err;
+		}
+
+		if (get_user(name, &ext->name))
+			return -EFAULT;
+
+		err = -EINVAL;
+		if (name < count) {
+			name = array_index_nospec(name, count);
+			if (tbl[name])
+				err = tbl[name](ext, data);
+		}
+		if (err)
+			return err;
+
+		if (get_user(next, &ext->next_extension) ||
+		    overflows_type(next, ext))
+			return -EFAULT;
+
+		ext = u64_to_user_ptr(next);
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_user_extensions.h b/drivers/gpu/drm/i915/i915_user_extensions.h
new file mode 100644
index 000000000..a14bf6bba
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_user_extensions.h
@@ -0,0 +1,20 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef I915_USER_EXTENSIONS_H
+#define I915_USER_EXTENSIONS_H
+
+struct i915_user_extension;
+
+typedef int (*i915_user_extension_fn)(struct i915_user_extension __user *ext,
+				      void *data);
+
+int i915_user_extensions(struct i915_user_extension __user *ext,
+			 const i915_user_extension_fn *tbl,
+			 unsigned int count,
+			 void *data);
+
+#endif /* I915_USER_EXTENSIONS_H */
diff --git a/drivers/gpu/drm/i915/i915_utils.c b/drivers/gpu/drm/i915/i915_utils.c
new file mode 100644
index 000000000..29fd02bf5
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_utils.c
@@ -0,0 +1,127 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/device.h>
+
+#include <drm/drm_drv.h>
+
+#include "i915_drv.h"
+#include "i915_utils.h"
+
+#define FDO_BUG_MSG "Please file a bug on drm/i915; see " FDO_BUG_URL " for details."
+
+void
+__i915_printk(struct drm_i915_private *dev_priv, const char *level,
+	      const char *fmt, ...)
+{
+	static bool shown_bug_once;
+	struct device *kdev = dev_priv->drm.dev;
+	bool is_error = level[1] <= KERN_ERR[1];
+	bool is_debug = level[1] == KERN_DEBUG[1];
+	struct va_format vaf;
+	va_list args;
+
+	if (is_debug && !drm_debug_enabled(DRM_UT_DRIVER))
+		return;
+
+	va_start(args, fmt);
+
+	vaf.fmt = fmt;
+	vaf.va = &args;
+
+	if (is_error)
+		dev_printk(level, kdev, "%pV", &vaf);
+	else
+		dev_printk(level, kdev, "[" DRM_NAME ":%ps] %pV",
+			   __builtin_return_address(0), &vaf);
+
+	va_end(args);
+
+	if (is_error && !shown_bug_once) {
+		/*
+		 * Ask the user to file a bug report for the error, except
+		 * if they may have caused the bug by fiddling with unsafe
+		 * module parameters.
+		 */
+		if (!test_taint(TAINT_USER))
+			dev_notice(kdev, "%s", FDO_BUG_MSG);
+		shown_bug_once = true;
+	}
+}
+
+void add_taint_for_CI(struct drm_i915_private *i915, unsigned int taint)
+{
+	__i915_printk(i915, KERN_NOTICE, "CI tainted:%#x by %pS\n",
+		      taint, (void *)_RET_IP_);
+
+	/* Failures that occur during fault injection testing are expected */
+	if (!i915_error_injected())
+		__add_taint_for_CI(taint);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG)
+static unsigned int i915_probe_fail_count;
+
+int __i915_inject_probe_error(struct drm_i915_private *i915, int err,
+			      const char *func, int line)
+{
+	if (i915_probe_fail_count >= i915_modparams.inject_probe_failure)
+		return 0;
+
+	if (++i915_probe_fail_count < i915_modparams.inject_probe_failure)
+		return 0;
+
+	__i915_printk(i915, KERN_INFO,
+		      "Injecting failure %d at checkpoint %u [%s:%d]\n",
+		      err, i915_modparams.inject_probe_failure, func, line);
+	i915_modparams.inject_probe_failure = 0;
+	return err;
+}
+
+bool i915_error_injected(void)
+{
+	return i915_probe_fail_count && !i915_modparams.inject_probe_failure;
+}
+
+#endif
+
+void cancel_timer(struct timer_list *t)
+{
+	if (!timer_active(t))
+		return;
+
+	del_timer(t);
+	WRITE_ONCE(t->expires, 0);
+}
+
+void set_timer_ms(struct timer_list *t, unsigned long timeout)
+{
+	if (!timeout) {
+		cancel_timer(t);
+		return;
+	}
+
+	timeout = msecs_to_jiffies(timeout);
+
+	/*
+	 * Paranoia to make sure the compiler computes the timeout before
+	 * loading 'jiffies' as jiffies is volatile and may be updated in
+	 * the background by a timer tick. All to reduce the complexity
+	 * of the addition and reduce the risk of losing a jiffie.
+	 */
+	barrier();
+
+	/* Keep t->expires = 0 reserved to indicate a canceled timer. */
+	mod_timer(t, jiffies + timeout ?: 1);
+}
+
+bool i915_vtd_active(struct drm_i915_private *i915)
+{
+	if (device_iommu_mapped(i915->drm.dev))
+		return true;
+
+	/* Running as a guest, we assume the host is enforcing VT'd */
+	return i915_run_as_guest();
+}
diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h
new file mode 100644
index 000000000..6c14d1336
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_utils.h
@@ -0,0 +1,400 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __I915_UTILS_H
+#define __I915_UTILS_H
+
+#include <linux/list.h>
+#include <linux/overflow.h>
+#include <linux/sched.h>
+#include <linux/string_helpers.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+#include <linux/sched/clock.h>
+
+#ifdef CONFIG_X86
+#include <asm/hypervisor.h>
+#endif
+
+struct drm_i915_private;
+struct timer_list;
+
+#define FDO_BUG_URL "https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs"
+
+#define MISSING_CASE(x) WARN(1, "Missing case (%s == %ld)\n", \
+			     __stringify(x), (long)(x))
+
+void __printf(3, 4)
+__i915_printk(struct drm_i915_private *dev_priv, const char *level,
+	      const char *fmt, ...);
+
+#define i915_report_error(dev_priv, fmt, ...)				   \
+	__i915_printk(dev_priv, KERN_ERR, fmt, ##__VA_ARGS__)
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG)
+
+int __i915_inject_probe_error(struct drm_i915_private *i915, int err,
+			      const char *func, int line);
+#define i915_inject_probe_error(_i915, _err) \
+	__i915_inject_probe_error((_i915), (_err), __func__, __LINE__)
+bool i915_error_injected(void);
+
+#else
+
+#define i915_inject_probe_error(i915, e) ({ BUILD_BUG_ON_INVALID(i915); 0; })
+#define i915_error_injected() false
+
+#endif
+
+#define i915_inject_probe_failure(i915) i915_inject_probe_error((i915), -ENODEV)
+
+#define i915_probe_error(i915, fmt, ...)				   \
+	__i915_printk(i915, i915_error_injected() ? KERN_DEBUG : KERN_ERR, \
+		      fmt, ##__VA_ARGS__)
+
+#if defined(GCC_VERSION) && GCC_VERSION >= 70000
+#define add_overflows_t(T, A, B) \
+	__builtin_add_overflow_p((A), (B), (T)0)
+#else
+#define add_overflows_t(T, A, B) ({ \
+	typeof(A) a = (A); \
+	typeof(B) b = (B); \
+	(T)(a + b) < a; \
+})
+#endif
+
+#define add_overflows(A, B) \
+	add_overflows_t(typeof((A) + (B)), (A), (B))
+
+#define range_overflows(start, size, max) ({ \
+	typeof(start) start__ = (start); \
+	typeof(size) size__ = (size); \
+	typeof(max) max__ = (max); \
+	(void)(&start__ == &size__); \
+	(void)(&start__ == &max__); \
+	start__ >= max__ || size__ > max__ - start__; \
+})
+
+#define range_overflows_t(type, start, size, max) \
+	range_overflows((type)(start), (type)(size), (type)(max))
+
+#define range_overflows_end(start, size, max) ({ \
+	typeof(start) start__ = (start); \
+	typeof(size) size__ = (size); \
+	typeof(max) max__ = (max); \
+	(void)(&start__ == &size__); \
+	(void)(&start__ == &max__); \
+	start__ > max__ || size__ > max__ - start__; \
+})
+
+#define range_overflows_end_t(type, start, size, max) \
+	range_overflows_end((type)(start), (type)(size), (type)(max))
+
+/* Note we don't consider signbits :| */
+#define overflows_type(x, T) \
+	(sizeof(x) > sizeof(T) && (x) >> BITS_PER_TYPE(T))
+
+#define ptr_mask_bits(ptr, n) ({					\
+	unsigned long __v = (unsigned long)(ptr);			\
+	(typeof(ptr))(__v & -BIT(n));					\
+})
+
+#define ptr_unmask_bits(ptr, n) ((unsigned long)(ptr) & (BIT(n) - 1))
+
+#define ptr_unpack_bits(ptr, bits, n) ({				\
+	unsigned long __v = (unsigned long)(ptr);			\
+	*(bits) = __v & (BIT(n) - 1);					\
+	(typeof(ptr))(__v & -BIT(n));					\
+})
+
+#define ptr_pack_bits(ptr, bits, n) ({					\
+	unsigned long __bits = (bits);					\
+	GEM_BUG_ON(__bits & -BIT(n));					\
+	((typeof(ptr))((unsigned long)(ptr) | __bits));			\
+})
+
+#define ptr_dec(ptr) ({							\
+	unsigned long __v = (unsigned long)(ptr);			\
+	(typeof(ptr))(__v - 1);						\
+})
+
+#define ptr_inc(ptr) ({							\
+	unsigned long __v = (unsigned long)(ptr);			\
+	(typeof(ptr))(__v + 1);						\
+})
+
+#define page_mask_bits(ptr) ptr_mask_bits(ptr, PAGE_SHIFT)
+#define page_unmask_bits(ptr) ptr_unmask_bits(ptr, PAGE_SHIFT)
+#define page_pack_bits(ptr, bits) ptr_pack_bits(ptr, bits, PAGE_SHIFT)
+#define page_unpack_bits(ptr, bits) ptr_unpack_bits(ptr, bits, PAGE_SHIFT)
+
+#define struct_member(T, member) (((T *)0)->member)
+
+#define fetch_and_zero(ptr) ({						\
+	typeof(*ptr) __T = *(ptr);					\
+	*(ptr) = (typeof(*ptr))0;					\
+	__T;								\
+})
+
+static __always_inline ptrdiff_t ptrdiff(const void *a, const void *b)
+{
+	return a - b;
+}
+
+/*
+ * container_of_user: Extract the superclass from a pointer to a member.
+ *
+ * Exactly like container_of() with the exception that it plays nicely
+ * with sparse for __user @ptr.
+ */
+#define container_of_user(ptr, type, member) ({				\
+	void __user *__mptr = (void __user *)(ptr);			\
+	BUILD_BUG_ON_MSG(!__same_type(*(ptr), struct_member(type, member)) && \
+			 !__same_type(*(ptr), void),			\
+			 "pointer type mismatch in container_of()");	\
+	((type __user *)(__mptr - offsetof(type, member))); })
+
+/*
+ * check_user_mbz: Check that a user value exists and is zero
+ *
+ * Frequently in our uABI we reserve space for future extensions, and
+ * two ensure that userspace is prepared we enforce that space must
+ * be zero. (Then any future extension can safely assume a default value
+ * of 0.)
+ *
+ * check_user_mbz() combines checking that the user pointer is accessible
+ * and that the contained value is zero.
+ *
+ * Returns: -EFAULT if not accessible, -EINVAL if !zero, or 0 on success.
+ */
+#define check_user_mbz(U) ({						\
+	typeof(*(U)) mbz__;						\
+	get_user(mbz__, (U)) ? -EFAULT : mbz__ ? -EINVAL : 0;		\
+})
+
+#define u64_to_ptr(T, x) ({						\
+	typecheck(u64, x);						\
+	(T *)(uintptr_t)(x);						\
+})
+
+#define __mask_next_bit(mask) ({					\
+	int __idx = ffs(mask) - 1;					\
+	mask &= ~BIT(__idx);						\
+	__idx;								\
+})
+
+static inline bool is_power_of_2_u64(u64 n)
+{
+	return (n != 0 && ((n & (n - 1)) == 0));
+}
+
+static inline void __list_del_many(struct list_head *head,
+				   struct list_head *first)
+{
+	first->prev = head;
+	WRITE_ONCE(head->next, first);
+}
+
+static inline int list_is_last_rcu(const struct list_head *list,
+				   const struct list_head *head)
+{
+	return READ_ONCE(list->next) == head;
+}
+
+static inline unsigned long msecs_to_jiffies_timeout(const unsigned int m)
+{
+	unsigned long j = msecs_to_jiffies(m);
+
+	return min_t(unsigned long, MAX_JIFFY_OFFSET, j + 1);
+}
+
+/*
+ * If you need to wait X milliseconds between events A and B, but event B
+ * doesn't happen exactly after event A, you record the timestamp (jiffies) of
+ * when event A happened, then just before event B you call this function and
+ * pass the timestamp as the first argument, and X as the second argument.
+ */
+static inline void
+wait_remaining_ms_from_jiffies(unsigned long timestamp_jiffies, int to_wait_ms)
+{
+	unsigned long target_jiffies, tmp_jiffies, remaining_jiffies;
+
+	/*
+	 * Don't re-read the value of "jiffies" every time since it may change
+	 * behind our back and break the math.
+	 */
+	tmp_jiffies = jiffies;
+	target_jiffies = timestamp_jiffies +
+			 msecs_to_jiffies_timeout(to_wait_ms);
+
+	if (time_after(target_jiffies, tmp_jiffies)) {
+		remaining_jiffies = target_jiffies - tmp_jiffies;
+		while (remaining_jiffies)
+			remaining_jiffies =
+			    schedule_timeout_uninterruptible(remaining_jiffies);
+	}
+}
+
+/**
+ * __wait_for - magic wait macro
+ *
+ * Macro to help avoid open coding check/wait/timeout patterns. Note that it's
+ * important that we check the condition again after having timed out, since the
+ * timeout could be due to preemption or similar and we've never had a chance to
+ * check the condition before the timeout.
+ */
+#define __wait_for(OP, COND, US, Wmin, Wmax) ({ \
+	const ktime_t end__ = ktime_add_ns(ktime_get_raw(), 1000ll * (US)); \
+	long wait__ = (Wmin); /* recommended min for usleep is 10 us */	\
+	int ret__;							\
+	might_sleep();							\
+	for (;;) {							\
+		const bool expired__ = ktime_after(ktime_get_raw(), end__); \
+		OP;							\
+		/* Guarantee COND check prior to timeout */		\
+		barrier();						\
+		if (COND) {						\
+			ret__ = 0;					\
+			break;						\
+		}							\
+		if (expired__) {					\
+			ret__ = -ETIMEDOUT;				\
+			break;						\
+		}							\
+		usleep_range(wait__, wait__ * 2);			\
+		if (wait__ < (Wmax))					\
+			wait__ <<= 1;					\
+	}								\
+	ret__;								\
+})
+
+#define _wait_for(COND, US, Wmin, Wmax)	__wait_for(, (COND), (US), (Wmin), \
+						   (Wmax))
+#define wait_for(COND, MS)		_wait_for((COND), (MS) * 1000, 10, 1000)
+
+/* If CONFIG_PREEMPT_COUNT is disabled, in_atomic() always reports false. */
+#if defined(CONFIG_DRM_I915_DEBUG) && defined(CONFIG_PREEMPT_COUNT)
+# define _WAIT_FOR_ATOMIC_CHECK(ATOMIC) WARN_ON_ONCE((ATOMIC) && !in_atomic())
+#else
+# define _WAIT_FOR_ATOMIC_CHECK(ATOMIC) do { } while (0)
+#endif
+
+#define _wait_for_atomic(COND, US, ATOMIC) \
+({ \
+	int cpu, ret, timeout = (US) * 1000; \
+	u64 base; \
+	_WAIT_FOR_ATOMIC_CHECK(ATOMIC); \
+	if (!(ATOMIC)) { \
+		preempt_disable(); \
+		cpu = smp_processor_id(); \
+	} \
+	base = local_clock(); \
+	for (;;) { \
+		u64 now = local_clock(); \
+		if (!(ATOMIC)) \
+			preempt_enable(); \
+		/* Guarantee COND check prior to timeout */ \
+		barrier(); \
+		if (COND) { \
+			ret = 0; \
+			break; \
+		} \
+		if (now - base >= timeout) { \
+			ret = -ETIMEDOUT; \
+			break; \
+		} \
+		cpu_relax(); \
+		if (!(ATOMIC)) { \
+			preempt_disable(); \
+			if (unlikely(cpu != smp_processor_id())) { \
+				timeout -= now - base; \
+				cpu = smp_processor_id(); \
+				base = local_clock(); \
+			} \
+		} \
+	} \
+	ret; \
+})
+
+#define wait_for_us(COND, US) \
+({ \
+	int ret__; \
+	BUILD_BUG_ON(!__builtin_constant_p(US)); \
+	if ((US) > 10) \
+		ret__ = _wait_for((COND), (US), 10, 10); \
+	else \
+		ret__ = _wait_for_atomic((COND), (US), 0); \
+	ret__; \
+})
+
+#define wait_for_atomic_us(COND, US) \
+({ \
+	BUILD_BUG_ON(!__builtin_constant_p(US)); \
+	BUILD_BUG_ON((US) > 50000); \
+	_wait_for_atomic((COND), (US), 1); \
+})
+
+#define wait_for_atomic(COND, MS) wait_for_atomic_us((COND), (MS) * 1000)
+
+#define KHz(x) (1000 * (x))
+#define MHz(x) KHz(1000 * (x))
+
+void add_taint_for_CI(struct drm_i915_private *i915, unsigned int taint);
+static inline void __add_taint_for_CI(unsigned int taint)
+{
+	/*
+	 * The system is "ok", just about surviving for the user, but
+	 * CI results are now unreliable as the HW is very suspect.
+	 * CI checks the taint state after every test and will reboot
+	 * the machine if the kernel is tainted.
+	 */
+	add_taint(taint, LOCKDEP_STILL_OK);
+}
+
+void cancel_timer(struct timer_list *t);
+void set_timer_ms(struct timer_list *t, unsigned long timeout);
+
+static inline bool timer_active(const struct timer_list *t)
+{
+	return READ_ONCE(t->expires);
+}
+
+static inline bool timer_expired(const struct timer_list *t)
+{
+	return timer_active(t) && !timer_pending(t);
+}
+
+static inline bool i915_run_as_guest(void)
+{
+#if IS_ENABLED(CONFIG_X86)
+	return !hypervisor_is_type(X86_HYPER_NATIVE);
+#else
+	/* Not supported yet */
+	return false;
+#endif
+}
+
+bool i915_vtd_active(struct drm_i915_private *i915);
+
+#endif /* !__I915_UTILS_H */
diff --git a/drivers/gpu/drm/i915/i915_vgpu.c b/drivers/gpu/drm/i915/i915_vgpu.c
new file mode 100644
index 000000000..c97323973
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_vgpu.c
@@ -0,0 +1,338 @@
+/*
+ * Copyright(c) 2011-2015 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include "i915_drv.h"
+#include "i915_pvinfo.h"
+#include "i915_vgpu.h"
+
+/**
+ * DOC: Intel GVT-g guest support
+ *
+ * Intel GVT-g is a graphics virtualization technology which shares the
+ * GPU among multiple virtual machines on a time-sharing basis. Each
+ * virtual machine is presented a virtual GPU (vGPU), which has equivalent
+ * features as the underlying physical GPU (pGPU), so i915 driver can run
+ * seamlessly in a virtual machine. This file provides vGPU specific
+ * optimizations when running in a virtual machine, to reduce the complexity
+ * of vGPU emulation and to improve the overall performance.
+ *
+ * A primary function introduced here is so-called "address space ballooning"
+ * technique. Intel GVT-g partitions global graphics memory among multiple VMs,
+ * so each VM can directly access a portion of the memory without hypervisor's
+ * intervention, e.g. filling textures or queuing commands. However with the
+ * partitioning an unmodified i915 driver would assume a smaller graphics
+ * memory starting from address ZERO, then requires vGPU emulation module to
+ * translate the graphics address between 'guest view' and 'host view', for
+ * all registers and command opcodes which contain a graphics memory address.
+ * To reduce the complexity, Intel GVT-g introduces "address space ballooning",
+ * by telling the exact partitioning knowledge to each guest i915 driver, which
+ * then reserves and prevents non-allocated portions from allocation. Thus vGPU
+ * emulation module only needs to scan and validate graphics addresses without
+ * complexity of address translation.
+ *
+ */
+
+/**
+ * intel_vgpu_detect - detect virtual GPU
+ * @dev_priv: i915 device private
+ *
+ * This function is called at the initialization stage, to detect whether
+ * running on a vGPU.
+ */
+void intel_vgpu_detect(struct drm_i915_private *dev_priv)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	u64 magic;
+	u16 version_major;
+	void __iomem *shared_area;
+
+	BUILD_BUG_ON(sizeof(struct vgt_if) != VGT_PVINFO_SIZE);
+
+	/*
+	 * This is called before we setup the main MMIO BAR mappings used via
+	 * the uncore structure, so we need to access the BAR directly. Since
+	 * we do not support VGT on older gens, return early so we don't have
+	 * to consider differently numbered or sized MMIO bars
+	 */
+	if (GRAPHICS_VER(dev_priv) < 6)
+		return;
+
+	shared_area = pci_iomap_range(pdev, 0, VGT_PVINFO_PAGE, VGT_PVINFO_SIZE);
+	if (!shared_area) {
+		drm_err(&dev_priv->drm,
+			"failed to map MMIO bar to check for VGT\n");
+		return;
+	}
+
+	magic = readq(shared_area + vgtif_offset(magic));
+	if (magic != VGT_MAGIC)
+		goto out;
+
+	version_major = readw(shared_area + vgtif_offset(version_major));
+	if (version_major < VGT_VERSION_MAJOR) {
+		drm_info(&dev_priv->drm, "VGT interface version mismatch!\n");
+		goto out;
+	}
+
+	dev_priv->vgpu.caps = readl(shared_area + vgtif_offset(vgt_caps));
+
+	dev_priv->vgpu.active = true;
+	mutex_init(&dev_priv->vgpu.lock);
+	drm_info(&dev_priv->drm, "Virtual GPU for Intel GVT-g detected.\n");
+
+out:
+	pci_iounmap(pdev, shared_area);
+}
+
+void intel_vgpu_register(struct drm_i915_private *i915)
+{
+	/*
+	 * Notify a valid surface after modesetting, when running inside a VM.
+	 */
+	if (intel_vgpu_active(i915))
+		intel_uncore_write(&i915->uncore, vgtif_reg(display_ready),
+				   VGT_DRV_DISPLAY_READY);
+}
+
+bool intel_vgpu_active(struct drm_i915_private *dev_priv)
+{
+	return dev_priv->vgpu.active;
+}
+
+bool intel_vgpu_has_full_ppgtt(struct drm_i915_private *dev_priv)
+{
+	return dev_priv->vgpu.caps & VGT_CAPS_FULL_PPGTT;
+}
+
+bool intel_vgpu_has_hwsp_emulation(struct drm_i915_private *dev_priv)
+{
+	return dev_priv->vgpu.caps & VGT_CAPS_HWSP_EMULATION;
+}
+
+bool intel_vgpu_has_huge_gtt(struct drm_i915_private *dev_priv)
+{
+	return dev_priv->vgpu.caps & VGT_CAPS_HUGE_GTT;
+}
+
+struct _balloon_info_ {
+	/*
+	 * There are up to 2 regions per mappable/unmappable graphic
+	 * memory that might be ballooned. Here, index 0/1 is for mappable
+	 * graphic memory, 2/3 for unmappable graphic memory.
+	 */
+	struct drm_mm_node space[4];
+};
+
+static struct _balloon_info_ bl_info;
+
+static void vgt_deballoon_space(struct i915_ggtt *ggtt,
+				struct drm_mm_node *node)
+{
+	struct drm_i915_private *dev_priv = ggtt->vm.i915;
+	if (!drm_mm_node_allocated(node))
+		return;
+
+	drm_dbg(&dev_priv->drm,
+		"deballoon space: range [0x%llx - 0x%llx] %llu KiB.\n",
+		node->start,
+		node->start + node->size,
+		node->size / 1024);
+
+	ggtt->vm.reserved -= node->size;
+	drm_mm_remove_node(node);
+}
+
+/**
+ * intel_vgt_deballoon - deballoon reserved graphics address trunks
+ * @ggtt: the global GGTT from which we reserved earlier
+ *
+ * This function is called to deallocate the ballooned-out graphic memory, when
+ * driver is unloaded or when ballooning fails.
+ */
+void intel_vgt_deballoon(struct i915_ggtt *ggtt)
+{
+	struct drm_i915_private *dev_priv = ggtt->vm.i915;
+	int i;
+
+	if (!intel_vgpu_active(ggtt->vm.i915))
+		return;
+
+	drm_dbg(&dev_priv->drm, "VGT deballoon.\n");
+
+	for (i = 0; i < 4; i++)
+		vgt_deballoon_space(ggtt, &bl_info.space[i]);
+}
+
+static int vgt_balloon_space(struct i915_ggtt *ggtt,
+			     struct drm_mm_node *node,
+			     unsigned long start, unsigned long end)
+{
+	struct drm_i915_private *dev_priv = ggtt->vm.i915;
+	unsigned long size = end - start;
+	int ret;
+
+	if (start >= end)
+		return -EINVAL;
+
+	drm_info(&dev_priv->drm,
+		 "balloon space: range [ 0x%lx - 0x%lx ] %lu KiB.\n",
+		 start, end, size / 1024);
+	ret = i915_gem_gtt_reserve(&ggtt->vm, NULL, node,
+				   size, start, I915_COLOR_UNEVICTABLE,
+				   0);
+	if (!ret)
+		ggtt->vm.reserved += size;
+
+	return ret;
+}
+
+/**
+ * intel_vgt_balloon - balloon out reserved graphics address trunks
+ * @ggtt: the global GGTT from which to reserve
+ *
+ * This function is called at the initialization stage, to balloon out the
+ * graphic address space allocated to other vGPUs, by marking these spaces as
+ * reserved. The ballooning related knowledge(starting address and size of
+ * the mappable/unmappable graphic memory) is described in the vgt_if structure
+ * in a reserved mmio range.
+ *
+ * To give an example, the drawing below depicts one typical scenario after
+ * ballooning. Here the vGPU1 has 2 pieces of graphic address spaces ballooned
+ * out each for the mappable and the non-mappable part. From the vGPU1 point of
+ * view, the total size is the same as the physical one, with the start address
+ * of its graphic space being zero. Yet there are some portions ballooned out(
+ * the shadow part, which are marked as reserved by drm allocator). From the
+ * host point of view, the graphic address space is partitioned by multiple
+ * vGPUs in different VMs. ::
+ *
+ *                         vGPU1 view         Host view
+ *              0 ------> +-----------+     +-----------+
+ *                ^       |###########|     |   vGPU3   |
+ *                |       |###########|     +-----------+
+ *                |       |###########|     |   vGPU2   |
+ *                |       +-----------+     +-----------+
+ *         mappable GM    | available | ==> |   vGPU1   |
+ *                |       +-----------+     +-----------+
+ *                |       |###########|     |           |
+ *                v       |###########|     |   Host    |
+ *                +=======+===========+     +===========+
+ *                ^       |###########|     |   vGPU3   |
+ *                |       |###########|     +-----------+
+ *                |       |###########|     |   vGPU2   |
+ *                |       +-----------+     +-----------+
+ *       unmappable GM    | available | ==> |   vGPU1   |
+ *                |       +-----------+     +-----------+
+ *                |       |###########|     |           |
+ *                |       |###########|     |   Host    |
+ *                v       |###########|     |           |
+ *  total GM size ------> +-----------+     +-----------+
+ *
+ * Returns:
+ * zero on success, non-zero if configuration invalid or ballooning failed
+ */
+int intel_vgt_balloon(struct i915_ggtt *ggtt)
+{
+	struct drm_i915_private *dev_priv = ggtt->vm.i915;
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	unsigned long ggtt_end = ggtt->vm.total;
+
+	unsigned long mappable_base, mappable_size, mappable_end;
+	unsigned long unmappable_base, unmappable_size, unmappable_end;
+	int ret;
+
+	if (!intel_vgpu_active(ggtt->vm.i915))
+		return 0;
+
+	mappable_base =
+	  intel_uncore_read(uncore, vgtif_reg(avail_rs.mappable_gmadr.base));
+	mappable_size =
+	  intel_uncore_read(uncore, vgtif_reg(avail_rs.mappable_gmadr.size));
+	unmappable_base =
+	  intel_uncore_read(uncore, vgtif_reg(avail_rs.nonmappable_gmadr.base));
+	unmappable_size =
+	  intel_uncore_read(uncore, vgtif_reg(avail_rs.nonmappable_gmadr.size));
+
+	mappable_end = mappable_base + mappable_size;
+	unmappable_end = unmappable_base + unmappable_size;
+
+	drm_info(&dev_priv->drm, "VGT ballooning configuration:\n");
+	drm_info(&dev_priv->drm,
+		 "Mappable graphic memory: base 0x%lx size %ldKiB\n",
+		 mappable_base, mappable_size / 1024);
+	drm_info(&dev_priv->drm,
+		 "Unmappable graphic memory: base 0x%lx size %ldKiB\n",
+		 unmappable_base, unmappable_size / 1024);
+
+	if (mappable_end > ggtt->mappable_end ||
+	    unmappable_base < ggtt->mappable_end ||
+	    unmappable_end > ggtt_end) {
+		drm_err(&dev_priv->drm, "Invalid ballooning configuration!\n");
+		return -EINVAL;
+	}
+
+	/* Unmappable graphic memory ballooning */
+	if (unmappable_base > ggtt->mappable_end) {
+		ret = vgt_balloon_space(ggtt, &bl_info.space[2],
+					ggtt->mappable_end, unmappable_base);
+
+		if (ret)
+			goto err;
+	}
+
+	if (unmappable_end < ggtt_end) {
+		ret = vgt_balloon_space(ggtt, &bl_info.space[3],
+					unmappable_end, ggtt_end);
+		if (ret)
+			goto err_upon_mappable;
+	}
+
+	/* Mappable graphic memory ballooning */
+	if (mappable_base) {
+		ret = vgt_balloon_space(ggtt, &bl_info.space[0],
+					0, mappable_base);
+
+		if (ret)
+			goto err_upon_unmappable;
+	}
+
+	if (mappable_end < ggtt->mappable_end) {
+		ret = vgt_balloon_space(ggtt, &bl_info.space[1],
+					mappable_end, ggtt->mappable_end);
+
+		if (ret)
+			goto err_below_mappable;
+	}
+
+	drm_info(&dev_priv->drm, "VGT balloon successfully\n");
+	return 0;
+
+err_below_mappable:
+	vgt_deballoon_space(ggtt, &bl_info.space[0]);
+err_upon_unmappable:
+	vgt_deballoon_space(ggtt, &bl_info.space[3]);
+err_upon_mappable:
+	vgt_deballoon_space(ggtt, &bl_info.space[2]);
+err:
+	drm_err(&dev_priv->drm, "VGT balloon fail\n");
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/i915_vgpu.h b/drivers/gpu/drm/i915/i915_vgpu.h
new file mode 100644
index 000000000..ffbb77d08
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_vgpu.h
@@ -0,0 +1,42 @@
+/*
+ * Copyright(c) 2011-2015 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef _I915_VGPU_H_
+#define _I915_VGPU_H_
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct i915_ggtt;
+
+void intel_vgpu_detect(struct drm_i915_private *i915);
+bool intel_vgpu_active(struct drm_i915_private *i915);
+void intel_vgpu_register(struct drm_i915_private *i915);
+bool intel_vgpu_has_full_ppgtt(struct drm_i915_private *i915);
+bool intel_vgpu_has_hwsp_emulation(struct drm_i915_private *i915);
+bool intel_vgpu_has_huge_gtt(struct drm_i915_private *i915);
+
+int intel_vgt_balloon(struct i915_ggtt *ggtt);
+void intel_vgt_deballoon(struct i915_ggtt *ggtt);
+
+#endif /* _I915_VGPU_H_ */
diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c
new file mode 100644
index 000000000..c8ad8f37e
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_vma.c
@@ -0,0 +1,2197 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/sched/mm.h>
+#include <linux/dma-fence-array.h>
+#include <drm/drm_gem.h>
+
+#include "display/intel_frontbuffer.h"
+#include "gem/i915_gem_lmem.h"
+#include "gem/i915_gem_tiling.h"
+#include "gt/intel_engine.h"
+#include "gt/intel_engine_heartbeat.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_requests.h"
+
+#include "i915_drv.h"
+#include "i915_gem_evict.h"
+#include "i915_sw_fence_work.h"
+#include "i915_trace.h"
+#include "i915_vma.h"
+#include "i915_vma_resource.h"
+
+static inline void assert_vma_held_evict(const struct i915_vma *vma)
+{
+	/*
+	 * We may be forced to unbind when the vm is dead, to clean it up.
+	 * This is the only exception to the requirement of the object lock
+	 * being held.
+	 */
+	if (kref_read(&vma->vm->ref))
+		assert_object_held_shared(vma->obj);
+}
+
+static struct kmem_cache *slab_vmas;
+
+static struct i915_vma *i915_vma_alloc(void)
+{
+	return kmem_cache_zalloc(slab_vmas, GFP_KERNEL);
+}
+
+static void i915_vma_free(struct i915_vma *vma)
+{
+	return kmem_cache_free(slab_vmas, vma);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_ERRLOG_GEM) && IS_ENABLED(CONFIG_DRM_DEBUG_MM)
+
+#include <linux/stackdepot.h>
+
+static void vma_print_allocator(struct i915_vma *vma, const char *reason)
+{
+	char buf[512];
+
+	if (!vma->node.stack) {
+		DRM_DEBUG_DRIVER("vma.node [%08llx + %08llx] %s: unknown owner\n",
+				 vma->node.start, vma->node.size, reason);
+		return;
+	}
+
+	stack_depot_snprint(vma->node.stack, buf, sizeof(buf), 0);
+	DRM_DEBUG_DRIVER("vma.node [%08llx + %08llx] %s: inserted at %s\n",
+			 vma->node.start, vma->node.size, reason, buf);
+}
+
+#else
+
+static void vma_print_allocator(struct i915_vma *vma, const char *reason)
+{
+}
+
+#endif
+
+static inline struct i915_vma *active_to_vma(struct i915_active *ref)
+{
+	return container_of(ref, typeof(struct i915_vma), active);
+}
+
+static int __i915_vma_active(struct i915_active *ref)
+{
+	return i915_vma_tryget(active_to_vma(ref)) ? 0 : -ENOENT;
+}
+
+static void __i915_vma_retire(struct i915_active *ref)
+{
+	i915_vma_put(active_to_vma(ref));
+}
+
+static struct i915_vma *
+vma_create(struct drm_i915_gem_object *obj,
+	   struct i915_address_space *vm,
+	   const struct i915_gtt_view *view)
+{
+	struct i915_vma *pos = ERR_PTR(-E2BIG);
+	struct i915_vma *vma;
+	struct rb_node *rb, **p;
+	int err;
+
+	/* The aliasing_ppgtt should never be used directly! */
+	GEM_BUG_ON(vm == &vm->gt->ggtt->alias->vm);
+
+	vma = i915_vma_alloc();
+	if (vma == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	vma->ops = &vm->vma_ops;
+	vma->obj = obj;
+	vma->size = obj->base.size;
+	vma->display_alignment = I915_GTT_MIN_ALIGNMENT;
+
+	i915_active_init(&vma->active, __i915_vma_active, __i915_vma_retire, 0);
+
+	/* Declare ourselves safe for use inside shrinkers */
+	if (IS_ENABLED(CONFIG_LOCKDEP)) {
+		fs_reclaim_acquire(GFP_KERNEL);
+		might_lock(&vma->active.mutex);
+		fs_reclaim_release(GFP_KERNEL);
+	}
+
+	INIT_LIST_HEAD(&vma->closed_link);
+	INIT_LIST_HEAD(&vma->obj_link);
+	RB_CLEAR_NODE(&vma->obj_node);
+
+	if (view && view->type != I915_GTT_VIEW_NORMAL) {
+		vma->gtt_view = *view;
+		if (view->type == I915_GTT_VIEW_PARTIAL) {
+			GEM_BUG_ON(range_overflows_t(u64,
+						     view->partial.offset,
+						     view->partial.size,
+						     obj->base.size >> PAGE_SHIFT));
+			vma->size = view->partial.size;
+			vma->size <<= PAGE_SHIFT;
+			GEM_BUG_ON(vma->size > obj->base.size);
+		} else if (view->type == I915_GTT_VIEW_ROTATED) {
+			vma->size = intel_rotation_info_size(&view->rotated);
+			vma->size <<= PAGE_SHIFT;
+		} else if (view->type == I915_GTT_VIEW_REMAPPED) {
+			vma->size = intel_remapped_info_size(&view->remapped);
+			vma->size <<= PAGE_SHIFT;
+		}
+	}
+
+	if (unlikely(vma->size > vm->total))
+		goto err_vma;
+
+	GEM_BUG_ON(!IS_ALIGNED(vma->size, I915_GTT_PAGE_SIZE));
+
+	err = mutex_lock_interruptible(&vm->mutex);
+	if (err) {
+		pos = ERR_PTR(err);
+		goto err_vma;
+	}
+
+	vma->vm = vm;
+	list_add_tail(&vma->vm_link, &vm->unbound_list);
+
+	spin_lock(&obj->vma.lock);
+	if (i915_is_ggtt(vm)) {
+		if (unlikely(overflows_type(vma->size, u32)))
+			goto err_unlock;
+
+		vma->fence_size = i915_gem_fence_size(vm->i915, vma->size,
+						      i915_gem_object_get_tiling(obj),
+						      i915_gem_object_get_stride(obj));
+		if (unlikely(vma->fence_size < vma->size || /* overflow */
+			     vma->fence_size > vm->total))
+			goto err_unlock;
+
+		GEM_BUG_ON(!IS_ALIGNED(vma->fence_size, I915_GTT_MIN_ALIGNMENT));
+
+		vma->fence_alignment = i915_gem_fence_alignment(vm->i915, vma->size,
+								i915_gem_object_get_tiling(obj),
+								i915_gem_object_get_stride(obj));
+		GEM_BUG_ON(!is_power_of_2(vma->fence_alignment));
+
+		__set_bit(I915_VMA_GGTT_BIT, __i915_vma_flags(vma));
+	}
+
+	rb = NULL;
+	p = &obj->vma.tree.rb_node;
+	while (*p) {
+		long cmp;
+
+		rb = *p;
+		pos = rb_entry(rb, struct i915_vma, obj_node);
+
+		/*
+		 * If the view already exists in the tree, another thread
+		 * already created a matching vma, so return the older instance
+		 * and dispose of ours.
+		 */
+		cmp = i915_vma_compare(pos, vm, view);
+		if (cmp < 0)
+			p = &rb->rb_right;
+		else if (cmp > 0)
+			p = &rb->rb_left;
+		else
+			goto err_unlock;
+	}
+	rb_link_node(&vma->obj_node, rb, p);
+	rb_insert_color(&vma->obj_node, &obj->vma.tree);
+
+	if (i915_vma_is_ggtt(vma))
+		/*
+		 * We put the GGTT vma at the start of the vma-list, followed
+		 * by the ppGGTT vma. This allows us to break early when
+		 * iterating over only the GGTT vma for an object, see
+		 * for_each_ggtt_vma()
+		 */
+		list_add(&vma->obj_link, &obj->vma.list);
+	else
+		list_add_tail(&vma->obj_link, &obj->vma.list);
+
+	spin_unlock(&obj->vma.lock);
+	mutex_unlock(&vm->mutex);
+
+	return vma;
+
+err_unlock:
+	spin_unlock(&obj->vma.lock);
+	list_del_init(&vma->vm_link);
+	mutex_unlock(&vm->mutex);
+err_vma:
+	i915_vma_free(vma);
+	return pos;
+}
+
+static struct i915_vma *
+i915_vma_lookup(struct drm_i915_gem_object *obj,
+	   struct i915_address_space *vm,
+	   const struct i915_gtt_view *view)
+{
+	struct rb_node *rb;
+
+	rb = obj->vma.tree.rb_node;
+	while (rb) {
+		struct i915_vma *vma = rb_entry(rb, struct i915_vma, obj_node);
+		long cmp;
+
+		cmp = i915_vma_compare(vma, vm, view);
+		if (cmp == 0)
+			return vma;
+
+		if (cmp < 0)
+			rb = rb->rb_right;
+		else
+			rb = rb->rb_left;
+	}
+
+	return NULL;
+}
+
+/**
+ * i915_vma_instance - return the singleton instance of the VMA
+ * @obj: parent &struct drm_i915_gem_object to be mapped
+ * @vm: address space in which the mapping is located
+ * @view: additional mapping requirements
+ *
+ * i915_vma_instance() looks up an existing VMA of the @obj in the @vm with
+ * the same @view characteristics. If a match is not found, one is created.
+ * Once created, the VMA is kept until either the object is freed, or the
+ * address space is closed.
+ *
+ * Returns the vma, or an error pointer.
+ */
+struct i915_vma *
+i915_vma_instance(struct drm_i915_gem_object *obj,
+		  struct i915_address_space *vm,
+		  const struct i915_gtt_view *view)
+{
+	struct i915_vma *vma;
+
+	GEM_BUG_ON(view && !i915_is_ggtt_or_dpt(vm));
+	GEM_BUG_ON(!kref_read(&vm->ref));
+
+	spin_lock(&obj->vma.lock);
+	vma = i915_vma_lookup(obj, vm, view);
+	spin_unlock(&obj->vma.lock);
+
+	/* vma_create() will resolve the race if another creates the vma */
+	if (unlikely(!vma))
+		vma = vma_create(obj, vm, view);
+
+	GEM_BUG_ON(!IS_ERR(vma) && i915_vma_compare(vma, vm, view));
+	return vma;
+}
+
+struct i915_vma_work {
+	struct dma_fence_work base;
+	struct i915_address_space *vm;
+	struct i915_vm_pt_stash stash;
+	struct i915_vma_resource *vma_res;
+	struct drm_i915_gem_object *obj;
+	struct i915_sw_dma_fence_cb cb;
+	enum i915_cache_level cache_level;
+	unsigned int flags;
+};
+
+static void __vma_bind(struct dma_fence_work *work)
+{
+	struct i915_vma_work *vw = container_of(work, typeof(*vw), base);
+	struct i915_vma_resource *vma_res = vw->vma_res;
+
+	/*
+	 * We are about the bind the object, which must mean we have already
+	 * signaled the work to potentially clear/move the pages underneath. If
+	 * something went wrong at that stage then the object should have
+	 * unknown_state set, in which case we need to skip the bind.
+	 */
+	if (i915_gem_object_has_unknown_state(vw->obj))
+		return;
+
+	vma_res->ops->bind_vma(vma_res->vm, &vw->stash,
+			       vma_res, vw->cache_level, vw->flags);
+}
+
+static void __vma_release(struct dma_fence_work *work)
+{
+	struct i915_vma_work *vw = container_of(work, typeof(*vw), base);
+
+	if (vw->obj)
+		i915_gem_object_put(vw->obj);
+
+	i915_vm_free_pt_stash(vw->vm, &vw->stash);
+	if (vw->vma_res)
+		i915_vma_resource_put(vw->vma_res);
+}
+
+static const struct dma_fence_work_ops bind_ops = {
+	.name = "bind",
+	.work = __vma_bind,
+	.release = __vma_release,
+};
+
+struct i915_vma_work *i915_vma_work(void)
+{
+	struct i915_vma_work *vw;
+
+	vw = kzalloc(sizeof(*vw), GFP_KERNEL);
+	if (!vw)
+		return NULL;
+
+	dma_fence_work_init(&vw->base, &bind_ops);
+	vw->base.dma.error = -EAGAIN; /* disable the worker by default */
+
+	return vw;
+}
+
+int i915_vma_wait_for_bind(struct i915_vma *vma)
+{
+	int err = 0;
+
+	if (rcu_access_pointer(vma->active.excl.fence)) {
+		struct dma_fence *fence;
+
+		rcu_read_lock();
+		fence = dma_fence_get_rcu_safe(&vma->active.excl.fence);
+		rcu_read_unlock();
+		if (fence) {
+			err = dma_fence_wait(fence, true);
+			dma_fence_put(fence);
+		}
+	}
+
+	return err;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)
+static int i915_vma_verify_bind_complete(struct i915_vma *vma)
+{
+	struct dma_fence *fence = i915_active_fence_get(&vma->active.excl);
+	int err;
+
+	if (!fence)
+		return 0;
+
+	if (dma_fence_is_signaled(fence))
+		err = fence->error;
+	else
+		err = -EBUSY;
+
+	dma_fence_put(fence);
+
+	return err;
+}
+#else
+#define i915_vma_verify_bind_complete(_vma) 0
+#endif
+
+I915_SELFTEST_EXPORT void
+i915_vma_resource_init_from_vma(struct i915_vma_resource *vma_res,
+				struct i915_vma *vma)
+{
+	struct drm_i915_gem_object *obj = vma->obj;
+
+	i915_vma_resource_init(vma_res, vma->vm, vma->pages, &vma->page_sizes,
+			       obj->mm.rsgt, i915_gem_object_is_readonly(obj),
+			       i915_gem_object_is_lmem(obj), obj->mm.region,
+			       vma->ops, vma->private, vma->node.start,
+			       vma->node.size, vma->size);
+}
+
+/**
+ * i915_vma_bind - Sets up PTEs for an VMA in it's corresponding address space.
+ * @vma: VMA to map
+ * @cache_level: mapping cache level
+ * @flags: flags like global or local mapping
+ * @work: preallocated worker for allocating and binding the PTE
+ * @vma_res: pointer to a preallocated vma resource. The resource is either
+ * consumed or freed.
+ *
+ * DMA addresses are taken from the scatter-gather table of this object (or of
+ * this VMA in case of non-default GGTT views) and PTE entries set up.
+ * Note that DMA addresses are also the only part of the SG table we care about.
+ */
+int i915_vma_bind(struct i915_vma *vma,
+		  enum i915_cache_level cache_level,
+		  u32 flags,
+		  struct i915_vma_work *work,
+		  struct i915_vma_resource *vma_res)
+{
+	u32 bind_flags;
+	u32 vma_flags;
+	int ret;
+
+	lockdep_assert_held(&vma->vm->mutex);
+	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	GEM_BUG_ON(vma->size > vma->node.size);
+
+	if (GEM_DEBUG_WARN_ON(range_overflows(vma->node.start,
+					      vma->node.size,
+					      vma->vm->total))) {
+		i915_vma_resource_free(vma_res);
+		return -ENODEV;
+	}
+
+	if (GEM_DEBUG_WARN_ON(!flags)) {
+		i915_vma_resource_free(vma_res);
+		return -EINVAL;
+	}
+
+	bind_flags = flags;
+	bind_flags &= I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND;
+
+	vma_flags = atomic_read(&vma->flags);
+	vma_flags &= I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND;
+
+	bind_flags &= ~vma_flags;
+	if (bind_flags == 0) {
+		i915_vma_resource_free(vma_res);
+		return 0;
+	}
+
+	GEM_BUG_ON(!atomic_read(&vma->pages_count));
+
+	/* Wait for or await async unbinds touching our range */
+	if (work && bind_flags & vma->vm->bind_async_flags)
+		ret = i915_vma_resource_bind_dep_await(vma->vm,
+						       &work->base.chain,
+						       vma->node.start,
+						       vma->node.size,
+						       true,
+						       GFP_NOWAIT |
+						       __GFP_RETRY_MAYFAIL |
+						       __GFP_NOWARN);
+	else
+		ret = i915_vma_resource_bind_dep_sync(vma->vm, vma->node.start,
+						      vma->node.size, true);
+	if (ret) {
+		i915_vma_resource_free(vma_res);
+		return ret;
+	}
+
+	if (vma->resource || !vma_res) {
+		/* Rebinding with an additional I915_VMA_*_BIND */
+		GEM_WARN_ON(!vma_flags);
+		i915_vma_resource_free(vma_res);
+	} else {
+		i915_vma_resource_init_from_vma(vma_res, vma);
+		vma->resource = vma_res;
+	}
+	trace_i915_vma_bind(vma, bind_flags);
+	if (work && bind_flags & vma->vm->bind_async_flags) {
+		struct dma_fence *prev;
+
+		work->vma_res = i915_vma_resource_get(vma->resource);
+		work->cache_level = cache_level;
+		work->flags = bind_flags;
+
+		/*
+		 * Note we only want to chain up to the migration fence on
+		 * the pages (not the object itself). As we don't track that,
+		 * yet, we have to use the exclusive fence instead.
+		 *
+		 * Also note that we do not want to track the async vma as
+		 * part of the obj->resv->excl_fence as it only affects
+		 * execution and not content or object's backing store lifetime.
+		 */
+		prev = i915_active_set_exclusive(&vma->active, &work->base.dma);
+		if (prev) {
+			__i915_sw_fence_await_dma_fence(&work->base.chain,
+							prev,
+							&work->cb);
+			dma_fence_put(prev);
+		}
+
+		work->base.dma.error = 0; /* enable the queue_work() */
+		work->obj = i915_gem_object_get(vma->obj);
+	} else {
+		ret = i915_gem_object_wait_moving_fence(vma->obj, true);
+		if (ret) {
+			i915_vma_resource_free(vma->resource);
+			vma->resource = NULL;
+
+			return ret;
+		}
+		vma->ops->bind_vma(vma->vm, NULL, vma->resource, cache_level,
+				   bind_flags);
+	}
+
+	atomic_or(bind_flags, &vma->flags);
+	return 0;
+}
+
+void __iomem *i915_vma_pin_iomap(struct i915_vma *vma)
+{
+	void __iomem *ptr;
+	int err;
+
+	if (WARN_ON_ONCE(vma->obj->flags & I915_BO_ALLOC_GPU_ONLY))
+		return IOMEM_ERR_PTR(-EINVAL);
+
+	GEM_BUG_ON(!i915_vma_is_ggtt(vma));
+	GEM_BUG_ON(!i915_vma_is_bound(vma, I915_VMA_GLOBAL_BIND));
+	GEM_BUG_ON(i915_vma_verify_bind_complete(vma));
+
+	ptr = READ_ONCE(vma->iomap);
+	if (ptr == NULL) {
+		/*
+		 * TODO: consider just using i915_gem_object_pin_map() for lmem
+		 * instead, which already supports mapping non-contiguous chunks
+		 * of pages, that way we can also drop the
+		 * I915_BO_ALLOC_CONTIGUOUS when allocating the object.
+		 */
+		if (i915_gem_object_is_lmem(vma->obj)) {
+			ptr = i915_gem_object_lmem_io_map(vma->obj, 0,
+							  vma->obj->base.size);
+		} else if (i915_vma_is_map_and_fenceable(vma)) {
+			ptr = io_mapping_map_wc(&i915_vm_to_ggtt(vma->vm)->iomap,
+						vma->node.start,
+						vma->node.size);
+		} else {
+			ptr = (void __iomem *)
+				i915_gem_object_pin_map(vma->obj, I915_MAP_WC);
+			if (IS_ERR(ptr)) {
+				err = PTR_ERR(ptr);
+				goto err;
+			}
+			ptr = page_pack_bits(ptr, 1);
+		}
+
+		if (ptr == NULL) {
+			err = -ENOMEM;
+			goto err;
+		}
+
+		if (unlikely(cmpxchg(&vma->iomap, NULL, ptr))) {
+			if (page_unmask_bits(ptr))
+				__i915_gem_object_release_map(vma->obj);
+			else
+				io_mapping_unmap(ptr);
+			ptr = vma->iomap;
+		}
+	}
+
+	__i915_vma_pin(vma);
+
+	err = i915_vma_pin_fence(vma);
+	if (err)
+		goto err_unpin;
+
+	i915_vma_set_ggtt_write(vma);
+
+	/* NB Access through the GTT requires the device to be awake. */
+	return page_mask_bits(ptr);
+
+err_unpin:
+	__i915_vma_unpin(vma);
+err:
+	return IOMEM_ERR_PTR(err);
+}
+
+void i915_vma_flush_writes(struct i915_vma *vma)
+{
+	if (i915_vma_unset_ggtt_write(vma))
+		intel_gt_flush_ggtt_writes(vma->vm->gt);
+}
+
+void i915_vma_unpin_iomap(struct i915_vma *vma)
+{
+	GEM_BUG_ON(vma->iomap == NULL);
+
+	/* XXX We keep the mapping until __i915_vma_unbind()/evict() */
+
+	i915_vma_flush_writes(vma);
+
+	i915_vma_unpin_fence(vma);
+	i915_vma_unpin(vma);
+}
+
+void i915_vma_unpin_and_release(struct i915_vma **p_vma, unsigned int flags)
+{
+	struct i915_vma *vma;
+	struct drm_i915_gem_object *obj;
+
+	vma = fetch_and_zero(p_vma);
+	if (!vma)
+		return;
+
+	obj = vma->obj;
+	GEM_BUG_ON(!obj);
+
+	i915_vma_unpin(vma);
+
+	if (flags & I915_VMA_RELEASE_MAP)
+		i915_gem_object_unpin_map(obj);
+
+	i915_gem_object_put(obj);
+}
+
+bool i915_vma_misplaced(const struct i915_vma *vma,
+			u64 size, u64 alignment, u64 flags)
+{
+	if (!drm_mm_node_allocated(&vma->node))
+		return false;
+
+	if (test_bit(I915_VMA_ERROR_BIT, __i915_vma_flags(vma)))
+		return true;
+
+	if (vma->node.size < size)
+		return true;
+
+	GEM_BUG_ON(alignment && !is_power_of_2(alignment));
+	if (alignment && !IS_ALIGNED(vma->node.start, alignment))
+		return true;
+
+	if (flags & PIN_MAPPABLE && !i915_vma_is_map_and_fenceable(vma))
+		return true;
+
+	if (flags & PIN_OFFSET_BIAS &&
+	    vma->node.start < (flags & PIN_OFFSET_MASK))
+		return true;
+
+	if (flags & PIN_OFFSET_FIXED &&
+	    vma->node.start != (flags & PIN_OFFSET_MASK))
+		return true;
+
+	return false;
+}
+
+void __i915_vma_set_map_and_fenceable(struct i915_vma *vma)
+{
+	bool mappable, fenceable;
+
+	GEM_BUG_ON(!i915_vma_is_ggtt(vma));
+	GEM_BUG_ON(!vma->fence_size);
+
+	fenceable = (vma->node.size >= vma->fence_size &&
+		     IS_ALIGNED(vma->node.start, vma->fence_alignment));
+
+	mappable = vma->node.start + vma->fence_size <= i915_vm_to_ggtt(vma->vm)->mappable_end;
+
+	if (mappable && fenceable)
+		set_bit(I915_VMA_CAN_FENCE_BIT, __i915_vma_flags(vma));
+	else
+		clear_bit(I915_VMA_CAN_FENCE_BIT, __i915_vma_flags(vma));
+}
+
+bool i915_gem_valid_gtt_space(struct i915_vma *vma, unsigned long color)
+{
+	struct drm_mm_node *node = &vma->node;
+	struct drm_mm_node *other;
+
+	/*
+	 * On some machines we have to be careful when putting differing types
+	 * of snoopable memory together to avoid the prefetcher crossing memory
+	 * domains and dying. During vm initialisation, we decide whether or not
+	 * these constraints apply and set the drm_mm.color_adjust
+	 * appropriately.
+	 */
+	if (!i915_vm_has_cache_coloring(vma->vm))
+		return true;
+
+	/* Only valid to be called on an already inserted vma */
+	GEM_BUG_ON(!drm_mm_node_allocated(node));
+	GEM_BUG_ON(list_empty(&node->node_list));
+
+	other = list_prev_entry(node, node_list);
+	if (i915_node_color_differs(other, color) &&
+	    !drm_mm_hole_follows(other))
+		return false;
+
+	other = list_next_entry(node, node_list);
+	if (i915_node_color_differs(other, color) &&
+	    !drm_mm_hole_follows(node))
+		return false;
+
+	return true;
+}
+
+/**
+ * i915_vma_insert - finds a slot for the vma in its address space
+ * @vma: the vma
+ * @size: requested size in bytes (can be larger than the VMA)
+ * @alignment: required alignment
+ * @flags: mask of PIN_* flags to use
+ *
+ * First we try to allocate some free space that meets the requirements for
+ * the VMA. Failiing that, if the flags permit, it will evict an old VMA,
+ * preferrably the oldest idle entry to make room for the new VMA.
+ *
+ * Returns:
+ * 0 on success, negative error code otherwise.
+ */
+static int
+i915_vma_insert(struct i915_vma *vma, struct i915_gem_ww_ctx *ww,
+		u64 size, u64 alignment, u64 flags)
+{
+	unsigned long color;
+	u64 start, end;
+	int ret;
+
+	GEM_BUG_ON(i915_vma_is_bound(vma, I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND));
+	GEM_BUG_ON(drm_mm_node_allocated(&vma->node));
+
+	size = max(size, vma->size);
+	alignment = max(alignment, vma->display_alignment);
+	if (flags & PIN_MAPPABLE) {
+		size = max_t(typeof(size), size, vma->fence_size);
+		alignment = max_t(typeof(alignment),
+				  alignment, vma->fence_alignment);
+	}
+
+	GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE));
+	GEM_BUG_ON(!IS_ALIGNED(alignment, I915_GTT_MIN_ALIGNMENT));
+	GEM_BUG_ON(!is_power_of_2(alignment));
+
+	start = flags & PIN_OFFSET_BIAS ? flags & PIN_OFFSET_MASK : 0;
+	GEM_BUG_ON(!IS_ALIGNED(start, I915_GTT_PAGE_SIZE));
+
+	end = vma->vm->total;
+	if (flags & PIN_MAPPABLE)
+		end = min_t(u64, end, i915_vm_to_ggtt(vma->vm)->mappable_end);
+	if (flags & PIN_ZONE_4G)
+		end = min_t(u64, end, (1ULL << 32) - I915_GTT_PAGE_SIZE);
+	GEM_BUG_ON(!IS_ALIGNED(end, I915_GTT_PAGE_SIZE));
+
+	alignment = max(alignment, i915_vm_obj_min_alignment(vma->vm, vma->obj));
+	/*
+	 * for compact-pt we round up the reservation to prevent
+	 * any smaller pages being used within the same PDE
+	 */
+	if (NEEDS_COMPACT_PT(vma->vm->i915))
+		size = round_up(size, alignment);
+
+	/* If binding the object/GGTT view requires more space than the entire
+	 * aperture has, reject it early before evicting everything in a vain
+	 * attempt to find space.
+	 */
+	if (size > end) {
+		DRM_DEBUG("Attempting to bind an object larger than the aperture: request=%llu > %s aperture=%llu\n",
+			  size, flags & PIN_MAPPABLE ? "mappable" : "total",
+			  end);
+		return -ENOSPC;
+	}
+
+	color = 0;
+
+	if (i915_vm_has_cache_coloring(vma->vm))
+		color = vma->obj->cache_level;
+
+	if (flags & PIN_OFFSET_FIXED) {
+		u64 offset = flags & PIN_OFFSET_MASK;
+		if (!IS_ALIGNED(offset, alignment) ||
+		    range_overflows(offset, size, end))
+			return -EINVAL;
+
+		ret = i915_gem_gtt_reserve(vma->vm, ww, &vma->node,
+					   size, offset, color,
+					   flags);
+		if (ret)
+			return ret;
+	} else {
+		/*
+		 * We only support huge gtt pages through the 48b PPGTT,
+		 * however we also don't want to force any alignment for
+		 * objects which need to be tightly packed into the low 32bits.
+		 *
+		 * Note that we assume that GGTT are limited to 4GiB for the
+		 * forseeable future. See also i915_ggtt_offset().
+		 */
+		if (upper_32_bits(end - 1) &&
+		    vma->page_sizes.sg > I915_GTT_PAGE_SIZE) {
+			/*
+			 * We can't mix 64K and 4K PTEs in the same page-table
+			 * (2M block), and so to avoid the ugliness and
+			 * complexity of coloring we opt for just aligning 64K
+			 * objects to 2M.
+			 */
+			u64 page_alignment =
+				rounddown_pow_of_two(vma->page_sizes.sg |
+						     I915_GTT_PAGE_SIZE_2M);
+
+			/*
+			 * Check we don't expand for the limited Global GTT
+			 * (mappable aperture is even more precious!). This
+			 * also checks that we exclude the aliasing-ppgtt.
+			 */
+			GEM_BUG_ON(i915_vma_is_ggtt(vma));
+
+			alignment = max(alignment, page_alignment);
+
+			if (vma->page_sizes.sg & I915_GTT_PAGE_SIZE_64K)
+				size = round_up(size, I915_GTT_PAGE_SIZE_2M);
+		}
+
+		ret = i915_gem_gtt_insert(vma->vm, ww, &vma->node,
+					  size, alignment, color,
+					  start, end, flags);
+		if (ret)
+			return ret;
+
+		GEM_BUG_ON(vma->node.start < start);
+		GEM_BUG_ON(vma->node.start + vma->node.size > end);
+	}
+	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	GEM_BUG_ON(!i915_gem_valid_gtt_space(vma, color));
+
+	list_move_tail(&vma->vm_link, &vma->vm->bound_list);
+
+	return 0;
+}
+
+static void
+i915_vma_detach(struct i915_vma *vma)
+{
+	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	GEM_BUG_ON(i915_vma_is_bound(vma, I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND));
+
+	/*
+	 * And finally now the object is completely decoupled from this
+	 * vma, we can drop its hold on the backing storage and allow
+	 * it to be reaped by the shrinker.
+	 */
+	list_move_tail(&vma->vm_link, &vma->vm->unbound_list);
+}
+
+static bool try_qad_pin(struct i915_vma *vma, unsigned int flags)
+{
+	unsigned int bound;
+
+	bound = atomic_read(&vma->flags);
+
+	if (flags & PIN_VALIDATE) {
+		flags &= I915_VMA_BIND_MASK;
+
+		return (flags & bound) == flags;
+	}
+
+	/* with the lock mandatory for unbind, we don't race here */
+	flags &= I915_VMA_BIND_MASK;
+	do {
+		if (unlikely(flags & ~bound))
+			return false;
+
+		if (unlikely(bound & (I915_VMA_OVERFLOW | I915_VMA_ERROR)))
+			return false;
+
+		GEM_BUG_ON(((bound + 1) & I915_VMA_PIN_MASK) == 0);
+	} while (!atomic_try_cmpxchg(&vma->flags, &bound, bound + 1));
+
+	return true;
+}
+
+static struct scatterlist *
+rotate_pages(struct drm_i915_gem_object *obj, unsigned int offset,
+	     unsigned int width, unsigned int height,
+	     unsigned int src_stride, unsigned int dst_stride,
+	     struct sg_table *st, struct scatterlist *sg)
+{
+	unsigned int column, row;
+	unsigned int src_idx;
+
+	for (column = 0; column < width; column++) {
+		unsigned int left;
+
+		src_idx = src_stride * (height - 1) + column + offset;
+		for (row = 0; row < height; row++) {
+			st->nents++;
+			/*
+			 * We don't need the pages, but need to initialize
+			 * the entries so the sg list can be happily traversed.
+			 * The only thing we need are DMA addresses.
+			 */
+			sg_set_page(sg, NULL, I915_GTT_PAGE_SIZE, 0);
+			sg_dma_address(sg) =
+				i915_gem_object_get_dma_address(obj, src_idx);
+			sg_dma_len(sg) = I915_GTT_PAGE_SIZE;
+			sg = sg_next(sg);
+			src_idx -= src_stride;
+		}
+
+		left = (dst_stride - height) * I915_GTT_PAGE_SIZE;
+
+		if (!left)
+			continue;
+
+		st->nents++;
+
+		/*
+		 * The DE ignores the PTEs for the padding tiles, the sg entry
+		 * here is just a conenience to indicate how many padding PTEs
+		 * to insert at this spot.
+		 */
+		sg_set_page(sg, NULL, left, 0);
+		sg_dma_address(sg) = 0;
+		sg_dma_len(sg) = left;
+		sg = sg_next(sg);
+	}
+
+	return sg;
+}
+
+static noinline struct sg_table *
+intel_rotate_pages(struct intel_rotation_info *rot_info,
+		   struct drm_i915_gem_object *obj)
+{
+	unsigned int size = intel_rotation_info_size(rot_info);
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct sg_table *st;
+	struct scatterlist *sg;
+	int ret = -ENOMEM;
+	int i;
+
+	/* Allocate target SG list. */
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (!st)
+		goto err_st_alloc;
+
+	ret = sg_alloc_table(st, size, GFP_KERNEL);
+	if (ret)
+		goto err_sg_alloc;
+
+	st->nents = 0;
+	sg = st->sgl;
+
+	for (i = 0 ; i < ARRAY_SIZE(rot_info->plane); i++)
+		sg = rotate_pages(obj, rot_info->plane[i].offset,
+				  rot_info->plane[i].width, rot_info->plane[i].height,
+				  rot_info->plane[i].src_stride,
+				  rot_info->plane[i].dst_stride,
+				  st, sg);
+
+	return st;
+
+err_sg_alloc:
+	kfree(st);
+err_st_alloc:
+
+	drm_dbg(&i915->drm, "Failed to create rotated mapping for object size %zu! (%ux%u tiles, %u pages)\n",
+		obj->base.size, rot_info->plane[0].width,
+		rot_info->plane[0].height, size);
+
+	return ERR_PTR(ret);
+}
+
+static struct scatterlist *
+add_padding_pages(unsigned int count,
+		  struct sg_table *st, struct scatterlist *sg)
+{
+	st->nents++;
+
+	/*
+	 * The DE ignores the PTEs for the padding tiles, the sg entry
+	 * here is just a convenience to indicate how many padding PTEs
+	 * to insert at this spot.
+	 */
+	sg_set_page(sg, NULL, count * I915_GTT_PAGE_SIZE, 0);
+	sg_dma_address(sg) = 0;
+	sg_dma_len(sg) = count * I915_GTT_PAGE_SIZE;
+	sg = sg_next(sg);
+
+	return sg;
+}
+
+static struct scatterlist *
+remap_tiled_color_plane_pages(struct drm_i915_gem_object *obj,
+			      unsigned int offset, unsigned int alignment_pad,
+			      unsigned int width, unsigned int height,
+			      unsigned int src_stride, unsigned int dst_stride,
+			      struct sg_table *st, struct scatterlist *sg,
+			      unsigned int *gtt_offset)
+{
+	unsigned int row;
+
+	if (!width || !height)
+		return sg;
+
+	if (alignment_pad)
+		sg = add_padding_pages(alignment_pad, st, sg);
+
+	for (row = 0; row < height; row++) {
+		unsigned int left = width * I915_GTT_PAGE_SIZE;
+
+		while (left) {
+			dma_addr_t addr;
+			unsigned int length;
+
+			/*
+			 * We don't need the pages, but need to initialize
+			 * the entries so the sg list can be happily traversed.
+			 * The only thing we need are DMA addresses.
+			 */
+
+			addr = i915_gem_object_get_dma_address_len(obj, offset, &length);
+
+			length = min(left, length);
+
+			st->nents++;
+
+			sg_set_page(sg, NULL, length, 0);
+			sg_dma_address(sg) = addr;
+			sg_dma_len(sg) = length;
+			sg = sg_next(sg);
+
+			offset += length / I915_GTT_PAGE_SIZE;
+			left -= length;
+		}
+
+		offset += src_stride - width;
+
+		left = (dst_stride - width) * I915_GTT_PAGE_SIZE;
+
+		if (!left)
+			continue;
+
+		sg = add_padding_pages(left >> PAGE_SHIFT, st, sg);
+	}
+
+	*gtt_offset += alignment_pad + dst_stride * height;
+
+	return sg;
+}
+
+static struct scatterlist *
+remap_contiguous_pages(struct drm_i915_gem_object *obj,
+		       unsigned int obj_offset,
+		       unsigned int count,
+		       struct sg_table *st, struct scatterlist *sg)
+{
+	struct scatterlist *iter;
+	unsigned int offset;
+
+	iter = i915_gem_object_get_sg_dma(obj, obj_offset, &offset);
+	GEM_BUG_ON(!iter);
+
+	do {
+		unsigned int len;
+
+		len = min(sg_dma_len(iter) - (offset << PAGE_SHIFT),
+			  count << PAGE_SHIFT);
+		sg_set_page(sg, NULL, len, 0);
+		sg_dma_address(sg) =
+			sg_dma_address(iter) + (offset << PAGE_SHIFT);
+		sg_dma_len(sg) = len;
+
+		st->nents++;
+		count -= len >> PAGE_SHIFT;
+		if (count == 0)
+			return sg;
+
+		sg = __sg_next(sg);
+		iter = __sg_next(iter);
+		offset = 0;
+	} while (1);
+}
+
+static struct scatterlist *
+remap_linear_color_plane_pages(struct drm_i915_gem_object *obj,
+			       unsigned int obj_offset, unsigned int alignment_pad,
+			       unsigned int size,
+			       struct sg_table *st, struct scatterlist *sg,
+			       unsigned int *gtt_offset)
+{
+	if (!size)
+		return sg;
+
+	if (alignment_pad)
+		sg = add_padding_pages(alignment_pad, st, sg);
+
+	sg = remap_contiguous_pages(obj, obj_offset, size, st, sg);
+	sg = sg_next(sg);
+
+	*gtt_offset += alignment_pad + size;
+
+	return sg;
+}
+
+static struct scatterlist *
+remap_color_plane_pages(const struct intel_remapped_info *rem_info,
+			struct drm_i915_gem_object *obj,
+			int color_plane,
+			struct sg_table *st, struct scatterlist *sg,
+			unsigned int *gtt_offset)
+{
+	unsigned int alignment_pad = 0;
+
+	if (rem_info->plane_alignment)
+		alignment_pad = ALIGN(*gtt_offset, rem_info->plane_alignment) - *gtt_offset;
+
+	if (rem_info->plane[color_plane].linear)
+		sg = remap_linear_color_plane_pages(obj,
+						    rem_info->plane[color_plane].offset,
+						    alignment_pad,
+						    rem_info->plane[color_plane].size,
+						    st, sg,
+						    gtt_offset);
+
+	else
+		sg = remap_tiled_color_plane_pages(obj,
+						   rem_info->plane[color_plane].offset,
+						   alignment_pad,
+						   rem_info->plane[color_plane].width,
+						   rem_info->plane[color_plane].height,
+						   rem_info->plane[color_plane].src_stride,
+						   rem_info->plane[color_plane].dst_stride,
+						   st, sg,
+						   gtt_offset);
+
+	return sg;
+}
+
+static noinline struct sg_table *
+intel_remap_pages(struct intel_remapped_info *rem_info,
+		  struct drm_i915_gem_object *obj)
+{
+	unsigned int size = intel_remapped_info_size(rem_info);
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct sg_table *st;
+	struct scatterlist *sg;
+	unsigned int gtt_offset = 0;
+	int ret = -ENOMEM;
+	int i;
+
+	/* Allocate target SG list. */
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (!st)
+		goto err_st_alloc;
+
+	ret = sg_alloc_table(st, size, GFP_KERNEL);
+	if (ret)
+		goto err_sg_alloc;
+
+	st->nents = 0;
+	sg = st->sgl;
+
+	for (i = 0 ; i < ARRAY_SIZE(rem_info->plane); i++)
+		sg = remap_color_plane_pages(rem_info, obj, i, st, sg, &gtt_offset);
+
+	i915_sg_trim(st);
+
+	return st;
+
+err_sg_alloc:
+	kfree(st);
+err_st_alloc:
+
+	drm_dbg(&i915->drm, "Failed to create remapped mapping for object size %zu! (%ux%u tiles, %u pages)\n",
+		obj->base.size, rem_info->plane[0].width,
+		rem_info->plane[0].height, size);
+
+	return ERR_PTR(ret);
+}
+
+static noinline struct sg_table *
+intel_partial_pages(const struct i915_gtt_view *view,
+		    struct drm_i915_gem_object *obj)
+{
+	struct sg_table *st;
+	struct scatterlist *sg;
+	unsigned int count = view->partial.size;
+	int ret = -ENOMEM;
+
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (!st)
+		goto err_st_alloc;
+
+	ret = sg_alloc_table(st, count, GFP_KERNEL);
+	if (ret)
+		goto err_sg_alloc;
+
+	st->nents = 0;
+
+	sg = remap_contiguous_pages(obj, view->partial.offset, count, st, st->sgl);
+
+	sg_mark_end(sg);
+	i915_sg_trim(st); /* Drop any unused tail entries. */
+
+	return st;
+
+err_sg_alloc:
+	kfree(st);
+err_st_alloc:
+	return ERR_PTR(ret);
+}
+
+static int
+__i915_vma_get_pages(struct i915_vma *vma)
+{
+	struct sg_table *pages;
+
+	/*
+	 * The vma->pages are only valid within the lifespan of the borrowed
+	 * obj->mm.pages. When the obj->mm.pages sg_table is regenerated, so
+	 * must be the vma->pages. A simple rule is that vma->pages must only
+	 * be accessed when the obj->mm.pages are pinned.
+	 */
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(vma->obj));
+
+	switch (vma->gtt_view.type) {
+	default:
+		GEM_BUG_ON(vma->gtt_view.type);
+		fallthrough;
+	case I915_GTT_VIEW_NORMAL:
+		pages = vma->obj->mm.pages;
+		break;
+
+	case I915_GTT_VIEW_ROTATED:
+		pages =
+			intel_rotate_pages(&vma->gtt_view.rotated, vma->obj);
+		break;
+
+	case I915_GTT_VIEW_REMAPPED:
+		pages =
+			intel_remap_pages(&vma->gtt_view.remapped, vma->obj);
+		break;
+
+	case I915_GTT_VIEW_PARTIAL:
+		pages = intel_partial_pages(&vma->gtt_view, vma->obj);
+		break;
+	}
+
+	if (IS_ERR(pages)) {
+		drm_err(&vma->vm->i915->drm,
+			"Failed to get pages for VMA view type %u (%ld)!\n",
+			vma->gtt_view.type, PTR_ERR(pages));
+		return PTR_ERR(pages);
+	}
+
+	vma->pages = pages;
+
+	return 0;
+}
+
+I915_SELFTEST_EXPORT int i915_vma_get_pages(struct i915_vma *vma)
+{
+	int err;
+
+	if (atomic_add_unless(&vma->pages_count, 1, 0))
+		return 0;
+
+	err = i915_gem_object_pin_pages(vma->obj);
+	if (err)
+		return err;
+
+	err = __i915_vma_get_pages(vma);
+	if (err)
+		goto err_unpin;
+
+	vma->page_sizes = vma->obj->mm.page_sizes;
+	atomic_inc(&vma->pages_count);
+
+	return 0;
+
+err_unpin:
+	__i915_gem_object_unpin_pages(vma->obj);
+
+	return err;
+}
+
+void vma_invalidate_tlb(struct i915_address_space *vm, u32 *tlb)
+{
+	/*
+	 * Before we release the pages that were bound by this vma, we
+	 * must invalidate all the TLBs that may still have a reference
+	 * back to our physical address. It only needs to be done once,
+	 * so after updating the PTE to point away from the pages, record
+	 * the most recent TLB invalidation seqno, and if we have not yet
+	 * flushed the TLBs upon release, perform a full invalidation.
+	 */
+	WRITE_ONCE(*tlb, intel_gt_next_invalidate_tlb_full(vm->gt));
+}
+
+static void __vma_put_pages(struct i915_vma *vma, unsigned int count)
+{
+	/* We allocate under vma_get_pages, so beware the shrinker */
+	GEM_BUG_ON(atomic_read(&vma->pages_count) < count);
+
+	if (atomic_sub_return(count, &vma->pages_count) == 0) {
+		if (vma->pages != vma->obj->mm.pages) {
+			sg_free_table(vma->pages);
+			kfree(vma->pages);
+		}
+		vma->pages = NULL;
+
+		i915_gem_object_unpin_pages(vma->obj);
+	}
+}
+
+I915_SELFTEST_EXPORT void i915_vma_put_pages(struct i915_vma *vma)
+{
+	if (atomic_add_unless(&vma->pages_count, -1, 1))
+		return;
+
+	__vma_put_pages(vma, 1);
+}
+
+static void vma_unbind_pages(struct i915_vma *vma)
+{
+	unsigned int count;
+
+	lockdep_assert_held(&vma->vm->mutex);
+
+	/* The upper portion of pages_count is the number of bindings */
+	count = atomic_read(&vma->pages_count);
+	count >>= I915_VMA_PAGES_BIAS;
+	GEM_BUG_ON(!count);
+
+	__vma_put_pages(vma, count | count << I915_VMA_PAGES_BIAS);
+}
+
+int i915_vma_pin_ww(struct i915_vma *vma, struct i915_gem_ww_ctx *ww,
+		    u64 size, u64 alignment, u64 flags)
+{
+	struct i915_vma_work *work = NULL;
+	struct dma_fence *moving = NULL;
+	struct i915_vma_resource *vma_res = NULL;
+	intel_wakeref_t wakeref = 0;
+	unsigned int bound;
+	int err;
+
+	assert_vma_held(vma);
+	GEM_BUG_ON(!ww);
+
+	BUILD_BUG_ON(PIN_GLOBAL != I915_VMA_GLOBAL_BIND);
+	BUILD_BUG_ON(PIN_USER != I915_VMA_LOCAL_BIND);
+
+	GEM_BUG_ON(!(flags & (PIN_USER | PIN_GLOBAL)));
+
+	/* First try and grab the pin without rebinding the vma */
+	if (try_qad_pin(vma, flags))
+		return 0;
+
+	err = i915_vma_get_pages(vma);
+	if (err)
+		return err;
+
+	if (flags & PIN_GLOBAL)
+		wakeref = intel_runtime_pm_get(&vma->vm->i915->runtime_pm);
+
+	if (flags & vma->vm->bind_async_flags) {
+		/* lock VM */
+		err = i915_vm_lock_objects(vma->vm, ww);
+		if (err)
+			goto err_rpm;
+
+		work = i915_vma_work();
+		if (!work) {
+			err = -ENOMEM;
+			goto err_rpm;
+		}
+
+		work->vm = vma->vm;
+
+		err = i915_gem_object_get_moving_fence(vma->obj, &moving);
+		if (err)
+			goto err_rpm;
+
+		dma_fence_work_chain(&work->base, moving);
+
+		/* Allocate enough page directories to used PTE */
+		if (vma->vm->allocate_va_range) {
+			err = i915_vm_alloc_pt_stash(vma->vm,
+						     &work->stash,
+						     vma->size);
+			if (err)
+				goto err_fence;
+
+			err = i915_vm_map_pt_stash(vma->vm, &work->stash);
+			if (err)
+				goto err_fence;
+		}
+	}
+
+	vma_res = i915_vma_resource_alloc();
+	if (IS_ERR(vma_res)) {
+		err = PTR_ERR(vma_res);
+		goto err_fence;
+	}
+
+	/*
+	 * Differentiate between user/kernel vma inside the aliasing-ppgtt.
+	 *
+	 * We conflate the Global GTT with the user's vma when using the
+	 * aliasing-ppgtt, but it is still vitally important to try and
+	 * keep the use cases distinct. For example, userptr objects are
+	 * not allowed inside the Global GTT as that will cause lock
+	 * inversions when we have to evict them the mmu_notifier callbacks -
+	 * but they are allowed to be part of the user ppGTT which can never
+	 * be mapped. As such we try to give the distinct users of the same
+	 * mutex, distinct lockclasses [equivalent to how we keep i915_ggtt
+	 * and i915_ppgtt separate].
+	 *
+	 * NB this may cause us to mask real lock inversions -- while the
+	 * code is safe today, lockdep may not be able to spot future
+	 * transgressions.
+	 */
+	err = mutex_lock_interruptible_nested(&vma->vm->mutex,
+					      !(flags & PIN_GLOBAL));
+	if (err)
+		goto err_vma_res;
+
+	/* No more allocations allowed now we hold vm->mutex */
+
+	if (unlikely(i915_vma_is_closed(vma))) {
+		err = -ENOENT;
+		goto err_unlock;
+	}
+
+	bound = atomic_read(&vma->flags);
+	if (unlikely(bound & I915_VMA_ERROR)) {
+		err = -ENOMEM;
+		goto err_unlock;
+	}
+
+	if (unlikely(!((bound + 1) & I915_VMA_PIN_MASK))) {
+		err = -EAGAIN; /* pins are meant to be fairly temporary */
+		goto err_unlock;
+	}
+
+	if (unlikely(!(flags & ~bound & I915_VMA_BIND_MASK))) {
+		if (!(flags & PIN_VALIDATE))
+			__i915_vma_pin(vma);
+		goto err_unlock;
+	}
+
+	err = i915_active_acquire(&vma->active);
+	if (err)
+		goto err_unlock;
+
+	if (!(bound & I915_VMA_BIND_MASK)) {
+		err = i915_vma_insert(vma, ww, size, alignment, flags);
+		if (err)
+			goto err_active;
+
+		if (i915_is_ggtt(vma->vm))
+			__i915_vma_set_map_and_fenceable(vma);
+	}
+
+	GEM_BUG_ON(!vma->pages);
+	err = i915_vma_bind(vma,
+			    vma->obj->cache_level,
+			    flags, work, vma_res);
+	vma_res = NULL;
+	if (err)
+		goto err_remove;
+
+	/* There should only be at most 2 active bindings (user, global) */
+	GEM_BUG_ON(bound + I915_VMA_PAGES_ACTIVE < bound);
+	atomic_add(I915_VMA_PAGES_ACTIVE, &vma->pages_count);
+	list_move_tail(&vma->vm_link, &vma->vm->bound_list);
+
+	if (!(flags & PIN_VALIDATE)) {
+		__i915_vma_pin(vma);
+		GEM_BUG_ON(!i915_vma_is_pinned(vma));
+	}
+	GEM_BUG_ON(!i915_vma_is_bound(vma, flags));
+	GEM_BUG_ON(i915_vma_misplaced(vma, size, alignment, flags));
+
+err_remove:
+	if (!i915_vma_is_bound(vma, I915_VMA_BIND_MASK)) {
+		i915_vma_detach(vma);
+		drm_mm_remove_node(&vma->node);
+	}
+err_active:
+	i915_active_release(&vma->active);
+err_unlock:
+	mutex_unlock(&vma->vm->mutex);
+err_vma_res:
+	i915_vma_resource_free(vma_res);
+err_fence:
+	if (work)
+		dma_fence_work_commit_imm(&work->base);
+err_rpm:
+	if (wakeref)
+		intel_runtime_pm_put(&vma->vm->i915->runtime_pm, wakeref);
+
+	if (moving)
+		dma_fence_put(moving);
+
+	i915_vma_put_pages(vma);
+	return err;
+}
+
+static void flush_idle_contexts(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id)
+		intel_engine_flush_barriers(engine);
+
+	intel_gt_wait_for_idle(gt, MAX_SCHEDULE_TIMEOUT);
+}
+
+static int __i915_ggtt_pin(struct i915_vma *vma, struct i915_gem_ww_ctx *ww,
+			   u32 align, unsigned int flags)
+{
+	struct i915_address_space *vm = vma->vm;
+	int err;
+
+	do {
+		err = i915_vma_pin_ww(vma, ww, 0, align, flags | PIN_GLOBAL);
+
+		if (err != -ENOSPC) {
+			if (!err) {
+				err = i915_vma_wait_for_bind(vma);
+				if (err)
+					i915_vma_unpin(vma);
+			}
+			return err;
+		}
+
+		/* Unlike i915_vma_pin, we don't take no for an answer! */
+		flush_idle_contexts(vm->gt);
+		if (mutex_lock_interruptible(&vm->mutex) == 0) {
+			/*
+			 * We pass NULL ww here, as we don't want to unbind
+			 * locked objects when called from execbuf when pinning
+			 * is removed. This would probably regress badly.
+			 */
+			i915_gem_evict_vm(vm, NULL, NULL);
+			mutex_unlock(&vm->mutex);
+		}
+	} while (1);
+}
+
+int i915_ggtt_pin(struct i915_vma *vma, struct i915_gem_ww_ctx *ww,
+		  u32 align, unsigned int flags)
+{
+	struct i915_gem_ww_ctx _ww;
+	int err;
+
+	GEM_BUG_ON(!i915_vma_is_ggtt(vma));
+
+	if (ww)
+		return __i915_ggtt_pin(vma, ww, align, flags);
+
+	lockdep_assert_not_held(&vma->obj->base.resv->lock.base);
+
+	for_i915_gem_ww(&_ww, err, true) {
+		err = i915_gem_object_lock(vma->obj, &_ww);
+		if (!err)
+			err = __i915_ggtt_pin(vma, &_ww, align, flags);
+	}
+
+	return err;
+}
+
+static void __vma_close(struct i915_vma *vma, struct intel_gt *gt)
+{
+	/*
+	 * We defer actually closing, unbinding and destroying the VMA until
+	 * the next idle point, or if the object is freed in the meantime. By
+	 * postponing the unbind, we allow for it to be resurrected by the
+	 * client, avoiding the work required to rebind the VMA. This is
+	 * advantageous for DRI, where the client/server pass objects
+	 * between themselves, temporarily opening a local VMA to the
+	 * object, and then closing it again. The same object is then reused
+	 * on the next frame (or two, depending on the depth of the swap queue)
+	 * causing us to rebind the VMA once more. This ends up being a lot
+	 * of wasted work for the steady state.
+	 */
+	GEM_BUG_ON(i915_vma_is_closed(vma));
+	list_add(&vma->closed_link, &gt->closed_vma);
+}
+
+void i915_vma_close(struct i915_vma *vma)
+{
+	struct intel_gt *gt = vma->vm->gt;
+	unsigned long flags;
+
+	if (i915_vma_is_ggtt(vma))
+		return;
+
+	GEM_BUG_ON(!atomic_read(&vma->open_count));
+	if (atomic_dec_and_lock_irqsave(&vma->open_count,
+					&gt->closed_lock,
+					flags)) {
+		__vma_close(vma, gt);
+		spin_unlock_irqrestore(&gt->closed_lock, flags);
+	}
+}
+
+static void __i915_vma_remove_closed(struct i915_vma *vma)
+{
+	list_del_init(&vma->closed_link);
+}
+
+void i915_vma_reopen(struct i915_vma *vma)
+{
+	struct intel_gt *gt = vma->vm->gt;
+
+	spin_lock_irq(&gt->closed_lock);
+	if (i915_vma_is_closed(vma))
+		__i915_vma_remove_closed(vma);
+	spin_unlock_irq(&gt->closed_lock);
+}
+
+static void force_unbind(struct i915_vma *vma)
+{
+	if (!drm_mm_node_allocated(&vma->node))
+		return;
+
+	atomic_and(~I915_VMA_PIN_MASK, &vma->flags);
+	WARN_ON(__i915_vma_unbind(vma));
+	GEM_BUG_ON(drm_mm_node_allocated(&vma->node));
+}
+
+static void release_references(struct i915_vma *vma, struct intel_gt *gt,
+			       bool vm_ddestroy)
+{
+	struct drm_i915_gem_object *obj = vma->obj;
+
+	GEM_BUG_ON(i915_vma_is_active(vma));
+
+	spin_lock(&obj->vma.lock);
+	list_del(&vma->obj_link);
+	if (!RB_EMPTY_NODE(&vma->obj_node))
+		rb_erase(&vma->obj_node, &obj->vma.tree);
+
+	spin_unlock(&obj->vma.lock);
+
+	spin_lock_irq(&gt->closed_lock);
+	__i915_vma_remove_closed(vma);
+	spin_unlock_irq(&gt->closed_lock);
+
+	if (vm_ddestroy)
+		i915_vm_resv_put(vma->vm);
+
+	i915_active_fini(&vma->active);
+	GEM_WARN_ON(vma->resource);
+	i915_vma_free(vma);
+}
+
+/**
+ * i915_vma_destroy_locked - Remove all weak reference to the vma and put
+ * the initial reference.
+ *
+ * This function should be called when it's decided the vma isn't needed
+ * anymore. The caller must assure that it doesn't race with another lookup
+ * plus destroy, typically by taking an appropriate reference.
+ *
+ * Current callsites are
+ * - __i915_gem_object_pages_fini()
+ * - __i915_vm_close() - Blocks the above function by taking a reference on
+ * the object.
+ * - __i915_vma_parked() - Blocks the above functions by taking a reference
+ * on the vm and a reference on the object. Also takes the object lock so
+ * destruction from __i915_vma_parked() can be blocked by holding the
+ * object lock. Since the object lock is only allowed from within i915 with
+ * an object refcount, holding the object lock also implicitly blocks the
+ * vma freeing from __i915_gem_object_pages_fini().
+ *
+ * Because of locks taken during destruction, a vma is also guaranteed to
+ * stay alive while the following locks are held if it was looked up while
+ * holding one of the locks:
+ * - vm->mutex
+ * - obj->vma.lock
+ * - gt->closed_lock
+ */
+void i915_vma_destroy_locked(struct i915_vma *vma)
+{
+	lockdep_assert_held(&vma->vm->mutex);
+
+	force_unbind(vma);
+	list_del_init(&vma->vm_link);
+	release_references(vma, vma->vm->gt, false);
+}
+
+void i915_vma_destroy(struct i915_vma *vma)
+{
+	struct intel_gt *gt;
+	bool vm_ddestroy;
+
+	mutex_lock(&vma->vm->mutex);
+	force_unbind(vma);
+	list_del_init(&vma->vm_link);
+	vm_ddestroy = vma->vm_ddestroy;
+	vma->vm_ddestroy = false;
+
+	/* vma->vm may be freed when releasing vma->vm->mutex. */
+	gt = vma->vm->gt;
+	mutex_unlock(&vma->vm->mutex);
+	release_references(vma, gt, vm_ddestroy);
+}
+
+void i915_vma_parked(struct intel_gt *gt)
+{
+	struct i915_vma *vma, *next;
+	LIST_HEAD(closed);
+
+	spin_lock_irq(&gt->closed_lock);
+	list_for_each_entry_safe(vma, next, &gt->closed_vma, closed_link) {
+		struct drm_i915_gem_object *obj = vma->obj;
+		struct i915_address_space *vm = vma->vm;
+
+		/* XXX All to avoid keeping a reference on i915_vma itself */
+
+		if (!kref_get_unless_zero(&obj->base.refcount))
+			continue;
+
+		if (!i915_vm_tryget(vm)) {
+			i915_gem_object_put(obj);
+			continue;
+		}
+
+		list_move(&vma->closed_link, &closed);
+	}
+	spin_unlock_irq(&gt->closed_lock);
+
+	/* As the GT is held idle, no vma can be reopened as we destroy them */
+	list_for_each_entry_safe(vma, next, &closed, closed_link) {
+		struct drm_i915_gem_object *obj = vma->obj;
+		struct i915_address_space *vm = vma->vm;
+
+		if (i915_gem_object_trylock(obj, NULL)) {
+			INIT_LIST_HEAD(&vma->closed_link);
+			i915_vma_destroy(vma);
+			i915_gem_object_unlock(obj);
+		} else {
+			/* back you go.. */
+			spin_lock_irq(&gt->closed_lock);
+			list_add(&vma->closed_link, &gt->closed_vma);
+			spin_unlock_irq(&gt->closed_lock);
+		}
+
+		i915_gem_object_put(obj);
+		i915_vm_put(vm);
+	}
+}
+
+static void __i915_vma_iounmap(struct i915_vma *vma)
+{
+	GEM_BUG_ON(i915_vma_is_pinned(vma));
+
+	if (vma->iomap == NULL)
+		return;
+
+	if (page_unmask_bits(vma->iomap))
+		__i915_gem_object_release_map(vma->obj);
+	else
+		io_mapping_unmap(vma->iomap);
+	vma->iomap = NULL;
+}
+
+void i915_vma_revoke_mmap(struct i915_vma *vma)
+{
+	struct drm_vma_offset_node *node;
+	u64 vma_offset;
+
+	if (!i915_vma_has_userfault(vma))
+		return;
+
+	GEM_BUG_ON(!i915_vma_is_map_and_fenceable(vma));
+	GEM_BUG_ON(!vma->obj->userfault_count);
+
+	node = &vma->mmo->vma_node;
+	vma_offset = vma->gtt_view.partial.offset << PAGE_SHIFT;
+	unmap_mapping_range(vma->vm->i915->drm.anon_inode->i_mapping,
+			    drm_vma_node_offset_addr(node) + vma_offset,
+			    vma->size,
+			    1);
+
+	i915_vma_unset_userfault(vma);
+	if (!--vma->obj->userfault_count)
+		list_del(&vma->obj->userfault_link);
+}
+
+static int
+__i915_request_await_bind(struct i915_request *rq, struct i915_vma *vma)
+{
+	return __i915_request_await_exclusive(rq, &vma->active);
+}
+
+static int __i915_vma_move_to_active(struct i915_vma *vma, struct i915_request *rq)
+{
+	int err;
+
+	/* Wait for the vma to be bound before we start! */
+	err = __i915_request_await_bind(rq, vma);
+	if (err)
+		return err;
+
+	return i915_active_add_request(&vma->active, rq);
+}
+
+int _i915_vma_move_to_active(struct i915_vma *vma,
+			     struct i915_request *rq,
+			     struct dma_fence *fence,
+			     unsigned int flags)
+{
+	struct drm_i915_gem_object *obj = vma->obj;
+	int err;
+
+	assert_object_held(obj);
+
+	GEM_BUG_ON(!vma->pages);
+
+	err = __i915_vma_move_to_active(vma, rq);
+	if (unlikely(err))
+		return err;
+
+	/*
+	 * Reserve fences slot early to prevent an allocation after preparing
+	 * the workload and associating fences with dma_resv.
+	 */
+	if (fence && !(flags & __EXEC_OBJECT_NO_RESERVE)) {
+		struct dma_fence *curr;
+		int idx;
+
+		dma_fence_array_for_each(curr, idx, fence)
+			;
+		err = dma_resv_reserve_fences(vma->obj->base.resv, idx);
+		if (unlikely(err))
+			return err;
+	}
+
+	if (flags & EXEC_OBJECT_WRITE) {
+		struct intel_frontbuffer *front;
+
+		front = __intel_frontbuffer_get(obj);
+		if (unlikely(front)) {
+			if (intel_frontbuffer_invalidate(front, ORIGIN_CS))
+				i915_active_add_request(&front->write, rq);
+			intel_frontbuffer_put(front);
+		}
+	}
+
+	if (fence) {
+		struct dma_fence *curr;
+		enum dma_resv_usage usage;
+		int idx;
+
+		if (flags & EXEC_OBJECT_WRITE) {
+			usage = DMA_RESV_USAGE_WRITE;
+			obj->write_domain = I915_GEM_DOMAIN_RENDER;
+			obj->read_domains = 0;
+		} else {
+			usage = DMA_RESV_USAGE_READ;
+			obj->write_domain = 0;
+		}
+
+		dma_fence_array_for_each(curr, idx, fence)
+			dma_resv_add_fence(vma->obj->base.resv, curr, usage);
+	}
+
+	if (flags & EXEC_OBJECT_NEEDS_FENCE && vma->fence)
+		i915_active_add_request(&vma->fence->active, rq);
+
+	obj->read_domains |= I915_GEM_GPU_DOMAINS;
+	obj->mm.dirty = true;
+
+	GEM_BUG_ON(!i915_vma_is_active(vma));
+	return 0;
+}
+
+struct dma_fence *__i915_vma_evict(struct i915_vma *vma, bool async)
+{
+	struct i915_vma_resource *vma_res = vma->resource;
+	struct dma_fence *unbind_fence;
+
+	GEM_BUG_ON(i915_vma_is_pinned(vma));
+	assert_vma_held_evict(vma);
+
+	if (i915_vma_is_map_and_fenceable(vma)) {
+		/* Force a pagefault for domain tracking on next user access */
+		i915_vma_revoke_mmap(vma);
+
+		/*
+		 * Check that we have flushed all writes through the GGTT
+		 * before the unbind, other due to non-strict nature of those
+		 * indirect writes they may end up referencing the GGTT PTE
+		 * after the unbind.
+		 *
+		 * Note that we may be concurrently poking at the GGTT_WRITE
+		 * bit from set-domain, as we mark all GGTT vma associated
+		 * with an object. We know this is for another vma, as we
+		 * are currently unbinding this one -- so if this vma will be
+		 * reused, it will be refaulted and have its dirty bit set
+		 * before the next write.
+		 */
+		i915_vma_flush_writes(vma);
+
+		/* release the fence reg _after_ flushing */
+		i915_vma_revoke_fence(vma);
+
+		clear_bit(I915_VMA_CAN_FENCE_BIT, __i915_vma_flags(vma));
+	}
+
+	__i915_vma_iounmap(vma);
+
+	GEM_BUG_ON(vma->fence);
+	GEM_BUG_ON(i915_vma_has_userfault(vma));
+
+	/* Object backend must be async capable. */
+	GEM_WARN_ON(async && !vma->resource->bi.pages_rsgt);
+
+	/* If vm is not open, unbind is a nop. */
+	vma_res->needs_wakeref = i915_vma_is_bound(vma, I915_VMA_GLOBAL_BIND) &&
+		kref_read(&vma->vm->ref);
+	vma_res->skip_pte_rewrite = !kref_read(&vma->vm->ref) ||
+		vma->vm->skip_pte_rewrite;
+	trace_i915_vma_unbind(vma);
+
+	if (async)
+		unbind_fence = i915_vma_resource_unbind(vma_res,
+							&vma->obj->mm.tlb);
+	else
+		unbind_fence = i915_vma_resource_unbind(vma_res, NULL);
+
+	vma->resource = NULL;
+
+	atomic_and(~(I915_VMA_BIND_MASK | I915_VMA_ERROR | I915_VMA_GGTT_WRITE),
+		   &vma->flags);
+
+	i915_vma_detach(vma);
+
+	if (!async) {
+		if (unbind_fence) {
+			dma_fence_wait(unbind_fence, false);
+			dma_fence_put(unbind_fence);
+			unbind_fence = NULL;
+		}
+		vma_invalidate_tlb(vma->vm, &vma->obj->mm.tlb);
+	}
+
+	/*
+	 * Binding itself may not have completed until the unbind fence signals,
+	 * so don't drop the pages until that happens, unless the resource is
+	 * async_capable.
+	 */
+
+	vma_unbind_pages(vma);
+	return unbind_fence;
+}
+
+int __i915_vma_unbind(struct i915_vma *vma)
+{
+	int ret;
+
+	lockdep_assert_held(&vma->vm->mutex);
+	assert_vma_held_evict(vma);
+
+	if (!drm_mm_node_allocated(&vma->node))
+		return 0;
+
+	if (i915_vma_is_pinned(vma)) {
+		vma_print_allocator(vma, "is pinned");
+		return -EAGAIN;
+	}
+
+	/*
+	 * After confirming that no one else is pinning this vma, wait for
+	 * any laggards who may have crept in during the wait (through
+	 * a residual pin skipping the vm->mutex) to complete.
+	 */
+	ret = i915_vma_sync(vma);
+	if (ret)
+		return ret;
+
+	GEM_BUG_ON(i915_vma_is_active(vma));
+	__i915_vma_evict(vma, false);
+
+	drm_mm_remove_node(&vma->node); /* pairs with i915_vma_release() */
+	return 0;
+}
+
+static struct dma_fence *__i915_vma_unbind_async(struct i915_vma *vma)
+{
+	struct dma_fence *fence;
+
+	lockdep_assert_held(&vma->vm->mutex);
+
+	if (!drm_mm_node_allocated(&vma->node))
+		return NULL;
+
+	if (i915_vma_is_pinned(vma) ||
+	    &vma->obj->mm.rsgt->table != vma->resource->bi.pages)
+		return ERR_PTR(-EAGAIN);
+
+	/*
+	 * We probably need to replace this with awaiting the fences of the
+	 * object's dma_resv when the vma active goes away. When doing that
+	 * we need to be careful to not add the vma_resource unbind fence
+	 * immediately to the object's dma_resv, because then unbinding
+	 * the next vma from the object, in case there are many, will
+	 * actually await the unbinding of the previous vmas, which is
+	 * undesirable.
+	 */
+	if (i915_sw_fence_await_active(&vma->resource->chain, &vma->active,
+				       I915_ACTIVE_AWAIT_EXCL |
+				       I915_ACTIVE_AWAIT_ACTIVE) < 0) {
+		return ERR_PTR(-EBUSY);
+	}
+
+	fence = __i915_vma_evict(vma, true);
+
+	drm_mm_remove_node(&vma->node); /* pairs with i915_vma_release() */
+
+	return fence;
+}
+
+int i915_vma_unbind(struct i915_vma *vma)
+{
+	struct i915_address_space *vm = vma->vm;
+	intel_wakeref_t wakeref = 0;
+	int err;
+
+	assert_object_held_shared(vma->obj);
+
+	/* Optimistic wait before taking the mutex */
+	err = i915_vma_sync(vma);
+	if (err)
+		return err;
+
+	if (!drm_mm_node_allocated(&vma->node))
+		return 0;
+
+	if (i915_vma_is_pinned(vma)) {
+		vma_print_allocator(vma, "is pinned");
+		return -EAGAIN;
+	}
+
+	if (i915_vma_is_bound(vma, I915_VMA_GLOBAL_BIND))
+		/* XXX not always required: nop_clear_range */
+		wakeref = intel_runtime_pm_get(&vm->i915->runtime_pm);
+
+	err = mutex_lock_interruptible_nested(&vma->vm->mutex, !wakeref);
+	if (err)
+		goto out_rpm;
+
+	err = __i915_vma_unbind(vma);
+	mutex_unlock(&vm->mutex);
+
+out_rpm:
+	if (wakeref)
+		intel_runtime_pm_put(&vm->i915->runtime_pm, wakeref);
+	return err;
+}
+
+int i915_vma_unbind_async(struct i915_vma *vma, bool trylock_vm)
+{
+	struct drm_i915_gem_object *obj = vma->obj;
+	struct i915_address_space *vm = vma->vm;
+	intel_wakeref_t wakeref = 0;
+	struct dma_fence *fence;
+	int err;
+
+	/*
+	 * We need the dma-resv lock since we add the
+	 * unbind fence to the dma-resv object.
+	 */
+	assert_object_held(obj);
+
+	if (!drm_mm_node_allocated(&vma->node))
+		return 0;
+
+	if (i915_vma_is_pinned(vma)) {
+		vma_print_allocator(vma, "is pinned");
+		return -EAGAIN;
+	}
+
+	if (!obj->mm.rsgt)
+		return -EBUSY;
+
+	err = dma_resv_reserve_fences(obj->base.resv, 2);
+	if (err)
+		return -EBUSY;
+
+	/*
+	 * It would be great if we could grab this wakeref from the
+	 * async unbind work if needed, but we can't because it uses
+	 * kmalloc and it's in the dma-fence signalling critical path.
+	 */
+	if (i915_vma_is_bound(vma, I915_VMA_GLOBAL_BIND))
+		wakeref = intel_runtime_pm_get(&vm->i915->runtime_pm);
+
+	if (trylock_vm && !mutex_trylock(&vm->mutex)) {
+		err = -EBUSY;
+		goto out_rpm;
+	} else if (!trylock_vm) {
+		err = mutex_lock_interruptible_nested(&vm->mutex, !wakeref);
+		if (err)
+			goto out_rpm;
+	}
+
+	fence = __i915_vma_unbind_async(vma);
+	mutex_unlock(&vm->mutex);
+	if (IS_ERR_OR_NULL(fence)) {
+		err = PTR_ERR_OR_ZERO(fence);
+		goto out_rpm;
+	}
+
+	dma_resv_add_fence(obj->base.resv, fence, DMA_RESV_USAGE_READ);
+	dma_fence_put(fence);
+
+out_rpm:
+	if (wakeref)
+		intel_runtime_pm_put(&vm->i915->runtime_pm, wakeref);
+	return err;
+}
+
+int i915_vma_unbind_unlocked(struct i915_vma *vma)
+{
+	int err;
+
+	i915_gem_object_lock(vma->obj, NULL);
+	err = i915_vma_unbind(vma);
+	i915_gem_object_unlock(vma->obj);
+
+	return err;
+}
+
+struct i915_vma *i915_vma_make_unshrinkable(struct i915_vma *vma)
+{
+	i915_gem_object_make_unshrinkable(vma->obj);
+	return vma;
+}
+
+void i915_vma_make_shrinkable(struct i915_vma *vma)
+{
+	i915_gem_object_make_shrinkable(vma->obj);
+}
+
+void i915_vma_make_purgeable(struct i915_vma *vma)
+{
+	i915_gem_object_make_purgeable(vma->obj);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/i915_vma.c"
+#endif
+
+void i915_vma_module_exit(void)
+{
+	kmem_cache_destroy(slab_vmas);
+}
+
+int __init i915_vma_module_init(void)
+{
+	slab_vmas = KMEM_CACHE(i915_vma, SLAB_HWCACHE_ALIGN);
+	if (!slab_vmas)
+		return -ENOMEM;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_vma.h b/drivers/gpu/drm/i915/i915_vma.h
new file mode 100644
index 000000000..aecd9c644
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_vma.h
@@ -0,0 +1,444 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __I915_VMA_H__
+#define __I915_VMA_H__
+
+#include <linux/io-mapping.h>
+#include <linux/rbtree.h>
+
+#include <drm/drm_mm.h>
+
+#include "gt/intel_ggtt_fencing.h"
+#include "gem/i915_gem_object.h"
+
+#include "i915_gem_gtt.h"
+
+#include "i915_active.h"
+#include "i915_request.h"
+#include "i915_vma_resource.h"
+#include "i915_vma_types.h"
+
+struct i915_vma *
+i915_vma_instance(struct drm_i915_gem_object *obj,
+		  struct i915_address_space *vm,
+		  const struct i915_gtt_view *view);
+
+void i915_vma_unpin_and_release(struct i915_vma **p_vma, unsigned int flags);
+#define I915_VMA_RELEASE_MAP BIT(0)
+
+static inline bool i915_vma_is_active(const struct i915_vma *vma)
+{
+	return !i915_active_is_idle(&vma->active);
+}
+
+/* do not reserve memory to prevent deadlocks */
+#define __EXEC_OBJECT_NO_RESERVE BIT(31)
+
+int __must_check _i915_vma_move_to_active(struct i915_vma *vma,
+					  struct i915_request *rq,
+					  struct dma_fence *fence,
+					  unsigned int flags);
+static inline int __must_check
+i915_vma_move_to_active(struct i915_vma *vma, struct i915_request *rq,
+			unsigned int flags)
+{
+	return _i915_vma_move_to_active(vma, rq, &rq->fence, flags);
+}
+
+#define __i915_vma_flags(v) ((unsigned long *)&(v)->flags.counter)
+
+static inline bool i915_vma_is_ggtt(const struct i915_vma *vma)
+{
+	return test_bit(I915_VMA_GGTT_BIT, __i915_vma_flags(vma));
+}
+
+static inline bool i915_vma_is_dpt(const struct i915_vma *vma)
+{
+	return i915_is_dpt(vma->vm);
+}
+
+static inline bool i915_vma_has_ggtt_write(const struct i915_vma *vma)
+{
+	return test_bit(I915_VMA_GGTT_WRITE_BIT, __i915_vma_flags(vma));
+}
+
+static inline void i915_vma_set_ggtt_write(struct i915_vma *vma)
+{
+	GEM_BUG_ON(!i915_vma_is_ggtt(vma));
+	set_bit(I915_VMA_GGTT_WRITE_BIT, __i915_vma_flags(vma));
+}
+
+static inline bool i915_vma_unset_ggtt_write(struct i915_vma *vma)
+{
+	return test_and_clear_bit(I915_VMA_GGTT_WRITE_BIT,
+				  __i915_vma_flags(vma));
+}
+
+void i915_vma_flush_writes(struct i915_vma *vma);
+
+static inline bool i915_vma_is_map_and_fenceable(const struct i915_vma *vma)
+{
+	return test_bit(I915_VMA_CAN_FENCE_BIT, __i915_vma_flags(vma));
+}
+
+static inline bool i915_vma_set_userfault(struct i915_vma *vma)
+{
+	GEM_BUG_ON(!i915_vma_is_map_and_fenceable(vma));
+	return test_and_set_bit(I915_VMA_USERFAULT_BIT, __i915_vma_flags(vma));
+}
+
+static inline void i915_vma_unset_userfault(struct i915_vma *vma)
+{
+	return clear_bit(I915_VMA_USERFAULT_BIT, __i915_vma_flags(vma));
+}
+
+static inline bool i915_vma_has_userfault(const struct i915_vma *vma)
+{
+	return test_bit(I915_VMA_USERFAULT_BIT, __i915_vma_flags(vma));
+}
+
+static inline bool i915_vma_is_closed(const struct i915_vma *vma)
+{
+	return !list_empty(&vma->closed_link);
+}
+
+static inline u32 i915_ggtt_offset(const struct i915_vma *vma)
+{
+	GEM_BUG_ON(!i915_vma_is_ggtt(vma));
+	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	GEM_BUG_ON(upper_32_bits(vma->node.start));
+	GEM_BUG_ON(upper_32_bits(vma->node.start + vma->node.size - 1));
+	return lower_32_bits(vma->node.start);
+}
+
+static inline u32 i915_ggtt_pin_bias(struct i915_vma *vma)
+{
+	return i915_vm_to_ggtt(vma->vm)->pin_bias;
+}
+
+static inline struct i915_vma *i915_vma_get(struct i915_vma *vma)
+{
+	i915_gem_object_get(vma->obj);
+	return vma;
+}
+
+static inline struct i915_vma *i915_vma_tryget(struct i915_vma *vma)
+{
+	if (likely(kref_get_unless_zero(&vma->obj->base.refcount)))
+		return vma;
+
+	return NULL;
+}
+
+static inline void i915_vma_put(struct i915_vma *vma)
+{
+	i915_gem_object_put(vma->obj);
+}
+
+static inline long
+i915_vma_compare(struct i915_vma *vma,
+		 struct i915_address_space *vm,
+		 const struct i915_gtt_view *view)
+{
+	ptrdiff_t cmp;
+
+	GEM_BUG_ON(view && !i915_is_ggtt_or_dpt(vm));
+
+	cmp = ptrdiff(vma->vm, vm);
+	if (cmp)
+		return cmp;
+
+	BUILD_BUG_ON(I915_GTT_VIEW_NORMAL != 0);
+	cmp = vma->gtt_view.type;
+	if (!view)
+		return cmp;
+
+	cmp -= view->type;
+	if (cmp)
+		return cmp;
+
+	assert_i915_gem_gtt_types();
+
+	/* gtt_view.type also encodes its size so that we both distinguish
+	 * different views using it as a "type" and also use a compact (no
+	 * accessing of uninitialised padding bytes) memcmp without storing
+	 * an extra parameter or adding more code.
+	 *
+	 * To ensure that the memcmp is valid for all branches of the union,
+	 * even though the code looks like it is just comparing one branch,
+	 * we assert above that all branches have the same address, and that
+	 * each branch has a unique type/size.
+	 */
+	BUILD_BUG_ON(I915_GTT_VIEW_NORMAL >= I915_GTT_VIEW_PARTIAL);
+	BUILD_BUG_ON(I915_GTT_VIEW_PARTIAL >= I915_GTT_VIEW_ROTATED);
+	BUILD_BUG_ON(I915_GTT_VIEW_ROTATED >= I915_GTT_VIEW_REMAPPED);
+	BUILD_BUG_ON(offsetof(typeof(*view), rotated) !=
+		     offsetof(typeof(*view), partial));
+	BUILD_BUG_ON(offsetof(typeof(*view), rotated) !=
+		     offsetof(typeof(*view), remapped));
+	return memcmp(&vma->gtt_view.partial, &view->partial, view->type);
+}
+
+struct i915_vma_work *i915_vma_work(void);
+int i915_vma_bind(struct i915_vma *vma,
+		  enum i915_cache_level cache_level,
+		  u32 flags,
+		  struct i915_vma_work *work,
+		  struct i915_vma_resource *vma_res);
+
+bool i915_gem_valid_gtt_space(struct i915_vma *vma, unsigned long color);
+bool i915_vma_misplaced(const struct i915_vma *vma,
+			u64 size, u64 alignment, u64 flags);
+void __i915_vma_set_map_and_fenceable(struct i915_vma *vma);
+void i915_vma_revoke_mmap(struct i915_vma *vma);
+void vma_invalidate_tlb(struct i915_address_space *vm, u32 *tlb);
+struct dma_fence *__i915_vma_evict(struct i915_vma *vma, bool async);
+int __i915_vma_unbind(struct i915_vma *vma);
+int __must_check i915_vma_unbind(struct i915_vma *vma);
+int __must_check i915_vma_unbind_async(struct i915_vma *vma, bool trylock_vm);
+int __must_check i915_vma_unbind_unlocked(struct i915_vma *vma);
+void i915_vma_unlink_ctx(struct i915_vma *vma);
+void i915_vma_close(struct i915_vma *vma);
+void i915_vma_reopen(struct i915_vma *vma);
+
+void i915_vma_destroy_locked(struct i915_vma *vma);
+void i915_vma_destroy(struct i915_vma *vma);
+
+#define assert_vma_held(vma) dma_resv_assert_held((vma)->obj->base.resv)
+
+static inline void i915_vma_lock(struct i915_vma *vma)
+{
+	dma_resv_lock(vma->obj->base.resv, NULL);
+}
+
+static inline void i915_vma_unlock(struct i915_vma *vma)
+{
+	dma_resv_unlock(vma->obj->base.resv);
+}
+
+int __must_check
+i915_vma_pin_ww(struct i915_vma *vma, struct i915_gem_ww_ctx *ww,
+		u64 size, u64 alignment, u64 flags);
+
+static inline int __must_check
+i915_vma_pin(struct i915_vma *vma, u64 size, u64 alignment, u64 flags)
+{
+	struct i915_gem_ww_ctx ww;
+	int err;
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	err = i915_gem_object_lock(vma->obj, &ww);
+	if (!err)
+		err = i915_vma_pin_ww(vma, &ww, size, alignment, flags);
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+
+	return err;
+}
+
+int i915_ggtt_pin(struct i915_vma *vma, struct i915_gem_ww_ctx *ww,
+		  u32 align, unsigned int flags);
+
+static inline int i915_vma_pin_count(const struct i915_vma *vma)
+{
+	return atomic_read(&vma->flags) & I915_VMA_PIN_MASK;
+}
+
+static inline bool i915_vma_is_pinned(const struct i915_vma *vma)
+{
+	return i915_vma_pin_count(vma);
+}
+
+static inline void __i915_vma_pin(struct i915_vma *vma)
+{
+	atomic_inc(&vma->flags);
+	GEM_BUG_ON(!i915_vma_is_pinned(vma));
+}
+
+static inline void __i915_vma_unpin(struct i915_vma *vma)
+{
+	GEM_BUG_ON(!i915_vma_is_pinned(vma));
+	atomic_dec(&vma->flags);
+}
+
+static inline void i915_vma_unpin(struct i915_vma *vma)
+{
+	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	__i915_vma_unpin(vma);
+}
+
+static inline bool i915_vma_is_bound(const struct i915_vma *vma,
+				     unsigned int where)
+{
+	return atomic_read(&vma->flags) & where;
+}
+
+static inline bool i915_node_color_differs(const struct drm_mm_node *node,
+					   unsigned long color)
+{
+	return drm_mm_node_allocated(node) && node->color != color;
+}
+
+/**
+ * i915_vma_pin_iomap - calls ioremap_wc to map the GGTT VMA via the aperture
+ * @vma: VMA to iomap
+ *
+ * The passed in VMA has to be pinned in the global GTT mappable region.
+ * An extra pinning of the VMA is acquired for the return iomapping,
+ * the caller must call i915_vma_unpin_iomap to relinquish the pinning
+ * after the iomapping is no longer required.
+ *
+ * Returns a valid iomapped pointer or ERR_PTR.
+ */
+void __iomem *i915_vma_pin_iomap(struct i915_vma *vma);
+
+/**
+ * i915_vma_unpin_iomap - unpins the mapping returned from i915_vma_iomap
+ * @vma: VMA to unpin
+ *
+ * Unpins the previously iomapped VMA from i915_vma_pin_iomap().
+ *
+ * This function is only valid to be called on a VMA previously
+ * iomapped by the caller with i915_vma_pin_iomap().
+ */
+void i915_vma_unpin_iomap(struct i915_vma *vma);
+
+/**
+ * i915_vma_pin_fence - pin fencing state
+ * @vma: vma to pin fencing for
+ *
+ * This pins the fencing state (whether tiled or untiled) to make sure the
+ * vma (and its object) is ready to be used as a scanout target. Fencing
+ * status must be synchronize first by calling i915_vma_get_fence():
+ *
+ * The resulting fence pin reference must be released again with
+ * i915_vma_unpin_fence().
+ *
+ * Returns:
+ *
+ * True if the vma has a fence, false otherwise.
+ */
+int __must_check i915_vma_pin_fence(struct i915_vma *vma);
+void i915_vma_revoke_fence(struct i915_vma *vma);
+
+int __i915_vma_pin_fence(struct i915_vma *vma);
+
+static inline void __i915_vma_unpin_fence(struct i915_vma *vma)
+{
+	GEM_BUG_ON(atomic_read(&vma->fence->pin_count) <= 0);
+	atomic_dec(&vma->fence->pin_count);
+}
+
+/**
+ * i915_vma_unpin_fence - unpin fencing state
+ * @vma: vma to unpin fencing for
+ *
+ * This releases the fence pin reference acquired through
+ * i915_vma_pin_fence. It will handle both objects with and without an
+ * attached fence correctly, callers do not need to distinguish this.
+ */
+static inline void
+i915_vma_unpin_fence(struct i915_vma *vma)
+{
+	if (vma->fence)
+		__i915_vma_unpin_fence(vma);
+}
+
+void i915_vma_parked(struct intel_gt *gt);
+
+static inline bool i915_vma_is_scanout(const struct i915_vma *vma)
+{
+	return test_bit(I915_VMA_SCANOUT_BIT, __i915_vma_flags(vma));
+}
+
+static inline void i915_vma_mark_scanout(struct i915_vma *vma)
+{
+	set_bit(I915_VMA_SCANOUT_BIT, __i915_vma_flags(vma));
+}
+
+static inline void i915_vma_clear_scanout(struct i915_vma *vma)
+{
+	clear_bit(I915_VMA_SCANOUT_BIT, __i915_vma_flags(vma));
+}
+
+#define for_each_until(cond) if (cond) break; else
+
+/**
+ * for_each_ggtt_vma - Iterate over the GGTT VMA belonging to an object.
+ * @V: the #i915_vma iterator
+ * @OBJ: the #drm_i915_gem_object
+ *
+ * GGTT VMA are placed at the being of the object's vma_list, see
+ * vma_create(), so we can stop our walk as soon as we see a ppgtt VMA,
+ * or the list is empty ofc.
+ */
+#define for_each_ggtt_vma(V, OBJ) \
+	list_for_each_entry(V, &(OBJ)->vma.list, obj_link)		\
+		for_each_until(!i915_vma_is_ggtt(V))
+
+struct i915_vma *i915_vma_make_unshrinkable(struct i915_vma *vma);
+void i915_vma_make_shrinkable(struct i915_vma *vma);
+void i915_vma_make_purgeable(struct i915_vma *vma);
+
+int i915_vma_wait_for_bind(struct i915_vma *vma);
+
+static inline int i915_vma_sync(struct i915_vma *vma)
+{
+	/* Wait for the asynchronous bindings and pending GPU reads */
+	return i915_active_wait(&vma->active);
+}
+
+/**
+ * i915_vma_get_current_resource - Get the current resource of the vma
+ * @vma: The vma to get the current resource from.
+ *
+ * It's illegal to call this function if the vma is not bound.
+ *
+ * Return: A refcounted pointer to the current vma resource
+ * of the vma, assuming the vma is bound.
+ */
+static inline struct i915_vma_resource *
+i915_vma_get_current_resource(struct i915_vma *vma)
+{
+	return i915_vma_resource_get(vma->resource);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+void i915_vma_resource_init_from_vma(struct i915_vma_resource *vma_res,
+				     struct i915_vma *vma);
+#endif
+
+void i915_vma_module_exit(void);
+int i915_vma_module_init(void);
+
+I915_SELFTEST_DECLARE(int i915_vma_get_pages(struct i915_vma *vma));
+I915_SELFTEST_DECLARE(void i915_vma_put_pages(struct i915_vma *vma));
+
+#endif
diff --git a/drivers/gpu/drm/i915/i915_vma_resource.c b/drivers/gpu/drm/i915/i915_vma_resource.c
new file mode 100644
index 000000000..de1342dbf
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_vma_resource.c
@@ -0,0 +1,425 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/interval_tree_generic.h>
+#include <linux/sched/mm.h>
+
+#include "i915_sw_fence.h"
+#include "i915_vma_resource.h"
+#include "i915_drv.h"
+#include "intel_memory_region.h"
+
+#include "gt/intel_gtt.h"
+
+static struct kmem_cache *slab_vma_resources;
+
+/**
+ * DOC:
+ * We use a per-vm interval tree to keep track of vma_resources
+ * scheduled for unbind but not yet unbound. The tree is protected by
+ * the vm mutex, and nodes are removed just after the unbind fence signals.
+ * The removal takes the vm mutex from a kernel thread which we need to
+ * keep in mind so that we don't grab the mutex and try to wait for all
+ * pending unbinds to complete, because that will temporaryily block many
+ * of the workqueue threads, and people will get angry.
+ *
+ * We should consider using a single ordered fence per VM instead but that
+ * requires ordering the unbinds and might introduce unnecessary waiting
+ * for unrelated unbinds. Amount of code will probably be roughly the same
+ * due to the simplicity of using the interval tree interface.
+ *
+ * Another drawback of this interval tree is that the complexity of insertion
+ * and removal of fences increases as O(ln(pending_unbinds)) instead of
+ * O(1) for a single fence without interval tree.
+ */
+#define VMA_RES_START(_node) ((_node)->start)
+#define VMA_RES_LAST(_node) ((_node)->start + (_node)->node_size - 1)
+INTERVAL_TREE_DEFINE(struct i915_vma_resource, rb,
+		     u64, __subtree_last,
+		     VMA_RES_START, VMA_RES_LAST, static, vma_res_itree);
+
+/* Callbacks for the unbind dma-fence. */
+
+/**
+ * i915_vma_resource_alloc - Allocate a vma resource
+ *
+ * Return: A pointer to a cleared struct i915_vma_resource or
+ * a -ENOMEM error pointer if allocation fails.
+ */
+struct i915_vma_resource *i915_vma_resource_alloc(void)
+{
+	struct i915_vma_resource *vma_res =
+		kmem_cache_zalloc(slab_vma_resources, GFP_KERNEL);
+
+	return vma_res ? vma_res : ERR_PTR(-ENOMEM);
+}
+
+/**
+ * i915_vma_resource_free - Free a vma resource
+ * @vma_res: The vma resource to free.
+ */
+void i915_vma_resource_free(struct i915_vma_resource *vma_res)
+{
+	if (vma_res)
+		kmem_cache_free(slab_vma_resources, vma_res);
+}
+
+static const char *get_driver_name(struct dma_fence *fence)
+{
+	return "vma unbind fence";
+}
+
+static const char *get_timeline_name(struct dma_fence *fence)
+{
+	return "unbound";
+}
+
+static void unbind_fence_free_rcu(struct rcu_head *head)
+{
+	struct i915_vma_resource *vma_res =
+		container_of(head, typeof(*vma_res), unbind_fence.rcu);
+
+	i915_vma_resource_free(vma_res);
+}
+
+static void unbind_fence_release(struct dma_fence *fence)
+{
+	struct i915_vma_resource *vma_res =
+		container_of(fence, typeof(*vma_res), unbind_fence);
+
+	i915_sw_fence_fini(&vma_res->chain);
+
+	call_rcu(&fence->rcu, unbind_fence_free_rcu);
+}
+
+static struct dma_fence_ops unbind_fence_ops = {
+	.get_driver_name = get_driver_name,
+	.get_timeline_name = get_timeline_name,
+	.release = unbind_fence_release,
+};
+
+static void __i915_vma_resource_unhold(struct i915_vma_resource *vma_res)
+{
+	struct i915_address_space *vm;
+
+	if (!refcount_dec_and_test(&vma_res->hold_count))
+		return;
+
+	dma_fence_signal(&vma_res->unbind_fence);
+
+	vm = vma_res->vm;
+	if (vma_res->wakeref)
+		intel_runtime_pm_put(&vm->i915->runtime_pm, vma_res->wakeref);
+
+	vma_res->vm = NULL;
+	if (!RB_EMPTY_NODE(&vma_res->rb)) {
+		mutex_lock(&vm->mutex);
+		vma_res_itree_remove(vma_res, &vm->pending_unbind);
+		mutex_unlock(&vm->mutex);
+	}
+
+	if (vma_res->bi.pages_rsgt)
+		i915_refct_sgt_put(vma_res->bi.pages_rsgt);
+}
+
+/**
+ * i915_vma_resource_unhold - Unhold the signaling of the vma resource unbind
+ * fence.
+ * @vma_res: The vma resource.
+ * @lockdep_cookie: The lockdep cookie returned from i915_vma_resource_hold.
+ *
+ * The function may leave a dma_fence critical section.
+ */
+void i915_vma_resource_unhold(struct i915_vma_resource *vma_res,
+			      bool lockdep_cookie)
+{
+	dma_fence_end_signalling(lockdep_cookie);
+
+	if (IS_ENABLED(CONFIG_PROVE_LOCKING)) {
+		unsigned long irq_flags;
+
+		/* Inefficient open-coded might_lock_irqsave() */
+		spin_lock_irqsave(&vma_res->lock, irq_flags);
+		spin_unlock_irqrestore(&vma_res->lock, irq_flags);
+	}
+
+	__i915_vma_resource_unhold(vma_res);
+}
+
+/**
+ * i915_vma_resource_hold - Hold the signaling of the vma resource unbind fence.
+ * @vma_res: The vma resource.
+ * @lockdep_cookie: Pointer to a bool serving as a lockdep cooke that should
+ * be given as an argument to the pairing i915_vma_resource_unhold.
+ *
+ * If returning true, the function enters a dma_fence signalling critical
+ * section if not in one already.
+ *
+ * Return: true if holding successful, false if not.
+ */
+bool i915_vma_resource_hold(struct i915_vma_resource *vma_res,
+			    bool *lockdep_cookie)
+{
+	bool held = refcount_inc_not_zero(&vma_res->hold_count);
+
+	if (held)
+		*lockdep_cookie = dma_fence_begin_signalling();
+
+	return held;
+}
+
+static void i915_vma_resource_unbind_work(struct work_struct *work)
+{
+	struct i915_vma_resource *vma_res =
+		container_of(work, typeof(*vma_res), work);
+	struct i915_address_space *vm = vma_res->vm;
+	bool lockdep_cookie;
+
+	lockdep_cookie = dma_fence_begin_signalling();
+	if (likely(!vma_res->skip_pte_rewrite))
+		vma_res->ops->unbind_vma(vm, vma_res);
+
+	dma_fence_end_signalling(lockdep_cookie);
+	__i915_vma_resource_unhold(vma_res);
+	i915_vma_resource_put(vma_res);
+}
+
+static int
+i915_vma_resource_fence_notify(struct i915_sw_fence *fence,
+			       enum i915_sw_fence_notify state)
+{
+	struct i915_vma_resource *vma_res =
+		container_of(fence, typeof(*vma_res), chain);
+	struct dma_fence *unbind_fence =
+		&vma_res->unbind_fence;
+
+	switch (state) {
+	case FENCE_COMPLETE:
+		dma_fence_get(unbind_fence);
+		if (vma_res->immediate_unbind) {
+			i915_vma_resource_unbind_work(&vma_res->work);
+		} else {
+			INIT_WORK(&vma_res->work, i915_vma_resource_unbind_work);
+			queue_work(system_unbound_wq, &vma_res->work);
+		}
+		break;
+	case FENCE_FREE:
+		i915_vma_resource_put(vma_res);
+		break;
+	}
+
+	return NOTIFY_DONE;
+}
+
+/**
+ * i915_vma_resource_unbind - Unbind a vma resource
+ * @vma_res: The vma resource to unbind.
+ * @tlb: pointer to vma->obj->mm.tlb associated with the resource
+ *	 to be stored at vma_res->tlb. When not-NULL, it will be used
+ *	 to do TLB cache invalidation before freeing a VMA resource.
+ *	 Used only for async unbind.
+ *
+ * At this point this function does little more than publish a fence that
+ * signals immediately unless signaling is held back.
+ *
+ * Return: A refcounted pointer to a dma-fence that signals when unbinding is
+ * complete.
+ */
+struct dma_fence *i915_vma_resource_unbind(struct i915_vma_resource *vma_res,
+					   u32 *tlb)
+{
+	struct i915_address_space *vm = vma_res->vm;
+
+	vma_res->tlb = tlb;
+
+	/* Reference for the sw fence */
+	i915_vma_resource_get(vma_res);
+
+	/* Caller must already have a wakeref in this case. */
+	if (vma_res->needs_wakeref)
+		vma_res->wakeref = intel_runtime_pm_get_if_in_use(&vm->i915->runtime_pm);
+
+	if (atomic_read(&vma_res->chain.pending) <= 1) {
+		RB_CLEAR_NODE(&vma_res->rb);
+		vma_res->immediate_unbind = 1;
+	} else {
+		vma_res_itree_insert(vma_res, &vma_res->vm->pending_unbind);
+	}
+
+	i915_sw_fence_commit(&vma_res->chain);
+
+	return &vma_res->unbind_fence;
+}
+
+/**
+ * __i915_vma_resource_init - Initialize a vma resource.
+ * @vma_res: The vma resource to initialize
+ *
+ * Initializes the private members of a vma resource.
+ */
+void __i915_vma_resource_init(struct i915_vma_resource *vma_res)
+{
+	spin_lock_init(&vma_res->lock);
+	dma_fence_init(&vma_res->unbind_fence, &unbind_fence_ops,
+		       &vma_res->lock, 0, 0);
+	refcount_set(&vma_res->hold_count, 1);
+	i915_sw_fence_init(&vma_res->chain, i915_vma_resource_fence_notify);
+}
+
+static void
+i915_vma_resource_color_adjust_range(struct i915_address_space *vm,
+				     u64 *start,
+				     u64 *end)
+{
+	if (i915_vm_has_cache_coloring(vm)) {
+		if (*start)
+			*start -= I915_GTT_PAGE_SIZE;
+		*end += I915_GTT_PAGE_SIZE;
+	}
+}
+
+/**
+ * i915_vma_resource_bind_dep_sync - Wait for / sync all unbinds touching a
+ * certain vm range.
+ * @vm: The vm to look at.
+ * @offset: The range start.
+ * @size: The range size.
+ * @intr: Whether to wait interrubtible.
+ *
+ * The function needs to be called with the vm lock held.
+ *
+ * Return: Zero on success, -ERESTARTSYS if interrupted and @intr==true
+ */
+int i915_vma_resource_bind_dep_sync(struct i915_address_space *vm,
+				    u64 offset,
+				    u64 size,
+				    bool intr)
+{
+	struct i915_vma_resource *node;
+	u64 last = offset + size - 1;
+
+	lockdep_assert_held(&vm->mutex);
+	might_sleep();
+
+	i915_vma_resource_color_adjust_range(vm, &offset, &last);
+	node = vma_res_itree_iter_first(&vm->pending_unbind, offset, last);
+	while (node) {
+		int ret = dma_fence_wait(&node->unbind_fence, intr);
+
+		if (ret)
+			return ret;
+
+		node = vma_res_itree_iter_next(node, offset, last);
+	}
+
+	return 0;
+}
+
+/**
+ * i915_vma_resource_bind_dep_sync_all - Wait for / sync all unbinds of a vm,
+ * releasing the vm lock while waiting.
+ * @vm: The vm to look at.
+ *
+ * The function may not be called with the vm lock held.
+ * Typically this is called at vm destruction to finish any pending
+ * unbind operations. The vm mutex is released while waiting to avoid
+ * stalling kernel workqueues trying to grab the mutex.
+ */
+void i915_vma_resource_bind_dep_sync_all(struct i915_address_space *vm)
+{
+	struct i915_vma_resource *node;
+	struct dma_fence *fence;
+
+	do {
+		fence = NULL;
+		mutex_lock(&vm->mutex);
+		node = vma_res_itree_iter_first(&vm->pending_unbind, 0,
+						U64_MAX);
+		if (node)
+			fence = dma_fence_get_rcu(&node->unbind_fence);
+		mutex_unlock(&vm->mutex);
+
+		if (fence) {
+			/*
+			 * The wait makes sure the node eventually removes
+			 * itself from the tree.
+			 */
+			dma_fence_wait(fence, false);
+			dma_fence_put(fence);
+		}
+	} while (node);
+}
+
+/**
+ * i915_vma_resource_bind_dep_await - Have a struct i915_sw_fence await all
+ * pending unbinds in a certain range of a vm.
+ * @vm: The vm to look at.
+ * @sw_fence: The struct i915_sw_fence that will be awaiting the unbinds.
+ * @offset: The range start.
+ * @size: The range size.
+ * @intr: Whether to wait interrubtible.
+ * @gfp: Allocation mode for memory allocations.
+ *
+ * The function makes @sw_fence await all pending unbinds in a certain
+ * vm range before calling the complete notifier. To be able to await
+ * each individual unbind, the function needs to allocate memory using
+ * the @gpf allocation mode. If that fails, the function will instead
+ * wait for the unbind fence to signal, using @intr to judge whether to
+ * wait interruptible or not. Note that @gfp should ideally be selected so
+ * as to avoid any expensive memory allocation stalls and rather fail and
+ * synchronize itself. For now the vm mutex is required when calling this
+ * function with means that @gfp can't call into direct reclaim. In reality
+ * this means that during heavy memory pressure, we will sync in this
+ * function.
+ *
+ * Return: Zero on success, -ERESTARTSYS if interrupted and @intr==true
+ */
+int i915_vma_resource_bind_dep_await(struct i915_address_space *vm,
+				     struct i915_sw_fence *sw_fence,
+				     u64 offset,
+				     u64 size,
+				     bool intr,
+				     gfp_t gfp)
+{
+	struct i915_vma_resource *node;
+	u64 last = offset + size - 1;
+
+	lockdep_assert_held(&vm->mutex);
+	might_alloc(gfp);
+	might_sleep();
+
+	i915_vma_resource_color_adjust_range(vm, &offset, &last);
+	node = vma_res_itree_iter_first(&vm->pending_unbind, offset, last);
+	while (node) {
+		int ret;
+
+		ret = i915_sw_fence_await_dma_fence(sw_fence,
+						    &node->unbind_fence,
+						    0, gfp);
+		if (ret < 0) {
+			ret = dma_fence_wait(&node->unbind_fence, intr);
+			if (ret)
+				return ret;
+		}
+
+		node = vma_res_itree_iter_next(node, offset, last);
+	}
+
+	return 0;
+}
+
+void i915_vma_resource_module_exit(void)
+{
+	kmem_cache_destroy(slab_vma_resources);
+}
+
+int __init i915_vma_resource_module_init(void)
+{
+	slab_vma_resources = KMEM_CACHE(i915_vma_resource, SLAB_HWCACHE_ALIGN);
+	if (!slab_vma_resources)
+		return -ENOMEM;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_vma_resource.h b/drivers/gpu/drm/i915/i915_vma_resource.h
new file mode 100644
index 000000000..06923d181
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_vma_resource.h
@@ -0,0 +1,244 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __I915_VMA_RESOURCE_H__
+#define __I915_VMA_RESOURCE_H__
+
+#include <linux/dma-fence.h>
+#include <linux/refcount.h>
+
+#include "i915_gem.h"
+#include "i915_scatterlist.h"
+#include "i915_sw_fence.h"
+#include "intel_runtime_pm.h"
+
+struct intel_memory_region;
+
+struct i915_page_sizes {
+	/**
+	 * The sg mask of the pages sg_table. i.e the mask of
+	 * the lengths for each sg entry.
+	 */
+	unsigned int phys;
+
+	/**
+	 * The gtt page sizes we are allowed to use given the
+	 * sg mask and the supported page sizes. This will
+	 * express the smallest unit we can use for the whole
+	 * object, as well as the larger sizes we may be able
+	 * to use opportunistically.
+	 */
+	unsigned int sg;
+};
+
+/**
+ * struct i915_vma_resource - Snapshotted unbind information.
+ * @unbind_fence: Fence to mark unbinding complete. Note that this fence
+ * is not considered published until unbind is scheduled, and as such it
+ * is illegal to access this fence before scheduled unbind other than
+ * for refcounting.
+ * @lock: The @unbind_fence lock.
+ * @hold_count: Number of holders blocking the fence from finishing.
+ * The vma itself is keeping a hold, which is released when unbind
+ * is scheduled.
+ * @work: Work struct for deferred unbind work.
+ * @chain: Pointer to struct i915_sw_fence used to await dependencies.
+ * @rb: Rb node for the vm's pending unbind interval tree.
+ * @__subtree_last: Interval tree private member.
+ * @vm: non-refcounted pointer to the vm. This is for internal use only and
+ * this member is cleared after vm_resource unbind.
+ * @mr: The memory region of the object pointed to by the vma.
+ * @ops: Pointer to the backend i915_vma_ops.
+ * @private: Bind backend private info.
+ * @start: Offset into the address space of bind range start.
+ * @node_size: Size of the allocated range manager node.
+ * @vma_size: Bind size.
+ * @page_sizes_gtt: Resulting page sizes from the bind operation.
+ * @bound_flags: Flags indicating binding status.
+ * @allocated: Backend private data. TODO: Should move into @private.
+ * @immediate_unbind: Unbind can be done immediately and doesn't need to be
+ * deferred to a work item awaiting unsignaled fences. This is a hack.
+ * (dma_fence_work uses a fence flag for this, but this seems slightly
+ * cleaner).
+ * @needs_wakeref: Whether a wakeref is needed during unbind. Since we can't
+ * take a wakeref in the dma-fence signalling critical path, it needs to be
+ * taken when the unbind is scheduled.
+ * @skip_pte_rewrite: During ggtt suspend and vm takedown pte rewriting
+ * needs to be skipped for unbind.
+ * @tlb: pointer for obj->mm.tlb, if async unbind. Otherwise, NULL
+ *
+ * The lifetime of a struct i915_vma_resource is from a binding request to
+ * the actual possible asynchronous unbind has completed.
+ */
+struct i915_vma_resource {
+	struct dma_fence unbind_fence;
+	/* See above for description of the lock. */
+	spinlock_t lock;
+	refcount_t hold_count;
+	struct work_struct work;
+	struct i915_sw_fence chain;
+	struct rb_node rb;
+	u64 __subtree_last;
+	struct i915_address_space *vm;
+	intel_wakeref_t wakeref;
+
+	/**
+	 * struct i915_vma_bindinfo - Information needed for async bind
+	 * only but that can be dropped after the bind has taken place.
+	 * Consider making this a separate argument to the bind_vma
+	 * op, coalescing with other arguments like vm, stash, cache_level
+	 * and flags
+	 * @pages: The pages sg-table.
+	 * @page_sizes: Page sizes of the pages.
+	 * @pages_rsgt: Refcounted sg-table when delayed object destruction
+	 * is supported. May be NULL.
+	 * @readonly: Whether the vma should be bound read-only.
+	 * @lmem: Whether the vma points to lmem.
+	 */
+	struct i915_vma_bindinfo {
+		struct sg_table *pages;
+		struct i915_page_sizes page_sizes;
+		struct i915_refct_sgt *pages_rsgt;
+		bool readonly:1;
+		bool lmem:1;
+	} bi;
+
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+	struct intel_memory_region *mr;
+#endif
+	const struct i915_vma_ops *ops;
+	void *private;
+	u64 start;
+	u64 node_size;
+	u64 vma_size;
+	u32 page_sizes_gtt;
+
+	u32 bound_flags;
+	bool allocated:1;
+	bool immediate_unbind:1;
+	bool needs_wakeref:1;
+	bool skip_pte_rewrite:1;
+
+	u32 *tlb;
+};
+
+bool i915_vma_resource_hold(struct i915_vma_resource *vma_res,
+			    bool *lockdep_cookie);
+
+void i915_vma_resource_unhold(struct i915_vma_resource *vma_res,
+			      bool lockdep_cookie);
+
+struct i915_vma_resource *i915_vma_resource_alloc(void);
+
+void i915_vma_resource_free(struct i915_vma_resource *vma_res);
+
+struct dma_fence *i915_vma_resource_unbind(struct i915_vma_resource *vma_res,
+					   u32 *tlb);
+
+void __i915_vma_resource_init(struct i915_vma_resource *vma_res);
+
+/**
+ * i915_vma_resource_get - Take a reference on a vma resource
+ * @vma_res: The vma resource on which to take a reference.
+ *
+ * Return: The @vma_res pointer
+ */
+static inline struct i915_vma_resource
+*i915_vma_resource_get(struct i915_vma_resource *vma_res)
+{
+	dma_fence_get(&vma_res->unbind_fence);
+	return vma_res;
+}
+
+/**
+ * i915_vma_resource_put - Release a reference to a struct i915_vma_resource
+ * @vma_res: The resource
+ */
+static inline void i915_vma_resource_put(struct i915_vma_resource *vma_res)
+{
+	dma_fence_put(&vma_res->unbind_fence);
+}
+
+/**
+ * i915_vma_resource_init - Initialize a vma resource.
+ * @vma_res: The vma resource to initialize
+ * @vm: Pointer to the vm.
+ * @pages: The pages sg-table.
+ * @page_sizes: Page sizes of the pages.
+ * @pages_rsgt: Pointer to a struct i915_refct_sgt of an object with
+ * delayed destruction.
+ * @readonly: Whether the vma should be bound read-only.
+ * @lmem: Whether the vma points to lmem.
+ * @mr: The memory region of the object the vma points to.
+ * @ops: The backend ops.
+ * @private: Bind backend private info.
+ * @start: Offset into the address space of bind range start.
+ * @node_size: Size of the allocated range manager node.
+ * @size: Bind size.
+ *
+ * Initializes a vma resource allocated using i915_vma_resource_alloc().
+ * The reason for having separate allocate and initialize function is that
+ * initialization may need to be performed from under a lock where
+ * allocation is not allowed.
+ */
+static inline void i915_vma_resource_init(struct i915_vma_resource *vma_res,
+					  struct i915_address_space *vm,
+					  struct sg_table *pages,
+					  const struct i915_page_sizes *page_sizes,
+					  struct i915_refct_sgt *pages_rsgt,
+					  bool readonly,
+					  bool lmem,
+					  struct intel_memory_region *mr,
+					  const struct i915_vma_ops *ops,
+					  void *private,
+					  u64 start,
+					  u64 node_size,
+					  u64 size)
+{
+	__i915_vma_resource_init(vma_res);
+	vma_res->vm = vm;
+	vma_res->bi.pages = pages;
+	vma_res->bi.page_sizes = *page_sizes;
+	if (pages_rsgt)
+		vma_res->bi.pages_rsgt = i915_refct_sgt_get(pages_rsgt);
+	vma_res->bi.readonly = readonly;
+	vma_res->bi.lmem = lmem;
+#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
+	vma_res->mr = mr;
+#endif
+	vma_res->ops = ops;
+	vma_res->private = private;
+	vma_res->start = start;
+	vma_res->node_size = node_size;
+	vma_res->vma_size = size;
+}
+
+static inline void i915_vma_resource_fini(struct i915_vma_resource *vma_res)
+{
+	GEM_BUG_ON(refcount_read(&vma_res->hold_count) != 1);
+	if (vma_res->bi.pages_rsgt)
+		i915_refct_sgt_put(vma_res->bi.pages_rsgt);
+	i915_sw_fence_fini(&vma_res->chain);
+}
+
+int i915_vma_resource_bind_dep_sync(struct i915_address_space *vm,
+				    u64 first,
+				    u64 last,
+				    bool intr);
+
+int i915_vma_resource_bind_dep_await(struct i915_address_space *vm,
+				     struct i915_sw_fence *sw_fence,
+				     u64 first,
+				     u64 last,
+				     bool intr,
+				     gfp_t gfp);
+
+void i915_vma_resource_bind_dep_sync_all(struct i915_address_space *vm);
+
+void i915_vma_resource_module_exit(void);
+
+int i915_vma_resource_module_init(void);
+
+#endif
diff --git a/drivers/gpu/drm/i915/i915_vma_types.h b/drivers/gpu/drm/i915/i915_vma_types.h
new file mode 100644
index 000000000..ec0f6c9f5
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_vma_types.h
@@ -0,0 +1,305 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __I915_VMA_TYPES_H__
+#define __I915_VMA_TYPES_H__
+
+#include <linux/rbtree.h>
+
+#include <drm/drm_mm.h>
+
+#include "gem/i915_gem_object_types.h"
+
+enum i915_cache_level;
+
+/**
+ * DOC: Global GTT views
+ *
+ * Background and previous state
+ *
+ * Historically objects could exists (be bound) in global GTT space only as
+ * singular instances with a view representing all of the object's backing pages
+ * in a linear fashion. This view will be called a normal view.
+ *
+ * To support multiple views of the same object, where the number of mapped
+ * pages is not equal to the backing store, or where the layout of the pages
+ * is not linear, concept of a GGTT view was added.
+ *
+ * One example of an alternative view is a stereo display driven by a single
+ * image. In this case we would have a framebuffer looking like this
+ * (2x2 pages):
+ *
+ *    12
+ *    34
+ *
+ * Above would represent a normal GGTT view as normally mapped for GPU or CPU
+ * rendering. In contrast, fed to the display engine would be an alternative
+ * view which could look something like this:
+ *
+ *   1212
+ *   3434
+ *
+ * In this example both the size and layout of pages in the alternative view is
+ * different from the normal view.
+ *
+ * Implementation and usage
+ *
+ * GGTT views are implemented using VMAs and are distinguished via enum
+ * i915_gtt_view_type and struct i915_gtt_view.
+ *
+ * A new flavour of core GEM functions which work with GGTT bound objects were
+ * added with the _ggtt_ infix, and sometimes with _view postfix to avoid
+ * renaming  in large amounts of code. They take the struct i915_gtt_view
+ * parameter encapsulating all metadata required to implement a view.
+ *
+ * As a helper for callers which are only interested in the normal view,
+ * globally const i915_gtt_view_normal singleton instance exists. All old core
+ * GEM API functions, the ones not taking the view parameter, are operating on,
+ * or with the normal GGTT view.
+ *
+ * Code wanting to add or use a new GGTT view needs to:
+ *
+ * 1. Add a new enum with a suitable name.
+ * 2. Extend the metadata in the i915_gtt_view structure if required.
+ * 3. Add support to i915_get_vma_pages().
+ *
+ * New views are required to build a scatter-gather table from within the
+ * i915_get_vma_pages function. This table is stored in the vma.gtt_view and
+ * exists for the lifetime of an VMA.
+ *
+ * Core API is designed to have copy semantics which means that passed in
+ * struct i915_gtt_view does not need to be persistent (left around after
+ * calling the core API functions).
+ *
+ */
+
+struct i915_vma_resource;
+
+struct intel_remapped_plane_info {
+	/* in gtt pages */
+	u32 offset:31;
+	u32 linear:1;
+	union {
+		/* in gtt pages for !linear */
+		struct {
+			u16 width;
+			u16 height;
+			u16 src_stride;
+			u16 dst_stride;
+		};
+
+		/* in gtt pages for linear */
+		u32 size;
+	};
+} __packed;
+
+struct intel_remapped_info {
+	struct intel_remapped_plane_info plane[4];
+	/* in gtt pages */
+	u32 plane_alignment;
+} __packed;
+
+struct intel_rotation_info {
+	struct intel_remapped_plane_info plane[2];
+} __packed;
+
+struct intel_partial_info {
+	u64 offset;
+	unsigned int size;
+} __packed;
+
+enum i915_gtt_view_type {
+	I915_GTT_VIEW_NORMAL = 0,
+	I915_GTT_VIEW_ROTATED = sizeof(struct intel_rotation_info),
+	I915_GTT_VIEW_PARTIAL = sizeof(struct intel_partial_info),
+	I915_GTT_VIEW_REMAPPED = sizeof(struct intel_remapped_info),
+};
+
+static inline void assert_i915_gem_gtt_types(void)
+{
+	BUILD_BUG_ON(sizeof(struct intel_rotation_info) != 2 * sizeof(u32) + 8 * sizeof(u16));
+	BUILD_BUG_ON(sizeof(struct intel_partial_info) != sizeof(u64) + sizeof(unsigned int));
+	BUILD_BUG_ON(sizeof(struct intel_remapped_info) != 5 * sizeof(u32) + 16 * sizeof(u16));
+
+	/* Check that rotation/remapped shares offsets for simplicity */
+	BUILD_BUG_ON(offsetof(struct intel_remapped_info, plane[0]) !=
+		     offsetof(struct intel_rotation_info, plane[0]));
+	BUILD_BUG_ON(offsetofend(struct intel_remapped_info, plane[1]) !=
+		     offsetofend(struct intel_rotation_info, plane[1]));
+
+	/* As we encode the size of each branch inside the union into its type,
+	 * we have to be careful that each branch has a unique size.
+	 */
+	switch ((enum i915_gtt_view_type)0) {
+	case I915_GTT_VIEW_NORMAL:
+	case I915_GTT_VIEW_PARTIAL:
+	case I915_GTT_VIEW_ROTATED:
+	case I915_GTT_VIEW_REMAPPED:
+		/* gcc complains if these are identical cases */
+		break;
+	}
+}
+
+struct i915_gtt_view {
+	enum i915_gtt_view_type type;
+	union {
+		/* Members need to contain no holes/padding */
+		struct intel_partial_info partial;
+		struct intel_rotation_info rotated;
+		struct intel_remapped_info remapped;
+	};
+};
+
+/**
+ * DOC: Virtual Memory Address
+ *
+ * A VMA represents a GEM BO that is bound into an address space. Therefore, a
+ * VMA's presence cannot be guaranteed before binding, or after unbinding the
+ * object into/from the address space.
+ *
+ * To make things as simple as possible (ie. no refcounting), a VMA's lifetime
+ * will always be <= an objects lifetime. So object refcounting should cover us.
+ */
+struct i915_vma {
+	struct drm_mm_node node;
+
+	struct i915_address_space *vm;
+	const struct i915_vma_ops *ops;
+
+	struct drm_i915_gem_object *obj;
+
+	struct sg_table *pages;
+	void __iomem *iomap;
+	void *private; /* owned by creator */
+
+	struct i915_fence_reg *fence;
+
+	u64 size;
+	u64 display_alignment;
+	struct i915_page_sizes page_sizes;
+
+	/* mmap-offset associated with fencing for this vma */
+	struct i915_mmap_offset	*mmo;
+
+	u32 fence_size;
+	u32 fence_alignment;
+
+	/**
+	 * Count of the number of times this vma has been opened by different
+	 * handles (but same file) for execbuf, i.e. the number of aliases
+	 * that exist in the ctx->handle_vmas LUT for this vma.
+	 */
+	atomic_t open_count;
+	atomic_t flags;
+	/**
+	 * How many users have pinned this object in GTT space.
+	 *
+	 * This is a tightly bound, fairly small number of users, so we
+	 * stuff inside the flags field so that we can both check for overflow
+	 * and detect a no-op i915_vma_pin() in a single check, while also
+	 * pinning the vma.
+	 *
+	 * The worst case display setup would have the same vma pinned for
+	 * use on each plane on each crtc, while also building the next atomic
+	 * state and holding a pin for the length of the cleanup queue. In the
+	 * future, the flip queue may be increased from 1.
+	 * Estimated worst case: 3 [qlen] * 4 [max crtcs] * 7 [max planes] = 84
+	 *
+	 * For GEM, the number of concurrent users for pwrite/pread is
+	 * unbounded. For execbuffer, it is currently one but will in future
+	 * be extended to allow multiple clients to pin vma concurrently.
+	 *
+	 * We also use suballocated pages, with each suballocation claiming
+	 * its own pin on the shared vma. At present, this is limited to
+	 * exclusive cachelines of a single page, so a maximum of 64 possible
+	 * users.
+	 */
+#define I915_VMA_PIN_MASK 0x3ff
+#define I915_VMA_OVERFLOW 0x200
+
+	/** Flags and address space this VMA is bound to */
+#define I915_VMA_GLOBAL_BIND_BIT 10
+#define I915_VMA_LOCAL_BIND_BIT  11
+
+#define I915_VMA_GLOBAL_BIND	((int)BIT(I915_VMA_GLOBAL_BIND_BIT))
+#define I915_VMA_LOCAL_BIND	((int)BIT(I915_VMA_LOCAL_BIND_BIT))
+
+#define I915_VMA_BIND_MASK (I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND)
+
+#define I915_VMA_ERROR_BIT	12
+#define I915_VMA_ERROR		((int)BIT(I915_VMA_ERROR_BIT))
+
+#define I915_VMA_GGTT_BIT	13
+#define I915_VMA_CAN_FENCE_BIT	14
+#define I915_VMA_USERFAULT_BIT	15
+#define I915_VMA_GGTT_WRITE_BIT	16
+
+#define I915_VMA_GGTT		((int)BIT(I915_VMA_GGTT_BIT))
+#define I915_VMA_CAN_FENCE	((int)BIT(I915_VMA_CAN_FENCE_BIT))
+#define I915_VMA_USERFAULT	((int)BIT(I915_VMA_USERFAULT_BIT))
+#define I915_VMA_GGTT_WRITE	((int)BIT(I915_VMA_GGTT_WRITE_BIT))
+
+#define I915_VMA_SCANOUT_BIT	17
+#define I915_VMA_SCANOUT	((int)BIT(I915_VMA_SCANOUT_BIT))
+
+	struct i915_active active;
+
+#define I915_VMA_PAGES_BIAS 24
+#define I915_VMA_PAGES_ACTIVE (BIT(24) | 1)
+	atomic_t pages_count; /* number of active binds to the pages */
+
+	/**
+	 * Whether we hold a reference on the vm dma_resv lock to temporarily
+	 * block vm freeing until the vma is destroyed.
+	 * Protected by the vm mutex.
+	 */
+	bool vm_ddestroy;
+
+	/**
+	 * Support different GGTT views into the same object.
+	 * This means there can be multiple VMA mappings per object and per VM.
+	 * i915_gtt_view_type is used to distinguish between those entries.
+	 * The default one of zero (I915_GTT_VIEW_NORMAL) is default and also
+	 * assumed in GEM functions which take no ggtt view parameter.
+	 */
+	struct i915_gtt_view gtt_view;
+
+	/** This object's place on the active/inactive lists */
+	struct list_head vm_link;
+
+	struct list_head obj_link; /* Link in the object's VMA list */
+	struct rb_node obj_node;
+	struct hlist_node obj_hash;
+
+	/** This vma's place in the eviction list */
+	struct list_head evict_link;
+
+	struct list_head closed_link;
+
+	/** The async vma resource. Protected by the vm_mutex */
+	struct i915_vma_resource *resource;
+};
+
+#endif
diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c
new file mode 100644
index 000000000..20575eb77
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_device_info.c
@@ -0,0 +1,453 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/string_helpers.h>
+
+#include <drm/drm_print.h>
+#include <drm/i915_pciids.h>
+
+#include "display/intel_cdclk.h"
+#include "display/intel_de.h"
+#include "intel_device_info.h"
+#include "i915_drv.h"
+#include "i915_utils.h"
+
+#define PLATFORM_NAME(x) [INTEL_##x] = #x
+static const char * const platform_names[] = {
+	PLATFORM_NAME(I830),
+	PLATFORM_NAME(I845G),
+	PLATFORM_NAME(I85X),
+	PLATFORM_NAME(I865G),
+	PLATFORM_NAME(I915G),
+	PLATFORM_NAME(I915GM),
+	PLATFORM_NAME(I945G),
+	PLATFORM_NAME(I945GM),
+	PLATFORM_NAME(G33),
+	PLATFORM_NAME(PINEVIEW),
+	PLATFORM_NAME(I965G),
+	PLATFORM_NAME(I965GM),
+	PLATFORM_NAME(G45),
+	PLATFORM_NAME(GM45),
+	PLATFORM_NAME(IRONLAKE),
+	PLATFORM_NAME(SANDYBRIDGE),
+	PLATFORM_NAME(IVYBRIDGE),
+	PLATFORM_NAME(VALLEYVIEW),
+	PLATFORM_NAME(HASWELL),
+	PLATFORM_NAME(BROADWELL),
+	PLATFORM_NAME(CHERRYVIEW),
+	PLATFORM_NAME(SKYLAKE),
+	PLATFORM_NAME(BROXTON),
+	PLATFORM_NAME(KABYLAKE),
+	PLATFORM_NAME(GEMINILAKE),
+	PLATFORM_NAME(COFFEELAKE),
+	PLATFORM_NAME(COMETLAKE),
+	PLATFORM_NAME(ICELAKE),
+	PLATFORM_NAME(ELKHARTLAKE),
+	PLATFORM_NAME(JASPERLAKE),
+	PLATFORM_NAME(TIGERLAKE),
+	PLATFORM_NAME(ROCKETLAKE),
+	PLATFORM_NAME(DG1),
+	PLATFORM_NAME(ALDERLAKE_S),
+	PLATFORM_NAME(ALDERLAKE_P),
+	PLATFORM_NAME(XEHPSDV),
+	PLATFORM_NAME(DG2),
+	PLATFORM_NAME(PONTEVECCHIO),
+	PLATFORM_NAME(METEORLAKE),
+};
+#undef PLATFORM_NAME
+
+const char *intel_platform_name(enum intel_platform platform)
+{
+	BUILD_BUG_ON(ARRAY_SIZE(platform_names) != INTEL_MAX_PLATFORMS);
+
+	if (WARN_ON_ONCE(platform >= ARRAY_SIZE(platform_names) ||
+			 platform_names[platform] == NULL))
+		return "<unknown>";
+
+	return platform_names[platform];
+}
+
+void intel_device_info_print(const struct intel_device_info *info,
+			     const struct intel_runtime_info *runtime,
+			     struct drm_printer *p)
+{
+	if (runtime->graphics.ip.rel)
+		drm_printf(p, "graphics version: %u.%02u\n",
+			   runtime->graphics.ip.ver,
+			   runtime->graphics.ip.rel);
+	else
+		drm_printf(p, "graphics version: %u\n",
+			   runtime->graphics.ip.ver);
+
+	if (runtime->media.ip.rel)
+		drm_printf(p, "media version: %u.%02u\n",
+			   runtime->media.ip.ver,
+			   runtime->media.ip.rel);
+	else
+		drm_printf(p, "media version: %u\n",
+			   runtime->media.ip.ver);
+
+	if (runtime->display.ip.rel)
+		drm_printf(p, "display version: %u.%02u\n",
+			   runtime->display.ip.ver,
+			   runtime->display.ip.rel);
+	else
+		drm_printf(p, "display version: %u\n",
+			   runtime->display.ip.ver);
+
+	drm_printf(p, "gt: %d\n", info->gt);
+	drm_printf(p, "memory-regions: %x\n", runtime->memory_regions);
+	drm_printf(p, "page-sizes: %x\n", runtime->page_sizes);
+	drm_printf(p, "platform: %s\n", intel_platform_name(info->platform));
+	drm_printf(p, "ppgtt-size: %d\n", runtime->ppgtt_size);
+	drm_printf(p, "ppgtt-type: %d\n", runtime->ppgtt_type);
+	drm_printf(p, "dma_mask_size: %u\n", info->dma_mask_size);
+
+#define PRINT_FLAG(name) drm_printf(p, "%s: %s\n", #name, str_yes_no(info->name))
+	DEV_INFO_FOR_EACH_FLAG(PRINT_FLAG);
+#undef PRINT_FLAG
+
+	drm_printf(p, "has_pooled_eu: %s\n", str_yes_no(runtime->has_pooled_eu));
+
+#define PRINT_FLAG(name) drm_printf(p, "%s: %s\n", #name, str_yes_no(info->display.name))
+	DEV_INFO_DISPLAY_FOR_EACH_FLAG(PRINT_FLAG);
+#undef PRINT_FLAG
+
+	drm_printf(p, "has_hdcp: %s\n", str_yes_no(runtime->has_hdcp));
+	drm_printf(p, "has_dmc: %s\n", str_yes_no(runtime->has_dmc));
+	drm_printf(p, "has_dsc: %s\n", str_yes_no(runtime->has_dsc));
+
+	drm_printf(p, "rawclk rate: %u kHz\n", runtime->rawclk_freq);
+}
+
+#undef INTEL_VGA_DEVICE
+#define INTEL_VGA_DEVICE(id, info) (id)
+
+static const u16 subplatform_ult_ids[] = {
+	INTEL_HSW_ULT_GT1_IDS(0),
+	INTEL_HSW_ULT_GT2_IDS(0),
+	INTEL_HSW_ULT_GT3_IDS(0),
+	INTEL_BDW_ULT_GT1_IDS(0),
+	INTEL_BDW_ULT_GT2_IDS(0),
+	INTEL_BDW_ULT_GT3_IDS(0),
+	INTEL_BDW_ULT_RSVD_IDS(0),
+	INTEL_SKL_ULT_GT1_IDS(0),
+	INTEL_SKL_ULT_GT2_IDS(0),
+	INTEL_SKL_ULT_GT3_IDS(0),
+	INTEL_KBL_ULT_GT1_IDS(0),
+	INTEL_KBL_ULT_GT2_IDS(0),
+	INTEL_KBL_ULT_GT3_IDS(0),
+	INTEL_CFL_U_GT2_IDS(0),
+	INTEL_CFL_U_GT3_IDS(0),
+	INTEL_WHL_U_GT1_IDS(0),
+	INTEL_WHL_U_GT2_IDS(0),
+	INTEL_WHL_U_GT3_IDS(0),
+	INTEL_CML_U_GT1_IDS(0),
+	INTEL_CML_U_GT2_IDS(0),
+};
+
+static const u16 subplatform_ulx_ids[] = {
+	INTEL_HSW_ULX_GT1_IDS(0),
+	INTEL_HSW_ULX_GT2_IDS(0),
+	INTEL_BDW_ULX_GT1_IDS(0),
+	INTEL_BDW_ULX_GT2_IDS(0),
+	INTEL_BDW_ULX_GT3_IDS(0),
+	INTEL_BDW_ULX_RSVD_IDS(0),
+	INTEL_SKL_ULX_GT1_IDS(0),
+	INTEL_SKL_ULX_GT2_IDS(0),
+	INTEL_KBL_ULX_GT1_IDS(0),
+	INTEL_KBL_ULX_GT2_IDS(0),
+	INTEL_AML_KBL_GT2_IDS(0),
+	INTEL_AML_CFL_GT2_IDS(0),
+};
+
+static const u16 subplatform_portf_ids[] = {
+	INTEL_ICL_PORT_F_IDS(0),
+};
+
+static const u16 subplatform_uy_ids[] = {
+	INTEL_TGL_12_GT2_IDS(0),
+};
+
+static const u16 subplatform_n_ids[] = {
+	INTEL_ADLN_IDS(0),
+};
+
+static const u16 subplatform_rpl_ids[] = {
+	INTEL_RPLS_IDS(0),
+	INTEL_RPLP_IDS(0),
+};
+
+static const u16 subplatform_g10_ids[] = {
+	INTEL_DG2_G10_IDS(0),
+	INTEL_ATS_M150_IDS(0),
+};
+
+static const u16 subplatform_g11_ids[] = {
+	INTEL_DG2_G11_IDS(0),
+	INTEL_ATS_M75_IDS(0),
+};
+
+static const u16 subplatform_g12_ids[] = {
+	INTEL_DG2_G12_IDS(0),
+};
+
+static const u16 subplatform_m_ids[] = {
+	INTEL_MTL_M_IDS(0),
+};
+
+static const u16 subplatform_p_ids[] = {
+	INTEL_MTL_P_IDS(0),
+};
+
+static bool find_devid(u16 id, const u16 *p, unsigned int num)
+{
+	for (; num; num--, p++) {
+		if (*p == id)
+			return true;
+	}
+
+	return false;
+}
+
+void intel_device_info_subplatform_init(struct drm_i915_private *i915)
+{
+	const struct intel_device_info *info = INTEL_INFO(i915);
+	const struct intel_runtime_info *rinfo = RUNTIME_INFO(i915);
+	const unsigned int pi = __platform_mask_index(rinfo, info->platform);
+	const unsigned int pb = __platform_mask_bit(rinfo, info->platform);
+	u16 devid = INTEL_DEVID(i915);
+	u32 mask = 0;
+
+	/* Make sure IS_<platform> checks are working. */
+	RUNTIME_INFO(i915)->platform_mask[pi] = BIT(pb);
+
+	/* Find and mark subplatform bits based on the PCI device id. */
+	if (find_devid(devid, subplatform_ult_ids,
+		       ARRAY_SIZE(subplatform_ult_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_ULT);
+	} else if (find_devid(devid, subplatform_ulx_ids,
+			      ARRAY_SIZE(subplatform_ulx_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_ULX);
+		if (IS_HASWELL(i915) || IS_BROADWELL(i915)) {
+			/* ULX machines are also considered ULT. */
+			mask |= BIT(INTEL_SUBPLATFORM_ULT);
+		}
+	} else if (find_devid(devid, subplatform_portf_ids,
+			      ARRAY_SIZE(subplatform_portf_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_PORTF);
+	} else if (find_devid(devid, subplatform_uy_ids,
+			   ARRAY_SIZE(subplatform_uy_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_UY);
+	} else if (find_devid(devid, subplatform_n_ids,
+				ARRAY_SIZE(subplatform_n_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_N);
+	} else if (find_devid(devid, subplatform_rpl_ids,
+			      ARRAY_SIZE(subplatform_rpl_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_RPL);
+	} else if (find_devid(devid, subplatform_g10_ids,
+			      ARRAY_SIZE(subplatform_g10_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_G10);
+	} else if (find_devid(devid, subplatform_g11_ids,
+			      ARRAY_SIZE(subplatform_g11_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_G11);
+	} else if (find_devid(devid, subplatform_g12_ids,
+			      ARRAY_SIZE(subplatform_g12_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_G12);
+	} else if (find_devid(devid, subplatform_m_ids,
+			      ARRAY_SIZE(subplatform_m_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_M);
+	} else if (find_devid(devid, subplatform_p_ids,
+			      ARRAY_SIZE(subplatform_p_ids))) {
+		mask = BIT(INTEL_SUBPLATFORM_P);
+	}
+
+	GEM_BUG_ON(mask & ~INTEL_SUBPLATFORM_MASK);
+
+	RUNTIME_INFO(i915)->platform_mask[pi] |= mask;
+}
+
+/**
+ * intel_device_info_runtime_init - initialize runtime info
+ * @dev_priv: the i915 device
+ *
+ * Determine various intel_device_info fields at runtime.
+ *
+ * Use it when either:
+ *   - it's judged too laborious to fill n static structures with the limit
+ *     when a simple if statement does the job,
+ *   - run-time checks (eg read fuse/strap registers) are needed.
+ *
+ * This function needs to be called:
+ *   - after the MMIO has been setup as we are reading registers,
+ *   - after the PCH has been detected,
+ *   - before the first usage of the fields it can tweak.
+ */
+void intel_device_info_runtime_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_device_info *info = mkwrite_device_info(dev_priv);
+	struct intel_runtime_info *runtime = RUNTIME_INFO(dev_priv);
+	enum pipe pipe;
+
+	/* Wa_14011765242: adl-s A0,A1 */
+	if (IS_ADLS_DISPLAY_STEP(dev_priv, STEP_A0, STEP_A2))
+		for_each_pipe(dev_priv, pipe)
+			runtime->num_scalers[pipe] = 0;
+	else if (DISPLAY_VER(dev_priv) >= 11) {
+		for_each_pipe(dev_priv, pipe)
+			runtime->num_scalers[pipe] = 2;
+	} else if (DISPLAY_VER(dev_priv) >= 9) {
+		runtime->num_scalers[PIPE_A] = 2;
+		runtime->num_scalers[PIPE_B] = 2;
+		runtime->num_scalers[PIPE_C] = 1;
+	}
+
+	BUILD_BUG_ON(BITS_PER_TYPE(intel_engine_mask_t) < I915_NUM_ENGINES);
+
+	if (DISPLAY_VER(dev_priv) >= 13 || HAS_D12_PLANE_MINIMIZATION(dev_priv))
+		for_each_pipe(dev_priv, pipe)
+			runtime->num_sprites[pipe] = 4;
+	else if (DISPLAY_VER(dev_priv) >= 11)
+		for_each_pipe(dev_priv, pipe)
+			runtime->num_sprites[pipe] = 6;
+	else if (DISPLAY_VER(dev_priv) == 10)
+		for_each_pipe(dev_priv, pipe)
+			runtime->num_sprites[pipe] = 3;
+	else if (IS_BROXTON(dev_priv)) {
+		/*
+		 * Skylake and Broxton currently don't expose the topmost plane as its
+		 * use is exclusive with the legacy cursor and we only want to expose
+		 * one of those, not both. Until we can safely expose the topmost plane
+		 * as a DRM_PLANE_TYPE_CURSOR with all the features exposed/supported,
+		 * we don't expose the topmost plane at all to prevent ABI breakage
+		 * down the line.
+		 */
+
+		runtime->num_sprites[PIPE_A] = 2;
+		runtime->num_sprites[PIPE_B] = 2;
+		runtime->num_sprites[PIPE_C] = 1;
+	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		for_each_pipe(dev_priv, pipe)
+			runtime->num_sprites[pipe] = 2;
+	} else if (DISPLAY_VER(dev_priv) >= 5 || IS_G4X(dev_priv)) {
+		for_each_pipe(dev_priv, pipe)
+			runtime->num_sprites[pipe] = 1;
+	}
+
+	if (HAS_DISPLAY(dev_priv) && IS_GRAPHICS_VER(dev_priv, 7, 8) &&
+	    HAS_PCH_SPLIT(dev_priv)) {
+		u32 fuse_strap = intel_de_read(dev_priv, FUSE_STRAP);
+		u32 sfuse_strap = intel_de_read(dev_priv, SFUSE_STRAP);
+
+		/*
+		 * SFUSE_STRAP is supposed to have a bit signalling the display
+		 * is fused off. Unfortunately it seems that, at least in
+		 * certain cases, fused off display means that PCH display
+		 * reads don't land anywhere. In that case, we read 0s.
+		 *
+		 * On CPT/PPT, we can detect this case as SFUSE_STRAP_FUSE_LOCK
+		 * should be set when taking over after the firmware.
+		 */
+		if (fuse_strap & ILK_INTERNAL_DISPLAY_DISABLE ||
+		    sfuse_strap & SFUSE_STRAP_DISPLAY_DISABLED ||
+		    (HAS_PCH_CPT(dev_priv) &&
+		     !(sfuse_strap & SFUSE_STRAP_FUSE_LOCK))) {
+			drm_info(&dev_priv->drm,
+				 "Display fused off, disabling\n");
+			runtime->pipe_mask = 0;
+			runtime->cpu_transcoder_mask = 0;
+			runtime->fbc_mask = 0;
+		} else if (fuse_strap & IVB_PIPE_C_DISABLE) {
+			drm_info(&dev_priv->drm, "PipeC fused off\n");
+			runtime->pipe_mask &= ~BIT(PIPE_C);
+			runtime->cpu_transcoder_mask &= ~BIT(TRANSCODER_C);
+		}
+	} else if (HAS_DISPLAY(dev_priv) && DISPLAY_VER(dev_priv) >= 9) {
+		u32 dfsm = intel_de_read(dev_priv, SKL_DFSM);
+
+		if (dfsm & SKL_DFSM_PIPE_A_DISABLE) {
+			runtime->pipe_mask &= ~BIT(PIPE_A);
+			runtime->cpu_transcoder_mask &= ~BIT(TRANSCODER_A);
+			runtime->fbc_mask &= ~BIT(INTEL_FBC_A);
+		}
+		if (dfsm & SKL_DFSM_PIPE_B_DISABLE) {
+			runtime->pipe_mask &= ~BIT(PIPE_B);
+			runtime->cpu_transcoder_mask &= ~BIT(TRANSCODER_B);
+		}
+		if (dfsm & SKL_DFSM_PIPE_C_DISABLE) {
+			runtime->pipe_mask &= ~BIT(PIPE_C);
+			runtime->cpu_transcoder_mask &= ~BIT(TRANSCODER_C);
+		}
+
+		if (DISPLAY_VER(dev_priv) >= 12 &&
+		    (dfsm & TGL_DFSM_PIPE_D_DISABLE)) {
+			runtime->pipe_mask &= ~BIT(PIPE_D);
+			runtime->cpu_transcoder_mask &= ~BIT(TRANSCODER_D);
+		}
+
+		if (dfsm & SKL_DFSM_DISPLAY_HDCP_DISABLE)
+			runtime->has_hdcp = 0;
+
+		if (dfsm & SKL_DFSM_DISPLAY_PM_DISABLE)
+			runtime->fbc_mask = 0;
+
+		if (DISPLAY_VER(dev_priv) >= 11 && (dfsm & ICL_DFSM_DMC_DISABLE))
+			runtime->has_dmc = 0;
+
+		if (DISPLAY_VER(dev_priv) >= 10 &&
+		    (dfsm & GLK_DFSM_DISPLAY_DSC_DISABLE))
+			runtime->has_dsc = 0;
+	}
+
+	if (GRAPHICS_VER(dev_priv) == 6 && i915_vtd_active(dev_priv)) {
+		drm_info(&dev_priv->drm,
+			 "Disabling ppGTT for VT-d support\n");
+		runtime->ppgtt_type = INTEL_PPGTT_NONE;
+	}
+
+	runtime->rawclk_freq = intel_read_rawclk(dev_priv);
+	drm_dbg(&dev_priv->drm, "rawclk rate: %d kHz\n", runtime->rawclk_freq);
+
+	if (!HAS_DISPLAY(dev_priv)) {
+		dev_priv->drm.driver_features &= ~(DRIVER_MODESET |
+						   DRIVER_ATOMIC);
+		memset(&info->display, 0, sizeof(info->display));
+
+		runtime->cpu_transcoder_mask = 0;
+		memset(runtime->num_sprites, 0, sizeof(runtime->num_sprites));
+		memset(runtime->num_scalers, 0, sizeof(runtime->num_scalers));
+		runtime->fbc_mask = 0;
+		runtime->has_hdcp = false;
+		runtime->has_dmc = false;
+		runtime->has_dsc = false;
+	}
+}
+
+void intel_driver_caps_print(const struct intel_driver_caps *caps,
+			     struct drm_printer *p)
+{
+	drm_printf(p, "Has logical contexts? %s\n",
+		   str_yes_no(caps->has_logical_contexts));
+	drm_printf(p, "scheduler: %x\n", caps->scheduler);
+}
diff --git a/drivers/gpu/drm/i915/intel_device_info.h b/drivers/gpu/drm/i915/intel_device_info.h
new file mode 100644
index 000000000..d638235e1
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_device_info.h
@@ -0,0 +1,321 @@
+/*
+ * Copyright © 2014-2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef _INTEL_DEVICE_INFO_H_
+#define _INTEL_DEVICE_INFO_H_
+
+#include <uapi/drm/i915_drm.h>
+
+#include "intel_step.h"
+
+#include "display/intel_display.h"
+
+#include "gt/intel_engine_types.h"
+#include "gt/intel_context_types.h"
+#include "gt/intel_sseu.h"
+
+struct drm_printer;
+struct drm_i915_private;
+struct intel_gt_definition;
+
+/* Keep in gen based order, and chronological order within a gen */
+enum intel_platform {
+	INTEL_PLATFORM_UNINITIALIZED = 0,
+	/* gen2 */
+	INTEL_I830,
+	INTEL_I845G,
+	INTEL_I85X,
+	INTEL_I865G,
+	/* gen3 */
+	INTEL_I915G,
+	INTEL_I915GM,
+	INTEL_I945G,
+	INTEL_I945GM,
+	INTEL_G33,
+	INTEL_PINEVIEW,
+	/* gen4 */
+	INTEL_I965G,
+	INTEL_I965GM,
+	INTEL_G45,
+	INTEL_GM45,
+	/* gen5 */
+	INTEL_IRONLAKE,
+	/* gen6 */
+	INTEL_SANDYBRIDGE,
+	/* gen7 */
+	INTEL_IVYBRIDGE,
+	INTEL_VALLEYVIEW,
+	INTEL_HASWELL,
+	/* gen8 */
+	INTEL_BROADWELL,
+	INTEL_CHERRYVIEW,
+	/* gen9 */
+	INTEL_SKYLAKE,
+	INTEL_BROXTON,
+	INTEL_KABYLAKE,
+	INTEL_GEMINILAKE,
+	INTEL_COFFEELAKE,
+	INTEL_COMETLAKE,
+	/* gen11 */
+	INTEL_ICELAKE,
+	INTEL_ELKHARTLAKE,
+	INTEL_JASPERLAKE,
+	/* gen12 */
+	INTEL_TIGERLAKE,
+	INTEL_ROCKETLAKE,
+	INTEL_DG1,
+	INTEL_ALDERLAKE_S,
+	INTEL_ALDERLAKE_P,
+	INTEL_XEHPSDV,
+	INTEL_DG2,
+	INTEL_PONTEVECCHIO,
+	INTEL_METEORLAKE,
+	INTEL_MAX_PLATFORMS
+};
+
+/*
+ * Subplatform bits share the same namespace per parent platform. In other words
+ * it is fine for the same bit to be used on multiple parent platforms.
+ */
+
+#define INTEL_SUBPLATFORM_BITS (3)
+#define INTEL_SUBPLATFORM_MASK (BIT(INTEL_SUBPLATFORM_BITS) - 1)
+
+/* HSW/BDW/SKL/KBL/CFL */
+#define INTEL_SUBPLATFORM_ULT	(0)
+#define INTEL_SUBPLATFORM_ULX	(1)
+
+/* ICL */
+#define INTEL_SUBPLATFORM_PORTF	(0)
+
+/* TGL */
+#define INTEL_SUBPLATFORM_UY	(0)
+
+/* DG2 */
+#define INTEL_SUBPLATFORM_G10	0
+#define INTEL_SUBPLATFORM_G11	1
+#define INTEL_SUBPLATFORM_G12	2
+
+/* ADL */
+#define INTEL_SUBPLATFORM_RPL	0
+
+/* ADL-P */
+/*
+ * As #define INTEL_SUBPLATFORM_RPL 0 will apply
+ * here too, SUBPLATFORM_N will have different
+ * bit set
+ */
+#define INTEL_SUBPLATFORM_N    1
+
+/* MTL */
+#define INTEL_SUBPLATFORM_M	0
+#define INTEL_SUBPLATFORM_P	1
+
+enum intel_ppgtt_type {
+	INTEL_PPGTT_NONE = I915_GEM_PPGTT_NONE,
+	INTEL_PPGTT_ALIASING = I915_GEM_PPGTT_ALIASING,
+	INTEL_PPGTT_FULL = I915_GEM_PPGTT_FULL,
+};
+
+#define DEV_INFO_FOR_EACH_FLAG(func) \
+	func(is_mobile); \
+	func(is_lp); \
+	func(require_force_probe); \
+	func(is_dgfx); \
+	/* Keep has_* in alphabetical order */ \
+	func(has_64bit_reloc); \
+	func(has_64k_pages); \
+	func(needs_compact_pt); \
+	func(gpu_reset_clobbers_display); \
+	func(has_reset_engine); \
+	func(has_3d_pipeline); \
+	func(has_4tile); \
+	func(has_flat_ccs); \
+	func(has_global_mocs); \
+	func(has_gt_uc); \
+	func(has_heci_pxp); \
+	func(has_heci_gscfi); \
+	func(has_guc_deprivilege); \
+	func(has_l3_ccs_read); \
+	func(has_l3_dpf); \
+	func(has_llc); \
+	func(has_logical_ring_contexts); \
+	func(has_logical_ring_elsq); \
+	func(has_media_ratio_mode); \
+	func(has_mslice_steering); \
+	func(has_one_eu_per_fuse_bit); \
+	func(has_pxp); \
+	func(has_rc6); \
+	func(has_rc6p); \
+	func(has_rps); \
+	func(has_runtime_pm); \
+	func(has_snoop); \
+	func(has_coherent_ggtt); \
+	func(tuning_thread_rr_after_dep); \
+	func(unfenced_needs_alignment); \
+	func(hws_needs_physical);
+
+#define DEV_INFO_DISPLAY_FOR_EACH_FLAG(func) \
+	/* Keep in alphabetical order */ \
+	func(cursor_needs_physical); \
+	func(has_cdclk_crawl); \
+	func(has_ddi); \
+	func(has_dp_mst); \
+	func(has_dsb); \
+	func(has_fpga_dbg); \
+	func(has_gmch); \
+	func(has_hotplug); \
+	func(has_hti); \
+	func(has_ipc); \
+	func(has_modular_fia); \
+	func(has_overlay); \
+	func(has_psr); \
+	func(has_psr_hw_tracking); \
+	func(overlay_needs_physical); \
+	func(supports_tv);
+
+struct ip_version {
+	u8 ver;
+	u8 rel;
+};
+
+struct intel_runtime_info {
+	struct {
+		struct ip_version ip;
+	} graphics;
+	struct {
+		struct ip_version ip;
+	} media;
+	struct {
+		struct ip_version ip;
+	} display;
+
+	/*
+	 * Platform mask is used for optimizing or-ed IS_PLATFORM calls into
+	 * single runtime conditionals, and also to provide groundwork for
+	 * future per platform, or per SKU build optimizations.
+	 *
+	 * Array can be extended when necessary if the corresponding
+	 * BUILD_BUG_ON is hit.
+	 */
+	u32 platform_mask[2];
+
+	u16 device_id;
+
+	intel_engine_mask_t platform_engine_mask; /* Engines supported by the HW */
+
+	u32 rawclk_freq;
+
+	struct intel_step_info step;
+
+	unsigned int page_sizes; /* page sizes supported by the HW */
+
+	enum intel_ppgtt_type ppgtt_type;
+	unsigned int ppgtt_size; /* log2, e.g. 31/32/48 bits */
+
+	u32 memory_regions; /* regions supported by the HW */
+
+	bool has_pooled_eu;
+
+	/* display */
+	struct {
+		u8 pipe_mask;
+		u8 cpu_transcoder_mask;
+
+		u8 num_sprites[I915_MAX_PIPES];
+		u8 num_scalers[I915_MAX_PIPES];
+
+		u8 fbc_mask;
+
+		bool has_hdcp;
+		bool has_dmc;
+		bool has_dsc;
+	};
+};
+
+struct intel_device_info {
+	enum intel_platform platform;
+
+	unsigned int dma_mask_size; /* available DMA address bits */
+
+	const struct intel_gt_definition *extra_gt_list;
+
+	u8 gt; /* GT number, 0 if undefined */
+
+#define DEFINE_FLAG(name) u8 name:1
+	DEV_INFO_FOR_EACH_FLAG(DEFINE_FLAG);
+#undef DEFINE_FLAG
+
+	struct {
+		u8 abox_mask;
+
+		struct {
+			u16 size; /* in blocks */
+			u8 slice_mask;
+		} dbuf;
+
+#define DEFINE_FLAG(name) u8 name:1
+		DEV_INFO_DISPLAY_FOR_EACH_FLAG(DEFINE_FLAG);
+#undef DEFINE_FLAG
+
+		/* Global register offset for the display engine */
+		u32 mmio_offset;
+
+		/* Register offsets for the various display pipes and transcoders */
+		u32 pipe_offsets[I915_MAX_TRANSCODERS];
+		u32 trans_offsets[I915_MAX_TRANSCODERS];
+		u32 cursor_offsets[I915_MAX_PIPES];
+
+		struct {
+			u32 degamma_lut_size;
+			u32 gamma_lut_size;
+			u32 degamma_lut_tests;
+			u32 gamma_lut_tests;
+		} color;
+	} display;
+
+	/*
+	 * Initial runtime info. Do not access outside of i915_driver_create().
+	 */
+	const struct intel_runtime_info __runtime;
+};
+
+struct intel_driver_caps {
+	unsigned int scheduler;
+	bool has_logical_contexts:1;
+};
+
+const char *intel_platform_name(enum intel_platform platform);
+
+void intel_device_info_subplatform_init(struct drm_i915_private *dev_priv);
+void intel_device_info_runtime_init(struct drm_i915_private *dev_priv);
+
+void intel_device_info_print(const struct intel_device_info *info,
+			     const struct intel_runtime_info *runtime,
+			     struct drm_printer *p);
+
+void intel_driver_caps_print(const struct intel_driver_caps *caps,
+			     struct drm_printer *p);
+
+#endif
diff --git a/drivers/gpu/drm/i915/intel_dram.c b/drivers/gpu/drm/i915/intel_dram.c
new file mode 100644
index 000000000..bba8cb6e8
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_dram.c
@@ -0,0 +1,572 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_dram.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pcode.h"
+
+struct dram_dimm_info {
+	u16 size;
+	u8 width, ranks;
+};
+
+struct dram_channel_info {
+	struct dram_dimm_info dimm_l, dimm_s;
+	u8 ranks;
+	bool is_16gb_dimm;
+};
+
+#define DRAM_TYPE_STR(type) [INTEL_DRAM_ ## type] = #type
+
+static const char *intel_dram_type_str(enum intel_dram_type type)
+{
+	static const char * const str[] = {
+		DRAM_TYPE_STR(UNKNOWN),
+		DRAM_TYPE_STR(DDR3),
+		DRAM_TYPE_STR(DDR4),
+		DRAM_TYPE_STR(LPDDR3),
+		DRAM_TYPE_STR(LPDDR4),
+	};
+
+	if (type >= ARRAY_SIZE(str))
+		type = INTEL_DRAM_UNKNOWN;
+
+	return str[type];
+}
+
+#undef DRAM_TYPE_STR
+
+static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
+{
+	return dimm->ranks * 64 / (dimm->width ?: 1);
+}
+
+/* Returns total Gb for the whole DIMM */
+static int skl_get_dimm_size(u16 val)
+{
+	return (val & SKL_DRAM_SIZE_MASK) * 8;
+}
+
+static int skl_get_dimm_width(u16 val)
+{
+	if (skl_get_dimm_size(val) == 0)
+		return 0;
+
+	switch (val & SKL_DRAM_WIDTH_MASK) {
+	case SKL_DRAM_WIDTH_X8:
+	case SKL_DRAM_WIDTH_X16:
+	case SKL_DRAM_WIDTH_X32:
+		val = (val & SKL_DRAM_WIDTH_MASK) >> SKL_DRAM_WIDTH_SHIFT;
+		return 8 << val;
+	default:
+		MISSING_CASE(val);
+		return 0;
+	}
+}
+
+static int skl_get_dimm_ranks(u16 val)
+{
+	if (skl_get_dimm_size(val) == 0)
+		return 0;
+
+	val = (val & SKL_DRAM_RANK_MASK) >> SKL_DRAM_RANK_SHIFT;
+
+	return val + 1;
+}
+
+/* Returns total Gb for the whole DIMM */
+static int icl_get_dimm_size(u16 val)
+{
+	return (val & ICL_DRAM_SIZE_MASK) * 8 / 2;
+}
+
+static int icl_get_dimm_width(u16 val)
+{
+	if (icl_get_dimm_size(val) == 0)
+		return 0;
+
+	switch (val & ICL_DRAM_WIDTH_MASK) {
+	case ICL_DRAM_WIDTH_X8:
+	case ICL_DRAM_WIDTH_X16:
+	case ICL_DRAM_WIDTH_X32:
+		val = (val & ICL_DRAM_WIDTH_MASK) >> ICL_DRAM_WIDTH_SHIFT;
+		return 8 << val;
+	default:
+		MISSING_CASE(val);
+		return 0;
+	}
+}
+
+static int icl_get_dimm_ranks(u16 val)
+{
+	if (icl_get_dimm_size(val) == 0)
+		return 0;
+
+	val = (val & ICL_DRAM_RANK_MASK) >> ICL_DRAM_RANK_SHIFT;
+
+	return val + 1;
+}
+
+static bool
+skl_is_16gb_dimm(const struct dram_dimm_info *dimm)
+{
+	/* Convert total Gb to Gb per DRAM device */
+	return dimm->size / (intel_dimm_num_devices(dimm) ?: 1) == 16;
+}
+
+static void
+skl_dram_get_dimm_info(struct drm_i915_private *i915,
+		       struct dram_dimm_info *dimm,
+		       int channel, char dimm_name, u16 val)
+{
+	if (GRAPHICS_VER(i915) >= 11) {
+		dimm->size = icl_get_dimm_size(val);
+		dimm->width = icl_get_dimm_width(val);
+		dimm->ranks = icl_get_dimm_ranks(val);
+	} else {
+		dimm->size = skl_get_dimm_size(val);
+		dimm->width = skl_get_dimm_width(val);
+		dimm->ranks = skl_get_dimm_ranks(val);
+	}
+
+	drm_dbg_kms(&i915->drm,
+		    "CH%u DIMM %c size: %u Gb, width: X%u, ranks: %u, 16Gb DIMMs: %s\n",
+		    channel, dimm_name, dimm->size, dimm->width, dimm->ranks,
+		    str_yes_no(skl_is_16gb_dimm(dimm)));
+}
+
+static int
+skl_dram_get_channel_info(struct drm_i915_private *i915,
+			  struct dram_channel_info *ch,
+			  int channel, u32 val)
+{
+	skl_dram_get_dimm_info(i915, &ch->dimm_l,
+			       channel, 'L', val & 0xffff);
+	skl_dram_get_dimm_info(i915, &ch->dimm_s,
+			       channel, 'S', val >> 16);
+
+	if (ch->dimm_l.size == 0 && ch->dimm_s.size == 0) {
+		drm_dbg_kms(&i915->drm, "CH%u not populated\n", channel);
+		return -EINVAL;
+	}
+
+	if (ch->dimm_l.ranks == 2 || ch->dimm_s.ranks == 2)
+		ch->ranks = 2;
+	else if (ch->dimm_l.ranks == 1 && ch->dimm_s.ranks == 1)
+		ch->ranks = 2;
+	else
+		ch->ranks = 1;
+
+	ch->is_16gb_dimm = skl_is_16gb_dimm(&ch->dimm_l) ||
+		skl_is_16gb_dimm(&ch->dimm_s);
+
+	drm_dbg_kms(&i915->drm, "CH%u ranks: %u, 16Gb DIMMs: %s\n",
+		    channel, ch->ranks, str_yes_no(ch->is_16gb_dimm));
+
+	return 0;
+}
+
+static bool
+intel_is_dram_symmetric(const struct dram_channel_info *ch0,
+			const struct dram_channel_info *ch1)
+{
+	return !memcmp(ch0, ch1, sizeof(*ch0)) &&
+		(ch0->dimm_s.size == 0 ||
+		 !memcmp(&ch0->dimm_l, &ch0->dimm_s, sizeof(ch0->dimm_l)));
+}
+
+static int
+skl_dram_get_channels_info(struct drm_i915_private *i915)
+{
+	struct dram_info *dram_info = &i915->dram_info;
+	struct dram_channel_info ch0 = {}, ch1 = {};
+	u32 val;
+	int ret;
+
+	val = intel_uncore_read(&i915->uncore,
+				SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
+	ret = skl_dram_get_channel_info(i915, &ch0, 0, val);
+	if (ret == 0)
+		dram_info->num_channels++;
+
+	val = intel_uncore_read(&i915->uncore,
+				SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
+	ret = skl_dram_get_channel_info(i915, &ch1, 1, val);
+	if (ret == 0)
+		dram_info->num_channels++;
+
+	if (dram_info->num_channels == 0) {
+		drm_info(&i915->drm, "Number of memory channels is zero\n");
+		return -EINVAL;
+	}
+
+	if (ch0.ranks == 0 && ch1.ranks == 0) {
+		drm_info(&i915->drm, "couldn't get memory rank information\n");
+		return -EINVAL;
+	}
+
+	dram_info->wm_lv_0_adjust_needed = ch0.is_16gb_dimm || ch1.is_16gb_dimm;
+
+	dram_info->symmetric_memory = intel_is_dram_symmetric(&ch0, &ch1);
+
+	drm_dbg_kms(&i915->drm, "Memory configuration is symmetric? %s\n",
+		    str_yes_no(dram_info->symmetric_memory));
+
+	return 0;
+}
+
+static enum intel_dram_type
+skl_get_dram_type(struct drm_i915_private *i915)
+{
+	u32 val;
+
+	val = intel_uncore_read(&i915->uncore,
+				SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
+
+	switch (val & SKL_DRAM_DDR_TYPE_MASK) {
+	case SKL_DRAM_DDR_TYPE_DDR3:
+		return INTEL_DRAM_DDR3;
+	case SKL_DRAM_DDR_TYPE_DDR4:
+		return INTEL_DRAM_DDR4;
+	case SKL_DRAM_DDR_TYPE_LPDDR3:
+		return INTEL_DRAM_LPDDR3;
+	case SKL_DRAM_DDR_TYPE_LPDDR4:
+		return INTEL_DRAM_LPDDR4;
+	default:
+		MISSING_CASE(val);
+		return INTEL_DRAM_UNKNOWN;
+	}
+}
+
+static int
+skl_get_dram_info(struct drm_i915_private *i915)
+{
+	struct dram_info *dram_info = &i915->dram_info;
+	int ret;
+
+	dram_info->type = skl_get_dram_type(i915);
+	drm_dbg_kms(&i915->drm, "DRAM type: %s\n",
+		    intel_dram_type_str(dram_info->type));
+
+	ret = skl_dram_get_channels_info(i915);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+/* Returns Gb per DRAM device */
+static int bxt_get_dimm_size(u32 val)
+{
+	switch (val & BXT_DRAM_SIZE_MASK) {
+	case BXT_DRAM_SIZE_4GBIT:
+		return 4;
+	case BXT_DRAM_SIZE_6GBIT:
+		return 6;
+	case BXT_DRAM_SIZE_8GBIT:
+		return 8;
+	case BXT_DRAM_SIZE_12GBIT:
+		return 12;
+	case BXT_DRAM_SIZE_16GBIT:
+		return 16;
+	default:
+		MISSING_CASE(val);
+		return 0;
+	}
+}
+
+static int bxt_get_dimm_width(u32 val)
+{
+	if (!bxt_get_dimm_size(val))
+		return 0;
+
+	val = (val & BXT_DRAM_WIDTH_MASK) >> BXT_DRAM_WIDTH_SHIFT;
+
+	return 8 << val;
+}
+
+static int bxt_get_dimm_ranks(u32 val)
+{
+	if (!bxt_get_dimm_size(val))
+		return 0;
+
+	switch (val & BXT_DRAM_RANK_MASK) {
+	case BXT_DRAM_RANK_SINGLE:
+		return 1;
+	case BXT_DRAM_RANK_DUAL:
+		return 2;
+	default:
+		MISSING_CASE(val);
+		return 0;
+	}
+}
+
+static enum intel_dram_type bxt_get_dimm_type(u32 val)
+{
+	if (!bxt_get_dimm_size(val))
+		return INTEL_DRAM_UNKNOWN;
+
+	switch (val & BXT_DRAM_TYPE_MASK) {
+	case BXT_DRAM_TYPE_DDR3:
+		return INTEL_DRAM_DDR3;
+	case BXT_DRAM_TYPE_LPDDR3:
+		return INTEL_DRAM_LPDDR3;
+	case BXT_DRAM_TYPE_DDR4:
+		return INTEL_DRAM_DDR4;
+	case BXT_DRAM_TYPE_LPDDR4:
+		return INTEL_DRAM_LPDDR4;
+	default:
+		MISSING_CASE(val);
+		return INTEL_DRAM_UNKNOWN;
+	}
+}
+
+static void bxt_get_dimm_info(struct dram_dimm_info *dimm, u32 val)
+{
+	dimm->width = bxt_get_dimm_width(val);
+	dimm->ranks = bxt_get_dimm_ranks(val);
+
+	/*
+	 * Size in register is Gb per DRAM device. Convert to total
+	 * Gb to match the way we report this for non-LP platforms.
+	 */
+	dimm->size = bxt_get_dimm_size(val) * intel_dimm_num_devices(dimm);
+}
+
+static int bxt_get_dram_info(struct drm_i915_private *i915)
+{
+	struct dram_info *dram_info = &i915->dram_info;
+	u32 val;
+	u8 valid_ranks = 0;
+	int i;
+
+	/*
+	 * Now read each DUNIT8/9/10/11 to check the rank of each dimms.
+	 */
+	for (i = BXT_D_CR_DRP0_DUNIT_START; i <= BXT_D_CR_DRP0_DUNIT_END; i++) {
+		struct dram_dimm_info dimm;
+		enum intel_dram_type type;
+
+		val = intel_uncore_read(&i915->uncore, BXT_D_CR_DRP0_DUNIT(i));
+		if (val == 0xFFFFFFFF)
+			continue;
+
+		dram_info->num_channels++;
+
+		bxt_get_dimm_info(&dimm, val);
+		type = bxt_get_dimm_type(val);
+
+		drm_WARN_ON(&i915->drm, type != INTEL_DRAM_UNKNOWN &&
+			    dram_info->type != INTEL_DRAM_UNKNOWN &&
+			    dram_info->type != type);
+
+		drm_dbg_kms(&i915->drm,
+			    "CH%u DIMM size: %u Gb, width: X%u, ranks: %u, type: %s\n",
+			    i - BXT_D_CR_DRP0_DUNIT_START,
+			    dimm.size, dimm.width, dimm.ranks,
+			    intel_dram_type_str(type));
+
+		if (valid_ranks == 0)
+			valid_ranks = dimm.ranks;
+
+		if (type != INTEL_DRAM_UNKNOWN)
+			dram_info->type = type;
+	}
+
+	if (dram_info->type == INTEL_DRAM_UNKNOWN || valid_ranks == 0) {
+		drm_info(&i915->drm, "couldn't get memory information\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
+{
+	struct dram_info *dram_info = &dev_priv->dram_info;
+	u32 val = 0;
+	int ret;
+
+	ret = snb_pcode_read(&dev_priv->uncore, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
+			     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
+	if (ret)
+		return ret;
+
+	if (GRAPHICS_VER(dev_priv) == 12) {
+		switch (val & 0xf) {
+		case 0:
+			dram_info->type = INTEL_DRAM_DDR4;
+			break;
+		case 1:
+			dram_info->type = INTEL_DRAM_DDR5;
+			break;
+		case 2:
+			dram_info->type = INTEL_DRAM_LPDDR5;
+			break;
+		case 3:
+			dram_info->type = INTEL_DRAM_LPDDR4;
+			break;
+		case 4:
+			dram_info->type = INTEL_DRAM_DDR3;
+			break;
+		case 5:
+			dram_info->type = INTEL_DRAM_LPDDR3;
+			break;
+		default:
+			MISSING_CASE(val & 0xf);
+			return -EINVAL;
+		}
+	} else {
+		switch (val & 0xf) {
+		case 0:
+			dram_info->type = INTEL_DRAM_DDR4;
+			break;
+		case 1:
+			dram_info->type = INTEL_DRAM_DDR3;
+			break;
+		case 2:
+			dram_info->type = INTEL_DRAM_LPDDR3;
+			break;
+		case 3:
+			dram_info->type = INTEL_DRAM_LPDDR4;
+			break;
+		default:
+			MISSING_CASE(val & 0xf);
+			return -EINVAL;
+		}
+	}
+
+	dram_info->num_channels = (val & 0xf0) >> 4;
+	dram_info->num_qgv_points = (val & 0xf00) >> 8;
+	dram_info->num_psf_gv_points = (val & 0x3000) >> 12;
+
+	return 0;
+}
+
+static int gen11_get_dram_info(struct drm_i915_private *i915)
+{
+	int ret = skl_get_dram_info(i915);
+
+	if (ret)
+		return ret;
+
+	return icl_pcode_read_mem_global_info(i915);
+}
+
+static int gen12_get_dram_info(struct drm_i915_private *i915)
+{
+	i915->dram_info.wm_lv_0_adjust_needed = false;
+
+	return icl_pcode_read_mem_global_info(i915);
+}
+
+static int xelpdp_get_dram_info(struct drm_i915_private *i915)
+{
+	u32 val = intel_uncore_read(&i915->uncore, MTL_MEM_SS_INFO_GLOBAL);
+	struct dram_info *dram_info = &i915->dram_info;
+
+	switch (REG_FIELD_GET(MTL_DDR_TYPE_MASK, val)) {
+	case 0:
+		dram_info->type = INTEL_DRAM_DDR4;
+		break;
+	case 1:
+		dram_info->type = INTEL_DRAM_DDR5;
+		break;
+	case 2:
+		dram_info->type = INTEL_DRAM_LPDDR5;
+		break;
+	case 3:
+		dram_info->type = INTEL_DRAM_LPDDR4;
+		break;
+	case 4:
+		dram_info->type = INTEL_DRAM_DDR3;
+		break;
+	case 5:
+		dram_info->type = INTEL_DRAM_LPDDR3;
+		break;
+	default:
+		MISSING_CASE(val);
+		return -EINVAL;
+	}
+
+	dram_info->num_channels = REG_FIELD_GET(MTL_N_OF_POPULATED_CH_MASK, val);
+	dram_info->num_qgv_points = REG_FIELD_GET(MTL_N_OF_ENABLED_QGV_POINTS_MASK, val);
+	/* PSF GV points not supported in D14+ */
+
+	return 0;
+}
+
+void intel_dram_detect(struct drm_i915_private *i915)
+{
+	struct dram_info *dram_info = &i915->dram_info;
+	int ret;
+
+	if (GRAPHICS_VER(i915) < 9 || IS_DG2(i915) || !HAS_DISPLAY(i915))
+		return;
+
+	/*
+	 * Assume level 0 watermark latency adjustment is needed until proven
+	 * otherwise, this w/a is not needed by bxt/glk.
+	 */
+	dram_info->wm_lv_0_adjust_needed = !IS_GEN9_LP(i915);
+
+	if (DISPLAY_VER(i915) >= 14)
+		ret = xelpdp_get_dram_info(i915);
+	else if (GRAPHICS_VER(i915) >= 12)
+		ret = gen12_get_dram_info(i915);
+	else if (GRAPHICS_VER(i915) >= 11)
+		ret = gen11_get_dram_info(i915);
+	else if (IS_GEN9_LP(i915))
+		ret = bxt_get_dram_info(i915);
+	else
+		ret = skl_get_dram_info(i915);
+	if (ret)
+		return;
+
+	drm_dbg_kms(&i915->drm, "DRAM channels: %u\n", dram_info->num_channels);
+
+	drm_dbg_kms(&i915->drm, "Watermark level 0 adjustment needed: %s\n",
+		    str_yes_no(dram_info->wm_lv_0_adjust_needed));
+}
+
+static u32 gen9_edram_size_mb(struct drm_i915_private *i915, u32 cap)
+{
+	static const u8 ways[8] = { 4, 8, 12, 16, 16, 16, 16, 16 };
+	static const u8 sets[4] = { 1, 1, 2, 2 };
+
+	return EDRAM_NUM_BANKS(cap) *
+		ways[EDRAM_WAYS_IDX(cap)] *
+		sets[EDRAM_SETS_IDX(cap)];
+}
+
+void intel_dram_edram_detect(struct drm_i915_private *i915)
+{
+	u32 edram_cap = 0;
+
+	if (!(IS_HASWELL(i915) || IS_BROADWELL(i915) || GRAPHICS_VER(i915) >= 9))
+		return;
+
+	edram_cap = __raw_uncore_read32(&i915->uncore, HSW_EDRAM_CAP);
+
+	/* NB: We can't write IDICR yet because we don't have gt funcs set up */
+
+	if (!(edram_cap & EDRAM_ENABLED))
+		return;
+
+	/*
+	 * The needed capability bits for size calculation are not there with
+	 * pre gen9 so return 128MB always.
+	 */
+	if (GRAPHICS_VER(i915) < 9)
+		i915->edram_size_mb = 128;
+	else
+		i915->edram_size_mb = gen9_edram_size_mb(i915, edram_cap);
+
+	drm_info(&i915->drm, "Found %uMB of eDRAM\n", i915->edram_size_mb);
+}
diff --git a/drivers/gpu/drm/i915/intel_dram.h b/drivers/gpu/drm/i915/intel_dram.h
new file mode 100644
index 000000000..4ba13c131
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_dram.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __INTEL_DRAM_H__
+#define __INTEL_DRAM_H__
+
+struct drm_i915_private;
+
+void intel_dram_edram_detect(struct drm_i915_private *i915);
+void intel_dram_detect(struct drm_i915_private *i915);
+
+#endif /* __INTEL_DRAM_H__ */
diff --git a/drivers/gpu/drm/i915/intel_gvt.c b/drivers/gpu/drm/i915/intel_gvt.c
new file mode 100644
index 000000000..e98b6d69a
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_gvt.c
@@ -0,0 +1,322 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include "i915_drv.h"
+#include "i915_vgpu.h"
+#include "intel_gvt.h"
+#include "gem/i915_gem_dmabuf.h"
+#include "gt/intel_context.h"
+#include "gt/intel_ring.h"
+#include "gt/shmem_utils.h"
+
+/**
+ * DOC: Intel GVT-g host support
+ *
+ * Intel GVT-g is a graphics virtualization technology which shares the
+ * GPU among multiple virtual machines on a time-sharing basis. Each
+ * virtual machine is presented a virtual GPU (vGPU), which has equivalent
+ * features as the underlying physical GPU (pGPU), so i915 driver can run
+ * seamlessly in a virtual machine.
+ *
+ * To virtualize GPU resources GVT-g driver depends on hypervisor technology
+ * e.g KVM/VFIO/mdev, Xen, etc. to provide resource access trapping capability
+ * and be virtualized within GVT-g device module. More architectural design
+ * doc is available on https://01.org/group/2230/documentation-list.
+ */
+
+static LIST_HEAD(intel_gvt_devices);
+static const struct intel_vgpu_ops *intel_gvt_ops;
+static DEFINE_MUTEX(intel_gvt_mutex);
+
+static bool is_supported_device(struct drm_i915_private *dev_priv)
+{
+	if (IS_BROADWELL(dev_priv))
+		return true;
+	if (IS_SKYLAKE(dev_priv))
+		return true;
+	if (IS_KABYLAKE(dev_priv))
+		return true;
+	if (IS_BROXTON(dev_priv))
+		return true;
+	if (IS_COFFEELAKE(dev_priv))
+		return true;
+	if (IS_COMETLAKE(dev_priv))
+		return true;
+
+	return false;
+}
+
+static void free_initial_hw_state(struct drm_i915_private *dev_priv)
+{
+	struct i915_virtual_gpu *vgpu = &dev_priv->vgpu;
+
+	vfree(vgpu->initial_mmio);
+	vgpu->initial_mmio = NULL;
+
+	kfree(vgpu->initial_cfg_space);
+	vgpu->initial_cfg_space = NULL;
+}
+
+static void save_mmio(struct intel_gvt_mmio_table_iter *iter, u32 offset,
+		      u32 size)
+{
+	struct drm_i915_private *dev_priv = iter->i915;
+	u32 *mmio, i;
+
+	for (i = offset; i < offset + size; i += 4) {
+		mmio = iter->data + i;
+		*mmio = intel_uncore_read_notrace(to_gt(dev_priv)->uncore,
+						  _MMIO(i));
+	}
+}
+
+static int handle_mmio(struct intel_gvt_mmio_table_iter *iter,
+		       u32 offset, u32 size)
+{
+	if (WARN_ON(!IS_ALIGNED(offset, 4)))
+		return -EINVAL;
+
+	save_mmio(iter, offset, size);
+	return 0;
+}
+
+static int save_initial_hw_state(struct drm_i915_private *dev_priv)
+{
+	struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
+	struct i915_virtual_gpu *vgpu = &dev_priv->vgpu;
+	struct intel_gvt_mmio_table_iter iter;
+	void *mem;
+	int i, ret;
+
+	mem = kzalloc(PCI_CFG_SPACE_EXP_SIZE, GFP_KERNEL);
+	if (!mem)
+		return -ENOMEM;
+
+	vgpu->initial_cfg_space = mem;
+
+	for (i = 0; i < PCI_CFG_SPACE_EXP_SIZE; i += 4)
+		pci_read_config_dword(pdev, i, mem + i);
+
+	mem = vzalloc(2 * SZ_1M);
+	if (!mem) {
+		ret = -ENOMEM;
+		goto err_mmio;
+	}
+
+	vgpu->initial_mmio = mem;
+
+	iter.i915 = dev_priv;
+	iter.data = vgpu->initial_mmio;
+	iter.handle_mmio_cb = handle_mmio;
+
+	ret = intel_gvt_iterate_mmio_table(&iter);
+	if (ret)
+		goto err_iterate;
+
+	return 0;
+
+err_iterate:
+	vfree(vgpu->initial_mmio);
+	vgpu->initial_mmio = NULL;
+err_mmio:
+	kfree(vgpu->initial_cfg_space);
+	vgpu->initial_cfg_space = NULL;
+
+	return ret;
+}
+
+static void intel_gvt_init_device(struct drm_i915_private *dev_priv)
+{
+	if (!dev_priv->params.enable_gvt) {
+		drm_dbg(&dev_priv->drm,
+			"GVT-g is disabled by kernel params\n");
+		return;
+	}
+
+	if (intel_vgpu_active(dev_priv)) {
+		drm_info(&dev_priv->drm, "GVT-g is disabled for guest\n");
+		return;
+	}
+
+	if (!is_supported_device(dev_priv)) {
+		drm_info(&dev_priv->drm,
+			 "Unsupported device. GVT-g is disabled\n");
+		return;
+	}
+
+	if (intel_uc_wants_guc_submission(&to_gt(dev_priv)->uc)) {
+		drm_err(&dev_priv->drm,
+			"Graphics virtualization is not yet supported with GuC submission\n");
+		return;
+	}
+
+	if (save_initial_hw_state(dev_priv)) {
+		drm_dbg(&dev_priv->drm, "Failed to save initial HW state\n");
+		return;
+	}
+
+	if (intel_gvt_ops->init_device(dev_priv))
+		drm_dbg(&dev_priv->drm, "Fail to init GVT device\n");
+}
+
+static void intel_gvt_clean_device(struct drm_i915_private *dev_priv)
+{
+	if (dev_priv->gvt)
+		intel_gvt_ops->clean_device(dev_priv);
+	free_initial_hw_state(dev_priv);
+}
+
+int intel_gvt_set_ops(const struct intel_vgpu_ops *ops)
+{
+	struct drm_i915_private *dev_priv;
+
+	mutex_lock(&intel_gvt_mutex);
+	if (intel_gvt_ops) {
+		mutex_unlock(&intel_gvt_mutex);
+		return -EINVAL;
+	}
+	intel_gvt_ops = ops;
+
+	list_for_each_entry(dev_priv, &intel_gvt_devices, vgpu.entry)
+		intel_gvt_init_device(dev_priv);
+	mutex_unlock(&intel_gvt_mutex);
+
+	return 0;
+}
+EXPORT_SYMBOL_NS_GPL(intel_gvt_set_ops, I915_GVT);
+
+void intel_gvt_clear_ops(const struct intel_vgpu_ops *ops)
+{
+	struct drm_i915_private *dev_priv;
+
+	mutex_lock(&intel_gvt_mutex);
+	if (intel_gvt_ops != ops) {
+		mutex_unlock(&intel_gvt_mutex);
+		return;
+	}
+
+	list_for_each_entry(dev_priv, &intel_gvt_devices, vgpu.entry)
+		intel_gvt_clean_device(dev_priv);
+
+	intel_gvt_ops = NULL;
+	mutex_unlock(&intel_gvt_mutex);
+}
+EXPORT_SYMBOL_NS_GPL(intel_gvt_clear_ops, I915_GVT);
+
+/**
+ * intel_gvt_init - initialize GVT components
+ * @dev_priv: drm i915 private data
+ *
+ * This function is called at the initialization stage to create a GVT device.
+ *
+ * Returns:
+ * Zero on success, negative error code if failed.
+ *
+ */
+int intel_gvt_init(struct drm_i915_private *dev_priv)
+{
+	if (i915_inject_probe_failure(dev_priv))
+		return -ENODEV;
+
+	mutex_lock(&intel_gvt_mutex);
+	list_add_tail(&dev_priv->vgpu.entry, &intel_gvt_devices);
+	if (intel_gvt_ops)
+		intel_gvt_init_device(dev_priv);
+	mutex_unlock(&intel_gvt_mutex);
+
+	return 0;
+}
+
+/**
+ * intel_gvt_driver_remove - cleanup GVT components when i915 driver is
+ *			     unbinding
+ * @dev_priv: drm i915 private *
+ *
+ * This function is called at the i915 driver unloading stage, to shutdown
+ * GVT components and release the related resources.
+ */
+void intel_gvt_driver_remove(struct drm_i915_private *dev_priv)
+{
+	mutex_lock(&intel_gvt_mutex);
+	intel_gvt_clean_device(dev_priv);
+	list_del(&dev_priv->vgpu.entry);
+	mutex_unlock(&intel_gvt_mutex);
+}
+
+/**
+ * intel_gvt_resume - GVT resume routine wapper
+ *
+ * @dev_priv: drm i915 private *
+ *
+ * This function is called at the i915 driver resume stage to restore required
+ * HW status for GVT so that vGPU can continue running after resumed.
+ */
+void intel_gvt_resume(struct drm_i915_private *dev_priv)
+{
+	mutex_lock(&intel_gvt_mutex);
+	if (dev_priv->gvt)
+		intel_gvt_ops->pm_resume(dev_priv);
+	mutex_unlock(&intel_gvt_mutex);
+}
+
+/*
+ * Exported here so that the exports only get created when GVT support is
+ * actually enabled.
+ */
+EXPORT_SYMBOL_NS_GPL(i915_gem_object_alloc, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_gem_object_create_shmem, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_gem_object_init, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_gem_object_ggtt_pin_ww, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_gem_object_pin_map, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_gem_object_set_to_cpu_domain, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(__i915_gem_object_flush_map, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(__i915_gem_object_set_pages, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_gem_gtt_insert, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_gem_prime_export, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_gem_ww_ctx_init, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_gem_ww_ctx_backoff, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_gem_ww_ctx_fini, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_ppgtt_create, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_request_add, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_request_create, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_request_wait, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_reserve_fence, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_unreserve_fence, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_vm_release, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(_i915_vma_move_to_active, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(intel_context_create, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(__intel_context_do_pin, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(__intel_context_do_unpin, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(intel_ring_begin, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(intel_runtime_pm_get, I915_GVT);
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+EXPORT_SYMBOL_NS_GPL(intel_runtime_pm_put, I915_GVT);
+#endif
+EXPORT_SYMBOL_NS_GPL(intel_runtime_pm_put_unchecked, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(intel_uncore_forcewake_for_reg, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(intel_uncore_forcewake_get, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(intel_uncore_forcewake_put, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(shmem_pin_map, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(shmem_unpin_map, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(__px_dma, I915_GVT);
+EXPORT_SYMBOL_NS_GPL(i915_fence_ops, I915_GVT);
diff --git a/drivers/gpu/drm/i915/intel_gvt.h b/drivers/gpu/drm/i915/intel_gvt.h
new file mode 100644
index 000000000..eb2a2be25
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_gvt.h
@@ -0,0 +1,78 @@
+/*
+ * Copyright(c) 2011-2016 Intel Corporation. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef _INTEL_GVT_H_
+#define _INTEL_GVT_H_
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+
+#ifdef CONFIG_DRM_I915_GVT
+
+struct intel_gvt_mmio_table_iter {
+	struct drm_i915_private *i915;
+	void *data;
+	int (*handle_mmio_cb)(struct intel_gvt_mmio_table_iter *iter,
+			      u32 offset, u32 size);
+};
+
+int intel_gvt_init(struct drm_i915_private *dev_priv);
+void intel_gvt_driver_remove(struct drm_i915_private *dev_priv);
+int intel_gvt_init_host(void);
+void intel_gvt_resume(struct drm_i915_private *dev_priv);
+int intel_gvt_iterate_mmio_table(struct intel_gvt_mmio_table_iter *iter);
+
+struct intel_vgpu_ops {
+	int (*init_device)(struct drm_i915_private *dev_priv);
+	void (*clean_device)(struct drm_i915_private *dev_priv);
+	void (*pm_resume)(struct drm_i915_private *i915);
+};
+
+int intel_gvt_set_ops(const struct intel_vgpu_ops *ops);
+void intel_gvt_clear_ops(const struct intel_vgpu_ops *ops);
+
+#else
+static inline int intel_gvt_init(struct drm_i915_private *dev_priv)
+{
+	return 0;
+}
+
+static inline void intel_gvt_driver_remove(struct drm_i915_private *dev_priv)
+{
+}
+
+static inline void intel_gvt_resume(struct drm_i915_private *dev_priv)
+{
+}
+
+struct intel_gvt_mmio_table_iter {
+};
+
+static inline int intel_gvt_iterate_mmio_table(struct intel_gvt_mmio_table_iter *iter)
+{
+	return 0;
+}
+#endif
+
+#endif /* _INTEL_GVT_H_ */
diff --git a/drivers/gpu/drm/i915/intel_gvt_mmio_table.c b/drivers/gpu/drm/i915/intel_gvt_mmio_table.c
new file mode 100644
index 000000000..8279dc580
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_gvt_mmio_table.c
@@ -0,0 +1,1296 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include "display/intel_audio_regs.h"
+#include "display/intel_backlight_regs.h"
+#include "display/intel_dmc_regs.h"
+#include "display/vlv_dsi_pll_regs.h"
+#include "gt/intel_gt_regs.h"
+#include "gvt/gvt.h"
+
+#include "i915_drv.h"
+#include "i915_pvinfo.h"
+#include "i915_reg.h"
+#include "intel_gvt.h"
+#include "intel_mchbar_regs.h"
+
+#define MMIO_F(reg, s) do { \
+	int ret; \
+	ret = iter->handle_mmio_cb(iter, i915_mmio_reg_offset(reg), s); \
+	if (ret) \
+		return ret; \
+} while (0)
+
+#define MMIO_D(reg) MMIO_F(reg, 4)
+
+#define MMIO_RING_F(prefix, s) do { \
+	MMIO_F(prefix(RENDER_RING_BASE), s); \
+	MMIO_F(prefix(BLT_RING_BASE), s); \
+	MMIO_F(prefix(GEN6_BSD_RING_BASE), s); \
+	MMIO_F(prefix(VEBOX_RING_BASE), s); \
+	if (HAS_ENGINE(to_gt(iter->i915), VCS1)) \
+		MMIO_F(prefix(GEN8_BSD2_RING_BASE), s); \
+} while (0)
+
+#define MMIO_RING_D(prefix) \
+	MMIO_RING_F(prefix, 4)
+
+static int iterate_generic_mmio(struct intel_gvt_mmio_table_iter *iter)
+{
+	struct drm_i915_private *dev_priv = iter->i915;
+
+	MMIO_RING_D(RING_IMR);
+	MMIO_D(SDEIMR);
+	MMIO_D(SDEIER);
+	MMIO_D(SDEIIR);
+	MMIO_D(SDEISR);
+	MMIO_RING_D(RING_HWSTAM);
+	MMIO_D(BSD_HWS_PGA_GEN7);
+	MMIO_D(BLT_HWS_PGA_GEN7);
+	MMIO_D(VEBOX_HWS_PGA_GEN7);
+
+#define RING_REG(base) _MMIO((base) + 0x28)
+	MMIO_RING_D(RING_REG);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x134)
+	MMIO_RING_D(RING_REG);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x6c)
+	MMIO_RING_D(RING_REG);
+#undef RING_REG
+	MMIO_D(_MMIO(0x2148));
+	MMIO_D(CCID(RENDER_RING_BASE));
+	MMIO_D(_MMIO(0x12198));
+	MMIO_D(GEN7_CXT_SIZE);
+	MMIO_RING_D(RING_TAIL);
+	MMIO_RING_D(RING_HEAD);
+	MMIO_RING_D(RING_CTL);
+	MMIO_RING_D(RING_ACTHD);
+	MMIO_RING_D(RING_START);
+
+	/* RING MODE */
+#define RING_REG(base) _MMIO((base) + 0x29c)
+	MMIO_RING_D(RING_REG);
+#undef RING_REG
+
+	MMIO_RING_D(RING_MI_MODE);
+	MMIO_RING_D(RING_INSTPM);
+	MMIO_RING_D(RING_TIMESTAMP);
+	MMIO_RING_D(RING_TIMESTAMP_UDW);
+	MMIO_D(GEN7_GT_MODE);
+	MMIO_D(CACHE_MODE_0_GEN7);
+	MMIO_D(CACHE_MODE_1);
+	MMIO_D(CACHE_MODE_0);
+	MMIO_D(_MMIO(0x2124));
+	MMIO_D(_MMIO(0x20dc));
+	MMIO_D(_3D_CHICKEN3);
+	MMIO_D(_MMIO(0x2088));
+	MMIO_D(FF_SLICE_CS_CHICKEN2);
+	MMIO_D(_MMIO(0x2470));
+	MMIO_D(GAM_ECOCHK);
+	MMIO_D(GEN7_COMMON_SLICE_CHICKEN1);
+	MMIO_D(COMMON_SLICE_CHICKEN2);
+	MMIO_D(_MMIO(0x9030));
+	MMIO_D(_MMIO(0x20a0));
+	MMIO_D(_MMIO(0x2420));
+	MMIO_D(_MMIO(0x2430));
+	MMIO_D(_MMIO(0x2434));
+	MMIO_D(_MMIO(0x2438));
+	MMIO_D(_MMIO(0x243c));
+	MMIO_D(_MMIO(0x7018));
+	MMIO_D(HALF_SLICE_CHICKEN3);
+	MMIO_D(GEN7_HALF_SLICE_CHICKEN1);
+	/* display */
+	MMIO_F(_MMIO(0x60220), 0x20);
+	MMIO_D(_MMIO(0x602a0));
+	MMIO_D(_MMIO(0x65050));
+	MMIO_D(_MMIO(0x650b4));
+	MMIO_D(_MMIO(0xc4040));
+	MMIO_D(DERRMR);
+	MMIO_D(PIPEDSL(PIPE_A));
+	MMIO_D(PIPEDSL(PIPE_B));
+	MMIO_D(PIPEDSL(PIPE_C));
+	MMIO_D(PIPEDSL(_PIPE_EDP));
+	MMIO_D(PIPECONF(PIPE_A));
+	MMIO_D(PIPECONF(PIPE_B));
+	MMIO_D(PIPECONF(PIPE_C));
+	MMIO_D(PIPECONF(_PIPE_EDP));
+	MMIO_D(PIPESTAT(PIPE_A));
+	MMIO_D(PIPESTAT(PIPE_B));
+	MMIO_D(PIPESTAT(PIPE_C));
+	MMIO_D(PIPESTAT(_PIPE_EDP));
+	MMIO_D(PIPE_FLIPCOUNT_G4X(PIPE_A));
+	MMIO_D(PIPE_FLIPCOUNT_G4X(PIPE_B));
+	MMIO_D(PIPE_FLIPCOUNT_G4X(PIPE_C));
+	MMIO_D(PIPE_FLIPCOUNT_G4X(_PIPE_EDP));
+	MMIO_D(PIPE_FRMCOUNT_G4X(PIPE_A));
+	MMIO_D(PIPE_FRMCOUNT_G4X(PIPE_B));
+	MMIO_D(PIPE_FRMCOUNT_G4X(PIPE_C));
+	MMIO_D(PIPE_FRMCOUNT_G4X(_PIPE_EDP));
+	MMIO_D(CURCNTR(PIPE_A));
+	MMIO_D(CURCNTR(PIPE_B));
+	MMIO_D(CURCNTR(PIPE_C));
+	MMIO_D(CURPOS(PIPE_A));
+	MMIO_D(CURPOS(PIPE_B));
+	MMIO_D(CURPOS(PIPE_C));
+	MMIO_D(CURBASE(PIPE_A));
+	MMIO_D(CURBASE(PIPE_B));
+	MMIO_D(CURBASE(PIPE_C));
+	MMIO_D(CUR_FBC_CTL(PIPE_A));
+	MMIO_D(CUR_FBC_CTL(PIPE_B));
+	MMIO_D(CUR_FBC_CTL(PIPE_C));
+	MMIO_D(_MMIO(0x700ac));
+	MMIO_D(_MMIO(0x710ac));
+	MMIO_D(_MMIO(0x720ac));
+	MMIO_D(_MMIO(0x70090));
+	MMIO_D(_MMIO(0x70094));
+	MMIO_D(_MMIO(0x70098));
+	MMIO_D(_MMIO(0x7009c));
+	MMIO_D(DSPCNTR(PIPE_A));
+	MMIO_D(DSPADDR(PIPE_A));
+	MMIO_D(DSPSTRIDE(PIPE_A));
+	MMIO_D(DSPPOS(PIPE_A));
+	MMIO_D(DSPSIZE(PIPE_A));
+	MMIO_D(DSPSURF(PIPE_A));
+	MMIO_D(DSPOFFSET(PIPE_A));
+	MMIO_D(DSPSURFLIVE(PIPE_A));
+	MMIO_D(REG_50080(PIPE_A, PLANE_PRIMARY));
+	MMIO_D(DSPCNTR(PIPE_B));
+	MMIO_D(DSPADDR(PIPE_B));
+	MMIO_D(DSPSTRIDE(PIPE_B));
+	MMIO_D(DSPPOS(PIPE_B));
+	MMIO_D(DSPSIZE(PIPE_B));
+	MMIO_D(DSPSURF(PIPE_B));
+	MMIO_D(DSPOFFSET(PIPE_B));
+	MMIO_D(DSPSURFLIVE(PIPE_B));
+	MMIO_D(REG_50080(PIPE_B, PLANE_PRIMARY));
+	MMIO_D(DSPCNTR(PIPE_C));
+	MMIO_D(DSPADDR(PIPE_C));
+	MMIO_D(DSPSTRIDE(PIPE_C));
+	MMIO_D(DSPPOS(PIPE_C));
+	MMIO_D(DSPSIZE(PIPE_C));
+	MMIO_D(DSPSURF(PIPE_C));
+	MMIO_D(DSPOFFSET(PIPE_C));
+	MMIO_D(DSPSURFLIVE(PIPE_C));
+	MMIO_D(REG_50080(PIPE_C, PLANE_PRIMARY));
+	MMIO_D(SPRCTL(PIPE_A));
+	MMIO_D(SPRLINOFF(PIPE_A));
+	MMIO_D(SPRSTRIDE(PIPE_A));
+	MMIO_D(SPRPOS(PIPE_A));
+	MMIO_D(SPRSIZE(PIPE_A));
+	MMIO_D(SPRKEYVAL(PIPE_A));
+	MMIO_D(SPRKEYMSK(PIPE_A));
+	MMIO_D(SPRSURF(PIPE_A));
+	MMIO_D(SPRKEYMAX(PIPE_A));
+	MMIO_D(SPROFFSET(PIPE_A));
+	MMIO_D(SPRSCALE(PIPE_A));
+	MMIO_D(SPRSURFLIVE(PIPE_A));
+	MMIO_D(REG_50080(PIPE_A, PLANE_SPRITE0));
+	MMIO_D(SPRCTL(PIPE_B));
+	MMIO_D(SPRLINOFF(PIPE_B));
+	MMIO_D(SPRSTRIDE(PIPE_B));
+	MMIO_D(SPRPOS(PIPE_B));
+	MMIO_D(SPRSIZE(PIPE_B));
+	MMIO_D(SPRKEYVAL(PIPE_B));
+	MMIO_D(SPRKEYMSK(PIPE_B));
+	MMIO_D(SPRSURF(PIPE_B));
+	MMIO_D(SPRKEYMAX(PIPE_B));
+	MMIO_D(SPROFFSET(PIPE_B));
+	MMIO_D(SPRSCALE(PIPE_B));
+	MMIO_D(SPRSURFLIVE(PIPE_B));
+	MMIO_D(REG_50080(PIPE_B, PLANE_SPRITE0));
+	MMIO_D(SPRCTL(PIPE_C));
+	MMIO_D(SPRLINOFF(PIPE_C));
+	MMIO_D(SPRSTRIDE(PIPE_C));
+	MMIO_D(SPRPOS(PIPE_C));
+	MMIO_D(SPRSIZE(PIPE_C));
+	MMIO_D(SPRKEYVAL(PIPE_C));
+	MMIO_D(SPRKEYMSK(PIPE_C));
+	MMIO_D(SPRSURF(PIPE_C));
+	MMIO_D(SPRKEYMAX(PIPE_C));
+	MMIO_D(SPROFFSET(PIPE_C));
+	MMIO_D(SPRSCALE(PIPE_C));
+	MMIO_D(SPRSURFLIVE(PIPE_C));
+	MMIO_D(REG_50080(PIPE_C, PLANE_SPRITE0));
+	MMIO_D(HTOTAL(TRANSCODER_A));
+	MMIO_D(HBLANK(TRANSCODER_A));
+	MMIO_D(HSYNC(TRANSCODER_A));
+	MMIO_D(VTOTAL(TRANSCODER_A));
+	MMIO_D(VBLANK(TRANSCODER_A));
+	MMIO_D(VSYNC(TRANSCODER_A));
+	MMIO_D(BCLRPAT(TRANSCODER_A));
+	MMIO_D(VSYNCSHIFT(TRANSCODER_A));
+	MMIO_D(PIPESRC(TRANSCODER_A));
+	MMIO_D(HTOTAL(TRANSCODER_B));
+	MMIO_D(HBLANK(TRANSCODER_B));
+	MMIO_D(HSYNC(TRANSCODER_B));
+	MMIO_D(VTOTAL(TRANSCODER_B));
+	MMIO_D(VBLANK(TRANSCODER_B));
+	MMIO_D(VSYNC(TRANSCODER_B));
+	MMIO_D(BCLRPAT(TRANSCODER_B));
+	MMIO_D(VSYNCSHIFT(TRANSCODER_B));
+	MMIO_D(PIPESRC(TRANSCODER_B));
+	MMIO_D(HTOTAL(TRANSCODER_C));
+	MMIO_D(HBLANK(TRANSCODER_C));
+	MMIO_D(HSYNC(TRANSCODER_C));
+	MMIO_D(VTOTAL(TRANSCODER_C));
+	MMIO_D(VBLANK(TRANSCODER_C));
+	MMIO_D(VSYNC(TRANSCODER_C));
+	MMIO_D(BCLRPAT(TRANSCODER_C));
+	MMIO_D(VSYNCSHIFT(TRANSCODER_C));
+	MMIO_D(PIPESRC(TRANSCODER_C));
+	MMIO_D(HTOTAL(TRANSCODER_EDP));
+	MMIO_D(HBLANK(TRANSCODER_EDP));
+	MMIO_D(HSYNC(TRANSCODER_EDP));
+	MMIO_D(VTOTAL(TRANSCODER_EDP));
+	MMIO_D(VBLANK(TRANSCODER_EDP));
+	MMIO_D(VSYNC(TRANSCODER_EDP));
+	MMIO_D(BCLRPAT(TRANSCODER_EDP));
+	MMIO_D(VSYNCSHIFT(TRANSCODER_EDP));
+	MMIO_D(PIPE_DATA_M1(TRANSCODER_A));
+	MMIO_D(PIPE_DATA_N1(TRANSCODER_A));
+	MMIO_D(PIPE_DATA_M2(TRANSCODER_A));
+	MMIO_D(PIPE_DATA_N2(TRANSCODER_A));
+	MMIO_D(PIPE_LINK_M1(TRANSCODER_A));
+	MMIO_D(PIPE_LINK_N1(TRANSCODER_A));
+	MMIO_D(PIPE_LINK_M2(TRANSCODER_A));
+	MMIO_D(PIPE_LINK_N2(TRANSCODER_A));
+	MMIO_D(PIPE_DATA_M1(TRANSCODER_B));
+	MMIO_D(PIPE_DATA_N1(TRANSCODER_B));
+	MMIO_D(PIPE_DATA_M2(TRANSCODER_B));
+	MMIO_D(PIPE_DATA_N2(TRANSCODER_B));
+	MMIO_D(PIPE_LINK_M1(TRANSCODER_B));
+	MMIO_D(PIPE_LINK_N1(TRANSCODER_B));
+	MMIO_D(PIPE_LINK_M2(TRANSCODER_B));
+	MMIO_D(PIPE_LINK_N2(TRANSCODER_B));
+	MMIO_D(PIPE_DATA_M1(TRANSCODER_C));
+	MMIO_D(PIPE_DATA_N1(TRANSCODER_C));
+	MMIO_D(PIPE_DATA_M2(TRANSCODER_C));
+	MMIO_D(PIPE_DATA_N2(TRANSCODER_C));
+	MMIO_D(PIPE_LINK_M1(TRANSCODER_C));
+	MMIO_D(PIPE_LINK_N1(TRANSCODER_C));
+	MMIO_D(PIPE_LINK_M2(TRANSCODER_C));
+	MMIO_D(PIPE_LINK_N2(TRANSCODER_C));
+	MMIO_D(PIPE_DATA_M1(TRANSCODER_EDP));
+	MMIO_D(PIPE_DATA_N1(TRANSCODER_EDP));
+	MMIO_D(PIPE_DATA_M2(TRANSCODER_EDP));
+	MMIO_D(PIPE_DATA_N2(TRANSCODER_EDP));
+	MMIO_D(PIPE_LINK_M1(TRANSCODER_EDP));
+	MMIO_D(PIPE_LINK_N1(TRANSCODER_EDP));
+	MMIO_D(PIPE_LINK_M2(TRANSCODER_EDP));
+	MMIO_D(PIPE_LINK_N2(TRANSCODER_EDP));
+	MMIO_D(PF_CTL(PIPE_A));
+	MMIO_D(PF_WIN_SZ(PIPE_A));
+	MMIO_D(PF_WIN_POS(PIPE_A));
+	MMIO_D(PF_VSCALE(PIPE_A));
+	MMIO_D(PF_HSCALE(PIPE_A));
+	MMIO_D(PF_CTL(PIPE_B));
+	MMIO_D(PF_WIN_SZ(PIPE_B));
+	MMIO_D(PF_WIN_POS(PIPE_B));
+	MMIO_D(PF_VSCALE(PIPE_B));
+	MMIO_D(PF_HSCALE(PIPE_B));
+	MMIO_D(PF_CTL(PIPE_C));
+	MMIO_D(PF_WIN_SZ(PIPE_C));
+	MMIO_D(PF_WIN_POS(PIPE_C));
+	MMIO_D(PF_VSCALE(PIPE_C));
+	MMIO_D(PF_HSCALE(PIPE_C));
+	MMIO_D(WM0_PIPE_ILK(PIPE_A));
+	MMIO_D(WM0_PIPE_ILK(PIPE_B));
+	MMIO_D(WM0_PIPE_ILK(PIPE_C));
+	MMIO_D(WM1_LP_ILK);
+	MMIO_D(WM2_LP_ILK);
+	MMIO_D(WM3_LP_ILK);
+	MMIO_D(WM1S_LP_ILK);
+	MMIO_D(WM2S_LP_IVB);
+	MMIO_D(WM3S_LP_IVB);
+	MMIO_D(BLC_PWM_CPU_CTL2);
+	MMIO_D(BLC_PWM_CPU_CTL);
+	MMIO_D(BLC_PWM_PCH_CTL1);
+	MMIO_D(BLC_PWM_PCH_CTL2);
+	MMIO_D(_MMIO(0x48268));
+	MMIO_F(PCH_GMBUS0, 4 * 4);
+	MMIO_F(PCH_GPIO_BASE, 6 * 4);
+	MMIO_F(_MMIO(0xe4f00), 0x28);
+	MMIO_D(_MMIO(_PCH_TRANSACONF));
+	MMIO_D(_MMIO(_PCH_TRANSBCONF));
+	MMIO_D(FDI_RX_IIR(PIPE_A));
+	MMIO_D(FDI_RX_IIR(PIPE_B));
+	MMIO_D(FDI_RX_IIR(PIPE_C));
+	MMIO_D(FDI_RX_IMR(PIPE_A));
+	MMIO_D(FDI_RX_IMR(PIPE_B));
+	MMIO_D(FDI_RX_IMR(PIPE_C));
+	MMIO_D(FDI_RX_CTL(PIPE_A));
+	MMIO_D(FDI_RX_CTL(PIPE_B));
+	MMIO_D(FDI_RX_CTL(PIPE_C));
+	MMIO_D(_MMIO(_PCH_TRANS_HTOTAL_A));
+	MMIO_D(_MMIO(_PCH_TRANS_HBLANK_A));
+	MMIO_D(_MMIO(_PCH_TRANS_HSYNC_A));
+	MMIO_D(_MMIO(_PCH_TRANS_VTOTAL_A));
+	MMIO_D(_MMIO(_PCH_TRANS_VBLANK_A));
+	MMIO_D(_MMIO(_PCH_TRANS_VSYNC_A));
+	MMIO_D(_MMIO(_PCH_TRANS_VSYNCSHIFT_A));
+	MMIO_D(_MMIO(_PCH_TRANS_HTOTAL_B));
+	MMIO_D(_MMIO(_PCH_TRANS_HBLANK_B));
+	MMIO_D(_MMIO(_PCH_TRANS_HSYNC_B));
+	MMIO_D(_MMIO(_PCH_TRANS_VTOTAL_B));
+	MMIO_D(_MMIO(_PCH_TRANS_VBLANK_B));
+	MMIO_D(_MMIO(_PCH_TRANS_VSYNC_B));
+	MMIO_D(_MMIO(_PCH_TRANS_VSYNCSHIFT_B));
+	MMIO_D(_MMIO(_PCH_TRANSA_DATA_M1));
+	MMIO_D(_MMIO(_PCH_TRANSA_DATA_N1));
+	MMIO_D(_MMIO(_PCH_TRANSA_DATA_M2));
+	MMIO_D(_MMIO(_PCH_TRANSA_DATA_N2));
+	MMIO_D(_MMIO(_PCH_TRANSA_LINK_M1));
+	MMIO_D(_MMIO(_PCH_TRANSA_LINK_N1));
+	MMIO_D(_MMIO(_PCH_TRANSA_LINK_M2));
+	MMIO_D(_MMIO(_PCH_TRANSA_LINK_N2));
+	MMIO_D(TRANS_DP_CTL(PIPE_A));
+	MMIO_D(TRANS_DP_CTL(PIPE_B));
+	MMIO_D(TRANS_DP_CTL(PIPE_C));
+	MMIO_D(TVIDEO_DIP_CTL(PIPE_A));
+	MMIO_D(TVIDEO_DIP_DATA(PIPE_A));
+	MMIO_D(TVIDEO_DIP_GCP(PIPE_A));
+	MMIO_D(TVIDEO_DIP_CTL(PIPE_B));
+	MMIO_D(TVIDEO_DIP_DATA(PIPE_B));
+	MMIO_D(TVIDEO_DIP_GCP(PIPE_B));
+	MMIO_D(TVIDEO_DIP_CTL(PIPE_C));
+	MMIO_D(TVIDEO_DIP_DATA(PIPE_C));
+	MMIO_D(TVIDEO_DIP_GCP(PIPE_C));
+	MMIO_D(_MMIO(_FDI_RXA_MISC));
+	MMIO_D(_MMIO(_FDI_RXB_MISC));
+	MMIO_D(_MMIO(_FDI_RXA_TUSIZE1));
+	MMIO_D(_MMIO(_FDI_RXA_TUSIZE2));
+	MMIO_D(_MMIO(_FDI_RXB_TUSIZE1));
+	MMIO_D(_MMIO(_FDI_RXB_TUSIZE2));
+	MMIO_D(PCH_PP_CONTROL);
+	MMIO_D(PCH_PP_DIVISOR);
+	MMIO_D(PCH_PP_STATUS);
+	MMIO_D(PCH_LVDS);
+	MMIO_D(_MMIO(_PCH_DPLL_A));
+	MMIO_D(_MMIO(_PCH_DPLL_B));
+	MMIO_D(_MMIO(_PCH_FPA0));
+	MMIO_D(_MMIO(_PCH_FPA1));
+	MMIO_D(_MMIO(_PCH_FPB0));
+	MMIO_D(_MMIO(_PCH_FPB1));
+	MMIO_D(PCH_DREF_CONTROL);
+	MMIO_D(PCH_RAWCLK_FREQ);
+	MMIO_D(PCH_DPLL_SEL);
+	MMIO_D(_MMIO(0x61208));
+	MMIO_D(_MMIO(0x6120c));
+	MMIO_D(PCH_PP_ON_DELAYS);
+	MMIO_D(PCH_PP_OFF_DELAYS);
+	MMIO_D(_MMIO(0xe651c));
+	MMIO_D(_MMIO(0xe661c));
+	MMIO_D(_MMIO(0xe671c));
+	MMIO_D(_MMIO(0xe681c));
+	MMIO_D(_MMIO(0xe6c04));
+	MMIO_D(_MMIO(0xe6e1c));
+	MMIO_D(PCH_PORT_HOTPLUG);
+	MMIO_D(LCPLL_CTL);
+	MMIO_D(FUSE_STRAP);
+	MMIO_D(DIGITAL_PORT_HOTPLUG_CNTRL);
+	MMIO_D(DISP_ARB_CTL);
+	MMIO_D(DISP_ARB_CTL2);
+	MMIO_D(ILK_DISPLAY_CHICKEN1);
+	MMIO_D(ILK_DISPLAY_CHICKEN2);
+	MMIO_D(ILK_DSPCLK_GATE_D);
+	MMIO_D(SOUTH_CHICKEN1);
+	MMIO_D(SOUTH_CHICKEN2);
+	MMIO_D(_MMIO(_TRANSA_CHICKEN1));
+	MMIO_D(_MMIO(_TRANSB_CHICKEN1));
+	MMIO_D(SOUTH_DSPCLK_GATE_D);
+	MMIO_D(_MMIO(_TRANSA_CHICKEN2));
+	MMIO_D(_MMIO(_TRANSB_CHICKEN2));
+	MMIO_D(ILK_DPFC_CB_BASE(INTEL_FBC_A));
+	MMIO_D(ILK_DPFC_CONTROL(INTEL_FBC_A));
+	MMIO_D(ILK_DPFC_RECOMP_CTL(INTEL_FBC_A));
+	MMIO_D(ILK_DPFC_STATUS(INTEL_FBC_A));
+	MMIO_D(ILK_DPFC_FENCE_YOFF(INTEL_FBC_A));
+	MMIO_D(ILK_DPFC_CHICKEN(INTEL_FBC_A));
+	MMIO_D(ILK_FBC_RT_BASE);
+	MMIO_D(IPS_CTL);
+	MMIO_D(PIPE_CSC_COEFF_RY_GY(PIPE_A));
+	MMIO_D(PIPE_CSC_COEFF_BY(PIPE_A));
+	MMIO_D(PIPE_CSC_COEFF_RU_GU(PIPE_A));
+	MMIO_D(PIPE_CSC_COEFF_BU(PIPE_A));
+	MMIO_D(PIPE_CSC_COEFF_RV_GV(PIPE_A));
+	MMIO_D(PIPE_CSC_COEFF_BV(PIPE_A));
+	MMIO_D(PIPE_CSC_MODE(PIPE_A));
+	MMIO_D(PIPE_CSC_PREOFF_HI(PIPE_A));
+	MMIO_D(PIPE_CSC_PREOFF_ME(PIPE_A));
+	MMIO_D(PIPE_CSC_PREOFF_LO(PIPE_A));
+	MMIO_D(PIPE_CSC_POSTOFF_HI(PIPE_A));
+	MMIO_D(PIPE_CSC_POSTOFF_ME(PIPE_A));
+	MMIO_D(PIPE_CSC_POSTOFF_LO(PIPE_A));
+	MMIO_D(PIPE_CSC_COEFF_RY_GY(PIPE_B));
+	MMIO_D(PIPE_CSC_COEFF_BY(PIPE_B));
+	MMIO_D(PIPE_CSC_COEFF_RU_GU(PIPE_B));
+	MMIO_D(PIPE_CSC_COEFF_BU(PIPE_B));
+	MMIO_D(PIPE_CSC_COEFF_RV_GV(PIPE_B));
+	MMIO_D(PIPE_CSC_COEFF_BV(PIPE_B));
+	MMIO_D(PIPE_CSC_MODE(PIPE_B));
+	MMIO_D(PIPE_CSC_PREOFF_HI(PIPE_B));
+	MMIO_D(PIPE_CSC_PREOFF_ME(PIPE_B));
+	MMIO_D(PIPE_CSC_PREOFF_LO(PIPE_B));
+	MMIO_D(PIPE_CSC_POSTOFF_HI(PIPE_B));
+	MMIO_D(PIPE_CSC_POSTOFF_ME(PIPE_B));
+	MMIO_D(PIPE_CSC_POSTOFF_LO(PIPE_B));
+	MMIO_D(PIPE_CSC_COEFF_RY_GY(PIPE_C));
+	MMIO_D(PIPE_CSC_COEFF_BY(PIPE_C));
+	MMIO_D(PIPE_CSC_COEFF_RU_GU(PIPE_C));
+	MMIO_D(PIPE_CSC_COEFF_BU(PIPE_C));
+	MMIO_D(PIPE_CSC_COEFF_RV_GV(PIPE_C));
+	MMIO_D(PIPE_CSC_COEFF_BV(PIPE_C));
+	MMIO_D(PIPE_CSC_MODE(PIPE_C));
+	MMIO_D(PIPE_CSC_PREOFF_HI(PIPE_C));
+	MMIO_D(PIPE_CSC_PREOFF_ME(PIPE_C));
+	MMIO_D(PIPE_CSC_PREOFF_LO(PIPE_C));
+	MMIO_D(PIPE_CSC_POSTOFF_HI(PIPE_C));
+	MMIO_D(PIPE_CSC_POSTOFF_ME(PIPE_C));
+	MMIO_D(PIPE_CSC_POSTOFF_LO(PIPE_C));
+	MMIO_D(PREC_PAL_INDEX(PIPE_A));
+	MMIO_D(PREC_PAL_DATA(PIPE_A));
+	MMIO_F(PREC_PAL_GC_MAX(PIPE_A, 0), 4 * 3);
+	MMIO_D(PREC_PAL_INDEX(PIPE_B));
+	MMIO_D(PREC_PAL_DATA(PIPE_B));
+	MMIO_F(PREC_PAL_GC_MAX(PIPE_B, 0), 4 * 3);
+	MMIO_D(PREC_PAL_INDEX(PIPE_C));
+	MMIO_D(PREC_PAL_DATA(PIPE_C));
+	MMIO_F(PREC_PAL_GC_MAX(PIPE_C, 0), 4 * 3);
+	MMIO_D(_MMIO(0x60110));
+	MMIO_D(_MMIO(0x61110));
+	MMIO_F(_MMIO(0x70400), 0x40);
+	MMIO_F(_MMIO(0x71400), 0x40);
+	MMIO_F(_MMIO(0x72400), 0x40);
+	MMIO_D(WM_LINETIME(PIPE_A));
+	MMIO_D(WM_LINETIME(PIPE_B));
+	MMIO_D(WM_LINETIME(PIPE_C));
+	MMIO_D(SPLL_CTL);
+	MMIO_D(_MMIO(_WRPLL_CTL1));
+	MMIO_D(_MMIO(_WRPLL_CTL2));
+	MMIO_D(PORT_CLK_SEL(PORT_A));
+	MMIO_D(PORT_CLK_SEL(PORT_B));
+	MMIO_D(PORT_CLK_SEL(PORT_C));
+	MMIO_D(PORT_CLK_SEL(PORT_D));
+	MMIO_D(PORT_CLK_SEL(PORT_E));
+	MMIO_D(TRANS_CLK_SEL(TRANSCODER_A));
+	MMIO_D(TRANS_CLK_SEL(TRANSCODER_B));
+	MMIO_D(TRANS_CLK_SEL(TRANSCODER_C));
+	MMIO_D(HSW_NDE_RSTWRN_OPT);
+	MMIO_D(_MMIO(0x46508));
+	MMIO_D(_MMIO(0x49080));
+	MMIO_D(_MMIO(0x49180));
+	MMIO_D(_MMIO(0x49280));
+	MMIO_F(_MMIO(0x49090), 0x14);
+	MMIO_F(_MMIO(0x49190), 0x14);
+	MMIO_F(_MMIO(0x49290), 0x14);
+	MMIO_D(GAMMA_MODE(PIPE_A));
+	MMIO_D(GAMMA_MODE(PIPE_B));
+	MMIO_D(GAMMA_MODE(PIPE_C));
+	MMIO_D(PIPE_MULT(PIPE_A));
+	MMIO_D(PIPE_MULT(PIPE_B));
+	MMIO_D(PIPE_MULT(PIPE_C));
+	MMIO_D(HSW_TVIDEO_DIP_CTL(TRANSCODER_A));
+	MMIO_D(HSW_TVIDEO_DIP_CTL(TRANSCODER_B));
+	MMIO_D(HSW_TVIDEO_DIP_CTL(TRANSCODER_C));
+	MMIO_D(SFUSE_STRAP);
+	MMIO_D(SBI_ADDR);
+	MMIO_D(SBI_DATA);
+	MMIO_D(SBI_CTL_STAT);
+	MMIO_D(PIXCLK_GATE);
+	MMIO_F(_MMIO(_DPA_AUX_CH_CTL), 6 * 4);
+	MMIO_D(DDI_BUF_CTL(PORT_A));
+	MMIO_D(DDI_BUF_CTL(PORT_B));
+	MMIO_D(DDI_BUF_CTL(PORT_C));
+	MMIO_D(DDI_BUF_CTL(PORT_D));
+	MMIO_D(DDI_BUF_CTL(PORT_E));
+	MMIO_D(DP_TP_CTL(PORT_A));
+	MMIO_D(DP_TP_CTL(PORT_B));
+	MMIO_D(DP_TP_CTL(PORT_C));
+	MMIO_D(DP_TP_CTL(PORT_D));
+	MMIO_D(DP_TP_CTL(PORT_E));
+	MMIO_D(DP_TP_STATUS(PORT_A));
+	MMIO_D(DP_TP_STATUS(PORT_B));
+	MMIO_D(DP_TP_STATUS(PORT_C));
+	MMIO_D(DP_TP_STATUS(PORT_D));
+	MMIO_D(DP_TP_STATUS(PORT_E));
+	MMIO_F(_MMIO(_DDI_BUF_TRANS_A), 0x50);
+	MMIO_F(_MMIO(0x64e60), 0x50);
+	MMIO_F(_MMIO(0x64eC0), 0x50);
+	MMIO_F(_MMIO(0x64f20), 0x50);
+	MMIO_F(_MMIO(0x64f80), 0x50);
+	MMIO_D(HSW_AUD_CFG(PIPE_A));
+	MMIO_D(HSW_AUD_PIN_ELD_CP_VLD);
+	MMIO_D(HSW_AUD_MISC_CTRL(PIPE_A));
+	MMIO_D(_MMIO(_TRANS_DDI_FUNC_CTL_A));
+	MMIO_D(_MMIO(_TRANS_DDI_FUNC_CTL_B));
+	MMIO_D(_MMIO(_TRANS_DDI_FUNC_CTL_C));
+	MMIO_D(_MMIO(_TRANS_DDI_FUNC_CTL_EDP));
+	MMIO_D(_MMIO(_TRANSA_MSA_MISC));
+	MMIO_D(_MMIO(_TRANSB_MSA_MISC));
+	MMIO_D(_MMIO(_TRANSC_MSA_MISC));
+	MMIO_D(_MMIO(_TRANS_EDP_MSA_MISC));
+	MMIO_D(FORCEWAKE);
+	MMIO_D(FORCEWAKE_ACK);
+	MMIO_D(GEN6_GT_CORE_STATUS);
+	MMIO_D(GEN6_GT_THREAD_STATUS_REG);
+	MMIO_D(GTFIFODBG);
+	MMIO_D(GTFIFOCTL);
+	MMIO_D(ECOBUS);
+	MMIO_D(GEN6_RC_CONTROL);
+	MMIO_D(GEN6_RC_STATE);
+	MMIO_D(GEN6_RPNSWREQ);
+	MMIO_D(GEN6_RC_VIDEO_FREQ);
+	MMIO_D(GEN6_RP_DOWN_TIMEOUT);
+	MMIO_D(GEN6_RP_INTERRUPT_LIMITS);
+	MMIO_D(GEN6_RPSTAT1);
+	MMIO_D(GEN6_RP_CONTROL);
+	MMIO_D(GEN6_RP_UP_THRESHOLD);
+	MMIO_D(GEN6_RP_DOWN_THRESHOLD);
+	MMIO_D(GEN6_RP_CUR_UP_EI);
+	MMIO_D(GEN6_RP_CUR_UP);
+	MMIO_D(GEN6_RP_PREV_UP);
+	MMIO_D(GEN6_RP_CUR_DOWN_EI);
+	MMIO_D(GEN6_RP_CUR_DOWN);
+	MMIO_D(GEN6_RP_PREV_DOWN);
+	MMIO_D(GEN6_RP_UP_EI);
+	MMIO_D(GEN6_RP_DOWN_EI);
+	MMIO_D(GEN6_RP_IDLE_HYSTERSIS);
+	MMIO_D(GEN6_RC1_WAKE_RATE_LIMIT);
+	MMIO_D(GEN6_RC6_WAKE_RATE_LIMIT);
+	MMIO_D(GEN6_RC6pp_WAKE_RATE_LIMIT);
+	MMIO_D(GEN6_RC_EVALUATION_INTERVAL);
+	MMIO_D(GEN6_RC_IDLE_HYSTERSIS);
+	MMIO_D(GEN6_RC_SLEEP);
+	MMIO_D(GEN6_RC1e_THRESHOLD);
+	MMIO_D(GEN6_RC6_THRESHOLD);
+	MMIO_D(GEN6_RC6p_THRESHOLD);
+	MMIO_D(GEN6_RC6pp_THRESHOLD);
+	MMIO_D(GEN6_PMINTRMSK);
+
+	MMIO_D(RSTDBYCTL);
+	MMIO_D(GEN6_GDRST);
+	MMIO_F(FENCE_REG_GEN6_LO(0), 0x80);
+	MMIO_D(CPU_VGACNTRL);
+	MMIO_D(TILECTL);
+	MMIO_D(GEN6_UCGCTL1);
+	MMIO_D(GEN6_UCGCTL2);
+	MMIO_F(_MMIO(0x4f000), 0x90);
+	MMIO_D(GEN6_PCODE_DATA);
+	MMIO_D(_MMIO(0x13812c));
+	MMIO_D(GEN7_ERR_INT);
+	MMIO_D(HSW_EDRAM_CAP);
+	MMIO_D(HSW_IDICR);
+	MMIO_D(GFX_FLSH_CNTL_GEN6);
+	MMIO_D(_MMIO(0x3c));
+	MMIO_D(_MMIO(0x860));
+	MMIO_D(ECOSKPD(RENDER_RING_BASE));
+	MMIO_D(_MMIO(0x121d0));
+	MMIO_D(ECOSKPD(BLT_RING_BASE));
+	MMIO_D(_MMIO(0x41d0));
+	MMIO_D(GAC_ECO_BITS);
+	MMIO_D(_MMIO(0x6200));
+	MMIO_D(_MMIO(0x6204));
+	MMIO_D(_MMIO(0x6208));
+	MMIO_D(_MMIO(0x7118));
+	MMIO_D(_MMIO(0x7180));
+	MMIO_D(_MMIO(0x7408));
+	MMIO_D(_MMIO(0x7c00));
+	MMIO_D(GEN6_MBCTL);
+	MMIO_D(_MMIO(0x911c));
+	MMIO_D(_MMIO(0x9120));
+	MMIO_D(GEN7_UCGCTL4);
+	MMIO_D(GAB_CTL);
+	MMIO_D(_MMIO(0x48800));
+	MMIO_D(_MMIO(0xce044));
+	MMIO_D(_MMIO(0xe6500));
+	MMIO_D(_MMIO(0xe6504));
+	MMIO_D(_MMIO(0xe6600));
+	MMIO_D(_MMIO(0xe6604));
+	MMIO_D(_MMIO(0xe6700));
+	MMIO_D(_MMIO(0xe6704));
+	MMIO_D(_MMIO(0xe6800));
+	MMIO_D(_MMIO(0xe6804));
+	MMIO_D(PCH_GMBUS4);
+	MMIO_D(PCH_GMBUS5);
+	MMIO_D(_MMIO(0x902c));
+	MMIO_D(_MMIO(0xec008));
+	MMIO_D(_MMIO(0xec00c));
+	MMIO_D(_MMIO(0xec008 + 0x18));
+	MMIO_D(_MMIO(0xec00c + 0x18));
+	MMIO_D(_MMIO(0xec008 + 0x18 * 2));
+	MMIO_D(_MMIO(0xec00c + 0x18 * 2));
+	MMIO_D(_MMIO(0xec008 + 0x18 * 3));
+	MMIO_D(_MMIO(0xec00c + 0x18 * 3));
+	MMIO_D(_MMIO(0xec408));
+	MMIO_D(_MMIO(0xec40c));
+	MMIO_D(_MMIO(0xec408 + 0x18));
+	MMIO_D(_MMIO(0xec40c + 0x18));
+	MMIO_D(_MMIO(0xec408 + 0x18 * 2));
+	MMIO_D(_MMIO(0xec40c + 0x18 * 2));
+	MMIO_D(_MMIO(0xec408 + 0x18 * 3));
+	MMIO_D(_MMIO(0xec40c + 0x18 * 3));
+	MMIO_D(_MMIO(0xfc810));
+	MMIO_D(_MMIO(0xfc81c));
+	MMIO_D(_MMIO(0xfc828));
+	MMIO_D(_MMIO(0xfc834));
+	MMIO_D(_MMIO(0xfcc00));
+	MMIO_D(_MMIO(0xfcc0c));
+	MMIO_D(_MMIO(0xfcc18));
+	MMIO_D(_MMIO(0xfcc24));
+	MMIO_D(_MMIO(0xfd000));
+	MMIO_D(_MMIO(0xfd00c));
+	MMIO_D(_MMIO(0xfd018));
+	MMIO_D(_MMIO(0xfd024));
+	MMIO_D(_MMIO(0xfd034));
+	MMIO_D(FPGA_DBG);
+	MMIO_D(_MMIO(0x2054));
+	MMIO_D(_MMIO(0x12054));
+	MMIO_D(_MMIO(0x22054));
+	MMIO_D(_MMIO(0x1a054));
+	MMIO_D(_MMIO(0x44070));
+	MMIO_D(_MMIO(0x2178));
+	MMIO_D(_MMIO(0x217c));
+	MMIO_D(_MMIO(0x12178));
+	MMIO_D(_MMIO(0x1217c));
+	MMIO_F(_MMIO(0x5200), 32);
+	MMIO_F(_MMIO(0x5240), 32);
+	MMIO_F(_MMIO(0x5280), 16);
+	MMIO_D(BCS_SWCTRL);
+	MMIO_F(HS_INVOCATION_COUNT, 8);
+	MMIO_F(DS_INVOCATION_COUNT, 8);
+	MMIO_F(IA_VERTICES_COUNT, 8);
+	MMIO_F(IA_PRIMITIVES_COUNT, 8);
+	MMIO_F(VS_INVOCATION_COUNT, 8);
+	MMIO_F(GS_INVOCATION_COUNT, 8);
+	MMIO_F(GS_PRIMITIVES_COUNT, 8);
+	MMIO_F(CL_INVOCATION_COUNT, 8);
+	MMIO_F(CL_PRIMITIVES_COUNT, 8);
+	MMIO_F(PS_INVOCATION_COUNT, 8);
+	MMIO_F(PS_DEPTH_COUNT, 8);
+	MMIO_D(ARB_MODE);
+	MMIO_RING_D(RING_BBADDR);
+	MMIO_D(_MMIO(0x2220));
+	MMIO_D(_MMIO(0x12220));
+	MMIO_D(_MMIO(0x22220));
+	MMIO_RING_D(RING_SYNC_1);
+	MMIO_RING_D(RING_SYNC_0);
+	MMIO_D(GUC_STATUS);
+
+	MMIO_F(_MMIO(MCHBAR_MIRROR_BASE_SNB), 0x40000);
+	MMIO_F(_MMIO(VGT_PVINFO_PAGE), VGT_PVINFO_SIZE);
+	MMIO_F(LGC_PALETTE(PIPE_A, 0), 1024);
+	MMIO_F(LGC_PALETTE(PIPE_B, 0), 1024);
+	MMIO_F(LGC_PALETTE(PIPE_C, 0), 1024);
+
+	return 0;
+}
+
+static int iterate_bdw_only_mmio(struct intel_gvt_mmio_table_iter *iter)
+{
+	MMIO_D(HSW_PWR_WELL_CTL1);
+	MMIO_D(HSW_PWR_WELL_CTL2);
+	MMIO_D(HSW_PWR_WELL_CTL3);
+	MMIO_D(HSW_PWR_WELL_CTL4);
+	MMIO_D(HSW_PWR_WELL_CTL5);
+	MMIO_D(HSW_PWR_WELL_CTL6);
+
+	MMIO_D(WM_MISC);
+	MMIO_D(_MMIO(_SRD_CTL_EDP));
+
+	MMIO_D(_MMIO(0xb1f0));
+	MMIO_D(_MMIO(0xb1c0));
+	MMIO_D(_MMIO(0xb100));
+	MMIO_D(_MMIO(0xb10c));
+	MMIO_D(_MMIO(0xb110));
+	MMIO_D(_MMIO(0x83a4));
+	MMIO_D(_MMIO(0x8430));
+	MMIO_D(_MMIO(0x2248));
+	MMIO_D(FORCEWAKE_ACK_HSW);
+
+	return 0;
+}
+
+static int iterate_bdw_plus_mmio(struct intel_gvt_mmio_table_iter *iter)
+{
+	struct drm_i915_private *dev_priv = iter->i915;
+
+	MMIO_D(GEN8_GT_IMR(0));
+	MMIO_D(GEN8_GT_IER(0));
+	MMIO_D(GEN8_GT_IIR(0));
+	MMIO_D(GEN8_GT_ISR(0));
+	MMIO_D(GEN8_GT_IMR(1));
+	MMIO_D(GEN8_GT_IER(1));
+	MMIO_D(GEN8_GT_IIR(1));
+	MMIO_D(GEN8_GT_ISR(1));
+	MMIO_D(GEN8_GT_IMR(2));
+	MMIO_D(GEN8_GT_IER(2));
+	MMIO_D(GEN8_GT_IIR(2));
+	MMIO_D(GEN8_GT_ISR(2));
+	MMIO_D(GEN8_GT_IMR(3));
+	MMIO_D(GEN8_GT_IER(3));
+	MMIO_D(GEN8_GT_IIR(3));
+	MMIO_D(GEN8_GT_ISR(3));
+	MMIO_D(GEN8_DE_PIPE_IMR(PIPE_A));
+	MMIO_D(GEN8_DE_PIPE_IER(PIPE_A));
+	MMIO_D(GEN8_DE_PIPE_IIR(PIPE_A));
+	MMIO_D(GEN8_DE_PIPE_ISR(PIPE_A));
+	MMIO_D(GEN8_DE_PIPE_IMR(PIPE_B));
+	MMIO_D(GEN8_DE_PIPE_IER(PIPE_B));
+	MMIO_D(GEN8_DE_PIPE_IIR(PIPE_B));
+	MMIO_D(GEN8_DE_PIPE_ISR(PIPE_B));
+	MMIO_D(GEN8_DE_PIPE_IMR(PIPE_C));
+	MMIO_D(GEN8_DE_PIPE_IER(PIPE_C));
+	MMIO_D(GEN8_DE_PIPE_IIR(PIPE_C));
+	MMIO_D(GEN8_DE_PIPE_ISR(PIPE_C));
+	MMIO_D(GEN8_DE_PORT_IMR);
+	MMIO_D(GEN8_DE_PORT_IER);
+	MMIO_D(GEN8_DE_PORT_IIR);
+	MMIO_D(GEN8_DE_PORT_ISR);
+	MMIO_D(GEN8_DE_MISC_IMR);
+	MMIO_D(GEN8_DE_MISC_IER);
+	MMIO_D(GEN8_DE_MISC_IIR);
+	MMIO_D(GEN8_DE_MISC_ISR);
+	MMIO_D(GEN8_PCU_IMR);
+	MMIO_D(GEN8_PCU_IER);
+	MMIO_D(GEN8_PCU_IIR);
+	MMIO_D(GEN8_PCU_ISR);
+	MMIO_D(GEN8_MASTER_IRQ);
+	MMIO_RING_D(RING_ACTHD_UDW);
+
+#define RING_REG(base) _MMIO((base) + 0xd0)
+	MMIO_RING_D(RING_REG);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x230)
+	MMIO_RING_D(RING_REG);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x234)
+	MMIO_RING_F(RING_REG, 8);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x244)
+	MMIO_RING_D(RING_REG);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x370)
+	MMIO_RING_F(RING_REG, 48);
+#undef RING_REG
+
+#define RING_REG(base) _MMIO((base) + 0x3a0)
+	MMIO_RING_D(RING_REG);
+#undef RING_REG
+
+	MMIO_D(PIPEMISC(PIPE_A));
+	MMIO_D(PIPEMISC(PIPE_B));
+	MMIO_D(PIPEMISC(PIPE_C));
+	MMIO_D(_MMIO(0x1c1d0));
+	MMIO_D(GEN6_MBCUNIT_SNPCR);
+	MMIO_D(GEN7_MISCCPCTL);
+	MMIO_D(_MMIO(0x1c054));
+	MMIO_D(GEN6_PCODE_MAILBOX);
+	if (!IS_BROXTON(dev_priv))
+		MMIO_D(GEN8_PRIVATE_PAT_LO);
+	MMIO_D(GEN8_PRIVATE_PAT_HI);
+	MMIO_D(GAMTARBMODE);
+
+#define RING_REG(base) _MMIO((base) + 0x270)
+	MMIO_RING_F(RING_REG, 32);
+#undef RING_REG
+
+	MMIO_RING_D(RING_HWS_PGA);
+	MMIO_D(HDC_CHICKEN0);
+	MMIO_D(CHICKEN_PIPESL_1(PIPE_A));
+	MMIO_D(CHICKEN_PIPESL_1(PIPE_B));
+	MMIO_D(CHICKEN_PIPESL_1(PIPE_C));
+	MMIO_D(_MMIO(0x6671c));
+	MMIO_D(_MMIO(0x66c00));
+	MMIO_D(_MMIO(0x66c04));
+	MMIO_D(HSW_GTT_CACHE_EN);
+	MMIO_D(GEN8_EU_DISABLE0);
+	MMIO_D(GEN8_EU_DISABLE1);
+	MMIO_D(GEN8_EU_DISABLE2);
+	MMIO_D(_MMIO(0xfdc));
+	MMIO_D(GEN8_ROW_CHICKEN);
+	MMIO_D(GEN7_ROW_CHICKEN2);
+	MMIO_D(GEN8_UCGCTL6);
+	MMIO_D(GEN8_L3SQCREG4);
+	MMIO_D(GEN9_SCRATCH_LNCF1);
+	MMIO_F(_MMIO(0x24d0), 48);
+	MMIO_D(_MMIO(0x44484));
+	MMIO_D(_MMIO(0x4448c));
+	MMIO_D(GEN8_L3_LRA_1_GPGPU);
+	MMIO_D(_MMIO(0x110000));
+	MMIO_D(_MMIO(0x48400));
+	MMIO_D(_MMIO(0x6e570));
+	MMIO_D(_MMIO(0x65f10));
+	MMIO_D(_MMIO(0xe194));
+	MMIO_D(_MMIO(0xe188));
+	MMIO_D(HALF_SLICE_CHICKEN2);
+	MMIO_D(_MMIO(0x2580));
+	MMIO_D(_MMIO(0xe220));
+	MMIO_D(_MMIO(0xe230));
+	MMIO_D(_MMIO(0xe240));
+	MMIO_D(_MMIO(0xe260));
+	MMIO_D(_MMIO(0xe270));
+	MMIO_D(_MMIO(0xe280));
+	MMIO_D(_MMIO(0xe2a0));
+	MMIO_D(_MMIO(0xe2b0));
+	MMIO_D(_MMIO(0xe2c0));
+	MMIO_D(_MMIO(0x21f0));
+	MMIO_D(GEN8_GAMW_ECO_DEV_RW_IA);
+	MMIO_D(_MMIO(0x215c));
+	MMIO_F(_MMIO(0x2290), 8);
+	MMIO_D(_MMIO(0x2b00));
+	MMIO_D(_MMIO(0x2360));
+	MMIO_D(_MMIO(0x1c17c));
+	MMIO_D(_MMIO(0x1c178));
+	MMIO_D(_MMIO(0x4260));
+	MMIO_D(_MMIO(0x4264));
+	MMIO_D(_MMIO(0x4268));
+	MMIO_D(_MMIO(0x426c));
+	MMIO_D(_MMIO(0x4270));
+	MMIO_D(_MMIO(0x4094));
+	MMIO_D(_MMIO(0x22178));
+	MMIO_D(_MMIO(0x1a178));
+	MMIO_D(_MMIO(0x1a17c));
+	MMIO_D(_MMIO(0x2217c));
+	MMIO_D(EDP_PSR_IMR);
+	MMIO_D(EDP_PSR_IIR);
+	MMIO_D(_MMIO(0xe4cc));
+	MMIO_D(GEN7_SC_INSTDONE);
+
+	return 0;
+}
+
+static int iterate_pre_skl_mmio(struct intel_gvt_mmio_table_iter *iter)
+{
+	MMIO_D(FORCEWAKE_MT);
+
+	MMIO_D(PCH_ADPA);
+	MMIO_F(_MMIO(_PCH_DPB_AUX_CH_CTL), 6 * 4);
+	MMIO_F(_MMIO(_PCH_DPC_AUX_CH_CTL), 6 * 4);
+	MMIO_F(_MMIO(_PCH_DPD_AUX_CH_CTL), 6 * 4);
+
+	MMIO_F(_MMIO(0x70440), 0xc);
+	MMIO_F(_MMIO(0x71440), 0xc);
+	MMIO_F(_MMIO(0x72440), 0xc);
+	MMIO_F(_MMIO(0x7044c), 0xc);
+	MMIO_F(_MMIO(0x7144c), 0xc);
+	MMIO_F(_MMIO(0x7244c), 0xc);
+
+	return 0;
+}
+
+static int iterate_skl_plus_mmio(struct intel_gvt_mmio_table_iter *iter)
+{
+	struct drm_i915_private *dev_priv = iter->i915;
+
+	MMIO_D(FORCEWAKE_RENDER_GEN9);
+	MMIO_D(FORCEWAKE_ACK_RENDER_GEN9);
+	MMIO_D(FORCEWAKE_GT_GEN9);
+	MMIO_D(FORCEWAKE_ACK_GT_GEN9);
+	MMIO_D(FORCEWAKE_MEDIA_GEN9);
+	MMIO_D(FORCEWAKE_ACK_MEDIA_GEN9);
+	MMIO_F(DP_AUX_CH_CTL(AUX_CH_B), 6 * 4);
+	MMIO_F(DP_AUX_CH_CTL(AUX_CH_C), 6 * 4);
+	MMIO_F(DP_AUX_CH_CTL(AUX_CH_D), 6 * 4);
+	MMIO_D(HSW_PWR_WELL_CTL1);
+	MMIO_D(HSW_PWR_WELL_CTL2);
+	MMIO_D(DBUF_CTL_S(0));
+	MMIO_D(GEN9_PG_ENABLE);
+	MMIO_D(GEN9_MEDIA_PG_IDLE_HYSTERESIS);
+	MMIO_D(GEN9_RENDER_PG_IDLE_HYSTERESIS);
+	MMIO_D(GEN9_GAMT_ECO_REG_RW_IA);
+	MMIO_D(MMCD_MISC_CTRL);
+	MMIO_D(CHICKEN_PAR1_1);
+	MMIO_D(DC_STATE_EN);
+	MMIO_D(DC_STATE_DEBUG);
+	MMIO_D(CDCLK_CTL);
+	MMIO_D(LCPLL1_CTL);
+	MMIO_D(LCPLL2_CTL);
+	MMIO_D(_MMIO(_DPLL1_CFGCR1));
+	MMIO_D(_MMIO(_DPLL2_CFGCR1));
+	MMIO_D(_MMIO(_DPLL3_CFGCR1));
+	MMIO_D(_MMIO(_DPLL1_CFGCR2));
+	MMIO_D(_MMIO(_DPLL2_CFGCR2));
+	MMIO_D(_MMIO(_DPLL3_CFGCR2));
+	MMIO_D(DPLL_CTRL1);
+	MMIO_D(DPLL_CTRL2);
+	MMIO_D(DPLL_STATUS);
+	MMIO_D(SKL_PS_WIN_POS(PIPE_A, 0));
+	MMIO_D(SKL_PS_WIN_POS(PIPE_A, 1));
+	MMIO_D(SKL_PS_WIN_POS(PIPE_B, 0));
+	MMIO_D(SKL_PS_WIN_POS(PIPE_B, 1));
+	MMIO_D(SKL_PS_WIN_POS(PIPE_C, 0));
+	MMIO_D(SKL_PS_WIN_POS(PIPE_C, 1));
+	MMIO_D(SKL_PS_WIN_SZ(PIPE_A, 0));
+	MMIO_D(SKL_PS_WIN_SZ(PIPE_A, 1));
+	MMIO_D(SKL_PS_WIN_SZ(PIPE_B, 0));
+	MMIO_D(SKL_PS_WIN_SZ(PIPE_B, 1));
+	MMIO_D(SKL_PS_WIN_SZ(PIPE_C, 0));
+	MMIO_D(SKL_PS_WIN_SZ(PIPE_C, 1));
+	MMIO_D(SKL_PS_CTRL(PIPE_A, 0));
+	MMIO_D(SKL_PS_CTRL(PIPE_A, 1));
+	MMIO_D(SKL_PS_CTRL(PIPE_B, 0));
+	MMIO_D(SKL_PS_CTRL(PIPE_B, 1));
+	MMIO_D(SKL_PS_CTRL(PIPE_C, 0));
+	MMIO_D(SKL_PS_CTRL(PIPE_C, 1));
+	MMIO_D(PLANE_BUF_CFG(PIPE_A, 0));
+	MMIO_D(PLANE_BUF_CFG(PIPE_A, 1));
+	MMIO_D(PLANE_BUF_CFG(PIPE_A, 2));
+	MMIO_D(PLANE_BUF_CFG(PIPE_A, 3));
+	MMIO_D(PLANE_BUF_CFG(PIPE_B, 0));
+	MMIO_D(PLANE_BUF_CFG(PIPE_B, 1));
+	MMIO_D(PLANE_BUF_CFG(PIPE_B, 2));
+	MMIO_D(PLANE_BUF_CFG(PIPE_B, 3));
+	MMIO_D(PLANE_BUF_CFG(PIPE_C, 0));
+	MMIO_D(PLANE_BUF_CFG(PIPE_C, 1));
+	MMIO_D(PLANE_BUF_CFG(PIPE_C, 2));
+	MMIO_D(PLANE_BUF_CFG(PIPE_C, 3));
+	MMIO_D(CUR_BUF_CFG(PIPE_A));
+	MMIO_D(CUR_BUF_CFG(PIPE_B));
+	MMIO_D(CUR_BUF_CFG(PIPE_C));
+	MMIO_F(PLANE_WM(PIPE_A, 0, 0), 4 * 8);
+	MMIO_F(PLANE_WM(PIPE_A, 1, 0), 4 * 8);
+	MMIO_F(PLANE_WM(PIPE_A, 2, 0), 4 * 8);
+	MMIO_F(PLANE_WM(PIPE_B, 0, 0), 4 * 8);
+	MMIO_F(PLANE_WM(PIPE_B, 1, 0), 4 * 8);
+	MMIO_F(PLANE_WM(PIPE_B, 2, 0), 4 * 8);
+	MMIO_F(PLANE_WM(PIPE_C, 0, 0), 4 * 8);
+	MMIO_F(PLANE_WM(PIPE_C, 1, 0), 4 * 8);
+	MMIO_F(PLANE_WM(PIPE_C, 2, 0), 4 * 8);
+	MMIO_F(CUR_WM(PIPE_A, 0), 4 * 8);
+	MMIO_F(CUR_WM(PIPE_B, 0), 4 * 8);
+	MMIO_F(CUR_WM(PIPE_C, 0), 4 * 8);
+	MMIO_D(PLANE_WM_TRANS(PIPE_A, 0));
+	MMIO_D(PLANE_WM_TRANS(PIPE_A, 1));
+	MMIO_D(PLANE_WM_TRANS(PIPE_A, 2));
+	MMIO_D(PLANE_WM_TRANS(PIPE_B, 0));
+	MMIO_D(PLANE_WM_TRANS(PIPE_B, 1));
+	MMIO_D(PLANE_WM_TRANS(PIPE_B, 2));
+	MMIO_D(PLANE_WM_TRANS(PIPE_C, 0));
+	MMIO_D(PLANE_WM_TRANS(PIPE_C, 1));
+	MMIO_D(PLANE_WM_TRANS(PIPE_C, 2));
+	MMIO_D(CUR_WM_TRANS(PIPE_A));
+	MMIO_D(CUR_WM_TRANS(PIPE_B));
+	MMIO_D(CUR_WM_TRANS(PIPE_C));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_A, 0));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_A, 1));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_A, 2));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_A, 3));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_B, 0));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_B, 1));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_B, 2));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_B, 3));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_C, 0));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_C, 1));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_C, 2));
+	MMIO_D(PLANE_NV12_BUF_CFG(PIPE_C, 3));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_A, 1)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_A, 2)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_A, 3)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_A, 4)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_B, 1)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_B, 2)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_B, 3)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_B, 4)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_C, 1)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_C, 2)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_C, 3)));
+	MMIO_D(_MMIO(_REG_701C0(PIPE_C, 4)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_A, 1)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_A, 2)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_A, 3)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_A, 4)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_B, 1)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_B, 2)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_B, 3)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_B, 4)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_C, 1)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_C, 2)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_C, 3)));
+	MMIO_D(_MMIO(_REG_701C4(PIPE_C, 4)));
+	MMIO_D(_MMIO(_PLANE_CTL_3_A));
+	MMIO_D(_MMIO(_PLANE_CTL_3_B));
+	MMIO_D(_MMIO(0x72380));
+	MMIO_D(_MMIO(0x7239c));
+	MMIO_D(_MMIO(_PLANE_SURF_3_A));
+	MMIO_D(_MMIO(_PLANE_SURF_3_B));
+	MMIO_D(DMC_SSP_BASE);
+	MMIO_D(DMC_HTP_SKL);
+	MMIO_D(DMC_LAST_WRITE);
+	MMIO_D(BDW_SCRATCH1);
+	MMIO_D(SKL_DFSM);
+	MMIO_D(DISPIO_CR_TX_BMU_CR0);
+	MMIO_F(GEN9_GFX_MOCS(0), 0x7f8);
+	MMIO_F(GEN7_L3CNTLREG2, 0x80);
+	MMIO_D(RPM_CONFIG0);
+	MMIO_D(_MMIO(0xd08));
+	MMIO_D(RC6_LOCATION);
+	MMIO_D(GEN7_FF_SLICE_CS_CHICKEN1);
+	MMIO_D(GEN9_CS_DEBUG_MODE1);
+	/* TRTT */
+	MMIO_D(TRVATTL3PTRDW(0));
+	MMIO_D(TRVATTL3PTRDW(1));
+	MMIO_D(TRVATTL3PTRDW(2));
+	MMIO_D(TRVATTL3PTRDW(3));
+	MMIO_D(TRVADR);
+	MMIO_D(TRTTE);
+	MMIO_D(_MMIO(0x4dfc));
+	MMIO_D(_MMIO(0x46430));
+	MMIO_D(_MMIO(0x46520));
+	MMIO_D(_MMIO(0xc403c));
+	MMIO_D(GEN8_GARBCNTL);
+	MMIO_D(DMA_CTRL);
+	MMIO_D(_MMIO(0x65900));
+	MMIO_D(GEN6_STOLEN_RESERVED);
+	MMIO_D(_MMIO(0x4068));
+	MMIO_D(_MMIO(0x67054));
+	MMIO_D(_MMIO(0x6e560));
+	MMIO_D(_MMIO(0x6e554));
+	MMIO_D(_MMIO(0x2b20));
+	MMIO_D(_MMIO(0x65f00));
+	MMIO_D(_MMIO(0x65f08));
+	MMIO_D(_MMIO(0x320f0));
+	MMIO_D(_MMIO(0x70034));
+	MMIO_D(_MMIO(0x71034));
+	MMIO_D(_MMIO(0x72034));
+	MMIO_D(_MMIO(_PLANE_KEYVAL_1(PIPE_A)));
+	MMIO_D(_MMIO(_PLANE_KEYVAL_1(PIPE_B)));
+	MMIO_D(_MMIO(_PLANE_KEYVAL_1(PIPE_C)));
+	MMIO_D(_MMIO(_PLANE_KEYMAX_1(PIPE_A)));
+	MMIO_D(_MMIO(_PLANE_KEYMAX_1(PIPE_B)));
+	MMIO_D(_MMIO(_PLANE_KEYMAX_1(PIPE_C)));
+	MMIO_D(_MMIO(_PLANE_KEYMSK_1(PIPE_A)));
+	MMIO_D(_MMIO(_PLANE_KEYMSK_1(PIPE_B)));
+	MMIO_D(_MMIO(_PLANE_KEYMSK_1(PIPE_C)));
+	MMIO_D(_MMIO(0x44500));
+#define CSFE_CHICKEN1_REG(base) _MMIO((base) + 0xD4)
+	MMIO_RING_D(CSFE_CHICKEN1_REG);
+#undef CSFE_CHICKEN1_REG
+	MMIO_D(GEN8_HDC_CHICKEN1);
+	MMIO_D(GEN9_WM_CHICKEN3);
+
+	if (IS_KABYLAKE(dev_priv) ||
+	    IS_COFFEELAKE(dev_priv) || IS_COMETLAKE(dev_priv))
+		MMIO_D(GAMT_CHKN_BIT_REG);
+	if (!IS_BROXTON(dev_priv))
+		MMIO_D(GEN9_CTX_PREEMPT_REG);
+	MMIO_F(_MMIO(DMC_MMIO_START_RANGE), 0x3000);
+	return 0;
+}
+
+static int iterate_bxt_mmio(struct intel_gvt_mmio_table_iter *iter)
+{
+	struct drm_i915_private *dev_priv = iter->i915;
+
+	MMIO_F(_MMIO(0x80000), 0x3000);
+	MMIO_D(GEN7_SAMPLER_INSTDONE);
+	MMIO_D(GEN7_ROW_INSTDONE);
+	MMIO_D(GEN8_FAULT_TLB_DATA0);
+	MMIO_D(GEN8_FAULT_TLB_DATA1);
+	MMIO_D(ERROR_GEN6);
+	MMIO_D(DONE_REG);
+	MMIO_D(EIR);
+	MMIO_D(PGTBL_ER);
+	MMIO_D(_MMIO(0x4194));
+	MMIO_D(_MMIO(0x4294));
+	MMIO_D(_MMIO(0x4494));
+	MMIO_RING_D(RING_PSMI_CTL);
+	MMIO_RING_D(RING_DMA_FADD);
+	MMIO_RING_D(RING_DMA_FADD_UDW);
+	MMIO_RING_D(RING_IPEHR);
+	MMIO_RING_D(RING_INSTPS);
+	MMIO_RING_D(RING_BBADDR_UDW);
+	MMIO_RING_D(RING_BBSTATE);
+	MMIO_RING_D(RING_IPEIR);
+	MMIO_F(SOFT_SCRATCH(0), 16 * 4);
+	MMIO_D(BXT_P_CR_GT_DISP_PWRON);
+	MMIO_D(BXT_RP_STATE_CAP);
+	MMIO_D(BXT_PHY_CTL_FAMILY(DPIO_PHY0));
+	MMIO_D(BXT_PHY_CTL_FAMILY(DPIO_PHY1));
+	MMIO_D(BXT_PHY_CTL(PORT_A));
+	MMIO_D(BXT_PHY_CTL(PORT_B));
+	MMIO_D(BXT_PHY_CTL(PORT_C));
+	MMIO_D(BXT_PORT_PLL_ENABLE(PORT_A));
+	MMIO_D(BXT_PORT_PLL_ENABLE(PORT_B));
+	MMIO_D(BXT_PORT_PLL_ENABLE(PORT_C));
+	MMIO_D(BXT_PORT_CL1CM_DW0(DPIO_PHY0));
+	MMIO_D(BXT_PORT_CL1CM_DW9(DPIO_PHY0));
+	MMIO_D(BXT_PORT_CL1CM_DW10(DPIO_PHY0));
+	MMIO_D(BXT_PORT_CL1CM_DW28(DPIO_PHY0));
+	MMIO_D(BXT_PORT_CL1CM_DW30(DPIO_PHY0));
+	MMIO_D(BXT_PORT_CL2CM_DW6(DPIO_PHY0));
+	MMIO_D(BXT_PORT_REF_DW3(DPIO_PHY0));
+	MMIO_D(BXT_PORT_REF_DW6(DPIO_PHY0));
+	MMIO_D(BXT_PORT_REF_DW8(DPIO_PHY0));
+	MMIO_D(BXT_PORT_CL1CM_DW0(DPIO_PHY1));
+	MMIO_D(BXT_PORT_CL1CM_DW9(DPIO_PHY1));
+	MMIO_D(BXT_PORT_CL1CM_DW10(DPIO_PHY1));
+	MMIO_D(BXT_PORT_CL1CM_DW28(DPIO_PHY1));
+	MMIO_D(BXT_PORT_CL1CM_DW30(DPIO_PHY1));
+	MMIO_D(BXT_PORT_CL2CM_DW6(DPIO_PHY1));
+	MMIO_D(BXT_PORT_REF_DW3(DPIO_PHY1));
+	MMIO_D(BXT_PORT_REF_DW6(DPIO_PHY1));
+	MMIO_D(BXT_PORT_REF_DW8(DPIO_PHY1));
+	MMIO_D(BXT_PORT_PLL_EBB_0(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_PLL_EBB_4(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_PCS_DW10_LN01(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_PCS_DW10_GRP(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_PCS_DW12_LN01(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_PCS_DW12_LN23(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_PCS_DW12_GRP(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW2_LN0(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW2_GRP(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW3_LN0(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW3_GRP(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW4_LN0(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW4_GRP(DPIO_PHY0, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY0, DPIO_CH0, 0));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY0, DPIO_CH0, 1));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY0, DPIO_CH0, 2));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY0, DPIO_CH0, 3));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH0, 0));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH0, 1));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH0, 2));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH0, 3));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH0, 6));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH0, 8));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH0, 9));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH0, 10));
+	MMIO_D(BXT_PORT_PLL_EBB_0(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_PLL_EBB_4(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_PCS_DW10_LN01(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_PCS_DW10_GRP(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_PCS_DW12_LN01(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_PCS_DW12_LN23(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_PCS_DW12_GRP(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_TX_DW2_LN0(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_TX_DW2_GRP(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_TX_DW3_LN0(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_TX_DW3_GRP(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_TX_DW4_LN0(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_TX_DW4_GRP(DPIO_PHY0, DPIO_CH1));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY0, DPIO_CH1, 0));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY0, DPIO_CH1, 1));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY0, DPIO_CH1, 2));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY0, DPIO_CH1, 3));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH1, 0));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH1, 1));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH1, 2));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH1, 3));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH1, 6));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH1, 8));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH1, 9));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY0, DPIO_CH1, 10));
+	MMIO_D(BXT_PORT_PLL_EBB_0(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_PLL_EBB_4(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_PCS_DW10_LN01(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_PCS_DW10_GRP(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_PCS_DW12_LN01(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_PCS_DW12_LN23(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_PCS_DW12_GRP(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW2_LN0(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW2_GRP(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW3_LN0(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW3_GRP(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW4_LN0(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW4_GRP(DPIO_PHY1, DPIO_CH0));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY1, DPIO_CH0, 0));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY1, DPIO_CH0, 1));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY1, DPIO_CH0, 2));
+	MMIO_D(BXT_PORT_TX_DW14_LN(DPIO_PHY1, DPIO_CH0, 3));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY1, DPIO_CH0, 0));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY1, DPIO_CH0, 1));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY1, DPIO_CH0, 2));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY1, DPIO_CH0, 3));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY1, DPIO_CH0, 6));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY1, DPIO_CH0, 8));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY1, DPIO_CH0, 9));
+	MMIO_D(BXT_PORT_PLL(DPIO_PHY1, DPIO_CH0, 10));
+	MMIO_D(BXT_DE_PLL_CTL);
+	MMIO_D(BXT_DE_PLL_ENABLE);
+	MMIO_D(BXT_DSI_PLL_CTL);
+	MMIO_D(BXT_DSI_PLL_ENABLE);
+	MMIO_D(GEN9_CLKGATE_DIS_0);
+	MMIO_D(GEN9_CLKGATE_DIS_4);
+	MMIO_D(HSW_TVIDEO_DIP_GCP(TRANSCODER_A));
+	MMIO_D(HSW_TVIDEO_DIP_GCP(TRANSCODER_B));
+	MMIO_D(HSW_TVIDEO_DIP_GCP(TRANSCODER_C));
+	MMIO_D(RC6_CTX_BASE);
+	MMIO_D(GEN8_PUSHBUS_CONTROL);
+	MMIO_D(GEN8_PUSHBUS_ENABLE);
+	MMIO_D(GEN8_PUSHBUS_SHIFT);
+	MMIO_D(GEN6_GFXPAUSE);
+	MMIO_D(GEN8_L3SQCREG1);
+	MMIO_D(GEN8_L3CNTLREG);
+	MMIO_D(_MMIO(0x20D8));
+	MMIO_F(GEN8_RING_CS_GPR(RENDER_RING_BASE, 0), 0x40);
+	MMIO_F(GEN8_RING_CS_GPR(GEN6_BSD_RING_BASE, 0), 0x40);
+	MMIO_F(GEN8_RING_CS_GPR(BLT_RING_BASE, 0), 0x40);
+	MMIO_F(GEN8_RING_CS_GPR(VEBOX_RING_BASE, 0), 0x40);
+	MMIO_D(GEN9_CTX_PREEMPT_REG);
+	MMIO_D(GEN8_PRIVATE_PAT_LO);
+
+	return 0;
+}
+
+/**
+ * intel_gvt_iterate_mmio_table - Iterate the GVT MMIO table
+ * @iter: the interator
+ *
+ * This function is called for iterating the GVT MMIO table when i915 is
+ * taking the snapshot of the HW and GVT is building MMIO tracking table.
+ */
+int intel_gvt_iterate_mmio_table(struct intel_gvt_mmio_table_iter *iter)
+{
+	struct drm_i915_private *i915 = iter->i915;
+	int ret;
+
+	ret = iterate_generic_mmio(iter);
+	if (ret)
+		goto err;
+
+	if (IS_BROADWELL(i915)) {
+		ret = iterate_bdw_only_mmio(iter);
+		if (ret)
+			goto err;
+		ret = iterate_bdw_plus_mmio(iter);
+		if (ret)
+			goto err;
+		ret = iterate_pre_skl_mmio(iter);
+		if (ret)
+			goto err;
+	} else if (IS_SKYLAKE(i915) ||
+		   IS_KABYLAKE(i915) ||
+		   IS_COFFEELAKE(i915) ||
+		   IS_COMETLAKE(i915)) {
+		ret = iterate_bdw_plus_mmio(iter);
+		if (ret)
+			goto err;
+		ret = iterate_skl_plus_mmio(iter);
+		if (ret)
+			goto err;
+	} else if (IS_BROXTON(i915)) {
+		ret = iterate_bdw_plus_mmio(iter);
+		if (ret)
+			goto err;
+		ret = iterate_skl_plus_mmio(iter);
+		if (ret)
+			goto err;
+		ret = iterate_bxt_mmio(iter);
+		if (ret)
+			goto err;
+	}
+
+	return 0;
+err:
+	return ret;
+}
+EXPORT_SYMBOL_NS_GPL(intel_gvt_iterate_mmio_table, I915_GVT);
diff --git a/drivers/gpu/drm/i915/intel_mchbar_regs.h b/drivers/gpu/drm/i915/intel_mchbar_regs.h
new file mode 100644
index 000000000..ffc702b79
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_mchbar_regs.h
@@ -0,0 +1,231 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_MCHBAR_REGS__
+#define __INTEL_MCHBAR_REGS__
+
+#include "i915_reg_defs.h"
+
+/*
+ * MCHBAR mirror.
+ *
+ * This mirrors the MCHBAR MMIO space whose location is determined by
+ * device 0 function 0's pci config register 0x44 or 0x48 and matches it in
+ * every way.  It is not accessible from the CP register read instructions.
+ *
+ * Starting from Haswell, you can't write registers using the MCHBAR mirror,
+ * just read.
+ */
+
+#define MCHBAR_MIRROR_BASE			0x10000
+#define MCHBAR_MIRROR_BASE_SNB			0x140000
+
+#define CTG_STOLEN_RESERVED			_MMIO(MCHBAR_MIRROR_BASE + 0x34)
+#define ELK_STOLEN_RESERVED			_MMIO(MCHBAR_MIRROR_BASE + 0x48)
+#define   G4X_STOLEN_RESERVED_ADDR1_MASK	(0xFFFF << 16)
+#define   G4X_STOLEN_RESERVED_ADDR2_MASK	(0xFFF << 4)
+#define   G4X_STOLEN_RESERVED_ENABLE		(1 << 0)
+
+/* Pineview MCH register contains DDR3 setting */
+#define CSHRDDR3CTL				_MMIO(MCHBAR_MIRROR_BASE + 0x1a8)
+#define   CSHRDDR3CTL_DDR3			(1 << 2)
+
+/* 915-945 and GM965 MCH register controlling DRAM channel access */
+#define DCC					_MMIO(MCHBAR_MIRROR_BASE + 0x200)
+#define   DCC_ADDRESSING_MODE_SINGLE_CHANNEL	(0 << 0)
+#define   DCC_ADDRESSING_MODE_DUAL_CHANNEL_ASYMMETRIC	(1 << 0)
+#define   DCC_ADDRESSING_MODE_DUAL_CHANNEL_INTERLEAVED	(2 << 0)
+#define   DCC_ADDRESSING_MODE_MASK		(3 << 0)
+#define   DCC_CHANNEL_XOR_DISABLE		(1 << 10)
+#define   DCC_CHANNEL_XOR_BIT_17		(1 << 9)
+#define DCC2					_MMIO(MCHBAR_MIRROR_BASE + 0x204)
+#define   DCC2_MODIFIED_ENHANCED_DISABLE	(1 << 20)
+
+/* 965 MCH register controlling DRAM channel configuration */
+#define C0DRB3_BW				_MMIO(MCHBAR_MIRROR_BASE + 0x206)
+#define C1DRB3_BW				_MMIO(MCHBAR_MIRROR_BASE + 0x606)
+
+/* Clocking configuration register */
+#define CLKCFG					_MMIO(MCHBAR_MIRROR_BASE + 0xc00)
+#define CLKCFG_FSB_400				(0 << 0)	/* hrawclk 100 */
+#define CLKCFG_FSB_400_ALT			(5 << 0)	/* hrawclk 100 */
+#define CLKCFG_FSB_533				(1 << 0)	/* hrawclk 133 */
+#define CLKCFG_FSB_667				(3 << 0)	/* hrawclk 166 */
+#define CLKCFG_FSB_800				(2 << 0)	/* hrawclk 200 */
+#define CLKCFG_FSB_1067				(6 << 0)	/* hrawclk 266 */
+#define CLKCFG_FSB_1067_ALT			(0 << 0)	/* hrawclk 266 */
+#define CLKCFG_FSB_1333				(7 << 0)	/* hrawclk 333 */
+#define CLKCFG_FSB_1333_ALT			(4 << 0)	/* hrawclk 333 */
+#define CLKCFG_FSB_1600_ALT			(6 << 0)	/* hrawclk 400 */
+#define CLKCFG_FSB_MASK				(7 << 0)
+#define CLKCFG_MEM_533				(1 << 4)
+#define CLKCFG_MEM_667				(2 << 4)
+#define CLKCFG_MEM_800				(3 << 4)
+#define CLKCFG_MEM_MASK				(7 << 4)
+
+#define HPLLVCO_MOBILE				_MMIO(MCHBAR_MIRROR_BASE + 0xc0f)
+#define HPLLVCO					_MMIO(MCHBAR_MIRROR_BASE + 0xc38)
+
+#define TSC1					_MMIO(MCHBAR_MIRROR_BASE + 0x1001)
+#define   TSE					(1 << 0)
+#define TR1					_MMIO(MCHBAR_MIRROR_BASE + 0x1006)
+#define TSFS					_MMIO(MCHBAR_MIRROR_BASE + 0x1020)
+#define   TSFS_SLOPE_MASK			0x0000ff00
+#define   TSFS_SLOPE_SHIFT			8
+#define   TSFS_INTR_MASK			0x000000ff
+
+/* Memory latency timer register */
+#define MLTR_ILK				_MMIO(MCHBAR_MIRROR_BASE + 0x1222)
+/* the unit of memory self-refresh latency time is 0.5us */
+#define   MLTR_WM2_MASK				REG_GENMASK(13, 8)
+#define   MLTR_WM1_MASK				REG_GENMASK(5, 0)
+
+#define CSIPLL0					_MMIO(MCHBAR_MIRROR_BASE + 0x2c10)
+#define DDRMPLL1				_MMIO(MCHBAR_MIRROR_BASE + 0x2c20)
+
+#define ILK_GDSR				_MMIO(MCHBAR_MIRROR_BASE + 0x2ca4)
+#define  ILK_GRDOM_FULL				(0 << 1)
+#define  ILK_GRDOM_RENDER			(1 << 1)
+#define  ILK_GRDOM_MEDIA			(3 << 1)
+#define  ILK_GRDOM_MASK				(3 << 1)
+#define  ILK_GRDOM_RESET_ENABLE			(1 << 0)
+
+#define BXT_D_CR_DRP0_DUNIT8			0x1000
+#define BXT_D_CR_DRP0_DUNIT9			0x1200
+#define   BXT_D_CR_DRP0_DUNIT_START		8
+#define   BXT_D_CR_DRP0_DUNIT_END		11
+#define BXT_D_CR_DRP0_DUNIT(x)			_MMIO(MCHBAR_MIRROR_BASE_SNB + \
+						      _PICK_EVEN((x) - 8, BXT_D_CR_DRP0_DUNIT8,\
+								 BXT_D_CR_DRP0_DUNIT9))
+#define   BXT_DRAM_RANK_MASK			0x3
+#define   BXT_DRAM_RANK_SINGLE			0x1
+#define   BXT_DRAM_RANK_DUAL			0x3
+#define   BXT_DRAM_WIDTH_MASK			(0x3 << 4)
+#define   BXT_DRAM_WIDTH_SHIFT			4
+#define   BXT_DRAM_WIDTH_X8			(0x0 << 4)
+#define   BXT_DRAM_WIDTH_X16			(0x1 << 4)
+#define   BXT_DRAM_WIDTH_X32			(0x2 << 4)
+#define   BXT_DRAM_WIDTH_X64			(0x3 << 4)
+#define   BXT_DRAM_SIZE_MASK			(0x7 << 6)
+#define   BXT_DRAM_SIZE_SHIFT			6
+#define   BXT_DRAM_SIZE_4GBIT			(0x0 << 6)
+#define   BXT_DRAM_SIZE_6GBIT			(0x1 << 6)
+#define   BXT_DRAM_SIZE_8GBIT			(0x2 << 6)
+#define   BXT_DRAM_SIZE_12GBIT			(0x3 << 6)
+#define   BXT_DRAM_SIZE_16GBIT			(0x4 << 6)
+#define   BXT_DRAM_TYPE_MASK			(0x7 << 22)
+#define   BXT_DRAM_TYPE_SHIFT			22
+#define   BXT_DRAM_TYPE_DDR3			(0x0 << 22)
+#define   BXT_DRAM_TYPE_LPDDR3			(0x1 << 22)
+#define   BXT_DRAM_TYPE_LPDDR4			(0x2 << 22)
+#define   BXT_DRAM_TYPE_DDR4			(0x4 << 22)
+
+#define MCHBAR_CH0_CR_TC_PRE_0_0_0_MCHBAR	_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x4000)
+#define   DG1_DRAM_T_RDPRE_MASK			REG_GENMASK(16, 11)
+#define   DG1_DRAM_T_RP_MASK			REG_GENMASK(6, 0)
+#define MCHBAR_CH0_CR_TC_PRE_0_0_0_MCHBAR_HIGH	_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x4004)
+#define   DG1_DRAM_T_RCD_MASK			REG_GENMASK(15, 9)
+#define   DG1_DRAM_T_RAS_MASK			REG_GENMASK(8, 1)
+
+#define SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN	_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5000)
+#define   SKL_DRAM_DDR_TYPE_MASK		(0x3 << 0)
+#define   SKL_DRAM_DDR_TYPE_DDR4		(0 << 0)
+#define   SKL_DRAM_DDR_TYPE_DDR3		(1 << 0)
+#define   SKL_DRAM_DDR_TYPE_LPDDR3		(2 << 0)
+#define   SKL_DRAM_DDR_TYPE_LPDDR4		(3 << 0)
+
+/* snb MCH registers for reading the DRAM channel configuration */
+#define MAD_DIMM_C0				_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5004)
+#define MAD_DIMM_C1				_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5008)
+#define MAD_DIMM_C2				_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x500C)
+#define   MAD_DIMM_ECC_MASK			(0x3 << 24)
+#define   MAD_DIMM_ECC_OFF			(0x0 << 24)
+#define   MAD_DIMM_ECC_IO_ON_LOGIC_OFF		(0x1 << 24)
+#define   MAD_DIMM_ECC_IO_OFF_LOGIC_ON		(0x2 << 24)
+#define   MAD_DIMM_ECC_ON			(0x3 << 24)
+#define   MAD_DIMM_ENH_INTERLEAVE		(0x1 << 22)
+#define   MAD_DIMM_RANK_INTERLEAVE		(0x1 << 21)
+#define   MAD_DIMM_B_WIDTH_X16			(0x1 << 20) /* X8 chips if unset */
+#define   MAD_DIMM_A_WIDTH_X16			(0x1 << 19) /* X8 chips if unset */
+#define   MAD_DIMM_B_DUAL_RANK			(0x1 << 18)
+#define   MAD_DIMM_A_DUAL_RANK			(0x1 << 17)
+#define   MAD_DIMM_A_SELECT			(0x1 << 16)
+/* DIMM sizes are in multiples of 256mb. */
+#define   MAD_DIMM_B_SIZE_SHIFT			8
+#define   MAD_DIMM_B_SIZE_MASK			(0xff << MAD_DIMM_B_SIZE_SHIFT)
+#define   MAD_DIMM_A_SIZE_SHIFT			0
+#define   MAD_DIMM_A_SIZE_MASK			(0xff << MAD_DIMM_A_SIZE_SHIFT)
+
+#define SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN	_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x500C)
+#define SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN	_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5010)
+#define   SKL_DRAM_S_SHIFT			16
+#define   SKL_DRAM_SIZE_MASK			0x3F
+#define   SKL_DRAM_WIDTH_MASK			(0x3 << 8)
+#define   SKL_DRAM_WIDTH_SHIFT			8
+#define   SKL_DRAM_WIDTH_X8			(0x0 << 8)
+#define   SKL_DRAM_WIDTH_X16			(0x1 << 8)
+#define   SKL_DRAM_WIDTH_X32			(0x2 << 8)
+#define   SKL_DRAM_RANK_MASK			(0x1 << 10)
+#define   SKL_DRAM_RANK_SHIFT			10
+#define   SKL_DRAM_RANK_1			(0x0 << 10)
+#define   SKL_DRAM_RANK_2			(0x1 << 10)
+#define   SKL_DRAM_RANK_MASK			(0x1 << 10)
+#define   ICL_DRAM_SIZE_MASK			0x7F
+#define   ICL_DRAM_WIDTH_MASK			(0x3 << 7)
+#define   ICL_DRAM_WIDTH_SHIFT			7
+#define   ICL_DRAM_WIDTH_X8			(0x0 << 7)
+#define   ICL_DRAM_WIDTH_X16			(0x1 << 7)
+#define   ICL_DRAM_WIDTH_X32			(0x2 << 7)
+#define   ICL_DRAM_RANK_MASK			(0x3 << 9)
+#define   ICL_DRAM_RANK_SHIFT			9
+#define   ICL_DRAM_RANK_1			(0x0 << 9)
+#define   ICL_DRAM_RANK_2			(0x1 << 9)
+#define   ICL_DRAM_RANK_3			(0x2 << 9)
+#define   ICL_DRAM_RANK_4			(0x3 << 9)
+
+#define SA_PERF_STATUS_0_0_0_MCHBAR_PC		_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5918)
+#define  DG1_QCLK_RATIO_MASK			REG_GENMASK(9, 2)
+#define  DG1_QCLK_REFERENCE			REG_BIT(10)
+
+#define GEN6_GT_PERF_STATUS			_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5948)
+#define GEN6_RP_STATE_LIMITS			_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5994)
+#define GEN6_RP_STATE_CAP			_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5998)
+#define   RP0_CAP_MASK				REG_GENMASK(7, 0)
+#define   RP1_CAP_MASK				REG_GENMASK(15, 8)
+#define   RPN_CAP_MASK				REG_GENMASK(23, 16)
+
+#define GEN10_FREQ_INFO_REC			_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5ef0)
+#define   RPE_MASK				REG_GENMASK(15, 8)
+
+/* snb MCH registers for priority tuning */
+#define MCH_SSKPD				_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5d10)
+#define   SSKPD_NEW_WM0_MASK_HSW		REG_GENMASK64(63, 56)
+#define   SSKPD_WM4_MASK_HSW			REG_GENMASK64(40, 32)
+#define   SSKPD_WM3_MASK_HSW			REG_GENMASK64(28, 20)
+#define   SSKPD_WM2_MASK_HSW			REG_GENMASK64(19, 12)
+#define   SSKPD_WM1_MASK_HSW			REG_GENMASK64(11, 4)
+#define   SSKPD_OLD_WM0_MASK_HSW		REG_GENMASK64(3, 0)
+#define   SSKPD_WM3_MASK_SNB			REG_GENMASK(29, 24)
+#define   SSKPD_WM2_MASK_SNB			REG_GENMASK(21, 16)
+#define   SSKPD_WM1_MASK_SNB			REG_GENMASK(13, 8)
+#define   SSKPD_WM0_MASK_SNB			REG_GENMASK(5, 0)
+
+/* Memory controller frequency in MCHBAR for Haswell (possible SNB+) */
+#define DCLK					_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5e04)
+#define SKL_MC_BIOS_DATA_0_0_0_MCHBAR_PCU	_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5e04)
+#define   DG1_GEAR_TYPE				REG_BIT(16)
+
+/*
+ * Please see hsw_read_dcomp() and hsw_write_dcomp() before using this register,
+ * since on HSW we can't write to it using intel_uncore_write.
+ */
+#define D_COMP_HSW				_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x5f0c)
+#define  D_COMP_RCOMP_IN_PROGRESS		(1 << 9)
+#define  D_COMP_COMP_FORCE			(1 << 8)
+#define  D_COMP_COMP_DISABLE			(1 << 0)
+
+#define BXT_GT_PERF_STATUS			_MMIO(MCHBAR_MIRROR_BASE_SNB + 0x7070)
+
+#endif /* __INTEL_MCHBAR_REGS */
diff --git a/drivers/gpu/drm/i915/intel_memory_region.c b/drivers/gpu/drm/i915/intel_memory_region.c
new file mode 100644
index 000000000..9a4a7fb55
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_memory_region.c
@@ -0,0 +1,380 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/prandom.h>
+
+#include <uapi/drm/i915_drm.h>
+
+#include "intel_memory_region.h"
+#include "i915_drv.h"
+#include "i915_ttm_buddy_manager.h"
+
+static const struct {
+	u16 class;
+	u16 instance;
+} intel_region_map[] = {
+	[INTEL_REGION_SMEM] = {
+		.class = INTEL_MEMORY_SYSTEM,
+		.instance = 0,
+	},
+	[INTEL_REGION_LMEM_0] = {
+		.class = INTEL_MEMORY_LOCAL,
+		.instance = 0,
+	},
+	[INTEL_REGION_STOLEN_SMEM] = {
+		.class = INTEL_MEMORY_STOLEN_SYSTEM,
+		.instance = 0,
+	},
+	[INTEL_REGION_STOLEN_LMEM] = {
+		.class = INTEL_MEMORY_STOLEN_LOCAL,
+		.instance = 0,
+	},
+};
+
+static int __iopagetest(struct intel_memory_region *mem,
+			u8 __iomem *va, int pagesize,
+			u8 value, resource_size_t offset,
+			const void *caller)
+{
+	int byte = prandom_u32_max(pagesize);
+	u8 result[3];
+
+	memset_io(va, value, pagesize); /* or GPF! */
+	wmb();
+
+	result[0] = ioread8(va);
+	result[1] = ioread8(va + byte);
+	result[2] = ioread8(va + pagesize - 1);
+	if (memchr_inv(result, value, sizeof(result))) {
+		dev_err(mem->i915->drm.dev,
+			"Failed to read back from memory region:%pR at [%pa + %pa] for %ps; wrote %x, read (%x, %x, %x)\n",
+			&mem->region, &mem->io_start, &offset, caller,
+			value, result[0], result[1], result[2]);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int iopagetest(struct intel_memory_region *mem,
+		      resource_size_t offset,
+		      const void *caller)
+{
+	const u8 val[] = { 0x0, 0xa5, 0xc3, 0xf0 };
+	void __iomem *va;
+	int err;
+	int i;
+
+	va = ioremap_wc(mem->io_start + offset, PAGE_SIZE);
+	if (!va) {
+		dev_err(mem->i915->drm.dev,
+			"Failed to ioremap memory region [%pa + %pa] for %ps\n",
+			&mem->io_start, &offset, caller);
+		return -EFAULT;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(val); i++) {
+		err = __iopagetest(mem, va, PAGE_SIZE, val[i], offset, caller);
+		if (err)
+			break;
+
+		err = __iopagetest(mem, va, PAGE_SIZE, ~val[i], offset, caller);
+		if (err)
+			break;
+	}
+
+	iounmap(va);
+	return err;
+}
+
+static resource_size_t random_page(resource_size_t last)
+{
+	/* Limited to low 44b (16TiB), but should suffice for a spot check */
+	return prandom_u32_max(last >> PAGE_SHIFT) << PAGE_SHIFT;
+}
+
+static int iomemtest(struct intel_memory_region *mem,
+		     bool test_all,
+		     const void *caller)
+{
+	resource_size_t last, page;
+	int err;
+
+	if (mem->io_size < PAGE_SIZE)
+		return 0;
+
+	last = mem->io_size - PAGE_SIZE;
+
+	/*
+	 * Quick test to check read/write access to the iomap (backing store).
+	 *
+	 * Write a byte, read it back. If the iomapping fails, we expect
+	 * a GPF preventing further execution. If the backing store does not
+	 * exist, the read back will return garbage. We check a couple of pages,
+	 * the first and last of the specified region to confirm the backing
+	 * store + iomap does cover the entire memory region; and we check
+	 * a random offset within as a quick spot check for bad memory.
+	 */
+
+	if (test_all) {
+		for (page = 0; page <= last; page += PAGE_SIZE) {
+			err = iopagetest(mem, page, caller);
+			if (err)
+				return err;
+		}
+	} else {
+		err = iopagetest(mem, 0, caller);
+		if (err)
+			return err;
+
+		err = iopagetest(mem, last, caller);
+		if (err)
+			return err;
+
+		err = iopagetest(mem, random_page(last), caller);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+struct intel_memory_region *
+intel_memory_region_lookup(struct drm_i915_private *i915,
+			   u16 class, u16 instance)
+{
+	struct intel_memory_region *mr;
+	int id;
+
+	/* XXX: consider maybe converting to an rb tree at some point */
+	for_each_memory_region(mr, i915, id) {
+		if (mr->type == class && mr->instance == instance)
+			return mr;
+	}
+
+	return NULL;
+}
+
+struct intel_memory_region *
+intel_memory_region_by_type(struct drm_i915_private *i915,
+			    enum intel_memory_type mem_type)
+{
+	struct intel_memory_region *mr;
+	int id;
+
+	for_each_memory_region(mr, i915, id)
+		if (mr->type == mem_type)
+			return mr;
+
+	return NULL;
+}
+
+/**
+ * intel_memory_region_reserve - Reserve a memory range
+ * @mem: The region for which we want to reserve a range.
+ * @offset: Start of the range to reserve.
+ * @size: The size of the range to reserve.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int intel_memory_region_reserve(struct intel_memory_region *mem,
+				resource_size_t offset,
+				resource_size_t size)
+{
+	struct ttm_resource_manager *man = mem->region_private;
+
+	GEM_BUG_ON(mem->is_range_manager);
+
+	return i915_ttm_buddy_man_reserve(man, offset, size);
+}
+
+void intel_memory_region_debug(struct intel_memory_region *mr,
+			       struct drm_printer *printer)
+{
+	drm_printf(printer, "%s: ", mr->name);
+
+	if (mr->region_private)
+		ttm_resource_manager_debug(mr->region_private, printer);
+	else
+		drm_printf(printer, "total:%pa bytes\n", &mr->total);
+}
+
+static int intel_memory_region_memtest(struct intel_memory_region *mem,
+				       void *caller)
+{
+	struct drm_i915_private *i915 = mem->i915;
+	int err = 0;
+
+	if (!mem->io_start)
+		return 0;
+
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM) || i915->params.memtest)
+		err = iomemtest(mem, i915->params.memtest, caller);
+
+	return err;
+}
+
+struct intel_memory_region *
+intel_memory_region_create(struct drm_i915_private *i915,
+			   resource_size_t start,
+			   resource_size_t size,
+			   resource_size_t min_page_size,
+			   resource_size_t io_start,
+			   resource_size_t io_size,
+			   u16 type,
+			   u16 instance,
+			   const struct intel_memory_region_ops *ops)
+{
+	struct intel_memory_region *mem;
+	int err;
+
+	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
+	if (!mem)
+		return ERR_PTR(-ENOMEM);
+
+	mem->i915 = i915;
+	mem->region = (struct resource)DEFINE_RES_MEM(start, size);
+	mem->io_start = io_start;
+	mem->io_size = io_size;
+	mem->min_page_size = min_page_size;
+	mem->ops = ops;
+	mem->total = size;
+	mem->type = type;
+	mem->instance = instance;
+
+	mutex_init(&mem->objects.lock);
+	INIT_LIST_HEAD(&mem->objects.list);
+
+	if (ops->init) {
+		err = ops->init(mem);
+		if (err)
+			goto err_free;
+	}
+
+	err = intel_memory_region_memtest(mem, (void *)_RET_IP_);
+	if (err)
+		goto err_release;
+
+	return mem;
+
+err_release:
+	if (mem->ops->release)
+		mem->ops->release(mem);
+err_free:
+	kfree(mem);
+	return ERR_PTR(err);
+}
+
+void intel_memory_region_set_name(struct intel_memory_region *mem,
+				  const char *fmt, ...)
+{
+	va_list ap;
+
+	va_start(ap, fmt);
+	vsnprintf(mem->name, sizeof(mem->name), fmt, ap);
+	va_end(ap);
+}
+
+void intel_memory_region_avail(struct intel_memory_region *mr,
+			       u64 *avail, u64 *visible_avail)
+{
+	if (mr->type == INTEL_MEMORY_LOCAL) {
+		i915_ttm_buddy_man_avail(mr->region_private,
+					 avail, visible_avail);
+		*avail <<= PAGE_SHIFT;
+		*visible_avail <<= PAGE_SHIFT;
+	} else {
+		*avail = mr->total;
+		*visible_avail = mr->total;
+	}
+}
+
+void intel_memory_region_destroy(struct intel_memory_region *mem)
+{
+	int ret = 0;
+
+	if (mem->ops->release)
+		ret = mem->ops->release(mem);
+
+	GEM_WARN_ON(!list_empty_careful(&mem->objects.list));
+	mutex_destroy(&mem->objects.lock);
+	if (!ret)
+		kfree(mem);
+}
+
+/* Global memory region registration -- only slight layer inversions! */
+
+int intel_memory_regions_hw_probe(struct drm_i915_private *i915)
+{
+	int err, i;
+
+	for (i = 0; i < ARRAY_SIZE(i915->mm.regions); i++) {
+		struct intel_memory_region *mem = ERR_PTR(-ENODEV);
+		u16 type, instance;
+
+		if (!HAS_REGION(i915, BIT(i)))
+			continue;
+
+		type = intel_region_map[i].class;
+		instance = intel_region_map[i].instance;
+		switch (type) {
+		case INTEL_MEMORY_SYSTEM:
+			if (IS_DGFX(i915))
+				mem = i915_gem_ttm_system_setup(i915, type,
+								instance);
+			else
+				mem = i915_gem_shmem_setup(i915, type,
+							   instance);
+			break;
+		case INTEL_MEMORY_STOLEN_LOCAL:
+			mem = i915_gem_stolen_lmem_setup(i915, type, instance);
+			if (!IS_ERR(mem))
+				i915->mm.stolen_region = mem;
+			break;
+		case INTEL_MEMORY_STOLEN_SYSTEM:
+			mem = i915_gem_stolen_smem_setup(i915, type, instance);
+			if (!IS_ERR(mem))
+				i915->mm.stolen_region = mem;
+			break;
+		default:
+			continue;
+		}
+
+		if (IS_ERR(mem)) {
+			err = PTR_ERR(mem);
+			drm_err(&i915->drm,
+				"Failed to setup region(%d) type=%d\n",
+				err, type);
+			goto out_cleanup;
+		}
+
+		mem->id = i;
+		i915->mm.regions[i] = mem;
+	}
+
+	return 0;
+
+out_cleanup:
+	intel_memory_regions_driver_release(i915);
+	return err;
+}
+
+void intel_memory_regions_driver_release(struct drm_i915_private *i915)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(i915->mm.regions); i++) {
+		struct intel_memory_region *region =
+			fetch_and_zero(&i915->mm.regions[i]);
+
+		if (region)
+			intel_memory_region_destroy(region);
+	}
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/intel_memory_region.c"
+#include "selftests/mock_region.c"
+#endif
diff --git a/drivers/gpu/drm/i915/intel_memory_region.h b/drivers/gpu/drm/i915/intel_memory_region.h
new file mode 100644
index 000000000..2953ed5c3
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_memory_region.h
@@ -0,0 +1,139 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_MEMORY_REGION_H__
+#define __INTEL_MEMORY_REGION_H__
+
+#include <linux/ioport.h>
+#include <linux/mutex.h>
+#include <linux/io-mapping.h>
+#include <drm/drm_mm.h>
+#include <uapi/drm/i915_drm.h>
+
+struct drm_i915_private;
+struct drm_i915_gem_object;
+struct drm_printer;
+struct intel_memory_region;
+struct sg_table;
+struct ttm_resource;
+
+enum intel_memory_type {
+	INTEL_MEMORY_SYSTEM = I915_MEMORY_CLASS_SYSTEM,
+	INTEL_MEMORY_LOCAL = I915_MEMORY_CLASS_DEVICE,
+	INTEL_MEMORY_STOLEN_SYSTEM,
+	INTEL_MEMORY_STOLEN_LOCAL,
+	INTEL_MEMORY_MOCK,
+};
+
+enum intel_region_id {
+	INTEL_REGION_SMEM = 0,
+	INTEL_REGION_LMEM_0,
+	INTEL_REGION_LMEM_1,
+	INTEL_REGION_LMEM_2,
+	INTEL_REGION_LMEM_3,
+	INTEL_REGION_STOLEN_SMEM,
+	INTEL_REGION_STOLEN_LMEM,
+	INTEL_REGION_UNKNOWN, /* Should be last */
+};
+
+#define REGION_SMEM     BIT(INTEL_REGION_SMEM)
+#define REGION_LMEM     BIT(INTEL_REGION_LMEM_0)
+#define REGION_STOLEN_SMEM   BIT(INTEL_REGION_STOLEN_SMEM)
+#define REGION_STOLEN_LMEM   BIT(INTEL_REGION_STOLEN_LMEM)
+
+#define I915_ALLOC_CONTIGUOUS     BIT(0)
+
+#define for_each_memory_region(mr, i915, id) \
+	for (id = 0; id < ARRAY_SIZE((i915)->mm.regions); id++) \
+		for_each_if((mr) = (i915)->mm.regions[id])
+
+struct intel_memory_region_ops {
+	unsigned int flags;
+
+	int (*init)(struct intel_memory_region *mem);
+	int (*release)(struct intel_memory_region *mem);
+
+	int (*init_object)(struct intel_memory_region *mem,
+			   struct drm_i915_gem_object *obj,
+			   resource_size_t offset,
+			   resource_size_t size,
+			   resource_size_t page_size,
+			   unsigned int flags);
+};
+
+struct intel_memory_region {
+	struct drm_i915_private *i915;
+
+	const struct intel_memory_region_ops *ops;
+
+	struct io_mapping iomap;
+	struct resource region;
+
+	resource_size_t io_start;
+	resource_size_t io_size;
+	resource_size_t min_page_size;
+	resource_size_t total;
+
+	u16 type;
+	u16 instance;
+	enum intel_region_id id;
+	char name[16];
+	bool private; /* not for userspace */
+
+	struct {
+		struct mutex lock; /* Protects access to objects */
+		struct list_head list;
+	} objects;
+
+	bool is_range_manager;
+
+	void *region_private;
+};
+
+struct intel_memory_region *
+intel_memory_region_lookup(struct drm_i915_private *i915,
+			   u16 class, u16 instance);
+
+struct intel_memory_region *
+intel_memory_region_create(struct drm_i915_private *i915,
+			   resource_size_t start,
+			   resource_size_t size,
+			   resource_size_t min_page_size,
+			   resource_size_t io_start,
+			   resource_size_t io_size,
+			   u16 type,
+			   u16 instance,
+			   const struct intel_memory_region_ops *ops);
+
+void intel_memory_region_destroy(struct intel_memory_region *mem);
+
+int intel_memory_regions_hw_probe(struct drm_i915_private *i915);
+void intel_memory_regions_driver_release(struct drm_i915_private *i915);
+struct intel_memory_region *
+intel_memory_region_by_type(struct drm_i915_private *i915,
+			    enum intel_memory_type mem_type);
+
+__printf(2, 3) void
+intel_memory_region_set_name(struct intel_memory_region *mem,
+			     const char *fmt, ...);
+
+int intel_memory_region_reserve(struct intel_memory_region *mem,
+				resource_size_t offset,
+				resource_size_t size);
+
+void intel_memory_region_debug(struct intel_memory_region *mr,
+			       struct drm_printer *printer);
+
+void intel_memory_region_avail(struct intel_memory_region *mr,
+			       u64 *avail, u64 *visible_avail);
+
+struct intel_memory_region *
+i915_gem_ttm_system_setup(struct drm_i915_private *i915,
+			  u16 type, u16 instance);
+struct intel_memory_region *
+i915_gem_shmem_setup(struct drm_i915_private *i915,
+		     u16 type, u16 instance);
+
+#endif
diff --git a/drivers/gpu/drm/i915/intel_pch.c b/drivers/gpu/drm/i915/intel_pch.c
new file mode 100644
index 000000000..ba9843cb1
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_pch.c
@@ -0,0 +1,279 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright 2019 Intel Corporation.
+ */
+
+#include "i915_drv.h"
+#include "i915_utils.h"
+#include "intel_pch.h"
+
+/* Map PCH device id to PCH type, or PCH_NONE if unknown. */
+static enum intel_pch
+intel_pch_type(const struct drm_i915_private *dev_priv, unsigned short id)
+{
+	switch (id) {
+	case INTEL_PCH_IBX_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Ibex Peak PCH\n");
+		drm_WARN_ON(&dev_priv->drm, GRAPHICS_VER(dev_priv) != 5);
+		return PCH_IBX;
+	case INTEL_PCH_CPT_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found CougarPoint PCH\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    GRAPHICS_VER(dev_priv) != 6 && !IS_IVYBRIDGE(dev_priv));
+		return PCH_CPT;
+	case INTEL_PCH_PPT_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found PantherPoint PCH\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    GRAPHICS_VER(dev_priv) != 6 && !IS_IVYBRIDGE(dev_priv));
+		/* PPT is CPT compatible */
+		return PCH_CPT;
+	case INTEL_PCH_LPT_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found LynxPoint PCH\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_HASWELL(dev_priv) && !IS_BROADWELL(dev_priv));
+		drm_WARN_ON(&dev_priv->drm,
+			    IS_HSW_ULT(dev_priv) || IS_BDW_ULT(dev_priv));
+		return PCH_LPT;
+	case INTEL_PCH_LPT_LP_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found LynxPoint LP PCH\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_HASWELL(dev_priv) && !IS_BROADWELL(dev_priv));
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_HSW_ULT(dev_priv) && !IS_BDW_ULT(dev_priv));
+		return PCH_LPT;
+	case INTEL_PCH_WPT_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found WildcatPoint PCH\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_HASWELL(dev_priv) && !IS_BROADWELL(dev_priv));
+		drm_WARN_ON(&dev_priv->drm,
+			    IS_HSW_ULT(dev_priv) || IS_BDW_ULT(dev_priv));
+		/* WPT is LPT compatible */
+		return PCH_LPT;
+	case INTEL_PCH_WPT_LP_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found WildcatPoint LP PCH\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_HASWELL(dev_priv) && !IS_BROADWELL(dev_priv));
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_HSW_ULT(dev_priv) && !IS_BDW_ULT(dev_priv));
+		/* WPT is LPT compatible */
+		return PCH_LPT;
+	case INTEL_PCH_SPT_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found SunrisePoint PCH\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_SKYLAKE(dev_priv) && !IS_KABYLAKE(dev_priv));
+		return PCH_SPT;
+	case INTEL_PCH_SPT_LP_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found SunrisePoint LP PCH\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_SKYLAKE(dev_priv) &&
+			    !IS_KABYLAKE(dev_priv) &&
+			    !IS_COFFEELAKE(dev_priv) &&
+			    !IS_COMETLAKE(dev_priv));
+		return PCH_SPT;
+	case INTEL_PCH_KBP_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Kaby Lake PCH (KBP)\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_SKYLAKE(dev_priv) &&
+			    !IS_KABYLAKE(dev_priv) &&
+			    !IS_COFFEELAKE(dev_priv) &&
+			    !IS_COMETLAKE(dev_priv));
+		/* KBP is SPT compatible */
+		return PCH_SPT;
+	case INTEL_PCH_CNP_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Cannon Lake PCH (CNP)\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_COFFEELAKE(dev_priv) &&
+			    !IS_COMETLAKE(dev_priv));
+		return PCH_CNP;
+	case INTEL_PCH_CNP_LP_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm,
+			    "Found Cannon Lake LP PCH (CNP-LP)\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_COFFEELAKE(dev_priv) &&
+			    !IS_COMETLAKE(dev_priv));
+		return PCH_CNP;
+	case INTEL_PCH_CMP_DEVICE_ID_TYPE:
+	case INTEL_PCH_CMP2_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Comet Lake PCH (CMP)\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_COFFEELAKE(dev_priv) &&
+			    !IS_COMETLAKE(dev_priv) &&
+			    !IS_ROCKETLAKE(dev_priv));
+		/* CMP is CNP compatible */
+		return PCH_CNP;
+	case INTEL_PCH_CMP_V_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Comet Lake V PCH (CMP-V)\n");
+		drm_WARN_ON(&dev_priv->drm,
+			    !IS_COFFEELAKE(dev_priv) &&
+			    !IS_COMETLAKE(dev_priv));
+		/* CMP-V is based on KBP, which is SPT compatible */
+		return PCH_SPT;
+	case INTEL_PCH_ICP_DEVICE_ID_TYPE:
+	case INTEL_PCH_ICP2_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Ice Lake PCH\n");
+		drm_WARN_ON(&dev_priv->drm, !IS_ICELAKE(dev_priv));
+		return PCH_ICP;
+	case INTEL_PCH_MCC_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Mule Creek Canyon PCH\n");
+		drm_WARN_ON(&dev_priv->drm, !IS_JSL_EHL(dev_priv));
+		/* MCC is TGP compatible */
+		return PCH_TGP;
+	case INTEL_PCH_TGP_DEVICE_ID_TYPE:
+	case INTEL_PCH_TGP2_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Tiger Lake LP PCH\n");
+		drm_WARN_ON(&dev_priv->drm, !IS_TIGERLAKE(dev_priv) &&
+			    !IS_ROCKETLAKE(dev_priv) &&
+			    !IS_GEN9_BC(dev_priv));
+		return PCH_TGP;
+	case INTEL_PCH_JSP_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Jasper Lake PCH\n");
+		drm_WARN_ON(&dev_priv->drm, !IS_JSL_EHL(dev_priv));
+		/* JSP is ICP compatible */
+		return PCH_ICP;
+	case INTEL_PCH_ADP_DEVICE_ID_TYPE:
+	case INTEL_PCH_ADP2_DEVICE_ID_TYPE:
+	case INTEL_PCH_ADP3_DEVICE_ID_TYPE:
+	case INTEL_PCH_ADP4_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Alder Lake PCH\n");
+		drm_WARN_ON(&dev_priv->drm, !IS_ALDERLAKE_S(dev_priv) &&
+			    !IS_ALDERLAKE_P(dev_priv));
+		return PCH_ADP;
+	case INTEL_PCH_MTP_DEVICE_ID_TYPE:
+	case INTEL_PCH_MTP2_DEVICE_ID_TYPE:
+		drm_dbg_kms(&dev_priv->drm, "Found Meteor Lake PCH\n");
+		drm_WARN_ON(&dev_priv->drm, !IS_METEORLAKE(dev_priv));
+		return PCH_MTP;
+	default:
+		return PCH_NONE;
+	}
+}
+
+static bool intel_is_virt_pch(unsigned short id,
+			      unsigned short svendor, unsigned short sdevice)
+{
+	return (id == INTEL_PCH_P2X_DEVICE_ID_TYPE ||
+		id == INTEL_PCH_P3X_DEVICE_ID_TYPE ||
+		(id == INTEL_PCH_QEMU_DEVICE_ID_TYPE &&
+		 svendor == PCI_SUBVENDOR_ID_REDHAT_QUMRANET &&
+		 sdevice == PCI_SUBDEVICE_ID_QEMU));
+}
+
+static void
+intel_virt_detect_pch(const struct drm_i915_private *dev_priv,
+		      unsigned short *pch_id, enum intel_pch *pch_type)
+{
+	unsigned short id = 0;
+
+	/*
+	 * In a virtualized passthrough environment we can be in a
+	 * setup where the ISA bridge is not able to be passed through.
+	 * In this case, a south bridge can be emulated and we have to
+	 * make an educated guess as to which PCH is really there.
+	 */
+
+	if (IS_METEORLAKE(dev_priv))
+		id = INTEL_PCH_MTP_DEVICE_ID_TYPE;
+	else if (IS_ALDERLAKE_S(dev_priv) || IS_ALDERLAKE_P(dev_priv))
+		id = INTEL_PCH_ADP_DEVICE_ID_TYPE;
+	else if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv))
+		id = INTEL_PCH_TGP_DEVICE_ID_TYPE;
+	else if (IS_JSL_EHL(dev_priv))
+		id = INTEL_PCH_MCC_DEVICE_ID_TYPE;
+	else if (IS_ICELAKE(dev_priv))
+		id = INTEL_PCH_ICP_DEVICE_ID_TYPE;
+	else if (IS_COFFEELAKE(dev_priv) ||
+		 IS_COMETLAKE(dev_priv))
+		id = INTEL_PCH_CNP_DEVICE_ID_TYPE;
+	else if (IS_KABYLAKE(dev_priv) || IS_SKYLAKE(dev_priv))
+		id = INTEL_PCH_SPT_DEVICE_ID_TYPE;
+	else if (IS_HSW_ULT(dev_priv) || IS_BDW_ULT(dev_priv))
+		id = INTEL_PCH_LPT_LP_DEVICE_ID_TYPE;
+	else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+		id = INTEL_PCH_LPT_DEVICE_ID_TYPE;
+	else if (GRAPHICS_VER(dev_priv) == 6 || IS_IVYBRIDGE(dev_priv))
+		id = INTEL_PCH_CPT_DEVICE_ID_TYPE;
+	else if (GRAPHICS_VER(dev_priv) == 5)
+		id = INTEL_PCH_IBX_DEVICE_ID_TYPE;
+
+	if (id)
+		drm_dbg_kms(&dev_priv->drm, "Assuming PCH ID %04x\n", id);
+	else
+		drm_dbg_kms(&dev_priv->drm, "Assuming no PCH\n");
+
+	*pch_type = intel_pch_type(dev_priv, id);
+
+	/* Sanity check virtual PCH id */
+	if (drm_WARN_ON(&dev_priv->drm,
+			id && *pch_type == PCH_NONE))
+		id = 0;
+
+	*pch_id = id;
+}
+
+void intel_detect_pch(struct drm_i915_private *dev_priv)
+{
+	struct pci_dev *pch = NULL;
+	unsigned short id;
+	enum intel_pch pch_type;
+
+	/* DG1 has south engine display on the same PCI device */
+	if (IS_DG1(dev_priv)) {
+		dev_priv->pch_type = PCH_DG1;
+		return;
+	} else if (IS_DG2(dev_priv)) {
+		dev_priv->pch_type = PCH_DG2;
+		return;
+	}
+
+	/*
+	 * The reason to probe ISA bridge instead of Dev31:Fun0 is to
+	 * make graphics device passthrough work easy for VMM, that only
+	 * need to expose ISA bridge to let driver know the real hardware
+	 * underneath. This is a requirement from virtualization team.
+	 *
+	 * In some virtualized environments (e.g. XEN), there is irrelevant
+	 * ISA bridge in the system. To work reliably, we should scan trhough
+	 * all the ISA bridge devices and check for the first match, instead
+	 * of only checking the first one.
+	 */
+	while ((pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, pch))) {
+		if (pch->vendor != PCI_VENDOR_ID_INTEL)
+			continue;
+
+		id = pch->device & INTEL_PCH_DEVICE_ID_MASK;
+
+		pch_type = intel_pch_type(dev_priv, id);
+		if (pch_type != PCH_NONE) {
+			dev_priv->pch_type = pch_type;
+			dev_priv->pch_id = id;
+			break;
+		} else if (intel_is_virt_pch(id, pch->subsystem_vendor,
+					     pch->subsystem_device)) {
+			intel_virt_detect_pch(dev_priv, &id, &pch_type);
+			dev_priv->pch_type = pch_type;
+			dev_priv->pch_id = id;
+			break;
+		}
+	}
+
+	/*
+	 * Use PCH_NOP (PCH but no South Display) for PCH platforms without
+	 * display.
+	 */
+	if (pch && !HAS_DISPLAY(dev_priv)) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Display disabled, reverting to NOP PCH\n");
+		dev_priv->pch_type = PCH_NOP;
+		dev_priv->pch_id = 0;
+	} else if (!pch) {
+		if (i915_run_as_guest() && HAS_DISPLAY(dev_priv)) {
+			intel_virt_detect_pch(dev_priv, &id, &pch_type);
+			dev_priv->pch_type = pch_type;
+			dev_priv->pch_id = id;
+		} else {
+			drm_dbg_kms(&dev_priv->drm, "No PCH found.\n");
+		}
+	}
+
+	pci_dev_put(pch);
+}
diff --git a/drivers/gpu/drm/i915/intel_pch.h b/drivers/gpu/drm/i915/intel_pch.h
new file mode 100644
index 000000000..32aff5a70
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_pch.h
@@ -0,0 +1,91 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2019 Intel Corporation.
+ */
+
+#ifndef __INTEL_PCH__
+#define __INTEL_PCH__
+
+struct drm_i915_private;
+
+/*
+ * Sorted by south display engine compatibility.
+ * If the new PCH comes with a south display engine that is not
+ * inherited from the latest item, please do not add it to the
+ * end. Instead, add it right after its "parent" PCH.
+ */
+enum intel_pch {
+	PCH_NOP = -1,	/* PCH without south display */
+	PCH_NONE = 0,	/* No PCH present */
+	PCH_IBX,	/* Ibexpeak PCH */
+	PCH_CPT,	/* Cougarpoint/Pantherpoint PCH */
+	PCH_LPT,	/* Lynxpoint/Wildcatpoint PCH */
+	PCH_SPT,        /* Sunrisepoint/Kaby Lake PCH */
+	PCH_CNP,        /* Cannon/Comet Lake PCH */
+	PCH_ICP,	/* Ice Lake/Jasper Lake PCH */
+	PCH_TGP,	/* Tiger Lake/Mule Creek Canyon PCH */
+	PCH_ADP,	/* Alder Lake PCH */
+	PCH_MTP,	/* Meteor Lake PCH */
+
+	/* Fake PCHs, functionality handled on the same PCI dev */
+	PCH_DG1 = 1024,
+	PCH_DG2,
+};
+
+#define INTEL_PCH_DEVICE_ID_MASK		0xff80
+#define INTEL_PCH_IBX_DEVICE_ID_TYPE		0x3b00
+#define INTEL_PCH_CPT_DEVICE_ID_TYPE		0x1c00
+#define INTEL_PCH_PPT_DEVICE_ID_TYPE		0x1e00
+#define INTEL_PCH_LPT_DEVICE_ID_TYPE		0x8c00
+#define INTEL_PCH_LPT_LP_DEVICE_ID_TYPE		0x9c00
+#define INTEL_PCH_WPT_DEVICE_ID_TYPE		0x8c80
+#define INTEL_PCH_WPT_LP_DEVICE_ID_TYPE		0x9c80
+#define INTEL_PCH_SPT_DEVICE_ID_TYPE		0xA100
+#define INTEL_PCH_SPT_LP_DEVICE_ID_TYPE		0x9D00
+#define INTEL_PCH_KBP_DEVICE_ID_TYPE		0xA280
+#define INTEL_PCH_CNP_DEVICE_ID_TYPE		0xA300
+#define INTEL_PCH_CNP_LP_DEVICE_ID_TYPE		0x9D80
+#define INTEL_PCH_CMP_DEVICE_ID_TYPE		0x0280
+#define INTEL_PCH_CMP2_DEVICE_ID_TYPE		0x0680
+#define INTEL_PCH_CMP_V_DEVICE_ID_TYPE		0xA380
+#define INTEL_PCH_ICP_DEVICE_ID_TYPE		0x3480
+#define INTEL_PCH_ICP2_DEVICE_ID_TYPE		0x3880
+#define INTEL_PCH_MCC_DEVICE_ID_TYPE		0x4B00
+#define INTEL_PCH_TGP_DEVICE_ID_TYPE		0xA080
+#define INTEL_PCH_TGP2_DEVICE_ID_TYPE		0x4380
+#define INTEL_PCH_JSP_DEVICE_ID_TYPE		0x4D80
+#define INTEL_PCH_ADP_DEVICE_ID_TYPE		0x7A80
+#define INTEL_PCH_ADP2_DEVICE_ID_TYPE		0x5180
+#define INTEL_PCH_ADP3_DEVICE_ID_TYPE		0x7A00
+#define INTEL_PCH_ADP4_DEVICE_ID_TYPE		0x5480
+#define INTEL_PCH_MTP_DEVICE_ID_TYPE		0x7E00
+#define INTEL_PCH_MTP2_DEVICE_ID_TYPE		0xAE00
+#define INTEL_PCH_P2X_DEVICE_ID_TYPE		0x7100
+#define INTEL_PCH_P3X_DEVICE_ID_TYPE		0x7000
+#define INTEL_PCH_QEMU_DEVICE_ID_TYPE		0x2900 /* qemu q35 has 2918 */
+
+#define INTEL_PCH_TYPE(dev_priv)		((dev_priv)->pch_type)
+#define INTEL_PCH_ID(dev_priv)			((dev_priv)->pch_id)
+#define HAS_PCH_MTP(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_MTP)
+#define HAS_PCH_DG2(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_DG2)
+#define HAS_PCH_ADP(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_ADP)
+#define HAS_PCH_DG1(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_DG1)
+#define HAS_PCH_TGP(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_TGP)
+#define HAS_PCH_ICP(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_ICP)
+#define HAS_PCH_CNP(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_CNP)
+#define HAS_PCH_SPT(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_SPT)
+#define HAS_PCH_LPT(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_LPT)
+#define HAS_PCH_LPT_LP(dev_priv) \
+	(INTEL_PCH_ID(dev_priv) == INTEL_PCH_LPT_LP_DEVICE_ID_TYPE || \
+	 INTEL_PCH_ID(dev_priv) == INTEL_PCH_WPT_LP_DEVICE_ID_TYPE)
+#define HAS_PCH_LPT_H(dev_priv) \
+	(INTEL_PCH_ID(dev_priv) == INTEL_PCH_LPT_DEVICE_ID_TYPE || \
+	 INTEL_PCH_ID(dev_priv) == INTEL_PCH_WPT_DEVICE_ID_TYPE)
+#define HAS_PCH_CPT(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_CPT)
+#define HAS_PCH_IBX(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_IBX)
+#define HAS_PCH_NOP(dev_priv)			(INTEL_PCH_TYPE(dev_priv) == PCH_NOP)
+#define HAS_PCH_SPLIT(dev_priv)			(INTEL_PCH_TYPE(dev_priv) != PCH_NONE)
+
+void intel_detect_pch(struct drm_i915_private *dev_priv);
+
+#endif /* __INTEL_PCH__ */
diff --git a/drivers/gpu/drm/i915/intel_pci_config.h b/drivers/gpu/drm/i915/intel_pci_config.h
new file mode 100644
index 000000000..4977a524c
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_pci_config.h
@@ -0,0 +1,92 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef __INTEL_PCI_CONFIG_H__
+#define __INTEL_PCI_CONFIG_H__
+
+/* PCI BARs */
+#define GTTMMADR_BAR				0
+#define GEN2_GTTMMADR_BAR			1
+#define GFXMEM_BAR				2
+#define GTT_APERTURE_BAR			GFXMEM_BAR
+#define GEN12_LMEM_BAR				GFXMEM_BAR
+
+/* BSM in include/drm/i915_drm.h */
+
+#define MCHBAR_I915				0x44
+#define MCHBAR_I965				0x48
+#define   MCHBAR_SIZE				(4 * 4096)
+
+#define DEVEN					0x54
+#define   DEVEN_MCHBAR_EN			(1 << 28)
+
+#define HPLLCC					0xc0 /* 85x only */
+#define   GC_CLOCK_CONTROL_MASK			(0x7 << 0)
+#define   GC_CLOCK_133_200			(0 << 0)
+#define   GC_CLOCK_100_200			(1 << 0)
+#define   GC_CLOCK_100_133			(2 << 0)
+#define   GC_CLOCK_133_266			(3 << 0)
+#define   GC_CLOCK_133_200_2			(4 << 0)
+#define   GC_CLOCK_133_266_2			(5 << 0)
+#define   GC_CLOCK_166_266			(6 << 0)
+#define   GC_CLOCK_166_250			(7 << 0)
+
+#define I915_GDRST				0xc0
+#define   GRDOM_FULL				(0 << 2)
+#define   GRDOM_RENDER				(1 << 2)
+#define   GRDOM_MEDIA				(3 << 2)
+#define   GRDOM_MASK				(3 << 2)
+#define   GRDOM_RESET_STATUS			(1 << 1)
+#define   GRDOM_RESET_ENABLE			(1 << 0)
+
+/* BSpec only has register offset, PCI device and bit found empirically */
+#define I830_CLOCK_GATE				0xc8 /* device 0 */
+#define   I830_L2_CACHE_CLOCK_GATE_DISABLE	(1 << 2)
+
+#define GCDGMBUS				0xcc
+
+#define GCFGC2					0xda
+#define GCFGC					0xf0 /* 915+ only */
+#define   GC_LOW_FREQUENCY_ENABLE		(1 << 7)
+#define   GC_DISPLAY_CLOCK_190_200_MHZ		(0 << 4)
+#define   GC_DISPLAY_CLOCK_333_320_MHZ		(4 << 4)
+#define   GC_DISPLAY_CLOCK_267_MHZ_PNV		(0 << 4)
+#define   GC_DISPLAY_CLOCK_333_MHZ_PNV		(1 << 4)
+#define   GC_DISPLAY_CLOCK_444_MHZ_PNV		(2 << 4)
+#define   GC_DISPLAY_CLOCK_200_MHZ_PNV		(5 << 4)
+#define   GC_DISPLAY_CLOCK_133_MHZ_PNV		(6 << 4)
+#define   GC_DISPLAY_CLOCK_167_MHZ_PNV		(7 << 4)
+#define   GC_DISPLAY_CLOCK_MASK			(7 << 4)
+#define   GM45_GC_RENDER_CLOCK_MASK		(0xf << 0)
+#define   GM45_GC_RENDER_CLOCK_266_MHZ		(8 << 0)
+#define   GM45_GC_RENDER_CLOCK_320_MHZ		(9 << 0)
+#define   GM45_GC_RENDER_CLOCK_400_MHZ		(0xb << 0)
+#define   GM45_GC_RENDER_CLOCK_533_MHZ		(0xc << 0)
+#define   I965_GC_RENDER_CLOCK_MASK		(0xf << 0)
+#define   I965_GC_RENDER_CLOCK_267_MHZ		(2 << 0)
+#define   I965_GC_RENDER_CLOCK_333_MHZ		(3 << 0)
+#define   I965_GC_RENDER_CLOCK_444_MHZ		(4 << 0)
+#define   I965_GC_RENDER_CLOCK_533_MHZ		(5 << 0)
+#define   I945_GC_RENDER_CLOCK_MASK		(7 << 0)
+#define   I945_GC_RENDER_CLOCK_166_MHZ		(0 << 0)
+#define   I945_GC_RENDER_CLOCK_200_MHZ		(1 << 0)
+#define   I945_GC_RENDER_CLOCK_250_MHZ		(3 << 0)
+#define   I945_GC_RENDER_CLOCK_400_MHZ		(5 << 0)
+#define   I915_GC_RENDER_CLOCK_MASK		(7 << 0)
+#define   I915_GC_RENDER_CLOCK_166_MHZ		(0 << 0)
+#define   I915_GC_RENDER_CLOCK_200_MHZ		(1 << 0)
+#define   I915_GC_RENDER_CLOCK_333_MHZ		(4 << 0)
+
+#define ASLE					0xe4
+#define ASLS					0xfc
+
+#define SWSCI					0xe8
+#define   SWSCI_SCISEL				(1 << 15)
+#define   SWSCI_GSSCIE				(1 << 0)
+
+/* legacy/combination backlight modes, also called LBB */
+#define LBPC					0xf4
+
+#endif /* __INTEL_PCI_CONFIG_H__ */
diff --git a/drivers/gpu/drm/i915/intel_pcode.c b/drivers/gpu/drm/i915/intel_pcode.c
new file mode 100644
index 000000000..a234d9b4e
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_pcode.c
@@ -0,0 +1,248 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2013-2021 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "intel_pcode.h"
+
+static int gen6_check_mailbox_status(u32 mbox)
+{
+	switch (mbox & GEN6_PCODE_ERROR_MASK) {
+	case GEN6_PCODE_SUCCESS:
+		return 0;
+	case GEN6_PCODE_UNIMPLEMENTED_CMD:
+		return -ENODEV;
+	case GEN6_PCODE_ILLEGAL_CMD:
+		return -ENXIO;
+	case GEN6_PCODE_MIN_FREQ_TABLE_GT_RATIO_OUT_OF_RANGE:
+	case GEN7_PCODE_MIN_FREQ_TABLE_GT_RATIO_OUT_OF_RANGE:
+		return -EOVERFLOW;
+	case GEN6_PCODE_TIMEOUT:
+		return -ETIMEDOUT;
+	default:
+		MISSING_CASE(mbox & GEN6_PCODE_ERROR_MASK);
+		return 0;
+	}
+}
+
+static int gen7_check_mailbox_status(u32 mbox)
+{
+	switch (mbox & GEN6_PCODE_ERROR_MASK) {
+	case GEN6_PCODE_SUCCESS:
+		return 0;
+	case GEN6_PCODE_ILLEGAL_CMD:
+		return -ENXIO;
+	case GEN7_PCODE_TIMEOUT:
+		return -ETIMEDOUT;
+	case GEN7_PCODE_ILLEGAL_DATA:
+		return -EINVAL;
+	case GEN11_PCODE_ILLEGAL_SUBCOMMAND:
+		return -ENXIO;
+	case GEN11_PCODE_LOCKED:
+		return -EBUSY;
+	case GEN11_PCODE_REJECTED:
+		return -EACCES;
+	case GEN7_PCODE_MIN_FREQ_TABLE_GT_RATIO_OUT_OF_RANGE:
+		return -EOVERFLOW;
+	default:
+		MISSING_CASE(mbox & GEN6_PCODE_ERROR_MASK);
+		return 0;
+	}
+}
+
+static int __snb_pcode_rw(struct intel_uncore *uncore, u32 mbox,
+			  u32 *val, u32 *val1,
+			  int fast_timeout_us, int slow_timeout_ms,
+			  bool is_read)
+{
+	lockdep_assert_held(&uncore->i915->sb_lock);
+
+	/*
+	 * GEN6_PCODE_* are outside of the forcewake domain, we can use
+	 * intel_uncore_read/write_fw variants to reduce the amount of work
+	 * required when reading/writing.
+	 */
+
+	if (intel_uncore_read_fw(uncore, GEN6_PCODE_MAILBOX) & GEN6_PCODE_READY)
+		return -EAGAIN;
+
+	intel_uncore_write_fw(uncore, GEN6_PCODE_DATA, *val);
+	intel_uncore_write_fw(uncore, GEN6_PCODE_DATA1, val1 ? *val1 : 0);
+	intel_uncore_write_fw(uncore,
+			      GEN6_PCODE_MAILBOX, GEN6_PCODE_READY | mbox);
+
+	if (__intel_wait_for_register_fw(uncore,
+					 GEN6_PCODE_MAILBOX,
+					 GEN6_PCODE_READY, 0,
+					 fast_timeout_us,
+					 slow_timeout_ms,
+					 &mbox))
+		return -ETIMEDOUT;
+
+	if (is_read)
+		*val = intel_uncore_read_fw(uncore, GEN6_PCODE_DATA);
+	if (is_read && val1)
+		*val1 = intel_uncore_read_fw(uncore, GEN6_PCODE_DATA1);
+
+	if (GRAPHICS_VER(uncore->i915) > 6)
+		return gen7_check_mailbox_status(mbox);
+	else
+		return gen6_check_mailbox_status(mbox);
+}
+
+int snb_pcode_read(struct intel_uncore *uncore, u32 mbox, u32 *val, u32 *val1)
+{
+	int err;
+
+	mutex_lock(&uncore->i915->sb_lock);
+	err = __snb_pcode_rw(uncore, mbox, val, val1, 500, 20, true);
+	mutex_unlock(&uncore->i915->sb_lock);
+
+	if (err) {
+		drm_dbg(&uncore->i915->drm,
+			"warning: pcode (read from mbox %x) mailbox access failed for %ps: %d\n",
+			mbox, __builtin_return_address(0), err);
+	}
+
+	return err;
+}
+
+int snb_pcode_write_timeout(struct intel_uncore *uncore, u32 mbox, u32 val,
+			    int fast_timeout_us, int slow_timeout_ms)
+{
+	int err;
+
+	mutex_lock(&uncore->i915->sb_lock);
+	err = __snb_pcode_rw(uncore, mbox, &val, NULL,
+			     fast_timeout_us, slow_timeout_ms, false);
+	mutex_unlock(&uncore->i915->sb_lock);
+
+	if (err) {
+		drm_dbg(&uncore->i915->drm,
+			"warning: pcode (write of 0x%08x to mbox %x) mailbox access failed for %ps: %d\n",
+			val, mbox, __builtin_return_address(0), err);
+	}
+
+	return err;
+}
+
+static bool skl_pcode_try_request(struct intel_uncore *uncore, u32 mbox,
+				  u32 request, u32 reply_mask, u32 reply,
+				  u32 *status)
+{
+	*status = __snb_pcode_rw(uncore, mbox, &request, NULL, 500, 0, true);
+
+	return (*status == 0) && ((request & reply_mask) == reply);
+}
+
+/**
+ * skl_pcode_request - send PCODE request until acknowledgment
+ * @uncore: uncore
+ * @mbox: PCODE mailbox ID the request is targeted for
+ * @request: request ID
+ * @reply_mask: mask used to check for request acknowledgment
+ * @reply: value used to check for request acknowledgment
+ * @timeout_base_ms: timeout for polling with preemption enabled
+ *
+ * Keep resending the @request to @mbox until PCODE acknowledges it, PCODE
+ * reports an error or an overall timeout of @timeout_base_ms+50 ms expires.
+ * The request is acknowledged once the PCODE reply dword equals @reply after
+ * applying @reply_mask. Polling is first attempted with preemption enabled
+ * for @timeout_base_ms and if this times out for another 50 ms with
+ * preemption disabled.
+ *
+ * Returns 0 on success, %-ETIMEDOUT in case of a timeout, <0 in case of some
+ * other error as reported by PCODE.
+ */
+int skl_pcode_request(struct intel_uncore *uncore, u32 mbox, u32 request,
+		      u32 reply_mask, u32 reply, int timeout_base_ms)
+{
+	u32 status;
+	int ret;
+
+	mutex_lock(&uncore->i915->sb_lock);
+
+#define COND \
+	skl_pcode_try_request(uncore, mbox, request, reply_mask, reply, &status)
+
+	/*
+	 * Prime the PCODE by doing a request first. Normally it guarantees
+	 * that a subsequent request, at most @timeout_base_ms later, succeeds.
+	 * _wait_for() doesn't guarantee when its passed condition is evaluated
+	 * first, so send the first request explicitly.
+	 */
+	if (COND) {
+		ret = 0;
+		goto out;
+	}
+	ret = _wait_for(COND, timeout_base_ms * 1000, 10, 10);
+	if (!ret)
+		goto out;
+
+	/*
+	 * The above can time out if the number of requests was low (2 in the
+	 * worst case) _and_ PCODE was busy for some reason even after a
+	 * (queued) request and @timeout_base_ms delay. As a workaround retry
+	 * the poll with preemption disabled to maximize the number of
+	 * requests. Increase the timeout from @timeout_base_ms to 50ms to
+	 * account for interrupts that could reduce the number of these
+	 * requests, and for any quirks of the PCODE firmware that delays
+	 * the request completion.
+	 */
+	drm_dbg_kms(&uncore->i915->drm,
+		    "PCODE timeout, retrying with preemption disabled\n");
+	drm_WARN_ON_ONCE(&uncore->i915->drm, timeout_base_ms > 3);
+	preempt_disable();
+	ret = wait_for_atomic(COND, 50);
+	preempt_enable();
+
+out:
+	mutex_unlock(&uncore->i915->sb_lock);
+	return status ? status : ret;
+#undef COND
+}
+
+int intel_pcode_init(struct intel_uncore *uncore)
+{
+	if (!IS_DGFX(uncore->i915))
+		return 0;
+
+	return skl_pcode_request(uncore, DG1_PCODE_STATUS,
+				 DG1_UNCORE_GET_INIT_STATUS,
+				 DG1_UNCORE_INIT_STATUS_COMPLETE,
+				 DG1_UNCORE_INIT_STATUS_COMPLETE, 180000);
+}
+
+int snb_pcode_read_p(struct intel_uncore *uncore, u32 mbcmd, u32 p1, u32 p2, u32 *val)
+{
+	intel_wakeref_t wakeref;
+	u32 mbox;
+	int err;
+
+	mbox = REG_FIELD_PREP(GEN6_PCODE_MB_COMMAND, mbcmd)
+		| REG_FIELD_PREP(GEN6_PCODE_MB_PARAM1, p1)
+		| REG_FIELD_PREP(GEN6_PCODE_MB_PARAM2, p2);
+
+	with_intel_runtime_pm(uncore->rpm, wakeref)
+		err = snb_pcode_read(uncore, mbox, val, NULL);
+
+	return err;
+}
+
+int snb_pcode_write_p(struct intel_uncore *uncore, u32 mbcmd, u32 p1, u32 p2, u32 val)
+{
+	intel_wakeref_t wakeref;
+	u32 mbox;
+	int err;
+
+	mbox = REG_FIELD_PREP(GEN6_PCODE_MB_COMMAND, mbcmd)
+		| REG_FIELD_PREP(GEN6_PCODE_MB_PARAM1, p1)
+		| REG_FIELD_PREP(GEN6_PCODE_MB_PARAM2, p2);
+
+	with_intel_runtime_pm(uncore->rpm, wakeref)
+		err = snb_pcode_write(uncore, mbox, val);
+
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/intel_pcode.h b/drivers/gpu/drm/i915/intel_pcode.h
new file mode 100644
index 000000000..8d2198e29
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_pcode.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2013-2021 Intel Corporation
+ */
+
+#ifndef _INTEL_PCODE_H_
+#define _INTEL_PCODE_H_
+
+#include <linux/types.h>
+
+struct intel_uncore;
+
+int snb_pcode_read(struct intel_uncore *uncore, u32 mbox, u32 *val, u32 *val1);
+int snb_pcode_write_timeout(struct intel_uncore *uncore, u32 mbox, u32 val,
+			    int fast_timeout_us, int slow_timeout_ms);
+#define snb_pcode_write(uncore, mbox, val) \
+	snb_pcode_write_timeout(uncore, mbox, val, 500, 0)
+
+int skl_pcode_request(struct intel_uncore *uncore, u32 mbox, u32 request,
+		      u32 reply_mask, u32 reply, int timeout_base_ms);
+
+int intel_pcode_init(struct intel_uncore *uncore);
+
+/*
+ * Helpers for dGfx PCODE mailbox command formatting
+ */
+int snb_pcode_read_p(struct intel_uncore *uncore, u32 mbcmd, u32 p1, u32 p2, u32 *val);
+int snb_pcode_write_p(struct intel_uncore *uncore, u32 mbcmd, u32 p1, u32 p2, u32 val);
+
+#endif /* _INTEL_PCODE_H */
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
new file mode 100644
index 000000000..8f86f56e7
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -0,0 +1,5052 @@
+/*
+ * Copyright © 2012 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eugeni Dodonov <eugeni.dodonov@intel.com>
+ *
+ */
+
+#include "display/intel_de.h"
+#include "display/intel_display_trace.h"
+#include "display/skl_watermark.h"
+
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gt_regs.h"
+
+#include "i915_drv.h"
+#include "intel_mchbar_regs.h"
+#include "intel_pm.h"
+#include "vlv_sideband.h"
+
+struct drm_i915_clock_gating_funcs {
+	void (*init_clock_gating)(struct drm_i915_private *i915);
+};
+
+/* used in computing the new watermarks state */
+struct intel_wm_config {
+	unsigned int num_pipes_active;
+	bool sprites_enabled;
+	bool sprites_scaled;
+};
+
+static void gen9_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	if (HAS_LLC(dev_priv)) {
+		/*
+		 * WaCompressedResourceDisplayNewHashMode:skl,kbl
+		 * Display WA #0390: skl,kbl
+		 *
+		 * Must match Sampler, Pixel Back End, and Media. See
+		 * WaCompressedResourceSamplerPbeMediaNewHashMode.
+		 */
+		intel_uncore_write(&dev_priv->uncore, CHICKEN_PAR1_1,
+			   intel_uncore_read(&dev_priv->uncore, CHICKEN_PAR1_1) |
+			   SKL_DE_COMPRESSED_HASH_MODE);
+	}
+
+	/* See Bspec note for PSR2_CTL bit 31, Wa#828:skl,bxt,kbl,cfl */
+	intel_uncore_write(&dev_priv->uncore, CHICKEN_PAR1_1,
+		   intel_uncore_read(&dev_priv->uncore, CHICKEN_PAR1_1) | SKL_EDP_PSR_FIX_RDWRAP);
+
+	/* WaEnableChickenDCPR:skl,bxt,kbl,glk,cfl */
+	intel_uncore_write(&dev_priv->uncore, GEN8_CHICKEN_DCPR_1,
+		   intel_uncore_read(&dev_priv->uncore, GEN8_CHICKEN_DCPR_1) | MASK_WAKEMEM);
+
+	/*
+	 * WaFbcWakeMemOn:skl,bxt,kbl,glk,cfl
+	 * Display WA #0859: skl,bxt,kbl,glk,cfl
+	 */
+	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
+		   DISP_FBC_MEMORY_WAKE);
+}
+
+static void bxt_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	gen9_init_clock_gating(dev_priv);
+
+	/* WaDisableSDEUnitClockGating:bxt */
+	intel_uncore_write(&dev_priv->uncore, GEN8_UCGCTL6, intel_uncore_read(&dev_priv->uncore, GEN8_UCGCTL6) |
+		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
+
+	/*
+	 * FIXME:
+	 * GEN8_HDCUNIT_CLOCK_GATE_DISABLE_HDCREQ applies on 3x6 GT SKUs only.
+	 */
+	intel_uncore_write(&dev_priv->uncore, GEN8_UCGCTL6, intel_uncore_read(&dev_priv->uncore, GEN8_UCGCTL6) |
+		   GEN8_HDCUNIT_CLOCK_GATE_DISABLE_HDCREQ);
+
+	/*
+	 * Wa: Backlight PWM may stop in the asserted state, causing backlight
+	 * to stay fully on.
+	 */
+	intel_uncore_write(&dev_priv->uncore, GEN9_CLKGATE_DIS_0, intel_uncore_read(&dev_priv->uncore, GEN9_CLKGATE_DIS_0) |
+		   PWM1_GATING_DIS | PWM2_GATING_DIS);
+
+	/*
+	 * Lower the display internal timeout.
+	 * This is needed to avoid any hard hangs when DSI port PLL
+	 * is off and a MMIO access is attempted by any privilege
+	 * application, using batch buffers or any other means.
+	 */
+	intel_uncore_write(&dev_priv->uncore, RM_TIMEOUT, MMIO_TIMEOUT_US(950));
+
+	/*
+	 * WaFbcTurnOffFbcWatermark:bxt
+	 * Display WA #0562: bxt
+	 */
+	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
+		   DISP_FBC_WM_DIS);
+
+	/*
+	 * WaFbcHighMemBwCorruptionAvoidance:bxt
+	 * Display WA #0883: bxt
+	 */
+	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
+			   intel_uncore_read(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A)) |
+			   DPFC_DISABLE_DUMMY0);
+}
+
+static void glk_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	gen9_init_clock_gating(dev_priv);
+
+	/*
+	 * WaDisablePWMClockGating:glk
+	 * Backlight PWM may stop in the asserted state, causing backlight
+	 * to stay fully on.
+	 */
+	intel_uncore_write(&dev_priv->uncore, GEN9_CLKGATE_DIS_0, intel_uncore_read(&dev_priv->uncore, GEN9_CLKGATE_DIS_0) |
+		   PWM1_GATING_DIS | PWM2_GATING_DIS);
+}
+
+static void pnv_get_mem_freq(struct drm_i915_private *dev_priv)
+{
+	u32 tmp;
+
+	tmp = intel_uncore_read(&dev_priv->uncore, CLKCFG);
+
+	switch (tmp & CLKCFG_FSB_MASK) {
+	case CLKCFG_FSB_533:
+		dev_priv->fsb_freq = 533; /* 133*4 */
+		break;
+	case CLKCFG_FSB_800:
+		dev_priv->fsb_freq = 800; /* 200*4 */
+		break;
+	case CLKCFG_FSB_667:
+		dev_priv->fsb_freq =  667; /* 167*4 */
+		break;
+	case CLKCFG_FSB_400:
+		dev_priv->fsb_freq = 400; /* 100*4 */
+		break;
+	}
+
+	switch (tmp & CLKCFG_MEM_MASK) {
+	case CLKCFG_MEM_533:
+		dev_priv->mem_freq = 533;
+		break;
+	case CLKCFG_MEM_667:
+		dev_priv->mem_freq = 667;
+		break;
+	case CLKCFG_MEM_800:
+		dev_priv->mem_freq = 800;
+		break;
+	}
+
+	/* detect pineview DDR3 setting */
+	tmp = intel_uncore_read(&dev_priv->uncore, CSHRDDR3CTL);
+	dev_priv->is_ddr3 = (tmp & CSHRDDR3CTL_DDR3) ? 1 : 0;
+}
+
+static void ilk_get_mem_freq(struct drm_i915_private *dev_priv)
+{
+	u16 ddrpll, csipll;
+
+	ddrpll = intel_uncore_read16(&dev_priv->uncore, DDRMPLL1);
+	csipll = intel_uncore_read16(&dev_priv->uncore, CSIPLL0);
+
+	switch (ddrpll & 0xff) {
+	case 0xc:
+		dev_priv->mem_freq = 800;
+		break;
+	case 0x10:
+		dev_priv->mem_freq = 1066;
+		break;
+	case 0x14:
+		dev_priv->mem_freq = 1333;
+		break;
+	case 0x18:
+		dev_priv->mem_freq = 1600;
+		break;
+	default:
+		drm_dbg(&dev_priv->drm, "unknown memory frequency 0x%02x\n",
+			ddrpll & 0xff);
+		dev_priv->mem_freq = 0;
+		break;
+	}
+
+	switch (csipll & 0x3ff) {
+	case 0x00c:
+		dev_priv->fsb_freq = 3200;
+		break;
+	case 0x00e:
+		dev_priv->fsb_freq = 3733;
+		break;
+	case 0x010:
+		dev_priv->fsb_freq = 4266;
+		break;
+	case 0x012:
+		dev_priv->fsb_freq = 4800;
+		break;
+	case 0x014:
+		dev_priv->fsb_freq = 5333;
+		break;
+	case 0x016:
+		dev_priv->fsb_freq = 5866;
+		break;
+	case 0x018:
+		dev_priv->fsb_freq = 6400;
+		break;
+	default:
+		drm_dbg(&dev_priv->drm, "unknown fsb frequency 0x%04x\n",
+			csipll & 0x3ff);
+		dev_priv->fsb_freq = 0;
+		break;
+	}
+}
+
+static const struct cxsr_latency cxsr_latency_table[] = {
+	{1, 0, 800, 400, 3382, 33382, 3983, 33983},    /* DDR2-400 SC */
+	{1, 0, 800, 667, 3354, 33354, 3807, 33807},    /* DDR2-667 SC */
+	{1, 0, 800, 800, 3347, 33347, 3763, 33763},    /* DDR2-800 SC */
+	{1, 1, 800, 667, 6420, 36420, 6873, 36873},    /* DDR3-667 SC */
+	{1, 1, 800, 800, 5902, 35902, 6318, 36318},    /* DDR3-800 SC */
+
+	{1, 0, 667, 400, 3400, 33400, 4021, 34021},    /* DDR2-400 SC */
+	{1, 0, 667, 667, 3372, 33372, 3845, 33845},    /* DDR2-667 SC */
+	{1, 0, 667, 800, 3386, 33386, 3822, 33822},    /* DDR2-800 SC */
+	{1, 1, 667, 667, 6438, 36438, 6911, 36911},    /* DDR3-667 SC */
+	{1, 1, 667, 800, 5941, 35941, 6377, 36377},    /* DDR3-800 SC */
+
+	{1, 0, 400, 400, 3472, 33472, 4173, 34173},    /* DDR2-400 SC */
+	{1, 0, 400, 667, 3443, 33443, 3996, 33996},    /* DDR2-667 SC */
+	{1, 0, 400, 800, 3430, 33430, 3946, 33946},    /* DDR2-800 SC */
+	{1, 1, 400, 667, 6509, 36509, 7062, 37062},    /* DDR3-667 SC */
+	{1, 1, 400, 800, 5985, 35985, 6501, 36501},    /* DDR3-800 SC */
+
+	{0, 0, 800, 400, 3438, 33438, 4065, 34065},    /* DDR2-400 SC */
+	{0, 0, 800, 667, 3410, 33410, 3889, 33889},    /* DDR2-667 SC */
+	{0, 0, 800, 800, 3403, 33403, 3845, 33845},    /* DDR2-800 SC */
+	{0, 1, 800, 667, 6476, 36476, 6955, 36955},    /* DDR3-667 SC */
+	{0, 1, 800, 800, 5958, 35958, 6400, 36400},    /* DDR3-800 SC */
+
+	{0, 0, 667, 400, 3456, 33456, 4103, 34106},    /* DDR2-400 SC */
+	{0, 0, 667, 667, 3428, 33428, 3927, 33927},    /* DDR2-667 SC */
+	{0, 0, 667, 800, 3443, 33443, 3905, 33905},    /* DDR2-800 SC */
+	{0, 1, 667, 667, 6494, 36494, 6993, 36993},    /* DDR3-667 SC */
+	{0, 1, 667, 800, 5998, 35998, 6460, 36460},    /* DDR3-800 SC */
+
+	{0, 0, 400, 400, 3528, 33528, 4255, 34255},    /* DDR2-400 SC */
+	{0, 0, 400, 667, 3500, 33500, 4079, 34079},    /* DDR2-667 SC */
+	{0, 0, 400, 800, 3487, 33487, 4029, 34029},    /* DDR2-800 SC */
+	{0, 1, 400, 667, 6566, 36566, 7145, 37145},    /* DDR3-667 SC */
+	{0, 1, 400, 800, 6042, 36042, 6584, 36584},    /* DDR3-800 SC */
+};
+
+static const struct cxsr_latency *intel_get_cxsr_latency(bool is_desktop,
+							 bool is_ddr3,
+							 int fsb,
+							 int mem)
+{
+	const struct cxsr_latency *latency;
+	int i;
+
+	if (fsb == 0 || mem == 0)
+		return NULL;
+
+	for (i = 0; i < ARRAY_SIZE(cxsr_latency_table); i++) {
+		latency = &cxsr_latency_table[i];
+		if (is_desktop == latency->is_desktop &&
+		    is_ddr3 == latency->is_ddr3 &&
+		    fsb == latency->fsb_freq && mem == latency->mem_freq)
+			return latency;
+	}
+
+	DRM_DEBUG_KMS("Unknown FSB/MEM found, disable CxSR\n");
+
+	return NULL;
+}
+
+static void chv_set_memory_dvfs(struct drm_i915_private *dev_priv, bool enable)
+{
+	u32 val;
+
+	vlv_punit_get(dev_priv);
+
+	val = vlv_punit_read(dev_priv, PUNIT_REG_DDR_SETUP2);
+	if (enable)
+		val &= ~FORCE_DDR_HIGH_FREQ;
+	else
+		val |= FORCE_DDR_HIGH_FREQ;
+	val &= ~FORCE_DDR_LOW_FREQ;
+	val |= FORCE_DDR_FREQ_REQ_ACK;
+	vlv_punit_write(dev_priv, PUNIT_REG_DDR_SETUP2, val);
+
+	if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DDR_SETUP2) &
+		      FORCE_DDR_FREQ_REQ_ACK) == 0, 3))
+		drm_err(&dev_priv->drm,
+			"timed out waiting for Punit DDR DVFS request\n");
+
+	vlv_punit_put(dev_priv);
+}
+
+static void chv_set_memory_pm5(struct drm_i915_private *dev_priv, bool enable)
+{
+	u32 val;
+
+	vlv_punit_get(dev_priv);
+
+	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM);
+	if (enable)
+		val |= DSP_MAXFIFO_PM5_ENABLE;
+	else
+		val &= ~DSP_MAXFIFO_PM5_ENABLE;
+	vlv_punit_write(dev_priv, PUNIT_REG_DSPSSPM, val);
+
+	vlv_punit_put(dev_priv);
+}
+
+#define FW_WM(value, plane) \
+	(((value) << DSPFW_ ## plane ## _SHIFT) & DSPFW_ ## plane ## _MASK)
+
+static bool _intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enable)
+{
+	bool was_enabled;
+	u32 val;
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		was_enabled = intel_uncore_read(&dev_priv->uncore, FW_BLC_SELF_VLV) & FW_CSPWRDWNEN;
+		intel_uncore_write(&dev_priv->uncore, FW_BLC_SELF_VLV, enable ? FW_CSPWRDWNEN : 0);
+		intel_uncore_posting_read(&dev_priv->uncore, FW_BLC_SELF_VLV);
+	} else if (IS_G4X(dev_priv) || IS_I965GM(dev_priv)) {
+		was_enabled = intel_uncore_read(&dev_priv->uncore, FW_BLC_SELF) & FW_BLC_SELF_EN;
+		intel_uncore_write(&dev_priv->uncore, FW_BLC_SELF, enable ? FW_BLC_SELF_EN : 0);
+		intel_uncore_posting_read(&dev_priv->uncore, FW_BLC_SELF);
+	} else if (IS_PINEVIEW(dev_priv)) {
+		val = intel_uncore_read(&dev_priv->uncore, DSPFW3);
+		was_enabled = val & PINEVIEW_SELF_REFRESH_EN;
+		if (enable)
+			val |= PINEVIEW_SELF_REFRESH_EN;
+		else
+			val &= ~PINEVIEW_SELF_REFRESH_EN;
+		intel_uncore_write(&dev_priv->uncore, DSPFW3, val);
+		intel_uncore_posting_read(&dev_priv->uncore, DSPFW3);
+	} else if (IS_I945G(dev_priv) || IS_I945GM(dev_priv)) {
+		was_enabled = intel_uncore_read(&dev_priv->uncore, FW_BLC_SELF) & FW_BLC_SELF_EN;
+		val = enable ? _MASKED_BIT_ENABLE(FW_BLC_SELF_EN) :
+			       _MASKED_BIT_DISABLE(FW_BLC_SELF_EN);
+		intel_uncore_write(&dev_priv->uncore, FW_BLC_SELF, val);
+		intel_uncore_posting_read(&dev_priv->uncore, FW_BLC_SELF);
+	} else if (IS_I915GM(dev_priv)) {
+		/*
+		 * FIXME can't find a bit like this for 915G, and
+		 * and yet it does have the related watermark in
+		 * FW_BLC_SELF. What's going on?
+		 */
+		was_enabled = intel_uncore_read(&dev_priv->uncore, INSTPM) & INSTPM_SELF_EN;
+		val = enable ? _MASKED_BIT_ENABLE(INSTPM_SELF_EN) :
+			       _MASKED_BIT_DISABLE(INSTPM_SELF_EN);
+		intel_uncore_write(&dev_priv->uncore, INSTPM, val);
+		intel_uncore_posting_read(&dev_priv->uncore, INSTPM);
+	} else {
+		return false;
+	}
+
+	trace_intel_memory_cxsr(dev_priv, was_enabled, enable);
+
+	drm_dbg_kms(&dev_priv->drm, "memory self-refresh is %s (was %s)\n",
+		    str_enabled_disabled(enable),
+		    str_enabled_disabled(was_enabled));
+
+	return was_enabled;
+}
+
+/**
+ * intel_set_memory_cxsr - Configure CxSR state
+ * @dev_priv: i915 device
+ * @enable: Allow vs. disallow CxSR
+ *
+ * Allow or disallow the system to enter a special CxSR
+ * (C-state self refresh) state. What typically happens in CxSR mode
+ * is that several display FIFOs may get combined into a single larger
+ * FIFO for a particular plane (so called max FIFO mode) to allow the
+ * system to defer memory fetches longer, and the memory will enter
+ * self refresh.
+ *
+ * Note that enabling CxSR does not guarantee that the system enter
+ * this special mode, nor does it guarantee that the system stays
+ * in that mode once entered. So this just allows/disallows the system
+ * to autonomously utilize the CxSR mode. Other factors such as core
+ * C-states will affect when/if the system actually enters/exits the
+ * CxSR mode.
+ *
+ * Note that on VLV/CHV this actually only controls the max FIFO mode,
+ * and the system is free to enter/exit memory self refresh at any time
+ * even when the use of CxSR has been disallowed.
+ *
+ * While the system is actually in the CxSR/max FIFO mode, some plane
+ * control registers will not get latched on vblank. Thus in order to
+ * guarantee the system will respond to changes in the plane registers
+ * we must always disallow CxSR prior to making changes to those registers.
+ * Unfortunately the system will re-evaluate the CxSR conditions at
+ * frame start which happens after vblank start (which is when the plane
+ * registers would get latched), so we can't proceed with the plane update
+ * during the same frame where we disallowed CxSR.
+ *
+ * Certain platforms also have a deeper HPLL SR mode. Fortunately the
+ * HPLL SR mode depends on CxSR itself, so we don't have to hand hold
+ * the hardware w.r.t. HPLL SR when writing to plane registers.
+ * Disallowing just CxSR is sufficient.
+ */
+bool intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enable)
+{
+	bool ret;
+
+	mutex_lock(&dev_priv->display.wm.wm_mutex);
+	ret = _intel_set_memory_cxsr(dev_priv, enable);
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		dev_priv->display.wm.vlv.cxsr = enable;
+	else if (IS_G4X(dev_priv))
+		dev_priv->display.wm.g4x.cxsr = enable;
+	mutex_unlock(&dev_priv->display.wm.wm_mutex);
+
+	return ret;
+}
+
+/*
+ * Latency for FIFO fetches is dependent on several factors:
+ *   - memory configuration (speed, channels)
+ *   - chipset
+ *   - current MCH state
+ * It can be fairly high in some situations, so here we assume a fairly
+ * pessimal value.  It's a tradeoff between extra memory fetches (if we
+ * set this value too high, the FIFO will fetch frequently to stay full)
+ * and power consumption (set it too low to save power and we might see
+ * FIFO underruns and display "flicker").
+ *
+ * A value of 5us seems to be a good balance; safe for very low end
+ * platforms but not overly aggressive on lower latency configs.
+ */
+static const int pessimal_latency_ns = 5000;
+
+#define VLV_FIFO_START(dsparb, dsparb2, lo_shift, hi_shift) \
+	((((dsparb) >> (lo_shift)) & 0xff) | ((((dsparb2) >> (hi_shift)) & 0x1) << 8))
+
+static void vlv_get_fifo_size(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct vlv_fifo_state *fifo_state = &crtc_state->wm.vlv.fifo_state;
+	enum pipe pipe = crtc->pipe;
+	int sprite0_start, sprite1_start;
+	u32 dsparb, dsparb2, dsparb3;
+
+	switch (pipe) {
+	case PIPE_A:
+		dsparb = intel_uncore_read(&dev_priv->uncore, DSPARB);
+		dsparb2 = intel_uncore_read(&dev_priv->uncore, DSPARB2);
+		sprite0_start = VLV_FIFO_START(dsparb, dsparb2, 0, 0);
+		sprite1_start = VLV_FIFO_START(dsparb, dsparb2, 8, 4);
+		break;
+	case PIPE_B:
+		dsparb = intel_uncore_read(&dev_priv->uncore, DSPARB);
+		dsparb2 = intel_uncore_read(&dev_priv->uncore, DSPARB2);
+		sprite0_start = VLV_FIFO_START(dsparb, dsparb2, 16, 8);
+		sprite1_start = VLV_FIFO_START(dsparb, dsparb2, 24, 12);
+		break;
+	case PIPE_C:
+		dsparb2 = intel_uncore_read(&dev_priv->uncore, DSPARB2);
+		dsparb3 = intel_uncore_read(&dev_priv->uncore, DSPARB3);
+		sprite0_start = VLV_FIFO_START(dsparb3, dsparb2, 0, 16);
+		sprite1_start = VLV_FIFO_START(dsparb3, dsparb2, 8, 20);
+		break;
+	default:
+		MISSING_CASE(pipe);
+		return;
+	}
+
+	fifo_state->plane[PLANE_PRIMARY] = sprite0_start;
+	fifo_state->plane[PLANE_SPRITE0] = sprite1_start - sprite0_start;
+	fifo_state->plane[PLANE_SPRITE1] = 511 - sprite1_start;
+	fifo_state->plane[PLANE_CURSOR] = 63;
+}
+
+static int i9xx_get_fifo_size(struct drm_i915_private *dev_priv,
+			      enum i9xx_plane_id i9xx_plane)
+{
+	u32 dsparb = intel_uncore_read(&dev_priv->uncore, DSPARB);
+	int size;
+
+	size = dsparb & 0x7f;
+	if (i9xx_plane == PLANE_B)
+		size = ((dsparb >> DSPARB_CSTART_SHIFT) & 0x7f) - size;
+
+	drm_dbg_kms(&dev_priv->drm, "FIFO size - (0x%08x) %c: %d\n",
+		    dsparb, plane_name(i9xx_plane), size);
+
+	return size;
+}
+
+static int i830_get_fifo_size(struct drm_i915_private *dev_priv,
+			      enum i9xx_plane_id i9xx_plane)
+{
+	u32 dsparb = intel_uncore_read(&dev_priv->uncore, DSPARB);
+	int size;
+
+	size = dsparb & 0x1ff;
+	if (i9xx_plane == PLANE_B)
+		size = ((dsparb >> DSPARB_BEND_SHIFT) & 0x1ff) - size;
+	size >>= 1; /* Convert to cachelines */
+
+	drm_dbg_kms(&dev_priv->drm, "FIFO size - (0x%08x) %c: %d\n",
+		    dsparb, plane_name(i9xx_plane), size);
+
+	return size;
+}
+
+static int i845_get_fifo_size(struct drm_i915_private *dev_priv,
+			      enum i9xx_plane_id i9xx_plane)
+{
+	u32 dsparb = intel_uncore_read(&dev_priv->uncore, DSPARB);
+	int size;
+
+	size = dsparb & 0x7f;
+	size >>= 2; /* Convert to cachelines */
+
+	drm_dbg_kms(&dev_priv->drm, "FIFO size - (0x%08x) %c: %d\n",
+		    dsparb, plane_name(i9xx_plane), size);
+
+	return size;
+}
+
+/* Pineview has different values for various configs */
+static const struct intel_watermark_params pnv_display_wm = {
+	.fifo_size = PINEVIEW_DISPLAY_FIFO,
+	.max_wm = PINEVIEW_MAX_WM,
+	.default_wm = PINEVIEW_DFT_WM,
+	.guard_size = PINEVIEW_GUARD_WM,
+	.cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
+};
+
+static const struct intel_watermark_params pnv_display_hplloff_wm = {
+	.fifo_size = PINEVIEW_DISPLAY_FIFO,
+	.max_wm = PINEVIEW_MAX_WM,
+	.default_wm = PINEVIEW_DFT_HPLLOFF_WM,
+	.guard_size = PINEVIEW_GUARD_WM,
+	.cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
+};
+
+static const struct intel_watermark_params pnv_cursor_wm = {
+	.fifo_size = PINEVIEW_CURSOR_FIFO,
+	.max_wm = PINEVIEW_CURSOR_MAX_WM,
+	.default_wm = PINEVIEW_CURSOR_DFT_WM,
+	.guard_size = PINEVIEW_CURSOR_GUARD_WM,
+	.cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
+};
+
+static const struct intel_watermark_params pnv_cursor_hplloff_wm = {
+	.fifo_size = PINEVIEW_CURSOR_FIFO,
+	.max_wm = PINEVIEW_CURSOR_MAX_WM,
+	.default_wm = PINEVIEW_CURSOR_DFT_WM,
+	.guard_size = PINEVIEW_CURSOR_GUARD_WM,
+	.cacheline_size = PINEVIEW_FIFO_LINE_SIZE,
+};
+
+static const struct intel_watermark_params i965_cursor_wm_info = {
+	.fifo_size = I965_CURSOR_FIFO,
+	.max_wm = I965_CURSOR_MAX_WM,
+	.default_wm = I965_CURSOR_DFT_WM,
+	.guard_size = 2,
+	.cacheline_size = I915_FIFO_LINE_SIZE,
+};
+
+static const struct intel_watermark_params i945_wm_info = {
+	.fifo_size = I945_FIFO_SIZE,
+	.max_wm = I915_MAX_WM,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I915_FIFO_LINE_SIZE,
+};
+
+static const struct intel_watermark_params i915_wm_info = {
+	.fifo_size = I915_FIFO_SIZE,
+	.max_wm = I915_MAX_WM,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I915_FIFO_LINE_SIZE,
+};
+
+static const struct intel_watermark_params i830_a_wm_info = {
+	.fifo_size = I855GM_FIFO_SIZE,
+	.max_wm = I915_MAX_WM,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I830_FIFO_LINE_SIZE,
+};
+
+static const struct intel_watermark_params i830_bc_wm_info = {
+	.fifo_size = I855GM_FIFO_SIZE,
+	.max_wm = I915_MAX_WM/2,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I830_FIFO_LINE_SIZE,
+};
+
+static const struct intel_watermark_params i845_wm_info = {
+	.fifo_size = I830_FIFO_SIZE,
+	.max_wm = I915_MAX_WM,
+	.default_wm = 1,
+	.guard_size = 2,
+	.cacheline_size = I830_FIFO_LINE_SIZE,
+};
+
+/**
+ * intel_wm_method1 - Method 1 / "small buffer" watermark formula
+ * @pixel_rate: Pipe pixel rate in kHz
+ * @cpp: Plane bytes per pixel
+ * @latency: Memory wakeup latency in 0.1us units
+ *
+ * Compute the watermark using the method 1 or "small buffer"
+ * formula. The caller may additonally add extra cachelines
+ * to account for TLB misses and clock crossings.
+ *
+ * This method is concerned with the short term drain rate
+ * of the FIFO, ie. it does not account for blanking periods
+ * which would effectively reduce the average drain rate across
+ * a longer period. The name "small" refers to the fact the
+ * FIFO is relatively small compared to the amount of data
+ * fetched.
+ *
+ * The FIFO level vs. time graph might look something like:
+ *
+ *   |\   |\
+ *   | \  | \
+ * __---__---__ (- plane active, _ blanking)
+ * -> time
+ *
+ * or perhaps like this:
+ *
+ *   |\|\  |\|\
+ * __----__----__ (- plane active, _ blanking)
+ * -> time
+ *
+ * Returns:
+ * The watermark in bytes
+ */
+static unsigned int intel_wm_method1(unsigned int pixel_rate,
+				     unsigned int cpp,
+				     unsigned int latency)
+{
+	u64 ret;
+
+	ret = mul_u32_u32(pixel_rate, cpp * latency);
+	ret = DIV_ROUND_UP_ULL(ret, 10000);
+
+	return ret;
+}
+
+/**
+ * intel_wm_method2 - Method 2 / "large buffer" watermark formula
+ * @pixel_rate: Pipe pixel rate in kHz
+ * @htotal: Pipe horizontal total
+ * @width: Plane width in pixels
+ * @cpp: Plane bytes per pixel
+ * @latency: Memory wakeup latency in 0.1us units
+ *
+ * Compute the watermark using the method 2 or "large buffer"
+ * formula. The caller may additonally add extra cachelines
+ * to account for TLB misses and clock crossings.
+ *
+ * This method is concerned with the long term drain rate
+ * of the FIFO, ie. it does account for blanking periods
+ * which effectively reduce the average drain rate across
+ * a longer period. The name "large" refers to the fact the
+ * FIFO is relatively large compared to the amount of data
+ * fetched.
+ *
+ * The FIFO level vs. time graph might look something like:
+ *
+ *    |\___       |\___
+ *    |    \___   |    \___
+ *    |        \  |        \
+ * __ --__--__--__--__--__--__ (- plane active, _ blanking)
+ * -> time
+ *
+ * Returns:
+ * The watermark in bytes
+ */
+static unsigned int intel_wm_method2(unsigned int pixel_rate,
+				     unsigned int htotal,
+				     unsigned int width,
+				     unsigned int cpp,
+				     unsigned int latency)
+{
+	unsigned int ret;
+
+	/*
+	 * FIXME remove once all users are computing
+	 * watermarks in the correct place.
+	 */
+	if (WARN_ON_ONCE(htotal == 0))
+		htotal = 1;
+
+	ret = (latency * pixel_rate) / (htotal * 10000);
+	ret = (ret + 1) * width * cpp;
+
+	return ret;
+}
+
+/**
+ * intel_calculate_wm - calculate watermark level
+ * @pixel_rate: pixel clock
+ * @wm: chip FIFO params
+ * @fifo_size: size of the FIFO buffer
+ * @cpp: bytes per pixel
+ * @latency_ns: memory latency for the platform
+ *
+ * Calculate the watermark level (the level at which the display plane will
+ * start fetching from memory again).  Each chip has a different display
+ * FIFO size and allocation, so the caller needs to figure that out and pass
+ * in the correct intel_watermark_params structure.
+ *
+ * As the pixel clock runs, the FIFO will be drained at a rate that depends
+ * on the pixel size.  When it reaches the watermark level, it'll start
+ * fetching FIFO line sized based chunks from memory until the FIFO fills
+ * past the watermark point.  If the FIFO drains completely, a FIFO underrun
+ * will occur, and a display engine hang could result.
+ */
+static unsigned int intel_calculate_wm(int pixel_rate,
+				       const struct intel_watermark_params *wm,
+				       int fifo_size, int cpp,
+				       unsigned int latency_ns)
+{
+	int entries, wm_size;
+
+	/*
+	 * Note: we need to make sure we don't overflow for various clock &
+	 * latency values.
+	 * clocks go from a few thousand to several hundred thousand.
+	 * latency is usually a few thousand
+	 */
+	entries = intel_wm_method1(pixel_rate, cpp,
+				   latency_ns / 100);
+	entries = DIV_ROUND_UP(entries, wm->cacheline_size) +
+		wm->guard_size;
+	DRM_DEBUG_KMS("FIFO entries required for mode: %d\n", entries);
+
+	wm_size = fifo_size - entries;
+	DRM_DEBUG_KMS("FIFO watermark level: %d\n", wm_size);
+
+	/* Don't promote wm_size to unsigned... */
+	if (wm_size > wm->max_wm)
+		wm_size = wm->max_wm;
+	if (wm_size <= 0)
+		wm_size = wm->default_wm;
+
+	/*
+	 * Bspec seems to indicate that the value shouldn't be lower than
+	 * 'burst size + 1'. Certainly 830 is quite unhappy with low values.
+	 * Lets go for 8 which is the burst size since certain platforms
+	 * already use a hardcoded 8 (which is what the spec says should be
+	 * done).
+	 */
+	if (wm_size <= 8)
+		wm_size = 8;
+
+	return wm_size;
+}
+
+static bool is_disabling(int old, int new, int threshold)
+{
+	return old >= threshold && new < threshold;
+}
+
+static bool is_enabling(int old, int new, int threshold)
+{
+	return old < threshold && new >= threshold;
+}
+
+static int intel_wm_num_levels(struct drm_i915_private *dev_priv)
+{
+	return dev_priv->display.wm.max_level + 1;
+}
+
+bool intel_wm_plane_visible(const struct intel_crtc_state *crtc_state,
+			    const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+
+	/* FIXME check the 'enable' instead */
+	if (!crtc_state->hw.active)
+		return false;
+
+	/*
+	 * Treat cursor with fb as always visible since cursor updates
+	 * can happen faster than the vrefresh rate, and the current
+	 * watermark code doesn't handle that correctly. Cursor updates
+	 * which set/clear the fb or change the cursor size are going
+	 * to get throttled by intel_legacy_cursor_update() to work
+	 * around this problem with the watermark code.
+	 */
+	if (plane->id == PLANE_CURSOR)
+		return plane_state->hw.fb != NULL;
+	else
+		return plane_state->uapi.visible;
+}
+
+static bool intel_crtc_active(struct intel_crtc *crtc)
+{
+	/* Be paranoid as we can arrive here with only partial
+	 * state retrieved from the hardware during setup.
+	 *
+	 * We can ditch the adjusted_mode.crtc_clock check as soon
+	 * as Haswell has gained clock readout/fastboot support.
+	 *
+	 * We can ditch the crtc->primary->state->fb check as soon as we can
+	 * properly reconstruct framebuffers.
+	 *
+	 * FIXME: The intel_crtc->active here should be switched to
+	 * crtc->state->active once we have proper CRTC states wired up
+	 * for atomic.
+	 */
+	return crtc && crtc->active && crtc->base.primary->state->fb &&
+		crtc->config->hw.adjusted_mode.crtc_clock;
+}
+
+static struct intel_crtc *single_enabled_crtc(struct drm_i915_private *dev_priv)
+{
+	struct intel_crtc *crtc, *enabled = NULL;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		if (intel_crtc_active(crtc)) {
+			if (enabled)
+				return NULL;
+			enabled = crtc;
+		}
+	}
+
+	return enabled;
+}
+
+static void pnv_update_wm(struct drm_i915_private *dev_priv)
+{
+	struct intel_crtc *crtc;
+	const struct cxsr_latency *latency;
+	u32 reg;
+	unsigned int wm;
+
+	latency = intel_get_cxsr_latency(!IS_MOBILE(dev_priv),
+					 dev_priv->is_ddr3,
+					 dev_priv->fsb_freq,
+					 dev_priv->mem_freq);
+	if (!latency) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "Unknown FSB/MEM found, disable CxSR\n");
+		intel_set_memory_cxsr(dev_priv, false);
+		return;
+	}
+
+	crtc = single_enabled_crtc(dev_priv);
+	if (crtc) {
+		const struct drm_framebuffer *fb =
+			crtc->base.primary->state->fb;
+		int pixel_rate = crtc->config->pixel_rate;
+		int cpp = fb->format->cpp[0];
+
+		/* Display SR */
+		wm = intel_calculate_wm(pixel_rate, &pnv_display_wm,
+					pnv_display_wm.fifo_size,
+					cpp, latency->display_sr);
+		reg = intel_uncore_read(&dev_priv->uncore, DSPFW1);
+		reg &= ~DSPFW_SR_MASK;
+		reg |= FW_WM(wm, SR);
+		intel_uncore_write(&dev_priv->uncore, DSPFW1, reg);
+		drm_dbg_kms(&dev_priv->drm, "DSPFW1 register is %x\n", reg);
+
+		/* cursor SR */
+		wm = intel_calculate_wm(pixel_rate, &pnv_cursor_wm,
+					pnv_display_wm.fifo_size,
+					4, latency->cursor_sr);
+		reg = intel_uncore_read(&dev_priv->uncore, DSPFW3);
+		reg &= ~DSPFW_CURSOR_SR_MASK;
+		reg |= FW_WM(wm, CURSOR_SR);
+		intel_uncore_write(&dev_priv->uncore, DSPFW3, reg);
+
+		/* Display HPLL off SR */
+		wm = intel_calculate_wm(pixel_rate, &pnv_display_hplloff_wm,
+					pnv_display_hplloff_wm.fifo_size,
+					cpp, latency->display_hpll_disable);
+		reg = intel_uncore_read(&dev_priv->uncore, DSPFW3);
+		reg &= ~DSPFW_HPLL_SR_MASK;
+		reg |= FW_WM(wm, HPLL_SR);
+		intel_uncore_write(&dev_priv->uncore, DSPFW3, reg);
+
+		/* cursor HPLL off SR */
+		wm = intel_calculate_wm(pixel_rate, &pnv_cursor_hplloff_wm,
+					pnv_display_hplloff_wm.fifo_size,
+					4, latency->cursor_hpll_disable);
+		reg = intel_uncore_read(&dev_priv->uncore, DSPFW3);
+		reg &= ~DSPFW_HPLL_CURSOR_MASK;
+		reg |= FW_WM(wm, HPLL_CURSOR);
+		intel_uncore_write(&dev_priv->uncore, DSPFW3, reg);
+		drm_dbg_kms(&dev_priv->drm, "DSPFW3 register is %x\n", reg);
+
+		intel_set_memory_cxsr(dev_priv, true);
+	} else {
+		intel_set_memory_cxsr(dev_priv, false);
+	}
+}
+
+/*
+ * Documentation says:
+ * "If the line size is small, the TLB fetches can get in the way of the
+ *  data fetches, causing some lag in the pixel data return which is not
+ *  accounted for in the above formulas. The following adjustment only
+ *  needs to be applied if eight whole lines fit in the buffer at once.
+ *  The WM is adjusted upwards by the difference between the FIFO size
+ *  and the size of 8 whole lines. This adjustment is always performed
+ *  in the actual pixel depth regardless of whether FBC is enabled or not."
+ */
+static unsigned int g4x_tlb_miss_wa(int fifo_size, int width, int cpp)
+{
+	int tlb_miss = fifo_size * 64 - width * cpp * 8;
+
+	return max(0, tlb_miss);
+}
+
+static void g4x_write_wm_values(struct drm_i915_private *dev_priv,
+				const struct g4x_wm_values *wm)
+{
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe)
+		trace_g4x_wm(intel_crtc_for_pipe(dev_priv, pipe), wm);
+
+	intel_uncore_write(&dev_priv->uncore, DSPFW1,
+		   FW_WM(wm->sr.plane, SR) |
+		   FW_WM(wm->pipe[PIPE_B].plane[PLANE_CURSOR], CURSORB) |
+		   FW_WM(wm->pipe[PIPE_B].plane[PLANE_PRIMARY], PLANEB) |
+		   FW_WM(wm->pipe[PIPE_A].plane[PLANE_PRIMARY], PLANEA));
+	intel_uncore_write(&dev_priv->uncore, DSPFW2,
+		   (wm->fbc_en ? DSPFW_FBC_SR_EN : 0) |
+		   FW_WM(wm->sr.fbc, FBC_SR) |
+		   FW_WM(wm->hpll.fbc, FBC_HPLL_SR) |
+		   FW_WM(wm->pipe[PIPE_B].plane[PLANE_SPRITE0], SPRITEB) |
+		   FW_WM(wm->pipe[PIPE_A].plane[PLANE_CURSOR], CURSORA) |
+		   FW_WM(wm->pipe[PIPE_A].plane[PLANE_SPRITE0], SPRITEA));
+	intel_uncore_write(&dev_priv->uncore, DSPFW3,
+		   (wm->hpll_en ? DSPFW_HPLL_SR_EN : 0) |
+		   FW_WM(wm->sr.cursor, CURSOR_SR) |
+		   FW_WM(wm->hpll.cursor, HPLL_CURSOR) |
+		   FW_WM(wm->hpll.plane, HPLL_SR));
+
+	intel_uncore_posting_read(&dev_priv->uncore, DSPFW1);
+}
+
+#define FW_WM_VLV(value, plane) \
+	(((value) << DSPFW_ ## plane ## _SHIFT) & DSPFW_ ## plane ## _MASK_VLV)
+
+static void vlv_write_wm_values(struct drm_i915_private *dev_priv,
+				const struct vlv_wm_values *wm)
+{
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe) {
+		trace_vlv_wm(intel_crtc_for_pipe(dev_priv, pipe), wm);
+
+		intel_uncore_write(&dev_priv->uncore, VLV_DDL(pipe),
+			   (wm->ddl[pipe].plane[PLANE_CURSOR] << DDL_CURSOR_SHIFT) |
+			   (wm->ddl[pipe].plane[PLANE_SPRITE1] << DDL_SPRITE_SHIFT(1)) |
+			   (wm->ddl[pipe].plane[PLANE_SPRITE0] << DDL_SPRITE_SHIFT(0)) |
+			   (wm->ddl[pipe].plane[PLANE_PRIMARY] << DDL_PLANE_SHIFT));
+	}
+
+	/*
+	 * Zero the (unused) WM1 watermarks, and also clear all the
+	 * high order bits so that there are no out of bounds values
+	 * present in the registers during the reprogramming.
+	 */
+	intel_uncore_write(&dev_priv->uncore, DSPHOWM, 0);
+	intel_uncore_write(&dev_priv->uncore, DSPHOWM1, 0);
+	intel_uncore_write(&dev_priv->uncore, DSPFW4, 0);
+	intel_uncore_write(&dev_priv->uncore, DSPFW5, 0);
+	intel_uncore_write(&dev_priv->uncore, DSPFW6, 0);
+
+	intel_uncore_write(&dev_priv->uncore, DSPFW1,
+		   FW_WM(wm->sr.plane, SR) |
+		   FW_WM(wm->pipe[PIPE_B].plane[PLANE_CURSOR], CURSORB) |
+		   FW_WM_VLV(wm->pipe[PIPE_B].plane[PLANE_PRIMARY], PLANEB) |
+		   FW_WM_VLV(wm->pipe[PIPE_A].plane[PLANE_PRIMARY], PLANEA));
+	intel_uncore_write(&dev_priv->uncore, DSPFW2,
+		   FW_WM_VLV(wm->pipe[PIPE_A].plane[PLANE_SPRITE1], SPRITEB) |
+		   FW_WM(wm->pipe[PIPE_A].plane[PLANE_CURSOR], CURSORA) |
+		   FW_WM_VLV(wm->pipe[PIPE_A].plane[PLANE_SPRITE0], SPRITEA));
+	intel_uncore_write(&dev_priv->uncore, DSPFW3,
+		   FW_WM(wm->sr.cursor, CURSOR_SR));
+
+	if (IS_CHERRYVIEW(dev_priv)) {
+		intel_uncore_write(&dev_priv->uncore, DSPFW7_CHV,
+			   FW_WM_VLV(wm->pipe[PIPE_B].plane[PLANE_SPRITE1], SPRITED) |
+			   FW_WM_VLV(wm->pipe[PIPE_B].plane[PLANE_SPRITE0], SPRITEC));
+		intel_uncore_write(&dev_priv->uncore, DSPFW8_CHV,
+			   FW_WM_VLV(wm->pipe[PIPE_C].plane[PLANE_SPRITE1], SPRITEF) |
+			   FW_WM_VLV(wm->pipe[PIPE_C].plane[PLANE_SPRITE0], SPRITEE));
+		intel_uncore_write(&dev_priv->uncore, DSPFW9_CHV,
+			   FW_WM_VLV(wm->pipe[PIPE_C].plane[PLANE_PRIMARY], PLANEC) |
+			   FW_WM(wm->pipe[PIPE_C].plane[PLANE_CURSOR], CURSORC));
+		intel_uncore_write(&dev_priv->uncore, DSPHOWM,
+			   FW_WM(wm->sr.plane >> 9, SR_HI) |
+			   FW_WM(wm->pipe[PIPE_C].plane[PLANE_SPRITE1] >> 8, SPRITEF_HI) |
+			   FW_WM(wm->pipe[PIPE_C].plane[PLANE_SPRITE0] >> 8, SPRITEE_HI) |
+			   FW_WM(wm->pipe[PIPE_C].plane[PLANE_PRIMARY] >> 8, PLANEC_HI) |
+			   FW_WM(wm->pipe[PIPE_B].plane[PLANE_SPRITE1] >> 8, SPRITED_HI) |
+			   FW_WM(wm->pipe[PIPE_B].plane[PLANE_SPRITE0] >> 8, SPRITEC_HI) |
+			   FW_WM(wm->pipe[PIPE_B].plane[PLANE_PRIMARY] >> 8, PLANEB_HI) |
+			   FW_WM(wm->pipe[PIPE_A].plane[PLANE_SPRITE1] >> 8, SPRITEB_HI) |
+			   FW_WM(wm->pipe[PIPE_A].plane[PLANE_SPRITE0] >> 8, SPRITEA_HI) |
+			   FW_WM(wm->pipe[PIPE_A].plane[PLANE_PRIMARY] >> 8, PLANEA_HI));
+	} else {
+		intel_uncore_write(&dev_priv->uncore, DSPFW7,
+			   FW_WM_VLV(wm->pipe[PIPE_B].plane[PLANE_SPRITE1], SPRITED) |
+			   FW_WM_VLV(wm->pipe[PIPE_B].plane[PLANE_SPRITE0], SPRITEC));
+		intel_uncore_write(&dev_priv->uncore, DSPHOWM,
+			   FW_WM(wm->sr.plane >> 9, SR_HI) |
+			   FW_WM(wm->pipe[PIPE_B].plane[PLANE_SPRITE1] >> 8, SPRITED_HI) |
+			   FW_WM(wm->pipe[PIPE_B].plane[PLANE_SPRITE0] >> 8, SPRITEC_HI) |
+			   FW_WM(wm->pipe[PIPE_B].plane[PLANE_PRIMARY] >> 8, PLANEB_HI) |
+			   FW_WM(wm->pipe[PIPE_A].plane[PLANE_SPRITE1] >> 8, SPRITEB_HI) |
+			   FW_WM(wm->pipe[PIPE_A].plane[PLANE_SPRITE0] >> 8, SPRITEA_HI) |
+			   FW_WM(wm->pipe[PIPE_A].plane[PLANE_PRIMARY] >> 8, PLANEA_HI));
+	}
+
+	intel_uncore_posting_read(&dev_priv->uncore, DSPFW1);
+}
+
+#undef FW_WM_VLV
+
+static void g4x_setup_wm_latency(struct drm_i915_private *dev_priv)
+{
+	/* all latencies in usec */
+	dev_priv->display.wm.pri_latency[G4X_WM_LEVEL_NORMAL] = 5;
+	dev_priv->display.wm.pri_latency[G4X_WM_LEVEL_SR] = 12;
+	dev_priv->display.wm.pri_latency[G4X_WM_LEVEL_HPLL] = 35;
+
+	dev_priv->display.wm.max_level = G4X_WM_LEVEL_HPLL;
+}
+
+static int g4x_plane_fifo_size(enum plane_id plane_id, int level)
+{
+	/*
+	 * DSPCNTR[13] supposedly controls whether the
+	 * primary plane can use the FIFO space otherwise
+	 * reserved for the sprite plane. It's not 100% clear
+	 * what the actual FIFO size is, but it looks like we
+	 * can happily set both primary and sprite watermarks
+	 * up to 127 cachelines. So that would seem to mean
+	 * that either DSPCNTR[13] doesn't do anything, or that
+	 * the total FIFO is >= 256 cachelines in size. Either
+	 * way, we don't seem to have to worry about this
+	 * repartitioning as the maximum watermark value the
+	 * register can hold for each plane is lower than the
+	 * minimum FIFO size.
+	 */
+	switch (plane_id) {
+	case PLANE_CURSOR:
+		return 63;
+	case PLANE_PRIMARY:
+		return level == G4X_WM_LEVEL_NORMAL ? 127 : 511;
+	case PLANE_SPRITE0:
+		return level == G4X_WM_LEVEL_NORMAL ? 127 : 0;
+	default:
+		MISSING_CASE(plane_id);
+		return 0;
+	}
+}
+
+static int g4x_fbc_fifo_size(int level)
+{
+	switch (level) {
+	case G4X_WM_LEVEL_SR:
+		return 7;
+	case G4X_WM_LEVEL_HPLL:
+		return 15;
+	default:
+		MISSING_CASE(level);
+		return 0;
+	}
+}
+
+static u16 g4x_compute_wm(const struct intel_crtc_state *crtc_state,
+			  const struct intel_plane_state *plane_state,
+			  int level)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_display_mode *pipe_mode =
+		&crtc_state->hw.pipe_mode;
+	unsigned int latency = dev_priv->display.wm.pri_latency[level] * 10;
+	unsigned int pixel_rate, htotal, cpp, width, wm;
+
+	if (latency == 0)
+		return USHRT_MAX;
+
+	if (!intel_wm_plane_visible(crtc_state, plane_state))
+		return 0;
+
+	cpp = plane_state->hw.fb->format->cpp[0];
+
+	/*
+	 * WaUse32BppForSRWM:ctg,elk
+	 *
+	 * The spec fails to list this restriction for the
+	 * HPLL watermark, which seems a little strange.
+	 * Let's use 32bpp for the HPLL watermark as well.
+	 */
+	if (plane->id == PLANE_PRIMARY &&
+	    level != G4X_WM_LEVEL_NORMAL)
+		cpp = max(cpp, 4u);
+
+	pixel_rate = crtc_state->pixel_rate;
+	htotal = pipe_mode->crtc_htotal;
+	width = drm_rect_width(&plane_state->uapi.src) >> 16;
+
+	if (plane->id == PLANE_CURSOR) {
+		wm = intel_wm_method2(pixel_rate, htotal, width, cpp, latency);
+	} else if (plane->id == PLANE_PRIMARY &&
+		   level == G4X_WM_LEVEL_NORMAL) {
+		wm = intel_wm_method1(pixel_rate, cpp, latency);
+	} else {
+		unsigned int small, large;
+
+		small = intel_wm_method1(pixel_rate, cpp, latency);
+		large = intel_wm_method2(pixel_rate, htotal, width, cpp, latency);
+
+		wm = min(small, large);
+	}
+
+	wm += g4x_tlb_miss_wa(g4x_plane_fifo_size(plane->id, level),
+			      width, cpp);
+
+	wm = DIV_ROUND_UP(wm, 64) + 2;
+
+	return min_t(unsigned int, wm, USHRT_MAX);
+}
+
+static bool g4x_raw_plane_wm_set(struct intel_crtc_state *crtc_state,
+				 int level, enum plane_id plane_id, u16 value)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	bool dirty = false;
+
+	for (; level < intel_wm_num_levels(dev_priv); level++) {
+		struct g4x_pipe_wm *raw = &crtc_state->wm.g4x.raw[level];
+
+		dirty |= raw->plane[plane_id] != value;
+		raw->plane[plane_id] = value;
+	}
+
+	return dirty;
+}
+
+static bool g4x_raw_fbc_wm_set(struct intel_crtc_state *crtc_state,
+			       int level, u16 value)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	bool dirty = false;
+
+	/* NORMAL level doesn't have an FBC watermark */
+	level = max(level, G4X_WM_LEVEL_SR);
+
+	for (; level < intel_wm_num_levels(dev_priv); level++) {
+		struct g4x_pipe_wm *raw = &crtc_state->wm.g4x.raw[level];
+
+		dirty |= raw->fbc != value;
+		raw->fbc = value;
+	}
+
+	return dirty;
+}
+
+static u32 ilk_compute_fbc_wm(const struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state,
+			      u32 pri_val);
+
+static bool g4x_raw_plane_wm_compute(struct intel_crtc_state *crtc_state,
+				     const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	int num_levels = intel_wm_num_levels(to_i915(plane->base.dev));
+	enum plane_id plane_id = plane->id;
+	bool dirty = false;
+	int level;
+
+	if (!intel_wm_plane_visible(crtc_state, plane_state)) {
+		dirty |= g4x_raw_plane_wm_set(crtc_state, 0, plane_id, 0);
+		if (plane_id == PLANE_PRIMARY)
+			dirty |= g4x_raw_fbc_wm_set(crtc_state, 0, 0);
+		goto out;
+	}
+
+	for (level = 0; level < num_levels; level++) {
+		struct g4x_pipe_wm *raw = &crtc_state->wm.g4x.raw[level];
+		int wm, max_wm;
+
+		wm = g4x_compute_wm(crtc_state, plane_state, level);
+		max_wm = g4x_plane_fifo_size(plane_id, level);
+
+		if (wm > max_wm)
+			break;
+
+		dirty |= raw->plane[plane_id] != wm;
+		raw->plane[plane_id] = wm;
+
+		if (plane_id != PLANE_PRIMARY ||
+		    level == G4X_WM_LEVEL_NORMAL)
+			continue;
+
+		wm = ilk_compute_fbc_wm(crtc_state, plane_state,
+					raw->plane[plane_id]);
+		max_wm = g4x_fbc_fifo_size(level);
+
+		/*
+		 * FBC wm is not mandatory as we
+		 * can always just disable its use.
+		 */
+		if (wm > max_wm)
+			wm = USHRT_MAX;
+
+		dirty |= raw->fbc != wm;
+		raw->fbc = wm;
+	}
+
+	/* mark watermarks as invalid */
+	dirty |= g4x_raw_plane_wm_set(crtc_state, level, plane_id, USHRT_MAX);
+
+	if (plane_id == PLANE_PRIMARY)
+		dirty |= g4x_raw_fbc_wm_set(crtc_state, level, USHRT_MAX);
+
+ out:
+	if (dirty) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "%s watermarks: normal=%d, SR=%d, HPLL=%d\n",
+			    plane->base.name,
+			    crtc_state->wm.g4x.raw[G4X_WM_LEVEL_NORMAL].plane[plane_id],
+			    crtc_state->wm.g4x.raw[G4X_WM_LEVEL_SR].plane[plane_id],
+			    crtc_state->wm.g4x.raw[G4X_WM_LEVEL_HPLL].plane[plane_id]);
+
+		if (plane_id == PLANE_PRIMARY)
+			drm_dbg_kms(&dev_priv->drm,
+				    "FBC watermarks: SR=%d, HPLL=%d\n",
+				    crtc_state->wm.g4x.raw[G4X_WM_LEVEL_SR].fbc,
+				    crtc_state->wm.g4x.raw[G4X_WM_LEVEL_HPLL].fbc);
+	}
+
+	return dirty;
+}
+
+static bool g4x_raw_plane_wm_is_valid(const struct intel_crtc_state *crtc_state,
+				      enum plane_id plane_id, int level)
+{
+	const struct g4x_pipe_wm *raw = &crtc_state->wm.g4x.raw[level];
+
+	return raw->plane[plane_id] <= g4x_plane_fifo_size(plane_id, level);
+}
+
+static bool g4x_raw_crtc_wm_is_valid(const struct intel_crtc_state *crtc_state,
+				     int level)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+
+	if (level > dev_priv->display.wm.max_level)
+		return false;
+
+	return g4x_raw_plane_wm_is_valid(crtc_state, PLANE_PRIMARY, level) &&
+		g4x_raw_plane_wm_is_valid(crtc_state, PLANE_SPRITE0, level) &&
+		g4x_raw_plane_wm_is_valid(crtc_state, PLANE_CURSOR, level);
+}
+
+/* mark all levels starting from 'level' as invalid */
+static void g4x_invalidate_wms(struct intel_crtc *crtc,
+			       struct g4x_wm_state *wm_state, int level)
+{
+	if (level <= G4X_WM_LEVEL_NORMAL) {
+		enum plane_id plane_id;
+
+		for_each_plane_id_on_crtc(crtc, plane_id)
+			wm_state->wm.plane[plane_id] = USHRT_MAX;
+	}
+
+	if (level <= G4X_WM_LEVEL_SR) {
+		wm_state->cxsr = false;
+		wm_state->sr.cursor = USHRT_MAX;
+		wm_state->sr.plane = USHRT_MAX;
+		wm_state->sr.fbc = USHRT_MAX;
+	}
+
+	if (level <= G4X_WM_LEVEL_HPLL) {
+		wm_state->hpll_en = false;
+		wm_state->hpll.cursor = USHRT_MAX;
+		wm_state->hpll.plane = USHRT_MAX;
+		wm_state->hpll.fbc = USHRT_MAX;
+	}
+}
+
+static bool g4x_compute_fbc_en(const struct g4x_wm_state *wm_state,
+			       int level)
+{
+	if (level < G4X_WM_LEVEL_SR)
+		return false;
+
+	if (level >= G4X_WM_LEVEL_SR &&
+	    wm_state->sr.fbc > g4x_fbc_fifo_size(G4X_WM_LEVEL_SR))
+		return false;
+
+	if (level >= G4X_WM_LEVEL_HPLL &&
+	    wm_state->hpll.fbc > g4x_fbc_fifo_size(G4X_WM_LEVEL_HPLL))
+		return false;
+
+	return true;
+}
+
+static int g4x_compute_pipe_wm(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct g4x_wm_state *wm_state = &crtc_state->wm.g4x.optimal;
+	u8 active_planes = crtc_state->active_planes & ~BIT(PLANE_CURSOR);
+	const struct g4x_pipe_wm *raw;
+	const struct intel_plane_state *old_plane_state;
+	const struct intel_plane_state *new_plane_state;
+	struct intel_plane *plane;
+	enum plane_id plane_id;
+	int i, level;
+	unsigned int dirty = 0;
+
+	for_each_oldnew_intel_plane_in_state(state, plane,
+					     old_plane_state,
+					     new_plane_state, i) {
+		if (new_plane_state->hw.crtc != &crtc->base &&
+		    old_plane_state->hw.crtc != &crtc->base)
+			continue;
+
+		if (g4x_raw_plane_wm_compute(crtc_state, new_plane_state))
+			dirty |= BIT(plane->id);
+	}
+
+	if (!dirty)
+		return 0;
+
+	level = G4X_WM_LEVEL_NORMAL;
+	if (!g4x_raw_crtc_wm_is_valid(crtc_state, level))
+		goto out;
+
+	raw = &crtc_state->wm.g4x.raw[level];
+	for_each_plane_id_on_crtc(crtc, plane_id)
+		wm_state->wm.plane[plane_id] = raw->plane[plane_id];
+
+	level = G4X_WM_LEVEL_SR;
+	if (!g4x_raw_crtc_wm_is_valid(crtc_state, level))
+		goto out;
+
+	raw = &crtc_state->wm.g4x.raw[level];
+	wm_state->sr.plane = raw->plane[PLANE_PRIMARY];
+	wm_state->sr.cursor = raw->plane[PLANE_CURSOR];
+	wm_state->sr.fbc = raw->fbc;
+
+	wm_state->cxsr = active_planes == BIT(PLANE_PRIMARY);
+
+	level = G4X_WM_LEVEL_HPLL;
+	if (!g4x_raw_crtc_wm_is_valid(crtc_state, level))
+		goto out;
+
+	raw = &crtc_state->wm.g4x.raw[level];
+	wm_state->hpll.plane = raw->plane[PLANE_PRIMARY];
+	wm_state->hpll.cursor = raw->plane[PLANE_CURSOR];
+	wm_state->hpll.fbc = raw->fbc;
+
+	wm_state->hpll_en = wm_state->cxsr;
+
+	level++;
+
+ out:
+	if (level == G4X_WM_LEVEL_NORMAL)
+		return -EINVAL;
+
+	/* invalidate the higher levels */
+	g4x_invalidate_wms(crtc, wm_state, level);
+
+	/*
+	 * Determine if the FBC watermark(s) can be used. IF
+	 * this isn't the case we prefer to disable the FBC
+	 * watermark(s) rather than disable the SR/HPLL
+	 * level(s) entirely. 'level-1' is the highest valid
+	 * level here.
+	 */
+	wm_state->fbc_en = g4x_compute_fbc_en(wm_state, level - 1);
+
+	return 0;
+}
+
+static int g4x_compute_intermediate_wm(struct intel_atomic_state *state,
+				       struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct g4x_wm_state *intermediate = &new_crtc_state->wm.g4x.intermediate;
+	const struct g4x_wm_state *optimal = &new_crtc_state->wm.g4x.optimal;
+	const struct g4x_wm_state *active = &old_crtc_state->wm.g4x.optimal;
+	enum plane_id plane_id;
+
+	if (!new_crtc_state->hw.active ||
+	    drm_atomic_crtc_needs_modeset(&new_crtc_state->uapi)) {
+		*intermediate = *optimal;
+
+		intermediate->cxsr = false;
+		intermediate->hpll_en = false;
+		goto out;
+	}
+
+	intermediate->cxsr = optimal->cxsr && active->cxsr &&
+		!new_crtc_state->disable_cxsr;
+	intermediate->hpll_en = optimal->hpll_en && active->hpll_en &&
+		!new_crtc_state->disable_cxsr;
+	intermediate->fbc_en = optimal->fbc_en && active->fbc_en;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		intermediate->wm.plane[plane_id] =
+			max(optimal->wm.plane[plane_id],
+			    active->wm.plane[plane_id]);
+
+		drm_WARN_ON(&dev_priv->drm, intermediate->wm.plane[plane_id] >
+			    g4x_plane_fifo_size(plane_id, G4X_WM_LEVEL_NORMAL));
+	}
+
+	intermediate->sr.plane = max(optimal->sr.plane,
+				     active->sr.plane);
+	intermediate->sr.cursor = max(optimal->sr.cursor,
+				      active->sr.cursor);
+	intermediate->sr.fbc = max(optimal->sr.fbc,
+				   active->sr.fbc);
+
+	intermediate->hpll.plane = max(optimal->hpll.plane,
+				       active->hpll.plane);
+	intermediate->hpll.cursor = max(optimal->hpll.cursor,
+					active->hpll.cursor);
+	intermediate->hpll.fbc = max(optimal->hpll.fbc,
+				     active->hpll.fbc);
+
+	drm_WARN_ON(&dev_priv->drm,
+		    (intermediate->sr.plane >
+		     g4x_plane_fifo_size(PLANE_PRIMARY, G4X_WM_LEVEL_SR) ||
+		     intermediate->sr.cursor >
+		     g4x_plane_fifo_size(PLANE_CURSOR, G4X_WM_LEVEL_SR)) &&
+		    intermediate->cxsr);
+	drm_WARN_ON(&dev_priv->drm,
+		    (intermediate->sr.plane >
+		     g4x_plane_fifo_size(PLANE_PRIMARY, G4X_WM_LEVEL_HPLL) ||
+		     intermediate->sr.cursor >
+		     g4x_plane_fifo_size(PLANE_CURSOR, G4X_WM_LEVEL_HPLL)) &&
+		    intermediate->hpll_en);
+
+	drm_WARN_ON(&dev_priv->drm,
+		    intermediate->sr.fbc > g4x_fbc_fifo_size(1) &&
+		    intermediate->fbc_en && intermediate->cxsr);
+	drm_WARN_ON(&dev_priv->drm,
+		    intermediate->hpll.fbc > g4x_fbc_fifo_size(2) &&
+		    intermediate->fbc_en && intermediate->hpll_en);
+
+out:
+	/*
+	 * If our intermediate WM are identical to the final WM, then we can
+	 * omit the post-vblank programming; only update if it's different.
+	 */
+	if (memcmp(intermediate, optimal, sizeof(*intermediate)) != 0)
+		new_crtc_state->wm.need_postvbl_update = true;
+
+	return 0;
+}
+
+static void g4x_merge_wm(struct drm_i915_private *dev_priv,
+			 struct g4x_wm_values *wm)
+{
+	struct intel_crtc *crtc;
+	int num_active_pipes = 0;
+
+	wm->cxsr = true;
+	wm->hpll_en = true;
+	wm->fbc_en = true;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		const struct g4x_wm_state *wm_state = &crtc->wm.active.g4x;
+
+		if (!crtc->active)
+			continue;
+
+		if (!wm_state->cxsr)
+			wm->cxsr = false;
+		if (!wm_state->hpll_en)
+			wm->hpll_en = false;
+		if (!wm_state->fbc_en)
+			wm->fbc_en = false;
+
+		num_active_pipes++;
+	}
+
+	if (num_active_pipes != 1) {
+		wm->cxsr = false;
+		wm->hpll_en = false;
+		wm->fbc_en = false;
+	}
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		const struct g4x_wm_state *wm_state = &crtc->wm.active.g4x;
+		enum pipe pipe = crtc->pipe;
+
+		wm->pipe[pipe] = wm_state->wm;
+		if (crtc->active && wm->cxsr)
+			wm->sr = wm_state->sr;
+		if (crtc->active && wm->hpll_en)
+			wm->hpll = wm_state->hpll;
+	}
+}
+
+static void g4x_program_watermarks(struct drm_i915_private *dev_priv)
+{
+	struct g4x_wm_values *old_wm = &dev_priv->display.wm.g4x;
+	struct g4x_wm_values new_wm = {};
+
+	g4x_merge_wm(dev_priv, &new_wm);
+
+	if (memcmp(old_wm, &new_wm, sizeof(new_wm)) == 0)
+		return;
+
+	if (is_disabling(old_wm->cxsr, new_wm.cxsr, true))
+		_intel_set_memory_cxsr(dev_priv, false);
+
+	g4x_write_wm_values(dev_priv, &new_wm);
+
+	if (is_enabling(old_wm->cxsr, new_wm.cxsr, true))
+		_intel_set_memory_cxsr(dev_priv, true);
+
+	*old_wm = new_wm;
+}
+
+static void g4x_initial_watermarks(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	mutex_lock(&dev_priv->display.wm.wm_mutex);
+	crtc->wm.active.g4x = crtc_state->wm.g4x.intermediate;
+	g4x_program_watermarks(dev_priv);
+	mutex_unlock(&dev_priv->display.wm.wm_mutex);
+}
+
+static void g4x_optimize_watermarks(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (!crtc_state->wm.need_postvbl_update)
+		return;
+
+	mutex_lock(&dev_priv->display.wm.wm_mutex);
+	crtc->wm.active.g4x = crtc_state->wm.g4x.optimal;
+	g4x_program_watermarks(dev_priv);
+	mutex_unlock(&dev_priv->display.wm.wm_mutex);
+}
+
+/* latency must be in 0.1us units. */
+static unsigned int vlv_wm_method2(unsigned int pixel_rate,
+				   unsigned int htotal,
+				   unsigned int width,
+				   unsigned int cpp,
+				   unsigned int latency)
+{
+	unsigned int ret;
+
+	ret = intel_wm_method2(pixel_rate, htotal,
+			       width, cpp, latency);
+	ret = DIV_ROUND_UP(ret, 64);
+
+	return ret;
+}
+
+static void vlv_setup_wm_latency(struct drm_i915_private *dev_priv)
+{
+	/* all latencies in usec */
+	dev_priv->display.wm.pri_latency[VLV_WM_LEVEL_PM2] = 3;
+
+	dev_priv->display.wm.max_level = VLV_WM_LEVEL_PM2;
+
+	if (IS_CHERRYVIEW(dev_priv)) {
+		dev_priv->display.wm.pri_latency[VLV_WM_LEVEL_PM5] = 12;
+		dev_priv->display.wm.pri_latency[VLV_WM_LEVEL_DDR_DVFS] = 33;
+
+		dev_priv->display.wm.max_level = VLV_WM_LEVEL_DDR_DVFS;
+	}
+}
+
+static u16 vlv_compute_wm_level(const struct intel_crtc_state *crtc_state,
+				const struct intel_plane_state *plane_state,
+				int level)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	const struct drm_display_mode *pipe_mode =
+		&crtc_state->hw.pipe_mode;
+	unsigned int pixel_rate, htotal, cpp, width, wm;
+
+	if (dev_priv->display.wm.pri_latency[level] == 0)
+		return USHRT_MAX;
+
+	if (!intel_wm_plane_visible(crtc_state, plane_state))
+		return 0;
+
+	cpp = plane_state->hw.fb->format->cpp[0];
+	pixel_rate = crtc_state->pixel_rate;
+	htotal = pipe_mode->crtc_htotal;
+	width = drm_rect_width(&plane_state->uapi.src) >> 16;
+
+	if (plane->id == PLANE_CURSOR) {
+		/*
+		 * FIXME the formula gives values that are
+		 * too big for the cursor FIFO, and hence we
+		 * would never be able to use cursors. For
+		 * now just hardcode the watermark.
+		 */
+		wm = 63;
+	} else {
+		wm = vlv_wm_method2(pixel_rate, htotal, width, cpp,
+				    dev_priv->display.wm.pri_latency[level] * 10);
+	}
+
+	return min_t(unsigned int, wm, USHRT_MAX);
+}
+
+static bool vlv_need_sprite0_fifo_workaround(unsigned int active_planes)
+{
+	return (active_planes & (BIT(PLANE_SPRITE0) |
+				 BIT(PLANE_SPRITE1))) == BIT(PLANE_SPRITE1);
+}
+
+static int vlv_compute_fifo(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct g4x_pipe_wm *raw =
+		&crtc_state->wm.vlv.raw[VLV_WM_LEVEL_PM2];
+	struct vlv_fifo_state *fifo_state = &crtc_state->wm.vlv.fifo_state;
+	u8 active_planes = crtc_state->active_planes & ~BIT(PLANE_CURSOR);
+	int num_active_planes = hweight8(active_planes);
+	const int fifo_size = 511;
+	int fifo_extra, fifo_left = fifo_size;
+	int sprite0_fifo_extra = 0;
+	unsigned int total_rate;
+	enum plane_id plane_id;
+
+	/*
+	 * When enabling sprite0 after sprite1 has already been enabled
+	 * we tend to get an underrun unless sprite0 already has some
+	 * FIFO space allcoated. Hence we always allocate at least one
+	 * cacheline for sprite0 whenever sprite1 is enabled.
+	 *
+	 * All other plane enable sequences appear immune to this problem.
+	 */
+	if (vlv_need_sprite0_fifo_workaround(active_planes))
+		sprite0_fifo_extra = 1;
+
+	total_rate = raw->plane[PLANE_PRIMARY] +
+		raw->plane[PLANE_SPRITE0] +
+		raw->plane[PLANE_SPRITE1] +
+		sprite0_fifo_extra;
+
+	if (total_rate > fifo_size)
+		return -EINVAL;
+
+	if (total_rate == 0)
+		total_rate = 1;
+
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		unsigned int rate;
+
+		if ((active_planes & BIT(plane_id)) == 0) {
+			fifo_state->plane[plane_id] = 0;
+			continue;
+		}
+
+		rate = raw->plane[plane_id];
+		fifo_state->plane[plane_id] = fifo_size * rate / total_rate;
+		fifo_left -= fifo_state->plane[plane_id];
+	}
+
+	fifo_state->plane[PLANE_SPRITE0] += sprite0_fifo_extra;
+	fifo_left -= sprite0_fifo_extra;
+
+	fifo_state->plane[PLANE_CURSOR] = 63;
+
+	fifo_extra = DIV_ROUND_UP(fifo_left, num_active_planes ?: 1);
+
+	/* spread the remainder evenly */
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		int plane_extra;
+
+		if (fifo_left == 0)
+			break;
+
+		if ((active_planes & BIT(plane_id)) == 0)
+			continue;
+
+		plane_extra = min(fifo_extra, fifo_left);
+		fifo_state->plane[plane_id] += plane_extra;
+		fifo_left -= plane_extra;
+	}
+
+	drm_WARN_ON(&dev_priv->drm, active_planes != 0 && fifo_left != 0);
+
+	/* give it all to the first plane if none are active */
+	if (active_planes == 0) {
+		drm_WARN_ON(&dev_priv->drm, fifo_left != fifo_size);
+		fifo_state->plane[PLANE_PRIMARY] = fifo_left;
+	}
+
+	return 0;
+}
+
+/* mark all levels starting from 'level' as invalid */
+static void vlv_invalidate_wms(struct intel_crtc *crtc,
+			       struct vlv_wm_state *wm_state, int level)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	for (; level < intel_wm_num_levels(dev_priv); level++) {
+		enum plane_id plane_id;
+
+		for_each_plane_id_on_crtc(crtc, plane_id)
+			wm_state->wm[level].plane[plane_id] = USHRT_MAX;
+
+		wm_state->sr[level].cursor = USHRT_MAX;
+		wm_state->sr[level].plane = USHRT_MAX;
+	}
+}
+
+static u16 vlv_invert_wm_value(u16 wm, u16 fifo_size)
+{
+	if (wm > fifo_size)
+		return USHRT_MAX;
+	else
+		return fifo_size - wm;
+}
+
+/*
+ * Starting from 'level' set all higher
+ * levels to 'value' in the "raw" watermarks.
+ */
+static bool vlv_raw_plane_wm_set(struct intel_crtc_state *crtc_state,
+				 int level, enum plane_id plane_id, u16 value)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	int num_levels = intel_wm_num_levels(dev_priv);
+	bool dirty = false;
+
+	for (; level < num_levels; level++) {
+		struct g4x_pipe_wm *raw = &crtc_state->wm.vlv.raw[level];
+
+		dirty |= raw->plane[plane_id] != value;
+		raw->plane[plane_id] = value;
+	}
+
+	return dirty;
+}
+
+static bool vlv_raw_plane_wm_compute(struct intel_crtc_state *crtc_state,
+				     const struct intel_plane_state *plane_state)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->uapi.plane);
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->uapi.crtc->dev);
+	enum plane_id plane_id = plane->id;
+	int num_levels = intel_wm_num_levels(to_i915(plane->base.dev));
+	int level;
+	bool dirty = false;
+
+	if (!intel_wm_plane_visible(crtc_state, plane_state)) {
+		dirty |= vlv_raw_plane_wm_set(crtc_state, 0, plane_id, 0);
+		goto out;
+	}
+
+	for (level = 0; level < num_levels; level++) {
+		struct g4x_pipe_wm *raw = &crtc_state->wm.vlv.raw[level];
+		int wm = vlv_compute_wm_level(crtc_state, plane_state, level);
+		int max_wm = plane_id == PLANE_CURSOR ? 63 : 511;
+
+		if (wm > max_wm)
+			break;
+
+		dirty |= raw->plane[plane_id] != wm;
+		raw->plane[plane_id] = wm;
+	}
+
+	/* mark all higher levels as invalid */
+	dirty |= vlv_raw_plane_wm_set(crtc_state, level, plane_id, USHRT_MAX);
+
+out:
+	if (dirty)
+		drm_dbg_kms(&dev_priv->drm,
+			    "%s watermarks: PM2=%d, PM5=%d, DDR DVFS=%d\n",
+			    plane->base.name,
+			    crtc_state->wm.vlv.raw[VLV_WM_LEVEL_PM2].plane[plane_id],
+			    crtc_state->wm.vlv.raw[VLV_WM_LEVEL_PM5].plane[plane_id],
+			    crtc_state->wm.vlv.raw[VLV_WM_LEVEL_DDR_DVFS].plane[plane_id]);
+
+	return dirty;
+}
+
+static bool vlv_raw_plane_wm_is_valid(const struct intel_crtc_state *crtc_state,
+				      enum plane_id plane_id, int level)
+{
+	const struct g4x_pipe_wm *raw =
+		&crtc_state->wm.vlv.raw[level];
+	const struct vlv_fifo_state *fifo_state =
+		&crtc_state->wm.vlv.fifo_state;
+
+	return raw->plane[plane_id] <= fifo_state->plane[plane_id];
+}
+
+static bool vlv_raw_crtc_wm_is_valid(const struct intel_crtc_state *crtc_state, int level)
+{
+	return vlv_raw_plane_wm_is_valid(crtc_state, PLANE_PRIMARY, level) &&
+		vlv_raw_plane_wm_is_valid(crtc_state, PLANE_SPRITE0, level) &&
+		vlv_raw_plane_wm_is_valid(crtc_state, PLANE_SPRITE1, level) &&
+		vlv_raw_plane_wm_is_valid(crtc_state, PLANE_CURSOR, level);
+}
+
+static int vlv_compute_pipe_wm(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct vlv_wm_state *wm_state = &crtc_state->wm.vlv.optimal;
+	const struct vlv_fifo_state *fifo_state =
+		&crtc_state->wm.vlv.fifo_state;
+	u8 active_planes = crtc_state->active_planes & ~BIT(PLANE_CURSOR);
+	int num_active_planes = hweight8(active_planes);
+	bool needs_modeset = drm_atomic_crtc_needs_modeset(&crtc_state->uapi);
+	const struct intel_plane_state *old_plane_state;
+	const struct intel_plane_state *new_plane_state;
+	struct intel_plane *plane;
+	enum plane_id plane_id;
+	int level, ret, i;
+	unsigned int dirty = 0;
+
+	for_each_oldnew_intel_plane_in_state(state, plane,
+					     old_plane_state,
+					     new_plane_state, i) {
+		if (new_plane_state->hw.crtc != &crtc->base &&
+		    old_plane_state->hw.crtc != &crtc->base)
+			continue;
+
+		if (vlv_raw_plane_wm_compute(crtc_state, new_plane_state))
+			dirty |= BIT(plane->id);
+	}
+
+	/*
+	 * DSPARB registers may have been reset due to the
+	 * power well being turned off. Make sure we restore
+	 * them to a consistent state even if no primary/sprite
+	 * planes are initially active.
+	 */
+	if (needs_modeset)
+		crtc_state->fifo_changed = true;
+
+	if (!dirty)
+		return 0;
+
+	/* cursor changes don't warrant a FIFO recompute */
+	if (dirty & ~BIT(PLANE_CURSOR)) {
+		const struct intel_crtc_state *old_crtc_state =
+			intel_atomic_get_old_crtc_state(state, crtc);
+		const struct vlv_fifo_state *old_fifo_state =
+			&old_crtc_state->wm.vlv.fifo_state;
+
+		ret = vlv_compute_fifo(crtc_state);
+		if (ret)
+			return ret;
+
+		if (needs_modeset ||
+		    memcmp(old_fifo_state, fifo_state,
+			   sizeof(*fifo_state)) != 0)
+			crtc_state->fifo_changed = true;
+	}
+
+	/* initially allow all levels */
+	wm_state->num_levels = intel_wm_num_levels(dev_priv);
+	/*
+	 * Note that enabling cxsr with no primary/sprite planes
+	 * enabled can wedge the pipe. Hence we only allow cxsr
+	 * with exactly one enabled primary/sprite plane.
+	 */
+	wm_state->cxsr = crtc->pipe != PIPE_C && num_active_planes == 1;
+
+	for (level = 0; level < wm_state->num_levels; level++) {
+		const struct g4x_pipe_wm *raw = &crtc_state->wm.vlv.raw[level];
+		const int sr_fifo_size = INTEL_NUM_PIPES(dev_priv) * 512 - 1;
+
+		if (!vlv_raw_crtc_wm_is_valid(crtc_state, level))
+			break;
+
+		for_each_plane_id_on_crtc(crtc, plane_id) {
+			wm_state->wm[level].plane[plane_id] =
+				vlv_invert_wm_value(raw->plane[plane_id],
+						    fifo_state->plane[plane_id]);
+		}
+
+		wm_state->sr[level].plane =
+			vlv_invert_wm_value(max3(raw->plane[PLANE_PRIMARY],
+						 raw->plane[PLANE_SPRITE0],
+						 raw->plane[PLANE_SPRITE1]),
+					    sr_fifo_size);
+
+		wm_state->sr[level].cursor =
+			vlv_invert_wm_value(raw->plane[PLANE_CURSOR],
+					    63);
+	}
+
+	if (level == 0)
+		return -EINVAL;
+
+	/* limit to only levels we can actually handle */
+	wm_state->num_levels = level;
+
+	/* invalidate the higher levels */
+	vlv_invalidate_wms(crtc, wm_state, level);
+
+	return 0;
+}
+
+#define VLV_FIFO(plane, value) \
+	(((value) << DSPARB_ ## plane ## _SHIFT_VLV) & DSPARB_ ## plane ## _MASK_VLV)
+
+static void vlv_atomic_update_fifo(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_uncore *uncore = &dev_priv->uncore;
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct vlv_fifo_state *fifo_state =
+		&crtc_state->wm.vlv.fifo_state;
+	int sprite0_start, sprite1_start, fifo_size;
+	u32 dsparb, dsparb2, dsparb3;
+
+	if (!crtc_state->fifo_changed)
+		return;
+
+	sprite0_start = fifo_state->plane[PLANE_PRIMARY];
+	sprite1_start = fifo_state->plane[PLANE_SPRITE0] + sprite0_start;
+	fifo_size = fifo_state->plane[PLANE_SPRITE1] + sprite1_start;
+
+	drm_WARN_ON(&dev_priv->drm, fifo_state->plane[PLANE_CURSOR] != 63);
+	drm_WARN_ON(&dev_priv->drm, fifo_size != 511);
+
+	trace_vlv_fifo_size(crtc, sprite0_start, sprite1_start, fifo_size);
+
+	/*
+	 * uncore.lock serves a double purpose here. It allows us to
+	 * use the less expensive I915_{READ,WRITE}_FW() functions, and
+	 * it protects the DSPARB registers from getting clobbered by
+	 * parallel updates from multiple pipes.
+	 *
+	 * intel_pipe_update_start() has already disabled interrupts
+	 * for us, so a plain spin_lock() is sufficient here.
+	 */
+	spin_lock(&uncore->lock);
+
+	switch (crtc->pipe) {
+	case PIPE_A:
+		dsparb = intel_uncore_read_fw(uncore, DSPARB);
+		dsparb2 = intel_uncore_read_fw(uncore, DSPARB2);
+
+		dsparb &= ~(VLV_FIFO(SPRITEA, 0xff) |
+			    VLV_FIFO(SPRITEB, 0xff));
+		dsparb |= (VLV_FIFO(SPRITEA, sprite0_start) |
+			   VLV_FIFO(SPRITEB, sprite1_start));
+
+		dsparb2 &= ~(VLV_FIFO(SPRITEA_HI, 0x1) |
+			     VLV_FIFO(SPRITEB_HI, 0x1));
+		dsparb2 |= (VLV_FIFO(SPRITEA_HI, sprite0_start >> 8) |
+			   VLV_FIFO(SPRITEB_HI, sprite1_start >> 8));
+
+		intel_uncore_write_fw(uncore, DSPARB, dsparb);
+		intel_uncore_write_fw(uncore, DSPARB2, dsparb2);
+		break;
+	case PIPE_B:
+		dsparb = intel_uncore_read_fw(uncore, DSPARB);
+		dsparb2 = intel_uncore_read_fw(uncore, DSPARB2);
+
+		dsparb &= ~(VLV_FIFO(SPRITEC, 0xff) |
+			    VLV_FIFO(SPRITED, 0xff));
+		dsparb |= (VLV_FIFO(SPRITEC, sprite0_start) |
+			   VLV_FIFO(SPRITED, sprite1_start));
+
+		dsparb2 &= ~(VLV_FIFO(SPRITEC_HI, 0xff) |
+			     VLV_FIFO(SPRITED_HI, 0xff));
+		dsparb2 |= (VLV_FIFO(SPRITEC_HI, sprite0_start >> 8) |
+			   VLV_FIFO(SPRITED_HI, sprite1_start >> 8));
+
+		intel_uncore_write_fw(uncore, DSPARB, dsparb);
+		intel_uncore_write_fw(uncore, DSPARB2, dsparb2);
+		break;
+	case PIPE_C:
+		dsparb3 = intel_uncore_read_fw(uncore, DSPARB3);
+		dsparb2 = intel_uncore_read_fw(uncore, DSPARB2);
+
+		dsparb3 &= ~(VLV_FIFO(SPRITEE, 0xff) |
+			     VLV_FIFO(SPRITEF, 0xff));
+		dsparb3 |= (VLV_FIFO(SPRITEE, sprite0_start) |
+			    VLV_FIFO(SPRITEF, sprite1_start));
+
+		dsparb2 &= ~(VLV_FIFO(SPRITEE_HI, 0xff) |
+			     VLV_FIFO(SPRITEF_HI, 0xff));
+		dsparb2 |= (VLV_FIFO(SPRITEE_HI, sprite0_start >> 8) |
+			   VLV_FIFO(SPRITEF_HI, sprite1_start >> 8));
+
+		intel_uncore_write_fw(uncore, DSPARB3, dsparb3);
+		intel_uncore_write_fw(uncore, DSPARB2, dsparb2);
+		break;
+	default:
+		break;
+	}
+
+	intel_uncore_posting_read_fw(uncore, DSPARB);
+
+	spin_unlock(&uncore->lock);
+}
+
+#undef VLV_FIFO
+
+static int vlv_compute_intermediate_wm(struct intel_atomic_state *state,
+				       struct intel_crtc *crtc)
+{
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct vlv_wm_state *intermediate = &new_crtc_state->wm.vlv.intermediate;
+	const struct vlv_wm_state *optimal = &new_crtc_state->wm.vlv.optimal;
+	const struct vlv_wm_state *active = &old_crtc_state->wm.vlv.optimal;
+	int level;
+
+	if (!new_crtc_state->hw.active ||
+	    drm_atomic_crtc_needs_modeset(&new_crtc_state->uapi)) {
+		*intermediate = *optimal;
+
+		intermediate->cxsr = false;
+		goto out;
+	}
+
+	intermediate->num_levels = min(optimal->num_levels, active->num_levels);
+	intermediate->cxsr = optimal->cxsr && active->cxsr &&
+		!new_crtc_state->disable_cxsr;
+
+	for (level = 0; level < intermediate->num_levels; level++) {
+		enum plane_id plane_id;
+
+		for_each_plane_id_on_crtc(crtc, plane_id) {
+			intermediate->wm[level].plane[plane_id] =
+				min(optimal->wm[level].plane[plane_id],
+				    active->wm[level].plane[plane_id]);
+		}
+
+		intermediate->sr[level].plane = min(optimal->sr[level].plane,
+						    active->sr[level].plane);
+		intermediate->sr[level].cursor = min(optimal->sr[level].cursor,
+						     active->sr[level].cursor);
+	}
+
+	vlv_invalidate_wms(crtc, intermediate, level);
+
+out:
+	/*
+	 * If our intermediate WM are identical to the final WM, then we can
+	 * omit the post-vblank programming; only update if it's different.
+	 */
+	if (memcmp(intermediate, optimal, sizeof(*intermediate)) != 0)
+		new_crtc_state->wm.need_postvbl_update = true;
+
+	return 0;
+}
+
+static void vlv_merge_wm(struct drm_i915_private *dev_priv,
+			 struct vlv_wm_values *wm)
+{
+	struct intel_crtc *crtc;
+	int num_active_pipes = 0;
+
+	wm->level = dev_priv->display.wm.max_level;
+	wm->cxsr = true;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		const struct vlv_wm_state *wm_state = &crtc->wm.active.vlv;
+
+		if (!crtc->active)
+			continue;
+
+		if (!wm_state->cxsr)
+			wm->cxsr = false;
+
+		num_active_pipes++;
+		wm->level = min_t(int, wm->level, wm_state->num_levels - 1);
+	}
+
+	if (num_active_pipes != 1)
+		wm->cxsr = false;
+
+	if (num_active_pipes > 1)
+		wm->level = VLV_WM_LEVEL_PM2;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		const struct vlv_wm_state *wm_state = &crtc->wm.active.vlv;
+		enum pipe pipe = crtc->pipe;
+
+		wm->pipe[pipe] = wm_state->wm[wm->level];
+		if (crtc->active && wm->cxsr)
+			wm->sr = wm_state->sr[wm->level];
+
+		wm->ddl[pipe].plane[PLANE_PRIMARY] = DDL_PRECISION_HIGH | 2;
+		wm->ddl[pipe].plane[PLANE_SPRITE0] = DDL_PRECISION_HIGH | 2;
+		wm->ddl[pipe].plane[PLANE_SPRITE1] = DDL_PRECISION_HIGH | 2;
+		wm->ddl[pipe].plane[PLANE_CURSOR] = DDL_PRECISION_HIGH | 2;
+	}
+}
+
+static void vlv_program_watermarks(struct drm_i915_private *dev_priv)
+{
+	struct vlv_wm_values *old_wm = &dev_priv->display.wm.vlv;
+	struct vlv_wm_values new_wm = {};
+
+	vlv_merge_wm(dev_priv, &new_wm);
+
+	if (memcmp(old_wm, &new_wm, sizeof(new_wm)) == 0)
+		return;
+
+	if (is_disabling(old_wm->level, new_wm.level, VLV_WM_LEVEL_DDR_DVFS))
+		chv_set_memory_dvfs(dev_priv, false);
+
+	if (is_disabling(old_wm->level, new_wm.level, VLV_WM_LEVEL_PM5))
+		chv_set_memory_pm5(dev_priv, false);
+
+	if (is_disabling(old_wm->cxsr, new_wm.cxsr, true))
+		_intel_set_memory_cxsr(dev_priv, false);
+
+	vlv_write_wm_values(dev_priv, &new_wm);
+
+	if (is_enabling(old_wm->cxsr, new_wm.cxsr, true))
+		_intel_set_memory_cxsr(dev_priv, true);
+
+	if (is_enabling(old_wm->level, new_wm.level, VLV_WM_LEVEL_PM5))
+		chv_set_memory_pm5(dev_priv, true);
+
+	if (is_enabling(old_wm->level, new_wm.level, VLV_WM_LEVEL_DDR_DVFS))
+		chv_set_memory_dvfs(dev_priv, true);
+
+	*old_wm = new_wm;
+}
+
+static void vlv_initial_watermarks(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	mutex_lock(&dev_priv->display.wm.wm_mutex);
+	crtc->wm.active.vlv = crtc_state->wm.vlv.intermediate;
+	vlv_program_watermarks(dev_priv);
+	mutex_unlock(&dev_priv->display.wm.wm_mutex);
+}
+
+static void vlv_optimize_watermarks(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (!crtc_state->wm.need_postvbl_update)
+		return;
+
+	mutex_lock(&dev_priv->display.wm.wm_mutex);
+	crtc->wm.active.vlv = crtc_state->wm.vlv.optimal;
+	vlv_program_watermarks(dev_priv);
+	mutex_unlock(&dev_priv->display.wm.wm_mutex);
+}
+
+static void i965_update_wm(struct drm_i915_private *dev_priv)
+{
+	struct intel_crtc *crtc;
+	int srwm = 1;
+	int cursor_sr = 16;
+	bool cxsr_enabled;
+
+	/* Calc sr entries for one plane configs */
+	crtc = single_enabled_crtc(dev_priv);
+	if (crtc) {
+		/* self-refresh has much higher latency */
+		static const int sr_latency_ns = 12000;
+		const struct drm_display_mode *pipe_mode =
+			&crtc->config->hw.pipe_mode;
+		const struct drm_framebuffer *fb =
+			crtc->base.primary->state->fb;
+		int pixel_rate = crtc->config->pixel_rate;
+		int htotal = pipe_mode->crtc_htotal;
+		int width = drm_rect_width(&crtc->base.primary->state->src) >> 16;
+		int cpp = fb->format->cpp[0];
+		int entries;
+
+		entries = intel_wm_method2(pixel_rate, htotal,
+					   width, cpp, sr_latency_ns / 100);
+		entries = DIV_ROUND_UP(entries, I915_FIFO_LINE_SIZE);
+		srwm = I965_FIFO_SIZE - entries;
+		if (srwm < 0)
+			srwm = 1;
+		srwm &= 0x1ff;
+		drm_dbg_kms(&dev_priv->drm,
+			    "self-refresh entries: %d, wm: %d\n",
+			    entries, srwm);
+
+		entries = intel_wm_method2(pixel_rate, htotal,
+					   crtc->base.cursor->state->crtc_w, 4,
+					   sr_latency_ns / 100);
+		entries = DIV_ROUND_UP(entries,
+				       i965_cursor_wm_info.cacheline_size) +
+			i965_cursor_wm_info.guard_size;
+
+		cursor_sr = i965_cursor_wm_info.fifo_size - entries;
+		if (cursor_sr > i965_cursor_wm_info.max_wm)
+			cursor_sr = i965_cursor_wm_info.max_wm;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "self-refresh watermark: display plane %d "
+			    "cursor %d\n", srwm, cursor_sr);
+
+		cxsr_enabled = true;
+	} else {
+		cxsr_enabled = false;
+		/* Turn off self refresh if both pipes are enabled */
+		intel_set_memory_cxsr(dev_priv, false);
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Setting FIFO watermarks - A: 8, B: 8, C: 8, SR %d\n",
+		    srwm);
+
+	/* 965 has limitations... */
+	intel_uncore_write(&dev_priv->uncore, DSPFW1, FW_WM(srwm, SR) |
+		   FW_WM(8, CURSORB) |
+		   FW_WM(8, PLANEB) |
+		   FW_WM(8, PLANEA));
+	intel_uncore_write(&dev_priv->uncore, DSPFW2, FW_WM(8, CURSORA) |
+		   FW_WM(8, PLANEC_OLD));
+	/* update cursor SR watermark */
+	intel_uncore_write(&dev_priv->uncore, DSPFW3, FW_WM(cursor_sr, CURSOR_SR));
+
+	if (cxsr_enabled)
+		intel_set_memory_cxsr(dev_priv, true);
+}
+
+#undef FW_WM
+
+static struct intel_crtc *intel_crtc_for_plane(struct drm_i915_private *i915,
+					       enum i9xx_plane_id i9xx_plane)
+{
+	struct intel_plane *plane;
+
+	for_each_intel_plane(&i915->drm, plane) {
+		if (plane->id == PLANE_PRIMARY &&
+		    plane->i9xx_plane == i9xx_plane)
+			return intel_crtc_for_pipe(i915, plane->pipe);
+	}
+
+	return NULL;
+}
+
+static void i9xx_update_wm(struct drm_i915_private *dev_priv)
+{
+	const struct intel_watermark_params *wm_info;
+	u32 fwater_lo;
+	u32 fwater_hi;
+	int cwm, srwm = 1;
+	int fifo_size;
+	int planea_wm, planeb_wm;
+	struct intel_crtc *crtc;
+
+	if (IS_I945GM(dev_priv))
+		wm_info = &i945_wm_info;
+	else if (DISPLAY_VER(dev_priv) != 2)
+		wm_info = &i915_wm_info;
+	else
+		wm_info = &i830_a_wm_info;
+
+	if (DISPLAY_VER(dev_priv) == 2)
+		fifo_size = i830_get_fifo_size(dev_priv, PLANE_A);
+	else
+		fifo_size = i9xx_get_fifo_size(dev_priv, PLANE_A);
+	crtc = intel_crtc_for_plane(dev_priv, PLANE_A);
+	if (intel_crtc_active(crtc)) {
+		const struct drm_framebuffer *fb =
+			crtc->base.primary->state->fb;
+		int cpp;
+
+		if (DISPLAY_VER(dev_priv) == 2)
+			cpp = 4;
+		else
+			cpp = fb->format->cpp[0];
+
+		planea_wm = intel_calculate_wm(crtc->config->pixel_rate,
+					       wm_info, fifo_size, cpp,
+					       pessimal_latency_ns);
+	} else {
+		planea_wm = fifo_size - wm_info->guard_size;
+		if (planea_wm > (long)wm_info->max_wm)
+			planea_wm = wm_info->max_wm;
+	}
+
+	if (DISPLAY_VER(dev_priv) == 2)
+		wm_info = &i830_bc_wm_info;
+
+	if (DISPLAY_VER(dev_priv) == 2)
+		fifo_size = i830_get_fifo_size(dev_priv, PLANE_B);
+	else
+		fifo_size = i9xx_get_fifo_size(dev_priv, PLANE_B);
+	crtc = intel_crtc_for_plane(dev_priv, PLANE_B);
+	if (intel_crtc_active(crtc)) {
+		const struct drm_framebuffer *fb =
+			crtc->base.primary->state->fb;
+		int cpp;
+
+		if (DISPLAY_VER(dev_priv) == 2)
+			cpp = 4;
+		else
+			cpp = fb->format->cpp[0];
+
+		planeb_wm = intel_calculate_wm(crtc->config->pixel_rate,
+					       wm_info, fifo_size, cpp,
+					       pessimal_latency_ns);
+	} else {
+		planeb_wm = fifo_size - wm_info->guard_size;
+		if (planeb_wm > (long)wm_info->max_wm)
+			planeb_wm = wm_info->max_wm;
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "FIFO watermarks - A: %d, B: %d\n", planea_wm, planeb_wm);
+
+	crtc = single_enabled_crtc(dev_priv);
+	if (IS_I915GM(dev_priv) && crtc) {
+		struct drm_i915_gem_object *obj;
+
+		obj = intel_fb_obj(crtc->base.primary->state->fb);
+
+		/* self-refresh seems busted with untiled */
+		if (!i915_gem_object_is_tiled(obj))
+			crtc = NULL;
+	}
+
+	/*
+	 * Overlay gets an aggressive default since video jitter is bad.
+	 */
+	cwm = 2;
+
+	/* Play safe and disable self-refresh before adjusting watermarks. */
+	intel_set_memory_cxsr(dev_priv, false);
+
+	/* Calc sr entries for one plane configs */
+	if (HAS_FW_BLC(dev_priv) && crtc) {
+		/* self-refresh has much higher latency */
+		static const int sr_latency_ns = 6000;
+		const struct drm_display_mode *pipe_mode =
+			&crtc->config->hw.pipe_mode;
+		const struct drm_framebuffer *fb =
+			crtc->base.primary->state->fb;
+		int pixel_rate = crtc->config->pixel_rate;
+		int htotal = pipe_mode->crtc_htotal;
+		int width = drm_rect_width(&crtc->base.primary->state->src) >> 16;
+		int cpp;
+		int entries;
+
+		if (IS_I915GM(dev_priv) || IS_I945GM(dev_priv))
+			cpp = 4;
+		else
+			cpp = fb->format->cpp[0];
+
+		entries = intel_wm_method2(pixel_rate, htotal, width, cpp,
+					   sr_latency_ns / 100);
+		entries = DIV_ROUND_UP(entries, wm_info->cacheline_size);
+		drm_dbg_kms(&dev_priv->drm,
+			    "self-refresh entries: %d\n", entries);
+		srwm = wm_info->fifo_size - entries;
+		if (srwm < 0)
+			srwm = 1;
+
+		if (IS_I945G(dev_priv) || IS_I945GM(dev_priv))
+			intel_uncore_write(&dev_priv->uncore, FW_BLC_SELF,
+				   FW_BLC_SELF_FIFO_MASK | (srwm & 0xff));
+		else
+			intel_uncore_write(&dev_priv->uncore, FW_BLC_SELF, srwm & 0x3f);
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Setting FIFO watermarks - A: %d, B: %d, C: %d, SR %d\n",
+		     planea_wm, planeb_wm, cwm, srwm);
+
+	fwater_lo = ((planeb_wm & 0x3f) << 16) | (planea_wm & 0x3f);
+	fwater_hi = (cwm & 0x1f);
+
+	/* Set request length to 8 cachelines per fetch */
+	fwater_lo = fwater_lo | (1 << 24) | (1 << 8);
+	fwater_hi = fwater_hi | (1 << 8);
+
+	intel_uncore_write(&dev_priv->uncore, FW_BLC, fwater_lo);
+	intel_uncore_write(&dev_priv->uncore, FW_BLC2, fwater_hi);
+
+	if (crtc)
+		intel_set_memory_cxsr(dev_priv, true);
+}
+
+static void i845_update_wm(struct drm_i915_private *dev_priv)
+{
+	struct intel_crtc *crtc;
+	u32 fwater_lo;
+	int planea_wm;
+
+	crtc = single_enabled_crtc(dev_priv);
+	if (crtc == NULL)
+		return;
+
+	planea_wm = intel_calculate_wm(crtc->config->pixel_rate,
+				       &i845_wm_info,
+				       i845_get_fifo_size(dev_priv, PLANE_A),
+				       4, pessimal_latency_ns);
+	fwater_lo = intel_uncore_read(&dev_priv->uncore, FW_BLC) & ~0xfff;
+	fwater_lo |= (3<<8) | planea_wm;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Setting FIFO watermarks - A: %d\n", planea_wm);
+
+	intel_uncore_write(&dev_priv->uncore, FW_BLC, fwater_lo);
+}
+
+/* latency must be in 0.1us units. */
+static unsigned int ilk_wm_method1(unsigned int pixel_rate,
+				   unsigned int cpp,
+				   unsigned int latency)
+{
+	unsigned int ret;
+
+	ret = intel_wm_method1(pixel_rate, cpp, latency);
+	ret = DIV_ROUND_UP(ret, 64) + 2;
+
+	return ret;
+}
+
+/* latency must be in 0.1us units. */
+static unsigned int ilk_wm_method2(unsigned int pixel_rate,
+				   unsigned int htotal,
+				   unsigned int width,
+				   unsigned int cpp,
+				   unsigned int latency)
+{
+	unsigned int ret;
+
+	ret = intel_wm_method2(pixel_rate, htotal,
+			       width, cpp, latency);
+	ret = DIV_ROUND_UP(ret, 64) + 2;
+
+	return ret;
+}
+
+static u32 ilk_wm_fbc(u32 pri_val, u32 horiz_pixels, u8 cpp)
+{
+	/*
+	 * Neither of these should be possible since this function shouldn't be
+	 * called if the CRTC is off or the plane is invisible.  But let's be
+	 * extra paranoid to avoid a potential divide-by-zero if we screw up
+	 * elsewhere in the driver.
+	 */
+	if (WARN_ON(!cpp))
+		return 0;
+	if (WARN_ON(!horiz_pixels))
+		return 0;
+
+	return DIV_ROUND_UP(pri_val * 64, horiz_pixels * cpp) + 2;
+}
+
+struct ilk_wm_maximums {
+	u16 pri;
+	u16 spr;
+	u16 cur;
+	u16 fbc;
+};
+
+/*
+ * For both WM_PIPE and WM_LP.
+ * mem_value must be in 0.1us units.
+ */
+static u32 ilk_compute_pri_wm(const struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state,
+			      u32 mem_value, bool is_lp)
+{
+	u32 method1, method2;
+	int cpp;
+
+	if (mem_value == 0)
+		return U32_MAX;
+
+	if (!intel_wm_plane_visible(crtc_state, plane_state))
+		return 0;
+
+	cpp = plane_state->hw.fb->format->cpp[0];
+
+	method1 = ilk_wm_method1(crtc_state->pixel_rate, cpp, mem_value);
+
+	if (!is_lp)
+		return method1;
+
+	method2 = ilk_wm_method2(crtc_state->pixel_rate,
+				 crtc_state->hw.pipe_mode.crtc_htotal,
+				 drm_rect_width(&plane_state->uapi.src) >> 16,
+				 cpp, mem_value);
+
+	return min(method1, method2);
+}
+
+/*
+ * For both WM_PIPE and WM_LP.
+ * mem_value must be in 0.1us units.
+ */
+static u32 ilk_compute_spr_wm(const struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state,
+			      u32 mem_value)
+{
+	u32 method1, method2;
+	int cpp;
+
+	if (mem_value == 0)
+		return U32_MAX;
+
+	if (!intel_wm_plane_visible(crtc_state, plane_state))
+		return 0;
+
+	cpp = plane_state->hw.fb->format->cpp[0];
+
+	method1 = ilk_wm_method1(crtc_state->pixel_rate, cpp, mem_value);
+	method2 = ilk_wm_method2(crtc_state->pixel_rate,
+				 crtc_state->hw.pipe_mode.crtc_htotal,
+				 drm_rect_width(&plane_state->uapi.src) >> 16,
+				 cpp, mem_value);
+	return min(method1, method2);
+}
+
+/*
+ * For both WM_PIPE and WM_LP.
+ * mem_value must be in 0.1us units.
+ */
+static u32 ilk_compute_cur_wm(const struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state,
+			      u32 mem_value)
+{
+	int cpp;
+
+	if (mem_value == 0)
+		return U32_MAX;
+
+	if (!intel_wm_plane_visible(crtc_state, plane_state))
+		return 0;
+
+	cpp = plane_state->hw.fb->format->cpp[0];
+
+	return ilk_wm_method2(crtc_state->pixel_rate,
+			      crtc_state->hw.pipe_mode.crtc_htotal,
+			      drm_rect_width(&plane_state->uapi.src) >> 16,
+			      cpp, mem_value);
+}
+
+/* Only for WM_LP. */
+static u32 ilk_compute_fbc_wm(const struct intel_crtc_state *crtc_state,
+			      const struct intel_plane_state *plane_state,
+			      u32 pri_val)
+{
+	int cpp;
+
+	if (!intel_wm_plane_visible(crtc_state, plane_state))
+		return 0;
+
+	cpp = plane_state->hw.fb->format->cpp[0];
+
+	return ilk_wm_fbc(pri_val, drm_rect_width(&plane_state->uapi.src) >> 16,
+			  cpp);
+}
+
+static unsigned int
+ilk_display_fifo_size(const struct drm_i915_private *dev_priv)
+{
+	if (DISPLAY_VER(dev_priv) >= 8)
+		return 3072;
+	else if (DISPLAY_VER(dev_priv) >= 7)
+		return 768;
+	else
+		return 512;
+}
+
+static unsigned int
+ilk_plane_wm_reg_max(const struct drm_i915_private *dev_priv,
+		     int level, bool is_sprite)
+{
+	if (DISPLAY_VER(dev_priv) >= 8)
+		/* BDW primary/sprite plane watermarks */
+		return level == 0 ? 255 : 2047;
+	else if (DISPLAY_VER(dev_priv) >= 7)
+		/* IVB/HSW primary/sprite plane watermarks */
+		return level == 0 ? 127 : 1023;
+	else if (!is_sprite)
+		/* ILK/SNB primary plane watermarks */
+		return level == 0 ? 127 : 511;
+	else
+		/* ILK/SNB sprite plane watermarks */
+		return level == 0 ? 63 : 255;
+}
+
+static unsigned int
+ilk_cursor_wm_reg_max(const struct drm_i915_private *dev_priv, int level)
+{
+	if (DISPLAY_VER(dev_priv) >= 7)
+		return level == 0 ? 63 : 255;
+	else
+		return level == 0 ? 31 : 63;
+}
+
+static unsigned int ilk_fbc_wm_reg_max(const struct drm_i915_private *dev_priv)
+{
+	if (DISPLAY_VER(dev_priv) >= 8)
+		return 31;
+	else
+		return 15;
+}
+
+/* Calculate the maximum primary/sprite plane watermark */
+static unsigned int ilk_plane_wm_max(const struct drm_i915_private *dev_priv,
+				     int level,
+				     const struct intel_wm_config *config,
+				     enum intel_ddb_partitioning ddb_partitioning,
+				     bool is_sprite)
+{
+	unsigned int fifo_size = ilk_display_fifo_size(dev_priv);
+
+	/* if sprites aren't enabled, sprites get nothing */
+	if (is_sprite && !config->sprites_enabled)
+		return 0;
+
+	/* HSW allows LP1+ watermarks even with multiple pipes */
+	if (level == 0 || config->num_pipes_active > 1) {
+		fifo_size /= INTEL_NUM_PIPES(dev_priv);
+
+		/*
+		 * For some reason the non self refresh
+		 * FIFO size is only half of the self
+		 * refresh FIFO size on ILK/SNB.
+		 */
+		if (DISPLAY_VER(dev_priv) <= 6)
+			fifo_size /= 2;
+	}
+
+	if (config->sprites_enabled) {
+		/* level 0 is always calculated with 1:1 split */
+		if (level > 0 && ddb_partitioning == INTEL_DDB_PART_5_6) {
+			if (is_sprite)
+				fifo_size *= 5;
+			fifo_size /= 6;
+		} else {
+			fifo_size /= 2;
+		}
+	}
+
+	/* clamp to max that the registers can hold */
+	return min(fifo_size, ilk_plane_wm_reg_max(dev_priv, level, is_sprite));
+}
+
+/* Calculate the maximum cursor plane watermark */
+static unsigned int ilk_cursor_wm_max(const struct drm_i915_private *dev_priv,
+				      int level,
+				      const struct intel_wm_config *config)
+{
+	/* HSW LP1+ watermarks w/ multiple pipes */
+	if (level > 0 && config->num_pipes_active > 1)
+		return 64;
+
+	/* otherwise just report max that registers can hold */
+	return ilk_cursor_wm_reg_max(dev_priv, level);
+}
+
+static void ilk_compute_wm_maximums(const struct drm_i915_private *dev_priv,
+				    int level,
+				    const struct intel_wm_config *config,
+				    enum intel_ddb_partitioning ddb_partitioning,
+				    struct ilk_wm_maximums *max)
+{
+	max->pri = ilk_plane_wm_max(dev_priv, level, config, ddb_partitioning, false);
+	max->spr = ilk_plane_wm_max(dev_priv, level, config, ddb_partitioning, true);
+	max->cur = ilk_cursor_wm_max(dev_priv, level, config);
+	max->fbc = ilk_fbc_wm_reg_max(dev_priv);
+}
+
+static void ilk_compute_wm_reg_maximums(const struct drm_i915_private *dev_priv,
+					int level,
+					struct ilk_wm_maximums *max)
+{
+	max->pri = ilk_plane_wm_reg_max(dev_priv, level, false);
+	max->spr = ilk_plane_wm_reg_max(dev_priv, level, true);
+	max->cur = ilk_cursor_wm_reg_max(dev_priv, level);
+	max->fbc = ilk_fbc_wm_reg_max(dev_priv);
+}
+
+static bool ilk_validate_wm_level(int level,
+				  const struct ilk_wm_maximums *max,
+				  struct intel_wm_level *result)
+{
+	bool ret;
+
+	/* already determined to be invalid? */
+	if (!result->enable)
+		return false;
+
+	result->enable = result->pri_val <= max->pri &&
+			 result->spr_val <= max->spr &&
+			 result->cur_val <= max->cur;
+
+	ret = result->enable;
+
+	/*
+	 * HACK until we can pre-compute everything,
+	 * and thus fail gracefully if LP0 watermarks
+	 * are exceeded...
+	 */
+	if (level == 0 && !result->enable) {
+		if (result->pri_val > max->pri)
+			DRM_DEBUG_KMS("Primary WM%d too large %u (max %u)\n",
+				      level, result->pri_val, max->pri);
+		if (result->spr_val > max->spr)
+			DRM_DEBUG_KMS("Sprite WM%d too large %u (max %u)\n",
+				      level, result->spr_val, max->spr);
+		if (result->cur_val > max->cur)
+			DRM_DEBUG_KMS("Cursor WM%d too large %u (max %u)\n",
+				      level, result->cur_val, max->cur);
+
+		result->pri_val = min_t(u32, result->pri_val, max->pri);
+		result->spr_val = min_t(u32, result->spr_val, max->spr);
+		result->cur_val = min_t(u32, result->cur_val, max->cur);
+		result->enable = true;
+	}
+
+	return ret;
+}
+
+static void ilk_compute_wm_level(const struct drm_i915_private *dev_priv,
+				 const struct intel_crtc *crtc,
+				 int level,
+				 struct intel_crtc_state *crtc_state,
+				 const struct intel_plane_state *pristate,
+				 const struct intel_plane_state *sprstate,
+				 const struct intel_plane_state *curstate,
+				 struct intel_wm_level *result)
+{
+	u16 pri_latency = dev_priv->display.wm.pri_latency[level];
+	u16 spr_latency = dev_priv->display.wm.spr_latency[level];
+	u16 cur_latency = dev_priv->display.wm.cur_latency[level];
+
+	/* WM1+ latency values stored in 0.5us units */
+	if (level > 0) {
+		pri_latency *= 5;
+		spr_latency *= 5;
+		cur_latency *= 5;
+	}
+
+	if (pristate) {
+		result->pri_val = ilk_compute_pri_wm(crtc_state, pristate,
+						     pri_latency, level);
+		result->fbc_val = ilk_compute_fbc_wm(crtc_state, pristate, result->pri_val);
+	}
+
+	if (sprstate)
+		result->spr_val = ilk_compute_spr_wm(crtc_state, sprstate, spr_latency);
+
+	if (curstate)
+		result->cur_val = ilk_compute_cur_wm(crtc_state, curstate, cur_latency);
+
+	result->enable = true;
+}
+
+static void hsw_read_wm_latency(struct drm_i915_private *i915, u16 wm[])
+{
+	u64 sskpd;
+
+	sskpd = intel_uncore_read64(&i915->uncore, MCH_SSKPD);
+
+	wm[0] = REG_FIELD_GET64(SSKPD_NEW_WM0_MASK_HSW, sskpd);
+	if (wm[0] == 0)
+		wm[0] = REG_FIELD_GET64(SSKPD_OLD_WM0_MASK_HSW, sskpd);
+	wm[1] = REG_FIELD_GET64(SSKPD_WM1_MASK_HSW, sskpd);
+	wm[2] = REG_FIELD_GET64(SSKPD_WM2_MASK_HSW, sskpd);
+	wm[3] = REG_FIELD_GET64(SSKPD_WM3_MASK_HSW, sskpd);
+	wm[4] = REG_FIELD_GET64(SSKPD_WM4_MASK_HSW, sskpd);
+}
+
+static void snb_read_wm_latency(struct drm_i915_private *i915, u16 wm[])
+{
+	u32 sskpd;
+
+	sskpd = intel_uncore_read(&i915->uncore, MCH_SSKPD);
+
+	wm[0] = REG_FIELD_GET(SSKPD_WM0_MASK_SNB, sskpd);
+	wm[1] = REG_FIELD_GET(SSKPD_WM1_MASK_SNB, sskpd);
+	wm[2] = REG_FIELD_GET(SSKPD_WM2_MASK_SNB, sskpd);
+	wm[3] = REG_FIELD_GET(SSKPD_WM3_MASK_SNB, sskpd);
+}
+
+static void ilk_read_wm_latency(struct drm_i915_private *i915, u16 wm[])
+{
+	u32 mltr;
+
+	mltr = intel_uncore_read(&i915->uncore, MLTR_ILK);
+
+	/* ILK primary LP0 latency is 700 ns */
+	wm[0] = 7;
+	wm[1] = REG_FIELD_GET(MLTR_WM1_MASK, mltr);
+	wm[2] = REG_FIELD_GET(MLTR_WM2_MASK, mltr);
+}
+
+static void intel_fixup_spr_wm_latency(struct drm_i915_private *dev_priv,
+				       u16 wm[5])
+{
+	/* ILK sprite LP0 latency is 1300 ns */
+	if (DISPLAY_VER(dev_priv) == 5)
+		wm[0] = 13;
+}
+
+static void intel_fixup_cur_wm_latency(struct drm_i915_private *dev_priv,
+				       u16 wm[5])
+{
+	/* ILK cursor LP0 latency is 1300 ns */
+	if (DISPLAY_VER(dev_priv) == 5)
+		wm[0] = 13;
+}
+
+int ilk_wm_max_level(const struct drm_i915_private *dev_priv)
+{
+	/* how many WM levels are we expecting */
+	if (HAS_HW_SAGV_WM(dev_priv))
+		return 5;
+	else if (DISPLAY_VER(dev_priv) >= 9)
+		return 7;
+	else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+		return 4;
+	else if (DISPLAY_VER(dev_priv) >= 6)
+		return 3;
+	else
+		return 2;
+}
+
+void intel_print_wm_latency(struct drm_i915_private *dev_priv,
+			    const char *name, const u16 wm[])
+{
+	int level, max_level = ilk_wm_max_level(dev_priv);
+
+	for (level = 0; level <= max_level; level++) {
+		unsigned int latency = wm[level];
+
+		if (latency == 0) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "%s WM%d latency not provided\n",
+				    name, level);
+			continue;
+		}
+
+		/*
+		 * - latencies are in us on gen9.
+		 * - before then, WM1+ latency values are in 0.5us units
+		 */
+		if (DISPLAY_VER(dev_priv) >= 9)
+			latency *= 10;
+		else if (level > 0)
+			latency *= 5;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "%s WM%d latency %u (%u.%u usec)\n", name, level,
+			    wm[level], latency / 10, latency % 10);
+	}
+}
+
+static bool ilk_increase_wm_latency(struct drm_i915_private *dev_priv,
+				    u16 wm[5], u16 min)
+{
+	int level, max_level = ilk_wm_max_level(dev_priv);
+
+	if (wm[0] >= min)
+		return false;
+
+	wm[0] = max(wm[0], min);
+	for (level = 1; level <= max_level; level++)
+		wm[level] = max_t(u16, wm[level], DIV_ROUND_UP(min, 5));
+
+	return true;
+}
+
+static void snb_wm_latency_quirk(struct drm_i915_private *dev_priv)
+{
+	bool changed;
+
+	/*
+	 * The BIOS provided WM memory latency values are often
+	 * inadequate for high resolution displays. Adjust them.
+	 */
+	changed = ilk_increase_wm_latency(dev_priv, dev_priv->display.wm.pri_latency, 12);
+	changed |= ilk_increase_wm_latency(dev_priv, dev_priv->display.wm.spr_latency, 12);
+	changed |= ilk_increase_wm_latency(dev_priv, dev_priv->display.wm.cur_latency, 12);
+
+	if (!changed)
+		return;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "WM latency values increased to avoid potential underruns\n");
+	intel_print_wm_latency(dev_priv, "Primary", dev_priv->display.wm.pri_latency);
+	intel_print_wm_latency(dev_priv, "Sprite", dev_priv->display.wm.spr_latency);
+	intel_print_wm_latency(dev_priv, "Cursor", dev_priv->display.wm.cur_latency);
+}
+
+static void snb_wm_lp3_irq_quirk(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * On some SNB machines (Thinkpad X220 Tablet at least)
+	 * LP3 usage can cause vblank interrupts to be lost.
+	 * The DEIIR bit will go high but it looks like the CPU
+	 * never gets interrupted.
+	 *
+	 * It's not clear whether other interrupt source could
+	 * be affected or if this is somehow limited to vblank
+	 * interrupts only. To play it safe we disable LP3
+	 * watermarks entirely.
+	 */
+	if (dev_priv->display.wm.pri_latency[3] == 0 &&
+	    dev_priv->display.wm.spr_latency[3] == 0 &&
+	    dev_priv->display.wm.cur_latency[3] == 0)
+		return;
+
+	dev_priv->display.wm.pri_latency[3] = 0;
+	dev_priv->display.wm.spr_latency[3] = 0;
+	dev_priv->display.wm.cur_latency[3] = 0;
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "LP3 watermarks disabled due to potential for lost interrupts\n");
+	intel_print_wm_latency(dev_priv, "Primary", dev_priv->display.wm.pri_latency);
+	intel_print_wm_latency(dev_priv, "Sprite", dev_priv->display.wm.spr_latency);
+	intel_print_wm_latency(dev_priv, "Cursor", dev_priv->display.wm.cur_latency);
+}
+
+static void ilk_setup_wm_latency(struct drm_i915_private *dev_priv)
+{
+	if (IS_BROADWELL(dev_priv) || IS_HASWELL(dev_priv))
+		hsw_read_wm_latency(dev_priv, dev_priv->display.wm.pri_latency);
+	else if (DISPLAY_VER(dev_priv) >= 6)
+		snb_read_wm_latency(dev_priv, dev_priv->display.wm.pri_latency);
+	else
+		ilk_read_wm_latency(dev_priv, dev_priv->display.wm.pri_latency);
+
+	memcpy(dev_priv->display.wm.spr_latency, dev_priv->display.wm.pri_latency,
+	       sizeof(dev_priv->display.wm.pri_latency));
+	memcpy(dev_priv->display.wm.cur_latency, dev_priv->display.wm.pri_latency,
+	       sizeof(dev_priv->display.wm.pri_latency));
+
+	intel_fixup_spr_wm_latency(dev_priv, dev_priv->display.wm.spr_latency);
+	intel_fixup_cur_wm_latency(dev_priv, dev_priv->display.wm.cur_latency);
+
+	intel_print_wm_latency(dev_priv, "Primary", dev_priv->display.wm.pri_latency);
+	intel_print_wm_latency(dev_priv, "Sprite", dev_priv->display.wm.spr_latency);
+	intel_print_wm_latency(dev_priv, "Cursor", dev_priv->display.wm.cur_latency);
+
+	if (DISPLAY_VER(dev_priv) == 6) {
+		snb_wm_latency_quirk(dev_priv);
+		snb_wm_lp3_irq_quirk(dev_priv);
+	}
+}
+
+static bool ilk_validate_pipe_wm(const struct drm_i915_private *dev_priv,
+				 struct intel_pipe_wm *pipe_wm)
+{
+	/* LP0 watermark maximums depend on this pipe alone */
+	const struct intel_wm_config config = {
+		.num_pipes_active = 1,
+		.sprites_enabled = pipe_wm->sprites_enabled,
+		.sprites_scaled = pipe_wm->sprites_scaled,
+	};
+	struct ilk_wm_maximums max;
+
+	/* LP0 watermarks always use 1/2 DDB partitioning */
+	ilk_compute_wm_maximums(dev_priv, 0, &config, INTEL_DDB_PART_1_2, &max);
+
+	/* At least LP0 must be valid */
+	if (!ilk_validate_wm_level(0, &max, &pipe_wm->wm[0])) {
+		drm_dbg_kms(&dev_priv->drm, "LP0 watermark invalid\n");
+		return false;
+	}
+
+	return true;
+}
+
+/* Compute new watermarks for the pipe */
+static int ilk_compute_pipe_wm(struct intel_atomic_state *state,
+			       struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
+	struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	struct intel_pipe_wm *pipe_wm;
+	struct intel_plane *plane;
+	const struct intel_plane_state *plane_state;
+	const struct intel_plane_state *pristate = NULL;
+	const struct intel_plane_state *sprstate = NULL;
+	const struct intel_plane_state *curstate = NULL;
+	int level, max_level = ilk_wm_max_level(dev_priv), usable_level;
+	struct ilk_wm_maximums max;
+
+	pipe_wm = &crtc_state->wm.ilk.optimal;
+
+	intel_atomic_crtc_state_for_each_plane_state(plane, plane_state, crtc_state) {
+		if (plane->base.type == DRM_PLANE_TYPE_PRIMARY)
+			pristate = plane_state;
+		else if (plane->base.type == DRM_PLANE_TYPE_OVERLAY)
+			sprstate = plane_state;
+		else if (plane->base.type == DRM_PLANE_TYPE_CURSOR)
+			curstate = plane_state;
+	}
+
+	pipe_wm->pipe_enabled = crtc_state->hw.active;
+	pipe_wm->sprites_enabled = crtc_state->active_planes & BIT(PLANE_SPRITE0);
+	pipe_wm->sprites_scaled = crtc_state->scaled_planes & BIT(PLANE_SPRITE0);
+
+	usable_level = max_level;
+
+	/* ILK/SNB: LP2+ watermarks only w/o sprites */
+	if (DISPLAY_VER(dev_priv) <= 6 && pipe_wm->sprites_enabled)
+		usable_level = 1;
+
+	/* ILK/SNB/IVB: LP1+ watermarks only w/o scaling */
+	if (pipe_wm->sprites_scaled)
+		usable_level = 0;
+
+	memset(&pipe_wm->wm, 0, sizeof(pipe_wm->wm));
+	ilk_compute_wm_level(dev_priv, crtc, 0, crtc_state,
+			     pristate, sprstate, curstate, &pipe_wm->wm[0]);
+
+	if (!ilk_validate_pipe_wm(dev_priv, pipe_wm))
+		return -EINVAL;
+
+	ilk_compute_wm_reg_maximums(dev_priv, 1, &max);
+
+	for (level = 1; level <= usable_level; level++) {
+		struct intel_wm_level *wm = &pipe_wm->wm[level];
+
+		ilk_compute_wm_level(dev_priv, crtc, level, crtc_state,
+				     pristate, sprstate, curstate, wm);
+
+		/*
+		 * Disable any watermark level that exceeds the
+		 * register maximums since such watermarks are
+		 * always invalid.
+		 */
+		if (!ilk_validate_wm_level(level, &max, wm)) {
+			memset(wm, 0, sizeof(*wm));
+			break;
+		}
+	}
+
+	return 0;
+}
+
+/*
+ * Build a set of 'intermediate' watermark values that satisfy both the old
+ * state and the new state.  These can be programmed to the hardware
+ * immediately.
+ */
+static int ilk_compute_intermediate_wm(struct intel_atomic_state *state,
+				       struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_crtc_state *new_crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+	const struct intel_crtc_state *old_crtc_state =
+		intel_atomic_get_old_crtc_state(state, crtc);
+	struct intel_pipe_wm *a = &new_crtc_state->wm.ilk.intermediate;
+	const struct intel_pipe_wm *b = &old_crtc_state->wm.ilk.optimal;
+	int level, max_level = ilk_wm_max_level(dev_priv);
+
+	/*
+	 * Start with the final, target watermarks, then combine with the
+	 * currently active watermarks to get values that are safe both before
+	 * and after the vblank.
+	 */
+	*a = new_crtc_state->wm.ilk.optimal;
+	if (!new_crtc_state->hw.active ||
+	    drm_atomic_crtc_needs_modeset(&new_crtc_state->uapi) ||
+	    state->skip_intermediate_wm)
+		return 0;
+
+	a->pipe_enabled |= b->pipe_enabled;
+	a->sprites_enabled |= b->sprites_enabled;
+	a->sprites_scaled |= b->sprites_scaled;
+
+	for (level = 0; level <= max_level; level++) {
+		struct intel_wm_level *a_wm = &a->wm[level];
+		const struct intel_wm_level *b_wm = &b->wm[level];
+
+		a_wm->enable &= b_wm->enable;
+		a_wm->pri_val = max(a_wm->pri_val, b_wm->pri_val);
+		a_wm->spr_val = max(a_wm->spr_val, b_wm->spr_val);
+		a_wm->cur_val = max(a_wm->cur_val, b_wm->cur_val);
+		a_wm->fbc_val = max(a_wm->fbc_val, b_wm->fbc_val);
+	}
+
+	/*
+	 * We need to make sure that these merged watermark values are
+	 * actually a valid configuration themselves.  If they're not,
+	 * there's no safe way to transition from the old state to
+	 * the new state, so we need to fail the atomic transaction.
+	 */
+	if (!ilk_validate_pipe_wm(dev_priv, a))
+		return -EINVAL;
+
+	/*
+	 * If our intermediate WM are identical to the final WM, then we can
+	 * omit the post-vblank programming; only update if it's different.
+	 */
+	if (memcmp(a, &new_crtc_state->wm.ilk.optimal, sizeof(*a)) != 0)
+		new_crtc_state->wm.need_postvbl_update = true;
+
+	return 0;
+}
+
+/*
+ * Merge the watermarks from all active pipes for a specific level.
+ */
+static void ilk_merge_wm_level(struct drm_i915_private *dev_priv,
+			       int level,
+			       struct intel_wm_level *ret_wm)
+{
+	const struct intel_crtc *crtc;
+
+	ret_wm->enable = true;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		const struct intel_pipe_wm *active = &crtc->wm.active.ilk;
+		const struct intel_wm_level *wm = &active->wm[level];
+
+		if (!active->pipe_enabled)
+			continue;
+
+		/*
+		 * The watermark values may have been used in the past,
+		 * so we must maintain them in the registers for some
+		 * time even if the level is now disabled.
+		 */
+		if (!wm->enable)
+			ret_wm->enable = false;
+
+		ret_wm->pri_val = max(ret_wm->pri_val, wm->pri_val);
+		ret_wm->spr_val = max(ret_wm->spr_val, wm->spr_val);
+		ret_wm->cur_val = max(ret_wm->cur_val, wm->cur_val);
+		ret_wm->fbc_val = max(ret_wm->fbc_val, wm->fbc_val);
+	}
+}
+
+/*
+ * Merge all low power watermarks for all active pipes.
+ */
+static void ilk_wm_merge(struct drm_i915_private *dev_priv,
+			 const struct intel_wm_config *config,
+			 const struct ilk_wm_maximums *max,
+			 struct intel_pipe_wm *merged)
+{
+	int level, max_level = ilk_wm_max_level(dev_priv);
+	int last_enabled_level = max_level;
+
+	/* ILK/SNB/IVB: LP1+ watermarks only w/ single pipe */
+	if ((DISPLAY_VER(dev_priv) <= 6 || IS_IVYBRIDGE(dev_priv)) &&
+	    config->num_pipes_active > 1)
+		last_enabled_level = 0;
+
+	/* ILK: FBC WM must be disabled always */
+	merged->fbc_wm_enabled = DISPLAY_VER(dev_priv) >= 6;
+
+	/* merge each WM1+ level */
+	for (level = 1; level <= max_level; level++) {
+		struct intel_wm_level *wm = &merged->wm[level];
+
+		ilk_merge_wm_level(dev_priv, level, wm);
+
+		if (level > last_enabled_level)
+			wm->enable = false;
+		else if (!ilk_validate_wm_level(level, max, wm))
+			/* make sure all following levels get disabled */
+			last_enabled_level = level - 1;
+
+		/*
+		 * The spec says it is preferred to disable
+		 * FBC WMs instead of disabling a WM level.
+		 */
+		if (wm->fbc_val > max->fbc) {
+			if (wm->enable)
+				merged->fbc_wm_enabled = false;
+			wm->fbc_val = 0;
+		}
+	}
+
+	/* ILK: LP2+ must be disabled when FBC WM is disabled but FBC enabled */
+	if (DISPLAY_VER(dev_priv) == 5 && HAS_FBC(dev_priv) &&
+	    dev_priv->params.enable_fbc && !merged->fbc_wm_enabled) {
+		for (level = 2; level <= max_level; level++) {
+			struct intel_wm_level *wm = &merged->wm[level];
+
+			wm->enable = false;
+		}
+	}
+}
+
+static int ilk_wm_lp_to_level(int wm_lp, const struct intel_pipe_wm *pipe_wm)
+{
+	/* LP1,LP2,LP3 levels are either 1,2,3 or 1,3,4 */
+	return wm_lp + (wm_lp >= 2 && pipe_wm->wm[4].enable);
+}
+
+/* The value we need to program into the WM_LPx latency field */
+static unsigned int ilk_wm_lp_latency(struct drm_i915_private *dev_priv,
+				      int level)
+{
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+		return 2 * level;
+	else
+		return dev_priv->display.wm.pri_latency[level];
+}
+
+static void ilk_compute_wm_results(struct drm_i915_private *dev_priv,
+				   const struct intel_pipe_wm *merged,
+				   enum intel_ddb_partitioning partitioning,
+				   struct ilk_wm_values *results)
+{
+	struct intel_crtc *crtc;
+	int level, wm_lp;
+
+	results->enable_fbc_wm = merged->fbc_wm_enabled;
+	results->partitioning = partitioning;
+
+	/* LP1+ register values */
+	for (wm_lp = 1; wm_lp <= 3; wm_lp++) {
+		const struct intel_wm_level *r;
+
+		level = ilk_wm_lp_to_level(wm_lp, merged);
+
+		r = &merged->wm[level];
+
+		/*
+		 * Maintain the watermark values even if the level is
+		 * disabled. Doing otherwise could cause underruns.
+		 */
+		results->wm_lp[wm_lp - 1] =
+			WM_LP_LATENCY(ilk_wm_lp_latency(dev_priv, level)) |
+			WM_LP_PRIMARY(r->pri_val) |
+			WM_LP_CURSOR(r->cur_val);
+
+		if (r->enable)
+			results->wm_lp[wm_lp - 1] |= WM_LP_ENABLE;
+
+		if (DISPLAY_VER(dev_priv) >= 8)
+			results->wm_lp[wm_lp - 1] |= WM_LP_FBC_BDW(r->fbc_val);
+		else
+			results->wm_lp[wm_lp - 1] |= WM_LP_FBC_ILK(r->fbc_val);
+
+		results->wm_lp_spr[wm_lp - 1] = WM_LP_SPRITE(r->spr_val);
+
+		/*
+		 * Always set WM_LP_SPRITE_EN when spr_val != 0, even if the
+		 * level is disabled. Doing otherwise could cause underruns.
+		 */
+		if (DISPLAY_VER(dev_priv) <= 6 && r->spr_val) {
+			drm_WARN_ON(&dev_priv->drm, wm_lp != 1);
+			results->wm_lp_spr[wm_lp - 1] |= WM_LP_SPRITE_ENABLE;
+		}
+	}
+
+	/* LP0 register values */
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		enum pipe pipe = crtc->pipe;
+		const struct intel_pipe_wm *pipe_wm = &crtc->wm.active.ilk;
+		const struct intel_wm_level *r = &pipe_wm->wm[0];
+
+		if (drm_WARN_ON(&dev_priv->drm, !r->enable))
+			continue;
+
+		results->wm_pipe[pipe] =
+			WM0_PIPE_PRIMARY(r->pri_val) |
+			WM0_PIPE_SPRITE(r->spr_val) |
+			WM0_PIPE_CURSOR(r->cur_val);
+	}
+}
+
+/* Find the result with the highest level enabled. Check for enable_fbc_wm in
+ * case both are at the same level. Prefer r1 in case they're the same. */
+static struct intel_pipe_wm *
+ilk_find_best_result(struct drm_i915_private *dev_priv,
+		     struct intel_pipe_wm *r1,
+		     struct intel_pipe_wm *r2)
+{
+	int level, max_level = ilk_wm_max_level(dev_priv);
+	int level1 = 0, level2 = 0;
+
+	for (level = 1; level <= max_level; level++) {
+		if (r1->wm[level].enable)
+			level1 = level;
+		if (r2->wm[level].enable)
+			level2 = level;
+	}
+
+	if (level1 == level2) {
+		if (r2->fbc_wm_enabled && !r1->fbc_wm_enabled)
+			return r2;
+		else
+			return r1;
+	} else if (level1 > level2) {
+		return r1;
+	} else {
+		return r2;
+	}
+}
+
+/* dirty bits used to track which watermarks need changes */
+#define WM_DIRTY_PIPE(pipe) (1 << (pipe))
+#define WM_DIRTY_LP(wm_lp) (1 << (15 + (wm_lp)))
+#define WM_DIRTY_LP_ALL (WM_DIRTY_LP(1) | WM_DIRTY_LP(2) | WM_DIRTY_LP(3))
+#define WM_DIRTY_FBC (1 << 24)
+#define WM_DIRTY_DDB (1 << 25)
+
+static unsigned int ilk_compute_wm_dirty(struct drm_i915_private *dev_priv,
+					 const struct ilk_wm_values *old,
+					 const struct ilk_wm_values *new)
+{
+	unsigned int dirty = 0;
+	enum pipe pipe;
+	int wm_lp;
+
+	for_each_pipe(dev_priv, pipe) {
+		if (old->wm_pipe[pipe] != new->wm_pipe[pipe]) {
+			dirty |= WM_DIRTY_PIPE(pipe);
+			/* Must disable LP1+ watermarks too */
+			dirty |= WM_DIRTY_LP_ALL;
+		}
+	}
+
+	if (old->enable_fbc_wm != new->enable_fbc_wm) {
+		dirty |= WM_DIRTY_FBC;
+		/* Must disable LP1+ watermarks too */
+		dirty |= WM_DIRTY_LP_ALL;
+	}
+
+	if (old->partitioning != new->partitioning) {
+		dirty |= WM_DIRTY_DDB;
+		/* Must disable LP1+ watermarks too */
+		dirty |= WM_DIRTY_LP_ALL;
+	}
+
+	/* LP1+ watermarks already deemed dirty, no need to continue */
+	if (dirty & WM_DIRTY_LP_ALL)
+		return dirty;
+
+	/* Find the lowest numbered LP1+ watermark in need of an update... */
+	for (wm_lp = 1; wm_lp <= 3; wm_lp++) {
+		if (old->wm_lp[wm_lp - 1] != new->wm_lp[wm_lp - 1] ||
+		    old->wm_lp_spr[wm_lp - 1] != new->wm_lp_spr[wm_lp - 1])
+			break;
+	}
+
+	/* ...and mark it and all higher numbered LP1+ watermarks as dirty */
+	for (; wm_lp <= 3; wm_lp++)
+		dirty |= WM_DIRTY_LP(wm_lp);
+
+	return dirty;
+}
+
+static bool _ilk_disable_lp_wm(struct drm_i915_private *dev_priv,
+			       unsigned int dirty)
+{
+	struct ilk_wm_values *previous = &dev_priv->display.wm.hw;
+	bool changed = false;
+
+	if (dirty & WM_DIRTY_LP(3) && previous->wm_lp[2] & WM_LP_ENABLE) {
+		previous->wm_lp[2] &= ~WM_LP_ENABLE;
+		intel_uncore_write(&dev_priv->uncore, WM3_LP_ILK, previous->wm_lp[2]);
+		changed = true;
+	}
+	if (dirty & WM_DIRTY_LP(2) && previous->wm_lp[1] & WM_LP_ENABLE) {
+		previous->wm_lp[1] &= ~WM_LP_ENABLE;
+		intel_uncore_write(&dev_priv->uncore, WM2_LP_ILK, previous->wm_lp[1]);
+		changed = true;
+	}
+	if (dirty & WM_DIRTY_LP(1) && previous->wm_lp[0] & WM_LP_ENABLE) {
+		previous->wm_lp[0] &= ~WM_LP_ENABLE;
+		intel_uncore_write(&dev_priv->uncore, WM1_LP_ILK, previous->wm_lp[0]);
+		changed = true;
+	}
+
+	/*
+	 * Don't touch WM_LP_SPRITE_ENABLE here.
+	 * Doing so could cause underruns.
+	 */
+
+	return changed;
+}
+
+/*
+ * The spec says we shouldn't write when we don't need, because every write
+ * causes WMs to be re-evaluated, expending some power.
+ */
+static void ilk_write_wm_values(struct drm_i915_private *dev_priv,
+				struct ilk_wm_values *results)
+{
+	struct ilk_wm_values *previous = &dev_priv->display.wm.hw;
+	unsigned int dirty;
+	u32 val;
+
+	dirty = ilk_compute_wm_dirty(dev_priv, previous, results);
+	if (!dirty)
+		return;
+
+	_ilk_disable_lp_wm(dev_priv, dirty);
+
+	if (dirty & WM_DIRTY_PIPE(PIPE_A))
+		intel_uncore_write(&dev_priv->uncore, WM0_PIPE_ILK(PIPE_A), results->wm_pipe[0]);
+	if (dirty & WM_DIRTY_PIPE(PIPE_B))
+		intel_uncore_write(&dev_priv->uncore, WM0_PIPE_ILK(PIPE_B), results->wm_pipe[1]);
+	if (dirty & WM_DIRTY_PIPE(PIPE_C))
+		intel_uncore_write(&dev_priv->uncore, WM0_PIPE_ILK(PIPE_C), results->wm_pipe[2]);
+
+	if (dirty & WM_DIRTY_DDB) {
+		if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
+			val = intel_uncore_read(&dev_priv->uncore, WM_MISC);
+			if (results->partitioning == INTEL_DDB_PART_1_2)
+				val &= ~WM_MISC_DATA_PARTITION_5_6;
+			else
+				val |= WM_MISC_DATA_PARTITION_5_6;
+			intel_uncore_write(&dev_priv->uncore, WM_MISC, val);
+		} else {
+			val = intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL2);
+			if (results->partitioning == INTEL_DDB_PART_1_2)
+				val &= ~DISP_DATA_PARTITION_5_6;
+			else
+				val |= DISP_DATA_PARTITION_5_6;
+			intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL2, val);
+		}
+	}
+
+	if (dirty & WM_DIRTY_FBC) {
+		val = intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL);
+		if (results->enable_fbc_wm)
+			val &= ~DISP_FBC_WM_DIS;
+		else
+			val |= DISP_FBC_WM_DIS;
+		intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, val);
+	}
+
+	if (dirty & WM_DIRTY_LP(1) &&
+	    previous->wm_lp_spr[0] != results->wm_lp_spr[0])
+		intel_uncore_write(&dev_priv->uncore, WM1S_LP_ILK, results->wm_lp_spr[0]);
+
+	if (DISPLAY_VER(dev_priv) >= 7) {
+		if (dirty & WM_DIRTY_LP(2) && previous->wm_lp_spr[1] != results->wm_lp_spr[1])
+			intel_uncore_write(&dev_priv->uncore, WM2S_LP_IVB, results->wm_lp_spr[1]);
+		if (dirty & WM_DIRTY_LP(3) && previous->wm_lp_spr[2] != results->wm_lp_spr[2])
+			intel_uncore_write(&dev_priv->uncore, WM3S_LP_IVB, results->wm_lp_spr[2]);
+	}
+
+	if (dirty & WM_DIRTY_LP(1) && previous->wm_lp[0] != results->wm_lp[0])
+		intel_uncore_write(&dev_priv->uncore, WM1_LP_ILK, results->wm_lp[0]);
+	if (dirty & WM_DIRTY_LP(2) && previous->wm_lp[1] != results->wm_lp[1])
+		intel_uncore_write(&dev_priv->uncore, WM2_LP_ILK, results->wm_lp[1]);
+	if (dirty & WM_DIRTY_LP(3) && previous->wm_lp[2] != results->wm_lp[2])
+		intel_uncore_write(&dev_priv->uncore, WM3_LP_ILK, results->wm_lp[2]);
+
+	dev_priv->display.wm.hw = *results;
+}
+
+bool ilk_disable_lp_wm(struct drm_i915_private *dev_priv)
+{
+	return _ilk_disable_lp_wm(dev_priv, WM_DIRTY_LP_ALL);
+}
+
+static void ilk_compute_wm_config(struct drm_i915_private *dev_priv,
+				  struct intel_wm_config *config)
+{
+	struct intel_crtc *crtc;
+
+	/* Compute the currently _active_ config */
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		const struct intel_pipe_wm *wm = &crtc->wm.active.ilk;
+
+		if (!wm->pipe_enabled)
+			continue;
+
+		config->sprites_enabled |= wm->sprites_enabled;
+		config->sprites_scaled |= wm->sprites_scaled;
+		config->num_pipes_active++;
+	}
+}
+
+static void ilk_program_watermarks(struct drm_i915_private *dev_priv)
+{
+	struct intel_pipe_wm lp_wm_1_2 = {}, lp_wm_5_6 = {}, *best_lp_wm;
+	struct ilk_wm_maximums max;
+	struct intel_wm_config config = {};
+	struct ilk_wm_values results = {};
+	enum intel_ddb_partitioning partitioning;
+
+	ilk_compute_wm_config(dev_priv, &config);
+
+	ilk_compute_wm_maximums(dev_priv, 1, &config, INTEL_DDB_PART_1_2, &max);
+	ilk_wm_merge(dev_priv, &config, &max, &lp_wm_1_2);
+
+	/* 5/6 split only in single pipe config on IVB+ */
+	if (DISPLAY_VER(dev_priv) >= 7 &&
+	    config.num_pipes_active == 1 && config.sprites_enabled) {
+		ilk_compute_wm_maximums(dev_priv, 1, &config, INTEL_DDB_PART_5_6, &max);
+		ilk_wm_merge(dev_priv, &config, &max, &lp_wm_5_6);
+
+		best_lp_wm = ilk_find_best_result(dev_priv, &lp_wm_1_2, &lp_wm_5_6);
+	} else {
+		best_lp_wm = &lp_wm_1_2;
+	}
+
+	partitioning = (best_lp_wm == &lp_wm_1_2) ?
+		       INTEL_DDB_PART_1_2 : INTEL_DDB_PART_5_6;
+
+	ilk_compute_wm_results(dev_priv, best_lp_wm, partitioning, &results);
+
+	ilk_write_wm_values(dev_priv, &results);
+}
+
+static void ilk_initial_watermarks(struct intel_atomic_state *state,
+				   struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	mutex_lock(&dev_priv->display.wm.wm_mutex);
+	crtc->wm.active.ilk = crtc_state->wm.ilk.intermediate;
+	ilk_program_watermarks(dev_priv);
+	mutex_unlock(&dev_priv->display.wm.wm_mutex);
+}
+
+static void ilk_optimize_watermarks(struct intel_atomic_state *state,
+				    struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct intel_crtc_state *crtc_state =
+		intel_atomic_get_new_crtc_state(state, crtc);
+
+	if (!crtc_state->wm.need_postvbl_update)
+		return;
+
+	mutex_lock(&dev_priv->display.wm.wm_mutex);
+	crtc->wm.active.ilk = crtc_state->wm.ilk.optimal;
+	ilk_program_watermarks(dev_priv);
+	mutex_unlock(&dev_priv->display.wm.wm_mutex);
+}
+
+static void ilk_pipe_wm_get_hw_state(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct ilk_wm_values *hw = &dev_priv->display.wm.hw;
+	struct intel_crtc_state *crtc_state = to_intel_crtc_state(crtc->base.state);
+	struct intel_pipe_wm *active = &crtc_state->wm.ilk.optimal;
+	enum pipe pipe = crtc->pipe;
+
+	hw->wm_pipe[pipe] = intel_uncore_read(&dev_priv->uncore, WM0_PIPE_ILK(pipe));
+
+	memset(active, 0, sizeof(*active));
+
+	active->pipe_enabled = crtc->active;
+
+	if (active->pipe_enabled) {
+		u32 tmp = hw->wm_pipe[pipe];
+
+		/*
+		 * For active pipes LP0 watermark is marked as
+		 * enabled, and LP1+ watermaks as disabled since
+		 * we can't really reverse compute them in case
+		 * multiple pipes are active.
+		 */
+		active->wm[0].enable = true;
+		active->wm[0].pri_val = REG_FIELD_GET(WM0_PIPE_PRIMARY_MASK, tmp);
+		active->wm[0].spr_val = REG_FIELD_GET(WM0_PIPE_SPRITE_MASK, tmp);
+		active->wm[0].cur_val = REG_FIELD_GET(WM0_PIPE_CURSOR_MASK, tmp);
+	} else {
+		int level, max_level = ilk_wm_max_level(dev_priv);
+
+		/*
+		 * For inactive pipes, all watermark levels
+		 * should be marked as enabled but zeroed,
+		 * which is what we'd compute them to.
+		 */
+		for (level = 0; level <= max_level; level++)
+			active->wm[level].enable = true;
+	}
+
+	crtc->wm.active.ilk = *active;
+}
+
+#define _FW_WM(value, plane) \
+	(((value) & DSPFW_ ## plane ## _MASK) >> DSPFW_ ## plane ## _SHIFT)
+#define _FW_WM_VLV(value, plane) \
+	(((value) & DSPFW_ ## plane ## _MASK_VLV) >> DSPFW_ ## plane ## _SHIFT)
+
+static void g4x_read_wm_values(struct drm_i915_private *dev_priv,
+			       struct g4x_wm_values *wm)
+{
+	u32 tmp;
+
+	tmp = intel_uncore_read(&dev_priv->uncore, DSPFW1);
+	wm->sr.plane = _FW_WM(tmp, SR);
+	wm->pipe[PIPE_B].plane[PLANE_CURSOR] = _FW_WM(tmp, CURSORB);
+	wm->pipe[PIPE_B].plane[PLANE_PRIMARY] = _FW_WM(tmp, PLANEB);
+	wm->pipe[PIPE_A].plane[PLANE_PRIMARY] = _FW_WM(tmp, PLANEA);
+
+	tmp = intel_uncore_read(&dev_priv->uncore, DSPFW2);
+	wm->fbc_en = tmp & DSPFW_FBC_SR_EN;
+	wm->sr.fbc = _FW_WM(tmp, FBC_SR);
+	wm->hpll.fbc = _FW_WM(tmp, FBC_HPLL_SR);
+	wm->pipe[PIPE_B].plane[PLANE_SPRITE0] = _FW_WM(tmp, SPRITEB);
+	wm->pipe[PIPE_A].plane[PLANE_CURSOR] = _FW_WM(tmp, CURSORA);
+	wm->pipe[PIPE_A].plane[PLANE_SPRITE0] = _FW_WM(tmp, SPRITEA);
+
+	tmp = intel_uncore_read(&dev_priv->uncore, DSPFW3);
+	wm->hpll_en = tmp & DSPFW_HPLL_SR_EN;
+	wm->sr.cursor = _FW_WM(tmp, CURSOR_SR);
+	wm->hpll.cursor = _FW_WM(tmp, HPLL_CURSOR);
+	wm->hpll.plane = _FW_WM(tmp, HPLL_SR);
+}
+
+static void vlv_read_wm_values(struct drm_i915_private *dev_priv,
+			       struct vlv_wm_values *wm)
+{
+	enum pipe pipe;
+	u32 tmp;
+
+	for_each_pipe(dev_priv, pipe) {
+		tmp = intel_uncore_read(&dev_priv->uncore, VLV_DDL(pipe));
+
+		wm->ddl[pipe].plane[PLANE_PRIMARY] =
+			(tmp >> DDL_PLANE_SHIFT) & (DDL_PRECISION_HIGH | DRAIN_LATENCY_MASK);
+		wm->ddl[pipe].plane[PLANE_CURSOR] =
+			(tmp >> DDL_CURSOR_SHIFT) & (DDL_PRECISION_HIGH | DRAIN_LATENCY_MASK);
+		wm->ddl[pipe].plane[PLANE_SPRITE0] =
+			(tmp >> DDL_SPRITE_SHIFT(0)) & (DDL_PRECISION_HIGH | DRAIN_LATENCY_MASK);
+		wm->ddl[pipe].plane[PLANE_SPRITE1] =
+			(tmp >> DDL_SPRITE_SHIFT(1)) & (DDL_PRECISION_HIGH | DRAIN_LATENCY_MASK);
+	}
+
+	tmp = intel_uncore_read(&dev_priv->uncore, DSPFW1);
+	wm->sr.plane = _FW_WM(tmp, SR);
+	wm->pipe[PIPE_B].plane[PLANE_CURSOR] = _FW_WM(tmp, CURSORB);
+	wm->pipe[PIPE_B].plane[PLANE_PRIMARY] = _FW_WM_VLV(tmp, PLANEB);
+	wm->pipe[PIPE_A].plane[PLANE_PRIMARY] = _FW_WM_VLV(tmp, PLANEA);
+
+	tmp = intel_uncore_read(&dev_priv->uncore, DSPFW2);
+	wm->pipe[PIPE_A].plane[PLANE_SPRITE1] = _FW_WM_VLV(tmp, SPRITEB);
+	wm->pipe[PIPE_A].plane[PLANE_CURSOR] = _FW_WM(tmp, CURSORA);
+	wm->pipe[PIPE_A].plane[PLANE_SPRITE0] = _FW_WM_VLV(tmp, SPRITEA);
+
+	tmp = intel_uncore_read(&dev_priv->uncore, DSPFW3);
+	wm->sr.cursor = _FW_WM(tmp, CURSOR_SR);
+
+	if (IS_CHERRYVIEW(dev_priv)) {
+		tmp = intel_uncore_read(&dev_priv->uncore, DSPFW7_CHV);
+		wm->pipe[PIPE_B].plane[PLANE_SPRITE1] = _FW_WM_VLV(tmp, SPRITED);
+		wm->pipe[PIPE_B].plane[PLANE_SPRITE0] = _FW_WM_VLV(tmp, SPRITEC);
+
+		tmp = intel_uncore_read(&dev_priv->uncore, DSPFW8_CHV);
+		wm->pipe[PIPE_C].plane[PLANE_SPRITE1] = _FW_WM_VLV(tmp, SPRITEF);
+		wm->pipe[PIPE_C].plane[PLANE_SPRITE0] = _FW_WM_VLV(tmp, SPRITEE);
+
+		tmp = intel_uncore_read(&dev_priv->uncore, DSPFW9_CHV);
+		wm->pipe[PIPE_C].plane[PLANE_PRIMARY] = _FW_WM_VLV(tmp, PLANEC);
+		wm->pipe[PIPE_C].plane[PLANE_CURSOR] = _FW_WM(tmp, CURSORC);
+
+		tmp = intel_uncore_read(&dev_priv->uncore, DSPHOWM);
+		wm->sr.plane |= _FW_WM(tmp, SR_HI) << 9;
+		wm->pipe[PIPE_C].plane[PLANE_SPRITE1] |= _FW_WM(tmp, SPRITEF_HI) << 8;
+		wm->pipe[PIPE_C].plane[PLANE_SPRITE0] |= _FW_WM(tmp, SPRITEE_HI) << 8;
+		wm->pipe[PIPE_C].plane[PLANE_PRIMARY] |= _FW_WM(tmp, PLANEC_HI) << 8;
+		wm->pipe[PIPE_B].plane[PLANE_SPRITE1] |= _FW_WM(tmp, SPRITED_HI) << 8;
+		wm->pipe[PIPE_B].plane[PLANE_SPRITE0] |= _FW_WM(tmp, SPRITEC_HI) << 8;
+		wm->pipe[PIPE_B].plane[PLANE_PRIMARY] |= _FW_WM(tmp, PLANEB_HI) << 8;
+		wm->pipe[PIPE_A].plane[PLANE_SPRITE1] |= _FW_WM(tmp, SPRITEB_HI) << 8;
+		wm->pipe[PIPE_A].plane[PLANE_SPRITE0] |= _FW_WM(tmp, SPRITEA_HI) << 8;
+		wm->pipe[PIPE_A].plane[PLANE_PRIMARY] |= _FW_WM(tmp, PLANEA_HI) << 8;
+	} else {
+		tmp = intel_uncore_read(&dev_priv->uncore, DSPFW7);
+		wm->pipe[PIPE_B].plane[PLANE_SPRITE1] = _FW_WM_VLV(tmp, SPRITED);
+		wm->pipe[PIPE_B].plane[PLANE_SPRITE0] = _FW_WM_VLV(tmp, SPRITEC);
+
+		tmp = intel_uncore_read(&dev_priv->uncore, DSPHOWM);
+		wm->sr.plane |= _FW_WM(tmp, SR_HI) << 9;
+		wm->pipe[PIPE_B].plane[PLANE_SPRITE1] |= _FW_WM(tmp, SPRITED_HI) << 8;
+		wm->pipe[PIPE_B].plane[PLANE_SPRITE0] |= _FW_WM(tmp, SPRITEC_HI) << 8;
+		wm->pipe[PIPE_B].plane[PLANE_PRIMARY] |= _FW_WM(tmp, PLANEB_HI) << 8;
+		wm->pipe[PIPE_A].plane[PLANE_SPRITE1] |= _FW_WM(tmp, SPRITEB_HI) << 8;
+		wm->pipe[PIPE_A].plane[PLANE_SPRITE0] |= _FW_WM(tmp, SPRITEA_HI) << 8;
+		wm->pipe[PIPE_A].plane[PLANE_PRIMARY] |= _FW_WM(tmp, PLANEA_HI) << 8;
+	}
+}
+
+#undef _FW_WM
+#undef _FW_WM_VLV
+
+void g4x_wm_get_hw_state(struct drm_i915_private *dev_priv)
+{
+	struct g4x_wm_values *wm = &dev_priv->display.wm.g4x;
+	struct intel_crtc *crtc;
+
+	g4x_read_wm_values(dev_priv, wm);
+
+	wm->cxsr = intel_uncore_read(&dev_priv->uncore, FW_BLC_SELF) & FW_BLC_SELF_EN;
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct g4x_wm_state *active = &crtc->wm.active.g4x;
+		struct g4x_pipe_wm *raw;
+		enum pipe pipe = crtc->pipe;
+		enum plane_id plane_id;
+		int level, max_level;
+
+		active->cxsr = wm->cxsr;
+		active->hpll_en = wm->hpll_en;
+		active->fbc_en = wm->fbc_en;
+
+		active->sr = wm->sr;
+		active->hpll = wm->hpll;
+
+		for_each_plane_id_on_crtc(crtc, plane_id) {
+			active->wm.plane[plane_id] =
+				wm->pipe[pipe].plane[plane_id];
+		}
+
+		if (wm->cxsr && wm->hpll_en)
+			max_level = G4X_WM_LEVEL_HPLL;
+		else if (wm->cxsr)
+			max_level = G4X_WM_LEVEL_SR;
+		else
+			max_level = G4X_WM_LEVEL_NORMAL;
+
+		level = G4X_WM_LEVEL_NORMAL;
+		raw = &crtc_state->wm.g4x.raw[level];
+		for_each_plane_id_on_crtc(crtc, plane_id)
+			raw->plane[plane_id] = active->wm.plane[plane_id];
+
+		level = G4X_WM_LEVEL_SR;
+		if (level > max_level)
+			goto out;
+
+		raw = &crtc_state->wm.g4x.raw[level];
+		raw->plane[PLANE_PRIMARY] = active->sr.plane;
+		raw->plane[PLANE_CURSOR] = active->sr.cursor;
+		raw->plane[PLANE_SPRITE0] = 0;
+		raw->fbc = active->sr.fbc;
+
+		level = G4X_WM_LEVEL_HPLL;
+		if (level > max_level)
+			goto out;
+
+		raw = &crtc_state->wm.g4x.raw[level];
+		raw->plane[PLANE_PRIMARY] = active->hpll.plane;
+		raw->plane[PLANE_CURSOR] = active->hpll.cursor;
+		raw->plane[PLANE_SPRITE0] = 0;
+		raw->fbc = active->hpll.fbc;
+
+		level++;
+	out:
+		for_each_plane_id_on_crtc(crtc, plane_id)
+			g4x_raw_plane_wm_set(crtc_state, level,
+					     plane_id, USHRT_MAX);
+		g4x_raw_fbc_wm_set(crtc_state, level, USHRT_MAX);
+
+		crtc_state->wm.g4x.optimal = *active;
+		crtc_state->wm.g4x.intermediate = *active;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "Initial watermarks: pipe %c, plane=%d, cursor=%d, sprite=%d\n",
+			    pipe_name(pipe),
+			    wm->pipe[pipe].plane[PLANE_PRIMARY],
+			    wm->pipe[pipe].plane[PLANE_CURSOR],
+			    wm->pipe[pipe].plane[PLANE_SPRITE0]);
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Initial SR watermarks: plane=%d, cursor=%d fbc=%d\n",
+		    wm->sr.plane, wm->sr.cursor, wm->sr.fbc);
+	drm_dbg_kms(&dev_priv->drm,
+		    "Initial HPLL watermarks: plane=%d, SR cursor=%d fbc=%d\n",
+		    wm->hpll.plane, wm->hpll.cursor, wm->hpll.fbc);
+	drm_dbg_kms(&dev_priv->drm, "Initial SR=%s HPLL=%s FBC=%s\n",
+		    str_yes_no(wm->cxsr), str_yes_no(wm->hpll_en),
+		    str_yes_no(wm->fbc_en));
+}
+
+void g4x_wm_sanitize(struct drm_i915_private *dev_priv)
+{
+	struct intel_plane *plane;
+	struct intel_crtc *crtc;
+
+	mutex_lock(&dev_priv->display.wm.wm_mutex);
+
+	for_each_intel_plane(&dev_priv->drm, plane) {
+		struct intel_crtc *crtc =
+			intel_crtc_for_pipe(dev_priv, plane->pipe);
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct intel_plane_state *plane_state =
+			to_intel_plane_state(plane->base.state);
+		struct g4x_wm_state *wm_state = &crtc_state->wm.g4x.optimal;
+		enum plane_id plane_id = plane->id;
+		int level;
+
+		if (plane_state->uapi.visible)
+			continue;
+
+		for (level = 0; level < 3; level++) {
+			struct g4x_pipe_wm *raw =
+				&crtc_state->wm.g4x.raw[level];
+
+			raw->plane[plane_id] = 0;
+			wm_state->wm.plane[plane_id] = 0;
+		}
+
+		if (plane_id == PLANE_PRIMARY) {
+			for (level = 0; level < 3; level++) {
+				struct g4x_pipe_wm *raw =
+					&crtc_state->wm.g4x.raw[level];
+				raw->fbc = 0;
+			}
+
+			wm_state->sr.fbc = 0;
+			wm_state->hpll.fbc = 0;
+			wm_state->fbc_en = false;
+		}
+	}
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		crtc_state->wm.g4x.intermediate =
+			crtc_state->wm.g4x.optimal;
+		crtc->wm.active.g4x = crtc_state->wm.g4x.optimal;
+	}
+
+	g4x_program_watermarks(dev_priv);
+
+	mutex_unlock(&dev_priv->display.wm.wm_mutex);
+}
+
+void vlv_wm_get_hw_state(struct drm_i915_private *dev_priv)
+{
+	struct vlv_wm_values *wm = &dev_priv->display.wm.vlv;
+	struct intel_crtc *crtc;
+	u32 val;
+
+	vlv_read_wm_values(dev_priv, wm);
+
+	wm->cxsr = intel_uncore_read(&dev_priv->uncore, FW_BLC_SELF_VLV) & FW_CSPWRDWNEN;
+	wm->level = VLV_WM_LEVEL_PM2;
+
+	if (IS_CHERRYVIEW(dev_priv)) {
+		vlv_punit_get(dev_priv);
+
+		val = vlv_punit_read(dev_priv, PUNIT_REG_DSPSSPM);
+		if (val & DSP_MAXFIFO_PM5_ENABLE)
+			wm->level = VLV_WM_LEVEL_PM5;
+
+		/*
+		 * If DDR DVFS is disabled in the BIOS, Punit
+		 * will never ack the request. So if that happens
+		 * assume we don't have to enable/disable DDR DVFS
+		 * dynamically. To test that just set the REQ_ACK
+		 * bit to poke the Punit, but don't change the
+		 * HIGH/LOW bits so that we don't actually change
+		 * the current state.
+		 */
+		val = vlv_punit_read(dev_priv, PUNIT_REG_DDR_SETUP2);
+		val |= FORCE_DDR_FREQ_REQ_ACK;
+		vlv_punit_write(dev_priv, PUNIT_REG_DDR_SETUP2, val);
+
+		if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DDR_SETUP2) &
+			      FORCE_DDR_FREQ_REQ_ACK) == 0, 3)) {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Punit not acking DDR DVFS request, "
+				    "assuming DDR DVFS is disabled\n");
+			dev_priv->display.wm.max_level = VLV_WM_LEVEL_PM5;
+		} else {
+			val = vlv_punit_read(dev_priv, PUNIT_REG_DDR_SETUP2);
+			if ((val & FORCE_DDR_HIGH_FREQ) == 0)
+				wm->level = VLV_WM_LEVEL_DDR_DVFS;
+		}
+
+		vlv_punit_put(dev_priv);
+	}
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct vlv_wm_state *active = &crtc->wm.active.vlv;
+		const struct vlv_fifo_state *fifo_state =
+			&crtc_state->wm.vlv.fifo_state;
+		enum pipe pipe = crtc->pipe;
+		enum plane_id plane_id;
+		int level;
+
+		vlv_get_fifo_size(crtc_state);
+
+		active->num_levels = wm->level + 1;
+		active->cxsr = wm->cxsr;
+
+		for (level = 0; level < active->num_levels; level++) {
+			struct g4x_pipe_wm *raw =
+				&crtc_state->wm.vlv.raw[level];
+
+			active->sr[level].plane = wm->sr.plane;
+			active->sr[level].cursor = wm->sr.cursor;
+
+			for_each_plane_id_on_crtc(crtc, plane_id) {
+				active->wm[level].plane[plane_id] =
+					wm->pipe[pipe].plane[plane_id];
+
+				raw->plane[plane_id] =
+					vlv_invert_wm_value(active->wm[level].plane[plane_id],
+							    fifo_state->plane[plane_id]);
+			}
+		}
+
+		for_each_plane_id_on_crtc(crtc, plane_id)
+			vlv_raw_plane_wm_set(crtc_state, level,
+					     plane_id, USHRT_MAX);
+		vlv_invalidate_wms(crtc, active, level);
+
+		crtc_state->wm.vlv.optimal = *active;
+		crtc_state->wm.vlv.intermediate = *active;
+
+		drm_dbg_kms(&dev_priv->drm,
+			    "Initial watermarks: pipe %c, plane=%d, cursor=%d, sprite0=%d, sprite1=%d\n",
+			    pipe_name(pipe),
+			    wm->pipe[pipe].plane[PLANE_PRIMARY],
+			    wm->pipe[pipe].plane[PLANE_CURSOR],
+			    wm->pipe[pipe].plane[PLANE_SPRITE0],
+			    wm->pipe[pipe].plane[PLANE_SPRITE1]);
+	}
+
+	drm_dbg_kms(&dev_priv->drm,
+		    "Initial watermarks: SR plane=%d, SR cursor=%d level=%d cxsr=%d\n",
+		    wm->sr.plane, wm->sr.cursor, wm->level, wm->cxsr);
+}
+
+void vlv_wm_sanitize(struct drm_i915_private *dev_priv)
+{
+	struct intel_plane *plane;
+	struct intel_crtc *crtc;
+
+	mutex_lock(&dev_priv->display.wm.wm_mutex);
+
+	for_each_intel_plane(&dev_priv->drm, plane) {
+		struct intel_crtc *crtc =
+			intel_crtc_for_pipe(dev_priv, plane->pipe);
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct intel_plane_state *plane_state =
+			to_intel_plane_state(plane->base.state);
+		struct vlv_wm_state *wm_state = &crtc_state->wm.vlv.optimal;
+		const struct vlv_fifo_state *fifo_state =
+			&crtc_state->wm.vlv.fifo_state;
+		enum plane_id plane_id = plane->id;
+		int level;
+
+		if (plane_state->uapi.visible)
+			continue;
+
+		for (level = 0; level < wm_state->num_levels; level++) {
+			struct g4x_pipe_wm *raw =
+				&crtc_state->wm.vlv.raw[level];
+
+			raw->plane[plane_id] = 0;
+
+			wm_state->wm[level].plane[plane_id] =
+				vlv_invert_wm_value(raw->plane[plane_id],
+						    fifo_state->plane[plane_id]);
+		}
+	}
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		crtc_state->wm.vlv.intermediate =
+			crtc_state->wm.vlv.optimal;
+		crtc->wm.active.vlv = crtc_state->wm.vlv.optimal;
+	}
+
+	vlv_program_watermarks(dev_priv);
+
+	mutex_unlock(&dev_priv->display.wm.wm_mutex);
+}
+
+/*
+ * FIXME should probably kill this and improve
+ * the real watermark readout/sanitation instead
+ */
+static void ilk_init_lp_watermarks(struct drm_i915_private *dev_priv)
+{
+	intel_uncore_write(&dev_priv->uncore, WM3_LP_ILK, intel_uncore_read(&dev_priv->uncore, WM3_LP_ILK) & ~WM_LP_ENABLE);
+	intel_uncore_write(&dev_priv->uncore, WM2_LP_ILK, intel_uncore_read(&dev_priv->uncore, WM2_LP_ILK) & ~WM_LP_ENABLE);
+	intel_uncore_write(&dev_priv->uncore, WM1_LP_ILK, intel_uncore_read(&dev_priv->uncore, WM1_LP_ILK) & ~WM_LP_ENABLE);
+
+	/*
+	 * Don't touch WM_LP_SPRITE_ENABLE here.
+	 * Doing so could cause underruns.
+	 */
+}
+
+void ilk_wm_get_hw_state(struct drm_i915_private *dev_priv)
+{
+	struct ilk_wm_values *hw = &dev_priv->display.wm.hw;
+	struct intel_crtc *crtc;
+
+	ilk_init_lp_watermarks(dev_priv);
+
+	for_each_intel_crtc(&dev_priv->drm, crtc)
+		ilk_pipe_wm_get_hw_state(crtc);
+
+	hw->wm_lp[0] = intel_uncore_read(&dev_priv->uncore, WM1_LP_ILK);
+	hw->wm_lp[1] = intel_uncore_read(&dev_priv->uncore, WM2_LP_ILK);
+	hw->wm_lp[2] = intel_uncore_read(&dev_priv->uncore, WM3_LP_ILK);
+
+	hw->wm_lp_spr[0] = intel_uncore_read(&dev_priv->uncore, WM1S_LP_ILK);
+	if (DISPLAY_VER(dev_priv) >= 7) {
+		hw->wm_lp_spr[1] = intel_uncore_read(&dev_priv->uncore, WM2S_LP_IVB);
+		hw->wm_lp_spr[2] = intel_uncore_read(&dev_priv->uncore, WM3S_LP_IVB);
+	}
+
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+		hw->partitioning = (intel_uncore_read(&dev_priv->uncore, WM_MISC) & WM_MISC_DATA_PARTITION_5_6) ?
+			INTEL_DDB_PART_5_6 : INTEL_DDB_PART_1_2;
+	else if (IS_IVYBRIDGE(dev_priv))
+		hw->partitioning = (intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL2) & DISP_DATA_PARTITION_5_6) ?
+			INTEL_DDB_PART_5_6 : INTEL_DDB_PART_1_2;
+
+	hw->enable_fbc_wm =
+		!(intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) & DISP_FBC_WM_DIS);
+}
+
+static void ibx_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * On Ibex Peak and Cougar Point, we need to disable clock
+	 * gating for the panel power sequencer or it will fail to
+	 * start up when no ports are active.
+	 */
+	intel_uncore_write(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D, PCH_DPLSUNIT_CLOCK_GATE_DISABLE);
+}
+
+static void g4x_disable_trickle_feed(struct drm_i915_private *dev_priv)
+{
+	enum pipe pipe;
+
+	for_each_pipe(dev_priv, pipe) {
+		intel_uncore_write(&dev_priv->uncore, DSPCNTR(pipe),
+			   intel_uncore_read(&dev_priv->uncore, DSPCNTR(pipe)) |
+			   DISP_TRICKLE_FEED_DISABLE);
+
+		intel_uncore_write(&dev_priv->uncore, DSPSURF(pipe), intel_uncore_read(&dev_priv->uncore, DSPSURF(pipe)));
+		intel_uncore_posting_read(&dev_priv->uncore, DSPSURF(pipe));
+	}
+}
+
+static void ilk_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	u32 dspclk_gate = ILK_VRHUNIT_CLOCK_GATE_DISABLE;
+
+	/*
+	 * Required for FBC
+	 * WaFbcDisableDpfcClockGating:ilk
+	 */
+	dspclk_gate |= ILK_DPFCRUNIT_CLOCK_GATE_DISABLE |
+		   ILK_DPFCUNIT_CLOCK_GATE_DISABLE |
+		   ILK_DPFDUNIT_CLOCK_GATE_ENABLE;
+
+	intel_uncore_write(&dev_priv->uncore, PCH_3DCGDIS0,
+		   MARIUNIT_CLOCK_GATE_DISABLE |
+		   SVSMUNIT_CLOCK_GATE_DISABLE);
+	intel_uncore_write(&dev_priv->uncore, PCH_3DCGDIS1,
+		   VFMUNIT_CLOCK_GATE_DISABLE);
+
+	/*
+	 * According to the spec the following bits should be set in
+	 * order to enable memory self-refresh
+	 * The bit 22/21 of 0x42004
+	 * The bit 5 of 0x42020
+	 * The bit 15 of 0x45000
+	 */
+	intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2,
+		   (intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2) |
+		    ILK_DPARB_GATE | ILK_VSDPFD_FULL));
+	dspclk_gate |= ILK_DPARBUNIT_CLOCK_GATE_ENABLE;
+	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL,
+		   (intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
+		    DISP_FBC_WM_DIS));
+
+	/*
+	 * Based on the document from hardware guys the following bits
+	 * should be set unconditionally in order to enable FBC.
+	 * The bit 22 of 0x42000
+	 * The bit 22 of 0x42004
+	 * The bit 7,8,9 of 0x42020.
+	 */
+	if (IS_IRONLAKE_M(dev_priv)) {
+		/* WaFbcAsynchFlipDisableFbcQueue:ilk */
+		intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1,
+			   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1) |
+			   ILK_FBCQ_DIS);
+		intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2,
+			   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2) |
+			   ILK_DPARB_GATE);
+	}
+
+	intel_uncore_write(&dev_priv->uncore, ILK_DSPCLK_GATE_D, dspclk_gate);
+
+	intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2,
+		   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2) |
+		   ILK_ELPIN_409_SELECT);
+
+	g4x_disable_trickle_feed(dev_priv);
+
+	ibx_init_clock_gating(dev_priv);
+}
+
+static void cpt_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	enum pipe pipe;
+	u32 val;
+
+	/*
+	 * On Ibex Peak and Cougar Point, we need to disable clock
+	 * gating for the panel power sequencer or it will fail to
+	 * start up when no ports are active.
+	 */
+	intel_uncore_write(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D, PCH_DPLSUNIT_CLOCK_GATE_DISABLE |
+		   PCH_DPLUNIT_CLOCK_GATE_DISABLE |
+		   PCH_CPUNIT_CLOCK_GATE_DISABLE);
+	intel_uncore_write(&dev_priv->uncore, SOUTH_CHICKEN2, intel_uncore_read(&dev_priv->uncore, SOUTH_CHICKEN2) |
+		   DPLS_EDP_PPS_FIX_DIS);
+	/* The below fixes the weird display corruption, a few pixels shifted
+	 * downward, on (only) LVDS of some HP laptops with IVY.
+	 */
+	for_each_pipe(dev_priv, pipe) {
+		val = intel_uncore_read(&dev_priv->uncore, TRANS_CHICKEN2(pipe));
+		val |= TRANS_CHICKEN2_TIMING_OVERRIDE;
+		val &= ~TRANS_CHICKEN2_FDI_POLARITY_REVERSED;
+		if (dev_priv->display.vbt.fdi_rx_polarity_inverted)
+			val |= TRANS_CHICKEN2_FDI_POLARITY_REVERSED;
+		val &= ~TRANS_CHICKEN2_DISABLE_DEEP_COLOR_COUNTER;
+		val &= ~TRANS_CHICKEN2_DISABLE_DEEP_COLOR_MODESWITCH;
+		intel_uncore_write(&dev_priv->uncore, TRANS_CHICKEN2(pipe), val);
+	}
+	/* WADP0ClockGatingDisable */
+	for_each_pipe(dev_priv, pipe) {
+		intel_uncore_write(&dev_priv->uncore, TRANS_CHICKEN1(pipe),
+			   TRANS_CHICKEN1_DP0UNIT_GC_DISABLE);
+	}
+}
+
+static void gen6_check_mch_setup(struct drm_i915_private *dev_priv)
+{
+	u32 tmp;
+
+	tmp = intel_uncore_read(&dev_priv->uncore, MCH_SSKPD);
+	if (REG_FIELD_GET(SSKPD_WM0_MASK_SNB, tmp) != 12)
+		drm_dbg_kms(&dev_priv->drm,
+			    "Wrong MCH_SSKPD value: 0x%08x This can cause underruns.\n",
+			    tmp);
+}
+
+static void gen6_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	u32 dspclk_gate = ILK_VRHUNIT_CLOCK_GATE_DISABLE;
+
+	intel_uncore_write(&dev_priv->uncore, ILK_DSPCLK_GATE_D, dspclk_gate);
+
+	intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2,
+		   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2) |
+		   ILK_ELPIN_409_SELECT);
+
+	intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL1,
+		   intel_uncore_read(&dev_priv->uncore, GEN6_UCGCTL1) |
+		   GEN6_BLBUNIT_CLOCK_GATE_DISABLE |
+		   GEN6_CSUNIT_CLOCK_GATE_DISABLE);
+
+	/* According to the BSpec vol1g, bit 12 (RCPBUNIT) clock
+	 * gating disable must be set.  Failure to set it results in
+	 * flickering pixels due to Z write ordering failures after
+	 * some amount of runtime in the Mesa "fire" demo, and Unigine
+	 * Sanctuary and Tropics, and apparently anything else with
+	 * alpha test or pixel discard.
+	 *
+	 * According to the spec, bit 11 (RCCUNIT) must also be set,
+	 * but we didn't debug actual testcases to find it out.
+	 *
+	 * WaDisableRCCUnitClockGating:snb
+	 * WaDisableRCPBUnitClockGating:snb
+	 */
+	intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL2,
+		   GEN6_RCPBUNIT_CLOCK_GATE_DISABLE |
+		   GEN6_RCCUNIT_CLOCK_GATE_DISABLE);
+
+	/*
+	 * According to the spec the following bits should be
+	 * set in order to enable memory self-refresh and fbc:
+	 * The bit21 and bit22 of 0x42000
+	 * The bit21 and bit22 of 0x42004
+	 * The bit5 and bit7 of 0x42020
+	 * The bit14 of 0x70180
+	 * The bit14 of 0x71180
+	 *
+	 * WaFbcAsynchFlipDisableFbcQueue:snb
+	 */
+	intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1,
+		   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1) |
+		   ILK_FBCQ_DIS | ILK_PABSTRETCH_DIS);
+	intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2,
+		   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN2) |
+		   ILK_DPARB_GATE | ILK_VSDPFD_FULL);
+	intel_uncore_write(&dev_priv->uncore, ILK_DSPCLK_GATE_D,
+		   intel_uncore_read(&dev_priv->uncore, ILK_DSPCLK_GATE_D) |
+		   ILK_DPARBUNIT_CLOCK_GATE_ENABLE  |
+		   ILK_DPFDUNIT_CLOCK_GATE_ENABLE);
+
+	g4x_disable_trickle_feed(dev_priv);
+
+	cpt_init_clock_gating(dev_priv);
+
+	gen6_check_mch_setup(dev_priv);
+}
+
+static void lpt_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * TODO: this bit should only be enabled when really needed, then
+	 * disabled when not needed anymore in order to save power.
+	 */
+	if (HAS_PCH_LPT_LP(dev_priv))
+		intel_uncore_write(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D,
+			   intel_uncore_read(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D) |
+			   PCH_LP_PARTITION_LEVEL_DISABLE);
+
+	/* WADPOClockGatingDisable:hsw */
+	intel_uncore_write(&dev_priv->uncore, TRANS_CHICKEN1(PIPE_A),
+		   intel_uncore_read(&dev_priv->uncore, TRANS_CHICKEN1(PIPE_A)) |
+		   TRANS_CHICKEN1_DP0UNIT_GC_DISABLE);
+}
+
+static void lpt_suspend_hw(struct drm_i915_private *dev_priv)
+{
+	if (HAS_PCH_LPT_LP(dev_priv)) {
+		u32 val = intel_uncore_read(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D);
+
+		val &= ~PCH_LP_PARTITION_LEVEL_DISABLE;
+		intel_uncore_write(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D, val);
+	}
+}
+
+static void gen8_set_l3sqc_credits(struct drm_i915_private *dev_priv,
+				   int general_prio_credits,
+				   int high_prio_credits)
+{
+	u32 misccpctl;
+	u32 val;
+
+	/* WaTempDisableDOPClkGating:bdw */
+	misccpctl = intel_uncore_read(&dev_priv->uncore, GEN7_MISCCPCTL);
+	intel_uncore_write(&dev_priv->uncore, GEN7_MISCCPCTL, misccpctl & ~GEN7_DOP_CLOCK_GATE_ENABLE);
+
+	val = intel_uncore_read(&dev_priv->uncore, GEN8_L3SQCREG1);
+	val &= ~L3_PRIO_CREDITS_MASK;
+	val |= L3_GENERAL_PRIO_CREDITS(general_prio_credits);
+	val |= L3_HIGH_PRIO_CREDITS(high_prio_credits);
+	intel_uncore_write(&dev_priv->uncore, GEN8_L3SQCREG1, val);
+
+	/*
+	 * Wait at least 100 clocks before re-enabling clock gating.
+	 * See the definition of L3SQCREG1 in BSpec.
+	 */
+	intel_uncore_posting_read(&dev_priv->uncore, GEN8_L3SQCREG1);
+	udelay(1);
+	intel_uncore_write(&dev_priv->uncore, GEN7_MISCCPCTL, misccpctl);
+}
+
+static void icl_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	/* Wa_1409120013:icl,ehl */
+	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
+			   DPFC_CHICKEN_COMP_DUMMY_PIXEL);
+
+	/*Wa_14010594013:icl, ehl */
+	intel_uncore_rmw(&dev_priv->uncore, GEN8_CHICKEN_DCPR_1,
+			 0, ICL_DELAY_PMRSP);
+}
+
+static void gen12lp_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	/* Wa_1409120013 */
+	if (DISPLAY_VER(dev_priv) == 12)
+		intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
+				   DPFC_CHICKEN_COMP_DUMMY_PIXEL);
+
+	/* Wa_1409825376:tgl (pre-prod)*/
+	if (IS_TGL_DISPLAY_STEP(dev_priv, STEP_A0, STEP_C0))
+		intel_uncore_write(&dev_priv->uncore, GEN9_CLKGATE_DIS_3, intel_uncore_read(&dev_priv->uncore, GEN9_CLKGATE_DIS_3) |
+			   TGL_VRH_GATING_DIS);
+
+	/* Wa_14013723622:tgl,rkl,dg1,adl-s */
+	if (DISPLAY_VER(dev_priv) == 12)
+		intel_uncore_rmw(&dev_priv->uncore, CLKREQ_POLICY,
+				 CLKREQ_POLICY_MEM_UP_OVRD, 0);
+}
+
+static void adlp_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	gen12lp_init_clock_gating(dev_priv);
+
+	/* Wa_22011091694:adlp */
+	intel_de_rmw(dev_priv, GEN9_CLKGATE_DIS_5, 0, DPCE_GATING_DIS);
+
+	/* Bspec/49189 Initialize Sequence */
+	intel_de_rmw(dev_priv, GEN8_CHICKEN_DCPR_1, DDI_CLOCK_REG_ACCESS, 0);
+}
+
+static void dg1_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	gen12lp_init_clock_gating(dev_priv);
+
+	/* Wa_1409836686:dg1[a0] */
+	if (IS_DG1_GRAPHICS_STEP(dev_priv, STEP_A0, STEP_B0))
+		intel_uncore_write(&dev_priv->uncore, GEN9_CLKGATE_DIS_3, intel_uncore_read(&dev_priv->uncore, GEN9_CLKGATE_DIS_3) |
+			   DPT_GATING_DIS);
+}
+
+static void xehpsdv_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	/* Wa_22010146351:xehpsdv */
+	if (IS_XEHPSDV_GRAPHICS_STEP(dev_priv, STEP_A0, STEP_B0))
+		intel_uncore_rmw(&dev_priv->uncore, XEHP_CLOCK_GATE_DIS, 0, SGR_DIS);
+}
+
+static void dg2_init_clock_gating(struct drm_i915_private *i915)
+{
+	/* Wa_22010954014:dg2 */
+	intel_uncore_rmw(&i915->uncore, XEHP_CLOCK_GATE_DIS, 0,
+			 SGSI_SIDECLK_DIS);
+
+	/*
+	 * Wa_14010733611:dg2_g10
+	 * Wa_22010146351:dg2_g10
+	 */
+	if (IS_DG2_GRAPHICS_STEP(i915, G10, STEP_A0, STEP_B0))
+		intel_uncore_rmw(&i915->uncore, XEHP_CLOCK_GATE_DIS, 0,
+				 SGR_DIS | SGGI_DIS);
+}
+
+static void pvc_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	/* Wa_14012385139:pvc */
+	if (IS_PVC_BD_STEP(dev_priv, STEP_A0, STEP_B0))
+		intel_uncore_rmw(&dev_priv->uncore, XEHP_CLOCK_GATE_DIS, 0, SGR_DIS);
+
+	/* Wa_22010954014:pvc */
+	if (IS_PVC_BD_STEP(dev_priv, STEP_A0, STEP_B0))
+		intel_uncore_rmw(&dev_priv->uncore, XEHP_CLOCK_GATE_DIS, 0, SGSI_SIDECLK_DIS);
+}
+
+static void cnp_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	if (!HAS_PCH_CNP(dev_priv))
+		return;
+
+	/* Display WA #1181 WaSouthDisplayDisablePWMCGEGating: cnp */
+	intel_uncore_write(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D, intel_uncore_read(&dev_priv->uncore, SOUTH_DSPCLK_GATE_D) |
+		   CNP_PWM_CGE_GATING_DISABLE);
+}
+
+static void cfl_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	cnp_init_clock_gating(dev_priv);
+	gen9_init_clock_gating(dev_priv);
+
+	/* WAC6entrylatency:cfl */
+	intel_uncore_write(&dev_priv->uncore, FBC_LLC_READ_CTRL, intel_uncore_read(&dev_priv->uncore, FBC_LLC_READ_CTRL) |
+		   FBC_LLC_FULLY_OPEN);
+
+	/*
+	 * WaFbcTurnOffFbcWatermark:cfl
+	 * Display WA #0562: cfl
+	 */
+	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
+		   DISP_FBC_WM_DIS);
+
+	/*
+	 * WaFbcNukeOnHostModify:cfl
+	 * Display WA #0873: cfl
+	 */
+	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
+			   intel_uncore_read(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A)) |
+			   DPFC_NUKE_ON_ANY_MODIFICATION);
+}
+
+static void kbl_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	gen9_init_clock_gating(dev_priv);
+
+	/* WAC6entrylatency:kbl */
+	intel_uncore_write(&dev_priv->uncore, FBC_LLC_READ_CTRL, intel_uncore_read(&dev_priv->uncore, FBC_LLC_READ_CTRL) |
+		   FBC_LLC_FULLY_OPEN);
+
+	/* WaDisableSDEUnitClockGating:kbl */
+	if (IS_KBL_GRAPHICS_STEP(dev_priv, 0, STEP_C0))
+		intel_uncore_write(&dev_priv->uncore, GEN8_UCGCTL6, intel_uncore_read(&dev_priv->uncore, GEN8_UCGCTL6) |
+			   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
+
+	/* WaDisableGamClockGating:kbl */
+	if (IS_KBL_GRAPHICS_STEP(dev_priv, 0, STEP_C0))
+		intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL1, intel_uncore_read(&dev_priv->uncore, GEN6_UCGCTL1) |
+			   GEN6_GAMUNIT_CLOCK_GATE_DISABLE);
+
+	/*
+	 * WaFbcTurnOffFbcWatermark:kbl
+	 * Display WA #0562: kbl
+	 */
+	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
+		   DISP_FBC_WM_DIS);
+
+	/*
+	 * WaFbcNukeOnHostModify:kbl
+	 * Display WA #0873: kbl
+	 */
+	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
+			   intel_uncore_read(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A)) |
+			   DPFC_NUKE_ON_ANY_MODIFICATION);
+}
+
+static void skl_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	gen9_init_clock_gating(dev_priv);
+
+	/* WaDisableDopClockGating:skl */
+	intel_uncore_write(&dev_priv->uncore, GEN7_MISCCPCTL, intel_uncore_read(&dev_priv->uncore, GEN7_MISCCPCTL) &
+		   ~GEN7_DOP_CLOCK_GATE_ENABLE);
+
+	/* WAC6entrylatency:skl */
+	intel_uncore_write(&dev_priv->uncore, FBC_LLC_READ_CTRL, intel_uncore_read(&dev_priv->uncore, FBC_LLC_READ_CTRL) |
+		   FBC_LLC_FULLY_OPEN);
+
+	/*
+	 * WaFbcTurnOffFbcWatermark:skl
+	 * Display WA #0562: skl
+	 */
+	intel_uncore_write(&dev_priv->uncore, DISP_ARB_CTL, intel_uncore_read(&dev_priv->uncore, DISP_ARB_CTL) |
+		   DISP_FBC_WM_DIS);
+
+	/*
+	 * WaFbcNukeOnHostModify:skl
+	 * Display WA #0873: skl
+	 */
+	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
+			   intel_uncore_read(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A)) |
+			   DPFC_NUKE_ON_ANY_MODIFICATION);
+
+	/*
+	 * WaFbcHighMemBwCorruptionAvoidance:skl
+	 * Display WA #0883: skl
+	 */
+	intel_uncore_write(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A),
+			   intel_uncore_read(&dev_priv->uncore, ILK_DPFC_CHICKEN(INTEL_FBC_A)) |
+			   DPFC_DISABLE_DUMMY0);
+}
+
+static void bdw_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	enum pipe pipe;
+
+	/* WaFbcAsynchFlipDisableFbcQueue:hsw,bdw */
+	intel_uncore_write(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A),
+		   intel_uncore_read(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A)) |
+		   HSW_FBCQ_DIS);
+
+	/* WaSwitchSolVfFArbitrationPriority:bdw */
+	intel_uncore_write(&dev_priv->uncore, GAM_ECOCHK, intel_uncore_read(&dev_priv->uncore, GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL);
+
+	/* WaPsrDPAMaskVBlankInSRD:bdw */
+	intel_uncore_write(&dev_priv->uncore, CHICKEN_PAR1_1,
+		   intel_uncore_read(&dev_priv->uncore, CHICKEN_PAR1_1) | DPA_MASK_VBLANK_SRD);
+
+	for_each_pipe(dev_priv, pipe) {
+		/* WaPsrDPRSUnmaskVBlankInSRD:bdw */
+		intel_uncore_write(&dev_priv->uncore, CHICKEN_PIPESL_1(pipe),
+			   intel_uncore_read(&dev_priv->uncore, CHICKEN_PIPESL_1(pipe)) |
+			   BDW_DPRS_MASK_VBLANK_SRD);
+	}
+
+	/* WaVSRefCountFullforceMissDisable:bdw */
+	/* WaDSRefCountFullforceMissDisable:bdw */
+	intel_uncore_write(&dev_priv->uncore, GEN7_FF_THREAD_MODE,
+		   intel_uncore_read(&dev_priv->uncore, GEN7_FF_THREAD_MODE) &
+		   ~(GEN8_FF_DS_REF_CNT_FFME | GEN7_FF_VS_REF_CNT_FFME));
+
+	intel_uncore_write(&dev_priv->uncore, RING_PSMI_CTL(RENDER_RING_BASE),
+		   _MASKED_BIT_ENABLE(GEN8_RC_SEMA_IDLE_MSG_DISABLE));
+
+	/* WaDisableSDEUnitClockGating:bdw */
+	intel_uncore_write(&dev_priv->uncore, GEN8_UCGCTL6, intel_uncore_read(&dev_priv->uncore, GEN8_UCGCTL6) |
+		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
+
+	/* WaProgramL3SqcReg1Default:bdw */
+	gen8_set_l3sqc_credits(dev_priv, 30, 2);
+
+	/* WaKVMNotificationOnConfigChange:bdw */
+	intel_uncore_write(&dev_priv->uncore, CHICKEN_PAR2_1, intel_uncore_read(&dev_priv->uncore, CHICKEN_PAR2_1)
+		   | KVM_CONFIG_CHANGE_NOTIFICATION_SELECT);
+
+	lpt_init_clock_gating(dev_priv);
+
+	/* WaDisableDopClockGating:bdw
+	 *
+	 * Also see the CHICKEN2 write in bdw_init_workarounds() to disable DOP
+	 * clock gating.
+	 */
+	intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL1,
+		   intel_uncore_read(&dev_priv->uncore, GEN6_UCGCTL1) | GEN6_EU_TCUNIT_CLOCK_GATE_DISABLE);
+}
+
+static void hsw_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	/* WaFbcAsynchFlipDisableFbcQueue:hsw,bdw */
+	intel_uncore_write(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A),
+		   intel_uncore_read(&dev_priv->uncore, CHICKEN_PIPESL_1(PIPE_A)) |
+		   HSW_FBCQ_DIS);
+
+	/* This is required by WaCatErrorRejectionIssue:hsw */
+	intel_uncore_write(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
+		   intel_uncore_read(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
+		   GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
+
+	/* WaSwitchSolVfFArbitrationPriority:hsw */
+	intel_uncore_write(&dev_priv->uncore, GAM_ECOCHK, intel_uncore_read(&dev_priv->uncore, GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL);
+
+	lpt_init_clock_gating(dev_priv);
+}
+
+static void ivb_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	u32 snpcr;
+
+	intel_uncore_write(&dev_priv->uncore, ILK_DSPCLK_GATE_D, ILK_VRHUNIT_CLOCK_GATE_DISABLE);
+
+	/* WaFbcAsynchFlipDisableFbcQueue:ivb */
+	intel_uncore_write(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1,
+		   intel_uncore_read(&dev_priv->uncore, ILK_DISPLAY_CHICKEN1) |
+		   ILK_FBCQ_DIS);
+
+	/* WaDisableBackToBackFlipFix:ivb */
+	intel_uncore_write(&dev_priv->uncore, IVB_CHICKEN3,
+		   CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE |
+		   CHICKEN3_DGMG_DONE_FIX_DISABLE);
+
+	if (IS_IVB_GT1(dev_priv))
+		intel_uncore_write(&dev_priv->uncore, GEN7_ROW_CHICKEN2,
+			   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
+	else {
+		/* must write both registers */
+		intel_uncore_write(&dev_priv->uncore, GEN7_ROW_CHICKEN2,
+			   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
+		intel_uncore_write(&dev_priv->uncore, GEN7_ROW_CHICKEN2_GT2,
+			   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
+	}
+
+	/*
+	 * According to the spec, bit 13 (RCZUNIT) must be set on IVB.
+	 * This implements the WaDisableRCZUnitClockGating:ivb workaround.
+	 */
+	intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL2,
+		   GEN6_RCZUNIT_CLOCK_GATE_DISABLE);
+
+	/* This is required by WaCatErrorRejectionIssue:ivb */
+	intel_uncore_write(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
+			intel_uncore_read(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
+			GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
+
+	g4x_disable_trickle_feed(dev_priv);
+
+	snpcr = intel_uncore_read(&dev_priv->uncore, GEN6_MBCUNIT_SNPCR);
+	snpcr &= ~GEN6_MBC_SNPCR_MASK;
+	snpcr |= GEN6_MBC_SNPCR_MED;
+	intel_uncore_write(&dev_priv->uncore, GEN6_MBCUNIT_SNPCR, snpcr);
+
+	if (!HAS_PCH_NOP(dev_priv))
+		cpt_init_clock_gating(dev_priv);
+
+	gen6_check_mch_setup(dev_priv);
+}
+
+static void vlv_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	/* WaDisableBackToBackFlipFix:vlv */
+	intel_uncore_write(&dev_priv->uncore, IVB_CHICKEN3,
+		   CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE |
+		   CHICKEN3_DGMG_DONE_FIX_DISABLE);
+
+	/* WaDisableDopClockGating:vlv */
+	intel_uncore_write(&dev_priv->uncore, GEN7_ROW_CHICKEN2,
+		   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
+
+	/* This is required by WaCatErrorRejectionIssue:vlv */
+	intel_uncore_write(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
+		   intel_uncore_read(&dev_priv->uncore, GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
+		   GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
+
+	/*
+	 * According to the spec, bit 13 (RCZUNIT) must be set on IVB.
+	 * This implements the WaDisableRCZUnitClockGating:vlv workaround.
+	 */
+	intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL2,
+		   GEN6_RCZUNIT_CLOCK_GATE_DISABLE);
+
+	/* WaDisableL3Bank2xClockGate:vlv
+	 * Disabling L3 clock gating- MMIO 940c[25] = 1
+	 * Set bit 25, to disable L3_BANK_2x_CLK_GATING */
+	intel_uncore_write(&dev_priv->uncore, GEN7_UCGCTL4,
+		   intel_uncore_read(&dev_priv->uncore, GEN7_UCGCTL4) | GEN7_L3BANK2X_CLOCK_GATE_DISABLE);
+
+	/*
+	 * WaDisableVLVClockGating_VBIIssue:vlv
+	 * Disable clock gating on th GCFG unit to prevent a delay
+	 * in the reporting of vblank events.
+	 */
+	intel_uncore_write(&dev_priv->uncore, VLV_GUNIT_CLOCK_GATE, GCFG_DIS);
+}
+
+static void chv_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	/* WaVSRefCountFullforceMissDisable:chv */
+	/* WaDSRefCountFullforceMissDisable:chv */
+	intel_uncore_write(&dev_priv->uncore, GEN7_FF_THREAD_MODE,
+		   intel_uncore_read(&dev_priv->uncore, GEN7_FF_THREAD_MODE) &
+		   ~(GEN8_FF_DS_REF_CNT_FFME | GEN7_FF_VS_REF_CNT_FFME));
+
+	/* WaDisableSemaphoreAndSyncFlipWait:chv */
+	intel_uncore_write(&dev_priv->uncore, RING_PSMI_CTL(RENDER_RING_BASE),
+		   _MASKED_BIT_ENABLE(GEN8_RC_SEMA_IDLE_MSG_DISABLE));
+
+	/* WaDisableCSUnitClockGating:chv */
+	intel_uncore_write(&dev_priv->uncore, GEN6_UCGCTL1, intel_uncore_read(&dev_priv->uncore, GEN6_UCGCTL1) |
+		   GEN6_CSUNIT_CLOCK_GATE_DISABLE);
+
+	/* WaDisableSDEUnitClockGating:chv */
+	intel_uncore_write(&dev_priv->uncore, GEN8_UCGCTL6, intel_uncore_read(&dev_priv->uncore, GEN8_UCGCTL6) |
+		   GEN8_SDEUNIT_CLOCK_GATE_DISABLE);
+
+	/*
+	 * WaProgramL3SqcReg1Default:chv
+	 * See gfxspecs/Related Documents/Performance Guide/
+	 * LSQC Setting Recommendations.
+	 */
+	gen8_set_l3sqc_credits(dev_priv, 38, 2);
+}
+
+static void g4x_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	u32 dspclk_gate;
+
+	intel_uncore_write(&dev_priv->uncore, RENCLK_GATE_D1, 0);
+	intel_uncore_write(&dev_priv->uncore, RENCLK_GATE_D2, VF_UNIT_CLOCK_GATE_DISABLE |
+		   GS_UNIT_CLOCK_GATE_DISABLE |
+		   CL_UNIT_CLOCK_GATE_DISABLE);
+	intel_uncore_write(&dev_priv->uncore, RAMCLK_GATE_D, 0);
+	dspclk_gate = VRHUNIT_CLOCK_GATE_DISABLE |
+		OVRUNIT_CLOCK_GATE_DISABLE |
+		OVCUNIT_CLOCK_GATE_DISABLE;
+	if (IS_GM45(dev_priv))
+		dspclk_gate |= DSSUNIT_CLOCK_GATE_DISABLE;
+	intel_uncore_write(&dev_priv->uncore, DSPCLK_GATE_D(dev_priv), dspclk_gate);
+
+	g4x_disable_trickle_feed(dev_priv);
+}
+
+static void i965gm_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	struct intel_uncore *uncore = &dev_priv->uncore;
+
+	intel_uncore_write(uncore, RENCLK_GATE_D1, I965_RCC_CLOCK_GATE_DISABLE);
+	intel_uncore_write(uncore, RENCLK_GATE_D2, 0);
+	intel_uncore_write(uncore, DSPCLK_GATE_D(dev_priv), 0);
+	intel_uncore_write(uncore, RAMCLK_GATE_D, 0);
+	intel_uncore_write16(uncore, DEUC, 0);
+	intel_uncore_write(uncore,
+			   MI_ARB_STATE,
+			   _MASKED_BIT_ENABLE(MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE));
+}
+
+static void i965g_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	intel_uncore_write(&dev_priv->uncore, RENCLK_GATE_D1, I965_RCZ_CLOCK_GATE_DISABLE |
+		   I965_RCC_CLOCK_GATE_DISABLE |
+		   I965_RCPB_CLOCK_GATE_DISABLE |
+		   I965_ISC_CLOCK_GATE_DISABLE |
+		   I965_FBC_CLOCK_GATE_DISABLE);
+	intel_uncore_write(&dev_priv->uncore, RENCLK_GATE_D2, 0);
+	intel_uncore_write(&dev_priv->uncore, MI_ARB_STATE,
+		   _MASKED_BIT_ENABLE(MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE));
+}
+
+static void gen3_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	u32 dstate = intel_uncore_read(&dev_priv->uncore, D_STATE);
+
+	dstate |= DSTATE_PLL_D3_OFF | DSTATE_GFX_CLOCK_GATING |
+		DSTATE_DOT_CLOCK_GATING;
+	intel_uncore_write(&dev_priv->uncore, D_STATE, dstate);
+
+	if (IS_PINEVIEW(dev_priv))
+		intel_uncore_write(&dev_priv->uncore, ECOSKPD(RENDER_RING_BASE),
+				   _MASKED_BIT_ENABLE(ECO_GATING_CX_ONLY));
+
+	/* IIR "flip pending" means done if this bit is set */
+	intel_uncore_write(&dev_priv->uncore, ECOSKPD(RENDER_RING_BASE),
+			   _MASKED_BIT_DISABLE(ECO_FLIP_DONE));
+
+	/* interrupts should cause a wake up from C3 */
+	intel_uncore_write(&dev_priv->uncore, INSTPM, _MASKED_BIT_ENABLE(INSTPM_AGPBUSY_INT_EN));
+
+	/* On GEN3 we really need to make sure the ARB C3 LP bit is set */
+	intel_uncore_write(&dev_priv->uncore, MI_ARB_STATE, _MASKED_BIT_ENABLE(MI_ARB_C3_LP_WRITE_ENABLE));
+
+	intel_uncore_write(&dev_priv->uncore, MI_ARB_STATE,
+		   _MASKED_BIT_ENABLE(MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE));
+}
+
+static void i85x_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	intel_uncore_write(&dev_priv->uncore, RENCLK_GATE_D1, SV_CLOCK_GATE_DISABLE);
+
+	/* interrupts should cause a wake up from C3 */
+	intel_uncore_write(&dev_priv->uncore, MI_STATE, _MASKED_BIT_ENABLE(MI_AGPBUSY_INT_EN) |
+		   _MASKED_BIT_DISABLE(MI_AGPBUSY_830_MODE));
+
+	intel_uncore_write(&dev_priv->uncore, MEM_MODE,
+		   _MASKED_BIT_ENABLE(MEM_DISPLAY_TRICKLE_FEED_DISABLE));
+
+	/*
+	 * Have FBC ignore 3D activity since we use software
+	 * render tracking, and otherwise a pure 3D workload
+	 * (even if it just renders a single frame and then does
+	 * abosultely nothing) would not allow FBC to recompress
+	 * until a 2D blit occurs.
+	 */
+	intel_uncore_write(&dev_priv->uncore, SCPD0,
+		   _MASKED_BIT_ENABLE(SCPD_FBC_IGNORE_3D));
+}
+
+static void i830_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	intel_uncore_write(&dev_priv->uncore, MEM_MODE,
+		   _MASKED_BIT_ENABLE(MEM_DISPLAY_A_TRICKLE_FEED_DISABLE) |
+		   _MASKED_BIT_ENABLE(MEM_DISPLAY_B_TRICKLE_FEED_DISABLE));
+}
+
+void intel_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	dev_priv->clock_gating_funcs->init_clock_gating(dev_priv);
+}
+
+void intel_suspend_hw(struct drm_i915_private *dev_priv)
+{
+	if (HAS_PCH_LPT(dev_priv))
+		lpt_suspend_hw(dev_priv);
+}
+
+static void nop_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	drm_dbg_kms(&dev_priv->drm,
+		    "No clock gating settings or workarounds applied.\n");
+}
+
+#define CG_FUNCS(platform)						\
+static const struct drm_i915_clock_gating_funcs platform##_clock_gating_funcs = { \
+	.init_clock_gating = platform##_init_clock_gating,		\
+}
+
+CG_FUNCS(pvc);
+CG_FUNCS(dg2);
+CG_FUNCS(xehpsdv);
+CG_FUNCS(adlp);
+CG_FUNCS(dg1);
+CG_FUNCS(gen12lp);
+CG_FUNCS(icl);
+CG_FUNCS(cfl);
+CG_FUNCS(skl);
+CG_FUNCS(kbl);
+CG_FUNCS(bxt);
+CG_FUNCS(glk);
+CG_FUNCS(bdw);
+CG_FUNCS(chv);
+CG_FUNCS(hsw);
+CG_FUNCS(ivb);
+CG_FUNCS(vlv);
+CG_FUNCS(gen6);
+CG_FUNCS(ilk);
+CG_FUNCS(g4x);
+CG_FUNCS(i965gm);
+CG_FUNCS(i965g);
+CG_FUNCS(gen3);
+CG_FUNCS(i85x);
+CG_FUNCS(i830);
+CG_FUNCS(nop);
+#undef CG_FUNCS
+
+/**
+ * intel_init_clock_gating_hooks - setup the clock gating hooks
+ * @dev_priv: device private
+ *
+ * Setup the hooks that configure which clocks of a given platform can be
+ * gated and also apply various GT and display specific workarounds for these
+ * platforms. Note that some GT specific workarounds are applied separately
+ * when GPU contexts or batchbuffers start their execution.
+ */
+void intel_init_clock_gating_hooks(struct drm_i915_private *dev_priv)
+{
+	if (IS_PONTEVECCHIO(dev_priv))
+		dev_priv->clock_gating_funcs = &pvc_clock_gating_funcs;
+	else if (IS_DG2(dev_priv))
+		dev_priv->clock_gating_funcs = &dg2_clock_gating_funcs;
+	else if (IS_XEHPSDV(dev_priv))
+		dev_priv->clock_gating_funcs = &xehpsdv_clock_gating_funcs;
+	else if (IS_ALDERLAKE_P(dev_priv))
+		dev_priv->clock_gating_funcs = &adlp_clock_gating_funcs;
+	else if (IS_DG1(dev_priv))
+		dev_priv->clock_gating_funcs = &dg1_clock_gating_funcs;
+	else if (GRAPHICS_VER(dev_priv) == 12)
+		dev_priv->clock_gating_funcs = &gen12lp_clock_gating_funcs;
+	else if (GRAPHICS_VER(dev_priv) == 11)
+		dev_priv->clock_gating_funcs = &icl_clock_gating_funcs;
+	else if (IS_COFFEELAKE(dev_priv) || IS_COMETLAKE(dev_priv))
+		dev_priv->clock_gating_funcs = &cfl_clock_gating_funcs;
+	else if (IS_SKYLAKE(dev_priv))
+		dev_priv->clock_gating_funcs = &skl_clock_gating_funcs;
+	else if (IS_KABYLAKE(dev_priv))
+		dev_priv->clock_gating_funcs = &kbl_clock_gating_funcs;
+	else if (IS_BROXTON(dev_priv))
+		dev_priv->clock_gating_funcs = &bxt_clock_gating_funcs;
+	else if (IS_GEMINILAKE(dev_priv))
+		dev_priv->clock_gating_funcs = &glk_clock_gating_funcs;
+	else if (IS_BROADWELL(dev_priv))
+		dev_priv->clock_gating_funcs = &bdw_clock_gating_funcs;
+	else if (IS_CHERRYVIEW(dev_priv))
+		dev_priv->clock_gating_funcs = &chv_clock_gating_funcs;
+	else if (IS_HASWELL(dev_priv))
+		dev_priv->clock_gating_funcs = &hsw_clock_gating_funcs;
+	else if (IS_IVYBRIDGE(dev_priv))
+		dev_priv->clock_gating_funcs = &ivb_clock_gating_funcs;
+	else if (IS_VALLEYVIEW(dev_priv))
+		dev_priv->clock_gating_funcs = &vlv_clock_gating_funcs;
+	else if (GRAPHICS_VER(dev_priv) == 6)
+		dev_priv->clock_gating_funcs = &gen6_clock_gating_funcs;
+	else if (GRAPHICS_VER(dev_priv) == 5)
+		dev_priv->clock_gating_funcs = &ilk_clock_gating_funcs;
+	else if (IS_G4X(dev_priv))
+		dev_priv->clock_gating_funcs = &g4x_clock_gating_funcs;
+	else if (IS_I965GM(dev_priv))
+		dev_priv->clock_gating_funcs = &i965gm_clock_gating_funcs;
+	else if (IS_I965G(dev_priv))
+		dev_priv->clock_gating_funcs = &i965g_clock_gating_funcs;
+	else if (GRAPHICS_VER(dev_priv) == 3)
+		dev_priv->clock_gating_funcs = &gen3_clock_gating_funcs;
+	else if (IS_I85X(dev_priv) || IS_I865G(dev_priv))
+		dev_priv->clock_gating_funcs = &i85x_clock_gating_funcs;
+	else if (GRAPHICS_VER(dev_priv) == 2)
+		dev_priv->clock_gating_funcs = &i830_clock_gating_funcs;
+	else {
+		MISSING_CASE(INTEL_DEVID(dev_priv));
+		dev_priv->clock_gating_funcs = &nop_clock_gating_funcs;
+	}
+}
+
+static const struct intel_wm_funcs ilk_wm_funcs = {
+	.compute_pipe_wm = ilk_compute_pipe_wm,
+	.compute_intermediate_wm = ilk_compute_intermediate_wm,
+	.initial_watermarks = ilk_initial_watermarks,
+	.optimize_watermarks = ilk_optimize_watermarks,
+};
+
+static const struct intel_wm_funcs vlv_wm_funcs = {
+	.compute_pipe_wm = vlv_compute_pipe_wm,
+	.compute_intermediate_wm = vlv_compute_intermediate_wm,
+	.initial_watermarks = vlv_initial_watermarks,
+	.optimize_watermarks = vlv_optimize_watermarks,
+	.atomic_update_watermarks = vlv_atomic_update_fifo,
+};
+
+static const struct intel_wm_funcs g4x_wm_funcs = {
+	.compute_pipe_wm = g4x_compute_pipe_wm,
+	.compute_intermediate_wm = g4x_compute_intermediate_wm,
+	.initial_watermarks = g4x_initial_watermarks,
+	.optimize_watermarks = g4x_optimize_watermarks,
+};
+
+static const struct intel_wm_funcs pnv_wm_funcs = {
+	.update_wm = pnv_update_wm,
+};
+
+static const struct intel_wm_funcs i965_wm_funcs = {
+	.update_wm = i965_update_wm,
+};
+
+static const struct intel_wm_funcs i9xx_wm_funcs = {
+	.update_wm = i9xx_update_wm,
+};
+
+static const struct intel_wm_funcs i845_wm_funcs = {
+	.update_wm = i845_update_wm,
+};
+
+static const struct intel_wm_funcs nop_funcs = {
+};
+
+/* Set up chip specific power management-related functions */
+void intel_init_pm(struct drm_i915_private *dev_priv)
+{
+	if (DISPLAY_VER(dev_priv) >= 9) {
+		skl_wm_init(dev_priv);
+		return;
+	}
+
+	/* For cxsr */
+	if (IS_PINEVIEW(dev_priv))
+		pnv_get_mem_freq(dev_priv);
+	else if (GRAPHICS_VER(dev_priv) == 5)
+		ilk_get_mem_freq(dev_priv);
+
+	/* For FIFO watermark updates */
+	if (HAS_PCH_SPLIT(dev_priv)) {
+		ilk_setup_wm_latency(dev_priv);
+
+		if ((DISPLAY_VER(dev_priv) == 5 && dev_priv->display.wm.pri_latency[1] &&
+		     dev_priv->display.wm.spr_latency[1] && dev_priv->display.wm.cur_latency[1]) ||
+		    (DISPLAY_VER(dev_priv) != 5 && dev_priv->display.wm.pri_latency[0] &&
+		     dev_priv->display.wm.spr_latency[0] && dev_priv->display.wm.cur_latency[0])) {
+			dev_priv->display.funcs.wm = &ilk_wm_funcs;
+		} else {
+			drm_dbg_kms(&dev_priv->drm,
+				    "Failed to read display plane latency. "
+				    "Disable CxSR\n");
+			dev_priv->display.funcs.wm = &nop_funcs;
+		}
+	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
+		vlv_setup_wm_latency(dev_priv);
+		dev_priv->display.funcs.wm = &vlv_wm_funcs;
+	} else if (IS_G4X(dev_priv)) {
+		g4x_setup_wm_latency(dev_priv);
+		dev_priv->display.funcs.wm = &g4x_wm_funcs;
+	} else if (IS_PINEVIEW(dev_priv)) {
+		if (!intel_get_cxsr_latency(!IS_MOBILE(dev_priv),
+					    dev_priv->is_ddr3,
+					    dev_priv->fsb_freq,
+					    dev_priv->mem_freq)) {
+			drm_info(&dev_priv->drm,
+				 "failed to find known CxSR latency "
+				 "(found ddr%s fsb freq %d, mem freq %d), "
+				 "disabling CxSR\n",
+				 (dev_priv->is_ddr3 == 1) ? "3" : "2",
+				 dev_priv->fsb_freq, dev_priv->mem_freq);
+			/* Disable CxSR and never update its watermark again */
+			intel_set_memory_cxsr(dev_priv, false);
+			dev_priv->display.funcs.wm = &nop_funcs;
+		} else
+			dev_priv->display.funcs.wm = &pnv_wm_funcs;
+	} else if (DISPLAY_VER(dev_priv) == 4) {
+		dev_priv->display.funcs.wm = &i965_wm_funcs;
+	} else if (DISPLAY_VER(dev_priv) == 3) {
+		dev_priv->display.funcs.wm = &i9xx_wm_funcs;
+	} else if (DISPLAY_VER(dev_priv) == 2) {
+		if (INTEL_NUM_PIPES(dev_priv) == 1)
+			dev_priv->display.funcs.wm = &i845_wm_funcs;
+		else
+			dev_priv->display.funcs.wm = &i9xx_wm_funcs;
+	} else {
+		drm_err(&dev_priv->drm,
+			"unexpected fall-through in %s\n", __func__);
+		dev_priv->display.funcs.wm = &nop_funcs;
+	}
+}
+
+void intel_pm_setup(struct drm_i915_private *dev_priv)
+{
+	dev_priv->runtime_pm.suspended = false;
+	atomic_set(&dev_priv->runtime_pm.wakeref_count, 0);
+}
diff --git a/drivers/gpu/drm/i915/intel_pm.h b/drivers/gpu/drm/i915/intel_pm.h
new file mode 100644
index 000000000..c09b872d6
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_pm.h
@@ -0,0 +1,34 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_PM_H__
+#define __INTEL_PM_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct intel_crtc_state;
+struct intel_plane_state;
+
+void intel_init_clock_gating(struct drm_i915_private *dev_priv);
+void intel_suspend_hw(struct drm_i915_private *dev_priv);
+int ilk_wm_max_level(const struct drm_i915_private *dev_priv);
+void intel_init_pm(struct drm_i915_private *dev_priv);
+void intel_init_clock_gating_hooks(struct drm_i915_private *dev_priv);
+void intel_pm_setup(struct drm_i915_private *dev_priv);
+void g4x_wm_get_hw_state(struct drm_i915_private *dev_priv);
+void vlv_wm_get_hw_state(struct drm_i915_private *dev_priv);
+void ilk_wm_get_hw_state(struct drm_i915_private *dev_priv);
+void g4x_wm_sanitize(struct drm_i915_private *dev_priv);
+void vlv_wm_sanitize(struct drm_i915_private *dev_priv);
+bool ilk_disable_lp_wm(struct drm_i915_private *dev_priv);
+bool intel_wm_plane_visible(const struct intel_crtc_state *crtc_state,
+			    const struct intel_plane_state *plane_state);
+void intel_print_wm_latency(struct drm_i915_private *dev_priv,
+			    const char *name, const u16 wm[]);
+
+bool intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enable);
+
+#endif /* __INTEL_PM_H__ */
diff --git a/drivers/gpu/drm/i915/intel_pm_types.h b/drivers/gpu/drm/i915/intel_pm_types.h
new file mode 100644
index 000000000..211632f58
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_pm_types.h
@@ -0,0 +1,76 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __INTEL_PM_TYPES_H__
+#define __INTEL_PM_TYPES_H__
+
+#include <linux/types.h>
+
+#include "display/intel_display.h"
+
+enum intel_ddb_partitioning {
+	INTEL_DDB_PART_1_2,
+	INTEL_DDB_PART_5_6, /* IVB+ */
+};
+
+struct ilk_wm_values {
+	u32 wm_pipe[3];
+	u32 wm_lp[3];
+	u32 wm_lp_spr[3];
+	bool enable_fbc_wm;
+	enum intel_ddb_partitioning partitioning;
+};
+
+struct g4x_pipe_wm {
+	u16 plane[I915_MAX_PLANES];
+	u16 fbc;
+};
+
+struct g4x_sr_wm {
+	u16 plane;
+	u16 cursor;
+	u16 fbc;
+};
+
+struct vlv_wm_ddl_values {
+	u8 plane[I915_MAX_PLANES];
+};
+
+struct vlv_wm_values {
+	struct g4x_pipe_wm pipe[3];
+	struct g4x_sr_wm sr;
+	struct vlv_wm_ddl_values ddl[3];
+	u8 level;
+	bool cxsr;
+};
+
+struct g4x_wm_values {
+	struct g4x_pipe_wm pipe[2];
+	struct g4x_sr_wm sr;
+	struct g4x_sr_wm hpll;
+	bool cxsr;
+	bool hpll_en;
+	bool fbc_en;
+};
+
+struct skl_ddb_entry {
+	u16 start, end;	/* in number of blocks, 'end' is exclusive */
+};
+
+static inline u16 skl_ddb_entry_size(const struct skl_ddb_entry *entry)
+{
+	return entry->end - entry->start;
+}
+
+static inline bool skl_ddb_entry_equal(const struct skl_ddb_entry *e1,
+				       const struct skl_ddb_entry *e2)
+{
+	if (e1->start == e2->start && e1->end == e2->end)
+		return true;
+
+	return false;
+}
+
+#endif /* __INTEL_PM_TYPES_H__ */
diff --git a/drivers/gpu/drm/i915/intel_region_ttm.c b/drivers/gpu/drm/i915/intel_region_ttm.c
new file mode 100644
index 000000000..575d67bc6
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_region_ttm.c
@@ -0,0 +1,252 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+#include <drm/ttm/ttm_bo_driver.h>
+#include <drm/ttm/ttm_device.h>
+#include <drm/ttm/ttm_range_manager.h>
+
+#include "i915_drv.h"
+#include "i915_scatterlist.h"
+#include "i915_ttm_buddy_manager.h"
+
+#include "intel_region_ttm.h"
+
+#include "gem/i915_gem_region.h"
+#include "gem/i915_gem_ttm.h" /* For the funcs/ops export only */
+/**
+ * DOC: TTM support structure
+ *
+ * The code in this file deals with setting up memory managers for TTM
+ * LMEM and MOCK regions and converting the output from
+ * the managers to struct sg_table, Basically providing the mapping from
+ * i915 GEM regions to TTM memory types and resource managers.
+ */
+
+/**
+ * intel_region_ttm_device_init - Initialize a TTM device
+ * @dev_priv: Pointer to an i915 device private structure.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int intel_region_ttm_device_init(struct drm_i915_private *dev_priv)
+{
+	struct drm_device *drm = &dev_priv->drm;
+
+	return ttm_device_init(&dev_priv->bdev, i915_ttm_driver(),
+			       drm->dev, drm->anon_inode->i_mapping,
+			       drm->vma_offset_manager, false, false);
+}
+
+/**
+ * intel_region_ttm_device_fini - Finalize a TTM device
+ * @dev_priv: Pointer to an i915 device private structure.
+ */
+void intel_region_ttm_device_fini(struct drm_i915_private *dev_priv)
+{
+	ttm_device_fini(&dev_priv->bdev);
+}
+
+/*
+ * Map the i915 memory regions to TTM memory types. We use the
+ * driver-private types for now, reserving TTM_PL_VRAM for stolen
+ * memory and TTM_PL_TT for GGTT use if decided to implement this.
+ */
+int intel_region_to_ttm_type(const struct intel_memory_region *mem)
+{
+	int type;
+
+	GEM_BUG_ON(mem->type != INTEL_MEMORY_LOCAL &&
+		   mem->type != INTEL_MEMORY_MOCK &&
+		   mem->type != INTEL_MEMORY_SYSTEM);
+
+	if (mem->type == INTEL_MEMORY_SYSTEM)
+		return TTM_PL_SYSTEM;
+
+	type = mem->instance + TTM_PL_PRIV;
+	GEM_BUG_ON(type >= TTM_NUM_MEM_TYPES);
+
+	return type;
+}
+
+/**
+ * intel_region_ttm_init - Initialize a memory region for TTM.
+ * @mem: The region to initialize.
+ *
+ * This function initializes a suitable TTM resource manager for the
+ * region, and if it's a LMEM region type, attaches it to the TTM
+ * device. MOCK regions are NOT attached to the TTM device, since we don't
+ * have one for the mock selftests.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int intel_region_ttm_init(struct intel_memory_region *mem)
+{
+	struct ttm_device *bdev = &mem->i915->bdev;
+	int mem_type = intel_region_to_ttm_type(mem);
+	int ret;
+
+	ret = i915_ttm_buddy_man_init(bdev, mem_type, false,
+				      resource_size(&mem->region),
+				      mem->io_size,
+				      mem->min_page_size, PAGE_SIZE);
+	if (ret)
+		return ret;
+
+	mem->region_private = ttm_manager_type(bdev, mem_type);
+
+	return 0;
+}
+
+/**
+ * intel_region_ttm_fini - Finalize a TTM region.
+ * @mem: The memory region
+ *
+ * This functions takes down the TTM resource manager associated with the
+ * memory region, and if it was registered with the TTM device,
+ * removes that registration.
+ */
+int intel_region_ttm_fini(struct intel_memory_region *mem)
+{
+	struct ttm_resource_manager *man = mem->region_private;
+	int ret = -EBUSY;
+	int count;
+
+	/*
+	 * Put the region's move fences. This releases requests that
+	 * may hold on to contexts and vms that may hold on to buffer
+	 * objects placed in this region.
+	 */
+	if (man)
+		ttm_resource_manager_cleanup(man);
+
+	/* Flush objects from region. */
+	for (count = 0; count < 10; ++count) {
+		i915_gem_flush_free_objects(mem->i915);
+
+		mutex_lock(&mem->objects.lock);
+		if (list_empty(&mem->objects.list))
+			ret = 0;
+		mutex_unlock(&mem->objects.lock);
+		if (!ret)
+			break;
+
+		msleep(20);
+		flush_delayed_work(&mem->i915->bdev.wq);
+	}
+
+	/* If we leaked objects, Don't free the region causing use after free */
+	if (ret || !man)
+		return ret;
+
+	ret = i915_ttm_buddy_man_fini(&mem->i915->bdev,
+				      intel_region_to_ttm_type(mem));
+	GEM_WARN_ON(ret);
+	mem->region_private = NULL;
+
+	return ret;
+}
+
+/**
+ * intel_region_ttm_resource_to_rsgt -
+ * Convert an opaque TTM resource manager resource to a refcounted sg_table.
+ * @mem: The memory region.
+ * @res: The resource manager resource obtained from the TTM resource manager.
+ * @page_alignment: Required page alignment for each sg entry. Power of two.
+ *
+ * The gem backends typically use sg-tables for operations on the underlying
+ * io_memory. So provide a way for the backends to translate the
+ * nodes they are handed from TTM to sg-tables.
+ *
+ * Return: A malloced sg_table on success, an error pointer on failure.
+ */
+struct i915_refct_sgt *
+intel_region_ttm_resource_to_rsgt(struct intel_memory_region *mem,
+				  struct ttm_resource *res,
+				  u32 page_alignment)
+{
+	if (mem->is_range_manager) {
+		struct ttm_range_mgr_node *range_node =
+			to_ttm_range_mgr_node(res);
+
+		return i915_rsgt_from_mm_node(&range_node->mm_nodes[0],
+					      mem->region.start,
+					      page_alignment);
+	} else {
+		return i915_rsgt_from_buddy_resource(res, mem->region.start,
+						     page_alignment);
+	}
+}
+
+#ifdef CONFIG_DRM_I915_SELFTEST
+/**
+ * intel_region_ttm_resource_alloc - Allocate memory resources from a region
+ * @mem: The memory region,
+ * @size: The requested size in bytes
+ * @flags: Allocation flags
+ *
+ * This functionality is provided only for callers that need to allocate
+ * memory from standalone TTM range managers, without the TTM eviction
+ * functionality. Don't use if you are not completely sure that's the
+ * case. The returned opaque node can be converted to an sg_table using
+ * intel_region_ttm_resource_to_st(), and can be freed using
+ * intel_region_ttm_resource_free().
+ *
+ * Return: A valid pointer on success, an error pointer on failure.
+ */
+struct ttm_resource *
+intel_region_ttm_resource_alloc(struct intel_memory_region *mem,
+				resource_size_t offset,
+				resource_size_t size,
+				unsigned int flags)
+{
+	struct ttm_resource_manager *man = mem->region_private;
+	struct ttm_place place = {};
+	struct ttm_buffer_object mock_bo = {};
+	struct ttm_resource *res;
+	int ret;
+
+	if (flags & I915_BO_ALLOC_CONTIGUOUS)
+		place.flags |= TTM_PL_FLAG_CONTIGUOUS;
+	if (offset != I915_BO_INVALID_OFFSET) {
+		place.fpfn = offset >> PAGE_SHIFT;
+		place.lpfn = place.fpfn + (size >> PAGE_SHIFT);
+	} else if (mem->io_size && mem->io_size < mem->total) {
+		if (flags & I915_BO_ALLOC_GPU_ONLY) {
+			place.flags |= TTM_PL_FLAG_TOPDOWN;
+		} else {
+			place.fpfn = 0;
+			place.lpfn = mem->io_size >> PAGE_SHIFT;
+		}
+	}
+
+	mock_bo.base.size = size;
+	mock_bo.bdev = &mem->i915->bdev;
+
+	ret = man->func->alloc(man, &mock_bo, &place, &res);
+	if (ret == -ENOSPC)
+		ret = -ENXIO;
+	if (!ret)
+		res->bo = NULL; /* Rather blow up, then some uaf */
+	return ret ? ERR_PTR(ret) : res;
+}
+
+#endif
+
+/**
+ * intel_region_ttm_resource_free - Free a resource allocated from a resource manager
+ * @mem: The region the resource was allocated from.
+ * @res: The opaque resource representing an allocation.
+ */
+void intel_region_ttm_resource_free(struct intel_memory_region *mem,
+				    struct ttm_resource *res)
+{
+	struct ttm_resource_manager *man = mem->region_private;
+	struct ttm_buffer_object mock_bo = {};
+
+	mock_bo.base.size = res->num_pages << PAGE_SHIFT;
+	mock_bo.bdev = &mem->i915->bdev;
+	res->bo = &mock_bo;
+
+	man->func->free(man, res);
+}
diff --git a/drivers/gpu/drm/i915/intel_region_ttm.h b/drivers/gpu/drm/i915/intel_region_ttm.h
new file mode 100644
index 000000000..5bb8d8b58
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_region_ttm.h
@@ -0,0 +1,44 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+#ifndef _INTEL_REGION_TTM_H_
+#define _INTEL_REGION_TTM_H_
+
+#include <linux/types.h>
+
+#include "i915_selftest.h"
+
+struct drm_i915_private;
+struct intel_memory_region;
+struct ttm_resource;
+struct ttm_device_funcs;
+
+int intel_region_ttm_device_init(struct drm_i915_private *dev_priv);
+
+void intel_region_ttm_device_fini(struct drm_i915_private *dev_priv);
+
+int intel_region_ttm_init(struct intel_memory_region *mem);
+
+int intel_region_ttm_fini(struct intel_memory_region *mem);
+
+struct i915_refct_sgt *
+intel_region_ttm_resource_to_rsgt(struct intel_memory_region *mem,
+				  struct ttm_resource *res,
+				  u32 page_alignment);
+
+void intel_region_ttm_resource_free(struct intel_memory_region *mem,
+				    struct ttm_resource *res);
+
+int intel_region_to_ttm_type(const struct intel_memory_region *mem);
+
+struct ttm_device_funcs *i915_ttm_driver(void);
+
+#ifdef CONFIG_DRM_I915_SELFTEST
+struct ttm_resource *
+intel_region_ttm_resource_alloc(struct intel_memory_region *mem,
+				resource_size_t offset,
+				resource_size_t size,
+				unsigned int flags);
+#endif
+#endif /* _INTEL_REGION_TTM_H_ */
diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
new file mode 100644
index 000000000..129746713
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
@@ -0,0 +1,660 @@
+/*
+ * Copyright © 2012-2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eugeni Dodonov <eugeni.dodonov@intel.com>
+ *    Daniel Vetter <daniel.vetter@ffwll.ch>
+ *
+ */
+
+#include <linux/pm_runtime.h>
+
+#include <drm/drm_print.h>
+
+#include "i915_drv.h"
+#include "i915_trace.h"
+
+/**
+ * DOC: runtime pm
+ *
+ * The i915 driver supports dynamic enabling and disabling of entire hardware
+ * blocks at runtime. This is especially important on the display side where
+ * software is supposed to control many power gates manually on recent hardware,
+ * since on the GT side a lot of the power management is done by the hardware.
+ * But even there some manual control at the device level is required.
+ *
+ * Since i915 supports a diverse set of platforms with a unified codebase and
+ * hardware engineers just love to shuffle functionality around between power
+ * domains there's a sizeable amount of indirection required. This file provides
+ * generic functions to the driver for grabbing and releasing references for
+ * abstract power domains. It then maps those to the actual power wells
+ * present for a given platform.
+ */
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+
+#include <linux/sort.h>
+
+#define STACKDEPTH 8
+
+static noinline depot_stack_handle_t __save_depot_stack(void)
+{
+	unsigned long entries[STACKDEPTH];
+	unsigned int n;
+
+	n = stack_trace_save(entries, ARRAY_SIZE(entries), 1);
+	return stack_depot_save(entries, n, GFP_NOWAIT | __GFP_NOWARN);
+}
+
+static void init_intel_runtime_pm_wakeref(struct intel_runtime_pm *rpm)
+{
+	spin_lock_init(&rpm->debug.lock);
+	stack_depot_init();
+}
+
+static noinline depot_stack_handle_t
+track_intel_runtime_pm_wakeref(struct intel_runtime_pm *rpm)
+{
+	depot_stack_handle_t stack, *stacks;
+	unsigned long flags;
+
+	if (rpm->no_wakeref_tracking)
+		return -1;
+
+	stack = __save_depot_stack();
+	if (!stack)
+		return -1;
+
+	spin_lock_irqsave(&rpm->debug.lock, flags);
+
+	if (!rpm->debug.count)
+		rpm->debug.last_acquire = stack;
+
+	stacks = krealloc(rpm->debug.owners,
+			  (rpm->debug.count + 1) * sizeof(*stacks),
+			  GFP_NOWAIT | __GFP_NOWARN);
+	if (stacks) {
+		stacks[rpm->debug.count++] = stack;
+		rpm->debug.owners = stacks;
+	} else {
+		stack = -1;
+	}
+
+	spin_unlock_irqrestore(&rpm->debug.lock, flags);
+
+	return stack;
+}
+
+static void untrack_intel_runtime_pm_wakeref(struct intel_runtime_pm *rpm,
+					     depot_stack_handle_t stack)
+{
+	struct drm_i915_private *i915 = container_of(rpm,
+						     struct drm_i915_private,
+						     runtime_pm);
+	unsigned long flags, n;
+	bool found = false;
+
+	if (unlikely(stack == -1))
+		return;
+
+	spin_lock_irqsave(&rpm->debug.lock, flags);
+	for (n = rpm->debug.count; n--; ) {
+		if (rpm->debug.owners[n] == stack) {
+			memmove(rpm->debug.owners + n,
+				rpm->debug.owners + n + 1,
+				(--rpm->debug.count - n) * sizeof(stack));
+			found = true;
+			break;
+		}
+	}
+	spin_unlock_irqrestore(&rpm->debug.lock, flags);
+
+	if (drm_WARN(&i915->drm, !found,
+		     "Unmatched wakeref (tracking %lu), count %u\n",
+		     rpm->debug.count, atomic_read(&rpm->wakeref_count))) {
+		char *buf;
+
+		buf = kmalloc(PAGE_SIZE, GFP_NOWAIT | __GFP_NOWARN);
+		if (!buf)
+			return;
+
+		stack_depot_snprint(stack, buf, PAGE_SIZE, 2);
+		DRM_DEBUG_DRIVER("wakeref %x from\n%s", stack, buf);
+
+		stack = READ_ONCE(rpm->debug.last_release);
+		if (stack) {
+			stack_depot_snprint(stack, buf, PAGE_SIZE, 2);
+			DRM_DEBUG_DRIVER("wakeref last released at\n%s", buf);
+		}
+
+		kfree(buf);
+	}
+}
+
+static int cmphandle(const void *_a, const void *_b)
+{
+	const depot_stack_handle_t * const a = _a, * const b = _b;
+
+	if (*a < *b)
+		return -1;
+	else if (*a > *b)
+		return 1;
+	else
+		return 0;
+}
+
+static void
+__print_intel_runtime_pm_wakeref(struct drm_printer *p,
+				 const struct intel_runtime_pm_debug *dbg)
+{
+	unsigned long i;
+	char *buf;
+
+	buf = kmalloc(PAGE_SIZE, GFP_NOWAIT | __GFP_NOWARN);
+	if (!buf)
+		return;
+
+	if (dbg->last_acquire) {
+		stack_depot_snprint(dbg->last_acquire, buf, PAGE_SIZE, 2);
+		drm_printf(p, "Wakeref last acquired:\n%s", buf);
+	}
+
+	if (dbg->last_release) {
+		stack_depot_snprint(dbg->last_release, buf, PAGE_SIZE, 2);
+		drm_printf(p, "Wakeref last released:\n%s", buf);
+	}
+
+	drm_printf(p, "Wakeref count: %lu\n", dbg->count);
+
+	sort(dbg->owners, dbg->count, sizeof(*dbg->owners), cmphandle, NULL);
+
+	for (i = 0; i < dbg->count; i++) {
+		depot_stack_handle_t stack = dbg->owners[i];
+		unsigned long rep;
+
+		rep = 1;
+		while (i + 1 < dbg->count && dbg->owners[i + 1] == stack)
+			rep++, i++;
+		stack_depot_snprint(stack, buf, PAGE_SIZE, 2);
+		drm_printf(p, "Wakeref x%lu taken at:\n%s", rep, buf);
+	}
+
+	kfree(buf);
+}
+
+static noinline void
+__untrack_all_wakerefs(struct intel_runtime_pm_debug *debug,
+		       struct intel_runtime_pm_debug *saved)
+{
+	*saved = *debug;
+
+	debug->owners = NULL;
+	debug->count = 0;
+	debug->last_release = __save_depot_stack();
+}
+
+static void
+dump_and_free_wakeref_tracking(struct intel_runtime_pm_debug *debug)
+{
+	if (debug->count) {
+		struct drm_printer p = drm_debug_printer("i915");
+
+		__print_intel_runtime_pm_wakeref(&p, debug);
+	}
+
+	kfree(debug->owners);
+}
+
+static noinline void
+__intel_wakeref_dec_and_check_tracking(struct intel_runtime_pm *rpm)
+{
+	struct intel_runtime_pm_debug dbg = {};
+	unsigned long flags;
+
+	if (!atomic_dec_and_lock_irqsave(&rpm->wakeref_count,
+					 &rpm->debug.lock,
+					 flags))
+		return;
+
+	__untrack_all_wakerefs(&rpm->debug, &dbg);
+	spin_unlock_irqrestore(&rpm->debug.lock, flags);
+
+	dump_and_free_wakeref_tracking(&dbg);
+}
+
+static noinline void
+untrack_all_intel_runtime_pm_wakerefs(struct intel_runtime_pm *rpm)
+{
+	struct intel_runtime_pm_debug dbg = {};
+	unsigned long flags;
+
+	spin_lock_irqsave(&rpm->debug.lock, flags);
+	__untrack_all_wakerefs(&rpm->debug, &dbg);
+	spin_unlock_irqrestore(&rpm->debug.lock, flags);
+
+	dump_and_free_wakeref_tracking(&dbg);
+}
+
+void print_intel_runtime_pm_wakeref(struct intel_runtime_pm *rpm,
+				    struct drm_printer *p)
+{
+	struct intel_runtime_pm_debug dbg = {};
+
+	do {
+		unsigned long alloc = dbg.count;
+		depot_stack_handle_t *s;
+
+		spin_lock_irq(&rpm->debug.lock);
+		dbg.count = rpm->debug.count;
+		if (dbg.count <= alloc) {
+			memcpy(dbg.owners,
+			       rpm->debug.owners,
+			       dbg.count * sizeof(*s));
+		}
+		dbg.last_acquire = rpm->debug.last_acquire;
+		dbg.last_release = rpm->debug.last_release;
+		spin_unlock_irq(&rpm->debug.lock);
+		if (dbg.count <= alloc)
+			break;
+
+		s = krealloc(dbg.owners,
+			     dbg.count * sizeof(*s),
+			     GFP_NOWAIT | __GFP_NOWARN);
+		if (!s)
+			goto out;
+
+		dbg.owners = s;
+	} while (1);
+
+	__print_intel_runtime_pm_wakeref(p, &dbg);
+
+out:
+	kfree(dbg.owners);
+}
+
+#else
+
+static void init_intel_runtime_pm_wakeref(struct intel_runtime_pm *rpm)
+{
+}
+
+static depot_stack_handle_t
+track_intel_runtime_pm_wakeref(struct intel_runtime_pm *rpm)
+{
+	return -1;
+}
+
+static void untrack_intel_runtime_pm_wakeref(struct intel_runtime_pm *rpm,
+					     intel_wakeref_t wref)
+{
+}
+
+static void
+__intel_wakeref_dec_and_check_tracking(struct intel_runtime_pm *rpm)
+{
+	atomic_dec(&rpm->wakeref_count);
+}
+
+static void
+untrack_all_intel_runtime_pm_wakerefs(struct intel_runtime_pm *rpm)
+{
+}
+
+#endif
+
+static void
+intel_runtime_pm_acquire(struct intel_runtime_pm *rpm, bool wakelock)
+{
+	if (wakelock) {
+		atomic_add(1 + INTEL_RPM_WAKELOCK_BIAS, &rpm->wakeref_count);
+		assert_rpm_wakelock_held(rpm);
+	} else {
+		atomic_inc(&rpm->wakeref_count);
+		assert_rpm_raw_wakeref_held(rpm);
+	}
+}
+
+static void
+intel_runtime_pm_release(struct intel_runtime_pm *rpm, int wakelock)
+{
+	if (wakelock) {
+		assert_rpm_wakelock_held(rpm);
+		atomic_sub(INTEL_RPM_WAKELOCK_BIAS, &rpm->wakeref_count);
+	} else {
+		assert_rpm_raw_wakeref_held(rpm);
+	}
+
+	__intel_wakeref_dec_and_check_tracking(rpm);
+}
+
+static intel_wakeref_t __intel_runtime_pm_get(struct intel_runtime_pm *rpm,
+					      bool wakelock)
+{
+	struct drm_i915_private *i915 = container_of(rpm,
+						     struct drm_i915_private,
+						     runtime_pm);
+	int ret;
+
+	ret = pm_runtime_get_sync(rpm->kdev);
+	drm_WARN_ONCE(&i915->drm, ret < 0,
+		      "pm_runtime_get_sync() failed: %d\n", ret);
+
+	intel_runtime_pm_acquire(rpm, wakelock);
+
+	return track_intel_runtime_pm_wakeref(rpm);
+}
+
+/**
+ * intel_runtime_pm_get_raw - grab a raw runtime pm reference
+ * @rpm: the intel_runtime_pm structure
+ *
+ * This is the unlocked version of intel_display_power_is_enabled() and should
+ * only be used from error capture and recovery code where deadlocks are
+ * possible.
+ * This function grabs a device-level runtime pm reference (mostly used for
+ * asynchronous PM management from display code) and ensures that it is powered
+ * up. Raw references are not considered during wakelock assert checks.
+ *
+ * Any runtime pm reference obtained by this function must have a symmetric
+ * call to intel_runtime_pm_put_raw() to release the reference again.
+ *
+ * Returns: the wakeref cookie to pass to intel_runtime_pm_put_raw(), evaluates
+ * as True if the wakeref was acquired, or False otherwise.
+ */
+intel_wakeref_t intel_runtime_pm_get_raw(struct intel_runtime_pm *rpm)
+{
+	return __intel_runtime_pm_get(rpm, false);
+}
+
+/**
+ * intel_runtime_pm_get - grab a runtime pm reference
+ * @rpm: the intel_runtime_pm structure
+ *
+ * This function grabs a device-level runtime pm reference (mostly used for GEM
+ * code to ensure the GTT or GT is on) and ensures that it is powered up.
+ *
+ * Any runtime pm reference obtained by this function must have a symmetric
+ * call to intel_runtime_pm_put() to release the reference again.
+ *
+ * Returns: the wakeref cookie to pass to intel_runtime_pm_put()
+ */
+intel_wakeref_t intel_runtime_pm_get(struct intel_runtime_pm *rpm)
+{
+	return __intel_runtime_pm_get(rpm, true);
+}
+
+/**
+ * __intel_runtime_pm_get_if_active - grab a runtime pm reference if device is active
+ * @rpm: the intel_runtime_pm structure
+ * @ignore_usecount: get a ref even if dev->power.usage_count is 0
+ *
+ * This function grabs a device-level runtime pm reference if the device is
+ * already active and ensures that it is powered up. It is illegal to try
+ * and access the HW should intel_runtime_pm_get_if_active() report failure.
+ *
+ * If @ignore_usecount is true, a reference will be acquired even if there is no
+ * user requiring the device to be powered up (dev->power.usage_count == 0).
+ * If the function returns false in this case then it's guaranteed that the
+ * device's runtime suspend hook has been called already or that it will be
+ * called (and hence it's also guaranteed that the device's runtime resume
+ * hook will be called eventually).
+ *
+ * Any runtime pm reference obtained by this function must have a symmetric
+ * call to intel_runtime_pm_put() to release the reference again.
+ *
+ * Returns: the wakeref cookie to pass to intel_runtime_pm_put(), evaluates
+ * as True if the wakeref was acquired, or False otherwise.
+ */
+static intel_wakeref_t __intel_runtime_pm_get_if_active(struct intel_runtime_pm *rpm,
+							bool ignore_usecount)
+{
+	if (IS_ENABLED(CONFIG_PM)) {
+		/*
+		 * In cases runtime PM is disabled by the RPM core and we get
+		 * an -EINVAL return value we are not supposed to call this
+		 * function, since the power state is undefined. This applies
+		 * atm to the late/early system suspend/resume handlers.
+		 */
+		if (pm_runtime_get_if_active(rpm->kdev, ignore_usecount) <= 0)
+			return 0;
+	}
+
+	intel_runtime_pm_acquire(rpm, true);
+
+	return track_intel_runtime_pm_wakeref(rpm);
+}
+
+intel_wakeref_t intel_runtime_pm_get_if_in_use(struct intel_runtime_pm *rpm)
+{
+	return __intel_runtime_pm_get_if_active(rpm, false);
+}
+
+intel_wakeref_t intel_runtime_pm_get_if_active(struct intel_runtime_pm *rpm)
+{
+	return __intel_runtime_pm_get_if_active(rpm, true);
+}
+
+/**
+ * intel_runtime_pm_get_noresume - grab a runtime pm reference
+ * @rpm: the intel_runtime_pm structure
+ *
+ * This function grabs a device-level runtime pm reference (mostly used for GEM
+ * code to ensure the GTT or GT is on).
+ *
+ * It will _not_ power up the device but instead only check that it's powered
+ * on.  Therefore it is only valid to call this functions from contexts where
+ * the device is known to be powered up and where trying to power it up would
+ * result in hilarity and deadlocks. That pretty much means only the system
+ * suspend/resume code where this is used to grab runtime pm references for
+ * delayed setup down in work items.
+ *
+ * Any runtime pm reference obtained by this function must have a symmetric
+ * call to intel_runtime_pm_put() to release the reference again.
+ *
+ * Returns: the wakeref cookie to pass to intel_runtime_pm_put()
+ */
+intel_wakeref_t intel_runtime_pm_get_noresume(struct intel_runtime_pm *rpm)
+{
+	assert_rpm_wakelock_held(rpm);
+	pm_runtime_get_noresume(rpm->kdev);
+
+	intel_runtime_pm_acquire(rpm, true);
+
+	return track_intel_runtime_pm_wakeref(rpm);
+}
+
+static void __intel_runtime_pm_put(struct intel_runtime_pm *rpm,
+				   intel_wakeref_t wref,
+				   bool wakelock)
+{
+	struct device *kdev = rpm->kdev;
+
+	untrack_intel_runtime_pm_wakeref(rpm, wref);
+
+	intel_runtime_pm_release(rpm, wakelock);
+
+	pm_runtime_mark_last_busy(kdev);
+	pm_runtime_put_autosuspend(kdev);
+}
+
+/**
+ * intel_runtime_pm_put_raw - release a raw runtime pm reference
+ * @rpm: the intel_runtime_pm structure
+ * @wref: wakeref acquired for the reference that is being released
+ *
+ * This function drops the device-level runtime pm reference obtained by
+ * intel_runtime_pm_get_raw() and might power down the corresponding
+ * hardware block right away if this is the last reference.
+ */
+void
+intel_runtime_pm_put_raw(struct intel_runtime_pm *rpm, intel_wakeref_t wref)
+{
+	__intel_runtime_pm_put(rpm, wref, false);
+}
+
+/**
+ * intel_runtime_pm_put_unchecked - release an unchecked runtime pm reference
+ * @rpm: the intel_runtime_pm structure
+ *
+ * This function drops the device-level runtime pm reference obtained by
+ * intel_runtime_pm_get() and might power down the corresponding
+ * hardware block right away if this is the last reference.
+ *
+ * This function exists only for historical reasons and should be avoided in
+ * new code, as the correctness of its use cannot be checked. Always use
+ * intel_runtime_pm_put() instead.
+ */
+void intel_runtime_pm_put_unchecked(struct intel_runtime_pm *rpm)
+{
+	__intel_runtime_pm_put(rpm, -1, true);
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+/**
+ * intel_runtime_pm_put - release a runtime pm reference
+ * @rpm: the intel_runtime_pm structure
+ * @wref: wakeref acquired for the reference that is being released
+ *
+ * This function drops the device-level runtime pm reference obtained by
+ * intel_runtime_pm_get() and might power down the corresponding
+ * hardware block right away if this is the last reference.
+ */
+void intel_runtime_pm_put(struct intel_runtime_pm *rpm, intel_wakeref_t wref)
+{
+	__intel_runtime_pm_put(rpm, wref, true);
+}
+#endif
+
+/**
+ * intel_runtime_pm_enable - enable runtime pm
+ * @rpm: the intel_runtime_pm structure
+ *
+ * This function enables runtime pm at the end of the driver load sequence.
+ *
+ * Note that this function does currently not enable runtime pm for the
+ * subordinate display power domains. That is done by
+ * intel_power_domains_enable().
+ */
+void intel_runtime_pm_enable(struct intel_runtime_pm *rpm)
+{
+	struct drm_i915_private *i915 = container_of(rpm,
+						     struct drm_i915_private,
+						     runtime_pm);
+	struct device *kdev = rpm->kdev;
+
+	/*
+	 * Disable the system suspend direct complete optimization, which can
+	 * leave the device suspended skipping the driver's suspend handlers
+	 * if the device was already runtime suspended. This is needed due to
+	 * the difference in our runtime and system suspend sequence and
+	 * becaue the HDA driver may require us to enable the audio power
+	 * domain during system suspend.
+	 */
+	dev_pm_set_driver_flags(kdev, DPM_FLAG_NO_DIRECT_COMPLETE);
+
+	pm_runtime_set_autosuspend_delay(kdev, 10000); /* 10s */
+	pm_runtime_mark_last_busy(kdev);
+
+	/*
+	 * Take a permanent reference to disable the RPM functionality and drop
+	 * it only when unloading the driver. Use the low level get/put helpers,
+	 * so the driver's own RPM reference tracking asserts also work on
+	 * platforms without RPM support.
+	 */
+	if (!rpm->available) {
+		int ret;
+
+		pm_runtime_dont_use_autosuspend(kdev);
+		ret = pm_runtime_get_sync(kdev);
+		drm_WARN(&i915->drm, ret < 0,
+			 "pm_runtime_get_sync() failed: %d\n", ret);
+	} else {
+		pm_runtime_use_autosuspend(kdev);
+	}
+
+	/*
+	 *  FIXME: Temp hammer to keep autosupend disable on lmem supported platforms.
+	 *  As per PCIe specs 5.3.1.4.1, all iomem read write request over a PCIe
+	 *  function will be unsupported in case PCIe endpoint function is in D3.
+	 *  Let's keep i915 autosuspend control 'on' till we fix all known issue
+	 *  with lmem access in D3.
+	 */
+	if (!IS_DGFX(i915))
+		pm_runtime_allow(kdev);
+
+	/*
+	 * The core calls the driver load handler with an RPM reference held.
+	 * We drop that here and will reacquire it during unloading in
+	 * intel_power_domains_fini().
+	 */
+	pm_runtime_put_autosuspend(kdev);
+}
+
+void intel_runtime_pm_disable(struct intel_runtime_pm *rpm)
+{
+	struct drm_i915_private *i915 = container_of(rpm,
+						     struct drm_i915_private,
+						     runtime_pm);
+	struct device *kdev = rpm->kdev;
+
+	/* Transfer rpm ownership back to core */
+	drm_WARN(&i915->drm, pm_runtime_get_sync(kdev) < 0,
+		 "Failed to pass rpm ownership back to core\n");
+
+	pm_runtime_dont_use_autosuspend(kdev);
+
+	if (!rpm->available)
+		pm_runtime_put(kdev);
+}
+
+void intel_runtime_pm_driver_release(struct intel_runtime_pm *rpm)
+{
+	struct drm_i915_private *i915 = container_of(rpm,
+						     struct drm_i915_private,
+						     runtime_pm);
+	int count = atomic_read(&rpm->wakeref_count);
+
+	intel_wakeref_auto_fini(&rpm->userfault_wakeref);
+
+	drm_WARN(&i915->drm, count,
+		 "i915 raw-wakerefs=%d wakelocks=%d on cleanup\n",
+		 intel_rpm_raw_wakeref_count(count),
+		 intel_rpm_wakelock_count(count));
+
+	untrack_all_intel_runtime_pm_wakerefs(rpm);
+}
+
+void intel_runtime_pm_init_early(struct intel_runtime_pm *rpm)
+{
+	struct drm_i915_private *i915 =
+			container_of(rpm, struct drm_i915_private, runtime_pm);
+	struct pci_dev *pdev = to_pci_dev(i915->drm.dev);
+	struct device *kdev = &pdev->dev;
+
+	rpm->kdev = kdev;
+	rpm->available = HAS_RUNTIME_PM(i915);
+
+	init_intel_runtime_pm_wakeref(rpm);
+	INIT_LIST_HEAD(&rpm->lmem_userfault_list);
+	spin_lock_init(&rpm->lmem_userfault_lock);
+	intel_wakeref_auto_init(&rpm->userfault_wakeref, rpm);
+}
diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.h b/drivers/gpu/drm/i915/intel_runtime_pm.h
new file mode 100644
index 000000000..98b8b28ba
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_runtime_pm.h
@@ -0,0 +1,239 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __INTEL_RUNTIME_PM_H__
+#define __INTEL_RUNTIME_PM_H__
+
+#include <linux/types.h>
+
+#include "intel_wakeref.h"
+
+#include "i915_utils.h"
+
+struct device;
+struct drm_i915_private;
+struct drm_printer;
+
+enum i915_drm_suspend_mode {
+	I915_DRM_SUSPEND_IDLE,
+	I915_DRM_SUSPEND_MEM,
+	I915_DRM_SUSPEND_HIBERNATE,
+};
+
+/*
+ * This struct helps tracking the state needed for runtime PM, which puts the
+ * device in PCI D3 state. Notice that when this happens, nothing on the
+ * graphics device works, even register access, so we don't get interrupts nor
+ * anything else.
+ *
+ * Every piece of our code that needs to actually touch the hardware needs to
+ * either call intel_runtime_pm_get or call intel_display_power_get with the
+ * appropriate power domain.
+ *
+ * Our driver uses the autosuspend delay feature, which means we'll only really
+ * suspend if we stay with zero refcount for a certain amount of time. The
+ * default value is currently very conservative (see intel_runtime_pm_enable), but
+ * it can be changed with the standard runtime PM files from sysfs.
+ *
+ * The irqs_disabled variable becomes true exactly after we disable the IRQs and
+ * goes back to false exactly before we reenable the IRQs. We use this variable
+ * to check if someone is trying to enable/disable IRQs while they're supposed
+ * to be disabled. This shouldn't happen and we'll print some error messages in
+ * case it happens.
+ *
+ * For more, read the Documentation/power/runtime_pm.rst.
+ */
+struct intel_runtime_pm {
+	atomic_t wakeref_count;
+	struct device *kdev; /* points to i915->drm.dev */
+	bool available;
+	bool suspended;
+	bool irqs_enabled;
+	bool no_wakeref_tracking;
+
+	/*
+	 *  Protects access to lmem usefault list.
+	 *  It is required, if we are outside of the runtime suspend path,
+	 *  access to @lmem_userfault_list requires always first grabbing the
+	 *  runtime pm, to ensure we can't race against runtime suspend.
+	 *  Once we have that we also need to grab @lmem_userfault_lock,
+	 *  at which point we have exclusive access.
+	 *  The runtime suspend path is special since it doesn't really hold any locks,
+	 *  but instead has exclusive access by virtue of all other accesses requiring
+	 *  holding the runtime pm wakeref.
+	 */
+	spinlock_t lmem_userfault_lock;
+
+	/*
+	 *  Keep list of userfaulted gem obj, which require to release their
+	 *  mmap mappings at runtime suspend path.
+	 */
+	struct list_head lmem_userfault_list;
+
+	/* Manual runtime pm autosuspend delay for user GGTT/lmem mmaps */
+	struct intel_wakeref_auto userfault_wakeref;
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+	/*
+	 * To aide detection of wakeref leaks and general misuse, we
+	 * track all wakeref holders. With manual markup (i.e. returning
+	 * a cookie to each rpm_get caller which they then supply to their
+	 * paired rpm_put) we can remove corresponding pairs of and keep
+	 * the array trimmed to active wakerefs.
+	 */
+	struct intel_runtime_pm_debug {
+		spinlock_t lock;
+
+		depot_stack_handle_t last_acquire;
+		depot_stack_handle_t last_release;
+
+		depot_stack_handle_t *owners;
+		unsigned long count;
+	} debug;
+#endif
+};
+
+#define BITS_PER_WAKEREF	\
+	BITS_PER_TYPE(struct_member(struct intel_runtime_pm, wakeref_count))
+#define INTEL_RPM_WAKELOCK_SHIFT	(BITS_PER_WAKEREF / 2)
+#define INTEL_RPM_WAKELOCK_BIAS		(1 << INTEL_RPM_WAKELOCK_SHIFT)
+#define INTEL_RPM_RAW_WAKEREF_MASK	(INTEL_RPM_WAKELOCK_BIAS - 1)
+
+static inline int
+intel_rpm_raw_wakeref_count(int wakeref_count)
+{
+	return wakeref_count & INTEL_RPM_RAW_WAKEREF_MASK;
+}
+
+static inline int
+intel_rpm_wakelock_count(int wakeref_count)
+{
+	return wakeref_count >> INTEL_RPM_WAKELOCK_SHIFT;
+}
+
+static inline void
+assert_rpm_device_not_suspended(struct intel_runtime_pm *rpm)
+{
+	WARN_ONCE(rpm->suspended,
+		  "Device suspended during HW access\n");
+}
+
+static inline void
+__assert_rpm_raw_wakeref_held(struct intel_runtime_pm *rpm, int wakeref_count)
+{
+	assert_rpm_device_not_suspended(rpm);
+	WARN_ONCE(!intel_rpm_raw_wakeref_count(wakeref_count),
+		  "RPM raw-wakeref not held\n");
+}
+
+static inline void
+__assert_rpm_wakelock_held(struct intel_runtime_pm *rpm, int wakeref_count)
+{
+	__assert_rpm_raw_wakeref_held(rpm, wakeref_count);
+	WARN_ONCE(!intel_rpm_wakelock_count(wakeref_count),
+		  "RPM wakelock ref not held during HW access\n");
+}
+
+static inline void
+assert_rpm_raw_wakeref_held(struct intel_runtime_pm *rpm)
+{
+	__assert_rpm_raw_wakeref_held(rpm, atomic_read(&rpm->wakeref_count));
+}
+
+static inline void
+assert_rpm_wakelock_held(struct intel_runtime_pm *rpm)
+{
+	__assert_rpm_wakelock_held(rpm, atomic_read(&rpm->wakeref_count));
+}
+
+/**
+ * disable_rpm_wakeref_asserts - disable the RPM assert checks
+ * @rpm: the intel_runtime_pm structure
+ *
+ * This function disable asserts that check if we hold an RPM wakelock
+ * reference, while keeping the device-not-suspended checks still enabled.
+ * It's meant to be used only in special circumstances where our rule about
+ * the wakelock refcount wrt. the device power state doesn't hold. According
+ * to this rule at any point where we access the HW or want to keep the HW in
+ * an active state we must hold an RPM wakelock reference acquired via one of
+ * the intel_runtime_pm_get() helpers. Currently there are a few special spots
+ * where this rule doesn't hold: the IRQ and suspend/resume handlers, the
+ * forcewake release timer, and the GPU RPS and hangcheck works. All other
+ * users should avoid using this function.
+ *
+ * Any calls to this function must have a symmetric call to
+ * enable_rpm_wakeref_asserts().
+ */
+static inline void
+disable_rpm_wakeref_asserts(struct intel_runtime_pm *rpm)
+{
+	atomic_add(INTEL_RPM_WAKELOCK_BIAS + 1,
+		   &rpm->wakeref_count);
+}
+
+/**
+ * enable_rpm_wakeref_asserts - re-enable the RPM assert checks
+ * @rpm: the intel_runtime_pm structure
+ *
+ * This function re-enables the RPM assert checks after disabling them with
+ * disable_rpm_wakeref_asserts. It's meant to be used only in special
+ * circumstances otherwise its use should be avoided.
+ *
+ * Any calls to this function must have a symmetric call to
+ * disable_rpm_wakeref_asserts().
+ */
+static inline void
+enable_rpm_wakeref_asserts(struct intel_runtime_pm *rpm)
+{
+	atomic_sub(INTEL_RPM_WAKELOCK_BIAS + 1,
+		   &rpm->wakeref_count);
+}
+
+void intel_runtime_pm_init_early(struct intel_runtime_pm *rpm);
+void intel_runtime_pm_enable(struct intel_runtime_pm *rpm);
+void intel_runtime_pm_disable(struct intel_runtime_pm *rpm);
+void intel_runtime_pm_driver_release(struct intel_runtime_pm *rpm);
+
+intel_wakeref_t intel_runtime_pm_get(struct intel_runtime_pm *rpm);
+intel_wakeref_t intel_runtime_pm_get_if_in_use(struct intel_runtime_pm *rpm);
+intel_wakeref_t intel_runtime_pm_get_if_active(struct intel_runtime_pm *rpm);
+intel_wakeref_t intel_runtime_pm_get_noresume(struct intel_runtime_pm *rpm);
+intel_wakeref_t intel_runtime_pm_get_raw(struct intel_runtime_pm *rpm);
+
+#define with_intel_runtime_pm(rpm, wf) \
+	for ((wf) = intel_runtime_pm_get(rpm); (wf); \
+	     intel_runtime_pm_put((rpm), (wf)), (wf) = 0)
+
+#define with_intel_runtime_pm_if_in_use(rpm, wf) \
+	for ((wf) = intel_runtime_pm_get_if_in_use(rpm); (wf); \
+	     intel_runtime_pm_put((rpm), (wf)), (wf) = 0)
+
+#define with_intel_runtime_pm_if_active(rpm, wf) \
+	for ((wf) = intel_runtime_pm_get_if_active(rpm); (wf); \
+	     intel_runtime_pm_put((rpm), (wf)), (wf) = 0)
+
+void intel_runtime_pm_put_unchecked(struct intel_runtime_pm *rpm);
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+void intel_runtime_pm_put(struct intel_runtime_pm *rpm, intel_wakeref_t wref);
+#else
+static inline void
+intel_runtime_pm_put(struct intel_runtime_pm *rpm, intel_wakeref_t wref)
+{
+	intel_runtime_pm_put_unchecked(rpm);
+}
+#endif
+void intel_runtime_pm_put_raw(struct intel_runtime_pm *rpm, intel_wakeref_t wref);
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)
+void print_intel_runtime_pm_wakeref(struct intel_runtime_pm *rpm,
+				    struct drm_printer *p);
+#else
+static inline void print_intel_runtime_pm_wakeref(struct intel_runtime_pm *rpm,
+						  struct drm_printer *p)
+{
+}
+#endif
+
+#endif /* __INTEL_RUNTIME_PM_H__ */
diff --git a/drivers/gpu/drm/i915/intel_sbi.c b/drivers/gpu/drm/i915/intel_sbi.c
new file mode 100644
index 000000000..5c6e517c7
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_sbi.c
@@ -0,0 +1,74 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2013-2021 Intel Corporation
+ *
+ * LPT/WPT IOSF sideband.
+ */
+
+#include "i915_drv.h"
+#include "intel_sbi.h"
+#include "i915_reg.h"
+
+/* SBI access */
+static int intel_sbi_rw(struct drm_i915_private *i915, u16 reg,
+			enum intel_sbi_destination destination,
+			u32 *val, bool is_read)
+{
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 cmd;
+
+	lockdep_assert_held(&i915->sb_lock);
+
+	if (intel_wait_for_register_fw(uncore,
+				       SBI_CTL_STAT, SBI_BUSY, 0,
+				       100)) {
+		drm_err(&i915->drm,
+			"timeout waiting for SBI to become ready\n");
+		return -EBUSY;
+	}
+
+	intel_uncore_write_fw(uncore, SBI_ADDR, (u32)reg << 16);
+	intel_uncore_write_fw(uncore, SBI_DATA, is_read ? 0 : *val);
+
+	if (destination == SBI_ICLK)
+		cmd = SBI_CTL_DEST_ICLK | SBI_CTL_OP_CRRD;
+	else
+		cmd = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IORD;
+	if (!is_read)
+		cmd |= BIT(8);
+	intel_uncore_write_fw(uncore, SBI_CTL_STAT, cmd | SBI_BUSY);
+
+	if (__intel_wait_for_register_fw(uncore,
+					 SBI_CTL_STAT, SBI_BUSY, 0,
+					 100, 100, &cmd)) {
+		drm_err(&i915->drm,
+			"timeout waiting for SBI to complete read\n");
+		return -ETIMEDOUT;
+	}
+
+	if (cmd & SBI_RESPONSE_FAIL) {
+		drm_err(&i915->drm, "error during SBI read of reg %x\n", reg);
+		return -ENXIO;
+	}
+
+	if (is_read)
+		*val = intel_uncore_read_fw(uncore, SBI_DATA);
+
+	return 0;
+}
+
+u32 intel_sbi_read(struct drm_i915_private *i915, u16 reg,
+		   enum intel_sbi_destination destination)
+{
+	u32 result = 0;
+
+	intel_sbi_rw(i915, reg, destination, &result, true);
+
+	return result;
+}
+
+void intel_sbi_write(struct drm_i915_private *i915, u16 reg, u32 value,
+		     enum intel_sbi_destination destination)
+{
+	intel_sbi_rw(i915, reg, destination, &value, false);
+}
diff --git a/drivers/gpu/drm/i915/intel_sbi.h b/drivers/gpu/drm/i915/intel_sbi.h
new file mode 100644
index 000000000..f5a862210
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_sbi.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2013-2021 Intel Corporation
+ */
+
+#ifndef _INTEL_SBI_H_
+#define _INTEL_SBI_H_
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+
+enum intel_sbi_destination {
+	SBI_ICLK,
+	SBI_MPHY,
+};
+
+u32 intel_sbi_read(struct drm_i915_private *i915, u16 reg,
+		   enum intel_sbi_destination destination);
+void intel_sbi_write(struct drm_i915_private *i915, u16 reg, u32 value,
+		     enum intel_sbi_destination destination);
+
+#endif /* _INTEL_SBI_H_ */
diff --git a/drivers/gpu/drm/i915/intel_step.c b/drivers/gpu/drm/i915/intel_step.c
new file mode 100644
index 000000000..42b3133d8
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_step.c
@@ -0,0 +1,318 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020,2021 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "intel_step.h"
+
+/*
+ * Some platforms have unusual ways of mapping PCI revision ID to GT/display
+ * steppings.  E.g., in some cases a higher PCI revision may translate to a
+ * lower stepping of the GT and/or display IP.  This file provides lookup
+ * tables to map the PCI revision into a standard set of stepping values that
+ * can be compared numerically.
+ *
+ * Also note that some revisions/steppings may have been set aside as
+ * placeholders but never materialized in real hardware; in those cases there
+ * may be jumps in the revision IDs or stepping values in the tables below.
+ */
+
+/*
+ * Some platforms always have the same stepping value for GT and display;
+ * use a macro to define these to make it easier to identify the platforms
+ * where the two steppings can deviate.
+ */
+#define COMMON_STEP(x)  .graphics_step = STEP_##x, .display_step = STEP_##x, .media_step = STEP_##x
+#define COMMON_GT_MEDIA_STEP(x)  .graphics_step = STEP_##x, .media_step = STEP_##x
+
+static const struct intel_step_info skl_revids[] = {
+	[0x6] = { COMMON_STEP(G0) },
+	[0x7] = { COMMON_STEP(H0) },
+	[0x9] = { COMMON_STEP(J0) },
+	[0xA] = { COMMON_STEP(I1) },
+};
+
+static const struct intel_step_info kbl_revids[] = {
+	[1] = { COMMON_GT_MEDIA_STEP(B0), .display_step = STEP_B0 },
+	[2] = { COMMON_GT_MEDIA_STEP(C0), .display_step = STEP_B0 },
+	[3] = { COMMON_GT_MEDIA_STEP(D0), .display_step = STEP_B0 },
+	[4] = { COMMON_GT_MEDIA_STEP(F0), .display_step = STEP_C0 },
+	[5] = { COMMON_GT_MEDIA_STEP(C0), .display_step = STEP_B1 },
+	[6] = { COMMON_GT_MEDIA_STEP(D1), .display_step = STEP_B1 },
+	[7] = { COMMON_GT_MEDIA_STEP(G0), .display_step = STEP_C0 },
+};
+
+static const struct intel_step_info bxt_revids[] = {
+	[0xA] = { COMMON_STEP(C0) },
+	[0xB] = { COMMON_STEP(C0) },
+	[0xC] = { COMMON_STEP(D0) },
+	[0xD] = { COMMON_STEP(E0) },
+};
+
+static const struct intel_step_info glk_revids[] = {
+	[3] = { COMMON_STEP(B0) },
+};
+
+static const struct intel_step_info icl_revids[] = {
+	[7] = { COMMON_STEP(D0) },
+};
+
+static const struct intel_step_info jsl_ehl_revids[] = {
+	[0] = { COMMON_STEP(A0) },
+	[1] = { COMMON_STEP(B0) },
+};
+
+static const struct intel_step_info tgl_uy_revids[] = {
+	[0] = { COMMON_GT_MEDIA_STEP(A0), .display_step = STEP_A0 },
+	[1] = { COMMON_GT_MEDIA_STEP(B0), .display_step = STEP_C0 },
+	[2] = { COMMON_GT_MEDIA_STEP(B1), .display_step = STEP_C0 },
+	[3] = { COMMON_GT_MEDIA_STEP(C0), .display_step = STEP_D0 },
+};
+
+/* Same GT stepping between tgl_uy_revids and tgl_revids don't mean the same HW */
+static const struct intel_step_info tgl_revids[] = {
+	[0] = { COMMON_GT_MEDIA_STEP(A0), .display_step = STEP_B0 },
+	[1] = { COMMON_GT_MEDIA_STEP(B0), .display_step = STEP_D0 },
+};
+
+static const struct intel_step_info rkl_revids[] = {
+	[0] = { COMMON_STEP(A0) },
+	[1] = { COMMON_STEP(B0) },
+	[4] = { COMMON_STEP(C0) },
+};
+
+static const struct intel_step_info dg1_revids[] = {
+	[0] = { COMMON_STEP(A0) },
+	[1] = { COMMON_STEP(B0) },
+};
+
+static const struct intel_step_info adls_revids[] = {
+	[0x0] = { COMMON_GT_MEDIA_STEP(A0), .display_step = STEP_A0 },
+	[0x1] = { COMMON_GT_MEDIA_STEP(A0), .display_step = STEP_A2 },
+	[0x4] = { COMMON_GT_MEDIA_STEP(B0), .display_step = STEP_B0 },
+	[0x8] = { COMMON_GT_MEDIA_STEP(C0), .display_step = STEP_B0 },
+	[0xC] = { COMMON_GT_MEDIA_STEP(D0), .display_step = STEP_C0 },
+};
+
+static const struct intel_step_info adlp_revids[] = {
+	[0x0] = { COMMON_GT_MEDIA_STEP(A0), .display_step = STEP_A0 },
+	[0x4] = { COMMON_GT_MEDIA_STEP(B0), .display_step = STEP_B0 },
+	[0x8] = { COMMON_GT_MEDIA_STEP(C0), .display_step = STEP_C0 },
+	[0xC] = { COMMON_GT_MEDIA_STEP(C0), .display_step = STEP_D0 },
+};
+
+static const struct intel_step_info xehpsdv_revids[] = {
+	[0x0] = { COMMON_GT_MEDIA_STEP(A0) },
+	[0x1] = { COMMON_GT_MEDIA_STEP(A1) },
+	[0x4] = { COMMON_GT_MEDIA_STEP(B0) },
+	[0x8] = { COMMON_GT_MEDIA_STEP(C0) },
+};
+
+static const struct intel_step_info dg2_g10_revid_step_tbl[] = {
+	[0x0] = { COMMON_GT_MEDIA_STEP(A0), .display_step = STEP_A0 },
+	[0x1] = { COMMON_GT_MEDIA_STEP(A1), .display_step = STEP_A0 },
+	[0x4] = { COMMON_GT_MEDIA_STEP(B0), .display_step = STEP_B0 },
+	[0x8] = { COMMON_GT_MEDIA_STEP(C0), .display_step = STEP_C0 },
+};
+
+static const struct intel_step_info dg2_g11_revid_step_tbl[] = {
+	[0x0] = { COMMON_GT_MEDIA_STEP(A0), .display_step = STEP_B0 },
+	[0x4] = { COMMON_GT_MEDIA_STEP(B0), .display_step = STEP_C0 },
+	[0x5] = { COMMON_GT_MEDIA_STEP(B1), .display_step = STEP_C0 },
+};
+
+static const struct intel_step_info dg2_g12_revid_step_tbl[] = {
+	[0x0] = { COMMON_GT_MEDIA_STEP(A0), .display_step = STEP_C0 },
+};
+
+static const struct intel_step_info adls_rpls_revids[] = {
+	[0x4] = { COMMON_GT_MEDIA_STEP(D0), .display_step = STEP_D0 },
+	[0xC] = { COMMON_GT_MEDIA_STEP(D0), .display_step = STEP_C0 },
+};
+
+static const struct intel_step_info adlp_n_revids[] = {
+	[0x0] = { COMMON_GT_MEDIA_STEP(A0), .display_step = STEP_D0 },
+};
+
+static void pvc_step_init(struct drm_i915_private *i915, int pci_revid);
+
+void intel_step_init(struct drm_i915_private *i915)
+{
+	const struct intel_step_info *revids = NULL;
+	int size = 0;
+	int revid = INTEL_REVID(i915);
+	struct intel_step_info step = {};
+
+	if (IS_PONTEVECCHIO(i915)) {
+		pvc_step_init(i915, revid);
+		return;
+	} else if (IS_DG2_G10(i915)) {
+		revids = dg2_g10_revid_step_tbl;
+		size = ARRAY_SIZE(dg2_g10_revid_step_tbl);
+	} else if (IS_DG2_G11(i915)) {
+		revids = dg2_g11_revid_step_tbl;
+		size = ARRAY_SIZE(dg2_g11_revid_step_tbl);
+	} else if (IS_DG2_G12(i915)) {
+		revids = dg2_g12_revid_step_tbl;
+		size = ARRAY_SIZE(dg2_g12_revid_step_tbl);
+	} else if (IS_XEHPSDV(i915)) {
+		revids = xehpsdv_revids;
+		size = ARRAY_SIZE(xehpsdv_revids);
+	} else if (IS_ADLP_N(i915)) {
+		revids = adlp_n_revids;
+		size = ARRAY_SIZE(adlp_n_revids);
+	} else if (IS_ALDERLAKE_P(i915)) {
+		revids = adlp_revids;
+		size = ARRAY_SIZE(adlp_revids);
+	} else if (IS_ADLS_RPLS(i915)) {
+		revids = adls_rpls_revids;
+		size = ARRAY_SIZE(adls_rpls_revids);
+	} else if (IS_ALDERLAKE_S(i915)) {
+		revids = adls_revids;
+		size = ARRAY_SIZE(adls_revids);
+	} else if (IS_DG1(i915)) {
+		revids = dg1_revids;
+		size = ARRAY_SIZE(dg1_revids);
+	} else if (IS_ROCKETLAKE(i915)) {
+		revids = rkl_revids;
+		size = ARRAY_SIZE(rkl_revids);
+	} else if (IS_TGL_UY(i915)) {
+		revids = tgl_uy_revids;
+		size = ARRAY_SIZE(tgl_uy_revids);
+	} else if (IS_TIGERLAKE(i915)) {
+		revids = tgl_revids;
+		size = ARRAY_SIZE(tgl_revids);
+	} else if (IS_JSL_EHL(i915)) {
+		revids = jsl_ehl_revids;
+		size = ARRAY_SIZE(jsl_ehl_revids);
+	} else if (IS_ICELAKE(i915)) {
+		revids = icl_revids;
+		size = ARRAY_SIZE(icl_revids);
+	} else if (IS_GEMINILAKE(i915)) {
+		revids = glk_revids;
+		size = ARRAY_SIZE(glk_revids);
+	} else if (IS_BROXTON(i915)) {
+		revids = bxt_revids;
+		size = ARRAY_SIZE(bxt_revids);
+	} else if (IS_KABYLAKE(i915)) {
+		revids = kbl_revids;
+		size = ARRAY_SIZE(kbl_revids);
+	} else if (IS_SKYLAKE(i915)) {
+		revids = skl_revids;
+		size = ARRAY_SIZE(skl_revids);
+	}
+
+	/* Not using the stepping scheme for the platform yet. */
+	if (!revids)
+		return;
+
+	if (revid < size && revids[revid].graphics_step != STEP_NONE) {
+		step = revids[revid];
+	} else {
+		drm_warn(&i915->drm, "Unknown revid 0x%02x\n", revid);
+
+		/*
+		 * If we hit a gap in the revid array, use the information for
+		 * the next revid.
+		 *
+		 * This may be wrong in all sorts of ways, especially if the
+		 * steppings in the array are not monotonically increasing, but
+		 * it's better than defaulting to 0.
+		 */
+		while (revid < size && revids[revid].graphics_step == STEP_NONE)
+			revid++;
+
+		if (revid < size) {
+			drm_dbg(&i915->drm, "Using steppings for revid 0x%02x\n",
+				revid);
+			step = revids[revid];
+		} else {
+			drm_dbg(&i915->drm, "Using future steppings\n");
+			step.graphics_step = STEP_FUTURE;
+			step.display_step = STEP_FUTURE;
+		}
+	}
+
+	if (drm_WARN_ON(&i915->drm, step.graphics_step == STEP_NONE))
+		return;
+
+	RUNTIME_INFO(i915)->step = step;
+}
+
+#define PVC_BD_REVID	GENMASK(5, 3)
+#define PVC_CT_REVID	GENMASK(2, 0)
+
+static const int pvc_bd_subids[] = {
+	[0x0] = STEP_A0,
+	[0x3] = STEP_B0,
+	[0x4] = STEP_B1,
+	[0x5] = STEP_B3,
+};
+
+static const int pvc_ct_subids[] = {
+	[0x3] = STEP_A0,
+	[0x5] = STEP_B0,
+	[0x6] = STEP_B1,
+	[0x7] = STEP_C0,
+};
+
+static int
+pvc_step_lookup(struct drm_i915_private *i915, const char *type,
+		const int *table, int size, int subid)
+{
+	if (subid < size && table[subid] != STEP_NONE)
+		return table[subid];
+
+	drm_warn(&i915->drm, "Unknown %s id 0x%02x\n", type, subid);
+
+	/*
+	 * As on other platforms, try to use the next higher ID if we land on a
+	 * gap in the table.
+	 */
+	while (subid < size && table[subid] == STEP_NONE)
+		subid++;
+
+	if (subid < size) {
+		drm_dbg(&i915->drm, "Using steppings for %s id 0x%02x\n",
+			type, subid);
+		return table[subid];
+	}
+
+	drm_dbg(&i915->drm, "Using future steppings\n");
+	return STEP_FUTURE;
+}
+
+/*
+ * PVC needs special handling since we don't lookup the
+ * revid in a table, but rather specific bitfields within
+ * the revid for various components.
+ */
+static void pvc_step_init(struct drm_i915_private *i915, int pci_revid)
+{
+	int ct_subid, bd_subid;
+
+	bd_subid = FIELD_GET(PVC_BD_REVID, pci_revid);
+	ct_subid = FIELD_GET(PVC_CT_REVID, pci_revid);
+
+	RUNTIME_INFO(i915)->step.basedie_step =
+		pvc_step_lookup(i915, "Base Die", pvc_bd_subids,
+				ARRAY_SIZE(pvc_bd_subids), bd_subid);
+	RUNTIME_INFO(i915)->step.graphics_step =
+		pvc_step_lookup(i915, "Compute Tile", pvc_ct_subids,
+				ARRAY_SIZE(pvc_ct_subids), ct_subid);
+}
+
+#define STEP_NAME_CASE(name)	\
+	case STEP_##name:	\
+		return #name;
+
+const char *intel_step_name(enum intel_step step)
+{
+	switch (step) {
+	STEP_NAME_LIST(STEP_NAME_CASE);
+
+	default:
+		return "**";
+	}
+}
diff --git a/drivers/gpu/drm/i915/intel_step.h b/drivers/gpu/drm/i915/intel_step.h
new file mode 100644
index 000000000..a6b12bfa9
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_step.h
@@ -0,0 +1,56 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020,2021 Intel Corporation
+ */
+
+#ifndef __INTEL_STEP_H__
+#define __INTEL_STEP_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+
+struct intel_step_info {
+	u8 graphics_step;	/* Represents the compute tile on Xe_HPC */
+	u8 display_step;
+	u8 media_step;
+	u8 basedie_step;
+};
+
+#define STEP_ENUM_VAL(name)  STEP_##name,
+
+#define STEP_NAME_LIST(func)		\
+	func(A0)			\
+	func(A1)			\
+	func(A2)			\
+	func(B0)			\
+	func(B1)			\
+	func(B2)			\
+	func(B3)			\
+	func(C0)			\
+	func(C1)			\
+	func(D0)			\
+	func(D1)			\
+	func(E0)			\
+	func(F0)			\
+	func(G0)			\
+	func(H0)			\
+	func(I0)			\
+	func(I1)			\
+	func(J0)
+
+/*
+ * Symbolic steppings that do not match the hardware. These are valid both as gt
+ * and display steppings as symbolic names.
+ */
+enum intel_step {
+	STEP_NONE = 0,
+	STEP_NAME_LIST(STEP_ENUM_VAL)
+	STEP_FUTURE,
+	STEP_FOREVER,
+};
+
+void intel_step_init(struct drm_i915_private *i915);
+const char *intel_step_name(enum intel_step step);
+
+#endif /* __INTEL_STEP_H__ */
diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
new file mode 100644
index 000000000..5cd423c7b
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_uncore.c
@@ -0,0 +1,2678 @@
+/*
+ * Copyright © 2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <drm/drm_managed.h>
+#include <linux/pm_runtime.h>
+
+#include "gt/intel_engine_regs.h"
+#include "gt/intel_gt_regs.h"
+
+#include "i915_drv.h"
+#include "i915_iosf_mbi.h"
+#include "i915_trace.h"
+#include "i915_vgpu.h"
+#include "intel_pm.h"
+
+#define FORCEWAKE_ACK_TIMEOUT_MS 50
+#define GT_FIFO_TIMEOUT_MS	 10
+
+#define __raw_posting_read(...) ((void)__raw_uncore_read32(__VA_ARGS__))
+
+static void
+fw_domains_get(struct intel_uncore *uncore, enum forcewake_domains fw_domains)
+{
+	uncore->fw_get_funcs->force_wake_get(uncore, fw_domains);
+}
+
+void
+intel_uncore_mmio_debug_init_early(struct drm_i915_private *i915)
+{
+	spin_lock_init(&i915->mmio_debug.lock);
+	i915->mmio_debug.unclaimed_mmio_check = 1;
+
+	i915->uncore.debug = &i915->mmio_debug;
+}
+
+static void mmio_debug_suspend(struct intel_uncore *uncore)
+{
+	if (!uncore->debug)
+		return;
+
+	spin_lock(&uncore->debug->lock);
+
+	/* Save and disable mmio debugging for the user bypass */
+	if (!uncore->debug->suspend_count++) {
+		uncore->debug->saved_mmio_check = uncore->debug->unclaimed_mmio_check;
+		uncore->debug->unclaimed_mmio_check = 0;
+	}
+
+	spin_unlock(&uncore->debug->lock);
+}
+
+static bool check_for_unclaimed_mmio(struct intel_uncore *uncore);
+
+static void mmio_debug_resume(struct intel_uncore *uncore)
+{
+	if (!uncore->debug)
+		return;
+
+	spin_lock(&uncore->debug->lock);
+
+	if (!--uncore->debug->suspend_count)
+		uncore->debug->unclaimed_mmio_check = uncore->debug->saved_mmio_check;
+
+	if (check_for_unclaimed_mmio(uncore))
+		drm_info(&uncore->i915->drm,
+			 "Invalid mmio detected during user access\n");
+
+	spin_unlock(&uncore->debug->lock);
+}
+
+static const char * const forcewake_domain_names[] = {
+	"render",
+	"gt",
+	"media",
+	"vdbox0",
+	"vdbox1",
+	"vdbox2",
+	"vdbox3",
+	"vdbox4",
+	"vdbox5",
+	"vdbox6",
+	"vdbox7",
+	"vebox0",
+	"vebox1",
+	"vebox2",
+	"vebox3",
+};
+
+const char *
+intel_uncore_forcewake_domain_to_str(const enum forcewake_domain_id id)
+{
+	BUILD_BUG_ON(ARRAY_SIZE(forcewake_domain_names) != FW_DOMAIN_ID_COUNT);
+
+	if (id >= 0 && id < FW_DOMAIN_ID_COUNT)
+		return forcewake_domain_names[id];
+
+	WARN_ON(id);
+
+	return "unknown";
+}
+
+#define fw_ack(d) readl((d)->reg_ack)
+#define fw_set(d, val) writel(_MASKED_BIT_ENABLE((val)), (d)->reg_set)
+#define fw_clear(d, val) writel(_MASKED_BIT_DISABLE((val)), (d)->reg_set)
+
+static inline void
+fw_domain_reset(const struct intel_uncore_forcewake_domain *d)
+{
+	/*
+	 * We don't really know if the powerwell for the forcewake domain we are
+	 * trying to reset here does exist at this point (engines could be fused
+	 * off in ICL+), so no waiting for acks
+	 */
+	/* WaRsClearFWBitsAtReset */
+	if (GRAPHICS_VER(d->uncore->i915) >= 12)
+		fw_clear(d, 0xefff);
+	else
+		fw_clear(d, 0xffff);
+}
+
+static inline void
+fw_domain_arm_timer(struct intel_uncore_forcewake_domain *d)
+{
+	GEM_BUG_ON(d->uncore->fw_domains_timer & d->mask);
+	d->uncore->fw_domains_timer |= d->mask;
+	d->wake_count++;
+	hrtimer_start_range_ns(&d->timer,
+			       NSEC_PER_MSEC,
+			       NSEC_PER_MSEC,
+			       HRTIMER_MODE_REL);
+}
+
+static inline int
+__wait_for_ack(const struct intel_uncore_forcewake_domain *d,
+	       const u32 ack,
+	       const u32 value)
+{
+	return wait_for_atomic((fw_ack(d) & ack) == value,
+			       FORCEWAKE_ACK_TIMEOUT_MS);
+}
+
+static inline int
+wait_ack_clear(const struct intel_uncore_forcewake_domain *d,
+	       const u32 ack)
+{
+	return __wait_for_ack(d, ack, 0);
+}
+
+static inline int
+wait_ack_set(const struct intel_uncore_forcewake_domain *d,
+	     const u32 ack)
+{
+	return __wait_for_ack(d, ack, ack);
+}
+
+static inline void
+fw_domain_wait_ack_clear(const struct intel_uncore_forcewake_domain *d)
+{
+	if (wait_ack_clear(d, FORCEWAKE_KERNEL)) {
+		DRM_ERROR("%s: timed out waiting for forcewake ack to clear.\n",
+			  intel_uncore_forcewake_domain_to_str(d->id));
+		add_taint_for_CI(d->uncore->i915, TAINT_WARN); /* CI now unreliable */
+	}
+}
+
+enum ack_type {
+	ACK_CLEAR = 0,
+	ACK_SET
+};
+
+static int
+fw_domain_wait_ack_with_fallback(const struct intel_uncore_forcewake_domain *d,
+				 const enum ack_type type)
+{
+	const u32 ack_bit = FORCEWAKE_KERNEL;
+	const u32 value = type == ACK_SET ? ack_bit : 0;
+	unsigned int pass;
+	bool ack_detected;
+
+	/*
+	 * There is a possibility of driver's wake request colliding
+	 * with hardware's own wake requests and that can cause
+	 * hardware to not deliver the driver's ack message.
+	 *
+	 * Use a fallback bit toggle to kick the gpu state machine
+	 * in the hope that the original ack will be delivered along with
+	 * the fallback ack.
+	 *
+	 * This workaround is described in HSDES #1604254524 and it's known as:
+	 * WaRsForcewakeAddDelayForAck:skl,bxt,kbl,glk,cfl,cnl,icl
+	 * although the name is a bit misleading.
+	 */
+
+	pass = 1;
+	do {
+		wait_ack_clear(d, FORCEWAKE_KERNEL_FALLBACK);
+
+		fw_set(d, FORCEWAKE_KERNEL_FALLBACK);
+		/* Give gt some time to relax before the polling frenzy */
+		udelay(10 * pass);
+		wait_ack_set(d, FORCEWAKE_KERNEL_FALLBACK);
+
+		ack_detected = (fw_ack(d) & ack_bit) == value;
+
+		fw_clear(d, FORCEWAKE_KERNEL_FALLBACK);
+	} while (!ack_detected && pass++ < 10);
+
+	DRM_DEBUG_DRIVER("%s had to use fallback to %s ack, 0x%x (passes %u)\n",
+			 intel_uncore_forcewake_domain_to_str(d->id),
+			 type == ACK_SET ? "set" : "clear",
+			 fw_ack(d),
+			 pass);
+
+	return ack_detected ? 0 : -ETIMEDOUT;
+}
+
+static inline void
+fw_domain_wait_ack_clear_fallback(const struct intel_uncore_forcewake_domain *d)
+{
+	if (likely(!wait_ack_clear(d, FORCEWAKE_KERNEL)))
+		return;
+
+	if (fw_domain_wait_ack_with_fallback(d, ACK_CLEAR))
+		fw_domain_wait_ack_clear(d);
+}
+
+static inline void
+fw_domain_get(const struct intel_uncore_forcewake_domain *d)
+{
+	fw_set(d, FORCEWAKE_KERNEL);
+}
+
+static inline void
+fw_domain_wait_ack_set(const struct intel_uncore_forcewake_domain *d)
+{
+	if (wait_ack_set(d, FORCEWAKE_KERNEL)) {
+		DRM_ERROR("%s: timed out waiting for forcewake ack request.\n",
+			  intel_uncore_forcewake_domain_to_str(d->id));
+		add_taint_for_CI(d->uncore->i915, TAINT_WARN); /* CI now unreliable */
+	}
+}
+
+static inline void
+fw_domain_wait_ack_set_fallback(const struct intel_uncore_forcewake_domain *d)
+{
+	if (likely(!wait_ack_set(d, FORCEWAKE_KERNEL)))
+		return;
+
+	if (fw_domain_wait_ack_with_fallback(d, ACK_SET))
+		fw_domain_wait_ack_set(d);
+}
+
+static inline void
+fw_domain_put(const struct intel_uncore_forcewake_domain *d)
+{
+	fw_clear(d, FORCEWAKE_KERNEL);
+}
+
+static void
+fw_domains_get_normal(struct intel_uncore *uncore, enum forcewake_domains fw_domains)
+{
+	struct intel_uncore_forcewake_domain *d;
+	unsigned int tmp;
+
+	GEM_BUG_ON(fw_domains & ~uncore->fw_domains);
+
+	for_each_fw_domain_masked(d, fw_domains, uncore, tmp) {
+		fw_domain_wait_ack_clear(d);
+		fw_domain_get(d);
+	}
+
+	for_each_fw_domain_masked(d, fw_domains, uncore, tmp)
+		fw_domain_wait_ack_set(d);
+
+	uncore->fw_domains_active |= fw_domains;
+}
+
+static void
+fw_domains_get_with_fallback(struct intel_uncore *uncore,
+			     enum forcewake_domains fw_domains)
+{
+	struct intel_uncore_forcewake_domain *d;
+	unsigned int tmp;
+
+	GEM_BUG_ON(fw_domains & ~uncore->fw_domains);
+
+	for_each_fw_domain_masked(d, fw_domains, uncore, tmp) {
+		fw_domain_wait_ack_clear_fallback(d);
+		fw_domain_get(d);
+	}
+
+	for_each_fw_domain_masked(d, fw_domains, uncore, tmp)
+		fw_domain_wait_ack_set_fallback(d);
+
+	uncore->fw_domains_active |= fw_domains;
+}
+
+static void
+fw_domains_put(struct intel_uncore *uncore, enum forcewake_domains fw_domains)
+{
+	struct intel_uncore_forcewake_domain *d;
+	unsigned int tmp;
+
+	GEM_BUG_ON(fw_domains & ~uncore->fw_domains);
+
+	for_each_fw_domain_masked(d, fw_domains, uncore, tmp)
+		fw_domain_put(d);
+
+	uncore->fw_domains_active &= ~fw_domains;
+}
+
+static void
+fw_domains_reset(struct intel_uncore *uncore,
+		 enum forcewake_domains fw_domains)
+{
+	struct intel_uncore_forcewake_domain *d;
+	unsigned int tmp;
+
+	if (!fw_domains)
+		return;
+
+	GEM_BUG_ON(fw_domains & ~uncore->fw_domains);
+
+	for_each_fw_domain_masked(d, fw_domains, uncore, tmp)
+		fw_domain_reset(d);
+}
+
+static inline u32 gt_thread_status(struct intel_uncore *uncore)
+{
+	u32 val;
+
+	val = __raw_uncore_read32(uncore, GEN6_GT_THREAD_STATUS_REG);
+	val &= GEN6_GT_THREAD_STATUS_CORE_MASK;
+
+	return val;
+}
+
+static void __gen6_gt_wait_for_thread_c0(struct intel_uncore *uncore)
+{
+	/*
+	 * w/a for a sporadic read returning 0 by waiting for the GT
+	 * thread to wake up.
+	 */
+	drm_WARN_ONCE(&uncore->i915->drm,
+		      wait_for_atomic_us(gt_thread_status(uncore) == 0, 5000),
+		      "GT thread status wait timed out\n");
+}
+
+static void fw_domains_get_with_thread_status(struct intel_uncore *uncore,
+					      enum forcewake_domains fw_domains)
+{
+	fw_domains_get_normal(uncore, fw_domains);
+
+	/* WaRsForcewakeWaitTC0:snb,ivb,hsw,bdw,vlv */
+	__gen6_gt_wait_for_thread_c0(uncore);
+}
+
+static inline u32 fifo_free_entries(struct intel_uncore *uncore)
+{
+	u32 count = __raw_uncore_read32(uncore, GTFIFOCTL);
+
+	return count & GT_FIFO_FREE_ENTRIES_MASK;
+}
+
+static void __gen6_gt_wait_for_fifo(struct intel_uncore *uncore)
+{
+	u32 n;
+
+	/* On VLV, FIFO will be shared by both SW and HW.
+	 * So, we need to read the FREE_ENTRIES everytime */
+	if (IS_VALLEYVIEW(uncore->i915))
+		n = fifo_free_entries(uncore);
+	else
+		n = uncore->fifo_count;
+
+	if (n <= GT_FIFO_NUM_RESERVED_ENTRIES) {
+		if (wait_for_atomic((n = fifo_free_entries(uncore)) >
+				    GT_FIFO_NUM_RESERVED_ENTRIES,
+				    GT_FIFO_TIMEOUT_MS)) {
+			drm_dbg(&uncore->i915->drm,
+				"GT_FIFO timeout, entries: %u\n", n);
+			return;
+		}
+	}
+
+	uncore->fifo_count = n - 1;
+}
+
+static enum hrtimer_restart
+intel_uncore_fw_release_timer(struct hrtimer *timer)
+{
+	struct intel_uncore_forcewake_domain *domain =
+	       container_of(timer, struct intel_uncore_forcewake_domain, timer);
+	struct intel_uncore *uncore = domain->uncore;
+	unsigned long irqflags;
+
+	assert_rpm_device_not_suspended(uncore->rpm);
+
+	if (xchg(&domain->active, false))
+		return HRTIMER_RESTART;
+
+	spin_lock_irqsave(&uncore->lock, irqflags);
+
+	uncore->fw_domains_timer &= ~domain->mask;
+
+	GEM_BUG_ON(!domain->wake_count);
+	if (--domain->wake_count == 0)
+		fw_domains_put(uncore, domain->mask);
+
+	spin_unlock_irqrestore(&uncore->lock, irqflags);
+
+	return HRTIMER_NORESTART;
+}
+
+/* Note callers must have acquired the PUNIT->PMIC bus, before calling this. */
+static unsigned int
+intel_uncore_forcewake_reset(struct intel_uncore *uncore)
+{
+	unsigned long irqflags;
+	struct intel_uncore_forcewake_domain *domain;
+	int retry_count = 100;
+	enum forcewake_domains fw, active_domains;
+
+	iosf_mbi_assert_punit_acquired();
+
+	/* Hold uncore.lock across reset to prevent any register access
+	 * with forcewake not set correctly. Wait until all pending
+	 * timers are run before holding.
+	 */
+	while (1) {
+		unsigned int tmp;
+
+		active_domains = 0;
+
+		for_each_fw_domain(domain, uncore, tmp) {
+			smp_store_mb(domain->active, false);
+			if (hrtimer_cancel(&domain->timer) == 0)
+				continue;
+
+			intel_uncore_fw_release_timer(&domain->timer);
+		}
+
+		spin_lock_irqsave(&uncore->lock, irqflags);
+
+		for_each_fw_domain(domain, uncore, tmp) {
+			if (hrtimer_active(&domain->timer))
+				active_domains |= domain->mask;
+		}
+
+		if (active_domains == 0)
+			break;
+
+		if (--retry_count == 0) {
+			drm_err(&uncore->i915->drm, "Timed out waiting for forcewake timers to finish\n");
+			break;
+		}
+
+		spin_unlock_irqrestore(&uncore->lock, irqflags);
+		cond_resched();
+	}
+
+	drm_WARN_ON(&uncore->i915->drm, active_domains);
+
+	fw = uncore->fw_domains_active;
+	if (fw)
+		fw_domains_put(uncore, fw);
+
+	fw_domains_reset(uncore, uncore->fw_domains);
+	assert_forcewakes_inactive(uncore);
+
+	spin_unlock_irqrestore(&uncore->lock, irqflags);
+
+	return fw; /* track the lost user forcewake domains */
+}
+
+static bool
+fpga_check_for_unclaimed_mmio(struct intel_uncore *uncore)
+{
+	u32 dbg;
+
+	dbg = __raw_uncore_read32(uncore, FPGA_DBG);
+	if (likely(!(dbg & FPGA_DBG_RM_NOCLAIM)))
+		return false;
+
+	/*
+	 * Bugs in PCI programming (or failing hardware) can occasionally cause
+	 * us to lose access to the MMIO BAR.  When this happens, register
+	 * reads will come back with 0xFFFFFFFF for every register and things
+	 * go bad very quickly.  Let's try to detect that special case and at
+	 * least try to print a more informative message about what has
+	 * happened.
+	 *
+	 * During normal operation the FPGA_DBG register has several unused
+	 * bits that will always read back as 0's so we can use them as canaries
+	 * to recognize when MMIO accesses are just busted.
+	 */
+	if (unlikely(dbg == ~0))
+		drm_err(&uncore->i915->drm,
+			"Lost access to MMIO BAR; all registers now read back as 0xFFFFFFFF!\n");
+
+	__raw_uncore_write32(uncore, FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
+
+	return true;
+}
+
+static bool
+vlv_check_for_unclaimed_mmio(struct intel_uncore *uncore)
+{
+	u32 cer;
+
+	cer = __raw_uncore_read32(uncore, CLAIM_ER);
+	if (likely(!(cer & (CLAIM_ER_OVERFLOW | CLAIM_ER_CTR_MASK))))
+		return false;
+
+	__raw_uncore_write32(uncore, CLAIM_ER, CLAIM_ER_CLR);
+
+	return true;
+}
+
+static bool
+gen6_check_for_fifo_debug(struct intel_uncore *uncore)
+{
+	u32 fifodbg;
+
+	fifodbg = __raw_uncore_read32(uncore, GTFIFODBG);
+
+	if (unlikely(fifodbg)) {
+		drm_dbg(&uncore->i915->drm, "GTFIFODBG = 0x08%x\n", fifodbg);
+		__raw_uncore_write32(uncore, GTFIFODBG, fifodbg);
+	}
+
+	return fifodbg;
+}
+
+static bool
+check_for_unclaimed_mmio(struct intel_uncore *uncore)
+{
+	bool ret = false;
+
+	lockdep_assert_held(&uncore->debug->lock);
+
+	if (uncore->debug->suspend_count)
+		return false;
+
+	if (intel_uncore_has_fpga_dbg_unclaimed(uncore))
+		ret |= fpga_check_for_unclaimed_mmio(uncore);
+
+	if (intel_uncore_has_dbg_unclaimed(uncore))
+		ret |= vlv_check_for_unclaimed_mmio(uncore);
+
+	if (intel_uncore_has_fifo(uncore))
+		ret |= gen6_check_for_fifo_debug(uncore);
+
+	return ret;
+}
+
+static void forcewake_early_sanitize(struct intel_uncore *uncore,
+				     unsigned int restore_forcewake)
+{
+	GEM_BUG_ON(!intel_uncore_has_forcewake(uncore));
+
+	/* WaDisableShadowRegForCpd:chv */
+	if (IS_CHERRYVIEW(uncore->i915)) {
+		__raw_uncore_write32(uncore, GTFIFOCTL,
+				     __raw_uncore_read32(uncore, GTFIFOCTL) |
+				     GT_FIFO_CTL_BLOCK_ALL_POLICY_STALL |
+				     GT_FIFO_CTL_RC6_POLICY_STALL);
+	}
+
+	iosf_mbi_punit_acquire();
+	intel_uncore_forcewake_reset(uncore);
+	if (restore_forcewake) {
+		spin_lock_irq(&uncore->lock);
+		fw_domains_get(uncore, restore_forcewake);
+
+		if (intel_uncore_has_fifo(uncore))
+			uncore->fifo_count = fifo_free_entries(uncore);
+		spin_unlock_irq(&uncore->lock);
+	}
+	iosf_mbi_punit_release();
+}
+
+void intel_uncore_suspend(struct intel_uncore *uncore)
+{
+	if (!intel_uncore_has_forcewake(uncore))
+		return;
+
+	iosf_mbi_punit_acquire();
+	iosf_mbi_unregister_pmic_bus_access_notifier_unlocked(
+		&uncore->pmic_bus_access_nb);
+	uncore->fw_domains_saved = intel_uncore_forcewake_reset(uncore);
+	iosf_mbi_punit_release();
+}
+
+void intel_uncore_resume_early(struct intel_uncore *uncore)
+{
+	unsigned int restore_forcewake;
+
+	if (intel_uncore_unclaimed_mmio(uncore))
+		drm_dbg(&uncore->i915->drm, "unclaimed mmio detected on resume, clearing\n");
+
+	if (!intel_uncore_has_forcewake(uncore))
+		return;
+
+	restore_forcewake = fetch_and_zero(&uncore->fw_domains_saved);
+	forcewake_early_sanitize(uncore, restore_forcewake);
+
+	iosf_mbi_register_pmic_bus_access_notifier(&uncore->pmic_bus_access_nb);
+}
+
+void intel_uncore_runtime_resume(struct intel_uncore *uncore)
+{
+	if (!intel_uncore_has_forcewake(uncore))
+		return;
+
+	iosf_mbi_register_pmic_bus_access_notifier(&uncore->pmic_bus_access_nb);
+}
+
+static void __intel_uncore_forcewake_get(struct intel_uncore *uncore,
+					 enum forcewake_domains fw_domains)
+{
+	struct intel_uncore_forcewake_domain *domain;
+	unsigned int tmp;
+
+	fw_domains &= uncore->fw_domains;
+
+	for_each_fw_domain_masked(domain, fw_domains, uncore, tmp) {
+		if (domain->wake_count++) {
+			fw_domains &= ~domain->mask;
+			domain->active = true;
+		}
+	}
+
+	if (fw_domains)
+		fw_domains_get(uncore, fw_domains);
+}
+
+/**
+ * intel_uncore_forcewake_get - grab forcewake domain references
+ * @uncore: the intel_uncore structure
+ * @fw_domains: forcewake domains to get reference on
+ *
+ * This function can be used get GT's forcewake domain references.
+ * Normal register access will handle the forcewake domains automatically.
+ * However if some sequence requires the GT to not power down a particular
+ * forcewake domains this function should be called at the beginning of the
+ * sequence. And subsequently the reference should be dropped by symmetric
+ * call to intel_unforce_forcewake_put(). Usually caller wants all the domains
+ * to be kept awake so the @fw_domains would be then FORCEWAKE_ALL.
+ */
+void intel_uncore_forcewake_get(struct intel_uncore *uncore,
+				enum forcewake_domains fw_domains)
+{
+	unsigned long irqflags;
+
+	if (!uncore->fw_get_funcs)
+		return;
+
+	assert_rpm_wakelock_held(uncore->rpm);
+
+	spin_lock_irqsave(&uncore->lock, irqflags);
+	__intel_uncore_forcewake_get(uncore, fw_domains);
+	spin_unlock_irqrestore(&uncore->lock, irqflags);
+}
+
+/**
+ * intel_uncore_forcewake_user_get - claim forcewake on behalf of userspace
+ * @uncore: the intel_uncore structure
+ *
+ * This function is a wrapper around intel_uncore_forcewake_get() to acquire
+ * the GT powerwell and in the process disable our debugging for the
+ * duration of userspace's bypass.
+ */
+void intel_uncore_forcewake_user_get(struct intel_uncore *uncore)
+{
+	spin_lock_irq(&uncore->lock);
+	if (!uncore->user_forcewake_count++) {
+		intel_uncore_forcewake_get__locked(uncore, FORCEWAKE_ALL);
+		mmio_debug_suspend(uncore);
+	}
+	spin_unlock_irq(&uncore->lock);
+}
+
+/**
+ * intel_uncore_forcewake_user_put - release forcewake on behalf of userspace
+ * @uncore: the intel_uncore structure
+ *
+ * This function complements intel_uncore_forcewake_user_get() and releases
+ * the GT powerwell taken on behalf of the userspace bypass.
+ */
+void intel_uncore_forcewake_user_put(struct intel_uncore *uncore)
+{
+	spin_lock_irq(&uncore->lock);
+	if (!--uncore->user_forcewake_count) {
+		mmio_debug_resume(uncore);
+		intel_uncore_forcewake_put__locked(uncore, FORCEWAKE_ALL);
+	}
+	spin_unlock_irq(&uncore->lock);
+}
+
+/**
+ * intel_uncore_forcewake_get__locked - grab forcewake domain references
+ * @uncore: the intel_uncore structure
+ * @fw_domains: forcewake domains to get reference on
+ *
+ * See intel_uncore_forcewake_get(). This variant places the onus
+ * on the caller to explicitly handle the dev_priv->uncore.lock spinlock.
+ */
+void intel_uncore_forcewake_get__locked(struct intel_uncore *uncore,
+					enum forcewake_domains fw_domains)
+{
+	lockdep_assert_held(&uncore->lock);
+
+	if (!uncore->fw_get_funcs)
+		return;
+
+	__intel_uncore_forcewake_get(uncore, fw_domains);
+}
+
+static void __intel_uncore_forcewake_put(struct intel_uncore *uncore,
+					 enum forcewake_domains fw_domains,
+					 bool delayed)
+{
+	struct intel_uncore_forcewake_domain *domain;
+	unsigned int tmp;
+
+	fw_domains &= uncore->fw_domains;
+
+	for_each_fw_domain_masked(domain, fw_domains, uncore, tmp) {
+		GEM_BUG_ON(!domain->wake_count);
+
+		if (--domain->wake_count) {
+			domain->active = true;
+			continue;
+		}
+
+		if (delayed &&
+		    !(domain->uncore->fw_domains_timer & domain->mask))
+			fw_domain_arm_timer(domain);
+		else
+			fw_domains_put(uncore, domain->mask);
+	}
+}
+
+/**
+ * intel_uncore_forcewake_put - release a forcewake domain reference
+ * @uncore: the intel_uncore structure
+ * @fw_domains: forcewake domains to put references
+ *
+ * This function drops the device-level forcewakes for specified
+ * domains obtained by intel_uncore_forcewake_get().
+ */
+void intel_uncore_forcewake_put(struct intel_uncore *uncore,
+				enum forcewake_domains fw_domains)
+{
+	unsigned long irqflags;
+
+	if (!uncore->fw_get_funcs)
+		return;
+
+	spin_lock_irqsave(&uncore->lock, irqflags);
+	__intel_uncore_forcewake_put(uncore, fw_domains, false);
+	spin_unlock_irqrestore(&uncore->lock, irqflags);
+}
+
+void intel_uncore_forcewake_put_delayed(struct intel_uncore *uncore,
+					enum forcewake_domains fw_domains)
+{
+	unsigned long irqflags;
+
+	if (!uncore->fw_get_funcs)
+		return;
+
+	spin_lock_irqsave(&uncore->lock, irqflags);
+	__intel_uncore_forcewake_put(uncore, fw_domains, true);
+	spin_unlock_irqrestore(&uncore->lock, irqflags);
+}
+
+/**
+ * intel_uncore_forcewake_flush - flush the delayed release
+ * @uncore: the intel_uncore structure
+ * @fw_domains: forcewake domains to flush
+ */
+void intel_uncore_forcewake_flush(struct intel_uncore *uncore,
+				  enum forcewake_domains fw_domains)
+{
+	struct intel_uncore_forcewake_domain *domain;
+	unsigned int tmp;
+
+	if (!uncore->fw_get_funcs)
+		return;
+
+	fw_domains &= uncore->fw_domains;
+	for_each_fw_domain_masked(domain, fw_domains, uncore, tmp) {
+		WRITE_ONCE(domain->active, false);
+		if (hrtimer_cancel(&domain->timer))
+			intel_uncore_fw_release_timer(&domain->timer);
+	}
+}
+
+/**
+ * intel_uncore_forcewake_put__locked - grab forcewake domain references
+ * @uncore: the intel_uncore structure
+ * @fw_domains: forcewake domains to get reference on
+ *
+ * See intel_uncore_forcewake_put(). This variant places the onus
+ * on the caller to explicitly handle the dev_priv->uncore.lock spinlock.
+ */
+void intel_uncore_forcewake_put__locked(struct intel_uncore *uncore,
+					enum forcewake_domains fw_domains)
+{
+	lockdep_assert_held(&uncore->lock);
+
+	if (!uncore->fw_get_funcs)
+		return;
+
+	__intel_uncore_forcewake_put(uncore, fw_domains, false);
+}
+
+void assert_forcewakes_inactive(struct intel_uncore *uncore)
+{
+	if (!uncore->fw_get_funcs)
+		return;
+
+	drm_WARN(&uncore->i915->drm, uncore->fw_domains_active,
+		 "Expected all fw_domains to be inactive, but %08x are still on\n",
+		 uncore->fw_domains_active);
+}
+
+void assert_forcewakes_active(struct intel_uncore *uncore,
+			      enum forcewake_domains fw_domains)
+{
+	struct intel_uncore_forcewake_domain *domain;
+	unsigned int tmp;
+
+	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM))
+		return;
+
+	if (!uncore->fw_get_funcs)
+		return;
+
+	spin_lock_irq(&uncore->lock);
+
+	assert_rpm_wakelock_held(uncore->rpm);
+
+	fw_domains &= uncore->fw_domains;
+	drm_WARN(&uncore->i915->drm, fw_domains & ~uncore->fw_domains_active,
+		 "Expected %08x fw_domains to be active, but %08x are off\n",
+		 fw_domains, fw_domains & ~uncore->fw_domains_active);
+
+	/*
+	 * Check that the caller has an explicit wakeref and we don't mistake
+	 * it for the auto wakeref.
+	 */
+	for_each_fw_domain_masked(domain, fw_domains, uncore, tmp) {
+		unsigned int actual = READ_ONCE(domain->wake_count);
+		unsigned int expect = 1;
+
+		if (uncore->fw_domains_timer & domain->mask)
+			expect++; /* pending automatic release */
+
+		if (drm_WARN(&uncore->i915->drm, actual < expect,
+			     "Expected domain %d to be held awake by caller, count=%d\n",
+			     domain->id, actual))
+			break;
+	}
+
+	spin_unlock_irq(&uncore->lock);
+}
+
+/* We give fast paths for the really cool registers */
+#define NEEDS_FORCE_WAKE(reg) ({ \
+	u32 __reg = (reg); \
+	__reg < 0x40000 || __reg >= GEN11_BSD_RING_BASE; \
+})
+
+static int fw_range_cmp(u32 offset, const struct intel_forcewake_range *entry)
+{
+	if (offset < entry->start)
+		return -1;
+	else if (offset > entry->end)
+		return 1;
+	else
+		return 0;
+}
+
+/* Copied and "macroized" from lib/bsearch.c */
+#define BSEARCH(key, base, num, cmp) ({                                 \
+	unsigned int start__ = 0, end__ = (num);                        \
+	typeof(base) result__ = NULL;                                   \
+	while (start__ < end__) {                                       \
+		unsigned int mid__ = start__ + (end__ - start__) / 2;   \
+		int ret__ = (cmp)((key), (base) + mid__);               \
+		if (ret__ < 0) {                                        \
+			end__ = mid__;                                  \
+		} else if (ret__ > 0) {                                 \
+			start__ = mid__ + 1;                            \
+		} else {                                                \
+			result__ = (base) + mid__;                      \
+			break;                                          \
+		}                                                       \
+	}                                                               \
+	result__;                                                       \
+})
+
+static enum forcewake_domains
+find_fw_domain(struct intel_uncore *uncore, u32 offset)
+{
+	const struct intel_forcewake_range *entry;
+
+	if (IS_GSI_REG(offset))
+		offset += uncore->gsi_offset;
+
+	entry = BSEARCH(offset,
+			uncore->fw_domains_table,
+			uncore->fw_domains_table_entries,
+			fw_range_cmp);
+
+	if (!entry)
+		return 0;
+
+	/*
+	 * The list of FW domains depends on the SKU in gen11+ so we
+	 * can't determine it statically. We use FORCEWAKE_ALL and
+	 * translate it here to the list of available domains.
+	 */
+	if (entry->domains == FORCEWAKE_ALL)
+		return uncore->fw_domains;
+
+	drm_WARN(&uncore->i915->drm, entry->domains & ~uncore->fw_domains,
+		 "Uninitialized forcewake domain(s) 0x%x accessed at 0x%x\n",
+		 entry->domains & ~uncore->fw_domains, offset);
+
+	return entry->domains;
+}
+
+/*
+ * Shadowed register tables describe special register ranges that i915 is
+ * allowed to write to without acquiring forcewake.  If these registers' power
+ * wells are down, the hardware will save values written by i915 to a shadow
+ * copy and automatically transfer them into the real register the next time
+ * the power well is woken up.  Shadowing only applies to writes; forcewake
+ * must still be acquired when reading from registers in these ranges.
+ *
+ * The documentation for shadowed registers is somewhat spotty on older
+ * platforms.  However missing registers from these lists is non-fatal; it just
+ * means we'll wake up the hardware for some register accesses where we didn't
+ * really need to.
+ *
+ * The ranges listed in these tables must be sorted by offset.
+ *
+ * When adding new tables here, please also add them to
+ * intel_shadow_table_check() in selftests/intel_uncore.c so that they will be
+ * scanned for obvious mistakes or typos by the selftests.
+ */
+
+static const struct i915_range gen8_shadowed_regs[] = {
+	{ .start =  0x2030, .end =  0x2030 },
+	{ .start =  0xA008, .end =  0xA00C },
+	{ .start = 0x12030, .end = 0x12030 },
+	{ .start = 0x1a030, .end = 0x1a030 },
+	{ .start = 0x22030, .end = 0x22030 },
+};
+
+static const struct i915_range gen11_shadowed_regs[] = {
+	{ .start =   0x2030, .end =   0x2030 },
+	{ .start =   0x2550, .end =   0x2550 },
+	{ .start =   0xA008, .end =   0xA00C },
+	{ .start =  0x22030, .end =  0x22030 },
+	{ .start =  0x22230, .end =  0x22230 },
+	{ .start =  0x22510, .end =  0x22550 },
+	{ .start = 0x1C0030, .end = 0x1C0030 },
+	{ .start = 0x1C0230, .end = 0x1C0230 },
+	{ .start = 0x1C0510, .end = 0x1C0550 },
+	{ .start = 0x1C4030, .end = 0x1C4030 },
+	{ .start = 0x1C4230, .end = 0x1C4230 },
+	{ .start = 0x1C4510, .end = 0x1C4550 },
+	{ .start = 0x1C8030, .end = 0x1C8030 },
+	{ .start = 0x1C8230, .end = 0x1C8230 },
+	{ .start = 0x1C8510, .end = 0x1C8550 },
+	{ .start = 0x1D0030, .end = 0x1D0030 },
+	{ .start = 0x1D0230, .end = 0x1D0230 },
+	{ .start = 0x1D0510, .end = 0x1D0550 },
+	{ .start = 0x1D4030, .end = 0x1D4030 },
+	{ .start = 0x1D4230, .end = 0x1D4230 },
+	{ .start = 0x1D4510, .end = 0x1D4550 },
+	{ .start = 0x1D8030, .end = 0x1D8030 },
+	{ .start = 0x1D8230, .end = 0x1D8230 },
+	{ .start = 0x1D8510, .end = 0x1D8550 },
+};
+
+static const struct i915_range gen12_shadowed_regs[] = {
+	{ .start =   0x2030, .end =   0x2030 },
+	{ .start =   0x2510, .end =   0x2550 },
+	{ .start =   0xA008, .end =   0xA00C },
+	{ .start =   0xA188, .end =   0xA188 },
+	{ .start =   0xA278, .end =   0xA278 },
+	{ .start =   0xA540, .end =   0xA56C },
+	{ .start =   0xC4C8, .end =   0xC4C8 },
+	{ .start =   0xC4D4, .end =   0xC4D4 },
+	{ .start =   0xC600, .end =   0xC600 },
+	{ .start =  0x22030, .end =  0x22030 },
+	{ .start =  0x22510, .end =  0x22550 },
+	{ .start = 0x1C0030, .end = 0x1C0030 },
+	{ .start = 0x1C0510, .end = 0x1C0550 },
+	{ .start = 0x1C4030, .end = 0x1C4030 },
+	{ .start = 0x1C4510, .end = 0x1C4550 },
+	{ .start = 0x1C8030, .end = 0x1C8030 },
+	{ .start = 0x1C8510, .end = 0x1C8550 },
+	{ .start = 0x1D0030, .end = 0x1D0030 },
+	{ .start = 0x1D0510, .end = 0x1D0550 },
+	{ .start = 0x1D4030, .end = 0x1D4030 },
+	{ .start = 0x1D4510, .end = 0x1D4550 },
+	{ .start = 0x1D8030, .end = 0x1D8030 },
+	{ .start = 0x1D8510, .end = 0x1D8550 },
+
+	/*
+	 * The rest of these ranges are specific to Xe_HP and beyond, but
+	 * are reserved/unused ranges on earlier gen12 platforms, so they can
+	 * be safely added to the gen12 table.
+	 */
+	{ .start = 0x1E0030, .end = 0x1E0030 },
+	{ .start = 0x1E0510, .end = 0x1E0550 },
+	{ .start = 0x1E4030, .end = 0x1E4030 },
+	{ .start = 0x1E4510, .end = 0x1E4550 },
+	{ .start = 0x1E8030, .end = 0x1E8030 },
+	{ .start = 0x1E8510, .end = 0x1E8550 },
+	{ .start = 0x1F0030, .end = 0x1F0030 },
+	{ .start = 0x1F0510, .end = 0x1F0550 },
+	{ .start = 0x1F4030, .end = 0x1F4030 },
+	{ .start = 0x1F4510, .end = 0x1F4550 },
+	{ .start = 0x1F8030, .end = 0x1F8030 },
+	{ .start = 0x1F8510, .end = 0x1F8550 },
+};
+
+static const struct i915_range dg2_shadowed_regs[] = {
+	{ .start =   0x2030, .end =   0x2030 },
+	{ .start =   0x2510, .end =   0x2550 },
+	{ .start =   0xA008, .end =   0xA00C },
+	{ .start =   0xA188, .end =   0xA188 },
+	{ .start =   0xA278, .end =   0xA278 },
+	{ .start =   0xA540, .end =   0xA56C },
+	{ .start =   0xC4C8, .end =   0xC4C8 },
+	{ .start =   0xC4E0, .end =   0xC4E0 },
+	{ .start =   0xC600, .end =   0xC600 },
+	{ .start =   0xC658, .end =   0xC658 },
+	{ .start =  0x22030, .end =  0x22030 },
+	{ .start =  0x22510, .end =  0x22550 },
+	{ .start = 0x1C0030, .end = 0x1C0030 },
+	{ .start = 0x1C0510, .end = 0x1C0550 },
+	{ .start = 0x1C4030, .end = 0x1C4030 },
+	{ .start = 0x1C4510, .end = 0x1C4550 },
+	{ .start = 0x1C8030, .end = 0x1C8030 },
+	{ .start = 0x1C8510, .end = 0x1C8550 },
+	{ .start = 0x1D0030, .end = 0x1D0030 },
+	{ .start = 0x1D0510, .end = 0x1D0550 },
+	{ .start = 0x1D4030, .end = 0x1D4030 },
+	{ .start = 0x1D4510, .end = 0x1D4550 },
+	{ .start = 0x1D8030, .end = 0x1D8030 },
+	{ .start = 0x1D8510, .end = 0x1D8550 },
+	{ .start = 0x1E0030, .end = 0x1E0030 },
+	{ .start = 0x1E0510, .end = 0x1E0550 },
+	{ .start = 0x1E4030, .end = 0x1E4030 },
+	{ .start = 0x1E4510, .end = 0x1E4550 },
+	{ .start = 0x1E8030, .end = 0x1E8030 },
+	{ .start = 0x1E8510, .end = 0x1E8550 },
+	{ .start = 0x1F0030, .end = 0x1F0030 },
+	{ .start = 0x1F0510, .end = 0x1F0550 },
+	{ .start = 0x1F4030, .end = 0x1F4030 },
+	{ .start = 0x1F4510, .end = 0x1F4550 },
+	{ .start = 0x1F8030, .end = 0x1F8030 },
+	{ .start = 0x1F8510, .end = 0x1F8550 },
+};
+
+static const struct i915_range pvc_shadowed_regs[] = {
+	{ .start =   0x2030, .end =   0x2030 },
+	{ .start =   0x2510, .end =   0x2550 },
+	{ .start =   0xA008, .end =   0xA00C },
+	{ .start =   0xA188, .end =   0xA188 },
+	{ .start =   0xA278, .end =   0xA278 },
+	{ .start =   0xA540, .end =   0xA56C },
+	{ .start =   0xC4C8, .end =   0xC4C8 },
+	{ .start =   0xC4E0, .end =   0xC4E0 },
+	{ .start =   0xC600, .end =   0xC600 },
+	{ .start =   0xC658, .end =   0xC658 },
+	{ .start =  0x22030, .end =  0x22030 },
+	{ .start =  0x22510, .end =  0x22550 },
+	{ .start = 0x1C0030, .end = 0x1C0030 },
+	{ .start = 0x1C0510, .end = 0x1C0550 },
+	{ .start = 0x1C4030, .end = 0x1C4030 },
+	{ .start = 0x1C4510, .end = 0x1C4550 },
+	{ .start = 0x1C8030, .end = 0x1C8030 },
+	{ .start = 0x1C8510, .end = 0x1C8550 },
+	{ .start = 0x1D0030, .end = 0x1D0030 },
+	{ .start = 0x1D0510, .end = 0x1D0550 },
+	{ .start = 0x1D4030, .end = 0x1D4030 },
+	{ .start = 0x1D4510, .end = 0x1D4550 },
+	{ .start = 0x1D8030, .end = 0x1D8030 },
+	{ .start = 0x1D8510, .end = 0x1D8550 },
+	{ .start = 0x1E0030, .end = 0x1E0030 },
+	{ .start = 0x1E0510, .end = 0x1E0550 },
+	{ .start = 0x1E4030, .end = 0x1E4030 },
+	{ .start = 0x1E4510, .end = 0x1E4550 },
+	{ .start = 0x1E8030, .end = 0x1E8030 },
+	{ .start = 0x1E8510, .end = 0x1E8550 },
+	{ .start = 0x1F0030, .end = 0x1F0030 },
+	{ .start = 0x1F0510, .end = 0x1F0550 },
+	{ .start = 0x1F4030, .end = 0x1F4030 },
+	{ .start = 0x1F4510, .end = 0x1F4550 },
+	{ .start = 0x1F8030, .end = 0x1F8030 },
+	{ .start = 0x1F8510, .end = 0x1F8550 },
+};
+
+static int mmio_range_cmp(u32 key, const struct i915_range *range)
+{
+	if (key < range->start)
+		return -1;
+	else if (key > range->end)
+		return 1;
+	else
+		return 0;
+}
+
+static bool is_shadowed(struct intel_uncore *uncore, u32 offset)
+{
+	if (drm_WARN_ON(&uncore->i915->drm, !uncore->shadowed_reg_table))
+		return false;
+
+	if (IS_GSI_REG(offset))
+		offset += uncore->gsi_offset;
+
+	return BSEARCH(offset,
+		       uncore->shadowed_reg_table,
+		       uncore->shadowed_reg_table_entries,
+		       mmio_range_cmp);
+}
+
+static enum forcewake_domains
+gen6_reg_write_fw_domains(struct intel_uncore *uncore, i915_reg_t reg)
+{
+	return FORCEWAKE_RENDER;
+}
+
+#define __fwtable_reg_read_fw_domains(uncore, offset) \
+({ \
+	enum forcewake_domains __fwd = 0; \
+	if (NEEDS_FORCE_WAKE((offset))) \
+		__fwd = find_fw_domain(uncore, offset); \
+	__fwd; \
+})
+
+#define __fwtable_reg_write_fw_domains(uncore, offset) \
+({ \
+	enum forcewake_domains __fwd = 0; \
+	const u32 __offset = (offset); \
+	if (NEEDS_FORCE_WAKE((__offset)) && !is_shadowed(uncore, __offset)) \
+		__fwd = find_fw_domain(uncore, __offset); \
+	__fwd; \
+})
+
+#define GEN_FW_RANGE(s, e, d) \
+	{ .start = (s), .end = (e), .domains = (d) }
+
+/*
+ * All platforms' forcewake tables below must be sorted by offset ranges.
+ * Furthermore, new forcewake tables added should be "watertight" and have
+ * no gaps between ranges.
+ *
+ * When there are multiple consecutive ranges listed in the bspec with
+ * the same forcewake domain, it is customary to combine them into a single
+ * row in the tables below to keep the tables small and lookups fast.
+ * Likewise, reserved/unused ranges may be combined with the preceding and/or
+ * following ranges since the driver will never be making MMIO accesses in
+ * those ranges.
+ *
+ * For example, if the bspec were to list:
+ *
+ *    ...
+ *    0x1000 - 0x1fff:  GT
+ *    0x2000 - 0x2cff:  GT
+ *    0x2d00 - 0x2fff:  unused/reserved
+ *    0x3000 - 0xffff:  GT
+ *    ...
+ *
+ * these could all be represented by a single line in the code:
+ *
+ *   GEN_FW_RANGE(0x1000, 0xffff, FORCEWAKE_GT)
+ *
+ * When adding new forcewake tables here, please also add them to
+ * intel_uncore_mock_selftests in selftests/intel_uncore.c so that they will be
+ * scanned for obvious mistakes or typos by the selftests.
+ */
+
+static const struct intel_forcewake_range __gen6_fw_ranges[] = {
+	GEN_FW_RANGE(0x0, 0x3ffff, FORCEWAKE_RENDER),
+};
+
+static const struct intel_forcewake_range __vlv_fw_ranges[] = {
+	GEN_FW_RANGE(0x2000, 0x3fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x5000, 0x7fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0xb000, 0x11fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x12000, 0x13fff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x22000, 0x23fff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x2e000, 0x2ffff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x30000, 0x3ffff, FORCEWAKE_MEDIA),
+};
+
+static const struct intel_forcewake_range __chv_fw_ranges[] = {
+	GEN_FW_RANGE(0x2000, 0x3fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x4000, 0x4fff, FORCEWAKE_RENDER | FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x5200, 0x7fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8000, 0x82ff, FORCEWAKE_RENDER | FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x8300, 0x84ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8500, 0x85ff, FORCEWAKE_RENDER | FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x8800, 0x88ff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x9000, 0xafff, FORCEWAKE_RENDER | FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0xb000, 0xb47f, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0xd000, 0xd7ff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0xe000, 0xe7ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0xf000, 0xffff, FORCEWAKE_RENDER | FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x12000, 0x13fff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x1a000, 0x1bfff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x1e800, 0x1e9ff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x30000, 0x37fff, FORCEWAKE_MEDIA),
+};
+
+static const struct intel_forcewake_range __gen9_fw_ranges[] = {
+	GEN_FW_RANGE(0x0, 0xaff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0xb00, 0x1fff, 0), /* uncore range */
+	GEN_FW_RANGE(0x2000, 0x26ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x2700, 0x2fff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x3000, 0x3fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x4000, 0x51ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x5200, 0x7fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8000, 0x812f, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x8130, 0x813f, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x8140, 0x815f, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8160, 0x82ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x8300, 0x84ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8500, 0x87ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x8800, 0x89ff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x8a00, 0x8bff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x8c00, 0x8cff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8d00, 0x93ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x9400, 0x97ff, FORCEWAKE_RENDER | FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x9800, 0xafff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0xb000, 0xb47f, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0xb480, 0xcfff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0xd000, 0xd7ff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0xd800, 0xdfff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0xe000, 0xe8ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0xe900, 0x11fff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x12000, 0x13fff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x14000, 0x19fff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x1a000, 0x1e9ff, FORCEWAKE_MEDIA),
+	GEN_FW_RANGE(0x1ea00, 0x243ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x24400, 0x247ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x24800, 0x2ffff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x30000, 0x3ffff, FORCEWAKE_MEDIA),
+};
+
+static const struct intel_forcewake_range __gen11_fw_ranges[] = {
+	GEN_FW_RANGE(0x0, 0x1fff, 0), /* uncore range */
+	GEN_FW_RANGE(0x2000, 0x26ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x2700, 0x2fff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x3000, 0x3fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x4000, 0x51ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x5200, 0x7fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8000, 0x813f, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x8140, 0x815f, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8160, 0x82ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x8300, 0x84ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8500, 0x87ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x8800, 0x8bff, 0),
+	GEN_FW_RANGE(0x8c00, 0x8cff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8d00, 0x94cf, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x94d0, 0x955f, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x9560, 0x95ff, 0),
+	GEN_FW_RANGE(0x9600, 0xafff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0xb000, 0xb47f, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0xb480, 0xdeff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0xdf00, 0xe8ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0xe900, 0x16dff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x16e00, 0x19fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x1a000, 0x23fff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x24000, 0x2407f, 0),
+	GEN_FW_RANGE(0x24080, 0x2417f, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x24180, 0x242ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x24300, 0x243ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x24400, 0x24fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x25000, 0x3ffff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x40000, 0x1bffff, 0),
+	GEN_FW_RANGE(0x1c0000, 0x1c3fff, FORCEWAKE_MEDIA_VDBOX0),
+	GEN_FW_RANGE(0x1c4000, 0x1c7fff, 0),
+	GEN_FW_RANGE(0x1c8000, 0x1cffff, FORCEWAKE_MEDIA_VEBOX0),
+	GEN_FW_RANGE(0x1d0000, 0x1d3fff, FORCEWAKE_MEDIA_VDBOX2),
+	GEN_FW_RANGE(0x1d4000, 0x1dbfff, 0)
+};
+
+static const struct intel_forcewake_range __gen12_fw_ranges[] = {
+	GEN_FW_RANGE(0x0, 0x1fff, 0), /*
+		0x0   -  0xaff: reserved
+		0xb00 - 0x1fff: always on */
+	GEN_FW_RANGE(0x2000, 0x26ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x2700, 0x27ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x2800, 0x2aff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x2b00, 0x2fff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x3000, 0x3fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x4000, 0x51ff, FORCEWAKE_GT), /*
+		0x4000 - 0x48ff: gt
+		0x4900 - 0x51ff: reserved */
+	GEN_FW_RANGE(0x5200, 0x7fff, FORCEWAKE_RENDER), /*
+		0x5200 - 0x53ff: render
+		0x5400 - 0x54ff: reserved
+		0x5500 - 0x7fff: render */
+	GEN_FW_RANGE(0x8000, 0x813f, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x8140, 0x815f, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8160, 0x81ff, 0), /*
+		0x8160 - 0x817f: reserved
+		0x8180 - 0x81ff: always on */
+	GEN_FW_RANGE(0x8200, 0x82ff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x8300, 0x84ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8500, 0x94cf, FORCEWAKE_GT), /*
+		0x8500 - 0x87ff: gt
+		0x8800 - 0x8fff: reserved
+		0x9000 - 0x947f: gt
+		0x9480 - 0x94cf: reserved */
+	GEN_FW_RANGE(0x94d0, 0x955f, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x9560, 0x97ff, 0), /*
+		0x9560 - 0x95ff: always on
+		0x9600 - 0x97ff: reserved */
+	GEN_FW_RANGE(0x9800, 0xafff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0xb000, 0xb3ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0xb400, 0xcfff, FORCEWAKE_GT), /*
+		0xb400 - 0xbf7f: gt
+		0xb480 - 0xbfff: reserved
+		0xc000 - 0xcfff: gt */
+	GEN_FW_RANGE(0xd000, 0xd7ff, 0),
+	GEN_FW_RANGE(0xd800, 0xd8ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0xd900, 0xdbff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0xdc00, 0xefff, FORCEWAKE_RENDER), /*
+		0xdc00 - 0xddff: render
+		0xde00 - 0xde7f: reserved
+		0xde80 - 0xe8ff: render
+		0xe900 - 0xefff: reserved */
+	GEN_FW_RANGE(0xf000, 0x147ff, FORCEWAKE_GT), /*
+		 0xf000 - 0xffff: gt
+		0x10000 - 0x147ff: reserved */
+	GEN_FW_RANGE(0x14800, 0x1ffff, FORCEWAKE_RENDER), /*
+		0x14800 - 0x14fff: render
+		0x15000 - 0x16dff: reserved
+		0x16e00 - 0x1bfff: render
+		0x1c000 - 0x1ffff: reserved */
+	GEN_FW_RANGE(0x20000, 0x20fff, FORCEWAKE_MEDIA_VDBOX0),
+	GEN_FW_RANGE(0x21000, 0x21fff, FORCEWAKE_MEDIA_VDBOX2),
+	GEN_FW_RANGE(0x22000, 0x23fff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x24000, 0x2417f, 0), /*
+		0x24000 - 0x2407f: always on
+		0x24080 - 0x2417f: reserved */
+	GEN_FW_RANGE(0x24180, 0x249ff, FORCEWAKE_GT), /*
+		0x24180 - 0x241ff: gt
+		0x24200 - 0x249ff: reserved */
+	GEN_FW_RANGE(0x24a00, 0x251ff, FORCEWAKE_RENDER), /*
+		0x24a00 - 0x24a7f: render
+		0x24a80 - 0x251ff: reserved */
+	GEN_FW_RANGE(0x25200, 0x255ff, FORCEWAKE_GT), /*
+		0x25200 - 0x252ff: gt
+		0x25300 - 0x255ff: reserved */
+	GEN_FW_RANGE(0x25600, 0x2567f, FORCEWAKE_MEDIA_VDBOX0),
+	GEN_FW_RANGE(0x25680, 0x259ff, FORCEWAKE_MEDIA_VDBOX2), /*
+		0x25680 - 0x256ff: VD2
+		0x25700 - 0x259ff: reserved */
+	GEN_FW_RANGE(0x25a00, 0x25a7f, FORCEWAKE_MEDIA_VDBOX0),
+	GEN_FW_RANGE(0x25a80, 0x2ffff, FORCEWAKE_MEDIA_VDBOX2), /*
+		0x25a80 - 0x25aff: VD2
+		0x25b00 - 0x2ffff: reserved */
+	GEN_FW_RANGE(0x30000, 0x3ffff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x40000, 0x1bffff, 0),
+	GEN_FW_RANGE(0x1c0000, 0x1c3fff, FORCEWAKE_MEDIA_VDBOX0), /*
+		0x1c0000 - 0x1c2bff: VD0
+		0x1c2c00 - 0x1c2cff: reserved
+		0x1c2d00 - 0x1c2dff: VD0
+		0x1c2e00 - 0x1c3eff: reserved
+		0x1c3f00 - 0x1c3fff: VD0 */
+	GEN_FW_RANGE(0x1c4000, 0x1c7fff, 0),
+	GEN_FW_RANGE(0x1c8000, 0x1cbfff, FORCEWAKE_MEDIA_VEBOX0), /*
+		0x1c8000 - 0x1ca0ff: VE0
+		0x1ca100 - 0x1cbeff: reserved
+		0x1cbf00 - 0x1cbfff: VE0 */
+	GEN_FW_RANGE(0x1cc000, 0x1cffff, FORCEWAKE_MEDIA_VDBOX0), /*
+		0x1cc000 - 0x1ccfff: VD0
+		0x1cd000 - 0x1cffff: reserved */
+	GEN_FW_RANGE(0x1d0000, 0x1d3fff, FORCEWAKE_MEDIA_VDBOX2), /*
+		0x1d0000 - 0x1d2bff: VD2
+		0x1d2c00 - 0x1d2cff: reserved
+		0x1d2d00 - 0x1d2dff: VD2
+		0x1d2e00 - 0x1d3eff: reserved
+		0x1d3f00 - 0x1d3fff: VD2 */
+};
+
+/*
+ * Graphics IP version 12.55 brings a slight change to the 0xd800 range,
+ * switching it from the GT domain to the render domain.
+ */
+#define XEHP_FWRANGES(FW_RANGE_D800)					\
+	GEN_FW_RANGE(0x0, 0x1fff, 0), /*					\
+		  0x0 -  0xaff: reserved					\
+		0xb00 - 0x1fff: always on */					\
+	GEN_FW_RANGE(0x2000, 0x26ff, FORCEWAKE_RENDER),				\
+	GEN_FW_RANGE(0x2700, 0x4aff, FORCEWAKE_GT),				\
+	GEN_FW_RANGE(0x4b00, 0x51ff, 0), /*					\
+		0x4b00 - 0x4fff: reserved					\
+		0x5000 - 0x51ff: always on */					\
+	GEN_FW_RANGE(0x5200, 0x7fff, FORCEWAKE_RENDER),				\
+	GEN_FW_RANGE(0x8000, 0x813f, FORCEWAKE_GT),				\
+	GEN_FW_RANGE(0x8140, 0x815f, FORCEWAKE_RENDER),				\
+	GEN_FW_RANGE(0x8160, 0x81ff, 0), /*					\
+		0x8160 - 0x817f: reserved					\
+		0x8180 - 0x81ff: always on */					\
+	GEN_FW_RANGE(0x8200, 0x82ff, FORCEWAKE_GT),				\
+	GEN_FW_RANGE(0x8300, 0x84ff, FORCEWAKE_RENDER),				\
+	GEN_FW_RANGE(0x8500, 0x8cff, FORCEWAKE_GT), /*				\
+		0x8500 - 0x87ff: gt						\
+		0x8800 - 0x8c7f: reserved					\
+		0x8c80 - 0x8cff: gt (DG2 only) */				\
+	GEN_FW_RANGE(0x8d00, 0x8fff, FORCEWAKE_RENDER), /*			\
+		0x8d00 - 0x8dff: render (DG2 only)				\
+		0x8e00 - 0x8fff: reserved */					\
+	GEN_FW_RANGE(0x9000, 0x94cf, FORCEWAKE_GT), /*				\
+		0x9000 - 0x947f: gt						\
+		0x9480 - 0x94cf: reserved */					\
+	GEN_FW_RANGE(0x94d0, 0x955f, FORCEWAKE_RENDER),				\
+	GEN_FW_RANGE(0x9560, 0x967f, 0), /*					\
+		0x9560 - 0x95ff: always on					\
+		0x9600 - 0x967f: reserved */					\
+	GEN_FW_RANGE(0x9680, 0x97ff, FORCEWAKE_RENDER), /*			\
+		0x9680 - 0x96ff: render (DG2 only)				\
+		0x9700 - 0x97ff: reserved */					\
+	GEN_FW_RANGE(0x9800, 0xcfff, FORCEWAKE_GT), /*				\
+		0x9800 - 0xb4ff: gt						\
+		0xb500 - 0xbfff: reserved					\
+		0xc000 - 0xcfff: gt */						\
+	GEN_FW_RANGE(0xd000, 0xd7ff, 0),					\
+	GEN_FW_RANGE(0xd800, 0xd87f, FW_RANGE_D800),			\
+	GEN_FW_RANGE(0xd880, 0xdbff, FORCEWAKE_GT),				\
+	GEN_FW_RANGE(0xdc00, 0xdcff, FORCEWAKE_RENDER),				\
+	GEN_FW_RANGE(0xdd00, 0xde7f, FORCEWAKE_GT), /*				\
+		0xdd00 - 0xddff: gt						\
+		0xde00 - 0xde7f: reserved */					\
+	GEN_FW_RANGE(0xde80, 0xe8ff, FORCEWAKE_RENDER), /*			\
+		0xde80 - 0xdfff: render						\
+		0xe000 - 0xe0ff: reserved					\
+		0xe100 - 0xe8ff: render */					\
+	GEN_FW_RANGE(0xe900, 0xffff, FORCEWAKE_GT), /*				\
+		0xe900 - 0xe9ff: gt						\
+		0xea00 - 0xefff: reserved					\
+		0xf000 - 0xffff: gt */						\
+	GEN_FW_RANGE(0x10000, 0x12fff, 0), /*					\
+		0x10000 - 0x11fff: reserved					\
+		0x12000 - 0x127ff: always on					\
+		0x12800 - 0x12fff: reserved */					\
+	GEN_FW_RANGE(0x13000, 0x131ff, FORCEWAKE_MEDIA_VDBOX0), /* DG2 only */	\
+	GEN_FW_RANGE(0x13200, 0x13fff, FORCEWAKE_MEDIA_VDBOX2), /*		\
+		0x13200 - 0x133ff: VD2 (DG2 only)				\
+		0x13400 - 0x13fff: reserved */					\
+	GEN_FW_RANGE(0x14000, 0x141ff, FORCEWAKE_MEDIA_VDBOX0), /* XEHPSDV only */	\
+	GEN_FW_RANGE(0x14200, 0x143ff, FORCEWAKE_MEDIA_VDBOX2), /* XEHPSDV only */	\
+	GEN_FW_RANGE(0x14400, 0x145ff, FORCEWAKE_MEDIA_VDBOX4), /* XEHPSDV only */	\
+	GEN_FW_RANGE(0x14600, 0x147ff, FORCEWAKE_MEDIA_VDBOX6), /* XEHPSDV only */	\
+	GEN_FW_RANGE(0x14800, 0x14fff, FORCEWAKE_RENDER),			\
+	GEN_FW_RANGE(0x15000, 0x16dff, FORCEWAKE_GT), /*			\
+		0x15000 - 0x15fff: gt (DG2 only)				\
+		0x16000 - 0x16dff: reserved */					\
+	GEN_FW_RANGE(0x16e00, 0x1ffff, FORCEWAKE_RENDER),			\
+	GEN_FW_RANGE(0x20000, 0x21fff, FORCEWAKE_MEDIA_VDBOX0), /*		\
+		0x20000 - 0x20fff: VD0 (XEHPSDV only)				\
+		0x21000 - 0x21fff: reserved */					\
+	GEN_FW_RANGE(0x22000, 0x23fff, FORCEWAKE_GT),				\
+	GEN_FW_RANGE(0x24000, 0x2417f, 0), /*					\
+		0x24000 - 0x2407f: always on					\
+		0x24080 - 0x2417f: reserved */					\
+	GEN_FW_RANGE(0x24180, 0x249ff, FORCEWAKE_GT), /*			\
+		0x24180 - 0x241ff: gt						\
+		0x24200 - 0x249ff: reserved */					\
+	GEN_FW_RANGE(0x24a00, 0x251ff, FORCEWAKE_RENDER), /*			\
+		0x24a00 - 0x24a7f: render					\
+		0x24a80 - 0x251ff: reserved */					\
+	GEN_FW_RANGE(0x25200, 0x25fff, FORCEWAKE_GT), /*			\
+		0x25200 - 0x252ff: gt						\
+		0x25300 - 0x25fff: reserved */					\
+	GEN_FW_RANGE(0x26000, 0x2ffff, FORCEWAKE_RENDER), /*			\
+		0x26000 - 0x27fff: render					\
+		0x28000 - 0x29fff: reserved					\
+		0x2a000 - 0x2ffff: undocumented */				\
+	GEN_FW_RANGE(0x30000, 0x3ffff, FORCEWAKE_GT),				\
+	GEN_FW_RANGE(0x40000, 0x1bffff, 0),					\
+	GEN_FW_RANGE(0x1c0000, 0x1c3fff, FORCEWAKE_MEDIA_VDBOX0), /*		\
+		0x1c0000 - 0x1c2bff: VD0					\
+		0x1c2c00 - 0x1c2cff: reserved					\
+		0x1c2d00 - 0x1c2dff: VD0					\
+		0x1c2e00 - 0x1c3eff: VD0 (DG2 only)				\
+		0x1c3f00 - 0x1c3fff: VD0 */					\
+	GEN_FW_RANGE(0x1c4000, 0x1c7fff, FORCEWAKE_MEDIA_VDBOX1), /*		\
+		0x1c4000 - 0x1c6bff: VD1					\
+		0x1c6c00 - 0x1c6cff: reserved					\
+		0x1c6d00 - 0x1c6dff: VD1					\
+		0x1c6e00 - 0x1c7fff: reserved */				\
+	GEN_FW_RANGE(0x1c8000, 0x1cbfff, FORCEWAKE_MEDIA_VEBOX0), /*		\
+		0x1c8000 - 0x1ca0ff: VE0					\
+		0x1ca100 - 0x1cbfff: reserved */				\
+	GEN_FW_RANGE(0x1cc000, 0x1ccfff, FORCEWAKE_MEDIA_VDBOX0),		\
+	GEN_FW_RANGE(0x1cd000, 0x1cdfff, FORCEWAKE_MEDIA_VDBOX2),		\
+	GEN_FW_RANGE(0x1ce000, 0x1cefff, FORCEWAKE_MEDIA_VDBOX4),		\
+	GEN_FW_RANGE(0x1cf000, 0x1cffff, FORCEWAKE_MEDIA_VDBOX6),		\
+	GEN_FW_RANGE(0x1d0000, 0x1d3fff, FORCEWAKE_MEDIA_VDBOX2), /*		\
+		0x1d0000 - 0x1d2bff: VD2					\
+		0x1d2c00 - 0x1d2cff: reserved					\
+		0x1d2d00 - 0x1d2dff: VD2					\
+		0x1d2e00 - 0x1d3dff: VD2 (DG2 only)				\
+		0x1d3e00 - 0x1d3eff: reserved					\
+		0x1d3f00 - 0x1d3fff: VD2 */					\
+	GEN_FW_RANGE(0x1d4000, 0x1d7fff, FORCEWAKE_MEDIA_VDBOX3), /*		\
+		0x1d4000 - 0x1d6bff: VD3					\
+		0x1d6c00 - 0x1d6cff: reserved					\
+		0x1d6d00 - 0x1d6dff: VD3					\
+		0x1d6e00 - 0x1d7fff: reserved */				\
+	GEN_FW_RANGE(0x1d8000, 0x1dffff, FORCEWAKE_MEDIA_VEBOX1), /*		\
+		0x1d8000 - 0x1da0ff: VE1					\
+		0x1da100 - 0x1dffff: reserved */				\
+	GEN_FW_RANGE(0x1e0000, 0x1e3fff, FORCEWAKE_MEDIA_VDBOX4), /*		\
+		0x1e0000 - 0x1e2bff: VD4					\
+		0x1e2c00 - 0x1e2cff: reserved					\
+		0x1e2d00 - 0x1e2dff: VD4					\
+		0x1e2e00 - 0x1e3eff: reserved					\
+		0x1e3f00 - 0x1e3fff: VD4 */					\
+	GEN_FW_RANGE(0x1e4000, 0x1e7fff, FORCEWAKE_MEDIA_VDBOX5), /*		\
+		0x1e4000 - 0x1e6bff: VD5					\
+		0x1e6c00 - 0x1e6cff: reserved					\
+		0x1e6d00 - 0x1e6dff: VD5					\
+		0x1e6e00 - 0x1e7fff: reserved */				\
+	GEN_FW_RANGE(0x1e8000, 0x1effff, FORCEWAKE_MEDIA_VEBOX2), /*		\
+		0x1e8000 - 0x1ea0ff: VE2					\
+		0x1ea100 - 0x1effff: reserved */				\
+	GEN_FW_RANGE(0x1f0000, 0x1f3fff, FORCEWAKE_MEDIA_VDBOX6), /*		\
+		0x1f0000 - 0x1f2bff: VD6					\
+		0x1f2c00 - 0x1f2cff: reserved					\
+		0x1f2d00 - 0x1f2dff: VD6					\
+		0x1f2e00 - 0x1f3eff: reserved					\
+		0x1f3f00 - 0x1f3fff: VD6 */					\
+	GEN_FW_RANGE(0x1f4000, 0x1f7fff, FORCEWAKE_MEDIA_VDBOX7), /*		\
+		0x1f4000 - 0x1f6bff: VD7					\
+		0x1f6c00 - 0x1f6cff: reserved					\
+		0x1f6d00 - 0x1f6dff: VD7					\
+		0x1f6e00 - 0x1f7fff: reserved */				\
+	GEN_FW_RANGE(0x1f8000, 0x1fa0ff, FORCEWAKE_MEDIA_VEBOX3),
+
+static const struct intel_forcewake_range __xehp_fw_ranges[] = {
+	XEHP_FWRANGES(FORCEWAKE_GT)
+};
+
+static const struct intel_forcewake_range __dg2_fw_ranges[] = {
+	XEHP_FWRANGES(FORCEWAKE_RENDER)
+};
+
+static const struct intel_forcewake_range __pvc_fw_ranges[] = {
+	GEN_FW_RANGE(0x0, 0xaff, 0),
+	GEN_FW_RANGE(0xb00, 0xbff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0xc00, 0xfff, 0),
+	GEN_FW_RANGE(0x1000, 0x1fff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x2000, 0x26ff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x2700, 0x2fff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0x3000, 0x3fff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x4000, 0x813f, FORCEWAKE_GT), /*
+		0x4000 - 0x4aff: gt
+		0x4b00 - 0x4fff: reserved
+		0x5000 - 0x51ff: gt
+		0x5200 - 0x52ff: reserved
+		0x5300 - 0x53ff: gt
+		0x5400 - 0x7fff: reserved
+		0x8000 - 0x813f: gt */
+	GEN_FW_RANGE(0x8140, 0x817f, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x8180, 0x81ff, 0),
+	GEN_FW_RANGE(0x8200, 0x94cf, FORCEWAKE_GT), /*
+		0x8200 - 0x82ff: gt
+		0x8300 - 0x84ff: reserved
+		0x8500 - 0x887f: gt
+		0x8880 - 0x8a7f: reserved
+		0x8a80 - 0x8aff: gt
+		0x8b00 - 0x8fff: reserved
+		0x9000 - 0x947f: gt
+		0x9480 - 0x94cf: reserved */
+	GEN_FW_RANGE(0x94d0, 0x955f, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0x9560, 0x967f, 0), /*
+		0x9560 - 0x95ff: always on
+		0x9600 - 0x967f: reserved */
+	GEN_FW_RANGE(0x9680, 0x97ff, FORCEWAKE_RENDER), /*
+		0x9680 - 0x96ff: render
+		0x9700 - 0x97ff: reserved */
+	GEN_FW_RANGE(0x9800, 0xcfff, FORCEWAKE_GT), /*
+		0x9800 - 0xb4ff: gt
+		0xb500 - 0xbfff: reserved
+		0xc000 - 0xcfff: gt */
+	GEN_FW_RANGE(0xd000, 0xd3ff, 0),
+	GEN_FW_RANGE(0xd400, 0xdbff, FORCEWAKE_GT),
+	GEN_FW_RANGE(0xdc00, 0xdcff, FORCEWAKE_RENDER),
+	GEN_FW_RANGE(0xdd00, 0xde7f, FORCEWAKE_GT), /*
+		0xdd00 - 0xddff: gt
+		0xde00 - 0xde7f: reserved */
+	GEN_FW_RANGE(0xde80, 0xe8ff, FORCEWAKE_RENDER), /*
+		0xde80 - 0xdeff: render
+		0xdf00 - 0xe1ff: reserved
+		0xe200 - 0xe7ff: render
+		0xe800 - 0xe8ff: reserved */
+	GEN_FW_RANGE(0xe900, 0x11fff, FORCEWAKE_GT), /*
+		 0xe900 -  0xe9ff: gt
+		 0xea00 -  0xebff: reserved
+		 0xec00 -  0xffff: gt
+		0x10000 - 0x11fff: reserved */
+	GEN_FW_RANGE(0x12000, 0x12fff, 0), /*
+		0x12000 - 0x127ff: always on
+		0x12800 - 0x12fff: reserved */
+	GEN_FW_RANGE(0x13000, 0x23fff, FORCEWAKE_GT), /*
+		0x13000 - 0x135ff: gt
+		0x13600 - 0x147ff: reserved
+		0x14800 - 0x153ff: gt
+		0x15400 - 0x19fff: reserved
+		0x1a000 - 0x1ffff: gt
+		0x20000 - 0x21fff: reserved
+		0x22000 - 0x23fff: gt */
+	GEN_FW_RANGE(0x24000, 0x2417f, 0), /*
+		24000 - 0x2407f: always on
+		24080 - 0x2417f: reserved */
+	GEN_FW_RANGE(0x24180, 0x3ffff, FORCEWAKE_GT), /*
+		0x24180 - 0x241ff: gt
+		0x24200 - 0x251ff: reserved
+		0x25200 - 0x252ff: gt
+		0x25300 - 0x25fff: reserved
+		0x26000 - 0x27fff: gt
+		0x28000 - 0x2ffff: reserved
+		0x30000 - 0x3ffff: gt */
+	GEN_FW_RANGE(0x40000, 0x1bffff, 0),
+	GEN_FW_RANGE(0x1c0000, 0x1c3fff, FORCEWAKE_MEDIA_VDBOX0), /*
+		0x1c0000 - 0x1c2bff: VD0
+		0x1c2c00 - 0x1c2cff: reserved
+		0x1c2d00 - 0x1c2dff: VD0
+		0x1c2e00 - 0x1c3eff: reserved
+		0x1c3f00 - 0x1c3fff: VD0 */
+	GEN_FW_RANGE(0x1c4000, 0x1cffff, FORCEWAKE_MEDIA_VDBOX1), /*
+		0x1c4000 - 0x1c6aff: VD1
+		0x1c6b00 - 0x1c7eff: reserved
+		0x1c7f00 - 0x1c7fff: VD1
+		0x1c8000 - 0x1cffff: reserved */
+	GEN_FW_RANGE(0x1d0000, 0x23ffff, FORCEWAKE_MEDIA_VDBOX2), /*
+		0x1d0000 - 0x1d2aff: VD2
+		0x1d2b00 - 0x1d3eff: reserved
+		0x1d3f00 - 0x1d3fff: VD2
+		0x1d4000 - 0x23ffff: reserved */
+	GEN_FW_RANGE(0x240000, 0x3dffff, 0),
+	GEN_FW_RANGE(0x3e0000, 0x3effff, FORCEWAKE_GT),
+};
+
+static void
+ilk_dummy_write(struct intel_uncore *uncore)
+{
+	/* WaIssueDummyWriteToWakeupFromRC6:ilk Issue a dummy write to wake up
+	 * the chip from rc6 before touching it for real. MI_MODE is masked,
+	 * hence harmless to write 0 into. */
+	__raw_uncore_write32(uncore, RING_MI_MODE(RENDER_RING_BASE), 0);
+}
+
+static void
+__unclaimed_reg_debug(struct intel_uncore *uncore,
+		      const i915_reg_t reg,
+		      const bool read)
+{
+	if (drm_WARN(&uncore->i915->drm,
+		     check_for_unclaimed_mmio(uncore),
+		     "Unclaimed %s register 0x%x\n",
+		     read ? "read from" : "write to",
+		     i915_mmio_reg_offset(reg)))
+		/* Only report the first N failures */
+		uncore->i915->params.mmio_debug--;
+}
+
+static void
+__unclaimed_previous_reg_debug(struct intel_uncore *uncore,
+			       const i915_reg_t reg,
+			       const bool read)
+{
+	if (check_for_unclaimed_mmio(uncore))
+		drm_dbg(&uncore->i915->drm,
+			"Unclaimed access detected before %s register 0x%x\n",
+			read ? "read from" : "write to",
+			i915_mmio_reg_offset(reg));
+}
+
+static inline void
+unclaimed_reg_debug(struct intel_uncore *uncore,
+		    const i915_reg_t reg,
+		    const bool read,
+		    const bool before)
+{
+	if (likely(!uncore->i915->params.mmio_debug) || !uncore->debug)
+		return;
+
+	/* interrupts are disabled and re-enabled around uncore->lock usage */
+	lockdep_assert_held(&uncore->lock);
+
+	if (before) {
+		spin_lock(&uncore->debug->lock);
+		__unclaimed_previous_reg_debug(uncore, reg, read);
+	} else {
+		__unclaimed_reg_debug(uncore, reg, read);
+		spin_unlock(&uncore->debug->lock);
+	}
+}
+
+#define __vgpu_read(x) \
+static u##x \
+vgpu_read##x(struct intel_uncore *uncore, i915_reg_t reg, bool trace) { \
+	u##x val = __raw_uncore_read##x(uncore, reg); \
+	trace_i915_reg_rw(false, reg, val, sizeof(val), trace); \
+	return val; \
+}
+__vgpu_read(8)
+__vgpu_read(16)
+__vgpu_read(32)
+__vgpu_read(64)
+
+#define GEN2_READ_HEADER(x) \
+	u##x val = 0; \
+	assert_rpm_wakelock_held(uncore->rpm);
+
+#define GEN2_READ_FOOTER \
+	trace_i915_reg_rw(false, reg, val, sizeof(val), trace); \
+	return val
+
+#define __gen2_read(x) \
+static u##x \
+gen2_read##x(struct intel_uncore *uncore, i915_reg_t reg, bool trace) { \
+	GEN2_READ_HEADER(x); \
+	val = __raw_uncore_read##x(uncore, reg); \
+	GEN2_READ_FOOTER; \
+}
+
+#define __gen5_read(x) \
+static u##x \
+gen5_read##x(struct intel_uncore *uncore, i915_reg_t reg, bool trace) { \
+	GEN2_READ_HEADER(x); \
+	ilk_dummy_write(uncore); \
+	val = __raw_uncore_read##x(uncore, reg); \
+	GEN2_READ_FOOTER; \
+}
+
+__gen5_read(8)
+__gen5_read(16)
+__gen5_read(32)
+__gen5_read(64)
+__gen2_read(8)
+__gen2_read(16)
+__gen2_read(32)
+__gen2_read(64)
+
+#undef __gen5_read
+#undef __gen2_read
+
+#undef GEN2_READ_FOOTER
+#undef GEN2_READ_HEADER
+
+#define GEN6_READ_HEADER(x) \
+	u32 offset = i915_mmio_reg_offset(reg); \
+	unsigned long irqflags; \
+	u##x val = 0; \
+	assert_rpm_wakelock_held(uncore->rpm); \
+	spin_lock_irqsave(&uncore->lock, irqflags); \
+	unclaimed_reg_debug(uncore, reg, true, true)
+
+#define GEN6_READ_FOOTER \
+	unclaimed_reg_debug(uncore, reg, true, false); \
+	spin_unlock_irqrestore(&uncore->lock, irqflags); \
+	trace_i915_reg_rw(false, reg, val, sizeof(val), trace); \
+	return val
+
+static noinline void ___force_wake_auto(struct intel_uncore *uncore,
+					enum forcewake_domains fw_domains)
+{
+	struct intel_uncore_forcewake_domain *domain;
+	unsigned int tmp;
+
+	GEM_BUG_ON(fw_domains & ~uncore->fw_domains);
+
+	for_each_fw_domain_masked(domain, fw_domains, uncore, tmp)
+		fw_domain_arm_timer(domain);
+
+	fw_domains_get(uncore, fw_domains);
+}
+
+static inline void __force_wake_auto(struct intel_uncore *uncore,
+				     enum forcewake_domains fw_domains)
+{
+	GEM_BUG_ON(!fw_domains);
+
+	/* Turn on all requested but inactive supported forcewake domains. */
+	fw_domains &= uncore->fw_domains;
+	fw_domains &= ~uncore->fw_domains_active;
+
+	if (fw_domains)
+		___force_wake_auto(uncore, fw_domains);
+}
+
+#define __gen_fwtable_read(x) \
+static u##x \
+fwtable_read##x(struct intel_uncore *uncore, i915_reg_t reg, bool trace) \
+{ \
+	enum forcewake_domains fw_engine; \
+	GEN6_READ_HEADER(x); \
+	fw_engine = __fwtable_reg_read_fw_domains(uncore, offset); \
+	if (fw_engine) \
+		__force_wake_auto(uncore, fw_engine); \
+	val = __raw_uncore_read##x(uncore, reg); \
+	GEN6_READ_FOOTER; \
+}
+
+static enum forcewake_domains
+fwtable_reg_read_fw_domains(struct intel_uncore *uncore, i915_reg_t reg) {
+	return __fwtable_reg_read_fw_domains(uncore, i915_mmio_reg_offset(reg));
+}
+
+__gen_fwtable_read(8)
+__gen_fwtable_read(16)
+__gen_fwtable_read(32)
+__gen_fwtable_read(64)
+
+#undef __gen_fwtable_read
+#undef GEN6_READ_FOOTER
+#undef GEN6_READ_HEADER
+
+#define GEN2_WRITE_HEADER \
+	trace_i915_reg_rw(true, reg, val, sizeof(val), trace); \
+	assert_rpm_wakelock_held(uncore->rpm); \
+
+#define GEN2_WRITE_FOOTER
+
+#define __gen2_write(x) \
+static void \
+gen2_write##x(struct intel_uncore *uncore, i915_reg_t reg, u##x val, bool trace) { \
+	GEN2_WRITE_HEADER; \
+	__raw_uncore_write##x(uncore, reg, val); \
+	GEN2_WRITE_FOOTER; \
+}
+
+#define __gen5_write(x) \
+static void \
+gen5_write##x(struct intel_uncore *uncore, i915_reg_t reg, u##x val, bool trace) { \
+	GEN2_WRITE_HEADER; \
+	ilk_dummy_write(uncore); \
+	__raw_uncore_write##x(uncore, reg, val); \
+	GEN2_WRITE_FOOTER; \
+}
+
+__gen5_write(8)
+__gen5_write(16)
+__gen5_write(32)
+__gen2_write(8)
+__gen2_write(16)
+__gen2_write(32)
+
+#undef __gen5_write
+#undef __gen2_write
+
+#undef GEN2_WRITE_FOOTER
+#undef GEN2_WRITE_HEADER
+
+#define GEN6_WRITE_HEADER \
+	u32 offset = i915_mmio_reg_offset(reg); \
+	unsigned long irqflags; \
+	trace_i915_reg_rw(true, reg, val, sizeof(val), trace); \
+	assert_rpm_wakelock_held(uncore->rpm); \
+	spin_lock_irqsave(&uncore->lock, irqflags); \
+	unclaimed_reg_debug(uncore, reg, false, true)
+
+#define GEN6_WRITE_FOOTER \
+	unclaimed_reg_debug(uncore, reg, false, false); \
+	spin_unlock_irqrestore(&uncore->lock, irqflags)
+
+#define __gen6_write(x) \
+static void \
+gen6_write##x(struct intel_uncore *uncore, i915_reg_t reg, u##x val, bool trace) { \
+	GEN6_WRITE_HEADER; \
+	if (NEEDS_FORCE_WAKE(offset)) \
+		__gen6_gt_wait_for_fifo(uncore); \
+	__raw_uncore_write##x(uncore, reg, val); \
+	GEN6_WRITE_FOOTER; \
+}
+__gen6_write(8)
+__gen6_write(16)
+__gen6_write(32)
+
+#define __gen_fwtable_write(x) \
+static void \
+fwtable_write##x(struct intel_uncore *uncore, i915_reg_t reg, u##x val, bool trace) { \
+	enum forcewake_domains fw_engine; \
+	GEN6_WRITE_HEADER; \
+	fw_engine = __fwtable_reg_write_fw_domains(uncore, offset); \
+	if (fw_engine) \
+		__force_wake_auto(uncore, fw_engine); \
+	__raw_uncore_write##x(uncore, reg, val); \
+	GEN6_WRITE_FOOTER; \
+}
+
+static enum forcewake_domains
+fwtable_reg_write_fw_domains(struct intel_uncore *uncore, i915_reg_t reg)
+{
+	return __fwtable_reg_write_fw_domains(uncore, i915_mmio_reg_offset(reg));
+}
+
+__gen_fwtable_write(8)
+__gen_fwtable_write(16)
+__gen_fwtable_write(32)
+
+#undef __gen_fwtable_write
+#undef GEN6_WRITE_FOOTER
+#undef GEN6_WRITE_HEADER
+
+#define __vgpu_write(x) \
+static void \
+vgpu_write##x(struct intel_uncore *uncore, i915_reg_t reg, u##x val, bool trace) { \
+	trace_i915_reg_rw(true, reg, val, sizeof(val), trace); \
+	__raw_uncore_write##x(uncore, reg, val); \
+}
+__vgpu_write(8)
+__vgpu_write(16)
+__vgpu_write(32)
+
+#define ASSIGN_RAW_WRITE_MMIO_VFUNCS(uncore, x) \
+do { \
+	(uncore)->funcs.mmio_writeb = x##_write8; \
+	(uncore)->funcs.mmio_writew = x##_write16; \
+	(uncore)->funcs.mmio_writel = x##_write32; \
+} while (0)
+
+#define ASSIGN_RAW_READ_MMIO_VFUNCS(uncore, x) \
+do { \
+	(uncore)->funcs.mmio_readb = x##_read8; \
+	(uncore)->funcs.mmio_readw = x##_read16; \
+	(uncore)->funcs.mmio_readl = x##_read32; \
+	(uncore)->funcs.mmio_readq = x##_read64; \
+} while (0)
+
+#define ASSIGN_WRITE_MMIO_VFUNCS(uncore, x) \
+do { \
+	ASSIGN_RAW_WRITE_MMIO_VFUNCS((uncore), x); \
+	(uncore)->funcs.write_fw_domains = x##_reg_write_fw_domains; \
+} while (0)
+
+#define ASSIGN_READ_MMIO_VFUNCS(uncore, x) \
+do { \
+	ASSIGN_RAW_READ_MMIO_VFUNCS(uncore, x); \
+	(uncore)->funcs.read_fw_domains = x##_reg_read_fw_domains; \
+} while (0)
+
+static int __fw_domain_init(struct intel_uncore *uncore,
+			    enum forcewake_domain_id domain_id,
+			    i915_reg_t reg_set,
+			    i915_reg_t reg_ack)
+{
+	struct intel_uncore_forcewake_domain *d;
+
+	GEM_BUG_ON(domain_id >= FW_DOMAIN_ID_COUNT);
+	GEM_BUG_ON(uncore->fw_domain[domain_id]);
+
+	if (i915_inject_probe_failure(uncore->i915))
+		return -ENOMEM;
+
+	d = kzalloc(sizeof(*d), GFP_KERNEL);
+	if (!d)
+		return -ENOMEM;
+
+	drm_WARN_ON(&uncore->i915->drm, !i915_mmio_reg_valid(reg_set));
+	drm_WARN_ON(&uncore->i915->drm, !i915_mmio_reg_valid(reg_ack));
+
+	d->uncore = uncore;
+	d->wake_count = 0;
+	d->reg_set = uncore->regs + i915_mmio_reg_offset(reg_set) + uncore->gsi_offset;
+	d->reg_ack = uncore->regs + i915_mmio_reg_offset(reg_ack) + uncore->gsi_offset;
+
+	d->id = domain_id;
+
+	BUILD_BUG_ON(FORCEWAKE_RENDER != (1 << FW_DOMAIN_ID_RENDER));
+	BUILD_BUG_ON(FORCEWAKE_GT != (1 << FW_DOMAIN_ID_GT));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA != (1 << FW_DOMAIN_ID_MEDIA));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VDBOX0 != (1 << FW_DOMAIN_ID_MEDIA_VDBOX0));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VDBOX1 != (1 << FW_DOMAIN_ID_MEDIA_VDBOX1));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VDBOX2 != (1 << FW_DOMAIN_ID_MEDIA_VDBOX2));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VDBOX3 != (1 << FW_DOMAIN_ID_MEDIA_VDBOX3));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VDBOX4 != (1 << FW_DOMAIN_ID_MEDIA_VDBOX4));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VDBOX5 != (1 << FW_DOMAIN_ID_MEDIA_VDBOX5));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VDBOX6 != (1 << FW_DOMAIN_ID_MEDIA_VDBOX6));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VDBOX7 != (1 << FW_DOMAIN_ID_MEDIA_VDBOX7));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VEBOX0 != (1 << FW_DOMAIN_ID_MEDIA_VEBOX0));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VEBOX1 != (1 << FW_DOMAIN_ID_MEDIA_VEBOX1));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VEBOX2 != (1 << FW_DOMAIN_ID_MEDIA_VEBOX2));
+	BUILD_BUG_ON(FORCEWAKE_MEDIA_VEBOX3 != (1 << FW_DOMAIN_ID_MEDIA_VEBOX3));
+
+	d->mask = BIT(domain_id);
+
+	hrtimer_init(&d->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+	d->timer.function = intel_uncore_fw_release_timer;
+
+	uncore->fw_domains |= BIT(domain_id);
+
+	fw_domain_reset(d);
+
+	uncore->fw_domain[domain_id] = d;
+
+	return 0;
+}
+
+static void fw_domain_fini(struct intel_uncore *uncore,
+			   enum forcewake_domain_id domain_id)
+{
+	struct intel_uncore_forcewake_domain *d;
+
+	GEM_BUG_ON(domain_id >= FW_DOMAIN_ID_COUNT);
+
+	d = fetch_and_zero(&uncore->fw_domain[domain_id]);
+	if (!d)
+		return;
+
+	uncore->fw_domains &= ~BIT(domain_id);
+	drm_WARN_ON(&uncore->i915->drm, d->wake_count);
+	drm_WARN_ON(&uncore->i915->drm, hrtimer_cancel(&d->timer));
+	kfree(d);
+}
+
+static void intel_uncore_fw_domains_fini(struct intel_uncore *uncore)
+{
+	struct intel_uncore_forcewake_domain *d;
+	int tmp;
+
+	for_each_fw_domain(d, uncore, tmp)
+		fw_domain_fini(uncore, d->id);
+}
+
+static const struct intel_uncore_fw_get uncore_get_fallback = {
+	.force_wake_get = fw_domains_get_with_fallback
+};
+
+static const struct intel_uncore_fw_get uncore_get_normal = {
+	.force_wake_get = fw_domains_get_normal,
+};
+
+static const struct intel_uncore_fw_get uncore_get_thread_status = {
+	.force_wake_get = fw_domains_get_with_thread_status
+};
+
+static int intel_uncore_fw_domains_init(struct intel_uncore *uncore)
+{
+	struct drm_i915_private *i915 = uncore->i915;
+	int ret = 0;
+
+	GEM_BUG_ON(!intel_uncore_has_forcewake(uncore));
+
+#define fw_domain_init(uncore__, id__, set__, ack__) \
+	(ret ?: (ret = __fw_domain_init((uncore__), (id__), (set__), (ack__))))
+
+	if (GRAPHICS_VER(i915) >= 11) {
+		/* we'll prune the domains of missing engines later */
+		intel_engine_mask_t emask = RUNTIME_INFO(i915)->platform_engine_mask;
+		int i;
+
+		uncore->fw_get_funcs = &uncore_get_fallback;
+		fw_domain_init(uncore, FW_DOMAIN_ID_RENDER,
+			       FORCEWAKE_RENDER_GEN9,
+			       FORCEWAKE_ACK_RENDER_GEN9);
+		fw_domain_init(uncore, FW_DOMAIN_ID_GT,
+			       FORCEWAKE_GT_GEN9,
+			       FORCEWAKE_ACK_GT_GEN9);
+
+		for (i = 0; i < I915_MAX_VCS; i++) {
+			if (!__HAS_ENGINE(emask, _VCS(i)))
+				continue;
+
+			fw_domain_init(uncore, FW_DOMAIN_ID_MEDIA_VDBOX0 + i,
+				       FORCEWAKE_MEDIA_VDBOX_GEN11(i),
+				       FORCEWAKE_ACK_MEDIA_VDBOX_GEN11(i));
+		}
+		for (i = 0; i < I915_MAX_VECS; i++) {
+			if (!__HAS_ENGINE(emask, _VECS(i)))
+				continue;
+
+			fw_domain_init(uncore, FW_DOMAIN_ID_MEDIA_VEBOX0 + i,
+				       FORCEWAKE_MEDIA_VEBOX_GEN11(i),
+				       FORCEWAKE_ACK_MEDIA_VEBOX_GEN11(i));
+		}
+	} else if (IS_GRAPHICS_VER(i915, 9, 10)) {
+		uncore->fw_get_funcs = &uncore_get_fallback;
+		fw_domain_init(uncore, FW_DOMAIN_ID_RENDER,
+			       FORCEWAKE_RENDER_GEN9,
+			       FORCEWAKE_ACK_RENDER_GEN9);
+		fw_domain_init(uncore, FW_DOMAIN_ID_GT,
+			       FORCEWAKE_GT_GEN9,
+			       FORCEWAKE_ACK_GT_GEN9);
+		fw_domain_init(uncore, FW_DOMAIN_ID_MEDIA,
+			       FORCEWAKE_MEDIA_GEN9, FORCEWAKE_ACK_MEDIA_GEN9);
+	} else if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) {
+		uncore->fw_get_funcs = &uncore_get_normal;
+		fw_domain_init(uncore, FW_DOMAIN_ID_RENDER,
+			       FORCEWAKE_VLV, FORCEWAKE_ACK_VLV);
+		fw_domain_init(uncore, FW_DOMAIN_ID_MEDIA,
+			       FORCEWAKE_MEDIA_VLV, FORCEWAKE_ACK_MEDIA_VLV);
+	} else if (IS_HASWELL(i915) || IS_BROADWELL(i915)) {
+		uncore->fw_get_funcs = &uncore_get_thread_status;
+		fw_domain_init(uncore, FW_DOMAIN_ID_RENDER,
+			       FORCEWAKE_MT, FORCEWAKE_ACK_HSW);
+	} else if (IS_IVYBRIDGE(i915)) {
+		u32 ecobus;
+
+		/* IVB configs may use multi-threaded forcewake */
+
+		/* A small trick here - if the bios hasn't configured
+		 * MT forcewake, and if the device is in RC6, then
+		 * force_wake_mt_get will not wake the device and the
+		 * ECOBUS read will return zero. Which will be
+		 * (correctly) interpreted by the test below as MT
+		 * forcewake being disabled.
+		 */
+		uncore->fw_get_funcs = &uncore_get_thread_status;
+
+		/* We need to init first for ECOBUS access and then
+		 * determine later if we want to reinit, in case of MT access is
+		 * not working. In this stage we don't know which flavour this
+		 * ivb is, so it is better to reset also the gen6 fw registers
+		 * before the ecobus check.
+		 */
+
+		__raw_uncore_write32(uncore, FORCEWAKE, 0);
+		__raw_posting_read(uncore, ECOBUS);
+
+		ret = __fw_domain_init(uncore, FW_DOMAIN_ID_RENDER,
+				       FORCEWAKE_MT, FORCEWAKE_MT_ACK);
+		if (ret)
+			goto out;
+
+		spin_lock_irq(&uncore->lock);
+		fw_domains_get_with_thread_status(uncore, FORCEWAKE_RENDER);
+		ecobus = __raw_uncore_read32(uncore, ECOBUS);
+		fw_domains_put(uncore, FORCEWAKE_RENDER);
+		spin_unlock_irq(&uncore->lock);
+
+		if (!(ecobus & FORCEWAKE_MT_ENABLE)) {
+			drm_info(&i915->drm, "No MT forcewake available on Ivybridge, this can result in issues\n");
+			drm_info(&i915->drm, "when using vblank-synced partial screen updates.\n");
+			fw_domain_fini(uncore, FW_DOMAIN_ID_RENDER);
+			fw_domain_init(uncore, FW_DOMAIN_ID_RENDER,
+				       FORCEWAKE, FORCEWAKE_ACK);
+		}
+	} else if (GRAPHICS_VER(i915) == 6) {
+		uncore->fw_get_funcs = &uncore_get_thread_status;
+		fw_domain_init(uncore, FW_DOMAIN_ID_RENDER,
+			       FORCEWAKE, FORCEWAKE_ACK);
+	}
+
+#undef fw_domain_init
+
+	/* All future platforms are expected to require complex power gating */
+	drm_WARN_ON(&i915->drm, !ret && uncore->fw_domains == 0);
+
+out:
+	if (ret)
+		intel_uncore_fw_domains_fini(uncore);
+
+	return ret;
+}
+
+#define ASSIGN_FW_DOMAINS_TABLE(uncore, d) \
+{ \
+	(uncore)->fw_domains_table = \
+			(struct intel_forcewake_range *)(d); \
+	(uncore)->fw_domains_table_entries = ARRAY_SIZE((d)); \
+}
+
+#define ASSIGN_SHADOW_TABLE(uncore, d) \
+{ \
+	(uncore)->shadowed_reg_table = d; \
+	(uncore)->shadowed_reg_table_entries = ARRAY_SIZE((d)); \
+}
+
+static int i915_pmic_bus_access_notifier(struct notifier_block *nb,
+					 unsigned long action, void *data)
+{
+	struct intel_uncore *uncore = container_of(nb,
+			struct intel_uncore, pmic_bus_access_nb);
+
+	switch (action) {
+	case MBI_PMIC_BUS_ACCESS_BEGIN:
+		/*
+		 * forcewake all now to make sure that we don't need to do a
+		 * forcewake later which on systems where this notifier gets
+		 * called requires the punit to access to the shared pmic i2c
+		 * bus, which will be busy after this notification, leading to:
+		 * "render: timed out waiting for forcewake ack request."
+		 * errors.
+		 *
+		 * The notifier is unregistered during intel_runtime_suspend(),
+		 * so it's ok to access the HW here without holding a RPM
+		 * wake reference -> disable wakeref asserts for the time of
+		 * the access.
+		 */
+		disable_rpm_wakeref_asserts(uncore->rpm);
+		intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+		enable_rpm_wakeref_asserts(uncore->rpm);
+		break;
+	case MBI_PMIC_BUS_ACCESS_END:
+		intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
+		break;
+	}
+
+	return NOTIFY_OK;
+}
+
+static void uncore_unmap_mmio(struct drm_device *drm, void *regs)
+{
+	iounmap(regs);
+}
+
+int intel_uncore_setup_mmio(struct intel_uncore *uncore, phys_addr_t phys_addr)
+{
+	struct drm_i915_private *i915 = uncore->i915;
+	int mmio_size;
+
+	/*
+	 * Before gen4, the registers and the GTT are behind different BARs.
+	 * However, from gen4 onwards, the registers and the GTT are shared
+	 * in the same BAR, so we want to restrict this ioremap from
+	 * clobbering the GTT which we want ioremap_wc instead. Fortunately,
+	 * the register BAR remains the same size for all the earlier
+	 * generations up to Ironlake.
+	 * For dgfx chips register range is expanded to 4MB, and this larger
+	 * range is also used for integrated gpus beginning with Meteor Lake.
+	 */
+	if (IS_DGFX(i915) || GRAPHICS_VER_FULL(i915) >= IP_VER(12, 70))
+		mmio_size = 4 * 1024 * 1024;
+	else if (GRAPHICS_VER(i915) >= 5)
+		mmio_size = 2 * 1024 * 1024;
+	else
+		mmio_size = 512 * 1024;
+
+	uncore->regs = ioremap(phys_addr, mmio_size);
+	if (uncore->regs == NULL) {
+		drm_err(&i915->drm, "failed to map registers\n");
+		return -EIO;
+	}
+
+	return drmm_add_action_or_reset(&i915->drm, uncore_unmap_mmio, uncore->regs);
+}
+
+void intel_uncore_init_early(struct intel_uncore *uncore,
+			     struct intel_gt *gt)
+{
+	spin_lock_init(&uncore->lock);
+	uncore->i915 = gt->i915;
+	uncore->gt = gt;
+	uncore->rpm = &gt->i915->runtime_pm;
+}
+
+static void uncore_raw_init(struct intel_uncore *uncore)
+{
+	GEM_BUG_ON(intel_uncore_has_forcewake(uncore));
+
+	if (intel_vgpu_active(uncore->i915)) {
+		ASSIGN_RAW_WRITE_MMIO_VFUNCS(uncore, vgpu);
+		ASSIGN_RAW_READ_MMIO_VFUNCS(uncore, vgpu);
+	} else if (GRAPHICS_VER(uncore->i915) == 5) {
+		ASSIGN_RAW_WRITE_MMIO_VFUNCS(uncore, gen5);
+		ASSIGN_RAW_READ_MMIO_VFUNCS(uncore, gen5);
+	} else {
+		ASSIGN_RAW_WRITE_MMIO_VFUNCS(uncore, gen2);
+		ASSIGN_RAW_READ_MMIO_VFUNCS(uncore, gen2);
+	}
+}
+
+static int uncore_forcewake_init(struct intel_uncore *uncore)
+{
+	struct drm_i915_private *i915 = uncore->i915;
+	int ret;
+
+	GEM_BUG_ON(!intel_uncore_has_forcewake(uncore));
+
+	ret = intel_uncore_fw_domains_init(uncore);
+	if (ret)
+		return ret;
+	forcewake_early_sanitize(uncore, 0);
+
+	ASSIGN_READ_MMIO_VFUNCS(uncore, fwtable);
+
+	if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 60)) {
+		ASSIGN_FW_DOMAINS_TABLE(uncore, __pvc_fw_ranges);
+		ASSIGN_SHADOW_TABLE(uncore, pvc_shadowed_regs);
+		ASSIGN_WRITE_MMIO_VFUNCS(uncore, fwtable);
+	} else if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 55)) {
+		ASSIGN_FW_DOMAINS_TABLE(uncore, __dg2_fw_ranges);
+		ASSIGN_SHADOW_TABLE(uncore, dg2_shadowed_regs);
+		ASSIGN_WRITE_MMIO_VFUNCS(uncore, fwtable);
+	} else if (GRAPHICS_VER_FULL(i915) >= IP_VER(12, 50)) {
+		ASSIGN_FW_DOMAINS_TABLE(uncore, __xehp_fw_ranges);
+		ASSIGN_SHADOW_TABLE(uncore, gen12_shadowed_regs);
+		ASSIGN_WRITE_MMIO_VFUNCS(uncore, fwtable);
+	} else if (GRAPHICS_VER(i915) >= 12) {
+		ASSIGN_FW_DOMAINS_TABLE(uncore, __gen12_fw_ranges);
+		ASSIGN_SHADOW_TABLE(uncore, gen12_shadowed_regs);
+		ASSIGN_WRITE_MMIO_VFUNCS(uncore, fwtable);
+	} else if (GRAPHICS_VER(i915) == 11) {
+		ASSIGN_FW_DOMAINS_TABLE(uncore, __gen11_fw_ranges);
+		ASSIGN_SHADOW_TABLE(uncore, gen11_shadowed_regs);
+		ASSIGN_WRITE_MMIO_VFUNCS(uncore, fwtable);
+	} else if (IS_GRAPHICS_VER(i915, 9, 10)) {
+		ASSIGN_FW_DOMAINS_TABLE(uncore, __gen9_fw_ranges);
+		ASSIGN_SHADOW_TABLE(uncore, gen8_shadowed_regs);
+		ASSIGN_WRITE_MMIO_VFUNCS(uncore, fwtable);
+	} else if (IS_CHERRYVIEW(i915)) {
+		ASSIGN_FW_DOMAINS_TABLE(uncore, __chv_fw_ranges);
+		ASSIGN_SHADOW_TABLE(uncore, gen8_shadowed_regs);
+		ASSIGN_WRITE_MMIO_VFUNCS(uncore, fwtable);
+	} else if (GRAPHICS_VER(i915) == 8) {
+		ASSIGN_FW_DOMAINS_TABLE(uncore, __gen6_fw_ranges);
+		ASSIGN_SHADOW_TABLE(uncore, gen8_shadowed_regs);
+		ASSIGN_WRITE_MMIO_VFUNCS(uncore, fwtable);
+	} else if (IS_VALLEYVIEW(i915)) {
+		ASSIGN_FW_DOMAINS_TABLE(uncore, __vlv_fw_ranges);
+		ASSIGN_WRITE_MMIO_VFUNCS(uncore, gen6);
+	} else if (IS_GRAPHICS_VER(i915, 6, 7)) {
+		ASSIGN_FW_DOMAINS_TABLE(uncore, __gen6_fw_ranges);
+		ASSIGN_WRITE_MMIO_VFUNCS(uncore, gen6);
+	}
+
+	uncore->pmic_bus_access_nb.notifier_call = i915_pmic_bus_access_notifier;
+	iosf_mbi_register_pmic_bus_access_notifier(&uncore->pmic_bus_access_nb);
+
+	return 0;
+}
+
+int intel_uncore_init_mmio(struct intel_uncore *uncore)
+{
+	struct drm_i915_private *i915 = uncore->i915;
+	int ret;
+
+	/*
+	 * The boot firmware initializes local memory and assesses its health.
+	 * If memory training fails, the punit will have been instructed to
+	 * keep the GT powered down; we won't be able to communicate with it
+	 * and we should not continue with driver initialization.
+	 */
+	if (IS_DGFX(i915) &&
+	    !(__raw_uncore_read32(uncore, GU_CNTL) & LMEM_INIT)) {
+		drm_err(&i915->drm, "LMEM not initialized by firmware\n");
+		return -ENODEV;
+	}
+
+	if (GRAPHICS_VER(i915) > 5 && !intel_vgpu_active(i915))
+		uncore->flags |= UNCORE_HAS_FORCEWAKE;
+
+	if (!intel_uncore_has_forcewake(uncore)) {
+		uncore_raw_init(uncore);
+	} else {
+		ret = uncore_forcewake_init(uncore);
+		if (ret)
+			return ret;
+	}
+
+	/* make sure fw funcs are set if and only if we have fw*/
+	GEM_BUG_ON(intel_uncore_has_forcewake(uncore) != !!uncore->fw_get_funcs);
+	GEM_BUG_ON(intel_uncore_has_forcewake(uncore) != !!uncore->funcs.read_fw_domains);
+	GEM_BUG_ON(intel_uncore_has_forcewake(uncore) != !!uncore->funcs.write_fw_domains);
+
+	if (HAS_FPGA_DBG_UNCLAIMED(i915))
+		uncore->flags |= UNCORE_HAS_FPGA_DBG_UNCLAIMED;
+
+	if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915))
+		uncore->flags |= UNCORE_HAS_DBG_UNCLAIMED;
+
+	if (IS_GRAPHICS_VER(i915, 6, 7))
+		uncore->flags |= UNCORE_HAS_FIFO;
+
+	/* clear out unclaimed reg detection bit */
+	if (intel_uncore_unclaimed_mmio(uncore))
+		drm_dbg(&i915->drm, "unclaimed mmio detected on uncore init, clearing\n");
+
+	return 0;
+}
+
+/*
+ * We might have detected that some engines are fused off after we initialized
+ * the forcewake domains. Prune them, to make sure they only reference existing
+ * engines.
+ */
+void intel_uncore_prune_engine_fw_domains(struct intel_uncore *uncore,
+					  struct intel_gt *gt)
+{
+	enum forcewake_domains fw_domains = uncore->fw_domains;
+	enum forcewake_domain_id domain_id;
+	int i;
+
+	if (!intel_uncore_has_forcewake(uncore) || GRAPHICS_VER(uncore->i915) < 11)
+		return;
+
+	for (i = 0; i < I915_MAX_VCS; i++) {
+		domain_id = FW_DOMAIN_ID_MEDIA_VDBOX0 + i;
+
+		if (HAS_ENGINE(gt, _VCS(i)))
+			continue;
+
+		/*
+		 * Starting with XeHP, the power well for an even-numbered
+		 * VDBOX is also used for shared units within the
+		 * media slice such as SFC.  So even if the engine
+		 * itself is fused off, we still need to initialize
+		 * the forcewake domain if any of the other engines
+		 * in the same media slice are present.
+		 */
+		if (GRAPHICS_VER_FULL(uncore->i915) >= IP_VER(12, 50) && i % 2 == 0) {
+			if ((i + 1 < I915_MAX_VCS) && HAS_ENGINE(gt, _VCS(i + 1)))
+				continue;
+
+			if (HAS_ENGINE(gt, _VECS(i / 2)))
+				continue;
+		}
+
+		if (fw_domains & BIT(domain_id))
+			fw_domain_fini(uncore, domain_id);
+	}
+
+	for (i = 0; i < I915_MAX_VECS; i++) {
+		domain_id = FW_DOMAIN_ID_MEDIA_VEBOX0 + i;
+
+		if (HAS_ENGINE(gt, _VECS(i)))
+			continue;
+
+		if (fw_domains & BIT(domain_id))
+			fw_domain_fini(uncore, domain_id);
+	}
+}
+
+/* Called via drm-managed action */
+void intel_uncore_fini_mmio(struct drm_device *dev, void *data)
+{
+	struct intel_uncore *uncore = data;
+
+	if (intel_uncore_has_forcewake(uncore)) {
+		iosf_mbi_punit_acquire();
+		iosf_mbi_unregister_pmic_bus_access_notifier_unlocked(
+			&uncore->pmic_bus_access_nb);
+		intel_uncore_forcewake_reset(uncore);
+		intel_uncore_fw_domains_fini(uncore);
+		iosf_mbi_punit_release();
+	}
+}
+
+/**
+ * __intel_wait_for_register_fw - wait until register matches expected state
+ * @uncore: the struct intel_uncore
+ * @reg: the register to read
+ * @mask: mask to apply to register value
+ * @value: expected value
+ * @fast_timeout_us: fast timeout in microsecond for atomic/tight wait
+ * @slow_timeout_ms: slow timeout in millisecond
+ * @out_value: optional placeholder to hold registry value
+ *
+ * This routine waits until the target register @reg contains the expected
+ * @value after applying the @mask, i.e. it waits until ::
+ *
+ *     (intel_uncore_read_fw(uncore, reg) & mask) == value
+ *
+ * Otherwise, the wait will timeout after @slow_timeout_ms milliseconds.
+ * For atomic context @slow_timeout_ms must be zero and @fast_timeout_us
+ * must be not larger than 20,0000 microseconds.
+ *
+ * Note that this routine assumes the caller holds forcewake asserted, it is
+ * not suitable for very long waits. See intel_wait_for_register() if you
+ * wish to wait without holding forcewake for the duration (i.e. you expect
+ * the wait to be slow).
+ *
+ * Return: 0 if the register matches the desired condition, or -ETIMEDOUT.
+ */
+int __intel_wait_for_register_fw(struct intel_uncore *uncore,
+				 i915_reg_t reg,
+				 u32 mask,
+				 u32 value,
+				 unsigned int fast_timeout_us,
+				 unsigned int slow_timeout_ms,
+				 u32 *out_value)
+{
+	u32 reg_value = 0;
+#define done (((reg_value = intel_uncore_read_fw(uncore, reg)) & mask) == value)
+	int ret;
+
+	/* Catch any overuse of this function */
+	might_sleep_if(slow_timeout_ms);
+	GEM_BUG_ON(fast_timeout_us > 20000);
+	GEM_BUG_ON(!fast_timeout_us && !slow_timeout_ms);
+
+	ret = -ETIMEDOUT;
+	if (fast_timeout_us && fast_timeout_us <= 20000)
+		ret = _wait_for_atomic(done, fast_timeout_us, 0);
+	if (ret && slow_timeout_ms)
+		ret = wait_for(done, slow_timeout_ms);
+
+	if (out_value)
+		*out_value = reg_value;
+
+	return ret;
+#undef done
+}
+
+/**
+ * __intel_wait_for_register - wait until register matches expected state
+ * @uncore: the struct intel_uncore
+ * @reg: the register to read
+ * @mask: mask to apply to register value
+ * @value: expected value
+ * @fast_timeout_us: fast timeout in microsecond for atomic/tight wait
+ * @slow_timeout_ms: slow timeout in millisecond
+ * @out_value: optional placeholder to hold registry value
+ *
+ * This routine waits until the target register @reg contains the expected
+ * @value after applying the @mask, i.e. it waits until ::
+ *
+ *     (intel_uncore_read(uncore, reg) & mask) == value
+ *
+ * Otherwise, the wait will timeout after @timeout_ms milliseconds.
+ *
+ * Return: 0 if the register matches the desired condition, or -ETIMEDOUT.
+ */
+int __intel_wait_for_register(struct intel_uncore *uncore,
+			      i915_reg_t reg,
+			      u32 mask,
+			      u32 value,
+			      unsigned int fast_timeout_us,
+			      unsigned int slow_timeout_ms,
+			      u32 *out_value)
+{
+	unsigned fw =
+		intel_uncore_forcewake_for_reg(uncore, reg, FW_REG_READ);
+	u32 reg_value;
+	int ret;
+
+	might_sleep_if(slow_timeout_ms);
+
+	spin_lock_irq(&uncore->lock);
+	intel_uncore_forcewake_get__locked(uncore, fw);
+
+	ret = __intel_wait_for_register_fw(uncore,
+					   reg, mask, value,
+					   fast_timeout_us, 0, &reg_value);
+
+	intel_uncore_forcewake_put__locked(uncore, fw);
+	spin_unlock_irq(&uncore->lock);
+
+	if (ret && slow_timeout_ms)
+		ret = __wait_for(reg_value = intel_uncore_read_notrace(uncore,
+								       reg),
+				 (reg_value & mask) == value,
+				 slow_timeout_ms * 1000, 10, 1000);
+
+	/* just trace the final value */
+	trace_i915_reg_rw(false, reg, reg_value, sizeof(reg_value), true);
+
+	if (out_value)
+		*out_value = reg_value;
+
+	return ret;
+}
+
+bool intel_uncore_unclaimed_mmio(struct intel_uncore *uncore)
+{
+	bool ret;
+
+	if (!uncore->debug)
+		return false;
+
+	spin_lock_irq(&uncore->debug->lock);
+	ret = check_for_unclaimed_mmio(uncore);
+	spin_unlock_irq(&uncore->debug->lock);
+
+	return ret;
+}
+
+bool
+intel_uncore_arm_unclaimed_mmio_detection(struct intel_uncore *uncore)
+{
+	bool ret = false;
+
+	if (drm_WARN_ON(&uncore->i915->drm, !uncore->debug))
+		return false;
+
+	spin_lock_irq(&uncore->debug->lock);
+
+	if (unlikely(uncore->debug->unclaimed_mmio_check <= 0))
+		goto out;
+
+	if (unlikely(check_for_unclaimed_mmio(uncore))) {
+		if (!uncore->i915->params.mmio_debug) {
+			drm_dbg(&uncore->i915->drm,
+				"Unclaimed register detected, "
+				"enabling oneshot unclaimed register reporting. "
+				"Please use i915.mmio_debug=N for more information.\n");
+			uncore->i915->params.mmio_debug++;
+		}
+		uncore->debug->unclaimed_mmio_check--;
+		ret = true;
+	}
+
+out:
+	spin_unlock_irq(&uncore->debug->lock);
+
+	return ret;
+}
+
+/**
+ * intel_uncore_forcewake_for_reg - which forcewake domains are needed to access
+ * 				    a register
+ * @uncore: pointer to struct intel_uncore
+ * @reg: register in question
+ * @op: operation bitmask of FW_REG_READ and/or FW_REG_WRITE
+ *
+ * Returns a set of forcewake domains required to be taken with for example
+ * intel_uncore_forcewake_get for the specified register to be accessible in the
+ * specified mode (read, write or read/write) with raw mmio accessors.
+ *
+ * NOTE: On Gen6 and Gen7 write forcewake domain (FORCEWAKE_RENDER) requires the
+ * callers to do FIFO management on their own or risk losing writes.
+ */
+enum forcewake_domains
+intel_uncore_forcewake_for_reg(struct intel_uncore *uncore,
+			       i915_reg_t reg, unsigned int op)
+{
+	enum forcewake_domains fw_domains = 0;
+
+	drm_WARN_ON(&uncore->i915->drm, !op);
+
+	if (!intel_uncore_has_forcewake(uncore))
+		return 0;
+
+	if (op & FW_REG_READ)
+		fw_domains = uncore->funcs.read_fw_domains(uncore, reg);
+
+	if (op & FW_REG_WRITE)
+		fw_domains |= uncore->funcs.write_fw_domains(uncore, reg);
+
+	drm_WARN_ON(&uncore->i915->drm, fw_domains & ~uncore->fw_domains);
+
+	return fw_domains;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_uncore.c"
+#include "selftests/intel_uncore.c"
+#endif
diff --git a/drivers/gpu/drm/i915/intel_uncore.h b/drivers/gpu/drm/i915/intel_uncore.h
new file mode 100644
index 000000000..5022bac80
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_uncore.h
@@ -0,0 +1,485 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __INTEL_UNCORE_H__
+#define __INTEL_UNCORE_H__
+
+#include <linux/spinlock.h>
+#include <linux/notifier.h>
+#include <linux/hrtimer.h>
+#include <linux/io-64-nonatomic-lo-hi.h>
+#include <linux/types.h>
+
+#include "i915_reg_defs.h"
+
+struct drm_device;
+struct drm_i915_private;
+struct intel_runtime_pm;
+struct intel_uncore;
+struct intel_gt;
+
+struct intel_uncore_mmio_debug {
+	spinlock_t lock; /** lock is also taken in irq contexts. */
+	int unclaimed_mmio_check;
+	int saved_mmio_check;
+	u32 suspend_count;
+};
+
+enum forcewake_domain_id {
+	FW_DOMAIN_ID_RENDER = 0,
+	FW_DOMAIN_ID_GT,        /* also includes blitter engine */
+	FW_DOMAIN_ID_MEDIA,
+	FW_DOMAIN_ID_MEDIA_VDBOX0,
+	FW_DOMAIN_ID_MEDIA_VDBOX1,
+	FW_DOMAIN_ID_MEDIA_VDBOX2,
+	FW_DOMAIN_ID_MEDIA_VDBOX3,
+	FW_DOMAIN_ID_MEDIA_VDBOX4,
+	FW_DOMAIN_ID_MEDIA_VDBOX5,
+	FW_DOMAIN_ID_MEDIA_VDBOX6,
+	FW_DOMAIN_ID_MEDIA_VDBOX7,
+	FW_DOMAIN_ID_MEDIA_VEBOX0,
+	FW_DOMAIN_ID_MEDIA_VEBOX1,
+	FW_DOMAIN_ID_MEDIA_VEBOX2,
+	FW_DOMAIN_ID_MEDIA_VEBOX3,
+
+	FW_DOMAIN_ID_COUNT
+};
+
+enum forcewake_domains {
+	FORCEWAKE_RENDER	= BIT(FW_DOMAIN_ID_RENDER),
+	FORCEWAKE_GT		= BIT(FW_DOMAIN_ID_GT),
+	FORCEWAKE_MEDIA		= BIT(FW_DOMAIN_ID_MEDIA),
+	FORCEWAKE_MEDIA_VDBOX0	= BIT(FW_DOMAIN_ID_MEDIA_VDBOX0),
+	FORCEWAKE_MEDIA_VDBOX1	= BIT(FW_DOMAIN_ID_MEDIA_VDBOX1),
+	FORCEWAKE_MEDIA_VDBOX2	= BIT(FW_DOMAIN_ID_MEDIA_VDBOX2),
+	FORCEWAKE_MEDIA_VDBOX3	= BIT(FW_DOMAIN_ID_MEDIA_VDBOX3),
+	FORCEWAKE_MEDIA_VDBOX4	= BIT(FW_DOMAIN_ID_MEDIA_VDBOX4),
+	FORCEWAKE_MEDIA_VDBOX5	= BIT(FW_DOMAIN_ID_MEDIA_VDBOX5),
+	FORCEWAKE_MEDIA_VDBOX6	= BIT(FW_DOMAIN_ID_MEDIA_VDBOX6),
+	FORCEWAKE_MEDIA_VDBOX7	= BIT(FW_DOMAIN_ID_MEDIA_VDBOX7),
+	FORCEWAKE_MEDIA_VEBOX0	= BIT(FW_DOMAIN_ID_MEDIA_VEBOX0),
+	FORCEWAKE_MEDIA_VEBOX1	= BIT(FW_DOMAIN_ID_MEDIA_VEBOX1),
+	FORCEWAKE_MEDIA_VEBOX2	= BIT(FW_DOMAIN_ID_MEDIA_VEBOX2),
+	FORCEWAKE_MEDIA_VEBOX3	= BIT(FW_DOMAIN_ID_MEDIA_VEBOX3),
+
+	FORCEWAKE_ALL = BIT(FW_DOMAIN_ID_COUNT) - 1,
+};
+
+struct intel_uncore_fw_get {
+	void (*force_wake_get)(struct intel_uncore *uncore,
+			       enum forcewake_domains domains);
+};
+
+struct intel_uncore_funcs {
+	enum forcewake_domains (*read_fw_domains)(struct intel_uncore *uncore,
+						  i915_reg_t r);
+	enum forcewake_domains (*write_fw_domains)(struct intel_uncore *uncore,
+						   i915_reg_t r);
+
+	u8 (*mmio_readb)(struct intel_uncore *uncore,
+			 i915_reg_t r, bool trace);
+	u16 (*mmio_readw)(struct intel_uncore *uncore,
+			  i915_reg_t r, bool trace);
+	u32 (*mmio_readl)(struct intel_uncore *uncore,
+			  i915_reg_t r, bool trace);
+	u64 (*mmio_readq)(struct intel_uncore *uncore,
+			  i915_reg_t r, bool trace);
+
+	void (*mmio_writeb)(struct intel_uncore *uncore,
+			    i915_reg_t r, u8 val, bool trace);
+	void (*mmio_writew)(struct intel_uncore *uncore,
+			    i915_reg_t r, u16 val, bool trace);
+	void (*mmio_writel)(struct intel_uncore *uncore,
+			    i915_reg_t r, u32 val, bool trace);
+};
+
+struct intel_forcewake_range {
+	u32 start;
+	u32 end;
+
+	enum forcewake_domains domains;
+};
+
+/* Other register ranges (e.g., shadow tables, MCR tables, etc.) */
+struct i915_range {
+	u32 start;
+	u32 end;
+};
+
+struct intel_uncore {
+	void __iomem *regs;
+
+	struct drm_i915_private *i915;
+	struct intel_gt *gt;
+	struct intel_runtime_pm *rpm;
+
+	spinlock_t lock; /** lock is also taken in irq contexts. */
+
+	/*
+	 * Do we need to apply an additional offset to reach the beginning
+	 * of the basic non-engine GT registers (referred to as "GSI" on
+	 * newer platforms, or "GT block" on older platforms)?  If so, we'll
+	 * track that here and apply it transparently to registers in the
+	 * appropriate range to maintain compatibility with our existing
+	 * register definitions and GT code.
+	 */
+	u32 gsi_offset;
+
+	unsigned int flags;
+#define UNCORE_HAS_FORCEWAKE		BIT(0)
+#define UNCORE_HAS_FPGA_DBG_UNCLAIMED	BIT(1)
+#define UNCORE_HAS_DBG_UNCLAIMED	BIT(2)
+#define UNCORE_HAS_FIFO			BIT(3)
+
+	const struct intel_forcewake_range *fw_domains_table;
+	unsigned int fw_domains_table_entries;
+
+	/*
+	 * Shadowed registers are special cases where we can safely write
+	 * to the register *without* grabbing forcewake.
+	 */
+	const struct i915_range *shadowed_reg_table;
+	unsigned int shadowed_reg_table_entries;
+
+	struct notifier_block pmic_bus_access_nb;
+	const struct intel_uncore_fw_get *fw_get_funcs;
+	struct intel_uncore_funcs funcs;
+
+	unsigned int fifo_count;
+
+	enum forcewake_domains fw_domains;
+	enum forcewake_domains fw_domains_active;
+	enum forcewake_domains fw_domains_timer;
+	enum forcewake_domains fw_domains_saved; /* user domains saved for S3 */
+
+	struct intel_uncore_forcewake_domain {
+		struct intel_uncore *uncore;
+		enum forcewake_domain_id id;
+		enum forcewake_domains mask;
+		unsigned int wake_count;
+		bool active;
+		struct hrtimer timer;
+		u32 __iomem *reg_set;
+		u32 __iomem *reg_ack;
+	} *fw_domain[FW_DOMAIN_ID_COUNT];
+
+	unsigned int user_forcewake_count;
+
+	struct intel_uncore_mmio_debug *debug;
+};
+
+/* Iterate over initialised fw domains */
+#define for_each_fw_domain_masked(domain__, mask__, uncore__, tmp__) \
+	for (tmp__ = (mask__); tmp__ ;) \
+		for_each_if(domain__ = (uncore__)->fw_domain[__mask_next_bit(tmp__)])
+
+#define for_each_fw_domain(domain__, uncore__, tmp__) \
+	for_each_fw_domain_masked(domain__, (uncore__)->fw_domains, uncore__, tmp__)
+
+static inline bool
+intel_uncore_has_forcewake(const struct intel_uncore *uncore)
+{
+	return uncore->flags & UNCORE_HAS_FORCEWAKE;
+}
+
+static inline bool
+intel_uncore_has_fpga_dbg_unclaimed(const struct intel_uncore *uncore)
+{
+	return uncore->flags & UNCORE_HAS_FPGA_DBG_UNCLAIMED;
+}
+
+static inline bool
+intel_uncore_has_dbg_unclaimed(const struct intel_uncore *uncore)
+{
+	return uncore->flags & UNCORE_HAS_DBG_UNCLAIMED;
+}
+
+static inline bool
+intel_uncore_has_fifo(const struct intel_uncore *uncore)
+{
+	return uncore->flags & UNCORE_HAS_FIFO;
+}
+
+void intel_uncore_mmio_debug_init_early(struct drm_i915_private *i915);
+void intel_uncore_init_early(struct intel_uncore *uncore,
+			     struct intel_gt *gt);
+int intel_uncore_setup_mmio(struct intel_uncore *uncore, phys_addr_t phys_addr);
+int intel_uncore_init_mmio(struct intel_uncore *uncore);
+void intel_uncore_prune_engine_fw_domains(struct intel_uncore *uncore,
+					  struct intel_gt *gt);
+bool intel_uncore_unclaimed_mmio(struct intel_uncore *uncore);
+bool intel_uncore_arm_unclaimed_mmio_detection(struct intel_uncore *uncore);
+void intel_uncore_cleanup_mmio(struct intel_uncore *uncore);
+void intel_uncore_fini_mmio(struct drm_device *dev, void *data);
+void intel_uncore_suspend(struct intel_uncore *uncore);
+void intel_uncore_resume_early(struct intel_uncore *uncore);
+void intel_uncore_runtime_resume(struct intel_uncore *uncore);
+
+void assert_forcewakes_inactive(struct intel_uncore *uncore);
+void assert_forcewakes_active(struct intel_uncore *uncore,
+			      enum forcewake_domains fw_domains);
+const char *intel_uncore_forcewake_domain_to_str(const enum forcewake_domain_id id);
+
+enum forcewake_domains
+intel_uncore_forcewake_for_reg(struct intel_uncore *uncore,
+			       i915_reg_t reg, unsigned int op);
+#define FW_REG_READ  (1)
+#define FW_REG_WRITE (2)
+
+void intel_uncore_forcewake_get(struct intel_uncore *uncore,
+				enum forcewake_domains domains);
+void intel_uncore_forcewake_put(struct intel_uncore *uncore,
+				enum forcewake_domains domains);
+void intel_uncore_forcewake_put_delayed(struct intel_uncore *uncore,
+					enum forcewake_domains domains);
+void intel_uncore_forcewake_flush(struct intel_uncore *uncore,
+				  enum forcewake_domains fw_domains);
+
+/*
+ * Like above but the caller must manage the uncore.lock itself.
+ * Must be used with intel_uncore_read_fw() and friends.
+ */
+void intel_uncore_forcewake_get__locked(struct intel_uncore *uncore,
+					enum forcewake_domains domains);
+void intel_uncore_forcewake_put__locked(struct intel_uncore *uncore,
+					enum forcewake_domains domains);
+
+void intel_uncore_forcewake_user_get(struct intel_uncore *uncore);
+void intel_uncore_forcewake_user_put(struct intel_uncore *uncore);
+
+int __intel_wait_for_register(struct intel_uncore *uncore,
+			      i915_reg_t reg,
+			      u32 mask,
+			      u32 value,
+			      unsigned int fast_timeout_us,
+			      unsigned int slow_timeout_ms,
+			      u32 *out_value);
+static inline int
+intel_wait_for_register(struct intel_uncore *uncore,
+			i915_reg_t reg,
+			u32 mask,
+			u32 value,
+			unsigned int timeout_ms)
+{
+	return __intel_wait_for_register(uncore, reg, mask, value, 2,
+					 timeout_ms, NULL);
+}
+
+int __intel_wait_for_register_fw(struct intel_uncore *uncore,
+				 i915_reg_t reg,
+				 u32 mask,
+				 u32 value,
+				 unsigned int fast_timeout_us,
+				 unsigned int slow_timeout_ms,
+				 u32 *out_value);
+static inline int
+intel_wait_for_register_fw(struct intel_uncore *uncore,
+			   i915_reg_t reg,
+			   u32 mask,
+			   u32 value,
+			       unsigned int timeout_ms)
+{
+	return __intel_wait_for_register_fw(uncore, reg, mask, value,
+					    2, timeout_ms, NULL);
+}
+
+#define IS_GSI_REG(reg) ((reg) < 0x40000)
+
+/* register access functions */
+#define __raw_read(x__, s__) \
+static inline u##x__ __raw_uncore_read##x__(const struct intel_uncore *uncore, \
+					    i915_reg_t reg) \
+{ \
+	u32 offset = i915_mmio_reg_offset(reg); \
+	if (IS_GSI_REG(offset)) \
+		offset += uncore->gsi_offset; \
+	return read##s__(uncore->regs + offset); \
+}
+
+#define __raw_write(x__, s__) \
+static inline void __raw_uncore_write##x__(const struct intel_uncore *uncore, \
+					   i915_reg_t reg, u##x__ val) \
+{ \
+	u32 offset = i915_mmio_reg_offset(reg); \
+	if (IS_GSI_REG(offset)) \
+		offset += uncore->gsi_offset; \
+	write##s__(val, uncore->regs + offset); \
+}
+__raw_read(8, b)
+__raw_read(16, w)
+__raw_read(32, l)
+__raw_read(64, q)
+
+__raw_write(8, b)
+__raw_write(16, w)
+__raw_write(32, l)
+__raw_write(64, q)
+
+#undef __raw_read
+#undef __raw_write
+
+#define __uncore_read(name__, x__, s__, trace__) \
+static inline u##x__ intel_uncore_##name__(struct intel_uncore *uncore, \
+					   i915_reg_t reg) \
+{ \
+	return uncore->funcs.mmio_read##s__(uncore, reg, (trace__)); \
+}
+
+#define __uncore_write(name__, x__, s__, trace__) \
+static inline void intel_uncore_##name__(struct intel_uncore *uncore, \
+					 i915_reg_t reg, u##x__ val) \
+{ \
+	uncore->funcs.mmio_write##s__(uncore, reg, val, (trace__)); \
+}
+
+__uncore_read(read8, 8, b, true)
+__uncore_read(read16, 16, w, true)
+__uncore_read(read, 32, l, true)
+__uncore_read(read16_notrace, 16, w, false)
+__uncore_read(read_notrace, 32, l, false)
+
+__uncore_write(write8, 8, b, true)
+__uncore_write(write16, 16, w, true)
+__uncore_write(write, 32, l, true)
+__uncore_write(write_notrace, 32, l, false)
+
+/* Be very careful with read/write 64-bit values. On 32-bit machines, they
+ * will be implemented using 2 32-bit writes in an arbitrary order with
+ * an arbitrary delay between them. This can cause the hardware to
+ * act upon the intermediate value, possibly leading to corruption and
+ * machine death. For this reason we do not support intel_uncore_write64,
+ * or uncore->funcs.mmio_writeq.
+ *
+ * When reading a 64-bit value as two 32-bit values, the delay may cause
+ * the two reads to mismatch, e.g. a timestamp overflowing. Also note that
+ * occasionally a 64-bit register does not actually support a full readq
+ * and must be read using two 32-bit reads.
+ *
+ * You have been warned.
+ */
+__uncore_read(read64, 64, q, true)
+
+static inline u64
+intel_uncore_read64_2x32(struct intel_uncore *uncore,
+			 i915_reg_t lower_reg, i915_reg_t upper_reg)
+{
+	u32 upper, lower, old_upper, loop = 0;
+	upper = intel_uncore_read(uncore, upper_reg);
+	do {
+		old_upper = upper;
+		lower = intel_uncore_read(uncore, lower_reg);
+		upper = intel_uncore_read(uncore, upper_reg);
+	} while (upper != old_upper && loop++ < 2);
+	return (u64)upper << 32 | lower;
+}
+
+#define intel_uncore_posting_read(...) ((void)intel_uncore_read_notrace(__VA_ARGS__))
+#define intel_uncore_posting_read16(...) ((void)intel_uncore_read16_notrace(__VA_ARGS__))
+
+#undef __uncore_read
+#undef __uncore_write
+
+/* These are untraced mmio-accessors that are only valid to be used inside
+ * critical sections, such as inside IRQ handlers, where forcewake is explicitly
+ * controlled.
+ *
+ * Think twice, and think again, before using these.
+ *
+ * As an example, these accessors can possibly be used between:
+ *
+ * spin_lock_irq(&uncore->lock);
+ * intel_uncore_forcewake_get__locked();
+ *
+ * and
+ *
+ * intel_uncore_forcewake_put__locked();
+ * spin_unlock_irq(&uncore->lock);
+ *
+ *
+ * Note: some registers may not need forcewake held, so
+ * intel_uncore_forcewake_{get,put} can be omitted, see
+ * intel_uncore_forcewake_for_reg().
+ *
+ * Certain architectures will die if the same cacheline is concurrently accessed
+ * by different clients (e.g. on Ivybridge). Access to registers should
+ * therefore generally be serialised, by either the dev_priv->uncore.lock or
+ * a more localised lock guarding all access to that bank of registers.
+ */
+#define intel_uncore_read_fw(...) __raw_uncore_read32(__VA_ARGS__)
+#define intel_uncore_write_fw(...) __raw_uncore_write32(__VA_ARGS__)
+#define intel_uncore_write64_fw(...) __raw_uncore_write64(__VA_ARGS__)
+#define intel_uncore_posting_read_fw(...) ((void)intel_uncore_read_fw(__VA_ARGS__))
+
+static inline void intel_uncore_rmw(struct intel_uncore *uncore,
+				    i915_reg_t reg, u32 clear, u32 set)
+{
+	u32 old, val;
+
+	old = intel_uncore_read(uncore, reg);
+	val = (old & ~clear) | set;
+	if (val != old)
+		intel_uncore_write(uncore, reg, val);
+}
+
+static inline void intel_uncore_rmw_fw(struct intel_uncore *uncore,
+				       i915_reg_t reg, u32 clear, u32 set)
+{
+	u32 old, val;
+
+	old = intel_uncore_read_fw(uncore, reg);
+	val = (old & ~clear) | set;
+	if (val != old)
+		intel_uncore_write_fw(uncore, reg, val);
+}
+
+static inline int intel_uncore_write_and_verify(struct intel_uncore *uncore,
+						i915_reg_t reg, u32 val,
+						u32 mask, u32 expected_val)
+{
+	u32 reg_val;
+
+	intel_uncore_write(uncore, reg, val);
+	reg_val = intel_uncore_read(uncore, reg);
+
+	return (reg_val & mask) != expected_val ? -EINVAL : 0;
+}
+
+/*
+ * The raw_reg_{read,write} macros are intended as a micro-optimization for
+ * interrupt handlers so that the pointer indirection on uncore->regs can
+ * be computed once (and presumably cached in a register) instead of generating
+ * extra load instructions for each MMIO access.
+ *
+ * Given that these macros are only intended for non-GSI interrupt registers
+ * (and the goal is to avoid extra instructions generated by the compiler),
+ * these macros do not account for uncore->gsi_offset.  Any caller that needs
+ * to use these macros on a GSI register is responsible for adding the
+ * appropriate GSI offset to the 'base' parameter.
+ */
+#define raw_reg_read(base, reg) \
+	readl(base + i915_mmio_reg_offset(reg))
+#define raw_reg_write(base, reg, value) \
+	writel(value, base + i915_mmio_reg_offset(reg))
+
+#endif /* !__INTEL_UNCORE_H__ */
diff --git a/drivers/gpu/drm/i915/intel_wakeref.c b/drivers/gpu/drm/i915/intel_wakeref.c
new file mode 100644
index 000000000..dfd87d082
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_wakeref.c
@@ -0,0 +1,189 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/wait_bit.h>
+
+#include "intel_runtime_pm.h"
+#include "intel_wakeref.h"
+
+static void rpm_get(struct intel_wakeref *wf)
+{
+	wf->wakeref = intel_runtime_pm_get(wf->rpm);
+}
+
+static void rpm_put(struct intel_wakeref *wf)
+{
+	intel_wakeref_t wakeref = fetch_and_zero(&wf->wakeref);
+
+	intel_runtime_pm_put(wf->rpm, wakeref);
+	INTEL_WAKEREF_BUG_ON(!wakeref);
+}
+
+int __intel_wakeref_get_first(struct intel_wakeref *wf)
+{
+	/*
+	 * Treat get/put as different subclasses, as we may need to run
+	 * the put callback from under the shrinker and do not want to
+	 * cross-contanimate that callback with any extra work performed
+	 * upon acquiring the wakeref.
+	 */
+	mutex_lock_nested(&wf->mutex, SINGLE_DEPTH_NESTING);
+	if (!atomic_read(&wf->count)) {
+		int err;
+
+		rpm_get(wf);
+
+		err = wf->ops->get(wf);
+		if (unlikely(err)) {
+			rpm_put(wf);
+			mutex_unlock(&wf->mutex);
+			return err;
+		}
+
+		smp_mb__before_atomic(); /* release wf->count */
+	}
+	atomic_inc(&wf->count);
+	mutex_unlock(&wf->mutex);
+
+	INTEL_WAKEREF_BUG_ON(atomic_read(&wf->count) <= 0);
+	return 0;
+}
+
+static void ____intel_wakeref_put_last(struct intel_wakeref *wf)
+{
+	INTEL_WAKEREF_BUG_ON(atomic_read(&wf->count) <= 0);
+	if (unlikely(!atomic_dec_and_test(&wf->count)))
+		goto unlock;
+
+	/* ops->put() must reschedule its own release on error/deferral */
+	if (likely(!wf->ops->put(wf))) {
+		rpm_put(wf);
+		wake_up_var(&wf->wakeref);
+	}
+
+unlock:
+	mutex_unlock(&wf->mutex);
+}
+
+void __intel_wakeref_put_last(struct intel_wakeref *wf, unsigned long flags)
+{
+	INTEL_WAKEREF_BUG_ON(delayed_work_pending(&wf->work));
+
+	/* Assume we are not in process context and so cannot sleep. */
+	if (flags & INTEL_WAKEREF_PUT_ASYNC || !mutex_trylock(&wf->mutex)) {
+		mod_delayed_work(system_wq, &wf->work,
+				 FIELD_GET(INTEL_WAKEREF_PUT_DELAY, flags));
+		return;
+	}
+
+	____intel_wakeref_put_last(wf);
+}
+
+static void __intel_wakeref_put_work(struct work_struct *wrk)
+{
+	struct intel_wakeref *wf = container_of(wrk, typeof(*wf), work.work);
+
+	if (atomic_add_unless(&wf->count, -1, 1))
+		return;
+
+	mutex_lock(&wf->mutex);
+	____intel_wakeref_put_last(wf);
+}
+
+void __intel_wakeref_init(struct intel_wakeref *wf,
+			  struct intel_runtime_pm *rpm,
+			  const struct intel_wakeref_ops *ops,
+			  struct intel_wakeref_lockclass *key)
+{
+	wf->rpm = rpm;
+	wf->ops = ops;
+
+	__mutex_init(&wf->mutex, "wakeref.mutex", &key->mutex);
+	atomic_set(&wf->count, 0);
+	wf->wakeref = 0;
+
+	INIT_DELAYED_WORK(&wf->work, __intel_wakeref_put_work);
+	lockdep_init_map(&wf->work.work.lockdep_map,
+			 "wakeref.work", &key->work, 0);
+}
+
+int intel_wakeref_wait_for_idle(struct intel_wakeref *wf)
+{
+	int err;
+
+	might_sleep();
+
+	err = wait_var_event_killable(&wf->wakeref,
+				      !intel_wakeref_is_active(wf));
+	if (err)
+		return err;
+
+	intel_wakeref_unlock_wait(wf);
+	return 0;
+}
+
+static void wakeref_auto_timeout(struct timer_list *t)
+{
+	struct intel_wakeref_auto *wf = from_timer(wf, t, timer);
+	intel_wakeref_t wakeref;
+	unsigned long flags;
+
+	if (!refcount_dec_and_lock_irqsave(&wf->count, &wf->lock, &flags))
+		return;
+
+	wakeref = fetch_and_zero(&wf->wakeref);
+	spin_unlock_irqrestore(&wf->lock, flags);
+
+	intel_runtime_pm_put(wf->rpm, wakeref);
+}
+
+void intel_wakeref_auto_init(struct intel_wakeref_auto *wf,
+			     struct intel_runtime_pm *rpm)
+{
+	spin_lock_init(&wf->lock);
+	timer_setup(&wf->timer, wakeref_auto_timeout, 0);
+	refcount_set(&wf->count, 0);
+	wf->wakeref = 0;
+	wf->rpm = rpm;
+}
+
+void intel_wakeref_auto(struct intel_wakeref_auto *wf, unsigned long timeout)
+{
+	unsigned long flags;
+
+	if (!timeout) {
+		if (del_timer_sync(&wf->timer))
+			wakeref_auto_timeout(&wf->timer);
+		return;
+	}
+
+	/* Our mission is that we only extend an already active wakeref */
+	assert_rpm_wakelock_held(wf->rpm);
+
+	if (!refcount_inc_not_zero(&wf->count)) {
+		spin_lock_irqsave(&wf->lock, flags);
+		if (!refcount_inc_not_zero(&wf->count)) {
+			INTEL_WAKEREF_BUG_ON(wf->wakeref);
+			wf->wakeref = intel_runtime_pm_get_if_in_use(wf->rpm);
+			refcount_set(&wf->count, 1);
+		}
+		spin_unlock_irqrestore(&wf->lock, flags);
+	}
+
+	/*
+	 * If we extend a pending timer, we will only get a single timer
+	 * callback and so need to cancel the local inc by running the
+	 * elided callback to keep the wf->count balanced.
+	 */
+	if (mod_timer(&wf->timer, jiffies + timeout))
+		wakeref_auto_timeout(&wf->timer);
+}
+
+void intel_wakeref_auto_fini(struct intel_wakeref_auto *wf)
+{
+	intel_wakeref_auto(wf, 0);
+	INTEL_WAKEREF_BUG_ON(wf->wakeref);
+}
diff --git a/drivers/gpu/drm/i915/intel_wakeref.h b/drivers/gpu/drm/i915/intel_wakeref.h
new file mode 100644
index 000000000..4f4c2e15e
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_wakeref.h
@@ -0,0 +1,294 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef INTEL_WAKEREF_H
+#define INTEL_WAKEREF_H
+
+#include <linux/atomic.h>
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <linux/lockdep.h>
+#include <linux/mutex.h>
+#include <linux/refcount.h>
+#include <linux/stackdepot.h>
+#include <linux/timer.h>
+#include <linux/workqueue.h>
+
+#if IS_ENABLED(CONFIG_DRM_I915_DEBUG)
+#define INTEL_WAKEREF_BUG_ON(expr) BUG_ON(expr)
+#else
+#define INTEL_WAKEREF_BUG_ON(expr) BUILD_BUG_ON_INVALID(expr)
+#endif
+
+struct intel_runtime_pm;
+struct intel_wakeref;
+
+typedef depot_stack_handle_t intel_wakeref_t;
+
+struct intel_wakeref_ops {
+	int (*get)(struct intel_wakeref *wf);
+	int (*put)(struct intel_wakeref *wf);
+};
+
+struct intel_wakeref {
+	atomic_t count;
+	struct mutex mutex;
+
+	intel_wakeref_t wakeref;
+
+	struct intel_runtime_pm *rpm;
+	const struct intel_wakeref_ops *ops;
+
+	struct delayed_work work;
+};
+
+struct intel_wakeref_lockclass {
+	struct lock_class_key mutex;
+	struct lock_class_key work;
+};
+
+void __intel_wakeref_init(struct intel_wakeref *wf,
+			  struct intel_runtime_pm *rpm,
+			  const struct intel_wakeref_ops *ops,
+			  struct intel_wakeref_lockclass *key);
+#define intel_wakeref_init(wf, rpm, ops) do {				\
+	static struct intel_wakeref_lockclass __key;			\
+									\
+	__intel_wakeref_init((wf), (rpm), (ops), &__key);		\
+} while (0)
+
+int __intel_wakeref_get_first(struct intel_wakeref *wf);
+void __intel_wakeref_put_last(struct intel_wakeref *wf, unsigned long flags);
+
+/**
+ * intel_wakeref_get: Acquire the wakeref
+ * @wf: the wakeref
+ *
+ * Acquire a hold on the wakeref. The first user to do so, will acquire
+ * the runtime pm wakeref and then call the @fn underneath the wakeref
+ * mutex.
+ *
+ * Note that @fn is allowed to fail, in which case the runtime-pm wakeref
+ * will be released and the acquisition unwound, and an error reported.
+ *
+ * Returns: 0 if the wakeref was acquired successfully, or a negative error
+ * code otherwise.
+ */
+static inline int
+intel_wakeref_get(struct intel_wakeref *wf)
+{
+	might_sleep();
+	if (unlikely(!atomic_inc_not_zero(&wf->count)))
+		return __intel_wakeref_get_first(wf);
+
+	return 0;
+}
+
+/**
+ * __intel_wakeref_get: Acquire the wakeref, again
+ * @wf: the wakeref
+ *
+ * Increment the wakeref counter, only valid if it is already held by
+ * the caller.
+ *
+ * See intel_wakeref_get().
+ */
+static inline void
+__intel_wakeref_get(struct intel_wakeref *wf)
+{
+	INTEL_WAKEREF_BUG_ON(atomic_read(&wf->count) <= 0);
+	atomic_inc(&wf->count);
+}
+
+/**
+ * intel_wakeref_get_if_in_use: Acquire the wakeref
+ * @wf: the wakeref
+ *
+ * Acquire a hold on the wakeref, but only if the wakeref is already
+ * active.
+ *
+ * Returns: true if the wakeref was acquired, false otherwise.
+ */
+static inline bool
+intel_wakeref_get_if_active(struct intel_wakeref *wf)
+{
+	return atomic_inc_not_zero(&wf->count);
+}
+
+enum {
+	INTEL_WAKEREF_PUT_ASYNC_BIT = 0,
+	__INTEL_WAKEREF_PUT_LAST_BIT__
+};
+
+static inline void
+intel_wakeref_might_get(struct intel_wakeref *wf)
+{
+	might_lock(&wf->mutex);
+}
+
+/**
+ * intel_wakeref_put_flags: Release the wakeref
+ * @wf: the wakeref
+ * @flags: control flags
+ *
+ * Release our hold on the wakeref. When there are no more users,
+ * the runtime pm wakeref will be released after the @fn callback is called
+ * underneath the wakeref mutex.
+ *
+ * Note that @fn is allowed to fail, in which case the runtime-pm wakeref
+ * is retained and an error reported.
+ *
+ * Returns: 0 if the wakeref was released successfully, or a negative error
+ * code otherwise.
+ */
+static inline void
+__intel_wakeref_put(struct intel_wakeref *wf, unsigned long flags)
+#define INTEL_WAKEREF_PUT_ASYNC BIT(INTEL_WAKEREF_PUT_ASYNC_BIT)
+#define INTEL_WAKEREF_PUT_DELAY \
+	GENMASK(BITS_PER_LONG - 1, __INTEL_WAKEREF_PUT_LAST_BIT__)
+{
+	INTEL_WAKEREF_BUG_ON(atomic_read(&wf->count) <= 0);
+	if (unlikely(!atomic_add_unless(&wf->count, -1, 1)))
+		__intel_wakeref_put_last(wf, flags);
+}
+
+static inline void
+intel_wakeref_put(struct intel_wakeref *wf)
+{
+	might_sleep();
+	__intel_wakeref_put(wf, 0);
+}
+
+static inline void
+intel_wakeref_put_async(struct intel_wakeref *wf)
+{
+	__intel_wakeref_put(wf, INTEL_WAKEREF_PUT_ASYNC);
+}
+
+static inline void
+intel_wakeref_put_delay(struct intel_wakeref *wf, unsigned long delay)
+{
+	__intel_wakeref_put(wf,
+			    INTEL_WAKEREF_PUT_ASYNC |
+			    FIELD_PREP(INTEL_WAKEREF_PUT_DELAY, delay));
+}
+
+static inline void
+intel_wakeref_might_put(struct intel_wakeref *wf)
+{
+	might_lock(&wf->mutex);
+}
+
+/**
+ * intel_wakeref_lock: Lock the wakeref (mutex)
+ * @wf: the wakeref
+ *
+ * Locks the wakeref to prevent it being acquired or released. New users
+ * can still adjust the counter, but the wakeref itself (and callback)
+ * cannot be acquired or released.
+ */
+static inline void
+intel_wakeref_lock(struct intel_wakeref *wf)
+	__acquires(wf->mutex)
+{
+	mutex_lock(&wf->mutex);
+}
+
+/**
+ * intel_wakeref_unlock: Unlock the wakeref
+ * @wf: the wakeref
+ *
+ * Releases a previously acquired intel_wakeref_lock().
+ */
+static inline void
+intel_wakeref_unlock(struct intel_wakeref *wf)
+	__releases(wf->mutex)
+{
+	mutex_unlock(&wf->mutex);
+}
+
+/**
+ * intel_wakeref_unlock_wait: Wait until the active callback is complete
+ * @wf: the wakeref
+ *
+ * Waits for the active callback (under the @wf->mutex or another CPU) is
+ * complete.
+ */
+static inline void
+intel_wakeref_unlock_wait(struct intel_wakeref *wf)
+{
+	mutex_lock(&wf->mutex);
+	mutex_unlock(&wf->mutex);
+	flush_delayed_work(&wf->work);
+}
+
+/**
+ * intel_wakeref_is_active: Query whether the wakeref is currently held
+ * @wf: the wakeref
+ *
+ * Returns: true if the wakeref is currently held.
+ */
+static inline bool
+intel_wakeref_is_active(const struct intel_wakeref *wf)
+{
+	return READ_ONCE(wf->wakeref);
+}
+
+/**
+ * __intel_wakeref_defer_park: Defer the current park callback
+ * @wf: the wakeref
+ */
+static inline void
+__intel_wakeref_defer_park(struct intel_wakeref *wf)
+{
+	lockdep_assert_held(&wf->mutex);
+	INTEL_WAKEREF_BUG_ON(atomic_read(&wf->count));
+	atomic_set_release(&wf->count, 1);
+}
+
+/**
+ * intel_wakeref_wait_for_idle: Wait until the wakeref is idle
+ * @wf: the wakeref
+ *
+ * Wait for the earlier asynchronous release of the wakeref. Note
+ * this will wait for any third party as well, so make sure you only wait
+ * when you have control over the wakeref and trust no one else is acquiring
+ * it.
+ *
+ * Return: 0 on success, error code if killed.
+ */
+int intel_wakeref_wait_for_idle(struct intel_wakeref *wf);
+
+struct intel_wakeref_auto {
+	struct intel_runtime_pm *rpm;
+	struct timer_list timer;
+	intel_wakeref_t wakeref;
+	spinlock_t lock;
+	refcount_t count;
+};
+
+/**
+ * intel_wakeref_auto: Delay the runtime-pm autosuspend
+ * @wf: the wakeref
+ * @timeout: relative timeout in jiffies
+ *
+ * The runtime-pm core uses a suspend delay after the last wakeref
+ * is released before triggering runtime suspend of the device. That
+ * delay is configurable via sysfs with little regard to the device
+ * characteristics. Instead, we want to tune the autosuspend based on our
+ * HW knowledge. intel_wakeref_auto() delays the sleep by the supplied
+ * timeout.
+ *
+ * Pass @timeout = 0 to cancel a previous autosuspend by executing the
+ * suspend immediately.
+ */
+void intel_wakeref_auto(struct intel_wakeref_auto *wf, unsigned long timeout);
+
+void intel_wakeref_auto_init(struct intel_wakeref_auto *wf,
+			     struct intel_runtime_pm *rpm);
+void intel_wakeref_auto_fini(struct intel_wakeref_auto *wf);
+
+#endif /* INTEL_WAKEREF_H */
diff --git a/drivers/gpu/drm/i915/intel_wopcm.c b/drivers/gpu/drm/i915/intel_wopcm.c
new file mode 100644
index 000000000..322fb9eeb
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_wopcm.c
@@ -0,0 +1,305 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2017-2019 Intel Corporation
+ */
+
+#include "intel_wopcm.h"
+#include "i915_drv.h"
+
+/**
+ * DOC: WOPCM Layout
+ *
+ * The layout of the WOPCM will be fixed after writing to GuC WOPCM size and
+ * offset registers whose values are calculated and determined by HuC/GuC
+ * firmware size and set of hardware requirements/restrictions as shown below:
+ *
+ * ::
+ *
+ *    +=========> +====================+ <== WOPCM Top
+ *    ^           |  HW contexts RSVD  |
+ *    |     +===> +====================+ <== GuC WOPCM Top
+ *    |     ^     |                    |
+ *    |     |     |                    |
+ *    |     |     |                    |
+ *    |    GuC    |                    |
+ *    |   WOPCM   |                    |
+ *    |    Size   +--------------------+
+ *  WOPCM   |     |    GuC FW RSVD     |
+ *    |     |     +--------------------+
+ *    |     |     |   GuC Stack RSVD   |
+ *    |     |     +------------------- +
+ *    |     v     |   GuC WOPCM RSVD   |
+ *    |     +===> +====================+ <== GuC WOPCM base
+ *    |           |     WOPCM RSVD     |
+ *    |           +------------------- + <== HuC Firmware Top
+ *    v           |      HuC FW        |
+ *    +=========> +====================+ <== WOPCM Base
+ *
+ * GuC accessible WOPCM starts at GuC WOPCM base and ends at GuC WOPCM top.
+ * The top part of the WOPCM is reserved for hardware contexts (e.g. RC6
+ * context).
+ */
+
+/* Default WOPCM size is 2MB from Gen11, 1MB on previous platforms */
+#define GEN11_WOPCM_SIZE		SZ_2M
+#define GEN9_WOPCM_SIZE			SZ_1M
+#define MAX_WOPCM_SIZE			SZ_8M
+/* 16KB WOPCM (RSVD WOPCM) is reserved from HuC firmware top. */
+#define WOPCM_RESERVED_SIZE		SZ_16K
+
+/* 16KB reserved at the beginning of GuC WOPCM. */
+#define GUC_WOPCM_RESERVED		SZ_16K
+/* 8KB from GUC_WOPCM_RESERVED is reserved for GuC stack. */
+#define GUC_WOPCM_STACK_RESERVED	SZ_8K
+
+/* GuC WOPCM Offset value needs to be aligned to 16KB. */
+#define GUC_WOPCM_OFFSET_ALIGNMENT	(1UL << GUC_WOPCM_OFFSET_SHIFT)
+
+/* 24KB at the end of WOPCM is reserved for RC6 CTX on BXT. */
+#define BXT_WOPCM_RC6_CTX_RESERVED	(SZ_16K + SZ_8K)
+/* 36KB WOPCM reserved at the end of WOPCM on ICL. */
+#define ICL_WOPCM_HW_CTX_RESERVED	(SZ_32K + SZ_4K)
+
+/* 128KB from GUC_WOPCM_RESERVED is reserved for FW on Gen9. */
+#define GEN9_GUC_FW_RESERVED	SZ_128K
+#define GEN9_GUC_WOPCM_OFFSET	(GUC_WOPCM_RESERVED + GEN9_GUC_FW_RESERVED)
+
+static inline struct drm_i915_private *wopcm_to_i915(struct intel_wopcm *wopcm)
+{
+	return container_of(wopcm, struct drm_i915_private, wopcm);
+}
+
+/**
+ * intel_wopcm_init_early() - Early initialization of the WOPCM.
+ * @wopcm: pointer to intel_wopcm.
+ *
+ * Setup the size of WOPCM which will be used by later on WOPCM partitioning.
+ */
+void intel_wopcm_init_early(struct intel_wopcm *wopcm)
+{
+	struct drm_i915_private *i915 = wopcm_to_i915(wopcm);
+
+	if (!HAS_GT_UC(i915))
+		return;
+
+	if (GRAPHICS_VER(i915) >= 11)
+		wopcm->size = GEN11_WOPCM_SIZE;
+	else
+		wopcm->size = GEN9_WOPCM_SIZE;
+
+	drm_dbg(&i915->drm, "WOPCM: %uK\n", wopcm->size / 1024);
+}
+
+static u32 context_reserved_size(struct drm_i915_private *i915)
+{
+	if (IS_GEN9_LP(i915))
+		return BXT_WOPCM_RC6_CTX_RESERVED;
+	else if (GRAPHICS_VER(i915) >= 11)
+		return ICL_WOPCM_HW_CTX_RESERVED;
+	else
+		return 0;
+}
+
+static bool gen9_check_dword_gap(struct drm_i915_private *i915,
+				 u32 guc_wopcm_base, u32 guc_wopcm_size)
+{
+	u32 offset;
+
+	/*
+	 * GuC WOPCM size shall be at least a dword larger than the offset from
+	 * WOPCM base (GuC WOPCM offset from WOPCM base + GEN9_GUC_WOPCM_OFFSET)
+	 * due to hardware limitation on Gen9.
+	 */
+	offset = guc_wopcm_base + GEN9_GUC_WOPCM_OFFSET;
+	if (offset > guc_wopcm_size ||
+	    (guc_wopcm_size - offset) < sizeof(u32)) {
+		drm_err(&i915->drm,
+			"WOPCM: invalid GuC region size: %uK < %uK\n",
+			guc_wopcm_size / SZ_1K,
+			(u32)(offset + sizeof(u32)) / SZ_1K);
+		return false;
+	}
+
+	return true;
+}
+
+static bool gen9_check_huc_fw_fits(struct drm_i915_private *i915,
+				   u32 guc_wopcm_size, u32 huc_fw_size)
+{
+	/*
+	 * On Gen9, hardware requires the total available GuC WOPCM
+	 * size to be larger than or equal to HuC firmware size. Otherwise,
+	 * firmware uploading would fail.
+	 */
+	if (huc_fw_size > guc_wopcm_size - GUC_WOPCM_RESERVED) {
+		drm_err(&i915->drm, "WOPCM: no space for %s: %uK < %uK\n",
+			intel_uc_fw_type_repr(INTEL_UC_FW_TYPE_HUC),
+			(guc_wopcm_size - GUC_WOPCM_RESERVED) / SZ_1K,
+			huc_fw_size / 1024);
+		return false;
+	}
+
+	return true;
+}
+
+static bool check_hw_restrictions(struct drm_i915_private *i915,
+				  u32 guc_wopcm_base, u32 guc_wopcm_size,
+				  u32 huc_fw_size)
+{
+	if (GRAPHICS_VER(i915) == 9 && !gen9_check_dword_gap(i915, guc_wopcm_base,
+							     guc_wopcm_size))
+		return false;
+
+	if (GRAPHICS_VER(i915) == 9 &&
+	    !gen9_check_huc_fw_fits(i915, guc_wopcm_size, huc_fw_size))
+		return false;
+
+	return true;
+}
+
+static bool __check_layout(struct drm_i915_private *i915, u32 wopcm_size,
+			   u32 guc_wopcm_base, u32 guc_wopcm_size,
+			   u32 guc_fw_size, u32 huc_fw_size)
+{
+	const u32 ctx_rsvd = context_reserved_size(i915);
+	u32 size;
+
+	size = wopcm_size - ctx_rsvd;
+	if (unlikely(range_overflows(guc_wopcm_base, guc_wopcm_size, size))) {
+		drm_err(&i915->drm,
+			"WOPCM: invalid GuC region layout: %uK + %uK > %uK\n",
+			guc_wopcm_base / SZ_1K, guc_wopcm_size / SZ_1K,
+			size / SZ_1K);
+		return false;
+	}
+
+	size = guc_fw_size + GUC_WOPCM_RESERVED + GUC_WOPCM_STACK_RESERVED;
+	if (unlikely(guc_wopcm_size < size)) {
+		drm_err(&i915->drm, "WOPCM: no space for %s: %uK < %uK\n",
+			intel_uc_fw_type_repr(INTEL_UC_FW_TYPE_GUC),
+			guc_wopcm_size / SZ_1K, size / SZ_1K);
+		return false;
+	}
+
+	size = huc_fw_size + WOPCM_RESERVED_SIZE;
+	if (unlikely(guc_wopcm_base < size)) {
+		drm_err(&i915->drm, "WOPCM: no space for %s: %uK < %uK\n",
+			intel_uc_fw_type_repr(INTEL_UC_FW_TYPE_HUC),
+			guc_wopcm_base / SZ_1K, size / SZ_1K);
+		return false;
+	}
+
+	return check_hw_restrictions(i915, guc_wopcm_base, guc_wopcm_size,
+				     huc_fw_size);
+}
+
+static bool __wopcm_regs_locked(struct intel_uncore *uncore,
+				u32 *guc_wopcm_base, u32 *guc_wopcm_size)
+{
+	u32 reg_base = intel_uncore_read(uncore, DMA_GUC_WOPCM_OFFSET);
+	u32 reg_size = intel_uncore_read(uncore, GUC_WOPCM_SIZE);
+
+	if (!(reg_size & GUC_WOPCM_SIZE_LOCKED) ||
+	    !(reg_base & GUC_WOPCM_OFFSET_VALID))
+		return false;
+
+	*guc_wopcm_base = reg_base & GUC_WOPCM_OFFSET_MASK;
+	*guc_wopcm_size = reg_size & GUC_WOPCM_SIZE_MASK;
+	return true;
+}
+
+static bool __wopcm_regs_writable(struct intel_uncore *uncore)
+{
+	if (!HAS_GUC_DEPRIVILEGE(uncore->i915))
+		return true;
+
+	return intel_uncore_read(uncore, GUC_SHIM_CONTROL2) & GUC_IS_PRIVILEGED;
+}
+
+/**
+ * intel_wopcm_init() - Initialize the WOPCM structure.
+ * @wopcm: pointer to intel_wopcm.
+ *
+ * This function will partition WOPCM space based on GuC and HuC firmware sizes
+ * and will allocate max remaining for use by GuC. This function will also
+ * enforce platform dependent hardware restrictions on GuC WOPCM offset and
+ * size. It will fail the WOPCM init if any of these checks fail, so that the
+ * following WOPCM registers setup and GuC firmware uploading would be aborted.
+ */
+void intel_wopcm_init(struct intel_wopcm *wopcm)
+{
+	struct drm_i915_private *i915 = wopcm_to_i915(wopcm);
+	struct intel_gt *gt = to_gt(i915);
+	u32 guc_fw_size = intel_uc_fw_get_upload_size(&gt->uc.guc.fw);
+	u32 huc_fw_size = intel_uc_fw_get_upload_size(&gt->uc.huc.fw);
+	u32 ctx_rsvd = context_reserved_size(i915);
+	u32 wopcm_size = wopcm->size;
+	u32 guc_wopcm_base;
+	u32 guc_wopcm_size;
+
+	if (!guc_fw_size)
+		return;
+
+	GEM_BUG_ON(!wopcm_size);
+	GEM_BUG_ON(wopcm->guc.base);
+	GEM_BUG_ON(wopcm->guc.size);
+	GEM_BUG_ON(guc_fw_size >= wopcm_size);
+	GEM_BUG_ON(huc_fw_size >= wopcm_size);
+	GEM_BUG_ON(ctx_rsvd + WOPCM_RESERVED_SIZE >= wopcm_size);
+
+	if (i915_inject_probe_failure(i915))
+		return;
+
+	if (__wopcm_regs_locked(gt->uncore, &guc_wopcm_base, &guc_wopcm_size)) {
+		drm_dbg(&i915->drm, "GuC WOPCM is already locked [%uK, %uK)\n",
+			guc_wopcm_base / SZ_1K, guc_wopcm_size / SZ_1K);
+		/*
+		 * Note that to keep things simple (i.e. avoid different
+		 * defines per platform) our WOPCM math doesn't always use the
+		 * actual WOPCM size, but a value that is less or equal to it.
+		 * This is perfectly fine when i915 programs the registers, but
+		 * on platforms with GuC deprivilege the registers are not
+		 * writable from i915 and are instead pre-programmed by the
+		 * bios/IFWI, so there might be a mismatch of sizes.
+		 * Instead of handling the size difference, we trust that the
+		 * programmed values make sense and disable the relevant check
+		 * by using the maximum possible WOPCM size in the verification
+		 * math. In the extremely unlikely case that the registers
+		 * were pre-programmed with an invalid value, we will still
+		 * gracefully fail later during the GuC/HuC dma.
+		 */
+		if (!__wopcm_regs_writable(gt->uncore))
+			wopcm_size = MAX_WOPCM_SIZE;
+
+		goto check;
+	}
+
+	/*
+	 * Aligned value of guc_wopcm_base will determine available WOPCM space
+	 * for HuC firmware and mandatory reserved area.
+	 */
+	guc_wopcm_base = huc_fw_size + WOPCM_RESERVED_SIZE;
+	guc_wopcm_base = ALIGN(guc_wopcm_base, GUC_WOPCM_OFFSET_ALIGNMENT);
+
+	/*
+	 * Need to clamp guc_wopcm_base now to make sure the following math is
+	 * correct. Formal check of whole WOPCM layout will be done below.
+	 */
+	guc_wopcm_base = min(guc_wopcm_base, wopcm_size - ctx_rsvd);
+
+	/* Aligned remainings of usable WOPCM space can be assigned to GuC. */
+	guc_wopcm_size = wopcm_size - ctx_rsvd - guc_wopcm_base;
+	guc_wopcm_size &= GUC_WOPCM_SIZE_MASK;
+
+	drm_dbg(&i915->drm, "Calculated GuC WOPCM [%uK, %uK)\n",
+		guc_wopcm_base / SZ_1K, guc_wopcm_size / SZ_1K);
+
+check:
+	if (__check_layout(i915, wopcm_size, guc_wopcm_base, guc_wopcm_size,
+			   guc_fw_size, huc_fw_size)) {
+		wopcm->guc.base = guc_wopcm_base;
+		wopcm->guc.size = guc_wopcm_size;
+		GEM_BUG_ON(!wopcm->guc.base);
+		GEM_BUG_ON(!wopcm->guc.size);
+	}
+}
diff --git a/drivers/gpu/drm/i915/intel_wopcm.h b/drivers/gpu/drm/i915/intel_wopcm.h
new file mode 100644
index 000000000..17d6aa860
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_wopcm.h
@@ -0,0 +1,60 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2017-2018 Intel Corporation
+ */
+
+#ifndef _INTEL_WOPCM_H_
+#define _INTEL_WOPCM_H_
+
+#include <linux/types.h>
+
+/**
+ * struct intel_wopcm - Overall WOPCM info and WOPCM regions.
+ * @size: Size of overall WOPCM.
+ * @guc: GuC WOPCM Region info.
+ * @guc.base: GuC WOPCM base which is offset from WOPCM base.
+ * @guc.size: Size of the GuC WOPCM region.
+ */
+struct intel_wopcm {
+	u32 size;
+	struct {
+		u32 base;
+		u32 size;
+	} guc;
+};
+
+/**
+ * intel_wopcm_guc_base()
+ * @wopcm:	intel_wopcm structure
+ *
+ * Returns the base of the WOPCM shadowed region.
+ *
+ * Returns:
+ * 0 if GuC is not present or not in use.
+ * Otherwise, the GuC WOPCM base.
+ */
+static inline u32 intel_wopcm_guc_base(struct intel_wopcm *wopcm)
+{
+	return wopcm->guc.base;
+}
+
+/**
+ * intel_wopcm_guc_size()
+ * @wopcm:	intel_wopcm structure
+ *
+ * Returns size of the WOPCM shadowed region.
+ *
+ * Returns:
+ * 0 if GuC is not present or not in use.
+ * Otherwise, the GuC WOPCM size.
+ */
+static inline u32 intel_wopcm_guc_size(struct intel_wopcm *wopcm)
+{
+	return wopcm->guc.size;
+}
+
+void intel_wopcm_init_early(struct intel_wopcm *wopcm);
+void intel_wopcm_init(struct intel_wopcm *wopcm);
+
+#endif
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp.c b/drivers/gpu/drm/i915/pxp/intel_pxp.c
new file mode 100644
index 000000000..69cdaaddc
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp.c
@@ -0,0 +1,319 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright(c) 2020 Intel Corporation.
+ */
+#include <linux/workqueue.h>
+#include "intel_pxp.h"
+#include "intel_pxp_irq.h"
+#include "intel_pxp_session.h"
+#include "intel_pxp_tee.h"
+#include "gem/i915_gem_context.h"
+#include "gt/intel_context.h"
+#include "i915_drv.h"
+
+/**
+ * DOC: PXP
+ *
+ * PXP (Protected Xe Path) is a feature available in Gen12 and newer platforms.
+ * It allows execution and flip to display of protected (i.e. encrypted)
+ * objects. The SW support is enabled via the CONFIG_DRM_I915_PXP kconfig.
+ *
+ * Objects can opt-in to PXP encryption at creation time via the
+ * I915_GEM_CREATE_EXT_PROTECTED_CONTENT create_ext flag. For objects to be
+ * correctly protected they must be used in conjunction with a context created
+ * with the I915_CONTEXT_PARAM_PROTECTED_CONTENT flag. See the documentation
+ * of those two uapi flags for details and restrictions.
+ *
+ * Protected objects are tied to a pxp session; currently we only support one
+ * session, which i915 manages and whose index is available in the uapi
+ * (I915_PROTECTED_CONTENT_DEFAULT_SESSION) for use in instructions targeting
+ * protected objects.
+ * The session is invalidated by the HW when certain events occur (e.g.
+ * suspend/resume). When this happens, all the objects that were used with the
+ * session are marked as invalid and all contexts marked as using protected
+ * content are banned. Any further attempt at using them in an execbuf call is
+ * rejected, while flips are converted to black frames.
+ *
+ * Some of the PXP setup operations are performed by the Management Engine,
+ * which is handled by the mei driver; communication between i915 and mei is
+ * performed via the mei_pxp component module.
+ */
+
+struct intel_gt *pxp_to_gt(const struct intel_pxp *pxp)
+{
+	return container_of(pxp, struct intel_gt, pxp);
+}
+
+bool intel_pxp_is_enabled(const struct intel_pxp *pxp)
+{
+	return pxp->ce;
+}
+
+bool intel_pxp_is_active(const struct intel_pxp *pxp)
+{
+	return pxp->arb_is_valid;
+}
+
+/* KCR register definitions */
+#define KCR_INIT _MMIO(0x320f0)
+/* Setting KCR Init bit is required after system boot */
+#define KCR_INIT_ALLOW_DISPLAY_ME_WRITES REG_BIT(14)
+
+static void kcr_pxp_enable(struct intel_gt *gt)
+{
+	intel_uncore_write(gt->uncore, KCR_INIT,
+			   _MASKED_BIT_ENABLE(KCR_INIT_ALLOW_DISPLAY_ME_WRITES));
+}
+
+static void kcr_pxp_disable(struct intel_gt *gt)
+{
+	intel_uncore_write(gt->uncore, KCR_INIT,
+			   _MASKED_BIT_DISABLE(KCR_INIT_ALLOW_DISPLAY_ME_WRITES));
+}
+
+static int create_vcs_context(struct intel_pxp *pxp)
+{
+	static struct lock_class_key pxp_lock;
+	struct intel_gt *gt = pxp_to_gt(pxp);
+	struct intel_engine_cs *engine;
+	struct intel_context *ce;
+	int i;
+
+	/*
+	 * Find the first VCS engine present. We're guaranteed there is one
+	 * if we're in this function due to the check in has_pxp
+	 */
+	for (i = 0, engine = NULL; !engine; i++)
+		engine = gt->engine_class[VIDEO_DECODE_CLASS][i];
+
+	GEM_BUG_ON(!engine || engine->class != VIDEO_DECODE_CLASS);
+
+	ce = intel_engine_create_pinned_context(engine, engine->gt->vm, SZ_4K,
+						I915_GEM_HWS_PXP_ADDR,
+						&pxp_lock, "pxp_context");
+	if (IS_ERR(ce)) {
+		drm_err(&gt->i915->drm, "failed to create VCS ctx for PXP\n");
+		return PTR_ERR(ce);
+	}
+
+	pxp->ce = ce;
+
+	return 0;
+}
+
+static void destroy_vcs_context(struct intel_pxp *pxp)
+{
+	intel_engine_destroy_pinned_context(fetch_and_zero(&pxp->ce));
+}
+
+void intel_pxp_init(struct intel_pxp *pxp)
+{
+	struct intel_gt *gt = pxp_to_gt(pxp);
+	int ret;
+
+	if (!HAS_PXP(gt->i915))
+		return;
+
+	mutex_init(&pxp->tee_mutex);
+
+	/*
+	 * we'll use the completion to check if there is a termination pending,
+	 * so we start it as completed and we reinit it when a termination
+	 * is triggered.
+	 */
+	init_completion(&pxp->termination);
+	complete_all(&pxp->termination);
+
+	mutex_init(&pxp->arb_mutex);
+	INIT_WORK(&pxp->session_work, intel_pxp_session_work);
+
+	ret = create_vcs_context(pxp);
+	if (ret)
+		return;
+
+	ret = intel_pxp_tee_component_init(pxp);
+	if (ret)
+		goto out_context;
+
+	drm_info(&gt->i915->drm, "Protected Xe Path (PXP) protected content support initialized\n");
+
+	return;
+
+out_context:
+	destroy_vcs_context(pxp);
+}
+
+void intel_pxp_fini(struct intel_pxp *pxp)
+{
+	if (!intel_pxp_is_enabled(pxp))
+		return;
+
+	pxp->arb_is_valid = false;
+
+	intel_pxp_tee_component_fini(pxp);
+
+	destroy_vcs_context(pxp);
+}
+
+void intel_pxp_mark_termination_in_progress(struct intel_pxp *pxp)
+{
+	pxp->arb_is_valid = false;
+	reinit_completion(&pxp->termination);
+}
+
+static void pxp_queue_termination(struct intel_pxp *pxp)
+{
+	struct intel_gt *gt = pxp_to_gt(pxp);
+
+	/*
+	 * We want to get the same effect as if we received a termination
+	 * interrupt, so just pretend that we did.
+	 */
+	spin_lock_irq(gt->irq_lock);
+	intel_pxp_mark_termination_in_progress(pxp);
+	pxp->session_events |= PXP_TERMINATION_REQUEST;
+	queue_work(system_unbound_wq, &pxp->session_work);
+	spin_unlock_irq(gt->irq_lock);
+}
+
+static bool pxp_component_bound(struct intel_pxp *pxp)
+{
+	bool bound = false;
+
+	mutex_lock(&pxp->tee_mutex);
+	if (pxp->pxp_component)
+		bound = true;
+	mutex_unlock(&pxp->tee_mutex);
+
+	return bound;
+}
+
+/*
+ * the arb session is restarted from the irq work when we receive the
+ * termination completion interrupt
+ */
+int intel_pxp_start(struct intel_pxp *pxp)
+{
+	int ret = 0;
+
+	if (!intel_pxp_is_enabled(pxp))
+		return -ENODEV;
+
+	if (wait_for(pxp_component_bound(pxp), 250))
+		return -ENXIO;
+
+	mutex_lock(&pxp->arb_mutex);
+
+	if (pxp->arb_is_valid)
+		goto unlock;
+
+	pxp_queue_termination(pxp);
+
+	if (!wait_for_completion_timeout(&pxp->termination,
+					msecs_to_jiffies(250))) {
+		ret = -ETIMEDOUT;
+		goto unlock;
+	}
+
+	/* make sure the compiler doesn't optimize the double access */
+	barrier();
+
+	if (!pxp->arb_is_valid)
+		ret = -EIO;
+
+unlock:
+	mutex_unlock(&pxp->arb_mutex);
+	return ret;
+}
+
+void intel_pxp_init_hw(struct intel_pxp *pxp)
+{
+	kcr_pxp_enable(pxp_to_gt(pxp));
+	intel_pxp_irq_enable(pxp);
+}
+
+void intel_pxp_fini_hw(struct intel_pxp *pxp)
+{
+	kcr_pxp_disable(pxp_to_gt(pxp));
+
+	intel_pxp_irq_disable(pxp);
+}
+
+int intel_pxp_key_check(struct intel_pxp *pxp,
+			struct drm_i915_gem_object *obj,
+			bool assign)
+{
+	if (!intel_pxp_is_active(pxp))
+		return -ENODEV;
+
+	if (!i915_gem_object_is_protected(obj))
+		return -EINVAL;
+
+	GEM_BUG_ON(!pxp->key_instance);
+
+	/*
+	 * If this is the first time we're using this object, it's not
+	 * encrypted yet; it will be encrypted with the current key, so mark it
+	 * as such. If the object is already encrypted, check instead if the
+	 * used key is still valid.
+	 */
+	if (!obj->pxp_key_instance && assign)
+		obj->pxp_key_instance = pxp->key_instance;
+
+	if (obj->pxp_key_instance != pxp->key_instance)
+		return -ENOEXEC;
+
+	return 0;
+}
+
+void intel_pxp_invalidate(struct intel_pxp *pxp)
+{
+	struct drm_i915_private *i915 = pxp_to_gt(pxp)->i915;
+	struct i915_gem_context *ctx, *cn;
+
+	/* ban all contexts marked as protected */
+	spin_lock_irq(&i915->gem.contexts.lock);
+	list_for_each_entry_safe(ctx, cn, &i915->gem.contexts.list, link) {
+		struct i915_gem_engines_iter it;
+		struct intel_context *ce;
+
+		if (!kref_get_unless_zero(&ctx->ref))
+			continue;
+
+		if (likely(!i915_gem_context_uses_protected_content(ctx))) {
+			i915_gem_context_put(ctx);
+			continue;
+		}
+
+		spin_unlock_irq(&i915->gem.contexts.lock);
+
+		/*
+		 * By the time we get here we are either going to suspend with
+		 * quiesced execution or the HW keys are already long gone and
+		 * in this case it is worthless to attempt to close the context
+		 * and wait for its execution. It will hang the GPU if it has
+		 * not already. So, as a fast mitigation, we can ban the
+		 * context as quick as we can. That might race with the
+		 * execbuffer, but currently this is the best that can be done.
+		 */
+		for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it)
+			intel_context_ban(ce, NULL);
+		i915_gem_context_unlock_engines(ctx);
+
+		/*
+		 * The context has been banned, no need to keep the wakeref.
+		 * This is safe from races because the only other place this
+		 * is touched is context_release and we're holding a ctx ref
+		 */
+		if (ctx->pxp_wakeref) {
+			intel_runtime_pm_put(&i915->runtime_pm,
+					     ctx->pxp_wakeref);
+			ctx->pxp_wakeref = 0;
+		}
+
+		spin_lock_irq(&i915->gem.contexts.lock);
+		list_safe_reset_next(ctx, cn, link);
+		i915_gem_context_put(ctx);
+	}
+	spin_unlock_irq(&i915->gem.contexts.lock);
+}
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp.h b/drivers/gpu/drm/i915/pxp/intel_pxp.h
new file mode 100644
index 000000000..73847e535
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp.h
@@ -0,0 +1,67 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright(c) 2020, Intel Corporation. All rights reserved.
+ */
+
+#ifndef __INTEL_PXP_H__
+#define __INTEL_PXP_H__
+
+#include <linux/errno.h>
+#include <linux/types.h>
+
+struct intel_pxp;
+struct drm_i915_gem_object;
+
+#ifdef CONFIG_DRM_I915_PXP
+struct intel_gt *pxp_to_gt(const struct intel_pxp *pxp);
+bool intel_pxp_is_enabled(const struct intel_pxp *pxp);
+bool intel_pxp_is_active(const struct intel_pxp *pxp);
+
+void intel_pxp_init(struct intel_pxp *pxp);
+void intel_pxp_fini(struct intel_pxp *pxp);
+
+void intel_pxp_init_hw(struct intel_pxp *pxp);
+void intel_pxp_fini_hw(struct intel_pxp *pxp);
+
+void intel_pxp_mark_termination_in_progress(struct intel_pxp *pxp);
+
+int intel_pxp_start(struct intel_pxp *pxp);
+
+int intel_pxp_key_check(struct intel_pxp *pxp,
+			struct drm_i915_gem_object *obj,
+			bool assign);
+
+void intel_pxp_invalidate(struct intel_pxp *pxp);
+#else
+static inline void intel_pxp_init(struct intel_pxp *pxp)
+{
+}
+
+static inline void intel_pxp_fini(struct intel_pxp *pxp)
+{
+}
+
+static inline int intel_pxp_start(struct intel_pxp *pxp)
+{
+	return -ENODEV;
+}
+
+static inline bool intel_pxp_is_enabled(const struct intel_pxp *pxp)
+{
+	return false;
+}
+
+static inline bool intel_pxp_is_active(const struct intel_pxp *pxp)
+{
+	return false;
+}
+
+static inline int intel_pxp_key_check(struct intel_pxp *pxp,
+				      struct drm_i915_gem_object *obj,
+				      bool assign)
+{
+	return -ENODEV;
+}
+#endif
+
+#endif /* __INTEL_PXP_H__ */
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_cmd.c b/drivers/gpu/drm/i915/pxp/intel_pxp_cmd.c
new file mode 100644
index 000000000..f41e45763
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_cmd.c
@@ -0,0 +1,141 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright(c) 2020, Intel Corporation. All rights reserved.
+ */
+
+#include "intel_pxp.h"
+#include "intel_pxp_cmd.h"
+#include "intel_pxp_session.h"
+#include "gt/intel_context.h"
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_ring.h"
+
+#include "i915_trace.h"
+
+/* stall until prior PXP and MFX/HCP/HUC objects are cmopleted */
+#define MFX_WAIT_PXP (MFX_WAIT | \
+		      MFX_WAIT_DW0_PXP_SYNC_CONTROL_FLAG | \
+		      MFX_WAIT_DW0_MFX_SYNC_CONTROL_FLAG)
+
+static u32 *pxp_emit_session_selection(u32 *cs, u32 idx)
+{
+	*cs++ = MFX_WAIT_PXP;
+
+	/* pxp off */
+	*cs++ = MI_FLUSH_DW;
+	*cs++ = 0;
+	*cs++ = 0;
+
+	/* select session */
+	*cs++ = MI_SET_APPID | MI_SET_APPID_SESSION_ID(idx);
+
+	*cs++ = MFX_WAIT_PXP;
+
+	/* pxp on */
+	*cs++ = MI_FLUSH_DW | MI_FLUSH_DW_PROTECTED_MEM_EN |
+		MI_FLUSH_DW_OP_STOREDW | MI_FLUSH_DW_STORE_INDEX;
+	*cs++ = I915_GEM_HWS_PXP_ADDR | MI_FLUSH_DW_USE_GTT;
+	*cs++ = 0;
+
+	*cs++ = MFX_WAIT_PXP;
+
+	return cs;
+}
+
+static u32 *pxp_emit_inline_termination(u32 *cs)
+{
+	/* session inline termination */
+	*cs++ = CRYPTO_KEY_EXCHANGE;
+	*cs++ = 0;
+
+	return cs;
+}
+
+static u32 *pxp_emit_session_termination(u32 *cs, u32 idx)
+{
+	cs = pxp_emit_session_selection(cs, idx);
+	cs = pxp_emit_inline_termination(cs);
+
+	return cs;
+}
+
+static u32 *pxp_emit_wait(u32 *cs)
+{
+	/* wait for cmds to go through */
+	*cs++ = MFX_WAIT_PXP;
+	*cs++ = 0;
+
+	return cs;
+}
+
+/*
+ * if we ever need to terminate more than one session, we can submit multiple
+ * selections and terminations back-to-back with a single wait at the end
+ */
+#define SELECTION_LEN 10
+#define TERMINATION_LEN 2
+#define SESSION_TERMINATION_LEN(x) ((SELECTION_LEN + TERMINATION_LEN) * (x))
+#define WAIT_LEN 2
+
+static void pxp_request_commit(struct i915_request *rq)
+{
+	struct i915_sched_attr attr = { .priority = I915_PRIORITY_MAX };
+	struct intel_timeline * const tl = i915_request_timeline(rq);
+
+	lockdep_unpin_lock(&tl->mutex, rq->cookie);
+
+	trace_i915_request_add(rq);
+	__i915_request_commit(rq);
+	__i915_request_queue(rq, &attr);
+
+	mutex_unlock(&tl->mutex);
+}
+
+int intel_pxp_terminate_session(struct intel_pxp *pxp, u32 id)
+{
+	struct i915_request *rq;
+	struct intel_context *ce = pxp->ce;
+	u32 *cs;
+	int err = 0;
+
+	if (!intel_pxp_is_enabled(pxp))
+		return 0;
+
+	rq = i915_request_create(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	if (ce->engine->emit_init_breadcrumb) {
+		err = ce->engine->emit_init_breadcrumb(rq);
+		if (err)
+			goto out_rq;
+	}
+
+	cs = intel_ring_begin(rq, SESSION_TERMINATION_LEN(1) + WAIT_LEN);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		goto out_rq;
+	}
+
+	cs = pxp_emit_session_termination(cs, id);
+	cs = pxp_emit_wait(cs);
+
+	intel_ring_advance(rq, cs);
+
+out_rq:
+	i915_request_get(rq);
+
+	if (unlikely(err))
+		i915_request_set_error_once(rq, err);
+
+	pxp_request_commit(rq);
+
+	if (!err && i915_request_wait(rq, 0, HZ / 5) < 0)
+		err = -ETIME;
+
+	i915_request_put(rq);
+
+	return err;
+}
+
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_cmd.h b/drivers/gpu/drm/i915/pxp/intel_pxp_cmd.h
new file mode 100644
index 000000000..6d6299543
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_cmd.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright(c) 2020, Intel Corporation. All rights reserved.
+ */
+
+#ifndef __INTEL_PXP_CMD_H__
+#define __INTEL_PXP_CMD_H__
+
+#include <linux/types.h>
+
+struct intel_pxp;
+
+int intel_pxp_terminate_session(struct intel_pxp *pxp, u32 idx);
+
+#endif /* __INTEL_PXP_CMD_H__ */
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_debugfs.c b/drivers/gpu/drm/i915/pxp/intel_pxp_debugfs.c
new file mode 100644
index 000000000..4359e8be4
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_debugfs.c
@@ -0,0 +1,81 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/debugfs.h>
+#include <linux/string_helpers.h>
+
+#include <drm/drm_print.h>
+
+#include "gt/intel_gt_debugfs.h"
+#include "i915_drv.h"
+#include "intel_pxp.h"
+#include "intel_pxp_debugfs.h"
+#include "intel_pxp_irq.h"
+
+static int pxp_info_show(struct seq_file *m, void *data)
+{
+	struct intel_pxp *pxp = m->private;
+	struct drm_printer p = drm_seq_file_printer(m);
+	bool enabled = intel_pxp_is_enabled(pxp);
+
+	if (!enabled) {
+		drm_printf(&p, "pxp disabled\n");
+		return 0;
+	}
+
+	drm_printf(&p, "active: %s\n", str_yes_no(intel_pxp_is_active(pxp)));
+	drm_printf(&p, "instance counter: %u\n", pxp->key_instance);
+
+	return 0;
+}
+DEFINE_INTEL_GT_DEBUGFS_ATTRIBUTE(pxp_info);
+
+static int pxp_terminate_get(void *data, u64 *val)
+{
+	/* nothing to read */
+	return -EPERM;
+}
+
+static int pxp_terminate_set(void *data, u64 val)
+{
+	struct intel_pxp *pxp = data;
+	struct intel_gt *gt = pxp_to_gt(pxp);
+
+	if (!intel_pxp_is_active(pxp))
+		return -ENODEV;
+
+	/* simulate a termination interrupt */
+	spin_lock_irq(gt->irq_lock);
+	intel_pxp_irq_handler(pxp, GEN12_DISPLAY_PXP_STATE_TERMINATED_INTERRUPT);
+	spin_unlock_irq(gt->irq_lock);
+
+	if (!wait_for_completion_timeout(&pxp->termination,
+					 msecs_to_jiffies(100)))
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+DEFINE_SIMPLE_ATTRIBUTE(pxp_terminate_fops, pxp_terminate_get, pxp_terminate_set, "%llx\n");
+void intel_pxp_debugfs_register(struct intel_pxp *pxp, struct dentry *gt_root)
+{
+	static const struct intel_gt_debugfs_file files[] = {
+		{ "info", &pxp_info_fops, NULL },
+		{ "terminate_state", &pxp_terminate_fops, NULL },
+	};
+	struct dentry *root;
+
+	if (!gt_root)
+		return;
+
+	if (!HAS_PXP((pxp_to_gt(pxp)->i915)))
+		return;
+
+	root = debugfs_create_dir("pxp", gt_root);
+	if (IS_ERR(root))
+		return;
+
+	intel_gt_debugfs_register_files(root, files, ARRAY_SIZE(files), pxp);
+}
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_debugfs.h b/drivers/gpu/drm/i915/pxp/intel_pxp_debugfs.h
new file mode 100644
index 000000000..7e0c3d2f5
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_debugfs.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __INTEL_PXP_DEBUGFS_H__
+#define __INTEL_PXP_DEBUGFS_H__
+
+struct intel_pxp;
+struct dentry;
+
+#ifdef CONFIG_DRM_I915_PXP
+void intel_pxp_debugfs_register(struct intel_pxp *pxp, struct dentry *root);
+#else
+static inline void
+intel_pxp_debugfs_register(struct intel_pxp *pxp, struct dentry *root)
+{
+}
+#endif
+
+#endif /* __INTEL_PXP_DEBUGFS_H__ */
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_irq.c b/drivers/gpu/drm/i915/pxp/intel_pxp_irq.c
new file mode 100644
index 000000000..c28be4307
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_irq.c
@@ -0,0 +1,102 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright(c) 2020 Intel Corporation.
+ */
+#include <linux/workqueue.h>
+#include "intel_pxp.h"
+#include "intel_pxp_irq.h"
+#include "intel_pxp_session.h"
+#include "gt/intel_gt_irq.h"
+#include "gt/intel_gt_regs.h"
+#include "gt/intel_gt_types.h"
+#include "i915_irq.h"
+#include "i915_reg.h"
+#include "intel_runtime_pm.h"
+
+/**
+ * intel_pxp_irq_handler - Handles PXP interrupts.
+ * @pxp: pointer to pxp struct
+ * @iir: interrupt vector
+ */
+void intel_pxp_irq_handler(struct intel_pxp *pxp, u16 iir)
+{
+	struct intel_gt *gt = pxp_to_gt(pxp);
+
+	if (GEM_WARN_ON(!intel_pxp_is_enabled(pxp)))
+		return;
+
+	lockdep_assert_held(gt->irq_lock);
+
+	if (unlikely(!iir))
+		return;
+
+	if (iir & (GEN12_DISPLAY_PXP_STATE_TERMINATED_INTERRUPT |
+		   GEN12_DISPLAY_APP_TERMINATED_PER_FW_REQ_INTERRUPT)) {
+		/* immediately mark PXP as inactive on termination */
+		intel_pxp_mark_termination_in_progress(pxp);
+		pxp->session_events |= PXP_TERMINATION_REQUEST | PXP_INVAL_REQUIRED;
+	}
+
+	if (iir & GEN12_DISPLAY_STATE_RESET_COMPLETE_INTERRUPT)
+		pxp->session_events |= PXP_TERMINATION_COMPLETE;
+
+	if (pxp->session_events)
+		queue_work(system_unbound_wq, &pxp->session_work);
+}
+
+static inline void __pxp_set_interrupts(struct intel_gt *gt, u32 interrupts)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	const u32 mask = interrupts << 16;
+
+	intel_uncore_write(uncore, GEN11_CRYPTO_RSVD_INTR_ENABLE, mask);
+	intel_uncore_write(uncore, GEN11_CRYPTO_RSVD_INTR_MASK,  ~mask);
+}
+
+static inline void pxp_irq_reset(struct intel_gt *gt)
+{
+	spin_lock_irq(gt->irq_lock);
+	gen11_gt_reset_one_iir(gt, 0, GEN11_KCR);
+	spin_unlock_irq(gt->irq_lock);
+}
+
+void intel_pxp_irq_enable(struct intel_pxp *pxp)
+{
+	struct intel_gt *gt = pxp_to_gt(pxp);
+
+	spin_lock_irq(gt->irq_lock);
+
+	if (!pxp->irq_enabled)
+		WARN_ON_ONCE(gen11_gt_reset_one_iir(gt, 0, GEN11_KCR));
+
+	__pxp_set_interrupts(gt, GEN12_PXP_INTERRUPTS);
+	pxp->irq_enabled = true;
+
+	spin_unlock_irq(gt->irq_lock);
+}
+
+void intel_pxp_irq_disable(struct intel_pxp *pxp)
+{
+	struct intel_gt *gt = pxp_to_gt(pxp);
+
+	/*
+	 * We always need to submit a global termination when we re-enable the
+	 * interrupts, so there is no need to make sure that the session state
+	 * makes sense at the end of this function. Just make sure this is not
+	 * called in a path were the driver consider the session as valid and
+	 * doesn't call a termination on restart.
+	 */
+	GEM_WARN_ON(intel_pxp_is_active(pxp));
+
+	spin_lock_irq(gt->irq_lock);
+
+	pxp->irq_enabled = false;
+	__pxp_set_interrupts(gt, 0);
+
+	spin_unlock_irq(gt->irq_lock);
+	intel_synchronize_irq(gt->i915);
+
+	pxp_irq_reset(gt);
+
+	flush_work(&pxp->session_work);
+}
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_irq.h b/drivers/gpu/drm/i915/pxp/intel_pxp_irq.h
new file mode 100644
index 000000000..8b5793654
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_irq.h
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright(c) 2020, Intel Corporation. All rights reserved.
+ */
+
+#ifndef __INTEL_PXP_IRQ_H__
+#define __INTEL_PXP_IRQ_H__
+
+#include <linux/types.h>
+
+struct intel_pxp;
+
+#define GEN12_DISPLAY_PXP_STATE_TERMINATED_INTERRUPT BIT(1)
+#define GEN12_DISPLAY_APP_TERMINATED_PER_FW_REQ_INTERRUPT BIT(2)
+#define GEN12_DISPLAY_STATE_RESET_COMPLETE_INTERRUPT BIT(3)
+
+#define GEN12_PXP_INTERRUPTS \
+	(GEN12_DISPLAY_PXP_STATE_TERMINATED_INTERRUPT | \
+	 GEN12_DISPLAY_APP_TERMINATED_PER_FW_REQ_INTERRUPT | \
+	 GEN12_DISPLAY_STATE_RESET_COMPLETE_INTERRUPT)
+
+#ifdef CONFIG_DRM_I915_PXP
+void intel_pxp_irq_enable(struct intel_pxp *pxp);
+void intel_pxp_irq_disable(struct intel_pxp *pxp);
+void intel_pxp_irq_handler(struct intel_pxp *pxp, u16 iir);
+#else
+static inline void intel_pxp_irq_handler(struct intel_pxp *pxp, u16 iir)
+{
+}
+#endif
+
+#endif /* __INTEL_PXP_IRQ_H__ */
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_pm.c b/drivers/gpu/drm/i915/pxp/intel_pxp_pm.c
new file mode 100644
index 000000000..6a7d4e2ee
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_pm.c
@@ -0,0 +1,61 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright(c) 2020 Intel Corporation.
+ */
+
+#include "intel_pxp.h"
+#include "intel_pxp_irq.h"
+#include "intel_pxp_pm.h"
+#include "intel_pxp_session.h"
+#include "i915_drv.h"
+
+void intel_pxp_suspend_prepare(struct intel_pxp *pxp)
+{
+	if (!intel_pxp_is_enabled(pxp))
+		return;
+
+	pxp->arb_is_valid = false;
+
+	intel_pxp_invalidate(pxp);
+}
+
+void intel_pxp_suspend(struct intel_pxp *pxp)
+{
+	intel_wakeref_t wakeref;
+
+	if (!intel_pxp_is_enabled(pxp))
+		return;
+
+	with_intel_runtime_pm(&pxp_to_gt(pxp)->i915->runtime_pm, wakeref) {
+		intel_pxp_fini_hw(pxp);
+		pxp->hw_state_invalidated = false;
+	}
+}
+
+void intel_pxp_resume(struct intel_pxp *pxp)
+{
+	if (!intel_pxp_is_enabled(pxp))
+		return;
+
+	/*
+	 * The PXP component gets automatically unbound when we go into S3 and
+	 * re-bound after we come out, so in that scenario we can defer the
+	 * hw init to the bind call.
+	 */
+	if (!pxp->pxp_component)
+		return;
+
+	intel_pxp_init_hw(pxp);
+}
+
+void intel_pxp_runtime_suspend(struct intel_pxp *pxp)
+{
+	if (!intel_pxp_is_enabled(pxp))
+		return;
+
+	pxp->arb_is_valid = false;
+
+	intel_pxp_fini_hw(pxp);
+
+	pxp->hw_state_invalidated = false;
+}
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_pm.h b/drivers/gpu/drm/i915/pxp/intel_pxp_pm.h
new file mode 100644
index 000000000..586be7691
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_pm.h
@@ -0,0 +1,37 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright(c) 2020, Intel Corporation. All rights reserved.
+ */
+
+#ifndef __INTEL_PXP_PM_H__
+#define __INTEL_PXP_PM_H__
+
+struct intel_pxp;
+
+#ifdef CONFIG_DRM_I915_PXP
+void intel_pxp_suspend_prepare(struct intel_pxp *pxp);
+void intel_pxp_suspend(struct intel_pxp *pxp);
+void intel_pxp_resume(struct intel_pxp *pxp);
+void intel_pxp_runtime_suspend(struct intel_pxp *pxp);
+#else
+static inline void intel_pxp_suspend_prepare(struct intel_pxp *pxp)
+{
+}
+
+static inline void intel_pxp_suspend(struct intel_pxp *pxp)
+{
+}
+
+static inline void intel_pxp_resume(struct intel_pxp *pxp)
+{
+}
+
+static inline void intel_pxp_runtime_suspend(struct intel_pxp *pxp)
+{
+}
+#endif
+static inline void intel_pxp_runtime_resume(struct intel_pxp *pxp)
+{
+	intel_pxp_resume(pxp);
+}
+#endif /* __INTEL_PXP_PM_H__ */
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_session.c b/drivers/gpu/drm/i915/pxp/intel_pxp_session.c
new file mode 100644
index 000000000..1bb5b5249
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_session.c
@@ -0,0 +1,174 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright(c) 2020, Intel Corporation. All rights reserved.
+ */
+
+#include "i915_drv.h"
+
+#include "intel_pxp.h"
+#include "intel_pxp_cmd.h"
+#include "intel_pxp_session.h"
+#include "intel_pxp_tee.h"
+#include "intel_pxp_types.h"
+
+#define ARB_SESSION I915_PROTECTED_CONTENT_DEFAULT_SESSION /* shorter define */
+
+#define GEN12_KCR_SIP _MMIO(0x32260) /* KCR hwdrm session in play 0-31 */
+
+/* PXP global terminate register for session termination */
+#define PXP_GLOBAL_TERMINATE _MMIO(0x320f8)
+
+static bool intel_pxp_session_is_in_play(struct intel_pxp *pxp, u32 id)
+{
+	struct intel_uncore *uncore = pxp_to_gt(pxp)->uncore;
+	intel_wakeref_t wakeref;
+	u32 sip = 0;
+
+	/* if we're suspended the session is considered off */
+	with_intel_runtime_pm_if_in_use(uncore->rpm, wakeref)
+		sip = intel_uncore_read(uncore, GEN12_KCR_SIP);
+
+	return sip & BIT(id);
+}
+
+static int pxp_wait_for_session_state(struct intel_pxp *pxp, u32 id, bool in_play)
+{
+	struct intel_uncore *uncore = pxp_to_gt(pxp)->uncore;
+	intel_wakeref_t wakeref;
+	u32 mask = BIT(id);
+	int ret;
+
+	/* if we're suspended the session is considered off */
+	wakeref = intel_runtime_pm_get_if_in_use(uncore->rpm);
+	if (!wakeref)
+		return in_play ? -ENODEV : 0;
+
+	ret = intel_wait_for_register(uncore,
+				      GEN12_KCR_SIP,
+				      mask,
+				      in_play ? mask : 0,
+				      100);
+
+	intel_runtime_pm_put(uncore->rpm, wakeref);
+
+	return ret;
+}
+
+static int pxp_create_arb_session(struct intel_pxp *pxp)
+{
+	struct intel_gt *gt = pxp_to_gt(pxp);
+	int ret;
+
+	pxp->arb_is_valid = false;
+
+	if (intel_pxp_session_is_in_play(pxp, ARB_SESSION)) {
+		drm_err(&gt->i915->drm, "arb session already in play at creation time\n");
+		return -EEXIST;
+	}
+
+	ret = intel_pxp_tee_cmd_create_arb_session(pxp, ARB_SESSION);
+	if (ret) {
+		drm_err(&gt->i915->drm, "tee cmd for arb session creation failed\n");
+		return ret;
+	}
+
+	ret = pxp_wait_for_session_state(pxp, ARB_SESSION, true);
+	if (ret) {
+		drm_err(&gt->i915->drm, "arb session failed to go in play\n");
+		return ret;
+	}
+
+	if (!++pxp->key_instance)
+		++pxp->key_instance;
+
+	pxp->arb_is_valid = true;
+
+	return 0;
+}
+
+static int pxp_terminate_arb_session_and_global(struct intel_pxp *pxp)
+{
+	int ret;
+	struct intel_gt *gt = pxp_to_gt(pxp);
+
+	/* must mark termination in progress calling this function */
+	GEM_WARN_ON(pxp->arb_is_valid);
+
+	/* terminate the hw sessions */
+	ret = intel_pxp_terminate_session(pxp, ARB_SESSION);
+	if (ret) {
+		drm_err(&gt->i915->drm, "Failed to submit session termination\n");
+		return ret;
+	}
+
+	ret = pxp_wait_for_session_state(pxp, ARB_SESSION, false);
+	if (ret) {
+		drm_err(&gt->i915->drm, "Session state did not clear\n");
+		return ret;
+	}
+
+	intel_uncore_write(gt->uncore, PXP_GLOBAL_TERMINATE, 1);
+
+	return ret;
+}
+
+static void pxp_terminate(struct intel_pxp *pxp)
+{
+	int ret;
+
+	pxp->hw_state_invalidated = true;
+
+	/*
+	 * if we fail to submit the termination there is no point in waiting for
+	 * it to complete. PXP will be marked as non-active until the next
+	 * termination is issued.
+	 */
+	ret = pxp_terminate_arb_session_and_global(pxp);
+	if (ret)
+		complete_all(&pxp->termination);
+}
+
+static void pxp_terminate_complete(struct intel_pxp *pxp)
+{
+	/* Re-create the arb session after teardown handle complete */
+	if (fetch_and_zero(&pxp->hw_state_invalidated))
+		pxp_create_arb_session(pxp);
+
+	complete_all(&pxp->termination);
+}
+
+void intel_pxp_session_work(struct work_struct *work)
+{
+	struct intel_pxp *pxp = container_of(work, typeof(*pxp), session_work);
+	struct intel_gt *gt = pxp_to_gt(pxp);
+	intel_wakeref_t wakeref;
+	u32 events = 0;
+
+	spin_lock_irq(gt->irq_lock);
+	events = fetch_and_zero(&pxp->session_events);
+	spin_unlock_irq(gt->irq_lock);
+
+	if (!events)
+		return;
+
+	if (events & PXP_INVAL_REQUIRED)
+		intel_pxp_invalidate(pxp);
+
+	/*
+	 * If we're processing an event while suspending then don't bother,
+	 * we're going to re-init everything on resume anyway.
+	 */
+	wakeref = intel_runtime_pm_get_if_in_use(gt->uncore->rpm);
+	if (!wakeref)
+		return;
+
+	if (events & PXP_TERMINATION_REQUEST) {
+		events &= ~PXP_TERMINATION_COMPLETE;
+		pxp_terminate(pxp);
+	}
+
+	if (events & PXP_TERMINATION_COMPLETE)
+		pxp_terminate_complete(pxp);
+
+	intel_runtime_pm_put(gt->uncore->rpm, wakeref);
+}
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_session.h b/drivers/gpu/drm/i915/pxp/intel_pxp_session.h
new file mode 100644
index 000000000..ba4c9d2b9
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_session.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright(c) 2020, Intel Corporation. All rights reserved.
+ */
+
+#ifndef __INTEL_PXP_SESSION_H__
+#define __INTEL_PXP_SESSION_H__
+
+#include <linux/types.h>
+
+struct work_struct;
+
+void intel_pxp_session_work(struct work_struct *work);
+
+#endif /* __INTEL_PXP_SESSION_H__ */
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_tee.c b/drivers/gpu/drm/i915/pxp/intel_pxp_tee.c
new file mode 100644
index 000000000..4b6f5655f
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_tee.c
@@ -0,0 +1,179 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright(c) 2020 Intel Corporation.
+ */
+
+#include <linux/component.h>
+
+#include <drm/i915_pxp_tee_interface.h>
+#include <drm/i915_component.h>
+
+#include "i915_drv.h"
+#include "intel_pxp.h"
+#include "intel_pxp_session.h"
+#include "intel_pxp_tee.h"
+#include "intel_pxp_tee_interface.h"
+
+static inline struct intel_pxp *i915_dev_to_pxp(struct device *i915_kdev)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(i915_kdev);
+
+	return &to_gt(i915)->pxp;
+}
+
+static int intel_pxp_tee_io_message(struct intel_pxp *pxp,
+				    void *msg_in, u32 msg_in_size,
+				    void *msg_out, u32 msg_out_max_size,
+				    u32 *msg_out_rcv_size)
+{
+	struct drm_i915_private *i915 = pxp_to_gt(pxp)->i915;
+	struct i915_pxp_component *pxp_component = pxp->pxp_component;
+	int ret = 0;
+
+	mutex_lock(&pxp->tee_mutex);
+
+	/*
+	 * The binding of the component is asynchronous from i915 probe, so we
+	 * can't be sure it has happened.
+	 */
+	if (!pxp_component) {
+		ret = -ENODEV;
+		goto unlock;
+	}
+
+	ret = pxp_component->ops->send(pxp_component->tee_dev, msg_in, msg_in_size);
+	if (ret) {
+		drm_err(&i915->drm, "Failed to send PXP TEE message\n");
+		goto unlock;
+	}
+
+	ret = pxp_component->ops->recv(pxp_component->tee_dev, msg_out, msg_out_max_size);
+	if (ret < 0) {
+		drm_err(&i915->drm, "Failed to receive PXP TEE message\n");
+		goto unlock;
+	}
+
+	if (ret > msg_out_max_size) {
+		drm_err(&i915->drm,
+			"Failed to receive PXP TEE message due to unexpected output size\n");
+		ret = -ENOSPC;
+		goto unlock;
+	}
+
+	if (msg_out_rcv_size)
+		*msg_out_rcv_size = ret;
+
+	ret = 0;
+unlock:
+	mutex_unlock(&pxp->tee_mutex);
+	return ret;
+}
+
+/**
+ * i915_pxp_tee_component_bind - bind function to pass the function pointers to pxp_tee
+ * @i915_kdev: pointer to i915 kernel device
+ * @tee_kdev: pointer to tee kernel device
+ * @data: pointer to pxp_tee_master containing the function pointers
+ *
+ * This bind function is called during the system boot or resume from system sleep.
+ *
+ * Return: return 0 if successful.
+ */
+static int i915_pxp_tee_component_bind(struct device *i915_kdev,
+				       struct device *tee_kdev, void *data)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(i915_kdev);
+	struct intel_pxp *pxp = i915_dev_to_pxp(i915_kdev);
+	intel_wakeref_t wakeref;
+
+	mutex_lock(&pxp->tee_mutex);
+	pxp->pxp_component = data;
+	pxp->pxp_component->tee_dev = tee_kdev;
+	mutex_unlock(&pxp->tee_mutex);
+
+	/* if we are suspended, the HW will be re-initialized on resume */
+	wakeref = intel_runtime_pm_get_if_in_use(&i915->runtime_pm);
+	if (!wakeref)
+		return 0;
+
+	/* the component is required to fully start the PXP HW */
+	intel_pxp_init_hw(pxp);
+
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+
+	return 0;
+}
+
+static void i915_pxp_tee_component_unbind(struct device *i915_kdev,
+					  struct device *tee_kdev, void *data)
+{
+	struct drm_i915_private *i915 = kdev_to_i915(i915_kdev);
+	struct intel_pxp *pxp = i915_dev_to_pxp(i915_kdev);
+	intel_wakeref_t wakeref;
+
+	with_intel_runtime_pm_if_in_use(&i915->runtime_pm, wakeref)
+		intel_pxp_fini_hw(pxp);
+
+	mutex_lock(&pxp->tee_mutex);
+	pxp->pxp_component = NULL;
+	mutex_unlock(&pxp->tee_mutex);
+}
+
+static const struct component_ops i915_pxp_tee_component_ops = {
+	.bind   = i915_pxp_tee_component_bind,
+	.unbind = i915_pxp_tee_component_unbind,
+};
+
+int intel_pxp_tee_component_init(struct intel_pxp *pxp)
+{
+	int ret;
+	struct intel_gt *gt = pxp_to_gt(pxp);
+	struct drm_i915_private *i915 = gt->i915;
+
+	ret = component_add_typed(i915->drm.dev, &i915_pxp_tee_component_ops,
+				  I915_COMPONENT_PXP);
+	if (ret < 0) {
+		drm_err(&i915->drm, "Failed to add PXP component (%d)\n", ret);
+		return ret;
+	}
+
+	pxp->pxp_component_added = true;
+
+	return 0;
+}
+
+void intel_pxp_tee_component_fini(struct intel_pxp *pxp)
+{
+	struct drm_i915_private *i915 = pxp_to_gt(pxp)->i915;
+
+	if (!pxp->pxp_component_added)
+		return;
+
+	component_del(i915->drm.dev, &i915_pxp_tee_component_ops);
+	pxp->pxp_component_added = false;
+}
+
+int intel_pxp_tee_cmd_create_arb_session(struct intel_pxp *pxp,
+					 int arb_session_id)
+{
+	struct drm_i915_private *i915 = pxp_to_gt(pxp)->i915;
+	struct pxp_tee_create_arb_in msg_in = {0};
+	struct pxp_tee_create_arb_out msg_out = {0};
+	int ret;
+
+	msg_in.header.api_version = PXP_TEE_APIVER;
+	msg_in.header.command_id = PXP_TEE_ARB_CMDID;
+	msg_in.header.buffer_len = sizeof(msg_in) - sizeof(msg_in.header);
+	msg_in.protection_mode = PXP_TEE_ARB_PROTECTION_MODE;
+	msg_in.session_id = arb_session_id;
+
+	ret = intel_pxp_tee_io_message(pxp,
+				       &msg_in, sizeof(msg_in),
+				       &msg_out, sizeof(msg_out),
+				       NULL);
+
+	if (ret)
+		drm_err(&i915->drm, "Failed to send tee msg ret=[%d]\n", ret);
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_tee.h b/drivers/gpu/drm/i915/pxp/intel_pxp_tee.h
new file mode 100644
index 000000000..c136053ce
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_tee.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright(c) 2020, Intel Corporation. All rights reserved.
+ */
+
+#ifndef __INTEL_PXP_TEE_H__
+#define __INTEL_PXP_TEE_H__
+
+#include "intel_pxp.h"
+
+int intel_pxp_tee_component_init(struct intel_pxp *pxp);
+void intel_pxp_tee_component_fini(struct intel_pxp *pxp);
+
+int intel_pxp_tee_cmd_create_arb_session(struct intel_pxp *pxp,
+					 int arb_session_id);
+
+#endif /* __INTEL_PXP_TEE_H__ */
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_tee_interface.h b/drivers/gpu/drm/i915/pxp/intel_pxp_tee_interface.h
new file mode 100644
index 000000000..36e9b0868
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_tee_interface.h
@@ -0,0 +1,36 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright(c) 2020, Intel Corporation. All rights reserved.
+ */
+
+#ifndef __INTEL_PXP_TEE_INTERFACE_H__
+#define __INTEL_PXP_TEE_INTERFACE_H__
+
+#include <linux/types.h>
+
+#define PXP_TEE_APIVER 0x40002
+#define PXP_TEE_ARB_CMDID 0x1e
+#define PXP_TEE_ARB_PROTECTION_MODE 0x2
+
+/* PXP TEE message header */
+struct pxp_tee_cmd_header {
+	u32 api_version;
+	u32 command_id;
+	u32 status;
+	/* Length of the message (excluding the header) */
+	u32 buffer_len;
+} __packed;
+
+/* PXP TEE message input to create a arbitrary session */
+struct pxp_tee_create_arb_in {
+	struct pxp_tee_cmd_header header;
+	u32 protection_mode;
+	u32 session_id;
+} __packed;
+
+/* PXP TEE message output to create a arbitrary session */
+struct pxp_tee_create_arb_out {
+	struct pxp_tee_cmd_header header;
+} __packed;
+
+#endif /* __INTEL_PXP_TEE_INTERFACE_H__ */
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_types.h b/drivers/gpu/drm/i915/pxp/intel_pxp_types.h
new file mode 100644
index 000000000..7ce5f37ee
--- /dev/null
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_types.h
@@ -0,0 +1,81 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright(c) 2020, Intel Corporation. All rights reserved.
+ */
+
+#ifndef __INTEL_PXP_TYPES_H__
+#define __INTEL_PXP_TYPES_H__
+
+#include <linux/completion.h>
+#include <linux/mutex.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+
+struct intel_context;
+struct i915_pxp_component;
+
+/**
+ * struct intel_pxp - pxp state
+ */
+struct intel_pxp {
+	/**
+	 * @pxp_component: i915_pxp_component struct of the bound mei_pxp
+	 * module. Only set and cleared inside component bind/unbind functions,
+	 * which are protected by &tee_mutex.
+	 */
+	struct i915_pxp_component *pxp_component;
+	/**
+	 * @pxp_component_added: track if the pxp component has been added.
+	 * Set and cleared in tee init and fini functions respectively.
+	 */
+	bool pxp_component_added;
+
+	/** @ce: kernel-owned context used for PXP operations */
+	struct intel_context *ce;
+
+	/** @arb_mutex: protects arb session start */
+	struct mutex arb_mutex;
+	/**
+	 * @arb_is_valid: tracks arb session status.
+	 * After a teardown, the arb session can still be in play on the HW
+	 * even if the keys are gone, so we can't rely on the HW state of the
+	 * session to know if it's valid and need to track the status in SW.
+	 */
+	bool arb_is_valid;
+
+	/**
+	 * @key_instance: tracks which key instance we're on, so we can use it
+	 * to determine if an object was created using the current key or a
+	 * previous one.
+	 */
+	u32 key_instance;
+
+	/** @tee_mutex: protects the tee channel binding and messaging. */
+	struct mutex tee_mutex;
+
+	/**
+	 * @hw_state_invalidated: if the HW perceives an attack on the integrity
+	 * of the encryption it will invalidate the keys and expect SW to
+	 * re-initialize the session. We keep track of this state to make sure
+	 * we only re-start the arb session when required.
+	 */
+	bool hw_state_invalidated;
+
+	/** @irq_enabled: tracks the status of the kcr irqs */
+	bool irq_enabled;
+	/**
+	 * @termination: tracks the status of a pending termination. Only
+	 * re-initialized under gt->irq_lock and completed in &session_work.
+	 */
+	struct completion termination;
+
+	/** @session_work: worker that manages session events. */
+	struct work_struct session_work;
+	/** @session_events: pending session events, protected with gt->irq_lock. */
+	u32 session_events;
+#define PXP_TERMINATION_REQUEST  BIT(0)
+#define PXP_TERMINATION_COMPLETE BIT(1)
+#define PXP_INVAL_REQUIRED       BIT(2)
+};
+
+#endif /* __INTEL_PXP_TYPES_H__ */
diff --git a/drivers/gpu/drm/i915/selftests/i915_active.c b/drivers/gpu/drm/i915/selftests/i915_active.c
new file mode 100644
index 000000000..b61fe850e
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_active.c
@@ -0,0 +1,352 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/kref.h>
+#include <linux/string_helpers.h>
+
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_gt.h"
+
+#include "i915_selftest.h"
+
+#include "igt_flush_test.h"
+#include "lib_sw_fence.h"
+
+struct live_active {
+	struct i915_active base;
+	struct kref ref;
+	bool retired;
+};
+
+static void __live_get(struct live_active *active)
+{
+	kref_get(&active->ref);
+}
+
+static void __live_free(struct live_active *active)
+{
+	i915_active_fini(&active->base);
+	kfree(active);
+}
+
+static void __live_release(struct kref *ref)
+{
+	struct live_active *active = container_of(ref, typeof(*active), ref);
+
+	__live_free(active);
+}
+
+static void __live_put(struct live_active *active)
+{
+	kref_put(&active->ref, __live_release);
+}
+
+static int __live_active(struct i915_active *base)
+{
+	struct live_active *active = container_of(base, typeof(*active), base);
+
+	__live_get(active);
+	return 0;
+}
+
+static void __live_retire(struct i915_active *base)
+{
+	struct live_active *active = container_of(base, typeof(*active), base);
+
+	active->retired = true;
+	__live_put(active);
+}
+
+static struct live_active *__live_alloc(struct drm_i915_private *i915)
+{
+	struct live_active *active;
+
+	active = kzalloc(sizeof(*active), GFP_KERNEL);
+	if (!active)
+		return NULL;
+
+	kref_init(&active->ref);
+	i915_active_init(&active->base, __live_active, __live_retire, 0);
+
+	return active;
+}
+
+static struct live_active *
+__live_active_setup(struct drm_i915_private *i915)
+{
+	struct intel_engine_cs *engine;
+	struct i915_sw_fence *submit;
+	struct live_active *active;
+	unsigned int count = 0;
+	int err = 0;
+
+	active = __live_alloc(i915);
+	if (!active)
+		return ERR_PTR(-ENOMEM);
+
+	submit = heap_fence_create(GFP_KERNEL);
+	if (!submit) {
+		kfree(active);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	err = i915_active_acquire(&active->base);
+	if (err)
+		goto out;
+
+	for_each_uabi_engine(engine, i915) {
+		struct i915_request *rq;
+
+		rq = intel_engine_create_kernel_request(engine);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		err = i915_sw_fence_await_sw_fence_gfp(&rq->submit,
+						       submit,
+						       GFP_KERNEL);
+		if (err >= 0)
+			err = i915_active_add_request(&active->base, rq);
+		i915_request_add(rq);
+		if (err) {
+			pr_err("Failed to track active ref!\n");
+			break;
+		}
+
+		count++;
+	}
+
+	i915_active_release(&active->base);
+	if (READ_ONCE(active->retired) && count) {
+		pr_err("i915_active retired before submission!\n");
+		err = -EINVAL;
+	}
+	if (atomic_read(&active->base.count) != count) {
+		pr_err("i915_active not tracking all requests, found %d, expected %d\n",
+		       atomic_read(&active->base.count), count);
+		err = -EINVAL;
+	}
+
+out:
+	i915_sw_fence_commit(submit);
+	heap_fence_put(submit);
+	if (err) {
+		__live_put(active);
+		active = ERR_PTR(err);
+	}
+
+	return active;
+}
+
+static int live_active_wait(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct live_active *active;
+	int err = 0;
+
+	/* Check that we get a callback when requests retire upon waiting */
+
+	active = __live_active_setup(i915);
+	if (IS_ERR(active))
+		return PTR_ERR(active);
+
+	__i915_active_wait(&active->base, TASK_UNINTERRUPTIBLE);
+	if (!READ_ONCE(active->retired)) {
+		struct drm_printer p = drm_err_printer(__func__);
+
+		pr_err("i915_active not retired after waiting!\n");
+		i915_active_print(&active->base, &p);
+
+		err = -EINVAL;
+	}
+
+	__live_put(active);
+
+	if (igt_flush_test(i915))
+		err = -EIO;
+
+	return err;
+}
+
+static int live_active_retire(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct live_active *active;
+	int err = 0;
+
+	/* Check that we get a callback when requests are indirectly retired */
+
+	active = __live_active_setup(i915);
+	if (IS_ERR(active))
+		return PTR_ERR(active);
+
+	/* waits for & retires all requests */
+	if (igt_flush_test(i915))
+		err = -EIO;
+
+	if (!READ_ONCE(active->retired)) {
+		struct drm_printer p = drm_err_printer(__func__);
+
+		pr_err("i915_active not retired after flushing!\n");
+		i915_active_print(&active->base, &p);
+
+		err = -EINVAL;
+	}
+
+	__live_put(active);
+
+	return err;
+}
+
+static int live_active_barrier(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	struct live_active *active;
+	int err = 0;
+
+	/* Check that we get a callback when requests retire upon waiting */
+
+	active = __live_alloc(i915);
+	if (!active)
+		return -ENOMEM;
+
+	err = i915_active_acquire(&active->base);
+	if (err)
+		goto out;
+
+	for_each_uabi_engine(engine, i915) {
+		err = i915_active_acquire_preallocate_barrier(&active->base,
+							      engine);
+		if (err)
+			break;
+
+		i915_active_acquire_barrier(&active->base);
+	}
+
+	i915_active_release(&active->base);
+	if (err)
+		goto out;
+
+	__i915_active_wait(&active->base, TASK_UNINTERRUPTIBLE);
+	if (!READ_ONCE(active->retired)) {
+		pr_err("i915_active not retired after flushing barriers!\n");
+		err = -EINVAL;
+	}
+
+out:
+	__live_put(active);
+
+	if (igt_flush_test(i915))
+		err = -EIO;
+
+	return err;
+}
+
+int i915_active_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_active_wait),
+		SUBTEST(live_active_retire),
+		SUBTEST(live_active_barrier),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return i915_subtests(tests, i915);
+}
+
+static struct intel_engine_cs *node_to_barrier(struct active_node *it)
+{
+	struct intel_engine_cs *engine;
+
+	if (!is_barrier(&it->base))
+		return NULL;
+
+	engine = __barrier_to_engine(it);
+	smp_rmb(); /* serialise with add_active_barriers */
+	if (!is_barrier(&it->base))
+		return NULL;
+
+	return engine;
+}
+
+void i915_active_print(struct i915_active *ref, struct drm_printer *m)
+{
+	drm_printf(m, "active %ps:%ps\n", ref->active, ref->retire);
+	drm_printf(m, "\tcount: %d\n", atomic_read(&ref->count));
+	drm_printf(m, "\tpreallocated barriers? %s\n",
+		   str_yes_no(!llist_empty(&ref->preallocated_barriers)));
+
+	if (i915_active_acquire_if_busy(ref)) {
+		struct active_node *it, *n;
+
+		rbtree_postorder_for_each_entry_safe(it, n, &ref->tree, node) {
+			struct intel_engine_cs *engine;
+
+			engine = node_to_barrier(it);
+			if (engine) {
+				drm_printf(m, "\tbarrier: %s\n", engine->name);
+				continue;
+			}
+
+			if (i915_active_fence_isset(&it->base)) {
+				drm_printf(m,
+					   "\ttimeline: %llx\n", it->timeline);
+				continue;
+			}
+		}
+
+		i915_active_release(ref);
+	}
+}
+
+static void spin_unlock_wait(spinlock_t *lock)
+{
+	spin_lock_irq(lock);
+	spin_unlock_irq(lock);
+}
+
+static void active_flush(struct i915_active *ref,
+			 struct i915_active_fence *active)
+{
+	struct dma_fence *fence;
+
+	fence = xchg(__active_fence_slot(active), NULL);
+	if (!fence)
+		return;
+
+	spin_lock_irq(fence->lock);
+	__list_del_entry(&active->cb.node);
+	spin_unlock_irq(fence->lock); /* serialise with fence->cb_list */
+	atomic_dec(&ref->count);
+
+	GEM_BUG_ON(!test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags));
+}
+
+void i915_active_unlock_wait(struct i915_active *ref)
+{
+	if (i915_active_acquire_if_busy(ref)) {
+		struct active_node *it, *n;
+
+		/* Wait for all active callbacks */
+		rcu_read_lock();
+		active_flush(ref, &ref->excl);
+		rbtree_postorder_for_each_entry_safe(it, n, &ref->tree, node)
+			active_flush(ref, &it->base);
+		rcu_read_unlock();
+
+		i915_active_release(ref);
+	}
+
+	/* And wait for the retire callback */
+	spin_unlock_wait(&ref->tree_lock);
+
+	/* ... which may have been on a thread instead */
+	flush_work(&ref->work);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem.c b/drivers/gpu/drm/i915/selftests/i915_gem.c
new file mode 100644
index 000000000..e5dd82e7e
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_gem.c
@@ -0,0 +1,256 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/random.h>
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_pm.h"
+#include "gem/selftests/igt_gem_utils.h"
+#include "gem/selftests/mock_context.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_pm.h"
+
+#include "i915_selftest.h"
+
+#include "igt_flush_test.h"
+#include "mock_drm.h"
+
+static int switch_to_context(struct i915_gem_context *ctx)
+{
+	struct i915_gem_engines_iter it;
+	struct intel_context *ce;
+	int err = 0;
+
+	for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it) {
+		struct i915_request *rq;
+
+		rq = intel_context_create_request(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_add(rq);
+	}
+	i915_gem_context_unlock_engines(ctx);
+
+	return err;
+}
+
+static void trash_stolen(struct drm_i915_private *i915)
+{
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	const u64 slot = ggtt->error_capture.start;
+	const resource_size_t size = resource_size(&i915->dsm);
+	unsigned long page;
+	u32 prng = 0x12345678;
+
+	/* XXX: fsck. needs some more thought... */
+	if (!i915_ggtt_has_aperture(ggtt))
+		return;
+
+	for (page = 0; page < size; page += PAGE_SIZE) {
+		const dma_addr_t dma = i915->dsm.start + page;
+		u32 __iomem *s;
+		int x;
+
+		ggtt->vm.insert_page(&ggtt->vm, dma, slot, I915_CACHE_NONE, 0);
+
+		s = io_mapping_map_atomic_wc(&ggtt->iomap, slot);
+		for (x = 0; x < PAGE_SIZE / sizeof(u32); x++) {
+			prng = next_pseudo_random32(prng);
+			iowrite32(prng, &s[x]);
+		}
+		io_mapping_unmap_atomic(s);
+	}
+
+	ggtt->vm.clear_range(&ggtt->vm, slot, PAGE_SIZE);
+}
+
+static void simulate_hibernate(struct drm_i915_private *i915)
+{
+	intel_wakeref_t wakeref;
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	/*
+	 * As a final sting in the tail, invalidate stolen. Under a real S4,
+	 * stolen is lost and needs to be refilled on resume. However, under
+	 * CI we merely do S4-device testing (as full S4 is too unreliable
+	 * for automated testing across a cluster), so to simulate the effect
+	 * of stolen being trashed across S4, we trash it ourselves.
+	 */
+	trash_stolen(i915);
+
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+}
+
+static int igt_pm_prepare(struct drm_i915_private *i915)
+{
+	i915_gem_suspend(i915);
+
+	return 0;
+}
+
+static void igt_pm_suspend(struct drm_i915_private *i915)
+{
+	intel_wakeref_t wakeref;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		i915_ggtt_suspend(to_gt(i915)->ggtt);
+		i915_gem_suspend_late(i915);
+	}
+}
+
+static void igt_pm_hibernate(struct drm_i915_private *i915)
+{
+	intel_wakeref_t wakeref;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		i915_ggtt_suspend(to_gt(i915)->ggtt);
+
+		i915_gem_freeze(i915);
+		i915_gem_freeze_late(i915);
+	}
+}
+
+static void igt_pm_resume(struct drm_i915_private *i915)
+{
+	intel_wakeref_t wakeref;
+
+	/*
+	 * Both suspend and hibernate follow the same wakeup path and assume
+	 * that runtime-pm just works.
+	 */
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref) {
+		i915_ggtt_resume(to_gt(i915)->ggtt);
+		i915_gem_resume(i915);
+	}
+}
+
+static int igt_gem_suspend(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_gem_context *ctx;
+	struct file *file;
+	int err;
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	err = -ENOMEM;
+	ctx = live_context(i915, file);
+	if (!IS_ERR(ctx))
+		err = switch_to_context(ctx);
+	if (err)
+		goto out;
+
+	err = igt_pm_prepare(i915);
+	if (err)
+		goto out;
+
+	igt_pm_suspend(i915);
+
+	/* Here be dragons! Note that with S3RST any S3 may become S4! */
+	simulate_hibernate(i915);
+
+	igt_pm_resume(i915);
+
+	err = switch_to_context(ctx);
+out:
+	fput(file);
+	return err;
+}
+
+static int igt_gem_hibernate(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_gem_context *ctx;
+	struct file *file;
+	int err;
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	err = -ENOMEM;
+	ctx = live_context(i915, file);
+	if (!IS_ERR(ctx))
+		err = switch_to_context(ctx);
+	if (err)
+		goto out;
+
+	err = igt_pm_prepare(i915);
+	if (err)
+		goto out;
+
+	igt_pm_hibernate(i915);
+
+	/* Here be dragons! */
+	simulate_hibernate(i915);
+
+	igt_pm_resume(i915);
+
+	err = switch_to_context(ctx);
+out:
+	fput(file);
+	return err;
+}
+
+static int igt_gem_ww_ctx(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj, *obj2;
+	struct i915_gem_ww_ctx ww;
+	int err = 0;
+
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	obj2 = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj2)) {
+		err = PTR_ERR(obj2);
+		goto put1;
+	}
+
+	i915_gem_ww_ctx_init(&ww, true);
+retry:
+	/* Lock the objects, twice for good measure (-EALREADY handling) */
+	err = i915_gem_object_lock(obj, &ww);
+	if (!err)
+		err = i915_gem_object_lock_interruptible(obj, &ww);
+	if (!err)
+		err = i915_gem_object_lock_interruptible(obj2, &ww);
+	if (!err)
+		err = i915_gem_object_lock(obj2, &ww);
+
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+	i915_gem_object_put(obj2);
+put1:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+int i915_gem_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_gem_suspend),
+		SUBTEST(igt_gem_hibernate),
+		SUBTEST(igt_gem_ww_ctx),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_evict.c b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c
new file mode 100644
index 000000000..37068542a
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c
@@ -0,0 +1,569 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_pm.h"
+#include "gem/selftests/igt_gem_utils.h"
+#include "gem/selftests/mock_context.h"
+#include "gt/intel_gt.h"
+
+#include "i915_selftest.h"
+
+#include "igt_flush_test.h"
+#include "lib_sw_fence.h"
+#include "mock_drm.h"
+#include "mock_gem_device.h"
+
+static void quirk_add(struct drm_i915_gem_object *obj,
+		      struct list_head *objects)
+{
+	/* quirk is only for live tiled objects, use it to declare ownership */
+	GEM_BUG_ON(i915_gem_object_has_tiling_quirk(obj));
+	i915_gem_object_set_tiling_quirk(obj);
+	list_add(&obj->st_link, objects);
+}
+
+static int populate_ggtt(struct i915_ggtt *ggtt, struct list_head *objects)
+{
+	struct drm_i915_gem_object *obj;
+	unsigned long count;
+
+	count = 0;
+	do {
+		struct i915_vma *vma;
+
+		obj = i915_gem_object_create_internal(ggtt->vm.i915,
+						      I915_GTT_PAGE_SIZE);
+		if (IS_ERR(obj))
+			return PTR_ERR(obj);
+
+		vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, 0);
+		if (IS_ERR(vma)) {
+			i915_gem_object_put(obj);
+			if (vma == ERR_PTR(-ENOSPC))
+				break;
+
+			return PTR_ERR(vma);
+		}
+
+		quirk_add(obj, objects);
+		count++;
+	} while (1);
+	pr_debug("Filled GGTT with %lu pages [%llu total]\n",
+		 count, ggtt->vm.total / PAGE_SIZE);
+
+	if (list_empty(&ggtt->vm.bound_list)) {
+		pr_err("No objects on the GGTT inactive list!\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void unpin_ggtt(struct i915_ggtt *ggtt)
+{
+	struct i915_vma *vma;
+
+	list_for_each_entry(vma, &ggtt->vm.bound_list, vm_link)
+		if (i915_gem_object_has_tiling_quirk(vma->obj))
+			i915_vma_unpin(vma);
+}
+
+static void cleanup_objects(struct i915_ggtt *ggtt, struct list_head *list)
+{
+	struct drm_i915_gem_object *obj, *on;
+
+	list_for_each_entry_safe(obj, on, list, st_link) {
+		GEM_BUG_ON(!i915_gem_object_has_tiling_quirk(obj));
+		i915_gem_object_set_tiling_quirk(obj);
+		i915_gem_object_put(obj);
+	}
+
+	i915_gem_drain_freed_objects(ggtt->vm.i915);
+}
+
+static int igt_evict_something(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_ggtt *ggtt = gt->ggtt;
+	LIST_HEAD(objects);
+	int err;
+
+	/* Fill the GGTT with pinned objects and try to evict one. */
+
+	err = populate_ggtt(ggtt, &objects);
+	if (err)
+		goto cleanup;
+
+	/* Everything is pinned, nothing should happen */
+	mutex_lock(&ggtt->vm.mutex);
+	err = i915_gem_evict_something(&ggtt->vm, NULL,
+				       I915_GTT_PAGE_SIZE, 0, 0,
+				       0, U64_MAX,
+				       0);
+	mutex_unlock(&ggtt->vm.mutex);
+	if (err != -ENOSPC) {
+		pr_err("i915_gem_evict_something failed on a full GGTT with err=%d\n",
+		       err);
+		goto cleanup;
+	}
+
+	unpin_ggtt(ggtt);
+
+	/* Everything is unpinned, we should be able to evict something */
+	mutex_lock(&ggtt->vm.mutex);
+	err = i915_gem_evict_something(&ggtt->vm, NULL,
+				       I915_GTT_PAGE_SIZE, 0, 0,
+				       0, U64_MAX,
+				       0);
+	mutex_unlock(&ggtt->vm.mutex);
+	if (err) {
+		pr_err("i915_gem_evict_something failed on a full GGTT with err=%d\n",
+		       err);
+		goto cleanup;
+	}
+
+cleanup:
+	cleanup_objects(ggtt, &objects);
+	return err;
+}
+
+static int igt_overcommit(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_ggtt *ggtt = gt->ggtt;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	LIST_HEAD(objects);
+	int err;
+
+	/* Fill the GGTT with pinned objects and then try to pin one more.
+	 * We expect it to fail.
+	 */
+
+	err = populate_ggtt(ggtt, &objects);
+	if (err)
+		goto cleanup;
+
+	obj = i915_gem_object_create_internal(gt->i915, I915_GTT_PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto cleanup;
+	}
+
+	quirk_add(obj, &objects);
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, 0);
+	if (vma != ERR_PTR(-ENOSPC)) {
+		pr_err("Failed to evict+insert, i915_gem_object_ggtt_pin returned err=%d\n", (int)PTR_ERR_OR_ZERO(vma));
+		err = -EINVAL;
+		goto cleanup;
+	}
+
+cleanup:
+	cleanup_objects(ggtt, &objects);
+	return err;
+}
+
+static int igt_evict_for_vma(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_ggtt *ggtt = gt->ggtt;
+	struct drm_mm_node target = {
+		.start = 0,
+		.size = 4096,
+	};
+	LIST_HEAD(objects);
+	int err;
+
+	/* Fill the GGTT with pinned objects and try to evict a range. */
+
+	err = populate_ggtt(ggtt, &objects);
+	if (err)
+		goto cleanup;
+
+	/* Everything is pinned, nothing should happen */
+	mutex_lock(&ggtt->vm.mutex);
+	err = i915_gem_evict_for_node(&ggtt->vm, NULL, &target, 0);
+	mutex_unlock(&ggtt->vm.mutex);
+	if (err != -ENOSPC) {
+		pr_err("i915_gem_evict_for_node on a full GGTT returned err=%d\n",
+		       err);
+		goto cleanup;
+	}
+
+	unpin_ggtt(ggtt);
+
+	/* Everything is unpinned, we should be able to evict the node */
+	mutex_lock(&ggtt->vm.mutex);
+	err = i915_gem_evict_for_node(&ggtt->vm, NULL, &target, 0);
+	mutex_unlock(&ggtt->vm.mutex);
+	if (err) {
+		pr_err("i915_gem_evict_for_node returned err=%d\n",
+		       err);
+		goto cleanup;
+	}
+
+cleanup:
+	cleanup_objects(ggtt, &objects);
+	return err;
+}
+
+static void mock_color_adjust(const struct drm_mm_node *node,
+			      unsigned long color,
+			      u64 *start,
+			      u64 *end)
+{
+}
+
+static int igt_evict_for_cache_color(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_ggtt *ggtt = gt->ggtt;
+	const unsigned long flags = PIN_OFFSET_FIXED;
+	struct drm_mm_node target = {
+		.start = I915_GTT_PAGE_SIZE * 2,
+		.size = I915_GTT_PAGE_SIZE,
+		.color = I915_CACHE_LLC,
+	};
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	LIST_HEAD(objects);
+	int err;
+
+	/*
+	 * Currently the use of color_adjust for the GGTT is limited to cache
+	 * coloring and guard pages, and so the presence of mm.color_adjust for
+	 * the GGTT is assumed to be i915_ggtt_color_adjust, hence using a mock
+	 * color adjust will work just fine for our purposes.
+	 */
+	ggtt->vm.mm.color_adjust = mock_color_adjust;
+	GEM_BUG_ON(!i915_vm_has_cache_coloring(&ggtt->vm));
+
+	obj = i915_gem_object_create_internal(gt->i915, I915_GTT_PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto cleanup;
+	}
+	i915_gem_object_set_cache_coherency(obj, I915_CACHE_LLC);
+	quirk_add(obj, &objects);
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0,
+				       I915_GTT_PAGE_SIZE | flags);
+	if (IS_ERR(vma)) {
+		pr_err("[0]i915_gem_object_ggtt_pin failed\n");
+		err = PTR_ERR(vma);
+		goto cleanup;
+	}
+
+	obj = i915_gem_object_create_internal(gt->i915, I915_GTT_PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto cleanup;
+	}
+	i915_gem_object_set_cache_coherency(obj, I915_CACHE_LLC);
+	quirk_add(obj, &objects);
+
+	/* Neighbouring; same colour - should fit */
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0,
+				       (I915_GTT_PAGE_SIZE * 2) | flags);
+	if (IS_ERR(vma)) {
+		pr_err("[1]i915_gem_object_ggtt_pin failed\n");
+		err = PTR_ERR(vma);
+		goto cleanup;
+	}
+
+	i915_vma_unpin(vma);
+
+	/* Remove just the second vma */
+	mutex_lock(&ggtt->vm.mutex);
+	err = i915_gem_evict_for_node(&ggtt->vm, NULL, &target, 0);
+	mutex_unlock(&ggtt->vm.mutex);
+	if (err) {
+		pr_err("[0]i915_gem_evict_for_node returned err=%d\n", err);
+		goto cleanup;
+	}
+
+	/* Attempt to remove the first *pinned* vma, by removing the (empty)
+	 * neighbour -- this should fail.
+	 */
+	target.color = I915_CACHE_L3_LLC;
+
+	mutex_lock(&ggtt->vm.mutex);
+	err = i915_gem_evict_for_node(&ggtt->vm, NULL, &target, 0);
+	mutex_unlock(&ggtt->vm.mutex);
+	if (!err) {
+		pr_err("[1]i915_gem_evict_for_node returned err=%d\n", err);
+		err = -EINVAL;
+		goto cleanup;
+	}
+
+	err = 0;
+
+cleanup:
+	unpin_ggtt(ggtt);
+	cleanup_objects(ggtt, &objects);
+	ggtt->vm.mm.color_adjust = NULL;
+	return err;
+}
+
+static int igt_evict_vm(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct i915_ggtt *ggtt = gt->ggtt;
+	struct i915_gem_ww_ctx ww;
+	LIST_HEAD(objects);
+	int err;
+
+	/* Fill the GGTT with pinned objects and try to evict everything. */
+
+	err = populate_ggtt(ggtt, &objects);
+	if (err)
+		goto cleanup;
+
+	/* Everything is pinned, nothing should happen */
+	mutex_lock(&ggtt->vm.mutex);
+	err = i915_gem_evict_vm(&ggtt->vm, NULL, NULL);
+	mutex_unlock(&ggtt->vm.mutex);
+	if (err) {
+		pr_err("i915_gem_evict_vm on a full GGTT returned err=%d]\n",
+		       err);
+		goto cleanup;
+	}
+
+	unpin_ggtt(ggtt);
+
+	for_i915_gem_ww(&ww, err, false) {
+		mutex_lock(&ggtt->vm.mutex);
+		err = i915_gem_evict_vm(&ggtt->vm, &ww, NULL);
+		mutex_unlock(&ggtt->vm.mutex);
+	}
+
+	if (err) {
+		pr_err("i915_gem_evict_vm on a full GGTT returned err=%d]\n",
+		       err);
+		goto cleanup;
+	}
+
+cleanup:
+	cleanup_objects(ggtt, &objects);
+	return err;
+}
+
+static int igt_evict_contexts(void *arg)
+{
+	const u64 PRETEND_GGTT_SIZE = 16ull << 20;
+	struct intel_gt *gt = arg;
+	struct i915_ggtt *ggtt = gt->ggtt;
+	struct drm_i915_private *i915 = gt->i915;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct reserved {
+		struct drm_mm_node node;
+		struct reserved *next;
+	} *reserved = NULL;
+	intel_wakeref_t wakeref;
+	struct drm_mm_node hole;
+	unsigned long count;
+	int err;
+
+	/*
+	 * The purpose of this test is to verify that we will trigger an
+	 * eviction in the GGTT when constructing a request that requires
+	 * additional space in the GGTT for pinning the context. This space
+	 * is not directly tied to the request so reclaiming it requires
+	 * extra work.
+	 *
+	 * As such this test is only meaningful for full-ppgtt environments
+	 * where the GTT space of the request is separate from the GGTT
+	 * allocation required to build the request.
+	 */
+	if (!HAS_FULL_PPGTT(i915))
+		return 0;
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	/* Reserve a block so that we know we have enough to fit a few rq */
+	memset(&hole, 0, sizeof(hole));
+	mutex_lock(&ggtt->vm.mutex);
+	err = i915_gem_gtt_insert(&ggtt->vm, NULL, &hole,
+				  PRETEND_GGTT_SIZE, 0, I915_COLOR_UNEVICTABLE,
+				  0, ggtt->vm.total,
+				  PIN_NOEVICT);
+	if (err)
+		goto out_locked;
+
+	/* Make the GGTT appear small by filling it with unevictable nodes */
+	count = 0;
+	do {
+		struct reserved *r;
+
+		mutex_unlock(&ggtt->vm.mutex);
+		r = kcalloc(1, sizeof(*r), GFP_KERNEL);
+		mutex_lock(&ggtt->vm.mutex);
+		if (!r) {
+			err = -ENOMEM;
+			goto out_locked;
+		}
+
+		if (i915_gem_gtt_insert(&ggtt->vm, NULL, &r->node,
+					1ul << 20, 0, I915_COLOR_UNEVICTABLE,
+					0, ggtt->vm.total,
+					PIN_NOEVICT)) {
+			kfree(r);
+			break;
+		}
+
+		r->next = reserved;
+		reserved = r;
+
+		count++;
+	} while (1);
+	drm_mm_remove_node(&hole);
+	mutex_unlock(&ggtt->vm.mutex);
+	pr_info("Filled GGTT with %lu 1MiB nodes\n", count);
+
+	/* Overfill the GGTT with context objects and so try to evict one. */
+	for_each_engine(engine, gt, id) {
+		struct i915_sw_fence fence;
+		struct i915_request *last = NULL;
+
+		count = 0;
+		onstack_fence_init(&fence);
+		do {
+			struct intel_context *ce;
+			struct i915_request *rq;
+
+			ce = intel_context_create(engine);
+			if (IS_ERR(ce))
+				break;
+
+			/* We will need some GGTT space for the rq's context */
+			igt_evict_ctl.fail_if_busy = true;
+			rq = intel_context_create_request(ce);
+			igt_evict_ctl.fail_if_busy = false;
+			intel_context_put(ce);
+
+			if (IS_ERR(rq)) {
+				/* When full, fail_if_busy will trigger EBUSY */
+				if (PTR_ERR(rq) != -EBUSY) {
+					pr_err("Unexpected error from request alloc (on %s): %d\n",
+					       engine->name,
+					       (int)PTR_ERR(rq));
+					err = PTR_ERR(rq);
+				}
+				break;
+			}
+
+			/* Keep every request/ctx pinned until we are full */
+			err = i915_sw_fence_await_sw_fence_gfp(&rq->submit,
+							       &fence,
+							       GFP_KERNEL);
+			if (err < 0)
+				break;
+
+			i915_request_add(rq);
+			count++;
+			if (last)
+				i915_request_put(last);
+			last = i915_request_get(rq);
+			err = 0;
+		} while(1);
+		onstack_fence_fini(&fence);
+		pr_info("Submitted %lu contexts/requests on %s\n",
+			count, engine->name);
+		if (err)
+			break;
+		if (last) {
+			if (i915_request_wait(last, 0, HZ) < 0) {
+				err = -EIO;
+				i915_request_put(last);
+				pr_err("Failed waiting for last request (on %s)",
+				       engine->name);
+				break;
+			}
+			i915_request_put(last);
+		}
+		err = intel_gt_wait_for_idle(engine->gt, HZ * 3);
+		if (err) {
+			pr_err("Failed to idle GT (on %s)", engine->name);
+			break;
+		}
+	}
+
+	mutex_lock(&ggtt->vm.mutex);
+out_locked:
+	if (igt_flush_test(i915))
+		err = -EIO;
+	while (reserved) {
+		struct reserved *next = reserved->next;
+
+		drm_mm_remove_node(&reserved->node);
+		kfree(reserved);
+
+		reserved = next;
+	}
+	if (drm_mm_node_allocated(&hole))
+		drm_mm_remove_node(&hole);
+	mutex_unlock(&ggtt->vm.mutex);
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+
+	return err;
+}
+
+int i915_gem_evict_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_evict_something),
+		SUBTEST(igt_evict_for_vma),
+		SUBTEST(igt_evict_for_cache_color),
+		SUBTEST(igt_evict_vm),
+		SUBTEST(igt_overcommit),
+	};
+	struct drm_i915_private *i915;
+	intel_wakeref_t wakeref;
+	int err = 0;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
+		err = i915_subtests(tests, to_gt(i915));
+
+	mock_destroy_device(i915);
+	return err;
+}
+
+int i915_gem_evict_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_evict_contexts),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
new file mode 100644
index 000000000..e050a2de5
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
@@ -0,0 +1,2328 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/list_sort.h>
+#include <linux/prime_numbers.h>
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_region.h"
+#include "gem/selftests/mock_context.h"
+#include "gt/intel_context.h"
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gtt.h"
+
+#include "i915_random.h"
+#include "i915_selftest.h"
+#include "i915_vma_resource.h"
+
+#include "mock_drm.h"
+#include "mock_gem_device.h"
+#include "mock_gtt.h"
+#include "igt_flush_test.h"
+
+static void cleanup_freed_objects(struct drm_i915_private *i915)
+{
+	i915_gem_drain_freed_objects(i915);
+}
+
+static void fake_free_pages(struct drm_i915_gem_object *obj,
+			    struct sg_table *pages)
+{
+	sg_free_table(pages);
+	kfree(pages);
+}
+
+static int fake_get_pages(struct drm_i915_gem_object *obj)
+{
+#define GFP (GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY)
+#define PFN_BIAS 0x1000
+	struct sg_table *pages;
+	struct scatterlist *sg;
+	unsigned int sg_page_sizes;
+	typeof(obj->base.size) rem;
+
+	pages = kmalloc(sizeof(*pages), GFP);
+	if (!pages)
+		return -ENOMEM;
+
+	rem = round_up(obj->base.size, BIT(31)) >> 31;
+	if (sg_alloc_table(pages, rem, GFP)) {
+		kfree(pages);
+		return -ENOMEM;
+	}
+
+	sg_page_sizes = 0;
+	rem = obj->base.size;
+	for (sg = pages->sgl; sg; sg = sg_next(sg)) {
+		unsigned long len = min_t(typeof(rem), rem, BIT(31));
+
+		GEM_BUG_ON(!len);
+		sg_set_page(sg, pfn_to_page(PFN_BIAS), len, 0);
+		sg_dma_address(sg) = page_to_phys(sg_page(sg));
+		sg_dma_len(sg) = len;
+		sg_page_sizes |= len;
+
+		rem -= len;
+	}
+	GEM_BUG_ON(rem);
+
+	__i915_gem_object_set_pages(obj, pages, sg_page_sizes);
+
+	return 0;
+#undef GFP
+}
+
+static void fake_put_pages(struct drm_i915_gem_object *obj,
+			   struct sg_table *pages)
+{
+	fake_free_pages(obj, pages);
+	obj->mm.dirty = false;
+}
+
+static const struct drm_i915_gem_object_ops fake_ops = {
+	.name = "fake-gem",
+	.flags = I915_GEM_OBJECT_IS_SHRINKABLE,
+	.get_pages = fake_get_pages,
+	.put_pages = fake_put_pages,
+};
+
+static struct drm_i915_gem_object *
+fake_dma_object(struct drm_i915_private *i915, u64 size)
+{
+	static struct lock_class_key lock_class;
+	struct drm_i915_gem_object *obj;
+
+	GEM_BUG_ON(!size);
+	GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE));
+
+	if (overflows_type(size, obj->base.size))
+		return ERR_PTR(-E2BIG);
+
+	obj = i915_gem_object_alloc();
+	if (!obj)
+		goto err;
+
+	drm_gem_private_object_init(&i915->drm, &obj->base, size);
+	i915_gem_object_init(obj, &fake_ops, &lock_class, 0);
+
+	i915_gem_object_set_volatile(obj);
+
+	obj->write_domain = I915_GEM_DOMAIN_CPU;
+	obj->read_domains = I915_GEM_DOMAIN_CPU;
+	obj->cache_level = I915_CACHE_NONE;
+
+	/* Preallocate the "backing storage" */
+	if (i915_gem_object_pin_pages_unlocked(obj))
+		goto err_obj;
+
+	i915_gem_object_unpin_pages(obj);
+	return obj;
+
+err_obj:
+	i915_gem_object_put(obj);
+err:
+	return ERR_PTR(-ENOMEM);
+}
+
+static int igt_ppgtt_alloc(void *arg)
+{
+	struct drm_i915_private *dev_priv = arg;
+	struct i915_ppgtt *ppgtt;
+	struct i915_gem_ww_ctx ww;
+	u64 size, last, limit;
+	int err = 0;
+
+	/* Allocate a ppggt and try to fill the entire range */
+
+	if (!HAS_PPGTT(dev_priv))
+		return 0;
+
+	ppgtt = i915_ppgtt_create(to_gt(dev_priv), 0);
+	if (IS_ERR(ppgtt))
+		return PTR_ERR(ppgtt);
+
+	if (!ppgtt->vm.allocate_va_range)
+		goto err_ppgtt_cleanup;
+
+	/*
+	 * While we only allocate the page tables here and so we could
+	 * address a much larger GTT than we could actually fit into
+	 * RAM, a practical limit is the amount of physical pages in the system.
+	 * This should ensure that we do not run into the oomkiller during
+	 * the test and take down the machine wilfully.
+	 */
+	limit = totalram_pages() << PAGE_SHIFT;
+	limit = min(ppgtt->vm.total, limit);
+
+	i915_gem_ww_ctx_init(&ww, false);
+retry:
+	err = i915_vm_lock_objects(&ppgtt->vm, &ww);
+	if (err)
+		goto err_ppgtt_cleanup;
+
+	/* Check we can allocate the entire range */
+	for (size = 4096; size <= limit; size <<= 2) {
+		struct i915_vm_pt_stash stash = {};
+
+		err = i915_vm_alloc_pt_stash(&ppgtt->vm, &stash, size);
+		if (err)
+			goto err_ppgtt_cleanup;
+
+		err = i915_vm_map_pt_stash(&ppgtt->vm, &stash);
+		if (err) {
+			i915_vm_free_pt_stash(&ppgtt->vm, &stash);
+			goto err_ppgtt_cleanup;
+		}
+
+		ppgtt->vm.allocate_va_range(&ppgtt->vm, &stash, 0, size);
+		cond_resched();
+
+		ppgtt->vm.clear_range(&ppgtt->vm, 0, size);
+
+		i915_vm_free_pt_stash(&ppgtt->vm, &stash);
+	}
+
+	/* Check we can incrementally allocate the entire range */
+	for (last = 0, size = 4096; size <= limit; last = size, size <<= 2) {
+		struct i915_vm_pt_stash stash = {};
+
+		err = i915_vm_alloc_pt_stash(&ppgtt->vm, &stash, size - last);
+		if (err)
+			goto err_ppgtt_cleanup;
+
+		err = i915_vm_map_pt_stash(&ppgtt->vm, &stash);
+		if (err) {
+			i915_vm_free_pt_stash(&ppgtt->vm, &stash);
+			goto err_ppgtt_cleanup;
+		}
+
+		ppgtt->vm.allocate_va_range(&ppgtt->vm, &stash,
+					    last, size - last);
+		cond_resched();
+
+		i915_vm_free_pt_stash(&ppgtt->vm, &stash);
+	}
+
+err_ppgtt_cleanup:
+	if (err == -EDEADLK) {
+		err = i915_gem_ww_ctx_backoff(&ww);
+		if (!err)
+			goto retry;
+	}
+	i915_gem_ww_ctx_fini(&ww);
+
+	i915_vm_put(&ppgtt->vm);
+	return err;
+}
+
+static int lowlevel_hole(struct i915_address_space *vm,
+			 u64 hole_start, u64 hole_end,
+			 unsigned long end_time)
+{
+	const unsigned int min_alignment =
+		i915_vm_min_alignment(vm, INTEL_MEMORY_SYSTEM);
+	I915_RND_STATE(seed_prng);
+	struct i915_vma_resource *mock_vma_res;
+	unsigned int size;
+
+	mock_vma_res = kzalloc(sizeof(*mock_vma_res), GFP_KERNEL);
+	if (!mock_vma_res)
+		return -ENOMEM;
+
+	/* Keep creating larger objects until one cannot fit into the hole */
+	for (size = 12; (hole_end - hole_start) >> size; size++) {
+		I915_RND_SUBSTATE(prng, seed_prng);
+		struct drm_i915_gem_object *obj;
+		unsigned int *order, count, n;
+		u64 hole_size, aligned_size;
+
+		aligned_size = max_t(u32, ilog2(min_alignment), size);
+		hole_size = (hole_end - hole_start) >> aligned_size;
+		if (hole_size > KMALLOC_MAX_SIZE / sizeof(u32))
+			hole_size = KMALLOC_MAX_SIZE / sizeof(u32);
+		count = hole_size >> 1;
+		if (!count) {
+			pr_debug("%s: hole is too small [%llx - %llx] >> %d: %lld\n",
+				 __func__, hole_start, hole_end, size, hole_size);
+			break;
+		}
+
+		do {
+			order = i915_random_order(count, &prng);
+			if (order)
+				break;
+		} while (count >>= 1);
+		if (!count) {
+			kfree(mock_vma_res);
+			return -ENOMEM;
+		}
+		GEM_BUG_ON(!order);
+
+		GEM_BUG_ON(count * BIT_ULL(aligned_size) > vm->total);
+		GEM_BUG_ON(hole_start + count * BIT_ULL(aligned_size) > hole_end);
+
+		/* Ignore allocation failures (i.e. don't report them as
+		 * a test failure) as we are purposefully allocating very
+		 * large objects without checking that we have sufficient
+		 * memory. We expect to hit -ENOMEM.
+		 */
+
+		obj = fake_dma_object(vm->i915, BIT_ULL(size));
+		if (IS_ERR(obj)) {
+			kfree(order);
+			break;
+		}
+
+		GEM_BUG_ON(obj->base.size != BIT_ULL(size));
+
+		if (i915_gem_object_pin_pages_unlocked(obj)) {
+			i915_gem_object_put(obj);
+			kfree(order);
+			break;
+		}
+
+		for (n = 0; n < count; n++) {
+			u64 addr = hole_start + order[n] * BIT_ULL(aligned_size);
+			intel_wakeref_t wakeref;
+
+			GEM_BUG_ON(addr + BIT_ULL(aligned_size) > vm->total);
+
+			if (igt_timeout(end_time,
+					"%s timed out before %d/%d\n",
+					__func__, n, count)) {
+				hole_end = hole_start; /* quit */
+				break;
+			}
+
+			if (vm->allocate_va_range) {
+				struct i915_vm_pt_stash stash = {};
+				struct i915_gem_ww_ctx ww;
+				int err;
+
+				i915_gem_ww_ctx_init(&ww, false);
+retry:
+				err = i915_vm_lock_objects(vm, &ww);
+				if (err)
+					goto alloc_vm_end;
+
+				err = -ENOMEM;
+				if (i915_vm_alloc_pt_stash(vm, &stash,
+							   BIT_ULL(size)))
+					goto alloc_vm_end;
+
+				err = i915_vm_map_pt_stash(vm, &stash);
+				if (!err)
+					vm->allocate_va_range(vm, &stash,
+							      addr, BIT_ULL(size));
+				i915_vm_free_pt_stash(vm, &stash);
+alloc_vm_end:
+				if (err == -EDEADLK) {
+					err = i915_gem_ww_ctx_backoff(&ww);
+					if (!err)
+						goto retry;
+				}
+				i915_gem_ww_ctx_fini(&ww);
+
+				if (err)
+					break;
+			}
+
+			mock_vma_res->bi.pages = obj->mm.pages;
+			mock_vma_res->node_size = BIT_ULL(aligned_size);
+			mock_vma_res->start = addr;
+
+			with_intel_runtime_pm(vm->gt->uncore->rpm, wakeref)
+			  vm->insert_entries(vm, mock_vma_res,
+						   I915_CACHE_NONE, 0);
+		}
+		count = n;
+
+		i915_random_reorder(order, count, &prng);
+		for (n = 0; n < count; n++) {
+			u64 addr = hole_start + order[n] * BIT_ULL(aligned_size);
+			intel_wakeref_t wakeref;
+
+			GEM_BUG_ON(addr + BIT_ULL(size) > vm->total);
+			with_intel_runtime_pm(vm->gt->uncore->rpm, wakeref)
+				vm->clear_range(vm, addr, BIT_ULL(size));
+		}
+
+		i915_gem_object_unpin_pages(obj);
+		i915_gem_object_put(obj);
+
+		kfree(order);
+
+		cleanup_freed_objects(vm->i915);
+	}
+
+	kfree(mock_vma_res);
+	return 0;
+}
+
+static void close_object_list(struct list_head *objects,
+			      struct i915_address_space *vm)
+{
+	struct drm_i915_gem_object *obj, *on;
+	int ignored;
+
+	list_for_each_entry_safe(obj, on, objects, st_link) {
+		struct i915_vma *vma;
+
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (!IS_ERR(vma))
+			ignored = i915_vma_unbind_unlocked(vma);
+
+		list_del(&obj->st_link);
+		i915_gem_object_put(obj);
+	}
+}
+
+static int fill_hole(struct i915_address_space *vm,
+		     u64 hole_start, u64 hole_end,
+		     unsigned long end_time)
+{
+	const u64 hole_size = hole_end - hole_start;
+	struct drm_i915_gem_object *obj;
+	const unsigned int min_alignment =
+		i915_vm_min_alignment(vm, INTEL_MEMORY_SYSTEM);
+	const unsigned long max_pages =
+		min_t(u64, ULONG_MAX - 1, (hole_size / 2) >> ilog2(min_alignment));
+	const unsigned long max_step = max(int_sqrt(max_pages), 2UL);
+	unsigned long npages, prime, flags;
+	struct i915_vma *vma;
+	LIST_HEAD(objects);
+	int err;
+
+	/* Try binding many VMA working inwards from either edge */
+
+	flags = PIN_OFFSET_FIXED | PIN_USER;
+	if (i915_is_ggtt(vm))
+		flags |= PIN_GLOBAL;
+
+	for_each_prime_number_from(prime, 2, max_step) {
+		for (npages = 1; npages <= max_pages; npages *= prime) {
+			const u64 full_size = npages << PAGE_SHIFT;
+			const struct {
+				const char *name;
+				u64 offset;
+				int step;
+			} phases[] = {
+				{ "top-down", hole_end, -1, },
+				{ "bottom-up", hole_start, 1, },
+				{ }
+			}, *p;
+
+			obj = fake_dma_object(vm->i915, full_size);
+			if (IS_ERR(obj))
+				break;
+
+			list_add(&obj->st_link, &objects);
+
+			/* Align differing sized objects against the edges, and
+			 * check we don't walk off into the void when binding
+			 * them into the GTT.
+			 */
+			for (p = phases; p->name; p++) {
+				u64 offset;
+
+				offset = p->offset;
+				list_for_each_entry(obj, &objects, st_link) {
+					u64 aligned_size = round_up(obj->base.size,
+								    min_alignment);
+
+					vma = i915_vma_instance(obj, vm, NULL);
+					if (IS_ERR(vma))
+						continue;
+
+					if (p->step < 0) {
+						if (offset < hole_start + aligned_size)
+							break;
+						offset -= aligned_size;
+					}
+
+					err = i915_vma_pin(vma, 0, 0, offset | flags);
+					if (err) {
+						pr_err("%s(%s) pin (forward) failed with err=%d on size=%lu pages (prime=%lu), offset=%llx\n",
+						       __func__, p->name, err, npages, prime, offset);
+						goto err;
+					}
+
+					if (!drm_mm_node_allocated(&vma->node) ||
+					    i915_vma_misplaced(vma, 0, 0, offset | flags)) {
+						pr_err("%s(%s) (forward) insert failed: vma.node=%llx + %llx [allocated? %d], expected offset %llx\n",
+						       __func__, p->name, vma->node.start, vma->node.size, drm_mm_node_allocated(&vma->node),
+						       offset);
+						err = -EINVAL;
+						goto err;
+					}
+
+					i915_vma_unpin(vma);
+
+					if (p->step > 0) {
+						if (offset + aligned_size > hole_end)
+							break;
+						offset += aligned_size;
+					}
+				}
+
+				offset = p->offset;
+				list_for_each_entry(obj, &objects, st_link) {
+					u64 aligned_size = round_up(obj->base.size,
+								    min_alignment);
+
+					vma = i915_vma_instance(obj, vm, NULL);
+					if (IS_ERR(vma))
+						continue;
+
+					if (p->step < 0) {
+						if (offset < hole_start + aligned_size)
+							break;
+						offset -= aligned_size;
+					}
+
+					if (!drm_mm_node_allocated(&vma->node) ||
+					    i915_vma_misplaced(vma, 0, 0, offset | flags)) {
+						pr_err("%s(%s) (forward) moved vma.node=%llx + %llx, expected offset %llx\n",
+						       __func__, p->name, vma->node.start, vma->node.size,
+						       offset);
+						err = -EINVAL;
+						goto err;
+					}
+
+					err = i915_vma_unbind_unlocked(vma);
+					if (err) {
+						pr_err("%s(%s) (forward) unbind of vma.node=%llx + %llx failed with err=%d\n",
+						       __func__, p->name, vma->node.start, vma->node.size,
+						       err);
+						goto err;
+					}
+
+					if (p->step > 0) {
+						if (offset + aligned_size > hole_end)
+							break;
+						offset += aligned_size;
+					}
+				}
+
+				offset = p->offset;
+				list_for_each_entry_reverse(obj, &objects, st_link) {
+					u64 aligned_size = round_up(obj->base.size,
+								    min_alignment);
+
+					vma = i915_vma_instance(obj, vm, NULL);
+					if (IS_ERR(vma))
+						continue;
+
+					if (p->step < 0) {
+						if (offset < hole_start + aligned_size)
+							break;
+						offset -= aligned_size;
+					}
+
+					err = i915_vma_pin(vma, 0, 0, offset | flags);
+					if (err) {
+						pr_err("%s(%s) pin (backward) failed with err=%d on size=%lu pages (prime=%lu), offset=%llx\n",
+						       __func__, p->name, err, npages, prime, offset);
+						goto err;
+					}
+
+					if (!drm_mm_node_allocated(&vma->node) ||
+					    i915_vma_misplaced(vma, 0, 0, offset | flags)) {
+						pr_err("%s(%s) (backward) insert failed: vma.node=%llx + %llx [allocated? %d], expected offset %llx\n",
+						       __func__, p->name, vma->node.start, vma->node.size, drm_mm_node_allocated(&vma->node),
+						       offset);
+						err = -EINVAL;
+						goto err;
+					}
+
+					i915_vma_unpin(vma);
+
+					if (p->step > 0) {
+						if (offset + aligned_size > hole_end)
+							break;
+						offset += aligned_size;
+					}
+				}
+
+				offset = p->offset;
+				list_for_each_entry_reverse(obj, &objects, st_link) {
+					u64 aligned_size = round_up(obj->base.size,
+								    min_alignment);
+
+					vma = i915_vma_instance(obj, vm, NULL);
+					if (IS_ERR(vma))
+						continue;
+
+					if (p->step < 0) {
+						if (offset < hole_start + aligned_size)
+							break;
+						offset -= aligned_size;
+					}
+
+					if (!drm_mm_node_allocated(&vma->node) ||
+					    i915_vma_misplaced(vma, 0, 0, offset | flags)) {
+						pr_err("%s(%s) (backward) moved vma.node=%llx + %llx [allocated? %d], expected offset %llx\n",
+						       __func__, p->name, vma->node.start, vma->node.size, drm_mm_node_allocated(&vma->node),
+						       offset);
+						err = -EINVAL;
+						goto err;
+					}
+
+					err = i915_vma_unbind_unlocked(vma);
+					if (err) {
+						pr_err("%s(%s) (backward) unbind of vma.node=%llx + %llx failed with err=%d\n",
+						       __func__, p->name, vma->node.start, vma->node.size,
+						       err);
+						goto err;
+					}
+
+					if (p->step > 0) {
+						if (offset + aligned_size > hole_end)
+							break;
+						offset += aligned_size;
+					}
+				}
+			}
+
+			if (igt_timeout(end_time, "%s timed out (npages=%lu, prime=%lu)\n",
+					__func__, npages, prime)) {
+				err = -EINTR;
+				goto err;
+			}
+		}
+
+		close_object_list(&objects, vm);
+		cleanup_freed_objects(vm->i915);
+	}
+
+	return 0;
+
+err:
+	close_object_list(&objects, vm);
+	return err;
+}
+
+static int walk_hole(struct i915_address_space *vm,
+		     u64 hole_start, u64 hole_end,
+		     unsigned long end_time)
+{
+	const u64 hole_size = hole_end - hole_start;
+	const unsigned long max_pages =
+		min_t(u64, ULONG_MAX - 1, hole_size >> PAGE_SHIFT);
+	unsigned long min_alignment;
+	unsigned long flags;
+	u64 size;
+
+	/* Try binding a single VMA in different positions within the hole */
+
+	flags = PIN_OFFSET_FIXED | PIN_USER;
+	if (i915_is_ggtt(vm))
+		flags |= PIN_GLOBAL;
+
+	min_alignment = i915_vm_min_alignment(vm, INTEL_MEMORY_SYSTEM);
+
+	for_each_prime_number_from(size, 1, max_pages) {
+		struct drm_i915_gem_object *obj;
+		struct i915_vma *vma;
+		u64 addr;
+		int err = 0;
+
+		obj = fake_dma_object(vm->i915, size << PAGE_SHIFT);
+		if (IS_ERR(obj))
+			break;
+
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto err_put;
+		}
+
+		for (addr = hole_start;
+		     addr + obj->base.size < hole_end;
+		     addr += round_up(obj->base.size, min_alignment)) {
+			err = i915_vma_pin(vma, 0, 0, addr | flags);
+			if (err) {
+				pr_err("%s bind failed at %llx + %llx [hole %llx- %llx] with err=%d\n",
+				       __func__, addr, vma->size,
+				       hole_start, hole_end, err);
+				goto err_put;
+			}
+			i915_vma_unpin(vma);
+
+			if (!drm_mm_node_allocated(&vma->node) ||
+			    i915_vma_misplaced(vma, 0, 0, addr | flags)) {
+				pr_err("%s incorrect at %llx + %llx\n",
+				       __func__, addr, vma->size);
+				err = -EINVAL;
+				goto err_put;
+			}
+
+			err = i915_vma_unbind_unlocked(vma);
+			if (err) {
+				pr_err("%s unbind failed at %llx + %llx  with err=%d\n",
+				       __func__, addr, vma->size, err);
+				goto err_put;
+			}
+
+			GEM_BUG_ON(drm_mm_node_allocated(&vma->node));
+
+			if (igt_timeout(end_time,
+					"%s timed out at %llx\n",
+					__func__, addr)) {
+				err = -EINTR;
+				goto err_put;
+			}
+		}
+
+err_put:
+		i915_gem_object_put(obj);
+		if (err)
+			return err;
+
+		cleanup_freed_objects(vm->i915);
+	}
+
+	return 0;
+}
+
+static int pot_hole(struct i915_address_space *vm,
+		    u64 hole_start, u64 hole_end,
+		    unsigned long end_time)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	unsigned int min_alignment;
+	unsigned long flags;
+	unsigned int pot;
+	int err = 0;
+
+	flags = PIN_OFFSET_FIXED | PIN_USER;
+	if (i915_is_ggtt(vm))
+		flags |= PIN_GLOBAL;
+
+	min_alignment = i915_vm_min_alignment(vm, INTEL_MEMORY_SYSTEM);
+
+	obj = i915_gem_object_create_internal(vm->i915, 2 * I915_GTT_PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	/* Insert a pair of pages across every pot boundary within the hole */
+	for (pot = fls64(hole_end - 1) - 1;
+	     pot > ilog2(2 * min_alignment);
+	     pot--) {
+		u64 step = BIT_ULL(pot);
+		u64 addr;
+
+		for (addr = round_up(hole_start + min_alignment, step) - min_alignment;
+		     hole_end > addr && hole_end - addr >= 2 * min_alignment;
+		     addr += step) {
+			err = i915_vma_pin(vma, 0, 0, addr | flags);
+			if (err) {
+				pr_err("%s failed to pin object at %llx in hole [%llx - %llx], with err=%d\n",
+				       __func__,
+				       addr,
+				       hole_start, hole_end,
+				       err);
+				goto err_obj;
+			}
+
+			if (!drm_mm_node_allocated(&vma->node) ||
+			    i915_vma_misplaced(vma, 0, 0, addr | flags)) {
+				pr_err("%s incorrect at %llx + %llx\n",
+				       __func__, addr, vma->size);
+				i915_vma_unpin(vma);
+				err = i915_vma_unbind_unlocked(vma);
+				err = -EINVAL;
+				goto err_obj;
+			}
+
+			i915_vma_unpin(vma);
+			err = i915_vma_unbind_unlocked(vma);
+			GEM_BUG_ON(err);
+		}
+
+		if (igt_timeout(end_time,
+				"%s timed out after %d/%d\n",
+				__func__, pot, fls64(hole_end - 1) - 1)) {
+			err = -EINTR;
+			goto err_obj;
+		}
+	}
+
+err_obj:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static int drunk_hole(struct i915_address_space *vm,
+		      u64 hole_start, u64 hole_end,
+		      unsigned long end_time)
+{
+	I915_RND_STATE(prng);
+	unsigned int min_alignment;
+	unsigned int size;
+	unsigned long flags;
+
+	flags = PIN_OFFSET_FIXED | PIN_USER;
+	if (i915_is_ggtt(vm))
+		flags |= PIN_GLOBAL;
+
+	min_alignment = i915_vm_min_alignment(vm, INTEL_MEMORY_SYSTEM);
+
+	/* Keep creating larger objects until one cannot fit into the hole */
+	for (size = 12; (hole_end - hole_start) >> size; size++) {
+		struct drm_i915_gem_object *obj;
+		unsigned int *order, count, n;
+		struct i915_vma *vma;
+		u64 hole_size, aligned_size;
+		int err = -ENODEV;
+
+		aligned_size = max_t(u32, ilog2(min_alignment), size);
+		hole_size = (hole_end - hole_start) >> aligned_size;
+		if (hole_size > KMALLOC_MAX_SIZE / sizeof(u32))
+			hole_size = KMALLOC_MAX_SIZE / sizeof(u32);
+		count = hole_size >> 1;
+		if (!count) {
+			pr_debug("%s: hole is too small [%llx - %llx] >> %d: %lld\n",
+				 __func__, hole_start, hole_end, size, hole_size);
+			break;
+		}
+
+		do {
+			order = i915_random_order(count, &prng);
+			if (order)
+				break;
+		} while (count >>= 1);
+		if (!count)
+			return -ENOMEM;
+		GEM_BUG_ON(!order);
+
+		/* Ignore allocation failures (i.e. don't report them as
+		 * a test failure) as we are purposefully allocating very
+		 * large objects without checking that we have sufficient
+		 * memory. We expect to hit -ENOMEM.
+		 */
+
+		obj = fake_dma_object(vm->i915, BIT_ULL(size));
+		if (IS_ERR(obj)) {
+			kfree(order);
+			break;
+		}
+
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto err_obj;
+		}
+
+		GEM_BUG_ON(vma->size != BIT_ULL(size));
+
+		for (n = 0; n < count; n++) {
+			u64 addr = hole_start + order[n] * BIT_ULL(aligned_size);
+
+			err = i915_vma_pin(vma, 0, 0, addr | flags);
+			if (err) {
+				pr_err("%s failed to pin object at %llx + %llx in hole [%llx - %llx], with err=%d\n",
+				       __func__,
+				       addr, BIT_ULL(size),
+				       hole_start, hole_end,
+				       err);
+				goto err_obj;
+			}
+
+			if (!drm_mm_node_allocated(&vma->node) ||
+			    i915_vma_misplaced(vma, 0, 0, addr | flags)) {
+				pr_err("%s incorrect at %llx + %llx\n",
+				       __func__, addr, BIT_ULL(size));
+				i915_vma_unpin(vma);
+				err = i915_vma_unbind_unlocked(vma);
+				err = -EINVAL;
+				goto err_obj;
+			}
+
+			i915_vma_unpin(vma);
+			err = i915_vma_unbind_unlocked(vma);
+			GEM_BUG_ON(err);
+
+			if (igt_timeout(end_time,
+					"%s timed out after %d/%d\n",
+					__func__, n, count)) {
+				err = -EINTR;
+				goto err_obj;
+			}
+		}
+
+err_obj:
+		i915_gem_object_put(obj);
+		kfree(order);
+		if (err)
+			return err;
+
+		cleanup_freed_objects(vm->i915);
+	}
+
+	return 0;
+}
+
+static int __shrink_hole(struct i915_address_space *vm,
+			 u64 hole_start, u64 hole_end,
+			 unsigned long end_time)
+{
+	struct drm_i915_gem_object *obj;
+	unsigned long flags = PIN_OFFSET_FIXED | PIN_USER;
+	unsigned int min_alignment;
+	unsigned int order = 12;
+	LIST_HEAD(objects);
+	int err = 0;
+	u64 addr;
+
+	min_alignment = i915_vm_min_alignment(vm, INTEL_MEMORY_SYSTEM);
+
+	/* Keep creating larger objects until one cannot fit into the hole */
+	for (addr = hole_start; addr < hole_end; ) {
+		struct i915_vma *vma;
+		u64 size = BIT_ULL(order++);
+
+		size = min(size, hole_end - addr);
+		obj = fake_dma_object(vm->i915, size);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			break;
+		}
+
+		list_add(&obj->st_link, &objects);
+
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			break;
+		}
+
+		GEM_BUG_ON(vma->size != size);
+
+		err = i915_vma_pin(vma, 0, 0, addr | flags);
+		if (err) {
+			pr_err("%s failed to pin object at %llx + %llx in hole [%llx - %llx], with err=%d\n",
+			       __func__, addr, size, hole_start, hole_end, err);
+			break;
+		}
+
+		if (!drm_mm_node_allocated(&vma->node) ||
+		    i915_vma_misplaced(vma, 0, 0, addr | flags)) {
+			pr_err("%s incorrect at %llx + %llx\n",
+			       __func__, addr, size);
+			i915_vma_unpin(vma);
+			err = i915_vma_unbind_unlocked(vma);
+			err = -EINVAL;
+			break;
+		}
+
+		i915_vma_unpin(vma);
+		addr += round_up(size, min_alignment);
+
+		/*
+		 * Since we are injecting allocation faults at random intervals,
+		 * wait for this allocation to complete before we change the
+		 * faultinjection.
+		 */
+		err = i915_vma_sync(vma);
+		if (err)
+			break;
+
+		if (igt_timeout(end_time,
+				"%s timed out at ofset %llx [%llx - %llx]\n",
+				__func__, addr, hole_start, hole_end)) {
+			err = -EINTR;
+			break;
+		}
+	}
+
+	close_object_list(&objects, vm);
+	cleanup_freed_objects(vm->i915);
+	return err;
+}
+
+static int shrink_hole(struct i915_address_space *vm,
+		       u64 hole_start, u64 hole_end,
+		       unsigned long end_time)
+{
+	unsigned long prime;
+	int err;
+
+	vm->fault_attr.probability = 999;
+	atomic_set(&vm->fault_attr.times, -1);
+
+	for_each_prime_number_from(prime, 0, ULONG_MAX - 1) {
+		vm->fault_attr.interval = prime;
+		err = __shrink_hole(vm, hole_start, hole_end, end_time);
+		if (err)
+			break;
+	}
+
+	memset(&vm->fault_attr, 0, sizeof(vm->fault_attr));
+
+	return err;
+}
+
+static int shrink_boom(struct i915_address_space *vm,
+		       u64 hole_start, u64 hole_end,
+		       unsigned long end_time)
+{
+	unsigned int sizes[] = { SZ_2M, SZ_1G };
+	struct drm_i915_gem_object *purge;
+	struct drm_i915_gem_object *explode;
+	int err;
+	int i;
+
+	/*
+	 * Catch the case which shrink_hole seems to miss. The setup here
+	 * requires invoking the shrinker as we do the alloc_pt/alloc_pd, while
+	 * ensuring that all vma assiocated with the respective pd/pdp are
+	 * unpinned at the time.
+	 */
+
+	for (i = 0; i < ARRAY_SIZE(sizes); ++i) {
+		unsigned int flags = PIN_USER | PIN_OFFSET_FIXED;
+		unsigned int size = sizes[i];
+		struct i915_vma *vma;
+
+		purge = fake_dma_object(vm->i915, size);
+		if (IS_ERR(purge))
+			return PTR_ERR(purge);
+
+		vma = i915_vma_instance(purge, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto err_purge;
+		}
+
+		err = i915_vma_pin(vma, 0, 0, flags);
+		if (err)
+			goto err_purge;
+
+		/* Should now be ripe for purging */
+		i915_vma_unpin(vma);
+
+		explode = fake_dma_object(vm->i915, size);
+		if (IS_ERR(explode)) {
+			err = PTR_ERR(explode);
+			goto err_purge;
+		}
+
+		vm->fault_attr.probability = 100;
+		vm->fault_attr.interval = 1;
+		atomic_set(&vm->fault_attr.times, -1);
+
+		vma = i915_vma_instance(explode, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto err_explode;
+		}
+
+		err = i915_vma_pin(vma, 0, 0, flags | size);
+		if (err)
+			goto err_explode;
+
+		i915_vma_unpin(vma);
+
+		i915_gem_object_put(purge);
+		i915_gem_object_put(explode);
+
+		memset(&vm->fault_attr, 0, sizeof(vm->fault_attr));
+		cleanup_freed_objects(vm->i915);
+	}
+
+	return 0;
+
+err_explode:
+	i915_gem_object_put(explode);
+err_purge:
+	i915_gem_object_put(purge);
+	memset(&vm->fault_attr, 0, sizeof(vm->fault_attr));
+	return err;
+}
+
+static int misaligned_case(struct i915_address_space *vm, struct intel_memory_region *mr,
+			   u64 addr, u64 size, unsigned long flags)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err = 0;
+	u64 expected_vma_size, expected_node_size;
+	bool is_stolen = mr->type == INTEL_MEMORY_STOLEN_SYSTEM ||
+			 mr->type == INTEL_MEMORY_STOLEN_LOCAL;
+
+	obj = i915_gem_object_create_region(mr, size, 0, I915_BO_ALLOC_GPU_ONLY);
+	if (IS_ERR(obj)) {
+		/* if iGVT-g or DMAR is active, stolen mem will be uninitialized */
+		if (PTR_ERR(obj) == -ENODEV && is_stolen)
+			return 0;
+		return PTR_ERR(obj);
+	}
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_put;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, addr | flags);
+	if (err)
+		goto err_put;
+	i915_vma_unpin(vma);
+
+	if (!drm_mm_node_allocated(&vma->node)) {
+		err = -EINVAL;
+		goto err_put;
+	}
+
+	if (i915_vma_misplaced(vma, 0, 0, addr | flags)) {
+		err = -EINVAL;
+		goto err_put;
+	}
+
+	expected_vma_size = round_up(size, 1 << (ffs(vma->resource->page_sizes_gtt) - 1));
+	expected_node_size = expected_vma_size;
+
+	if (HAS_64K_PAGES(vm->i915) && i915_gem_object_is_lmem(obj)) {
+		/*
+		 * The compact-pt should expand lmem node to 2MB for the ppGTT,
+		 * for all other cases we should only expect 64K.
+		 */
+		expected_vma_size = round_up(size, I915_GTT_PAGE_SIZE_64K);
+		if (NEEDS_COMPACT_PT(vm->i915) && !i915_is_ggtt(vm))
+			expected_node_size = round_up(size, I915_GTT_PAGE_SIZE_2M);
+		else
+			expected_node_size = round_up(size, I915_GTT_PAGE_SIZE_64K);
+	}
+
+	if (vma->size != expected_vma_size || vma->node.size != expected_node_size) {
+		err = i915_vma_unbind_unlocked(vma);
+		err = -EBADSLT;
+		goto err_put;
+	}
+
+	err = i915_vma_unbind_unlocked(vma);
+	if (err)
+		goto err_put;
+
+	GEM_BUG_ON(drm_mm_node_allocated(&vma->node));
+
+err_put:
+	i915_gem_object_put(obj);
+	cleanup_freed_objects(vm->i915);
+	return err;
+}
+
+static int misaligned_pin(struct i915_address_space *vm,
+			  u64 hole_start, u64 hole_end,
+			  unsigned long end_time)
+{
+	struct intel_memory_region *mr;
+	enum intel_region_id id;
+	unsigned long flags = PIN_OFFSET_FIXED | PIN_USER;
+	int err = 0;
+	u64 hole_size = hole_end - hole_start;
+
+	if (i915_is_ggtt(vm))
+		flags |= PIN_GLOBAL;
+
+	for_each_memory_region(mr, vm->i915, id) {
+		u64 min_alignment = i915_vm_min_alignment(vm, mr->type);
+		u64 size = min_alignment;
+		u64 addr = round_down(hole_start + (hole_size / 2), min_alignment);
+
+		/* avoid -ENOSPC on very small hole setups */
+		if (hole_size < 3 * min_alignment)
+			continue;
+
+		/* we can't test < 4k alignment due to flags being encoded in lower bits */
+		if (min_alignment != I915_GTT_PAGE_SIZE_4K) {
+			err = misaligned_case(vm, mr, addr + (min_alignment / 2), size, flags);
+			/* misaligned should error with -EINVAL*/
+			if (!err)
+				err = -EBADSLT;
+			if (err != -EINVAL)
+				return err;
+		}
+
+		/* test for vma->size expansion to min page size */
+		err = misaligned_case(vm, mr, addr, PAGE_SIZE, flags);
+		if (err)
+			return err;
+
+		/* test for intermediate size not expanding vma->size for large alignments */
+		err = misaligned_case(vm, mr, addr, size / 2, flags);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int exercise_ppgtt(struct drm_i915_private *dev_priv,
+			  int (*func)(struct i915_address_space *vm,
+				      u64 hole_start, u64 hole_end,
+				      unsigned long end_time))
+{
+	struct i915_ppgtt *ppgtt;
+	IGT_TIMEOUT(end_time);
+	struct file *file;
+	int err;
+
+	if (!HAS_FULL_PPGTT(dev_priv))
+		return 0;
+
+	file = mock_file(dev_priv);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	ppgtt = i915_ppgtt_create(to_gt(dev_priv), 0);
+	if (IS_ERR(ppgtt)) {
+		err = PTR_ERR(ppgtt);
+		goto out_free;
+	}
+	GEM_BUG_ON(offset_in_page(ppgtt->vm.total));
+	assert_vm_alive(&ppgtt->vm);
+
+	err = func(&ppgtt->vm, 0, ppgtt->vm.total, end_time);
+
+	i915_vm_put(&ppgtt->vm);
+
+out_free:
+	fput(file);
+	return err;
+}
+
+static int igt_ppgtt_fill(void *arg)
+{
+	return exercise_ppgtt(arg, fill_hole);
+}
+
+static int igt_ppgtt_walk(void *arg)
+{
+	return exercise_ppgtt(arg, walk_hole);
+}
+
+static int igt_ppgtt_pot(void *arg)
+{
+	return exercise_ppgtt(arg, pot_hole);
+}
+
+static int igt_ppgtt_drunk(void *arg)
+{
+	return exercise_ppgtt(arg, drunk_hole);
+}
+
+static int igt_ppgtt_lowlevel(void *arg)
+{
+	return exercise_ppgtt(arg, lowlevel_hole);
+}
+
+static int igt_ppgtt_shrink(void *arg)
+{
+	return exercise_ppgtt(arg, shrink_hole);
+}
+
+static int igt_ppgtt_shrink_boom(void *arg)
+{
+	return exercise_ppgtt(arg, shrink_boom);
+}
+
+static int igt_ppgtt_misaligned_pin(void *arg)
+{
+	return exercise_ppgtt(arg, misaligned_pin);
+}
+
+static int sort_holes(void *priv, const struct list_head *A,
+		      const struct list_head *B)
+{
+	struct drm_mm_node *a = list_entry(A, typeof(*a), hole_stack);
+	struct drm_mm_node *b = list_entry(B, typeof(*b), hole_stack);
+
+	if (a->start < b->start)
+		return -1;
+	else
+		return 1;
+}
+
+static int exercise_ggtt(struct drm_i915_private *i915,
+			 int (*func)(struct i915_address_space *vm,
+				     u64 hole_start, u64 hole_end,
+				     unsigned long end_time))
+{
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	u64 hole_start, hole_end, last = 0;
+	struct drm_mm_node *node;
+	IGT_TIMEOUT(end_time);
+	int err = 0;
+
+restart:
+	list_sort(NULL, &ggtt->vm.mm.hole_stack, sort_holes);
+	drm_mm_for_each_hole(node, &ggtt->vm.mm, hole_start, hole_end) {
+		if (hole_start < last)
+			continue;
+
+		if (ggtt->vm.mm.color_adjust)
+			ggtt->vm.mm.color_adjust(node, 0,
+						 &hole_start, &hole_end);
+		if (hole_start >= hole_end)
+			continue;
+
+		err = func(&ggtt->vm, hole_start, hole_end, end_time);
+		if (err)
+			break;
+
+		/* As we have manipulated the drm_mm, the list may be corrupt */
+		last = hole_end;
+		goto restart;
+	}
+
+	return err;
+}
+
+static int igt_ggtt_fill(void *arg)
+{
+	return exercise_ggtt(arg, fill_hole);
+}
+
+static int igt_ggtt_walk(void *arg)
+{
+	return exercise_ggtt(arg, walk_hole);
+}
+
+static int igt_ggtt_pot(void *arg)
+{
+	return exercise_ggtt(arg, pot_hole);
+}
+
+static int igt_ggtt_drunk(void *arg)
+{
+	return exercise_ggtt(arg, drunk_hole);
+}
+
+static int igt_ggtt_lowlevel(void *arg)
+{
+	return exercise_ggtt(arg, lowlevel_hole);
+}
+
+static int igt_ggtt_misaligned_pin(void *arg)
+{
+	return exercise_ggtt(arg, misaligned_pin);
+}
+
+static int igt_ggtt_page(void *arg)
+{
+	const unsigned int count = PAGE_SIZE/sizeof(u32);
+	I915_RND_STATE(prng);
+	struct drm_i915_private *i915 = arg;
+	struct i915_ggtt *ggtt = to_gt(i915)->ggtt;
+	struct drm_i915_gem_object *obj;
+	intel_wakeref_t wakeref;
+	struct drm_mm_node tmp;
+	unsigned int *order, n;
+	int err;
+
+	if (!i915_ggtt_has_aperture(ggtt))
+		return 0;
+
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	err = i915_gem_object_pin_pages_unlocked(obj);
+	if (err)
+		goto out_free;
+
+	memset(&tmp, 0, sizeof(tmp));
+	mutex_lock(&ggtt->vm.mutex);
+	err = drm_mm_insert_node_in_range(&ggtt->vm.mm, &tmp,
+					  count * PAGE_SIZE, 0,
+					  I915_COLOR_UNEVICTABLE,
+					  0, ggtt->mappable_end,
+					  DRM_MM_INSERT_LOW);
+	mutex_unlock(&ggtt->vm.mutex);
+	if (err)
+		goto out_unpin;
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	for (n = 0; n < count; n++) {
+		u64 offset = tmp.start + n * PAGE_SIZE;
+
+		ggtt->vm.insert_page(&ggtt->vm,
+				     i915_gem_object_get_dma_address(obj, 0),
+				     offset, I915_CACHE_NONE, 0);
+	}
+
+	order = i915_random_order(count, &prng);
+	if (!order) {
+		err = -ENOMEM;
+		goto out_remove;
+	}
+
+	for (n = 0; n < count; n++) {
+		u64 offset = tmp.start + order[n] * PAGE_SIZE;
+		u32 __iomem *vaddr;
+
+		vaddr = io_mapping_map_atomic_wc(&ggtt->iomap, offset);
+		iowrite32(n, vaddr + n);
+		io_mapping_unmap_atomic(vaddr);
+	}
+	intel_gt_flush_ggtt_writes(ggtt->vm.gt);
+
+	i915_random_reorder(order, count, &prng);
+	for (n = 0; n < count; n++) {
+		u64 offset = tmp.start + order[n] * PAGE_SIZE;
+		u32 __iomem *vaddr;
+		u32 val;
+
+		vaddr = io_mapping_map_atomic_wc(&ggtt->iomap, offset);
+		val = ioread32(vaddr + n);
+		io_mapping_unmap_atomic(vaddr);
+
+		if (val != n) {
+			pr_err("insert page failed: found %d, expected %d\n",
+			       val, n);
+			err = -EINVAL;
+			break;
+		}
+	}
+
+	kfree(order);
+out_remove:
+	ggtt->vm.clear_range(&ggtt->vm, tmp.start, tmp.size);
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+	mutex_lock(&ggtt->vm.mutex);
+	drm_mm_remove_node(&tmp);
+	mutex_unlock(&ggtt->vm.mutex);
+out_unpin:
+	i915_gem_object_unpin_pages(obj);
+out_free:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static void track_vma_bind(struct i915_vma *vma)
+{
+	struct drm_i915_gem_object *obj = vma->obj;
+
+	__i915_gem_object_pin_pages(obj);
+
+	GEM_BUG_ON(atomic_read(&vma->pages_count));
+	atomic_set(&vma->pages_count, I915_VMA_PAGES_ACTIVE);
+	__i915_gem_object_pin_pages(obj);
+	vma->pages = obj->mm.pages;
+	vma->resource->bi.pages = vma->pages;
+
+	mutex_lock(&vma->vm->mutex);
+	list_move_tail(&vma->vm_link, &vma->vm->bound_list);
+	mutex_unlock(&vma->vm->mutex);
+}
+
+static int exercise_mock(struct drm_i915_private *i915,
+			 int (*func)(struct i915_address_space *vm,
+				     u64 hole_start, u64 hole_end,
+				     unsigned long end_time))
+{
+	const u64 limit = totalram_pages() << PAGE_SHIFT;
+	struct i915_address_space *vm;
+	struct i915_gem_context *ctx;
+	IGT_TIMEOUT(end_time);
+	int err;
+
+	ctx = mock_context(i915, "mock");
+	if (!ctx)
+		return -ENOMEM;
+
+	vm = i915_gem_context_get_eb_vm(ctx);
+	err = func(vm, 0, min(vm->total, limit), end_time);
+	i915_vm_put(vm);
+
+	mock_context_close(ctx);
+	return err;
+}
+
+static int igt_mock_fill(void *arg)
+{
+	struct i915_ggtt *ggtt = arg;
+
+	return exercise_mock(ggtt->vm.i915, fill_hole);
+}
+
+static int igt_mock_walk(void *arg)
+{
+	struct i915_ggtt *ggtt = arg;
+
+	return exercise_mock(ggtt->vm.i915, walk_hole);
+}
+
+static int igt_mock_pot(void *arg)
+{
+	struct i915_ggtt *ggtt = arg;
+
+	return exercise_mock(ggtt->vm.i915, pot_hole);
+}
+
+static int igt_mock_drunk(void *arg)
+{
+	struct i915_ggtt *ggtt = arg;
+
+	return exercise_mock(ggtt->vm.i915, drunk_hole);
+}
+
+static int reserve_gtt_with_resource(struct i915_vma *vma, u64 offset)
+{
+	struct i915_address_space *vm = vma->vm;
+	struct i915_vma_resource *vma_res;
+	struct drm_i915_gem_object *obj = vma->obj;
+	int err;
+
+	vma_res = i915_vma_resource_alloc();
+	if (IS_ERR(vma_res))
+		return PTR_ERR(vma_res);
+
+	mutex_lock(&vm->mutex);
+	err = i915_gem_gtt_reserve(vm, NULL, &vma->node, obj->base.size,
+				   offset,
+				   obj->cache_level,
+				   0);
+	if (!err) {
+		i915_vma_resource_init_from_vma(vma_res, vma);
+		vma->resource = vma_res;
+	} else {
+		kfree(vma_res);
+	}
+	mutex_unlock(&vm->mutex);
+
+	return err;
+}
+
+static int igt_gtt_reserve(void *arg)
+{
+	struct i915_ggtt *ggtt = arg;
+	struct drm_i915_gem_object *obj, *on;
+	I915_RND_STATE(prng);
+	LIST_HEAD(objects);
+	u64 total;
+	int err = -ENODEV;
+
+	/* i915_gem_gtt_reserve() tries to reserve the precise range
+	 * for the node, and evicts if it has to. So our test checks that
+	 * it can give us the requsted space and prevent overlaps.
+	 */
+
+	/* Start by filling the GGTT */
+	for (total = 0;
+	     total + 2 * I915_GTT_PAGE_SIZE <= ggtt->vm.total;
+	     total += 2 * I915_GTT_PAGE_SIZE) {
+		struct i915_vma *vma;
+
+		obj = i915_gem_object_create_internal(ggtt->vm.i915,
+						      2 * PAGE_SIZE);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto out;
+		}
+
+		err = i915_gem_object_pin_pages_unlocked(obj);
+		if (err) {
+			i915_gem_object_put(obj);
+			goto out;
+		}
+
+		list_add(&obj->st_link, &objects);
+		vma = i915_vma_instance(obj, &ggtt->vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		err = reserve_gtt_with_resource(vma, total);
+		if (err) {
+			pr_err("i915_gem_gtt_reserve (pass 1) failed at %llu/%llu with err=%d\n",
+			       total, ggtt->vm.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+		if (vma->node.start != total ||
+		    vma->node.size != 2*I915_GTT_PAGE_SIZE) {
+			pr_err("i915_gem_gtt_reserve (pass 1) placement failed, found (%llx + %llx), expected (%llx + %llx)\n",
+			       vma->node.start, vma->node.size,
+			       total, 2*I915_GTT_PAGE_SIZE);
+			err = -EINVAL;
+			goto out;
+		}
+	}
+
+	/* Now we start forcing evictions */
+	for (total = I915_GTT_PAGE_SIZE;
+	     total + 2 * I915_GTT_PAGE_SIZE <= ggtt->vm.total;
+	     total += 2 * I915_GTT_PAGE_SIZE) {
+		struct i915_vma *vma;
+
+		obj = i915_gem_object_create_internal(ggtt->vm.i915,
+						      2 * PAGE_SIZE);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto out;
+		}
+
+		err = i915_gem_object_pin_pages_unlocked(obj);
+		if (err) {
+			i915_gem_object_put(obj);
+			goto out;
+		}
+
+		list_add(&obj->st_link, &objects);
+
+		vma = i915_vma_instance(obj, &ggtt->vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		err = reserve_gtt_with_resource(vma, total);
+		if (err) {
+			pr_err("i915_gem_gtt_reserve (pass 2) failed at %llu/%llu with err=%d\n",
+			       total, ggtt->vm.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+		if (vma->node.start != total ||
+		    vma->node.size != 2*I915_GTT_PAGE_SIZE) {
+			pr_err("i915_gem_gtt_reserve (pass 2) placement failed, found (%llx + %llx), expected (%llx + %llx)\n",
+			       vma->node.start, vma->node.size,
+			       total, 2*I915_GTT_PAGE_SIZE);
+			err = -EINVAL;
+			goto out;
+		}
+	}
+
+	/* And then try at random */
+	list_for_each_entry_safe(obj, on, &objects, st_link) {
+		struct i915_vma *vma;
+		u64 offset;
+
+		vma = i915_vma_instance(obj, &ggtt->vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		err = i915_vma_unbind_unlocked(vma);
+		if (err) {
+			pr_err("i915_vma_unbind failed with err=%d!\n", err);
+			goto out;
+		}
+
+		offset = igt_random_offset(&prng,
+					   0, ggtt->vm.total,
+					   2 * I915_GTT_PAGE_SIZE,
+					   I915_GTT_MIN_ALIGNMENT);
+
+		err = reserve_gtt_with_resource(vma, offset);
+		if (err) {
+			pr_err("i915_gem_gtt_reserve (pass 3) failed at %llu/%llu with err=%d\n",
+			       total, ggtt->vm.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+		if (vma->node.start != offset ||
+		    vma->node.size != 2*I915_GTT_PAGE_SIZE) {
+			pr_err("i915_gem_gtt_reserve (pass 3) placement failed, found (%llx + %llx), expected (%llx + %llx)\n",
+			       vma->node.start, vma->node.size,
+			       offset, 2*I915_GTT_PAGE_SIZE);
+			err = -EINVAL;
+			goto out;
+		}
+	}
+
+out:
+	list_for_each_entry_safe(obj, on, &objects, st_link) {
+		i915_gem_object_unpin_pages(obj);
+		i915_gem_object_put(obj);
+	}
+	return err;
+}
+
+static int insert_gtt_with_resource(struct i915_vma *vma)
+{
+	struct i915_address_space *vm = vma->vm;
+	struct i915_vma_resource *vma_res;
+	struct drm_i915_gem_object *obj = vma->obj;
+	int err;
+
+	vma_res = i915_vma_resource_alloc();
+	if (IS_ERR(vma_res))
+		return PTR_ERR(vma_res);
+
+	mutex_lock(&vm->mutex);
+	err = i915_gem_gtt_insert(vm, NULL, &vma->node, obj->base.size, 0,
+				  obj->cache_level, 0, vm->total, 0);
+	if (!err) {
+		i915_vma_resource_init_from_vma(vma_res, vma);
+		vma->resource = vma_res;
+	} else {
+		kfree(vma_res);
+	}
+	mutex_unlock(&vm->mutex);
+
+	return err;
+}
+
+static int igt_gtt_insert(void *arg)
+{
+	struct i915_ggtt *ggtt = arg;
+	struct drm_i915_gem_object *obj, *on;
+	struct drm_mm_node tmp = {};
+	const struct invalid_insert {
+		u64 size;
+		u64 alignment;
+		u64 start, end;
+	} invalid_insert[] = {
+		{
+			ggtt->vm.total + I915_GTT_PAGE_SIZE, 0,
+			0, ggtt->vm.total,
+		},
+		{
+			2*I915_GTT_PAGE_SIZE, 0,
+			0, I915_GTT_PAGE_SIZE,
+		},
+		{
+			-(u64)I915_GTT_PAGE_SIZE, 0,
+			0, 4*I915_GTT_PAGE_SIZE,
+		},
+		{
+			-(u64)2*I915_GTT_PAGE_SIZE, 2*I915_GTT_PAGE_SIZE,
+			0, 4*I915_GTT_PAGE_SIZE,
+		},
+		{
+			I915_GTT_PAGE_SIZE, I915_GTT_MIN_ALIGNMENT << 1,
+			I915_GTT_MIN_ALIGNMENT, I915_GTT_MIN_ALIGNMENT << 1,
+		},
+		{}
+	}, *ii;
+	LIST_HEAD(objects);
+	u64 total;
+	int err = -ENODEV;
+
+	/* i915_gem_gtt_insert() tries to allocate some free space in the GTT
+	 * to the node, evicting if required.
+	 */
+
+	/* Check a couple of obviously invalid requests */
+	for (ii = invalid_insert; ii->size; ii++) {
+		mutex_lock(&ggtt->vm.mutex);
+		err = i915_gem_gtt_insert(&ggtt->vm, NULL, &tmp,
+					  ii->size, ii->alignment,
+					  I915_COLOR_UNEVICTABLE,
+					  ii->start, ii->end,
+					  0);
+		mutex_unlock(&ggtt->vm.mutex);
+		if (err != -ENOSPC) {
+			pr_err("Invalid i915_gem_gtt_insert(.size=%llx, .alignment=%llx, .start=%llx, .end=%llx) succeeded (err=%d)\n",
+			       ii->size, ii->alignment, ii->start, ii->end,
+			       err);
+			return -EINVAL;
+		}
+	}
+
+	/* Start by filling the GGTT */
+	for (total = 0;
+	     total + I915_GTT_PAGE_SIZE <= ggtt->vm.total;
+	     total += I915_GTT_PAGE_SIZE) {
+		struct i915_vma *vma;
+
+		obj = i915_gem_object_create_internal(ggtt->vm.i915,
+						      I915_GTT_PAGE_SIZE);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto out;
+		}
+
+		err = i915_gem_object_pin_pages_unlocked(obj);
+		if (err) {
+			i915_gem_object_put(obj);
+			goto out;
+		}
+
+		list_add(&obj->st_link, &objects);
+
+		vma = i915_vma_instance(obj, &ggtt->vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		err = insert_gtt_with_resource(vma);
+		if (err == -ENOSPC) {
+			/* maxed out the GGTT space */
+			i915_gem_object_put(obj);
+			break;
+		}
+		if (err) {
+			pr_err("i915_gem_gtt_insert (pass 1) failed at %llu/%llu with err=%d\n",
+			       total, ggtt->vm.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+		__i915_vma_pin(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	}
+
+	list_for_each_entry(obj, &objects, st_link) {
+		struct i915_vma *vma;
+
+		vma = i915_vma_instance(obj, &ggtt->vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		if (!drm_mm_node_allocated(&vma->node)) {
+			pr_err("VMA was unexpectedly evicted!\n");
+			err = -EINVAL;
+			goto out;
+		}
+
+		__i915_vma_unpin(vma);
+	}
+
+	/* If we then reinsert, we should find the same hole */
+	list_for_each_entry_safe(obj, on, &objects, st_link) {
+		struct i915_vma *vma;
+		u64 offset;
+
+		vma = i915_vma_instance(obj, &ggtt->vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+		offset = vma->node.start;
+
+		err = i915_vma_unbind_unlocked(vma);
+		if (err) {
+			pr_err("i915_vma_unbind failed with err=%d!\n", err);
+			goto out;
+		}
+
+		err = insert_gtt_with_resource(vma);
+		if (err) {
+			pr_err("i915_gem_gtt_insert (pass 2) failed at %llu/%llu with err=%d\n",
+			       total, ggtt->vm.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+		if (vma->node.start != offset) {
+			pr_err("i915_gem_gtt_insert did not return node to its previous location (the only hole), expected address %llx, found %llx\n",
+			       offset, vma->node.start);
+			err = -EINVAL;
+			goto out;
+		}
+	}
+
+	/* And then force evictions */
+	for (total = 0;
+	     total + 2 * I915_GTT_PAGE_SIZE <= ggtt->vm.total;
+	     total += 2 * I915_GTT_PAGE_SIZE) {
+		struct i915_vma *vma;
+
+		obj = i915_gem_object_create_internal(ggtt->vm.i915,
+						      2 * I915_GTT_PAGE_SIZE);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto out;
+		}
+
+		err = i915_gem_object_pin_pages_unlocked(obj);
+		if (err) {
+			i915_gem_object_put(obj);
+			goto out;
+		}
+
+		list_add(&obj->st_link, &objects);
+
+		vma = i915_vma_instance(obj, &ggtt->vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		err = insert_gtt_with_resource(vma);
+		if (err) {
+			pr_err("i915_gem_gtt_insert (pass 3) failed at %llu/%llu with err=%d\n",
+			       total, ggtt->vm.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	}
+
+out:
+	list_for_each_entry_safe(obj, on, &objects, st_link) {
+		i915_gem_object_unpin_pages(obj);
+		i915_gem_object_put(obj);
+	}
+	return err;
+}
+
+int i915_gem_gtt_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_mock_drunk),
+		SUBTEST(igt_mock_walk),
+		SUBTEST(igt_mock_pot),
+		SUBTEST(igt_mock_fill),
+		SUBTEST(igt_gtt_reserve),
+		SUBTEST(igt_gtt_insert),
+	};
+	struct drm_i915_private *i915;
+	struct intel_gt *gt;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	/* allocate the ggtt */
+	err = intel_gt_assign_ggtt(to_gt(i915));
+	if (err)
+		goto out_put;
+
+	gt = to_gt(i915);
+
+	mock_init_ggtt(gt);
+
+	err = i915_subtests(tests, gt->ggtt);
+
+	mock_device_flush(i915);
+	i915_gem_drain_freed_objects(i915);
+	mock_fini_ggtt(gt->ggtt);
+
+out_put:
+	mock_destroy_device(i915);
+	return err;
+}
+
+static int context_sync(struct intel_context *ce)
+{
+	struct i915_request *rq;
+	long timeout;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	timeout = i915_request_wait(rq, 0, HZ / 5);
+	i915_request_put(rq);
+
+	return timeout < 0 ? -EIO : 0;
+}
+
+static struct i915_request *
+submit_batch(struct intel_context *ce, u64 addr)
+{
+	struct i915_request *rq;
+	int err;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return rq;
+
+	err = 0;
+	if (rq->engine->emit_init_breadcrumb) /* detect a hang */
+		err = rq->engine->emit_init_breadcrumb(rq);
+	if (err == 0)
+		err = rq->engine->emit_bb_start(rq, addr, 0, 0);
+
+	if (err == 0)
+		i915_request_get(rq);
+	i915_request_add(rq);
+
+	return err ? ERR_PTR(err) : rq;
+}
+
+static u32 *spinner(u32 *batch, int i)
+{
+	return batch + i * 64 / sizeof(*batch) + 4;
+}
+
+static void end_spin(u32 *batch, int i)
+{
+	*spinner(batch, i) = MI_BATCH_BUFFER_END;
+	wmb();
+}
+
+static int igt_cs_tlb(void *arg)
+{
+	const unsigned int count = PAGE_SIZE / 64;
+	const unsigned int chunk_size = count * PAGE_SIZE;
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *bbe, *act, *out;
+	struct i915_gem_engines_iter it;
+	struct i915_address_space *vm;
+	struct i915_gem_context *ctx;
+	struct intel_context *ce;
+	struct i915_vma *vma;
+	I915_RND_STATE(prng);
+	struct file *file;
+	unsigned int i;
+	u32 *result;
+	u32 *batch;
+	int err = 0;
+
+	/*
+	 * Our mission here is to fool the hardware to execute something
+	 * from scratch as it has not seen the batch move (due to missing
+	 * the TLB invalidate).
+	 */
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	ctx = live_context(i915, file);
+	if (IS_ERR(ctx)) {
+		err = PTR_ERR(ctx);
+		goto out_unlock;
+	}
+
+	vm = i915_gem_context_get_eb_vm(ctx);
+	if (i915_is_ggtt(vm))
+		goto out_vm;
+
+	/* Create two pages; dummy we prefill the TLB, and intended */
+	bbe = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(bbe)) {
+		err = PTR_ERR(bbe);
+		goto out_vm;
+	}
+
+	batch = i915_gem_object_pin_map_unlocked(bbe, I915_MAP_WC);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		goto out_put_bbe;
+	}
+	memset32(batch, MI_BATCH_BUFFER_END, PAGE_SIZE / sizeof(u32));
+	i915_gem_object_flush_map(bbe);
+	i915_gem_object_unpin_map(bbe);
+
+	act = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(act)) {
+		err = PTR_ERR(act);
+		goto out_put_bbe;
+	}
+
+	/* Track the execution of each request by writing into different slot */
+	batch = i915_gem_object_pin_map_unlocked(act, I915_MAP_WC);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		goto out_put_act;
+	}
+	for (i = 0; i < count; i++) {
+		u32 *cs = batch + i * 64 / sizeof(*cs);
+		u64 addr = (vm->total - PAGE_SIZE) + i * sizeof(u32);
+
+		GEM_BUG_ON(GRAPHICS_VER(i915) < 6);
+		cs[0] = MI_STORE_DWORD_IMM_GEN4;
+		if (GRAPHICS_VER(i915) >= 8) {
+			cs[1] = lower_32_bits(addr);
+			cs[2] = upper_32_bits(addr);
+			cs[3] = i;
+			cs[4] = MI_NOOP;
+			cs[5] = MI_BATCH_BUFFER_START_GEN8;
+		} else {
+			cs[1] = 0;
+			cs[2] = lower_32_bits(addr);
+			cs[3] = i;
+			cs[4] = MI_NOOP;
+			cs[5] = MI_BATCH_BUFFER_START;
+		}
+	}
+
+	out = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(out)) {
+		err = PTR_ERR(out);
+		goto out_put_batch;
+	}
+	i915_gem_object_set_cache_coherency(out, I915_CACHING_CACHED);
+
+	vma = i915_vma_instance(out, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto out_put_out;
+	}
+
+	err = i915_vma_pin(vma, 0, 0,
+			   PIN_USER |
+			   PIN_OFFSET_FIXED |
+			   (vm->total - PAGE_SIZE));
+	if (err)
+		goto out_put_out;
+	GEM_BUG_ON(vma->node.start != vm->total - PAGE_SIZE);
+
+	result = i915_gem_object_pin_map_unlocked(out, I915_MAP_WB);
+	if (IS_ERR(result)) {
+		err = PTR_ERR(result);
+		goto out_put_out;
+	}
+
+	for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it) {
+		IGT_TIMEOUT(end_time);
+		unsigned long pass = 0;
+
+		if (!intel_engine_can_store_dword(ce->engine))
+			continue;
+
+		while (!__igt_timeout(end_time, NULL)) {
+			struct i915_vm_pt_stash stash = {};
+			struct i915_request *rq;
+			struct i915_gem_ww_ctx ww;
+			struct i915_vma_resource *vma_res;
+			u64 offset;
+
+			offset = igt_random_offset(&prng,
+						   0, vm->total - PAGE_SIZE,
+						   chunk_size, PAGE_SIZE);
+
+			memset32(result, STACK_MAGIC, PAGE_SIZE / sizeof(u32));
+
+			vma = i915_vma_instance(bbe, vm, NULL);
+			if (IS_ERR(vma)) {
+				err = PTR_ERR(vma);
+				goto end;
+			}
+
+			i915_gem_object_lock(bbe, NULL);
+			err = i915_vma_get_pages(vma);
+			i915_gem_object_unlock(bbe);
+			if (err)
+				goto end;
+
+			vma_res = i915_vma_resource_alloc();
+			if (IS_ERR(vma_res)) {
+				i915_vma_put_pages(vma);
+				err = PTR_ERR(vma_res);
+				goto end;
+			}
+
+			i915_gem_ww_ctx_init(&ww, false);
+retry:
+			err = i915_vm_lock_objects(vm, &ww);
+			if (err)
+				goto end_ww;
+
+			err = i915_vm_alloc_pt_stash(vm, &stash, chunk_size);
+			if (err)
+				goto end_ww;
+
+			err = i915_vm_map_pt_stash(vm, &stash);
+			if (!err)
+				vm->allocate_va_range(vm, &stash, offset, chunk_size);
+			i915_vm_free_pt_stash(vm, &stash);
+end_ww:
+			if (err == -EDEADLK) {
+				err = i915_gem_ww_ctx_backoff(&ww);
+				if (!err)
+					goto retry;
+			}
+			i915_gem_ww_ctx_fini(&ww);
+			if (err) {
+				kfree(vma_res);
+				goto end;
+			}
+
+			i915_vma_resource_init_from_vma(vma_res, vma);
+			/* Prime the TLB with the dummy pages */
+			for (i = 0; i < count; i++) {
+				vma_res->start = offset + i * PAGE_SIZE;
+				vm->insert_entries(vm, vma_res, I915_CACHE_NONE,
+						   0);
+
+				rq = submit_batch(ce, vma_res->start);
+				if (IS_ERR(rq)) {
+					err = PTR_ERR(rq);
+					i915_vma_resource_fini(vma_res);
+					kfree(vma_res);
+					goto end;
+				}
+				i915_request_put(rq);
+			}
+			i915_vma_resource_fini(vma_res);
+			i915_vma_put_pages(vma);
+
+			err = context_sync(ce);
+			if (err) {
+				pr_err("%s: dummy setup timed out\n",
+				       ce->engine->name);
+				kfree(vma_res);
+				goto end;
+			}
+
+			vma = i915_vma_instance(act, vm, NULL);
+			if (IS_ERR(vma)) {
+				kfree(vma_res);
+				err = PTR_ERR(vma);
+				goto end;
+			}
+
+			i915_gem_object_lock(act, NULL);
+			err = i915_vma_get_pages(vma);
+			i915_gem_object_unlock(act);
+			if (err) {
+				kfree(vma_res);
+				goto end;
+			}
+
+			i915_vma_resource_init_from_vma(vma_res, vma);
+			/* Replace the TLB with target batches */
+			for (i = 0; i < count; i++) {
+				struct i915_request *rq;
+				u32 *cs = batch + i * 64 / sizeof(*cs);
+				u64 addr;
+
+				vma_res->start = offset + i * PAGE_SIZE;
+				vm->insert_entries(vm, vma_res, I915_CACHE_NONE, 0);
+
+				addr = vma_res->start + i * 64;
+				cs[4] = MI_NOOP;
+				cs[6] = lower_32_bits(addr);
+				cs[7] = upper_32_bits(addr);
+				wmb();
+
+				rq = submit_batch(ce, addr);
+				if (IS_ERR(rq)) {
+					err = PTR_ERR(rq);
+					i915_vma_resource_fini(vma_res);
+					kfree(vma_res);
+					goto end;
+				}
+
+				/* Wait until the context chain has started */
+				if (i == 0) {
+					while (READ_ONCE(result[i]) &&
+					       !i915_request_completed(rq))
+						cond_resched();
+				} else {
+					end_spin(batch, i - 1);
+				}
+
+				i915_request_put(rq);
+			}
+			end_spin(batch, count - 1);
+
+			i915_vma_resource_fini(vma_res);
+			kfree(vma_res);
+			i915_vma_put_pages(vma);
+
+			err = context_sync(ce);
+			if (err) {
+				pr_err("%s: writes timed out\n",
+				       ce->engine->name);
+				goto end;
+			}
+
+			for (i = 0; i < count; i++) {
+				if (result[i] != i) {
+					pr_err("%s: Write lost on pass %lu, at offset %llx, index %d, found %x, expected %x\n",
+					       ce->engine->name, pass,
+					       offset, i, result[i], i);
+					err = -EINVAL;
+					goto end;
+				}
+			}
+
+			vm->clear_range(vm, offset, chunk_size);
+			pass++;
+		}
+	}
+end:
+	if (igt_flush_test(i915))
+		err = -EIO;
+	i915_gem_context_unlock_engines(ctx);
+	i915_gem_object_unpin_map(out);
+out_put_out:
+	i915_gem_object_put(out);
+out_put_batch:
+	i915_gem_object_unpin_map(act);
+out_put_act:
+	i915_gem_object_put(act);
+out_put_bbe:
+	i915_gem_object_put(bbe);
+out_vm:
+	i915_vm_put(vm);
+out_unlock:
+	fput(file);
+	return err;
+}
+
+int i915_gem_gtt_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_ppgtt_alloc),
+		SUBTEST(igt_ppgtt_lowlevel),
+		SUBTEST(igt_ppgtt_drunk),
+		SUBTEST(igt_ppgtt_walk),
+		SUBTEST(igt_ppgtt_pot),
+		SUBTEST(igt_ppgtt_fill),
+		SUBTEST(igt_ppgtt_shrink),
+		SUBTEST(igt_ppgtt_shrink_boom),
+		SUBTEST(igt_ppgtt_misaligned_pin),
+		SUBTEST(igt_ggtt_lowlevel),
+		SUBTEST(igt_ggtt_drunk),
+		SUBTEST(igt_ggtt_walk),
+		SUBTEST(igt_ggtt_pot),
+		SUBTEST(igt_ggtt_fill),
+		SUBTEST(igt_ggtt_page),
+		SUBTEST(igt_ggtt_misaligned_pin),
+		SUBTEST(igt_cs_tlb),
+	};
+
+	GEM_BUG_ON(offset_in_page(to_gt(i915)->ggtt->vm.total));
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_live_selftests.h b/drivers/gpu/drm/i915/selftests/i915_live_selftests.h
new file mode 100644
index 000000000..aaf8a380e
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_live_selftests.h
@@ -0,0 +1,54 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef selftest
+#define selftest(x, y)
+#endif
+
+/*
+ * List each unit test as selftest(name, function)
+ *
+ * The name is used as both an enum and expanded as subtest__name to create
+ * a module parameter. It must be unique and legal for a C identifier.
+ *
+ * The function should be of type int function(void). It may be conditionally
+ * compiled using #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST).
+ *
+ * Tests are executed in order by igt/i915_selftest
+ */
+selftest(sanitycheck, i915_live_sanitycheck) /* keep first (igt selfcheck) */
+selftest(uncore, intel_uncore_live_selftests)
+selftest(workarounds, intel_workarounds_live_selftests)
+selftest(gt_engines, intel_engine_live_selftests)
+selftest(gt_timelines, intel_timeline_live_selftests)
+selftest(gt_contexts, intel_context_live_selftests)
+selftest(gt_lrc, intel_lrc_live_selftests)
+selftest(gt_mocs, intel_mocs_live_selftests)
+selftest(gt_pm, intel_gt_pm_live_selftests)
+selftest(gt_heartbeat, intel_heartbeat_live_selftests)
+selftest(requests, i915_request_live_selftests)
+selftest(migrate, intel_migrate_live_selftests)
+selftest(active, i915_active_live_selftests)
+selftest(objects, i915_gem_object_live_selftests)
+selftest(mman, i915_gem_mman_live_selftests)
+selftest(dmabuf, i915_gem_dmabuf_live_selftests)
+selftest(vma, i915_vma_live_selftests)
+selftest(coherency, i915_gem_coherency_live_selftests)
+selftest(gtt, i915_gem_gtt_live_selftests)
+selftest(gem, i915_gem_live_selftests)
+selftest(evict, i915_gem_evict_live_selftests)
+selftest(hugepages, i915_gem_huge_page_live_selftests)
+selftest(gem_contexts, i915_gem_context_live_selftests)
+selftest(client, i915_gem_client_blt_live_selftests)
+selftest(gem_migrate, i915_gem_migrate_live_selftests)
+selftest(reset, intel_reset_live_selftests)
+selftest(memory_region, intel_memory_region_live_selftests)
+selftest(hangcheck, intel_hangcheck_live_selftests)
+selftest(execlists, intel_execlists_live_selftests)
+selftest(ring_submission, intel_ring_submission_live_selftests)
+selftest(perf, i915_perf_live_selftests)
+selftest(slpc, intel_slpc_live_selftests)
+selftest(guc, intel_guc_live_selftests)
+selftest(guc_multi_lrc, intel_guc_multi_lrc_live_selftests)
+selftest(guc_hang, intel_guc_hang_check)
+/* Here be dragons: keep last to run last! */
+selftest(late_gt_pm, intel_gt_pm_late_selftests)
diff --git a/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h b/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h
new file mode 100644
index 000000000..0c22e0fc9
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h
@@ -0,0 +1,35 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef selftest
+#define selftest(x, y)
+#endif
+
+/*
+ * List each unit test as selftest(name, function)
+ *
+ * The name is used as both an enum and expanded as subtest__name to create
+ * a module parameter. It must be unique and legal for a C identifier.
+ *
+ * The function should be of type int function(void). It may be conditionally
+ * compiled using #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST).
+ *
+ * Tests are executed in order by igt/i915_selftest
+ */
+selftest(sanitycheck, i915_mock_sanitycheck) /* keep first (igt selfcheck) */
+selftest(shmem, shmem_utils_mock_selftests)
+selftest(fence, i915_sw_fence_mock_selftests)
+selftest(scatterlist, scatterlist_mock_selftests)
+selftest(syncmap, i915_syncmap_mock_selftests)
+selftest(uncore, intel_uncore_mock_selftests)
+selftest(ring, intel_ring_mock_selftests)
+selftest(engine, intel_engine_cs_mock_selftests)
+selftest(timelines, intel_timeline_mock_selftests)
+selftest(requests, i915_request_mock_selftests)
+selftest(objects, i915_gem_object_mock_selftests)
+selftest(phys, i915_gem_phys_mock_selftests)
+selftest(dmabuf, i915_gem_dmabuf_mock_selftests)
+selftest(vma, i915_vma_mock_selftests)
+selftest(evict, i915_gem_evict_mock_selftests)
+selftest(gtt, i915_gem_gtt_mock_selftests)
+selftest(hugepages, i915_gem_huge_page_mock_selftests)
+selftest(memory_region, intel_memory_region_mock_selftests)
diff --git a/drivers/gpu/drm/i915/selftests/i915_perf.c b/drivers/gpu/drm/i915/selftests/i915_perf.c
new file mode 100644
index 000000000..429c6d73b
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_perf.c
@@ -0,0 +1,439 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/kref.h>
+
+#include "gem/i915_gem_pm.h"
+#include "gt/intel_gt.h"
+
+#include "i915_selftest.h"
+
+#include "igt_flush_test.h"
+#include "lib_sw_fence.h"
+
+#define TEST_OA_CONFIG_UUID "12345678-1234-1234-1234-1234567890ab"
+
+static int
+alloc_empty_config(struct i915_perf *perf)
+{
+	struct i915_oa_config *oa_config;
+
+	oa_config = kzalloc(sizeof(*oa_config), GFP_KERNEL);
+	if (!oa_config)
+		return -ENOMEM;
+
+	oa_config->perf = perf;
+	kref_init(&oa_config->ref);
+
+	strlcpy(oa_config->uuid, TEST_OA_CONFIG_UUID, sizeof(oa_config->uuid));
+
+	mutex_lock(&perf->metrics_lock);
+
+	oa_config->id = idr_alloc(&perf->metrics_idr, oa_config, 2, 0, GFP_KERNEL);
+	if (oa_config->id < 0)  {
+		mutex_unlock(&perf->metrics_lock);
+		i915_oa_config_put(oa_config);
+		return -ENOMEM;
+	}
+
+	mutex_unlock(&perf->metrics_lock);
+
+	return 0;
+}
+
+static void
+destroy_empty_config(struct i915_perf *perf)
+{
+	struct i915_oa_config *oa_config = NULL, *tmp;
+	int id;
+
+	mutex_lock(&perf->metrics_lock);
+
+	idr_for_each_entry(&perf->metrics_idr, tmp, id) {
+		if (!strcmp(tmp->uuid, TEST_OA_CONFIG_UUID)) {
+			oa_config = tmp;
+			break;
+		}
+	}
+
+	if (oa_config)
+		idr_remove(&perf->metrics_idr, oa_config->id);
+
+	mutex_unlock(&perf->metrics_lock);
+
+	if (oa_config)
+		i915_oa_config_put(oa_config);
+}
+
+static struct i915_oa_config *
+get_empty_config(struct i915_perf *perf)
+{
+	struct i915_oa_config *oa_config = NULL, *tmp;
+	int id;
+
+	mutex_lock(&perf->metrics_lock);
+
+	idr_for_each_entry(&perf->metrics_idr, tmp, id) {
+		if (!strcmp(tmp->uuid, TEST_OA_CONFIG_UUID)) {
+			oa_config = i915_oa_config_get(tmp);
+			break;
+		}
+	}
+
+	mutex_unlock(&perf->metrics_lock);
+
+	return oa_config;
+}
+
+static struct i915_perf_stream *
+test_stream(struct i915_perf *perf)
+{
+	struct drm_i915_perf_open_param param = {};
+	struct i915_oa_config *oa_config = get_empty_config(perf);
+	struct perf_open_properties props = {
+		.engine = intel_engine_lookup_user(perf->i915,
+						   I915_ENGINE_CLASS_RENDER,
+						   0),
+		.sample_flags = SAMPLE_OA_REPORT,
+		.oa_format = GRAPHICS_VER(perf->i915) == 12 ?
+		I915_OA_FORMAT_A32u40_A4u32_B8_C8 : I915_OA_FORMAT_C4_B8,
+	};
+	struct i915_perf_stream *stream;
+
+	if (!oa_config)
+		return NULL;
+
+	props.metrics_set = oa_config->id;
+
+	stream = kzalloc(sizeof(*stream), GFP_KERNEL);
+	if (!stream) {
+		i915_oa_config_put(oa_config);
+		return NULL;
+	}
+
+	stream->perf = perf;
+
+	mutex_lock(&perf->lock);
+	if (i915_oa_stream_init(stream, &param, &props)) {
+		kfree(stream);
+		stream =  NULL;
+	}
+	mutex_unlock(&perf->lock);
+
+	i915_oa_config_put(oa_config);
+
+	return stream;
+}
+
+static void stream_destroy(struct i915_perf_stream *stream)
+{
+	struct i915_perf *perf = stream->perf;
+
+	mutex_lock(&perf->lock);
+	i915_perf_destroy_locked(stream);
+	mutex_unlock(&perf->lock);
+}
+
+static int live_sanitycheck(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_perf_stream *stream;
+
+	/* Quick check we can create a perf stream */
+
+	stream = test_stream(&i915->perf);
+	if (!stream)
+		return -EINVAL;
+
+	stream_destroy(stream);
+	return 0;
+}
+
+static int write_timestamp(struct i915_request *rq, int slot)
+{
+	u32 *cs;
+	int len;
+
+	cs = intel_ring_begin(rq, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	len = 5;
+	if (GRAPHICS_VER(rq->engine->i915) >= 8)
+		len++;
+
+	*cs++ = GFX_OP_PIPE_CONTROL(len);
+	*cs++ = PIPE_CONTROL_GLOBAL_GTT_IVB |
+		PIPE_CONTROL_STORE_DATA_INDEX |
+		PIPE_CONTROL_WRITE_TIMESTAMP;
+	*cs++ = slot * sizeof(u32);
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = 0;
+
+	intel_ring_advance(rq, cs);
+
+	return 0;
+}
+
+static ktime_t poll_status(struct i915_request *rq, int slot)
+{
+	while (!intel_read_status_page(rq->engine, slot) &&
+	       !i915_request_completed(rq))
+		cpu_relax();
+
+	return ktime_get();
+}
+
+static int live_noa_delay(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_perf_stream *stream;
+	struct i915_request *rq;
+	ktime_t t0, t1;
+	u64 expected;
+	u32 delay;
+	int err;
+	int i;
+
+	/* Check that the GPU delays matches expectations */
+
+	stream = test_stream(&i915->perf);
+	if (!stream)
+		return -ENOMEM;
+
+	expected = atomic64_read(&stream->perf->noa_programming_delay);
+
+	if (stream->engine->class != RENDER_CLASS) {
+		err = -ENODEV;
+		goto out;
+	}
+
+	for (i = 0; i < 4; i++)
+		intel_write_status_page(stream->engine, 0x100 + i, 0);
+
+	rq = intel_engine_create_kernel_request(stream->engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out;
+	}
+
+	if (rq->engine->emit_init_breadcrumb) {
+		err = rq->engine->emit_init_breadcrumb(rq);
+		if (err) {
+			i915_request_add(rq);
+			goto out;
+		}
+	}
+
+	err = write_timestamp(rq, 0x100);
+	if (err) {
+		i915_request_add(rq);
+		goto out;
+	}
+
+	err = rq->engine->emit_bb_start(rq,
+					i915_ggtt_offset(stream->noa_wait), 0,
+					I915_DISPATCH_SECURE);
+	if (err) {
+		i915_request_add(rq);
+		goto out;
+	}
+
+	err = write_timestamp(rq, 0x102);
+	if (err) {
+		i915_request_add(rq);
+		goto out;
+	}
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	preempt_disable();
+	t0 = poll_status(rq, 0x100);
+	t1 = poll_status(rq, 0x102);
+	preempt_enable();
+
+	pr_info("CPU delay: %lluns, expected %lluns\n",
+		ktime_sub(t1, t0), expected);
+
+	delay = intel_read_status_page(stream->engine, 0x102);
+	delay -= intel_read_status_page(stream->engine, 0x100);
+	delay = intel_gt_clock_interval_to_ns(stream->engine->gt, delay);
+	pr_info("GPU delay: %uns, expected %lluns\n",
+		delay, expected);
+
+	if (4 * delay < 3 * expected || 2 * delay > 3 * expected) {
+		pr_err("GPU delay [%uus] outside of expected threshold! [%lluus, %lluus]\n",
+		       delay / 1000,
+		       div_u64(3 * expected, 4000),
+		       div_u64(3 * expected, 2000));
+		err = -EINVAL;
+	}
+
+	i915_request_put(rq);
+out:
+	stream_destroy(stream);
+	return err;
+}
+
+static int live_noa_gpr(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_perf_stream *stream;
+	struct intel_context *ce;
+	struct i915_request *rq;
+	u32 *cs, *store;
+	void *scratch;
+	u32 gpr0;
+	int err;
+	int i;
+
+	/* Check that the delay does not clobber user context state (GPR) */
+
+	stream = test_stream(&i915->perf);
+	if (!stream)
+		return -ENOMEM;
+
+	gpr0 = i915_mmio_reg_offset(GEN8_RING_CS_GPR(stream->engine->mmio_base, 0));
+
+	ce = intel_context_create(stream->engine);
+	if (IS_ERR(ce)) {
+		err = PTR_ERR(ce);
+		goto out;
+	}
+
+	/* Poison the ce->vm so we detect writes not to the GGTT gt->scratch */
+	scratch = __px_vaddr(ce->vm->scratch[0]);
+	memset(scratch, POISON_FREE, PAGE_SIZE);
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_ce;
+	}
+	i915_request_get(rq);
+
+	if (rq->engine->emit_init_breadcrumb) {
+		err = rq->engine->emit_init_breadcrumb(rq);
+		if (err) {
+			i915_request_add(rq);
+			goto out_rq;
+		}
+	}
+
+	/* Fill the 16 qword [32 dword] GPR with a known unlikely value */
+	cs = intel_ring_begin(rq, 2 * 32 + 2);
+	if (IS_ERR(cs)) {
+		err = PTR_ERR(cs);
+		i915_request_add(rq);
+		goto out_rq;
+	}
+
+	*cs++ = MI_LOAD_REGISTER_IMM(32);
+	for (i = 0; i < 32; i++) {
+		*cs++ = gpr0 + i * sizeof(u32);
+		*cs++ = STACK_MAGIC;
+	}
+	*cs++ = MI_NOOP;
+	intel_ring_advance(rq, cs);
+
+	/* Execute the GPU delay */
+	err = rq->engine->emit_bb_start(rq,
+					i915_ggtt_offset(stream->noa_wait), 0,
+					I915_DISPATCH_SECURE);
+	if (err) {
+		i915_request_add(rq);
+		goto out_rq;
+	}
+
+	/* Read the GPR back, using the pinned global HWSP for convenience */
+	store = memset32(rq->engine->status_page.addr + 512, 0, 32);
+	for (i = 0; i < 32; i++) {
+		u32 cmd;
+
+		cs = intel_ring_begin(rq, 4);
+		if (IS_ERR(cs)) {
+			err = PTR_ERR(cs);
+			i915_request_add(rq);
+			goto out_rq;
+		}
+
+		cmd = MI_STORE_REGISTER_MEM;
+		if (GRAPHICS_VER(i915) >= 8)
+			cmd++;
+		cmd |= MI_USE_GGTT;
+
+		*cs++ = cmd;
+		*cs++ = gpr0 + i * sizeof(u32);
+		*cs++ = i915_ggtt_offset(rq->engine->status_page.vma) +
+			offset_in_page(store) +
+			i * sizeof(u32);
+		*cs++ = 0;
+		intel_ring_advance(rq, cs);
+	}
+
+	i915_request_add(rq);
+
+	if (i915_request_wait(rq, I915_WAIT_INTERRUPTIBLE, HZ / 2) < 0) {
+		pr_err("noa_wait timed out\n");
+		intel_gt_set_wedged(stream->engine->gt);
+		err = -EIO;
+		goto out_rq;
+	}
+
+	/* Verify that the GPR contain our expected values */
+	for (i = 0; i < 32; i++) {
+		if (store[i] == STACK_MAGIC)
+			continue;
+
+		pr_err("GPR[%d] lost, found:%08x, expected:%08x!\n",
+		       i, store[i], STACK_MAGIC);
+		err = -EINVAL;
+	}
+
+	/* Verify that the user's scratch page was not used for GPR storage */
+	if (memchr_inv(scratch, POISON_FREE, PAGE_SIZE)) {
+		pr_err("Scratch page overwritten!\n");
+		igt_hexdump(scratch, 4096);
+		err = -EINVAL;
+	}
+
+out_rq:
+	i915_request_put(rq);
+out_ce:
+	intel_context_put(ce);
+out:
+	stream_destroy(stream);
+	return err;
+}
+
+int i915_perf_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_sanitycheck),
+		SUBTEST(live_noa_delay),
+		SUBTEST(live_noa_gpr),
+	};
+	struct i915_perf *perf = &i915->perf;
+	int err;
+
+	if (!perf->metrics_kobj || !perf->ops.enable_metric_set)
+		return 0;
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	err = alloc_empty_config(&i915->perf);
+	if (err)
+		return err;
+
+	err = i915_live_subtests(tests, i915);
+
+	destroy_empty_config(&i915->perf);
+
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_perf_selftests.h b/drivers/gpu/drm/i915/selftests/i915_perf_selftests.h
new file mode 100644
index 000000000..058450d35
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_perf_selftests.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef selftest
+#define selftest(x, y)
+#endif
+
+/*
+ * List each unit test as selftest(name, function)
+ *
+ * The name is used as both an enum and expanded as subtest__name to create
+ * a module parameter. It must be unique and legal for a C identifier.
+ *
+ * The function should be of type int function(void). It may be conditionally
+ * compiled using #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST).
+ *
+ * Tests are executed in order by igt/i915_selftest
+ */
+selftest(engine_cs, intel_engine_cs_perf_selftests)
+selftest(request, i915_request_perf_selftests)
+selftest(migrate, intel_migrate_perf_selftests)
+selftest(region, intel_memory_region_perf_selftests)
diff --git a/drivers/gpu/drm/i915/selftests/i915_random.c b/drivers/gpu/drm/i915/selftests/i915_random.c
new file mode 100644
index 000000000..abdfadcf6
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_random.c
@@ -0,0 +1,109 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/bitops.h>
+#include <linux/kernel.h>
+#include <linux/random.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+
+#include "i915_random.h"
+#include "i915_utils.h"
+
+u64 i915_prandom_u64_state(struct rnd_state *rnd)
+{
+	u64 x;
+
+	x = prandom_u32_state(rnd);
+	x <<= 32;
+	x |= prandom_u32_state(rnd);
+
+	return x;
+}
+
+void i915_prandom_shuffle(void *arr, size_t elsz, size_t count,
+			  struct rnd_state *state)
+{
+	char stack[128];
+
+	if (WARN_ON(elsz > sizeof(stack) || count > U32_MAX))
+		return;
+
+	if (!elsz || !count)
+		return;
+
+	/* Fisher-Yates shuffle courtesy of Knuth */
+	while (--count) {
+		size_t swp;
+
+		swp = i915_prandom_u32_max_state(count + 1, state);
+		if (swp == count)
+			continue;
+
+		memcpy(stack, arr + count * elsz, elsz);
+		memcpy(arr + count * elsz, arr + swp * elsz, elsz);
+		memcpy(arr + swp * elsz, stack, elsz);
+	}
+}
+
+void i915_random_reorder(unsigned int *order, unsigned int count,
+			 struct rnd_state *state)
+{
+	i915_prandom_shuffle(order, sizeof(*order), count, state);
+}
+
+unsigned int *i915_random_order(unsigned int count, struct rnd_state *state)
+{
+	unsigned int *order, i;
+
+	order = kmalloc_array(count, sizeof(*order),
+			      GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN);
+	if (!order)
+		return order;
+
+	for (i = 0; i < count; i++)
+		order[i] = i;
+
+	i915_random_reorder(order, count, state);
+	return order;
+}
+
+u64 igt_random_offset(struct rnd_state *state,
+		      u64 start, u64 end,
+		      u64 len, u64 align)
+{
+	u64 range, addr;
+
+	BUG_ON(range_overflows(start, len, end));
+	BUG_ON(round_up(start, align) > round_down(end - len, align));
+
+	range = round_down(end - len, align) - round_up(start, align);
+	if (range) {
+		addr = i915_prandom_u64_state(state);
+		div64_u64_rem(addr, range, &addr);
+		start += addr;
+	}
+
+	return round_up(start, align);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_random.h b/drivers/gpu/drm/i915/selftests/i915_random.h
new file mode 100644
index 000000000..05364eca2
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_random.h
@@ -0,0 +1,65 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __I915_SELFTESTS_RANDOM_H__
+#define __I915_SELFTESTS_RANDOM_H__
+
+#include <linux/math64.h>
+#include <linux/random.h>
+
+#include "../i915_selftest.h"
+
+#define I915_RND_STATE_INITIALIZER(x) ({				\
+	struct rnd_state state__;					\
+	prandom_seed_state(&state__, (x));				\
+	state__;							\
+})
+
+#define I915_RND_STATE(name__) \
+	struct rnd_state name__ = I915_RND_STATE_INITIALIZER(i915_selftest.random_seed)
+
+#define I915_RND_SUBSTATE(name__, parent__) \
+	struct rnd_state name__ = I915_RND_STATE_INITIALIZER(prandom_u32_state(&(parent__)))
+
+u64 i915_prandom_u64_state(struct rnd_state *rnd);
+
+static inline u32 i915_prandom_u32_max_state(u32 ep_ro, struct rnd_state *state)
+{
+	return upper_32_bits(mul_u32_u32(prandom_u32_state(state), ep_ro));
+}
+
+unsigned int *i915_random_order(unsigned int count,
+				struct rnd_state *state);
+void i915_random_reorder(unsigned int *order,
+			 unsigned int count,
+			 struct rnd_state *state);
+
+void i915_prandom_shuffle(void *arr, size_t elsz, size_t count,
+			  struct rnd_state *state);
+
+u64 igt_random_offset(struct rnd_state *state,
+		      u64 start, u64 end,
+		      u64 len, u64 align);
+
+#endif /* !__I915_SELFTESTS_RANDOM_H__ */
diff --git a/drivers/gpu/drm/i915/selftests/i915_request.c b/drivers/gpu/drm/i915/selftests/i915_request.c
new file mode 100644
index 000000000..a46350c37
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_request.c
@@ -0,0 +1,3288 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/prime_numbers.h>
+#include <linux/pm_qos.h>
+#include <linux/sort.h>
+
+#include "gem/i915_gem_internal.h"
+#include "gem/i915_gem_pm.h"
+#include "gem/selftests/mock_context.h"
+
+#include "gt/intel_engine_heartbeat.h"
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_engine_user.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_clock_utils.h"
+#include "gt/intel_gt_requests.h"
+#include "gt/selftest_engine_heartbeat.h"
+
+#include "i915_random.h"
+#include "i915_selftest.h"
+#include "igt_flush_test.h"
+#include "igt_live_test.h"
+#include "igt_spinner.h"
+#include "lib_sw_fence.h"
+
+#include "mock_drm.h"
+#include "mock_gem_device.h"
+
+static unsigned int num_uabi_engines(struct drm_i915_private *i915)
+{
+	struct intel_engine_cs *engine;
+	unsigned int count;
+
+	count = 0;
+	for_each_uabi_engine(engine, i915)
+		count++;
+
+	return count;
+}
+
+static struct intel_engine_cs *rcs0(struct drm_i915_private *i915)
+{
+	return intel_engine_lookup_user(i915, I915_ENGINE_CLASS_RENDER, 0);
+}
+
+static int igt_add_request(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_request *request;
+
+	/* Basic preliminary test to create a request and let it loose! */
+
+	request = mock_request(rcs0(i915)->kernel_context, HZ / 10);
+	if (!request)
+		return -ENOMEM;
+
+	i915_request_add(request);
+
+	return 0;
+}
+
+static int igt_wait_request(void *arg)
+{
+	const long T = HZ / 4;
+	struct drm_i915_private *i915 = arg;
+	struct i915_request *request;
+	int err = -EINVAL;
+
+	/* Submit a request, then wait upon it */
+
+	request = mock_request(rcs0(i915)->kernel_context, T);
+	if (!request)
+		return -ENOMEM;
+
+	i915_request_get(request);
+
+	if (i915_request_wait(request, 0, 0) != -ETIME) {
+		pr_err("request wait (busy query) succeeded (expected timeout before submit!)\n");
+		goto out_request;
+	}
+
+	if (i915_request_wait(request, 0, T) != -ETIME) {
+		pr_err("request wait succeeded (expected timeout before submit!)\n");
+		goto out_request;
+	}
+
+	if (i915_request_completed(request)) {
+		pr_err("request completed before submit!!\n");
+		goto out_request;
+	}
+
+	i915_request_add(request);
+
+	if (i915_request_wait(request, 0, 0) != -ETIME) {
+		pr_err("request wait (busy query) succeeded (expected timeout after submit!)\n");
+		goto out_request;
+	}
+
+	if (i915_request_completed(request)) {
+		pr_err("request completed immediately!\n");
+		goto out_request;
+	}
+
+	if (i915_request_wait(request, 0, T / 2) != -ETIME) {
+		pr_err("request wait succeeded (expected timeout!)\n");
+		goto out_request;
+	}
+
+	if (i915_request_wait(request, 0, T) == -ETIME) {
+		pr_err("request wait timed out!\n");
+		goto out_request;
+	}
+
+	if (!i915_request_completed(request)) {
+		pr_err("request not complete after waiting!\n");
+		goto out_request;
+	}
+
+	if (i915_request_wait(request, 0, T) == -ETIME) {
+		pr_err("request wait timed out when already complete!\n");
+		goto out_request;
+	}
+
+	err = 0;
+out_request:
+	i915_request_put(request);
+	mock_device_flush(i915);
+	return err;
+}
+
+static int igt_fence_wait(void *arg)
+{
+	const long T = HZ / 4;
+	struct drm_i915_private *i915 = arg;
+	struct i915_request *request;
+	int err = -EINVAL;
+
+	/* Submit a request, treat it as a fence and wait upon it */
+
+	request = mock_request(rcs0(i915)->kernel_context, T);
+	if (!request)
+		return -ENOMEM;
+
+	if (dma_fence_wait_timeout(&request->fence, false, T) != -ETIME) {
+		pr_err("fence wait success before submit (expected timeout)!\n");
+		goto out;
+	}
+
+	i915_request_add(request);
+
+	if (dma_fence_is_signaled(&request->fence)) {
+		pr_err("fence signaled immediately!\n");
+		goto out;
+	}
+
+	if (dma_fence_wait_timeout(&request->fence, false, T / 2) != -ETIME) {
+		pr_err("fence wait success after submit (expected timeout)!\n");
+		goto out;
+	}
+
+	if (dma_fence_wait_timeout(&request->fence, false, T) <= 0) {
+		pr_err("fence wait timed out (expected success)!\n");
+		goto out;
+	}
+
+	if (!dma_fence_is_signaled(&request->fence)) {
+		pr_err("fence unsignaled after waiting!\n");
+		goto out;
+	}
+
+	if (dma_fence_wait_timeout(&request->fence, false, T) <= 0) {
+		pr_err("fence wait timed out when complete (expected success)!\n");
+		goto out;
+	}
+
+	err = 0;
+out:
+	mock_device_flush(i915);
+	return err;
+}
+
+static int igt_request_rewind(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_request *request, *vip;
+	struct i915_gem_context *ctx[2];
+	struct intel_context *ce;
+	int err = -EINVAL;
+
+	ctx[0] = mock_context(i915, "A");
+	if (!ctx[0]) {
+		err = -ENOMEM;
+		goto err_ctx_0;
+	}
+
+	ce = i915_gem_context_get_engine(ctx[0], RCS0);
+	GEM_BUG_ON(IS_ERR(ce));
+	request = mock_request(ce, 2 * HZ);
+	intel_context_put(ce);
+	if (!request) {
+		err = -ENOMEM;
+		goto err_context_0;
+	}
+
+	i915_request_get(request);
+	i915_request_add(request);
+
+	ctx[1] = mock_context(i915, "B");
+	if (!ctx[1]) {
+		err = -ENOMEM;
+		goto err_ctx_1;
+	}
+
+	ce = i915_gem_context_get_engine(ctx[1], RCS0);
+	GEM_BUG_ON(IS_ERR(ce));
+	vip = mock_request(ce, 0);
+	intel_context_put(ce);
+	if (!vip) {
+		err = -ENOMEM;
+		goto err_context_1;
+	}
+
+	/* Simulate preemption by manual reordering */
+	if (!mock_cancel_request(request)) {
+		pr_err("failed to cancel request (already executed)!\n");
+		i915_request_add(vip);
+		goto err_context_1;
+	}
+	i915_request_get(vip);
+	i915_request_add(vip);
+	rcu_read_lock();
+	request->engine->submit_request(request);
+	rcu_read_unlock();
+
+
+	if (i915_request_wait(vip, 0, HZ) == -ETIME) {
+		pr_err("timed out waiting for high priority request\n");
+		goto err;
+	}
+
+	if (i915_request_completed(request)) {
+		pr_err("low priority request already completed\n");
+		goto err;
+	}
+
+	err = 0;
+err:
+	i915_request_put(vip);
+err_context_1:
+	mock_context_close(ctx[1]);
+err_ctx_1:
+	i915_request_put(request);
+err_context_0:
+	mock_context_close(ctx[0]);
+err_ctx_0:
+	mock_device_flush(i915);
+	return err;
+}
+
+struct smoketest {
+	struct intel_engine_cs *engine;
+	struct i915_gem_context **contexts;
+	atomic_long_t num_waits, num_fences;
+	int ncontexts, max_batch;
+	struct i915_request *(*request_alloc)(struct intel_context *ce);
+};
+
+static struct i915_request *
+__mock_request_alloc(struct intel_context *ce)
+{
+	return mock_request(ce, 0);
+}
+
+static struct i915_request *
+__live_request_alloc(struct intel_context *ce)
+{
+	return intel_context_create_request(ce);
+}
+
+struct smoke_thread {
+	struct kthread_worker *worker;
+	struct kthread_work work;
+	struct smoketest *t;
+	bool stop;
+	int result;
+};
+
+static void __igt_breadcrumbs_smoketest(struct kthread_work *work)
+{
+	struct smoke_thread *thread = container_of(work, typeof(*thread), work);
+	struct smoketest *t = thread->t;
+	const unsigned int max_batch = min(t->ncontexts, t->max_batch) - 1;
+	const unsigned int total = 4 * t->ncontexts + 1;
+	unsigned int num_waits = 0, num_fences = 0;
+	struct i915_request **requests;
+	I915_RND_STATE(prng);
+	unsigned int *order;
+	int err = 0;
+
+	/*
+	 * A very simple test to catch the most egregious of list handling bugs.
+	 *
+	 * At its heart, we simply create oodles of requests running across
+	 * multiple kthreads and enable signaling on them, for the sole purpose
+	 * of stressing our breadcrumb handling. The only inspection we do is
+	 * that the fences were marked as signaled.
+	 */
+
+	requests = kcalloc(total, sizeof(*requests), GFP_KERNEL);
+	if (!requests) {
+		thread->result = -ENOMEM;
+		return;
+	}
+
+	order = i915_random_order(total, &prng);
+	if (!order) {
+		err = -ENOMEM;
+		goto out_requests;
+	}
+
+	while (!READ_ONCE(thread->stop)) {
+		struct i915_sw_fence *submit, *wait;
+		unsigned int n, count;
+
+		submit = heap_fence_create(GFP_KERNEL);
+		if (!submit) {
+			err = -ENOMEM;
+			break;
+		}
+
+		wait = heap_fence_create(GFP_KERNEL);
+		if (!wait) {
+			i915_sw_fence_commit(submit);
+			heap_fence_put(submit);
+			err = -ENOMEM;
+			break;
+		}
+
+		i915_random_reorder(order, total, &prng);
+		count = 1 + i915_prandom_u32_max_state(max_batch, &prng);
+
+		for (n = 0; n < count; n++) {
+			struct i915_gem_context *ctx =
+				t->contexts[order[n] % t->ncontexts];
+			struct i915_request *rq;
+			struct intel_context *ce;
+
+			ce = i915_gem_context_get_engine(ctx, t->engine->legacy_idx);
+			GEM_BUG_ON(IS_ERR(ce));
+			rq = t->request_alloc(ce);
+			intel_context_put(ce);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				count = n;
+				break;
+			}
+
+			err = i915_sw_fence_await_sw_fence_gfp(&rq->submit,
+							       submit,
+							       GFP_KERNEL);
+
+			requests[n] = i915_request_get(rq);
+			i915_request_add(rq);
+
+			if (err >= 0)
+				err = i915_sw_fence_await_dma_fence(wait,
+								    &rq->fence,
+								    0,
+								    GFP_KERNEL);
+
+			if (err < 0) {
+				i915_request_put(rq);
+				count = n;
+				break;
+			}
+		}
+
+		i915_sw_fence_commit(submit);
+		i915_sw_fence_commit(wait);
+
+		if (!wait_event_timeout(wait->wait,
+					i915_sw_fence_done(wait),
+					5 * HZ)) {
+			struct i915_request *rq = requests[count - 1];
+
+			pr_err("waiting for %d/%d fences (last %llx:%lld) on %s timed out!\n",
+			       atomic_read(&wait->pending), count,
+			       rq->fence.context, rq->fence.seqno,
+			       t->engine->name);
+			GEM_TRACE_DUMP();
+
+			intel_gt_set_wedged(t->engine->gt);
+			GEM_BUG_ON(!i915_request_completed(rq));
+			i915_sw_fence_wait(wait);
+			err = -EIO;
+		}
+
+		for (n = 0; n < count; n++) {
+			struct i915_request *rq = requests[n];
+
+			if (!test_bit(DMA_FENCE_FLAG_SIGNALED_BIT,
+				      &rq->fence.flags)) {
+				pr_err("%llu:%llu was not signaled!\n",
+				       rq->fence.context, rq->fence.seqno);
+				err = -EINVAL;
+			}
+
+			i915_request_put(rq);
+		}
+
+		heap_fence_put(wait);
+		heap_fence_put(submit);
+
+		if (err < 0)
+			break;
+
+		num_fences += count;
+		num_waits++;
+
+		cond_resched();
+	}
+
+	atomic_long_add(num_fences, &t->num_fences);
+	atomic_long_add(num_waits, &t->num_waits);
+
+	kfree(order);
+out_requests:
+	kfree(requests);
+	thread->result = err;
+}
+
+static int mock_breadcrumbs_smoketest(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct smoketest t = {
+		.engine = rcs0(i915),
+		.ncontexts = 1024,
+		.max_batch = 1024,
+		.request_alloc = __mock_request_alloc
+	};
+	unsigned int ncpus = num_online_cpus();
+	struct smoke_thread *threads;
+	unsigned int n;
+	int ret = 0;
+
+	/*
+	 * Smoketest our breadcrumb/signal handling for requests across multiple
+	 * threads. A very simple test to only catch the most egregious of bugs.
+	 * See __igt_breadcrumbs_smoketest();
+	 */
+
+	threads = kcalloc(ncpus, sizeof(*threads), GFP_KERNEL);
+	if (!threads)
+		return -ENOMEM;
+
+	t.contexts = kcalloc(t.ncontexts, sizeof(*t.contexts), GFP_KERNEL);
+	if (!t.contexts) {
+		ret = -ENOMEM;
+		goto out_threads;
+	}
+
+	for (n = 0; n < t.ncontexts; n++) {
+		t.contexts[n] = mock_context(t.engine->i915, "mock");
+		if (!t.contexts[n]) {
+			ret = -ENOMEM;
+			goto out_contexts;
+		}
+	}
+
+	for (n = 0; n < ncpus; n++) {
+		struct kthread_worker *worker;
+
+		worker = kthread_create_worker(0, "igt/%d", n);
+		if (IS_ERR(worker)) {
+			ret = PTR_ERR(worker);
+			ncpus = n;
+			break;
+		}
+
+		threads[n].worker = worker;
+		threads[n].t = &t;
+		threads[n].stop = false;
+		threads[n].result = 0;
+
+		kthread_init_work(&threads[n].work,
+				  __igt_breadcrumbs_smoketest);
+		kthread_queue_work(worker, &threads[n].work);
+	}
+
+	msleep(jiffies_to_msecs(i915_selftest.timeout_jiffies));
+
+	for (n = 0; n < ncpus; n++) {
+		int err;
+
+		WRITE_ONCE(threads[n].stop, true);
+		kthread_flush_work(&threads[n].work);
+		err = READ_ONCE(threads[n].result);
+		if (err < 0 && !ret)
+			ret = err;
+
+		kthread_destroy_worker(threads[n].worker);
+	}
+	pr_info("Completed %lu waits for %lu fence across %d cpus\n",
+		atomic_long_read(&t.num_waits),
+		atomic_long_read(&t.num_fences),
+		ncpus);
+
+out_contexts:
+	for (n = 0; n < t.ncontexts; n++) {
+		if (!t.contexts[n])
+			break;
+		mock_context_close(t.contexts[n]);
+	}
+	kfree(t.contexts);
+out_threads:
+	kfree(threads);
+	return ret;
+}
+
+int i915_request_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_add_request),
+		SUBTEST(igt_wait_request),
+		SUBTEST(igt_fence_wait),
+		SUBTEST(igt_request_rewind),
+		SUBTEST(mock_breadcrumbs_smoketest),
+	};
+	struct drm_i915_private *i915;
+	intel_wakeref_t wakeref;
+	int err = 0;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	with_intel_runtime_pm(&i915->runtime_pm, wakeref)
+		err = i915_subtests(tests, i915);
+
+	mock_destroy_device(i915);
+
+	return err;
+}
+
+static int live_nop_request(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	struct igt_live_test t;
+	int err = -ENODEV;
+
+	/*
+	 * Submit various sized batches of empty requests, to each engine
+	 * (individually), and wait for the batch to complete. We can check
+	 * the overhead of submitting requests to the hardware.
+	 */
+
+	for_each_uabi_engine(engine, i915) {
+		unsigned long n, prime;
+		IGT_TIMEOUT(end_time);
+		ktime_t times[2] = {};
+
+		err = igt_live_test_begin(&t, i915, __func__, engine->name);
+		if (err)
+			return err;
+
+		intel_engine_pm_get(engine);
+		for_each_prime_number_from(prime, 1, 8192) {
+			struct i915_request *request = NULL;
+
+			times[1] = ktime_get_raw();
+
+			for (n = 0; n < prime; n++) {
+				i915_request_put(request);
+				request = i915_request_create(engine->kernel_context);
+				if (IS_ERR(request))
+					return PTR_ERR(request);
+
+				/*
+				 * This space is left intentionally blank.
+				 *
+				 * We do not actually want to perform any
+				 * action with this request, we just want
+				 * to measure the latency in allocation
+				 * and submission of our breadcrumbs -
+				 * ensuring that the bare request is sufficient
+				 * for the system to work (i.e. proper HEAD
+				 * tracking of the rings, interrupt handling,
+				 * etc). It also gives us the lowest bounds
+				 * for latency.
+				 */
+
+				i915_request_get(request);
+				i915_request_add(request);
+			}
+			i915_request_wait(request, 0, MAX_SCHEDULE_TIMEOUT);
+			i915_request_put(request);
+
+			times[1] = ktime_sub(ktime_get_raw(), times[1]);
+			if (prime == 1)
+				times[0] = times[1];
+
+			if (__igt_timeout(end_time, NULL))
+				break;
+		}
+		intel_engine_pm_put(engine);
+
+		err = igt_live_test_end(&t);
+		if (err)
+			return err;
+
+		pr_info("Request latencies on %s: 1 = %lluns, %lu = %lluns\n",
+			engine->name,
+			ktime_to_ns(times[0]),
+			prime, div64_u64(ktime_to_ns(times[1]), prime));
+	}
+
+	return err;
+}
+
+static int __cancel_inactive(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+	struct igt_spinner spin;
+	struct i915_request *rq;
+	int err = 0;
+
+	if (igt_spinner_init(&spin, engine->gt))
+		return -ENOMEM;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce)) {
+		err = PTR_ERR(ce);
+		goto out_spin;
+	}
+
+	rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_ce;
+	}
+
+	pr_debug("%s: Cancelling inactive request\n", engine->name);
+	i915_request_cancel(rq, -EINTR);
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+		struct drm_printer p = drm_info_printer(engine->i915->drm.dev);
+
+		pr_err("%s: Failed to cancel inactive request\n", engine->name);
+		intel_engine_dump(engine, &p, "%s\n", engine->name);
+		err = -ETIME;
+		goto out_rq;
+	}
+
+	if (rq->fence.error != -EINTR) {
+		pr_err("%s: fence not cancelled (%u)\n",
+		       engine->name, rq->fence.error);
+		err = -EINVAL;
+	}
+
+out_rq:
+	i915_request_put(rq);
+out_ce:
+	intel_context_put(ce);
+out_spin:
+	igt_spinner_fini(&spin);
+	if (err)
+		pr_err("%s: %s error %d\n", __func__, engine->name, err);
+	return err;
+}
+
+static int __cancel_active(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+	struct igt_spinner spin;
+	struct i915_request *rq;
+	int err = 0;
+
+	if (igt_spinner_init(&spin, engine->gt))
+		return -ENOMEM;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce)) {
+		err = PTR_ERR(ce);
+		goto out_spin;
+	}
+
+	rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_ce;
+	}
+
+	pr_debug("%s: Cancelling active request\n", engine->name);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (!igt_wait_for_spinner(&spin, rq)) {
+		struct drm_printer p = drm_info_printer(engine->i915->drm.dev);
+
+		pr_err("Failed to start spinner on %s\n", engine->name);
+		intel_engine_dump(engine, &p, "%s\n", engine->name);
+		err = -ETIME;
+		goto out_rq;
+	}
+	i915_request_cancel(rq, -EINTR);
+
+	if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+		struct drm_printer p = drm_info_printer(engine->i915->drm.dev);
+
+		pr_err("%s: Failed to cancel active request\n", engine->name);
+		intel_engine_dump(engine, &p, "%s\n", engine->name);
+		err = -ETIME;
+		goto out_rq;
+	}
+
+	if (rq->fence.error != -EINTR) {
+		pr_err("%s: fence not cancelled (%u)\n",
+		       engine->name, rq->fence.error);
+		err = -EINVAL;
+	}
+
+out_rq:
+	i915_request_put(rq);
+out_ce:
+	intel_context_put(ce);
+out_spin:
+	igt_spinner_fini(&spin);
+	if (err)
+		pr_err("%s: %s error %d\n", __func__, engine->name, err);
+	return err;
+}
+
+static int __cancel_completed(struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+	struct igt_spinner spin;
+	struct i915_request *rq;
+	int err = 0;
+
+	if (igt_spinner_init(&spin, engine->gt))
+		return -ENOMEM;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce)) {
+		err = PTR_ERR(ce);
+		goto out_spin;
+	}
+
+	rq = igt_spinner_create_request(&spin, ce, MI_ARB_CHECK);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_ce;
+	}
+	igt_spinner_end(&spin);
+	i915_request_get(rq);
+	i915_request_add(rq);
+
+	if (i915_request_wait(rq, 0, HZ / 5) < 0) {
+		err = -ETIME;
+		goto out_rq;
+	}
+
+	pr_debug("%s: Cancelling completed request\n", engine->name);
+	i915_request_cancel(rq, -EINTR);
+	if (rq->fence.error) {
+		pr_err("%s: fence not cancelled (%u)\n",
+		       engine->name, rq->fence.error);
+		err = -EINVAL;
+	}
+
+out_rq:
+	i915_request_put(rq);
+out_ce:
+	intel_context_put(ce);
+out_spin:
+	igt_spinner_fini(&spin);
+	if (err)
+		pr_err("%s: %s error %d\n", __func__, engine->name, err);
+	return err;
+}
+
+/*
+ * Test to prove a non-preemptable request can be cancelled and a subsequent
+ * request on the same context can successfully complete after cancellation.
+ *
+ * Testing methodology is to create a non-preemptible request and submit it,
+ * wait for spinner to start, create a NOP request and submit it, cancel the
+ * spinner, wait for spinner to complete and verify it failed with an error,
+ * finally wait for NOP request to complete verify it succeeded without an
+ * error. Preemption timeout also reduced / restored so test runs in a timely
+ * maner.
+ */
+static int __cancel_reset(struct drm_i915_private *i915,
+			  struct intel_engine_cs *engine)
+{
+	struct intel_context *ce;
+	struct igt_spinner spin;
+	struct i915_request *rq, *nop;
+	unsigned long preempt_timeout_ms;
+	int err = 0;
+
+	if (!CONFIG_DRM_I915_PREEMPT_TIMEOUT ||
+	    !intel_has_reset_engine(engine->gt))
+		return 0;
+
+	preempt_timeout_ms = engine->props.preempt_timeout_ms;
+	engine->props.preempt_timeout_ms = 100;
+
+	if (igt_spinner_init(&spin, engine->gt))
+		goto out_restore;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce)) {
+		err = PTR_ERR(ce);
+		goto out_spin;
+	}
+
+	rq = igt_spinner_create_request(&spin, ce, MI_NOOP);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto out_ce;
+	}
+
+	pr_debug("%s: Cancelling active non-preemptable request\n",
+		 engine->name);
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (!igt_wait_for_spinner(&spin, rq)) {
+		struct drm_printer p = drm_info_printer(engine->i915->drm.dev);
+
+		pr_err("Failed to start spinner on %s\n", engine->name);
+		intel_engine_dump(engine, &p, "%s\n", engine->name);
+		err = -ETIME;
+		goto out_rq;
+	}
+
+	nop = intel_context_create_request(ce);
+	if (IS_ERR(nop))
+		goto out_rq;
+	i915_request_get(nop);
+	i915_request_add(nop);
+
+	i915_request_cancel(rq, -EINTR);
+
+	if (i915_request_wait(rq, 0, HZ) < 0) {
+		struct drm_printer p = drm_info_printer(engine->i915->drm.dev);
+
+		pr_err("%s: Failed to cancel hung request\n", engine->name);
+		intel_engine_dump(engine, &p, "%s\n", engine->name);
+		err = -ETIME;
+		goto out_nop;
+	}
+
+	if (rq->fence.error != -EINTR) {
+		pr_err("%s: fence not cancelled (%u)\n",
+		       engine->name, rq->fence.error);
+		err = -EINVAL;
+		goto out_nop;
+	}
+
+	if (i915_request_wait(nop, 0, HZ) < 0) {
+		struct drm_printer p = drm_info_printer(engine->i915->drm.dev);
+
+		pr_err("%s: Failed to complete nop request\n", engine->name);
+		intel_engine_dump(engine, &p, "%s\n", engine->name);
+		err = -ETIME;
+		goto out_nop;
+	}
+
+	if (nop->fence.error != 0) {
+		pr_err("%s: Nop request errored (%u)\n",
+		       engine->name, nop->fence.error);
+		err = -EINVAL;
+	}
+
+out_nop:
+	i915_request_put(nop);
+out_rq:
+	i915_request_put(rq);
+out_ce:
+	intel_context_put(ce);
+out_spin:
+	igt_spinner_fini(&spin);
+out_restore:
+	engine->props.preempt_timeout_ms = preempt_timeout_ms;
+	if (err)
+		pr_err("%s: %s error %d\n", __func__, engine->name, err);
+	return err;
+}
+
+static int live_cancel_request(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+
+	/*
+	 * Check cancellation of requests. We expect to be able to immediately
+	 * cancel active requests, even if they are currently on the GPU.
+	 */
+
+	for_each_uabi_engine(engine, i915) {
+		struct igt_live_test t;
+		int err, err2;
+
+		if (!intel_engine_has_preemption(engine))
+			continue;
+
+		err = igt_live_test_begin(&t, i915, __func__, engine->name);
+		if (err)
+			return err;
+
+		err = __cancel_inactive(engine);
+		if (err == 0)
+			err = __cancel_active(engine);
+		if (err == 0)
+			err = __cancel_completed(engine);
+
+		err2 = igt_live_test_end(&t);
+		if (err)
+			return err;
+		if (err2)
+			return err2;
+
+		/* Expects reset so call outside of igt_live_test_* */
+		err = __cancel_reset(i915, engine);
+		if (err)
+			return err;
+
+		if (igt_flush_test(i915))
+			return -EIO;
+	}
+
+	return 0;
+}
+
+static struct i915_vma *empty_batch(struct drm_i915_private *i915)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u32 *cmd;
+	int err;
+
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	cmd = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WB);
+	if (IS_ERR(cmd)) {
+		err = PTR_ERR(cmd);
+		goto err;
+	}
+
+	*cmd = MI_BATCH_BUFFER_END;
+
+	__i915_gem_object_flush_map(obj, 0, 64);
+	i915_gem_object_unpin_map(obj);
+
+	intel_gt_chipset_flush(to_gt(i915));
+
+	vma = i915_vma_instance(obj, &to_gt(i915)->ggtt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER | PIN_GLOBAL);
+	if (err)
+		goto err;
+
+	/* Force the wait now to avoid including it in the benchmark */
+	err = i915_vma_sync(vma);
+	if (err)
+		goto err_pin;
+
+	return vma;
+
+err_pin:
+	i915_vma_unpin(vma);
+err:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static struct i915_request *
+empty_request(struct intel_engine_cs *engine,
+	      struct i915_vma *batch)
+{
+	struct i915_request *request;
+	int err;
+
+	request = i915_request_create(engine->kernel_context);
+	if (IS_ERR(request))
+		return request;
+
+	err = engine->emit_bb_start(request,
+				    batch->node.start,
+				    batch->node.size,
+				    I915_DISPATCH_SECURE);
+	if (err)
+		goto out_request;
+
+	i915_request_get(request);
+out_request:
+	i915_request_add(request);
+	return err ? ERR_PTR(err) : request;
+}
+
+static int live_empty_request(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	struct igt_live_test t;
+	struct i915_vma *batch;
+	int err = 0;
+
+	/*
+	 * Submit various sized batches of empty requests, to each engine
+	 * (individually), and wait for the batch to complete. We can check
+	 * the overhead of submitting requests to the hardware.
+	 */
+
+	batch = empty_batch(i915);
+	if (IS_ERR(batch))
+		return PTR_ERR(batch);
+
+	for_each_uabi_engine(engine, i915) {
+		IGT_TIMEOUT(end_time);
+		struct i915_request *request;
+		unsigned long n, prime;
+		ktime_t times[2] = {};
+
+		err = igt_live_test_begin(&t, i915, __func__, engine->name);
+		if (err)
+			goto out_batch;
+
+		intel_engine_pm_get(engine);
+
+		/* Warmup / preload */
+		request = empty_request(engine, batch);
+		if (IS_ERR(request)) {
+			err = PTR_ERR(request);
+			intel_engine_pm_put(engine);
+			goto out_batch;
+		}
+		i915_request_wait(request, 0, MAX_SCHEDULE_TIMEOUT);
+
+		for_each_prime_number_from(prime, 1, 8192) {
+			times[1] = ktime_get_raw();
+
+			for (n = 0; n < prime; n++) {
+				i915_request_put(request);
+				request = empty_request(engine, batch);
+				if (IS_ERR(request)) {
+					err = PTR_ERR(request);
+					intel_engine_pm_put(engine);
+					goto out_batch;
+				}
+			}
+			i915_request_wait(request, 0, MAX_SCHEDULE_TIMEOUT);
+
+			times[1] = ktime_sub(ktime_get_raw(), times[1]);
+			if (prime == 1)
+				times[0] = times[1];
+
+			if (__igt_timeout(end_time, NULL))
+				break;
+		}
+		i915_request_put(request);
+		intel_engine_pm_put(engine);
+
+		err = igt_live_test_end(&t);
+		if (err)
+			goto out_batch;
+
+		pr_info("Batch latencies on %s: 1 = %lluns, %lu = %lluns\n",
+			engine->name,
+			ktime_to_ns(times[0]),
+			prime, div64_u64(ktime_to_ns(times[1]), prime));
+	}
+
+out_batch:
+	i915_vma_unpin(batch);
+	i915_vma_put(batch);
+	return err;
+}
+
+static struct i915_vma *recursive_batch(struct drm_i915_private *i915)
+{
+	struct drm_i915_gem_object *obj;
+	const int ver = GRAPHICS_VER(i915);
+	struct i915_vma *vma;
+	u32 *cmd;
+	int err;
+
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, to_gt(i915)->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto err;
+
+	cmd = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(cmd)) {
+		err = PTR_ERR(cmd);
+		goto err;
+	}
+
+	if (ver >= 8) {
+		*cmd++ = MI_BATCH_BUFFER_START | 1 << 8 | 1;
+		*cmd++ = lower_32_bits(vma->node.start);
+		*cmd++ = upper_32_bits(vma->node.start);
+	} else if (ver >= 6) {
+		*cmd++ = MI_BATCH_BUFFER_START | 1 << 8;
+		*cmd++ = lower_32_bits(vma->node.start);
+	} else {
+		*cmd++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT;
+		*cmd++ = lower_32_bits(vma->node.start);
+	}
+	*cmd++ = MI_BATCH_BUFFER_END; /* terminate early in case of error */
+
+	__i915_gem_object_flush_map(obj, 0, 64);
+	i915_gem_object_unpin_map(obj);
+
+	intel_gt_chipset_flush(to_gt(i915));
+
+	return vma;
+
+err:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static int recursive_batch_resolve(struct i915_vma *batch)
+{
+	u32 *cmd;
+
+	cmd = i915_gem_object_pin_map_unlocked(batch->obj, I915_MAP_WC);
+	if (IS_ERR(cmd))
+		return PTR_ERR(cmd);
+
+	*cmd = MI_BATCH_BUFFER_END;
+
+	__i915_gem_object_flush_map(batch->obj, 0, sizeof(*cmd));
+	i915_gem_object_unpin_map(batch->obj);
+
+	intel_gt_chipset_flush(batch->vm->gt);
+
+	return 0;
+}
+
+static int live_all_engines(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	const unsigned int nengines = num_uabi_engines(i915);
+	struct intel_engine_cs *engine;
+	struct i915_request **request;
+	struct igt_live_test t;
+	struct i915_vma *batch;
+	unsigned int idx;
+	int err;
+
+	/*
+	 * Check we can submit requests to all engines simultaneously. We
+	 * send a recursive batch to each engine - checking that we don't
+	 * block doing so, and that they don't complete too soon.
+	 */
+
+	request = kcalloc(nengines, sizeof(*request), GFP_KERNEL);
+	if (!request)
+		return -ENOMEM;
+
+	err = igt_live_test_begin(&t, i915, __func__, "");
+	if (err)
+		goto out_free;
+
+	batch = recursive_batch(i915);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		pr_err("%s: Unable to create batch, err=%d\n", __func__, err);
+		goto out_free;
+	}
+
+	i915_vma_lock(batch);
+
+	idx = 0;
+	for_each_uabi_engine(engine, i915) {
+		request[idx] = intel_engine_create_kernel_request(engine);
+		if (IS_ERR(request[idx])) {
+			err = PTR_ERR(request[idx]);
+			pr_err("%s: Request allocation failed with err=%d\n",
+			       __func__, err);
+			goto out_request;
+		}
+
+		err = i915_request_await_object(request[idx], batch->obj, 0);
+		if (err == 0)
+			err = i915_vma_move_to_active(batch, request[idx], 0);
+		GEM_BUG_ON(err);
+
+		err = engine->emit_bb_start(request[idx],
+					    batch->node.start,
+					    batch->node.size,
+					    0);
+		GEM_BUG_ON(err);
+		request[idx]->batch = batch;
+
+		i915_request_get(request[idx]);
+		i915_request_add(request[idx]);
+		idx++;
+	}
+
+	i915_vma_unlock(batch);
+
+	idx = 0;
+	for_each_uabi_engine(engine, i915) {
+		if (i915_request_completed(request[idx])) {
+			pr_err("%s(%s): request completed too early!\n",
+			       __func__, engine->name);
+			err = -EINVAL;
+			goto out_request;
+		}
+		idx++;
+	}
+
+	err = recursive_batch_resolve(batch);
+	if (err) {
+		pr_err("%s: failed to resolve batch, err=%d\n", __func__, err);
+		goto out_request;
+	}
+
+	idx = 0;
+	for_each_uabi_engine(engine, i915) {
+		long timeout;
+
+		timeout = i915_request_wait(request[idx], 0,
+					    MAX_SCHEDULE_TIMEOUT);
+		if (timeout < 0) {
+			err = timeout;
+			pr_err("%s: error waiting for request on %s, err=%d\n",
+			       __func__, engine->name, err);
+			goto out_request;
+		}
+
+		GEM_BUG_ON(!i915_request_completed(request[idx]));
+		i915_request_put(request[idx]);
+		request[idx] = NULL;
+		idx++;
+	}
+
+	err = igt_live_test_end(&t);
+
+out_request:
+	idx = 0;
+	for_each_uabi_engine(engine, i915) {
+		if (request[idx])
+			i915_request_put(request[idx]);
+		idx++;
+	}
+	i915_vma_unpin(batch);
+	i915_vma_put(batch);
+out_free:
+	kfree(request);
+	return err;
+}
+
+static int live_sequential_engines(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	const unsigned int nengines = num_uabi_engines(i915);
+	struct i915_request **request;
+	struct i915_request *prev = NULL;
+	struct intel_engine_cs *engine;
+	struct igt_live_test t;
+	unsigned int idx;
+	int err;
+
+	/*
+	 * Check we can submit requests to all engines sequentially, such
+	 * that each successive request waits for the earlier ones. This
+	 * tests that we don't execute requests out of order, even though
+	 * they are running on independent engines.
+	 */
+
+	request = kcalloc(nengines, sizeof(*request), GFP_KERNEL);
+	if (!request)
+		return -ENOMEM;
+
+	err = igt_live_test_begin(&t, i915, __func__, "");
+	if (err)
+		goto out_free;
+
+	idx = 0;
+	for_each_uabi_engine(engine, i915) {
+		struct i915_vma *batch;
+
+		batch = recursive_batch(i915);
+		if (IS_ERR(batch)) {
+			err = PTR_ERR(batch);
+			pr_err("%s: Unable to create batch for %s, err=%d\n",
+			       __func__, engine->name, err);
+			goto out_free;
+		}
+
+		i915_vma_lock(batch);
+		request[idx] = intel_engine_create_kernel_request(engine);
+		if (IS_ERR(request[idx])) {
+			err = PTR_ERR(request[idx]);
+			pr_err("%s: Request allocation failed for %s with err=%d\n",
+			       __func__, engine->name, err);
+			goto out_unlock;
+		}
+
+		if (prev) {
+			err = i915_request_await_dma_fence(request[idx],
+							   &prev->fence);
+			if (err) {
+				i915_request_add(request[idx]);
+				pr_err("%s: Request await failed for %s with err=%d\n",
+				       __func__, engine->name, err);
+				goto out_unlock;
+			}
+		}
+
+		err = i915_request_await_object(request[idx],
+						batch->obj, false);
+		if (err == 0)
+			err = i915_vma_move_to_active(batch, request[idx], 0);
+		GEM_BUG_ON(err);
+
+		err = engine->emit_bb_start(request[idx],
+					    batch->node.start,
+					    batch->node.size,
+					    0);
+		GEM_BUG_ON(err);
+		request[idx]->batch = batch;
+
+		i915_request_get(request[idx]);
+		i915_request_add(request[idx]);
+
+		prev = request[idx];
+		idx++;
+
+out_unlock:
+		i915_vma_unlock(batch);
+		if (err)
+			goto out_request;
+	}
+
+	idx = 0;
+	for_each_uabi_engine(engine, i915) {
+		long timeout;
+
+		if (i915_request_completed(request[idx])) {
+			pr_err("%s(%s): request completed too early!\n",
+			       __func__, engine->name);
+			err = -EINVAL;
+			goto out_request;
+		}
+
+		err = recursive_batch_resolve(request[idx]->batch);
+		if (err) {
+			pr_err("%s: failed to resolve batch, err=%d\n",
+			       __func__, err);
+			goto out_request;
+		}
+
+		timeout = i915_request_wait(request[idx], 0,
+					    MAX_SCHEDULE_TIMEOUT);
+		if (timeout < 0) {
+			err = timeout;
+			pr_err("%s: error waiting for request on %s, err=%d\n",
+			       __func__, engine->name, err);
+			goto out_request;
+		}
+
+		GEM_BUG_ON(!i915_request_completed(request[idx]));
+		idx++;
+	}
+
+	err = igt_live_test_end(&t);
+
+out_request:
+	idx = 0;
+	for_each_uabi_engine(engine, i915) {
+		u32 *cmd;
+
+		if (!request[idx])
+			break;
+
+		cmd = i915_gem_object_pin_map_unlocked(request[idx]->batch->obj,
+						       I915_MAP_WC);
+		if (!IS_ERR(cmd)) {
+			*cmd = MI_BATCH_BUFFER_END;
+
+			__i915_gem_object_flush_map(request[idx]->batch->obj,
+						    0, sizeof(*cmd));
+			i915_gem_object_unpin_map(request[idx]->batch->obj);
+
+			intel_gt_chipset_flush(engine->gt);
+		}
+
+		i915_vma_put(request[idx]->batch);
+		i915_request_put(request[idx]);
+		idx++;
+	}
+out_free:
+	kfree(request);
+	return err;
+}
+
+struct parallel_thread {
+	struct kthread_worker *worker;
+	struct kthread_work work;
+	struct intel_engine_cs *engine;
+	int result;
+};
+
+static void __live_parallel_engine1(struct kthread_work *work)
+{
+	struct parallel_thread *thread =
+		container_of(work, typeof(*thread), work);
+	struct intel_engine_cs *engine = thread->engine;
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+	int err = 0;
+
+	count = 0;
+	intel_engine_pm_get(engine);
+	do {
+		struct i915_request *rq;
+
+		rq = i915_request_create(engine->kernel_context);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		err = 0;
+		if (i915_request_wait(rq, 0, HZ) < 0)
+			err = -ETIME;
+		i915_request_put(rq);
+		if (err)
+			break;
+
+		count++;
+	} while (!__igt_timeout(end_time, NULL));
+	intel_engine_pm_put(engine);
+
+	pr_info("%s: %lu request + sync\n", engine->name, count);
+	thread->result = err;
+}
+
+static void __live_parallel_engineN(struct kthread_work *work)
+{
+	struct parallel_thread *thread =
+		container_of(work, typeof(*thread), work);
+	struct intel_engine_cs *engine = thread->engine;
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+	int err = 0;
+
+	count = 0;
+	intel_engine_pm_get(engine);
+	do {
+		struct i915_request *rq;
+
+		rq = i915_request_create(engine->kernel_context);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_add(rq);
+		count++;
+	} while (!__igt_timeout(end_time, NULL));
+	intel_engine_pm_put(engine);
+
+	pr_info("%s: %lu requests\n", engine->name, count);
+	thread->result = err;
+}
+
+static bool wake_all(struct drm_i915_private *i915)
+{
+	if (atomic_dec_and_test(&i915->selftest.counter)) {
+		wake_up_var(&i915->selftest.counter);
+		return true;
+	}
+
+	return false;
+}
+
+static int wait_for_all(struct drm_i915_private *i915)
+{
+	if (wake_all(i915))
+		return 0;
+
+	if (wait_var_event_timeout(&i915->selftest.counter,
+				   !atomic_read(&i915->selftest.counter),
+				   i915_selftest.timeout_jiffies))
+		return 0;
+
+	return -ETIME;
+}
+
+static void __live_parallel_spin(struct kthread_work *work)
+{
+	struct parallel_thread *thread =
+		container_of(work, typeof(*thread), work);
+	struct intel_engine_cs *engine = thread->engine;
+	struct igt_spinner spin;
+	struct i915_request *rq;
+	int err = 0;
+
+	/*
+	 * Create a spinner running for eternity on each engine. If a second
+	 * spinner is incorrectly placed on the same engine, it will not be
+	 * able to start in time.
+	 */
+
+	if (igt_spinner_init(&spin, engine->gt)) {
+		wake_all(engine->i915);
+		thread->result = -ENOMEM;
+		return;
+	}
+
+	intel_engine_pm_get(engine);
+	rq = igt_spinner_create_request(&spin,
+					engine->kernel_context,
+					MI_NOOP); /* no preemption */
+	intel_engine_pm_put(engine);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		if (err == -ENODEV)
+			err = 0;
+		wake_all(engine->i915);
+		goto out_spin;
+	}
+
+	i915_request_get(rq);
+	i915_request_add(rq);
+	if (igt_wait_for_spinner(&spin, rq)) {
+		/* Occupy this engine for the whole test */
+		err = wait_for_all(engine->i915);
+	} else {
+		pr_err("Failed to start spinner on %s\n", engine->name);
+		err = -EINVAL;
+	}
+	igt_spinner_end(&spin);
+
+	if (err == 0 && i915_request_wait(rq, 0, HZ) < 0)
+		err = -EIO;
+	i915_request_put(rq);
+
+out_spin:
+	igt_spinner_fini(&spin);
+	thread->result = err;
+}
+
+static int live_parallel_engines(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	static void (* const func[])(struct kthread_work *) = {
+		__live_parallel_engine1,
+		__live_parallel_engineN,
+		__live_parallel_spin,
+		NULL,
+	};
+	const unsigned int nengines = num_uabi_engines(i915);
+	struct parallel_thread *threads;
+	struct intel_engine_cs *engine;
+	void (* const *fn)(struct kthread_work *);
+	int err = 0;
+
+	/*
+	 * Check we can submit requests to all engines concurrently. This
+	 * tests that we load up the system maximally.
+	 */
+
+	threads = kcalloc(nengines, sizeof(*threads), GFP_KERNEL);
+	if (!threads)
+		return -ENOMEM;
+
+	for (fn = func; !err && *fn; fn++) {
+		char name[KSYM_NAME_LEN];
+		struct igt_live_test t;
+		unsigned int idx;
+
+		snprintf(name, sizeof(name), "%ps", *fn);
+		err = igt_live_test_begin(&t, i915, __func__, name);
+		if (err)
+			break;
+
+		atomic_set(&i915->selftest.counter, nengines);
+
+		idx = 0;
+		for_each_uabi_engine(engine, i915) {
+			struct kthread_worker *worker;
+
+			worker = kthread_create_worker(0, "igt/parallel:%s",
+						       engine->name);
+			if (IS_ERR(worker)) {
+				err = PTR_ERR(worker);
+				break;
+			}
+
+			threads[idx].worker = worker;
+			threads[idx].result = 0;
+			threads[idx].engine = engine;
+
+			kthread_init_work(&threads[idx].work, *fn);
+			kthread_queue_work(worker, &threads[idx].work);
+			idx++;
+		}
+
+		idx = 0;
+		for_each_uabi_engine(engine, i915) {
+			int status;
+
+			if (!threads[idx].worker)
+				break;
+
+			kthread_flush_work(&threads[idx].work);
+			status = READ_ONCE(threads[idx].result);
+			if (status && !err)
+				err = status;
+
+			kthread_destroy_worker(threads[idx++].worker);
+		}
+
+		if (igt_live_test_end(&t))
+			err = -EIO;
+	}
+
+	kfree(threads);
+	return err;
+}
+
+static int
+max_batches(struct i915_gem_context *ctx, struct intel_engine_cs *engine)
+{
+	struct i915_request *rq;
+	int ret;
+
+	/*
+	 * Before execlists, all contexts share the same ringbuffer. With
+	 * execlists, each context/engine has a separate ringbuffer and
+	 * for the purposes of this test, inexhaustible.
+	 *
+	 * For the global ringbuffer though, we have to be very careful
+	 * that we do not wrap while preventing the execution of requests
+	 * with a unsignaled fence.
+	 */
+	if (HAS_EXECLISTS(ctx->i915))
+		return INT_MAX;
+
+	rq = igt_request_alloc(ctx, engine);
+	if (IS_ERR(rq)) {
+		ret = PTR_ERR(rq);
+	} else {
+		int sz;
+
+		ret = rq->ring->size - rq->reserved_space;
+		i915_request_add(rq);
+
+		sz = rq->ring->emit - rq->head;
+		if (sz < 0)
+			sz += rq->ring->size;
+		ret /= sz;
+		ret /= 2; /* leave half spare, in case of emergency! */
+	}
+
+	return ret;
+}
+
+static int live_breadcrumbs_smoketest(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	const unsigned int nengines = num_uabi_engines(i915);
+	const unsigned int ncpus = num_online_cpus();
+	unsigned long num_waits, num_fences;
+	struct intel_engine_cs *engine;
+	struct smoke_thread *threads;
+	struct igt_live_test live;
+	intel_wakeref_t wakeref;
+	struct smoketest *smoke;
+	unsigned int n, idx;
+	struct file *file;
+	int ret = 0;
+
+	/*
+	 * Smoketest our breadcrumb/signal handling for requests across multiple
+	 * threads. A very simple test to only catch the most egregious of bugs.
+	 * See __igt_breadcrumbs_smoketest();
+	 *
+	 * On real hardware this time.
+	 */
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	file = mock_file(i915);
+	if (IS_ERR(file)) {
+		ret = PTR_ERR(file);
+		goto out_rpm;
+	}
+
+	smoke = kcalloc(nengines, sizeof(*smoke), GFP_KERNEL);
+	if (!smoke) {
+		ret = -ENOMEM;
+		goto out_file;
+	}
+
+	threads = kcalloc(ncpus * nengines, sizeof(*threads), GFP_KERNEL);
+	if (!threads) {
+		ret = -ENOMEM;
+		goto out_smoke;
+	}
+
+	smoke[0].request_alloc = __live_request_alloc;
+	smoke[0].ncontexts = 64;
+	smoke[0].contexts = kcalloc(smoke[0].ncontexts,
+				    sizeof(*smoke[0].contexts),
+				    GFP_KERNEL);
+	if (!smoke[0].contexts) {
+		ret = -ENOMEM;
+		goto out_threads;
+	}
+
+	for (n = 0; n < smoke[0].ncontexts; n++) {
+		smoke[0].contexts[n] = live_context(i915, file);
+		if (IS_ERR(smoke[0].contexts[n])) {
+			ret = PTR_ERR(smoke[0].contexts[n]);
+			goto out_contexts;
+		}
+	}
+
+	ret = igt_live_test_begin(&live, i915, __func__, "");
+	if (ret)
+		goto out_contexts;
+
+	idx = 0;
+	for_each_uabi_engine(engine, i915) {
+		smoke[idx] = smoke[0];
+		smoke[idx].engine = engine;
+		smoke[idx].max_batch =
+			max_batches(smoke[0].contexts[0], engine);
+		if (smoke[idx].max_batch < 0) {
+			ret = smoke[idx].max_batch;
+			goto out_flush;
+		}
+		/* One ring interleaved between requests from all cpus */
+		smoke[idx].max_batch /= num_online_cpus() + 1;
+		pr_debug("Limiting batches to %d requests on %s\n",
+			 smoke[idx].max_batch, engine->name);
+
+		for (n = 0; n < ncpus; n++) {
+			unsigned int i = idx * ncpus + n;
+			struct kthread_worker *worker;
+
+			worker = kthread_create_worker(0, "igt/%d.%d", idx, n);
+			if (IS_ERR(worker)) {
+				ret = PTR_ERR(worker);
+				goto out_flush;
+			}
+
+			threads[i].worker = worker;
+			threads[i].t = &smoke[idx];
+
+			kthread_init_work(&threads[i].work,
+					  __igt_breadcrumbs_smoketest);
+			kthread_queue_work(worker, &threads[i].work);
+		}
+
+		idx++;
+	}
+
+	msleep(jiffies_to_msecs(i915_selftest.timeout_jiffies));
+
+out_flush:
+	idx = 0;
+	num_waits = 0;
+	num_fences = 0;
+	for_each_uabi_engine(engine, i915) {
+		for (n = 0; n < ncpus; n++) {
+			unsigned int i = idx * ncpus + n;
+			int err;
+
+			if (!threads[i].worker)
+				continue;
+
+			WRITE_ONCE(threads[i].stop, true);
+			kthread_flush_work(&threads[i].work);
+			err = READ_ONCE(threads[i].result);
+			if (err < 0 && !ret)
+				ret = err;
+
+			kthread_destroy_worker(threads[i].worker);
+		}
+
+		num_waits += atomic_long_read(&smoke[idx].num_waits);
+		num_fences += atomic_long_read(&smoke[idx].num_fences);
+		idx++;
+	}
+	pr_info("Completed %lu waits for %lu fences across %d engines and %d cpus\n",
+		num_waits, num_fences, idx, ncpus);
+
+	ret = igt_live_test_end(&live) ?: ret;
+out_contexts:
+	kfree(smoke[0].contexts);
+out_threads:
+	kfree(threads);
+out_smoke:
+	kfree(smoke);
+out_file:
+	fput(file);
+out_rpm:
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+
+	return ret;
+}
+
+int i915_request_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_nop_request),
+		SUBTEST(live_all_engines),
+		SUBTEST(live_sequential_engines),
+		SUBTEST(live_parallel_engines),
+		SUBTEST(live_empty_request),
+		SUBTEST(live_cancel_request),
+		SUBTEST(live_breadcrumbs_smoketest),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return i915_live_subtests(tests, i915);
+}
+
+static int switch_to_kernel_sync(struct intel_context *ce, int err)
+{
+	struct i915_request *rq;
+	struct dma_fence *fence;
+
+	rq = intel_engine_create_kernel_request(ce->engine);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	fence = i915_active_fence_get(&ce->timeline->last_request);
+	if (fence) {
+		i915_request_await_dma_fence(rq, fence);
+		dma_fence_put(fence);
+	}
+
+	rq = i915_request_get(rq);
+	i915_request_add(rq);
+	if (i915_request_wait(rq, 0, HZ / 2) < 0 && !err)
+		err = -ETIME;
+	i915_request_put(rq);
+
+	while (!err && !intel_engine_is_idle(ce->engine))
+		intel_engine_flush_submission(ce->engine);
+
+	return err;
+}
+
+struct perf_stats {
+	struct intel_engine_cs *engine;
+	unsigned long count;
+	ktime_t time;
+	ktime_t busy;
+	u64 runtime;
+};
+
+struct perf_series {
+	struct drm_i915_private *i915;
+	unsigned int nengines;
+	struct intel_context *ce[];
+};
+
+static int cmp_u32(const void *A, const void *B)
+{
+	const u32 *a = A, *b = B;
+
+	return *a - *b;
+}
+
+static u32 trifilter(u32 *a)
+{
+	u64 sum;
+
+#define TF_COUNT 5
+	sort(a, TF_COUNT, sizeof(*a), cmp_u32, NULL);
+
+	sum = mul_u32_u32(a[2], 2);
+	sum += a[1];
+	sum += a[3];
+
+	GEM_BUG_ON(sum > U32_MAX);
+	return sum;
+#define TF_BIAS 2
+}
+
+static u64 cycles_to_ns(struct intel_engine_cs *engine, u32 cycles)
+{
+	u64 ns = intel_gt_clock_interval_to_ns(engine->gt, cycles);
+
+	return DIV_ROUND_CLOSEST(ns, 1 << TF_BIAS);
+}
+
+static u32 *emit_timestamp_store(u32 *cs, struct intel_context *ce, u32 offset)
+{
+	*cs++ = MI_STORE_REGISTER_MEM_GEN8 | MI_USE_GGTT;
+	*cs++ = i915_mmio_reg_offset(RING_TIMESTAMP((ce->engine->mmio_base)));
+	*cs++ = offset;
+	*cs++ = 0;
+
+	return cs;
+}
+
+static u32 *emit_store_dw(u32 *cs, u32 offset, u32 value)
+{
+	*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+	*cs++ = offset;
+	*cs++ = 0;
+	*cs++ = value;
+
+	return cs;
+}
+
+static u32 *emit_semaphore_poll(u32 *cs, u32 mode, u32 value, u32 offset)
+{
+	*cs++ = MI_SEMAPHORE_WAIT |
+		MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_POLL |
+		mode;
+	*cs++ = value;
+	*cs++ = offset;
+	*cs++ = 0;
+
+	return cs;
+}
+
+static u32 *emit_semaphore_poll_until(u32 *cs, u32 offset, u32 value)
+{
+	return emit_semaphore_poll(cs, MI_SEMAPHORE_SAD_EQ_SDD, value, offset);
+}
+
+static void semaphore_set(u32 *sema, u32 value)
+{
+	WRITE_ONCE(*sema, value);
+	wmb(); /* flush the update to the cache, and beyond */
+}
+
+static u32 *hwsp_scratch(const struct intel_context *ce)
+{
+	return memset32(ce->engine->status_page.addr + 1000, 0, 21);
+}
+
+static u32 hwsp_offset(const struct intel_context *ce, u32 *dw)
+{
+	return (i915_ggtt_offset(ce->engine->status_page.vma) +
+		offset_in_page(dw));
+}
+
+static int measure_semaphore_response(struct intel_context *ce)
+{
+	u32 *sema = hwsp_scratch(ce);
+	const u32 offset = hwsp_offset(ce, sema);
+	u32 elapsed[TF_COUNT], cycles;
+	struct i915_request *rq;
+	u32 *cs;
+	int err;
+	int i;
+
+	/*
+	 * Measure how many cycles it takes for the HW to detect the change
+	 * in a semaphore value.
+	 *
+	 *    A: read CS_TIMESTAMP from CPU
+	 *    poke semaphore
+	 *    B: read CS_TIMESTAMP on GPU
+	 *
+	 * Semaphore latency: B - A
+	 */
+
+	semaphore_set(sema, -1);
+
+	rq = i915_request_create(ce);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 4 + 12 * ARRAY_SIZE(elapsed));
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		err = PTR_ERR(cs);
+		goto err;
+	}
+
+	cs = emit_store_dw(cs, offset, 0);
+	for (i = 1; i <= ARRAY_SIZE(elapsed); i++) {
+		cs = emit_semaphore_poll_until(cs, offset, i);
+		cs = emit_timestamp_store(cs, ce, offset + i * sizeof(u32));
+		cs = emit_store_dw(cs, offset, 0);
+	}
+
+	intel_ring_advance(rq, cs);
+	i915_request_add(rq);
+
+	if (wait_for(READ_ONCE(*sema) == 0, 50)) {
+		err = -EIO;
+		goto err;
+	}
+
+	for (i = 1; i <= ARRAY_SIZE(elapsed); i++) {
+		preempt_disable();
+		cycles = ENGINE_READ_FW(ce->engine, RING_TIMESTAMP);
+		semaphore_set(sema, i);
+		preempt_enable();
+
+		if (wait_for(READ_ONCE(*sema) == 0, 50)) {
+			err = -EIO;
+			goto err;
+		}
+
+		elapsed[i - 1] = sema[i] - cycles;
+	}
+
+	cycles = trifilter(elapsed);
+	pr_info("%s: semaphore response %d cycles, %lluns\n",
+		ce->engine->name, cycles >> TF_BIAS,
+		cycles_to_ns(ce->engine, cycles));
+
+	return intel_gt_wait_for_idle(ce->engine->gt, HZ);
+
+err:
+	intel_gt_set_wedged(ce->engine->gt);
+	return err;
+}
+
+static int measure_idle_dispatch(struct intel_context *ce)
+{
+	u32 *sema = hwsp_scratch(ce);
+	const u32 offset = hwsp_offset(ce, sema);
+	u32 elapsed[TF_COUNT], cycles;
+	u32 *cs;
+	int err;
+	int i;
+
+	/*
+	 * Measure how long it takes for us to submit a request while the
+	 * engine is idle, but is resting in our context.
+	 *
+	 *    A: read CS_TIMESTAMP from CPU
+	 *    submit request
+	 *    B: read CS_TIMESTAMP on GPU
+	 *
+	 * Submission latency: B - A
+	 */
+
+	for (i = 0; i < ARRAY_SIZE(elapsed); i++) {
+		struct i915_request *rq;
+
+		err = intel_gt_wait_for_idle(ce->engine->gt, HZ / 2);
+		if (err)
+			return err;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err;
+		}
+
+		cs = intel_ring_begin(rq, 4);
+		if (IS_ERR(cs)) {
+			i915_request_add(rq);
+			err = PTR_ERR(cs);
+			goto err;
+		}
+
+		cs = emit_timestamp_store(cs, ce, offset + i * sizeof(u32));
+
+		intel_ring_advance(rq, cs);
+
+		preempt_disable();
+		local_bh_disable();
+		elapsed[i] = ENGINE_READ_FW(ce->engine, RING_TIMESTAMP);
+		i915_request_add(rq);
+		local_bh_enable();
+		preempt_enable();
+	}
+
+	err = intel_gt_wait_for_idle(ce->engine->gt, HZ / 2);
+	if (err)
+		goto err;
+
+	for (i = 0; i < ARRAY_SIZE(elapsed); i++)
+		elapsed[i] = sema[i] - elapsed[i];
+
+	cycles = trifilter(elapsed);
+	pr_info("%s: idle dispatch latency %d cycles, %lluns\n",
+		ce->engine->name, cycles >> TF_BIAS,
+		cycles_to_ns(ce->engine, cycles));
+
+	return intel_gt_wait_for_idle(ce->engine->gt, HZ);
+
+err:
+	intel_gt_set_wedged(ce->engine->gt);
+	return err;
+}
+
+static int measure_busy_dispatch(struct intel_context *ce)
+{
+	u32 *sema = hwsp_scratch(ce);
+	const u32 offset = hwsp_offset(ce, sema);
+	u32 elapsed[TF_COUNT + 1], cycles;
+	u32 *cs;
+	int err;
+	int i;
+
+	/*
+	 * Measure how long it takes for us to submit a request while the
+	 * engine is busy, polling on a semaphore in our context. With
+	 * direct submission, this will include the cost of a lite restore.
+	 *
+	 *    A: read CS_TIMESTAMP from CPU
+	 *    submit request
+	 *    B: read CS_TIMESTAMP on GPU
+	 *
+	 * Submission latency: B - A
+	 */
+
+	for (i = 1; i <= ARRAY_SIZE(elapsed); i++) {
+		struct i915_request *rq;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err;
+		}
+
+		cs = intel_ring_begin(rq, 12);
+		if (IS_ERR(cs)) {
+			i915_request_add(rq);
+			err = PTR_ERR(cs);
+			goto err;
+		}
+
+		cs = emit_store_dw(cs, offset + i * sizeof(u32), -1);
+		cs = emit_semaphore_poll_until(cs, offset, i);
+		cs = emit_timestamp_store(cs, ce, offset + i * sizeof(u32));
+
+		intel_ring_advance(rq, cs);
+
+		if (i > 1 && wait_for(READ_ONCE(sema[i - 1]), 500)) {
+			err = -EIO;
+			goto err;
+		}
+
+		preempt_disable();
+		local_bh_disable();
+		elapsed[i - 1] = ENGINE_READ_FW(ce->engine, RING_TIMESTAMP);
+		i915_request_add(rq);
+		local_bh_enable();
+		semaphore_set(sema, i - 1);
+		preempt_enable();
+	}
+
+	wait_for(READ_ONCE(sema[i - 1]), 500);
+	semaphore_set(sema, i - 1);
+
+	for (i = 1; i <= TF_COUNT; i++) {
+		GEM_BUG_ON(sema[i] == -1);
+		elapsed[i - 1] = sema[i] - elapsed[i];
+	}
+
+	cycles = trifilter(elapsed);
+	pr_info("%s: busy dispatch latency %d cycles, %lluns\n",
+		ce->engine->name, cycles >> TF_BIAS,
+		cycles_to_ns(ce->engine, cycles));
+
+	return intel_gt_wait_for_idle(ce->engine->gt, HZ);
+
+err:
+	intel_gt_set_wedged(ce->engine->gt);
+	return err;
+}
+
+static int plug(struct intel_engine_cs *engine, u32 *sema, u32 mode, int value)
+{
+	const u32 offset =
+		i915_ggtt_offset(engine->status_page.vma) +
+		offset_in_page(sema);
+	struct i915_request *rq;
+	u32 *cs;
+
+	rq = i915_request_create(engine->kernel_context);
+	if (IS_ERR(rq))
+		return PTR_ERR(rq);
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs)) {
+		i915_request_add(rq);
+		return PTR_ERR(cs);
+	}
+
+	cs = emit_semaphore_poll(cs, mode, value, offset);
+
+	intel_ring_advance(rq, cs);
+	i915_request_add(rq);
+
+	return 0;
+}
+
+static int measure_inter_request(struct intel_context *ce)
+{
+	u32 *sema = hwsp_scratch(ce);
+	const u32 offset = hwsp_offset(ce, sema);
+	u32 elapsed[TF_COUNT + 1], cycles;
+	struct i915_sw_fence *submit;
+	int i, err;
+
+	/*
+	 * Measure how long it takes to advance from one request into the
+	 * next. Between each request we flush the GPU caches to memory,
+	 * update the breadcrumbs, and then invalidate those caches.
+	 * We queue up all the requests to be submitted in one batch so
+	 * it should be one set of contiguous measurements.
+	 *
+	 *    A: read CS_TIMESTAMP on GPU
+	 *    advance request
+	 *    B: read CS_TIMESTAMP on GPU
+	 *
+	 * Request latency: B - A
+	 */
+
+	err = plug(ce->engine, sema, MI_SEMAPHORE_SAD_NEQ_SDD, 0);
+	if (err)
+		return err;
+
+	submit = heap_fence_create(GFP_KERNEL);
+	if (!submit) {
+		semaphore_set(sema, 1);
+		return -ENOMEM;
+	}
+
+	intel_engine_flush_submission(ce->engine);
+	for (i = 1; i <= ARRAY_SIZE(elapsed); i++) {
+		struct i915_request *rq;
+		u32 *cs;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err_submit;
+		}
+
+		err = i915_sw_fence_await_sw_fence_gfp(&rq->submit,
+						       submit,
+						       GFP_KERNEL);
+		if (err < 0) {
+			i915_request_add(rq);
+			goto err_submit;
+		}
+
+		cs = intel_ring_begin(rq, 4);
+		if (IS_ERR(cs)) {
+			i915_request_add(rq);
+			err = PTR_ERR(cs);
+			goto err_submit;
+		}
+
+		cs = emit_timestamp_store(cs, ce, offset + i * sizeof(u32));
+
+		intel_ring_advance(rq, cs);
+		i915_request_add(rq);
+	}
+	i915_sw_fence_commit(submit);
+	intel_engine_flush_submission(ce->engine);
+	heap_fence_put(submit);
+
+	semaphore_set(sema, 1);
+	err = intel_gt_wait_for_idle(ce->engine->gt, HZ / 2);
+	if (err)
+		goto err;
+
+	for (i = 1; i <= TF_COUNT; i++)
+		elapsed[i - 1] = sema[i + 1] - sema[i];
+
+	cycles = trifilter(elapsed);
+	pr_info("%s: inter-request latency %d cycles, %lluns\n",
+		ce->engine->name, cycles >> TF_BIAS,
+		cycles_to_ns(ce->engine, cycles));
+
+	return intel_gt_wait_for_idle(ce->engine->gt, HZ);
+
+err_submit:
+	i915_sw_fence_commit(submit);
+	heap_fence_put(submit);
+	semaphore_set(sema, 1);
+err:
+	intel_gt_set_wedged(ce->engine->gt);
+	return err;
+}
+
+static int measure_context_switch(struct intel_context *ce)
+{
+	u32 *sema = hwsp_scratch(ce);
+	const u32 offset = hwsp_offset(ce, sema);
+	struct i915_request *fence = NULL;
+	u32 elapsed[TF_COUNT + 1], cycles;
+	int i, j, err;
+	u32 *cs;
+
+	/*
+	 * Measure how long it takes to advance from one request in one
+	 * context to a request in another context. This allows us to
+	 * measure how long the context save/restore take, along with all
+	 * the inter-context setup we require.
+	 *
+	 *    A: read CS_TIMESTAMP on GPU
+	 *    switch context
+	 *    B: read CS_TIMESTAMP on GPU
+	 *
+	 * Context switch latency: B - A
+	 */
+
+	err = plug(ce->engine, sema, MI_SEMAPHORE_SAD_NEQ_SDD, 0);
+	if (err)
+		return err;
+
+	for (i = 1; i <= ARRAY_SIZE(elapsed); i++) {
+		struct intel_context *arr[] = {
+			ce, ce->engine->kernel_context
+		};
+		u32 addr = offset + ARRAY_SIZE(arr) * i * sizeof(u32);
+
+		for (j = 0; j < ARRAY_SIZE(arr); j++) {
+			struct i915_request *rq;
+
+			rq = i915_request_create(arr[j]);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				goto err_fence;
+			}
+
+			if (fence) {
+				err = i915_request_await_dma_fence(rq,
+								   &fence->fence);
+				if (err) {
+					i915_request_add(rq);
+					goto err_fence;
+				}
+			}
+
+			cs = intel_ring_begin(rq, 4);
+			if (IS_ERR(cs)) {
+				i915_request_add(rq);
+				err = PTR_ERR(cs);
+				goto err_fence;
+			}
+
+			cs = emit_timestamp_store(cs, ce, addr);
+			addr += sizeof(u32);
+
+			intel_ring_advance(rq, cs);
+
+			i915_request_put(fence);
+			fence = i915_request_get(rq);
+
+			i915_request_add(rq);
+		}
+	}
+	i915_request_put(fence);
+	intel_engine_flush_submission(ce->engine);
+
+	semaphore_set(sema, 1);
+	err = intel_gt_wait_for_idle(ce->engine->gt, HZ / 2);
+	if (err)
+		goto err;
+
+	for (i = 1; i <= TF_COUNT; i++)
+		elapsed[i - 1] = sema[2 * i + 2] - sema[2 * i + 1];
+
+	cycles = trifilter(elapsed);
+	pr_info("%s: context switch latency %d cycles, %lluns\n",
+		ce->engine->name, cycles >> TF_BIAS,
+		cycles_to_ns(ce->engine, cycles));
+
+	return intel_gt_wait_for_idle(ce->engine->gt, HZ);
+
+err_fence:
+	i915_request_put(fence);
+	semaphore_set(sema, 1);
+err:
+	intel_gt_set_wedged(ce->engine->gt);
+	return err;
+}
+
+static int measure_preemption(struct intel_context *ce)
+{
+	u32 *sema = hwsp_scratch(ce);
+	const u32 offset = hwsp_offset(ce, sema);
+	u32 elapsed[TF_COUNT], cycles;
+	u32 *cs;
+	int err;
+	int i;
+
+	/*
+	 * We measure two latencies while triggering preemption. The first
+	 * latency is how long it takes for us to submit a preempting request.
+	 * The second latency is how it takes for us to return from the
+	 * preemption back to the original context.
+	 *
+	 *    A: read CS_TIMESTAMP from CPU
+	 *    submit preemption
+	 *    B: read CS_TIMESTAMP on GPU (in preempting context)
+	 *    context switch
+	 *    C: read CS_TIMESTAMP on GPU (in original context)
+	 *
+	 * Preemption dispatch latency: B - A
+	 * Preemption switch latency: C - B
+	 */
+
+	if (!intel_engine_has_preemption(ce->engine))
+		return 0;
+
+	for (i = 1; i <= ARRAY_SIZE(elapsed); i++) {
+		u32 addr = offset + 2 * i * sizeof(u32);
+		struct i915_request *rq;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err;
+		}
+
+		cs = intel_ring_begin(rq, 12);
+		if (IS_ERR(cs)) {
+			i915_request_add(rq);
+			err = PTR_ERR(cs);
+			goto err;
+		}
+
+		cs = emit_store_dw(cs, addr, -1);
+		cs = emit_semaphore_poll_until(cs, offset, i);
+		cs = emit_timestamp_store(cs, ce, addr + sizeof(u32));
+
+		intel_ring_advance(rq, cs);
+		i915_request_add(rq);
+
+		if (wait_for(READ_ONCE(sema[2 * i]) == -1, 500)) {
+			err = -EIO;
+			goto err;
+		}
+
+		rq = i915_request_create(ce->engine->kernel_context);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err;
+		}
+
+		cs = intel_ring_begin(rq, 8);
+		if (IS_ERR(cs)) {
+			i915_request_add(rq);
+			err = PTR_ERR(cs);
+			goto err;
+		}
+
+		cs = emit_timestamp_store(cs, ce, addr);
+		cs = emit_store_dw(cs, offset, i);
+
+		intel_ring_advance(rq, cs);
+		rq->sched.attr.priority = I915_PRIORITY_BARRIER;
+
+		elapsed[i - 1] = ENGINE_READ_FW(ce->engine, RING_TIMESTAMP);
+		i915_request_add(rq);
+	}
+
+	if (wait_for(READ_ONCE(sema[2 * i - 2]) != -1, 500)) {
+		err = -EIO;
+		goto err;
+	}
+
+	for (i = 1; i <= TF_COUNT; i++)
+		elapsed[i - 1] = sema[2 * i + 0] - elapsed[i - 1];
+
+	cycles = trifilter(elapsed);
+	pr_info("%s: preemption dispatch latency %d cycles, %lluns\n",
+		ce->engine->name, cycles >> TF_BIAS,
+		cycles_to_ns(ce->engine, cycles));
+
+	for (i = 1; i <= TF_COUNT; i++)
+		elapsed[i - 1] = sema[2 * i + 1] - sema[2 * i + 0];
+
+	cycles = trifilter(elapsed);
+	pr_info("%s: preemption switch latency %d cycles, %lluns\n",
+		ce->engine->name, cycles >> TF_BIAS,
+		cycles_to_ns(ce->engine, cycles));
+
+	return intel_gt_wait_for_idle(ce->engine->gt, HZ);
+
+err:
+	intel_gt_set_wedged(ce->engine->gt);
+	return err;
+}
+
+struct signal_cb {
+	struct dma_fence_cb base;
+	bool seen;
+};
+
+static void signal_cb(struct dma_fence *fence, struct dma_fence_cb *cb)
+{
+	struct signal_cb *s = container_of(cb, typeof(*s), base);
+
+	smp_store_mb(s->seen, true); /* be safe, be strong */
+}
+
+static int measure_completion(struct intel_context *ce)
+{
+	u32 *sema = hwsp_scratch(ce);
+	const u32 offset = hwsp_offset(ce, sema);
+	u32 elapsed[TF_COUNT], cycles;
+	u32 *cs;
+	int err;
+	int i;
+
+	/*
+	 * Measure how long it takes for the signal (interrupt) to be
+	 * sent from the GPU to be processed by the CPU.
+	 *
+	 *    A: read CS_TIMESTAMP on GPU
+	 *    signal
+	 *    B: read CS_TIMESTAMP from CPU
+	 *
+	 * Completion latency: B - A
+	 */
+
+	for (i = 1; i <= ARRAY_SIZE(elapsed); i++) {
+		struct signal_cb cb = { .seen = false };
+		struct i915_request *rq;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto err;
+		}
+
+		cs = intel_ring_begin(rq, 12);
+		if (IS_ERR(cs)) {
+			i915_request_add(rq);
+			err = PTR_ERR(cs);
+			goto err;
+		}
+
+		cs = emit_store_dw(cs, offset + i * sizeof(u32), -1);
+		cs = emit_semaphore_poll_until(cs, offset, i);
+		cs = emit_timestamp_store(cs, ce, offset + i * sizeof(u32));
+
+		intel_ring_advance(rq, cs);
+
+		dma_fence_add_callback(&rq->fence, &cb.base, signal_cb);
+		i915_request_add(rq);
+
+		intel_engine_flush_submission(ce->engine);
+		if (wait_for(READ_ONCE(sema[i]) == -1, 50)) {
+			err = -EIO;
+			goto err;
+		}
+
+		preempt_disable();
+		semaphore_set(sema, i);
+		while (!READ_ONCE(cb.seen))
+			cpu_relax();
+
+		elapsed[i - 1] = ENGINE_READ_FW(ce->engine, RING_TIMESTAMP);
+		preempt_enable();
+	}
+
+	err = intel_gt_wait_for_idle(ce->engine->gt, HZ / 2);
+	if (err)
+		goto err;
+
+	for (i = 0; i < ARRAY_SIZE(elapsed); i++) {
+		GEM_BUG_ON(sema[i + 1] == -1);
+		elapsed[i] = elapsed[i] - sema[i + 1];
+	}
+
+	cycles = trifilter(elapsed);
+	pr_info("%s: completion latency %d cycles, %lluns\n",
+		ce->engine->name, cycles >> TF_BIAS,
+		cycles_to_ns(ce->engine, cycles));
+
+	return intel_gt_wait_for_idle(ce->engine->gt, HZ);
+
+err:
+	intel_gt_set_wedged(ce->engine->gt);
+	return err;
+}
+
+static void rps_pin(struct intel_gt *gt)
+{
+	/* Pin the frequency to max */
+	atomic_inc(&gt->rps.num_waiters);
+	intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);
+
+	mutex_lock(&gt->rps.lock);
+	intel_rps_set(&gt->rps, gt->rps.max_freq);
+	mutex_unlock(&gt->rps.lock);
+}
+
+static void rps_unpin(struct intel_gt *gt)
+{
+	intel_uncore_forcewake_put(gt->uncore, FORCEWAKE_ALL);
+	atomic_dec(&gt->rps.num_waiters);
+}
+
+static int perf_request_latency(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	struct pm_qos_request qos;
+	int err = 0;
+
+	if (GRAPHICS_VER(i915) < 8) /* per-engine CS timestamp, semaphores */
+		return 0;
+
+	cpu_latency_qos_add_request(&qos, 0); /* disable cstates */
+
+	for_each_uabi_engine(engine, i915) {
+		struct intel_context *ce;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out;
+		}
+
+		err = intel_context_pin(ce);
+		if (err) {
+			intel_context_put(ce);
+			goto out;
+		}
+
+		st_engine_heartbeat_disable(engine);
+		rps_pin(engine->gt);
+
+		if (err == 0)
+			err = measure_semaphore_response(ce);
+		if (err == 0)
+			err = measure_idle_dispatch(ce);
+		if (err == 0)
+			err = measure_busy_dispatch(ce);
+		if (err == 0)
+			err = measure_inter_request(ce);
+		if (err == 0)
+			err = measure_context_switch(ce);
+		if (err == 0)
+			err = measure_preemption(ce);
+		if (err == 0)
+			err = measure_completion(ce);
+
+		rps_unpin(engine->gt);
+		st_engine_heartbeat_enable(engine);
+
+		intel_context_unpin(ce);
+		intel_context_put(ce);
+		if (err)
+			goto out;
+	}
+
+out:
+	if (igt_flush_test(i915))
+		err = -EIO;
+
+	cpu_latency_qos_remove_request(&qos);
+	return err;
+}
+
+static int s_sync0(void *arg)
+{
+	struct perf_series *ps = arg;
+	IGT_TIMEOUT(end_time);
+	unsigned int idx = 0;
+	int err = 0;
+
+	GEM_BUG_ON(!ps->nengines);
+	do {
+		struct i915_request *rq;
+
+		rq = i915_request_create(ps->ce[idx]);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		if (i915_request_wait(rq, 0, HZ / 5) < 0)
+			err = -ETIME;
+		i915_request_put(rq);
+		if (err)
+			break;
+
+		if (++idx == ps->nengines)
+			idx = 0;
+	} while (!__igt_timeout(end_time, NULL));
+
+	return err;
+}
+
+static int s_sync1(void *arg)
+{
+	struct perf_series *ps = arg;
+	struct i915_request *prev = NULL;
+	IGT_TIMEOUT(end_time);
+	unsigned int idx = 0;
+	int err = 0;
+
+	GEM_BUG_ON(!ps->nengines);
+	do {
+		struct i915_request *rq;
+
+		rq = i915_request_create(ps->ce[idx]);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		if (prev && i915_request_wait(prev, 0, HZ / 5) < 0)
+			err = -ETIME;
+		i915_request_put(prev);
+		prev = rq;
+		if (err)
+			break;
+
+		if (++idx == ps->nengines)
+			idx = 0;
+	} while (!__igt_timeout(end_time, NULL));
+	i915_request_put(prev);
+
+	return err;
+}
+
+static int s_many(void *arg)
+{
+	struct perf_series *ps = arg;
+	IGT_TIMEOUT(end_time);
+	unsigned int idx = 0;
+
+	GEM_BUG_ON(!ps->nengines);
+	do {
+		struct i915_request *rq;
+
+		rq = i915_request_create(ps->ce[idx]);
+		if (IS_ERR(rq))
+			return PTR_ERR(rq);
+
+		i915_request_add(rq);
+
+		if (++idx == ps->nengines)
+			idx = 0;
+	} while (!__igt_timeout(end_time, NULL));
+
+	return 0;
+}
+
+static int perf_series_engines(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	static int (* const func[])(void *arg) = {
+		s_sync0,
+		s_sync1,
+		s_many,
+		NULL,
+	};
+	const unsigned int nengines = num_uabi_engines(i915);
+	struct intel_engine_cs *engine;
+	int (* const *fn)(void *arg);
+	struct pm_qos_request qos;
+	struct perf_stats *stats;
+	struct perf_series *ps;
+	unsigned int idx;
+	int err = 0;
+
+	stats = kcalloc(nengines, sizeof(*stats), GFP_KERNEL);
+	if (!stats)
+		return -ENOMEM;
+
+	ps = kzalloc(struct_size(ps, ce, nengines), GFP_KERNEL);
+	if (!ps) {
+		kfree(stats);
+		return -ENOMEM;
+	}
+
+	cpu_latency_qos_add_request(&qos, 0); /* disable cstates */
+
+	ps->i915 = i915;
+	ps->nengines = nengines;
+
+	idx = 0;
+	for_each_uabi_engine(engine, i915) {
+		struct intel_context *ce;
+
+		ce = intel_context_create(engine);
+		if (IS_ERR(ce)) {
+			err = PTR_ERR(ce);
+			goto out;
+		}
+
+		err = intel_context_pin(ce);
+		if (err) {
+			intel_context_put(ce);
+			goto out;
+		}
+
+		ps->ce[idx++] = ce;
+	}
+	GEM_BUG_ON(idx != ps->nengines);
+
+	for (fn = func; *fn && !err; fn++) {
+		char name[KSYM_NAME_LEN];
+		struct igt_live_test t;
+
+		snprintf(name, sizeof(name), "%ps", *fn);
+		err = igt_live_test_begin(&t, i915, __func__, name);
+		if (err)
+			break;
+
+		for (idx = 0; idx < nengines; idx++) {
+			struct perf_stats *p =
+				memset(&stats[idx], 0, sizeof(stats[idx]));
+			struct intel_context *ce = ps->ce[idx];
+
+			p->engine = ps->ce[idx]->engine;
+			intel_engine_pm_get(p->engine);
+
+			if (intel_engine_supports_stats(p->engine))
+				p->busy = intel_engine_get_busy_time(p->engine,
+								     &p->time) + 1;
+			else
+				p->time = ktime_get();
+			p->runtime = -intel_context_get_total_runtime_ns(ce);
+		}
+
+		err = (*fn)(ps);
+		if (igt_live_test_end(&t))
+			err = -EIO;
+
+		for (idx = 0; idx < nengines; idx++) {
+			struct perf_stats *p = &stats[idx];
+			struct intel_context *ce = ps->ce[idx];
+			int integer, decimal;
+			u64 busy, dt, now;
+
+			if (p->busy)
+				p->busy = ktime_sub(intel_engine_get_busy_time(p->engine,
+									       &now),
+						    p->busy - 1);
+			else
+				now = ktime_get();
+			p->time = ktime_sub(now, p->time);
+
+			err = switch_to_kernel_sync(ce, err);
+			p->runtime += intel_context_get_total_runtime_ns(ce);
+			intel_engine_pm_put(p->engine);
+
+			busy = 100 * ktime_to_ns(p->busy);
+			dt = ktime_to_ns(p->time);
+			if (dt) {
+				integer = div64_u64(busy, dt);
+				busy -= integer * dt;
+				decimal = div64_u64(100 * busy, dt);
+			} else {
+				integer = 0;
+				decimal = 0;
+			}
+
+			pr_info("%s %5s: { seqno:%d, busy:%d.%02d%%, runtime:%lldms, walltime:%lldms }\n",
+				name, p->engine->name, ce->timeline->seqno,
+				integer, decimal,
+				div_u64(p->runtime, 1000 * 1000),
+				div_u64(ktime_to_ns(p->time), 1000 * 1000));
+		}
+	}
+
+out:
+	for (idx = 0; idx < nengines; idx++) {
+		if (IS_ERR_OR_NULL(ps->ce[idx]))
+			break;
+
+		intel_context_unpin(ps->ce[idx]);
+		intel_context_put(ps->ce[idx]);
+	}
+	kfree(ps);
+
+	cpu_latency_qos_remove_request(&qos);
+	kfree(stats);
+	return err;
+}
+
+struct p_thread {
+	struct perf_stats p;
+	struct kthread_worker *worker;
+	struct kthread_work work;
+	struct intel_engine_cs *engine;
+	int result;
+};
+
+static void p_sync0(struct kthread_work *work)
+{
+	struct p_thread *thread = container_of(work, typeof(*thread), work);
+	struct perf_stats *p = &thread->p;
+	struct intel_engine_cs *engine = p->engine;
+	struct intel_context *ce;
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+	bool busy;
+	int err = 0;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce)) {
+		thread->result = PTR_ERR(ce);
+		return;
+	}
+
+	err = intel_context_pin(ce);
+	if (err) {
+		intel_context_put(ce);
+		thread->result = err;
+		return;
+	}
+
+	if (intel_engine_supports_stats(engine)) {
+		p->busy = intel_engine_get_busy_time(engine, &p->time);
+		busy = true;
+	} else {
+		p->time = ktime_get();
+		busy = false;
+	}
+
+	count = 0;
+	do {
+		struct i915_request *rq;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		err = 0;
+		if (i915_request_wait(rq, 0, HZ) < 0)
+			err = -ETIME;
+		i915_request_put(rq);
+		if (err)
+			break;
+
+		count++;
+	} while (!__igt_timeout(end_time, NULL));
+
+	if (busy) {
+		ktime_t now;
+
+		p->busy = ktime_sub(intel_engine_get_busy_time(engine, &now),
+				    p->busy);
+		p->time = ktime_sub(now, p->time);
+	} else {
+		p->time = ktime_sub(ktime_get(), p->time);
+	}
+
+	err = switch_to_kernel_sync(ce, err);
+	p->runtime = intel_context_get_total_runtime_ns(ce);
+	p->count = count;
+
+	intel_context_unpin(ce);
+	intel_context_put(ce);
+	thread->result = err;
+}
+
+static void p_sync1(struct kthread_work *work)
+{
+	struct p_thread *thread = container_of(work, typeof(*thread), work);
+	struct perf_stats *p = &thread->p;
+	struct intel_engine_cs *engine = p->engine;
+	struct i915_request *prev = NULL;
+	struct intel_context *ce;
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+	bool busy;
+	int err = 0;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce)) {
+		thread->result = PTR_ERR(ce);
+		return;
+	}
+
+	err = intel_context_pin(ce);
+	if (err) {
+		intel_context_put(ce);
+		thread->result = err;
+		return;
+	}
+
+	if (intel_engine_supports_stats(engine)) {
+		p->busy = intel_engine_get_busy_time(engine, &p->time);
+		busy = true;
+	} else {
+		p->time = ktime_get();
+		busy = false;
+	}
+
+	count = 0;
+	do {
+		struct i915_request *rq;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_get(rq);
+		i915_request_add(rq);
+
+		err = 0;
+		if (prev && i915_request_wait(prev, 0, HZ) < 0)
+			err = -ETIME;
+		i915_request_put(prev);
+		prev = rq;
+		if (err)
+			break;
+
+		count++;
+	} while (!__igt_timeout(end_time, NULL));
+	i915_request_put(prev);
+
+	if (busy) {
+		ktime_t now;
+
+		p->busy = ktime_sub(intel_engine_get_busy_time(engine, &now),
+				    p->busy);
+		p->time = ktime_sub(now, p->time);
+	} else {
+		p->time = ktime_sub(ktime_get(), p->time);
+	}
+
+	err = switch_to_kernel_sync(ce, err);
+	p->runtime = intel_context_get_total_runtime_ns(ce);
+	p->count = count;
+
+	intel_context_unpin(ce);
+	intel_context_put(ce);
+	thread->result = err;
+}
+
+static void p_many(struct kthread_work *work)
+{
+	struct p_thread *thread = container_of(work, typeof(*thread), work);
+	struct perf_stats *p = &thread->p;
+	struct intel_engine_cs *engine = p->engine;
+	struct intel_context *ce;
+	IGT_TIMEOUT(end_time);
+	unsigned long count;
+	int err = 0;
+	bool busy;
+
+	ce = intel_context_create(engine);
+	if (IS_ERR(ce)) {
+		thread->result = PTR_ERR(ce);
+		return;
+	}
+
+	err = intel_context_pin(ce);
+	if (err) {
+		intel_context_put(ce);
+		thread->result = err;
+		return;
+	}
+
+	if (intel_engine_supports_stats(engine)) {
+		p->busy = intel_engine_get_busy_time(engine, &p->time);
+		busy = true;
+	} else {
+		p->time = ktime_get();
+		busy = false;
+	}
+
+	count = 0;
+	do {
+		struct i915_request *rq;
+
+		rq = i915_request_create(ce);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			break;
+		}
+
+		i915_request_add(rq);
+		count++;
+	} while (!__igt_timeout(end_time, NULL));
+
+	if (busy) {
+		ktime_t now;
+
+		p->busy = ktime_sub(intel_engine_get_busy_time(engine, &now),
+				    p->busy);
+		p->time = ktime_sub(now, p->time);
+	} else {
+		p->time = ktime_sub(ktime_get(), p->time);
+	}
+
+	err = switch_to_kernel_sync(ce, err);
+	p->runtime = intel_context_get_total_runtime_ns(ce);
+	p->count = count;
+
+	intel_context_unpin(ce);
+	intel_context_put(ce);
+	thread->result = err;
+}
+
+static int perf_parallel_engines(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	static void (* const func[])(struct kthread_work *) = {
+		p_sync0,
+		p_sync1,
+		p_many,
+		NULL,
+	};
+	const unsigned int nengines = num_uabi_engines(i915);
+	void (* const *fn)(struct kthread_work *);
+	struct intel_engine_cs *engine;
+	struct pm_qos_request qos;
+	struct p_thread *engines;
+	int err = 0;
+
+	engines = kcalloc(nengines, sizeof(*engines), GFP_KERNEL);
+	if (!engines)
+		return -ENOMEM;
+
+	cpu_latency_qos_add_request(&qos, 0);
+
+	for (fn = func; *fn; fn++) {
+		char name[KSYM_NAME_LEN];
+		struct igt_live_test t;
+		unsigned int idx;
+
+		snprintf(name, sizeof(name), "%ps", *fn);
+		err = igt_live_test_begin(&t, i915, __func__, name);
+		if (err)
+			break;
+
+		atomic_set(&i915->selftest.counter, nengines);
+
+		idx = 0;
+		for_each_uabi_engine(engine, i915) {
+			struct kthread_worker *worker;
+
+			intel_engine_pm_get(engine);
+
+			memset(&engines[idx].p, 0, sizeof(engines[idx].p));
+
+			worker = kthread_create_worker(0, "igt:%s",
+						       engine->name);
+			if (IS_ERR(worker)) {
+				err = PTR_ERR(worker);
+				intel_engine_pm_put(engine);
+				break;
+			}
+			engines[idx].worker = worker;
+			engines[idx].result = 0;
+			engines[idx].p.engine = engine;
+			engines[idx].engine = engine;
+
+			kthread_init_work(&engines[idx].work, *fn);
+			kthread_queue_work(worker, &engines[idx].work);
+			idx++;
+		}
+
+		idx = 0;
+		for_each_uabi_engine(engine, i915) {
+			int status;
+
+			if (!engines[idx].worker)
+				break;
+
+			kthread_flush_work(&engines[idx].work);
+			status = READ_ONCE(engines[idx].result);
+			if (status && !err)
+				err = status;
+
+			intel_engine_pm_put(engine);
+
+			kthread_destroy_worker(engines[idx].worker);
+			idx++;
+		}
+
+		if (igt_live_test_end(&t))
+			err = -EIO;
+		if (err)
+			break;
+
+		idx = 0;
+		for_each_uabi_engine(engine, i915) {
+			struct perf_stats *p = &engines[idx].p;
+			u64 busy = 100 * ktime_to_ns(p->busy);
+			u64 dt = ktime_to_ns(p->time);
+			int integer, decimal;
+
+			if (dt) {
+				integer = div64_u64(busy, dt);
+				busy -= integer * dt;
+				decimal = div64_u64(100 * busy, dt);
+			} else {
+				integer = 0;
+				decimal = 0;
+			}
+
+			GEM_BUG_ON(engine != p->engine);
+			pr_info("%s %5s: { count:%lu, busy:%d.%02d%%, runtime:%lldms, walltime:%lldms }\n",
+				name, engine->name, p->count, integer, decimal,
+				div_u64(p->runtime, 1000 * 1000),
+				div_u64(ktime_to_ns(p->time), 1000 * 1000));
+			idx++;
+		}
+	}
+
+	cpu_latency_qos_remove_request(&qos);
+	kfree(engines);
+	return err;
+}
+
+int i915_request_perf_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(perf_request_latency),
+		SUBTEST(perf_series_engines),
+		SUBTEST(perf_parallel_engines),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return i915_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_selftest.c b/drivers/gpu/drm/i915/selftests/i915_selftest.c
new file mode 100644
index 000000000..39da0fb0d
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_selftest.c
@@ -0,0 +1,440 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/random.h>
+
+#include "gt/intel_gt_pm.h"
+#include "i915_driver.h"
+#include "i915_drv.h"
+#include "i915_selftest.h"
+
+#include "igt_flush_test.h"
+
+struct i915_selftest i915_selftest __read_mostly = {
+	.timeout_ms = 500,
+};
+
+int i915_mock_sanitycheck(void)
+{
+	pr_info(DRIVER_NAME ": %s() - ok!\n", __func__);
+	return 0;
+}
+
+int i915_live_sanitycheck(struct drm_i915_private *i915)
+{
+	pr_info("%s: %s() - ok!\n", i915->drm.driver->name, __func__);
+	return 0;
+}
+
+enum {
+#define selftest(name, func) mock_##name,
+#include "i915_mock_selftests.h"
+#undef selftest
+};
+
+enum {
+#define selftest(name, func) live_##name,
+#include "i915_live_selftests.h"
+#undef selftest
+};
+
+enum {
+#define selftest(name, func) perf_##name,
+#include "i915_perf_selftests.h"
+#undef selftest
+};
+
+struct selftest {
+	bool enabled;
+	const char *name;
+	union {
+		int (*mock)(void);
+		int (*live)(struct drm_i915_private *);
+	};
+};
+
+#define selftest(n, f) [mock_##n] = { .name = #n, { .mock = f } },
+static struct selftest mock_selftests[] = {
+#include "i915_mock_selftests.h"
+};
+#undef selftest
+
+#define selftest(n, f) [live_##n] = { .name = #n, { .live = f } },
+static struct selftest live_selftests[] = {
+#include "i915_live_selftests.h"
+};
+#undef selftest
+
+#define selftest(n, f) [perf_##n] = { .name = #n, { .live = f } },
+static struct selftest perf_selftests[] = {
+#include "i915_perf_selftests.h"
+};
+#undef selftest
+
+/* Embed the line number into the parameter name so that we can order tests */
+#define selftest(n, func) selftest_0(n, func, param(n))
+#define param(n) __PASTE(igt__, __PASTE(__LINE__, __mock_##n))
+#define selftest_0(n, func, id) \
+module_param_named(id, mock_selftests[mock_##n].enabled, bool, 0400);
+#include "i915_mock_selftests.h"
+#undef selftest_0
+#undef param
+
+#define param(n) __PASTE(igt__, __PASTE(__LINE__, __live_##n))
+#define selftest_0(n, func, id) \
+module_param_named(id, live_selftests[live_##n].enabled, bool, 0400);
+#include "i915_live_selftests.h"
+#undef selftest_0
+#undef param
+
+#define param(n) __PASTE(igt__, __PASTE(__LINE__, __perf_##n))
+#define selftest_0(n, func, id) \
+module_param_named(id, perf_selftests[perf_##n].enabled, bool, 0400);
+#include "i915_perf_selftests.h"
+#undef selftest_0
+#undef param
+#undef selftest
+
+static void set_default_test_all(struct selftest *st, unsigned int count)
+{
+	unsigned int i;
+
+	for (i = 0; i < count; i++)
+		if (st[i].enabled)
+			return;
+
+	for (i = 0; i < count; i++)
+		st[i].enabled = true;
+}
+
+static int __run_selftests(const char *name,
+			   struct selftest *st,
+			   unsigned int count,
+			   void *data)
+{
+	int err = 0;
+
+	while (!i915_selftest.random_seed)
+		i915_selftest.random_seed = get_random_u32();
+
+	i915_selftest.timeout_jiffies =
+		i915_selftest.timeout_ms ?
+		msecs_to_jiffies_timeout(i915_selftest.timeout_ms) :
+		MAX_SCHEDULE_TIMEOUT;
+
+	set_default_test_all(st, count);
+
+	pr_info(DRIVER_NAME ": Performing %s selftests with st_random_seed=0x%x st_timeout=%u\n",
+		name, i915_selftest.random_seed, i915_selftest.timeout_ms);
+
+	/* Tests are listed in order in i915_*_selftests.h */
+	for (; count--; st++) {
+		if (!st->enabled)
+			continue;
+
+		cond_resched();
+		if (signal_pending(current))
+			return -EINTR;
+
+		pr_info(DRIVER_NAME ": Running %s\n", st->name);
+		if (data)
+			err = st->live(data);
+		else
+			err = st->mock();
+		if (err == -EINTR && !signal_pending(current))
+			err = 0;
+		if (err)
+			break;
+	}
+
+	if (WARN(err > 0 || err == -ENOTTY,
+		 "%s returned %d, conflicting with selftest's magic values!\n",
+		 st->name, err))
+		err = -1;
+
+	return err;
+}
+
+#define run_selftests(x, data) \
+	__run_selftests(#x, x##_selftests, ARRAY_SIZE(x##_selftests), data)
+
+int i915_mock_selftests(void)
+{
+	int err;
+
+	if (!i915_selftest.mock)
+		return 0;
+
+	err = run_selftests(mock, NULL);
+	if (err) {
+		i915_selftest.mock = err;
+		return 1;
+	}
+
+	if (i915_selftest.mock < 0) {
+		i915_selftest.mock = -ENOTTY;
+		return 1;
+	}
+
+	return 0;
+}
+
+int i915_live_selftests(struct pci_dev *pdev)
+{
+	int err;
+
+	if (!i915_selftest.live)
+		return 0;
+
+	err = run_selftests(live, pdev_to_i915(pdev));
+	if (err) {
+		i915_selftest.live = err;
+		return err;
+	}
+
+	if (i915_selftest.live < 0) {
+		i915_selftest.live = -ENOTTY;
+		return 1;
+	}
+
+	return 0;
+}
+
+int i915_perf_selftests(struct pci_dev *pdev)
+{
+	int err;
+
+	if (!i915_selftest.perf)
+		return 0;
+
+	err = run_selftests(perf, pdev_to_i915(pdev));
+	if (err) {
+		i915_selftest.perf = err;
+		return err;
+	}
+
+	if (i915_selftest.perf < 0) {
+		i915_selftest.perf = -ENOTTY;
+		return 1;
+	}
+
+	return 0;
+}
+
+static bool apply_subtest_filter(const char *caller, const char *name)
+{
+	char *filter, *sep, *tok;
+	bool result = true;
+
+	filter = kstrdup(i915_selftest.filter, GFP_KERNEL);
+	for (sep = filter; (tok = strsep(&sep, ","));) {
+		bool allow = true;
+		char *sl;
+
+		if (*tok == '!') {
+			allow = false;
+			tok++;
+		}
+
+		if (*tok == '\0')
+			continue;
+
+		sl = strchr(tok, '/');
+		if (sl) {
+			*sl++ = '\0';
+			if (strcmp(tok, caller)) {
+				if (allow)
+					result = false;
+				continue;
+			}
+			tok = sl;
+		}
+
+		if (strcmp(tok, name)) {
+			if (allow)
+				result = false;
+			continue;
+		}
+
+		result = allow;
+		break;
+	}
+	kfree(filter);
+
+	return result;
+}
+
+int __i915_nop_setup(void *data)
+{
+	return 0;
+}
+
+int __i915_nop_teardown(int err, void *data)
+{
+	return err;
+}
+
+int __i915_live_setup(void *data)
+{
+	struct drm_i915_private *i915 = data;
+
+	/* The selftests expect an idle system */
+	if (intel_gt_pm_wait_for_idle(to_gt(i915)))
+		return -EIO;
+
+	return intel_gt_terminally_wedged(to_gt(i915));
+}
+
+int __i915_live_teardown(int err, void *data)
+{
+	struct drm_i915_private *i915 = data;
+
+	if (igt_flush_test(i915))
+		err = -EIO;
+
+	i915_gem_drain_freed_objects(i915);
+
+	return err;
+}
+
+int __intel_gt_live_setup(void *data)
+{
+	struct intel_gt *gt = data;
+
+	/* The selftests expect an idle system */
+	if (intel_gt_pm_wait_for_idle(gt))
+		return -EIO;
+
+	return intel_gt_terminally_wedged(gt);
+}
+
+int __intel_gt_live_teardown(int err, void *data)
+{
+	struct intel_gt *gt = data;
+
+	if (igt_flush_test(gt->i915))
+		err = -EIO;
+
+	i915_gem_drain_freed_objects(gt->i915);
+
+	return err;
+}
+
+int __i915_subtests(const char *caller,
+		    int (*setup)(void *data),
+		    int (*teardown)(int err, void *data),
+		    const struct i915_subtest *st,
+		    unsigned int count,
+		    void *data)
+{
+	int err;
+
+	for (; count--; st++) {
+		cond_resched();
+		if (signal_pending(current))
+			return -EINTR;
+
+		if (!apply_subtest_filter(caller, st->name))
+			continue;
+
+		err = setup(data);
+		if (err) {
+			pr_err(DRIVER_NAME "/%s: setup failed for %s\n",
+			       caller, st->name);
+			return err;
+		}
+
+		pr_info(DRIVER_NAME ": Running %s/%s\n", caller, st->name);
+		GEM_TRACE("Running %s/%s\n", caller, st->name);
+
+		err = teardown(st->func(data), data);
+		if (err && err != -EINTR) {
+			pr_err(DRIVER_NAME "/%s: %s failed with error %d\n",
+			       caller, st->name, err);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+bool __igt_timeout(unsigned long timeout, const char *fmt, ...)
+{
+	va_list va;
+
+	if (!signal_pending(current)) {
+		cond_resched();
+		if (time_before(jiffies, timeout))
+			return false;
+	}
+
+	if (fmt) {
+		va_start(va, fmt);
+		vprintk(fmt, va);
+		va_end(va);
+	}
+
+	return true;
+}
+
+void igt_hexdump(const void *buf, size_t len)
+{
+	const size_t rowsize = 8 * sizeof(u32);
+	const void *prev = NULL;
+	bool skip = false;
+	size_t pos;
+
+	for (pos = 0; pos < len; pos += rowsize) {
+		char line[128];
+
+		if (prev && !memcmp(prev, buf + pos, rowsize)) {
+			if (!skip) {
+				pr_info("*\n");
+				skip = true;
+			}
+			continue;
+		}
+
+		WARN_ON_ONCE(hex_dump_to_buffer(buf + pos, len - pos,
+						rowsize, sizeof(u32),
+						line, sizeof(line),
+						false) >= sizeof(line));
+		pr_info("[%04zx] %s\n", pos, line);
+
+		prev = buf + pos;
+		skip = false;
+	}
+}
+
+module_param_named(st_random_seed, i915_selftest.random_seed, uint, 0400);
+module_param_named(st_timeout, i915_selftest.timeout_ms, uint, 0400);
+module_param_named(st_filter, i915_selftest.filter, charp, 0400);
+
+module_param_named_unsafe(mock_selftests, i915_selftest.mock, int, 0400);
+MODULE_PARM_DESC(mock_selftests, "Run selftests before loading, using mock hardware (0:disabled [default], 1:run tests then load driver, -1:run tests then leave dummy module)");
+
+module_param_named_unsafe(live_selftests, i915_selftest.live, int, 0400);
+MODULE_PARM_DESC(live_selftests, "Run selftests after driver initialisation on the live system (0:disabled [default], 1:run tests then continue, -1:run tests then exit module)");
+
+module_param_named_unsafe(perf_selftests, i915_selftest.perf, int, 0400);
+MODULE_PARM_DESC(perf_selftests, "Run performance orientated selftests after driver initialisation on the live system (0:disabled [default], 1:run tests then continue, -1:run tests then exit module)");
diff --git a/drivers/gpu/drm/i915/selftests/i915_sw_fence.c b/drivers/gpu/drm/i915/selftests/i915_sw_fence.c
new file mode 100644
index 000000000..daa985e5a
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_sw_fence.c
@@ -0,0 +1,757 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/completion.h>
+#include <linux/delay.h>
+#include <linux/prime_numbers.h>
+
+#include "../i915_selftest.h"
+
+static int
+fence_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
+{
+	switch (state) {
+	case FENCE_COMPLETE:
+		break;
+
+	case FENCE_FREE:
+		/* Leave the fence for the caller to free it after testing */
+		break;
+	}
+
+	return NOTIFY_DONE;
+}
+
+static struct i915_sw_fence *alloc_fence(void)
+{
+	struct i915_sw_fence *fence;
+
+	fence = kmalloc(sizeof(*fence), GFP_KERNEL);
+	if (!fence)
+		return NULL;
+
+	i915_sw_fence_init(fence, fence_notify);
+	return fence;
+}
+
+static void free_fence(struct i915_sw_fence *fence)
+{
+	i915_sw_fence_fini(fence);
+	kfree(fence);
+}
+
+static int __test_self(struct i915_sw_fence *fence)
+{
+	if (i915_sw_fence_done(fence))
+		return -EINVAL;
+
+	i915_sw_fence_commit(fence);
+	if (!i915_sw_fence_done(fence))
+		return -EINVAL;
+
+	i915_sw_fence_wait(fence);
+	if (!i915_sw_fence_done(fence))
+		return -EINVAL;
+
+	return 0;
+}
+
+static int test_self(void *arg)
+{
+	struct i915_sw_fence *fence;
+	int ret;
+
+	/* Test i915_sw_fence signaling and completion testing */
+	fence = alloc_fence();
+	if (!fence)
+		return -ENOMEM;
+
+	ret = __test_self(fence);
+
+	free_fence(fence);
+	return ret;
+}
+
+static int test_dag(void *arg)
+{
+	struct i915_sw_fence *A, *B, *C;
+	int ret = -EINVAL;
+
+	/* Test detection of cycles within the i915_sw_fence graphs */
+	if (!IS_ENABLED(CONFIG_DRM_I915_SW_FENCE_CHECK_DAG))
+		return 0;
+
+	A = alloc_fence();
+	if (!A)
+		return -ENOMEM;
+
+	if (i915_sw_fence_await_sw_fence_gfp(A, A, GFP_KERNEL) != -EINVAL) {
+		pr_err("recursive cycle not detected (AA)\n");
+		goto err_A;
+	}
+
+	B = alloc_fence();
+	if (!B) {
+		ret = -ENOMEM;
+		goto err_A;
+	}
+
+	i915_sw_fence_await_sw_fence_gfp(A, B, GFP_KERNEL);
+	if (i915_sw_fence_await_sw_fence_gfp(B, A, GFP_KERNEL) != -EINVAL) {
+		pr_err("single depth cycle not detected (BAB)\n");
+		goto err_B;
+	}
+
+	C = alloc_fence();
+	if (!C) {
+		ret = -ENOMEM;
+		goto err_B;
+	}
+
+	if (i915_sw_fence_await_sw_fence_gfp(B, C, GFP_KERNEL) == -EINVAL) {
+		pr_err("invalid cycle detected\n");
+		goto err_C;
+	}
+	if (i915_sw_fence_await_sw_fence_gfp(C, B, GFP_KERNEL) != -EINVAL) {
+		pr_err("single depth cycle not detected (CBC)\n");
+		goto err_C;
+	}
+	if (i915_sw_fence_await_sw_fence_gfp(C, A, GFP_KERNEL) != -EINVAL) {
+		pr_err("cycle not detected (BA, CB, AC)\n");
+		goto err_C;
+	}
+	if (i915_sw_fence_await_sw_fence_gfp(A, C, GFP_KERNEL) == -EINVAL) {
+		pr_err("invalid cycle detected\n");
+		goto err_C;
+	}
+
+	i915_sw_fence_commit(A);
+	i915_sw_fence_commit(B);
+	i915_sw_fence_commit(C);
+
+	ret = 0;
+	if (!i915_sw_fence_done(C)) {
+		pr_err("fence C not done\n");
+		ret = -EINVAL;
+	}
+	if (!i915_sw_fence_done(B)) {
+		pr_err("fence B not done\n");
+		ret = -EINVAL;
+	}
+	if (!i915_sw_fence_done(A)) {
+		pr_err("fence A not done\n");
+		ret = -EINVAL;
+	}
+err_C:
+	free_fence(C);
+err_B:
+	free_fence(B);
+err_A:
+	free_fence(A);
+	return ret;
+}
+
+static int test_AB(void *arg)
+{
+	struct i915_sw_fence *A, *B;
+	int ret;
+
+	/* Test i915_sw_fence (A) waiting on an event source (B) */
+	A = alloc_fence();
+	if (!A)
+		return -ENOMEM;
+	B = alloc_fence();
+	if (!B) {
+		ret = -ENOMEM;
+		goto err_A;
+	}
+
+	ret = i915_sw_fence_await_sw_fence_gfp(A, B, GFP_KERNEL);
+	if (ret < 0)
+		goto err_B;
+	if (ret == 0) {
+		pr_err("Incorrectly reported fence A was complete before await\n");
+		ret = -EINVAL;
+		goto err_B;
+	}
+
+	ret = -EINVAL;
+	i915_sw_fence_commit(A);
+	if (i915_sw_fence_done(A))
+		goto err_B;
+
+	i915_sw_fence_commit(B);
+	if (!i915_sw_fence_done(B)) {
+		pr_err("Fence B is not done\n");
+		goto err_B;
+	}
+
+	if (!i915_sw_fence_done(A)) {
+		pr_err("Fence A is not done\n");
+		goto err_B;
+	}
+
+	ret = 0;
+err_B:
+	free_fence(B);
+err_A:
+	free_fence(A);
+	return ret;
+}
+
+static int test_ABC(void *arg)
+{
+	struct i915_sw_fence *A, *B, *C;
+	int ret;
+
+	/* Test a chain of fences, A waits on B who waits on C */
+	A = alloc_fence();
+	if (!A)
+		return -ENOMEM;
+
+	B = alloc_fence();
+	if (!B) {
+		ret = -ENOMEM;
+		goto err_A;
+	}
+
+	C = alloc_fence();
+	if (!C) {
+		ret = -ENOMEM;
+		goto err_B;
+	}
+
+	ret = i915_sw_fence_await_sw_fence_gfp(A, B, GFP_KERNEL);
+	if (ret < 0)
+		goto err_C;
+	if (ret == 0) {
+		pr_err("Incorrectly reported fence B was complete before await\n");
+		goto err_C;
+	}
+
+	ret = i915_sw_fence_await_sw_fence_gfp(B, C, GFP_KERNEL);
+	if (ret < 0)
+		goto err_C;
+	if (ret == 0) {
+		pr_err("Incorrectly reported fence C was complete before await\n");
+		goto err_C;
+	}
+
+	ret = -EINVAL;
+	i915_sw_fence_commit(A);
+	if (i915_sw_fence_done(A)) {
+		pr_err("Fence A completed early\n");
+		goto err_C;
+	}
+
+	i915_sw_fence_commit(B);
+	if (i915_sw_fence_done(B)) {
+		pr_err("Fence B completed early\n");
+		goto err_C;
+	}
+
+	if (i915_sw_fence_done(A)) {
+		pr_err("Fence A completed early (after signaling B)\n");
+		goto err_C;
+	}
+
+	i915_sw_fence_commit(C);
+
+	ret = 0;
+	if (!i915_sw_fence_done(C)) {
+		pr_err("Fence C not done\n");
+		ret = -EINVAL;
+	}
+	if (!i915_sw_fence_done(B)) {
+		pr_err("Fence B not done\n");
+		ret = -EINVAL;
+	}
+	if (!i915_sw_fence_done(A)) {
+		pr_err("Fence A not done\n");
+		ret = -EINVAL;
+	}
+err_C:
+	free_fence(C);
+err_B:
+	free_fence(B);
+err_A:
+	free_fence(A);
+	return ret;
+}
+
+static int test_AB_C(void *arg)
+{
+	struct i915_sw_fence *A, *B, *C;
+	int ret = -EINVAL;
+
+	/* Test multiple fences (AB) waiting on a single event (C) */
+	A = alloc_fence();
+	if (!A)
+		return -ENOMEM;
+
+	B = alloc_fence();
+	if (!B) {
+		ret = -ENOMEM;
+		goto err_A;
+	}
+
+	C = alloc_fence();
+	if (!C) {
+		ret = -ENOMEM;
+		goto err_B;
+	}
+
+	ret = i915_sw_fence_await_sw_fence_gfp(A, C, GFP_KERNEL);
+	if (ret < 0)
+		goto err_C;
+	if (ret == 0) {
+		ret = -EINVAL;
+		goto err_C;
+	}
+
+	ret = i915_sw_fence_await_sw_fence_gfp(B, C, GFP_KERNEL);
+	if (ret < 0)
+		goto err_C;
+	if (ret == 0) {
+		ret = -EINVAL;
+		goto err_C;
+	}
+
+	i915_sw_fence_commit(A);
+	i915_sw_fence_commit(B);
+
+	ret = 0;
+	if (i915_sw_fence_done(A)) {
+		pr_err("Fence A completed early\n");
+		ret = -EINVAL;
+	}
+
+	if (i915_sw_fence_done(B)) {
+		pr_err("Fence B completed early\n");
+		ret = -EINVAL;
+	}
+
+	i915_sw_fence_commit(C);
+	if (!i915_sw_fence_done(C)) {
+		pr_err("Fence C not done\n");
+		ret = -EINVAL;
+	}
+
+	if (!i915_sw_fence_done(B)) {
+		pr_err("Fence B not done\n");
+		ret = -EINVAL;
+	}
+
+	if (!i915_sw_fence_done(A)) {
+		pr_err("Fence A not done\n");
+		ret = -EINVAL;
+	}
+
+err_C:
+	free_fence(C);
+err_B:
+	free_fence(B);
+err_A:
+	free_fence(A);
+	return ret;
+}
+
+static int test_C_AB(void *arg)
+{
+	struct i915_sw_fence *A, *B, *C;
+	int ret;
+
+	/* Test multiple event sources (A,B) for a single fence (C) */
+	A = alloc_fence();
+	if (!A)
+		return -ENOMEM;
+
+	B = alloc_fence();
+	if (!B) {
+		ret = -ENOMEM;
+		goto err_A;
+	}
+
+	C = alloc_fence();
+	if (!C) {
+		ret = -ENOMEM;
+		goto err_B;
+	}
+
+	ret = i915_sw_fence_await_sw_fence_gfp(C, A, GFP_KERNEL);
+	if (ret < 0)
+		goto err_C;
+	if (ret == 0) {
+		ret = -EINVAL;
+		goto err_C;
+	}
+
+	ret = i915_sw_fence_await_sw_fence_gfp(C, B, GFP_KERNEL);
+	if (ret < 0)
+		goto err_C;
+	if (ret == 0) {
+		ret = -EINVAL;
+		goto err_C;
+	}
+
+	ret = 0;
+	i915_sw_fence_commit(C);
+	if (i915_sw_fence_done(C))
+		ret = -EINVAL;
+
+	i915_sw_fence_commit(A);
+	i915_sw_fence_commit(B);
+
+	if (!i915_sw_fence_done(A)) {
+		pr_err("Fence A not done\n");
+		ret = -EINVAL;
+	}
+
+	if (!i915_sw_fence_done(B)) {
+		pr_err("Fence B not done\n");
+		ret = -EINVAL;
+	}
+
+	if (!i915_sw_fence_done(C)) {
+		pr_err("Fence C not done\n");
+		ret = -EINVAL;
+	}
+
+err_C:
+	free_fence(C);
+err_B:
+	free_fence(B);
+err_A:
+	free_fence(A);
+	return ret;
+}
+
+static int test_chain(void *arg)
+{
+	int nfences = 4096;
+	struct i915_sw_fence **fences;
+	int ret, i;
+
+	/* Test a long chain of fences */
+	fences = kmalloc_array(nfences, sizeof(*fences), GFP_KERNEL);
+	if (!fences)
+		return -ENOMEM;
+
+	for (i = 0; i < nfences; i++) {
+		fences[i] = alloc_fence();
+		if (!fences[i]) {
+			nfences = i;
+			ret = -ENOMEM;
+			goto err;
+		}
+
+		if (i > 0) {
+			ret = i915_sw_fence_await_sw_fence_gfp(fences[i],
+							       fences[i - 1],
+							       GFP_KERNEL);
+			if (ret < 0) {
+				nfences = i + 1;
+				goto err;
+			}
+
+			i915_sw_fence_commit(fences[i]);
+		}
+	}
+
+	ret = 0;
+	for (i = nfences; --i; ) {
+		if (i915_sw_fence_done(fences[i])) {
+			if (ret == 0)
+				pr_err("Fence[%d] completed early\n", i);
+			ret = -EINVAL;
+		}
+	}
+	i915_sw_fence_commit(fences[0]);
+	for (i = 0; ret == 0 && i < nfences; i++) {
+		if (!i915_sw_fence_done(fences[i])) {
+			pr_err("Fence[%d] is not done\n", i);
+			ret = -EINVAL;
+		}
+	}
+
+err:
+	for (i = 0; i < nfences; i++)
+		free_fence(fences[i]);
+	kfree(fences);
+	return ret;
+}
+
+struct task_ipc {
+	struct work_struct work;
+	struct completion started;
+	struct i915_sw_fence *in, *out;
+	int value;
+};
+
+static void task_ipc(struct work_struct *work)
+{
+	struct task_ipc *ipc = container_of(work, typeof(*ipc), work);
+
+	complete(&ipc->started);
+
+	i915_sw_fence_wait(ipc->in);
+	smp_store_mb(ipc->value, 1);
+	i915_sw_fence_commit(ipc->out);
+}
+
+static int test_ipc(void *arg)
+{
+	struct task_ipc ipc;
+	int ret = 0;
+
+	/* Test use of i915_sw_fence as an interprocess signaling mechanism */
+	ipc.in = alloc_fence();
+	if (!ipc.in)
+		return -ENOMEM;
+	ipc.out = alloc_fence();
+	if (!ipc.out) {
+		ret = -ENOMEM;
+		goto err_in;
+	}
+
+	/* use a completion to avoid chicken-and-egg testing */
+	init_completion(&ipc.started);
+
+	ipc.value = 0;
+	INIT_WORK_ONSTACK(&ipc.work, task_ipc);
+	schedule_work(&ipc.work);
+
+	wait_for_completion(&ipc.started);
+
+	usleep_range(1000, 2000);
+	if (READ_ONCE(ipc.value)) {
+		pr_err("worker updated value before i915_sw_fence was signaled\n");
+		ret = -EINVAL;
+	}
+
+	i915_sw_fence_commit(ipc.in);
+	i915_sw_fence_wait(ipc.out);
+
+	if (!READ_ONCE(ipc.value)) {
+		pr_err("worker signaled i915_sw_fence before value was posted\n");
+		ret = -EINVAL;
+	}
+
+	flush_work(&ipc.work);
+	destroy_work_on_stack(&ipc.work);
+	free_fence(ipc.out);
+err_in:
+	free_fence(ipc.in);
+	return ret;
+}
+
+static int test_timer(void *arg)
+{
+	unsigned long target, delay;
+	struct timed_fence tf;
+
+	preempt_disable();
+	timed_fence_init(&tf, target = jiffies);
+	if (!i915_sw_fence_done(&tf.fence)) {
+		pr_err("Fence with immediate expiration not signaled\n");
+		goto err;
+	}
+	preempt_enable();
+	timed_fence_fini(&tf);
+
+	for_each_prime_number(delay, i915_selftest.timeout_jiffies/2) {
+		preempt_disable();
+		timed_fence_init(&tf, target = jiffies + delay);
+		if (i915_sw_fence_done(&tf.fence)) {
+			pr_err("Fence with future expiration (%lu jiffies) already signaled\n", delay);
+			goto err;
+		}
+		preempt_enable();
+
+		i915_sw_fence_wait(&tf.fence);
+
+		preempt_disable();
+		if (!i915_sw_fence_done(&tf.fence)) {
+			pr_err("Fence not signaled after wait\n");
+			goto err;
+		}
+		if (time_before(jiffies, target)) {
+			pr_err("Fence signaled too early, target=%lu, now=%lu\n",
+			       target, jiffies);
+			goto err;
+		}
+		preempt_enable();
+		timed_fence_fini(&tf);
+	}
+
+	return 0;
+
+err:
+	preempt_enable();
+	timed_fence_fini(&tf);
+	return -EINVAL;
+}
+
+static const char *mock_name(struct dma_fence *fence)
+{
+	return "mock";
+}
+
+static const struct dma_fence_ops mock_fence_ops = {
+	.get_driver_name = mock_name,
+	.get_timeline_name = mock_name,
+};
+
+static DEFINE_SPINLOCK(mock_fence_lock);
+
+static struct dma_fence *alloc_dma_fence(void)
+{
+	struct dma_fence *dma;
+
+	dma = kmalloc(sizeof(*dma), GFP_KERNEL);
+	if (dma)
+		dma_fence_init(dma, &mock_fence_ops, &mock_fence_lock, 0, 0);
+
+	return dma;
+}
+
+static struct i915_sw_fence *
+wrap_dma_fence(struct dma_fence *dma, unsigned long delay)
+{
+	struct i915_sw_fence *fence;
+	int err;
+
+	fence = alloc_fence();
+	if (!fence)
+		return ERR_PTR(-ENOMEM);
+
+	err = i915_sw_fence_await_dma_fence(fence, dma, delay, GFP_NOWAIT);
+	i915_sw_fence_commit(fence);
+	if (err < 0) {
+		free_fence(fence);
+		return ERR_PTR(err);
+	}
+
+	return fence;
+}
+
+static int test_dma_fence(void *arg)
+{
+	struct i915_sw_fence *timeout = NULL, *not = NULL;
+	unsigned long delay = i915_selftest.timeout_jiffies;
+	unsigned long end, sleep;
+	struct dma_fence *dma;
+	int err;
+
+	dma = alloc_dma_fence();
+	if (!dma)
+		return -ENOMEM;
+
+	timeout = wrap_dma_fence(dma, delay);
+	if (IS_ERR(timeout)) {
+		err = PTR_ERR(timeout);
+		goto err;
+	}
+
+	not = wrap_dma_fence(dma, 0);
+	if (IS_ERR(not)) {
+		err = PTR_ERR(not);
+		goto err;
+	}
+
+	err = -EINVAL;
+	if (i915_sw_fence_done(timeout) || i915_sw_fence_done(not)) {
+		pr_err("Fences immediately signaled\n");
+		goto err;
+	}
+
+	/* We round the timeout for the fence up to the next second */
+	end = round_jiffies_up(jiffies + delay);
+
+	sleep = jiffies_to_usecs(delay) / 3;
+	usleep_range(sleep, 2 * sleep);
+	if (time_after(jiffies, end)) {
+		pr_debug("Slept too long, delay=%lu, (target=%lu, now=%lu) skipping\n",
+			 delay, end, jiffies);
+		goto skip;
+	}
+
+	if (i915_sw_fence_done(timeout) || i915_sw_fence_done(not)) {
+		pr_err("Fences signaled too early\n");
+		goto err;
+	}
+
+	if (!wait_event_timeout(timeout->wait,
+				i915_sw_fence_done(timeout),
+				2 * (end - jiffies) + 1)) {
+		pr_err("Timeout fence unsignaled!\n");
+		goto err;
+	}
+
+	if (i915_sw_fence_done(not)) {
+		pr_err("No timeout fence signaled!\n");
+		goto err;
+	}
+
+skip:
+	dma_fence_signal(dma);
+
+	if (!i915_sw_fence_done(timeout) || !i915_sw_fence_done(not)) {
+		pr_err("Fences unsignaled\n");
+		goto err;
+	}
+
+	free_fence(not);
+	free_fence(timeout);
+	dma_fence_put(dma);
+
+	return 0;
+
+err:
+	dma_fence_signal(dma);
+	if (!IS_ERR_OR_NULL(timeout))
+		free_fence(timeout);
+	if (!IS_ERR_OR_NULL(not))
+		free_fence(not);
+	dma_fence_put(dma);
+	return err;
+}
+
+int i915_sw_fence_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(test_self),
+		SUBTEST(test_dag),
+		SUBTEST(test_AB),
+		SUBTEST(test_ABC),
+		SUBTEST(test_AB_C),
+		SUBTEST(test_C_AB),
+		SUBTEST(test_chain),
+		SUBTEST(test_ipc),
+		SUBTEST(test_timer),
+		SUBTEST(test_dma_fence),
+	};
+
+	return i915_subtests(tests, NULL);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_syncmap.c b/drivers/gpu/drm/i915/selftests/i915_syncmap.c
new file mode 100644
index 000000000..47f4ae18a
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_syncmap.c
@@ -0,0 +1,616 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "../i915_selftest.h"
+#include "i915_random.h"
+
+static char *
+__sync_print(struct i915_syncmap *p,
+	     char *buf, unsigned long *sz,
+	     unsigned int depth,
+	     unsigned int last,
+	     unsigned int idx)
+{
+	unsigned long len;
+	unsigned int i, X;
+
+	if (depth) {
+		unsigned int d;
+
+		for (d = 0; d < depth - 1; d++) {
+			if (last & BIT(depth - d - 1))
+				len = scnprintf(buf, *sz, "|   ");
+			else
+				len = scnprintf(buf, *sz, "    ");
+			buf += len;
+			*sz -= len;
+		}
+		len = scnprintf(buf, *sz, "%x-> ", idx);
+		buf += len;
+		*sz -= len;
+	}
+
+	/* We mark bits after the prefix as "X" */
+	len = scnprintf(buf, *sz, "0x%016llx", p->prefix << p->height << SHIFT);
+	buf += len;
+	*sz -= len;
+	X = (p->height + SHIFT) / 4;
+	scnprintf(buf - X, *sz + X, "%*s", X, "XXXXXXXXXXXXXXXXX");
+
+	if (!p->height) {
+		for_each_set_bit(i, (unsigned long *)&p->bitmap, KSYNCMAP) {
+			len = scnprintf(buf, *sz, " %x:%x,",
+					i, __sync_seqno(p)[i]);
+			buf += len;
+			*sz -= len;
+		}
+		buf -= 1;
+		*sz += 1;
+	}
+
+	len = scnprintf(buf, *sz, "\n");
+	buf += len;
+	*sz -= len;
+
+	if (p->height) {
+		for_each_set_bit(i, (unsigned long *)&p->bitmap, KSYNCMAP) {
+			buf = __sync_print(__sync_child(p)[i], buf, sz,
+					   depth + 1,
+					   last << 1 | !!(p->bitmap >> (i + 1)),
+					   i);
+		}
+	}
+
+	return buf;
+}
+
+static bool
+i915_syncmap_print_to_buf(struct i915_syncmap *p, char *buf, unsigned long sz)
+{
+	if (!p)
+		return false;
+
+	while (p->parent)
+		p = p->parent;
+
+	__sync_print(p, buf, &sz, 0, 1, 0);
+	return true;
+}
+
+static int check_syncmap_free(struct i915_syncmap **sync)
+{
+	i915_syncmap_free(sync);
+	if (*sync) {
+		pr_err("sync not cleared after free\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int dump_syncmap(struct i915_syncmap *sync, int err)
+{
+	char *buf;
+
+	if (!err)
+		return check_syncmap_free(&sync);
+
+	buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
+	if (!buf)
+		goto skip;
+
+	if (i915_syncmap_print_to_buf(sync, buf, PAGE_SIZE))
+		pr_err("%s", buf);
+
+	kfree(buf);
+
+skip:
+	i915_syncmap_free(&sync);
+	return err;
+}
+
+static int igt_syncmap_init(void *arg)
+{
+	struct i915_syncmap *sync = (void *)~0ul;
+
+	/*
+	 * Cursory check that we can initialise a random pointer and transform
+	 * it into the root pointer of a syncmap.
+	 */
+
+	i915_syncmap_init(&sync);
+	return check_syncmap_free(&sync);
+}
+
+static int check_seqno(struct i915_syncmap *leaf, unsigned int idx, u32 seqno)
+{
+	if (leaf->height) {
+		pr_err("%s: not a leaf, height is %d\n",
+		       __func__, leaf->height);
+		return -EINVAL;
+	}
+
+	if (__sync_seqno(leaf)[idx] != seqno) {
+		pr_err("%s: seqno[%d], found %x, expected %x\n",
+		       __func__, idx, __sync_seqno(leaf)[idx], seqno);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int check_one(struct i915_syncmap **sync, u64 context, u32 seqno)
+{
+	int err;
+
+	err = i915_syncmap_set(sync, context, seqno);
+	if (err)
+		return err;
+
+	if ((*sync)->height) {
+		pr_err("Inserting first context=%llx did not return leaf (height=%d, prefix=%llx\n",
+		       context, (*sync)->height, (*sync)->prefix);
+		return -EINVAL;
+	}
+
+	if ((*sync)->parent) {
+		pr_err("Inserting first context=%llx created branches!\n",
+		       context);
+		return -EINVAL;
+	}
+
+	if (hweight32((*sync)->bitmap) != 1) {
+		pr_err("First bitmap does not contain a single entry, found %x (count=%d)!\n",
+		       (*sync)->bitmap, hweight32((*sync)->bitmap));
+		return -EINVAL;
+	}
+
+	err = check_seqno((*sync), ilog2((*sync)->bitmap), seqno);
+	if (err)
+		return err;
+
+	if (!i915_syncmap_is_later(sync, context, seqno)) {
+		pr_err("Lookup of first context=%llx/seqno=%x failed!\n",
+		       context, seqno);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int igt_syncmap_one(void *arg)
+{
+	I915_RND_STATE(prng);
+	IGT_TIMEOUT(end_time);
+	struct i915_syncmap *sync;
+	unsigned long max = 1;
+	int err;
+
+	/*
+	 * Check that inserting a new id, creates a leaf and only that leaf.
+	 */
+
+	i915_syncmap_init(&sync);
+
+	do {
+		u64 context = i915_prandom_u64_state(&prng);
+		unsigned long loop;
+
+		err = check_syncmap_free(&sync);
+		if (err)
+			goto out;
+
+		for (loop = 0; loop <= max; loop++) {
+			err = check_one(&sync, context,
+					prandom_u32_state(&prng));
+			if (err)
+				goto out;
+		}
+		max++;
+	} while (!__igt_timeout(end_time, NULL));
+	pr_debug("%s: Completed %lu single insertions\n",
+		 __func__, max * (max - 1) / 2);
+out:
+	return dump_syncmap(sync, err);
+}
+
+static int check_leaf(struct i915_syncmap **sync, u64 context, u32 seqno)
+{
+	int err;
+
+	err = i915_syncmap_set(sync, context, seqno);
+	if (err)
+		return err;
+
+	if ((*sync)->height) {
+		pr_err("Inserting context=%llx did not return leaf (height=%d, prefix=%llx\n",
+		       context, (*sync)->height, (*sync)->prefix);
+		return -EINVAL;
+	}
+
+	if (hweight32((*sync)->bitmap) != 1) {
+		pr_err("First entry into leaf (context=%llx) does not contain a single entry, found %x (count=%d)!\n",
+		       context, (*sync)->bitmap, hweight32((*sync)->bitmap));
+		return -EINVAL;
+	}
+
+	err = check_seqno((*sync), ilog2((*sync)->bitmap), seqno);
+	if (err)
+		return err;
+
+	if (!i915_syncmap_is_later(sync, context, seqno)) {
+		pr_err("Lookup of first entry context=%llx/seqno=%x failed!\n",
+		       context, seqno);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int igt_syncmap_join_above(void *arg)
+{
+	struct i915_syncmap *sync;
+	unsigned int pass, order;
+	int err;
+
+	i915_syncmap_init(&sync);
+
+	/*
+	 * When we have a new id that doesn't fit inside the existing tree,
+	 * we need to add a new layer above.
+	 *
+	 * 1: 0x00000001
+	 * 2: 0x00000010
+	 * 3: 0x00000100
+	 * 4: 0x00001000
+	 * ...
+	 * Each pass the common prefix shrinks and we have to insert a join.
+	 * Each join will only contain two branches, the latest of which
+	 * is always a leaf.
+	 *
+	 * If we then reuse the same set of contexts, we expect to build an
+	 * identical tree.
+	 */
+	for (pass = 0; pass < 3; pass++) {
+		for (order = 0; order < 64; order += SHIFT) {
+			u64 context = BIT_ULL(order);
+			struct i915_syncmap *join;
+
+			err = check_leaf(&sync, context, 0);
+			if (err)
+				goto out;
+
+			join = sync->parent;
+			if (!join) /* very first insert will have no parents */
+				continue;
+
+			if (!join->height) {
+				pr_err("Parent with no height!\n");
+				err = -EINVAL;
+				goto out;
+			}
+
+			if (hweight32(join->bitmap) != 2) {
+				pr_err("Join does not have 2 children: %x (%d)\n",
+				       join->bitmap, hweight32(join->bitmap));
+				err = -EINVAL;
+				goto out;
+			}
+
+			if (__sync_child(join)[__sync_branch_idx(join, context)] != sync) {
+				pr_err("Leaf misplaced in parent!\n");
+				err = -EINVAL;
+				goto out;
+			}
+		}
+	}
+out:
+	return dump_syncmap(sync, err);
+}
+
+static int igt_syncmap_join_below(void *arg)
+{
+	struct i915_syncmap *sync;
+	unsigned int step, order, idx;
+	int err = -ENODEV;
+
+	i915_syncmap_init(&sync);
+
+	/*
+	 * Check that we can split a compacted branch by replacing it with
+	 * a join.
+	 */
+	for (step = 0; step < KSYNCMAP; step++) {
+		for (order = 64 - SHIFT; order > 0; order -= SHIFT) {
+			u64 context = step * BIT_ULL(order);
+
+			err = i915_syncmap_set(&sync, context, 0);
+			if (err)
+				goto out;
+
+			if (sync->height) {
+				pr_err("Inserting context=%llx (order=%d, step=%d) did not return leaf (height=%d, prefix=%llx\n",
+				       context, order, step, sync->height, sync->prefix);
+				err = -EINVAL;
+				goto out;
+			}
+		}
+	}
+
+	for (step = 0; step < KSYNCMAP; step++) {
+		for (order = SHIFT; order < 64; order += SHIFT) {
+			u64 context = step * BIT_ULL(order);
+
+			if (!i915_syncmap_is_later(&sync, context, 0)) {
+				pr_err("1: context %llx (order=%d, step=%d) not found\n",
+				       context, order, step);
+				err = -EINVAL;
+				goto out;
+			}
+
+			for (idx = 1; idx < KSYNCMAP; idx++) {
+				if (i915_syncmap_is_later(&sync, context + idx, 0)) {
+					pr_err("1: context %llx (order=%d, step=%d) should not exist\n",
+					       context + idx, order, step);
+					err = -EINVAL;
+					goto out;
+				}
+			}
+		}
+	}
+
+	for (order = SHIFT; order < 64; order += SHIFT) {
+		for (step = 0; step < KSYNCMAP; step++) {
+			u64 context = step * BIT_ULL(order);
+
+			if (!i915_syncmap_is_later(&sync, context, 0)) {
+				pr_err("2: context %llx (order=%d, step=%d) not found\n",
+				       context, order, step);
+				err = -EINVAL;
+				goto out;
+			}
+		}
+	}
+
+out:
+	return dump_syncmap(sync, err);
+}
+
+static int igt_syncmap_neighbours(void *arg)
+{
+	I915_RND_STATE(prng);
+	IGT_TIMEOUT(end_time);
+	struct i915_syncmap *sync;
+	int err = -ENODEV;
+
+	/*
+	 * Each leaf holds KSYNCMAP seqno. Check that when we create KSYNCMAP
+	 * neighbouring ids, they all fit into the same leaf.
+	 */
+
+	i915_syncmap_init(&sync);
+	do {
+		u64 context = i915_prandom_u64_state(&prng) & ~MASK;
+		unsigned int idx;
+
+		if (i915_syncmap_is_later(&sync, context, 0)) /* Skip repeats */
+			continue;
+
+		for (idx = 0; idx < KSYNCMAP; idx++) {
+			err = i915_syncmap_set(&sync, context + idx, 0);
+			if (err)
+				goto out;
+
+			if (sync->height) {
+				pr_err("Inserting context=%llx did not return leaf (height=%d, prefix=%llx\n",
+				       context, sync->height, sync->prefix);
+				err = -EINVAL;
+				goto out;
+			}
+
+			if (sync->bitmap != BIT(idx + 1) - 1) {
+				pr_err("Inserting neighbouring context=0x%llx+%d, did not fit into the same leaf bitmap=%x (%d), expected %lx (%d)\n",
+				       context, idx,
+				       sync->bitmap, hweight32(sync->bitmap),
+				       BIT(idx + 1) - 1, idx + 1);
+				err = -EINVAL;
+				goto out;
+			}
+		}
+	} while (!__igt_timeout(end_time, NULL));
+out:
+	return dump_syncmap(sync, err);
+}
+
+static int igt_syncmap_compact(void *arg)
+{
+	struct i915_syncmap *sync;
+	unsigned int idx, order;
+	int err = -ENODEV;
+
+	i915_syncmap_init(&sync);
+
+	/*
+	 * The syncmap are "space efficient" compressed radix trees - any
+	 * branch with only one child is skipped and replaced by the child.
+	 *
+	 * If we construct a tree with ids that are neighbouring at a non-zero
+	 * height, we form a join but each child of that join is directly a
+	 * leaf holding the single id.
+	 */
+	for (order = SHIFT; order < 64; order += SHIFT) {
+		err = check_syncmap_free(&sync);
+		if (err)
+			goto out;
+
+		/* Create neighbours in the parent */
+		for (idx = 0; idx < KSYNCMAP; idx++) {
+			u64 context = idx * BIT_ULL(order) + idx;
+
+			err = i915_syncmap_set(&sync, context, 0);
+			if (err)
+				goto out;
+
+			if (sync->height) {
+				pr_err("Inserting context=%llx (order=%d, idx=%d) did not return leaf (height=%d, prefix=%llx\n",
+				       context, order, idx,
+				       sync->height, sync->prefix);
+				err = -EINVAL;
+				goto out;
+			}
+		}
+
+		sync = sync->parent;
+		if (sync->parent) {
+			pr_err("Parent (join) of last leaf was not the sync!\n");
+			err = -EINVAL;
+			goto out;
+		}
+
+		if (sync->height != order) {
+			pr_err("Join does not have the expected height, found %d, expected %d\n",
+			       sync->height, order);
+			err = -EINVAL;
+			goto out;
+		}
+
+		if (sync->bitmap != BIT(KSYNCMAP) - 1) {
+			pr_err("Join is not full!, found %x (%d) expected %lx (%d)\n",
+			       sync->bitmap, hweight32(sync->bitmap),
+			       BIT(KSYNCMAP) - 1, KSYNCMAP);
+			err = -EINVAL;
+			goto out;
+		}
+
+		/* Each of our children should be a leaf */
+		for (idx = 0; idx < KSYNCMAP; idx++) {
+			struct i915_syncmap *leaf = __sync_child(sync)[idx];
+
+			if (leaf->height) {
+				pr_err("Child %d is a not leaf!\n", idx);
+				err = -EINVAL;
+				goto out;
+			}
+
+			if (leaf->parent != sync) {
+				pr_err("Child %d is not attached to us!\n",
+				       idx);
+				err = -EINVAL;
+				goto out;
+			}
+
+			if (!is_power_of_2(leaf->bitmap)) {
+				pr_err("Child %d holds more than one id, found %x (%d)\n",
+				       idx, leaf->bitmap, hweight32(leaf->bitmap));
+				err = -EINVAL;
+				goto out;
+			}
+
+			if (leaf->bitmap != BIT(idx)) {
+				pr_err("Child %d has wrong seqno idx, found %d, expected %d\n",
+				       idx, ilog2(leaf->bitmap), idx);
+				err = -EINVAL;
+				goto out;
+			}
+		}
+	}
+out:
+	return dump_syncmap(sync, err);
+}
+
+static int igt_syncmap_random(void *arg)
+{
+	I915_RND_STATE(prng);
+	IGT_TIMEOUT(end_time);
+	struct i915_syncmap *sync;
+	unsigned long count, phase, i;
+	u32 seqno;
+	int err;
+
+	i915_syncmap_init(&sync);
+
+	/*
+	 * Having tried to test the individual operations within i915_syncmap,
+	 * run a smoketest exploring the entire u64 space with random
+	 * insertions.
+	 */
+
+	count = 0;
+	phase = jiffies + HZ/100 + 1;
+	do {
+		u64 context = i915_prandom_u64_state(&prng);
+
+		err = i915_syncmap_set(&sync, context, 0);
+		if (err)
+			goto out;
+
+		count++;
+	} while (!time_after(jiffies, phase));
+	seqno = 0;
+
+	phase = 0;
+	do {
+		I915_RND_STATE(ctx);
+		u32 last_seqno = seqno;
+		bool expect;
+
+		seqno = prandom_u32_state(&prng);
+		expect = seqno_later(last_seqno, seqno);
+
+		for (i = 0; i < count; i++) {
+			u64 context = i915_prandom_u64_state(&ctx);
+
+			if (i915_syncmap_is_later(&sync, context, seqno) != expect) {
+				pr_err("context=%llu, last=%u this=%u did not match expectation (%d)\n",
+				       context, last_seqno, seqno, expect);
+				err = -EINVAL;
+				goto out;
+			}
+
+			err = i915_syncmap_set(&sync, context, seqno);
+			if (err)
+				goto out;
+		}
+
+		phase++;
+	} while (!__igt_timeout(end_time, NULL));
+	pr_debug("Completed %lu passes, each of %lu contexts\n", phase, count);
+out:
+	return dump_syncmap(sync, err);
+}
+
+int i915_syncmap_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_syncmap_init),
+		SUBTEST(igt_syncmap_one),
+		SUBTEST(igt_syncmap_join_above),
+		SUBTEST(igt_syncmap_join_below),
+		SUBTEST(igt_syncmap_neighbours),
+		SUBTEST(igt_syncmap_compact),
+		SUBTEST(igt_syncmap_random),
+	};
+
+	return i915_subtests(tests, NULL);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_vma.c b/drivers/gpu/drm/i915/selftests/i915_vma.c
new file mode 100644
index 000000000..71b52d5ef
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_vma.c
@@ -0,0 +1,1107 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/prime_numbers.h>
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_internal.h"
+#include "gem/selftests/mock_context.h"
+
+#include "i915_scatterlist.h"
+#include "i915_selftest.h"
+
+#include "mock_gem_device.h"
+#include "mock_gtt.h"
+
+static bool assert_vma(struct i915_vma *vma,
+		       struct drm_i915_gem_object *obj,
+		       struct i915_gem_context *ctx)
+{
+	bool ok = true;
+
+	if (vma->vm != ctx->vm) {
+		pr_err("VMA created with wrong VM\n");
+		ok = false;
+	}
+
+	if (vma->size != obj->base.size) {
+		pr_err("VMA created with wrong size, found %llu, expected %zu\n",
+		       vma->size, obj->base.size);
+		ok = false;
+	}
+
+	if (vma->gtt_view.type != I915_GTT_VIEW_NORMAL) {
+		pr_err("VMA created with wrong type [%d]\n",
+		       vma->gtt_view.type);
+		ok = false;
+	}
+
+	return ok;
+}
+
+static struct i915_vma *
+checked_vma_instance(struct drm_i915_gem_object *obj,
+		     struct i915_address_space *vm,
+		     const struct i915_gtt_view *view)
+{
+	struct i915_vma *vma;
+	bool ok = true;
+
+	vma = i915_vma_instance(obj, vm, view);
+	if (IS_ERR(vma))
+		return vma;
+
+	/* Manual checks, will be reinforced by i915_vma_compare! */
+	if (vma->vm != vm) {
+		pr_err("VMA's vm [%p] does not match request [%p]\n",
+		       vma->vm, vm);
+		ok = false;
+	}
+
+	if (i915_is_ggtt(vm) != i915_vma_is_ggtt(vma)) {
+		pr_err("VMA ggtt status [%d] does not match parent [%d]\n",
+		       i915_vma_is_ggtt(vma), i915_is_ggtt(vm));
+		ok = false;
+	}
+
+	if (i915_vma_compare(vma, vm, view)) {
+		pr_err("i915_vma_compare failed with create parameters!\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (i915_vma_compare(vma, vma->vm,
+			     i915_vma_is_ggtt(vma) ? &vma->gtt_view : NULL)) {
+		pr_err("i915_vma_compare failed with itself\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (!ok) {
+		pr_err("i915_vma_compare failed to detect the difference!\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	return vma;
+}
+
+static int create_vmas(struct drm_i915_private *i915,
+		       struct list_head *objects,
+		       struct list_head *contexts)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_gem_context *ctx;
+	int pinned;
+
+	list_for_each_entry(obj, objects, st_link) {
+		for (pinned = 0; pinned <= 1; pinned++) {
+			list_for_each_entry(ctx, contexts, link) {
+				struct i915_address_space *vm;
+				struct i915_vma *vma;
+				int err;
+
+				vm = i915_gem_context_get_eb_vm(ctx);
+				vma = checked_vma_instance(obj, vm, NULL);
+				i915_vm_put(vm);
+				if (IS_ERR(vma))
+					return PTR_ERR(vma);
+
+				if (!assert_vma(vma, obj, ctx)) {
+					pr_err("VMA lookup/create failed\n");
+					return -EINVAL;
+				}
+
+				if (!pinned) {
+					err = i915_vma_pin(vma, 0, 0, PIN_USER);
+					if (err) {
+						pr_err("Failed to pin VMA\n");
+						return err;
+					}
+				} else {
+					i915_vma_unpin(vma);
+				}
+			}
+		}
+	}
+
+	return 0;
+}
+
+static int igt_vma_create(void *arg)
+{
+	struct i915_ggtt *ggtt = arg;
+	struct drm_i915_private *i915 = ggtt->vm.i915;
+	struct drm_i915_gem_object *obj, *on;
+	struct i915_gem_context *ctx, *cn;
+	unsigned long num_obj, num_ctx;
+	unsigned long no, nc;
+	IGT_TIMEOUT(end_time);
+	LIST_HEAD(contexts);
+	LIST_HEAD(objects);
+	int err = -ENOMEM;
+
+	/* Exercise creating many vma amonst many objections, checking the
+	 * vma creation and lookup routines.
+	 */
+
+	no = 0;
+	for_each_prime_number(num_obj, ULONG_MAX - 1) {
+		for (; no < num_obj; no++) {
+			obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+			if (IS_ERR(obj))
+				goto out;
+
+			list_add(&obj->st_link, &objects);
+		}
+
+		nc = 0;
+		for_each_prime_number(num_ctx, 2 * BITS_PER_LONG) {
+			for (; nc < num_ctx; nc++) {
+				ctx = mock_context(i915, "mock");
+				if (!ctx)
+					goto out;
+
+				list_move(&ctx->link, &contexts);
+			}
+
+			err = create_vmas(i915, &objects, &contexts);
+			if (err)
+				goto out;
+
+			if (igt_timeout(end_time,
+					"%s timed out: after %lu objects in %lu contexts\n",
+					__func__, no, nc))
+				goto end;
+		}
+
+		list_for_each_entry_safe(ctx, cn, &contexts, link) {
+			list_del_init(&ctx->link);
+			mock_context_close(ctx);
+		}
+
+		cond_resched();
+	}
+
+end:
+	/* Final pass to lookup all created contexts */
+	err = create_vmas(i915, &objects, &contexts);
+out:
+	list_for_each_entry_safe(ctx, cn, &contexts, link) {
+		list_del_init(&ctx->link);
+		mock_context_close(ctx);
+	}
+
+	list_for_each_entry_safe(obj, on, &objects, st_link)
+		i915_gem_object_put(obj);
+	return err;
+}
+
+struct pin_mode {
+	u64 size;
+	u64 flags;
+	bool (*assert)(const struct i915_vma *,
+		       const struct pin_mode *mode,
+		       int result);
+	const char *string;
+};
+
+static bool assert_pin_valid(const struct i915_vma *vma,
+			     const struct pin_mode *mode,
+			     int result)
+{
+	if (result)
+		return false;
+
+	if (i915_vma_misplaced(vma, mode->size, 0, mode->flags))
+		return false;
+
+	return true;
+}
+
+__maybe_unused
+static bool assert_pin_enospc(const struct i915_vma *vma,
+			      const struct pin_mode *mode,
+			      int result)
+{
+	return result == -ENOSPC;
+}
+
+__maybe_unused
+static bool assert_pin_einval(const struct i915_vma *vma,
+			      const struct pin_mode *mode,
+			      int result)
+{
+	return result == -EINVAL;
+}
+
+static int igt_vma_pin1(void *arg)
+{
+	struct i915_ggtt *ggtt = arg;
+	const struct pin_mode modes[] = {
+#define VALID(sz, fl) { .size = (sz), .flags = (fl), .assert = assert_pin_valid, .string = #sz ", " #fl ", (valid) " }
+#define __INVALID(sz, fl, check, eval) { .size = (sz), .flags = (fl), .assert = (check), .string = #sz ", " #fl ", (invalid " #eval ")" }
+#define INVALID(sz, fl) __INVALID(sz, fl, assert_pin_einval, EINVAL)
+#define NOSPACE(sz, fl) __INVALID(sz, fl, assert_pin_enospc, ENOSPC)
+		VALID(0, PIN_GLOBAL),
+		VALID(0, PIN_GLOBAL | PIN_MAPPABLE),
+
+		VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | 4096),
+		VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | 8192),
+		VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | (ggtt->mappable_end - 4096)),
+		VALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | (ggtt->mappable_end - 4096)),
+		VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | (ggtt->vm.total - 4096)),
+
+		VALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | (ggtt->mappable_end - 4096)),
+		INVALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | ggtt->mappable_end),
+		VALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | (ggtt->vm.total - 4096)),
+		INVALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | ggtt->vm.total),
+		INVALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | round_down(U64_MAX, PAGE_SIZE)),
+
+		VALID(4096, PIN_GLOBAL),
+		VALID(8192, PIN_GLOBAL),
+		VALID(ggtt->mappable_end - 4096, PIN_GLOBAL | PIN_MAPPABLE),
+		VALID(ggtt->mappable_end, PIN_GLOBAL | PIN_MAPPABLE),
+		NOSPACE(ggtt->mappable_end + 4096, PIN_GLOBAL | PIN_MAPPABLE),
+		VALID(ggtt->vm.total - 4096, PIN_GLOBAL),
+		VALID(ggtt->vm.total, PIN_GLOBAL),
+		NOSPACE(ggtt->vm.total + 4096, PIN_GLOBAL),
+		NOSPACE(round_down(U64_MAX, PAGE_SIZE), PIN_GLOBAL),
+		INVALID(8192, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | (ggtt->mappable_end - 4096)),
+		INVALID(8192, PIN_GLOBAL | PIN_OFFSET_FIXED | (ggtt->vm.total - 4096)),
+		INVALID(8192, PIN_GLOBAL | PIN_OFFSET_FIXED | (round_down(U64_MAX, PAGE_SIZE) - 4096)),
+
+		VALID(8192, PIN_GLOBAL | PIN_OFFSET_BIAS | (ggtt->mappable_end - 4096)),
+
+#if !IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)
+		/* Misusing BIAS is a programming error (it is not controllable
+		 * from userspace) so when debugging is enabled, it explodes.
+		 * However, the tests are still quite interesting for checking
+		 * variable start, end and size.
+		 */
+		NOSPACE(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | ggtt->mappable_end),
+		NOSPACE(0, PIN_GLOBAL | PIN_OFFSET_BIAS | ggtt->vm.total),
+		NOSPACE(8192, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | (ggtt->mappable_end - 4096)),
+		NOSPACE(8192, PIN_GLOBAL | PIN_OFFSET_BIAS | (ggtt->vm.total - 4096)),
+#endif
+		{ },
+#undef NOSPACE
+#undef INVALID
+#undef __INVALID
+#undef VALID
+	}, *m;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err = -EINVAL;
+
+	/* Exercise all the weird and wonderful i915_vma_pin requests,
+	 * focusing on error handling of boundary conditions.
+	 */
+
+	GEM_BUG_ON(!drm_mm_clean(&ggtt->vm.mm));
+
+	obj = i915_gem_object_create_internal(ggtt->vm.i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = checked_vma_instance(obj, &ggtt->vm, NULL);
+	if (IS_ERR(vma))
+		goto out;
+
+	for (m = modes; m->assert; m++) {
+		err = i915_vma_pin(vma, m->size, 0, m->flags);
+		if (!m->assert(vma, m, err)) {
+			pr_err("%s to pin single page into GGTT with mode[%d:%s]: size=%llx flags=%llx, err=%d\n",
+			       m->assert == assert_pin_valid ? "Failed" : "Unexpectedly succeeded",
+			       (int)(m - modes), m->string, m->size, m->flags,
+			       err);
+			if (!err)
+				i915_vma_unpin(vma);
+			err = -EINVAL;
+			goto out;
+		}
+
+		if (!err) {
+			i915_vma_unpin(vma);
+			err = i915_vma_unbind_unlocked(vma);
+			if (err) {
+				pr_err("Failed to unbind single page from GGTT, err=%d\n", err);
+				goto out;
+			}
+		}
+
+		cond_resched();
+	}
+
+	err = 0;
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static unsigned long rotated_index(const struct intel_rotation_info *r,
+				   unsigned int n,
+				   unsigned int x,
+				   unsigned int y)
+{
+	return (r->plane[n].src_stride * (r->plane[n].height - y - 1) +
+		r->plane[n].offset + x);
+}
+
+static struct scatterlist *
+assert_rotated(struct drm_i915_gem_object *obj,
+	       const struct intel_rotation_info *r, unsigned int n,
+	       struct scatterlist *sg)
+{
+	unsigned int x, y;
+
+	for (x = 0; x < r->plane[n].width; x++) {
+		unsigned int left;
+
+		for (y = 0; y < r->plane[n].height; y++) {
+			unsigned long src_idx;
+			dma_addr_t src;
+
+			if (!sg) {
+				pr_err("Invalid sg table: too short at plane %d, (%d, %d)!\n",
+				       n, x, y);
+				return ERR_PTR(-EINVAL);
+			}
+
+			src_idx = rotated_index(r, n, x, y);
+			src = i915_gem_object_get_dma_address(obj, src_idx);
+
+			if (sg_dma_len(sg) != PAGE_SIZE) {
+				pr_err("Invalid sg.length, found %d, expected %lu for rotated page (%d, %d) [src index %lu]\n",
+				       sg_dma_len(sg), PAGE_SIZE,
+				       x, y, src_idx);
+				return ERR_PTR(-EINVAL);
+			}
+
+			if (sg_dma_address(sg) != src) {
+				pr_err("Invalid address for rotated page (%d, %d) [src index %lu]\n",
+				       x, y, src_idx);
+				return ERR_PTR(-EINVAL);
+			}
+
+			sg = sg_next(sg);
+		}
+
+		left = (r->plane[n].dst_stride - y) * PAGE_SIZE;
+
+		if (!left)
+			continue;
+
+		if (!sg) {
+			pr_err("Invalid sg table: too short at plane %d, (%d, %d)!\n",
+			       n, x, y);
+			return ERR_PTR(-EINVAL);
+		}
+
+		if (sg_dma_len(sg) != left) {
+			pr_err("Invalid sg.length, found %d, expected %u for rotated page (%d, %d)\n",
+			       sg_dma_len(sg), left, x, y);
+			return ERR_PTR(-EINVAL);
+		}
+
+		if (sg_dma_address(sg) != 0) {
+			pr_err("Invalid address, found %pad, expected 0 for remapped page (%d, %d)\n",
+			       &sg_dma_address(sg), x, y);
+			return ERR_PTR(-EINVAL);
+		}
+
+		sg = sg_next(sg);
+	}
+
+	return sg;
+}
+
+static unsigned long remapped_index(const struct intel_remapped_info *r,
+				    unsigned int n,
+				    unsigned int x,
+				    unsigned int y)
+{
+	return (r->plane[n].src_stride * y +
+		r->plane[n].offset + x);
+}
+
+static struct scatterlist *
+assert_remapped(struct drm_i915_gem_object *obj,
+		const struct intel_remapped_info *r, unsigned int n,
+		struct scatterlist *sg)
+{
+	unsigned int x, y;
+	unsigned int left = 0;
+	unsigned int offset;
+
+	for (y = 0; y < r->plane[n].height; y++) {
+		for (x = 0; x < r->plane[n].width; x++) {
+			unsigned long src_idx;
+			dma_addr_t src;
+
+			if (!sg) {
+				pr_err("Invalid sg table: too short at plane %d, (%d, %d)!\n",
+				       n, x, y);
+				return ERR_PTR(-EINVAL);
+			}
+			if (!left) {
+				offset = 0;
+				left = sg_dma_len(sg);
+			}
+
+			src_idx = remapped_index(r, n, x, y);
+			src = i915_gem_object_get_dma_address(obj, src_idx);
+
+			if (left < PAGE_SIZE || left & (PAGE_SIZE-1)) {
+				pr_err("Invalid sg.length, found %d, expected %lu for remapped page (%d, %d) [src index %lu]\n",
+				       sg_dma_len(sg), PAGE_SIZE,
+				       x, y, src_idx);
+				return ERR_PTR(-EINVAL);
+			}
+
+			if (sg_dma_address(sg) + offset != src) {
+				pr_err("Invalid address for remapped page (%d, %d) [src index %lu]\n",
+				       x, y, src_idx);
+				return ERR_PTR(-EINVAL);
+			}
+
+			left -= PAGE_SIZE;
+			offset += PAGE_SIZE;
+
+
+			if (!left)
+				sg = sg_next(sg);
+		}
+
+		if (left) {
+			pr_err("Unexpected sg tail with %d size for remapped page (%d, %d)\n",
+			       left,
+			       x, y);
+			return ERR_PTR(-EINVAL);
+		}
+
+		left = (r->plane[n].dst_stride - r->plane[n].width) * PAGE_SIZE;
+
+		if (!left)
+			continue;
+
+		if (!sg) {
+			pr_err("Invalid sg table: too short at plane %d, (%d, %d)!\n",
+			       n, x, y);
+			return ERR_PTR(-EINVAL);
+		}
+
+		if (sg_dma_len(sg) != left) {
+			pr_err("Invalid sg.length, found %u, expected %u for remapped page (%d, %d)\n",
+			       sg_dma_len(sg), left,
+			       x, y);
+			return ERR_PTR(-EINVAL);
+		}
+
+		if (sg_dma_address(sg) != 0) {
+			pr_err("Invalid address, found %pad, expected 0 for remapped page (%d, %d)\n",
+			       &sg_dma_address(sg),
+			       x, y);
+			return ERR_PTR(-EINVAL);
+		}
+
+		sg = sg_next(sg);
+		left = 0;
+	}
+
+	return sg;
+}
+
+static unsigned int remapped_size(enum i915_gtt_view_type view_type,
+				  const struct intel_remapped_plane_info *a,
+				  const struct intel_remapped_plane_info *b)
+{
+
+	if (view_type == I915_GTT_VIEW_ROTATED)
+		return a->dst_stride * a->width + b->dst_stride * b->width;
+	else
+		return a->dst_stride * a->height + b->dst_stride * b->height;
+}
+
+static int igt_vma_rotate_remap(void *arg)
+{
+	struct i915_ggtt *ggtt = arg;
+	struct i915_address_space *vm = &ggtt->vm;
+	struct drm_i915_gem_object *obj;
+	const struct intel_remapped_plane_info planes[] = {
+		{ .width = 1, .height = 1, .src_stride = 1 },
+		{ .width = 2, .height = 2, .src_stride = 2 },
+		{ .width = 4, .height = 4, .src_stride = 4 },
+		{ .width = 8, .height = 8, .src_stride = 8 },
+
+		{ .width = 3, .height = 5, .src_stride = 3 },
+		{ .width = 3, .height = 5, .src_stride = 4 },
+		{ .width = 3, .height = 5, .src_stride = 5 },
+
+		{ .width = 5, .height = 3, .src_stride = 5 },
+		{ .width = 5, .height = 3, .src_stride = 7 },
+		{ .width = 5, .height = 3, .src_stride = 9 },
+
+		{ .width = 4, .height = 6, .src_stride = 6 },
+		{ .width = 6, .height = 4, .src_stride = 6 },
+
+		{ .width = 2, .height = 2, .src_stride = 2, .dst_stride = 2 },
+		{ .width = 3, .height = 3, .src_stride = 3, .dst_stride = 4 },
+		{ .width = 5, .height = 6, .src_stride = 7, .dst_stride = 8 },
+
+		{ }
+	}, *a, *b;
+	enum i915_gtt_view_type types[] = {
+		I915_GTT_VIEW_ROTATED,
+		I915_GTT_VIEW_REMAPPED,
+		0,
+	}, *t;
+	const unsigned int max_pages = 64;
+	int err = -ENOMEM;
+
+	/* Create VMA for many different combinations of planes and check
+	 * that the page layout within the rotated VMA match our expectations.
+	 */
+
+	obj = i915_gem_object_create_internal(vm->i915, max_pages * PAGE_SIZE);
+	if (IS_ERR(obj))
+		goto out;
+
+	for (t = types; *t; t++) {
+	for (a = planes; a->width; a++) {
+		for (b = planes + ARRAY_SIZE(planes); b-- != planes; ) {
+			struct i915_gtt_view view = {
+				.type = *t,
+				.remapped.plane[0] = *a,
+				.remapped.plane[1] = *b,
+			};
+			struct intel_remapped_plane_info *plane_info = view.remapped.plane;
+			unsigned int n, max_offset;
+
+			max_offset = max(plane_info[0].src_stride * plane_info[0].height,
+					 plane_info[1].src_stride * plane_info[1].height);
+			GEM_BUG_ON(max_offset > max_pages);
+			max_offset = max_pages - max_offset;
+
+			if (!plane_info[0].dst_stride)
+				plane_info[0].dst_stride = view.type == I915_GTT_VIEW_ROTATED ?
+									plane_info[0].height :
+									plane_info[0].width;
+			if (!plane_info[1].dst_stride)
+				plane_info[1].dst_stride = view.type == I915_GTT_VIEW_ROTATED ?
+									plane_info[1].height :
+									plane_info[1].width;
+
+			for_each_prime_number_from(plane_info[0].offset, 0, max_offset) {
+				for_each_prime_number_from(plane_info[1].offset, 0, max_offset) {
+					struct scatterlist *sg;
+					struct i915_vma *vma;
+					unsigned int expected_pages;
+
+					vma = checked_vma_instance(obj, vm, &view);
+					if (IS_ERR(vma)) {
+						err = PTR_ERR(vma);
+						goto out_object;
+					}
+
+					err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+					if (err) {
+						pr_err("Failed to pin VMA, err=%d\n", err);
+						goto out_object;
+					}
+
+					expected_pages = remapped_size(view.type, &plane_info[0], &plane_info[1]);
+
+					if (view.type == I915_GTT_VIEW_ROTATED &&
+					    vma->size != expected_pages * PAGE_SIZE) {
+						pr_err("VMA is wrong size, expected %lu, found %llu\n",
+						       PAGE_SIZE * expected_pages, vma->size);
+						err = -EINVAL;
+						goto out_object;
+					}
+
+					if (view.type == I915_GTT_VIEW_REMAPPED &&
+					    vma->size > expected_pages * PAGE_SIZE) {
+						pr_err("VMA is wrong size, expected %lu, found %llu\n",
+						       PAGE_SIZE * expected_pages, vma->size);
+						err = -EINVAL;
+						goto out_object;
+					}
+
+					if (vma->pages->nents > expected_pages) {
+						pr_err("sg table is wrong sizeo, expected %u, found %u nents\n",
+						       expected_pages, vma->pages->nents);
+						err = -EINVAL;
+						goto out_object;
+					}
+
+					if (vma->node.size < vma->size) {
+						pr_err("VMA binding too small, expected %llu, found %llu\n",
+						       vma->size, vma->node.size);
+						err = -EINVAL;
+						goto out_object;
+					}
+
+					if (vma->pages == obj->mm.pages) {
+						pr_err("VMA using unrotated object pages!\n");
+						err = -EINVAL;
+						goto out_object;
+					}
+
+					sg = vma->pages->sgl;
+					for (n = 0; n < ARRAY_SIZE(view.rotated.plane); n++) {
+						if (view.type == I915_GTT_VIEW_ROTATED)
+							sg = assert_rotated(obj, &view.rotated, n, sg);
+						else
+							sg = assert_remapped(obj, &view.remapped, n, sg);
+						if (IS_ERR(sg)) {
+							pr_err("Inconsistent %s VMA pages for plane %d: [(%d, %d, %d, %d, %d), (%d, %d, %d, %d, %d)]\n",
+							       view.type == I915_GTT_VIEW_ROTATED ?
+							       "rotated" : "remapped", n,
+							       plane_info[0].width,
+							       plane_info[0].height,
+							       plane_info[0].src_stride,
+							       plane_info[0].dst_stride,
+							       plane_info[0].offset,
+							       plane_info[1].width,
+							       plane_info[1].height,
+							       plane_info[1].src_stride,
+							       plane_info[1].dst_stride,
+							       plane_info[1].offset);
+							err = -EINVAL;
+							goto out_object;
+						}
+					}
+
+					i915_vma_unpin(vma);
+					err = i915_vma_unbind_unlocked(vma);
+					if (err) {
+						pr_err("Unbinding returned %i\n", err);
+						goto out_object;
+					}
+					cond_resched();
+				}
+			}
+		}
+	}
+	}
+
+out_object:
+	i915_gem_object_put(obj);
+out:
+	return err;
+}
+
+static bool assert_partial(struct drm_i915_gem_object *obj,
+			   struct i915_vma *vma,
+			   unsigned long offset,
+			   unsigned long size)
+{
+	struct sgt_iter sgt;
+	dma_addr_t dma;
+
+	for_each_sgt_daddr(dma, sgt, vma->pages) {
+		dma_addr_t src;
+
+		if (!size) {
+			pr_err("Partial scattergather list too long\n");
+			return false;
+		}
+
+		src = i915_gem_object_get_dma_address(obj, offset);
+		if (src != dma) {
+			pr_err("DMA mismatch for partial page offset %lu\n",
+			       offset);
+			return false;
+		}
+
+		offset++;
+		size--;
+	}
+
+	return true;
+}
+
+static bool assert_pin(struct i915_vma *vma,
+		       struct i915_gtt_view *view,
+		       u64 size,
+		       const char *name)
+{
+	bool ok = true;
+
+	if (vma->size != size) {
+		pr_err("(%s) VMA is wrong size, expected %llu, found %llu\n",
+		       name, size, vma->size);
+		ok = false;
+	}
+
+	if (vma->node.size < vma->size) {
+		pr_err("(%s) VMA binding too small, expected %llu, found %llu\n",
+		       name, vma->size, vma->node.size);
+		ok = false;
+	}
+
+	if (view && view->type != I915_GTT_VIEW_NORMAL) {
+		if (memcmp(&vma->gtt_view, view, sizeof(*view))) {
+			pr_err("(%s) VMA mismatch upon creation!\n",
+			       name);
+			ok = false;
+		}
+
+		if (vma->pages == vma->obj->mm.pages) {
+			pr_err("(%s) VMA using original object pages!\n",
+			       name);
+			ok = false;
+		}
+	} else {
+		if (vma->gtt_view.type != I915_GTT_VIEW_NORMAL) {
+			pr_err("Not the normal ggtt view! Found %d\n",
+			       vma->gtt_view.type);
+			ok = false;
+		}
+
+		if (vma->pages != vma->obj->mm.pages) {
+			pr_err("VMA not using object pages!\n");
+			ok = false;
+		}
+	}
+
+	return ok;
+}
+
+static int igt_vma_partial(void *arg)
+{
+	struct i915_ggtt *ggtt = arg;
+	struct i915_address_space *vm = &ggtt->vm;
+	const unsigned int npages = 1021; /* prime! */
+	struct drm_i915_gem_object *obj;
+	const struct phase {
+		const char *name;
+	} phases[] = {
+		{ "create" },
+		{ "lookup" },
+		{ },
+	}, *p;
+	unsigned int sz, offset;
+	struct i915_vma *vma;
+	int err = -ENOMEM;
+
+	/* Create lots of different VMA for the object and check that
+	 * we are returned the same VMA when we later request the same range.
+	 */
+
+	obj = i915_gem_object_create_internal(vm->i915, npages * PAGE_SIZE);
+	if (IS_ERR(obj))
+		goto out;
+
+	for (p = phases; p->name; p++) { /* exercise both create/lookup */
+		unsigned int count, nvma;
+
+		nvma = 0;
+		for_each_prime_number_from(sz, 1, npages) {
+			for_each_prime_number_from(offset, 0, npages - sz) {
+				struct i915_gtt_view view;
+
+				view.type = I915_GTT_VIEW_PARTIAL;
+				view.partial.offset = offset;
+				view.partial.size = sz;
+
+				if (sz == npages)
+					view.type = I915_GTT_VIEW_NORMAL;
+
+				vma = checked_vma_instance(obj, vm, &view);
+				if (IS_ERR(vma)) {
+					err = PTR_ERR(vma);
+					goto out_object;
+				}
+
+				err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+				if (err)
+					goto out_object;
+
+				if (!assert_pin(vma, &view, sz*PAGE_SIZE, p->name)) {
+					pr_err("(%s) Inconsistent partial pinning for (offset=%d, size=%d)\n",
+					       p->name, offset, sz);
+					err = -EINVAL;
+					goto out_object;
+				}
+
+				if (!assert_partial(obj, vma, offset, sz)) {
+					pr_err("(%s) Inconsistent partial pages for (offset=%d, size=%d)\n",
+					       p->name, offset, sz);
+					err = -EINVAL;
+					goto out_object;
+				}
+
+				i915_vma_unpin(vma);
+				nvma++;
+				err = i915_vma_unbind_unlocked(vma);
+				if (err) {
+					pr_err("Unbinding returned %i\n", err);
+					goto out_object;
+				}
+
+				cond_resched();
+			}
+		}
+
+		count = 0;
+		list_for_each_entry(vma, &obj->vma.list, obj_link)
+			count++;
+		if (count != nvma) {
+			pr_err("(%s) All partial vma were not recorded on the obj->vma_list: found %u, expected %u\n",
+			       p->name, count, nvma);
+			err = -EINVAL;
+			goto out_object;
+		}
+
+		/* Check that we did create the whole object mapping */
+		vma = checked_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out_object;
+		}
+
+		err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+		if (err)
+			goto out_object;
+
+		if (!assert_pin(vma, NULL, obj->base.size, p->name)) {
+			pr_err("(%s) inconsistent full pin\n", p->name);
+			err = -EINVAL;
+			goto out_object;
+		}
+
+		i915_vma_unpin(vma);
+
+		err = i915_vma_unbind_unlocked(vma);
+		if (err) {
+			pr_err("Unbinding returned %i\n", err);
+			goto out_object;
+		}
+
+		count = 0;
+		list_for_each_entry(vma, &obj->vma.list, obj_link)
+			count++;
+		if (count != nvma) {
+			pr_err("(%s) allocated an extra full vma!\n", p->name);
+			err = -EINVAL;
+			goto out_object;
+		}
+	}
+
+out_object:
+	i915_gem_object_put(obj);
+out:
+	return err;
+}
+
+int i915_vma_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_vma_create),
+		SUBTEST(igt_vma_pin1),
+		SUBTEST(igt_vma_rotate_remap),
+		SUBTEST(igt_vma_partial),
+	};
+	struct drm_i915_private *i915;
+	struct intel_gt *gt;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	/* allocate the ggtt */
+	err = intel_gt_assign_ggtt(to_gt(i915));
+	if (err)
+		goto out_put;
+
+	gt = to_gt(i915);
+
+	mock_init_ggtt(gt);
+
+	err = i915_subtests(tests, gt->ggtt);
+
+	mock_device_flush(i915);
+	i915_gem_drain_freed_objects(i915);
+	mock_fini_ggtt(gt->ggtt);
+
+out_put:
+	mock_destroy_device(i915);
+	return err;
+}
+
+static int igt_vma_remapped_gtt(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	const struct intel_remapped_plane_info planes[] = {
+		{ .width = 1, .height = 1, .src_stride = 1 },
+		{ .width = 2, .height = 2, .src_stride = 2 },
+		{ .width = 4, .height = 4, .src_stride = 4 },
+		{ .width = 8, .height = 8, .src_stride = 8 },
+
+		{ .width = 3, .height = 5, .src_stride = 3 },
+		{ .width = 3, .height = 5, .src_stride = 4 },
+		{ .width = 3, .height = 5, .src_stride = 5 },
+
+		{ .width = 5, .height = 3, .src_stride = 5 },
+		{ .width = 5, .height = 3, .src_stride = 7 },
+		{ .width = 5, .height = 3, .src_stride = 9 },
+
+		{ .width = 4, .height = 6, .src_stride = 6 },
+		{ .width = 6, .height = 4, .src_stride = 6 },
+
+		{ .width = 2, .height = 2, .src_stride = 2, .dst_stride = 2 },
+		{ .width = 3, .height = 3, .src_stride = 3, .dst_stride = 4 },
+		{ .width = 5, .height = 6, .src_stride = 7, .dst_stride = 8 },
+
+		{ }
+	}, *p;
+	enum i915_gtt_view_type types[] = {
+		I915_GTT_VIEW_ROTATED,
+		I915_GTT_VIEW_REMAPPED,
+		0,
+	}, *t;
+	struct drm_i915_gem_object *obj;
+	intel_wakeref_t wakeref;
+	int err = 0;
+
+	if (!i915_ggtt_has_aperture(to_gt(i915)->ggtt))
+		return 0;
+
+	obj = i915_gem_object_create_internal(i915, 10 * 10 * PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	wakeref = intel_runtime_pm_get(&i915->runtime_pm);
+
+	for (t = types; *t; t++) {
+		for (p = planes; p->width; p++) {
+			struct i915_gtt_view view = {
+				.type = *t,
+				.rotated.plane[0] = *p,
+			};
+			struct intel_remapped_plane_info *plane_info = view.rotated.plane;
+			struct i915_vma *vma;
+			u32 __iomem *map;
+			unsigned int x, y;
+
+			i915_gem_object_lock(obj, NULL);
+			err = i915_gem_object_set_to_gtt_domain(obj, true);
+			i915_gem_object_unlock(obj);
+			if (err)
+				goto out;
+
+			if (!plane_info[0].dst_stride)
+				plane_info[0].dst_stride = *t == I915_GTT_VIEW_ROTATED ?
+								 p->height : p->width;
+
+			vma = i915_gem_object_ggtt_pin(obj, &view, 0, 0, PIN_MAPPABLE);
+			if (IS_ERR(vma)) {
+				err = PTR_ERR(vma);
+				goto out;
+			}
+
+			GEM_BUG_ON(vma->gtt_view.type != *t);
+
+			map = i915_vma_pin_iomap(vma);
+			i915_vma_unpin(vma);
+			if (IS_ERR(map)) {
+				err = PTR_ERR(map);
+				goto out;
+			}
+
+			for (y = 0 ; y < plane_info[0].height; y++) {
+				for (x = 0 ; x < plane_info[0].width; x++) {
+					unsigned int offset;
+					u32 val = y << 16 | x;
+
+					if (*t == I915_GTT_VIEW_ROTATED)
+						offset = (x * plane_info[0].dst_stride + y) * PAGE_SIZE;
+					else
+						offset = (y * plane_info[0].dst_stride + x) * PAGE_SIZE;
+
+					iowrite32(val, &map[offset / sizeof(*map)]);
+				}
+			}
+
+			i915_vma_unpin_iomap(vma);
+
+			vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE);
+			if (IS_ERR(vma)) {
+				err = PTR_ERR(vma);
+				goto out;
+			}
+
+			GEM_BUG_ON(vma->gtt_view.type != I915_GTT_VIEW_NORMAL);
+
+			map = i915_vma_pin_iomap(vma);
+			i915_vma_unpin(vma);
+			if (IS_ERR(map)) {
+				err = PTR_ERR(map);
+				goto out;
+			}
+
+			for (y = 0 ; y < plane_info[0].height; y++) {
+				for (x = 0 ; x < plane_info[0].width; x++) {
+					unsigned int offset, src_idx;
+					u32 exp = y << 16 | x;
+					u32 val;
+
+					if (*t == I915_GTT_VIEW_ROTATED)
+						src_idx = rotated_index(&view.rotated, 0, x, y);
+					else
+						src_idx = remapped_index(&view.remapped, 0, x, y);
+					offset = src_idx * PAGE_SIZE;
+
+					val = ioread32(&map[offset / sizeof(*map)]);
+					if (val != exp) {
+						pr_err("%s VMA write test failed, expected 0x%x, found 0x%x\n",
+						       *t == I915_GTT_VIEW_ROTATED ? "Rotated" : "Remapped",
+						       exp, val);
+						i915_vma_unpin_iomap(vma);
+						err = -EINVAL;
+						goto out;
+					}
+				}
+			}
+			i915_vma_unpin_iomap(vma);
+
+			cond_resched();
+		}
+	}
+
+out:
+	intel_runtime_pm_put(&i915->runtime_pm, wakeref);
+	i915_gem_object_put(obj);
+
+	return err;
+}
+
+int i915_vma_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_vma_remapped_gtt),
+	};
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/igt_atomic.c b/drivers/gpu/drm/i915/selftests/igt_atomic.c
new file mode 100644
index 000000000..fb506b699
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_atomic.c
@@ -0,0 +1,47 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include <linux/preempt.h>
+#include <linux/bottom_half.h>
+#include <linux/irqflags.h>
+
+#include "igt_atomic.h"
+
+static void __preempt_begin(void)
+{
+	preempt_disable();
+}
+
+static void __preempt_end(void)
+{
+	preempt_enable();
+}
+
+static void __softirq_begin(void)
+{
+	local_bh_disable();
+}
+
+static void __softirq_end(void)
+{
+	local_bh_enable();
+}
+
+static void __hardirq_begin(void)
+{
+	local_irq_disable();
+}
+
+static void __hardirq_end(void)
+{
+	local_irq_enable();
+}
+
+const struct igt_atomic_section igt_atomic_phases[] = {
+	{ "preempt", __preempt_begin, __preempt_end },
+	{ "softirq", __softirq_begin, __softirq_end },
+	{ "hardirq", __hardirq_begin, __hardirq_end },
+	{ }
+};
diff --git a/drivers/gpu/drm/i915/selftests/igt_atomic.h b/drivers/gpu/drm/i915/selftests/igt_atomic.h
new file mode 100644
index 000000000..1991798ab
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_atomic.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef IGT_ATOMIC_H
+#define IGT_ATOMIC_H
+
+struct igt_atomic_section {
+	const char *name;
+	void (*critical_section_begin)(void);
+	void (*critical_section_end)(void);
+};
+
+extern const struct igt_atomic_section igt_atomic_phases[];
+
+#endif /* IGT_ATOMIC_H */
diff --git a/drivers/gpu/drm/i915/selftests/igt_flush_test.c b/drivers/gpu/drm/i915/selftests/igt_flush_test.c
new file mode 100644
index 000000000..b484e12df
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_flush_test.c
@@ -0,0 +1,35 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_requests.h"
+
+#include "i915_drv.h"
+#include "i915_selftest.h"
+
+#include "igt_flush_test.h"
+
+int igt_flush_test(struct drm_i915_private *i915)
+{
+	struct intel_gt *gt = to_gt(i915);
+	int ret = intel_gt_is_wedged(gt) ? -EIO : 0;
+
+	cond_resched();
+
+	if (intel_gt_wait_for_idle(gt, HZ * 3) == -ETIME) {
+		pr_err("%pS timed out, cancelling all further testing.\n",
+		       __builtin_return_address(0));
+
+		GEM_TRACE("%pS timed out.\n",
+			  __builtin_return_address(0));
+		GEM_TRACE_DUMP();
+
+		intel_gt_set_wedged(gt);
+		ret = -EIO;
+	}
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/i915/selftests/igt_flush_test.h b/drivers/gpu/drm/i915/selftests/igt_flush_test.h
new file mode 100644
index 000000000..7541fa74e
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_flush_test.h
@@ -0,0 +1,14 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef IGT_FLUSH_TEST_H
+#define IGT_FLUSH_TEST_H
+
+struct drm_i915_private;
+
+int igt_flush_test(struct drm_i915_private *i915);
+
+#endif /* IGT_FLUSH_TEST_H */
diff --git a/drivers/gpu/drm/i915/selftests/igt_live_test.c b/drivers/gpu/drm/i915/selftests/igt_live_test.c
new file mode 100644
index 000000000..72b58b666
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_live_test.c
@@ -0,0 +1,73 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "gt/intel_gt.h"
+
+#include "../i915_selftest.h"
+#include "igt_flush_test.h"
+#include "igt_live_test.h"
+
+int igt_live_test_begin(struct igt_live_test *t,
+			struct drm_i915_private *i915,
+			const char *func,
+			const char *name)
+{
+	struct intel_gt *gt = to_gt(i915);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int err;
+
+	t->i915 = i915;
+	t->func = func;
+	t->name = name;
+
+	err = intel_gt_wait_for_idle(gt, MAX_SCHEDULE_TIMEOUT);
+	if (err) {
+		pr_err("%s(%s): failed to idle before, with err=%d!",
+		       func, name, err);
+		return err;
+	}
+
+	t->reset_global = i915_reset_count(&i915->gpu_error);
+
+	for_each_engine(engine, gt, id)
+		t->reset_engine[id] =
+			i915_reset_engine_count(&i915->gpu_error, engine);
+
+	return 0;
+}
+
+int igt_live_test_end(struct igt_live_test *t)
+{
+	struct drm_i915_private *i915 = t->i915;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	if (igt_flush_test(i915))
+		return -EIO;
+
+	if (t->reset_global != i915_reset_count(&i915->gpu_error)) {
+		pr_err("%s(%s): GPU was reset %d times!\n",
+		       t->func, t->name,
+		       i915_reset_count(&i915->gpu_error) - t->reset_global);
+		return -EIO;
+	}
+
+	for_each_engine(engine, to_gt(i915), id) {
+		if (t->reset_engine[id] ==
+		    i915_reset_engine_count(&i915->gpu_error, engine))
+			continue;
+
+		pr_err("%s(%s): engine '%s' was reset %d times!\n",
+		       t->func, t->name, engine->name,
+		       i915_reset_engine_count(&i915->gpu_error, engine) -
+		       t->reset_engine[id]);
+		return -EIO;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/selftests/igt_live_test.h b/drivers/gpu/drm/i915/selftests/igt_live_test.h
new file mode 100644
index 000000000..36ed42736
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_live_test.h
@@ -0,0 +1,35 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef IGT_LIVE_TEST_H
+#define IGT_LIVE_TEST_H
+
+#include "gt/intel_engine.h" /* for I915_NUM_ENGINES */
+
+struct drm_i915_private;
+
+struct igt_live_test {
+	struct drm_i915_private *i915;
+	const char *func;
+	const char *name;
+
+	unsigned int reset_global;
+	unsigned int reset_engine[I915_NUM_ENGINES];
+};
+
+/*
+ * Flush the GPU state before and after the test to ensure that no residual
+ * code is running on the GPU that may affect this test. Also compare the
+ * state before and after the test and alert if it unexpectedly changes,
+ * e.g. if the GPU was reset.
+ */
+int igt_live_test_begin(struct igt_live_test *t,
+			struct drm_i915_private *i915,
+			const char *func,
+			const char *name);
+int igt_live_test_end(struct igt_live_test *t);
+
+#endif /* IGT_LIVE_TEST_H */
diff --git a/drivers/gpu/drm/i915/selftests/igt_mmap.c b/drivers/gpu/drm/i915/selftests/igt_mmap.c
new file mode 100644
index 000000000..e920a461b
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_mmap.c
@@ -0,0 +1,52 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <drm/drm_file.h>
+
+#include "i915_drv.h"
+#include "igt_mmap.h"
+
+unsigned long igt_mmap_offset(struct drm_i915_private *i915,
+			      u64 offset,
+			      unsigned long size,
+			      unsigned long prot,
+			      unsigned long flags)
+{
+	struct drm_vma_offset_node *node;
+	struct file *file;
+	unsigned long addr;
+	int err;
+
+	/* no need to refcount, we own this object */
+	drm_vma_offset_lock_lookup(i915->drm.vma_offset_manager);
+	node = drm_vma_offset_exact_lookup_locked(i915->drm.vma_offset_manager,
+						  offset / PAGE_SIZE, size / PAGE_SIZE);
+	drm_vma_offset_unlock_lookup(i915->drm.vma_offset_manager);
+
+	if (GEM_WARN_ON(!node)) {
+		pr_info("Failed to lookup %llx\n", offset);
+		return -ENOENT;
+	}
+
+	/* Pretend to open("/dev/dri/card0") */
+	file = mock_drm_getfile(i915->drm.primary, O_RDWR);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	err = drm_vma_node_allow(node, file->private_data);
+	if (err) {
+		addr = err;
+		goto out_file;
+	}
+
+	addr = vm_mmap(file, 0, drm_vma_node_size(node) << PAGE_SHIFT,
+		       prot, flags, drm_vma_node_offset_addr(node));
+
+	drm_vma_node_revoke(node, file->private_data);
+out_file:
+	fput(file);
+	return addr;
+}
diff --git a/drivers/gpu/drm/i915/selftests/igt_mmap.h b/drivers/gpu/drm/i915/selftests/igt_mmap.h
new file mode 100644
index 000000000..acbe34d81
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_mmap.h
@@ -0,0 +1,21 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef IGT_MMAP_H
+#define IGT_MMAP_H
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct drm_vma_offset_node;
+
+unsigned long igt_mmap_offset(struct drm_i915_private *i915,
+			      u64 offset,
+			      unsigned long size,
+			      unsigned long prot,
+			      unsigned long flags);
+
+#endif /* IGT_MMAP_H */
diff --git a/drivers/gpu/drm/i915/selftests/igt_reset.c b/drivers/gpu/drm/i915/selftests/igt_reset.c
new file mode 100644
index 000000000..a2838c65f
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_reset.c
@@ -0,0 +1,51 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include "igt_reset.h"
+
+#include "gt/intel_engine.h"
+#include "gt/intel_gt.h"
+
+#include "../i915_drv.h"
+
+void igt_global_reset_lock(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	pr_debug("%s: current gpu_error=%08lx\n", __func__, gt->reset.flags);
+
+	while (test_and_set_bit(I915_RESET_BACKOFF, &gt->reset.flags))
+		wait_event(gt->reset.queue,
+			   !test_bit(I915_RESET_BACKOFF, &gt->reset.flags));
+
+	for_each_engine(engine, gt, id) {
+		while (test_and_set_bit(I915_RESET_ENGINE + id,
+					&gt->reset.flags))
+			wait_on_bit(&gt->reset.flags, I915_RESET_ENGINE + id,
+				    TASK_UNINTERRUPTIBLE);
+	}
+}
+
+void igt_global_reset_unlock(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id)
+		clear_and_wake_up_bit(I915_RESET_ENGINE + id, &gt->reset.flags);
+
+	clear_bit(I915_RESET_BACKOFF, &gt->reset.flags);
+	wake_up_all(&gt->reset.queue);
+}
+
+bool igt_force_reset(struct intel_gt *gt)
+{
+	intel_gt_set_wedged(gt);
+	intel_gt_reset(gt, 0, NULL);
+
+	return !intel_gt_is_wedged(gt);
+}
diff --git a/drivers/gpu/drm/i915/selftests/igt_reset.h b/drivers/gpu/drm/i915/selftests/igt_reset.h
new file mode 100644
index 000000000..851873b67
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_reset.h
@@ -0,0 +1,18 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef __I915_SELFTESTS_IGT_RESET_H__
+#define __I915_SELFTESTS_IGT_RESET_H__
+
+#include <linux/types.h>
+
+struct intel_gt;
+
+void igt_global_reset_lock(struct intel_gt *gt);
+void igt_global_reset_unlock(struct intel_gt *gt);
+bool igt_force_reset(struct intel_gt *gt);
+
+#endif
diff --git a/drivers/gpu/drm/i915/selftests/igt_spinner.c b/drivers/gpu/drm/i915/selftests/igt_spinner.c
new file mode 100644
index 000000000..0c22594ae
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_spinner.c
@@ -0,0 +1,277 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+#include "gt/intel_gpu_commands.h"
+#include "gt/intel_gt.h"
+
+#include "gem/i915_gem_internal.h"
+#include "gem/selftests/igt_gem_utils.h"
+
+#include "igt_spinner.h"
+
+int igt_spinner_init(struct igt_spinner *spin, struct intel_gt *gt)
+{
+	int err;
+
+	memset(spin, 0, sizeof(*spin));
+	spin->gt = gt;
+
+	spin->hws = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(spin->hws)) {
+		err = PTR_ERR(spin->hws);
+		goto err;
+	}
+	i915_gem_object_set_cache_coherency(spin->hws, I915_CACHE_LLC);
+
+	spin->obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
+	if (IS_ERR(spin->obj)) {
+		err = PTR_ERR(spin->obj);
+		goto err_hws;
+	}
+
+	return 0;
+
+err_hws:
+	i915_gem_object_put(spin->hws);
+err:
+	return err;
+}
+
+static void *igt_spinner_pin_obj(struct intel_context *ce,
+				 struct i915_gem_ww_ctx *ww,
+				 struct drm_i915_gem_object *obj,
+				 unsigned int mode, struct i915_vma **vma)
+{
+	void *vaddr;
+	int ret;
+
+	*vma = i915_vma_instance(obj, ce->vm, NULL);
+	if (IS_ERR(*vma))
+		return ERR_CAST(*vma);
+
+	ret = i915_gem_object_lock(obj, ww);
+	if (ret)
+		return ERR_PTR(ret);
+
+	vaddr = i915_gem_object_pin_map(obj, mode);
+
+	if (!ww)
+		i915_gem_object_unlock(obj);
+
+	if (IS_ERR(vaddr))
+		return vaddr;
+
+	if (ww)
+		ret = i915_vma_pin_ww(*vma, ww, 0, 0, PIN_USER);
+	else
+		ret = i915_vma_pin(*vma, 0, 0, PIN_USER);
+
+	if (ret) {
+		i915_gem_object_unpin_map(obj);
+		return ERR_PTR(ret);
+	}
+
+	return vaddr;
+}
+
+int igt_spinner_pin(struct igt_spinner *spin,
+		    struct intel_context *ce,
+		    struct i915_gem_ww_ctx *ww)
+{
+	void *vaddr;
+
+	if (spin->ce && WARN_ON(spin->ce != ce))
+		return -ENODEV;
+	spin->ce = ce;
+
+	if (!spin->seqno) {
+		vaddr = igt_spinner_pin_obj(ce, ww, spin->hws, I915_MAP_WB, &spin->hws_vma);
+		if (IS_ERR(vaddr))
+			return PTR_ERR(vaddr);
+
+		spin->seqno = memset(vaddr, 0xff, PAGE_SIZE);
+	}
+
+	if (!spin->batch) {
+		unsigned int mode;
+
+		mode = i915_coherent_map_type(spin->gt->i915, spin->obj, false);
+		vaddr = igt_spinner_pin_obj(ce, ww, spin->obj, mode, &spin->batch_vma);
+		if (IS_ERR(vaddr))
+			return PTR_ERR(vaddr);
+
+		spin->batch = vaddr;
+	}
+
+	return 0;
+}
+
+static unsigned int seqno_offset(u64 fence)
+{
+	return offset_in_page(sizeof(u32) * fence);
+}
+
+static u64 hws_address(const struct i915_vma *hws,
+		       const struct i915_request *rq)
+{
+	return hws->node.start + seqno_offset(rq->fence.context);
+}
+
+static int move_to_active(struct i915_vma *vma,
+			  struct i915_request *rq,
+			  unsigned int flags)
+{
+	int err;
+
+	i915_vma_lock(vma);
+	err = i915_request_await_object(rq, vma->obj,
+					flags & EXEC_OBJECT_WRITE);
+	if (err == 0)
+		err = i915_vma_move_to_active(vma, rq, flags);
+	i915_vma_unlock(vma);
+
+	return err;
+}
+
+struct i915_request *
+igt_spinner_create_request(struct igt_spinner *spin,
+			   struct intel_context *ce,
+			   u32 arbitration_command)
+{
+	struct intel_engine_cs *engine = ce->engine;
+	struct i915_request *rq = NULL;
+	struct i915_vma *hws, *vma;
+	unsigned int flags;
+	u32 *batch;
+	int err;
+
+	GEM_BUG_ON(spin->gt != ce->vm->gt);
+
+	if (!intel_engine_can_store_dword(ce->engine))
+		return ERR_PTR(-ENODEV);
+
+	if (!spin->batch) {
+		err = igt_spinner_pin(spin, ce, NULL);
+		if (err)
+			return ERR_PTR(err);
+	}
+
+	hws = spin->hws_vma;
+	vma = spin->batch_vma;
+
+	rq = intel_context_create_request(ce);
+	if (IS_ERR(rq))
+		return ERR_CAST(rq);
+
+	err = move_to_active(vma, rq, 0);
+	if (err)
+		goto cancel_rq;
+
+	err = move_to_active(hws, rq, 0);
+	if (err)
+		goto cancel_rq;
+
+	batch = spin->batch;
+
+	if (GRAPHICS_VER(rq->engine->i915) >= 8) {
+		*batch++ = MI_STORE_DWORD_IMM_GEN4;
+		*batch++ = lower_32_bits(hws_address(hws, rq));
+		*batch++ = upper_32_bits(hws_address(hws, rq));
+	} else if (GRAPHICS_VER(rq->engine->i915) >= 6) {
+		*batch++ = MI_STORE_DWORD_IMM_GEN4;
+		*batch++ = 0;
+		*batch++ = hws_address(hws, rq);
+	} else if (GRAPHICS_VER(rq->engine->i915) >= 4) {
+		*batch++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
+		*batch++ = 0;
+		*batch++ = hws_address(hws, rq);
+	} else {
+		*batch++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL;
+		*batch++ = hws_address(hws, rq);
+	}
+	*batch++ = rq->fence.seqno;
+
+	*batch++ = arbitration_command;
+
+	if (GRAPHICS_VER(rq->engine->i915) >= 8)
+		*batch++ = MI_BATCH_BUFFER_START | BIT(8) | 1;
+	else if (IS_HASWELL(rq->engine->i915))
+		*batch++ = MI_BATCH_BUFFER_START | MI_BATCH_PPGTT_HSW;
+	else if (GRAPHICS_VER(rq->engine->i915) >= 6)
+		*batch++ = MI_BATCH_BUFFER_START;
+	else
+		*batch++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT;
+	*batch++ = lower_32_bits(vma->node.start);
+	*batch++ = upper_32_bits(vma->node.start);
+
+	*batch++ = MI_BATCH_BUFFER_END; /* not reached */
+
+	intel_gt_chipset_flush(engine->gt);
+
+	if (engine->emit_init_breadcrumb) {
+		err = engine->emit_init_breadcrumb(rq);
+		if (err)
+			goto cancel_rq;
+	}
+
+	flags = 0;
+	if (GRAPHICS_VER(rq->engine->i915) <= 5)
+		flags |= I915_DISPATCH_SECURE;
+	err = engine->emit_bb_start(rq, vma->node.start, PAGE_SIZE, flags);
+
+cancel_rq:
+	if (err) {
+		i915_request_set_error_once(rq, err);
+		i915_request_add(rq);
+	}
+	return err ? ERR_PTR(err) : rq;
+}
+
+static u32
+hws_seqno(const struct igt_spinner *spin, const struct i915_request *rq)
+{
+	u32 *seqno = spin->seqno + seqno_offset(rq->fence.context);
+
+	return READ_ONCE(*seqno);
+}
+
+void igt_spinner_end(struct igt_spinner *spin)
+{
+	if (!spin->batch)
+		return;
+
+	*spin->batch = MI_BATCH_BUFFER_END;
+	intel_gt_chipset_flush(spin->gt);
+}
+
+void igt_spinner_fini(struct igt_spinner *spin)
+{
+	igt_spinner_end(spin);
+
+	if (spin->batch) {
+		i915_vma_unpin(spin->batch_vma);
+		i915_gem_object_unpin_map(spin->obj);
+	}
+	i915_gem_object_put(spin->obj);
+
+	if (spin->seqno) {
+		i915_vma_unpin(spin->hws_vma);
+		i915_gem_object_unpin_map(spin->hws);
+	}
+	i915_gem_object_put(spin->hws);
+}
+
+bool igt_wait_for_spinner(struct igt_spinner *spin, struct i915_request *rq)
+{
+	if (i915_request_is_ready(rq))
+		intel_engine_flush_submission(rq->engine);
+
+	return !(wait_for_us(i915_seqno_passed(hws_seqno(spin, rq),
+					       rq->fence.seqno),
+			     100) &&
+		 wait_for(i915_seqno_passed(hws_seqno(spin, rq),
+					    rq->fence.seqno),
+			  50));
+}
diff --git a/drivers/gpu/drm/i915/selftests/igt_spinner.h b/drivers/gpu/drm/i915/selftests/igt_spinner.h
new file mode 100644
index 000000000..fbe5b1625
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/igt_spinner.h
@@ -0,0 +1,43 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#ifndef __I915_SELFTESTS_IGT_SPINNER_H__
+#define __I915_SELFTESTS_IGT_SPINNER_H__
+
+#include "gem/i915_gem_context.h"
+#include "gt/intel_engine.h"
+
+#include "i915_drv.h"
+#include "i915_request.h"
+#include "i915_selftest.h"
+
+struct intel_gt;
+
+struct igt_spinner {
+	struct intel_gt *gt;
+	struct drm_i915_gem_object *hws;
+	struct drm_i915_gem_object *obj;
+	struct intel_context *ce;
+	struct i915_vma *hws_vma, *batch_vma;
+	u32 *batch;
+	void *seqno;
+};
+
+int igt_spinner_init(struct igt_spinner *spin, struct intel_gt *gt);
+int igt_spinner_pin(struct igt_spinner *spin,
+		    struct intel_context *ce,
+		    struct i915_gem_ww_ctx *ww);
+void igt_spinner_fini(struct igt_spinner *spin);
+
+struct i915_request *
+igt_spinner_create_request(struct igt_spinner *spin,
+			   struct intel_context *ce,
+			   u32 arbitration_command);
+void igt_spinner_end(struct igt_spinner *spin);
+
+bool igt_wait_for_spinner(struct igt_spinner *spin, struct i915_request *rq);
+
+#endif
diff --git a/drivers/gpu/drm/i915/selftests/intel_memory_region.c b/drivers/gpu/drm/i915/selftests/intel_memory_region.c
new file mode 100644
index 000000000..3b18e5905
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/intel_memory_region.c
@@ -0,0 +1,1411 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#include <linux/prime_numbers.h>
+#include <linux/sort.h>
+
+#include <drm/drm_buddy.h>
+
+#include "../i915_selftest.h"
+
+#include "mock_drm.h"
+#include "mock_gem_device.h"
+#include "mock_region.h"
+
+#include "gem/i915_gem_context.h"
+#include "gem/i915_gem_lmem.h"
+#include "gem/i915_gem_region.h"
+#include "gem/i915_gem_ttm.h"
+#include "gem/selftests/igt_gem_utils.h"
+#include "gem/selftests/mock_context.h"
+#include "gt/intel_engine_pm.h"
+#include "gt/intel_engine_user.h"
+#include "gt/intel_gt.h"
+#include "gt/intel_migrate.h"
+#include "i915_memcpy.h"
+#include "i915_ttm_buddy_manager.h"
+#include "selftests/igt_flush_test.h"
+#include "selftests/i915_random.h"
+
+static void close_objects(struct intel_memory_region *mem,
+			  struct list_head *objects)
+{
+	struct drm_i915_private *i915 = mem->i915;
+	struct drm_i915_gem_object *obj, *on;
+
+	list_for_each_entry_safe(obj, on, objects, st_link) {
+		i915_gem_object_lock(obj, NULL);
+		if (i915_gem_object_has_pinned_pages(obj))
+			i915_gem_object_unpin_pages(obj);
+		/* No polluting the memory region between tests */
+		__i915_gem_object_put_pages(obj);
+		i915_gem_object_unlock(obj);
+		list_del(&obj->st_link);
+		i915_gem_object_put(obj);
+	}
+
+	cond_resched();
+
+	i915_gem_drain_freed_objects(i915);
+}
+
+static int igt_mock_fill(void *arg)
+{
+	struct intel_memory_region *mem = arg;
+	resource_size_t total = resource_size(&mem->region);
+	resource_size_t page_size;
+	resource_size_t rem;
+	unsigned long max_pages;
+	unsigned long page_num;
+	LIST_HEAD(objects);
+	int err = 0;
+
+	page_size = PAGE_SIZE;
+	max_pages = div64_u64(total, page_size);
+	rem = total;
+
+	for_each_prime_number_from(page_num, 1, max_pages) {
+		resource_size_t size = page_num * page_size;
+		struct drm_i915_gem_object *obj;
+
+		obj = i915_gem_object_create_region(mem, size, 0, 0);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			break;
+		}
+
+		err = i915_gem_object_pin_pages_unlocked(obj);
+		if (err) {
+			i915_gem_object_put(obj);
+			break;
+		}
+
+		list_add(&obj->st_link, &objects);
+		rem -= size;
+	}
+
+	if (err == -ENOMEM)
+		err = 0;
+	if (err == -ENXIO) {
+		if (page_num * page_size <= rem) {
+			pr_err("%s failed, space still left in region\n",
+			       __func__);
+			err = -EINVAL;
+		} else {
+			err = 0;
+		}
+	}
+
+	close_objects(mem, &objects);
+
+	return err;
+}
+
+static struct drm_i915_gem_object *
+igt_object_create(struct intel_memory_region *mem,
+		  struct list_head *objects,
+		  u64 size,
+		  unsigned int flags)
+{
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	obj = i915_gem_object_create_region(mem, size, 0, flags);
+	if (IS_ERR(obj))
+		return obj;
+
+	err = i915_gem_object_pin_pages_unlocked(obj);
+	if (err)
+		goto put;
+
+	list_add(&obj->st_link, objects);
+	return obj;
+
+put:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static void igt_object_release(struct drm_i915_gem_object *obj)
+{
+	i915_gem_object_lock(obj, NULL);
+	i915_gem_object_unpin_pages(obj);
+	__i915_gem_object_put_pages(obj);
+	i915_gem_object_unlock(obj);
+	list_del(&obj->st_link);
+	i915_gem_object_put(obj);
+}
+
+static bool is_contiguous(struct drm_i915_gem_object *obj)
+{
+	struct scatterlist *sg;
+	dma_addr_t addr = -1;
+
+	for (sg = obj->mm.pages->sgl; sg; sg = sg_next(sg)) {
+		if (addr != -1 && sg_dma_address(sg) != addr)
+			return false;
+
+		addr = sg_dma_address(sg) + sg_dma_len(sg);
+	}
+
+	return true;
+}
+
+static int igt_mock_reserve(void *arg)
+{
+	struct intel_memory_region *mem = arg;
+	struct drm_i915_private *i915 = mem->i915;
+	resource_size_t avail = resource_size(&mem->region);
+	struct drm_i915_gem_object *obj;
+	const u32 chunk_size = SZ_32M;
+	u32 i, offset, count, *order;
+	u64 allocated, cur_avail;
+	I915_RND_STATE(prng);
+	LIST_HEAD(objects);
+	int err = 0;
+
+	count = avail / chunk_size;
+	order = i915_random_order(count, &prng);
+	if (!order)
+		return 0;
+
+	mem = mock_region_create(i915, 0, SZ_2G, I915_GTT_PAGE_SIZE_4K, 0, 0);
+	if (IS_ERR(mem)) {
+		pr_err("failed to create memory region\n");
+		err = PTR_ERR(mem);
+		goto out_free_order;
+	}
+
+	/* Reserve a bunch of ranges within the region */
+	for (i = 0; i < count; ++i) {
+		u64 start = order[i] * chunk_size;
+		u64 size = i915_prandom_u32_max_state(chunk_size, &prng);
+
+		/* Allow for some really big holes */
+		if (!size)
+			continue;
+
+		size = round_up(size, PAGE_SIZE);
+		offset = igt_random_offset(&prng, 0, chunk_size, size,
+					   PAGE_SIZE);
+
+		err = intel_memory_region_reserve(mem, start + offset, size);
+		if (err) {
+			pr_err("%s failed to reserve range", __func__);
+			goto out_close;
+		}
+
+		/* XXX: maybe sanity check the block range here? */
+		avail -= size;
+	}
+
+	/* Try to see if we can allocate from the remaining space */
+	allocated = 0;
+	cur_avail = avail;
+	do {
+		u32 size = i915_prandom_u32_max_state(cur_avail, &prng);
+
+		size = max_t(u32, round_up(size, PAGE_SIZE), PAGE_SIZE);
+		obj = igt_object_create(mem, &objects, size, 0);
+		if (IS_ERR(obj)) {
+			if (PTR_ERR(obj) == -ENXIO)
+				break;
+
+			err = PTR_ERR(obj);
+			goto out_close;
+		}
+		cur_avail -= size;
+		allocated += size;
+	} while (1);
+
+	if (allocated != avail) {
+		pr_err("%s mismatch between allocation and free space", __func__);
+		err = -EINVAL;
+	}
+
+out_close:
+	close_objects(mem, &objects);
+	intel_memory_region_destroy(mem);
+out_free_order:
+	kfree(order);
+	return err;
+}
+
+static int igt_mock_contiguous(void *arg)
+{
+	struct intel_memory_region *mem = arg;
+	struct drm_i915_gem_object *obj;
+	unsigned long n_objects;
+	LIST_HEAD(objects);
+	LIST_HEAD(holes);
+	I915_RND_STATE(prng);
+	resource_size_t total;
+	resource_size_t min;
+	u64 target;
+	int err = 0;
+
+	total = resource_size(&mem->region);
+
+	/* Min size */
+	obj = igt_object_create(mem, &objects, PAGE_SIZE,
+				I915_BO_ALLOC_CONTIGUOUS);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	if (!is_contiguous(obj)) {
+		pr_err("%s min object spans disjoint sg entries\n", __func__);
+		err = -EINVAL;
+		goto err_close_objects;
+	}
+
+	igt_object_release(obj);
+
+	/* Max size */
+	obj = igt_object_create(mem, &objects, total, I915_BO_ALLOC_CONTIGUOUS);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	if (!is_contiguous(obj)) {
+		pr_err("%s max object spans disjoint sg entries\n", __func__);
+		err = -EINVAL;
+		goto err_close_objects;
+	}
+
+	igt_object_release(obj);
+
+	/* Internal fragmentation should not bleed into the object size */
+	target = i915_prandom_u64_state(&prng);
+	div64_u64_rem(target, total, &target);
+	target = round_up(target, PAGE_SIZE);
+	target = max_t(u64, PAGE_SIZE, target);
+
+	obj = igt_object_create(mem, &objects, target,
+				I915_BO_ALLOC_CONTIGUOUS);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	if (obj->base.size != target) {
+		pr_err("%s obj->base.size(%zx) != target(%llx)\n", __func__,
+		       obj->base.size, target);
+		err = -EINVAL;
+		goto err_close_objects;
+	}
+
+	if (!is_contiguous(obj)) {
+		pr_err("%s object spans disjoint sg entries\n", __func__);
+		err = -EINVAL;
+		goto err_close_objects;
+	}
+
+	igt_object_release(obj);
+
+	/*
+	 * Try to fragment the address space, such that half of it is free, but
+	 * the max contiguous block size is SZ_64K.
+	 */
+
+	target = SZ_64K;
+	n_objects = div64_u64(total, target);
+
+	while (n_objects--) {
+		struct list_head *list;
+
+		if (n_objects % 2)
+			list = &holes;
+		else
+			list = &objects;
+
+		obj = igt_object_create(mem, list, target,
+					I915_BO_ALLOC_CONTIGUOUS);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto err_close_objects;
+		}
+	}
+
+	close_objects(mem, &holes);
+
+	min = target;
+	target = total >> 1;
+
+	/* Make sure we can still allocate all the fragmented space */
+	obj = igt_object_create(mem, &objects, target, 0);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto err_close_objects;
+	}
+
+	igt_object_release(obj);
+
+	/*
+	 * Even though we have enough free space, we don't have a big enough
+	 * contiguous block. Make sure that holds true.
+	 */
+
+	do {
+		bool should_fail = target > min;
+
+		obj = igt_object_create(mem, &objects, target,
+					I915_BO_ALLOC_CONTIGUOUS);
+		if (should_fail != IS_ERR(obj)) {
+			pr_err("%s target allocation(%llx) mismatch\n",
+			       __func__, target);
+			err = -EINVAL;
+			goto err_close_objects;
+		}
+
+		target >>= 1;
+	} while (target >= PAGE_SIZE);
+
+err_close_objects:
+	list_splice_tail(&holes, &objects);
+	close_objects(mem, &objects);
+	return err;
+}
+
+static int igt_mock_splintered_region(void *arg)
+{
+	struct intel_memory_region *mem = arg;
+	struct drm_i915_private *i915 = mem->i915;
+	struct i915_ttm_buddy_resource *res;
+	struct drm_i915_gem_object *obj;
+	struct drm_buddy *mm;
+	unsigned int expected_order;
+	LIST_HEAD(objects);
+	u64 size;
+	int err = 0;
+
+	/*
+	 * Sanity check we can still allocate everything even if the
+	 * mm.max_order != mm.size. i.e our starting address space size is not a
+	 * power-of-two.
+	 */
+
+	size = (SZ_4G - 1) & PAGE_MASK;
+	mem = mock_region_create(i915, 0, size, PAGE_SIZE, 0, 0);
+	if (IS_ERR(mem))
+		return PTR_ERR(mem);
+
+	obj = igt_object_create(mem, &objects, size, 0);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto out_close;
+	}
+
+	res = to_ttm_buddy_resource(obj->mm.res);
+	mm = res->mm;
+	if (mm->size != size) {
+		pr_err("%s size mismatch(%llu != %llu)\n",
+		       __func__, mm->size, size);
+		err = -EINVAL;
+		goto out_put;
+	}
+
+	expected_order = get_order(rounddown_pow_of_two(size));
+	if (mm->max_order != expected_order) {
+		pr_err("%s order mismatch(%u != %u)\n",
+		       __func__, mm->max_order, expected_order);
+		err = -EINVAL;
+		goto out_put;
+	}
+
+	close_objects(mem, &objects);
+
+	/*
+	 * While we should be able allocate everything without any flag
+	 * restrictions, if we consider I915_BO_ALLOC_CONTIGUOUS then we are
+	 * actually limited to the largest power-of-two for the region size i.e
+	 * max_order, due to the inner workings of the buddy allocator. So make
+	 * sure that does indeed hold true.
+	 */
+
+	obj = igt_object_create(mem, &objects, size, I915_BO_ALLOC_CONTIGUOUS);
+	if (!IS_ERR(obj)) {
+		pr_err("%s too large contiguous allocation was not rejected\n",
+		       __func__);
+		err = -EINVAL;
+		goto out_close;
+	}
+
+	obj = igt_object_create(mem, &objects, rounddown_pow_of_two(size),
+				I915_BO_ALLOC_CONTIGUOUS);
+	if (IS_ERR(obj)) {
+		pr_err("%s largest possible contiguous allocation failed\n",
+		       __func__);
+		err = PTR_ERR(obj);
+		goto out_close;
+	}
+
+out_close:
+	close_objects(mem, &objects);
+out_put:
+	intel_memory_region_destroy(mem);
+	return err;
+}
+
+#ifndef SZ_8G
+#define SZ_8G BIT_ULL(33)
+#endif
+
+static int igt_mock_max_segment(void *arg)
+{
+	struct intel_memory_region *mem = arg;
+	struct drm_i915_private *i915 = mem->i915;
+	struct i915_ttm_buddy_resource *res;
+	struct drm_i915_gem_object *obj;
+	struct drm_buddy_block *block;
+	struct drm_buddy *mm;
+	struct list_head *blocks;
+	struct scatterlist *sg;
+	I915_RND_STATE(prng);
+	LIST_HEAD(objects);
+	unsigned int max_segment;
+	unsigned int ps;
+	u64 size;
+	int err = 0;
+
+	/*
+	 * While we may create very large contiguous blocks, we may need
+	 * to break those down for consumption elsewhere. In particular,
+	 * dma-mapping with scatterlist elements have an implicit limit of
+	 * UINT_MAX on each element.
+	 */
+
+	size = SZ_8G;
+	ps = PAGE_SIZE;
+	if (i915_prandom_u64_state(&prng) & 1)
+		ps = SZ_64K; /* For something like DG2 */
+
+	max_segment = round_down(UINT_MAX, ps);
+
+	mem = mock_region_create(i915, 0, size, ps, 0, 0);
+	if (IS_ERR(mem))
+		return PTR_ERR(mem);
+
+	obj = igt_object_create(mem, &objects, size, 0);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto out_put;
+	}
+
+	res = to_ttm_buddy_resource(obj->mm.res);
+	blocks = &res->blocks;
+	mm = res->mm;
+	size = 0;
+	list_for_each_entry(block, blocks, link) {
+		if (drm_buddy_block_size(mm, block) > size)
+			size = drm_buddy_block_size(mm, block);
+	}
+	if (size < max_segment) {
+		pr_err("%s: Failed to create a huge contiguous block [> %u], largest block %lld\n",
+		       __func__, max_segment, size);
+		err = -EINVAL;
+		goto out_close;
+	}
+
+	for (sg = obj->mm.pages->sgl; sg; sg = sg_next(sg)) {
+		dma_addr_t daddr = sg_dma_address(sg);
+
+		if (sg->length > max_segment) {
+			pr_err("%s: Created an oversized scatterlist entry, %u > %u\n",
+			       __func__, sg->length, max_segment);
+			err = -EINVAL;
+			goto out_close;
+		}
+
+		if (!IS_ALIGNED(daddr, ps)) {
+			pr_err("%s: Created an unaligned scatterlist entry, addr=%pa, ps=%u\n",
+			       __func__,  &daddr, ps);
+			err = -EINVAL;
+			goto out_close;
+		}
+	}
+
+out_close:
+	close_objects(mem, &objects);
+out_put:
+	intel_memory_region_destroy(mem);
+	return err;
+}
+
+static u64 igt_object_mappable_total(struct drm_i915_gem_object *obj)
+{
+	struct intel_memory_region *mr = obj->mm.region;
+	struct i915_ttm_buddy_resource *bman_res =
+		to_ttm_buddy_resource(obj->mm.res);
+	struct drm_buddy *mm = bman_res->mm;
+	struct drm_buddy_block *block;
+	u64 total;
+
+	total = 0;
+	list_for_each_entry(block, &bman_res->blocks, link) {
+		u64 start = drm_buddy_block_offset(block);
+		u64 end = start + drm_buddy_block_size(mm, block);
+
+		if (start < mr->io_size)
+			total += min_t(u64, end, mr->io_size) - start;
+	}
+
+	return total;
+}
+
+static int igt_mock_io_size(void *arg)
+{
+	struct intel_memory_region *mr = arg;
+	struct drm_i915_private *i915 = mr->i915;
+	struct drm_i915_gem_object *obj;
+	u64 mappable_theft_total;
+	u64 io_size;
+	u64 total;
+	u64 ps;
+	u64 rem;
+	u64 size;
+	I915_RND_STATE(prng);
+	LIST_HEAD(objects);
+	int err = 0;
+
+	ps = SZ_4K;
+	if (i915_prandom_u64_state(&prng) & 1)
+		ps = SZ_64K; /* For something like DG2 */
+
+	div64_u64_rem(i915_prandom_u64_state(&prng), SZ_8G, &total);
+	total = round_down(total, ps);
+	total = max_t(u64, total, SZ_1G);
+
+	div64_u64_rem(i915_prandom_u64_state(&prng), total - ps, &io_size);
+	io_size = round_down(io_size, ps);
+	io_size = max_t(u64, io_size, SZ_256M); /* 256M seems to be the common lower limit */
+
+	pr_info("%s with ps=%llx, io_size=%llx, total=%llx\n",
+		__func__, ps, io_size, total);
+
+	mr = mock_region_create(i915, 0, total, ps, 0, io_size);
+	if (IS_ERR(mr)) {
+		err = PTR_ERR(mr);
+		goto out_err;
+	}
+
+	mappable_theft_total = 0;
+	rem = total - io_size;
+	do {
+		div64_u64_rem(i915_prandom_u64_state(&prng), rem, &size);
+		size = round_down(size, ps);
+		size = max(size, ps);
+
+		obj = igt_object_create(mr, &objects, size,
+					I915_BO_ALLOC_GPU_ONLY);
+		if (IS_ERR(obj)) {
+			pr_err("%s TOPDOWN failed with rem=%llx, size=%llx\n",
+			       __func__, rem, size);
+			err = PTR_ERR(obj);
+			goto out_close;
+		}
+
+		mappable_theft_total += igt_object_mappable_total(obj);
+		rem -= size;
+	} while (rem);
+
+	pr_info("%s mappable theft=(%lluMiB/%lluMiB), total=%lluMiB\n",
+		__func__,
+		(u64)mappable_theft_total >> 20,
+		(u64)io_size >> 20,
+		(u64)total >> 20);
+
+	/*
+	 * Even if we allocate all of the non-mappable portion, we should still
+	 * be able to dip into the mappable portion.
+	 */
+	obj = igt_object_create(mr, &objects, io_size,
+				I915_BO_ALLOC_GPU_ONLY);
+	if (IS_ERR(obj)) {
+		pr_err("%s allocation unexpectedly failed\n", __func__);
+		err = PTR_ERR(obj);
+		goto out_close;
+	}
+
+	close_objects(mr, &objects);
+
+	rem = io_size;
+	do {
+		div64_u64_rem(i915_prandom_u64_state(&prng), rem, &size);
+		size = round_down(size, ps);
+		size = max(size, ps);
+
+		obj = igt_object_create(mr, &objects, size, 0);
+		if (IS_ERR(obj)) {
+			pr_err("%s MAPPABLE failed with rem=%llx, size=%llx\n",
+			       __func__, rem, size);
+			err = PTR_ERR(obj);
+			goto out_close;
+		}
+
+		if (igt_object_mappable_total(obj) != size) {
+			pr_err("%s allocation is not mappable(size=%llx)\n",
+			       __func__, size);
+			err = -EINVAL;
+			goto out_close;
+		}
+		rem -= size;
+	} while (rem);
+
+	/*
+	 * We assume CPU access is required by default, which should result in a
+	 * failure here, even though the non-mappable portion is free.
+	 */
+	obj = igt_object_create(mr, &objects, ps, 0);
+	if (!IS_ERR(obj)) {
+		pr_err("%s allocation unexpectedly succeeded\n", __func__);
+		err = -EINVAL;
+		goto out_close;
+	}
+
+out_close:
+	close_objects(mr, &objects);
+	intel_memory_region_destroy(mr);
+out_err:
+	if (err == -ENOMEM)
+		err = 0;
+
+	return err;
+}
+
+static int igt_gpu_write_dw(struct intel_context *ce,
+			    struct i915_vma *vma,
+			    u32 dword,
+			    u32 value)
+{
+	return igt_gpu_fill_dw(ce, vma, dword * sizeof(u32),
+			       vma->size >> PAGE_SHIFT, value);
+}
+
+static int igt_cpu_check(struct drm_i915_gem_object *obj, u32 dword, u32 val)
+{
+	unsigned long n = obj->base.size >> PAGE_SHIFT;
+	u32 *ptr;
+	int err;
+
+	err = i915_gem_object_wait(obj, 0, MAX_SCHEDULE_TIMEOUT);
+	if (err)
+		return err;
+
+	ptr = i915_gem_object_pin_map(obj, I915_MAP_WC);
+	if (IS_ERR(ptr))
+		return PTR_ERR(ptr);
+
+	ptr += dword;
+	while (n--) {
+		if (*ptr != val) {
+			pr_err("base[%u]=%08x, val=%08x\n",
+			       dword, *ptr, val);
+			err = -EINVAL;
+			break;
+		}
+
+		ptr += PAGE_SIZE / sizeof(*ptr);
+	}
+
+	i915_gem_object_unpin_map(obj);
+	return err;
+}
+
+static int igt_gpu_write(struct i915_gem_context *ctx,
+			 struct drm_i915_gem_object *obj)
+{
+	struct i915_gem_engines *engines;
+	struct i915_gem_engines_iter it;
+	struct i915_address_space *vm;
+	struct intel_context *ce;
+	I915_RND_STATE(prng);
+	IGT_TIMEOUT(end_time);
+	unsigned int count;
+	struct i915_vma *vma;
+	int *order;
+	int i, n;
+	int err = 0;
+
+	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
+
+	n = 0;
+	count = 0;
+	for_each_gem_engine(ce, i915_gem_context_lock_engines(ctx), it) {
+		count++;
+		if (!intel_engine_can_store_dword(ce->engine))
+			continue;
+
+		vm = ce->vm;
+		n++;
+	}
+	i915_gem_context_unlock_engines(ctx);
+	if (!n)
+		return 0;
+
+	order = i915_random_order(count * count, &prng);
+	if (!order)
+		return -ENOMEM;
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto out_free;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto out_free;
+
+	i = 0;
+	engines = i915_gem_context_lock_engines(ctx);
+	do {
+		u32 rng = prandom_u32_state(&prng);
+		u32 dword = offset_in_page(rng) / 4;
+
+		ce = engines->engines[order[i] % engines->num_engines];
+		i = (i + 1) % (count * count);
+		if (!ce || !intel_engine_can_store_dword(ce->engine))
+			continue;
+
+		err = igt_gpu_write_dw(ce, vma, dword, rng);
+		if (err)
+			break;
+
+		i915_gem_object_lock(obj, NULL);
+		err = igt_cpu_check(obj, dword, rng);
+		i915_gem_object_unlock(obj);
+		if (err)
+			break;
+	} while (!__igt_timeout(end_time, NULL));
+	i915_gem_context_unlock_engines(ctx);
+
+out_free:
+	kfree(order);
+
+	if (err == -ENOMEM)
+		err = 0;
+
+	return err;
+}
+
+static int igt_lmem_create(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	int err = 0;
+
+	obj = i915_gem_object_create_lmem(i915, PAGE_SIZE, 0);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	err = i915_gem_object_pin_pages_unlocked(obj);
+	if (err)
+		goto out_put;
+
+	i915_gem_object_unpin_pages(obj);
+out_put:
+	i915_gem_object_put(obj);
+
+	return err;
+}
+
+static int igt_lmem_create_with_ps(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	int err = 0;
+	u32 ps;
+
+	for (ps = PAGE_SIZE; ps <= SZ_1G; ps <<= 1) {
+		struct drm_i915_gem_object *obj;
+		dma_addr_t daddr;
+
+		obj = __i915_gem_object_create_lmem_with_ps(i915, ps, ps, 0);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			if (err == -ENXIO || err == -E2BIG) {
+				pr_info("%s not enough lmem for ps(%u) err=%d\n",
+					__func__, ps, err);
+				err = 0;
+			}
+
+			break;
+		}
+
+		if (obj->base.size != ps) {
+			pr_err("%s size(%zu) != ps(%u)\n",
+			       __func__, obj->base.size, ps);
+			err = -EINVAL;
+			goto out_put;
+		}
+
+		i915_gem_object_lock(obj, NULL);
+		err = i915_gem_object_pin_pages(obj);
+		if (err) {
+			if (err == -ENXIO || err == -E2BIG || err == -ENOMEM) {
+				pr_info("%s not enough lmem for ps(%u) err=%d\n",
+					__func__, ps, err);
+				err = 0;
+			}
+			goto out_put;
+		}
+
+		daddr = i915_gem_object_get_dma_address(obj, 0);
+		if (!IS_ALIGNED(daddr, ps)) {
+			pr_err("%s daddr(%pa) not aligned with ps(%u)\n",
+			       __func__, &daddr, ps);
+			err = -EINVAL;
+			goto out_unpin;
+		}
+
+out_unpin:
+		i915_gem_object_unpin_pages(obj);
+		__i915_gem_object_put_pages(obj);
+out_put:
+		i915_gem_object_unlock(obj);
+		i915_gem_object_put(obj);
+
+		if (err)
+			break;
+	}
+
+	return err;
+}
+
+static int igt_lmem_create_cleared_cpu(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	I915_RND_STATE(prng);
+	IGT_TIMEOUT(end_time);
+	u32 size, i;
+	int err;
+
+	i915_gem_drain_freed_objects(i915);
+
+	size = max_t(u32, PAGE_SIZE, i915_prandom_u32_max_state(SZ_32M, &prng));
+	size = round_up(size, PAGE_SIZE);
+	i = 0;
+
+	do {
+		struct drm_i915_gem_object *obj;
+		unsigned int flags;
+		u32 dword, val;
+		void *vaddr;
+
+		/*
+		 * Alternate between cleared and uncleared allocations, while
+		 * also dirtying the pages each time to check that the pages are
+		 * always cleared if requested, since we should get some overlap
+		 * of the underlying pages, if not all, since we are the only
+		 * user.
+		 */
+
+		flags = I915_BO_ALLOC_CPU_CLEAR;
+		if (i & 1)
+			flags = 0;
+
+		obj = i915_gem_object_create_lmem(i915, size, flags);
+		if (IS_ERR(obj))
+			return PTR_ERR(obj);
+
+		i915_gem_object_lock(obj, NULL);
+		err = i915_gem_object_pin_pages(obj);
+		if (err)
+			goto out_put;
+
+		dword = i915_prandom_u32_max_state(PAGE_SIZE / sizeof(u32),
+						   &prng);
+
+		if (flags & I915_BO_ALLOC_CPU_CLEAR) {
+			err = igt_cpu_check(obj, dword, 0);
+			if (err) {
+				pr_err("%s failed with size=%u, flags=%u\n",
+				       __func__, size, flags);
+				goto out_unpin;
+			}
+		}
+
+		vaddr = i915_gem_object_pin_map(obj, I915_MAP_WC);
+		if (IS_ERR(vaddr)) {
+			err = PTR_ERR(vaddr);
+			goto out_unpin;
+		}
+
+		val = prandom_u32_state(&prng);
+
+		memset32(vaddr, val, obj->base.size / sizeof(u32));
+
+		i915_gem_object_flush_map(obj);
+		i915_gem_object_unpin_map(obj);
+out_unpin:
+		i915_gem_object_unpin_pages(obj);
+		__i915_gem_object_put_pages(obj);
+out_put:
+		i915_gem_object_unlock(obj);
+		i915_gem_object_put(obj);
+
+		if (err)
+			break;
+		++i;
+	} while (!__igt_timeout(end_time, NULL));
+
+	pr_info("%s completed (%u) iterations\n", __func__, i);
+
+	return err;
+}
+
+static int igt_lmem_write_gpu(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct i915_gem_context *ctx;
+	struct file *file;
+	I915_RND_STATE(prng);
+	u32 sz;
+	int err;
+
+	file = mock_file(i915);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	ctx = live_context(i915, file);
+	if (IS_ERR(ctx)) {
+		err = PTR_ERR(ctx);
+		goto out_file;
+	}
+
+	sz = round_up(prandom_u32_state(&prng) % SZ_32M, PAGE_SIZE);
+
+	obj = i915_gem_object_create_lmem(i915, sz, 0);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto out_file;
+	}
+
+	err = i915_gem_object_pin_pages_unlocked(obj);
+	if (err)
+		goto out_put;
+
+	err = igt_gpu_write(ctx, obj);
+	if (err)
+		pr_err("igt_gpu_write failed(%d)\n", err);
+
+	i915_gem_object_unpin_pages(obj);
+out_put:
+	i915_gem_object_put(obj);
+out_file:
+	fput(file);
+	return err;
+}
+
+static struct intel_engine_cs *
+random_engine_class(struct drm_i915_private *i915,
+		    unsigned int class,
+		    struct rnd_state *prng)
+{
+	struct intel_engine_cs *engine;
+	unsigned int count;
+
+	count = 0;
+	for (engine = intel_engine_lookup_user(i915, class, 0);
+	     engine && engine->uabi_class == class;
+	     engine = rb_entry_safe(rb_next(&engine->uabi_node),
+				    typeof(*engine), uabi_node))
+		count++;
+
+	count = i915_prandom_u32_max_state(count, prng);
+	return intel_engine_lookup_user(i915, class, count);
+}
+
+static int igt_lmem_write_cpu(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	I915_RND_STATE(prng);
+	IGT_TIMEOUT(end_time);
+	u32 bytes[] = {
+		0, /* rng placeholder */
+		sizeof(u32),
+		sizeof(u64),
+		64, /* cl */
+		PAGE_SIZE,
+		PAGE_SIZE - sizeof(u32),
+		PAGE_SIZE - sizeof(u64),
+		PAGE_SIZE - 64,
+	};
+	struct intel_engine_cs *engine;
+	struct i915_request *rq;
+	u32 *vaddr;
+	u32 sz;
+	u32 i;
+	int *order;
+	int count;
+	int err;
+
+	engine = random_engine_class(i915, I915_ENGINE_CLASS_COPY, &prng);
+	if (!engine)
+		return 0;
+
+	pr_info("%s: using %s\n", __func__, engine->name);
+
+	sz = round_up(prandom_u32_state(&prng) % SZ_32M, PAGE_SIZE);
+	sz = max_t(u32, 2 * PAGE_SIZE, sz);
+
+	obj = i915_gem_object_create_lmem(i915, sz, I915_BO_ALLOC_CONTIGUOUS);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vaddr = i915_gem_object_pin_map_unlocked(obj, I915_MAP_WC);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto out_put;
+	}
+
+	i915_gem_object_lock(obj, NULL);
+
+	err = dma_resv_reserve_fences(obj->base.resv, 1);
+	if (err) {
+		i915_gem_object_unlock(obj);
+		goto out_put;
+	}
+
+	/* Put the pages into a known state -- from the gpu for added fun */
+	intel_engine_pm_get(engine);
+	err = intel_context_migrate_clear(engine->gt->migrate.context, NULL,
+					  obj->mm.pages->sgl, I915_CACHE_NONE,
+					  true, 0xdeadbeaf, &rq);
+	if (rq) {
+		dma_resv_add_fence(obj->base.resv, &rq->fence,
+				   DMA_RESV_USAGE_WRITE);
+		i915_request_put(rq);
+	}
+
+	intel_engine_pm_put(engine);
+	if (!err)
+		err = i915_gem_object_set_to_wc_domain(obj, true);
+	i915_gem_object_unlock(obj);
+	if (err)
+		goto out_unpin;
+
+	count = ARRAY_SIZE(bytes);
+	order = i915_random_order(count * count, &prng);
+	if (!order) {
+		err = -ENOMEM;
+		goto out_unpin;
+	}
+
+	/* A random multiple of u32, picked between [64, PAGE_SIZE - 64] */
+	bytes[0] = igt_random_offset(&prng, 64, PAGE_SIZE - 64, 0, sizeof(u32));
+	GEM_BUG_ON(!IS_ALIGNED(bytes[0], sizeof(u32)));
+
+	i = 0;
+	do {
+		u32 offset;
+		u32 align;
+		u32 dword;
+		u32 size;
+		u32 val;
+
+		size = bytes[order[i] % count];
+		i = (i + 1) % (count * count);
+
+		align = bytes[order[i] % count];
+		i = (i + 1) % (count * count);
+
+		align = max_t(u32, sizeof(u32), rounddown_pow_of_two(align));
+
+		offset = igt_random_offset(&prng, 0, obj->base.size,
+					   size, align);
+
+		val = prandom_u32_state(&prng);
+		memset32(vaddr + offset / sizeof(u32), val ^ 0xdeadbeaf,
+			 size / sizeof(u32));
+
+		/*
+		 * Sample random dw -- don't waste precious time reading every
+		 * single dw.
+		 */
+		dword = igt_random_offset(&prng, offset,
+					  offset + size,
+					  sizeof(u32), sizeof(u32));
+		dword /= sizeof(u32);
+		if (vaddr[dword] != (val ^ 0xdeadbeaf)) {
+			pr_err("%s vaddr[%u]=%u, val=%u, size=%u, align=%u, offset=%u\n",
+			       __func__, dword, vaddr[dword], val ^ 0xdeadbeaf,
+			       size, align, offset);
+			err = -EINVAL;
+			break;
+		}
+	} while (!__igt_timeout(end_time, NULL));
+
+out_unpin:
+	i915_gem_object_unpin_map(obj);
+out_put:
+	i915_gem_object_put(obj);
+
+	return err;
+}
+
+static const char *repr_type(u32 type)
+{
+	switch (type) {
+	case I915_MAP_WB:
+		return "WB";
+	case I915_MAP_WC:
+		return "WC";
+	}
+
+	return "";
+}
+
+static struct drm_i915_gem_object *
+create_region_for_mapping(struct intel_memory_region *mr, u64 size, u32 type,
+			  void **out_addr)
+{
+	struct drm_i915_gem_object *obj;
+	void *addr;
+
+	obj = i915_gem_object_create_region(mr, size, 0, 0);
+	if (IS_ERR(obj)) {
+		if (PTR_ERR(obj) == -ENOSPC) /* Stolen memory */
+			return ERR_PTR(-ENODEV);
+		return obj;
+	}
+
+	addr = i915_gem_object_pin_map_unlocked(obj, type);
+	if (IS_ERR(addr)) {
+		i915_gem_object_put(obj);
+		if (PTR_ERR(addr) == -ENXIO)
+			return ERR_PTR(-ENODEV);
+		return addr;
+	}
+
+	*out_addr = addr;
+	return obj;
+}
+
+static int wrap_ktime_compare(const void *A, const void *B)
+{
+	const ktime_t *a = A, *b = B;
+
+	return ktime_compare(*a, *b);
+}
+
+static void igt_memcpy_long(void *dst, const void *src, size_t size)
+{
+	unsigned long *tmp = dst;
+	const unsigned long *s = src;
+
+	size = size / sizeof(unsigned long);
+	while (size--)
+		*tmp++ = *s++;
+}
+
+static inline void igt_memcpy(void *dst, const void *src, size_t size)
+{
+	memcpy(dst, src, size);
+}
+
+static inline void igt_memcpy_from_wc(void *dst, const void *src, size_t size)
+{
+	i915_memcpy_from_wc(dst, src, size);
+}
+
+static int _perf_memcpy(struct intel_memory_region *src_mr,
+			struct intel_memory_region *dst_mr,
+			u64 size, u32 src_type, u32 dst_type)
+{
+	struct drm_i915_private *i915 = src_mr->i915;
+	const struct {
+		const char *name;
+		void (*copy)(void *dst, const void *src, size_t size);
+		bool skip;
+	} tests[] = {
+		{
+			"memcpy",
+			igt_memcpy,
+		},
+		{
+			"memcpy_long",
+			igt_memcpy_long,
+		},
+		{
+			"memcpy_from_wc",
+			igt_memcpy_from_wc,
+			!i915_has_memcpy_from_wc(),
+		},
+	};
+	struct drm_i915_gem_object *src, *dst;
+	void *src_addr, *dst_addr;
+	int ret = 0;
+	int i;
+
+	src = create_region_for_mapping(src_mr, size, src_type, &src_addr);
+	if (IS_ERR(src)) {
+		ret = PTR_ERR(src);
+		goto out;
+	}
+
+	dst = create_region_for_mapping(dst_mr, size, dst_type, &dst_addr);
+	if (IS_ERR(dst)) {
+		ret = PTR_ERR(dst);
+		goto out_unpin_src;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(tests); ++i) {
+		ktime_t t[5];
+		int pass;
+
+		if (tests[i].skip)
+			continue;
+
+		for (pass = 0; pass < ARRAY_SIZE(t); pass++) {
+			ktime_t t0, t1;
+
+			t0 = ktime_get();
+
+			tests[i].copy(dst_addr, src_addr, size);
+
+			t1 = ktime_get();
+			t[pass] = ktime_sub(t1, t0);
+		}
+
+		sort(t, ARRAY_SIZE(t), sizeof(*t), wrap_ktime_compare, NULL);
+		if (t[0] <= 0) {
+			/* ignore the impossible to protect our sanity */
+			pr_debug("Skipping %s src(%s, %s) -> dst(%s, %s) %14s %4lluKiB copy, unstable measurement [%lld, %lld]\n",
+				 __func__,
+				 src_mr->name, repr_type(src_type),
+				 dst_mr->name, repr_type(dst_type),
+				 tests[i].name, size >> 10,
+				 t[0], t[4]);
+			continue;
+		}
+
+		pr_info("%s src(%s, %s) -> dst(%s, %s) %14s %4llu KiB copy: %5lld MiB/s\n",
+			__func__,
+			src_mr->name, repr_type(src_type),
+			dst_mr->name, repr_type(dst_type),
+			tests[i].name, size >> 10,
+			div64_u64(mul_u32_u32(4 * size,
+					      1000 * 1000 * 1000),
+				  t[1] + 2 * t[2] + t[3]) >> 20);
+
+		cond_resched();
+	}
+
+	i915_gem_object_unpin_map(dst);
+	i915_gem_object_put(dst);
+out_unpin_src:
+	i915_gem_object_unpin_map(src);
+	i915_gem_object_put(src);
+
+	i915_gem_drain_freed_objects(i915);
+out:
+	if (ret == -ENODEV)
+		ret = 0;
+
+	return ret;
+}
+
+static int perf_memcpy(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	static const u32 types[] = {
+		I915_MAP_WB,
+		I915_MAP_WC,
+	};
+	static const u32 sizes[] = {
+		SZ_4K,
+		SZ_64K,
+		SZ_4M,
+	};
+	struct intel_memory_region *src_mr, *dst_mr;
+	int src_id, dst_id;
+	int i, j, k;
+	int ret;
+
+	for_each_memory_region(src_mr, i915, src_id) {
+		for_each_memory_region(dst_mr, i915, dst_id) {
+			for (i = 0; i < ARRAY_SIZE(sizes); ++i) {
+				for (j = 0; j < ARRAY_SIZE(types); ++j) {
+					for (k = 0; k < ARRAY_SIZE(types); ++k) {
+						ret = _perf_memcpy(src_mr,
+								   dst_mr,
+								   sizes[i],
+								   types[j],
+								   types[k]);
+						if (ret)
+							return ret;
+					}
+				}
+			}
+		}
+	}
+
+	return 0;
+}
+
+int intel_memory_region_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_mock_reserve),
+		SUBTEST(igt_mock_fill),
+		SUBTEST(igt_mock_contiguous),
+		SUBTEST(igt_mock_splintered_region),
+		SUBTEST(igt_mock_max_segment),
+		SUBTEST(igt_mock_io_size),
+	};
+	struct intel_memory_region *mem;
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	mem = mock_region_create(i915, 0, SZ_2G, I915_GTT_PAGE_SIZE_4K, 0, 0);
+	if (IS_ERR(mem)) {
+		pr_err("failed to create memory region\n");
+		err = PTR_ERR(mem);
+		goto out_unref;
+	}
+
+	err = i915_subtests(tests, mem);
+
+	intel_memory_region_destroy(mem);
+out_unref:
+	mock_destroy_device(i915);
+	return err;
+}
+
+int intel_memory_region_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_lmem_create),
+		SUBTEST(igt_lmem_create_with_ps),
+		SUBTEST(igt_lmem_create_cleared_cpu),
+		SUBTEST(igt_lmem_write_cpu),
+		SUBTEST(igt_lmem_write_gpu),
+	};
+
+	if (!HAS_LMEM(i915)) {
+		pr_info("device lacks LMEM support, skipping\n");
+		return 0;
+	}
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return i915_live_subtests(tests, i915);
+}
+
+int intel_memory_region_perf_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(perf_memcpy),
+	};
+
+	if (intel_gt_is_wedged(to_gt(i915)))
+		return 0;
+
+	return i915_live_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/intel_scheduler_helpers.c b/drivers/gpu/drm/i915/selftests/intel_scheduler_helpers.c
new file mode 100644
index 000000000..2990dd4d4
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/intel_scheduler_helpers.c
@@ -0,0 +1,99 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+//#include "gt/intel_engine_user.h"
+#include "gt/intel_gt.h"
+#include "i915_drv.h"
+#include "i915_selftest.h"
+
+#include "selftests/intel_scheduler_helpers.h"
+
+#define REDUCED_TIMESLICE	5
+#define REDUCED_PREEMPT		10
+#define WAIT_FOR_RESET_TIME	10000
+
+struct intel_engine_cs *intel_selftest_find_any_engine(struct intel_gt *gt)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, gt, id)
+		return engine;
+
+	pr_err("No valid engine found!\n");
+	return NULL;
+}
+
+int intel_selftest_modify_policy(struct intel_engine_cs *engine,
+				 struct intel_selftest_saved_policy *saved,
+				 enum selftest_scheduler_modify modify_type)
+{
+	int err;
+
+	saved->reset = engine->i915->params.reset;
+	saved->flags = engine->flags;
+	saved->timeslice = engine->props.timeslice_duration_ms;
+	saved->preempt_timeout = engine->props.preempt_timeout_ms;
+
+	switch (modify_type) {
+	case SELFTEST_SCHEDULER_MODIFY_FAST_RESET:
+		/*
+		 * Enable force pre-emption on time slice expiration
+		 * together with engine reset on pre-emption timeout.
+		 * This is required to make the GuC notice and reset
+		 * the single hanging context.
+		 * Also, reduce the preemption timeout to something
+		 * small to speed the test up.
+		 */
+		engine->i915->params.reset = 2;
+		engine->flags |= I915_ENGINE_WANT_FORCED_PREEMPTION;
+		engine->props.timeslice_duration_ms = REDUCED_TIMESLICE;
+		engine->props.preempt_timeout_ms = REDUCED_PREEMPT;
+		break;
+
+	case SELFTEST_SCHEDULER_MODIFY_NO_HANGCHECK:
+		engine->props.preempt_timeout_ms = 0;
+		break;
+
+	default:
+		pr_err("Invalid scheduler policy modification type: %d!\n", modify_type);
+		return -EINVAL;
+	}
+
+	if (!intel_engine_uses_guc(engine))
+		return 0;
+
+	err = intel_guc_global_policies_update(&engine->gt->uc.guc);
+	if (err)
+		intel_selftest_restore_policy(engine, saved);
+
+	return err;
+}
+
+int intel_selftest_restore_policy(struct intel_engine_cs *engine,
+				  struct intel_selftest_saved_policy *saved)
+{
+	/* Restore the original policies */
+	engine->i915->params.reset = saved->reset;
+	engine->flags = saved->flags;
+	engine->props.timeslice_duration_ms = saved->timeslice;
+	engine->props.preempt_timeout_ms = saved->preempt_timeout;
+
+	if (!intel_engine_uses_guc(engine))
+		return 0;
+
+	return intel_guc_global_policies_update(&engine->gt->uc.guc);
+}
+
+int intel_selftest_wait_for_rq(struct i915_request *rq)
+{
+	long ret;
+
+	ret = i915_request_wait(rq, 0, WAIT_FOR_RESET_TIME);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/selftests/intel_scheduler_helpers.h b/drivers/gpu/drm/i915/selftests/intel_scheduler_helpers.h
new file mode 100644
index 000000000..ae60bb507
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/intel_scheduler_helpers.h
@@ -0,0 +1,35 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _INTEL_SELFTEST_SCHEDULER_HELPERS_H_
+#define _INTEL_SELFTEST_SCHEDULER_HELPERS_H_
+
+#include <linux/types.h>
+
+struct i915_request;
+struct intel_engine_cs;
+struct intel_gt;
+
+struct intel_selftest_saved_policy {
+	u32 flags;
+	u32 reset;
+	u64 timeslice;
+	u64 preempt_timeout;
+};
+
+enum selftest_scheduler_modify {
+	SELFTEST_SCHEDULER_MODIFY_NO_HANGCHECK = 0,
+	SELFTEST_SCHEDULER_MODIFY_FAST_RESET,
+};
+
+struct intel_engine_cs *intel_selftest_find_any_engine(struct intel_gt *gt);
+int intel_selftest_modify_policy(struct intel_engine_cs *engine,
+				 struct intel_selftest_saved_policy *saved,
+				 enum selftest_scheduler_modify modify_type);
+int intel_selftest_restore_policy(struct intel_engine_cs *engine,
+				  struct intel_selftest_saved_policy *saved);
+int intel_selftest_wait_for_rq(struct i915_request *rq);
+
+#endif
diff --git a/drivers/gpu/drm/i915/selftests/intel_uncore.c b/drivers/gpu/drm/i915/selftests/intel_uncore.c
new file mode 100644
index 000000000..fda9bb79c
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/intel_uncore.c
@@ -0,0 +1,350 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "../i915_selftest.h"
+
+static int intel_fw_table_check(const struct intel_forcewake_range *ranges,
+				unsigned int num_ranges,
+				bool is_watertight)
+{
+	unsigned int i;
+	s32 prev;
+
+	for (i = 0, prev = -1; i < num_ranges; i++, ranges++) {
+		/* Check that the table is watertight */
+		if (is_watertight && (prev + 1) != (s32)ranges->start) {
+			pr_err("%s: entry[%d]:(%x, %x) is not watertight to previous (%x)\n",
+			       __func__, i, ranges->start, ranges->end, prev);
+			return -EINVAL;
+		}
+
+		/* Check that the table never goes backwards */
+		if (prev >= (s32)ranges->start) {
+			pr_err("%s: entry[%d]:(%x, %x) is less than the previous (%x)\n",
+			       __func__, i, ranges->start, ranges->end, prev);
+			return -EINVAL;
+		}
+
+		/* Check that the entry is valid */
+		if (ranges->start >= ranges->end) {
+			pr_err("%s: entry[%d]:(%x, %x) has negative length\n",
+			       __func__, i, ranges->start, ranges->end);
+			return -EINVAL;
+		}
+
+		prev = ranges->end;
+	}
+
+	return 0;
+}
+
+static int intel_shadow_table_check(void)
+{
+	struct {
+		const struct i915_range *regs;
+		unsigned int size;
+	} range_lists[] = {
+		{ gen8_shadowed_regs, ARRAY_SIZE(gen8_shadowed_regs) },
+		{ gen11_shadowed_regs, ARRAY_SIZE(gen11_shadowed_regs) },
+		{ gen12_shadowed_regs, ARRAY_SIZE(gen12_shadowed_regs) },
+		{ dg2_shadowed_regs, ARRAY_SIZE(dg2_shadowed_regs) },
+		{ pvc_shadowed_regs, ARRAY_SIZE(pvc_shadowed_regs) },
+	};
+	const struct i915_range *range;
+	unsigned int i, j;
+	s32 prev;
+
+	for (j = 0; j < ARRAY_SIZE(range_lists); ++j) {
+		range = range_lists[j].regs;
+		for (i = 0, prev = -1; i < range_lists[j].size; i++, range++) {
+			if (range->end < range->start) {
+				pr_err("%s: range[%d]:(%06x-%06x) has end before start\n",
+				       __func__, i, range->start, range->end);
+				return -EINVAL;
+			}
+
+			if (prev >= (s32)range->start) {
+				pr_err("%s: range[%d]:(%06x-%06x) is before end of previous (%06x)\n",
+				       __func__, i, range->start, range->end, prev);
+				return -EINVAL;
+			}
+
+			if (range->start % 4) {
+				pr_err("%s: range[%d]:(%06x-%06x) has non-dword-aligned start\n",
+				       __func__, i, range->start, range->end);
+				return -EINVAL;
+			}
+
+			prev = range->end;
+		}
+	}
+
+	return 0;
+}
+
+int intel_uncore_mock_selftests(void)
+{
+	struct {
+		const struct intel_forcewake_range *ranges;
+		unsigned int num_ranges;
+		bool is_watertight;
+	} fw[] = {
+		{ __vlv_fw_ranges, ARRAY_SIZE(__vlv_fw_ranges), false },
+		{ __chv_fw_ranges, ARRAY_SIZE(__chv_fw_ranges), false },
+		{ __gen9_fw_ranges, ARRAY_SIZE(__gen9_fw_ranges), true },
+		{ __gen11_fw_ranges, ARRAY_SIZE(__gen11_fw_ranges), true },
+		{ __gen12_fw_ranges, ARRAY_SIZE(__gen12_fw_ranges), true },
+		{ __xehp_fw_ranges, ARRAY_SIZE(__xehp_fw_ranges), true },
+		{ __pvc_fw_ranges, ARRAY_SIZE(__pvc_fw_ranges), true },
+	};
+	int err, i;
+
+	for (i = 0; i < ARRAY_SIZE(fw); i++) {
+		err = intel_fw_table_check(fw[i].ranges,
+					   fw[i].num_ranges,
+					   fw[i].is_watertight);
+		if (err)
+			return err;
+	}
+
+	err = intel_shadow_table_check();
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static int live_forcewake_ops(void *arg)
+{
+	static const struct reg {
+		const char *name;
+		u8 min_graphics_ver;
+		u8 max_graphics_ver;
+		unsigned long platforms;
+		unsigned int offset;
+	} registers[] = {
+		{
+			"RING_START",
+			6, 7,
+			0x38,
+		},
+		{
+			"RING_MI_MODE",
+			8, U8_MAX,
+			0x9c,
+		}
+	};
+	const struct reg *r;
+	struct intel_gt *gt = arg;
+	struct intel_uncore_forcewake_domain *domain;
+	struct intel_uncore *uncore = gt->uncore;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	intel_wakeref_t wakeref;
+	unsigned int tmp;
+	int err = 0;
+
+	GEM_BUG_ON(gt->awake);
+
+	/* vlv/chv with their pcu behave differently wrt reads */
+	if (IS_VALLEYVIEW(gt->i915) || IS_CHERRYVIEW(gt->i915)) {
+		pr_debug("PCU fakes forcewake badly; skipping\n");
+		return 0;
+	}
+
+	/*
+	 * Not quite as reliable across the gen as one would hope.
+	 *
+	 * Either our theory of operation is incorrect, or there remain
+	 * external parties interfering with the powerwells.
+	 *
+	 * https://bugs.freedesktop.org/show_bug.cgi?id=110210
+	 */
+	if (!IS_ENABLED(CONFIG_DRM_I915_SELFTEST_BROKEN))
+		return 0;
+
+	/* We have to pick carefully to get the exact behaviour we need */
+	for (r = registers; r->name; r++)
+		if (IS_GRAPHICS_VER(gt->i915, r->min_graphics_ver, r->max_graphics_ver))
+			break;
+	if (!r->name) {
+		pr_debug("Forcewaked register not known for %s; skipping\n",
+			 intel_platform_name(INTEL_INFO(gt->i915)->platform));
+		return 0;
+	}
+
+	wakeref = intel_runtime_pm_get(uncore->rpm);
+
+	for_each_fw_domain(domain, uncore, tmp) {
+		smp_store_mb(domain->active, false);
+		if (!hrtimer_cancel(&domain->timer))
+			continue;
+
+		intel_uncore_fw_release_timer(&domain->timer);
+	}
+
+	for_each_engine(engine, gt, id) {
+		i915_reg_t mmio = _MMIO(engine->mmio_base + r->offset);
+		u32 __iomem *reg = uncore->regs + engine->mmio_base + r->offset;
+		enum forcewake_domains fw_domains;
+		u32 val;
+
+		if (!engine->default_state)
+			continue;
+
+		fw_domains = intel_uncore_forcewake_for_reg(uncore, mmio,
+							    FW_REG_READ);
+		if (!fw_domains)
+			continue;
+
+		for_each_fw_domain_masked(domain, fw_domains, uncore, tmp) {
+			if (!domain->wake_count)
+				continue;
+
+			pr_err("fw_domain %s still active, aborting test!\n",
+			       intel_uncore_forcewake_domain_to_str(domain->id));
+			err = -EINVAL;
+			goto out_rpm;
+		}
+
+		intel_uncore_forcewake_get(uncore, fw_domains);
+		val = readl(reg);
+		intel_uncore_forcewake_put(uncore, fw_domains);
+
+		/* Flush the forcewake release (delayed onto a timer) */
+		for_each_fw_domain_masked(domain, fw_domains, uncore, tmp) {
+			smp_store_mb(domain->active, false);
+			if (hrtimer_cancel(&domain->timer))
+				intel_uncore_fw_release_timer(&domain->timer);
+
+			preempt_disable();
+			err = wait_ack_clear(domain, FORCEWAKE_KERNEL);
+			preempt_enable();
+			if (err) {
+				pr_err("Failed to clear fw_domain %s\n",
+				       intel_uncore_forcewake_domain_to_str(domain->id));
+				goto out_rpm;
+			}
+		}
+
+		if (!val) {
+			pr_err("%s:%s was zero while fw was held!\n",
+			       engine->name, r->name);
+			err = -EINVAL;
+			goto out_rpm;
+		}
+
+		/* We then expect the read to return 0 outside of the fw */
+		if (wait_for(readl(reg) == 0, 100)) {
+			pr_err("%s:%s=%0x, fw_domains 0x%x still up after 100ms!\n",
+			       engine->name, r->name, readl(reg), fw_domains);
+			err = -ETIMEDOUT;
+			goto out_rpm;
+		}
+	}
+
+out_rpm:
+	intel_runtime_pm_put(uncore->rpm, wakeref);
+	return err;
+}
+
+static int live_forcewake_domains(void *arg)
+{
+#define FW_RANGE 0x40000
+	struct intel_gt *gt = arg;
+	struct intel_uncore *uncore = gt->uncore;
+	unsigned long *valid;
+	u32 offset;
+	int err;
+
+	if (!HAS_FPGA_DBG_UNCLAIMED(gt->i915) &&
+	    !IS_VALLEYVIEW(gt->i915) &&
+	    !IS_CHERRYVIEW(gt->i915))
+		return 0;
+
+	/*
+	 * This test may lockup the machine or cause GPU hangs afterwards.
+	 */
+	if (!IS_ENABLED(CONFIG_DRM_I915_SELFTEST_BROKEN))
+		return 0;
+
+	valid = bitmap_zalloc(FW_RANGE, GFP_KERNEL);
+	if (!valid)
+		return -ENOMEM;
+
+	intel_uncore_forcewake_get(uncore, FORCEWAKE_ALL);
+
+	check_for_unclaimed_mmio(uncore);
+	for (offset = 0; offset < FW_RANGE; offset += 4) {
+		i915_reg_t reg = { offset };
+
+		intel_uncore_posting_read_fw(uncore, reg);
+		if (!check_for_unclaimed_mmio(uncore))
+			set_bit(offset, valid);
+	}
+
+	intel_uncore_forcewake_put(uncore, FORCEWAKE_ALL);
+
+	err = 0;
+	for_each_set_bit(offset, valid, FW_RANGE) {
+		i915_reg_t reg = { offset };
+
+		iosf_mbi_punit_acquire();
+		intel_uncore_forcewake_reset(uncore);
+		iosf_mbi_punit_release();
+
+		check_for_unclaimed_mmio(uncore);
+
+		intel_uncore_posting_read_fw(uncore, reg);
+		if (check_for_unclaimed_mmio(uncore)) {
+			pr_err("Unclaimed mmio read to register 0x%04x\n",
+			       offset);
+			err = -EINVAL;
+		}
+	}
+
+	bitmap_free(valid);
+	return err;
+}
+
+static int live_fw_table(void *arg)
+{
+	struct intel_gt *gt = arg;
+
+	/* Confirm the table we load is still valid */
+	return intel_fw_table_check(gt->uncore->fw_domains_table,
+				    gt->uncore->fw_domains_table_entries,
+				    GRAPHICS_VER(gt->i915) >= 9);
+}
+
+int intel_uncore_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_fw_table),
+		SUBTEST(live_forcewake_ops),
+		SUBTEST(live_forcewake_domains),
+	};
+
+	return intel_gt_live_subtests(tests, to_gt(i915));
+}
diff --git a/drivers/gpu/drm/i915/selftests/lib_sw_fence.c b/drivers/gpu/drm/i915/selftests/lib_sw_fence.c
new file mode 100644
index 000000000..bf2752cc1
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/lib_sw_fence.c
@@ -0,0 +1,136 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "lib_sw_fence.h"
+
+/* Small library of different fence types useful for writing tests */
+
+static int
+nop_fence_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
+{
+	return NOTIFY_DONE;
+}
+
+void __onstack_fence_init(struct i915_sw_fence *fence,
+			  const char *name,
+			  struct lock_class_key *key)
+{
+	debug_fence_init_onstack(fence);
+
+	__init_waitqueue_head(&fence->wait, name, key);
+	atomic_set(&fence->pending, 1);
+	fence->error = 0;
+	fence->fn = nop_fence_notify;
+}
+
+void onstack_fence_fini(struct i915_sw_fence *fence)
+{
+	if (!fence->fn)
+		return;
+
+	i915_sw_fence_commit(fence);
+	i915_sw_fence_fini(fence);
+}
+
+static void timed_fence_wake(struct timer_list *t)
+{
+	struct timed_fence *tf = from_timer(tf, t, timer);
+
+	i915_sw_fence_commit(&tf->fence);
+}
+
+void timed_fence_init(struct timed_fence *tf, unsigned long expires)
+{
+	onstack_fence_init(&tf->fence);
+
+	timer_setup_on_stack(&tf->timer, timed_fence_wake, 0);
+
+	if (time_after(expires, jiffies))
+		mod_timer(&tf->timer, expires);
+	else
+		i915_sw_fence_commit(&tf->fence);
+}
+
+void timed_fence_fini(struct timed_fence *tf)
+{
+	if (del_timer_sync(&tf->timer))
+		i915_sw_fence_commit(&tf->fence);
+
+	destroy_timer_on_stack(&tf->timer);
+	i915_sw_fence_fini(&tf->fence);
+}
+
+struct heap_fence {
+	struct i915_sw_fence fence;
+	union {
+		struct kref ref;
+		struct rcu_head rcu;
+	};
+};
+
+static int
+heap_fence_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
+{
+	struct heap_fence *h = container_of(fence, typeof(*h), fence);
+
+	switch (state) {
+	case FENCE_COMPLETE:
+		break;
+
+	case FENCE_FREE:
+		heap_fence_put(&h->fence);
+	}
+
+	return NOTIFY_DONE;
+}
+
+struct i915_sw_fence *heap_fence_create(gfp_t gfp)
+{
+	struct heap_fence *h;
+
+	h = kmalloc(sizeof(*h), gfp);
+	if (!h)
+		return NULL;
+
+	i915_sw_fence_init(&h->fence, heap_fence_notify);
+	refcount_set(&h->ref.refcount, 2);
+
+	return &h->fence;
+}
+
+static void heap_fence_release(struct kref *ref)
+{
+	struct heap_fence *h = container_of(ref, typeof(*h), ref);
+
+	i915_sw_fence_fini(&h->fence);
+
+	kfree_rcu(h, rcu);
+}
+
+void heap_fence_put(struct i915_sw_fence *fence)
+{
+	struct heap_fence *h = container_of(fence, typeof(*h), fence);
+
+	kref_put(&h->ref, heap_fence_release);
+}
diff --git a/drivers/gpu/drm/i915/selftests/lib_sw_fence.h b/drivers/gpu/drm/i915/selftests/lib_sw_fence.h
new file mode 100644
index 000000000..e54d6bc23
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/lib_sw_fence.h
@@ -0,0 +1,43 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * lib_sw_fence.h - library routines for testing N:M synchronisation points
+ *
+ * Copyright (C) 2017 Intel Corporation
+ */
+
+#ifndef _LIB_SW_FENCE_H_
+#define _LIB_SW_FENCE_H_
+
+#include <linux/timer.h>
+
+#include "../i915_sw_fence.h"
+
+#ifdef CONFIG_LOCKDEP
+#define onstack_fence_init(fence)				\
+do {								\
+	static struct lock_class_key __key;			\
+								\
+	__onstack_fence_init((fence), #fence, &__key);	\
+} while (0)
+#else
+#define onstack_fence_init(fence)				\
+	__onstack_fence_init((fence), NULL, NULL)
+#endif
+
+void __onstack_fence_init(struct i915_sw_fence *fence,
+			  const char *name,
+			  struct lock_class_key *key);
+void onstack_fence_fini(struct i915_sw_fence *fence);
+
+struct timed_fence {
+	struct i915_sw_fence fence;
+	struct timer_list timer;
+};
+
+void timed_fence_init(struct timed_fence *tf, unsigned long expires);
+void timed_fence_fini(struct timed_fence *tf);
+
+struct i915_sw_fence *heap_fence_create(gfp_t gfp);
+void heap_fence_put(struct i915_sw_fence *fence);
+
+#endif /* _LIB_SW_FENCE_H_ */
diff --git a/drivers/gpu/drm/i915/selftests/librapl.c b/drivers/gpu/drm/i915/selftests/librapl.c
new file mode 100644
index 000000000..eb03b5b28
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/librapl.c
@@ -0,0 +1,34 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <asm/msr.h>
+
+#include "i915_drv.h"
+#include "librapl.h"
+
+bool librapl_supported(const struct drm_i915_private *i915)
+{
+	/* Discrete cards require hwmon integration */
+	if (IS_DGFX(i915))
+		return false;
+
+	return librapl_energy_uJ();
+}
+
+u64 librapl_energy_uJ(void)
+{
+	unsigned long long power;
+	u32 units;
+
+	if (rdmsrl_safe(MSR_RAPL_POWER_UNIT, &power))
+		return 0;
+
+	units = (power & 0x1f00) >> 8;
+
+	if (rdmsrl_safe(MSR_PP1_ENERGY_STATUS, &power))
+		return 0;
+
+	return (1000000 * power) >> units; /* convert to uJ */
+}
diff --git a/drivers/gpu/drm/i915/selftests/librapl.h b/drivers/gpu/drm/i915/selftests/librapl.h
new file mode 100644
index 000000000..e3b24fad0
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/librapl.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef SELFTEST_LIBRAPL_H
+#define SELFTEST_LIBRAPL_H
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+
+bool librapl_supported(const struct drm_i915_private *i915);
+
+u64 librapl_energy_uJ(void);
+
+#endif /* SELFTEST_LIBRAPL_H */
diff --git a/drivers/gpu/drm/i915/selftests/mock_drm.h b/drivers/gpu/drm/i915/selftests/mock_drm.h
new file mode 100644
index 000000000..9916b6f95
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_drm.h
@@ -0,0 +1,45 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MOCK_DRM_H
+#define __MOCK_DRM_H
+
+#include <drm/drm_file.h>
+
+#include "i915_drv.h"
+
+struct drm_file;
+struct file;
+
+static inline struct file *mock_file(struct drm_i915_private *i915)
+{
+	return mock_drm_getfile(i915->drm.primary, O_RDWR);
+}
+
+static inline struct drm_file *to_drm_file(struct file *f)
+{
+	return f->private_data;
+}
+
+#endif /* !__MOCK_DRM_H */
diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.c b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
new file mode 100644
index 000000000..fff11c90f
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
@@ -0,0 +1,252 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/pm_domain.h>
+#include <linux/pm_runtime.h>
+#include <linux/iommu.h>
+
+#include <drm/drm_managed.h>
+
+#include "gt/intel_gt.h"
+#include "gt/intel_gt_requests.h"
+#include "gt/mock_engine.h"
+#include "intel_memory_region.h"
+#include "intel_region_ttm.h"
+
+#include "mock_request.h"
+#include "mock_gem_device.h"
+#include "mock_gtt.h"
+#include "mock_uncore.h"
+#include "mock_region.h"
+
+#include "gem/selftests/mock_context.h"
+#include "gem/selftests/mock_gem_object.h"
+
+void mock_device_flush(struct drm_i915_private *i915)
+{
+	struct intel_gt *gt = to_gt(i915);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	do {
+		for_each_engine(engine, gt, id)
+			mock_engine_flush(engine);
+	} while (intel_gt_retire_requests_timeout(gt, MAX_SCHEDULE_TIMEOUT,
+						  NULL));
+}
+
+static void mock_device_release(struct drm_device *dev)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+
+	if (!i915->do_release)
+		goto out;
+
+	mock_device_flush(i915);
+	intel_gt_driver_remove(to_gt(i915));
+
+	i915_gem_drain_workqueue(i915);
+	i915_gem_drain_freed_objects(i915);
+
+	mock_fini_ggtt(to_gt(i915)->ggtt);
+	destroy_workqueue(i915->wq);
+
+	intel_region_ttm_device_fini(i915);
+	intel_gt_driver_late_release_all(i915);
+	intel_memory_regions_driver_release(i915);
+
+	drm_mode_config_cleanup(&i915->drm);
+
+out:
+	i915_params_free(&i915->params);
+}
+
+static const struct drm_driver mock_driver = {
+	.name = "mock",
+	.driver_features = DRIVER_GEM,
+	.release = mock_device_release,
+};
+
+static void release_dev(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+
+	kfree(pdev);
+}
+
+static int pm_domain_resume(struct device *dev)
+{
+	return pm_generic_runtime_resume(dev);
+}
+
+static int pm_domain_suspend(struct device *dev)
+{
+	return pm_generic_runtime_suspend(dev);
+}
+
+static struct dev_pm_domain pm_domain = {
+	.ops = {
+		.runtime_suspend = pm_domain_suspend,
+		.runtime_resume = pm_domain_resume,
+	},
+};
+
+static void mock_gt_probe(struct drm_i915_private *i915)
+{
+	i915->gt[0] = &i915->gt0;
+	i915->gt[0]->name = "Mock GT";
+}
+
+struct drm_i915_private *mock_gem_device(void)
+{
+#if IS_ENABLED(CONFIG_IOMMU_API) && defined(CONFIG_INTEL_IOMMU)
+	static struct dev_iommu fake_iommu = { .priv = (void *)-1 };
+#endif
+	struct drm_i915_private *i915;
+	struct pci_dev *pdev;
+	int ret;
+
+	pdev = kzalloc(sizeof(*pdev), GFP_KERNEL);
+	if (!pdev)
+		return NULL;
+	device_initialize(&pdev->dev);
+	pdev->class = PCI_BASE_CLASS_DISPLAY << 16;
+	pdev->dev.release = release_dev;
+	dev_set_name(&pdev->dev, "mock");
+	dma_coerce_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64));
+
+#if IS_ENABLED(CONFIG_IOMMU_API) && defined(CONFIG_INTEL_IOMMU)
+	/* HACK to disable iommu for the fake device; force identity mapping */
+	pdev->dev.iommu = &fake_iommu;
+#endif
+	if (!devres_open_group(&pdev->dev, NULL, GFP_KERNEL)) {
+		put_device(&pdev->dev);
+		return NULL;
+	}
+
+	i915 = devm_drm_dev_alloc(&pdev->dev, &mock_driver,
+				  struct drm_i915_private, drm);
+	if (IS_ERR(i915)) {
+		pr_err("Failed to allocate mock GEM device: err=%ld\n", PTR_ERR(i915));
+		devres_release_group(&pdev->dev, NULL);
+		put_device(&pdev->dev);
+
+		return NULL;
+	}
+
+	pci_set_drvdata(pdev, i915);
+
+	dev_pm_domain_set(&pdev->dev, &pm_domain);
+	pm_runtime_enable(&pdev->dev);
+	pm_runtime_dont_use_autosuspend(&pdev->dev);
+	if (pm_runtime_enabled(&pdev->dev))
+		WARN_ON(pm_runtime_get_sync(&pdev->dev));
+
+
+	i915_params_copy(&i915->params, &i915_modparams);
+
+	intel_runtime_pm_init_early(&i915->runtime_pm);
+	/* wakeref tracking has significant overhead */
+	i915->runtime_pm.no_wakeref_tracking = true;
+
+	/* Using the global GTT may ask questions about KMS users, so prepare */
+	drm_mode_config_init(&i915->drm);
+
+	RUNTIME_INFO(i915)->graphics.ip.ver = -1;
+
+	RUNTIME_INFO(i915)->page_sizes =
+		I915_GTT_PAGE_SIZE_4K |
+		I915_GTT_PAGE_SIZE_64K |
+		I915_GTT_PAGE_SIZE_2M;
+
+	RUNTIME_INFO(i915)->memory_regions = REGION_SMEM;
+	intel_memory_regions_hw_probe(i915);
+
+	spin_lock_init(&i915->gpu_error.lock);
+
+	i915_gem_init__mm(i915);
+	intel_root_gt_init_early(i915);
+	mock_uncore_init(&i915->uncore, i915);
+	atomic_inc(&to_gt(i915)->wakeref.count); /* disable; no hw support */
+	to_gt(i915)->awake = -ENODEV;
+	mock_gt_probe(i915);
+
+	ret = intel_region_ttm_device_init(i915);
+	if (ret)
+		goto err_ttm;
+
+	i915->wq = alloc_ordered_workqueue("mock", 0);
+	if (!i915->wq)
+		goto err_drv;
+
+	mock_init_contexts(i915);
+
+	/* allocate the ggtt */
+	ret = intel_gt_assign_ggtt(to_gt(i915));
+	if (ret)
+		goto err_unlock;
+
+	mock_init_ggtt(to_gt(i915));
+	to_gt(i915)->vm = i915_vm_get(&to_gt(i915)->ggtt->vm);
+
+	RUNTIME_INFO(i915)->platform_engine_mask = BIT(0);
+	to_gt(i915)->info.engine_mask = BIT(0);
+
+	to_gt(i915)->engine[RCS0] = mock_engine(i915, "mock", RCS0);
+	if (!to_gt(i915)->engine[RCS0])
+		goto err_unlock;
+
+	if (mock_engine_init(to_gt(i915)->engine[RCS0]))
+		goto err_context;
+
+	__clear_bit(I915_WEDGED, &to_gt(i915)->reset.flags);
+	intel_engines_driver_register(i915);
+
+	i915->do_release = true;
+	ida_init(&i915->selftest.mock_region_instances);
+
+	return i915;
+
+err_context:
+	intel_gt_driver_remove(to_gt(i915));
+err_unlock:
+	destroy_workqueue(i915->wq);
+err_drv:
+	intel_region_ttm_device_fini(i915);
+err_ttm:
+	intel_gt_driver_late_release_all(i915);
+	intel_memory_regions_driver_release(i915);
+	drm_mode_config_cleanup(&i915->drm);
+	mock_destroy_device(i915);
+
+	return NULL;
+}
+
+void mock_destroy_device(struct drm_i915_private *i915)
+{
+	struct device *dev = i915->drm.dev;
+
+	devres_release_group(dev, NULL);
+	put_device(dev);
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.h b/drivers/gpu/drm/i915/selftests/mock_gem_device.h
new file mode 100644
index 000000000..953cfe4fa
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __MOCK_GEM_DEVICE_H__
+#define __MOCK_GEM_DEVICE_H__
+
+struct drm_i915_private;
+
+struct drm_i915_private *mock_gem_device(void);
+void mock_device_flush(struct drm_i915_private *i915);
+
+void mock_destroy_device(struct drm_i915_private *i915);
+
+#endif /* !__MOCK_GEM_DEVICE_H__ */
diff --git a/drivers/gpu/drm/i915/selftests/mock_gtt.c b/drivers/gpu/drm/i915/selftests/mock_gtt.c
new file mode 100644
index 000000000..568840e7c
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_gtt.c
@@ -0,0 +1,136 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "mock_gtt.h"
+
+static void mock_insert_page(struct i915_address_space *vm,
+			     dma_addr_t addr,
+			     u64 offset,
+			     enum i915_cache_level level,
+			     u32 flags)
+{
+}
+
+static void mock_insert_entries(struct i915_address_space *vm,
+				struct i915_vma_resource *vma_res,
+				enum i915_cache_level level, u32 flags)
+{
+}
+
+static void mock_bind_ppgtt(struct i915_address_space *vm,
+			    struct i915_vm_pt_stash *stash,
+			    struct i915_vma_resource *vma_res,
+			    enum i915_cache_level cache_level,
+			    u32 flags)
+{
+	GEM_BUG_ON(flags & I915_VMA_GLOBAL_BIND);
+	vma_res->bound_flags |= flags;
+}
+
+static void mock_unbind_ppgtt(struct i915_address_space *vm,
+			      struct i915_vma_resource *vma_res)
+{
+}
+
+static void mock_cleanup(struct i915_address_space *vm)
+{
+}
+
+static void mock_clear_range(struct i915_address_space *vm,
+			     u64 start, u64 length)
+{
+}
+
+struct i915_ppgtt *mock_ppgtt(struct drm_i915_private *i915, const char *name)
+{
+	struct i915_ppgtt *ppgtt;
+
+	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+	if (!ppgtt)
+		return NULL;
+
+	ppgtt->vm.gt = to_gt(i915);
+	ppgtt->vm.i915 = i915;
+	ppgtt->vm.total = round_down(U64_MAX, PAGE_SIZE);
+	ppgtt->vm.dma = i915->drm.dev;
+
+	i915_address_space_init(&ppgtt->vm, VM_CLASS_PPGTT);
+
+	ppgtt->vm.alloc_pt_dma = alloc_pt_dma;
+	ppgtt->vm.alloc_scratch_dma = alloc_pt_dma;
+
+	ppgtt->vm.clear_range = mock_clear_range;
+	ppgtt->vm.insert_page = mock_insert_page;
+	ppgtt->vm.insert_entries = mock_insert_entries;
+	ppgtt->vm.cleanup = mock_cleanup;
+
+	ppgtt->vm.vma_ops.bind_vma    = mock_bind_ppgtt;
+	ppgtt->vm.vma_ops.unbind_vma  = mock_unbind_ppgtt;
+
+	return ppgtt;
+}
+
+static void mock_bind_ggtt(struct i915_address_space *vm,
+			   struct i915_vm_pt_stash *stash,
+			   struct i915_vma_resource *vma_res,
+			   enum i915_cache_level cache_level,
+			   u32 flags)
+{
+}
+
+static void mock_unbind_ggtt(struct i915_address_space *vm,
+			     struct i915_vma_resource *vma_res)
+{
+}
+
+void mock_init_ggtt(struct intel_gt *gt)
+{
+	struct i915_ggtt *ggtt = gt->ggtt;
+
+	ggtt->vm.gt = gt;
+	ggtt->vm.i915 = gt->i915;
+	ggtt->vm.is_ggtt = true;
+
+	ggtt->gmadr = (struct resource) DEFINE_RES_MEM(0, 2048 * PAGE_SIZE);
+	ggtt->mappable_end = resource_size(&ggtt->gmadr);
+	ggtt->vm.total = 4096 * PAGE_SIZE;
+
+	ggtt->vm.alloc_pt_dma = alloc_pt_dma;
+	ggtt->vm.alloc_scratch_dma = alloc_pt_dma;
+
+	ggtt->vm.clear_range = mock_clear_range;
+	ggtt->vm.insert_page = mock_insert_page;
+	ggtt->vm.insert_entries = mock_insert_entries;
+	ggtt->vm.cleanup = mock_cleanup;
+
+	ggtt->vm.vma_ops.bind_vma    = mock_bind_ggtt;
+	ggtt->vm.vma_ops.unbind_vma  = mock_unbind_ggtt;
+
+	i915_address_space_init(&ggtt->vm, VM_CLASS_GGTT);
+}
+
+void mock_fini_ggtt(struct i915_ggtt *ggtt)
+{
+	i915_address_space_fini(&ggtt->vm);
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_gtt.h b/drivers/gpu/drm/i915/selftests/mock_gtt.h
new file mode 100644
index 000000000..d6eb90bd7
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_gtt.h
@@ -0,0 +1,37 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MOCK_GTT_H
+#define __MOCK_GTT_H
+
+struct drm_i915_private;
+struct i915_ggtt;
+struct intel_gt;
+
+void mock_init_ggtt(struct intel_gt *gt);
+void mock_fini_ggtt(struct i915_ggtt *ggtt);
+
+struct i915_ppgtt *mock_ppgtt(struct drm_i915_private *i915, const char *name);
+
+#endif /* !__MOCK_GTT_H */
diff --git a/drivers/gpu/drm/i915/selftests/mock_region.c b/drivers/gpu/drm/i915/selftests/mock_region.c
new file mode 100644
index 000000000..bac21fe84
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_region.c
@@ -0,0 +1,124 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2019-2021 Intel Corporation
+ */
+
+#include <drm/ttm/ttm_placement.h>
+#include <linux/scatterlist.h>
+
+#include "gem/i915_gem_region.h"
+#include "intel_memory_region.h"
+#include "intel_region_ttm.h"
+
+#include "mock_region.h"
+
+static void mock_region_put_pages(struct drm_i915_gem_object *obj,
+				  struct sg_table *pages)
+{
+	i915_refct_sgt_put(obj->mm.rsgt);
+	obj->mm.rsgt = NULL;
+	intel_region_ttm_resource_free(obj->mm.region, obj->mm.res);
+}
+
+static int mock_region_get_pages(struct drm_i915_gem_object *obj)
+{
+	struct sg_table *pages;
+	int err;
+
+	obj->mm.res = intel_region_ttm_resource_alloc(obj->mm.region,
+						      obj->bo_offset,
+						      obj->base.size,
+						      obj->flags);
+	if (IS_ERR(obj->mm.res))
+		return PTR_ERR(obj->mm.res);
+
+	obj->mm.rsgt = intel_region_ttm_resource_to_rsgt(obj->mm.region,
+							 obj->mm.res,
+							 obj->mm.region->min_page_size);
+	if (IS_ERR(obj->mm.rsgt)) {
+		err = PTR_ERR(obj->mm.rsgt);
+		goto err_free_resource;
+	}
+
+	pages = &obj->mm.rsgt->table;
+	__i915_gem_object_set_pages(obj, pages, i915_sg_dma_sizes(pages->sgl));
+
+	return 0;
+
+err_free_resource:
+	intel_region_ttm_resource_free(obj->mm.region, obj->mm.res);
+	return err;
+}
+
+static const struct drm_i915_gem_object_ops mock_region_obj_ops = {
+	.name = "mock-region",
+	.get_pages = mock_region_get_pages,
+	.put_pages = mock_region_put_pages,
+	.release = i915_gem_object_release_memory_region,
+};
+
+static int mock_object_init(struct intel_memory_region *mem,
+			    struct drm_i915_gem_object *obj,
+			    resource_size_t offset,
+			    resource_size_t size,
+			    resource_size_t page_size,
+			    unsigned int flags)
+{
+	static struct lock_class_key lock_class;
+	struct drm_i915_private *i915 = mem->i915;
+
+	if (size > resource_size(&mem->region))
+		return -E2BIG;
+
+	drm_gem_private_object_init(&i915->drm, &obj->base, size);
+	i915_gem_object_init(obj, &mock_region_obj_ops, &lock_class, flags);
+
+	obj->bo_offset = offset;
+
+	obj->read_domains = I915_GEM_DOMAIN_CPU | I915_GEM_DOMAIN_GTT;
+
+	i915_gem_object_set_cache_coherency(obj, I915_CACHE_NONE);
+
+	i915_gem_object_init_memory_region(obj, mem);
+
+	return 0;
+}
+
+static int mock_region_fini(struct intel_memory_region *mem)
+{
+	struct drm_i915_private *i915 = mem->i915;
+	int instance = mem->instance;
+	int ret;
+
+	ret = intel_region_ttm_fini(mem);
+	ida_free(&i915->selftest.mock_region_instances, instance);
+
+	return ret;
+}
+
+static const struct intel_memory_region_ops mock_region_ops = {
+	.init = intel_region_ttm_init,
+	.release = mock_region_fini,
+	.init_object = mock_object_init,
+};
+
+struct intel_memory_region *
+mock_region_create(struct drm_i915_private *i915,
+		   resource_size_t start,
+		   resource_size_t size,
+		   resource_size_t min_page_size,
+		   resource_size_t io_start,
+		   resource_size_t io_size)
+{
+	int instance = ida_alloc_max(&i915->selftest.mock_region_instances,
+				     TTM_NUM_MEM_TYPES - TTM_PL_PRIV - 1,
+				     GFP_KERNEL);
+
+	if (instance < 0)
+		return ERR_PTR(instance);
+
+	return intel_memory_region_create(i915, start, size, min_page_size,
+					  io_start, io_size,
+					  INTEL_MEMORY_MOCK, instance,
+					  &mock_region_ops);
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_region.h b/drivers/gpu/drm/i915/selftests/mock_region.h
new file mode 100644
index 000000000..e36c3a433
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_region.h
@@ -0,0 +1,22 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2019 Intel Corporation
+ */
+
+#ifndef __MOCK_REGION_H
+#define __MOCK_REGION_H
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+struct intel_memory_region;
+
+struct intel_memory_region *
+mock_region_create(struct drm_i915_private *i915,
+		   resource_size_t start,
+		   resource_size_t size,
+		   resource_size_t min_page_size,
+		   resource_size_t io_start,
+		   resource_size_t io_size);
+
+#endif /* !__MOCK_REGION_H */
diff --git a/drivers/gpu/drm/i915/selftests/mock_request.c b/drivers/gpu/drm/i915/selftests/mock_request.c
new file mode 100644
index 000000000..09f747228
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_request.c
@@ -0,0 +1,59 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "gem/selftests/igt_gem_utils.h"
+#include "gt/mock_engine.h"
+
+#include "mock_request.h"
+
+struct i915_request *
+mock_request(struct intel_context *ce, unsigned long delay)
+{
+	struct i915_request *request;
+
+	/* NB the i915->requests slab cache is enlarged to fit mock_request */
+	request = intel_context_create_request(ce);
+	if (IS_ERR(request))
+		return NULL;
+
+	request->mock.delay = delay;
+	return request;
+}
+
+bool mock_cancel_request(struct i915_request *request)
+{
+	struct mock_engine *engine =
+		container_of(request->engine, typeof(*engine), base);
+	bool was_queued;
+
+	spin_lock_irq(&engine->hw_lock);
+	was_queued = !list_empty(&request->mock.link);
+	list_del_init(&request->mock.link);
+	spin_unlock_irq(&engine->hw_lock);
+
+	if (was_queued)
+		i915_request_unsubmit(request);
+
+	return was_queued;
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_request.h b/drivers/gpu/drm/i915/selftests/mock_request.h
new file mode 100644
index 000000000..8907b60c2
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_request.h
@@ -0,0 +1,37 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MOCK_REQUEST__
+#define __MOCK_REQUEST__
+
+#include <linux/list.h>
+
+#include "../i915_request.h"
+
+struct i915_request *
+mock_request(struct intel_context *ce, unsigned long delay);
+
+bool mock_cancel_request(struct i915_request *request);
+
+#endif /* !__MOCK_REQUEST__ */
diff --git a/drivers/gpu/drm/i915/selftests/mock_uncore.c b/drivers/gpu/drm/i915/selftests/mock_uncore.c
new file mode 100644
index 000000000..f2d6be5e1
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_uncore.c
@@ -0,0 +1,49 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "mock_uncore.h"
+
+#define __nop_write(x) \
+static void \
+nop_write##x(struct intel_uncore *uncore, i915_reg_t reg, u##x val, bool trace) { }
+__nop_write(8)
+__nop_write(16)
+__nop_write(32)
+
+#define __nop_read(x) \
+static u##x \
+nop_read##x(struct intel_uncore *uncore, i915_reg_t reg, bool trace) { return 0; }
+__nop_read(8)
+__nop_read(16)
+__nop_read(32)
+__nop_read(64)
+
+void mock_uncore_init(struct intel_uncore *uncore,
+		      struct drm_i915_private *i915)
+{
+	intel_uncore_init_early(uncore, to_gt(i915));
+
+	ASSIGN_RAW_WRITE_MMIO_VFUNCS(uncore, nop);
+	ASSIGN_RAW_READ_MMIO_VFUNCS(uncore, nop);
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_uncore.h b/drivers/gpu/drm/i915/selftests/mock_uncore.h
new file mode 100644
index 000000000..7acf1ef4d
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_uncore.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MOCK_UNCORE_H
+#define __MOCK_UNCORE_H
+
+struct drm_i915_private;
+struct intel_uncore;
+
+void mock_uncore_init(struct intel_uncore *uncore,
+		      struct drm_i915_private *i915);
+
+#endif /* !__MOCK_UNCORE_H */
diff --git a/drivers/gpu/drm/i915/selftests/scatterlist.c b/drivers/gpu/drm/i915/selftests/scatterlist.c
new file mode 100644
index 000000000..d599186d5
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/scatterlist.c
@@ -0,0 +1,380 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/prime_numbers.h>
+#include <linux/random.h>
+
+#include "i915_selftest.h"
+#include "i915_utils.h"
+
+#define PFN_BIAS (1 << 10)
+
+struct pfn_table {
+	struct sg_table st;
+	unsigned long start, end;
+};
+
+typedef unsigned int (*npages_fn_t)(unsigned long n,
+				    unsigned long count,
+				    struct rnd_state *rnd);
+
+static noinline int expect_pfn_sg(struct pfn_table *pt,
+				  npages_fn_t npages_fn,
+				  struct rnd_state *rnd,
+				  const char *who,
+				  unsigned long timeout)
+{
+	struct scatterlist *sg;
+	unsigned long pfn, n;
+
+	pfn = pt->start;
+	for_each_sg(pt->st.sgl, sg, pt->st.nents, n) {
+		struct page *page = sg_page(sg);
+		unsigned int npages = npages_fn(n, pt->st.nents, rnd);
+
+		if (page_to_pfn(page) != pfn) {
+			pr_err("%s: %s left pages out of order, expected pfn %lu, found pfn %lu (using for_each_sg)\n",
+			       __func__, who, pfn, page_to_pfn(page));
+			return -EINVAL;
+		}
+
+		if (sg->length != npages * PAGE_SIZE) {
+			pr_err("%s: %s copied wrong sg length, expected size %lu, found %u (using for_each_sg)\n",
+			       __func__, who, npages * PAGE_SIZE, sg->length);
+			return -EINVAL;
+		}
+
+		if (igt_timeout(timeout, "%s timed out\n", who))
+			return -EINTR;
+
+		pfn += npages;
+	}
+	if (pfn != pt->end) {
+		pr_err("%s: %s finished on wrong pfn, expected %lu, found %lu\n",
+		       __func__, who, pt->end, pfn);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static noinline int expect_pfn_sg_page_iter(struct pfn_table *pt,
+					    const char *who,
+					    unsigned long timeout)
+{
+	struct sg_page_iter sgiter;
+	unsigned long pfn;
+
+	pfn = pt->start;
+	for_each_sg_page(pt->st.sgl, &sgiter, pt->st.nents, 0) {
+		struct page *page = sg_page_iter_page(&sgiter);
+
+		if (page != pfn_to_page(pfn)) {
+			pr_err("%s: %s left pages out of order, expected pfn %lu, found pfn %lu (using for_each_sg_page)\n",
+			       __func__, who, pfn, page_to_pfn(page));
+			return -EINVAL;
+		}
+
+		if (igt_timeout(timeout, "%s timed out\n", who))
+			return -EINTR;
+
+		pfn++;
+	}
+	if (pfn != pt->end) {
+		pr_err("%s: %s finished on wrong pfn, expected %lu, found %lu\n",
+		       __func__, who, pt->end, pfn);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static noinline int expect_pfn_sgtiter(struct pfn_table *pt,
+				       const char *who,
+				       unsigned long timeout)
+{
+	struct sgt_iter sgt;
+	struct page *page;
+	unsigned long pfn;
+
+	pfn = pt->start;
+	for_each_sgt_page(page, sgt, &pt->st) {
+		if (page != pfn_to_page(pfn)) {
+			pr_err("%s: %s left pages out of order, expected pfn %lu, found pfn %lu (using for_each_sgt_page)\n",
+			       __func__, who, pfn, page_to_pfn(page));
+			return -EINVAL;
+		}
+
+		if (igt_timeout(timeout, "%s timed out\n", who))
+			return -EINTR;
+
+		pfn++;
+	}
+	if (pfn != pt->end) {
+		pr_err("%s: %s finished on wrong pfn, expected %lu, found %lu\n",
+		       __func__, who, pt->end, pfn);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int expect_pfn_sgtable(struct pfn_table *pt,
+			      npages_fn_t npages_fn,
+			      struct rnd_state *rnd,
+			      const char *who,
+			      unsigned long timeout)
+{
+	int err;
+
+	err = expect_pfn_sg(pt, npages_fn, rnd, who, timeout);
+	if (err)
+		return err;
+
+	err = expect_pfn_sg_page_iter(pt, who, timeout);
+	if (err)
+		return err;
+
+	err = expect_pfn_sgtiter(pt, who, timeout);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static unsigned int one(unsigned long n,
+			unsigned long count,
+			struct rnd_state *rnd)
+{
+	return 1;
+}
+
+static unsigned int grow(unsigned long n,
+			 unsigned long count,
+			 struct rnd_state *rnd)
+{
+	return n + 1;
+}
+
+static unsigned int shrink(unsigned long n,
+			   unsigned long count,
+			   struct rnd_state *rnd)
+{
+	return count - n;
+}
+
+static unsigned int random(unsigned long n,
+			   unsigned long count,
+			   struct rnd_state *rnd)
+{
+	return 1 + (prandom_u32_state(rnd) % 1024);
+}
+
+static unsigned int random_page_size_pages(unsigned long n,
+					   unsigned long count,
+					   struct rnd_state *rnd)
+{
+	/* 4K, 64K, 2M */
+	static unsigned int page_count[] = {
+		BIT(12) >> PAGE_SHIFT,
+		BIT(16) >> PAGE_SHIFT,
+		BIT(21) >> PAGE_SHIFT,
+	};
+
+	return page_count[(prandom_u32_state(rnd) % 3)];
+}
+
+static inline bool page_contiguous(struct page *first,
+				   struct page *last,
+				   unsigned long npages)
+{
+	return first + npages == last;
+}
+
+static int alloc_table(struct pfn_table *pt,
+		       unsigned long count, unsigned long max,
+		       npages_fn_t npages_fn,
+		       struct rnd_state *rnd,
+		       int alloc_error)
+{
+	struct scatterlist *sg;
+	unsigned long n, pfn;
+
+	if (sg_alloc_table(&pt->st, max,
+			   GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN))
+		return alloc_error;
+
+	/* count should be less than 20 to prevent overflowing sg->length */
+	GEM_BUG_ON(overflows_type(count * PAGE_SIZE, sg->length));
+
+	/* Construct a table where each scatterlist contains different number
+	 * of entries. The idea is to check that we can iterate the individual
+	 * pages from inside the coalesced lists.
+	 */
+	pt->start = PFN_BIAS;
+	pfn = pt->start;
+	sg = pt->st.sgl;
+	for (n = 0; n < count; n++) {
+		unsigned long npages = npages_fn(n, count, rnd);
+
+		/* Nobody expects the Sparse Memmap! */
+		if (!page_contiguous(pfn_to_page(pfn),
+				     pfn_to_page(pfn + npages),
+				     npages)) {
+			sg_free_table(&pt->st);
+			return -ENOSPC;
+		}
+
+		if (n)
+			sg = sg_next(sg);
+		sg_set_page(sg, pfn_to_page(pfn), npages * PAGE_SIZE, 0);
+
+		GEM_BUG_ON(page_to_pfn(sg_page(sg)) != pfn);
+		GEM_BUG_ON(sg->length != npages * PAGE_SIZE);
+		GEM_BUG_ON(sg->offset != 0);
+
+		pfn += npages;
+	}
+	sg_mark_end(sg);
+	pt->st.nents = n;
+	pt->end = pfn;
+
+	return 0;
+}
+
+static const npages_fn_t npages_funcs[] = {
+	one,
+	grow,
+	shrink,
+	random,
+	random_page_size_pages,
+	NULL,
+};
+
+static int igt_sg_alloc(void *ignored)
+{
+	IGT_TIMEOUT(end_time);
+	const unsigned long max_order = 20; /* approximating a 4GiB object */
+	struct rnd_state prng;
+	unsigned long prime;
+	int alloc_error = -ENOMEM;
+
+	for_each_prime_number(prime, max_order) {
+		unsigned long size = BIT(prime);
+		int offset;
+
+		for (offset = -1; offset <= 1; offset++) {
+			unsigned long sz = size + offset;
+			const npages_fn_t *npages;
+			struct pfn_table pt;
+			int err;
+
+			for (npages = npages_funcs; *npages; npages++) {
+				prandom_seed_state(&prng,
+						   i915_selftest.random_seed);
+				err = alloc_table(&pt, sz, sz, *npages, &prng,
+						  alloc_error);
+				if (err == -ENOSPC)
+					break;
+				if (err)
+					return err;
+
+				prandom_seed_state(&prng,
+						   i915_selftest.random_seed);
+				err = expect_pfn_sgtable(&pt, *npages, &prng,
+							 "sg_alloc_table",
+							 end_time);
+				sg_free_table(&pt.st);
+				if (err)
+					return err;
+			}
+		}
+
+		/* Test at least one continuation before accepting oom */
+		if (size > SG_MAX_SINGLE_ALLOC)
+			alloc_error = -ENOSPC;
+	}
+
+	return 0;
+}
+
+static int igt_sg_trim(void *ignored)
+{
+	IGT_TIMEOUT(end_time);
+	const unsigned long max = PAGE_SIZE; /* not prime! */
+	struct pfn_table pt;
+	unsigned long prime;
+	int alloc_error = -ENOMEM;
+
+	for_each_prime_number(prime, max) {
+		const npages_fn_t *npages;
+		int err;
+
+		for (npages = npages_funcs; *npages; npages++) {
+			struct rnd_state prng;
+
+			prandom_seed_state(&prng, i915_selftest.random_seed);
+			err = alloc_table(&pt, prime, max, *npages, &prng,
+					  alloc_error);
+			if (err == -ENOSPC)
+				break;
+			if (err)
+				return err;
+
+			if (i915_sg_trim(&pt.st)) {
+				if (pt.st.orig_nents != prime ||
+				    pt.st.nents != prime) {
+					pr_err("i915_sg_trim failed (nents %u, orig_nents %u), expected %lu\n",
+					       pt.st.nents, pt.st.orig_nents, prime);
+					err = -EINVAL;
+				} else {
+					prandom_seed_state(&prng,
+							   i915_selftest.random_seed);
+					err = expect_pfn_sgtable(&pt,
+								 *npages, &prng,
+								 "i915_sg_trim",
+								 end_time);
+				}
+			}
+			sg_free_table(&pt.st);
+			if (err)
+				return err;
+		}
+
+		/* Test at least one continuation before accepting oom */
+		if (prime > SG_MAX_SINGLE_ALLOC)
+			alloc_error = -ENOSPC;
+	}
+
+	return 0;
+}
+
+int scatterlist_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_sg_alloc),
+		SUBTEST(igt_sg_trim),
+	};
+
+	return i915_subtests(tests, NULL);
+}
diff --git a/drivers/gpu/drm/i915/vlv_sideband.c b/drivers/gpu/drm/i915/vlv_sideband.c
new file mode 100644
index 000000000..c26001300
--- /dev/null
+++ b/drivers/gpu/drm/i915/vlv_sideband.c
@@ -0,0 +1,266 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2013-2021 Intel Corporation
+ */
+
+#include "i915_drv.h"
+#include "i915_iosf_mbi.h"
+#include "i915_reg.h"
+#include "vlv_sideband.h"
+
+/*
+ * IOSF sideband, see VLV2_SidebandMsg_HAS.docx and
+ * VLV_VLV2_PUNIT_HAS_0.8.docx
+ */
+
+/* Standard MMIO read, non-posted */
+#define SB_MRD_NP	0x00
+/* Standard MMIO write, non-posted */
+#define SB_MWR_NP	0x01
+/* Private register read, double-word addressing, non-posted */
+#define SB_CRRDDA_NP	0x06
+/* Private register write, double-word addressing, non-posted */
+#define SB_CRWRDA_NP	0x07
+
+static void ping(void *info)
+{
+}
+
+static void __vlv_punit_get(struct drm_i915_private *i915)
+{
+	iosf_mbi_punit_acquire();
+
+	/*
+	 * Prevent the cpu from sleeping while we use this sideband, otherwise
+	 * the punit may cause a machine hang. The issue appears to be isolated
+	 * with changing the power state of the CPU package while changing
+	 * the power state via the punit, and we have only observed it
+	 * reliably on 4-core Baytail systems suggesting the issue is in the
+	 * power delivery mechanism and likely to be board/function
+	 * specific. Hence we presume the workaround needs only be applied
+	 * to the Valleyview P-unit and not all sideband communications.
+	 */
+	if (IS_VALLEYVIEW(i915)) {
+		cpu_latency_qos_update_request(&i915->sb_qos, 0);
+		on_each_cpu(ping, NULL, 1);
+	}
+}
+
+static void __vlv_punit_put(struct drm_i915_private *i915)
+{
+	if (IS_VALLEYVIEW(i915))
+		cpu_latency_qos_update_request(&i915->sb_qos,
+					       PM_QOS_DEFAULT_VALUE);
+
+	iosf_mbi_punit_release();
+}
+
+void vlv_iosf_sb_get(struct drm_i915_private *i915, unsigned long ports)
+{
+	if (ports & BIT(VLV_IOSF_SB_PUNIT))
+		__vlv_punit_get(i915);
+
+	mutex_lock(&i915->sb_lock);
+}
+
+void vlv_iosf_sb_put(struct drm_i915_private *i915, unsigned long ports)
+{
+	mutex_unlock(&i915->sb_lock);
+
+	if (ports & BIT(VLV_IOSF_SB_PUNIT))
+		__vlv_punit_put(i915);
+}
+
+static int vlv_sideband_rw(struct drm_i915_private *i915,
+			   u32 devfn, u32 port, u32 opcode,
+			   u32 addr, u32 *val)
+{
+	struct intel_uncore *uncore = &i915->uncore;
+	const bool is_read = (opcode == SB_MRD_NP || opcode == SB_CRRDDA_NP);
+	int err;
+
+	lockdep_assert_held(&i915->sb_lock);
+	if (port == IOSF_PORT_PUNIT)
+		iosf_mbi_assert_punit_acquired();
+
+	/* Flush the previous comms, just in case it failed last time. */
+	if (intel_wait_for_register(uncore,
+				    VLV_IOSF_DOORBELL_REQ, IOSF_SB_BUSY, 0,
+				    5)) {
+		drm_dbg(&i915->drm, "IOSF sideband idle wait (%s) timed out\n",
+			is_read ? "read" : "write");
+		return -EAGAIN;
+	}
+
+	preempt_disable();
+
+	intel_uncore_write_fw(uncore, VLV_IOSF_ADDR, addr);
+	intel_uncore_write_fw(uncore, VLV_IOSF_DATA, is_read ? 0 : *val);
+	intel_uncore_write_fw(uncore, VLV_IOSF_DOORBELL_REQ,
+			      (devfn << IOSF_DEVFN_SHIFT) |
+			      (opcode << IOSF_OPCODE_SHIFT) |
+			      (port << IOSF_PORT_SHIFT) |
+			      (0xf << IOSF_BYTE_ENABLES_SHIFT) |
+			      (0 << IOSF_BAR_SHIFT) |
+			      IOSF_SB_BUSY);
+
+	if (__intel_wait_for_register_fw(uncore,
+					 VLV_IOSF_DOORBELL_REQ, IOSF_SB_BUSY, 0,
+					 10000, 0, NULL) == 0) {
+		if (is_read)
+			*val = intel_uncore_read_fw(uncore, VLV_IOSF_DATA);
+		err = 0;
+	} else {
+		drm_dbg(&i915->drm, "IOSF sideband finish wait (%s) timed out\n",
+			is_read ? "read" : "write");
+		err = -ETIMEDOUT;
+	}
+
+	preempt_enable();
+
+	return err;
+}
+
+u32 vlv_punit_read(struct drm_i915_private *i915, u32 addr)
+{
+	u32 val = 0;
+
+	vlv_sideband_rw(i915, PCI_DEVFN(0, 0), IOSF_PORT_PUNIT,
+			SB_CRRDDA_NP, addr, &val);
+
+	return val;
+}
+
+int vlv_punit_write(struct drm_i915_private *i915, u32 addr, u32 val)
+{
+	return vlv_sideband_rw(i915, PCI_DEVFN(0, 0), IOSF_PORT_PUNIT,
+			       SB_CRWRDA_NP, addr, &val);
+}
+
+u32 vlv_bunit_read(struct drm_i915_private *i915, u32 reg)
+{
+	u32 val = 0;
+
+	vlv_sideband_rw(i915, PCI_DEVFN(0, 0), IOSF_PORT_BUNIT,
+			SB_CRRDDA_NP, reg, &val);
+
+	return val;
+}
+
+void vlv_bunit_write(struct drm_i915_private *i915, u32 reg, u32 val)
+{
+	vlv_sideband_rw(i915, PCI_DEVFN(0, 0), IOSF_PORT_BUNIT,
+			SB_CRWRDA_NP, reg, &val);
+}
+
+u32 vlv_nc_read(struct drm_i915_private *i915, u8 addr)
+{
+	u32 val = 0;
+
+	vlv_sideband_rw(i915, PCI_DEVFN(0, 0), IOSF_PORT_NC,
+			SB_CRRDDA_NP, addr, &val);
+
+	return val;
+}
+
+u32 vlv_iosf_sb_read(struct drm_i915_private *i915, u8 port, u32 reg)
+{
+	u32 val = 0;
+
+	vlv_sideband_rw(i915, PCI_DEVFN(0, 0), port,
+			SB_CRRDDA_NP, reg, &val);
+
+	return val;
+}
+
+void vlv_iosf_sb_write(struct drm_i915_private *i915,
+		       u8 port, u32 reg, u32 val)
+{
+	vlv_sideband_rw(i915, PCI_DEVFN(0, 0), port,
+			SB_CRWRDA_NP, reg, &val);
+}
+
+u32 vlv_cck_read(struct drm_i915_private *i915, u32 reg)
+{
+	u32 val = 0;
+
+	vlv_sideband_rw(i915, PCI_DEVFN(0, 0), IOSF_PORT_CCK,
+			SB_CRRDDA_NP, reg, &val);
+
+	return val;
+}
+
+void vlv_cck_write(struct drm_i915_private *i915, u32 reg, u32 val)
+{
+	vlv_sideband_rw(i915, PCI_DEVFN(0, 0), IOSF_PORT_CCK,
+			SB_CRWRDA_NP, reg, &val);
+}
+
+u32 vlv_ccu_read(struct drm_i915_private *i915, u32 reg)
+{
+	u32 val = 0;
+
+	vlv_sideband_rw(i915, PCI_DEVFN(0, 0), IOSF_PORT_CCU,
+			SB_CRRDDA_NP, reg, &val);
+
+	return val;
+}
+
+void vlv_ccu_write(struct drm_i915_private *i915, u32 reg, u32 val)
+{
+	vlv_sideband_rw(i915, PCI_DEVFN(0, 0), IOSF_PORT_CCU,
+			SB_CRWRDA_NP, reg, &val);
+}
+
+static u32 vlv_dpio_phy_iosf_port(struct drm_i915_private *i915, enum dpio_phy phy)
+{
+	/*
+	 * IOSF_PORT_DPIO: VLV x2 PHY (DP/HDMI B and C), CHV x1 PHY (DP/HDMI D)
+	 * IOSF_PORT_DPIO_2: CHV x2 PHY (DP/HDMI B and C)
+	 */
+	if (IS_CHERRYVIEW(i915))
+		return phy == DPIO_PHY0 ? IOSF_PORT_DPIO_2 : IOSF_PORT_DPIO;
+	else
+		return IOSF_PORT_DPIO;
+}
+
+u32 vlv_dpio_read(struct drm_i915_private *i915, enum pipe pipe, int reg)
+{
+	u32 port = vlv_dpio_phy_iosf_port(i915, DPIO_PHY(pipe));
+	u32 val = 0;
+
+	vlv_sideband_rw(i915, DPIO_DEVFN, port, SB_MRD_NP, reg, &val);
+
+	/*
+	 * FIXME: There might be some registers where all 1's is a valid value,
+	 * so ideally we should check the register offset instead...
+	 */
+	drm_WARN(&i915->drm, val == 0xffffffff,
+		 "DPIO read pipe %c reg 0x%x == 0x%x\n",
+		 pipe_name(pipe), reg, val);
+
+	return val;
+}
+
+void vlv_dpio_write(struct drm_i915_private *i915,
+		    enum pipe pipe, int reg, u32 val)
+{
+	u32 port = vlv_dpio_phy_iosf_port(i915, DPIO_PHY(pipe));
+
+	vlv_sideband_rw(i915, DPIO_DEVFN, port, SB_MWR_NP, reg, &val);
+}
+
+u32 vlv_flisdsi_read(struct drm_i915_private *i915, u32 reg)
+{
+	u32 val = 0;
+
+	vlv_sideband_rw(i915, DPIO_DEVFN, IOSF_PORT_FLISDSI, SB_CRRDDA_NP,
+			reg, &val);
+	return val;
+}
+
+void vlv_flisdsi_write(struct drm_i915_private *i915, u32 reg, u32 val)
+{
+	vlv_sideband_rw(i915, DPIO_DEVFN, IOSF_PORT_FLISDSI, SB_CRWRDA_NP,
+			reg, &val);
+}
diff --git a/drivers/gpu/drm/i915/vlv_sideband.h b/drivers/gpu/drm/i915/vlv_sideband.h
new file mode 100644
index 000000000..9ce283d96
--- /dev/null
+++ b/drivers/gpu/drm/i915/vlv_sideband.h
@@ -0,0 +1,125 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2013-2021 Intel Corporation
+ */
+
+#ifndef _VLV_SIDEBAND_H_
+#define _VLV_SIDEBAND_H_
+
+#include <linux/bitops.h>
+#include <linux/types.h>
+
+#include "vlv_sideband_reg.h"
+
+enum pipe;
+struct drm_i915_private;
+
+enum {
+	VLV_IOSF_SB_BUNIT,
+	VLV_IOSF_SB_CCK,
+	VLV_IOSF_SB_CCU,
+	VLV_IOSF_SB_DPIO,
+	VLV_IOSF_SB_FLISDSI,
+	VLV_IOSF_SB_GPIO,
+	VLV_IOSF_SB_NC,
+	VLV_IOSF_SB_PUNIT,
+};
+
+void vlv_iosf_sb_get(struct drm_i915_private *i915, unsigned long ports);
+u32 vlv_iosf_sb_read(struct drm_i915_private *i915, u8 port, u32 reg);
+void vlv_iosf_sb_write(struct drm_i915_private *i915,
+		       u8 port, u32 reg, u32 val);
+void vlv_iosf_sb_put(struct drm_i915_private *i915, unsigned long ports);
+
+static inline void vlv_bunit_get(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_get(i915, BIT(VLV_IOSF_SB_BUNIT));
+}
+
+u32 vlv_bunit_read(struct drm_i915_private *i915, u32 reg);
+void vlv_bunit_write(struct drm_i915_private *i915, u32 reg, u32 val);
+
+static inline void vlv_bunit_put(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_put(i915, BIT(VLV_IOSF_SB_BUNIT));
+}
+
+static inline void vlv_cck_get(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_get(i915, BIT(VLV_IOSF_SB_CCK));
+}
+
+u32 vlv_cck_read(struct drm_i915_private *i915, u32 reg);
+void vlv_cck_write(struct drm_i915_private *i915, u32 reg, u32 val);
+
+static inline void vlv_cck_put(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_put(i915, BIT(VLV_IOSF_SB_CCK));
+}
+
+static inline void vlv_ccu_get(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_get(i915, BIT(VLV_IOSF_SB_CCU));
+}
+
+u32 vlv_ccu_read(struct drm_i915_private *i915, u32 reg);
+void vlv_ccu_write(struct drm_i915_private *i915, u32 reg, u32 val);
+
+static inline void vlv_ccu_put(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_put(i915, BIT(VLV_IOSF_SB_CCU));
+}
+
+static inline void vlv_dpio_get(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_get(i915, BIT(VLV_IOSF_SB_DPIO));
+}
+
+u32 vlv_dpio_read(struct drm_i915_private *i915, enum pipe pipe, int reg);
+void vlv_dpio_write(struct drm_i915_private *i915,
+		    enum pipe pipe, int reg, u32 val);
+
+static inline void vlv_dpio_put(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_put(i915, BIT(VLV_IOSF_SB_DPIO));
+}
+
+static inline void vlv_flisdsi_get(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_get(i915, BIT(VLV_IOSF_SB_FLISDSI));
+}
+
+u32 vlv_flisdsi_read(struct drm_i915_private *i915, u32 reg);
+void vlv_flisdsi_write(struct drm_i915_private *i915, u32 reg, u32 val);
+
+static inline void vlv_flisdsi_put(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_put(i915, BIT(VLV_IOSF_SB_FLISDSI));
+}
+
+static inline void vlv_nc_get(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_get(i915, BIT(VLV_IOSF_SB_NC));
+}
+
+u32 vlv_nc_read(struct drm_i915_private *i915, u8 addr);
+
+static inline void vlv_nc_put(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_put(i915, BIT(VLV_IOSF_SB_NC));
+}
+
+static inline void vlv_punit_get(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_get(i915, BIT(VLV_IOSF_SB_PUNIT));
+}
+
+u32 vlv_punit_read(struct drm_i915_private *i915, u32 addr);
+int vlv_punit_write(struct drm_i915_private *i915, u32 addr, u32 val);
+
+static inline void vlv_punit_put(struct drm_i915_private *i915)
+{
+	vlv_iosf_sb_put(i915, BIT(VLV_IOSF_SB_PUNIT));
+}
+
+#endif /* _VLV_SIDEBAND_H_ */
diff --git a/drivers/gpu/drm/i915/vlv_sideband_reg.h b/drivers/gpu/drm/i915/vlv_sideband_reg.h
new file mode 100644
index 000000000..b7fbff3d0
--- /dev/null
+++ b/drivers/gpu/drm/i915/vlv_sideband_reg.h
@@ -0,0 +1,180 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2022 Intel Corporation
+ */
+
+#ifndef _VLV_SIDEBAND_REG_H_
+#define _VLV_SIDEBAND_REG_H_
+
+/* See configdb bunit SB addr map */
+#define BUNIT_REG_BISOC				0x11
+
+/* PUNIT_REG_*SSPM0 */
+#define   _SSPM0_SSC(val)			((val) << 0)
+#define   SSPM0_SSC_MASK			_SSPM0_SSC(0x3)
+#define   SSPM0_SSC_PWR_ON			_SSPM0_SSC(0x0)
+#define   SSPM0_SSC_CLK_GATE			_SSPM0_SSC(0x1)
+#define   SSPM0_SSC_RESET			_SSPM0_SSC(0x2)
+#define   SSPM0_SSC_PWR_GATE			_SSPM0_SSC(0x3)
+#define   _SSPM0_SSS(val)			((val) << 24)
+#define   SSPM0_SSS_MASK			_SSPM0_SSS(0x3)
+#define   SSPM0_SSS_PWR_ON			_SSPM0_SSS(0x0)
+#define   SSPM0_SSS_CLK_GATE			_SSPM0_SSS(0x1)
+#define   SSPM0_SSS_RESET			_SSPM0_SSS(0x2)
+#define   SSPM0_SSS_PWR_GATE			_SSPM0_SSS(0x3)
+
+/* PUNIT_REG_*SSPM1 */
+#define   SSPM1_FREQSTAT_SHIFT			24
+#define   SSPM1_FREQSTAT_MASK			(0x1f << SSPM1_FREQSTAT_SHIFT)
+#define   SSPM1_FREQGUAR_SHIFT			8
+#define   SSPM1_FREQGUAR_MASK			(0x1f << SSPM1_FREQGUAR_SHIFT)
+#define   SSPM1_FREQ_SHIFT			0
+#define   SSPM1_FREQ_MASK			(0x1f << SSPM1_FREQ_SHIFT)
+
+#define PUNIT_REG_VEDSSPM0			0x32
+#define PUNIT_REG_VEDSSPM1			0x33
+
+#define PUNIT_REG_DSPSSPM			0x36
+#define   DSPFREQSTAT_SHIFT_CHV			24
+#define   DSPFREQSTAT_MASK_CHV			(0x1f << DSPFREQSTAT_SHIFT_CHV)
+#define   DSPFREQGUAR_SHIFT_CHV			8
+#define   DSPFREQGUAR_MASK_CHV			(0x1f << DSPFREQGUAR_SHIFT_CHV)
+#define   DSPFREQSTAT_SHIFT			30
+#define   DSPFREQSTAT_MASK			(0x3 << DSPFREQSTAT_SHIFT)
+#define   DSPFREQGUAR_SHIFT			14
+#define   DSPFREQGUAR_MASK			(0x3 << DSPFREQGUAR_SHIFT)
+#define   DSP_MAXFIFO_PM5_STATUS		(1 << 22) /* chv */
+#define   DSP_AUTO_CDCLK_GATE_DISABLE		(1 << 7) /* chv */
+#define   DSP_MAXFIFO_PM5_ENABLE		(1 << 6) /* chv */
+#define   _DP_SSC(val, pipe)			((val) << (2 * (pipe)))
+#define   DP_SSC_MASK(pipe)			_DP_SSC(0x3, (pipe))
+#define   DP_SSC_PWR_ON(pipe)			_DP_SSC(0x0, (pipe))
+#define   DP_SSC_CLK_GATE(pipe)			_DP_SSC(0x1, (pipe))
+#define   DP_SSC_RESET(pipe)			_DP_SSC(0x2, (pipe))
+#define   DP_SSC_PWR_GATE(pipe)			_DP_SSC(0x3, (pipe))
+#define   _DP_SSS(val, pipe)			((val) << (2 * (pipe) + 16))
+#define   DP_SSS_MASK(pipe)			_DP_SSS(0x3, (pipe))
+#define   DP_SSS_PWR_ON(pipe)			_DP_SSS(0x0, (pipe))
+#define   DP_SSS_CLK_GATE(pipe)			_DP_SSS(0x1, (pipe))
+#define   DP_SSS_RESET(pipe)			_DP_SSS(0x2, (pipe))
+#define   DP_SSS_PWR_GATE(pipe)			_DP_SSS(0x3, (pipe))
+
+#define PUNIT_REG_ISPSSPM0			0x39
+#define PUNIT_REG_ISPSSPM1			0x3a
+
+#define PUNIT_REG_PWRGT_CTRL			0x60
+#define PUNIT_REG_PWRGT_STATUS			0x61
+#define   PUNIT_PWRGT_MASK(pw_idx)		(3 << ((pw_idx) * 2))
+#define   PUNIT_PWRGT_PWR_ON(pw_idx)		(0 << ((pw_idx) * 2))
+#define   PUNIT_PWRGT_CLK_GATE(pw_idx)		(1 << ((pw_idx) * 2))
+#define   PUNIT_PWRGT_RESET(pw_idx)		(2 << ((pw_idx) * 2))
+#define   PUNIT_PWRGT_PWR_GATE(pw_idx)		(3 << ((pw_idx) * 2))
+
+#define PUNIT_PWGT_IDX_RENDER			0
+#define PUNIT_PWGT_IDX_MEDIA			1
+#define PUNIT_PWGT_IDX_DISP2D			3
+#define PUNIT_PWGT_IDX_DPIO_CMN_BC		5
+#define PUNIT_PWGT_IDX_DPIO_TX_B_LANES_01	6
+#define PUNIT_PWGT_IDX_DPIO_TX_B_LANES_23	7
+#define PUNIT_PWGT_IDX_DPIO_TX_C_LANES_01	8
+#define PUNIT_PWGT_IDX_DPIO_TX_C_LANES_23	9
+#define PUNIT_PWGT_IDX_DPIO_RX0			10
+#define PUNIT_PWGT_IDX_DPIO_RX1			11
+#define PUNIT_PWGT_IDX_DPIO_CMN_D		12
+
+#define PUNIT_REG_GPU_LFM			0xd3
+#define PUNIT_REG_GPU_FREQ_REQ			0xd4
+#define PUNIT_REG_GPU_FREQ_STS			0xd8
+#define   GPLLENABLE				(1 << 4)
+#define   GENFREQSTATUS				(1 << 0)
+#define PUNIT_REG_MEDIA_TURBO_FREQ_REQ		0xdc
+#define PUNIT_REG_CZ_TIMESTAMP			0xce
+
+#define PUNIT_FUSE_BUS2				0xf6 /* bits 47:40 */
+#define PUNIT_FUSE_BUS1				0xf5 /* bits 55:48 */
+
+#define FB_GFX_FMAX_AT_VMAX_FUSE		0x136
+#define FB_GFX_FREQ_FUSE_MASK			0xff
+#define FB_GFX_FMAX_AT_VMAX_2SS4EU_FUSE_SHIFT	24
+#define FB_GFX_FMAX_AT_VMAX_2SS6EU_FUSE_SHIFT	16
+#define FB_GFX_FMAX_AT_VMAX_2SS8EU_FUSE_SHIFT	8
+
+#define FB_GFX_FMIN_AT_VMIN_FUSE		0x137
+#define FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT		8
+
+#define PUNIT_REG_DDR_SETUP2			0x139
+#define   FORCE_DDR_FREQ_REQ_ACK		(1 << 8)
+#define   FORCE_DDR_LOW_FREQ			(1 << 1)
+#define   FORCE_DDR_HIGH_FREQ			(1 << 0)
+
+#define PUNIT_GPU_STATUS_REG			0xdb
+#define PUNIT_GPU_STATUS_MAX_FREQ_SHIFT	16
+#define PUNIT_GPU_STATUS_MAX_FREQ_MASK		0xff
+#define PUNIT_GPU_STATIS_GFX_MIN_FREQ_SHIFT	8
+#define PUNIT_GPU_STATUS_GFX_MIN_FREQ_MASK	0xff
+
+#define PUNIT_GPU_DUTYCYCLE_REG		0xdf
+#define PUNIT_GPU_DUTYCYCLE_RPE_FREQ_SHIFT	8
+#define PUNIT_GPU_DUTYCYCLE_RPE_FREQ_MASK	0xff
+
+#define IOSF_NC_FB_GFX_FREQ_FUSE		0x1c
+#define   FB_GFX_MAX_FREQ_FUSE_SHIFT		3
+#define   FB_GFX_MAX_FREQ_FUSE_MASK		0x000007f8
+#define   FB_GFX_FGUARANTEED_FREQ_FUSE_SHIFT	11
+#define   FB_GFX_FGUARANTEED_FREQ_FUSE_MASK	0x0007f800
+#define IOSF_NC_FB_GFX_FMAX_FUSE_HI		0x34
+#define   FB_FMAX_VMIN_FREQ_HI_MASK		0x00000007
+#define IOSF_NC_FB_GFX_FMAX_FUSE_LO		0x30
+#define   FB_FMAX_VMIN_FREQ_LO_SHIFT		27
+#define   FB_FMAX_VMIN_FREQ_LO_MASK		0xf8000000
+
+#define VLV_TURBO_SOC_OVERRIDE		0x04
+#define   VLV_OVERRIDE_EN		1
+#define   VLV_SOC_TDP_EN		(1 << 1)
+#define   VLV_BIAS_CPU_125_SOC_875	(6 << 2)
+#define   CHV_BIAS_CPU_50_SOC_50	(3 << 2)
+
+/* vlv2 north clock has */
+#define CCK_FUSE_REG				0x8
+#define  CCK_FUSE_HPLL_FREQ_MASK		0x3
+#define CCK_REG_DSI_PLL_FUSE			0x44
+#define CCK_REG_DSI_PLL_CONTROL			0x48
+#define  DSI_PLL_VCO_EN				(1 << 31)
+#define  DSI_PLL_LDO_GATE			(1 << 30)
+#define  DSI_PLL_P1_POST_DIV_SHIFT		17
+#define  DSI_PLL_P1_POST_DIV_MASK		(0x1ff << 17)
+#define  DSI_PLL_P2_MUX_DSI0_DIV2		(1 << 13)
+#define  DSI_PLL_P3_MUX_DSI1_DIV2		(1 << 12)
+#define  DSI_PLL_MUX_MASK			(3 << 9)
+#define  DSI_PLL_MUX_DSI0_DSIPLL		(0 << 10)
+#define  DSI_PLL_MUX_DSI0_CCK			(1 << 10)
+#define  DSI_PLL_MUX_DSI1_DSIPLL		(0 << 9)
+#define  DSI_PLL_MUX_DSI1_CCK			(1 << 9)
+#define  DSI_PLL_CLK_GATE_MASK			(0xf << 5)
+#define  DSI_PLL_CLK_GATE_DSI0_DSIPLL		(1 << 8)
+#define  DSI_PLL_CLK_GATE_DSI1_DSIPLL		(1 << 7)
+#define  DSI_PLL_CLK_GATE_DSI0_CCK		(1 << 6)
+#define  DSI_PLL_CLK_GATE_DSI1_CCK		(1 << 5)
+#define  DSI_PLL_LOCK				(1 << 0)
+#define CCK_REG_DSI_PLL_DIVIDER			0x4c
+#define  DSI_PLL_LFSR				(1 << 31)
+#define  DSI_PLL_FRACTION_EN			(1 << 30)
+#define  DSI_PLL_FRAC_COUNTER_SHIFT		27
+#define  DSI_PLL_FRAC_COUNTER_MASK		(7 << 27)
+#define  DSI_PLL_USYNC_CNT_SHIFT		18
+#define  DSI_PLL_USYNC_CNT_MASK			(0x1ff << 18)
+#define  DSI_PLL_N1_DIV_SHIFT			16
+#define  DSI_PLL_N1_DIV_MASK			(3 << 16)
+#define  DSI_PLL_M1_DIV_SHIFT			0
+#define  DSI_PLL_M1_DIV_MASK			(0x1ff << 0)
+#define CCK_CZ_CLOCK_CONTROL			0x62
+#define CCK_GPLL_CLOCK_CONTROL			0x67
+#define CCK_DISPLAY_CLOCK_CONTROL		0x6b
+#define CCK_DISPLAY_REF_CLOCK_CONTROL		0x6c
+#define  CCK_TRUNK_FORCE_ON			(1 << 17)
+#define  CCK_TRUNK_FORCE_OFF			(1 << 16)
+#define  CCK_FREQUENCY_STATUS			(0x1f << 8)
+#define  CCK_FREQUENCY_STATUS_SHIFT		8
+#define  CCK_FREQUENCY_VALUES			(0x1f << 0)
+
+#endif /* _VLV_SIDEBAND_REG_H_ */
diff --git a/drivers/gpu/drm/i915/vlv_suspend.c b/drivers/gpu/drm/i915/vlv_suspend.c
new file mode 100644
index 000000000..664fde244
--- /dev/null
+++ b/drivers/gpu/drm/i915/vlv_suspend.c
@@ -0,0 +1,492 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#include <linux/string_helpers.h>
+#include <linux/kernel.h>
+
+#include <drm/drm_print.h>
+
+#include "i915_drv.h"
+#include "i915_reg.h"
+#include "i915_trace.h"
+#include "i915_utils.h"
+#include "intel_pm.h"
+#include "vlv_suspend.h"
+
+#include "gt/intel_gt_regs.h"
+
+struct vlv_s0ix_state {
+	/* GAM */
+	u32 wr_watermark;
+	u32 gfx_prio_ctrl;
+	u32 arb_mode;
+	u32 gfx_pend_tlb0;
+	u32 gfx_pend_tlb1;
+	u32 lra_limits[GEN7_LRA_LIMITS_REG_NUM];
+	u32 media_max_req_count;
+	u32 gfx_max_req_count;
+	u32 render_hwsp;
+	u32 ecochk;
+	u32 bsd_hwsp;
+	u32 blt_hwsp;
+	u32 tlb_rd_addr;
+
+	/* MBC */
+	u32 g3dctl;
+	u32 gsckgctl;
+	u32 mbctl;
+
+	/* GCP */
+	u32 ucgctl1;
+	u32 ucgctl3;
+	u32 rcgctl1;
+	u32 rcgctl2;
+	u32 rstctl;
+	u32 misccpctl;
+
+	/* GPM */
+	u32 gfxpause;
+	u32 rpdeuhwtc;
+	u32 rpdeuc;
+	u32 ecobus;
+	u32 pwrdwnupctl;
+	u32 rp_down_timeout;
+	u32 rp_deucsw;
+	u32 rcubmabdtmr;
+	u32 rcedata;
+	u32 spare2gh;
+
+	/* Display 1 CZ domain */
+	u32 gt_imr;
+	u32 gt_ier;
+	u32 pm_imr;
+	u32 pm_ier;
+	u32 gt_scratch[GEN7_GT_SCRATCH_REG_NUM];
+
+	/* GT SA CZ domain */
+	u32 tilectl;
+	u32 gt_fifoctl;
+	u32 gtlc_wake_ctrl;
+	u32 gtlc_survive;
+	u32 pmwgicz;
+
+	/* Display 2 CZ domain */
+	u32 gu_ctl0;
+	u32 gu_ctl1;
+	u32 pcbr;
+	u32 clock_gate_dis2;
+};
+
+/*
+ * Save all Gunit registers that may be lost after a D3 and a subsequent
+ * S0i[R123] transition. The list of registers needing a save/restore is
+ * defined in the VLV2_S0IXRegs document. This documents marks all Gunit
+ * registers in the following way:
+ * - Driver: saved/restored by the driver
+ * - Punit : saved/restored by the Punit firmware
+ * - No, w/o marking: no need to save/restore, since the register is R/O or
+ *                    used internally by the HW in a way that doesn't depend
+ *                    keeping the content across a suspend/resume.
+ * - Debug : used for debugging
+ *
+ * We save/restore all registers marked with 'Driver', with the following
+ * exceptions:
+ * - Registers out of use, including also registers marked with 'Debug'.
+ *   These have no effect on the driver's operation, so we don't save/restore
+ *   them to reduce the overhead.
+ * - Registers that are fully setup by an initialization function called from
+ *   the resume path. For example many clock gating and RPS/RC6 registers.
+ * - Registers that provide the right functionality with their reset defaults.
+ *
+ * TODO: Except for registers that based on the above 3 criteria can be safely
+ * ignored, we save/restore all others, practically treating the HW context as
+ * a black-box for the driver. Further investigation is needed to reduce the
+ * saved/restored registers even further, by following the same 3 criteria.
+ */
+static void vlv_save_gunit_s0ix_state(struct drm_i915_private *i915)
+{
+	struct vlv_s0ix_state *s = i915->vlv_s0ix_state;
+	struct intel_uncore *uncore = &i915->uncore;
+	int i;
+
+	if (!s)
+		return;
+
+	/* GAM 0x4000-0x4770 */
+	s->wr_watermark = intel_uncore_read(uncore, GEN7_WR_WATERMARK);
+	s->gfx_prio_ctrl = intel_uncore_read(uncore, GEN7_GFX_PRIO_CTRL);
+	s->arb_mode = intel_uncore_read(uncore, ARB_MODE);
+	s->gfx_pend_tlb0 = intel_uncore_read(uncore, GEN7_GFX_PEND_TLB0);
+	s->gfx_pend_tlb1 = intel_uncore_read(uncore, GEN7_GFX_PEND_TLB1);
+
+	for (i = 0; i < ARRAY_SIZE(s->lra_limits); i++)
+		s->lra_limits[i] = intel_uncore_read(uncore, GEN7_LRA_LIMITS(i));
+
+	s->media_max_req_count = intel_uncore_read(uncore, GEN7_MEDIA_MAX_REQ_COUNT);
+	s->gfx_max_req_count = intel_uncore_read(uncore, GEN7_GFX_MAX_REQ_COUNT);
+
+	s->render_hwsp = intel_uncore_read(uncore, RENDER_HWS_PGA_GEN7);
+	s->ecochk = intel_uncore_read(uncore, GAM_ECOCHK);
+	s->bsd_hwsp = intel_uncore_read(uncore, BSD_HWS_PGA_GEN7);
+	s->blt_hwsp = intel_uncore_read(uncore, BLT_HWS_PGA_GEN7);
+
+	s->tlb_rd_addr = intel_uncore_read(uncore, GEN7_TLB_RD_ADDR);
+
+	/* MBC 0x9024-0x91D0, 0x8500 */
+	s->g3dctl = intel_uncore_read(uncore, VLV_G3DCTL);
+	s->gsckgctl = intel_uncore_read(uncore, VLV_GSCKGCTL);
+	s->mbctl = intel_uncore_read(uncore, GEN6_MBCTL);
+
+	/* GCP 0x9400-0x9424, 0x8100-0x810C */
+	s->ucgctl1 = intel_uncore_read(uncore, GEN6_UCGCTL1);
+	s->ucgctl3 = intel_uncore_read(uncore, GEN6_UCGCTL3);
+	s->rcgctl1 = intel_uncore_read(uncore, GEN6_RCGCTL1);
+	s->rcgctl2 = intel_uncore_read(uncore, GEN6_RCGCTL2);
+	s->rstctl = intel_uncore_read(uncore, GEN6_RSTCTL);
+	s->misccpctl = intel_uncore_read(uncore, GEN7_MISCCPCTL);
+
+	/* GPM 0xA000-0xAA84, 0x8000-0x80FC */
+	s->gfxpause = intel_uncore_read(uncore, GEN6_GFXPAUSE);
+	s->rpdeuhwtc = intel_uncore_read(uncore, GEN6_RPDEUHWTC);
+	s->rpdeuc = intel_uncore_read(uncore, GEN6_RPDEUC);
+	s->ecobus = intel_uncore_read(uncore, ECOBUS);
+	s->pwrdwnupctl = intel_uncore_read(uncore, VLV_PWRDWNUPCTL);
+	s->rp_down_timeout = intel_uncore_read(uncore, GEN6_RP_DOWN_TIMEOUT);
+	s->rp_deucsw = intel_uncore_read(uncore, GEN6_RPDEUCSW);
+	s->rcubmabdtmr = intel_uncore_read(uncore, GEN6_RCUBMABDTMR);
+	s->rcedata = intel_uncore_read(uncore, VLV_RCEDATA);
+	s->spare2gh = intel_uncore_read(uncore, VLV_SPAREG2H);
+
+	/* Display CZ domain, 0x4400C-0x4402C, 0x4F000-0x4F11F */
+	s->gt_imr = intel_uncore_read(uncore, GTIMR);
+	s->gt_ier = intel_uncore_read(uncore, GTIER);
+	s->pm_imr = intel_uncore_read(uncore, GEN6_PMIMR);
+	s->pm_ier = intel_uncore_read(uncore, GEN6_PMIER);
+
+	for (i = 0; i < ARRAY_SIZE(s->gt_scratch); i++)
+		s->gt_scratch[i] = intel_uncore_read(uncore, GEN7_GT_SCRATCH(i));
+
+	/* GT SA CZ domain, 0x100000-0x138124 */
+	s->tilectl = intel_uncore_read(uncore, TILECTL);
+	s->gt_fifoctl = intel_uncore_read(uncore, GTFIFOCTL);
+	s->gtlc_wake_ctrl = intel_uncore_read(uncore, VLV_GTLC_WAKE_CTRL);
+	s->gtlc_survive = intel_uncore_read(uncore, VLV_GTLC_SURVIVABILITY_REG);
+	s->pmwgicz = intel_uncore_read(uncore, VLV_PMWGICZ);
+
+	/* Gunit-Display CZ domain, 0x182028-0x1821CF */
+	s->gu_ctl0 = intel_uncore_read(uncore, VLV_GU_CTL0);
+	s->gu_ctl1 = intel_uncore_read(uncore, VLV_GU_CTL1);
+	s->pcbr = intel_uncore_read(uncore, VLV_PCBR);
+	s->clock_gate_dis2 = intel_uncore_read(uncore, VLV_GUNIT_CLOCK_GATE2);
+
+	/*
+	 * Not saving any of:
+	 * DFT,		0x9800-0x9EC0
+	 * SARB,	0xB000-0xB1FC
+	 * GAC,		0x5208-0x524C, 0x14000-0x14C000
+	 * PCI CFG
+	 */
+}
+
+static void vlv_restore_gunit_s0ix_state(struct drm_i915_private *i915)
+{
+	struct vlv_s0ix_state *s = i915->vlv_s0ix_state;
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 val;
+	int i;
+
+	if (!s)
+		return;
+
+	/* GAM 0x4000-0x4770 */
+	intel_uncore_write(uncore, GEN7_WR_WATERMARK, s->wr_watermark);
+	intel_uncore_write(uncore, GEN7_GFX_PRIO_CTRL, s->gfx_prio_ctrl);
+	intel_uncore_write(uncore, ARB_MODE, s->arb_mode | (0xffff << 16));
+	intel_uncore_write(uncore, GEN7_GFX_PEND_TLB0, s->gfx_pend_tlb0);
+	intel_uncore_write(uncore, GEN7_GFX_PEND_TLB1, s->gfx_pend_tlb1);
+
+	for (i = 0; i < ARRAY_SIZE(s->lra_limits); i++)
+		intel_uncore_write(uncore, GEN7_LRA_LIMITS(i), s->lra_limits[i]);
+
+	intel_uncore_write(uncore, GEN7_MEDIA_MAX_REQ_COUNT, s->media_max_req_count);
+	intel_uncore_write(uncore, GEN7_GFX_MAX_REQ_COUNT, s->gfx_max_req_count);
+
+	intel_uncore_write(uncore, RENDER_HWS_PGA_GEN7, s->render_hwsp);
+	intel_uncore_write(uncore, GAM_ECOCHK, s->ecochk);
+	intel_uncore_write(uncore, BSD_HWS_PGA_GEN7, s->bsd_hwsp);
+	intel_uncore_write(uncore, BLT_HWS_PGA_GEN7, s->blt_hwsp);
+
+	intel_uncore_write(uncore, GEN7_TLB_RD_ADDR, s->tlb_rd_addr);
+
+	/* MBC 0x9024-0x91D0, 0x8500 */
+	intel_uncore_write(uncore, VLV_G3DCTL, s->g3dctl);
+	intel_uncore_write(uncore, VLV_GSCKGCTL, s->gsckgctl);
+	intel_uncore_write(uncore, GEN6_MBCTL, s->mbctl);
+
+	/* GCP 0x9400-0x9424, 0x8100-0x810C */
+	intel_uncore_write(uncore, GEN6_UCGCTL1, s->ucgctl1);
+	intel_uncore_write(uncore, GEN6_UCGCTL3, s->ucgctl3);
+	intel_uncore_write(uncore, GEN6_RCGCTL1, s->rcgctl1);
+	intel_uncore_write(uncore, GEN6_RCGCTL2, s->rcgctl2);
+	intel_uncore_write(uncore, GEN6_RSTCTL, s->rstctl);
+	intel_uncore_write(uncore, GEN7_MISCCPCTL, s->misccpctl);
+
+	/* GPM 0xA000-0xAA84, 0x8000-0x80FC */
+	intel_uncore_write(uncore, GEN6_GFXPAUSE, s->gfxpause);
+	intel_uncore_write(uncore, GEN6_RPDEUHWTC, s->rpdeuhwtc);
+	intel_uncore_write(uncore, GEN6_RPDEUC, s->rpdeuc);
+	intel_uncore_write(uncore, ECOBUS, s->ecobus);
+	intel_uncore_write(uncore, VLV_PWRDWNUPCTL, s->pwrdwnupctl);
+	intel_uncore_write(uncore, GEN6_RP_DOWN_TIMEOUT, s->rp_down_timeout);
+	intel_uncore_write(uncore, GEN6_RPDEUCSW, s->rp_deucsw);
+	intel_uncore_write(uncore, GEN6_RCUBMABDTMR, s->rcubmabdtmr);
+	intel_uncore_write(uncore, VLV_RCEDATA, s->rcedata);
+	intel_uncore_write(uncore, VLV_SPAREG2H, s->spare2gh);
+
+	/* Display CZ domain, 0x4400C-0x4402C, 0x4F000-0x4F11F */
+	intel_uncore_write(uncore, GTIMR, s->gt_imr);
+	intel_uncore_write(uncore, GTIER, s->gt_ier);
+	intel_uncore_write(uncore, GEN6_PMIMR, s->pm_imr);
+	intel_uncore_write(uncore, GEN6_PMIER, s->pm_ier);
+
+	for (i = 0; i < ARRAY_SIZE(s->gt_scratch); i++)
+		intel_uncore_write(uncore, GEN7_GT_SCRATCH(i), s->gt_scratch[i]);
+
+	/* GT SA CZ domain, 0x100000-0x138124 */
+	intel_uncore_write(uncore, TILECTL, s->tilectl);
+	intel_uncore_write(uncore, GTFIFOCTL, s->gt_fifoctl);
+	/*
+	 * Preserve the GT allow wake and GFX force clock bit, they are not
+	 * be restored, as they are used to control the s0ix suspend/resume
+	 * sequence by the caller.
+	 */
+	val = intel_uncore_read(uncore, VLV_GTLC_WAKE_CTRL);
+	val &= VLV_GTLC_ALLOWWAKEREQ;
+	val |= s->gtlc_wake_ctrl & ~VLV_GTLC_ALLOWWAKEREQ;
+	intel_uncore_write(uncore, VLV_GTLC_WAKE_CTRL, val);
+
+	val = intel_uncore_read(uncore, VLV_GTLC_SURVIVABILITY_REG);
+	val &= VLV_GFX_CLK_FORCE_ON_BIT;
+	val |= s->gtlc_survive & ~VLV_GFX_CLK_FORCE_ON_BIT;
+	intel_uncore_write(uncore, VLV_GTLC_SURVIVABILITY_REG, val);
+
+	intel_uncore_write(uncore, VLV_PMWGICZ, s->pmwgicz);
+
+	/* Gunit-Display CZ domain, 0x182028-0x1821CF */
+	intel_uncore_write(uncore, VLV_GU_CTL0, s->gu_ctl0);
+	intel_uncore_write(uncore, VLV_GU_CTL1, s->gu_ctl1);
+	intel_uncore_write(uncore, VLV_PCBR, s->pcbr);
+	intel_uncore_write(uncore, VLV_GUNIT_CLOCK_GATE2, s->clock_gate_dis2);
+}
+
+static int vlv_wait_for_pw_status(struct drm_i915_private *i915,
+				  u32 mask, u32 val)
+{
+	i915_reg_t reg = VLV_GTLC_PW_STATUS;
+	u32 reg_value;
+	int ret;
+
+	/* The HW does not like us polling for PW_STATUS frequently, so
+	 * use the sleeping loop rather than risk the busy spin within
+	 * intel_wait_for_register().
+	 *
+	 * Transitioning between RC6 states should be at most 2ms (see
+	 * valleyview_enable_rps) so use a 3ms timeout.
+	 */
+	ret = wait_for(((reg_value =
+			 intel_uncore_read_notrace(&i915->uncore, reg)) & mask)
+		       == val, 3);
+
+	/* just trace the final value */
+	trace_i915_reg_rw(false, reg, reg_value, sizeof(reg_value), true);
+
+	return ret;
+}
+
+static int vlv_force_gfx_clock(struct drm_i915_private *i915, bool force_on)
+{
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 val;
+	int err;
+
+	val = intel_uncore_read(uncore, VLV_GTLC_SURVIVABILITY_REG);
+	val &= ~VLV_GFX_CLK_FORCE_ON_BIT;
+	if (force_on)
+		val |= VLV_GFX_CLK_FORCE_ON_BIT;
+	intel_uncore_write(uncore, VLV_GTLC_SURVIVABILITY_REG, val);
+
+	if (!force_on)
+		return 0;
+
+	err = intel_wait_for_register(uncore,
+				      VLV_GTLC_SURVIVABILITY_REG,
+				      VLV_GFX_CLK_STATUS_BIT,
+				      VLV_GFX_CLK_STATUS_BIT,
+				      20);
+	if (err)
+		drm_err(&i915->drm,
+			"timeout waiting for GFX clock force-on (%08x)\n",
+			intel_uncore_read(uncore, VLV_GTLC_SURVIVABILITY_REG));
+
+	return err;
+}
+
+static int vlv_allow_gt_wake(struct drm_i915_private *i915, bool allow)
+{
+	struct intel_uncore *uncore = &i915->uncore;
+	u32 mask;
+	u32 val;
+	int err;
+
+	val = intel_uncore_read(uncore, VLV_GTLC_WAKE_CTRL);
+	val &= ~VLV_GTLC_ALLOWWAKEREQ;
+	if (allow)
+		val |= VLV_GTLC_ALLOWWAKEREQ;
+	intel_uncore_write(uncore, VLV_GTLC_WAKE_CTRL, val);
+	intel_uncore_posting_read(uncore, VLV_GTLC_WAKE_CTRL);
+
+	mask = VLV_GTLC_ALLOWWAKEACK;
+	val = allow ? mask : 0;
+
+	err = vlv_wait_for_pw_status(i915, mask, val);
+	if (err)
+		drm_err(&i915->drm, "timeout disabling GT waking\n");
+
+	return err;
+}
+
+static void vlv_wait_for_gt_wells(struct drm_i915_private *dev_priv,
+				  bool wait_for_on)
+{
+	u32 mask;
+	u32 val;
+
+	mask = VLV_GTLC_PW_MEDIA_STATUS_MASK | VLV_GTLC_PW_RENDER_STATUS_MASK;
+	val = wait_for_on ? mask : 0;
+
+	/*
+	 * RC6 transitioning can be delayed up to 2 msec (see
+	 * valleyview_enable_rps), use 3 msec for safety.
+	 *
+	 * This can fail to turn off the rc6 if the GPU is stuck after a failed
+	 * reset and we are trying to force the machine to sleep.
+	 */
+	if (vlv_wait_for_pw_status(dev_priv, mask, val))
+		drm_dbg(&dev_priv->drm,
+			"timeout waiting for GT wells to go %s\n",
+			str_on_off(wait_for_on));
+}
+
+static void vlv_check_no_gt_access(struct drm_i915_private *i915)
+{
+	struct intel_uncore *uncore = &i915->uncore;
+
+	if (!(intel_uncore_read(uncore, VLV_GTLC_PW_STATUS) & VLV_GTLC_ALLOWWAKEERR))
+		return;
+
+	drm_dbg(&i915->drm, "GT register access while GT waking disabled\n");
+	intel_uncore_write(uncore, VLV_GTLC_PW_STATUS, VLV_GTLC_ALLOWWAKEERR);
+}
+
+int vlv_suspend_complete(struct drm_i915_private *dev_priv)
+{
+	u32 mask;
+	int err;
+
+	if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv))
+		return 0;
+
+	/*
+	 * Bspec defines the following GT well on flags as debug only, so
+	 * don't treat them as hard failures.
+	 */
+	vlv_wait_for_gt_wells(dev_priv, false);
+
+	mask = VLV_GTLC_RENDER_CTX_EXISTS | VLV_GTLC_MEDIA_CTX_EXISTS;
+	drm_WARN_ON(&dev_priv->drm,
+		    (intel_uncore_read(&dev_priv->uncore, VLV_GTLC_WAKE_CTRL) & mask) != mask);
+
+	vlv_check_no_gt_access(dev_priv);
+
+	err = vlv_force_gfx_clock(dev_priv, true);
+	if (err)
+		goto err1;
+
+	err = vlv_allow_gt_wake(dev_priv, false);
+	if (err)
+		goto err2;
+
+	vlv_save_gunit_s0ix_state(dev_priv);
+
+	err = vlv_force_gfx_clock(dev_priv, false);
+	if (err)
+		goto err2;
+
+	return 0;
+
+err2:
+	/* For safety always re-enable waking and disable gfx clock forcing */
+	vlv_allow_gt_wake(dev_priv, true);
+err1:
+	vlv_force_gfx_clock(dev_priv, false);
+
+	return err;
+}
+
+int vlv_resume_prepare(struct drm_i915_private *dev_priv, bool rpm_resume)
+{
+	int err;
+	int ret;
+
+	if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv))
+		return 0;
+
+	/*
+	 * If any of the steps fail just try to continue, that's the best we
+	 * can do at this point. Return the first error code (which will also
+	 * leave RPM permanently disabled).
+	 */
+	ret = vlv_force_gfx_clock(dev_priv, true);
+
+	vlv_restore_gunit_s0ix_state(dev_priv);
+
+	err = vlv_allow_gt_wake(dev_priv, true);
+	if (!ret)
+		ret = err;
+
+	err = vlv_force_gfx_clock(dev_priv, false);
+	if (!ret)
+		ret = err;
+
+	vlv_check_no_gt_access(dev_priv);
+
+	if (rpm_resume)
+		intel_init_clock_gating(dev_priv);
+
+	return ret;
+}
+
+int vlv_suspend_init(struct drm_i915_private *i915)
+{
+	if (!IS_VALLEYVIEW(i915))
+		return 0;
+
+	/* we write all the values in the struct, so no need to zero it out */
+	i915->vlv_s0ix_state = kmalloc(sizeof(*i915->vlv_s0ix_state),
+				       GFP_KERNEL);
+	if (!i915->vlv_s0ix_state)
+		return -ENOMEM;
+
+	return 0;
+}
+
+void vlv_suspend_cleanup(struct drm_i915_private *i915)
+{
+	if (!i915->vlv_s0ix_state)
+		return;
+
+	kfree(i915->vlv_s0ix_state);
+	i915->vlv_s0ix_state = NULL;
+}
diff --git a/drivers/gpu/drm/i915/vlv_suspend.h b/drivers/gpu/drm/i915/vlv_suspend.h
new file mode 100644
index 000000000..895091cb1
--- /dev/null
+++ b/drivers/gpu/drm/i915/vlv_suspend.h
@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2020 Intel Corporation
+ */
+
+#ifndef __VLV_SUSPEND_H__
+#define __VLV_SUSPEND_H__
+
+#include <linux/types.h>
+
+struct drm_i915_private;
+
+int vlv_suspend_init(struct drm_i915_private *i915);
+void vlv_suspend_cleanup(struct drm_i915_private *i915);
+int vlv_suspend_complete(struct drm_i915_private *i915);
+int vlv_resume_prepare(struct drm_i915_private *i915, bool rpm_resume);
+
+#endif /* __VLV_SUSPEND_H__ */